diff --git a/latency-aware/deadly-corridor/small-policy/memory_rollout_teacher:deadly_corridor:teacher:fixed_l6:fs4:obs8p75:stride4:seed0/train/memory_rollout_teacher:deadly_corridor:teacher:fixed_l6:fs4:obs8p75:stride4:seed0/checkpoint_p0/checkpoint_000048836_25004032.pth b/latency-aware/deadly-corridor/small-policy/memory_rollout_teacher:deadly_corridor:teacher:fixed_l6:fs4:obs8p75:stride4:seed0/train/memory_rollout_teacher:deadly_corridor:teacher:fixed_l6:fs4:obs8p75:stride4:seed0/checkpoint_p0/checkpoint_000048836_25004032.pth new file mode 100644 index 0000000000000000000000000000000000000000..9501f0104961c1d513df0a69269fe7be504304e7 --- /dev/null +++ b/latency-aware/deadly-corridor/small-policy/memory_rollout_teacher:deadly_corridor:teacher:fixed_l6:fs4:obs8p75:stride4:seed0/train/memory_rollout_teacher:deadly_corridor:teacher:fixed_l6:fs4:obs8p75:stride4:seed0/checkpoint_p0/checkpoint_000048836_25004032.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6786d84ce663078e7cf02e7032035c4d899843fe06ca3d6d07e666c8d606644f +size 13352656 diff --git a/latency-aware/deadly-corridor/small-policy/memory_rollout_teacher:deadly_corridor:teacher:fixed_l6:fs4:obs8p75:stride4:seed0/train/memory_rollout_teacher:deadly_corridor:teacher:fixed_l6:fs4:obs8p75:stride4:seed0/config.json b/latency-aware/deadly-corridor/small-policy/memory_rollout_teacher:deadly_corridor:teacher:fixed_l6:fs4:obs8p75:stride4:seed0/train/memory_rollout_teacher:deadly_corridor:teacher:fixed_l6:fs4:obs8p75:stride4:seed0/config.json new file mode 100644 index 0000000000000000000000000000000000000000..a9038572b28a71e0b6bcd969f49c813d690603ee --- /dev/null +++ b/latency-aware/deadly-corridor/small-policy/memory_rollout_teacher:deadly_corridor:teacher:fixed_l6:fs4:obs8p75:stride4:seed0/train/memory_rollout_teacher:deadly_corridor:teacher:fixed_l6:fs4:obs8p75:stride4:seed0/config.json @@ -0,0 +1,260 @@ +{ + "help": false, + "algo": "APPO", + "env": "doom_deadly_corridor", + "experiment": "deadly_corridor_teacher_fixed_l6_obs8p75_seed0", + "train_dir": "results/checkpoints/memory/rollout_teacher", + "restart_behavior": "resume", + "device": "gpu", + "seed": 0, + "num_policies": 1, + "async_rl": true, + "serial_mode": false, + "batched_sampling": false, + "num_batches_to_accumulate": 2, + "worker_num_splits": 2, + "policy_workers_per_policy": 1, + "max_policy_lag": 1000, + "num_workers": 32, + "num_envs_per_worker": 4, + "batch_size": 1024, + "num_batches_per_epoch": 2, + "num_epochs": 2, + "rollout": 128, + "recurrence": 128, + "shuffle_minibatches": false, + "gamma": 0.99, + "reward_scale": 1.0, + "reward_clip": 1000.0, + "value_bootstrap": false, + "normalize_returns": true, + "exploration_loss_coeff": 0.001, + "value_loss_coeff": 0.5, + "kl_loss_coeff": 0.0, + "exploration_loss": "symmetric_kl", + "gae_lambda": 0.95, + "ppo_clip_ratio": 0.1, + "ppo_clip_value": 0.2, + "with_vtrace": false, + "vtrace_rho": 1.0, + "vtrace_c": 1.0, + "optimizer": "adam", + "adam_eps": 1e-06, + "adam_beta1": 0.9, + "adam_beta2": 0.999, + "max_grad_norm": 4.0, + "learning_rate": 0.0001, + "lr_schedule": "constant", + "lr_schedule_kl_threshold": 0.008, + "lr_adaptive_min": 1e-06, + "lr_adaptive_max": 0.01, + "obs_subtract_mean": 0.0, + "obs_scale": 255.0, + "normalize_input": true, + "normalize_input_keys": null, + "decorrelate_experience_max_seconds": 0, + "decorrelate_envs_on_one_worker": true, + "actor_worker_gpus": [], + "set_workers_cpu_affinity": true, + "force_envs_single_thread": false, + "default_niceness": 0, + "log_to_file": true, + "experiment_summaries_interval": 1, + "flush_summaries_interval": 30, + "stats_avg": 100, + "summaries_use_frameskip": true, + "heartbeat_interval": 20, + "heartbeat_reporting_interval": 600, + "train_for_env_steps": 25000000, + "train_for_seconds": 10000000000, + "save_every_sec": 600, + "keep_checkpoints": 5, + "load_checkpoint_kind": "latest", + "save_milestones_sec": -1, + "save_best_every_sec": 5, + "save_best_metric": "reward", + "save_best_after": 100000, + "benchmark": false, + "encoder_mlp_layers": [ + 512, + 512 + ], + "encoder_conv_architecture": "convnet_simple", + "encoder_conv_mlp_layers": [ + 512 + ], + "use_rnn": true, + "rnn_size": 512, + "rnn_type": "gru", + "rnn_num_layers": 1, + "decoder_mlp_layers": [], + "nonlinearity": "elu", + "policy_initialization": "orthogonal", + "policy_init_gain": 1.0, + "actor_critic_share_weights": true, + "adaptive_stddev": true, + "continuous_tanh_scale": 0.0, + "initial_stddev": 1.0, + "use_env_info_cache": false, + "env_gpu_actions": false, + "env_gpu_observations": true, + "env_frameskip": 1, + "env_framestack": 1, + "pixel_format": "CHW", + "use_record_episode_statistics": false, + "with_wandb": true, + "wandb_user": null, + "wandb_project": "latency-sensitive-bench", + "wandb_group": "deadly-corridor-obs8p75-fixed_l6", + "wandb_job_type": "sample_factory", + "wandb_tags": [ + "memory_rollout_teacher", + "deadly-corridor", + "teacher", + "fixed", + "fixed_l6", + "obs8p75", + "seed0" + ], + "with_pbt": false, + "pbt_mix_policies_in_one_env": true, + "pbt_period_env_steps": 5000000, + "pbt_start_mutation": 20000000, + "pbt_replace_fraction": 0.3, + "pbt_mutation_rate": 0.15, + "pbt_replace_reward_gap": 0.1, + "pbt_replace_reward_gap_absolute": 1e-06, + "pbt_optimize_gamma": false, + "pbt_target_objective": "true_objective", + "pbt_perturb_min": 1.1, + "pbt_perturb_max": 1.5, + "fps": 35, + "eval_env_frameskip": 1, + "no_render": false, + "save_video": false, + "video_frames": 1000000000.0, + "video_name": null, + "max_num_frames": 1000000000.0, + "max_num_episodes": 1000000000.0, + "push_to_hub": false, + "hf_repository": null, + "policy_index": 0, + "eval_deterministic": true, + "train_script": null, + "enjoy_script": null, + "num_agents": -1, + "num_humans": 0, + "num_bots": -1, + "start_bot_difficulty": null, + "timelimit": null, + "res_w": 128, + "res_h": 72, + "wide_aspect_ratio": false, + "record_to": null, + "mode": "train", + "env_fps": 35.0, + "obs_fps": 8.75, + "latency_type": "fixed", + "fixed_latency_ms": 171.42857142857144, + "mean_latency_ms": null, + "std_latency_ms": null, + "min_latency_ms": null, + "max_latency_ms": null, + "latency_seed": null, + "add_latency_info": false, + "max_pending_actions": null, + "hold_policy": "hold", + "ordering_policy": "latest_ready", + "eval_episodes": 100, + "eval_parallel_envs": 1, + "eval_latency_raw_frame_values": "0,1,2,3,4,5,10,15,40", + "eval_max_steps": 30000, + "eval_raw_reward": true, + "disable_reward_scaling": true, + "frame_stack": 4, + "max_episode_steps": 0, + "simulator": "cpu", + "episode_metrics_path": "results/checkpoints/memory/rollout_teacher/deadly_corridor_teacher_fixed_l6_obs8p75_seed0/episode_metrics.jsonl", + "vizdoom_init_max_parallel": 10, + "vizdoom_init_lock_timeout_sec": 5.0, + "vizdoom_init_lock_max_wait_sec": 60.0, + "vizdoom_init_retry_jitter_sec": 0.25, + "command_line": "--mode train --algo APPO --env doom_deadly_corridor --experiment deadly_corridor_teacher_fixed_l6_obs8p75_seed0 --train_dir results/checkpoints/memory/rollout_teacher --restart_behavior resume --device gpu --seed 0 --episode_metrics_path results/checkpoints/memory/rollout_teacher/deadly_corridor_teacher_fixed_l6_obs8p75_seed0/episode_metrics.jsonl --train_for_env_steps 25000000 --num_workers 32 --num_envs_per_worker 4 --num_policies 1 --batch_size 1024 --rollout 128 --recurrence 128 --num_epochs 2 --num_batches_per_epoch 2 --worker_num_splits 2 --learning_rate 0.0001 --gamma 0.99 --gae_lambda 0.95 --ppo_clip_ratio 0.1 --ppo_clip_value 0.2 --exploration_loss symmetric_kl --exploration_loss_coeff 0.001 --value_loss_coeff 0.5 --max_grad_norm 4.0 --rnn_type gru --rnn_size 512 --save_every_sec 600 --keep_checkpoints 5 --stats_avg 100 --experiment_summaries_interval 1 --async_rl True --use_rnn True --normalize_returns True --normalize_input True --latency-type fixed --fixed-latency-ms 171.42857142857144 --add-latency-info False --eval-episodes 100 --eval-parallel-envs 1 --eval-max-steps 30000 --eval-deterministic True --eval-raw-reward --with_wandb True --wandb_project latency-sensitive-bench --wandb_group deadly-corridor-obs8p75-fixed_l6 --wandb_job_type sample_factory --wandb_tags memory_rollout_teacher deadly-corridor teacher fixed fixed_l6 obs8p75 seed0 --env-fps 35 --obs-fps 8.75 --res_w 128 --res_h 72 --wide_aspect_ratio False --frame_stack 4 --simulator cpu --disable-reward-scaling True --hold-policy hold --ordering-policy latest_ready", + "cli_args": { + "algo": "APPO", + "env": "doom_deadly_corridor", + "experiment": "deadly_corridor_teacher_fixed_l6_obs8p75_seed0", + "train_dir": "results/checkpoints/memory/rollout_teacher", + "restart_behavior": "resume", + "device": "gpu", + "seed": 0, + "num_policies": 1, + "async_rl": true, + "worker_num_splits": 2, + "num_workers": 32, + "num_envs_per_worker": 4, + "batch_size": 1024, + "num_batches_per_epoch": 2, + "num_epochs": 2, + "rollout": 128, + "recurrence": 128, + "gamma": 0.99, + "normalize_returns": true, + "exploration_loss_coeff": 0.001, + "value_loss_coeff": 0.5, + "exploration_loss": "symmetric_kl", + "gae_lambda": 0.95, + "ppo_clip_ratio": 0.1, + "ppo_clip_value": 0.2, + "max_grad_norm": 4.0, + "learning_rate": 0.0001, + "normalize_input": true, + "experiment_summaries_interval": 1, + "stats_avg": 100, + "train_for_env_steps": 25000000, + "save_every_sec": 600, + "keep_checkpoints": 5, + "use_rnn": true, + "rnn_size": 512, + "rnn_type": "gru", + "with_wandb": true, + "wandb_project": "latency-sensitive-bench", + "wandb_group": "deadly-corridor-obs8p75-fixed_l6", + "wandb_job_type": "sample_factory", + "wandb_tags": [ + "memory_rollout_teacher", + "deadly-corridor", + "teacher", + "fixed", + "fixed_l6", + "obs8p75", + "seed0" + ], + "eval_deterministic": true, + "res_w": 128, + "res_h": 72, + "wide_aspect_ratio": false, + "mode": "train", + "env_fps": 35.0, + "obs_fps": 8.75, + "latency_type": "fixed", + "fixed_latency_ms": 171.42857142857144, + "add_latency_info": false, + "hold_policy": "hold", + "ordering_policy": "latest_ready", + "eval_episodes": 100, + "eval_parallel_envs": 1, + "eval_max_steps": 30000, + "eval_raw_reward": true, + "disable_reward_scaling": true, + "frame_stack": 4, + "simulator": "cpu", + "episode_metrics_path": "results/checkpoints/memory/rollout_teacher/deadly_corridor_teacher_fixed_l6_obs8p75_seed0/episode_metrics.jsonl", + "env_frameskip": 1, + "eval_env_frameskip": 1 + }, + "git_hash": "bedb62f3a97309f6dacdb6a2953e3312d9776765", + "git_repo_name": "git@github.com:ZihanWang314/latency-sensitive-bench.git", + "output_dir": "outputs/memory/rollout_teacher/deadly_corridor/train/teacher/fixed_l6/obs8p75/seed_0", + "wandb_unique_id": "deadly-corridor-obs8p75-fixed_l6-s0" +} \ No newline at end of file diff --git a/latency-aware/deadly-corridor/vla/starvla-wanoft-h8/wan_oft_deadly_corridor_fix_latency_6_context5_standard_sft_1000_effbs128_224_currentbce/checkpoints/steps_1000_pytorch_model.pt b/latency-aware/deadly-corridor/vla/starvla-wanoft-h8/wan_oft_deadly_corridor_fix_latency_6_context5_standard_sft_1000_effbs128_224_currentbce/checkpoints/steps_1000_pytorch_model.pt new file mode 100644 index 0000000000000000000000000000000000000000..13408b13d0a32ee8405cc88e1013373ec3af157e --- /dev/null +++ b/latency-aware/deadly-corridor/vla/starvla-wanoft-h8/wan_oft_deadly_corridor_fix_latency_6_context5_standard_sft_1000_effbs128_224_currentbce/checkpoints/steps_1000_pytorch_model.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a8bd1415990de5a8616a203b15d58f105dbcc57e5ac10ffff0ea70e2220d9c5c +size 25211531190 diff --git a/latency-aware/deadly-corridor/vla/starvla-wanoft-h8/wan_oft_deadly_corridor_fix_latency_6_context5_standard_sft_1000_effbs128_224_currentbce/config.full.yaml b/latency-aware/deadly-corridor/vla/starvla-wanoft-h8/wan_oft_deadly_corridor_fix_latency_6_context5_standard_sft_1000_effbs128_224_currentbce/config.full.yaml new file mode 100644 index 0000000000000000000000000000000000000000..0234b05f697c1a7f5c1d010a9fa6815789ed3d9e --- /dev/null +++ b/latency-aware/deadly-corridor/vla/starvla-wanoft-h8/wan_oft_deadly_corridor_fix_latency_6_context5_standard_sft_1000_effbs128_224_currentbce/config.full.yaml @@ -0,0 +1,286 @@ +framework: + name: WanOFT + world_model: + base_wm: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/wanoft_bench/third_party/starVLA/playground/Pretrained_models/Wan-AI/Wan2.2-TI2V-5B-Diffusers + extract_layers: + - -1 + num_frames: 5 + qwenvl: + base_vlm: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/wanoft_bench/third_party/starVLA/playground/Pretrained_models/Wan-AI/Wan2.2-TI2V-5B-Diffusers + attn_implementation: sdpa + flex_backend: triton + enable_gradient_checkpointing: true + action_model: + action_model_type: MLP + action_dim: 7 + action_hidden_dim: 3072 + action_horizon: 8 + future_action_window_size: 7 + past_action_window_size: 0 + loss_type: current_multibinary_bce + class_weights: null + future_loss_weight: null + action_query_source: mean + action_query_num_heads: 24 + state_dim: 7 + action_env_dim: 7 + kv_memory: + enabled: false + window: 4 + rollout_len: 8 + packed_train: false + rebased_sink: true +datasets: + vla_data: + dataset_py: lerobot_datasets + include_state: true + data_root_dir: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/wanoft_bench/third_party/starVLA/data/deadly_corridor_fix_latency_6_1000ep_context5 + data_mix: deadly_corridor_train__bridge + eval_data_mix: null + custom_mixtures_path: null + action_type: discrete + sequential_step_sampling: false + eval_sequential_step_sampling: null + num_workers: 8 + eval_num_workers: 8 + prefetch_factor: 4 + persistent_workers: true + pin_memory: true + shuffle: true + action_balance: + enabled: false + strategy: balanced_epoch + action_key: action_id + target_flap_fraction: 0.3 + noop_id: 0 + flap_id: 1 + latency_curriculum: + enabled: false + strategy: exclusive + latencies: null + phase_steps: null + phase_distributions: null + new_latency_passes: 1.0 + replay_passes: 0.25 + target_total_passes: 2.0 + final_equalization: true + step_budget_mode: auto + eval_at_phase_end: false + save_at_phase_end: false + computed_plan: null + per_device_batch_size: 1 + load_all_data_for_training: true + num_obs_frames: 1 + image_mode: single + prompt_mode: raw + stitch_grid: + - 2 + - 2 + obs_image_size: + - 224 + - 224 + video_backend: torchvision_av + observation_indices: + - -4 + - -3 + - -2 + - -1 + - 0 + language_indices: + - 0 + state_indices: + - 0 + action_indices: + - 0 + - 1 + - 2 + - 3 + - 4 + - 5 + - 6 + - 7 + pack_image_sequence: true + image_sequence_length: 5 + context_images_column: observation.context_images +dataset: + source_hf: latency-sensitive-bench/memory-data + config_name: deadly_corridor_fixed_latency_6_1000ep_7k2steps + source_subdir: null + converted_name: deadly_corridor_train__bridge + single_source_hf: '' + mixed_source_hf: '' + single_converted_name: deadly_corridor_train__bridge + mixed_converted_name: deadly_corridor_mixed_latency_train__bridge + single_latency_filter: null + mixed_latency_filter: null + force_download: false + setup_force: false + skip_verification: false + target_latency_unit: observation_steps + verify_rows: 200 + max_episodes: 1000 + episodes_per_latency: null + latency_filter: null + debug_subset: + enabled: false + max_episodes: 5 + suffix: debug +base_model: + repo_id: Wan-AI/Wan2.2-TI2V-5B-Diffusers +initialization: + checkpoint_local_dir: playground/Pretrained_models/WM4A-Wan2d2-OFT-LIBERO-4in1 + checkpoint_hf_repo_id: StarVLA/WM4A-Wan2d2-OFT-LIBERO-4in1 + checkpoint_filename: checkpoints/steps_60000_pytorch_model.pt +trainer: + max_train_steps: 1000 + num_warmup_steps: 0 + save_interval: 500 + eval_interval: 100 + eval_num_batches: 100 + per_latency_eval_num_batches: null + eval_action_classification: false + eval_action_classification_interval: null + cc_f1_tolerance: 1 + learning_rate: + base: 2.5e-05 + qwen_vl_interface: 1.0e-05 + action_model: 0.0001 + action_query_proj: 0.0001 + lr_scheduler_type: cosine_with_min_lr + scheduler_specific_kwargs: + min_lr: 1.0e-06 + freeze_modules: '' + freeze_vit: false + freeze_tied_embedding: false + freeze_llm_layers: [] + loss_scale: + vla: 1.0 + vlm: 0.1 + max_grad_norm: 1.0 + weight_decay: 0.0 + logging_frequency: 1 + profile_timing: + enabled: false + log_interval: 10 + gradient_clipping: 1.0 + gradient_accumulation_steps: 16 + distributed_backend: deepspeed + is_resume: false + pretrained_checkpoint: null + resume_step: 0 + reload_modules: null + optimizer: + name: AdamW + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + weight_decay: 1.0e-08 + fused: true + save_format: pt +workspace_dir: WORKSPACE_DIR +run_root_dir: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/wanoft_bench/third_party/starVLA/results/Checkpoints +seed: 42 +wandb_entity: zihanwang-ai-northwestern-university +wandb_project: starVLA_rl_games +auth: + env_file: null + hf_token_env: HF_TOKEN + wandb_api_key_env: WANDB_API_KEY +paths: + run_root_dir: results/Checkpoints + dataset_local_dir: data/deadly_corridor_fix_latency_6_1000ep_context5 + dataset_cache_dir: null + base_model_dir: playground/Pretrained_models/Wan-AI/Wan2.2-TI2V-5B-Diffusers + accelerate_config: starVLA/config/deepseeds/deepspeed_zero2.yaml +launch: + use_accelerate: true + gpus: null + num_processes: 1 + dry_run: false +conda: + enabled: true + env_name: null +rl_games: + model_alias: wan_oft + env_eval: + image_size: 224 + frameskip: 4 + image_transform: raw_rgb + prompt_mode: raw + ghost_trail: + history_frames: 5 + gamma: 1.3 + min_alpha: 35 + scroll_px_per_step: 4.0 + ground_fraction: 0.22 + seed: 42 + fixed_episode_seeds: true + latency_seed_stride: 0 + task_seed_stride: 0 + task_description: You are playing Deadly Corridor in VizDoom. Choose actions from + MOVE_FORWARD, MOVE_BACKWARD, MOVE_LEFT, MOVE_RIGHT, TURN_LEFT, TURN_RIGHT, ATTACK. + eval_parallel_envs: 5 + action_chunk_execution: + enabled: false + chunk_size: null + deadly: + action_layout: multibinary_7 + multibinary_threshold: 0.0 + enabled: false + eval_backend: eval_core + distributed_mode: rank_sharded + vectorized: + enabled: false + batch_size: 1 + latency: + prompt_map_path: data/deadly_corridor_fix_latency_6_1000ep_context5/deadly_corridor_train__bridge/latency_prompt_map.json + mode: single + values: + - 0 + mid_train: + enabled: false + interval_steps: 100 + latencies: + - 0 + num_episodes: 5 + max_steps_per_episode: 3600 + post_train: + enabled: false + latencies: + - 0 + - 1 + - 2 + - 3 + - 4 + num_episodes: 5 + max_steps_per_episode: 3600 + task: deadly_corridor + deadly_corridor_loss_type: current_multibinary_bce + initialization_mode: bridge + action_carrier: bridge +model: wan_oft +env: deadly_corridor +init: wan_oft_libero +mode: single +checkpoint: + load: auto + hf_repo_id: null + save_best_model: false + save_final_model: true + save_pt_file: true + save_training_state: true + save_safetensors_file: false + local: + keep_last_n: 1 + sync: + enabled: false + repo_id: null + keep_last_n: 0 + sync_every_n_checkpoints: 1 + resume_policy: local_latest +run_id: wan_oft_deadly_corridor_fix_latency_6_context5_standard_sft_1000_effbs128_224_currentbce +output_dir: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/wanoft_bench/third_party/starVLA/results/Checkpoints/wan_oft_deadly_corridor_fix_latency_6_context5_standard_sft_1000_effbs128_224_currentbce +config_yaml: null +is_debug: false +version_id: '0.21' diff --git a/latency-aware/deadly-corridor/vla/starvla-wanoft-h8/wan_oft_deadly_corridor_fix_latency_6_context5_standard_sft_1000_effbs128_224_currentbce/config.yaml b/latency-aware/deadly-corridor/vla/starvla-wanoft-h8/wan_oft_deadly_corridor_fix_latency_6_context5_standard_sft_1000_effbs128_224_currentbce/config.yaml new file mode 100644 index 0000000000000000000000000000000000000000..dd6263879abddf89ffe1937981fd15fee3538680 --- /dev/null +++ b/latency-aware/deadly-corridor/vla/starvla-wanoft-h8/wan_oft_deadly_corridor_fix_latency_6_context5_standard_sft_1000_effbs128_224_currentbce/config.yaml @@ -0,0 +1,83 @@ +checkpoint: + local: + keep_last_n: 1 + save_best_model: false + save_final_model: true + save_pt_file: true + save_safetensors_file: false + save_training_state: true + sync: + enabled: false + keep_last_n: 0 + repo_id: null +datasets: + vla_data: + data_mix: deadly_corridor_train__bridge + dataset_py: lerobot_datasets + eval_data_mix: null + latency_curriculum: + enabled: false + obs_image_size: + - 224 + - 224 + per_device_batch_size: 1 +framework: + action_model: + action_dim: 7 + action_env_dim: 7 + action_hidden_dim: 3072 + action_horizon: 8 + action_model_type: MLP + action_query_source: mean + loss_type: current_multibinary_bce + name: WanOFT + qwenvl: + base_vlm: /workspace/runtime/models/Wan2.2-TI2V-5B-Diffusers + world_model: + base_wm: /workspace/runtime/models/Wan2.2-TI2V-5B-Diffusers + extract_layers: + - -1 + num_frames: 5 +output_dir: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/wanoft_bench/third_party/starVLA/results/Checkpoints/wan_oft_deadly_corridor_fix_latency_6_context5_standard_sft_1000_effbs128_224_currentbce +rl_games: + env_eval: + enabled: false +run_id: wan_oft_deadly_corridor_fix_latency_6_context5_standard_sft_1000_effbs128_224_currentbce +run_root_dir: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/wanoft_bench/third_party/starVLA/results/Checkpoints +seed: 42 +trainer: + distributed_backend: deepspeed + eval_action_classification: false + eval_action_classification_interval: null + eval_interval: 100 + freeze_llm_layers: [] + freeze_modules: '' + freeze_tied_embedding: false + freeze_vit: false + gradient_accumulation_steps: 16 + is_resume: false + learning_rate: + action_model: 0.0001 + action_query_proj: 0.0001 + base: 2.5e-05 + qwen_vl_interface: 1.0e-05 + logging_frequency: 1 + lr_scheduler_type: cosine_with_min_lr + max_train_steps: 1000 + num_warmup_steps: 0 + optimizer: + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + fused: true + weight_decay: 1.0e-08 + pretrained_checkpoint: null + profile_timing: + enabled: false + reload_modules: null + save_interval: 500 + scheduler_specific_kwargs: + min_lr: 1.0e-06 +wandb_entity: zihanwang-ai-northwestern-university +wandb_project: starVLA_rl_games diff --git a/latency-aware/deadly-corridor/vla/starvla-wanoft-h8/wan_oft_deadly_corridor_fix_latency_6_context5_standard_sft_1000_effbs128_224_currentbce/dataset_statistics.json b/latency-aware/deadly-corridor/vla/starvla-wanoft-h8/wan_oft_deadly_corridor_fix_latency_6_context5_standard_sft_1000_effbs128_224_currentbce/dataset_statistics.json new file mode 100644 index 0000000000000000000000000000000000000000..9bdac9bda0eca8b9675179e245a47acde1efd126 --- /dev/null +++ b/latency-aware/deadly-corridor/vla/starvla-wanoft-h8/wan_oft_deadly_corridor_fix_latency_6_context5_standard_sft_1000_effbs128_224_currentbce/dataset_statistics.json @@ -0,0 +1,127 @@ +{ + "new_embodiment": { + "action": { + "mean": [ + 0.9434846043586731, + 0.023302890360355377, + 0.21685810387134552, + 0.5315648317337036, + 0.18562282621860504, + 0.09055173397064209, + 0.8273231983184814 + ], + "std": [ + 0.2308816909790039, + 0.15088047087192535, + 0.4120523929595947, + 0.4989914894104004, + 0.388867050409317, + 0.287028968334198, + 0.37795618176460266 + ], + "max": [ + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + 1.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + 1.0 + ], + "mask": [ + true, + true, + true, + true, + true, + true, + true + ] + }, + "state": { + "mean": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ] + }, + "num_transitions": 42484, + "num_trajectories": 990 + } +} \ No newline at end of file diff --git a/latency-aware/demon-attack/small-policy/memory_rollout_teacher:demon_attack:teacher:fixed_l6:fs4:obs15:stride4:seed0/train/memory_rollout_teacher:demon_attack:teacher:fixed_l6:fs4:obs15:stride4:seed0/checkpoint_p0/checkpoint_000048880_25026560.pth b/latency-aware/demon-attack/small-policy/memory_rollout_teacher:demon_attack:teacher:fixed_l6:fs4:obs15:stride4:seed0/train/memory_rollout_teacher:demon_attack:teacher:fixed_l6:fs4:obs15:stride4:seed0/checkpoint_p0/checkpoint_000048880_25026560.pth new file mode 100644 index 0000000000000000000000000000000000000000..c72c2960d75c685fe0ea5e6fd5a407dd42833720 --- /dev/null +++ b/latency-aware/demon-attack/small-policy/memory_rollout_teacher:demon_attack:teacher:fixed_l6:fs4:obs15:stride4:seed0/train/memory_rollout_teacher:demon_attack:teacher:fixed_l6:fs4:obs15:stride4:seed0/checkpoint_p0/checkpoint_000048880_25026560.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:131b3bc73aab996f9154c1fb40b632b77afd46fe1cf1cf9733b11d29870234bd +size 7210813 diff --git a/latency-aware/demon-attack/small-policy/memory_rollout_teacher:demon_attack:teacher:fixed_l6:fs4:obs15:stride4:seed0/train/memory_rollout_teacher:demon_attack:teacher:fixed_l6:fs4:obs15:stride4:seed0/config.json b/latency-aware/demon-attack/small-policy/memory_rollout_teacher:demon_attack:teacher:fixed_l6:fs4:obs15:stride4:seed0/train/memory_rollout_teacher:demon_attack:teacher:fixed_l6:fs4:obs15:stride4:seed0/config.json new file mode 100644 index 0000000000000000000000000000000000000000..4e6f8fc640fa1fb137b52a248221f970750100ee --- /dev/null +++ b/latency-aware/demon-attack/small-policy/memory_rollout_teacher:demon_attack:teacher:fixed_l6:fs4:obs15:stride4:seed0/train/memory_rollout_teacher:demon_attack:teacher:fixed_l6:fs4:obs15:stride4:seed0/config.json @@ -0,0 +1,255 @@ +{ + "help": false, + "algo": "APPO", + "env": "latency_demon_attack", + "experiment": "demon_attack_teacher_fixed_l6_obs15_seed0", + "train_dir": "results/checkpoints/memory/rollout_teacher", + "restart_behavior": "resume", + "device": "gpu", + "seed": 0, + "num_policies": 1, + "async_rl": true, + "serial_mode": false, + "batched_sampling": true, + "num_batches_to_accumulate": 2, + "worker_num_splits": 1, + "policy_workers_per_policy": 1, + "max_policy_lag": 300, + "num_workers": 1, + "num_envs_per_worker": 1, + "batch_size": 1024, + "num_batches_per_epoch": 8, + "num_epochs": 2, + "rollout": 64, + "recurrence": 32, + "shuffle_minibatches": false, + "gamma": 0.99, + "reward_scale": 1.0, + "reward_clip": 1000.0, + "value_bootstrap": false, + "normalize_returns": true, + "exploration_loss_coeff": 0.003, + "value_loss_coeff": 0.5, + "kl_loss_coeff": 0.0, + "exploration_loss": "entropy", + "gae_lambda": 0.95, + "ppo_clip_ratio": 0.1, + "ppo_clip_value": 0.2, + "with_vtrace": false, + "vtrace_rho": 1.0, + "vtrace_c": 1.0, + "optimizer": "adam", + "adam_eps": 1e-05, + "adam_beta1": 0.9, + "adam_beta2": 0.999, + "max_grad_norm": 0.0, + "learning_rate": 0.000303, + "lr_schedule": "linear_decay", + "lr_schedule_kl_threshold": 0.008, + "lr_adaptive_min": 1e-06, + "lr_adaptive_max": 0.01, + "obs_subtract_mean": 0.0, + "obs_scale": 255.0, + "normalize_input": true, + "normalize_input_keys": null, + "decorrelate_experience_max_seconds": 0, + "decorrelate_envs_on_one_worker": true, + "actor_worker_gpus": [ + 0 + ], + "set_workers_cpu_affinity": true, + "force_envs_single_thread": false, + "default_niceness": 0, + "log_to_file": true, + "experiment_summaries_interval": 1, + "flush_summaries_interval": 30, + "stats_avg": 100, + "summaries_use_frameskip": true, + "heartbeat_interval": 20, + "heartbeat_reporting_interval": 180, + "train_for_env_steps": 25000000, + "train_for_seconds": 10000000000, + "save_every_sec": 600, + "keep_checkpoints": 5, + "load_checkpoint_kind": "latest", + "save_milestones_sec": -1, + "save_best_every_sec": 5, + "save_best_metric": "reward", + "save_best_after": 100000, + "benchmark": false, + "encoder_mlp_layers": [ + 512, + 512 + ], + "encoder_conv_architecture": "convnet_atari", + "encoder_conv_mlp_layers": [ + 512 + ], + "use_rnn": false, + "rnn_size": 512, + "rnn_type": "gru", + "rnn_num_layers": 1, + "decoder_mlp_layers": [], + "nonlinearity": "elu", + "policy_initialization": "orthogonal", + "policy_init_gain": 1.0, + "actor_critic_share_weights": true, + "adaptive_stddev": true, + "continuous_tanh_scale": 0.0, + "initial_stddev": 1.0, + "use_env_info_cache": false, + "env_gpu_actions": true, + "env_gpu_observations": true, + "env_frameskip": 1, + "env_framestack": 1, + "pixel_format": "CHW", + "use_record_episode_statistics": false, + "with_wandb": true, + "wandb_user": null, + "wandb_project": "latency-sensitive-bench", + "wandb_group": "demon-attack-obs15-fixed_l6", + "wandb_job_type": "sample_factory", + "wandb_tags": [ + "memory_rollout_teacher", + "demon-attack", + "teacher", + "fixed", + "fixed_l6", + "obs15", + "seed0" + ], + "with_pbt": false, + "pbt_mix_policies_in_one_env": true, + "pbt_period_env_steps": 5000000, + "pbt_start_mutation": 20000000, + "pbt_replace_fraction": 0.3, + "pbt_mutation_rate": 0.15, + "pbt_replace_reward_gap": 0.1, + "pbt_replace_reward_gap_absolute": 1e-06, + "pbt_optimize_gamma": false, + "pbt_target_objective": "true_objective", + "pbt_perturb_min": 1.1, + "pbt_perturb_max": 1.5, + "env_id": "ALE/DemonAttack-v5", + "env_fps": 60.0, + "obs_fps": 15.0, + "screen_size": 84, + "noop_max": 30, + "reset_mode": "fire", + "frame_stack": 4, + "terminal_on_life_loss": true, + "clip_reward": true, + "repeat_action_probability": 0.0, + "full_action_space": false, + "max_episode_steps": 25000, + "simulator": "gpu", + "gpu_env_device": "auto", + "gpu_env_batch_size": 512, + "export_env_raw_rgb_frames": false, + "mode": "train", + "latency_type": "fixed", + "fixed_latency_ms": 100.0, + "mean_latency_ms": null, + "std_latency_ms": null, + "min_latency_ms": null, + "max_latency_ms": null, + "latency_seed": null, + "add_latency_info": false, + "max_pending_actions": null, + "hold_policy": "hold", + "ordering_policy": "latest_ready", + "eval_episodes": 100, + "eval_parallel_envs": 50, + "eval_latency_raw_frame_values": "0,1,2,3,4,5", + "eval_max_steps": 7200, + "eval_deterministic": true, + "eval_raw_reward": true, + "episode_metrics_path": "results/checkpoints/memory/rollout_teacher/demon_attack_teacher_fixed_l6_obs15_seed0/episode_metrics.jsonl", + "command_line": "--mode train --algo APPO --env latency_demon_attack --experiment demon_attack_teacher_fixed_l6_obs15_seed0 --train_dir results/checkpoints/memory/rollout_teacher --restart_behavior resume --device gpu --actor_worker_gpus 0 --env_gpu_observations True --env_gpu_actions True --gpu-env-batch-size 512 --seed 0 --episode_metrics_path results/checkpoints/memory/rollout_teacher/demon_attack_teacher_fixed_l6_obs15_seed0/episode_metrics.jsonl --train_for_env_steps 25000000 --num_workers 1 --num_envs_per_worker 1 --num_policies 1 --batch_size 1024 --rollout 64 --recurrence 32 --num_epochs 2 --num_batches_per_epoch 8 --worker_num_splits 1 --max_policy_lag 300 --learning_rate 0.000303 --gamma 0.99 --gae_lambda 0.95 --ppo_clip_ratio 0.1 --ppo_clip_value 0.2 --value_loss_coeff 0.5 --max_grad_norm 0.0 --save_every_sec 600 --keep_checkpoints 5 --stats_avg 100 --experiment_summaries_interval 1 --batched_sampling True --async_rl True --batched_sampling True --use_rnn False --normalize_returns True --normalize_input True --latency-type fixed --fixed-latency-ms 100.0 --add-latency-info False --eval-episodes 100 --eval-parallel-envs 50 --eval-max-steps 7200 --eval-deterministic True --eval-raw-reward --with_wandb True --wandb_project latency-sensitive-bench --wandb_group demon-attack-obs15-fixed_l6 --wandb_job_type sample_factory --wandb_tags memory_rollout_teacher demon-attack teacher fixed fixed_l6 obs15 seed0 --env_id ALE/DemonAttack-v5 --env-fps 60 --obs-fps 15.0 --screen_size 84 --noop_max 30 --reset-mode fire --frame_stack 4 --terminal_on_life_loss True --clip_reward True --repeat_action_probability 0.0 --full_action_space False --max_episode_steps 25000 --simulator gpu --gpu-env-device auto --hold-policy hold --ordering-policy latest_ready", + "cli_args": { + "algo": "APPO", + "env": "latency_demon_attack", + "experiment": "demon_attack_teacher_fixed_l6_obs15_seed0", + "train_dir": "results/checkpoints/memory/rollout_teacher", + "restart_behavior": "resume", + "device": "gpu", + "seed": 0, + "num_policies": 1, + "async_rl": true, + "batched_sampling": true, + "worker_num_splits": 1, + "max_policy_lag": 300, + "num_workers": 1, + "num_envs_per_worker": 1, + "batch_size": 1024, + "num_batches_per_epoch": 8, + "num_epochs": 2, + "rollout": 64, + "recurrence": 32, + "gamma": 0.99, + "normalize_returns": true, + "value_loss_coeff": 0.5, + "gae_lambda": 0.95, + "ppo_clip_ratio": 0.1, + "ppo_clip_value": 0.2, + "max_grad_norm": 0.0, + "learning_rate": 0.000303, + "normalize_input": true, + "actor_worker_gpus": [ + 0 + ], + "experiment_summaries_interval": 1, + "stats_avg": 100, + "train_for_env_steps": 25000000, + "save_every_sec": 600, + "keep_checkpoints": 5, + "use_rnn": false, + "env_gpu_actions": true, + "env_gpu_observations": true, + "with_wandb": true, + "wandb_project": "latency-sensitive-bench", + "wandb_group": "demon-attack-obs15-fixed_l6", + "wandb_job_type": "sample_factory", + "wandb_tags": [ + "memory_rollout_teacher", + "demon-attack", + "teacher", + "fixed", + "fixed_l6", + "obs15", + "seed0" + ], + "env_id": "ALE/DemonAttack-v5", + "env_fps": 60.0, + "obs_fps": 15.0, + "screen_size": 84, + "noop_max": 30, + "reset_mode": "fire", + "frame_stack": 4, + "terminal_on_life_loss": true, + "clip_reward": true, + "repeat_action_probability": 0.0, + "full_action_space": false, + "max_episode_steps": 25000, + "simulator": "gpu", + "gpu_env_device": "auto", + "gpu_env_batch_size": 512, + "mode": "train", + "latency_type": "fixed", + "fixed_latency_ms": 100.0, + "add_latency_info": false, + "hold_policy": "hold", + "ordering_policy": "latest_ready", + "eval_episodes": 100, + "eval_parallel_envs": 50, + "eval_max_steps": 7200, + "eval_deterministic": true, + "eval_raw_reward": true, + "episode_metrics_path": "results/checkpoints/memory/rollout_teacher/demon_attack_teacher_fixed_l6_obs15_seed0/episode_metrics.jsonl" + }, + "git_hash": "bedb62f3a97309f6dacdb6a2953e3312d9776765", + "git_repo_name": "git@github.com:ZihanWang314/latency-sensitive-bench.git", + "eval_env_frameskip": 1, + "output_dir": "outputs/memory/rollout_teacher/demon_attack/train/teacher/fixed_l6/obs15/seed_0", + "wandb_unique_id": "demon-attack-obs15-fixed_l6-s0" +} \ No newline at end of file diff --git a/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_2_200ep/checkpoints/steps_7000_model.safetensors b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_2_200ep/checkpoints/steps_7000_model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..d7b9a75500c7337d23907558c037120e370d6d96 --- /dev/null +++ b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_2_200ep/checkpoints/steps_7000_model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2ab6d4d4eb795a9a09da9c63d489771039440689877642e3f9ec48dd9112119a +size 9138230516 diff --git a/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_2_200ep/config.full.yaml b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_2_200ep/config.full.yaml new file mode 100644 index 0000000000000000000000000000000000000000..c4d160238ea9c743e2653a1b331d1f1b78fd0d17 --- /dev/null +++ b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_2_200ep/config.full.yaml @@ -0,0 +1,209 @@ +framework: + name: QwenOFT + qwenvl: + base_vlm: /workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + attn_implementation: flash_attention_2 + enable_gradient_checkpointing: true + action_model: + action_model_type: MLP + action_dim: 7 + action_hidden_dim: 2560 + future_action_window_size: 0 + past_action_window_size: 0 + loss_type: discrete_ce + state_dim: 7 + action_horizon: 1 + action_env_dim: 6 +datasets: + vla_data: + dataset_py: lerobot_datasets + include_state: true + data_root_dir: /workspace/starVLA/data/demon_attack_fix_latency_2_200ep + data_mix: demon_attack_train__bridge + eval_data_mix: demon_attack_train__bridge__val + custom_mixtures_path: null + action_type: discrete + sequential_step_sampling: false + shuffle: true + action_balance: + enabled: false + strategy: balanced_epoch + action_key: action_id + target_flap_fraction: 0.3 + noop_id: 0 + flap_id: 1 + latency_curriculum: + enabled: false + strategy: exclusive + latencies: null + phase_steps: null + per_device_batch_size: 32 + load_all_data_for_training: true + obs_image_size: null + video_backend: torchvision_av +dataset: + source_hf: '' + config_name: null + source_subdir: null + converted_name: demon_attack_train + single_source_hf: '' + mixed_source_hf: '' + single_converted_name: demon_attack_train + mixed_converted_name: demon_attack_mixed_latency_train + single_latency_filter: null + mixed_latency_filter: null + force_download: false + setup_force: false + skip_verification: false + verify_rows: 200 + max_episodes: null + episodes_per_latency: null + latency_filter: null + debug_subset: + enabled: false + max_episodes: 5 + suffix: debug +base_model: + repo_id: Qwen/Qwen3-VL-4B-Instruct +initialization: + checkpoint_local_dir: playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_hf_repo_id: StarVLA/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_filename: checkpoints/steps_5000_pytorch_model.pt +trainer: + max_train_steps: 7000 + num_warmup_steps: 100 + save_interval: 100 + eval_interval: 100 + eval_num_batches: 100 + per_latency_eval_num_batches: null + eval_action_classification: true + cc_f1_tolerance: 1 + learning_rate: + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + action_model: 0.0001 + lr_scheduler_type: cosine_with_min_lr + scheduler_specific_kwargs: + min_lr: 1.0e-06 + freeze_modules: '' + loss_scale: + vla: 1.0 + vlm: 0.1 + max_grad_norm: 1.0 + weight_decay: 0.0 + logging_frequency: 1 + gradient_clipping: 1.0 + gradient_accumulation_steps: 4 + distributed_backend: none + is_resume: false + pretrained_checkpoint: null + resume_step: 0 + reload_modules: null + optimizer: + name: AdamW + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + weight_decay: 1.0e-08 + fused: false + save_format: pt +workspace_dir: WORKSPACE_DIR +run_root_dir: /workspace/starVLA/results/Checkpoints +seed: 42 +wandb_entity: dongqianyu99-zhejiang-university +wandb_project: starVLA_rl_games +auth: + env_file: null + hf_token_env: HF_TOKEN + wandb_api_key_env: WANDB_API_KEY +paths: + run_root_dir: results/Checkpoints + dataset_local_dir: playground/Datasets/rl_games + dataset_cache_dir: null + base_model_dir: playground/Pretrained_models/Qwen3-VL-4B-Instruct + accelerate_config: starVLA/config/deepseeds/deepspeed_zero2.yaml +launch: + use_accelerate: true + gpus: null + num_processes: 1 + dry_run: false +conda: + enabled: true + env_name: null +rl_games: + model_alias: openvla + env_eval: + image_size: 224 + frameskip: 4 + seed: 42 + fixed_episode_seeds: true + latency_seed_stride: 0 + task_seed_stride: 0 + task_description: 'You are playing Demon Attack from a single game image. Choose + exactly one action from: NOOP, FIRE, RIGHT, LEFT, RIGHTFIRE, LEFTFIRE.' + demon_attack: + noop_max: 30 + enabled: true + distributed_mode: none + vectorized: + enabled: false + batch_size: 1 + latency: + prompt_map_path: /workspace/starVLA/data/demon_attack_fix_latency_2_200ep/demon_attack_train__bridge/latency_prompt_map.json + mode: single + values: + - 0 + mid_train: + enabled: false + interval_steps: 100 + latencies: + - 0 + num_episodes: 5 + max_steps_per_episode: 3600 + post_train: + enabled: false + latencies: + - 0 + - 1 + - 2 + - 3 + - 4 + num_episodes: 5 + max_steps_per_episode: 3600 + task: demon_attack + initialization_mode: bridge + action_carrier: bridge +model: openvla +env: demon_attack +init: bridge +bridge_base_model: + repo_id: + openvla: Qwen/Qwen3-VL-4B-Instruct + pi0: StarVLA/Qwen2.5-VL-3B-Instruct-Action + pi05: Qwen/Qwen3-VL-4B-Instruct + gr00t: Qwen/Qwen3-VL-4B-Instruct + local_dir: + openvla: playground/Pretrained_models/Qwen3-VL-4B-Instruct + pi0: playground/Pretrained_models/Qwen2.5-VL-3B-Instruct-Action + pi05: playground/Pretrained_models/Qwen3-VL-4B-Instruct + gr00t: playground/Pretrained_models/Qwen3-VL-4B-Instruct +mode: single +checkpoint: + load: auto + hf_repo_id: null + save_best_model: false + save_pt_file: false + local: + keep_last_n: 1 + sync: + enabled: false + repo_id: null + keep_last_n: 0 + sync_every_n_checkpoints: 1 + resume_policy: local_latest +run_id: demon_attack_fix_latency_2_200ep +output_dir: /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_2_200ep +config_yaml: null +is_debug: false +version_id: '0.21' diff --git a/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_2_200ep/config.yaml b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_2_200ep/config.yaml new file mode 100644 index 0000000000000000000000000000000000000000..ab81e313a52d83a6b660d7578813657e0cc99c4c --- /dev/null +++ b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_2_200ep/config.yaml @@ -0,0 +1,75 @@ +checkpoint: + local: + keep_last_n: 1 + save_best_model: false + save_pt_file: false + sync: + enabled: false + keep_last_n: 0 + repo_id: null +datasets: + vla_data: + data_mix: demon_attack_train__bridge + dataset_py: lerobot_datasets + eval_data_mix: demon_attack_train__bridge__val + latency_curriculum: + enabled: false + obs_image_size: null + per_device_batch_size: 32 +framework: + action_model: + action_dim: 7 + action_env_dim: 6 + action_hidden_dim: 2560 + action_horizon: 1 + action_model_type: MLP + loss_type: discrete_ce + name: QwenOFT + qwenvl: + attn_implementation: flash_attention_2 + base_vlm: /workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + enable_gradient_checkpointing: true +output_dir: /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_2_200ep +rl_games: + env_eval: + enabled: true + mid_train: + enabled: false + interval_steps: 100 + task: demon_attack +run_id: demon_attack_fix_latency_2_200ep +run_root_dir: /workspace/starVLA/results/Checkpoints +seed: 42 +trainer: + cc_f1_tolerance: 1 + distributed_backend: none + eval_action_classification: true + eval_interval: 100 + eval_num_batches: 100 + freeze_modules: '' + gradient_accumulation_steps: 4 + gradient_clipping: 1.0 + is_resume: false + learning_rate: + action_model: 0.0001 + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + logging_frequency: 1 + lr_scheduler_type: cosine_with_min_lr + max_train_steps: 7000 + num_warmup_steps: 100 + optimizer: + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + fused: false + weight_decay: 1.0e-08 + per_latency_eval_num_batches: null + pretrained_checkpoint: null + reload_modules: null + save_interval: 100 + scheduler_specific_kwargs: + min_lr: 1.0e-06 +wandb_entity: dongqianyu99-zhejiang-university +wandb_project: starVLA_rl_games diff --git a/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_2_200ep/dataset_statistics.json b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_2_200ep/dataset_statistics.json new file mode 100644 index 0000000000000000000000000000000000000000..fe6cef2f95beaa722344f7ada808ee5f59d12237 --- /dev/null +++ b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_2_200ep/dataset_statistics.json @@ -0,0 +1,127 @@ +{ + "new_embodiment": { + "action": { + "mean": [ + 0.06905864179134369, + 0.37893569469451904, + 0.13702447712421417, + 0.1011248454451561, + 0.12969808280467987, + 0.18415828049182892, + 0.0 + ], + "std": [ + 0.2532638907432556, + 0.4849342703819275, + 0.3443855047225952, + 0.3010394871234894, + 0.3360588848590851, + 0.3878043293952942, + 0.0 + ], + "max": [ + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + 0.0 + ], + "mask": [ + true, + true, + true, + true, + true, + true, + true + ] + }, + "state": { + "mean": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ] + }, + "num_transitions": 349558, + "num_trajectories": 180 + } +} \ No newline at end of file diff --git a/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_2_200ep/dataset_statistics_eval.json b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_2_200ep/dataset_statistics_eval.json new file mode 100644 index 0000000000000000000000000000000000000000..fc5bc71b1470a8fa935f98ea221f1d7e3c26f7a1 --- /dev/null +++ b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_2_200ep/dataset_statistics_eval.json @@ -0,0 +1,127 @@ +{ + "new_embodiment": { + "action": { + "mean": [ + 0.07371096312999725, + 0.3668539524078369, + 0.13910865783691406, + 0.1111111119389534, + 0.13123494386672974, + 0.17798034846782684, + 0.0 + ], + "std": [ + 0.2612755298614502, + 0.48196175694465637, + 0.3461320102214813, + 0.3142721354961395, + 0.3376607596874237, + 0.38254332542419434, + 0.0 + ], + "max": [ + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + 0.0 + ], + "mask": [ + true, + true, + true, + true, + true, + true, + true + ] + }, + "state": { + "mean": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ] + }, + "num_transitions": 52326, + "num_trajectories": 20 + } +} \ No newline at end of file diff --git a/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_2_200ep/hydra/.hydra/config.yaml b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_2_200ep/hydra/.hydra/config.yaml new file mode 100644 index 0000000000000000000000000000000000000000..2cd07f40e047d72a4c7e2670c5a25dbf6300b002 --- /dev/null +++ b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_2_200ep/hydra/.hydra/config.yaml @@ -0,0 +1,207 @@ +framework: + qwenvl: + base_vlm: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/playground/Pretrained_models/Qwen3-VL-4B-Instruct + attn_implementation: flash_attention_2 + enable_gradient_checkpointing: true + action_model: + state_dim: 7 + loss_type: discrete_ce + action_horizon: 1 + future_action_window_size: 0 + past_action_window_size: 0 + action_dim: 7 + action_env_dim: 6 + name: QwenOFT +datasets: + vla_data: + dataset_py: lerobot_datasets + include_state: true + data_root_dir: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/data/demon_attack_fix_latency_2_200ep + data_mix: demon_attack_train__bridge + eval_data_mix: demon_attack_train__bridge__val + custom_mixtures_path: null + action_type: discrete + sequential_step_sampling: false + shuffle: true + action_balance: + enabled: false + strategy: balanced_epoch + action_key: action_id + target_flap_fraction: 0.3 + noop_id: 0 + flap_id: 1 + latency_curriculum: + enabled: false + strategy: exclusive + latencies: null + phase_steps: null + per_device_batch_size: 32 + load_all_data_for_training: true + obs_image_size: null + video_backend: torchvision_av +dataset: + source_hf: ${dataset.single_source_hf} + config_name: null + source_subdir: null + converted_name: ${dataset.single_converted_name} + single_source_hf: '' + mixed_source_hf: '' + single_converted_name: demon_attack_train + mixed_converted_name: demon_attack_mixed_latency_train + single_latency_filter: null + mixed_latency_filter: null + force_download: false + setup_force: false + skip_verification: false + verify_rows: 200 + max_episodes: null + episodes_per_latency: null + latency_filter: ${dataset.single_latency_filter} + debug_subset: + enabled: false + max_episodes: 5 + suffix: debug +base_model: + repo_id: ${bridge_base_model.repo_id.${model}} +initialization: + checkpoint_local_dir: playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_hf_repo_id: StarVLA/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_filename: checkpoints/steps_5000_pytorch_model.pt +trainer: + max_train_steps: 7000 + num_warmup_steps: 100 + save_interval: 100 + eval_interval: 100 + eval_num_batches: 100 + per_latency_eval_num_batches: null + eval_action_classification: true + cc_f1_tolerance: 1 + learning_rate: + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + action_model: 0.0001 + lr_scheduler_type: cosine_with_min_lr + scheduler_specific_kwargs: + min_lr: 1.0e-06 + freeze_modules: '' + loss_scale: + vla: 1.0 + vlm: 0.1 + max_grad_norm: 1.0 + weight_decay: 0.0 + logging_frequency: 1 + gradient_clipping: 1.0 + gradient_accumulation_steps: 4 + distributed_backend: none + is_resume: false + pretrained_checkpoint: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt + resume_step: 0 + reload_modules: null + optimizer: + name: AdamW + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + weight_decay: 1.0e-08 + fused: false + save_format: pt +workspace_dir: WORKSPACE_DIR +run_root_dir: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints +seed: 42 +wandb_entity: dongqianyu99-zhejiang-university +wandb_project: starVLA_rl_games +auth: + env_file: null + hf_token_env: HF_TOKEN + wandb_api_key_env: WANDB_API_KEY +paths: + run_root_dir: results/Checkpoints + dataset_local_dir: playground/Datasets/rl_games + dataset_cache_dir: null + base_model_dir: ${bridge_base_model.local_dir.${model}} + accelerate_config: starVLA/config/deepseeds/deepspeed_zero2.yaml +launch: + use_accelerate: true + gpus: null + num_processes: 1 + dry_run: false +conda: + enabled: true + env_name: null +rl_games: + model_alias: openvla + env_eval: + image_size: 224 + frameskip: 4 + seed: 42 + fixed_episode_seeds: true + latency_seed_stride: 0 + task_seed_stride: 0 + task_description: 'You are playing Demon Attack from a single game image. Choose + exactly one action from: NOOP, FIRE, RIGHT, LEFT, RIGHTFIRE, LEFTFIRE.' + demon_attack: + noop_max: 30 + enabled: true + distributed_mode: none + vectorized: + enabled: false + batch_size: 1 + latency: + prompt_map_path: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/data/demon_attack_fix_latency_2_200ep/demon_attack_train__bridge/latency_prompt_map.json + mode: single + values: + - 0 + mid_train: + enabled: false + interval_steps: 100 + latencies: + - 0 + num_episodes: 5 + max_steps_per_episode: 3600 + post_train: + enabled: false + latencies: + - 0 + - 1 + - 2 + - 3 + - 4 + num_episodes: 5 + max_steps_per_episode: 3600 + task: demon_attack + initialization_mode: bridge + action_carrier: bridge +model: openvla +env: demon_attack +init: bridge +bridge_base_model: + repo_id: + openvla: Qwen/Qwen3-VL-4B-Instruct + pi0: StarVLA/Qwen2.5-VL-3B-Instruct-Action + pi05: Qwen/Qwen3-VL-4B-Instruct + gr00t: Qwen/Qwen3-VL-4B-Instruct + local_dir: + openvla: playground/Pretrained_models/Qwen3-VL-4B-Instruct + pi0: playground/Pretrained_models/Qwen2.5-VL-3B-Instruct-Action + pi05: playground/Pretrained_models/Qwen3-VL-4B-Instruct + gr00t: playground/Pretrained_models/Qwen3-VL-4B-Instruct +mode: single +checkpoint: + load: auto + hf_repo_id: null + save_best_model: false + save_pt_file: false + local: + keep_last_n: 1 + sync: + enabled: false + repo_id: null + keep_last_n: 0 + sync_every_n_checkpoints: 1 + resume_policy: local_latest +run_id: demon_attack_fix_latency_2_200ep +output_dir: null +config_yaml: null +is_debug: false +version_id: '0.21' diff --git a/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_2_200ep/hydra/.hydra/hydra.yaml b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_2_200ep/hydra/.hydra/hydra.yaml new file mode 100644 index 0000000000000000000000000000000000000000..e72222b2c7b9bd2d436d79e3c00f96c61763344e --- /dev/null +++ b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_2_200ep/hydra/.hydra/hydra.yaml @@ -0,0 +1,259 @@ +hydra: + run: + dir: ${run_root_dir}/${run_id}/hydra + sweep: + dir: multirun/${now:%Y-%m-%d}/${now:%H-%M-%S} + subdir: ${hydra.job.num} + launcher: + _target_: hydra._internal.core_plugins.basic_launcher.BasicLauncher + sweeper: + _target_: hydra._internal.core_plugins.basic_sweeper.BasicSweeper + max_batch_size: null + params: null + help: + app_name: ${hydra.job.name} + header: '${hydra.help.app_name} is powered by Hydra. + + ' + footer: 'Powered by Hydra (https://hydra.cc) + + Use --hydra-help to view Hydra specific help + + ' + template: '${hydra.help.header} + + == Configuration groups == + + Compose your configuration from those groups (group=option) + + + $APP_CONFIG_GROUPS + + + == Config == + + Override anything in the config (foo.bar=value) + + + $CONFIG + + + ${hydra.help.footer} + + ' + hydra_help: + template: 'Hydra (${hydra.runtime.version}) + + See https://hydra.cc for more info. + + + == Flags == + + $FLAGS_HELP + + + == Configuration groups == + + Compose your configuration from those groups (For example, append hydra/job_logging=disabled + to command line) + + + $HYDRA_CONFIG_GROUPS + + + Use ''--cfg hydra'' to Show the Hydra config. + + ' + hydra_help: ??? + hydra_logging: + version: 1 + formatters: + simple: + format: '[%(asctime)s][HYDRA] %(message)s' + handlers: + console: + class: logging.StreamHandler + formatter: simple + stream: ext://sys.stdout + root: + level: INFO + handlers: + - console + loggers: + logging_example: + level: DEBUG + disable_existing_loggers: false + job_logging: + version: 1 + formatters: + simple: + format: '[%(asctime)s][%(name)s][%(levelname)s] - %(message)s' + handlers: + console: + class: logging.StreamHandler + formatter: simple + stream: ext://sys.stdout + file: + class: logging.FileHandler + formatter: simple + filename: ${hydra.runtime.output_dir}/${hydra.job.name}.log + root: + level: INFO + handlers: + - console + - file + disable_existing_loggers: false + env: {} + mode: RUN + searchpath: [] + callbacks: {} + output_subdir: .hydra + overrides: + hydra: + - hydra.mode=RUN + task: + - model=openvla + - env=demon_attack + - init=bridge + - mode=single + - run_id=demon_attack_fix_latency_2_200ep + - run_root_dir=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints + - seed=42 + - wandb_entity=dongqianyu99-zhejiang-university + - wandb_project=starVLA_rl_games + - rl_games.env_eval.enabled=true + - checkpoint.sync.enabled=false + - checkpoint.sync.keep_last_n=0 + - checkpoint.local.keep_last_n=1 + - checkpoint.save_best_model=false + - checkpoint.save_pt_file=false + - trainer.is_resume=false + - trainer.pretrained_checkpoint=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt + - trainer.resume_step=0 + - trainer.max_train_steps=7000 + - trainer.num_warmup_steps=100 + - trainer.save_interval=100 + - trainer.eval_interval=100 + - trainer.eval_num_batches=100 + - trainer.logging_frequency=1 + - trainer.gradient_accumulation_steps=4 + - trainer.distributed_backend=none + - trainer.learning_rate.base=2e-05 + - trainer.learning_rate.qwen_vl_interface=1e-05 + - trainer.learning_rate.action_model=0.0001 + - trainer.lr_scheduler_type=cosine_with_min_lr + - trainer.scheduler_specific_kwargs.min_lr=1e-06 + - trainer.loss_scale.vla=1.0 + - trainer.loss_scale.vlm=0.1 + - trainer.max_grad_norm=1.0 + - trainer.weight_decay=0.0 + - trainer.gradient_clipping=1.0 + - trainer.optimizer.name=AdamW + - trainer.optimizer.betas=[0.9,0.95] + - trainer.optimizer.eps=1e-08 + - trainer.optimizer.weight_decay=1e-08 + - trainer.optimizer.fused=false + - trainer.save_format=pt + - framework.name=QwenOFT + - framework.qwenvl.attn_implementation=flash_attention_2 + - framework.qwenvl.enable_gradient_checkpointing=true + - framework.action_model.action_dim=7 + - framework.action_model.action_env_dim=6 + - framework.action_model.state_dim=7 + - framework.action_model.loss_type=discrete_ce + - framework.action_model.action_horizon=1 + - framework.action_model.future_action_window_size=0 + - framework.action_model.past_action_window_size=0 + - datasets.vla_data.include_state=true + - datasets.vla_data.action_type=discrete + - datasets.vla_data.sequential_step_sampling=false + - datasets.vla_data.shuffle=true + - datasets.vla_data.action_balance.enabled=false + - datasets.vla_data.action_balance.strategy=balanced_epoch + - datasets.vla_data.action_balance.action_key=action_id + - datasets.vla_data.action_balance.target_flap_fraction=0.3 + - datasets.vla_data.action_balance.noop_id=0 + - datasets.vla_data.action_balance.flap_id=1 + - datasets.vla_data.latency_curriculum.enabled=false + - datasets.vla_data.latency_curriculum.strategy=exclusive + - datasets.vla_data.per_device_batch_size=32 + - datasets.vla_data.load_all_data_for_training=true + - datasets.vla_data.video_backend=torchvision_av + - datasets.vla_data.data_root_dir=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/data/demon_attack_fix_latency_2_200ep + - datasets.vla_data.data_mix=demon_attack_train__bridge + - datasets.vla_data.eval_data_mix=demon_attack_train__bridge__val + - framework.qwenvl.base_vlm=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/playground/Pretrained_models/Qwen3-VL-4B-Instruct + - rl_games.task=demon_attack + - rl_games.model_alias=openvla + - rl_games.initialization_mode=bridge + - rl_games.action_carrier=bridge + - rl_games.env_eval.distributed_mode=none + - rl_games.env_eval.latency.mode=single + - rl_games.env_eval.frameskip=4 + - rl_games.env_eval.image_size=224 + - rl_games.env_eval.seed=42 + - rl_games.env_eval.fixed_episode_seeds=true + - rl_games.env_eval.latency_seed_stride=0 + - rl_games.env_eval.task_seed_stride=0 + - 'rl_games.env_eval.task_description=''You are playing Demon Attack from a single + game image. Choose exactly one action from: NOOP, FIRE, RIGHT, LEFT, RIGHTFIRE, + LEFTFIRE.''' + - rl_games.env_eval.latency.values=[0] + - rl_games.env_eval.latency.prompt_map_path=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/data/demon_attack_fix_latency_2_200ep/demon_attack_train__bridge/latency_prompt_map.json + - rl_games.env_eval.mid_train.enabled=false + - rl_games.env_eval.mid_train.interval_steps=100 + - rl_games.env_eval.mid_train.num_episodes=5 + - rl_games.env_eval.mid_train.max_steps_per_episode=3600 + - rl_games.env_eval.mid_train.latencies=[0] + - rl_games.env_eval.post_train.enabled=false + - rl_games.env_eval.post_train.num_episodes=5 + - rl_games.env_eval.post_train.max_steps_per_episode=3600 + - rl_games.env_eval.post_train.latencies=[0,1,2,3,4] + job: + name: train_starvla_hydra + chdir: false + override_dirname: 'checkpoint.local.keep_last_n=1,checkpoint.save_best_model=false,checkpoint.save_pt_file=false,checkpoint.sync.enabled=false,checkpoint.sync.keep_last_n=0,datasets.vla_data.action_balance.action_key=action_id,datasets.vla_data.action_balance.enabled=false,datasets.vla_data.action_balance.flap_id=1,datasets.vla_data.action_balance.noop_id=0,datasets.vla_data.action_balance.strategy=balanced_epoch,datasets.vla_data.action_balance.target_flap_fraction=0.3,datasets.vla_data.action_type=discrete,datasets.vla_data.data_mix=demon_attack_train__bridge,datasets.vla_data.data_root_dir=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/data/demon_attack_fix_latency_2_200ep,datasets.vla_data.eval_data_mix=demon_attack_train__bridge__val,datasets.vla_data.include_state=true,datasets.vla_data.latency_curriculum.enabled=false,datasets.vla_data.latency_curriculum.strategy=exclusive,datasets.vla_data.load_all_data_for_training=true,datasets.vla_data.per_device_batch_size=32,datasets.vla_data.sequential_step_sampling=false,datasets.vla_data.shuffle=true,datasets.vla_data.video_backend=torchvision_av,env=demon_attack,framework.action_model.action_dim=7,framework.action_model.action_env_dim=6,framework.action_model.action_horizon=1,framework.action_model.future_action_window_size=0,framework.action_model.loss_type=discrete_ce,framework.action_model.past_action_window_size=0,framework.action_model.state_dim=7,framework.name=QwenOFT,framework.qwenvl.attn_implementation=flash_attention_2,framework.qwenvl.base_vlm=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/playground/Pretrained_models/Qwen3-VL-4B-Instruct,framework.qwenvl.enable_gradient_checkpointing=true,init=bridge,mode=single,model=openvla,rl_games.action_carrier=bridge,rl_games.env_eval.distributed_mode=none,rl_games.env_eval.enabled=true,rl_games.env_eval.fixed_episode_seeds=true,rl_games.env_eval.frameskip=4,rl_games.env_eval.image_size=224,rl_games.env_eval.latency.mode=single,rl_games.env_eval.latency.prompt_map_path=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/data/demon_attack_fix_latency_2_200ep/demon_attack_train__bridge/latency_prompt_map.json,rl_games.env_eval.latency.values=[0],rl_games.env_eval.latency_seed_stride=0,rl_games.env_eval.mid_train.enabled=false,rl_games.env_eval.mid_train.interval_steps=100,rl_games.env_eval.mid_train.latencies=[0],rl_games.env_eval.mid_train.max_steps_per_episode=3600,rl_games.env_eval.mid_train.num_episodes=5,rl_games.env_eval.post_train.enabled=false,rl_games.env_eval.post_train.latencies=[0,1,2,3,4],rl_games.env_eval.post_train.max_steps_per_episode=3600,rl_games.env_eval.post_train.num_episodes=5,rl_games.env_eval.seed=42,rl_games.env_eval.task_description=''You + are playing Demon Attack from a single game image. Choose exactly one action + from: NOOP, FIRE, RIGHT, LEFT, RIGHTFIRE, LEFTFIRE.'',rl_games.env_eval.task_seed_stride=0,rl_games.initialization_mode=bridge,rl_games.model_alias=openvla,rl_games.task=demon_attack,run_id=demon_attack_fix_latency_2_200ep,run_root_dir=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints,seed=42,trainer.distributed_backend=none,trainer.eval_interval=100,trainer.eval_num_batches=100,trainer.gradient_accumulation_steps=4,trainer.gradient_clipping=1.0,trainer.is_resume=false,trainer.learning_rate.action_model=0.0001,trainer.learning_rate.base=2e-05,trainer.learning_rate.qwen_vl_interface=1e-05,trainer.logging_frequency=1,trainer.loss_scale.vla=1.0,trainer.loss_scale.vlm=0.1,trainer.lr_scheduler_type=cosine_with_min_lr,trainer.max_grad_norm=1.0,trainer.max_train_steps=7000,trainer.num_warmup_steps=100,trainer.optimizer.betas=[0.9,0.95],trainer.optimizer.eps=1e-08,trainer.optimizer.fused=false,trainer.optimizer.name=AdamW,trainer.optimizer.weight_decay=1e-08,trainer.pretrained_checkpoint=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt,trainer.resume_step=0,trainer.save_format=pt,trainer.save_interval=100,trainer.scheduler_specific_kwargs.min_lr=1e-06,trainer.weight_decay=0.0,wandb_entity=dongqianyu99-zhejiang-university,wandb_project=starVLA_rl_games' + id: ??? + num: ??? + config_name: train + env_set: {} + env_copy: [] + config: + override_dirname: + kv_sep: '=' + item_sep: ',' + exclude_keys: [] + runtime: + version: 1.3.2 + version_base: '1.1' + cwd: /inspire/hdd/project/spatialintelligence/public/lzj/starVLA + config_sources: + - path: hydra.conf + schema: pkg + provider: hydra + - path: /inspire/hdd/project/spatialintelligence/public/lzj/starVLA/examples/rl_games/config + schema: file + provider: main + - path: '' + schema: structured + provider: schema + output_dir: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/hydra + choices: + cross_task_setup: null + checkpoint: default + mode: single + init: bridge + env: demon_attack + model: openvla + hydra/env: default + hydra/callbacks: null + hydra/job_logging: default + hydra/hydra_logging: default + hydra/hydra_help: default + hydra/help: default + hydra/sweeper: basic + hydra/launcher: basic + hydra/output: default + verbose: false diff --git a/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_2_200ep/hydra/.hydra/overrides.yaml b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_2_200ep/hydra/.hydra/overrides.yaml new file mode 100644 index 0000000000000000000000000000000000000000..93828610c2c9574182880b038223b7f5ca41ea3e --- /dev/null +++ b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_2_200ep/hydra/.hydra/overrides.yaml @@ -0,0 +1,97 @@ +- model=openvla +- env=demon_attack +- init=bridge +- mode=single +- run_id=demon_attack_fix_latency_2_200ep +- run_root_dir=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints +- seed=42 +- wandb_entity=dongqianyu99-zhejiang-university +- wandb_project=starVLA_rl_games +- rl_games.env_eval.enabled=true +- checkpoint.sync.enabled=false +- checkpoint.sync.keep_last_n=0 +- checkpoint.local.keep_last_n=1 +- checkpoint.save_best_model=false +- checkpoint.save_pt_file=false +- trainer.is_resume=false +- trainer.pretrained_checkpoint=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt +- trainer.resume_step=0 +- trainer.max_train_steps=7000 +- trainer.num_warmup_steps=100 +- trainer.save_interval=100 +- trainer.eval_interval=100 +- trainer.eval_num_batches=100 +- trainer.logging_frequency=1 +- trainer.gradient_accumulation_steps=4 +- trainer.distributed_backend=none +- trainer.learning_rate.base=2e-05 +- trainer.learning_rate.qwen_vl_interface=1e-05 +- trainer.learning_rate.action_model=0.0001 +- trainer.lr_scheduler_type=cosine_with_min_lr +- trainer.scheduler_specific_kwargs.min_lr=1e-06 +- trainer.loss_scale.vla=1.0 +- trainer.loss_scale.vlm=0.1 +- trainer.max_grad_norm=1.0 +- trainer.weight_decay=0.0 +- trainer.gradient_clipping=1.0 +- trainer.optimizer.name=AdamW +- trainer.optimizer.betas=[0.9,0.95] +- trainer.optimizer.eps=1e-08 +- trainer.optimizer.weight_decay=1e-08 +- trainer.optimizer.fused=false +- trainer.save_format=pt +- framework.name=QwenOFT +- framework.qwenvl.attn_implementation=flash_attention_2 +- framework.qwenvl.enable_gradient_checkpointing=true +- framework.action_model.action_dim=7 +- framework.action_model.action_env_dim=6 +- framework.action_model.state_dim=7 +- framework.action_model.loss_type=discrete_ce +- framework.action_model.action_horizon=1 +- framework.action_model.future_action_window_size=0 +- framework.action_model.past_action_window_size=0 +- datasets.vla_data.include_state=true +- datasets.vla_data.action_type=discrete +- datasets.vla_data.sequential_step_sampling=false +- datasets.vla_data.shuffle=true +- datasets.vla_data.action_balance.enabled=false +- datasets.vla_data.action_balance.strategy=balanced_epoch +- datasets.vla_data.action_balance.action_key=action_id +- datasets.vla_data.action_balance.target_flap_fraction=0.3 +- datasets.vla_data.action_balance.noop_id=0 +- datasets.vla_data.action_balance.flap_id=1 +- datasets.vla_data.latency_curriculum.enabled=false +- datasets.vla_data.latency_curriculum.strategy=exclusive +- datasets.vla_data.per_device_batch_size=32 +- datasets.vla_data.load_all_data_for_training=true +- datasets.vla_data.video_backend=torchvision_av +- datasets.vla_data.data_root_dir=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/data/demon_attack_fix_latency_2_200ep +- datasets.vla_data.data_mix=demon_attack_train__bridge +- datasets.vla_data.eval_data_mix=demon_attack_train__bridge__val +- framework.qwenvl.base_vlm=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/playground/Pretrained_models/Qwen3-VL-4B-Instruct +- rl_games.task=demon_attack +- rl_games.model_alias=openvla +- rl_games.initialization_mode=bridge +- rl_games.action_carrier=bridge +- rl_games.env_eval.distributed_mode=none +- rl_games.env_eval.latency.mode=single +- rl_games.env_eval.frameskip=4 +- rl_games.env_eval.image_size=224 +- rl_games.env_eval.seed=42 +- rl_games.env_eval.fixed_episode_seeds=true +- rl_games.env_eval.latency_seed_stride=0 +- rl_games.env_eval.task_seed_stride=0 +- 'rl_games.env_eval.task_description=''You are playing Demon Attack from a single + game image. Choose exactly one action from: NOOP, FIRE, RIGHT, LEFT, RIGHTFIRE, + LEFTFIRE.''' +- rl_games.env_eval.latency.values=[0] +- rl_games.env_eval.latency.prompt_map_path=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/data/demon_attack_fix_latency_2_200ep/demon_attack_train__bridge/latency_prompt_map.json +- rl_games.env_eval.mid_train.enabled=false +- rl_games.env_eval.mid_train.interval_steps=100 +- rl_games.env_eval.mid_train.num_episodes=5 +- rl_games.env_eval.mid_train.max_steps_per_episode=3600 +- rl_games.env_eval.mid_train.latencies=[0] +- rl_games.env_eval.post_train.enabled=false +- rl_games.env_eval.post_train.num_episodes=5 +- rl_games.env_eval.post_train.max_steps_per_episode=3600 +- rl_games.env_eval.post_train.latencies=[0,1,2,3,4] diff --git a/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_2_200ep/hydra/train_starvla_hydra.log b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_2_200ep/hydra/train_starvla_hydra.log new file mode 100644 index 0000000000000000000000000000000000000000..93fd244872b187c772b8a3f12975cb66f4afe041 --- /dev/null +++ b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_2_200ep/hydra/train_starvla_hydra.log @@ -0,0 +1,7752 @@ +[2026-06-17 16:14:52,625][starVLA.training.train_starvla][INFO] - VLA Training :: Warming Up +[2026-06-17 16:14:52,643][starVLA.training.train_starvla][INFO] - ✅ Configuration wrapped for access tracking +[2026-06-17 16:16:20,171][starVLA.training.train_starvla][INFO] - Preloaded model checkpoint before Accelerator/DeepSpeed initialization: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt +[2026-06-17 16:16:20,386][starVLA.training.train_starvla][INFO] - Creating VLA Dataset with Mixture `demon_attack_train__bridge` +[2026-06-17 16:16:30,521][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Behavior (data_config): ['BEHAVIOR_challenge'] +[2026-06-17 16:16:30,522][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from DOMINO: ['robotwin'] +[2026-06-17 16:16:30,522][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from DOMINO (data_config): [] +[2026-06-17 16:16:30,522][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from DOMINO (data_config): ['domino', 'domino_clean', 'domino_random', 'domino_cotrain'] +[2026-06-17 16:16:30,522][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Franka: ['custom_robot_config', 'demo_sim_franka_delta_joints', 'SO101'] +[2026-06-17 16:16:30,522][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Franka (data_config): [] +[2026-06-17 16:16:30,522][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Franka (data_config): ['custom_dataset', 'custom_dataset_2', 'demo_sim_pick_place', 'SO101_pick'] +[2026-06-17 16:16:30,522][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from LIBERO: ['libero_franka'] +[2026-06-17 16:16:30,523][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from LIBERO (data_config): [] +[2026-06-17 16:16:30,523][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from LIBERO (data_config): ['libero_all', 'libero_goal', 'multi_robot'] +[2026-06-17 16:16:30,523][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from RoboChallenge_table30v2: ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-17 16:16:30,523][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from RoboChallenge_table30v2 (data_config): ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-17 16:16:30,523][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from RoboChallenge_table30v2 (data_config): ['robochallenge_table30v2_shred_paper', 'robochallenge_table30v2_ur5_all', 'robochallenge_table30v2_arx5_all', 'robochallenge_table30v2_dosw1_all'] +[2026-06-17 16:16:30,523][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_365: ['panda_omron_robocasa365'] +[2026-06-17 16:16:30,523][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_365 (data_config): [] +[2026-06-17 16:16:30,523][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_365 (data_config): ['robocasa365_open_drawer_target_human', 'robocasa365_atomic_target_human_all', 'robocasa365_composite_target_human_all', 'robocasa365_target_human_all'] +[2026-06-17 16:16:30,524][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_tabletop: ['fourier_gr1_arms_waist'] +[2026-06-17 16:16:30,524][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_tabletop (data_config): [] +[2026-06-17 16:16:30,524][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_tabletop (data_config): ['fourier_gr1_unified_1000'] +[2026-06-17 16:16:30,533][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robotwin: ['robotwin', 'robotwin50', 'arx_x5'] +[2026-06-17 16:16:30,533][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robotwin (data_config): [] +[2026-06-17 16:16:30,533][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robotwin (data_config): ['robotwin_all', 'robotwin_all_50', 'robotwin', 'robotwin_task1', 'robotwin_task2', 'arx_x5'] +[2026-06-17 16:16:30,534][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from SimplerEnv: ['oxe_droid', 'oxe_bridge', 'oxe_rt1'] +[2026-06-17 16:16:30,534][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from SimplerEnv (data_config): [] +[2026-06-17 16:16:30,534][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from SimplerEnv (data_config): ['bridge', 'bridge_rt_1'] +[2026-06-17 16:16:30,534][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from VLA-Arena: ['vla_arena_franka'] +[2026-06-17 16:16:30,534][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from VLA-Arena (data_config): [] +[2026-06-17 16:16:30,534][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from VLA-Arena (data_config): ['vla_arena_L0_S', 'vla_arena_L0_M', 'vla_arena_L0_L'] +[2026-06-17 16:16:30,534][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from rl_games: ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-17 16:16:30,534][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from rl_games (data_config): ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-17 16:16:30,534][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from rl_games (data_config): ['flappy_train', 'flappy_train__bridge', 'flappy_mixed_latency_train', 'flappy_mixed_latency_train__bridge', 'demon_attack_train', 'demon_attack_train__bridge', 'demon_attack_mixed_latency_train', 'demon_attack_mixed_latency_train__bridge', 'deadly_corridor_train', 'deadly_corridor_train__bridge', 'deadly_corridor_mixed_latency_train', 'deadly_corridor_mixed_latency_train__bridge'] +[2026-06-17 16:16:30,684][starVLA.training.train_starvla][INFO] - Creating VLA Eval Dataset with Mixture `demon_attack_train__bridge__val` +[2026-06-17 16:16:30,947][starVLA.training.train_starvla][INFO] - 📝 Full config saved at /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/config.full.yaml +[2026-06-17 16:16:30,961][starVLA.training.train_starvla][INFO] - 📊 Accessed config snapshot saved at /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/config.yaml +[2026-06-17 16:16:30,961][starVLA.training.train_starvla][INFO] - No pretrained checkpoint provided. Starting training from scratch. +[2026-06-17 16:16:40,899][starVLA.training.train_starvla][INFO] - ***** Training Configuration ***** +[2026-06-17 16:16:40,900][starVLA.training.train_starvla][INFO] - Total optimization steps = 7000 +[2026-06-17 16:16:40,901][starVLA.training.train_starvla][INFO] - Per device batch size = 32 +[2026-06-17 16:16:40,901][starVLA.training.train_starvla][INFO] - Gradient accumulation steps = 4 +[2026-06-17 16:16:40,901][starVLA.training.train_starvla][INFO] - Total batch size = 128 +[2026-06-17 16:17:58,285][starVLA.training.train_starvla][INFO] - Step 1, Loss: {'train/loss': 1.792243480682373, 'train/grad_norm_pre_clip': 4.3253703117370605, 'timing/data': 0.04041376989334822, 'timing/model': 0.9859112780541182, 'learning_rate/qwen_vl_interface': 1.0000000000000001e-07, 'learning_rate/action_model': 1.0000000000000002e-06, 'epoch': 0.0}) +[2026-06-17 16:18:01,490][starVLA.training.train_starvla][INFO] - Step 2, Loss: {'train/loss': 1.7802510261535645, 'train/grad_norm_pre_clip': 4.615568161010742, 'timing/data': 0.037709640339016914, 'timing/model': 0.87055822648108, 'learning_rate/qwen_vl_interface': 2.0000000000000002e-07, 'learning_rate/action_model': 2.0000000000000003e-06, 'epoch': 0.0}) +[2026-06-17 16:18:04,655][starVLA.training.train_starvla][INFO] - Step 3, Loss: {'train/loss': 1.7611647844314575, 'train/grad_norm_pre_clip': 5.65817403793335, 'timing/data': 0.03811107948422432, 'timing/model': 0.8261321680620313, 'learning_rate/qwen_vl_interface': 3.0000000000000004e-07, 'learning_rate/action_model': 3e-06, 'epoch': 0.0}) +[2026-06-17 16:18:07,788][starVLA.training.train_starvla][INFO] - Step 4, Loss: {'train/loss': 1.8256126642227173, 'train/grad_norm_pre_clip': 5.852148056030273, 'timing/data': 0.03672692272812128, 'timing/model': 0.8310963148251176, 'learning_rate/qwen_vl_interface': 4.0000000000000003e-07, 'learning_rate/action_model': 4.000000000000001e-06, 'epoch': 0.0}) +[2026-06-17 16:18:11,192][starVLA.training.train_starvla][INFO] - Step 5, Loss: {'train/loss': 1.781976342201233, 'train/grad_norm_pre_clip': 3.315664291381836, 'timing/data': 0.036838886328041553, 'timing/model': 0.9313817685469985, 'learning_rate/qwen_vl_interface': 5.000000000000001e-07, 'learning_rate/action_model': 5e-06, 'epoch': 0.0}) +[2026-06-17 16:18:14,366][starVLA.training.train_starvla][INFO] - Step 6, Loss: {'train/loss': 1.7249813079833984, 'train/grad_norm_pre_clip': 6.870468616485596, 'timing/data': 0.03292835131287575, 'timing/model': 0.8099550921469927, 'learning_rate/qwen_vl_interface': 6.000000000000001e-07, 'learning_rate/action_model': 6e-06, 'epoch': 0.0}) +[2026-06-17 16:18:17,492][starVLA.training.train_starvla][INFO] - Step 7, Loss: {'train/loss': 1.7435076236724854, 'train/grad_norm_pre_clip': 5.001789569854736, 'timing/data': 0.03319021873176098, 'timing/model': 0.7880974821746349, 'learning_rate/qwen_vl_interface': 7.000000000000001e-07, 'learning_rate/action_model': 7.000000000000001e-06, 'epoch': 0.0}) +[2026-06-17 16:18:20,599][starVLA.training.train_starvla][INFO] - Step 8, Loss: {'train/loss': 1.7135062217712402, 'train/grad_norm_pre_clip': 4.445444583892822, 'timing/data': 0.04233686067163944, 'timing/model': 0.8326930329203606, 'learning_rate/qwen_vl_interface': 8.000000000000001e-07, 'learning_rate/action_model': 8.000000000000001e-06, 'epoch': 0.0}) +[2026-06-17 16:18:23,654][starVLA.training.train_starvla][INFO] - Step 9, Loss: {'train/loss': 1.6159999370574951, 'train/grad_norm_pre_clip': 6.324630260467529, 'timing/data': 0.03360306657850742, 'timing/model': 0.7898134542629123, 'learning_rate/qwen_vl_interface': 9.000000000000001e-07, 'learning_rate/action_model': 9e-06, 'epoch': 0.0}) +[2026-06-17 16:18:26,743][starVLA.training.train_starvla][INFO] - Step 10, Loss: {'train/loss': 1.7419759035110474, 'train/grad_norm_pre_clip': 1.5386171340942383, 'timing/data': 0.020537983626127243, 'timing/model': 0.8230848228558898, 'learning_rate/qwen_vl_interface': 1.0000000000000002e-06, 'learning_rate/action_model': 1e-05, 'epoch': 0.0}) +[2026-06-17 16:18:29,827][starVLA.training.train_starvla][INFO] - Step 11, Loss: {'train/loss': 1.6450773477554321, 'train/grad_norm_pre_clip': 3.1033079624176025, 'timing/data': 0.02555249910801649, 'timing/model': 0.8204378820955753, 'learning_rate/qwen_vl_interface': 1.1e-06, 'learning_rate/action_model': 1.1000000000000001e-05, 'epoch': 0.0}) +[2026-06-17 16:18:33,045][starVLA.training.train_starvla][INFO] - Step 12, Loss: {'train/loss': 1.6153218746185303, 'train/grad_norm_pre_clip': 2.2485897541046143, 'timing/data': 0.02552107162773609, 'timing/model': 0.8548289658501744, 'learning_rate/qwen_vl_interface': 1.2000000000000002e-06, 'learning_rate/action_model': 1.2e-05, 'epoch': 0.0}) +[2026-06-17 16:18:36,264][starVLA.training.train_starvla][INFO] - Step 13, Loss: {'train/loss': 1.6656020879745483, 'train/grad_norm_pre_clip': 2.2646219730377197, 'timing/data': 0.022281505167484283, 'timing/model': 0.8568652821704745, 'learning_rate/qwen_vl_interface': 1.3e-06, 'learning_rate/action_model': 1.3000000000000001e-05, 'epoch': 0.0}) +[2026-06-17 16:18:39,491][starVLA.training.train_starvla][INFO] - Step 14, Loss: {'train/loss': 1.5443183183670044, 'train/grad_norm_pre_clip': 2.563469171524048, 'timing/data': 0.0234421594068408, 'timing/model': 0.8036459675058722, 'learning_rate/qwen_vl_interface': 1.4000000000000001e-06, 'learning_rate/action_model': 1.4000000000000001e-05, 'epoch': 0.01}) +[2026-06-17 16:18:42,726][starVLA.training.train_starvla][INFO] - Step 15, Loss: {'train/loss': 1.7710343599319458, 'train/grad_norm_pre_clip': 3.142991542816162, 'timing/data': 0.025240108370780945, 'timing/model': 0.8097073789685965, 'learning_rate/qwen_vl_interface': 1.5e-06, 'learning_rate/action_model': 1.5e-05, 'epoch': 0.01}) +[2026-06-17 16:18:45,828][starVLA.training.train_starvla][INFO] - Step 16, Loss: {'train/loss': 1.6914364099502563, 'train/grad_norm_pre_clip': 2.425793170928955, 'timing/data': 0.032343728467822075, 'timing/model': 0.8186677228659391, 'learning_rate/qwen_vl_interface': 1.6000000000000001e-06, 'learning_rate/action_model': 1.6000000000000003e-05, 'epoch': 0.01}) +[2026-06-17 16:19:08,433][starVLA.training.train_starvla][INFO] - Step 17, Loss: {'train/loss': 1.7686151266098022, 'train/grad_norm_pre_clip': 3.10892915725708, 'timing/data': 0.09302416816353798, 'timing/model': 6.02606850117445, 'learning_rate/qwen_vl_interface': 1.7000000000000002e-06, 'learning_rate/action_model': 1.7000000000000003e-05, 'epoch': 0.01}) +[2026-06-17 16:19:12,338][starVLA.training.train_starvla][INFO] - Step 18, Loss: {'train/loss': 1.7111928462982178, 'train/grad_norm_pre_clip': 2.904247999191284, 'timing/data': 0.022512677125632763, 'timing/model': 0.7858115108683705, 'learning_rate/qwen_vl_interface': 1.8000000000000001e-06, 'learning_rate/action_model': 1.8e-05, 'epoch': 0.01}) +[2026-06-17 16:19:15,401][starVLA.training.train_starvla][INFO] - Step 19, Loss: {'train/loss': 1.6629067659378052, 'train/grad_norm_pre_clip': 2.5133821964263916, 'timing/data': 0.022764590568840504, 'timing/model': 0.7945792162790895, 'learning_rate/qwen_vl_interface': 1.9000000000000002e-06, 'learning_rate/action_model': 1.9e-05, 'epoch': 0.01}) +[2026-06-17 16:19:18,604][starVLA.training.train_starvla][INFO] - Step 20, Loss: {'train/loss': 1.6369028091430664, 'train/grad_norm_pre_clip': 2.572242021560669, 'timing/data': 0.02409195341169834, 'timing/model': 0.805889219045639, 'learning_rate/qwen_vl_interface': 2.0000000000000003e-06, 'learning_rate/action_model': 2e-05, 'epoch': 0.01}) +[2026-06-17 16:19:21,743][starVLA.training.train_starvla][INFO] - Step 21, Loss: {'train/loss': 1.5394610166549683, 'train/grad_norm_pre_clip': 2.152561664581299, 'timing/data': 0.023247169330716133, 'timing/model': 0.7965377550572157, 'learning_rate/qwen_vl_interface': 2.1000000000000002e-06, 'learning_rate/action_model': 2.1e-05, 'epoch': 0.01}) +[2026-06-17 16:19:24,936][starVLA.training.train_starvla][INFO] - Step 22, Loss: {'train/loss': 1.7398689985275269, 'train/grad_norm_pre_clip': 2.3478434085845947, 'timing/data': 0.03167872503399849, 'timing/model': 0.8603290840983391, 'learning_rate/qwen_vl_interface': 2.2e-06, 'learning_rate/action_model': 2.2000000000000003e-05, 'epoch': 0.01}) +[2026-06-17 16:19:28,146][starVLA.training.train_starvla][INFO] - Step 23, Loss: {'train/loss': 1.6484382152557373, 'train/grad_norm_pre_clip': 1.1665070056915283, 'timing/data': 0.02182970754802227, 'timing/model': 0.7931582871824503, 'learning_rate/qwen_vl_interface': 2.3000000000000004e-06, 'learning_rate/action_model': 2.3000000000000003e-05, 'epoch': 0.01}) +[2026-06-17 16:19:31,287][starVLA.training.train_starvla][INFO] - Step 24, Loss: {'train/loss': 1.7011473178863525, 'train/grad_norm_pre_clip': 2.2543740272521973, 'timing/data': 0.023552171885967255, 'timing/model': 0.7978473864495754, 'learning_rate/qwen_vl_interface': 2.4000000000000003e-06, 'learning_rate/action_model': 2.4e-05, 'epoch': 0.01}) +[2026-06-17 16:19:34,362][starVLA.training.train_starvla][INFO] - Step 25, Loss: {'train/loss': 1.7032147645950317, 'train/grad_norm_pre_clip': 3.3796772956848145, 'timing/data': 0.024511823430657387, 'timing/model': 0.7963272901251912, 'learning_rate/qwen_vl_interface': 2.5e-06, 'learning_rate/action_model': 2.5e-05, 'epoch': 0.01}) +[2026-06-17 16:19:37,575][starVLA.training.train_starvla][INFO] - Step 26, Loss: {'train/loss': 1.551931619644165, 'train/grad_norm_pre_clip': 2.7524166107177734, 'timing/data': 0.02565260138362646, 'timing/model': 0.7917309049516916, 'learning_rate/qwen_vl_interface': 2.6e-06, 'learning_rate/action_model': 2.6000000000000002e-05, 'epoch': 0.01}) +[2026-06-17 16:19:41,053][starVLA.training.train_starvla][INFO] - Step 27, Loss: {'train/loss': 1.6516607999801636, 'train/grad_norm_pre_clip': 4.043429374694824, 'timing/data': 0.02257411554455757, 'timing/model': 0.8314378326758742, 'learning_rate/qwen_vl_interface': 2.7000000000000004e-06, 'learning_rate/action_model': 2.7000000000000002e-05, 'epoch': 0.01}) +[2026-06-17 16:19:44,194][starVLA.training.train_starvla][INFO] - Step 28, Loss: {'train/loss': 1.7210437059402466, 'train/grad_norm_pre_clip': 2.853471517562866, 'timing/data': 0.021971612237393856, 'timing/model': 0.8370262645184994, 'learning_rate/qwen_vl_interface': 2.8000000000000003e-06, 'learning_rate/action_model': 2.8000000000000003e-05, 'epoch': 0.01}) +[2026-06-17 16:19:47,287][starVLA.training.train_starvla][INFO] - Step 29, Loss: {'train/loss': 1.7695800065994263, 'train/grad_norm_pre_clip': 2.757753849029541, 'timing/data': 0.020823461934924126, 'timing/model': 0.829136909916997, 'learning_rate/qwen_vl_interface': 2.9e-06, 'learning_rate/action_model': 2.9e-05, 'epoch': 0.01}) +[2026-06-17 16:19:50,406][starVLA.training.train_starvla][INFO] - Step 30, Loss: {'train/loss': 1.5597997903823853, 'train/grad_norm_pre_clip': 2.7949483394622803, 'timing/data': 0.02725819405168295, 'timing/model': 0.7958221323788166, 'learning_rate/qwen_vl_interface': 3e-06, 'learning_rate/action_model': 3e-05, 'epoch': 0.01}) +[2026-06-17 16:19:53,520][starVLA.training.train_starvla][INFO] - Step 31, Loss: {'train/loss': 1.619009256362915, 'train/grad_norm_pre_clip': 2.333723783493042, 'timing/data': 0.022523638792335987, 'timing/model': 0.789181693457067, 'learning_rate/qwen_vl_interface': 3.1000000000000004e-06, 'learning_rate/action_model': 3.1e-05, 'epoch': 0.01}) +[2026-06-17 16:19:56,621][starVLA.training.train_starvla][INFO] - Step 32, Loss: {'train/loss': 1.791359782218933, 'train/grad_norm_pre_clip': 3.3878369331359863, 'timing/data': 0.028630780056118965, 'timing/model': 0.7964108809828758, 'learning_rate/qwen_vl_interface': 3.2000000000000003e-06, 'learning_rate/action_model': 3.2000000000000005e-05, 'epoch': 0.01}) +[2026-06-17 16:19:59,732][starVLA.training.train_starvla][INFO] - Step 33, Loss: {'train/loss': 1.708150029182434, 'train/grad_norm_pre_clip': 3.4409406185150146, 'timing/data': 0.021845166571438313, 'timing/model': 0.7921892711892724, 'learning_rate/qwen_vl_interface': 3.3000000000000006e-06, 'learning_rate/action_model': 3.3e-05, 'epoch': 0.01}) +[2026-06-17 16:20:02,927][starVLA.training.train_starvla][INFO] - Step 34, Loss: {'train/loss': 1.608240008354187, 'train/grad_norm_pre_clip': 3.020017147064209, 'timing/data': 0.02197064831852913, 'timing/model': 0.7965751998126507, 'learning_rate/qwen_vl_interface': 3.4000000000000005e-06, 'learning_rate/action_model': 3.4000000000000007e-05, 'epoch': 0.01}) +[2026-06-17 16:20:06,043][starVLA.training.train_starvla][INFO] - Step 35, Loss: {'train/loss': 1.7323870658874512, 'train/grad_norm_pre_clip': 4.107939720153809, 'timing/data': 0.022481353022158146, 'timing/model': 0.7853721855208278, 'learning_rate/qwen_vl_interface': 3.5e-06, 'learning_rate/action_model': 3.5e-05, 'epoch': 0.01}) +[2026-06-17 16:20:09,321][starVLA.training.train_starvla][INFO] - Step 36, Loss: {'train/loss': 1.6420992612838745, 'train/grad_norm_pre_clip': 3.2538869380950928, 'timing/data': 0.024688546545803547, 'timing/model': 0.8083127159625292, 'learning_rate/qwen_vl_interface': 3.6000000000000003e-06, 'learning_rate/action_model': 3.6e-05, 'epoch': 0.01}) +[2026-06-17 16:20:12,416][starVLA.training.train_starvla][INFO] - Step 37, Loss: {'train/loss': 1.7925009727478027, 'train/grad_norm_pre_clip': 3.2257397174835205, 'timing/data': 0.025790436193346977, 'timing/model': 0.7925497852265835, 'learning_rate/qwen_vl_interface': 3.7e-06, 'learning_rate/action_model': 3.7e-05, 'epoch': 0.01}) +[2026-06-17 16:20:15,472][starVLA.training.train_starvla][INFO] - Step 38, Loss: {'train/loss': 1.6384283304214478, 'train/grad_norm_pre_clip': 1.9777384996414185, 'timing/data': 0.021660366095602512, 'timing/model': 0.7979597970843315, 'learning_rate/qwen_vl_interface': 3.8000000000000005e-06, 'learning_rate/action_model': 3.8e-05, 'epoch': 0.01}) +[2026-06-17 16:20:18,678][starVLA.training.train_starvla][INFO] - Step 39, Loss: {'train/loss': 1.6269830465316772, 'train/grad_norm_pre_clip': 2.9694175720214844, 'timing/data': 0.023189950734376907, 'timing/model': 0.7998761031776667, 'learning_rate/qwen_vl_interface': 3.900000000000001e-06, 'learning_rate/action_model': 3.9000000000000006e-05, 'epoch': 0.01}) +[2026-06-17 16:20:21,763][starVLA.training.train_starvla][INFO] - Step 40, Loss: {'train/loss': 1.6542255878448486, 'train/grad_norm_pre_clip': 2.10082745552063, 'timing/data': 0.020791521295905113, 'timing/model': 0.7845291569828987, 'learning_rate/qwen_vl_interface': 4.000000000000001e-06, 'learning_rate/action_model': 4e-05, 'epoch': 0.01}) +[2026-06-17 16:20:24,936][starVLA.training.train_starvla][INFO] - Step 41, Loss: {'train/loss': 1.5474095344543457, 'train/grad_norm_pre_clip': 2.9135522842407227, 'timing/data': 0.024010934866964817, 'timing/model': 0.7894744221121073, 'learning_rate/qwen_vl_interface': 4.1e-06, 'learning_rate/action_model': 4.1e-05, 'epoch': 0.02}) +[2026-06-17 16:20:28,053][starVLA.training.train_starvla][INFO] - Step 42, Loss: {'train/loss': 1.6709822416305542, 'train/grad_norm_pre_clip': 3.869130849838257, 'timing/data': 0.025839395821094513, 'timing/model': 0.805668537504971, 'learning_rate/qwen_vl_interface': 4.2000000000000004e-06, 'learning_rate/action_model': 4.2e-05, 'epoch': 0.02}) +[2026-06-17 16:20:31,224][starVLA.training.train_starvla][INFO] - Step 43, Loss: {'train/loss': 1.7150821685791016, 'train/grad_norm_pre_clip': 2.432732343673706, 'timing/data': 0.023812091909348965, 'timing/model': 0.7868789965286851, 'learning_rate/qwen_vl_interface': 4.3e-06, 'learning_rate/action_model': 4.3e-05, 'epoch': 0.02}) +[2026-06-17 16:20:34,406][starVLA.training.train_starvla][INFO] - Step 44, Loss: {'train/loss': 1.5697675943374634, 'train/grad_norm_pre_clip': 2.4737980365753174, 'timing/data': 0.022669443860650063, 'timing/model': 0.8958319928497076, 'learning_rate/qwen_vl_interface': 4.4e-06, 'learning_rate/action_model': 4.4000000000000006e-05, 'epoch': 0.02}) +[2026-06-17 16:20:37,533][starVLA.training.train_starvla][INFO] - Step 45, Loss: {'train/loss': 1.5734748840332031, 'train/grad_norm_pre_clip': 4.757632255554199, 'timing/data': 0.02303670533001423, 'timing/model': 0.7918567731976509, 'learning_rate/qwen_vl_interface': 4.5e-06, 'learning_rate/action_model': 4.5e-05, 'epoch': 0.02}) +[2026-06-17 16:20:40,581][starVLA.training.train_starvla][INFO] - Step 46, Loss: {'train/loss': 1.9403833150863647, 'train/grad_norm_pre_clip': 6.796453952789307, 'timing/data': 0.022983728908002377, 'timing/model': 0.7942574862390757, 'learning_rate/qwen_vl_interface': 4.600000000000001e-06, 'learning_rate/action_model': 4.600000000000001e-05, 'epoch': 0.02}) +[2026-06-17 16:20:43,784][starVLA.training.train_starvla][INFO] - Step 47, Loss: {'train/loss': 1.5575501918792725, 'train/grad_norm_pre_clip': 3.1492624282836914, 'timing/data': 0.02351243793964386, 'timing/model': 0.8291018074378371, 'learning_rate/qwen_vl_interface': 4.7e-06, 'learning_rate/action_model': 4.7e-05, 'epoch': 0.02}) +[2026-06-17 16:20:47,000][starVLA.training.train_starvla][INFO] - Step 48, Loss: {'train/loss': 1.6810085773468018, 'train/grad_norm_pre_clip': 3.556525945663452, 'timing/data': 0.02799451258033514, 'timing/model': 0.8518491368740797, 'learning_rate/qwen_vl_interface': 4.800000000000001e-06, 'learning_rate/action_model': 4.8e-05, 'epoch': 0.02}) +[2026-06-17 16:20:50,199][starVLA.training.train_starvla][INFO] - Step 49, Loss: {'train/loss': 1.5318090915679932, 'train/grad_norm_pre_clip': 2.421844005584717, 'timing/data': 0.02113322727382183, 'timing/model': 0.7905780579894781, 'learning_rate/qwen_vl_interface': 4.9000000000000005e-06, 'learning_rate/action_model': 4.9e-05, 'epoch': 0.02}) +[2026-06-17 16:20:53,482][starVLA.training.train_starvla][INFO] - Step 50, Loss: {'train/loss': 1.7437641620635986, 'train/grad_norm_pre_clip': 3.451775312423706, 'timing/data': 0.020535713993012905, 'timing/model': 0.8021724857389927, 'learning_rate/qwen_vl_interface': 5e-06, 'learning_rate/action_model': 5e-05, 'epoch': 0.02}) +[2026-06-17 16:20:59,836][starVLA.training.train_starvla][INFO] - Step 51, Loss: {'train/loss': 1.8568860292434692, 'train/grad_norm_pre_clip': 3.6676735877990723, 'timing/data': 0.027126388624310493, 'timing/model': 2.498167695477605, 'learning_rate/qwen_vl_interface': 5.1e-06, 'learning_rate/action_model': 5.1000000000000006e-05, 'epoch': 0.02}) +[2026-06-17 16:21:04,650][starVLA.training.train_starvla][INFO] - Step 52, Loss: {'train/loss': 1.6864784955978394, 'train/grad_norm_pre_clip': 4.0756635665893555, 'timing/data': 0.03245226759463549, 'timing/model': 0.8702517785131931, 'learning_rate/qwen_vl_interface': 5.2e-06, 'learning_rate/action_model': 5.2000000000000004e-05, 'epoch': 0.02}) +[2026-06-17 16:21:07,955][starVLA.training.train_starvla][INFO] - Step 53, Loss: {'train/loss': 1.7869892120361328, 'train/grad_norm_pre_clip': 4.100557327270508, 'timing/data': 0.033850052393972874, 'timing/model': 0.8651962419971824, 'learning_rate/qwen_vl_interface': 5.300000000000001e-06, 'learning_rate/action_model': 5.300000000000001e-05, 'epoch': 0.02}) +[2026-06-17 16:21:11,067][starVLA.training.train_starvla][INFO] - Step 54, Loss: {'train/loss': 1.646984338760376, 'train/grad_norm_pre_clip': 4.943227767944336, 'timing/data': 0.026288647204637527, 'timing/model': 0.8060258533805609, 'learning_rate/qwen_vl_interface': 5.400000000000001e-06, 'learning_rate/action_model': 5.4000000000000005e-05, 'epoch': 0.02}) +[2026-06-17 16:21:14,192][starVLA.training.train_starvla][INFO] - Step 55, Loss: {'train/loss': 1.699790120124817, 'train/grad_norm_pre_clip': 4.9769415855407715, 'timing/data': 0.02296334132552147, 'timing/model': 0.8088477896526456, 'learning_rate/qwen_vl_interface': 5.500000000000001e-06, 'learning_rate/action_model': 5.500000000000001e-05, 'epoch': 0.02}) +[2026-06-17 16:21:18,172][starVLA.training.train_starvla][INFO] - Step 56, Loss: {'train/loss': 1.763013243675232, 'train/grad_norm_pre_clip': 5.666271686553955, 'timing/data': 0.022836645133793354, 'timing/model': 0.8749919896945357, 'learning_rate/qwen_vl_interface': 5.600000000000001e-06, 'learning_rate/action_model': 5.6000000000000006e-05, 'epoch': 0.02}) +[2026-06-17 16:21:21,454][starVLA.training.train_starvla][INFO] - Step 57, Loss: {'train/loss': 1.6662951707839966, 'train/grad_norm_pre_clip': 3.612802505493164, 'timing/data': 0.021935065276920795, 'timing/model': 0.7877859547734261, 'learning_rate/qwen_vl_interface': 5.7e-06, 'learning_rate/action_model': 5.6999999999999996e-05, 'epoch': 0.02}) +[2026-06-17 16:21:24,569][starVLA.training.train_starvla][INFO] - Step 58, Loss: {'train/loss': 1.6304197311401367, 'train/grad_norm_pre_clip': 2.5226526260375977, 'timing/data': 0.02566283755004406, 'timing/model': 0.8181494688615203, 'learning_rate/qwen_vl_interface': 5.8e-06, 'learning_rate/action_model': 5.8e-05, 'epoch': 0.02}) +[2026-06-17 16:21:27,758][starVLA.training.train_starvla][INFO] - Step 59, Loss: {'train/loss': 1.7437186241149902, 'train/grad_norm_pre_clip': 5.000603199005127, 'timing/data': 0.026824483647942543, 'timing/model': 0.8641260033473372, 'learning_rate/qwen_vl_interface': 5.9e-06, 'learning_rate/action_model': 5.9e-05, 'epoch': 0.02}) +[2026-06-17 16:21:31,048][starVLA.training.train_starvla][INFO] - Step 60, Loss: {'train/loss': 1.5347330570220947, 'train/grad_norm_pre_clip': 4.34968900680542, 'timing/data': 0.026952693238854408, 'timing/model': 0.8712187213823199, 'learning_rate/qwen_vl_interface': 6e-06, 'learning_rate/action_model': 6e-05, 'epoch': 0.02}) +[2026-06-17 16:21:34,256][starVLA.training.train_starvla][INFO] - Step 61, Loss: {'train/loss': 1.7708845138549805, 'train/grad_norm_pre_clip': 6.123332977294922, 'timing/data': 0.02780220191925764, 'timing/model': 0.8584810765460134, 'learning_rate/qwen_vl_interface': 6.1e-06, 'learning_rate/action_model': 6.1e-05, 'epoch': 0.02}) +[2026-06-17 16:21:37,355][starVLA.training.train_starvla][INFO] - Step 62, Loss: {'train/loss': 1.6952179670333862, 'train/grad_norm_pre_clip': 6.605747222900391, 'timing/data': 0.02238460537046194, 'timing/model': 0.8118153167888522, 'learning_rate/qwen_vl_interface': 6.200000000000001e-06, 'learning_rate/action_model': 6.2e-05, 'epoch': 0.02}) +[2026-06-17 16:21:40,575][starVLA.training.train_starvla][INFO] - Step 63, Loss: {'train/loss': 1.5279414653778076, 'train/grad_norm_pre_clip': 2.8175406455993652, 'timing/data': 0.024439027532935143, 'timing/model': 0.847420652396977, 'learning_rate/qwen_vl_interface': 6.300000000000001e-06, 'learning_rate/action_model': 6.3e-05, 'epoch': 0.02}) +[2026-06-17 16:21:43,771][starVLA.training.train_starvla][INFO] - Step 64, Loss: {'train/loss': 1.48135244846344, 'train/grad_norm_pre_clip': 2.716461658477783, 'timing/data': 0.024211661890149117, 'timing/model': 0.8970412351191044, 'learning_rate/qwen_vl_interface': 6.4000000000000006e-06, 'learning_rate/action_model': 6.400000000000001e-05, 'epoch': 0.02}) +[2026-06-17 16:21:46,899][starVLA.training.train_starvla][INFO] - Step 65, Loss: {'train/loss': 1.635460615158081, 'train/grad_norm_pre_clip': 4.135998725891113, 'timing/data': 0.0215902216732502, 'timing/model': 0.8033611159771681, 'learning_rate/qwen_vl_interface': 6.5000000000000004e-06, 'learning_rate/action_model': 6.500000000000001e-05, 'epoch': 0.02}) +[2026-06-17 16:21:50,039][starVLA.training.train_starvla][INFO] - Step 66, Loss: {'train/loss': 1.7056502103805542, 'train/grad_norm_pre_clip': 3.535521984100342, 'timing/data': 0.024058165960013866, 'timing/model': 0.7983755981549621, 'learning_rate/qwen_vl_interface': 6.600000000000001e-06, 'learning_rate/action_model': 6.6e-05, 'epoch': 0.02}) +[2026-06-17 16:21:54,497][starVLA.training.train_starvla][INFO] - Step 67, Loss: {'train/loss': 1.7558696269989014, 'train/grad_norm_pre_clip': 4.209866523742676, 'timing/data': 0.021195556968450546, 'timing/model': 2.065952750854194, 'learning_rate/qwen_vl_interface': 6.700000000000001e-06, 'learning_rate/action_model': 6.7e-05, 'epoch': 0.02}) +[2026-06-17 16:22:02,375][starVLA.training.train_starvla][INFO] - Step 68, Loss: {'train/loss': 1.5238902568817139, 'train/grad_norm_pre_clip': 3.1115007400512695, 'timing/data': 0.027554795145988464, 'timing/model': 1.592174748890102, 'learning_rate/qwen_vl_interface': 6.800000000000001e-06, 'learning_rate/action_model': 6.800000000000001e-05, 'epoch': 0.02}) +[2026-06-17 16:22:07,218][starVLA.training.train_starvla][INFO] - Step 69, Loss: {'train/loss': 1.7991797924041748, 'train/grad_norm_pre_clip': 5.771025657653809, 'timing/data': 0.022957815788686275, 'timing/model': 0.8097862983122468, 'learning_rate/qwen_vl_interface': 6.9e-06, 'learning_rate/action_model': 6.9e-05, 'epoch': 0.03}) +[2026-06-17 16:22:10,307][starVLA.training.train_starvla][INFO] - Step 70, Loss: {'train/loss': 1.6224994659423828, 'train/grad_norm_pre_clip': 4.180534839630127, 'timing/data': 0.02416158840060234, 'timing/model': 0.7944886591285467, 'learning_rate/qwen_vl_interface': 7e-06, 'learning_rate/action_model': 7e-05, 'epoch': 0.03}) +[2026-06-17 16:22:13,396][starVLA.training.train_starvla][INFO] - Step 71, Loss: {'train/loss': 1.7083908319473267, 'train/grad_norm_pre_clip': 5.385356426239014, 'timing/data': 0.020782340317964554, 'timing/model': 0.8292536893859506, 'learning_rate/qwen_vl_interface': 7.100000000000001e-06, 'learning_rate/action_model': 7.1e-05, 'epoch': 0.03}) +[2026-06-17 16:22:16,689][starVLA.training.train_starvla][INFO] - Step 72, Loss: {'train/loss': 1.6412774324417114, 'train/grad_norm_pre_clip': 4.940762042999268, 'timing/data': 0.024267803877592087, 'timing/model': 0.8554329518228769, 'learning_rate/qwen_vl_interface': 7.2000000000000005e-06, 'learning_rate/action_model': 7.2e-05, 'epoch': 0.03}) +[2026-06-17 16:22:19,946][starVLA.training.train_starvla][INFO] - Step 73, Loss: {'train/loss': 1.5463061332702637, 'train/grad_norm_pre_clip': 4.878565311431885, 'timing/data': 0.022648312151432037, 'timing/model': 0.8283470505848527, 'learning_rate/qwen_vl_interface': 7.3e-06, 'learning_rate/action_model': 7.3e-05, 'epoch': 0.03}) +[2026-06-17 16:22:25,071][starVLA.training.train_starvla][INFO] - Step 74, Loss: {'train/loss': 1.6530342102050781, 'train/grad_norm_pre_clip': 9.739480972290039, 'timing/data': 0.02160919737070799, 'timing/model': 2.342924471013248, 'learning_rate/qwen_vl_interface': 7.4e-06, 'learning_rate/action_model': 7.4e-05, 'epoch': 0.03}) +[2026-06-17 16:22:30,204][starVLA.training.train_starvla][INFO] - Step 75, Loss: {'train/loss': 1.4686312675476074, 'train/grad_norm_pre_clip': 4.246958255767822, 'timing/data': 0.02310222014784813, 'timing/model': 0.8404973465949297, 'learning_rate/qwen_vl_interface': 7.500000000000001e-06, 'learning_rate/action_model': 7.500000000000001e-05, 'epoch': 0.03}) +[2026-06-17 16:22:33,275][starVLA.training.train_starvla][INFO] - Step 76, Loss: {'train/loss': 1.5560868978500366, 'train/grad_norm_pre_clip': 12.390497207641602, 'timing/data': 0.02425209805369377, 'timing/model': 0.7895642509683967, 'learning_rate/qwen_vl_interface': 7.600000000000001e-06, 'learning_rate/action_model': 7.6e-05, 'epoch': 0.03}) +[2026-06-17 16:22:36,336][starVLA.training.train_starvla][INFO] - Step 77, Loss: {'train/loss': 1.7954655885696411, 'train/grad_norm_pre_clip': 7.022336006164551, 'timing/data': 0.02615372370928526, 'timing/model': 0.7981143556535244, 'learning_rate/qwen_vl_interface': 7.7e-06, 'learning_rate/action_model': 7.7e-05, 'epoch': 0.03}) +[2026-06-17 16:22:39,574][starVLA.training.train_starvla][INFO] - Step 78, Loss: {'train/loss': 1.853243112564087, 'train/grad_norm_pre_clip': 8.982224464416504, 'timing/data': 0.022673252038657665, 'timing/model': 0.8308469010517001, 'learning_rate/qwen_vl_interface': 7.800000000000002e-06, 'learning_rate/action_model': 7.800000000000001e-05, 'epoch': 0.03}) +[2026-06-17 16:22:43,928][starVLA.training.train_starvla][INFO] - Step 79, Loss: {'train/loss': 1.6926002502441406, 'train/grad_norm_pre_clip': 7.552641868591309, 'timing/data': 0.021635604090988636, 'timing/model': 2.044260563328862, 'learning_rate/qwen_vl_interface': 7.9e-06, 'learning_rate/action_model': 7.900000000000001e-05, 'epoch': 0.03}) +[2026-06-17 16:22:47,170][starVLA.training.train_starvla][INFO] - Step 80, Loss: {'train/loss': 1.379408836364746, 'train/grad_norm_pre_clip': 5.394141674041748, 'timing/data': 0.02534778881818056, 'timing/model': 0.8438224261626601, 'learning_rate/qwen_vl_interface': 8.000000000000001e-06, 'learning_rate/action_model': 8e-05, 'epoch': 0.03}) +[2026-06-17 16:22:50,371][starVLA.training.train_starvla][INFO] - Step 81, Loss: {'train/loss': 1.8190690279006958, 'train/grad_norm_pre_clip': 9.736157417297363, 'timing/data': 0.02739798743277788, 'timing/model': 0.8904277104884386, 'learning_rate/qwen_vl_interface': 8.1e-06, 'learning_rate/action_model': 8.1e-05, 'epoch': 0.03}) +[2026-06-17 16:22:53,459][starVLA.training.train_starvla][INFO] - Step 82, Loss: {'train/loss': 1.5545865297317505, 'train/grad_norm_pre_clip': 6.8974223136901855, 'timing/data': 0.023864060640335083, 'timing/model': 0.7910584909841418, 'learning_rate/qwen_vl_interface': 8.2e-06, 'learning_rate/action_model': 8.2e-05, 'epoch': 0.03}) +[2026-06-17 16:22:56,531][starVLA.training.train_starvla][INFO] - Step 83, Loss: {'train/loss': 1.6773147583007812, 'train/grad_norm_pre_clip': 6.5771684646606445, 'timing/data': 0.02170472964644432, 'timing/model': 0.7972289556637406, 'learning_rate/qwen_vl_interface': 8.3e-06, 'learning_rate/action_model': 8.3e-05, 'epoch': 0.03}) +[2026-06-17 16:22:59,654][starVLA.training.train_starvla][INFO] - Step 84, Loss: {'train/loss': 1.6990679502487183, 'train/grad_norm_pre_clip': 5.621453285217285, 'timing/data': 0.02723066322505474, 'timing/model': 0.797598103992641, 'learning_rate/qwen_vl_interface': 8.400000000000001e-06, 'learning_rate/action_model': 8.4e-05, 'epoch': 0.03}) +[2026-06-17 16:23:03,064][starVLA.training.train_starvla][INFO] - Step 85, Loss: {'train/loss': 1.74220871925354, 'train/grad_norm_pre_clip': 7.366023063659668, 'timing/data': 0.02102075796574354, 'timing/model': 0.7947074519470334, 'learning_rate/qwen_vl_interface': 8.5e-06, 'learning_rate/action_model': 8.5e-05, 'epoch': 0.03}) +[2026-06-17 16:23:07,246][starVLA.training.train_starvla][INFO] - Step 86, Loss: {'train/loss': 1.7728632688522339, 'train/grad_norm_pre_clip': 3.2729530334472656, 'timing/data': 0.02531088050454855, 'timing/model': 0.8696981677785516, 'learning_rate/qwen_vl_interface': 8.6e-06, 'learning_rate/action_model': 8.6e-05, 'epoch': 0.03}) +[2026-06-17 16:23:12,239][starVLA.training.train_starvla][INFO] - Step 87, Loss: {'train/loss': 1.7488709688186646, 'train/grad_norm_pre_clip': 8.792497634887695, 'timing/data': 0.024219156242907047, 'timing/model': 0.8345595337450504, 'learning_rate/qwen_vl_interface': 8.700000000000001e-06, 'learning_rate/action_model': 8.7e-05, 'epoch': 0.03}) +[2026-06-17 16:23:15,360][starVLA.training.train_starvla][INFO] - Step 88, Loss: {'train/loss': 1.6709814071655273, 'train/grad_norm_pre_clip': 3.8280537128448486, 'timing/data': 0.02529300097376108, 'timing/model': 0.8050066949799657, 'learning_rate/qwen_vl_interface': 8.8e-06, 'learning_rate/action_model': 8.800000000000001e-05, 'epoch': 0.03}) +[2026-06-17 16:23:18,614][starVLA.training.train_starvla][INFO] - Step 89, Loss: {'train/loss': 1.5433275699615479, 'train/grad_norm_pre_clip': 3.8964710235595703, 'timing/data': 0.027350242249667645, 'timing/model': 0.8793905582278967, 'learning_rate/qwen_vl_interface': 8.900000000000001e-06, 'learning_rate/action_model': 8.900000000000001e-05, 'epoch': 0.03}) +[2026-06-17 16:23:21,701][starVLA.training.train_starvla][INFO] - Step 90, Loss: {'train/loss': 1.7946288585662842, 'train/grad_norm_pre_clip': 5.034336566925049, 'timing/data': 0.026290835812687874, 'timing/model': 0.8197533786296844, 'learning_rate/qwen_vl_interface': 9e-06, 'learning_rate/action_model': 9e-05, 'epoch': 0.03}) +[2026-06-17 16:23:24,934][starVLA.training.train_starvla][INFO] - Step 91, Loss: {'train/loss': 1.5132111310958862, 'train/grad_norm_pre_clip': 3.72821044921875, 'timing/data': 0.023827344179153442, 'timing/model': 0.864637597464025, 'learning_rate/qwen_vl_interface': 9.100000000000001e-06, 'learning_rate/action_model': 9.1e-05, 'epoch': 0.03}) +[2026-06-17 16:23:28,048][starVLA.training.train_starvla][INFO] - Step 92, Loss: {'train/loss': 1.5738757848739624, 'train/grad_norm_pre_clip': 4.426362991333008, 'timing/data': 0.02305196411907673, 'timing/model': 0.8035360788926482, 'learning_rate/qwen_vl_interface': 9.200000000000002e-06, 'learning_rate/action_model': 9.200000000000001e-05, 'epoch': 0.03}) +[2026-06-17 16:23:31,256][starVLA.training.train_starvla][INFO] - Step 93, Loss: {'train/loss': 1.7326769828796387, 'train/grad_norm_pre_clip': 7.117413520812988, 'timing/data': 0.029002790339291096, 'timing/model': 0.8655656948685646, 'learning_rate/qwen_vl_interface': 9.3e-06, 'learning_rate/action_model': 9.300000000000001e-05, 'epoch': 0.03}) +[2026-06-17 16:23:34,361][starVLA.training.train_starvla][INFO] - Step 94, Loss: {'train/loss': 1.4742579460144043, 'train/grad_norm_pre_clip': 3.3917205333709717, 'timing/data': 0.024448184296488762, 'timing/model': 0.8302502054721117, 'learning_rate/qwen_vl_interface': 9.4e-06, 'learning_rate/action_model': 9.4e-05, 'epoch': 0.03}) +[2026-06-17 16:23:37,552][starVLA.training.train_starvla][INFO] - Step 95, Loss: {'train/loss': 1.7031787633895874, 'train/grad_norm_pre_clip': 6.3037543296813965, 'timing/data': 0.02339362818747759, 'timing/model': 0.8941332697868347, 'learning_rate/qwen_vl_interface': 9.5e-06, 'learning_rate/action_model': 9.5e-05, 'epoch': 0.03}) +[2026-06-17 16:23:40,696][starVLA.training.train_starvla][INFO] - Step 96, Loss: {'train/loss': 1.6016380786895752, 'train/grad_norm_pre_clip': 6.73072624206543, 'timing/data': 0.024804183281958103, 'timing/model': 0.7962460136041045, 'learning_rate/qwen_vl_interface': 9.600000000000001e-06, 'learning_rate/action_model': 9.6e-05, 'epoch': 0.04}) +[2026-06-17 16:23:43,818][starVLA.training.train_starvla][INFO] - Step 97, Loss: {'train/loss': 1.614386796951294, 'train/grad_norm_pre_clip': 6.971299648284912, 'timing/data': 0.02515214029699564, 'timing/model': 0.7956121060997248, 'learning_rate/qwen_vl_interface': 9.7e-06, 'learning_rate/action_model': 9.7e-05, 'epoch': 0.04}) +[2026-06-17 16:23:47,037][starVLA.training.train_starvla][INFO] - Step 98, Loss: {'train/loss': 1.6632912158966064, 'train/grad_norm_pre_clip': 7.695398330688477, 'timing/data': 0.022829116322100163, 'timing/model': 0.7978403251618147, 'learning_rate/qwen_vl_interface': 9.800000000000001e-06, 'learning_rate/action_model': 9.8e-05, 'epoch': 0.04}) +[2026-06-17 16:23:50,295][starVLA.training.train_starvla][INFO] - Step 99, Loss: {'train/loss': 1.6339178085327148, 'train/grad_norm_pre_clip': 9.313483238220215, 'timing/data': 0.02633517049252987, 'timing/model': 0.9054313097149134, 'learning_rate/qwen_vl_interface': 9.9e-06, 'learning_rate/action_model': 9.900000000000001e-05, 'epoch': 0.04}) +[2026-06-17 16:25:35,239][root][INFO] - gcc -pthread -B /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/include -fPIC -O2 -isystem /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/include -fPIC -c /tmp/tmpfdghbefs/test.c -o /tmp/tmpfdghbefs/test.o +[2026-06-17 16:25:35,265][root][INFO] - gcc -pthread -B /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/compiler_compat /tmp/tmpfdghbefs/test.o -laio -o /tmp/tmpfdghbefs/a.out +[2026-06-17 16:25:36,137][root][INFO] - gcc -pthread -B /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/include -fPIC -O2 -isystem /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/include -fPIC -c /tmp/tmpp8c6zmdc/test.c -o /tmp/tmpp8c6zmdc/test.o +[2026-06-17 16:25:36,156][root][INFO] - gcc -pthread -B /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/compiler_compat /tmp/tmpp8c6zmdc/test.o -L/usr/local/cuda -L/usr/local/cuda/lib64 -lcufile -o /tmp/tmpp8c6zmdc/a.out +[2026-06-17 16:25:36,207][root][INFO] - gcc -pthread -B /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/include -fPIC -O2 -isystem /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/include -fPIC -c /tmp/tmp1y61kp17/test.c -o /tmp/tmp1y61kp17/test.o +[2026-06-17 16:25:36,225][root][INFO] - gcc -pthread -B /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/compiler_compat /tmp/tmp1y61kp17/test.o -laio -o /tmp/tmp1y61kp17/a.out +[2026-06-17 16:27:26,855][starVLA.training.train_starvla][INFO] - Step 100, Loss: {'train/loss': 1.4966994524002075, 'train/grad_norm_pre_clip': 8.127405166625977, 'eval/loss': 1.6123631000518799, 'eval/latency_2/loss': 1.6123631000518799, 'eval/demon_fire_precision': 0.7179165359272043, 'eval/demon_fire_recall': 0.9573221757322176, 'eval/demon_fire_f1': 0.8205128205128205, 'eval/demon_fire_rate_model': 0.8867557039510295, 'eval/demon_fire_rate_teacher': 0.6649972175848636, 'eval/demon_left_precision': 0.0, 'eval/demon_left_recall': 0.0, 'eval/demon_left_f1': 0.0, 'eval/demon_left_rate_model': 0.0, 'eval/demon_left_rate_teacher': 0.2643294379521425, 'eval/demon_right_precision': 0.538083538083538, 'eval/demon_right_recall': 0.24745762711864408, 'eval/demon_right_f1': 0.3390092879256966, 'eval/demon_right_rate_model': 0.1132442960489705, 'eval/demon_right_rate_teacher': 0.24624373956594323, 'eval/demon_move_f1': 0.1695046439628483, 'eval/demon_cc_f1': 0.3812445938667929, 'eval/latency_2/demon_fire_precision': 0.7179165359272043, 'eval/latency_2/demon_fire_recall': 0.9573221757322176, 'eval/latency_2/demon_fire_f1': 0.8205128205128205, 'eval/latency_2/demon_fire_rate_model': 0.8867557039510295, 'eval/latency_2/demon_fire_rate_teacher': 0.6649972175848636, 'eval/latency_2/demon_left_precision': 0.0, 'eval/latency_2/demon_left_recall': 0.0, 'eval/latency_2/demon_left_f1': 0.0, 'eval/latency_2/demon_left_rate_model': 0.0, 'eval/latency_2/demon_left_rate_teacher': 0.2643294379521425, 'eval/latency_2/demon_right_precision': 0.538083538083538, 'eval/latency_2/demon_right_recall': 0.24745762711864408, 'eval/latency_2/demon_right_f1': 0.3390092879256966, 'eval/latency_2/demon_right_rate_model': 0.1132442960489705, 'eval/latency_2/demon_right_rate_teacher': 0.24624373956594323, 'eval/latency_2/demon_move_f1': 0.1695046439628483, 'eval/latency_2/demon_cc_f1': 0.3812445938667929, 'timing/data': 0.022475861944258213, 'timing/model': 0.7936836024746299, 'learning_rate/qwen_vl_interface': 1e-05, 'learning_rate/action_model': 0.0001, 'epoch': 0.04}) +[2026-06-17 16:27:26,880][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_100_state +[2026-06-17 16:27:26,887][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 16:27:36,843][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_100_state/model.safetensors +[2026-06-17 16:27:51,419][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_100_state/optimizer.bin +[2026-06-17 16:27:51,423][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_100_state/sampler.bin +[2026-06-17 16:27:51,423][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_100_state/sampler_1.bin +[2026-06-17 16:27:51,427][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_100_state/random_states_0.pkl +[2026-06-17 16:27:51,434][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 100 +[2026-06-17 16:27:51,435][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 16:27:51,452][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 16:27:59,738][starVLA.training.train_starvla][INFO] - Step 101, Loss: {'train/loss': 1.6528866291046143, 'train/grad_norm_pre_clip': 6.462955951690674, 'timing/data': 0.025691884569823742, 'timing/model': 1.1592171024531126, 'learning_rate/qwen_vl_interface': 9.999999507659945e-06, 'learning_rate/action_model': 9.999999507659945e-05, 'epoch': 0.04}) +[2026-06-17 16:28:04,262][starVLA.training.train_starvla][INFO] - Step 102, Loss: {'train/loss': 1.5109602212905884, 'train/grad_norm_pre_clip': 9.699865341186523, 'timing/data': 0.02285944204777479, 'timing/model': 1.4849164634943008, 'learning_rate/qwen_vl_interface': 9.99999803063988e-06, 'learning_rate/action_model': 9.99999803063988e-05, 'epoch': 0.04}) +[2026-06-17 16:28:08,873][starVLA.training.train_starvla][INFO] - Step 103, Loss: {'train/loss': 1.5244789123535156, 'train/grad_norm_pre_clip': 8.333252906799316, 'timing/data': 0.0258086659014225, 'timing/model': 1.497555493377149, 'learning_rate/qwen_vl_interface': 9.999995568940113e-06, 'learning_rate/action_model': 9.999995568940113e-05, 'epoch': 0.04}) +[2026-06-17 16:28:13,051][starVLA.training.train_starvla][INFO] - Step 104, Loss: {'train/loss': 1.6584199666976929, 'train/grad_norm_pre_clip': 9.384261131286621, 'timing/data': 0.020576628856360912, 'timing/model': 1.119282964617014, 'learning_rate/qwen_vl_interface': 9.999992122561152e-06, 'learning_rate/action_model': 9.999992122561152e-05, 'epoch': 0.04}) +[2026-06-17 16:28:17,631][starVLA.training.train_starvla][INFO] - Step 105, Loss: {'train/loss': 1.6709537506103516, 'train/grad_norm_pre_clip': 8.540838241577148, 'timing/data': 0.020662525668740273, 'timing/model': 1.107022387906909, 'learning_rate/qwen_vl_interface': 9.999987691503714e-06, 'learning_rate/action_model': 9.999987691503714e-05, 'epoch': 0.04}) +[2026-06-17 16:28:22,124][starVLA.training.train_starvla][INFO] - Step 106, Loss: {'train/loss': 1.477189540863037, 'train/grad_norm_pre_clip': 8.94442367553711, 'timing/data': 0.0214961227029562, 'timing/model': 1.0341777252033353, 'learning_rate/qwen_vl_interface': 9.999982275768716e-06, 'learning_rate/action_model': 9.999982275768715e-05, 'epoch': 0.04}) +[2026-06-17 16:28:26,614][starVLA.training.train_starvla][INFO] - Step 107, Loss: {'train/loss': 1.6286585330963135, 'train/grad_norm_pre_clip': 8.27634334564209, 'timing/data': 0.019442527554929256, 'timing/model': 1.1058123363181949, 'learning_rate/qwen_vl_interface': 9.99997587535728e-06, 'learning_rate/action_model': 9.99997587535728e-05, 'epoch': 0.04}) +[2026-06-17 16:28:30,743][starVLA.training.train_starvla][INFO] - Step 108, Loss: {'train/loss': 1.4940645694732666, 'train/grad_norm_pre_clip': 10.14163589477539, 'timing/data': 0.020231859758496284, 'timing/model': 1.0341834630817175, 'learning_rate/qwen_vl_interface': 9.999968490270735e-06, 'learning_rate/action_model': 9.999968490270735e-05, 'epoch': 0.04}) +[2026-06-17 16:28:35,066][starVLA.training.train_starvla][INFO] - Step 109, Loss: {'train/loss': 1.7085579633712769, 'train/grad_norm_pre_clip': 13.185009956359863, 'timing/data': 0.021094737574458122, 'timing/model': 1.041815709322691, 'learning_rate/qwen_vl_interface': 9.99996012051061e-06, 'learning_rate/action_model': 9.99996012051061e-05, 'epoch': 0.04}) +[2026-06-17 16:28:39,474][starVLA.training.train_starvla][INFO] - Step 110, Loss: {'train/loss': 1.8132392168045044, 'train/grad_norm_pre_clip': 9.438896179199219, 'timing/data': 0.020304004661738873, 'timing/model': 1.0322597660124302, 'learning_rate/qwen_vl_interface': 9.999950766078641e-06, 'learning_rate/action_model': 9.99995076607864e-05, 'epoch': 0.04}) +[2026-06-17 16:28:43,911][starVLA.training.train_starvla][INFO] - Step 111, Loss: {'train/loss': 1.6485178470611572, 'train/grad_norm_pre_clip': 11.960792541503906, 'timing/data': 0.01991551462560892, 'timing/model': 1.0926348995417356, 'learning_rate/qwen_vl_interface': 9.999940426976768e-06, 'learning_rate/action_model': 9.999940426976768e-05, 'epoch': 0.04}) +[2026-06-17 16:28:48,428][starVLA.training.train_starvla][INFO] - Step 112, Loss: {'train/loss': 1.4848521947860718, 'train/grad_norm_pre_clip': 6.898587703704834, 'timing/data': 0.019226315431296825, 'timing/model': 1.4444121746346354, 'learning_rate/qwen_vl_interface': 9.999929103207133e-06, 'learning_rate/action_model': 9.999929103207132e-05, 'epoch': 0.04}) +[2026-06-17 16:28:53,015][starVLA.training.train_starvla][INFO] - Step 113, Loss: {'train/loss': 1.5786000490188599, 'train/grad_norm_pre_clip': 7.609907627105713, 'timing/data': 0.02038678340613842, 'timing/model': 1.509980564005673, 'learning_rate/qwen_vl_interface': 9.999916794772083e-06, 'learning_rate/action_model': 9.999916794772082e-05, 'epoch': 0.04}) +[2026-06-17 16:28:57,129][starVLA.training.train_starvla][INFO] - Step 114, Loss: {'train/loss': 1.94257652759552, 'train/grad_norm_pre_clip': 11.206734657287598, 'timing/data': 0.020700113847851753, 'timing/model': 1.0568117620423436, 'learning_rate/qwen_vl_interface': 9.99990350167417e-06, 'learning_rate/action_model': 9.999903501674169e-05, 'epoch': 0.04}) +[2026-06-17 16:29:01,519][starVLA.training.train_starvla][INFO] - Step 115, Loss: {'train/loss': 1.6161280870437622, 'train/grad_norm_pre_clip': 5.773682594299316, 'timing/data': 0.023374421522021294, 'timing/model': 1.0501992208883166, 'learning_rate/qwen_vl_interface': 9.99988922391615e-06, 'learning_rate/action_model': 9.99988922391615e-05, 'epoch': 0.04}) +[2026-06-17 16:29:06,071][starVLA.training.train_starvla][INFO] - Step 116, Loss: {'train/loss': 1.5570783615112305, 'train/grad_norm_pre_clip': 4.879621505737305, 'timing/data': 0.02048734575510025, 'timing/model': 1.0385186458006501, 'learning_rate/qwen_vl_interface': 9.999873961500985e-06, 'learning_rate/action_model': 9.999873961500984e-05, 'epoch': 0.04}) +[2026-06-17 16:29:10,638][starVLA.training.train_starvla][INFO] - Step 117, Loss: {'train/loss': 1.4713480472564697, 'train/grad_norm_pre_clip': 3.665884494781494, 'timing/data': 0.02436453104019165, 'timing/model': 1.0994842443615198, 'learning_rate/qwen_vl_interface': 9.999857714431835e-06, 'learning_rate/action_model': 9.999857714431834e-05, 'epoch': 0.04}) +[2026-06-17 16:29:15,970][starVLA.training.train_starvla][INFO] - Step 118, Loss: {'train/loss': 1.7398525476455688, 'train/grad_norm_pre_clip': 7.800239086151123, 'timing/data': 0.020563621073961258, 'timing/model': 2.249376431107521, 'learning_rate/qwen_vl_interface': 9.999840482712069e-06, 'learning_rate/action_model': 9.999840482712068e-05, 'epoch': 0.04}) +[2026-06-17 16:29:21,071][starVLA.training.train_starvla][INFO] - Step 119, Loss: {'train/loss': 1.6748896837234497, 'train/grad_norm_pre_clip': 6.752143383026123, 'timing/data': 0.025938390754163265, 'timing/model': 1.092333272099495, 'learning_rate/qwen_vl_interface': 9.99982226634526e-06, 'learning_rate/action_model': 9.99982226634526e-05, 'epoch': 0.04}) +[2026-06-17 16:29:27,051][starVLA.training.train_starvla][INFO] - Step 120, Loss: {'train/loss': 1.7439732551574707, 'train/grad_norm_pre_clip': 7.484488010406494, 'timing/data': 0.028674479573965073, 'timing/model': 2.115248172543943, 'learning_rate/qwen_vl_interface': 9.999803065335185e-06, 'learning_rate/action_model': 9.999803065335185e-05, 'epoch': 0.04}) +[2026-06-17 16:29:33,377][starVLA.training.train_starvla][INFO] - Step 121, Loss: {'train/loss': 1.4189409017562866, 'train/grad_norm_pre_clip': 6.830794334411621, 'timing/data': 0.023052592761814594, 'timing/model': 1.0310673397034407, 'learning_rate/qwen_vl_interface': 9.999782879685825e-06, 'learning_rate/action_model': 9.999782879685824e-05, 'epoch': 0.04}) +[2026-06-17 16:29:38,020][starVLA.training.train_starvla][INFO] - Step 122, Loss: {'train/loss': 1.4433739185333252, 'train/grad_norm_pre_clip': 7.552727222442627, 'timing/data': 0.02151287253946066, 'timing/model': 1.1636616671457887, 'learning_rate/qwen_vl_interface': 9.99976170940136e-06, 'learning_rate/action_model': 9.99976170940136e-05, 'epoch': 0.04}) +[2026-06-17 16:29:43,841][starVLA.training.train_starvla][INFO] - Step 123, Loss: {'train/loss': 1.4076628684997559, 'train/grad_norm_pre_clip': 5.811940670013428, 'timing/data': 0.020739993080496788, 'timing/model': 1.5489404276013374, 'learning_rate/qwen_vl_interface': 9.999739554486184e-06, 'learning_rate/action_model': 9.999739554486182e-05, 'epoch': 0.05}) +[2026-06-17 16:29:48,002][starVLA.training.train_starvla][INFO] - Step 124, Loss: {'train/loss': 1.31047785282135, 'train/grad_norm_pre_clip': 5.389739036560059, 'timing/data': 0.021538703702390194, 'timing/model': 1.0745561812072992, 'learning_rate/qwen_vl_interface': 9.999716414944887e-06, 'learning_rate/action_model': 9.999716414944886e-05, 'epoch': 0.05}) +[2026-06-17 16:29:53,878][starVLA.training.train_starvla][INFO] - Step 125, Loss: {'train/loss': 1.6190388202667236, 'train/grad_norm_pre_clip': 8.308838844299316, 'timing/data': 0.020814260467886925, 'timing/model': 1.0404261518269777, 'learning_rate/qwen_vl_interface': 9.999692290782264e-06, 'learning_rate/action_model': 9.999692290782263e-05, 'epoch': 0.05}) +[2026-06-17 16:29:58,424][starVLA.training.train_starvla][INFO] - Step 126, Loss: {'train/loss': 1.609254240989685, 'train/grad_norm_pre_clip': 6.3175482749938965, 'timing/data': 0.020671097561717033, 'timing/model': 1.044785455800593, 'learning_rate/qwen_vl_interface': 9.999667182003321e-06, 'learning_rate/action_model': 9.999667182003321e-05, 'epoch': 0.05}) +[2026-06-17 16:30:02,818][starVLA.training.train_starvla][INFO] - Step 127, Loss: {'train/loss': 1.5247108936309814, 'train/grad_norm_pre_clip': 5.689061641693115, 'timing/data': 0.020141752436757088, 'timing/model': 1.3541545392945409, 'learning_rate/qwen_vl_interface': 9.99964108861326e-06, 'learning_rate/action_model': 9.99964108861326e-05, 'epoch': 0.05}) +[2026-06-17 16:30:07,173][starVLA.training.train_starvla][INFO] - Step 128, Loss: {'train/loss': 1.3214874267578125, 'train/grad_norm_pre_clip': 5.801607131958008, 'timing/data': 0.020306126214563847, 'timing/model': 1.3165500229224563, 'learning_rate/qwen_vl_interface': 9.99961401061749e-06, 'learning_rate/action_model': 9.99961401061749e-05, 'epoch': 0.05}) +[2026-06-17 16:30:11,325][starVLA.training.train_starvla][INFO] - Step 129, Loss: {'train/loss': 1.6779240369796753, 'train/grad_norm_pre_clip': 7.672790050506592, 'timing/data': 0.020400812849402428, 'timing/model': 1.1293362388387322, 'learning_rate/qwen_vl_interface': 9.999585948021623e-06, 'learning_rate/action_model': 9.999585948021624e-05, 'epoch': 0.05}) +[2026-06-17 16:30:15,825][starVLA.training.train_starvla][INFO] - Step 130, Loss: {'train/loss': 1.5588841438293457, 'train/grad_norm_pre_clip': 14.143230438232422, 'timing/data': 0.019866522401571274, 'timing/model': 1.0420051794499159, 'learning_rate/qwen_vl_interface': 9.99955690083148e-06, 'learning_rate/action_model': 9.99955690083148e-05, 'epoch': 0.05}) +[2026-06-17 16:30:20,344][starVLA.training.train_starvla][INFO] - Step 131, Loss: {'train/loss': 1.7107138633728027, 'train/grad_norm_pre_clip': 8.532184600830078, 'timing/data': 0.019774257205426693, 'timing/model': 1.102878000587225, 'learning_rate/qwen_vl_interface': 9.99952686905308e-06, 'learning_rate/action_model': 9.999526869053081e-05, 'epoch': 0.05}) +[2026-06-17 16:30:24,760][starVLA.training.train_starvla][INFO] - Step 132, Loss: {'train/loss': 1.4515550136566162, 'train/grad_norm_pre_clip': 7.791313171386719, 'timing/data': 0.02338321413844824, 'timing/model': 1.1165310284122825, 'learning_rate/qwen_vl_interface': 9.99949585269265e-06, 'learning_rate/action_model': 9.99949585269265e-05, 'epoch': 0.05}) +[2026-06-17 16:30:29,876][starVLA.training.train_starvla][INFO] - Step 133, Loss: {'train/loss': 1.7641867399215698, 'train/grad_norm_pre_clip': 32.01121139526367, 'timing/data': 0.023518788628280163, 'timing/model': 2.044361885637045, 'learning_rate/qwen_vl_interface': 9.99946385175662e-06, 'learning_rate/action_model': 9.999463851756619e-05, 'epoch': 0.05}) +[2026-06-17 16:30:33,958][starVLA.training.train_starvla][INFO] - Step 134, Loss: {'train/loss': 1.5766217708587646, 'train/grad_norm_pre_clip': 7.905345439910889, 'timing/data': 0.020902076736092567, 'timing/model': 1.02480214741081, 'learning_rate/qwen_vl_interface': 9.999430866251623e-06, 'learning_rate/action_model': 9.999430866251622e-05, 'epoch': 0.05}) +[2026-06-17 16:30:38,411][starVLA.training.train_starvla][INFO] - Step 135, Loss: {'train/loss': 1.5544939041137695, 'train/grad_norm_pre_clip': 7.274721145629883, 'timing/data': 0.02063822280615568, 'timing/model': 1.0402381364256144, 'learning_rate/qwen_vl_interface': 9.999396896184494e-06, 'learning_rate/action_model': 9.999396896184494e-05, 'epoch': 0.05}) +[2026-06-17 16:30:42,971][starVLA.training.train_starvla][INFO] - Step 136, Loss: {'train/loss': 1.4171000719070435, 'train/grad_norm_pre_clip': 7.159008979797363, 'timing/data': 0.02191950846463442, 'timing/model': 1.0799857089295983, 'learning_rate/qwen_vl_interface': 9.99936194156228e-06, 'learning_rate/action_model': 9.99936194156228e-05, 'epoch': 0.05}) +[2026-06-17 16:30:47,723][starVLA.training.train_starvla][INFO] - Step 137, Loss: {'train/loss': 1.3601267337799072, 'train/grad_norm_pre_clip': 10.144196510314941, 'timing/data': 0.02266242541372776, 'timing/model': 1.3549311328679323, 'learning_rate/qwen_vl_interface': 9.999326002392225e-06, 'learning_rate/action_model': 9.999326002392224e-05, 'epoch': 0.05}) +[2026-06-17 16:30:52,316][starVLA.training.train_starvla][INFO] - Step 138, Loss: {'train/loss': 1.4463523626327515, 'train/grad_norm_pre_clip': 12.495966911315918, 'timing/data': 0.026737552136182785, 'timing/model': 1.5920477695763111, 'learning_rate/qwen_vl_interface': 9.999289078681779e-06, 'learning_rate/action_model': 9.999289078681778e-05, 'epoch': 0.05}) +[2026-06-17 16:30:56,526][starVLA.training.train_starvla][INFO] - Step 139, Loss: {'train/loss': 1.3680907487869263, 'train/grad_norm_pre_clip': 6.683152675628662, 'timing/data': 0.020704942755401134, 'timing/model': 1.090132280252874, 'learning_rate/qwen_vl_interface': 9.999251170438594e-06, 'learning_rate/action_model': 9.999251170438594e-05, 'epoch': 0.05}) +[2026-06-17 16:31:01,143][starVLA.training.train_starvla][INFO] - Step 140, Loss: {'train/loss': 1.5376367568969727, 'train/grad_norm_pre_clip': 12.344992637634277, 'timing/data': 0.02952989563345909, 'timing/model': 1.0982414474710822, 'learning_rate/qwen_vl_interface': 9.999212277670533e-06, 'learning_rate/action_model': 9.999212277670533e-05, 'epoch': 0.05}) +[2026-06-17 16:31:05,534][starVLA.training.train_starvla][INFO] - Step 141, Loss: {'train/loss': 1.2744029760360718, 'train/grad_norm_pre_clip': 8.132094383239746, 'timing/data': 0.025482074357569218, 'timing/model': 1.0259067425504327, 'learning_rate/qwen_vl_interface': 9.999172400385658e-06, 'learning_rate/action_model': 9.999172400385658e-05, 'epoch': 0.05}) +[2026-06-17 16:31:10,106][starVLA.training.train_starvla][INFO] - Step 142, Loss: {'train/loss': 1.7509299516677856, 'train/grad_norm_pre_clip': 13.040061950683594, 'timing/data': 0.025516883470118046, 'timing/model': 1.166447913274169, 'learning_rate/qwen_vl_interface': 9.999131538592232e-06, 'learning_rate/action_model': 9.999131538592232e-05, 'epoch': 0.05}) +[2026-06-17 16:31:14,582][starVLA.training.train_starvla][INFO] - Step 143, Loss: {'train/loss': 1.6480573415756226, 'train/grad_norm_pre_clip': 10.424324035644531, 'timing/data': 0.022556607611477375, 'timing/model': 1.4350536856800318, 'learning_rate/qwen_vl_interface': 9.999089692298729e-06, 'learning_rate/action_model': 9.99908969229873e-05, 'epoch': 0.05}) +[2026-06-17 16:31:18,703][starVLA.training.train_starvla][INFO] - Step 144, Loss: {'train/loss': 1.6073263883590698, 'train/grad_norm_pre_clip': 10.619404792785645, 'timing/data': 0.024985789321362972, 'timing/model': 1.1137947719544172, 'learning_rate/qwen_vl_interface': 9.999046861513822e-06, 'learning_rate/action_model': 9.999046861513821e-05, 'epoch': 0.05}) +[2026-06-17 16:31:23,271][starVLA.training.train_starvla][INFO] - Step 145, Loss: {'train/loss': 1.3691643476486206, 'train/grad_norm_pre_clip': 12.34545612335205, 'timing/data': 0.019389862194657326, 'timing/model': 1.0420231958851218, 'learning_rate/qwen_vl_interface': 9.99900304624639e-06, 'learning_rate/action_model': 9.99900304624639e-05, 'epoch': 0.05}) +[2026-06-17 16:31:27,702][starVLA.training.train_starvla][INFO] - Step 146, Loss: {'train/loss': 1.5537241697311401, 'train/grad_norm_pre_clip': 10.081299781799316, 'timing/data': 0.020398318767547607, 'timing/model': 1.0354271866381168, 'learning_rate/qwen_vl_interface': 9.998958246505518e-06, 'learning_rate/action_model': 9.998958246505516e-05, 'epoch': 0.05}) +[2026-06-17 16:31:32,185][starVLA.training.train_starvla][INFO] - Step 147, Loss: {'train/loss': 1.456811547279358, 'train/grad_norm_pre_clip': 9.913576126098633, 'timing/data': 0.024939103983342648, 'timing/model': 1.3644924899563193, 'learning_rate/qwen_vl_interface': 9.998912462300489e-06, 'learning_rate/action_model': 9.998912462300488e-05, 'epoch': 0.05}) +[2026-06-17 16:31:36,598][starVLA.training.train_starvla][INFO] - Step 148, Loss: {'train/loss': 1.19246244430542, 'train/grad_norm_pre_clip': 9.155888557434082, 'timing/data': 0.020302779972553253, 'timing/model': 1.4247625777497888, 'learning_rate/qwen_vl_interface': 9.998865693640797e-06, 'learning_rate/action_model': 9.998865693640796e-05, 'epoch': 0.05}) +[2026-06-17 16:31:40,664][starVLA.training.train_starvla][INFO] - Step 149, Loss: {'train/loss': 1.4024348258972168, 'train/grad_norm_pre_clip': 11.600177764892578, 'timing/data': 0.021202736534178257, 'timing/model': 1.0523889670148492, 'learning_rate/qwen_vl_interface': 9.998817940536138e-06, 'learning_rate/action_model': 9.998817940536137e-05, 'epoch': 0.05}) +[2026-06-17 16:31:45,331][starVLA.training.train_starvla][INFO] - Step 150, Loss: {'train/loss': 1.8185911178588867, 'train/grad_norm_pre_clip': 16.493375778198242, 'timing/data': 0.021488869562745094, 'timing/model': 1.0968359364196658, 'learning_rate/qwen_vl_interface': 9.99876920299641e-06, 'learning_rate/action_model': 9.998769202996408e-05, 'epoch': 0.05}) +[2026-06-17 16:31:50,653][starVLA.training.train_starvla][INFO] - Step 151, Loss: {'train/loss': 1.5561107397079468, 'train/grad_norm_pre_clip': 21.84815216064453, 'timing/data': 0.01912882924079895, 'timing/model': 1.9065986424684525, 'learning_rate/qwen_vl_interface': 9.998719481031715e-06, 'learning_rate/action_model': 9.998719481031715e-05, 'epoch': 0.06}) +[2026-06-17 16:31:55,221][starVLA.training.train_starvla][INFO] - Step 152, Loss: {'train/loss': 1.418130874633789, 'train/grad_norm_pre_clip': 12.990785598754883, 'timing/data': 0.022436274215579033, 'timing/model': 1.0989420944824815, 'learning_rate/qwen_vl_interface': 9.998668774652363e-06, 'learning_rate/action_model': 9.998668774652362e-05, 'epoch': 0.06}) +[2026-06-17 16:31:59,579][starVLA.training.train_starvla][INFO] - Step 153, Loss: {'train/loss': 1.4463404417037964, 'train/grad_norm_pre_clip': 6.242889881134033, 'timing/data': 0.02034074254333973, 'timing/model': 1.375665683299303, 'learning_rate/qwen_vl_interface': 9.998617083868863e-06, 'learning_rate/action_model': 9.998617083868863e-05, 'epoch': 0.06}) +[2026-06-17 16:32:03,601][starVLA.training.train_starvla][INFO] - Step 154, Loss: {'train/loss': 1.3201874494552612, 'train/grad_norm_pre_clip': 10.124101638793945, 'timing/data': 0.019293121993541718, 'timing/model': 1.0278350878506899, 'learning_rate/qwen_vl_interface': 9.998564408691933e-06, 'learning_rate/action_model': 9.998564408691932e-05, 'epoch': 0.06}) +[2026-06-17 16:32:08,191][starVLA.training.train_starvla][INFO] - Step 155, Loss: {'train/loss': 1.7830396890640259, 'train/grad_norm_pre_clip': 12.292317390441895, 'timing/data': 0.030311882495880127, 'timing/model': 1.123646054416895, 'learning_rate/qwen_vl_interface': 9.998510749132491e-06, 'learning_rate/action_model': 9.998510749132491e-05, 'epoch': 0.06}) +[2026-06-17 16:32:12,768][starVLA.training.train_starvla][INFO] - Step 156, Loss: {'train/loss': 1.4487193822860718, 'train/grad_norm_pre_clip': 18.16716766357422, 'timing/data': 0.023229586891829967, 'timing/model': 1.039366920478642, 'learning_rate/qwen_vl_interface': 9.998456105201659e-06, 'learning_rate/action_model': 9.99845610520166e-05, 'epoch': 0.06}) +[2026-06-17 16:32:18,927][starVLA.training.train_starvla][INFO] - Step 157, Loss: {'train/loss': 1.3084224462509155, 'train/grad_norm_pre_clip': 9.060694694519043, 'timing/data': 0.03184577077627182, 'timing/model': 1.4652139144018292, 'learning_rate/qwen_vl_interface': 9.99840047691077e-06, 'learning_rate/action_model': 9.99840047691077e-05, 'epoch': 0.06}) +[2026-06-17 16:32:23,553][starVLA.training.train_starvla][INFO] - Step 158, Loss: {'train/loss': 1.587438702583313, 'train/grad_norm_pre_clip': 8.87014102935791, 'timing/data': 0.024014180526137352, 'timing/model': 1.3267403673380613, 'learning_rate/qwen_vl_interface': 9.998343864271352e-06, 'learning_rate/action_model': 9.998343864271351e-05, 'epoch': 0.06}) +[2026-06-17 16:32:27,628][starVLA.training.train_starvla][INFO] - Step 159, Loss: {'train/loss': 1.4154415130615234, 'train/grad_norm_pre_clip': 13.29587173461914, 'timing/data': 0.019602546468377113, 'timing/model': 1.0294175371527672, 'learning_rate/qwen_vl_interface': 9.99828626729514e-06, 'learning_rate/action_model': 9.998286267295139e-05, 'epoch': 0.06}) +[2026-06-17 16:32:33,414][starVLA.training.train_starvla][INFO] - Step 160, Loss: {'train/loss': 1.3104840517044067, 'train/grad_norm_pre_clip': 8.701961517333984, 'timing/data': 0.022347006015479565, 'timing/model': 1.0724990190938115, 'learning_rate/qwen_vl_interface': 9.998227685994077e-06, 'learning_rate/action_model': 9.998227685994075e-05, 'epoch': 0.06}) +[2026-06-17 16:32:37,877][starVLA.training.train_starvla][INFO] - Step 161, Loss: {'train/loss': 1.4695502519607544, 'train/grad_norm_pre_clip': 7.158535480499268, 'timing/data': 0.025585319846868515, 'timing/model': 1.0676624868065119, 'learning_rate/qwen_vl_interface': 9.998168120380303e-06, 'learning_rate/action_model': 9.998168120380303e-05, 'epoch': 0.06}) +[2026-06-17 16:32:42,398][starVLA.training.train_starvla][INFO] - Step 162, Loss: {'train/loss': 1.4783563613891602, 'train/grad_norm_pre_clip': 6.645670413970947, 'timing/data': 0.020740116015076637, 'timing/model': 1.0294175408780575, 'learning_rate/qwen_vl_interface': 9.99810757046617e-06, 'learning_rate/action_model': 9.99810757046617e-05, 'epoch': 0.06}) +[2026-06-17 16:32:46,846][starVLA.training.train_starvla][INFO] - Step 163, Loss: {'train/loss': 1.2668887376785278, 'train/grad_norm_pre_clip': 8.730359077453613, 'timing/data': 0.01997553091496229, 'timing/model': 1.3726269183680415, 'learning_rate/qwen_vl_interface': 9.998046036264229e-06, 'learning_rate/action_model': 9.99804603626423e-05, 'epoch': 0.06}) +[2026-06-17 16:32:50,876][starVLA.training.train_starvla][INFO] - Step 164, Loss: {'train/loss': 1.5005215406417847, 'train/grad_norm_pre_clip': 11.546540260314941, 'timing/data': 0.022033190354704857, 'timing/model': 1.0322561552748084, 'learning_rate/qwen_vl_interface': 9.997983517787236e-06, 'learning_rate/action_model': 9.997983517787236e-05, 'epoch': 0.06}) +[2026-06-17 16:32:55,446][starVLA.training.train_starvla][INFO] - Step 165, Loss: {'train/loss': 1.4704821109771729, 'train/grad_norm_pre_clip': 9.765777587890625, 'timing/data': 0.019881674088537693, 'timing/model': 1.0408032964915037, 'learning_rate/qwen_vl_interface': 9.99792001504815e-06, 'learning_rate/action_model': 9.997920015048149e-05, 'epoch': 0.06}) +[2026-06-17 16:32:59,973][starVLA.training.train_starvla][INFO] - Step 166, Loss: {'train/loss': 1.3234004974365234, 'train/grad_norm_pre_clip': 7.4427618980407715, 'timing/data': 0.052998410537838936, 'timing/model': 1.0411504227668047, 'learning_rate/qwen_vl_interface': 9.997855528060136e-06, 'learning_rate/action_model': 9.997855528060135e-05, 'epoch': 0.06}) +[2026-06-17 16:33:04,331][starVLA.training.train_starvla][INFO] - Step 167, Loss: {'train/loss': 1.4454503059387207, 'train/grad_norm_pre_clip': 12.028170585632324, 'timing/data': 0.01878522615879774, 'timing/model': 1.3326207781210542, 'learning_rate/qwen_vl_interface': 9.997790056836562e-06, 'learning_rate/action_model': 9.997790056836562e-05, 'epoch': 0.06}) +[2026-06-17 16:33:09,401][starVLA.training.train_starvla][INFO] - Step 168, Loss: {'train/loss': 1.4442064762115479, 'train/grad_norm_pre_clip': 15.181782722473145, 'timing/data': 0.023140937089920044, 'timing/model': 1.4051144272089005, 'learning_rate/qwen_vl_interface': 9.997723601390998e-06, 'learning_rate/action_model': 9.997723601390998e-05, 'epoch': 0.06}) +[2026-06-17 16:33:13,745][starVLA.training.train_starvla][INFO] - Step 169, Loss: {'train/loss': 1.4362270832061768, 'train/grad_norm_pre_clip': 11.3538179397583, 'timing/data': 0.023441641591489315, 'timing/model': 1.0989596154540777, 'learning_rate/qwen_vl_interface': 9.997656161737227e-06, 'learning_rate/action_model': 9.997656161737226e-05, 'epoch': 0.06}) +[2026-06-17 16:33:18,250][starVLA.training.train_starvla][INFO] - Step 170, Loss: {'train/loss': 1.50021231174469, 'train/grad_norm_pre_clip': 13.70458698272705, 'timing/data': 0.021879234351217747, 'timing/model': 1.0816171457991004, 'learning_rate/qwen_vl_interface': 9.997587737889222e-06, 'learning_rate/action_model': 9.997587737889221e-05, 'epoch': 0.06}) +[2026-06-17 16:33:22,751][starVLA.training.train_starvla][INFO] - Step 171, Loss: {'train/loss': 1.801353096961975, 'train/grad_norm_pre_clip': 10.925819396972656, 'timing/data': 0.0212256433442235, 'timing/model': 1.0319126779213548, 'learning_rate/qwen_vl_interface': 9.997518329861172e-06, 'learning_rate/action_model': 9.997518329861171e-05, 'epoch': 0.06}) +[2026-06-17 16:33:27,212][starVLA.training.train_starvla][INFO] - Step 172, Loss: {'train/loss': 1.5878995656967163, 'train/grad_norm_pre_clip': 16.492258071899414, 'timing/data': 0.03225706331431866, 'timing/model': 1.0989883970469236, 'learning_rate/qwen_vl_interface': 9.997447937667462e-06, 'learning_rate/action_model': 9.997447937667462e-05, 'epoch': 0.06}) +[2026-06-17 16:33:31,733][starVLA.training.train_starvla][INFO] - Step 173, Loss: {'train/loss': 1.4783936738967896, 'train/grad_norm_pre_clip': 17.638885498046875, 'timing/data': 0.02204915415495634, 'timing/model': 1.3733014473691583, 'learning_rate/qwen_vl_interface': 9.997376561322686e-06, 'learning_rate/action_model': 9.997376561322686e-05, 'epoch': 0.06}) +[2026-06-17 16:33:35,839][starVLA.training.train_starvla][INFO] - Step 174, Loss: {'train/loss': 1.327582836151123, 'train/grad_norm_pre_clip': 8.639373779296875, 'timing/data': 0.026536379009485245, 'timing/model': 1.0684823784977198, 'learning_rate/qwen_vl_interface': 9.997304200841642e-06, 'learning_rate/action_model': 9.997304200841641e-05, 'epoch': 0.06}) +[2026-06-17 16:33:40,434][starVLA.training.train_starvla][INFO] - Step 175, Loss: {'train/loss': 1.4363224506378174, 'train/grad_norm_pre_clip': 9.302727699279785, 'timing/data': 0.021664472296833992, 'timing/model': 1.0893539618700743, 'learning_rate/qwen_vl_interface': 9.997230856239327e-06, 'learning_rate/action_model': 9.997230856239327e-05, 'epoch': 0.06}) +[2026-06-17 16:33:44,990][starVLA.training.train_starvla][INFO] - Step 176, Loss: {'train/loss': 1.67946195602417, 'train/grad_norm_pre_clip': 17.55670166015625, 'timing/data': 0.0217279139906168, 'timing/model': 1.0760235637426376, 'learning_rate/qwen_vl_interface': 9.997156527530948e-06, 'learning_rate/action_model': 9.997156527530948e-05, 'epoch': 0.06}) +[2026-06-17 16:33:49,695][starVLA.training.train_starvla][INFO] - Step 177, Loss: {'train/loss': 1.4389475584030151, 'train/grad_norm_pre_clip': 9.621557235717773, 'timing/data': 0.0228475509211421, 'timing/model': 1.348833467811346, 'learning_rate/qwen_vl_interface': 9.997081214731915e-06, 'learning_rate/action_model': 9.997081214731914e-05, 'epoch': 0.06}) +[2026-06-17 16:33:54,083][starVLA.training.train_starvla][INFO] - Step 178, Loss: {'train/loss': 1.5267730951309204, 'train/grad_norm_pre_clip': 9.225396156311035, 'timing/data': 0.020065007731318474, 'timing/model': 1.385185131803155, 'learning_rate/qwen_vl_interface': 9.997004917857834e-06, 'learning_rate/action_model': 9.997004917857834e-05, 'epoch': 0.07}) +[2026-06-17 16:33:58,211][starVLA.training.train_starvla][INFO] - Step 179, Loss: {'train/loss': 1.172296404838562, 'train/grad_norm_pre_clip': 12.489436149597168, 'timing/data': 0.020141021348536015, 'timing/model': 1.058940626680851, 'learning_rate/qwen_vl_interface': 9.996927636924529e-06, 'learning_rate/action_model': 9.996927636924528e-05, 'epoch': 0.07}) +[2026-06-17 16:34:02,653][starVLA.training.train_starvla][INFO] - Step 180, Loss: {'train/loss': 1.1729741096496582, 'train/grad_norm_pre_clip': 8.456265449523926, 'timing/data': 0.01989168394356966, 'timing/model': 1.0309367189183831, 'learning_rate/qwen_vl_interface': 9.996849371948016e-06, 'learning_rate/action_model': 9.996849371948016e-05, 'epoch': 0.07}) +[2026-06-17 16:34:08,016][starVLA.training.train_starvla][INFO] - Step 181, Loss: {'train/loss': 1.6134884357452393, 'train/grad_norm_pre_clip': 8.60451602935791, 'timing/data': 0.020059323869645596, 'timing/model': 2.05572222918272, 'learning_rate/qwen_vl_interface': 9.996770122944518e-06, 'learning_rate/action_model': 9.996770122944518e-05, 'epoch': 0.07}) +[2026-06-17 16:34:14,308][starVLA.training.train_starvla][INFO] - Step 182, Loss: {'train/loss': 1.3295280933380127, 'train/grad_norm_pre_clip': 8.466900825500488, 'timing/data': 0.027795453555881977, 'timing/model': 1.0844972366467118, 'learning_rate/qwen_vl_interface': 9.99668988993047e-06, 'learning_rate/action_model': 9.99668988993047e-05, 'epoch': 0.07}) +[2026-06-17 16:34:18,772][starVLA.training.train_starvla][INFO] - Step 183, Loss: {'train/loss': 1.1627320051193237, 'train/grad_norm_pre_clip': 40.524810791015625, 'timing/data': 0.022344546392560005, 'timing/model': 1.4443515529856086, 'learning_rate/qwen_vl_interface': 9.996608672922496e-06, 'learning_rate/action_model': 9.996608672922496e-05, 'epoch': 0.07}) +[2026-06-17 16:34:22,835][starVLA.training.train_starvla][INFO] - Step 184, Loss: {'train/loss': 1.5366724729537964, 'train/grad_norm_pre_clip': 20.334253311157227, 'timing/data': 0.01961371023207903, 'timing/model': 1.0463786330074072, 'learning_rate/qwen_vl_interface': 9.996526471937438e-06, 'learning_rate/action_model': 9.996526471937439e-05, 'epoch': 0.07}) +[2026-06-17 16:34:27,258][starVLA.training.train_starvla][INFO] - Step 185, Loss: {'train/loss': 1.3842804431915283, 'train/grad_norm_pre_clip': 7.838594913482666, 'timing/data': 0.021836371161043644, 'timing/model': 1.0612661968916655, 'learning_rate/qwen_vl_interface': 9.996443286992337e-06, 'learning_rate/action_model': 9.996443286992336e-05, 'epoch': 0.07}) +[2026-06-17 16:34:31,763][starVLA.training.train_starvla][INFO] - Step 186, Loss: {'train/loss': 1.3743196725845337, 'train/grad_norm_pre_clip': 7.229007720947266, 'timing/data': 0.021253762766718864, 'timing/model': 1.0696720825508237, 'learning_rate/qwen_vl_interface': 9.996359118104432e-06, 'learning_rate/action_model': 9.996359118104432e-05, 'epoch': 0.07}) +[2026-06-17 16:34:36,424][starVLA.training.train_starvla][INFO] - Step 187, Loss: {'train/loss': 1.1708909273147583, 'train/grad_norm_pre_clip': 6.77536678314209, 'timing/data': 0.03125787153840065, 'timing/model': 1.4814748167991638, 'learning_rate/qwen_vl_interface': 9.996273965291177e-06, 'learning_rate/action_model': 9.996273965291176e-05, 'epoch': 0.07}) +[2026-06-17 16:34:40,963][starVLA.training.train_starvla][INFO] - Step 188, Loss: {'train/loss': 1.3154371976852417, 'train/grad_norm_pre_clip': 8.425618171691895, 'timing/data': 0.023565576411783695, 'timing/model': 1.41571971680969, 'learning_rate/qwen_vl_interface': 9.996187828570221e-06, 'learning_rate/action_model': 9.996187828570221e-05, 'epoch': 0.07}) +[2026-06-17 16:34:44,984][starVLA.training.train_starvla][INFO] - Step 189, Loss: {'train/loss': 1.2493469715118408, 'train/grad_norm_pre_clip': 5.831063747406006, 'timing/data': 0.021066789515316486, 'timing/model': 1.026079187169671, 'learning_rate/qwen_vl_interface': 9.996100707959422e-06, 'learning_rate/action_model': 9.996100707959421e-05, 'epoch': 0.07}) +[2026-06-17 16:34:49,579][starVLA.training.train_starvla][INFO] - Step 190, Loss: {'train/loss': 1.354018211364746, 'train/grad_norm_pre_clip': 10.759421348571777, 'timing/data': 0.020337440073490143, 'timing/model': 1.2084222817793489, 'learning_rate/qwen_vl_interface': 9.996012603476836e-06, 'learning_rate/action_model': 9.996012603476837e-05, 'epoch': 0.07}) +[2026-06-17 16:34:54,044][starVLA.training.train_starvla][INFO] - Step 191, Loss: {'train/loss': 1.258671760559082, 'train/grad_norm_pre_clip': 11.081815719604492, 'timing/data': 0.025787359103560448, 'timing/model': 1.0675995396450162, 'learning_rate/qwen_vl_interface': 9.995923515140734e-06, 'learning_rate/action_model': 9.995923515140732e-05, 'epoch': 0.07}) +[2026-06-17 16:34:58,507][starVLA.training.train_starvla][INFO] - Step 192, Loss: {'train/loss': 1.4856992959976196, 'train/grad_norm_pre_clip': 9.141610145568848, 'timing/data': 0.02082823868840933, 'timing/model': 1.032498188316822, 'learning_rate/qwen_vl_interface': 9.995833442969578e-06, 'learning_rate/action_model': 9.995833442969578e-05, 'epoch': 0.07}) +[2026-06-17 16:35:03,186][starVLA.training.train_starvla][INFO] - Step 193, Loss: {'train/loss': 1.3740094900131226, 'train/grad_norm_pre_clip': 8.870905876159668, 'timing/data': 0.024766726419329643, 'timing/model': 1.484824862331152, 'learning_rate/qwen_vl_interface': 9.995742386982045e-06, 'learning_rate/action_model': 9.995742386982045e-05, 'epoch': 0.07}) +[2026-06-17 16:35:07,228][starVLA.training.train_starvla][INFO] - Step 194, Loss: {'train/loss': 1.4499695301055908, 'train/grad_norm_pre_clip': 12.328436851501465, 'timing/data': 0.021623325534164906, 'timing/model': 1.0638093277812004, 'learning_rate/qwen_vl_interface': 9.995650347197006e-06, 'learning_rate/action_model': 9.995650347197005e-05, 'epoch': 0.07}) +[2026-06-17 16:35:11,762][starVLA.training.train_starvla][INFO] - Step 195, Loss: {'train/loss': 1.5551103353500366, 'train/grad_norm_pre_clip': 11.69092082977295, 'timing/data': 0.021234861575067043, 'timing/model': 1.044932210817933, 'learning_rate/qwen_vl_interface': 9.995557323633544e-06, 'learning_rate/action_model': 9.995557323633543e-05, 'epoch': 0.07}) +[2026-06-17 16:35:16,261][starVLA.training.train_starvla][INFO] - Step 196, Loss: {'train/loss': 1.4015291929244995, 'train/grad_norm_pre_clip': 17.409446716308594, 'timing/data': 0.022096536122262478, 'timing/model': 1.042082192376256, 'learning_rate/qwen_vl_interface': 9.995463316310942e-06, 'learning_rate/action_model': 9.995463316310943e-05, 'epoch': 0.07}) +[2026-06-17 16:35:21,103][starVLA.training.train_starvla][INFO] - Step 197, Loss: {'train/loss': 1.4654788970947266, 'train/grad_norm_pre_clip': 10.351492881774902, 'timing/data': 0.02792664710432291, 'timing/model': 1.437152232043445, 'learning_rate/qwen_vl_interface': 9.995368325248688e-06, 'learning_rate/action_model': 9.995368325248688e-05, 'epoch': 0.07}) +[2026-06-17 16:35:26,086][starVLA.training.train_starvla][INFO] - Step 198, Loss: {'train/loss': 1.4050443172454834, 'train/grad_norm_pre_clip': 8.131853103637695, 'timing/data': 0.02016741130501032, 'timing/model': 1.41679666377604, 'learning_rate/qwen_vl_interface': 9.995272350466476e-06, 'learning_rate/action_model': 9.995272350466476e-05, 'epoch': 0.07}) +[2026-06-17 16:35:30,249][starVLA.training.train_starvla][INFO] - Step 199, Loss: {'train/loss': 1.2572380304336548, 'train/grad_norm_pre_clip': 9.84299373626709, 'timing/data': 0.027177945710718632, 'timing/model': 1.1070133708417416, 'learning_rate/qwen_vl_interface': 9.995175391984199e-06, 'learning_rate/action_model': 9.995175391984198e-05, 'epoch': 0.07}) +[2026-06-17 16:37:59,536][starVLA.training.train_starvla][INFO] - Step 200, Loss: {'train/loss': 1.0545707941055298, 'train/grad_norm_pre_clip': 8.127081871032715, 'eval/loss': 1.4237730503082275, 'eval/latency_2/loss': 1.4237730503082275, 'eval/demon_fire_precision': 0.8306283811901789, 'eval/demon_fire_recall': 0.8351464435146444, 'eval/demon_fire_f1': 0.8328812852075943, 'eval/demon_fire_rate_model': 0.6686143572621035, 'eval/demon_fire_rate_teacher': 0.6649972175848636, 'eval/demon_left_precision': 0.7170138888888888, 'eval/demon_left_recall': 0.43473684210526314, 'eval/demon_left_f1': 0.5412844036697247, 'eval/demon_left_rate_model': 0.16026711185308848, 'eval/demon_left_rate_teacher': 0.2643294379521425, 'eval/demon_right_precision': 0.4645239539114615, 'eval/demon_right_recall': 0.8655367231638418, 'eval/demon_right_f1': 0.6045777426992895, 'eval/demon_right_rate_model': 0.45882025598219256, 'eval/demon_right_rate_teacher': 0.24624373956594323, 'eval/demon_move_f1': 0.5729310731845072, 'eval/demon_cc_f1': 0.73309531228011, 'eval/latency_2/demon_fire_precision': 0.8306283811901789, 'eval/latency_2/demon_fire_recall': 0.8351464435146444, 'eval/latency_2/demon_fire_f1': 0.8328812852075943, 'eval/latency_2/demon_fire_rate_model': 0.6686143572621035, 'eval/latency_2/demon_fire_rate_teacher': 0.6649972175848636, 'eval/latency_2/demon_left_precision': 0.7170138888888888, 'eval/latency_2/demon_left_recall': 0.43473684210526314, 'eval/latency_2/demon_left_f1': 0.5412844036697247, 'eval/latency_2/demon_left_rate_model': 0.16026711185308848, 'eval/latency_2/demon_left_rate_teacher': 0.2643294379521425, 'eval/latency_2/demon_right_precision': 0.4645239539114615, 'eval/latency_2/demon_right_recall': 0.8655367231638418, 'eval/latency_2/demon_right_f1': 0.6045777426992895, 'eval/latency_2/demon_right_rate_model': 0.45882025598219256, 'eval/latency_2/demon_right_rate_teacher': 0.24624373956594323, 'eval/latency_2/demon_move_f1': 0.5729310731845072, 'eval/latency_2/demon_cc_f1': 0.73309531228011, 'timing/data': 0.020515572279691696, 'timing/model': 1.0250440873205662, 'learning_rate/qwen_vl_interface': 9.995077449821957e-06, 'learning_rate/action_model': 9.995077449821956e-05, 'epoch': 0.07}) +[2026-06-17 16:37:59,545][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_200_state +[2026-06-17 16:37:59,552][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 16:38:17,561][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_200_state/model.safetensors +[2026-06-17 16:38:32,201][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_200_state/optimizer.bin +[2026-06-17 16:38:32,203][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_200_state/sampler.bin +[2026-06-17 16:38:32,204][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_200_state/sampler_1.bin +[2026-06-17 16:38:32,207][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_200_state/random_states_0.pkl +[2026-06-17 16:38:32,220][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 200 +[2026-06-17 16:38:32,221][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 16:38:32,239][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 16:38:36,804][starVLA.training.train_starvla][INFO] - Step 201, Loss: {'train/loss': 1.2011173963546753, 'train/grad_norm_pre_clip': 10.555339813232422, 'timing/data': 0.019365688785910606, 'timing/model': 1.092972961254418, 'learning_rate/qwen_vl_interface': 9.994978524000053e-06, 'learning_rate/action_model': 9.994978524000052e-05, 'epoch': 0.07}) +[2026-06-17 16:38:42,863][starVLA.training.train_starvla][INFO] - Step 202, Loss: {'train/loss': 1.8832221031188965, 'train/grad_norm_pre_clip': 9.745604515075684, 'timing/data': 0.018246717751026154, 'timing/model': 1.0888953655958176, 'learning_rate/qwen_vl_interface': 9.994878614538996e-06, 'learning_rate/action_model': 9.994878614538996e-05, 'epoch': 0.07}) +[2026-06-17 16:38:47,398][starVLA.training.train_starvla][INFO] - Step 203, Loss: {'train/loss': 1.469145655632019, 'train/grad_norm_pre_clip': 19.61475944519043, 'timing/data': 0.01995256543159485, 'timing/model': 1.0353635484352708, 'learning_rate/qwen_vl_interface': 9.994777721459496e-06, 'learning_rate/action_model': 9.994777721459497e-05, 'epoch': 0.07}) +[2026-06-17 16:38:51,830][starVLA.training.train_starvla][INFO] - Step 204, Loss: {'train/loss': 1.2959163188934326, 'train/grad_norm_pre_clip': 14.40942668914795, 'timing/data': 0.019713137298822403, 'timing/model': 1.3657656386494637, 'learning_rate/qwen_vl_interface': 9.994675844782468e-06, 'learning_rate/action_model': 9.994675844782468e-05, 'epoch': 0.07}) +[2026-06-17 16:38:56,011][starVLA.training.train_starvla][INFO] - Step 205, Loss: {'train/loss': 1.3682575225830078, 'train/grad_norm_pre_clip': 8.317523956298828, 'timing/data': 0.019185210578143597, 'timing/model': 1.0818995581939816, 'learning_rate/qwen_vl_interface': 9.994572984529033e-06, 'learning_rate/action_model': 9.994572984529034e-05, 'epoch': 0.08}) +[2026-06-17 16:39:00,472][starVLA.training.train_starvla][INFO] - Step 206, Loss: {'train/loss': 1.3019943237304688, 'train/grad_norm_pre_clip': 10.058140754699707, 'timing/data': 0.018392181023955345, 'timing/model': 1.0725178699940443, 'learning_rate/qwen_vl_interface': 9.994469140720513e-06, 'learning_rate/action_model': 9.994469140720512e-05, 'epoch': 0.08}) +[2026-06-17 16:39:04,857][starVLA.training.train_starvla][INFO] - Step 207, Loss: {'train/loss': 1.352639079093933, 'train/grad_norm_pre_clip': 13.30119800567627, 'timing/data': 0.019398449920117855, 'timing/model': 1.0430165082216263, 'learning_rate/qwen_vl_interface': 9.994364313378435e-06, 'learning_rate/action_model': 9.994364313378433e-05, 'epoch': 0.08}) +[2026-06-17 16:39:09,258][starVLA.training.train_starvla][INFO] - Step 208, Loss: {'train/loss': 1.1237562894821167, 'train/grad_norm_pre_clip': 7.532413482666016, 'timing/data': 0.017687426880002022, 'timing/model': 1.457251077517867, 'learning_rate/qwen_vl_interface': 9.99425850252453e-06, 'learning_rate/action_model': 9.994258502524529e-05, 'epoch': 0.08}) +[2026-06-17 16:39:13,657][starVLA.training.train_starvla][INFO] - Step 209, Loss: {'train/loss': 1.3689249753952026, 'train/grad_norm_pre_clip': 5.507537364959717, 'timing/data': 0.01837960910052061, 'timing/model': 1.391061213798821, 'learning_rate/qwen_vl_interface': 9.994151708180731e-06, 'learning_rate/action_model': 9.99415170818073e-05, 'epoch': 0.08}) +[2026-06-17 16:39:17,718][starVLA.training.train_starvla][INFO] - Step 210, Loss: {'train/loss': 1.5379136800765991, 'train/grad_norm_pre_clip': 12.037633895874023, 'timing/data': 0.021024136804044247, 'timing/model': 1.0504684392362833, 'learning_rate/qwen_vl_interface': 9.994043930369179e-06, 'learning_rate/action_model': 9.994043930369178e-05, 'epoch': 0.08}) +[2026-06-17 16:39:22,176][starVLA.training.train_starvla][INFO] - Step 211, Loss: {'train/loss': 1.1903860569000244, 'train/grad_norm_pre_clip': 10.40177059173584, 'timing/data': 0.01984833087772131, 'timing/model': 1.025442212820053, 'learning_rate/qwen_vl_interface': 9.993935169112213e-06, 'learning_rate/action_model': 9.993935169112213e-05, 'epoch': 0.08}) +[2026-06-17 16:39:26,650][starVLA.training.train_starvla][INFO] - Step 212, Loss: {'train/loss': 1.0010052919387817, 'train/grad_norm_pre_clip': 8.14352798461914, 'timing/data': 0.01913660578429699, 'timing/model': 1.051921982318163, 'learning_rate/qwen_vl_interface': 9.993825424432385e-06, 'learning_rate/action_model': 9.993825424432385e-05, 'epoch': 0.08}) +[2026-06-17 16:39:31,131][starVLA.training.train_starvla][INFO] - Step 213, Loss: {'train/loss': 1.4307150840759277, 'train/grad_norm_pre_clip': 13.24194622039795, 'timing/data': 0.018632998690009117, 'timing/model': 1.029303758405149, 'learning_rate/qwen_vl_interface': 9.99371469635244e-06, 'learning_rate/action_model': 9.99371469635244e-05, 'epoch': 0.08}) +[2026-06-17 16:39:35,576][starVLA.training.train_starvla][INFO] - Step 214, Loss: {'train/loss': 1.4949028491973877, 'train/grad_norm_pre_clip': 8.182601928710938, 'timing/data': 0.01920226775109768, 'timing/model': 1.3725325604900718, 'learning_rate/qwen_vl_interface': 9.993602984895336e-06, 'learning_rate/action_model': 9.993602984895335e-05, 'epoch': 0.08}) +[2026-06-17 16:39:39,784][starVLA.training.train_starvla][INFO] - Step 215, Loss: {'train/loss': 1.3599478006362915, 'train/grad_norm_pre_clip': 12.90252685546875, 'timing/data': 0.020615612156689167, 'timing/model': 1.0591168496757746, 'learning_rate/qwen_vl_interface': 9.993490290084225e-06, 'learning_rate/action_model': 9.993490290084225e-05, 'epoch': 0.08}) +[2026-06-17 16:39:44,130][starVLA.training.train_starvla][INFO] - Step 216, Loss: {'train/loss': 1.3549250364303589, 'train/grad_norm_pre_clip': 12.753436088562012, 'timing/data': 0.01796671561896801, 'timing/model': 1.0225299149751663, 'learning_rate/qwen_vl_interface': 9.993376611942477e-06, 'learning_rate/action_model': 9.993376611942477e-05, 'epoch': 0.08}) +[2026-06-17 16:39:48,460][starVLA.training.train_starvla][INFO] - Step 217, Loss: {'train/loss': 1.2841356992721558, 'train/grad_norm_pre_clip': 8.95715045928955, 'timing/data': 0.02014720719307661, 'timing/model': 1.0334271099418402, 'learning_rate/qwen_vl_interface': 9.99326195049365e-06, 'learning_rate/action_model': 9.99326195049365e-05, 'epoch': 0.08}) +[2026-06-17 16:39:52,871][starVLA.training.train_starvla][INFO] - Step 218, Loss: {'train/loss': 1.1514567136764526, 'train/grad_norm_pre_clip': 8.195832252502441, 'timing/data': 0.018253719434142113, 'timing/model': 1.4241539491340518, 'learning_rate/qwen_vl_interface': 9.99314630576152e-06, 'learning_rate/action_model': 9.993146305761519e-05, 'epoch': 0.08}) +[2026-06-17 16:39:57,291][starVLA.training.train_starvla][INFO] - Step 219, Loss: {'train/loss': 1.1813112497329712, 'train/grad_norm_pre_clip': 11.120454788208008, 'timing/data': 0.018535702489316463, 'timing/model': 1.345865087583661, 'learning_rate/qwen_vl_interface': 9.993029677770055e-06, 'learning_rate/action_model': 9.993029677770055e-05, 'epoch': 0.08}) +[2026-06-17 16:40:01,365][starVLA.training.train_starvla][INFO] - Step 220, Loss: {'train/loss': 1.2307509183883667, 'train/grad_norm_pre_clip': 12.306347846984863, 'timing/data': 0.021317596547305584, 'timing/model': 1.0735400905832648, 'learning_rate/qwen_vl_interface': 9.992912066543433e-06, 'learning_rate/action_model': 9.992912066543433e-05, 'epoch': 0.08}) +[2026-06-17 16:40:09,031][starVLA.training.train_starvla][INFO] - Step 221, Loss: {'train/loss': 0.9772780537605286, 'train/grad_norm_pre_clip': 8.147872924804688, 'timing/data': 0.019115260802209377, 'timing/model': 2.4935932233929634, 'learning_rate/qwen_vl_interface': 9.992793472106039e-06, 'learning_rate/action_model': 9.992793472106038e-05, 'epoch': 0.08}) +[2026-06-17 16:40:13,548][starVLA.training.train_starvla][INFO] - Step 222, Loss: {'train/loss': 1.5178937911987305, 'train/grad_norm_pre_clip': 12.8884916305542, 'timing/data': 0.022251009941101074, 'timing/model': 1.0514016430824995, 'learning_rate/qwen_vl_interface': 9.992673894482452e-06, 'learning_rate/action_model': 9.992673894482451e-05, 'epoch': 0.08}) +[2026-06-17 16:40:18,356][starVLA.training.train_starvla][INFO] - Step 223, Loss: {'train/loss': 1.3084543943405151, 'train/grad_norm_pre_clip': 19.242916107177734, 'timing/data': 0.017732955515384674, 'timing/model': 1.2031815750524402, 'learning_rate/qwen_vl_interface': 9.992553333697464e-06, 'learning_rate/action_model': 9.992553333697463e-05, 'epoch': 0.08}) +[2026-06-17 16:40:22,957][starVLA.training.train_starvla][INFO] - Step 224, Loss: {'train/loss': 1.6106741428375244, 'train/grad_norm_pre_clip': 15.314715385437012, 'timing/data': 0.019800331443548203, 'timing/model': 1.4133876245468855, 'learning_rate/qwen_vl_interface': 9.992431789776067e-06, 'learning_rate/action_model': 9.992431789776066e-05, 'epoch': 0.08}) +[2026-06-17 16:40:28,419][starVLA.training.train_starvla][INFO] - Step 225, Loss: {'train/loss': 1.472246766090393, 'train/grad_norm_pre_clip': 9.626615524291992, 'timing/data': 0.01996842958033085, 'timing/model': 1.0563180996105075, 'learning_rate/qwen_vl_interface': 9.992309262743458e-06, 'learning_rate/action_model': 9.992309262743457e-05, 'epoch': 0.08}) +[2026-06-17 16:40:33,016][starVLA.training.train_starvla][INFO] - Step 226, Loss: {'train/loss': 1.4623814821243286, 'train/grad_norm_pre_clip': 10.740212440490723, 'timing/data': 0.021036459133028984, 'timing/model': 1.041211323812604, 'learning_rate/qwen_vl_interface': 9.992185752625035e-06, 'learning_rate/action_model': 9.992185752625033e-05, 'epoch': 0.08}) +[2026-06-17 16:40:37,551][starVLA.training.train_starvla][INFO] - Step 227, Loss: {'train/loss': 1.220119833946228, 'train/grad_norm_pre_clip': 7.730826377868652, 'timing/data': 0.02391899097710848, 'timing/model': 1.1578466780483723, 'learning_rate/qwen_vl_interface': 9.992061259446402e-06, 'learning_rate/action_model': 9.992061259446401e-05, 'epoch': 0.08}) +[2026-06-17 16:40:42,018][starVLA.training.train_starvla][INFO] - Step 228, Loss: {'train/loss': 1.451904535293579, 'train/grad_norm_pre_clip': 9.678773880004883, 'timing/data': 0.01888681761920452, 'timing/model': 1.434882190078497, 'learning_rate/qwen_vl_interface': 9.991935783233368e-06, 'learning_rate/action_model': 9.991935783233367e-05, 'epoch': 0.08}) +[2026-06-17 16:40:46,628][starVLA.training.train_starvla][INFO] - Step 229, Loss: {'train/loss': 1.226872444152832, 'train/grad_norm_pre_clip': 6.906081199645996, 'timing/data': 0.022100580856204033, 'timing/model': 1.4794420516118407, 'learning_rate/qwen_vl_interface': 9.991809324011943e-06, 'learning_rate/action_model': 9.991809324011943e-05, 'epoch': 0.08}) +[2026-06-17 16:40:50,851][starVLA.training.train_starvla][INFO] - Step 230, Loss: {'train/loss': 1.039343237876892, 'train/grad_norm_pre_clip': 32.188438415527344, 'timing/data': 0.01923767477273941, 'timing/model': 1.050393633544445, 'learning_rate/qwen_vl_interface': 9.991681881808343e-06, 'learning_rate/action_model': 9.991681881808343e-05, 'epoch': 0.08}) +[2026-06-17 16:40:55,248][starVLA.training.train_starvla][INFO] - Step 231, Loss: {'train/loss': 1.4017510414123535, 'train/grad_norm_pre_clip': 8.25125503540039, 'timing/data': 0.020242982544004917, 'timing/model': 1.0390386004000902, 'learning_rate/qwen_vl_interface': 9.991553456648986e-06, 'learning_rate/action_model': 9.991553456648985e-05, 'epoch': 0.08}) +[2026-06-17 16:40:59,832][starVLA.training.train_starvla][INFO] - Step 232, Loss: {'train/loss': 1.3710066080093384, 'train/grad_norm_pre_clip': 10.485749244689941, 'timing/data': 0.023013041354715824, 'timing/model': 1.0569491246715188, 'learning_rate/qwen_vl_interface': 9.991424048560497e-06, 'learning_rate/action_model': 9.991424048560496e-05, 'epoch': 0.08}) +[2026-06-17 16:41:04,373][starVLA.training.train_starvla][INFO] - Step 233, Loss: {'train/loss': 1.310539722442627, 'train/grad_norm_pre_clip': 7.468189239501953, 'timing/data': 0.02091295551508665, 'timing/model': 1.151600817218423, 'learning_rate/qwen_vl_interface': 9.991293657569698e-06, 'learning_rate/action_model': 9.991293657569698e-05, 'epoch': 0.09}) +[2026-06-17 16:41:09,216][starVLA.training.train_starvla][INFO] - Step 234, Loss: {'train/loss': 1.1694889068603516, 'train/grad_norm_pre_clip': 8.27806568145752, 'timing/data': 0.020020103082060814, 'timing/model': 1.470035900361836, 'learning_rate/qwen_vl_interface': 9.991162283703624e-06, 'learning_rate/action_model': 9.991162283703624e-05, 'epoch': 0.09}) +[2026-06-17 16:41:13,336][starVLA.training.train_starvla][INFO] - Step 235, Loss: {'train/loss': 1.194572925567627, 'train/grad_norm_pre_clip': 9.4283447265625, 'timing/data': 0.022762956097722054, 'timing/model': 1.0492434175685048, 'learning_rate/qwen_vl_interface': 9.991029926989506e-06, 'learning_rate/action_model': 9.991029926989506e-05, 'epoch': 0.09}) +[2026-06-17 16:41:17,949][starVLA.training.train_starvla][INFO] - Step 236, Loss: {'train/loss': 1.1788359880447388, 'train/grad_norm_pre_clip': 13.024962425231934, 'timing/data': 0.017263876274228096, 'timing/model': 1.0348101276904345, 'learning_rate/qwen_vl_interface': 9.990896587454783e-06, 'learning_rate/action_model': 9.990896587454783e-05, 'epoch': 0.09}) +[2026-06-17 16:41:22,403][starVLA.training.train_starvla][INFO] - Step 237, Loss: {'train/loss': 0.9624062180519104, 'train/grad_norm_pre_clip': 7.881287574768066, 'timing/data': 0.01775598619133234, 'timing/model': 1.0677618375048041, 'learning_rate/qwen_vl_interface': 9.990762265127096e-06, 'learning_rate/action_model': 9.990762265127096e-05, 'epoch': 0.09}) +[2026-06-17 16:41:27,102][starVLA.training.train_starvla][INFO] - Step 238, Loss: {'train/loss': 1.4114139080047607, 'train/grad_norm_pre_clip': 11.151866912841797, 'timing/data': 0.024169557727873325, 'timing/model': 1.4735839860513806, 'learning_rate/qwen_vl_interface': 9.990626960034291e-06, 'learning_rate/action_model': 9.990626960034291e-05, 'epoch': 0.09}) +[2026-06-17 16:41:31,582][starVLA.training.train_starvla][INFO] - Step 239, Loss: {'train/loss': 1.3952860832214355, 'train/grad_norm_pre_clip': 12.549079895019531, 'timing/data': 0.019650958478450775, 'timing/model': 1.3363809594884515, 'learning_rate/qwen_vl_interface': 9.990490672204417e-06, 'learning_rate/action_model': 9.990490672204415e-05, 'epoch': 0.09}) +[2026-06-17 16:41:35,721][starVLA.training.train_starvla][INFO] - Step 240, Loss: {'train/loss': 1.462239384651184, 'train/grad_norm_pre_clip': 13.365047454833984, 'timing/data': 0.018795699812471867, 'timing/model': 1.047412695363164, 'learning_rate/qwen_vl_interface': 9.99035340166572e-06, 'learning_rate/action_model': 9.990353401665721e-05, 'epoch': 0.09}) +[2026-06-17 16:41:40,308][starVLA.training.train_starvla][INFO] - Step 241, Loss: {'train/loss': 1.228817343711853, 'train/grad_norm_pre_clip': 9.113723754882812, 'timing/data': 0.02163618803024292, 'timing/model': 1.0690416153520346, 'learning_rate/qwen_vl_interface': 9.990215148446667e-06, 'learning_rate/action_model': 9.990215148446666e-05, 'epoch': 0.09}) +[2026-06-17 16:41:44,840][starVLA.training.train_starvla][INFO] - Step 242, Loss: {'train/loss': 1.0109270811080933, 'train/grad_norm_pre_clip': 32.991024017333984, 'timing/data': 0.02274433057755232, 'timing/model': 1.0609033117070794, 'learning_rate/qwen_vl_interface': 9.990075912575912e-06, 'learning_rate/action_model': 9.990075912575911e-05, 'epoch': 0.09}) +[2026-06-17 16:41:49,387][starVLA.training.train_starvla][INFO] - Step 243, Loss: {'train/loss': 1.0314033031463623, 'train/grad_norm_pre_clip': 8.319636344909668, 'timing/data': 0.01959955506026745, 'timing/model': 1.0345841571688652, 'learning_rate/qwen_vl_interface': 9.98993569408232e-06, 'learning_rate/action_model': 9.989935694082321e-05, 'epoch': 0.09}) +[2026-06-17 16:41:53,796][starVLA.training.train_starvla][INFO] - Step 244, Loss: {'train/loss': 1.6455583572387695, 'train/grad_norm_pre_clip': 16.89745330810547, 'timing/data': 0.023899742402136326, 'timing/model': 1.4141507474705577, 'learning_rate/qwen_vl_interface': 9.989794492994958e-06, 'learning_rate/action_model': 9.989794492994958e-05, 'epoch': 0.09}) +[2026-06-17 16:41:57,972][starVLA.training.train_starvla][INFO] - Step 245, Loss: {'train/loss': 1.2569621801376343, 'train/grad_norm_pre_clip': 14.307209014892578, 'timing/data': 0.017644304782152176, 'timing/model': 1.0434254994615912, 'learning_rate/qwen_vl_interface': 9.989652309343097e-06, 'learning_rate/action_model': 9.989652309343098e-05, 'epoch': 0.09}) +[2026-06-17 16:42:02,378][starVLA.training.train_starvla][INFO] - Step 246, Loss: {'train/loss': 1.1921601295471191, 'train/grad_norm_pre_clip': 16.852922439575195, 'timing/data': 0.01812308467924595, 'timing/model': 1.0286144334822893, 'learning_rate/qwen_vl_interface': 9.989509143156213e-06, 'learning_rate/action_model': 9.989509143156213e-05, 'epoch': 0.09}) +[2026-06-17 16:42:07,487][starVLA.training.train_starvla][INFO] - Step 247, Loss: {'train/loss': 1.4353153705596924, 'train/grad_norm_pre_clip': 10.184931755065918, 'timing/data': 0.025105613283813, 'timing/model': 1.6467594858258963, 'learning_rate/qwen_vl_interface': 9.989364994463983e-06, 'learning_rate/action_model': 9.989364994463983e-05, 'epoch': 0.09}) +[2026-06-17 16:42:12,018][starVLA.training.train_starvla][INFO] - Step 248, Loss: {'train/loss': 1.327021598815918, 'train/grad_norm_pre_clip': 10.829435348510742, 'timing/data': 0.02626930270344019, 'timing/model': 1.4617978688329458, 'learning_rate/qwen_vl_interface': 9.989219863296291e-06, 'learning_rate/action_model': 9.989219863296291e-05, 'epoch': 0.09}) +[2026-06-17 16:42:16,524][starVLA.training.train_starvla][INFO] - Step 249, Loss: {'train/loss': 1.3385151624679565, 'train/grad_norm_pre_clip': 21.3526668548584, 'timing/data': 0.024142488837242126, 'timing/model': 1.3678339812904596, 'learning_rate/qwen_vl_interface': 9.98907374968322e-06, 'learning_rate/action_model': 9.989073749683219e-05, 'epoch': 0.09}) +[2026-06-17 16:42:20,701][starVLA.training.train_starvla][INFO] - Step 250, Loss: {'train/loss': 1.2759144306182861, 'train/grad_norm_pre_clip': 7.180746555328369, 'timing/data': 0.020010659471154213, 'timing/model': 1.1145686665549874, 'learning_rate/qwen_vl_interface': 9.988926653655062e-06, 'learning_rate/action_model': 9.988926653655062e-05, 'epoch': 0.09}) +[2026-06-17 16:42:25,028][starVLA.training.train_starvla][INFO] - Step 251, Loss: {'train/loss': 1.287552833557129, 'train/grad_norm_pre_clip': 9.288775444030762, 'timing/data': 0.023240976966917515, 'timing/model': 1.022827334702015, 'learning_rate/qwen_vl_interface': 9.988778575242309e-06, 'learning_rate/action_model': 9.988778575242309e-05, 'epoch': 0.09}) +[2026-06-17 16:42:30,840][starVLA.training.train_starvla][INFO] - Step 252, Loss: {'train/loss': 1.486104130744934, 'train/grad_norm_pre_clip': 10.601223945617676, 'timing/data': 0.017992272973060608, 'timing/model': 1.0663575874641538, 'learning_rate/qwen_vl_interface': 9.98862951447566e-06, 'learning_rate/action_model': 9.988629514475659e-05, 'epoch': 0.09}) +[2026-06-17 16:42:36,896][starVLA.training.train_starvla][INFO] - Step 253, Loss: {'train/loss': 1.3447731733322144, 'train/grad_norm_pre_clip': 14.936062812805176, 'timing/data': 0.02198015619069338, 'timing/model': 2.7303446447476745, 'learning_rate/qwen_vl_interface': 9.98847947138601e-06, 'learning_rate/action_model': 9.98847947138601e-05, 'epoch': 0.09}) +[2026-06-17 16:42:41,643][starVLA.training.train_starvla][INFO] - Step 254, Loss: {'train/loss': 1.1618252992630005, 'train/grad_norm_pre_clip': 9.555729866027832, 'timing/data': 0.019778831861913204, 'timing/model': 1.3394630942493677, 'learning_rate/qwen_vl_interface': 9.98832844600447e-06, 'learning_rate/action_model': 9.98832844600447e-05, 'epoch': 0.09}) +[2026-06-17 16:42:45,781][starVLA.training.train_starvla][INFO] - Step 255, Loss: {'train/loss': 1.3193864822387695, 'train/grad_norm_pre_clip': 8.594738960266113, 'timing/data': 0.021475081332027912, 'timing/model': 1.0426550237461925, 'learning_rate/qwen_vl_interface': 9.988176438362344e-06, 'learning_rate/action_model': 9.988176438362343e-05, 'epoch': 0.09}) +[2026-06-17 16:42:50,297][starVLA.training.train_starvla][INFO] - Step 256, Loss: {'train/loss': 1.187424659729004, 'train/grad_norm_pre_clip': 15.13784408569336, 'timing/data': 0.026794061064720154, 'timing/model': 1.0782237434759736, 'learning_rate/qwen_vl_interface': 9.988023448491143e-06, 'learning_rate/action_model': 9.988023448491143e-05, 'epoch': 0.09}) +[2026-06-17 16:42:54,670][starVLA.training.train_starvla][INFO] - Step 257, Loss: {'train/loss': 1.3263241052627563, 'train/grad_norm_pre_clip': 8.109904289245605, 'timing/data': 0.01843644492328167, 'timing/model': 1.0234491154551506, 'learning_rate/qwen_vl_interface': 9.987869476422582e-06, 'learning_rate/action_model': 9.987869476422583e-05, 'epoch': 0.09}) +[2026-06-17 16:42:59,264][starVLA.training.train_starvla][INFO] - Step 258, Loss: {'train/loss': 0.8763518333435059, 'train/grad_norm_pre_clip': 9.881486892700195, 'timing/data': 0.01950627937912941, 'timing/model': 1.51963781286031, 'learning_rate/qwen_vl_interface': 9.987714522188583e-06, 'learning_rate/action_model': 9.987714522188582e-05, 'epoch': 0.09}) +[2026-06-17 16:43:04,033][starVLA.training.train_starvla][INFO] - Step 259, Loss: {'train/loss': 1.030173659324646, 'train/grad_norm_pre_clip': 7.861474990844727, 'timing/data': 0.0241128196939826, 'timing/model': 1.3792606750503182, 'learning_rate/qwen_vl_interface': 9.987558585821264e-06, 'learning_rate/action_model': 9.987558585821263e-05, 'epoch': 0.09}) +[2026-06-17 16:43:09,443][starVLA.training.train_starvla][INFO] - Step 260, Loss: {'train/loss': 1.323501467704773, 'train/grad_norm_pre_clip': 10.15136432647705, 'timing/data': 0.01802821271121502, 'timing/model': 1.0486095556989312, 'learning_rate/qwen_vl_interface': 9.987401667352952e-06, 'learning_rate/action_model': 9.987401667352951e-05, 'epoch': 0.1}) +[2026-06-17 16:43:14,338][starVLA.training.train_starvla][INFO] - Step 261, Loss: {'train/loss': 1.3310476541519165, 'train/grad_norm_pre_clip': 8.232331275939941, 'timing/data': 0.020947802811861038, 'timing/model': 1.0770499743521214, 'learning_rate/qwen_vl_interface': 9.987243766816175e-06, 'learning_rate/action_model': 9.987243766816176e-05, 'epoch': 0.1}) +[2026-06-17 16:43:18,731][starVLA.training.train_starvla][INFO] - Step 262, Loss: {'train/loss': 1.0916619300842285, 'train/grad_norm_pre_clip': 9.459064483642578, 'timing/data': 0.018580111674964428, 'timing/model': 1.0389123298227787, 'learning_rate/qwen_vl_interface': 9.987084884243671e-06, 'learning_rate/action_model': 9.987084884243672e-05, 'epoch': 0.1}) +[2026-06-17 16:43:23,060][starVLA.training.train_starvla][INFO] - Step 263, Loss: {'train/loss': 1.1397985219955444, 'train/grad_norm_pre_clip': 8.067449569702148, 'timing/data': 0.02148453611880541, 'timing/model': 1.0295685678720474, 'learning_rate/qwen_vl_interface': 9.986925019668374e-06, 'learning_rate/action_model': 9.986925019668373e-05, 'epoch': 0.1}) +[2026-06-17 16:43:27,425][starVLA.training.train_starvla][INFO] - Step 264, Loss: {'train/loss': 1.0716643333435059, 'train/grad_norm_pre_clip': 8.619841575622559, 'timing/data': 0.021765022538602352, 'timing/model': 1.3718104306608438, 'learning_rate/qwen_vl_interface': 9.98676417312342e-06, 'learning_rate/action_model': 9.986764173123419e-05, 'epoch': 0.1}) +[2026-06-17 16:43:31,567][starVLA.training.train_starvla][INFO] - Step 265, Loss: {'train/loss': 1.4111756086349487, 'train/grad_norm_pre_clip': 17.889551162719727, 'timing/data': 0.019885382615029812, 'timing/model': 1.041888801380992, 'learning_rate/qwen_vl_interface': 9.986602344642157e-06, 'learning_rate/action_model': 9.986602344642156e-05, 'epoch': 0.1}) +[2026-06-17 16:43:35,988][starVLA.training.train_starvla][INFO] - Step 266, Loss: {'train/loss': 1.5200378894805908, 'train/grad_norm_pre_clip': 13.862203598022461, 'timing/data': 0.018173920921981335, 'timing/model': 1.022709065116942, 'learning_rate/qwen_vl_interface': 9.986439534258132e-06, 'learning_rate/action_model': 9.986439534258131e-05, 'epoch': 0.1}) +[2026-06-17 16:43:40,623][starVLA.training.train_starvla][INFO] - Step 267, Loss: {'train/loss': 1.5686228275299072, 'train/grad_norm_pre_clip': 12.851421356201172, 'timing/data': 0.02368111629039049, 'timing/model': 1.113292271271348, 'learning_rate/qwen_vl_interface': 9.986275742005093e-06, 'learning_rate/action_model': 9.986275742005093e-05, 'epoch': 0.1}) +[2026-06-17 16:43:45,226][starVLA.training.train_starvla][INFO] - Step 268, Loss: {'train/loss': 1.5011793375015259, 'train/grad_norm_pre_clip': 14.216830253601074, 'timing/data': 0.031578632071614265, 'timing/model': 1.4447006117552519, 'learning_rate/qwen_vl_interface': 9.986110967916997e-06, 'learning_rate/action_model': 9.986110967916996e-05, 'epoch': 0.1}) +[2026-06-17 16:43:50,951][starVLA.training.train_starvla][INFO] - Step 269, Loss: {'train/loss': 1.335407018661499, 'train/grad_norm_pre_clip': 9.479150772094727, 'timing/data': 0.02021370269358158, 'timing/model': 1.5678961891680956, 'learning_rate/qwen_vl_interface': 9.985945212028002e-06, 'learning_rate/action_model': 9.985945212028001e-05, 'epoch': 0.1}) +[2026-06-17 16:43:56,534][starVLA.training.train_starvla][INFO] - Step 270, Loss: {'train/loss': 1.352330207824707, 'train/grad_norm_pre_clip': 9.821697235107422, 'timing/data': 0.02003518585115671, 'timing/model': 1.034951894544065, 'learning_rate/qwen_vl_interface': 9.985778474372467e-06, 'learning_rate/action_model': 9.985778474372466e-05, 'epoch': 0.1}) +[2026-06-17 16:44:01,256][starVLA.training.train_starvla][INFO] - Step 271, Loss: {'train/loss': 0.9946938753128052, 'train/grad_norm_pre_clip': 8.450236320495605, 'timing/data': 0.02491983026266098, 'timing/model': 1.065192592330277, 'learning_rate/qwen_vl_interface': 9.98561075498496e-06, 'learning_rate/action_model': 9.98561075498496e-05, 'epoch': 0.1}) +[2026-06-17 16:44:05,718][starVLA.training.train_starvla][INFO] - Step 272, Loss: {'train/loss': 1.4415109157562256, 'train/grad_norm_pre_clip': 10.77132797241211, 'timing/data': 0.018494075164198875, 'timing/model': 1.0626534819602966, 'learning_rate/qwen_vl_interface': 9.985442053900246e-06, 'learning_rate/action_model': 9.985442053900246e-05, 'epoch': 0.1}) +[2026-06-17 16:44:11,628][starVLA.training.train_starvla][INFO] - Step 273, Loss: {'train/loss': 1.0931683778762817, 'train/grad_norm_pre_clip': 11.27292251586914, 'timing/data': 0.020030870102345943, 'timing/model': 1.042377196252346, 'learning_rate/qwen_vl_interface': 9.9852723711533e-06, 'learning_rate/action_model': 9.9852723711533e-05, 'epoch': 0.1}) +[2026-06-17 16:44:16,132][starVLA.training.train_starvla][INFO] - Step 274, Loss: {'train/loss': 1.6468467712402344, 'train/grad_norm_pre_clip': 18.561473846435547, 'timing/data': 0.02117234282195568, 'timing/model': 1.4397391229867935, 'learning_rate/qwen_vl_interface': 9.985101706779297e-06, 'learning_rate/action_model': 9.985101706779296e-05, 'epoch': 0.1}) +[2026-06-17 16:44:20,289][starVLA.training.train_starvla][INFO] - Step 275, Loss: {'train/loss': 1.438976764678955, 'train/grad_norm_pre_clip': 11.157441139221191, 'timing/data': 0.021528447046875954, 'timing/model': 1.0401407834142447, 'learning_rate/qwen_vl_interface': 9.984930060813613e-06, 'learning_rate/action_model': 9.984930060813613e-05, 'epoch': 0.1}) +[2026-06-17 16:44:24,610][starVLA.training.train_starvla][INFO] - Step 276, Loss: {'train/loss': 1.2774332761764526, 'train/grad_norm_pre_clip': 8.909632682800293, 'timing/data': 0.018656447529792786, 'timing/model': 1.0350521923974156, 'learning_rate/qwen_vl_interface': 9.984757433291832e-06, 'learning_rate/action_model': 9.984757433291832e-05, 'epoch': 0.1}) +[2026-06-17 16:44:29,997][starVLA.training.train_starvla][INFO] - Step 277, Loss: {'train/loss': 1.0448265075683594, 'train/grad_norm_pre_clip': 12.12403678894043, 'timing/data': 0.021544141694903374, 'timing/model': 1.0653739431872964, 'learning_rate/qwen_vl_interface': 9.984583824249742e-06, 'learning_rate/action_model': 9.984583824249742e-05, 'epoch': 0.1}) +[2026-06-17 16:44:34,501][starVLA.training.train_starvla][INFO] - Step 278, Loss: {'train/loss': 1.004251480102539, 'train/grad_norm_pre_clip': 10.445634841918945, 'timing/data': 0.020077913999557495, 'timing/model': 1.3865151703357697, 'learning_rate/qwen_vl_interface': 9.98440923372333e-06, 'learning_rate/action_model': 9.984409233723328e-05, 'epoch': 0.1}) +[2026-06-17 16:44:39,107][starVLA.training.train_starvla][INFO] - Step 279, Loss: {'train/loss': 1.3233617544174194, 'train/grad_norm_pre_clip': 9.27926254272461, 'timing/data': 0.018735546618700027, 'timing/model': 1.5923644546419382, 'learning_rate/qwen_vl_interface': 9.984233661748788e-06, 'learning_rate/action_model': 9.984233661748788e-05, 'epoch': 0.1}) +[2026-06-17 16:44:43,200][starVLA.training.train_starvla][INFO] - Step 280, Loss: {'train/loss': 1.0996711254119873, 'train/grad_norm_pre_clip': 8.881486892700195, 'timing/data': 0.018739488907158375, 'timing/model': 1.1139379600062966, 'learning_rate/qwen_vl_interface': 9.984057108362516e-06, 'learning_rate/action_model': 9.984057108362515e-05, 'epoch': 0.1}) +[2026-06-17 16:44:48,718][starVLA.training.train_starvla][INFO] - Step 281, Loss: {'train/loss': 1.2403892278671265, 'train/grad_norm_pre_clip': 12.443300247192383, 'timing/data': 0.019925905391573906, 'timing/model': 2.2371337423101068, 'learning_rate/qwen_vl_interface': 9.98387957360111e-06, 'learning_rate/action_model': 9.98387957360111e-05, 'epoch': 0.1}) +[2026-06-17 16:44:53,620][starVLA.training.train_starvla][INFO] - Step 282, Loss: {'train/loss': 1.3251315355300903, 'train/grad_norm_pre_clip': 10.154308319091797, 'timing/data': 0.020687013864517212, 'timing/model': 1.0512979915365577, 'learning_rate/qwen_vl_interface': 9.983701057501375e-06, 'learning_rate/action_model': 9.983701057501375e-05, 'epoch': 0.1}) +[2026-06-17 16:44:57,995][starVLA.training.train_starvla][INFO] - Step 283, Loss: {'train/loss': 1.2635624408721924, 'train/grad_norm_pre_clip': 13.313843727111816, 'timing/data': 0.020554437302052975, 'timing/model': 1.0411002160981297, 'learning_rate/qwen_vl_interface': 9.983521560100317e-06, 'learning_rate/action_model': 9.983521560100316e-05, 'epoch': 0.1}) +[2026-06-17 16:45:02,451][starVLA.training.train_starvla][INFO] - Step 284, Loss: {'train/loss': 1.4061411619186401, 'train/grad_norm_pre_clip': 7.4975786209106445, 'timing/data': 0.01838146522641182, 'timing/model': 1.3367540892213583, 'learning_rate/qwen_vl_interface': 9.983341081435145e-06, 'learning_rate/action_model': 9.983341081435145e-05, 'epoch': 0.1}) +[2026-06-17 16:45:06,453][starVLA.training.train_starvla][INFO] - Step 285, Loss: {'train/loss': 1.3154574632644653, 'train/grad_norm_pre_clip': 11.357002258300781, 'timing/data': 0.017460140399634838, 'timing/model': 1.0553448414430022, 'learning_rate/qwen_vl_interface': 9.983159621543275e-06, 'learning_rate/action_model': 9.983159621543275e-05, 'epoch': 0.1}) +[2026-06-17 16:45:10,914][starVLA.training.train_starvla][INFO] - Step 286, Loss: {'train/loss': 1.2911806106567383, 'train/grad_norm_pre_clip': 9.384349822998047, 'timing/data': 0.019330281764268875, 'timing/model': 1.0393771547824144, 'learning_rate/qwen_vl_interface': 9.982977180462324e-06, 'learning_rate/action_model': 9.982977180462323e-05, 'epoch': 0.1}) +[2026-06-17 16:45:15,317][starVLA.training.train_starvla][INFO] - Step 287, Loss: {'train/loss': 1.1828505992889404, 'train/grad_norm_pre_clip': 10.518908500671387, 'timing/data': 0.01710444502532482, 'timing/model': 1.0161370290443301, 'learning_rate/qwen_vl_interface': 9.982793758230107e-06, 'learning_rate/action_model': 9.982793758230108e-05, 'epoch': 0.11}) +[2026-06-17 16:45:19,937][starVLA.training.train_starvla][INFO] - Step 288, Loss: {'train/loss': 1.3399662971496582, 'train/grad_norm_pre_clip': 11.546113967895508, 'timing/data': 0.016631344333291054, 'timing/model': 1.3315364765003324, 'learning_rate/qwen_vl_interface': 9.982609354884653e-06, 'learning_rate/action_model': 9.982609354884653e-05, 'epoch': 0.11}) +[2026-06-17 16:45:24,488][starVLA.training.train_starvla][INFO] - Step 289, Loss: {'train/loss': 1.2278469800949097, 'train/grad_norm_pre_clip': 9.382172584533691, 'timing/data': 0.020533515140414238, 'timing/model': 1.4417502423748374, 'learning_rate/qwen_vl_interface': 9.982423970464187e-06, 'learning_rate/action_model': 9.982423970464186e-05, 'epoch': 0.11}) +[2026-06-17 16:45:28,679][starVLA.training.train_starvla][INFO] - Step 290, Loss: {'train/loss': 1.3908662796020508, 'train/grad_norm_pre_clip': 12.478296279907227, 'timing/data': 0.019967776723206043, 'timing/model': 1.0877296859398484, 'learning_rate/qwen_vl_interface': 9.98223760500714e-06, 'learning_rate/action_model': 9.982237605007139e-05, 'epoch': 0.11}) +[2026-06-17 16:45:33,272][starVLA.training.train_starvla][INFO] - Step 291, Loss: {'train/loss': 1.4522300958633423, 'train/grad_norm_pre_clip': 14.358949661254883, 'timing/data': 0.02018686104565859, 'timing/model': 1.0438761347904801, 'learning_rate/qwen_vl_interface': 9.982050258552146e-06, 'learning_rate/action_model': 9.982050258552145e-05, 'epoch': 0.11}) +[2026-06-17 16:45:38,015][starVLA.training.train_starvla][INFO] - Step 292, Loss: {'train/loss': 1.267259955406189, 'train/grad_norm_pre_clip': 8.730685234069824, 'timing/data': 0.01910051330924034, 'timing/model': 1.2828215770423412, 'learning_rate/qwen_vl_interface': 9.981861931138039e-06, 'learning_rate/action_model': 9.981861931138039e-05, 'epoch': 0.11}) +[2026-06-17 16:45:42,660][starVLA.training.train_starvla][INFO] - Step 293, Loss: {'train/loss': 1.0455846786499023, 'train/grad_norm_pre_clip': 7.9706621170043945, 'timing/data': 0.017468532547354698, 'timing/model': 1.0983629981055856, 'learning_rate/qwen_vl_interface': 9.981672622803863e-06, 'learning_rate/action_model': 9.981672622803862e-05, 'epoch': 0.11}) +[2026-06-17 16:45:47,226][starVLA.training.train_starvla][INFO] - Step 294, Loss: {'train/loss': 1.2434899806976318, 'train/grad_norm_pre_clip': 8.20085334777832, 'timing/data': 0.018676063045859337, 'timing/model': 1.486864710226655, 'learning_rate/qwen_vl_interface': 9.981482333588863e-06, 'learning_rate/action_model': 9.981482333588862e-05, 'epoch': 0.11}) +[2026-06-17 16:45:51,315][starVLA.training.train_starvla][INFO] - Step 295, Loss: {'train/loss': 1.0689102411270142, 'train/grad_norm_pre_clip': 9.825865745544434, 'timing/data': 0.020916716195642948, 'timing/model': 1.0378757985308766, 'learning_rate/qwen_vl_interface': 9.981291063532481e-06, 'learning_rate/action_model': 9.981291063532481e-05, 'epoch': 0.11}) +[2026-06-17 16:45:55,693][starVLA.training.train_starvla][INFO] - Step 296, Loss: {'train/loss': 1.0203135013580322, 'train/grad_norm_pre_clip': 10.620128631591797, 'timing/data': 0.020335829816758633, 'timing/model': 1.022194853052497, 'learning_rate/qwen_vl_interface': 9.981098812674372e-06, 'learning_rate/action_model': 9.981098812674371e-05, 'epoch': 0.11}) +[2026-06-17 16:46:00,538][starVLA.training.train_starvla][INFO] - Step 297, Loss: {'train/loss': 1.3517861366271973, 'train/grad_norm_pre_clip': 8.026640892028809, 'timing/data': 0.018831402994692326, 'timing/model': 1.544625292532146, 'learning_rate/qwen_vl_interface': 9.980905581054388e-06, 'learning_rate/action_model': 9.980905581054387e-05, 'epoch': 0.11}) +[2026-06-17 16:46:05,031][starVLA.training.train_starvla][INFO] - Step 298, Loss: {'train/loss': 1.2038960456848145, 'train/grad_norm_pre_clip': 8.220602035522461, 'timing/data': 0.02211402077227831, 'timing/model': 1.4426035806536674, 'learning_rate/qwen_vl_interface': 9.980711368712586e-06, 'learning_rate/action_model': 9.980711368712585e-05, 'epoch': 0.11}) +[2026-06-17 16:46:09,370][starVLA.training.train_starvla][INFO] - Step 299, Loss: {'train/loss': 1.3002350330352783, 'train/grad_norm_pre_clip': 10.136574745178223, 'timing/data': 0.019451008178293705, 'timing/model': 1.3417269000783563, 'learning_rate/qwen_vl_interface': 9.980516175689228e-06, 'learning_rate/action_model': 9.980516175689227e-05, 'epoch': 0.11}) +[2026-06-17 16:48:45,485][starVLA.training.train_starvla][INFO] - Step 300, Loss: {'train/loss': 1.4031895399093628, 'train/grad_norm_pre_clip': 10.576218605041504, 'eval/loss': 1.1902402639389038, 'eval/latency_2/loss': 1.1902402639389038, 'eval/demon_fire_precision': 0.7170212765957447, 'eval/demon_fire_recall': 0.9870292887029288, 'eval/demon_fire_f1': 0.8306338028169015, 'eval/demon_fire_rate_model': 0.9154145798553144, 'eval/demon_fire_rate_teacher': 0.6649972175848636, 'eval/demon_left_precision': 0.8146853146853147, 'eval/demon_left_recall': 0.4905263157894737, 'eval/demon_left_f1': 0.6123521681997373, 'eval/demon_left_rate_model': 0.15915414579855314, 'eval/demon_left_rate_teacher': 0.2643294379521425, 'eval/demon_right_precision': 0.7486338797814208, 'eval/demon_right_recall': 0.46440677966101696, 'eval/demon_right_f1': 0.5732217573221756, 'eval/demon_right_rate_model': 0.15275459098497496, 'eval/demon_right_rate_teacher': 0.24624373956594323, 'eval/demon_move_f1': 0.5927869627609564, 'eval/demon_cc_f1': 0.7413920811729962, 'eval/latency_2/demon_fire_precision': 0.7170212765957447, 'eval/latency_2/demon_fire_recall': 0.9870292887029288, 'eval/latency_2/demon_fire_f1': 0.8306338028169015, 'eval/latency_2/demon_fire_rate_model': 0.9154145798553144, 'eval/latency_2/demon_fire_rate_teacher': 0.6649972175848636, 'eval/latency_2/demon_left_precision': 0.8146853146853147, 'eval/latency_2/demon_left_recall': 0.4905263157894737, 'eval/latency_2/demon_left_f1': 0.6123521681997373, 'eval/latency_2/demon_left_rate_model': 0.15915414579855314, 'eval/latency_2/demon_left_rate_teacher': 0.2643294379521425, 'eval/latency_2/demon_right_precision': 0.7486338797814208, 'eval/latency_2/demon_right_recall': 0.46440677966101696, 'eval/latency_2/demon_right_f1': 0.5732217573221756, 'eval/latency_2/demon_right_rate_model': 0.15275459098497496, 'eval/latency_2/demon_right_rate_teacher': 0.24624373956594323, 'eval/latency_2/demon_move_f1': 0.5927869627609564, 'eval/latency_2/demon_cc_f1': 0.7413920811729962, 'timing/data': 0.02050892263650894, 'timing/model': 1.0387649135664105, 'learning_rate/qwen_vl_interface': 9.980320002024774e-06, 'learning_rate/action_model': 9.980320002024774e-05, 'epoch': 0.11}) +[2026-06-17 16:48:45,490][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_300_state +[2026-06-17 16:48:45,497][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 16:48:54,202][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_300_state/model.safetensors +[2026-06-17 16:49:08,592][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_300_state/optimizer.bin +[2026-06-17 16:49:08,593][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_300_state/sampler.bin +[2026-06-17 16:49:08,594][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_300_state/sampler_1.bin +[2026-06-17 16:49:08,596][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_300_state/random_states_0.pkl +[2026-06-17 16:49:08,601][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 300 +[2026-06-17 16:49:08,602][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 16:49:08,618][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 16:49:12,689][starVLA.training.train_starvla][INFO] - Step 301, Loss: {'train/loss': 1.1711692810058594, 'train/grad_norm_pre_clip': 12.290803909301758, 'timing/data': 0.02197389118373394, 'timing/model': 1.062127823010087, 'learning_rate/qwen_vl_interface': 9.980122847759896e-06, 'learning_rate/action_model': 9.980122847759896e-05, 'epoch': 0.11}) +[2026-06-17 16:49:18,672][starVLA.training.train_starvla][INFO] - Step 302, Loss: {'train/loss': 1.177423119544983, 'train/grad_norm_pre_clip': 7.786303997039795, 'timing/data': 0.02061640378087759, 'timing/model': 1.1099088517948985, 'learning_rate/qwen_vl_interface': 9.97992471293546e-06, 'learning_rate/action_model': 9.97992471293546e-05, 'epoch': 0.11}) +[2026-06-17 16:49:23,250][starVLA.training.train_starvla][INFO] - Step 303, Loss: {'train/loss': 1.135927438735962, 'train/grad_norm_pre_clip': 9.356779098510742, 'timing/data': 0.01929857674986124, 'timing/model': 1.1315160254016519, 'learning_rate/qwen_vl_interface': 9.979725597592543e-06, 'learning_rate/action_model': 9.979725597592541e-05, 'epoch': 0.11}) +[2026-06-17 16:49:29,882][starVLA.training.train_starvla][INFO] - Step 304, Loss: {'train/loss': 1.250555396080017, 'train/grad_norm_pre_clip': 54.405670166015625, 'timing/data': 0.02089452650398016, 'timing/model': 1.9868948496878147, 'learning_rate/qwen_vl_interface': 9.979525501772418e-06, 'learning_rate/action_model': 9.979525501772417e-05, 'epoch': 0.11}) +[2026-06-17 16:49:36,571][starVLA.training.train_starvla][INFO] - Step 305, Loss: {'train/loss': 1.1493542194366455, 'train/grad_norm_pre_clip': 13.505485534667969, 'timing/data': 0.019520653411746025, 'timing/model': 1.6041869865730405, 'learning_rate/qwen_vl_interface': 9.97932442551657e-06, 'learning_rate/action_model': 9.979324425516569e-05, 'epoch': 0.11}) +[2026-06-17 16:49:42,856][starVLA.training.train_starvla][INFO] - Step 306, Loss: {'train/loss': 1.1753811836242676, 'train/grad_norm_pre_clip': 10.919175148010254, 'timing/data': 0.021001403219997883, 'timing/model': 1.0301212146878242, 'learning_rate/qwen_vl_interface': 9.979122368866678e-06, 'learning_rate/action_model': 9.979122368866679e-05, 'epoch': 0.11}) +[2026-06-17 16:49:48,523][starVLA.training.train_starvla][INFO] - Step 307, Loss: {'train/loss': 1.8898504972457886, 'train/grad_norm_pre_clip': 15.559965133666992, 'timing/data': 0.023189746774733067, 'timing/model': 1.1048526018857956, 'learning_rate/qwen_vl_interface': 9.978919331864631e-06, 'learning_rate/action_model': 9.97891933186463e-05, 'epoch': 0.11}) +[2026-06-17 16:49:52,943][starVLA.training.train_starvla][INFO] - Step 308, Loss: {'train/loss': 1.3053933382034302, 'train/grad_norm_pre_clip': 13.644519805908203, 'timing/data': 0.021087494678795338, 'timing/model': 1.0308454036712646, 'learning_rate/qwen_vl_interface': 9.978715314552518e-06, 'learning_rate/action_model': 9.978715314552518e-05, 'epoch': 0.11}) +[2026-06-17 16:49:57,772][starVLA.training.train_starvla][INFO] - Step 309, Loss: {'train/loss': 1.1829241514205933, 'train/grad_norm_pre_clip': 8.752924919128418, 'timing/data': 0.01881922036409378, 'timing/model': 1.0257620410993695, 'learning_rate/qwen_vl_interface': 9.978510316972631e-06, 'learning_rate/action_model': 9.978510316972631e-05, 'epoch': 0.11}) +[2026-06-17 16:50:02,300][starVLA.training.train_starvla][INFO] - Step 310, Loss: {'train/loss': 1.2027720212936401, 'train/grad_norm_pre_clip': 11.938055038452148, 'timing/data': 0.02182506676763296, 'timing/model': 1.4965572925284505, 'learning_rate/qwen_vl_interface': 9.978304339167467e-06, 'learning_rate/action_model': 9.978304339167467e-05, 'epoch': 0.11}) +[2026-06-17 16:50:06,403][starVLA.training.train_starvla][INFO] - Step 311, Loss: {'train/loss': 1.0485141277313232, 'train/grad_norm_pre_clip': 12.81250286102295, 'timing/data': 0.021826249547302723, 'timing/model': 1.0364530682563782, 'learning_rate/qwen_vl_interface': 9.978097381179726e-06, 'learning_rate/action_model': 9.978097381179726e-05, 'epoch': 0.11}) +[2026-06-17 16:50:10,814][starVLA.training.train_starvla][INFO] - Step 312, Loss: {'train/loss': 1.289702296257019, 'train/grad_norm_pre_clip': 10.966578483581543, 'timing/data': 0.019797705113887787, 'timing/model': 1.0773055786266923, 'learning_rate/qwen_vl_interface': 9.977889443052312e-06, 'learning_rate/action_model': 9.977889443052311e-05, 'epoch': 0.11}) +[2026-06-17 16:50:15,141][starVLA.training.train_starvla][INFO] - Step 313, Loss: {'train/loss': 1.4515302181243896, 'train/grad_norm_pre_clip': 10.584626197814941, 'timing/data': 0.022995577193796635, 'timing/model': 1.0253751846030354, 'learning_rate/qwen_vl_interface': 9.977680524828327e-06, 'learning_rate/action_model': 9.977680524828327e-05, 'epoch': 0.11}) +[2026-06-17 16:50:19,587][starVLA.training.train_starvla][INFO] - Step 314, Loss: {'train/loss': 1.1209359169006348, 'train/grad_norm_pre_clip': 9.81175708770752, 'timing/data': 0.023195287212729454, 'timing/model': 1.0487257270142436, 'learning_rate/qwen_vl_interface': 9.977470626551083e-06, 'learning_rate/action_model': 9.977470626551083e-05, 'epoch': 0.11}) +[2026-06-17 16:50:23,979][starVLA.training.train_starvla][INFO] - Step 315, Loss: {'train/loss': 1.219191074371338, 'train/grad_norm_pre_clip': 10.310972213745117, 'timing/data': 0.025622612796723843, 'timing/model': 1.3829148197546601, 'learning_rate/qwen_vl_interface': 9.977259748264089e-06, 'learning_rate/action_model': 9.97725974826409e-05, 'epoch': 0.12}) +[2026-06-17 16:50:28,284][starVLA.training.train_starvla][INFO] - Step 316, Loss: {'train/loss': 1.1766657829284668, 'train/grad_norm_pre_clip': 11.133209228515625, 'timing/data': 0.020669694058597088, 'timing/model': 1.0394538762047887, 'learning_rate/qwen_vl_interface': 9.977047890011065e-06, 'learning_rate/action_model': 9.977047890011065e-05, 'epoch': 0.12}) +[2026-06-17 16:50:32,809][starVLA.training.train_starvla][INFO] - Step 317, Loss: {'train/loss': 1.2929203510284424, 'train/grad_norm_pre_clip': 12.44112491607666, 'timing/data': 0.02046610228717327, 'timing/model': 1.081039591692388, 'learning_rate/qwen_vl_interface': 9.976835051835928e-06, 'learning_rate/action_model': 9.976835051835926e-05, 'epoch': 0.12}) +[2026-06-17 16:50:37,278][starVLA.training.train_starvla][INFO] - Step 318, Loss: {'train/loss': 1.1864093542099, 'train/grad_norm_pre_clip': 10.901395797729492, 'timing/data': 0.028756314888596535, 'timing/model': 1.0631457082927227, 'learning_rate/qwen_vl_interface': 9.976621233782795e-06, 'learning_rate/action_model': 9.976621233782795e-05, 'epoch': 0.12}) +[2026-06-17 16:50:41,649][starVLA.training.train_starvla][INFO] - Step 319, Loss: {'train/loss': 0.9659988880157471, 'train/grad_norm_pre_clip': 11.668373107910156, 'timing/data': 0.020947867073118687, 'timing/model': 1.0225384179502726, 'learning_rate/qwen_vl_interface': 9.976406435895995e-06, 'learning_rate/action_model': 9.976406435895995e-05, 'epoch': 0.12}) +[2026-06-17 16:50:46,083][starVLA.training.train_starvla][INFO] - Step 320, Loss: {'train/loss': 1.1244451999664307, 'train/grad_norm_pre_clip': 13.09703254699707, 'timing/data': 0.019314422272145748, 'timing/model': 1.4314342252910137, 'learning_rate/qwen_vl_interface': 9.976190658220056e-06, 'learning_rate/action_model': 9.976190658220056e-05, 'epoch': 0.12}) +[2026-06-17 16:50:50,184][starVLA.training.train_starvla][INFO] - Step 321, Loss: {'train/loss': 1.2167456150054932, 'train/grad_norm_pre_clip': 6.399777412414551, 'timing/data': 0.019215716049075127, 'timing/model': 1.058990702033043, 'learning_rate/qwen_vl_interface': 9.975973900799709e-06, 'learning_rate/action_model': 9.975973900799708e-05, 'epoch': 0.12}) +[2026-06-17 16:50:54,668][starVLA.training.train_starvla][INFO] - Step 322, Loss: {'train/loss': 1.097534418106079, 'train/grad_norm_pre_clip': 8.58115005493164, 'timing/data': 0.019324510358273983, 'timing/model': 1.0657350067049265, 'learning_rate/qwen_vl_interface': 9.975756163679884e-06, 'learning_rate/action_model': 9.975756163679885e-05, 'epoch': 0.12}) +[2026-06-17 16:50:59,098][starVLA.training.train_starvla][INFO] - Step 323, Loss: {'train/loss': 1.0798718929290771, 'train/grad_norm_pre_clip': 21.913881301879883, 'timing/data': 0.020194356329739094, 'timing/model': 1.0868572993203998, 'learning_rate/qwen_vl_interface': 9.975537446905722e-06, 'learning_rate/action_model': 9.975537446905723e-05, 'epoch': 0.12}) +[2026-06-17 16:51:03,445][starVLA.training.train_starvla][INFO] - Step 324, Loss: {'train/loss': 1.247667908668518, 'train/grad_norm_pre_clip': 10.293198585510254, 'timing/data': 0.022408793680369854, 'timing/model': 1.0352821052074432, 'learning_rate/qwen_vl_interface': 9.975317750522565e-06, 'learning_rate/action_model': 9.975317750522564e-05, 'epoch': 0.12}) +[2026-06-17 16:51:08,644][starVLA.training.train_starvla][INFO] - Step 325, Loss: {'train/loss': 1.3523917198181152, 'train/grad_norm_pre_clip': 10.929548263549805, 'timing/data': 0.021373090334236622, 'timing/model': 1.3734142519533634, 'learning_rate/qwen_vl_interface': 9.97509707457595e-06, 'learning_rate/action_model': 9.97509707457595e-05, 'epoch': 0.12}) +[2026-06-17 16:51:14,642][starVLA.training.train_starvla][INFO] - Step 326, Loss: {'train/loss': 1.3373483419418335, 'train/grad_norm_pre_clip': 10.4685697555542, 'timing/data': 0.03003396186977625, 'timing/model': 2.362243801355362, 'learning_rate/qwen_vl_interface': 9.974875419111631e-06, 'learning_rate/action_model': 9.97487541911163e-05, 'epoch': 0.12}) +[2026-06-17 16:51:21,090][starVLA.training.train_starvla][INFO] - Step 327, Loss: {'train/loss': 0.9902666807174683, 'train/grad_norm_pre_clip': 19.571157455444336, 'timing/data': 0.028123692609369755, 'timing/model': 1.0556574333459139, 'learning_rate/qwen_vl_interface': 9.974652784175552e-06, 'learning_rate/action_model': 9.974652784175551e-05, 'epoch': 0.12}) +[2026-06-17 16:51:25,723][starVLA.training.train_starvla][INFO] - Step 328, Loss: {'train/loss': 0.8542802333831787, 'train/grad_norm_pre_clip': 9.53815746307373, 'timing/data': 0.026541970670223236, 'timing/model': 1.1691160015761852, 'learning_rate/qwen_vl_interface': 9.974429169813865e-06, 'learning_rate/action_model': 9.974429169813864e-05, 'epoch': 0.12}) +[2026-06-17 16:51:30,277][starVLA.training.train_starvla][INFO] - Step 329, Loss: {'train/loss': 1.301160454750061, 'train/grad_norm_pre_clip': 11.513134002685547, 'timing/data': 0.020059024915099144, 'timing/model': 1.0593221364542842, 'learning_rate/qwen_vl_interface': 9.974204576072928e-06, 'learning_rate/action_model': 9.974204576072928e-05, 'epoch': 0.12}) +[2026-06-17 16:51:36,374][starVLA.training.train_starvla][INFO] - Step 330, Loss: {'train/loss': 0.8131970167160034, 'train/grad_norm_pre_clip': 11.851226806640625, 'timing/data': 0.019066897220909595, 'timing/model': 1.9778140941634774, 'learning_rate/qwen_vl_interface': 9.973979002999299e-06, 'learning_rate/action_model': 9.973979002999299e-05, 'epoch': 0.12}) +[2026-06-17 16:51:40,567][starVLA.training.train_starvla][INFO] - Step 331, Loss: {'train/loss': 1.3449815511703491, 'train/grad_norm_pre_clip': 13.166958808898926, 'timing/data': 0.02130997274070978, 'timing/model': 1.0495092114433646, 'learning_rate/qwen_vl_interface': 9.97375245063974e-06, 'learning_rate/action_model': 9.97375245063974e-05, 'epoch': 0.12}) +[2026-06-17 16:51:45,138][starVLA.training.train_starvla][INFO] - Step 332, Loss: {'train/loss': 1.3973803520202637, 'train/grad_norm_pre_clip': 18.07870864868164, 'timing/data': 0.02056363131850958, 'timing/model': 1.0347320372238755, 'learning_rate/qwen_vl_interface': 9.973524919041215e-06, 'learning_rate/action_model': 9.973524919041215e-05, 'epoch': 0.12}) +[2026-06-17 16:51:49,611][starVLA.training.train_starvla][INFO] - Step 333, Loss: {'train/loss': 1.3166519403457642, 'train/grad_norm_pre_clip': 9.605111122131348, 'timing/data': 0.019345810636878014, 'timing/model': 1.0490504074841738, 'learning_rate/qwen_vl_interface': 9.97329640825089e-06, 'learning_rate/action_model': 9.973296408250891e-05, 'epoch': 0.12}) +[2026-06-17 16:51:54,194][starVLA.training.train_starvla][INFO] - Step 334, Loss: {'train/loss': 1.4116581678390503, 'train/grad_norm_pre_clip': 10.081107139587402, 'timing/data': 0.02128457836806774, 'timing/model': 1.0450736423954368, 'learning_rate/qwen_vl_interface': 9.97306691831614e-06, 'learning_rate/action_model': 9.97306691831614e-05, 'epoch': 0.12}) +[2026-06-17 16:51:58,720][starVLA.training.train_starvla][INFO] - Step 335, Loss: {'train/loss': 1.3035286664962769, 'train/grad_norm_pre_clip': 12.579167366027832, 'timing/data': 0.020779146812856197, 'timing/model': 1.5014147982001305, 'learning_rate/qwen_vl_interface': 9.972836449284533e-06, 'learning_rate/action_model': 9.972836449284532e-05, 'epoch': 0.12}) +[2026-06-17 16:52:02,755][starVLA.training.train_starvla][INFO] - Step 336, Loss: {'train/loss': 1.1746253967285156, 'train/grad_norm_pre_clip': 17.168668746948242, 'timing/data': 0.01928664557635784, 'timing/model': 1.0274145044386387, 'learning_rate/qwen_vl_interface': 9.97260500120385e-06, 'learning_rate/action_model': 9.97260500120385e-05, 'epoch': 0.12}) +[2026-06-17 16:52:07,178][starVLA.training.train_starvla][INFO] - Step 337, Loss: {'train/loss': 1.461150884628296, 'train/grad_norm_pre_clip': 16.47650909423828, 'timing/data': 0.019400586374104023, 'timing/model': 1.0634359000250697, 'learning_rate/qwen_vl_interface': 9.972372574122067e-06, 'learning_rate/action_model': 9.972372574122068e-05, 'epoch': 0.12}) +[2026-06-17 16:52:11,599][starVLA.training.train_starvla][INFO] - Step 338, Loss: {'train/loss': 1.2727837562561035, 'train/grad_norm_pre_clip': 13.530950546264648, 'timing/data': 0.024887877516448498, 'timing/model': 1.0304523892700672, 'learning_rate/qwen_vl_interface': 9.972139168087368e-06, 'learning_rate/action_model': 9.972139168087369e-05, 'epoch': 0.12}) +[2026-06-17 16:52:16,095][starVLA.training.train_starvla][INFO] - Step 339, Loss: {'train/loss': 1.422445297241211, 'train/grad_norm_pre_clip': 15.590438842773438, 'timing/data': 0.018696491606533527, 'timing/model': 1.1119914297014475, 'learning_rate/qwen_vl_interface': 9.97190478314814e-06, 'learning_rate/action_model': 9.97190478314814e-05, 'epoch': 0.12}) +[2026-06-17 16:52:20,478][starVLA.training.train_starvla][INFO] - Step 340, Loss: {'train/loss': 1.4776277542114258, 'train/grad_norm_pre_clip': 14.214816093444824, 'timing/data': 0.02011178247630596, 'timing/model': 1.034511748701334, 'learning_rate/qwen_vl_interface': 9.971669419352968e-06, 'learning_rate/action_model': 9.971669419352966e-05, 'epoch': 0.12}) +[2026-06-17 16:52:24,815][starVLA.training.train_starvla][INFO] - Step 341, Loss: {'train/loss': 1.246656894683838, 'train/grad_norm_pre_clip': 9.787111282348633, 'timing/data': 0.021969880908727646, 'timing/model': 1.3685146290808916, 'learning_rate/qwen_vl_interface': 9.971433076750644e-06, 'learning_rate/action_model': 9.971433076750643e-05, 'epoch': 0.12}) +[2026-06-17 16:52:29,031][starVLA.training.train_starvla][INFO] - Step 342, Loss: {'train/loss': 0.9720404148101807, 'train/grad_norm_pre_clip': 11.26638126373291, 'timing/data': 0.018850166350603104, 'timing/model': 1.1303896373137832, 'learning_rate/qwen_vl_interface': 9.971195755390166e-06, 'learning_rate/action_model': 9.971195755390164e-05, 'epoch': 0.13}) +[2026-06-17 16:52:34,421][starVLA.training.train_starvla][INFO] - Step 343, Loss: {'train/loss': 1.4029181003570557, 'train/grad_norm_pre_clip': 11.141135215759277, 'timing/data': 0.023520507849752903, 'timing/model': 1.0580713925883174, 'learning_rate/qwen_vl_interface': 9.970957455320723e-06, 'learning_rate/action_model': 9.970957455320723e-05, 'epoch': 0.13}) +[2026-06-17 16:52:39,015][starVLA.training.train_starvla][INFO] - Step 344, Loss: {'train/loss': 1.0888433456420898, 'train/grad_norm_pre_clip': 15.727767944335938, 'timing/data': 0.01919236220419407, 'timing/model': 1.13691748958081, 'learning_rate/qwen_vl_interface': 9.970718176591723e-06, 'learning_rate/action_model': 9.970718176591722e-05, 'epoch': 0.13}) +[2026-06-17 16:52:43,491][starVLA.training.train_starvla][INFO] - Step 345, Loss: {'train/loss': 0.9980985522270203, 'train/grad_norm_pre_clip': 7.46705961227417, 'timing/data': 0.019599835388362408, 'timing/model': 1.0430702166631818, 'learning_rate/qwen_vl_interface': 9.970477919252763e-06, 'learning_rate/action_model': 9.970477919252763e-05, 'epoch': 0.13}) +[2026-06-17 16:52:48,159][starVLA.training.train_starvla][INFO] - Step 346, Loss: {'train/loss': 1.2224164009094238, 'train/grad_norm_pre_clip': 11.099157333374023, 'timing/data': 0.019795788452029228, 'timing/model': 1.5193040147423744, 'learning_rate/qwen_vl_interface': 9.970236683353652e-06, 'learning_rate/action_model': 9.970236683353652e-05, 'epoch': 0.13}) +[2026-06-17 16:52:52,463][starVLA.training.train_starvla][INFO] - Step 347, Loss: {'train/loss': 1.1704126596450806, 'train/grad_norm_pre_clip': 9.14962100982666, 'timing/data': 0.020643816329538822, 'timing/model': 1.1194550339132547, 'learning_rate/qwen_vl_interface': 9.969994468944399e-06, 'learning_rate/action_model': 9.969994468944398e-05, 'epoch': 0.13}) +[2026-06-17 16:52:57,009][starVLA.training.train_starvla][INFO] - Step 348, Loss: {'train/loss': 1.2704830169677734, 'train/grad_norm_pre_clip': 8.205894470214844, 'timing/data': 0.019857769832015038, 'timing/model': 1.197070823982358, 'learning_rate/qwen_vl_interface': 9.969751276075212e-06, 'learning_rate/action_model': 9.969751276075212e-05, 'epoch': 0.13}) +[2026-06-17 16:53:01,598][starVLA.training.train_starvla][INFO] - Step 349, Loss: {'train/loss': 1.0346505641937256, 'train/grad_norm_pre_clip': 10.440866470336914, 'timing/data': 0.019955401308834553, 'timing/model': 1.0264638932421803, 'learning_rate/qwen_vl_interface': 9.969507104796506e-06, 'learning_rate/action_model': 9.969507104796505e-05, 'epoch': 0.13}) +[2026-06-17 16:53:06,020][starVLA.training.train_starvla][INFO] - Step 350, Loss: {'train/loss': 1.2140320539474487, 'train/grad_norm_pre_clip': 9.804434776306152, 'timing/data': 0.019002407789230347, 'timing/model': 1.0353473145514727, 'learning_rate/qwen_vl_interface': 9.969261955158898e-06, 'learning_rate/action_model': 9.969261955158898e-05, 'epoch': 0.13}) +[2026-06-17 16:53:10,485][starVLA.training.train_starvla][INFO] - Step 351, Loss: {'train/loss': 0.9996349811553955, 'train/grad_norm_pre_clip': 11.027738571166992, 'timing/data': 0.020311822183430195, 'timing/model': 1.436624406836927, 'learning_rate/qwen_vl_interface': 9.96901582721321e-06, 'learning_rate/action_model': 9.96901582721321e-05, 'epoch': 0.13}) +[2026-06-17 16:53:14,645][starVLA.training.train_starvla][INFO] - Step 352, Loss: {'train/loss': 1.2478262186050415, 'train/grad_norm_pre_clip': 12.639451026916504, 'timing/data': 0.023741649463772774, 'timing/model': 1.0772577077150345, 'learning_rate/qwen_vl_interface': 9.968768721010463e-06, 'learning_rate/action_model': 9.968768721010463e-05, 'epoch': 0.13}) +[2026-06-17 16:53:19,085][starVLA.training.train_starvla][INFO] - Step 353, Loss: {'train/loss': 1.0725561380386353, 'train/grad_norm_pre_clip': 11.420647621154785, 'timing/data': 0.02011137269437313, 'timing/model': 1.0378134483471513, 'learning_rate/qwen_vl_interface': 9.968520636601883e-06, 'learning_rate/action_model': 9.968520636601883e-05, 'epoch': 0.13}) +[2026-06-17 16:53:23,605][starVLA.training.train_starvla][INFO] - Step 354, Loss: {'train/loss': 1.181917428970337, 'train/grad_norm_pre_clip': 9.30182933807373, 'timing/data': 0.018824311904609203, 'timing/model': 1.1207366343587637, 'learning_rate/qwen_vl_interface': 9.968271574038899e-06, 'learning_rate/action_model': 9.968271574038898e-05, 'epoch': 0.13}) +[2026-06-17 16:53:28,057][starVLA.training.train_starvla][INFO] - Step 355, Loss: {'train/loss': 1.468764066696167, 'train/grad_norm_pre_clip': 13.206159591674805, 'timing/data': 0.024345530197024345, 'timing/model': 1.1059451326727867, 'learning_rate/qwen_vl_interface': 9.968021533373138e-06, 'learning_rate/action_model': 9.968021533373138e-05, 'epoch': 0.13}) +[2026-06-17 16:53:32,447][starVLA.training.train_starvla][INFO] - Step 356, Loss: {'train/loss': 1.079321026802063, 'train/grad_norm_pre_clip': 10.171799659729004, 'timing/data': 0.0198317626491189, 'timing/model': 1.34523431584239, 'learning_rate/qwen_vl_interface': 9.967770514656439e-06, 'learning_rate/action_model': 9.967770514656438e-05, 'epoch': 0.13}) +[2026-06-17 16:53:36,445][starVLA.training.train_starvla][INFO] - Step 357, Loss: {'train/loss': 1.286310076713562, 'train/grad_norm_pre_clip': 11.494880676269531, 'timing/data': 0.02013146784156561, 'timing/model': 1.0222084932029247, 'learning_rate/qwen_vl_interface': 9.967518517940832e-06, 'learning_rate/action_model': 9.967518517940832e-05, 'epoch': 0.13}) +[2026-06-17 16:53:42,145][starVLA.training.train_starvla][INFO] - Step 358, Loss: {'train/loss': 1.1381080150604248, 'train/grad_norm_pre_clip': 80.40670776367188, 'timing/data': 0.023280218243598938, 'timing/model': 2.3771170787513256, 'learning_rate/qwen_vl_interface': 9.967265543278563e-06, 'learning_rate/action_model': 9.967265543278562e-05, 'epoch': 0.13}) +[2026-06-17 16:53:48,188][starVLA.training.train_starvla][INFO] - Step 359, Loss: {'train/loss': 1.521470308303833, 'train/grad_norm_pre_clip': 12.802715301513672, 'timing/data': 0.019137060269713402, 'timing/model': 1.075429417192936, 'learning_rate/qwen_vl_interface': 9.96701159072207e-06, 'learning_rate/action_model': 9.96701159072207e-05, 'epoch': 0.13}) +[2026-06-17 16:53:52,794][starVLA.training.train_starvla][INFO] - Step 360, Loss: {'train/loss': 1.296952724456787, 'train/grad_norm_pre_clip': 10.258952140808105, 'timing/data': 0.02308248821645975, 'timing/model': 1.3973401738330722, 'learning_rate/qwen_vl_interface': 9.966756660323998e-06, 'learning_rate/action_model': 9.966756660323999e-05, 'epoch': 0.13}) +[2026-06-17 16:53:57,150][starVLA.training.train_starvla][INFO] - Step 361, Loss: {'train/loss': 1.502084493637085, 'train/grad_norm_pre_clip': 9.3262300491333, 'timing/data': 0.022089875303208828, 'timing/model': 1.327868984080851, 'learning_rate/qwen_vl_interface': 9.966500752137195e-06, 'learning_rate/action_model': 9.966500752137196e-05, 'epoch': 0.13}) +[2026-06-17 16:54:01,388][starVLA.training.train_starvla][INFO] - Step 362, Loss: {'train/loss': 1.388278841972351, 'train/grad_norm_pre_clip': 9.591658592224121, 'timing/data': 0.021827010437846184, 'timing/model': 1.1695112520828843, 'learning_rate/qwen_vl_interface': 9.966243866214712e-06, 'learning_rate/action_model': 9.966243866214712e-05, 'epoch': 0.13}) +[2026-06-17 16:54:05,853][starVLA.training.train_starvla][INFO] - Step 363, Loss: {'train/loss': 1.1936190128326416, 'train/grad_norm_pre_clip': 12.671957015991211, 'timing/data': 0.020038887858390808, 'timing/model': 1.066382727585733, 'learning_rate/qwen_vl_interface': 9.965986002609799e-06, 'learning_rate/action_model': 9.965986002609799e-05, 'epoch': 0.13}) +[2026-06-17 16:54:10,290][starVLA.training.train_starvla][INFO] - Step 364, Loss: {'train/loss': 0.950579047203064, 'train/grad_norm_pre_clip': 8.873348236083984, 'timing/data': 0.019825209863483906, 'timing/model': 1.0351416440680623, 'learning_rate/qwen_vl_interface': 9.965727161375915e-06, 'learning_rate/action_model': 9.965727161375914e-05, 'epoch': 0.13}) +[2026-06-17 16:54:14,682][starVLA.training.train_starvla][INFO] - Step 365, Loss: {'train/loss': 1.0576863288879395, 'train/grad_norm_pre_clip': 9.250356674194336, 'timing/data': 0.02651053946465254, 'timing/model': 1.0371183082461357, 'learning_rate/qwen_vl_interface': 9.965467342566715e-06, 'learning_rate/action_model': 9.965467342566715e-05, 'epoch': 0.13}) +[2026-06-17 16:54:20,606][starVLA.training.train_starvla][INFO] - Step 366, Loss: {'train/loss': 1.195521354675293, 'train/grad_norm_pre_clip': 8.480365753173828, 'timing/data': 0.021128359250724316, 'timing/model': 1.4622718952596188, 'learning_rate/qwen_vl_interface': 9.965206546236061e-06, 'learning_rate/action_model': 9.96520654623606e-05, 'epoch': 0.13}) +[2026-06-17 16:54:24,762][starVLA.training.train_starvla][INFO] - Step 367, Loss: {'train/loss': 1.2254915237426758, 'train/grad_norm_pre_clip': 11.148277282714844, 'timing/data': 0.02054920606315136, 'timing/model': 1.0319249276071787, 'learning_rate/qwen_vl_interface': 9.964944772438014e-06, 'learning_rate/action_model': 9.964944772438013e-05, 'epoch': 0.13}) +[2026-06-17 16:54:30,541][starVLA.training.train_starvla][INFO] - Step 368, Loss: {'train/loss': 0.9549626111984253, 'train/grad_norm_pre_clip': 8.430066108703613, 'timing/data': 0.021270846016705036, 'timing/model': 1.0502682775259018, 'learning_rate/qwen_vl_interface': 9.964682021226844e-06, 'learning_rate/action_model': 9.964682021226844e-05, 'epoch': 0.13}) +[2026-06-17 16:54:34,991][starVLA.training.train_starvla][INFO] - Step 369, Loss: {'train/loss': 1.2455872297286987, 'train/grad_norm_pre_clip': 11.028894424438477, 'timing/data': 0.022233259864151478, 'timing/model': 1.072351879440248, 'learning_rate/qwen_vl_interface': 9.964418292657019e-06, 'learning_rate/action_model': 9.964418292657018e-05, 'epoch': 0.14}) +[2026-06-17 16:54:39,639][starVLA.training.train_starvla][INFO] - Step 370, Loss: {'train/loss': 1.2513805627822876, 'train/grad_norm_pre_clip': 9.271860122680664, 'timing/data': 0.022102574817836285, 'timing/model': 1.5185782620683312, 'learning_rate/qwen_vl_interface': 9.96415358678321e-06, 'learning_rate/action_model': 9.964153586783209e-05, 'epoch': 0.14}) +[2026-06-17 16:54:44,094][starVLA.training.train_starvla][INFO] - Step 371, Loss: {'train/loss': 0.9855420589447021, 'train/grad_norm_pre_clip': 12.325096130371094, 'timing/data': 0.018675897270441055, 'timing/model': 1.3651481503620744, 'learning_rate/qwen_vl_interface': 9.963887903660285e-06, 'learning_rate/action_model': 9.963887903660285e-05, 'epoch': 0.14}) +[2026-06-17 16:54:48,218][starVLA.training.train_starvla][INFO] - Step 372, Loss: {'train/loss': 1.0100659132003784, 'train/grad_norm_pre_clip': 11.730514526367188, 'timing/data': 0.022071093320846558, 'timing/model': 1.0989983854815364, 'learning_rate/qwen_vl_interface': 9.96362124334333e-06, 'learning_rate/action_model': 9.963621243343329e-05, 'epoch': 0.14}) +[2026-06-17 16:54:53,952][starVLA.training.train_starvla][INFO] - Step 373, Loss: {'train/loss': 1.1330841779708862, 'train/grad_norm_pre_clip': 9.140787124633789, 'timing/data': 0.01980624906718731, 'timing/model': 1.0286848209798336, 'learning_rate/qwen_vl_interface': 9.963353605887617e-06, 'learning_rate/action_model': 9.963353605887616e-05, 'epoch': 0.14}) +[2026-06-17 16:54:58,462][starVLA.training.train_starvla][INFO] - Step 374, Loss: {'train/loss': 1.1115648746490479, 'train/grad_norm_pre_clip': 12.381871223449707, 'timing/data': 0.020609448663890362, 'timing/model': 1.0339795984327793, 'learning_rate/qwen_vl_interface': 9.963084991348628e-06, 'learning_rate/action_model': 9.963084991348628e-05, 'epoch': 0.14}) +[2026-06-17 16:55:02,928][starVLA.training.train_starvla][INFO] - Step 375, Loss: {'train/loss': 1.2526228427886963, 'train/grad_norm_pre_clip': 8.86418342590332, 'timing/data': 0.020236541517078876, 'timing/model': 1.027710571885109, 'learning_rate/qwen_vl_interface': 9.962815399782051e-06, 'learning_rate/action_model': 9.96281539978205e-05, 'epoch': 0.14}) +[2026-06-17 16:55:07,515][starVLA.training.train_starvla][INFO] - Step 376, Loss: {'train/loss': 1.2918691635131836, 'train/grad_norm_pre_clip': 11.317044258117676, 'timing/data': 0.01917179673910141, 'timing/model': 1.4150234824046493, 'learning_rate/qwen_vl_interface': 9.96254483124377e-06, 'learning_rate/action_model': 9.96254483124377e-05, 'epoch': 0.14}) +[2026-06-17 16:55:11,667][starVLA.training.train_starvla][INFO] - Step 377, Loss: {'train/loss': 1.241560459136963, 'train/grad_norm_pre_clip': 8.912923812866211, 'timing/data': 0.018831945024430752, 'timing/model': 1.107383576221764, 'learning_rate/qwen_vl_interface': 9.962273285789876e-06, 'learning_rate/action_model': 9.962273285789875e-05, 'epoch': 0.14}) +[2026-06-17 16:55:16,215][starVLA.training.train_starvla][INFO] - Step 378, Loss: {'train/loss': 0.8350830078125, 'train/grad_norm_pre_clip': 11.957317352294922, 'timing/data': 0.01852413173764944, 'timing/model': 1.0266721099615097, 'learning_rate/qwen_vl_interface': 9.962000763476658e-06, 'learning_rate/action_model': 9.962000763476658e-05, 'epoch': 0.14}) +[2026-06-17 16:55:20,671][starVLA.training.train_starvla][INFO] - Step 379, Loss: {'train/loss': 0.9566681981086731, 'train/grad_norm_pre_clip': 7.926413536071777, 'timing/data': 0.03211522847414017, 'timing/model': 1.0866608219221234, 'learning_rate/qwen_vl_interface': 9.96172726436061e-06, 'learning_rate/action_model': 9.96172726436061e-05, 'epoch': 0.14}) +[2026-06-17 16:55:25,566][starVLA.training.train_starvla][INFO] - Step 380, Loss: {'train/loss': 1.1123687028884888, 'train/grad_norm_pre_clip': 11.260783195495605, 'timing/data': 0.019669278524816036, 'timing/model': 1.855736967176199, 'learning_rate/qwen_vl_interface': 9.961452788498432e-06, 'learning_rate/action_model': 9.961452788498433e-05, 'epoch': 0.14}) +[2026-06-17 16:55:29,887][starVLA.training.train_starvla][INFO] - Step 381, Loss: {'train/loss': 1.199035406112671, 'train/grad_norm_pre_clip': 10.853241920471191, 'timing/data': 0.020170223899185658, 'timing/model': 1.3201363608241081, 'learning_rate/qwen_vl_interface': 9.96117733594702e-06, 'learning_rate/action_model': 9.96117733594702e-05, 'epoch': 0.14}) +[2026-06-17 16:55:33,978][starVLA.training.train_starvla][INFO] - Step 382, Loss: {'train/loss': 1.3294627666473389, 'train/grad_norm_pre_clip': 16.902114868164062, 'timing/data': 0.02021070383489132, 'timing/model': 1.0409650579094887, 'learning_rate/qwen_vl_interface': 9.960900906763478e-06, 'learning_rate/action_model': 9.960900906763477e-05, 'epoch': 0.14}) +[2026-06-17 16:55:38,446][starVLA.training.train_starvla][INFO] - Step 383, Loss: {'train/loss': 1.4094274044036865, 'train/grad_norm_pre_clip': 11.00835132598877, 'timing/data': 0.024137591011822224, 'timing/model': 1.032045834697783, 'learning_rate/qwen_vl_interface': 9.960623501005108e-06, 'learning_rate/action_model': 9.960623501005107e-05, 'epoch': 0.14}) +[2026-06-17 16:55:42,762][starVLA.training.train_starvla][INFO] - Step 384, Loss: {'train/loss': 1.0509461164474487, 'train/grad_norm_pre_clip': 12.589012145996094, 'timing/data': 0.019609532319009304, 'timing/model': 1.0198692558333278, 'learning_rate/qwen_vl_interface': 9.960345118729416e-06, 'learning_rate/action_model': 9.960345118729417e-05, 'epoch': 0.14}) +[2026-06-17 16:55:46,979][starVLA.training.train_starvla][INFO] - Step 385, Loss: {'train/loss': 1.0579442977905273, 'train/grad_norm_pre_clip': 15.024463653564453, 'timing/data': 0.018087205477058887, 'timing/model': 1.0078969616442919, 'learning_rate/qwen_vl_interface': 9.960065759994116e-06, 'learning_rate/action_model': 9.960065759994114e-05, 'epoch': 0.14}) +[2026-06-17 16:55:51,131][starVLA.training.train_starvla][INFO] - Step 386, Loss: {'train/loss': 1.2160335779190063, 'train/grad_norm_pre_clip': 9.878379821777344, 'timing/data': 0.0164212454110384, 'timing/model': 1.2547345850616693, 'learning_rate/qwen_vl_interface': 9.959785424857113e-06, 'learning_rate/action_model': 9.959785424857112e-05, 'epoch': 0.14}) +[2026-06-17 16:55:55,029][starVLA.training.train_starvla][INFO] - Step 387, Loss: {'train/loss': 0.954893171787262, 'train/grad_norm_pre_clip': 9.548003196716309, 'timing/data': 0.016315823420882225, 'timing/model': 1.0067507857456803, 'learning_rate/qwen_vl_interface': 9.959504113376523e-06, 'learning_rate/action_model': 9.959504113376523e-05, 'epoch': 0.14}) +[2026-06-17 16:55:59,203][starVLA.training.train_starvla][INFO] - Step 388, Loss: {'train/loss': 1.0360872745513916, 'train/grad_norm_pre_clip': 11.363664627075195, 'timing/data': 0.017377999611198902, 'timing/model': 1.0120917884632945, 'learning_rate/qwen_vl_interface': 9.959221825610664e-06, 'learning_rate/action_model': 9.959221825610663e-05, 'epoch': 0.14}) +[2026-06-17 16:56:03,457][starVLA.training.train_starvla][INFO] - Step 389, Loss: {'train/loss': 0.933947741985321, 'train/grad_norm_pre_clip': 8.588292121887207, 'timing/data': 0.018238257616758347, 'timing/model': 1.008222601376474, 'learning_rate/qwen_vl_interface': 9.958938561618053e-06, 'learning_rate/action_model': 9.958938561618053e-05, 'epoch': 0.14}) +[2026-06-17 16:56:07,616][starVLA.training.train_starvla][INFO] - Step 390, Loss: {'train/loss': 0.9541072249412537, 'train/grad_norm_pre_clip': 8.486592292785645, 'timing/data': 0.017775043845176697, 'timing/model': 1.265262889675796, 'learning_rate/qwen_vl_interface': 9.958654321457412e-06, 'learning_rate/action_model': 9.958654321457412e-05, 'epoch': 0.14}) +[2026-06-17 16:56:11,785][starVLA.training.train_starvla][INFO] - Step 391, Loss: {'train/loss': 1.1025093793869019, 'train/grad_norm_pre_clip': 10.03567886352539, 'timing/data': 0.017812861129641533, 'timing/model': 1.2625835370272398, 'learning_rate/qwen_vl_interface': 9.958369105187661e-06, 'learning_rate/action_model': 9.958369105187661e-05, 'epoch': 0.14}) +[2026-06-17 16:56:15,706][starVLA.training.train_starvla][INFO] - Step 392, Loss: {'train/loss': 1.0179524421691895, 'train/grad_norm_pre_clip': 8.869461059570312, 'timing/data': 0.017004188150167465, 'timing/model': 1.0091484123840928, 'learning_rate/qwen_vl_interface': 9.958082912867932e-06, 'learning_rate/action_model': 9.958082912867931e-05, 'epoch': 0.14}) +[2026-06-17 16:56:19,836][starVLA.training.train_starvla][INFO] - Step 393, Loss: {'train/loss': 1.392362117767334, 'train/grad_norm_pre_clip': 12.781990051269531, 'timing/data': 0.01676066406071186, 'timing/model': 1.0086066257208586, 'learning_rate/qwen_vl_interface': 9.957795744557545e-06, 'learning_rate/action_model': 9.957795744557546e-05, 'epoch': 0.14}) +[2026-06-17 16:56:23,996][starVLA.training.train_starvla][INFO] - Step 394, Loss: {'train/loss': 1.0475527048110962, 'train/grad_norm_pre_clip': 9.368742942810059, 'timing/data': 0.01668579876422882, 'timing/model': 1.0082555133849382, 'learning_rate/qwen_vl_interface': 9.957507600316036e-06, 'learning_rate/action_model': 9.957507600316036e-05, 'epoch': 0.14}) +[2026-06-17 16:56:28,169][starVLA.training.train_starvla][INFO] - Step 395, Loss: {'train/loss': 1.4756311178207397, 'train/grad_norm_pre_clip': 15.41281795501709, 'timing/data': 0.01805441826581955, 'timing/model': 1.0283241970464587, 'learning_rate/qwen_vl_interface': 9.957218480203137e-06, 'learning_rate/action_model': 9.957218480203136e-05, 'epoch': 0.14}) +[2026-06-17 16:56:32,793][starVLA.training.train_starvla][INFO] - Step 396, Loss: {'train/loss': 1.2331149578094482, 'train/grad_norm_pre_clip': 12.481048583984375, 'timing/data': 0.017967160791158676, 'timing/model': 1.3954772092401981, 'learning_rate/qwen_vl_interface': 9.95692838427878e-06, 'learning_rate/action_model': 9.956928384278778e-05, 'epoch': 0.15}) +[2026-06-17 16:56:36,806][starVLA.training.train_starvla][INFO] - Step 397, Loss: {'train/loss': 0.769783079624176, 'train/grad_norm_pre_clip': 9.27366828918457, 'timing/data': 0.01760256290435791, 'timing/model': 1.0291530527174473, 'learning_rate/qwen_vl_interface': 9.956637312603105e-06, 'learning_rate/action_model': 9.956637312603104e-05, 'epoch': 0.15}) +[2026-06-17 16:56:41,288][starVLA.training.train_starvla][INFO] - Step 398, Loss: {'train/loss': 0.8751251697540283, 'train/grad_norm_pre_clip': 12.648375511169434, 'timing/data': 0.02020847611129284, 'timing/model': 1.053628875873983, 'learning_rate/qwen_vl_interface': 9.95634526523645e-06, 'learning_rate/action_model': 9.956345265236449e-05, 'epoch': 0.15}) +[2026-06-17 16:56:45,869][starVLA.training.train_starvla][INFO] - Step 399, Loss: {'train/loss': 1.4282050132751465, 'train/grad_norm_pre_clip': 10.757465362548828, 'timing/data': 0.020362976007163525, 'timing/model': 1.0307523775845766, 'learning_rate/qwen_vl_interface': 9.956052242239358e-06, 'learning_rate/action_model': 9.956052242239358e-05, 'epoch': 0.15}) +[2026-06-17 16:59:20,535][starVLA.training.train_starvla][INFO] - Step 400, Loss: {'train/loss': 1.0584771633148193, 'train/grad_norm_pre_clip': 12.799132347106934, 'eval/loss': 1.1826013326644897, 'eval/latency_2/loss': 1.1826013326644897, 'eval/demon_fire_precision': 0.8772009029345372, 'eval/demon_fire_recall': 0.8129707112970711, 'eval/demon_fire_f1': 0.8438653637350706, 'eval/demon_fire_rate_model': 0.6163049526989427, 'eval/demon_fire_rate_teacher': 0.6649972175848636, 'eval/demon_left_precision': 0.6210801393728222, 'eval/demon_left_recall': 0.7505263157894737, 'eval/demon_left_f1': 0.6796949475691134, 'eval/demon_left_rate_model': 0.31942125765164164, 'eval/demon_left_rate_teacher': 0.2643294379521425, 'eval/demon_right_precision': 0.5767477203647416, 'eval/demon_right_recall': 0.8576271186440678, 'eval/demon_right_f1': 0.6896865061335756, 'eval/demon_right_rate_model': 0.36616583194212576, 'eval/demon_right_rate_teacher': 0.24624373956594323, 'eval/demon_move_f1': 0.6846907268513445, 'eval/demon_cc_f1': 0.7888486834363817, 'eval/latency_2/demon_fire_precision': 0.8772009029345372, 'eval/latency_2/demon_fire_recall': 0.8129707112970711, 'eval/latency_2/demon_fire_f1': 0.8438653637350706, 'eval/latency_2/demon_fire_rate_model': 0.6163049526989427, 'eval/latency_2/demon_fire_rate_teacher': 0.6649972175848636, 'eval/latency_2/demon_left_precision': 0.6210801393728222, 'eval/latency_2/demon_left_recall': 0.7505263157894737, 'eval/latency_2/demon_left_f1': 0.6796949475691134, 'eval/latency_2/demon_left_rate_model': 0.31942125765164164, 'eval/latency_2/demon_left_rate_teacher': 0.2643294379521425, 'eval/latency_2/demon_right_precision': 0.5767477203647416, 'eval/latency_2/demon_right_recall': 0.8576271186440678, 'eval/latency_2/demon_right_f1': 0.6896865061335756, 'eval/latency_2/demon_right_rate_model': 0.36616583194212576, 'eval/latency_2/demon_right_rate_teacher': 0.24624373956594323, 'eval/latency_2/demon_move_f1': 0.6846907268513445, 'eval/latency_2/demon_cc_f1': 0.7888486834363817, 'timing/data': 0.01979204174131155, 'timing/model': 1.3727364698424935, 'learning_rate/qwen_vl_interface': 9.955758243672572e-06, 'learning_rate/action_model': 9.955758243672572e-05, 'epoch': 0.15}) +[2026-06-17 16:59:20,537][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_400_state +[2026-06-17 16:59:20,545][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 16:59:29,410][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_400_state/model.safetensors +[2026-06-17 16:59:43,530][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_400_state/optimizer.bin +[2026-06-17 16:59:43,531][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_400_state/sampler.bin +[2026-06-17 16:59:43,532][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_400_state/sampler_1.bin +[2026-06-17 16:59:43,534][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_400_state/random_states_0.pkl +[2026-06-17 16:59:43,539][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 400 +[2026-06-17 16:59:43,539][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 16:59:43,556][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 16:59:49,718][starVLA.training.train_starvla][INFO] - Step 401, Loss: {'train/loss': 1.2032333612442017, 'train/grad_norm_pre_clip': 10.48668098449707, 'timing/data': 0.023656467907130718, 'timing/model': 1.0371733931824565, 'learning_rate/qwen_vl_interface': 9.955463269597038e-06, 'learning_rate/action_model': 9.955463269597038e-05, 'epoch': 0.15}) +[2026-06-17 16:59:54,192][starVLA.training.train_starvla][INFO] - Step 402, Loss: {'train/loss': 1.4663504362106323, 'train/grad_norm_pre_clip': 10.85206413269043, 'timing/data': 0.020709062926471233, 'timing/model': 1.3543239142745733, 'learning_rate/qwen_vl_interface': 9.955167320073905e-06, 'learning_rate/action_model': 9.955167320073905e-05, 'epoch': 0.15}) +[2026-06-17 17:00:00,917][starVLA.training.train_starvla][INFO] - Step 403, Loss: {'train/loss': 1.2076771259307861, 'train/grad_norm_pre_clip': 11.24959945678711, 'timing/data': 0.025333251804113388, 'timing/model': 1.0492622442543507, 'learning_rate/qwen_vl_interface': 9.954870395164524e-06, 'learning_rate/action_model': 9.954870395164524e-05, 'epoch': 0.15}) +[2026-06-17 17:00:05,193][starVLA.training.train_starvla][INFO] - Step 404, Loss: {'train/loss': 1.2303208112716675, 'train/grad_norm_pre_clip': 15.277338981628418, 'timing/data': 0.01980547234416008, 'timing/model': 1.0382037349045277, 'learning_rate/qwen_vl_interface': 9.954572494930448e-06, 'learning_rate/action_model': 9.954572494930448e-05, 'epoch': 0.15}) +[2026-06-17 17:00:15,114][starVLA.training.train_starvla][INFO] - Step 405, Loss: {'train/loss': 1.1752443313598633, 'train/grad_norm_pre_clip': 7.704478740692139, 'timing/data': 0.02721276879310608, 'timing/model': 3.8098208140581846, 'learning_rate/qwen_vl_interface': 9.95427361943343e-06, 'learning_rate/action_model': 9.954273619433431e-05, 'epoch': 0.15}) +[2026-06-17 17:00:19,799][starVLA.training.train_starvla][INFO] - Step 406, Loss: {'train/loss': 1.0228065252304077, 'train/grad_norm_pre_clip': 13.551774978637695, 'timing/data': 0.02324875071644783, 'timing/model': 1.0457081943750381, 'learning_rate/qwen_vl_interface': 9.95397376873543e-06, 'learning_rate/action_model': 9.95397376873543e-05, 'epoch': 0.15}) +[2026-06-17 17:00:26,146][starVLA.training.train_starvla][INFO] - Step 407, Loss: {'train/loss': 1.3596593141555786, 'train/grad_norm_pre_clip': 12.186027526855469, 'timing/data': 0.01890317816287279, 'timing/model': 1.6456702472642064, 'learning_rate/qwen_vl_interface': 9.953672942898606e-06, 'learning_rate/action_model': 9.953672942898605e-05, 'epoch': 0.15}) +[2026-06-17 17:00:32,556][starVLA.training.train_starvla][INFO] - Step 408, Loss: {'train/loss': 1.110634684562683, 'train/grad_norm_pre_clip': 14.411352157592773, 'timing/data': 0.026661859825253487, 'timing/model': 1.1160321943461895, 'learning_rate/qwen_vl_interface': 9.953371141985318e-06, 'learning_rate/action_model': 9.953371141985319e-05, 'epoch': 0.15}) +[2026-06-17 17:00:42,512][starVLA.training.train_starvla][INFO] - Step 409, Loss: {'train/loss': 0.8386374711990356, 'train/grad_norm_pre_clip': 10.587057113647461, 'timing/data': 0.02880196925252676, 'timing/model': 4.365180867724121, 'learning_rate/qwen_vl_interface': 9.953068366058133e-06, 'learning_rate/action_model': 9.953068366058133e-05, 'epoch': 0.15}) +[2026-06-17 17:00:54,883][starVLA.training.train_starvla][INFO] - Step 410, Loss: {'train/loss': 1.3217957019805908, 'train/grad_norm_pre_clip': 21.125728607177734, 'timing/data': 0.02919970266520977, 'timing/model': 5.540575383231044, 'learning_rate/qwen_vl_interface': 9.952764615179813e-06, 'learning_rate/action_model': 9.952764615179812e-05, 'epoch': 0.15}) +[2026-06-17 17:01:05,793][starVLA.training.train_starvla][INFO] - Step 411, Loss: {'train/loss': 1.004347801208496, 'train/grad_norm_pre_clip': 9.719449043273926, 'timing/data': 0.03434819635003805, 'timing/model': 1.5296414820477366, 'learning_rate/qwen_vl_interface': 9.95245988941333e-06, 'learning_rate/action_model': 9.95245988941333e-05, 'epoch': 0.15}) +[2026-06-17 17:01:15,616][starVLA.training.train_starvla][INFO] - Step 412, Loss: {'train/loss': 1.2041300535202026, 'train/grad_norm_pre_clip': 11.264825820922852, 'timing/data': 0.022974270395934582, 'timing/model': 6.285530934110284, 'learning_rate/qwen_vl_interface': 9.952154188821852e-06, 'learning_rate/action_model': 9.952154188821851e-05, 'epoch': 0.15}) +[2026-06-17 17:01:30,229][starVLA.training.train_starvla][INFO] - Step 413, Loss: {'train/loss': 1.1936508417129517, 'train/grad_norm_pre_clip': 8.919822692871094, 'timing/data': 0.03570243902504444, 'timing/model': 1.1948600728064775, 'learning_rate/qwen_vl_interface': 9.95184751346875e-06, 'learning_rate/action_model': 9.95184751346875e-05, 'epoch': 0.15}) +[2026-06-17 17:01:37,802][starVLA.training.train_starvla][INFO] - Step 414, Loss: {'train/loss': 1.1836771965026855, 'train/grad_norm_pre_clip': 10.896770477294922, 'timing/data': 0.019739771261811256, 'timing/model': 1.3219859460368752, 'learning_rate/qwen_vl_interface': 9.951539863417598e-06, 'learning_rate/action_model': 9.951539863417599e-05, 'epoch': 0.15}) +[2026-06-17 17:01:48,642][starVLA.training.train_starvla][INFO] - Step 415, Loss: {'train/loss': 1.020533800125122, 'train/grad_norm_pre_clip': 10.23153305053711, 'timing/data': 0.028949559666216373, 'timing/model': 1.3046788172796369, 'learning_rate/qwen_vl_interface': 9.951231238732174e-06, 'learning_rate/action_model': 9.951231238732173e-05, 'epoch': 0.15}) +[2026-06-17 17:01:53,227][starVLA.training.train_starvla][INFO] - Step 416, Loss: {'train/loss': 1.0911086797714233, 'train/grad_norm_pre_clip': 9.748270988464355, 'timing/data': 0.022335181944072247, 'timing/model': 1.1335097141563892, 'learning_rate/qwen_vl_interface': 9.950921639476455e-06, 'learning_rate/action_model': 9.950921639476455e-05, 'epoch': 0.15}) +[2026-06-17 17:01:57,669][starVLA.training.train_starvla][INFO] - Step 417, Loss: {'train/loss': 1.1692754030227661, 'train/grad_norm_pre_clip': 11.372596740722656, 'timing/data': 0.017575339414179325, 'timing/model': 1.4258283879607916, 'learning_rate/qwen_vl_interface': 9.950611065714623e-06, 'learning_rate/action_model': 9.950611065714623e-05, 'epoch': 0.15}) +[2026-06-17 17:02:01,789][starVLA.training.train_starvla][INFO] - Step 418, Loss: {'train/loss': 0.9888869524002075, 'train/grad_norm_pre_clip': 7.550706386566162, 'timing/data': 0.02097652293741703, 'timing/model': 1.0683034574612975, 'learning_rate/qwen_vl_interface': 9.950299517511058e-06, 'learning_rate/action_model': 9.950299517511058e-05, 'epoch': 0.15}) +[2026-06-17 17:02:06,299][starVLA.training.train_starvla][INFO] - Step 419, Loss: {'train/loss': 0.9591285586357117, 'train/grad_norm_pre_clip': 9.049429893493652, 'timing/data': 0.029661372303962708, 'timing/model': 1.058948122896254, 'learning_rate/qwen_vl_interface': 9.949986994930346e-06, 'learning_rate/action_model': 9.949986994930346e-05, 'epoch': 0.15}) +[2026-06-17 17:02:10,722][starVLA.training.train_starvla][INFO] - Step 420, Loss: {'train/loss': 1.7003902196884155, 'train/grad_norm_pre_clip': 9.400917053222656, 'timing/data': 0.020729729905724525, 'timing/model': 1.0393957942724228, 'learning_rate/qwen_vl_interface': 9.949673498037272e-06, 'learning_rate/action_model': 9.949673498037272e-05, 'epoch': 0.15}) +[2026-06-17 17:02:15,339][starVLA.training.train_starvla][INFO] - Step 421, Loss: {'train/loss': 0.9852898716926575, 'train/grad_norm_pre_clip': 9.050091743469238, 'timing/data': 0.02001513261348009, 'timing/model': 1.5202263919636607, 'learning_rate/qwen_vl_interface': 9.949359026896826e-06, 'learning_rate/action_model': 9.949359026896826e-05, 'epoch': 0.15}) +[2026-06-17 17:02:19,712][starVLA.training.train_starvla][INFO] - Step 422, Loss: {'train/loss': 0.9171653985977173, 'train/grad_norm_pre_clip': 8.240192413330078, 'timing/data': 0.019490135833621025, 'timing/model': 1.363694685511291, 'learning_rate/qwen_vl_interface': 9.949043581574196e-06, 'learning_rate/action_model': 9.949043581574196e-05, 'epoch': 0.15}) +[2026-06-17 17:02:23,960][starVLA.training.train_starvla][INFO] - Step 423, Loss: {'train/loss': 1.0761643648147583, 'train/grad_norm_pre_clip': 9.213967323303223, 'timing/data': 0.02127247955650091, 'timing/model': 1.1267513362690806, 'learning_rate/qwen_vl_interface': 9.948727162134775e-06, 'learning_rate/action_model': 9.948727162134775e-05, 'epoch': 0.15}) +[2026-06-17 17:02:28,432][starVLA.training.train_starvla][INFO] - Step 424, Loss: {'train/loss': 1.0328480005264282, 'train/grad_norm_pre_clip': 15.8936185836792, 'timing/data': 0.0177239952608943, 'timing/model': 1.1433184510096908, 'learning_rate/qwen_vl_interface': 9.948409768644157e-06, 'learning_rate/action_model': 9.948409768644156e-05, 'epoch': 0.16}) +[2026-06-17 17:02:32,872][starVLA.training.train_starvla][INFO] - Step 425, Loss: {'train/loss': 1.0157743692398071, 'train/grad_norm_pre_clip': 11.254278182983398, 'timing/data': 0.02108688559383154, 'timing/model': 1.029542344622314, 'learning_rate/qwen_vl_interface': 9.948091401168139e-06, 'learning_rate/action_model': 9.948091401168138e-05, 'epoch': 0.16}) +[2026-06-17 17:02:37,208][starVLA.training.train_starvla][INFO] - Step 426, Loss: {'train/loss': 1.2364357709884644, 'train/grad_norm_pre_clip': 7.615171909332275, 'timing/data': 0.018816476687788963, 'timing/model': 1.028346593491733, 'learning_rate/qwen_vl_interface': 9.947772059772718e-06, 'learning_rate/action_model': 9.947772059772717e-05, 'epoch': 0.16}) +[2026-06-17 17:02:41,736][starVLA.training.train_starvla][INFO] - Step 427, Loss: {'train/loss': 0.9728720784187317, 'train/grad_norm_pre_clip': 11.411595344543457, 'timing/data': 0.019523395225405693, 'timing/model': 1.3955266131088138, 'learning_rate/qwen_vl_interface': 9.947451744524094e-06, 'learning_rate/action_model': 9.947451744524093e-05, 'epoch': 0.16}) +[2026-06-17 17:02:45,971][starVLA.training.train_starvla][INFO] - Step 428, Loss: {'train/loss': 0.9790072441101074, 'train/grad_norm_pre_clip': 9.251545906066895, 'timing/data': 0.048287936486303806, 'timing/model': 1.0977697307243943, 'learning_rate/qwen_vl_interface': 9.947130455488668e-06, 'learning_rate/action_model': 9.947130455488666e-05, 'epoch': 0.16}) +[2026-06-17 17:02:50,315][starVLA.training.train_starvla][INFO] - Step 429, Loss: {'train/loss': 0.9934110045433044, 'train/grad_norm_pre_clip': 11.622074127197266, 'timing/data': 0.019591959193348885, 'timing/model': 1.0440818686038256, 'learning_rate/qwen_vl_interface': 9.946808192733045e-06, 'learning_rate/action_model': 9.946808192733044e-05, 'epoch': 0.16}) +[2026-06-17 17:02:54,673][starVLA.training.train_starvla][INFO] - Step 430, Loss: {'train/loss': 1.2862999439239502, 'train/grad_norm_pre_clip': 12.849979400634766, 'timing/data': 0.019494454376399517, 'timing/model': 1.0790381208062172, 'learning_rate/qwen_vl_interface': 9.946484956324028e-06, 'learning_rate/action_model': 9.946484956324028e-05, 'epoch': 0.16}) +[2026-06-17 17:03:00,102][starVLA.training.train_starvla][INFO] - Step 431, Loss: {'train/loss': 1.0405781269073486, 'train/grad_norm_pre_clip': 11.236438751220703, 'timing/data': 0.021901268512010574, 'timing/model': 2.28055408783257, 'learning_rate/qwen_vl_interface': 9.946160746328628e-06, 'learning_rate/action_model': 9.946160746328627e-05, 'epoch': 0.16}) +[2026-06-17 17:03:04,423][starVLA.training.train_starvla][INFO] - Step 432, Loss: {'train/loss': 1.1602904796600342, 'train/grad_norm_pre_clip': 10.548171997070312, 'timing/data': 0.01900789700448513, 'timing/model': 1.3393833665177226, 'learning_rate/qwen_vl_interface': 9.945835562814052e-06, 'learning_rate/action_model': 9.945835562814051e-05, 'epoch': 0.16}) +[2026-06-17 17:03:08,731][starVLA.training.train_starvla][INFO] - Step 433, Loss: {'train/loss': 1.2008178234100342, 'train/grad_norm_pre_clip': 11.405056953430176, 'timing/data': 0.021481266245245934, 'timing/model': 1.1643693270161748, 'learning_rate/qwen_vl_interface': 9.94550940584771e-06, 'learning_rate/action_model': 9.945509405847709e-05, 'epoch': 0.16}) +[2026-06-17 17:03:14,132][starVLA.training.train_starvla][INFO] - Step 434, Loss: {'train/loss': 0.9421463012695312, 'train/grad_norm_pre_clip': 8.957818031311035, 'timing/data': 0.01979917846620083, 'timing/model': 1.0870776399970055, 'learning_rate/qwen_vl_interface': 9.945182275497215e-06, 'learning_rate/action_model': 9.945182275497215e-05, 'epoch': 0.16}) +[2026-06-17 17:03:18,813][starVLA.training.train_starvla][INFO] - Step 435, Loss: {'train/loss': 0.9751580953598022, 'train/grad_norm_pre_clip': 11.447574615478516, 'timing/data': 0.02577373757958412, 'timing/model': 1.0999632775783539, 'learning_rate/qwen_vl_interface': 9.944854171830382e-06, 'learning_rate/action_model': 9.944854171830383e-05, 'epoch': 0.16}) +[2026-06-17 17:03:23,405][starVLA.training.train_starvla][INFO] - Step 436, Loss: {'train/loss': 1.2262574434280396, 'train/grad_norm_pre_clip': 10.54857349395752, 'timing/data': 0.01908130943775177, 'timing/model': 1.1744147660210729, 'learning_rate/qwen_vl_interface': 9.94452509491523e-06, 'learning_rate/action_model': 9.94452509491523e-05, 'epoch': 0.16}) +[2026-06-17 17:03:27,979][starVLA.training.train_starvla][INFO] - Step 437, Loss: {'train/loss': 1.1980807781219482, 'train/grad_norm_pre_clip': 9.520925521850586, 'timing/data': 0.020124532282352448, 'timing/model': 1.5707535911351442, 'learning_rate/qwen_vl_interface': 9.944195044819973e-06, 'learning_rate/action_model': 9.944195044819972e-05, 'epoch': 0.16}) +[2026-06-17 17:03:32,105][starVLA.training.train_starvla][INFO] - Step 438, Loss: {'train/loss': 1.0509403944015503, 'train/grad_norm_pre_clip': 11.646806716918945, 'timing/data': 0.021445448510348797, 'timing/model': 1.038232684135437, 'learning_rate/qwen_vl_interface': 9.94386402161303e-06, 'learning_rate/action_model': 9.943864021613029e-05, 'epoch': 0.16}) +[2026-06-17 17:03:36,567][starVLA.training.train_starvla][INFO] - Step 439, Loss: {'train/loss': 1.0630466938018799, 'train/grad_norm_pre_clip': 11.908037185668945, 'timing/data': 0.020290391519665718, 'timing/model': 1.08993602078408, 'learning_rate/qwen_vl_interface': 9.943532025363026e-06, 'learning_rate/action_model': 9.943532025363025e-05, 'epoch': 0.16}) +[2026-06-17 17:03:40,991][starVLA.training.train_starvla][INFO] - Step 440, Loss: {'train/loss': 1.1288421154022217, 'train/grad_norm_pre_clip': 9.344803810119629, 'timing/data': 0.01933290157467127, 'timing/model': 1.0215617408975959, 'learning_rate/qwen_vl_interface': 9.943199056138781e-06, 'learning_rate/action_model': 9.943199056138782e-05, 'epoch': 0.16}) +[2026-06-17 17:03:45,356][starVLA.training.train_starvla][INFO] - Step 441, Loss: {'train/loss': 0.8642454743385315, 'train/grad_norm_pre_clip': 7.858716011047363, 'timing/data': 0.02230890467762947, 'timing/model': 1.389293111860752, 'learning_rate/qwen_vl_interface': 9.942865114009325e-06, 'learning_rate/action_model': 9.942865114009324e-05, 'epoch': 0.16}) +[2026-06-17 17:03:49,804][starVLA.training.train_starvla][INFO] - Step 442, Loss: {'train/loss': 1.2278610467910767, 'train/grad_norm_pre_clip': 10.287467002868652, 'timing/data': 0.021985705941915512, 'timing/model': 1.3992125699296594, 'learning_rate/qwen_vl_interface': 9.94253019904388e-06, 'learning_rate/action_model': 9.942530199043879e-05, 'epoch': 0.16}) +[2026-06-17 17:03:53,797][starVLA.training.train_starvla][INFO] - Step 443, Loss: {'train/loss': 1.1976449489593506, 'train/grad_norm_pre_clip': 8.63174057006836, 'timing/data': 0.02201987337321043, 'timing/model': 1.0160153051838279, 'learning_rate/qwen_vl_interface': 9.942194311311874e-06, 'learning_rate/action_model': 9.942194311311874e-05, 'epoch': 0.16}) +[2026-06-17 17:03:58,064][starVLA.training.train_starvla][INFO] - Step 444, Loss: {'train/loss': 0.9572306275367737, 'train/grad_norm_pre_clip': 12.606206893920898, 'timing/data': 0.0197969488799572, 'timing/model': 1.0219121053814888, 'learning_rate/qwen_vl_interface': 9.941857450882939e-06, 'learning_rate/action_model': 9.941857450882937e-05, 'epoch': 0.16}) +[2026-06-17 17:04:02,555][starVLA.training.train_starvla][INFO] - Step 445, Loss: {'train/loss': 0.9477835297584534, 'train/grad_norm_pre_clip': 16.930965423583984, 'timing/data': 0.02019591722637415, 'timing/model': 1.034189340658486, 'learning_rate/qwen_vl_interface': 9.941519617826906e-06, 'learning_rate/action_model': 9.941519617826906e-05, 'epoch': 0.16}) +[2026-06-17 17:04:08,922][starVLA.training.train_starvla][INFO] - Step 446, Loss: {'train/loss': 1.0399576425552368, 'train/grad_norm_pre_clip': 11.824332237243652, 'timing/data': 0.027755283750593662, 'timing/model': 1.1333493767306209, 'learning_rate/qwen_vl_interface': 9.941180812213807e-06, 'learning_rate/action_model': 9.941180812213806e-05, 'epoch': 0.16}) +[2026-06-17 17:04:13,353][starVLA.training.train_starvla][INFO] - Step 447, Loss: {'train/loss': 1.025921106338501, 'train/grad_norm_pre_clip': 9.845277786254883, 'timing/data': 0.019893469288945198, 'timing/model': 1.3273918805643916, 'learning_rate/qwen_vl_interface': 9.940841034113878e-06, 'learning_rate/action_model': 9.940841034113877e-05, 'epoch': 0.16}) +[2026-06-17 17:04:17,623][starVLA.training.train_starvla][INFO] - Step 448, Loss: {'train/loss': 1.0287671089172363, 'train/grad_norm_pre_clip': 11.52161693572998, 'timing/data': 0.023860741406679153, 'timing/model': 1.1009416738525033, 'learning_rate/qwen_vl_interface': 9.940500283597557e-06, 'learning_rate/action_model': 9.940500283597556e-05, 'epoch': 0.16}) +[2026-06-17 17:04:22,323][starVLA.training.train_starvla][INFO] - Step 449, Loss: {'train/loss': 0.9750997424125671, 'train/grad_norm_pre_clip': 7.970271587371826, 'timing/data': 0.018345901742577553, 'timing/model': 1.0155770936980844, 'learning_rate/qwen_vl_interface': 9.940158560735478e-06, 'learning_rate/action_model': 9.940158560735477e-05, 'epoch': 0.16}) +[2026-06-17 17:04:28,244][starVLA.training.train_starvla][INFO] - Step 450, Loss: {'train/loss': 1.4740391969680786, 'train/grad_norm_pre_clip': 9.952220916748047, 'timing/data': 0.019268427975475788, 'timing/model': 1.0433427738025784, 'learning_rate/qwen_vl_interface': 9.939815865598482e-06, 'learning_rate/action_model': 9.93981586559848e-05, 'epoch': 0.16}) +[2026-06-17 17:04:32,536][starVLA.training.train_starvla][INFO] - Step 451, Loss: {'train/loss': 1.331352710723877, 'train/grad_norm_pre_clip': 14.467923164367676, 'timing/data': 0.01842053420841694, 'timing/model': 1.3290617130696774, 'learning_rate/qwen_vl_interface': 9.939472198257611e-06, 'learning_rate/action_model': 9.93947219825761e-05, 'epoch': 0.17}) +[2026-06-17 17:04:36,915][starVLA.training.train_starvla][INFO] - Step 452, Loss: {'train/loss': 0.7878012657165527, 'train/grad_norm_pre_clip': 10.514585494995117, 'timing/data': 0.019524697214365005, 'timing/model': 1.3604126619175076, 'learning_rate/qwen_vl_interface': 9.939127558784108e-06, 'learning_rate/action_model': 9.939127558784107e-05, 'epoch': 0.17}) +[2026-06-17 17:04:40,969][starVLA.training.train_starvla][INFO] - Step 453, Loss: {'train/loss': 1.1068919897079468, 'train/grad_norm_pre_clip': 9.149413108825684, 'timing/data': 0.02012033201754093, 'timing/model': 1.043469524011016, 'learning_rate/qwen_vl_interface': 9.938781947249417e-06, 'learning_rate/action_model': 9.938781947249416e-05, 'epoch': 0.17}) +[2026-06-17 17:04:45,387][starVLA.training.train_starvla][INFO] - Step 454, Loss: {'train/loss': 1.1520622968673706, 'train/grad_norm_pre_clip': 15.707913398742676, 'timing/data': 0.0193772092461586, 'timing/model': 1.0327627742663026, 'learning_rate/qwen_vl_interface': 9.93843536372518e-06, 'learning_rate/action_model': 9.93843536372518e-05, 'epoch': 0.17}) +[2026-06-17 17:04:49,942][starVLA.training.train_starvla][INFO] - Step 455, Loss: {'train/loss': 1.0048089027404785, 'train/grad_norm_pre_clip': 13.020991325378418, 'timing/data': 0.031866119243204594, 'timing/model': 1.10241355933249, 'learning_rate/qwen_vl_interface': 9.938087808283248e-06, 'learning_rate/action_model': 9.938087808283248e-05, 'epoch': 0.17}) +[2026-06-17 17:04:54,199][starVLA.training.train_starvla][INFO] - Step 456, Loss: {'train/loss': 1.212895154953003, 'train/grad_norm_pre_clip': 10.824097633361816, 'timing/data': 0.020899142138659954, 'timing/model': 1.032137980684638, 'learning_rate/qwen_vl_interface': 9.937739280995672e-06, 'learning_rate/action_model': 9.937739280995672e-05, 'epoch': 0.17}) +[2026-06-17 17:04:58,811][starVLA.training.train_starvla][INFO] - Step 457, Loss: {'train/loss': 1.0430468320846558, 'train/grad_norm_pre_clip': 8.282275199890137, 'timing/data': 0.021836262196302414, 'timing/model': 1.518542030826211, 'learning_rate/qwen_vl_interface': 9.937389781934697e-06, 'learning_rate/action_model': 9.937389781934696e-05, 'epoch': 0.17}) +[2026-06-17 17:05:02,908][starVLA.training.train_starvla][INFO] - Step 458, Loss: {'train/loss': 1.0979264974594116, 'train/grad_norm_pre_clip': 7.18075704574585, 'timing/data': 0.019376692362129688, 'timing/model': 1.037953651510179, 'learning_rate/qwen_vl_interface': 9.937039311172776e-06, 'learning_rate/action_model': 9.937039311172776e-05, 'epoch': 0.17}) +[2026-06-17 17:05:07,372][starVLA.training.train_starvla][INFO] - Step 459, Loss: {'train/loss': 0.7479822039604187, 'train/grad_norm_pre_clip': 9.657946586608887, 'timing/data': 0.020915471017360687, 'timing/model': 1.051799695007503, 'learning_rate/qwen_vl_interface': 9.936687868782562e-06, 'learning_rate/action_model': 9.936687868782561e-05, 'epoch': 0.17}) +[2026-06-17 17:05:11,877][starVLA.training.train_starvla][INFO] - Step 460, Loss: {'train/loss': 0.9043889045715332, 'train/grad_norm_pre_clip': 8.27604866027832, 'timing/data': 0.03308678511530161, 'timing/model': 1.088721171952784, 'learning_rate/qwen_vl_interface': 9.93633545483691e-06, 'learning_rate/action_model': 9.93633545483691e-05, 'epoch': 0.17}) +[2026-06-17 17:05:16,317][starVLA.training.train_starvla][INFO] - Step 461, Loss: {'train/loss': 1.1739158630371094, 'train/grad_norm_pre_clip': 20.323570251464844, 'timing/data': 0.022676879540085793, 'timing/model': 1.3643512427806854, 'learning_rate/qwen_vl_interface': 9.935982069408877e-06, 'learning_rate/action_model': 9.935982069408877e-05, 'epoch': 0.17}) +[2026-06-17 17:05:21,851][starVLA.training.train_starvla][INFO] - Step 462, Loss: {'train/loss': 1.036393165588379, 'train/grad_norm_pre_clip': 10.073607444763184, 'timing/data': 0.018816102296113968, 'timing/model': 2.465302735567093, 'learning_rate/qwen_vl_interface': 9.935627712571716e-06, 'learning_rate/action_model': 9.935627712571715e-05, 'epoch': 0.17}) +[2026-06-17 17:05:25,902][starVLA.training.train_starvla][INFO] - Step 463, Loss: {'train/loss': 1.2940597534179688, 'train/grad_norm_pre_clip': 10.975186347961426, 'timing/data': 0.022694796323776245, 'timing/model': 1.0654983464628458, 'learning_rate/qwen_vl_interface': 9.93527238439889e-06, 'learning_rate/action_model': 9.935272384398889e-05, 'epoch': 0.17}) +[2026-06-17 17:05:30,372][starVLA.training.train_starvla][INFO] - Step 464, Loss: {'train/loss': 1.2920713424682617, 'train/grad_norm_pre_clip': 11.983929634094238, 'timing/data': 0.01987764611840248, 'timing/model': 1.0228041308000684, 'learning_rate/qwen_vl_interface': 9.934916084964056e-06, 'learning_rate/action_model': 9.934916084964056e-05, 'epoch': 0.17}) +[2026-06-17 17:05:34,768][starVLA.training.train_starvla][INFO] - Step 465, Loss: {'train/loss': 0.9786317944526672, 'train/grad_norm_pre_clip': 11.379566192626953, 'timing/data': 0.019304240122437477, 'timing/model': 1.059942756779492, 'learning_rate/qwen_vl_interface': 9.934558814341077e-06, 'learning_rate/action_model': 9.934558814341077e-05, 'epoch': 0.17}) +[2026-06-17 17:05:39,170][starVLA.training.train_starvla][INFO] - Step 466, Loss: {'train/loss': 0.9955081939697266, 'train/grad_norm_pre_clip': 11.84821605682373, 'timing/data': 0.019336123019456863, 'timing/model': 1.0325705092400312, 'learning_rate/qwen_vl_interface': 9.934200572604015e-06, 'learning_rate/action_model': 9.934200572604015e-05, 'epoch': 0.17}) +[2026-06-17 17:05:43,530][starVLA.training.train_starvla][INFO] - Step 467, Loss: {'train/loss': 1.1322693824768066, 'train/grad_norm_pre_clip': 10.883112907409668, 'timing/data': 0.020228630863130093, 'timing/model': 1.3758764937520027, 'learning_rate/qwen_vl_interface': 9.933841359827133e-06, 'learning_rate/action_model': 9.933841359827133e-05, 'epoch': 0.17}) +[2026-06-17 17:05:47,606][starVLA.training.train_starvla][INFO] - Step 468, Loss: {'train/loss': 1.1232084035873413, 'train/grad_norm_pre_clip': 9.434732437133789, 'timing/data': 0.020636706613004208, 'timing/model': 1.0292179826647043, 'learning_rate/qwen_vl_interface': 9.933481176084899e-06, 'learning_rate/action_model': 9.933481176084899e-05, 'epoch': 0.17}) +[2026-06-17 17:05:53,232][starVLA.training.train_starvla][INFO] - Step 469, Loss: {'train/loss': 1.1568493843078613, 'train/grad_norm_pre_clip': 9.206536293029785, 'timing/data': 0.019164705649018288, 'timing/model': 1.0368308210745454, 'learning_rate/qwen_vl_interface': 9.933120021451976e-06, 'learning_rate/action_model': 9.933120021451976e-05, 'epoch': 0.17}) +[2026-06-17 17:05:57,728][starVLA.training.train_starvla][INFO] - Step 470, Loss: {'train/loss': 1.3499257564544678, 'train/grad_norm_pre_clip': 10.389280319213867, 'timing/data': 0.02063007280230522, 'timing/model': 1.0605449201539159, 'learning_rate/qwen_vl_interface': 9.932757896003235e-06, 'learning_rate/action_model': 9.932757896003235e-05, 'epoch': 0.17}) +[2026-06-17 17:06:02,288][starVLA.training.train_starvla][INFO] - Step 471, Loss: {'train/loss': 1.0732853412628174, 'train/grad_norm_pre_clip': 9.885515213012695, 'timing/data': 0.019633501768112183, 'timing/model': 1.4889125572517514, 'learning_rate/qwen_vl_interface': 9.932394799813742e-06, 'learning_rate/action_model': 9.932394799813741e-05, 'epoch': 0.17}) +[2026-06-17 17:06:06,821][starVLA.training.train_starvla][INFO] - Step 472, Loss: {'train/loss': 1.266188383102417, 'train/grad_norm_pre_clip': 37.759037017822266, 'timing/data': 0.021874848753213882, 'timing/model': 1.316489053890109, 'learning_rate/qwen_vl_interface': 9.932030732958768e-06, 'learning_rate/action_model': 9.932030732958769e-05, 'epoch': 0.17}) +[2026-06-17 17:06:11,048][starVLA.training.train_starvla][INFO] - Step 473, Loss: {'train/loss': 1.015833854675293, 'train/grad_norm_pre_clip': 14.547627449035645, 'timing/data': 0.02265447936952114, 'timing/model': 1.1096538044512272, 'learning_rate/qwen_vl_interface': 9.931665695513786e-06, 'learning_rate/action_model': 9.931665695513786e-05, 'epoch': 0.17}) +[2026-06-17 17:06:15,489][starVLA.training.train_starvla][INFO] - Step 474, Loss: {'train/loss': 1.3195915222167969, 'train/grad_norm_pre_clip': 8.542572021484375, 'timing/data': 0.0196422366425395, 'timing/model': 1.082127932459116, 'learning_rate/qwen_vl_interface': 9.931299687554466e-06, 'learning_rate/action_model': 9.931299687554466e-05, 'epoch': 0.17}) +[2026-06-17 17:06:19,877][starVLA.training.train_starvla][INFO] - Step 475, Loss: {'train/loss': 1.113008975982666, 'train/grad_norm_pre_clip': 8.649917602539062, 'timing/data': 0.021267910487949848, 'timing/model': 1.0555859059095383, 'learning_rate/qwen_vl_interface': 9.930932709156686e-06, 'learning_rate/action_model': 9.930932709156685e-05, 'epoch': 0.17}) +[2026-06-17 17:06:24,330][starVLA.training.train_starvla][INFO] - Step 476, Loss: {'train/loss': 0.8801599144935608, 'train/grad_norm_pre_clip': 8.352463722229004, 'timing/data': 0.02293036598712206, 'timing/model': 1.028165191411972, 'learning_rate/qwen_vl_interface': 9.930564760396518e-06, 'learning_rate/action_model': 9.930564760396518e-05, 'epoch': 0.17}) +[2026-06-17 17:06:28,717][starVLA.training.train_starvla][INFO] - Step 477, Loss: {'train/loss': 1.0989383459091187, 'train/grad_norm_pre_clip': 9.999788284301758, 'timing/data': 0.021450014784932137, 'timing/model': 1.420454729348421, 'learning_rate/qwen_vl_interface': 9.930195841350237e-06, 'learning_rate/action_model': 9.930195841350236e-05, 'epoch': 0.17}) +[2026-06-17 17:06:33,021][starVLA.training.train_starvla][INFO] - Step 478, Loss: {'train/loss': 1.1041353940963745, 'train/grad_norm_pre_clip': 9.969280242919922, 'timing/data': 0.021374154835939407, 'timing/model': 1.1064926395192742, 'learning_rate/qwen_vl_interface': 9.929825952094323e-06, 'learning_rate/action_model': 9.929825952094323e-05, 'epoch': 0.18}) +[2026-06-17 17:06:37,380][starVLA.training.train_starvla][INFO] - Step 479, Loss: {'train/loss': 1.120593547821045, 'train/grad_norm_pre_clip': 9.215652465820312, 'timing/data': 0.020537308417260647, 'timing/model': 1.0878358958289027, 'learning_rate/qwen_vl_interface': 9.929455092705453e-06, 'learning_rate/action_model': 9.929455092705453e-05, 'epoch': 0.18}) +[2026-06-17 17:06:41,826][starVLA.training.train_starvla][INFO] - Step 480, Loss: {'train/loss': 1.1623413562774658, 'train/grad_norm_pre_clip': 8.06666088104248, 'timing/data': 0.020322349853813648, 'timing/model': 1.0388316689059138, 'learning_rate/qwen_vl_interface': 9.929083263260508e-06, 'learning_rate/action_model': 9.929083263260507e-05, 'epoch': 0.18}) +[2026-06-17 17:06:46,196][starVLA.training.train_starvla][INFO] - Step 481, Loss: {'train/loss': 0.7132042646408081, 'train/grad_norm_pre_clip': 7.591670513153076, 'timing/data': 0.02018317859619856, 'timing/model': 1.338778618723154, 'learning_rate/qwen_vl_interface': 9.928710463836566e-06, 'learning_rate/action_model': 9.928710463836565e-05, 'epoch': 0.18}) +[2026-06-17 17:06:50,558][starVLA.training.train_starvla][INFO] - Step 482, Loss: {'train/loss': 0.869500994682312, 'train/grad_norm_pre_clip': 8.213979721069336, 'timing/data': 0.021651672199368477, 'timing/model': 1.3835042621940374, 'learning_rate/qwen_vl_interface': 9.928336694510912e-06, 'learning_rate/action_model': 9.928336694510911e-05, 'epoch': 0.18}) +[2026-06-17 17:06:55,104][starVLA.training.train_starvla][INFO] - Step 483, Loss: {'train/loss': 1.5140864849090576, 'train/grad_norm_pre_clip': 12.428789138793945, 'timing/data': 0.022509869188070297, 'timing/model': 1.0995228150859475, 'learning_rate/qwen_vl_interface': 9.927961955361025e-06, 'learning_rate/action_model': 9.927961955361025e-05, 'epoch': 0.18}) +[2026-06-17 17:06:59,507][starVLA.training.train_starvla][INFO] - Step 484, Loss: {'train/loss': 1.0718300342559814, 'train/grad_norm_pre_clip': 7.496310234069824, 'timing/data': 0.02074038702994585, 'timing/model': 1.1336352713406086, 'learning_rate/qwen_vl_interface': 9.927586246464593e-06, 'learning_rate/action_model': 9.927586246464593e-05, 'epoch': 0.18}) +[2026-06-17 17:07:04,010][starVLA.training.train_starvla][INFO] - Step 485, Loss: {'train/loss': 1.2324446439743042, 'train/grad_norm_pre_clip': 10.542070388793945, 'timing/data': 0.02056429348886013, 'timing/model': 1.1251579327508807, 'learning_rate/qwen_vl_interface': 9.927209567899499e-06, 'learning_rate/action_model': 9.927209567899498e-05, 'epoch': 0.18}) +[2026-06-17 17:07:08,303][starVLA.training.train_starvla][INFO] - Step 486, Loss: {'train/loss': 1.26580810546875, 'train/grad_norm_pre_clip': 9.575407981872559, 'timing/data': 0.025051147677004337, 'timing/model': 1.073925037868321, 'learning_rate/qwen_vl_interface': 9.926831919743827e-06, 'learning_rate/action_model': 9.926831919743827e-05, 'epoch': 0.18}) +[2026-06-17 17:07:12,779][starVLA.training.train_starvla][INFO] - Step 487, Loss: {'train/loss': 1.2466691732406616, 'train/grad_norm_pre_clip': 8.841318130493164, 'timing/data': 0.020186985842883587, 'timing/model': 1.3408198822289705, 'learning_rate/qwen_vl_interface': 9.926453302075866e-06, 'learning_rate/action_model': 9.926453302075867e-05, 'epoch': 0.18}) +[2026-06-17 17:07:16,865][starVLA.training.train_starvla][INFO] - Step 488, Loss: {'train/loss': 0.921149730682373, 'train/grad_norm_pre_clip': 8.843260765075684, 'timing/data': 0.02298772521317005, 'timing/model': 1.0309605859220028, 'learning_rate/qwen_vl_interface': 9.926073714974103e-06, 'learning_rate/action_model': 9.926073714974104e-05, 'epoch': 0.18}) +[2026-06-17 17:07:21,154][starVLA.training.train_starvla][INFO] - Step 489, Loss: {'train/loss': 1.4369648694992065, 'train/grad_norm_pre_clip': 12.117246627807617, 'timing/data': 0.018961094319820404, 'timing/model': 1.018201588653028, 'learning_rate/qwen_vl_interface': 9.925693158517228e-06, 'learning_rate/action_model': 9.925693158517228e-05, 'epoch': 0.18}) +[2026-06-17 17:07:25,526][starVLA.training.train_starvla][INFO] - Step 490, Loss: {'train/loss': 1.257442593574524, 'train/grad_norm_pre_clip': 11.053895950317383, 'timing/data': 0.019455982372164726, 'timing/model': 1.04142673779279, 'learning_rate/qwen_vl_interface': 9.92531163278413e-06, 'learning_rate/action_model': 9.92531163278413e-05, 'epoch': 0.18}) +[2026-06-17 17:07:29,997][starVLA.training.train_starvla][INFO] - Step 491, Loss: {'train/loss': 1.0808606147766113, 'train/grad_norm_pre_clip': 8.805867195129395, 'timing/data': 0.022753732278943062, 'timing/model': 1.3613905850797892, 'learning_rate/qwen_vl_interface': 9.924929137853899e-06, 'learning_rate/action_model': 9.924929137853898e-05, 'epoch': 0.18}) +[2026-06-17 17:07:34,263][starVLA.training.train_starvla][INFO] - Step 492, Loss: {'train/loss': 0.9969719648361206, 'train/grad_norm_pre_clip': 9.783205032348633, 'timing/data': 0.01861459296196699, 'timing/model': 1.293175152502954, 'learning_rate/qwen_vl_interface': 9.92454567380583e-06, 'learning_rate/action_model': 9.924545673805829e-05, 'epoch': 0.18}) +[2026-06-17 17:07:38,291][starVLA.training.train_starvla][INFO] - Step 493, Loss: {'train/loss': 1.2242138385772705, 'train/grad_norm_pre_clip': 12.08823013305664, 'timing/data': 0.018612723797559738, 'timing/model': 1.0416093803942204, 'learning_rate/qwen_vl_interface': 9.924161240719409e-06, 'learning_rate/action_model': 9.924161240719408e-05, 'epoch': 0.18}) +[2026-06-17 17:07:44,021][starVLA.training.train_starvla][INFO] - Step 494, Loss: {'train/loss': 1.5298773050308228, 'train/grad_norm_pre_clip': 11.55123519897461, 'timing/data': 0.023105362430214882, 'timing/model': 1.0466701406985521, 'learning_rate/qwen_vl_interface': 9.923775838674336e-06, 'learning_rate/action_model': 9.923775838674335e-05, 'epoch': 0.18}) +[2026-06-17 17:07:48,517][starVLA.training.train_starvla][INFO] - Step 495, Loss: {'train/loss': 1.0265616178512573, 'train/grad_norm_pre_clip': 8.128798484802246, 'timing/data': 0.01964481081813574, 'timing/model': 1.0898581771180034, 'learning_rate/qwen_vl_interface': 9.9233894677505e-06, 'learning_rate/action_model': 9.9233894677505e-05, 'epoch': 0.18}) +[2026-06-17 17:07:53,488][starVLA.training.train_starvla][INFO] - Step 496, Loss: {'train/loss': 1.1060668230056763, 'train/grad_norm_pre_clip': 13.722424507141113, 'timing/data': 0.021673496812582016, 'timing/model': 1.028475328348577, 'learning_rate/qwen_vl_interface': 9.923002128028e-06, 'learning_rate/action_model': 9.923002128028e-05, 'epoch': 0.18}) +[2026-06-17 17:07:57,891][starVLA.training.train_starvla][INFO] - Step 497, Loss: {'train/loss': 1.0166839361190796, 'train/grad_norm_pre_clip': 10.618631362915039, 'timing/data': 0.022118061780929565, 'timing/model': 1.3687515556812286, 'learning_rate/qwen_vl_interface': 9.922613819587131e-06, 'learning_rate/action_model': 9.92261381958713e-05, 'epoch': 0.18}) +[2026-06-17 17:08:01,957][starVLA.training.train_starvla][INFO] - Step 498, Loss: {'train/loss': 1.3144805431365967, 'train/grad_norm_pre_clip': 10.286073684692383, 'timing/data': 0.021403211168944836, 'timing/model': 1.0742594227194786, 'learning_rate/qwen_vl_interface': 9.922224542508389e-06, 'learning_rate/action_model': 9.922224542508388e-05, 'epoch': 0.18}) +[2026-06-17 17:08:07,589][starVLA.training.train_starvla][INFO] - Step 499, Loss: {'train/loss': 1.3149720430374146, 'train/grad_norm_pre_clip': 11.236574172973633, 'timing/data': 0.0213850736618042, 'timing/model': 1.09714316111058, 'learning_rate/qwen_vl_interface': 9.92183429687247e-06, 'learning_rate/action_model': 9.92183429687247e-05, 'epoch': 0.18}) +[2026-06-17 17:10:48,279][starVLA.training.train_starvla][INFO] - Step 500, Loss: {'train/loss': 1.4291484355926514, 'train/grad_norm_pre_clip': 14.90717887878418, 'eval/loss': 1.0869797468185425, 'eval/latency_2/loss': 1.0869797468185425, 'eval/demon_fire_precision': 0.7651888341543513, 'eval/demon_fire_recall': 0.9748953974895398, 'eval/demon_fire_f1': 0.8574057037718491, 'eval/demon_fire_rate_model': 0.8472454090150251, 'eval/demon_fire_rate_teacher': 0.6649972175848636, 'eval/demon_left_precision': 0.7208182912154031, 'eval/demon_left_recall': 0.6305263157894737, 'eval/demon_left_f1': 0.6726558113419427, 'eval/demon_left_rate_model': 0.23121869782971619, 'eval/demon_left_rate_teacher': 0.2643294379521425, 'eval/demon_right_precision': 0.6283095723014257, 'eval/demon_right_recall': 0.6971751412429379, 'eval/demon_right_f1': 0.6609534011783611, 'eval/demon_right_rate_model': 0.27323316638842515, 'eval/demon_right_rate_teacher': 0.24624373956594323, 'eval/demon_move_f1': 0.6668046062601519, 'eval/demon_cc_f1': 0.7896878024169998, 'eval/latency_2/demon_fire_precision': 0.7651888341543513, 'eval/latency_2/demon_fire_recall': 0.9748953974895398, 'eval/latency_2/demon_fire_f1': 0.8574057037718491, 'eval/latency_2/demon_fire_rate_model': 0.8472454090150251, 'eval/latency_2/demon_fire_rate_teacher': 0.6649972175848636, 'eval/latency_2/demon_left_precision': 0.7208182912154031, 'eval/latency_2/demon_left_recall': 0.6305263157894737, 'eval/latency_2/demon_left_f1': 0.6726558113419427, 'eval/latency_2/demon_left_rate_model': 0.23121869782971619, 'eval/latency_2/demon_left_rate_teacher': 0.2643294379521425, 'eval/latency_2/demon_right_precision': 0.6283095723014257, 'eval/latency_2/demon_right_recall': 0.6971751412429379, 'eval/latency_2/demon_right_f1': 0.6609534011783611, 'eval/latency_2/demon_right_rate_model': 0.27323316638842515, 'eval/latency_2/demon_right_rate_teacher': 0.24624373956594323, 'eval/latency_2/demon_move_f1': 0.6668046062601519, 'eval/latency_2/demon_cc_f1': 0.7896878024169998, 'timing/data': 0.021990230306982994, 'timing/model': 1.1000912161543965, 'learning_rate/qwen_vl_interface': 9.921443082760276e-06, 'learning_rate/action_model': 9.921443082760276e-05, 'epoch': 0.18}) +[2026-06-17 17:10:48,280][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_500_state +[2026-06-17 17:10:48,288][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 17:10:57,234][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_500_state/model.safetensors +[2026-06-17 17:11:11,202][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_500_state/optimizer.bin +[2026-06-17 17:11:11,203][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_500_state/sampler.bin +[2026-06-17 17:11:11,204][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_500_state/sampler_1.bin +[2026-06-17 17:11:11,206][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_500_state/random_states_0.pkl +[2026-06-17 17:11:11,211][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 500 +[2026-06-17 17:11:11,228][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 17:11:11,244][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 17:11:15,798][starVLA.training.train_starvla][INFO] - Step 501, Loss: {'train/loss': 1.287290334701538, 'train/grad_norm_pre_clip': 11.040101051330566, 'timing/data': 0.01913260854780674, 'timing/model': 1.0375110525637865, 'learning_rate/qwen_vl_interface': 9.921050900252904e-06, 'learning_rate/action_model': 9.921050900252903e-05, 'epoch': 0.18}) +[2026-06-17 17:11:20,240][starVLA.training.train_starvla][INFO] - Step 502, Loss: {'train/loss': 0.9428252577781677, 'train/grad_norm_pre_clip': 8.758962631225586, 'timing/data': 0.023104517720639706, 'timing/model': 1.0622558081522584, 'learning_rate/qwen_vl_interface': 9.920657749431654e-06, 'learning_rate/action_model': 9.920657749431654e-05, 'epoch': 0.18}) +[2026-06-17 17:11:24,778][starVLA.training.train_starvla][INFO] - Step 503, Loss: {'train/loss': 1.272578477859497, 'train/grad_norm_pre_clip': 11.644164085388184, 'timing/data': 0.01978009007871151, 'timing/model': 1.5013719219714403, 'learning_rate/qwen_vl_interface': 9.920263630378026e-06, 'learning_rate/action_model': 9.920263630378024e-05, 'epoch': 0.18}) +[2026-06-17 17:11:28,896][starVLA.training.train_starvla][INFO] - Step 504, Loss: {'train/loss': 0.8827171325683594, 'train/grad_norm_pre_clip': 9.24775218963623, 'timing/data': 0.018969709053635597, 'timing/model': 1.0493711438030005, 'learning_rate/qwen_vl_interface': 9.919868543173722e-06, 'learning_rate/action_model': 9.919868543173721e-05, 'epoch': 0.18}) +[2026-06-17 17:11:33,497][starVLA.training.train_starvla][INFO] - Step 505, Loss: {'train/loss': 1.1066701412200928, 'train/grad_norm_pre_clip': 10.994174003601074, 'timing/data': 0.021512407809495926, 'timing/model': 1.0962824700400233, 'learning_rate/qwen_vl_interface': 9.919472487900645e-06, 'learning_rate/action_model': 9.919472487900644e-05, 'epoch': 0.18}) +[2026-06-17 17:11:38,117][starVLA.training.train_starvla][INFO] - Step 506, Loss: {'train/loss': 1.3683509826660156, 'train/grad_norm_pre_clip': 10.093681335449219, 'timing/data': 0.022105801850557327, 'timing/model': 1.0623994199559093, 'learning_rate/qwen_vl_interface': 9.919075464640895e-06, 'learning_rate/action_model': 9.919075464640895e-05, 'epoch': 0.19}) +[2026-06-17 17:11:42,526][starVLA.training.train_starvla][INFO] - Step 507, Loss: {'train/loss': 1.2219953536987305, 'train/grad_norm_pre_clip': 12.106948852539062, 'timing/data': 0.019428051076829433, 'timing/model': 1.0222023893147707, 'learning_rate/qwen_vl_interface': 9.91867747347678e-06, 'learning_rate/action_model': 9.918677473476779e-05, 'epoch': 0.19}) +[2026-06-17 17:11:47,313][starVLA.training.train_starvla][INFO] - Step 508, Loss: {'train/loss': 0.9644173383712769, 'train/grad_norm_pre_clip': 10.318230628967285, 'timing/data': 0.020647373981773853, 'timing/model': 1.5167256407439709, 'learning_rate/qwen_vl_interface': 9.918278514490798e-06, 'learning_rate/action_model': 9.918278514490799e-05, 'epoch': 0.19}) +[2026-06-17 17:11:51,631][starVLA.training.train_starvla][INFO] - Step 509, Loss: {'train/loss': 0.935226321220398, 'train/grad_norm_pre_clip': 7.572553634643555, 'timing/data': 0.03952877223491669, 'timing/model': 1.1568211112171412, 'learning_rate/qwen_vl_interface': 9.917878587765658e-06, 'learning_rate/action_model': 9.917878587765658e-05, 'epoch': 0.19}) +[2026-06-17 17:11:56,226][starVLA.training.train_starvla][INFO] - Step 510, Loss: {'train/loss': 1.096762776374817, 'train/grad_norm_pre_clip': 13.933797836303711, 'timing/data': 0.018701577559113503, 'timing/model': 1.0498195001855493, 'learning_rate/qwen_vl_interface': 9.917477693384264e-06, 'learning_rate/action_model': 9.917477693384264e-05, 'epoch': 0.19}) +[2026-06-17 17:12:00,729][starVLA.training.train_starvla][INFO] - Step 511, Loss: {'train/loss': 0.9240381717681885, 'train/grad_norm_pre_clip': 13.522148132324219, 'timing/data': 0.02191825769841671, 'timing/model': 1.121411982923746, 'learning_rate/qwen_vl_interface': 9.917075831429721e-06, 'learning_rate/action_model': 9.917075831429722e-05, 'epoch': 0.19}) +[2026-06-17 17:12:05,430][starVLA.training.train_starvla][INFO] - Step 512, Loss: {'train/loss': 0.9079833626747131, 'train/grad_norm_pre_clip': 7.033409118652344, 'timing/data': 0.020507020875811577, 'timing/model': 1.495917578227818, 'learning_rate/qwen_vl_interface': 9.916673001985339e-06, 'learning_rate/action_model': 9.916673001985338e-05, 'epoch': 0.19}) +[2026-06-17 17:12:13,931][starVLA.training.train_starvla][INFO] - Step 513, Loss: {'train/loss': 1.1685818433761597, 'train/grad_norm_pre_clip': 9.733805656433105, 'timing/data': 0.021690713241696358, 'timing/model': 1.334966579452157, 'learning_rate/qwen_vl_interface': 9.91626920513462e-06, 'learning_rate/action_model': 9.916269205134619e-05, 'epoch': 0.19}) +[2026-06-17 17:12:20,931][starVLA.training.train_starvla][INFO] - Step 514, Loss: {'train/loss': 1.3969647884368896, 'train/grad_norm_pre_clip': 14.561269760131836, 'timing/data': 0.031047459691762924, 'timing/model': 1.5021094614639878, 'learning_rate/qwen_vl_interface': 9.915864440961272e-06, 'learning_rate/action_model': 9.915864440961272e-05, 'epoch': 0.19}) +[2026-06-17 17:12:26,278][starVLA.training.train_starvla][INFO] - Step 515, Loss: {'train/loss': 0.57793128490448, 'train/grad_norm_pre_clip': 7.2474446296691895, 'timing/data': 0.022300262935459614, 'timing/model': 1.2396590942516923, 'learning_rate/qwen_vl_interface': 9.915458709549205e-06, 'learning_rate/action_model': 9.915458709549206e-05, 'epoch': 0.19}) +[2026-06-17 17:12:32,261][starVLA.training.train_starvla][INFO] - Step 516, Loss: {'train/loss': 1.1522022485733032, 'train/grad_norm_pre_clip': 9.64298152923584, 'timing/data': 0.020528786815702915, 'timing/model': 2.5789246624335647, 'learning_rate/qwen_vl_interface': 9.91505201098253e-06, 'learning_rate/action_model': 9.91505201098253e-05, 'epoch': 0.19}) +[2026-06-17 17:12:39,340][starVLA.training.train_starvla][INFO] - Step 517, Loss: {'train/loss': 0.9018872976303101, 'train/grad_norm_pre_clip': 8.60446548461914, 'timing/data': 0.02217667829245329, 'timing/model': 1.5910478290170431, 'learning_rate/qwen_vl_interface': 9.914644345345551e-06, 'learning_rate/action_model': 9.914644345345552e-05, 'epoch': 0.19}) +[2026-06-17 17:12:54,097][starVLA.training.train_starvla][INFO] - Step 518, Loss: {'train/loss': 1.1047695875167847, 'train/grad_norm_pre_clip': 9.139427185058594, 'timing/data': 0.0253269886597991, 'timing/model': 4.268004843965173, 'learning_rate/qwen_vl_interface': 9.914235712722782e-06, 'learning_rate/action_model': 9.914235712722781e-05, 'epoch': 0.19}) +[2026-06-17 17:13:03,095][starVLA.training.train_starvla][INFO] - Step 519, Loss: {'train/loss': 1.04206383228302, 'train/grad_norm_pre_clip': 7.895787239074707, 'timing/data': 0.023319187574088573, 'timing/model': 1.3957525929436088, 'learning_rate/qwen_vl_interface': 9.913826113198927e-06, 'learning_rate/action_model': 9.913826113198927e-05, 'epoch': 0.19}) +[2026-06-17 17:13:14,353][starVLA.training.train_starvla][INFO] - Step 520, Loss: {'train/loss': 1.9402283430099487, 'train/grad_norm_pre_clip': 17.4179744720459, 'timing/data': 0.019568814896047115, 'timing/model': 4.38449486438185, 'learning_rate/qwen_vl_interface': 9.913415546858904e-06, 'learning_rate/action_model': 9.913415546858904e-05, 'epoch': 0.19}) +[2026-06-17 17:13:19,111][starVLA.training.train_starvla][INFO] - Step 521, Loss: {'train/loss': 0.9999784827232361, 'train/grad_norm_pre_clip': 10.795857429504395, 'timing/data': 0.022891215980052948, 'timing/model': 1.0546610560268164, 'learning_rate/qwen_vl_interface': 9.913004013787818e-06, 'learning_rate/action_model': 9.913004013787819e-05, 'epoch': 0.19}) +[2026-06-17 17:13:23,915][starVLA.training.train_starvla][INFO] - Step 522, Loss: {'train/loss': 0.9816137552261353, 'train/grad_norm_pre_clip': 9.829666137695312, 'timing/data': 0.023139390163123608, 'timing/model': 1.6115249125286937, 'learning_rate/qwen_vl_interface': 9.912591514070983e-06, 'learning_rate/action_model': 9.912591514070983e-05, 'epoch': 0.19}) +[2026-06-17 17:13:28,351][starVLA.training.train_starvla][INFO] - Step 523, Loss: {'train/loss': 0.70357346534729, 'train/grad_norm_pre_clip': 8.4360990524292, 'timing/data': 0.02665605302900076, 'timing/model': 1.4177013067528605, 'learning_rate/qwen_vl_interface': 9.912178047793912e-06, 'learning_rate/action_model': 9.91217804779391e-05, 'epoch': 0.19}) +[2026-06-17 17:13:32,523][starVLA.training.train_starvla][INFO] - Step 524, Loss: {'train/loss': 0.9696853160858154, 'train/grad_norm_pre_clip': 8.87833309173584, 'timing/data': 0.019802751019597054, 'timing/model': 1.024357257410884, 'learning_rate/qwen_vl_interface': 9.911763615042311e-06, 'learning_rate/action_model': 9.91176361504231e-05, 'epoch': 0.19}) +[2026-06-17 17:13:36,951][starVLA.training.train_starvla][INFO] - Step 525, Loss: {'train/loss': 1.568108320236206, 'train/grad_norm_pre_clip': 12.867558479309082, 'timing/data': 0.028992448933422565, 'timing/model': 1.1108825784176588, 'learning_rate/qwen_vl_interface': 9.911348215902099e-06, 'learning_rate/action_model': 9.911348215902099e-05, 'epoch': 0.19}) +[2026-06-17 17:13:41,479][starVLA.training.train_starvla][INFO] - Step 526, Loss: {'train/loss': 0.9695823788642883, 'train/grad_norm_pre_clip': 10.456868171691895, 'timing/data': 0.02368586789816618, 'timing/model': 1.0659350100904703, 'learning_rate/qwen_vl_interface': 9.910931850459385e-06, 'learning_rate/action_model': 9.910931850459385e-05, 'epoch': 0.19}) +[2026-06-17 17:13:45,953][starVLA.training.train_starvla][INFO] - Step 527, Loss: {'train/loss': 1.18576180934906, 'train/grad_norm_pre_clip': 13.917154312133789, 'timing/data': 0.020486213266849518, 'timing/model': 1.065661832690239, 'learning_rate/qwen_vl_interface': 9.910514518800485e-06, 'learning_rate/action_model': 9.910514518800484e-05, 'epoch': 0.19}) +[2026-06-17 17:13:50,314][starVLA.training.train_starvla][INFO] - Step 528, Loss: {'train/loss': 0.947296679019928, 'train/grad_norm_pre_clip': 8.21003246307373, 'timing/data': 0.020815696567296982, 'timing/model': 1.3590732831507921, 'learning_rate/qwen_vl_interface': 9.910096221011908e-06, 'learning_rate/action_model': 9.910096221011907e-05, 'epoch': 0.19}) +[2026-06-17 17:13:54,490][starVLA.training.train_starvla][INFO] - Step 529, Loss: {'train/loss': 1.1941890716552734, 'train/grad_norm_pre_clip': 9.133654594421387, 'timing/data': 0.019819794222712517, 'timing/model': 1.0711324717849493, 'learning_rate/qwen_vl_interface': 9.90967695718037e-06, 'learning_rate/action_model': 9.90967695718037e-05, 'epoch': 0.19}) +[2026-06-17 17:13:59,065][starVLA.training.train_starvla][INFO] - Step 530, Loss: {'train/loss': 0.9206918478012085, 'train/grad_norm_pre_clip': 8.423666000366211, 'timing/data': 0.02072122972458601, 'timing/model': 1.117416582070291, 'learning_rate/qwen_vl_interface': 9.909256727392786e-06, 'learning_rate/action_model': 9.909256727392784e-05, 'epoch': 0.19}) +[2026-06-17 17:14:03,618][starVLA.training.train_starvla][INFO] - Step 531, Loss: {'train/loss': 1.3922561407089233, 'train/grad_norm_pre_clip': 10.81475830078125, 'timing/data': 0.022794765420258045, 'timing/model': 1.0877454532310367, 'learning_rate/qwen_vl_interface': 9.908835531736269e-06, 'learning_rate/action_model': 9.908835531736267e-05, 'epoch': 0.19}) +[2026-06-17 17:14:08,175][starVLA.training.train_starvla][INFO] - Step 532, Loss: {'train/loss': 1.0312296152114868, 'train/grad_norm_pre_clip': 12.858447074890137, 'timing/data': 0.028044655919075012, 'timing/model': 1.441145516000688, 'learning_rate/qwen_vl_interface': 9.908413370298131e-06, 'learning_rate/action_model': 9.90841337029813e-05, 'epoch': 0.19}) +[2026-06-17 17:14:12,593][starVLA.training.train_starvla][INFO] - Step 533, Loss: {'train/loss': 0.9570356607437134, 'train/grad_norm_pre_clip': 10.382996559143066, 'timing/data': 0.022367309778928757, 'timing/model': 1.3653250318020582, 'learning_rate/qwen_vl_interface': 9.90799024316589e-06, 'learning_rate/action_model': 9.90799024316589e-05, 'epoch': 0.2}) +[2026-06-17 17:14:16,682][starVLA.training.train_starvla][INFO] - Step 534, Loss: {'train/loss': 0.8980559706687927, 'train/grad_norm_pre_clip': 10.323969841003418, 'timing/data': 0.01961064152419567, 'timing/model': 1.0271295169368386, 'learning_rate/qwen_vl_interface': 9.90756615042726e-06, 'learning_rate/action_model': 9.907566150427259e-05, 'epoch': 0.2}) +[2026-06-17 17:14:22,075][starVLA.training.train_starvla][INFO] - Step 535, Loss: {'train/loss': 1.224387288093567, 'train/grad_norm_pre_clip': 9.094449996948242, 'timing/data': 0.021650027483701706, 'timing/model': 1.1077172653749585, 'learning_rate/qwen_vl_interface': 9.907141092170154e-06, 'learning_rate/action_model': 9.907141092170153e-05, 'epoch': 0.2}) +[2026-06-17 17:14:30,082][starVLA.training.train_starvla][INFO] - Step 536, Loss: {'train/loss': 0.9491677284240723, 'train/grad_norm_pre_clip': 8.234261512756348, 'timing/data': 0.026915067806839943, 'timing/model': 1.1860776664689183, 'learning_rate/qwen_vl_interface': 9.906715068482689e-06, 'learning_rate/action_model': 9.90671506848269e-05, 'epoch': 0.2}) +[2026-06-17 17:14:35,623][starVLA.training.train_starvla][INFO] - Step 537, Loss: {'train/loss': 1.0537803173065186, 'train/grad_norm_pre_clip': 34.1346321105957, 'timing/data': 0.02370862103998661, 'timing/model': 1.0590276755392551, 'learning_rate/qwen_vl_interface': 9.90628807945318e-06, 'learning_rate/action_model': 9.90628807945318e-05, 'epoch': 0.2}) +[2026-06-17 17:14:46,283][starVLA.training.train_starvla][INFO] - Step 538, Loss: {'train/loss': 1.1165485382080078, 'train/grad_norm_pre_clip': 8.428092002868652, 'timing/data': 0.021642770618200302, 'timing/model': 2.8936454635113478, 'learning_rate/qwen_vl_interface': 9.90586012517014e-06, 'learning_rate/action_model': 9.90586012517014e-05, 'epoch': 0.2}) +[2026-06-17 17:14:56,726][starVLA.training.train_starvla][INFO] - Step 539, Loss: {'train/loss': 1.2316558361053467, 'train/grad_norm_pre_clip': 9.8134183883667, 'timing/data': 0.02383127622306347, 'timing/model': 4.534325499087572, 'learning_rate/qwen_vl_interface': 9.905431205722288e-06, 'learning_rate/action_model': 9.905431205722288e-05, 'epoch': 0.2}) +[2026-06-17 17:15:01,305][starVLA.training.train_starvla][INFO] - Step 540, Loss: {'train/loss': 1.0649840831756592, 'train/grad_norm_pre_clip': 12.529535293579102, 'timing/data': 0.01989759784191847, 'timing/model': 1.1053481241688132, 'learning_rate/qwen_vl_interface': 9.905001321198536e-06, 'learning_rate/action_model': 9.905001321198536e-05, 'epoch': 0.2}) +[2026-06-17 17:15:12,309][starVLA.training.train_starvla][INFO] - Step 541, Loss: {'train/loss': 1.0379951000213623, 'train/grad_norm_pre_clip': 9.31039047241211, 'timing/data': 0.020249685272574425, 'timing/model': 1.1404264299198985, 'learning_rate/qwen_vl_interface': 9.904570471688003e-06, 'learning_rate/action_model': 9.904570471688002e-05, 'epoch': 0.2}) +[2026-06-17 17:15:23,423][starVLA.training.train_starvla][INFO] - Step 542, Loss: {'train/loss': 1.0083496570587158, 'train/grad_norm_pre_clip': 11.268582344055176, 'timing/data': 0.02455669641494751, 'timing/model': 3.4308904698118567, 'learning_rate/qwen_vl_interface': 9.90413865728e-06, 'learning_rate/action_model': 9.90413865728e-05, 'epoch': 0.2}) +[2026-06-17 17:15:35,362][starVLA.training.train_starvla][INFO] - Step 543, Loss: {'train/loss': 0.8895502686500549, 'train/grad_norm_pre_clip': 17.548839569091797, 'timing/data': 0.03489668574184179, 'timing/model': 1.8062497014179826, 'learning_rate/qwen_vl_interface': 9.903705878064047e-06, 'learning_rate/action_model': 9.903705878064046e-05, 'epoch': 0.2}) +[2026-06-17 17:15:39,688][starVLA.training.train_starvla][INFO] - Step 544, Loss: {'train/loss': 1.1641789674758911, 'train/grad_norm_pre_clip': 9.794082641601562, 'timing/data': 0.018693728372454643, 'timing/model': 1.0483096558600664, 'learning_rate/qwen_vl_interface': 9.903272134129858e-06, 'learning_rate/action_model': 9.903272134129858e-05, 'epoch': 0.2}) +[2026-06-17 17:15:44,310][starVLA.training.train_starvla][INFO] - Step 545, Loss: {'train/loss': 1.4372159242630005, 'train/grad_norm_pre_clip': 11.468127250671387, 'timing/data': 0.02326066605746746, 'timing/model': 1.1407560678198934, 'learning_rate/qwen_vl_interface': 9.90283742556735e-06, 'learning_rate/action_model': 9.902837425567349e-05, 'epoch': 0.2}) +[2026-06-17 17:15:48,937][starVLA.training.train_starvla][INFO] - Step 546, Loss: {'train/loss': 1.3305450677871704, 'train/grad_norm_pre_clip': 9.970451354980469, 'timing/data': 0.028154452331364155, 'timing/model': 1.091037523932755, 'learning_rate/qwen_vl_interface': 9.902401752466635e-06, 'learning_rate/action_model': 9.902401752466634e-05, 'epoch': 0.2}) +[2026-06-17 17:15:53,369][starVLA.training.train_starvla][INFO] - Step 547, Loss: {'train/loss': 1.0153709650039673, 'train/grad_norm_pre_clip': 8.676091194152832, 'timing/data': 0.018464943394064903, 'timing/model': 1.0240781269967556, 'learning_rate/qwen_vl_interface': 9.90196511491803e-06, 'learning_rate/action_model': 9.901965114918029e-05, 'epoch': 0.2}) +[2026-06-17 17:15:57,985][starVLA.training.train_starvla][INFO] - Step 548, Loss: {'train/loss': 1.0075346231460571, 'train/grad_norm_pre_clip': 7.707160949707031, 'timing/data': 0.022748091258108616, 'timing/model': 1.3719006441533566, 'learning_rate/qwen_vl_interface': 9.901527513012053e-06, 'learning_rate/action_model': 9.901527513012052e-05, 'epoch': 0.2}) +[2026-06-17 17:16:02,263][starVLA.training.train_starvla][INFO] - Step 549, Loss: {'train/loss': 0.8213682770729065, 'train/grad_norm_pre_clip': 9.886584281921387, 'timing/data': 0.024839614517986774, 'timing/model': 1.0420287484303117, 'learning_rate/qwen_vl_interface': 9.901088946839415e-06, 'learning_rate/action_model': 9.901088946839414e-05, 'epoch': 0.2}) +[2026-06-17 17:16:07,030][starVLA.training.train_starvla][INFO] - Step 550, Loss: {'train/loss': 1.0584920644760132, 'train/grad_norm_pre_clip': 10.38399887084961, 'timing/data': 0.03341889660805464, 'timing/model': 1.2655152399092913, 'learning_rate/qwen_vl_interface': 9.900649416491035e-06, 'learning_rate/action_model': 9.900649416491035e-05, 'epoch': 0.2}) +[2026-06-17 17:16:11,607][starVLA.training.train_starvla][INFO] - Step 551, Loss: {'train/loss': 0.8689774870872498, 'train/grad_norm_pre_clip': 7.602765083312988, 'timing/data': 0.019604426808655262, 'timing/model': 1.1122755771502852, 'learning_rate/qwen_vl_interface': 9.900208922058026e-06, 'learning_rate/action_model': 9.900208922058025e-05, 'epoch': 0.2}) +[2026-06-17 17:16:19,221][starVLA.training.train_starvla][INFO] - Step 552, Loss: {'train/loss': 1.037548542022705, 'train/grad_norm_pre_clip': 11.767572402954102, 'timing/data': 0.024410570040345192, 'timing/model': 4.590326016768813, 'learning_rate/qwen_vl_interface': 9.899767463631704e-06, 'learning_rate/action_model': 9.899767463631704e-05, 'epoch': 0.2}) +[2026-06-17 17:16:23,561][starVLA.training.train_starvla][INFO] - Step 553, Loss: {'train/loss': 0.846267580986023, 'train/grad_norm_pre_clip': 13.90396499633789, 'timing/data': 0.023264405317604542, 'timing/model': 1.3102715434506536, 'learning_rate/qwen_vl_interface': 9.899325041303582e-06, 'learning_rate/action_model': 9.899325041303583e-05, 'epoch': 0.2}) +[2026-06-17 17:16:27,702][starVLA.training.train_starvla][INFO] - Step 554, Loss: {'train/loss': 1.1251978874206543, 'train/grad_norm_pre_clip': 12.154162406921387, 'timing/data': 0.021603473462164402, 'timing/model': 1.0411515310406685, 'learning_rate/qwen_vl_interface': 9.89888165516538e-06, 'learning_rate/action_model': 9.898881655165379e-05, 'epoch': 0.2}) +[2026-06-17 17:16:32,197][starVLA.training.train_starvla][INFO] - Step 555, Loss: {'train/loss': 1.2164746522903442, 'train/grad_norm_pre_clip': 12.302247047424316, 'timing/data': 0.021097838878631592, 'timing/model': 1.068021671846509, 'learning_rate/qwen_vl_interface': 9.898437305309004e-06, 'learning_rate/action_model': 9.898437305309003e-05, 'epoch': 0.2}) +[2026-06-17 17:16:36,838][starVLA.training.train_starvla][INFO] - Step 556, Loss: {'train/loss': 0.8394044637680054, 'train/grad_norm_pre_clip': 9.336030006408691, 'timing/data': 0.03759469464421272, 'timing/model': 1.0970881627872586, 'learning_rate/qwen_vl_interface': 9.897991991826575e-06, 'learning_rate/action_model': 9.897991991826574e-05, 'epoch': 0.2}) +[2026-06-17 17:16:42,518][starVLA.training.train_starvla][INFO] - Step 557, Loss: {'train/loss': 0.963670015335083, 'train/grad_norm_pre_clip': 8.842543601989746, 'timing/data': 0.024864613078534603, 'timing/model': 1.4132974427193403, 'learning_rate/qwen_vl_interface': 9.897545714810404e-06, 'learning_rate/action_model': 9.897545714810403e-05, 'epoch': 0.2}) +[2026-06-17 17:16:47,219][starVLA.training.train_starvla][INFO] - Step 558, Loss: {'train/loss': 1.3250983953475952, 'train/grad_norm_pre_clip': 9.969608306884766, 'timing/data': 0.02384754829108715, 'timing/model': 1.4058708641678095, 'learning_rate/qwen_vl_interface': 9.897098474353004e-06, 'learning_rate/action_model': 9.897098474353004e-05, 'epoch': 0.2}) +[2026-06-17 17:16:51,518][starVLA.training.train_starvla][INFO] - Step 559, Loss: {'train/loss': 0.8775867819786072, 'train/grad_norm_pre_clip': 10.895586013793945, 'timing/data': 0.019802317023277283, 'timing/model': 1.0610778219997883, 'learning_rate/qwen_vl_interface': 9.896650270547092e-06, 'learning_rate/action_model': 9.89665027054709e-05, 'epoch': 0.2}) +[2026-06-17 17:16:55,989][starVLA.training.train_starvla][INFO] - Step 560, Loss: {'train/loss': 0.9214940071105957, 'train/grad_norm_pre_clip': 11.221832275390625, 'timing/data': 0.018845507875084877, 'timing/model': 1.0349567970260978, 'learning_rate/qwen_vl_interface': 9.896201103485577e-06, 'learning_rate/action_model': 9.896201103485577e-05, 'epoch': 0.21}) +[2026-06-17 17:17:03,855][starVLA.training.train_starvla][INFO] - Step 561, Loss: {'train/loss': 1.1401275396347046, 'train/grad_norm_pre_clip': 9.840187072753906, 'timing/data': 0.021887837909162045, 'timing/model': 1.0727938497439027, 'learning_rate/qwen_vl_interface': 9.895750973261576e-06, 'learning_rate/action_model': 9.895750973261574e-05, 'epoch': 0.21}) +[2026-06-17 17:17:08,489][starVLA.training.train_starvla][INFO] - Step 562, Loss: {'train/loss': 0.9439969658851624, 'train/grad_norm_pre_clip': 9.810684204101562, 'timing/data': 0.0297697139903903, 'timing/model': 1.5764915058389306, 'learning_rate/qwen_vl_interface': 9.895299879968398e-06, 'learning_rate/action_model': 9.895299879968398e-05, 'epoch': 0.21}) +[2026-06-17 17:17:12,905][starVLA.training.train_starvla][INFO] - Step 563, Loss: {'train/loss': 1.0086562633514404, 'train/grad_norm_pre_clip': 12.728636741638184, 'timing/data': 0.02010186668485403, 'timing/model': 1.359375306405127, 'learning_rate/qwen_vl_interface': 9.894847823699555e-06, 'learning_rate/action_model': 9.894847823699555e-05, 'epoch': 0.21}) +[2026-06-17 17:17:17,203][starVLA.training.train_starvla][INFO] - Step 564, Loss: {'train/loss': 0.9566246271133423, 'train/grad_norm_pre_clip': 11.554666519165039, 'timing/data': 0.024281674064695835, 'timing/model': 1.2235744018107653, 'learning_rate/qwen_vl_interface': 9.894394804548763e-06, 'learning_rate/action_model': 9.894394804548762e-05, 'epoch': 0.21}) +[2026-06-17 17:17:21,657][starVLA.training.train_starvla][INFO] - Step 565, Loss: {'train/loss': 0.8766858577728271, 'train/grad_norm_pre_clip': 10.0487060546875, 'timing/data': 0.021897506900131702, 'timing/model': 1.0381867978721857, 'learning_rate/qwen_vl_interface': 9.893940822609928e-06, 'learning_rate/action_model': 9.893940822609927e-05, 'epoch': 0.21}) +[2026-06-17 17:17:26,810][starVLA.training.train_starvla][INFO] - Step 566, Loss: {'train/loss': 1.0530989170074463, 'train/grad_norm_pre_clip': 10.388522148132324, 'timing/data': 0.02510916069149971, 'timing/model': 1.437673700042069, 'learning_rate/qwen_vl_interface': 9.893485877977163e-06, 'learning_rate/action_model': 9.893485877977163e-05, 'epoch': 0.21}) +[2026-06-17 17:17:33,019][starVLA.training.train_starvla][INFO] - Step 567, Loss: {'train/loss': 1.0340087413787842, 'train/grad_norm_pre_clip': 13.49599552154541, 'timing/data': 0.020746523514389992, 'timing/model': 2.5493333488702774, 'learning_rate/qwen_vl_interface': 9.89302997074478e-06, 'learning_rate/action_model': 9.89302997074478e-05, 'epoch': 0.21}) +[2026-06-17 17:17:37,710][starVLA.training.train_starvla][INFO] - Step 568, Loss: {'train/loss': 1.309747576713562, 'train/grad_norm_pre_clip': 11.665705680847168, 'timing/data': 0.02308914251625538, 'timing/model': 1.4067572886124253, 'learning_rate/qwen_vl_interface': 9.892573101007288e-06, 'learning_rate/action_model': 9.892573101007289e-05, 'epoch': 0.21}) +[2026-06-17 17:17:41,733][starVLA.training.train_starvla][INFO] - Step 569, Loss: {'train/loss': 1.1659619808197021, 'train/grad_norm_pre_clip': 19.658870697021484, 'timing/data': 0.02293007541447878, 'timing/model': 1.0249086804687977, 'learning_rate/qwen_vl_interface': 9.892115268859398e-06, 'learning_rate/action_model': 9.892115268859397e-05, 'epoch': 0.21}) +[2026-06-17 17:17:46,272][starVLA.training.train_starvla][INFO] - Step 570, Loss: {'train/loss': 1.257131576538086, 'train/grad_norm_pre_clip': 11.106587409973145, 'timing/data': 0.019462023861706257, 'timing/model': 1.0373595785349607, 'learning_rate/qwen_vl_interface': 9.891656474396015e-06, 'learning_rate/action_model': 9.891656474396015e-05, 'epoch': 0.21}) +[2026-06-17 17:17:50,680][starVLA.training.train_starvla][INFO] - Step 571, Loss: {'train/loss': 1.0584995746612549, 'train/grad_norm_pre_clip': 14.913363456726074, 'timing/data': 0.018110256642103195, 'timing/model': 1.0209474815055728, 'learning_rate/qwen_vl_interface': 9.891196717712253e-06, 'learning_rate/action_model': 9.891196717712252e-05, 'epoch': 0.21}) +[2026-06-17 17:17:55,345][starVLA.training.train_starvla][INFO] - Step 572, Loss: {'train/loss': 1.1953848600387573, 'train/grad_norm_pre_clip': 12.164716720581055, 'timing/data': 0.020565366372466087, 'timing/model': 1.4334209617227316, 'learning_rate/qwen_vl_interface': 9.890735998903414e-06, 'learning_rate/action_model': 9.890735998903414e-05, 'epoch': 0.21}) +[2026-06-17 17:18:00,517][starVLA.training.train_starvla][INFO] - Step 573, Loss: {'train/loss': 1.0657163858413696, 'train/grad_norm_pre_clip': 9.401843070983887, 'timing/data': 0.019900920800864697, 'timing/model': 1.627137959934771, 'learning_rate/qwen_vl_interface': 9.890274318065012e-06, 'learning_rate/action_model': 9.890274318065011e-05, 'epoch': 0.21}) +[2026-06-17 17:18:04,679][starVLA.training.train_starvla][INFO] - Step 574, Loss: {'train/loss': 1.1712583303451538, 'train/grad_norm_pre_clip': 12.517924308776855, 'timing/data': 0.024114572443068027, 'timing/model': 1.0294202668592334, 'learning_rate/qwen_vl_interface': 9.889811675292748e-06, 'learning_rate/action_model': 9.889811675292749e-05, 'epoch': 0.21}) +[2026-06-17 17:18:09,362][starVLA.training.train_starvla][INFO] - Step 575, Loss: {'train/loss': 1.3236907720565796, 'train/grad_norm_pre_clip': 13.775891304016113, 'timing/data': 0.01812317781150341, 'timing/model': 1.1880441922694445, 'learning_rate/qwen_vl_interface': 9.889348070682533e-06, 'learning_rate/action_model': 9.889348070682533e-05, 'epoch': 0.21}) +[2026-06-17 17:18:20,265][starVLA.training.train_starvla][INFO] - Step 576, Loss: {'train/loss': 0.9401376247406006, 'train/grad_norm_pre_clip': 11.309714317321777, 'timing/data': 0.10039473790675402, 'timing/model': 1.049333117902279, 'learning_rate/qwen_vl_interface': 9.888883504330469e-06, 'learning_rate/action_model': 9.888883504330468e-05, 'epoch': 0.21}) +[2026-06-17 17:18:25,509][starVLA.training.train_starvla][INFO] - Step 577, Loss: {'train/loss': 1.1489975452423096, 'train/grad_norm_pre_clip': 8.732254028320312, 'timing/data': 0.03155475668609142, 'timing/model': 1.0638624588027596, 'learning_rate/qwen_vl_interface': 9.888417976332865e-06, 'learning_rate/action_model': 9.888417976332865e-05, 'epoch': 0.21}) +[2026-06-17 17:18:30,725][starVLA.training.train_starvla][INFO] - Step 578, Loss: {'train/loss': 1.0886621475219727, 'train/grad_norm_pre_clip': 8.865368843078613, 'timing/data': 0.022526733577251434, 'timing/model': 1.4696520809084177, 'learning_rate/qwen_vl_interface': 9.887951486786222e-06, 'learning_rate/action_model': 9.887951486786221e-05, 'epoch': 0.21}) +[2026-06-17 17:18:36,056][starVLA.training.train_starvla][INFO] - Step 579, Loss: {'train/loss': 1.2973512411117554, 'train/grad_norm_pre_clip': 9.732525825500488, 'timing/data': 0.023961182683706284, 'timing/model': 1.1838331324979663, 'learning_rate/qwen_vl_interface': 9.887484035787246e-06, 'learning_rate/action_model': 9.887484035787246e-05, 'epoch': 0.21}) +[2026-06-17 17:18:40,531][starVLA.training.train_starvla][INFO] - Step 580, Loss: {'train/loss': 0.7253910303115845, 'train/grad_norm_pre_clip': 10.77261734008789, 'timing/data': 0.029807575047016144, 'timing/model': 1.0793449943885207, 'learning_rate/qwen_vl_interface': 9.887015623432837e-06, 'learning_rate/action_model': 9.887015623432837e-05, 'epoch': 0.21}) +[2026-06-17 17:18:46,553][starVLA.training.train_starvla][INFO] - Step 581, Loss: {'train/loss': 0.738353967666626, 'train/grad_norm_pre_clip': 7.324125289916992, 'timing/data': 0.019301168620586395, 'timing/model': 1.0501197259873152, 'learning_rate/qwen_vl_interface': 9.886546249820101e-06, 'learning_rate/action_model': 9.886546249820102e-05, 'epoch': 0.21}) +[2026-06-17 17:18:52,624][starVLA.training.train_starvla][INFO] - Step 582, Loss: {'train/loss': 0.7248641848564148, 'train/grad_norm_pre_clip': 7.536610126495361, 'timing/data': 0.02080798614770174, 'timing/model': 1.5592650435864925, 'learning_rate/qwen_vl_interface': 9.886075915046339e-06, 'learning_rate/action_model': 9.886075915046338e-05, 'epoch': 0.21}) +[2026-06-17 17:18:57,064][starVLA.training.train_starvla][INFO] - Step 583, Loss: {'train/loss': 0.6303195357322693, 'train/grad_norm_pre_clip': 7.790208339691162, 'timing/data': 0.02198007609695196, 'timing/model': 1.3463938469067216, 'learning_rate/qwen_vl_interface': 9.885604619209052e-06, 'learning_rate/action_model': 9.885604619209051e-05, 'epoch': 0.21}) +[2026-06-17 17:19:01,632][starVLA.training.train_starvla][INFO] - Step 584, Loss: {'train/loss': 0.9427518248558044, 'train/grad_norm_pre_clip': 8.754922866821289, 'timing/data': 0.026024981401860714, 'timing/model': 1.053083992563188, 'learning_rate/qwen_vl_interface': 9.885132362405935e-06, 'learning_rate/action_model': 9.885132362405936e-05, 'epoch': 0.21}) +[2026-06-17 17:19:06,128][starVLA.training.train_starvla][INFO] - Step 585, Loss: {'train/loss': 1.1585538387298584, 'train/grad_norm_pre_clip': 11.319167137145996, 'timing/data': 0.033753846772015095, 'timing/model': 1.0379396872594953, 'learning_rate/qwen_vl_interface': 9.884659144734896e-06, 'learning_rate/action_model': 9.884659144734896e-05, 'epoch': 0.21}) +[2026-06-17 17:19:13,860][starVLA.training.train_starvla][INFO] - Step 586, Loss: {'train/loss': 1.3490983247756958, 'train/grad_norm_pre_clip': 13.600818634033203, 'timing/data': 0.02098015695810318, 'timing/model': 1.0916940961033106, 'learning_rate/qwen_vl_interface': 9.884184966294028e-06, 'learning_rate/action_model': 9.884184966294028e-05, 'epoch': 0.21}) +[2026-06-17 17:19:20,007][starVLA.training.train_starvla][INFO] - Step 587, Loss: {'train/loss': 0.9512361288070679, 'train/grad_norm_pre_clip': 11.895182609558105, 'timing/data': 0.020679854787886143, 'timing/model': 1.1262607173994184, 'learning_rate/qwen_vl_interface': 9.883709827181632e-06, 'learning_rate/action_model': 9.88370982718163e-05, 'epoch': 0.21}) +[2026-06-17 17:19:24,604][starVLA.training.train_starvla][INFO] - Step 588, Loss: {'train/loss': 1.2455661296844482, 'train/grad_norm_pre_clip': 12.812755584716797, 'timing/data': 0.01992864068597555, 'timing/model': 1.44645672198385, 'learning_rate/qwen_vl_interface': 9.8832337274962e-06, 'learning_rate/action_model': 9.883233727496198e-05, 'epoch': 0.22}) +[2026-06-17 17:19:28,765][starVLA.training.train_starvla][INFO] - Step 589, Loss: {'train/loss': 0.9474503993988037, 'train/grad_norm_pre_clip': 14.727198600769043, 'timing/data': 0.020023740828037262, 'timing/model': 1.05244982521981, 'learning_rate/qwen_vl_interface': 9.88275666733643e-06, 'learning_rate/action_model': 9.882756667336429e-05, 'epoch': 0.22}) +[2026-06-17 17:19:33,164][starVLA.training.train_starvla][INFO] - Step 590, Loss: {'train/loss': 1.0605688095092773, 'train/grad_norm_pre_clip': 10.051602363586426, 'timing/data': 0.022036375477910042, 'timing/model': 1.0494744488969445, 'learning_rate/qwen_vl_interface': 9.88227864680122e-06, 'learning_rate/action_model': 9.88227864680122e-05, 'epoch': 0.22}) +[2026-06-17 17:19:37,466][starVLA.training.train_starvla][INFO] - Step 591, Loss: {'train/loss': 1.1922292709350586, 'train/grad_norm_pre_clip': 20.899654388427734, 'timing/data': 0.01938180159777403, 'timing/model': 1.0225733779370785, 'learning_rate/qwen_vl_interface': 9.88179966598966e-06, 'learning_rate/action_model': 9.88179966598966e-05, 'epoch': 0.22}) +[2026-06-17 17:19:41,816][starVLA.training.train_starvla][INFO] - Step 592, Loss: {'train/loss': 1.0969717502593994, 'train/grad_norm_pre_clip': 8.772209167480469, 'timing/data': 0.01979004219174385, 'timing/model': 1.3469482185319066, 'learning_rate/qwen_vl_interface': 9.881319725001046e-06, 'learning_rate/action_model': 9.881319725001046e-05, 'epoch': 0.22}) +[2026-06-17 17:19:46,160][starVLA.training.train_starvla][INFO] - Step 593, Loss: {'train/loss': 0.9426378607749939, 'train/grad_norm_pre_clip': 9.59130859375, 'timing/data': 0.019013497978448868, 'timing/model': 1.367925920523703, 'learning_rate/qwen_vl_interface': 9.88083882393487e-06, 'learning_rate/action_model': 9.880838823934869e-05, 'epoch': 0.22}) +[2026-06-17 17:19:50,176][starVLA.training.train_starvla][INFO] - Step 594, Loss: {'train/loss': 1.0481029748916626, 'train/grad_norm_pre_clip': 9.03452205657959, 'timing/data': 0.01891427394002676, 'timing/model': 1.0314593827351928, 'learning_rate/qwen_vl_interface': 9.880356962890823e-06, 'learning_rate/action_model': 9.880356962890822e-05, 'epoch': 0.22}) +[2026-06-17 17:19:54,555][starVLA.training.train_starvla][INFO] - Step 595, Loss: {'train/loss': 1.400119662284851, 'train/grad_norm_pre_clip': 13.543107986450195, 'timing/data': 0.021009018644690514, 'timing/model': 1.0198467709124088, 'learning_rate/qwen_vl_interface': 9.879874141968792e-06, 'learning_rate/action_model': 9.879874141968792e-05, 'epoch': 0.22}) +[2026-06-17 17:19:59,249][starVLA.training.train_starvla][INFO] - Step 596, Loss: {'train/loss': 1.147786021232605, 'train/grad_norm_pre_clip': 11.169400215148926, 'timing/data': 0.02187153697013855, 'timing/model': 1.0450297882780433, 'learning_rate/qwen_vl_interface': 9.879390361268873e-06, 'learning_rate/action_model': 9.879390361268873e-05, 'epoch': 0.22}) +[2026-06-17 17:20:03,653][starVLA.training.train_starvla][INFO] - Step 597, Loss: {'train/loss': 0.8801038265228271, 'train/grad_norm_pre_clip': 9.188981056213379, 'timing/data': 0.018217459321022034, 'timing/model': 1.0698822243139148, 'learning_rate/qwen_vl_interface': 9.878905620891349e-06, 'learning_rate/action_model': 9.878905620891349e-05, 'epoch': 0.22}) +[2026-06-17 17:20:08,102][starVLA.training.train_starvla][INFO] - Step 598, Loss: {'train/loss': 0.8874760270118713, 'train/grad_norm_pre_clip': 9.102721214294434, 'timing/data': 0.018454699777066708, 'timing/model': 1.3578014662489295, 'learning_rate/qwen_vl_interface': 9.87841992093671e-06, 'learning_rate/action_model': 9.87841992093671e-05, 'epoch': 0.22}) +[2026-06-17 17:20:12,307][starVLA.training.train_starvla][INFO] - Step 599, Loss: {'train/loss': 0.9149947762489319, 'train/grad_norm_pre_clip': 10.681318283081055, 'timing/data': 0.02239503152668476, 'timing/model': 1.0924043785780668, 'learning_rate/qwen_vl_interface': 9.877933261505642e-06, 'learning_rate/action_model': 9.877933261505641e-05, 'epoch': 0.22}) +[2026-06-17 17:24:02,107][starVLA.training.train_starvla][INFO] - Step 600, Loss: {'train/loss': 0.9365050196647644, 'train/grad_norm_pre_clip': 8.719137191772461, 'eval/loss': 1.0355645418167114, 'eval/latency_2/loss': 1.0355645418167114, 'eval/demon_fire_precision': 0.8151439744045503, 'eval/demon_fire_recall': 0.9594142259414226, 'eval/demon_fire_f1': 0.8814145685181625, 'eval/demon_fire_rate_model': 0.7826933778519756, 'eval/demon_fire_rate_teacher': 0.6649972175848636, 'eval/demon_left_precision': 0.6613616268788682, 'eval/demon_left_recall': 0.7873684210526316, 'eval/demon_left_f1': 0.7188851513695338, 'eval/demon_left_rate_model': 0.31469115191986646, 'eval/demon_left_rate_teacher': 0.2643294379521425, 'eval/demon_right_precision': 0.8155038759689922, 'eval/demon_right_recall': 0.5943502824858757, 'eval/demon_right_f1': 0.6875816993464052, 'eval/demon_right_rate_model': 0.17946577629382304, 'eval/demon_right_rate_teacher': 0.24624373956594323, 'eval/demon_move_f1': 0.7032334253579695, 'eval/demon_cc_f1': 0.8191492080902278, 'eval/latency_2/demon_fire_precision': 0.8151439744045503, 'eval/latency_2/demon_fire_recall': 0.9594142259414226, 'eval/latency_2/demon_fire_f1': 0.8814145685181625, 'eval/latency_2/demon_fire_rate_model': 0.7826933778519756, 'eval/latency_2/demon_fire_rate_teacher': 0.6649972175848636, 'eval/latency_2/demon_left_precision': 0.6613616268788682, 'eval/latency_2/demon_left_recall': 0.7873684210526316, 'eval/latency_2/demon_left_f1': 0.7188851513695338, 'eval/latency_2/demon_left_rate_model': 0.31469115191986646, 'eval/latency_2/demon_left_rate_teacher': 0.2643294379521425, 'eval/latency_2/demon_right_precision': 0.8155038759689922, 'eval/latency_2/demon_right_recall': 0.5943502824858757, 'eval/latency_2/demon_right_f1': 0.6875816993464052, 'eval/latency_2/demon_right_rate_model': 0.17946577629382304, 'eval/latency_2/demon_right_rate_teacher': 0.24624373956594323, 'eval/latency_2/demon_move_f1': 0.7032334253579695, 'eval/latency_2/demon_cc_f1': 0.8191492080902278, 'timing/data': 0.02462458424270153, 'timing/model': 1.0719320103526115, 'learning_rate/qwen_vl_interface': 9.877445642699026e-06, 'learning_rate/action_model': 9.877445642699026e-05, 'epoch': 0.22}) +[2026-06-17 17:24:02,121][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_600_state +[2026-06-17 17:24:02,129][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 17:24:11,675][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_600_state/model.safetensors +[2026-06-17 17:24:25,765][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_600_state/optimizer.bin +[2026-06-17 17:24:25,767][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_600_state/sampler.bin +[2026-06-17 17:24:25,768][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_600_state/sampler_1.bin +[2026-06-17 17:24:25,770][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_600_state/random_states_0.pkl +[2026-06-17 17:24:25,775][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 600 +[2026-06-17 17:24:25,776][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 17:24:25,793][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 17:24:30,317][starVLA.training.train_starvla][INFO] - Step 601, Loss: {'train/loss': 0.8887234330177307, 'train/grad_norm_pre_clip': 10.797003746032715, 'timing/data': 0.022091664373874664, 'timing/model': 1.0418251501396298, 'learning_rate/qwen_vl_interface': 9.876957064617951e-06, 'learning_rate/action_model': 9.876957064617951e-05, 'epoch': 0.22}) +[2026-06-17 17:24:40,471][starVLA.training.train_starvla][INFO] - Step 602, Loss: {'train/loss': 1.0254385471343994, 'train/grad_norm_pre_clip': 17.0272216796875, 'timing/data': 0.03166564926505089, 'timing/model': 2.2062123231589794, 'learning_rate/qwen_vl_interface': 9.876467527363696e-06, 'learning_rate/action_model': 9.876467527363697e-05, 'epoch': 0.22}) +[2026-06-17 17:24:49,625][starVLA.training.train_starvla][INFO] - Step 603, Loss: {'train/loss': 1.100206971168518, 'train/grad_norm_pre_clip': 10.210905075073242, 'timing/data': 0.028970579616725445, 'timing/model': 5.653771985322237, 'learning_rate/qwen_vl_interface': 9.875977031037746e-06, 'learning_rate/action_model': 9.875977031037745e-05, 'epoch': 0.22}) +[2026-06-17 17:25:00,799][starVLA.training.train_starvla][INFO] - Step 604, Loss: {'train/loss': 0.9143043756484985, 'train/grad_norm_pre_clip': 10.14862060546875, 'timing/data': 0.021268202923238277, 'timing/model': 1.4679215466603637, 'learning_rate/qwen_vl_interface': 9.875485575741779e-06, 'learning_rate/action_model': 9.875485575741779e-05, 'epoch': 0.22}) +[2026-06-17 17:25:05,229][starVLA.training.train_starvla][INFO] - Step 605, Loss: {'train/loss': 1.1275103092193604, 'train/grad_norm_pre_clip': 9.368342399597168, 'timing/data': 0.024571510963141918, 'timing/model': 1.2210385529324412, 'learning_rate/qwen_vl_interface': 9.874993161577674e-06, 'learning_rate/action_model': 9.874993161577675e-05, 'epoch': 0.22}) +[2026-06-17 17:25:09,902][starVLA.training.train_starvla][INFO] - Step 606, Loss: {'train/loss': 1.1983113288879395, 'train/grad_norm_pre_clip': 9.8831205368042, 'timing/data': 0.024667659774422646, 'timing/model': 1.0569874159991741, 'learning_rate/qwen_vl_interface': 9.874499788647512e-06, 'learning_rate/action_model': 9.874499788647513e-05, 'epoch': 0.22}) +[2026-06-17 17:25:19,207][starVLA.training.train_starvla][INFO] - Step 607, Loss: {'train/loss': 1.029252529144287, 'train/grad_norm_pre_clip': 10.625373840332031, 'timing/data': 0.028713406063616276, 'timing/model': 1.17527870927006, 'learning_rate/qwen_vl_interface': 9.874005457053565e-06, 'learning_rate/action_model': 9.874005457053565e-05, 'epoch': 0.22}) +[2026-06-17 17:25:27,116][starVLA.training.train_starvla][INFO] - Step 608, Loss: {'train/loss': 1.3544782400131226, 'train/grad_norm_pre_clip': 13.719735145568848, 'timing/data': 0.01849720161408186, 'timing/model': 1.0596987958997488, 'learning_rate/qwen_vl_interface': 9.873510166898315e-06, 'learning_rate/action_model': 9.873510166898314e-05, 'epoch': 0.22}) +[2026-06-17 17:25:31,772][starVLA.training.train_starvla][INFO] - Step 609, Loss: {'train/loss': 1.2667081356048584, 'train/grad_norm_pre_clip': 9.974019050598145, 'timing/data': 0.02049951907247305, 'timing/model': 1.029587241820991, 'learning_rate/qwen_vl_interface': 9.87301391828443e-06, 'learning_rate/action_model': 9.873013918284428e-05, 'epoch': 0.22}) +[2026-06-17 17:25:37,924][starVLA.training.train_starvla][INFO] - Step 610, Loss: {'train/loss': 1.115404486656189, 'train/grad_norm_pre_clip': 10.834269523620605, 'timing/data': 0.07137695420533419, 'timing/model': 1.3904441120103002, 'learning_rate/qwen_vl_interface': 9.872516711314784e-06, 'learning_rate/action_model': 9.872516711314784e-05, 'epoch': 0.22}) +[2026-06-17 17:25:42,145][starVLA.training.train_starvla][INFO] - Step 611, Loss: {'train/loss': 1.006632924079895, 'train/grad_norm_pre_clip': 11.164201736450195, 'timing/data': 0.02160400804132223, 'timing/model': 1.0456900503486395, 'learning_rate/qwen_vl_interface': 9.872018546092452e-06, 'learning_rate/action_model': 9.872018546092451e-05, 'epoch': 0.22}) +[2026-06-17 17:25:46,610][starVLA.training.train_starvla][INFO] - Step 612, Loss: {'train/loss': 1.01473069190979, 'train/grad_norm_pre_clip': 8.765393257141113, 'timing/data': 0.021920660510659218, 'timing/model': 1.0673175314441323, 'learning_rate/qwen_vl_interface': 9.871519422720699e-06, 'learning_rate/action_model': 9.871519422720698e-05, 'epoch': 0.22}) +[2026-06-17 17:25:51,226][starVLA.training.train_starvla][INFO] - Step 613, Loss: {'train/loss': 0.9541326761245728, 'train/grad_norm_pre_clip': 8.285472869873047, 'timing/data': 0.02017376944422722, 'timing/model': 1.055969025939703, 'learning_rate/qwen_vl_interface': 9.871019341303e-06, 'learning_rate/action_model': 9.871019341303e-05, 'epoch': 0.22}) +[2026-06-17 17:25:55,656][starVLA.training.train_starvla][INFO] - Step 614, Loss: {'train/loss': 0.9570810794830322, 'train/grad_norm_pre_clip': 9.450338363647461, 'timing/data': 0.021079150028526783, 'timing/model': 1.0300487270578742, 'learning_rate/qwen_vl_interface': 9.870518301943018e-06, 'learning_rate/action_model': 9.870518301943018e-05, 'epoch': 0.22}) +[2026-06-17 17:25:59,988][starVLA.training.train_starvla][INFO] - Step 615, Loss: {'train/loss': 1.0839651823043823, 'train/grad_norm_pre_clip': 10.305806159973145, 'timing/data': 0.020801358856260777, 'timing/model': 1.3468949990347028, 'learning_rate/qwen_vl_interface': 9.870016304744617e-06, 'learning_rate/action_model': 9.870016304744616e-05, 'epoch': 0.23}) +[2026-06-17 17:26:04,234][starVLA.training.train_starvla][INFO] - Step 616, Loss: {'train/loss': 0.6899428367614746, 'train/grad_norm_pre_clip': 6.561985015869141, 'timing/data': 0.023586220107972622, 'timing/model': 1.209974060766399, 'learning_rate/qwen_vl_interface': 9.869513349811868e-06, 'learning_rate/action_model': 9.869513349811868e-05, 'epoch': 0.23}) +[2026-06-17 17:26:08,727][starVLA.training.train_starvla][INFO] - Step 617, Loss: {'train/loss': 0.8794146776199341, 'train/grad_norm_pre_clip': 13.375307083129883, 'timing/data': 0.022099805064499378, 'timing/model': 1.0498751839622855, 'learning_rate/qwen_vl_interface': 9.86900943724903e-06, 'learning_rate/action_model': 9.869009437249031e-05, 'epoch': 0.23}) +[2026-06-17 17:26:13,214][starVLA.training.train_starvla][INFO] - Step 618, Loss: {'train/loss': 0.9790565371513367, 'train/grad_norm_pre_clip': 9.050399780273438, 'timing/data': 0.020014754496514797, 'timing/model': 1.0642893239855766, 'learning_rate/qwen_vl_interface': 9.868504567160566e-06, 'learning_rate/action_model': 9.868504567160565e-05, 'epoch': 0.23}) +[2026-06-17 17:26:17,904][starVLA.training.train_starvla][INFO] - Step 619, Loss: {'train/loss': 0.8087159991264343, 'train/grad_norm_pre_clip': 9.114714622497559, 'timing/data': 0.019832467660307884, 'timing/model': 1.6893929792568088, 'learning_rate/qwen_vl_interface': 9.867998739651134e-06, 'learning_rate/action_model': 9.867998739651134e-05, 'epoch': 0.23}) +[2026-06-17 17:26:27,406][starVLA.training.train_starvla][INFO] - Step 620, Loss: {'train/loss': 0.9857364296913147, 'train/grad_norm_pre_clip': 10.026531219482422, 'timing/data': 0.019212116487324238, 'timing/model': 1.5541393738240004, 'learning_rate/qwen_vl_interface': 9.867491954825595e-06, 'learning_rate/action_model': 9.867491954825595e-05, 'epoch': 0.23}) +[2026-06-17 17:26:32,210][starVLA.training.train_starvla][INFO] - Step 621, Loss: {'train/loss': 1.054827332496643, 'train/grad_norm_pre_clip': 11.6088285446167, 'timing/data': 0.022626284509897232, 'timing/model': 1.2112140832468867, 'learning_rate/qwen_vl_interface': 9.866984212789007e-06, 'learning_rate/action_model': 9.866984212789006e-05, 'epoch': 0.23}) +[2026-06-17 17:26:39,157][starVLA.training.train_starvla][INFO] - Step 622, Loss: {'train/loss': 1.1960638761520386, 'train/grad_norm_pre_clip': 27.246213912963867, 'timing/data': 0.0223105913028121, 'timing/model': 1.1721396008506417, 'learning_rate/qwen_vl_interface': 9.86647551364662e-06, 'learning_rate/action_model': 9.86647551364662e-05, 'epoch': 0.23}) +[2026-06-17 17:26:50,806][starVLA.training.train_starvla][INFO] - Step 623, Loss: {'train/loss': 0.904057502746582, 'train/grad_norm_pre_clip': 7.824310779571533, 'timing/data': 0.0269218347966671, 'timing/model': 1.0417493032291532, 'learning_rate/qwen_vl_interface': 9.865965857503893e-06, 'learning_rate/action_model': 9.865965857503894e-05, 'epoch': 0.23}) +[2026-06-17 17:26:57,829][starVLA.training.train_starvla][INFO] - Step 624, Loss: {'train/loss': 1.093704342842102, 'train/grad_norm_pre_clip': 10.539999961853027, 'timing/data': 0.020357973873615265, 'timing/model': 1.2463995991274714, 'learning_rate/qwen_vl_interface': 9.865455244466478e-06, 'learning_rate/action_model': 9.865455244466477e-05, 'epoch': 0.23}) +[2026-06-17 17:27:07,242][starVLA.training.train_starvla][INFO] - Step 625, Loss: {'train/loss': 1.0156422853469849, 'train/grad_norm_pre_clip': 10.672554969787598, 'timing/data': 0.02232247218489647, 'timing/model': 1.5092617580667138, 'learning_rate/qwen_vl_interface': 9.864943674640223e-06, 'learning_rate/action_model': 9.864943674640222e-05, 'epoch': 0.23}) +[2026-06-17 17:27:17,209][starVLA.training.train_starvla][INFO] - Step 626, Loss: {'train/loss': 1.3031071424484253, 'train/grad_norm_pre_clip': 10.185940742492676, 'timing/data': 0.023513185791671276, 'timing/model': 1.1126957470551133, 'learning_rate/qwen_vl_interface': 9.864431148131178e-06, 'learning_rate/action_model': 9.864431148131178e-05, 'epoch': 0.23}) +[2026-06-17 17:27:23,781][starVLA.training.train_starvla][INFO] - Step 627, Loss: {'train/loss': 0.9281281232833862, 'train/grad_norm_pre_clip': 11.589739799499512, 'timing/data': 0.02529445756226778, 'timing/model': 2.6400879761204123, 'learning_rate/qwen_vl_interface': 9.863917665045591e-06, 'learning_rate/action_model': 9.863917665045592e-05, 'epoch': 0.23}) +[2026-06-17 17:27:33,019][starVLA.training.train_starvla][INFO] - Step 628, Loss: {'train/loss': 1.3114373683929443, 'train/grad_norm_pre_clip': 10.946816444396973, 'timing/data': 0.020241590216755867, 'timing/model': 1.0652529019862413, 'learning_rate/qwen_vl_interface': 9.863403225489909e-06, 'learning_rate/action_model': 9.863403225489909e-05, 'epoch': 0.23}) +[2026-06-17 17:27:40,662][starVLA.training.train_starvla][INFO] - Step 629, Loss: {'train/loss': 0.8227211236953735, 'train/grad_norm_pre_clip': 8.564746856689453, 'timing/data': 0.020937456749379635, 'timing/model': 1.375661070458591, 'learning_rate/qwen_vl_interface': 9.862887829570771e-06, 'learning_rate/action_model': 9.862887829570772e-05, 'epoch': 0.23}) +[2026-06-17 17:27:45,156][starVLA.training.train_starvla][INFO] - Step 630, Loss: {'train/loss': 1.1690419912338257, 'train/grad_norm_pre_clip': 9.599908828735352, 'timing/data': 0.01949946489185095, 'timing/model': 1.3798270942643285, 'learning_rate/qwen_vl_interface': 9.862371477395024e-06, 'learning_rate/action_model': 9.862371477395025e-05, 'epoch': 0.23}) +[2026-06-17 17:27:49,466][starVLA.training.train_starvla][INFO] - Step 631, Loss: {'train/loss': 0.9119678139686584, 'train/grad_norm_pre_clip': 8.022979736328125, 'timing/data': 0.018589324317872524, 'timing/model': 1.2848235499113798, 'learning_rate/qwen_vl_interface': 9.861854169069708e-06, 'learning_rate/action_model': 9.861854169069707e-05, 'epoch': 0.23}) +[2026-06-17 17:27:53,967][starVLA.training.train_starvla][INFO] - Step 632, Loss: {'train/loss': 0.7420274019241333, 'train/grad_norm_pre_clip': 9.089730262756348, 'timing/data': 0.021373323164880276, 'timing/model': 1.0183464959263802, 'learning_rate/qwen_vl_interface': 9.861335904702058e-06, 'learning_rate/action_model': 9.861335904702058e-05, 'epoch': 0.23}) +[2026-06-17 17:27:58,613][starVLA.training.train_starvla][INFO] - Step 633, Loss: {'train/loss': 1.0893728733062744, 'train/grad_norm_pre_clip': 11.801934242248535, 'timing/data': 0.022487347945570946, 'timing/model': 1.0300716888159513, 'learning_rate/qwen_vl_interface': 9.860816684399514e-06, 'learning_rate/action_model': 9.860816684399514e-05, 'epoch': 0.23}) +[2026-06-17 17:28:03,132][starVLA.training.train_starvla][INFO] - Step 634, Loss: {'train/loss': 1.249190330505371, 'train/grad_norm_pre_clip': 10.42459774017334, 'timing/data': 0.026666715741157532, 'timing/model': 1.1372399646788836, 'learning_rate/qwen_vl_interface': 9.86029650826971e-06, 'learning_rate/action_model': 9.86029650826971e-05, 'epoch': 0.23}) +[2026-06-17 17:28:07,813][starVLA.training.train_starvla][INFO] - Step 635, Loss: {'train/loss': 1.1612324714660645, 'train/grad_norm_pre_clip': 11.882965087890625, 'timing/data': 0.018867559731006622, 'timing/model': 1.4406961295753717, 'learning_rate/qwen_vl_interface': 9.859775376420478e-06, 'learning_rate/action_model': 9.859775376420478e-05, 'epoch': 0.23}) +[2026-06-17 17:28:12,159][starVLA.training.train_starvla][INFO] - Step 636, Loss: {'train/loss': 1.1176321506500244, 'train/grad_norm_pre_clip': 18.159149169921875, 'timing/data': 0.02131206076592207, 'timing/model': 1.0363572556525469, 'learning_rate/qwen_vl_interface': 9.85925328895985e-06, 'learning_rate/action_model': 9.85925328895985e-05, 'epoch': 0.23}) +[2026-06-17 17:28:16,533][starVLA.training.train_starvla][INFO] - Step 637, Loss: {'train/loss': 0.8223778009414673, 'train/grad_norm_pre_clip': 12.463562965393066, 'timing/data': 0.02162829414010048, 'timing/model': 1.0872741537168622, 'learning_rate/qwen_vl_interface': 9.858730245996057e-06, 'learning_rate/action_model': 9.858730245996057e-05, 'epoch': 0.23}) +[2026-06-17 17:28:21,086][starVLA.training.train_starvla][INFO] - Step 638, Loss: {'train/loss': 0.8729778528213501, 'train/grad_norm_pre_clip': 12.667720794677734, 'timing/data': 0.029140292666852474, 'timing/model': 1.0819388153031468, 'learning_rate/qwen_vl_interface': 9.858206247637522e-06, 'learning_rate/action_model': 9.858206247637523e-05, 'epoch': 0.23}) +[2026-06-17 17:28:25,483][starVLA.training.train_starvla][INFO] - Step 639, Loss: {'train/loss': 0.6423929929733276, 'train/grad_norm_pre_clip': 7.2220139503479, 'timing/data': 0.0213849525898695, 'timing/model': 1.3709658421576023, 'learning_rate/qwen_vl_interface': 9.857681293992876e-06, 'learning_rate/action_model': 9.857681293992877e-05, 'epoch': 0.23}) +[2026-06-17 17:28:29,988][starVLA.training.train_starvla][INFO] - Step 640, Loss: {'train/loss': 0.830906867980957, 'train/grad_norm_pre_clip': 9.396641731262207, 'timing/data': 0.027408745139837265, 'timing/model': 1.4017101805657148, 'learning_rate/qwen_vl_interface': 9.85715538517094e-06, 'learning_rate/action_model': 9.857155385170938e-05, 'epoch': 0.23}) +[2026-06-17 17:28:34,155][starVLA.training.train_starvla][INFO] - Step 641, Loss: {'train/loss': 0.9544891119003296, 'train/grad_norm_pre_clip': 8.377038955688477, 'timing/data': 0.020656468346714973, 'timing/model': 1.0289499713107944, 'learning_rate/qwen_vl_interface': 9.856628521280733e-06, 'learning_rate/action_model': 9.856628521280733e-05, 'epoch': 0.23}) +[2026-06-17 17:28:38,604][starVLA.training.train_starvla][INFO] - Step 642, Loss: {'train/loss': 0.7061123847961426, 'train/grad_norm_pre_clip': 7.898867607116699, 'timing/data': 0.020933330059051514, 'timing/model': 1.077559256926179, 'learning_rate/qwen_vl_interface': 9.856100702431477e-06, 'learning_rate/action_model': 9.856100702431477e-05, 'epoch': 0.24}) +[2026-06-17 17:28:43,395][starVLA.training.train_starvla][INFO] - Step 643, Loss: {'train/loss': 0.8614134192466736, 'train/grad_norm_pre_clip': 11.211005210876465, 'timing/data': 0.08624549396336079, 'timing/model': 1.274218874052167, 'learning_rate/qwen_vl_interface': 9.85557192873259e-06, 'learning_rate/action_model': 9.85557192873259e-05, 'epoch': 0.24}) +[2026-06-17 17:28:54,434][starVLA.training.train_starvla][INFO] - Step 644, Loss: {'train/loss': 0.6878277063369751, 'train/grad_norm_pre_clip': 7.750011444091797, 'timing/data': 0.024784616194665432, 'timing/model': 4.109561240300536, 'learning_rate/qwen_vl_interface': 9.855042200293684e-06, 'learning_rate/action_model': 9.855042200293685e-05, 'epoch': 0.24}) +[2026-06-17 17:28:59,911][starVLA.training.train_starvla][INFO] - Step 645, Loss: {'train/loss': 1.1090672016143799, 'train/grad_norm_pre_clip': 13.994937896728516, 'timing/data': 0.027823195792734623, 'timing/model': 1.7653226051479578, 'learning_rate/qwen_vl_interface': 9.854511517224578e-06, 'learning_rate/action_model': 9.854511517224578e-05, 'epoch': 0.24}) +[2026-06-17 17:29:07,445][starVLA.training.train_starvla][INFO] - Step 646, Loss: {'train/loss': 1.3333673477172852, 'train/grad_norm_pre_clip': 13.198164939880371, 'timing/data': 0.0224884282797575, 'timing/model': 4.220958311110735, 'learning_rate/qwen_vl_interface': 9.853979879635278e-06, 'learning_rate/action_model': 9.853979879635278e-05, 'epoch': 0.24}) +[2026-06-17 17:29:15,321][starVLA.training.train_starvla][INFO] - Step 647, Loss: {'train/loss': 1.0933265686035156, 'train/grad_norm_pre_clip': 12.436195373535156, 'timing/data': 0.020801459439098835, 'timing/model': 1.1445921510457993, 'learning_rate/qwen_vl_interface': 9.853447287635995e-06, 'learning_rate/action_model': 9.853447287635995e-05, 'epoch': 0.24}) +[2026-06-17 17:29:22,216][starVLA.training.train_starvla][INFO] - Step 648, Loss: {'train/loss': 1.2556108236312866, 'train/grad_norm_pre_clip': 12.000990867614746, 'timing/data': 0.022647212259471416, 'timing/model': 1.1277749491855502, 'learning_rate/qwen_vl_interface': 9.852913741337136e-06, 'learning_rate/action_model': 9.852913741337136e-05, 'epoch': 0.24}) +[2026-06-17 17:29:26,663][starVLA.training.train_starvla][INFO] - Step 649, Loss: {'train/loss': 1.12654447555542, 'train/grad_norm_pre_clip': 13.55984115600586, 'timing/data': 0.020536838099360466, 'timing/model': 1.3823760487139225, 'learning_rate/qwen_vl_interface': 9.852379240849306e-06, 'learning_rate/action_model': 9.852379240849305e-05, 'epoch': 0.24}) +[2026-06-17 17:29:31,159][starVLA.training.train_starvla][INFO] - Step 650, Loss: {'train/loss': 1.3523972034454346, 'train/grad_norm_pre_clip': 14.874080657958984, 'timing/data': 0.03384463395923376, 'timing/model': 1.4439226277172565, 'learning_rate/qwen_vl_interface': 9.851843786283306e-06, 'learning_rate/action_model': 9.851843786283306e-05, 'epoch': 0.24}) +[2026-06-17 17:29:37,615][starVLA.training.train_starvla][INFO] - Step 651, Loss: {'train/loss': 1.0221309661865234, 'train/grad_norm_pre_clip': 10.607979774475098, 'timing/data': 0.024463968351483345, 'timing/model': 1.376504865474999, 'learning_rate/qwen_vl_interface': 9.851307377750137e-06, 'learning_rate/action_model': 9.851307377750137e-05, 'epoch': 0.24}) +[2026-06-17 17:29:42,252][starVLA.training.train_starvla][INFO] - Step 652, Loss: {'train/loss': 1.411927580833435, 'train/grad_norm_pre_clip': 16.444751739501953, 'timing/data': 0.020858771167695522, 'timing/model': 1.039785078726709, 'learning_rate/qwen_vl_interface': 9.850770015360999e-06, 'learning_rate/action_model': 9.850770015360999e-05, 'epoch': 0.24}) +[2026-06-17 17:29:46,662][starVLA.training.train_starvla][INFO] - Step 653, Loss: {'train/loss': 0.8590906858444214, 'train/grad_norm_pre_clip': 7.69363260269165, 'timing/data': 0.0196683993563056, 'timing/model': 1.0380301252007484, 'learning_rate/qwen_vl_interface': 9.850231699227284e-06, 'learning_rate/action_model': 9.850231699227284e-05, 'epoch': 0.24}) +[2026-06-17 17:29:51,205][starVLA.training.train_starvla][INFO] - Step 654, Loss: {'train/loss': 1.0309306383132935, 'train/grad_norm_pre_clip': 15.663768768310547, 'timing/data': 0.020985804498195648, 'timing/model': 1.0731359850615263, 'learning_rate/qwen_vl_interface': 9.84969242946059e-06, 'learning_rate/action_model': 9.84969242946059e-05, 'epoch': 0.24}) +[2026-06-17 17:29:55,653][starVLA.training.train_starvla][INFO] - Step 655, Loss: {'train/loss': 0.9560694098472595, 'train/grad_norm_pre_clip': 10.580312728881836, 'timing/data': 0.020672883838415146, 'timing/model': 1.4490454448387027, 'learning_rate/qwen_vl_interface': 9.849152206172704e-06, 'learning_rate/action_model': 9.849152206172704e-05, 'epoch': 0.24}) +[2026-06-17 17:29:59,668][starVLA.training.train_starvla][INFO] - Step 656, Loss: {'train/loss': 0.9082782864570618, 'train/grad_norm_pre_clip': 13.015084266662598, 'timing/data': 0.019482537172734737, 'timing/model': 1.0345009416341782, 'learning_rate/qwen_vl_interface': 9.84861102947562e-06, 'learning_rate/action_model': 9.848611029475619e-05, 'epoch': 0.24}) +[2026-06-17 17:30:04,383][starVLA.training.train_starvla][INFO] - Step 657, Loss: {'train/loss': 0.7578959465026855, 'train/grad_norm_pre_clip': 13.245368003845215, 'timing/data': 0.020835776813328266, 'timing/model': 1.048310725018382, 'learning_rate/qwen_vl_interface': 9.848068899481516e-06, 'learning_rate/action_model': 9.848068899481516e-05, 'epoch': 0.24}) +[2026-06-17 17:30:08,850][starVLA.training.train_starvla][INFO] - Step 658, Loss: {'train/loss': 0.8174154162406921, 'train/grad_norm_pre_clip': 8.818976402282715, 'timing/data': 0.019932132214307785, 'timing/model': 1.0242432421073318, 'learning_rate/qwen_vl_interface': 9.847525816302784e-06, 'learning_rate/action_model': 9.847525816302784e-05, 'epoch': 0.24}) +[2026-06-17 17:30:13,325][starVLA.training.train_starvla][INFO] - Step 659, Loss: {'train/loss': 0.8612310290336609, 'train/grad_norm_pre_clip': 8.945205688476562, 'timing/data': 0.024226127192378044, 'timing/model': 1.404030429199338, 'learning_rate/qwen_vl_interface': 9.846981780052005e-06, 'learning_rate/action_model': 9.846981780052003e-05, 'epoch': 0.24}) +[2026-06-17 17:30:17,853][starVLA.training.train_starvla][INFO] - Step 660, Loss: {'train/loss': 1.3202261924743652, 'train/grad_norm_pre_clip': 9.75903034210205, 'timing/data': 0.019129646010696888, 'timing/model': 1.465088359080255, 'learning_rate/qwen_vl_interface': 9.846436790841955e-06, 'learning_rate/action_model': 9.846436790841955e-05, 'epoch': 0.24}) +[2026-06-17 17:30:21,996][starVLA.training.train_starvla][INFO] - Step 661, Loss: {'train/loss': 0.8271220326423645, 'train/grad_norm_pre_clip': 18.27916145324707, 'timing/data': 0.02100243978202343, 'timing/model': 1.0324836876243353, 'learning_rate/qwen_vl_interface': 9.84589084878561e-06, 'learning_rate/action_model': 9.845890848785611e-05, 'epoch': 0.24}) +[2026-06-17 17:30:26,366][starVLA.training.train_starvla][INFO] - Step 662, Loss: {'train/loss': 1.1296919584274292, 'train/grad_norm_pre_clip': 17.077592849731445, 'timing/data': 0.02046738751232624, 'timing/model': 1.0840620938688517, 'learning_rate/qwen_vl_interface': 9.84534395399615e-06, 'learning_rate/action_model': 9.84534395399615e-05, 'epoch': 0.24}) +[2026-06-17 17:30:31,053][starVLA.training.train_starvla][INFO] - Step 663, Loss: {'train/loss': 1.1202507019042969, 'train/grad_norm_pre_clip': 11.53305721282959, 'timing/data': 0.02044784650206566, 'timing/model': 1.156192610040307, 'learning_rate/qwen_vl_interface': 9.844796106586942e-06, 'learning_rate/action_model': 9.844796106586941e-05, 'epoch': 0.24}) +[2026-06-17 17:30:36,252][starVLA.training.train_starvla][INFO] - Step 664, Loss: {'train/loss': 1.23921799659729, 'train/grad_norm_pre_clip': 20.137601852416992, 'timing/data': 0.026556359604001045, 'timing/model': 1.1767797376960516, 'learning_rate/qwen_vl_interface': 9.844247306671557e-06, 'learning_rate/action_model': 9.844247306671555e-05, 'epoch': 0.24}) +[2026-06-17 17:30:40,771][starVLA.training.train_starvla][INFO] - Step 665, Loss: {'train/loss': 0.8181849718093872, 'train/grad_norm_pre_clip': 7.866286754608154, 'timing/data': 0.01949777826666832, 'timing/model': 1.4140829527750611, 'learning_rate/qwen_vl_interface': 9.843697554363762e-06, 'learning_rate/action_model': 9.843697554363762e-05, 'epoch': 0.24}) +[2026-06-17 17:30:44,998][starVLA.training.train_starvla][INFO] - Step 666, Loss: {'train/loss': 1.0552642345428467, 'train/grad_norm_pre_clip': 11.30548095703125, 'timing/data': 0.01948403473943472, 'timing/model': 1.0930992979556322, 'learning_rate/qwen_vl_interface': 9.84314684977752e-06, 'learning_rate/action_model': 9.84314684977752e-05, 'epoch': 0.24}) +[2026-06-17 17:30:49,569][starVLA.training.train_starvla][INFO] - Step 667, Loss: {'train/loss': 1.1638836860656738, 'train/grad_norm_pre_clip': 11.978632926940918, 'timing/data': 0.09153967350721359, 'timing/model': 1.13093629386276, 'learning_rate/qwen_vl_interface': 9.842595193026993e-06, 'learning_rate/action_model': 9.842595193026993e-05, 'epoch': 0.24}) +[2026-06-17 17:31:03,612][starVLA.training.train_starvla][INFO] - Step 668, Loss: {'train/loss': 1.1638708114624023, 'train/grad_norm_pre_clip': 12.910313606262207, 'timing/data': 0.019256511703133583, 'timing/model': 1.5332222916185856, 'learning_rate/qwen_vl_interface': 9.842042584226545e-06, 'learning_rate/action_model': 9.842042584226544e-05, 'epoch': 0.24}) +[2026-06-17 17:31:08,192][starVLA.training.train_starvla][INFO] - Step 669, Loss: {'train/loss': 1.2326034307479858, 'train/grad_norm_pre_clip': 9.47160816192627, 'timing/data': 0.020762069150805473, 'timing/model': 1.461515891365707, 'learning_rate/qwen_vl_interface': 9.841489023490723e-06, 'learning_rate/action_model': 9.841489023490723e-05, 'epoch': 0.24}) +[2026-06-17 17:31:12,913][starVLA.training.train_starvla][INFO] - Step 670, Loss: {'train/loss': 1.1477818489074707, 'train/grad_norm_pre_clip': 11.67863941192627, 'timing/data': 0.022247211076319218, 'timing/model': 1.4659527642652392, 'learning_rate/qwen_vl_interface': 9.840934510934288e-06, 'learning_rate/action_model': 9.840934510934288e-05, 'epoch': 0.25}) +[2026-06-17 17:31:17,375][starVLA.training.train_starvla][INFO] - Step 671, Loss: {'train/loss': 0.7550649046897888, 'train/grad_norm_pre_clip': 8.288402557373047, 'timing/data': 0.019602426327764988, 'timing/model': 1.1469329157844186, 'learning_rate/qwen_vl_interface': 9.840379046672189e-06, 'learning_rate/action_model': 9.840379046672189e-05, 'epoch': 0.25}) +[2026-06-17 17:31:21,869][starVLA.training.train_starvla][INFO] - Step 672, Loss: {'train/loss': 0.767061173915863, 'train/grad_norm_pre_clip': 8.920614242553711, 'timing/data': 0.01919155940413475, 'timing/model': 1.0295255417004228, 'learning_rate/qwen_vl_interface': 9.839822630819573e-06, 'learning_rate/action_model': 9.839822630819572e-05, 'epoch': 0.25}) +[2026-06-17 17:31:26,470][starVLA.training.train_starvla][INFO] - Step 673, Loss: {'train/loss': 0.8092700242996216, 'train/grad_norm_pre_clip': 13.089136123657227, 'timing/data': 0.02515689842402935, 'timing/model': 1.0754559086635709, 'learning_rate/qwen_vl_interface': 9.839265263491789e-06, 'learning_rate/action_model': 9.839265263491789e-05, 'epoch': 0.25}) +[2026-06-17 17:31:31,317][starVLA.training.train_starvla][INFO] - Step 674, Loss: {'train/loss': 1.0812160968780518, 'train/grad_norm_pre_clip': 9.54635238647461, 'timing/data': 0.021208166144788265, 'timing/model': 1.0860989587381482, 'learning_rate/qwen_vl_interface': 9.838706944804376e-06, 'learning_rate/action_model': 9.838706944804374e-05, 'epoch': 0.25}) +[2026-06-17 17:31:36,186][starVLA.training.train_starvla][INFO] - Step 675, Loss: {'train/loss': 0.8712316155433655, 'train/grad_norm_pre_clip': 8.613469123840332, 'timing/data': 0.019442525692284107, 'timing/model': 1.5193559024482965, 'learning_rate/qwen_vl_interface': 9.838147674873076e-06, 'learning_rate/action_model': 9.838147674873075e-05, 'epoch': 0.25}) +[2026-06-17 17:31:40,363][starVLA.training.train_starvla][INFO] - Step 676, Loss: {'train/loss': 0.8933358192443848, 'train/grad_norm_pre_clip': 10.436421394348145, 'timing/data': 0.02508420217782259, 'timing/model': 1.1500777173787355, 'learning_rate/qwen_vl_interface': 9.837587453813826e-06, 'learning_rate/action_model': 9.837587453813824e-05, 'epoch': 0.25}) +[2026-06-17 17:31:44,930][starVLA.training.train_starvla][INFO] - Step 677, Loss: {'train/loss': 0.9211423993110657, 'train/grad_norm_pre_clip': 8.58486270904541, 'timing/data': 0.023477043956518173, 'timing/model': 1.1025999151170254, 'learning_rate/qwen_vl_interface': 9.83702628174276e-06, 'learning_rate/action_model': 9.837026281742759e-05, 'epoch': 0.25}) +[2026-06-17 17:31:49,875][starVLA.training.train_starvla][INFO] - Step 678, Loss: {'train/loss': 0.8995309472084045, 'train/grad_norm_pre_clip': 10.34205150604248, 'timing/data': 0.0009651500731706619, 'timing/model': 1.033282095566392, 'learning_rate/qwen_vl_interface': 9.83646415877621e-06, 'learning_rate/action_model': 9.83646415877621e-05, 'epoch': 0.25}) +[2026-06-17 17:31:58,373][starVLA.training.train_starvla][INFO] - Step 679, Loss: {'train/loss': 0.8650745749473572, 'train/grad_norm_pre_clip': 9.717997550964355, 'timing/data': 0.001347200945019722, 'timing/model': 1.0886244559660554, 'learning_rate/qwen_vl_interface': 9.835901085030703e-06, 'learning_rate/action_model': 9.835901085030703e-05, 'epoch': 0.25}) +[2026-06-17 17:32:02,975][starVLA.training.train_starvla][INFO] - Step 680, Loss: {'train/loss': 1.153335690498352, 'train/grad_norm_pre_clip': 10.024694442749023, 'timing/data': 0.001128964126110077, 'timing/model': 1.355763616040349, 'learning_rate/qwen_vl_interface': 9.83533706062297e-06, 'learning_rate/action_model': 9.835337060622969e-05, 'epoch': 0.25}) +[2026-06-17 17:32:07,477][starVLA.training.train_starvla][INFO] - Step 681, Loss: {'train/loss': 1.0966793298721313, 'train/grad_norm_pre_clip': 10.969101905822754, 'timing/data': 0.0008602850139141083, 'timing/model': 1.045717230066657, 'learning_rate/qwen_vl_interface': 9.834772085669927e-06, 'learning_rate/action_model': 9.834772085669927e-05, 'epoch': 0.25}) +[2026-06-17 17:32:13,117][starVLA.training.train_starvla][INFO] - Step 682, Loss: {'train/loss': 0.7285168170928955, 'train/grad_norm_pre_clip': 10.111359596252441, 'timing/data': 0.0012601958587765694, 'timing/model': 1.1784055205062032, 'learning_rate/qwen_vl_interface': 9.834206160288698e-06, 'learning_rate/action_model': 9.834206160288698e-05, 'epoch': 0.25}) +[2026-06-17 17:32:17,805][starVLA.training.train_starvla][INFO] - Step 683, Loss: {'train/loss': 1.0574140548706055, 'train/grad_norm_pre_clip': 11.7440824508667, 'timing/data': 0.0011879382655024529, 'timing/model': 1.0994872320443392, 'learning_rate/qwen_vl_interface': 9.8336392845966e-06, 'learning_rate/action_model': 9.833639284596599e-05, 'epoch': 0.25}) +[2026-06-17 17:32:27,337][starVLA.training.train_starvla][INFO] - Step 684, Loss: {'train/loss': 0.7645302414894104, 'train/grad_norm_pre_clip': 8.777069091796875, 'timing/data': 0.001062881201505661, 'timing/model': 1.2326887985691428, 'learning_rate/qwen_vl_interface': 9.833071458711146e-06, 'learning_rate/action_model': 9.833071458711146e-05, 'epoch': 0.25}) +[2026-06-17 17:32:32,185][starVLA.training.train_starvla][INFO] - Step 685, Loss: {'train/loss': 0.9942209720611572, 'train/grad_norm_pre_clip': 13.418195724487305, 'timing/data': 0.0013861330226063728, 'timing/model': 1.1337780756875873, 'learning_rate/qwen_vl_interface': 9.832502682750048e-06, 'learning_rate/action_model': 9.832502682750047e-05, 'epoch': 0.25}) +[2026-06-17 17:32:37,026][starVLA.training.train_starvla][INFO] - Step 686, Loss: {'train/loss': 1.1270477771759033, 'train/grad_norm_pre_clip': 13.220375061035156, 'timing/data': 0.0007899319753050804, 'timing/model': 1.1094471151009202, 'learning_rate/qwen_vl_interface': 9.831932956831212e-06, 'learning_rate/action_model': 9.831932956831211e-05, 'epoch': 0.25}) +[2026-06-17 17:32:44,299][starVLA.training.train_starvla][INFO] - Step 687, Loss: {'train/loss': 0.9231350421905518, 'train/grad_norm_pre_clip': 10.807673454284668, 'timing/data': 0.0011802827939391136, 'timing/model': 1.0443765874952078, 'learning_rate/qwen_vl_interface': 9.831362281072743e-06, 'learning_rate/action_model': 9.831362281072742e-05, 'epoch': 0.25}) +[2026-06-17 17:32:48,846][starVLA.training.train_starvla][INFO] - Step 688, Loss: {'train/loss': 1.0851818323135376, 'train/grad_norm_pre_clip': 9.754137992858887, 'timing/data': 0.001304461620748043, 'timing/model': 1.038998430594802, 'learning_rate/qwen_vl_interface': 9.830790655592944e-06, 'learning_rate/action_model': 9.830790655592943e-05, 'epoch': 0.25}) +[2026-06-17 17:32:57,515][starVLA.training.train_starvla][INFO] - Step 689, Loss: {'train/loss': 1.210988998413086, 'train/grad_norm_pre_clip': 11.778996467590332, 'timing/data': 0.0009467164054512978, 'timing/model': 1.0837581781670451, 'learning_rate/qwen_vl_interface': 9.830218080510313e-06, 'learning_rate/action_model': 9.830218080510312e-05, 'epoch': 0.25}) +[2026-06-17 17:33:05,048][starVLA.training.train_starvla][INFO] - Step 690, Loss: {'train/loss': 1.2021363973617554, 'train/grad_norm_pre_clip': 19.353023529052734, 'timing/data': 0.0016411198303103447, 'timing/model': 1.4113066662102938, 'learning_rate/qwen_vl_interface': 9.829644555943545e-06, 'learning_rate/action_model': 9.829644555943545e-05, 'epoch': 0.25}) +[2026-06-17 17:33:09,093][starVLA.training.train_starvla][INFO] - Step 691, Loss: {'train/loss': 0.6649283170700073, 'train/grad_norm_pre_clip': 11.395458221435547, 'timing/data': 0.0010268250480294228, 'timing/model': 1.0532059064134955, 'learning_rate/qwen_vl_interface': 9.82907008201153e-06, 'learning_rate/action_model': 9.829070082011531e-05, 'epoch': 0.25}) +[2026-06-17 17:33:14,067][starVLA.training.train_starvla][INFO] - Step 692, Loss: {'train/loss': 1.0002901554107666, 'train/grad_norm_pre_clip': 9.89005184173584, 'timing/data': 0.0008960766717791557, 'timing/model': 1.0199995571747422, 'learning_rate/qwen_vl_interface': 9.828494658833366e-06, 'learning_rate/action_model': 9.828494658833365e-05, 'epoch': 0.25}) +[2026-06-17 17:33:18,589][starVLA.training.train_starvla][INFO] - Step 693, Loss: {'train/loss': 1.0959701538085938, 'train/grad_norm_pre_clip': 12.157973289489746, 'timing/data': 0.0017058951780200005, 'timing/model': 1.0986528871580958, 'learning_rate/qwen_vl_interface': 9.82791828652833e-06, 'learning_rate/action_model': 9.827918286528329e-05, 'epoch': 0.25}) +[2026-06-17 17:33:23,178][starVLA.training.train_starvla][INFO] - Step 694, Loss: {'train/loss': 0.7689855694770813, 'train/grad_norm_pre_clip': 10.071372032165527, 'timing/data': 0.0009018750861287117, 'timing/model': 1.045409801416099, 'learning_rate/qwen_vl_interface': 9.827340965215903e-06, 'learning_rate/action_model': 9.827340965215904e-05, 'epoch': 0.25}) +[2026-06-17 17:33:27,649][starVLA.training.train_starvla][INFO] - Step 695, Loss: {'train/loss': 0.7251216173171997, 'train/grad_norm_pre_clip': 9.45955753326416, 'timing/data': 0.0009028250351548195, 'timing/model': 1.0479527143761516, 'learning_rate/qwen_vl_interface': 9.826762695015773e-06, 'learning_rate/action_model': 9.826762695015772e-05, 'epoch': 0.25}) +[2026-06-17 17:33:32,378][starVLA.training.train_starvla][INFO] - Step 696, Loss: {'train/loss': 0.5762797594070435, 'train/grad_norm_pre_clip': 9.64277458190918, 'timing/data': 0.001304827630519867, 'timing/model': 1.1176952701061964, 'learning_rate/qwen_vl_interface': 9.82618347604781e-06, 'learning_rate/action_model': 9.82618347604781e-05, 'epoch': 0.25}) +[2026-06-17 17:33:40,173][starVLA.training.train_starvla][INFO] - Step 697, Loss: {'train/loss': 0.8156794905662537, 'train/grad_norm_pre_clip': 11.158976554870605, 'timing/data': 0.001287182793021202, 'timing/model': 1.1042943745851517, 'learning_rate/qwen_vl_interface': 9.825603308432089e-06, 'learning_rate/action_model': 9.825603308432088e-05, 'epoch': 0.26}) +[2026-06-17 17:33:45,039][starVLA.training.train_starvla][INFO] - Step 698, Loss: {'train/loss': 0.7624971866607666, 'train/grad_norm_pre_clip': 9.933258056640625, 'timing/data': 0.0013613123446702957, 'timing/model': 1.191237366758287, 'learning_rate/qwen_vl_interface': 9.825022192288878e-06, 'learning_rate/action_model': 9.825022192288877e-05, 'epoch': 0.26}) +[2026-06-17 17:33:49,720][starVLA.training.train_starvla][INFO] - Step 699, Loss: {'train/loss': 0.9855507016181946, 'train/grad_norm_pre_clip': 15.855751037597656, 'timing/data': 0.0011949213221669197, 'timing/model': 1.5030599208548665, 'learning_rate/qwen_vl_interface': 9.824440127738643e-06, 'learning_rate/action_model': 9.824440127738642e-05, 'epoch': 0.26}) +[2026-06-17 17:37:07,049][starVLA.training.train_starvla][INFO] - Step 700, Loss: {'train/loss': 0.8645768165588379, 'train/grad_norm_pre_clip': 13.881725311279297, 'eval/loss': 1.0085334777832031, 'eval/latency_2/loss': 1.0085334777832031, 'eval/demon_fire_precision': 0.8242446573323507, 'eval/demon_fire_recall': 0.9359832635983264, 'eval/demon_fire_f1': 0.8765673981191223, 'eval/demon_fire_rate_model': 0.755147468002226, 'eval/demon_fire_rate_teacher': 0.6649972175848636, 'eval/demon_left_precision': 0.7571743929359823, 'eval/demon_left_recall': 0.7221052631578947, 'eval/demon_left_f1': 0.7392241379310345, 'eval/demon_left_rate_model': 0.25208681135225375, 'eval/demon_left_rate_teacher': 0.2643294379521425, 'eval/demon_right_precision': 0.7720797720797721, 'eval/demon_right_recall': 0.6124293785310735, 'eval/demon_right_f1': 0.6830497794580971, 'eval/demon_right_rate_model': 0.19532554257095158, 'eval/demon_right_rate_teacher': 0.24624373956594323, 'eval/demon_move_f1': 0.7111369586945657, 'eval/demon_cc_f1': 0.819383883371534, 'eval/latency_2/demon_fire_precision': 0.8242446573323507, 'eval/latency_2/demon_fire_recall': 0.9359832635983264, 'eval/latency_2/demon_fire_f1': 0.8765673981191223, 'eval/latency_2/demon_fire_rate_model': 0.755147468002226, 'eval/latency_2/demon_fire_rate_teacher': 0.6649972175848636, 'eval/latency_2/demon_left_precision': 0.7571743929359823, 'eval/latency_2/demon_left_recall': 0.7221052631578947, 'eval/latency_2/demon_left_f1': 0.7392241379310345, 'eval/latency_2/demon_left_rate_model': 0.25208681135225375, 'eval/latency_2/demon_left_rate_teacher': 0.2643294379521425, 'eval/latency_2/demon_right_precision': 0.7720797720797721, 'eval/latency_2/demon_right_recall': 0.6124293785310735, 'eval/latency_2/demon_right_f1': 0.6830497794580971, 'eval/latency_2/demon_right_rate_model': 0.19532554257095158, 'eval/latency_2/demon_right_rate_teacher': 0.24624373956594323, 'eval/latency_2/demon_move_f1': 0.7111369586945657, 'eval/latency_2/demon_cc_f1': 0.819383883371534, 'timing/data': 0.001300731673836708, 'timing/model': 1.524925691075623, 'learning_rate/qwen_vl_interface': 9.823857114902047e-06, 'learning_rate/action_model': 9.823857114902047e-05, 'epoch': 0.26}) +[2026-06-17 17:37:07,064][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_700_state +[2026-06-17 17:37:07,072][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 17:37:15,446][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_700_state/model.safetensors +[2026-06-17 17:37:30,416][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_700_state/optimizer.bin +[2026-06-17 17:37:30,417][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_700_state/sampler.bin +[2026-06-17 17:37:30,418][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_700_state/sampler_1.bin +[2026-06-17 17:37:30,420][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_700_state/random_states_0.pkl +[2026-06-17 17:37:30,425][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 700 +[2026-06-17 17:37:30,425][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 17:37:30,442][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 17:37:35,227][starVLA.training.train_starvla][INFO] - Step 701, Loss: {'train/loss': 1.2310373783111572, 'train/grad_norm_pre_clip': 30.67075538635254, 'timing/data': 0.0010095415636897087, 'timing/model': 1.7375184502452612, 'learning_rate/qwen_vl_interface': 9.823273153899951e-06, 'learning_rate/action_model': 9.82327315389995e-05, 'epoch': 0.26}) +[2026-06-17 17:37:39,677][starVLA.training.train_starvla][INFO] - Step 702, Loss: {'train/loss': 0.810506284236908, 'train/grad_norm_pre_clip': 20.429380416870117, 'timing/data': 0.0009095696732401848, 'timing/model': 1.205808718688786, 'learning_rate/qwen_vl_interface': 9.822688244853407e-06, 'learning_rate/action_model': 9.822688244853408e-05, 'epoch': 0.26}) +[2026-06-17 17:37:44,523][starVLA.training.train_starvla][INFO] - Step 703, Loss: {'train/loss': 0.7820857167243958, 'train/grad_norm_pre_clip': 11.135741233825684, 'timing/data': 0.0011771004647016525, 'timing/model': 1.050930061377585, 'learning_rate/qwen_vl_interface': 9.822102387883672e-06, 'learning_rate/action_model': 9.822102387883671e-05, 'epoch': 0.26}) +[2026-06-17 17:37:52,218][starVLA.training.train_starvla][INFO] - Step 704, Loss: {'train/loss': 1.042427659034729, 'train/grad_norm_pre_clip': 11.400845527648926, 'timing/data': 0.0010290639474987984, 'timing/model': 1.0880710370838642, 'learning_rate/qwen_vl_interface': 9.82151558311219e-06, 'learning_rate/action_model': 9.821515583112191e-05, 'epoch': 0.26}) +[2026-06-17 17:38:00,001][starVLA.training.train_starvla][INFO] - Step 705, Loss: {'train/loss': 0.9374603033065796, 'train/grad_norm_pre_clip': 12.254064559936523, 'timing/data': 0.001067240722477436, 'timing/model': 4.128863321617246, 'learning_rate/qwen_vl_interface': 9.82092783066061e-06, 'learning_rate/action_model': 9.82092783066061e-05, 'epoch': 0.26}) +[2026-06-17 17:38:05,013][starVLA.training.train_starvla][INFO] - Step 706, Loss: {'train/loss': 1.076063632965088, 'train/grad_norm_pre_clip': 12.547825813293457, 'timing/data': 0.0009577013552188873, 'timing/model': 1.8221678985282779, 'learning_rate/qwen_vl_interface': 9.820339130650771e-06, 'learning_rate/action_model': 9.820339130650771e-05, 'epoch': 0.26}) +[2026-06-17 17:38:15,688][starVLA.training.train_starvla][INFO] - Step 707, Loss: {'train/loss': 0.7698710560798645, 'train/grad_norm_pre_clip': 10.610733985900879, 'timing/data': 0.001550561748445034, 'timing/model': 4.5345867816358805, 'learning_rate/qwen_vl_interface': 9.819749483204715e-06, 'learning_rate/action_model': 9.819749483204715e-05, 'epoch': 0.26}) +[2026-06-17 17:38:22,168][starVLA.training.train_starvla][INFO] - Step 708, Loss: {'train/loss': 1.053079605102539, 'train/grad_norm_pre_clip': 23.579004287719727, 'timing/data': 0.0009186314418911934, 'timing/model': 1.0401401109993458, 'learning_rate/qwen_vl_interface': 9.81915888844467e-06, 'learning_rate/action_model': 9.81915888844467e-05, 'epoch': 0.26}) +[2026-06-17 17:38:26,750][starVLA.training.train_starvla][INFO] - Step 709, Loss: {'train/loss': 0.9243817329406738, 'train/grad_norm_pre_clip': 12.528881072998047, 'timing/data': 0.0012402096763253212, 'timing/model': 1.0785991726443172, 'learning_rate/qwen_vl_interface': 9.818567346493072e-06, 'learning_rate/action_model': 9.818567346493071e-05, 'epoch': 0.26}) +[2026-06-17 17:38:31,405][starVLA.training.train_starvla][INFO] - Step 710, Loss: {'train/loss': 1.0598630905151367, 'train/grad_norm_pre_clip': 13.51202392578125, 'timing/data': 0.0009570419788360596, 'timing/model': 1.1311520859599113, 'learning_rate/qwen_vl_interface': 9.817974857472549e-06, 'learning_rate/action_model': 9.817974857472548e-05, 'epoch': 0.26}) +[2026-06-17 17:38:40,979][starVLA.training.train_starvla][INFO] - Step 711, Loss: {'train/loss': 0.654512882232666, 'train/grad_norm_pre_clip': 7.1621599197387695, 'timing/data': 0.0010018125176429749, 'timing/model': 5.949492931365967, 'learning_rate/qwen_vl_interface': 9.817381421505922e-06, 'learning_rate/action_model': 9.817381421505921e-05, 'epoch': 0.26}) +[2026-06-17 17:38:45,105][starVLA.training.train_starvla][INFO] - Step 712, Loss: {'train/loss': 1.281159520149231, 'train/grad_norm_pre_clip': 13.219993591308594, 'timing/data': 0.0008418643847107887, 'timing/model': 1.0728190001100302, 'learning_rate/qwen_vl_interface': 9.81678703871621e-06, 'learning_rate/action_model': 9.81678703871621e-05, 'epoch': 0.26}) +[2026-06-17 17:38:49,541][starVLA.training.train_starvla][INFO] - Step 713, Loss: {'train/loss': 0.7653716802597046, 'train/grad_norm_pre_clip': 12.289382934570312, 'timing/data': 0.0011484352871775627, 'timing/model': 1.020992262288928, 'learning_rate/qwen_vl_interface': 9.816191709226631e-06, 'learning_rate/action_model': 9.816191709226632e-05, 'epoch': 0.26}) +[2026-06-17 17:38:54,365][starVLA.training.train_starvla][INFO] - Step 714, Loss: {'train/loss': 0.9914069175720215, 'train/grad_norm_pre_clip': 13.968193054199219, 'timing/data': 0.0012378431856632233, 'timing/model': 1.0280924560502172, 'learning_rate/qwen_vl_interface': 9.815595433160599e-06, 'learning_rate/action_model': 9.815595433160598e-05, 'epoch': 0.26}) +[2026-06-17 17:39:03,186][starVLA.training.train_starvla][INFO] - Step 715, Loss: {'train/loss': 1.0872547626495361, 'train/grad_norm_pre_clip': 10.282485961914062, 'timing/data': 0.0008908296003937721, 'timing/model': 1.0684330817312002, 'learning_rate/qwen_vl_interface': 9.81499821064172e-06, 'learning_rate/action_model': 9.81499821064172e-05, 'epoch': 0.26}) +[2026-06-17 17:39:07,908][starVLA.training.train_starvla][INFO] - Step 716, Loss: {'train/loss': 1.177018404006958, 'train/grad_norm_pre_clip': 12.484865188598633, 'timing/data': 0.0009540524333715439, 'timing/model': 1.5729955891147256, 'learning_rate/qwen_vl_interface': 9.814400041793799e-06, 'learning_rate/action_model': 9.814400041793798e-05, 'epoch': 0.26}) +[2026-06-17 17:39:13,016][starVLA.training.train_starvla][INFO] - Step 717, Loss: {'train/loss': 0.6420682668685913, 'train/grad_norm_pre_clip': 9.015435218811035, 'timing/data': 0.0008857911452651024, 'timing/model': 1.0720517188310623, 'learning_rate/qwen_vl_interface': 9.813800926740839e-06, 'learning_rate/action_model': 9.813800926740838e-05, 'epoch': 0.26}) +[2026-06-17 17:39:17,518][starVLA.training.train_starvla][INFO] - Step 718, Loss: {'train/loss': 0.5949299335479736, 'train/grad_norm_pre_clip': 6.875156402587891, 'timing/data': 0.0007899841293692589, 'timing/model': 1.0533301820978522, 'learning_rate/qwen_vl_interface': 9.813200865607034e-06, 'learning_rate/action_model': 9.813200865607034e-05, 'epoch': 0.26}) +[2026-06-17 17:39:21,835][starVLA.training.train_starvla][INFO] - Step 719, Loss: {'train/loss': 0.8504424095153809, 'train/grad_norm_pre_clip': 9.360286712646484, 'timing/data': 0.0011067893356084824, 'timing/model': 1.040527575649321, 'learning_rate/qwen_vl_interface': 9.812599858516782e-06, 'learning_rate/action_model': 9.812599858516782e-05, 'epoch': 0.26}) +[2026-06-17 17:39:30,031][starVLA.training.train_starvla][INFO] - Step 720, Loss: {'train/loss': 1.0600913763046265, 'train/grad_norm_pre_clip': 11.434370040893555, 'timing/data': 0.0010718274861574173, 'timing/model': 1.7542515164241195, 'learning_rate/qwen_vl_interface': 9.81199790559467e-06, 'learning_rate/action_model': 9.811997905594668e-05, 'epoch': 0.26}) +[2026-06-17 17:39:37,270][starVLA.training.train_starvla][INFO] - Step 721, Loss: {'train/loss': 1.4499354362487793, 'train/grad_norm_pre_clip': 12.17367935180664, 'timing/data': 0.0007769390940666199, 'timing/model': 1.3061139518395066, 'learning_rate/qwen_vl_interface': 9.811395006965481e-06, 'learning_rate/action_model': 9.81139500696548e-05, 'epoch': 0.26}) +[2026-06-17 17:39:41,362][starVLA.training.train_starvla][INFO] - Step 722, Loss: {'train/loss': 0.7775442600250244, 'train/grad_norm_pre_clip': 9.661409378051758, 'timing/data': 0.0008646706119179726, 'timing/model': 1.0251101134344935, 'learning_rate/qwen_vl_interface': 9.810791162754199e-06, 'learning_rate/action_model': 9.810791162754198e-05, 'epoch': 0.26}) +[2026-06-17 17:39:45,699][starVLA.training.train_starvla][INFO] - Step 723, Loss: {'train/loss': 1.4493986368179321, 'train/grad_norm_pre_clip': 14.599129676818848, 'timing/data': 0.0013522179797291756, 'timing/model': 1.0654211696237326, 'learning_rate/qwen_vl_interface': 9.810186373086002e-06, 'learning_rate/action_model': 9.810186373086002e-05, 'epoch': 0.26}) +[2026-06-17 17:39:50,053][starVLA.training.train_starvla][INFO] - Step 724, Loss: {'train/loss': 0.9672217965126038, 'train/grad_norm_pre_clip': 8.900249481201172, 'timing/data': 0.0010323049500584602, 'timing/model': 1.017239891923964, 'learning_rate/qwen_vl_interface': 9.809580638086264e-06, 'learning_rate/action_model': 9.809580638086263e-05, 'epoch': 0.27}) +[2026-06-17 17:39:54,520][starVLA.training.train_starvla][INFO] - Step 725, Loss: {'train/loss': 0.894027829170227, 'train/grad_norm_pre_clip': 8.71680736541748, 'timing/data': 0.0011992789804935455, 'timing/model': 1.0440823258832097, 'learning_rate/qwen_vl_interface': 9.80897395788055e-06, 'learning_rate/action_model': 9.808973957880551e-05, 'epoch': 0.27}) +[2026-06-17 17:40:01,200][starVLA.training.train_starvla][INFO] - Step 726, Loss: {'train/loss': 1.0495864152908325, 'train/grad_norm_pre_clip': 12.117110252380371, 'timing/data': 0.0008808616548776627, 'timing/model': 3.695077445358038, 'learning_rate/qwen_vl_interface': 9.808366332594632e-06, 'learning_rate/action_model': 9.808366332594632e-05, 'epoch': 0.27}) +[2026-06-17 17:40:13,017][starVLA.training.train_starvla][INFO] - Step 727, Loss: {'train/loss': 0.909386932849884, 'train/grad_norm_pre_clip': 9.132278442382812, 'timing/data': 0.0008964361622929573, 'timing/model': 5.611837657168508, 'learning_rate/qwen_vl_interface': 9.807757762354467e-06, 'learning_rate/action_model': 9.807757762354467e-05, 'epoch': 0.27}) +[2026-06-17 17:40:19,030][starVLA.training.train_starvla][INFO] - Step 728, Loss: {'train/loss': 0.772027313709259, 'train/grad_norm_pre_clip': 8.225181579589844, 'timing/data': 0.0008731856942176819, 'timing/model': 1.073910360224545, 'learning_rate/qwen_vl_interface': 9.807148247286215e-06, 'learning_rate/action_model': 9.807148247286215e-05, 'epoch': 0.27}) +[2026-06-17 17:40:28,218][starVLA.training.train_starvla][INFO] - Step 729, Loss: {'train/loss': 0.9916606545448303, 'train/grad_norm_pre_clip': 9.231536865234375, 'timing/data': 0.0010084696114063263, 'timing/model': 1.0856058858335018, 'learning_rate/qwen_vl_interface': 9.806537787516226e-06, 'learning_rate/action_model': 9.806537787516226e-05, 'epoch': 0.27}) +[2026-06-17 17:40:33,802][starVLA.training.train_starvla][INFO] - Step 730, Loss: {'train/loss': 1.0854426622390747, 'train/grad_norm_pre_clip': 9.33370304107666, 'timing/data': 0.0010311184450984001, 'timing/model': 1.7794629065319896, 'learning_rate/qwen_vl_interface': 9.80592638317105e-06, 'learning_rate/action_model': 9.80592638317105e-05, 'epoch': 0.27}) +[2026-06-17 17:40:42,619][starVLA.training.train_starvla][INFO] - Step 731, Loss: {'train/loss': 1.3581788539886475, 'train/grad_norm_pre_clip': 11.920809745788574, 'timing/data': 0.00098478514701128, 'timing/model': 1.4012823076918721, 'learning_rate/qwen_vl_interface': 9.805314034377434e-06, 'learning_rate/action_model': 9.805314034377434e-05, 'epoch': 0.27}) +[2026-06-17 17:40:49,208][starVLA.training.train_starvla][INFO] - Step 732, Loss: {'train/loss': 0.7252365350723267, 'train/grad_norm_pre_clip': 10.609944343566895, 'timing/data': 0.0009903982281684875, 'timing/model': 1.1491003027185798, 'learning_rate/qwen_vl_interface': 9.804700741262318e-06, 'learning_rate/action_model': 9.804700741262317e-05, 'epoch': 0.27}) +[2026-06-17 17:40:56,212][starVLA.training.train_starvla][INFO] - Step 733, Loss: {'train/loss': 0.9599598050117493, 'train/grad_norm_pre_clip': 31.050918579101562, 'timing/data': 0.007869623601436615, 'timing/model': 1.609773537144065, 'learning_rate/qwen_vl_interface': 9.804086503952835e-06, 'learning_rate/action_model': 9.804086503952835e-05, 'epoch': 0.27}) +[2026-06-17 17:41:03,939][starVLA.training.train_starvla][INFO] - Step 734, Loss: {'train/loss': 0.9634745717048645, 'train/grad_norm_pre_clip': 10.874214172363281, 'timing/data': 0.001524028368294239, 'timing/model': 1.070187228731811, 'learning_rate/qwen_vl_interface': 9.803471322576319e-06, 'learning_rate/action_model': 9.803471322576319e-05, 'epoch': 0.27}) +[2026-06-17 17:41:08,535][starVLA.training.train_starvla][INFO] - Step 735, Loss: {'train/loss': 0.8851491808891296, 'train/grad_norm_pre_clip': 14.293601036071777, 'timing/data': 0.0012671370059251785, 'timing/model': 1.0302641596645117, 'learning_rate/qwen_vl_interface': 9.8028551972603e-06, 'learning_rate/action_model': 9.802855197260299e-05, 'epoch': 0.27}) +[2026-06-17 17:41:13,104][starVLA.training.train_starvla][INFO] - Step 736, Loss: {'train/loss': 1.0087205171585083, 'train/grad_norm_pre_clip': 9.885668754577637, 'timing/data': 0.0008610785007476807, 'timing/model': 1.4749314785003662, 'learning_rate/qwen_vl_interface': 9.802238128132498e-06, 'learning_rate/action_model': 9.802238128132498e-05, 'epoch': 0.27}) +[2026-06-17 17:41:17,532][starVLA.training.train_starvla][INFO] - Step 737, Loss: {'train/loss': 0.952205240726471, 'train/grad_norm_pre_clip': 12.65179443359375, 'timing/data': 0.0011107642203569412, 'timing/model': 1.1672147791832685, 'learning_rate/qwen_vl_interface': 9.801620115320836e-06, 'learning_rate/action_model': 9.801620115320835e-05, 'epoch': 0.27}) +[2026-06-17 17:41:21,928][starVLA.training.train_starvla][INFO] - Step 738, Loss: {'train/loss': 1.132841944694519, 'train/grad_norm_pre_clip': 13.166354179382324, 'timing/data': 0.0015639783814549446, 'timing/model': 1.122265757061541, 'learning_rate/qwen_vl_interface': 9.801001158953425e-06, 'learning_rate/action_model': 9.801001158953423e-05, 'epoch': 0.27}) +[2026-06-17 17:41:26,533][starVLA.training.train_starvla][INFO] - Step 739, Loss: {'train/loss': 1.0986279249191284, 'train/grad_norm_pre_clip': 14.05884075164795, 'timing/data': 0.0011096429079771042, 'timing/model': 1.1169944247230887, 'learning_rate/qwen_vl_interface': 9.800381259158575e-06, 'learning_rate/action_model': 9.800381259158574e-05, 'epoch': 0.27}) +[2026-06-17 17:41:30,890][starVLA.training.train_starvla][INFO] - Step 740, Loss: {'train/loss': 0.7211622595787048, 'train/grad_norm_pre_clip': 10.255760192871094, 'timing/data': 0.001142638735473156, 'timing/model': 1.34989620372653, 'learning_rate/qwen_vl_interface': 9.799760416064794e-06, 'learning_rate/action_model': 9.799760416064793e-05, 'epoch': 0.27}) +[2026-06-17 17:41:35,386][starVLA.training.train_starvla][INFO] - Step 741, Loss: {'train/loss': 0.7775458097457886, 'train/grad_norm_pre_clip': 10.427228927612305, 'timing/data': 0.0021373387426137924, 'timing/model': 1.3689973298460245, 'learning_rate/qwen_vl_interface': 9.799138629800785e-06, 'learning_rate/action_model': 9.799138629800785e-05, 'epoch': 0.27}) +[2026-06-17 17:41:39,643][starVLA.training.train_starvla][INFO] - Step 742, Loss: {'train/loss': 1.0343542098999023, 'train/grad_norm_pre_clip': 9.093170166015625, 'timing/data': 0.0008648373186588287, 'timing/model': 1.1467233328148723, 'learning_rate/qwen_vl_interface': 9.798515900495441e-06, 'learning_rate/action_model': 9.79851590049544e-05, 'epoch': 0.27}) +[2026-06-17 17:41:43,952][starVLA.training.train_starvla][INFO] - Step 743, Loss: {'train/loss': 1.2518552541732788, 'train/grad_norm_pre_clip': 10.580398559570312, 'timing/data': 0.0010118735954165459, 'timing/model': 1.0382505673915148, 'learning_rate/qwen_vl_interface': 9.797892228277857e-06, 'learning_rate/action_model': 9.797892228277857e-05, 'epoch': 0.27}) +[2026-06-17 17:41:48,424][starVLA.training.train_starvla][INFO] - Step 744, Loss: {'train/loss': 1.2469878196716309, 'train/grad_norm_pre_clip': 14.618731498718262, 'timing/data': 0.0011328579857945442, 'timing/model': 1.0487964795902371, 'learning_rate/qwen_vl_interface': 9.797267613277319e-06, 'learning_rate/action_model': 9.797267613277319e-05, 'epoch': 0.27}) +[2026-06-17 17:41:52,798][starVLA.training.train_starvla][INFO] - Step 745, Loss: {'train/loss': 1.3872425556182861, 'train/grad_norm_pre_clip': 14.338205337524414, 'timing/data': 0.0009360592812299728, 'timing/model': 1.035872577689588, 'learning_rate/qwen_vl_interface': 9.796642055623314e-06, 'learning_rate/action_model': 9.796642055623314e-05, 'epoch': 0.27}) +[2026-06-17 17:41:57,239][starVLA.training.train_starvla][INFO] - Step 746, Loss: {'train/loss': 1.1069334745407104, 'train/grad_norm_pre_clip': 12.492823600769043, 'timing/data': 0.0008060522377490997, 'timing/model': 1.377203168347478, 'learning_rate/qwen_vl_interface': 9.796015555445518e-06, 'learning_rate/action_model': 9.796015555445517e-05, 'epoch': 0.27}) +[2026-06-17 17:42:01,365][starVLA.training.train_starvla][INFO] - Step 747, Loss: {'train/loss': 0.9008079767227173, 'train/grad_norm_pre_clip': 10.265508651733398, 'timing/data': 0.0009103072807192802, 'timing/model': 1.0408579651266336, 'learning_rate/qwen_vl_interface': 9.795388112873806e-06, 'learning_rate/action_model': 9.795388112873806e-05, 'epoch': 0.27}) +[2026-06-17 17:42:05,811][starVLA.training.train_starvla][INFO] - Step 748, Loss: {'train/loss': 1.0108134746551514, 'train/grad_norm_pre_clip': 16.979249954223633, 'timing/data': 0.0012351907789707184, 'timing/model': 1.0497544361278415, 'learning_rate/qwen_vl_interface': 9.794759728038246e-06, 'learning_rate/action_model': 9.794759728038246e-05, 'epoch': 0.27}) +[2026-06-17 17:42:10,433][starVLA.training.train_starvla][INFO] - Step 749, Loss: {'train/loss': 0.8663166165351868, 'train/grad_norm_pre_clip': 9.503520011901855, 'timing/data': 0.0010321279987692833, 'timing/model': 1.1373567869886756, 'learning_rate/qwen_vl_interface': 9.794130401069105e-06, 'learning_rate/action_model': 9.794130401069104e-05, 'epoch': 0.27}) +[2026-06-17 17:42:19,433][starVLA.training.train_starvla][INFO] - Step 750, Loss: {'train/loss': 1.0091519355773926, 'train/grad_norm_pre_clip': 12.061758995056152, 'timing/data': 0.001346871256828308, 'timing/model': 1.0863442197442055, 'learning_rate/qwen_vl_interface': 9.79350013209684e-06, 'learning_rate/action_model': 9.79350013209684e-05, 'epoch': 0.27}) +[2026-06-17 17:42:24,633][starVLA.training.train_starvla][INFO] - Step 751, Loss: {'train/loss': 0.8499398827552795, 'train/grad_norm_pre_clip': 12.398650169372559, 'timing/data': 0.0010969750583171844, 'timing/model': 1.4611258944496512, 'learning_rate/qwen_vl_interface': 9.792868921252112e-06, 'learning_rate/action_model': 9.792868921252111e-05, 'epoch': 0.27}) +[2026-06-17 17:42:28,855][starVLA.training.train_starvla][INFO] - Step 752, Loss: {'train/loss': 0.8704848289489746, 'train/grad_norm_pre_clip': 8.6002197265625, 'timing/data': 0.0009241830557584763, 'timing/model': 1.1114659272134304, 'learning_rate/qwen_vl_interface': 9.792236768665765e-06, 'learning_rate/action_model': 9.792236768665764e-05, 'epoch': 0.28}) +[2026-06-17 17:42:36,020][starVLA.training.train_starvla][INFO] - Step 753, Loss: {'train/loss': 0.9341263771057129, 'train/grad_norm_pre_clip': 36.04135513305664, 'timing/data': 0.0011273883283138275, 'timing/model': 2.611876674927771, 'learning_rate/qwen_vl_interface': 9.79160367446885e-06, 'learning_rate/action_model': 9.79160367446885e-05, 'epoch': 0.28}) +[2026-06-17 17:42:40,595][starVLA.training.train_starvla][INFO] - Step 754, Loss: {'train/loss': 1.34199857711792, 'train/grad_norm_pre_clip': 12.224617004394531, 'timing/data': 0.002140018157660961, 'timing/model': 1.1053048018366098, 'learning_rate/qwen_vl_interface': 9.790969638792606e-06, 'learning_rate/action_model': 9.790969638792606e-05, 'epoch': 0.28}) +[2026-06-17 17:42:48,606][starVLA.training.train_starvla][INFO] - Step 755, Loss: {'train/loss': 0.9790046811103821, 'train/grad_norm_pre_clip': 14.020153045654297, 'timing/data': 0.0008908957242965698, 'timing/model': 4.493783997371793, 'learning_rate/qwen_vl_interface': 9.790334661768471e-06, 'learning_rate/action_model': 9.79033466176847e-05, 'epoch': 0.28}) +[2026-06-17 17:42:56,799][starVLA.training.train_starvla][INFO] - Step 756, Loss: {'train/loss': 1.0990025997161865, 'train/grad_norm_pre_clip': 10.634258270263672, 'timing/data': 0.0017255255952477455, 'timing/model': 4.561367218382657, 'learning_rate/qwen_vl_interface': 9.78969874352807e-06, 'learning_rate/action_model': 9.78969874352807e-05, 'epoch': 0.28}) +[2026-06-17 17:43:01,705][starVLA.training.train_starvla][INFO] - Step 757, Loss: {'train/loss': 1.023272156715393, 'train/grad_norm_pre_clip': 8.411210060119629, 'timing/data': 0.0010374365374445915, 'timing/model': 1.2534840004518628, 'learning_rate/qwen_vl_interface': 9.78906188420324e-06, 'learning_rate/action_model': 9.78906188420324e-05, 'epoch': 0.28}) +[2026-06-17 17:43:07,595][starVLA.training.train_starvla][INFO] - Step 758, Loss: {'train/loss': 0.9584890007972717, 'train/grad_norm_pre_clip': 22.367525100708008, 'timing/data': 0.000977853313088417, 'timing/model': 1.0473263626918197, 'learning_rate/qwen_vl_interface': 9.788424083925996e-06, 'learning_rate/action_model': 9.788424083925995e-05, 'epoch': 0.28}) +[2026-06-17 17:43:12,273][starVLA.training.train_starvla][INFO] - Step 759, Loss: {'train/loss': 0.8135600090026855, 'train/grad_norm_pre_clip': 10.330526351928711, 'timing/data': 0.0009842552244663239, 'timing/model': 1.227677726186812, 'learning_rate/qwen_vl_interface': 9.787785342828554e-06, 'learning_rate/action_model': 9.787785342828553e-05, 'epoch': 0.28}) +[2026-06-17 17:43:16,923][starVLA.training.train_starvla][INFO] - Step 760, Loss: {'train/loss': 0.8498422503471375, 'train/grad_norm_pre_clip': 9.151909828186035, 'timing/data': 0.0014078523963689804, 'timing/model': 1.5173851680010557, 'learning_rate/qwen_vl_interface': 9.78714566104333e-06, 'learning_rate/action_model': 9.787145661043329e-05, 'epoch': 0.28}) +[2026-06-17 17:43:21,449][starVLA.training.train_starvla][INFO] - Step 761, Loss: {'train/loss': 0.90440833568573, 'train/grad_norm_pre_clip': 9.508562088012695, 'timing/data': 0.001085522584617138, 'timing/model': 1.386461122892797, 'learning_rate/qwen_vl_interface': 9.786505038702926e-06, 'learning_rate/action_model': 9.786505038702925e-05, 'epoch': 0.28}) +[2026-06-17 17:43:25,884][starVLA.training.train_starvla][INFO] - Step 762, Loss: {'train/loss': 0.7663896083831787, 'train/grad_norm_pre_clip': 9.644837379455566, 'timing/data': 0.0009160023182630539, 'timing/model': 1.193868805654347, 'learning_rate/qwen_vl_interface': 9.785863475940147e-06, 'learning_rate/action_model': 9.785863475940145e-05, 'epoch': 0.28}) +[2026-06-17 17:43:30,341][starVLA.training.train_starvla][INFO] - Step 763, Loss: {'train/loss': 1.4400495290756226, 'train/grad_norm_pre_clip': 12.910411834716797, 'timing/data': 0.0009471885859966278, 'timing/model': 1.0843345867469907, 'learning_rate/qwen_vl_interface': 9.785220972887988e-06, 'learning_rate/action_model': 9.785220972887989e-05, 'epoch': 0.28}) +[2026-06-17 17:43:34,957][starVLA.training.train_starvla][INFO] - Step 764, Loss: {'train/loss': 0.9626464247703552, 'train/grad_norm_pre_clip': 10.365528106689453, 'timing/data': 0.0008730031549930573, 'timing/model': 1.0582462809979916, 'learning_rate/qwen_vl_interface': 9.784577529679643e-06, 'learning_rate/action_model': 9.784577529679643e-05, 'epoch': 0.28}) +[2026-06-17 17:43:40,541][starVLA.training.train_starvla][INFO] - Step 765, Loss: {'train/loss': 1.0999739170074463, 'train/grad_norm_pre_clip': 11.13131046295166, 'timing/data': 0.0008619707077741623, 'timing/model': 2.208722115494311, 'learning_rate/qwen_vl_interface': 9.783933146448495e-06, 'learning_rate/action_model': 9.783933146448495e-05, 'epoch': 0.28}) +[2026-06-17 17:43:49,229][starVLA.training.train_starvla][INFO] - Step 766, Loss: {'train/loss': 0.7771046161651611, 'train/grad_norm_pre_clip': 8.15423583984375, 'timing/data': 0.0013623833656311035, 'timing/model': 1.4220481412485242, 'learning_rate/qwen_vl_interface': 9.783287823328127e-06, 'learning_rate/action_model': 9.783287823328127e-05, 'epoch': 0.28}) +[2026-06-17 17:43:53,378][starVLA.training.train_starvla][INFO] - Step 767, Loss: {'train/loss': 0.7309585809707642, 'train/grad_norm_pre_clip': 7.884757041931152, 'timing/data': 0.0013695638626813889, 'timing/model': 1.1237294906750321, 'learning_rate/qwen_vl_interface': 9.782641560452317e-06, 'learning_rate/action_model': 9.782641560452316e-05, 'epoch': 0.28}) +[2026-06-17 17:44:05,338][starVLA.training.train_starvla][INFO] - Step 768, Loss: {'train/loss': 1.1221579313278198, 'train/grad_norm_pre_clip': 10.607842445373535, 'timing/data': 0.0011769086122512817, 'timing/model': 1.0556794926524162, 'learning_rate/qwen_vl_interface': 9.781994357955032e-06, 'learning_rate/action_model': 9.781994357955032e-05, 'epoch': 0.28}) +[2026-06-17 17:44:10,196][starVLA.training.train_starvla][INFO] - Step 769, Loss: {'train/loss': 1.106516718864441, 'train/grad_norm_pre_clip': 9.473536491394043, 'timing/data': 0.000867762602865696, 'timing/model': 1.0521500445902348, 'learning_rate/qwen_vl_interface': 9.78134621597044e-06, 'learning_rate/action_model': 9.781346215970441e-05, 'epoch': 0.28}) +[2026-06-17 17:44:14,960][starVLA.training.train_starvla][INFO] - Step 770, Loss: {'train/loss': 0.918144702911377, 'train/grad_norm_pre_clip': 8.223576545715332, 'timing/data': 0.0010668272152543068, 'timing/model': 1.1686676517128944, 'learning_rate/qwen_vl_interface': 9.780697134632903e-06, 'learning_rate/action_model': 9.780697134632902e-05, 'epoch': 0.28}) +[2026-06-17 17:44:19,670][starVLA.training.train_starvla][INFO] - Step 771, Loss: {'train/loss': 0.7398858666419983, 'train/grad_norm_pre_clip': 16.08806610107422, 'timing/data': 0.001070961356163025, 'timing/model': 1.5332339899614453, 'learning_rate/qwen_vl_interface': 9.780047114076974e-06, 'learning_rate/action_model': 9.780047114076973e-05, 'epoch': 0.28}) +[2026-06-17 17:44:24,012][starVLA.training.train_starvla][INFO] - Step 772, Loss: {'train/loss': 0.8870595097541809, 'train/grad_norm_pre_clip': 10.287087440490723, 'timing/data': 0.0009836778044700623, 'timing/model': 1.1437364760786295, 'learning_rate/qwen_vl_interface': 9.779396154437404e-06, 'learning_rate/action_model': 9.779396154437403e-05, 'epoch': 0.28}) +[2026-06-17 17:44:28,728][starVLA.training.train_starvla][INFO] - Step 773, Loss: {'train/loss': 0.8339181542396545, 'train/grad_norm_pre_clip': 14.252846717834473, 'timing/data': 0.0009963903576135635, 'timing/model': 1.0598951168358326, 'learning_rate/qwen_vl_interface': 9.778744255849138e-06, 'learning_rate/action_model': 9.778744255849138e-05, 'epoch': 0.28}) +[2026-06-17 17:44:33,429][starVLA.training.train_starvla][INFO] - Step 774, Loss: {'train/loss': 1.2986141443252563, 'train/grad_norm_pre_clip': 12.761898040771484, 'timing/data': 0.0011004293337464333, 'timing/model': 1.1570403352379799, 'learning_rate/qwen_vl_interface': 9.778091418447313e-06, 'learning_rate/action_model': 9.778091418447312e-05, 'epoch': 0.28}) +[2026-06-17 17:44:45,641][starVLA.training.train_starvla][INFO] - Step 775, Loss: {'train/loss': 1.238217830657959, 'train/grad_norm_pre_clip': 12.18608283996582, 'timing/data': 0.0012751184403896332, 'timing/model': 5.059095500968397, 'learning_rate/qwen_vl_interface': 9.777437642367265e-06, 'learning_rate/action_model': 9.777437642367264e-05, 'epoch': 0.28}) +[2026-06-17 17:44:52,040][starVLA.training.train_starvla][INFO] - Step 776, Loss: {'train/loss': 0.7009708881378174, 'train/grad_norm_pre_clip': 9.275503158569336, 'timing/data': 0.001773330383002758, 'timing/model': 1.3821062734350562, 'learning_rate/qwen_vl_interface': 9.776782927744523e-06, 'learning_rate/action_model': 9.776782927744522e-05, 'epoch': 0.28}) +[2026-06-17 17:44:56,228][starVLA.training.train_starvla][INFO] - Step 777, Loss: {'train/loss': 1.0382426977157593, 'train/grad_norm_pre_clip': 9.146988868713379, 'timing/data': 0.001255583018064499, 'timing/model': 1.0622385805472732, 'learning_rate/qwen_vl_interface': 9.776127274714808e-06, 'learning_rate/action_model': 9.776127274714807e-05, 'epoch': 0.28}) +[2026-06-17 17:45:01,049][starVLA.training.train_starvla][INFO] - Step 778, Loss: {'train/loss': 0.842518150806427, 'train/grad_norm_pre_clip': 11.363494873046875, 'timing/data': 0.0010844497010111809, 'timing/model': 1.1849072985351086, 'learning_rate/qwen_vl_interface': 9.775470683414038e-06, 'learning_rate/action_model': 9.775470683414038e-05, 'epoch': 0.28}) +[2026-06-17 17:45:06,145][starVLA.training.train_starvla][INFO] - Step 779, Loss: {'train/loss': 0.7574163675308228, 'train/grad_norm_pre_clip': 8.839938163757324, 'timing/data': 0.0010495949536561966, 'timing/model': 1.033917005173862, 'learning_rate/qwen_vl_interface': 9.774813153978327e-06, 'learning_rate/action_model': 9.774813153978327e-05, 'epoch': 0.29}) +[2026-06-17 17:45:10,714][starVLA.training.train_starvla][INFO] - Step 780, Loss: {'train/loss': 1.3105969429016113, 'train/grad_norm_pre_clip': 12.402137756347656, 'timing/data': 0.0011422047391533852, 'timing/model': 1.1166358767077327, 'learning_rate/qwen_vl_interface': 9.77415468654398e-06, 'learning_rate/action_model': 9.77415468654398e-05, 'epoch': 0.29}) +[2026-06-17 17:45:15,178][starVLA.training.train_starvla][INFO] - Step 781, Loss: {'train/loss': 0.9074145555496216, 'train/grad_norm_pre_clip': 10.009857177734375, 'timing/data': 0.0016715507954359055, 'timing/model': 1.3966072984039783, 'learning_rate/qwen_vl_interface': 9.773495281247497e-06, 'learning_rate/action_model': 9.773495281247497e-05, 'epoch': 0.29}) +[2026-06-17 17:45:23,939][starVLA.training.train_starvla][INFO] - Step 782, Loss: {'train/loss': 1.2733145952224731, 'train/grad_norm_pre_clip': 12.419938087463379, 'timing/data': 0.0019243955612182617, 'timing/model': 1.3617253387346864, 'learning_rate/qwen_vl_interface': 9.772834938225577e-06, 'learning_rate/action_model': 9.772834938225576e-05, 'epoch': 0.29}) +[2026-06-17 17:45:32,049][starVLA.training.train_starvla][INFO] - Step 783, Loss: {'train/loss': 1.3285064697265625, 'train/grad_norm_pre_clip': 16.820711135864258, 'timing/data': 0.0011735660955309868, 'timing/model': 1.1926368055865169, 'learning_rate/qwen_vl_interface': 9.772173657615107e-06, 'learning_rate/action_model': 9.772173657615106e-05, 'epoch': 0.29}) +[2026-06-17 17:45:36,467][starVLA.training.train_starvla][INFO] - Step 784, Loss: {'train/loss': 0.7674809098243713, 'train/grad_norm_pre_clip': 8.341108322143555, 'timing/data': 0.0009764302521944046, 'timing/model': 1.042935430072248, 'learning_rate/qwen_vl_interface': 9.77151143955317e-06, 'learning_rate/action_model': 9.77151143955317e-05, 'epoch': 0.29}) +[2026-06-17 17:45:43,739][starVLA.training.train_starvla][INFO] - Step 785, Loss: {'train/loss': 0.8684049844741821, 'train/grad_norm_pre_clip': 10.557819366455078, 'timing/data': 0.0013700192794203758, 'timing/model': 1.2427839916199446, 'learning_rate/qwen_vl_interface': 9.770848284177048e-06, 'learning_rate/action_model': 9.770848284177047e-05, 'epoch': 0.29}) +[2026-06-17 17:45:48,380][starVLA.training.train_starvla][INFO] - Step 786, Loss: {'train/loss': 1.0107344388961792, 'train/grad_norm_pre_clip': 9.191425323486328, 'timing/data': 0.0010760575532913208, 'timing/model': 1.363161887973547, 'learning_rate/qwen_vl_interface': 9.77018419162421e-06, 'learning_rate/action_model': 9.77018419162421e-05, 'epoch': 0.29}) +[2026-06-17 17:45:52,894][starVLA.training.train_starvla][INFO] - Step 787, Loss: {'train/loss': 0.8915736675262451, 'train/grad_norm_pre_clip': 12.479645729064941, 'timing/data': 0.0009084371849894524, 'timing/model': 1.0274695483967662, 'learning_rate/qwen_vl_interface': 9.769519162032328e-06, 'learning_rate/action_model': 9.769519162032327e-05, 'epoch': 0.29}) +[2026-06-17 17:45:57,524][starVLA.training.train_starvla][INFO] - Step 788, Loss: {'train/loss': 0.8409537076950073, 'train/grad_norm_pre_clip': 8.963459968566895, 'timing/data': 0.0010175108909606934, 'timing/model': 1.1470369324088097, 'learning_rate/qwen_vl_interface': 9.768853195539259e-06, 'learning_rate/action_model': 9.768853195539258e-05, 'epoch': 0.29}) +[2026-06-17 17:46:02,732][starVLA.training.train_starvla][INFO] - Step 789, Loss: {'train/loss': 0.7390252351760864, 'train/grad_norm_pre_clip': 9.587227821350098, 'timing/data': 0.0014139749109745026, 'timing/model': 1.7220488209277391, 'learning_rate/qwen_vl_interface': 9.76818629228306e-06, 'learning_rate/action_model': 9.768186292283059e-05, 'epoch': 0.29}) +[2026-06-17 17:46:09,648][starVLA.training.train_starvla][INFO] - Step 790, Loss: {'train/loss': 0.7268503308296204, 'train/grad_norm_pre_clip': 9.076427459716797, 'timing/data': 0.000856923870742321, 'timing/model': 1.0688830446451902, 'learning_rate/qwen_vl_interface': 9.767518452401981e-06, 'learning_rate/action_model': 9.76751845240198e-05, 'epoch': 0.29}) +[2026-06-17 17:46:14,553][starVLA.training.train_starvla][INFO] - Step 791, Loss: {'train/loss': 1.5531774759292603, 'train/grad_norm_pre_clip': 13.929156303405762, 'timing/data': 0.0011275671422481537, 'timing/model': 1.577530624344945, 'learning_rate/qwen_vl_interface': 9.766849676034465e-06, 'learning_rate/action_model': 9.766849676034464e-05, 'epoch': 0.29}) +[2026-06-17 17:46:18,757][starVLA.training.train_starvla][INFO] - Step 792, Loss: {'train/loss': 0.861882209777832, 'train/grad_norm_pre_clip': 9.803342819213867, 'timing/data': 0.0010645901784300804, 'timing/model': 1.0215178290382028, 'learning_rate/qwen_vl_interface': 9.766179963319152e-06, 'learning_rate/action_model': 9.76617996331915e-05, 'epoch': 0.29}) +[2026-06-17 17:46:23,179][starVLA.training.train_starvla][INFO] - Step 793, Loss: {'train/loss': 1.2272168397903442, 'train/grad_norm_pre_clip': 11.39306640625, 'timing/data': 0.0009827418252825737, 'timing/model': 1.0231542875990272, 'learning_rate/qwen_vl_interface': 9.765509314394872e-06, 'learning_rate/action_model': 9.765509314394872e-05, 'epoch': 0.29}) +[2026-06-17 17:46:27,699][starVLA.training.train_starvla][INFO] - Step 794, Loss: {'train/loss': 1.3947641849517822, 'train/grad_norm_pre_clip': 14.801368713378906, 'timing/data': 0.0017626723274588585, 'timing/model': 1.0502121997997165, 'learning_rate/qwen_vl_interface': 9.764837729400652e-06, 'learning_rate/action_model': 9.764837729400653e-05, 'epoch': 0.29}) +[2026-06-17 17:46:32,203][starVLA.training.train_starvla][INFO] - Step 795, Loss: {'train/loss': 0.8567902445793152, 'train/grad_norm_pre_clip': 49.783287048339844, 'timing/data': 0.000982784666121006, 'timing/model': 1.4492477914318442, 'learning_rate/qwen_vl_interface': 9.764165208475714e-06, 'learning_rate/action_model': 9.764165208475714e-05, 'epoch': 0.29}) +[2026-06-17 17:46:36,863][starVLA.training.train_starvla][INFO] - Step 796, Loss: {'train/loss': 1.0134282112121582, 'train/grad_norm_pre_clip': 13.613147735595703, 'timing/data': 0.0010702358558773994, 'timing/model': 1.3832745105028152, 'learning_rate/qwen_vl_interface': 9.76349175175947e-06, 'learning_rate/action_model': 9.763491751759469e-05, 'epoch': 0.29}) +[2026-06-17 17:46:41,403][starVLA.training.train_starvla][INFO] - Step 797, Loss: {'train/loss': 0.8856855630874634, 'train/grad_norm_pre_clip': 11.757887840270996, 'timing/data': 0.0009078970178961754, 'timing/model': 1.0918447999283671, 'learning_rate/qwen_vl_interface': 9.76281735939153e-06, 'learning_rate/action_model': 9.762817359391529e-05, 'epoch': 0.29}) +[2026-06-17 17:46:45,955][starVLA.training.train_starvla][INFO] - Step 798, Loss: {'train/loss': 1.042588472366333, 'train/grad_norm_pre_clip': 9.787665367126465, 'timing/data': 0.0011918330565094948, 'timing/model': 1.1443531746044755, 'learning_rate/qwen_vl_interface': 9.762142031511695e-06, 'learning_rate/action_model': 9.762142031511695e-05, 'epoch': 0.29}) +[2026-06-17 17:46:52,416][starVLA.training.train_starvla][INFO] - Step 799, Loss: {'train/loss': 0.9242709875106812, 'train/grad_norm_pre_clip': 11.772394180297852, 'timing/data': 0.0008148103952407837, 'timing/model': 1.2462065545842052, 'learning_rate/qwen_vl_interface': 9.76146576825996e-06, 'learning_rate/action_model': 9.761465768259959e-05, 'epoch': 0.29}) +[2026-06-17 17:50:13,349][starVLA.training.train_starvla][INFO] - Step 800, Loss: {'train/loss': 1.0798660516738892, 'train/grad_norm_pre_clip': 17.566072463989258, 'eval/loss': 0.9643787145614624, 'eval/latency_2/loss': 0.9643787145614624, 'eval/demon_fire_precision': 0.8301886792452831, 'eval/demon_fire_recall': 0.9573221757322176, 'eval/demon_fire_f1': 0.8892343567819666, 'eval/demon_fire_rate_model': 0.766833611574847, 'eval/demon_fire_rate_teacher': 0.6649972175848636, 'eval/demon_left_precision': 0.7494305239179955, 'eval/demon_left_recall': 0.6926315789473684, 'eval/demon_left_f1': 0.7199124726477024, 'eval/demon_left_rate_model': 0.2442960489705064, 'eval/demon_left_rate_teacher': 0.2643294379521425, 'eval/demon_right_precision': 0.75, 'eval/demon_right_recall': 0.6169491525423729, 'eval/demon_right_f1': 0.6769993800371978, 'eval/demon_right_rate_model': 0.2025598219254313, 'eval/demon_right_rate_teacher': 0.24624373956594323, 'eval/demon_move_f1': 0.6984559263424501, 'eval/demon_cc_f1': 0.8218865592240616, 'eval/latency_2/demon_fire_precision': 0.8301886792452831, 'eval/latency_2/demon_fire_recall': 0.9573221757322176, 'eval/latency_2/demon_fire_f1': 0.8892343567819666, 'eval/latency_2/demon_fire_rate_model': 0.766833611574847, 'eval/latency_2/demon_fire_rate_teacher': 0.6649972175848636, 'eval/latency_2/demon_left_precision': 0.7494305239179955, 'eval/latency_2/demon_left_recall': 0.6926315789473684, 'eval/latency_2/demon_left_f1': 0.7199124726477024, 'eval/latency_2/demon_left_rate_model': 0.2442960489705064, 'eval/latency_2/demon_left_rate_teacher': 0.2643294379521425, 'eval/latency_2/demon_right_precision': 0.75, 'eval/latency_2/demon_right_recall': 0.6169491525423729, 'eval/latency_2/demon_right_f1': 0.6769993800371978, 'eval/latency_2/demon_right_rate_model': 0.2025598219254313, 'eval/latency_2/demon_right_rate_teacher': 0.24624373956594323, 'eval/latency_2/demon_move_f1': 0.6984559263424501, 'eval/latency_2/demon_cc_f1': 0.8218865592240616, 'timing/data': 0.0014303028583526611, 'timing/model': 3.1275605531409383, 'learning_rate/qwen_vl_interface': 9.760788569776518e-06, 'learning_rate/action_model': 9.760788569776517e-05, 'epoch': 0.29}) +[2026-06-17 17:50:13,366][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_800_state +[2026-06-17 17:50:13,384][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 17:50:22,759][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_800_state/model.safetensors +[2026-06-17 17:50:37,635][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_800_state/optimizer.bin +[2026-06-17 17:50:37,636][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_800_state/sampler.bin +[2026-06-17 17:50:37,636][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_800_state/sampler_1.bin +[2026-06-17 17:50:37,639][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_800_state/random_states_0.pkl +[2026-06-17 17:50:37,644][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 800 +[2026-06-17 17:50:37,645][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 17:50:37,662][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 17:50:42,398][starVLA.training.train_starvla][INFO] - Step 801, Loss: {'train/loss': 0.8765641450881958, 'train/grad_norm_pre_clip': 11.483075141906738, 'timing/data': 0.001225275918841362, 'timing/model': 1.1269696336239576, 'learning_rate/qwen_vl_interface': 9.760110436201752e-06, 'learning_rate/action_model': 9.76011043620175e-05, 'epoch': 0.29}) +[2026-06-17 17:50:46,980][starVLA.training.train_starvla][INFO] - Step 802, Loss: {'train/loss': 1.0181868076324463, 'train/grad_norm_pre_clip': 7.845815181732178, 'timing/data': 0.0009551625698804855, 'timing/model': 1.3578812777996063, 'learning_rate/qwen_vl_interface': 9.759431367676238e-06, 'learning_rate/action_model': 9.759431367676237e-05, 'epoch': 0.29}) +[2026-06-17 17:50:51,063][starVLA.training.train_starvla][INFO] - Step 803, Loss: {'train/loss': 1.2125983238220215, 'train/grad_norm_pre_clip': 12.177746772766113, 'timing/data': 0.017240460962057114, 'timing/model': 1.033856519497931, 'learning_rate/qwen_vl_interface': 9.758751364340751e-06, 'learning_rate/action_model': 9.75875136434075e-05, 'epoch': 0.29}) +[2026-06-17 17:51:00,353][starVLA.training.train_starvla][INFO] - Step 804, Loss: {'train/loss': 1.0238114595413208, 'train/grad_norm_pre_clip': 10.81242847442627, 'timing/data': 0.012777803465723991, 'timing/model': 2.3842870369553566, 'learning_rate/qwen_vl_interface': 9.758070426336252e-06, 'learning_rate/action_model': 9.758070426336252e-05, 'epoch': 0.29}) +[2026-06-17 17:51:04,893][starVLA.training.train_starvla][INFO] - Step 805, Loss: {'train/loss': 1.1253331899642944, 'train/grad_norm_pre_clip': 9.07354736328125, 'timing/data': 0.011961800046265125, 'timing/model': 1.089931644499302, 'learning_rate/qwen_vl_interface': 9.757388553803905e-06, 'learning_rate/action_model': 9.757388553803905e-05, 'epoch': 0.29}) +[2026-06-17 17:51:09,322][starVLA.training.train_starvla][INFO] - Step 806, Loss: {'train/loss': 0.9519449472427368, 'train/grad_norm_pre_clip': 8.289199829101562, 'timing/data': 0.011582400649785995, 'timing/model': 1.06421541236341, 'learning_rate/qwen_vl_interface': 9.75670574688506e-06, 'learning_rate/action_model': 9.75670574688506e-05, 'epoch': 0.3}) +[2026-06-17 17:51:17,018][starVLA.training.train_starvla][INFO] - Step 807, Loss: {'train/loss': 1.2427070140838623, 'train/grad_norm_pre_clip': 13.612125396728516, 'timing/data': 0.017777882516384125, 'timing/model': 1.5158641692250967, 'learning_rate/qwen_vl_interface': 9.756022005721265e-06, 'learning_rate/action_model': 9.756022005721265e-05, 'epoch': 0.3}) +[2026-06-17 17:51:22,475][starVLA.training.train_starvla][INFO] - Step 808, Loss: {'train/loss': 0.757327139377594, 'train/grad_norm_pre_clip': 8.975356101989746, 'timing/data': 0.011149157769978046, 'timing/model': 1.0300908731296659, 'learning_rate/qwen_vl_interface': 9.755337330454257e-06, 'learning_rate/action_model': 9.755337330454257e-05, 'epoch': 0.3}) +[2026-06-17 17:51:26,849][starVLA.training.train_starvla][INFO] - Step 809, Loss: {'train/loss': 1.1484814882278442, 'train/grad_norm_pre_clip': 12.1996488571167, 'timing/data': 0.011583341285586357, 'timing/model': 1.0312430588528514, 'learning_rate/qwen_vl_interface': 9.754651721225975e-06, 'learning_rate/action_model': 9.754651721225975e-05, 'epoch': 0.3}) +[2026-06-17 17:51:31,297][starVLA.training.train_starvla][INFO] - Step 810, Loss: {'train/loss': 0.9251454472541809, 'train/grad_norm_pre_clip': 11.604898452758789, 'timing/data': 0.010618667118251324, 'timing/model': 1.052902857773006, 'learning_rate/qwen_vl_interface': 9.753965178178543e-06, 'learning_rate/action_model': 9.753965178178543e-05, 'epoch': 0.3}) +[2026-06-17 17:51:35,834][starVLA.training.train_starvla][INFO] - Step 811, Loss: {'train/loss': 0.9352920055389404, 'train/grad_norm_pre_clip': 9.354071617126465, 'timing/data': 0.01550384797155857, 'timing/model': 1.0293631497770548, 'learning_rate/qwen_vl_interface': 9.753277701454285e-06, 'learning_rate/action_model': 9.753277701454284e-05, 'epoch': 0.3}) +[2026-06-17 17:51:42,533][starVLA.training.train_starvla][INFO] - Step 812, Loss: {'train/loss': 0.9590646028518677, 'train/grad_norm_pre_clip': 8.677351951599121, 'timing/data': 0.01113216858357191, 'timing/model': 2.1646269662305713, 'learning_rate/qwen_vl_interface': 9.752589291195711e-06, 'learning_rate/action_model': 9.75258929119571e-05, 'epoch': 0.3}) +[2026-06-17 17:51:46,576][starVLA.training.train_starvla][INFO] - Step 813, Loss: {'train/loss': 1.0168800354003906, 'train/grad_norm_pre_clip': 8.423524856567383, 'timing/data': 0.014231227338314056, 'timing/model': 1.036867923103273, 'learning_rate/qwen_vl_interface': 9.751899947545534e-06, 'learning_rate/action_model': 9.751899947545533e-05, 'epoch': 0.3}) +[2026-06-17 17:51:50,988][starVLA.training.train_starvla][INFO] - Step 814, Loss: {'train/loss': 0.7552645802497864, 'train/grad_norm_pre_clip': 6.832289218902588, 'timing/data': 0.013029227033257484, 'timing/model': 1.0735816899687052, 'learning_rate/qwen_vl_interface': 9.751209670646653e-06, 'learning_rate/action_model': 9.751209670646652e-05, 'epoch': 0.3}) +[2026-06-17 17:51:55,410][starVLA.training.train_starvla][INFO] - Step 815, Loss: {'train/loss': 1.1263785362243652, 'train/grad_norm_pre_clip': 9.455887794494629, 'timing/data': 0.011097355745732784, 'timing/model': 1.064144122414291, 'learning_rate/qwen_vl_interface': 9.750518460642162e-06, 'learning_rate/action_model': 9.75051846064216e-05, 'epoch': 0.3}) +[2026-06-17 17:51:59,796][starVLA.training.train_starvla][INFO] - Step 816, Loss: {'train/loss': 0.8322159051895142, 'train/grad_norm_pre_clip': 11.25029182434082, 'timing/data': 0.012235763482749462, 'timing/model': 1.13502939324826, 'learning_rate/qwen_vl_interface': 9.749826317675352e-06, 'learning_rate/action_model': 9.749826317675352e-05, 'epoch': 0.3}) +[2026-06-17 17:52:04,614][starVLA.training.train_starvla][INFO] - Step 817, Loss: {'train/loss': 1.0486536026000977, 'train/grad_norm_pre_clip': 10.342031478881836, 'timing/data': 0.012780250050127506, 'timing/model': 1.535865195095539, 'learning_rate/qwen_vl_interface': 9.749133241889706e-06, 'learning_rate/action_model': 9.749133241889705e-05, 'epoch': 0.3}) +[2026-06-17 17:52:08,789][starVLA.training.train_starvla][INFO] - Step 818, Loss: {'train/loss': 0.6770079731941223, 'train/grad_norm_pre_clip': 7.866754055023193, 'timing/data': 0.015539322048425674, 'timing/model': 1.0506712207570672, 'learning_rate/qwen_vl_interface': 9.748439233428897e-06, 'learning_rate/action_model': 9.748439233428896e-05, 'epoch': 0.3}) +[2026-06-17 17:52:13,283][starVLA.training.train_starvla][INFO] - Step 819, Loss: {'train/loss': 1.3399921655654907, 'train/grad_norm_pre_clip': 12.3770112991333, 'timing/data': 0.011433462612330914, 'timing/model': 1.102330157533288, 'learning_rate/qwen_vl_interface': 9.747744292436794e-06, 'learning_rate/action_model': 9.747744292436793e-05, 'epoch': 0.3}) +[2026-06-17 17:52:17,737][starVLA.training.train_starvla][INFO] - Step 820, Loss: {'train/loss': 0.902644157409668, 'train/grad_norm_pre_clip': 9.331132888793945, 'timing/data': 0.0111915348097682, 'timing/model': 1.0680649923160672, 'learning_rate/qwen_vl_interface': 9.747048419057458e-06, 'learning_rate/action_model': 9.747048419057458e-05, 'epoch': 0.3}) +[2026-06-17 17:52:22,231][starVLA.training.train_starvla][INFO] - Step 821, Loss: {'train/loss': 1.0334092378616333, 'train/grad_norm_pre_clip': 11.601950645446777, 'timing/data': 0.011146364733576775, 'timing/model': 1.2154505848884583, 'learning_rate/qwen_vl_interface': 9.746351613435148e-06, 'learning_rate/action_model': 9.746351613435147e-05, 'epoch': 0.3}) +[2026-06-17 17:52:26,553][starVLA.training.train_starvla][INFO] - Step 822, Loss: {'train/loss': 0.8950122594833374, 'train/grad_norm_pre_clip': 7.813165664672852, 'timing/data': 0.018446835689246655, 'timing/model': 1.3310883324593306, 'learning_rate/qwen_vl_interface': 9.74565387571431e-06, 'learning_rate/action_model': 9.745653875714309e-05, 'epoch': 0.3}) +[2026-06-17 17:52:30,835][starVLA.training.train_starvla][INFO] - Step 823, Loss: {'train/loss': 0.8154041767120361, 'train/grad_norm_pre_clip': 12.677748680114746, 'timing/data': 0.013117307797074318, 'timing/model': 1.180972863920033, 'learning_rate/qwen_vl_interface': 9.744955206039585e-06, 'learning_rate/action_model': 9.744955206039585e-05, 'epoch': 0.3}) +[2026-06-17 17:52:35,304][starVLA.training.train_starvla][INFO] - Step 824, Loss: {'train/loss': 0.7439048886299133, 'train/grad_norm_pre_clip': 9.696566581726074, 'timing/data': 0.022336142137646675, 'timing/model': 1.0857906872406602, 'learning_rate/qwen_vl_interface': 9.74425560455581e-06, 'learning_rate/action_model': 9.74425560455581e-05, 'epoch': 0.3}) +[2026-06-17 17:52:39,596][starVLA.training.train_starvla][INFO] - Step 825, Loss: {'train/loss': 1.3518741130828857, 'train/grad_norm_pre_clip': 15.860892295837402, 'timing/data': 0.012524213641881943, 'timing/model': 1.0667722951620817, 'learning_rate/qwen_vl_interface': 9.743555071408011e-06, 'learning_rate/action_model': 9.743555071408011e-05, 'epoch': 0.3}) +[2026-06-17 17:52:44,059][starVLA.training.train_starvla][INFO] - Step 826, Loss: {'train/loss': 0.986050546169281, 'train/grad_norm_pre_clip': 10.864737510681152, 'timing/data': 0.010825211182236671, 'timing/model': 1.0235370239242911, 'learning_rate/qwen_vl_interface': 9.742853606741412e-06, 'learning_rate/action_model': 9.742853606741411e-05, 'epoch': 0.3}) +[2026-06-17 17:52:48,524][starVLA.training.train_starvla][INFO] - Step 827, Loss: {'train/loss': 0.826569139957428, 'train/grad_norm_pre_clip': 11.523345947265625, 'timing/data': 0.014052994549274445, 'timing/model': 1.409676417708397, 'learning_rate/qwen_vl_interface': 9.742151210701423e-06, 'learning_rate/action_model': 9.742151210701424e-05, 'epoch': 0.3}) +[2026-06-17 17:52:52,603][starVLA.training.train_starvla][INFO] - Step 828, Loss: {'train/loss': 1.0518769025802612, 'train/grad_norm_pre_clip': 11.763015747070312, 'timing/data': 0.010350274853408337, 'timing/model': 1.0340371737256646, 'learning_rate/qwen_vl_interface': 9.741447883433656e-06, 'learning_rate/action_model': 9.741447883433656e-05, 'epoch': 0.3}) +[2026-06-17 17:53:00,258][starVLA.training.train_starvla][INFO] - Step 829, Loss: {'train/loss': 1.0624070167541504, 'train/grad_norm_pre_clip': 15.97321891784668, 'timing/data': 0.010800018906593323, 'timing/model': 1.1681540738791227, 'learning_rate/qwen_vl_interface': 9.74074362508391e-06, 'learning_rate/action_model': 9.740743625083911e-05, 'epoch': 0.3}) +[2026-06-17 17:53:04,613][starVLA.training.train_starvla][INFO] - Step 830, Loss: {'train/loss': 0.6820756196975708, 'train/grad_norm_pre_clip': 9.81755256652832, 'timing/data': 0.015100476332008839, 'timing/model': 1.0396842071786523, 'learning_rate/qwen_vl_interface': 9.74003843579818e-06, 'learning_rate/action_model': 9.740038435798178e-05, 'epoch': 0.3}) +[2026-06-17 17:53:09,317][starVLA.training.train_starvla][INFO] - Step 831, Loss: {'train/loss': 0.8924475908279419, 'train/grad_norm_pre_clip': 8.712940216064453, 'timing/data': 0.0127536216750741, 'timing/model': 1.0561304539442062, 'learning_rate/qwen_vl_interface': 9.739332315722647e-06, 'learning_rate/action_model': 9.739332315722646e-05, 'epoch': 0.3}) +[2026-06-17 17:53:14,745][starVLA.training.train_starvla][INFO] - Step 832, Loss: {'train/loss': 1.1182398796081543, 'train/grad_norm_pre_clip': 10.220032691955566, 'timing/data': 0.010831889696419239, 'timing/model': 1.375797945074737, 'learning_rate/qwen_vl_interface': 9.738625265003695e-06, 'learning_rate/action_model': 9.738625265003695e-05, 'epoch': 0.3}) +[2026-06-17 17:53:23,365][starVLA.training.train_starvla][INFO] - Step 833, Loss: {'train/loss': 0.9904526472091675, 'train/grad_norm_pre_clip': 8.668439865112305, 'timing/data': 0.016283798031508923, 'timing/model': 2.690469266846776, 'learning_rate/qwen_vl_interface': 9.7379172837879e-06, 'learning_rate/action_model': 9.737917283787898e-05, 'epoch': 0.31}) +[2026-06-17 17:53:28,111][starVLA.training.train_starvla][INFO] - Step 834, Loss: {'train/loss': 0.9152529239654541, 'train/grad_norm_pre_clip': 11.352760314941406, 'timing/data': 0.013238433748483658, 'timing/model': 1.045972090214491, 'learning_rate/qwen_vl_interface': 9.737208372222018e-06, 'learning_rate/action_model': 9.737208372222017e-05, 'epoch': 0.31}) +[2026-06-17 17:53:36,308][starVLA.training.train_starvla][INFO] - Step 835, Loss: {'train/loss': 1.2064882516860962, 'train/grad_norm_pre_clip': 12.755698204040527, 'timing/data': 0.014387953095138073, 'timing/model': 1.3142748065292835, 'learning_rate/qwen_vl_interface': 9.736498530453015e-06, 'learning_rate/action_model': 9.736498530453014e-05, 'epoch': 0.31}) +[2026-06-17 17:53:46,407][starVLA.training.train_starvla][INFO] - Step 836, Loss: {'train/loss': 0.8634564876556396, 'train/grad_norm_pre_clip': 10.793668746948242, 'timing/data': 0.01661327015608549, 'timing/model': 2.290504502132535, 'learning_rate/qwen_vl_interface': 9.735787758628036e-06, 'learning_rate/action_model': 9.735787758628035e-05, 'epoch': 0.31}) +[2026-06-17 17:53:52,100][starVLA.training.train_starvla][INFO] - Step 837, Loss: {'train/loss': 0.7151817083358765, 'train/grad_norm_pre_clip': 10.601226806640625, 'timing/data': 0.011954443529248238, 'timing/model': 1.4726747246459126, 'learning_rate/qwen_vl_interface': 9.73507605689443e-06, 'learning_rate/action_model': 9.735076056894429e-05, 'epoch': 0.31}) +[2026-06-17 17:53:59,495][starVLA.training.train_starvla][INFO] - Step 838, Loss: {'train/loss': 1.014496922492981, 'train/grad_norm_pre_clip': 10.219795227050781, 'timing/data': 0.012395150028169155, 'timing/model': 1.1373518463224173, 'learning_rate/qwen_vl_interface': 9.73436342539973e-06, 'learning_rate/action_model': 9.734363425399728e-05, 'epoch': 0.31}) +[2026-06-17 17:54:07,425][starVLA.training.train_starvla][INFO] - Step 839, Loss: {'train/loss': 1.0029916763305664, 'train/grad_norm_pre_clip': 10.043240547180176, 'timing/data': 0.015227887779474258, 'timing/model': 1.3612663066014647, 'learning_rate/qwen_vl_interface': 9.733649864291667e-06, 'learning_rate/action_model': 9.733649864291666e-05, 'epoch': 0.31}) +[2026-06-17 17:54:12,101][starVLA.training.train_starvla][INFO] - Step 840, Loss: {'train/loss': 1.237772822380066, 'train/grad_norm_pre_clip': 16.39181137084961, 'timing/data': 0.024781874381005764, 'timing/model': 1.1478895340114832, 'learning_rate/qwen_vl_interface': 9.732935373718162e-06, 'learning_rate/action_model': 9.732935373718161e-05, 'epoch': 0.31}) +[2026-06-17 17:54:23,221][starVLA.training.train_starvla][INFO] - Step 841, Loss: {'train/loss': 1.0612099170684814, 'train/grad_norm_pre_clip': 8.44089126586914, 'timing/data': 0.02195722423493862, 'timing/model': 5.701297355815768, 'learning_rate/qwen_vl_interface': 9.73221995382733e-06, 'learning_rate/action_model': 9.73221995382733e-05, 'epoch': 0.31}) +[2026-06-17 17:54:38,265][starVLA.training.train_starvla][INFO] - Step 842, Loss: {'train/loss': 1.126022219657898, 'train/grad_norm_pre_clip': 13.64140510559082, 'timing/data': 0.015608592890202999, 'timing/model': 5.206475110724568, 'learning_rate/qwen_vl_interface': 9.73150360476748e-06, 'learning_rate/action_model': 9.73150360476748e-05, 'epoch': 0.31}) +[2026-06-17 17:54:50,567][starVLA.training.train_starvla][INFO] - Step 843, Loss: {'train/loss': 1.1643925905227661, 'train/grad_norm_pre_clip': 11.001655578613281, 'timing/data': 0.014741018414497375, 'timing/model': 1.1298294002190232, 'learning_rate/qwen_vl_interface': 9.73078632668711e-06, 'learning_rate/action_model': 9.73078632668711e-05, 'epoch': 0.31}) +[2026-06-17 17:54:54,959][starVLA.training.train_starvla][INFO] - Step 844, Loss: {'train/loss': 1.0731853246688843, 'train/grad_norm_pre_clip': 12.611642837524414, 'timing/data': 0.01391938142478466, 'timing/model': 1.024554361589253, 'learning_rate/qwen_vl_interface': 9.730068119734912e-06, 'learning_rate/action_model': 9.730068119734912e-05, 'epoch': 0.31}) +[2026-06-17 17:54:59,914][starVLA.training.train_starvla][INFO] - Step 845, Loss: {'train/loss': 0.7714079022407532, 'train/grad_norm_pre_clip': 8.088655471801758, 'timing/data': 0.011349870823323727, 'timing/model': 1.2106634275987744, 'learning_rate/qwen_vl_interface': 9.729348984059773e-06, 'learning_rate/action_model': 9.729348984059774e-05, 'epoch': 0.31}) +[2026-06-17 17:55:04,444][starVLA.training.train_starvla][INFO] - Step 846, Loss: {'train/loss': 1.1468456983566284, 'train/grad_norm_pre_clip': 11.315229415893555, 'timing/data': 0.014241334050893784, 'timing/model': 1.0349132334813476, 'learning_rate/qwen_vl_interface': 9.728628919810769e-06, 'learning_rate/action_model': 9.728628919810768e-05, 'epoch': 0.31}) +[2026-06-17 17:55:09,079][starVLA.training.train_starvla][INFO] - Step 847, Loss: {'train/loss': 0.8316787481307983, 'train/grad_norm_pre_clip': 7.428349018096924, 'timing/data': 0.011533113196492195, 'timing/model': 1.3779197307303548, 'learning_rate/qwen_vl_interface': 9.727907927137171e-06, 'learning_rate/action_model': 9.727907927137171e-05, 'epoch': 0.31}) +[2026-06-17 17:55:13,362][starVLA.training.train_starvla][INFO] - Step 848, Loss: {'train/loss': 1.2031222581863403, 'train/grad_norm_pre_clip': 8.798129081726074, 'timing/data': 0.013446815311908722, 'timing/model': 1.0274726189672947, 'learning_rate/qwen_vl_interface': 9.72718600618844e-06, 'learning_rate/action_model': 9.727186006188441e-05, 'epoch': 0.31}) +[2026-06-17 17:55:17,823][starVLA.training.train_starvla][INFO] - Step 849, Loss: {'train/loss': 1.3098808526992798, 'train/grad_norm_pre_clip': 11.722162246704102, 'timing/data': 0.011384617537260056, 'timing/model': 1.0366314444690943, 'learning_rate/qwen_vl_interface': 9.726463157114233e-06, 'learning_rate/action_model': 9.726463157114234e-05, 'epoch': 0.31}) +[2026-06-17 17:55:22,291][starVLA.training.train_starvla][INFO] - Step 850, Loss: {'train/loss': 0.9051558375358582, 'train/grad_norm_pre_clip': 7.861241340637207, 'timing/data': 0.012217112816870213, 'timing/model': 1.0346491001546383, 'learning_rate/qwen_vl_interface': 9.725739380064399e-06, 'learning_rate/action_model': 9.725739380064398e-05, 'epoch': 0.31}) +[2026-06-17 17:55:26,678][starVLA.training.train_starvla][INFO] - Step 851, Loss: {'train/loss': 0.8230786919593811, 'train/grad_norm_pre_clip': 9.939591407775879, 'timing/data': 0.010742640122771263, 'timing/model': 1.0294932266697288, 'learning_rate/qwen_vl_interface': 9.725014675188972e-06, 'learning_rate/action_model': 9.725014675188973e-05, 'epoch': 0.31}) +[2026-06-17 17:55:31,483][starVLA.training.train_starvla][INFO] - Step 852, Loss: {'train/loss': 0.7826758623123169, 'train/grad_norm_pre_clip': 7.9818034172058105, 'timing/data': 0.018244953826069832, 'timing/model': 1.3546504927799106, 'learning_rate/qwen_vl_interface': 9.72428904263819e-06, 'learning_rate/action_model': 9.72428904263819e-05, 'epoch': 0.31}) +[2026-06-17 17:55:35,549][starVLA.training.train_starvla][INFO] - Step 853, Loss: {'train/loss': 0.9197903871536255, 'train/grad_norm_pre_clip': 8.405945777893066, 'timing/data': 0.011357606388628483, 'timing/model': 1.0286606829613447, 'learning_rate/qwen_vl_interface': 9.723562482562474e-06, 'learning_rate/action_model': 9.723562482562473e-05, 'epoch': 0.31}) +[2026-06-17 17:55:40,173][starVLA.training.train_starvla][INFO] - Step 854, Loss: {'train/loss': 0.9472106099128723, 'train/grad_norm_pre_clip': 9.020529747009277, 'timing/data': 0.01203209813684225, 'timing/model': 1.0737202558666468, 'learning_rate/qwen_vl_interface': 9.72283499511244e-06, 'learning_rate/action_model': 9.722834995112439e-05, 'epoch': 0.31}) +[2026-06-17 17:55:44,792][starVLA.training.train_starvla][INFO] - Step 855, Loss: {'train/loss': 1.18761146068573, 'train/grad_norm_pre_clip': 9.44408893585205, 'timing/data': 0.01244535855948925, 'timing/model': 1.1205511903390288, 'learning_rate/qwen_vl_interface': 9.722106580438901e-06, 'learning_rate/action_model': 9.7221065804389e-05, 'epoch': 0.31}) +[2026-06-17 17:55:49,193][starVLA.training.train_starvla][INFO] - Step 856, Loss: {'train/loss': 0.7586427927017212, 'train/grad_norm_pre_clip': 7.199925422668457, 'timing/data': 0.011570571921765804, 'timing/model': 1.0246518608182669, 'learning_rate/qwen_vl_interface': 9.721377238692855e-06, 'learning_rate/action_model': 9.721377238692854e-05, 'epoch': 0.31}) +[2026-06-17 17:55:53,745][starVLA.training.train_starvla][INFO] - Step 857, Loss: {'train/loss': 0.941458523273468, 'train/grad_norm_pre_clip': 10.14294147491455, 'timing/data': 0.013277099467813969, 'timing/model': 1.3476661071181297, 'learning_rate/qwen_vl_interface': 9.720646970025494e-06, 'learning_rate/action_model': 9.720646970025494e-05, 'epoch': 0.31}) +[2026-06-17 17:56:01,661][starVLA.training.train_starvla][INFO] - Step 858, Loss: {'train/loss': 1.027824878692627, 'train/grad_norm_pre_clip': 8.741289138793945, 'timing/data': 0.013929784297943115, 'timing/model': 4.489348819479346, 'learning_rate/qwen_vl_interface': 9.719915774588209e-06, 'learning_rate/action_model': 9.719915774588209e-05, 'epoch': 0.31}) +[2026-06-17 17:56:10,640][starVLA.training.train_starvla][INFO] - Step 859, Loss: {'train/loss': 0.7908792495727539, 'train/grad_norm_pre_clip': 10.910865783691406, 'timing/data': 0.01175592839717865, 'timing/model': 1.1314225662499666, 'learning_rate/qwen_vl_interface': 9.719183652532572e-06, 'learning_rate/action_model': 9.719183652532572e-05, 'epoch': 0.31}) +[2026-06-17 17:56:17,198][starVLA.training.train_starvla][INFO] - Step 860, Loss: {'train/loss': 1.04136061668396, 'train/grad_norm_pre_clip': 9.480477333068848, 'timing/data': 0.013485457748174667, 'timing/model': 1.0469570783898234, 'learning_rate/qwen_vl_interface': 9.718450604010355e-06, 'learning_rate/action_model': 9.718450604010355e-05, 'epoch': 0.31}) +[2026-06-17 17:56:22,393][starVLA.training.train_starvla][INFO] - Step 861, Loss: {'train/loss': 0.8143160939216614, 'train/grad_norm_pre_clip': 7.904932022094727, 'timing/data': 0.011180351488292217, 'timing/model': 1.4065331378951669, 'learning_rate/qwen_vl_interface': 9.717716629173521e-06, 'learning_rate/action_model': 9.71771662917352e-05, 'epoch': 0.32}) +[2026-06-17 17:56:27,186][starVLA.training.train_starvla][INFO] - Step 862, Loss: {'train/loss': 0.893575131893158, 'train/grad_norm_pre_clip': 11.986361503601074, 'timing/data': 0.012426513247191906, 'timing/model': 1.418159299530089, 'learning_rate/qwen_vl_interface': 9.71698172817422e-06, 'learning_rate/action_model': 9.716981728174219e-05, 'epoch': 0.32}) +[2026-06-17 17:56:31,340][starVLA.training.train_starvla][INFO] - Step 863, Loss: {'train/loss': 0.858243465423584, 'train/grad_norm_pre_clip': 10.90666389465332, 'timing/data': 0.010284020565450191, 'timing/model': 1.0320134572684765, 'learning_rate/qwen_vl_interface': 9.7162459011648e-06, 'learning_rate/action_model': 9.7162459011648e-05, 'epoch': 0.32}) +[2026-06-17 17:56:37,816][starVLA.training.train_starvla][INFO] - Step 864, Loss: {'train/loss': 0.691247284412384, 'train/grad_norm_pre_clip': 8.253978729248047, 'timing/data': 0.01540900208055973, 'timing/model': 1.0478141009807587, 'learning_rate/qwen_vl_interface': 9.7155091482978e-06, 'learning_rate/action_model': 9.7155091482978e-05, 'epoch': 0.32}) +[2026-06-17 17:56:42,203][starVLA.training.train_starvla][INFO] - Step 865, Loss: {'train/loss': 1.1713013648986816, 'train/grad_norm_pre_clip': 11.560797691345215, 'timing/data': 0.013361676596105099, 'timing/model': 1.0351614188402891, 'learning_rate/qwen_vl_interface': 9.714771469725947e-06, 'learning_rate/action_model': 9.714771469725947e-05, 'epoch': 0.32}) +[2026-06-17 17:56:48,238][starVLA.training.train_starvla][INFO] - Step 866, Loss: {'train/loss': 0.9048929214477539, 'train/grad_norm_pre_clip': 9.594532012939453, 'timing/data': 0.012095180340111256, 'timing/model': 1.082784190773964, 'learning_rate/qwen_vl_interface': 9.714032865602167e-06, 'learning_rate/action_model': 9.714032865602166e-05, 'epoch': 0.32}) +[2026-06-17 17:56:52,687][starVLA.training.train_starvla][INFO] - Step 867, Loss: {'train/loss': 0.8113067150115967, 'train/grad_norm_pre_clip': 10.375831604003906, 'timing/data': 0.011141913011670113, 'timing/model': 1.321551044471562, 'learning_rate/qwen_vl_interface': 9.713293336079568e-06, 'learning_rate/action_model': 9.713293336079568e-05, 'epoch': 0.32}) +[2026-06-17 17:56:56,827][starVLA.training.train_starvla][INFO] - Step 868, Loss: {'train/loss': 0.8247404098510742, 'train/grad_norm_pre_clip': 8.42602252960205, 'timing/data': 0.02019677311182022, 'timing/model': 1.0888654608279467, 'learning_rate/qwen_vl_interface': 9.712552881311459e-06, 'learning_rate/action_model': 9.712552881311458e-05, 'epoch': 0.32}) +[2026-06-17 17:57:01,093][starVLA.training.train_starvla][INFO] - Step 869, Loss: {'train/loss': 0.6653397083282471, 'train/grad_norm_pre_clip': 7.689490795135498, 'timing/data': 0.01079353503882885, 'timing/model': 1.0194865362718701, 'learning_rate/qwen_vl_interface': 9.711811501451336e-06, 'learning_rate/action_model': 9.711811501451335e-05, 'epoch': 0.32}) +[2026-06-17 17:57:05,561][starVLA.training.train_starvla][INFO] - Step 870, Loss: {'train/loss': 0.9461852312088013, 'train/grad_norm_pre_clip': 8.150985717773438, 'timing/data': 0.017725185491144657, 'timing/model': 1.0419418402016163, 'learning_rate/qwen_vl_interface': 9.711069196652886e-06, 'learning_rate/action_model': 9.711069196652886e-05, 'epoch': 0.32}) +[2026-06-17 17:57:09,907][starVLA.training.train_starvla][INFO] - Step 871, Loss: {'train/loss': 1.2895004749298096, 'train/grad_norm_pre_clip': 14.68254566192627, 'timing/data': 0.010394967161118984, 'timing/model': 1.0593385389074683, 'learning_rate/qwen_vl_interface': 9.710325967069992e-06, 'learning_rate/action_model': 9.710325967069991e-05, 'epoch': 0.32}) +[2026-06-17 17:57:14,273][starVLA.training.train_starvla][INFO] - Step 872, Loss: {'train/loss': 1.1619561910629272, 'train/grad_norm_pre_clip': 11.744097709655762, 'timing/data': 0.011438318528234959, 'timing/model': 1.351657367311418, 'learning_rate/qwen_vl_interface': 9.709581812856725e-06, 'learning_rate/action_model': 9.709581812856726e-05, 'epoch': 0.32}) +[2026-06-17 17:57:18,322][starVLA.training.train_starvla][INFO] - Step 873, Loss: {'train/loss': 0.7537936568260193, 'train/grad_norm_pre_clip': 9.634549140930176, 'timing/data': 0.01193790603429079, 'timing/model': 1.0273227663710713, 'learning_rate/qwen_vl_interface': 9.708836734167351e-06, 'learning_rate/action_model': 9.708836734167351e-05, 'epoch': 0.32}) +[2026-06-17 17:57:24,583][starVLA.training.train_starvla][INFO] - Step 874, Loss: {'train/loss': 0.9140032529830933, 'train/grad_norm_pre_clip': 8.032390594482422, 'timing/data': 0.010830719023942947, 'timing/model': 1.0284818215295672, 'learning_rate/qwen_vl_interface': 9.708090731156321e-06, 'learning_rate/action_model': 9.708090731156321e-05, 'epoch': 0.32}) +[2026-06-17 17:57:29,138][starVLA.training.train_starvla][INFO] - Step 875, Loss: {'train/loss': 0.8758654594421387, 'train/grad_norm_pre_clip': 9.705172538757324, 'timing/data': 0.019306758418679237, 'timing/model': 1.069858313538134, 'learning_rate/qwen_vl_interface': 9.707343803978288e-06, 'learning_rate/action_model': 9.707343803978288e-05, 'epoch': 0.32}) +[2026-06-17 17:57:33,714][starVLA.training.train_starvla][INFO] - Step 876, Loss: {'train/loss': 1.16489577293396, 'train/grad_norm_pre_clip': 10.446846961975098, 'timing/data': 0.014174294658005238, 'timing/model': 1.0227440698072314, 'learning_rate/qwen_vl_interface': 9.706595952788088e-06, 'learning_rate/action_model': 9.706595952788088e-05, 'epoch': 0.32}) +[2026-06-17 17:57:38,122][starVLA.training.train_starvla][INFO] - Step 877, Loss: {'train/loss': 1.1158570051193237, 'train/grad_norm_pre_clip': 11.183319091796875, 'timing/data': 0.012080783024430275, 'timing/model': 1.332783074118197, 'learning_rate/qwen_vl_interface': 9.705847177740751e-06, 'learning_rate/action_model': 9.705847177740751e-05, 'epoch': 0.32}) +[2026-06-17 17:57:42,240][starVLA.training.train_starvla][INFO] - Step 878, Loss: {'train/loss': 1.0661755800247192, 'train/grad_norm_pre_clip': 10.748946189880371, 'timing/data': 0.010908544063568115, 'timing/model': 1.0623289877548814, 'learning_rate/qwen_vl_interface': 9.7050974789915e-06, 'learning_rate/action_model': 9.7050974789915e-05, 'epoch': 0.32}) +[2026-06-17 17:57:46,620][starVLA.training.train_starvla][INFO] - Step 879, Loss: {'train/loss': 1.2609186172485352, 'train/grad_norm_pre_clip': 13.146276473999023, 'timing/data': 0.010747634805738926, 'timing/model': 1.0578510453924537, 'learning_rate/qwen_vl_interface': 9.704346856695747e-06, 'learning_rate/action_model': 9.704346856695746e-05, 'epoch': 0.32}) +[2026-06-17 17:57:50,813][starVLA.training.train_starvla][INFO] - Step 880, Loss: {'train/loss': 1.1133190393447876, 'train/grad_norm_pre_clip': 9.582773208618164, 'timing/data': 0.0106950169429183, 'timing/model': 0.9957501534372568, 'learning_rate/qwen_vl_interface': 9.703595311009098e-06, 'learning_rate/action_model': 9.703595311009097e-05, 'epoch': 0.32}) +[2026-06-17 17:57:54,903][starVLA.training.train_starvla][INFO] - Step 881, Loss: {'train/loss': 1.2742048501968384, 'train/grad_norm_pre_clip': 11.318220138549805, 'timing/data': 0.010651922784745693, 'timing/model': 1.021350072696805, 'learning_rate/qwen_vl_interface': 9.702842842087348e-06, 'learning_rate/action_model': 9.702842842087348e-05, 'epoch': 0.32}) +[2026-06-17 17:57:59,045][starVLA.training.train_starvla][INFO] - Step 882, Loss: {'train/loss': 0.8026957511901855, 'train/grad_norm_pre_clip': 8.186445236206055, 'timing/data': 0.015385067090392113, 'timing/model': 1.2325104791671038, 'learning_rate/qwen_vl_interface': 9.702089450086486e-06, 'learning_rate/action_model': 9.702089450086486e-05, 'epoch': 0.32}) +[2026-06-17 17:58:02,921][starVLA.training.train_starvla][INFO] - Step 883, Loss: {'train/loss': 1.089526653289795, 'train/grad_norm_pre_clip': 11.07852840423584, 'timing/data': 0.010493318550288677, 'timing/model': 0.9946049330756068, 'learning_rate/qwen_vl_interface': 9.70133513516269e-06, 'learning_rate/action_model': 9.701335135162688e-05, 'epoch': 0.32}) +[2026-06-17 17:58:07,932][starVLA.training.train_starvla][INFO] - Step 884, Loss: {'train/loss': 1.0574272871017456, 'train/grad_norm_pre_clip': 9.496848106384277, 'timing/data': 0.009119139984250069, 'timing/model': 1.0037339394912124, 'learning_rate/qwen_vl_interface': 9.70057989747233e-06, 'learning_rate/action_model': 9.700579897472329e-05, 'epoch': 0.32}) +[2026-06-17 17:58:12,022][starVLA.training.train_starvla][INFO] - Step 885, Loss: {'train/loss': 0.8757904171943665, 'train/grad_norm_pre_clip': 11.647747993469238, 'timing/data': 0.00987770315259695, 'timing/model': 0.9931763140484691, 'learning_rate/qwen_vl_interface': 9.699823737171969e-06, 'learning_rate/action_model': 9.699823737171969e-05, 'epoch': 0.32}) +[2026-06-17 17:58:16,227][starVLA.training.train_starvla][INFO] - Step 886, Loss: {'train/loss': 0.730492353439331, 'train/grad_norm_pre_clip': 116.79192352294922, 'timing/data': 0.01008832361549139, 'timing/model': 1.0463871574029326, 'learning_rate/qwen_vl_interface': 9.699066654418358e-06, 'learning_rate/action_model': 9.699066654418358e-05, 'epoch': 0.32}) +[2026-06-17 17:58:20,345][starVLA.training.train_starvla][INFO] - Step 887, Loss: {'train/loss': 1.0601050853729248, 'train/grad_norm_pre_clip': 12.080177307128906, 'timing/data': 0.00978525634855032, 'timing/model': 1.226222568191588, 'learning_rate/qwen_vl_interface': 9.698308649368444e-06, 'learning_rate/action_model': 9.698308649368444e-05, 'epoch': 0.32}) +[2026-06-17 17:58:24,211][starVLA.training.train_starvla][INFO] - Step 888, Loss: {'train/loss': 0.9676638245582581, 'train/grad_norm_pre_clip': 10.649382591247559, 'timing/data': 0.01002653781324625, 'timing/model': 1.0023635104298592, 'learning_rate/qwen_vl_interface': 9.69754972217936e-06, 'learning_rate/action_model': 9.69754972217936e-05, 'epoch': 0.33}) +[2026-06-17 17:58:28,309][starVLA.training.train_starvla][INFO] - Step 889, Loss: {'train/loss': 0.9568396806716919, 'train/grad_norm_pre_clip': 10.194828987121582, 'timing/data': 0.009730972349643707, 'timing/model': 1.0328578278422356, 'learning_rate/qwen_vl_interface': 9.696789873008432e-06, 'learning_rate/action_model': 9.696789873008433e-05, 'epoch': 0.33}) +[2026-06-17 17:58:32,663][starVLA.training.train_starvla][INFO] - Step 890, Loss: {'train/loss': 1.1103142499923706, 'train/grad_norm_pre_clip': 10.96174430847168, 'timing/data': 0.01357983984053135, 'timing/model': 1.0130814667791128, 'learning_rate/qwen_vl_interface': 9.69602910201318e-06, 'learning_rate/action_model': 9.696029102013179e-05, 'epoch': 0.33}) +[2026-06-17 17:58:37,156][starVLA.training.train_starvla][INFO] - Step 891, Loss: {'train/loss': 1.003326416015625, 'train/grad_norm_pre_clip': 9.930232048034668, 'timing/data': 0.010288490913808346, 'timing/model': 1.0204448197036982, 'learning_rate/qwen_vl_interface': 9.695267409351308e-06, 'learning_rate/action_model': 9.695267409351307e-05, 'epoch': 0.33}) +[2026-06-17 17:58:41,576][starVLA.training.train_starvla][INFO] - Step 892, Loss: {'train/loss': 0.7694538831710815, 'train/grad_norm_pre_clip': 8.182167053222656, 'timing/data': 0.010937883518636227, 'timing/model': 1.399208844639361, 'learning_rate/qwen_vl_interface': 9.694504795180722e-06, 'learning_rate/action_model': 9.694504795180721e-05, 'epoch': 0.33}) +[2026-06-17 17:58:47,357][starVLA.training.train_starvla][INFO] - Step 893, Loss: {'train/loss': 1.0422241687774658, 'train/grad_norm_pre_clip': 10.354642868041992, 'timing/data': 0.015650520101189613, 'timing/model': 1.1189794316887856, 'learning_rate/qwen_vl_interface': 9.693741259659507e-06, 'learning_rate/action_model': 9.693741259659507e-05, 'epoch': 0.33}) +[2026-06-17 17:58:51,638][starVLA.training.train_starvla][INFO] - Step 894, Loss: {'train/loss': 0.6343986988067627, 'train/grad_norm_pre_clip': 9.363310813903809, 'timing/data': 0.012675709091126919, 'timing/model': 1.022760869935155, 'learning_rate/qwen_vl_interface': 9.692976802945947e-06, 'learning_rate/action_model': 9.692976802945948e-05, 'epoch': 0.33}) +[2026-06-17 17:58:55,946][starVLA.training.train_starvla][INFO] - Step 895, Loss: {'train/loss': 0.8106216192245483, 'train/grad_norm_pre_clip': 7.5931878089904785, 'timing/data': 0.013118172064423561, 'timing/model': 1.0454492745921016, 'learning_rate/qwen_vl_interface': 9.692211425198517e-06, 'learning_rate/action_model': 9.692211425198517e-05, 'epoch': 0.33}) +[2026-06-17 17:59:00,484][starVLA.training.train_starvla][INFO] - Step 896, Loss: {'train/loss': 1.0178532600402832, 'train/grad_norm_pre_clip': 11.602764129638672, 'timing/data': 0.0138164097443223, 'timing/model': 1.11660878919065, 'learning_rate/qwen_vl_interface': 9.691445126575879e-06, 'learning_rate/action_model': 9.691445126575877e-05, 'epoch': 0.33}) +[2026-06-17 17:59:04,855][starVLA.training.train_starvla][INFO] - Step 897, Loss: {'train/loss': 1.2140015363693237, 'train/grad_norm_pre_clip': 10.98363971710205, 'timing/data': 0.010222223587334156, 'timing/model': 1.3760318234562874, 'learning_rate/qwen_vl_interface': 9.690677907236884e-06, 'learning_rate/action_model': 9.690677907236884e-05, 'epoch': 0.33}) +[2026-06-17 17:59:08,882][starVLA.training.train_starvla][INFO] - Step 898, Loss: {'train/loss': 1.072554111480713, 'train/grad_norm_pre_clip': 8.415872573852539, 'timing/data': 0.014696966856718063, 'timing/model': 1.0279841544106603, 'learning_rate/qwen_vl_interface': 9.689909767340582e-06, 'learning_rate/action_model': 9.689909767340583e-05, 'epoch': 0.33}) +[2026-06-17 17:59:13,447][starVLA.training.train_starvla][INFO] - Step 899, Loss: {'train/loss': 1.2678509950637817, 'train/grad_norm_pre_clip': 12.617173194885254, 'timing/data': 0.015102666802704334, 'timing/model': 1.1221009958535433, 'learning_rate/qwen_vl_interface': 9.689140707046209e-06, 'learning_rate/action_model': 9.689140707046209e-05, 'epoch': 0.33}) +[2026-06-17 18:01:56,714][starVLA.training.train_starvla][INFO] - Step 900, Loss: {'train/loss': 0.723626971244812, 'train/grad_norm_pre_clip': 8.572879791259766, 'eval/loss': 0.9780584573745728, 'eval/latency_2/loss': 0.9780584573745728, 'eval/demon_fire_precision': 0.815751960085531, 'eval/demon_fire_recall': 0.9577405857740586, 'eval/demon_fire_f1': 0.8810623556581987, 'eval/demon_fire_rate_model': 0.7807456872565387, 'eval/demon_fire_rate_teacher': 0.6649972175848636, 'eval/demon_left_precision': 0.8228404099560761, 'eval/demon_left_recall': 0.5915789473684211, 'eval/demon_left_f1': 0.6883037354562156, 'eval/demon_left_rate_model': 0.19003895381190875, 'eval/demon_left_rate_teacher': 0.2643294379521425, 'eval/demon_right_precision': 0.8014492753623188, 'eval/demon_right_recall': 0.6248587570621469, 'eval/demon_right_f1': 0.7022222222222223, 'eval/demon_right_rate_model': 0.19198664440734559, 'eval/demon_right_rate_teacher': 0.24624373956594323, 'eval/demon_move_f1': 0.695262978839219, 'eval/demon_cc_f1': 0.815669454970404, 'eval/latency_2/demon_fire_precision': 0.815751960085531, 'eval/latency_2/demon_fire_recall': 0.9577405857740586, 'eval/latency_2/demon_fire_f1': 0.8810623556581987, 'eval/latency_2/demon_fire_rate_model': 0.7807456872565387, 'eval/latency_2/demon_fire_rate_teacher': 0.6649972175848636, 'eval/latency_2/demon_left_precision': 0.8228404099560761, 'eval/latency_2/demon_left_recall': 0.5915789473684211, 'eval/latency_2/demon_left_f1': 0.6883037354562156, 'eval/latency_2/demon_left_rate_model': 0.19003895381190875, 'eval/latency_2/demon_left_rate_teacher': 0.2643294379521425, 'eval/latency_2/demon_right_precision': 0.8014492753623188, 'eval/latency_2/demon_right_recall': 0.6248587570621469, 'eval/latency_2/demon_right_f1': 0.7022222222222223, 'eval/latency_2/demon_right_rate_model': 0.19198664440734559, 'eval/latency_2/demon_right_rate_teacher': 0.24624373956594323, 'eval/latency_2/demon_move_f1': 0.695262978839219, 'eval/latency_2/demon_cc_f1': 0.815669454970404, 'timing/data': 0.012060683220624924, 'timing/model': 1.0227827839553356, 'learning_rate/qwen_vl_interface': 9.68837072651319e-06, 'learning_rate/action_model': 9.688370726513188e-05, 'epoch': 0.33}) +[2026-06-17 18:01:56,716][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_900_state +[2026-06-17 18:01:56,723][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 18:02:05,186][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_900_state/model.safetensors +[2026-06-17 18:02:20,175][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_900_state/optimizer.bin +[2026-06-17 18:02:20,200][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_900_state/sampler.bin +[2026-06-17 18:02:20,201][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_900_state/sampler_1.bin +[2026-06-17 18:02:20,204][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_900_state/random_states_0.pkl +[2026-06-17 18:02:20,208][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 900 +[2026-06-17 18:02:20,209][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 18:02:20,226][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 18:02:25,115][starVLA.training.train_starvla][INFO] - Step 901, Loss: {'train/loss': 1.1893001794815063, 'train/grad_norm_pre_clip': 14.9992094039917, 'timing/data': 0.011599929071962833, 'timing/model': 1.3393180165439844, 'learning_rate/qwen_vl_interface': 9.687599825901141e-06, 'learning_rate/action_model': 9.687599825901142e-05, 'epoch': 0.33}) +[2026-06-17 18:02:29,650][starVLA.training.train_starvla][INFO] - Step 902, Loss: {'train/loss': 0.9189640879631042, 'train/grad_norm_pre_clip': 15.434539794921875, 'timing/data': 0.016081400215625763, 'timing/model': 1.0823661284521222, 'learning_rate/qwen_vl_interface': 9.686828005369876e-06, 'learning_rate/action_model': 9.686828005369875e-05, 'epoch': 0.33}) +[2026-06-17 18:02:36,588][starVLA.training.train_starvla][INFO] - Step 903, Loss: {'train/loss': 1.1883686780929565, 'train/grad_norm_pre_clip': 11.983675956726074, 'timing/data': 0.017520719207823277, 'timing/model': 3.87159959692508, 'learning_rate/qwen_vl_interface': 9.68605526507939e-06, 'learning_rate/action_model': 9.686055265079389e-05, 'epoch': 0.33}) +[2026-06-17 18:02:40,776][starVLA.training.train_starvla][INFO] - Step 904, Loss: {'train/loss': 0.7022789716720581, 'train/grad_norm_pre_clip': 9.211250305175781, 'timing/data': 0.011209800839424133, 'timing/model': 1.0267579518258572, 'learning_rate/qwen_vl_interface': 9.685281605189874e-06, 'learning_rate/action_model': 9.685281605189874e-05, 'epoch': 0.33}) +[2026-06-17 18:02:45,493][starVLA.training.train_starvla][INFO] - Step 905, Loss: {'train/loss': 0.9378650784492493, 'train/grad_norm_pre_clip': 9.509248733520508, 'timing/data': 0.013261795043945312, 'timing/model': 1.1378588788211346, 'learning_rate/qwen_vl_interface': 9.68450702586171e-06, 'learning_rate/action_model': 9.684507025861709e-05, 'epoch': 0.33}) +[2026-06-17 18:02:50,016][starVLA.training.train_starvla][INFO] - Step 906, Loss: {'train/loss': 0.9778465628623962, 'train/grad_norm_pre_clip': 9.27680778503418, 'timing/data': 0.011705436743795872, 'timing/model': 1.0426857005804777, 'learning_rate/qwen_vl_interface': 9.683731527255467e-06, 'learning_rate/action_model': 9.683731527255467e-05, 'epoch': 0.33}) +[2026-06-17 18:02:54,748][starVLA.training.train_starvla][INFO] - Step 907, Loss: {'train/loss': 1.138241171836853, 'train/grad_norm_pre_clip': 14.066658020019531, 'timing/data': 0.012389135546982288, 'timing/model': 1.16894471924752, 'learning_rate/qwen_vl_interface': 9.682955109531906e-06, 'learning_rate/action_model': 9.682955109531905e-05, 'epoch': 0.33}) +[2026-06-17 18:03:00,392][starVLA.training.train_starvla][INFO] - Step 908, Loss: {'train/loss': 0.9166134595870972, 'train/grad_norm_pre_clip': 39.55315399169922, 'timing/data': 0.01292511448264122, 'timing/model': 1.5060658184811473, 'learning_rate/qwen_vl_interface': 9.682177772851983e-06, 'learning_rate/action_model': 9.682177772851982e-05, 'epoch': 0.33}) +[2026-06-17 18:03:04,676][starVLA.training.train_starvla][INFO] - Step 909, Loss: {'train/loss': 0.9184909462928772, 'train/grad_norm_pre_clip': 11.704510688781738, 'timing/data': 0.011448044329881668, 'timing/model': 1.0736888032406569, 'learning_rate/qwen_vl_interface': 9.681399517376837e-06, 'learning_rate/action_model': 9.681399517376837e-05, 'epoch': 0.33}) +[2026-06-17 18:03:09,286][starVLA.training.train_starvla][INFO] - Step 910, Loss: {'train/loss': 0.9402940273284912, 'train/grad_norm_pre_clip': 10.192523956298828, 'timing/data': 0.013599456287920475, 'timing/model': 1.0307716885581613, 'learning_rate/qwen_vl_interface': 9.680620343267803e-06, 'learning_rate/action_model': 9.680620343267802e-05, 'epoch': 0.33}) +[2026-06-17 18:03:13,902][starVLA.training.train_starvla][INFO] - Step 911, Loss: {'train/loss': 1.0067532062530518, 'train/grad_norm_pre_clip': 11.894857406616211, 'timing/data': 0.011784864589571953, 'timing/model': 1.0546260299161077, 'learning_rate/qwen_vl_interface': 9.679840250686403e-06, 'learning_rate/action_model': 9.679840250686404e-05, 'epoch': 0.33}) +[2026-06-17 18:03:21,551][starVLA.training.train_starvla][INFO] - Step 912, Loss: {'train/loss': 0.9843918681144714, 'train/grad_norm_pre_clip': 12.268887519836426, 'timing/data': 0.010941286571323872, 'timing/model': 1.3834384875372052, 'learning_rate/qwen_vl_interface': 9.679059239794355e-06, 'learning_rate/action_model': 9.679059239794355e-05, 'epoch': 0.33}) +[2026-06-17 18:03:26,113][starVLA.training.train_starvla][INFO] - Step 913, Loss: {'train/loss': 0.9571453928947449, 'train/grad_norm_pre_clip': 11.295634269714355, 'timing/data': 0.011494409292936325, 'timing/model': 1.0838610315695405, 'learning_rate/qwen_vl_interface': 9.67827731075356e-06, 'learning_rate/action_model': 9.67827731075356e-05, 'epoch': 0.33}) +[2026-06-17 18:03:30,512][starVLA.training.train_starvla][INFO] - Step 914, Loss: {'train/loss': 1.4402436017990112, 'train/grad_norm_pre_clip': 40.316036224365234, 'timing/data': 0.012175499461591244, 'timing/model': 1.0804513795301318, 'learning_rate/qwen_vl_interface': 9.677494463726112e-06, 'learning_rate/action_model': 9.677494463726112e-05, 'epoch': 0.33}) +[2026-06-17 18:03:34,809][starVLA.training.train_starvla][INFO] - Step 915, Loss: {'train/loss': 0.7418903112411499, 'train/grad_norm_pre_clip': 11.196112632751465, 'timing/data': 0.012455610558390617, 'timing/model': 1.0219754772260785, 'learning_rate/qwen_vl_interface': 9.676710698874299e-06, 'learning_rate/action_model': 9.676710698874298e-05, 'epoch': 0.34}) +[2026-06-17 18:03:39,212][starVLA.training.train_starvla][INFO] - Step 916, Loss: {'train/loss': 0.8442214131355286, 'train/grad_norm_pre_clip': 8.445097923278809, 'timing/data': 0.01605482306331396, 'timing/model': 1.053974324837327, 'learning_rate/qwen_vl_interface': 9.675926016360595e-06, 'learning_rate/action_model': 9.675926016360595e-05, 'epoch': 0.34}) +[2026-06-17 18:03:44,645][starVLA.training.train_starvla][INFO] - Step 917, Loss: {'train/loss': 1.1103368997573853, 'train/grad_norm_pre_clip': 9.946599006652832, 'timing/data': 0.014104409143328667, 'timing/model': 1.038337741047144, 'learning_rate/qwen_vl_interface': 9.675140416347664e-06, 'learning_rate/action_model': 9.675140416347663e-05, 'epoch': 0.34}) +[2026-06-17 18:03:49,112][starVLA.training.train_starvla][INFO] - Step 918, Loss: {'train/loss': 1.0484261512756348, 'train/grad_norm_pre_clip': 11.251388549804688, 'timing/data': 0.014475377276539803, 'timing/model': 1.343057706952095, 'learning_rate/qwen_vl_interface': 9.674353898998362e-06, 'learning_rate/action_model': 9.674353898998363e-05, 'epoch': 0.34}) +[2026-06-17 18:03:55,171][starVLA.training.train_starvla][INFO] - Step 919, Loss: {'train/loss': 0.7755177617073059, 'train/grad_norm_pre_clip': 11.34393310546875, 'timing/data': 0.011467020027339458, 'timing/model': 1.0621319264173508, 'learning_rate/qwen_vl_interface': 9.673566464475738e-06, 'learning_rate/action_model': 9.673566464475739e-05, 'epoch': 0.34}) +[2026-06-17 18:03:59,807][starVLA.training.train_starvla][INFO] - Step 920, Loss: {'train/loss': 1.2124035358428955, 'train/grad_norm_pre_clip': 12.789381980895996, 'timing/data': 0.01905122399330139, 'timing/model': 1.0616549579426646, 'learning_rate/qwen_vl_interface': 9.672778112943026e-06, 'learning_rate/action_model': 9.672778112943026e-05, 'epoch': 0.34}) +[2026-06-17 18:04:04,084][starVLA.training.train_starvla][INFO] - Step 921, Loss: {'train/loss': 0.9749798774719238, 'train/grad_norm_pre_clip': 13.10033130645752, 'timing/data': 0.010445178486406803, 'timing/model': 1.0534600708633661, 'learning_rate/qwen_vl_interface': 9.671988844563652e-06, 'learning_rate/action_model': 9.671988844563652e-05, 'epoch': 0.34}) +[2026-06-17 18:04:08,768][starVLA.training.train_starvla][INFO] - Step 922, Loss: {'train/loss': 1.0638742446899414, 'train/grad_norm_pre_clip': 17.36373519897461, 'timing/data': 0.015495925210416317, 'timing/model': 1.0410010870546103, 'learning_rate/qwen_vl_interface': 9.671198659501233e-06, 'learning_rate/action_model': 9.671198659501231e-05, 'epoch': 0.34}) +[2026-06-17 18:04:13,259][starVLA.training.train_starvla][INFO] - Step 923, Loss: {'train/loss': 1.0028152465820312, 'train/grad_norm_pre_clip': 9.856432914733887, 'timing/data': 0.01063575316220522, 'timing/model': 1.4611949296668172, 'learning_rate/qwen_vl_interface': 9.670407557919573e-06, 'learning_rate/action_model': 9.670407557919572e-05, 'epoch': 0.34}) +[2026-06-17 18:04:17,434][starVLA.training.train_starvla][INFO] - Step 924, Loss: {'train/loss': 0.8657914400100708, 'train/grad_norm_pre_clip': 10.287715911865234, 'timing/data': 0.01668324787169695, 'timing/model': 1.1476537063717842, 'learning_rate/qwen_vl_interface': 9.669615539982668e-06, 'learning_rate/action_model': 9.669615539982668e-05, 'epoch': 0.34}) +[2026-06-17 18:04:21,907][starVLA.training.train_starvla][INFO] - Step 925, Loss: {'train/loss': 0.9097200036048889, 'train/grad_norm_pre_clip': 9.223382949829102, 'timing/data': 0.010989591479301453, 'timing/model': 1.1231073681265116, 'learning_rate/qwen_vl_interface': 9.668822605854708e-06, 'learning_rate/action_model': 9.668822605854706e-05, 'epoch': 0.34}) +[2026-06-17 18:04:26,309][starVLA.training.train_starvla][INFO] - Step 926, Loss: {'train/loss': 0.9192983508110046, 'train/grad_norm_pre_clip': 7.776662349700928, 'timing/data': 0.013129987753927708, 'timing/model': 1.0614602668210864, 'learning_rate/qwen_vl_interface': 9.668028755700067e-06, 'learning_rate/action_model': 9.668028755700067e-05, 'epoch': 0.34}) +[2026-06-17 18:04:30,856][starVLA.training.train_starvla][INFO] - Step 927, Loss: {'train/loss': 0.822144091129303, 'train/grad_norm_pre_clip': 9.612934112548828, 'timing/data': 0.01351435761898756, 'timing/model': 1.0736463451758027, 'learning_rate/qwen_vl_interface': 9.66723398968331e-06, 'learning_rate/action_model': 9.66723398968331e-05, 'epoch': 0.34}) +[2026-06-17 18:04:35,331][starVLA.training.train_starvla][INFO] - Step 928, Loss: {'train/loss': 0.9506554007530212, 'train/grad_norm_pre_clip': 9.138736724853516, 'timing/data': 0.01184970885515213, 'timing/model': 1.3695495473220944, 'learning_rate/qwen_vl_interface': 9.666438307969194e-06, 'learning_rate/action_model': 9.666438307969195e-05, 'epoch': 0.34}) +[2026-06-17 18:04:39,408][starVLA.training.train_starvla][INFO] - Step 929, Loss: {'train/loss': 0.7335756421089172, 'train/grad_norm_pre_clip': 9.257335662841797, 'timing/data': 0.03764175437390804, 'timing/model': 1.0268368627876043, 'learning_rate/qwen_vl_interface': 9.665641710722666e-06, 'learning_rate/action_model': 9.665641710722665e-05, 'epoch': 0.34}) +[2026-06-17 18:04:43,677][starVLA.training.train_starvla][INFO] - Step 930, Loss: {'train/loss': 1.1170943975448608, 'train/grad_norm_pre_clip': 10.692550659179688, 'timing/data': 0.03848003875464201, 'timing/model': 1.0132452826946974, 'learning_rate/qwen_vl_interface': 9.664844198108858e-06, 'learning_rate/action_model': 9.664844198108858e-05, 'epoch': 0.34}) +[2026-06-17 18:04:49,197][starVLA.training.train_starvla][INFO] - Step 931, Loss: {'train/loss': 0.7884868383407593, 'train/grad_norm_pre_clip': 8.487523078918457, 'timing/data': 0.036324755288660526, 'timing/model': 1.0412158919498324, 'learning_rate/qwen_vl_interface': 9.6640457702931e-06, 'learning_rate/action_model': 9.664045770293099e-05, 'epoch': 0.34}) +[2026-06-17 18:04:53,709][starVLA.training.train_starvla][INFO] - Step 932, Loss: {'train/loss': 0.9176535606384277, 'train/grad_norm_pre_clip': 9.37562084197998, 'timing/data': 0.04454992711544037, 'timing/model': 1.2041033310815692, 'learning_rate/qwen_vl_interface': 9.663246427440901e-06, 'learning_rate/action_model': 9.6632464274409e-05, 'epoch': 0.34}) +[2026-06-17 18:04:58,211][starVLA.training.train_starvla][INFO] - Step 933, Loss: {'train/loss': 0.8830446004867554, 'train/grad_norm_pre_clip': 10.367717742919922, 'timing/data': 0.039928561076521873, 'timing/model': 1.342868528328836, 'learning_rate/qwen_vl_interface': 9.662446169717969e-06, 'learning_rate/action_model': 9.662446169717969e-05, 'epoch': 0.34}) +[2026-06-17 18:05:02,322][starVLA.training.train_starvla][INFO] - Step 934, Loss: {'train/loss': 1.0559221506118774, 'train/grad_norm_pre_clip': 9.631749153137207, 'timing/data': 0.04217646550387144, 'timing/model': 1.0755740823224187, 'learning_rate/qwen_vl_interface': 9.6616449972902e-06, 'learning_rate/action_model': 9.6616449972902e-05, 'epoch': 0.34}) +[2026-06-17 18:05:06,848][starVLA.training.train_starvla][INFO] - Step 935, Loss: {'train/loss': 1.2356748580932617, 'train/grad_norm_pre_clip': 12.237859725952148, 'timing/data': 0.04129733704030514, 'timing/model': 1.109835415147245, 'learning_rate/qwen_vl_interface': 9.660842910323675e-06, 'learning_rate/action_model': 9.660842910323674e-05, 'epoch': 0.34}) +[2026-06-17 18:05:12,274][starVLA.training.train_starvla][INFO] - Step 936, Loss: {'train/loss': 1.060668706893921, 'train/grad_norm_pre_clip': 9.236343383789062, 'timing/data': 0.044547852128744125, 'timing/model': 1.04835819080472, 'learning_rate/qwen_vl_interface': 9.660039908984667e-06, 'learning_rate/action_model': 9.660039908984666e-05, 'epoch': 0.34}) +[2026-06-17 18:05:16,881][starVLA.training.train_starvla][INFO] - Step 937, Loss: {'train/loss': 0.9633055925369263, 'train/grad_norm_pre_clip': 41.05470275878906, 'timing/data': 0.047292061150074005, 'timing/model': 1.031828106380999, 'learning_rate/qwen_vl_interface': 9.659235993439643e-06, 'learning_rate/action_model': 9.659235993439643e-05, 'epoch': 0.34}) +[2026-06-17 18:05:28,692][starVLA.training.train_starvla][INFO] - Step 938, Loss: {'train/loss': 1.0078095197677612, 'train/grad_norm_pre_clip': 10.339173316955566, 'timing/data': 0.05114362295717001, 'timing/model': 5.582087859511375, 'learning_rate/qwen_vl_interface': 9.658431163855251e-06, 'learning_rate/action_model': 9.658431163855251e-05, 'epoch': 0.34}) +[2026-06-17 18:05:36,706][starVLA.training.train_starvla][INFO] - Step 939, Loss: {'train/loss': 0.5715228915214539, 'train/grad_norm_pre_clip': 8.582784652709961, 'timing/data': 0.041006747633218765, 'timing/model': 4.519414566457272, 'learning_rate/qwen_vl_interface': 9.657625420398334e-06, 'learning_rate/action_model': 9.657625420398333e-05, 'epoch': 0.34}) +[2026-06-17 18:05:41,544][starVLA.training.train_starvla][INFO] - Step 940, Loss: {'train/loss': 0.7889337539672852, 'train/grad_norm_pre_clip': 11.740036964416504, 'timing/data': 0.09811725467443466, 'timing/model': 1.2174455309286714, 'learning_rate/qwen_vl_interface': 9.656818763235927e-06, 'learning_rate/action_model': 9.656818763235926e-05, 'epoch': 0.34}) +[2026-06-17 18:05:46,114][starVLA.training.train_starvla][INFO] - Step 941, Loss: {'train/loss': 1.099202036857605, 'train/grad_norm_pre_clip': 10.996489524841309, 'timing/data': 0.044765934348106384, 'timing/model': 1.0778473541140556, 'learning_rate/qwen_vl_interface': 9.656011192535248e-06, 'learning_rate/action_model': 9.656011192535247e-05, 'epoch': 0.34}) +[2026-06-17 18:05:50,930][starVLA.training.train_starvla][INFO] - Step 942, Loss: {'train/loss': 1.0629173517227173, 'train/grad_norm_pre_clip': 10.545608520507812, 'timing/data': 0.040695332922041416, 'timing/model': 1.1324086235836148, 'learning_rate/qwen_vl_interface': 9.655202708463706e-06, 'learning_rate/action_model': 9.655202708463707e-05, 'epoch': 0.34}) +[2026-06-17 18:05:55,596][starVLA.training.train_starvla][INFO] - Step 943, Loss: {'train/loss': 0.949765145778656, 'train/grad_norm_pre_clip': 11.072797775268555, 'timing/data': 0.04395685996860266, 'timing/model': 1.3942247088998556, 'learning_rate/qwen_vl_interface': 9.654393311188905e-06, 'learning_rate/action_model': 9.654393311188905e-05, 'epoch': 0.35}) +[2026-06-17 18:06:01,373][starVLA.training.train_starvla][INFO] - Step 944, Loss: {'train/loss': 1.1014548540115356, 'train/grad_norm_pre_clip': 11.659842491149902, 'timing/data': 0.12132460065186024, 'timing/model': 2.53502778429538, 'learning_rate/qwen_vl_interface': 9.653583000878631e-06, 'learning_rate/action_model': 9.65358300087863e-05, 'epoch': 0.35}) +[2026-06-17 18:06:06,604][starVLA.training.train_starvla][INFO] - Step 945, Loss: {'train/loss': 0.9668247699737549, 'train/grad_norm_pre_clip': 12.834473609924316, 'timing/data': 0.10077218804508448, 'timing/model': 1.1975432867184281, 'learning_rate/qwen_vl_interface': 9.65277177770086e-06, 'learning_rate/action_model': 9.65277177770086e-05, 'epoch': 0.35}) +[2026-06-17 18:06:16,434][starVLA.training.train_starvla][INFO] - Step 946, Loss: {'train/loss': 0.869422972202301, 'train/grad_norm_pre_clip': 11.413973808288574, 'timing/data': 0.03772685304284096, 'timing/model': 4.633789937011898, 'learning_rate/qwen_vl_interface': 9.651959641823764e-06, 'learning_rate/action_model': 9.651959641823763e-05, 'epoch': 0.35}) +[2026-06-17 18:06:21,180][starVLA.training.train_starvla][INFO] - Step 947, Loss: {'train/loss': 0.7506346106529236, 'train/grad_norm_pre_clip': 9.598124504089355, 'timing/data': 0.042660477571189404, 'timing/model': 1.1510938042774796, 'learning_rate/qwen_vl_interface': 9.651146593415696e-06, 'learning_rate/action_model': 9.651146593415695e-05, 'epoch': 0.35}) +[2026-06-17 18:06:29,511][starVLA.training.train_starvla][INFO] - Step 948, Loss: {'train/loss': 0.9727461934089661, 'train/grad_norm_pre_clip': 10.757107734680176, 'timing/data': 0.043074870482087135, 'timing/model': 1.359434432350099, 'learning_rate/qwen_vl_interface': 9.650332632645206e-06, 'learning_rate/action_model': 9.650332632645205e-05, 'epoch': 0.35}) +[2026-06-17 18:06:34,371][starVLA.training.train_starvla][INFO] - Step 949, Loss: {'train/loss': 1.1347302198410034, 'train/grad_norm_pre_clip': 13.6275053024292, 'timing/data': 0.04036126472055912, 'timing/model': 1.127735085785389, 'learning_rate/qwen_vl_interface': 9.649517759681022e-06, 'learning_rate/action_model': 9.649517759681023e-05, 'epoch': 0.35}) +[2026-06-17 18:06:42,277][starVLA.training.train_starvla][INFO] - Step 950, Loss: {'train/loss': 1.0670071840286255, 'train/grad_norm_pre_clip': 12.22091007232666, 'timing/data': 0.04912151303142309, 'timing/model': 1.18061713129282, 'learning_rate/qwen_vl_interface': 9.648701974692075e-06, 'learning_rate/action_model': 9.648701974692075e-05, 'epoch': 0.35}) +[2026-06-17 18:06:46,996][starVLA.training.train_starvla][INFO] - Step 951, Loss: {'train/loss': 0.7698935270309448, 'train/grad_norm_pre_clip': 7.462150573730469, 'timing/data': 0.04666382074356079, 'timing/model': 1.2538911942392588, 'learning_rate/qwen_vl_interface': 9.647885277847476e-06, 'learning_rate/action_model': 9.647885277847476e-05, 'epoch': 0.35}) +[2026-06-17 18:06:52,561][starVLA.training.train_starvla][INFO] - Step 952, Loss: {'train/loss': 1.005205750465393, 'train/grad_norm_pre_clip': 9.77520751953125, 'timing/data': 0.04290372133255005, 'timing/model': 1.1167774256318808, 'learning_rate/qwen_vl_interface': 9.647067669316525e-06, 'learning_rate/action_model': 9.647067669316525e-05, 'epoch': 0.35}) +[2026-06-17 18:06:57,215][starVLA.training.train_starvla][INFO] - Step 953, Loss: {'train/loss': 1.0025832653045654, 'train/grad_norm_pre_clip': 14.326139450073242, 'timing/data': 0.04248928278684616, 'timing/model': 1.3131229085847735, 'learning_rate/qwen_vl_interface': 9.646249149268716e-06, 'learning_rate/action_model': 9.646249149268715e-05, 'epoch': 0.35}) +[2026-06-17 18:07:01,375][starVLA.training.train_starvla][INFO] - Step 954, Loss: {'train/loss': 0.7143807411193848, 'train/grad_norm_pre_clip': 10.120884895324707, 'timing/data': 0.03895787335932255, 'timing/model': 1.029020861722529, 'learning_rate/qwen_vl_interface': 9.645429717873728e-06, 'learning_rate/action_model': 9.645429717873727e-05, 'epoch': 0.35}) +[2026-06-17 18:07:05,917][starVLA.training.train_starvla][INFO] - Step 955, Loss: {'train/loss': 0.6449514627456665, 'train/grad_norm_pre_clip': 7.637849807739258, 'timing/data': 0.04244587663561106, 'timing/model': 1.128320710733533, 'learning_rate/qwen_vl_interface': 9.644609375301429e-06, 'learning_rate/action_model': 9.64460937530143e-05, 'epoch': 0.35}) +[2026-06-17 18:07:10,415][starVLA.training.train_starvla][INFO] - Step 956, Loss: {'train/loss': 1.0235283374786377, 'train/grad_norm_pre_clip': 10.915735244750977, 'timing/data': 0.0412450572475791, 'timing/model': 1.1308911368250847, 'learning_rate/qwen_vl_interface': 9.643788121721877e-06, 'learning_rate/action_model': 9.643788121721876e-05, 'epoch': 0.35}) +[2026-06-17 18:07:14,816][starVLA.training.train_starvla][INFO] - Step 957, Loss: {'train/loss': 0.9202315211296082, 'train/grad_norm_pre_clip': 9.250683784484863, 'timing/data': 0.042822981253266335, 'timing/model': 1.1286487188190222, 'learning_rate/qwen_vl_interface': 9.64296595730532e-06, 'learning_rate/action_model': 9.64296595730532e-05, 'epoch': 0.35}) +[2026-06-17 18:07:19,456][starVLA.training.train_starvla][INFO] - Step 958, Loss: {'train/loss': 1.0723718404769897, 'train/grad_norm_pre_clip': 10.846312522888184, 'timing/data': 0.03923780284821987, 'timing/model': 1.4570507360622287, 'learning_rate/qwen_vl_interface': 9.642142882222193e-06, 'learning_rate/action_model': 9.642142882222193e-05, 'epoch': 0.35}) +[2026-06-17 18:07:23,759][starVLA.training.train_starvla][INFO] - Step 959, Loss: {'train/loss': 0.9126356244087219, 'train/grad_norm_pre_clip': 9.238466262817383, 'timing/data': 0.06218421645462513, 'timing/model': 1.1257806038483977, 'learning_rate/qwen_vl_interface': 9.641318896643123e-06, 'learning_rate/action_model': 9.641318896643122e-05, 'epoch': 0.35}) +[2026-06-17 18:07:28,141][starVLA.training.train_starvla][INFO] - Step 960, Loss: {'train/loss': 0.8302826285362244, 'train/grad_norm_pre_clip': 9.85767936706543, 'timing/data': 0.04044162109494209, 'timing/model': 1.0439793830737472, 'learning_rate/qwen_vl_interface': 9.640494000738917e-06, 'learning_rate/action_model': 9.640494000738917e-05, 'epoch': 0.35}) +[2026-06-17 18:07:32,578][starVLA.training.train_starvla][INFO] - Step 961, Loss: {'train/loss': 0.8119900822639465, 'train/grad_norm_pre_clip': 9.702649116516113, 'timing/data': 0.04557373747229576, 'timing/model': 1.0216927453875542, 'learning_rate/qwen_vl_interface': 9.63966819468058e-06, 'learning_rate/action_model': 9.63966819468058e-05, 'epoch': 0.35}) +[2026-06-17 18:07:37,172][starVLA.training.train_starvla][INFO] - Step 962, Loss: {'train/loss': 0.8073103427886963, 'train/grad_norm_pre_clip': 9.934326171875, 'timing/data': 0.040681508369743824, 'timing/model': 1.038812829181552, 'learning_rate/qwen_vl_interface': 9.638841478639306e-06, 'learning_rate/action_model': 9.638841478639306e-05, 'epoch': 0.35}) +[2026-06-17 18:07:41,957][starVLA.training.train_starvla][INFO] - Step 963, Loss: {'train/loss': 1.2835311889648438, 'train/grad_norm_pre_clip': 11.980247497558594, 'timing/data': 0.05786530300974846, 'timing/model': 1.5674489121884108, 'learning_rate/qwen_vl_interface': 9.638013852786468e-06, 'learning_rate/action_model': 9.638013852786467e-05, 'epoch': 0.35}) +[2026-06-17 18:07:46,001][starVLA.training.train_starvla][INFO] - Step 964, Loss: {'train/loss': 0.8423565030097961, 'train/grad_norm_pre_clip': 9.715947151184082, 'timing/data': 0.04537427891045809, 'timing/model': 1.0331412134692073, 'learning_rate/qwen_vl_interface': 9.637185317293638e-06, 'learning_rate/action_model': 9.637185317293638e-05, 'epoch': 0.35}) +[2026-06-17 18:07:50,679][starVLA.training.train_starvla][INFO] - Step 965, Loss: {'train/loss': 0.7106943130493164, 'train/grad_norm_pre_clip': 8.01213264465332, 'timing/data': 0.04245064314454794, 'timing/model': 1.1021882193163037, 'learning_rate/qwen_vl_interface': 9.63635587233257e-06, 'learning_rate/action_model': 9.63635587233257e-05, 'epoch': 0.35}) +[2026-06-17 18:07:55,126][starVLA.training.train_starvla][INFO] - Step 966, Loss: {'train/loss': 0.88908451795578, 'train/grad_norm_pre_clip': 9.090853691101074, 'timing/data': 0.04142476059496403, 'timing/model': 1.0407922891899943, 'learning_rate/qwen_vl_interface': 9.63552551807521e-06, 'learning_rate/action_model': 9.63552551807521e-05, 'epoch': 0.35}) +[2026-06-17 18:08:00,434][starVLA.training.train_starvla][INFO] - Step 967, Loss: {'train/loss': 0.8654416799545288, 'train/grad_norm_pre_clip': 10.473362922668457, 'timing/data': 0.0393324950709939, 'timing/model': 2.0371974194422364, 'learning_rate/qwen_vl_interface': 9.634694254693691e-06, 'learning_rate/action_model': 9.634694254693692e-05, 'epoch': 0.35}) +[2026-06-17 18:08:06,496][starVLA.training.train_starvla][INFO] - Step 968, Loss: {'train/loss': 0.7982006669044495, 'train/grad_norm_pre_clip': 11.3764066696167, 'timing/data': 0.04069965519011021, 'timing/model': 1.3789451522752643, 'learning_rate/qwen_vl_interface': 9.633862082360336e-06, 'learning_rate/action_model': 9.633862082360335e-05, 'epoch': 0.35}) +[2026-06-17 18:08:14,120][starVLA.training.train_starvla][INFO] - Step 969, Loss: {'train/loss': 0.791663646697998, 'train/grad_norm_pre_clip': 10.171143531799316, 'timing/data': 0.046869173645973206, 'timing/model': 1.105478529818356, 'learning_rate/qwen_vl_interface': 9.633029001247654e-06, 'learning_rate/action_model': 9.633029001247653e-05, 'epoch': 0.35}) +[2026-06-17 18:08:18,493][starVLA.training.train_starvla][INFO] - Step 970, Loss: {'train/loss': 0.980736255645752, 'train/grad_norm_pre_clip': 12.147969245910645, 'timing/data': 0.04663263075053692, 'timing/model': 1.0729532642289996, 'learning_rate/qwen_vl_interface': 9.63219501152834e-06, 'learning_rate/action_model': 9.632195011528341e-05, 'epoch': 0.36}) +[2026-06-17 18:08:26,448][starVLA.training.train_starvla][INFO] - Step 971, Loss: {'train/loss': 0.8732982873916626, 'train/grad_norm_pre_clip': 10.443941116333008, 'timing/data': 0.049491177313029766, 'timing/model': 1.1400355007499456, 'learning_rate/qwen_vl_interface': 9.63136011337529e-06, 'learning_rate/action_model': 9.63136011337529e-05, 'epoch': 0.36}) +[2026-06-17 18:08:30,988][starVLA.training.train_starvla][INFO] - Step 972, Loss: {'train/loss': 0.9860720634460449, 'train/grad_norm_pre_clip': 10.657692909240723, 'timing/data': 0.04134642984718084, 'timing/model': 1.1906393775716424, 'learning_rate/qwen_vl_interface': 9.63052430696157e-06, 'learning_rate/action_model': 9.63052430696157e-05, 'epoch': 0.36}) +[2026-06-17 18:08:37,114][starVLA.training.train_starvla][INFO] - Step 973, Loss: {'train/loss': 0.8060486912727356, 'train/grad_norm_pre_clip': 11.562755584716797, 'timing/data': 0.04389039985835552, 'timing/model': 3.020634835585952, 'learning_rate/qwen_vl_interface': 9.629687592460449e-06, 'learning_rate/action_model': 9.62968759246045e-05, 'epoch': 0.36}) +[2026-06-17 18:08:42,582][starVLA.training.train_starvla][INFO] - Step 974, Loss: {'train/loss': 0.8742403984069824, 'train/grad_norm_pre_clip': 14.485472679138184, 'timing/data': 0.03979714214801788, 'timing/model': 1.0241273157298565, 'learning_rate/qwen_vl_interface': 9.628849970045378e-06, 'learning_rate/action_model': 9.628849970045378e-05, 'epoch': 0.36}) +[2026-06-17 18:08:47,598][starVLA.training.train_starvla][INFO] - Step 975, Loss: {'train/loss': 1.0684055089950562, 'train/grad_norm_pre_clip': 14.57805347442627, 'timing/data': 0.07449853979051113, 'timing/model': 1.038153750821948, 'learning_rate/qwen_vl_interface': 9.628011439889997e-06, 'learning_rate/action_model': 9.628011439889996e-05, 'epoch': 0.36}) +[2026-06-17 18:08:52,083][starVLA.training.train_starvla][INFO] - Step 976, Loss: {'train/loss': 1.2500431537628174, 'train/grad_norm_pre_clip': 16.925832748413086, 'timing/data': 0.03835865668952465, 'timing/model': 1.0739370258525014, 'learning_rate/qwen_vl_interface': 9.62717200216813e-06, 'learning_rate/action_model': 9.62717200216813e-05, 'epoch': 0.36}) +[2026-06-17 18:08:56,340][starVLA.training.train_starvla][INFO] - Step 977, Loss: {'train/loss': 1.1499019861221313, 'train/grad_norm_pre_clip': 12.903468132019043, 'timing/data': 0.03950256574898958, 'timing/model': 1.0188549682497978, 'learning_rate/qwen_vl_interface': 9.6263316570538e-06, 'learning_rate/action_model': 9.6263316570538e-05, 'epoch': 0.36}) +[2026-06-17 18:09:00,765][starVLA.training.train_starvla][INFO] - Step 978, Loss: {'train/loss': 1.1941581964492798, 'train/grad_norm_pre_clip': 15.66463851928711, 'timing/data': 0.04657704662531614, 'timing/model': 1.3628853121772408, 'learning_rate/qwen_vl_interface': 9.625490404721208e-06, 'learning_rate/action_model': 9.625490404721208e-05, 'epoch': 0.36}) +[2026-06-17 18:09:04,789][starVLA.training.train_starvla][INFO] - Step 979, Loss: {'train/loss': 0.9598749876022339, 'train/grad_norm_pre_clip': 16.418537139892578, 'timing/data': 0.03762239683419466, 'timing/model': 1.057525347918272, 'learning_rate/qwen_vl_interface': 9.624648245344747e-06, 'learning_rate/action_model': 9.624648245344746e-05, 'epoch': 0.36}) +[2026-06-17 18:09:09,102][starVLA.training.train_starvla][INFO] - Step 980, Loss: {'train/loss': 1.41819167137146, 'train/grad_norm_pre_clip': 12.009950637817383, 'timing/data': 0.04192875698208809, 'timing/model': 1.0196007741615176, 'learning_rate/qwen_vl_interface': 9.623805179098996e-06, 'learning_rate/action_model': 9.623805179098996e-05, 'epoch': 0.36}) +[2026-06-17 18:09:13,532][starVLA.training.train_starvla][INFO] - Step 981, Loss: {'train/loss': 1.0336028337478638, 'train/grad_norm_pre_clip': 11.632628440856934, 'timing/data': 0.04359641298651695, 'timing/model': 1.0222405605018139, 'learning_rate/qwen_vl_interface': 9.622961206158726e-06, 'learning_rate/action_model': 9.622961206158726e-05, 'epoch': 0.36}) +[2026-06-17 18:09:17,946][starVLA.training.train_starvla][INFO] - Step 982, Loss: {'train/loss': 1.1798025369644165, 'train/grad_norm_pre_clip': 16.785062789916992, 'timing/data': 0.042444515973329544, 'timing/model': 1.020671023055911, 'learning_rate/qwen_vl_interface': 9.622116326698894e-06, 'learning_rate/action_model': 9.622116326698893e-05, 'epoch': 0.36}) +[2026-06-17 18:09:22,207][starVLA.training.train_starvla][INFO] - Step 983, Loss: {'train/loss': 0.8901058435440063, 'train/grad_norm_pre_clip': 11.361884117126465, 'timing/data': 0.03816376347094774, 'timing/model': 1.3166226716712117, 'learning_rate/qwen_vl_interface': 9.621270540894639e-06, 'learning_rate/action_model': 9.621270540894639e-05, 'epoch': 0.36}) +[2026-06-17 18:09:26,332][starVLA.training.train_starvla][INFO] - Step 984, Loss: {'train/loss': 0.9273363947868347, 'train/grad_norm_pre_clip': 10.211050033569336, 'timing/data': 0.04390178807079792, 'timing/model': 1.0872358260676265, 'learning_rate/qwen_vl_interface': 9.6204238489213e-06, 'learning_rate/action_model': 9.6204238489213e-05, 'epoch': 0.36}) +[2026-06-17 18:09:31,888][starVLA.training.train_starvla][INFO] - Step 985, Loss: {'train/loss': 1.2338756322860718, 'train/grad_norm_pre_clip': 14.311583518981934, 'timing/data': 0.042056869715452194, 'timing/model': 1.064174716360867, 'learning_rate/qwen_vl_interface': 9.619576250954393e-06, 'learning_rate/action_model': 9.619576250954393e-05, 'epoch': 0.36}) +[2026-06-17 18:09:36,507][starVLA.training.train_starvla][INFO] - Step 986, Loss: {'train/loss': 0.9114664196968079, 'train/grad_norm_pre_clip': 9.67786693572998, 'timing/data': 0.04758071806281805, 'timing/model': 1.050258838571608, 'learning_rate/qwen_vl_interface': 9.618727747169628e-06, 'learning_rate/action_model': 9.618727747169627e-05, 'epoch': 0.36}) +[2026-06-17 18:09:40,950][starVLA.training.train_starvla][INFO] - Step 987, Loss: {'train/loss': 1.0230202674865723, 'train/grad_norm_pre_clip': 10.151883125305176, 'timing/data': 0.03963928949087858, 'timing/model': 1.0385527815669775, 'learning_rate/qwen_vl_interface': 9.617878337742902e-06, 'learning_rate/action_model': 9.617878337742902e-05, 'epoch': 0.36}) +[2026-06-17 18:09:45,581][starVLA.training.train_starvla][INFO] - Step 988, Loss: {'train/loss': 0.960711658000946, 'train/grad_norm_pre_clip': 12.510005950927734, 'timing/data': 0.044357613660395145, 'timing/model': 1.502042654901743, 'learning_rate/qwen_vl_interface': 9.617028022850294e-06, 'learning_rate/action_model': 9.617028022850293e-05, 'epoch': 0.36}) +[2026-06-17 18:09:49,892][starVLA.training.train_starvla][INFO] - Step 989, Loss: {'train/loss': 1.2190523147583008, 'train/grad_norm_pre_clip': 11.767693519592285, 'timing/data': 0.038944208063185215, 'timing/model': 1.0487529579550028, 'learning_rate/qwen_vl_interface': 9.616176802668078e-06, 'learning_rate/action_model': 9.616176802668078e-05, 'epoch': 0.36}) +[2026-06-17 18:09:54,564][starVLA.training.train_starvla][INFO] - Step 990, Loss: {'train/loss': 0.9879220128059387, 'train/grad_norm_pre_clip': 12.287003517150879, 'timing/data': 0.038900201208889484, 'timing/model': 1.323858874849975, 'learning_rate/qwen_vl_interface': 9.615324677372712e-06, 'learning_rate/action_model': 9.615324677372713e-05, 'epoch': 0.36}) +[2026-06-17 18:09:59,062][starVLA.training.train_starvla][INFO] - Step 991, Loss: {'train/loss': 0.9983266592025757, 'train/grad_norm_pre_clip': 10.906732559204102, 'timing/data': 0.04415036924183369, 'timing/model': 1.0598299549892545, 'learning_rate/qwen_vl_interface': 9.614471647140844e-06, 'learning_rate/action_model': 9.614471647140844e-05, 'epoch': 0.36}) +[2026-06-17 18:10:03,350][starVLA.training.train_starvla][INFO] - Step 992, Loss: {'train/loss': 0.732530951499939, 'train/grad_norm_pre_clip': 10.477814674377441, 'timing/data': 0.043803444132208824, 'timing/model': 1.0514288330450654, 'learning_rate/qwen_vl_interface': 9.613617712149309e-06, 'learning_rate/action_model': 9.613617712149308e-05, 'epoch': 0.36}) +[2026-06-17 18:10:07,809][starVLA.training.train_starvla][INFO] - Step 993, Loss: {'train/loss': 0.9287348985671997, 'train/grad_norm_pre_clip': 14.638524055480957, 'timing/data': 0.05642438028007746, 'timing/model': 1.340108129195869, 'learning_rate/qwen_vl_interface': 9.612762872575124e-06, 'learning_rate/action_model': 9.612762872575124e-05, 'epoch': 0.36}) +[2026-06-17 18:10:12,905][starVLA.training.train_starvla][INFO] - Step 994, Loss: {'train/loss': 0.7219423651695251, 'train/grad_norm_pre_clip': 9.948741912841797, 'timing/data': 0.04125869460403919, 'timing/model': 1.0872118547558784, 'learning_rate/qwen_vl_interface': 9.611907128595503e-06, 'learning_rate/action_model': 9.611907128595502e-05, 'epoch': 0.36}) +[2026-06-17 18:10:17,388][starVLA.training.train_starvla][INFO] - Step 995, Loss: {'train/loss': 1.1809669733047485, 'train/grad_norm_pre_clip': 9.229206085205078, 'timing/data': 0.04501454904675484, 'timing/model': 1.035498403944075, 'learning_rate/qwen_vl_interface': 9.61105048038784e-06, 'learning_rate/action_model': 9.611050480387839e-05, 'epoch': 0.36}) +[2026-06-17 18:10:25,122][starVLA.training.train_starvla][INFO] - Step 996, Loss: {'train/loss': 0.9478445649147034, 'train/grad_norm_pre_clip': 11.116229057312012, 'timing/data': 0.041217808611691, 'timing/model': 4.141180318780243, 'learning_rate/qwen_vl_interface': 9.610192928129716e-06, 'learning_rate/action_model': 9.610192928129717e-05, 'epoch': 0.36}) +[2026-06-17 18:10:31,380][starVLA.training.train_starvla][INFO] - Step 997, Loss: {'train/loss': 0.8279128670692444, 'train/grad_norm_pre_clip': 9.366066932678223, 'timing/data': 0.04309805482625961, 'timing/model': 1.19565689843148, 'learning_rate/qwen_vl_interface': 9.609334471998911e-06, 'learning_rate/action_model': 9.609334471998912e-05, 'epoch': 0.37}) +[2026-06-17 18:10:37,013][starVLA.training.train_starvla][INFO] - Step 998, Loss: {'train/loss': 0.843408465385437, 'train/grad_norm_pre_clip': 12.29306697845459, 'timing/data': 0.05036493390798569, 'timing/model': 1.47255892585963, 'learning_rate/qwen_vl_interface': 9.608475112173379e-06, 'learning_rate/action_model': 9.608475112173379e-05, 'epoch': 0.37}) +[2026-06-17 18:10:41,222][starVLA.training.train_starvla][INFO] - Step 999, Loss: {'train/loss': 0.9078732132911682, 'train/grad_norm_pre_clip': 9.949462890625, 'timing/data': 0.04709874000400305, 'timing/model': 1.0660569807514548, 'learning_rate/qwen_vl_interface': 9.607614848831264e-06, 'learning_rate/action_model': 9.607614848831264e-05, 'epoch': 0.37}) +[2026-06-17 18:13:59,730][starVLA.training.train_starvla][INFO] - Step 1000, Loss: {'train/loss': 0.6437767148017883, 'train/grad_norm_pre_clip': 7.45102596282959, 'eval/loss': 0.942751944065094, 'eval/latency_2/loss': 0.942751944065094, 'eval/demon_fire_precision': 0.8286649690796654, 'eval/demon_fire_recall': 0.9531380753138076, 'eval/demon_fire_f1': 0.8865538042420704, 'eval/demon_fire_rate_model': 0.7648859209794101, 'eval/demon_fire_rate_teacher': 0.6649972175848636, 'eval/demon_left_precision': 0.8372434017595308, 'eval/demon_left_recall': 0.6010526315789474, 'eval/demon_left_f1': 0.6997549019607843, 'eval/demon_left_rate_model': 0.1897607122982749, 'eval/demon_left_rate_teacher': 0.2643294379521425, 'eval/demon_right_precision': 0.722158438576349, 'eval/demon_right_recall': 0.7107344632768362, 'eval/demon_right_f1': 0.7164009111617311, 'eval/demon_right_rate_model': 0.24234835837506957, 'eval/demon_right_rate_teacher': 0.24624373956594323, 'eval/demon_move_f1': 0.7080779065612577, 'eval/demon_cc_f1': 0.8242280989132466, 'eval/latency_2/demon_fire_precision': 0.8286649690796654, 'eval/latency_2/demon_fire_recall': 0.9531380753138076, 'eval/latency_2/demon_fire_f1': 0.8865538042420704, 'eval/latency_2/demon_fire_rate_model': 0.7648859209794101, 'eval/latency_2/demon_fire_rate_teacher': 0.6649972175848636, 'eval/latency_2/demon_left_precision': 0.8372434017595308, 'eval/latency_2/demon_left_recall': 0.6010526315789474, 'eval/latency_2/demon_left_f1': 0.6997549019607843, 'eval/latency_2/demon_left_rate_model': 0.1897607122982749, 'eval/latency_2/demon_left_rate_teacher': 0.2643294379521425, 'eval/latency_2/demon_right_precision': 0.722158438576349, 'eval/latency_2/demon_right_recall': 0.7107344632768362, 'eval/latency_2/demon_right_f1': 0.7164009111617311, 'eval/latency_2/demon_right_rate_model': 0.24234835837506957, 'eval/latency_2/demon_right_rate_teacher': 0.24624373956594323, 'eval/latency_2/demon_move_f1': 0.7080779065612577, 'eval/latency_2/demon_cc_f1': 0.8242280989132466, 'timing/data': 0.04460127279162407, 'timing/model': 1.0438632629811764, 'learning_rate/qwen_vl_interface': 9.606753682150904e-06, 'learning_rate/action_model': 9.606753682150903e-05, 'epoch': 0.37}) +[2026-06-17 18:13:59,733][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_1000_state +[2026-06-17 18:13:59,742][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 18:14:08,187][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_1000_state/model.safetensors +[2026-06-17 18:14:22,243][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_1000_state/optimizer.bin +[2026-06-17 18:14:22,243][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_1000_state/sampler.bin +[2026-06-17 18:14:22,244][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_1000_state/sampler_1.bin +[2026-06-17 18:14:22,247][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_1000_state/random_states_0.pkl +[2026-06-17 18:14:22,304][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1000 +[2026-06-17 18:14:22,304][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 18:14:22,321][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 18:14:28,703][starVLA.training.train_starvla][INFO] - Step 1001, Loss: {'train/loss': 0.7679046988487244, 'train/grad_norm_pre_clip': 11.638825416564941, 'timing/data': 0.06762078311294317, 'timing/model': 1.0395624497905374, 'learning_rate/qwen_vl_interface': 9.605891612310815e-06, 'learning_rate/action_model': 9.605891612310814e-05, 'epoch': 0.37}) +[2026-06-17 18:14:33,832][starVLA.training.train_starvla][INFO] - Step 1002, Loss: {'train/loss': 0.8351244330406189, 'train/grad_norm_pre_clip': 9.242891311645508, 'timing/data': 0.09882614202797413, 'timing/model': 1.3985161585733294, 'learning_rate/qwen_vl_interface': 9.60502863948971e-06, 'learning_rate/action_model': 9.60502863948971e-05, 'epoch': 0.37}) +[2026-06-17 18:14:40,185][starVLA.training.train_starvla][INFO] - Step 1003, Loss: {'train/loss': 0.9829500317573547, 'train/grad_norm_pre_clip': 8.705347061157227, 'timing/data': 0.04738644324243069, 'timing/model': 1.0419782809913158, 'learning_rate/qwen_vl_interface': 9.604164763866479e-06, 'learning_rate/action_model': 9.60416476386648e-05, 'epoch': 0.37}) +[2026-06-17 18:14:44,613][starVLA.training.train_starvla][INFO] - Step 1004, Loss: {'train/loss': 1.156143069267273, 'train/grad_norm_pre_clip': 10.465620994567871, 'timing/data': 0.0452893553301692, 'timing/model': 1.3331759301945567, 'learning_rate/qwen_vl_interface': 9.60329998562021e-06, 'learning_rate/action_model': 9.60329998562021e-05, 'epoch': 0.37}) +[2026-06-17 18:14:48,947][starVLA.training.train_starvla][INFO] - Step 1005, Loss: {'train/loss': 1.1294360160827637, 'train/grad_norm_pre_clip': 13.678503036499023, 'timing/data': 0.04728280287235975, 'timing/model': 1.0365998642519116, 'learning_rate/qwen_vl_interface': 9.602434304930168e-06, 'learning_rate/action_model': 9.602434304930168e-05, 'epoch': 0.37}) +[2026-06-17 18:14:55,034][starVLA.training.train_starvla][INFO] - Step 1006, Loss: {'train/loss': 1.0138704776763916, 'train/grad_norm_pre_clip': 9.338358879089355, 'timing/data': 0.04949239734560251, 'timing/model': 1.0750660197809339, 'learning_rate/qwen_vl_interface': 9.601567721975811e-06, 'learning_rate/action_model': 9.601567721975812e-05, 'epoch': 0.37}) +[2026-06-17 18:14:59,525][starVLA.training.train_starvla][INFO] - Step 1007, Loss: {'train/loss': 1.0369014739990234, 'train/grad_norm_pre_clip': 9.47046947479248, 'timing/data': 0.04631149675697088, 'timing/model': 1.1415615109726787, 'learning_rate/qwen_vl_interface': 9.600700236936785e-06, 'learning_rate/action_model': 9.600700236936785e-05, 'epoch': 0.37}) +[2026-06-17 18:15:03,850][starVLA.training.train_starvla][INFO] - Step 1008, Loss: {'train/loss': 0.8799895644187927, 'train/grad_norm_pre_clip': 9.804922103881836, 'timing/data': 0.044656965881586075, 'timing/model': 1.018397694453597, 'learning_rate/qwen_vl_interface': 9.599831849992916e-06, 'learning_rate/action_model': 9.599831849992916e-05, 'epoch': 0.37}) +[2026-06-17 18:15:08,633][starVLA.training.train_starvla][INFO] - Step 1009, Loss: {'train/loss': 0.9918866753578186, 'train/grad_norm_pre_clip': 11.790761947631836, 'timing/data': 0.04264301899820566, 'timing/model': 1.6550971008837223, 'learning_rate/qwen_vl_interface': 9.598962561324225e-06, 'learning_rate/action_model': 9.598962561324225e-05, 'epoch': 0.37}) +[2026-06-17 18:15:13,000][starVLA.training.train_starvla][INFO] - Step 1010, Loss: {'train/loss': 0.8370401263237, 'train/grad_norm_pre_clip': 8.595226287841797, 'timing/data': 0.04902927577495575, 'timing/model': 1.2124350164085627, 'learning_rate/qwen_vl_interface': 9.598092371110916e-06, 'learning_rate/action_model': 9.598092371110916e-05, 'epoch': 0.37}) +[2026-06-17 18:15:17,435][starVLA.training.train_starvla][INFO] - Step 1011, Loss: {'train/loss': 1.1440516710281372, 'train/grad_norm_pre_clip': 11.684041976928711, 'timing/data': 0.051310510374605656, 'timing/model': 1.0733564803376794, 'learning_rate/qwen_vl_interface': 9.59722127953338e-06, 'learning_rate/action_model': 9.597221279533379e-05, 'epoch': 0.37}) +[2026-06-17 18:15:26,219][starVLA.training.train_starvla][INFO] - Step 1012, Loss: {'train/loss': 1.0348833799362183, 'train/grad_norm_pre_clip': 11.726415634155273, 'timing/data': 0.12947022262960672, 'timing/model': 4.106706359423697, 'learning_rate/qwen_vl_interface': 9.596349286772194e-06, 'learning_rate/action_model': 9.596349286772193e-05, 'epoch': 0.37}) +[2026-06-17 18:15:34,894][starVLA.training.train_starvla][INFO] - Step 1013, Loss: {'train/loss': 1.0650010108947754, 'train/grad_norm_pre_clip': 11.140048027038574, 'timing/data': 0.012726345099508762, 'timing/model': 1.0228946758434176, 'learning_rate/qwen_vl_interface': 9.595476393008123e-06, 'learning_rate/action_model': 9.595476393008123e-05, 'epoch': 0.37}) +[2026-06-17 18:15:40,918][starVLA.training.train_starvla][INFO] - Step 1014, Loss: {'train/loss': 1.0861538648605347, 'train/grad_norm_pre_clip': 12.429500579833984, 'timing/data': 0.051894962787628174, 'timing/model': 1.3779598455876112, 'learning_rate/qwen_vl_interface': 9.59460259842212e-06, 'learning_rate/action_model': 9.59460259842212e-05, 'epoch': 0.37}) +[2026-06-17 18:15:45,150][starVLA.training.train_starvla][INFO] - Step 1015, Loss: {'train/loss': 0.9120551347732544, 'train/grad_norm_pre_clip': 10.941944122314453, 'timing/data': 0.05257245060056448, 'timing/model': 1.028700791299343, 'learning_rate/qwen_vl_interface': 9.593727903195323e-06, 'learning_rate/action_model': 9.593727903195324e-05, 'epoch': 0.37}) +[2026-06-17 18:15:49,568][starVLA.training.train_starvla][INFO] - Step 1016, Loss: {'train/loss': 1.1849833726882935, 'train/grad_norm_pre_clip': 10.864523887634277, 'timing/data': 0.03953450080007315, 'timing/model': 1.1468339255079627, 'learning_rate/qwen_vl_interface': 9.59285230750906e-06, 'learning_rate/action_model': 9.592852307509059e-05, 'epoch': 0.37}) +[2026-06-17 18:15:58,680][starVLA.training.train_starvla][INFO] - Step 1017, Loss: {'train/loss': 0.9548633694648743, 'train/grad_norm_pre_clip': 15.806410789489746, 'timing/data': 0.04516854044049978, 'timing/model': 5.38938738591969, 'learning_rate/qwen_vl_interface': 9.591975811544839e-06, 'learning_rate/action_model': 9.591975811544839e-05, 'epoch': 0.37}) +[2026-06-17 18:16:05,205][starVLA.training.train_starvla][INFO] - Step 1018, Loss: {'train/loss': 0.8257431387901306, 'train/grad_norm_pre_clip': 8.5367431640625, 'timing/data': 0.04626335483044386, 'timing/model': 1.3228208720684052, 'learning_rate/qwen_vl_interface': 9.59109841548436e-06, 'learning_rate/action_model': 9.591098415484359e-05, 'epoch': 0.37}) +[2026-06-17 18:16:09,936][starVLA.training.train_starvla][INFO] - Step 1019, Loss: {'train/loss': 0.9179769158363342, 'train/grad_norm_pre_clip': 11.83267879486084, 'timing/data': 0.05547386407852173, 'timing/model': 1.313607795163989, 'learning_rate/qwen_vl_interface': 9.590220119509508e-06, 'learning_rate/action_model': 9.590220119509508e-05, 'epoch': 0.37}) +[2026-06-17 18:16:14,272][starVLA.training.train_starvla][INFO] - Step 1020, Loss: {'train/loss': 1.0969597101211548, 'train/grad_norm_pre_clip': 14.255087852478027, 'timing/data': 0.03772688936442137, 'timing/model': 1.0513585628941655, 'learning_rate/qwen_vl_interface': 9.589340923802355e-06, 'learning_rate/action_model': 9.589340923802354e-05, 'epoch': 0.37}) +[2026-06-17 18:16:19,073][starVLA.training.train_starvla][INFO] - Step 1021, Loss: {'train/loss': 0.9991323947906494, 'train/grad_norm_pre_clip': 10.627488136291504, 'timing/data': 0.04438985791057348, 'timing/model': 1.02592777274549, 'learning_rate/qwen_vl_interface': 9.58846082854516e-06, 'learning_rate/action_model': 9.58846082854516e-05, 'epoch': 0.37}) +[2026-06-17 18:16:24,319][starVLA.training.train_starvla][INFO] - Step 1022, Loss: {'train/loss': 0.8206233382225037, 'train/grad_norm_pre_clip': 9.022098541259766, 'timing/data': 0.04991656634956598, 'timing/model': 1.2079781517386436, 'learning_rate/qwen_vl_interface': 9.587579833920366e-06, 'learning_rate/action_model': 9.587579833920365e-05, 'epoch': 0.37}) +[2026-06-17 18:16:29,076][starVLA.training.train_starvla][INFO] - Step 1023, Loss: {'train/loss': 1.170052409172058, 'train/grad_norm_pre_clip': 13.11557388305664, 'timing/data': 0.05474381800740957, 'timing/model': 1.1182866049930453, 'learning_rate/qwen_vl_interface': 9.586697940110605e-06, 'learning_rate/action_model': 9.586697940110605e-05, 'epoch': 0.37}) +[2026-06-17 18:16:33,772][starVLA.training.train_starvla][INFO] - Step 1024, Loss: {'train/loss': 1.1215823888778687, 'train/grad_norm_pre_clip': 10.01009464263916, 'timing/data': 0.04033909551799297, 'timing/model': 1.3470782078802586, 'learning_rate/qwen_vl_interface': 9.585815147298696e-06, 'learning_rate/action_model': 9.585815147298696e-05, 'epoch': 0.37}) +[2026-06-17 18:16:42,082][starVLA.training.train_starvla][INFO] - Step 1025, Loss: {'train/loss': 1.1825093030929565, 'train/grad_norm_pre_clip': 10.963176727294922, 'timing/data': 0.0388789027929306, 'timing/model': 1.0485943807289004, 'learning_rate/qwen_vl_interface': 9.58493145566764e-06, 'learning_rate/action_model': 9.58493145566764e-05, 'epoch': 0.38}) +[2026-06-17 18:16:46,603][starVLA.training.train_starvla][INFO] - Step 1026, Loss: {'train/loss': 0.7603065967559814, 'train/grad_norm_pre_clip': 15.142120361328125, 'timing/data': 0.04758862406015396, 'timing/model': 1.0775623610243201, 'learning_rate/qwen_vl_interface': 9.584046865400627e-06, 'learning_rate/action_model': 9.584046865400627e-05, 'epoch': 0.38}) +[2026-06-17 18:16:51,146][starVLA.training.train_starvla][INFO] - Step 1027, Loss: {'train/loss': 0.8857142329216003, 'train/grad_norm_pre_clip': 9.50666332244873, 'timing/data': 0.0788765586912632, 'timing/model': 1.0266119195148349, 'learning_rate/qwen_vl_interface': 9.583161376681039e-06, 'learning_rate/action_model': 9.583161376681038e-05, 'epoch': 0.38}) +[2026-06-17 18:16:55,689][starVLA.training.train_starvla][INFO] - Step 1028, Loss: {'train/loss': 0.8434826731681824, 'train/grad_norm_pre_clip': 11.33797550201416, 'timing/data': 0.04271486774086952, 'timing/model': 1.2770995637401938, 'learning_rate/qwen_vl_interface': 9.582274989692432e-06, 'learning_rate/action_model': 9.582274989692432e-05, 'epoch': 0.38}) +[2026-06-17 18:17:00,160][starVLA.training.train_starvla][INFO] - Step 1029, Loss: {'train/loss': 1.1079962253570557, 'train/grad_norm_pre_clip': 12.040177345275879, 'timing/data': 0.04500129912048578, 'timing/model': 1.3376731220632792, 'learning_rate/qwen_vl_interface': 9.58138770461856e-06, 'learning_rate/action_model': 9.58138770461856e-05, 'epoch': 0.38}) +[2026-06-17 18:17:04,869][starVLA.training.train_starvla][INFO] - Step 1030, Loss: {'train/loss': 1.0138760805130005, 'train/grad_norm_pre_clip': 10.8082914352417, 'timing/data': 0.03784042689949274, 'timing/model': 1.2900462122634053, 'learning_rate/qwen_vl_interface': 9.580499521643355e-06, 'learning_rate/action_model': 9.580499521643355e-05, 'epoch': 0.38}) +[2026-06-17 18:17:09,536][starVLA.training.train_starvla][INFO] - Step 1031, Loss: {'train/loss': 0.7718320488929749, 'train/grad_norm_pre_clip': 7.704963684082031, 'timing/data': 0.05324147455394268, 'timing/model': 1.2316743051633239, 'learning_rate/qwen_vl_interface': 9.579610440950939e-06, 'learning_rate/action_model': 9.579610440950937e-05, 'epoch': 0.38}) +[2026-06-17 18:17:15,102][starVLA.training.train_starvla][INFO] - Step 1032, Loss: {'train/loss': 0.9036181569099426, 'train/grad_norm_pre_clip': 9.432272911071777, 'timing/data': 0.035810413770377636, 'timing/model': 1.8213566187769175, 'learning_rate/qwen_vl_interface': 9.57872046272562e-06, 'learning_rate/action_model': 9.57872046272562e-05, 'epoch': 0.38}) +[2026-06-17 18:17:19,461][starVLA.training.train_starvla][INFO] - Step 1033, Loss: {'train/loss': 0.815908670425415, 'train/grad_norm_pre_clip': 10.769959449768066, 'timing/data': 0.04242485109716654, 'timing/model': 1.0238526118919253, 'learning_rate/qwen_vl_interface': 9.577829587151892e-06, 'learning_rate/action_model': 9.577829587151891e-05, 'epoch': 0.38}) +[2026-06-17 18:17:25,707][starVLA.training.train_starvla][INFO] - Step 1034, Loss: {'train/loss': 0.9222283959388733, 'train/grad_norm_pre_clip': 9.51379680633545, 'timing/data': 0.0830354942008853, 'timing/model': 2.883879190310836, 'learning_rate/qwen_vl_interface': 9.576937814414433e-06, 'learning_rate/action_model': 9.576937814414432e-05, 'epoch': 0.38}) +[2026-06-17 18:17:37,707][starVLA.training.train_starvla][INFO] - Step 1035, Loss: {'train/loss': 0.7925607562065125, 'train/grad_norm_pre_clip': 10.319158554077148, 'timing/data': 0.05155555531382561, 'timing/model': 1.1058033788576722, 'learning_rate/qwen_vl_interface': 9.576045144698109e-06, 'learning_rate/action_model': 9.576045144698108e-05, 'epoch': 0.38}) +[2026-06-17 18:17:42,650][starVLA.training.train_starvla][INFO] - Step 1036, Loss: {'train/loss': 0.8800839185714722, 'train/grad_norm_pre_clip': 9.78993034362793, 'timing/data': 0.051103467121720314, 'timing/model': 1.0641897954046726, 'learning_rate/qwen_vl_interface': 9.57515157818797e-06, 'learning_rate/action_model': 9.57515157818797e-05, 'epoch': 0.38}) +[2026-06-17 18:17:47,272][starVLA.training.train_starvla][INFO] - Step 1037, Loss: {'train/loss': 0.8798264265060425, 'train/grad_norm_pre_clip': 12.13564395904541, 'timing/data': 0.044317579828202724, 'timing/model': 1.1452876580879092, 'learning_rate/qwen_vl_interface': 9.574257115069256e-06, 'learning_rate/action_model': 9.574257115069256e-05, 'epoch': 0.38}) +[2026-06-17 18:17:52,254][starVLA.training.train_starvla][INFO] - Step 1038, Loss: {'train/loss': 0.6817413568496704, 'train/grad_norm_pre_clip': 8.093542098999023, 'timing/data': 0.056178295984864235, 'timing/model': 1.2163305105641484, 'learning_rate/qwen_vl_interface': 9.573361755527389e-06, 'learning_rate/action_model': 9.573361755527389e-05, 'epoch': 0.38}) +[2026-06-17 18:17:56,477][starVLA.training.train_starvla][INFO] - Step 1039, Loss: {'train/loss': 0.9089042544364929, 'train/grad_norm_pre_clip': 10.058534622192383, 'timing/data': 0.04708575550466776, 'timing/model': 1.0386544466018677, 'learning_rate/qwen_vl_interface': 9.572465499747978e-06, 'learning_rate/action_model': 9.572465499747978e-05, 'epoch': 0.38}) +[2026-06-17 18:18:07,171][starVLA.training.train_starvla][INFO] - Step 1040, Loss: {'train/loss': 1.0166045427322388, 'train/grad_norm_pre_clip': 10.651217460632324, 'timing/data': 0.058482563123106956, 'timing/model': 1.204893640242517, 'learning_rate/qwen_vl_interface': 9.571568347916819e-06, 'learning_rate/action_model': 9.571568347916819e-05, 'epoch': 0.38}) +[2026-06-17 18:18:12,437][starVLA.training.train_starvla][INFO] - Step 1041, Loss: {'train/loss': 0.8220081925392151, 'train/grad_norm_pre_clip': 8.904889106750488, 'timing/data': 0.042106032371520996, 'timing/model': 1.2646284997463226, 'learning_rate/qwen_vl_interface': 9.570670300219889e-06, 'learning_rate/action_model': 9.570670300219889e-05, 'epoch': 0.38}) +[2026-06-17 18:18:20,211][starVLA.training.train_starvla][INFO] - Step 1042, Loss: {'train/loss': 0.9357768893241882, 'train/grad_norm_pre_clip': 11.003104209899902, 'timing/data': 0.04725081566721201, 'timing/model': 1.0306135704740882, 'learning_rate/qwen_vl_interface': 9.56977135684336e-06, 'learning_rate/action_model': 9.569771356843359e-05, 'epoch': 0.38}) +[2026-06-17 18:18:24,882][starVLA.training.train_starvla][INFO] - Step 1043, Loss: {'train/loss': 1.257198452949524, 'train/grad_norm_pre_clip': 10.327126502990723, 'timing/data': 0.04052351601421833, 'timing/model': 1.0249771028757095, 'learning_rate/qwen_vl_interface': 9.568871517973576e-06, 'learning_rate/action_model': 9.568871517973577e-05, 'epoch': 0.38}) +[2026-06-17 18:18:29,707][starVLA.training.train_starvla][INFO] - Step 1044, Loss: {'train/loss': 0.9304652214050293, 'train/grad_norm_pre_clip': 11.561235427856445, 'timing/data': 0.041020698845386505, 'timing/model': 1.620370619930327, 'learning_rate/qwen_vl_interface': 9.567970783797081e-06, 'learning_rate/action_model': 9.567970783797082e-05, 'epoch': 0.38}) +[2026-06-17 18:18:37,066][starVLA.training.train_starvla][INFO] - Step 1045, Loss: {'train/loss': 1.0605260133743286, 'train/grad_norm_pre_clip': 9.655052185058594, 'timing/data': 0.052625550888478756, 'timing/model': 1.03961475007236, 'learning_rate/qwen_vl_interface': 9.567069154500597e-06, 'learning_rate/action_model': 9.567069154500596e-05, 'epoch': 0.38}) +[2026-06-17 18:18:41,415][starVLA.training.train_starvla][INFO] - Step 1046, Loss: {'train/loss': 0.9097487330436707, 'train/grad_norm_pre_clip': 10.906375885009766, 'timing/data': 0.04238448292016983, 'timing/model': 1.0475951889529824, 'learning_rate/qwen_vl_interface': 9.56616663027103e-06, 'learning_rate/action_model': 9.566166630271029e-05, 'epoch': 0.38}) +[2026-06-17 18:18:45,768][starVLA.training.train_starvla][INFO] - Step 1047, Loss: {'train/loss': 0.9333236813545227, 'train/grad_norm_pre_clip': 10.355889320373535, 'timing/data': 0.04294877499341965, 'timing/model': 1.0226011658087373, 'learning_rate/qwen_vl_interface': 9.56526321129548e-06, 'learning_rate/action_model': 9.565263211295479e-05, 'epoch': 0.38}) +[2026-06-17 18:18:50,235][starVLA.training.train_starvla][INFO] - Step 1048, Loss: {'train/loss': 1.0136109590530396, 'train/grad_norm_pre_clip': 12.374311447143555, 'timing/data': 0.043704014271497726, 'timing/model': 1.0643481146544218, 'learning_rate/qwen_vl_interface': 9.56435889776122e-06, 'learning_rate/action_model': 9.564358897761218e-05, 'epoch': 0.38}) +[2026-06-17 18:18:54,675][starVLA.training.train_starvla][INFO] - Step 1049, Loss: {'train/loss': 0.8361619114875793, 'train/grad_norm_pre_clip': 10.014612197875977, 'timing/data': 0.04324078559875488, 'timing/model': 1.4366826266050339, 'learning_rate/qwen_vl_interface': 9.563453689855717e-06, 'learning_rate/action_model': 9.563453689855716e-05, 'epoch': 0.38}) +[2026-06-17 18:19:03,265][starVLA.training.train_starvla][INFO] - Step 1050, Loss: {'train/loss': 0.865162193775177, 'train/grad_norm_pre_clip': 8.226152420043945, 'timing/data': 0.04549393057823181, 'timing/model': 1.0677971243858337, 'learning_rate/qwen_vl_interface': 9.562547587766624e-06, 'learning_rate/action_model': 9.562547587766623e-05, 'epoch': 0.38}) +[2026-06-17 18:19:08,228][starVLA.training.train_starvla][INFO] - Step 1051, Loss: {'train/loss': 0.7822195291519165, 'train/grad_norm_pre_clip': 12.307512283325195, 'timing/data': 0.04407349694520235, 'timing/model': 1.2451208280399442, 'learning_rate/qwen_vl_interface': 9.561640591681775e-06, 'learning_rate/action_model': 9.561640591681774e-05, 'epoch': 0.38}) +[2026-06-17 18:19:13,131][starVLA.training.train_starvla][INFO] - Step 1052, Loss: {'train/loss': 0.9093707799911499, 'train/grad_norm_pre_clip': 9.432769775390625, 'timing/data': 0.046257736161351204, 'timing/model': 1.158066337928176, 'learning_rate/qwen_vl_interface': 9.560732701789193e-06, 'learning_rate/action_model': 9.560732701789193e-05, 'epoch': 0.39}) +[2026-06-17 18:19:23,898][starVLA.training.train_starvla][INFO] - Step 1053, Loss: {'train/loss': 0.7273436188697815, 'train/grad_norm_pre_clip': 9.347413063049316, 'timing/data': 0.056276327930390835, 'timing/model': 1.2035063728690147, 'learning_rate/qwen_vl_interface': 9.55982391827708e-06, 'learning_rate/action_model': 9.559823918277081e-05, 'epoch': 0.39}) +[2026-06-17 18:19:28,794][starVLA.training.train_starvla][INFO] - Step 1054, Loss: {'train/loss': 0.829276442527771, 'train/grad_norm_pre_clip': 9.800372123718262, 'timing/data': 0.041797847487032413, 'timing/model': 1.6485181981697679, 'learning_rate/qwen_vl_interface': 9.558914241333835e-06, 'learning_rate/action_model': 9.558914241333835e-05, 'epoch': 0.39}) +[2026-06-17 18:19:35,001][starVLA.training.train_starvla][INFO] - Step 1055, Loss: {'train/loss': 1.1186274290084839, 'train/grad_norm_pre_clip': 10.657951354980469, 'timing/data': 0.05992131493985653, 'timing/model': 1.3138748481869698, 'learning_rate/qwen_vl_interface': 9.55800367114803e-06, 'learning_rate/action_model': 9.55800367114803e-05, 'epoch': 0.39}) +[2026-06-17 18:19:47,483][starVLA.training.train_starvla][INFO] - Step 1056, Loss: {'train/loss': 0.8742797374725342, 'train/grad_norm_pre_clip': 10.588970184326172, 'timing/data': 0.06398049183189869, 'timing/model': 1.0907190255820751, 'learning_rate/qwen_vl_interface': 9.557092207908428e-06, 'learning_rate/action_model': 9.557092207908428e-05, 'epoch': 0.39}) +[2026-06-17 18:19:53,905][starVLA.training.train_starvla][INFO] - Step 1057, Loss: {'train/loss': 0.7931081056594849, 'train/grad_norm_pre_clip': 9.02820873260498, 'timing/data': 0.04106385167688131, 'timing/model': 2.727006770670414, 'learning_rate/qwen_vl_interface': 9.556179851803978e-06, 'learning_rate/action_model': 9.556179851803978e-05, 'epoch': 0.39}) +[2026-06-17 18:19:59,039][starVLA.training.train_starvla][INFO] - Step 1058, Loss: {'train/loss': 0.9405195116996765, 'train/grad_norm_pre_clip': 11.31093692779541, 'timing/data': 0.03782945219427347, 'timing/model': 1.1062679151073098, 'learning_rate/qwen_vl_interface': 9.555266603023809e-06, 'learning_rate/action_model': 9.555266603023808e-05, 'epoch': 0.39}) +[2026-06-17 18:20:05,653][starVLA.training.train_starvla][INFO] - Step 1059, Loss: {'train/loss': 0.80361407995224, 'train/grad_norm_pre_clip': 14.515007019042969, 'timing/data': 0.05939867068082094, 'timing/model': 2.8938844576478004, 'learning_rate/qwen_vl_interface': 9.554352461757243e-06, 'learning_rate/action_model': 9.554352461757242e-05, 'epoch': 0.39}) +[2026-06-17 18:20:10,380][starVLA.training.train_starvla][INFO] - Step 1060, Loss: {'train/loss': 1.2025766372680664, 'train/grad_norm_pre_clip': 21.830739974975586, 'timing/data': 0.05646566301584244, 'timing/model': 1.1389310043305159, 'learning_rate/qwen_vl_interface': 9.553437428193778e-06, 'learning_rate/action_model': 9.553437428193778e-05, 'epoch': 0.39}) +[2026-06-17 18:20:20,807][starVLA.training.train_starvla][INFO] - Step 1061, Loss: {'train/loss': 0.8575097918510437, 'train/grad_norm_pre_clip': 13.288294792175293, 'timing/data': 0.11240559443831444, 'timing/model': 2.9872508058324456, 'learning_rate/qwen_vl_interface': 9.552521502523104e-06, 'learning_rate/action_model': 9.552521502523102e-05, 'epoch': 0.39}) +[2026-06-17 18:20:33,205][starVLA.training.train_starvla][INFO] - Step 1062, Loss: {'train/loss': 1.2140976190567017, 'train/grad_norm_pre_clip': 11.002643585205078, 'timing/data': 0.04572938196361065, 'timing/model': 4.7071927934885025, 'learning_rate/qwen_vl_interface': 9.551604684935093e-06, 'learning_rate/action_model': 9.551604684935092e-05, 'epoch': 0.39}) +[2026-06-17 18:20:38,394][starVLA.training.train_starvla][INFO] - Step 1063, Loss: {'train/loss': 0.8714727759361267, 'train/grad_norm_pre_clip': 11.121723175048828, 'timing/data': 0.04386814497411251, 'timing/model': 1.2015584465116262, 'learning_rate/qwen_vl_interface': 9.550686975619803e-06, 'learning_rate/action_model': 9.550686975619802e-05, 'epoch': 0.39}) +[2026-06-17 18:20:43,087][starVLA.training.train_starvla][INFO] - Step 1064, Loss: {'train/loss': 1.0937511920928955, 'train/grad_norm_pre_clip': 15.758537292480469, 'timing/data': 0.05170530918985605, 'timing/model': 1.384039863012731, 'learning_rate/qwen_vl_interface': 9.549768374767473e-06, 'learning_rate/action_model': 9.549768374767473e-05, 'epoch': 0.39}) +[2026-06-17 18:20:47,278][starVLA.training.train_starvla][INFO] - Step 1065, Loss: {'train/loss': 1.0559111833572388, 'train/grad_norm_pre_clip': 11.485640525817871, 'timing/data': 0.039741418324410915, 'timing/model': 1.1080000754445791, 'learning_rate/qwen_vl_interface': 9.548848882568533e-06, 'learning_rate/action_model': 9.548848882568532e-05, 'epoch': 0.39}) +[2026-06-17 18:20:51,903][starVLA.training.train_starvla][INFO] - Step 1066, Loss: {'train/loss': 1.0765557289123535, 'train/grad_norm_pre_clip': 17.66425895690918, 'timing/data': 0.04623216297477484, 'timing/model': 1.2195746218785644, 'learning_rate/qwen_vl_interface': 9.547928499213593e-06, 'learning_rate/action_model': 9.547928499213593e-05, 'epoch': 0.39}) +[2026-06-17 18:20:56,293][starVLA.training.train_starvla][INFO] - Step 1067, Loss: {'train/loss': 1.2663830518722534, 'train/grad_norm_pre_clip': 15.560914993286133, 'timing/data': 0.04720566235482693, 'timing/model': 1.0227857641875744, 'learning_rate/qwen_vl_interface': 9.547007224893452e-06, 'learning_rate/action_model': 9.547007224893452e-05, 'epoch': 0.39}) +[2026-06-17 18:21:00,928][starVLA.training.train_starvla][INFO] - Step 1068, Loss: {'train/loss': 1.0234286785125732, 'train/grad_norm_pre_clip': 12.188887596130371, 'timing/data': 0.04206667095422745, 'timing/model': 1.0489841140806675, 'learning_rate/qwen_vl_interface': 9.546085059799088e-06, 'learning_rate/action_model': 9.546085059799088e-05, 'epoch': 0.39}) +[2026-06-17 18:21:05,410][starVLA.training.train_starvla][INFO] - Step 1069, Loss: {'train/loss': 1.1669479608535767, 'train/grad_norm_pre_clip': 13.21911334991455, 'timing/data': 0.05050198920071125, 'timing/model': 1.3806217815726995, 'learning_rate/qwen_vl_interface': 9.545162004121672e-06, 'learning_rate/action_model': 9.54516200412167e-05, 'epoch': 0.39}) +[2026-06-17 18:21:09,611][starVLA.training.train_starvla][INFO] - Step 1070, Loss: {'train/loss': 1.1041409969329834, 'train/grad_norm_pre_clip': 11.617043495178223, 'timing/data': 0.03908785991370678, 'timing/model': 1.1323862662538886, 'learning_rate/qwen_vl_interface': 9.544238058052546e-06, 'learning_rate/action_model': 9.544238058052545e-05, 'epoch': 0.39}) +[2026-06-17 18:21:14,057][starVLA.training.train_starvla][INFO] - Step 1071, Loss: {'train/loss': 0.8956157565116882, 'train/grad_norm_pre_clip': 10.585119247436523, 'timing/data': 0.04364695120602846, 'timing/model': 1.0348169123753905, 'learning_rate/qwen_vl_interface': 9.54331322178325e-06, 'learning_rate/action_model': 9.543313221783251e-05, 'epoch': 0.39}) +[2026-06-17 18:21:18,417][starVLA.training.train_starvla][INFO] - Step 1072, Loss: {'train/loss': 1.054460048675537, 'train/grad_norm_pre_clip': 11.639994621276855, 'timing/data': 0.03896793071180582, 'timing/model': 1.0516530852764845, 'learning_rate/qwen_vl_interface': 9.542387495505505e-06, 'learning_rate/action_model': 9.542387495505506e-05, 'epoch': 0.39}) +[2026-06-17 18:21:22,777][starVLA.training.train_starvla][INFO] - Step 1073, Loss: {'train/loss': 1.0607683658599854, 'train/grad_norm_pre_clip': 10.420818328857422, 'timing/data': 0.037133656442165375, 'timing/model': 1.0216541085392237, 'learning_rate/qwen_vl_interface': 9.541460879411214e-06, 'learning_rate/action_model': 9.541460879411215e-05, 'epoch': 0.39}) +[2026-06-17 18:21:27,037][starVLA.training.train_starvla][INFO] - Step 1074, Loss: {'train/loss': 1.1005735397338867, 'train/grad_norm_pre_clip': 10.91850757598877, 'timing/data': 0.040827710181474686, 'timing/model': 1.3072256380692124, 'learning_rate/qwen_vl_interface': 9.540533373692466e-06, 'learning_rate/action_model': 9.540533373692466e-05, 'epoch': 0.39}) +[2026-06-17 18:21:31,077][starVLA.training.train_starvla][INFO] - Step 1075, Loss: {'train/loss': 1.158779263496399, 'train/grad_norm_pre_clip': 15.277965545654297, 'timing/data': 0.03773710876703262, 'timing/model': 1.0798400277271867, 'learning_rate/qwen_vl_interface': 9.539604978541531e-06, 'learning_rate/action_model': 9.539604978541531e-05, 'epoch': 0.39}) +[2026-06-17 18:21:35,489][starVLA.training.train_starvla][INFO] - Step 1076, Loss: {'train/loss': 0.9025579690933228, 'train/grad_norm_pre_clip': 14.920943260192871, 'timing/data': 0.04442686028778553, 'timing/model': 1.0870683444663882, 'learning_rate/qwen_vl_interface': 9.538675694150868e-06, 'learning_rate/action_model': 9.538675694150868e-05, 'epoch': 0.39}) +[2026-06-17 18:21:40,084][starVLA.training.train_starvla][INFO] - Step 1077, Loss: {'train/loss': 0.8427988886833191, 'train/grad_norm_pre_clip': 10.051285743713379, 'timing/data': 0.041039781644940376, 'timing/model': 1.089217383414507, 'learning_rate/qwen_vl_interface': 9.53774552071312e-06, 'learning_rate/action_model': 9.537745520713119e-05, 'epoch': 0.39}) +[2026-06-17 18:21:44,902][starVLA.training.train_starvla][INFO] - Step 1078, Loss: {'train/loss': 0.9384318590164185, 'train/grad_norm_pre_clip': 10.676924705505371, 'timing/data': 0.05164094548672438, 'timing/model': 1.0671345042064786, 'learning_rate/qwen_vl_interface': 9.53681445842111e-06, 'learning_rate/action_model': 9.53681445842111e-05, 'epoch': 0.39}) +[2026-06-17 18:21:49,417][starVLA.training.train_starvla][INFO] - Step 1079, Loss: {'train/loss': 0.8247570395469666, 'train/grad_norm_pre_clip': 9.195972442626953, 'timing/data': 0.03922777529805899, 'timing/model': 1.3860591892153025, 'learning_rate/qwen_vl_interface': 9.53588250746785e-06, 'learning_rate/action_model': 9.53588250746785e-05, 'epoch': 0.4}) +[2026-06-17 18:22:00,229][starVLA.training.train_starvla][INFO] - Step 1080, Loss: {'train/loss': 1.0828875303268433, 'train/grad_norm_pre_clip': 9.859399795532227, 'timing/data': 0.046459619887173176, 'timing/model': 1.1731008226051927, 'learning_rate/qwen_vl_interface': 9.534949668046537e-06, 'learning_rate/action_model': 9.534949668046536e-05, 'epoch': 0.4}) +[2026-06-17 18:22:08,507][starVLA.training.train_starvla][INFO] - Step 1081, Loss: {'train/loss': 0.7232193350791931, 'train/grad_norm_pre_clip': 8.286986351013184, 'timing/data': 0.04533530864864588, 'timing/model': 4.407801633700728, 'learning_rate/qwen_vl_interface': 9.534015940350545e-06, 'learning_rate/action_model': 9.534015940350544e-05, 'epoch': 0.4}) +[2026-06-17 18:22:23,691][starVLA.training.train_starvla][INFO] - Step 1082, Loss: {'train/loss': 0.9914410710334778, 'train/grad_norm_pre_clip': 12.520739555358887, 'timing/data': 0.05297346878796816, 'timing/model': 5.516827594488859, 'learning_rate/qwen_vl_interface': 9.533081324573437e-06, 'learning_rate/action_model': 9.533081324573437e-05, 'epoch': 0.4}) +[2026-06-17 18:22:33,049][starVLA.training.train_starvla][INFO] - Step 1083, Loss: {'train/loss': 0.8862811923027039, 'train/grad_norm_pre_clip': 11.915770530700684, 'timing/data': 0.06509094312787056, 'timing/model': 2.827019221149385, 'learning_rate/qwen_vl_interface': 9.532145820908965e-06, 'learning_rate/action_model': 9.532145820908964e-05, 'epoch': 0.4}) +[2026-06-17 18:22:41,039][starVLA.training.train_starvla][INFO] - Step 1084, Loss: {'train/loss': 1.1235601902008057, 'train/grad_norm_pre_clip': 12.362839698791504, 'timing/data': 0.09050209447741508, 'timing/model': 3.5174676906317472, 'learning_rate/qwen_vl_interface': 9.531209429551056e-06, 'learning_rate/action_model': 9.531209429551054e-05, 'epoch': 0.4}) +[2026-06-17 18:22:53,120][starVLA.training.train_starvla][INFO] - Step 1085, Loss: {'train/loss': 0.5041443109512329, 'train/grad_norm_pre_clip': 5.926629066467285, 'timing/data': 0.04301758576184511, 'timing/model': 5.131718397140503, 'learning_rate/qwen_vl_interface': 9.530272150693823e-06, 'learning_rate/action_model': 9.530272150693824e-05, 'epoch': 0.4}) +[2026-06-17 18:23:00,075][starVLA.training.train_starvla][INFO] - Step 1086, Loss: {'train/loss': 0.6013709902763367, 'train/grad_norm_pre_clip': 6.905246734619141, 'timing/data': 0.04681982286274433, 'timing/model': 1.2550445748493075, 'learning_rate/qwen_vl_interface': 9.52933398453157e-06, 'learning_rate/action_model': 9.52933398453157e-05, 'epoch': 0.4}) +[2026-06-17 18:23:09,578][starVLA.training.train_starvla][INFO] - Step 1087, Loss: {'train/loss': 0.7437606453895569, 'train/grad_norm_pre_clip': 7.983479022979736, 'timing/data': 0.04558504745364189, 'timing/model': 1.1274851998314261, 'learning_rate/qwen_vl_interface': 9.528394931258777e-06, 'learning_rate/action_model': 9.528394931258776e-05, 'epoch': 0.4}) +[2026-06-17 18:23:14,056][starVLA.training.train_starvla][INFO] - Step 1088, Loss: {'train/loss': 0.8064404129981995, 'train/grad_norm_pre_clip': 8.206828117370605, 'timing/data': 0.05038599297404289, 'timing/model': 1.0403998391702771, 'learning_rate/qwen_vl_interface': 9.527454991070108e-06, 'learning_rate/action_model': 9.527454991070108e-05, 'epoch': 0.4}) +[2026-06-17 18:23:18,598][starVLA.training.train_starvla][INFO] - Step 1089, Loss: {'train/loss': 0.6615638732910156, 'train/grad_norm_pre_clip': 15.261545181274414, 'timing/data': 0.056362904608249664, 'timing/model': 1.5199370617046952, 'learning_rate/qwen_vl_interface': 9.52651416416042e-06, 'learning_rate/action_model': 9.526514164160421e-05, 'epoch': 0.4}) +[2026-06-17 18:23:22,731][starVLA.training.train_starvla][INFO] - Step 1090, Loss: {'train/loss': 0.9727381467819214, 'train/grad_norm_pre_clip': 9.794966697692871, 'timing/data': 0.0439594155177474, 'timing/model': 1.0788802420720458, 'learning_rate/qwen_vl_interface': 9.525572450724743e-06, 'learning_rate/action_model': 9.525572450724743e-05, 'epoch': 0.4}) +[2026-06-17 18:23:27,321][starVLA.training.train_starvla][INFO] - Step 1091, Loss: {'train/loss': 0.936062216758728, 'train/grad_norm_pre_clip': 12.81898021697998, 'timing/data': 0.052936640568077564, 'timing/model': 1.1358806816861033, 'learning_rate/qwen_vl_interface': 9.524629850958297e-06, 'learning_rate/action_model': 9.524629850958297e-05, 'epoch': 0.4}) +[2026-06-17 18:23:31,694][starVLA.training.train_starvla][INFO] - Step 1092, Loss: {'train/loss': 0.950358510017395, 'train/grad_norm_pre_clip': 10.236984252929688, 'timing/data': 0.05275475513190031, 'timing/model': 1.0675039598718286, 'learning_rate/qwen_vl_interface': 9.523686365056482e-06, 'learning_rate/action_model': 9.523686365056481e-05, 'epoch': 0.4}) +[2026-06-17 18:23:36,157][starVLA.training.train_starvla][INFO] - Step 1093, Loss: {'train/loss': 0.8676921725273132, 'train/grad_norm_pre_clip': 9.385440826416016, 'timing/data': 0.04562380723655224, 'timing/model': 1.0334248868748546, 'learning_rate/qwen_vl_interface': 9.522741993214886e-06, 'learning_rate/action_model': 9.522741993214886e-05, 'epoch': 0.4}) +[2026-06-17 18:23:40,630][starVLA.training.train_starvla][INFO] - Step 1094, Loss: {'train/loss': 0.7548169493675232, 'train/grad_norm_pre_clip': 18.70599365234375, 'timing/data': 0.036608717404305935, 'timing/model': 1.3453350309282541, 'learning_rate/qwen_vl_interface': 9.521796735629275e-06, 'learning_rate/action_model': 9.521796735629275e-05, 'epoch': 0.4}) +[2026-06-17 18:23:44,674][starVLA.training.train_starvla][INFO] - Step 1095, Loss: {'train/loss': 0.9907962679862976, 'train/grad_norm_pre_clip': 11.423991203308105, 'timing/data': 0.038456627167761326, 'timing/model': 1.090433944016695, 'learning_rate/qwen_vl_interface': 9.520850592495603e-06, 'learning_rate/action_model': 9.520850592495604e-05, 'epoch': 0.4}) +[2026-06-17 18:23:49,025][starVLA.training.train_starvla][INFO] - Step 1096, Loss: {'train/loss': 0.9936355352401733, 'train/grad_norm_pre_clip': 11.613801956176758, 'timing/data': 0.05381085351109505, 'timing/model': 1.0371228326112032, 'learning_rate/qwen_vl_interface': 9.519903564010012e-06, 'learning_rate/action_model': 9.51990356401001e-05, 'epoch': 0.4}) +[2026-06-17 18:23:53,395][starVLA.training.train_starvla][INFO] - Step 1097, Loss: {'train/loss': 0.9866945147514343, 'train/grad_norm_pre_clip': 12.656781196594238, 'timing/data': 0.03985547088086605, 'timing/model': 1.1113082244992256, 'learning_rate/qwen_vl_interface': 9.518955650368816e-06, 'learning_rate/action_model': 9.518955650368814e-05, 'epoch': 0.4}) +[2026-06-17 18:23:58,600][starVLA.training.train_starvla][INFO] - Step 1098, Loss: {'train/loss': 0.7909483909606934, 'train/grad_norm_pre_clip': 8.80527114868164, 'timing/data': 0.046560514718294144, 'timing/model': 1.790668341331184, 'learning_rate/qwen_vl_interface': 9.51800685176852e-06, 'learning_rate/action_model': 9.518006851768518e-05, 'epoch': 0.4}) +[2026-06-17 18:24:05,624][starVLA.training.train_starvla][INFO] - Step 1099, Loss: {'train/loss': 0.940143346786499, 'train/grad_norm_pre_clip': 10.876364707946777, 'timing/data': 0.04565168358385563, 'timing/model': 1.438988815061748, 'learning_rate/qwen_vl_interface': 9.51705716840581e-06, 'learning_rate/action_model': 9.517057168405808e-05, 'epoch': 0.4}) +[2026-06-17 18:27:45,259][starVLA.training.train_starvla][INFO] - Step 1100, Loss: {'train/loss': 0.9687687158584595, 'train/grad_norm_pre_clip': 9.568918228149414, 'eval/loss': 0.9164586067199707, 'eval/latency_2/loss': 0.9164586067199707, 'eval/demon_fire_precision': 0.8580669710806698, 'eval/demon_fire_recall': 0.9435146443514645, 'eval/demon_fire_f1': 0.898764447987246, 'eval/demon_fire_rate_model': 0.7312186978297162, 'eval/demon_fire_rate_teacher': 0.6649972175848636, 'eval/demon_left_precision': 0.7906976744186046, 'eval/demon_left_recall': 0.68, 'eval/demon_left_f1': 0.7311827956989247, 'eval/demon_left_rate_model': 0.2273233166388425, 'eval/demon_left_rate_teacher': 0.2643294379521425, 'eval/demon_right_precision': 0.7874331550802139, 'eval/demon_right_recall': 0.6655367231638418, 'eval/demon_right_f1': 0.7213717085119413, 'eval/demon_right_rate_model': 0.20812465219810797, 'eval/demon_right_rate_teacher': 0.24624373956594323, 'eval/demon_move_f1': 0.726277252105433, 'eval/demon_cc_f1': 0.8389878305131377, 'eval/latency_2/demon_fire_precision': 0.8580669710806698, 'eval/latency_2/demon_fire_recall': 0.9435146443514645, 'eval/latency_2/demon_fire_f1': 0.898764447987246, 'eval/latency_2/demon_fire_rate_model': 0.7312186978297162, 'eval/latency_2/demon_fire_rate_teacher': 0.6649972175848636, 'eval/latency_2/demon_left_precision': 0.7906976744186046, 'eval/latency_2/demon_left_recall': 0.68, 'eval/latency_2/demon_left_f1': 0.7311827956989247, 'eval/latency_2/demon_left_rate_model': 0.2273233166388425, 'eval/latency_2/demon_left_rate_teacher': 0.2643294379521425, 'eval/latency_2/demon_right_precision': 0.7874331550802139, 'eval/latency_2/demon_right_recall': 0.6655367231638418, 'eval/latency_2/demon_right_f1': 0.7213717085119413, 'eval/latency_2/demon_right_rate_model': 0.20812465219810797, 'eval/latency_2/demon_right_rate_teacher': 0.24624373956594323, 'eval/latency_2/demon_move_f1': 0.726277252105433, 'eval/latency_2/demon_cc_f1': 0.8389878305131377, 'timing/data': 0.04299548827111721, 'timing/model': 1.025068635120988, 'learning_rate/qwen_vl_interface': 9.516106600477557e-06, 'learning_rate/action_model': 9.516106600477557e-05, 'epoch': 0.4}) +[2026-06-17 18:27:45,276][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_1100_state +[2026-06-17 18:27:45,285][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 18:27:54,615][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_1100_state/model.safetensors +[2026-06-17 18:28:09,516][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_1100_state/optimizer.bin +[2026-06-17 18:28:09,517][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_1100_state/sampler.bin +[2026-06-17 18:28:09,518][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_1100_state/sampler_1.bin +[2026-06-17 18:28:09,520][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_1100_state/random_states_0.pkl +[2026-06-17 18:28:09,525][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1100 +[2026-06-17 18:28:09,525][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 18:28:09,541][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 18:28:15,474][starVLA.training.train_starvla][INFO] - Step 1101, Loss: {'train/loss': 0.829179048538208, 'train/grad_norm_pre_clip': 9.974331855773926, 'timing/data': 0.05002064164727926, 'timing/model': 1.036392985843122, 'learning_rate/qwen_vl_interface': 9.515155148180816e-06, 'learning_rate/action_model': 9.515155148180816e-05, 'epoch': 0.4}) +[2026-06-17 18:28:20,116][starVLA.training.train_starvla][INFO] - Step 1102, Loss: {'train/loss': 1.1732516288757324, 'train/grad_norm_pre_clip': 12.933764457702637, 'timing/data': 0.04509477596729994, 'timing/model': 1.0323982844129205, 'learning_rate/qwen_vl_interface': 9.514202811712825e-06, 'learning_rate/action_model': 9.514202811712825e-05, 'epoch': 0.4}) +[2026-06-17 18:28:29,331][starVLA.training.train_starvla][INFO] - Step 1103, Loss: {'train/loss': 1.1195118427276611, 'train/grad_norm_pre_clip': 16.648460388183594, 'timing/data': 0.04135285783559084, 'timing/model': 5.619615209288895, 'learning_rate/qwen_vl_interface': 9.513249591271e-06, 'learning_rate/action_model': 9.513249591271e-05, 'epoch': 0.4}) +[2026-06-17 18:28:37,070][starVLA.training.train_starvla][INFO] - Step 1104, Loss: {'train/loss': 0.9258173108100891, 'train/grad_norm_pre_clip': 10.86066722869873, 'timing/data': 0.03254545573145151, 'timing/model': 1.0308089088648558, 'learning_rate/qwen_vl_interface': 9.512295487052949e-06, 'learning_rate/action_model': 9.512295487052949e-05, 'epoch': 0.4}) +[2026-06-17 18:28:41,362][starVLA.training.train_starvla][INFO] - Step 1105, Loss: {'train/loss': 0.9294167757034302, 'train/grad_norm_pre_clip': 11.369150161743164, 'timing/data': 0.038907986134290695, 'timing/model': 1.0258106729015708, 'learning_rate/qwen_vl_interface': 9.511340499256455e-06, 'learning_rate/action_model': 9.511340499256455e-05, 'epoch': 0.4}) +[2026-06-17 18:28:47,578][starVLA.training.train_starvla][INFO] - Step 1106, Loss: {'train/loss': 0.8521751761436462, 'train/grad_norm_pre_clip': 8.421906471252441, 'timing/data': 0.029221071861684322, 'timing/model': 1.0231527546420693, 'learning_rate/qwen_vl_interface': 9.510384628079493e-06, 'learning_rate/action_model': 9.510384628079493e-05, 'epoch': 0.4}) +[2026-06-17 18:28:51,917][starVLA.training.train_starvla][INFO] - Step 1107, Loss: {'train/loss': 0.6686413288116455, 'train/grad_norm_pre_clip': 8.985587120056152, 'timing/data': 0.02787258755415678, 'timing/model': 1.0095595084130764, 'learning_rate/qwen_vl_interface': 9.50942787372021e-06, 'learning_rate/action_model': 9.50942787372021e-05, 'epoch': 0.41}) +[2026-06-17 18:28:55,967][starVLA.training.train_starvla][INFO] - Step 1108, Loss: {'train/loss': 0.7164382934570312, 'train/grad_norm_pre_clip': 9.260941505432129, 'timing/data': 0.02661775704473257, 'timing/model': 0.9970638426020741, 'learning_rate/qwen_vl_interface': 9.508470236376948e-06, 'learning_rate/action_model': 9.508470236376948e-05, 'epoch': 0.41}) +[2026-06-17 18:29:00,113][starVLA.training.train_starvla][INFO] - Step 1109, Loss: {'train/loss': 0.5851055383682251, 'train/grad_norm_pre_clip': 7.556893348693848, 'timing/data': 0.030887849628925323, 'timing/model': 1.2456438187509775, 'learning_rate/qwen_vl_interface': 9.507511716248224e-06, 'learning_rate/action_model': 9.507511716248224e-05, 'epoch': 0.41}) +[2026-06-17 18:29:04,153][starVLA.training.train_starvla][INFO] - Step 1110, Loss: {'train/loss': 0.98261958360672, 'train/grad_norm_pre_clip': 16.998390197753906, 'timing/data': 0.025808069854974747, 'timing/model': 1.178006541915238, 'learning_rate/qwen_vl_interface': 9.506552313532739e-06, 'learning_rate/action_model': 9.506552313532739e-05, 'epoch': 0.41}) +[2026-06-17 18:29:08,151][starVLA.training.train_starvla][INFO] - Step 1111, Loss: {'train/loss': 0.9445818066596985, 'train/grad_norm_pre_clip': 11.12675666809082, 'timing/data': 0.0294107748195529, 'timing/model': 1.024953274987638, 'learning_rate/qwen_vl_interface': 9.505592028429377e-06, 'learning_rate/action_model': 9.505592028429377e-05, 'epoch': 0.41}) +[2026-06-17 18:29:12,506][starVLA.training.train_starvla][INFO] - Step 1112, Loss: {'train/loss': 0.793070375919342, 'train/grad_norm_pre_clip': 11.989752769470215, 'timing/data': 0.025439178571105003, 'timing/model': 1.0037456899881363, 'learning_rate/qwen_vl_interface': 9.504630861137212e-06, 'learning_rate/action_model': 9.504630861137212e-05, 'epoch': 0.41}) +[2026-06-17 18:29:16,654][starVLA.training.train_starvla][INFO] - Step 1113, Loss: {'train/loss': 0.652127206325531, 'train/grad_norm_pre_clip': 11.153420448303223, 'timing/data': 0.03463076427578926, 'timing/model': 1.0067085092887282, 'learning_rate/qwen_vl_interface': 9.503668811855489e-06, 'learning_rate/action_model': 9.503668811855488e-05, 'epoch': 0.41}) +[2026-06-17 18:29:20,797][starVLA.training.train_starvla][INFO] - Step 1114, Loss: {'train/loss': 0.8042859435081482, 'train/grad_norm_pre_clip': 13.822744369506836, 'timing/data': 0.027787030674517155, 'timing/model': 1.0146455308422446, 'learning_rate/qwen_vl_interface': 9.502705880783644e-06, 'learning_rate/action_model': 9.502705880783645e-05, 'epoch': 0.41}) +[2026-06-17 18:29:24,893][starVLA.training.train_starvla][INFO] - Step 1115, Loss: {'train/loss': 0.7502663135528564, 'train/grad_norm_pre_clip': 12.247648239135742, 'timing/data': 0.02708328701555729, 'timing/model': 1.2071078782901168, 'learning_rate/qwen_vl_interface': 9.501742068121296e-06, 'learning_rate/action_model': 9.501742068121296e-05, 'epoch': 0.41}) +[2026-06-17 18:29:28,835][starVLA.training.train_starvla][INFO] - Step 1116, Loss: {'train/loss': 0.7930204272270203, 'train/grad_norm_pre_clip': 16.59238052368164, 'timing/data': 0.028063694015145302, 'timing/model': 1.006010846234858, 'learning_rate/qwen_vl_interface': 9.500777374068241e-06, 'learning_rate/action_model': 9.50077737406824e-05, 'epoch': 0.41}) +[2026-06-17 18:29:32,928][starVLA.training.train_starvla][INFO] - Step 1117, Loss: {'train/loss': 0.6497328877449036, 'train/grad_norm_pre_clip': 12.257843971252441, 'timing/data': 0.027445019222795963, 'timing/model': 1.028787748888135, 'learning_rate/qwen_vl_interface': 9.49981179882446e-06, 'learning_rate/action_model': 9.499811798824461e-05, 'epoch': 0.41}) +[2026-06-17 18:29:38,010][starVLA.training.train_starvla][INFO] - Step 1118, Loss: {'train/loss': 0.7249748110771179, 'train/grad_norm_pre_clip': 12.927314758300781, 'timing/data': 0.026710488833487034, 'timing/model': 0.9998669093474746, 'learning_rate/qwen_vl_interface': 9.498845342590124e-06, 'learning_rate/action_model': 9.498845342590124e-05, 'epoch': 0.41}) +[2026-06-17 18:29:42,161][starVLA.training.train_starvla][INFO] - Step 1119, Loss: {'train/loss': 1.1305186748504639, 'train/grad_norm_pre_clip': 14.29682445526123, 'timing/data': 0.026312440633773804, 'timing/model': 0.9990487769246101, 'learning_rate/qwen_vl_interface': 9.497878005565575e-06, 'learning_rate/action_model': 9.497878005565574e-05, 'epoch': 0.41}) +[2026-06-17 18:29:50,212][starVLA.training.train_starvla][INFO] - Step 1120, Loss: {'train/loss': 0.8422845602035522, 'train/grad_norm_pre_clip': 11.474625587463379, 'timing/data': 0.03200453706085682, 'timing/model': 1.480845594778657, 'learning_rate/qwen_vl_interface': 9.496909787951344e-06, 'learning_rate/action_model': 9.496909787951343e-05, 'epoch': 0.41}) +[2026-06-17 18:29:54,461][starVLA.training.train_starvla][INFO] - Step 1121, Loss: {'train/loss': 1.0196224451065063, 'train/grad_norm_pre_clip': 13.489996910095215, 'timing/data': 0.03471003472805023, 'timing/model': 1.0924483835697174, 'learning_rate/qwen_vl_interface': 9.495940689948144e-06, 'learning_rate/action_model': 9.495940689948144e-05, 'epoch': 0.41}) +[2026-06-17 18:29:59,189][starVLA.training.train_starvla][INFO] - Step 1122, Loss: {'train/loss': 0.770870566368103, 'train/grad_norm_pre_clip': 11.800889015197754, 'timing/data': 0.03053854499012232, 'timing/model': 1.1531409248709679, 'learning_rate/qwen_vl_interface': 9.49497071175687e-06, 'learning_rate/action_model': 9.49497071175687e-05, 'epoch': 0.41}) +[2026-06-17 18:30:06,114][starVLA.training.train_starvla][INFO] - Step 1123, Loss: {'train/loss': 0.662865936756134, 'train/grad_norm_pre_clip': 16.882226943969727, 'timing/data': 0.03487939853221178, 'timing/model': 1.0294396821409464, 'learning_rate/qwen_vl_interface': 9.493999853578601e-06, 'learning_rate/action_model': 9.493999853578602e-05, 'epoch': 0.41}) +[2026-06-17 18:30:10,715][starVLA.training.train_starvla][INFO] - Step 1124, Loss: {'train/loss': 0.8190128207206726, 'train/grad_norm_pre_clip': 10.398070335388184, 'timing/data': 0.033746689558029175, 'timing/model': 1.1152102760970592, 'learning_rate/qwen_vl_interface': 9.493028115614595e-06, 'learning_rate/action_model': 9.493028115614595e-05, 'epoch': 0.41}) +[2026-06-17 18:30:15,336][starVLA.training.train_starvla][INFO] - Step 1125, Loss: {'train/loss': 0.7917624115943909, 'train/grad_norm_pre_clip': 13.9341459274292, 'timing/data': 0.03236284293234348, 'timing/model': 1.4509325502440333, 'learning_rate/qwen_vl_interface': 9.492055498066295e-06, 'learning_rate/action_model': 9.492055498066294e-05, 'epoch': 0.41}) +[2026-06-17 18:30:19,765][starVLA.training.train_starvla][INFO] - Step 1126, Loss: {'train/loss': 0.8625532388687134, 'train/grad_norm_pre_clip': 10.82672119140625, 'timing/data': 0.03126058727502823, 'timing/model': 1.0654337331652641, 'learning_rate/qwen_vl_interface': 9.491082001135326e-06, 'learning_rate/action_model': 9.491082001135325e-05, 'epoch': 0.41}) +[2026-06-17 18:30:24,256][starVLA.training.train_starvla][INFO] - Step 1127, Loss: {'train/loss': 0.8227024674415588, 'train/grad_norm_pre_clip': 11.439752578735352, 'timing/data': 0.03189373668283224, 'timing/model': 1.0636236546561122, 'learning_rate/qwen_vl_interface': 9.490107625023492e-06, 'learning_rate/action_model': 9.490107625023492e-05, 'epoch': 0.41}) +[2026-06-17 18:30:30,728][starVLA.training.train_starvla][INFO] - Step 1128, Loss: {'train/loss': 0.7120510339736938, 'train/grad_norm_pre_clip': 8.91397762298584, 'timing/data': 0.03442181460559368, 'timing/model': 1.0794620085507631, 'learning_rate/qwen_vl_interface': 9.489132369932788e-06, 'learning_rate/action_model': 9.489132369932788e-05, 'epoch': 0.41}) +[2026-06-17 18:30:35,396][starVLA.training.train_starvla][INFO] - Step 1129, Loss: {'train/loss': 0.6743304133415222, 'train/grad_norm_pre_clip': 14.204151153564453, 'timing/data': 0.03860718663781881, 'timing/model': 1.1625931151211262, 'learning_rate/qwen_vl_interface': 9.488156236065378e-06, 'learning_rate/action_model': 9.488156236065379e-05, 'epoch': 0.41}) +[2026-06-17 18:30:40,200][starVLA.training.train_starvla][INFO] - Step 1130, Loss: {'train/loss': 0.8668181300163269, 'train/grad_norm_pre_clip': 9.743617057800293, 'timing/data': 0.03718963824212551, 'timing/model': 1.4358726609498262, 'learning_rate/qwen_vl_interface': 9.487179223623626e-06, 'learning_rate/action_model': 9.487179223623624e-05, 'epoch': 0.41}) +[2026-06-17 18:30:44,443][starVLA.training.train_starvla][INFO] - Step 1131, Loss: {'train/loss': 0.7290607690811157, 'train/grad_norm_pre_clip': 9.898545265197754, 'timing/data': 0.03221971541643143, 'timing/model': 1.0974840391427279, 'learning_rate/qwen_vl_interface': 9.486201332810058e-06, 'learning_rate/action_model': 9.486201332810057e-05, 'epoch': 0.41}) +[2026-06-17 18:30:49,284][starVLA.training.train_starvla][INFO] - Step 1132, Loss: {'train/loss': 0.9016034007072449, 'train/grad_norm_pre_clip': 12.167519569396973, 'timing/data': 0.03360749315470457, 'timing/model': 1.212407436221838, 'learning_rate/qwen_vl_interface': 9.485222563827395e-06, 'learning_rate/action_model': 9.485222563827395e-05, 'epoch': 0.41}) +[2026-06-17 18:30:57,680][starVLA.training.train_starvla][INFO] - Step 1133, Loss: {'train/loss': 0.7764526009559631, 'train/grad_norm_pre_clip': 9.916696548461914, 'timing/data': 0.04016860481351614, 'timing/model': 1.0955830002203584, 'learning_rate/qwen_vl_interface': 9.484242916878537e-06, 'learning_rate/action_model': 9.484242916878536e-05, 'epoch': 0.41}) +[2026-06-17 18:31:02,807][starVLA.training.train_starvla][INFO] - Step 1134, Loss: {'train/loss': 0.825651228427887, 'train/grad_norm_pre_clip': 12.235694885253906, 'timing/data': 0.03561058733612299, 'timing/model': 1.2219478720799088, 'learning_rate/qwen_vl_interface': 9.483262392166567e-06, 'learning_rate/action_model': 9.483262392166566e-05, 'epoch': 0.42}) +[2026-06-17 18:31:07,561][starVLA.training.train_starvla][INFO] - Step 1135, Loss: {'train/loss': 1.029301404953003, 'train/grad_norm_pre_clip': 13.498821258544922, 'timing/data': 0.03779760468751192, 'timing/model': 1.4478816855698824, 'learning_rate/qwen_vl_interface': 9.482280989894748e-06, 'learning_rate/action_model': 9.482280989894748e-05, 'epoch': 0.42}) +[2026-06-17 18:31:13,730][starVLA.training.train_starvla][INFO] - Step 1136, Loss: {'train/loss': 1.0778952836990356, 'train/grad_norm_pre_clip': 14.524222373962402, 'timing/data': 0.0844624862074852, 'timing/model': 2.5980672435835004, 'learning_rate/qwen_vl_interface': 9.481298710266525e-06, 'learning_rate/action_model': 9.481298710266526e-05, 'epoch': 0.42}) +[2026-06-17 18:31:26,570][starVLA.training.train_starvla][INFO] - Step 1137, Loss: {'train/loss': 0.8285880088806152, 'train/grad_norm_pre_clip': 13.100943565368652, 'timing/data': 0.04231424443423748, 'timing/model': 1.3624151851981878, 'learning_rate/qwen_vl_interface': 9.480315553485527e-06, 'learning_rate/action_model': 9.480315553485527e-05, 'epoch': 0.42}) +[2026-06-17 18:31:33,929][starVLA.training.train_starvla][INFO] - Step 1138, Loss: {'train/loss': 0.8841521143913269, 'train/grad_norm_pre_clip': 13.449224472045898, 'timing/data': 0.03737882059067488, 'timing/model': 1.4589588204398751, 'learning_rate/qwen_vl_interface': 9.479331519755563e-06, 'learning_rate/action_model': 9.479331519755562e-05, 'epoch': 0.42}) +[2026-06-17 18:31:38,900][starVLA.training.train_starvla][INFO] - Step 1139, Loss: {'train/loss': 0.6017556190490723, 'train/grad_norm_pre_clip': 9.37901782989502, 'timing/data': 0.04638290777802467, 'timing/model': 1.7565842978656292, 'learning_rate/qwen_vl_interface': 9.478346609280622e-06, 'learning_rate/action_model': 9.478346609280623e-05, 'epoch': 0.42}) +[2026-06-17 18:31:50,205][starVLA.training.train_starvla][INFO] - Step 1140, Loss: {'train/loss': 0.7987974882125854, 'train/grad_norm_pre_clip': 9.003450393676758, 'timing/data': 0.04469637479633093, 'timing/model': 1.3387101339176297, 'learning_rate/qwen_vl_interface': 9.477360822264883e-06, 'learning_rate/action_model': 9.477360822264882e-05, 'epoch': 0.42}) +[2026-06-17 18:31:54,805][starVLA.training.train_starvla][INFO] - Step 1141, Loss: {'train/loss': 0.8660200238227844, 'train/grad_norm_pre_clip': 10.635963439941406, 'timing/data': 0.045235306955873966, 'timing/model': 1.0928116301074624, 'learning_rate/qwen_vl_interface': 9.476374158912692e-06, 'learning_rate/action_model': 9.476374158912691e-05, 'epoch': 0.42}) +[2026-06-17 18:32:04,801][starVLA.training.train_starvla][INFO] - Step 1142, Loss: {'train/loss': 0.7562849521636963, 'train/grad_norm_pre_clip': 13.791186332702637, 'timing/data': 0.036556653678417206, 'timing/model': 2.3163969423621893, 'learning_rate/qwen_vl_interface': 9.475386619428593e-06, 'learning_rate/action_model': 9.475386619428594e-05, 'epoch': 0.42}) +[2026-06-17 18:32:09,904][starVLA.training.train_starvla][INFO] - Step 1143, Loss: {'train/loss': 0.5984029769897461, 'train/grad_norm_pre_clip': 8.002846717834473, 'timing/data': 0.03673850558698177, 'timing/model': 1.3259841799736023, 'learning_rate/qwen_vl_interface': 9.4743982040173e-06, 'learning_rate/action_model': 9.4743982040173e-05, 'epoch': 0.42}) +[2026-06-17 18:32:14,628][starVLA.training.train_starvla][INFO] - Step 1144, Loss: {'train/loss': 1.2348321676254272, 'train/grad_norm_pre_clip': 17.502483367919922, 'timing/data': 0.03152161743491888, 'timing/model': 1.5924229891970754, 'learning_rate/qwen_vl_interface': 9.473408912883714e-06, 'learning_rate/action_model': 9.473408912883714e-05, 'epoch': 0.42}) +[2026-06-17 18:32:21,799][starVLA.training.train_starvla][INFO] - Step 1145, Loss: {'train/loss': 0.9443081617355347, 'train/grad_norm_pre_clip': 12.350251197814941, 'timing/data': 0.0333041762933135, 'timing/model': 1.3517083348706365, 'learning_rate/qwen_vl_interface': 9.472418746232916e-06, 'learning_rate/action_model': 9.472418746232916e-05, 'epoch': 0.42}) +[2026-06-17 18:32:26,066][starVLA.training.train_starvla][INFO] - Step 1146, Loss: {'train/loss': 1.0917139053344727, 'train/grad_norm_pre_clip': 12.3367338180542, 'timing/data': 0.033849477767944336, 'timing/model': 1.1289583202451468, 'learning_rate/qwen_vl_interface': 9.471427704270168e-06, 'learning_rate/action_model': 9.471427704270169e-05, 'epoch': 0.42}) +[2026-06-17 18:32:33,346][starVLA.training.train_starvla][INFO] - Step 1147, Loss: {'train/loss': 1.3476754426956177, 'train/grad_norm_pre_clip': 17.828231811523438, 'timing/data': 0.035011288709938526, 'timing/model': 1.0755923707038164, 'learning_rate/qwen_vl_interface': 9.470435787200914e-06, 'learning_rate/action_model': 9.470435787200915e-05, 'epoch': 0.42}) +[2026-06-17 18:32:40,704][starVLA.training.train_starvla][INFO] - Step 1148, Loss: {'train/loss': 0.7633754014968872, 'train/grad_norm_pre_clip': 10.92077922821045, 'timing/data': 0.032305190339684486, 'timing/model': 3.713378544896841, 'learning_rate/qwen_vl_interface': 9.469442995230781e-06, 'learning_rate/action_model': 9.469442995230782e-05, 'epoch': 0.42}) +[2026-06-17 18:32:45,222][starVLA.training.train_starvla][INFO] - Step 1149, Loss: {'train/loss': 0.9280017018318176, 'train/grad_norm_pre_clip': 14.028518676757812, 'timing/data': 0.03328929375857115, 'timing/model': 1.0635288879275322, 'learning_rate/qwen_vl_interface': 9.468449328565574e-06, 'learning_rate/action_model': 9.468449328565573e-05, 'epoch': 0.42}) +[2026-06-17 18:32:49,718][starVLA.training.train_starvla][INFO] - Step 1150, Loss: {'train/loss': 1.12661874294281, 'train/grad_norm_pre_clip': 12.887545585632324, 'timing/data': 0.0314791239798069, 'timing/model': 1.452088455669582, 'learning_rate/qwen_vl_interface': 9.467454787411282e-06, 'learning_rate/action_model': 9.467454787411281e-05, 'epoch': 0.42}) +[2026-06-17 18:32:53,747][starVLA.training.train_starvla][INFO] - Step 1151, Loss: {'train/loss': 0.7276303172111511, 'train/grad_norm_pre_clip': 13.612868309020996, 'timing/data': 0.03562164679169655, 'timing/model': 1.041888971813023, 'learning_rate/qwen_vl_interface': 9.466459371974076e-06, 'learning_rate/action_model': 9.466459371974076e-05, 'epoch': 0.42}) +[2026-06-17 18:32:58,105][starVLA.training.train_starvla][INFO] - Step 1152, Loss: {'train/loss': 0.9225580096244812, 'train/grad_norm_pre_clip': 13.445830345153809, 'timing/data': 0.03312185499817133, 'timing/model': 1.05031097214669, 'learning_rate/qwen_vl_interface': 9.465463082460304e-06, 'learning_rate/action_model': 9.465463082460304e-05, 'epoch': 0.42}) +[2026-06-17 18:33:02,417][starVLA.training.train_starvla][INFO] - Step 1153, Loss: {'train/loss': 1.0036053657531738, 'train/grad_norm_pre_clip': 10.952298164367676, 'timing/data': 0.03209285344928503, 'timing/model': 1.0175342680886388, 'learning_rate/qwen_vl_interface': 9.4644659190765e-06, 'learning_rate/action_model': 9.464465919076499e-05, 'epoch': 0.42}) +[2026-06-17 18:33:07,822][starVLA.training.train_starvla][INFO] - Step 1154, Loss: {'train/loss': 1.2188971042633057, 'train/grad_norm_pre_clip': 11.813915252685547, 'timing/data': 0.02994250226765871, 'timing/model': 1.0352882565930486, 'learning_rate/qwen_vl_interface': 9.463467882029374e-06, 'learning_rate/action_model': 9.463467882029373e-05, 'epoch': 0.42}) +[2026-06-17 18:33:16,592][starVLA.training.train_starvla][INFO] - Step 1155, Loss: {'train/loss': 0.7698391675949097, 'train/grad_norm_pre_clip': 11.212586402893066, 'timing/data': 0.02978928852826357, 'timing/model': 1.335722784511745, 'learning_rate/qwen_vl_interface': 9.462468971525822e-06, 'learning_rate/action_model': 9.462468971525821e-05, 'epoch': 0.42}) +[2026-06-17 18:33:20,743][starVLA.training.train_starvla][INFO] - Step 1156, Loss: {'train/loss': 0.9991040229797363, 'train/grad_norm_pre_clip': 15.913947105407715, 'timing/data': 0.03408662509173155, 'timing/model': 1.06228188239038, 'learning_rate/qwen_vl_interface': 9.461469187772921e-06, 'learning_rate/action_model': 9.46146918777292e-05, 'epoch': 0.42}) +[2026-06-17 18:33:25,289][starVLA.training.train_starvla][INFO] - Step 1157, Loss: {'train/loss': 1.0437085628509521, 'train/grad_norm_pre_clip': 13.096818923950195, 'timing/data': 0.031480262987315655, 'timing/model': 1.0482632592320442, 'learning_rate/qwen_vl_interface': 9.460468530977924e-06, 'learning_rate/action_model': 9.460468530977924e-05, 'epoch': 0.42}) +[2026-06-17 18:33:29,789][starVLA.training.train_starvla][INFO] - Step 1158, Loss: {'train/loss': 0.9601609706878662, 'train/grad_norm_pre_clip': 11.228496551513672, 'timing/data': 0.03140755835920572, 'timing/model': 1.0507668750360608, 'learning_rate/qwen_vl_interface': 9.459467001348271e-06, 'learning_rate/action_model': 9.45946700134827e-05, 'epoch': 0.42}) +[2026-06-17 18:33:34,361][starVLA.training.train_starvla][INFO] - Step 1159, Loss: {'train/loss': 0.8039278984069824, 'train/grad_norm_pre_clip': 10.642253875732422, 'timing/data': 0.03366898652166128, 'timing/model': 1.0792916398495436, 'learning_rate/qwen_vl_interface': 9.458464599091578e-06, 'learning_rate/action_model': 9.458464599091578e-05, 'epoch': 0.42}) +[2026-06-17 18:33:39,119][starVLA.training.train_starvla][INFO] - Step 1160, Loss: {'train/loss': 1.4639333486557007, 'train/grad_norm_pre_clip': 23.74625015258789, 'timing/data': 0.12922509387135506, 'timing/model': 1.5843664351850748, 'learning_rate/qwen_vl_interface': 9.457461324415647e-06, 'learning_rate/action_model': 9.457461324415646e-05, 'epoch': 0.42}) +[2026-06-17 18:33:57,268][starVLA.training.train_starvla][INFO] - Step 1161, Loss: {'train/loss': 0.7330243587493896, 'train/grad_norm_pre_clip': 8.909444808959961, 'timing/data': 0.029536155052483082, 'timing/model': 5.484511326998472, 'learning_rate/qwen_vl_interface': 9.456457177528454e-06, 'learning_rate/action_model': 9.456457177528453e-05, 'epoch': 0.43}) +[2026-06-17 18:34:05,477][starVLA.training.train_starvla][INFO] - Step 1162, Loss: {'train/loss': 0.7492550611495972, 'train/grad_norm_pre_clip': 9.722101211547852, 'timing/data': 0.04749458562582731, 'timing/model': 1.0235663680359721, 'learning_rate/qwen_vl_interface': 9.455452158638163e-06, 'learning_rate/action_model': 9.455452158638163e-05, 'epoch': 0.43}) +[2026-06-17 18:34:10,232][starVLA.training.train_starvla][INFO] - Step 1163, Loss: {'train/loss': 0.682867169380188, 'train/grad_norm_pre_clip': 9.066988945007324, 'timing/data': 0.04617369640618563, 'timing/model': 1.070440943352878, 'learning_rate/qwen_vl_interface': 9.454446267953114e-06, 'learning_rate/action_model': 9.454446267953114e-05, 'epoch': 0.43}) +[2026-06-17 18:34:17,143][starVLA.training.train_starvla][INFO] - Step 1164, Loss: {'train/loss': 0.9500823020935059, 'train/grad_norm_pre_clip': 12.109138488769531, 'timing/data': 0.04326233733445406, 'timing/model': 3.359255843795836, 'learning_rate/qwen_vl_interface': 9.453439505681829e-06, 'learning_rate/action_model': 9.453439505681829e-05, 'epoch': 0.43}) +[2026-06-17 18:34:24,014][starVLA.training.train_starvla][INFO] - Step 1165, Loss: {'train/loss': 1.0069528818130493, 'train/grad_norm_pre_clip': 12.058023452758789, 'timing/data': 0.16533921379595995, 'timing/model': 1.4955815803259611, 'learning_rate/qwen_vl_interface': 9.452431872033012e-06, 'learning_rate/action_model': 9.452431872033012e-05, 'epoch': 0.43}) +[2026-06-17 18:34:33,413][starVLA.training.train_starvla][INFO] - Step 1166, Loss: {'train/loss': 1.1634339094161987, 'train/grad_norm_pre_clip': 13.26209831237793, 'timing/data': 0.042340013198554516, 'timing/model': 5.040517578832805, 'learning_rate/qwen_vl_interface': 9.451423367215546e-06, 'learning_rate/action_model': 9.451423367215545e-05, 'epoch': 0.43}) +[2026-06-17 18:34:42,878][starVLA.training.train_starvla][INFO] - Step 1167, Loss: {'train/loss': 1.0689948797225952, 'train/grad_norm_pre_clip': 11.209518432617188, 'timing/data': 0.04800200276076794, 'timing/model': 1.0489417873322964, 'learning_rate/qwen_vl_interface': 9.450413991438494e-06, 'learning_rate/action_model': 9.450413991438494e-05, 'epoch': 0.43}) +[2026-06-17 18:34:55,685][starVLA.training.train_starvla][INFO] - Step 1168, Loss: {'train/loss': 1.0475951433181763, 'train/grad_norm_pre_clip': 11.876263618469238, 'timing/data': 0.04754401929676533, 'timing/model': 1.2130036978051066, 'learning_rate/qwen_vl_interface': 9.449403744911104e-06, 'learning_rate/action_model': 9.449403744911103e-05, 'epoch': 0.43}) +[2026-06-17 18:35:02,625][starVLA.training.train_starvla][INFO] - Step 1169, Loss: {'train/loss': 1.1551766395568848, 'train/grad_norm_pre_clip': 12.206430435180664, 'timing/data': 0.04165264591574669, 'timing/model': 1.1573658967390656, 'learning_rate/qwen_vl_interface': 9.448392627842797e-06, 'learning_rate/action_model': 9.448392627842795e-05, 'epoch': 0.43}) +[2026-06-17 18:35:11,722][starVLA.training.train_starvla][INFO] - Step 1170, Loss: {'train/loss': 0.9724050760269165, 'train/grad_norm_pre_clip': 12.169062614440918, 'timing/data': 0.04331290163099766, 'timing/model': 1.312568194232881, 'learning_rate/qwen_vl_interface': 9.44738064044318e-06, 'learning_rate/action_model': 9.44738064044318e-05, 'epoch': 0.43}) +[2026-06-17 18:35:15,874][starVLA.training.train_starvla][INFO] - Step 1171, Loss: {'train/loss': 0.9395083785057068, 'train/grad_norm_pre_clip': 11.161541938781738, 'timing/data': 0.04572909139096737, 'timing/model': 1.0554791269823909, 'learning_rate/qwen_vl_interface': 9.44636778292204e-06, 'learning_rate/action_model': 9.44636778292204e-05, 'epoch': 0.43}) +[2026-06-17 18:35:20,537][starVLA.training.train_starvla][INFO] - Step 1172, Loss: {'train/loss': 0.9128010272979736, 'train/grad_norm_pre_clip': 9.57964038848877, 'timing/data': 0.06898657325655222, 'timing/model': 1.1004559183493257, 'learning_rate/qwen_vl_interface': 9.445354055489345e-06, 'learning_rate/action_model': 9.445354055489345e-05, 'epoch': 0.43}) +[2026-06-17 18:35:24,948][starVLA.training.train_starvla][INFO] - Step 1173, Loss: {'train/loss': 0.802817165851593, 'train/grad_norm_pre_clip': 9.700119018554688, 'timing/data': 0.04402558971196413, 'timing/model': 1.0518158115446568, 'learning_rate/qwen_vl_interface': 9.444339458355239e-06, 'learning_rate/action_model': 9.444339458355239e-05, 'epoch': 0.43}) +[2026-06-17 18:35:29,356][starVLA.training.train_starvla][INFO] - Step 1174, Loss: {'train/loss': 0.8868101239204407, 'train/grad_norm_pre_clip': 10.144886016845703, 'timing/data': 0.0390285337343812, 'timing/model': 1.025835707783699, 'learning_rate/qwen_vl_interface': 9.443323991730049e-06, 'learning_rate/action_model': 9.443323991730049e-05, 'epoch': 0.43}) +[2026-06-17 18:35:33,781][starVLA.training.train_starvla][INFO] - Step 1175, Loss: {'train/loss': 0.9832445979118347, 'train/grad_norm_pre_clip': 11.806954383850098, 'timing/data': 0.045307490043342113, 'timing/model': 1.40846386551857, 'learning_rate/qwen_vl_interface': 9.442307655824284e-06, 'learning_rate/action_model': 9.442307655824285e-05, 'epoch': 0.43}) +[2026-06-17 18:35:37,916][starVLA.training.train_starvla][INFO] - Step 1176, Loss: {'train/loss': 0.7147232890129089, 'train/grad_norm_pre_clip': 8.638559341430664, 'timing/data': 0.05166422110050917, 'timing/model': 1.1807798650115728, 'learning_rate/qwen_vl_interface': 9.441290450848631e-06, 'learning_rate/action_model': 9.44129045084863e-05, 'epoch': 0.43}) +[2026-06-17 18:35:42,314][starVLA.training.train_starvla][INFO] - Step 1177, Loss: {'train/loss': 0.8811633586883545, 'train/grad_norm_pre_clip': 8.785270690917969, 'timing/data': 0.04989695083349943, 'timing/model': 1.0537435859441757, 'learning_rate/qwen_vl_interface': 9.440272377013957e-06, 'learning_rate/action_model': 9.440272377013956e-05, 'epoch': 0.43}) +[2026-06-17 18:35:46,684][starVLA.training.train_starvla][INFO] - Step 1178, Loss: {'train/loss': 1.17984139919281, 'train/grad_norm_pre_clip': 13.641374588012695, 'timing/data': 0.043653834611177444, 'timing/model': 1.0341317160055041, 'learning_rate/qwen_vl_interface': 9.43925343453131e-06, 'learning_rate/action_model': 9.43925343453131e-05, 'epoch': 0.43}) +[2026-06-17 18:35:50,960][starVLA.training.train_starvla][INFO] - Step 1179, Loss: {'train/loss': 0.5958899855613708, 'train/grad_norm_pre_clip': 8.224358558654785, 'timing/data': 0.041944487020373344, 'timing/model': 1.0298776412382722, 'learning_rate/qwen_vl_interface': 9.438233623611918e-06, 'learning_rate/action_model': 9.438233623611918e-05, 'epoch': 0.43}) +[2026-06-17 18:35:55,236][starVLA.training.train_starvla][INFO] - Step 1180, Loss: {'train/loss': 0.7775846719741821, 'train/grad_norm_pre_clip': 11.486392974853516, 'timing/data': 0.03817243967205286, 'timing/model': 1.2911509843543172, 'learning_rate/qwen_vl_interface': 9.43721294446719e-06, 'learning_rate/action_model': 9.437212944467191e-05, 'epoch': 0.43}) +[2026-06-17 18:35:59,348][starVLA.training.train_starvla][INFO] - Step 1181, Loss: {'train/loss': 0.8883888125419617, 'train/grad_norm_pre_clip': 14.41526985168457, 'timing/data': 0.046532352454960346, 'timing/model': 1.049354345537722, 'learning_rate/qwen_vl_interface': 9.436191397308713e-06, 'learning_rate/action_model': 9.436191397308713e-05, 'epoch': 0.43}) +[2026-06-17 18:36:03,660][starVLA.training.train_starvla][INFO] - Step 1182, Loss: {'train/loss': 0.7566357254981995, 'train/grad_norm_pre_clip': 11.876022338867188, 'timing/data': 0.044621302746236324, 'timing/model': 1.0415002666413784, 'learning_rate/qwen_vl_interface': 9.435168982348254e-06, 'learning_rate/action_model': 9.435168982348255e-05, 'epoch': 0.43}) +[2026-06-17 18:36:08,114][starVLA.training.train_starvla][INFO] - Step 1183, Loss: {'train/loss': 1.1972286701202393, 'train/grad_norm_pre_clip': 24.855215072631836, 'timing/data': 0.04008140601217747, 'timing/model': 1.1019063163548708, 'learning_rate/qwen_vl_interface': 9.434145699797763e-06, 'learning_rate/action_model': 9.434145699797764e-05, 'epoch': 0.43}) +[2026-06-17 18:36:12,608][starVLA.training.train_starvla][INFO] - Step 1184, Loss: {'train/loss': 0.8479220271110535, 'train/grad_norm_pre_clip': 8.744598388671875, 'timing/data': 0.04487017076462507, 'timing/model': 1.111404387280345, 'learning_rate/qwen_vl_interface': 9.433121549869365e-06, 'learning_rate/action_model': 9.433121549869365e-05, 'epoch': 0.43}) +[2026-06-17 18:36:16,928][starVLA.training.train_starvla][INFO] - Step 1185, Loss: {'train/loss': 0.6181869506835938, 'train/grad_norm_pre_clip': 10.043922424316406, 'timing/data': 0.042518327943980694, 'timing/model': 1.2925102962180972, 'learning_rate/qwen_vl_interface': 9.43209653277537e-06, 'learning_rate/action_model': 9.432096532775369e-05, 'epoch': 0.43}) +[2026-06-17 18:36:21,106][starVLA.training.train_starvla][INFO] - Step 1186, Loss: {'train/loss': 1.0593960285186768, 'train/grad_norm_pre_clip': 11.012337684631348, 'timing/data': 0.039211658760905266, 'timing/model': 1.0282096434384584, 'learning_rate/qwen_vl_interface': 9.431070648728262e-06, 'learning_rate/action_model': 9.431070648728261e-05, 'epoch': 0.43}) +[2026-06-17 18:36:25,653][starVLA.training.train_starvla][INFO] - Step 1187, Loss: {'train/loss': 0.9806240200996399, 'train/grad_norm_pre_clip': 12.724571228027344, 'timing/data': 0.0377311035990715, 'timing/model': 1.1406992608681321, 'learning_rate/qwen_vl_interface': 9.430043897940709e-06, 'learning_rate/action_model': 9.43004389794071e-05, 'epoch': 0.43}) +[2026-06-17 18:36:37,181][starVLA.training.train_starvla][INFO] - Step 1188, Loss: {'train/loss': 0.9395615458488464, 'train/grad_norm_pre_clip': 12.371216773986816, 'timing/data': 0.050484638661146164, 'timing/model': 5.176512602716684, 'learning_rate/qwen_vl_interface': 9.429016280625559e-06, 'learning_rate/action_model': 9.429016280625559e-05, 'epoch': 0.44}) +[2026-06-17 18:36:42,146][starVLA.training.train_starvla][INFO] - Step 1189, Loss: {'train/loss': 1.0520782470703125, 'train/grad_norm_pre_clip': 12.062973022460938, 'timing/data': 0.05222300719469786, 'timing/model': 1.1947673801332712, 'learning_rate/qwen_vl_interface': 9.427987796995838e-06, 'learning_rate/action_model': 9.427987796995837e-05, 'epoch': 0.44}) +[2026-06-17 18:36:50,607][starVLA.training.train_starvla][INFO] - Step 1190, Loss: {'train/loss': 0.9607232809066772, 'train/grad_norm_pre_clip': 13.51518726348877, 'timing/data': 0.039593931287527084, 'timing/model': 5.296717626973987, 'learning_rate/qwen_vl_interface': 9.42695844726475e-06, 'learning_rate/action_model': 9.426958447264749e-05, 'epoch': 0.44}) +[2026-06-17 18:36:56,445][starVLA.training.train_starvla][INFO] - Step 1191, Loss: {'train/loss': 0.6815750598907471, 'train/grad_norm_pre_clip': 9.628329277038574, 'timing/data': 0.050877002999186516, 'timing/model': 2.719257959164679, 'learning_rate/qwen_vl_interface': 9.42592823164568e-06, 'learning_rate/action_model': 9.42592823164568e-05, 'epoch': 0.44}) +[2026-06-17 18:37:04,535][starVLA.training.train_starvla][INFO] - Step 1192, Loss: {'train/loss': 0.8813126087188721, 'train/grad_norm_pre_clip': 13.823585510253906, 'timing/data': 0.039428217336535454, 'timing/model': 4.442863837815821, 'learning_rate/qwen_vl_interface': 9.424897150352197e-06, 'learning_rate/action_model': 9.424897150352197e-05, 'epoch': 0.44}) +[2026-06-17 18:37:09,720][starVLA.training.train_starvla][INFO] - Step 1193, Loss: {'train/loss': 0.6573750376701355, 'train/grad_norm_pre_clip': 12.08028507232666, 'timing/data': 0.04705998208373785, 'timing/model': 1.2030852809548378, 'learning_rate/qwen_vl_interface': 9.42386520359804e-06, 'learning_rate/action_model': 9.42386520359804e-05, 'epoch': 0.44}) +[2026-06-17 18:37:20,044][starVLA.training.train_starvla][INFO] - Step 1194, Loss: {'train/loss': 0.7111653089523315, 'train/grad_norm_pre_clip': 10.298148155212402, 'timing/data': 0.040717494674026966, 'timing/model': 1.0494634555652738, 'learning_rate/qwen_vl_interface': 9.422832391597136e-06, 'learning_rate/action_model': 9.422832391597136e-05, 'epoch': 0.44}) +[2026-06-17 18:37:25,283][starVLA.training.train_starvla][INFO] - Step 1195, Loss: {'train/loss': 0.984043300151825, 'train/grad_norm_pre_clip': 12.90408706665039, 'timing/data': 0.04226134344935417, 'timing/model': 1.4893678640946746, 'learning_rate/qwen_vl_interface': 9.421798714563586e-06, 'learning_rate/action_model': 9.421798714563586e-05, 'epoch': 0.44}) +[2026-06-17 18:37:29,550][starVLA.training.train_starvla][INFO] - Step 1196, Loss: {'train/loss': 0.9263761043548584, 'train/grad_norm_pre_clip': 12.358380317687988, 'timing/data': 0.04741180036216974, 'timing/model': 1.0864477967843413, 'learning_rate/qwen_vl_interface': 9.420764172711676e-06, 'learning_rate/action_model': 9.420764172711676e-05, 'epoch': 0.44}) +[2026-06-17 18:37:33,983][starVLA.training.train_starvla][INFO] - Step 1197, Loss: {'train/loss': 0.7443047761917114, 'train/grad_norm_pre_clip': 10.98271656036377, 'timing/data': 0.041712744161486626, 'timing/model': 1.0213858438655734, 'learning_rate/qwen_vl_interface': 9.419728766255863e-06, 'learning_rate/action_model': 9.419728766255862e-05, 'epoch': 0.44}) +[2026-06-17 18:37:38,322][starVLA.training.train_starvla][INFO] - Step 1198, Loss: {'train/loss': 0.803640604019165, 'train/grad_norm_pre_clip': 9.153635025024414, 'timing/data': 0.04743661638349295, 'timing/model': 1.064234558492899, 'learning_rate/qwen_vl_interface': 9.418692495410789e-06, 'learning_rate/action_model': 9.418692495410789e-05, 'epoch': 0.44}) +[2026-06-17 18:37:42,664][starVLA.training.train_starvla][INFO] - Step 1199, Loss: {'train/loss': 1.1024647951126099, 'train/grad_norm_pre_clip': 12.060795783996582, 'timing/data': 0.03938794136047363, 'timing/model': 1.0362885287031531, 'learning_rate/qwen_vl_interface': 9.417655360391277e-06, 'learning_rate/action_model': 9.417655360391277e-05, 'epoch': 0.44}) +[2026-06-17 18:40:27,253][starVLA.training.train_starvla][INFO] - Step 1200, Loss: {'train/loss': 0.871864378452301, 'train/grad_norm_pre_clip': 15.124364852905273, 'eval/loss': 0.9479005932807922, 'eval/latency_2/loss': 0.9479005932807922, 'eval/demon_fire_precision': 0.9203096539162113, 'eval/demon_fire_recall': 0.8456066945606695, 'eval/demon_fire_f1': 0.8813781072830353, 'eval/demon_fire_rate_model': 0.6110183639398998, 'eval/demon_fire_rate_teacher': 0.6649972175848636, 'eval/demon_left_precision': 0.8330804248861912, 'eval/demon_left_recall': 0.5778947368421052, 'eval/demon_left_f1': 0.6824114356743319, 'eval/demon_left_rate_model': 0.18336115748469672, 'eval/demon_left_rate_teacher': 0.2643294379521425, 'eval/demon_right_precision': 0.6568160152526216, 'eval/demon_right_recall': 0.7785310734463277, 'eval/demon_right_f1': 0.7125129265770425, 'eval/demon_right_rate_model': 0.29187534780189206, 'eval/demon_right_rate_teacher': 0.24624373956594323, 'eval/demon_move_f1': 0.6974621811256871, 'eval/demon_cc_f1': 0.8167654933408565, 'eval/latency_2/demon_fire_precision': 0.9203096539162113, 'eval/latency_2/demon_fire_recall': 0.8456066945606695, 'eval/latency_2/demon_fire_f1': 0.8813781072830353, 'eval/latency_2/demon_fire_rate_model': 0.6110183639398998, 'eval/latency_2/demon_fire_rate_teacher': 0.6649972175848636, 'eval/latency_2/demon_left_precision': 0.8330804248861912, 'eval/latency_2/demon_left_recall': 0.5778947368421052, 'eval/latency_2/demon_left_f1': 0.6824114356743319, 'eval/latency_2/demon_left_rate_model': 0.18336115748469672, 'eval/latency_2/demon_left_rate_teacher': 0.2643294379521425, 'eval/latency_2/demon_right_precision': 0.6568160152526216, 'eval/latency_2/demon_right_recall': 0.7785310734463277, 'eval/latency_2/demon_right_f1': 0.7125129265770425, 'eval/latency_2/demon_right_rate_model': 0.29187534780189206, 'eval/latency_2/demon_right_rate_teacher': 0.24624373956594323, 'eval/latency_2/demon_move_f1': 0.6974621811256871, 'eval/latency_2/demon_cc_f1': 0.8167654933408565, 'timing/data': 0.04509792663156986, 'timing/model': 1.356858391314745, 'learning_rate/qwen_vl_interface': 9.416617361412325e-06, 'learning_rate/action_model': 9.416617361412325e-05, 'epoch': 0.44}) +[2026-06-17 18:40:27,284][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_1200_state +[2026-06-17 18:40:27,291][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 18:40:36,044][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_1200_state/model.safetensors +[2026-06-17 18:40:51,055][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_1200_state/optimizer.bin +[2026-06-17 18:40:51,056][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_1200_state/sampler.bin +[2026-06-17 18:40:51,056][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_1200_state/sampler_1.bin +[2026-06-17 18:40:51,059][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_1200_state/random_states_0.pkl +[2026-06-17 18:40:51,062][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1200 +[2026-06-17 18:40:51,063][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 18:40:51,079][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 18:40:55,725][starVLA.training.train_starvla][INFO] - Step 1201, Loss: {'train/loss': 0.6824963092803955, 'train/grad_norm_pre_clip': 10.242818832397461, 'timing/data': 0.053880225867033005, 'timing/model': 1.3628462497144938, 'learning_rate/qwen_vl_interface': 9.415578498689106e-06, 'learning_rate/action_model': 9.415578498689106e-05, 'epoch': 0.44}) +[2026-06-17 18:41:00,121][starVLA.training.train_starvla][INFO] - Step 1202, Loss: {'train/loss': 0.9869419932365417, 'train/grad_norm_pre_clip': 12.39946174621582, 'timing/data': 0.04231814667582512, 'timing/model': 1.1335983593016863, 'learning_rate/qwen_vl_interface': 9.414538772436987e-06, 'learning_rate/action_model': 9.414538772436986e-05, 'epoch': 0.44}) +[2026-06-17 18:41:04,718][starVLA.training.train_starvla][INFO] - Step 1203, Loss: {'train/loss': 0.6395591497421265, 'train/grad_norm_pre_clip': 8.877397537231445, 'timing/data': 0.10172920301556587, 'timing/model': 1.070805775001645, 'learning_rate/qwen_vl_interface': 9.413498182871496e-06, 'learning_rate/action_model': 9.413498182871496e-05, 'epoch': 0.44}) +[2026-06-17 18:41:09,489][starVLA.training.train_starvla][INFO] - Step 1204, Loss: {'train/loss': 0.8614562153816223, 'train/grad_norm_pre_clip': 29.335359573364258, 'timing/data': 0.052982863038778305, 'timing/model': 1.201857440173626, 'learning_rate/qwen_vl_interface': 9.412456730208352e-06, 'learning_rate/action_model': 9.412456730208352e-05, 'epoch': 0.44}) +[2026-06-17 18:41:14,250][starVLA.training.train_starvla][INFO] - Step 1205, Loss: {'train/loss': 0.6740168333053589, 'train/grad_norm_pre_clip': 10.2219877243042, 'timing/data': 0.04207540024071932, 'timing/model': 1.0918727163225412, 'learning_rate/qwen_vl_interface': 9.41141441466345e-06, 'learning_rate/action_model': 9.41141441466345e-05, 'epoch': 0.44}) +[2026-06-17 18:41:18,589][starVLA.training.train_starvla][INFO] - Step 1206, Loss: {'train/loss': 0.9380730390548706, 'train/grad_norm_pre_clip': 15.994696617126465, 'timing/data': 0.0389487287029624, 'timing/model': 1.3322464460507035, 'learning_rate/qwen_vl_interface': 9.41037123645286e-06, 'learning_rate/action_model': 9.41037123645286e-05, 'epoch': 0.44}) +[2026-06-17 18:41:22,840][starVLA.training.train_starvla][INFO] - Step 1207, Loss: {'train/loss': 0.9042413830757141, 'train/grad_norm_pre_clip': 13.087238311767578, 'timing/data': 0.04373085964471102, 'timing/model': 1.1112478552386165, 'learning_rate/qwen_vl_interface': 9.409327195792838e-06, 'learning_rate/action_model': 9.409327195792838e-05, 'epoch': 0.44}) +[2026-06-17 18:41:28,131][starVLA.training.train_starvla][INFO] - Step 1208, Loss: {'train/loss': 0.8856225609779358, 'train/grad_norm_pre_clip': 11.713644981384277, 'timing/data': 0.04034368973225355, 'timing/model': 1.0531211327761412, 'learning_rate/qwen_vl_interface': 9.408282292899812e-06, 'learning_rate/action_model': 9.408282292899812e-05, 'epoch': 0.44}) +[2026-06-17 18:41:32,853][starVLA.training.train_starvla][INFO] - Step 1209, Loss: {'train/loss': 1.1209636926651, 'train/grad_norm_pre_clip': 12.64219856262207, 'timing/data': 0.10052397567778826, 'timing/model': 1.030969644896686, 'learning_rate/qwen_vl_interface': 9.407236527990394e-06, 'learning_rate/action_model': 9.407236527990393e-05, 'epoch': 0.44}) +[2026-06-17 18:41:37,202][starVLA.training.train_starvla][INFO] - Step 1210, Loss: {'train/loss': 0.7978469729423523, 'train/grad_norm_pre_clip': 12.87028980255127, 'timing/data': 0.04435041267424822, 'timing/model': 1.0389489457011223, 'learning_rate/qwen_vl_interface': 9.406189901281369e-06, 'learning_rate/action_model': 9.406189901281368e-05, 'epoch': 0.44}) +[2026-06-17 18:41:41,706][starVLA.training.train_starvla][INFO] - Step 1211, Loss: {'train/loss': 1.1283169984817505, 'train/grad_norm_pre_clip': 21.706342697143555, 'timing/data': 0.04218602739274502, 'timing/model': 1.4012858597561717, 'learning_rate/qwen_vl_interface': 9.405142412989705e-06, 'learning_rate/action_model': 9.405142412989704e-05, 'epoch': 0.44}) +[2026-06-17 18:41:45,955][starVLA.training.train_starvla][INFO] - Step 1212, Loss: {'train/loss': 0.957841694355011, 'train/grad_norm_pre_clip': 11.884666442871094, 'timing/data': 0.045473781414330006, 'timing/model': 1.0357263395562768, 'learning_rate/qwen_vl_interface': 9.404094063332546e-06, 'learning_rate/action_model': 9.404094063332546e-05, 'epoch': 0.44}) +[2026-06-17 18:41:50,360][starVLA.training.train_starvla][INFO] - Step 1213, Loss: {'train/loss': 0.626569390296936, 'train/grad_norm_pre_clip': 10.063584327697754, 'timing/data': 0.044524677097797394, 'timing/model': 1.0686578322201967, 'learning_rate/qwen_vl_interface': 9.40304485252722e-06, 'learning_rate/action_model': 9.403044852527219e-05, 'epoch': 0.44}) +[2026-06-17 18:41:54,763][starVLA.training.train_starvla][INFO] - Step 1214, Loss: {'train/loss': 0.9508417844772339, 'train/grad_norm_pre_clip': 13.309670448303223, 'timing/data': 0.036133202724158764, 'timing/model': 1.1465970920398831, 'learning_rate/qwen_vl_interface': 9.401994780791226e-06, 'learning_rate/action_model': 9.401994780791226e-05, 'epoch': 0.44}) +[2026-06-17 18:41:59,271][starVLA.training.train_starvla][INFO] - Step 1215, Loss: {'train/loss': 0.7771314382553101, 'train/grad_norm_pre_clip': 10.541804313659668, 'timing/data': 0.045156000182032585, 'timing/model': 1.067152375355363, 'learning_rate/qwen_vl_interface': 9.400943848342246e-06, 'learning_rate/action_model': 9.400943848342246e-05, 'epoch': 0.44}) +[2026-06-17 18:42:03,566][starVLA.training.train_starvla][INFO] - Step 1216, Loss: {'train/loss': 0.8823255896568298, 'train/grad_norm_pre_clip': 12.220048904418945, 'timing/data': 0.04190206900238991, 'timing/model': 1.3205678425729275, 'learning_rate/qwen_vl_interface': 9.39989205539814e-06, 'learning_rate/action_model': 9.39989205539814e-05, 'epoch': 0.45}) +[2026-06-17 18:42:09,362][starVLA.training.train_starvla][INFO] - Step 1217, Loss: {'train/loss': 0.9231011867523193, 'train/grad_norm_pre_clip': 13.13967514038086, 'timing/data': 0.04892593156546354, 'timing/model': 1.024638599716127, 'learning_rate/qwen_vl_interface': 9.398839402176946e-06, 'learning_rate/action_model': 9.398839402176946e-05, 'epoch': 0.45}) +[2026-06-17 18:42:13,841][starVLA.training.train_starvla][INFO] - Step 1218, Loss: {'train/loss': 0.6184191107749939, 'train/grad_norm_pre_clip': 11.221162796020508, 'timing/data': 0.042030563578009605, 'timing/model': 1.075552868656814, 'learning_rate/qwen_vl_interface': 9.39778588889688e-06, 'learning_rate/action_model': 9.39778588889688e-05, 'epoch': 0.45}) +[2026-06-17 18:42:21,551][starVLA.training.train_starvla][INFO] - Step 1219, Loss: {'train/loss': 1.0882328748703003, 'train/grad_norm_pre_clip': 12.566652297973633, 'timing/data': 0.037422445602715015, 'timing/model': 1.07023448869586, 'learning_rate/qwen_vl_interface': 9.396731515776335e-06, 'learning_rate/action_model': 9.396731515776333e-05, 'epoch': 0.45}) +[2026-06-17 18:42:26,540][starVLA.training.train_starvla][INFO] - Step 1220, Loss: {'train/loss': 0.8976194858551025, 'train/grad_norm_pre_clip': 12.943140983581543, 'timing/data': 0.047644300386309624, 'timing/model': 1.4653813960030675, 'learning_rate/qwen_vl_interface': 9.395676283033884e-06, 'learning_rate/action_model': 9.395676283033885e-05, 'epoch': 0.45}) +[2026-06-17 18:42:34,349][starVLA.training.train_starvla][INFO] - Step 1221, Loss: {'train/loss': 0.7814903855323792, 'train/grad_norm_pre_clip': 9.393535614013672, 'timing/data': 0.04713441524654627, 'timing/model': 1.421584821306169, 'learning_rate/qwen_vl_interface': 9.39462019088828e-06, 'learning_rate/action_model': 9.394620190888279e-05, 'epoch': 0.45}) +[2026-06-17 18:42:38,658][starVLA.training.train_starvla][INFO] - Step 1222, Loss: {'train/loss': 1.0529804229736328, 'train/grad_norm_pre_clip': 12.87759017944336, 'timing/data': 0.04287244472652674, 'timing/model': 1.1377562480047345, 'learning_rate/qwen_vl_interface': 9.393563239558449e-06, 'learning_rate/action_model': 9.39356323955845e-05, 'epoch': 0.45}) +[2026-06-17 18:42:49,308][starVLA.training.train_starvla][INFO] - Step 1223, Loss: {'train/loss': 0.9653605222702026, 'train/grad_norm_pre_clip': 12.602791786193848, 'timing/data': 0.04733758885413408, 'timing/model': 5.575346154160798, 'learning_rate/qwen_vl_interface': 9.392505429263501e-06, 'learning_rate/action_model': 9.392505429263501e-05, 'epoch': 0.45}) +[2026-06-17 18:42:54,152][starVLA.training.train_starvla][INFO] - Step 1224, Loss: {'train/loss': 0.9007222056388855, 'train/grad_norm_pre_clip': 12.515177726745605, 'timing/data': 0.04961917642503977, 'timing/model': 1.0379819916561246, 'learning_rate/qwen_vl_interface': 9.39144676022272e-06, 'learning_rate/action_model': 9.39144676022272e-05, 'epoch': 0.45}) +[2026-06-17 18:42:58,496][starVLA.training.train_starvla][INFO] - Step 1225, Loss: {'train/loss': 0.7150869369506836, 'train/grad_norm_pre_clip': 10.768134117126465, 'timing/data': 0.0428236173465848, 'timing/model': 1.0460470505058765, 'learning_rate/qwen_vl_interface': 9.390387232655568e-06, 'learning_rate/action_model': 9.390387232655569e-05, 'epoch': 0.45}) +[2026-06-17 18:43:03,296][starVLA.training.train_starvla][INFO] - Step 1226, Loss: {'train/loss': 0.8480385541915894, 'train/grad_norm_pre_clip': 10.088272094726562, 'timing/data': 0.040431175380945206, 'timing/model': 1.5199316963553429, 'learning_rate/qwen_vl_interface': 9.38932684678169e-06, 'learning_rate/action_model': 9.38932684678169e-05, 'epoch': 0.45}) +[2026-06-17 18:43:07,320][starVLA.training.train_starvla][INFO] - Step 1227, Loss: {'train/loss': 0.8844214081764221, 'train/grad_norm_pre_clip': 11.692538261413574, 'timing/data': 0.0422873692587018, 'timing/model': 1.0437756953760982, 'learning_rate/qwen_vl_interface': 9.388265602820901e-06, 'learning_rate/action_model': 9.3882656028209e-05, 'epoch': 0.45}) +[2026-06-17 18:43:11,720][starVLA.training.train_starvla][INFO] - Step 1228, Loss: {'train/loss': 0.8679608106613159, 'train/grad_norm_pre_clip': 11.50027847290039, 'timing/data': 0.04879925213754177, 'timing/model': 1.043511789292097, 'learning_rate/qwen_vl_interface': 9.387203500993199e-06, 'learning_rate/action_model': 9.387203500993199e-05, 'epoch': 0.45}) +[2026-06-17 18:43:16,174][starVLA.training.train_starvla][INFO] - Step 1229, Loss: {'train/loss': 0.6992629766464233, 'train/grad_norm_pre_clip': 10.319314002990723, 'timing/data': 0.0461597703397274, 'timing/model': 1.0261045647785068, 'learning_rate/qwen_vl_interface': 9.38614054151876e-06, 'learning_rate/action_model': 9.38614054151876e-05, 'epoch': 0.45}) +[2026-06-17 18:43:22,319][starVLA.training.train_starvla][INFO] - Step 1230, Loss: {'train/loss': 0.8892424702644348, 'train/grad_norm_pre_clip': 14.536870002746582, 'timing/data': 0.12131388485431671, 'timing/model': 1.3748736148700118, 'learning_rate/qwen_vl_interface': 9.385076724617937e-06, 'learning_rate/action_model': 9.385076724617935e-05, 'epoch': 0.45}) +[2026-06-17 18:43:27,186][starVLA.training.train_starvla][INFO] - Step 1231, Loss: {'train/loss': 0.650343120098114, 'train/grad_norm_pre_clip': 11.062828063964844, 'timing/data': 0.013147656805813313, 'timing/model': 1.589498565532267, 'learning_rate/qwen_vl_interface': 9.384012050511258e-06, 'learning_rate/action_model': 9.384012050511258e-05, 'epoch': 0.45}) +[2026-06-17 18:43:34,976][starVLA.training.train_starvla][INFO] - Step 1232, Loss: {'train/loss': 0.7472177147865295, 'train/grad_norm_pre_clip': 11.60098934173584, 'timing/data': 0.04921284690499306, 'timing/model': 1.180658409371972, 'learning_rate/qwen_vl_interface': 9.382946519419433e-06, 'learning_rate/action_model': 9.382946519419434e-05, 'epoch': 0.45}) +[2026-06-17 18:43:39,985][starVLA.training.train_starvla][INFO] - Step 1233, Loss: {'train/loss': 0.8696939945220947, 'train/grad_norm_pre_clip': 13.361412048339844, 'timing/data': 0.04877392202615738, 'timing/model': 1.1925048297271132, 'learning_rate/qwen_vl_interface': 9.381880131563347e-06, 'learning_rate/action_model': 9.381880131563346e-05, 'epoch': 0.45}) +[2026-06-17 18:43:44,636][starVLA.training.train_starvla][INFO] - Step 1234, Loss: {'train/loss': 0.815425455570221, 'train/grad_norm_pre_clip': 10.223007202148438, 'timing/data': 0.03979965206235647, 'timing/model': 1.133354158140719, 'learning_rate/qwen_vl_interface': 9.380812887164065e-06, 'learning_rate/action_model': 9.380812887164065e-05, 'epoch': 0.45}) +[2026-06-17 18:43:52,624][starVLA.training.train_starvla][INFO] - Step 1235, Loss: {'train/loss': 0.5103610157966614, 'train/grad_norm_pre_clip': 8.049752235412598, 'timing/data': 0.041852792724967, 'timing/model': 4.666016369126737, 'learning_rate/qwen_vl_interface': 9.379744786442824e-06, 'learning_rate/action_model': 9.379744786442824e-05, 'epoch': 0.45}) +[2026-06-17 18:43:57,306][starVLA.training.train_starvla][INFO] - Step 1236, Loss: {'train/loss': 0.9366986155509949, 'train/grad_norm_pre_clip': 10.812959671020508, 'timing/data': 0.03976790700107813, 'timing/model': 1.4343648785725236, 'learning_rate/qwen_vl_interface': 9.378675829621047e-06, 'learning_rate/action_model': 9.378675829621046e-05, 'epoch': 0.45}) +[2026-06-17 18:44:03,041][starVLA.training.train_starvla][INFO] - Step 1237, Loss: {'train/loss': 1.1419039964675903, 'train/grad_norm_pre_clip': 13.553759574890137, 'timing/data': 0.045526036992669106, 'timing/model': 1.0818959986791015, 'learning_rate/qwen_vl_interface': 9.377606016920328e-06, 'learning_rate/action_model': 9.377606016920327e-05, 'epoch': 0.45}) +[2026-06-17 18:44:07,374][starVLA.training.train_starvla][INFO] - Step 1238, Loss: {'train/loss': 0.5807620286941528, 'train/grad_norm_pre_clip': 12.012042045593262, 'timing/data': 0.04103440325707197, 'timing/model': 1.0244986778125167, 'learning_rate/qwen_vl_interface': 9.376535348562438e-06, 'learning_rate/action_model': 9.376535348562439e-05, 'epoch': 0.45}) +[2026-06-17 18:44:11,728][starVLA.training.train_starvla][INFO] - Step 1239, Loss: {'train/loss': 0.6766160130500793, 'train/grad_norm_pre_clip': 10.389690399169922, 'timing/data': 0.045969923958182335, 'timing/model': 1.050897279754281, 'learning_rate/qwen_vl_interface': 9.375463824769332e-06, 'learning_rate/action_model': 9.375463824769333e-05, 'epoch': 0.45}) +[2026-06-17 18:44:16,285][starVLA.training.train_starvla][INFO] - Step 1240, Loss: {'train/loss': 0.8158975839614868, 'train/grad_norm_pre_clip': 13.318171501159668, 'timing/data': 0.04298245348036289, 'timing/model': 1.122471290640533, 'learning_rate/qwen_vl_interface': 9.374391445763135e-06, 'learning_rate/action_model': 9.374391445763135e-05, 'epoch': 0.45}) +[2026-06-17 18:44:20,991][starVLA.training.train_starvla][INFO] - Step 1241, Loss: {'train/loss': 1.177555799484253, 'train/grad_norm_pre_clip': 21.287492752075195, 'timing/data': 0.04078123904764652, 'timing/model': 1.4866602951660752, 'learning_rate/qwen_vl_interface': 9.373318211766155e-06, 'learning_rate/action_model': 9.373318211766155e-05, 'epoch': 0.45}) +[2026-06-17 18:44:25,379][starVLA.training.train_starvla][INFO] - Step 1242, Loss: {'train/loss': 1.0278096199035645, 'train/grad_norm_pre_clip': 12.898940086364746, 'timing/data': 0.04563355166465044, 'timing/model': 1.2034852914512157, 'learning_rate/qwen_vl_interface': 9.37224412300087e-06, 'learning_rate/action_model': 9.37224412300087e-05, 'epoch': 0.45}) +[2026-06-17 18:44:29,864][starVLA.training.train_starvla][INFO] - Step 1243, Loss: {'train/loss': 1.2602102756500244, 'train/grad_norm_pre_clip': 16.3052921295166, 'timing/data': 0.04283720813691616, 'timing/model': 1.0748430388048291, 'learning_rate/qwen_vl_interface': 9.371169179689945e-06, 'learning_rate/action_model': 9.371169179689945e-05, 'epoch': 0.46}) +[2026-06-17 18:44:34,746][starVLA.training.train_starvla][INFO] - Step 1244, Loss: {'train/loss': 0.7981777191162109, 'train/grad_norm_pre_clip': 9.229944229125977, 'timing/data': 0.04402249213308096, 'timing/model': 1.1316145630553365, 'learning_rate/qwen_vl_interface': 9.370093382056213e-06, 'learning_rate/action_model': 9.370093382056214e-05, 'epoch': 0.46}) +[2026-06-17 18:44:39,552][starVLA.training.train_starvla][INFO] - Step 1245, Loss: {'train/loss': 0.7973420023918152, 'train/grad_norm_pre_clip': 11.021828651428223, 'timing/data': 0.0444391593337059, 'timing/model': 1.0994837693870068, 'learning_rate/qwen_vl_interface': 9.369016730322692e-06, 'learning_rate/action_model': 9.36901673032269e-05, 'epoch': 0.46}) +[2026-06-17 18:44:44,291][starVLA.training.train_starvla][INFO] - Step 1246, Loss: {'train/loss': 1.0941044092178345, 'train/grad_norm_pre_clip': 14.269164085388184, 'timing/data': 0.04983496107161045, 'timing/model': 1.6870687240734696, 'learning_rate/qwen_vl_interface': 9.367939224712566e-06, 'learning_rate/action_model': 9.367939224712566e-05, 'epoch': 0.46}) +[2026-06-17 18:44:48,446][starVLA.training.train_starvla][INFO] - Step 1247, Loss: {'train/loss': 0.9752513766288757, 'train/grad_norm_pre_clip': 9.422800064086914, 'timing/data': 0.03993123862892389, 'timing/model': 1.0442998567596078, 'learning_rate/qwen_vl_interface': 9.366860865449213e-06, 'learning_rate/action_model': 9.366860865449212e-05, 'epoch': 0.46}) +[2026-06-17 18:44:53,180][starVLA.training.train_starvla][INFO] - Step 1248, Loss: {'train/loss': 0.6555604338645935, 'train/grad_norm_pre_clip': 10.45487117767334, 'timing/data': 0.05101699382066727, 'timing/model': 1.0476531740278006, 'learning_rate/qwen_vl_interface': 9.365781652756168e-06, 'learning_rate/action_model': 9.365781652756168e-05, 'epoch': 0.46}) +[2026-06-17 18:44:57,799][starVLA.training.train_starvla][INFO] - Step 1249, Loss: {'train/loss': 0.7286990284919739, 'train/grad_norm_pre_clip': 9.852645874023438, 'timing/data': 0.04169107135385275, 'timing/model': 1.0944950869306922, 'learning_rate/qwen_vl_interface': 9.36470158685716e-06, 'learning_rate/action_model': 9.36470158685716e-05, 'epoch': 0.46}) +[2026-06-17 18:45:02,469][starVLA.training.train_starvla][INFO] - Step 1250, Loss: {'train/loss': 0.8928737044334412, 'train/grad_norm_pre_clip': 11.433928489685059, 'timing/data': 0.04170997254550457, 'timing/model': 1.2714397264644504, 'learning_rate/qwen_vl_interface': 9.363620667976085e-06, 'learning_rate/action_model': 9.363620667976085e-05, 'epoch': 0.46}) +[2026-06-17 18:45:07,022][starVLA.training.train_starvla][INFO] - Step 1251, Loss: {'train/loss': 1.0622591972351074, 'train/grad_norm_pre_clip': 15.0189208984375, 'timing/data': 0.03933075163513422, 'timing/model': 1.3527776971459389, 'learning_rate/qwen_vl_interface': 9.362538896337019e-06, 'learning_rate/action_model': 9.362538896337019e-05, 'epoch': 0.46}) +[2026-06-17 18:45:17,777][starVLA.training.train_starvla][INFO] - Step 1252, Loss: {'train/loss': 0.822381317615509, 'train/grad_norm_pre_clip': 11.111505508422852, 'timing/data': 0.04748423211276531, 'timing/model': 1.036764102987945, 'learning_rate/qwen_vl_interface': 9.361456272164215e-06, 'learning_rate/action_model': 9.361456272164214e-05, 'epoch': 0.46}) +[2026-06-17 18:45:22,250][starVLA.training.train_starvla][INFO] - Step 1253, Loss: {'train/loss': 0.9486652612686157, 'train/grad_norm_pre_clip': 31.052751541137695, 'timing/data': 0.043018653988838196, 'timing/model': 1.0869622165337205, 'learning_rate/qwen_vl_interface': 9.3603727956821e-06, 'learning_rate/action_model': 9.3603727956821e-05, 'epoch': 0.46}) +[2026-06-17 18:45:26,891][starVLA.training.train_starvla][INFO] - Step 1254, Loss: {'train/loss': 0.6815980076789856, 'train/grad_norm_pre_clip': 11.043231010437012, 'timing/data': 0.03867632057517767, 'timing/model': 1.1969044050201774, 'learning_rate/qwen_vl_interface': 9.359288467115283e-06, 'learning_rate/action_model': 9.359288467115283e-05, 'epoch': 0.46}) +[2026-06-17 18:45:35,832][starVLA.training.train_starvla][INFO] - Step 1255, Loss: {'train/loss': 1.0585901737213135, 'train/grad_norm_pre_clip': 10.625167846679688, 'timing/data': 0.04512578155845404, 'timing/model': 5.514957147650421, 'learning_rate/qwen_vl_interface': 9.358203286688544e-06, 'learning_rate/action_model': 9.358203286688544e-05, 'epoch': 0.46}) +[2026-06-17 18:45:40,840][starVLA.training.train_starvla][INFO] - Step 1256, Loss: {'train/loss': 0.8242944478988647, 'train/grad_norm_pre_clip': 10.327407836914062, 'timing/data': 0.04488419368863106, 'timing/model': 1.3791292747482657, 'learning_rate/qwen_vl_interface': 9.357117254626844e-06, 'learning_rate/action_model': 9.357117254626844e-05, 'epoch': 0.46}) +[2026-06-17 18:45:47,985][starVLA.training.train_starvla][INFO] - Step 1257, Loss: {'train/loss': 1.1507126092910767, 'train/grad_norm_pre_clip': 13.613508224487305, 'timing/data': 0.07032398879528046, 'timing/model': 1.1121965199708939, 'learning_rate/qwen_vl_interface': 9.356030371155316e-06, 'learning_rate/action_model': 9.356030371155316e-05, 'epoch': 0.46}) +[2026-06-17 18:45:52,645][starVLA.training.train_starvla][INFO] - Step 1258, Loss: {'train/loss': 0.6167322993278503, 'train/grad_norm_pre_clip': 10.897737503051758, 'timing/data': 0.04734709393233061, 'timing/model': 1.079432349652052, 'learning_rate/qwen_vl_interface': 9.354942636499275e-06, 'learning_rate/action_model': 9.354942636499276e-05, 'epoch': 0.46}) +[2026-06-17 18:46:01,605][starVLA.training.train_starvla][INFO] - Step 1259, Loss: {'train/loss': 0.6277689337730408, 'train/grad_norm_pre_clip': 9.617257118225098, 'timing/data': 0.038973210379481316, 'timing/model': 5.605184485204518, 'learning_rate/qwen_vl_interface': 9.35385405088421e-06, 'learning_rate/action_model': 9.353854050884208e-05, 'epoch': 0.46}) +[2026-06-17 18:46:09,906][starVLA.training.train_starvla][INFO] - Step 1260, Loss: {'train/loss': 1.2722970247268677, 'train/grad_norm_pre_clip': 12.908705711364746, 'timing/data': 0.048348418436944485, 'timing/model': 1.1578500252217054, 'learning_rate/qwen_vl_interface': 9.352764614535782e-06, 'learning_rate/action_model': 9.352764614535781e-05, 'epoch': 0.46}) +[2026-06-17 18:46:14,238][starVLA.training.train_starvla][INFO] - Step 1261, Loss: {'train/loss': 0.8345999121665955, 'train/grad_norm_pre_clip': 10.70256519317627, 'timing/data': 0.039111580699682236, 'timing/model': 1.2995080724358559, 'learning_rate/qwen_vl_interface': 9.351674327679837e-06, 'learning_rate/action_model': 9.351674327679836e-05, 'epoch': 0.46}) +[2026-06-17 18:46:18,549][starVLA.training.train_starvla][INFO] - Step 1262, Loss: {'train/loss': 0.7085641622543335, 'train/grad_norm_pre_clip': 10.004979133605957, 'timing/data': 0.045274932868778706, 'timing/model': 1.0747599601745605, 'learning_rate/qwen_vl_interface': 9.350583190542387e-06, 'learning_rate/action_model': 9.350583190542387e-05, 'epoch': 0.46}) +[2026-06-17 18:46:23,382][starVLA.training.train_starvla][INFO] - Step 1263, Loss: {'train/loss': 0.7359296679496765, 'train/grad_norm_pre_clip': 12.184527397155762, 'timing/data': 0.05170549172908068, 'timing/model': 1.039858765900135, 'learning_rate/qwen_vl_interface': 9.34949120334963e-06, 'learning_rate/action_model': 9.34949120334963e-05, 'epoch': 0.46}) +[2026-06-17 18:46:32,453][starVLA.training.train_starvla][INFO] - Step 1264, Loss: {'train/loss': 0.8252752423286438, 'train/grad_norm_pre_clip': 10.363682746887207, 'timing/data': 0.0513142729178071, 'timing/model': 5.513972700573504, 'learning_rate/qwen_vl_interface': 9.348398366327938e-06, 'learning_rate/action_model': 9.348398366327937e-05, 'epoch': 0.46}) +[2026-06-17 18:46:40,932][starVLA.training.train_starvla][INFO] - Step 1265, Loss: {'train/loss': 0.7407881617546082, 'train/grad_norm_pre_clip': 9.861047744750977, 'timing/data': 0.04960283450782299, 'timing/model': 4.8928557112813, 'learning_rate/qwen_vl_interface': 9.347304679703852e-06, 'learning_rate/action_model': 9.347304679703851e-05, 'epoch': 0.46}) +[2026-06-17 18:46:45,579][starVLA.training.train_starvla][INFO] - Step 1266, Loss: {'train/loss': 1.069631576538086, 'train/grad_norm_pre_clip': 11.28839111328125, 'timing/data': 0.048637584783136845, 'timing/model': 1.4373652627691627, 'learning_rate/qwen_vl_interface': 9.346210143704099e-06, 'learning_rate/action_model': 9.346210143704098e-05, 'epoch': 0.46}) +[2026-06-17 18:46:52,401][starVLA.training.train_starvla][INFO] - Step 1267, Loss: {'train/loss': 0.7637850046157837, 'train/grad_norm_pre_clip': 9.913192749023438, 'timing/data': 0.05006987415254116, 'timing/model': 3.670612433925271, 'learning_rate/qwen_vl_interface': 9.345114758555573e-06, 'learning_rate/action_model': 9.345114758555573e-05, 'epoch': 0.46}) +[2026-06-17 18:46:58,539][starVLA.training.train_starvla][INFO] - Step 1268, Loss: {'train/loss': 0.954531729221344, 'train/grad_norm_pre_clip': 18.067949295043945, 'timing/data': 0.05862634442746639, 'timing/model': 2.641232750378549, 'learning_rate/qwen_vl_interface': 9.344018524485352e-06, 'learning_rate/action_model': 9.344018524485351e-05, 'epoch': 0.46}) +[2026-06-17 18:47:06,348][starVLA.training.train_starvla][INFO] - Step 1269, Loss: {'train/loss': 0.9917047023773193, 'train/grad_norm_pre_clip': 12.400315284729004, 'timing/data': 0.06246063578873873, 'timing/model': 1.0880679599940777, 'learning_rate/qwen_vl_interface': 9.342921441720686e-06, 'learning_rate/action_model': 9.342921441720685e-05, 'epoch': 0.46}) +[2026-06-17 18:47:10,797][starVLA.training.train_starvla][INFO] - Step 1270, Loss: {'train/loss': 1.0251191854476929, 'train/grad_norm_pre_clip': 12.30415153503418, 'timing/data': 0.04076694417744875, 'timing/model': 1.0501876417547464, 'learning_rate/qwen_vl_interface': 9.341823510489e-06, 'learning_rate/action_model': 9.341823510489e-05, 'epoch': 0.47}) +[2026-06-17 18:47:15,388][starVLA.training.train_starvla][INFO] - Step 1271, Loss: {'train/loss': 0.5757533311843872, 'train/grad_norm_pre_clip': 7.633093357086182, 'timing/data': 0.0379560524597764, 'timing/model': 1.4940497912466526, 'learning_rate/qwen_vl_interface': 9.340724731017898e-06, 'learning_rate/action_model': 9.340724731017898e-05, 'epoch': 0.47}) +[2026-06-17 18:47:19,723][starVLA.training.train_starvla][INFO] - Step 1272, Loss: {'train/loss': 0.7709372639656067, 'train/grad_norm_pre_clip': 11.126900672912598, 'timing/data': 0.048676139675080776, 'timing/model': 1.1339554889127612, 'learning_rate/qwen_vl_interface': 9.339625103535158e-06, 'learning_rate/action_model': 9.339625103535156e-05, 'epoch': 0.47}) +[2026-06-17 18:47:24,165][starVLA.training.train_starvla][INFO] - Step 1273, Loss: {'train/loss': 0.6174789667129517, 'train/grad_norm_pre_clip': 10.53641414642334, 'timing/data': 0.10947462450712919, 'timing/model': 1.0497055603191257, 'learning_rate/qwen_vl_interface': 9.338524628268734e-06, 'learning_rate/action_model': 9.338524628268733e-05, 'epoch': 0.47}) +[2026-06-17 18:47:29,209][starVLA.training.train_starvla][INFO] - Step 1274, Loss: {'train/loss': 0.7269473075866699, 'train/grad_norm_pre_clip': 9.567201614379883, 'timing/data': 0.045448687858879566, 'timing/model': 1.39024752471596, 'learning_rate/qwen_vl_interface': 9.337423305446753e-06, 'learning_rate/action_model': 9.337423305446753e-05, 'epoch': 0.47}) +[2026-06-17 18:47:34,672][starVLA.training.train_starvla][INFO] - Step 1275, Loss: {'train/loss': 0.8267743587493896, 'train/grad_norm_pre_clip': 15.487139701843262, 'timing/data': 0.10868643689900637, 'timing/model': 1.354207506403327, 'learning_rate/qwen_vl_interface': 9.336321135297525e-06, 'learning_rate/action_model': 9.336321135297525e-05, 'epoch': 0.47}) +[2026-06-17 18:47:43,181][starVLA.training.train_starvla][INFO] - Step 1276, Loss: {'train/loss': 0.756008505821228, 'train/grad_norm_pre_clip': 9.370710372924805, 'timing/data': 0.05175085365772247, 'timing/model': 1.5645760810002685, 'learning_rate/qwen_vl_interface': 9.335218118049526e-06, 'learning_rate/action_model': 9.335218118049526e-05, 'epoch': 0.47}) +[2026-06-17 18:47:47,617][starVLA.training.train_starvla][INFO] - Step 1277, Loss: {'train/loss': 0.7702024579048157, 'train/grad_norm_pre_clip': 12.47612476348877, 'timing/data': 0.05870293453335762, 'timing/model': 1.0566602414473891, 'learning_rate/qwen_vl_interface': 9.334114253931417e-06, 'learning_rate/action_model': 9.334114253931416e-05, 'epoch': 0.47}) +[2026-06-17 18:47:52,169][starVLA.training.train_starvla][INFO] - Step 1278, Loss: {'train/loss': 0.933856725692749, 'train/grad_norm_pre_clip': 14.094192504882812, 'timing/data': 0.04154070559889078, 'timing/model': 1.038605460897088, 'learning_rate/qwen_vl_interface': 9.333009543172027e-06, 'learning_rate/action_model': 9.333009543172027e-05, 'epoch': 0.47}) +[2026-06-17 18:47:58,067][starVLA.training.train_starvla][INFO] - Step 1279, Loss: {'train/loss': 1.1063916683197021, 'train/grad_norm_pre_clip': 14.783759117126465, 'timing/data': 0.0543231638148427, 'timing/model': 1.3832337632775307, 'learning_rate/qwen_vl_interface': 9.331903986000365e-06, 'learning_rate/action_model': 9.331903986000365e-05, 'epoch': 0.47}) +[2026-06-17 18:48:02,685][starVLA.training.train_starvla][INFO] - Step 1280, Loss: {'train/loss': 0.7552647590637207, 'train/grad_norm_pre_clip': 10.738537788391113, 'timing/data': 0.05451726075261831, 'timing/model': 1.0277845822274685, 'learning_rate/qwen_vl_interface': 9.330797582645615e-06, 'learning_rate/action_model': 9.330797582645615e-05, 'epoch': 0.47}) +[2026-06-17 18:48:08,184][starVLA.training.train_starvla][INFO] - Step 1281, Loss: {'train/loss': 0.7788583636283875, 'train/grad_norm_pre_clip': 10.685733795166016, 'timing/data': 0.04392548557370901, 'timing/model': 1.4235233720391989, 'learning_rate/qwen_vl_interface': 9.329690333337137e-06, 'learning_rate/action_model': 9.329690333337137e-05, 'epoch': 0.47}) +[2026-06-17 18:48:12,679][starVLA.training.train_starvla][INFO] - Step 1282, Loss: {'train/loss': 0.8234718441963196, 'train/grad_norm_pre_clip': 14.209980964660645, 'timing/data': 0.04627616982907057, 'timing/model': 1.027565729804337, 'learning_rate/qwen_vl_interface': 9.328582238304461e-06, 'learning_rate/action_model': 9.328582238304461e-05, 'epoch': 0.47}) +[2026-06-17 18:48:17,392][starVLA.training.train_starvla][INFO] - Step 1283, Loss: {'train/loss': 0.7482554912567139, 'train/grad_norm_pre_clip': 10.844950675964355, 'timing/data': 0.04451209492981434, 'timing/model': 1.1038824301213026, 'learning_rate/qwen_vl_interface': 9.327473297777297e-06, 'learning_rate/action_model': 9.327473297777297e-05, 'epoch': 0.47}) +[2026-06-17 18:48:21,883][starVLA.training.train_starvla][INFO] - Step 1284, Loss: {'train/loss': 0.7559683322906494, 'train/grad_norm_pre_clip': 18.912405014038086, 'timing/data': 0.07021302357316017, 'timing/model': 1.0956252915784717, 'learning_rate/qwen_vl_interface': 9.326363511985533e-06, 'learning_rate/action_model': 9.326363511985533e-05, 'epoch': 0.47}) +[2026-06-17 18:48:26,416][starVLA.training.train_starvla][INFO] - Step 1285, Loss: {'train/loss': 0.6047112941741943, 'train/grad_norm_pre_clip': 11.21469497680664, 'timing/data': 0.05519981775432825, 'timing/model': 1.087016575038433, 'learning_rate/qwen_vl_interface': 9.325252881159226e-06, 'learning_rate/action_model': 9.325252881159226e-05, 'epoch': 0.47}) +[2026-06-17 18:48:30,712][starVLA.training.train_starvla][INFO] - Step 1286, Loss: {'train/loss': 0.8622410297393799, 'train/grad_norm_pre_clip': 9.905749320983887, 'timing/data': 0.04359045345336199, 'timing/model': 1.3010885855183005, 'learning_rate/qwen_vl_interface': 9.324141405528612e-06, 'learning_rate/action_model': 9.324141405528612e-05, 'epoch': 0.47}) +[2026-06-17 18:48:34,815][starVLA.training.train_starvla][INFO] - Step 1287, Loss: {'train/loss': 1.0428904294967651, 'train/grad_norm_pre_clip': 13.855469703674316, 'timing/data': 0.04091065842658281, 'timing/model': 1.1075465073809028, 'learning_rate/qwen_vl_interface': 9.3230290853241e-06, 'learning_rate/action_model': 9.3230290853241e-05, 'epoch': 0.47}) +[2026-06-17 18:48:39,446][starVLA.training.train_starvla][INFO] - Step 1288, Loss: {'train/loss': 0.8344210982322693, 'train/grad_norm_pre_clip': 10.367759704589844, 'timing/data': 0.050368004478514194, 'timing/model': 1.044755881652236, 'learning_rate/qwen_vl_interface': 9.321915920776278e-06, 'learning_rate/action_model': 9.321915920776278e-05, 'epoch': 0.47}) +[2026-06-17 18:48:43,893][starVLA.training.train_starvla][INFO] - Step 1289, Loss: {'train/loss': 1.2965725660324097, 'train/grad_norm_pre_clip': 14.734964370727539, 'timing/data': 0.04004942998290062, 'timing/model': 1.0357290701940656, 'learning_rate/qwen_vl_interface': 9.320801912115901e-06, 'learning_rate/action_model': 9.320801912115901e-05, 'epoch': 0.47}) +[2026-06-17 18:48:48,389][starVLA.training.train_starvla][INFO] - Step 1290, Loss: {'train/loss': 0.8460085391998291, 'train/grad_norm_pre_clip': 11.064239501953125, 'timing/data': 0.04803184885531664, 'timing/model': 1.0757853807881474, 'learning_rate/qwen_vl_interface': 9.319687059573909e-06, 'learning_rate/action_model': 9.319687059573909e-05, 'epoch': 0.47}) +[2026-06-17 18:48:52,913][starVLA.training.train_starvla][INFO] - Step 1291, Loss: {'train/loss': 0.8081862330436707, 'train/grad_norm_pre_clip': 11.663727760314941, 'timing/data': 0.05092151090502739, 'timing/model': 1.2766955206170678, 'learning_rate/qwen_vl_interface': 9.318571363381408e-06, 'learning_rate/action_model': 9.318571363381408e-05, 'epoch': 0.47}) +[2026-06-17 18:48:57,067][starVLA.training.train_starvla][INFO] - Step 1292, Loss: {'train/loss': 0.6385007500648499, 'train/grad_norm_pre_clip': 8.836649894714355, 'timing/data': 0.049696934409439564, 'timing/model': 1.1412487020716071, 'learning_rate/qwen_vl_interface': 9.317454823769686e-06, 'learning_rate/action_model': 9.317454823769687e-05, 'epoch': 0.47}) +[2026-06-17 18:49:01,501][starVLA.training.train_starvla][INFO] - Step 1293, Loss: {'train/loss': 0.82660973072052, 'train/grad_norm_pre_clip': 12.914461135864258, 'timing/data': 0.048324478790163994, 'timing/model': 1.065436533652246, 'learning_rate/qwen_vl_interface': 9.316337440970206e-06, 'learning_rate/action_model': 9.316337440970206e-05, 'epoch': 0.47}) +[2026-06-17 18:49:06,036][starVLA.training.train_starvla][INFO] - Step 1294, Loss: {'train/loss': 0.9381935000419617, 'train/grad_norm_pre_clip': 14.493385314941406, 'timing/data': 0.041109600104391575, 'timing/model': 1.0294820927083492, 'learning_rate/qwen_vl_interface': 9.315219215214596e-06, 'learning_rate/action_model': 9.315219215214595e-05, 'epoch': 0.47}) +[2026-06-17 18:49:10,708][starVLA.training.train_starvla][INFO] - Step 1295, Loss: {'train/loss': 0.7795265913009644, 'train/grad_norm_pre_clip': 10.555155754089355, 'timing/data': 0.044203720055520535, 'timing/model': 1.0536008644849062, 'learning_rate/qwen_vl_interface': 9.314100146734668e-06, 'learning_rate/action_model': 9.314100146734668e-05, 'epoch': 0.47}) +[2026-06-17 18:49:15,219][starVLA.training.train_starvla][INFO] - Step 1296, Loss: {'train/loss': 0.7326449155807495, 'train/grad_norm_pre_clip': 13.31725025177002, 'timing/data': 0.05020004231482744, 'timing/model': 1.3910314124077559, 'learning_rate/qwen_vl_interface': 9.312980235762408e-06, 'learning_rate/action_model': 9.312980235762408e-05, 'epoch': 0.47}) +[2026-06-17 18:49:19,344][starVLA.training.train_starvla][INFO] - Step 1297, Loss: {'train/loss': 0.7707500457763672, 'train/grad_norm_pre_clip': 10.10479736328125, 'timing/data': 0.04553858656436205, 'timing/model': 1.0637230332940817, 'learning_rate/qwen_vl_interface': 9.311859482529971e-06, 'learning_rate/action_model': 9.31185948252997e-05, 'epoch': 0.47}) +[2026-06-17 18:49:26,004][starVLA.training.train_starvla][INFO] - Step 1298, Loss: {'train/loss': 1.1071544885635376, 'train/grad_norm_pre_clip': 12.544201850891113, 'timing/data': 0.04231402184814215, 'timing/model': 1.1674255141988397, 'learning_rate/qwen_vl_interface': 9.310737887269694e-06, 'learning_rate/action_model': 9.310737887269694e-05, 'epoch': 0.48}) +[2026-06-17 18:49:32,303][starVLA.training.train_starvla][INFO] - Step 1299, Loss: {'train/loss': 0.9644901156425476, 'train/grad_norm_pre_clip': 12.075445175170898, 'timing/data': 0.053916667588055134, 'timing/model': 1.6335963811725378, 'learning_rate/qwen_vl_interface': 9.309615450214084e-06, 'learning_rate/action_model': 9.309615450214084e-05, 'epoch': 0.48}) +[2026-06-17 18:53:17,928][starVLA.training.train_starvla][INFO] - Step 1300, Loss: {'train/loss': 0.8863177299499512, 'train/grad_norm_pre_clip': 29.49139404296875, 'eval/loss': 0.8804256319999695, 'eval/latency_2/loss': 0.8804256319999695, 'eval/demon_fire_precision': 0.8567659253675085, 'eval/demon_fire_recall': 0.9510460251046026, 'eval/demon_fire_f1': 0.9014475510608765, 'eval/demon_fire_rate_model': 0.738174735670562, 'eval/demon_fire_rate_teacher': 0.6649972175848636, 'eval/demon_left_precision': 0.714148219441771, 'eval/demon_left_recall': 0.7810526315789473, 'eval/demon_left_f1': 0.7461035696329813, 'eval/demon_left_rate_model': 0.2890929326655537, 'eval/demon_left_rate_teacher': 0.2643294379521425, 'eval/demon_right_precision': 0.7234262125902993, 'eval/demon_right_recall': 0.792090395480226, 'eval/demon_right_f1': 0.756202804746494, 'eval/demon_right_rate_model': 0.2696160267111853, 'eval/demon_right_rate_teacher': 0.24624373956594323, 'eval/demon_move_f1': 0.7511531871897377, 'eval/demon_cc_f1': 0.8504017991611966, 'eval/latency_2/demon_fire_precision': 0.8567659253675085, 'eval/latency_2/demon_fire_recall': 0.9510460251046026, 'eval/latency_2/demon_fire_f1': 0.9014475510608765, 'eval/latency_2/demon_fire_rate_model': 0.738174735670562, 'eval/latency_2/demon_fire_rate_teacher': 0.6649972175848636, 'eval/latency_2/demon_left_precision': 0.714148219441771, 'eval/latency_2/demon_left_recall': 0.7810526315789473, 'eval/latency_2/demon_left_f1': 0.7461035696329813, 'eval/latency_2/demon_left_rate_model': 0.2890929326655537, 'eval/latency_2/demon_left_rate_teacher': 0.2643294379521425, 'eval/latency_2/demon_right_precision': 0.7234262125902993, 'eval/latency_2/demon_right_recall': 0.792090395480226, 'eval/latency_2/demon_right_f1': 0.756202804746494, 'eval/latency_2/demon_right_rate_model': 0.2696160267111853, 'eval/latency_2/demon_right_rate_teacher': 0.24624373956594323, 'eval/latency_2/demon_move_f1': 0.7511531871897377, 'eval/latency_2/demon_cc_f1': 0.8504017991611966, 'timing/data': 0.056989945471286774, 'timing/model': 1.0995215373113751, 'learning_rate/qwen_vl_interface': 9.308492171595821e-06, 'learning_rate/action_model': 9.30849217159582e-05, 'epoch': 0.48}) +[2026-06-17 18:53:17,948][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_1300_state +[2026-06-17 18:53:17,955][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 18:53:26,022][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_1300_state/model.safetensors +[2026-06-17 18:53:41,041][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_1300_state/optimizer.bin +[2026-06-17 18:53:41,042][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_1300_state/sampler.bin +[2026-06-17 18:53:41,043][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_1300_state/sampler_1.bin +[2026-06-17 18:53:41,045][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_1300_state/random_states_0.pkl +[2026-06-17 18:53:41,050][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1300 +[2026-06-17 18:53:41,051][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 18:53:41,072][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 18:53:45,735][starVLA.training.train_starvla][INFO] - Step 1301, Loss: {'train/loss': 0.9366483092308044, 'train/grad_norm_pre_clip': 14.402046203613281, 'timing/data': 0.03621367458254099, 'timing/model': 1.0868783965706825, 'learning_rate/qwen_vl_interface': 9.307368051647765e-06, 'learning_rate/action_model': 9.307368051647765e-05, 'epoch': 0.48}) +[2026-06-17 18:53:50,119][starVLA.training.train_starvla][INFO] - Step 1302, Loss: {'train/loss': 0.7765122056007385, 'train/grad_norm_pre_clip': 11.980783462524414, 'timing/data': 0.03206414636224508, 'timing/model': 1.2875260515138507, 'learning_rate/qwen_vl_interface': 9.306243090602945e-06, 'learning_rate/action_model': 9.306243090602946e-05, 'epoch': 0.48}) +[2026-06-17 18:53:54,353][starVLA.training.train_starvla][INFO] - Step 1303, Loss: {'train/loss': 0.7508398294448853, 'train/grad_norm_pre_clip': 11.9087553024292, 'timing/data': 0.03139710798859596, 'timing/model': 1.080583724193275, 'learning_rate/qwen_vl_interface': 9.30511728869457e-06, 'learning_rate/action_model': 9.305117288694569e-05, 'epoch': 0.48}) +[2026-06-17 18:53:58,848][starVLA.training.train_starvla][INFO] - Step 1304, Loss: {'train/loss': 0.9077948331832886, 'train/grad_norm_pre_clip': 11.121620178222656, 'timing/data': 0.0330310445278883, 'timing/model': 1.1519161565229297, 'learning_rate/qwen_vl_interface': 9.303990646156015e-06, 'learning_rate/action_model': 9.303990646156014e-05, 'epoch': 0.48}) +[2026-06-17 18:54:03,530][starVLA.training.train_starvla][INFO] - Step 1305, Loss: {'train/loss': 1.0275132656097412, 'train/grad_norm_pre_clip': 12.032208442687988, 'timing/data': 0.035945020616054535, 'timing/model': 1.1184951569885015, 'learning_rate/qwen_vl_interface': 9.302863163220837e-06, 'learning_rate/action_model': 9.302863163220837e-05, 'epoch': 0.48}) +[2026-06-17 18:54:08,100][starVLA.training.train_starvla][INFO] - Step 1306, Loss: {'train/loss': 0.9862173199653625, 'train/grad_norm_pre_clip': 14.167837142944336, 'timing/data': 0.03454635851085186, 'timing/model': 1.0795632898807526, 'learning_rate/qwen_vl_interface': 9.301734840122765e-06, 'learning_rate/action_model': 9.301734840122766e-05, 'epoch': 0.48}) +[2026-06-17 18:54:12,266][starVLA.training.train_starvla][INFO] - Step 1307, Loss: {'train/loss': 1.0929076671600342, 'train/grad_norm_pre_clip': 12.825639724731445, 'timing/data': 0.029866740107536316, 'timing/model': 1.0332174664363265, 'learning_rate/qwen_vl_interface': 9.300605677095701e-06, 'learning_rate/action_model': 9.3006056770957e-05, 'epoch': 0.48}) +[2026-06-17 18:54:16,667][starVLA.training.train_starvla][INFO] - Step 1308, Loss: {'train/loss': 0.8512713313102722, 'train/grad_norm_pre_clip': 10.908613204956055, 'timing/data': 0.028589525260031223, 'timing/model': 1.0260619753971696, 'learning_rate/qwen_vl_interface': 9.299475674373722e-06, 'learning_rate/action_model': 9.299475674373721e-05, 'epoch': 0.48}) +[2026-06-17 18:54:21,155][starVLA.training.train_starvla][INFO] - Step 1309, Loss: {'train/loss': 0.9698367714881897, 'train/grad_norm_pre_clip': 11.871926307678223, 'timing/data': 0.034368316642940044, 'timing/model': 1.0290367482230067, 'learning_rate/qwen_vl_interface': 9.298344832191076e-06, 'learning_rate/action_model': 9.298344832191076e-05, 'epoch': 0.48}) +[2026-06-17 18:54:25,621][starVLA.training.train_starvla][INFO] - Step 1310, Loss: {'train/loss': 1.0737981796264648, 'train/grad_norm_pre_clip': 11.581080436706543, 'timing/data': 0.028471133671700954, 'timing/model': 1.0378099409863353, 'learning_rate/qwen_vl_interface': 9.297213150782192e-06, 'learning_rate/action_model': 9.297213150782192e-05, 'epoch': 0.48}) +[2026-06-17 18:54:30,205][starVLA.training.train_starvla][INFO] - Step 1311, Loss: {'train/loss': 0.6899705529212952, 'train/grad_norm_pre_clip': 12.02767562866211, 'timing/data': 0.03667911794036627, 'timing/model': 1.3875636206939816, 'learning_rate/qwen_vl_interface': 9.296080630381666e-06, 'learning_rate/action_model': 9.296080630381666e-05, 'epoch': 0.48}) +[2026-06-17 18:54:34,669][starVLA.training.train_starvla][INFO] - Step 1312, Loss: {'train/loss': 0.8136577010154724, 'train/grad_norm_pre_clip': 12.257177352905273, 'timing/data': 0.03550674766302109, 'timing/model': 1.4751073336228728, 'learning_rate/qwen_vl_interface': 9.294947271224271e-06, 'learning_rate/action_model': 9.294947271224272e-05, 'epoch': 0.48}) +[2026-06-17 18:54:38,967][starVLA.training.train_starvla][INFO] - Step 1313, Loss: {'train/loss': 0.8914231061935425, 'train/grad_norm_pre_clip': 10.975510597229004, 'timing/data': 0.030840198509395123, 'timing/model': 1.1300615780055523, 'learning_rate/qwen_vl_interface': 9.293813073544957e-06, 'learning_rate/action_model': 9.293813073544956e-05, 'epoch': 0.48}) +[2026-06-17 18:54:43,286][starVLA.training.train_starvla][INFO] - Step 1314, Loss: {'train/loss': 0.7598613500595093, 'train/grad_norm_pre_clip': 13.426923751831055, 'timing/data': 0.037748025730252266, 'timing/model': 1.0468121264129877, 'learning_rate/qwen_vl_interface': 9.292678037578838e-06, 'learning_rate/action_model': 9.292678037578837e-05, 'epoch': 0.48}) +[2026-06-17 18:54:47,720][starVLA.training.train_starvla][INFO] - Step 1315, Loss: {'train/loss': 0.8986246585845947, 'train/grad_norm_pre_clip': 12.312274932861328, 'timing/data': 0.027993563562631607, 'timing/model': 1.0974306724965572, 'learning_rate/qwen_vl_interface': 9.291542163561214e-06, 'learning_rate/action_model': 9.291542163561213e-05, 'epoch': 0.48}) +[2026-06-17 18:54:52,276][starVLA.training.train_starvla][INFO] - Step 1316, Loss: {'train/loss': 0.715073823928833, 'train/grad_norm_pre_clip': 8.79065227508545, 'timing/data': 0.034112341701984406, 'timing/model': 1.0298815974965692, 'learning_rate/qwen_vl_interface': 9.29040545172755e-06, 'learning_rate/action_model': 9.290405451727549e-05, 'epoch': 0.48}) +[2026-06-17 18:54:56,988][starVLA.training.train_starvla][INFO] - Step 1317, Loss: {'train/loss': 0.9121764898300171, 'train/grad_norm_pre_clip': 13.723764419555664, 'timing/data': 0.03581725433468819, 'timing/model': 1.5017628008499742, 'learning_rate/qwen_vl_interface': 9.289267902313486e-06, 'learning_rate/action_model': 9.289267902313487e-05, 'epoch': 0.48}) +[2026-06-17 18:55:01,539][starVLA.training.train_starvla][INFO] - Step 1318, Loss: {'train/loss': 0.8365491032600403, 'train/grad_norm_pre_clip': 53.429412841796875, 'timing/data': 0.034341455437242985, 'timing/model': 1.258428548462689, 'learning_rate/qwen_vl_interface': 9.288129515554843e-06, 'learning_rate/action_model': 9.288129515554843e-05, 'epoch': 0.48}) +[2026-06-17 18:55:06,477][starVLA.training.train_starvla][INFO] - Step 1319, Loss: {'train/loss': 0.6734505295753479, 'train/grad_norm_pre_clip': 10.707452774047852, 'timing/data': 0.034058703109622, 'timing/model': 1.5430789655074477, 'learning_rate/qwen_vl_interface': 9.286990291687605e-06, 'learning_rate/action_model': 9.286990291687604e-05, 'epoch': 0.48}) +[2026-06-17 18:55:12,559][starVLA.training.train_starvla][INFO] - Step 1320, Loss: {'train/loss': 0.7611769437789917, 'train/grad_norm_pre_clip': 12.535431861877441, 'timing/data': 0.02872585505247116, 'timing/model': 1.0997905861586332, 'learning_rate/qwen_vl_interface': 9.285850230947934e-06, 'learning_rate/action_model': 9.285850230947934e-05, 'epoch': 0.48}) +[2026-06-17 18:55:17,157][starVLA.training.train_starvla][INFO] - Step 1321, Loss: {'train/loss': 0.6596088409423828, 'train/grad_norm_pre_clip': 10.353772163391113, 'timing/data': 0.03885790053755045, 'timing/model': 1.1515159914270043, 'learning_rate/qwen_vl_interface': 9.28470933357217e-06, 'learning_rate/action_model': 9.28470933357217e-05, 'epoch': 0.48}) +[2026-06-17 18:55:21,664][starVLA.training.train_starvla][INFO] - Step 1322, Loss: {'train/loss': 1.0345109701156616, 'train/grad_norm_pre_clip': 11.723955154418945, 'timing/data': 0.02973010018467903, 'timing/model': 1.4742264952510595, 'learning_rate/qwen_vl_interface': 9.283567599796819e-06, 'learning_rate/action_model': 9.283567599796818e-05, 'epoch': 0.48}) +[2026-06-17 18:55:25,811][starVLA.training.train_starvla][INFO] - Step 1323, Loss: {'train/loss': 0.8371288776397705, 'train/grad_norm_pre_clip': 12.353337287902832, 'timing/data': 0.03341332916170359, 'timing/model': 1.0323733994737267, 'learning_rate/qwen_vl_interface': 9.282425029858561e-06, 'learning_rate/action_model': 9.282425029858561e-05, 'epoch': 0.48}) +[2026-06-17 18:55:30,235][starVLA.training.train_starvla][INFO] - Step 1324, Loss: {'train/loss': 0.9310422539710999, 'train/grad_norm_pre_clip': 12.677413940429688, 'timing/data': 0.03314196411520243, 'timing/model': 1.0630120672285557, 'learning_rate/qwen_vl_interface': 9.281281623994258e-06, 'learning_rate/action_model': 9.281281623994258e-05, 'epoch': 0.48}) +[2026-06-17 18:55:34,701][starVLA.training.train_starvla][INFO] - Step 1325, Loss: {'train/loss': 0.6255802512168884, 'train/grad_norm_pre_clip': 9.888131141662598, 'timing/data': 0.036489629186689854, 'timing/model': 1.040536431595683, 'learning_rate/qwen_vl_interface': 9.280137382440936e-06, 'learning_rate/action_model': 9.280137382440937e-05, 'epoch': 0.49}) +[2026-06-17 18:55:44,706][starVLA.training.train_starvla][INFO] - Step 1326, Loss: {'train/loss': 0.6089840531349182, 'train/grad_norm_pre_clip': 8.546220779418945, 'timing/data': 0.03302103281021118, 'timing/model': 1.0282386532053351, 'learning_rate/qwen_vl_interface': 9.278992305435798e-06, 'learning_rate/action_model': 9.278992305435797e-05, 'epoch': 0.49}) +[2026-06-17 18:55:49,252][starVLA.training.train_starvla][INFO] - Step 1327, Loss: {'train/loss': 0.7012940049171448, 'train/grad_norm_pre_clip': 10.461472511291504, 'timing/data': 0.04391487222164869, 'timing/model': 1.4144468838348985, 'learning_rate/qwen_vl_interface': 9.27784639321622e-06, 'learning_rate/action_model': 9.27784639321622e-05, 'epoch': 0.49}) +[2026-06-17 18:55:53,458][starVLA.training.train_starvla][INFO] - Step 1328, Loss: {'train/loss': 0.9488977789878845, 'train/grad_norm_pre_clip': 14.645676612854004, 'timing/data': 0.03268803935497999, 'timing/model': 1.0268236622214317, 'learning_rate/qwen_vl_interface': 9.27669964601975e-06, 'learning_rate/action_model': 9.27669964601975e-05, 'epoch': 0.49}) +[2026-06-17 18:55:58,000][starVLA.training.train_starvla][INFO] - Step 1329, Loss: {'train/loss': 0.78813236951828, 'train/grad_norm_pre_clip': 10.186603546142578, 'timing/data': 0.032295445911586285, 'timing/model': 1.059050184674561, 'learning_rate/qwen_vl_interface': 9.27555206408411e-06, 'learning_rate/action_model': 9.275552064084111e-05, 'epoch': 0.49}) +[2026-06-17 18:56:02,362][starVLA.training.train_starvla][INFO] - Step 1330, Loss: {'train/loss': 0.9168780446052551, 'train/grad_norm_pre_clip': 12.818531036376953, 'timing/data': 0.034371708519756794, 'timing/model': 1.0291033573448658, 'learning_rate/qwen_vl_interface': 9.274403647647197e-06, 'learning_rate/action_model': 9.274403647647197e-05, 'epoch': 0.49}) +[2026-06-17 18:56:06,835][starVLA.training.train_starvla][INFO] - Step 1331, Loss: {'train/loss': 0.7205784320831299, 'train/grad_norm_pre_clip': 12.280366897583008, 'timing/data': 0.03331563714891672, 'timing/model': 1.1052902769297361, 'learning_rate/qwen_vl_interface': 9.273254396947079e-06, 'learning_rate/action_model': 9.273254396947078e-05, 'epoch': 0.49}) +[2026-06-17 18:56:11,339][starVLA.training.train_starvla][INFO] - Step 1332, Loss: {'train/loss': 0.6662853360176086, 'train/grad_norm_pre_clip': 9.955892562866211, 'timing/data': 0.035360963083803654, 'timing/model': 1.483006913214922, 'learning_rate/qwen_vl_interface': 9.27210431222199e-06, 'learning_rate/action_model': 9.27210431222199e-05, 'epoch': 0.49}) +[2026-06-17 18:56:15,948][starVLA.training.train_starvla][INFO] - Step 1333, Loss: {'train/loss': 0.8733704686164856, 'train/grad_norm_pre_clip': 12.217720031738281, 'timing/data': 0.0321937371045351, 'timing/model': 1.0266677727922797, 'learning_rate/qwen_vl_interface': 9.270953393710353e-06, 'learning_rate/action_model': 9.270953393710352e-05, 'epoch': 0.49}) +[2026-06-17 18:56:20,288][starVLA.training.train_starvla][INFO] - Step 1334, Loss: {'train/loss': 0.888633668422699, 'train/grad_norm_pre_clip': 13.949268341064453, 'timing/data': 0.044132993556559086, 'timing/model': 1.0216524070128798, 'learning_rate/qwen_vl_interface': 9.269801641650751e-06, 'learning_rate/action_model': 9.26980164165075e-05, 'epoch': 0.49}) +[2026-06-17 18:56:24,731][starVLA.training.train_starvla][INFO] - Step 1335, Loss: {'train/loss': 0.8772851228713989, 'train/grad_norm_pre_clip': 11.542566299438477, 'timing/data': 0.031715523451566696, 'timing/model': 1.0727401431649923, 'learning_rate/qwen_vl_interface': 9.268649056281941e-06, 'learning_rate/action_model': 9.268649056281941e-05, 'epoch': 0.49}) +[2026-06-17 18:56:29,183][starVLA.training.train_starvla][INFO] - Step 1336, Loss: {'train/loss': 0.9309024214744568, 'train/grad_norm_pre_clip': 11.15234375, 'timing/data': 0.030868541449308395, 'timing/model': 1.1176022067666054, 'learning_rate/qwen_vl_interface': 9.267495637842859e-06, 'learning_rate/action_model': 9.267495637842857e-05, 'epoch': 0.49}) +[2026-06-17 18:56:33,483][starVLA.training.train_starvla][INFO] - Step 1337, Loss: {'train/loss': 0.7386569380760193, 'train/grad_norm_pre_clip': 12.12341594696045, 'timing/data': 0.030737757682800293, 'timing/model': 1.2990952786058187, 'learning_rate/qwen_vl_interface': 9.266341386572607e-06, 'learning_rate/action_model': 9.266341386572605e-05, 'epoch': 0.49}) +[2026-06-17 18:56:37,591][starVLA.training.train_starvla][INFO] - Step 1338, Loss: {'train/loss': 0.49461179971694946, 'train/grad_norm_pre_clip': 12.65359115600586, 'timing/data': 0.032391210086643696, 'timing/model': 1.0220096074044704, 'learning_rate/qwen_vl_interface': 9.265186302710465e-06, 'learning_rate/action_model': 9.265186302710465e-05, 'epoch': 0.49}) +[2026-06-17 18:56:42,345][starVLA.training.train_starvla][INFO] - Step 1339, Loss: {'train/loss': 0.9094140529632568, 'train/grad_norm_pre_clip': 13.29706859588623, 'timing/data': 0.03357633389532566, 'timing/model': 1.1703869793564081, 'learning_rate/qwen_vl_interface': 9.26403038649588e-06, 'learning_rate/action_model': 9.26403038649588e-05, 'epoch': 0.49}) +[2026-06-17 18:56:46,656][starVLA.training.train_starvla][INFO] - Step 1340, Loss: {'train/loss': 0.8423495292663574, 'train/grad_norm_pre_clip': 13.109501838684082, 'timing/data': 0.030094177462160587, 'timing/model': 1.0217367894947529, 'learning_rate/qwen_vl_interface': 9.262873638168478e-06, 'learning_rate/action_model': 9.262873638168478e-05, 'epoch': 0.49}) +[2026-06-17 18:56:51,210][starVLA.training.train_starvla][INFO] - Step 1341, Loss: {'train/loss': 0.7036483287811279, 'train/grad_norm_pre_clip': 12.310809135437012, 'timing/data': 0.02911705058068037, 'timing/model': 1.4461015164852142, 'learning_rate/qwen_vl_interface': 9.261716057968054e-06, 'learning_rate/action_model': 9.261716057968053e-05, 'epoch': 0.49}) +[2026-06-17 18:56:59,606][starVLA.training.train_starvla][INFO] - Step 1342, Loss: {'train/loss': 0.793205201625824, 'train/grad_norm_pre_clip': 10.927213668823242, 'timing/data': 0.03431448247283697, 'timing/model': 1.0661502806469798, 'learning_rate/qwen_vl_interface': 9.260557646134572e-06, 'learning_rate/action_model': 9.260557646134572e-05, 'epoch': 0.49}) +[2026-06-17 18:57:04,072][starVLA.training.train_starvla][INFO] - Step 1343, Loss: {'train/loss': 1.213815450668335, 'train/grad_norm_pre_clip': 17.89804458618164, 'timing/data': 0.03164200484752655, 'timing/model': 1.0874660015106201, 'learning_rate/qwen_vl_interface': 9.259398402908175e-06, 'learning_rate/action_model': 9.259398402908174e-05, 'epoch': 0.49}) +[2026-06-17 18:57:08,442][starVLA.training.train_starvla][INFO] - Step 1344, Loss: {'train/loss': 0.8704582452774048, 'train/grad_norm_pre_clip': 12.951944351196289, 'timing/data': 0.034884422086179256, 'timing/model': 1.0710557624697685, 'learning_rate/qwen_vl_interface': 9.258238328529177e-06, 'learning_rate/action_model': 9.258238328529176e-05, 'epoch': 0.49}) +[2026-06-17 18:57:12,876][starVLA.training.train_starvla][INFO] - Step 1345, Loss: {'train/loss': 0.5534888505935669, 'train/grad_norm_pre_clip': 12.577394485473633, 'timing/data': 0.02853390946984291, 'timing/model': 1.0245317621156573, 'learning_rate/qwen_vl_interface': 9.257077423238058e-06, 'learning_rate/action_model': 9.257077423238058e-05, 'epoch': 0.49}) +[2026-06-17 18:57:17,320][starVLA.training.train_starvla][INFO] - Step 1346, Loss: {'train/loss': 0.7399468421936035, 'train/grad_norm_pre_clip': 10.107165336608887, 'timing/data': 0.03154574427753687, 'timing/model': 1.3841404980048537, 'learning_rate/qwen_vl_interface': 9.25591568727548e-06, 'learning_rate/action_model': 9.255915687275478e-05, 'epoch': 0.49}) +[2026-06-17 18:57:22,018][starVLA.training.train_starvla][INFO] - Step 1347, Loss: {'train/loss': 0.6420629024505615, 'train/grad_norm_pre_clip': 10.700796127319336, 'timing/data': 0.03043826576322317, 'timing/model': 1.4903943073004484, 'learning_rate/qwen_vl_interface': 9.254753120882265e-06, 'learning_rate/action_model': 9.254753120882266e-05, 'epoch': 0.49}) +[2026-06-17 18:57:26,126][starVLA.training.train_starvla][INFO] - Step 1348, Loss: {'train/loss': 0.509917676448822, 'train/grad_norm_pre_clip': 9.572672843933105, 'timing/data': 0.034912655130028725, 'timing/model': 1.0566441472619772, 'learning_rate/qwen_vl_interface': 9.25358972429942e-06, 'learning_rate/action_model': 9.253589724299421e-05, 'epoch': 0.49}) +[2026-06-17 18:57:30,589][starVLA.training.train_starvla][INFO] - Step 1349, Loss: {'train/loss': 0.8509604334831238, 'train/grad_norm_pre_clip': 10.95561408996582, 'timing/data': 0.030284106731414795, 'timing/model': 1.0618491219356656, 'learning_rate/qwen_vl_interface': 9.25242549776812e-06, 'learning_rate/action_model': 9.252425497768119e-05, 'epoch': 0.49}) +[2026-06-17 18:57:35,095][starVLA.training.train_starvla][INFO] - Step 1350, Loss: {'train/loss': 0.6874441504478455, 'train/grad_norm_pre_clip': 11.875994682312012, 'timing/data': 0.034351774491369724, 'timing/model': 1.0339314164593816, 'learning_rate/qwen_vl_interface': 9.251260441529706e-06, 'learning_rate/action_model': 9.251260441529705e-05, 'epoch': 0.49}) +[2026-06-17 18:57:41,322][starVLA.training.train_starvla][INFO] - Step 1351, Loss: {'train/loss': 0.9312986731529236, 'train/grad_norm_pre_clip': 12.93307876586914, 'timing/data': 0.0346783772110939, 'timing/model': 1.04023855086416, 'learning_rate/qwen_vl_interface': 9.250094555825697e-06, 'learning_rate/action_model': 9.250094555825696e-05, 'epoch': 0.49}) +[2026-06-17 18:57:45,588][starVLA.training.train_starvla][INFO] - Step 1352, Loss: {'train/loss': 1.201539397239685, 'train/grad_norm_pre_clip': 17.568689346313477, 'timing/data': 0.028573814779520035, 'timing/model': 1.3098482144996524, 'learning_rate/qwen_vl_interface': 9.248927840897782e-06, 'learning_rate/action_model': 9.248927840897781e-05, 'epoch': 0.5}) +[2026-06-17 18:57:49,879][starVLA.training.train_starvla][INFO] - Step 1353, Loss: {'train/loss': 0.9138703942298889, 'train/grad_norm_pre_clip': 17.80723762512207, 'timing/data': 0.030279180966317654, 'timing/model': 1.0434092823415995, 'learning_rate/qwen_vl_interface': 9.247760296987822e-06, 'learning_rate/action_model': 9.247760296987821e-05, 'epoch': 0.5}) +[2026-06-17 18:57:54,307][starVLA.training.train_starvla][INFO] - Step 1354, Loss: {'train/loss': 0.644149661064148, 'train/grad_norm_pre_clip': 13.170084953308105, 'timing/data': 0.03230288624763489, 'timing/model': 1.0567347928881645, 'learning_rate/qwen_vl_interface': 9.246591924337852e-06, 'learning_rate/action_model': 9.246591924337852e-05, 'epoch': 0.5}) +[2026-06-17 18:57:58,770][starVLA.training.train_starvla][INFO] - Step 1355, Loss: {'train/loss': 0.8496648669242859, 'train/grad_norm_pre_clip': 14.702723503112793, 'timing/data': 0.03169854357838631, 'timing/model': 1.108985218219459, 'learning_rate/qwen_vl_interface': 9.245422723190076e-06, 'learning_rate/action_model': 9.245422723190074e-05, 'epoch': 0.5}) +[2026-06-17 18:58:03,198][starVLA.training.train_starvla][INFO] - Step 1356, Loss: {'train/loss': 0.7586786150932312, 'train/grad_norm_pre_clip': 11.136033058166504, 'timing/data': 0.03179758694022894, 'timing/model': 1.0536920949816704, 'learning_rate/qwen_vl_interface': 9.244252693786868e-06, 'learning_rate/action_model': 9.244252693786869e-05, 'epoch': 0.5}) +[2026-06-17 18:58:07,631][starVLA.training.train_starvla][INFO] - Step 1357, Loss: {'train/loss': 0.6429435610771179, 'train/grad_norm_pre_clip': 12.601375579833984, 'timing/data': 0.02789068967103958, 'timing/model': 1.4569868026301265, 'learning_rate/qwen_vl_interface': 9.24308183637078e-06, 'learning_rate/action_model': 9.24308183637078e-05, 'epoch': 0.5}) +[2026-06-17 18:58:11,807][starVLA.training.train_starvla][INFO] - Step 1358, Loss: {'train/loss': 0.6846776604652405, 'train/grad_norm_pre_clip': 9.46191692352295, 'timing/data': 0.0370703237131238, 'timing/model': 1.0985273607075214, 'learning_rate/qwen_vl_interface': 9.24191015118453e-06, 'learning_rate/action_model': 9.24191015118453e-05, 'epoch': 0.5}) +[2026-06-17 18:58:16,288][starVLA.training.train_starvla][INFO] - Step 1359, Loss: {'train/loss': 0.7334972620010376, 'train/grad_norm_pre_clip': 12.933324813842773, 'timing/data': 0.05430828407406807, 'timing/model': 1.0897411927580833, 'learning_rate/qwen_vl_interface': 9.24073763847101e-06, 'learning_rate/action_model': 9.24073763847101e-05, 'epoch': 0.5}) +[2026-06-17 18:58:20,579][starVLA.training.train_starvla][INFO] - Step 1360, Loss: {'train/loss': 0.8276509046554565, 'train/grad_norm_pre_clip': 10.797972679138184, 'timing/data': 0.03396415151655674, 'timing/model': 1.0356853771954775, 'learning_rate/qwen_vl_interface': 9.239564298473284e-06, 'learning_rate/action_model': 9.239564298473284e-05, 'epoch': 0.5}) +[2026-06-17 18:58:25,001][starVLA.training.train_starvla][INFO] - Step 1361, Loss: {'train/loss': 0.8540713787078857, 'train/grad_norm_pre_clip': 12.021079063415527, 'timing/data': 0.03096150793135166, 'timing/model': 1.055851778946817, 'learning_rate/qwen_vl_interface': 9.238390131434586e-06, 'learning_rate/action_model': 9.238390131434586e-05, 'epoch': 0.5}) +[2026-06-17 18:58:29,722][starVLA.training.train_starvla][INFO] - Step 1362, Loss: {'train/loss': 0.7646099328994751, 'train/grad_norm_pre_clip': 12.756381034851074, 'timing/data': 0.03212716802954674, 'timing/model': 1.530098456889391, 'learning_rate/qwen_vl_interface': 9.237215137598322e-06, 'learning_rate/action_model': 9.237215137598322e-05, 'epoch': 0.5}) +[2026-06-17 18:58:36,006][starVLA.training.train_starvla][INFO] - Step 1363, Loss: {'train/loss': 0.505821943283081, 'train/grad_norm_pre_clip': 12.33495044708252, 'timing/data': 0.04088359512388706, 'timing/model': 1.0909820441156626, 'learning_rate/qwen_vl_interface': 9.236039317208068e-06, 'learning_rate/action_model': 9.236039317208068e-05, 'epoch': 0.5}) +[2026-06-17 18:58:42,300][starVLA.training.train_starvla][INFO] - Step 1364, Loss: {'train/loss': 0.932568371295929, 'train/grad_norm_pre_clip': 13.336316108703613, 'timing/data': 0.0328134847804904, 'timing/model': 1.106575888581574, 'learning_rate/qwen_vl_interface': 9.234862670507577e-06, 'learning_rate/action_model': 9.234862670507576e-05, 'epoch': 0.5}) +[2026-06-17 18:58:46,805][starVLA.training.train_starvla][INFO] - Step 1365, Loss: {'train/loss': 0.8746949434280396, 'train/grad_norm_pre_clip': 14.828598022460938, 'timing/data': 0.034246714785695076, 'timing/model': 1.0657425625249743, 'learning_rate/qwen_vl_interface': 9.233685197740765e-06, 'learning_rate/action_model': 9.233685197740764e-05, 'epoch': 0.5}) +[2026-06-17 18:58:51,455][starVLA.training.train_starvla][INFO] - Step 1366, Loss: {'train/loss': 0.8587326407432556, 'train/grad_norm_pre_clip': 26.8243408203125, 'timing/data': 0.0399877829477191, 'timing/model': 1.0476317927241325, 'learning_rate/qwen_vl_interface': 9.232506899151725e-06, 'learning_rate/action_model': 9.232506899151725e-05, 'epoch': 0.5}) +[2026-06-17 18:58:57,798][starVLA.training.train_starvla][INFO] - Step 1367, Loss: {'train/loss': 0.7715703248977661, 'train/grad_norm_pre_clip': 10.064345359802246, 'timing/data': 0.032883268781006336, 'timing/model': 1.4514525262638927, 'learning_rate/qwen_vl_interface': 9.23132777498472e-06, 'learning_rate/action_model': 9.231327774984719e-05, 'epoch': 0.5}) +[2026-06-17 18:59:01,915][starVLA.training.train_starvla][INFO] - Step 1368, Loss: {'train/loss': 1.1249325275421143, 'train/grad_norm_pre_clip': 13.410038948059082, 'timing/data': 0.034030708484351635, 'timing/model': 1.0651902565732598, 'learning_rate/qwen_vl_interface': 9.230147825484182e-06, 'learning_rate/action_model': 9.230147825484183e-05, 'epoch': 0.5}) +[2026-06-17 18:59:06,543][starVLA.training.train_starvla][INFO] - Step 1369, Loss: {'train/loss': 0.7351748943328857, 'train/grad_norm_pre_clip': 13.698771476745605, 'timing/data': 0.04163663741201162, 'timing/model': 1.1980469692498446, 'learning_rate/qwen_vl_interface': 9.228967050894717e-06, 'learning_rate/action_model': 9.228967050894716e-05, 'epoch': 0.5}) +[2026-06-17 18:59:12,764][starVLA.training.train_starvla][INFO] - Step 1370, Loss: {'train/loss': 0.8884472250938416, 'train/grad_norm_pre_clip': 16.460180282592773, 'timing/data': 0.030062773264944553, 'timing/model': 1.0716592129319906, 'learning_rate/qwen_vl_interface': 9.227785451461103e-06, 'learning_rate/action_model': 9.227785451461102e-05, 'epoch': 0.5}) +[2026-06-17 18:59:17,093][starVLA.training.train_starvla][INFO] - Step 1371, Loss: {'train/loss': 1.043670415878296, 'train/grad_norm_pre_clip': 40.98452377319336, 'timing/data': 0.029084070585668087, 'timing/model': 1.0271135615184903, 'learning_rate/qwen_vl_interface': 9.226603027428281e-06, 'learning_rate/action_model': 9.226603027428281e-05, 'epoch': 0.5}) +[2026-06-17 18:59:21,692][starVLA.training.train_starvla][INFO] - Step 1372, Loss: {'train/loss': 0.6831846833229065, 'train/grad_norm_pre_clip': 15.487289428710938, 'timing/data': 0.03269463591277599, 'timing/model': 1.4032046580687165, 'learning_rate/qwen_vl_interface': 9.225419779041376e-06, 'learning_rate/action_model': 9.225419779041376e-05, 'epoch': 0.5}) +[2026-06-17 18:59:25,748][starVLA.training.train_starvla][INFO] - Step 1373, Loss: {'train/loss': 1.1159032583236694, 'train/grad_norm_pre_clip': 14.781580924987793, 'timing/data': 0.03742889128625393, 'timing/model': 1.0388508941978216, 'learning_rate/qwen_vl_interface': 9.224235706545672e-06, 'learning_rate/action_model': 9.224235706545671e-05, 'epoch': 0.5}) +[2026-06-17 18:59:31,146][starVLA.training.train_starvla][INFO] - Step 1374, Loss: {'train/loss': 0.6486409902572632, 'train/grad_norm_pre_clip': 24.454570770263672, 'timing/data': 0.03885449469089508, 'timing/model': 1.0322352619841695, 'learning_rate/qwen_vl_interface': 9.22305081018663e-06, 'learning_rate/action_model': 9.223050810186629e-05, 'epoch': 0.5}) +[2026-06-17 18:59:35,504][starVLA.training.train_starvla][INFO] - Step 1375, Loss: {'train/loss': 1.033170223236084, 'train/grad_norm_pre_clip': 27.579631805419922, 'timing/data': 0.03325275331735611, 'timing/model': 1.0546249141916633, 'learning_rate/qwen_vl_interface': 9.221865090209878e-06, 'learning_rate/action_model': 9.221865090209878e-05, 'epoch': 0.5}) +[2026-06-17 18:59:40,158][starVLA.training.train_starvla][INFO] - Step 1376, Loss: {'train/loss': 0.8506433963775635, 'train/grad_norm_pre_clip': 15.189522743225098, 'timing/data': 0.029995991848409176, 'timing/model': 1.137936177663505, 'learning_rate/qwen_vl_interface': 9.22067854686122e-06, 'learning_rate/action_model': 9.22067854686122e-05, 'epoch': 0.5}) +[2026-06-17 18:59:44,584][starVLA.training.train_starvla][INFO] - Step 1377, Loss: {'train/loss': 0.8153848648071289, 'train/grad_norm_pre_clip': 11.331587791442871, 'timing/data': 0.031319670379161835, 'timing/model': 1.3773204302415252, 'learning_rate/qwen_vl_interface': 9.219491180386626e-06, 'learning_rate/action_model': 9.219491180386626e-05, 'epoch': 0.5}) +[2026-06-17 18:59:48,691][starVLA.training.train_starvla][INFO] - Step 1378, Loss: {'train/loss': 0.6013140678405762, 'train/grad_norm_pre_clip': 8.295807838439941, 'timing/data': 0.034351103939116, 'timing/model': 1.050250917673111, 'learning_rate/qwen_vl_interface': 9.21830299103224e-06, 'learning_rate/action_model': 9.21830299103224e-05, 'epoch': 0.5}) +[2026-06-17 18:59:53,460][starVLA.training.train_starvla][INFO] - Step 1379, Loss: {'train/loss': 0.7645726203918457, 'train/grad_norm_pre_clip': 11.905899047851562, 'timing/data': 0.030246312730014324, 'timing/model': 1.1913908813148737, 'learning_rate/qwen_vl_interface': 9.217113979044373e-06, 'learning_rate/action_model': 9.217113979044373e-05, 'epoch': 0.5}) +[2026-06-17 18:59:57,882][starVLA.training.train_starvla][INFO] - Step 1380, Loss: {'train/loss': 1.0194379091262817, 'train/grad_norm_pre_clip': 13.987650871276855, 'timing/data': 0.03574559651315212, 'timing/model': 1.0934764528647065, 'learning_rate/qwen_vl_interface': 9.215924144669509e-06, 'learning_rate/action_model': 9.215924144669509e-05, 'epoch': 0.51}) +[2026-06-17 19:00:02,136][starVLA.training.train_starvla][INFO] - Step 1381, Loss: {'train/loss': 0.5658153891563416, 'train/grad_norm_pre_clip': 10.171656608581543, 'timing/data': 0.031555394642055035, 'timing/model': 1.014129894785583, 'learning_rate/qwen_vl_interface': 9.214733488154304e-06, 'learning_rate/action_model': 9.214733488154304e-05, 'epoch': 0.51}) +[2026-06-17 19:00:06,272][starVLA.training.train_starvla][INFO] - Step 1382, Loss: {'train/loss': 1.163071870803833, 'train/grad_norm_pre_clip': 17.37321662902832, 'timing/data': 0.03181572165340185, 'timing/model': 1.2249934989959002, 'learning_rate/qwen_vl_interface': 9.213542009745579e-06, 'learning_rate/action_model': 9.213542009745578e-05, 'epoch': 0.51}) +[2026-06-17 19:00:10,226][starVLA.training.train_starvla][INFO] - Step 1383, Loss: {'train/loss': 0.6789734959602356, 'train/grad_norm_pre_clip': 11.131339073181152, 'timing/data': 0.028294537216424942, 'timing/model': 1.0118254916742444, 'learning_rate/qwen_vl_interface': 9.21234970969033e-06, 'learning_rate/action_model': 9.21234970969033e-05, 'epoch': 0.51}) +[2026-06-17 19:00:14,434][starVLA.training.train_starvla][INFO] - Step 1384, Loss: {'train/loss': 0.9602515697479248, 'train/grad_norm_pre_clip': 15.642549514770508, 'timing/data': 0.02536530327051878, 'timing/model': 1.0088659236207604, 'learning_rate/qwen_vl_interface': 9.211156588235724e-06, 'learning_rate/action_model': 9.211156588235725e-05, 'epoch': 0.51}) +[2026-06-17 19:00:18,703][starVLA.training.train_starvla][INFO] - Step 1385, Loss: {'train/loss': 1.1088863611221313, 'train/grad_norm_pre_clip': 12.86655330657959, 'timing/data': 0.02640835102647543, 'timing/model': 1.0175897171720862, 'learning_rate/qwen_vl_interface': 9.209962645629095e-06, 'learning_rate/action_model': 9.209962645629095e-05, 'epoch': 0.51}) +[2026-06-17 19:00:22,890][starVLA.training.train_starvla][INFO] - Step 1386, Loss: {'train/loss': 0.854997992515564, 'train/grad_norm_pre_clip': 20.987594604492188, 'timing/data': 0.028013364411890507, 'timing/model': 1.2872249381616712, 'learning_rate/qwen_vl_interface': 9.208767882117947e-06, 'learning_rate/action_model': 9.208767882117947e-05, 'epoch': 0.51}) +[2026-06-17 19:00:26,988][starVLA.training.train_starvla][INFO] - Step 1387, Loss: {'train/loss': 0.7521778345108032, 'train/grad_norm_pre_clip': 10.742197036743164, 'timing/data': 0.028562480583786964, 'timing/model': 1.1782734580338001, 'learning_rate/qwen_vl_interface': 9.207572297949958e-06, 'learning_rate/action_model': 9.207572297949958e-05, 'epoch': 0.51}) +[2026-06-17 19:00:31,161][starVLA.training.train_starvla][INFO] - Step 1388, Loss: {'train/loss': 0.5616086721420288, 'train/grad_norm_pre_clip': 8.788656234741211, 'timing/data': 0.02738228626549244, 'timing/model': 1.0128497490659356, 'learning_rate/qwen_vl_interface': 9.206375893372973e-06, 'learning_rate/action_model': 9.206375893372973e-05, 'epoch': 0.51}) +[2026-06-17 19:00:35,347][starVLA.training.train_starvla][INFO] - Step 1389, Loss: {'train/loss': 1.0420194864273071, 'train/grad_norm_pre_clip': 13.081075668334961, 'timing/data': 0.03039263840764761, 'timing/model': 1.0221853470429778, 'learning_rate/qwen_vl_interface': 9.20517866863501e-06, 'learning_rate/action_model': 9.205178668635009e-05, 'epoch': 0.51}) +[2026-06-17 19:00:39,560][starVLA.training.train_starvla][INFO] - Step 1390, Loss: {'train/loss': 1.2601171731948853, 'train/grad_norm_pre_clip': 13.872231483459473, 'timing/data': 0.02597868163138628, 'timing/model': 1.0545644639059901, 'learning_rate/qwen_vl_interface': 9.20398062398425e-06, 'learning_rate/action_model': 9.20398062398425e-05, 'epoch': 0.51}) +[2026-06-17 19:00:43,763][starVLA.training.train_starvla][INFO] - Step 1391, Loss: {'train/loss': 0.9798772931098938, 'train/grad_norm_pre_clip': 11.852460861206055, 'timing/data': 0.026682550087571144, 'timing/model': 1.2761327689513564, 'learning_rate/qwen_vl_interface': 9.202781759669053e-06, 'learning_rate/action_model': 9.202781759669052e-05, 'epoch': 0.51}) +[2026-06-17 19:00:47,921][starVLA.training.train_starvla][INFO] - Step 1392, Loss: {'train/loss': 0.50921630859375, 'train/grad_norm_pre_clip': 9.03744125366211, 'timing/data': 0.026632838882505894, 'timing/model': 1.2323330854997039, 'learning_rate/qwen_vl_interface': 9.201582075937943e-06, 'learning_rate/action_model': 9.201582075937944e-05, 'epoch': 0.51}) +[2026-06-17 19:00:51,885][starVLA.training.train_starvla][INFO] - Step 1393, Loss: {'train/loss': 0.6576969623565674, 'train/grad_norm_pre_clip': 13.422738075256348, 'timing/data': 0.028893964365124702, 'timing/model': 1.0216664858162403, 'learning_rate/qwen_vl_interface': 9.200381573039619e-06, 'learning_rate/action_model': 9.200381573039619e-05, 'epoch': 0.51}) +[2026-06-17 19:00:56,343][starVLA.training.train_starvla][INFO] - Step 1394, Loss: {'train/loss': 0.8328812122344971, 'train/grad_norm_pre_clip': 12.850776672363281, 'timing/data': 0.035444969311356544, 'timing/model': 1.1373096872121096, 'learning_rate/qwen_vl_interface': 9.199180251222942e-06, 'learning_rate/action_model': 9.199180251222942e-05, 'epoch': 0.51}) +[2026-06-17 19:01:00,689][starVLA.training.train_starvla][INFO] - Step 1395, Loss: {'train/loss': 0.682577908039093, 'train/grad_norm_pre_clip': 10.66959285736084, 'timing/data': 0.033549677580595016, 'timing/model': 1.0506886951625347, 'learning_rate/qwen_vl_interface': 9.197978110736951e-06, 'learning_rate/action_model': 9.19797811073695e-05, 'epoch': 0.51}) +[2026-06-17 19:01:05,032][starVLA.training.train_starvla][INFO] - Step 1396, Loss: {'train/loss': 0.7550933957099915, 'train/grad_norm_pre_clip': 11.122956275939941, 'timing/data': 0.03298918064683676, 'timing/model': 1.0345400059595704, 'learning_rate/qwen_vl_interface': 9.196775151830849e-06, 'learning_rate/action_model': 9.196775151830849e-05, 'epoch': 0.51}) +[2026-06-17 19:01:09,439][starVLA.training.train_starvla][INFO] - Step 1397, Loss: {'train/loss': 0.949864387512207, 'train/grad_norm_pre_clip': 19.227886199951172, 'timing/data': 0.03215007949620485, 'timing/model': 1.3462706916034222, 'learning_rate/qwen_vl_interface': 9.19557137475401e-06, 'learning_rate/action_model': 9.19557137475401e-05, 'epoch': 0.51}) +[2026-06-17 19:01:13,612][starVLA.training.train_starvla][INFO] - Step 1398, Loss: {'train/loss': 0.9826037287712097, 'train/grad_norm_pre_clip': 11.783202171325684, 'timing/data': 0.031030213460326195, 'timing/model': 1.0568106696009636, 'learning_rate/qwen_vl_interface': 9.19436677975598e-06, 'learning_rate/action_model': 9.19436677975598e-05, 'epoch': 0.51}) +[2026-06-17 19:01:17,956][starVLA.training.train_starvla][INFO] - Step 1399, Loss: {'train/loss': 0.7869167327880859, 'train/grad_norm_pre_clip': 9.528254508972168, 'timing/data': 0.03341388329863548, 'timing/model': 1.022840915247798, 'learning_rate/qwen_vl_interface': 9.193161367086472e-06, 'learning_rate/action_model': 9.193161367086472e-05, 'epoch': 0.51}) +[2026-06-17 19:03:51,457][starVLA.training.train_starvla][INFO] - Step 1400, Loss: {'train/loss': 0.9490264058113098, 'train/grad_norm_pre_clip': 11.914741516113281, 'eval/loss': 0.9013091921806335, 'eval/latency_2/loss': 0.9013091921806335, 'eval/demon_fire_precision': 0.8084291187739464, 'eval/demon_fire_recall': 0.9711297071129708, 'eval/demon_fire_f1': 0.8823417601216499, 'eval/demon_fire_rate_model': 0.7988313856427379, 'eval/demon_fire_rate_teacher': 0.6649972175848636, 'eval/demon_left_precision': 0.7851335656213705, 'eval/demon_left_recall': 0.7115789473684211, 'eval/demon_left_f1': 0.7465488680287133, 'eval/demon_left_rate_model': 0.2395659432387312, 'eval/demon_left_rate_teacher': 0.2643294379521425, 'eval/demon_right_precision': 0.8263565891472868, 'eval/demon_right_recall': 0.6022598870056497, 'eval/demon_right_f1': 0.6967320261437908, 'eval/demon_right_rate_model': 0.17946577629382304, 'eval/demon_right_rate_teacher': 0.24624373956594323, 'eval/demon_move_f1': 0.7216404470862521, 'eval/demon_cc_f1': 0.8270868437742547, 'eval/latency_2/demon_fire_precision': 0.8084291187739464, 'eval/latency_2/demon_fire_recall': 0.9711297071129708, 'eval/latency_2/demon_fire_f1': 0.8823417601216499, 'eval/latency_2/demon_fire_rate_model': 0.7988313856427379, 'eval/latency_2/demon_fire_rate_teacher': 0.6649972175848636, 'eval/latency_2/demon_left_precision': 0.7851335656213705, 'eval/latency_2/demon_left_recall': 0.7115789473684211, 'eval/latency_2/demon_left_f1': 0.7465488680287133, 'eval/latency_2/demon_left_rate_model': 0.2395659432387312, 'eval/latency_2/demon_left_rate_teacher': 0.2643294379521425, 'eval/latency_2/demon_right_precision': 0.8263565891472868, 'eval/latency_2/demon_right_recall': 0.6022598870056497, 'eval/latency_2/demon_right_f1': 0.6967320261437908, 'eval/latency_2/demon_right_rate_model': 0.17946577629382304, 'eval/latency_2/demon_right_rate_teacher': 0.24624373956594323, 'eval/latency_2/demon_move_f1': 0.7216404470862521, 'eval/latency_2/demon_cc_f1': 0.8270868437742547, 'timing/data': 0.0383252901956439, 'timing/model': 1.0408819736912847, 'learning_rate/qwen_vl_interface': 9.191955136995371e-06, 'learning_rate/action_model': 9.191955136995371e-05, 'epoch': 0.51}) +[2026-06-17 19:03:51,458][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_1400_state +[2026-06-17 19:03:51,530][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 19:04:00,135][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_1400_state/model.safetensors +[2026-06-17 19:04:14,219][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_1400_state/optimizer.bin +[2026-06-17 19:04:14,220][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_1400_state/sampler.bin +[2026-06-17 19:04:14,221][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_1400_state/sampler_1.bin +[2026-06-17 19:04:14,223][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_1400_state/random_states_0.pkl +[2026-06-17 19:04:14,227][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1400 +[2026-06-17 19:04:14,228][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 19:04:14,248][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 19:04:18,794][starVLA.training.train_starvla][INFO] - Step 1401, Loss: {'train/loss': 0.6821375489234924, 'train/grad_norm_pre_clip': 13.925751686096191, 'timing/data': 0.032249702140688896, 'timing/model': 1.0362958880141377, 'learning_rate/qwen_vl_interface': 9.190748089732727e-06, 'learning_rate/action_model': 9.190748089732728e-05, 'epoch': 0.51}) +[2026-06-17 19:04:23,281][starVLA.training.train_starvla][INFO] - Step 1402, Loss: {'train/loss': 0.8326992392539978, 'train/grad_norm_pre_clip': 10.894451141357422, 'timing/data': 0.03134012594819069, 'timing/model': 1.10337835829705, 'learning_rate/qwen_vl_interface': 9.189540225548763e-06, 'learning_rate/action_model': 9.189540225548764e-05, 'epoch': 0.51}) +[2026-06-17 19:04:27,600][starVLA.training.train_starvla][INFO] - Step 1403, Loss: {'train/loss': 0.9118896126747131, 'train/grad_norm_pre_clip': 10.49122428894043, 'timing/data': 0.02908241841942072, 'timing/model': 1.347243675030768, 'learning_rate/qwen_vl_interface': 9.188331544693872e-06, 'learning_rate/action_model': 9.188331544693873e-05, 'epoch': 0.51}) +[2026-06-17 19:04:32,828][starVLA.training.train_starvla][INFO] - Step 1404, Loss: {'train/loss': 0.9182208776473999, 'train/grad_norm_pre_clip': 10.605646133422852, 'timing/data': 0.030510366894304752, 'timing/model': 1.0813126992434263, 'learning_rate/qwen_vl_interface': 9.187122047418614e-06, 'learning_rate/action_model': 9.187122047418614e-05, 'epoch': 0.51}) +[2026-06-17 19:04:37,332][starVLA.training.train_starvla][INFO] - Step 1405, Loss: {'train/loss': 0.7996065616607666, 'train/grad_norm_pre_clip': 12.077476501464844, 'timing/data': 0.029144540429115295, 'timing/model': 1.0447809305042028, 'learning_rate/qwen_vl_interface': 9.18591173397372e-06, 'learning_rate/action_model': 9.185911733973719e-05, 'epoch': 0.51}) +[2026-06-17 19:04:41,659][starVLA.training.train_starvla][INFO] - Step 1406, Loss: {'train/loss': 0.6867218613624573, 'train/grad_norm_pre_clip': 11.415421485900879, 'timing/data': 0.03489313740283251, 'timing/model': 1.0240659266710281, 'learning_rate/qwen_vl_interface': 9.184700604610087e-06, 'learning_rate/action_model': 9.184700604610085e-05, 'epoch': 0.51}) +[2026-06-17 19:04:46,975][starVLA.training.train_starvla][INFO] - Step 1407, Loss: {'train/loss': 1.0219069719314575, 'train/grad_norm_pre_clip': 11.645164489746094, 'timing/data': 0.029370096512138844, 'timing/model': 2.0627004466950893, 'learning_rate/qwen_vl_interface': 9.183488659578786e-06, 'learning_rate/action_model': 9.183488659578785e-05, 'epoch': 0.52}) +[2026-06-17 19:04:51,425][starVLA.training.train_starvla][INFO] - Step 1408, Loss: {'train/loss': 0.9257204532623291, 'train/grad_norm_pre_clip': 12.337970733642578, 'timing/data': 0.032645056024193764, 'timing/model': 1.3470069197937846, 'learning_rate/qwen_vl_interface': 9.18227589913105e-06, 'learning_rate/action_model': 9.18227589913105e-05, 'epoch': 0.52}) +[2026-06-17 19:04:55,532][starVLA.training.train_starvla][INFO] - Step 1409, Loss: {'train/loss': 0.8084383010864258, 'train/grad_norm_pre_clip': 9.008495330810547, 'timing/data': 0.03003445267677307, 'timing/model': 1.03767114225775, 'learning_rate/qwen_vl_interface': 9.181062323518293e-06, 'learning_rate/action_model': 9.181062323518292e-05, 'epoch': 0.52}) +[2026-06-17 19:05:02,501][starVLA.training.train_starvla][INFO] - Step 1410, Loss: {'train/loss': 0.7851336002349854, 'train/grad_norm_pre_clip': 10.600346565246582, 'timing/data': 0.031867941841483116, 'timing/model': 1.1038025990128517, 'learning_rate/qwen_vl_interface': 9.179847932992084e-06, 'learning_rate/action_model': 9.179847932992083e-05, 'epoch': 0.52}) +[2026-06-17 19:05:06,863][starVLA.training.train_starvla][INFO] - Step 1411, Loss: {'train/loss': 0.8568106293678284, 'train/grad_norm_pre_clip': 9.865581512451172, 'timing/data': 0.030597454868257046, 'timing/model': 1.026070736348629, 'learning_rate/qwen_vl_interface': 9.17863272780417e-06, 'learning_rate/action_model': 9.178632727804168e-05, 'epoch': 0.52}) +[2026-06-17 19:05:13,070][starVLA.training.train_starvla][INFO] - Step 1412, Loss: {'train/loss': 0.7074083089828491, 'train/grad_norm_pre_clip': 9.411057472229004, 'timing/data': 0.03126817476004362, 'timing/model': 1.234096072614193, 'learning_rate/qwen_vl_interface': 9.177416708206463e-06, 'learning_rate/action_model': 9.177416708206463e-05, 'epoch': 0.52}) +[2026-06-17 19:05:19,467][starVLA.training.train_starvla][INFO] - Step 1413, Loss: {'train/loss': 0.9123409390449524, 'train/grad_norm_pre_clip': 11.88646411895752, 'timing/data': 0.032219658605754375, 'timing/model': 1.3832476809620857, 'learning_rate/qwen_vl_interface': 9.176199874451049e-06, 'learning_rate/action_model': 9.176199874451048e-05, 'epoch': 0.52}) +[2026-06-17 19:05:23,465][starVLA.training.train_starvla][INFO] - Step 1414, Loss: {'train/loss': 0.8354833126068115, 'train/grad_norm_pre_clip': 13.725031852722168, 'timing/data': 0.031131897121667862, 'timing/model': 1.0301944762468338, 'learning_rate/qwen_vl_interface': 9.174982226790175e-06, 'learning_rate/action_model': 9.174982226790175e-05, 'epoch': 0.52}) +[2026-06-17 19:05:29,593][starVLA.training.train_starvla][INFO] - Step 1415, Loss: {'train/loss': 1.0830894708633423, 'train/grad_norm_pre_clip': 12.628706932067871, 'timing/data': 0.03050308022648096, 'timing/model': 1.032714694738388, 'learning_rate/qwen_vl_interface': 9.173763765476262e-06, 'learning_rate/action_model': 9.17376376547626e-05, 'epoch': 0.52}) +[2026-06-17 19:05:33,985][starVLA.training.train_starvla][INFO] - Step 1416, Loss: {'train/loss': 0.9712753295898438, 'train/grad_norm_pre_clip': 10.71994400024414, 'timing/data': 0.030712096951901913, 'timing/model': 1.052196329459548, 'learning_rate/qwen_vl_interface': 9.172544490761897e-06, 'learning_rate/action_model': 9.172544490761897e-05, 'epoch': 0.52}) +[2026-06-17 19:05:38,507][starVLA.training.train_starvla][INFO] - Step 1417, Loss: {'train/loss': 0.810308575630188, 'train/grad_norm_pre_clip': 12.59400463104248, 'timing/data': 0.029909138567745686, 'timing/model': 1.0667703384533525, 'learning_rate/qwen_vl_interface': 9.171324402899841e-06, 'learning_rate/action_model': 9.17132440289984e-05, 'epoch': 0.52}) +[2026-06-17 19:05:44,472][starVLA.training.train_starvla][INFO] - Step 1418, Loss: {'train/loss': 0.7756257057189941, 'train/grad_norm_pre_clip': 11.921788215637207, 'timing/data': 0.03076461609452963, 'timing/model': 1.3841030606999993, 'learning_rate/qwen_vl_interface': 9.170103502143014e-06, 'learning_rate/action_model': 9.170103502143013e-05, 'epoch': 0.52}) +[2026-06-17 19:05:48,620][starVLA.training.train_starvla][INFO] - Step 1419, Loss: {'train/loss': 0.7398670315742493, 'train/grad_norm_pre_clip': 9.584405899047852, 'timing/data': 0.03066322673112154, 'timing/model': 1.0290262689813972, 'learning_rate/qwen_vl_interface': 9.168881788744513e-06, 'learning_rate/action_model': 9.168881788744513e-05, 'epoch': 0.52}) +[2026-06-17 19:05:53,182][starVLA.training.train_starvla][INFO] - Step 1420, Loss: {'train/loss': 0.7364880442619324, 'train/grad_norm_pre_clip': 11.361165046691895, 'timing/data': 0.03511285316199064, 'timing/model': 1.0424475697800517, 'learning_rate/qwen_vl_interface': 9.1676592629576e-06, 'learning_rate/action_model': 9.167659262957601e-05, 'epoch': 0.52}) +[2026-06-17 19:05:59,060][starVLA.training.train_starvla][INFO] - Step 1421, Loss: {'train/loss': 0.6809909343719482, 'train/grad_norm_pre_clip': 29.01495933532715, 'timing/data': 0.030154659412801266, 'timing/model': 1.0586115401238203, 'learning_rate/qwen_vl_interface': 9.16643592503571e-06, 'learning_rate/action_model': 9.166435925035709e-05, 'epoch': 0.52}) +[2026-06-17 19:06:03,521][starVLA.training.train_starvla][INFO] - Step 1422, Loss: {'train/loss': 0.5724897980690002, 'train/grad_norm_pre_clip': 9.449589729309082, 'timing/data': 0.03227314259856939, 'timing/model': 1.4704153658822179, 'learning_rate/qwen_vl_interface': 9.165211775232436e-06, 'learning_rate/action_model': 9.165211775232435e-05, 'epoch': 0.52}) +[2026-06-17 19:06:07,746][starVLA.training.train_starvla][INFO] - Step 1423, Loss: {'train/loss': 0.6008864045143127, 'train/grad_norm_pre_clip': 10.333343505859375, 'timing/data': 0.03577305842190981, 'timing/model': 1.1139083122834563, 'learning_rate/qwen_vl_interface': 9.163986813801548e-06, 'learning_rate/action_model': 9.163986813801548e-05, 'epoch': 0.52}) +[2026-06-17 19:06:12,216][starVLA.training.train_starvla][INFO] - Step 1424, Loss: {'train/loss': 0.7760998010635376, 'train/grad_norm_pre_clip': 11.77849006652832, 'timing/data': 0.03311455249786377, 'timing/model': 1.0407891785725951, 'learning_rate/qwen_vl_interface': 9.162761040996979e-06, 'learning_rate/action_model': 9.162761040996978e-05, 'epoch': 0.52}) +[2026-06-17 19:06:16,886][starVLA.training.train_starvla][INFO] - Step 1425, Loss: {'train/loss': 0.8065808415412903, 'train/grad_norm_pre_clip': 11.171868324279785, 'timing/data': 0.03177565522491932, 'timing/model': 1.1130248429253697, 'learning_rate/qwen_vl_interface': 9.161534457072839e-06, 'learning_rate/action_model': 9.161534457072839e-05, 'epoch': 0.52}) +[2026-06-17 19:06:22,983][starVLA.training.train_starvla][INFO] - Step 1426, Loss: {'train/loss': 0.9114906191825867, 'train/grad_norm_pre_clip': 14.17132568359375, 'timing/data': 0.037329237908124924, 'timing/model': 1.2925869198516011, 'learning_rate/qwen_vl_interface': 9.160307062283397e-06, 'learning_rate/action_model': 9.160307062283397e-05, 'epoch': 0.52}) +[2026-06-17 19:06:27,563][starVLA.training.train_starvla][INFO] - Step 1427, Loss: {'train/loss': 0.7922322750091553, 'train/grad_norm_pre_clip': 12.79690170288086, 'timing/data': 0.030218511819839478, 'timing/model': 1.4757537692785263, 'learning_rate/qwen_vl_interface': 9.159078856883093e-06, 'learning_rate/action_model': 9.159078856883093e-05, 'epoch': 0.52}) +[2026-06-17 19:06:32,027][starVLA.training.train_starvla][INFO] - Step 1428, Loss: {'train/loss': 0.7034875750541687, 'train/grad_norm_pre_clip': 12.988043785095215, 'timing/data': 0.03057398647069931, 'timing/model': 1.3311539068818092, 'learning_rate/qwen_vl_interface': 9.157849841126535e-06, 'learning_rate/action_model': 9.157849841126535e-05, 'epoch': 0.52}) +[2026-06-17 19:06:36,163][starVLA.training.train_starvla][INFO] - Step 1429, Loss: {'train/loss': 1.0400190353393555, 'train/grad_norm_pre_clip': 13.539956092834473, 'timing/data': 0.03123610746115446, 'timing/model': 1.0477683898061514, 'learning_rate/qwen_vl_interface': 9.156620015268502e-06, 'learning_rate/action_model': 9.1566200152685e-05, 'epoch': 0.52}) +[2026-06-17 19:06:40,559][starVLA.training.train_starvla][INFO] - Step 1430, Loss: {'train/loss': 0.6832180023193359, 'train/grad_norm_pre_clip': 10.26926326751709, 'timing/data': 0.03120442572981119, 'timing/model': 1.030708134174347, 'learning_rate/qwen_vl_interface': 9.155389379563934e-06, 'learning_rate/action_model': 9.155389379563933e-05, 'epoch': 0.52}) +[2026-06-17 19:06:44,993][starVLA.training.train_starvla][INFO] - Step 1431, Loss: {'train/loss': 0.6546742916107178, 'train/grad_norm_pre_clip': 10.765028953552246, 'timing/data': 0.035884411074221134, 'timing/model': 1.1321126148104668, 'learning_rate/qwen_vl_interface': 9.154157934267946e-06, 'learning_rate/action_model': 9.154157934267944e-05, 'epoch': 0.52}) +[2026-06-17 19:06:49,447][starVLA.training.train_starvla][INFO] - Step 1432, Loss: {'train/loss': 1.1964572668075562, 'train/grad_norm_pre_clip': 16.601070404052734, 'timing/data': 0.031374734826385975, 'timing/model': 1.128523570485413, 'learning_rate/qwen_vl_interface': 9.152925679635817e-06, 'learning_rate/action_model': 9.152925679635817e-05, 'epoch': 0.52}) +[2026-06-17 19:06:53,759][starVLA.training.train_starvla][INFO] - Step 1433, Loss: {'train/loss': 0.843686044216156, 'train/grad_norm_pre_clip': 15.254827499389648, 'timing/data': 0.032837074249982834, 'timing/model': 1.3112119631841779, 'learning_rate/qwen_vl_interface': 9.151692615922995e-06, 'learning_rate/action_model': 9.151692615922994e-05, 'epoch': 0.52}) +[2026-06-17 19:06:57,939][starVLA.training.train_starvla][INFO] - Step 1434, Loss: {'train/loss': 0.8868582248687744, 'train/grad_norm_pre_clip': 14.040237426757812, 'timing/data': 0.028613979928195477, 'timing/model': 1.0202636662870646, 'learning_rate/qwen_vl_interface': 9.150458743385095e-06, 'learning_rate/action_model': 9.150458743385094e-05, 'epoch': 0.53}) +[2026-06-17 19:07:02,379][starVLA.training.train_starvla][INFO] - Step 1435, Loss: {'train/loss': 0.8335379958152771, 'train/grad_norm_pre_clip': 13.73949909210205, 'timing/data': 0.029418370686471462, 'timing/model': 1.062914608977735, 'learning_rate/qwen_vl_interface': 9.1492240622779e-06, 'learning_rate/action_model': 9.1492240622779e-05, 'epoch': 0.53}) +[2026-06-17 19:07:06,930][starVLA.training.train_starvla][INFO] - Step 1436, Loss: {'train/loss': 0.7593342065811157, 'train/grad_norm_pre_clip': 13.717330932617188, 'timing/data': 0.03264834266155958, 'timing/model': 1.0432604113593698, 'learning_rate/qwen_vl_interface': 9.147988572857362e-06, 'learning_rate/action_model': 9.147988572857362e-05, 'epoch': 0.53}) +[2026-06-17 19:07:11,218][starVLA.training.train_starvla][INFO] - Step 1437, Loss: {'train/loss': 0.9476178288459778, 'train/grad_norm_pre_clip': 121.65113067626953, 'timing/data': 0.03090895153582096, 'timing/model': 1.0368683086708188, 'learning_rate/qwen_vl_interface': 9.146752275379599e-06, 'learning_rate/action_model': 9.146752275379598e-05, 'epoch': 0.53}) +[2026-06-17 19:07:15,804][starVLA.training.train_starvla][INFO] - Step 1438, Loss: {'train/loss': 0.7928213477134705, 'train/grad_norm_pre_clip': 15.689992904663086, 'timing/data': 0.033633298240602016, 'timing/model': 1.4013344319537282, 'learning_rate/qwen_vl_interface': 9.145515170100894e-06, 'learning_rate/action_model': 9.145515170100894e-05, 'epoch': 0.53}) +[2026-06-17 19:07:19,851][starVLA.training.train_starvla][INFO] - Step 1439, Loss: {'train/loss': 0.8319206237792969, 'train/grad_norm_pre_clip': 12.236283302307129, 'timing/data': 0.030881199054419994, 'timing/model': 1.04713340010494, 'learning_rate/qwen_vl_interface': 9.144277257277705e-06, 'learning_rate/action_model': 9.144277257277703e-05, 'epoch': 0.53}) +[2026-06-17 19:07:24,128][starVLA.training.train_starvla][INFO] - Step 1440, Loss: {'train/loss': 0.7082923650741577, 'train/grad_norm_pre_clip': 20.236221313476562, 'timing/data': 0.030003733932971954, 'timing/model': 1.0411800416186452, 'learning_rate/qwen_vl_interface': 9.14303853716665e-06, 'learning_rate/action_model': 9.143038537166649e-05, 'epoch': 0.53}) +[2026-06-17 19:07:28,620][starVLA.training.train_starvla][INFO] - Step 1441, Loss: {'train/loss': 0.7578786611557007, 'train/grad_norm_pre_clip': 11.612004280090332, 'timing/data': 0.03274821862578392, 'timing/model': 1.0453985165804625, 'learning_rate/qwen_vl_interface': 9.141799010024516e-06, 'learning_rate/action_model': 9.141799010024515e-05, 'epoch': 0.53}) +[2026-06-17 19:07:32,968][starVLA.training.train_starvla][INFO] - Step 1442, Loss: {'train/loss': 0.6390109062194824, 'train/grad_norm_pre_clip': 9.493945121765137, 'timing/data': 0.03243697341531515, 'timing/model': 1.0300219748169184, 'learning_rate/qwen_vl_interface': 9.140558676108258e-06, 'learning_rate/action_model': 9.140558676108259e-05, 'epoch': 0.53}) +[2026-06-17 19:07:37,409][starVLA.training.train_starvla][INFO] - Step 1443, Loss: {'train/loss': 0.6887524724006653, 'train/grad_norm_pre_clip': 10.805132865905762, 'timing/data': 0.029233059845864773, 'timing/model': 1.4148272322490811, 'learning_rate/qwen_vl_interface': 9.139317535675002e-06, 'learning_rate/action_model': 9.139317535675002e-05, 'epoch': 0.53}) +[2026-06-17 19:07:41,507][starVLA.training.train_starvla][INFO] - Step 1444, Loss: {'train/loss': 0.7269537448883057, 'train/grad_norm_pre_clip': 12.548218727111816, 'timing/data': 0.03145584091544151, 'timing/model': 1.063206348568201, 'learning_rate/qwen_vl_interface': 9.138075588982035e-06, 'learning_rate/action_model': 9.138075588982036e-05, 'epoch': 0.53}) +[2026-06-17 19:07:45,996][starVLA.training.train_starvla][INFO] - Step 1445, Loss: {'train/loss': 0.6602019667625427, 'train/grad_norm_pre_clip': 9.317337036132812, 'timing/data': 0.029925983399152756, 'timing/model': 1.0606170613318682, 'learning_rate/qwen_vl_interface': 9.136832836286816e-06, 'learning_rate/action_model': 9.136832836286816e-05, 'epoch': 0.53}) +[2026-06-17 19:07:50,487][starVLA.training.train_starvla][INFO] - Step 1446, Loss: {'train/loss': 0.6525959372520447, 'train/grad_norm_pre_clip': 11.278389930725098, 'timing/data': 0.034400083124637604, 'timing/model': 1.0995237547904253, 'learning_rate/qwen_vl_interface': 9.135589277846967e-06, 'learning_rate/action_model': 9.135589277846966e-05, 'epoch': 0.53}) +[2026-06-17 19:07:54,922][starVLA.training.train_starvla][INFO] - Step 1447, Loss: {'train/loss': 1.14168119430542, 'train/grad_norm_pre_clip': 11.56252384185791, 'timing/data': 0.03041219897568226, 'timing/model': 1.0336816562339664, 'learning_rate/qwen_vl_interface': 9.134344913920278e-06, 'learning_rate/action_model': 9.134344913920278e-05, 'epoch': 0.53}) +[2026-06-17 19:08:01,408][starVLA.training.train_starvla][INFO] - Step 1448, Loss: {'train/loss': 0.6873679161071777, 'train/grad_norm_pre_clip': 9.298929214477539, 'timing/data': 0.029484347440302372, 'timing/model': 1.33781682793051, 'learning_rate/qwen_vl_interface': 9.133099744764711e-06, 'learning_rate/action_model': 9.133099744764711e-05, 'epoch': 0.53}) +[2026-06-17 19:08:05,462][starVLA.training.train_starvla][INFO] - Step 1449, Loss: {'train/loss': 0.8260385394096375, 'train/grad_norm_pre_clip': 20.655738830566406, 'timing/data': 0.031389071606099606, 'timing/model': 1.0634322771802545, 'learning_rate/qwen_vl_interface': 9.131853770638386e-06, 'learning_rate/action_model': 9.131853770638386e-05, 'epoch': 0.53}) +[2026-06-17 19:08:09,875][starVLA.training.train_starvla][INFO] - Step 1450, Loss: {'train/loss': 0.684616208076477, 'train/grad_norm_pre_clip': 9.785805702209473, 'timing/data': 0.032167939469218254, 'timing/model': 1.127108795568347, 'learning_rate/qwen_vl_interface': 9.130606991799599e-06, 'learning_rate/action_model': 9.130606991799599e-05, 'epoch': 0.53}) +[2026-06-17 19:08:17,035][starVLA.training.train_starvla][INFO] - Step 1451, Loss: {'train/loss': 1.1226146221160889, 'train/grad_norm_pre_clip': 15.066446304321289, 'timing/data': 0.034447258338332176, 'timing/model': 1.028420384041965, 'learning_rate/qwen_vl_interface': 9.129359408506808e-06, 'learning_rate/action_model': 9.129359408506808e-05, 'epoch': 0.53}) +[2026-06-17 19:08:21,403][starVLA.training.train_starvla][INFO] - Step 1452, Loss: {'train/loss': 1.0966482162475586, 'train/grad_norm_pre_clip': 14.933083534240723, 'timing/data': 0.03317859023809433, 'timing/model': 1.0333345048129559, 'learning_rate/qwen_vl_interface': 9.128111021018636e-06, 'learning_rate/action_model': 9.128111021018636e-05, 'epoch': 0.53}) +[2026-06-17 19:08:25,983][starVLA.training.train_starvla][INFO] - Step 1453, Loss: {'train/loss': 0.7103243470191956, 'train/grad_norm_pre_clip': 10.437580108642578, 'timing/data': 0.030041822232306004, 'timing/model': 1.395472171716392, 'learning_rate/qwen_vl_interface': 9.126861829593876e-06, 'learning_rate/action_model': 9.126861829593876e-05, 'epoch': 0.53}) +[2026-06-17 19:08:30,306][starVLA.training.train_starvla][INFO] - Step 1454, Loss: {'train/loss': 0.7063582539558411, 'train/grad_norm_pre_clip': 11.060938835144043, 'timing/data': 0.07581185735762119, 'timing/model': 1.196440969593823, 'learning_rate/qwen_vl_interface': 9.125611834491487e-06, 'learning_rate/action_model': 9.125611834491486e-05, 'epoch': 0.53}) +[2026-06-17 19:08:34,773][starVLA.training.train_starvla][INFO] - Step 1455, Loss: {'train/loss': 0.9015368819236755, 'train/grad_norm_pre_clip': 9.994333267211914, 'timing/data': 0.03106766566634178, 'timing/model': 1.0404694844037294, 'learning_rate/qwen_vl_interface': 9.124361035970594e-06, 'learning_rate/action_model': 9.124361035970594e-05, 'epoch': 0.53}) +[2026-06-17 19:08:39,169][starVLA.training.train_starvla][INFO] - Step 1456, Loss: {'train/loss': 0.520781934261322, 'train/grad_norm_pre_clip': 8.706615447998047, 'timing/data': 0.030724473297595978, 'timing/model': 1.0416434528306127, 'learning_rate/qwen_vl_interface': 9.123109434290491e-06, 'learning_rate/action_model': 9.123109434290492e-05, 'epoch': 0.53}) +[2026-06-17 19:08:43,636][starVLA.training.train_starvla][INFO] - Step 1457, Loss: {'train/loss': 0.5151724815368652, 'train/grad_norm_pre_clip': 9.54293155670166, 'timing/data': 0.035471098497509956, 'timing/model': 1.0594524452462792, 'learning_rate/qwen_vl_interface': 9.121857029710632e-06, 'learning_rate/action_model': 9.121857029710633e-05, 'epoch': 0.53}) +[2026-06-17 19:08:48,047][starVLA.training.train_starvla][INFO] - Step 1458, Loss: {'train/loss': 1.1877050399780273, 'train/grad_norm_pre_clip': 13.591150283813477, 'timing/data': 0.03115720394998789, 'timing/model': 1.3236992228776217, 'learning_rate/qwen_vl_interface': 9.120603822490647e-06, 'learning_rate/action_model': 9.120603822490647e-05, 'epoch': 0.53}) +[2026-06-17 19:08:52,087][starVLA.training.train_starvla][INFO] - Step 1459, Loss: {'train/loss': 0.8575260639190674, 'train/grad_norm_pre_clip': 10.233731269836426, 'timing/data': 0.03059930633753538, 'timing/model': 1.0491550900042057, 'learning_rate/qwen_vl_interface': 9.119349812890323e-06, 'learning_rate/action_model': 9.119349812890323e-05, 'epoch': 0.53}) +[2026-06-17 19:08:56,525][starVLA.training.train_starvla][INFO] - Step 1460, Loss: {'train/loss': 0.8206380009651184, 'train/grad_norm_pre_clip': 13.533490180969238, 'timing/data': 0.039579764008522034, 'timing/model': 1.0398095957934856, 'learning_rate/qwen_vl_interface': 9.11809500116962e-06, 'learning_rate/action_model': 9.11809500116962e-05, 'epoch': 0.53}) +[2026-06-17 19:09:01,150][starVLA.training.train_starvla][INFO] - Step 1461, Loss: {'train/loss': 0.6875351667404175, 'train/grad_norm_pre_clip': 14.558802604675293, 'timing/data': 0.03243573196232319, 'timing/model': 1.025262976065278, 'learning_rate/qwen_vl_interface': 9.116839387588662e-06, 'learning_rate/action_model': 9.116839387588661e-05, 'epoch': 0.53}) +[2026-06-17 19:09:05,505][starVLA.training.train_starvla][INFO] - Step 1462, Loss: {'train/loss': 0.7154668569564819, 'train/grad_norm_pre_clip': 11.013533592224121, 'timing/data': 0.034219417721033096, 'timing/model': 1.0804299684241414, 'learning_rate/qwen_vl_interface': 9.115582972407735e-06, 'learning_rate/action_model': 9.115582972407736e-05, 'epoch': 0.54}) +[2026-06-17 19:09:09,852][starVLA.training.train_starvla][INFO] - Step 1463, Loss: {'train/loss': 1.0739128589630127, 'train/grad_norm_pre_clip': 12.401204109191895, 'timing/data': 0.036084034480154514, 'timing/model': 1.3088603029027581, 'learning_rate/qwen_vl_interface': 9.114325755887301e-06, 'learning_rate/action_model': 9.114325755887301e-05, 'epoch': 0.54}) +[2026-06-17 19:09:15,926][starVLA.training.train_starvla][INFO] - Step 1464, Loss: {'train/loss': 0.9401842951774597, 'train/grad_norm_pre_clip': 12.158440589904785, 'timing/data': 0.029451935552060604, 'timing/model': 1.030604020692408, 'learning_rate/qwen_vl_interface': 9.113067738287978e-06, 'learning_rate/action_model': 9.113067738287979e-05, 'epoch': 0.54}) +[2026-06-17 19:09:20,330][starVLA.training.train_starvla][INFO] - Step 1465, Loss: {'train/loss': 0.6788198351860046, 'train/grad_norm_pre_clip': 9.75898551940918, 'timing/data': 0.029530256055295467, 'timing/model': 1.0269350809976459, 'learning_rate/qwen_vl_interface': 9.111808919870557e-06, 'learning_rate/action_model': 9.111808919870557e-05, 'epoch': 0.54}) +[2026-06-17 19:09:24,641][starVLA.training.train_starvla][INFO] - Step 1466, Loss: {'train/loss': 0.7518202066421509, 'train/grad_norm_pre_clip': 23.41824722290039, 'timing/data': 0.030063017271459103, 'timing/model': 1.0220832712948322, 'learning_rate/qwen_vl_interface': 9.11054930089599e-06, 'learning_rate/action_model': 9.110549300895991e-05, 'epoch': 0.54}) +[2026-06-17 19:09:31,011][starVLA.training.train_starvla][INFO] - Step 1467, Loss: {'train/loss': 1.0786116123199463, 'train/grad_norm_pre_clip': 14.932477951049805, 'timing/data': 0.0348039697855711, 'timing/model': 1.4131328538060188, 'learning_rate/qwen_vl_interface': 9.109288881625402e-06, 'learning_rate/action_model': 9.109288881625402e-05, 'epoch': 0.54}) +[2026-06-17 19:09:35,176][starVLA.training.train_starvla][INFO] - Step 1468, Loss: {'train/loss': 0.7009021043777466, 'train/grad_norm_pre_clip': 11.127385139465332, 'timing/data': 0.032649542205035686, 'timing/model': 1.0331137962639332, 'learning_rate/qwen_vl_interface': 9.108027662320073e-06, 'learning_rate/action_model': 9.108027662320073e-05, 'epoch': 0.54}) +[2026-06-17 19:09:39,607][starVLA.training.train_starvla][INFO] - Step 1469, Loss: {'train/loss': 0.7029830813407898, 'train/grad_norm_pre_clip': 12.583349227905273, 'timing/data': 0.029032894410192966, 'timing/model': 1.0736741470173001, 'learning_rate/qwen_vl_interface': 9.10676564324146e-06, 'learning_rate/action_model': 9.10676564324146e-05, 'epoch': 0.54}) +[2026-06-17 19:09:43,929][starVLA.training.train_starvla][INFO] - Step 1470, Loss: {'train/loss': 1.018639087677002, 'train/grad_norm_pre_clip': 32.577980041503906, 'timing/data': 0.032034906558692455, 'timing/model': 1.0345147978514433, 'learning_rate/qwen_vl_interface': 9.105502824651179e-06, 'learning_rate/action_model': 9.105502824651178e-05, 'epoch': 0.54}) +[2026-06-17 19:09:48,435][starVLA.training.train_starvla][INFO] - Step 1471, Loss: {'train/loss': 1.2278286218643188, 'train/grad_norm_pre_clip': 15.412908554077148, 'timing/data': 0.034179736860096455, 'timing/model': 1.0219463119283319, 'learning_rate/qwen_vl_interface': 9.104239206811012e-06, 'learning_rate/action_model': 9.104239206811011e-05, 'epoch': 0.54}) +[2026-06-17 19:09:52,739][starVLA.training.train_starvla][INFO] - Step 1472, Loss: {'train/loss': 0.5557032823562622, 'train/grad_norm_pre_clip': 10.140212059020996, 'timing/data': 0.02778482437133789, 'timing/model': 1.3594678482040763, 'learning_rate/qwen_vl_interface': 9.102974789982911e-06, 'learning_rate/action_model': 9.102974789982912e-05, 'epoch': 0.54}) +[2026-06-17 19:09:57,133][starVLA.training.train_starvla][INFO] - Step 1473, Loss: {'train/loss': 1.2316932678222656, 'train/grad_norm_pre_clip': 14.326142311096191, 'timing/data': 0.03435521014034748, 'timing/model': 1.3472613785415888, 'learning_rate/qwen_vl_interface': 9.10170957442899e-06, 'learning_rate/action_model': 9.101709574428991e-05, 'epoch': 0.54}) +[2026-06-17 19:10:01,149][starVLA.training.train_starvla][INFO] - Step 1474, Loss: {'train/loss': 0.8485438823699951, 'train/grad_norm_pre_clip': 12.393241882324219, 'timing/data': 0.036040362901985645, 'timing/model': 1.039928468875587, 'learning_rate/qwen_vl_interface': 9.10044356041153e-06, 'learning_rate/action_model': 9.10044356041153e-05, 'epoch': 0.54}) +[2026-06-17 19:10:05,724][starVLA.training.train_starvla][INFO] - Step 1475, Loss: {'train/loss': 0.8056230545043945, 'train/grad_norm_pre_clip': 10.779232025146484, 'timing/data': 0.04126294795423746, 'timing/model': 1.130912248045206, 'learning_rate/qwen_vl_interface': 9.099176748192976e-06, 'learning_rate/action_model': 9.099176748192976e-05, 'epoch': 0.54}) +[2026-06-17 19:10:10,347][starVLA.training.train_starvla][INFO] - Step 1476, Loss: {'train/loss': 0.8231797218322754, 'train/grad_norm_pre_clip': 9.822495460510254, 'timing/data': 0.041114844381809235, 'timing/model': 1.141997391358018, 'learning_rate/qwen_vl_interface': 9.097909138035943e-06, 'learning_rate/action_model': 9.097909138035941e-05, 'epoch': 0.54}) +[2026-06-17 19:10:14,952][starVLA.training.train_starvla][INFO] - Step 1477, Loss: {'train/loss': 0.7587025165557861, 'train/grad_norm_pre_clip': 9.55008602142334, 'timing/data': 0.03176325932145119, 'timing/model': 1.0823293551802635, 'learning_rate/qwen_vl_interface': 9.096640730203202e-06, 'learning_rate/action_model': 9.096640730203202e-05, 'epoch': 0.54}) +[2026-06-17 19:10:19,383][starVLA.training.train_starvla][INFO] - Step 1478, Loss: {'train/loss': 0.8951961398124695, 'train/grad_norm_pre_clip': 12.773171424865723, 'timing/data': 0.0288524916395545, 'timing/model': 1.4198567243292928, 'learning_rate/qwen_vl_interface': 9.095371524957698e-06, 'learning_rate/action_model': 9.095371524957698e-05, 'epoch': 0.54}) +[2026-06-17 19:10:23,586][starVLA.training.train_starvla][INFO] - Step 1479, Loss: {'train/loss': 1.0133455991744995, 'train/grad_norm_pre_clip': 11.811334609985352, 'timing/data': 0.03109095338732004, 'timing/model': 1.0279645519331098, 'learning_rate/qwen_vl_interface': 9.094101522562541e-06, 'learning_rate/action_model': 9.09410152256254e-05, 'epoch': 0.54}) +[2026-06-17 19:10:29,042][starVLA.training.train_starvla][INFO] - Step 1480, Loss: {'train/loss': 0.8949469327926636, 'train/grad_norm_pre_clip': 15.716550827026367, 'timing/data': 0.02984779141843319, 'timing/model': 1.0294393971562386, 'learning_rate/qwen_vl_interface': 9.092830723281e-06, 'learning_rate/action_model': 9.092830723281e-05, 'epoch': 0.54}) +[2026-06-17 19:10:33,564][starVLA.training.train_starvla][INFO] - Step 1481, Loss: {'train/loss': 1.0902732610702515, 'train/grad_norm_pre_clip': 12.608743667602539, 'timing/data': 0.036623197607696056, 'timing/model': 1.062700710259378, 'learning_rate/qwen_vl_interface': 9.091559127376519e-06, 'learning_rate/action_model': 9.091559127376517e-05, 'epoch': 0.54}) +[2026-06-17 19:10:41,021][starVLA.training.train_starvla][INFO] - Step 1482, Loss: {'train/loss': 1.0164388418197632, 'train/grad_norm_pre_clip': 13.690289497375488, 'timing/data': 0.03072226606309414, 'timing/model': 1.2961881821975112, 'learning_rate/qwen_vl_interface': 9.090286735112693e-06, 'learning_rate/action_model': 9.090286735112693e-05, 'epoch': 0.54}) +[2026-06-17 19:10:45,529][starVLA.training.train_starvla][INFO] - Step 1483, Loss: {'train/loss': 0.7725770473480225, 'train/grad_norm_pre_clip': 11.915515899658203, 'timing/data': 0.03402448818087578, 'timing/model': 1.3851521406322718, 'learning_rate/qwen_vl_interface': 9.089013546753296e-06, 'learning_rate/action_model': 9.089013546753295e-05, 'epoch': 0.54}) +[2026-06-17 19:10:49,597][starVLA.training.train_starvla][INFO] - Step 1484, Loss: {'train/loss': 0.6721908450126648, 'train/grad_norm_pre_clip': 12.56443977355957, 'timing/data': 0.03544485103338957, 'timing/model': 1.0471418974921107, 'learning_rate/qwen_vl_interface': 9.08773956256226e-06, 'learning_rate/action_model': 9.087739562562259e-05, 'epoch': 0.54}) +[2026-06-17 19:10:54,199][starVLA.training.train_starvla][INFO] - Step 1485, Loss: {'train/loss': 1.1442177295684814, 'train/grad_norm_pre_clip': 13.184809684753418, 'timing/data': 0.03292834199965, 'timing/model': 1.1500228038057685, 'learning_rate/qwen_vl_interface': 9.086464782803682e-06, 'learning_rate/action_model': 9.086464782803682e-05, 'epoch': 0.54}) +[2026-06-17 19:10:58,769][starVLA.training.train_starvla][INFO] - Step 1486, Loss: {'train/loss': 1.2914165258407593, 'train/grad_norm_pre_clip': 15.266260147094727, 'timing/data': 0.03140703774988651, 'timing/model': 1.1332741482183337, 'learning_rate/qwen_vl_interface': 9.085189207741827e-06, 'learning_rate/action_model': 9.085189207741827e-05, 'epoch': 0.54}) +[2026-06-17 19:11:03,038][starVLA.training.train_starvla][INFO] - Step 1487, Loss: {'train/loss': 1.0775249004364014, 'train/grad_norm_pre_clip': 13.651721000671387, 'timing/data': 0.03284144587814808, 'timing/model': 1.038026018999517, 'learning_rate/qwen_vl_interface': 9.083912837641123e-06, 'learning_rate/action_model': 9.083912837641123e-05, 'epoch': 0.54}) +[2026-06-17 19:11:07,702][starVLA.training.train_starvla][INFO] - Step 1488, Loss: {'train/loss': 0.8058013916015625, 'train/grad_norm_pre_clip': 11.81207275390625, 'timing/data': 0.03796364367008209, 'timing/model': 1.4350112183019519, 'learning_rate/qwen_vl_interface': 9.082635672766162e-06, 'learning_rate/action_model': 9.082635672766161e-05, 'epoch': 0.54}) +[2026-06-17 19:11:11,985][starVLA.training.train_starvla][INFO] - Step 1489, Loss: {'train/loss': 0.4682302474975586, 'train/grad_norm_pre_clip': 10.13573169708252, 'timing/data': 0.03926308546215296, 'timing/model': 1.0738598443567753, 'learning_rate/qwen_vl_interface': 9.081357713381702e-06, 'learning_rate/action_model': 9.081357713381701e-05, 'epoch': 0.55}) +[2026-06-17 19:11:16,440][starVLA.training.train_starvla][INFO] - Step 1490, Loss: {'train/loss': 0.655197024345398, 'train/grad_norm_pre_clip': 11.06411075592041, 'timing/data': 0.031296830624341965, 'timing/model': 1.056377675384283, 'learning_rate/qwen_vl_interface': 9.080078959752664e-06, 'learning_rate/action_model': 9.080078959752664e-05, 'epoch': 0.55}) +[2026-06-17 19:11:21,016][starVLA.training.train_starvla][INFO] - Step 1491, Loss: {'train/loss': 0.6592519283294678, 'train/grad_norm_pre_clip': 9.812173843383789, 'timing/data': 0.032567087560892105, 'timing/model': 1.029189051128924, 'learning_rate/qwen_vl_interface': 9.078799412144137e-06, 'learning_rate/action_model': 9.078799412144137e-05, 'epoch': 0.55}) +[2026-06-17 19:11:25,528][starVLA.training.train_starvla][INFO] - Step 1492, Loss: {'train/loss': 0.735500693321228, 'train/grad_norm_pre_clip': 12.007132530212402, 'timing/data': 0.030494002625346184, 'timing/model': 1.0822384050115943, 'learning_rate/qwen_vl_interface': 9.077519070821373e-06, 'learning_rate/action_model': 9.077519070821373e-05, 'epoch': 0.55}) +[2026-06-17 19:11:29,858][starVLA.training.train_starvla][INFO] - Step 1493, Loss: {'train/loss': 0.9397719502449036, 'train/grad_norm_pre_clip': 11.403770446777344, 'timing/data': 0.02913007792085409, 'timing/model': 1.1634971983730793, 'learning_rate/qwen_vl_interface': 9.076237936049787e-06, 'learning_rate/action_model': 9.076237936049786e-05, 'epoch': 0.55}) +[2026-06-17 19:11:34,646][starVLA.training.train_starvla][INFO] - Step 1494, Loss: {'train/loss': 0.8036560416221619, 'train/grad_norm_pre_clip': 15.317991256713867, 'timing/data': 0.03244335576891899, 'timing/model': 1.1931839101016521, 'learning_rate/qwen_vl_interface': 9.074956008094958e-06, 'learning_rate/action_model': 9.074956008094958e-05, 'epoch': 0.55}) +[2026-06-17 19:11:39,209][starVLA.training.train_starvla][INFO] - Step 1495, Loss: {'train/loss': 0.9394525289535522, 'train/grad_norm_pre_clip': 11.878615379333496, 'timing/data': 0.031934501603245735, 'timing/model': 1.0750523610040545, 'learning_rate/qwen_vl_interface': 9.073673287222634e-06, 'learning_rate/action_model': 9.073673287222633e-05, 'epoch': 0.55}) +[2026-06-17 19:11:43,549][starVLA.training.train_starvla][INFO] - Step 1496, Loss: {'train/loss': 0.8688735365867615, 'train/grad_norm_pre_clip': 11.480048179626465, 'timing/data': 0.033450606279075146, 'timing/model': 1.0437939567491412, 'learning_rate/qwen_vl_interface': 9.072389773698723e-06, 'learning_rate/action_model': 9.072389773698723e-05, 'epoch': 0.55}) +[2026-06-17 19:11:51,429][starVLA.training.train_starvla][INFO] - Step 1497, Loss: {'train/loss': 0.8727430701255798, 'train/grad_norm_pre_clip': 12.981284141540527, 'timing/data': 0.03260885551571846, 'timing/model': 3.3342353720217943, 'learning_rate/qwen_vl_interface': 9.071105467789299e-06, 'learning_rate/action_model': 9.071105467789298e-05, 'epoch': 0.55}) +[2026-06-17 19:12:00,214][starVLA.training.train_starvla][INFO] - Step 1498, Loss: {'train/loss': 0.8523238301277161, 'train/grad_norm_pre_clip': 12.098206520080566, 'timing/data': 0.03520679380744696, 'timing/model': 3.1055249040946364, 'learning_rate/qwen_vl_interface': 9.0698203697606e-06, 'learning_rate/action_model': 9.0698203697606e-05, 'epoch': 0.55}) +[2026-06-17 19:12:05,378][starVLA.training.train_starvla][INFO] - Step 1499, Loss: {'train/loss': 1.1589360237121582, 'train/grad_norm_pre_clip': 11.845351219177246, 'timing/data': 0.03119524009525776, 'timing/model': 1.101069682277739, 'learning_rate/qwen_vl_interface': 9.068534479879026e-06, 'learning_rate/action_model': 9.068534479879024e-05, 'epoch': 0.55}) +[2026-06-17 19:14:31,280][starVLA.training.train_starvla][INFO] - Step 1500, Loss: {'train/loss': 0.8379648327827454, 'train/grad_norm_pre_clip': 10.889679908752441, 'eval/loss': 0.853726863861084, 'eval/latency_2/loss': 0.853726863861084, 'eval/demon_fire_precision': 0.8759842519685039, 'eval/demon_fire_recall': 0.9309623430962343, 'eval/demon_fire_f1': 0.9026369168356998, 'eval/demon_fire_rate_model': 0.7067334446299388, 'eval/demon_fire_rate_teacher': 0.6649972175848636, 'eval/demon_left_precision': 0.8480113636363636, 'eval/demon_left_recall': 0.628421052631579, 'eval/demon_left_f1': 0.7218863361547764, 'eval/demon_left_rate_model': 0.19588202559821927, 'eval/demon_left_rate_teacher': 0.2643294379521425, 'eval/demon_right_precision': 0.8465909090909091, 'eval/demon_right_recall': 0.6734463276836158, 'eval/demon_right_f1': 0.750157331655129, 'eval/demon_right_rate_model': 0.19588202559821927, 'eval/demon_right_rate_teacher': 0.24624373956594323, 'eval/demon_move_f1': 0.7360218339049527, 'eval/demon_cc_f1': 0.84523550153892, 'eval/latency_2/demon_fire_precision': 0.8759842519685039, 'eval/latency_2/demon_fire_recall': 0.9309623430962343, 'eval/latency_2/demon_fire_f1': 0.9026369168356998, 'eval/latency_2/demon_fire_rate_model': 0.7067334446299388, 'eval/latency_2/demon_fire_rate_teacher': 0.6649972175848636, 'eval/latency_2/demon_left_precision': 0.8480113636363636, 'eval/latency_2/demon_left_recall': 0.628421052631579, 'eval/latency_2/demon_left_f1': 0.7218863361547764, 'eval/latency_2/demon_left_rate_model': 0.19588202559821927, 'eval/latency_2/demon_left_rate_teacher': 0.2643294379521425, 'eval/latency_2/demon_right_precision': 0.8465909090909091, 'eval/latency_2/demon_right_recall': 0.6734463276836158, 'eval/latency_2/demon_right_f1': 0.750157331655129, 'eval/latency_2/demon_right_rate_model': 0.19588202559821927, 'eval/latency_2/demon_right_rate_teacher': 0.24624373956594323, 'eval/latency_2/demon_move_f1': 0.7360218339049527, 'eval/latency_2/demon_cc_f1': 0.84523550153892, 'timing/data': 0.03085933532565832, 'timing/model': 1.1191236237064004, 'learning_rate/qwen_vl_interface': 9.067247798411147e-06, 'learning_rate/action_model': 9.067247798411147e-05, 'epoch': 0.55}) +[2026-06-17 19:14:31,282][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_1500_state +[2026-06-17 19:14:31,289][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 19:14:39,763][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_1500_state/model.safetensors +[2026-06-17 19:14:54,219][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_1500_state/optimizer.bin +[2026-06-17 19:14:54,235][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_1500_state/sampler.bin +[2026-06-17 19:14:54,236][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_1500_state/sampler_1.bin +[2026-06-17 19:14:54,238][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_1500_state/random_states_0.pkl +[2026-06-17 19:14:54,243][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1500 +[2026-06-17 19:14:54,243][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 19:14:54,261][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 19:14:58,832][starVLA.training.train_starvla][INFO] - Step 1501, Loss: {'train/loss': 0.9277939796447754, 'train/grad_norm_pre_clip': 10.788884162902832, 'timing/data': 0.05113935749977827, 'timing/model': 1.1238475134596229, 'learning_rate/qwen_vl_interface': 9.065960325623692e-06, 'learning_rate/action_model': 9.065960325623692e-05, 'epoch': 0.55}) +[2026-06-17 19:15:03,407][starVLA.training.train_starvla][INFO] - Step 1502, Loss: {'train/loss': 1.0947489738464355, 'train/grad_norm_pre_clip': 16.00867462158203, 'timing/data': 0.030569988302886486, 'timing/model': 1.0950094750151038, 'learning_rate/qwen_vl_interface': 9.064672061783555e-06, 'learning_rate/action_model': 9.064672061783554e-05, 'epoch': 0.55}) +[2026-06-17 19:15:07,867][starVLA.training.train_starvla][INFO] - Step 1503, Loss: {'train/loss': 0.6011435985565186, 'train/grad_norm_pre_clip': 8.844498634338379, 'timing/data': 0.03197997435927391, 'timing/model': 1.420790672302246, 'learning_rate/qwen_vl_interface': 9.063383007157795e-06, 'learning_rate/action_model': 9.063383007157796e-05, 'epoch': 0.55}) +[2026-06-17 19:15:11,946][starVLA.training.train_starvla][INFO] - Step 1504, Loss: {'train/loss': 0.7935783267021179, 'train/grad_norm_pre_clip': 12.160447120666504, 'timing/data': 0.03317858651280403, 'timing/model': 1.0321236411109567, 'learning_rate/qwen_vl_interface': 9.062093162013635e-06, 'learning_rate/action_model': 9.062093162013634e-05, 'epoch': 0.55}) +[2026-06-17 19:15:19,544][starVLA.training.train_starvla][INFO] - Step 1505, Loss: {'train/loss': 0.627343475818634, 'train/grad_norm_pre_clip': 10.242488861083984, 'timing/data': 0.034023579210042953, 'timing/model': 4.262013762257993, 'learning_rate/qwen_vl_interface': 9.06080252661846e-06, 'learning_rate/action_model': 9.060802526618458e-05, 'epoch': 0.55}) +[2026-06-17 19:15:26,764][starVLA.training.train_starvla][INFO] - Step 1506, Loss: {'train/loss': 0.9027576446533203, 'train/grad_norm_pre_clip': 12.791940689086914, 'timing/data': 0.03470880910754204, 'timing/model': 1.0333473850041628, 'learning_rate/qwen_vl_interface': 9.059511101239818e-06, 'learning_rate/action_model': 9.059511101239818e-05, 'epoch': 0.55}) +[2026-06-17 19:15:31,065][starVLA.training.train_starvla][INFO] - Step 1507, Loss: {'train/loss': 0.5417485237121582, 'train/grad_norm_pre_clip': 8.638419151306152, 'timing/data': 0.034002359956502914, 'timing/model': 1.0569265922531486, 'learning_rate/qwen_vl_interface': 9.058218886145427e-06, 'learning_rate/action_model': 9.058218886145427e-05, 'epoch': 0.55}) +[2026-06-17 19:15:35,491][starVLA.training.train_starvla][INFO] - Step 1508, Loss: {'train/loss': 0.8943391442298889, 'train/grad_norm_pre_clip': 109.2715835571289, 'timing/data': 0.03511543292552233, 'timing/model': 1.3689457532018423, 'learning_rate/qwen_vl_interface': 9.056925881603165e-06, 'learning_rate/action_model': 9.056925881603166e-05, 'epoch': 0.55}) +[2026-06-17 19:15:39,839][starVLA.training.train_starvla][INFO] - Step 1509, Loss: {'train/loss': 0.7161557674407959, 'train/grad_norm_pre_clip': 10.753713607788086, 'timing/data': 0.031243912875652313, 'timing/model': 1.2966980868950486, 'learning_rate/qwen_vl_interface': 9.05563208788107e-06, 'learning_rate/action_model': 9.05563208788107e-05, 'epoch': 0.55}) +[2026-06-17 19:15:43,881][starVLA.training.train_starvla][INFO] - Step 1510, Loss: {'train/loss': 0.7527211904525757, 'train/grad_norm_pre_clip': 12.179370880126953, 'timing/data': 0.0368178766220808, 'timing/model': 1.0291882818564773, 'learning_rate/qwen_vl_interface': 9.054337505247349e-06, 'learning_rate/action_model': 9.054337505247348e-05, 'epoch': 0.55}) +[2026-06-17 19:15:48,282][starVLA.training.train_starvla][INFO] - Step 1511, Loss: {'train/loss': 0.7408933639526367, 'train/grad_norm_pre_clip': 12.286377906799316, 'timing/data': 0.032058751210570335, 'timing/model': 1.0911780297756195, 'learning_rate/qwen_vl_interface': 9.05304213397037e-06, 'learning_rate/action_model': 9.053042133970369e-05, 'epoch': 0.55}) +[2026-06-17 19:15:52,646][starVLA.training.train_starvla][INFO] - Step 1512, Loss: {'train/loss': 0.7719024419784546, 'train/grad_norm_pre_clip': 12.552775382995605, 'timing/data': 0.05034400708973408, 'timing/model': 1.033675691112876, 'learning_rate/qwen_vl_interface': 9.051745974318663e-06, 'learning_rate/action_model': 9.051745974318662e-05, 'epoch': 0.55}) +[2026-06-17 19:15:56,962][starVLA.training.train_starvla][INFO] - Step 1513, Loss: {'train/loss': 0.7504525780677795, 'train/grad_norm_pre_clip': 12.699633598327637, 'timing/data': 0.034414131194353104, 'timing/model': 1.0305582266300917, 'learning_rate/qwen_vl_interface': 9.050449026560923e-06, 'learning_rate/action_model': 9.050449026560924e-05, 'epoch': 0.55}) +[2026-06-17 19:16:01,439][starVLA.training.train_starvla][INFO] - Step 1514, Loss: {'train/loss': 0.8639519810676575, 'train/grad_norm_pre_clip': 12.320463180541992, 'timing/data': 0.03403371572494507, 'timing/model': 1.3997619720175862, 'learning_rate/qwen_vl_interface': 9.049151290966013e-06, 'learning_rate/action_model': 9.049151290966012e-05, 'epoch': 0.55}) +[2026-06-17 19:16:05,619][starVLA.training.train_starvla][INFO] - Step 1515, Loss: {'train/loss': 0.5652942657470703, 'train/grad_norm_pre_clip': 8.44500732421875, 'timing/data': 0.030847995541989803, 'timing/model': 1.0979494517669082, 'learning_rate/qwen_vl_interface': 9.04785276780295e-06, 'learning_rate/action_model': 9.04785276780295e-05, 'epoch': 0.55}) +[2026-06-17 19:16:10,042][starVLA.training.train_starvla][INFO] - Step 1516, Loss: {'train/loss': 0.8122379183769226, 'train/grad_norm_pre_clip': 13.6676607131958, 'timing/data': 0.03894964046776295, 'timing/model': 1.0385948605835438, 'learning_rate/qwen_vl_interface': 9.046553457340924e-06, 'learning_rate/action_model': 9.046553457340924e-05, 'epoch': 0.56}) +[2026-06-17 19:16:14,424][starVLA.training.train_starvla][INFO] - Step 1517, Loss: {'train/loss': 0.7906363606452942, 'train/grad_norm_pre_clip': 14.627455711364746, 'timing/data': 0.031256516464054585, 'timing/model': 1.0708884056657553, 'learning_rate/qwen_vl_interface': 9.04525335984928e-06, 'learning_rate/action_model': 9.04525335984928e-05, 'epoch': 0.56}) +[2026-06-17 19:16:19,236][starVLA.training.train_starvla][INFO] - Step 1518, Loss: {'train/loss': 0.6777027249336243, 'train/grad_norm_pre_clip': 16.946914672851562, 'timing/data': 0.03977327328175306, 'timing/model': 1.1656020116060972, 'learning_rate/qwen_vl_interface': 9.043952475597532e-06, 'learning_rate/action_model': 9.043952475597532e-05, 'epoch': 0.56}) +[2026-06-17 19:16:23,537][starVLA.training.train_starvla][INFO] - Step 1519, Loss: {'train/loss': 0.5737916827201843, 'train/grad_norm_pre_clip': 10.355452537536621, 'timing/data': 0.03268574643880129, 'timing/model': 1.3324210550636053, 'learning_rate/qwen_vl_interface': 9.04265080485535e-06, 'learning_rate/action_model': 9.042650804855351e-05, 'epoch': 0.56}) +[2026-06-17 19:16:27,572][starVLA.training.train_starvla][INFO] - Step 1520, Loss: {'train/loss': 0.8623608946800232, 'train/grad_norm_pre_clip': 12.120803833007812, 'timing/data': 0.03186419326812029, 'timing/model': 1.033068722113967, 'learning_rate/qwen_vl_interface': 9.041348347892578e-06, 'learning_rate/action_model': 9.041348347892579e-05, 'epoch': 0.56}) +[2026-06-17 19:16:32,170][starVLA.training.train_starvla][INFO] - Step 1521, Loss: {'train/loss': 0.6829545497894287, 'train/grad_norm_pre_clip': 11.877241134643555, 'timing/data': 0.030645588412880898, 'timing/model': 1.1057199873030186, 'learning_rate/qwen_vl_interface': 9.040045104979213e-06, 'learning_rate/action_model': 9.040045104979214e-05, 'epoch': 0.56}) +[2026-06-17 19:16:36,606][starVLA.training.train_starvla][INFO] - Step 1522, Loss: {'train/loss': 1.0934993028640747, 'train/grad_norm_pre_clip': 22.252614974975586, 'timing/data': 0.061022549867630005, 'timing/model': 1.0994321005418897, 'learning_rate/qwen_vl_interface': 9.03874107638542e-06, 'learning_rate/action_model': 9.038741076385419e-05, 'epoch': 0.56}) +[2026-06-17 19:16:42,474][starVLA.training.train_starvla][INFO] - Step 1523, Loss: {'train/loss': 0.9781442284584045, 'train/grad_norm_pre_clip': 13.8031587600708, 'timing/data': 0.037672556936740875, 'timing/model': 1.0256370743736625, 'learning_rate/qwen_vl_interface': 9.037436262381525e-06, 'learning_rate/action_model': 9.037436262381525e-05, 'epoch': 0.56}) +[2026-06-17 19:16:46,792][starVLA.training.train_starvla][INFO] - Step 1524, Loss: {'train/loss': 0.7542113661766052, 'train/grad_norm_pre_clip': 11.0255765914917, 'timing/data': 0.03280505444854498, 'timing/model': 1.3311027456074953, 'learning_rate/qwen_vl_interface': 9.036130663238018e-06, 'learning_rate/action_model': 9.036130663238017e-05, 'epoch': 0.56}) +[2026-06-17 19:16:50,926][starVLA.training.train_starvla][INFO] - Step 1525, Loss: {'train/loss': 0.5802357792854309, 'train/grad_norm_pre_clip': 12.726428985595703, 'timing/data': 0.033652735874056816, 'timing/model': 1.0488077644258738, 'learning_rate/qwen_vl_interface': 9.034824279225547e-06, 'learning_rate/action_model': 9.034824279225547e-05, 'epoch': 0.56}) +[2026-06-17 19:16:55,331][starVLA.training.train_starvla][INFO] - Step 1526, Loss: {'train/loss': 0.7527050971984863, 'train/grad_norm_pre_clip': 11.094708442687988, 'timing/data': 0.035053253173828125, 'timing/model': 1.034003796055913, 'learning_rate/qwen_vl_interface': 9.033517110614933e-06, 'learning_rate/action_model': 9.033517110614933e-05, 'epoch': 0.56}) +[2026-06-17 19:16:59,703][starVLA.training.train_starvla][INFO] - Step 1527, Loss: {'train/loss': 1.1296306848526, 'train/grad_norm_pre_clip': 18.74451446533203, 'timing/data': 0.03234332427382469, 'timing/model': 1.0404616668820381, 'learning_rate/qwen_vl_interface': 9.032209157677151e-06, 'learning_rate/action_model': 9.032209157677151e-05, 'epoch': 0.56}) +[2026-06-17 19:17:04,256][starVLA.training.train_starvla][INFO] - Step 1528, Loss: {'train/loss': 0.6977509260177612, 'train/grad_norm_pre_clip': 11.459519386291504, 'timing/data': 0.035551224835217, 'timing/model': 1.0472376318648458, 'learning_rate/qwen_vl_interface': 9.03090042068334e-06, 'learning_rate/action_model': 9.030900420683339e-05, 'epoch': 0.56}) +[2026-06-17 19:17:12,095][starVLA.training.train_starvla][INFO] - Step 1529, Loss: {'train/loss': 1.0450929403305054, 'train/grad_norm_pre_clip': 15.224881172180176, 'timing/data': 0.029651662334799767, 'timing/model': 3.6029783887788653, 'learning_rate/qwen_vl_interface': 9.029590899904803e-06, 'learning_rate/action_model': 9.029590899904803e-05, 'epoch': 0.56}) +[2026-06-17 19:17:16,226][starVLA.training.train_starvla][INFO] - Step 1530, Loss: {'train/loss': 1.1176997423171997, 'train/grad_norm_pre_clip': 13.418387413024902, 'timing/data': 0.03250661864876747, 'timing/model': 1.1263987822458148, 'learning_rate/qwen_vl_interface': 9.028280595613006e-06, 'learning_rate/action_model': 9.028280595613005e-05, 'epoch': 0.56}) +[2026-06-17 19:17:24,438][starVLA.training.train_starvla][INFO] - Step 1531, Loss: {'train/loss': 1.0539706945419312, 'train/grad_norm_pre_clip': 36.74406051635742, 'timing/data': 0.032462140545248985, 'timing/model': 3.7310846718028188, 'learning_rate/qwen_vl_interface': 9.026969508079575e-06, 'learning_rate/action_model': 9.026969508079575e-05, 'epoch': 0.56}) +[2026-06-17 19:17:28,842][starVLA.training.train_starvla][INFO] - Step 1532, Loss: {'train/loss': 1.1261855363845825, 'train/grad_norm_pre_clip': 14.077609062194824, 'timing/data': 0.031841122545301914, 'timing/model': 1.0822216626256704, 'learning_rate/qwen_vl_interface': 9.025657637576302e-06, 'learning_rate/action_model': 9.025657637576302e-05, 'epoch': 0.56}) +[2026-06-17 19:17:33,270][starVLA.training.train_starvla][INFO] - Step 1533, Loss: {'train/loss': 0.6490848660469055, 'train/grad_norm_pre_clip': 14.990797996520996, 'timing/data': 0.03656884375959635, 'timing/model': 1.121514300815761, 'learning_rate/qwen_vl_interface': 9.024344984375138e-06, 'learning_rate/action_model': 9.024344984375137e-05, 'epoch': 0.56}) +[2026-06-17 19:17:37,558][starVLA.training.train_starvla][INFO] - Step 1534, Loss: {'train/loss': 1.0307157039642334, 'train/grad_norm_pre_clip': 15.490409851074219, 'timing/data': 0.03253647033125162, 'timing/model': 1.3046122509986162, 'learning_rate/qwen_vl_interface': 9.023031548748197e-06, 'learning_rate/action_model': 9.023031548748196e-05, 'epoch': 0.56}) +[2026-06-17 19:17:41,657][starVLA.training.train_starvla][INFO] - Step 1535, Loss: {'train/loss': 1.2490314245224, 'train/grad_norm_pre_clip': 16.236536026000977, 'timing/data': 0.031937905587255955, 'timing/model': 1.117409293539822, 'learning_rate/qwen_vl_interface': 9.021717330967755e-06, 'learning_rate/action_model': 9.021717330967754e-05, 'epoch': 0.56}) +[2026-06-17 19:17:46,061][starVLA.training.train_starvla][INFO] - Step 1536, Loss: {'train/loss': 0.7766448259353638, 'train/grad_norm_pre_clip': 13.1571044921875, 'timing/data': 0.03295086324214935, 'timing/model': 1.0615814756602049, 'learning_rate/qwen_vl_interface': 9.020402331306254e-06, 'learning_rate/action_model': 9.020402331306253e-05, 'epoch': 0.56}) +[2026-06-17 19:17:50,569][starVLA.training.train_starvla][INFO] - Step 1537, Loss: {'train/loss': 0.9155820608139038, 'train/grad_norm_pre_clip': 15.259967803955078, 'timing/data': 0.04227658826857805, 'timing/model': 1.1366628017276525, 'learning_rate/qwen_vl_interface': 9.01908655003629e-06, 'learning_rate/action_model': 9.01908655003629e-05, 'epoch': 0.56}) +[2026-06-17 19:17:54,935][starVLA.training.train_starvla][INFO] - Step 1538, Loss: {'train/loss': 1.0154099464416504, 'train/grad_norm_pre_clip': 13.891695022583008, 'timing/data': 0.0213545523583889, 'timing/model': 1.380808656103909, 'learning_rate/qwen_vl_interface': 9.017769987430629e-06, 'learning_rate/action_model': 9.017769987430628e-05, 'epoch': 0.56}) +[2026-06-17 19:17:59,028][starVLA.training.train_starvla][INFO] - Step 1539, Loss: {'train/loss': 0.9944078326225281, 'train/grad_norm_pre_clip': 13.415794372558594, 'timing/data': 0.023285308852791786, 'timing/model': 1.0762113304808736, 'learning_rate/qwen_vl_interface': 9.016452643762197e-06, 'learning_rate/action_model': 9.016452643762197e-05, 'epoch': 0.56}) +[2026-06-17 19:18:03,509][starVLA.training.train_starvla][INFO] - Step 1540, Loss: {'train/loss': 0.7121023535728455, 'train/grad_norm_pre_clip': 12.8051118850708, 'timing/data': 0.021915940567851067, 'timing/model': 1.0742540806531906, 'learning_rate/qwen_vl_interface': 9.015134519304078e-06, 'learning_rate/action_model': 9.015134519304078e-05, 'epoch': 0.56}) +[2026-06-17 19:18:08,013][starVLA.training.train_starvla][INFO] - Step 1541, Loss: {'train/loss': 0.7125842571258545, 'train/grad_norm_pre_clip': 11.81849193572998, 'timing/data': 0.021043891087174416, 'timing/model': 1.0752293467521667, 'learning_rate/qwen_vl_interface': 9.013815614329521e-06, 'learning_rate/action_model': 9.013815614329522e-05, 'epoch': 0.56}) +[2026-06-17 19:18:12,430][starVLA.training.train_starvla][INFO] - Step 1542, Loss: {'train/loss': 0.5593873858451843, 'train/grad_norm_pre_clip': 10.143768310546875, 'timing/data': 0.022464005276560783, 'timing/model': 1.0886494228616357, 'learning_rate/qwen_vl_interface': 9.01249592911194e-06, 'learning_rate/action_model': 9.012495929111938e-05, 'epoch': 0.56}) +[2026-06-17 19:18:16,894][starVLA.training.train_starvla][INFO] - Step 1543, Loss: {'train/loss': 0.7644730806350708, 'train/grad_norm_pre_clip': 27.815719604492188, 'timing/data': 0.024173048324882984, 'timing/model': 1.3992475513368845, 'learning_rate/qwen_vl_interface': 9.0111754639249e-06, 'learning_rate/action_model': 9.0111754639249e-05, 'epoch': 0.57}) +[2026-06-17 19:18:24,546][starVLA.training.train_starvla][INFO] - Step 1544, Loss: {'train/loss': 0.6877801418304443, 'train/grad_norm_pre_clip': 24.223018646240234, 'timing/data': 0.021391739137470722, 'timing/model': 4.636159246787429, 'learning_rate/qwen_vl_interface': 9.009854219042142e-06, 'learning_rate/action_model': 9.009854219042142e-05, 'epoch': 0.57}) +[2026-06-17 19:18:28,600][starVLA.training.train_starvla][INFO] - Step 1545, Loss: {'train/loss': 0.83879154920578, 'train/grad_norm_pre_clip': 14.94367790222168, 'timing/data': 0.0266775107011199, 'timing/model': 1.0481949001550674, 'learning_rate/qwen_vl_interface': 9.008532194737556e-06, 'learning_rate/action_model': 9.008532194737557e-05, 'epoch': 0.57}) +[2026-06-17 19:18:32,979][starVLA.training.train_starvla][INFO] - Step 1546, Loss: {'train/loss': 0.5969136357307434, 'train/grad_norm_pre_clip': 10.565438270568848, 'timing/data': 0.023686259053647518, 'timing/model': 1.0546164168044925, 'learning_rate/qwen_vl_interface': 9.007209391285205e-06, 'learning_rate/action_model': 9.007209391285205e-05, 'epoch': 0.57}) +[2026-06-17 19:18:37,316][starVLA.training.train_starvla][INFO] - Step 1547, Loss: {'train/loss': 1.2030655145645142, 'train/grad_norm_pre_clip': 16.65045166015625, 'timing/data': 0.024964372627437115, 'timing/model': 1.0688507333397865, 'learning_rate/qwen_vl_interface': 9.005885808959301e-06, 'learning_rate/action_model': 9.005885808959302e-05, 'epoch': 0.57}) +[2026-06-17 19:18:42,149][starVLA.training.train_starvla][INFO] - Step 1548, Loss: {'train/loss': 0.9578495621681213, 'train/grad_norm_pre_clip': 12.05491828918457, 'timing/data': 0.0228508235886693, 'timing/model': 1.0896237595006824, 'learning_rate/qwen_vl_interface': 9.00456144803423e-06, 'learning_rate/action_model': 9.00456144803423e-05, 'epoch': 0.57}) +[2026-06-17 19:18:46,488][starVLA.training.train_starvla][INFO] - Step 1549, Loss: {'train/loss': 0.9732919931411743, 'train/grad_norm_pre_clip': 17.20800018310547, 'timing/data': 0.020949913188815117, 'timing/model': 1.3311982369050384, 'learning_rate/qwen_vl_interface': 9.00323630878453e-06, 'learning_rate/action_model': 9.00323630878453e-05, 'epoch': 0.57}) +[2026-06-17 19:18:50,550][starVLA.training.train_starvla][INFO] - Step 1550, Loss: {'train/loss': 0.7420139908790588, 'train/grad_norm_pre_clip': 11.622516632080078, 'timing/data': 0.02121184580028057, 'timing/model': 1.0226276302710176, 'learning_rate/qwen_vl_interface': 9.001910391484903e-06, 'learning_rate/action_model': 9.001910391484904e-05, 'epoch': 0.57}) +[2026-06-17 19:18:54,859][starVLA.training.train_starvla][INFO] - Step 1551, Loss: {'train/loss': 0.6800782680511475, 'train/grad_norm_pre_clip': 10.30514144897461, 'timing/data': 0.02647004835307598, 'timing/model': 1.0381513051688671, 'learning_rate/qwen_vl_interface': 9.000583696410216e-06, 'learning_rate/action_model': 9.000583696410217e-05, 'epoch': 0.57}) +[2026-06-17 19:19:00,380][starVLA.training.train_starvla][INFO] - Step 1552, Loss: {'train/loss': 0.698193371295929, 'train/grad_norm_pre_clip': 10.091890335083008, 'timing/data': 0.02626799512654543, 'timing/model': 1.072813594713807, 'learning_rate/qwen_vl_interface': 8.999256223835493e-06, 'learning_rate/action_model': 8.999256223835494e-05, 'epoch': 0.57}) +[2026-06-17 19:19:07,251][starVLA.training.train_starvla][INFO] - Step 1553, Loss: {'train/loss': 0.6457648873329163, 'train/grad_norm_pre_clip': 11.072620391845703, 'timing/data': 0.02146462630480528, 'timing/model': 3.5459882710129023, 'learning_rate/qwen_vl_interface': 8.997927974035922e-06, 'learning_rate/action_model': 8.997927974035921e-05, 'epoch': 0.57}) +[2026-06-17 19:19:13,110][starVLA.training.train_starvla][INFO] - Step 1554, Loss: {'train/loss': 0.6917676329612732, 'train/grad_norm_pre_clip': 11.74174976348877, 'timing/data': 0.019575522281229496, 'timing/model': 1.3347787009552121, 'learning_rate/qwen_vl_interface': 8.996598947286846e-06, 'learning_rate/action_model': 8.996598947286846e-05, 'epoch': 0.57}) +[2026-06-17 19:19:17,308][starVLA.training.train_starvla][INFO] - Step 1555, Loss: {'train/loss': 0.7792641520500183, 'train/grad_norm_pre_clip': 11.05040168762207, 'timing/data': 0.02147846855223179, 'timing/model': 1.109742034226656, 'learning_rate/qwen_vl_interface': 8.99526914386378e-06, 'learning_rate/action_model': 8.995269143863779e-05, 'epoch': 0.57}) +[2026-06-17 19:19:21,713][starVLA.training.train_starvla][INFO] - Step 1556, Loss: {'train/loss': 0.7204140424728394, 'train/grad_norm_pre_clip': 10.757938385009766, 'timing/data': 0.03154610190540552, 'timing/model': 1.0611754823476076, 'learning_rate/qwen_vl_interface': 8.993938564042386e-06, 'learning_rate/action_model': 8.993938564042387e-05, 'epoch': 0.57}) +[2026-06-17 19:19:26,093][starVLA.training.train_starvla][INFO] - Step 1557, Loss: {'train/loss': 0.5786038041114807, 'train/grad_norm_pre_clip': 10.360745429992676, 'timing/data': 0.022156551480293274, 'timing/model': 1.026300380937755, 'learning_rate/qwen_vl_interface': 8.992607208098502e-06, 'learning_rate/action_model': 8.992607208098503e-05, 'epoch': 0.57}) +[2026-06-17 19:19:31,733][starVLA.training.train_starvla][INFO] - Step 1558, Loss: {'train/loss': 0.6427636742591858, 'train/grad_norm_pre_clip': 12.201345443725586, 'timing/data': 0.02283241879194975, 'timing/model': 1.05510373134166, 'learning_rate/qwen_vl_interface': 8.991275076308117e-06, 'learning_rate/action_model': 8.991275076308117e-05, 'epoch': 0.57}) +[2026-06-17 19:19:36,170][starVLA.training.train_starvla][INFO] - Step 1559, Loss: {'train/loss': 0.798724889755249, 'train/grad_norm_pre_clip': 13.704798698425293, 'timing/data': 0.022593485191464424, 'timing/model': 1.3881346294656396, 'learning_rate/qwen_vl_interface': 8.989942168947382e-06, 'learning_rate/action_model': 8.989942168947381e-05, 'epoch': 0.57}) +[2026-06-17 19:19:41,831][starVLA.training.train_starvla][INFO] - Step 1560, Loss: {'train/loss': 0.467153936624527, 'train/grad_norm_pre_clip': 9.06859302520752, 'timing/data': 0.03109237365424633, 'timing/model': 1.120614280924201, 'learning_rate/qwen_vl_interface': 8.98860848629261e-06, 'learning_rate/action_model': 8.988608486292611e-05, 'epoch': 0.57}) +[2026-06-17 19:19:46,919][starVLA.training.train_starvla][INFO] - Step 1561, Loss: {'train/loss': 0.6611450910568237, 'train/grad_norm_pre_clip': 15.435952186584473, 'timing/data': 0.020381983369588852, 'timing/model': 1.034352826885879, 'learning_rate/qwen_vl_interface': 8.987274028620276e-06, 'learning_rate/action_model': 8.987274028620276e-05, 'epoch': 0.57}) +[2026-06-17 19:19:51,323][starVLA.training.train_starvla][INFO] - Step 1562, Loss: {'train/loss': 0.5636929869651794, 'train/grad_norm_pre_clip': 10.688820838928223, 'timing/data': 0.02375544235110283, 'timing/model': 1.0346885407343507, 'learning_rate/qwen_vl_interface': 8.985938796207017e-06, 'learning_rate/action_model': 8.985938796207017e-05, 'epoch': 0.57}) +[2026-06-17 19:19:55,815][starVLA.training.train_starvla][INFO] - Step 1563, Loss: {'train/loss': 0.771367073059082, 'train/grad_norm_pre_clip': 12.845582008361816, 'timing/data': 0.02020927146077156, 'timing/model': 1.065373652614653, 'learning_rate/qwen_vl_interface': 8.984602789329623e-06, 'learning_rate/action_model': 8.984602789329623e-05, 'epoch': 0.57}) +[2026-06-17 19:20:00,287][starVLA.training.train_starvla][INFO] - Step 1564, Loss: {'train/loss': 0.9095942378044128, 'train/grad_norm_pre_clip': 13.104735374450684, 'timing/data': 0.021923083811998367, 'timing/model': 1.3512973431497812, 'learning_rate/qwen_vl_interface': 8.983266008265052e-06, 'learning_rate/action_model': 8.983266008265052e-05, 'epoch': 0.57}) +[2026-06-17 19:20:04,582][starVLA.training.train_starvla][INFO] - Step 1565, Loss: {'train/loss': 0.46706894040107727, 'train/grad_norm_pre_clip': 19.46123695373535, 'timing/data': 0.020009208470582962, 'timing/model': 1.1063693389296532, 'learning_rate/qwen_vl_interface': 8.981928453290421e-06, 'learning_rate/action_model': 8.981928453290421e-05, 'epoch': 0.57}) +[2026-06-17 19:20:09,100][starVLA.training.train_starvla][INFO] - Step 1566, Loss: {'train/loss': 0.5818426609039307, 'train/grad_norm_pre_clip': 12.325495719909668, 'timing/data': 0.02090784627944231, 'timing/model': 1.040220646187663, 'learning_rate/qwen_vl_interface': 8.980590124683007e-06, 'learning_rate/action_model': 8.980590124683007e-05, 'epoch': 0.57}) +[2026-06-17 19:20:15,897][starVLA.training.train_starvla][INFO] - Step 1567, Loss: {'train/loss': 0.9338195323944092, 'train/grad_norm_pre_clip': 12.524950981140137, 'timing/data': 0.026900931261479855, 'timing/model': 1.067070122808218, 'learning_rate/qwen_vl_interface': 8.979251022720244e-06, 'learning_rate/action_model': 8.979251022720244e-05, 'epoch': 0.57}) +[2026-06-17 19:20:20,354][starVLA.training.train_starvla][INFO] - Step 1568, Loss: {'train/loss': 0.8783343434333801, 'train/grad_norm_pre_clip': 11.213855743408203, 'timing/data': 0.026389681734144688, 'timing/model': 1.0465043699368834, 'learning_rate/qwen_vl_interface': 8.977911147679733e-06, 'learning_rate/action_model': 8.977911147679733e-05, 'epoch': 0.57}) +[2026-06-17 19:20:24,711][starVLA.training.train_starvla][INFO] - Step 1569, Loss: {'train/loss': 0.6045761704444885, 'train/grad_norm_pre_clip': 11.442867279052734, 'timing/data': 0.023918893188238144, 'timing/model': 1.323548604734242, 'learning_rate/qwen_vl_interface': 8.976570499839229e-06, 'learning_rate/action_model': 8.976570499839228e-05, 'epoch': 0.57}) +[2026-06-17 19:20:28,769][starVLA.training.train_starvla][INFO] - Step 1570, Loss: {'train/loss': 0.6469810009002686, 'train/grad_norm_pre_clip': 11.087852478027344, 'timing/data': 0.026384098455309868, 'timing/model': 1.0557696064934134, 'learning_rate/qwen_vl_interface': 8.97522907947665e-06, 'learning_rate/action_model': 8.975229079476649e-05, 'epoch': 0.57}) +[2026-06-17 19:20:33,221][starVLA.training.train_starvla][INFO] - Step 1571, Loss: {'train/loss': 0.827466607093811, 'train/grad_norm_pre_clip': 16.04332160949707, 'timing/data': 0.02896732185035944, 'timing/model': 1.077799511142075, 'learning_rate/qwen_vl_interface': 8.973886886870075e-06, 'learning_rate/action_model': 8.973886886870074e-05, 'epoch': 0.58}) +[2026-06-17 19:20:40,417][starVLA.training.train_starvla][INFO] - Step 1572, Loss: {'train/loss': 0.917513906955719, 'train/grad_norm_pre_clip': 13.607632637023926, 'timing/data': 0.021831811405718327, 'timing/model': 1.0575379971414804, 'learning_rate/qwen_vl_interface': 8.972543922297739e-06, 'learning_rate/action_model': 8.972543922297739e-05, 'epoch': 0.58}) +[2026-06-17 19:20:44,929][starVLA.training.train_starvla][INFO] - Step 1573, Loss: {'train/loss': 0.8782903552055359, 'train/grad_norm_pre_clip': 13.301880836486816, 'timing/data': 0.02091181930154562, 'timing/model': 1.0289668822661042, 'learning_rate/qwen_vl_interface': 8.971200186038046e-06, 'learning_rate/action_model': 8.971200186038045e-05, 'epoch': 0.58}) +[2026-06-17 19:20:48,987][starVLA.training.train_starvla][INFO] - Step 1574, Loss: {'train/loss': 0.5722538232803345, 'train/grad_norm_pre_clip': 9.931195259094238, 'timing/data': 0.02166130766272545, 'timing/model': 1.0558262718841434, 'learning_rate/qwen_vl_interface': 8.969855678369547e-06, 'learning_rate/action_model': 8.969855678369546e-05, 'epoch': 0.58}) +[2026-06-17 19:20:53,451][starVLA.training.train_starvla][INFO] - Step 1575, Loss: {'train/loss': 1.2956048250198364, 'train/grad_norm_pre_clip': 13.727704048156738, 'timing/data': 0.022363080643117428, 'timing/model': 1.1032684342935681, 'learning_rate/qwen_vl_interface': 8.968510399570964e-06, 'learning_rate/action_model': 8.968510399570965e-05, 'epoch': 0.58}) +[2026-06-17 19:20:57,806][starVLA.training.train_starvla][INFO] - Step 1576, Loss: {'train/loss': 0.9153268933296204, 'train/grad_norm_pre_clip': 88.90823364257812, 'timing/data': 0.021830706857144833, 'timing/model': 1.0636170515790582, 'learning_rate/qwen_vl_interface': 8.967164349921173e-06, 'learning_rate/action_model': 8.967164349921174e-05, 'epoch': 0.58}) +[2026-06-17 19:21:02,439][starVLA.training.train_starvla][INFO] - Step 1577, Loss: {'train/loss': 0.8952429890632629, 'train/grad_norm_pre_clip': 12.08688735961914, 'timing/data': 0.02190300729125738, 'timing/model': 1.0993754686787724, 'learning_rate/qwen_vl_interface': 8.965817529699216e-06, 'learning_rate/action_model': 8.965817529699215e-05, 'epoch': 0.58}) +[2026-06-17 19:21:07,032][starVLA.training.train_starvla][INFO] - Step 1578, Loss: {'train/loss': 1.1721227169036865, 'train/grad_norm_pre_clip': 15.557610511779785, 'timing/data': 0.02174554578959942, 'timing/model': 1.3661478478461504, 'learning_rate/qwen_vl_interface': 8.964469939184283e-06, 'learning_rate/action_model': 8.964469939184282e-05, 'epoch': 0.58}) +[2026-06-17 19:21:11,845][starVLA.training.train_starvla][INFO] - Step 1579, Loss: {'train/loss': 0.8325466513633728, 'train/grad_norm_pre_clip': 15.236441612243652, 'timing/data': 0.06768803019076586, 'timing/model': 1.5240496573969722, 'learning_rate/qwen_vl_interface': 8.963121578655736e-06, 'learning_rate/action_model': 8.963121578655736e-05, 'epoch': 0.58}) +[2026-06-17 19:21:15,946][starVLA.training.train_starvla][INFO] - Step 1580, Loss: {'train/loss': 0.6017863750457764, 'train/grad_norm_pre_clip': 12.225998878479004, 'timing/data': 0.02073372434824705, 'timing/model': 1.028901919722557, 'learning_rate/qwen_vl_interface': 8.96177244839309e-06, 'learning_rate/action_model': 8.961772448393092e-05, 'epoch': 0.58}) +[2026-06-17 19:21:20,427][starVLA.training.train_starvla][INFO] - Step 1581, Loss: {'train/loss': 0.796197235584259, 'train/grad_norm_pre_clip': 19.82257652282715, 'timing/data': 0.02151775173842907, 'timing/model': 1.1350276172161102, 'learning_rate/qwen_vl_interface': 8.960422548676024e-06, 'learning_rate/action_model': 8.960422548676022e-05, 'epoch': 0.58}) +[2026-06-17 19:21:25,116][starVLA.training.train_starvla][INFO] - Step 1582, Loss: {'train/loss': 0.7230079770088196, 'train/grad_norm_pre_clip': 10.318863868713379, 'timing/data': 0.021516202948987484, 'timing/model': 1.0959346303716302, 'learning_rate/qwen_vl_interface': 8.959071879784368e-06, 'learning_rate/action_model': 8.959071879784368e-05, 'epoch': 0.58}) +[2026-06-17 19:21:29,603][starVLA.training.train_starvla][INFO] - Step 1583, Loss: {'train/loss': 0.6483997106552124, 'train/grad_norm_pre_clip': 10.341242790222168, 'timing/data': 0.02861845400184393, 'timing/model': 1.0608337121084332, 'learning_rate/qwen_vl_interface': 8.95772044199812e-06, 'learning_rate/action_model': 8.95772044199812e-05, 'epoch': 0.58}) +[2026-06-17 19:21:37,200][starVLA.training.train_starvla][INFO] - Step 1584, Loss: {'train/loss': 0.6669144630432129, 'train/grad_norm_pre_clip': 13.666779518127441, 'timing/data': 0.022206200286746025, 'timing/model': 1.3821626994758844, 'learning_rate/qwen_vl_interface': 8.956368235597437e-06, 'learning_rate/action_model': 8.956368235597436e-05, 'epoch': 0.58}) +[2026-06-17 19:21:41,554][starVLA.training.train_starvla][INFO] - Step 1585, Loss: {'train/loss': 0.6728125214576721, 'train/grad_norm_pre_clip': 11.371964454650879, 'timing/data': 0.028226369060575962, 'timing/model': 1.1024366691708565, 'learning_rate/qwen_vl_interface': 8.955015260862627e-06, 'learning_rate/action_model': 8.955015260862627e-05, 'epoch': 0.58}) +[2026-06-17 19:21:45,978][starVLA.training.train_starvla][INFO] - Step 1586, Loss: {'train/loss': 0.6435917019844055, 'train/grad_norm_pre_clip': 11.212189674377441, 'timing/data': 0.019363832660019398, 'timing/model': 1.0207260949537158, 'learning_rate/qwen_vl_interface': 8.95366151807417e-06, 'learning_rate/action_model': 8.95366151807417e-05, 'epoch': 0.58}) +[2026-06-17 19:21:50,382][starVLA.training.train_starvla][INFO] - Step 1587, Loss: {'train/loss': 0.8897873759269714, 'train/grad_norm_pre_clip': 14.439849853515625, 'timing/data': 0.020738404244184494, 'timing/model': 1.0250127529725432, 'learning_rate/qwen_vl_interface': 8.952307007512694e-06, 'learning_rate/action_model': 8.952307007512694e-05, 'epoch': 0.58}) +[2026-06-17 19:21:55,087][starVLA.training.train_starvla][INFO] - Step 1588, Loss: {'train/loss': 1.132322907447815, 'train/grad_norm_pre_clip': 17.139385223388672, 'timing/data': 0.026575569063425064, 'timing/model': 1.1916103679686785, 'learning_rate/qwen_vl_interface': 8.950951729458991e-06, 'learning_rate/action_model': 8.950951729458991e-05, 'epoch': 0.58}) +[2026-06-17 19:21:59,680][starVLA.training.train_starvla][INFO] - Step 1589, Loss: {'train/loss': 0.9788693189620972, 'train/grad_norm_pre_clip': 12.106536865234375, 'timing/data': 0.025130835361778736, 'timing/model': 1.3452621512115002, 'learning_rate/qwen_vl_interface': 8.949595684194012e-06, 'learning_rate/action_model': 8.949595684194012e-05, 'epoch': 0.58}) +[2026-06-17 19:22:03,798][starVLA.training.train_starvla][INFO] - Step 1590, Loss: {'train/loss': 0.5296226143836975, 'train/grad_norm_pre_clip': 11.651960372924805, 'timing/data': 0.020694708451628685, 'timing/model': 1.0291575621813536, 'learning_rate/qwen_vl_interface': 8.948238871998866e-06, 'learning_rate/action_model': 8.948238871998866e-05, 'epoch': 0.58}) +[2026-06-17 19:22:08,288][starVLA.training.train_starvla][INFO] - Step 1591, Loss: {'train/loss': 0.8483760356903076, 'train/grad_norm_pre_clip': 10.54002571105957, 'timing/data': 0.02180574554949999, 'timing/model': 1.1684715738520026, 'learning_rate/qwen_vl_interface': 8.946881293154825e-06, 'learning_rate/action_model': 8.946881293154825e-05, 'epoch': 0.58}) +[2026-06-17 19:22:12,920][starVLA.training.train_starvla][INFO] - Step 1592, Loss: {'train/loss': 0.6161388754844666, 'train/grad_norm_pre_clip': 10.224026679992676, 'timing/data': 0.025127382948994637, 'timing/model': 1.0506416326388717, 'learning_rate/qwen_vl_interface': 8.945522947943312e-06, 'learning_rate/action_model': 8.945522947943313e-05, 'epoch': 0.58}) +[2026-06-17 19:22:17,327][starVLA.training.train_starvla][INFO] - Step 1593, Loss: {'train/loss': 0.5355005264282227, 'train/grad_norm_pre_clip': 10.367616653442383, 'timing/data': 0.022681154310703278, 'timing/model': 1.3556781150400639, 'learning_rate/qwen_vl_interface': 8.944163836645916e-06, 'learning_rate/action_model': 8.944163836645915e-05, 'epoch': 0.58}) +[2026-06-17 19:22:21,691][starVLA.training.train_starvla][INFO] - Step 1594, Loss: {'train/loss': 0.5893871784210205, 'train/grad_norm_pre_clip': 13.128668785095215, 'timing/data': 0.0220640255138278, 'timing/model': 1.3812420722097158, 'learning_rate/qwen_vl_interface': 8.942803959544381e-06, 'learning_rate/action_model': 8.942803959544381e-05, 'epoch': 0.58}) +[2026-06-17 19:22:25,928][starVLA.training.train_starvla][INFO] - Step 1595, Loss: {'train/loss': 1.0452349185943604, 'train/grad_norm_pre_clip': 15.719182014465332, 'timing/data': 0.020641903392970562, 'timing/model': 1.0326011488214135, 'learning_rate/qwen_vl_interface': 8.941443316920613e-06, 'learning_rate/action_model': 8.941443316920612e-05, 'epoch': 0.58}) +[2026-06-17 19:22:31,091][starVLA.training.train_starvla][INFO] - Step 1596, Loss: {'train/loss': 1.2056465148925781, 'train/grad_norm_pre_clip': 13.385054588317871, 'timing/data': 0.02674392331391573, 'timing/model': 1.8291054926812649, 'learning_rate/qwen_vl_interface': 8.940081909056671e-06, 'learning_rate/action_model': 8.940081909056671e-05, 'epoch': 0.58}) +[2026-06-17 19:22:35,552][starVLA.training.train_starvla][INFO] - Step 1597, Loss: {'train/loss': 0.6558667421340942, 'train/grad_norm_pre_clip': 10.112190246582031, 'timing/data': 0.020984736271202564, 'timing/model': 1.108665021136403, 'learning_rate/qwen_vl_interface': 8.93871973623478e-06, 'learning_rate/action_model': 8.93871973623478e-05, 'epoch': 0.58}) +[2026-06-17 19:22:40,303][starVLA.training.train_starvla][INFO] - Step 1598, Loss: {'train/loss': 0.8117072582244873, 'train/grad_norm_pre_clip': 11.054816246032715, 'timing/data': 0.023098547011613846, 'timing/model': 1.1137194065377116, 'learning_rate/qwen_vl_interface': 8.937356798737317e-06, 'learning_rate/action_model': 8.937356798737317e-05, 'epoch': 0.59}) +[2026-06-17 19:22:44,825][starVLA.training.train_starvla][INFO] - Step 1599, Loss: {'train/loss': 0.7460546493530273, 'train/grad_norm_pre_clip': 24.43157958984375, 'timing/data': 0.02158822864294052, 'timing/model': 1.363505246117711, 'learning_rate/qwen_vl_interface': 8.935993096846825e-06, 'learning_rate/action_model': 8.935993096846825e-05, 'epoch': 0.59}) +[2026-06-17 19:25:51,850][starVLA.training.train_starvla][INFO] - Step 1600, Loss: {'train/loss': 0.6899636387825012, 'train/grad_norm_pre_clip': 12.103705406188965, 'eval/loss': 0.8185341358184814, 'eval/latency_2/loss': 0.8185341358184814, 'eval/demon_fire_precision': 0.9062890462307371, 'eval/demon_fire_recall': 0.9104602510460251, 'eval/demon_fire_f1': 0.9083698601544563, 'eval/demon_fire_rate_model': 0.6680578742348359, 'eval/demon_fire_rate_teacher': 0.6649972175848636, 'eval/demon_left_precision': 0.7444237918215614, 'eval/demon_left_recall': 0.8431578947368421, 'eval/demon_left_f1': 0.7907206317867719, 'eval/demon_left_rate_model': 0.29938786867000555, 'eval/demon_left_rate_teacher': 0.2643294379521425, 'eval/demon_right_precision': 0.787515006002401, 'eval/demon_right_recall': 0.7412429378531074, 'eval/demon_right_f1': 0.7636786961583236, 'eval/demon_right_rate_model': 0.23177518085698387, 'eval/demon_right_rate_teacher': 0.24624373956594323, 'eval/demon_move_f1': 0.7771996639725478, 'eval/demon_cc_f1': 0.8645938707034914, 'eval/latency_2/demon_fire_precision': 0.9062890462307371, 'eval/latency_2/demon_fire_recall': 0.9104602510460251, 'eval/latency_2/demon_fire_f1': 0.9083698601544563, 'eval/latency_2/demon_fire_rate_model': 0.6680578742348359, 'eval/latency_2/demon_fire_rate_teacher': 0.6649972175848636, 'eval/latency_2/demon_left_precision': 0.7444237918215614, 'eval/latency_2/demon_left_recall': 0.8431578947368421, 'eval/latency_2/demon_left_f1': 0.7907206317867719, 'eval/latency_2/demon_left_rate_model': 0.29938786867000555, 'eval/latency_2/demon_left_rate_teacher': 0.2643294379521425, 'eval/latency_2/demon_right_precision': 0.787515006002401, 'eval/latency_2/demon_right_recall': 0.7412429378531074, 'eval/latency_2/demon_right_f1': 0.7636786961583236, 'eval/latency_2/demon_right_rate_model': 0.23177518085698387, 'eval/latency_2/demon_right_rate_teacher': 0.24624373956594323, 'eval/latency_2/demon_move_f1': 0.7771996639725478, 'eval/latency_2/demon_cc_f1': 0.8645938707034914, 'timing/data': 0.022055012173950672, 'timing/model': 1.0466136373579502, 'learning_rate/qwen_vl_interface': 8.934628630845994e-06, 'learning_rate/action_model': 8.934628630845995e-05, 'epoch': 0.59}) +[2026-06-17 19:25:51,851][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_1600_state +[2026-06-17 19:25:51,858][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 19:25:59,895][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_1600_state/model.safetensors +[2026-06-17 19:26:14,919][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_1600_state/optimizer.bin +[2026-06-17 19:26:14,923][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_1600_state/sampler.bin +[2026-06-17 19:26:14,923][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_1600_state/sampler_1.bin +[2026-06-17 19:26:14,926][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_1600_state/random_states_0.pkl +[2026-06-17 19:26:14,940][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1600 +[2026-06-17 19:26:14,940][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 19:26:14,962][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 19:26:19,246][starVLA.training.train_starvla][INFO] - Step 1601, Loss: {'train/loss': 0.7911773920059204, 'train/grad_norm_pre_clip': 13.004132270812988, 'timing/data': 0.024538240395486355, 'timing/model': 1.2390689998865128, 'learning_rate/qwen_vl_interface': 8.933263401017685e-06, 'learning_rate/action_model': 8.933263401017684e-05, 'epoch': 0.59}) +[2026-06-17 19:26:23,752][starVLA.training.train_starvla][INFO] - Step 1602, Loss: {'train/loss': 0.513919472694397, 'train/grad_norm_pre_clip': 10.04446029663086, 'timing/data': 0.02300079632550478, 'timing/model': 1.0424549393355846, 'learning_rate/qwen_vl_interface': 8.93189740764491e-06, 'learning_rate/action_model': 8.931897407644909e-05, 'epoch': 0.59}) +[2026-06-17 19:26:28,246][starVLA.training.train_starvla][INFO] - Step 1603, Loss: {'train/loss': 0.8594726324081421, 'train/grad_norm_pre_clip': 17.8432674407959, 'timing/data': 0.023297988809645176, 'timing/model': 1.044219865463674, 'learning_rate/qwen_vl_interface': 8.93053065101084e-06, 'learning_rate/action_model': 8.93053065101084e-05, 'epoch': 0.59}) +[2026-06-17 19:26:34,834][starVLA.training.train_starvla][INFO] - Step 1604, Loss: {'train/loss': 0.7610661387443542, 'train/grad_norm_pre_clip': 10.63732624053955, 'timing/data': 0.02467063721269369, 'timing/model': 3.315685810521245, 'learning_rate/qwen_vl_interface': 8.929163131398804e-06, 'learning_rate/action_model': 8.929163131398805e-05, 'epoch': 0.59}) +[2026-06-17 19:26:41,718][starVLA.training.train_starvla][INFO] - Step 1605, Loss: {'train/loss': 1.0235300064086914, 'train/grad_norm_pre_clip': 12.80614948272705, 'timing/data': 0.029736099764704704, 'timing/model': 1.4392619170248508, 'learning_rate/qwen_vl_interface': 8.927794849092294e-06, 'learning_rate/action_model': 8.927794849092293e-05, 'epoch': 0.59}) +[2026-06-17 19:26:46,124][starVLA.training.train_starvla][INFO] - Step 1606, Loss: {'train/loss': 0.5411748886108398, 'train/grad_norm_pre_clip': 10.199646949768066, 'timing/data': 0.028925884515047073, 'timing/model': 1.2015569200739264, 'learning_rate/qwen_vl_interface': 8.926425804374953e-06, 'learning_rate/action_model': 8.926425804374954e-05, 'epoch': 0.59}) +[2026-06-17 19:26:50,628][starVLA.training.train_starvla][INFO] - Step 1607, Loss: {'train/loss': 0.8769612908363342, 'train/grad_norm_pre_clip': 22.502363204956055, 'timing/data': 0.027213267982006073, 'timing/model': 1.0284824995324016, 'learning_rate/qwen_vl_interface': 8.925055997530587e-06, 'learning_rate/action_model': 8.925055997530587e-05, 'epoch': 0.59}) +[2026-06-17 19:26:57,821][starVLA.training.train_starvla][INFO] - Step 1608, Loss: {'train/loss': 1.155037522315979, 'train/grad_norm_pre_clip': 14.829754829406738, 'timing/data': 0.022454584017395973, 'timing/model': 1.0233096573501825, 'learning_rate/qwen_vl_interface': 8.92368542884316e-06, 'learning_rate/action_model': 8.923685428843159e-05, 'epoch': 0.59}) +[2026-06-17 19:27:03,162][starVLA.training.train_starvla][INFO] - Step 1609, Loss: {'train/loss': 0.7873626351356506, 'train/grad_norm_pre_clip': 13.10058879852295, 'timing/data': 0.02300423849374056, 'timing/model': 1.0607486888766289, 'learning_rate/qwen_vl_interface': 8.922314098596789e-06, 'learning_rate/action_model': 8.922314098596789e-05, 'epoch': 0.59}) +[2026-06-17 19:27:07,740][starVLA.training.train_starvla][INFO] - Step 1610, Loss: {'train/loss': 0.9261779189109802, 'train/grad_norm_pre_clip': 12.512801170349121, 'timing/data': 0.03168273065239191, 'timing/model': 1.5203724838793278, 'learning_rate/qwen_vl_interface': 8.920942007075752e-06, 'learning_rate/action_model': 8.920942007075752e-05, 'epoch': 0.59}) +[2026-06-17 19:27:12,218][starVLA.training.train_starvla][INFO] - Step 1611, Loss: {'train/loss': 1.0264484882354736, 'train/grad_norm_pre_clip': 13.765545845031738, 'timing/data': 0.025001753121614456, 'timing/model': 1.0994245875626802, 'learning_rate/qwen_vl_interface': 8.91956915456449e-06, 'learning_rate/action_model': 8.91956915456449e-05, 'epoch': 0.59}) +[2026-06-17 19:27:19,330][starVLA.training.train_starvla][INFO] - Step 1612, Loss: {'train/loss': 0.9333939552307129, 'train/grad_norm_pre_clip': 12.295635223388672, 'timing/data': 0.020216732285916805, 'timing/model': 1.0949736321344972, 'learning_rate/qwen_vl_interface': 8.918195541347593e-06, 'learning_rate/action_model': 8.918195541347593e-05, 'epoch': 0.59}) +[2026-06-17 19:27:23,771][starVLA.training.train_starvla][INFO] - Step 1613, Loss: {'train/loss': 0.7294255495071411, 'train/grad_norm_pre_clip': 11.3329439163208, 'timing/data': 0.02641908824443817, 'timing/model': 1.039837864227593, 'learning_rate/qwen_vl_interface': 8.916821167709811e-06, 'learning_rate/action_model': 8.916821167709812e-05, 'epoch': 0.59}) +[2026-06-17 19:27:28,239][starVLA.training.train_starvla][INFO] - Step 1614, Loss: {'train/loss': 1.1055476665496826, 'train/grad_norm_pre_clip': 15.135663986206055, 'timing/data': 0.027313029393553734, 'timing/model': 1.028191989287734, 'learning_rate/qwen_vl_interface': 8.915446033936057e-06, 'learning_rate/action_model': 8.915446033936056e-05, 'epoch': 0.59}) +[2026-06-17 19:27:32,840][starVLA.training.train_starvla][INFO] - Step 1615, Loss: {'train/loss': 0.8874590396881104, 'train/grad_norm_pre_clip': 12.607818603515625, 'timing/data': 0.02173919416964054, 'timing/model': 1.4709368534386158, 'learning_rate/qwen_vl_interface': 8.914070140311395e-06, 'learning_rate/action_model': 8.914070140311395e-05, 'epoch': 0.59}) +[2026-06-17 19:27:36,948][starVLA.training.train_starvla][INFO] - Step 1616, Loss: {'train/loss': 0.6461292505264282, 'train/grad_norm_pre_clip': 9.712628364562988, 'timing/data': 0.020455426536500454, 'timing/model': 1.030268277041614, 'learning_rate/qwen_vl_interface': 8.912693487121053e-06, 'learning_rate/action_model': 8.912693487121052e-05, 'epoch': 0.59}) +[2026-06-17 19:27:41,912][starVLA.training.train_starvla][INFO] - Step 1617, Loss: {'train/loss': 0.9523879289627075, 'train/grad_norm_pre_clip': 15.469643592834473, 'timing/data': 0.021463481709361076, 'timing/model': 1.0282739661633968, 'learning_rate/qwen_vl_interface': 8.911316074650407e-06, 'learning_rate/action_model': 8.911316074650406e-05, 'epoch': 0.59}) +[2026-06-17 19:27:48,705][starVLA.training.train_starvla][INFO] - Step 1618, Loss: {'train/loss': 0.8700529932975769, 'train/grad_norm_pre_clip': 11.991532325744629, 'timing/data': 0.021170993335545063, 'timing/model': 1.097211703658104, 'learning_rate/qwen_vl_interface': 8.909937903185e-06, 'learning_rate/action_model': 8.909937903185e-05, 'epoch': 0.59}) +[2026-06-17 19:27:53,475][starVLA.training.train_starvla][INFO] - Step 1619, Loss: {'train/loss': 0.7148924469947815, 'train/grad_norm_pre_clip': 10.342002868652344, 'timing/data': 0.030916293151676655, 'timing/model': 1.0378544861450791, 'learning_rate/qwen_vl_interface': 8.908558973010528e-06, 'learning_rate/action_model': 8.908558973010526e-05, 'epoch': 0.59}) +[2026-06-17 19:27:57,997][starVLA.training.train_starvla][INFO] - Step 1620, Loss: {'train/loss': 0.767073392868042, 'train/grad_norm_pre_clip': 14.662605285644531, 'timing/data': 0.021199091337621212, 'timing/model': 1.5223582107573748, 'learning_rate/qwen_vl_interface': 8.907179284412841e-06, 'learning_rate/action_model': 8.90717928441284e-05, 'epoch': 0.59}) +[2026-06-17 19:28:02,193][starVLA.training.train_starvla][INFO] - Step 1621, Loss: {'train/loss': 0.7039235830307007, 'train/grad_norm_pre_clip': 9.60749626159668, 'timing/data': 0.019969599321484566, 'timing/model': 1.0318982526659966, 'learning_rate/qwen_vl_interface': 8.905798837677956e-06, 'learning_rate/action_model': 8.905798837677956e-05, 'epoch': 0.59}) +[2026-06-17 19:28:06,580][starVLA.training.train_starvla][INFO] - Step 1622, Loss: {'train/loss': 0.7544372081756592, 'train/grad_norm_pre_clip': 13.116204261779785, 'timing/data': 0.02384006790816784, 'timing/model': 1.048720602877438, 'learning_rate/qwen_vl_interface': 8.904417633092037e-06, 'learning_rate/action_model': 8.904417633092037e-05, 'epoch': 0.59}) +[2026-06-17 19:28:10,905][starVLA.training.train_starvla][INFO] - Step 1623, Loss: {'train/loss': 0.8349637389183044, 'train/grad_norm_pre_clip': 13.448858261108398, 'timing/data': 0.020854776725172997, 'timing/model': 1.0606234790757298, 'learning_rate/qwen_vl_interface': 8.903035670941411e-06, 'learning_rate/action_model': 8.903035670941412e-05, 'epoch': 0.59}) +[2026-06-17 19:28:15,425][starVLA.training.train_starvla][INFO] - Step 1624, Loss: {'train/loss': 0.6644432544708252, 'train/grad_norm_pre_clip': 11.820111274719238, 'timing/data': 0.021967164240777493, 'timing/model': 1.1292329598218203, 'learning_rate/qwen_vl_interface': 8.901652951512559e-06, 'learning_rate/action_model': 8.90165295151256e-05, 'epoch': 0.59}) +[2026-06-17 19:28:19,994][starVLA.training.train_starvla][INFO] - Step 1625, Loss: {'train/loss': 0.7415863275527954, 'train/grad_norm_pre_clip': 10.21917724609375, 'timing/data': 0.020086548291146755, 'timing/model': 1.405574425123632, 'learning_rate/qwen_vl_interface': 8.900269475092122e-06, 'learning_rate/action_model': 8.900269475092122e-05, 'epoch': 0.6}) +[2026-06-17 19:28:24,110][starVLA.training.train_starvla][INFO] - Step 1626, Loss: {'train/loss': 0.8190487623214722, 'train/grad_norm_pre_clip': 10.528799057006836, 'timing/data': 0.028489409014582634, 'timing/model': 1.0797644583508372, 'learning_rate/qwen_vl_interface': 8.898885241966894e-06, 'learning_rate/action_model': 8.898885241966893e-05, 'epoch': 0.6}) +[2026-06-17 19:28:29,295][starVLA.training.train_starvla][INFO] - Step 1627, Loss: {'train/loss': 0.6925413608551025, 'train/grad_norm_pre_clip': 11.021937370300293, 'timing/data': 0.022147808223962784, 'timing/model': 1.7575566191226244, 'learning_rate/qwen_vl_interface': 8.89750025242383e-06, 'learning_rate/action_model': 8.89750025242383e-05, 'epoch': 0.6}) +[2026-06-17 19:28:33,811][starVLA.training.train_starvla][INFO] - Step 1628, Loss: {'train/loss': 0.6569419503211975, 'train/grad_norm_pre_clip': 11.526588439941406, 'timing/data': 0.03284956328570843, 'timing/model': 1.0554158678278327, 'learning_rate/qwen_vl_interface': 8.896114506750037e-06, 'learning_rate/action_model': 8.896114506750036e-05, 'epoch': 0.6}) +[2026-06-17 19:28:38,213][starVLA.training.train_starvla][INFO] - Step 1629, Loss: {'train/loss': 0.8259190917015076, 'train/grad_norm_pre_clip': 12.991575241088867, 'timing/data': 0.02104303054511547, 'timing/model': 1.098389495164156, 'learning_rate/qwen_vl_interface': 8.894728005232785e-06, 'learning_rate/action_model': 8.894728005232785e-05, 'epoch': 0.6}) +[2026-06-17 19:28:42,372][starVLA.training.train_starvla][INFO] - Step 1630, Loss: {'train/loss': 0.9057356715202332, 'train/grad_norm_pre_clip': 13.958431243896484, 'timing/data': 0.020913496613502502, 'timing/model': 1.0269908104091883, 'learning_rate/qwen_vl_interface': 8.893340748159494e-06, 'learning_rate/action_model': 8.893340748159493e-05, 'epoch': 0.6}) +[2026-06-17 19:28:46,806][starVLA.training.train_starvla][INFO] - Step 1631, Loss: {'train/loss': 0.7580442428588867, 'train/grad_norm_pre_clip': 14.790887832641602, 'timing/data': 0.027137959375977516, 'timing/model': 1.1180510055273771, 'learning_rate/qwen_vl_interface': 8.891952735817746e-06, 'learning_rate/action_model': 8.891952735817745e-05, 'epoch': 0.6}) +[2026-06-17 19:28:52,514][starVLA.training.train_starvla][INFO] - Step 1632, Loss: {'train/loss': 0.8481354713439941, 'train/grad_norm_pre_clip': 13.491140365600586, 'timing/data': 0.01962924748659134, 'timing/model': 1.0455936351791024, 'learning_rate/qwen_vl_interface': 8.890563968495278e-06, 'learning_rate/action_model': 8.890563968495278e-05, 'epoch': 0.6}) +[2026-06-17 19:29:03,723][starVLA.training.train_starvla][INFO] - Step 1633, Loss: {'train/loss': 0.6841814517974854, 'train/grad_norm_pre_clip': 11.105000495910645, 'timing/data': 0.02798561565577984, 'timing/model': 5.012929363176227, 'learning_rate/qwen_vl_interface': 8.88917444647998e-06, 'learning_rate/action_model': 8.889174446479981e-05, 'epoch': 0.6}) +[2026-06-17 19:29:08,188][starVLA.training.train_starvla][INFO] - Step 1634, Loss: {'train/loss': 0.7711678743362427, 'train/grad_norm_pre_clip': 11.29692268371582, 'timing/data': 0.021924955770373344, 'timing/model': 1.4189797202125192, 'learning_rate/qwen_vl_interface': 8.887784170059905e-06, 'learning_rate/action_model': 8.887784170059905e-05, 'epoch': 0.6}) +[2026-06-17 19:29:12,684][starVLA.training.train_starvla][INFO] - Step 1635, Loss: {'train/loss': 0.6862826943397522, 'train/grad_norm_pre_clip': 16.42254638671875, 'timing/data': 0.02068010065704584, 'timing/model': 1.317937906831503, 'learning_rate/qwen_vl_interface': 8.886393139523258e-06, 'learning_rate/action_model': 8.886393139523257e-05, 'epoch': 0.6}) +[2026-06-17 19:29:21,108][starVLA.training.train_starvla][INFO] - Step 1636, Loss: {'train/loss': 0.5692455172538757, 'train/grad_norm_pre_clip': 11.228482246398926, 'timing/data': 0.02662445418536663, 'timing/model': 2.626324627548456, 'learning_rate/qwen_vl_interface': 8.885001355158397e-06, 'learning_rate/action_model': 8.885001355158397e-05, 'epoch': 0.6}) +[2026-06-17 19:29:25,652][starVLA.training.train_starvla][INFO] - Step 1637, Loss: {'train/loss': 0.8161905407905579, 'train/grad_norm_pre_clip': 11.73520565032959, 'timing/data': 0.022080501541495323, 'timing/model': 1.0323467813432217, 'learning_rate/qwen_vl_interface': 8.883608817253847e-06, 'learning_rate/action_model': 8.883608817253847e-05, 'epoch': 0.6}) +[2026-06-17 19:29:30,366][starVLA.training.train_starvla][INFO] - Step 1638, Loss: {'train/loss': 0.8397438526153564, 'train/grad_norm_pre_clip': 13.737334251403809, 'timing/data': 0.02014781441539526, 'timing/model': 1.0852157827466726, 'learning_rate/qwen_vl_interface': 8.882215526098276e-06, 'learning_rate/action_model': 8.882215526098277e-05, 'epoch': 0.6}) +[2026-06-17 19:29:34,971][starVLA.training.train_starvla][INFO] - Step 1639, Loss: {'train/loss': 0.7605466842651367, 'train/grad_norm_pre_clip': 12.091571807861328, 'timing/data': 0.033314839005470276, 'timing/model': 1.1331355832517147, 'learning_rate/qwen_vl_interface': 8.880821481980521e-06, 'learning_rate/action_model': 8.880821481980521e-05, 'epoch': 0.6}) +[2026-06-17 19:29:39,523][starVLA.training.train_starvla][INFO] - Step 1640, Loss: {'train/loss': 0.7557788491249084, 'train/grad_norm_pre_clip': 11.738189697265625, 'timing/data': 0.020556493662297726, 'timing/model': 1.3353965962305665, 'learning_rate/qwen_vl_interface': 8.879426685189565e-06, 'learning_rate/action_model': 8.879426685189565e-05, 'epoch': 0.6}) +[2026-06-17 19:29:43,834][starVLA.training.train_starvla][INFO] - Step 1641, Loss: {'train/loss': 1.068162202835083, 'train/grad_norm_pre_clip': 14.6605806350708, 'timing/data': 0.02122611179947853, 'timing/model': 1.1883823554962873, 'learning_rate/qwen_vl_interface': 8.878031136014552e-06, 'learning_rate/action_model': 8.878031136014551e-05, 'epoch': 0.6}) +[2026-06-17 19:29:48,213][starVLA.training.train_starvla][INFO] - Step 1642, Loss: {'train/loss': 0.5990281701087952, 'train/grad_norm_pre_clip': 12.065337181091309, 'timing/data': 0.022722957655787468, 'timing/model': 1.0485587306320667, 'learning_rate/qwen_vl_interface': 8.87663483474478e-06, 'learning_rate/action_model': 8.87663483474478e-05, 'epoch': 0.6}) +[2026-06-17 19:29:54,310][starVLA.training.train_starvla][INFO] - Step 1643, Loss: {'train/loss': 0.6266099810600281, 'train/grad_norm_pre_clip': 24.86519432067871, 'timing/data': 0.020135038532316685, 'timing/model': 1.2213693344965577, 'learning_rate/qwen_vl_interface': 8.875237781669704e-06, 'learning_rate/action_model': 8.875237781669703e-05, 'epoch': 0.6}) +[2026-06-17 19:29:58,988][starVLA.training.train_starvla][INFO] - Step 1644, Loss: {'train/loss': 0.8256840705871582, 'train/grad_norm_pre_clip': 13.113378524780273, 'timing/data': 0.025530003011226654, 'timing/model': 1.1488320315256715, 'learning_rate/qwen_vl_interface': 8.873839977078933e-06, 'learning_rate/action_model': 8.873839977078934e-05, 'epoch': 0.6}) +[2026-06-17 19:30:03,485][starVLA.training.train_starvla][INFO] - Step 1645, Loss: {'train/loss': 0.6901575326919556, 'train/grad_norm_pre_clip': 12.198992729187012, 'timing/data': 0.025086717680096626, 'timing/model': 1.3538675289601088, 'learning_rate/qwen_vl_interface': 8.87244142126224e-06, 'learning_rate/action_model': 8.87244142126224e-05, 'epoch': 0.6}) +[2026-06-17 19:30:07,661][starVLA.training.train_starvla][INFO] - Step 1646, Loss: {'train/loss': 0.4857994019985199, 'train/grad_norm_pre_clip': 10.626646041870117, 'timing/data': 0.024078761227428913, 'timing/model': 1.0360444020479918, 'learning_rate/qwen_vl_interface': 8.871042114509539e-06, 'learning_rate/action_model': 8.87104211450954e-05, 'epoch': 0.6}) +[2026-06-17 19:30:12,188][starVLA.training.train_starvla][INFO] - Step 1647, Loss: {'train/loss': 0.6107779741287231, 'train/grad_norm_pre_clip': 23.74482536315918, 'timing/data': 0.025043511763215065, 'timing/model': 1.0866138096898794, 'learning_rate/qwen_vl_interface': 8.869642057110912e-06, 'learning_rate/action_model': 8.869642057110911e-05, 'epoch': 0.6}) +[2026-06-17 19:30:16,585][starVLA.training.train_starvla][INFO] - Step 1648, Loss: {'train/loss': 0.9444153904914856, 'train/grad_norm_pre_clip': 14.25501537322998, 'timing/data': 0.021939929574728012, 'timing/model': 1.0464320359751582, 'learning_rate/qwen_vl_interface': 8.868241249356592e-06, 'learning_rate/action_model': 8.868241249356592e-05, 'epoch': 0.6}) +[2026-06-17 19:30:21,060][starVLA.training.train_starvla][INFO] - Step 1649, Loss: {'train/loss': 0.5418545603752136, 'train/grad_norm_pre_clip': 10.67920207977295, 'timing/data': 0.021203389391303062, 'timing/model': 1.0178885757923126, 'learning_rate/qwen_vl_interface': 8.866839691536968e-06, 'learning_rate/action_model': 8.866839691536968e-05, 'epoch': 0.6}) +[2026-06-17 19:30:25,580][starVLA.training.train_starvla][INFO] - Step 1650, Loss: {'train/loss': 0.9758944511413574, 'train/grad_norm_pre_clip': 12.252483367919922, 'timing/data': 0.02016430627554655, 'timing/model': 1.3410733928903937, 'learning_rate/qwen_vl_interface': 8.865437383942584e-06, 'learning_rate/action_model': 8.865437383942583e-05, 'epoch': 0.6}) +[2026-06-17 19:30:29,725][starVLA.training.train_starvla][INFO] - Step 1651, Loss: {'train/loss': 0.697872519493103, 'train/grad_norm_pre_clip': 11.632469177246094, 'timing/data': 0.02045429591089487, 'timing/model': 1.0260993437841535, 'learning_rate/qwen_vl_interface': 8.86403432686414e-06, 'learning_rate/action_model': 8.86403432686414e-05, 'epoch': 0.6}) +[2026-06-17 19:30:34,195][starVLA.training.train_starvla][INFO] - Step 1652, Loss: {'train/loss': 0.5960620045661926, 'train/grad_norm_pre_clip': 9.193955421447754, 'timing/data': 0.023011604323983192, 'timing/model': 1.122805212624371, 'learning_rate/qwen_vl_interface': 8.86263052059249e-06, 'learning_rate/action_model': 8.862630520592489e-05, 'epoch': 0.6}) +[2026-06-17 19:30:38,616][starVLA.training.train_starvla][INFO] - Step 1653, Loss: {'train/loss': 0.6066881418228149, 'train/grad_norm_pre_clip': 14.955896377563477, 'timing/data': 0.026982413604855537, 'timing/model': 1.037676239386201, 'learning_rate/qwen_vl_interface': 8.861225965418647e-06, 'learning_rate/action_model': 8.861225965418646e-05, 'epoch': 0.61}) +[2026-06-17 19:30:45,885][starVLA.training.train_starvla][INFO] - Step 1654, Loss: {'train/loss': 0.8935036659240723, 'train/grad_norm_pre_clip': 23.078258514404297, 'timing/data': 0.02431060839444399, 'timing/model': 1.0409463960677385, 'learning_rate/qwen_vl_interface': 8.859820661633775e-06, 'learning_rate/action_model': 8.859820661633775e-05, 'epoch': 0.61}) +[2026-06-17 19:30:56,315][starVLA.training.train_starvla][INFO] - Step 1655, Loss: {'train/loss': 0.8259588479995728, 'train/grad_norm_pre_clip': 14.27446174621582, 'timing/data': 0.021141979843378067, 'timing/model': 3.5228803930804133, 'learning_rate/qwen_vl_interface': 8.858414609529193e-06, 'learning_rate/action_model': 8.858414609529193e-05, 'epoch': 0.61}) +[2026-06-17 19:31:00,493][starVLA.training.train_starvla][INFO] - Step 1656, Loss: {'train/loss': 0.9494128823280334, 'train/grad_norm_pre_clip': 15.721809387207031, 'timing/data': 0.02459335047751665, 'timing/model': 1.045791975222528, 'learning_rate/qwen_vl_interface': 8.857007809396384e-06, 'learning_rate/action_model': 8.857007809396383e-05, 'epoch': 0.61}) +[2026-06-17 19:31:07,779][starVLA.training.train_starvla][INFO] - Step 1657, Loss: {'train/loss': 0.7143293619155884, 'train/grad_norm_pre_clip': 12.85887622833252, 'timing/data': 0.02234673500061035, 'timing/model': 1.3176645142957568, 'learning_rate/qwen_vl_interface': 8.855600261526973e-06, 'learning_rate/action_model': 8.855600261526973e-05, 'epoch': 0.61}) +[2026-06-17 19:31:12,418][starVLA.training.train_starvla][INFO] - Step 1658, Loss: {'train/loss': 0.8779585361480713, 'train/grad_norm_pre_clip': 18.065767288208008, 'timing/data': 0.02272822242230177, 'timing/model': 1.1063767774030566, 'learning_rate/qwen_vl_interface': 8.854191966212747e-06, 'learning_rate/action_model': 8.854191966212748e-05, 'epoch': 0.61}) +[2026-06-17 19:31:16,794][starVLA.training.train_starvla][INFO] - Step 1659, Loss: {'train/loss': 0.5280014872550964, 'train/grad_norm_pre_clip': 11.404546737670898, 'timing/data': 0.020291980355978012, 'timing/model': 1.0285955742001534, 'learning_rate/qwen_vl_interface': 8.852782923745648e-06, 'learning_rate/action_model': 8.852782923745648e-05, 'epoch': 0.61}) +[2026-06-17 19:31:21,153][starVLA.training.train_starvla][INFO] - Step 1660, Loss: {'train/loss': 0.4433911144733429, 'train/grad_norm_pre_clip': 8.461291313171387, 'timing/data': 0.021492660976946354, 'timing/model': 1.3314437111839652, 'learning_rate/qwen_vl_interface': 8.851373134417773e-06, 'learning_rate/action_model': 8.851373134417773e-05, 'epoch': 0.61}) +[2026-06-17 19:31:25,375][starVLA.training.train_starvla][INFO] - Step 1661, Loss: {'train/loss': 1.0443555116653442, 'train/grad_norm_pre_clip': 14.327853202819824, 'timing/data': 0.02221635077148676, 'timing/model': 1.0233731986954808, 'learning_rate/qwen_vl_interface': 8.849962598521373e-06, 'learning_rate/action_model': 8.849962598521373e-05, 'epoch': 0.61}) +[2026-06-17 19:31:29,883][starVLA.training.train_starvla][INFO] - Step 1662, Loss: {'train/loss': 0.7175456285476685, 'train/grad_norm_pre_clip': 12.726357460021973, 'timing/data': 0.03544791508466005, 'timing/model': 1.1154401358217, 'learning_rate/qwen_vl_interface': 8.848551316348852e-06, 'learning_rate/action_model': 8.848551316348853e-05, 'epoch': 0.61}) +[2026-06-17 19:31:34,741][starVLA.training.train_starvla][INFO] - Step 1663, Loss: {'train/loss': 1.0772345066070557, 'train/grad_norm_pre_clip': 15.925111770629883, 'timing/data': 0.028233994729816914, 'timing/model': 1.210888740606606, 'learning_rate/qwen_vl_interface': 8.847139288192771e-06, 'learning_rate/action_model': 8.84713928819277e-05, 'epoch': 0.61}) +[2026-06-17 19:31:39,315][starVLA.training.train_starvla][INFO] - Step 1664, Loss: {'train/loss': 0.7614436745643616, 'train/grad_norm_pre_clip': 11.914738655090332, 'timing/data': 0.025388801470398903, 'timing/model': 1.0992968836799264, 'learning_rate/qwen_vl_interface': 8.845726514345845e-06, 'learning_rate/action_model': 8.845726514345843e-05, 'epoch': 0.61}) +[2026-06-17 19:31:43,373][starVLA.training.train_starvla][INFO] - Step 1665, Loss: {'train/loss': 0.720745861530304, 'train/grad_norm_pre_clip': 14.694055557250977, 'timing/data': 0.022777743637561798, 'timing/model': 1.0534298485144973, 'learning_rate/qwen_vl_interface': 8.844312995100943e-06, 'learning_rate/action_model': 8.844312995100942e-05, 'epoch': 0.61}) +[2026-06-17 19:31:47,758][starVLA.training.train_starvla][INFO] - Step 1666, Loss: {'train/loss': 0.8108026385307312, 'train/grad_norm_pre_clip': 13.804866790771484, 'timing/data': 0.02189776860177517, 'timing/model': 1.0341011136770248, 'learning_rate/qwen_vl_interface': 8.842898730751088e-06, 'learning_rate/action_model': 8.842898730751087e-05, 'epoch': 0.61}) +[2026-06-17 19:31:52,247][starVLA.training.train_starvla][INFO] - Step 1667, Loss: {'train/loss': 0.5755285024642944, 'train/grad_norm_pre_clip': 12.368972778320312, 'timing/data': 0.03270103316754103, 'timing/model': 1.0967023018747568, 'learning_rate/qwen_vl_interface': 8.841483721589462e-06, 'learning_rate/action_model': 8.84148372158946e-05, 'epoch': 0.61}) +[2026-06-17 19:31:56,706][starVLA.training.train_starvla][INFO] - Step 1668, Loss: {'train/loss': 0.5789861679077148, 'train/grad_norm_pre_clip': 11.287018775939941, 'timing/data': 0.02147568203508854, 'timing/model': 1.0363919669762254, 'learning_rate/qwen_vl_interface': 8.840067967909396e-06, 'learning_rate/action_model': 8.840067967909394e-05, 'epoch': 0.61}) +[2026-06-17 19:32:01,273][starVLA.training.train_starvla][INFO] - Step 1669, Loss: {'train/loss': 0.6471482515335083, 'train/grad_norm_pre_clip': 13.652498245239258, 'timing/data': 0.036388056352734566, 'timing/model': 1.4406160740181804, 'learning_rate/qwen_vl_interface': 8.838651470004376e-06, 'learning_rate/action_model': 8.838651470004375e-05, 'epoch': 0.61}) +[2026-06-17 19:32:05,801][starVLA.training.train_starvla][INFO] - Step 1670, Loss: {'train/loss': 0.870342493057251, 'train/grad_norm_pre_clip': 13.493043899536133, 'timing/data': 0.026041343808174133, 'timing/model': 1.2974331378936768, 'learning_rate/qwen_vl_interface': 8.837234228168045e-06, 'learning_rate/action_model': 8.837234228168044e-05, 'epoch': 0.61}) +[2026-06-17 19:32:11,777][starVLA.training.train_starvla][INFO] - Step 1671, Loss: {'train/loss': 0.9429859519004822, 'train/grad_norm_pre_clip': 15.701787948608398, 'timing/data': 0.02430069539695978, 'timing/model': 1.0598763078451157, 'learning_rate/qwen_vl_interface': 8.835816242694197e-06, 'learning_rate/action_model': 8.835816242694197e-05, 'epoch': 0.61}) +[2026-06-17 19:32:16,263][starVLA.training.train_starvla][INFO] - Step 1672, Loss: {'train/loss': 1.107602596282959, 'train/grad_norm_pre_clip': 16.43222999572754, 'timing/data': 0.02718986663967371, 'timing/model': 1.1270153960213065, 'learning_rate/qwen_vl_interface': 8.834397513876786e-06, 'learning_rate/action_model': 8.834397513876785e-05, 'epoch': 0.61}) +[2026-06-17 19:32:20,646][starVLA.training.train_starvla][INFO] - Step 1673, Loss: {'train/loss': 0.6443745493888855, 'train/grad_norm_pre_clip': 13.438790321350098, 'timing/data': 0.02387729100883007, 'timing/model': 1.0388913424685597, 'learning_rate/qwen_vl_interface': 8.832978042009913e-06, 'learning_rate/action_model': 8.832978042009912e-05, 'epoch': 0.61}) +[2026-06-17 19:32:25,115][starVLA.training.train_starvla][INFO] - Step 1674, Loss: {'train/loss': 0.8052170276641846, 'train/grad_norm_pre_clip': 14.129961967468262, 'timing/data': 0.03592708520591259, 'timing/model': 1.1276216404512525, 'learning_rate/qwen_vl_interface': 8.831557827387835e-06, 'learning_rate/action_model': 8.831557827387834e-05, 'epoch': 0.61}) +[2026-06-17 19:32:29,562][starVLA.training.train_starvla][INFO] - Step 1675, Loss: {'train/loss': 0.5597712993621826, 'train/grad_norm_pre_clip': 12.406609535217285, 'timing/data': 0.022441954351961613, 'timing/model': 1.3506991881877184, 'learning_rate/qwen_vl_interface': 8.830136870304966e-06, 'learning_rate/action_model': 8.830136870304965e-05, 'epoch': 0.61}) +[2026-06-17 19:32:34,656][starVLA.training.train_starvla][INFO] - Step 1676, Loss: {'train/loss': 0.6761451363563538, 'train/grad_norm_pre_clip': 11.146389961242676, 'timing/data': 0.02280537039041519, 'timing/model': 1.0488852299749851, 'learning_rate/qwen_vl_interface': 8.828715171055871e-06, 'learning_rate/action_model': 8.828715171055871e-05, 'epoch': 0.61}) +[2026-06-17 19:32:39,217][starVLA.training.train_starvla][INFO] - Step 1677, Loss: {'train/loss': 1.0139186382293701, 'train/grad_norm_pre_clip': 17.024555206298828, 'timing/data': 0.0214053550735116, 'timing/model': 1.036053916439414, 'learning_rate/qwen_vl_interface': 8.827292729935271e-06, 'learning_rate/action_model': 8.827292729935271e-05, 'epoch': 0.61}) +[2026-06-17 19:32:43,704][starVLA.training.train_starvla][INFO] - Step 1678, Loss: {'train/loss': 0.8416684865951538, 'train/grad_norm_pre_clip': 13.650146484375, 'timing/data': 0.023507495410740376, 'timing/model': 1.056703943759203, 'learning_rate/qwen_vl_interface': 8.825869547238038e-06, 'learning_rate/action_model': 8.825869547238038e-05, 'epoch': 0.61}) +[2026-06-17 19:32:48,073][starVLA.training.train_starvla][INFO] - Step 1679, Loss: {'train/loss': 0.7989048361778259, 'train/grad_norm_pre_clip': 12.891388893127441, 'timing/data': 0.022279830649495125, 'timing/model': 1.0349666569381952, 'learning_rate/qwen_vl_interface': 8.824445623259202e-06, 'learning_rate/action_model': 8.8244456232592e-05, 'epoch': 0.61}) +[2026-06-17 19:32:55,665][starVLA.training.train_starvla][INFO] - Step 1680, Loss: {'train/loss': 0.5737625360488892, 'train/grad_norm_pre_clip': 10.485382080078125, 'timing/data': 0.02474383357912302, 'timing/model': 1.4588261805474758, 'learning_rate/qwen_vl_interface': 8.82302095829394e-06, 'learning_rate/action_model': 8.82302095829394e-05, 'epoch': 0.62}) +[2026-06-17 19:32:59,970][starVLA.training.train_starvla][INFO] - Step 1681, Loss: {'train/loss': 0.7906075119972229, 'train/grad_norm_pre_clip': 13.900525093078613, 'timing/data': 0.020917372778058052, 'timing/model': 1.0533615127205849, 'learning_rate/qwen_vl_interface': 8.82159555263759e-06, 'learning_rate/action_model': 8.821595552637589e-05, 'epoch': 0.62}) +[2026-06-17 19:33:07,295][starVLA.training.train_starvla][INFO] - Step 1682, Loss: {'train/loss': 0.668530285358429, 'train/grad_norm_pre_clip': 10.835278511047363, 'timing/data': 0.03108085412532091, 'timing/model': 1.0549363028258085, 'learning_rate/qwen_vl_interface': 8.820169406585638e-06, 'learning_rate/action_model': 8.820169406585637e-05, 'epoch': 0.62}) +[2026-06-17 19:33:11,794][starVLA.training.train_starvla][INFO] - Step 1683, Loss: {'train/loss': 0.541111171245575, 'train/grad_norm_pre_clip': 10.028279304504395, 'timing/data': 0.021754475310444832, 'timing/model': 1.0751098403707147, 'learning_rate/qwen_vl_interface': 8.818742520433727e-06, 'learning_rate/action_model': 8.818742520433726e-05, 'epoch': 0.62}) +[2026-06-17 19:33:18,533][starVLA.training.train_starvla][INFO] - Step 1684, Loss: {'train/loss': 0.6523945331573486, 'train/grad_norm_pre_clip': 14.494497299194336, 'timing/data': 0.022715953178703785, 'timing/model': 1.118691404350102, 'learning_rate/qwen_vl_interface': 8.81731489447765e-06, 'learning_rate/action_model': 8.81731489447765e-05, 'epoch': 0.62}) +[2026-06-17 19:33:22,836][starVLA.training.train_starvla][INFO] - Step 1685, Loss: {'train/loss': 0.8628342747688293, 'train/grad_norm_pre_clip': 14.938234329223633, 'timing/data': 0.020908672362565994, 'timing/model': 1.3355530481785536, 'learning_rate/qwen_vl_interface': 8.815886529013356e-06, 'learning_rate/action_model': 8.815886529013356e-05, 'epoch': 0.62}) +[2026-06-17 19:33:30,242][starVLA.training.train_starvla][INFO] - Step 1686, Loss: {'train/loss': 0.6151569485664368, 'train/grad_norm_pre_clip': 12.310032844543457, 'timing/data': 0.023408202454447746, 'timing/model': 1.0627667773514986, 'learning_rate/qwen_vl_interface': 8.814457424336951e-06, 'learning_rate/action_model': 8.81445742433695e-05, 'epoch': 0.62}) +[2026-06-17 19:33:34,837][starVLA.training.train_starvla][INFO] - Step 1687, Loss: {'train/loss': 0.9453202486038208, 'train/grad_norm_pre_clip': 15.73455810546875, 'timing/data': 0.021721210330724716, 'timing/model': 1.0395235111936927, 'learning_rate/qwen_vl_interface': 8.813027580744685e-06, 'learning_rate/action_model': 8.813027580744684e-05, 'epoch': 0.62}) +[2026-06-17 19:33:39,210][starVLA.training.train_starvla][INFO] - Step 1688, Loss: {'train/loss': 0.6550833582878113, 'train/grad_norm_pre_clip': 11.00476360321045, 'timing/data': 0.024173266254365444, 'timing/model': 1.0389342736452818, 'learning_rate/qwen_vl_interface': 8.811596998532965e-06, 'learning_rate/action_model': 8.811596998532965e-05, 'epoch': 0.62}) +[2026-06-17 19:33:43,622][starVLA.training.train_starvla][INFO] - Step 1689, Loss: {'train/loss': 0.9180960059165955, 'train/grad_norm_pre_clip': 14.796453475952148, 'timing/data': 0.020947893150150776, 'timing/model': 1.0254437988623977, 'learning_rate/qwen_vl_interface': 8.810165677998356e-06, 'learning_rate/action_model': 8.810165677998356e-05, 'epoch': 0.62}) +[2026-06-17 19:33:48,257][starVLA.training.train_starvla][INFO] - Step 1690, Loss: {'train/loss': 0.9831990599632263, 'train/grad_norm_pre_clip': 13.3004150390625, 'timing/data': 0.024731026962399483, 'timing/model': 1.4984099147841334, 'learning_rate/qwen_vl_interface': 8.808733619437572e-06, 'learning_rate/action_model': 8.808733619437572e-05, 'epoch': 0.62}) +[2026-06-17 19:33:52,340][starVLA.training.train_starvla][INFO] - Step 1691, Loss: {'train/loss': 0.5316370129585266, 'train/grad_norm_pre_clip': 11.040224075317383, 'timing/data': 0.021861235611140728, 'timing/model': 1.0312133934348822, 'learning_rate/qwen_vl_interface': 8.80730082314748e-06, 'learning_rate/action_model': 8.80730082314748e-05, 'epoch': 0.62}) +[2026-06-17 19:33:57,721][starVLA.training.train_starvla][INFO] - Step 1692, Loss: {'train/loss': 1.0103709697723389, 'train/grad_norm_pre_clip': 25.40291976928711, 'timing/data': 0.02384176291525364, 'timing/model': 1.053779543377459, 'learning_rate/qwen_vl_interface': 8.805867289425098e-06, 'learning_rate/action_model': 8.805867289425098e-05, 'epoch': 0.62}) +[2026-06-17 19:34:03,970][starVLA.training.train_starvla][INFO] - Step 1693, Loss: {'train/loss': 0.8888933062553406, 'train/grad_norm_pre_clip': 13.24434757232666, 'timing/data': 0.026858951896429062, 'timing/model': 1.1314350049942732, 'learning_rate/qwen_vl_interface': 8.804433018567601e-06, 'learning_rate/action_model': 8.8044330185676e-05, 'epoch': 0.62}) +[2026-06-17 19:34:08,432][starVLA.training.train_starvla][INFO] - Step 1694, Loss: {'train/loss': 0.5077512860298157, 'train/grad_norm_pre_clip': 10.522377014160156, 'timing/data': 0.02264945674687624, 'timing/model': 1.0895780259743333, 'learning_rate/qwen_vl_interface': 8.802998010872316e-06, 'learning_rate/action_model': 8.802998010872316e-05, 'epoch': 0.62}) +[2026-06-17 19:34:12,791][starVLA.training.train_starvla][INFO] - Step 1695, Loss: {'train/loss': 0.7407088875770569, 'train/grad_norm_pre_clip': 14.916669845581055, 'timing/data': 0.02317999117076397, 'timing/model': 1.3595479438081384, 'learning_rate/qwen_vl_interface': 8.801562266636719e-06, 'learning_rate/action_model': 8.801562266636719e-05, 'epoch': 0.62}) +[2026-06-17 19:34:16,826][starVLA.training.train_starvla][INFO] - Step 1696, Loss: {'train/loss': 0.9859091639518738, 'train/grad_norm_pre_clip': 13.677783012390137, 'timing/data': 0.02058990951627493, 'timing/model': 1.023378225043416, 'learning_rate/qwen_vl_interface': 8.800125786158442e-06, 'learning_rate/action_model': 8.800125786158442e-05, 'epoch': 0.62}) +[2026-06-17 19:34:21,294][starVLA.training.train_starvla][INFO] - Step 1697, Loss: {'train/loss': 1.0042178630828857, 'train/grad_norm_pre_clip': 83.65347290039062, 'timing/data': 0.021557116881012917, 'timing/model': 1.0445155650377274, 'learning_rate/qwen_vl_interface': 8.798688569735268e-06, 'learning_rate/action_model': 8.798688569735268e-05, 'epoch': 0.62}) +[2026-06-17 19:34:25,800][starVLA.training.train_starvla][INFO] - Step 1698, Loss: {'train/loss': 0.9758726954460144, 'train/grad_norm_pre_clip': 14.344839096069336, 'timing/data': 0.023942173458635807, 'timing/model': 1.147967902943492, 'learning_rate/qwen_vl_interface': 8.797250617665139e-06, 'learning_rate/action_model': 8.797250617665138e-05, 'epoch': 0.62}) +[2026-06-17 19:34:30,191][starVLA.training.train_starvla][INFO] - Step 1699, Loss: {'train/loss': 0.816676914691925, 'train/grad_norm_pre_clip': 15.83871841430664, 'timing/data': 0.02556560281664133, 'timing/model': 1.0295026395469904, 'learning_rate/qwen_vl_interface': 8.795811930246136e-06, 'learning_rate/action_model': 8.795811930246136e-05, 'epoch': 0.62}) +[2026-06-17 19:37:39,711][starVLA.training.train_starvla][INFO] - Step 1700, Loss: {'train/loss': 0.9621554017066956, 'train/grad_norm_pre_clip': 11.594625473022461, 'eval/loss': 0.8432295322418213, 'eval/latency_2/loss': 0.8432295322418213, 'eval/demon_fire_precision': 0.8647081561654426, 'eval/demon_fire_recall': 0.9359832635983264, 'eval/demon_fire_f1': 0.8989351014667469, 'eval/demon_fire_rate_model': 0.719810795770729, 'eval/demon_fire_rate_teacher': 0.6649972175848636, 'eval/demon_left_precision': 0.8043478260869565, 'eval/demon_left_recall': 0.6231578947368421, 'eval/demon_left_f1': 0.7022538552787664, 'eval/demon_left_rate_model': 0.20478575403450194, 'eval/demon_left_rate_teacher': 0.2643294379521425, 'eval/demon_right_precision': 0.7519908987485779, 'eval/demon_right_recall': 0.7468926553672316, 'eval/demon_right_f1': 0.7494331065759638, 'eval/demon_right_rate_model': 0.24457429048414023, 'eval/demon_right_rate_teacher': 0.24624373956594323, 'eval/demon_move_f1': 0.7258434809273651, 'eval/demon_cc_f1': 0.8389181530114936, 'eval/latency_2/demon_fire_precision': 0.8647081561654426, 'eval/latency_2/demon_fire_recall': 0.9359832635983264, 'eval/latency_2/demon_fire_f1': 0.8989351014667469, 'eval/latency_2/demon_fire_rate_model': 0.719810795770729, 'eval/latency_2/demon_fire_rate_teacher': 0.6649972175848636, 'eval/latency_2/demon_left_precision': 0.8043478260869565, 'eval/latency_2/demon_left_recall': 0.6231578947368421, 'eval/latency_2/demon_left_f1': 0.7022538552787664, 'eval/latency_2/demon_left_rate_model': 0.20478575403450194, 'eval/latency_2/demon_left_rate_teacher': 0.2643294379521425, 'eval/latency_2/demon_right_precision': 0.7519908987485779, 'eval/latency_2/demon_right_recall': 0.7468926553672316, 'eval/latency_2/demon_right_f1': 0.7494331065759638, 'eval/latency_2/demon_right_rate_model': 0.24457429048414023, 'eval/latency_2/demon_right_rate_teacher': 0.24624373956594323, 'eval/latency_2/demon_move_f1': 0.7258434809273651, 'eval/latency_2/demon_cc_f1': 0.8389181530114936, 'timing/data': 0.02251610904932022, 'timing/model': 1.0565841989591718, 'learning_rate/qwen_vl_interface': 8.794372507776508e-06, 'learning_rate/action_model': 8.794372507776508e-05, 'epoch': 0.62}) +[2026-06-17 19:37:39,727][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_1700_state +[2026-06-17 19:37:39,735][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 19:37:48,415][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_1700_state/model.safetensors +[2026-06-17 19:38:02,902][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_1700_state/optimizer.bin +[2026-06-17 19:38:02,903][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_1700_state/sampler.bin +[2026-06-17 19:38:02,903][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_1700_state/sampler_1.bin +[2026-06-17 19:38:02,906][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_1700_state/random_states_0.pkl +[2026-06-17 19:38:02,910][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1700 +[2026-06-17 19:38:02,910][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 19:38:02,926][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 19:38:07,454][starVLA.training.train_starvla][INFO] - Step 1701, Loss: {'train/loss': 1.0465116500854492, 'train/grad_norm_pre_clip': 18.00114631652832, 'timing/data': 0.02507737185806036, 'timing/model': 1.4663170659914613, 'learning_rate/qwen_vl_interface': 8.792932350554642e-06, 'learning_rate/action_model': 8.792932350554643e-05, 'epoch': 0.62}) +[2026-06-17 19:38:11,507][starVLA.training.train_starvla][INFO] - Step 1702, Loss: {'train/loss': 0.8484938740730286, 'train/grad_norm_pre_clip': 15.859899520874023, 'timing/data': 0.021720089949667454, 'timing/model': 1.0322503075003624, 'learning_rate/qwen_vl_interface': 8.791491458879089e-06, 'learning_rate/action_model': 8.791491458879089e-05, 'epoch': 0.62}) +[2026-06-17 19:38:15,876][starVLA.training.train_starvla][INFO] - Step 1703, Loss: {'train/loss': 0.7603923678398132, 'train/grad_norm_pre_clip': 14.787311553955078, 'timing/data': 0.022038886323571205, 'timing/model': 1.027399556711316, 'learning_rate/qwen_vl_interface': 8.790049833048545e-06, 'learning_rate/action_model': 8.790049833048545e-05, 'epoch': 0.62}) +[2026-06-17 19:38:20,330][starVLA.training.train_starvla][INFO] - Step 1704, Loss: {'train/loss': 0.8138455748558044, 'train/grad_norm_pre_clip': 13.401114463806152, 'timing/data': 0.020968060940504074, 'timing/model': 1.0239656195044518, 'learning_rate/qwen_vl_interface': 8.788607473361862e-06, 'learning_rate/action_model': 8.788607473361862e-05, 'epoch': 0.62}) +[2026-06-17 19:38:24,742][starVLA.training.train_starvla][INFO] - Step 1705, Loss: {'train/loss': 0.9142777323722839, 'train/grad_norm_pre_clip': 90.03724670410156, 'timing/data': 0.019668024964630604, 'timing/model': 1.04036779794842, 'learning_rate/qwen_vl_interface': 8.787164380118044e-06, 'learning_rate/action_model': 8.787164380118043e-05, 'epoch': 0.62}) +[2026-06-17 19:38:29,052][starVLA.training.train_starvla][INFO] - Step 1706, Loss: {'train/loss': 0.6516612768173218, 'train/grad_norm_pre_clip': 19.05191421508789, 'timing/data': 0.02063450589776039, 'timing/model': 1.3200066899880767, 'learning_rate/qwen_vl_interface': 8.785720553616241e-06, 'learning_rate/action_model': 8.78572055361624e-05, 'epoch': 0.62}) +[2026-06-17 19:38:33,100][starVLA.training.train_starvla][INFO] - Step 1707, Loss: {'train/loss': 0.7495223879814148, 'train/grad_norm_pre_clip': 13.055525779724121, 'timing/data': 0.02326720580458641, 'timing/model': 1.0882294559851289, 'learning_rate/qwen_vl_interface': 8.784275994155762e-06, 'learning_rate/action_model': 8.784275994155762e-05, 'epoch': 0.63}) +[2026-06-17 19:38:37,688][starVLA.training.train_starvla][INFO] - Step 1708, Loss: {'train/loss': 0.6856842637062073, 'train/grad_norm_pre_clip': 10.497086524963379, 'timing/data': 0.02395919058471918, 'timing/model': 1.1091386750340462, 'learning_rate/qwen_vl_interface': 8.782830702036067e-06, 'learning_rate/action_model': 8.782830702036067e-05, 'epoch': 0.63}) +[2026-06-17 19:38:42,119][starVLA.training.train_starvla][INFO] - Step 1709, Loss: {'train/loss': 0.7516873478889465, 'train/grad_norm_pre_clip': 10.56217098236084, 'timing/data': 0.021923601627349854, 'timing/model': 1.0310831777751446, 'learning_rate/qwen_vl_interface': 8.781384677556767e-06, 'learning_rate/action_model': 8.781384677556767e-05, 'epoch': 0.63}) +[2026-06-17 19:38:46,868][starVLA.training.train_starvla][INFO] - Step 1710, Loss: {'train/loss': 0.6962845325469971, 'train/grad_norm_pre_clip': 11.639822006225586, 'timing/data': 0.023272416554391384, 'timing/model': 1.3603841867297888, 'learning_rate/qwen_vl_interface': 8.779937921017624e-06, 'learning_rate/action_model': 8.779937921017623e-05, 'epoch': 0.63}) +[2026-06-17 19:38:51,364][starVLA.training.train_starvla][INFO] - Step 1711, Loss: {'train/loss': 0.7635190486907959, 'train/grad_norm_pre_clip': 13.52315616607666, 'timing/data': 0.02245517075061798, 'timing/model': 1.3514197720214725, 'learning_rate/qwen_vl_interface': 8.77849043271855e-06, 'learning_rate/action_model': 8.778490432718551e-05, 'epoch': 0.63}) +[2026-06-17 19:38:55,605][starVLA.training.train_starvla][INFO] - Step 1712, Loss: {'train/loss': 0.653041422367096, 'train/grad_norm_pre_clip': 10.472506523132324, 'timing/data': 0.022960537113249302, 'timing/model': 1.026933421380818, 'learning_rate/qwen_vl_interface': 8.777042212959615e-06, 'learning_rate/action_model': 8.777042212959614e-05, 'epoch': 0.63}) +[2026-06-17 19:39:00,067][starVLA.training.train_starvla][INFO] - Step 1713, Loss: {'train/loss': 0.45162948966026306, 'train/grad_norm_pre_clip': 11.781301498413086, 'timing/data': 0.019508128985762596, 'timing/model': 1.1042828895151615, 'learning_rate/qwen_vl_interface': 8.775593262041032e-06, 'learning_rate/action_model': 8.77559326204103e-05, 'epoch': 0.63}) +[2026-06-17 19:39:05,917][starVLA.training.train_starvla][INFO] - Step 1714, Loss: {'train/loss': 0.5820672512054443, 'train/grad_norm_pre_clip': 10.243853569030762, 'timing/data': 0.022164203226566315, 'timing/model': 2.3865595143288374, 'learning_rate/qwen_vl_interface': 8.774143580263172e-06, 'learning_rate/action_model': 8.77414358026317e-05, 'epoch': 0.63}) +[2026-06-17 19:39:10,426][starVLA.training.train_starvla][INFO] - Step 1715, Loss: {'train/loss': 0.8447731733322144, 'train/grad_norm_pre_clip': 11.520356178283691, 'timing/data': 0.02257414348423481, 'timing/model': 1.034119070507586, 'learning_rate/qwen_vl_interface': 8.772693167926554e-06, 'learning_rate/action_model': 8.772693167926555e-05, 'epoch': 0.63}) +[2026-06-17 19:39:14,829][starVLA.training.train_starvla][INFO] - Step 1716, Loss: {'train/loss': 0.5062931180000305, 'train/grad_norm_pre_clip': 8.9428071975708, 'timing/data': 0.02075585350394249, 'timing/model': 1.3418689919635653, 'learning_rate/qwen_vl_interface': 8.771242025331855e-06, 'learning_rate/action_model': 8.771242025331854e-05, 'epoch': 0.63}) +[2026-06-17 19:39:18,963][starVLA.training.train_starvla][INFO] - Step 1717, Loss: {'train/loss': 0.736096203327179, 'train/grad_norm_pre_clip': 12.419352531433105, 'timing/data': 0.021295377984642982, 'timing/model': 1.0852397363632917, 'learning_rate/qwen_vl_interface': 8.769790152779893e-06, 'learning_rate/action_model': 8.769790152779892e-05, 'epoch': 0.63}) +[2026-06-17 19:39:25,474][starVLA.training.train_starvla][INFO] - Step 1718, Loss: {'train/loss': 0.5134996771812439, 'train/grad_norm_pre_clip': 11.226903915405273, 'timing/data': 0.02282350603491068, 'timing/model': 1.1074819201603532, 'learning_rate/qwen_vl_interface': 8.768337550571644e-06, 'learning_rate/action_model': 8.768337550571645e-05, 'epoch': 0.63}) +[2026-06-17 19:39:30,057][starVLA.training.train_starvla][INFO] - Step 1719, Loss: {'train/loss': 0.8389679789543152, 'train/grad_norm_pre_clip': 9.675209045410156, 'timing/data': 0.021842039190232754, 'timing/model': 1.034166063182056, 'learning_rate/qwen_vl_interface': 8.766884219008238e-06, 'learning_rate/action_model': 8.766884219008238e-05, 'epoch': 0.63}) +[2026-06-17 19:39:34,416][starVLA.training.train_starvla][INFO] - Step 1720, Loss: {'train/loss': 0.7825822830200195, 'train/grad_norm_pre_clip': 18.835676193237305, 'timing/data': 0.028417144902050495, 'timing/model': 1.0629108333960176, 'learning_rate/qwen_vl_interface': 8.765430158390948e-06, 'learning_rate/action_model': 8.765430158390947e-05, 'epoch': 0.63}) +[2026-06-17 19:39:38,968][starVLA.training.train_starvla][INFO] - Step 1721, Loss: {'train/loss': 0.532477080821991, 'train/grad_norm_pre_clip': 10.689424514770508, 'timing/data': 0.020618046633899212, 'timing/model': 1.395940076559782, 'learning_rate/qwen_vl_interface': 8.763975369021202e-06, 'learning_rate/action_model': 8.763975369021203e-05, 'epoch': 0.63}) +[2026-06-17 19:39:43,108][starVLA.training.train_starvla][INFO] - Step 1722, Loss: {'train/loss': 0.6671178340911865, 'train/grad_norm_pre_clip': 12.814624786376953, 'timing/data': 0.0206133546307683, 'timing/model': 1.1240091267973185, 'learning_rate/qwen_vl_interface': 8.762519851200581e-06, 'learning_rate/action_model': 8.762519851200581e-05, 'epoch': 0.63}) +[2026-06-17 19:39:49,816][starVLA.training.train_starvla][INFO] - Step 1723, Loss: {'train/loss': 0.49851977825164795, 'train/grad_norm_pre_clip': 12.487761497497559, 'timing/data': 0.02103761862963438, 'timing/model': 2.932413648813963, 'learning_rate/qwen_vl_interface': 8.761063605230818e-06, 'learning_rate/action_model': 8.761063605230818e-05, 'epoch': 0.63}) +[2026-06-17 19:39:55,388][starVLA.training.train_starvla][INFO] - Step 1724, Loss: {'train/loss': 0.5638434290885925, 'train/grad_norm_pre_clip': 11.319488525390625, 'timing/data': 0.026440289802849293, 'timing/model': 1.061842929571867, 'learning_rate/qwen_vl_interface': 8.759606631413788e-06, 'learning_rate/action_model': 8.759606631413788e-05, 'epoch': 0.63}) +[2026-06-17 19:40:03,136][starVLA.training.train_starvla][INFO] - Step 1725, Loss: {'train/loss': 0.39145129919052124, 'train/grad_norm_pre_clip': 10.186100959777832, 'timing/data': 0.0207055751234293, 'timing/model': 2.3828524574637413, 'learning_rate/qwen_vl_interface': 8.75814893005153e-06, 'learning_rate/action_model': 8.75814893005153e-05, 'epoch': 0.63}) +[2026-06-17 19:40:08,020][starVLA.training.train_starvla][INFO] - Step 1726, Loss: {'train/loss': 0.43511754274368286, 'train/grad_norm_pre_clip': 9.019240379333496, 'timing/data': 0.027829963713884354, 'timing/model': 1.6149331266060472, 'learning_rate/qwen_vl_interface': 8.756690501446222e-06, 'learning_rate/action_model': 8.756690501446221e-05, 'epoch': 0.63}) +[2026-06-17 19:40:14,093][starVLA.training.train_starvla][INFO] - Step 1727, Loss: {'train/loss': 0.5739015340805054, 'train/grad_norm_pre_clip': 10.585387229919434, 'timing/data': 0.02072280365973711, 'timing/model': 1.0320871071889997, 'learning_rate/qwen_vl_interface': 8.755231345900199e-06, 'learning_rate/action_model': 8.755231345900198e-05, 'epoch': 0.63}) +[2026-06-17 19:40:18,516][starVLA.training.train_starvla][INFO] - Step 1728, Loss: {'train/loss': 0.5190710425376892, 'train/grad_norm_pre_clip': 9.412820816040039, 'timing/data': 0.02047665696591139, 'timing/model': 1.0901586739346385, 'learning_rate/qwen_vl_interface': 8.753771463715947e-06, 'learning_rate/action_model': 8.753771463715947e-05, 'epoch': 0.63}) +[2026-06-17 19:40:22,952][starVLA.training.train_starvla][INFO] - Step 1729, Loss: {'train/loss': 0.8419937491416931, 'train/grad_norm_pre_clip': 13.824109077453613, 'timing/data': 0.06042338069528341, 'timing/model': 1.0379573535174131, 'learning_rate/qwen_vl_interface': 8.7523108551961e-06, 'learning_rate/action_model': 8.7523108551961e-05, 'epoch': 0.63}) +[2026-06-17 19:40:27,450][starVLA.training.train_starvla][INFO] - Step 1730, Loss: {'train/loss': 0.7725167274475098, 'train/grad_norm_pre_clip': 16.331350326538086, 'timing/data': 0.019971825182437897, 'timing/model': 1.0703189233317971, 'learning_rate/qwen_vl_interface': 8.750849520643445e-06, 'learning_rate/action_model': 8.750849520643444e-05, 'epoch': 0.63}) +[2026-06-17 19:40:31,947][starVLA.training.train_starvla][INFO] - Step 1731, Loss: {'train/loss': 0.9420254230499268, 'train/grad_norm_pre_clip': 15.334017753601074, 'timing/data': 0.029815870337188244, 'timing/model': 1.3752300553023815, 'learning_rate/qwen_vl_interface': 8.749387460360917e-06, 'learning_rate/action_model': 8.749387460360916e-05, 'epoch': 0.63}) +[2026-06-17 19:40:36,007][starVLA.training.train_starvla][INFO] - Step 1732, Loss: {'train/loss': 0.34723734855651855, 'train/grad_norm_pre_clip': 9.429642677307129, 'timing/data': 0.02025442384183407, 'timing/model': 1.0942398607730865, 'learning_rate/qwen_vl_interface': 8.747924674651602e-06, 'learning_rate/action_model': 8.747924674651602e-05, 'epoch': 0.63}) +[2026-06-17 19:40:40,411][starVLA.training.train_starvla][INFO] - Step 1733, Loss: {'train/loss': 0.5732365846633911, 'train/grad_norm_pre_clip': 14.343025207519531, 'timing/data': 0.02082109823822975, 'timing/model': 1.1139215994626284, 'learning_rate/qwen_vl_interface': 8.74646116381874e-06, 'learning_rate/action_model': 8.74646116381874e-05, 'epoch': 0.63}) +[2026-06-17 19:40:44,955][starVLA.training.train_starvla][INFO] - Step 1734, Loss: {'train/loss': 0.6977452039718628, 'train/grad_norm_pre_clip': 14.52429485321045, 'timing/data': 0.02027500793337822, 'timing/model': 1.0845473567023873, 'learning_rate/qwen_vl_interface': 8.744996928165714e-06, 'learning_rate/action_model': 8.744996928165714e-05, 'epoch': 0.63}) +[2026-06-17 19:40:49,347][starVLA.training.train_starvla][INFO] - Step 1735, Loss: {'train/loss': 1.1135104894638062, 'train/grad_norm_pre_clip': 16.036548614501953, 'timing/data': 0.020089318975806236, 'timing/model': 1.017125567421317, 'learning_rate/qwen_vl_interface': 8.743531967996066e-06, 'learning_rate/action_model': 8.743531967996065e-05, 'epoch': 0.64}) +[2026-06-17 19:40:53,378][starVLA.training.train_starvla][INFO] - Step 1736, Loss: {'train/loss': 0.9157419800758362, 'train/grad_norm_pre_clip': 14.135661125183105, 'timing/data': 0.020983602851629257, 'timing/model': 1.049911117181182, 'learning_rate/qwen_vl_interface': 8.742066283613481e-06, 'learning_rate/action_model': 8.74206628361348e-05, 'epoch': 0.64}) +[2026-06-17 19:40:57,745][starVLA.training.train_starvla][INFO] - Step 1737, Loss: {'train/loss': 0.4511566162109375, 'train/grad_norm_pre_clip': 11.994183540344238, 'timing/data': 0.019553180783987045, 'timing/model': 1.051945143379271, 'learning_rate/qwen_vl_interface': 8.740599875321797e-06, 'learning_rate/action_model': 8.740599875321797e-05, 'epoch': 0.64}) +[2026-06-17 19:41:02,170][starVLA.training.train_starvla][INFO] - Step 1738, Loss: {'train/loss': 0.555945634841919, 'train/grad_norm_pre_clip': 15.174464225769043, 'timing/data': 0.023270828649401665, 'timing/model': 1.0688795316964388, 'learning_rate/qwen_vl_interface': 8.739132743425004e-06, 'learning_rate/action_model': 8.739132743425002e-05, 'epoch': 0.64}) +[2026-06-17 19:41:06,668][starVLA.training.train_starvla][INFO] - Step 1739, Loss: {'train/loss': 0.42296454310417175, 'train/grad_norm_pre_clip': 14.617244720458984, 'timing/data': 0.019541804678738117, 'timing/model': 1.0251503167673945, 'learning_rate/qwen_vl_interface': 8.737664888227239e-06, 'learning_rate/action_model': 8.737664888227239e-05, 'epoch': 0.64}) +[2026-06-17 19:41:11,135][starVLA.training.train_starvla][INFO] - Step 1740, Loss: {'train/loss': 0.849898099899292, 'train/grad_norm_pre_clip': 19.06865119934082, 'timing/data': 0.02166967000812292, 'timing/model': 1.3669032156467438, 'learning_rate/qwen_vl_interface': 8.736196310032787e-06, 'learning_rate/action_model': 8.736196310032787e-05, 'epoch': 0.64}) +[2026-06-17 19:41:20,500][starVLA.training.train_starvla][INFO] - Step 1741, Loss: {'train/loss': 0.781487226486206, 'train/grad_norm_pre_clip': 16.738706588745117, 'timing/data': 0.021107174456119537, 'timing/model': 3.8424776028841734, 'learning_rate/qwen_vl_interface': 8.734727009146092e-06, 'learning_rate/action_model': 8.734727009146092e-05, 'epoch': 0.64}) +[2026-06-17 19:41:24,582][starVLA.training.train_starvla][INFO] - Step 1742, Loss: {'train/loss': 0.9542290568351746, 'train/grad_norm_pre_clip': 15.159026145935059, 'timing/data': 0.02372719906270504, 'timing/model': 1.0392482709139585, 'learning_rate/qwen_vl_interface': 8.733256985871737e-06, 'learning_rate/action_model': 8.733256985871737e-05, 'epoch': 0.64}) +[2026-06-17 19:41:32,712][starVLA.training.train_starvla][INFO] - Step 1743, Loss: {'train/loss': 1.0775216817855835, 'train/grad_norm_pre_clip': 22.200607299804688, 'timing/data': 0.01936358120292425, 'timing/model': 2.854714493267238, 'learning_rate/qwen_vl_interface': 8.73178624051446e-06, 'learning_rate/action_model': 8.73178624051446e-05, 'epoch': 0.64}) +[2026-06-17 19:41:37,316][starVLA.training.train_starvla][INFO] - Step 1744, Loss: {'train/loss': 0.908002495765686, 'train/grad_norm_pre_clip': 19.349836349487305, 'timing/data': 0.02045554295182228, 'timing/model': 1.0458238432183862, 'learning_rate/qwen_vl_interface': 8.73031477337915e-06, 'learning_rate/action_model': 8.73031477337915e-05, 'epoch': 0.64}) +[2026-06-17 19:41:42,778][starVLA.training.train_starvla][INFO] - Step 1745, Loss: {'train/loss': 0.6835575103759766, 'train/grad_norm_pre_clip': 13.68877124786377, 'timing/data': 0.02182080317288637, 'timing/model': 1.2228241553530097, 'learning_rate/qwen_vl_interface': 8.728842584770842e-06, 'learning_rate/action_model': 8.728842584770842e-05, 'epoch': 0.64}) +[2026-06-17 19:41:49,274][starVLA.training.train_starvla][INFO] - Step 1746, Loss: {'train/loss': 0.8102730512619019, 'train/grad_norm_pre_clip': 18.63381004333496, 'timing/data': 0.021182449534535408, 'timing/model': 3.3618671847507358, 'learning_rate/qwen_vl_interface': 8.727369674994722e-06, 'learning_rate/action_model': 8.727369674994722e-05, 'epoch': 0.64}) +[2026-06-17 19:41:55,635][starVLA.training.train_starvla][INFO] - Step 1747, Loss: {'train/loss': 0.6060442924499512, 'train/grad_norm_pre_clip': 14.809818267822266, 'timing/data': 0.02208597306162119, 'timing/model': 3.348525364883244, 'learning_rate/qwen_vl_interface': 8.725896044356128e-06, 'learning_rate/action_model': 8.725896044356128e-05, 'epoch': 0.64}) +[2026-06-17 19:42:02,628][starVLA.training.train_starvla][INFO] - Step 1748, Loss: {'train/loss': 0.5874845385551453, 'train/grad_norm_pre_clip': 18.364788055419922, 'timing/data': 0.020025416277348995, 'timing/model': 1.0305077657103539, 'learning_rate/qwen_vl_interface': 8.724421693160544e-06, 'learning_rate/action_model': 8.724421693160543e-05, 'epoch': 0.64}) +[2026-06-17 19:42:07,078][starVLA.training.train_starvla][INFO] - Step 1749, Loss: {'train/loss': 0.747783899307251, 'train/grad_norm_pre_clip': 13.173233985900879, 'timing/data': 0.020051862113177776, 'timing/model': 1.0301550645381212, 'learning_rate/qwen_vl_interface': 8.722946621713603e-06, 'learning_rate/action_model': 8.722946621713602e-05, 'epoch': 0.64}) +[2026-06-17 19:42:11,533][starVLA.training.train_starvla][INFO] - Step 1750, Loss: {'train/loss': 0.8291351199150085, 'train/grad_norm_pre_clip': 14.311079025268555, 'timing/data': 0.02587110735476017, 'timing/model': 1.1027290681377053, 'learning_rate/qwen_vl_interface': 8.721470830321091e-06, 'learning_rate/action_model': 8.72147083032109e-05, 'epoch': 0.64}) +[2026-06-17 19:42:16,104][starVLA.training.train_starvla][INFO] - Step 1751, Loss: {'train/loss': 0.7596500515937805, 'train/grad_norm_pre_clip': 25.73183250427246, 'timing/data': 0.02250855602324009, 'timing/model': 1.3375440165400505, 'learning_rate/qwen_vl_interface': 8.719994319288939e-06, 'learning_rate/action_model': 8.71999431928894e-05, 'epoch': 0.64}) +[2026-06-17 19:42:20,346][starVLA.training.train_starvla][INFO] - Step 1752, Loss: {'train/loss': 0.7760269641876221, 'train/grad_norm_pre_clip': 12.724588394165039, 'timing/data': 0.020221699960529804, 'timing/model': 1.0751027865335345, 'learning_rate/qwen_vl_interface': 8.718517088923233e-06, 'learning_rate/action_model': 8.718517088923233e-05, 'epoch': 0.64}) +[2026-06-17 19:42:24,896][starVLA.training.train_starvla][INFO] - Step 1753, Loss: {'train/loss': 0.9476093649864197, 'train/grad_norm_pre_clip': 19.905670166015625, 'timing/data': 0.020315798930823803, 'timing/model': 1.0728656072169542, 'learning_rate/qwen_vl_interface': 8.7170391395302e-06, 'learning_rate/action_model': 8.7170391395302e-05, 'epoch': 0.64}) +[2026-06-17 19:42:29,300][starVLA.training.train_starvla][INFO] - Step 1754, Loss: {'train/loss': 0.8672108054161072, 'train/grad_norm_pre_clip': 14.798794746398926, 'timing/data': 0.02173548750579357, 'timing/model': 1.0470590759068727, 'learning_rate/qwen_vl_interface': 8.715560471416223e-06, 'learning_rate/action_model': 8.715560471416224e-05, 'epoch': 0.64}) +[2026-06-17 19:42:33,821][starVLA.training.train_starvla][INFO] - Step 1755, Loss: {'train/loss': 0.9195138812065125, 'train/grad_norm_pre_clip': 15.081740379333496, 'timing/data': 0.019711630418896675, 'timing/model': 1.4558977633714676, 'learning_rate/qwen_vl_interface': 8.714081084887833e-06, 'learning_rate/action_model': 8.714081084887833e-05, 'epoch': 0.64}) +[2026-06-17 19:42:38,215][starVLA.training.train_starvla][INFO] - Step 1756, Loss: {'train/loss': 1.3955016136169434, 'train/grad_norm_pre_clip': 17.577363967895508, 'timing/data': 0.0253695510327816, 'timing/model': 1.4241278730332851, 'learning_rate/qwen_vl_interface': 8.712600980251706e-06, 'learning_rate/action_model': 8.712600980251705e-05, 'epoch': 0.64}) +[2026-06-17 19:42:42,396][starVLA.training.train_starvla][INFO] - Step 1757, Loss: {'train/loss': 0.9493764042854309, 'train/grad_norm_pre_clip': 14.009344100952148, 'timing/data': 0.02060545701533556, 'timing/model': 1.048546627163887, 'learning_rate/qwen_vl_interface': 8.71112015781467e-06, 'learning_rate/action_model': 8.71112015781467e-05, 'epoch': 0.64}) +[2026-06-17 19:42:46,915][starVLA.training.train_starvla][INFO] - Step 1758, Loss: {'train/loss': 1.1449778079986572, 'train/grad_norm_pre_clip': 16.314044952392578, 'timing/data': 0.021874086931347847, 'timing/model': 1.0425186064094305, 'learning_rate/qwen_vl_interface': 8.709638617883699e-06, 'learning_rate/action_model': 8.709638617883699e-05, 'epoch': 0.64}) +[2026-06-17 19:42:51,297][starVLA.training.train_starvla][INFO] - Step 1759, Loss: {'train/loss': 0.6065181493759155, 'train/grad_norm_pre_clip': 10.05344295501709, 'timing/data': 0.02059894986450672, 'timing/model': 1.102525444701314, 'learning_rate/qwen_vl_interface': 8.708156360765922e-06, 'learning_rate/action_model': 8.708156360765922e-05, 'epoch': 0.64}) +[2026-06-17 19:42:55,763][starVLA.training.train_starvla][INFO] - Step 1760, Loss: {'train/loss': 0.6495103240013123, 'train/grad_norm_pre_clip': 31.40751838684082, 'timing/data': 0.021656488068401814, 'timing/model': 1.0391702251508832, 'learning_rate/qwen_vl_interface': 8.706673386768609e-06, 'learning_rate/action_model': 8.706673386768609e-05, 'epoch': 0.64}) +[2026-06-17 19:43:00,309][starVLA.training.train_starvla][INFO] - Step 1761, Loss: {'train/loss': 0.8730976581573486, 'train/grad_norm_pre_clip': 14.716450691223145, 'timing/data': 0.02177367825061083, 'timing/model': 1.3414234118536115, 'learning_rate/qwen_vl_interface': 8.705189696199184e-06, 'learning_rate/action_model': 8.705189696199183e-05, 'epoch': 0.64}) +[2026-06-17 19:43:05,681][starVLA.training.train_starvla][INFO] - Step 1762, Loss: {'train/loss': 0.6017475724220276, 'train/grad_norm_pre_clip': 18.50811767578125, 'timing/data': 0.02134348824620247, 'timing/model': 2.3618184719234705, 'learning_rate/qwen_vl_interface': 8.703705289365216e-06, 'learning_rate/action_model': 8.703705289365216e-05, 'epoch': 0.65}) +[2026-06-17 19:43:12,488][starVLA.training.train_starvla][INFO] - Step 1763, Loss: {'train/loss': 0.6668696403503418, 'train/grad_norm_pre_clip': 15.333067893981934, 'timing/data': 0.02060198411345482, 'timing/model': 2.7235310170799494, 'learning_rate/qwen_vl_interface': 8.702220166574426e-06, 'learning_rate/action_model': 8.702220166574425e-05, 'epoch': 0.65}) +[2026-06-17 19:43:17,107][starVLA.training.train_starvla][INFO] - Step 1764, Loss: {'train/loss': 1.0608259439468384, 'train/grad_norm_pre_clip': 18.706138610839844, 'timing/data': 0.021015798673033714, 'timing/model': 1.114813000895083, 'learning_rate/qwen_vl_interface': 8.700734328134681e-06, 'learning_rate/action_model': 8.70073432813468e-05, 'epoch': 0.65}) +[2026-06-17 19:43:21,663][starVLA.training.train_starvla][INFO] - Step 1765, Loss: {'train/loss': 0.7905035018920898, 'train/grad_norm_pre_clip': 13.345664978027344, 'timing/data': 0.022846011444926262, 'timing/model': 1.0961942933499813, 'learning_rate/qwen_vl_interface': 8.699247774353997e-06, 'learning_rate/action_model': 8.699247774353996e-05, 'epoch': 0.65}) +[2026-06-17 19:43:28,069][starVLA.training.train_starvla][INFO] - Step 1766, Loss: {'train/loss': 0.6527664661407471, 'train/grad_norm_pre_clip': 12.914539337158203, 'timing/data': 0.022161255590617657, 'timing/model': 3.448528696782887, 'learning_rate/qwen_vl_interface': 8.697760505540536e-06, 'learning_rate/action_model': 8.697760505540536e-05, 'epoch': 0.65}) +[2026-06-17 19:43:32,339][starVLA.training.train_starvla][INFO] - Step 1767, Loss: {'train/loss': 0.7996333837509155, 'train/grad_norm_pre_clip': 11.257174491882324, 'timing/data': 0.02291432023048401, 'timing/model': 1.1315995575860143, 'learning_rate/qwen_vl_interface': 8.696272522002614e-06, 'learning_rate/action_model': 8.696272522002615e-05, 'epoch': 0.65}) +[2026-06-17 19:43:40,072][starVLA.training.train_starvla][INFO] - Step 1768, Loss: {'train/loss': 0.8550218939781189, 'train/grad_norm_pre_clip': 12.174678802490234, 'timing/data': 0.020669876597821712, 'timing/model': 1.1585210850462317, 'learning_rate/qwen_vl_interface': 8.69478382404869e-06, 'learning_rate/action_model': 8.69478382404869e-05, 'epoch': 0.65}) +[2026-06-17 19:43:44,567][starVLA.training.train_starvla][INFO] - Step 1769, Loss: {'train/loss': 1.2907615900039673, 'train/grad_norm_pre_clip': 16.40612030029297, 'timing/data': 0.023753518238663673, 'timing/model': 1.033564436249435, 'learning_rate/qwen_vl_interface': 8.693294411987371e-06, 'learning_rate/action_model': 8.693294411987371e-05, 'epoch': 0.65}) +[2026-06-17 19:43:49,125][starVLA.training.train_starvla][INFO] - Step 1770, Loss: {'train/loss': 0.9059585332870483, 'train/grad_norm_pre_clip': 12.045554161071777, 'timing/data': 0.021594494581222534, 'timing/model': 1.1268986286595464, 'learning_rate/qwen_vl_interface': 8.691804286127416e-06, 'learning_rate/action_model': 8.691804286127415e-05, 'epoch': 0.65}) +[2026-06-17 19:43:53,789][starVLA.training.train_starvla][INFO] - Step 1771, Loss: {'train/loss': 0.5395404100418091, 'train/grad_norm_pre_clip': 10.156872749328613, 'timing/data': 0.020722966641187668, 'timing/model': 1.5213495790958405, 'learning_rate/qwen_vl_interface': 8.690313446777729e-06, 'learning_rate/action_model': 8.690313446777728e-05, 'epoch': 0.65}) +[2026-06-17 19:43:59,162][starVLA.training.train_starvla][INFO] - Step 1772, Loss: {'train/loss': 0.7958534359931946, 'train/grad_norm_pre_clip': 16.765823364257812, 'timing/data': 0.05640921648591757, 'timing/model': 1.047221357934177, 'learning_rate/qwen_vl_interface': 8.688821894247363e-06, 'learning_rate/action_model': 8.688821894247362e-05, 'epoch': 0.65}) +[2026-06-17 19:44:06,473][starVLA.training.train_starvla][INFO] - Step 1773, Loss: {'train/loss': 1.3177680969238281, 'train/grad_norm_pre_clip': 13.726178169250488, 'timing/data': 0.028103874064981937, 'timing/model': 1.168565976433456, 'learning_rate/qwen_vl_interface': 8.687329628845517e-06, 'learning_rate/action_model': 8.687329628845516e-05, 'epoch': 0.65}) +[2026-06-17 19:44:11,081][starVLA.training.train_starvla][INFO] - Step 1774, Loss: {'train/loss': 0.8164808750152588, 'train/grad_norm_pre_clip': 10.60921859741211, 'timing/data': 0.020269704051315784, 'timing/model': 1.0272468198090792, 'learning_rate/qwen_vl_interface': 8.68583665088154e-06, 'learning_rate/action_model': 8.68583665088154e-05, 'epoch': 0.65}) +[2026-06-17 19:44:15,436][starVLA.training.train_starvla][INFO] - Step 1775, Loss: {'train/loss': 0.7110936045646667, 'train/grad_norm_pre_clip': 13.757024765014648, 'timing/data': 0.019645840860903263, 'timing/model': 1.0254712281748652, 'learning_rate/qwen_vl_interface': 8.684342960664928e-06, 'learning_rate/action_model': 8.684342960664928e-05, 'epoch': 0.65}) +[2026-06-17 19:44:20,432][starVLA.training.train_starvla][INFO] - Step 1776, Loss: {'train/loss': 0.7840887904167175, 'train/grad_norm_pre_clip': 10.156180381774902, 'timing/data': 0.02042084652930498, 'timing/model': 1.2711963970214128, 'learning_rate/qwen_vl_interface': 8.682848558505327e-06, 'learning_rate/action_model': 8.682848558505326e-05, 'epoch': 0.65}) +[2026-06-17 19:44:24,464][starVLA.training.train_starvla][INFO] - Step 1777, Loss: {'train/loss': 0.9290231466293335, 'train/grad_norm_pre_clip': 11.270525932312012, 'timing/data': 0.02146976813673973, 'timing/model': 1.0384945319965482, 'learning_rate/qwen_vl_interface': 8.681353444712525e-06, 'learning_rate/action_model': 8.681353444712524e-05, 'epoch': 0.65}) +[2026-06-17 19:44:28,978][starVLA.training.train_starvla][INFO] - Step 1778, Loss: {'train/loss': 0.9442143440246582, 'train/grad_norm_pre_clip': 13.268369674682617, 'timing/data': 0.021093780174851418, 'timing/model': 1.272578464820981, 'learning_rate/qwen_vl_interface': 8.67985761959646e-06, 'learning_rate/action_model': 8.67985761959646e-05, 'epoch': 0.65}) +[2026-06-17 19:44:33,538][starVLA.training.train_starvla][INFO] - Step 1779, Loss: {'train/loss': 1.0569303035736084, 'train/grad_norm_pre_clip': 22.587553024291992, 'timing/data': 0.024489483796060085, 'timing/model': 1.0595025410875678, 'learning_rate/qwen_vl_interface': 8.678361083467221e-06, 'learning_rate/action_model': 8.67836108346722e-05, 'epoch': 0.65}) +[2026-06-17 19:44:39,928][starVLA.training.train_starvla][INFO] - Step 1780, Loss: {'train/loss': 0.743317723274231, 'train/grad_norm_pre_clip': 11.111849784851074, 'timing/data': 0.019577614031732082, 'timing/model': 2.9926237808540463, 'learning_rate/qwen_vl_interface': 8.67686383663504e-06, 'learning_rate/action_model': 8.67686383663504e-05, 'epoch': 0.65}) +[2026-06-17 19:44:44,043][starVLA.training.train_starvla][INFO] - Step 1781, Loss: {'train/loss': 1.0573216676712036, 'train/grad_norm_pre_clip': 15.695751190185547, 'timing/data': 0.02138519287109375, 'timing/model': 1.0860652215778828, 'learning_rate/qwen_vl_interface': 8.6753658794103e-06, 'learning_rate/action_model': 8.675365879410298e-05, 'epoch': 0.65}) +[2026-06-17 19:44:48,448][starVLA.training.train_starvla][INFO] - Step 1782, Loss: {'train/loss': 0.6680095791816711, 'train/grad_norm_pre_clip': 11.813183784484863, 'timing/data': 0.023801326751708984, 'timing/model': 1.0330185936763883, 'learning_rate/qwen_vl_interface': 8.673867212103525e-06, 'learning_rate/action_model': 8.673867212103524e-05, 'epoch': 0.65}) +[2026-06-17 19:44:52,858][starVLA.training.train_starvla][INFO] - Step 1783, Loss: {'train/loss': 0.63808274269104, 'train/grad_norm_pre_clip': 9.774473190307617, 'timing/data': 0.023919335566461086, 'timing/model': 1.0300727412104607, 'learning_rate/qwen_vl_interface': 8.672367835025395e-06, 'learning_rate/action_model': 8.672367835025393e-05, 'epoch': 0.65}) +[2026-06-17 19:44:57,406][starVLA.training.train_starvla][INFO] - Step 1784, Loss: {'train/loss': 0.8589888215065002, 'train/grad_norm_pre_clip': 14.174013137817383, 'timing/data': 0.026763206347823143, 'timing/model': 1.1331056728959084, 'learning_rate/qwen_vl_interface': 8.670867748486727e-06, 'learning_rate/action_model': 8.670867748486726e-05, 'epoch': 0.65}) +[2026-06-17 19:45:01,885][starVLA.training.train_starvla][INFO] - Step 1785, Loss: {'train/loss': 0.8466274738311768, 'train/grad_norm_pre_clip': 13.004631042480469, 'timing/data': 0.025369230657815933, 'timing/model': 1.4083016570657492, 'learning_rate/qwen_vl_interface': 8.669366952798494e-06, 'learning_rate/action_model': 8.669366952798494e-05, 'epoch': 0.65}) +[2026-06-17 19:45:08,050][starVLA.training.train_starvla][INFO] - Step 1786, Loss: {'train/loss': 0.5067897439002991, 'train/grad_norm_pre_clip': 8.160924911499023, 'timing/data': 0.020962483249604702, 'timing/model': 1.4999284502118826, 'learning_rate/qwen_vl_interface': 8.667865448271814e-06, 'learning_rate/action_model': 8.667865448271814e-05, 'epoch': 0.65}) +[2026-06-17 19:45:12,150][starVLA.training.train_starvla][INFO] - Step 1787, Loss: {'train/loss': 0.7343822717666626, 'train/grad_norm_pre_clip': 10.843705177307129, 'timing/data': 0.02001460548490286, 'timing/model': 1.041913965716958, 'learning_rate/qwen_vl_interface': 8.666363235217949e-06, 'learning_rate/action_model': 8.66636323521795e-05, 'epoch': 0.65}) +[2026-06-17 19:45:16,494][starVLA.training.train_starvla][INFO] - Step 1788, Loss: {'train/loss': 0.8352506160736084, 'train/grad_norm_pre_clip': 12.113877296447754, 'timing/data': 0.022380158305168152, 'timing/model': 1.094677645713091, 'learning_rate/qwen_vl_interface': 8.66486031394831e-06, 'learning_rate/action_model': 8.664860313948309e-05, 'epoch': 0.65}) +[2026-06-17 19:45:23,005][starVLA.training.train_starvla][INFO] - Step 1789, Loss: {'train/loss': 1.0276669263839722, 'train/grad_norm_pre_clip': 12.515904426574707, 'timing/data': 0.020753558725118637, 'timing/model': 3.1084527606144547, 'learning_rate/qwen_vl_interface': 8.66335668477445e-06, 'learning_rate/action_model': 8.66335668477445e-05, 'epoch': 0.66}) +[2026-06-17 19:45:27,671][starVLA.training.train_starvla][INFO] - Step 1790, Loss: {'train/loss': 0.6639597415924072, 'train/grad_norm_pre_clip': 11.243520736694336, 'timing/data': 0.022397820837795734, 'timing/model': 1.1059739058837295, 'learning_rate/qwen_vl_interface': 8.661852348008079e-06, 'learning_rate/action_model': 8.661852348008078e-05, 'epoch': 0.66}) +[2026-06-17 19:45:31,953][starVLA.training.train_starvla][INFO] - Step 1791, Loss: {'train/loss': 0.8665200471878052, 'train/grad_norm_pre_clip': 16.039005279541016, 'timing/data': 0.023962450213730335, 'timing/model': 1.3061982514336705, 'learning_rate/qwen_vl_interface': 8.660347303961044e-06, 'learning_rate/action_model': 8.660347303961044e-05, 'epoch': 0.66}) +[2026-06-17 19:45:38,476][starVLA.training.train_starvla][INFO] - Step 1792, Loss: {'train/loss': 0.846379280090332, 'train/grad_norm_pre_clip': 15.034896850585938, 'timing/data': 0.023030211217701435, 'timing/model': 2.813733628951013, 'learning_rate/qwen_vl_interface': 8.658841552945342e-06, 'learning_rate/action_model': 8.658841552945342e-05, 'epoch': 0.66}) +[2026-06-17 19:45:42,931][starVLA.training.train_starvla][INFO] - Step 1793, Loss: {'train/loss': 0.5327152013778687, 'train/grad_norm_pre_clip': 11.24649429321289, 'timing/data': 0.022151648998260498, 'timing/model': 1.0254178503528237, 'learning_rate/qwen_vl_interface': 8.65733509527312e-06, 'learning_rate/action_model': 8.65733509527312e-05, 'epoch': 0.66}) +[2026-06-17 19:45:47,479][starVLA.training.train_starvla][INFO] - Step 1794, Loss: {'train/loss': 0.5052626729011536, 'train/grad_norm_pre_clip': 13.379817008972168, 'timing/data': 0.031419276259839535, 'timing/model': 1.1118561401963234, 'learning_rate/qwen_vl_interface': 8.655827931256667e-06, 'learning_rate/action_model': 8.655827931256667e-05, 'epoch': 0.66}) +[2026-06-17 19:45:52,047][starVLA.training.train_starvla][INFO] - Step 1795, Loss: {'train/loss': 0.8004867434501648, 'train/grad_norm_pre_clip': 12.092386245727539, 'timing/data': 0.022638102993369102, 'timing/model': 1.0375471794977784, 'learning_rate/qwen_vl_interface': 8.654320061208418e-06, 'learning_rate/action_model': 8.654320061208417e-05, 'epoch': 0.66}) +[2026-06-17 19:45:56,493][starVLA.training.train_starvla][INFO] - Step 1796, Loss: {'train/loss': 0.7926382422447205, 'train/grad_norm_pre_clip': 21.61372947692871, 'timing/data': 0.0223229993134737, 'timing/model': 1.3204788211733103, 'learning_rate/qwen_vl_interface': 8.652811485440956e-06, 'learning_rate/action_model': 8.652811485440956e-05, 'epoch': 0.66}) +[2026-06-17 19:46:00,513][starVLA.training.train_starvla][INFO] - Step 1797, Loss: {'train/loss': 0.5970134735107422, 'train/grad_norm_pre_clip': 8.943182945251465, 'timing/data': 0.019869470968842506, 'timing/model': 1.047082582488656, 'learning_rate/qwen_vl_interface': 8.651302204267013e-06, 'learning_rate/action_model': 8.651302204267013e-05, 'epoch': 0.66}) +[2026-06-17 19:46:04,981][starVLA.training.train_starvla][INFO] - Step 1798, Loss: {'train/loss': 0.6538674831390381, 'train/grad_norm_pre_clip': 9.898107528686523, 'timing/data': 0.020519557408988476, 'timing/model': 1.0244821710512042, 'learning_rate/qwen_vl_interface': 8.649792217999461e-06, 'learning_rate/action_model': 8.649792217999462e-05, 'epoch': 0.66}) +[2026-06-17 19:46:09,328][starVLA.training.train_starvla][INFO] - Step 1799, Loss: {'train/loss': 0.6647958755493164, 'train/grad_norm_pre_clip': 27.204051971435547, 'timing/data': 0.01982094906270504, 'timing/model': 1.0507412869483232, 'learning_rate/qwen_vl_interface': 8.648281526951327e-06, 'learning_rate/action_model': 8.648281526951327e-05, 'epoch': 0.66}) +[2026-06-17 19:48:43,867][starVLA.training.train_starvla][INFO] - Step 1800, Loss: {'train/loss': 0.8271918892860413, 'train/grad_norm_pre_clip': 18.594440460205078, 'eval/loss': 0.8534319400787354, 'eval/latency_2/loss': 0.8534319400787354, 'eval/demon_fire_precision': 0.8809051210797936, 'eval/demon_fire_recall': 0.9284518828451883, 'eval/demon_fire_f1': 0.904053778773681, 'eval/demon_fire_rate_model': 0.7008903728436283, 'eval/demon_fire_rate_teacher': 0.6649972175848636, 'eval/demon_left_precision': 0.7162661737523105, 'eval/demon_left_recall': 0.8157894736842105, 'eval/demon_left_f1': 0.7627952755905512, 'eval/demon_left_rate_model': 0.30105731775180855, 'eval/demon_left_rate_teacher': 0.2643294379521425, 'eval/demon_right_precision': 0.7123681687440077, 'eval/demon_right_recall': 0.8395480225988701, 'eval/demon_right_f1': 0.7707468879668049, 'eval/demon_right_rate_model': 0.29020589872008906, 'eval/demon_right_rate_teacher': 0.24624373956594323, 'eval/demon_move_f1': 0.766771081778678, 'eval/demon_cc_f1': 0.8579704681449917, 'eval/latency_2/demon_fire_precision': 0.8809051210797936, 'eval/latency_2/demon_fire_recall': 0.9284518828451883, 'eval/latency_2/demon_fire_f1': 0.904053778773681, 'eval/latency_2/demon_fire_rate_model': 0.7008903728436283, 'eval/latency_2/demon_fire_rate_teacher': 0.6649972175848636, 'eval/latency_2/demon_left_precision': 0.7162661737523105, 'eval/latency_2/demon_left_recall': 0.8157894736842105, 'eval/latency_2/demon_left_f1': 0.7627952755905512, 'eval/latency_2/demon_left_rate_model': 0.30105731775180855, 'eval/latency_2/demon_left_rate_teacher': 0.2643294379521425, 'eval/latency_2/demon_right_precision': 0.7123681687440077, 'eval/latency_2/demon_right_recall': 0.8395480225988701, 'eval/latency_2/demon_right_f1': 0.7707468879668049, 'eval/latency_2/demon_right_rate_model': 0.29020589872008906, 'eval/latency_2/demon_right_rate_teacher': 0.24624373956594323, 'eval/latency_2/demon_move_f1': 0.766771081778678, 'eval/latency_2/demon_cc_f1': 0.8579704681449917, 'timing/data': 0.019619532860815525, 'timing/model': 1.0542260259389877, 'learning_rate/qwen_vl_interface': 8.646770131435774e-06, 'learning_rate/action_model': 8.646770131435773e-05, 'epoch': 0.66}) +[2026-06-17 19:48:43,902][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_1800_state +[2026-06-17 19:48:43,910][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 19:48:52,830][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_1800_state/model.safetensors +[2026-06-17 19:49:07,649][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_1800_state/optimizer.bin +[2026-06-17 19:49:07,650][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_1800_state/sampler.bin +[2026-06-17 19:49:07,651][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_1800_state/sampler_1.bin +[2026-06-17 19:49:07,653][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_1800_state/random_states_0.pkl +[2026-06-17 19:49:07,657][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1800 +[2026-06-17 19:49:07,657][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 19:49:07,674][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 19:49:15,067][starVLA.training.train_starvla][INFO] - Step 1801, Loss: {'train/loss': 1.0317829847335815, 'train/grad_norm_pre_clip': 15.681199073791504, 'timing/data': 0.024207566864788532, 'timing/model': 1.0417231153696775, 'learning_rate/qwen_vl_interface': 8.645258031766116e-06, 'learning_rate/action_model': 8.645258031766116e-05, 'epoch': 0.66}) +[2026-06-17 19:49:21,045][starVLA.training.train_starvla][INFO] - Step 1802, Loss: {'train/loss': 0.36668434739112854, 'train/grad_norm_pre_clip': 8.879165649414062, 'timing/data': 0.02186413574963808, 'timing/model': 1.3424782315269113, 'learning_rate/qwen_vl_interface': 8.643745228255816e-06, 'learning_rate/action_model': 8.643745228255817e-05, 'epoch': 0.66}) +[2026-06-17 19:49:25,386][starVLA.training.train_starvla][INFO] - Step 1803, Loss: {'train/loss': 0.905482828617096, 'train/grad_norm_pre_clip': 13.355778694152832, 'timing/data': 0.02064923755824566, 'timing/model': 1.0578028047457337, 'learning_rate/qwen_vl_interface': 8.64223172121848e-06, 'learning_rate/action_model': 8.642231721218479e-05, 'epoch': 0.66}) +[2026-06-17 19:49:30,516][starVLA.training.train_starvla][INFO] - Step 1804, Loss: {'train/loss': 0.7345325350761414, 'train/grad_norm_pre_clip': 13.991185188293457, 'timing/data': 0.029668118804693222, 'timing/model': 1.1375405034050345, 'learning_rate/qwen_vl_interface': 8.640717510967857e-06, 'learning_rate/action_model': 8.640717510967856e-05, 'epoch': 0.66}) +[2026-06-17 19:49:35,150][starVLA.training.train_starvla][INFO] - Step 1805, Loss: {'train/loss': 0.5446573495864868, 'train/grad_norm_pre_clip': 11.135255813598633, 'timing/data': 0.024249272421002388, 'timing/model': 1.0262931874021888, 'learning_rate/qwen_vl_interface': 8.639202597817844e-06, 'learning_rate/action_model': 8.639202597817844e-05, 'epoch': 0.66}) +[2026-06-17 19:49:39,816][starVLA.training.train_starvla][INFO] - Step 1806, Loss: {'train/loss': 0.5943170189857483, 'train/grad_norm_pre_clip': 11.183501243591309, 'timing/data': 0.02605010662227869, 'timing/model': 1.0670745121315122, 'learning_rate/qwen_vl_interface': 8.637686982082487e-06, 'learning_rate/action_model': 8.637686982082487e-05, 'epoch': 0.66}) +[2026-06-17 19:49:44,106][starVLA.training.train_starvla][INFO] - Step 1807, Loss: {'train/loss': 0.6086452603340149, 'train/grad_norm_pre_clip': 10.959749221801758, 'timing/data': 0.02227039262652397, 'timing/model': 1.1201247610151768, 'learning_rate/qwen_vl_interface': 8.636170664075972e-06, 'learning_rate/action_model': 8.636170664075972e-05, 'epoch': 0.66}) +[2026-06-17 19:49:48,486][starVLA.training.train_starvla][INFO] - Step 1808, Loss: {'train/loss': 0.9286558628082275, 'train/grad_norm_pre_clip': 13.822484016418457, 'timing/data': 0.021228128112852573, 'timing/model': 1.0547701604664326, 'learning_rate/qwen_vl_interface': 8.634653644112636e-06, 'learning_rate/action_model': 8.634653644112635e-05, 'epoch': 0.66}) +[2026-06-17 19:49:53,088][starVLA.training.train_starvla][INFO] - Step 1809, Loss: {'train/loss': 0.7887710332870483, 'train/grad_norm_pre_clip': 13.309646606445312, 'timing/data': 0.020251724869012833, 'timing/model': 1.1130178347229958, 'learning_rate/qwen_vl_interface': 8.633135922506957e-06, 'learning_rate/action_model': 8.633135922506957e-05, 'epoch': 0.66}) +[2026-06-17 19:49:57,711][starVLA.training.train_starvla][INFO] - Step 1810, Loss: {'train/loss': 0.7331168055534363, 'train/grad_norm_pre_clip': 11.727325439453125, 'timing/data': 0.020655494183301926, 'timing/model': 1.207053603604436, 'learning_rate/qwen_vl_interface': 8.63161749957356e-06, 'learning_rate/action_model': 8.63161749957356e-05, 'epoch': 0.66}) +[2026-06-17 19:50:02,113][starVLA.training.train_starvla][INFO] - Step 1811, Loss: {'train/loss': 0.7640330791473389, 'train/grad_norm_pre_clip': 11.901056289672852, 'timing/data': 0.026506444439291954, 'timing/model': 1.4316033674404025, 'learning_rate/qwen_vl_interface': 8.630098375627216e-06, 'learning_rate/action_model': 8.630098375627217e-05, 'epoch': 0.66}) +[2026-06-17 19:50:06,874][starVLA.training.train_starvla][INFO] - Step 1812, Loss: {'train/loss': 0.5748326778411865, 'train/grad_norm_pre_clip': 10.923060417175293, 'timing/data': 0.03316861763596535, 'timing/model': 1.6086916271597147, 'learning_rate/qwen_vl_interface': 8.628578550982846e-06, 'learning_rate/action_model': 8.628578550982845e-05, 'epoch': 0.66}) +[2026-06-17 19:50:10,924][starVLA.training.train_starvla][INFO] - Step 1813, Loss: {'train/loss': 0.4808081388473511, 'train/grad_norm_pre_clip': 12.359750747680664, 'timing/data': 0.024331851862370968, 'timing/model': 1.0348634533584118, 'learning_rate/qwen_vl_interface': 8.627058025955504e-06, 'learning_rate/action_model': 8.627058025955503e-05, 'epoch': 0.66}) +[2026-06-17 19:50:20,280][starVLA.training.train_starvla][INFO] - Step 1814, Loss: {'train/loss': 0.9130160808563232, 'train/grad_norm_pre_clip': 15.909341812133789, 'timing/data': 0.020451262593269348, 'timing/model': 3.2431951332837343, 'learning_rate/qwen_vl_interface': 8.6255368008604e-06, 'learning_rate/action_model': 8.625536800860398e-05, 'epoch': 0.66}) +[2026-06-17 19:50:24,995][starVLA.training.train_starvla][INFO] - Step 1815, Loss: {'train/loss': 0.9030261039733887, 'train/grad_norm_pre_clip': 12.997721672058105, 'timing/data': 0.02101538795977831, 'timing/model': 1.041581442579627, 'learning_rate/qwen_vl_interface': 8.624014876012883e-06, 'learning_rate/action_model': 8.624014876012882e-05, 'epoch': 0.66}) +[2026-06-17 19:50:29,512][starVLA.training.train_starvla][INFO] - Step 1816, Loss: {'train/loss': 0.8170953989028931, 'train/grad_norm_pre_clip': 14.352030754089355, 'timing/data': 0.025090958923101425, 'timing/model': 1.1744392281398177, 'learning_rate/qwen_vl_interface': 8.622492251728453e-06, 'learning_rate/action_model': 8.622492251728452e-05, 'epoch': 0.66}) +[2026-06-17 19:50:33,955][starVLA.training.train_starvla][INFO] - Step 1817, Loss: {'train/loss': 0.7372880578041077, 'train/grad_norm_pre_clip': 12.379417419433594, 'timing/data': 0.022508231922984123, 'timing/model': 1.4495838452130556, 'learning_rate/qwen_vl_interface': 8.62096892832275e-06, 'learning_rate/action_model': 8.620968928322749e-05, 'epoch': 0.67}) +[2026-06-17 19:50:38,106][starVLA.training.train_starvla][INFO] - Step 1818, Loss: {'train/loss': 0.5797772407531738, 'train/grad_norm_pre_clip': 12.928906440734863, 'timing/data': 0.02545384131371975, 'timing/model': 1.0448634708300233, 'learning_rate/qwen_vl_interface': 8.619444906111562e-06, 'learning_rate/action_model': 8.619444906111561e-05, 'epoch': 0.67}) +[2026-06-17 19:50:42,494][starVLA.training.train_starvla][INFO] - Step 1819, Loss: {'train/loss': 0.5120853185653687, 'train/grad_norm_pre_clip': 9.494959831237793, 'timing/data': 0.02025773376226425, 'timing/model': 1.0329360701143742, 'learning_rate/qwen_vl_interface': 8.617920185410817e-06, 'learning_rate/action_model': 8.617920185410817e-05, 'epoch': 0.67}) +[2026-06-17 19:50:47,099][starVLA.training.train_starvla][INFO] - Step 1820, Loss: {'train/loss': 1.1571855545043945, 'train/grad_norm_pre_clip': 13.46182918548584, 'timing/data': 0.02531738579273224, 'timing/model': 1.127611362375319, 'learning_rate/qwen_vl_interface': 8.616394766536594e-06, 'learning_rate/action_model': 8.616394766536593e-05, 'epoch': 0.67}) +[2026-06-17 19:50:51,532][starVLA.training.train_starvla][INFO] - Step 1821, Loss: {'train/loss': 0.9101342558860779, 'train/grad_norm_pre_clip': 17.156600952148438, 'timing/data': 0.024938922375440598, 'timing/model': 1.0225938390940428, 'learning_rate/qwen_vl_interface': 8.614868649805115e-06, 'learning_rate/action_model': 8.614868649805114e-05, 'epoch': 0.67}) +[2026-06-17 19:50:58,270][starVLA.training.train_starvla][INFO] - Step 1822, Loss: {'train/loss': 0.6601517200469971, 'train/grad_norm_pre_clip': 12.622220039367676, 'timing/data': 0.023261425085365772, 'timing/model': 1.3428790345788002, 'learning_rate/qwen_vl_interface': 8.613341835532743e-06, 'learning_rate/action_model': 8.613341835532743e-05, 'epoch': 0.67}) +[2026-06-17 19:51:02,574][starVLA.training.train_starvla][INFO] - Step 1823, Loss: {'train/loss': 0.5663524866104126, 'train/grad_norm_pre_clip': 13.487174987792969, 'timing/data': 0.028905910439789295, 'timing/model': 1.0658152075484395, 'learning_rate/qwen_vl_interface': 8.611814324035989e-06, 'learning_rate/action_model': 8.61181432403599e-05, 'epoch': 0.67}) +[2026-06-17 19:51:07,057][starVLA.training.train_starvla][INFO] - Step 1824, Loss: {'train/loss': 0.5706176161766052, 'train/grad_norm_pre_clip': 11.145676612854004, 'timing/data': 0.021517368964850903, 'timing/model': 1.0800916133448482, 'learning_rate/qwen_vl_interface': 8.610286115631508e-06, 'learning_rate/action_model': 8.610286115631509e-05, 'epoch': 0.67}) +[2026-06-17 19:51:11,703][starVLA.training.train_starvla][INFO] - Step 1825, Loss: {'train/loss': 0.7718448638916016, 'train/grad_norm_pre_clip': 11.728477478027344, 'timing/data': 0.029370004311203957, 'timing/model': 1.104752803221345, 'learning_rate/qwen_vl_interface': 8.608757210636101e-06, 'learning_rate/action_model': 8.608757210636101e-05, 'epoch': 0.67}) +[2026-06-17 19:51:17,472][starVLA.training.train_starvla][INFO] - Step 1826, Loss: {'train/loss': 0.722663938999176, 'train/grad_norm_pre_clip': 13.384105682373047, 'timing/data': 0.02114691026508808, 'timing/model': 2.237035342492163, 'learning_rate/qwen_vl_interface': 8.60722760936671e-06, 'learning_rate/action_model': 8.60722760936671e-05, 'epoch': 0.67}) +[2026-06-17 19:51:22,001][starVLA.training.train_starvla][INFO] - Step 1827, Loss: {'train/loss': 0.8022003173828125, 'train/grad_norm_pre_clip': 15.546689987182617, 'timing/data': 0.020863373763859272, 'timing/model': 1.3802091870456934, 'learning_rate/qwen_vl_interface': 8.605697312140421e-06, 'learning_rate/action_model': 8.605697312140422e-05, 'epoch': 0.67}) +[2026-06-17 19:51:26,043][starVLA.training.train_starvla][INFO] - Step 1828, Loss: {'train/loss': 0.7117701768875122, 'train/grad_norm_pre_clip': 13.190550804138184, 'timing/data': 0.02189852111041546, 'timing/model': 1.0747475791722536, 'learning_rate/qwen_vl_interface': 8.604166319274472e-06, 'learning_rate/action_model': 8.604166319274471e-05, 'epoch': 0.67}) +[2026-06-17 19:51:30,401][starVLA.training.train_starvla][INFO] - Step 1829, Loss: {'train/loss': 0.6035366058349609, 'train/grad_norm_pre_clip': 9.341647148132324, 'timing/data': 0.021751295775175095, 'timing/model': 1.0477736499160528, 'learning_rate/qwen_vl_interface': 8.602634631086233e-06, 'learning_rate/action_model': 8.602634631086233e-05, 'epoch': 0.67}) +[2026-06-17 19:51:36,025][starVLA.training.train_starvla][INFO] - Step 1830, Loss: {'train/loss': 0.9196500182151794, 'train/grad_norm_pre_clip': 16.67430877685547, 'timing/data': 0.022073399275541306, 'timing/model': 1.051010956056416, 'learning_rate/qwen_vl_interface': 8.601102247893227e-06, 'learning_rate/action_model': 8.601102247893226e-05, 'epoch': 0.67}) +[2026-06-17 19:51:41,582][starVLA.training.train_starvla][INFO] - Step 1831, Loss: {'train/loss': 0.5823948383331299, 'train/grad_norm_pre_clip': 12.760139465332031, 'timing/data': 0.02277906145900488, 'timing/model': 1.9684023829177022, 'learning_rate/qwen_vl_interface': 8.599569170013121e-06, 'learning_rate/action_model': 8.59956917001312e-05, 'epoch': 0.67}) +[2026-06-17 19:51:45,978][starVLA.training.train_starvla][INFO] - Step 1832, Loss: {'train/loss': 0.7494455575942993, 'train/grad_norm_pre_clip': 12.192488670349121, 'timing/data': 0.020290758460760117, 'timing/model': 1.3614573888480663, 'learning_rate/qwen_vl_interface': 8.59803539776372e-06, 'learning_rate/action_model': 8.59803539776372e-05, 'epoch': 0.67}) +[2026-06-17 19:51:50,066][starVLA.training.train_starvla][INFO] - Step 1833, Loss: {'train/loss': 0.5758930444717407, 'train/grad_norm_pre_clip': 10.059954643249512, 'timing/data': 0.021324566565454006, 'timing/model': 1.0447815684601665, 'learning_rate/qwen_vl_interface': 8.596500931462979e-06, 'learning_rate/action_model': 8.596500931462979e-05, 'epoch': 0.67}) +[2026-06-17 19:51:56,805][starVLA.training.train_starvla][INFO] - Step 1834, Loss: {'train/loss': 0.9084370136260986, 'train/grad_norm_pre_clip': 16.60007667541504, 'timing/data': 0.02243809588253498, 'timing/model': 1.0994458869099617, 'learning_rate/qwen_vl_interface': 8.594965771428993e-06, 'learning_rate/action_model': 8.594965771428993e-05, 'epoch': 0.67}) +[2026-06-17 19:52:01,192][starVLA.training.train_starvla][INFO] - Step 1835, Loss: {'train/loss': 0.6862971186637878, 'train/grad_norm_pre_clip': 10.685795783996582, 'timing/data': 0.021648989990353584, 'timing/model': 1.0346884755417705, 'learning_rate/qwen_vl_interface': 8.593429917980006e-06, 'learning_rate/action_model': 8.593429917980006e-05, 'epoch': 0.67}) +[2026-06-17 19:52:05,684][starVLA.training.train_starvla][INFO] - Step 1836, Loss: {'train/loss': 1.1516242027282715, 'train/grad_norm_pre_clip': 15.961139678955078, 'timing/data': 0.023787789046764374, 'timing/model': 1.0881208898499608, 'learning_rate/qwen_vl_interface': 8.591893371434397e-06, 'learning_rate/action_model': 8.591893371434396e-05, 'epoch': 0.67}) +[2026-06-17 19:52:10,326][starVLA.training.train_starvla][INFO] - Step 1837, Loss: {'train/loss': 0.7884481549263, 'train/grad_norm_pre_clip': 19.712051391601562, 'timing/data': 0.020384115166962147, 'timing/model': 1.4669127007946372, 'learning_rate/qwen_vl_interface': 8.590356132110695e-06, 'learning_rate/action_model': 8.590356132110695e-05, 'epoch': 0.67}) +[2026-06-17 19:52:14,673][starVLA.training.train_starvla][INFO] - Step 1838, Loss: {'train/loss': 0.8644999265670776, 'train/grad_norm_pre_clip': 14.937925338745117, 'timing/data': 0.021260494366288185, 'timing/model': 1.2305478416383266, 'learning_rate/qwen_vl_interface': 8.588818200327575e-06, 'learning_rate/action_model': 8.588818200327575e-05, 'epoch': 0.67}) +[2026-06-17 19:52:19,284][starVLA.training.train_starvla][INFO] - Step 1839, Loss: {'train/loss': 0.7996457815170288, 'train/grad_norm_pre_clip': 13.94688892364502, 'timing/data': 0.021562759764492512, 'timing/model': 1.0950152091681957, 'learning_rate/qwen_vl_interface': 8.587279576403848e-06, 'learning_rate/action_model': 8.587279576403847e-05, 'epoch': 0.67}) +[2026-06-17 19:52:23,716][starVLA.training.train_starvla][INFO] - Step 1840, Loss: {'train/loss': 0.8906955718994141, 'train/grad_norm_pre_clip': 13.714653015136719, 'timing/data': 0.020912387408316135, 'timing/model': 1.0509153669700027, 'learning_rate/qwen_vl_interface': 8.585740260658475e-06, 'learning_rate/action_model': 8.585740260658475e-05, 'epoch': 0.67}) +[2026-06-17 19:52:28,139][starVLA.training.train_starvla][INFO] - Step 1841, Loss: {'train/loss': 0.6279063820838928, 'train/grad_norm_pre_clip': 11.112289428710938, 'timing/data': 0.02256591245532036, 'timing/model': 1.049234109930694, 'learning_rate/qwen_vl_interface': 8.584200253410557e-06, 'learning_rate/action_model': 8.584200253410556e-05, 'epoch': 0.67}) +[2026-06-17 19:52:32,410][starVLA.training.train_starvla][INFO] - Step 1842, Loss: {'train/loss': 0.6961316466331482, 'train/grad_norm_pre_clip': 12.611786842346191, 'timing/data': 0.020625390112400055, 'timing/model': 1.1198001569136977, 'learning_rate/qwen_vl_interface': 8.582659554979336e-06, 'learning_rate/action_model': 8.582659554979336e-05, 'epoch': 0.67}) +[2026-06-17 19:52:37,006][starVLA.training.train_starvla][INFO] - Step 1843, Loss: {'train/loss': 0.7059768438339233, 'train/grad_norm_pre_clip': 14.465286254882812, 'timing/data': 0.021942592225968838, 'timing/model': 1.032180741429329, 'learning_rate/qwen_vl_interface': 8.581118165684204e-06, 'learning_rate/action_model': 8.581118165684204e-05, 'epoch': 0.67}) +[2026-06-17 19:52:41,550][starVLA.training.train_starvla][INFO] - Step 1844, Loss: {'train/loss': 0.8823834657669067, 'train/grad_norm_pre_clip': 13.3371000289917, 'timing/data': 0.024678055197000504, 'timing/model': 1.0571504114195704, 'learning_rate/qwen_vl_interface': 8.579576085844695e-06, 'learning_rate/action_model': 8.579576085844694e-05, 'epoch': 0.68}) +[2026-06-17 19:52:45,950][starVLA.training.train_starvla][INFO] - Step 1845, Loss: {'train/loss': 0.7563950419425964, 'train/grad_norm_pre_clip': 12.675030708312988, 'timing/data': 0.021250768564641476, 'timing/model': 1.0454215873032808, 'learning_rate/qwen_vl_interface': 8.578033315780474e-06, 'learning_rate/action_model': 8.578033315780475e-05, 'epoch': 0.68}) +[2026-06-17 19:52:50,721][starVLA.training.train_starvla][INFO] - Step 1846, Loss: {'train/loss': 0.7848904132843018, 'train/grad_norm_pre_clip': 11.777917861938477, 'timing/data': 0.023186239413917065, 'timing/model': 1.5742900371551514, 'learning_rate/qwen_vl_interface': 8.576489855811371e-06, 'learning_rate/action_model': 8.57648985581137e-05, 'epoch': 0.68}) +[2026-06-17 19:52:55,186][starVLA.training.train_starvla][INFO] - Step 1847, Loss: {'train/loss': 0.4136909544467926, 'train/grad_norm_pre_clip': 10.036345481872559, 'timing/data': 0.0222325436770916, 'timing/model': 1.4332874091342092, 'learning_rate/qwen_vl_interface': 8.57494570625734e-06, 'learning_rate/action_model': 8.57494570625734e-05, 'epoch': 0.68}) +[2026-06-17 19:53:00,592][starVLA.training.train_starvla][INFO] - Step 1848, Loss: {'train/loss': 0.7437028288841248, 'train/grad_norm_pre_clip': 12.405914306640625, 'timing/data': 0.02139344345778227, 'timing/model': 1.1643797680735588, 'learning_rate/qwen_vl_interface': 8.573400867438486e-06, 'learning_rate/action_model': 8.573400867438486e-05, 'epoch': 0.68}) +[2026-06-17 19:53:05,073][starVLA.training.train_starvla][INFO] - Step 1849, Loss: {'train/loss': 0.5787872076034546, 'train/grad_norm_pre_clip': 10.295899391174316, 'timing/data': 0.020352079533040524, 'timing/model': 1.088994107209146, 'learning_rate/qwen_vl_interface': 8.571855339675057e-06, 'learning_rate/action_model': 8.571855339675057e-05, 'epoch': 0.68}) +[2026-06-17 19:53:09,566][starVLA.training.train_starvla][INFO] - Step 1850, Loss: {'train/loss': 0.8109418749809265, 'train/grad_norm_pre_clip': 13.261299133300781, 'timing/data': 0.02600891888141632, 'timing/model': 1.0342309847474098, 'learning_rate/qwen_vl_interface': 8.57030912328744e-06, 'learning_rate/action_model': 8.57030912328744e-05, 'epoch': 0.68}) +[2026-06-17 19:53:14,016][starVLA.training.train_starvla][INFO] - Step 1851, Loss: {'train/loss': 0.577387273311615, 'train/grad_norm_pre_clip': 13.718905448913574, 'timing/data': 0.020539771765470505, 'timing/model': 1.0265995068475604, 'learning_rate/qwen_vl_interface': 8.568762218596172e-06, 'learning_rate/action_model': 8.568762218596171e-05, 'epoch': 0.68}) +[2026-06-17 19:53:21,206][starVLA.training.train_starvla][INFO] - Step 1852, Loss: {'train/loss': 0.6646799445152283, 'train/grad_norm_pre_clip': 15.517446517944336, 'timing/data': 0.024856816977262497, 'timing/model': 4.1007820796221495, 'learning_rate/qwen_vl_interface': 8.567214625921922e-06, 'learning_rate/action_model': 8.567214625921922e-05, 'epoch': 0.68}) +[2026-06-17 19:53:25,597][starVLA.training.train_starvla][INFO] - Step 1853, Loss: {'train/loss': 0.7207340598106384, 'train/grad_norm_pre_clip': 12.391047477722168, 'timing/data': 0.020882091484963894, 'timing/model': 1.104296219535172, 'learning_rate/qwen_vl_interface': 8.565666345585511e-06, 'learning_rate/action_model': 8.56566634558551e-05, 'epoch': 0.68}) +[2026-06-17 19:53:30,123][starVLA.training.train_starvla][INFO] - Step 1854, Loss: {'train/loss': 0.6734055876731873, 'train/grad_norm_pre_clip': 12.104921340942383, 'timing/data': 0.021612598560750484, 'timing/model': 1.0650318693369627, 'learning_rate/qwen_vl_interface': 8.564117377907898e-06, 'learning_rate/action_model': 8.564117377907898e-05, 'epoch': 0.68}) +[2026-06-17 19:53:34,495][starVLA.training.train_starvla][INFO] - Step 1855, Loss: {'train/loss': 0.9379417300224304, 'train/grad_norm_pre_clip': 15.884217262268066, 'timing/data': 0.022505571134388447, 'timing/model': 1.0265223998576403, 'learning_rate/qwen_vl_interface': 8.56256772321019e-06, 'learning_rate/action_model': 8.562567723210188e-05, 'epoch': 0.68}) +[2026-06-17 19:53:39,027][starVLA.training.train_starvla][INFO] - Step 1856, Loss: {'train/loss': 0.5242384076118469, 'train/grad_norm_pre_clip': 9.773857116699219, 'timing/data': 0.020508669316768646, 'timing/model': 1.2296893568709493, 'learning_rate/qwen_vl_interface': 8.561017381813624e-06, 'learning_rate/action_model': 8.561017381813624e-05, 'epoch': 0.68}) +[2026-06-17 19:53:43,642][starVLA.training.train_starvla][INFO] - Step 1857, Loss: {'train/loss': 0.6293091773986816, 'train/grad_norm_pre_clip': 13.736739158630371, 'timing/data': 0.02729166392236948, 'timing/model': 1.4750253297388554, 'learning_rate/qwen_vl_interface': 8.559466354039594e-06, 'learning_rate/action_model': 8.559466354039594e-05, 'epoch': 0.68}) +[2026-06-17 19:53:47,804][starVLA.training.train_starvla][INFO] - Step 1858, Loss: {'train/loss': 0.9732372164726257, 'train/grad_norm_pre_clip': 20.765962600708008, 'timing/data': 0.023192908614873886, 'timing/model': 1.083254843018949, 'learning_rate/qwen_vl_interface': 8.557914640209628e-06, 'learning_rate/action_model': 8.557914640209627e-05, 'epoch': 0.68}) +[2026-06-17 19:53:52,258][starVLA.training.train_starvla][INFO] - Step 1859, Loss: {'train/loss': 0.8341982960700989, 'train/grad_norm_pre_clip': 14.05089282989502, 'timing/data': 0.026745223440229893, 'timing/model': 1.0420685913413763, 'learning_rate/qwen_vl_interface': 8.556362240645397e-06, 'learning_rate/action_model': 8.556362240645397e-05, 'epoch': 0.68}) +[2026-06-17 19:53:56,756][starVLA.training.train_starvla][INFO] - Step 1860, Loss: {'train/loss': 0.728979766368866, 'train/grad_norm_pre_clip': 12.913349151611328, 'timing/data': 0.0226186104118824, 'timing/model': 1.071367010474205, 'learning_rate/qwen_vl_interface': 8.554809155668716e-06, 'learning_rate/action_model': 8.554809155668716e-05, 'epoch': 0.68}) +[2026-06-17 19:54:02,816][starVLA.training.train_starvla][INFO] - Step 1861, Loss: {'train/loss': 0.796066403388977, 'train/grad_norm_pre_clip': 13.61827278137207, 'timing/data': 0.021464018151164055, 'timing/model': 1.5332739213481545, 'learning_rate/qwen_vl_interface': 8.553255385601541e-06, 'learning_rate/action_model': 8.55325538560154e-05, 'epoch': 0.68}) +[2026-06-17 19:54:07,262][starVLA.training.train_starvla][INFO] - Step 1862, Loss: {'train/loss': 0.5568937659263611, 'train/grad_norm_pre_clip': 11.55280590057373, 'timing/data': 0.02229208406060934, 'timing/model': 1.4368926351889968, 'learning_rate/qwen_vl_interface': 8.551700930765969e-06, 'learning_rate/action_model': 8.551700930765969e-05, 'epoch': 0.68}) +[2026-06-17 19:54:11,646][starVLA.training.train_starvla][INFO] - Step 1863, Loss: {'train/loss': 0.8145684599876404, 'train/grad_norm_pre_clip': 15.85020923614502, 'timing/data': 0.06988511700183153, 'timing/model': 1.1256902301684022, 'learning_rate/qwen_vl_interface': 8.550145791484243e-06, 'learning_rate/action_model': 8.550145791484242e-05, 'epoch': 0.68}) +[2026-06-17 19:54:16,108][starVLA.training.train_starvla][INFO] - Step 1864, Loss: {'train/loss': 0.8278127312660217, 'train/grad_norm_pre_clip': 14.382225036621094, 'timing/data': 0.020717205479741096, 'timing/model': 1.0665104174986482, 'learning_rate/qwen_vl_interface': 8.548589968078743e-06, 'learning_rate/action_model': 8.548589968078743e-05, 'epoch': 0.68}) +[2026-06-17 19:54:22,985][starVLA.training.train_starvla][INFO] - Step 1865, Loss: {'train/loss': 0.8056464791297913, 'train/grad_norm_pre_clip': 14.472749710083008, 'timing/data': 0.022314520552754402, 'timing/model': 2.7011571917682886, 'learning_rate/qwen_vl_interface': 8.547033460871992e-06, 'learning_rate/action_model': 8.547033460871992e-05, 'epoch': 0.68}) +[2026-06-17 19:54:27,646][starVLA.training.train_starvla][INFO] - Step 1866, Loss: {'train/loss': 0.489090234041214, 'train/grad_norm_pre_clip': 14.494867324829102, 'timing/data': 0.028817400336265564, 'timing/model': 1.0984012642875314, 'learning_rate/qwen_vl_interface': 8.545476270186658e-06, 'learning_rate/action_model': 8.545476270186659e-05, 'epoch': 0.68}) +[2026-06-17 19:54:33,477][starVLA.training.train_starvla][INFO] - Step 1867, Loss: {'train/loss': 0.7849369049072266, 'train/grad_norm_pre_clip': 13.632680892944336, 'timing/data': 0.026812206953763962, 'timing/model': 1.4449503077194095, 'learning_rate/qwen_vl_interface': 8.543918396345546e-06, 'learning_rate/action_model': 8.543918396345546e-05, 'epoch': 0.68}) +[2026-06-17 19:54:37,673][starVLA.training.train_starvla][INFO] - Step 1868, Loss: {'train/loss': 0.8112380504608154, 'train/grad_norm_pre_clip': 17.459068298339844, 'timing/data': 0.024520810693502426, 'timing/model': 1.1558719128370285, 'learning_rate/qwen_vl_interface': 8.542359839671607e-06, 'learning_rate/action_model': 8.542359839671606e-05, 'epoch': 0.68}) +[2026-06-17 19:54:42,263][starVLA.training.train_starvla][INFO] - Step 1869, Loss: {'train/loss': 0.6174426674842834, 'train/grad_norm_pre_clip': 11.105854034423828, 'timing/data': 0.023830853402614594, 'timing/model': 1.0737751610577106, 'learning_rate/qwen_vl_interface': 8.540800600487931e-06, 'learning_rate/action_model': 8.54080060048793e-05, 'epoch': 0.68}) +[2026-06-17 19:54:46,737][starVLA.training.train_starvla][INFO] - Step 1870, Loss: {'train/loss': 1.217227578163147, 'train/grad_norm_pre_clip': 15.0059814453125, 'timing/data': 0.021845565177500248, 'timing/model': 1.0284692011773586, 'learning_rate/qwen_vl_interface': 8.539240679117748e-06, 'learning_rate/action_model': 8.539240679117748e-05, 'epoch': 0.68}) +[2026-06-17 19:54:51,146][starVLA.training.train_starvla][INFO] - Step 1871, Loss: {'train/loss': 1.1572093963623047, 'train/grad_norm_pre_clip': 15.430850982666016, 'timing/data': 0.021538632921874523, 'timing/model': 1.0564433010295033, 'learning_rate/qwen_vl_interface': 8.537680075884433e-06, 'learning_rate/action_model': 8.537680075884433e-05, 'epoch': 0.69}) +[2026-06-17 19:55:00,499][starVLA.training.train_starvla][INFO] - Step 1872, Loss: {'train/loss': 0.7052357792854309, 'train/grad_norm_pre_clip': 13.147937774658203, 'timing/data': 0.02162688411772251, 'timing/model': 1.3213851805776358, 'learning_rate/qwen_vl_interface': 8.5361187911115e-06, 'learning_rate/action_model': 8.5361187911115e-05, 'epoch': 0.69}) +[2026-06-17 19:55:04,648][starVLA.training.train_starvla][INFO] - Step 1873, Loss: {'train/loss': 0.8116673231124878, 'train/grad_norm_pre_clip': 13.951924324035645, 'timing/data': 0.02014935575425625, 'timing/model': 1.1062535354867578, 'learning_rate/qwen_vl_interface': 8.534556825122608e-06, 'learning_rate/action_model': 8.534556825122607e-05, 'epoch': 0.69}) +[2026-06-17 19:55:09,389][starVLA.training.train_starvla][INFO] - Step 1874, Loss: {'train/loss': 0.6804594993591309, 'train/grad_norm_pre_clip': 15.258919715881348, 'timing/data': 0.023270762525498867, 'timing/model': 1.1348960446193814, 'learning_rate/qwen_vl_interface': 8.532994178241551e-06, 'learning_rate/action_model': 8.53299417824155e-05, 'epoch': 0.69}) +[2026-06-17 19:55:13,869][starVLA.training.train_starvla][INFO] - Step 1875, Loss: {'train/loss': 0.6691452264785767, 'train/grad_norm_pre_clip': 15.54604434967041, 'timing/data': 0.02480575628578663, 'timing/model': 1.061868385411799, 'learning_rate/qwen_vl_interface': 8.531430850792267e-06, 'learning_rate/action_model': 8.531430850792266e-05, 'epoch': 0.69}) +[2026-06-17 19:55:18,258][starVLA.training.train_starvla][INFO] - Step 1876, Loss: {'train/loss': 0.7310897707939148, 'train/grad_norm_pre_clip': 11.91794204711914, 'timing/data': 0.021118255332112312, 'timing/model': 1.0340634025633335, 'learning_rate/qwen_vl_interface': 8.529866843098838e-06, 'learning_rate/action_model': 8.529866843098837e-05, 'epoch': 0.69}) +[2026-06-17 19:55:24,614][starVLA.training.train_starvla][INFO] - Step 1877, Loss: {'train/loss': 0.8833509683609009, 'train/grad_norm_pre_clip': 15.289629936218262, 'timing/data': 0.022086555138230324, 'timing/model': 1.1650627786293626, 'learning_rate/qwen_vl_interface': 8.528302155485482e-06, 'learning_rate/action_model': 8.528302155485481e-05, 'epoch': 0.69}) +[2026-06-17 19:55:29,279][starVLA.training.train_starvla][INFO] - Step 1878, Loss: {'train/loss': 0.7220128774642944, 'train/grad_norm_pre_clip': 13.106026649475098, 'timing/data': 0.020017790608108044, 'timing/model': 1.100240359082818, 'learning_rate/qwen_vl_interface': 8.526736788276563e-06, 'learning_rate/action_model': 8.526736788276563e-05, 'epoch': 0.69}) +[2026-06-17 19:55:33,916][starVLA.training.train_starvla][INFO] - Step 1879, Loss: {'train/loss': 0.4881575405597687, 'train/grad_norm_pre_clip': 11.69123363494873, 'timing/data': 0.021891859360039234, 'timing/model': 1.113508972339332, 'learning_rate/qwen_vl_interface': 8.525170741796578e-06, 'learning_rate/action_model': 8.525170741796579e-05, 'epoch': 0.69}) +[2026-06-17 19:55:38,561][starVLA.training.train_starvla][INFO] - Step 1880, Loss: {'train/loss': 0.8804976344108582, 'train/grad_norm_pre_clip': 13.334916114807129, 'timing/data': 0.022692049853503704, 'timing/model': 1.134580608457327, 'learning_rate/qwen_vl_interface': 8.523604016370179e-06, 'learning_rate/action_model': 8.523604016370179e-05, 'epoch': 0.69}) +[2026-06-17 19:55:43,015][starVLA.training.train_starvla][INFO] - Step 1881, Loss: {'train/loss': 0.8155738115310669, 'train/grad_norm_pre_clip': 19.039287567138672, 'timing/data': 0.0232896339148283, 'timing/model': 1.0275164805352688, 'learning_rate/qwen_vl_interface': 8.522036612322143e-06, 'learning_rate/action_model': 8.522036612322142e-05, 'epoch': 0.69}) +[2026-06-17 19:55:48,468][starVLA.training.train_starvla][INFO] - Step 1882, Loss: {'train/loss': 1.1762185096740723, 'train/grad_norm_pre_clip': 59.87508010864258, 'timing/data': 0.021629774011671543, 'timing/model': 1.9565077349543571, 'learning_rate/qwen_vl_interface': 8.520468529977395e-06, 'learning_rate/action_model': 8.520468529977395e-05, 'epoch': 0.69}) +[2026-06-17 19:55:54,982][starVLA.training.train_starvla][INFO] - Step 1883, Loss: {'train/loss': 0.5843126177787781, 'train/grad_norm_pre_clip': 11.136757850646973, 'timing/data': 0.031184034422039986, 'timing/model': 1.1278584599494934, 'learning_rate/qwen_vl_interface': 8.518899769661002e-06, 'learning_rate/action_model': 8.518899769661002e-05, 'epoch': 0.69}) +[2026-06-17 19:55:59,539][starVLA.training.train_starvla][INFO] - Step 1884, Loss: {'train/loss': 0.7289248108863831, 'train/grad_norm_pre_clip': 11.882101058959961, 'timing/data': 0.019726919941604137, 'timing/model': 1.0729428427293897, 'learning_rate/qwen_vl_interface': 8.517330331698171e-06, 'learning_rate/action_model': 8.51733033169817e-05, 'epoch': 0.69}) +[2026-06-17 19:56:04,179][starVLA.training.train_starvla][INFO] - Step 1885, Loss: {'train/loss': 0.58750981092453, 'train/grad_norm_pre_clip': 13.818575859069824, 'timing/data': 0.022516464814543724, 'timing/model': 1.079251753166318, 'learning_rate/qwen_vl_interface': 8.515760216414244e-06, 'learning_rate/action_model': 8.515760216414244e-05, 'epoch': 0.69}) +[2026-06-17 19:56:08,610][starVLA.training.train_starvla][INFO] - Step 1886, Loss: {'train/loss': 0.5601668357849121, 'train/grad_norm_pre_clip': 11.414628028869629, 'timing/data': 0.02221305575221777, 'timing/model': 1.070573735050857, 'learning_rate/qwen_vl_interface': 8.514189424134712e-06, 'learning_rate/action_model': 8.514189424134713e-05, 'epoch': 0.69}) +[2026-06-17 19:56:15,135][starVLA.training.train_starvla][INFO] - Step 1887, Loss: {'train/loss': 0.5925711989402771, 'train/grad_norm_pre_clip': 15.115729331970215, 'timing/data': 0.023545497097074986, 'timing/model': 3.464054769836366, 'learning_rate/qwen_vl_interface': 8.512617955185199e-06, 'learning_rate/action_model': 8.5126179551852e-05, 'epoch': 0.69}) +[2026-06-17 19:56:21,630][starVLA.training.train_starvla][INFO] - Step 1888, Loss: {'train/loss': 1.066107988357544, 'train/grad_norm_pre_clip': 16.987295150756836, 'timing/data': 0.024365411140024662, 'timing/model': 1.5289236884564161, 'learning_rate/qwen_vl_interface': 8.511045809891472e-06, 'learning_rate/action_model': 8.511045809891472e-05, 'epoch': 0.69}) +[2026-06-17 19:56:26,416][starVLA.training.train_starvla][INFO] - Step 1889, Loss: {'train/loss': 0.9884309768676758, 'train/grad_norm_pre_clip': 17.4622745513916, 'timing/data': 0.024921175092458725, 'timing/model': 1.0450114700943232, 'learning_rate/qwen_vl_interface': 8.509472988579443e-06, 'learning_rate/action_model': 8.509472988579443e-05, 'epoch': 0.69}) +[2026-06-17 19:56:31,001][starVLA.training.train_starvla][INFO] - Step 1890, Loss: {'train/loss': 0.5606884956359863, 'train/grad_norm_pre_clip': 13.465984344482422, 'timing/data': 0.02306924294680357, 'timing/model': 1.032355540432036, 'learning_rate/qwen_vl_interface': 8.507899491575153e-06, 'learning_rate/action_model': 8.507899491575154e-05, 'epoch': 0.69}) +[2026-06-17 19:56:36,989][starVLA.training.train_starvla][INFO] - Step 1891, Loss: {'train/loss': 1.0488853454589844, 'train/grad_norm_pre_clip': 15.323118209838867, 'timing/data': 0.03192935138940811, 'timing/model': 2.9351839711889625, 'learning_rate/qwen_vl_interface': 8.506325319204795e-06, 'learning_rate/action_model': 8.506325319204796e-05, 'epoch': 0.69}) +[2026-06-17 19:56:41,719][starVLA.training.train_starvla][INFO] - Step 1892, Loss: {'train/loss': 0.6335946321487427, 'train/grad_norm_pre_clip': 13.640849113464355, 'timing/data': 0.02372147887945175, 'timing/model': 1.4017919385805726, 'learning_rate/qwen_vl_interface': 8.504750471794694e-06, 'learning_rate/action_model': 8.504750471794693e-05, 'epoch': 0.69}) +[2026-06-17 19:56:47,604][starVLA.training.train_starvla][INFO] - Step 1893, Loss: {'train/loss': 0.7787553668022156, 'train/grad_norm_pre_clip': 14.525915145874023, 'timing/data': 0.021723204292356968, 'timing/model': 2.7154176319018006, 'learning_rate/qwen_vl_interface': 8.503174949671315e-06, 'learning_rate/action_model': 8.503174949671315e-05, 'epoch': 0.69}) +[2026-06-17 19:56:52,820][starVLA.training.train_starvla][INFO] - Step 1894, Loss: {'train/loss': 0.8450074791908264, 'train/grad_norm_pre_clip': 14.838542938232422, 'timing/data': 0.021838776767253876, 'timing/model': 1.0302522713318467, 'learning_rate/qwen_vl_interface': 8.501598753161273e-06, 'learning_rate/action_model': 8.501598753161272e-05, 'epoch': 0.69}) +[2026-06-17 19:56:57,640][starVLA.training.train_starvla][INFO] - Step 1895, Loss: {'train/loss': 0.8247668147087097, 'train/grad_norm_pre_clip': 13.783108711242676, 'timing/data': 0.02307269722223282, 'timing/model': 1.1316686812788248, 'learning_rate/qwen_vl_interface': 8.500021882591308e-06, 'learning_rate/action_model': 8.500021882591308e-05, 'epoch': 0.69}) +[2026-06-17 19:57:02,133][starVLA.training.train_starvla][INFO] - Step 1896, Loss: {'train/loss': 0.9432129263877869, 'train/grad_norm_pre_clip': 13.769347190856934, 'timing/data': 0.021439519710838795, 'timing/model': 1.0333397602662444, 'learning_rate/qwen_vl_interface': 8.498444338288308e-06, 'learning_rate/action_model': 8.498444338288309e-05, 'epoch': 0.69}) +[2026-06-17 19:57:12,796][starVLA.training.train_starvla][INFO] - Step 1897, Loss: {'train/loss': 0.5682161450386047, 'train/grad_norm_pre_clip': 11.373738288879395, 'timing/data': 0.023862257599830627, 'timing/model': 1.3322979668155313, 'learning_rate/qwen_vl_interface': 8.496866120579304e-06, 'learning_rate/action_model': 8.496866120579303e-05, 'epoch': 0.69}) +[2026-06-17 19:57:17,077][starVLA.training.train_starvla][INFO] - Step 1898, Loss: {'train/loss': 0.9308332204818726, 'train/grad_norm_pre_clip': 13.34084701538086, 'timing/data': 0.02149911690503359, 'timing/model': 1.0715119494125247, 'learning_rate/qwen_vl_interface': 8.495287229791459e-06, 'learning_rate/action_model': 8.495287229791459e-05, 'epoch': 0.7}) +[2026-06-17 19:57:23,909][starVLA.training.train_starvla][INFO] - Step 1899, Loss: {'train/loss': 0.8903933763504028, 'train/grad_norm_pre_clip': 12.074711799621582, 'timing/data': 0.030097507871687412, 'timing/model': 3.469396765343845, 'learning_rate/qwen_vl_interface': 8.493707666252078e-06, 'learning_rate/action_model': 8.493707666252078e-05, 'epoch': 0.7}) +[2026-06-17 20:01:22,549][starVLA.training.train_starvla][INFO] - Step 1900, Loss: {'train/loss': 0.6139743328094482, 'train/grad_norm_pre_clip': 11.640478134155273, 'eval/loss': 0.7972376346588135, 'eval/latency_2/loss': 0.7972376346588135, 'eval/demon_fire_precision': 0.897196261682243, 'eval/demon_fire_recall': 0.9238493723849373, 'eval/demon_fire_f1': 0.9103277674706245, 'eval/demon_fire_rate_model': 0.6847523650528659, 'eval/demon_fire_rate_teacher': 0.6649972175848636, 'eval/demon_left_precision': 0.8835098335854765, 'eval/demon_left_recall': 0.6147368421052631, 'eval/demon_left_f1': 0.7250155183116076, 'eval/demon_left_rate_model': 0.18391764051196438, 'eval/demon_left_rate_teacher': 0.2643294379521425, 'eval/demon_right_precision': 0.8395860284605433, 'eval/demon_right_recall': 0.7333333333333333, 'eval/demon_right_f1': 0.7828709288299155, 'eval/demon_right_rate_model': 0.2150806900389538, 'eval/demon_right_rate_teacher': 0.24624373956594323, 'eval/demon_move_f1': 0.7539432235707615, 'eval/demon_cc_f1': 0.8569995538238261, 'eval/latency_2/demon_fire_precision': 0.897196261682243, 'eval/latency_2/demon_fire_recall': 0.9238493723849373, 'eval/latency_2/demon_fire_f1': 0.9103277674706245, 'eval/latency_2/demon_fire_rate_model': 0.6847523650528659, 'eval/latency_2/demon_fire_rate_teacher': 0.6649972175848636, 'eval/latency_2/demon_left_precision': 0.8835098335854765, 'eval/latency_2/demon_left_recall': 0.6147368421052631, 'eval/latency_2/demon_left_f1': 0.7250155183116076, 'eval/latency_2/demon_left_rate_model': 0.18391764051196438, 'eval/latency_2/demon_left_rate_teacher': 0.2643294379521425, 'eval/latency_2/demon_right_precision': 0.8395860284605433, 'eval/latency_2/demon_right_recall': 0.7333333333333333, 'eval/latency_2/demon_right_f1': 0.7828709288299155, 'eval/latency_2/demon_right_rate_model': 0.2150806900389538, 'eval/latency_2/demon_right_rate_teacher': 0.24624373956594323, 'eval/latency_2/demon_move_f1': 0.7539432235707615, 'eval/latency_2/demon_cc_f1': 0.8569995538238261, 'timing/data': 0.02162482775747776, 'timing/model': 1.1773548116907477, 'learning_rate/qwen_vl_interface': 8.492127430288609e-06, 'learning_rate/action_model': 8.492127430288608e-05, 'epoch': 0.7}) +[2026-06-17 20:01:22,562][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_1900_state +[2026-06-17 20:01:22,569][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 20:01:31,185][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_1900_state/model.safetensors +[2026-06-17 20:01:45,268][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_1900_state/optimizer.bin +[2026-06-17 20:01:45,269][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_1900_state/sampler.bin +[2026-06-17 20:01:45,270][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_1900_state/sampler_1.bin +[2026-06-17 20:01:45,272][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_1900_state/random_states_0.pkl +[2026-06-17 20:01:45,276][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1900 +[2026-06-17 20:01:45,276][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 20:01:45,293][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 20:01:49,726][starVLA.training.train_starvla][INFO] - Step 1901, Loss: {'train/loss': 0.8859124779701233, 'train/grad_norm_pre_clip': 14.28500747680664, 'timing/data': 0.023108686320483685, 'timing/model': 1.0232707038521767, 'learning_rate/qwen_vl_interface': 8.490546522228633e-06, 'learning_rate/action_model': 8.490546522228633e-05, 'epoch': 0.7}) +[2026-06-17 20:01:54,269][starVLA.training.train_starvla][INFO] - Step 1902, Loss: {'train/loss': 0.6009731292724609, 'train/grad_norm_pre_clip': 12.445987701416016, 'timing/data': 0.023459622636437416, 'timing/model': 1.43725645262748, 'learning_rate/qwen_vl_interface': 8.488964942399876e-06, 'learning_rate/action_model': 8.488964942399877e-05, 'epoch': 0.7}) +[2026-06-17 20:01:58,567][starVLA.training.train_starvla][INFO] - Step 1903, Loss: {'train/loss': 0.6063592433929443, 'train/grad_norm_pre_clip': 10.637709617614746, 'timing/data': 0.020056970417499542, 'timing/model': 1.2900749342516065, 'learning_rate/qwen_vl_interface': 8.487382691130203e-06, 'learning_rate/action_model': 8.487382691130202e-05, 'epoch': 0.7}) +[2026-06-17 20:02:02,480][starVLA.training.train_starvla][INFO] - Step 1904, Loss: {'train/loss': 0.8963440656661987, 'train/grad_norm_pre_clip': 12.524182319641113, 'timing/data': 0.018869399093091488, 'timing/model': 1.0051078349351883, 'learning_rate/qwen_vl_interface': 8.485799768747614e-06, 'learning_rate/action_model': 8.485799768747614e-05, 'epoch': 0.7}) +[2026-06-17 20:02:06,540][starVLA.training.train_starvla][INFO] - Step 1905, Loss: {'train/loss': 0.7552967071533203, 'train/grad_norm_pre_clip': 12.03471851348877, 'timing/data': 0.019225834868848324, 'timing/model': 1.000802882015705, 'learning_rate/qwen_vl_interface': 8.484216175580248e-06, 'learning_rate/action_model': 8.484216175580249e-05, 'epoch': 0.7}) +[2026-06-17 20:02:10,644][starVLA.training.train_starvla][INFO] - Step 1906, Loss: {'train/loss': 0.7290719747543335, 'train/grad_norm_pre_clip': 17.086734771728516, 'timing/data': 0.01781820598989725, 'timing/model': 1.0051511526107788, 'learning_rate/qwen_vl_interface': 8.482631911956395e-06, 'learning_rate/action_model': 8.482631911956395e-05, 'epoch': 0.7}) +[2026-06-17 20:02:14,742][starVLA.training.train_starvla][INFO] - Step 1907, Loss: {'train/loss': 0.8866801857948303, 'train/grad_norm_pre_clip': 15.488173484802246, 'timing/data': 0.01806564535945654, 'timing/model': 1.0176940420642495, 'learning_rate/qwen_vl_interface': 8.481046978204462e-06, 'learning_rate/action_model': 8.481046978204463e-05, 'epoch': 0.7}) +[2026-06-17 20:02:19,001][starVLA.training.train_starvla][INFO] - Step 1908, Loss: {'train/loss': 0.9951480031013489, 'train/grad_norm_pre_clip': 17.376096725463867, 'timing/data': 0.018155284225940704, 'timing/model': 1.2112317029386759, 'learning_rate/qwen_vl_interface': 8.479461374653017e-06, 'learning_rate/action_model': 8.479461374653018e-05, 'epoch': 0.7}) +[2026-06-17 20:02:22,908][starVLA.training.train_starvla][INFO] - Step 1909, Loss: {'train/loss': 0.5934031009674072, 'train/grad_norm_pre_clip': 9.611408233642578, 'timing/data': 0.01779849175363779, 'timing/model': 1.005319464020431, 'learning_rate/qwen_vl_interface': 8.477875101630752e-06, 'learning_rate/action_model': 8.477875101630753e-05, 'epoch': 0.7}) +[2026-06-17 20:02:27,082][starVLA.training.train_starvla][INFO] - Step 1910, Loss: {'train/loss': 0.7498235702514648, 'train/grad_norm_pre_clip': 14.076170921325684, 'timing/data': 0.01817262638360262, 'timing/model': 1.0031180595979095, 'learning_rate/qwen_vl_interface': 8.476288159466508e-06, 'learning_rate/action_model': 8.476288159466508e-05, 'epoch': 0.7}) +[2026-06-17 20:02:31,264][starVLA.training.train_starvla][INFO] - Step 1911, Loss: {'train/loss': 0.7617266178131104, 'train/grad_norm_pre_clip': 14.522350311279297, 'timing/data': 0.021410896442830563, 'timing/model': 1.0655457992106676, 'learning_rate/qwen_vl_interface': 8.474700548489255e-06, 'learning_rate/action_model': 8.474700548489255e-05, 'epoch': 0.7}) +[2026-06-17 20:02:35,415][starVLA.training.train_starvla][INFO] - Step 1912, Loss: {'train/loss': 0.5282595157623291, 'train/grad_norm_pre_clip': 11.025739669799805, 'timing/data': 0.018911249935626984, 'timing/model': 1.0033418405801058, 'learning_rate/qwen_vl_interface': 8.473112269028108e-06, 'learning_rate/action_model': 8.473112269028107e-05, 'epoch': 0.7}) +[2026-06-17 20:02:40,611][starVLA.training.train_starvla][INFO] - Step 1913, Loss: {'train/loss': 0.773305356502533, 'train/grad_norm_pre_clip': 16.086414337158203, 'timing/data': 0.01844126731157303, 'timing/model': 1.250316645950079, 'learning_rate/qwen_vl_interface': 8.47152332141232e-06, 'learning_rate/action_model': 8.47152332141232e-05, 'epoch': 0.7}) +[2026-06-17 20:02:44,971][starVLA.training.train_starvla][INFO] - Step 1914, Loss: {'train/loss': 1.0555461645126343, 'train/grad_norm_pre_clip': 18.17699432373047, 'timing/data': 0.019399192184209824, 'timing/model': 1.0192940020933747, 'learning_rate/qwen_vl_interface': 8.469933705971282e-06, 'learning_rate/action_model': 8.469933705971282e-05, 'epoch': 0.7}) +[2026-06-17 20:02:49,113][starVLA.training.train_starvla][INFO] - Step 1915, Loss: {'train/loss': 0.8439525961875916, 'train/grad_norm_pre_clip': 16.13544273376465, 'timing/data': 0.020890465937554836, 'timing/model': 1.0394718321040273, 'learning_rate/qwen_vl_interface': 8.468343423034521e-06, 'learning_rate/action_model': 8.46834342303452e-05, 'epoch': 0.7}) +[2026-06-17 20:02:53,313][starVLA.training.train_starvla][INFO] - Step 1916, Loss: {'train/loss': 0.92516028881073, 'train/grad_norm_pre_clip': 15.349472999572754, 'timing/data': 0.01875862292945385, 'timing/model': 1.0152791449800134, 'learning_rate/qwen_vl_interface': 8.466752472931704e-06, 'learning_rate/action_model': 8.466752472931704e-05, 'epoch': 0.7}) +[2026-06-17 20:02:57,476][starVLA.training.train_starvla][INFO] - Step 1917, Loss: {'train/loss': 0.7066773772239685, 'train/grad_norm_pre_clip': 15.58298110961914, 'timing/data': 0.020679909735918045, 'timing/model': 1.018360586836934, 'learning_rate/qwen_vl_interface': 8.465160855992639e-06, 'learning_rate/action_model': 8.465160855992638e-05, 'epoch': 0.7}) +[2026-06-17 20:03:01,741][starVLA.training.train_starvla][INFO] - Step 1918, Loss: {'train/loss': 0.7751713991165161, 'train/grad_norm_pre_clip': 12.84054183959961, 'timing/data': 0.021286409348249435, 'timing/model': 1.3217122657224536, 'learning_rate/qwen_vl_interface': 8.46356857254727e-06, 'learning_rate/action_model': 8.46356857254727e-05, 'epoch': 0.7}) +[2026-06-17 20:03:05,848][starVLA.training.train_starvla][INFO] - Step 1919, Loss: {'train/loss': 0.8440389633178711, 'train/grad_norm_pre_clip': 14.55400562286377, 'timing/data': 0.019843525253236294, 'timing/model': 1.0694139879196882, 'learning_rate/qwen_vl_interface': 8.461975622925676e-06, 'learning_rate/action_model': 8.461975622925676e-05, 'epoch': 0.7}) +[2026-06-17 20:03:10,162][starVLA.training.train_starvla][INFO] - Step 1920, Loss: {'train/loss': 0.8644701838493347, 'train/grad_norm_pre_clip': 14.98257827758789, 'timing/data': 0.020650501362979412, 'timing/model': 1.0332045685499907, 'learning_rate/qwen_vl_interface': 8.460382007458081e-06, 'learning_rate/action_model': 8.460382007458081e-05, 'epoch': 0.7}) +[2026-06-17 20:03:14,507][starVLA.training.train_starvla][INFO] - Step 1921, Loss: {'train/loss': 0.8275449275970459, 'train/grad_norm_pre_clip': 14.581169128417969, 'timing/data': 0.019510526210069656, 'timing/model': 1.0350492242723703, 'learning_rate/qwen_vl_interface': 8.45878772647484e-06, 'learning_rate/action_model': 8.458787726474839e-05, 'epoch': 0.7}) +[2026-06-17 20:03:18,862][starVLA.training.train_starvla][INFO] - Step 1922, Loss: {'train/loss': 0.6510826945304871, 'train/grad_norm_pre_clip': 12.214868545532227, 'timing/data': 0.024195761419832706, 'timing/model': 1.0575228612869978, 'learning_rate/qwen_vl_interface': 8.457192780306452e-06, 'learning_rate/action_model': 8.457192780306451e-05, 'epoch': 0.7}) +[2026-06-17 20:03:23,358][starVLA.training.train_starvla][INFO] - Step 1923, Loss: {'train/loss': 0.5821258425712585, 'train/grad_norm_pre_clip': 10.405978202819824, 'timing/data': 0.023899707943201065, 'timing/model': 1.4330722019076347, 'learning_rate/qwen_vl_interface': 8.455597169283549e-06, 'learning_rate/action_model': 8.455597169283548e-05, 'epoch': 0.7}) +[2026-06-17 20:03:27,462][starVLA.training.train_starvla][INFO] - Step 1924, Loss: {'train/loss': 0.5893347859382629, 'train/grad_norm_pre_clip': 11.736028671264648, 'timing/data': 0.020804244093596935, 'timing/model': 1.0298858527094126, 'learning_rate/qwen_vl_interface': 8.454000893736901e-06, 'learning_rate/action_model': 8.4540008937369e-05, 'epoch': 0.7}) +[2026-06-17 20:03:36,536][starVLA.training.train_starvla][INFO] - Step 1925, Loss: {'train/loss': 0.7943456172943115, 'train/grad_norm_pre_clip': 12.813864707946777, 'timing/data': 0.021058346144855022, 'timing/model': 1.1804138207808137, 'learning_rate/qwen_vl_interface': 8.452403953997422e-06, 'learning_rate/action_model': 8.452403953997422e-05, 'epoch': 0.7}) +[2026-06-17 20:03:40,867][starVLA.training.train_starvla][INFO] - Step 1926, Loss: {'train/loss': 0.8641120791435242, 'train/grad_norm_pre_clip': 15.501651763916016, 'timing/data': 0.022021539509296417, 'timing/model': 1.0271355574950576, 'learning_rate/qwen_vl_interface': 8.450806350396158e-06, 'learning_rate/action_model': 8.450806350396158e-05, 'epoch': 0.71}) +[2026-06-17 20:03:45,369][starVLA.training.train_starvla][INFO] - Step 1927, Loss: {'train/loss': 0.5991291403770447, 'train/grad_norm_pre_clip': 12.652729988098145, 'timing/data': 0.03034414816647768, 'timing/model': 1.035786365158856, 'learning_rate/qwen_vl_interface': 8.449208083264291e-06, 'learning_rate/action_model': 8.449208083264292e-05, 'epoch': 0.71}) +[2026-06-17 20:03:50,087][starVLA.training.train_starvla][INFO] - Step 1928, Loss: {'train/loss': 0.5096077919006348, 'train/grad_norm_pre_clip': 9.938386917114258, 'timing/data': 0.022347151301801205, 'timing/model': 1.5172170251607895, 'learning_rate/qwen_vl_interface': 8.447609152933149e-06, 'learning_rate/action_model': 8.447609152933149e-05, 'epoch': 0.71}) +[2026-06-17 20:03:54,678][starVLA.training.train_starvla][INFO] - Step 1929, Loss: {'train/loss': 0.8936886787414551, 'train/grad_norm_pre_clip': 36.441749572753906, 'timing/data': 0.02624228410422802, 'timing/model': 1.0888263266533613, 'learning_rate/qwen_vl_interface': 8.44600955973419e-06, 'learning_rate/action_model': 8.446009559734188e-05, 'epoch': 0.71}) +[2026-06-17 20:04:01,496][starVLA.training.train_starvla][INFO] - Step 1930, Loss: {'train/loss': 0.732783854007721, 'train/grad_norm_pre_clip': 12.6818265914917, 'timing/data': 0.022829938679933548, 'timing/model': 3.4296736493706703, 'learning_rate/qwen_vl_interface': 8.444409303999005e-06, 'learning_rate/action_model': 8.444409303999005e-05, 'epoch': 0.71}) +[2026-06-17 20:04:06,125][starVLA.training.train_starvla][INFO] - Step 1931, Loss: {'train/loss': 0.9110787510871887, 'train/grad_norm_pre_clip': 15.87809944152832, 'timing/data': 0.02113521285355091, 'timing/model': 1.0685507114976645, 'learning_rate/qwen_vl_interface': 8.442808386059337e-06, 'learning_rate/action_model': 8.442808386059337e-05, 'epoch': 0.71}) +[2026-06-17 20:04:10,590][starVLA.training.train_starvla][INFO] - Step 1932, Loss: {'train/loss': 0.9419505596160889, 'train/grad_norm_pre_clip': 13.104711532592773, 'timing/data': 0.0270543172955513, 'timing/model': 1.089956291951239, 'learning_rate/qwen_vl_interface': 8.441206806247055e-06, 'learning_rate/action_model': 8.441206806247054e-05, 'epoch': 0.71}) +[2026-06-17 20:04:15,052][starVLA.training.train_starvla][INFO] - Step 1933, Loss: {'train/loss': 0.8222126960754395, 'train/grad_norm_pre_clip': 17.14810562133789, 'timing/data': 0.019889689050614834, 'timing/model': 1.3870001463219523, 'learning_rate/qwen_vl_interface': 8.439604564894167e-06, 'learning_rate/action_model': 8.439604564894167e-05, 'epoch': 0.71}) +[2026-06-17 20:04:19,333][starVLA.training.train_starvla][INFO] - Step 1934, Loss: {'train/loss': 0.6969242691993713, 'train/grad_norm_pre_clip': 12.100808143615723, 'timing/data': 0.021428702399134636, 'timing/model': 1.0666918950155377, 'learning_rate/qwen_vl_interface': 8.438001662332821e-06, 'learning_rate/action_model': 8.43800166233282e-05, 'epoch': 0.71}) +[2026-06-17 20:04:23,960][starVLA.training.train_starvla][INFO] - Step 1935, Loss: {'train/loss': 0.4852481484413147, 'train/grad_norm_pre_clip': 8.990778923034668, 'timing/data': 0.021536474116146564, 'timing/model': 1.1714888531714678, 'learning_rate/qwen_vl_interface': 8.4363980988953e-06, 'learning_rate/action_model': 8.436398098895301e-05, 'epoch': 0.71}) +[2026-06-17 20:04:33,298][starVLA.training.train_starvla][INFO] - Step 1936, Loss: {'train/loss': 0.8690274357795715, 'train/grad_norm_pre_clip': 16.01963233947754, 'timing/data': 0.020835534669458866, 'timing/model': 4.995318474248052, 'learning_rate/qwen_vl_interface': 8.434793874914026e-06, 'learning_rate/action_model': 8.434793874914026e-05, 'epoch': 0.71}) +[2026-06-17 20:04:37,793][starVLA.training.train_starvla][INFO] - Step 1937, Loss: {'train/loss': 0.6118839383125305, 'train/grad_norm_pre_clip': 11.631457328796387, 'timing/data': 0.025831030681729317, 'timing/model': 1.0601895488798618, 'learning_rate/qwen_vl_interface': 8.433188990721553e-06, 'learning_rate/action_model': 8.433188990721552e-05, 'epoch': 0.71}) +[2026-06-17 20:04:42,345][starVLA.training.train_starvla][INFO] - Step 1938, Loss: {'train/loss': 0.8720851540565491, 'train/grad_norm_pre_clip': 12.229860305786133, 'timing/data': 0.021782688796520233, 'timing/model': 1.300947790965438, 'learning_rate/qwen_vl_interface': 8.431583446650577e-06, 'learning_rate/action_model': 8.431583446650578e-05, 'epoch': 0.71}) +[2026-06-17 20:04:46,559][starVLA.training.train_starvla][INFO] - Step 1939, Loss: {'train/loss': 0.6862410306930542, 'train/grad_norm_pre_clip': 11.207082748413086, 'timing/data': 0.022650957107543945, 'timing/model': 1.1543761771172285, 'learning_rate/qwen_vl_interface': 8.429977243033928e-06, 'learning_rate/action_model': 8.429977243033928e-05, 'epoch': 0.71}) +[2026-06-17 20:04:51,097][starVLA.training.train_starvla][INFO] - Step 1940, Loss: {'train/loss': 0.8903945088386536, 'train/grad_norm_pre_clip': 14.634628295898438, 'timing/data': 0.02155798301100731, 'timing/model': 1.0838038315996528, 'learning_rate/qwen_vl_interface': 8.428370380204577e-06, 'learning_rate/action_model': 8.428370380204577e-05, 'epoch': 0.71}) +[2026-06-17 20:04:58,944][starVLA.training.train_starvla][INFO] - Step 1941, Loss: {'train/loss': 0.9708185195922852, 'train/grad_norm_pre_clip': 12.665163040161133, 'timing/data': 0.022770028561353683, 'timing/model': 1.084400575608015, 'learning_rate/qwen_vl_interface': 8.426762858495625e-06, 'learning_rate/action_model': 8.426762858495625e-05, 'epoch': 0.71}) +[2026-06-17 20:05:03,418][starVLA.training.train_starvla][INFO] - Step 1942, Loss: {'train/loss': 0.7124214172363281, 'train/grad_norm_pre_clip': 20.21217155456543, 'timing/data': 0.026280218735337257, 'timing/model': 1.1027531791478395, 'learning_rate/qwen_vl_interface': 8.425154678240315e-06, 'learning_rate/action_model': 8.425154678240314e-05, 'epoch': 0.71}) +[2026-06-17 20:05:08,083][starVLA.training.train_starvla][INFO] - Step 1943, Loss: {'train/loss': 0.7683680653572083, 'train/grad_norm_pre_clip': 11.9185209274292, 'timing/data': 0.020481380634009838, 'timing/model': 1.5068361405283213, 'learning_rate/qwen_vl_interface': 8.423545839772021e-06, 'learning_rate/action_model': 8.423545839772021e-05, 'epoch': 0.71}) +[2026-06-17 20:05:12,261][starVLA.training.train_starvla][INFO] - Step 1944, Loss: {'train/loss': 0.46059471368789673, 'train/grad_norm_pre_clip': 11.163606643676758, 'timing/data': 0.020301404409110546, 'timing/model': 1.0350979510694742, 'learning_rate/qwen_vl_interface': 8.421936343424264e-06, 'learning_rate/action_model': 8.421936343424262e-05, 'epoch': 0.71}) +[2026-06-17 20:05:17,604][starVLA.training.train_starvla][INFO] - Step 1945, Loss: {'train/loss': 0.5255900621414185, 'train/grad_norm_pre_clip': 10.378576278686523, 'timing/data': 0.02172889281064272, 'timing/model': 1.0254472950473428, 'learning_rate/qwen_vl_interface': 8.420326189530685e-06, 'learning_rate/action_model': 8.420326189530685e-05, 'epoch': 0.71}) +[2026-06-17 20:05:22,233][starVLA.training.train_starvla][INFO] - Step 1946, Loss: {'train/loss': 1.030925989151001, 'train/grad_norm_pre_clip': 15.076464653015137, 'timing/data': 0.02201798651367426, 'timing/model': 1.0342626357451081, 'learning_rate/qwen_vl_interface': 8.418715378425078e-06, 'learning_rate/action_model': 8.418715378425078e-05, 'epoch': 0.71}) +[2026-06-17 20:05:26,679][starVLA.training.train_starvla][INFO] - Step 1947, Loss: {'train/loss': 0.697109043598175, 'train/grad_norm_pre_clip': 14.38564682006836, 'timing/data': 0.021359321661293507, 'timing/model': 1.1032616896554828, 'learning_rate/qwen_vl_interface': 8.417103910441364e-06, 'learning_rate/action_model': 8.417103910441363e-05, 'epoch': 0.71}) +[2026-06-17 20:05:31,173][starVLA.training.train_starvla][INFO] - Step 1948, Loss: {'train/loss': 0.7673987150192261, 'train/grad_norm_pre_clip': 11.436407089233398, 'timing/data': 0.023760799318552017, 'timing/model': 1.4675195384770632, 'learning_rate/qwen_vl_interface': 8.415491785913599e-06, 'learning_rate/action_model': 8.4154917859136e-05, 'epoch': 0.71}) +[2026-06-17 20:05:36,731][starVLA.training.train_starvla][INFO] - Step 1949, Loss: {'train/loss': 0.8456889390945435, 'train/grad_norm_pre_clip': 13.885916709899902, 'timing/data': 0.021567387506365776, 'timing/model': 2.431510880589485, 'learning_rate/qwen_vl_interface': 8.413879005175983e-06, 'learning_rate/action_model': 8.413879005175983e-05, 'epoch': 0.71}) +[2026-06-17 20:05:41,902][starVLA.training.train_starvla][INFO] - Step 1950, Loss: {'train/loss': 0.8629153966903687, 'train/grad_norm_pre_clip': 14.410090446472168, 'timing/data': 0.026350650005042553, 'timing/model': 1.0474232900887728, 'learning_rate/qwen_vl_interface': 8.412265568562843e-06, 'learning_rate/action_model': 8.412265568562842e-05, 'epoch': 0.71}) +[2026-06-17 20:05:46,445][starVLA.training.train_starvla][INFO] - Step 1951, Loss: {'train/loss': 1.1464173793792725, 'train/grad_norm_pre_clip': 15.489956855773926, 'timing/data': 0.022409985773265362, 'timing/model': 1.0800308119505644, 'learning_rate/qwen_vl_interface': 8.41065147640865e-06, 'learning_rate/action_model': 8.41065147640865e-05, 'epoch': 0.71}) +[2026-06-17 20:05:50,884][starVLA.training.train_starvla][INFO] - Step 1952, Loss: {'train/loss': 0.7383943796157837, 'train/grad_norm_pre_clip': 15.251667976379395, 'timing/data': 0.02082399744540453, 'timing/model': 1.0969970421865582, 'learning_rate/qwen_vl_interface': 8.409036729048002e-06, 'learning_rate/action_model': 8.409036729048002e-05, 'epoch': 0.71}) +[2026-06-17 20:05:56,621][starVLA.training.train_starvla][INFO] - Step 1953, Loss: {'train/loss': 0.7446579933166504, 'train/grad_norm_pre_clip': 12.943598747253418, 'timing/data': 0.021768076345324516, 'timing/model': 1.4880927428603172, 'learning_rate/qwen_vl_interface': 8.407421326815644e-06, 'learning_rate/action_model': 8.407421326815644e-05, 'epoch': 0.72}) +[2026-06-17 20:06:06,111][starVLA.training.train_starvla][INFO] - Step 1954, Loss: {'train/loss': 0.7709665894508362, 'train/grad_norm_pre_clip': 14.438053131103516, 'timing/data': 0.020953300409018993, 'timing/model': 1.0342591255903244, 'learning_rate/qwen_vl_interface': 8.405805270046445e-06, 'learning_rate/action_model': 8.405805270046445e-05, 'epoch': 0.72}) +[2026-06-17 20:06:10,863][starVLA.training.train_starvla][INFO] - Step 1955, Loss: {'train/loss': 1.0585484504699707, 'train/grad_norm_pre_clip': 19.684938430786133, 'timing/data': 0.02010220568627119, 'timing/model': 1.0831891587004066, 'learning_rate/qwen_vl_interface': 8.404188559075419e-06, 'learning_rate/action_model': 8.404188559075419e-05, 'epoch': 0.72}) +[2026-06-17 20:06:15,412][starVLA.training.train_starvla][INFO] - Step 1956, Loss: {'train/loss': 0.7189643979072571, 'train/grad_norm_pre_clip': 13.282105445861816, 'timing/data': 0.02081870660185814, 'timing/model': 1.025628473609686, 'learning_rate/qwen_vl_interface': 8.40257119423771e-06, 'learning_rate/action_model': 8.402571194237709e-05, 'epoch': 0.72}) +[2026-06-17 20:06:19,872][starVLA.training.train_starvla][INFO] - Step 1957, Loss: {'train/loss': 0.5680103302001953, 'train/grad_norm_pre_clip': 10.503754615783691, 'timing/data': 0.023966746404767036, 'timing/model': 1.0355264861136675, 'learning_rate/qwen_vl_interface': 8.4009531758686e-06, 'learning_rate/action_model': 8.4009531758686e-05, 'epoch': 0.72}) +[2026-06-17 20:06:24,201][starVLA.training.train_starvla][INFO] - Step 1958, Loss: {'train/loss': 0.8977997899055481, 'train/grad_norm_pre_clip': 13.75369930267334, 'timing/data': 0.019980428740382195, 'timing/model': 1.370099044404924, 'learning_rate/qwen_vl_interface': 8.399334504303507e-06, 'learning_rate/action_model': 8.399334504303507e-05, 'epoch': 0.72}) +[2026-06-17 20:06:28,427][starVLA.training.train_starvla][INFO] - Step 1959, Loss: {'train/loss': 0.500095009803772, 'train/grad_norm_pre_clip': 9.420393943786621, 'timing/data': 0.030480321496725082, 'timing/model': 1.089597376063466, 'learning_rate/qwen_vl_interface': 8.397715179877982e-06, 'learning_rate/action_model': 8.397715179877982e-05, 'epoch': 0.72}) +[2026-06-17 20:06:32,949][starVLA.training.train_starvla][INFO] - Step 1960, Loss: {'train/loss': 1.0218157768249512, 'train/grad_norm_pre_clip': 17.005605697631836, 'timing/data': 0.021983572281897068, 'timing/model': 1.0803494052961469, 'learning_rate/qwen_vl_interface': 8.396095202927712e-06, 'learning_rate/action_model': 8.396095202927712e-05, 'epoch': 0.72}) +[2026-06-17 20:06:37,560][starVLA.training.train_starvla][INFO] - Step 1961, Loss: {'train/loss': 0.5397263169288635, 'train/grad_norm_pre_clip': 9.74244213104248, 'timing/data': 0.02644984796643257, 'timing/model': 1.1513328161090612, 'learning_rate/qwen_vl_interface': 8.394474573788522e-06, 'learning_rate/action_model': 8.394474573788522e-05, 'epoch': 0.72}) +[2026-06-17 20:06:42,062][starVLA.training.train_starvla][INFO] - Step 1962, Loss: {'train/loss': 0.4292093813419342, 'train/grad_norm_pre_clip': 10.436805725097656, 'timing/data': 0.020196842029690742, 'timing/model': 1.0715447682887316, 'learning_rate/qwen_vl_interface': 8.39285329279637e-06, 'learning_rate/action_model': 8.39285329279637e-05, 'epoch': 0.72}) +[2026-06-17 20:06:46,712][starVLA.training.train_starvla][INFO] - Step 1963, Loss: {'train/loss': 0.5075064301490784, 'train/grad_norm_pre_clip': 12.311514854431152, 'timing/data': 0.025234716944396496, 'timing/model': 1.4290234483778477, 'learning_rate/qwen_vl_interface': 8.391231360287348e-06, 'learning_rate/action_model': 8.391231360287347e-05, 'epoch': 0.72}) +[2026-06-17 20:06:50,770][starVLA.training.train_starvla][INFO] - Step 1964, Loss: {'train/loss': 0.6171386241912842, 'train/grad_norm_pre_clip': 11.466254234313965, 'timing/data': 0.022126332856714725, 'timing/model': 1.0408486239612103, 'learning_rate/qwen_vl_interface': 8.389608776597684e-06, 'learning_rate/action_model': 8.389608776597684e-05, 'epoch': 0.72}) +[2026-06-17 20:06:55,206][starVLA.training.train_starvla][INFO] - Step 1965, Loss: {'train/loss': 0.9861413240432739, 'train/grad_norm_pre_clip': 15.881061553955078, 'timing/data': 0.020057937130331993, 'timing/model': 1.0413066875189543, 'learning_rate/qwen_vl_interface': 8.387985542063743e-06, 'learning_rate/action_model': 8.387985542063743e-05, 'epoch': 0.72}) +[2026-06-17 20:06:59,732][starVLA.training.train_starvla][INFO] - Step 1966, Loss: {'train/loss': 0.8446695804595947, 'train/grad_norm_pre_clip': 14.902097702026367, 'timing/data': 0.020747880451381207, 'timing/model': 1.0906746182590723, 'learning_rate/qwen_vl_interface': 8.386361657022024e-06, 'learning_rate/action_model': 8.386361657022023e-05, 'epoch': 0.72}) +[2026-06-17 20:07:04,412][starVLA.training.train_starvla][INFO] - Step 1967, Loss: {'train/loss': 0.5457751750946045, 'train/grad_norm_pre_clip': 26.063217163085938, 'timing/data': 0.022941204719245434, 'timing/model': 1.041223581880331, 'learning_rate/qwen_vl_interface': 8.384737121809156e-06, 'learning_rate/action_model': 8.384737121809156e-05, 'epoch': 0.72}) +[2026-06-17 20:07:14,607][starVLA.training.train_starvla][INFO] - Step 1968, Loss: {'train/loss': 1.1641243696212769, 'train/grad_norm_pre_clip': 17.26093864440918, 'timing/data': 0.022416451014578342, 'timing/model': 2.8446073811501265, 'learning_rate/qwen_vl_interface': 8.38311193676191e-06, 'learning_rate/action_model': 8.383111936761909e-05, 'epoch': 0.72}) +[2026-06-17 20:07:18,719][starVLA.training.train_starvla][INFO] - Step 1969, Loss: {'train/loss': 0.8474987745285034, 'train/grad_norm_pre_clip': 21.749629974365234, 'timing/data': 0.020506116561591625, 'timing/model': 1.089329369366169, 'learning_rate/qwen_vl_interface': 8.381486102217188e-06, 'learning_rate/action_model': 8.381486102217188e-05, 'epoch': 0.72}) +[2026-06-17 20:07:23,180][starVLA.training.train_starvla][INFO] - Step 1970, Loss: {'train/loss': 0.27647268772125244, 'train/grad_norm_pre_clip': 10.312933921813965, 'timing/data': 0.0204001534730196, 'timing/model': 1.0210351524874568, 'learning_rate/qwen_vl_interface': 8.379859618512029e-06, 'learning_rate/action_model': 8.379859618512029e-05, 'epoch': 0.72}) +[2026-06-17 20:07:27,832][starVLA.training.train_starvla][INFO] - Step 1971, Loss: {'train/loss': 0.5903212428092957, 'train/grad_norm_pre_clip': 15.880270004272461, 'timing/data': 0.023006025701761246, 'timing/model': 1.0489799864590168, 'learning_rate/qwen_vl_interface': 8.378232485983603e-06, 'learning_rate/action_model': 8.378232485983603e-05, 'epoch': 0.72}) +[2026-06-17 20:07:32,269][starVLA.training.train_starvla][INFO] - Step 1972, Loss: {'train/loss': 0.7870069146156311, 'train/grad_norm_pre_clip': 14.944771766662598, 'timing/data': 0.021501660346984863, 'timing/model': 1.0939522050321102, 'learning_rate/qwen_vl_interface': 8.376604704969216e-06, 'learning_rate/action_model': 8.376604704969216e-05, 'epoch': 0.72}) +[2026-06-17 20:07:37,006][starVLA.training.train_starvla][INFO] - Step 1973, Loss: {'train/loss': 0.8519002795219421, 'train/grad_norm_pre_clip': 14.420845031738281, 'timing/data': 0.02660108357667923, 'timing/model': 1.6601224150508642, 'learning_rate/qwen_vl_interface': 8.374976275806307e-06, 'learning_rate/action_model': 8.374976275806308e-05, 'epoch': 0.72}) +[2026-06-17 20:07:41,347][starVLA.training.train_starvla][INFO] - Step 1974, Loss: {'train/loss': 0.5959876179695129, 'train/grad_norm_pre_clip': 10.867406845092773, 'timing/data': 0.023853253573179245, 'timing/model': 1.0340441977605224, 'learning_rate/qwen_vl_interface': 8.373347198832457e-06, 'learning_rate/action_model': 8.373347198832457e-05, 'epoch': 0.72}) +[2026-06-17 20:07:46,102][starVLA.training.train_starvla][INFO] - Step 1975, Loss: {'train/loss': 0.9577754139900208, 'train/grad_norm_pre_clip': 14.912466049194336, 'timing/data': 0.033950481563806534, 'timing/model': 1.1906195571646094, 'learning_rate/qwen_vl_interface': 8.37171747438537e-06, 'learning_rate/action_model': 8.37171747438537e-05, 'epoch': 0.72}) +[2026-06-17 20:07:50,651][starVLA.training.train_starvla][INFO] - Step 1976, Loss: {'train/loss': 0.7074434757232666, 'train/grad_norm_pre_clip': 22.48126792907715, 'timing/data': 0.021037500351667404, 'timing/model': 1.048862244002521, 'learning_rate/qwen_vl_interface': 8.370087102802892e-06, 'learning_rate/action_model': 8.370087102802892e-05, 'epoch': 0.72}) +[2026-06-17 20:07:55,019][starVLA.training.train_starvla][INFO] - Step 1977, Loss: {'train/loss': 0.9145601391792297, 'train/grad_norm_pre_clip': 13.86445140838623, 'timing/data': 0.020904215052723885, 'timing/model': 1.0753394486382604, 'learning_rate/qwen_vl_interface': 8.368456084423e-06, 'learning_rate/action_model': 8.368456084422999e-05, 'epoch': 0.72}) +[2026-06-17 20:07:59,511][starVLA.training.train_starvla][INFO] - Step 1978, Loss: {'train/loss': 0.7183972001075745, 'train/grad_norm_pre_clip': 15.594886779785156, 'timing/data': 0.021523592062294483, 'timing/model': 1.3638706374913454, 'learning_rate/qwen_vl_interface': 8.366824419583806e-06, 'learning_rate/action_model': 8.366824419583806e-05, 'epoch': 0.72}) +[2026-06-17 20:08:03,616][starVLA.training.train_starvla][INFO] - Step 1979, Loss: {'train/loss': 0.6685919165611267, 'train/grad_norm_pre_clip': 12.880249977111816, 'timing/data': 0.022207139991223812, 'timing/model': 1.0674493815749884, 'learning_rate/qwen_vl_interface': 8.365192108623555e-06, 'learning_rate/action_model': 8.365192108623555e-05, 'epoch': 0.72}) +[2026-06-17 20:08:08,133][starVLA.training.train_starvla][INFO] - Step 1980, Loss: {'train/loss': 0.7066010236740112, 'train/grad_norm_pre_clip': 12.56011962890625, 'timing/data': 0.02470429614186287, 'timing/model': 1.0910319425165653, 'learning_rate/qwen_vl_interface': 8.363559151880631e-06, 'learning_rate/action_model': 8.36355915188063e-05, 'epoch': 0.73}) +[2026-06-17 20:08:12,840][starVLA.training.train_starvla][INFO] - Step 1981, Loss: {'train/loss': 0.5857687592506409, 'train/grad_norm_pre_clip': 10.537483215332031, 'timing/data': 0.022750855423510075, 'timing/model': 1.1222595302388072, 'learning_rate/qwen_vl_interface': 8.361925549693541e-06, 'learning_rate/action_model': 8.36192554969354e-05, 'epoch': 0.73}) +[2026-06-17 20:08:17,441][starVLA.training.train_starvla][INFO] - Step 1982, Loss: {'train/loss': 0.6944474577903748, 'train/grad_norm_pre_clip': 11.040732383728027, 'timing/data': 0.023757951334118843, 'timing/model': 1.0261893589049578, 'learning_rate/qwen_vl_interface': 8.360291302400937e-06, 'learning_rate/action_model': 8.360291302400937e-05, 'epoch': 0.73}) +[2026-06-17 20:08:21,878][starVLA.training.train_starvla][INFO] - Step 1983, Loss: {'train/loss': 0.8655036687850952, 'train/grad_norm_pre_clip': 15.212702751159668, 'timing/data': 0.021000156179070473, 'timing/model': 1.3181853387504816, 'learning_rate/qwen_vl_interface': 8.3586564103416e-06, 'learning_rate/action_model': 8.358656410341599e-05, 'epoch': 0.73}) +[2026-06-17 20:08:26,054][starVLA.training.train_starvla][INFO] - Step 1984, Loss: {'train/loss': 0.7046453356742859, 'train/grad_norm_pre_clip': 12.43289566040039, 'timing/data': 0.021810345351696014, 'timing/model': 1.0883081499487162, 'learning_rate/qwen_vl_interface': 8.357020873854445e-06, 'learning_rate/action_model': 8.357020873854444e-05, 'epoch': 0.73}) +[2026-06-17 20:08:30,489][starVLA.training.train_starvla][INFO] - Step 1985, Loss: {'train/loss': 0.7607046961784363, 'train/grad_norm_pre_clip': 13.03911018371582, 'timing/data': 0.02057298645377159, 'timing/model': 1.0375538766384125, 'learning_rate/qwen_vl_interface': 8.355384693278516e-06, 'learning_rate/action_model': 8.355384693278517e-05, 'epoch': 0.73}) +[2026-06-17 20:08:35,098][starVLA.training.train_starvla][INFO] - Step 1986, Loss: {'train/loss': 0.6440778970718384, 'train/grad_norm_pre_clip': 12.626002311706543, 'timing/data': 0.02454548329114914, 'timing/model': 1.1209122287109494, 'learning_rate/qwen_vl_interface': 8.353747868953002e-06, 'learning_rate/action_model': 8.353747868953003e-05, 'epoch': 0.73}) +[2026-06-17 20:08:39,523][starVLA.training.train_starvla][INFO] - Step 1987, Loss: {'train/loss': 0.8711785674095154, 'train/grad_norm_pre_clip': 15.299238204956055, 'timing/data': 0.023488120175898075, 'timing/model': 1.0934866163879633, 'learning_rate/qwen_vl_interface': 8.352110401217213e-06, 'learning_rate/action_model': 8.352110401217213e-05, 'epoch': 0.73}) +[2026-06-17 20:08:44,122][starVLA.training.train_starvla][INFO] - Step 1988, Loss: {'train/loss': 0.6569919586181641, 'train/grad_norm_pre_clip': 11.072473526000977, 'timing/data': 0.022794971242547035, 'timing/model': 1.4421462109312415, 'learning_rate/qwen_vl_interface': 8.350472290410602e-06, 'learning_rate/action_model': 8.350472290410602e-05, 'epoch': 0.73}) +[2026-06-17 20:08:48,206][starVLA.training.train_starvla][INFO] - Step 1989, Loss: {'train/loss': 0.5140093564987183, 'train/grad_norm_pre_clip': 10.482754707336426, 'timing/data': 0.030375681817531586, 'timing/model': 1.0526079451665282, 'learning_rate/qwen_vl_interface': 8.348833536872747e-06, 'learning_rate/action_model': 8.348833536872747e-05, 'epoch': 0.73}) +[2026-06-17 20:08:58,800][starVLA.training.train_starvla][INFO] - Step 1990, Loss: {'train/loss': 0.6422566175460815, 'train/grad_norm_pre_clip': 11.820321083068848, 'timing/data': 0.024714913219213486, 'timing/model': 4.2488998509943485, 'learning_rate/qwen_vl_interface': 8.347194140943366e-06, 'learning_rate/action_model': 8.347194140943366e-05, 'epoch': 0.73}) +[2026-06-17 20:09:12,084][starVLA.training.train_starvla][INFO] - Step 1991, Loss: {'train/loss': 0.5990908741950989, 'train/grad_norm_pre_clip': 10.677960395812988, 'timing/data': 0.0206596152856946, 'timing/model': 1.2147002425044775, 'learning_rate/qwen_vl_interface': 8.345554102962306e-06, 'learning_rate/action_model': 8.345554102962305e-05, 'epoch': 0.73}) +[2026-06-17 20:09:16,791][starVLA.training.train_starvla][INFO] - Step 1992, Loss: {'train/loss': 0.8955692052841187, 'train/grad_norm_pre_clip': 14.514382362365723, 'timing/data': 0.02119291014969349, 'timing/model': 1.0349110569804907, 'learning_rate/qwen_vl_interface': 8.34391342326955e-06, 'learning_rate/action_model': 8.343913423269549e-05, 'epoch': 0.73}) +[2026-06-17 20:09:21,302][starVLA.training.train_starvla][INFO] - Step 1993, Loss: {'train/loss': 0.8216884136199951, 'train/grad_norm_pre_clip': 12.245298385620117, 'timing/data': 0.022089666686952114, 'timing/model': 1.3715732488781214, 'learning_rate/qwen_vl_interface': 8.342272102205213e-06, 'learning_rate/action_model': 8.342272102205212e-05, 'epoch': 0.73}) +[2026-06-17 20:09:25,616][starVLA.training.train_starvla][INFO] - Step 1994, Loss: {'train/loss': 0.6437509655952454, 'train/grad_norm_pre_clip': 10.857220649719238, 'timing/data': 0.022947381250560284, 'timing/model': 1.137586997821927, 'learning_rate/qwen_vl_interface': 8.340630140109539e-06, 'learning_rate/action_model': 8.340630140109539e-05, 'epoch': 0.73}) +[2026-06-17 20:09:31,292][starVLA.training.train_starvla][INFO] - Step 1995, Loss: {'train/loss': 0.6990893483161926, 'train/grad_norm_pre_clip': 15.00549030303955, 'timing/data': 0.025549033656716347, 'timing/model': 2.2457213643938303, 'learning_rate/qwen_vl_interface': 8.338987537322916e-06, 'learning_rate/action_model': 8.338987537322916e-05, 'epoch': 0.73}) +[2026-06-17 20:09:35,968][starVLA.training.train_starvla][INFO] - Step 1996, Loss: {'train/loss': 0.547336220741272, 'train/grad_norm_pre_clip': 10.282794952392578, 'timing/data': 0.020101728849112988, 'timing/model': 1.1105225458741188, 'learning_rate/qwen_vl_interface': 8.337344294185848e-06, 'learning_rate/action_model': 8.337344294185849e-05, 'epoch': 0.73}) +[2026-06-17 20:09:41,490][starVLA.training.train_starvla][INFO] - Step 1997, Loss: {'train/loss': 0.7578168511390686, 'train/grad_norm_pre_clip': 12.738399505615234, 'timing/data': 0.03220310993492603, 'timing/model': 1.0971603896468878, 'learning_rate/qwen_vl_interface': 8.335700411038988e-06, 'learning_rate/action_model': 8.335700411038988e-05, 'epoch': 0.73}) +[2026-06-17 20:09:45,908][starVLA.training.train_starvla][INFO] - Step 1998, Loss: {'train/loss': 0.8872625231742859, 'train/grad_norm_pre_clip': 14.817364692687988, 'timing/data': 0.028249982744455338, 'timing/model': 1.3487829770892859, 'learning_rate/qwen_vl_interface': 8.334055888223111e-06, 'learning_rate/action_model': 8.334055888223111e-05, 'epoch': 0.73}) +[2026-06-17 20:09:50,146][starVLA.training.train_starvla][INFO] - Step 1999, Loss: {'train/loss': 0.7109673619270325, 'train/grad_norm_pre_clip': 12.950043678283691, 'timing/data': 0.011840379796922207, 'timing/model': 1.033849318511784, 'learning_rate/qwen_vl_interface': 8.332410726079132e-06, 'learning_rate/action_model': 8.332410726079132e-05, 'epoch': 0.73}) +[2026-06-17 20:13:17,777][starVLA.training.train_starvla][INFO] - Step 2000, Loss: {'train/loss': 1.0162255764007568, 'train/grad_norm_pre_clip': 14.766885757446289, 'eval/loss': 0.774057149887085, 'eval/latency_2/loss': 0.774057149887085, 'eval/demon_fire_precision': 0.8947156111335216, 'eval/demon_fire_recall': 0.9280334728033472, 'eval/demon_fire_f1': 0.9110700349147668, 'eval/demon_fire_rate_model': 0.6897607122982748, 'eval/demon_fire_rate_teacher': 0.6649972175848636, 'eval/demon_left_precision': 0.8146186440677966, 'eval/demon_left_recall': 0.8094736842105263, 'eval/demon_left_f1': 0.812038014783527, 'eval/demon_left_rate_model': 0.2626599888703395, 'eval/demon_left_rate_teacher': 0.2643294379521425, 'eval/demon_right_precision': 0.740234375, 'eval/demon_right_recall': 0.8564971751412429, 'eval/demon_right_f1': 0.7941330539549502, 'eval/demon_right_rate_model': 0.2849193099610462, 'eval/demon_right_rate_teacher': 0.24624373956594323, 'eval/demon_move_f1': 0.8030855343692386, 'eval/demon_cc_f1': 0.8757437535466642, 'eval/latency_2/demon_fire_precision': 0.8947156111335216, 'eval/latency_2/demon_fire_recall': 0.9280334728033472, 'eval/latency_2/demon_fire_f1': 0.9110700349147668, 'eval/latency_2/demon_fire_rate_model': 0.6897607122982748, 'eval/latency_2/demon_fire_rate_teacher': 0.6649972175848636, 'eval/latency_2/demon_left_precision': 0.8146186440677966, 'eval/latency_2/demon_left_recall': 0.8094736842105263, 'eval/latency_2/demon_left_f1': 0.812038014783527, 'eval/latency_2/demon_left_rate_model': 0.2626599888703395, 'eval/latency_2/demon_left_rate_teacher': 0.2643294379521425, 'eval/latency_2/demon_right_precision': 0.740234375, 'eval/latency_2/demon_right_recall': 0.8564971751412429, 'eval/latency_2/demon_right_f1': 0.7941330539549502, 'eval/latency_2/demon_right_rate_model': 0.2849193099610462, 'eval/latency_2/demon_right_rate_teacher': 0.24624373956594323, 'eval/latency_2/demon_move_f1': 0.8030855343692386, 'eval/latency_2/demon_cc_f1': 0.8757437535466642, 'timing/data': 0.01214901078492403, 'timing/model': 1.0682656103745103, 'learning_rate/qwen_vl_interface': 8.330764924948093e-06, 'learning_rate/action_model': 8.330764924948093e-05, 'epoch': 0.73}) +[2026-06-17 20:13:17,804][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_2000_state +[2026-06-17 20:13:17,812][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 20:13:27,529][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_2000_state/model.safetensors +[2026-06-17 20:13:42,175][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_2000_state/optimizer.bin +[2026-06-17 20:13:42,176][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_2000_state/sampler.bin +[2026-06-17 20:13:42,176][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_2000_state/sampler_1.bin +[2026-06-17 20:13:42,178][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_2000_state/random_states_0.pkl +[2026-06-17 20:13:42,183][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2000 +[2026-06-17 20:13:42,183][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 20:13:42,200][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 20:13:46,764][starVLA.training.train_starvla][INFO] - Step 2001, Loss: {'train/loss': 1.1987181901931763, 'train/grad_norm_pre_clip': 24.20020866394043, 'timing/data': 0.017186111770570278, 'timing/model': 1.0217161681503057, 'learning_rate/qwen_vl_interface': 8.329118485171167e-06, 'learning_rate/action_model': 8.329118485171167e-05, 'epoch': 0.73}) +[2026-06-17 20:13:51,218][starVLA.training.train_starvla][INFO] - Step 2002, Loss: {'train/loss': 0.7095680832862854, 'train/grad_norm_pre_clip': 14.250768661499023, 'timing/data': 0.012583710253238678, 'timing/model': 1.0844258554279804, 'learning_rate/qwen_vl_interface': 8.327471407089667e-06, 'learning_rate/action_model': 8.327471407089668e-05, 'epoch': 0.73}) +[2026-06-17 20:13:55,585][starVLA.training.train_starvla][INFO] - Step 2003, Loss: {'train/loss': 0.530916690826416, 'train/grad_norm_pre_clip': 11.033941268920898, 'timing/data': 0.011110512539744377, 'timing/model': 1.0391243314370513, 'learning_rate/qwen_vl_interface': 8.325823691045032e-06, 'learning_rate/action_model': 8.325823691045033e-05, 'epoch': 0.73}) +[2026-06-17 20:13:59,895][starVLA.training.train_starvla][INFO] - Step 2004, Loss: {'train/loss': 0.8247007131576538, 'train/grad_norm_pre_clip': 15.117152214050293, 'timing/data': 0.011757253669202328, 'timing/model': 1.2672393219545484, 'learning_rate/qwen_vl_interface': 8.32417533737884e-06, 'learning_rate/action_model': 8.324175337378839e-05, 'epoch': 0.73}) +[2026-06-17 20:14:04,030][starVLA.training.train_starvla][INFO] - Step 2005, Loss: {'train/loss': 0.9104347229003906, 'train/grad_norm_pre_clip': 18.52704620361328, 'timing/data': 0.011743699200451374, 'timing/model': 1.0739647056907415, 'learning_rate/qwen_vl_interface': 8.322526346432789e-06, 'learning_rate/action_model': 8.322526346432789e-05, 'epoch': 0.73}) +[2026-06-17 20:14:08,519][starVLA.training.train_starvla][INFO] - Step 2006, Loss: {'train/loss': 0.7053831219673157, 'train/grad_norm_pre_clip': 16.570545196533203, 'timing/data': 0.01265527680516243, 'timing/model': 1.138188710436225, 'learning_rate/qwen_vl_interface': 8.32087671854872e-06, 'learning_rate/action_model': 8.32087671854872e-05, 'epoch': 0.73}) +[2026-06-17 20:14:12,982][starVLA.training.train_starvla][INFO] - Step 2007, Loss: {'train/loss': 0.6194016933441162, 'train/grad_norm_pre_clip': 12.165139198303223, 'timing/data': 0.010614325292408466, 'timing/model': 1.1050051907077432, 'learning_rate/qwen_vl_interface': 8.319226454068604e-06, 'learning_rate/action_model': 8.319226454068604e-05, 'epoch': 0.73}) +[2026-06-17 20:14:17,428][starVLA.training.train_starvla][INFO] - Step 2008, Loss: {'train/loss': 0.7628931999206543, 'train/grad_norm_pre_clip': 13.287644386291504, 'timing/data': 0.015842432156205177, 'timing/model': 1.214746038429439, 'learning_rate/qwen_vl_interface': 8.317575553334542e-06, 'learning_rate/action_model': 8.317575553334542e-05, 'epoch': 0.74}) +[2026-06-17 20:14:21,942][starVLA.training.train_starvla][INFO] - Step 2009, Loss: {'train/loss': 0.927538275718689, 'train/grad_norm_pre_clip': 16.90860939025879, 'timing/data': 0.011951982975006104, 'timing/model': 1.348355207592249, 'learning_rate/qwen_vl_interface': 8.315924016688767e-06, 'learning_rate/action_model': 8.315924016688766e-05, 'epoch': 0.74}) +[2026-06-17 20:14:25,939][starVLA.training.train_starvla][INFO] - Step 2010, Loss: {'train/loss': 0.6694533824920654, 'train/grad_norm_pre_clip': 9.850193977355957, 'timing/data': 0.01167901512235403, 'timing/model': 1.0243360167369246, 'learning_rate/qwen_vl_interface': 8.314271844473644e-06, 'learning_rate/action_model': 8.314271844473644e-05, 'epoch': 0.74}) +[2026-06-17 20:14:30,455][starVLA.training.train_starvla][INFO] - Step 2011, Loss: {'train/loss': 0.5361205339431763, 'train/grad_norm_pre_clip': 61.75752639770508, 'timing/data': 0.010713892988860607, 'timing/model': 1.0221167048439384, 'learning_rate/qwen_vl_interface': 8.312619037031668e-06, 'learning_rate/action_model': 8.312619037031668e-05, 'epoch': 0.74}) +[2026-06-17 20:14:34,802][starVLA.training.train_starvla][INFO] - Step 2012, Loss: {'train/loss': 0.764539897441864, 'train/grad_norm_pre_clip': 13.772741317749023, 'timing/data': 0.011836390011012554, 'timing/model': 1.0891225161030889, 'learning_rate/qwen_vl_interface': 8.310965594705473e-06, 'learning_rate/action_model': 8.310965594705472e-05, 'epoch': 0.74}) +[2026-06-17 20:14:39,357][starVLA.training.train_starvla][INFO] - Step 2013, Loss: {'train/loss': 0.6357806324958801, 'train/grad_norm_pre_clip': 11.519989013671875, 'timing/data': 0.014522219076752663, 'timing/model': 1.0913380337879062, 'learning_rate/qwen_vl_interface': 8.309311517837814e-06, 'learning_rate/action_model': 8.309311517837813e-05, 'epoch': 0.74}) +[2026-06-17 20:14:43,501][starVLA.training.train_starvla][INFO] - Step 2014, Loss: {'train/loss': 1.1211535930633545, 'train/grad_norm_pre_clip': 31.165576934814453, 'timing/data': 0.010679495520889759, 'timing/model': 1.0504994997754693, 'learning_rate/qwen_vl_interface': 8.307656806771587e-06, 'learning_rate/action_model': 8.307656806771588e-05, 'epoch': 0.74}) +[2026-06-17 20:14:48,671][starVLA.training.train_starvla][INFO] - Step 2015, Loss: {'train/loss': 0.6641516089439392, 'train/grad_norm_pre_clip': 11.737244606018066, 'timing/data': 0.012001441791653633, 'timing/model': 1.1501752929762006, 'learning_rate/qwen_vl_interface': 8.306001461849814e-06, 'learning_rate/action_model': 8.306001461849814e-05, 'epoch': 0.74}) +[2026-06-17 20:14:53,226][starVLA.training.train_starvla][INFO] - Step 2016, Loss: {'train/loss': 0.6146767735481262, 'train/grad_norm_pre_clip': 9.684731483459473, 'timing/data': 0.012183358892798424, 'timing/model': 1.0720130754634738, 'learning_rate/qwen_vl_interface': 8.30434548341565e-06, 'learning_rate/action_model': 8.304345483415649e-05, 'epoch': 0.74}) +[2026-06-17 20:14:59,714][starVLA.training.train_starvla][INFO] - Step 2017, Loss: {'train/loss': 0.6501287817955017, 'train/grad_norm_pre_clip': 12.828710556030273, 'timing/data': 0.010837092064321041, 'timing/model': 1.1035576183348894, 'learning_rate/qwen_vl_interface': 8.30268887181238e-06, 'learning_rate/action_model': 8.302688871812378e-05, 'epoch': 0.74}) +[2026-06-17 20:15:09,009][starVLA.training.train_starvla][INFO] - Step 2018, Loss: {'train/loss': 0.9053521752357483, 'train/grad_norm_pre_clip': 14.605895042419434, 'timing/data': 0.015247529372572899, 'timing/model': 4.096350774168968, 'learning_rate/qwen_vl_interface': 8.301031627383422e-06, 'learning_rate/action_model': 8.301031627383421e-05, 'epoch': 0.74}) +[2026-06-17 20:15:13,423][starVLA.training.train_starvla][INFO] - Step 2019, Loss: {'train/loss': 0.5964910984039307, 'train/grad_norm_pre_clip': 10.105420112609863, 'timing/data': 0.014162003993988037, 'timing/model': 1.366225415840745, 'learning_rate/qwen_vl_interface': 8.299373750472323e-06, 'learning_rate/action_model': 8.299373750472324e-05, 'epoch': 0.74}) +[2026-06-17 20:15:18,007][starVLA.training.train_starvla][INFO] - Step 2020, Loss: {'train/loss': 0.6940786838531494, 'train/grad_norm_pre_clip': 17.88897132873535, 'timing/data': 0.012090479023754597, 'timing/model': 1.131426028907299, 'learning_rate/qwen_vl_interface': 8.297715241422765e-06, 'learning_rate/action_model': 8.297715241422766e-05, 'epoch': 0.74}) +[2026-06-17 20:15:24,162][starVLA.training.train_starvla][INFO] - Step 2021, Loss: {'train/loss': 0.6587038040161133, 'train/grad_norm_pre_clip': 9.736605644226074, 'timing/data': 0.011456048116087914, 'timing/model': 1.1973903179168701, 'learning_rate/qwen_vl_interface': 8.29605610057856e-06, 'learning_rate/action_model': 8.29605610057856e-05, 'epoch': 0.74}) +[2026-06-17 20:15:28,856][starVLA.training.train_starvla][INFO] - Step 2022, Loss: {'train/loss': 0.9675506353378296, 'train/grad_norm_pre_clip': 16.177085876464844, 'timing/data': 0.010766497813165188, 'timing/model': 1.1055490663275123, 'learning_rate/qwen_vl_interface': 8.294396328283647e-06, 'learning_rate/action_model': 8.294396328283645e-05, 'epoch': 0.74}) +[2026-06-17 20:15:36,471][starVLA.training.train_starvla][INFO] - Step 2023, Loss: {'train/loss': 1.0687649250030518, 'train/grad_norm_pre_clip': 17.480024337768555, 'timing/data': 0.015382304787635803, 'timing/model': 3.975678705610335, 'learning_rate/qwen_vl_interface': 8.292735924882097e-06, 'learning_rate/action_model': 8.292735924882097e-05, 'epoch': 0.74}) +[2026-06-17 20:15:40,587][starVLA.training.train_starvla][INFO] - Step 2024, Loss: {'train/loss': 0.5196114182472229, 'train/grad_norm_pre_clip': 10.169041633605957, 'timing/data': 0.022979529574513435, 'timing/model': 1.030921759083867, 'learning_rate/qwen_vl_interface': 8.291074890718117e-06, 'learning_rate/action_model': 8.291074890718117e-05, 'epoch': 0.74}) +[2026-06-17 20:15:45,127][starVLA.training.train_starvla][INFO] - Step 2025, Loss: {'train/loss': 0.6424887776374817, 'train/grad_norm_pre_clip': 14.638862609863281, 'timing/data': 0.03390737250447273, 'timing/model': 1.1200571516528726, 'learning_rate/qwen_vl_interface': 8.28941322613604e-06, 'learning_rate/action_model': 8.289413226136039e-05, 'epoch': 0.74}) +[2026-06-17 20:15:49,763][starVLA.training.train_starvla][INFO] - Step 2026, Loss: {'train/loss': 0.5982610583305359, 'train/grad_norm_pre_clip': 10.356561660766602, 'timing/data': 0.011579612269997597, 'timing/model': 1.0500233881175518, 'learning_rate/qwen_vl_interface': 8.287750931480332e-06, 'learning_rate/action_model': 8.287750931480332e-05, 'epoch': 0.74}) +[2026-06-17 20:15:54,246][starVLA.training.train_starvla][INFO] - Step 2027, Loss: {'train/loss': 0.45180997252464294, 'train/grad_norm_pre_clip': 8.689699172973633, 'timing/data': 0.01179229561239481, 'timing/model': 1.0818744571879506, 'learning_rate/qwen_vl_interface': 8.286088007095584e-06, 'learning_rate/action_model': 8.286088007095584e-05, 'epoch': 0.74}) +[2026-06-17 20:15:58,707][starVLA.training.train_starvla][INFO] - Step 2028, Loss: {'train/loss': 0.5023033022880554, 'train/grad_norm_pre_clip': 16.89796257019043, 'timing/data': 0.010587260127067566, 'timing/model': 1.4286005245521665, 'learning_rate/qwen_vl_interface': 8.284424453326528e-06, 'learning_rate/action_model': 8.284424453326528e-05, 'epoch': 0.74}) +[2026-06-17 20:16:03,023][starVLA.training.train_starvla][INFO] - Step 2029, Loss: {'train/loss': 0.4314289689064026, 'train/grad_norm_pre_clip': 10.53249740600586, 'timing/data': 0.011284168809652328, 'timing/model': 1.3272612942382693, 'learning_rate/qwen_vl_interface': 8.282760270518017e-06, 'learning_rate/action_model': 8.282760270518017e-05, 'epoch': 0.74}) +[2026-06-17 20:16:07,091][starVLA.training.train_starvla][INFO] - Step 2030, Loss: {'train/loss': 0.45244842767715454, 'train/grad_norm_pre_clip': 8.829285621643066, 'timing/data': 0.011742680333554745, 'timing/model': 1.0695719039067626, 'learning_rate/qwen_vl_interface': 8.281095459015037e-06, 'learning_rate/action_model': 8.281095459015036e-05, 'epoch': 0.74}) +[2026-06-17 20:16:11,608][starVLA.training.train_starvla][INFO] - Step 2031, Loss: {'train/loss': 0.6762465238571167, 'train/grad_norm_pre_clip': 13.414344787597656, 'timing/data': 0.015069556422531605, 'timing/model': 1.101363431662321, 'learning_rate/qwen_vl_interface': 8.279430019162709e-06, 'learning_rate/action_model': 8.279430019162708e-05, 'epoch': 0.74}) +[2026-06-17 20:16:16,245][starVLA.training.train_starvla][INFO] - Step 2032, Loss: {'train/loss': 0.5686619281768799, 'train/grad_norm_pre_clip': 11.063602447509766, 'timing/data': 0.018518651835620403, 'timing/model': 1.0549995880573988, 'learning_rate/qwen_vl_interface': 8.277763951306277e-06, 'learning_rate/action_model': 8.277763951306277e-05, 'epoch': 0.74}) +[2026-06-17 20:16:23,684][starVLA.training.train_starvla][INFO] - Step 2033, Loss: {'train/loss': 0.7241930365562439, 'train/grad_norm_pre_clip': 12.815360069274902, 'timing/data': 0.014438875019550323, 'timing/model': 1.0783829521387815, 'learning_rate/qwen_vl_interface': 8.27609725579112e-06, 'learning_rate/action_model': 8.27609725579112e-05, 'epoch': 0.74}) +[2026-06-17 20:16:28,072][starVLA.training.train_starvla][INFO] - Step 2034, Loss: {'train/loss': 0.6654602885246277, 'train/grad_norm_pre_clip': 16.41158676147461, 'timing/data': 0.019308618269860744, 'timing/model': 1.3462656335905194, 'learning_rate/qwen_vl_interface': 8.274429932962745e-06, 'learning_rate/action_model': 8.274429932962746e-05, 'epoch': 0.74}) +[2026-06-17 20:16:32,220][starVLA.training.train_starvla][INFO] - Step 2035, Loss: {'train/loss': 0.7622685432434082, 'train/grad_norm_pre_clip': 15.721137046813965, 'timing/data': 0.010443148203194141, 'timing/model': 1.0614834632724524, 'learning_rate/qwen_vl_interface': 8.272761983166793e-06, 'learning_rate/action_model': 8.272761983166792e-05, 'epoch': 0.75}) +[2026-06-17 20:16:42,122][starVLA.training.train_starvla][INFO] - Step 2036, Loss: {'train/loss': 0.8443771004676819, 'train/grad_norm_pre_clip': 14.374327659606934, 'timing/data': 0.012794921174645424, 'timing/model': 1.0945502864196897, 'learning_rate/qwen_vl_interface': 8.271093406749025e-06, 'learning_rate/action_model': 8.271093406749025e-05, 'epoch': 0.75}) +[2026-06-17 20:16:46,874][starVLA.training.train_starvla][INFO] - Step 2037, Loss: {'train/loss': 0.6336965560913086, 'train/grad_norm_pre_clip': 11.413802146911621, 'timing/data': 0.012546535581350327, 'timing/model': 1.1908417595550418, 'learning_rate/qwen_vl_interface': 8.269424204055346e-06, 'learning_rate/action_model': 8.269424204055346e-05, 'epoch': 0.75}) +[2026-06-17 20:16:51,563][starVLA.training.train_starvla][INFO] - Step 2038, Loss: {'train/loss': 0.5060081481933594, 'train/grad_norm_pre_clip': 10.71201229095459, 'timing/data': 0.01338935736566782, 'timing/model': 1.0862227268517017, 'learning_rate/qwen_vl_interface': 8.26775437543178e-06, 'learning_rate/action_model': 8.26775437543178e-05, 'epoch': 0.75}) +[2026-06-17 20:16:57,873][starVLA.training.train_starvla][INFO] - Step 2039, Loss: {'train/loss': 0.5577821731567383, 'train/grad_norm_pre_clip': 16.06657600402832, 'timing/data': 0.010856998153030872, 'timing/model': 1.3995686210691929, 'learning_rate/qwen_vl_interface': 8.266083921224484e-06, 'learning_rate/action_model': 8.266083921224485e-05, 'epoch': 0.75}) +[2026-06-17 20:17:02,084][starVLA.training.train_starvla][INFO] - Step 2040, Loss: {'train/loss': 0.8504457473754883, 'train/grad_norm_pre_clip': 16.814563751220703, 'timing/data': 0.013710157945752144, 'timing/model': 1.0438343612477183, 'learning_rate/qwen_vl_interface': 8.264412841779747e-06, 'learning_rate/action_model': 8.264412841779746e-05, 'epoch': 0.75}) +[2026-06-17 20:17:06,633][starVLA.training.train_starvla][INFO] - Step 2041, Loss: {'train/loss': 0.8781364560127258, 'train/grad_norm_pre_clip': 15.693085670471191, 'timing/data': 0.011942326091229916, 'timing/model': 1.0673518311232328, 'learning_rate/qwen_vl_interface': 8.262741137443984e-06, 'learning_rate/action_model': 8.262741137443985e-05, 'epoch': 0.75}) +[2026-06-17 20:17:11,986][starVLA.training.train_starvla][INFO] - Step 2042, Loss: {'train/loss': 0.5292798280715942, 'train/grad_norm_pre_clip': 13.065884590148926, 'timing/data': 0.01373568456619978, 'timing/model': 1.0873232102021575, 'learning_rate/qwen_vl_interface': 8.26106880856374e-06, 'learning_rate/action_model': 8.26106880856374e-05, 'epoch': 0.75}) +[2026-06-17 20:17:22,433][starVLA.training.train_starvla][INFO] - Step 2043, Loss: {'train/loss': 0.7106803059577942, 'train/grad_norm_pre_clip': 15.134330749511719, 'timing/data': 0.011875014752149582, 'timing/model': 2.481117518618703, 'learning_rate/qwen_vl_interface': 8.259395855485693e-06, 'learning_rate/action_model': 8.259395855485694e-05, 'epoch': 0.75}) +[2026-06-17 20:17:30,001][starVLA.training.train_starvla][INFO] - Step 2044, Loss: {'train/loss': 0.7922206521034241, 'train/grad_norm_pre_clip': 16.847389221191406, 'timing/data': 0.011831007897853851, 'timing/model': 1.3547260677441955, 'learning_rate/qwen_vl_interface': 8.257722278556647e-06, 'learning_rate/action_model': 8.257722278556646e-05, 'epoch': 0.75}) +[2026-06-17 20:17:34,464][starVLA.training.train_starvla][INFO] - Step 2045, Loss: {'train/loss': 0.6335632801055908, 'train/grad_norm_pre_clip': 11.307172775268555, 'timing/data': 0.014735444448888302, 'timing/model': 1.1407624250277877, 'learning_rate/qwen_vl_interface': 8.256048078123536e-06, 'learning_rate/action_model': 8.256048078123535e-05, 'epoch': 0.75}) +[2026-06-17 20:17:39,018][starVLA.training.train_starvla][INFO] - Step 2046, Loss: {'train/loss': 0.6614722013473511, 'train/grad_norm_pre_clip': 13.79708194732666, 'timing/data': 0.013887534849345684, 'timing/model': 1.0930859399959445, 'learning_rate/qwen_vl_interface': 8.254373254533424e-06, 'learning_rate/action_model': 8.254373254533423e-05, 'epoch': 0.75}) +[2026-06-17 20:17:43,502][starVLA.training.train_starvla][INFO] - Step 2047, Loss: {'train/loss': 1.077428936958313, 'train/grad_norm_pre_clip': 15.728652000427246, 'timing/data': 0.01130211353302002, 'timing/model': 1.1213319189846516, 'learning_rate/qwen_vl_interface': 8.252697808133501e-06, 'learning_rate/action_model': 8.252697808133501e-05, 'epoch': 0.75}) +[2026-06-17 20:17:47,915][starVLA.training.train_starvla][INFO] - Step 2048, Loss: {'train/loss': 0.6358395218849182, 'train/grad_norm_pre_clip': 10.412761688232422, 'timing/data': 0.02107179630547762, 'timing/model': 1.0761879850178957, 'learning_rate/qwen_vl_interface': 8.251021739271095e-06, 'learning_rate/action_model': 8.251021739271094e-05, 'epoch': 0.75}) +[2026-06-17 20:17:52,613][starVLA.training.train_starvla][INFO] - Step 2049, Loss: {'train/loss': 0.9342638850212097, 'train/grad_norm_pre_clip': 103.95999908447266, 'timing/data': 0.01075592078268528, 'timing/model': 1.4401408433914185, 'learning_rate/qwen_vl_interface': 8.249345048293652e-06, 'learning_rate/action_model': 8.24934504829365e-05, 'epoch': 0.75}) +[2026-06-17 20:17:56,890][starVLA.training.train_starvla][INFO] - Step 2050, Loss: {'train/loss': 0.9560872912406921, 'train/grad_norm_pre_clip': 17.67560386657715, 'timing/data': 0.014273170381784439, 'timing/model': 1.0419534109532833, 'learning_rate/qwen_vl_interface': 8.247667735548753e-06, 'learning_rate/action_model': 8.247667735548752e-05, 'epoch': 0.75}) +[2026-06-17 20:18:01,414][starVLA.training.train_starvla][INFO] - Step 2051, Loss: {'train/loss': 0.8906638622283936, 'train/grad_norm_pre_clip': 13.609088897705078, 'timing/data': 0.016924467869102955, 'timing/model': 1.1710479026660323, 'learning_rate/qwen_vl_interface': 8.245989801384104e-06, 'learning_rate/action_model': 8.245989801384104e-05, 'epoch': 0.75}) +[2026-06-17 20:18:06,055][starVLA.training.train_starvla][INFO] - Step 2052, Loss: {'train/loss': 0.731614887714386, 'train/grad_norm_pre_clip': 22.16348648071289, 'timing/data': 0.029199245385825634, 'timing/model': 1.256852657534182, 'learning_rate/qwen_vl_interface': 8.24431124614755e-06, 'learning_rate/action_model': 8.244311246147549e-05, 'epoch': 0.75}) +[2026-06-17 20:18:10,830][starVLA.training.train_starvla][INFO] - Step 2053, Loss: {'train/loss': 0.7297414541244507, 'train/grad_norm_pre_clip': 14.660475730895996, 'timing/data': 0.01343852374702692, 'timing/model': 1.0965170310810208, 'learning_rate/qwen_vl_interface': 8.24263207018705e-06, 'learning_rate/action_model': 8.24263207018705e-05, 'epoch': 0.75}) +[2026-06-17 20:18:17,072][starVLA.training.train_starvla][INFO] - Step 2054, Loss: {'train/loss': 0.7870874404907227, 'train/grad_norm_pre_clip': 13.643780708312988, 'timing/data': 0.01755034551024437, 'timing/model': 1.3912693252786994, 'learning_rate/qwen_vl_interface': 8.2409522738507e-06, 'learning_rate/action_model': 8.240952273850701e-05, 'epoch': 0.75}) +[2026-06-17 20:18:21,266][starVLA.training.train_starvla][INFO] - Step 2055, Loss: {'train/loss': 0.8585503697395325, 'train/grad_norm_pre_clip': 15.12085247039795, 'timing/data': 0.01336771622300148, 'timing/model': 1.0422133691608906, 'learning_rate/qwen_vl_interface': 8.23927185748673e-06, 'learning_rate/action_model': 8.239271857486729e-05, 'epoch': 0.75}) +[2026-06-17 20:18:25,776][starVLA.training.train_starvla][INFO] - Step 2056, Loss: {'train/loss': 0.9634584784507751, 'train/grad_norm_pre_clip': 15.627074241638184, 'timing/data': 0.014441275969147682, 'timing/model': 1.0742869013920426, 'learning_rate/qwen_vl_interface': 8.237590821443484e-06, 'learning_rate/action_model': 8.237590821443483e-05, 'epoch': 0.75}) +[2026-06-17 20:18:32,484][starVLA.training.train_starvla][INFO] - Step 2057, Loss: {'train/loss': 0.6923859119415283, 'train/grad_norm_pre_clip': 13.649271965026855, 'timing/data': 0.012273448519408703, 'timing/model': 1.0299652311950922, 'learning_rate/qwen_vl_interface': 8.235909166069447e-06, 'learning_rate/action_model': 8.235909166069447e-05, 'epoch': 0.75}) +[2026-06-17 20:18:37,113][starVLA.training.train_starvla][INFO] - Step 2058, Loss: {'train/loss': 0.7719312906265259, 'train/grad_norm_pre_clip': 12.336577415466309, 'timing/data': 0.012728296220302582, 'timing/model': 1.0939093260094523, 'learning_rate/qwen_vl_interface': 8.234226891713225e-06, 'learning_rate/action_model': 8.234226891713225e-05, 'epoch': 0.75}) +[2026-06-17 20:18:41,647][starVLA.training.train_starvla][INFO] - Step 2059, Loss: {'train/loss': 0.8762111067771912, 'train/grad_norm_pre_clip': 14.343474388122559, 'timing/data': 0.010238113813102245, 'timing/model': 1.3807480446994305, 'learning_rate/qwen_vl_interface': 8.232543998723559e-06, 'learning_rate/action_model': 8.232543998723559e-05, 'epoch': 0.75}) +[2026-06-17 20:18:47,844][starVLA.training.train_starvla][INFO] - Step 2060, Loss: {'train/loss': 0.7707005739212036, 'train/grad_norm_pre_clip': 17.100976943969727, 'timing/data': 0.014249465428292751, 'timing/model': 1.1012167939916253, 'learning_rate/qwen_vl_interface': 8.230860487449314e-06, 'learning_rate/action_model': 8.230860487449313e-05, 'epoch': 0.75}) +[2026-06-17 20:18:55,508][starVLA.training.train_starvla][INFO] - Step 2061, Loss: {'train/loss': 0.6147454977035522, 'train/grad_norm_pre_clip': 12.530399322509766, 'timing/data': 0.011540569365024567, 'timing/model': 1.0226651635020971, 'learning_rate/qwen_vl_interface': 8.22917635823948e-06, 'learning_rate/action_model': 8.229176358239478e-05, 'epoch': 0.75}) +[2026-06-17 20:19:00,128][starVLA.training.train_starvla][INFO] - Step 2062, Loss: {'train/loss': 0.7308361530303955, 'train/grad_norm_pre_clip': 14.28463077545166, 'timing/data': 0.019875275902450085, 'timing/model': 1.2550350641831756, 'learning_rate/qwen_vl_interface': 8.227491611443182e-06, 'learning_rate/action_model': 8.227491611443182e-05, 'epoch': 0.76}) +[2026-06-17 20:19:04,638][starVLA.training.train_starvla][INFO] - Step 2063, Loss: {'train/loss': 0.5562272071838379, 'train/grad_norm_pre_clip': 10.975690841674805, 'timing/data': 0.011808118782937527, 'timing/model': 1.2560027781873941, 'learning_rate/qwen_vl_interface': 8.225806247409667e-06, 'learning_rate/action_model': 8.225806247409668e-05, 'epoch': 0.76}) +[2026-06-17 20:19:09,257][starVLA.training.train_starvla][INFO] - Step 2064, Loss: {'train/loss': 0.8706955909729004, 'train/grad_norm_pre_clip': 12.995144844055176, 'timing/data': 0.011190938763320446, 'timing/model': 1.3749610520899296, 'learning_rate/qwen_vl_interface': 8.224120266488318e-06, 'learning_rate/action_model': 8.224120266488318e-05, 'epoch': 0.76}) +[2026-06-17 20:19:13,277][starVLA.training.train_starvla][INFO] - Step 2065, Loss: {'train/loss': 0.69185870885849, 'train/grad_norm_pre_clip': 15.903862953186035, 'timing/data': 0.011182614602148533, 'timing/model': 1.078163300640881, 'learning_rate/qwen_vl_interface': 8.222433669028636e-06, 'learning_rate/action_model': 8.222433669028635e-05, 'epoch': 0.76}) +[2026-06-17 20:19:19,563][starVLA.training.train_starvla][INFO] - Step 2066, Loss: {'train/loss': 0.6555567383766174, 'train/grad_norm_pre_clip': 13.496993064880371, 'timing/data': 0.01630602963268757, 'timing/model': 1.1573538454249501, 'learning_rate/qwen_vl_interface': 8.220746455380257e-06, 'learning_rate/action_model': 8.220746455380257e-05, 'epoch': 0.76}) +[2026-06-17 20:19:24,235][starVLA.training.train_starvla][INFO] - Step 2067, Loss: {'train/loss': 0.7340000867843628, 'train/grad_norm_pre_clip': 14.111265182495117, 'timing/data': 0.011983483098447323, 'timing/model': 1.0841139443218708, 'learning_rate/qwen_vl_interface': 8.219058625892939e-06, 'learning_rate/action_model': 8.21905862589294e-05, 'epoch': 0.76}) +[2026-06-17 20:19:28,642][starVLA.training.train_starvla][INFO] - Step 2068, Loss: {'train/loss': 0.8200210928916931, 'train/grad_norm_pre_clip': 13.418957710266113, 'timing/data': 0.011079578660428524, 'timing/model': 1.1549659688025713, 'learning_rate/qwen_vl_interface': 8.217370180916574e-06, 'learning_rate/action_model': 8.217370180916575e-05, 'epoch': 0.76}) +[2026-06-17 20:19:32,983][starVLA.training.train_starvla][INFO] - Step 2069, Loss: {'train/loss': 0.8107872605323792, 'train/grad_norm_pre_clip': 15.655858993530273, 'timing/data': 0.009826880879700184, 'timing/model': 1.3156625693663955, 'learning_rate/qwen_vl_interface': 8.215681120801179e-06, 'learning_rate/action_model': 8.215681120801179e-05, 'epoch': 0.76}) +[2026-06-17 20:19:39,797][starVLA.training.train_starvla][INFO] - Step 2070, Loss: {'train/loss': 0.833268404006958, 'train/grad_norm_pre_clip': 15.783102035522461, 'timing/data': 0.013104535639286041, 'timing/model': 1.0526682380586863, 'learning_rate/qwen_vl_interface': 8.213991445896895e-06, 'learning_rate/action_model': 8.213991445896895e-05, 'epoch': 0.76}) +[2026-06-17 20:19:44,464][starVLA.training.train_starvla][INFO] - Step 2071, Loss: {'train/loss': 0.8348766565322876, 'train/grad_norm_pre_clip': 12.704032897949219, 'timing/data': 0.013769520446658134, 'timing/model': 1.0951231913641095, 'learning_rate/qwen_vl_interface': 8.212301156553995e-06, 'learning_rate/action_model': 8.212301156553995e-05, 'epoch': 0.76}) +[2026-06-17 20:19:49,252][starVLA.training.train_starvla][INFO] - Step 2072, Loss: {'train/loss': 0.6046759486198425, 'train/grad_norm_pre_clip': 11.637422561645508, 'timing/data': 0.011260569095611572, 'timing/model': 1.1118823057040572, 'learning_rate/qwen_vl_interface': 8.210610253122879e-06, 'learning_rate/action_model': 8.210610253122879e-05, 'epoch': 0.76}) +[2026-06-17 20:19:53,746][starVLA.training.train_starvla][INFO] - Step 2073, Loss: {'train/loss': 0.5261065363883972, 'train/grad_norm_pre_clip': 11.760096549987793, 'timing/data': 0.014763502404093742, 'timing/model': 1.1091120056807995, 'learning_rate/qwen_vl_interface': 8.20891873595407e-06, 'learning_rate/action_model': 8.20891873595407e-05, 'epoch': 0.76}) +[2026-06-17 20:19:58,352][starVLA.training.train_starvla][INFO] - Step 2074, Loss: {'train/loss': 0.6555507183074951, 'train/grad_norm_pre_clip': 11.541224479675293, 'timing/data': 0.011168400757014751, 'timing/model': 1.451128063723445, 'learning_rate/qwen_vl_interface': 8.207226605398225e-06, 'learning_rate/action_model': 8.207226605398224e-05, 'epoch': 0.76}) +[2026-06-17 20:20:02,461][starVLA.training.train_starvla][INFO] - Step 2075, Loss: {'train/loss': 0.8807172775268555, 'train/grad_norm_pre_clip': 13.772315979003906, 'timing/data': 0.011372774839401245, 'timing/model': 1.120717910118401, 'learning_rate/qwen_vl_interface': 8.20553386180612e-06, 'learning_rate/action_model': 8.205533861806119e-05, 'epoch': 0.76}) +[2026-06-17 20:20:12,407][starVLA.training.train_starvla][INFO] - Step 2076, Loss: {'train/loss': 0.6861075758934021, 'train/grad_norm_pre_clip': 12.801895141601562, 'timing/data': 0.011877263896167278, 'timing/model': 4.139338277280331, 'learning_rate/qwen_vl_interface': 8.203840505528665e-06, 'learning_rate/action_model': 8.203840505528665e-05, 'epoch': 0.76}) +[2026-06-17 20:20:24,414][starVLA.training.train_starvla][INFO] - Step 2077, Loss: {'train/loss': 0.6324232220649719, 'train/grad_norm_pre_clip': 13.665388107299805, 'timing/data': 0.01366026233881712, 'timing/model': 1.1222263285890222, 'learning_rate/qwen_vl_interface': 8.202146536916899e-06, 'learning_rate/action_model': 8.202146536916899e-05, 'epoch': 0.76}) +[2026-06-17 20:20:28,990][starVLA.training.train_starvla][INFO] - Step 2078, Loss: {'train/loss': 0.6287594437599182, 'train/grad_norm_pre_clip': 13.45140552520752, 'timing/data': 0.013627187348902225, 'timing/model': 1.076914275996387, 'learning_rate/qwen_vl_interface': 8.200451956321974e-06, 'learning_rate/action_model': 8.200451956321974e-05, 'epoch': 0.76}) +[2026-06-17 20:20:33,488][starVLA.training.train_starvla][INFO] - Step 2079, Loss: {'train/loss': 0.5982665419578552, 'train/grad_norm_pre_clip': 11.485801696777344, 'timing/data': 0.011942511424422264, 'timing/model': 1.4414411159232259, 'learning_rate/qwen_vl_interface': 8.198756764095187e-06, 'learning_rate/action_model': 8.198756764095187e-05, 'epoch': 0.76}) +[2026-06-17 20:20:37,878][starVLA.training.train_starvla][INFO] - Step 2080, Loss: {'train/loss': 0.6115727424621582, 'train/grad_norm_pre_clip': 11.42465591430664, 'timing/data': 0.012712615542113781, 'timing/model': 1.085982359945774, 'learning_rate/qwen_vl_interface': 8.19706096058795e-06, 'learning_rate/action_model': 8.197060960587949e-05, 'epoch': 0.76}) +[2026-06-17 20:20:42,499][starVLA.training.train_starvla][INFO] - Step 2081, Loss: {'train/loss': 0.6811162233352661, 'train/grad_norm_pre_clip': 12.32113265991211, 'timing/data': 0.014695316553115845, 'timing/model': 1.1048588724806905, 'learning_rate/qwen_vl_interface': 8.195364546151802e-06, 'learning_rate/action_model': 8.195364546151801e-05, 'epoch': 0.76}) +[2026-06-17 20:20:47,935][starVLA.training.train_starvla][INFO] - Step 2082, Loss: {'train/loss': 0.652884304523468, 'train/grad_norm_pre_clip': 12.193320274353027, 'timing/data': 0.013112395070493221, 'timing/model': 1.0481518665328622, 'learning_rate/qwen_vl_interface': 8.193667521138416e-06, 'learning_rate/action_model': 8.193667521138416e-05, 'epoch': 0.76}) +[2026-06-17 20:20:52,284][starVLA.training.train_starvla][INFO] - Step 2083, Loss: {'train/loss': 0.8473365902900696, 'train/grad_norm_pre_clip': 18.05620574951172, 'timing/data': 0.011717699468135834, 'timing/model': 1.0409886734560132, 'learning_rate/qwen_vl_interface': 8.191969885899585e-06, 'learning_rate/action_model': 8.191969885899585e-05, 'epoch': 0.76}) +[2026-06-17 20:20:57,028][starVLA.training.train_starvla][INFO] - Step 2084, Loss: {'train/loss': 0.8378729820251465, 'train/grad_norm_pre_clip': 17.029653549194336, 'timing/data': 0.01056963112205267, 'timing/model': 1.4330240469425917, 'learning_rate/qwen_vl_interface': 8.190271640787232e-06, 'learning_rate/action_model': 8.190271640787232e-05, 'epoch': 0.76}) +[2026-06-17 20:21:01,054][starVLA.training.train_starvla][INFO] - Step 2085, Loss: {'train/loss': 0.48376524448394775, 'train/grad_norm_pre_clip': 10.746832847595215, 'timing/data': 0.012064198032021523, 'timing/model': 1.0355117358267307, 'learning_rate/qwen_vl_interface': 8.188572786153403e-06, 'learning_rate/action_model': 8.188572786153402e-05, 'epoch': 0.76}) +[2026-06-17 20:21:08,805][starVLA.training.train_starvla][INFO] - Step 2086, Loss: {'train/loss': 0.4729846715927124, 'train/grad_norm_pre_clip': 11.871196746826172, 'timing/data': 0.016510589979588985, 'timing/model': 4.376123298890889, 'learning_rate/qwen_vl_interface': 8.186873322350274e-06, 'learning_rate/action_model': 8.186873322350273e-05, 'epoch': 0.76}) +[2026-06-17 20:21:13,425][starVLA.training.train_starvla][INFO] - Step 2087, Loss: {'train/loss': 0.9729926586151123, 'train/grad_norm_pre_clip': 13.959222793579102, 'timing/data': 0.019124886021018028, 'timing/model': 1.0773727893829346, 'learning_rate/qwen_vl_interface': 8.185173249730143e-06, 'learning_rate/action_model': 8.185173249730142e-05, 'epoch': 0.76}) +[2026-06-17 20:21:20,871][starVLA.training.train_starvla][INFO] - Step 2088, Loss: {'train/loss': 0.6358382701873779, 'train/grad_norm_pre_clip': 11.693033218383789, 'timing/data': 0.013612998649477959, 'timing/model': 4.163002890534699, 'learning_rate/qwen_vl_interface': 8.183472568645441e-06, 'learning_rate/action_model': 8.18347256864544e-05, 'epoch': 0.76}) +[2026-06-17 20:21:28,378][starVLA.training.train_starvla][INFO] - Step 2089, Loss: {'train/loss': 0.690726637840271, 'train/grad_norm_pre_clip': 14.721960067749023, 'timing/data': 0.01039876975119114, 'timing/model': 1.4288971228525043, 'learning_rate/qwen_vl_interface': 8.181771279448718e-06, 'learning_rate/action_model': 8.181771279448718e-05, 'epoch': 0.76}) +[2026-06-17 20:21:32,730][starVLA.training.train_starvla][INFO] - Step 2090, Loss: {'train/loss': 0.7439017295837402, 'train/grad_norm_pre_clip': 16.33782386779785, 'timing/data': 0.012158308178186417, 'timing/model': 1.1305909510701895, 'learning_rate/qwen_vl_interface': 8.180069382492655e-06, 'learning_rate/action_model': 8.180069382492654e-05, 'epoch': 0.77}) +[2026-06-17 20:21:37,367][starVLA.training.train_starvla][INFO] - Step 2091, Loss: {'train/loss': 0.5461228489875793, 'train/grad_norm_pre_clip': 12.590340614318848, 'timing/data': 0.01087542250752449, 'timing/model': 1.086541653610766, 'learning_rate/qwen_vl_interface': 8.178366878130056e-06, 'learning_rate/action_model': 8.178366878130056e-05, 'epoch': 0.77}) +[2026-06-17 20:21:43,751][starVLA.training.train_starvla][INFO] - Step 2092, Loss: {'train/loss': 0.704253613948822, 'train/grad_norm_pre_clip': 18.073081970214844, 'timing/data': 0.013537464663386345, 'timing/model': 1.1991972355172038, 'learning_rate/qwen_vl_interface': 8.176663766713852e-06, 'learning_rate/action_model': 8.176663766713852e-05, 'epoch': 0.77}) +[2026-06-17 20:21:50,303][starVLA.training.train_starvla][INFO] - Step 2093, Loss: {'train/loss': 0.6390923261642456, 'train/grad_norm_pre_clip': 11.441837310791016, 'timing/data': 0.020845726132392883, 'timing/model': 1.0668873460963368, 'learning_rate/qwen_vl_interface': 8.174960048597099e-06, 'learning_rate/action_model': 8.174960048597099e-05, 'epoch': 0.77}) +[2026-06-17 20:21:58,816][starVLA.training.train_starvla][INFO] - Step 2094, Loss: {'train/loss': 0.8620901107788086, 'train/grad_norm_pre_clip': 14.425440788269043, 'timing/data': 0.021654099225997925, 'timing/model': 3.15745606739074, 'learning_rate/qwen_vl_interface': 8.173255724132982e-06, 'learning_rate/action_model': 8.17325572413298e-05, 'epoch': 0.77}) +[2026-06-17 20:22:02,898][starVLA.training.train_starvla][INFO] - Step 2095, Loss: {'train/loss': 0.5487748384475708, 'train/grad_norm_pre_clip': 11.670701026916504, 'timing/data': 0.011161230504512787, 'timing/model': 1.047026696614921, 'learning_rate/qwen_vl_interface': 8.171550793674806e-06, 'learning_rate/action_model': 8.171550793674805e-05, 'epoch': 0.77}) +[2026-06-17 20:22:07,430][starVLA.training.train_starvla][INFO] - Step 2096, Loss: {'train/loss': 0.7511873245239258, 'train/grad_norm_pre_clip': 15.643453598022461, 'timing/data': 0.011862128041684628, 'timing/model': 1.110492437146604, 'learning_rate/qwen_vl_interface': 8.169845257576008e-06, 'learning_rate/action_model': 8.169845257576007e-05, 'epoch': 0.77}) +[2026-06-17 20:22:12,182][starVLA.training.train_starvla][INFO] - Step 2097, Loss: {'train/loss': 0.6441870927810669, 'train/grad_norm_pre_clip': 14.009493827819824, 'timing/data': 0.01203213632106781, 'timing/model': 1.093052021227777, 'learning_rate/qwen_vl_interface': 8.168139116190146e-06, 'learning_rate/action_model': 8.168139116190146e-05, 'epoch': 0.77}) +[2026-06-17 20:22:16,990][starVLA.training.train_starvla][INFO] - Step 2098, Loss: {'train/loss': 0.9945574998855591, 'train/grad_norm_pre_clip': 26.874725341796875, 'timing/data': 0.013216640800237656, 'timing/model': 1.1164369210600853, 'learning_rate/qwen_vl_interface': 8.166432369870905e-06, 'learning_rate/action_model': 8.166432369870904e-05, 'epoch': 0.77}) +[2026-06-17 20:22:21,539][starVLA.training.train_starvla][INFO] - Step 2099, Loss: {'train/loss': 0.7129694819450378, 'train/grad_norm_pre_clip': 14.30089282989502, 'timing/data': 0.011835051700472832, 'timing/model': 1.46162177529186, 'learning_rate/qwen_vl_interface': 8.164725018972096e-06, 'learning_rate/action_model': 8.164725018972096e-05, 'epoch': 0.77}) +[2026-06-17 20:26:15,884][starVLA.training.train_starvla][INFO] - Step 2100, Loss: {'train/loss': 0.7494968175888062, 'train/grad_norm_pre_clip': 15.521967887878418, 'eval/loss': 0.7963147759437561, 'eval/latency_2/loss': 0.7963147759437561, 'eval/demon_fire_precision': 0.878515625, 'eval/demon_fire_recall': 0.9410041841004184, 'eval/demon_fire_f1': 0.9086868686868688, 'eval/demon_fire_rate_model': 0.7122982749026154, 'eval/demon_fire_rate_teacher': 0.6649972175848636, 'eval/demon_left_precision': 0.8195232690124858, 'eval/demon_left_recall': 0.76, 'eval/demon_left_f1': 0.7886400873839433, 'eval/demon_left_rate_model': 0.2451307735114079, 'eval/demon_left_rate_teacher': 0.2643294379521425, 'eval/demon_right_precision': 0.7536800785083415, 'eval/demon_right_recall': 0.8677966101694915, 'eval/demon_right_f1': 0.8067226890756303, 'eval/demon_right_rate_model': 0.283528102392877, 'eval/demon_right_rate_teacher': 0.24624373956594323, 'eval/demon_move_f1': 0.7976813882297868, 'eval/demon_cc_f1': 0.8722712433919749, 'eval/latency_2/demon_fire_precision': 0.878515625, 'eval/latency_2/demon_fire_recall': 0.9410041841004184, 'eval/latency_2/demon_fire_f1': 0.9086868686868688, 'eval/latency_2/demon_fire_rate_model': 0.7122982749026154, 'eval/latency_2/demon_fire_rate_teacher': 0.6649972175848636, 'eval/latency_2/demon_left_precision': 0.8195232690124858, 'eval/latency_2/demon_left_recall': 0.76, 'eval/latency_2/demon_left_f1': 0.7886400873839433, 'eval/latency_2/demon_left_rate_model': 0.2451307735114079, 'eval/latency_2/demon_left_rate_teacher': 0.2643294379521425, 'eval/latency_2/demon_right_precision': 0.7536800785083415, 'eval/latency_2/demon_right_recall': 0.8677966101694915, 'eval/latency_2/demon_right_f1': 0.8067226890756303, 'eval/latency_2/demon_right_rate_model': 0.283528102392877, 'eval/latency_2/demon_right_rate_teacher': 0.24624373956594323, 'eval/latency_2/demon_move_f1': 0.7976813882297868, 'eval/latency_2/demon_cc_f1': 0.8722712433919749, 'timing/data': 0.011026457883417606, 'timing/model': 1.181684236973524, 'learning_rate/qwen_vl_interface': 8.163017063847654e-06, 'learning_rate/action_model': 8.163017063847654e-05, 'epoch': 0.77}) +[2026-06-17 20:26:15,886][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_2100_state +[2026-06-17 20:26:15,893][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 20:26:24,551][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_2100_state/model.safetensors +[2026-06-17 20:26:39,729][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_2100_state/optimizer.bin +[2026-06-17 20:26:39,730][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_2100_state/sampler.bin +[2026-06-17 20:26:39,731][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_2100_state/sampler_1.bin +[2026-06-17 20:26:39,733][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_2100_state/random_states_0.pkl +[2026-06-17 20:26:39,739][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2100 +[2026-06-17 20:26:39,739][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 20:26:39,755][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 20:26:44,199][starVLA.training.train_starvla][INFO] - Step 2101, Loss: {'train/loss': 0.6001425981521606, 'train/grad_norm_pre_clip': 14.335444450378418, 'timing/data': 0.012012714520096779, 'timing/model': 1.128343333490193, 'learning_rate/qwen_vl_interface': 8.16130850485164e-06, 'learning_rate/action_model': 8.161308504851639e-05, 'epoch': 0.77}) +[2026-06-17 20:26:52,122][starVLA.training.train_starvla][INFO] - Step 2102, Loss: {'train/loss': 0.8083817958831787, 'train/grad_norm_pre_clip': 27.184123992919922, 'timing/data': 0.015888302586972713, 'timing/model': 1.0599063020199537, 'learning_rate/qwen_vl_interface': 8.15959934233824e-06, 'learning_rate/action_model': 8.15959934233824e-05, 'epoch': 0.77}) +[2026-06-17 20:26:57,031][starVLA.training.train_starvla][INFO] - Step 2103, Loss: {'train/loss': 0.5658068060874939, 'train/grad_norm_pre_clip': 13.551197052001953, 'timing/data': 0.021405712701380253, 'timing/model': 1.0945126973092556, 'learning_rate/qwen_vl_interface': 8.157889576661765e-06, 'learning_rate/action_model': 8.157889576661764e-05, 'epoch': 0.77}) +[2026-06-17 20:27:02,617][starVLA.training.train_starvla][INFO] - Step 2104, Loss: {'train/loss': 0.6241689324378967, 'train/grad_norm_pre_clip': 12.496708869934082, 'timing/data': 0.013338646851480007, 'timing/model': 1.3138540964573622, 'learning_rate/qwen_vl_interface': 8.156179208176652e-06, 'learning_rate/action_model': 8.156179208176652e-05, 'epoch': 0.77}) +[2026-06-17 20:27:11,223][starVLA.training.train_starvla][INFO] - Step 2105, Loss: {'train/loss': 0.6279549598693848, 'train/grad_norm_pre_clip': 16.393495559692383, 'timing/data': 0.011082012206315994, 'timing/model': 5.0436685383319855, 'learning_rate/qwen_vl_interface': 8.154468237237463e-06, 'learning_rate/action_model': 8.154468237237462e-05, 'epoch': 0.77}) +[2026-06-17 20:27:22,852][starVLA.training.train_starvla][INFO] - Step 2106, Loss: {'train/loss': 0.8873980045318604, 'train/grad_norm_pre_clip': 17.16205406188965, 'timing/data': 0.013975047506392002, 'timing/model': 1.2162947980687022, 'learning_rate/qwen_vl_interface': 8.152756664198882e-06, 'learning_rate/action_model': 8.152756664198882e-05, 'epoch': 0.77}) +[2026-06-17 20:27:31,628][starVLA.training.train_starvla][INFO] - Step 2107, Loss: {'train/loss': 0.684876561164856, 'train/grad_norm_pre_clip': 17.385150909423828, 'timing/data': 0.012728141620755196, 'timing/model': 3.337970376946032, 'learning_rate/qwen_vl_interface': 8.151044489415722e-06, 'learning_rate/action_model': 8.151044489415722e-05, 'epoch': 0.77}) +[2026-06-17 20:27:36,928][starVLA.training.train_starvla][INFO] - Step 2108, Loss: {'train/loss': 0.7629207372665405, 'train/grad_norm_pre_clip': 15.808602333068848, 'timing/data': 0.011653413064777851, 'timing/model': 1.0550924884155393, 'learning_rate/qwen_vl_interface': 8.149331713242916e-06, 'learning_rate/action_model': 8.149331713242916e-05, 'epoch': 0.77}) +[2026-06-17 20:27:46,631][starVLA.training.train_starvla][INFO] - Step 2109, Loss: {'train/loss': 0.6365696787834167, 'train/grad_norm_pre_clip': 15.152188301086426, 'timing/data': 0.01515771821141243, 'timing/model': 1.2116742972284555, 'learning_rate/qwen_vl_interface': 8.147618336035526e-06, 'learning_rate/action_model': 8.147618336035526e-05, 'epoch': 0.77}) +[2026-06-17 20:27:51,288][starVLA.training.train_starvla][INFO] - Step 2110, Loss: {'train/loss': 0.9021296501159668, 'train/grad_norm_pre_clip': 21.327056884765625, 'timing/data': 0.01151991542428732, 'timing/model': 1.417136968113482, 'learning_rate/qwen_vl_interface': 8.14590435814874e-06, 'learning_rate/action_model': 8.14590435814874e-05, 'epoch': 0.77}) +[2026-06-17 20:27:55,879][starVLA.training.train_starvla][INFO] - Step 2111, Loss: {'train/loss': 0.9080686569213867, 'train/grad_norm_pre_clip': 14.757164001464844, 'timing/data': 0.012238935567438602, 'timing/model': 1.4028403041884303, 'learning_rate/qwen_vl_interface': 8.144189779937862e-06, 'learning_rate/action_model': 8.144189779937862e-05, 'epoch': 0.77}) +[2026-06-17 20:27:59,969][starVLA.training.train_starvla][INFO] - Step 2112, Loss: {'train/loss': 1.0103739500045776, 'train/grad_norm_pre_clip': 20.711181640625, 'timing/data': 0.01074235513806343, 'timing/model': 1.0471640089526772, 'learning_rate/qwen_vl_interface': 8.142474601758331e-06, 'learning_rate/action_model': 8.142474601758329e-05, 'epoch': 0.77}) +[2026-06-17 20:28:04,391][starVLA.training.train_starvla][INFO] - Step 2113, Loss: {'train/loss': 0.6155321598052979, 'train/grad_norm_pre_clip': 15.882864952087402, 'timing/data': 0.012080361135303974, 'timing/model': 1.0517492229118943, 'learning_rate/qwen_vl_interface': 8.1407588239657e-06, 'learning_rate/action_model': 8.140758823965698e-05, 'epoch': 0.77}) +[2026-06-17 20:28:09,024][starVLA.training.train_starvla][INFO] - Step 2114, Loss: {'train/loss': 0.7269859313964844, 'train/grad_norm_pre_clip': 15.49377155303955, 'timing/data': 0.01542675495147705, 'timing/model': 1.3678311984986067, 'learning_rate/qwen_vl_interface': 8.139042446915653e-06, 'learning_rate/action_model': 8.139042446915654e-05, 'epoch': 0.77}) +[2026-06-17 20:28:13,326][starVLA.training.train_starvla][INFO] - Step 2115, Loss: {'train/loss': 0.8006561994552612, 'train/grad_norm_pre_clip': 18.341184616088867, 'timing/data': 0.011070908978581429, 'timing/model': 1.0304479338228703, 'learning_rate/qwen_vl_interface': 8.137325470964001e-06, 'learning_rate/action_model': 8.137325470964e-05, 'epoch': 0.77}) +[2026-06-17 20:28:17,750][starVLA.training.train_starvla][INFO] - Step 2116, Loss: {'train/loss': 0.9188422560691833, 'train/grad_norm_pre_clip': 18.010454177856445, 'timing/data': 0.010747898370027542, 'timing/model': 1.2882467675954103, 'learning_rate/qwen_vl_interface': 8.135607896466669e-06, 'learning_rate/action_model': 8.13560789646667e-05, 'epoch': 0.77}) +[2026-06-17 20:28:21,757][starVLA.training.train_starvla][INFO] - Step 2117, Loss: {'train/loss': 0.7675708532333374, 'train/grad_norm_pre_clip': 13.129059791564941, 'timing/data': 0.011274370364844799, 'timing/model': 1.0279405368492007, 'learning_rate/qwen_vl_interface': 8.133889723779718e-06, 'learning_rate/action_model': 8.133889723779717e-05, 'epoch': 0.78}) +[2026-06-17 20:28:26,292][starVLA.training.train_starvla][INFO] - Step 2118, Loss: {'train/loss': 0.6805241703987122, 'train/grad_norm_pre_clip': 14.856040954589844, 'timing/data': 0.011189182288944721, 'timing/model': 1.2010747361928225, 'learning_rate/qwen_vl_interface': 8.132170953259321e-06, 'learning_rate/action_model': 8.132170953259322e-05, 'epoch': 0.78}) +[2026-06-17 20:28:30,760][starVLA.training.train_starvla][INFO] - Step 2119, Loss: {'train/loss': 0.8688370585441589, 'train/grad_norm_pre_clip': 14.631180763244629, 'timing/data': 0.013058450073003769, 'timing/model': 1.0275342259556055, 'learning_rate/qwen_vl_interface': 8.130451585261786e-06, 'learning_rate/action_model': 8.130451585261786e-05, 'epoch': 0.78}) +[2026-06-17 20:28:35,280][starVLA.training.train_starvla][INFO] - Step 2120, Loss: {'train/loss': 0.3542771637439728, 'train/grad_norm_pre_clip': 7.960671424865723, 'timing/data': 0.03166446276009083, 'timing/model': 1.427337340079248, 'learning_rate/qwen_vl_interface': 8.128731620143538e-06, 'learning_rate/action_model': 8.128731620143538e-05, 'epoch': 0.78}) +[2026-06-17 20:28:39,636][starVLA.training.train_starvla][INFO] - Step 2121, Loss: {'train/loss': 0.6855514645576477, 'train/grad_norm_pre_clip': 14.511877059936523, 'timing/data': 0.012449883855879307, 'timing/model': 1.0415408872067928, 'learning_rate/qwen_vl_interface': 8.127011058261129e-06, 'learning_rate/action_model': 8.127011058261128e-05, 'epoch': 0.78}) +[2026-06-17 20:28:44,188][starVLA.training.train_starvla][INFO] - Step 2122, Loss: {'train/loss': 0.7803635597229004, 'train/grad_norm_pre_clip': 14.244767189025879, 'timing/data': 0.01603571604937315, 'timing/model': 1.0422430764883757, 'learning_rate/qwen_vl_interface': 8.125289899971229e-06, 'learning_rate/action_model': 8.125289899971228e-05, 'epoch': 0.78}) +[2026-06-17 20:28:48,658][starVLA.training.train_starvla][INFO] - Step 2123, Loss: {'train/loss': 0.5416364073753357, 'train/grad_norm_pre_clip': 83.36079406738281, 'timing/data': 0.012566988356411457, 'timing/model': 1.0368299167603254, 'learning_rate/qwen_vl_interface': 8.123568145630641e-06, 'learning_rate/action_model': 8.12356814563064e-05, 'epoch': 0.78}) +[2026-06-17 20:28:53,172][starVLA.training.train_starvla][INFO] - Step 2124, Loss: {'train/loss': 0.8300545811653137, 'train/grad_norm_pre_clip': 14.580573081970215, 'timing/data': 0.011137359775602818, 'timing/model': 1.1047760965302587, 'learning_rate/qwen_vl_interface': 8.121845795596282e-06, 'learning_rate/action_model': 8.121845795596282e-05, 'epoch': 0.78}) +[2026-06-17 20:29:03,809][starVLA.training.train_starvla][INFO] - Step 2125, Loss: {'train/loss': 0.7133802175521851, 'train/grad_norm_pre_clip': 13.567159652709961, 'timing/data': 0.018633521161973476, 'timing/model': 5.588049467653036, 'learning_rate/qwen_vl_interface': 8.1201228502252e-06, 'learning_rate/action_model': 8.120122850225199e-05, 'epoch': 0.78}) +[2026-06-17 20:29:10,906][starVLA.training.train_starvla][INFO] - Step 2126, Loss: {'train/loss': 0.74812912940979, 'train/grad_norm_pre_clip': 14.671875, 'timing/data': 0.0010658185929059982, 'timing/model': 3.5123352771624923, 'learning_rate/qwen_vl_interface': 8.118399309874563e-06, 'learning_rate/action_model': 8.118399309874563e-05, 'epoch': 0.78}) +[2026-06-17 20:29:23,744][starVLA.training.train_starvla][INFO] - Step 2127, Loss: {'train/loss': 0.6315212249755859, 'train/grad_norm_pre_clip': 16.037750244140625, 'timing/data': 0.001151176169514656, 'timing/model': 5.452411721460521, 'learning_rate/qwen_vl_interface': 8.116675174901663e-06, 'learning_rate/action_model': 8.116675174901662e-05, 'epoch': 0.78}) +[2026-06-17 20:29:28,796][starVLA.training.train_starvla][INFO] - Step 2128, Loss: {'train/loss': 0.7903677821159363, 'train/grad_norm_pre_clip': 15.834342956542969, 'timing/data': 0.0014235125854611397, 'timing/model': 1.4043404152616858, 'learning_rate/qwen_vl_interface': 8.114950445663915e-06, 'learning_rate/action_model': 8.114950445663913e-05, 'epoch': 0.78}) +[2026-06-17 20:29:34,599][starVLA.training.train_starvla][INFO] - Step 2129, Loss: {'train/loss': 0.5456390380859375, 'train/grad_norm_pre_clip': 12.464375495910645, 'timing/data': 0.0011504637077450752, 'timing/model': 1.0671370709314942, 'learning_rate/qwen_vl_interface': 8.113225122518854e-06, 'learning_rate/action_model': 8.113225122518854e-05, 'epoch': 0.78}) +[2026-06-17 20:29:46,628][starVLA.training.train_starvla][INFO] - Step 2130, Loss: {'train/loss': 0.7759006023406982, 'train/grad_norm_pre_clip': 32.20113754272461, 'timing/data': 0.0012772390618920326, 'timing/model': 3.204087609425187, 'learning_rate/qwen_vl_interface': 8.111499205824148e-06, 'learning_rate/action_model': 8.111499205824147e-05, 'epoch': 0.78}) +[2026-06-17 20:29:55,820][starVLA.training.train_starvla][INFO] - Step 2131, Loss: {'train/loss': 0.5506445169448853, 'train/grad_norm_pre_clip': 14.537932395935059, 'timing/data': 0.0008606612682342529, 'timing/model': 4.575146401301026, 'learning_rate/qwen_vl_interface': 8.109772695937575e-06, 'learning_rate/action_model': 8.109772695937574e-05, 'epoch': 0.78}) +[2026-06-17 20:30:02,341][starVLA.training.train_starvla][INFO] - Step 2132, Loss: {'train/loss': 0.7365846037864685, 'train/grad_norm_pre_clip': 16.144088745117188, 'timing/data': 0.001299324445426464, 'timing/model': 1.252328421920538, 'learning_rate/qwen_vl_interface': 8.108045593217045e-06, 'learning_rate/action_model': 8.108045593217045e-05, 'epoch': 0.78}) +[2026-06-17 20:30:11,837][starVLA.training.train_starvla][INFO] - Step 2133, Loss: {'train/loss': 0.536584198474884, 'train/grad_norm_pre_clip': 14.4695463180542, 'timing/data': 0.0017419997602701187, 'timing/model': 1.0669193658977747, 'learning_rate/qwen_vl_interface': 8.106317898020589e-06, 'learning_rate/action_model': 8.106317898020589e-05, 'epoch': 0.78}) +[2026-06-17 20:30:19,706][starVLA.training.train_starvla][INFO] - Step 2134, Loss: {'train/loss': 0.621996283531189, 'train/grad_norm_pre_clip': 13.235133171081543, 'timing/data': 0.000980868935585022, 'timing/model': 1.0360285947099328, 'learning_rate/qwen_vl_interface': 8.104589610706359e-06, 'learning_rate/action_model': 8.104589610706359e-05, 'epoch': 0.78}) +[2026-06-17 20:30:24,053][starVLA.training.train_starvla][INFO] - Step 2135, Loss: {'train/loss': 0.6372410655021667, 'train/grad_norm_pre_clip': 13.480966567993164, 'timing/data': 0.0007719807326793671, 'timing/model': 1.025541485287249, 'learning_rate/qwen_vl_interface': 8.102860731632632e-06, 'learning_rate/action_model': 8.102860731632631e-05, 'epoch': 0.78}) +[2026-06-17 20:30:28,470][starVLA.training.train_starvla][INFO] - Step 2136, Loss: {'train/loss': 0.5446330904960632, 'train/grad_norm_pre_clip': 14.297674179077148, 'timing/data': 0.001082257367670536, 'timing/model': 1.3206711616367102, 'learning_rate/qwen_vl_interface': 8.101131261157802e-06, 'learning_rate/action_model': 8.101131261157802e-05, 'epoch': 0.78}) +[2026-06-17 20:30:32,774][starVLA.training.train_starvla][INFO] - Step 2137, Loss: {'train/loss': 1.0722421407699585, 'train/grad_norm_pre_clip': 17.488773345947266, 'timing/data': 0.0011421190574765205, 'timing/model': 1.0692525384947658, 'learning_rate/qwen_vl_interface': 8.099401199640397e-06, 'learning_rate/action_model': 8.099401199640397e-05, 'epoch': 0.78}) +[2026-06-17 20:30:37,182][starVLA.training.train_starvla][INFO] - Step 2138, Loss: {'train/loss': 0.8507179617881775, 'train/grad_norm_pre_clip': 14.695796012878418, 'timing/data': 0.0010723443701863289, 'timing/model': 1.038759021088481, 'learning_rate/qwen_vl_interface': 8.097670547439057e-06, 'learning_rate/action_model': 8.097670547439058e-05, 'epoch': 0.78}) +[2026-06-17 20:30:41,828][starVLA.training.train_starvla][INFO] - Step 2139, Loss: {'train/loss': 0.6950390934944153, 'train/grad_norm_pre_clip': 16.96487045288086, 'timing/data': 0.0018463507294654846, 'timing/model': 1.0794569021090865, 'learning_rate/qwen_vl_interface': 8.095939304912549e-06, 'learning_rate/action_model': 8.095939304912548e-05, 'epoch': 0.78}) +[2026-06-17 20:30:46,211][starVLA.training.train_starvla][INFO] - Step 2140, Loss: {'train/loss': 0.634013831615448, 'train/grad_norm_pre_clip': 13.845762252807617, 'timing/data': 0.000832737423479557, 'timing/model': 1.419208343140781, 'learning_rate/qwen_vl_interface': 8.094207472419759e-06, 'learning_rate/action_model': 8.094207472419759e-05, 'epoch': 0.78}) +[2026-06-17 20:30:50,780][starVLA.training.train_starvla][INFO] - Step 2141, Loss: {'train/loss': 0.8839463591575623, 'train/grad_norm_pre_clip': 19.471439361572266, 'timing/data': 0.0010599233210086823, 'timing/model': 1.5036919880658388, 'learning_rate/qwen_vl_interface': 8.092475050319699e-06, 'learning_rate/action_model': 8.092475050319698e-05, 'epoch': 0.78}) +[2026-06-17 20:30:58,739][starVLA.training.train_starvla][INFO] - Step 2142, Loss: {'train/loss': 0.8901836276054382, 'train/grad_norm_pre_clip': 15.349727630615234, 'timing/data': 0.0007907506078481674, 'timing/model': 4.4986590491607785, 'learning_rate/qwen_vl_interface': 8.090742038971503e-06, 'learning_rate/action_model': 8.090742038971503e-05, 'epoch': 0.78}) +[2026-06-17 20:31:03,606][starVLA.training.train_starvla][INFO] - Step 2143, Loss: {'train/loss': 0.6732602119445801, 'train/grad_norm_pre_clip': 15.05815315246582, 'timing/data': 0.0019063865765929222, 'timing/model': 1.26627445127815, 'learning_rate/qwen_vl_interface': 8.089008438734427e-06, 'learning_rate/action_model': 8.089008438734427e-05, 'epoch': 0.78}) +[2026-06-17 20:31:08,454][starVLA.training.train_starvla][INFO] - Step 2144, Loss: {'train/loss': 0.7149603366851807, 'train/grad_norm_pre_clip': 14.209904670715332, 'timing/data': 0.0012269308790564537, 'timing/model': 1.0337324934080243, 'learning_rate/qwen_vl_interface': 8.087274249967844e-06, 'learning_rate/action_model': 8.087274249967845e-05, 'epoch': 0.79}) +[2026-06-17 20:31:17,767][starVLA.training.train_starvla][INFO] - Step 2145, Loss: {'train/loss': 0.8393812775611877, 'train/grad_norm_pre_clip': 18.017366409301758, 'timing/data': 0.0013171043246984482, 'timing/model': 1.1251250533387065, 'learning_rate/qwen_vl_interface': 8.08553947303126e-06, 'learning_rate/action_model': 8.085539473031259e-05, 'epoch': 0.79}) +[2026-06-17 20:31:22,536][starVLA.training.train_starvla][INFO] - Step 2146, Loss: {'train/loss': 0.5999075770378113, 'train/grad_norm_pre_clip': 11.868304252624512, 'timing/data': 0.0008608987554907799, 'timing/model': 1.4306706357747316, 'learning_rate/qwen_vl_interface': 8.083804108284288e-06, 'learning_rate/action_model': 8.083804108284287e-05, 'epoch': 0.79}) +[2026-06-17 20:31:32,211][starVLA.training.train_starvla][INFO] - Step 2147, Loss: {'train/loss': 0.7572610974311829, 'train/grad_norm_pre_clip': 16.219741821289062, 'timing/data': 0.0010106069967150688, 'timing/model': 1.1488895555958152, 'learning_rate/qwen_vl_interface': 8.082068156086675e-06, 'learning_rate/action_model': 8.082068156086675e-05, 'epoch': 0.79}) +[2026-06-17 20:31:37,005][starVLA.training.train_starvla][INFO] - Step 2148, Loss: {'train/loss': 0.7210990190505981, 'train/grad_norm_pre_clip': 15.409102439880371, 'timing/data': 0.001063082367181778, 'timing/model': 1.3161289626732469, 'learning_rate/qwen_vl_interface': 8.080331616798289e-06, 'learning_rate/action_model': 8.080331616798287e-05, 'epoch': 0.79}) +[2026-06-17 20:31:41,826][starVLA.training.train_starvla][INFO] - Step 2149, Loss: {'train/loss': 0.6859143376350403, 'train/grad_norm_pre_clip': 23.285663604736328, 'timing/data': 0.0015526339411735535, 'timing/model': 1.0921449372544885, 'learning_rate/qwen_vl_interface': 8.078594490779112e-06, 'learning_rate/action_model': 8.078594490779111e-05, 'epoch': 0.79}) +[2026-06-17 20:31:46,547][starVLA.training.train_starvla][INFO] - Step 2150, Loss: {'train/loss': 0.7301638126373291, 'train/grad_norm_pre_clip': 16.08550262451172, 'timing/data': 0.0010483190417289734, 'timing/model': 1.3704524971544743, 'learning_rate/qwen_vl_interface': 8.076856778389253e-06, 'learning_rate/action_model': 8.076856778389252e-05, 'epoch': 0.79}) +[2026-06-17 20:31:52,359][starVLA.training.train_starvla][INFO] - Step 2151, Loss: {'train/loss': 0.7061419486999512, 'train/grad_norm_pre_clip': 16.03599739074707, 'timing/data': 0.0017759660258889198, 'timing/model': 1.3848250275477767, 'learning_rate/qwen_vl_interface': 8.075118479988942e-06, 'learning_rate/action_model': 8.075118479988941e-05, 'epoch': 0.79}) +[2026-06-17 20:31:56,722][starVLA.training.train_starvla][INFO] - Step 2152, Loss: {'train/loss': 0.6765465140342712, 'train/grad_norm_pre_clip': 13.81844711303711, 'timing/data': 0.0009887097403407097, 'timing/model': 1.0333469929173589, 'learning_rate/qwen_vl_interface': 8.07337959593853e-06, 'learning_rate/action_model': 8.073379595938529e-05, 'epoch': 0.79}) +[2026-06-17 20:32:01,669][starVLA.training.train_starvla][INFO] - Step 2153, Loss: {'train/loss': 0.6811169981956482, 'train/grad_norm_pre_clip': 13.197993278503418, 'timing/data': 0.0021516168490052223, 'timing/model': 1.182715562172234, 'learning_rate/qwen_vl_interface': 8.071640126598488e-06, 'learning_rate/action_model': 8.071640126598488e-05, 'epoch': 0.79}) +[2026-06-17 20:32:06,205][starVLA.training.train_starvla][INFO] - Step 2154, Loss: {'train/loss': 0.733927845954895, 'train/grad_norm_pre_clip': 15.82205867767334, 'timing/data': 0.0010437723249197006, 'timing/model': 1.1152803162112832, 'learning_rate/qwen_vl_interface': 8.069900072329414e-06, 'learning_rate/action_model': 8.069900072329415e-05, 'epoch': 0.79}) +[2026-06-17 20:32:10,682][starVLA.training.train_starvla][INFO] - Step 2155, Loss: {'train/loss': 0.954018771648407, 'train/grad_norm_pre_clip': 17.055225372314453, 'timing/data': 0.0009485911577939987, 'timing/model': 1.0408082837238908, 'learning_rate/qwen_vl_interface': 8.068159433492021e-06, 'learning_rate/action_model': 8.068159433492021e-05, 'epoch': 0.79}) +[2026-06-17 20:32:15,179][starVLA.training.train_starvla][INFO] - Step 2156, Loss: {'train/loss': 0.7937315106391907, 'train/grad_norm_pre_clip': 16.988412857055664, 'timing/data': 0.0010888781398534775, 'timing/model': 1.3989871637895703, 'learning_rate/qwen_vl_interface': 8.066418210447144e-06, 'learning_rate/action_model': 8.066418210447143e-05, 'epoch': 0.79}) +[2026-06-17 20:32:23,566][starVLA.training.train_starvla][INFO] - Step 2157, Loss: {'train/loss': 0.7774134874343872, 'train/grad_norm_pre_clip': 16.727890014648438, 'timing/data': 0.000860174186527729, 'timing/model': 3.951529396697879, 'learning_rate/qwen_vl_interface': 8.064676403555741e-06, 'learning_rate/action_model': 8.064676403555741e-05, 'epoch': 0.79}) +[2026-06-17 20:32:35,441][starVLA.training.train_starvla][INFO] - Step 2158, Loss: {'train/loss': 0.44036298990249634, 'train/grad_norm_pre_clip': 20.042509078979492, 'timing/data': 0.01102594193071127, 'timing/model': 1.0769044188782573, 'learning_rate/qwen_vl_interface': 8.06293401317889e-06, 'learning_rate/action_model': 8.062934013178891e-05, 'epoch': 0.79}) +[2026-06-17 20:32:39,857][starVLA.training.train_starvla][INFO] - Step 2159, Loss: {'train/loss': 0.5527464151382446, 'train/grad_norm_pre_clip': 12.823023796081543, 'timing/data': 0.012902540154755116, 'timing/model': 1.0413351757451892, 'learning_rate/qwen_vl_interface': 8.061191039677793e-06, 'learning_rate/action_model': 8.061191039677793e-05, 'epoch': 0.79}) +[2026-06-17 20:32:44,275][starVLA.training.train_starvla][INFO] - Step 2160, Loss: {'train/loss': 0.83982253074646, 'train/grad_norm_pre_clip': 13.769268035888672, 'timing/data': 0.013259679079055786, 'timing/model': 1.0340354843065143, 'learning_rate/qwen_vl_interface': 8.059447483413768e-06, 'learning_rate/action_model': 8.059447483413767e-05, 'epoch': 0.79}) +[2026-06-17 20:32:49,007][starVLA.training.train_starvla][INFO] - Step 2161, Loss: {'train/loss': 0.6016286611557007, 'train/grad_norm_pre_clip': 13.438360214233398, 'timing/data': 0.012273835949599743, 'timing/model': 1.4458675486966968, 'learning_rate/qwen_vl_interface': 8.057703344748258e-06, 'learning_rate/action_model': 8.057703344748257e-05, 'epoch': 0.79}) +[2026-06-17 20:32:53,115][starVLA.training.train_starvla][INFO] - Step 2162, Loss: {'train/loss': 0.7626926302909851, 'train/grad_norm_pre_clip': 16.835590362548828, 'timing/data': 0.011223710142076015, 'timing/model': 1.0979114221408963, 'learning_rate/qwen_vl_interface': 8.05595862404282e-06, 'learning_rate/action_model': 8.05595862404282e-05, 'epoch': 0.79}) +[2026-06-17 20:32:57,759][starVLA.training.train_starvla][INFO] - Step 2163, Loss: {'train/loss': 0.956590473651886, 'train/grad_norm_pre_clip': 41.90517044067383, 'timing/data': 0.011325472965836525, 'timing/model': 1.0915036527439952, 'learning_rate/qwen_vl_interface': 8.054213321659143e-06, 'learning_rate/action_model': 8.054213321659142e-05, 'epoch': 0.79}) +[2026-06-17 20:33:02,426][starVLA.training.train_starvla][INFO] - Step 2164, Loss: {'train/loss': 0.5969350934028625, 'train/grad_norm_pre_clip': 11.374788284301758, 'timing/data': 0.012018523178994656, 'timing/model': 1.1109042409807444, 'learning_rate/qwen_vl_interface': 8.052467437959025e-06, 'learning_rate/action_model': 8.052467437959025e-05, 'epoch': 0.79}) +[2026-06-17 20:33:06,893][starVLA.training.train_starvla][INFO] - Step 2165, Loss: {'train/loss': 0.7643327713012695, 'train/grad_norm_pre_clip': 13.509321212768555, 'timing/data': 0.012053024023771286, 'timing/model': 1.047649729065597, 'learning_rate/qwen_vl_interface': 8.050720973304393e-06, 'learning_rate/action_model': 8.050720973304391e-05, 'epoch': 0.79}) +[2026-06-17 20:33:11,312][starVLA.training.train_starvla][INFO] - Step 2166, Loss: {'train/loss': 0.7539157271385193, 'train/grad_norm_pre_clip': 13.920809745788574, 'timing/data': 0.02000776119530201, 'timing/model': 1.0877723675221205, 'learning_rate/qwen_vl_interface': 8.048973928057288e-06, 'learning_rate/action_model': 8.048973928057288e-05, 'epoch': 0.79}) +[2026-06-17 20:33:15,882][starVLA.training.train_starvla][INFO] - Step 2167, Loss: {'train/loss': 0.8987517356872559, 'train/grad_norm_pre_clip': 34.399288177490234, 'timing/data': 0.015727649442851543, 'timing/model': 1.0587582532316446, 'learning_rate/qwen_vl_interface': 8.047226302579877e-06, 'learning_rate/action_model': 8.047226302579877e-05, 'epoch': 0.79}) +[2026-06-17 20:33:20,488][starVLA.training.train_starvla][INFO] - Step 2168, Loss: {'train/loss': 0.9572034478187561, 'train/grad_norm_pre_clip': 16.242389678955078, 'timing/data': 0.010567392222583294, 'timing/model': 1.025271750986576, 'learning_rate/qwen_vl_interface': 8.045478097234443e-06, 'learning_rate/action_model': 8.045478097234444e-05, 'epoch': 0.79}) +[2026-06-17 20:33:28,051][starVLA.training.train_starvla][INFO] - Step 2169, Loss: {'train/loss': 0.7324715256690979, 'train/grad_norm_pre_clip': 18.13459587097168, 'timing/data': 0.01292339526116848, 'timing/model': 4.145011834800243, 'learning_rate/qwen_vl_interface': 8.043729312383393e-06, 'learning_rate/action_model': 8.043729312383392e-05, 'epoch': 0.79}) +[2026-06-17 20:33:33,883][starVLA.training.train_starvla][INFO] - Step 2170, Loss: {'train/loss': 0.7070439457893372, 'train/grad_norm_pre_clip': 11.755163192749023, 'timing/data': 0.01467933040112257, 'timing/model': 1.4655790803954005, 'learning_rate/qwen_vl_interface': 8.041979948389248e-06, 'learning_rate/action_model': 8.041979948389248e-05, 'epoch': 0.79}) +[2026-06-17 20:33:38,378][starVLA.training.train_starvla][INFO] - Step 2171, Loss: {'train/loss': 0.9843831062316895, 'train/grad_norm_pre_clip': 15.13992691040039, 'timing/data': 0.015907224267721176, 'timing/model': 1.3280931087210774, 'learning_rate/qwen_vl_interface': 8.040230005614656e-06, 'learning_rate/action_model': 8.040230005614656e-05, 'epoch': 0.79}) +[2026-06-17 20:33:42,455][starVLA.training.train_starvla][INFO] - Step 2172, Loss: {'train/loss': 0.8225009441375732, 'train/grad_norm_pre_clip': 16.813770294189453, 'timing/data': 0.010727502405643463, 'timing/model': 1.0224141189828515, 'learning_rate/qwen_vl_interface': 8.038479484422383e-06, 'learning_rate/action_model': 8.038479484422383e-05, 'epoch': 0.8}) +[2026-06-17 20:33:46,896][starVLA.training.train_starvla][INFO] - Step 2173, Loss: {'train/loss': 0.8201123476028442, 'train/grad_norm_pre_clip': 13.024637222290039, 'timing/data': 0.010950990952551365, 'timing/model': 1.026269169524312, 'learning_rate/qwen_vl_interface': 8.03672838517531e-06, 'learning_rate/action_model': 8.036728385175311e-05, 'epoch': 0.8}) +[2026-06-17 20:33:51,578][starVLA.training.train_starvla][INFO] - Step 2174, Loss: {'train/loss': 0.6502194404602051, 'train/grad_norm_pre_clip': 11.772605895996094, 'timing/data': 0.019429834559559822, 'timing/model': 1.0606207959353924, 'learning_rate/qwen_vl_interface': 8.034976708236446e-06, 'learning_rate/action_model': 8.034976708236445e-05, 'epoch': 0.8}) +[2026-06-17 20:33:56,118][starVLA.training.train_starvla][INFO] - Step 2175, Loss: {'train/loss': 0.7668182253837585, 'train/grad_norm_pre_clip': 15.624874114990234, 'timing/data': 0.011514822021126747, 'timing/model': 1.1267718616873026, 'learning_rate/qwen_vl_interface': 8.033224453968913e-06, 'learning_rate/action_model': 8.033224453968913e-05, 'epoch': 0.8}) +[2026-06-17 20:34:01,159][starVLA.training.train_starvla][INFO] - Step 2176, Loss: {'train/loss': 0.9030041694641113, 'train/grad_norm_pre_clip': 14.020905494689941, 'timing/data': 0.011274359188973904, 'timing/model': 1.3148870142176747, 'learning_rate/qwen_vl_interface': 8.031471622735956e-06, 'learning_rate/action_model': 8.031471622735956e-05, 'epoch': 0.8}) +[2026-06-17 20:34:05,700][starVLA.training.train_starvla][INFO] - Step 2177, Loss: {'train/loss': 1.0376534461975098, 'train/grad_norm_pre_clip': 15.261929512023926, 'timing/data': 0.0171198146417737, 'timing/model': 1.110535155981779, 'learning_rate/qwen_vl_interface': 8.029718214900937e-06, 'learning_rate/action_model': 8.029718214900937e-05, 'epoch': 0.8}) +[2026-06-17 20:34:10,487][starVLA.training.train_starvla][INFO] - Step 2178, Loss: {'train/loss': 1.1505664587020874, 'train/grad_norm_pre_clip': 16.9392147064209, 'timing/data': 0.010799874551594257, 'timing/model': 1.1717279758304358, 'learning_rate/qwen_vl_interface': 8.027964230827342e-06, 'learning_rate/action_model': 8.027964230827342e-05, 'epoch': 0.8}) +[2026-06-17 20:34:14,942][starVLA.training.train_starvla][INFO] - Step 2179, Loss: {'train/loss': 1.0151034593582153, 'train/grad_norm_pre_clip': 16.30404281616211, 'timing/data': 0.0156136779114604, 'timing/model': 1.0791512252762914, 'learning_rate/qwen_vl_interface': 8.026209670878776e-06, 'learning_rate/action_model': 8.026209670878775e-05, 'epoch': 0.8}) +[2026-06-17 20:34:24,465][starVLA.training.train_starvla][INFO] - Step 2180, Loss: {'train/loss': 0.83646559715271, 'train/grad_norm_pre_clip': 22.20747947692871, 'timing/data': 0.013253224082291126, 'timing/model': 1.3922046450898051, 'learning_rate/qwen_vl_interface': 8.024454535418952e-06, 'learning_rate/action_model': 8.024454535418951e-05, 'epoch': 0.8}) +[2026-06-17 20:34:28,996][starVLA.training.train_starvla][INFO] - Step 2181, Loss: {'train/loss': 0.6101812124252319, 'train/grad_norm_pre_clip': 12.079747200012207, 'timing/data': 0.011985259130597115, 'timing/model': 1.369280501268804, 'learning_rate/qwen_vl_interface': 8.02269882481172e-06, 'learning_rate/action_model': 8.022698824811721e-05, 'epoch': 0.8}) +[2026-06-17 20:34:38,108][starVLA.training.train_starvla][INFO] - Step 2182, Loss: {'train/loss': 0.6295558214187622, 'train/grad_norm_pre_clip': 12.45712661743164, 'timing/data': 0.010985538363456726, 'timing/model': 1.292472092434764, 'learning_rate/qwen_vl_interface': 8.02094253942104e-06, 'learning_rate/action_model': 8.02094253942104e-05, 'epoch': 0.8}) +[2026-06-17 20:34:42,527][starVLA.training.train_starvla][INFO] - Step 2183, Loss: {'train/loss': 0.7970526814460754, 'train/grad_norm_pre_clip': 13.221359252929688, 'timing/data': 0.012275723740458488, 'timing/model': 1.1079855151474476, 'learning_rate/qwen_vl_interface': 8.019185679610988e-06, 'learning_rate/action_model': 8.019185679610988e-05, 'epoch': 0.8}) +[2026-06-17 20:34:47,595][starVLA.training.train_starvla][INFO] - Step 2184, Loss: {'train/loss': 0.5091837644577026, 'train/grad_norm_pre_clip': 20.71551513671875, 'timing/data': 0.020706776529550552, 'timing/model': 1.0667043067514896, 'learning_rate/qwen_vl_interface': 8.017428245745765e-06, 'learning_rate/action_model': 8.017428245745763e-05, 'epoch': 0.8}) +[2026-06-17 20:34:59,144][starVLA.training.train_starvla][INFO] - Step 2185, Loss: {'train/loss': 0.6646293997764587, 'train/grad_norm_pre_clip': 12.684103012084961, 'timing/data': 0.011424871161580086, 'timing/model': 1.0570321679115295, 'learning_rate/qwen_vl_interface': 8.015670238189688e-06, 'learning_rate/action_model': 8.015670238189688e-05, 'epoch': 0.8}) +[2026-06-17 20:35:03,826][starVLA.training.train_starvla][INFO] - Step 2186, Loss: {'train/loss': 0.49104785919189453, 'train/grad_norm_pre_clip': 13.24801254272461, 'timing/data': 0.012013955973088741, 'timing/model': 1.449440885335207, 'learning_rate/qwen_vl_interface': 8.013911657307196e-06, 'learning_rate/action_model': 8.013911657307196e-05, 'epoch': 0.8}) +[2026-06-17 20:35:08,017][starVLA.training.train_starvla][INFO] - Step 2187, Loss: {'train/loss': 0.743750810623169, 'train/grad_norm_pre_clip': 16.307634353637695, 'timing/data': 0.013369797728955746, 'timing/model': 1.0998875675722957, 'learning_rate/qwen_vl_interface': 8.012152503462843e-06, 'learning_rate/action_model': 8.012152503462842e-05, 'epoch': 0.8}) +[2026-06-17 20:35:12,783][starVLA.training.train_starvla][INFO] - Step 2188, Loss: {'train/loss': 0.4782223403453827, 'train/grad_norm_pre_clip': 10.56068229675293, 'timing/data': 0.04713159892708063, 'timing/model': 1.081532808020711, 'learning_rate/qwen_vl_interface': 8.010392777021304e-06, 'learning_rate/action_model': 8.010392777021303e-05, 'epoch': 0.8}) +[2026-06-17 20:35:21,422][starVLA.training.train_starvla][INFO] - Step 2189, Loss: {'train/loss': 0.46007421612739563, 'train/grad_norm_pre_clip': 19.936939239501953, 'timing/data': 0.01463715173304081, 'timing/model': 1.3464271938428283, 'learning_rate/qwen_vl_interface': 8.008632478347372e-06, 'learning_rate/action_model': 8.008632478347371e-05, 'epoch': 0.8}) +[2026-06-17 20:35:27,595][starVLA.training.train_starvla][INFO] - Step 2190, Loss: {'train/loss': 0.6740403771400452, 'train/grad_norm_pre_clip': 15.274767875671387, 'timing/data': 0.016079282388091087, 'timing/model': 2.8292464818805456, 'learning_rate/qwen_vl_interface': 8.006871607805957e-06, 'learning_rate/action_model': 8.006871607805957e-05, 'epoch': 0.8}) +[2026-06-17 20:35:32,623][starVLA.training.train_starvla][INFO] - Step 2191, Loss: {'train/loss': 0.6696691513061523, 'train/grad_norm_pre_clip': 15.660011291503906, 'timing/data': 0.011778876185417175, 'timing/model': 1.5926657021045685, 'learning_rate/qwen_vl_interface': 8.005110165762093e-06, 'learning_rate/action_model': 8.005110165762093e-05, 'epoch': 0.8}) +[2026-06-17 20:35:36,752][starVLA.training.train_starvla][INFO] - Step 2192, Loss: {'train/loss': 0.7199481129646301, 'train/grad_norm_pre_clip': 13.355164527893066, 'timing/data': 0.011122994124889374, 'timing/model': 1.0381044950336218, 'learning_rate/qwen_vl_interface': 8.003348152580927e-06, 'learning_rate/action_model': 8.003348152580927e-05, 'epoch': 0.8}) +[2026-06-17 20:35:41,378][starVLA.training.train_starvla][INFO] - Step 2193, Loss: {'train/loss': 0.7731963396072388, 'train/grad_norm_pre_clip': 14.116292953491211, 'timing/data': 0.012676656246185303, 'timing/model': 1.0305587658658624, 'learning_rate/qwen_vl_interface': 8.001585568627724e-06, 'learning_rate/action_model': 8.001585568627725e-05, 'epoch': 0.8}) +[2026-06-17 20:35:46,002][starVLA.training.train_starvla][INFO] - Step 2194, Loss: {'train/loss': 0.744935154914856, 'train/grad_norm_pre_clip': 17.451475143432617, 'timing/data': 0.010943968780338764, 'timing/model': 1.027303146198392, 'learning_rate/qwen_vl_interface': 7.999822414267874e-06, 'learning_rate/action_model': 7.999822414267874e-05, 'epoch': 0.8}) +[2026-06-17 20:35:50,499][starVLA.training.train_starvla][INFO] - Step 2195, Loss: {'train/loss': 0.8822910785675049, 'train/grad_norm_pre_clip': 18.009214401245117, 'timing/data': 0.014355954714119434, 'timing/model': 1.085730087943375, 'learning_rate/qwen_vl_interface': 7.998058689866879e-06, 'learning_rate/action_model': 7.998058689866879e-05, 'epoch': 0.8}) +[2026-06-17 20:35:54,975][starVLA.training.train_starvla][INFO] - Step 2196, Loss: {'train/loss': 0.8284125924110413, 'train/grad_norm_pre_clip': 17.510263442993164, 'timing/data': 0.02081172540783882, 'timing/model': 1.3655554568395019, 'learning_rate/qwen_vl_interface': 7.99629439579036e-06, 'learning_rate/action_model': 7.99629439579036e-05, 'epoch': 0.8}) +[2026-06-17 20:35:59,255][starVLA.training.train_starvla][INFO] - Step 2197, Loss: {'train/loss': 0.7810466289520264, 'train/grad_norm_pre_clip': 14.594762802124023, 'timing/data': 0.011176745407283306, 'timing/model': 1.0507121905684471, 'learning_rate/qwen_vl_interface': 7.994529532404059e-06, 'learning_rate/action_model': 7.994529532404058e-05, 'epoch': 0.8}) +[2026-06-17 20:36:04,293][starVLA.training.train_starvla][INFO] - Step 2198, Loss: {'train/loss': 0.850071370601654, 'train/grad_norm_pre_clip': 14.626555442810059, 'timing/data': 0.011803330853581429, 'timing/model': 1.241923050954938, 'learning_rate/qwen_vl_interface': 7.99276410007383e-06, 'learning_rate/action_model': 7.99276410007383e-05, 'epoch': 0.8}) +[2026-06-17 20:36:08,999][starVLA.training.train_starvla][INFO] - Step 2199, Loss: {'train/loss': 0.861034631729126, 'train/grad_norm_pre_clip': 14.811626434326172, 'timing/data': 0.012383220717310905, 'timing/model': 1.098711283877492, 'learning_rate/qwen_vl_interface': 7.990998099165652e-06, 'learning_rate/action_model': 7.990998099165652e-05, 'epoch': 0.81}) +[2026-06-17 20:39:44,739][starVLA.training.train_starvla][INFO] - Step 2200, Loss: {'train/loss': 0.6221818923950195, 'train/grad_norm_pre_clip': 12.367345809936523, 'eval/loss': 0.7287123799324036, 'eval/latency_2/loss': 0.7287123799324036, 'eval/demon_fire_precision': 0.9162829635830891, 'eval/demon_fire_recall': 0.9158995815899582, 'eval/demon_fire_f1': 0.9160912324754134, 'eval/demon_fire_rate_model': 0.6647189760712299, 'eval/demon_fire_rate_teacher': 0.6649972175848636, 'eval/demon_left_precision': 0.7814702920443102, 'eval/demon_left_recall': 0.8168421052631579, 'eval/demon_left_f1': 0.7987647967061245, 'eval/demon_left_rate_model': 0.27629382303839733, 'eval/demon_left_rate_teacher': 0.2643294379521425, 'eval/demon_right_precision': 0.7908704883227177, 'eval/demon_right_recall': 0.8418079096045198, 'eval/demon_right_f1': 0.815544608648057, 'eval/demon_right_rate_model': 0.2621035058430718, 'eval/demon_right_rate_teacher': 0.24624373956594323, 'eval/demon_move_f1': 0.8071547026770908, 'eval/demon_cc_f1': 0.8804429011152834, 'eval/latency_2/demon_fire_precision': 0.9162829635830891, 'eval/latency_2/demon_fire_recall': 0.9158995815899582, 'eval/latency_2/demon_fire_f1': 0.9160912324754134, 'eval/latency_2/demon_fire_rate_model': 0.6647189760712299, 'eval/latency_2/demon_fire_rate_teacher': 0.6649972175848636, 'eval/latency_2/demon_left_precision': 0.7814702920443102, 'eval/latency_2/demon_left_recall': 0.8168421052631579, 'eval/latency_2/demon_left_f1': 0.7987647967061245, 'eval/latency_2/demon_left_rate_model': 0.27629382303839733, 'eval/latency_2/demon_left_rate_teacher': 0.2643294379521425, 'eval/latency_2/demon_right_precision': 0.7908704883227177, 'eval/latency_2/demon_right_recall': 0.8418079096045198, 'eval/latency_2/demon_right_f1': 0.815544608648057, 'eval/latency_2/demon_right_rate_model': 0.2621035058430718, 'eval/latency_2/demon_right_rate_teacher': 0.24624373956594323, 'eval/latency_2/demon_move_f1': 0.8071547026770908, 'eval/latency_2/demon_cc_f1': 0.8804429011152834, 'timing/data': 0.01165843103080988, 'timing/model': 1.150035372003913, 'learning_rate/qwen_vl_interface': 7.98923153004562e-06, 'learning_rate/action_model': 7.98923153004562e-05, 'epoch': 0.81}) +[2026-06-17 20:39:44,741][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_2200_state +[2026-06-17 20:39:44,748][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 20:39:53,232][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_2200_state/model.safetensors +[2026-06-17 20:40:08,065][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_2200_state/optimizer.bin +[2026-06-17 20:40:08,066][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_2200_state/sampler.bin +[2026-06-17 20:40:08,067][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_2200_state/sampler_1.bin +[2026-06-17 20:40:08,069][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_2200_state/random_states_0.pkl +[2026-06-17 20:40:08,074][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2200 +[2026-06-17 20:40:08,074][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 20:40:08,091][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 20:40:12,758][starVLA.training.train_starvla][INFO] - Step 2201, Loss: {'train/loss': 0.5684821605682373, 'train/grad_norm_pre_clip': 12.894694328308105, 'timing/data': 0.012602729722857475, 'timing/model': 1.1062399968504906, 'learning_rate/qwen_vl_interface': 7.987464393079943e-06, 'learning_rate/action_model': 7.987464393079943e-05, 'epoch': 0.81}) +[2026-06-17 20:40:18,257][starVLA.training.train_starvla][INFO] - Step 2202, Loss: {'train/loss': 0.7569907307624817, 'train/grad_norm_pre_clip': 17.76835060119629, 'timing/data': 0.017000422812998295, 'timing/model': 1.7004072442650795, 'learning_rate/qwen_vl_interface': 7.985696688634953e-06, 'learning_rate/action_model': 7.985696688634952e-05, 'epoch': 0.81}) +[2026-06-17 20:40:29,819][starVLA.training.train_starvla][INFO] - Step 2203, Loss: {'train/loss': 0.4967420697212219, 'train/grad_norm_pre_clip': 10.45032024383545, 'timing/data': 0.018030624836683273, 'timing/model': 2.544830727390945, 'learning_rate/qwen_vl_interface': 7.983928417077095e-06, 'learning_rate/action_model': 7.983928417077095e-05, 'epoch': 0.81}) +[2026-06-17 20:40:34,278][starVLA.training.train_starvla][INFO] - Step 2204, Loss: {'train/loss': 1.1410707235336304, 'train/grad_norm_pre_clip': 19.651308059692383, 'timing/data': 0.013554339297115803, 'timing/model': 1.0724506489932537, 'learning_rate/qwen_vl_interface': 7.982159578772935e-06, 'learning_rate/action_model': 7.982159578772934e-05, 'epoch': 0.81}) +[2026-06-17 20:40:39,183][starVLA.training.train_starvla][INFO] - Step 2205, Loss: {'train/loss': 0.5882595181465149, 'train/grad_norm_pre_clip': 11.795143127441406, 'timing/data': 0.010653120465576649, 'timing/model': 1.0422433484345675, 'learning_rate/qwen_vl_interface': 7.980390174089154e-06, 'learning_rate/action_model': 7.980390174089153e-05, 'epoch': 0.81}) +[2026-06-17 20:40:43,793][starVLA.training.train_starvla][INFO] - Step 2206, Loss: {'train/loss': 0.6646245718002319, 'train/grad_norm_pre_clip': 16.73076820373535, 'timing/data': 0.015227860771119595, 'timing/model': 1.1356144258752465, 'learning_rate/qwen_vl_interface': 7.978620203392551e-06, 'learning_rate/action_model': 7.97862020339255e-05, 'epoch': 0.81}) +[2026-06-17 20:40:48,468][starVLA.training.train_starvla][INFO] - Step 2207, Loss: {'train/loss': 0.7256847023963928, 'train/grad_norm_pre_clip': 13.349991798400879, 'timing/data': 0.011818945407867432, 'timing/model': 1.4357622526586056, 'learning_rate/qwen_vl_interface': 7.976849667050044e-06, 'learning_rate/action_model': 7.976849667050044e-05, 'epoch': 0.81}) +[2026-06-17 20:40:55,827][starVLA.training.train_starvla][INFO] - Step 2208, Loss: {'train/loss': 0.9599945545196533, 'train/grad_norm_pre_clip': 14.852310180664062, 'timing/data': 0.014151942916214466, 'timing/model': 4.1942896619439125, 'learning_rate/qwen_vl_interface': 7.975078565428667e-06, 'learning_rate/action_model': 7.975078565428668e-05, 'epoch': 0.81}) +[2026-06-17 20:41:00,862][starVLA.training.train_starvla][INFO] - Step 2209, Loss: {'train/loss': 0.6527765989303589, 'train/grad_norm_pre_clip': 12.79326057434082, 'timing/data': 0.01200153585523367, 'timing/model': 1.071977692656219, 'learning_rate/qwen_vl_interface': 7.97330689889557e-06, 'learning_rate/action_model': 7.97330689889557e-05, 'epoch': 0.81}) +[2026-06-17 20:41:05,643][starVLA.training.train_starvla][INFO] - Step 2210, Loss: {'train/loss': 0.9477414488792419, 'train/grad_norm_pre_clip': 13.38537883758545, 'timing/data': 0.0215515224263072, 'timing/model': 1.0292688347399235, 'learning_rate/qwen_vl_interface': 7.971534667818026e-06, 'learning_rate/action_model': 7.971534667818025e-05, 'epoch': 0.81}) +[2026-06-17 20:41:10,195][starVLA.training.train_starvla][INFO] - Step 2211, Loss: {'train/loss': 0.5405131578445435, 'train/grad_norm_pre_clip': 11.49637508392334, 'timing/data': 0.011731524951756, 'timing/model': 1.1441756524145603, 'learning_rate/qwen_vl_interface': 7.969761872563413e-06, 'learning_rate/action_model': 7.969761872563413e-05, 'epoch': 0.81}) +[2026-06-17 20:41:21,215][starVLA.training.train_starvla][INFO] - Step 2212, Loss: {'train/loss': 1.0205875635147095, 'train/grad_norm_pre_clip': 14.69249439239502, 'timing/data': 0.011220327578485012, 'timing/model': 4.641002133488655, 'learning_rate/qwen_vl_interface': 7.96798851349924e-06, 'learning_rate/action_model': 7.96798851349924e-05, 'epoch': 0.81}) +[2026-06-17 20:41:32,329][starVLA.training.train_starvla][INFO] - Step 2213, Loss: {'train/loss': 0.7454692721366882, 'train/grad_norm_pre_clip': 16.84851837158203, 'timing/data': 0.013226481154561043, 'timing/model': 1.205310307443142, 'learning_rate/qwen_vl_interface': 7.966214590993122e-06, 'learning_rate/action_model': 7.966214590993122e-05, 'epoch': 0.81}) +[2026-06-17 20:41:44,223][starVLA.training.train_starvla][INFO] - Step 2214, Loss: {'train/loss': 0.8566264510154724, 'train/grad_norm_pre_clip': 17.07354736328125, 'timing/data': 0.013515494763851166, 'timing/model': 1.1981368819251657, 'learning_rate/qwen_vl_interface': 7.964440105412797e-06, 'learning_rate/action_model': 7.964440105412796e-05, 'epoch': 0.81}) +[2026-06-17 20:41:54,827][starVLA.training.train_starvla][INFO] - Step 2215, Loss: {'train/loss': 0.8248327374458313, 'train/grad_norm_pre_clip': 15.602682113647461, 'timing/data': 0.011619790457189083, 'timing/model': 1.0536567876115441, 'learning_rate/qwen_vl_interface': 7.962665057126116e-06, 'learning_rate/action_model': 7.962665057126117e-05, 'epoch': 0.81}) +[2026-06-17 20:42:04,209][starVLA.training.train_starvla][INFO] - Step 2216, Loss: {'train/loss': 0.683803379535675, 'train/grad_norm_pre_clip': 13.300626754760742, 'timing/data': 0.011333686299622059, 'timing/model': 1.7363599482923746, 'learning_rate/qwen_vl_interface': 7.960889446501052e-06, 'learning_rate/action_model': 7.960889446501052e-05, 'epoch': 0.81}) +[2026-06-17 20:42:09,768][starVLA.training.train_starvla][INFO] - Step 2217, Loss: {'train/loss': 0.6289726495742798, 'train/grad_norm_pre_clip': 13.429178237915039, 'timing/data': 0.011059194803237915, 'timing/model': 1.5985838128253818, 'learning_rate/qwen_vl_interface': 7.959113273905688e-06, 'learning_rate/action_model': 7.959113273905688e-05, 'epoch': 0.81}) +[2026-06-17 20:42:21,616][starVLA.training.train_starvla][INFO] - Step 2218, Loss: {'train/loss': 0.6376397609710693, 'train/grad_norm_pre_clip': 12.295395851135254, 'timing/data': 0.016619802452623844, 'timing/model': 2.866689722985029, 'learning_rate/qwen_vl_interface': 7.957336539708226e-06, 'learning_rate/action_model': 7.957336539708225e-05, 'epoch': 0.81}) +[2026-06-17 20:42:31,519][starVLA.training.train_starvla][INFO] - Step 2219, Loss: {'train/loss': 0.6844353079795837, 'train/grad_norm_pre_clip': 15.250703811645508, 'timing/data': 0.01794310472905636, 'timing/model': 5.036423560231924, 'learning_rate/qwen_vl_interface': 7.955559244276987e-06, 'learning_rate/action_model': 7.955559244276987e-05, 'epoch': 0.81}) +[2026-06-17 20:42:41,123][starVLA.training.train_starvla][INFO] - Step 2220, Loss: {'train/loss': 0.8468137383460999, 'train/grad_norm_pre_clip': 12.079276084899902, 'timing/data': 0.012781795114278793, 'timing/model': 1.1072846492752433, 'learning_rate/qwen_vl_interface': 7.953781387980404e-06, 'learning_rate/action_model': 7.953781387980404e-05, 'epoch': 0.81}) +[2026-06-17 20:42:47,363][starVLA.training.train_starvla][INFO] - Step 2221, Loss: {'train/loss': 0.6598838567733765, 'train/grad_norm_pre_clip': 12.717811584472656, 'timing/data': 0.016610734164714813, 'timing/model': 1.0343319047242403, 'learning_rate/qwen_vl_interface': 7.952002971187032e-06, 'learning_rate/action_model': 7.952002971187032e-05, 'epoch': 0.81}) +[2026-06-17 20:42:51,577][starVLA.training.train_starvla][INFO] - Step 2222, Loss: {'train/loss': 0.6294743418693542, 'train/grad_norm_pre_clip': 12.39509105682373, 'timing/data': 0.012157149612903595, 'timing/model': 1.0342311756685376, 'learning_rate/qwen_vl_interface': 7.950223994265534e-06, 'learning_rate/action_model': 7.950223994265533e-05, 'epoch': 0.81}) +[2026-06-17 20:42:56,173][starVLA.training.train_starvla][INFO] - Step 2223, Loss: {'train/loss': 0.7714399099349976, 'train/grad_norm_pre_clip': 13.73373031616211, 'timing/data': 0.015847541391849518, 'timing/model': 1.042147028259933, 'learning_rate/qwen_vl_interface': 7.9484444575847e-06, 'learning_rate/action_model': 7.948444457584699e-05, 'epoch': 0.81}) +[2026-06-17 20:43:00,692][starVLA.training.train_starvla][INFO] - Step 2224, Loss: {'train/loss': 0.8737902045249939, 'train/grad_norm_pre_clip': 14.920083045959473, 'timing/data': 0.01272838469594717, 'timing/model': 1.1165986517444253, 'learning_rate/qwen_vl_interface': 7.946664361513423e-06, 'learning_rate/action_model': 7.946664361513421e-05, 'epoch': 0.81}) +[2026-06-17 20:43:05,137][starVLA.training.train_starvla][INFO] - Step 2225, Loss: {'train/loss': 0.7831318378448486, 'train/grad_norm_pre_clip': 13.019999504089355, 'timing/data': 0.012104258872568607, 'timing/model': 1.0249109771102667, 'learning_rate/qwen_vl_interface': 7.944883706420724e-06, 'learning_rate/action_model': 7.944883706420724e-05, 'epoch': 0.81}) +[2026-06-17 20:43:09,581][starVLA.training.train_starvla][INFO] - Step 2226, Loss: {'train/loss': 0.6422722339630127, 'train/grad_norm_pre_clip': 13.94908332824707, 'timing/data': 0.01155401486903429, 'timing/model': 1.082360478118062, 'learning_rate/qwen_vl_interface': 7.943102492675732e-06, 'learning_rate/action_model': 7.943102492675732e-05, 'epoch': 0.82}) +[2026-06-17 20:43:14,344][starVLA.training.train_starvla][INFO] - Step 2227, Loss: {'train/loss': 0.8966672420501709, 'train/grad_norm_pre_clip': 14.240080833435059, 'timing/data': 0.013189863413572311, 'timing/model': 1.581179739907384, 'learning_rate/qwen_vl_interface': 7.941320720647694e-06, 'learning_rate/action_model': 7.941320720647695e-05, 'epoch': 0.82}) +[2026-06-17 20:43:18,604][starVLA.training.train_starvla][INFO] - Step 2228, Loss: {'train/loss': 0.6433273553848267, 'train/grad_norm_pre_clip': 12.438827514648438, 'timing/data': 0.01380115281790495, 'timing/model': 1.0484233349561691, 'learning_rate/qwen_vl_interface': 7.939538390705976e-06, 'learning_rate/action_model': 7.939538390705976e-05, 'epoch': 0.82}) +[2026-06-17 20:43:23,194][starVLA.training.train_starvla][INFO] - Step 2229, Loss: {'train/loss': 0.8910346627235413, 'train/grad_norm_pre_clip': 12.696438789367676, 'timing/data': 0.012781349010765553, 'timing/model': 1.0427401009947062, 'learning_rate/qwen_vl_interface': 7.937755503220057e-06, 'learning_rate/action_model': 7.937755503220056e-05, 'epoch': 0.82}) +[2026-06-17 20:43:27,606][starVLA.training.train_starvla][INFO] - Step 2230, Loss: {'train/loss': 0.4563300609588623, 'train/grad_norm_pre_clip': 10.136463165283203, 'timing/data': 0.011097889393568039, 'timing/model': 1.089186911471188, 'learning_rate/qwen_vl_interface': 7.93597205855953e-06, 'learning_rate/action_model': 7.935972058559529e-05, 'epoch': 0.82}) +[2026-06-17 20:43:32,201][starVLA.training.train_starvla][INFO] - Step 2231, Loss: {'train/loss': 0.6514063477516174, 'train/grad_norm_pre_clip': 9.134175300598145, 'timing/data': 0.01387572567909956, 'timing/model': 1.122271598316729, 'learning_rate/qwen_vl_interface': 7.934188057094103e-06, 'learning_rate/action_model': 7.934188057094103e-05, 'epoch': 0.82}) +[2026-06-17 20:43:36,544][starVLA.training.train_starvla][INFO] - Step 2232, Loss: {'train/loss': 0.7463945150375366, 'train/grad_norm_pre_clip': 13.534106254577637, 'timing/data': 0.015373656526207924, 'timing/model': 1.0336155574768782, 'learning_rate/qwen_vl_interface': 7.932403499193604e-06, 'learning_rate/action_model': 7.932403499193605e-05, 'epoch': 0.82}) +[2026-06-17 20:43:41,333][starVLA.training.train_starvla][INFO] - Step 2233, Loss: {'train/loss': 0.9853332042694092, 'train/grad_norm_pre_clip': 15.445713996887207, 'timing/data': 0.011074033565819263, 'timing/model': 1.1380807720124722, 'learning_rate/qwen_vl_interface': 7.930618385227975e-06, 'learning_rate/action_model': 7.930618385227973e-05, 'epoch': 0.82}) +[2026-06-17 20:43:45,718][starVLA.training.train_starvla][INFO] - Step 2234, Loss: {'train/loss': 0.43009844422340393, 'train/grad_norm_pre_clip': 11.987021446228027, 'timing/data': 0.011785202659666538, 'timing/model': 1.0317579684779048, 'learning_rate/qwen_vl_interface': 7.928832715567268e-06, 'learning_rate/action_model': 7.928832715567267e-05, 'epoch': 0.82}) +[2026-06-17 20:43:49,905][starVLA.training.train_starvla][INFO] - Step 2235, Loss: {'train/loss': 0.9286928772926331, 'train/grad_norm_pre_clip': 15.617544174194336, 'timing/data': 0.009429309517145157, 'timing/model': 1.0378639129921794, 'learning_rate/qwen_vl_interface': 7.927046490581655e-06, 'learning_rate/action_model': 7.927046490581655e-05, 'epoch': 0.82}) +[2026-06-17 20:43:54,042][starVLA.training.train_starvla][INFO] - Step 2236, Loss: {'train/loss': 0.5142039060592651, 'train/grad_norm_pre_clip': 10.256169319152832, 'timing/data': 0.011739864945411682, 'timing/model': 1.2546538496389985, 'learning_rate/qwen_vl_interface': 7.925259710641428e-06, 'learning_rate/action_model': 7.925259710641427e-05, 'epoch': 0.82}) +[2026-06-17 20:43:59,584][starVLA.training.train_starvla][INFO] - Step 2237, Loss: {'train/loss': 0.6493460536003113, 'train/grad_norm_pre_clip': 12.150156021118164, 'timing/data': 0.010014481842517853, 'timing/model': 2.6020449493080378, 'learning_rate/qwen_vl_interface': 7.923472376116982e-06, 'learning_rate/action_model': 7.923472376116982e-05, 'epoch': 0.82}) +[2026-06-17 20:44:04,564][starVLA.training.train_starvla][INFO] - Step 2238, Loss: {'train/loss': 0.6841410398483276, 'train/grad_norm_pre_clip': 12.345983505249023, 'timing/data': 0.010126232169568539, 'timing/model': 1.0087893195450306, 'learning_rate/qwen_vl_interface': 7.921684487378836e-06, 'learning_rate/action_model': 7.921684487378837e-05, 'epoch': 0.82}) +[2026-06-17 20:44:10,901][starVLA.training.train_starvla][INFO] - Step 2239, Loss: {'train/loss': 0.8219444155693054, 'train/grad_norm_pre_clip': 13.781756401062012, 'timing/data': 0.012387228198349476, 'timing/model': 3.2747665541246533, 'learning_rate/qwen_vl_interface': 7.919896044797622e-06, 'learning_rate/action_model': 7.91989604479762e-05, 'epoch': 0.82}) +[2026-06-17 20:44:14,989][starVLA.training.train_starvla][INFO] - Step 2240, Loss: {'train/loss': 0.6184022426605225, 'train/grad_norm_pre_clip': 12.730843544006348, 'timing/data': 0.010928214527666569, 'timing/model': 1.0071972077712417, 'learning_rate/qwen_vl_interface': 7.918107048744083e-06, 'learning_rate/action_model': 7.918107048744083e-05, 'epoch': 0.82}) +[2026-06-17 20:44:19,273][starVLA.training.train_starvla][INFO] - Step 2241, Loss: {'train/loss': 0.5870122909545898, 'train/grad_norm_pre_clip': 11.493906021118164, 'timing/data': 0.010503467172384262, 'timing/model': 1.0053343595936894, 'learning_rate/qwen_vl_interface': 7.916317499589083e-06, 'learning_rate/action_model': 7.916317499589083e-05, 'epoch': 0.82}) +[2026-06-17 20:44:25,633][starVLA.training.train_starvla][INFO] - Step 2242, Loss: {'train/loss': 0.5645292401313782, 'train/grad_norm_pre_clip': 12.520437240600586, 'timing/data': 0.009644238278269768, 'timing/model': 1.2196627473458648, 'learning_rate/qwen_vl_interface': 7.914527397703596e-06, 'learning_rate/action_model': 7.914527397703595e-05, 'epoch': 0.82}) +[2026-06-17 20:44:30,337][starVLA.training.train_starvla][INFO] - Step 2243, Loss: {'train/loss': 0.5890389084815979, 'train/grad_norm_pre_clip': 12.0232515335083, 'timing/data': 0.01038864441215992, 'timing/model': 0.9997635167092085, 'learning_rate/qwen_vl_interface': 7.912736743458712e-06, 'learning_rate/action_model': 7.912736743458711e-05, 'epoch': 0.82}) +[2026-06-17 20:44:34,419][starVLA.training.train_starvla][INFO] - Step 2244, Loss: {'train/loss': 0.417995810508728, 'train/grad_norm_pre_clip': 9.489258766174316, 'timing/data': 0.010261964052915573, 'timing/model': 1.0044516334310174, 'learning_rate/qwen_vl_interface': 7.910945537225638e-06, 'learning_rate/action_model': 7.910945537225637e-05, 'epoch': 0.82}) +[2026-06-17 20:44:38,807][starVLA.training.train_starvla][INFO] - Step 2245, Loss: {'train/loss': 0.7019230127334595, 'train/grad_norm_pre_clip': 14.187345504760742, 'timing/data': 0.014839490875601768, 'timing/model': 1.206108957529068, 'learning_rate/qwen_vl_interface': 7.909153779375688e-06, 'learning_rate/action_model': 7.909153779375688e-05, 'epoch': 0.82}) +[2026-06-17 20:44:43,380][starVLA.training.train_starvla][INFO] - Step 2246, Loss: {'train/loss': 0.7923613786697388, 'train/grad_norm_pre_clip': 14.005204200744629, 'timing/data': 0.013105207122862339, 'timing/model': 1.0434975335374475, 'learning_rate/qwen_vl_interface': 7.907361470280299e-06, 'learning_rate/action_model': 7.907361470280299e-05, 'epoch': 0.82}) +[2026-06-17 20:44:48,165][starVLA.training.train_starvla][INFO] - Step 2247, Loss: {'train/loss': 0.5622174739837646, 'train/grad_norm_pre_clip': 12.427321434020996, 'timing/data': 0.014874286018311977, 'timing/model': 1.4404396992176771, 'learning_rate/qwen_vl_interface': 7.905568610311018e-06, 'learning_rate/action_model': 7.905568610311017e-05, 'epoch': 0.82}) +[2026-06-17 20:44:53,016][starVLA.training.train_starvla][INFO] - Step 2248, Loss: {'train/loss': 0.4845925569534302, 'train/grad_norm_pre_clip': 10.044868469238281, 'timing/data': 0.012025772593915462, 'timing/model': 1.3517889771610498, 'learning_rate/qwen_vl_interface': 7.903775199839506e-06, 'learning_rate/action_model': 7.903775199839505e-05, 'epoch': 0.82}) +[2026-06-17 20:44:57,565][starVLA.training.train_starvla][INFO] - Step 2249, Loss: {'train/loss': 0.7575204372406006, 'train/grad_norm_pre_clip': 14.8782958984375, 'timing/data': 0.012711388058960438, 'timing/model': 1.0380870187655091, 'learning_rate/qwen_vl_interface': 7.901981239237539e-06, 'learning_rate/action_model': 7.901981239237539e-05, 'epoch': 0.82}) +[2026-06-17 20:45:02,758][starVLA.training.train_starvla][INFO] - Step 2250, Loss: {'train/loss': 0.6388605237007141, 'train/grad_norm_pre_clip': 13.447471618652344, 'timing/data': 0.014614409767091274, 'timing/model': 1.0486441776156425, 'learning_rate/qwen_vl_interface': 7.900186728877009e-06, 'learning_rate/action_model': 7.90018672887701e-05, 'epoch': 0.82}) +[2026-06-17 20:45:07,378][starVLA.training.train_starvla][INFO] - Step 2251, Loss: {'train/loss': 0.7238845229148865, 'train/grad_norm_pre_clip': 12.112765312194824, 'timing/data': 0.01207059808075428, 'timing/model': 1.200060235336423, 'learning_rate/qwen_vl_interface': 7.898391669129919e-06, 'learning_rate/action_model': 7.898391669129918e-05, 'epoch': 0.82}) +[2026-06-17 20:45:15,810][starVLA.training.train_starvla][INFO] - Step 2252, Loss: {'train/loss': 0.9320793747901917, 'train/grad_norm_pre_clip': 15.717369079589844, 'timing/data': 0.014437669888138771, 'timing/model': 4.7163742277771235, 'learning_rate/qwen_vl_interface': 7.896596060368384e-06, 'learning_rate/action_model': 7.896596060368384e-05, 'epoch': 0.82}) +[2026-06-17 20:45:20,219][starVLA.training.train_starvla][INFO] - Step 2253, Loss: {'train/loss': 0.6197516322135925, 'train/grad_norm_pre_clip': 14.444674491882324, 'timing/data': 0.014665528200566769, 'timing/model': 1.0916904043406248, 'learning_rate/qwen_vl_interface': 7.89479990296464e-06, 'learning_rate/action_model': 7.894799902964639e-05, 'epoch': 0.82}) +[2026-06-17 20:45:25,137][starVLA.training.train_starvla][INFO] - Step 2254, Loss: {'train/loss': 0.6816195249557495, 'train/grad_norm_pre_clip': 13.430497169494629, 'timing/data': 0.012403825297951698, 'timing/model': 1.3410439053550363, 'learning_rate/qwen_vl_interface': 7.893003197291028e-06, 'learning_rate/action_model': 7.893003197291027e-05, 'epoch': 0.83}) +[2026-06-17 20:45:30,175][starVLA.training.train_starvla][INFO] - Step 2255, Loss: {'train/loss': 0.6481758952140808, 'train/grad_norm_pre_clip': 18.965633392333984, 'timing/data': 0.01384667493402958, 'timing/model': 1.0424664989113808, 'learning_rate/qwen_vl_interface': 7.89120594372001e-06, 'learning_rate/action_model': 7.89120594372001e-05, 'epoch': 0.83}) +[2026-06-17 20:45:35,391][starVLA.training.train_starvla][INFO] - Step 2256, Loss: {'train/loss': 0.528904378414154, 'train/grad_norm_pre_clip': 12.252402305603027, 'timing/data': 0.013112988322973251, 'timing/model': 1.2098632901906967, 'learning_rate/qwen_vl_interface': 7.88940814262416e-06, 'learning_rate/action_model': 7.889408142624158e-05, 'epoch': 0.83}) +[2026-06-17 20:45:40,362][starVLA.training.train_starvla][INFO] - Step 2257, Loss: {'train/loss': 0.7553123235702515, 'train/grad_norm_pre_clip': 11.709953308105469, 'timing/data': 0.01354182418435812, 'timing/model': 1.4393625734373927, 'learning_rate/qwen_vl_interface': 7.887609794376158e-06, 'learning_rate/action_model': 7.887609794376158e-05, 'epoch': 0.83}) +[2026-06-17 20:45:44,938][starVLA.training.train_starvla][INFO] - Step 2258, Loss: {'train/loss': 0.6795592904090881, 'train/grad_norm_pre_clip': 14.500455856323242, 'timing/data': 0.014219868928194046, 'timing/model': 1.0761816212907434, 'learning_rate/qwen_vl_interface': 7.88581089934881e-06, 'learning_rate/action_model': 7.88581089934881e-05, 'epoch': 0.83}) +[2026-06-17 20:45:51,940][starVLA.training.train_starvla][INFO] - Step 2259, Loss: {'train/loss': 0.27851441502571106, 'train/grad_norm_pre_clip': 7.045115947723389, 'timing/data': 0.012128264643251896, 'timing/model': 1.133557322435081, 'learning_rate/qwen_vl_interface': 7.884011457915025e-06, 'learning_rate/action_model': 7.884011457915024e-05, 'epoch': 0.83}) +[2026-06-17 20:46:05,693][starVLA.training.train_starvla][INFO] - Step 2260, Loss: {'train/loss': 0.6497730612754822, 'train/grad_norm_pre_clip': 11.559554100036621, 'timing/data': 0.014935738407075405, 'timing/model': 1.1255726218223572, 'learning_rate/qwen_vl_interface': 7.882211470447832e-06, 'learning_rate/action_model': 7.882211470447831e-05, 'epoch': 0.83}) +[2026-06-17 20:46:12,494][starVLA.training.train_starvla][INFO] - Step 2261, Loss: {'train/loss': 0.7924342751502991, 'train/grad_norm_pre_clip': 16.758544921875, 'timing/data': 0.014907498843967915, 'timing/model': 1.0853163311257958, 'learning_rate/qwen_vl_interface': 7.880410937320367e-06, 'learning_rate/action_model': 7.880410937320368e-05, 'epoch': 0.83}) +[2026-06-17 20:46:26,736][starVLA.training.train_starvla][INFO] - Step 2262, Loss: {'train/loss': 0.5960139632225037, 'train/grad_norm_pre_clip': 13.253105163574219, 'timing/data': 0.01610054261982441, 'timing/model': 1.8438565880060196, 'learning_rate/qwen_vl_interface': 7.878609858905885e-06, 'learning_rate/action_model': 7.878609858905886e-05, 'epoch': 0.83}) +[2026-06-17 20:46:36,244][starVLA.training.train_starvla][INFO] - Step 2263, Loss: {'train/loss': 0.6425042748451233, 'train/grad_norm_pre_clip': 13.077825546264648, 'timing/data': 0.011647013016045094, 'timing/model': 1.2186055546626449, 'learning_rate/qwen_vl_interface': 7.876808235577751e-06, 'learning_rate/action_model': 7.87680823557775e-05, 'epoch': 0.83}) +[2026-06-17 20:46:41,130][starVLA.training.train_starvla][INFO] - Step 2264, Loss: {'train/loss': 0.6298997402191162, 'train/grad_norm_pre_clip': 12.153966903686523, 'timing/data': 0.01398477703332901, 'timing/model': 1.199181723408401, 'learning_rate/qwen_vl_interface': 7.875006067709444e-06, 'learning_rate/action_model': 7.875006067709443e-05, 'epoch': 0.83}) +[2026-06-17 20:46:46,134][starVLA.training.train_starvla][INFO] - Step 2265, Loss: {'train/loss': 0.5799254179000854, 'train/grad_norm_pre_clip': 15.119152069091797, 'timing/data': 0.013294263742864132, 'timing/model': 1.0979466577991843, 'learning_rate/qwen_vl_interface': 7.873203355674553e-06, 'learning_rate/action_model': 7.873203355674552e-05, 'epoch': 0.83}) +[2026-06-17 20:46:51,863][starVLA.training.train_starvla][INFO] - Step 2266, Loss: {'train/loss': 0.27711009979248047, 'train/grad_norm_pre_clip': 12.362038612365723, 'timing/data': 0.0010883109644055367, 'timing/model': 1.3670370075851679, 'learning_rate/qwen_vl_interface': 7.871400099846784e-06, 'learning_rate/action_model': 7.871400099846783e-05, 'epoch': 0.83}) +[2026-06-17 20:46:56,575][starVLA.training.train_starvla][INFO] - Step 2267, Loss: {'train/loss': 0.2742370665073395, 'train/grad_norm_pre_clip': 8.664177894592285, 'timing/data': 0.000776851549744606, 'timing/model': 1.5746287070214748, 'learning_rate/qwen_vl_interface': 7.869596300599954e-06, 'learning_rate/action_model': 7.869596300599953e-05, 'epoch': 0.83}) +[2026-06-17 20:47:02,662][starVLA.training.train_starvla][INFO] - Step 2268, Loss: {'train/loss': 0.2741478979587555, 'train/grad_norm_pre_clip': 9.944464683532715, 'timing/data': 0.002246263436973095, 'timing/model': 1.0418669618666172, 'learning_rate/qwen_vl_interface': 7.867791958307992e-06, 'learning_rate/action_model': 7.867791958307992e-05, 'epoch': 0.83}) +[2026-06-17 20:47:11,835][starVLA.training.train_starvla][INFO] - Step 2269, Loss: {'train/loss': 0.6427359580993652, 'train/grad_norm_pre_clip': 13.045106887817383, 'timing/data': 0.0019675521180033684, 'timing/model': 1.126075522042811, 'learning_rate/qwen_vl_interface': 7.865987073344942e-06, 'learning_rate/action_model': 7.865987073344942e-05, 'epoch': 0.83}) +[2026-06-17 20:47:16,906][starVLA.training.train_starvla][INFO] - Step 2270, Loss: {'train/loss': 0.7098934054374695, 'train/grad_norm_pre_clip': 17.108356475830078, 'timing/data': 0.001381746493279934, 'timing/model': 1.2132172249257565, 'learning_rate/qwen_vl_interface': 7.864181646084953e-06, 'learning_rate/action_model': 7.864181646084953e-05, 'epoch': 0.83}) +[2026-06-17 20:47:21,781][starVLA.training.train_starvla][INFO] - Step 2271, Loss: {'train/loss': 0.7811508774757385, 'train/grad_norm_pre_clip': 15.357735633850098, 'timing/data': 0.000904853455722332, 'timing/model': 1.216684709303081, 'learning_rate/qwen_vl_interface': 7.862375676902298e-06, 'learning_rate/action_model': 7.862375676902297e-05, 'epoch': 0.83}) +[2026-06-17 20:47:28,503][starVLA.training.train_starvla][INFO] - Step 2272, Loss: {'train/loss': 0.6420020461082458, 'train/grad_norm_pre_clip': 15.70606517791748, 'timing/data': 0.001336636021733284, 'timing/model': 1.6444008089601994, 'learning_rate/qwen_vl_interface': 7.860569166171353e-06, 'learning_rate/action_model': 7.860569166171354e-05, 'epoch': 0.83}) +[2026-06-17 20:47:38,929][starVLA.training.train_starvla][INFO] - Step 2273, Loss: {'train/loss': 0.42353057861328125, 'train/grad_norm_pre_clip': 15.631303787231445, 'timing/data': 0.0019092485308647156, 'timing/model': 1.1156322723254561, 'learning_rate/qwen_vl_interface': 7.858762114266613e-06, 'learning_rate/action_model': 7.858762114266612e-05, 'epoch': 0.83}) +[2026-06-17 20:47:43,660][starVLA.training.train_starvla][INFO] - Step 2274, Loss: {'train/loss': 0.9263452291488647, 'train/grad_norm_pre_clip': 18.822460174560547, 'timing/data': 0.0011619506403803825, 'timing/model': 1.1179834557697177, 'learning_rate/qwen_vl_interface': 7.856954521562677e-06, 'learning_rate/action_model': 7.856954521562677e-05, 'epoch': 0.83}) +[2026-06-17 20:47:50,465][starVLA.training.train_starvla][INFO] - Step 2275, Loss: {'train/loss': 0.6224720478057861, 'train/grad_norm_pre_clip': 12.298503875732422, 'timing/data': 0.001443847082555294, 'timing/model': 1.113616325892508, 'learning_rate/qwen_vl_interface': 7.855146388434264e-06, 'learning_rate/action_model': 7.855146388434264e-05, 'epoch': 0.83}) +[2026-06-17 20:47:58,303][starVLA.training.train_starvla][INFO] - Step 2276, Loss: {'train/loss': 0.8446208834648132, 'train/grad_norm_pre_clip': 14.301750183105469, 'timing/data': 0.0009463950991630554, 'timing/model': 1.054900567047298, 'learning_rate/qwen_vl_interface': 7.853337715256202e-06, 'learning_rate/action_model': 7.853337715256203e-05, 'epoch': 0.83}) +[2026-06-17 20:48:03,478][starVLA.training.train_starvla][INFO] - Step 2277, Loss: {'train/loss': 0.9180824756622314, 'train/grad_norm_pre_clip': 20.567461013793945, 'timing/data': 0.0013628723099827766, 'timing/model': 1.6472135903313756, 'learning_rate/qwen_vl_interface': 7.851528502403431e-06, 'learning_rate/action_model': 7.851528502403431e-05, 'epoch': 0.83}) +[2026-06-17 20:48:08,011][starVLA.training.train_starvla][INFO] - Step 2278, Loss: {'train/loss': 0.5621031522750854, 'train/grad_norm_pre_clip': 35.19254684448242, 'timing/data': 0.0011152755469083786, 'timing/model': 1.0449106581509113, 'learning_rate/qwen_vl_interface': 7.849718750251001e-06, 'learning_rate/action_model': 7.849718750251002e-05, 'epoch': 0.83}) +[2026-06-17 20:48:20,546][starVLA.training.train_starvla][INFO] - Step 2279, Loss: {'train/loss': 0.42945215106010437, 'train/grad_norm_pre_clip': 11.54494857788086, 'timing/data': 0.0009997785091400146, 'timing/model': 1.3628799794241786, 'learning_rate/qwen_vl_interface': 7.847908459174077e-06, 'learning_rate/action_model': 7.847908459174078e-05, 'epoch': 0.83}) +[2026-06-17 20:48:25,135][starVLA.training.train_starvla][INFO] - Step 2280, Loss: {'train/loss': 0.7876271605491638, 'train/grad_norm_pre_clip': 13.317644119262695, 'timing/data': 0.0012096650898456573, 'timing/model': 1.144699557684362, 'learning_rate/qwen_vl_interface': 7.846097629547934e-06, 'learning_rate/action_model': 7.846097629547933e-05, 'epoch': 0.83}) +[2026-06-17 20:48:29,614][starVLA.training.train_starvla][INFO] - Step 2281, Loss: {'train/loss': 0.5737797021865845, 'train/grad_norm_pre_clip': 12.424551010131836, 'timing/data': 0.0007850043475627899, 'timing/model': 1.0927625084295869, 'learning_rate/qwen_vl_interface': 7.844286261747959e-06, 'learning_rate/action_model': 7.844286261747958e-05, 'epoch': 0.84}) +[2026-06-17 20:48:34,617][starVLA.training.train_starvla][INFO] - Step 2282, Loss: {'train/loss': 0.8027631044387817, 'train/grad_norm_pre_clip': 13.878935813903809, 'timing/data': 0.0010383892804384232, 'timing/model': 1.591536171734333, 'learning_rate/qwen_vl_interface': 7.84247435614965e-06, 'learning_rate/action_model': 7.84247435614965e-05, 'epoch': 0.84}) +[2026-06-17 20:48:39,631][starVLA.training.train_starvla][INFO] - Step 2283, Loss: {'train/loss': 0.6207088828086853, 'train/grad_norm_pre_clip': 14.354546546936035, 'timing/data': 0.0008892929181456566, 'timing/model': 1.1714110858738422, 'learning_rate/qwen_vl_interface': 7.840661913128618e-06, 'learning_rate/action_model': 7.840661913128619e-05, 'epoch': 0.84}) +[2026-06-17 20:48:44,550][starVLA.training.train_starvla][INFO] - Step 2284, Loss: {'train/loss': 0.6605440974235535, 'train/grad_norm_pre_clip': 13.839569091796875, 'timing/data': 0.001097293570637703, 'timing/model': 1.1723701125010848, 'learning_rate/qwen_vl_interface': 7.838848933060583e-06, 'learning_rate/action_model': 7.838848933060583e-05, 'epoch': 0.84}) +[2026-06-17 20:48:49,537][starVLA.training.train_starvla][INFO] - Step 2285, Loss: {'train/loss': 0.8331118822097778, 'train/grad_norm_pre_clip': 13.021513938903809, 'timing/data': 0.000993405468761921, 'timing/model': 1.2514560399577022, 'learning_rate/qwen_vl_interface': 7.83703541632138e-06, 'learning_rate/action_model': 7.83703541632138e-05, 'epoch': 0.84}) +[2026-06-17 20:48:55,785][starVLA.training.train_starvla][INFO] - Step 2286, Loss: {'train/loss': 0.5108500123023987, 'train/grad_norm_pre_clip': 13.586285591125488, 'timing/data': 0.0012890482321381569, 'timing/model': 1.1127838212996721, 'learning_rate/qwen_vl_interface': 7.835221363286949e-06, 'learning_rate/action_model': 7.835221363286949e-05, 'epoch': 0.84}) +[2026-06-17 20:49:00,380][starVLA.training.train_starvla][INFO] - Step 2287, Loss: {'train/loss': 0.520912230014801, 'train/grad_norm_pre_clip': 17.715755462646484, 'timing/data': 0.0011421404778957367, 'timing/model': 1.4643712295219302, 'learning_rate/qwen_vl_interface': 7.83340677433335e-06, 'learning_rate/action_model': 7.83340677433335e-05, 'epoch': 0.84}) +[2026-06-17 20:49:04,543][starVLA.training.train_starvla][INFO] - Step 2288, Loss: {'train/loss': 0.6863988637924194, 'train/grad_norm_pre_clip': 13.483454704284668, 'timing/data': 0.0011173458769917488, 'timing/model': 1.0765391066670418, 'learning_rate/qwen_vl_interface': 7.831591649836745e-06, 'learning_rate/action_model': 7.831591649836744e-05, 'epoch': 0.84}) +[2026-06-17 20:49:14,668][starVLA.training.train_starvla][INFO] - Step 2289, Loss: {'train/loss': 0.526612401008606, 'train/grad_norm_pre_clip': 11.75000286102295, 'timing/data': 0.0007630633190274239, 'timing/model': 1.0462772902101278, 'learning_rate/qwen_vl_interface': 7.829775990173413e-06, 'learning_rate/action_model': 7.829775990173413e-05, 'epoch': 0.84}) +[2026-06-17 20:49:19,499][starVLA.training.train_starvla][INFO] - Step 2290, Loss: {'train/loss': 0.8971198797225952, 'train/grad_norm_pre_clip': 17.015640258789062, 'timing/data': 0.0012952480465173721, 'timing/model': 1.0657858736813068, 'learning_rate/qwen_vl_interface': 7.827959795719744e-06, 'learning_rate/action_model': 7.827959795719744e-05, 'epoch': 0.84}) +[2026-06-17 20:49:27,304][starVLA.training.train_starvla][INFO] - Step 2291, Loss: {'train/loss': 0.7479790449142456, 'train/grad_norm_pre_clip': 15.272514343261719, 'timing/data': 0.0011946894228458405, 'timing/model': 4.369197613559663, 'learning_rate/qwen_vl_interface': 7.826143066852233e-06, 'learning_rate/action_model': 7.826143066852233e-05, 'epoch': 0.84}) +[2026-06-17 20:49:36,249][starVLA.training.train_starvla][INFO] - Step 2292, Loss: {'train/loss': 0.5520063042640686, 'train/grad_norm_pre_clip': 15.060190200805664, 'timing/data': 0.0013067685067653656, 'timing/model': 1.410783159546554, 'learning_rate/qwen_vl_interface': 7.824325803947494e-06, 'learning_rate/action_model': 7.824325803947495e-05, 'epoch': 0.84}) +[2026-06-17 20:49:40,417][starVLA.training.train_starvla][INFO] - Step 2293, Loss: {'train/loss': 0.9455338716506958, 'train/grad_norm_pre_clip': 15.409310340881348, 'timing/data': 0.0012620408087968826, 'timing/model': 1.0776301743462682, 'learning_rate/qwen_vl_interface': 7.822508007382246e-06, 'learning_rate/action_model': 7.822508007382245e-05, 'epoch': 0.84}) +[2026-06-17 20:49:45,333][starVLA.training.train_starvla][INFO] - Step 2294, Loss: {'train/loss': 0.5920009613037109, 'train/grad_norm_pre_clip': 12.18079662322998, 'timing/data': 0.0013160016387701035, 'timing/model': 1.494125572964549, 'learning_rate/qwen_vl_interface': 7.820689677533319e-06, 'learning_rate/action_model': 7.82068967753332e-05, 'epoch': 0.84}) +[2026-06-17 20:49:52,804][starVLA.training.train_starvla][INFO] - Step 2295, Loss: {'train/loss': 0.9172078967094421, 'train/grad_norm_pre_clip': 15.771194458007812, 'timing/data': 0.0011014863848686218, 'timing/model': 1.02601055521518, 'learning_rate/qwen_vl_interface': 7.818870814777655e-06, 'learning_rate/action_model': 7.818870814777655e-05, 'epoch': 0.84}) +[2026-06-17 20:49:57,345][starVLA.training.train_starvla][INFO] - Step 2296, Loss: {'train/loss': 0.8548704981803894, 'train/grad_norm_pre_clip': 17.546798706054688, 'timing/data': 0.0010634567588567734, 'timing/model': 1.0293323723599315, 'learning_rate/qwen_vl_interface': 7.81705141949231e-06, 'learning_rate/action_model': 7.817051419492308e-05, 'epoch': 0.84}) +[2026-06-17 20:50:02,085][starVLA.training.train_starvla][INFO] - Step 2297, Loss: {'train/loss': 0.8066915273666382, 'train/grad_norm_pre_clip': 12.898377418518066, 'timing/data': 0.0008687637746334076, 'timing/model': 1.5313821928575635, 'learning_rate/qwen_vl_interface': 7.81523149205444e-06, 'learning_rate/action_model': 7.81523149205444e-05, 'epoch': 0.84}) +[2026-06-17 20:50:06,928][starVLA.training.train_starvla][INFO] - Step 2298, Loss: {'train/loss': 0.8059448599815369, 'train/grad_norm_pre_clip': 19.28586769104004, 'timing/data': 0.0011184243485331535, 'timing/model': 1.260721543803811, 'learning_rate/qwen_vl_interface': 7.813411032841323e-06, 'learning_rate/action_model': 7.813411032841323e-05, 'epoch': 0.84}) +[2026-06-17 20:50:11,340][starVLA.training.train_starvla][INFO] - Step 2299, Loss: {'train/loss': 0.6570338606834412, 'train/grad_norm_pre_clip': 12.357080459594727, 'timing/data': 0.0009017782285809517, 'timing/model': 1.0261156363412738, 'learning_rate/qwen_vl_interface': 7.81159004223034e-06, 'learning_rate/action_model': 7.81159004223034e-05, 'epoch': 0.84}) +[2026-06-17 20:53:34,669][starVLA.training.train_starvla][INFO] - Step 2300, Loss: {'train/loss': 0.7254390120506287, 'train/grad_norm_pre_clip': 14.16989803314209, 'eval/loss': 0.7317513823509216, 'eval/latency_2/loss': 0.7317513823509216, 'eval/demon_fire_precision': 0.8868813825608798, 'eval/demon_fire_recall': 0.9447698744769875, 'eval/demon_fire_f1': 0.9149108589951377, 'eval/demon_fire_rate_model': 0.7084028937117418, 'eval/demon_fire_rate_teacher': 0.6649972175848636, 'eval/demon_left_precision': 0.8506024096385543, 'eval/demon_left_recall': 0.7431578947368421, 'eval/demon_left_f1': 0.7932584269662921, 'eval/demon_left_rate_model': 0.23094045631608237, 'eval/demon_left_rate_teacher': 0.2643294379521425, 'eval/demon_right_precision': 0.7546798029556651, 'eval/demon_right_recall': 0.8655367231638418, 'eval/demon_right_f1': 0.8063157894736843, 'eval/demon_right_rate_model': 0.2824151363383417, 'eval/demon_right_rate_teacher': 0.24624373956594323, 'eval/demon_move_f1': 0.7997871082199882, 'eval/demon_cc_f1': 0.8770378239725635, 'eval/latency_2/demon_fire_precision': 0.8868813825608798, 'eval/latency_2/demon_fire_recall': 0.9447698744769875, 'eval/latency_2/demon_fire_f1': 0.9149108589951377, 'eval/latency_2/demon_fire_rate_model': 0.7084028937117418, 'eval/latency_2/demon_fire_rate_teacher': 0.6649972175848636, 'eval/latency_2/demon_left_precision': 0.8506024096385543, 'eval/latency_2/demon_left_recall': 0.7431578947368421, 'eval/latency_2/demon_left_f1': 0.7932584269662921, 'eval/latency_2/demon_left_rate_model': 0.23094045631608237, 'eval/latency_2/demon_left_rate_teacher': 0.2643294379521425, 'eval/latency_2/demon_right_precision': 0.7546798029556651, 'eval/latency_2/demon_right_recall': 0.8655367231638418, 'eval/latency_2/demon_right_f1': 0.8063157894736843, 'eval/latency_2/demon_right_rate_model': 0.2824151363383417, 'eval/latency_2/demon_right_rate_teacher': 0.24624373956594323, 'eval/latency_2/demon_move_f1': 0.7997871082199882, 'eval/latency_2/demon_cc_f1': 0.8770378239725635, 'timing/data': 0.0012473370879888535, 'timing/model': 1.027102030813694, 'learning_rate/qwen_vl_interface': 7.809768520598986e-06, 'learning_rate/action_model': 7.809768520598984e-05, 'epoch': 0.84}) +[2026-06-17 20:53:34,671][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_2300_state +[2026-06-17 20:53:34,679][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 20:53:43,175][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_2300_state/model.safetensors +[2026-06-17 20:53:57,564][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_2300_state/optimizer.bin +[2026-06-17 20:53:57,565][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_2300_state/sampler.bin +[2026-06-17 20:53:57,566][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_2300_state/sampler_1.bin +[2026-06-17 20:53:57,568][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_2300_state/random_states_0.pkl +[2026-06-17 20:53:57,597][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2300 +[2026-06-17 20:53:57,597][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 20:53:57,614][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 20:54:02,775][starVLA.training.train_starvla][INFO] - Step 2301, Loss: {'train/loss': 0.7766382694244385, 'train/grad_norm_pre_clip': 13.885343551635742, 'timing/data': 0.0010390682145953178, 'timing/model': 1.172685799188912, 'learning_rate/qwen_vl_interface': 7.80794646832486e-06, 'learning_rate/action_model': 7.80794646832486e-05, 'epoch': 0.84}) +[2026-06-17 20:54:09,298][starVLA.training.train_starvla][INFO] - Step 2302, Loss: {'train/loss': 0.7180058360099792, 'train/grad_norm_pre_clip': 15.098349571228027, 'timing/data': 0.0012478232383728027, 'timing/model': 2.7407568730413914, 'learning_rate/qwen_vl_interface': 7.80612388578568e-06, 'learning_rate/action_model': 7.80612388578568e-05, 'epoch': 0.84}) +[2026-06-17 20:54:13,996][starVLA.training.train_starvla][INFO] - Step 2303, Loss: {'train/loss': 0.6839803457260132, 'train/grad_norm_pre_clip': 11.035399436950684, 'timing/data': 0.0012771263718605042, 'timing/model': 1.3968660160899162, 'learning_rate/qwen_vl_interface': 7.80430077335927e-06, 'learning_rate/action_model': 7.80430077335927e-05, 'epoch': 0.84}) +[2026-06-17 20:54:18,215][starVLA.training.train_starvla][INFO] - Step 2304, Loss: {'train/loss': 0.7002794146537781, 'train/grad_norm_pre_clip': 14.982848167419434, 'timing/data': 0.0009894492104649544, 'timing/model': 1.1270369123667479, 'learning_rate/qwen_vl_interface': 7.802477131423561e-06, 'learning_rate/action_model': 7.80247713142356e-05, 'epoch': 0.84}) +[2026-06-17 20:54:22,730][starVLA.training.train_starvla][INFO] - Step 2305, Loss: {'train/loss': 0.8044792413711548, 'train/grad_norm_pre_clip': 16.067296981811523, 'timing/data': 0.0007683634757995605, 'timing/model': 1.131714797578752, 'learning_rate/qwen_vl_interface': 7.800652960356594e-06, 'learning_rate/action_model': 7.800652960356594e-05, 'epoch': 0.84}) +[2026-06-17 20:54:27,477][starVLA.training.train_starvla][INFO] - Step 2306, Loss: {'train/loss': 0.4055963158607483, 'train/grad_norm_pre_clip': 10.738723754882812, 'timing/data': 0.0008970014750957489, 'timing/model': 1.0844859555363655, 'learning_rate/qwen_vl_interface': 7.798828260536524e-06, 'learning_rate/action_model': 7.798828260536524e-05, 'epoch': 0.84}) +[2026-06-17 20:54:32,350][starVLA.training.train_starvla][INFO] - Step 2307, Loss: {'train/loss': 0.5083659291267395, 'train/grad_norm_pre_clip': 12.212681770324707, 'timing/data': 0.0011022724211215973, 'timing/model': 1.0239037750288844, 'learning_rate/qwen_vl_interface': 7.797003032341615e-06, 'learning_rate/action_model': 7.797003032341614e-05, 'epoch': 0.84}) +[2026-06-17 20:54:40,012][starVLA.training.train_starvla][INFO] - Step 2308, Loss: {'train/loss': 0.7725055813789368, 'train/grad_norm_pre_clip': 16.717906951904297, 'timing/data': 0.0022530052810907364, 'timing/model': 4.41418325714767, 'learning_rate/qwen_vl_interface': 7.795177276150234e-06, 'learning_rate/action_model': 7.795177276150234e-05, 'epoch': 0.85}) +[2026-06-17 20:54:47,778][starVLA.training.train_starvla][INFO] - Step 2309, Loss: {'train/loss': 0.6381253600120544, 'train/grad_norm_pre_clip': 13.729537010192871, 'timing/data': 0.0010639447718858719, 'timing/model': 4.587443149648607, 'learning_rate/qwen_vl_interface': 7.793350992340867e-06, 'learning_rate/action_model': 7.793350992340867e-05, 'epoch': 0.85}) +[2026-06-17 20:55:00,188][starVLA.training.train_starvla][INFO] - Step 2310, Loss: {'train/loss': 0.7816944718360901, 'train/grad_norm_pre_clip': 19.814435958862305, 'timing/data': 0.001040767878293991, 'timing/model': 1.0333517603576183, 'learning_rate/qwen_vl_interface': 7.791524181292102e-06, 'learning_rate/action_model': 7.791524181292101e-05, 'epoch': 0.85}) +[2026-06-17 20:55:04,790][starVLA.training.train_starvla][INFO] - Step 2311, Loss: {'train/loss': 0.7822957038879395, 'train/grad_norm_pre_clip': 14.475554466247559, 'timing/data': 0.0008932324126362801, 'timing/model': 1.0563680538907647, 'learning_rate/qwen_vl_interface': 7.789696843382638e-06, 'learning_rate/action_model': 7.789696843382639e-05, 'epoch': 0.85}) +[2026-06-17 20:55:09,315][starVLA.training.train_starvla][INFO] - Step 2312, Loss: {'train/loss': 1.0707708597183228, 'train/grad_norm_pre_clip': 18.434362411499023, 'timing/data': 0.000764031894505024, 'timing/model': 1.0912958458065987, 'learning_rate/qwen_vl_interface': 7.787868978991288e-06, 'learning_rate/action_model': 7.787868978991287e-05, 'epoch': 0.85}) +[2026-06-17 20:55:14,193][starVLA.training.train_starvla][INFO] - Step 2313, Loss: {'train/loss': 0.6158562302589417, 'train/grad_norm_pre_clip': 12.611063003540039, 'timing/data': 0.000808357261121273, 'timing/model': 1.5533797992393374, 'learning_rate/qwen_vl_interface': 7.786040588496966e-06, 'learning_rate/action_model': 7.786040588496965e-05, 'epoch': 0.85}) +[2026-06-17 20:55:20,243][starVLA.training.train_starvla][INFO] - Step 2314, Loss: {'train/loss': 0.7569523453712463, 'train/grad_norm_pre_clip': 13.150136947631836, 'timing/data': 0.0010723713785409927, 'timing/model': 1.0341236405074596, 'learning_rate/qwen_vl_interface': 7.784211672278703e-06, 'learning_rate/action_model': 7.784211672278703e-05, 'epoch': 0.85}) +[2026-06-17 20:55:28,184][starVLA.training.train_starvla][INFO] - Step 2315, Loss: {'train/loss': 0.4580719470977783, 'train/grad_norm_pre_clip': 10.968043327331543, 'timing/data': 0.000781206414103508, 'timing/model': 1.1921143494546413, 'learning_rate/qwen_vl_interface': 7.78238223071563e-06, 'learning_rate/action_model': 7.78238223071563e-05, 'epoch': 0.85}) +[2026-06-17 20:55:32,965][starVLA.training.train_starvla][INFO] - Step 2316, Loss: {'train/loss': 0.5595022439956665, 'train/grad_norm_pre_clip': 12.336291313171387, 'timing/data': 0.0017725620418787003, 'timing/model': 1.180682641454041, 'learning_rate/qwen_vl_interface': 7.780552264187e-06, 'learning_rate/action_model': 7.780552264186999e-05, 'epoch': 0.85}) +[2026-06-17 20:55:37,612][starVLA.training.train_starvla][INFO] - Step 2317, Loss: {'train/loss': 0.6493220925331116, 'train/grad_norm_pre_clip': 13.256556510925293, 'timing/data': 0.0008521145209670067, 'timing/model': 1.0544378058984876, 'learning_rate/qwen_vl_interface': 7.77872177307216e-06, 'learning_rate/action_model': 7.778721773072159e-05, 'epoch': 0.85}) +[2026-06-17 20:55:47,019][starVLA.training.train_starvla][INFO] - Step 2318, Loss: {'train/loss': 0.83292555809021, 'train/grad_norm_pre_clip': 15.821427345275879, 'timing/data': 0.00108232069760561, 'timing/model': 1.4021398359909654, 'learning_rate/qwen_vl_interface': 7.776890757750575e-06, 'learning_rate/action_model': 7.776890757750575e-05, 'epoch': 0.85}) +[2026-06-17 20:55:51,150][starVLA.training.train_starvla][INFO] - Step 2319, Loss: {'train/loss': 0.7883690595626831, 'train/grad_norm_pre_clip': 17.586315155029297, 'timing/data': 0.000875517725944519, 'timing/model': 1.0385027546435595, 'learning_rate/qwen_vl_interface': 7.775059218601816e-06, 'learning_rate/action_model': 7.775059218601816e-05, 'epoch': 0.85}) +[2026-06-17 20:55:55,862][starVLA.training.train_starvla][INFO] - Step 2320, Loss: {'train/loss': 0.49493664503097534, 'train/grad_norm_pre_clip': 11.476031303405762, 'timing/data': 0.0012128334492444992, 'timing/model': 1.134986744262278, 'learning_rate/qwen_vl_interface': 7.773227156005567e-06, 'learning_rate/action_model': 7.773227156005567e-05, 'epoch': 0.85}) +[2026-06-17 20:56:00,638][starVLA.training.train_starvla][INFO] - Step 2321, Loss: {'train/loss': 0.9543431997299194, 'train/grad_norm_pre_clip': 18.651683807373047, 'timing/data': 0.001658877357840538, 'timing/model': 1.2417781222611666, 'learning_rate/qwen_vl_interface': 7.771394570341612e-06, 'learning_rate/action_model': 7.771394570341612e-05, 'epoch': 0.85}) +[2026-06-17 20:56:05,456][starVLA.training.train_starvla][INFO] - Step 2322, Loss: {'train/loss': 0.47422513365745544, 'train/grad_norm_pre_clip': 11.307100296020508, 'timing/data': 0.0008166469633579254, 'timing/model': 1.0512913288548589, 'learning_rate/qwen_vl_interface': 7.769561461989848e-06, 'learning_rate/action_model': 7.769561461989848e-05, 'epoch': 0.85}) +[2026-06-17 20:56:11,805][starVLA.training.train_starvla][INFO] - Step 2323, Loss: {'train/loss': 0.8834088444709778, 'train/grad_norm_pre_clip': 22.931804656982422, 'timing/data': 0.0010332092642784119, 'timing/model': 1.4520640028640628, 'learning_rate/qwen_vl_interface': 7.767727831330285e-06, 'learning_rate/action_model': 7.767727831330285e-05, 'epoch': 0.85}) +[2026-06-17 20:56:20,420][starVLA.training.train_starvla][INFO] - Step 2324, Loss: {'train/loss': 1.0573254823684692, 'train/grad_norm_pre_clip': 18.8299503326416, 'timing/data': 0.0008756425231695175, 'timing/model': 1.2456592554226518, 'learning_rate/qwen_vl_interface': 7.765893678743029e-06, 'learning_rate/action_model': 7.76589367874303e-05, 'epoch': 0.85}) +[2026-06-17 20:56:28,078][starVLA.training.train_starvla][INFO] - Step 2325, Loss: {'train/loss': 0.6387016177177429, 'train/grad_norm_pre_clip': 13.982425689697266, 'timing/data': 0.005462978966534138, 'timing/model': 1.964146751910448, 'learning_rate/qwen_vl_interface': 7.76405900460831e-06, 'learning_rate/action_model': 7.76405900460831e-05, 'epoch': 0.85}) +[2026-06-17 20:56:38,220][starVLA.training.train_starvla][INFO] - Step 2326, Loss: {'train/loss': 0.5451692938804626, 'train/grad_norm_pre_clip': 13.635032653808594, 'timing/data': 0.0014260094612836838, 'timing/model': 2.9535262174904346, 'learning_rate/qwen_vl_interface': 7.762223809306453e-06, 'learning_rate/action_model': 7.762223809306452e-05, 'epoch': 0.85}) +[2026-06-17 20:56:48,309][starVLA.training.train_starvla][INFO] - Step 2327, Loss: {'train/loss': 0.623863160610199, 'train/grad_norm_pre_clip': 14.273041725158691, 'timing/data': 0.007969086989760399, 'timing/model': 4.679267758503556, 'learning_rate/qwen_vl_interface': 7.760388093217897e-06, 'learning_rate/action_model': 7.760388093217898e-05, 'epoch': 0.85}) +[2026-06-17 20:56:58,391][starVLA.training.train_starvla][INFO] - Step 2328, Loss: {'train/loss': 0.8384034633636475, 'train/grad_norm_pre_clip': 18.57666778564453, 'timing/data': 0.0012074559926986694, 'timing/model': 6.287347197532654, 'learning_rate/qwen_vl_interface': 7.758551856723187e-06, 'learning_rate/action_model': 7.758551856723188e-05, 'epoch': 0.85}) +[2026-06-17 20:57:10,311][starVLA.training.train_starvla][INFO] - Step 2329, Loss: {'train/loss': 0.6886416077613831, 'train/grad_norm_pre_clip': 16.585338592529297, 'timing/data': 0.0009468505159020424, 'timing/model': 5.680550291202962, 'learning_rate/qwen_vl_interface': 7.756715100202979e-06, 'learning_rate/action_model': 7.75671510020298e-05, 'epoch': 0.85}) +[2026-06-17 20:57:15,556][starVLA.training.train_starvla][INFO] - Step 2330, Loss: {'train/loss': 0.6742678284645081, 'train/grad_norm_pre_clip': 14.101813316345215, 'timing/data': 0.0008699214085936546, 'timing/model': 1.0772982463240623, 'learning_rate/qwen_vl_interface': 7.754877824038035e-06, 'learning_rate/action_model': 7.754877824038035e-05, 'epoch': 0.85}) +[2026-06-17 20:57:20,128][starVLA.training.train_starvla][INFO] - Step 2331, Loss: {'train/loss': 0.5745722055435181, 'train/grad_norm_pre_clip': 12.39336109161377, 'timing/data': 0.0011650528758764267, 'timing/model': 1.07113247923553, 'learning_rate/qwen_vl_interface': 7.75304002860922e-06, 'learning_rate/action_model': 7.75304002860922e-05, 'epoch': 0.85}) +[2026-06-17 20:57:24,826][starVLA.training.train_starvla][INFO] - Step 2332, Loss: {'train/loss': 0.705520510673523, 'train/grad_norm_pre_clip': 12.435516357421875, 'timing/data': 0.0012516146525740623, 'timing/model': 1.5258844941854477, 'learning_rate/qwen_vl_interface': 7.751201714297513e-06, 'learning_rate/action_model': 7.751201714297512e-05, 'epoch': 0.85}) +[2026-06-17 20:57:29,602][starVLA.training.train_starvla][INFO] - Step 2333, Loss: {'train/loss': 0.8433133363723755, 'train/grad_norm_pre_clip': 18.407089233398438, 'timing/data': 0.0008643688634037971, 'timing/model': 1.3598221791908145, 'learning_rate/qwen_vl_interface': 7.749362881484002e-06, 'learning_rate/action_model': 7.749362881484001e-05, 'epoch': 0.85}) +[2026-06-17 20:57:33,713][starVLA.training.train_starvla][INFO] - Step 2334, Loss: {'train/loss': 0.5926420092582703, 'train/grad_norm_pre_clip': 13.035684585571289, 'timing/data': 0.000859295018017292, 'timing/model': 1.1155966212972999, 'learning_rate/qwen_vl_interface': 7.747523530549874e-06, 'learning_rate/action_model': 7.747523530549875e-05, 'epoch': 0.85}) +[2026-06-17 20:57:38,082][starVLA.training.train_starvla][INFO] - Step 2335, Loss: {'train/loss': 0.4332510232925415, 'train/grad_norm_pre_clip': 10.51098918914795, 'timing/data': 0.0018517374992370605, 'timing/model': 1.0501075582578778, 'learning_rate/qwen_vl_interface': 7.745683661876432e-06, 'learning_rate/action_model': 7.745683661876432e-05, 'epoch': 0.86}) +[2026-06-17 20:57:42,552][starVLA.training.train_starvla][INFO] - Step 2336, Loss: {'train/loss': 0.5736571550369263, 'train/grad_norm_pre_clip': 125.9189453125, 'timing/data': 0.0008756723254919052, 'timing/model': 1.034846062771976, 'learning_rate/qwen_vl_interface': 7.743843275845081e-06, 'learning_rate/action_model': 7.74384327584508e-05, 'epoch': 0.86}) +[2026-06-17 20:57:47,002][starVLA.training.train_starvla][INFO] - Step 2337, Loss: {'train/loss': 0.6516950130462646, 'train/grad_norm_pre_clip': 14.314945220947266, 'timing/data': 0.001277468167245388, 'timing/model': 1.0686884000897408, 'learning_rate/qwen_vl_interface': 7.742002372837336e-06, 'learning_rate/action_model': 7.742002372837335e-05, 'epoch': 0.86}) +[2026-06-17 20:57:51,578][starVLA.training.train_starvla][INFO] - Step 2338, Loss: {'train/loss': 0.8705958127975464, 'train/grad_norm_pre_clip': 17.28321075439453, 'timing/data': 0.0008463198319077492, 'timing/model': 1.3942431267350912, 'learning_rate/qwen_vl_interface': 7.740160953234815e-06, 'learning_rate/action_model': 7.740160953234815e-05, 'epoch': 0.86}) +[2026-06-17 20:57:55,831][starVLA.training.train_starvla][INFO] - Step 2339, Loss: {'train/loss': 1.1647768020629883, 'train/grad_norm_pre_clip': 21.972209930419922, 'timing/data': 0.0011169034987688065, 'timing/model': 1.1426953878253698, 'learning_rate/qwen_vl_interface': 7.738319017419252e-06, 'learning_rate/action_model': 7.738319017419252e-05, 'epoch': 0.86}) +[2026-06-17 20:58:00,399][starVLA.training.train_starvla][INFO] - Step 2340, Loss: {'train/loss': 0.36954736709594727, 'train/grad_norm_pre_clip': 11.767149925231934, 'timing/data': 0.0010600071400403976, 'timing/model': 1.1863529477268457, 'learning_rate/qwen_vl_interface': 7.736476565772478e-06, 'learning_rate/action_model': 7.736476565772478e-05, 'epoch': 0.86}) +[2026-06-17 20:58:04,993][starVLA.training.train_starvla][INFO] - Step 2341, Loss: {'train/loss': 0.6297692656517029, 'train/grad_norm_pre_clip': 16.696388244628906, 'timing/data': 0.0009030653163790703, 'timing/model': 1.0461841402575374, 'learning_rate/qwen_vl_interface': 7.734633598676435e-06, 'learning_rate/action_model': 7.734633598676434e-05, 'epoch': 0.86}) +[2026-06-17 20:58:09,764][starVLA.training.train_starvla][INFO] - Step 2342, Loss: {'train/loss': 0.6170594692230225, 'train/grad_norm_pre_clip': 15.636459350585938, 'timing/data': 0.0010147877037525177, 'timing/model': 1.0773481335490942, 'learning_rate/qwen_vl_interface': 7.732790116513176e-06, 'learning_rate/action_model': 7.732790116513176e-05, 'epoch': 0.86}) +[2026-06-17 20:58:14,179][starVLA.training.train_starvla][INFO] - Step 2343, Loss: {'train/loss': 0.7761425375938416, 'train/grad_norm_pre_clip': 15.249346733093262, 'timing/data': 0.0010816529393196106, 'timing/model': 1.3413924481719732, 'learning_rate/qwen_vl_interface': 7.730946119664852e-06, 'learning_rate/action_model': 7.730946119664853e-05, 'epoch': 0.86}) +[2026-06-17 20:58:18,726][starVLA.training.train_starvla][INFO] - Step 2344, Loss: {'train/loss': 0.7706810832023621, 'train/grad_norm_pre_clip': 14.757963180541992, 'timing/data': 0.0007338272407650948, 'timing/model': 1.1335875233635306, 'learning_rate/qwen_vl_interface': 7.729101608513729e-06, 'learning_rate/action_model': 7.72910160851373e-05, 'epoch': 0.86}) +[2026-06-17 20:58:28,938][starVLA.training.train_starvla][INFO] - Step 2345, Loss: {'train/loss': 0.8244583606719971, 'train/grad_norm_pre_clip': 15.387151718139648, 'timing/data': 0.0010504508391022682, 'timing/model': 2.104129744693637, 'learning_rate/qwen_vl_interface': 7.727256583442175e-06, 'learning_rate/action_model': 7.727256583442175e-05, 'epoch': 0.86}) +[2026-06-17 20:58:33,795][starVLA.training.train_starvla][INFO] - Step 2346, Loss: {'train/loss': 0.5783945322036743, 'train/grad_norm_pre_clip': 14.142110824584961, 'timing/data': 0.0010812105610966682, 'timing/model': 1.0567705379799008, 'learning_rate/qwen_vl_interface': 7.725411044832665e-06, 'learning_rate/action_model': 7.725411044832665e-05, 'epoch': 0.86}) +[2026-06-17 20:58:42,816][starVLA.training.train_starvla][INFO] - Step 2347, Loss: {'train/loss': 0.8618989586830139, 'train/grad_norm_pre_clip': 19.138694763183594, 'timing/data': 0.0008785789832472801, 'timing/model': 1.1857352508231997, 'learning_rate/qwen_vl_interface': 7.723564993067783e-06, 'learning_rate/action_model': 7.723564993067783e-05, 'epoch': 0.86}) +[2026-06-17 20:58:51,224][starVLA.training.train_starvla][INFO] - Step 2348, Loss: {'train/loss': 0.77642822265625, 'train/grad_norm_pre_clip': 14.400835037231445, 'timing/data': 0.001023680903017521, 'timing/model': 3.7629575692117214, 'learning_rate/qwen_vl_interface': 7.721718428530214e-06, 'learning_rate/action_model': 7.721718428530214e-05, 'epoch': 0.86}) +[2026-06-17 20:58:59,875][starVLA.training.train_starvla][INFO] - Step 2349, Loss: {'train/loss': 0.5178269147872925, 'train/grad_norm_pre_clip': 11.442418098449707, 'timing/data': 0.0013236990198493004, 'timing/model': 2.6746330335736275, 'learning_rate/qwen_vl_interface': 7.719871351602757e-06, 'learning_rate/action_model': 7.719871351602757e-05, 'epoch': 0.86}) +[2026-06-17 20:59:12,046][starVLA.training.train_starvla][INFO] - Step 2350, Loss: {'train/loss': 0.779514491558075, 'train/grad_norm_pre_clip': 15.02751350402832, 'timing/data': 0.0010321130976080894, 'timing/model': 4.17290507350117, 'learning_rate/qwen_vl_interface': 7.718023762668309e-06, 'learning_rate/action_model': 7.718023762668309e-05, 'epoch': 0.86}) +[2026-06-17 20:59:20,124][starVLA.training.train_starvla][INFO] - Step 2351, Loss: {'train/loss': 0.3479509949684143, 'train/grad_norm_pre_clip': 19.567453384399414, 'timing/data': 0.0010034218430519104, 'timing/model': 1.1883792877197266, 'learning_rate/qwen_vl_interface': 7.716175662109882e-06, 'learning_rate/action_model': 7.716175662109881e-05, 'epoch': 0.86}) +[2026-06-17 20:59:25,223][starVLA.training.train_starvla][INFO] - Step 2352, Loss: {'train/loss': 0.605731725692749, 'train/grad_norm_pre_clip': 16.1644229888916, 'timing/data': 0.0014864960685372353, 'timing/model': 1.2140355510637164, 'learning_rate/qwen_vl_interface': 7.714327050310586e-06, 'learning_rate/action_model': 7.714327050310585e-05, 'epoch': 0.86}) +[2026-06-17 20:59:30,040][starVLA.training.train_starvla][INFO] - Step 2353, Loss: {'train/loss': 0.8100349307060242, 'train/grad_norm_pre_clip': 15.427497863769531, 'timing/data': 0.0011914316564798355, 'timing/model': 1.4792560152709484, 'learning_rate/qwen_vl_interface': 7.71247792765364e-06, 'learning_rate/action_model': 7.712477927653638e-05, 'epoch': 0.86}) +[2026-06-17 20:59:34,130][starVLA.training.train_starvla][INFO] - Step 2354, Loss: {'train/loss': 0.6585466265678406, 'train/grad_norm_pre_clip': 17.446260452270508, 'timing/data': 0.0016099242493510246, 'timing/model': 1.0812977096065879, 'learning_rate/qwen_vl_interface': 7.71062829452237e-06, 'learning_rate/action_model': 7.710628294522369e-05, 'epoch': 0.86}) +[2026-06-17 20:59:38,665][starVLA.training.train_starvla][INFO] - Step 2355, Loss: {'train/loss': 0.6540372371673584, 'train/grad_norm_pre_clip': 13.92119312286377, 'timing/data': 0.0009230487048625946, 'timing/model': 1.0397780584171414, 'learning_rate/qwen_vl_interface': 7.708778151300206e-06, 'learning_rate/action_model': 7.708778151300206e-05, 'epoch': 0.86}) +[2026-06-17 20:59:43,038][starVLA.training.train_starvla][INFO] - Step 2356, Loss: {'train/loss': 0.6926655173301697, 'train/grad_norm_pre_clip': 14.083197593688965, 'timing/data': 0.0012066122144460678, 'timing/model': 1.0243584895506501, 'learning_rate/qwen_vl_interface': 7.706927498370687e-06, 'learning_rate/action_model': 7.706927498370686e-05, 'epoch': 0.86}) +[2026-06-17 20:59:47,604][starVLA.training.train_starvla][INFO] - Step 2357, Loss: {'train/loss': 0.812306821346283, 'train/grad_norm_pre_clip': 13.711419105529785, 'timing/data': 0.0015262477099895477, 'timing/model': 1.134022269397974, 'learning_rate/qwen_vl_interface': 7.705076336117452e-06, 'learning_rate/action_model': 7.705076336117451e-05, 'epoch': 0.86}) +[2026-06-17 20:59:51,965][starVLA.training.train_starvla][INFO] - Step 2358, Loss: {'train/loss': 0.7151044011116028, 'train/grad_norm_pre_clip': 15.411161422729492, 'timing/data': 0.0009067095816135406, 'timing/model': 1.3407342499122024, 'learning_rate/qwen_vl_interface': 7.703224664924252e-06, 'learning_rate/action_model': 7.703224664924251e-05, 'epoch': 0.86}) +[2026-06-17 20:59:56,078][starVLA.training.train_starvla][INFO] - Step 2359, Loss: {'train/loss': 0.516252338886261, 'train/grad_norm_pre_clip': 12.076282501220703, 'timing/data': 0.001135147176682949, 'timing/model': 1.0253326306119561, 'learning_rate/qwen_vl_interface': 7.70137248517494e-06, 'learning_rate/action_model': 7.701372485174939e-05, 'epoch': 0.86}) +[2026-06-17 21:00:00,680][starVLA.training.train_starvla][INFO] - Step 2360, Loss: {'train/loss': 0.7555113434791565, 'train/grad_norm_pre_clip': 14.871285438537598, 'timing/data': 0.0008958661928772926, 'timing/model': 1.1383235761895776, 'learning_rate/qwen_vl_interface': 7.699519797253471e-06, 'learning_rate/action_model': 7.699519797253472e-05, 'epoch': 0.86}) +[2026-06-17 21:00:05,413][starVLA.training.train_starvla][INFO] - Step 2361, Loss: {'train/loss': 0.6201817393302917, 'train/grad_norm_pre_clip': 14.277771949768066, 'timing/data': 0.0008655143901705742, 'timing/model': 1.134879526682198, 'learning_rate/qwen_vl_interface': 7.697666601543915e-06, 'learning_rate/action_model': 7.697666601543915e-05, 'epoch': 0.86}) +[2026-06-17 21:00:14,122][starVLA.training.train_starvla][INFO] - Step 2362, Loss: {'train/loss': 1.2290761470794678, 'train/grad_norm_pre_clip': 16.492549896240234, 'timing/data': 0.0009139515459537506, 'timing/model': 1.7708999002352357, 'learning_rate/qwen_vl_interface': 7.695812898430438e-06, 'learning_rate/action_model': 7.695812898430437e-05, 'epoch': 0.86}) +[2026-06-17 21:00:19,677][starVLA.training.train_starvla][INFO] - Step 2363, Loss: {'train/loss': 0.5907232761383057, 'train/grad_norm_pre_clip': 13.144747734069824, 'timing/data': 0.001474057324230671, 'timing/model': 1.3566277781501412, 'learning_rate/qwen_vl_interface': 7.693958688297316e-06, 'learning_rate/action_model': 7.693958688297314e-05, 'epoch': 0.87}) +[2026-06-17 21:00:23,828][starVLA.training.train_starvla][INFO] - Step 2364, Loss: {'train/loss': 0.6611164212226868, 'train/grad_norm_pre_clip': 14.44583511352539, 'timing/data': 0.0012472458183765411, 'timing/model': 1.0422836849465966, 'learning_rate/qwen_vl_interface': 7.692103971528925e-06, 'learning_rate/action_model': 7.692103971528926e-05, 'epoch': 0.87}) +[2026-06-17 21:00:31,811][starVLA.training.train_starvla][INFO] - Step 2365, Loss: {'train/loss': 0.5835518836975098, 'train/grad_norm_pre_clip': 13.870882987976074, 'timing/data': 0.0010656686499714851, 'timing/model': 1.1298990827053785, 'learning_rate/qwen_vl_interface': 7.690248748509757e-06, 'learning_rate/action_model': 7.690248748509757e-05, 'epoch': 0.87}) +[2026-06-17 21:00:38,151][starVLA.training.train_starvla][INFO] - Step 2366, Loss: {'train/loss': 0.7482579946517944, 'train/grad_norm_pre_clip': 17.25244903564453, 'timing/data': 0.001294698566198349, 'timing/model': 1.040476412512362, 'learning_rate/qwen_vl_interface': 7.688393019624396e-06, 'learning_rate/action_model': 7.688393019624395e-05, 'epoch': 0.87}) +[2026-06-17 21:00:42,513][starVLA.training.train_starvla][INFO] - Step 2367, Loss: {'train/loss': 0.6675395965576172, 'train/grad_norm_pre_clip': 11.202957153320312, 'timing/data': 0.0007474729791283607, 'timing/model': 1.3985611610114574, 'learning_rate/qwen_vl_interface': 7.686536785257539e-06, 'learning_rate/action_model': 7.686536785257538e-05, 'epoch': 0.87}) +[2026-06-17 21:00:47,000][starVLA.training.train_starvla][INFO] - Step 2368, Loss: {'train/loss': 0.5077128410339355, 'train/grad_norm_pre_clip': 14.39350414276123, 'timing/data': 0.0010987808927893639, 'timing/model': 1.36444700229913, 'learning_rate/qwen_vl_interface': 7.684680045793982e-06, 'learning_rate/action_model': 7.684680045793983e-05, 'epoch': 0.87}) +[2026-06-17 21:01:01,150][starVLA.training.train_starvla][INFO] - Step 2369, Loss: {'train/loss': 0.6067289710044861, 'train/grad_norm_pre_clip': 17.02479362487793, 'timing/data': 0.0013762563467025757, 'timing/model': 4.205947056412697, 'learning_rate/qwen_vl_interface': 7.682822801618634e-06, 'learning_rate/action_model': 7.682822801618633e-05, 'epoch': 0.87}) +[2026-06-17 21:01:18,867][starVLA.training.train_starvla][INFO] - Step 2370, Loss: {'train/loss': 0.5720212459564209, 'train/grad_norm_pre_clip': 12.9772310256958, 'timing/data': 0.0012761754915118217, 'timing/model': 4.489098993130028, 'learning_rate/qwen_vl_interface': 7.680965053116498e-06, 'learning_rate/action_model': 7.680965053116499e-05, 'epoch': 0.87}) +[2026-06-17 21:01:23,618][starVLA.training.train_starvla][INFO] - Step 2371, Loss: {'train/loss': 0.6404989361763, 'train/grad_norm_pre_clip': 15.539063453674316, 'timing/data': 0.001137884333729744, 'timing/model': 1.0632692920044065, 'learning_rate/qwen_vl_interface': 7.679106800672695e-06, 'learning_rate/action_model': 7.679106800672695e-05, 'epoch': 0.87}) +[2026-06-17 21:01:28,144][starVLA.training.train_starvla][INFO] - Step 2372, Loss: {'train/loss': 0.8526483774185181, 'train/grad_norm_pre_clip': 16.77243995666504, 'timing/data': 0.0012519480660557747, 'timing/model': 1.0288613177835941, 'learning_rate/qwen_vl_interface': 7.677248044672438e-06, 'learning_rate/action_model': 7.677248044672437e-05, 'epoch': 0.87}) +[2026-06-17 21:01:43,312][starVLA.training.train_starvla][INFO] - Step 2373, Loss: {'train/loss': 0.6532724499702454, 'train/grad_norm_pre_clip': 12.794403076171875, 'timing/data': 0.0009161289781332016, 'timing/model': 2.7014936301857233, 'learning_rate/qwen_vl_interface': 7.675388785501048e-06, 'learning_rate/action_model': 7.675388785501047e-05, 'epoch': 0.87}) +[2026-06-17 21:01:54,616][starVLA.training.train_starvla][INFO] - Step 2374, Loss: {'train/loss': 0.810295820236206, 'train/grad_norm_pre_clip': 16.00604820251465, 'timing/data': 0.001070437952876091, 'timing/model': 1.0945965284481645, 'learning_rate/qwen_vl_interface': 7.673529023543951e-06, 'learning_rate/action_model': 7.673529023543951e-05, 'epoch': 0.87}) +[2026-06-17 21:02:03,571][starVLA.training.train_starvla][INFO] - Step 2375, Loss: {'train/loss': 0.5105183720588684, 'train/grad_norm_pre_clip': 14.07469654083252, 'timing/data': 0.0010850625112652779, 'timing/model': 1.120872807689011, 'learning_rate/qwen_vl_interface': 7.671668759186683e-06, 'learning_rate/action_model': 7.671668759186683e-05, 'epoch': 0.87}) +[2026-06-17 21:02:08,408][starVLA.training.train_starvla][INFO] - Step 2376, Loss: {'train/loss': 0.463692307472229, 'train/grad_norm_pre_clip': 16.693042755126953, 'timing/data': 0.0011578109115362167, 'timing/model': 1.1401413399726152, 'learning_rate/qwen_vl_interface': 7.669807992814875e-06, 'learning_rate/action_model': 7.669807992814875e-05, 'epoch': 0.87}) +[2026-06-17 21:02:12,731][starVLA.training.train_starvla][INFO] - Step 2377, Loss: {'train/loss': 1.177464485168457, 'train/grad_norm_pre_clip': 25.976722717285156, 'timing/data': 0.001100161112844944, 'timing/model': 1.0599177097901702, 'learning_rate/qwen_vl_interface': 7.667946724814264e-06, 'learning_rate/action_model': 7.667946724814265e-05, 'epoch': 0.87}) +[2026-06-17 21:02:21,814][starVLA.training.train_starvla][INFO] - Step 2378, Loss: {'train/loss': 0.717315137386322, 'train/grad_norm_pre_clip': 13.789958953857422, 'timing/data': 0.00116000697016716, 'timing/model': 3.165161171928048, 'learning_rate/qwen_vl_interface': 7.666084955570698e-06, 'learning_rate/action_model': 7.666084955570698e-05, 'epoch': 0.87}) +[2026-06-17 21:02:26,025][starVLA.training.train_starvla][INFO] - Step 2379, Loss: {'train/loss': 0.8435562252998352, 'train/grad_norm_pre_clip': 17.880281448364258, 'timing/data': 0.0010314928367733955, 'timing/model': 1.0706307981163263, 'learning_rate/qwen_vl_interface': 7.664222685470119e-06, 'learning_rate/action_model': 7.664222685470118e-05, 'epoch': 0.87}) +[2026-06-17 21:02:30,885][starVLA.training.train_starvla][INFO] - Step 2380, Loss: {'train/loss': 0.8940573930740356, 'train/grad_norm_pre_clip': 16.14278221130371, 'timing/data': 0.0012416290119290352, 'timing/model': 1.0345742190256715, 'learning_rate/qwen_vl_interface': 7.66235991489858e-06, 'learning_rate/action_model': 7.66235991489858e-05, 'epoch': 0.87}) +[2026-06-17 21:02:35,628][starVLA.training.train_starvla][INFO] - Step 2381, Loss: {'train/loss': 1.003150463104248, 'train/grad_norm_pre_clip': 23.970184326171875, 'timing/data': 0.0009192386642098427, 'timing/model': 1.300288979895413, 'learning_rate/qwen_vl_interface': 7.660496644242233e-06, 'learning_rate/action_model': 7.660496644242233e-05, 'epoch': 0.87}) +[2026-06-17 21:02:40,378][starVLA.training.train_starvla][INFO] - Step 2382, Loss: {'train/loss': 0.7826608419418335, 'train/grad_norm_pre_clip': 17.42380714416504, 'timing/data': 0.001468287780880928, 'timing/model': 1.28957096952945, 'learning_rate/qwen_vl_interface': 7.658632873887342e-06, 'learning_rate/action_model': 7.658632873887342e-05, 'epoch': 0.87}) +[2026-06-17 21:02:45,083][starVLA.training.train_starvla][INFO] - Step 2383, Loss: {'train/loss': 0.7141495943069458, 'train/grad_norm_pre_clip': 16.510175704956055, 'timing/data': 0.0008871033787727356, 'timing/model': 1.5678312378004193, 'learning_rate/qwen_vl_interface': 7.656768604220262e-06, 'learning_rate/action_model': 7.656768604220261e-05, 'epoch': 0.87}) +[2026-06-17 21:02:49,433][starVLA.training.train_starvla][INFO] - Step 2384, Loss: {'train/loss': 0.791069507598877, 'train/grad_norm_pre_clip': 14.017769813537598, 'timing/data': 0.001077624037861824, 'timing/model': 1.0973560428246856, 'learning_rate/qwen_vl_interface': 7.654903835627459e-06, 'learning_rate/action_model': 7.654903835627458e-05, 'epoch': 0.87}) +[2026-06-17 21:02:53,902][starVLA.training.train_starvla][INFO] - Step 2385, Loss: {'train/loss': 0.4793829321861267, 'train/grad_norm_pre_clip': 11.227391242980957, 'timing/data': 0.0008597411215305328, 'timing/model': 1.030167113058269, 'learning_rate/qwen_vl_interface': 7.653038568495505e-06, 'learning_rate/action_model': 7.653038568495506e-05, 'epoch': 0.87}) +[2026-06-17 21:02:59,826][starVLA.training.train_starvla][INFO] - Step 2386, Loss: {'train/loss': 0.6429298520088196, 'train/grad_norm_pre_clip': 13.599711418151855, 'timing/data': 0.0017828121781349182, 'timing/model': 1.1681437166407704, 'learning_rate/qwen_vl_interface': 7.651172803211068e-06, 'learning_rate/action_model': 7.651172803211068e-05, 'epoch': 0.87}) +[2026-06-17 21:03:04,355][starVLA.training.train_starvla][INFO] - Step 2387, Loss: {'train/loss': 0.604319155216217, 'train/grad_norm_pre_clip': 11.72163200378418, 'timing/data': 0.0011181309819221497, 'timing/model': 1.101919368840754, 'learning_rate/qwen_vl_interface': 7.649306540160928e-06, 'learning_rate/action_model': 7.649306540160927e-05, 'epoch': 0.87}) +[2026-06-17 21:03:08,944][starVLA.training.train_starvla][INFO] - Step 2388, Loss: {'train/loss': 0.9480146169662476, 'train/grad_norm_pre_clip': 16.00019645690918, 'timing/data': 0.0009409934282302856, 'timing/model': 1.4299695380032063, 'learning_rate/qwen_vl_interface': 7.647439779731959e-06, 'learning_rate/action_model': 7.647439779731958e-05, 'epoch': 0.87}) +[2026-06-17 21:03:13,119][starVLA.training.train_starvla][INFO] - Step 2389, Loss: {'train/loss': 0.712247371673584, 'train/grad_norm_pre_clip': 14.764490127563477, 'timing/data': 0.0007439861074090004, 'timing/model': 1.066390079446137, 'learning_rate/qwen_vl_interface': 7.645572522311144e-06, 'learning_rate/action_model': 7.645572522311144e-05, 'epoch': 0.87}) +[2026-06-17 21:03:22,247][starVLA.training.train_starvla][INFO] - Step 2390, Loss: {'train/loss': 0.6122456192970276, 'train/grad_norm_pre_clip': 15.014822006225586, 'timing/data': 0.0009250389412045479, 'timing/model': 1.132907317019999, 'learning_rate/qwen_vl_interface': 7.643704768285568e-06, 'learning_rate/action_model': 7.643704768285568e-05, 'epoch': 0.88}) +[2026-06-17 21:03:26,662][starVLA.training.train_starvla][INFO] - Step 2391, Loss: {'train/loss': 0.6398694515228271, 'train/grad_norm_pre_clip': 15.060903549194336, 'timing/data': 0.0010991701856255531, 'timing/model': 1.0488916775211692, 'learning_rate/qwen_vl_interface': 7.641836518042417e-06, 'learning_rate/action_model': 7.641836518042416e-05, 'epoch': 0.88}) +[2026-06-17 21:03:31,217][starVLA.training.train_starvla][INFO] - Step 2392, Loss: {'train/loss': 0.6077788472175598, 'train/grad_norm_pre_clip': 13.782999992370605, 'timing/data': 0.0011023478582501411, 'timing/model': 1.1007361561059952, 'learning_rate/qwen_vl_interface': 7.639967771968983e-06, 'learning_rate/action_model': 7.639967771968983e-05, 'epoch': 0.88}) +[2026-06-17 21:03:43,983][starVLA.training.train_starvla][INFO] - Step 2393, Loss: {'train/loss': 0.7888301014900208, 'train/grad_norm_pre_clip': 14.390944480895996, 'timing/data': 0.011626560240983963, 'timing/model': 5.084753951989114, 'learning_rate/qwen_vl_interface': 7.638098530452657e-06, 'learning_rate/action_model': 7.638098530452658e-05, 'epoch': 0.88}) +[2026-06-17 21:03:48,279][starVLA.training.train_starvla][INFO] - Step 2394, Loss: {'train/loss': 0.5873976349830627, 'train/grad_norm_pre_clip': 14.317062377929688, 'timing/data': 0.01559476088732481, 'timing/model': 1.1368289338424802, 'learning_rate/qwen_vl_interface': 7.636228793880935e-06, 'learning_rate/action_model': 7.636228793880935e-05, 'epoch': 0.88}) +[2026-06-17 21:03:53,201][starVLA.training.train_starvla][INFO] - Step 2395, Loss: {'train/loss': 0.8723393678665161, 'train/grad_norm_pre_clip': 14.398106575012207, 'timing/data': 0.019302736036479473, 'timing/model': 1.269676255993545, 'learning_rate/qwen_vl_interface': 7.63435856264142e-06, 'learning_rate/action_model': 7.63435856264142e-05, 'epoch': 0.88}) +[2026-06-17 21:03:57,799][starVLA.training.train_starvla][INFO] - Step 2396, Loss: {'train/loss': 0.7793688178062439, 'train/grad_norm_pre_clip': 18.802915573120117, 'timing/data': 0.011624282225966454, 'timing/model': 1.0614702831953764, 'learning_rate/qwen_vl_interface': 7.632487837121806e-06, 'learning_rate/action_model': 7.632487837121806e-05, 'epoch': 0.88}) +[2026-06-17 21:04:02,647][starVLA.training.train_starvla][INFO] - Step 2397, Loss: {'train/loss': 0.6608343720436096, 'train/grad_norm_pre_clip': 11.644341468811035, 'timing/data': 0.012917125597596169, 'timing/model': 1.211238669231534, 'learning_rate/qwen_vl_interface': 7.630616617709902e-06, 'learning_rate/action_model': 7.630616617709902e-05, 'epoch': 0.88}) +[2026-06-17 21:04:06,926][starVLA.training.train_starvla][INFO] - Step 2398, Loss: {'train/loss': 1.0150851011276245, 'train/grad_norm_pre_clip': 19.26639175415039, 'timing/data': 0.012007948011159897, 'timing/model': 1.1424239231273532, 'learning_rate/qwen_vl_interface': 7.6287449047936095e-06, 'learning_rate/action_model': 7.628744904793609e-05, 'epoch': 0.88}) +[2026-06-17 21:04:14,761][starVLA.training.train_starvla][INFO] - Step 2399, Loss: {'train/loss': 0.5811690092086792, 'train/grad_norm_pre_clip': 12.13769245147705, 'timing/data': 0.011692306958138943, 'timing/model': 1.030218381434679, 'learning_rate/qwen_vl_interface': 7.6268726987609384e-06, 'learning_rate/action_model': 7.626872698760939e-05, 'epoch': 0.88}) +[2026-06-17 21:07:24,279][starVLA.training.train_starvla][INFO] - Step 2400, Loss: {'train/loss': 0.7360916137695312, 'train/grad_norm_pre_clip': 18.640226364135742, 'eval/loss': 0.7375178337097168, 'eval/latency_2/loss': 0.7375178337097168, 'eval/demon_fire_precision': 0.8695321414986686, 'eval/demon_fire_recall': 0.9564853556485355, 'eval/demon_fire_f1': 0.9109384339509863, 'eval/demon_fire_rate_model': 0.73149693934335, 'eval/demon_fire_rate_teacher': 0.6649972175848636, 'eval/demon_left_precision': 0.8231981981981982, 'eval/demon_left_recall': 0.7694736842105263, 'eval/demon_left_f1': 0.7954298150163222, 'eval/demon_left_rate_model': 0.24707846410684475, 'eval/demon_left_rate_teacher': 0.2643294379521425, 'eval/demon_right_precision': 0.8041804180418042, 'eval/demon_right_recall': 0.8259887005649718, 'eval/demon_right_f1': 0.8149386845039018, 'eval/demon_right_rate_model': 0.25292153589315525, 'eval/demon_right_rate_teacher': 0.24624373956594323, 'eval/demon_move_f1': 0.805184249760112, 'eval/demon_cc_f1': 0.876405870306586, 'eval/latency_2/demon_fire_precision': 0.8695321414986686, 'eval/latency_2/demon_fire_recall': 0.9564853556485355, 'eval/latency_2/demon_fire_f1': 0.9109384339509863, 'eval/latency_2/demon_fire_rate_model': 0.73149693934335, 'eval/latency_2/demon_fire_rate_teacher': 0.6649972175848636, 'eval/latency_2/demon_left_precision': 0.8231981981981982, 'eval/latency_2/demon_left_recall': 0.7694736842105263, 'eval/latency_2/demon_left_f1': 0.7954298150163222, 'eval/latency_2/demon_left_rate_model': 0.24707846410684475, 'eval/latency_2/demon_left_rate_teacher': 0.2643294379521425, 'eval/latency_2/demon_right_precision': 0.8041804180418042, 'eval/latency_2/demon_right_recall': 0.8259887005649718, 'eval/latency_2/demon_right_f1': 0.8149386845039018, 'eval/latency_2/demon_right_rate_model': 0.25292153589315525, 'eval/latency_2/demon_right_rate_teacher': 0.24624373956594323, 'eval/latency_2/demon_move_f1': 0.805184249760112, 'eval/latency_2/demon_cc_f1': 0.876405870306586, 'timing/data': 0.015432968735694885, 'timing/model': 1.0933162281289697, 'learning_rate/qwen_vl_interface': 7.625e-06, 'learning_rate/action_model': 7.625e-05, 'epoch': 0.88}) +[2026-06-17 21:07:24,293][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_2400_state +[2026-06-17 21:07:24,301][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 21:07:33,126][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_2400_state/model.safetensors +[2026-06-17 21:07:47,872][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_2400_state/optimizer.bin +[2026-06-17 21:07:47,873][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_2400_state/sampler.bin +[2026-06-17 21:07:47,873][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_2400_state/sampler_1.bin +[2026-06-17 21:07:47,876][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_2400_state/random_states_0.pkl +[2026-06-17 21:07:47,880][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2400 +[2026-06-17 21:07:47,880][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 21:07:47,897][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 21:07:52,676][starVLA.training.train_starvla][INFO] - Step 2401, Loss: {'train/loss': 0.6411458253860474, 'train/grad_norm_pre_clip': 15.65797233581543, 'timing/data': 0.011768399737775326, 'timing/model': 1.1050705006346107, 'learning_rate/qwen_vl_interface': 7.6231268088990075e-06, 'learning_rate/action_model': 7.623126808899008e-05, 'epoch': 0.88}) +[2026-06-17 21:08:01,901][starVLA.training.train_starvla][INFO] - Step 2402, Loss: {'train/loss': 0.47407427430152893, 'train/grad_norm_pre_clip': 16.739572525024414, 'timing/data': 0.01155280601233244, 'timing/model': 4.3128099497407675, 'learning_rate/qwen_vl_interface': 7.621253125846273e-06, 'learning_rate/action_model': 7.621253125846273e-05, 'epoch': 0.88}) +[2026-06-17 21:08:06,516][starVLA.training.train_starvla][INFO] - Step 2403, Loss: {'train/loss': 0.6168460249900818, 'train/grad_norm_pre_clip': 21.862449645996094, 'timing/data': 0.01313607394695282, 'timing/model': 1.0855990471318364, 'learning_rate/qwen_vl_interface': 7.619378951230213e-06, 'learning_rate/action_model': 7.619378951230213e-05, 'epoch': 0.88}) +[2026-06-17 21:08:11,286][starVLA.training.train_starvla][INFO] - Step 2404, Loss: {'train/loss': 0.6857472658157349, 'train/grad_norm_pre_clip': 16.177038192749023, 'timing/data': 0.012109718285501003, 'timing/model': 1.5156476153060794, 'learning_rate/qwen_vl_interface': 7.617504285439346e-06, 'learning_rate/action_model': 7.617504285439346e-05, 'epoch': 0.88}) +[2026-06-17 21:08:15,576][starVLA.training.train_starvla][INFO] - Step 2405, Loss: {'train/loss': 0.6143731474876404, 'train/grad_norm_pre_clip': 11.84697437286377, 'timing/data': 0.012672730721533298, 'timing/model': 1.086247150786221, 'learning_rate/qwen_vl_interface': 7.615629128862295e-06, 'learning_rate/action_model': 7.615629128862295e-05, 'epoch': 0.88}) +[2026-06-17 21:08:20,097][starVLA.training.train_starvla][INFO] - Step 2406, Loss: {'train/loss': 0.5913283228874207, 'train/grad_norm_pre_clip': 11.779923439025879, 'timing/data': 0.01274253148585558, 'timing/model': 1.0757158398628235, 'learning_rate/qwen_vl_interface': 7.61375348188778e-06, 'learning_rate/action_model': 7.61375348188778e-05, 'epoch': 0.88}) +[2026-06-17 21:08:24,951][starVLA.training.train_starvla][INFO] - Step 2407, Loss: {'train/loss': 0.7652516961097717, 'train/grad_norm_pre_clip': 12.121826171875, 'timing/data': 0.011586385779082775, 'timing/model': 1.0205272287130356, 'learning_rate/qwen_vl_interface': 7.611877344904624e-06, 'learning_rate/action_model': 7.611877344904624e-05, 'epoch': 0.88}) +[2026-06-17 21:08:29,994][starVLA.training.train_starvla][INFO] - Step 2408, Loss: {'train/loss': 0.6931810975074768, 'train/grad_norm_pre_clip': 16.32686996459961, 'timing/data': 0.056868902407586575, 'timing/model': 1.075058514252305, 'learning_rate/qwen_vl_interface': 7.610000718301753e-06, 'learning_rate/action_model': 7.610000718301753e-05, 'epoch': 0.88}) +[2026-06-17 21:08:35,239][starVLA.training.train_starvla][INFO] - Step 2409, Loss: {'train/loss': 0.5245550274848938, 'train/grad_norm_pre_clip': 13.608415603637695, 'timing/data': 0.04085767641663551, 'timing/model': 1.3598343823105097, 'learning_rate/qwen_vl_interface': 7.608123602468195e-06, 'learning_rate/action_model': 7.608123602468194e-05, 'epoch': 0.88}) +[2026-06-17 21:08:42,510][starVLA.training.train_starvla][INFO] - Step 2410, Loss: {'train/loss': 0.4934794306755066, 'train/grad_norm_pre_clip': 11.510254859924316, 'timing/data': 0.048186768777668476, 'timing/model': 1.0849953824654222, 'learning_rate/qwen_vl_interface': 7.606245997793074e-06, 'learning_rate/action_model': 7.606245997793075e-05, 'epoch': 0.88}) +[2026-06-17 21:08:47,166][starVLA.training.train_starvla][INFO] - Step 2411, Loss: {'train/loss': 1.175002932548523, 'train/grad_norm_pre_clip': 16.916627883911133, 'timing/data': 0.0391089441254735, 'timing/model': 1.0466828579083085, 'learning_rate/qwen_vl_interface': 7.604367904665625e-06, 'learning_rate/action_model': 7.604367904665624e-05, 'epoch': 0.88}) +[2026-06-17 21:08:51,638][starVLA.training.train_starvla][INFO] - Step 2412, Loss: {'train/loss': 0.74570232629776, 'train/grad_norm_pre_clip': 14.536208152770996, 'timing/data': 0.04011531639844179, 'timing/model': 1.0388477826490998, 'learning_rate/qwen_vl_interface': 7.602489323475174e-06, 'learning_rate/action_model': 7.602489323475174e-05, 'epoch': 0.88}) +[2026-06-17 21:08:56,468][starVLA.training.train_starvla][INFO] - Step 2413, Loss: {'train/loss': 0.9385864734649658, 'train/grad_norm_pre_clip': 29.701574325561523, 'timing/data': 0.04340770281851292, 'timing/model': 1.1199742909520864, 'learning_rate/qwen_vl_interface': 7.600610254611155e-06, 'learning_rate/action_model': 7.600610254611155e-05, 'epoch': 0.88}) +[2026-06-17 21:09:01,106][starVLA.training.train_starvla][INFO] - Step 2414, Loss: {'train/loss': 0.34791678190231323, 'train/grad_norm_pre_clip': 9.235193252563477, 'timing/data': 0.037980168126523495, 'timing/model': 1.4122320367023349, 'learning_rate/qwen_vl_interface': 7.5987306984631025e-06, 'learning_rate/action_model': 7.598730698463102e-05, 'epoch': 0.88}) +[2026-06-17 21:09:05,416][starVLA.training.train_starvla][INFO] - Step 2415, Loss: {'train/loss': 0.7696523070335388, 'train/grad_norm_pre_clip': 13.14050579071045, 'timing/data': 0.045016899704933167, 'timing/model': 1.0676953736692667, 'learning_rate/qwen_vl_interface': 7.596850655420647e-06, 'learning_rate/action_model': 7.596850655420646e-05, 'epoch': 0.88}) +[2026-06-17 21:09:12,607][starVLA.training.train_starvla][INFO] - Step 2416, Loss: {'train/loss': 0.639533281326294, 'train/grad_norm_pre_clip': 14.984244346618652, 'timing/data': 0.1089312992990017, 'timing/model': 1.381376982666552, 'learning_rate/qwen_vl_interface': 7.5949701258735265e-06, 'learning_rate/action_model': 7.594970125873526e-05, 'epoch': 0.88}) +[2026-06-17 21:09:20,657][starVLA.training.train_starvla][INFO] - Step 2417, Loss: {'train/loss': 0.5348681211471558, 'train/grad_norm_pre_clip': 13.205093383789062, 'timing/data': 0.024651167914271355, 'timing/model': 1.1021170234307647, 'learning_rate/qwen_vl_interface': 7.5930891102115735e-06, 'learning_rate/action_model': 7.593089110211573e-05, 'epoch': 0.89}) +[2026-06-17 21:09:25,326][starVLA.training.train_starvla][INFO] - Step 2418, Loss: {'train/loss': 0.6748200058937073, 'train/grad_norm_pre_clip': 13.943106651306152, 'timing/data': 0.03702163230627775, 'timing/model': 1.1839047502726316, 'learning_rate/qwen_vl_interface': 7.59120760882473e-06, 'learning_rate/action_model': 7.59120760882473e-05, 'epoch': 0.89}) +[2026-06-17 21:09:30,541][starVLA.training.train_starvla][INFO] - Step 2419, Loss: {'train/loss': 0.6354530453681946, 'train/grad_norm_pre_clip': 19.54041862487793, 'timing/data': 0.04437100701034069, 'timing/model': 1.6523631839081645, 'learning_rate/qwen_vl_interface': 7.5893256221030275e-06, 'learning_rate/action_model': 7.589325622103027e-05, 'epoch': 0.89}) +[2026-06-17 21:09:34,829][starVLA.training.train_starvla][INFO] - Step 2420, Loss: {'train/loss': 0.7228516936302185, 'train/grad_norm_pre_clip': 12.16392993927002, 'timing/data': 0.03588396497070789, 'timing/model': 1.1604894418269396, 'learning_rate/qwen_vl_interface': 7.587443150436608e-06, 'learning_rate/action_model': 7.587443150436608e-05, 'epoch': 0.89}) +[2026-06-17 21:09:45,394][starVLA.training.train_starvla][INFO] - Step 2421, Loss: {'train/loss': 0.5125159025192261, 'train/grad_norm_pre_clip': 13.246529579162598, 'timing/data': 0.0394363384693861, 'timing/model': 4.156391806900501, 'learning_rate/qwen_vl_interface': 7.585560194215706e-06, 'learning_rate/action_model': 7.585560194215706e-05, 'epoch': 0.89}) +[2026-06-17 21:09:53,063][starVLA.training.train_starvla][INFO] - Step 2422, Loss: {'train/loss': 0.7105432152748108, 'train/grad_norm_pre_clip': 17.329885482788086, 'timing/data': 0.03209769818931818, 'timing/model': 1.0740377884358168, 'learning_rate/qwen_vl_interface': 7.583676753830665e-06, 'learning_rate/action_model': 7.583676753830664e-05, 'epoch': 0.89}) +[2026-06-17 21:09:57,612][starVLA.training.train_starvla][INFO] - Step 2423, Loss: {'train/loss': 0.8581631779670715, 'train/grad_norm_pre_clip': 16.513286590576172, 'timing/data': 0.029589936137199402, 'timing/model': 1.3627195041626692, 'learning_rate/qwen_vl_interface': 7.58179282967192e-06, 'learning_rate/action_model': 7.58179282967192e-05, 'epoch': 0.89}) +[2026-06-17 21:10:02,094][starVLA.training.train_starvla][INFO] - Step 2424, Loss: {'train/loss': 0.7307682037353516, 'train/grad_norm_pre_clip': 14.991812705993652, 'timing/data': 0.044311048462986946, 'timing/model': 1.4211975680664182, 'learning_rate/qwen_vl_interface': 7.579908422130013e-06, 'learning_rate/action_model': 7.579908422130014e-05, 'epoch': 0.89}) +[2026-06-17 21:10:06,242][starVLA.training.train_starvla][INFO] - Step 2425, Loss: {'train/loss': 0.7501804828643799, 'train/grad_norm_pre_clip': 16.996015548706055, 'timing/data': 0.031079895794391632, 'timing/model': 1.0279035801067948, 'learning_rate/qwen_vl_interface': 7.578023531595583e-06, 'learning_rate/action_model': 7.578023531595583e-05, 'epoch': 0.89}) +[2026-06-17 21:10:10,813][starVLA.training.train_starvla][INFO] - Step 2426, Loss: {'train/loss': 0.4527284801006317, 'train/grad_norm_pre_clip': 10.298698425292969, 'timing/data': 0.03777635097503662, 'timing/model': 1.1381371458992362, 'learning_rate/qwen_vl_interface': 7.576138158459369e-06, 'learning_rate/action_model': 7.576138158459368e-05, 'epoch': 0.89}) +[2026-06-17 21:10:16,306][starVLA.training.train_starvla][INFO] - Step 2427, Loss: {'train/loss': 0.6733364462852478, 'train/grad_norm_pre_clip': 14.055645942687988, 'timing/data': 0.03303731605410576, 'timing/model': 1.080466072075069, 'learning_rate/qwen_vl_interface': 7.574252303112213e-06, 'learning_rate/action_model': 7.574252303112212e-05, 'epoch': 0.89}) +[2026-06-17 21:10:26,397][starVLA.training.train_starvla][INFO] - Step 2428, Loss: {'train/loss': 0.8728510737419128, 'train/grad_norm_pre_clip': 15.45972728729248, 'timing/data': 0.031590803526341915, 'timing/model': 1.2027131272479892, 'learning_rate/qwen_vl_interface': 7.572365965945053e-06, 'learning_rate/action_model': 7.572365965945052e-05, 'epoch': 0.89}) +[2026-06-17 21:10:30,870][starVLA.training.train_starvla][INFO] - Step 2429, Loss: {'train/loss': 0.5849741101264954, 'train/grad_norm_pre_clip': 12.701127052307129, 'timing/data': 0.031961035914719105, 'timing/model': 1.3430173648521304, 'learning_rate/qwen_vl_interface': 7.570479147348929e-06, 'learning_rate/action_model': 7.570479147348929e-05, 'epoch': 0.89}) +[2026-06-17 21:10:42,959][starVLA.training.train_starvla][INFO] - Step 2430, Loss: {'train/loss': 0.5115225315093994, 'train/grad_norm_pre_clip': 11.643060684204102, 'timing/data': 0.04414971265941858, 'timing/model': 1.217305501922965, 'learning_rate/qwen_vl_interface': 7.568591847714983e-06, 'learning_rate/action_model': 7.568591847714982e-05, 'epoch': 0.89}) +[2026-06-17 21:10:47,331][starVLA.training.train_starvla][INFO] - Step 2431, Loss: {'train/loss': 0.5327205657958984, 'train/grad_norm_pre_clip': 14.287325859069824, 'timing/data': 0.034676119685173035, 'timing/model': 1.099253517575562, 'learning_rate/qwen_vl_interface': 7.5667040674344495e-06, 'learning_rate/action_model': 7.566704067434449e-05, 'epoch': 0.89}) +[2026-06-17 21:10:56,935][starVLA.training.train_starvla][INFO] - Step 2432, Loss: {'train/loss': 0.6251873970031738, 'train/grad_norm_pre_clip': 14.544085502624512, 'timing/data': 0.17003689147531986, 'timing/model': 1.0994300646707416, 'learning_rate/qwen_vl_interface': 7.564815806898669e-06, 'learning_rate/action_model': 7.564815806898668e-05, 'epoch': 0.89}) +[2026-06-17 21:11:06,584][starVLA.training.train_starvla][INFO] - Step 2433, Loss: {'train/loss': 0.6266775131225586, 'train/grad_norm_pre_clip': 13.465171813964844, 'timing/data': 0.0253722770139575, 'timing/model': 5.44059969112277, 'learning_rate/qwen_vl_interface': 7.562927066499084e-06, 'learning_rate/action_model': 7.562927066499083e-05, 'epoch': 0.89}) +[2026-06-17 21:11:16,894][starVLA.training.train_starvla][INFO] - Step 2434, Loss: {'train/loss': 0.3890439569950104, 'train/grad_norm_pre_clip': 9.772821426391602, 'timing/data': 0.112006233073771, 'timing/model': 1.4363225381821394, 'learning_rate/qwen_vl_interface': 7.5610378466272274e-06, 'learning_rate/action_model': 7.561037846627227e-05, 'epoch': 0.89}) +[2026-06-17 21:11:26,920][starVLA.training.train_starvla][INFO] - Step 2435, Loss: {'train/loss': 0.48675772547721863, 'train/grad_norm_pre_clip': 12.010457992553711, 'timing/data': 0.038393912836909294, 'timing/model': 2.626200362108648, 'learning_rate/qwen_vl_interface': 7.5591481476747384e-06, 'learning_rate/action_model': 7.559148147674739e-05, 'epoch': 0.89}) +[2026-06-17 21:11:36,217][starVLA.training.train_starvla][INFO] - Step 2436, Loss: {'train/loss': 0.8200437426567078, 'train/grad_norm_pre_clip': 17.46014404296875, 'timing/data': 0.06083924230188131, 'timing/model': 1.1676351558417082, 'learning_rate/qwen_vl_interface': 7.557257970033355e-06, 'learning_rate/action_model': 7.557257970033354e-05, 'epoch': 0.89}) +[2026-06-17 21:11:41,044][starVLA.training.train_starvla][INFO] - Step 2437, Loss: {'train/loss': 0.7331068515777588, 'train/grad_norm_pre_clip': 15.49932861328125, 'timing/data': 0.050731114111840725, 'timing/model': 1.1821673391386867, 'learning_rate/qwen_vl_interface': 7.55536731409491e-06, 'learning_rate/action_model': 7.55536731409491e-05, 'epoch': 0.89}) +[2026-06-17 21:11:47,038][starVLA.training.train_starvla][INFO] - Step 2438, Loss: {'train/loss': 0.5874849557876587, 'train/grad_norm_pre_clip': 13.318160057067871, 'timing/data': 0.043874556198716164, 'timing/model': 2.6947225956246257, 'learning_rate/qwen_vl_interface': 7.553476180251341e-06, 'learning_rate/action_model': 7.55347618025134e-05, 'epoch': 0.89}) +[2026-06-17 21:11:53,666][starVLA.training.train_starvla][INFO] - Step 2439, Loss: {'train/loss': 0.6419168710708618, 'train/grad_norm_pre_clip': 15.078410148620605, 'timing/data': 0.08592230826616287, 'timing/model': 1.431773690506816, 'learning_rate/qwen_vl_interface': 7.551584568894681e-06, 'learning_rate/action_model': 7.551584568894681e-05, 'epoch': 0.89}) +[2026-06-17 21:11:58,542][starVLA.training.train_starvla][INFO] - Step 2440, Loss: {'train/loss': 0.5893204212188721, 'train/grad_norm_pre_clip': 12.006440162658691, 'timing/data': 0.04606068693101406, 'timing/model': 1.0231664003804326, 'learning_rate/qwen_vl_interface': 7.5496924804170635e-06, 'learning_rate/action_model': 7.549692480417063e-05, 'epoch': 0.89}) +[2026-06-17 21:12:03,142][starVLA.training.train_starvla][INFO] - Step 2441, Loss: {'train/loss': 0.9182575941085815, 'train/grad_norm_pre_clip': 18.859952926635742, 'timing/data': 0.04442847892642021, 'timing/model': 1.0291581442579627, 'learning_rate/qwen_vl_interface': 7.5477999152107195e-06, 'learning_rate/action_model': 7.54779991521072e-05, 'epoch': 0.89}) +[2026-06-17 21:12:07,602][starVLA.training.train_starvla][INFO] - Step 2442, Loss: {'train/loss': 0.43032121658325195, 'train/grad_norm_pre_clip': 13.079890251159668, 'timing/data': 0.05519583262503147, 'timing/model': 1.049525453709066, 'learning_rate/qwen_vl_interface': 7.545906873667982e-06, 'learning_rate/action_model': 7.545906873667981e-05, 'epoch': 0.89}) +[2026-06-17 21:12:11,966][starVLA.training.train_starvla][INFO] - Step 2443, Loss: {'train/loss': 0.5150113105773926, 'train/grad_norm_pre_clip': 12.085956573486328, 'timing/data': 0.044586120173335075, 'timing/model': 1.0262634018436074, 'learning_rate/qwen_vl_interface': 7.544013356181277e-06, 'learning_rate/action_model': 7.544013356181277e-05, 'epoch': 0.89}) +[2026-06-17 21:12:16,357][starVLA.training.train_starvla][INFO] - Step 2444, Loss: {'train/loss': 0.8258525133132935, 'train/grad_norm_pre_clip': 14.650097846984863, 'timing/data': 0.047515721060335636, 'timing/model': 1.370569839142263, 'learning_rate/qwen_vl_interface': 7.542119363143135e-06, 'learning_rate/action_model': 7.542119363143136e-05, 'epoch': 0.89}) +[2026-06-17 21:12:20,365][starVLA.training.train_starvla][INFO] - Step 2445, Loss: {'train/loss': 0.4848997890949249, 'train/grad_norm_pre_clip': 11.872421264648438, 'timing/data': 0.04216532036662102, 'timing/model': 1.0333411823958158, 'learning_rate/qwen_vl_interface': 7.540224894946184e-06, 'learning_rate/action_model': 7.540224894946184e-05, 'epoch': 0.9}) +[2026-06-17 21:12:24,822][starVLA.training.train_starvla][INFO] - Step 2446, Loss: {'train/loss': 0.5556003451347351, 'train/grad_norm_pre_clip': 12.801141738891602, 'timing/data': 0.03996897488832474, 'timing/model': 1.0342676900327206, 'learning_rate/qwen_vl_interface': 7.538329951983148e-06, 'learning_rate/action_model': 7.538329951983147e-05, 'epoch': 0.9}) +[2026-06-17 21:12:29,192][starVLA.training.train_starvla][INFO] - Step 2447, Loss: {'train/loss': 0.4935874044895172, 'train/grad_norm_pre_clip': 10.428693771362305, 'timing/data': 0.047991896979510784, 'timing/model': 1.0589833110570908, 'learning_rate/qwen_vl_interface': 7.536434534646851e-06, 'learning_rate/action_model': 7.53643453464685e-05, 'epoch': 0.9}) +[2026-06-17 21:12:33,688][starVLA.training.train_starvla][INFO] - Step 2448, Loss: {'train/loss': 0.6878064870834351, 'train/grad_norm_pre_clip': 14.61860466003418, 'timing/data': 0.04023602232336998, 'timing/model': 1.131871409714222, 'learning_rate/qwen_vl_interface': 7.534538643330213e-06, 'learning_rate/action_model': 7.534538643330214e-05, 'epoch': 0.9}) +[2026-06-17 21:12:38,114][starVLA.training.train_starvla][INFO] - Step 2449, Loss: {'train/loss': 0.5792487859725952, 'train/grad_norm_pre_clip': 16.748952865600586, 'timing/data': 0.046686457470059395, 'timing/model': 1.3264763038605452, 'learning_rate/qwen_vl_interface': 7.5326422784262566e-06, 'learning_rate/action_model': 7.532642278426256e-05, 'epoch': 0.9}) +[2026-06-17 21:12:42,557][starVLA.training.train_starvla][INFO] - Step 2450, Loss: {'train/loss': 0.7926082015037537, 'train/grad_norm_pre_clip': 13.854416847229004, 'timing/data': 0.05048598162829876, 'timing/model': 1.0268124528229237, 'learning_rate/qwen_vl_interface': 7.5307454403281e-06, 'learning_rate/action_model': 7.5307454403281e-05, 'epoch': 0.9}) +[2026-06-17 21:12:46,956][starVLA.training.train_starvla][INFO] - Step 2451, Loss: {'train/loss': 0.6147429347038269, 'train/grad_norm_pre_clip': 15.106049537658691, 'timing/data': 0.03726731147617102, 'timing/model': 1.0305202836170793, 'learning_rate/qwen_vl_interface': 7.528848129428958e-06, 'learning_rate/action_model': 7.528848129428958e-05, 'epoch': 0.9}) +[2026-06-17 21:12:51,417][starVLA.training.train_starvla][INFO] - Step 2452, Loss: {'train/loss': 0.8099570274353027, 'train/grad_norm_pre_clip': 22.730066299438477, 'timing/data': 0.04700987692922354, 'timing/model': 1.0495793269947171, 'learning_rate/qwen_vl_interface': 7.526950346122149e-06, 'learning_rate/action_model': 7.526950346122148e-05, 'epoch': 0.9}) +[2026-06-17 21:12:55,689][starVLA.training.train_starvla][INFO] - Step 2453, Loss: {'train/loss': 0.7364040613174438, 'train/grad_norm_pre_clip': 13.615528106689453, 'timing/data': 0.038769783452153206, 'timing/model': 1.0380282485857606, 'learning_rate/qwen_vl_interface': 7.525052090801083e-06, 'learning_rate/action_model': 7.525052090801083e-05, 'epoch': 0.9}) +[2026-06-17 21:13:00,124][starVLA.training.train_starvla][INFO] - Step 2454, Loss: {'train/loss': 0.643448531627655, 'train/grad_norm_pre_clip': 15.711132049560547, 'timing/data': 0.04407769162207842, 'timing/model': 1.3709707697853446, 'learning_rate/qwen_vl_interface': 7.523153363859269e-06, 'learning_rate/action_model': 7.523153363859269e-05, 'epoch': 0.9}) +[2026-06-17 21:13:07,379][starVLA.training.train_starvla][INFO] - Step 2455, Loss: {'train/loss': 0.7490171790122986, 'train/grad_norm_pre_clip': 15.124253273010254, 'timing/data': 0.09214852098375559, 'timing/model': 3.947519921697676, 'learning_rate/qwen_vl_interface': 7.521254165690317e-06, 'learning_rate/action_model': 7.521254165690317e-05, 'epoch': 0.9}) +[2026-06-17 21:13:14,821][starVLA.training.train_starvla][INFO] - Step 2456, Loss: {'train/loss': 0.979729175567627, 'train/grad_norm_pre_clip': 16.193344116210938, 'timing/data': 0.011343524791300297, 'timing/model': 3.619535408914089, 'learning_rate/qwen_vl_interface': 7.519354496687933e-06, 'learning_rate/action_model': 7.519354496687933e-05, 'epoch': 0.9}) +[2026-06-17 21:13:23,016][starVLA.training.train_starvla][INFO] - Step 2457, Loss: {'train/loss': 0.9775770902633667, 'train/grad_norm_pre_clip': 17.641307830810547, 'timing/data': 0.03031579591333866, 'timing/model': 1.1300004133954644, 'learning_rate/qwen_vl_interface': 7.517454357245918e-06, 'learning_rate/action_model': 7.517454357245917e-05, 'epoch': 0.9}) +[2026-06-17 21:13:29,613][starVLA.training.train_starvla][INFO] - Step 2458, Loss: {'train/loss': 0.8923126459121704, 'train/grad_norm_pre_clip': 23.853553771972656, 'timing/data': 0.04015587829053402, 'timing/model': 1.471409047022462, 'learning_rate/qwen_vl_interface': 7.515553747758176e-06, 'learning_rate/action_model': 7.515553747758176e-05, 'epoch': 0.9}) +[2026-06-17 21:13:38,509][starVLA.training.train_starvla][INFO] - Step 2459, Loss: {'train/loss': 0.6986211538314819, 'train/grad_norm_pre_clip': 12.65759563446045, 'timing/data': 0.02450631931424141, 'timing/model': 2.014025886543095, 'learning_rate/qwen_vl_interface': 7.513652668618705e-06, 'learning_rate/action_model': 7.513652668618705e-05, 'epoch': 0.9}) +[2026-06-17 21:13:43,790][starVLA.training.train_starvla][INFO] - Step 2460, Loss: {'train/loss': 0.6006937026977539, 'train/grad_norm_pre_clip': 15.303388595581055, 'timing/data': 0.01444652583450079, 'timing/model': 1.1396938813850284, 'learning_rate/qwen_vl_interface': 7.5117511202215994e-06, 'learning_rate/action_model': 7.5117511202216e-05, 'epoch': 0.9}) +[2026-06-17 21:13:50,669][starVLA.training.train_starvla][INFO] - Step 2461, Loss: {'train/loss': 0.7184322476387024, 'train/grad_norm_pre_clip': 15.421943664550781, 'timing/data': 0.01732679456472397, 'timing/model': 1.2375195669010282, 'learning_rate/qwen_vl_interface': 7.509849102961053e-06, 'learning_rate/action_model': 7.509849102961053e-05, 'epoch': 0.9}) +[2026-06-17 21:14:01,605][starVLA.training.train_starvla][INFO] - Step 2462, Loss: {'train/loss': 0.7680546045303345, 'train/grad_norm_pre_clip': 14.234503746032715, 'timing/data': 0.011644554324448109, 'timing/model': 1.1452763127163053, 'learning_rate/qwen_vl_interface': 7.507946617231357e-06, 'learning_rate/action_model': 7.507946617231356e-05, 'epoch': 0.9}) +[2026-06-17 21:14:07,800][starVLA.training.train_starvla][INFO] - Step 2463, Loss: {'train/loss': 0.6585752964019775, 'train/grad_norm_pre_clip': 11.709030151367188, 'timing/data': 0.012563465163111687, 'timing/model': 1.2512732055038214, 'learning_rate/qwen_vl_interface': 7.5060436634268965e-06, 'learning_rate/action_model': 7.506043663426896e-05, 'epoch': 0.9}) +[2026-06-17 21:14:13,320][starVLA.training.train_starvla][INFO] - Step 2464, Loss: {'train/loss': 0.6055970788002014, 'train/grad_norm_pre_clip': 13.265425682067871, 'timing/data': 0.012201633304357529, 'timing/model': 1.3857939951121807, 'learning_rate/qwen_vl_interface': 7.504140241942157e-06, 'learning_rate/action_model': 7.504140241942157e-05, 'epoch': 0.9}) +[2026-06-17 21:14:17,462][starVLA.training.train_starvla][INFO] - Step 2465, Loss: {'train/loss': 0.8193308711051941, 'train/grad_norm_pre_clip': 12.694074630737305, 'timing/data': 0.014448306523263454, 'timing/model': 1.0372602622956038, 'learning_rate/qwen_vl_interface': 7.502236353171718e-06, 'learning_rate/action_model': 7.502236353171718e-05, 'epoch': 0.9}) +[2026-06-17 21:14:22,287][starVLA.training.train_starvla][INFO] - Step 2466, Loss: {'train/loss': 0.8848572969436646, 'train/grad_norm_pre_clip': 15.002333641052246, 'timing/data': 0.015505963005125523, 'timing/model': 1.074701233766973, 'learning_rate/qwen_vl_interface': 7.500331997510262e-06, 'learning_rate/action_model': 7.500331997510262e-05, 'epoch': 0.9}) +[2026-06-17 21:14:26,940][starVLA.training.train_starvla][INFO] - Step 2467, Loss: {'train/loss': 0.752232551574707, 'train/grad_norm_pre_clip': 16.023807525634766, 'timing/data': 0.013470173813402653, 'timing/model': 1.0860894452780485, 'learning_rate/qwen_vl_interface': 7.498427175352561e-06, 'learning_rate/action_model': 7.49842717535256e-05, 'epoch': 0.9}) +[2026-06-17 21:14:31,533][starVLA.training.train_starvla][INFO] - Step 2468, Loss: {'train/loss': 0.7596360445022583, 'train/grad_norm_pre_clip': 20.267900466918945, 'timing/data': 0.012365313246846199, 'timing/model': 1.0644197259098291, 'learning_rate/qwen_vl_interface': 7.496521887093487e-06, 'learning_rate/action_model': 7.496521887093486e-05, 'epoch': 0.9}) +[2026-06-17 21:14:36,479][starVLA.training.train_starvla][INFO] - Step 2469, Loss: {'train/loss': 0.7185774445533752, 'train/grad_norm_pre_clip': 14.559414863586426, 'timing/data': 0.015606952831149101, 'timing/model': 1.7063752096146345, 'learning_rate/qwen_vl_interface': 7.494616133128008e-06, 'learning_rate/action_model': 7.494616133128008e-05, 'epoch': 0.9}) +[2026-06-17 21:14:40,556][starVLA.training.train_starvla][INFO] - Step 2470, Loss: {'train/loss': 0.711875319480896, 'train/grad_norm_pre_clip': 15.145578384399414, 'timing/data': 0.011706986464560032, 'timing/model': 1.057183237746358, 'learning_rate/qwen_vl_interface': 7.4927099138511895e-06, 'learning_rate/action_model': 7.492709913851189e-05, 'epoch': 0.9}) +[2026-06-17 21:14:45,278][starVLA.training.train_starvla][INFO] - Step 2471, Loss: {'train/loss': 0.7228167057037354, 'train/grad_norm_pre_clip': 12.892403602600098, 'timing/data': 0.011014995165169239, 'timing/model': 1.1450857669115067, 'learning_rate/qwen_vl_interface': 7.490803229658192e-06, 'learning_rate/action_model': 7.490803229658192e-05, 'epoch': 0.9}) +[2026-06-17 21:14:49,821][starVLA.training.train_starvla][INFO] - Step 2472, Loss: {'train/loss': 0.6362186074256897, 'train/grad_norm_pre_clip': 12.279006958007812, 'timing/data': 0.013773837126791477, 'timing/model': 1.0996818831190467, 'learning_rate/qwen_vl_interface': 7.488896080944274e-06, 'learning_rate/action_model': 7.488896080944273e-05, 'epoch': 0.91}) +[2026-06-17 21:14:54,283][starVLA.training.train_starvla][INFO] - Step 2473, Loss: {'train/loss': 0.4719124138355255, 'train/grad_norm_pre_clip': 11.744853973388672, 'timing/data': 0.012850325554609299, 'timing/model': 1.0298085119575262, 'learning_rate/qwen_vl_interface': 7.4869884681047896e-06, 'learning_rate/action_model': 7.486988468104789e-05, 'epoch': 0.91}) +[2026-06-17 21:14:58,729][starVLA.training.train_starvla][INFO] - Step 2474, Loss: {'train/loss': 0.6773369312286377, 'train/grad_norm_pre_clip': 13.27592658996582, 'timing/data': 0.014669319614768028, 'timing/model': 1.3275086907669902, 'learning_rate/qwen_vl_interface': 7.485080391535189e-06, 'learning_rate/action_model': 7.485080391535188e-05, 'epoch': 0.91}) +[2026-06-17 21:15:03,157][starVLA.training.train_starvla][INFO] - Step 2475, Loss: {'train/loss': 0.6149316430091858, 'train/grad_norm_pre_clip': 15.5321044921875, 'timing/data': 0.015720929950475693, 'timing/model': 1.098479044623673, 'learning_rate/qwen_vl_interface': 7.483171851631018e-06, 'learning_rate/action_model': 7.483171851631017e-05, 'epoch': 0.91}) +[2026-06-17 21:15:07,547][starVLA.training.train_starvla][INFO] - Step 2476, Loss: {'train/loss': 0.8055019974708557, 'train/grad_norm_pre_clip': 16.723936080932617, 'timing/data': 0.010971265845000744, 'timing/model': 1.06084209587425, 'learning_rate/qwen_vl_interface': 7.481262848787919e-06, 'learning_rate/action_model': 7.481262848787919e-05, 'epoch': 0.91}) +[2026-06-17 21:15:13,552][starVLA.training.train_starvla][INFO] - Step 2477, Loss: {'train/loss': 0.5796473622322083, 'train/grad_norm_pre_clip': 13.739514350891113, 'timing/data': 0.011732677929103374, 'timing/model': 2.497913694009185, 'learning_rate/qwen_vl_interface': 7.479353383401631e-06, 'learning_rate/action_model': 7.47935338340163e-05, 'epoch': 0.91}) +[2026-06-17 21:15:18,035][starVLA.training.train_starvla][INFO] - Step 2478, Loss: {'train/loss': 0.41014957427978516, 'train/grad_norm_pre_clip': 10.573878288269043, 'timing/data': 0.014117810875177383, 'timing/model': 1.365970915183425, 'learning_rate/qwen_vl_interface': 7.477443455867987e-06, 'learning_rate/action_model': 7.477443455867988e-05, 'epoch': 0.91}) +[2026-06-17 21:15:29,086][starVLA.training.train_starvla][INFO] - Step 2479, Loss: {'train/loss': 0.7328175902366638, 'train/grad_norm_pre_clip': 14.322102546691895, 'timing/data': 0.01135675236582756, 'timing/model': 5.171759442426264, 'learning_rate/qwen_vl_interface': 7.475533066582917e-06, 'learning_rate/action_model': 7.475533066582918e-05, 'epoch': 0.91}) +[2026-06-17 21:15:33,373][starVLA.training.train_starvla][INFO] - Step 2480, Loss: {'train/loss': 0.547697901725769, 'train/grad_norm_pre_clip': 14.995965003967285, 'timing/data': 0.011231377720832825, 'timing/model': 1.0481412773951888, 'learning_rate/qwen_vl_interface': 7.47362221594245e-06, 'learning_rate/action_model': 7.47362221594245e-05, 'epoch': 0.91}) +[2026-06-17 21:15:38,041][starVLA.training.train_starvla][INFO] - Step 2481, Loss: {'train/loss': 0.5624322295188904, 'train/grad_norm_pre_clip': 12.550430297851562, 'timing/data': 0.014076693914830685, 'timing/model': 1.1388702932745218, 'learning_rate/qwen_vl_interface': 7.4717109043427025e-06, 'learning_rate/action_model': 7.471710904342702e-05, 'epoch': 0.91}) +[2026-06-17 21:15:42,834][starVLA.training.train_starvla][INFO] - Step 2482, Loss: {'train/loss': 0.5391061305999756, 'train/grad_norm_pre_clip': 12.443763732910156, 'timing/data': 0.011652535758912563, 'timing/model': 1.106090996414423, 'learning_rate/qwen_vl_interface': 7.469799132179894e-06, 'learning_rate/action_model': 7.469799132179895e-05, 'epoch': 0.91}) +[2026-06-17 21:15:47,349][starVLA.training.train_starvla][INFO] - Step 2483, Loss: {'train/loss': 0.5372498035430908, 'train/grad_norm_pre_clip': 12.119999885559082, 'timing/data': 0.015425383113324642, 'timing/model': 1.0446674032136798, 'learning_rate/qwen_vl_interface': 7.467886899850338e-06, 'learning_rate/action_model': 7.467886899850338e-05, 'epoch': 0.91}) +[2026-06-17 21:15:51,804][starVLA.training.train_starvla][INFO] - Step 2484, Loss: {'train/loss': 0.6284905076026917, 'train/grad_norm_pre_clip': 10.836681365966797, 'timing/data': 0.01067474763840437, 'timing/model': 1.3889378029853106, 'learning_rate/qwen_vl_interface': 7.465974207750442e-06, 'learning_rate/action_model': 7.465974207750442e-05, 'epoch': 0.91}) +[2026-06-17 21:15:55,981][starVLA.training.train_starvla][INFO] - Step 2485, Loss: {'train/loss': 0.8323009610176086, 'train/grad_norm_pre_clip': 17.44647216796875, 'timing/data': 0.012949489057064056, 'timing/model': 1.0333566395565867, 'learning_rate/qwen_vl_interface': 7.464061056276707e-06, 'learning_rate/action_model': 7.464061056276707e-05, 'epoch': 0.91}) +[2026-06-17 21:16:03,902][starVLA.training.train_starvla][INFO] - Step 2486, Loss: {'train/loss': 0.6244108080863953, 'train/grad_norm_pre_clip': 15.778133392333984, 'timing/data': 0.012751013040542603, 'timing/model': 1.0728999897837639, 'learning_rate/qwen_vl_interface': 7.462147445825733e-06, 'learning_rate/action_model': 7.462147445825733e-05, 'epoch': 0.91}) +[2026-06-17 21:16:08,424][starVLA.training.train_starvla][INFO] - Step 2487, Loss: {'train/loss': 0.6533518433570862, 'train/grad_norm_pre_clip': 13.377143859863281, 'timing/data': 0.011630491353571415, 'timing/model': 1.0816825553774834, 'learning_rate/qwen_vl_interface': 7.460233376794211e-06, 'learning_rate/action_model': 7.46023337679421e-05, 'epoch': 0.91}) +[2026-06-17 21:16:16,278][starVLA.training.train_starvla][INFO] - Step 2488, Loss: {'train/loss': 0.7179291248321533, 'train/grad_norm_pre_clip': 15.417508125305176, 'timing/data': 0.014502618461847305, 'timing/model': 1.5389228723943233, 'learning_rate/qwen_vl_interface': 7.458318849578934e-06, 'learning_rate/action_model': 7.458318849578933e-05, 'epoch': 0.91}) +[2026-06-17 21:16:22,618][starVLA.training.train_starvla][INFO] - Step 2489, Loss: {'train/loss': 0.34103506803512573, 'train/grad_norm_pre_clip': 11.494889259338379, 'timing/data': 0.016435285098850727, 'timing/model': 1.4077915018424392, 'learning_rate/qwen_vl_interface': 7.456403864576782e-06, 'learning_rate/action_model': 7.456403864576782e-05, 'epoch': 0.91}) +[2026-06-17 21:16:26,674][starVLA.training.train_starvla][INFO] - Step 2490, Loss: {'train/loss': 1.0169636011123657, 'train/grad_norm_pre_clip': 15.731926918029785, 'timing/data': 0.011145984753966331, 'timing/model': 1.0367786744609475, 'learning_rate/qwen_vl_interface': 7.4544884221847354e-06, 'learning_rate/action_model': 7.454488422184735e-05, 'epoch': 0.91}) +[2026-06-17 21:16:31,732][starVLA.training.train_starvla][INFO] - Step 2491, Loss: {'train/loss': 0.7993045449256897, 'train/grad_norm_pre_clip': 18.8089656829834, 'timing/data': 0.013029227033257484, 'timing/model': 1.1327013624832034, 'learning_rate/qwen_vl_interface': 7.452572522799867e-06, 'learning_rate/action_model': 7.452572522799866e-05, 'epoch': 0.91}) +[2026-06-17 21:16:36,330][starVLA.training.train_starvla][INFO] - Step 2492, Loss: {'train/loss': 0.6137754917144775, 'train/grad_norm_pre_clip': 14.56477165222168, 'timing/data': 0.010406498797237873, 'timing/model': 1.0628811968490481, 'learning_rate/qwen_vl_interface': 7.450656166819344e-06, 'learning_rate/action_model': 7.450656166819344e-05, 'epoch': 0.91}) +[2026-06-17 21:16:40,851][starVLA.training.train_starvla][INFO] - Step 2493, Loss: {'train/loss': 0.45818278193473816, 'train/grad_norm_pre_clip': 12.54742431640625, 'timing/data': 0.014226244762539864, 'timing/model': 1.0235101729631424, 'learning_rate/qwen_vl_interface': 7.4487393546404295e-06, 'learning_rate/action_model': 7.44873935464043e-05, 'epoch': 0.91}) +[2026-06-17 21:16:48,652][starVLA.training.train_starvla][INFO] - Step 2494, Loss: {'train/loss': 0.6358216404914856, 'train/grad_norm_pre_clip': 14.617679595947266, 'timing/data': 0.012367918156087399, 'timing/model': 1.396085012704134, 'learning_rate/qwen_vl_interface': 7.446822086660481e-06, 'learning_rate/action_model': 7.446822086660481e-05, 'epoch': 0.91}) +[2026-06-17 21:16:52,671][starVLA.training.train_starvla][INFO] - Step 2495, Loss: {'train/loss': 0.6197354197502136, 'train/grad_norm_pre_clip': 16.43606948852539, 'timing/data': 0.011311848647892475, 'timing/model': 1.0209250524640083, 'learning_rate/qwen_vl_interface': 7.444904363276951e-06, 'learning_rate/action_model': 7.444904363276951e-05, 'epoch': 0.91}) +[2026-06-17 21:16:57,478][starVLA.training.train_starvla][INFO] - Step 2496, Loss: {'train/loss': 0.6649510264396667, 'train/grad_norm_pre_clip': 19.32822036743164, 'timing/data': 0.018751083873212337, 'timing/model': 1.145331813953817, 'learning_rate/qwen_vl_interface': 7.442986184887383e-06, 'learning_rate/action_model': 7.442986184887383e-05, 'epoch': 0.91}) +[2026-06-17 21:17:05,210][starVLA.training.train_starvla][INFO] - Step 2497, Loss: {'train/loss': 0.7268563508987427, 'train/grad_norm_pre_clip': 15.587738037109375, 'timing/data': 0.014384379610419273, 'timing/model': 4.181569434702396, 'learning_rate/qwen_vl_interface': 7.4410675518894205e-06, 'learning_rate/action_model': 7.44106755188942e-05, 'epoch': 0.91}) +[2026-06-17 21:17:09,984][starVLA.training.train_starvla][INFO] - Step 2498, Loss: {'train/loss': 0.8362263441085815, 'train/grad_norm_pre_clip': 19.000019073486328, 'timing/data': 0.012673457153141499, 'timing/model': 1.5001148469746113, 'learning_rate/qwen_vl_interface': 7.439148464680795e-06, 'learning_rate/action_model': 7.439148464680796e-05, 'epoch': 0.91}) +[2026-06-17 21:17:14,572][starVLA.training.train_starvla][INFO] - Step 2499, Loss: {'train/loss': 0.6261772513389587, 'train/grad_norm_pre_clip': 12.923213958740234, 'timing/data': 0.011687223799526691, 'timing/model': 1.394545778632164, 'learning_rate/qwen_vl_interface': 7.4372289236593395e-06, 'learning_rate/action_model': 7.43722892365934e-05, 'epoch': 0.92}) +[2026-06-17 21:20:43,632][starVLA.training.train_starvla][INFO] - Step 2500, Loss: {'train/loss': 0.7299816012382507, 'train/grad_norm_pre_clip': 14.903854370117188, 'eval/loss': 0.7230304479598999, 'eval/latency_2/loss': 0.7230304479598999, 'eval/demon_fire_precision': 0.8795690650250096, 'eval/demon_fire_recall': 0.9564853556485355, 'eval/demon_fire_f1': 0.9164161154539988, 'eval/demon_fire_rate_model': 0.723149693934335, 'eval/demon_fire_rate_teacher': 0.6649972175848636, 'eval/demon_left_precision': 0.8824383164005806, 'eval/demon_left_recall': 0.64, 'eval/demon_left_f1': 0.7419158023184869, 'eval/demon_left_rate_model': 0.19170840289371174, 'eval/demon_left_rate_teacher': 0.2643294379521425, 'eval/demon_right_precision': 0.8186874304783093, 'eval/demon_right_recall': 0.831638418079096, 'eval/demon_right_f1': 0.8251121076233183, 'eval/demon_right_rate_model': 0.25013912075681694, 'eval/demon_right_rate_teacher': 0.24624373956594323, 'eval/demon_move_f1': 0.7835139549709026, 'eval/demon_cc_f1': 0.8720406141932974, 'eval/latency_2/demon_fire_precision': 0.8795690650250096, 'eval/latency_2/demon_fire_recall': 0.9564853556485355, 'eval/latency_2/demon_fire_f1': 0.9164161154539988, 'eval/latency_2/demon_fire_rate_model': 0.723149693934335, 'eval/latency_2/demon_fire_rate_teacher': 0.6649972175848636, 'eval/latency_2/demon_left_precision': 0.8824383164005806, 'eval/latency_2/demon_left_recall': 0.64, 'eval/latency_2/demon_left_f1': 0.7419158023184869, 'eval/latency_2/demon_left_rate_model': 0.19170840289371174, 'eval/latency_2/demon_left_rate_teacher': 0.2643294379521425, 'eval/latency_2/demon_right_precision': 0.8186874304783093, 'eval/latency_2/demon_right_recall': 0.831638418079096, 'eval/latency_2/demon_right_f1': 0.8251121076233183, 'eval/latency_2/demon_right_rate_model': 0.25013912075681694, 'eval/latency_2/demon_right_rate_teacher': 0.24624373956594323, 'eval/latency_2/demon_move_f1': 0.7835139549709026, 'eval/latency_2/demon_cc_f1': 0.8720406141932974, 'timing/data': 0.012591663748025894, 'timing/model': 1.2246402436867356, 'learning_rate/qwen_vl_interface': 7.435308929222973e-06, 'learning_rate/action_model': 7.435308929222972e-05, 'epoch': 0.92}) +[2026-06-17 21:20:43,645][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_2500_state +[2026-06-17 21:20:43,653][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 21:20:52,747][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_2500_state/model.safetensors +[2026-06-17 21:21:07,590][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_2500_state/optimizer.bin +[2026-06-17 21:21:07,591][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_2500_state/sampler.bin +[2026-06-17 21:21:07,592][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_2500_state/sampler_1.bin +[2026-06-17 21:21:07,594][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_2500_state/random_states_0.pkl +[2026-06-17 21:21:07,598][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2500 +[2026-06-17 21:21:07,599][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 21:21:07,615][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 21:21:14,953][starVLA.training.train_starvla][INFO] - Step 2501, Loss: {'train/loss': 0.7299574613571167, 'train/grad_norm_pre_clip': 16.544456481933594, 'timing/data': 0.05710188578814268, 'timing/model': 3.951633899472654, 'learning_rate/qwen_vl_interface': 7.433388481769716e-06, 'learning_rate/action_model': 7.433388481769715e-05, 'epoch': 0.92}) +[2026-06-17 21:21:29,848][starVLA.training.train_starvla][INFO] - Step 2502, Loss: {'train/loss': 0.553524374961853, 'train/grad_norm_pre_clip': 11.4782133102417, 'timing/data': 0.001163933426141739, 'timing/model': 1.0657196883112192, 'learning_rate/qwen_vl_interface': 7.431467581697675e-06, 'learning_rate/action_model': 7.431467581697674e-05, 'epoch': 0.92}) +[2026-06-17 21:21:34,893][starVLA.training.train_starvla][INFO] - Step 2503, Loss: {'train/loss': 0.43070459365844727, 'train/grad_norm_pre_clip': 15.099431991577148, 'timing/data': 0.0012670811265707016, 'timing/model': 1.0634644953534007, 'learning_rate/qwen_vl_interface': 7.429546229405057e-06, 'learning_rate/action_model': 7.429546229405057e-05, 'epoch': 0.92}) +[2026-06-17 21:21:39,296][starVLA.training.train_starvla][INFO] - Step 2504, Loss: {'train/loss': 0.6212326884269714, 'train/grad_norm_pre_clip': 12.36311149597168, 'timing/data': 0.0009008739143610001, 'timing/model': 1.0792371230199933, 'learning_rate/qwen_vl_interface': 7.427624425290162e-06, 'learning_rate/action_model': 7.427624425290161e-05, 'epoch': 0.92}) +[2026-06-17 21:21:47,089][starVLA.training.train_starvla][INFO] - Step 2505, Loss: {'train/loss': 0.5766691565513611, 'train/grad_norm_pre_clip': 12.948838233947754, 'timing/data': 0.000733768567442894, 'timing/model': 1.345801887102425, 'learning_rate/qwen_vl_interface': 7.425702169751379e-06, 'learning_rate/action_model': 7.425702169751378e-05, 'epoch': 0.92}) +[2026-06-17 21:21:51,188][starVLA.training.train_starvla][INFO] - Step 2506, Loss: {'train/loss': 0.5108039975166321, 'train/grad_norm_pre_clip': 13.696145057678223, 'timing/data': 0.0011208895593881607, 'timing/model': 1.0475602056831121, 'learning_rate/qwen_vl_interface': 7.4237794631871965e-06, 'learning_rate/action_model': 7.423779463187196e-05, 'epoch': 0.92}) +[2026-06-17 21:21:55,915][starVLA.training.train_starvla][INFO] - Step 2507, Loss: {'train/loss': 0.5187841057777405, 'train/grad_norm_pre_clip': 14.22365951538086, 'timing/data': 0.0012656375765800476, 'timing/model': 1.2227002624422312, 'learning_rate/qwen_vl_interface': 7.42185630599619e-06, 'learning_rate/action_model': 7.42185630599619e-05, 'epoch': 0.92}) +[2026-06-17 21:22:00,483][starVLA.training.train_starvla][INFO] - Step 2508, Loss: {'train/loss': 0.6870840191841125, 'train/grad_norm_pre_clip': 16.4385929107666, 'timing/data': 0.001048397272825241, 'timing/model': 1.0617204569280148, 'learning_rate/qwen_vl_interface': 7.419932698577036e-06, 'learning_rate/action_model': 7.419932698577035e-05, 'epoch': 0.92}) +[2026-06-17 21:22:05,138][starVLA.training.train_starvla][INFO] - Step 2509, Loss: {'train/loss': 0.8469393849372864, 'train/grad_norm_pre_clip': 19.753692626953125, 'timing/data': 0.0011345148086547852, 'timing/model': 1.0663164891302586, 'learning_rate/qwen_vl_interface': 7.418008641328498e-06, 'learning_rate/action_model': 7.418008641328497e-05, 'epoch': 0.92}) +[2026-06-17 21:22:11,298][starVLA.training.train_starvla][INFO] - Step 2510, Loss: {'train/loss': 0.5794467329978943, 'train/grad_norm_pre_clip': 15.77381706237793, 'timing/data': 0.0009807664901018143, 'timing/model': 1.3869604161009192, 'learning_rate/qwen_vl_interface': 7.416084134649435e-06, 'learning_rate/action_model': 7.416084134649435e-05, 'epoch': 0.92}) +[2026-06-17 21:22:15,464][starVLA.training.train_starvla][INFO] - Step 2511, Loss: {'train/loss': 0.6804776787757874, 'train/grad_norm_pre_clip': 15.805682182312012, 'timing/data': 0.0010842038318514824, 'timing/model': 1.023699838668108, 'learning_rate/qwen_vl_interface': 7.414159178938801e-06, 'learning_rate/action_model': 7.414159178938801e-05, 'epoch': 0.92}) +[2026-06-17 21:22:20,207][starVLA.training.train_starvla][INFO] - Step 2512, Loss: {'train/loss': 0.6952360272407532, 'train/grad_norm_pre_clip': 15.262557983398438, 'timing/data': 0.0014649871736764908, 'timing/model': 1.1605334570631385, 'learning_rate/qwen_vl_interface': 7.412233774595639e-06, 'learning_rate/action_model': 7.41223377459564e-05, 'epoch': 0.92}) +[2026-06-17 21:22:31,343][starVLA.training.train_starvla][INFO] - Step 2513, Loss: {'train/loss': 0.7140275239944458, 'train/grad_norm_pre_clip': 14.896540641784668, 'timing/data': 0.013262223452329636, 'timing/model': 4.2970348400995135, 'learning_rate/qwen_vl_interface': 7.410307922019091e-06, 'learning_rate/action_model': 7.41030792201909e-05, 'epoch': 0.92}) +[2026-06-17 21:22:39,969][starVLA.training.train_starvla][INFO] - Step 2514, Loss: {'train/loss': 0.4312993288040161, 'train/grad_norm_pre_clip': 11.012255668640137, 'timing/data': 0.02250922005623579, 'timing/model': 1.0270833307877183, 'learning_rate/qwen_vl_interface': 7.408381621608387e-06, 'learning_rate/action_model': 7.408381621608387e-05, 'epoch': 0.92}) +[2026-06-17 21:22:44,675][starVLA.training.train_starvla][INFO] - Step 2515, Loss: {'train/loss': 0.593302845954895, 'train/grad_norm_pre_clip': 13.63753890991211, 'timing/data': 0.027102489955723286, 'timing/model': 1.3913486376404762, 'learning_rate/qwen_vl_interface': 7.406454873762848e-06, 'learning_rate/action_model': 7.406454873762848e-05, 'epoch': 0.92}) +[2026-06-17 21:22:54,545][starVLA.training.train_starvla][INFO] - Step 2516, Loss: {'train/loss': 0.83484947681427, 'train/grad_norm_pre_clip': 16.05217170715332, 'timing/data': 0.07617244496941566, 'timing/model': 1.0996087305247784, 'learning_rate/qwen_vl_interface': 7.404527678881894e-06, 'learning_rate/action_model': 7.404527678881894e-05, 'epoch': 0.92}) +[2026-06-17 21:23:06,889][starVLA.training.train_starvla][INFO] - Step 2517, Loss: {'train/loss': 0.815494954586029, 'train/grad_norm_pre_clip': 18.080707550048828, 'timing/data': 0.001067381352186203, 'timing/model': 4.594388327561319, 'learning_rate/qwen_vl_interface': 7.402600037365034e-06, 'learning_rate/action_model': 7.402600037365034e-05, 'epoch': 0.92}) +[2026-06-17 21:23:11,899][starVLA.training.train_starvla][INFO] - Step 2518, Loss: {'train/loss': 0.9344126582145691, 'train/grad_norm_pre_clip': 21.873981475830078, 'timing/data': 0.0017009563744068146, 'timing/model': 1.285428248345852, 'learning_rate/qwen_vl_interface': 7.400671949611869e-06, 'learning_rate/action_model': 7.40067194961187e-05, 'epoch': 0.92}) +[2026-06-17 21:23:16,298][starVLA.training.train_starvla][INFO] - Step 2519, Loss: {'train/loss': 0.6458116173744202, 'train/grad_norm_pre_clip': 14.397234916687012, 'timing/data': 0.000992502085864544, 'timing/model': 1.0292701544240117, 'learning_rate/qwen_vl_interface': 7.398743416022095e-06, 'learning_rate/action_model': 7.398743416022095e-05, 'epoch': 0.92}) +[2026-06-17 21:23:24,442][starVLA.training.train_starvla][INFO] - Step 2520, Loss: {'train/loss': 1.062426209449768, 'train/grad_norm_pre_clip': 20.18305206298828, 'timing/data': 0.001621129922568798, 'timing/model': 1.4984298944473267, 'learning_rate/qwen_vl_interface': 7.3968144369954995e-06, 'learning_rate/action_model': 7.396814436995499e-05, 'epoch': 0.92}) +[2026-06-17 21:23:30,217][starVLA.training.train_starvla][INFO] - Step 2521, Loss: {'train/loss': 0.6165140271186829, 'train/grad_norm_pre_clip': 13.621102333068848, 'timing/data': 0.0012995153665542603, 'timing/model': 2.7724757473915815, 'learning_rate/qwen_vl_interface': 7.39488501293196e-06, 'learning_rate/action_model': 7.39488501293196e-05, 'epoch': 0.92}) +[2026-06-17 21:23:35,623][starVLA.training.train_starvla][INFO] - Step 2522, Loss: {'train/loss': 0.8639854192733765, 'train/grad_norm_pre_clip': 15.85265064239502, 'timing/data': 0.000973530113697052, 'timing/model': 2.042771420441568, 'learning_rate/qwen_vl_interface': 7.392955144231452e-06, 'learning_rate/action_model': 7.392955144231452e-05, 'epoch': 0.92}) +[2026-06-17 21:23:40,078][starVLA.training.train_starvla][INFO] - Step 2523, Loss: {'train/loss': 0.8020588755607605, 'train/grad_norm_pre_clip': 42.74749755859375, 'timing/data': 0.0007966319099068642, 'timing/model': 1.045902851037681, 'learning_rate/qwen_vl_interface': 7.391024831294034e-06, 'learning_rate/action_model': 7.391024831294033e-05, 'epoch': 0.92}) +[2026-06-17 21:23:44,500][starVLA.training.train_starvla][INFO] - Step 2524, Loss: {'train/loss': 0.7364506721496582, 'train/grad_norm_pre_clip': 14.228196144104004, 'timing/data': 0.0010635172948241234, 'timing/model': 1.0296488581225276, 'learning_rate/qwen_vl_interface': 7.389094074519866e-06, 'learning_rate/action_model': 7.389094074519866e-05, 'epoch': 0.92}) +[2026-06-17 21:23:48,954][starVLA.training.train_starvla][INFO] - Step 2525, Loss: {'train/loss': 0.4728303551673889, 'train/grad_norm_pre_clip': 13.351696968078613, 'timing/data': 0.0008879601955413818, 'timing/model': 1.3950268356129527, 'learning_rate/qwen_vl_interface': 7.387162874309194e-06, 'learning_rate/action_model': 7.387162874309193e-05, 'epoch': 0.92}) +[2026-06-17 21:23:53,363][starVLA.training.train_starvla][INFO] - Step 2526, Loss: {'train/loss': 0.8328019380569458, 'train/grad_norm_pre_clip': 19.924650192260742, 'timing/data': 0.0009883148595690727, 'timing/model': 1.0264607109129429, 'learning_rate/qwen_vl_interface': 7.385231231062359e-06, 'learning_rate/action_model': 7.385231231062358e-05, 'epoch': 0.92}) +[2026-06-17 21:23:57,889][starVLA.training.train_starvla][INFO] - Step 2527, Loss: {'train/loss': 0.7929521203041077, 'train/grad_norm_pre_clip': 11.555231094360352, 'timing/data': 0.0012048305943608284, 'timing/model': 1.074697194620967, 'learning_rate/qwen_vl_interface': 7.383299145179789e-06, 'learning_rate/action_model': 7.383299145179789e-05, 'epoch': 0.93}) +[2026-06-17 21:24:07,001][starVLA.training.train_starvla][INFO] - Step 2528, Loss: {'train/loss': 0.7942320108413696, 'train/grad_norm_pre_clip': 18.405080795288086, 'timing/data': 0.0011782590299844742, 'timing/model': 1.126286894083023, 'learning_rate/qwen_vl_interface': 7.381366617062014e-06, 'learning_rate/action_model': 7.381366617062014e-05, 'epoch': 0.93}) +[2026-06-17 21:24:11,612][starVLA.training.train_starvla][INFO] - Step 2529, Loss: {'train/loss': 0.7285792231559753, 'train/grad_norm_pre_clip': 14.02968692779541, 'timing/data': 0.0008926969021558762, 'timing/model': 1.1051550963893533, 'learning_rate/qwen_vl_interface': 7.379433647109645e-06, 'learning_rate/action_model': 7.379433647109645e-05, 'epoch': 0.93}) +[2026-06-17 21:24:16,145][starVLA.training.train_starvla][INFO] - Step 2530, Loss: {'train/loss': 0.7581056952476501, 'train/grad_norm_pre_clip': 15.810344696044922, 'timing/data': 0.0011092955246567726, 'timing/model': 1.4147938843816519, 'learning_rate/qwen_vl_interface': 7.37750023572339e-06, 'learning_rate/action_model': 7.37750023572339e-05, 'epoch': 0.93}) +[2026-06-17 21:24:25,102][starVLA.training.train_starvla][INFO] - Step 2531, Loss: {'train/loss': 0.605282187461853, 'train/grad_norm_pre_clip': 17.43116569519043, 'timing/data': 0.0017858343198895454, 'timing/model': 1.304999217391014, 'learning_rate/qwen_vl_interface': 7.375566383304047e-06, 'learning_rate/action_model': 7.375566383304047e-05, 'epoch': 0.93}) +[2026-06-17 21:24:33,861][starVLA.training.train_starvla][INFO] - Step 2532, Loss: {'train/loss': 0.35206517577171326, 'train/grad_norm_pre_clip': 11.529106140136719, 'timing/data': 0.0011342987418174744, 'timing/model': 1.0627274410799146, 'learning_rate/qwen_vl_interface': 7.373632090252505e-06, 'learning_rate/action_model': 7.373632090252505e-05, 'epoch': 0.93}) +[2026-06-17 21:24:38,566][starVLA.training.train_starvla][INFO] - Step 2533, Loss: {'train/loss': 0.5101174712181091, 'train/grad_norm_pre_clip': 11.326042175292969, 'timing/data': 0.0012570768594741821, 'timing/model': 1.1125948345288634, 'learning_rate/qwen_vl_interface': 7.371697356969746e-06, 'learning_rate/action_model': 7.371697356969746e-05, 'epoch': 0.93}) +[2026-06-17 21:24:42,919][starVLA.training.train_starvla][INFO] - Step 2534, Loss: {'train/loss': 0.8809367418289185, 'train/grad_norm_pre_clip': 14.29650592803955, 'timing/data': 0.0009614462032914162, 'timing/model': 1.0392110850661993, 'learning_rate/qwen_vl_interface': 7.36976218385684e-06, 'learning_rate/action_model': 7.36976218385684e-05, 'epoch': 0.93}) +[2026-06-17 21:24:47,520][starVLA.training.train_starvla][INFO] - Step 2535, Loss: {'train/loss': 0.5386739373207092, 'train/grad_norm_pre_clip': 13.407075881958008, 'timing/data': 0.0012932363897562027, 'timing/model': 1.4278557682409883, 'learning_rate/qwen_vl_interface': 7.367826571314955e-06, 'learning_rate/action_model': 7.367826571314954e-05, 'epoch': 0.93}) +[2026-06-17 21:24:58,095][starVLA.training.train_starvla][INFO] - Step 2536, Loss: {'train/loss': 0.41166701912879944, 'train/grad_norm_pre_clip': 10.425204277038574, 'timing/data': 0.0012435801327228546, 'timing/model': 1.075721644796431, 'learning_rate/qwen_vl_interface': 7.36589051974534e-06, 'learning_rate/action_model': 7.36589051974534e-05, 'epoch': 0.93}) +[2026-06-17 21:25:02,861][starVLA.training.train_starvla][INFO] - Step 2537, Loss: {'train/loss': 0.806872546672821, 'train/grad_norm_pre_clip': 15.785730361938477, 'timing/data': 0.0009021293371915817, 'timing/model': 1.0603337604552507, 'learning_rate/qwen_vl_interface': 7.363954029549348e-06, 'learning_rate/action_model': 7.363954029549348e-05, 'epoch': 0.93}) +[2026-06-17 21:25:07,389][starVLA.training.train_starvla][INFO] - Step 2538, Loss: {'train/loss': 0.7711675763130188, 'train/grad_norm_pre_clip': 11.324346542358398, 'timing/data': 0.0011967336758971214, 'timing/model': 1.0750467767938972, 'learning_rate/qwen_vl_interface': 7.362017101128406e-06, 'learning_rate/action_model': 7.362017101128406e-05, 'epoch': 0.93}) +[2026-06-17 21:25:11,972][starVLA.training.train_starvla][INFO] - Step 2539, Loss: {'train/loss': 0.5410125255584717, 'train/grad_norm_pre_clip': 13.60043716430664, 'timing/data': 0.0012234095484018326, 'timing/model': 1.1732074776664376, 'learning_rate/qwen_vl_interface': 7.360079734884049e-06, 'learning_rate/action_model': 7.360079734884049e-05, 'epoch': 0.93}) +[2026-06-17 21:25:16,394][starVLA.training.train_starvla][INFO] - Step 2540, Loss: {'train/loss': 0.38377252221107483, 'train/grad_norm_pre_clip': 11.277286529541016, 'timing/data': 0.0007734848186373711, 'timing/model': 1.3449254864826798, 'learning_rate/qwen_vl_interface': 7.358141931217893e-06, 'learning_rate/action_model': 7.358141931217892e-05, 'epoch': 0.93}) +[2026-06-17 21:25:21,700][starVLA.training.train_starvla][INFO] - Step 2541, Loss: {'train/loss': 0.37661343812942505, 'train/grad_norm_pre_clip': 19.23186492919922, 'timing/data': 0.0008939700201153755, 'timing/model': 1.215901062823832, 'learning_rate/qwen_vl_interface': 7.356203690531643e-06, 'learning_rate/action_model': 7.356203690531643e-05, 'epoch': 0.93}) +[2026-06-17 21:25:29,699][starVLA.training.train_starvla][INFO] - Step 2542, Loss: {'train/loss': 0.47680437564849854, 'train/grad_norm_pre_clip': 11.894903182983398, 'timing/data': 0.002173592336475849, 'timing/model': 1.2320343051105738, 'learning_rate/qwen_vl_interface': 7.354265013227104e-06, 'learning_rate/action_model': 7.354265013227104e-05, 'epoch': 0.93}) +[2026-06-17 21:25:37,804][starVLA.training.train_starvla][INFO] - Step 2543, Loss: {'train/loss': 0.577957272529602, 'train/grad_norm_pre_clip': 737.7547607421875, 'timing/data': 0.0009116986766457558, 'timing/model': 1.0699017867445946, 'learning_rate/qwen_vl_interface': 7.352325899706163e-06, 'learning_rate/action_model': 7.352325899706162e-05, 'epoch': 0.93}) +[2026-06-17 21:25:45,488][starVLA.training.train_starvla][INFO] - Step 2544, Loss: {'train/loss': 0.8735854029655457, 'train/grad_norm_pre_clip': 15.718295097351074, 'timing/data': 0.0010769544169306755, 'timing/model': 1.1183213694021106, 'learning_rate/qwen_vl_interface': 7.350386350370802e-06, 'learning_rate/action_model': 7.350386350370801e-05, 'epoch': 0.93}) +[2026-06-17 21:25:50,643][starVLA.training.train_starvla][INFO] - Step 2545, Loss: {'train/loss': 0.46854129433631897, 'train/grad_norm_pre_clip': 12.483267784118652, 'timing/data': 0.00135883130133152, 'timing/model': 1.6757493140175939, 'learning_rate/qwen_vl_interface': 7.348446365623089e-06, 'learning_rate/action_model': 7.348446365623088e-05, 'epoch': 0.93}) +[2026-06-17 21:25:55,273][starVLA.training.train_starvla][INFO] - Step 2546, Loss: {'train/loss': 0.5197116136550903, 'train/grad_norm_pre_clip': 11.333666801452637, 'timing/data': 0.0012235427275300026, 'timing/model': 1.218682263046503, 'learning_rate/qwen_vl_interface': 7.346505945865186e-06, 'learning_rate/action_model': 7.346505945865186e-05, 'epoch': 0.93}) +[2026-06-17 21:26:10,549][starVLA.training.train_starvla][INFO] - Step 2547, Loss: {'train/loss': 0.5499695539474487, 'train/grad_norm_pre_clip': 17.612764358520508, 'timing/data': 0.001719210296869278, 'timing/model': 5.625177084468305, 'learning_rate/qwen_vl_interface': 7.344565091499345e-06, 'learning_rate/action_model': 7.344565091499344e-05, 'epoch': 0.93}) +[2026-06-17 21:26:22,477][starVLA.training.train_starvla][INFO] - Step 2548, Loss: {'train/loss': 0.6870511770248413, 'train/grad_norm_pre_clip': 17.55155372619629, 'timing/data': 0.0025119101628661156, 'timing/model': 1.1599798034876585, 'learning_rate/qwen_vl_interface': 7.342623802927907e-06, 'learning_rate/action_model': 7.342623802927907e-05, 'epoch': 0.93}) +[2026-06-17 21:26:31,751][starVLA.training.train_starvla][INFO] - Step 2549, Loss: {'train/loss': 0.7571754455566406, 'train/grad_norm_pre_clip': 18.224063873291016, 'timing/data': 0.0010099681094288826, 'timing/model': 1.641482312232256, 'learning_rate/qwen_vl_interface': 7.340682080553303e-06, 'learning_rate/action_model': 7.340682080553303e-05, 'epoch': 0.93}) +[2026-06-17 21:26:41,331][starVLA.training.train_starvla][INFO] - Step 2550, Loss: {'train/loss': 0.7585901021957397, 'train/grad_norm_pre_clip': 15.518835067749023, 'timing/data': 0.001150348223745823, 'timing/model': 1.5929547203704715, 'learning_rate/qwen_vl_interface': 7.338739924778053e-06, 'learning_rate/action_model': 7.338739924778053e-05, 'epoch': 0.93}) +[2026-06-17 21:26:45,656][starVLA.training.train_starvla][INFO] - Step 2551, Loss: {'train/loss': 0.6017962694168091, 'train/grad_norm_pre_clip': 14.429110527038574, 'timing/data': 0.0015341592952609062, 'timing/model': 1.0408737864345312, 'learning_rate/qwen_vl_interface': 7.336797336004771e-06, 'learning_rate/action_model': 7.33679733600477e-05, 'epoch': 0.93}) +[2026-06-17 21:26:50,336][starVLA.training.train_starvla][INFO] - Step 2552, Loss: {'train/loss': 0.9710792899131775, 'train/grad_norm_pre_clip': 16.13109588623047, 'timing/data': 0.0008975565433502197, 'timing/model': 1.0470074769109488, 'learning_rate/qwen_vl_interface': 7.334854314636153e-06, 'learning_rate/action_model': 7.334854314636152e-05, 'epoch': 0.93}) +[2026-06-17 21:26:54,801][starVLA.training.train_starvla][INFO] - Step 2553, Loss: {'train/loss': 0.8154045343399048, 'train/grad_norm_pre_clip': 16.942785263061523, 'timing/data': 0.0010147886350750923, 'timing/model': 1.1954906331375241, 'learning_rate/qwen_vl_interface': 7.3329108610749955e-06, 'learning_rate/action_model': 7.332910861074995e-05, 'epoch': 0.93}) +[2026-06-17 21:26:59,199][starVLA.training.train_starvla][INFO] - Step 2554, Loss: {'train/loss': 0.6146830320358276, 'train/grad_norm_pre_clip': 13.548521995544434, 'timing/data': 0.0008754711598157883, 'timing/model': 1.1370921898633242, 'learning_rate/qwen_vl_interface': 7.330966975724173e-06, 'learning_rate/action_model': 7.330966975724173e-05, 'epoch': 0.94}) +[2026-06-17 21:27:03,601][starVLA.training.train_starvla][INFO] - Step 2555, Loss: {'train/loss': 0.6232253909111023, 'train/grad_norm_pre_clip': 16.9276123046875, 'timing/data': 0.0010819919407367706, 'timing/model': 1.3579774852842093, 'learning_rate/qwen_vl_interface': 7.329022658986657e-06, 'learning_rate/action_model': 7.329022658986658e-05, 'epoch': 0.94}) +[2026-06-17 21:27:07,944][starVLA.training.train_starvla][INFO] - Step 2556, Loss: {'train/loss': 0.8351110816001892, 'train/grad_norm_pre_clip': 16.307960510253906, 'timing/data': 0.0010966509580612183, 'timing/model': 1.3478040993213654, 'learning_rate/qwen_vl_interface': 7.327077911265508e-06, 'learning_rate/action_model': 7.327077911265507e-05, 'epoch': 0.94}) +[2026-06-17 21:27:12,479][starVLA.training.train_starvla][INFO] - Step 2557, Loss: {'train/loss': 0.6761252284049988, 'train/grad_norm_pre_clip': 15.217133522033691, 'timing/data': 0.0007935930043458939, 'timing/model': 1.1919983439147472, 'learning_rate/qwen_vl_interface': 7.325132732963872e-06, 'learning_rate/action_model': 7.325132732963872e-05, 'epoch': 0.94}) +[2026-06-17 21:27:17,084][starVLA.training.train_starvla][INFO] - Step 2558, Loss: {'train/loss': 0.5601232051849365, 'train/grad_norm_pre_clip': 13.48865032196045, 'timing/data': 0.0008487021550536156, 'timing/model': 1.1975897457450628, 'learning_rate/qwen_vl_interface': 7.323187124484988e-06, 'learning_rate/action_model': 7.323187124484988e-05, 'epoch': 0.94}) +[2026-06-17 21:27:21,743][starVLA.training.train_starvla][INFO] - Step 2559, Loss: {'train/loss': 0.4981265664100647, 'train/grad_norm_pre_clip': 10.885604858398438, 'timing/data': 0.0008019832894206047, 'timing/model': 1.1447950387373567, 'learning_rate/qwen_vl_interface': 7.3212410862321815e-06, 'learning_rate/action_model': 7.321241086232182e-05, 'epoch': 0.94}) +[2026-06-17 21:27:26,163][starVLA.training.train_starvla][INFO] - Step 2560, Loss: {'train/loss': 0.7408360242843628, 'train/grad_norm_pre_clip': 20.738418579101562, 'timing/data': 0.0012262817472219467, 'timing/model': 1.3942764569073915, 'learning_rate/qwen_vl_interface': 7.319294618608869e-06, 'learning_rate/action_model': 7.319294618608868e-05, 'epoch': 0.94}) +[2026-06-17 21:27:30,760][starVLA.training.train_starvla][INFO] - Step 2561, Loss: {'train/loss': 0.5985115766525269, 'train/grad_norm_pre_clip': 14.469461441040039, 'timing/data': 0.0010970626026391983, 'timing/model': 1.1842417530715466, 'learning_rate/qwen_vl_interface': 7.317347722018557e-06, 'learning_rate/action_model': 7.317347722018557e-05, 'epoch': 0.94}) +[2026-06-17 21:27:35,071][starVLA.training.train_starvla][INFO] - Step 2562, Loss: {'train/loss': 0.8333226442337036, 'train/grad_norm_pre_clip': 16.658409118652344, 'timing/data': 0.0010996852070093155, 'timing/model': 1.049483334645629, 'learning_rate/qwen_vl_interface': 7.3154003968648355e-06, 'learning_rate/action_model': 7.315400396864835e-05, 'epoch': 0.94}) +[2026-06-17 21:27:39,661][starVLA.training.train_starvla][INFO] - Step 2563, Loss: {'train/loss': 0.7612578868865967, 'train/grad_norm_pre_clip': 14.855765342712402, 'timing/data': 0.0012952983379364014, 'timing/model': 1.0269952630624175, 'learning_rate/qwen_vl_interface': 7.313452643551392e-06, 'learning_rate/action_model': 7.313452643551392e-05, 'epoch': 0.94}) +[2026-06-17 21:27:44,263][starVLA.training.train_starvla][INFO] - Step 2564, Loss: {'train/loss': 0.8803308010101318, 'train/grad_norm_pre_clip': 17.95248794555664, 'timing/data': 0.0008879182860255241, 'timing/model': 1.0277299648150802, 'learning_rate/qwen_vl_interface': 7.311504462481993e-06, 'learning_rate/action_model': 7.311504462481993e-05, 'epoch': 0.94}) +[2026-06-17 21:27:51,692][starVLA.training.train_starvla][INFO] - Step 2565, Loss: {'train/loss': 0.4622704088687897, 'train/grad_norm_pre_clip': 11.042219161987305, 'timing/data': 0.00427749939262867, 'timing/model': 4.382791821844876, 'learning_rate/qwen_vl_interface': 7.3095558540605e-06, 'learning_rate/action_model': 7.309555854060499e-05, 'epoch': 0.94}) +[2026-06-17 21:28:04,537][starVLA.training.train_starvla][INFO] - Step 2566, Loss: {'train/loss': 0.4657556414604187, 'train/grad_norm_pre_clip': 11.335357666015625, 'timing/data': 0.002229996956884861, 'timing/model': 1.2832543952390552, 'learning_rate/qwen_vl_interface': 7.307606818690865e-06, 'learning_rate/action_model': 7.307606818690864e-05, 'epoch': 0.94}) +[2026-06-17 21:28:13,839][starVLA.training.train_starvla][INFO] - Step 2567, Loss: {'train/loss': 0.8398866653442383, 'train/grad_norm_pre_clip': 14.524589538574219, 'timing/data': 0.0008024517446756363, 'timing/model': 4.01678304001689, 'learning_rate/qwen_vl_interface': 7.30565735677712e-06, 'learning_rate/action_model': 7.305657356777119e-05, 'epoch': 0.94}) +[2026-06-17 21:28:24,489][starVLA.training.train_starvla][INFO] - Step 2568, Loss: {'train/loss': 0.47603943943977356, 'train/grad_norm_pre_clip': 13.74581241607666, 'timing/data': 0.0008263438940048218, 'timing/model': 2.569125718437135, 'learning_rate/qwen_vl_interface': 7.303707468723395e-06, 'learning_rate/action_model': 7.303707468723395e-05, 'epoch': 0.94}) +[2026-06-17 21:28:39,715][starVLA.training.train_starvla][INFO] - Step 2569, Loss: {'train/loss': 0.6477643251419067, 'train/grad_norm_pre_clip': 18.128677368164062, 'timing/data': 0.0009215893223881721, 'timing/model': 2.8740057265385985, 'learning_rate/qwen_vl_interface': 7.301757154933899e-06, 'learning_rate/action_model': 7.3017571549339e-05, 'epoch': 0.94}) +[2026-06-17 21:28:47,949][starVLA.training.train_starvla][INFO] - Step 2570, Loss: {'train/loss': 0.4296172857284546, 'train/grad_norm_pre_clip': 11.381546974182129, 'timing/data': 0.0008770106360316277, 'timing/model': 1.5444225454702973, 'learning_rate/qwen_vl_interface': 7.29980641581294e-06, 'learning_rate/action_model': 7.299806415812939e-05, 'epoch': 0.94}) +[2026-06-17 21:28:54,235][starVLA.training.train_starvla][INFO] - Step 2571, Loss: {'train/loss': 0.3505142331123352, 'train/grad_norm_pre_clip': 11.222711563110352, 'timing/data': 0.0009739268571138382, 'timing/model': 1.1307738749310374, 'learning_rate/qwen_vl_interface': 7.297855251764903e-06, 'learning_rate/action_model': 7.297855251764903e-05, 'epoch': 0.94}) +[2026-06-17 21:29:01,019][starVLA.training.train_starvla][INFO] - Step 2572, Loss: {'train/loss': 0.8200790882110596, 'train/grad_norm_pre_clip': 18.314620971679688, 'timing/data': 0.0008749719709157944, 'timing/model': 1.2004200397059321, 'learning_rate/qwen_vl_interface': 7.295903663194269e-06, 'learning_rate/action_model': 7.29590366319427e-05, 'epoch': 0.94}) +[2026-06-17 21:29:08,284][starVLA.training.train_starvla][INFO] - Step 2573, Loss: {'train/loss': 0.5179957747459412, 'train/grad_norm_pre_clip': 15.310758590698242, 'timing/data': 0.0011008698493242264, 'timing/model': 1.1280329339206219, 'learning_rate/qwen_vl_interface': 7.293951650505606e-06, 'learning_rate/action_model': 7.293951650505607e-05, 'epoch': 0.94}) +[2026-06-17 21:29:22,162][starVLA.training.train_starvla][INFO] - Step 2574, Loss: {'train/loss': 0.923037052154541, 'train/grad_norm_pre_clip': 19.97423553466797, 'timing/data': 0.0008697118610143661, 'timing/model': 1.2057432774454355, 'learning_rate/qwen_vl_interface': 7.291999214103565e-06, 'learning_rate/action_model': 7.291999214103564e-05, 'epoch': 0.94}) +[2026-06-17 21:29:26,674][starVLA.training.train_starvla][INFO] - Step 2575, Loss: {'train/loss': 0.7314765453338623, 'train/grad_norm_pre_clip': 18.55128288269043, 'timing/data': 0.0010591046884655952, 'timing/model': 1.4407452307641506, 'learning_rate/qwen_vl_interface': 7.29004635439289e-06, 'learning_rate/action_model': 7.29004635439289e-05, 'epoch': 0.94}) +[2026-06-17 21:29:30,892][starVLA.training.train_starvla][INFO] - Step 2576, Loss: {'train/loss': 0.7028144001960754, 'train/grad_norm_pre_clip': 18.721445083618164, 'timing/data': 0.002173001877963543, 'timing/model': 1.0694292597472668, 'learning_rate/qwen_vl_interface': 7.28809307177841e-06, 'learning_rate/action_model': 7.28809307177841e-05, 'epoch': 0.94}) +[2026-06-17 21:29:35,256][starVLA.training.train_starvla][INFO] - Step 2577, Loss: {'train/loss': 0.6206581592559814, 'train/grad_norm_pre_clip': 16.334863662719727, 'timing/data': 0.0011154785752296448, 'timing/model': 1.0544551741331816, 'learning_rate/qwen_vl_interface': 7.286139366665044e-06, 'learning_rate/action_model': 7.286139366665043e-05, 'epoch': 0.94}) +[2026-06-17 21:29:39,765][starVLA.training.train_starvla][INFO] - Step 2578, Loss: {'train/loss': 0.9417601823806763, 'train/grad_norm_pre_clip': 22.030736923217773, 'timing/data': 0.0011225054040551186, 'timing/model': 1.038454624824226, 'learning_rate/qwen_vl_interface': 7.284185239457796e-06, 'learning_rate/action_model': 7.284185239457795e-05, 'epoch': 0.94}) +[2026-06-17 21:29:44,383][starVLA.training.train_starvla][INFO] - Step 2579, Loss: {'train/loss': 0.7656091451644897, 'train/grad_norm_pre_clip': 17.231311798095703, 'timing/data': 0.0007500015199184418, 'timing/model': 1.0865670815110207, 'learning_rate/qwen_vl_interface': 7.282230690561759e-06, 'learning_rate/action_model': 7.28223069056176e-05, 'epoch': 0.94}) +[2026-06-17 21:29:48,732][starVLA.training.train_starvla][INFO] - Step 2580, Loss: {'train/loss': 0.6166437268257141, 'train/grad_norm_pre_clip': 12.252837181091309, 'timing/data': 0.0013317856937646866, 'timing/model': 1.3102081594988704, 'learning_rate/qwen_vl_interface': 7.280275720382113e-06, 'learning_rate/action_model': 7.280275720382113e-05, 'epoch': 0.94}) +[2026-06-17 21:29:52,755][starVLA.training.train_starvla][INFO] - Step 2581, Loss: {'train/loss': 0.47495314478874207, 'train/grad_norm_pre_clip': 10.454691886901855, 'timing/data': 0.000862407498061657, 'timing/model': 1.018440023995936, 'learning_rate/qwen_vl_interface': 7.2783203293241246e-06, 'learning_rate/action_model': 7.278320329324125e-05, 'epoch': 0.95}) +[2026-06-17 21:29:57,561][starVLA.training.train_starvla][INFO] - Step 2582, Loss: {'train/loss': 0.5762574672698975, 'train/grad_norm_pre_clip': 10.904901504516602, 'timing/data': 0.001228049397468567, 'timing/model': 1.3096644403412938, 'learning_rate/qwen_vl_interface': 7.276364517793152e-06, 'learning_rate/action_model': 7.276364517793151e-05, 'epoch': 0.95}) +[2026-06-17 21:30:02,909][starVLA.training.train_starvla][INFO] - Step 2583, Loss: {'train/loss': 0.5141718983650208, 'train/grad_norm_pre_clip': 13.333572387695312, 'timing/data': 0.001006326638162136, 'timing/model': 1.8231695275753736, 'learning_rate/qwen_vl_interface': 7.274408286194631e-06, 'learning_rate/action_model': 7.27440828619463e-05, 'epoch': 0.95}) +[2026-06-17 21:30:07,821][starVLA.training.train_starvla][INFO] - Step 2584, Loss: {'train/loss': 0.6876319646835327, 'train/grad_norm_pre_clip': 17.925905227661133, 'timing/data': 0.0010257735848426819, 'timing/model': 1.0727286245673895, 'learning_rate/qwen_vl_interface': 7.272451634934096e-06, 'learning_rate/action_model': 7.272451634934096e-05, 'epoch': 0.95}) +[2026-06-17 21:30:12,258][starVLA.training.train_starvla][INFO] - Step 2585, Loss: {'train/loss': 0.46662336587905884, 'train/grad_norm_pre_clip': 17.425596237182617, 'timing/data': 0.0009012417867779732, 'timing/model': 1.415795162320137, 'learning_rate/qwen_vl_interface': 7.270494564417161e-06, 'learning_rate/action_model': 7.270494564417161e-05, 'epoch': 0.95}) +[2026-06-17 21:30:16,406][starVLA.training.train_starvla][INFO] - Step 2586, Loss: {'train/loss': 0.49787646532058716, 'train/grad_norm_pre_clip': 12.209672927856445, 'timing/data': 0.0008995132520794868, 'timing/model': 1.101327657699585, 'learning_rate/qwen_vl_interface': 7.268537075049527e-06, 'learning_rate/action_model': 7.268537075049526e-05, 'epoch': 0.95}) +[2026-06-17 21:30:23,827][starVLA.training.train_starvla][INFO] - Step 2587, Loss: {'train/loss': 0.9673547148704529, 'train/grad_norm_pre_clip': 24.945079803466797, 'timing/data': 0.0008751489222049713, 'timing/model': 1.2702684570103884, 'learning_rate/qwen_vl_interface': 7.2665791672369875e-06, 'learning_rate/action_model': 7.266579167236988e-05, 'epoch': 0.95}) +[2026-06-17 21:30:28,189][starVLA.training.train_starvla][INFO] - Step 2588, Loss: {'train/loss': 0.5504613518714905, 'train/grad_norm_pre_clip': 29.804845809936523, 'timing/data': 0.0015792977064847946, 'timing/model': 1.0463675381615758, 'learning_rate/qwen_vl_interface': 7.264620841385415e-06, 'learning_rate/action_model': 7.264620841385415e-05, 'epoch': 0.95}) +[2026-06-17 21:30:32,710][starVLA.training.train_starvla][INFO] - Step 2589, Loss: {'train/loss': 0.4749917984008789, 'train/grad_norm_pre_clip': 13.45346736907959, 'timing/data': 0.0008988985791802406, 'timing/model': 1.109415713697672, 'learning_rate/qwen_vl_interface': 7.262662097900775e-06, 'learning_rate/action_model': 7.262662097900774e-05, 'epoch': 0.95}) +[2026-06-17 21:30:39,908][starVLA.training.train_starvla][INFO] - Step 2590, Loss: {'train/loss': 0.8444005846977234, 'train/grad_norm_pre_clip': 19.003095626831055, 'timing/data': 0.0008859140798449516, 'timing/model': 4.052267104387283, 'learning_rate/qwen_vl_interface': 7.260702937189116e-06, 'learning_rate/action_model': 7.260702937189115e-05, 'epoch': 0.95}) +[2026-06-17 21:30:45,823][starVLA.training.train_starvla][INFO] - Step 2591, Loss: {'train/loss': 0.6934302449226379, 'train/grad_norm_pre_clip': 15.384060859680176, 'timing/data': 0.0010173441842198372, 'timing/model': 1.085694015957415, 'learning_rate/qwen_vl_interface': 7.258743359656574e-06, 'learning_rate/action_model': 7.258743359656573e-05, 'epoch': 0.95}) +[2026-06-17 21:30:50,296][starVLA.training.train_starvla][INFO] - Step 2592, Loss: {'train/loss': 0.5210111141204834, 'train/grad_norm_pre_clip': 11.952963829040527, 'timing/data': 0.0008890451863408089, 'timing/model': 1.1135998032987118, 'learning_rate/qwen_vl_interface': 7.2567833657093725e-06, 'learning_rate/action_model': 7.256783365709371e-05, 'epoch': 0.95}) +[2026-06-17 21:30:54,965][starVLA.training.train_starvla][INFO] - Step 2593, Loss: {'train/loss': 0.5922988057136536, 'train/grad_norm_pre_clip': 12.301854133605957, 'timing/data': 0.0016242852434515953, 'timing/model': 1.0318715311586857, 'learning_rate/qwen_vl_interface': 7.254822955753821e-06, 'learning_rate/action_model': 7.25482295575382e-05, 'epoch': 0.95}) +[2026-06-17 21:31:03,505][starVLA.training.train_starvla][INFO] - Step 2594, Loss: {'train/loss': 0.43371129035949707, 'train/grad_norm_pre_clip': 13.207756996154785, 'timing/data': 0.0010527530685067177, 'timing/model': 4.983413400128484, 'learning_rate/qwen_vl_interface': 7.2528621301963135e-06, 'learning_rate/action_model': 7.252862130196313e-05, 'epoch': 0.95}) +[2026-06-17 21:31:07,818][starVLA.training.train_starvla][INFO] - Step 2595, Loss: {'train/loss': 1.0651493072509766, 'train/grad_norm_pre_clip': 18.812305450439453, 'timing/data': 0.0032031750306487083, 'timing/model': 1.148086660541594, 'learning_rate/qwen_vl_interface': 7.25090088944333e-06, 'learning_rate/action_model': 7.250900889443329e-05, 'epoch': 0.95}) +[2026-06-17 21:31:13,788][starVLA.training.train_starvla][INFO] - Step 2596, Loss: {'train/loss': 0.700214684009552, 'train/grad_norm_pre_clip': 19.31800079345703, 'timing/data': 0.00121286790817976, 'timing/model': 1.100869250483811, 'learning_rate/qwen_vl_interface': 7.248939233901438e-06, 'learning_rate/action_model': 7.248939233901439e-05, 'epoch': 0.95}) +[2026-06-17 21:31:18,248][starVLA.training.train_starvla][INFO] - Step 2597, Loss: {'train/loss': 0.9091529250144958, 'train/grad_norm_pre_clip': 17.99858283996582, 'timing/data': 0.0010154256597161293, 'timing/model': 1.034778174944222, 'learning_rate/qwen_vl_interface': 7.2469771639772944e-06, 'learning_rate/action_model': 7.246977163977294e-05, 'epoch': 0.95}) +[2026-06-17 21:31:22,882][starVLA.training.train_starvla][INFO] - Step 2598, Loss: {'train/loss': 0.5472292304039001, 'train/grad_norm_pre_clip': 13.567306518554688, 'timing/data': 0.0011941678822040558, 'timing/model': 1.020286844111979, 'learning_rate/qwen_vl_interface': 7.245014680077633e-06, 'learning_rate/action_model': 7.245014680077633e-05, 'epoch': 0.95}) +[2026-06-17 21:31:28,416][starVLA.training.train_starvla][INFO] - Step 2599, Loss: {'train/loss': 0.6472560167312622, 'train/grad_norm_pre_clip': 14.929705619812012, 'timing/data': 0.0011191293597221375, 'timing/model': 2.0366702703759074, 'learning_rate/qwen_vl_interface': 7.243051782609285e-06, 'learning_rate/action_model': 7.243051782609285e-05, 'epoch': 0.95}) +[2026-06-17 21:35:18,227][starVLA.training.train_starvla][INFO] - Step 2600, Loss: {'train/loss': 0.7890176177024841, 'train/grad_norm_pre_clip': 20.215030670166016, 'eval/loss': 0.6835198998451233, 'eval/latency_2/loss': 0.6835198998451233, 'eval/demon_fire_precision': 0.9070616883116883, 'eval/demon_fire_recall': 0.9351464435146444, 'eval/demon_fire_f1': 0.9208899876390606, 'eval/demon_fire_rate_model': 0.6855870895937674, 'eval/demon_fire_rate_teacher': 0.6649972175848636, 'eval/demon_left_precision': 0.8879310344827587, 'eval/demon_left_recall': 0.7589473684210526, 'eval/demon_left_f1': 0.81838819523269, 'eval/demon_left_rate_model': 0.22593210907067335, 'eval/demon_left_rate_teacher': 0.2643294379521425, 'eval/demon_right_precision': 0.8348946135831382, 'eval/demon_right_recall': 0.8056497175141243, 'eval/demon_right_f1': 0.8200115008625648, 'eval/demon_right_rate_model': 0.2376182526432944, 'eval/demon_right_rate_teacher': 0.24624373956594323, 'eval/demon_move_f1': 0.8191998480476275, 'eval/demon_cc_f1': 0.8878272577135818, 'eval/latency_2/demon_fire_precision': 0.9070616883116883, 'eval/latency_2/demon_fire_recall': 0.9351464435146444, 'eval/latency_2/demon_fire_f1': 0.9208899876390606, 'eval/latency_2/demon_fire_rate_model': 0.6855870895937674, 'eval/latency_2/demon_fire_rate_teacher': 0.6649972175848636, 'eval/latency_2/demon_left_precision': 0.8879310344827587, 'eval/latency_2/demon_left_recall': 0.7589473684210526, 'eval/latency_2/demon_left_f1': 0.81838819523269, 'eval/latency_2/demon_left_rate_model': 0.22593210907067335, 'eval/latency_2/demon_left_rate_teacher': 0.2643294379521425, 'eval/latency_2/demon_right_precision': 0.8348946135831382, 'eval/latency_2/demon_right_recall': 0.8056497175141243, 'eval/latency_2/demon_right_f1': 0.8200115008625648, 'eval/latency_2/demon_right_rate_model': 0.2376182526432944, 'eval/latency_2/demon_right_rate_teacher': 0.24624373956594323, 'eval/latency_2/demon_move_f1': 0.8191998480476275, 'eval/latency_2/demon_cc_f1': 0.8878272577135818, 'timing/data': 0.0013921288773417473, 'timing/model': 1.5208173086866736, 'learning_rate/qwen_vl_interface': 7.241088471979155e-06, 'learning_rate/action_model': 7.241088471979155e-05, 'epoch': 0.95}) +[2026-06-17 21:35:18,248][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_2600_state +[2026-06-17 21:35:18,255][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 21:35:26,860][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_2600_state/model.safetensors +[2026-06-17 21:35:41,889][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_2600_state/optimizer.bin +[2026-06-17 21:35:41,890][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_2600_state/sampler.bin +[2026-06-17 21:35:41,891][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_2600_state/sampler_1.bin +[2026-06-17 21:35:41,894][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_2600_state/random_states_0.pkl +[2026-06-17 21:35:41,897][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2600 +[2026-06-17 21:35:41,898][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 21:35:41,915][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 21:35:46,681][starVLA.training.train_starvla][INFO] - Step 2601, Loss: {'train/loss': 0.592250406742096, 'train/grad_norm_pre_clip': 14.989239692687988, 'timing/data': 0.0018173577263951302, 'timing/model': 1.4348810110241175, 'learning_rate/qwen_vl_interface': 7.239124748594244e-06, 'learning_rate/action_model': 7.239124748594244e-05, 'epoch': 0.95}) +[2026-06-17 21:35:50,740][starVLA.training.train_starvla][INFO] - Step 2602, Loss: {'train/loss': 0.5557940602302551, 'train/grad_norm_pre_clip': 14.60982608795166, 'timing/data': 0.0009135613217949867, 'timing/model': 1.0328901084139943, 'learning_rate/qwen_vl_interface': 7.237160612861632e-06, 'learning_rate/action_model': 7.237160612861632e-05, 'epoch': 0.95}) +[2026-06-17 21:35:55,369][starVLA.training.train_starvla][INFO] - Step 2603, Loss: {'train/loss': 0.556189239025116, 'train/grad_norm_pre_clip': 11.690971374511719, 'timing/data': 0.0012879911810159683, 'timing/model': 1.039864239282906, 'learning_rate/qwen_vl_interface': 7.235196065188485e-06, 'learning_rate/action_model': 7.235196065188485e-05, 'epoch': 0.95}) +[2026-06-17 21:36:00,011][starVLA.training.train_starvla][INFO] - Step 2604, Loss: {'train/loss': 0.6300477981567383, 'train/grad_norm_pre_clip': 30.264427185058594, 'timing/data': 0.0009104255586862564, 'timing/model': 1.1059278044849634, 'learning_rate/qwen_vl_interface': 7.23323110598206e-06, 'learning_rate/action_model': 7.23323110598206e-05, 'epoch': 0.95}) +[2026-06-17 21:36:06,123][starVLA.training.train_starvla][INFO] - Step 2605, Loss: {'train/loss': 0.7314413785934448, 'train/grad_norm_pre_clip': 14.119278907775879, 'timing/data': 0.0009383577853441238, 'timing/model': 2.6889043897390366, 'learning_rate/qwen_vl_interface': 7.231265735649691e-06, 'learning_rate/action_model': 7.231265735649691e-05, 'epoch': 0.95}) +[2026-06-17 21:36:10,687][starVLA.training.train_starvla][INFO] - Step 2606, Loss: {'train/loss': 0.6483144164085388, 'train/grad_norm_pre_clip': 13.958993911743164, 'timing/data': 0.0010498091578483582, 'timing/model': 1.373642971739173, 'learning_rate/qwen_vl_interface': 7.229299954598805e-06, 'learning_rate/action_model': 7.229299954598804e-05, 'epoch': 0.95}) +[2026-06-17 21:36:15,007][starVLA.training.train_starvla][INFO] - Step 2607, Loss: {'train/loss': 0.4991413354873657, 'train/grad_norm_pre_clip': 11.461586952209473, 'timing/data': 0.00120620708912611, 'timing/model': 1.2723345579579473, 'learning_rate/qwen_vl_interface': 7.227333763236907e-06, 'learning_rate/action_model': 7.227333763236906e-05, 'epoch': 0.95}) +[2026-06-17 21:36:19,799][starVLA.training.train_starvla][INFO] - Step 2608, Loss: {'train/loss': 0.5515707731246948, 'train/grad_norm_pre_clip': 13.17420482635498, 'timing/data': 0.0009052492678165436, 'timing/model': 1.1227103900164366, 'learning_rate/qwen_vl_interface': 7.225367161971592e-06, 'learning_rate/action_model': 7.225367161971591e-05, 'epoch': 0.95}) +[2026-06-17 21:36:24,286][starVLA.training.train_starvla][INFO] - Step 2609, Loss: {'train/loss': 0.7917957305908203, 'train/grad_norm_pre_clip': 15.857118606567383, 'timing/data': 0.0009273160248994827, 'timing/model': 1.1163228275254369, 'learning_rate/qwen_vl_interface': 7.223400151210539e-06, 'learning_rate/action_model': 7.223400151210538e-05, 'epoch': 0.96}) +[2026-06-17 21:36:28,641][starVLA.training.train_starvla][INFO] - Step 2610, Loss: {'train/loss': 0.5734406113624573, 'train/grad_norm_pre_clip': 13.256993293762207, 'timing/data': 0.0007641240954399109, 'timing/model': 1.0317986970767379, 'learning_rate/qwen_vl_interface': 7.2214327313615105e-06, 'learning_rate/action_model': 7.22143273136151e-05, 'epoch': 0.96}) +[2026-06-17 21:36:33,314][starVLA.training.train_starvla][INFO] - Step 2611, Loss: {'train/loss': 0.9626860618591309, 'train/grad_norm_pre_clip': 16.78003692626953, 'timing/data': 0.000988369807600975, 'timing/model': 1.4559884602203965, 'learning_rate/qwen_vl_interface': 7.219464902832357e-06, 'learning_rate/action_model': 7.219464902832356e-05, 'epoch': 0.96}) +[2026-06-17 21:36:38,719][starVLA.training.train_starvla][INFO] - Step 2612, Loss: {'train/loss': 0.3639122247695923, 'train/grad_norm_pre_clip': 12.831216812133789, 'timing/data': 0.0015261052176356316, 'timing/model': 2.3280041767284274, 'learning_rate/qwen_vl_interface': 7.2174966660310085e-06, 'learning_rate/action_model': 7.217496666031008e-05, 'epoch': 0.96}) +[2026-06-17 21:36:46,964][starVLA.training.train_starvla][INFO] - Step 2613, Loss: {'train/loss': 0.6015304327011108, 'train/grad_norm_pre_clip': 14.079427719116211, 'timing/data': 0.002932569943368435, 'timing/model': 4.3629682986065745, 'learning_rate/qwen_vl_interface': 7.215528021365487e-06, 'learning_rate/action_model': 7.215528021365487e-05, 'epoch': 0.96}) +[2026-06-17 21:36:51,390][starVLA.training.train_starvla][INFO] - Step 2614, Loss: {'train/loss': 0.4438799023628235, 'train/grad_norm_pre_clip': 13.931764602661133, 'timing/data': 0.0012933770194649696, 'timing/model': 1.0432178042829037, 'learning_rate/qwen_vl_interface': 7.213558969243888e-06, 'learning_rate/action_model': 7.213558969243888e-05, 'epoch': 0.96}) +[2026-06-17 21:36:56,143][starVLA.training.train_starvla][INFO] - Step 2615, Loss: {'train/loss': 0.8325578570365906, 'train/grad_norm_pre_clip': 16.448801040649414, 'timing/data': 0.0008746739476919174, 'timing/model': 1.3312491094693542, 'learning_rate/qwen_vl_interface': 7.211589510074404e-06, 'learning_rate/action_model': 7.211589510074404e-05, 'epoch': 0.96}) +[2026-06-17 21:37:00,406][starVLA.training.train_starvla][INFO] - Step 2616, Loss: {'train/loss': 0.41997843980789185, 'train/grad_norm_pre_clip': 14.538777351379395, 'timing/data': 0.0009666653349995613, 'timing/model': 1.0850457660853863, 'learning_rate/qwen_vl_interface': 7.209619644265302e-06, 'learning_rate/action_model': 7.209619644265302e-05, 'epoch': 0.96}) +[2026-06-17 21:37:08,185][starVLA.training.train_starvla][INFO] - Step 2617, Loss: {'train/loss': 0.8201071619987488, 'train/grad_norm_pre_clip': 16.41543960571289, 'timing/data': 0.003340691328048706, 'timing/model': 4.48214378207922, 'learning_rate/qwen_vl_interface': 7.20764937222494e-06, 'learning_rate/action_model': 7.20764937222494e-05, 'epoch': 0.96}) +[2026-06-17 21:37:16,242][starVLA.training.train_starvla][INFO] - Step 2618, Loss: {'train/loss': 0.7092803716659546, 'train/grad_norm_pre_clip': 21.859922409057617, 'timing/data': 0.0012250347062945366, 'timing/model': 1.1572758052498102, 'learning_rate/qwen_vl_interface': 7.205678694361757e-06, 'learning_rate/action_model': 7.205678694361757e-05, 'epoch': 0.96}) +[2026-06-17 21:37:21,110][starVLA.training.train_starvla][INFO] - Step 2619, Loss: {'train/loss': 0.6244359612464905, 'train/grad_norm_pre_clip': 15.086899757385254, 'timing/data': 0.0011225873604416847, 'timing/model': 1.2098065856844187, 'learning_rate/qwen_vl_interface': 7.203707611084276e-06, 'learning_rate/action_model': 7.203707611084275e-05, 'epoch': 0.96}) +[2026-06-17 21:37:25,790][starVLA.training.train_starvla][INFO] - Step 2620, Loss: {'train/loss': 0.9450147151947021, 'train/grad_norm_pre_clip': 17.07469367980957, 'timing/data': 0.001626109704375267, 'timing/model': 1.064193757250905, 'learning_rate/qwen_vl_interface': 7.2017361228011055e-06, 'learning_rate/action_model': 7.201736122801104e-05, 'epoch': 0.96}) +[2026-06-17 21:37:30,324][starVLA.training.train_starvla][INFO] - Step 2621, Loss: {'train/loss': 0.6200984120368958, 'train/grad_norm_pre_clip': 15.746769905090332, 'timing/data': 0.0009056609123945236, 'timing/model': 1.3251133412122726, 'learning_rate/qwen_vl_interface': 7.199764229920935e-06, 'learning_rate/action_model': 7.199764229920934e-05, 'epoch': 0.96}) +[2026-06-17 21:37:38,301][starVLA.training.train_starvla][INFO] - Step 2622, Loss: {'train/loss': 0.7359908223152161, 'train/grad_norm_pre_clip': 15.368474960327148, 'timing/data': 0.0010219728574156761, 'timing/model': 1.2807276649400592, 'learning_rate/qwen_vl_interface': 7.197791932852543e-06, 'learning_rate/action_model': 7.197791932852542e-05, 'epoch': 0.96}) +[2026-06-17 21:37:42,873][starVLA.training.train_starvla][INFO] - Step 2623, Loss: {'train/loss': 0.5751202702522278, 'train/grad_norm_pre_clip': 45.91202926635742, 'timing/data': 0.0010948656126856804, 'timing/model': 1.0184138575568795, 'learning_rate/qwen_vl_interface': 7.195819232004787e-06, 'learning_rate/action_model': 7.195819232004787e-05, 'epoch': 0.96}) +[2026-06-17 21:37:48,210][starVLA.training.train_starvla][INFO] - Step 2624, Loss: {'train/loss': 0.8023760914802551, 'train/grad_norm_pre_clip': 14.281978607177734, 'timing/data': 0.0011503435671329498, 'timing/model': 1.8976200558245182, 'learning_rate/qwen_vl_interface': 7.1938461277866105e-06, 'learning_rate/action_model': 7.19384612778661e-05, 'epoch': 0.96}) +[2026-06-17 21:37:55,894][starVLA.training.train_starvla][INFO] - Step 2625, Loss: {'train/loss': 0.7873952984809875, 'train/grad_norm_pre_clip': 18.20305633544922, 'timing/data': 0.0010098740458488464, 'timing/model': 1.331530204974115, 'learning_rate/qwen_vl_interface': 7.19187262060704e-06, 'learning_rate/action_model': 7.19187262060704e-05, 'epoch': 0.96}) +[2026-06-17 21:38:01,292][starVLA.training.train_starvla][INFO] - Step 2626, Loss: {'train/loss': 0.8840802311897278, 'train/grad_norm_pre_clip': 18.484718322753906, 'timing/data': 0.0013722795993089676, 'timing/model': 1.367692556232214, 'learning_rate/qwen_vl_interface': 7.189898710875185e-06, 'learning_rate/action_model': 7.189898710875184e-05, 'epoch': 0.96}) +[2026-06-17 21:38:05,565][starVLA.training.train_starvla][INFO] - Step 2627, Loss: {'train/loss': 0.4391546845436096, 'train/grad_norm_pre_clip': 16.93408203125, 'timing/data': 0.0010367901995778084, 'timing/model': 1.2675825376063585, 'learning_rate/qwen_vl_interface': 7.187924399000241e-06, 'learning_rate/action_model': 7.18792439900024e-05, 'epoch': 0.96}) +[2026-06-17 21:38:14,541][starVLA.training.train_starvla][INFO] - Step 2628, Loss: {'train/loss': 0.6342364549636841, 'train/grad_norm_pre_clip': 15.592098236083984, 'timing/data': 0.001276366412639618, 'timing/model': 1.0542881898581982, 'learning_rate/qwen_vl_interface': 7.185949685391483e-06, 'learning_rate/action_model': 7.185949685391483e-05, 'epoch': 0.96}) +[2026-06-17 21:38:19,140][starVLA.training.train_starvla][INFO] - Step 2629, Loss: {'train/loss': 0.7632887959480286, 'train/grad_norm_pre_clip': 19.637653350830078, 'timing/data': 0.0010513300076127052, 'timing/model': 1.1179353017359972, 'learning_rate/qwen_vl_interface': 7.183974570458272e-06, 'learning_rate/action_model': 7.183974570458272e-05, 'epoch': 0.96}) +[2026-06-17 21:38:23,772][starVLA.training.train_starvla][INFO] - Step 2630, Loss: {'train/loss': 0.5661532282829285, 'train/grad_norm_pre_clip': 14.688158988952637, 'timing/data': 0.0010285330936312675, 'timing/model': 1.2020936086773872, 'learning_rate/qwen_vl_interface': 7.181999054610053e-06, 'learning_rate/action_model': 7.181999054610052e-05, 'epoch': 0.96}) +[2026-06-17 21:38:28,288][starVLA.training.train_starvla][INFO] - Step 2631, Loss: {'train/loss': 0.6191149950027466, 'train/grad_norm_pre_clip': 12.741914749145508, 'timing/data': 0.0009022178128361702, 'timing/model': 1.3618830675259233, 'learning_rate/qwen_vl_interface': 7.180023138256349e-06, 'learning_rate/action_model': 7.180023138256348e-05, 'epoch': 0.96}) +[2026-06-17 21:38:39,828][starVLA.training.train_starvla][INFO] - Step 2632, Loss: {'train/loss': 0.8546151518821716, 'train/grad_norm_pre_clip': 18.208866119384766, 'timing/data': 0.0007861824706196785, 'timing/model': 1.0883647920563817, 'learning_rate/qwen_vl_interface': 7.178046821806771e-06, 'learning_rate/action_model': 7.17804682180677e-05, 'epoch': 0.96}) +[2026-06-17 21:38:44,367][starVLA.training.train_starvla][INFO] - Step 2633, Loss: {'train/loss': 0.6400948762893677, 'train/grad_norm_pre_clip': 26.68004035949707, 'timing/data': 0.000924503430724144, 'timing/model': 1.1290949266403913, 'learning_rate/qwen_vl_interface': 7.176070105671016e-06, 'learning_rate/action_model': 7.176070105671015e-05, 'epoch': 0.96}) +[2026-06-17 21:38:49,927][starVLA.training.train_starvla][INFO] - Step 2634, Loss: {'train/loss': 0.4077983796596527, 'train/grad_norm_pre_clip': 12.685136795043945, 'timing/data': 0.0009093703702092171, 'timing/model': 2.199867338873446, 'learning_rate/qwen_vl_interface': 7.174092990258852e-06, 'learning_rate/action_model': 7.174092990258851e-05, 'epoch': 0.96}) +[2026-06-17 21:38:54,439][starVLA.training.train_starvla][INFO] - Step 2635, Loss: {'train/loss': 0.8724616169929504, 'train/grad_norm_pre_clip': 16.825021743774414, 'timing/data': 0.0008707158267498016, 'timing/model': 1.0859836898744106, 'learning_rate/qwen_vl_interface': 7.172115475980142e-06, 'learning_rate/action_model': 7.172115475980142e-05, 'epoch': 0.96}) +[2026-06-17 21:38:59,119][starVLA.training.train_starvla][INFO] - Step 2636, Loss: {'train/loss': 0.49035921692848206, 'train/grad_norm_pre_clip': 12.553354263305664, 'timing/data': 0.0015522511675953865, 'timing/model': 1.503166907466948, 'learning_rate/qwen_vl_interface': 7.170137563244825e-06, 'learning_rate/action_model': 7.170137563244824e-05, 'epoch': 0.97}) +[2026-06-17 21:39:06,134][starVLA.training.train_starvla][INFO] - Step 2637, Loss: {'train/loss': 0.2954813241958618, 'train/grad_norm_pre_clip': 8.394363403320312, 'timing/data': 0.0007974216714501381, 'timing/model': 3.17878545448184, 'learning_rate/qwen_vl_interface': 7.168159252462927e-06, 'learning_rate/action_model': 7.168159252462926e-05, 'epoch': 0.97}) +[2026-06-17 21:39:10,772][starVLA.training.train_starvla][INFO] - Step 2638, Loss: {'train/loss': 0.39613935351371765, 'train/grad_norm_pre_clip': 12.592642784118652, 'timing/data': 0.0010677175596356392, 'timing/model': 1.1028403267264366, 'learning_rate/qwen_vl_interface': 7.166180544044549e-06, 'learning_rate/action_model': 7.166180544044549e-05, 'epoch': 0.97}) +[2026-06-17 21:39:17,118][starVLA.training.train_starvla][INFO] - Step 2639, Loss: {'train/loss': 0.6723440289497375, 'train/grad_norm_pre_clip': 17.585670471191406, 'timing/data': 0.0009651007130742073, 'timing/model': 3.024011490866542, 'learning_rate/qwen_vl_interface': 7.164201438399884e-06, 'learning_rate/action_model': 7.164201438399884e-05, 'epoch': 0.97}) +[2026-06-17 21:39:24,874][starVLA.training.train_starvla][INFO] - Step 2640, Loss: {'train/loss': 1.0397257804870605, 'train/grad_norm_pre_clip': 21.851707458496094, 'timing/data': 0.001304742880165577, 'timing/model': 1.0336709218099713, 'learning_rate/qwen_vl_interface': 7.1622219359392e-06, 'learning_rate/action_model': 7.162221935939199e-05, 'epoch': 0.97}) +[2026-06-17 21:39:31,151][starVLA.training.train_starvla][INFO] - Step 2641, Loss: {'train/loss': 0.8467397689819336, 'train/grad_norm_pre_clip': 28.602510452270508, 'timing/data': 0.0010502394288778305, 'timing/model': 1.4042598297819495, 'learning_rate/qwen_vl_interface': 7.160242037072852e-06, 'learning_rate/action_model': 7.160242037072851e-05, 'epoch': 0.97}) +[2026-06-17 21:39:35,436][starVLA.training.train_starvla][INFO] - Step 2642, Loss: {'train/loss': 0.8585691452026367, 'train/grad_norm_pre_clip': 17.1589298248291, 'timing/data': 0.0012986110523343086, 'timing/model': 1.0752572575584054, 'learning_rate/qwen_vl_interface': 7.158261742211275e-06, 'learning_rate/action_model': 7.158261742211274e-05, 'epoch': 0.97}) +[2026-06-17 21:39:45,037][starVLA.training.train_starvla][INFO] - Step 2643, Loss: {'train/loss': 0.6574892401695251, 'train/grad_norm_pre_clip': 18.98116111755371, 'timing/data': 0.0012606577947735786, 'timing/model': 1.161170057952404, 'learning_rate/qwen_vl_interface': 7.156281051764983e-06, 'learning_rate/action_model': 7.156281051764983e-05, 'epoch': 0.97}) +[2026-06-17 21:39:49,344][starVLA.training.train_starvla][INFO] - Step 2644, Loss: {'train/loss': 0.4793564975261688, 'train/grad_norm_pre_clip': 12.267342567443848, 'timing/data': 0.0013817595317959785, 'timing/model': 1.0465538306161761, 'learning_rate/qwen_vl_interface': 7.15429996614458e-06, 'learning_rate/action_model': 7.15429996614458e-05, 'epoch': 0.97}) +[2026-06-17 21:39:54,243][starVLA.training.train_starvla][INFO] - Step 2645, Loss: {'train/loss': 0.5009007453918457, 'train/grad_norm_pre_clip': 14.60254955291748, 'timing/data': 0.0014294451102614403, 'timing/model': 1.140541666187346, 'learning_rate/qwen_vl_interface': 7.152318485760743e-06, 'learning_rate/action_model': 7.152318485760742e-05, 'epoch': 0.97}) +[2026-06-17 21:40:00,539][starVLA.training.train_starvla][INFO] - Step 2646, Loss: {'train/loss': 0.6419588327407837, 'train/grad_norm_pre_clip': 16.351421356201172, 'timing/data': 0.0010100025683641434, 'timing/model': 1.4448298085480928, 'learning_rate/qwen_vl_interface': 7.150336611024239e-06, 'learning_rate/action_model': 7.150336611024239e-05, 'epoch': 0.97}) +[2026-06-17 21:40:05,021][starVLA.training.train_starvla][INFO] - Step 2647, Loss: {'train/loss': 0.7523629069328308, 'train/grad_norm_pre_clip': 12.724956512451172, 'timing/data': 0.0011760089546442032, 'timing/model': 1.1402365965768695, 'learning_rate/qwen_vl_interface': 7.1483543423459086e-06, 'learning_rate/action_model': 7.148354342345908e-05, 'epoch': 0.97}) +[2026-06-17 21:40:09,588][starVLA.training.train_starvla][INFO] - Step 2648, Loss: {'train/loss': 0.6153480410575867, 'train/grad_norm_pre_clip': 14.371773719787598, 'timing/data': 0.0008647684007883072, 'timing/model': 1.1487634228542447, 'learning_rate/qwen_vl_interface': 7.146371680136682e-06, 'learning_rate/action_model': 7.146371680136681e-05, 'epoch': 0.97}) +[2026-06-17 21:40:14,241][starVLA.training.train_starvla][INFO] - Step 2649, Loss: {'train/loss': 0.5201928019523621, 'train/grad_norm_pre_clip': 14.09052848815918, 'timing/data': 0.0014139050617814064, 'timing/model': 1.120212067849934, 'learning_rate/qwen_vl_interface': 7.144388624807565e-06, 'learning_rate/action_model': 7.144388624807566e-05, 'epoch': 0.97}) +[2026-06-17 21:40:19,309][starVLA.training.train_starvla][INFO] - Step 2650, Loss: {'train/loss': 0.64638352394104, 'train/grad_norm_pre_clip': 13.720455169677734, 'timing/data': 0.0018412945792078972, 'timing/model': 1.1243513198569417, 'learning_rate/qwen_vl_interface': 7.142405176769648e-06, 'learning_rate/action_model': 7.142405176769648e-05, 'epoch': 0.97}) +[2026-06-17 21:40:24,008][starVLA.training.train_starvla][INFO] - Step 2651, Loss: {'train/loss': 0.6230536103248596, 'train/grad_norm_pre_clip': 14.10531997680664, 'timing/data': 0.0009787017479538918, 'timing/model': 1.411104523576796, 'learning_rate/qwen_vl_interface': 7.140421336434102e-06, 'learning_rate/action_model': 7.140421336434102e-05, 'epoch': 0.97}) +[2026-06-17 21:40:29,206][starVLA.training.train_starvla][INFO] - Step 2652, Loss: {'train/loss': 0.33582326769828796, 'train/grad_norm_pre_clip': 9.725865364074707, 'timing/data': 0.0010064374655485153, 'timing/model': 1.3188715539872646, 'learning_rate/qwen_vl_interface': 7.138437104212176e-06, 'learning_rate/action_model': 7.138437104212176e-05, 'epoch': 0.97}) +[2026-06-17 21:40:41,935][starVLA.training.train_starvla][INFO] - Step 2653, Loss: {'train/loss': 1.048100471496582, 'train/grad_norm_pre_clip': 17.140310287475586, 'timing/data': 0.003097125329077244, 'timing/model': 4.30682372674346, 'learning_rate/qwen_vl_interface': 7.136452480515208e-06, 'learning_rate/action_model': 7.136452480515208e-05, 'epoch': 0.97}) +[2026-06-17 21:40:48,196][starVLA.training.train_starvla][INFO] - Step 2654, Loss: {'train/loss': 0.6169120669364929, 'train/grad_norm_pre_clip': 14.953134536743164, 'timing/data': 0.001270974986255169, 'timing/model': 1.1871292842552066, 'learning_rate/qwen_vl_interface': 7.134467465754612e-06, 'learning_rate/action_model': 7.134467465754612e-05, 'epoch': 0.97}) +[2026-06-17 21:40:58,113][starVLA.training.train_starvla][INFO] - Step 2655, Loss: {'train/loss': 0.5475658774375916, 'train/grad_norm_pre_clip': 12.212778091430664, 'timing/data': 0.0011865729466080666, 'timing/model': 1.1816677367314696, 'learning_rate/qwen_vl_interface': 7.132482060341881e-06, 'learning_rate/action_model': 7.13248206034188e-05, 'epoch': 0.97}) +[2026-06-17 21:41:03,176][starVLA.training.train_starvla][INFO] - Step 2656, Loss: {'train/loss': 0.8140557408332825, 'train/grad_norm_pre_clip': 17.914142608642578, 'timing/data': 0.0025687403976917267, 'timing/model': 1.759328975342214, 'learning_rate/qwen_vl_interface': 7.1304962646885935e-06, 'learning_rate/action_model': 7.130496264688593e-05, 'epoch': 0.97}) +[2026-06-17 21:41:11,421][starVLA.training.train_starvla][INFO] - Step 2657, Loss: {'train/loss': 0.9388512372970581, 'train/grad_norm_pre_clip': 17.45393943786621, 'timing/data': 0.0013558659702539444, 'timing/model': 3.0113603873178363, 'learning_rate/qwen_vl_interface': 7.128510079206408e-06, 'learning_rate/action_model': 7.128510079206407e-05, 'epoch': 0.97}) +[2026-06-17 21:41:16,234][starVLA.training.train_starvla][INFO] - Step 2658, Loss: {'train/loss': 0.5402709245681763, 'train/grad_norm_pre_clip': 16.128835678100586, 'timing/data': 0.0008088955655694008, 'timing/model': 1.2228067694231868, 'learning_rate/qwen_vl_interface': 7.126523504307061e-06, 'learning_rate/action_model': 7.126523504307061e-05, 'epoch': 0.97}) +[2026-06-17 21:41:25,845][starVLA.training.train_starvla][INFO] - Step 2659, Loss: {'train/loss': 0.6183025240898132, 'train/grad_norm_pre_clip': 16.328067779541016, 'timing/data': 0.0009997133165597916, 'timing/model': 1.1364266872406006, 'learning_rate/qwen_vl_interface': 7.124536540402371e-06, 'learning_rate/action_model': 7.124536540402372e-05, 'epoch': 0.97}) +[2026-06-17 21:41:35,628][starVLA.training.train_starvla][INFO] - Step 2660, Loss: {'train/loss': 0.9186463356018066, 'train/grad_norm_pre_clip': 20.29481315612793, 'timing/data': 0.0009652990847826004, 'timing/model': 4.223148124292493, 'learning_rate/qwen_vl_interface': 7.1225491879042405e-06, 'learning_rate/action_model': 7.12254918790424e-05, 'epoch': 0.97}) +[2026-06-17 21:41:40,259][starVLA.training.train_starvla][INFO] - Step 2661, Loss: {'train/loss': 0.5218163728713989, 'train/grad_norm_pre_clip': 12.879222869873047, 'timing/data': 0.001468028873205185, 'timing/model': 1.3049255823716521, 'learning_rate/qwen_vl_interface': 7.120561447224649e-06, 'learning_rate/action_model': 7.120561447224649e-05, 'epoch': 0.97}) +[2026-06-17 21:41:44,736][starVLA.training.train_starvla][INFO] - Step 2662, Loss: {'train/loss': 0.6809813976287842, 'train/grad_norm_pre_clip': 17.209875106811523, 'timing/data': 0.0009912597015500069, 'timing/model': 1.1389272902160883, 'learning_rate/qwen_vl_interface': 7.118573318775656e-06, 'learning_rate/action_model': 7.118573318775655e-05, 'epoch': 0.97}) +[2026-06-17 21:41:49,190][starVLA.training.train_starvla][INFO] - Step 2663, Loss: {'train/loss': 0.6075433492660522, 'train/grad_norm_pre_clip': 13.86070728302002, 'timing/data': 0.0011289920657873154, 'timing/model': 1.089023587293923, 'learning_rate/qwen_vl_interface': 7.116584802969403e-06, 'learning_rate/action_model': 7.116584802969402e-05, 'epoch': 0.98}) +[2026-06-17 21:41:53,518][starVLA.training.train_starvla][INFO] - Step 2664, Loss: {'train/loss': 0.4566161334514618, 'train/grad_norm_pre_clip': 12.860808372497559, 'timing/data': 0.0011524287983775139, 'timing/model': 1.0499480403959751, 'learning_rate/qwen_vl_interface': 7.114595900218112e-06, 'learning_rate/action_model': 7.114595900218112e-05, 'epoch': 0.98}) +[2026-06-17 21:41:57,978][starVLA.training.train_starvla][INFO] - Step 2665, Loss: {'train/loss': 0.5082018971443176, 'train/grad_norm_pre_clip': 12.542895317077637, 'timing/data': 0.0009358255192637444, 'timing/model': 1.076726345345378, 'learning_rate/qwen_vl_interface': 7.112606610934084e-06, 'learning_rate/action_model': 7.112606610934084e-05, 'epoch': 0.98}) +[2026-06-17 21:42:02,406][starVLA.training.train_starvla][INFO] - Step 2666, Loss: {'train/loss': 0.6305949091911316, 'train/grad_norm_pre_clip': 13.129826545715332, 'timing/data': 0.001068473793566227, 'timing/model': 1.3623695885762572, 'learning_rate/qwen_vl_interface': 7.110616935529702e-06, 'learning_rate/action_model': 7.110616935529702e-05, 'epoch': 0.98}) +[2026-06-17 21:42:06,492][starVLA.training.train_starvla][INFO] - Step 2667, Loss: {'train/loss': 0.7842857241630554, 'train/grad_norm_pre_clip': 15.97630500793457, 'timing/data': 0.0008725039660930634, 'timing/model': 1.0266386466100812, 'learning_rate/qwen_vl_interface': 7.108626874417426e-06, 'learning_rate/action_model': 7.108626874417426e-05, 'epoch': 0.98}) +[2026-06-17 21:42:11,133][starVLA.training.train_starvla][INFO] - Step 2668, Loss: {'train/loss': 0.48673900961875916, 'train/grad_norm_pre_clip': 13.988317489624023, 'timing/data': 0.0018668044358491898, 'timing/model': 1.1112449504435062, 'learning_rate/qwen_vl_interface': 7.106636428009801e-06, 'learning_rate/action_model': 7.1066364280098e-05, 'epoch': 0.98}) +[2026-06-17 21:42:15,464][starVLA.training.train_starvla][INFO] - Step 2669, Loss: {'train/loss': 0.7992221117019653, 'train/grad_norm_pre_clip': 18.206161499023438, 'timing/data': 0.0010465150699019432, 'timing/model': 1.0186638366430998, 'learning_rate/qwen_vl_interface': 7.104645596719445e-06, 'learning_rate/action_model': 7.104645596719445e-05, 'epoch': 0.98}) +[2026-06-17 21:42:19,857][starVLA.training.train_starvla][INFO] - Step 2670, Loss: {'train/loss': 0.733238697052002, 'train/grad_norm_pre_clip': 15.100296974182129, 'timing/data': 0.0008840933442115784, 'timing/model': 1.0597824081778526, 'learning_rate/qwen_vl_interface': 7.102654380959063e-06, 'learning_rate/action_model': 7.102654380959063e-05, 'epoch': 0.98}) +[2026-06-17 21:42:24,314][starVLA.training.train_starvla][INFO] - Step 2671, Loss: {'train/loss': 0.5658835172653198, 'train/grad_norm_pre_clip': 15.43511962890625, 'timing/data': 0.000975840725004673, 'timing/model': 1.4139417139813304, 'learning_rate/qwen_vl_interface': 7.100662781141432e-06, 'learning_rate/action_model': 7.100662781141432e-05, 'epoch': 0.98}) +[2026-06-17 21:42:28,399][starVLA.training.train_starvla][INFO] - Step 2672, Loss: {'train/loss': 1.0908409357070923, 'train/grad_norm_pre_clip': 22.545270919799805, 'timing/data': 0.0013206405565142632, 'timing/model': 1.0756042804569006, 'learning_rate/qwen_vl_interface': 7.0986707976794155e-06, 'learning_rate/action_model': 7.098670797679415e-05, 'epoch': 0.98}) +[2026-06-17 21:42:32,764][starVLA.training.train_starvla][INFO] - Step 2673, Loss: {'train/loss': 0.524101734161377, 'train/grad_norm_pre_clip': 14.096770286560059, 'timing/data': 0.0014876676723361015, 'timing/model': 1.0217859651893377, 'learning_rate/qwen_vl_interface': 7.096678430985955e-06, 'learning_rate/action_model': 7.096678430985954e-05, 'epoch': 0.98}) +[2026-06-17 21:42:37,339][starVLA.training.train_starvla][INFO] - Step 2674, Loss: {'train/loss': 1.2308976650238037, 'train/grad_norm_pre_clip': 22.001220703125, 'timing/data': 0.001393650658428669, 'timing/model': 1.194599336013198, 'learning_rate/qwen_vl_interface': 7.094685681474066e-06, 'learning_rate/action_model': 7.094685681474065e-05, 'epoch': 0.98}) +[2026-06-17 21:42:41,938][starVLA.training.train_starvla][INFO] - Step 2675, Loss: {'train/loss': 0.4544558823108673, 'train/grad_norm_pre_clip': 20.085947036743164, 'timing/data': 0.0011893566697835922, 'timing/model': 1.1798309590667486, 'learning_rate/qwen_vl_interface': 7.092692549556853e-06, 'learning_rate/action_model': 7.092692549556853e-05, 'epoch': 0.98}) +[2026-06-17 21:42:49,406][starVLA.training.train_starvla][INFO] - Step 2676, Loss: {'train/loss': 0.8344186544418335, 'train/grad_norm_pre_clip': 17.4029541015625, 'timing/data': 0.0011310381814837456, 'timing/model': 4.166954726912081, 'learning_rate/qwen_vl_interface': 7.0906990356474895e-06, 'learning_rate/action_model': 7.090699035647489e-05, 'epoch': 0.98}) +[2026-06-17 21:42:54,801][starVLA.training.train_starvla][INFO] - Step 2677, Loss: {'train/loss': 0.9127275943756104, 'train/grad_norm_pre_clip': 20.672073364257812, 'timing/data': 0.001118633896112442, 'timing/model': 1.2449734890833497, 'learning_rate/qwen_vl_interface': 7.088705140159238e-06, 'learning_rate/action_model': 7.088705140159237e-05, 'epoch': 0.98}) +[2026-06-17 21:43:02,678][starVLA.training.train_starvla][INFO] - Step 2678, Loss: {'train/loss': 0.7249085307121277, 'train/grad_norm_pre_clip': 17.238908767700195, 'timing/data': 0.0010895440354943275, 'timing/model': 2.633781290613115, 'learning_rate/qwen_vl_interface': 7.086710863505428e-06, 'learning_rate/action_model': 7.086710863505428e-05, 'epoch': 0.98}) +[2026-06-17 21:43:14,843][starVLA.training.train_starvla][INFO] - Step 2679, Loss: {'train/loss': 0.819872260093689, 'train/grad_norm_pre_clip': 17.46231460571289, 'timing/data': 0.0010627694427967072, 'timing/model': 1.176748426631093, 'learning_rate/qwen_vl_interface': 7.084716206099485e-06, 'learning_rate/action_model': 7.084716206099484e-05, 'epoch': 0.98}) +[2026-06-17 21:43:20,303][starVLA.training.train_starvla][INFO] - Step 2680, Loss: {'train/loss': 0.3677544593811035, 'train/grad_norm_pre_clip': 12.950221061706543, 'timing/data': 0.0009171674028038979, 'timing/model': 1.1033740798011422, 'learning_rate/qwen_vl_interface': 7.082721168354894e-06, 'learning_rate/action_model': 7.082721168354894e-05, 'epoch': 0.98}) +[2026-06-17 21:43:28,970][starVLA.training.train_starvla][INFO] - Step 2681, Loss: {'train/loss': 0.7248139977455139, 'train/grad_norm_pre_clip': 15.346517562866211, 'timing/data': 0.0010070335119962692, 'timing/model': 2.0314066922292113, 'learning_rate/qwen_vl_interface': 7.080725750685235e-06, 'learning_rate/action_model': 7.080725750685234e-05, 'epoch': 0.98}) +[2026-06-17 21:43:36,583][starVLA.training.train_starvla][INFO] - Step 2682, Loss: {'train/loss': 0.486660897731781, 'train/grad_norm_pre_clip': 12.385259628295898, 'timing/data': 0.0013568317517638206, 'timing/model': 1.156522100791335, 'learning_rate/qwen_vl_interface': 7.078729953504157e-06, 'learning_rate/action_model': 7.078729953504156e-05, 'epoch': 0.98}) +[2026-06-17 21:43:46,492][starVLA.training.train_starvla][INFO] - Step 2683, Loss: {'train/loss': 0.763092041015625, 'train/grad_norm_pre_clip': 15.550494194030762, 'timing/data': 0.0008810535073280334, 'timing/model': 4.379742625169456, 'learning_rate/qwen_vl_interface': 7.076733777225391e-06, 'learning_rate/action_model': 7.07673377722539e-05, 'epoch': 0.98}) +[2026-06-17 21:43:54,769][starVLA.training.train_starvla][INFO] - Step 2684, Loss: {'train/loss': 0.5374534726142883, 'train/grad_norm_pre_clip': 26.898937225341797, 'timing/data': 0.007548618130385876, 'timing/model': 1.2531419266015291, 'learning_rate/qwen_vl_interface': 7.074737222262749e-06, 'learning_rate/action_model': 7.074737222262748e-05, 'epoch': 0.98}) +[2026-06-17 21:43:59,464][starVLA.training.train_starvla][INFO] - Step 2685, Loss: {'train/loss': 0.7007853388786316, 'train/grad_norm_pre_clip': 19.271501541137695, 'timing/data': 0.001333010382950306, 'timing/model': 1.1666696444153786, 'learning_rate/qwen_vl_interface': 7.072740289030115e-06, 'learning_rate/action_model': 7.072740289030115e-05, 'epoch': 0.98}) +[2026-06-17 21:44:03,928][starVLA.training.train_starvla][INFO] - Step 2686, Loss: {'train/loss': 0.6456438899040222, 'train/grad_norm_pre_clip': 16.672069549560547, 'timing/data': 0.0008964324370026588, 'timing/model': 1.3548931293189526, 'learning_rate/qwen_vl_interface': 7.07074297794146e-06, 'learning_rate/action_model': 7.07074297794146e-05, 'epoch': 0.98}) +[2026-06-17 21:44:08,032][starVLA.training.train_starvla][INFO] - Step 2687, Loss: {'train/loss': 0.589594841003418, 'train/grad_norm_pre_clip': 13.334715843200684, 'timing/data': 0.00107529666274786, 'timing/model': 1.0521691404283047, 'learning_rate/qwen_vl_interface': 7.068745289410826e-06, 'learning_rate/action_model': 7.068745289410827e-05, 'epoch': 0.98}) +[2026-06-17 21:44:12,651][starVLA.training.train_starvla][INFO] - Step 2688, Loss: {'train/loss': 0.7417519688606262, 'train/grad_norm_pre_clip': 17.515188217163086, 'timing/data': 0.000942084938287735, 'timing/model': 1.117947063408792, 'learning_rate/qwen_vl_interface': 7.066747223852337e-06, 'learning_rate/action_model': 7.066747223852337e-05, 'epoch': 0.98}) +[2026-06-17 21:44:17,084][starVLA.training.train_starvla][INFO] - Step 2689, Loss: {'train/loss': 0.7775309085845947, 'train/grad_norm_pre_clip': 15.22586441040039, 'timing/data': 0.0007613841444253922, 'timing/model': 1.0311444839462638, 'learning_rate/qwen_vl_interface': 7.064748781680192e-06, 'learning_rate/action_model': 7.064748781680192e-05, 'epoch': 0.98}) +[2026-06-17 21:44:21,482][starVLA.training.train_starvla][INFO] - Step 2690, Loss: {'train/loss': 0.7499046921730042, 'train/grad_norm_pre_clip': 16.645841598510742, 'timing/data': 0.0010915640741586685, 'timing/model': 1.0251582134515047, 'learning_rate/qwen_vl_interface': 7.062749963308673e-06, 'learning_rate/action_model': 7.062749963308672e-05, 'epoch': 0.99}) +[2026-06-17 21:44:25,950][starVLA.training.train_starvla][INFO] - Step 2691, Loss: {'train/loss': 0.7756950259208679, 'train/grad_norm_pre_clip': 13.418593406677246, 'timing/data': 0.0008378801867365837, 'timing/model': 1.4499542182311416, 'learning_rate/qwen_vl_interface': 7.060750769152135e-06, 'learning_rate/action_model': 7.060750769152135e-05, 'epoch': 0.99}) +[2026-06-17 21:44:30,213][starVLA.training.train_starvla][INFO] - Step 2692, Loss: {'train/loss': 0.5673559904098511, 'train/grad_norm_pre_clip': 11.164134979248047, 'timing/data': 0.0008695106953382492, 'timing/model': 1.0564164873212576, 'learning_rate/qwen_vl_interface': 7.058751199625014e-06, 'learning_rate/action_model': 7.058751199625015e-05, 'epoch': 0.99}) +[2026-06-17 21:44:35,957][starVLA.training.train_starvla][INFO] - Step 2693, Loss: {'train/loss': 0.36227917671203613, 'train/grad_norm_pre_clip': 9.957803726196289, 'timing/data': 0.0008443081751465797, 'timing/model': 2.2073682276532054, 'learning_rate/qwen_vl_interface': 7.056751255141824e-06, 'learning_rate/action_model': 7.056751255141824e-05, 'epoch': 0.99}) +[2026-06-17 21:44:40,791][starVLA.training.train_starvla][INFO] - Step 2694, Loss: {'train/loss': 0.5367689728736877, 'train/grad_norm_pre_clip': 12.140649795532227, 'timing/data': 0.0008760783821344376, 'timing/model': 1.2783537982031703, 'learning_rate/qwen_vl_interface': 7.054750936117155e-06, 'learning_rate/action_model': 7.054750936117156e-05, 'epoch': 0.99}) +[2026-06-17 21:44:52,842][starVLA.training.train_starvla][INFO] - Step 2695, Loss: {'train/loss': 0.6422589421272278, 'train/grad_norm_pre_clip': 12.850310325622559, 'timing/data': 0.003933251835405827, 'timing/model': 1.532319844700396, 'learning_rate/qwen_vl_interface': 7.0527502429656745e-06, 'learning_rate/action_model': 7.052750242965674e-05, 'epoch': 0.99}) +[2026-06-17 21:44:57,768][starVLA.training.train_starvla][INFO] - Step 2696, Loss: {'train/loss': 0.39798420667648315, 'train/grad_norm_pre_clip': 12.91672134399414, 'timing/data': 0.0011455826461315155, 'timing/model': 1.506404148414731, 'learning_rate/qwen_vl_interface': 7.050749176102128e-06, 'learning_rate/action_model': 7.050749176102129e-05, 'epoch': 0.99}) +[2026-06-17 21:45:05,291][starVLA.training.train_starvla][INFO] - Step 2697, Loss: {'train/loss': 0.585464596748352, 'train/grad_norm_pre_clip': 11.589644432067871, 'timing/data': 0.0013452405110001564, 'timing/model': 1.0335420034825802, 'learning_rate/qwen_vl_interface': 7.048747735941342e-06, 'learning_rate/action_model': 7.048747735941341e-05, 'epoch': 0.99}) +[2026-06-17 21:45:10,121][starVLA.training.train_starvla][INFO] - Step 2698, Loss: {'train/loss': 0.5286599397659302, 'train/grad_norm_pre_clip': 13.244712829589844, 'timing/data': 0.0010956395417451859, 'timing/model': 1.118776905350387, 'learning_rate/qwen_vl_interface': 7.046745922898212e-06, 'learning_rate/action_model': 7.046745922898211e-05, 'epoch': 0.99}) +[2026-06-17 21:45:14,794][starVLA.training.train_starvla][INFO] - Step 2699, Loss: {'train/loss': 0.5650758743286133, 'train/grad_norm_pre_clip': 14.997289657592773, 'timing/data': 0.0010872455313801765, 'timing/model': 1.1094022691249847, 'learning_rate/qwen_vl_interface': 7.04474373738772e-06, 'learning_rate/action_model': 7.04474373738772e-05, 'epoch': 0.99}) +[2026-06-17 21:48:42,487][starVLA.training.train_starvla][INFO] - Step 2700, Loss: {'train/loss': 0.48453933000564575, 'train/grad_norm_pre_clip': 10.078662872314453, 'eval/loss': 0.6960732936859131, 'eval/latency_2/loss': 0.6960732936859131, 'eval/demon_fire_precision': 0.8977635782747604, 'eval/demon_fire_recall': 0.9405857740585774, 'eval/demon_fire_f1': 0.9186759297098488, 'eval/demon_fire_rate_model': 0.6967167501391207, 'eval/demon_fire_rate_teacher': 0.6649972175848636, 'eval/demon_left_precision': 0.7760467380720545, 'eval/demon_left_recall': 0.8389473684210527, 'eval/demon_left_f1': 0.806272129489125, 'eval/demon_left_rate_model': 0.2857540345019477, 'eval/demon_left_rate_teacher': 0.2643294379521425, 'eval/demon_right_precision': 0.8256467941507312, 'eval/demon_right_recall': 0.8293785310734463, 'eval/demon_right_f1': 0.8275084554678691, 'eval/demon_right_rate_model': 0.24735670562047857, 'eval/demon_right_rate_teacher': 0.24624373956594323, 'eval/demon_move_f1': 0.8168902924784971, 'eval/demon_cc_f1': 0.8855728563846368, 'eval/latency_2/demon_fire_precision': 0.8977635782747604, 'eval/latency_2/demon_fire_recall': 0.9405857740585774, 'eval/latency_2/demon_fire_f1': 0.9186759297098488, 'eval/latency_2/demon_fire_rate_model': 0.6967167501391207, 'eval/latency_2/demon_fire_rate_teacher': 0.6649972175848636, 'eval/latency_2/demon_left_precision': 0.7760467380720545, 'eval/latency_2/demon_left_recall': 0.8389473684210527, 'eval/latency_2/demon_left_f1': 0.806272129489125, 'eval/latency_2/demon_left_rate_model': 0.2857540345019477, 'eval/latency_2/demon_left_rate_teacher': 0.2643294379521425, 'eval/latency_2/demon_right_precision': 0.8256467941507312, 'eval/latency_2/demon_right_recall': 0.8293785310734463, 'eval/latency_2/demon_right_f1': 0.8275084554678691, 'eval/latency_2/demon_right_rate_model': 0.24735670562047857, 'eval/latency_2/demon_right_rate_teacher': 0.24624373956594323, 'eval/latency_2/demon_move_f1': 0.8168902924784971, 'eval/latency_2/demon_cc_f1': 0.8855728563846368, 'timing/data': 0.001180458813905716, 'timing/model': 1.1947660716250539, 'learning_rate/qwen_vl_interface': 7.042741179824919e-06, 'learning_rate/action_model': 7.042741179824918e-05, 'epoch': 0.99}) +[2026-06-17 21:48:42,520][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_2700_state +[2026-06-17 21:48:42,530][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 21:48:50,929][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_2700_state/model.safetensors +[2026-06-17 21:49:06,252][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_2700_state/optimizer.bin +[2026-06-17 21:49:06,253][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_2700_state/sampler.bin +[2026-06-17 21:49:06,254][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_2700_state/sampler_1.bin +[2026-06-17 21:49:06,256][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_2700_state/random_states_0.pkl +[2026-06-17 21:49:06,261][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2700 +[2026-06-17 21:49:06,262][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 21:49:06,278][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 21:49:10,988][starVLA.training.train_starvla][INFO] - Step 2701, Loss: {'train/loss': 0.9507283568382263, 'train/grad_norm_pre_clip': 15.930864334106445, 'timing/data': 0.0009098881855607033, 'timing/model': 1.0702254204079509, 'learning_rate/qwen_vl_interface': 7.040738250624944e-06, 'learning_rate/action_model': 7.040738250624944e-05, 'epoch': 0.99}) +[2026-06-17 21:49:16,063][starVLA.training.train_starvla][INFO] - Step 2702, Loss: {'train/loss': 0.6351925134658813, 'train/grad_norm_pre_clip': 13.706668853759766, 'timing/data': 0.001255640760064125, 'timing/model': 1.5067015970125794, 'learning_rate/qwen_vl_interface': 7.038734950203001e-06, 'learning_rate/action_model': 7.038734950203001e-05, 'epoch': 0.99}) +[2026-06-17 21:49:24,626][starVLA.training.train_starvla][INFO] - Step 2703, Loss: {'train/loss': 0.6574082374572754, 'train/grad_norm_pre_clip': 14.425851821899414, 'timing/data': 0.0014006933197379112, 'timing/model': 1.2718293955549598, 'learning_rate/qwen_vl_interface': 7.036731278974379e-06, 'learning_rate/action_model': 7.036731278974378e-05, 'epoch': 0.99}) +[2026-06-17 21:49:29,253][starVLA.training.train_starvla][INFO] - Step 2704, Loss: {'train/loss': 0.5948262810707092, 'train/grad_norm_pre_clip': 12.392507553100586, 'timing/data': 0.0011706538498401642, 'timing/model': 1.0299086198210716, 'learning_rate/qwen_vl_interface': 7.034727237354442e-06, 'learning_rate/action_model': 7.034727237354441e-05, 'epoch': 0.99}) +[2026-06-17 21:49:37,800][starVLA.training.train_starvla][INFO] - Step 2705, Loss: {'train/loss': 0.6735976338386536, 'train/grad_norm_pre_clip': 14.317061424255371, 'timing/data': 0.000851173885166645, 'timing/model': 5.0940136993303895, 'learning_rate/qwen_vl_interface': 7.0327228257586244e-06, 'learning_rate/action_model': 7.032722825758625e-05, 'epoch': 0.99}) +[2026-06-17 21:49:54,226][starVLA.training.train_starvla][INFO] - Step 2706, Loss: {'train/loss': 0.6058582067489624, 'train/grad_norm_pre_clip': 14.222147941589355, 'timing/data': 0.0009254282340407372, 'timing/model': 1.2432822585105896, 'learning_rate/qwen_vl_interface': 7.030718044602451e-06, 'learning_rate/action_model': 7.03071804460245e-05, 'epoch': 0.99}) +[2026-06-17 21:50:04,590][starVLA.training.train_starvla][INFO] - Step 2707, Loss: {'train/loss': 0.41056451201438904, 'train/grad_norm_pre_clip': 10.299699783325195, 'timing/data': 0.0012278761714696884, 'timing/model': 1.545163574628532, 'learning_rate/qwen_vl_interface': 7.028712894301507e-06, 'learning_rate/action_model': 7.028712894301506e-05, 'epoch': 0.99}) +[2026-06-17 21:50:08,758][starVLA.training.train_starvla][INFO] - Step 2708, Loss: {'train/loss': 0.697425127029419, 'train/grad_norm_pre_clip': 14.839652061462402, 'timing/data': 0.0009100977331399918, 'timing/model': 1.0349386585876346, 'learning_rate/qwen_vl_interface': 7.026707375271469e-06, 'learning_rate/action_model': 7.026707375271468e-05, 'epoch': 0.99}) +[2026-06-17 21:50:13,307][starVLA.training.train_starvla][INFO] - Step 2709, Loss: {'train/loss': 0.7880703806877136, 'train/grad_norm_pre_clip': 15.63615894317627, 'timing/data': 0.0010716430842876434, 'timing/model': 1.0473012579604983, 'learning_rate/qwen_vl_interface': 7.024701487928079e-06, 'learning_rate/action_model': 7.024701487928078e-05, 'epoch': 0.99}) +[2026-06-17 21:50:17,962][starVLA.training.train_starvla][INFO] - Step 2710, Loss: {'train/loss': 0.411693274974823, 'train/grad_norm_pre_clip': 11.570229530334473, 'timing/data': 0.0007881950587034225, 'timing/model': 1.0386016741394997, 'learning_rate/qwen_vl_interface': 7.02269523268716e-06, 'learning_rate/action_model': 7.022695232687159e-05, 'epoch': 0.99}) +[2026-06-17 21:50:22,460][starVLA.training.train_starvla][INFO] - Step 2711, Loss: {'train/loss': 0.8921445608139038, 'train/grad_norm_pre_clip': 16.65962791442871, 'timing/data': 0.002230696380138397, 'timing/model': 1.0511877248063684, 'learning_rate/qwen_vl_interface': 7.020688609964612e-06, 'learning_rate/action_model': 7.020688609964612e-05, 'epoch': 0.99}) +[2026-06-17 21:50:26,701][starVLA.training.train_starvla][INFO] - Step 2712, Loss: {'train/loss': 0.6031169295310974, 'train/grad_norm_pre_clip': 18.359058380126953, 'timing/data': 0.0009121531620621681, 'timing/model': 1.1207345621660352, 'learning_rate/qwen_vl_interface': 7.018681620176412e-06, 'learning_rate/action_model': 7.018681620176412e-05, 'epoch': 0.99}) +[2026-06-17 21:50:31,108][starVLA.training.train_starvla][INFO] - Step 2713, Loss: {'train/loss': 0.8192559480667114, 'train/grad_norm_pre_clip': 18.487253189086914, 'timing/data': 0.0008640913292765617, 'timing/model': 1.035424618050456, 'learning_rate/qwen_vl_interface': 7.016674263738606e-06, 'learning_rate/action_model': 7.016674263738607e-05, 'epoch': 0.99}) +[2026-06-17 21:50:35,615][starVLA.training.train_starvla][INFO] - Step 2714, Loss: {'train/loss': 0.8235979080200195, 'train/grad_norm_pre_clip': 15.988710403442383, 'timing/data': 0.0011908737942576408, 'timing/model': 1.0381772769615054, 'learning_rate/qwen_vl_interface': 7.014666541067325e-06, 'learning_rate/action_model': 7.014666541067325e-05, 'epoch': 0.99}) +[2026-06-17 21:50:40,049][starVLA.training.train_starvla][INFO] - Step 2715, Loss: {'train/loss': 0.5765824317932129, 'train/grad_norm_pre_clip': 11.844475746154785, 'timing/data': 0.0008109509944915771, 'timing/model': 1.031192533671856, 'learning_rate/qwen_vl_interface': 7.012658452578774e-06, 'learning_rate/action_model': 7.012658452578773e-05, 'epoch': 0.99}) +[2026-06-17 21:50:44,542][starVLA.training.train_starvla][INFO] - Step 2716, Loss: {'train/loss': 0.5434644222259521, 'train/grad_norm_pre_clip': 12.772039413452148, 'timing/data': 0.001069716177880764, 'timing/model': 1.064420928247273, 'learning_rate/qwen_vl_interface': 7.010649998689225e-06, 'learning_rate/action_model': 7.010649998689224e-05, 'epoch': 0.99}) +[2026-06-17 21:50:49,036][starVLA.training.train_starvla][INFO] - Step 2717, Loss: {'train/loss': 0.3009635806083679, 'train/grad_norm_pre_clip': 11.146123886108398, 'timing/data': 0.0009735589846968651, 'timing/model': 1.4208698375150561, 'learning_rate/qwen_vl_interface': 7.0086411798150385e-06, 'learning_rate/action_model': 7.008641179815038e-05, 'epoch': 0.99}) +[2026-06-17 21:50:58,228][starVLA.training.train_starvla][INFO] - Step 2718, Loss: {'train/loss': 0.4918747842311859, 'train/grad_norm_pre_clip': 11.517027854919434, 'timing/data': 0.0011554686352610588, 'timing/model': 1.243162195198238, 'learning_rate/qwen_vl_interface': 7.006631996372642e-06, 'learning_rate/action_model': 7.006631996372643e-05, 'epoch': 1.0}) +[2026-06-17 21:51:02,877][starVLA.training.train_starvla][INFO] - Step 2719, Loss: {'train/loss': 0.5570533275604248, 'train/grad_norm_pre_clip': 13.382562637329102, 'timing/data': 0.0024516861885786057, 'timing/model': 1.3343789977952838, 'learning_rate/qwen_vl_interface': 7.004622448778545e-06, 'learning_rate/action_model': 7.004622448778544e-05, 'epoch': 1.0}) +[2026-06-17 21:51:10,532][starVLA.training.train_starvla][INFO] - Step 2720, Loss: {'train/loss': 0.5054532885551453, 'train/grad_norm_pre_clip': 11.894979476928711, 'timing/data': 0.0008877208456397057, 'timing/model': 4.124082433991134, 'learning_rate/qwen_vl_interface': 7.0026125374493244e-06, 'learning_rate/action_model': 7.002612537449324e-05, 'epoch': 1.0}) +[2026-06-17 21:51:17,275][starVLA.training.train_starvla][INFO] - Step 2721, Loss: {'train/loss': 0.9508956670761108, 'train/grad_norm_pre_clip': 21.82118034362793, 'timing/data': 0.0010333983227610588, 'timing/model': 1.1957153137773275, 'learning_rate/qwen_vl_interface': 7.000602262801639e-06, 'learning_rate/action_model': 7.000602262801638e-05, 'epoch': 1.0}) +[2026-06-17 21:51:22,190][starVLA.training.train_starvla][INFO] - Step 2722, Loss: {'train/loss': 0.59462970495224, 'train/grad_norm_pre_clip': 16.05466651916504, 'timing/data': 0.0009463932365179062, 'timing/model': 1.4985768562182784, 'learning_rate/qwen_vl_interface': 6.9985916252522225e-06, 'learning_rate/action_model': 6.998591625252222e-05, 'epoch': 1.0}) +[2026-06-17 21:51:26,632][starVLA.training.train_starvla][INFO] - Step 2723, Loss: {'train/loss': 0.6043609976768494, 'train/grad_norm_pre_clip': 15.474818229675293, 'timing/data': 0.001046590507030487, 'timing/model': 1.1518167350441217, 'learning_rate/qwen_vl_interface': 6.9965806252178785e-06, 'learning_rate/action_model': 6.996580625217877e-05, 'epoch': 1.0}) +[2026-06-17 21:51:31,199][starVLA.training.train_starvla][INFO] - Step 2724, Loss: {'train/loss': 0.4165036976337433, 'train/grad_norm_pre_clip': 13.360236167907715, 'timing/data': 0.0013248072937130928, 'timing/model': 1.0688558164983988, 'learning_rate/qwen_vl_interface': 6.994569263115491e-06, 'learning_rate/action_model': 6.994569263115491e-05, 'epoch': 1.0}) +[2026-06-17 21:51:36,004][starVLA.training.train_starvla][INFO] - Step 2725, Loss: {'train/loss': 0.5943920016288757, 'train/grad_norm_pre_clip': 15.709977149963379, 'timing/data': 0.0008897027000784874, 'timing/model': 1.1244076248258352, 'learning_rate/qwen_vl_interface': 6.992557539362019e-06, 'learning_rate/action_model': 6.99255753936202e-05, 'epoch': 1.0}) +[2026-06-17 21:51:40,564][starVLA.training.train_starvla][INFO] - Step 2726, Loss: {'train/loss': 0.5987687110900879, 'train/grad_norm_pre_clip': 13.880043983459473, 'timing/data': 0.0010991934686899185, 'timing/model': 1.1553540891036391, 'learning_rate/qwen_vl_interface': 6.990545454374495e-06, 'learning_rate/action_model': 6.990545454374494e-05, 'epoch': 1.0}) +[2026-06-17 21:51:45,213][starVLA.training.train_starvla][INFO] - Step 2727, Loss: {'train/loss': 0.568863034248352, 'train/grad_norm_pre_clip': 16.473417282104492, 'timing/data': 0.0010852469131350517, 'timing/model': 1.4365016398951411, 'learning_rate/qwen_vl_interface': 6.988533008570024e-06, 'learning_rate/action_model': 6.988533008570024e-05, 'epoch': 1.0}) +[2026-06-17 21:51:55,267][starVLA.training.train_starvla][INFO] - Step 2728, Loss: {'train/loss': 0.8462345600128174, 'train/grad_norm_pre_clip': 16.06578826904297, 'timing/data': 0.0008915718644857407, 'timing/model': 1.049528836272657, 'learning_rate/qwen_vl_interface': 6.98652020236579e-06, 'learning_rate/action_model': 6.986520202365789e-05, 'epoch': 1.0}) +[2026-06-17 21:51:59,971][starVLA.training.train_starvla][INFO] - Step 2729, Loss: {'train/loss': 0.6995176076889038, 'train/grad_norm_pre_clip': 15.224125862121582, 'timing/data': 0.0009636431932449341, 'timing/model': 1.181351994164288, 'learning_rate/qwen_vl_interface': 6.98450703617905e-06, 'learning_rate/action_model': 6.98450703617905e-05, 'epoch': 1.0}) +[2026-06-17 21:52:08,513][starVLA.training.train_starvla][INFO] - Step 2730, Loss: {'train/loss': 0.7688318490982056, 'train/grad_norm_pre_clip': 16.919208526611328, 'timing/data': 0.0012656766921281815, 'timing/model': 2.7053551487624645, 'learning_rate/qwen_vl_interface': 6.982493510427134e-06, 'learning_rate/action_model': 6.982493510427133e-05, 'epoch': 1.0}) +[2026-06-17 21:52:14,426][starVLA.training.train_starvla][INFO] - Step 2731, Loss: {'train/loss': 0.8107951283454895, 'train/grad_norm_pre_clip': 24.825580596923828, 'timing/data': 0.0008293837308883667, 'timing/model': 2.3894889540970325, 'learning_rate/qwen_vl_interface': 6.980479625527449e-06, 'learning_rate/action_model': 6.980479625527448e-05, 'epoch': 1.0}) +[2026-06-17 21:52:19,564][starVLA.training.train_starvla][INFO] - Step 2732, Loss: {'train/loss': 0.6785500049591064, 'train/grad_norm_pre_clip': 21.300676345825195, 'timing/data': 0.03152504097670317, 'timing/model': 1.4587172772735357, 'learning_rate/qwen_vl_interface': 6.978465381897477e-06, 'learning_rate/action_model': 6.978465381897477e-05, 'epoch': 1.0}) +[2026-06-17 21:52:25,721][starVLA.training.train_starvla][INFO] - Step 2733, Loss: {'train/loss': 0.762044370174408, 'train/grad_norm_pre_clip': 17.64040184020996, 'timing/data': 0.035682535730302334, 'timing/model': 1.3456778666004539, 'learning_rate/qwen_vl_interface': 6.97645077995477e-06, 'learning_rate/action_model': 6.97645077995477e-05, 'epoch': 1.0}) +[2026-06-17 21:52:30,154][starVLA.training.train_starvla][INFO] - Step 2734, Loss: {'train/loss': 0.8177937269210815, 'train/grad_norm_pre_clip': 19.54365348815918, 'timing/data': 0.03009114135056734, 'timing/model': 1.0494998674839735, 'learning_rate/qwen_vl_interface': 6.97443582011696e-06, 'learning_rate/action_model': 6.974435820116959e-05, 'epoch': 1.0}) +[2026-06-17 21:52:34,806][starVLA.training.train_starvla][INFO] - Step 2735, Loss: {'train/loss': 0.6743097901344299, 'train/grad_norm_pre_clip': 15.375171661376953, 'timing/data': 0.06722180917859077, 'timing/model': 1.0375614939257503, 'learning_rate/qwen_vl_interface': 6.972420502801748e-06, 'learning_rate/action_model': 6.972420502801747e-05, 'epoch': 1.0}) +[2026-06-17 21:52:40,480][starVLA.training.train_starvla][INFO] - Step 2736, Loss: {'train/loss': 0.3001753091812134, 'train/grad_norm_pre_clip': 12.297032356262207, 'timing/data': 0.0315617760643363, 'timing/model': 1.0782429706305265, 'learning_rate/qwen_vl_interface': 6.970404828426911e-06, 'learning_rate/action_model': 6.97040482842691e-05, 'epoch': 1.0}) +[2026-06-17 21:52:45,205][starVLA.training.train_starvla][INFO] - Step 2737, Loss: {'train/loss': 0.8303126096725464, 'train/grad_norm_pre_clip': 18.206859588623047, 'timing/data': 0.03244903311133385, 'timing/model': 1.435961646027863, 'learning_rate/qwen_vl_interface': 6.9683887974103044e-06, 'learning_rate/action_model': 6.968388797410305e-05, 'epoch': 1.0}) +[2026-06-17 21:52:49,453][starVLA.training.train_starvla][INFO] - Step 2738, Loss: {'train/loss': 0.5288435816764832, 'train/grad_norm_pre_clip': 11.710087776184082, 'timing/data': 0.03606446925550699, 'timing/model': 1.0529612014070153, 'learning_rate/qwen_vl_interface': 6.966372410169848e-06, 'learning_rate/action_model': 6.966372410169848e-05, 'epoch': 1.0}) +[2026-06-17 21:52:55,988][starVLA.training.train_starvla][INFO] - Step 2739, Loss: {'train/loss': 0.6663784384727478, 'train/grad_norm_pre_clip': 16.47821807861328, 'timing/data': 0.09403201937675476, 'timing/model': 2.3562262980267406, 'learning_rate/qwen_vl_interface': 6.964355667123548e-06, 'learning_rate/action_model': 6.964355667123547e-05, 'epoch': 1.0}) +[2026-06-17 21:53:00,456][starVLA.training.train_starvla][INFO] - Step 2740, Loss: {'train/loss': 0.5802486538887024, 'train/grad_norm_pre_clip': 16.561138153076172, 'timing/data': 0.000988071784377098, 'timing/model': 1.0241060722619295, 'learning_rate/qwen_vl_interface': 6.96233856868947e-06, 'learning_rate/action_model': 6.962338568689469e-05, 'epoch': 1.0}) +[2026-06-17 21:53:05,039][starVLA.training.train_starvla][INFO] - Step 2741, Loss: {'train/loss': 0.6628992557525635, 'train/grad_norm_pre_clip': 17.064699172973633, 'timing/data': 0.0010561281815171242, 'timing/model': 1.1452092239633203, 'learning_rate/qwen_vl_interface': 6.960321115285767e-06, 'learning_rate/action_model': 6.960321115285767e-05, 'epoch': 1.0}) +[2026-06-17 21:53:17,239][starVLA.training.train_starvla][INFO] - Step 2742, Loss: {'train/loss': 0.4052315056324005, 'train/grad_norm_pre_clip': 13.851786613464355, 'timing/data': 0.0014785295352339745, 'timing/model': 5.323092211037874, 'learning_rate/qwen_vl_interface': 6.958303307330655e-06, 'learning_rate/action_model': 6.958303307330655e-05, 'epoch': 1.0}) +[2026-06-17 21:53:25,867][starVLA.training.train_starvla][INFO] - Step 2743, Loss: {'train/loss': 0.5380624532699585, 'train/grad_norm_pre_clip': 13.473482131958008, 'timing/data': 0.0008559795096516609, 'timing/model': 1.185803446918726, 'learning_rate/qwen_vl_interface': 6.956285145242429e-06, 'learning_rate/action_model': 6.956285145242428e-05, 'epoch': 1.0}) +[2026-06-17 21:53:30,581][starVLA.training.train_starvla][INFO] - Step 2744, Loss: {'train/loss': 0.556522786617279, 'train/grad_norm_pre_clip': 15.651129722595215, 'timing/data': 0.0011065565049648285, 'timing/model': 1.1761179463937879, 'learning_rate/qwen_vl_interface': 6.954266629439456e-06, 'learning_rate/action_model': 6.954266629439457e-05, 'epoch': 1.0}) +[2026-06-17 21:53:39,917][starVLA.training.train_starvla][INFO] - Step 2745, Loss: {'train/loss': 0.2804670035839081, 'train/grad_norm_pre_clip': 10.005321502685547, 'timing/data': 0.0012035863474011421, 'timing/model': 5.510649035684764, 'learning_rate/qwen_vl_interface': 6.952247760340176e-06, 'learning_rate/action_model': 6.952247760340176e-05, 'epoch': 1.01}) +[2026-06-17 21:53:46,669][starVLA.training.train_starvla][INFO] - Step 2746, Loss: {'train/loss': 0.4146806299686432, 'train/grad_norm_pre_clip': 12.042667388916016, 'timing/data': 0.0011254390701651573, 'timing/model': 1.213853127323091, 'learning_rate/qwen_vl_interface': 6.950228538363108e-06, 'learning_rate/action_model': 6.950228538363107e-05, 'epoch': 1.01}) +[2026-06-17 21:53:51,284][starVLA.training.train_starvla][INFO] - Step 2747, Loss: {'train/loss': 0.8728424906730652, 'train/grad_norm_pre_clip': 22.06655502319336, 'timing/data': 0.0009645335376262665, 'timing/model': 1.2114203851670027, 'learning_rate/qwen_vl_interface': 6.948208963926829e-06, 'learning_rate/action_model': 6.94820896392683e-05, 'epoch': 1.01}) +[2026-06-17 21:53:56,113][starVLA.training.train_starvla][INFO] - Step 2748, Loss: {'train/loss': 0.6384519338607788, 'train/grad_norm_pre_clip': 14.461377143859863, 'timing/data': 0.0009702639654278755, 'timing/model': 1.1570478193461895, 'learning_rate/qwen_vl_interface': 6.946189037450007e-06, 'learning_rate/action_model': 6.946189037450007e-05, 'epoch': 1.01}) +[2026-06-17 21:54:01,619][starVLA.training.train_starvla][INFO] - Step 2749, Loss: {'train/loss': 0.8509435653686523, 'train/grad_norm_pre_clip': 21.80186653137207, 'timing/data': 0.0009331405162811279, 'timing/model': 1.2329513495787978, 'learning_rate/qwen_vl_interface': 6.9441687593513715e-06, 'learning_rate/action_model': 6.944168759351372e-05, 'epoch': 1.01}) +[2026-06-17 21:54:06,878][starVLA.training.train_starvla][INFO] - Step 2750, Loss: {'train/loss': 0.6418091654777527, 'train/grad_norm_pre_clip': 17.479942321777344, 'timing/data': 0.0009554494172334671, 'timing/model': 1.6182146864011884, 'learning_rate/qwen_vl_interface': 6.942148130049732e-06, 'learning_rate/action_model': 6.942148130049731e-05, 'epoch': 1.01}) +[2026-06-17 21:54:11,678][starVLA.training.train_starvla][INFO] - Step 2751, Loss: {'train/loss': 0.6311442852020264, 'train/grad_norm_pre_clip': 18.22473907470703, 'timing/data': 0.0010598506778478622, 'timing/model': 1.0291021764278412, 'learning_rate/qwen_vl_interface': 6.940127149963961e-06, 'learning_rate/action_model': 6.94012714996396e-05, 'epoch': 1.01}) +[2026-06-17 21:54:16,313][starVLA.training.train_starvla][INFO] - Step 2752, Loss: {'train/loss': 0.9475975632667542, 'train/grad_norm_pre_clip': 19.44857406616211, 'timing/data': 0.0008815377950668335, 'timing/model': 1.3956485744565725, 'learning_rate/qwen_vl_interface': 6.9381058195130146e-06, 'learning_rate/action_model': 6.938105819513014e-05, 'epoch': 1.01}) +[2026-06-17 21:54:20,507][starVLA.training.train_starvla][INFO] - Step 2753, Loss: {'train/loss': 0.7628475427627563, 'train/grad_norm_pre_clip': 22.903533935546875, 'timing/data': 0.0010266024619340897, 'timing/model': 1.0618731146678329, 'learning_rate/qwen_vl_interface': 6.936084139115915e-06, 'learning_rate/action_model': 6.936084139115915e-05, 'epoch': 1.01}) +[2026-06-17 21:54:33,723][starVLA.training.train_starvla][INFO] - Step 2754, Loss: {'train/loss': 0.5886655449867249, 'train/grad_norm_pre_clip': 15.629141807556152, 'timing/data': 0.00087688397616148, 'timing/model': 1.1393736628815532, 'learning_rate/qwen_vl_interface': 6.934062109191759e-06, 'learning_rate/action_model': 6.934062109191759e-05, 'epoch': 1.01}) +[2026-06-17 21:54:38,160][starVLA.training.train_starvla][INFO] - Step 2755, Loss: {'train/loss': 0.6625742316246033, 'train/grad_norm_pre_clip': 17.706439971923828, 'timing/data': 0.03446758911013603, 'timing/model': 1.0554904285818338, 'learning_rate/qwen_vl_interface': 6.9320397301597155e-06, 'learning_rate/action_model': 6.932039730159715e-05, 'epoch': 1.01}) +[2026-06-17 21:54:44,104][starVLA.training.train_starvla][INFO] - Step 2756, Loss: {'train/loss': 0.9979097843170166, 'train/grad_norm_pre_clip': 19.23924446105957, 'timing/data': 0.03191171493381262, 'timing/model': 1.0330153247341514, 'learning_rate/qwen_vl_interface': 6.930017002439027e-06, 'learning_rate/action_model': 6.930017002439027e-05, 'epoch': 1.01}) +[2026-06-17 21:54:53,306][starVLA.training.train_starvla][INFO] - Step 2757, Loss: {'train/loss': 0.4766177833080292, 'train/grad_norm_pre_clip': 12.315011978149414, 'timing/data': 0.03243550565093756, 'timing/model': 1.5088232019916177, 'learning_rate/qwen_vl_interface': 6.927993926449005e-06, 'learning_rate/action_model': 6.927993926449005e-05, 'epoch': 1.01}) +[2026-06-17 21:54:57,363][starVLA.training.train_starvla][INFO] - Step 2758, Loss: {'train/loss': 0.8171367645263672, 'train/grad_norm_pre_clip': 17.679031372070312, 'timing/data': 0.03808485344052315, 'timing/model': 1.0334626361727715, 'learning_rate/qwen_vl_interface': 6.925970502609036e-06, 'learning_rate/action_model': 6.925970502609036e-05, 'epoch': 1.01}) +[2026-06-17 21:55:01,771][starVLA.training.train_starvla][INFO] - Step 2759, Loss: {'train/loss': 0.7136830687522888, 'train/grad_norm_pre_clip': 19.49648094177246, 'timing/data': 0.030440185219049454, 'timing/model': 1.0228656325489283, 'learning_rate/qwen_vl_interface': 6.92394673133858e-06, 'learning_rate/action_model': 6.92394673133858e-05, 'epoch': 1.01}) +[2026-06-17 21:55:06,151][starVLA.training.train_starvla][INFO] - Step 2760, Loss: {'train/loss': 0.6664132475852966, 'train/grad_norm_pre_clip': 36.13313293457031, 'timing/data': 0.0342209804803133, 'timing/model': 1.0378585495054722, 'learning_rate/qwen_vl_interface': 6.921922613057165e-06, 'learning_rate/action_model': 6.921922613057164e-05, 'epoch': 1.01}) +[2026-06-17 21:55:14,071][starVLA.training.train_starvla][INFO] - Step 2761, Loss: {'train/loss': 0.6785827875137329, 'train/grad_norm_pre_clip': 16.38535499572754, 'timing/data': 0.035899097099900246, 'timing/model': 1.2067104633897543, 'learning_rate/qwen_vl_interface': 6.9198981481843915e-06, 'learning_rate/action_model': 6.919898148184392e-05, 'epoch': 1.01}) +[2026-06-17 21:55:18,605][starVLA.training.train_starvla][INFO] - Step 2762, Loss: {'train/loss': 0.8246461749076843, 'train/grad_norm_pre_clip': 16.349645614624023, 'timing/data': 0.03371543809771538, 'timing/model': 1.5433719828724861, 'learning_rate/qwen_vl_interface': 6.917873337139936e-06, 'learning_rate/action_model': 6.917873337139937e-05, 'epoch': 1.01}) +[2026-06-17 21:55:24,055][starVLA.training.train_starvla][INFO] - Step 2763, Loss: {'train/loss': 0.5430944561958313, 'train/grad_norm_pre_clip': 15.299420356750488, 'timing/data': 0.033902998082339764, 'timing/model': 1.212136685848236, 'learning_rate/qwen_vl_interface': 6.915848180343545e-06, 'learning_rate/action_model': 6.915848180343545e-05, 'epoch': 1.01}) +[2026-06-17 21:55:29,616][starVLA.training.train_starvla][INFO] - Step 2764, Loss: {'train/loss': 0.5762099623680115, 'train/grad_norm_pre_clip': 14.478001594543457, 'timing/data': 0.04134074039757252, 'timing/model': 1.4823400005698204, 'learning_rate/qwen_vl_interface': 6.913822678215031e-06, 'learning_rate/action_model': 6.913822678215031e-05, 'epoch': 1.01}) +[2026-06-17 21:55:35,839][starVLA.training.train_starvla][INFO] - Step 2765, Loss: {'train/loss': 0.8007575273513794, 'train/grad_norm_pre_clip': 20.308570861816406, 'timing/data': 0.032204038463532925, 'timing/model': 1.086276637390256, 'learning_rate/qwen_vl_interface': 6.911796831174287e-06, 'learning_rate/action_model': 6.911796831174286e-05, 'epoch': 1.01}) +[2026-06-17 21:55:43,649][starVLA.training.train_starvla][INFO] - Step 2766, Loss: {'train/loss': 0.6246484518051147, 'train/grad_norm_pre_clip': 12.672070503234863, 'timing/data': 0.07993255089968443, 'timing/model': 3.9415629524737597, 'learning_rate/qwen_vl_interface': 6.909770639641271e-06, 'learning_rate/action_model': 6.909770639641271e-05, 'epoch': 1.01}) +[2026-06-17 21:55:48,903][starVLA.training.train_starvla][INFO] - Step 2767, Loss: {'train/loss': 0.45767930150032043, 'train/grad_norm_pre_clip': 11.52156925201416, 'timing/data': 0.09010123927146196, 'timing/model': 2.07332571875304, 'learning_rate/qwen_vl_interface': 6.907744104036015e-06, 'learning_rate/action_model': 6.907744104036015e-05, 'epoch': 1.01}) +[2026-06-17 21:55:55,232][starVLA.training.train_starvla][INFO] - Step 2768, Loss: {'train/loss': 0.6281266212463379, 'train/grad_norm_pre_clip': 15.25210189819336, 'timing/data': 0.036982277408242226, 'timing/model': 1.1610212503001094, 'learning_rate/qwen_vl_interface': 6.905717224778623e-06, 'learning_rate/action_model': 6.905717224778622e-05, 'epoch': 1.01}) +[2026-06-17 21:56:03,770][starVLA.training.train_starvla][INFO] - Step 2769, Loss: {'train/loss': 0.8340900540351868, 'train/grad_norm_pre_clip': 18.2839298248291, 'timing/data': 0.0342134665697813, 'timing/model': 1.2609986821189523, 'learning_rate/qwen_vl_interface': 6.903690002289268e-06, 'learning_rate/action_model': 6.903690002289267e-05, 'epoch': 1.01}) +[2026-06-17 21:56:13,102][starVLA.training.train_starvla][INFO] - Step 2770, Loss: {'train/loss': 0.66983962059021, 'train/grad_norm_pre_clip': 14.512726783752441, 'timing/data': 0.038885483518242836, 'timing/model': 2.317151096649468, 'learning_rate/qwen_vl_interface': 6.901662436988199e-06, 'learning_rate/action_model': 6.901662436988199e-05, 'epoch': 1.01}) +[2026-06-17 21:56:18,351][starVLA.training.train_starvla][INFO] - Step 2771, Loss: {'train/loss': 0.3493185043334961, 'train/grad_norm_pre_clip': 11.5387601852417, 'timing/data': 0.03987778164446354, 'timing/model': 1.6652009105309844, 'learning_rate/qwen_vl_interface': 6.899634529295726e-06, 'learning_rate/action_model': 6.899634529295726e-05, 'epoch': 1.01}) +[2026-06-17 21:56:22,758][starVLA.training.train_starvla][INFO] - Step 2772, Loss: {'train/loss': 0.9549753665924072, 'train/grad_norm_pre_clip': 19.434295654296875, 'timing/data': 0.032486872747540474, 'timing/model': 1.410040463320911, 'learning_rate/qwen_vl_interface': 6.897606279632241e-06, 'learning_rate/action_model': 6.897606279632241e-05, 'epoch': 1.02}) +[2026-06-17 21:56:27,077][starVLA.training.train_starvla][INFO] - Step 2773, Loss: {'train/loss': 0.8510292172431946, 'train/grad_norm_pre_clip': 18.414011001586914, 'timing/data': 0.029311825521290302, 'timing/model': 1.0254219705238938, 'learning_rate/qwen_vl_interface': 6.895577688418202e-06, 'learning_rate/action_model': 6.895577688418202e-05, 'epoch': 1.02}) +[2026-06-17 21:56:31,770][starVLA.training.train_starvla][INFO] - Step 2774, Loss: {'train/loss': 0.8038691282272339, 'train/grad_norm_pre_clip': 16.0488338470459, 'timing/data': 0.03349755238741636, 'timing/model': 1.0454671420156956, 'learning_rate/qwen_vl_interface': 6.893548756074136e-06, 'learning_rate/action_model': 6.893548756074136e-05, 'epoch': 1.02}) +[2026-06-17 21:56:36,261][starVLA.training.train_starvla][INFO] - Step 2775, Loss: {'train/loss': 0.5903215408325195, 'train/grad_norm_pre_clip': 19.80867576599121, 'timing/data': 0.03249749913811684, 'timing/model': 1.0439204452559352, 'learning_rate/qwen_vl_interface': 6.891519483020647e-06, 'learning_rate/action_model': 6.891519483020647e-05, 'epoch': 1.02}) +[2026-06-17 21:56:40,688][starVLA.training.train_starvla][INFO] - Step 2776, Loss: {'train/loss': 0.7273125648498535, 'train/grad_norm_pre_clip': 14.02334213256836, 'timing/data': 0.029971612617373466, 'timing/model': 1.0437005693092942, 'learning_rate/qwen_vl_interface': 6.8894898696784e-06, 'learning_rate/action_model': 6.889489869678399e-05, 'epoch': 1.02}) +[2026-06-17 21:56:45,042][starVLA.training.train_starvla][INFO] - Step 2777, Loss: {'train/loss': 0.7613067030906677, 'train/grad_norm_pre_clip': 17.934898376464844, 'timing/data': 0.03119009919464588, 'timing/model': 1.3168853931128979, 'learning_rate/qwen_vl_interface': 6.887459916468141e-06, 'learning_rate/action_model': 6.88745991646814e-05, 'epoch': 1.02}) +[2026-06-17 21:56:49,137][starVLA.training.train_starvla][INFO] - Step 2778, Loss: {'train/loss': 0.5977703332901001, 'train/grad_norm_pre_clip': 18.199752807617188, 'timing/data': 0.03244986571371555, 'timing/model': 1.0194564145058393, 'learning_rate/qwen_vl_interface': 6.885429623810678e-06, 'learning_rate/action_model': 6.885429623810677e-05, 'epoch': 1.02}) +[2026-06-17 21:56:53,620][starVLA.training.train_starvla][INFO] - Step 2779, Loss: {'train/loss': 0.4196006953716278, 'train/grad_norm_pre_clip': 9.495738983154297, 'timing/data': 0.03628836199641228, 'timing/model': 1.1064397133886814, 'learning_rate/qwen_vl_interface': 6.883398992126894e-06, 'learning_rate/action_model': 6.883398992126894e-05, 'epoch': 1.02}) +[2026-06-17 21:56:57,983][starVLA.training.train_starvla][INFO] - Step 2780, Loss: {'train/loss': 0.6417085528373718, 'train/grad_norm_pre_clip': 15.8351469039917, 'timing/data': 0.038790768943727016, 'timing/model': 1.0253175543621182, 'learning_rate/qwen_vl_interface': 6.8813680218377425e-06, 'learning_rate/action_model': 6.881368021837741e-05, 'epoch': 1.02}) +[2026-06-17 21:57:02,453][starVLA.training.train_starvla][INFO] - Step 2781, Loss: {'train/loss': 0.5465666651725769, 'train/grad_norm_pre_clip': 14.602438926696777, 'timing/data': 0.02992825396358967, 'timing/model': 1.0287320353090763, 'learning_rate/qwen_vl_interface': 6.8793367133642445e-06, 'learning_rate/action_model': 6.879336713364243e-05, 'epoch': 1.02}) +[2026-06-17 21:57:06,839][starVLA.training.train_starvla][INFO] - Step 2782, Loss: {'train/loss': 0.5295866131782532, 'train/grad_norm_pre_clip': 20.577056884765625, 'timing/data': 0.034233723767101765, 'timing/model': 1.3289811732247472, 'learning_rate/qwen_vl_interface': 6.877305067127492e-06, 'learning_rate/action_model': 6.877305067127492e-05, 'epoch': 1.02}) +[2026-06-17 21:57:11,051][starVLA.training.train_starvla][INFO] - Step 2783, Loss: {'train/loss': 0.9013776183128357, 'train/grad_norm_pre_clip': 25.950897216796875, 'timing/data': 0.03820385504513979, 'timing/model': 1.0469463309273124, 'learning_rate/qwen_vl_interface': 6.875273083548648e-06, 'learning_rate/action_model': 6.875273083548648e-05, 'epoch': 1.02}) +[2026-06-17 21:57:15,378][starVLA.training.train_starvla][INFO] - Step 2784, Loss: {'train/loss': 0.4803868234157562, 'train/grad_norm_pre_clip': 12.754701614379883, 'timing/data': 0.031219245865941048, 'timing/model': 1.0423066914081573, 'learning_rate/qwen_vl_interface': 6.873240763048945e-06, 'learning_rate/action_model': 6.873240763048945e-05, 'epoch': 1.02}) +[2026-06-17 21:57:20,012][starVLA.training.train_starvla][INFO] - Step 2785, Loss: {'train/loss': 0.5553451776504517, 'train/grad_norm_pre_clip': 11.63891315460205, 'timing/data': 0.03100731037557125, 'timing/model': 1.0682830568403006, 'learning_rate/qwen_vl_interface': 6.871208106049685e-06, 'learning_rate/action_model': 6.871208106049685e-05, 'epoch': 1.02}) +[2026-06-17 21:57:28,706][starVLA.training.train_starvla][INFO] - Step 2786, Loss: {'train/loss': 0.7171792387962341, 'train/grad_norm_pre_clip': 16.32120704650879, 'timing/data': 0.032783168368041515, 'timing/model': 5.392000123858452, 'learning_rate/qwen_vl_interface': 6.869175112972239e-06, 'learning_rate/action_model': 6.86917511297224e-05, 'epoch': 1.02}) +[2026-06-17 21:57:37,509][starVLA.training.train_starvla][INFO] - Step 2787, Loss: {'train/loss': 0.5786591172218323, 'train/grad_norm_pre_clip': 15.252358436584473, 'timing/data': 0.03547295089811087, 'timing/model': 1.6762926895171404, 'learning_rate/qwen_vl_interface': 6.867141784238052e-06, 'learning_rate/action_model': 6.867141784238051e-05, 'epoch': 1.02}) +[2026-06-17 21:57:46,489][starVLA.training.train_starvla][INFO] - Step 2788, Loss: {'train/loss': 0.7823193669319153, 'train/grad_norm_pre_clip': 13.819975852966309, 'timing/data': 0.03147755190730095, 'timing/model': 2.9007237013429403, 'learning_rate/qwen_vl_interface': 6.865108120268632e-06, 'learning_rate/action_model': 6.865108120268632e-05, 'epoch': 1.02}) +[2026-06-17 21:57:51,825][starVLA.training.train_starvla][INFO] - Step 2789, Loss: {'train/loss': 0.7023688554763794, 'train/grad_norm_pre_clip': 13.584270477294922, 'timing/data': 0.03157379012554884, 'timing/model': 1.0833849925547838, 'learning_rate/qwen_vl_interface': 6.863074121485561e-06, 'learning_rate/action_model': 6.86307412148556e-05, 'epoch': 1.02}) +[2026-06-17 21:58:03,751][starVLA.training.train_starvla][INFO] - Step 2790, Loss: {'train/loss': 0.5060545802116394, 'train/grad_norm_pre_clip': 11.654412269592285, 'timing/data': 0.041710732504725456, 'timing/model': 2.4230921268463135, 'learning_rate/qwen_vl_interface': 6.861039788310489e-06, 'learning_rate/action_model': 6.861039788310489e-05, 'epoch': 1.02}) +[2026-06-17 21:58:13,617][starVLA.training.train_starvla][INFO] - Step 2791, Loss: {'train/loss': 0.4586016833782196, 'train/grad_norm_pre_clip': 14.425274848937988, 'timing/data': 0.0431696567684412, 'timing/model': 3.580416343174875, 'learning_rate/qwen_vl_interface': 6.859005121165135e-06, 'learning_rate/action_model': 6.859005121165135e-05, 'epoch': 1.02}) +[2026-06-17 21:58:21,370][starVLA.training.train_starvla][INFO] - Step 2792, Loss: {'train/loss': 0.4993384778499603, 'train/grad_norm_pre_clip': 13.328383445739746, 'timing/data': 0.03604013007134199, 'timing/model': 1.4236703189089894, 'learning_rate/qwen_vl_interface': 6.856970120471289e-06, 'learning_rate/action_model': 6.856970120471289e-05, 'epoch': 1.02}) +[2026-06-17 21:58:31,710][starVLA.training.train_starvla][INFO] - Step 2793, Loss: {'train/loss': 0.8260895609855652, 'train/grad_norm_pre_clip': 15.29918384552002, 'timing/data': 0.02966408245265484, 'timing/model': 1.1384991323575377, 'learning_rate/qwen_vl_interface': 6.854934786650805e-06, 'learning_rate/action_model': 6.854934786650806e-05, 'epoch': 1.02}) +[2026-06-17 21:58:40,162][starVLA.training.train_starvla][INFO] - Step 2794, Loss: {'train/loss': 0.5856391787528992, 'train/grad_norm_pre_clip': 13.225744247436523, 'timing/data': 0.08576976787298918, 'timing/model': 2.8427946576848626, 'learning_rate/qwen_vl_interface': 6.852899120125615e-06, 'learning_rate/action_model': 6.852899120125616e-05, 'epoch': 1.02}) +[2026-06-17 21:58:50,566][starVLA.training.train_starvla][INFO] - Step 2795, Loss: {'train/loss': 0.8389942049980164, 'train/grad_norm_pre_clip': 17.19841957092285, 'timing/data': 0.03169535845518112, 'timing/model': 4.370465130545199, 'learning_rate/qwen_vl_interface': 6.850863121317712e-06, 'learning_rate/action_model': 6.850863121317712e-05, 'epoch': 1.02}) +[2026-06-17 21:58:55,319][starVLA.training.train_starvla][INFO] - Step 2796, Loss: {'train/loss': 0.5975393652915955, 'train/grad_norm_pre_clip': 11.828315734863281, 'timing/data': 0.030232718214392662, 'timing/model': 1.0654041450470686, 'learning_rate/qwen_vl_interface': 6.848826790649162e-06, 'learning_rate/action_model': 6.848826790649163e-05, 'epoch': 1.02}) +[2026-06-17 21:58:59,732][starVLA.training.train_starvla][INFO] - Step 2797, Loss: {'train/loss': 0.5754367113113403, 'train/grad_norm_pre_clip': 11.335013389587402, 'timing/data': 0.03675073757767677, 'timing/model': 1.3810099177062511, 'learning_rate/qwen_vl_interface': 6.846790128542098e-06, 'learning_rate/action_model': 6.846790128542098e-05, 'epoch': 1.02}) +[2026-06-17 21:59:03,854][starVLA.training.train_starvla][INFO] - Step 2798, Loss: {'train/loss': 0.7194114327430725, 'train/grad_norm_pre_clip': 13.954326629638672, 'timing/data': 0.03065238893032074, 'timing/model': 1.0908715091645718, 'learning_rate/qwen_vl_interface': 6.8447531354187214e-06, 'learning_rate/action_model': 6.844753135418722e-05, 'epoch': 1.02}) +[2026-06-17 21:59:08,585][starVLA.training.train_starvla][INFO] - Step 2799, Loss: {'train/loss': 0.6474918127059937, 'train/grad_norm_pre_clip': 14.037156105041504, 'timing/data': 0.03607056010514498, 'timing/model': 1.0952409524470568, 'learning_rate/qwen_vl_interface': 6.842715811701305e-06, 'learning_rate/action_model': 6.842715811701304e-05, 'epoch': 1.02}) +[2026-06-17 22:02:56,198][starVLA.training.train_starvla][INFO] - Step 2800, Loss: {'train/loss': 0.4540272057056427, 'train/grad_norm_pre_clip': 18.755266189575195, 'eval/loss': 0.6510800123214722, 'eval/latency_2/loss': 0.6510800123214722, 'eval/demon_fire_precision': 0.8964159117762899, 'eval/demon_fire_recall': 0.9523012552301255, 'eval/demon_fire_f1': 0.92351389734226, 'eval/demon_fire_rate_model': 0.7064552031163049, 'eval/demon_fire_rate_teacher': 0.6649972175848636, 'eval/demon_left_precision': 0.8250517598343685, 'eval/demon_left_recall': 0.8389473684210527, 'eval/demon_left_f1': 0.8319415448851775, 'eval/demon_left_rate_model': 0.2687813021702838, 'eval/demon_left_rate_teacher': 0.2643294379521425, 'eval/demon_right_precision': 0.8177826564215148, 'eval/demon_right_recall': 0.8418079096045198, 'eval/demon_right_f1': 0.8296213808463252, 'eval/demon_right_rate_model': 0.25347801892042293, 'eval/demon_right_rate_teacher': 0.24624373956594323, 'eval/demon_move_f1': 0.8307814628657513, 'eval/demon_cc_f1': 0.8935909158751404, 'eval/latency_2/demon_fire_precision': 0.8964159117762899, 'eval/latency_2/demon_fire_recall': 0.9523012552301255, 'eval/latency_2/demon_fire_f1': 0.92351389734226, 'eval/latency_2/demon_fire_rate_model': 0.7064552031163049, 'eval/latency_2/demon_fire_rate_teacher': 0.6649972175848636, 'eval/latency_2/demon_left_precision': 0.8250517598343685, 'eval/latency_2/demon_left_recall': 0.8389473684210527, 'eval/latency_2/demon_left_f1': 0.8319415448851775, 'eval/latency_2/demon_left_rate_model': 0.2687813021702838, 'eval/latency_2/demon_left_rate_teacher': 0.2643294379521425, 'eval/latency_2/demon_right_precision': 0.8177826564215148, 'eval/latency_2/demon_right_recall': 0.8418079096045198, 'eval/latency_2/demon_right_f1': 0.8296213808463252, 'eval/latency_2/demon_right_rate_model': 0.25347801892042293, 'eval/latency_2/demon_right_rate_teacher': 0.24624373956594323, 'eval/latency_2/demon_move_f1': 0.8307814628657513, 'eval/latency_2/demon_cc_f1': 0.8935909158751404, 'timing/data': 0.03252979088574648, 'timing/model': 1.097657361999154, 'learning_rate/qwen_vl_interface': 6.840678157812184e-06, 'learning_rate/action_model': 6.840678157812184e-05, 'epoch': 1.03}) +[2026-06-17 22:02:56,200][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_2800_state +[2026-06-17 22:02:56,207][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 22:03:04,822][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_2800_state/model.safetensors +[2026-06-17 22:03:19,234][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_2800_state/optimizer.bin +[2026-06-17 22:03:19,236][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_2800_state/sampler.bin +[2026-06-17 22:03:19,236][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_2800_state/sampler_1.bin +[2026-06-17 22:03:19,239][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_2800_state/random_states_0.pkl +[2026-06-17 22:03:19,244][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2800 +[2026-06-17 22:03:19,245][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 22:03:19,276][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 22:03:24,240][starVLA.training.train_starvla][INFO] - Step 2801, Loss: {'train/loss': 0.6000536680221558, 'train/grad_norm_pre_clip': 12.720261573791504, 'timing/data': 0.042478181421756744, 'timing/model': 1.1411452870815992, 'learning_rate/qwen_vl_interface': 6.8386401741737725e-06, 'learning_rate/action_model': 6.838640174173773e-05, 'epoch': 1.03}) +[2026-06-17 22:03:28,723][starVLA.training.train_starvla][INFO] - Step 2802, Loss: {'train/loss': 0.7434313893318176, 'train/grad_norm_pre_clip': 18.604135513305664, 'timing/data': 0.03341285511851311, 'timing/model': 1.0837646685540676, 'learning_rate/qwen_vl_interface': 6.836601861208542e-06, 'learning_rate/action_model': 6.836601861208541e-05, 'epoch': 1.03}) +[2026-06-17 22:03:33,169][starVLA.training.train_starvla][INFO] - Step 2803, Loss: {'train/loss': 0.6457605957984924, 'train/grad_norm_pre_clip': 14.902382850646973, 'timing/data': 0.031073695048689842, 'timing/model': 1.410400753840804, 'learning_rate/qwen_vl_interface': 6.834563219339039e-06, 'learning_rate/action_model': 6.834563219339038e-05, 'epoch': 1.03}) +[2026-06-17 22:03:37,683][starVLA.training.train_starvla][INFO] - Step 2804, Loss: {'train/loss': 0.8456117510795593, 'train/grad_norm_pre_clip': 15.288785934448242, 'timing/data': 0.03129652235656977, 'timing/model': 1.1275680139660835, 'learning_rate/qwen_vl_interface': 6.832524248987874e-06, 'learning_rate/action_model': 6.832524248987874e-05, 'epoch': 1.03}) +[2026-06-17 22:03:45,396][starVLA.training.train_starvla][INFO] - Step 2805, Loss: {'train/loss': 0.5693623423576355, 'train/grad_norm_pre_clip': 12.923699378967285, 'timing/data': 0.032739042304456234, 'timing/model': 1.0387392016127706, 'learning_rate/qwen_vl_interface': 6.8304849505777314e-06, 'learning_rate/action_model': 6.830484950577732e-05, 'epoch': 1.03}) +[2026-06-17 22:03:49,949][starVLA.training.train_starvla][INFO] - Step 2806, Loss: {'train/loss': 0.4909511208534241, 'train/grad_norm_pre_clip': 13.363181114196777, 'timing/data': 0.03314180765300989, 'timing/model': 1.0323416516184807, 'learning_rate/qwen_vl_interface': 6.828445324531357e-06, 'learning_rate/action_model': 6.828445324531357e-05, 'epoch': 1.03}) +[2026-06-17 22:03:54,446][starVLA.training.train_starvla][INFO] - Step 2807, Loss: {'train/loss': 0.5172527432441711, 'train/grad_norm_pre_clip': 11.870257377624512, 'timing/data': 0.0353584224358201, 'timing/model': 1.0868230760097504, 'learning_rate/qwen_vl_interface': 6.826405371271567e-06, 'learning_rate/action_model': 6.826405371271567e-05, 'epoch': 1.03}) +[2026-06-17 22:03:59,010][starVLA.training.train_starvla][INFO] - Step 2808, Loss: {'train/loss': 0.44458532333374023, 'train/grad_norm_pre_clip': 16.374664306640625, 'timing/data': 0.03488155733793974, 'timing/model': 1.4266156703233719, 'learning_rate/qwen_vl_interface': 6.82436509122125e-06, 'learning_rate/action_model': 6.82436509122125e-05, 'epoch': 1.03}) +[2026-06-17 22:04:06,414][starVLA.training.train_starvla][INFO] - Step 2809, Loss: {'train/loss': 0.5987475514411926, 'train/grad_norm_pre_clip': 11.253613471984863, 'timing/data': 0.03257262986153364, 'timing/model': 1.3315312638878822, 'learning_rate/qwen_vl_interface': 6.822324484803353e-06, 'learning_rate/action_model': 6.822324484803353e-05, 'epoch': 1.03}) +[2026-06-17 22:04:11,011][starVLA.training.train_starvla][INFO] - Step 2810, Loss: {'train/loss': 0.4150553345680237, 'train/grad_norm_pre_clip': 11.29725456237793, 'timing/data': 0.03577638417482376, 'timing/model': 1.0506253326311707, 'learning_rate/qwen_vl_interface': 6.8202835524409e-06, 'learning_rate/action_model': 6.8202835524409e-05, 'epoch': 1.03}) +[2026-06-17 22:04:15,734][starVLA.training.train_starvla][INFO] - Step 2811, Loss: {'train/loss': 0.7161263823509216, 'train/grad_norm_pre_clip': 15.887839317321777, 'timing/data': 0.034719571471214294, 'timing/model': 1.1213903427124023, 'learning_rate/qwen_vl_interface': 6.818242294556977e-06, 'learning_rate/action_model': 6.818242294556977e-05, 'epoch': 1.03}) +[2026-06-17 22:04:20,488][starVLA.training.train_starvla][INFO] - Step 2812, Loss: {'train/loss': 0.6378846764564514, 'train/grad_norm_pre_clip': 14.554821968078613, 'timing/data': 0.0342536810785532, 'timing/model': 1.1970623712986708, 'learning_rate/qwen_vl_interface': 6.816200711574738e-06, 'learning_rate/action_model': 6.816200711574738e-05, 'epoch': 1.03}) +[2026-06-17 22:04:25,383][starVLA.training.train_starvla][INFO] - Step 2813, Loss: {'train/loss': 0.5598571300506592, 'train/grad_norm_pre_clip': 15.5083646774292, 'timing/data': 0.03160449955612421, 'timing/model': 1.4835715796798468, 'learning_rate/qwen_vl_interface': 6.814158803917406e-06, 'learning_rate/action_model': 6.814158803917406e-05, 'epoch': 1.03}) +[2026-06-17 22:04:29,498][starVLA.training.train_starvla][INFO] - Step 2814, Loss: {'train/loss': 0.42945048213005066, 'train/grad_norm_pre_clip': 15.839674949645996, 'timing/data': 0.02910572476685047, 'timing/model': 1.1191239543259144, 'learning_rate/qwen_vl_interface': 6.812116572008271e-06, 'learning_rate/action_model': 6.812116572008271e-05, 'epoch': 1.03}) +[2026-06-17 22:04:34,006][starVLA.training.train_starvla][INFO] - Step 2815, Loss: {'train/loss': 0.6472196578979492, 'train/grad_norm_pre_clip': 16.823190689086914, 'timing/data': 0.04024420212954283, 'timing/model': 1.0560135170817375, 'learning_rate/qwen_vl_interface': 6.8100740162706906e-06, 'learning_rate/action_model': 6.81007401627069e-05, 'epoch': 1.03}) +[2026-06-17 22:04:40,073][starVLA.training.train_starvla][INFO] - Step 2816, Loss: {'train/loss': 0.65278160572052, 'train/grad_norm_pre_clip': 17.34186363220215, 'timing/data': 0.0357852503657341, 'timing/model': 1.1947960956022143, 'learning_rate/qwen_vl_interface': 6.808031137128087e-06, 'learning_rate/action_model': 6.808031137128086e-05, 'epoch': 1.03}) +[2026-06-17 22:04:44,566][starVLA.training.train_starvla][INFO] - Step 2817, Loss: {'train/loss': 0.8667513132095337, 'train/grad_norm_pre_clip': 15.229927062988281, 'timing/data': 0.030034220777451992, 'timing/model': 1.0841080928221345, 'learning_rate/qwen_vl_interface': 6.805987935003951e-06, 'learning_rate/action_model': 6.805987935003951e-05, 'epoch': 1.03}) +[2026-06-17 22:04:48,980][starVLA.training.train_starvla][INFO] - Step 2818, Loss: {'train/loss': 0.7543260455131531, 'train/grad_norm_pre_clip': 14.44738483428955, 'timing/data': 0.03439457435160875, 'timing/model': 1.026916935108602, 'learning_rate/qwen_vl_interface': 6.803944410321843e-06, 'learning_rate/action_model': 6.803944410321844e-05, 'epoch': 1.03}) +[2026-06-17 22:04:54,160][starVLA.training.train_starvla][INFO] - Step 2819, Loss: {'train/loss': 0.5589780807495117, 'train/grad_norm_pre_clip': 13.424569129943848, 'timing/data': 0.04191331472247839, 'timing/model': 1.5236741015687585, 'learning_rate/qwen_vl_interface': 6.801900563505388e-06, 'learning_rate/action_model': 6.801900563505388e-05, 'epoch': 1.03}) +[2026-06-17 22:04:58,299][starVLA.training.train_starvla][INFO] - Step 2820, Loss: {'train/loss': 0.608457624912262, 'train/grad_norm_pre_clip': 27.328506469726562, 'timing/data': 0.0286891870200634, 'timing/model': 1.0490376930683851, 'learning_rate/qwen_vl_interface': 6.7998563949782755e-06, 'learning_rate/action_model': 6.799856394978276e-05, 'epoch': 1.03}) +[2026-06-17 22:05:02,695][starVLA.training.train_starvla][INFO] - Step 2821, Loss: {'train/loss': 0.6694160103797913, 'train/grad_norm_pre_clip': 14.771803855895996, 'timing/data': 0.03126952797174454, 'timing/model': 1.0446333717554808, 'learning_rate/qwen_vl_interface': 6.797811905164263e-06, 'learning_rate/action_model': 6.797811905164263e-05, 'epoch': 1.03}) +[2026-06-17 22:05:07,047][starVLA.training.train_starvla][INFO] - Step 2822, Loss: {'train/loss': 0.46856141090393066, 'train/grad_norm_pre_clip': 10.709634780883789, 'timing/data': 0.03763832990080118, 'timing/model': 1.0451964046806097, 'learning_rate/qwen_vl_interface': 6.795767094487179e-06, 'learning_rate/action_model': 6.795767094487179e-05, 'epoch': 1.03}) +[2026-06-17 22:05:14,186][starVLA.training.train_starvla][INFO] - Step 2823, Loss: {'train/loss': 0.6034917831420898, 'train/grad_norm_pre_clip': 21.22842788696289, 'timing/data': 0.03268229868263006, 'timing/model': 1.0445282189175487, 'learning_rate/qwen_vl_interface': 6.793721963370913e-06, 'learning_rate/action_model': 6.793721963370913e-05, 'epoch': 1.03}) +[2026-06-17 22:05:18,713][starVLA.training.train_starvla][INFO] - Step 2824, Loss: {'train/loss': 0.6901339888572693, 'train/grad_norm_pre_clip': 13.262473106384277, 'timing/data': 0.035403584130108356, 'timing/model': 1.3731182646006346, 'learning_rate/qwen_vl_interface': 6.7916765122394215e-06, 'learning_rate/action_model': 6.791676512239421e-05, 'epoch': 1.03}) +[2026-06-17 22:05:22,967][starVLA.training.train_starvla][INFO] - Step 2825, Loss: {'train/loss': 0.6711404919624329, 'train/grad_norm_pre_clip': 15.222249031066895, 'timing/data': 0.03224895615130663, 'timing/model': 1.0674600526690483, 'learning_rate/qwen_vl_interface': 6.7896307415167305e-06, 'learning_rate/action_model': 6.78963074151673e-05, 'epoch': 1.03}) +[2026-06-17 22:05:27,573][starVLA.training.train_starvla][INFO] - Step 2826, Loss: {'train/loss': 0.6235702633857727, 'train/grad_norm_pre_clip': 12.407991409301758, 'timing/data': 0.03558202739804983, 'timing/model': 1.152099885046482, 'learning_rate/qwen_vl_interface': 6.787584651626929e-06, 'learning_rate/action_model': 6.78758465162693e-05, 'epoch': 1.03}) +[2026-06-17 22:05:33,887][starVLA.training.train_starvla][INFO] - Step 2827, Loss: {'train/loss': 0.5826888084411621, 'train/grad_norm_pre_clip': 14.853344917297363, 'timing/data': 0.031647467985749245, 'timing/model': 1.0769128650426865, 'learning_rate/qwen_vl_interface': 6.785538242994177e-06, 'learning_rate/action_model': 6.785538242994177e-05, 'epoch': 1.04}) +[2026-06-17 22:05:38,492][starVLA.training.train_starvla][INFO] - Step 2828, Loss: {'train/loss': 0.813402533531189, 'train/grad_norm_pre_clip': 16.84271812438965, 'timing/data': 0.0369719872251153, 'timing/model': 1.2040057657286525, 'learning_rate/qwen_vl_interface': 6.783491516042694e-06, 'learning_rate/action_model': 6.783491516042693e-05, 'epoch': 1.04}) +[2026-06-17 22:05:42,806][starVLA.training.train_starvla][INFO] - Step 2829, Loss: {'train/loss': 1.0746798515319824, 'train/grad_norm_pre_clip': 20.917682647705078, 'timing/data': 0.03210063278675079, 'timing/model': 1.322619435377419, 'learning_rate/qwen_vl_interface': 6.781444471196768e-06, 'learning_rate/action_model': 6.781444471196769e-05, 'epoch': 1.04}) +[2026-06-17 22:05:47,097][starVLA.training.train_starvla][INFO] - Step 2830, Loss: {'train/loss': 0.7743804454803467, 'train/grad_norm_pre_clip': 15.95676040649414, 'timing/data': 0.029331482015550137, 'timing/model': 1.118437954224646, 'learning_rate/qwen_vl_interface': 6.7793971088807575e-06, 'learning_rate/action_model': 6.779397108880757e-05, 'epoch': 1.04}) +[2026-06-17 22:05:52,311][starVLA.training.train_starvla][INFO] - Step 2831, Loss: {'train/loss': 0.5222882032394409, 'train/grad_norm_pre_clip': 16.0153865814209, 'timing/data': 0.04559687711298466, 'timing/model': 1.1542012114077806, 'learning_rate/qwen_vl_interface': 6.77734942951908e-06, 'learning_rate/action_model': 6.77734942951908e-05, 'epoch': 1.04}) +[2026-06-17 22:05:57,418][starVLA.training.train_starvla][INFO] - Step 2832, Loss: {'train/loss': 0.8262295126914978, 'train/grad_norm_pre_clip': 21.492738723754883, 'timing/data': 0.028057859279215336, 'timing/model': 1.062521512620151, 'learning_rate/qwen_vl_interface': 6.7753014335362224e-06, 'learning_rate/action_model': 6.775301433536222e-05, 'epoch': 1.04}) +[2026-06-17 22:06:02,002][starVLA.training.train_starvla][INFO] - Step 2833, Loss: {'train/loss': 0.30183669924736023, 'train/grad_norm_pre_clip': 9.0983304977417, 'timing/data': 0.03683878295123577, 'timing/model': 1.1229730555787683, 'learning_rate/qwen_vl_interface': 6.773253121356735e-06, 'learning_rate/action_model': 6.773253121356735e-05, 'epoch': 1.04}) +[2026-06-17 22:06:06,456][starVLA.training.train_starvla][INFO] - Step 2834, Loss: {'train/loss': 0.909885585308075, 'train/grad_norm_pre_clip': 14.929376602172852, 'timing/data': 0.030353745445609093, 'timing/model': 1.4370112316682935, 'learning_rate/qwen_vl_interface': 6.771204493405239e-06, 'learning_rate/action_model': 6.771204493405239e-05, 'epoch': 1.04}) +[2026-06-17 22:06:10,981][starVLA.training.train_starvla][INFO] - Step 2835, Loss: {'train/loss': 0.3753887712955475, 'train/grad_norm_pre_clip': 10.125805854797363, 'timing/data': 0.03254873584955931, 'timing/model': 1.03395289555192, 'learning_rate/qwen_vl_interface': 6.769155550106413e-06, 'learning_rate/action_model': 6.769155550106413e-05, 'epoch': 1.04}) +[2026-06-17 22:06:15,485][starVLA.training.train_starvla][INFO] - Step 2836, Loss: {'train/loss': 0.5940097570419312, 'train/grad_norm_pre_clip': 13.845205307006836, 'timing/data': 0.03164474479854107, 'timing/model': 1.0380144072696567, 'learning_rate/qwen_vl_interface': 6.767106291885006e-06, 'learning_rate/action_model': 6.767106291885006e-05, 'epoch': 1.04}) +[2026-06-17 22:06:20,252][starVLA.training.train_starvla][INFO] - Step 2837, Loss: {'train/loss': 0.5363712310791016, 'train/grad_norm_pre_clip': 13.253110885620117, 'timing/data': 0.03154600132256746, 'timing/model': 1.1283728992566466, 'learning_rate/qwen_vl_interface': 6.765056719165833e-06, 'learning_rate/action_model': 6.765056719165834e-05, 'epoch': 1.04}) +[2026-06-17 22:06:24,723][starVLA.training.train_starvla][INFO] - Step 2838, Loss: {'train/loss': 0.7744945287704468, 'train/grad_norm_pre_clip': 19.9788875579834, 'timing/data': 0.030334405601024628, 'timing/model': 1.377896286547184, 'learning_rate/qwen_vl_interface': 6.763006832373772e-06, 'learning_rate/action_model': 6.763006832373771e-05, 'epoch': 1.04}) +[2026-06-17 22:06:29,433][starVLA.training.train_starvla][INFO] - Step 2839, Loss: {'train/loss': 0.6262075304985046, 'train/grad_norm_pre_clip': 17.154102325439453, 'timing/data': 0.029420793056488037, 'timing/model': 1.4490486541762948, 'learning_rate/qwen_vl_interface': 6.760956631933766e-06, 'learning_rate/action_model': 6.760956631933765e-05, 'epoch': 1.04}) +[2026-06-17 22:06:33,899][starVLA.training.train_starvla][INFO] - Step 2840, Loss: {'train/loss': 1.072011947631836, 'train/grad_norm_pre_clip': 17.46825408935547, 'timing/data': 0.03463330678641796, 'timing/model': 1.06119847856462, 'learning_rate/qwen_vl_interface': 6.758906118270826e-06, 'learning_rate/action_model': 6.758906118270825e-05, 'epoch': 1.04}) +[2026-06-17 22:06:39,965][starVLA.training.train_starvla][INFO] - Step 2841, Loss: {'train/loss': 0.5532825589179993, 'train/grad_norm_pre_clip': 15.279997825622559, 'timing/data': 0.03270329814404249, 'timing/model': 1.0336303859949112, 'learning_rate/qwen_vl_interface': 6.756855291810024e-06, 'learning_rate/action_model': 6.756855291810024e-05, 'epoch': 1.04}) +[2026-06-17 22:06:44,658][starVLA.training.train_starvla][INFO] - Step 2842, Loss: {'train/loss': 0.6790420413017273, 'train/grad_norm_pre_clip': 15.548295021057129, 'timing/data': 0.03044019266963005, 'timing/model': 1.0217513125389814, 'learning_rate/qwen_vl_interface': 6.754804152976495e-06, 'learning_rate/action_model': 6.754804152976495e-05, 'epoch': 1.04}) +[2026-06-17 22:06:49,074][starVLA.training.train_starvla][INFO] - Step 2843, Loss: {'train/loss': 0.6714539527893066, 'train/grad_norm_pre_clip': 16.079620361328125, 'timing/data': 0.0284185241907835, 'timing/model': 1.0712436409667134, 'learning_rate/qwen_vl_interface': 6.752752702195448e-06, 'learning_rate/action_model': 6.752752702195448e-05, 'epoch': 1.04}) +[2026-06-17 22:06:54,865][starVLA.training.train_starvla][INFO] - Step 2844, Loss: {'train/loss': 0.6300683617591858, 'train/grad_norm_pre_clip': 14.594889640808105, 'timing/data': 0.029489311389625072, 'timing/model': 1.217449433170259, 'learning_rate/qwen_vl_interface': 6.750700939892151e-06, 'learning_rate/action_model': 6.75070093989215e-05, 'epoch': 1.04}) +[2026-06-17 22:06:58,736][starVLA.training.train_starvla][INFO] - Step 2845, Loss: {'train/loss': 0.7776740789413452, 'train/grad_norm_pre_clip': 16.000959396362305, 'timing/data': 0.027189123444259167, 'timing/model': 1.0130370864644647, 'learning_rate/qwen_vl_interface': 6.748648866491931e-06, 'learning_rate/action_model': 6.748648866491932e-05, 'epoch': 1.04}) +[2026-06-17 22:07:02,840][starVLA.training.train_starvla][INFO] - Step 2846, Loss: {'train/loss': 0.5368287563323975, 'train/grad_norm_pre_clip': 14.08301067352295, 'timing/data': 0.029432300478219986, 'timing/model': 1.0085803447291255, 'learning_rate/qwen_vl_interface': 6.7465964824201915e-06, 'learning_rate/action_model': 6.746596482420192e-05, 'epoch': 1.04}) +[2026-06-17 22:07:07,013][starVLA.training.train_starvla][INFO] - Step 2847, Loss: {'train/loss': 0.8553301692008972, 'train/grad_norm_pre_clip': 17.214046478271484, 'timing/data': 0.025028754025697708, 'timing/model': 1.0099469032138586, 'learning_rate/qwen_vl_interface': 6.744543788102388e-06, 'learning_rate/action_model': 6.744543788102388e-05, 'epoch': 1.04}) +[2026-06-17 22:07:11,114][starVLA.training.train_starvla][INFO] - Step 2848, Loss: {'train/loss': 0.5639788508415222, 'train/grad_norm_pre_clip': 17.21388816833496, 'timing/data': 0.030176297761499882, 'timing/model': 0.9976877942681313, 'learning_rate/qwen_vl_interface': 6.742490783964051e-06, 'learning_rate/action_model': 6.742490783964051e-05, 'epoch': 1.04}) +[2026-06-17 22:07:15,264][starVLA.training.train_starvla][INFO] - Step 2849, Loss: {'train/loss': 0.5331600308418274, 'train/grad_norm_pre_clip': 12.188721656799316, 'timing/data': 0.034183830954134464, 'timing/model': 1.2677699644118547, 'learning_rate/qwen_vl_interface': 6.740437470430766e-06, 'learning_rate/action_model': 6.740437470430766e-05, 'epoch': 1.04}) +[2026-06-17 22:07:19,138][starVLA.training.train_starvla][INFO] - Step 2850, Loss: {'train/loss': 0.39568984508514404, 'train/grad_norm_pre_clip': 13.570253372192383, 'timing/data': 0.02816107962280512, 'timing/model': 1.0009632119908929, 'learning_rate/qwen_vl_interface': 6.7383838479281915e-06, 'learning_rate/action_model': 6.738383847928191e-05, 'epoch': 1.04}) +[2026-06-17 22:07:23,227][starVLA.training.train_starvla][INFO] - Step 2851, Loss: {'train/loss': 0.6004431843757629, 'train/grad_norm_pre_clip': 17.401731491088867, 'timing/data': 0.025794828310608864, 'timing/model': 0.9971966752782464, 'learning_rate/qwen_vl_interface': 6.736329916882044e-06, 'learning_rate/action_model': 6.736329916882044e-05, 'epoch': 1.04}) +[2026-06-17 22:07:27,396][starVLA.training.train_starvla][INFO] - Step 2852, Loss: {'train/loss': 0.40154656767845154, 'train/grad_norm_pre_clip': 14.591618537902832, 'timing/data': 0.027378244325518608, 'timing/model': 1.0175621854141355, 'learning_rate/qwen_vl_interface': 6.734275677718105e-06, 'learning_rate/action_model': 6.734275677718105e-05, 'epoch': 1.04}) +[2026-06-17 22:07:31,506][starVLA.training.train_starvla][INFO] - Step 2853, Loss: {'train/loss': 0.6864994764328003, 'train/grad_norm_pre_clip': 14.92033863067627, 'timing/data': 0.025345115922391415, 'timing/model': 0.9996535293757915, 'learning_rate/qwen_vl_interface': 6.73222113086222e-06, 'learning_rate/action_model': 6.73222113086222e-05, 'epoch': 1.04}) +[2026-06-17 22:07:35,550][starVLA.training.train_starvla][INFO] - Step 2854, Loss: {'train/loss': 0.5292589664459229, 'train/grad_norm_pre_clip': 13.057897567749023, 'timing/data': 0.02309866063296795, 'timing/model': 1.1994887487962842, 'learning_rate/qwen_vl_interface': 6.7301662767403e-06, 'learning_rate/action_model': 6.7301662767403e-05, 'epoch': 1.05}) +[2026-06-17 22:07:39,646][starVLA.training.train_starvla][INFO] - Step 2855, Loss: {'train/loss': 0.3942255973815918, 'train/grad_norm_pre_clip': 12.442093849182129, 'timing/data': 0.0317060649394989, 'timing/model': 1.1281839394941926, 'learning_rate/qwen_vl_interface': 6.728111115778319e-06, 'learning_rate/action_model': 6.728111115778318e-05, 'epoch': 1.05}) +[2026-06-17 22:07:45,838][starVLA.training.train_starvla][INFO] - Step 2856, Loss: {'train/loss': 0.6933227777481079, 'train/grad_norm_pre_clip': 19.321067810058594, 'timing/data': 0.04065278451889753, 'timing/model': 1.3196765501052141, 'learning_rate/qwen_vl_interface': 6.7260556484023135e-06, 'learning_rate/action_model': 6.726055648402313e-05, 'epoch': 1.05}) +[2026-06-17 22:07:50,569][starVLA.training.train_starvla][INFO] - Step 2857, Loss: {'train/loss': 0.7361480593681335, 'train/grad_norm_pre_clip': 16.3469181060791, 'timing/data': 0.02938634715974331, 'timing/model': 1.1205638088285923, 'learning_rate/qwen_vl_interface': 6.723999875038381e-06, 'learning_rate/action_model': 6.72399987503838e-05, 'epoch': 1.05}) +[2026-06-17 22:07:55,322][starVLA.training.train_starvla][INFO] - Step 2858, Loss: {'train/loss': 0.43058982491493225, 'train/grad_norm_pre_clip': 13.390717506408691, 'timing/data': 0.03072839230298996, 'timing/model': 1.2465792391449213, 'learning_rate/qwen_vl_interface': 6.721943796112691e-06, 'learning_rate/action_model': 6.721943796112691e-05, 'epoch': 1.05}) +[2026-06-17 22:07:59,803][starVLA.training.train_starvla][INFO] - Step 2859, Loss: {'train/loss': 0.3456268608570099, 'train/grad_norm_pre_clip': 13.405277252197266, 'timing/data': 0.028307202272117138, 'timing/model': 1.4339610626921058, 'learning_rate/qwen_vl_interface': 6.719887412051468e-06, 'learning_rate/action_model': 6.719887412051468e-05, 'epoch': 1.05}) +[2026-06-17 22:08:13,108][starVLA.training.train_starvla][INFO] - Step 2860, Loss: {'train/loss': 0.6877318620681763, 'train/grad_norm_pre_clip': 19.38311195373535, 'timing/data': 0.03507177159190178, 'timing/model': 3.5215380927547812, 'learning_rate/qwen_vl_interface': 6.7178307232810015e-06, 'learning_rate/action_model': 6.717830723281e-05, 'epoch': 1.05}) +[2026-06-17 22:08:17,701][starVLA.training.train_starvla][INFO] - Step 2861, Loss: {'train/loss': 0.7735649943351746, 'train/grad_norm_pre_clip': 18.39731216430664, 'timing/data': 0.03563880082219839, 'timing/model': 1.111010080203414, 'learning_rate/qwen_vl_interface': 6.715773730227646e-06, 'learning_rate/action_model': 6.715773730227646e-05, 'epoch': 1.05}) +[2026-06-17 22:08:22,269][starVLA.training.train_starvla][INFO] - Step 2862, Loss: {'train/loss': 0.6697779893875122, 'train/grad_norm_pre_clip': 16.26483154296875, 'timing/data': 0.030241141095757484, 'timing/model': 1.0742443641647696, 'learning_rate/qwen_vl_interface': 6.7137164333178216e-06, 'learning_rate/action_model': 6.713716433317821e-05, 'epoch': 1.05}) +[2026-06-17 22:08:27,053][starVLA.training.train_starvla][INFO] - Step 2863, Loss: {'train/loss': 0.5246126651763916, 'train/grad_norm_pre_clip': 14.447308540344238, 'timing/data': 0.03796951565891504, 'timing/model': 1.2802295964211226, 'learning_rate/qwen_vl_interface': 6.711658832978004e-06, 'learning_rate/action_model': 6.711658832978003e-05, 'epoch': 1.05}) +[2026-06-17 22:08:31,712][starVLA.training.train_starvla][INFO] - Step 2864, Loss: {'train/loss': 0.9667697548866272, 'train/grad_norm_pre_clip': 20.62165069580078, 'timing/data': 0.0293243620544672, 'timing/model': 1.340702885761857, 'learning_rate/qwen_vl_interface': 6.709600929634737e-06, 'learning_rate/action_model': 6.709600929634738e-05, 'epoch': 1.05}) +[2026-06-17 22:08:36,016][starVLA.training.train_starvla][INFO] - Step 2865, Loss: {'train/loss': 0.6894096732139587, 'train/grad_norm_pre_clip': 16.610187530517578, 'timing/data': 0.03113289549946785, 'timing/model': 1.1272044517099857, 'learning_rate/qwen_vl_interface': 6.707542723714629e-06, 'learning_rate/action_model': 6.707542723714629e-05, 'epoch': 1.05}) +[2026-06-17 22:08:40,646][starVLA.training.train_starvla][INFO] - Step 2866, Loss: {'train/loss': 0.36607980728149414, 'train/grad_norm_pre_clip': 14.298803329467773, 'timing/data': 0.03644030820578337, 'timing/model': 1.042805423028767, 'learning_rate/qwen_vl_interface': 6.7054842156443435e-06, 'learning_rate/action_model': 6.705484215644343e-05, 'epoch': 1.05}) +[2026-06-17 22:08:45,323][starVLA.training.train_starvla][INFO] - Step 2867, Loss: {'train/loss': 0.3632465600967407, 'train/grad_norm_pre_clip': 12.785712242126465, 'timing/data': 0.14155634585767984, 'timing/model': 1.0903062568977475, 'learning_rate/qwen_vl_interface': 6.703425405850614e-06, 'learning_rate/action_model': 6.703425405850613e-05, 'epoch': 1.05}) +[2026-06-17 22:08:57,673][starVLA.training.train_starvla][INFO] - Step 2868, Loss: {'train/loss': 0.48881885409355164, 'train/grad_norm_pre_clip': 12.191707611083984, 'timing/data': 0.020835344679653645, 'timing/model': 3.1346377935260534, 'learning_rate/qwen_vl_interface': 6.701366294760236e-06, 'learning_rate/action_model': 6.701366294760235e-05, 'epoch': 1.05}) +[2026-06-17 22:09:02,052][starVLA.training.train_starvla][INFO] - Step 2869, Loss: {'train/loss': 0.5342887043952942, 'train/grad_norm_pre_clip': 13.177138328552246, 'timing/data': 0.04835613816976547, 'timing/model': 1.289690239354968, 'learning_rate/qwen_vl_interface': 6.6993068828000605e-06, 'learning_rate/action_model': 6.69930688280006e-05, 'epoch': 1.05}) +[2026-06-17 22:09:10,740][starVLA.training.train_starvla][INFO] - Step 2870, Loss: {'train/loss': 0.33863934874534607, 'train/grad_norm_pre_clip': 12.731281280517578, 'timing/data': 0.048935010097920895, 'timing/model': 4.2075430219992995, 'learning_rate/qwen_vl_interface': 6.697247170397011e-06, 'learning_rate/action_model': 6.697247170397011e-05, 'epoch': 1.05}) +[2026-06-17 22:09:18,080][starVLA.training.train_starvla][INFO] - Step 2871, Loss: {'train/loss': 0.524919867515564, 'train/grad_norm_pre_clip': 14.677885055541992, 'timing/data': 0.0426401486620307, 'timing/model': 1.1367877926677465, 'learning_rate/qwen_vl_interface': 6.6951871579780645e-06, 'learning_rate/action_model': 6.695187157978064e-05, 'epoch': 1.05}) +[2026-06-17 22:09:22,690][starVLA.training.train_starvla][INFO] - Step 2872, Loss: {'train/loss': 0.6119387745857239, 'train/grad_norm_pre_clip': 16.571184158325195, 'timing/data': 0.049569413997232914, 'timing/model': 1.054997838102281, 'learning_rate/qwen_vl_interface': 6.693126845970267e-06, 'learning_rate/action_model': 6.693126845970267e-05, 'epoch': 1.05}) +[2026-06-17 22:09:27,178][starVLA.training.train_starvla][INFO] - Step 2873, Loss: {'train/loss': 0.749852180480957, 'train/grad_norm_pre_clip': 18.485946655273438, 'timing/data': 0.05085236672312021, 'timing/model': 1.4379860823974013, 'learning_rate/qwen_vl_interface': 6.691066234800719e-06, 'learning_rate/action_model': 6.691066234800719e-05, 'epoch': 1.05}) +[2026-06-17 22:09:38,242][starVLA.training.train_starvla][INFO] - Step 2874, Loss: {'train/loss': 0.501164972782135, 'train/grad_norm_pre_clip': 15.19911003112793, 'timing/data': 0.10745949670672417, 'timing/model': 1.880664606578648, 'learning_rate/qwen_vl_interface': 6.689005324896591e-06, 'learning_rate/action_model': 6.689005324896591e-05, 'epoch': 1.05}) +[2026-06-17 22:09:46,425][starVLA.training.train_starvla][INFO] - Step 2875, Loss: {'train/loss': 0.4711807370185852, 'train/grad_norm_pre_clip': 12.948412895202637, 'timing/data': 0.05299768131226301, 'timing/model': 1.4773579752072692, 'learning_rate/qwen_vl_interface': 6.686944116685111e-06, 'learning_rate/action_model': 6.686944116685111e-05, 'epoch': 1.05}) +[2026-06-17 22:09:55,604][starVLA.training.train_starvla][INFO] - Step 2876, Loss: {'train/loss': 0.8277912735939026, 'train/grad_norm_pre_clip': 18.953182220458984, 'timing/data': 0.046798269264400005, 'timing/model': 1.0897077871486545, 'learning_rate/qwen_vl_interface': 6.684882610593568e-06, 'learning_rate/action_model': 6.684882610593567e-05, 'epoch': 1.05}) +[2026-06-17 22:10:02,672][starVLA.training.train_starvla][INFO] - Step 2877, Loss: {'train/loss': 1.138118028640747, 'train/grad_norm_pre_clip': 19.082860946655273, 'timing/data': 0.04816955979913473, 'timing/model': 1.3907890962436795, 'learning_rate/qwen_vl_interface': 6.682820807049317e-06, 'learning_rate/action_model': 6.682820807049317e-05, 'epoch': 1.05}) +[2026-06-17 22:10:07,727][starVLA.training.train_starvla][INFO] - Step 2878, Loss: {'train/loss': 0.6735659837722778, 'train/grad_norm_pre_clip': 16.61842155456543, 'timing/data': 0.039313952438533306, 'timing/model': 1.2181603033095598, 'learning_rate/qwen_vl_interface': 6.680758706479772e-06, 'learning_rate/action_model': 6.680758706479772e-05, 'epoch': 1.05}) +[2026-06-17 22:10:15,966][starVLA.training.train_starvla][INFO] - Step 2879, Loss: {'train/loss': 0.47747573256492615, 'train/grad_norm_pre_clip': 12.644501686096191, 'timing/data': 0.04507367592304945, 'timing/model': 1.5736084934324026, 'learning_rate/qwen_vl_interface': 6.678696309312405e-06, 'learning_rate/action_model': 6.678696309312406e-05, 'epoch': 1.05}) +[2026-06-17 22:10:29,713][starVLA.training.train_starvla][INFO] - Step 2880, Loss: {'train/loss': 0.5705469250679016, 'train/grad_norm_pre_clip': 15.093255996704102, 'timing/data': 0.09756297711282969, 'timing/model': 1.4091069139540195, 'learning_rate/qwen_vl_interface': 6.676633615974758e-06, 'learning_rate/action_model': 6.676633615974757e-05, 'epoch': 1.05}) +[2026-06-17 22:10:40,827][starVLA.training.train_starvla][INFO] - Step 2881, Loss: {'train/loss': 0.7118323445320129, 'train/grad_norm_pre_clip': 14.94443416595459, 'timing/data': 0.057987166568636894, 'timing/model': 1.3129640268161893, 'learning_rate/qwen_vl_interface': 6.6745706268944265e-06, 'learning_rate/action_model': 6.674570626894426e-05, 'epoch': 1.05}) +[2026-06-17 22:10:47,274][starVLA.training.train_starvla][INFO] - Step 2882, Loss: {'train/loss': 0.6637865900993347, 'train/grad_norm_pre_clip': 13.67199420928955, 'timing/data': 0.04348207917064428, 'timing/model': 1.1046288534998894, 'learning_rate/qwen_vl_interface': 6.672507342499071e-06, 'learning_rate/action_model': 6.67250734249907e-05, 'epoch': 1.06}) +[2026-06-17 22:10:52,866][starVLA.training.train_starvla][INFO] - Step 2883, Loss: {'train/loss': 0.5250831246376038, 'train/grad_norm_pre_clip': 12.676066398620605, 'timing/data': 0.04389350488781929, 'timing/model': 1.0459525594487786, 'learning_rate/qwen_vl_interface': 6.670443763216412e-06, 'learning_rate/action_model': 6.670443763216411e-05, 'epoch': 1.06}) +[2026-06-17 22:10:57,433][starVLA.training.train_starvla][INFO] - Step 2884, Loss: {'train/loss': 0.5868827700614929, 'train/grad_norm_pre_clip': 14.683181762695312, 'timing/data': 0.04634941183030605, 'timing/model': 1.3698995923623443, 'learning_rate/qwen_vl_interface': 6.668379889474233e-06, 'learning_rate/action_model': 6.668379889474234e-05, 'epoch': 1.06}) +[2026-06-17 22:11:01,575][starVLA.training.train_starvla][INFO] - Step 2885, Loss: {'train/loss': 0.5267800092697144, 'train/grad_norm_pre_clip': 15.72213363647461, 'timing/data': 0.04284566454589367, 'timing/model': 1.0271739913150668, 'learning_rate/qwen_vl_interface': 6.666315721700378e-06, 'learning_rate/action_model': 6.666315721700377e-05, 'epoch': 1.06}) +[2026-06-17 22:11:06,266][starVLA.training.train_starvla][INFO] - Step 2886, Loss: {'train/loss': 0.5321308374404907, 'train/grad_norm_pre_clip': 14.64744758605957, 'timing/data': 0.060692546889185905, 'timing/model': 1.0732351243495941, 'learning_rate/qwen_vl_interface': 6.664251260322748e-06, 'learning_rate/action_model': 6.664251260322749e-05, 'epoch': 1.06}) +[2026-06-17 22:11:10,791][starVLA.training.train_starvla][INFO] - Step 2887, Loss: {'train/loss': 0.6024599671363831, 'train/grad_norm_pre_clip': 16.363149642944336, 'timing/data': 0.04035494290292263, 'timing/model': 1.1066590119153261, 'learning_rate/qwen_vl_interface': 6.66218650576931e-06, 'learning_rate/action_model': 6.66218650576931e-05, 'epoch': 1.06}) +[2026-06-17 22:11:15,192][starVLA.training.train_starvla][INFO] - Step 2888, Loss: {'train/loss': 0.41997310519218445, 'train/grad_norm_pre_clip': 13.630759239196777, 'timing/data': 0.03813744056969881, 'timing/model': 1.0260297022759914, 'learning_rate/qwen_vl_interface': 6.66012145846809e-06, 'learning_rate/action_model': 6.660121458468091e-05, 'epoch': 1.06}) +[2026-06-17 22:11:19,679][starVLA.training.train_starvla][INFO] - Step 2889, Loss: {'train/loss': 0.7646633386611938, 'train/grad_norm_pre_clip': 18.139041900634766, 'timing/data': 0.04133452568203211, 'timing/model': 1.3174252677708864, 'learning_rate/qwen_vl_interface': 6.658056118847176e-06, 'learning_rate/action_model': 6.658056118847175e-05, 'epoch': 1.06}) +[2026-06-17 22:11:23,890][starVLA.training.train_starvla][INFO] - Step 2890, Loss: {'train/loss': 0.4137047231197357, 'train/grad_norm_pre_clip': 17.409631729125977, 'timing/data': 0.04132875334471464, 'timing/model': 1.1129918647930026, 'learning_rate/qwen_vl_interface': 6.655990487334713e-06, 'learning_rate/action_model': 6.655990487334712e-05, 'epoch': 1.06}) +[2026-06-17 22:11:28,210][starVLA.training.train_starvla][INFO] - Step 2891, Loss: {'train/loss': 0.4817127287387848, 'train/grad_norm_pre_clip': 13.190804481506348, 'timing/data': 0.03988752793520689, 'timing/model': 1.0392644125968218, 'learning_rate/qwen_vl_interface': 6.653924564358907e-06, 'learning_rate/action_model': 6.653924564358907e-05, 'epoch': 1.06}) +[2026-06-17 22:11:32,935][starVLA.training.train_starvla][INFO] - Step 2892, Loss: {'train/loss': 0.4047081768512726, 'train/grad_norm_pre_clip': 16.115821838378906, 'timing/data': 0.05284229014068842, 'timing/model': 1.1423859046772122, 'learning_rate/qwen_vl_interface': 6.6518583503480294e-06, 'learning_rate/action_model': 6.651858350348029e-05, 'epoch': 1.06}) +[2026-06-17 22:11:37,310][starVLA.training.train_starvla][INFO] - Step 2893, Loss: {'train/loss': 0.5956509113311768, 'train/grad_norm_pre_clip': 17.580705642700195, 'timing/data': 0.049986276775598526, 'timing/model': 1.0533125484362245, 'learning_rate/qwen_vl_interface': 6.649791845730408e-06, 'learning_rate/action_model': 6.649791845730408e-05, 'epoch': 1.06}) +[2026-06-17 22:11:42,070][starVLA.training.train_starvla][INFO] - Step 2894, Loss: {'train/loss': 0.515786349773407, 'train/grad_norm_pre_clip': 19.12138557434082, 'timing/data': 0.0427316315472126, 'timing/model': 1.5845894115045667, 'learning_rate/qwen_vl_interface': 6.64772505093443e-06, 'learning_rate/action_model': 6.647725050934429e-05, 'epoch': 1.06}) +[2026-06-17 22:11:46,132][starVLA.training.train_starvla][INFO] - Step 2895, Loss: {'train/loss': 0.6478363275527954, 'train/grad_norm_pre_clip': 16.571504592895508, 'timing/data': 0.0389041006565094, 'timing/model': 1.1010672161355615, 'learning_rate/qwen_vl_interface': 6.645657966388543e-06, 'learning_rate/action_model': 6.645657966388543e-05, 'epoch': 1.06}) +[2026-06-17 22:11:50,600][starVLA.training.train_starvla][INFO] - Step 2896, Loss: {'train/loss': 0.6220496892929077, 'train/grad_norm_pre_clip': 16.983327865600586, 'timing/data': 0.04599608015269041, 'timing/model': 1.1268235128372908, 'learning_rate/qwen_vl_interface': 6.643590592521261e-06, 'learning_rate/action_model': 6.643590592521261e-05, 'epoch': 1.06}) +[2026-06-17 22:11:55,256][starVLA.training.train_starvla][INFO] - Step 2897, Loss: {'train/loss': 0.32541441917419434, 'train/grad_norm_pre_clip': 11.301417350769043, 'timing/data': 0.04232178442180157, 'timing/model': 1.0598845025524497, 'learning_rate/qwen_vl_interface': 6.641522929761148e-06, 'learning_rate/action_model': 6.641522929761148e-05, 'epoch': 1.06}) +[2026-06-17 22:12:04,220][starVLA.training.train_starvla][INFO] - Step 2898, Loss: {'train/loss': 0.44126859307289124, 'train/grad_norm_pre_clip': 16.82002067565918, 'timing/data': 0.040847742930054665, 'timing/model': 1.0526396064087749, 'learning_rate/qwen_vl_interface': 6.639454978536835e-06, 'learning_rate/action_model': 6.639454978536834e-05, 'epoch': 1.06}) +[2026-06-17 22:12:10,148][starVLA.training.train_starvla][INFO] - Step 2899, Loss: {'train/loss': 0.6217300891876221, 'train/grad_norm_pre_clip': 16.983295440673828, 'timing/data': 0.041589909233152866, 'timing/model': 1.549014369957149, 'learning_rate/qwen_vl_interface': 6.63738673927701e-06, 'learning_rate/action_model': 6.63738673927701e-05, 'epoch': 1.06}) +[2026-06-17 22:16:16,834][starVLA.training.train_starvla][INFO] - Step 2900, Loss: {'train/loss': 0.8450590372085571, 'train/grad_norm_pre_clip': 19.353469848632812, 'eval/loss': 0.6512429714202881, 'eval/latency_2/loss': 0.6512429714202881, 'eval/demon_fire_precision': 0.9112474437627812, 'eval/demon_fire_recall': 0.9322175732217574, 'eval/demon_fire_f1': 0.9216132368148914, 'eval/demon_fire_rate_model': 0.6803005008347245, 'eval/demon_fire_rate_teacher': 0.6649972175848636, 'eval/demon_left_precision': 0.7887725975261656, 'eval/demon_left_recall': 0.8726315789473684, 'eval/demon_left_f1': 0.8285857071464268, 'eval/demon_left_rate_model': 0.2924318308291597, 'eval/demon_left_rate_teacher': 0.2643294379521425, 'eval/demon_right_precision': 0.8344671201814059, 'eval/demon_right_recall': 0.831638418079096, 'eval/demon_right_f1': 0.8330503678551217, 'eval/demon_right_rate_model': 0.24540901502504173, 'eval/demon_right_rate_teacher': 0.24624373956594323, 'eval/demon_move_f1': 0.8308180375007742, 'eval/demon_cc_f1': 0.8923571046002118, 'eval/latency_2/demon_fire_precision': 0.9112474437627812, 'eval/latency_2/demon_fire_recall': 0.9322175732217574, 'eval/latency_2/demon_fire_f1': 0.9216132368148914, 'eval/latency_2/demon_fire_rate_model': 0.6803005008347245, 'eval/latency_2/demon_fire_rate_teacher': 0.6649972175848636, 'eval/latency_2/demon_left_precision': 0.7887725975261656, 'eval/latency_2/demon_left_recall': 0.8726315789473684, 'eval/latency_2/demon_left_f1': 0.8285857071464268, 'eval/latency_2/demon_left_rate_model': 0.2924318308291597, 'eval/latency_2/demon_left_rate_teacher': 0.2643294379521425, 'eval/latency_2/demon_right_precision': 0.8344671201814059, 'eval/latency_2/demon_right_recall': 0.831638418079096, 'eval/latency_2/demon_right_f1': 0.8330503678551217, 'eval/latency_2/demon_right_rate_model': 0.24540901502504173, 'eval/latency_2/demon_right_rate_teacher': 0.24624373956594323, 'eval/latency_2/demon_move_f1': 0.8308180375007742, 'eval/latency_2/demon_cc_f1': 0.8923571046002118, 'timing/data': 0.043115449137985706, 'timing/model': 1.1108079142868519, 'learning_rate/qwen_vl_interface': 6.63531821241042e-06, 'learning_rate/action_model': 6.63531821241042e-05, 'epoch': 1.06}) +[2026-06-17 22:16:16,837][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_2900_state +[2026-06-17 22:16:16,847][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 22:16:25,702][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_2900_state/model.safetensors +[2026-06-17 22:16:41,258][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_2900_state/optimizer.bin +[2026-06-17 22:16:41,277][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_2900_state/sampler.bin +[2026-06-17 22:16:41,277][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_2900_state/sampler_1.bin +[2026-06-17 22:16:41,281][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_2900_state/random_states_0.pkl +[2026-06-17 22:16:41,285][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2900 +[2026-06-17 22:16:41,285][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 22:16:41,302][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 22:16:45,681][starVLA.training.train_starvla][INFO] - Step 2901, Loss: {'train/loss': 0.5082923769950867, 'train/grad_norm_pre_clip': 10.840113639831543, 'timing/data': 0.03954323846846819, 'timing/model': 1.059107480570674, 'learning_rate/qwen_vl_interface': 6.633249398365874e-06, 'learning_rate/action_model': 6.633249398365873e-05, 'epoch': 1.06}) +[2026-06-17 22:16:53,935][starVLA.training.train_starvla][INFO] - Step 2902, Loss: {'train/loss': 0.37363386154174805, 'train/grad_norm_pre_clip': 14.435691833496094, 'timing/data': 0.03935057297348976, 'timing/model': 1.059730731882155, 'learning_rate/qwen_vl_interface': 6.6311802975722395e-06, 'learning_rate/action_model': 6.63118029757224e-05, 'epoch': 1.06}) +[2026-06-17 22:16:58,592][starVLA.training.train_starvla][INFO] - Step 2903, Loss: {'train/loss': 0.8962112665176392, 'train/grad_norm_pre_clip': 20.491626739501953, 'timing/data': 0.045877791941165924, 'timing/model': 1.187553416006267, 'learning_rate/qwen_vl_interface': 6.629110910458444e-06, 'learning_rate/action_model': 6.629110910458444e-05, 'epoch': 1.06}) +[2026-06-17 22:17:03,407][starVLA.training.train_starvla][INFO] - Step 2904, Loss: {'train/loss': 0.304437518119812, 'train/grad_norm_pre_clip': 9.393231391906738, 'timing/data': 0.04373249597847462, 'timing/model': 1.2403589878231287, 'learning_rate/qwen_vl_interface': 6.627041237453471e-06, 'learning_rate/action_model': 6.627041237453471e-05, 'epoch': 1.06}) +[2026-06-17 22:17:08,325][starVLA.training.train_starvla][INFO] - Step 2905, Loss: {'train/loss': 0.47254782915115356, 'train/grad_norm_pre_clip': 14.844720840454102, 'timing/data': 0.04553218558430672, 'timing/model': 1.5523747270926833, 'learning_rate/qwen_vl_interface': 6.624971278986369e-06, 'learning_rate/action_model': 6.62497127898637e-05, 'epoch': 1.06}) +[2026-06-17 22:17:14,057][starVLA.training.train_starvla][INFO] - Step 2906, Loss: {'train/loss': 0.7974047660827637, 'train/grad_norm_pre_clip': 18.131805419921875, 'timing/data': 0.03944364469498396, 'timing/model': 1.0332416985183954, 'learning_rate/qwen_vl_interface': 6.622901035486241e-06, 'learning_rate/action_model': 6.622901035486241e-05, 'epoch': 1.06}) +[2026-06-17 22:17:18,500][starVLA.training.train_starvla][INFO] - Step 2907, Loss: {'train/loss': 0.5100205540657043, 'train/grad_norm_pre_clip': 14.729812622070312, 'timing/data': 0.05518912058323622, 'timing/model': 1.0509044332429767, 'learning_rate/qwen_vl_interface': 6.620830507382251e-06, 'learning_rate/action_model': 6.620830507382251e-05, 'epoch': 1.06}) +[2026-06-17 22:17:23,036][starVLA.training.train_starvla][INFO] - Step 2908, Loss: {'train/loss': 0.5983687043190002, 'train/grad_norm_pre_clip': 16.183425903320312, 'timing/data': 0.04726310260593891, 'timing/model': 1.085385812446475, 'learning_rate/qwen_vl_interface': 6.618759695103623e-06, 'learning_rate/action_model': 6.618759695103622e-05, 'epoch': 1.06}) +[2026-06-17 22:17:31,218][starVLA.training.train_starvla][INFO] - Step 2909, Loss: {'train/loss': 0.5380551218986511, 'train/grad_norm_pre_clip': 12.937660217285156, 'timing/data': 0.036878881976008415, 'timing/model': 4.634226729162037, 'learning_rate/qwen_vl_interface': 6.616688599079635e-06, 'learning_rate/action_model': 6.616688599079634e-05, 'epoch': 1.07}) +[2026-06-17 22:17:39,527][starVLA.training.train_starvla][INFO] - Step 2910, Loss: {'train/loss': 0.3486926853656769, 'train/grad_norm_pre_clip': 13.935603141784668, 'timing/data': 0.035785285755991936, 'timing/model': 1.6240882873535156, 'learning_rate/qwen_vl_interface': 6.614617219739632e-06, 'learning_rate/action_model': 6.614617219739632e-05, 'epoch': 1.07}) +[2026-06-17 22:17:43,748][starVLA.training.train_starvla][INFO] - Step 2911, Loss: {'train/loss': 0.5258346199989319, 'train/grad_norm_pre_clip': 13.081271171569824, 'timing/data': 0.044017137959599495, 'timing/model': 1.093491641804576, 'learning_rate/qwen_vl_interface': 6.612545557513008e-06, 'learning_rate/action_model': 6.612545557513008e-05, 'epoch': 1.07}) +[2026-06-17 22:17:48,405][starVLA.training.train_starvla][INFO] - Step 2912, Loss: {'train/loss': 0.9332010746002197, 'train/grad_norm_pre_clip': 19.12940788269043, 'timing/data': 0.04028244409710169, 'timing/model': 1.159489474259317, 'learning_rate/qwen_vl_interface': 6.610473612829227e-06, 'learning_rate/action_model': 6.610473612829226e-05, 'epoch': 1.07}) +[2026-06-17 22:17:52,979][starVLA.training.train_starvla][INFO] - Step 2913, Loss: {'train/loss': 0.4701346158981323, 'train/grad_norm_pre_clip': 18.12583351135254, 'timing/data': 0.03326867055147886, 'timing/model': 1.1364087527617812, 'learning_rate/qwen_vl_interface': 6.608401386117801e-06, 'learning_rate/action_model': 6.6084013861178e-05, 'epoch': 1.07}) +[2026-06-17 22:17:58,271][starVLA.training.train_starvla][INFO] - Step 2914, Loss: {'train/loss': 0.628858745098114, 'train/grad_norm_pre_clip': 17.487998962402344, 'timing/data': 0.02917620725929737, 'timing/model': 1.2777921659871936, 'learning_rate/qwen_vl_interface': 6.606328877808306e-06, 'learning_rate/action_model': 6.606328877808305e-05, 'epoch': 1.07}) +[2026-06-17 22:18:10,470][starVLA.training.train_starvla][INFO] - Step 2915, Loss: {'train/loss': 0.7893069982528687, 'train/grad_norm_pre_clip': 31.105634689331055, 'timing/data': 0.03612882923334837, 'timing/model': 5.579700178466737, 'learning_rate/qwen_vl_interface': 6.604256088330374e-06, 'learning_rate/action_model': 6.604256088330374e-05, 'epoch': 1.07}) +[2026-06-17 22:18:14,771][starVLA.training.train_starvla][INFO] - Step 2916, Loss: {'train/loss': 0.45682838559150696, 'train/grad_norm_pre_clip': 12.119171142578125, 'timing/data': 0.029663238674402237, 'timing/model': 1.0392408119514585, 'learning_rate/qwen_vl_interface': 6.602183018113698e-06, 'learning_rate/action_model': 6.602183018113698e-05, 'epoch': 1.07}) +[2026-06-17 22:18:19,405][starVLA.training.train_starvla][INFO] - Step 2917, Loss: {'train/loss': 0.5391657948493958, 'train/grad_norm_pre_clip': 16.623708724975586, 'timing/data': 0.03265553619712591, 'timing/model': 1.0829005744308233, 'learning_rate/qwen_vl_interface': 6.600109667588029e-06, 'learning_rate/action_model': 6.60010966758803e-05, 'epoch': 1.07}) +[2026-06-17 22:18:24,434][starVLA.training.train_starvla][INFO] - Step 2918, Loss: {'train/loss': 0.6983661651611328, 'train/grad_norm_pre_clip': 16.775846481323242, 'timing/data': 0.02817703690379858, 'timing/model': 1.2551033589988947, 'learning_rate/qwen_vl_interface': 6.598036037183172e-06, 'learning_rate/action_model': 6.598036037183172e-05, 'epoch': 1.07}) +[2026-06-17 22:18:33,309][starVLA.training.train_starvla][INFO] - Step 2919, Loss: {'train/loss': 0.33452296257019043, 'train/grad_norm_pre_clip': 11.006326675415039, 'timing/data': 0.029678862541913986, 'timing/model': 5.414690265432, 'learning_rate/qwen_vl_interface': 6.595962127328993e-06, 'learning_rate/action_model': 6.595962127328993e-05, 'epoch': 1.07}) +[2026-06-17 22:18:45,321][starVLA.training.train_starvla][INFO] - Step 2920, Loss: {'train/loss': 0.6746422648429871, 'train/grad_norm_pre_clip': 16.96982192993164, 'timing/data': 0.03322392608970404, 'timing/model': 1.3720662761479616, 'learning_rate/qwen_vl_interface': 6.593887938455418e-06, 'learning_rate/action_model': 6.593887938455418e-05, 'epoch': 1.07}) +[2026-06-17 22:18:49,608][starVLA.training.train_starvla][INFO] - Step 2921, Loss: {'train/loss': 0.5829049944877625, 'train/grad_norm_pre_clip': 17.113656997680664, 'timing/data': 0.033560484647750854, 'timing/model': 1.0817066729068756, 'learning_rate/qwen_vl_interface': 6.591813470992428e-06, 'learning_rate/action_model': 6.591813470992427e-05, 'epoch': 1.07}) +[2026-06-17 22:18:55,644][starVLA.training.train_starvla][INFO] - Step 2922, Loss: {'train/loss': 0.4941418170928955, 'train/grad_norm_pre_clip': 13.57015609741211, 'timing/data': 0.0355796180665493, 'timing/model': 1.0408543990924954, 'learning_rate/qwen_vl_interface': 6.58973872537006e-06, 'learning_rate/action_model': 6.589738725370059e-05, 'epoch': 1.07}) +[2026-06-17 22:19:03,691][starVLA.training.train_starvla][INFO] - Step 2923, Loss: {'train/loss': 0.5841147899627686, 'train/grad_norm_pre_clip': 13.634235382080078, 'timing/data': 0.038221622817218304, 'timing/model': 1.2337881485000253, 'learning_rate/qwen_vl_interface': 6.587663702018413e-06, 'learning_rate/action_model': 6.587663702018412e-05, 'epoch': 1.07}) +[2026-06-17 22:19:08,883][starVLA.training.train_starvla][INFO] - Step 2924, Loss: {'train/loss': 0.42841288447380066, 'train/grad_norm_pre_clip': 18.305091857910156, 'timing/data': 0.028112669475376606, 'timing/model': 1.2030128184705973, 'learning_rate/qwen_vl_interface': 6.585588401367641e-06, 'learning_rate/action_model': 6.58558840136764e-05, 'epoch': 1.07}) +[2026-06-17 22:19:13,426][starVLA.training.train_starvla][INFO] - Step 2925, Loss: {'train/loss': 0.7859103679656982, 'train/grad_norm_pre_clip': 14.335820198059082, 'timing/data': 0.02865943219512701, 'timing/model': 1.3892110707238317, 'learning_rate/qwen_vl_interface': 6.583512823847957e-06, 'learning_rate/action_model': 6.583512823847958e-05, 'epoch': 1.07}) +[2026-06-17 22:19:17,580][starVLA.training.train_starvla][INFO] - Step 2926, Loss: {'train/loss': 0.758773684501648, 'train/grad_norm_pre_clip': 16.838314056396484, 'timing/data': 0.04256669245660305, 'timing/model': 1.0319101735949516, 'learning_rate/qwen_vl_interface': 6.5814369698896285e-06, 'learning_rate/action_model': 6.581436969889628e-05, 'epoch': 1.07}) +[2026-06-17 22:19:22,572][starVLA.training.train_starvla][INFO] - Step 2927, Loss: {'train/loss': 0.29328101873397827, 'train/grad_norm_pre_clip': 11.885376930236816, 'timing/data': 0.038924651220440865, 'timing/model': 1.1926405457779765, 'learning_rate/qwen_vl_interface': 6.5793608399229855e-06, 'learning_rate/action_model': 6.579360839922985e-05, 'epoch': 1.07}) +[2026-06-17 22:19:27,101][starVLA.training.train_starvla][INFO] - Step 2928, Loss: {'train/loss': 0.7905523777008057, 'train/grad_norm_pre_clip': 19.712711334228516, 'timing/data': 0.031316651962697506, 'timing/model': 1.1252077417448163, 'learning_rate/qwen_vl_interface': 6.577284434378408e-06, 'learning_rate/action_model': 6.577284434378408e-05, 'epoch': 1.07}) +[2026-06-17 22:19:31,465][starVLA.training.train_starvla][INFO] - Step 2929, Loss: {'train/loss': 0.6001394391059875, 'train/grad_norm_pre_clip': 14.357847213745117, 'timing/data': 0.02891316544264555, 'timing/model': 1.0213178154081106, 'learning_rate/qwen_vl_interface': 6.5752077536863404e-06, 'learning_rate/action_model': 6.57520775368634e-05, 'epoch': 1.07}) +[2026-06-17 22:19:35,912][starVLA.training.train_starvla][INFO] - Step 2930, Loss: {'train/loss': 0.6865147948265076, 'train/grad_norm_pre_clip': 16.459924697875977, 'timing/data': 0.028494280762970448, 'timing/model': 1.3910826118662953, 'learning_rate/qwen_vl_interface': 6.573130798277277e-06, 'learning_rate/action_model': 6.573130798277277e-05, 'epoch': 1.07}) +[2026-06-17 22:19:40,102][starVLA.training.train_starvla][INFO] - Step 2931, Loss: {'train/loss': 0.5396637916564941, 'train/grad_norm_pre_clip': 22.10042381286621, 'timing/data': 0.03250441234558821, 'timing/model': 1.080452635884285, 'learning_rate/qwen_vl_interface': 6.571053568581778e-06, 'learning_rate/action_model': 6.571053568581777e-05, 'epoch': 1.07}) +[2026-06-17 22:19:44,734][starVLA.training.train_starvla][INFO] - Step 2932, Loss: {'train/loss': 0.3384479582309723, 'train/grad_norm_pre_clip': 12.43608570098877, 'timing/data': 0.02920982986688614, 'timing/model': 1.1655273921787739, 'learning_rate/qwen_vl_interface': 6.568976065030453e-06, 'learning_rate/action_model': 6.568976065030453e-05, 'epoch': 1.07}) +[2026-06-17 22:19:49,347][starVLA.training.train_starvla][INFO] - Step 2933, Loss: {'train/loss': 0.6341797709465027, 'train/grad_norm_pre_clip': 17.332609176635742, 'timing/data': 0.029314100742340088, 'timing/model': 1.0488463342189789, 'learning_rate/qwen_vl_interface': 6.566898288053968e-06, 'learning_rate/action_model': 6.566898288053968e-05, 'epoch': 1.07}) +[2026-06-17 22:19:53,941][starVLA.training.train_starvla][INFO] - Step 2934, Loss: {'train/loss': 0.5421633720397949, 'train/grad_norm_pre_clip': 14.369152069091797, 'timing/data': 0.04098101146519184, 'timing/model': 1.1095815598964691, 'learning_rate/qwen_vl_interface': 6.564820238083053e-06, 'learning_rate/action_model': 6.564820238083052e-05, 'epoch': 1.07}) +[2026-06-17 22:20:00,045][starVLA.training.train_starvla][INFO] - Step 2935, Loss: {'train/loss': 0.6952170729637146, 'train/grad_norm_pre_clip': 16.61365509033203, 'timing/data': 0.028231022879481316, 'timing/model': 1.506084868684411, 'learning_rate/qwen_vl_interface': 6.562741915548486e-06, 'learning_rate/action_model': 6.562741915548486e-05, 'epoch': 1.07}) +[2026-06-17 22:20:04,093][starVLA.training.train_starvla][INFO] - Step 2936, Loss: {'train/loss': 0.6165042519569397, 'train/grad_norm_pre_clip': 18.1435604095459, 'timing/data': 0.03263192530721426, 'timing/model': 1.0450021028518677, 'learning_rate/qwen_vl_interface': 6.5606633208811096e-06, 'learning_rate/action_model': 6.560663320881109e-05, 'epoch': 1.08}) +[2026-06-17 22:20:10,354][starVLA.training.train_starvla][INFO] - Step 2937, Loss: {'train/loss': 0.8645418882369995, 'train/grad_norm_pre_clip': 21.56896209716797, 'timing/data': 0.03304638899862766, 'timing/model': 1.0440752813592553, 'learning_rate/qwen_vl_interface': 6.558584454511814e-06, 'learning_rate/action_model': 6.558584454511814e-05, 'epoch': 1.08}) +[2026-06-17 22:20:14,752][starVLA.training.train_starvla][INFO] - Step 2938, Loss: {'train/loss': 0.6805824041366577, 'train/grad_norm_pre_clip': 17.916996002197266, 'timing/data': 0.03422651160508394, 'timing/model': 1.0534208873286843, 'learning_rate/qwen_vl_interface': 6.556505316871555e-06, 'learning_rate/action_model': 6.556505316871554e-05, 'epoch': 1.08}) +[2026-06-17 22:20:19,320][starVLA.training.train_starvla][INFO] - Step 2939, Loss: {'train/loss': 0.7501336336135864, 'train/grad_norm_pre_clip': 16.34001350402832, 'timing/data': 0.11844686605036259, 'timing/model': 1.0536882458254695, 'learning_rate/qwen_vl_interface': 6.554425908391337e-06, 'learning_rate/action_model': 6.554425908391336e-05, 'epoch': 1.08}) +[2026-06-17 22:20:27,428][starVLA.training.train_starvla][INFO] - Step 2940, Loss: {'train/loss': 0.4465658664703369, 'train/grad_norm_pre_clip': 14.399558067321777, 'timing/data': 0.022085603326559067, 'timing/model': 1.4351263642311096, 'learning_rate/qwen_vl_interface': 6.552346229502224e-06, 'learning_rate/action_model': 6.552346229502224e-05, 'epoch': 1.08}) +[2026-06-17 22:20:31,589][starVLA.training.train_starvla][INFO] - Step 2941, Loss: {'train/loss': 1.0516599416732788, 'train/grad_norm_pre_clip': 18.629520416259766, 'timing/data': 0.0402127867564559, 'timing/model': 1.0302593149244785, 'learning_rate/qwen_vl_interface': 6.550266280635337e-06, 'learning_rate/action_model': 6.550266280635336e-05, 'epoch': 1.08}) +[2026-06-17 22:20:39,887][starVLA.training.train_starvla][INFO] - Step 2942, Loss: {'train/loss': 0.3908810615539551, 'train/grad_norm_pre_clip': 11.921026229858398, 'timing/data': 0.0500941202044487, 'timing/model': 1.0487475069239736, 'learning_rate/qwen_vl_interface': 6.54818606222185e-06, 'learning_rate/action_model': 6.548186062221849e-05, 'epoch': 1.08}) +[2026-06-17 22:20:44,690][starVLA.training.train_starvla][INFO] - Step 2943, Loss: {'train/loss': 0.41104912757873535, 'train/grad_norm_pre_clip': 11.838561058044434, 'timing/data': 0.0394987128674984, 'timing/model': 1.1949779940769076, 'learning_rate/qwen_vl_interface': 6.546105574692997e-06, 'learning_rate/action_model': 6.546105574692997e-05, 'epoch': 1.08}) +[2026-06-17 22:20:49,236][starVLA.training.train_starvla][INFO] - Step 2944, Loss: {'train/loss': 0.9077635407447815, 'train/grad_norm_pre_clip': 19.742782592773438, 'timing/data': 0.04039504751563072, 'timing/model': 1.0504579842090607, 'learning_rate/qwen_vl_interface': 6.544024818480061e-06, 'learning_rate/action_model': 6.544024818480062e-05, 'epoch': 1.08}) +[2026-06-17 22:20:53,703][starVLA.training.train_starvla][INFO] - Step 2945, Loss: {'train/loss': 1.016641616821289, 'train/grad_norm_pre_clip': 18.774362564086914, 'timing/data': 0.04018347151577473, 'timing/model': 1.370581435970962, 'learning_rate/qwen_vl_interface': 6.5419437940143905e-06, 'learning_rate/action_model': 6.54194379401439e-05, 'epoch': 1.08}) +[2026-06-17 22:20:58,065][starVLA.training.train_starvla][INFO] - Step 2946, Loss: {'train/loss': 0.4783455431461334, 'train/grad_norm_pre_clip': 14.750489234924316, 'timing/data': 0.041204389184713364, 'timing/model': 1.084926108829677, 'learning_rate/qwen_vl_interface': 6.5398625017273795e-06, 'learning_rate/action_model': 6.539862501727379e-05, 'epoch': 1.08}) +[2026-06-17 22:21:02,693][starVLA.training.train_starvla][INFO] - Step 2947, Loss: {'train/loss': 0.8655214905738831, 'train/grad_norm_pre_clip': 17.999887466430664, 'timing/data': 0.044541011564433575, 'timing/model': 1.0939307399094105, 'learning_rate/qwen_vl_interface': 6.537780942050486e-06, 'learning_rate/action_model': 6.537780942050486e-05, 'epoch': 1.08}) +[2026-06-17 22:21:07,254][starVLA.training.train_starvla][INFO] - Step 2948, Loss: {'train/loss': 0.6128063797950745, 'train/grad_norm_pre_clip': 16.518657684326172, 'timing/data': 0.040579731576144695, 'timing/model': 1.1068344572559, 'learning_rate/qwen_vl_interface': 6.535699115415216e-06, 'learning_rate/action_model': 6.535699115415216e-05, 'epoch': 1.08}) +[2026-06-17 22:21:14,918][starVLA.training.train_starvla][INFO] - Step 2949, Loss: {'train/loss': 0.3779343068599701, 'train/grad_norm_pre_clip': 11.854541778564453, 'timing/data': 0.051446583122015, 'timing/model': 1.1015312215313315, 'learning_rate/qwen_vl_interface': 6.533617022253137e-06, 'learning_rate/action_model': 6.533617022253137e-05, 'epoch': 1.08}) +[2026-06-17 22:21:19,301][starVLA.training.train_starvla][INFO] - Step 2950, Loss: {'train/loss': 0.42463013529777527, 'train/grad_norm_pre_clip': 41.41982650756836, 'timing/data': 0.03913930896669626, 'timing/model': 1.285518259741366, 'learning_rate/qwen_vl_interface': 6.531534662995866e-06, 'learning_rate/action_model': 6.531534662995866e-05, 'epoch': 1.08}) +[2026-06-17 22:21:23,365][starVLA.training.train_starvla][INFO] - Step 2951, Loss: {'train/loss': 0.7541838884353638, 'train/grad_norm_pre_clip': 16.210243225097656, 'timing/data': 0.03802680782973766, 'timing/model': 1.0246880073100328, 'learning_rate/qwen_vl_interface': 6.529452038075082e-06, 'learning_rate/action_model': 6.529452038075081e-05, 'epoch': 1.08}) +[2026-06-17 22:21:27,922][starVLA.training.train_starvla][INFO] - Step 2952, Loss: {'train/loss': 0.49077582359313965, 'train/grad_norm_pre_clip': 12.557367324829102, 'timing/data': 0.036831362172961235, 'timing/model': 1.0507814455777407, 'learning_rate/qwen_vl_interface': 6.5273691479225124e-06, 'learning_rate/action_model': 6.527369147922511e-05, 'epoch': 1.08}) +[2026-06-17 22:21:36,839][starVLA.training.train_starvla][INFO] - Step 2953, Loss: {'train/loss': 0.8287625312805176, 'train/grad_norm_pre_clip': 18.823997497558594, 'timing/data': 0.05482547730207443, 'timing/model': 1.1802630480378866, 'learning_rate/qwen_vl_interface': 6.525285992969945e-06, 'learning_rate/action_model': 6.525285992969945e-05, 'epoch': 1.08}) +[2026-06-17 22:21:48,836][starVLA.training.train_starvla][INFO] - Step 2954, Loss: {'train/loss': 0.4814087450504303, 'train/grad_norm_pre_clip': 11.498064994812012, 'timing/data': 0.040594582445919514, 'timing/model': 1.076607727445662, 'learning_rate/qwen_vl_interface': 6.523202573649219e-06, 'learning_rate/action_model': 6.523202573649219e-05, 'epoch': 1.08}) +[2026-06-17 22:21:53,332][starVLA.training.train_starvla][INFO] - Step 2955, Loss: {'train/loss': 0.7082771062850952, 'train/grad_norm_pre_clip': 15.511168479919434, 'timing/data': 0.03906696755439043, 'timing/model': 1.4601819459348917, 'learning_rate/qwen_vl_interface': 6.5211188903922294e-06, 'learning_rate/action_model': 6.52111889039223e-05, 'epoch': 1.08}) +[2026-06-17 22:21:57,673][starVLA.training.train_starvla][INFO] - Step 2956, Loss: {'train/loss': 0.7143460512161255, 'train/grad_norm_pre_clip': 21.346939086914062, 'timing/data': 0.03790897689759731, 'timing/model': 1.0841216556727886, 'learning_rate/qwen_vl_interface': 6.519034943630926e-06, 'learning_rate/action_model': 6.519034943630927e-05, 'epoch': 1.08}) +[2026-06-17 22:22:02,092][starVLA.training.train_starvla][INFO] - Step 2957, Loss: {'train/loss': 0.7036586999893188, 'train/grad_norm_pre_clip': 19.367332458496094, 'timing/data': 0.03644593060016632, 'timing/model': 1.0754320714622736, 'learning_rate/qwen_vl_interface': 6.5169507337973125e-06, 'learning_rate/action_model': 6.516950733797312e-05, 'epoch': 1.08}) +[2026-06-17 22:22:11,706][starVLA.training.train_starvla][INFO] - Step 2958, Loss: {'train/loss': 0.4512818157672882, 'train/grad_norm_pre_clip': 12.780128479003906, 'timing/data': 0.045375644229352474, 'timing/model': 4.731625461019576, 'learning_rate/qwen_vl_interface': 6.51486626132345e-06, 'learning_rate/action_model': 6.51486626132345e-05, 'epoch': 1.08}) +[2026-06-17 22:22:20,754][starVLA.training.train_starvla][INFO] - Step 2959, Loss: {'train/loss': 0.7024632096290588, 'train/grad_norm_pre_clip': 14.818846702575684, 'timing/data': 0.04648020304739475, 'timing/model': 1.0296068312600255, 'learning_rate/qwen_vl_interface': 6.512781526641449e-06, 'learning_rate/action_model': 6.512781526641449e-05, 'epoch': 1.08}) +[2026-06-17 22:22:25,270][starVLA.training.train_starvla][INFO] - Step 2960, Loss: {'train/loss': 0.7896921038627625, 'train/grad_norm_pre_clip': 15.651565551757812, 'timing/data': 0.039285000413656235, 'timing/model': 1.314730972982943, 'learning_rate/qwen_vl_interface': 6.51069653018348e-06, 'learning_rate/action_model': 6.51069653018348e-05, 'epoch': 1.08}) +[2026-06-17 22:22:32,829][starVLA.training.train_starvla][INFO] - Step 2961, Loss: {'train/loss': 0.6852044463157654, 'train/grad_norm_pre_clip': 15.95016860961914, 'timing/data': 0.03949455823749304, 'timing/model': 4.239610097371042, 'learning_rate/qwen_vl_interface': 6.508611272381763e-06, 'learning_rate/action_model': 6.508611272381763e-05, 'epoch': 1.08}) +[2026-06-17 22:22:42,631][starVLA.training.train_starvla][INFO] - Step 2962, Loss: {'train/loss': 0.49521568417549133, 'train/grad_norm_pre_clip': 13.36966323852539, 'timing/data': 0.044840590097010136, 'timing/model': 1.0849888836964965, 'learning_rate/qwen_vl_interface': 6.506525753668576e-06, 'learning_rate/action_model': 6.506525753668576e-05, 'epoch': 1.08}) +[2026-06-17 22:22:47,446][starVLA.training.train_starvla][INFO] - Step 2963, Loss: {'train/loss': 0.6376315951347351, 'train/grad_norm_pre_clip': 15.295429229736328, 'timing/data': 0.04932300467044115, 'timing/model': 1.1560625825077295, 'learning_rate/qwen_vl_interface': 6.504439974476246e-06, 'learning_rate/action_model': 6.504439974476246e-05, 'epoch': 1.08}) +[2026-06-17 22:22:51,845][starVLA.training.train_starvla][INFO] - Step 2964, Loss: {'train/loss': 0.46403923630714417, 'train/grad_norm_pre_clip': 17.268590927124023, 'timing/data': 0.03595637995749712, 'timing/model': 1.036357051692903, 'learning_rate/qwen_vl_interface': 6.50235393523716e-06, 'learning_rate/action_model': 6.502353935237159e-05, 'epoch': 1.09}) +[2026-06-17 22:22:56,672][starVLA.training.train_starvla][INFO] - Step 2965, Loss: {'train/loss': 0.6622786521911621, 'train/grad_norm_pre_clip': 13.790690422058105, 'timing/data': 0.04589149076491594, 'timing/model': 1.5183925358578563, 'learning_rate/qwen_vl_interface': 6.500267636383758e-06, 'learning_rate/action_model': 6.500267636383758e-05, 'epoch': 1.09}) +[2026-06-17 22:23:01,671][starVLA.training.train_starvla][INFO] - Step 2966, Loss: {'train/loss': 0.5779446363449097, 'train/grad_norm_pre_clip': 15.339909553527832, 'timing/data': 0.03876884561032057, 'timing/model': 1.069760457612574, 'learning_rate/qwen_vl_interface': 6.498181078348527e-06, 'learning_rate/action_model': 6.498181078348527e-05, 'epoch': 1.09}) +[2026-06-17 22:23:11,822][starVLA.training.train_starvla][INFO] - Step 2967, Loss: {'train/loss': 0.7174906134605408, 'train/grad_norm_pre_clip': 17.810537338256836, 'timing/data': 0.0399467833340168, 'timing/model': 1.1349041415378451, 'learning_rate/qwen_vl_interface': 6.496094261564018e-06, 'learning_rate/action_model': 6.496094261564018e-05, 'epoch': 1.09}) +[2026-06-17 22:23:16,438][starVLA.training.train_starvla][INFO] - Step 2968, Loss: {'train/loss': 0.30684271454811096, 'train/grad_norm_pre_clip': 12.856664657592773, 'timing/data': 0.04452885314822197, 'timing/model': 1.078874435275793, 'learning_rate/qwen_vl_interface': 6.494007186462824e-06, 'learning_rate/action_model': 6.494007186462823e-05, 'epoch': 1.09}) +[2026-06-17 22:23:26,807][starVLA.training.train_starvla][INFO] - Step 2969, Loss: {'train/loss': 0.9037020802497864, 'train/grad_norm_pre_clip': 18.659149169921875, 'timing/data': 0.03916650917381048, 'timing/model': 4.0346889132633805, 'learning_rate/qwen_vl_interface': 6.491919853477604e-06, 'learning_rate/action_model': 6.491919853477604e-05, 'epoch': 1.09}) +[2026-06-17 22:23:31,618][starVLA.training.train_starvla][INFO] - Step 2970, Loss: {'train/loss': 0.4842800796031952, 'train/grad_norm_pre_clip': 11.922235488891602, 'timing/data': 0.04580753669142723, 'timing/model': 1.4535890072584152, 'learning_rate/qwen_vl_interface': 6.4898322630410615e-06, 'learning_rate/action_model': 6.489832263041062e-05, 'epoch': 1.09}) +[2026-06-17 22:23:35,849][starVLA.training.train_starvla][INFO] - Step 2971, Loss: {'train/loss': 0.795936107635498, 'train/grad_norm_pre_clip': 17.21491050720215, 'timing/data': 0.04193422757089138, 'timing/model': 1.0505764000117779, 'learning_rate/qwen_vl_interface': 6.487744415585958e-06, 'learning_rate/action_model': 6.487744415585958e-05, 'epoch': 1.09}) +[2026-06-17 22:23:40,489][starVLA.training.train_starvla][INFO] - Step 2972, Loss: {'train/loss': 0.5074283480644226, 'train/grad_norm_pre_clip': 15.503372192382812, 'timing/data': 0.04335892014205456, 'timing/model': 1.0551564255729318, 'learning_rate/qwen_vl_interface': 6.485656311545103e-06, 'learning_rate/action_model': 6.485656311545103e-05, 'epoch': 1.09}) +[2026-06-17 22:23:45,117][starVLA.training.train_starvla][INFO] - Step 2973, Loss: {'train/loss': 0.4664030373096466, 'train/grad_norm_pre_clip': 12.252731323242188, 'timing/data': 0.036513577215373516, 'timing/model': 1.0430586747825146, 'learning_rate/qwen_vl_interface': 6.483567951351366e-06, 'learning_rate/action_model': 6.483567951351365e-05, 'epoch': 1.09}) +[2026-06-17 22:23:49,762][starVLA.training.train_starvla][INFO] - Step 2974, Loss: {'train/loss': 0.4365729093551636, 'train/grad_norm_pre_clip': 13.59524154663086, 'timing/data': 0.04483702126890421, 'timing/model': 1.1734255095943809, 'learning_rate/qwen_vl_interface': 6.481479335437667e-06, 'learning_rate/action_model': 6.481479335437667e-05, 'epoch': 1.09}) +[2026-06-17 22:23:54,291][starVLA.training.train_starvla][INFO] - Step 2975, Loss: {'train/loss': 0.7238317131996155, 'train/grad_norm_pre_clip': 17.071245193481445, 'timing/data': 0.035281178541481495, 'timing/model': 1.3208021577447653, 'learning_rate/qwen_vl_interface': 6.479390464236974e-06, 'learning_rate/action_model': 6.479390464236974e-05, 'epoch': 1.09}) +[2026-06-17 22:24:01,694][starVLA.training.train_starvla][INFO] - Step 2976, Loss: {'train/loss': 0.7490572929382324, 'train/grad_norm_pre_clip': 16.305355072021484, 'timing/data': 0.04358530975878239, 'timing/model': 1.1704849246889353, 'learning_rate/qwen_vl_interface': 6.477301338182317e-06, 'learning_rate/action_model': 6.477301338182317e-05, 'epoch': 1.09}) +[2026-06-17 22:24:06,257][starVLA.training.train_starvla][INFO] - Step 2977, Loss: {'train/loss': 0.3457324802875519, 'train/grad_norm_pre_clip': 17.218461990356445, 'timing/data': 0.0438344432041049, 'timing/model': 1.0324697913601995, 'learning_rate/qwen_vl_interface': 6.47521195770677e-06, 'learning_rate/action_model': 6.47521195770677e-05, 'epoch': 1.09}) +[2026-06-17 22:24:13,711][starVLA.training.train_starvla][INFO] - Step 2978, Loss: {'train/loss': 0.46498724818229675, 'train/grad_norm_pre_clip': 12.787415504455566, 'timing/data': 0.038523921743035316, 'timing/model': 1.134387637488544, 'learning_rate/qwen_vl_interface': 6.473122323243467e-06, 'learning_rate/action_model': 6.473122323243467e-05, 'epoch': 1.09}) +[2026-06-17 22:24:18,399][starVLA.training.train_starvla][INFO] - Step 2979, Loss: {'train/loss': 0.8664596676826477, 'train/grad_norm_pre_clip': 19.542112350463867, 'timing/data': 0.03646787069737911, 'timing/model': 1.1205913797020912, 'learning_rate/qwen_vl_interface': 6.471032435225591e-06, 'learning_rate/action_model': 6.47103243522559e-05, 'epoch': 1.09}) +[2026-06-17 22:24:23,266][starVLA.training.train_starvla][INFO] - Step 2980, Loss: {'train/loss': 0.8319273591041565, 'train/grad_norm_pre_clip': 21.915721893310547, 'timing/data': 0.04208654072135687, 'timing/model': 1.6158163761720061, 'learning_rate/qwen_vl_interface': 6.468942294086376e-06, 'learning_rate/action_model': 6.468942294086376e-05, 'epoch': 1.09}) +[2026-06-17 22:24:27,674][starVLA.training.train_starvla][INFO] - Step 2981, Loss: {'train/loss': 0.5957810282707214, 'train/grad_norm_pre_clip': 14.95609188079834, 'timing/data': 0.05813252925872803, 'timing/model': 1.2136667324230075, 'learning_rate/qwen_vl_interface': 6.466851900259113e-06, 'learning_rate/action_model': 6.466851900259112e-05, 'epoch': 1.09}) +[2026-06-17 22:24:32,131][starVLA.training.train_starvla][INFO] - Step 2982, Loss: {'train/loss': 0.5138002634048462, 'train/grad_norm_pre_clip': 15.36172103881836, 'timing/data': 0.03980696201324463, 'timing/model': 1.203728104941547, 'learning_rate/qwen_vl_interface': 6.464761254177139e-06, 'learning_rate/action_model': 6.464761254177138e-05, 'epoch': 1.09}) +[2026-06-17 22:24:37,020][starVLA.training.train_starvla][INFO] - Step 2983, Loss: {'train/loss': 0.5772387385368347, 'train/grad_norm_pre_clip': 22.27501106262207, 'timing/data': 0.05574908573180437, 'timing/model': 1.2793305590748787, 'learning_rate/qwen_vl_interface': 6.462670356273852e-06, 'learning_rate/action_model': 6.462670356273852e-05, 'epoch': 1.09}) +[2026-06-17 22:24:48,614][starVLA.training.train_starvla][INFO] - Step 2984, Loss: {'train/loss': 0.3136943578720093, 'train/grad_norm_pre_clip': 13.315868377685547, 'timing/data': 0.04491074662655592, 'timing/model': 1.0802854942157865, 'learning_rate/qwen_vl_interface': 6.460579206982695e-06, 'learning_rate/action_model': 6.460579206982696e-05, 'epoch': 1.09}) +[2026-06-17 22:24:57,369][starVLA.training.train_starvla][INFO] - Step 2985, Loss: {'train/loss': 0.4868805706501007, 'train/grad_norm_pre_clip': 13.658897399902344, 'timing/data': 0.04460453614592552, 'timing/model': 3.9709672573953867, 'learning_rate/qwen_vl_interface': 6.458487806737166e-06, 'learning_rate/action_model': 6.458487806737165e-05, 'epoch': 1.09}) +[2026-06-17 22:25:05,133][starVLA.training.train_starvla][INFO] - Step 2986, Loss: {'train/loss': 0.9065859913825989, 'train/grad_norm_pre_clip': 19.289365768432617, 'timing/data': 0.044533505104482174, 'timing/model': 1.2016817638650537, 'learning_rate/qwen_vl_interface': 6.456396155970814e-06, 'learning_rate/action_model': 6.456396155970814e-05, 'epoch': 1.09}) +[2026-06-17 22:25:10,251][starVLA.training.train_starvla][INFO] - Step 2987, Loss: {'train/loss': 0.8191487193107605, 'train/grad_norm_pre_clip': 18.93255043029785, 'timing/data': 0.04712165705859661, 'timing/model': 1.2248451057821512, 'learning_rate/qwen_vl_interface': 6.4543042551172404e-06, 'learning_rate/action_model': 6.45430425511724e-05, 'epoch': 1.09}) +[2026-06-17 22:25:24,448][starVLA.training.train_starvla][INFO] - Step 2988, Loss: {'train/loss': 0.6185438632965088, 'train/grad_norm_pre_clip': 15.40628433227539, 'timing/data': 0.10416750516742468, 'timing/model': 1.098624630831182, 'learning_rate/qwen_vl_interface': 6.4522121046101e-06, 'learning_rate/action_model': 6.452212104610099e-05, 'epoch': 1.09}) +[2026-06-17 22:25:29,542][starVLA.training.train_starvla][INFO] - Step 2989, Loss: {'train/loss': 0.3142179250717163, 'train/grad_norm_pre_clip': 11.116558074951172, 'timing/data': 0.042150745168328285, 'timing/model': 1.3423357298597693, 'learning_rate/qwen_vl_interface': 6.450119704883095e-06, 'learning_rate/action_model': 6.450119704883096e-05, 'epoch': 1.09}) +[2026-06-17 22:25:37,437][starVLA.training.train_starvla][INFO] - Step 2990, Loss: {'train/loss': 0.44333571195602417, 'train/grad_norm_pre_clip': 13.998940467834473, 'timing/data': 0.0776792149990797, 'timing/model': 1.7811200534924865, 'learning_rate/qwen_vl_interface': 6.448027056369983e-06, 'learning_rate/action_model': 6.448027056369983e-05, 'epoch': 1.09}) +[2026-06-17 22:25:42,231][starVLA.training.train_starvla][INFO] - Step 2991, Loss: {'train/loss': 0.8314012289047241, 'train/grad_norm_pre_clip': 18.174842834472656, 'timing/data': 0.03976218122988939, 'timing/model': 1.158879560418427, 'learning_rate/qwen_vl_interface': 6.445934159504576e-06, 'learning_rate/action_model': 6.445934159504576e-05, 'epoch': 1.1}) +[2026-06-17 22:25:48,803][starVLA.training.train_starvla][INFO] - Step 2992, Loss: {'train/loss': 0.4963792860507965, 'train/grad_norm_pre_clip': 18.875696182250977, 'timing/data': 0.05109735019505024, 'timing/model': 1.2851028377190232, 'learning_rate/qwen_vl_interface': 6.44384101472073e-06, 'learning_rate/action_model': 6.443841014720729e-05, 'epoch': 1.1}) +[2026-06-17 22:25:53,550][starVLA.training.train_starvla][INFO] - Step 2993, Loss: {'train/loss': 0.5215142965316772, 'train/grad_norm_pre_clip': 14.449530601501465, 'timing/data': 0.04213784448802471, 'timing/model': 1.035419194959104, 'learning_rate/qwen_vl_interface': 6.441747622452357e-06, 'learning_rate/action_model': 6.441747622452356e-05, 'epoch': 1.1}) +[2026-06-17 22:25:58,122][starVLA.training.train_starvla][INFO] - Step 2994, Loss: {'train/loss': 0.5612762570381165, 'train/grad_norm_pre_clip': 13.052640914916992, 'timing/data': 0.04391597304493189, 'timing/model': 1.218069870956242, 'learning_rate/qwen_vl_interface': 6.4396539831334195e-06, 'learning_rate/action_model': 6.439653983133419e-05, 'epoch': 1.1}) +[2026-06-17 22:26:02,646][starVLA.training.train_starvla][INFO] - Step 2995, Loss: {'train/loss': 0.41988882422447205, 'train/grad_norm_pre_clip': 10.51060676574707, 'timing/data': 0.048901330679655075, 'timing/model': 1.4231417858973145, 'learning_rate/qwen_vl_interface': 6.437560097197928e-06, 'learning_rate/action_model': 6.437560097197927e-05, 'epoch': 1.1}) +[2026-06-17 22:26:06,837][starVLA.training.train_starvla][INFO] - Step 2996, Loss: {'train/loss': 0.9227285981178284, 'train/grad_norm_pre_clip': 24.846017837524414, 'timing/data': 0.035267459228634834, 'timing/model': 1.1456342656165361, 'learning_rate/qwen_vl_interface': 6.4354659650799546e-06, 'learning_rate/action_model': 6.435465965079955e-05, 'epoch': 1.1}) +[2026-06-17 22:26:11,239][starVLA.training.train_starvla][INFO] - Step 2997, Loss: {'train/loss': 0.5246860980987549, 'train/grad_norm_pre_clip': 17.45313835144043, 'timing/data': 0.040527017787098885, 'timing/model': 1.019127949140966, 'learning_rate/qwen_vl_interface': 6.433371587213609e-06, 'learning_rate/action_model': 6.433371587213607e-05, 'epoch': 1.1}) +[2026-06-17 22:26:15,712][starVLA.training.train_starvla][INFO] - Step 2998, Loss: {'train/loss': 0.870294451713562, 'train/grad_norm_pre_clip': 20.02613067626953, 'timing/data': 0.039294296875596046, 'timing/model': 1.0301733585074544, 'learning_rate/qwen_vl_interface': 6.431276964033062e-06, 'learning_rate/action_model': 6.431276964033062e-05, 'epoch': 1.1}) +[2026-06-17 22:26:20,111][starVLA.training.train_starvla][INFO] - Step 2999, Loss: {'train/loss': 0.6602540016174316, 'train/grad_norm_pre_clip': 17.1947021484375, 'timing/data': 0.044716848991811275, 'timing/model': 1.0815388979390264, 'learning_rate/qwen_vl_interface': 6.429182095972526e-06, 'learning_rate/action_model': 6.429182095972526e-05, 'epoch': 1.1}) +[2026-06-17 22:29:51,797][starVLA.training.train_starvla][INFO] - Step 3000, Loss: {'train/loss': 0.6805081963539124, 'train/grad_norm_pre_clip': 16.538719177246094, 'eval/loss': 0.6401004195213318, 'eval/latency_2/loss': 0.6401004195213318, 'eval/demon_fire_precision': 0.9285417532197756, 'eval/demon_fire_recall': 0.9351464435146444, 'eval/demon_fire_f1': 0.9318323952470293, 'eval/demon_fire_rate_model': 0.6697273233166389, 'eval/demon_fire_rate_teacher': 0.6649972175848636, 'eval/demon_left_precision': 0.8694074969770254, 'eval/demon_left_recall': 0.7568421052631579, 'eval/demon_left_f1': 0.8092290377039956, 'eval/demon_left_rate_model': 0.23010573177518084, 'eval/demon_left_rate_teacher': 0.2643294379521425, 'eval/demon_right_precision': 0.8108941418293937, 'eval/demon_right_recall': 0.8915254237288136, 'eval/demon_right_f1': 0.8493003229278795, 'eval/demon_right_rate_model': 0.27072899276572066, 'eval/demon_right_rate_teacher': 0.24624373956594323, 'eval/demon_move_f1': 0.8292646803159376, 'eval/demon_cc_f1': 0.8984933148406538, 'eval/latency_2/demon_fire_precision': 0.9285417532197756, 'eval/latency_2/demon_fire_recall': 0.9351464435146444, 'eval/latency_2/demon_fire_f1': 0.9318323952470293, 'eval/latency_2/demon_fire_rate_model': 0.6697273233166389, 'eval/latency_2/demon_fire_rate_teacher': 0.6649972175848636, 'eval/latency_2/demon_left_precision': 0.8694074969770254, 'eval/latency_2/demon_left_recall': 0.7568421052631579, 'eval/latency_2/demon_left_f1': 0.8092290377039956, 'eval/latency_2/demon_left_rate_model': 0.23010573177518084, 'eval/latency_2/demon_left_rate_teacher': 0.2643294379521425, 'eval/latency_2/demon_right_precision': 0.8108941418293937, 'eval/latency_2/demon_right_recall': 0.8915254237288136, 'eval/latency_2/demon_right_f1': 0.8493003229278795, 'eval/latency_2/demon_right_rate_model': 0.27072899276572066, 'eval/latency_2/demon_right_rate_teacher': 0.24624373956594323, 'eval/latency_2/demon_move_f1': 0.8292646803159376, 'eval/latency_2/demon_cc_f1': 0.8984933148406538, 'timing/data': 0.044348619878292084, 'timing/model': 1.3487658509984612, 'learning_rate/qwen_vl_interface': 6.427086983466275e-06, 'learning_rate/action_model': 6.427086983466275e-05, 'epoch': 1.1}) +[2026-06-17 22:29:51,799][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_3000_state +[2026-06-17 22:29:51,814][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 22:30:00,441][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_3000_state/model.safetensors +[2026-06-17 22:30:15,903][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_3000_state/optimizer.bin +[2026-06-17 22:30:15,904][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_3000_state/sampler.bin +[2026-06-17 22:30:15,904][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_3000_state/sampler_1.bin +[2026-06-17 22:30:15,907][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_3000_state/random_states_0.pkl +[2026-06-17 22:30:15,931][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3000 +[2026-06-17 22:30:15,931][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 22:30:15,948][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 22:30:20,702][starVLA.training.train_starvla][INFO] - Step 3001, Loss: {'train/loss': 0.655888020992279, 'train/grad_norm_pre_clip': 12.748757362365723, 'timing/data': 0.042211693711578846, 'timing/model': 1.3788997046649456, 'learning_rate/qwen_vl_interface': 6.4249916269486255e-06, 'learning_rate/action_model': 6.424991626948626e-05, 'epoch': 1.1}) +[2026-06-17 22:30:25,169][starVLA.training.train_starvla][INFO] - Step 3002, Loss: {'train/loss': 0.4159301519393921, 'train/grad_norm_pre_clip': 9.654074668884277, 'timing/data': 0.035800750367343426, 'timing/model': 1.020486505702138, 'learning_rate/qwen_vl_interface': 6.422896026853948e-06, 'learning_rate/action_model': 6.422896026853947e-05, 'epoch': 1.1}) +[2026-06-17 22:30:30,098][starVLA.training.train_starvla][INFO] - Step 3003, Loss: {'train/loss': 0.4003945589065552, 'train/grad_norm_pre_clip': 10.499143600463867, 'timing/data': 0.039644946344196796, 'timing/model': 1.2920050360262394, 'learning_rate/qwen_vl_interface': 6.420800183616661e-06, 'learning_rate/action_model': 6.420800183616661e-05, 'epoch': 1.1}) +[2026-06-17 22:30:39,239][starVLA.training.train_starvla][INFO] - Step 3004, Loss: {'train/loss': 0.39449918270111084, 'train/grad_norm_pre_clip': 12.010382652282715, 'timing/data': 0.03943227045238018, 'timing/model': 1.0276571540161967, 'learning_rate/qwen_vl_interface': 6.4187040976712355e-06, 'learning_rate/action_model': 6.418704097671235e-05, 'epoch': 1.1}) +[2026-06-17 22:30:43,741][starVLA.training.train_starvla][INFO] - Step 3005, Loss: {'train/loss': 0.649040699005127, 'train/grad_norm_pre_clip': 15.43419361114502, 'timing/data': 0.04014988988637924, 'timing/model': 1.1468477891758084, 'learning_rate/qwen_vl_interface': 6.416607769452195e-06, 'learning_rate/action_model': 6.416607769452194e-05, 'epoch': 1.1}) +[2026-06-17 22:30:48,426][starVLA.training.train_starvla][INFO] - Step 3006, Loss: {'train/loss': 0.5734633207321167, 'train/grad_norm_pre_clip': 14.149182319641113, 'timing/data': 0.04797875229269266, 'timing/model': 1.5752863297238946, 'learning_rate/qwen_vl_interface': 6.414511199394106e-06, 'learning_rate/action_model': 6.414511199394105e-05, 'epoch': 1.1}) +[2026-06-17 22:30:52,748][starVLA.training.train_starvla][INFO] - Step 3007, Loss: {'train/loss': 0.45923635363578796, 'train/grad_norm_pre_clip': 12.557677268981934, 'timing/data': 0.04477890767157078, 'timing/model': 1.0321049010381103, 'learning_rate/qwen_vl_interface': 6.4124143879315935e-06, 'learning_rate/action_model': 6.412414387931593e-05, 'epoch': 1.1}) +[2026-06-17 22:30:57,572][starVLA.training.train_starvla][INFO] - Step 3008, Loss: {'train/loss': 0.6383810639381409, 'train/grad_norm_pre_clip': 15.542591094970703, 'timing/data': 0.04070576932281256, 'timing/model': 1.2590493950992823, 'learning_rate/qwen_vl_interface': 6.410317335499328e-06, 'learning_rate/action_model': 6.410317335499328e-05, 'epoch': 1.1}) +[2026-06-17 22:31:02,081][starVLA.training.train_starvla][INFO] - Step 3009, Loss: {'train/loss': 0.3831827640533447, 'train/grad_norm_pre_clip': 8.99498176574707, 'timing/data': 0.09074910916388035, 'timing/model': 1.0426941188052297, 'learning_rate/qwen_vl_interface': 6.40822004253203e-06, 'learning_rate/action_model': 6.40822004253203e-05, 'epoch': 1.1}) +[2026-06-17 22:31:08,492][starVLA.training.train_starvla][INFO] - Step 3010, Loss: {'train/loss': 0.6846150755882263, 'train/grad_norm_pre_clip': 16.828231811523438, 'timing/data': 0.046710711903870106, 'timing/model': 1.1375332092866302, 'learning_rate/qwen_vl_interface': 6.406122509464471e-06, 'learning_rate/action_model': 6.40612250946447e-05, 'epoch': 1.1}) +[2026-06-17 22:31:12,956][starVLA.training.train_starvla][INFO] - Step 3011, Loss: {'train/loss': 0.8113297820091248, 'train/grad_norm_pre_clip': 20.1717586517334, 'timing/data': 0.04185785725712776, 'timing/model': 1.3091768911108375, 'learning_rate/qwen_vl_interface': 6.404024736731472e-06, 'learning_rate/action_model': 6.404024736731472e-05, 'epoch': 1.1}) +[2026-06-17 22:31:17,335][starVLA.training.train_starvla][INFO] - Step 3012, Loss: {'train/loss': 0.6461365222930908, 'train/grad_norm_pre_clip': 15.042248725891113, 'timing/data': 0.053030249662697315, 'timing/model': 1.0891471719369292, 'learning_rate/qwen_vl_interface': 6.401926724767903e-06, 'learning_rate/action_model': 6.401926724767902e-05, 'epoch': 1.1}) +[2026-06-17 22:31:22,039][starVLA.training.train_starvla][INFO] - Step 3013, Loss: {'train/loss': 0.6262356042861938, 'train/grad_norm_pre_clip': 21.30500030517578, 'timing/data': 0.03996819630265236, 'timing/model': 1.1157303731888533, 'learning_rate/qwen_vl_interface': 6.399828474008683e-06, 'learning_rate/action_model': 6.399828474008682e-05, 'epoch': 1.1}) +[2026-06-17 22:31:32,678][starVLA.training.train_starvla][INFO] - Step 3014, Loss: {'train/loss': 0.47180357575416565, 'train/grad_norm_pre_clip': 14.294881820678711, 'timing/data': 0.04341753479093313, 'timing/model': 1.1275800988078117, 'learning_rate/qwen_vl_interface': 6.397729984888785e-06, 'learning_rate/action_model': 6.397729984888785e-05, 'epoch': 1.1}) +[2026-06-17 22:31:37,426][starVLA.training.train_starvla][INFO] - Step 3015, Loss: {'train/loss': 0.8056908249855042, 'train/grad_norm_pre_clip': 16.486717224121094, 'timing/data': 0.048419639468193054, 'timing/model': 1.0754252495244145, 'learning_rate/qwen_vl_interface': 6.395631257843224e-06, 'learning_rate/action_model': 6.395631257843224e-05, 'epoch': 1.1}) +[2026-06-17 22:31:42,112][starVLA.training.train_starvla][INFO] - Step 3016, Loss: {'train/loss': 0.5332682132720947, 'train/grad_norm_pre_clip': 16.79468536376953, 'timing/data': 0.041031054221093655, 'timing/model': 1.4789259769022465, 'learning_rate/qwen_vl_interface': 6.393532293307073e-06, 'learning_rate/action_model': 6.393532293307072e-05, 'epoch': 1.1}) +[2026-06-17 22:31:46,520][starVLA.training.train_starvla][INFO] - Step 3017, Loss: {'train/loss': 0.49670320749282837, 'train/grad_norm_pre_clip': 12.60812759399414, 'timing/data': 0.04083655681461096, 'timing/model': 1.0904469909146428, 'learning_rate/qwen_vl_interface': 6.391433091715447e-06, 'learning_rate/action_model': 6.391433091715446e-05, 'epoch': 1.1}) +[2026-06-17 22:31:51,210][starVLA.training.train_starvla][INFO] - Step 3018, Loss: {'train/loss': 0.4211522042751312, 'train/grad_norm_pre_clip': 15.728096961975098, 'timing/data': 0.03710189554840326, 'timing/model': 1.03018035646528, 'learning_rate/qwen_vl_interface': 6.389333653503511e-06, 'learning_rate/action_model': 6.38933365350351e-05, 'epoch': 1.11}) +[2026-06-17 22:31:59,151][starVLA.training.train_starvla][INFO] - Step 3019, Loss: {'train/loss': 0.663105309009552, 'train/grad_norm_pre_clip': 15.403847694396973, 'timing/data': 0.11570333782583475, 'timing/model': 1.1107233325019479, 'learning_rate/qwen_vl_interface': 6.387233979106484e-06, 'learning_rate/action_model': 6.387233979106483e-05, 'epoch': 1.11}) +[2026-06-17 22:32:09,403][starVLA.training.train_starvla][INFO] - Step 3020, Loss: {'train/loss': 0.45882609486579895, 'train/grad_norm_pre_clip': 13.591191291809082, 'timing/data': 0.0122502027079463, 'timing/model': 1.0479722581803799, 'learning_rate/qwen_vl_interface': 6.385134068959629e-06, 'learning_rate/action_model': 6.385134068959628e-05, 'epoch': 1.11}) +[2026-06-17 22:32:13,758][starVLA.training.train_starvla][INFO] - Step 3021, Loss: {'train/loss': 0.6932231783866882, 'train/grad_norm_pre_clip': 14.131444931030273, 'timing/data': 0.01121770404279232, 'timing/model': 1.3000004040077329, 'learning_rate/qwen_vl_interface': 6.3830339234982596e-06, 'learning_rate/action_model': 6.38303392349826e-05, 'epoch': 1.11}) +[2026-06-17 22:32:17,842][starVLA.training.train_starvla][INFO] - Step 3022, Loss: {'train/loss': 0.633216917514801, 'train/grad_norm_pre_clip': 32.09650421142578, 'timing/data': 0.0137404166162014, 'timing/model': 1.0575874922797084, 'learning_rate/qwen_vl_interface': 6.38093354315774e-06, 'learning_rate/action_model': 6.38093354315774e-05, 'epoch': 1.11}) +[2026-06-17 22:32:22,273][starVLA.training.train_starvla][INFO] - Step 3023, Loss: {'train/loss': 0.5694279670715332, 'train/grad_norm_pre_clip': 13.687315940856934, 'timing/data': 0.013848905451595783, 'timing/model': 1.0502023426815867, 'learning_rate/qwen_vl_interface': 6.378832928373479e-06, 'learning_rate/action_model': 6.378832928373478e-05, 'epoch': 1.11}) +[2026-06-17 22:32:27,054][starVLA.training.train_starvla][INFO] - Step 3024, Loss: {'train/loss': 0.3939337134361267, 'train/grad_norm_pre_clip': 12.075942039489746, 'timing/data': 0.014179402962327003, 'timing/model': 1.1214878344908357, 'learning_rate/qwen_vl_interface': 6.376732079580936e-06, 'learning_rate/action_model': 6.376732079580936e-05, 'epoch': 1.11}) +[2026-06-17 22:32:35,803][starVLA.training.train_starvla][INFO] - Step 3025, Loss: {'train/loss': 0.7223290801048279, 'train/grad_norm_pre_clip': 18.981361389160156, 'timing/data': 0.011485337279736996, 'timing/model': 1.2410581316798925, 'learning_rate/qwen_vl_interface': 6.374630997215623e-06, 'learning_rate/action_model': 6.374630997215623e-05, 'epoch': 1.11}) +[2026-06-17 22:32:40,423][starVLA.training.train_starvla][INFO] - Step 3026, Loss: {'train/loss': 0.5918391346931458, 'train/grad_norm_pre_clip': 35.33667755126953, 'timing/data': 0.011627125553786755, 'timing/model': 1.3403564039617777, 'learning_rate/qwen_vl_interface': 6.372529681713093e-06, 'learning_rate/action_model': 6.372529681713094e-05, 'epoch': 1.11}) +[2026-06-17 22:32:44,492][starVLA.training.train_starvla][INFO] - Step 3027, Loss: {'train/loss': 0.3228350281715393, 'train/grad_norm_pre_clip': 10.429974555969238, 'timing/data': 0.011943479999899864, 'timing/model': 1.0815506586804986, 'learning_rate/qwen_vl_interface': 6.370428133508953e-06, 'learning_rate/action_model': 6.370428133508953e-05, 'epoch': 1.11}) +[2026-06-17 22:32:50,790][starVLA.training.train_starvla][INFO] - Step 3028, Loss: {'train/loss': 0.7433357238769531, 'train/grad_norm_pre_clip': 20.04144287109375, 'timing/data': 0.011941798962652683, 'timing/model': 1.2152126356959343, 'learning_rate/qwen_vl_interface': 6.368326353038852e-06, 'learning_rate/action_model': 6.368326353038853e-05, 'epoch': 1.11}) +[2026-06-17 22:32:58,814][starVLA.training.train_starvla][INFO] - Step 3029, Loss: {'train/loss': 0.5292418599128723, 'train/grad_norm_pre_clip': 18.05280303955078, 'timing/data': 0.012998195365071297, 'timing/model': 1.0530015164986253, 'learning_rate/qwen_vl_interface': 6.366224340738499e-06, 'learning_rate/action_model': 6.366224340738499e-05, 'epoch': 1.11}) +[2026-06-17 22:33:03,434][starVLA.training.train_starvla][INFO] - Step 3030, Loss: {'train/loss': 0.7599844336509705, 'train/grad_norm_pre_clip': 19.010854721069336, 'timing/data': 0.01371077261865139, 'timing/model': 1.5067997137084603, 'learning_rate/qwen_vl_interface': 6.3641220970436355e-06, 'learning_rate/action_model': 6.364122097043635e-05, 'epoch': 1.11}) +[2026-06-17 22:33:12,347][starVLA.training.train_starvla][INFO] - Step 3031, Loss: {'train/loss': 0.6104832291603088, 'train/grad_norm_pre_clip': 17.086458206176758, 'timing/data': 0.013008086010813713, 'timing/model': 1.4179437905550003, 'learning_rate/qwen_vl_interface': 6.362019622390065e-06, 'learning_rate/action_model': 6.362019622390066e-05, 'epoch': 1.11}) +[2026-06-17 22:33:18,176][starVLA.training.train_starvla][INFO] - Step 3032, Loss: {'train/loss': 0.6087501049041748, 'train/grad_norm_pre_clip': 17.603031158447266, 'timing/data': 0.01428287010639906, 'timing/model': 1.0306651452556252, 'learning_rate/qwen_vl_interface': 6.359916917213629e-06, 'learning_rate/action_model': 6.359916917213629e-05, 'epoch': 1.11}) +[2026-06-17 22:33:22,553][starVLA.training.train_starvla][INFO] - Step 3033, Loss: {'train/loss': 0.7021151781082153, 'train/grad_norm_pre_clip': 18.986793518066406, 'timing/data': 0.013399320654571056, 'timing/model': 1.025425181724131, 'learning_rate/qwen_vl_interface': 6.357813981950223e-06, 'learning_rate/action_model': 6.357813981950223e-05, 'epoch': 1.11}) +[2026-06-17 22:33:26,900][starVLA.training.train_starvla][INFO] - Step 3034, Loss: {'train/loss': 0.5282028913497925, 'train/grad_norm_pre_clip': 15.088921546936035, 'timing/data': 0.011323734186589718, 'timing/model': 1.0488935466855764, 'learning_rate/qwen_vl_interface': 6.355710817035784e-06, 'learning_rate/action_model': 6.355710817035783e-05, 'epoch': 1.11}) +[2026-06-17 22:33:34,720][starVLA.training.train_starvla][INFO] - Step 3035, Loss: {'train/loss': 0.8176973462104797, 'train/grad_norm_pre_clip': 16.60610008239746, 'timing/data': 0.010920505039393902, 'timing/model': 4.312046238221228, 'learning_rate/qwen_vl_interface': 6.353607422906304e-06, 'learning_rate/action_model': 6.353607422906304e-05, 'epoch': 1.11}) +[2026-06-17 22:33:41,636][starVLA.training.train_starvla][INFO] - Step 3036, Loss: {'train/loss': 0.4135507345199585, 'train/grad_norm_pre_clip': 17.309768676757812, 'timing/data': 0.012309357523918152, 'timing/model': 1.5139499278739095, 'learning_rate/qwen_vl_interface': 6.351503799997819e-06, 'learning_rate/action_model': 6.351503799997818e-05, 'epoch': 1.11}) +[2026-06-17 22:33:45,829][starVLA.training.train_starvla][INFO] - Step 3037, Loss: {'train/loss': 0.5410304069519043, 'train/grad_norm_pre_clip': 19.570180892944336, 'timing/data': 0.02049457747489214, 'timing/model': 1.153385067358613, 'learning_rate/qwen_vl_interface': 6.349399948746409e-06, 'learning_rate/action_model': 6.349399948746409e-05, 'epoch': 1.11}) +[2026-06-17 22:33:50,408][starVLA.training.train_starvla][INFO] - Step 3038, Loss: {'train/loss': 1.179290771484375, 'train/grad_norm_pre_clip': 23.664905548095703, 'timing/data': 0.010962639935314655, 'timing/model': 1.1821777857840061, 'learning_rate/qwen_vl_interface': 6.347295869588208e-06, 'learning_rate/action_model': 6.347295869588208e-05, 'epoch': 1.11}) +[2026-06-17 22:33:55,258][starVLA.training.train_starvla][INFO] - Step 3039, Loss: {'train/loss': 0.5827406644821167, 'train/grad_norm_pre_clip': 18.98118782043457, 'timing/data': 0.013204770162701607, 'timing/model': 1.0660321125760674, 'learning_rate/qwen_vl_interface': 6.345191562959392e-06, 'learning_rate/action_model': 6.345191562959391e-05, 'epoch': 1.11}) +[2026-06-17 22:34:00,022][starVLA.training.train_starvla][INFO] - Step 3040, Loss: {'train/loss': 0.5025241374969482, 'train/grad_norm_pre_clip': 19.361934661865234, 'timing/data': 0.01135498657822609, 'timing/model': 1.3554498022422194, 'learning_rate/qwen_vl_interface': 6.343087029296187e-06, 'learning_rate/action_model': 6.343087029296186e-05, 'epoch': 1.11}) +[2026-06-17 22:34:09,935][starVLA.training.train_starvla][INFO] - Step 3041, Loss: {'train/loss': 0.7237038016319275, 'train/grad_norm_pre_clip': 17.940902709960938, 'timing/data': 0.010597982443869114, 'timing/model': 5.82132310140878, 'learning_rate/qwen_vl_interface': 6.340982269034863e-06, 'learning_rate/action_model': 6.340982269034863e-05, 'epoch': 1.11}) +[2026-06-17 22:34:20,260][starVLA.training.train_starvla][INFO] - Step 3042, Loss: {'train/loss': 0.6431981921195984, 'train/grad_norm_pre_clip': 21.26363754272461, 'timing/data': 0.0015085218474268913, 'timing/model': 1.2607114361599088, 'learning_rate/qwen_vl_interface': 6.3388772826117415e-06, 'learning_rate/action_model': 6.338877282611742e-05, 'epoch': 1.11}) +[2026-06-17 22:34:25,233][starVLA.training.train_starvla][INFO] - Step 3043, Loss: {'train/loss': 0.40413305163383484, 'train/grad_norm_pre_clip': 14.177884101867676, 'timing/data': 0.003992816433310509, 'timing/model': 1.328682598657906, 'learning_rate/qwen_vl_interface': 6.3367720704631895e-06, 'learning_rate/action_model': 6.33677207046319e-05, 'epoch': 1.11}) +[2026-06-17 22:34:29,565][starVLA.training.train_starvla][INFO] - Step 3044, Loss: {'train/loss': 1.3761616945266724, 'train/grad_norm_pre_clip': 25.921098709106445, 'timing/data': 0.000931725837290287, 'timing/model': 1.0409181071445346, 'learning_rate/qwen_vl_interface': 6.334666633025616e-06, 'learning_rate/action_model': 6.334666633025615e-05, 'epoch': 1.11}) +[2026-06-17 22:34:34,113][starVLA.training.train_starvla][INFO] - Step 3045, Loss: {'train/loss': 0.4132526218891144, 'train/grad_norm_pre_clip': 16.593923568725586, 'timing/data': 0.0014659324660897255, 'timing/model': 1.17893807310611, 'learning_rate/qwen_vl_interface': 6.332560970735482e-06, 'learning_rate/action_model': 6.332560970735482e-05, 'epoch': 1.12}) +[2026-06-17 22:34:38,710][starVLA.training.train_starvla][INFO] - Step 3046, Loss: {'train/loss': 0.5990078449249268, 'train/grad_norm_pre_clip': 16.397539138793945, 'timing/data': 0.0010329028591513634, 'timing/model': 1.4707393432036042, 'learning_rate/qwen_vl_interface': 6.3304550840292945e-06, 'learning_rate/action_model': 6.330455084029294e-05, 'epoch': 1.12}) +[2026-06-17 22:34:43,078][starVLA.training.train_starvla][INFO] - Step 3047, Loss: {'train/loss': 0.4513389468193054, 'train/grad_norm_pre_clip': 14.269646644592285, 'timing/data': 0.0010772524401545525, 'timing/model': 1.1082306690514088, 'learning_rate/qwen_vl_interface': 6.328348973343605e-06, 'learning_rate/action_model': 6.328348973343605e-05, 'epoch': 1.12}) +[2026-06-17 22:34:47,613][starVLA.training.train_starvla][INFO] - Step 3048, Loss: {'train/loss': 0.45190003514289856, 'train/grad_norm_pre_clip': 13.119199752807617, 'timing/data': 0.00099116750061512, 'timing/model': 1.0536266760900617, 'learning_rate/qwen_vl_interface': 6.326242639115014e-06, 'learning_rate/action_model': 6.326242639115015e-05, 'epoch': 1.12}) +[2026-06-17 22:34:55,509][starVLA.training.train_starvla][INFO] - Step 3049, Loss: {'train/loss': 0.5373172760009766, 'train/grad_norm_pre_clip': 14.802114486694336, 'timing/data': 0.0009960159659385681, 'timing/model': 1.1131421467289329, 'learning_rate/qwen_vl_interface': 6.324136081780166e-06, 'learning_rate/action_model': 6.324136081780165e-05, 'epoch': 1.12}) +[2026-06-17 22:35:01,081][starVLA.training.train_starvla][INFO] - Step 3050, Loss: {'train/loss': 0.6630986928939819, 'train/grad_norm_pre_clip': 15.784605026245117, 'timing/data': 0.07342579681426287, 'timing/model': 1.666894607245922, 'learning_rate/qwen_vl_interface': 6.322029301775751e-06, 'learning_rate/action_model': 6.32202930177575e-05, 'epoch': 1.12}) +[2026-06-17 22:35:05,252][starVLA.training.train_starvla][INFO] - Step 3051, Loss: {'train/loss': 0.523848295211792, 'train/grad_norm_pre_clip': 14.922693252563477, 'timing/data': 0.03546627704054117, 'timing/model': 1.02185610961169, 'learning_rate/qwen_vl_interface': 6.3199222995385105e-06, 'learning_rate/action_model': 6.31992229953851e-05, 'epoch': 1.12}) +[2026-06-17 22:35:09,707][starVLA.training.train_starvla][INFO] - Step 3052, Loss: {'train/loss': 0.6489930152893066, 'train/grad_norm_pre_clip': 14.647055625915527, 'timing/data': 0.020425422117114067, 'timing/model': 1.035109193995595, 'learning_rate/qwen_vl_interface': 6.317815075505223e-06, 'learning_rate/action_model': 6.317815075505222e-05, 'epoch': 1.12}) +[2026-06-17 22:35:14,395][starVLA.training.train_starvla][INFO] - Step 3053, Loss: {'train/loss': 0.27484801411628723, 'train/grad_norm_pre_clip': 9.29019546508789, 'timing/data': 0.021865577436983585, 'timing/model': 1.1405933806672692, 'learning_rate/qwen_vl_interface': 6.315707630112724e-06, 'learning_rate/action_model': 6.315707630112723e-05, 'epoch': 1.12}) +[2026-06-17 22:35:20,817][starVLA.training.train_starvla][INFO] - Step 3054, Loss: {'train/loss': 0.5370057821273804, 'train/grad_norm_pre_clip': 15.065921783447266, 'timing/data': 0.026405763812363148, 'timing/model': 2.809578544460237, 'learning_rate/qwen_vl_interface': 6.313599963797884e-06, 'learning_rate/action_model': 6.313599963797884e-05, 'epoch': 1.12}) +[2026-06-17 22:35:28,766][starVLA.training.train_starvla][INFO] - Step 3055, Loss: {'train/loss': 0.7208505272865295, 'train/grad_norm_pre_clip': 19.961320877075195, 'timing/data': 0.03081580251455307, 'timing/model': 1.0417519453912973, 'learning_rate/qwen_vl_interface': 6.311492076997629e-06, 'learning_rate/action_model': 6.311492076997629e-05, 'epoch': 1.12}) +[2026-06-17 22:35:33,244][starVLA.training.train_starvla][INFO] - Step 3056, Loss: {'train/loss': 0.7361369729042053, 'train/grad_norm_pre_clip': 17.929882049560547, 'timing/data': 0.03363882377743721, 'timing/model': 1.3556996686384082, 'learning_rate/qwen_vl_interface': 6.309383970148922e-06, 'learning_rate/action_model': 6.309383970148922e-05, 'epoch': 1.12}) +[2026-06-17 22:35:42,033][starVLA.training.train_starvla][INFO] - Step 3057, Loss: {'train/loss': 0.5521343350410461, 'train/grad_norm_pre_clip': 17.945749282836914, 'timing/data': 0.03573527280241251, 'timing/model': 1.1010836381465197, 'learning_rate/qwen_vl_interface': 6.3072756436887805e-06, 'learning_rate/action_model': 6.30727564368878e-05, 'epoch': 1.12}) +[2026-06-17 22:35:46,563][starVLA.training.train_starvla][INFO] - Step 3058, Loss: {'train/loss': 0.34892237186431885, 'train/grad_norm_pre_clip': 13.744318008422852, 'timing/data': 0.021658985875546932, 'timing/model': 1.1528179980814457, 'learning_rate/qwen_vl_interface': 6.30516709805426e-06, 'learning_rate/action_model': 6.30516709805426e-05, 'epoch': 1.12}) +[2026-06-17 22:35:53,553][starVLA.training.train_starvla][INFO] - Step 3059, Loss: {'train/loss': 0.4525830149650574, 'train/grad_norm_pre_clip': 12.463205337524414, 'timing/data': 0.026105107739567757, 'timing/model': 1.2184590669348836, 'learning_rate/qwen_vl_interface': 6.303058333682464e-06, 'learning_rate/action_model': 6.303058333682463e-05, 'epoch': 1.12}) +[2026-06-17 22:35:58,296][starVLA.training.train_starvla][INFO] - Step 3060, Loss: {'train/loss': 0.6994925737380981, 'train/grad_norm_pre_clip': 21.158353805541992, 'timing/data': 0.024288262240588665, 'timing/model': 1.0677678836509585, 'learning_rate/qwen_vl_interface': 6.300949351010542e-06, 'learning_rate/action_model': 6.300949351010543e-05, 'epoch': 1.12}) +[2026-06-17 22:36:06,522][starVLA.training.train_starvla][INFO] - Step 3061, Loss: {'train/loss': 0.3456001877784729, 'train/grad_norm_pre_clip': 12.020514488220215, 'timing/data': 0.02219767216593027, 'timing/model': 1.0797617649659514, 'learning_rate/qwen_vl_interface': 6.298840150475689e-06, 'learning_rate/action_model': 6.298840150475689e-05, 'epoch': 1.12}) +[2026-06-17 22:36:12,907][starVLA.training.train_starvla][INFO] - Step 3062, Loss: {'train/loss': 0.5925280451774597, 'train/grad_norm_pre_clip': 20.63367462158203, 'timing/data': 0.02287157904356718, 'timing/model': 3.0936384201049805, 'learning_rate/qwen_vl_interface': 6.296730732515145e-06, 'learning_rate/action_model': 6.296730732515145e-05, 'epoch': 1.12}) +[2026-06-17 22:36:20,690][starVLA.training.train_starvla][INFO] - Step 3063, Loss: {'train/loss': 0.7694657444953918, 'train/grad_norm_pre_clip': 14.759214401245117, 'timing/data': 0.07719355542212725, 'timing/model': 2.671022211201489, 'learning_rate/qwen_vl_interface': 6.294621097566193e-06, 'learning_rate/action_model': 6.294621097566193e-05, 'epoch': 1.12}) +[2026-06-17 22:36:25,248][starVLA.training.train_starvla][INFO] - Step 3064, Loss: {'train/loss': 0.5921026468276978, 'train/grad_norm_pre_clip': 14.416603088378906, 'timing/data': 0.02319438476115465, 'timing/model': 1.167291048914194, 'learning_rate/qwen_vl_interface': 6.292511246066163e-06, 'learning_rate/action_model': 6.292511246066163e-05, 'epoch': 1.12}) +[2026-06-17 22:36:29,758][starVLA.training.train_starvla][INFO] - Step 3065, Loss: {'train/loss': 0.9207272529602051, 'train/grad_norm_pre_clip': 18.143321990966797, 'timing/data': 0.02964052651077509, 'timing/model': 1.3816687110811472, 'learning_rate/qwen_vl_interface': 6.290401178452431e-06, 'learning_rate/action_model': 6.29040117845243e-05, 'epoch': 1.12}) +[2026-06-17 22:36:34,224][starVLA.training.train_starvla][INFO] - Step 3066, Loss: {'train/loss': 0.550514280796051, 'train/grad_norm_pre_clip': 18.525726318359375, 'timing/data': 0.02670516725629568, 'timing/model': 1.506465413607657, 'learning_rate/qwen_vl_interface': 6.288290895162414e-06, 'learning_rate/action_model': 6.288290895162413e-05, 'epoch': 1.12}) +[2026-06-17 22:36:42,890][starVLA.training.train_starvla][INFO] - Step 3067, Loss: {'train/loss': 0.4828496277332306, 'train/grad_norm_pre_clip': 13.181059837341309, 'timing/data': 0.02282528020441532, 'timing/model': 1.0384812820702791, 'learning_rate/qwen_vl_interface': 6.286180396633579e-06, 'learning_rate/action_model': 6.286180396633579e-05, 'epoch': 1.12}) +[2026-06-17 22:36:47,443][starVLA.training.train_starvla][INFO] - Step 3068, Loss: {'train/loss': 0.5858014822006226, 'train/grad_norm_pre_clip': 16.766517639160156, 'timing/data': 0.030528713017702103, 'timing/model': 1.1037069521844387, 'learning_rate/qwen_vl_interface': 6.28406968330343e-06, 'learning_rate/action_model': 6.28406968330343e-05, 'epoch': 1.12}) +[2026-06-17 22:36:52,334][starVLA.training.train_starvla][INFO] - Step 3069, Loss: {'train/loss': 0.46020346879959106, 'train/grad_norm_pre_clip': 13.25261116027832, 'timing/data': 0.035999737679958344, 'timing/model': 1.2639538794755936, 'learning_rate/qwen_vl_interface': 6.281958755609526e-06, 'learning_rate/action_model': 6.281958755609526e-05, 'epoch': 1.12}) +[2026-06-17 22:37:00,141][starVLA.training.train_starvla][INFO] - Step 3070, Loss: {'train/loss': 0.5754808783531189, 'train/grad_norm_pre_clip': 15.349861145019531, 'timing/data': 0.029986985959112644, 'timing/model': 4.265564536675811, 'learning_rate/qwen_vl_interface': 6.279847613989459e-06, 'learning_rate/action_model': 6.279847613989459e-05, 'epoch': 1.12}) +[2026-06-17 22:37:04,580][starVLA.training.train_starvla][INFO] - Step 3071, Loss: {'train/loss': 0.7027800679206848, 'train/grad_norm_pre_clip': 16.495607376098633, 'timing/data': 0.02414050977677107, 'timing/model': 1.384264400228858, 'learning_rate/qwen_vl_interface': 6.2777362588808745e-06, 'learning_rate/action_model': 6.277736258880874e-05, 'epoch': 1.12}) +[2026-06-17 22:37:09,009][starVLA.training.train_starvla][INFO] - Step 3072, Loss: {'train/loss': 0.5265960097312927, 'train/grad_norm_pre_clip': 17.287731170654297, 'timing/data': 0.024120019748806953, 'timing/model': 1.089804783463478, 'learning_rate/qwen_vl_interface': 6.275624690721458e-06, 'learning_rate/action_model': 6.275624690721458e-05, 'epoch': 1.12}) +[2026-06-17 22:37:16,142][starVLA.training.train_starvla][INFO] - Step 3073, Loss: {'train/loss': 0.4023045301437378, 'train/grad_norm_pre_clip': 24.432193756103516, 'timing/data': 0.020061563700437546, 'timing/model': 1.1519878432154655, 'learning_rate/qwen_vl_interface': 6.27351290994894e-06, 'learning_rate/action_model': 6.27351290994894e-05, 'epoch': 1.13}) +[2026-06-17 22:37:20,469][starVLA.training.train_starvla][INFO] - Step 3074, Loss: {'train/loss': 0.7488346099853516, 'train/grad_norm_pre_clip': 19.928274154663086, 'timing/data': 0.025418567471206188, 'timing/model': 1.0355484876781702, 'learning_rate/qwen_vl_interface': 6.271400917001092e-06, 'learning_rate/action_model': 6.271400917001092e-05, 'epoch': 1.13}) +[2026-06-17 22:37:25,199][starVLA.training.train_starvla][INFO] - Step 3075, Loss: {'train/loss': 0.5019956231117249, 'train/grad_norm_pre_clip': 16.971582412719727, 'timing/data': 0.027739674784243107, 'timing/model': 1.1075633307918906, 'learning_rate/qwen_vl_interface': 6.269288712315737e-06, 'learning_rate/action_model': 6.269288712315737e-05, 'epoch': 1.13}) +[2026-06-17 22:37:29,810][starVLA.training.train_starvla][INFO] - Step 3076, Loss: {'train/loss': 0.3533986508846283, 'train/grad_norm_pre_clip': 13.74056339263916, 'timing/data': 0.023708435706794262, 'timing/model': 1.4296655114740133, 'learning_rate/qwen_vl_interface': 6.267176296330734e-06, 'learning_rate/action_model': 6.267176296330733e-05, 'epoch': 1.13}) +[2026-06-17 22:37:34,064][starVLA.training.train_starvla][INFO] - Step 3077, Loss: {'train/loss': 0.5827744603157043, 'train/grad_norm_pre_clip': 17.276994705200195, 'timing/data': 0.02937955502420664, 'timing/model': 1.0934757366776466, 'learning_rate/qwen_vl_interface': 6.265063669483992e-06, 'learning_rate/action_model': 6.265063669483991e-05, 'epoch': 1.13}) +[2026-06-17 22:37:38,762][starVLA.training.train_starvla][INFO] - Step 3078, Loss: {'train/loss': 0.6514115929603577, 'train/grad_norm_pre_clip': 16.69872283935547, 'timing/data': 0.026126031763851643, 'timing/model': 1.0477098673582077, 'learning_rate/qwen_vl_interface': 6.262950832213457e-06, 'learning_rate/action_model': 6.262950832213456e-05, 'epoch': 1.13}) +[2026-06-17 22:37:48,137][starVLA.training.train_starvla][INFO] - Step 3079, Loss: {'train/loss': 0.57918381690979, 'train/grad_norm_pre_clip': 17.823097229003906, 'timing/data': 0.02247515693306923, 'timing/model': 1.2466887682676315, 'learning_rate/qwen_vl_interface': 6.260837784957125e-06, 'learning_rate/action_model': 6.260837784957126e-05, 'epoch': 1.13}) +[2026-06-17 22:37:52,563][starVLA.training.train_starvla][INFO] - Step 3080, Loss: {'train/loss': 0.5521952509880066, 'train/grad_norm_pre_clip': 17.57973289489746, 'timing/data': 0.02421177737414837, 'timing/model': 1.1112942527979612, 'learning_rate/qwen_vl_interface': 6.258724528153034e-06, 'learning_rate/action_model': 6.258724528153033e-05, 'epoch': 1.13}) +[2026-06-17 22:37:57,138][starVLA.training.train_starvla][INFO] - Step 3081, Loss: {'train/loss': 0.7128149271011353, 'train/grad_norm_pre_clip': 16.157329559326172, 'timing/data': 0.052956415340304375, 'timing/model': 1.3871705550700426, 'learning_rate/qwen_vl_interface': 6.256611062239261e-06, 'learning_rate/action_model': 6.25661106223926e-05, 'epoch': 1.13}) +[2026-06-17 22:38:01,290][starVLA.training.train_starvla][INFO] - Step 3082, Loss: {'train/loss': 0.5049956440925598, 'train/grad_norm_pre_clip': 13.36157512664795, 'timing/data': 0.046932551078498363, 'timing/model': 1.1280593257397413, 'learning_rate/qwen_vl_interface': 6.254497387653933e-06, 'learning_rate/action_model': 6.254497387653932e-05, 'epoch': 1.13}) +[2026-06-17 22:38:13,211][starVLA.training.train_starvla][INFO] - Step 3083, Loss: {'train/loss': 0.6276820302009583, 'train/grad_norm_pre_clip': 15.359198570251465, 'timing/data': 0.09684831742197275, 'timing/model': 1.1890751477330923, 'learning_rate/qwen_vl_interface': 6.252383504835213e-06, 'learning_rate/action_model': 6.252383504835213e-05, 'epoch': 1.13}) +[2026-06-17 22:38:17,777][starVLA.training.train_starvla][INFO] - Step 3084, Loss: {'train/loss': 0.4408547878265381, 'train/grad_norm_pre_clip': 13.511553764343262, 'timing/data': 0.04831751994788647, 'timing/model': 1.0539601482450962, 'learning_rate/qwen_vl_interface': 6.250269414221317e-06, 'learning_rate/action_model': 6.250269414221317e-05, 'epoch': 1.13}) +[2026-06-17 22:38:22,354][starVLA.training.train_starvla][INFO] - Step 3085, Loss: {'train/loss': 0.5185653567314148, 'train/grad_norm_pre_clip': 16.19390296936035, 'timing/data': 0.036000944674015045, 'timing/model': 1.0621359534561634, 'learning_rate/qwen_vl_interface': 6.248155116250494e-06, 'learning_rate/action_model': 6.248155116250493e-05, 'epoch': 1.13}) +[2026-06-17 22:38:28,307][starVLA.training.train_starvla][INFO] - Step 3086, Loss: {'train/loss': 0.6118680238723755, 'train/grad_norm_pre_clip': 14.981924057006836, 'timing/data': 0.04303555376827717, 'timing/model': 2.9802495734766126, 'learning_rate/qwen_vl_interface': 6.246040611361042e-06, 'learning_rate/action_model': 6.246040611361041e-05, 'epoch': 1.13}) +[2026-06-17 22:38:36,851][starVLA.training.train_starvla][INFO] - Step 3087, Loss: {'train/loss': 0.3495462238788605, 'train/grad_norm_pre_clip': 11.17869758605957, 'timing/data': 0.05402623489499092, 'timing/model': 3.724920094013214, 'learning_rate/qwen_vl_interface': 6.2439258999912976e-06, 'learning_rate/action_model': 6.243925899991298e-05, 'epoch': 1.13}) +[2026-06-17 22:38:41,591][starVLA.training.train_starvla][INFO] - Step 3088, Loss: {'train/loss': 0.6983776092529297, 'train/grad_norm_pre_clip': 13.631771087646484, 'timing/data': 0.03966070804744959, 'timing/model': 1.032935448922217, 'learning_rate/qwen_vl_interface': 6.241810982579646e-06, 'learning_rate/action_model': 6.241810982579646e-05, 'epoch': 1.13}) +[2026-06-17 22:38:46,169][starVLA.training.train_starvla][INFO] - Step 3089, Loss: {'train/loss': 0.5773516893386841, 'train/grad_norm_pre_clip': 18.104747772216797, 'timing/data': 0.05432417243719101, 'timing/model': 1.059489275328815, 'learning_rate/qwen_vl_interface': 6.239695859564511e-06, 'learning_rate/action_model': 6.23969585956451e-05, 'epoch': 1.13}) +[2026-06-17 22:38:53,468][starVLA.training.train_starvla][INFO] - Step 3090, Loss: {'train/loss': 0.8228683471679688, 'train/grad_norm_pre_clip': 19.85236358642578, 'timing/data': 0.04701662342995405, 'timing/model': 1.551210168749094, 'learning_rate/qwen_vl_interface': 6.2375805313843564e-06, 'learning_rate/action_model': 6.237580531384356e-05, 'epoch': 1.13}) +[2026-06-17 22:38:57,942][starVLA.training.train_starvla][INFO] - Step 3091, Loss: {'train/loss': 0.6264362931251526, 'train/grad_norm_pre_clip': 15.805412292480469, 'timing/data': 0.04801123309880495, 'timing/model': 1.2525983620435, 'learning_rate/qwen_vl_interface': 6.2354649984777e-06, 'learning_rate/action_model': 6.235464998477699e-05, 'epoch': 1.13}) +[2026-06-17 22:39:02,412][starVLA.training.train_starvla][INFO] - Step 3092, Loss: {'train/loss': 0.9672355055809021, 'train/grad_norm_pre_clip': 19.785646438598633, 'timing/data': 0.040852660313248634, 'timing/model': 1.066697628237307, 'learning_rate/qwen_vl_interface': 6.2333492612830835e-06, 'learning_rate/action_model': 6.233349261283083e-05, 'epoch': 1.13}) +[2026-06-17 22:39:07,042][starVLA.training.train_starvla][INFO] - Step 3093, Loss: {'train/loss': 0.55184406042099, 'train/grad_norm_pre_clip': 15.647940635681152, 'timing/data': 0.03920835815370083, 'timing/model': 1.2888966733589768, 'learning_rate/qwen_vl_interface': 6.231233320239112e-06, 'learning_rate/action_model': 6.231233320239112e-05, 'epoch': 1.13}) +[2026-06-17 22:39:11,807][starVLA.training.train_starvla][INFO] - Step 3094, Loss: {'train/loss': 0.540513813495636, 'train/grad_norm_pre_clip': 17.179025650024414, 'timing/data': 0.05396994575858116, 'timing/model': 1.1914940224960446, 'learning_rate/qwen_vl_interface': 6.229117175784414e-06, 'learning_rate/action_model': 6.229117175784413e-05, 'epoch': 1.13}) +[2026-06-17 22:39:18,107][starVLA.training.train_starvla][INFO] - Step 3095, Loss: {'train/loss': 0.4651830494403839, 'train/grad_norm_pre_clip': 13.961999893188477, 'timing/data': 0.03858288377523422, 'timing/model': 1.4723570756614208, 'learning_rate/qwen_vl_interface': 6.227000828357672e-06, 'learning_rate/action_model': 6.227000828357672e-05, 'epoch': 1.13}) +[2026-06-17 22:39:28,517][starVLA.training.train_starvla][INFO] - Step 3096, Loss: {'train/loss': 0.4010198712348938, 'train/grad_norm_pre_clip': 14.311802864074707, 'timing/data': 0.0383741594851017, 'timing/model': 4.239410691894591, 'learning_rate/qwen_vl_interface': 6.2248842783976085e-06, 'learning_rate/action_model': 6.224884278397608e-05, 'epoch': 1.13}) +[2026-06-17 22:39:35,299][starVLA.training.train_starvla][INFO] - Step 3097, Loss: {'train/loss': 0.41485658288002014, 'train/grad_norm_pre_clip': 12.348760604858398, 'timing/data': 0.07513744290918112, 'timing/model': 1.1954160341992974, 'learning_rate/qwen_vl_interface': 6.222767526342983e-06, 'learning_rate/action_model': 6.222767526342983e-05, 'epoch': 1.13}) +[2026-06-17 22:39:44,378][starVLA.training.train_starvla][INFO] - Step 3098, Loss: {'train/loss': 0.692888081073761, 'train/grad_norm_pre_clip': 18.179737091064453, 'timing/data': 0.04142704512923956, 'timing/model': 1.0945427222177386, 'learning_rate/qwen_vl_interface': 6.220650572632605e-06, 'learning_rate/action_model': 6.220650572632604e-05, 'epoch': 1.13}) +[2026-06-17 22:39:51,133][starVLA.training.train_starvla][INFO] - Step 3099, Loss: {'train/loss': 0.5584661364555359, 'train/grad_norm_pre_clip': 14.858942031860352, 'timing/data': 0.051552820950746536, 'timing/model': 1.3258872693404555, 'learning_rate/qwen_vl_interface': 6.218533417705318e-06, 'learning_rate/action_model': 6.218533417705318e-05, 'epoch': 1.13}) +[2026-06-17 22:43:32,260][starVLA.training.train_starvla][INFO] - Step 3100, Loss: {'train/loss': 0.5630806684494019, 'train/grad_norm_pre_clip': 15.976326942443848, 'eval/loss': 0.6264866590499878, 'eval/latency_2/loss': 0.6264866590499878, 'eval/demon_fire_precision': 0.9191714053614947, 'eval/demon_fire_recall': 0.9468619246861925, 'eval/demon_fire_f1': 0.9328112118713932, 'eval/demon_fire_rate_model': 0.6850306065664997, 'eval/demon_fire_rate_teacher': 0.6649972175848636, 'eval/demon_left_precision': 0.8878048780487805, 'eval/demon_left_recall': 0.7663157894736842, 'eval/demon_left_f1': 0.8225988700564971, 'eval/demon_left_rate_model': 0.22815804117974403, 'eval/demon_left_rate_teacher': 0.2643294379521425, 'eval/demon_right_precision': 0.8365937859608745, 'eval/demon_right_recall': 0.8214689265536723, 'eval/demon_right_f1': 0.8289623717217788, 'eval/demon_right_rate_model': 0.2417918753478019, 'eval/demon_right_rate_teacher': 0.24624373956594323, 'eval/demon_move_f1': 0.825780620889138, 'eval/demon_cc_f1': 0.8978978690995711, 'eval/latency_2/demon_fire_precision': 0.9191714053614947, 'eval/latency_2/demon_fire_recall': 0.9468619246861925, 'eval/latency_2/demon_fire_f1': 0.9328112118713932, 'eval/latency_2/demon_fire_rate_model': 0.6850306065664997, 'eval/latency_2/demon_fire_rate_teacher': 0.6649972175848636, 'eval/latency_2/demon_left_precision': 0.8878048780487805, 'eval/latency_2/demon_left_recall': 0.7663157894736842, 'eval/latency_2/demon_left_f1': 0.8225988700564971, 'eval/latency_2/demon_left_rate_model': 0.22815804117974403, 'eval/latency_2/demon_left_rate_teacher': 0.2643294379521425, 'eval/latency_2/demon_right_precision': 0.8365937859608745, 'eval/latency_2/demon_right_recall': 0.8214689265536723, 'eval/latency_2/demon_right_f1': 0.8289623717217788, 'eval/latency_2/demon_right_rate_model': 0.2417918753478019, 'eval/latency_2/demon_right_rate_teacher': 0.24624373956594323, 'eval/latency_2/demon_move_f1': 0.825780620889138, 'eval/latency_2/demon_cc_f1': 0.8978978690995711, 'timing/data': 0.04453606717288494, 'timing/model': 1.0390838850289583, 'learning_rate/qwen_vl_interface': 6.216416062000012e-06, 'learning_rate/action_model': 6.216416062000013e-05, 'epoch': 1.14}) +[2026-06-17 22:43:32,281][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_3100_state +[2026-06-17 22:43:32,289][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 22:43:40,934][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_3100_state/model.safetensors +[2026-06-17 22:43:56,137][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_3100_state/optimizer.bin +[2026-06-17 22:43:56,139][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_3100_state/sampler.bin +[2026-06-17 22:43:56,140][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_3100_state/sampler_1.bin +[2026-06-17 22:43:56,142][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_3100_state/random_states_0.pkl +[2026-06-17 22:43:56,149][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3100 +[2026-06-17 22:43:56,150][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 22:43:56,220][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 22:44:04,332][starVLA.training.train_starvla][INFO] - Step 3101, Loss: {'train/loss': 0.5388824939727783, 'train/grad_norm_pre_clip': 17.142257690429688, 'timing/data': 0.0443026227876544, 'timing/model': 4.125720119103789, 'learning_rate/qwen_vl_interface': 6.214298505955615e-06, 'learning_rate/action_model': 6.214298505955614e-05, 'epoch': 1.14}) +[2026-06-17 22:44:14,520][starVLA.training.train_starvla][INFO] - Step 3102, Loss: {'train/loss': 0.46288394927978516, 'train/grad_norm_pre_clip': 19.849308013916016, 'timing/data': 0.011917787604033947, 'timing/model': 1.4917053133249283, 'learning_rate/qwen_vl_interface': 6.2121807500111e-06, 'learning_rate/action_model': 6.212180750011099e-05, 'epoch': 1.14}) +[2026-06-17 22:44:18,810][starVLA.training.train_starvla][INFO] - Step 3103, Loss: {'train/loss': 0.4581824243068695, 'train/grad_norm_pre_clip': 15.024641036987305, 'timing/data': 0.011074203066527843, 'timing/model': 1.1216787956655025, 'learning_rate/qwen_vl_interface': 6.210062794605482e-06, 'learning_rate/action_model': 6.21006279460548e-05, 'epoch': 1.14}) +[2026-06-17 22:44:23,366][starVLA.training.train_starvla][INFO] - Step 3104, Loss: {'train/loss': 0.3437489867210388, 'train/grad_norm_pre_clip': 12.254083633422852, 'timing/data': 0.0154110724106431, 'timing/model': 1.0488037448376417, 'learning_rate/qwen_vl_interface': 6.207944640177812e-06, 'learning_rate/action_model': 6.207944640177812e-05, 'epoch': 1.14}) +[2026-06-17 22:44:30,617][starVLA.training.train_starvla][INFO] - Step 3105, Loss: {'train/loss': 0.28716933727264404, 'train/grad_norm_pre_clip': 10.72336196899414, 'timing/data': 0.022162403911352158, 'timing/model': 2.514582305215299, 'learning_rate/qwen_vl_interface': 6.205826287167188e-06, 'learning_rate/action_model': 6.205826287167188e-05, 'epoch': 1.14}) +[2026-06-17 22:44:35,152][starVLA.training.train_starvla][INFO] - Step 3106, Loss: {'train/loss': 0.6771818399429321, 'train/grad_norm_pre_clip': 19.061567306518555, 'timing/data': 0.026800213381648064, 'timing/model': 1.051687995903194, 'learning_rate/qwen_vl_interface': 6.203707736012745e-06, 'learning_rate/action_model': 6.203707736012745e-05, 'epoch': 1.14}) +[2026-06-17 22:44:39,624][starVLA.training.train_starvla][INFO] - Step 3107, Loss: {'train/loss': 0.45922043919563293, 'train/grad_norm_pre_clip': 17.083908081054688, 'timing/data': 0.021987012587487698, 'timing/model': 1.4126791041344404, 'learning_rate/qwen_vl_interface': 6.201588987153665e-06, 'learning_rate/action_model': 6.201588987153664e-05, 'epoch': 1.14}) +[2026-06-17 22:44:43,915][starVLA.training.train_starvla][INFO] - Step 3108, Loss: {'train/loss': 0.5642573237419128, 'train/grad_norm_pre_clip': 13.888755798339844, 'timing/data': 0.021927569061517715, 'timing/model': 1.0449425345286727, 'learning_rate/qwen_vl_interface': 6.19947004102916e-06, 'learning_rate/action_model': 6.199470041029159e-05, 'epoch': 1.14}) +[2026-06-17 22:44:48,651][starVLA.training.train_starvla][INFO] - Step 3109, Loss: {'train/loss': 0.30243271589279175, 'train/grad_norm_pre_clip': 28.554203033447266, 'timing/data': 0.021651344373822212, 'timing/model': 1.1475448599085212, 'learning_rate/qwen_vl_interface': 6.197350898078496e-06, 'learning_rate/action_model': 6.197350898078495e-05, 'epoch': 1.14}) +[2026-06-17 22:44:53,397][starVLA.training.train_starvla][INFO] - Step 3110, Loss: {'train/loss': 0.47525233030319214, 'train/grad_norm_pre_clip': 14.848752975463867, 'timing/data': 0.024056081660091877, 'timing/model': 1.1274183578789234, 'learning_rate/qwen_vl_interface': 6.195231558740972e-06, 'learning_rate/action_model': 6.195231558740971e-05, 'epoch': 1.14}) +[2026-06-17 22:44:58,080][starVLA.training.train_starvla][INFO] - Step 3111, Loss: {'train/loss': 0.480920672416687, 'train/grad_norm_pre_clip': 15.856710433959961, 'timing/data': 0.02934631146490574, 'timing/model': 1.1406823322176933, 'learning_rate/qwen_vl_interface': 6.1931120234559275e-06, 'learning_rate/action_model': 6.193112023455927e-05, 'epoch': 1.14}) +[2026-06-17 22:45:02,676][starVLA.training.train_starvla][INFO] - Step 3112, Loss: {'train/loss': 0.5943025946617126, 'train/grad_norm_pre_clip': 19.407760620117188, 'timing/data': 0.02999019343405962, 'timing/model': 1.4208285044878721, 'learning_rate/qwen_vl_interface': 6.190992292662747e-06, 'learning_rate/action_model': 6.190992292662746e-05, 'epoch': 1.14}) +[2026-06-17 22:45:13,520][starVLA.training.train_starvla][INFO] - Step 3113, Loss: {'train/loss': 0.44858211278915405, 'train/grad_norm_pre_clip': 17.083759307861328, 'timing/data': 0.031515891663730145, 'timing/model': 1.0884254220873117, 'learning_rate/qwen_vl_interface': 6.18887236680085e-06, 'learning_rate/action_model': 6.18887236680085e-05, 'epoch': 1.14}) +[2026-06-17 22:45:18,309][starVLA.training.train_starvla][INFO] - Step 3114, Loss: {'train/loss': 0.8685828447341919, 'train/grad_norm_pre_clip': 22.056529998779297, 'timing/data': 0.0199982738122344, 'timing/model': 1.266887622885406, 'learning_rate/qwen_vl_interface': 6.186752246309702e-06, 'learning_rate/action_model': 6.186752246309702e-05, 'epoch': 1.14}) +[2026-06-17 22:45:23,386][starVLA.training.train_starvla][INFO] - Step 3115, Loss: {'train/loss': 0.4620891213417053, 'train/grad_norm_pre_clip': 20.624910354614258, 'timing/data': 0.02386653795838356, 'timing/model': 1.150336392223835, 'learning_rate/qwen_vl_interface': 6.184631931628806e-06, 'learning_rate/action_model': 6.184631931628806e-05, 'epoch': 1.14}) +[2026-06-17 22:45:32,712][starVLA.training.train_starvla][INFO] - Step 3116, Loss: {'train/loss': 0.4794416129589081, 'train/grad_norm_pre_clip': 14.168912887573242, 'timing/data': 0.02668710332363844, 'timing/model': 1.0606270544230938, 'learning_rate/qwen_vl_interface': 6.1825114231977036e-06, 'learning_rate/action_model': 6.182511423197704e-05, 'epoch': 1.14}) +[2026-06-17 22:45:41,948][starVLA.training.train_starvla][INFO] - Step 3117, Loss: {'train/loss': 0.655290961265564, 'train/grad_norm_pre_clip': 22.839366912841797, 'timing/data': 0.023557069711387157, 'timing/model': 1.4637155132368207, 'learning_rate/qwen_vl_interface': 6.18039072145598e-06, 'learning_rate/action_model': 6.18039072145598e-05, 'epoch': 1.14}) +[2026-06-17 22:45:45,972][starVLA.training.train_starvla][INFO] - Step 3118, Loss: {'train/loss': 0.4956165850162506, 'train/grad_norm_pre_clip': 16.060380935668945, 'timing/data': 0.02794536855071783, 'timing/model': 1.0440985672175884, 'learning_rate/qwen_vl_interface': 6.17826982684326e-06, 'learning_rate/action_model': 6.178269826843259e-05, 'epoch': 1.14}) +[2026-06-17 22:45:50,706][starVLA.training.train_starvla][INFO] - Step 3119, Loss: {'train/loss': 0.3584084212779999, 'train/grad_norm_pre_clip': 17.23280906677246, 'timing/data': 0.02053683903068304, 'timing/model': 1.0325285522267222, 'learning_rate/qwen_vl_interface': 6.176148739799205e-06, 'learning_rate/action_model': 6.176148739799205e-05, 'epoch': 1.14}) +[2026-06-17 22:45:55,463][starVLA.training.train_starvla][INFO] - Step 3120, Loss: {'train/loss': 0.7356231212615967, 'train/grad_norm_pre_clip': 17.017627716064453, 'timing/data': 0.01910959742963314, 'timing/model': 1.1629931926727295, 'learning_rate/qwen_vl_interface': 6.1740274607635195e-06, 'learning_rate/action_model': 6.174027460763519e-05, 'epoch': 1.14}) +[2026-06-17 22:46:00,035][starVLA.training.train_starvla][INFO] - Step 3121, Loss: {'train/loss': 0.5988547801971436, 'train/grad_norm_pre_clip': 16.96259307861328, 'timing/data': 0.024006962776184082, 'timing/model': 1.024657528847456, 'learning_rate/qwen_vl_interface': 6.1719059901759474e-06, 'learning_rate/action_model': 6.171905990175947e-05, 'epoch': 1.14}) +[2026-06-17 22:46:04,557][starVLA.training.train_starvla][INFO] - Step 3122, Loss: {'train/loss': 0.3939559757709503, 'train/grad_norm_pre_clip': 15.488035202026367, 'timing/data': 0.02332811802625656, 'timing/model': 1.4260026393458247, 'learning_rate/qwen_vl_interface': 6.169784328476272e-06, 'learning_rate/action_model': 6.169784328476271e-05, 'epoch': 1.14}) +[2026-06-17 22:46:08,578][starVLA.training.train_starvla][INFO] - Step 3123, Loss: {'train/loss': 0.571918785572052, 'train/grad_norm_pre_clip': 19.010574340820312, 'timing/data': 0.02063499204814434, 'timing/model': 1.0377100612968206, 'learning_rate/qwen_vl_interface': 6.167662476104314e-06, 'learning_rate/action_model': 6.167662476104314e-05, 'epoch': 1.14}) +[2026-06-17 22:46:13,300][starVLA.training.train_starvla][INFO] - Step 3124, Loss: {'train/loss': 0.47392329573631287, 'train/grad_norm_pre_clip': 16.889326095581055, 'timing/data': 0.020113850943744183, 'timing/model': 1.0767472451552749, 'learning_rate/qwen_vl_interface': 6.165540433499939e-06, 'learning_rate/action_model': 6.165540433499939e-05, 'epoch': 1.14}) +[2026-06-17 22:46:21,034][starVLA.training.train_starvla][INFO] - Step 3125, Loss: {'train/loss': 0.4559544622898102, 'train/grad_norm_pre_clip': 15.070260047912598, 'timing/data': 0.029741795733571053, 'timing/model': 1.023471936583519, 'learning_rate/qwen_vl_interface': 6.163418201103045e-06, 'learning_rate/action_model': 6.163418201103045e-05, 'epoch': 1.14}) +[2026-06-17 22:46:25,497][starVLA.training.train_starvla][INFO] - Step 3126, Loss: {'train/loss': 0.40126219391822815, 'train/grad_norm_pre_clip': 14.210675239562988, 'timing/data': 0.020540980622172356, 'timing/model': 1.1126004438847303, 'learning_rate/qwen_vl_interface': 6.161295779353578e-06, 'learning_rate/action_model': 6.161295779353577e-05, 'epoch': 1.14}) +[2026-06-17 22:46:30,149][starVLA.training.train_starvla][INFO] - Step 3127, Loss: {'train/loss': 0.74793541431427, 'train/grad_norm_pre_clip': 19.667503356933594, 'timing/data': 0.025687430053949356, 'timing/model': 1.4779701810330153, 'learning_rate/qwen_vl_interface': 6.159173168691512e-06, 'learning_rate/action_model': 6.159173168691512e-05, 'epoch': 1.15}) +[2026-06-17 22:46:36,074][starVLA.training.train_starvla][INFO] - Step 3128, Loss: {'train/loss': 0.39591169357299805, 'train/grad_norm_pre_clip': 14.60509204864502, 'timing/data': 0.027894859202206135, 'timing/model': 1.1653217943385243, 'learning_rate/qwen_vl_interface': 6.157050369556871e-06, 'learning_rate/action_model': 6.157050369556872e-05, 'epoch': 1.15}) +[2026-06-17 22:46:40,648][starVLA.training.train_starvla][INFO] - Step 3129, Loss: {'train/loss': 0.42844492197036743, 'train/grad_norm_pre_clip': 18.22929573059082, 'timing/data': 0.0219921562820673, 'timing/model': 1.1063585197553039, 'learning_rate/qwen_vl_interface': 6.154927382389714e-06, 'learning_rate/action_model': 6.154927382389714e-05, 'epoch': 1.15}) +[2026-06-17 22:46:45,406][starVLA.training.train_starvla][INFO] - Step 3130, Loss: {'train/loss': 0.9345154166221619, 'train/grad_norm_pre_clip': 17.841886520385742, 'timing/data': 0.019137246534228325, 'timing/model': 1.049506695009768, 'learning_rate/qwen_vl_interface': 6.152804207630135e-06, 'learning_rate/action_model': 6.152804207630136e-05, 'epoch': 1.15}) +[2026-06-17 22:46:51,465][starVLA.training.train_starvla][INFO] - Step 3131, Loss: {'train/loss': 0.47751447558403015, 'train/grad_norm_pre_clip': 15.822705268859863, 'timing/data': 0.021647906862199306, 'timing/model': 1.02896776702255, 'learning_rate/qwen_vl_interface': 6.150680845718274e-06, 'learning_rate/action_model': 6.150680845718274e-05, 'epoch': 1.15}) +[2026-06-17 22:46:57,464][starVLA.training.train_starvla][INFO] - Step 3132, Loss: {'train/loss': 0.4706560969352722, 'train/grad_norm_pre_clip': 17.420207977294922, 'timing/data': 0.0217109527438879, 'timing/model': 1.5159413600340486, 'learning_rate/qwen_vl_interface': 6.148557297094303e-06, 'learning_rate/action_model': 6.148557297094303e-05, 'epoch': 1.15}) +[2026-06-17 22:47:01,581][starVLA.training.train_starvla][INFO] - Step 3133, Loss: {'train/loss': 0.6370185613632202, 'train/grad_norm_pre_clip': 19.957197189331055, 'timing/data': 0.022048606537282467, 'timing/model': 1.029321382753551, 'learning_rate/qwen_vl_interface': 6.1464335621984375e-06, 'learning_rate/action_model': 6.146433562198437e-05, 'epoch': 1.15}) +[2026-06-17 22:47:06,057][starVLA.training.train_starvla][INFO] - Step 3134, Loss: {'train/loss': 0.5424199104309082, 'train/grad_norm_pre_clip': 14.857008934020996, 'timing/data': 0.025702957063913345, 'timing/model': 1.0248290142044425, 'learning_rate/qwen_vl_interface': 6.14430964147093e-06, 'learning_rate/action_model': 6.14430964147093e-05, 'epoch': 1.15}) +[2026-06-17 22:47:17,519][starVLA.training.train_starvla][INFO] - Step 3135, Loss: {'train/loss': 0.5675703883171082, 'train/grad_norm_pre_clip': 18.412553787231445, 'timing/data': 0.029240595176815987, 'timing/model': 4.270912436768413, 'learning_rate/qwen_vl_interface': 6.1421855353520735e-06, 'learning_rate/action_model': 6.142185535352073e-05, 'epoch': 1.15}) +[2026-06-17 22:47:26,779][starVLA.training.train_starvla][INFO] - Step 3136, Loss: {'train/loss': 0.6484789252281189, 'train/grad_norm_pre_clip': 19.238080978393555, 'timing/data': 0.026609936729073524, 'timing/model': 1.3915052181109786, 'learning_rate/qwen_vl_interface': 6.140061244282194e-06, 'learning_rate/action_model': 6.140061244282194e-05, 'epoch': 1.15}) +[2026-06-17 22:47:31,390][starVLA.training.train_starvla][INFO] - Step 3137, Loss: {'train/loss': 0.484774112701416, 'train/grad_norm_pre_clip': 18.167436599731445, 'timing/data': 0.023908957839012146, 'timing/model': 1.3696494968608022, 'learning_rate/qwen_vl_interface': 6.137936768701659e-06, 'learning_rate/action_model': 6.137936768701659e-05, 'epoch': 1.15}) +[2026-06-17 22:47:35,751][starVLA.training.train_starvla][INFO] - Step 3138, Loss: {'train/loss': 0.41311952471733093, 'train/grad_norm_pre_clip': 16.3126220703125, 'timing/data': 0.021161268465220928, 'timing/model': 1.030532251112163, 'learning_rate/qwen_vl_interface': 6.135812109050879e-06, 'learning_rate/action_model': 6.135812109050879e-05, 'epoch': 1.15}) +[2026-06-17 22:47:40,410][starVLA.training.train_starvla][INFO] - Step 3139, Loss: {'train/loss': 0.15295356512069702, 'train/grad_norm_pre_clip': 7.659632205963135, 'timing/data': 0.025142934173345566, 'timing/model': 1.1357207614928484, 'learning_rate/qwen_vl_interface': 6.1336872657702936e-06, 'learning_rate/action_model': 6.133687265770293e-05, 'epoch': 1.15}) +[2026-06-17 22:47:45,036][starVLA.training.train_starvla][INFO] - Step 3140, Loss: {'train/loss': 0.3456651270389557, 'train/grad_norm_pre_clip': 14.799924850463867, 'timing/data': 0.02346996683627367, 'timing/model': 1.1646205205470324, 'learning_rate/qwen_vl_interface': 6.131562239300389e-06, 'learning_rate/action_model': 6.13156223930039e-05, 'epoch': 1.15}) +[2026-06-17 22:47:50,094][starVLA.training.train_starvla][INFO] - Step 3141, Loss: {'train/loss': 0.6009923219680786, 'train/grad_norm_pre_clip': 20.239879608154297, 'timing/data': 0.029919488355517387, 'timing/model': 1.4023816110566258, 'learning_rate/qwen_vl_interface': 6.129437030081683e-06, 'learning_rate/action_model': 6.129437030081682e-05, 'epoch': 1.15}) +[2026-06-17 22:47:55,004][starVLA.training.train_starvla][INFO] - Step 3142, Loss: {'train/loss': 0.3804464638233185, 'train/grad_norm_pre_clip': 14.827760696411133, 'timing/data': 0.08165296353399754, 'timing/model': 1.5808215299621224, 'learning_rate/qwen_vl_interface': 6.1273116385547336e-06, 'learning_rate/action_model': 6.127311638554734e-05, 'epoch': 1.15}) +[2026-06-17 22:48:02,176][starVLA.training.train_starvla][INFO] - Step 3143, Loss: {'train/loss': 0.7013304829597473, 'train/grad_norm_pre_clip': 16.57203483581543, 'timing/data': 0.001766490750014782, 'timing/model': 1.0511940261349082, 'learning_rate/qwen_vl_interface': 6.125186065160138e-06, 'learning_rate/action_model': 6.125186065160138e-05, 'epoch': 1.15}) +[2026-06-17 22:48:06,574][starVLA.training.train_starvla][INFO] - Step 3144, Loss: {'train/loss': 0.41605326533317566, 'train/grad_norm_pre_clip': 13.774298667907715, 'timing/data': 0.001091943122446537, 'timing/model': 1.0237144278362393, 'learning_rate/qwen_vl_interface': 6.123060310338529e-06, 'learning_rate/action_model': 6.123060310338529e-05, 'epoch': 1.15}) +[2026-06-17 22:48:11,210][starVLA.training.train_starvla][INFO] - Step 3145, Loss: {'train/loss': 0.9320597648620605, 'train/grad_norm_pre_clip': 23.74606704711914, 'timing/data': 0.0010650446638464928, 'timing/model': 1.121693504974246, 'learning_rate/qwen_vl_interface': 6.120934374530579e-06, 'learning_rate/action_model': 6.12093437453058e-05, 'epoch': 1.15}) +[2026-06-17 22:48:18,937][starVLA.training.train_starvla][INFO] - Step 3146, Loss: {'train/loss': 0.38296395540237427, 'train/grad_norm_pre_clip': 16.00931739807129, 'timing/data': 0.0012839604169130325, 'timing/model': 1.0846088416874409, 'learning_rate/qwen_vl_interface': 6.118808258176996e-06, 'learning_rate/action_model': 6.118808258176995e-05, 'epoch': 1.15}) +[2026-06-17 22:48:23,419][starVLA.training.train_starvla][INFO] - Step 3147, Loss: {'train/loss': 0.5827668905258179, 'train/grad_norm_pre_clip': 17.883596420288086, 'timing/data': 0.0015869708731770515, 'timing/model': 1.0574632044881582, 'learning_rate/qwen_vl_interface': 6.116681961718525e-06, 'learning_rate/action_model': 6.116681961718524e-05, 'epoch': 1.15}) +[2026-06-17 22:48:29,005][starVLA.training.train_starvla][INFO] - Step 3148, Loss: {'train/loss': 0.7555450201034546, 'train/grad_norm_pre_clip': 23.189395904541016, 'timing/data': 0.0021606069058179855, 'timing/model': 2.129314647987485, 'learning_rate/qwen_vl_interface': 6.114555485595952e-06, 'learning_rate/action_model': 6.114555485595952e-05, 'epoch': 1.15}) +[2026-06-17 22:48:33,724][starVLA.training.train_starvla][INFO] - Step 3149, Loss: {'train/loss': 0.6569891571998596, 'train/grad_norm_pre_clip': 16.11435317993164, 'timing/data': 0.0012300238013267517, 'timing/model': 1.0350093636661768, 'learning_rate/qwen_vl_interface': 6.112428830250096e-06, 'learning_rate/action_model': 6.112428830250095e-05, 'epoch': 1.15}) +[2026-06-17 22:48:39,995][starVLA.training.train_starvla][INFO] - Step 3150, Loss: {'train/loss': 0.6094474196434021, 'train/grad_norm_pre_clip': 16.039016723632812, 'timing/data': 0.0008834656327962875, 'timing/model': 2.7598013123497367, 'learning_rate/qwen_vl_interface': 6.110301996121816e-06, 'learning_rate/action_model': 6.110301996121816e-05, 'epoch': 1.15}) +[2026-06-17 22:48:49,043][starVLA.training.train_starvla][INFO] - Step 3151, Loss: {'train/loss': 1.0844008922576904, 'train/grad_norm_pre_clip': 21.212995529174805, 'timing/data': 0.0014298269525170326, 'timing/model': 1.0208186879754066, 'learning_rate/qwen_vl_interface': 6.108174983652007e-06, 'learning_rate/action_model': 6.108174983652006e-05, 'epoch': 1.15}) +[2026-06-17 22:48:53,734][starVLA.training.train_starvla][INFO] - Step 3152, Loss: {'train/loss': 0.6486960053443909, 'train/grad_norm_pre_clip': 19.82797622680664, 'timing/data': 0.0008844863623380661, 'timing/model': 1.4769901279360056, 'learning_rate/qwen_vl_interface': 6.106047793281598e-06, 'learning_rate/action_model': 6.106047793281598e-05, 'epoch': 1.15}) +[2026-06-17 22:49:02,915][starVLA.training.train_starvla][INFO] - Step 3153, Loss: {'train/loss': 0.33603692054748535, 'train/grad_norm_pre_clip': 19.130374908447266, 'timing/data': 0.021492784842848778, 'timing/model': 1.0405704332515597, 'learning_rate/qwen_vl_interface': 6.1039204254515626e-06, 'learning_rate/action_model': 6.103920425451563e-05, 'epoch': 1.15}) +[2026-06-17 22:49:07,401][starVLA.training.train_starvla][INFO] - Step 3154, Loss: {'train/loss': 0.49274158477783203, 'train/grad_norm_pre_clip': 19.39247703552246, 'timing/data': 0.023567745462059975, 'timing/model': 1.1542923524975777, 'learning_rate/qwen_vl_interface': 6.101792880602901e-06, 'learning_rate/action_model': 6.101792880602901e-05, 'epoch': 1.15}) +[2026-06-17 22:49:12,068][starVLA.training.train_starvla][INFO] - Step 3155, Loss: {'train/loss': 0.4782717227935791, 'train/grad_norm_pre_clip': 16.230470657348633, 'timing/data': 0.019261419773101807, 'timing/model': 1.1201823959127069, 'learning_rate/qwen_vl_interface': 6.099665159176663e-06, 'learning_rate/action_model': 6.099665159176663e-05, 'epoch': 1.16}) +[2026-06-17 22:49:16,362][starVLA.training.train_starvla][INFO] - Step 3156, Loss: {'train/loss': 0.7052564024925232, 'train/grad_norm_pre_clip': 21.139734268188477, 'timing/data': 0.03162002097815275, 'timing/model': 1.3307528821751475, 'learning_rate/qwen_vl_interface': 6.09753726161392e-06, 'learning_rate/action_model': 6.09753726161392e-05, 'epoch': 1.16}) +[2026-06-17 22:49:20,846][starVLA.training.train_starvla][INFO] - Step 3157, Loss: {'train/loss': 0.7287856936454773, 'train/grad_norm_pre_clip': 19.020246505737305, 'timing/data': 0.02707590628415346, 'timing/model': 1.3252715449780226, 'learning_rate/qwen_vl_interface': 6.095409188355793e-06, 'learning_rate/action_model': 6.095409188355793e-05, 'epoch': 1.16}) +[2026-06-17 22:49:29,450][starVLA.training.train_starvla][INFO] - Step 3158, Loss: {'train/loss': 0.5940098762512207, 'train/grad_norm_pre_clip': 16.907318115234375, 'timing/data': 0.020004410296678543, 'timing/model': 1.1718551786616445, 'learning_rate/qwen_vl_interface': 6.0932809398434315e-06, 'learning_rate/action_model': 6.093280939843431e-05, 'epoch': 1.16}) +[2026-06-17 22:49:35,645][starVLA.training.train_starvla][INFO] - Step 3159, Loss: {'train/loss': 0.4560401141643524, 'train/grad_norm_pre_clip': 14.742195129394531, 'timing/data': 0.020710238255560398, 'timing/model': 2.9132064878940582, 'learning_rate/qwen_vl_interface': 6.091152516518022e-06, 'learning_rate/action_model': 6.0911525165180216e-05, 'epoch': 1.16}) +[2026-06-17 22:49:40,196][starVLA.training.train_starvla][INFO] - Step 3160, Loss: {'train/loss': 0.3493315577507019, 'train/grad_norm_pre_clip': 12.843513488769531, 'timing/data': 0.021124262362718582, 'timing/model': 1.029378592967987, 'learning_rate/qwen_vl_interface': 6.089023918820792e-06, 'learning_rate/action_model': 6.089023918820792e-05, 'epoch': 1.16}) +[2026-06-17 22:49:45,174][starVLA.training.train_starvla][INFO] - Step 3161, Loss: {'train/loss': 0.5284263491630554, 'train/grad_norm_pre_clip': 13.178725242614746, 'timing/data': 0.024088316597044468, 'timing/model': 1.0261227143928409, 'learning_rate/qwen_vl_interface': 6.0868951471929986e-06, 'learning_rate/action_model': 6.086895147192998e-05, 'epoch': 1.16}) +[2026-06-17 22:49:49,428][starVLA.training.train_starvla][INFO] - Step 3162, Loss: {'train/loss': 0.2983705401420593, 'train/grad_norm_pre_clip': 10.98969554901123, 'timing/data': 0.01928630843758583, 'timing/model': 1.3136651068925858, 'learning_rate/qwen_vl_interface': 6.084766202075944e-06, 'learning_rate/action_model': 6.0847662020759435e-05, 'epoch': 1.16}) +[2026-06-17 22:49:53,718][starVLA.training.train_starvla][INFO] - Step 3163, Loss: {'train/loss': 0.5606175661087036, 'train/grad_norm_pre_clip': 14.69177532196045, 'timing/data': 0.023119980469346046, 'timing/model': 1.1342283505946398, 'learning_rate/qwen_vl_interface': 6.082637083910955e-06, 'learning_rate/action_model': 6.0826370839109546e-05, 'epoch': 1.16}) +[2026-06-17 22:50:01,489][starVLA.training.train_starvla][INFO] - Step 3164, Loss: {'train/loss': 0.7704529762268066, 'train/grad_norm_pre_clip': 19.036640167236328, 'timing/data': 0.019833771511912346, 'timing/model': 4.39356857072562, 'learning_rate/qwen_vl_interface': 6.080507793139403e-06, 'learning_rate/action_model': 6.080507793139403e-05, 'epoch': 1.16}) +[2026-06-17 22:50:10,451][starVLA.training.train_starvla][INFO] - Step 3165, Loss: {'train/loss': 0.5985482335090637, 'train/grad_norm_pre_clip': 15.511833190917969, 'timing/data': 0.021988393738865852, 'timing/model': 1.038786068558693, 'learning_rate/qwen_vl_interface': 6.0783783302026925e-06, 'learning_rate/action_model': 6.078378330202692e-05, 'epoch': 1.16}) +[2026-06-17 22:50:14,907][starVLA.training.train_starvla][INFO] - Step 3166, Loss: {'train/loss': 0.5855821371078491, 'train/grad_norm_pre_clip': 37.69247055053711, 'timing/data': 0.02074639592319727, 'timing/model': 1.0317027689889073, 'learning_rate/qwen_vl_interface': 6.076248695542262e-06, 'learning_rate/action_model': 6.076248695542261e-05, 'epoch': 1.16}) +[2026-06-17 22:50:30,954][starVLA.training.train_starvla][INFO] - Step 3167, Loss: {'train/loss': 0.45698124170303345, 'train/grad_norm_pre_clip': 13.96932601928711, 'timing/data': 0.026851474307477474, 'timing/model': 1.4110571406781673, 'learning_rate/qwen_vl_interface': 6.074118889599589e-06, 'learning_rate/action_model': 6.0741188895995884e-05, 'epoch': 1.16}) +[2026-06-17 22:50:35,052][starVLA.training.train_starvla][INFO] - Step 3168, Loss: {'train/loss': 1.12228262424469, 'train/grad_norm_pre_clip': 24.054779052734375, 'timing/data': 0.029485195875167847, 'timing/model': 1.0175811806693673, 'learning_rate/qwen_vl_interface': 6.07198891281618e-06, 'learning_rate/action_model': 6.07198891281618e-05, 'epoch': 1.16}) +[2026-06-17 22:50:39,507][starVLA.training.train_starvla][INFO] - Step 3169, Loss: {'train/loss': 0.6506813764572144, 'train/grad_norm_pre_clip': 22.110027313232422, 'timing/data': 0.019381479360163212, 'timing/model': 1.0476300306618214, 'learning_rate/qwen_vl_interface': 6.0698587656335895e-06, 'learning_rate/action_model': 6.069858765633589e-05, 'epoch': 1.16}) +[2026-06-17 22:50:44,060][starVLA.training.train_starvla][INFO] - Step 3170, Loss: {'train/loss': 0.8006299138069153, 'train/grad_norm_pre_clip': 18.182300567626953, 'timing/data': 0.021307080052793026, 'timing/model': 1.0746571188792586, 'learning_rate/qwen_vl_interface': 6.067728448493392e-06, 'learning_rate/action_model': 6.067728448493392e-05, 'epoch': 1.16}) +[2026-06-17 22:50:50,004][starVLA.training.train_starvla][INFO] - Step 3171, Loss: {'train/loss': 0.513185441493988, 'train/grad_norm_pre_clip': 14.848925590515137, 'timing/data': 0.019932988099753857, 'timing/model': 1.1641681231558323, 'learning_rate/qwen_vl_interface': 6.065597961837209e-06, 'learning_rate/action_model': 6.0655979618372086e-05, 'epoch': 1.16}) +[2026-06-17 22:50:54,695][starVLA.training.train_starvla][INFO] - Step 3172, Loss: {'train/loss': 0.6275551319122314, 'train/grad_norm_pre_clip': 16.329851150512695, 'timing/data': 0.03631971310824156, 'timing/model': 1.5244188122451305, 'learning_rate/qwen_vl_interface': 6.063467306106688e-06, 'learning_rate/action_model': 6.0634673061066885e-05, 'epoch': 1.16}) +[2026-06-17 22:50:58,871][starVLA.training.train_starvla][INFO] - Step 3173, Loss: {'train/loss': 0.7003187537193298, 'train/grad_norm_pre_clip': 20.48796844482422, 'timing/data': 0.02074693702161312, 'timing/model': 1.096051693893969, 'learning_rate/qwen_vl_interface': 6.061336481743522e-06, 'learning_rate/action_model': 6.061336481743522e-05, 'epoch': 1.16}) +[2026-06-17 22:51:03,435][starVLA.training.train_starvla][INFO] - Step 3174, Loss: {'train/loss': 1.1408882141113281, 'train/grad_norm_pre_clip': 21.2700138092041, 'timing/data': 0.025236343033611774, 'timing/model': 1.12566839158535, 'learning_rate/qwen_vl_interface': 6.059205489189429e-06, 'learning_rate/action_model': 6.059205489189429e-05, 'epoch': 1.16}) +[2026-06-17 22:51:11,682][starVLA.training.train_starvla][INFO] - Step 3175, Loss: {'train/loss': 0.5067525506019592, 'train/grad_norm_pre_clip': 14.560285568237305, 'timing/data': 0.020054589956998825, 'timing/model': 1.0363542810082436, 'learning_rate/qwen_vl_interface': 6.057074328886168e-06, 'learning_rate/action_model': 6.057074328886168e-05, 'epoch': 1.16}) +[2026-06-17 22:51:16,198][starVLA.training.train_starvla][INFO] - Step 3176, Loss: {'train/loss': 0.36010125279426575, 'train/grad_norm_pre_clip': 15.935891151428223, 'timing/data': 0.02127815317362547, 'timing/model': 1.0794410929083824, 'learning_rate/qwen_vl_interface': 6.0549430012755295e-06, 'learning_rate/action_model': 6.0549430012755295e-05, 'epoch': 1.16}) +[2026-06-17 22:51:20,778][starVLA.training.train_starvla][INFO] - Step 3177, Loss: {'train/loss': 0.40499669313430786, 'train/grad_norm_pre_clip': 13.535755157470703, 'timing/data': 0.022898451425135136, 'timing/model': 1.427015796303749, 'learning_rate/qwen_vl_interface': 6.052811506799341e-06, 'learning_rate/action_model': 6.052811506799341e-05, 'epoch': 1.16}) +[2026-06-17 22:51:25,207][starVLA.training.train_starvla][INFO] - Step 3178, Loss: {'train/loss': 0.5915132164955139, 'train/grad_norm_pre_clip': 18.61801528930664, 'timing/data': 0.029964103363454342, 'timing/model': 1.2833118690177798, 'learning_rate/qwen_vl_interface': 6.050679845899462e-06, 'learning_rate/action_model': 6.050679845899462e-05, 'epoch': 1.16}) +[2026-06-17 22:51:30,005][starVLA.training.train_starvla][INFO] - Step 3179, Loss: {'train/loss': 0.2998247444629669, 'train/grad_norm_pre_clip': 11.021466255187988, 'timing/data': 0.028939751908183098, 'timing/model': 1.0879424074664712, 'learning_rate/qwen_vl_interface': 6.048548019017791e-06, 'learning_rate/action_model': 6.04854801901779e-05, 'epoch': 1.16}) +[2026-06-17 22:51:34,588][starVLA.training.train_starvla][INFO] - Step 3180, Loss: {'train/loss': 0.753976047039032, 'train/grad_norm_pre_clip': 14.762541770935059, 'timing/data': 0.01880253665149212, 'timing/model': 1.038621542043984, 'learning_rate/qwen_vl_interface': 6.046416026596256e-06, 'learning_rate/action_model': 6.046416026596255e-05, 'epoch': 1.16}) +[2026-06-17 22:51:41,076][starVLA.training.train_starvla][INFO] - Step 3181, Loss: {'train/loss': 0.7824872732162476, 'train/grad_norm_pre_clip': 25.778127670288086, 'timing/data': 0.020986788906157017, 'timing/model': 3.099513624794781, 'learning_rate/qwen_vl_interface': 6.044283869076822e-06, 'learning_rate/action_model': 6.0442838690768224e-05, 'epoch': 1.16}) +[2026-06-17 22:51:45,548][starVLA.training.train_starvla][INFO] - Step 3182, Loss: {'train/loss': 0.8089306950569153, 'train/grad_norm_pre_clip': 16.55045509338379, 'timing/data': 0.017995750531554222, 'timing/model': 1.3973658625036478, 'learning_rate/qwen_vl_interface': 6.042151546901487e-06, 'learning_rate/action_model': 6.042151546901487e-05, 'epoch': 1.17}) +[2026-06-17 22:51:49,698][starVLA.training.train_starvla][INFO] - Step 3183, Loss: {'train/loss': 0.5636452436447144, 'train/grad_norm_pre_clip': 18.21397590637207, 'timing/data': 0.022320969961583614, 'timing/model': 1.1145860338583589, 'learning_rate/qwen_vl_interface': 6.0400190605122846e-06, 'learning_rate/action_model': 6.040019060512284e-05, 'epoch': 1.17}) +[2026-06-17 22:52:08,711][starVLA.training.train_starvla][INFO] - Step 3184, Loss: {'train/loss': 0.42780593037605286, 'train/grad_norm_pre_clip': 14.951765060424805, 'timing/data': 0.026073982939124107, 'timing/model': 4.336240606382489, 'learning_rate/qwen_vl_interface': 6.0378864103512805e-06, 'learning_rate/action_model': 6.03788641035128e-05, 'epoch': 1.17}) +[2026-06-17 22:52:13,173][starVLA.training.train_starvla][INFO] - Step 3185, Loss: {'train/loss': 0.40253767371177673, 'train/grad_norm_pre_clip': 13.247786521911621, 'timing/data': 0.010833101347088814, 'timing/model': 1.0213090423494577, 'learning_rate/qwen_vl_interface': 6.035753596860573e-06, 'learning_rate/action_model': 6.0357535968605725e-05, 'epoch': 1.17}) +[2026-06-17 22:52:17,951][starVLA.training.train_starvla][INFO] - Step 3186, Loss: {'train/loss': 0.49915486574172974, 'train/grad_norm_pre_clip': 13.036362648010254, 'timing/data': 0.011407668702304363, 'timing/model': 1.3972613606601954, 'learning_rate/qwen_vl_interface': 6.033620620482303e-06, 'learning_rate/action_model': 6.033620620482303e-05, 'epoch': 1.17}) +[2026-06-17 22:52:22,551][starVLA.training.train_starvla][INFO] - Step 3187, Loss: {'train/loss': 0.5407986640930176, 'train/grad_norm_pre_clip': 15.064386367797852, 'timing/data': 0.012157159857451916, 'timing/model': 1.4775192383676767, 'learning_rate/qwen_vl_interface': 6.031487481658633e-06, 'learning_rate/action_model': 6.031487481658633e-05, 'epoch': 1.17}) +[2026-06-17 22:52:26,695][starVLA.training.train_starvla][INFO] - Step 3188, Loss: {'train/loss': 0.6206449270248413, 'train/grad_norm_pre_clip': 13.704535484313965, 'timing/data': 0.01495693251490593, 'timing/model': 1.132917563430965, 'learning_rate/qwen_vl_interface': 6.029354180831769e-06, 'learning_rate/action_model': 6.029354180831769e-05, 'epoch': 1.17}) +[2026-06-17 22:52:31,213][starVLA.training.train_starvla][INFO] - Step 3189, Loss: {'train/loss': 0.26995179057121277, 'train/grad_norm_pre_clip': 11.01807975769043, 'timing/data': 0.012094486504793167, 'timing/model': 1.047312025912106, 'learning_rate/qwen_vl_interface': 6.027220718443943e-06, 'learning_rate/action_model': 6.027220718443943e-05, 'epoch': 1.17}) +[2026-06-17 22:52:37,692][starVLA.training.train_starvla][INFO] - Step 3190, Loss: {'train/loss': 0.41792333126068115, 'train/grad_norm_pre_clip': 13.822032928466797, 'timing/data': 0.011231952346861362, 'timing/model': 1.0376585442572832, 'learning_rate/qwen_vl_interface': 6.025087094937426e-06, 'learning_rate/action_model': 6.025087094937426e-05, 'epoch': 1.17}) +[2026-06-17 22:52:42,187][starVLA.training.train_starvla][INFO] - Step 3191, Loss: {'train/loss': 0.374561607837677, 'train/grad_norm_pre_clip': 15.062804222106934, 'timing/data': 0.01153582613915205, 'timing/model': 1.0444393707439303, 'learning_rate/qwen_vl_interface': 6.022953310754521e-06, 'learning_rate/action_model': 6.022953310754521e-05, 'epoch': 1.17}) +[2026-06-17 22:52:46,723][starVLA.training.train_starvla][INFO] - Step 3192, Loss: {'train/loss': 0.5344120264053345, 'train/grad_norm_pre_clip': 14.640666007995605, 'timing/data': 0.010656544007360935, 'timing/model': 1.3805195530876517, 'learning_rate/qwen_vl_interface': 6.020819366337562e-06, 'learning_rate/action_model': 6.020819366337562e-05, 'epoch': 1.17}) +[2026-06-17 22:52:50,994][starVLA.training.train_starvla][INFO] - Step 3193, Loss: {'train/loss': 0.5660908818244934, 'train/grad_norm_pre_clip': 16.776973724365234, 'timing/data': 0.01688381191343069, 'timing/model': 1.0531175090000033, 'learning_rate/qwen_vl_interface': 6.018685262128921e-06, 'learning_rate/action_model': 6.01868526212892e-05, 'epoch': 1.17}) +[2026-06-17 22:52:55,549][starVLA.training.train_starvla][INFO] - Step 3194, Loss: {'train/loss': 0.389303982257843, 'train/grad_norm_pre_clip': 17.771652221679688, 'timing/data': 0.017175694927573204, 'timing/model': 1.1155744725838304, 'learning_rate/qwen_vl_interface': 6.016550998570995e-06, 'learning_rate/action_model': 6.0165509985709944e-05, 'epoch': 1.17}) +[2026-06-17 22:53:00,022][starVLA.training.train_starvla][INFO] - Step 3195, Loss: {'train/loss': 0.7289713621139526, 'train/grad_norm_pre_clip': 19.180383682250977, 'timing/data': 0.01126968301832676, 'timing/model': 1.0804587425664067, 'learning_rate/qwen_vl_interface': 6.014416576106224e-06, 'learning_rate/action_model': 6.0144165761062244e-05, 'epoch': 1.17}) +[2026-06-17 22:53:04,584][starVLA.training.train_starvla][INFO] - Step 3196, Loss: {'train/loss': 0.6878077983856201, 'train/grad_norm_pre_clip': 19.390954971313477, 'timing/data': 0.011562151834368706, 'timing/model': 1.5688910773023963, 'learning_rate/qwen_vl_interface': 6.012281995177074e-06, 'learning_rate/action_model': 6.012281995177074e-05, 'epoch': 1.17}) +[2026-06-17 22:53:16,762][starVLA.training.train_starvla][INFO] - Step 3197, Loss: {'train/loss': 0.5421746373176575, 'train/grad_norm_pre_clip': 19.957534790039062, 'timing/data': 0.013318550772964954, 'timing/model': 1.3131342083215714, 'learning_rate/qwen_vl_interface': 6.010147256226046e-06, 'learning_rate/action_model': 6.0101472562260454e-05, 'epoch': 1.17}) +[2026-06-17 22:53:21,433][starVLA.training.train_starvla][INFO] - Step 3198, Loss: {'train/loss': 0.6796087622642517, 'train/grad_norm_pre_clip': 23.44610595703125, 'timing/data': 0.0104738874360919, 'timing/model': 1.2581213219091296, 'learning_rate/qwen_vl_interface': 6.008012359695673e-06, 'learning_rate/action_model': 6.0080123596956726e-05, 'epoch': 1.17}) +[2026-06-17 22:53:30,928][starVLA.training.train_starvla][INFO] - Step 3199, Loss: {'train/loss': 0.642992377281189, 'train/grad_norm_pre_clip': 17.280282974243164, 'timing/data': 0.013238389976322651, 'timing/model': 4.466116786003113, 'learning_rate/qwen_vl_interface': 6.005877306028522e-06, 'learning_rate/action_model': 6.005877306028522e-05, 'epoch': 1.17}) +[2026-06-17 22:58:22,287][starVLA.training.train_starvla][INFO] - Step 3200, Loss: {'train/loss': 0.5583533048629761, 'train/grad_norm_pre_clip': 17.45632553100586, 'eval/loss': 0.6327798366546631, 'eval/latency_2/loss': 0.6327798366546631, 'eval/demon_fire_precision': 0.9101391650099404, 'eval/demon_fire_recall': 0.9577405857740586, 'eval/demon_fire_f1': 0.9333333333333335, 'eval/demon_fire_rate_model': 0.699777406789093, 'eval/demon_fire_rate_teacher': 0.6649972175848636, 'eval/demon_left_precision': 0.8292929292929293, 'eval/demon_left_recall': 0.8642105263157894, 'eval/demon_left_f1': 0.8463917525773196, 'eval/demon_left_rate_model': 0.27545909849749584, 'eval/demon_left_rate_teacher': 0.2643294379521425, 'eval/demon_right_precision': 0.8426966292134831, 'eval/demon_right_recall': 0.847457627118644, 'eval/demon_right_f1': 0.8450704225352111, 'eval/demon_right_rate_model': 0.2476349471341124, 'eval/demon_right_rate_teacher': 0.24624373956594323, 'eval/demon_move_f1': 0.8457310875562654, 'eval/demon_cc_f1': 0.9052045512383418, 'eval/latency_2/demon_fire_precision': 0.9101391650099404, 'eval/latency_2/demon_fire_recall': 0.9577405857740586, 'eval/latency_2/demon_fire_f1': 0.9333333333333335, 'eval/latency_2/demon_fire_rate_model': 0.699777406789093, 'eval/latency_2/demon_fire_rate_teacher': 0.6649972175848636, 'eval/latency_2/demon_left_precision': 0.8292929292929293, 'eval/latency_2/demon_left_recall': 0.8642105263157894, 'eval/latency_2/demon_left_f1': 0.8463917525773196, 'eval/latency_2/demon_left_rate_model': 0.27545909849749584, 'eval/latency_2/demon_left_rate_teacher': 0.2643294379521425, 'eval/latency_2/demon_right_precision': 0.8426966292134831, 'eval/latency_2/demon_right_recall': 0.847457627118644, 'eval/latency_2/demon_right_f1': 0.8450704225352111, 'eval/latency_2/demon_right_rate_model': 0.2476349471341124, 'eval/latency_2/demon_right_rate_teacher': 0.24624373956594323, 'eval/latency_2/demon_move_f1': 0.8457310875562654, 'eval/latency_2/demon_cc_f1': 0.9052045512383418, 'timing/data': 0.011866386979818344, 'timing/model': 1.1173683051019907, 'learning_rate/qwen_vl_interface': 6.003742095667194e-06, 'learning_rate/action_model': 6.003742095667194e-05, 'epoch': 1.17}) +[2026-06-17 22:58:22,316][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_3200_state +[2026-06-17 22:58:22,324][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 22:58:31,842][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_3200_state/model.safetensors +[2026-06-17 22:58:48,980][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_3200_state/optimizer.bin +[2026-06-17 22:58:49,005][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_3200_state/sampler.bin +[2026-06-17 22:58:49,006][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_3200_state/sampler_1.bin +[2026-06-17 22:58:49,008][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_3200_state/random_states_0.pkl +[2026-06-17 22:58:49,013][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3200 +[2026-06-17 22:58:49,013][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 22:58:49,030][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 22:58:53,400][starVLA.training.train_starvla][INFO] - Step 3201, Loss: {'train/loss': 0.7441793084144592, 'train/grad_norm_pre_clip': 17.08009910583496, 'timing/data': 0.009650602005422115, 'timing/model': 1.0263072224333882, 'learning_rate/qwen_vl_interface': 6.001606729054318e-06, 'learning_rate/action_model': 6.0016067290543176e-05, 'epoch': 1.17}) +[2026-06-17 22:58:57,621][starVLA.training.train_starvla][INFO] - Step 3202, Loss: {'train/loss': 0.4969231188297272, 'train/grad_norm_pre_clip': 15.494768142700195, 'timing/data': 0.010675719939172268, 'timing/model': 1.0422540828585625, 'learning_rate/qwen_vl_interface': 5.999471206632559e-06, 'learning_rate/action_model': 5.999471206632559e-05, 'epoch': 1.17}) +[2026-06-17 22:59:01,856][starVLA.training.train_starvla][INFO] - Step 3203, Loss: {'train/loss': 0.5982502102851868, 'train/grad_norm_pre_clip': 15.723908424377441, 'timing/data': 0.009925726801156998, 'timing/model': 1.2917960742488503, 'learning_rate/qwen_vl_interface': 5.99733552884461e-06, 'learning_rate/action_model': 5.99733552884461e-05, 'epoch': 1.17}) +[2026-06-17 22:59:05,770][starVLA.training.train_starvla][INFO] - Step 3204, Loss: {'train/loss': 0.7072315812110901, 'train/grad_norm_pre_clip': 16.651891708374023, 'timing/data': 0.010775795206427574, 'timing/model': 1.0092693101614714, 'learning_rate/qwen_vl_interface': 5.995199696133204e-06, 'learning_rate/action_model': 5.9951996961332034e-05, 'epoch': 1.17}) +[2026-06-17 22:59:09,862][starVLA.training.train_starvla][INFO] - Step 3205, Loss: {'train/loss': 0.5076422095298767, 'train/grad_norm_pre_clip': 15.861350059509277, 'timing/data': 0.01039995439350605, 'timing/model': 1.0005747023969889, 'learning_rate/qwen_vl_interface': 5.993063708941098e-06, 'learning_rate/action_model': 5.993063708941098e-05, 'epoch': 1.17}) +[2026-06-17 22:59:14,016][starVLA.training.train_starvla][INFO] - Step 3206, Loss: {'train/loss': 0.44009774923324585, 'train/grad_norm_pre_clip': 12.561025619506836, 'timing/data': 0.008909294381737709, 'timing/model': 1.0240452131256461, 'learning_rate/qwen_vl_interface': 5.990927567711085e-06, 'learning_rate/action_model': 5.990927567711085e-05, 'epoch': 1.17}) +[2026-06-17 22:59:18,183][starVLA.training.train_starvla][INFO] - Step 3207, Loss: {'train/loss': 0.4559776186943054, 'train/grad_norm_pre_clip': 10.867635726928711, 'timing/data': 0.014876734465360641, 'timing/model': 1.0462241936475039, 'learning_rate/qwen_vl_interface': 5.988791272885991e-06, 'learning_rate/action_model': 5.988791272885991e-05, 'epoch': 1.17}) +[2026-06-17 22:59:22,104][starVLA.training.train_starvla][INFO] - Step 3208, Loss: {'train/loss': 0.42793476581573486, 'train/grad_norm_pre_clip': 14.586437225341797, 'timing/data': 0.008960852399468422, 'timing/model': 1.008150665089488, 'learning_rate/qwen_vl_interface': 5.986654824908671e-06, 'learning_rate/action_model': 5.9866548249086706e-05, 'epoch': 1.17}) +[2026-06-17 22:59:26,224][starVLA.training.train_starvla][INFO] - Step 3209, Loss: {'train/loss': 0.5254927277565002, 'train/grad_norm_pre_clip': 14.895214080810547, 'timing/data': 0.009230582043528557, 'timing/model': 1.0134410075843334, 'learning_rate/qwen_vl_interface': 5.9845182242220125e-06, 'learning_rate/action_model': 5.984518224222012e-05, 'epoch': 1.18}) +[2026-06-17 22:59:30,416][starVLA.training.train_starvla][INFO] - Step 3210, Loss: {'train/loss': 0.5027150511741638, 'train/grad_norm_pre_clip': 14.993549346923828, 'timing/data': 0.00977540947496891, 'timing/model': 1.0307641373947263, 'learning_rate/qwen_vl_interface': 5.982381471268935e-06, 'learning_rate/action_model': 5.9823814712689354e-05, 'epoch': 1.18}) +[2026-06-17 22:59:34,730][starVLA.training.train_starvla][INFO] - Step 3211, Loss: {'train/loss': 0.4754248857498169, 'train/grad_norm_pre_clip': 13.871349334716797, 'timing/data': 0.009057939052581787, 'timing/model': 0.9969900557771325, 'learning_rate/qwen_vl_interface': 5.980244566492392e-06, 'learning_rate/action_model': 5.9802445664923925e-05, 'epoch': 1.18}) +[2026-06-17 22:59:38,820][starVLA.training.train_starvla][INFO] - Step 3212, Loss: {'train/loss': 0.10794448852539062, 'train/grad_norm_pre_clip': 6.078107833862305, 'timing/data': 0.009866585023701191, 'timing/model': 1.0001484490931034, 'learning_rate/qwen_vl_interface': 5.978107510335365e-06, 'learning_rate/action_model': 5.9781075103353647e-05, 'epoch': 1.18}) +[2026-06-17 22:59:42,925][starVLA.training.train_starvla][INFO] - Step 3213, Loss: {'train/loss': 0.7383609414100647, 'train/grad_norm_pre_clip': 17.553564071655273, 'timing/data': 0.011033598333597183, 'timing/model': 1.2097672829404473, 'learning_rate/qwen_vl_interface': 5.975970303240867e-06, 'learning_rate/action_model': 5.9759703032408664e-05, 'epoch': 1.18}) +[2026-06-17 22:59:46,819][starVLA.training.train_starvla][INFO] - Step 3214, Loss: {'train/loss': 0.7258782386779785, 'train/grad_norm_pre_clip': 18.984817504882812, 'timing/data': 0.009944445453584194, 'timing/model': 0.9952434962615371, 'learning_rate/qwen_vl_interface': 5.973832945651944e-06, 'learning_rate/action_model': 5.9738329456519435e-05, 'epoch': 1.18}) +[2026-06-17 22:59:55,203][starVLA.training.train_starvla][INFO] - Step 3215, Loss: {'train/loss': 0.6547484993934631, 'train/grad_norm_pre_clip': 15.771550178527832, 'timing/data': 0.011609806679189205, 'timing/model': 1.7035542167723179, 'learning_rate/qwen_vl_interface': 5.971695438011672e-06, 'learning_rate/action_model': 5.971695438011672e-05, 'epoch': 1.18}) +[2026-06-17 22:59:59,750][starVLA.training.train_starvla][INFO] - Step 3216, Loss: {'train/loss': 0.613247811794281, 'train/grad_norm_pre_clip': 22.673748016357422, 'timing/data': 0.01098898146301508, 'timing/model': 1.034834809601307, 'learning_rate/qwen_vl_interface': 5.969557780763159e-06, 'learning_rate/action_model': 5.9695577807631585e-05, 'epoch': 1.18}) +[2026-06-17 23:00:04,552][starVLA.training.train_starvla][INFO] - Step 3217, Loss: {'train/loss': 0.8638338446617126, 'train/grad_norm_pre_clip': 19.274667739868164, 'timing/data': 0.016930624842643738, 'timing/model': 1.611475022509694, 'learning_rate/qwen_vl_interface': 5.9674199743495456e-06, 'learning_rate/action_model': 5.967419974349545e-05, 'epoch': 1.18}) +[2026-06-17 23:00:09,810][starVLA.training.train_starvla][INFO] - Step 3218, Loss: {'train/loss': 0.6241284012794495, 'train/grad_norm_pre_clip': 19.31573486328125, 'timing/data': 0.011487063020467758, 'timing/model': 1.467358778230846, 'learning_rate/qwen_vl_interface': 5.965282019213996e-06, 'learning_rate/action_model': 5.965282019213996e-05, 'epoch': 1.18}) +[2026-06-17 23:00:14,032][starVLA.training.train_starvla][INFO] - Step 3219, Loss: {'train/loss': 0.298794150352478, 'train/grad_norm_pre_clip': 13.498981475830078, 'timing/data': 0.012707050889730453, 'timing/model': 1.096004898659885, 'learning_rate/qwen_vl_interface': 5.963143915799718e-06, 'learning_rate/action_model': 5.9631439157997184e-05, 'epoch': 1.18}) +[2026-06-17 23:00:19,049][starVLA.training.train_starvla][INFO] - Step 3220, Loss: {'train/loss': 1.0663294792175293, 'train/grad_norm_pre_clip': 25.878610610961914, 'timing/data': 0.01745221670717001, 'timing/model': 1.190860048867762, 'learning_rate/qwen_vl_interface': 5.961005664549936e-06, 'learning_rate/action_model': 5.9610056645499354e-05, 'epoch': 1.18}) +[2026-06-17 23:00:23,764][starVLA.training.train_starvla][INFO] - Step 3221, Loss: {'train/loss': 0.5881226658821106, 'train/grad_norm_pre_clip': 16.635995864868164, 'timing/data': 0.01148984581232071, 'timing/model': 1.1073834029957652, 'learning_rate/qwen_vl_interface': 5.958867265907916e-06, 'learning_rate/action_model': 5.958867265907916e-05, 'epoch': 1.18}) +[2026-06-17 23:00:28,916][starVLA.training.train_starvla][INFO] - Step 3222, Loss: {'train/loss': 0.5387926697731018, 'train/grad_norm_pre_clip': 19.04364013671875, 'timing/data': 0.018379121087491512, 'timing/model': 1.3489128146320581, 'learning_rate/qwen_vl_interface': 5.956728720316949e-06, 'learning_rate/action_model': 5.9567287203169485e-05, 'epoch': 1.18}) +[2026-06-17 23:00:36,680][starVLA.training.train_starvla][INFO] - Step 3223, Loss: {'train/loss': 0.4544821083545685, 'train/grad_norm_pre_clip': 16.43285369873047, 'timing/data': 0.012838060036301613, 'timing/model': 1.5646896716207266, 'learning_rate/qwen_vl_interface': 5.954590028220357e-06, 'learning_rate/action_model': 5.954590028220357e-05, 'epoch': 1.18}) +[2026-06-17 23:00:43,147][starVLA.training.train_starvla][INFO] - Step 3224, Loss: {'train/loss': 0.5931910872459412, 'train/grad_norm_pre_clip': 18.88329315185547, 'timing/data': 0.013137249276041985, 'timing/model': 2.590847388841212, 'learning_rate/qwen_vl_interface': 5.952451190061496e-06, 'learning_rate/action_model': 5.9524511900614956e-05, 'epoch': 1.18}) +[2026-06-17 23:00:48,885][starVLA.training.train_starvla][INFO] - Step 3225, Loss: {'train/loss': 0.7503785490989685, 'train/grad_norm_pre_clip': 27.549028396606445, 'timing/data': 0.010635661892592907, 'timing/model': 1.0331440092995763, 'learning_rate/qwen_vl_interface': 5.950312206283743e-06, 'learning_rate/action_model': 5.9503122062837434e-05, 'epoch': 1.18}) +[2026-06-17 23:00:54,058][starVLA.training.train_starvla][INFO] - Step 3226, Loss: {'train/loss': 0.6182565689086914, 'train/grad_norm_pre_clip': 15.405502319335938, 'timing/data': 0.014273503795266151, 'timing/model': 1.3644593795761466, 'learning_rate/qwen_vl_interface': 5.948173077330522e-06, 'learning_rate/action_model': 5.948173077330522e-05, 'epoch': 1.18}) +[2026-06-17 23:01:00,726][starVLA.training.train_starvla][INFO] - Step 3227, Loss: {'train/loss': 0.3722259998321533, 'train/grad_norm_pre_clip': 19.48956871032715, 'timing/data': 0.014814251102507114, 'timing/model': 3.201167678460479, 'learning_rate/qwen_vl_interface': 5.9460338036452665e-06, 'learning_rate/action_model': 5.946033803645267e-05, 'epoch': 1.18}) +[2026-06-17 23:01:05,598][starVLA.training.train_starvla][INFO] - Step 3228, Loss: {'train/loss': 0.3708324432373047, 'train/grad_norm_pre_clip': 11.427240371704102, 'timing/data': 0.012867127545177937, 'timing/model': 1.6587770162150264, 'learning_rate/qwen_vl_interface': 5.943894385671456e-06, 'learning_rate/action_model': 5.9438943856714556e-05, 'epoch': 1.18}) +[2026-06-17 23:01:09,867][starVLA.training.train_starvla][INFO] - Step 3229, Loss: {'train/loss': 0.6127832531929016, 'train/grad_norm_pre_clip': 18.83441734313965, 'timing/data': 0.014029675163328648, 'timing/model': 1.0345308277755976, 'learning_rate/qwen_vl_interface': 5.941754823852593e-06, 'learning_rate/action_model': 5.9417548238525924e-05, 'epoch': 1.18}) +[2026-06-17 23:01:14,976][starVLA.training.train_starvla][INFO] - Step 3230, Loss: {'train/loss': 0.6286095380783081, 'train/grad_norm_pre_clip': 17.68895149230957, 'timing/data': 0.011357513256371021, 'timing/model': 1.3866348331794143, 'learning_rate/qwen_vl_interface': 5.939615118632208e-06, 'learning_rate/action_model': 5.939615118632208e-05, 'epoch': 1.18}) +[2026-06-17 23:01:22,856][starVLA.training.train_starvla][INFO] - Step 3231, Loss: {'train/loss': 0.6710104942321777, 'train/grad_norm_pre_clip': 18.532217025756836, 'timing/data': 0.01594411861151457, 'timing/model': 1.2599739963188767, 'learning_rate/qwen_vl_interface': 5.937475270453872e-06, 'learning_rate/action_model': 5.937475270453872e-05, 'epoch': 1.18}) +[2026-06-17 23:01:29,012][starVLA.training.train_starvla][INFO] - Step 3232, Loss: {'train/loss': 0.49344369769096375, 'train/grad_norm_pre_clip': 26.9930477142334, 'timing/data': 0.012283368967473507, 'timing/model': 1.0570373656228185, 'learning_rate/qwen_vl_interface': 5.935335279761169e-06, 'learning_rate/action_model': 5.935335279761168e-05, 'epoch': 1.18}) +[2026-06-17 23:01:33,759][starVLA.training.train_starvla][INFO] - Step 3233, Loss: {'train/loss': 0.22397762537002563, 'train/grad_norm_pre_clip': 9.864301681518555, 'timing/data': 0.055183819495141506, 'timing/model': 1.5562557308003306, 'learning_rate/qwen_vl_interface': 5.933195146997725e-06, 'learning_rate/action_model': 5.9331951469977254e-05, 'epoch': 1.18}) +[2026-06-17 23:01:37,895][starVLA.training.train_starvla][INFO] - Step 3234, Loss: {'train/loss': 0.6713993549346924, 'train/grad_norm_pre_clip': 23.39638328552246, 'timing/data': 0.011930418200790882, 'timing/model': 1.061770910397172, 'learning_rate/qwen_vl_interface': 5.931054872607193e-06, 'learning_rate/action_model': 5.9310548726071925e-05, 'epoch': 1.18}) +[2026-06-17 23:01:42,453][starVLA.training.train_starvla][INFO] - Step 3235, Loss: {'train/loss': 0.3521542251110077, 'train/grad_norm_pre_clip': 16.224071502685547, 'timing/data': 0.01704496331512928, 'timing/model': 1.07449047639966, 'learning_rate/qwen_vl_interface': 5.928914457033252e-06, 'learning_rate/action_model': 5.928914457033252e-05, 'epoch': 1.18}) +[2026-06-17 23:01:47,166][starVLA.training.train_starvla][INFO] - Step 3236, Loss: {'train/loss': 0.39164263010025024, 'train/grad_norm_pre_clip': 16.378660202026367, 'timing/data': 0.011567715555429459, 'timing/model': 1.0580923883244395, 'learning_rate/qwen_vl_interface': 5.926773900719616e-06, 'learning_rate/action_model': 5.9267739007196157e-05, 'epoch': 1.18}) +[2026-06-17 23:02:03,296][starVLA.training.train_starvla][INFO] - Step 3237, Loss: {'train/loss': 0.8668370842933655, 'train/grad_norm_pre_clip': 24.066612243652344, 'timing/data': 0.014152556657791138, 'timing/model': 1.4928267821669579, 'learning_rate/qwen_vl_interface': 5.92463320411002e-06, 'learning_rate/action_model': 5.92463320411002e-05, 'epoch': 1.19}) +[2026-06-17 23:02:08,132][starVLA.training.train_starvla][INFO] - Step 3238, Loss: {'train/loss': 0.9078736305236816, 'train/grad_norm_pre_clip': 19.98682403564453, 'timing/data': 0.015243102796375751, 'timing/model': 1.4677330320701003, 'learning_rate/qwen_vl_interface': 5.922492367648236e-06, 'learning_rate/action_model': 5.922492367648236e-05, 'epoch': 1.19}) +[2026-06-17 23:02:12,899][starVLA.training.train_starvla][INFO] - Step 3239, Loss: {'train/loss': 0.6380669474601746, 'train/grad_norm_pre_clip': 27.826513290405273, 'timing/data': 0.014701307751238346, 'timing/model': 1.3080647783353925, 'learning_rate/qwen_vl_interface': 5.920351391778059e-06, 'learning_rate/action_model': 5.920351391778059e-05, 'epoch': 1.19}) +[2026-06-17 23:02:23,129][starVLA.training.train_starvla][INFO] - Step 3240, Loss: {'train/loss': 0.7701876759529114, 'train/grad_norm_pre_clip': 20.796531677246094, 'timing/data': 0.011713413521647453, 'timing/model': 1.4108532201498747, 'learning_rate/qwen_vl_interface': 5.918210276943318e-06, 'learning_rate/action_model': 5.918210276943318e-05, 'epoch': 1.19}) +[2026-06-17 23:02:27,996][starVLA.training.train_starvla][INFO] - Step 3241, Loss: {'train/loss': 0.2817462384700775, 'train/grad_norm_pre_clip': 15.513673782348633, 'timing/data': 0.01586237922310829, 'timing/model': 1.215901905670762, 'learning_rate/qwen_vl_interface': 5.916069023587867e-06, 'learning_rate/action_model': 5.916069023587867e-05, 'epoch': 1.19}) +[2026-06-17 23:02:36,649][starVLA.training.train_starvla][INFO] - Step 3242, Loss: {'train/loss': 0.40273478627204895, 'train/grad_norm_pre_clip': 13.335831642150879, 'timing/data': 0.014293987303972244, 'timing/model': 1.133409266360104, 'learning_rate/qwen_vl_interface': 5.913927632155592e-06, 'learning_rate/action_model': 5.913927632155591e-05, 'epoch': 1.19}) +[2026-06-17 23:02:41,529][starVLA.training.train_starvla][INFO] - Step 3243, Loss: {'train/loss': 0.41954129934310913, 'train/grad_norm_pre_clip': 11.599919319152832, 'timing/data': 0.011541243642568588, 'timing/model': 1.56856332346797, 'learning_rate/qwen_vl_interface': 5.911786103090402e-06, 'learning_rate/action_model': 5.911786103090402e-05, 'epoch': 1.19}) +[2026-06-17 23:02:46,117][starVLA.training.train_starvla][INFO] - Step 3244, Loss: {'train/loss': 0.9082580804824829, 'train/grad_norm_pre_clip': 22.010272979736328, 'timing/data': 0.01105425599962473, 'timing/model': 1.2772246124222875, 'learning_rate/qwen_vl_interface': 5.909644436836243e-06, 'learning_rate/action_model': 5.909644436836244e-05, 'epoch': 1.19}) +[2026-06-17 23:02:54,092][starVLA.training.train_starvla][INFO] - Step 3245, Loss: {'train/loss': 0.7127939462661743, 'train/grad_norm_pre_clip': 22.614303588867188, 'timing/data': 0.01458288636058569, 'timing/model': 1.3488864684477448, 'learning_rate/qwen_vl_interface': 5.907502633837082e-06, 'learning_rate/action_model': 5.907502633837082e-05, 'epoch': 1.19}) +[2026-06-17 23:02:59,509][starVLA.training.train_starvla][INFO] - Step 3246, Loss: {'train/loss': 0.27604609727859497, 'train/grad_norm_pre_clip': 15.612374305725098, 'timing/data': 0.01947695668786764, 'timing/model': 1.5343615412712097, 'learning_rate/qwen_vl_interface': 5.905360694536918e-06, 'learning_rate/action_model': 5.905360694536918e-05, 'epoch': 1.19}) +[2026-06-17 23:03:05,979][starVLA.training.train_starvla][INFO] - Step 3247, Loss: {'train/loss': 0.8656244277954102, 'train/grad_norm_pre_clip': 21.680810928344727, 'timing/data': 0.014808714389801025, 'timing/model': 1.054988894611597, 'learning_rate/qwen_vl_interface': 5.903218619379776e-06, 'learning_rate/action_model': 5.903218619379776e-05, 'epoch': 1.19}) +[2026-06-17 23:03:10,545][starVLA.training.train_starvla][INFO] - Step 3248, Loss: {'train/loss': 0.3394598960876465, 'train/grad_norm_pre_clip': 13.825610160827637, 'timing/data': 0.014050329104065895, 'timing/model': 1.4027059627696872, 'learning_rate/qwen_vl_interface': 5.901076408809712e-06, 'learning_rate/action_model': 5.9010764088097114e-05, 'epoch': 1.19}) +[2026-06-17 23:03:14,818][starVLA.training.train_starvla][INFO] - Step 3249, Loss: {'train/loss': 0.5259475111961365, 'train/grad_norm_pre_clip': 16.61214828491211, 'timing/data': 0.010814135894179344, 'timing/model': 1.079108165577054, 'learning_rate/qwen_vl_interface': 5.898934063270806e-06, 'learning_rate/action_model': 5.8989340632708056e-05, 'epoch': 1.19}) +[2026-06-17 23:03:19,717][starVLA.training.train_starvla][INFO] - Step 3250, Loss: {'train/loss': 0.5990010499954224, 'train/grad_norm_pre_clip': 19.989078521728516, 'timing/data': 0.017122001387178898, 'timing/model': 1.225400649011135, 'learning_rate/qwen_vl_interface': 5.896791583207173e-06, 'learning_rate/action_model': 5.896791583207173e-05, 'epoch': 1.19}) +[2026-06-17 23:03:24,155][starVLA.training.train_starvla][INFO] - Step 3251, Loss: {'train/loss': 0.6808560490608215, 'train/grad_norm_pre_clip': 21.155109405517578, 'timing/data': 0.012935910373926163, 'timing/model': 1.0336830392479897, 'learning_rate/qwen_vl_interface': 5.894648969062947e-06, 'learning_rate/action_model': 5.894648969062947e-05, 'epoch': 1.19}) +[2026-06-17 23:03:28,885][starVLA.training.train_starvla][INFO] - Step 3252, Loss: {'train/loss': 0.28907933831214905, 'train/grad_norm_pre_clip': 14.241174697875977, 'timing/data': 0.014846310019493103, 'timing/model': 1.310725630261004, 'learning_rate/qwen_vl_interface': 5.892506221282297e-06, 'learning_rate/action_model': 5.892506221282297e-05, 'epoch': 1.19}) +[2026-06-17 23:03:35,769][starVLA.training.train_starvla][INFO] - Step 3253, Loss: {'train/loss': 0.5284655094146729, 'train/grad_norm_pre_clip': 21.291330337524414, 'timing/data': 0.0111062778159976, 'timing/model': 1.567281031049788, 'learning_rate/qwen_vl_interface': 5.890363340309415e-06, 'learning_rate/action_model': 5.890363340309415e-05, 'epoch': 1.19}) +[2026-06-17 23:03:40,353][starVLA.training.train_starvla][INFO] - Step 3254, Loss: {'train/loss': 0.24358895421028137, 'train/grad_norm_pre_clip': 12.166003227233887, 'timing/data': 0.012466141022741795, 'timing/model': 1.1542251678183675, 'learning_rate/qwen_vl_interface': 5.888220326588523e-06, 'learning_rate/action_model': 5.8882203265885226e-05, 'epoch': 1.19}) +[2026-06-17 23:03:46,622][starVLA.training.train_starvla][INFO] - Step 3255, Loss: {'train/loss': 0.6163684725761414, 'train/grad_norm_pre_clip': 16.924833297729492, 'timing/data': 0.01599812973290682, 'timing/model': 2.4715039897710085, 'learning_rate/qwen_vl_interface': 5.8860771805638725e-06, 'learning_rate/action_model': 5.886077180563872e-05, 'epoch': 1.19}) +[2026-06-17 23:03:57,202][starVLA.training.train_starvla][INFO] - Step 3256, Loss: {'train/loss': 0.5985082983970642, 'train/grad_norm_pre_clip': 16.181806564331055, 'timing/data': 0.011987692676484585, 'timing/model': 1.0267706457525492, 'learning_rate/qwen_vl_interface': 5.883933902679736e-06, 'learning_rate/action_model': 5.8839339026797355e-05, 'epoch': 1.19}) +[2026-06-17 23:04:02,117][starVLA.training.train_starvla][INFO] - Step 3257, Loss: {'train/loss': 0.5862788558006287, 'train/grad_norm_pre_clip': 15.652250289916992, 'timing/data': 0.01173233613371849, 'timing/model': 1.5307281073182821, 'learning_rate/qwen_vl_interface': 5.8817904933804214e-06, 'learning_rate/action_model': 5.881790493380421e-05, 'epoch': 1.19}) +[2026-06-17 23:04:06,512][starVLA.training.train_starvla][INFO] - Step 3258, Loss: {'train/loss': 0.7029001712799072, 'train/grad_norm_pre_clip': 30.983684539794922, 'timing/data': 0.011266211047768593, 'timing/model': 1.3200580645352602, 'learning_rate/qwen_vl_interface': 5.8796469531102565e-06, 'learning_rate/action_model': 5.879646953110256e-05, 'epoch': 1.19}) +[2026-06-17 23:04:10,856][starVLA.training.train_starvla][INFO] - Step 3259, Loss: {'train/loss': 0.40292099118232727, 'train/grad_norm_pre_clip': 14.866521835327148, 'timing/data': 0.01127446349710226, 'timing/model': 1.0160287097096443, 'learning_rate/qwen_vl_interface': 5.877503282313601e-06, 'learning_rate/action_model': 5.877503282313601e-05, 'epoch': 1.19}) +[2026-06-17 23:04:15,290][starVLA.training.train_starvla][INFO] - Step 3260, Loss: {'train/loss': 0.8519325852394104, 'train/grad_norm_pre_clip': 20.832672119140625, 'timing/data': 0.01305785309523344, 'timing/model': 1.0429583573713899, 'learning_rate/qwen_vl_interface': 5.875359481434839e-06, 'learning_rate/action_model': 5.875359481434839e-05, 'epoch': 1.19}) +[2026-06-17 23:04:19,781][starVLA.training.train_starvla][INFO] - Step 3261, Loss: {'train/loss': 0.4994131028652191, 'train/grad_norm_pre_clip': 32.7510871887207, 'timing/data': 0.011824600398540497, 'timing/model': 1.0247490843757987, 'learning_rate/qwen_vl_interface': 5.873215550918384e-06, 'learning_rate/action_model': 5.873215550918384e-05, 'epoch': 1.19}) +[2026-06-17 23:04:24,235][starVLA.training.train_starvla][INFO] - Step 3262, Loss: {'train/loss': 0.6208748817443848, 'train/grad_norm_pre_clip': 16.01344871520996, 'timing/data': 0.010490340180695057, 'timing/model': 1.0540993576869369, 'learning_rate/qwen_vl_interface': 5.871071491208676e-06, 'learning_rate/action_model': 5.871071491208676e-05, 'epoch': 1.19}) +[2026-06-17 23:04:28,880][starVLA.training.train_starvla][INFO] - Step 3263, Loss: {'train/loss': 0.7234771251678467, 'train/grad_norm_pre_clip': 19.24954605102539, 'timing/data': 0.011375057511031628, 'timing/model': 1.4748644530773163, 'learning_rate/qwen_vl_interface': 5.868927302750178e-06, 'learning_rate/action_model': 5.868927302750178e-05, 'epoch': 1.19}) +[2026-06-17 23:04:33,300][starVLA.training.train_starvla][INFO] - Step 3264, Loss: {'train/loss': 0.4746739864349365, 'train/grad_norm_pre_clip': 16.012426376342773, 'timing/data': 0.01197268906980753, 'timing/model': 1.2021340169012547, 'learning_rate/qwen_vl_interface': 5.866782985987386e-06, 'learning_rate/action_model': 5.866782985987386e-05, 'epoch': 1.2}) +[2026-06-17 23:04:38,000][starVLA.training.train_starvla][INFO] - Step 3265, Loss: {'train/loss': 0.30112630128860474, 'train/grad_norm_pre_clip': 14.488099098205566, 'timing/data': 0.019276877865195274, 'timing/model': 1.1185786835849285, 'learning_rate/qwen_vl_interface': 5.864638541364815e-06, 'learning_rate/action_model': 5.8646385413648154e-05, 'epoch': 1.2}) +[2026-06-17 23:04:42,616][starVLA.training.train_starvla][INFO] - Step 3266, Loss: {'train/loss': 0.8344038724899292, 'train/grad_norm_pre_clip': 19.114763259887695, 'timing/data': 0.010751288384199142, 'timing/model': 1.091576125472784, 'learning_rate/qwen_vl_interface': 5.862493969327015e-06, 'learning_rate/action_model': 5.862493969327014e-05, 'epoch': 1.2}) +[2026-06-17 23:04:56,041][starVLA.training.train_starvla][INFO] - Step 3267, Loss: {'train/loss': 0.4290443956851959, 'train/grad_norm_pre_clip': 14.709161758422852, 'timing/data': 0.01774885132908821, 'timing/model': 4.57309082057327, 'learning_rate/qwen_vl_interface': 5.860349270318555e-06, 'learning_rate/action_model': 5.860349270318555e-05, 'epoch': 1.2}) +[2026-06-17 23:05:00,623][starVLA.training.train_starvla][INFO] - Step 3268, Loss: {'train/loss': 0.4107082188129425, 'train/grad_norm_pre_clip': 11.823604583740234, 'timing/data': 0.012796591967344284, 'timing/model': 1.4429325126111507, 'learning_rate/qwen_vl_interface': 5.8582044447840346e-06, 'learning_rate/action_model': 5.858204444784034e-05, 'epoch': 1.2}) +[2026-06-17 23:05:07,992][starVLA.training.train_starvla][INFO] - Step 3269, Loss: {'train/loss': 0.3678162693977356, 'train/grad_norm_pre_clip': 14.183589935302734, 'timing/data': 0.011937105096876621, 'timing/model': 1.018894069828093, 'learning_rate/qwen_vl_interface': 5.8560594931680785e-06, 'learning_rate/action_model': 5.856059493168078e-05, 'epoch': 1.2}) +[2026-06-17 23:05:12,387][starVLA.training.train_starvla][INFO] - Step 3270, Loss: {'train/loss': 0.5147632360458374, 'train/grad_norm_pre_clip': 14.620392799377441, 'timing/data': 0.010537246242165565, 'timing/model': 1.0251034647226334, 'learning_rate/qwen_vl_interface': 5.853914415915337e-06, 'learning_rate/action_model': 5.853914415915337e-05, 'epoch': 1.2}) +[2026-06-17 23:05:17,174][starVLA.training.train_starvla][INFO] - Step 3271, Loss: {'train/loss': 0.5556589365005493, 'train/grad_norm_pre_clip': 19.127756118774414, 'timing/data': 0.02252873033285141, 'timing/model': 1.1509801968932152, 'learning_rate/qwen_vl_interface': 5.8517692134704875e-06, 'learning_rate/action_model': 5.851769213470487e-05, 'epoch': 1.2}) +[2026-06-17 23:05:21,721][starVLA.training.train_starvla][INFO] - Step 3272, Loss: {'train/loss': 0.5933881998062134, 'train/grad_norm_pre_clip': 14.632112503051758, 'timing/data': 0.01228458620607853, 'timing/model': 1.0574605837464333, 'learning_rate/qwen_vl_interface': 5.849623886278232e-06, 'learning_rate/action_model': 5.849623886278231e-05, 'epoch': 1.2}) +[2026-06-17 23:05:26,342][starVLA.training.train_starvla][INFO] - Step 3273, Loss: {'train/loss': 0.4634629487991333, 'train/grad_norm_pre_clip': 18.61355972290039, 'timing/data': 0.015346178784966469, 'timing/model': 1.577393195591867, 'learning_rate/qwen_vl_interface': 5.8474784347833e-06, 'learning_rate/action_model': 5.8474784347833e-05, 'epoch': 1.2}) +[2026-06-17 23:05:37,270][starVLA.training.train_starvla][INFO] - Step 3274, Loss: {'train/loss': 0.3967735469341278, 'train/grad_norm_pre_clip': 13.135489463806152, 'timing/data': 0.012785591185092926, 'timing/model': 2.9003689689561725, 'learning_rate/qwen_vl_interface': 5.845332859430446e-06, 'learning_rate/action_model': 5.845332859430446e-05, 'epoch': 1.2}) +[2026-06-17 23:05:41,805][starVLA.training.train_starvla][INFO] - Step 3275, Loss: {'train/loss': 0.5239078998565674, 'train/grad_norm_pre_clip': 14.151200294494629, 'timing/data': 0.011425942182540894, 'timing/model': 1.0528595885261893, 'learning_rate/qwen_vl_interface': 5.843187160664449e-06, 'learning_rate/action_model': 5.8431871606644495e-05, 'epoch': 1.2}) +[2026-06-17 23:05:46,411][starVLA.training.train_starvla][INFO] - Step 3276, Loss: {'train/loss': 0.842735230922699, 'train/grad_norm_pre_clip': 20.472797393798828, 'timing/data': 0.011157773435115814, 'timing/model': 1.0375369098037481, 'learning_rate/qwen_vl_interface': 5.841041338930115e-06, 'learning_rate/action_model': 5.8410413389301155e-05, 'epoch': 1.2}) +[2026-06-17 23:05:50,899][starVLA.training.train_starvla][INFO] - Step 3277, Loss: {'train/loss': 0.6922587156295776, 'train/grad_norm_pre_clip': 19.12775230407715, 'timing/data': 0.0114910202100873, 'timing/model': 1.332125966437161, 'learning_rate/qwen_vl_interface': 5.8388953946722775e-06, 'learning_rate/action_model': 5.8388953946722766e-05, 'epoch': 1.2}) +[2026-06-17 23:05:59,060][starVLA.training.train_starvla][INFO] - Step 3278, Loss: {'train/loss': 0.5873814821243286, 'train/grad_norm_pre_clip': 14.790816307067871, 'timing/data': 0.014944413676857948, 'timing/model': 3.023314831778407, 'learning_rate/qwen_vl_interface': 5.836749328335789e-06, 'learning_rate/action_model': 5.8367493283357885e-05, 'epoch': 1.2}) +[2026-06-17 23:06:06,245][starVLA.training.train_starvla][INFO] - Step 3279, Loss: {'train/loss': 0.38827595114707947, 'train/grad_norm_pre_clip': 11.944276809692383, 'timing/data': 0.011893661692738533, 'timing/model': 1.0335228703916073, 'learning_rate/qwen_vl_interface': 5.8346031403655345e-06, 'learning_rate/action_model': 5.8346031403655345e-05, 'epoch': 1.2}) +[2026-06-17 23:06:11,612][starVLA.training.train_starvla][INFO] - Step 3280, Loss: {'train/loss': 0.34476080536842346, 'train/grad_norm_pre_clip': 10.348237037658691, 'timing/data': 0.010756907984614372, 'timing/model': 1.512785724364221, 'learning_rate/qwen_vl_interface': 5.832456831206418e-06, 'learning_rate/action_model': 5.8324568312064176e-05, 'epoch': 1.2}) +[2026-06-17 23:06:16,241][starVLA.training.train_starvla][INFO] - Step 3281, Loss: {'train/loss': 0.5088785886764526, 'train/grad_norm_pre_clip': 14.174942016601562, 'timing/data': 0.014007695950567722, 'timing/model': 1.1997421188279986, 'learning_rate/qwen_vl_interface': 5.830310401303377e-06, 'learning_rate/action_model': 5.830310401303377e-05, 'epoch': 1.2}) +[2026-06-17 23:06:25,710][starVLA.training.train_starvla][INFO] - Step 3282, Loss: {'train/loss': 0.5677276849746704, 'train/grad_norm_pre_clip': 23.03063201904297, 'timing/data': 0.014583968557417393, 'timing/model': 1.0950693460181355, 'learning_rate/qwen_vl_interface': 5.8281638511013625e-06, 'learning_rate/action_model': 5.828163851101363e-05, 'epoch': 1.2}) +[2026-06-17 23:06:34,269][starVLA.training.train_starvla][INFO] - Step 3283, Loss: {'train/loss': 0.32463449239730835, 'train/grad_norm_pre_clip': 9.8107271194458, 'timing/data': 0.012847994454205036, 'timing/model': 1.5697241835296154, 'learning_rate/qwen_vl_interface': 5.8260171810453605e-06, 'learning_rate/action_model': 5.82601718104536e-05, 'epoch': 1.2}) +[2026-06-17 23:06:44,719][starVLA.training.train_starvla][INFO] - Step 3284, Loss: {'train/loss': 0.26552945375442505, 'train/grad_norm_pre_clip': 10.937657356262207, 'timing/data': 0.013206321746110916, 'timing/model': 4.034677630290389, 'learning_rate/qwen_vl_interface': 5.823870391580376e-06, 'learning_rate/action_model': 5.823870391580376e-05, 'epoch': 1.2}) +[2026-06-17 23:06:49,412][starVLA.training.train_starvla][INFO] - Step 3285, Loss: {'train/loss': 0.5900207161903381, 'train/grad_norm_pre_clip': 24.291715621948242, 'timing/data': 0.010408220812678337, 'timing/model': 1.0611450420692563, 'learning_rate/qwen_vl_interface': 5.821723483151444e-06, 'learning_rate/action_model': 5.821723483151443e-05, 'epoch': 1.2}) +[2026-06-17 23:06:54,171][starVLA.training.train_starvla][INFO] - Step 3286, Loss: {'train/loss': 0.6890708208084106, 'train/grad_norm_pre_clip': 16.869657516479492, 'timing/data': 0.010488620959222317, 'timing/model': 1.105505465529859, 'learning_rate/qwen_vl_interface': 5.819576456203617e-06, 'learning_rate/action_model': 5.8195764562036174e-05, 'epoch': 1.2}) +[2026-06-17 23:07:04,684][starVLA.training.train_starvla][INFO] - Step 3287, Loss: {'train/loss': 0.2905047535896301, 'train/grad_norm_pre_clip': 13.551213264465332, 'timing/data': 0.01224343292415142, 'timing/model': 1.4838894121348858, 'learning_rate/qwen_vl_interface': 5.817429311181977e-06, 'learning_rate/action_model': 5.8174293111819764e-05, 'epoch': 1.2}) +[2026-06-17 23:07:13,813][starVLA.training.train_starvla][INFO] - Step 3288, Loss: {'train/loss': 0.5170843005180359, 'train/grad_norm_pre_clip': 15.819464683532715, 'timing/data': 0.014852329157292843, 'timing/model': 1.5124355144798756, 'learning_rate/qwen_vl_interface': 5.815282048531633e-06, 'learning_rate/action_model': 5.815282048531633e-05, 'epoch': 1.2}) +[2026-06-17 23:07:17,930][starVLA.training.train_starvla][INFO] - Step 3289, Loss: {'train/loss': 0.5198253393173218, 'train/grad_norm_pre_clip': 32.6502685546875, 'timing/data': 0.010625061579048634, 'timing/model': 1.126109172590077, 'learning_rate/qwen_vl_interface': 5.81313466869771e-06, 'learning_rate/action_model': 5.8131346686977096e-05, 'epoch': 1.2}) +[2026-06-17 23:07:22,507][starVLA.training.train_starvla][INFO] - Step 3290, Loss: {'train/loss': 0.2598476707935333, 'train/grad_norm_pre_clip': 11.660573959350586, 'timing/data': 0.01435044500976801, 'timing/model': 1.1872582761570811, 'learning_rate/qwen_vl_interface': 5.810987172125364e-06, 'learning_rate/action_model': 5.810987172125364e-05, 'epoch': 1.2}) +[2026-06-17 23:07:26,989][starVLA.training.train_starvla][INFO] - Step 3291, Loss: {'train/loss': 0.3653945028781891, 'train/grad_norm_pre_clip': 13.685026168823242, 'timing/data': 0.010583322495222092, 'timing/model': 1.0442510461434722, 'learning_rate/qwen_vl_interface': 5.808839559259774e-06, 'learning_rate/action_model': 5.8088395592597746e-05, 'epoch': 1.21}) +[2026-06-17 23:07:31,605][starVLA.training.train_starvla][INFO] - Step 3292, Loss: {'train/loss': 0.5516014695167542, 'train/grad_norm_pre_clip': 14.358367919921875, 'timing/data': 0.00996759906411171, 'timing/model': 1.0386262107640505, 'learning_rate/qwen_vl_interface': 5.806691830546143e-06, 'learning_rate/action_model': 5.8066918305461427e-05, 'epoch': 1.21}) +[2026-06-17 23:07:37,133][starVLA.training.train_starvla][INFO] - Step 3293, Loss: {'train/loss': 0.3867035210132599, 'train/grad_norm_pre_clip': 15.158074378967285, 'timing/data': 0.01402690913528204, 'timing/model': 1.5376636888831854, 'learning_rate/qwen_vl_interface': 5.804543986429696e-06, 'learning_rate/action_model': 5.804543986429696e-05, 'epoch': 1.21}) +[2026-06-17 23:07:45,305][starVLA.training.train_starvla][INFO] - Step 3294, Loss: {'train/loss': 0.6141894459724426, 'train/grad_norm_pre_clip': 18.472869873046875, 'timing/data': 0.010781427845358849, 'timing/model': 1.070401311852038, 'learning_rate/qwen_vl_interface': 5.802396027355683e-06, 'learning_rate/action_model': 5.8023960273556835e-05, 'epoch': 1.21}) +[2026-06-17 23:07:49,665][starVLA.training.train_starvla][INFO] - Step 3295, Loss: {'train/loss': 0.6690823435783386, 'train/grad_norm_pre_clip': 17.549942016601562, 'timing/data': 0.011048388667404652, 'timing/model': 1.062522742897272, 'learning_rate/qwen_vl_interface': 5.800247953769381e-06, 'learning_rate/action_model': 5.800247953769381e-05, 'epoch': 1.21}) +[2026-06-17 23:07:54,602][starVLA.training.train_starvla][INFO] - Step 3296, Loss: {'train/loss': 0.4441594183444977, 'train/grad_norm_pre_clip': 16.364492416381836, 'timing/data': 0.013929339125752449, 'timing/model': 1.0776559133082628, 'learning_rate/qwen_vl_interface': 5.798099766116084e-06, 'learning_rate/action_model': 5.798099766116084e-05, 'epoch': 1.21}) +[2026-06-17 23:07:59,230][starVLA.training.train_starvla][INFO] - Step 3297, Loss: {'train/loss': 0.8057790994644165, 'train/grad_norm_pre_clip': 22.008769989013672, 'timing/data': 0.01019427552819252, 'timing/model': 1.4814148861914873, 'learning_rate/qwen_vl_interface': 5.7959514648411175e-06, 'learning_rate/action_model': 5.795951464841117e-05, 'epoch': 1.21}) +[2026-06-17 23:08:03,712][starVLA.training.train_starvla][INFO] - Step 3298, Loss: {'train/loss': 0.5763375163078308, 'train/grad_norm_pre_clip': 21.98434066772461, 'timing/data': 0.010726502165198326, 'timing/model': 1.365052241832018, 'learning_rate/qwen_vl_interface': 5.793803050389824e-06, 'learning_rate/action_model': 5.793803050389824e-05, 'epoch': 1.21}) +[2026-06-17 23:08:08,537][starVLA.training.train_starvla][INFO] - Step 3299, Loss: {'train/loss': 0.5770501494407654, 'train/grad_norm_pre_clip': 19.742664337158203, 'timing/data': 0.011521638371050358, 'timing/model': 1.4852507021278143, 'learning_rate/qwen_vl_interface': 5.7916545232075725e-06, 'learning_rate/action_model': 5.791654523207573e-05, 'epoch': 1.21}) +[2026-06-17 23:10:53,485][starVLA.training.train_starvla][INFO] - Step 3300, Loss: {'train/loss': 0.48262399435043335, 'train/grad_norm_pre_clip': 19.579988479614258, 'eval/loss': 0.6334459185600281, 'eval/latency_2/loss': 0.6334459185600281, 'eval/demon_fire_precision': 0.9385072094995759, 'eval/demon_fire_recall': 0.9259414225941422, 'eval/demon_fire_f1': 0.9321819713563605, 'eval/demon_fire_rate_model': 0.656093489148581, 'eval/demon_fire_rate_teacher': 0.6649972175848636, 'eval/demon_left_precision': 0.8710762331838565, 'eval/demon_left_recall': 0.8178947368421052, 'eval/demon_left_f1': 0.8436482084690554, 'eval/demon_left_rate_model': 0.24819143016138007, 'eval/demon_left_rate_teacher': 0.2643294379521425, 'eval/demon_right_precision': 0.908843537414966, 'eval/demon_right_recall': 0.7548022598870057, 'eval/demon_right_f1': 0.8246913580246913, 'eval/demon_right_rate_model': 0.2045075125208681, 'eval/demon_right_rate_teacher': 0.24624373956594323, 'eval/demon_move_f1': 0.8341697832468733, 'eval/demon_cc_f1': 0.9004422753916245, 'eval/latency_2/demon_fire_precision': 0.9385072094995759, 'eval/latency_2/demon_fire_recall': 0.9259414225941422, 'eval/latency_2/demon_fire_f1': 0.9321819713563605, 'eval/latency_2/demon_fire_rate_model': 0.656093489148581, 'eval/latency_2/demon_fire_rate_teacher': 0.6649972175848636, 'eval/latency_2/demon_left_precision': 0.8710762331838565, 'eval/latency_2/demon_left_recall': 0.8178947368421052, 'eval/latency_2/demon_left_f1': 0.8436482084690554, 'eval/latency_2/demon_left_rate_model': 0.24819143016138007, 'eval/latency_2/demon_left_rate_teacher': 0.2643294379521425, 'eval/latency_2/demon_right_precision': 0.908843537414966, 'eval/latency_2/demon_right_recall': 0.7548022598870057, 'eval/latency_2/demon_right_f1': 0.8246913580246913, 'eval/latency_2/demon_right_rate_model': 0.2045075125208681, 'eval/latency_2/demon_right_rate_teacher': 0.24624373956594323, 'eval/latency_2/demon_move_f1': 0.8341697832468733, 'eval/latency_2/demon_cc_f1': 0.9004422753916245, 'timing/data': 0.010568605735898018, 'timing/model': 3.567331686615944, 'learning_rate/qwen_vl_interface': 5.789505883739756e-06, 'learning_rate/action_model': 5.7895058837397554e-05, 'epoch': 1.21}) +[2026-06-17 23:10:53,589][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_3300_state +[2026-06-17 23:10:53,597][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 23:11:08,894][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_3300_state/model.safetensors +[2026-06-17 23:11:27,866][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_3300_state/optimizer.bin +[2026-06-17 23:11:27,867][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_3300_state/sampler.bin +[2026-06-17 23:11:27,867][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_3300_state/sampler_1.bin +[2026-06-17 23:11:27,870][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_3300_state/random_states_0.pkl +[2026-06-17 23:11:27,875][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3300 +[2026-06-17 23:11:27,875][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 23:11:27,892][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 23:11:41,274][starVLA.training.train_starvla][INFO] - Step 3301, Loss: {'train/loss': 0.7272566556930542, 'train/grad_norm_pre_clip': 21.998165130615234, 'timing/data': 0.014511125162243843, 'timing/model': 2.9338081600144506, 'learning_rate/qwen_vl_interface': 5.78735713243179e-06, 'learning_rate/action_model': 5.78735713243179e-05, 'epoch': 1.21}) +[2026-06-17 23:11:54,034][starVLA.training.train_starvla][INFO] - Step 3302, Loss: {'train/loss': 0.8096444606781006, 'train/grad_norm_pre_clip': 23.47176742553711, 'timing/data': 0.0011986885219812393, 'timing/model': 4.354556445032358, 'learning_rate/qwen_vl_interface': 5.785208269729111e-06, 'learning_rate/action_model': 5.785208269729111e-05, 'epoch': 1.21}) +[2026-06-17 23:12:01,435][starVLA.training.train_starvla][INFO] - Step 3303, Loss: {'train/loss': 0.8697846531867981, 'train/grad_norm_pre_clip': 21.358388900756836, 'timing/data': 0.0012129191309213638, 'timing/model': 1.6003522258251905, 'learning_rate/qwen_vl_interface': 5.783059296077183e-06, 'learning_rate/action_model': 5.7830592960771826e-05, 'epoch': 1.21}) +[2026-06-17 23:12:09,771][starVLA.training.train_starvla][INFO] - Step 3304, Loss: {'train/loss': 0.8308653831481934, 'train/grad_norm_pre_clip': 39.45259094238281, 'timing/data': 0.001035497523844242, 'timing/model': 1.7359614865854383, 'learning_rate/qwen_vl_interface': 5.780910211921488e-06, 'learning_rate/action_model': 5.7809102119214874e-05, 'epoch': 1.21}) +[2026-06-17 23:12:20,418][starVLA.training.train_starvla][INFO] - Step 3305, Loss: {'train/loss': 0.48586708307266235, 'train/grad_norm_pre_clip': 15.584321022033691, 'timing/data': 0.0009130286052823067, 'timing/model': 1.2113732546567917, 'learning_rate/qwen_vl_interface': 5.778761017707536e-06, 'learning_rate/action_model': 5.778761017707536e-05, 'epoch': 1.21}) +[2026-06-17 23:12:30,575][starVLA.training.train_starvla][INFO] - Step 3306, Loss: {'train/loss': 0.7744710445404053, 'train/grad_norm_pre_clip': 19.32332992553711, 'timing/data': 0.0011105062440037727, 'timing/model': 1.4076003842055798, 'learning_rate/qwen_vl_interface': 5.7766117138808545e-06, 'learning_rate/action_model': 5.776611713880854e-05, 'epoch': 1.21}) +[2026-06-17 23:12:40,422][starVLA.training.train_starvla][INFO] - Step 3307, Loss: {'train/loss': 0.5104484558105469, 'train/grad_norm_pre_clip': 18.378406524658203, 'timing/data': 0.0009808680042624474, 'timing/model': 1.2764518558979034, 'learning_rate/qwen_vl_interface': 5.774462300886999e-06, 'learning_rate/action_model': 5.774462300886999e-05, 'epoch': 1.21}) +[2026-06-17 23:12:44,990][starVLA.training.train_starvla][INFO] - Step 3308, Loss: {'train/loss': 0.5740986466407776, 'train/grad_norm_pre_clip': 15.81457233428955, 'timing/data': 0.0020359698683023453, 'timing/model': 1.0713678766041994, 'learning_rate/qwen_vl_interface': 5.772312779171545e-06, 'learning_rate/action_model': 5.7723127791715445e-05, 'epoch': 1.21}) +[2026-06-17 23:12:49,095][starVLA.training.train_starvla][INFO] - Step 3309, Loss: {'train/loss': 0.5127440690994263, 'train/grad_norm_pre_clip': 15.137004852294922, 'timing/data': 0.000818522647023201, 'timing/model': 1.0479237847030163, 'learning_rate/qwen_vl_interface': 5.770163149180089e-06, 'learning_rate/action_model': 5.770163149180089e-05, 'epoch': 1.21}) +[2026-06-17 23:12:53,678][starVLA.training.train_starvla][INFO] - Step 3310, Loss: {'train/loss': 0.6169427037239075, 'train/grad_norm_pre_clip': 18.48514747619629, 'timing/data': 0.001033150590956211, 'timing/model': 1.0299612237140536, 'learning_rate/qwen_vl_interface': 5.768013411358252e-06, 'learning_rate/action_model': 5.768013411358252e-05, 'epoch': 1.21}) +[2026-06-17 23:12:58,599][starVLA.training.train_starvla][INFO] - Step 3311, Loss: {'train/loss': 0.34849798679351807, 'train/grad_norm_pre_clip': 11.381749153137207, 'timing/data': 0.0007681269198656082, 'timing/model': 1.3882702644914389, 'learning_rate/qwen_vl_interface': 5.765863566151677e-06, 'learning_rate/action_model': 5.7658635661516767e-05, 'epoch': 1.21}) +[2026-06-17 23:13:02,963][starVLA.training.train_starvla][INFO] - Step 3312, Loss: {'train/loss': 0.3317832052707672, 'train/grad_norm_pre_clip': 11.365324020385742, 'timing/data': 0.0008855639025568962, 'timing/model': 1.03573109023273, 'learning_rate/qwen_vl_interface': 5.763713614006033e-06, 'learning_rate/action_model': 5.763713614006032e-05, 'epoch': 1.21}) +[2026-06-17 23:13:07,491][starVLA.training.train_starvla][INFO] - Step 3313, Loss: {'train/loss': 0.4064522981643677, 'train/grad_norm_pre_clip': 23.55255699157715, 'timing/data': 0.0012366753071546555, 'timing/model': 1.0369659829884768, 'learning_rate/qwen_vl_interface': 5.761563555367e-06, 'learning_rate/action_model': 5.761563555367e-05, 'epoch': 1.21}) +[2026-06-17 23:13:11,838][starVLA.training.train_starvla][INFO] - Step 3314, Loss: {'train/loss': 0.15877743065357208, 'train/grad_norm_pre_clip': 8.867121696472168, 'timing/data': 0.0011005457490682602, 'timing/model': 1.3161321450024843, 'learning_rate/qwen_vl_interface': 5.759413390680293e-06, 'learning_rate/action_model': 5.759413390680293e-05, 'epoch': 1.21}) +[2026-06-17 23:13:17,635][starVLA.training.train_starvla][INFO] - Step 3315, Loss: {'train/loss': 0.6101161241531372, 'train/grad_norm_pre_clip': 21.638938903808594, 'timing/data': 0.001482001505792141, 'timing/model': 2.587808703072369, 'learning_rate/qwen_vl_interface': 5.757263120391643e-06, 'learning_rate/action_model': 5.757263120391643e-05, 'epoch': 1.21}) +[2026-06-17 23:13:22,491][starVLA.training.train_starvla][INFO] - Step 3316, Loss: {'train/loss': 0.4381544291973114, 'train/grad_norm_pre_clip': 16.74506187438965, 'timing/data': 0.0010342681780457497, 'timing/model': 1.04997365642339, 'learning_rate/qwen_vl_interface': 5.7551127449468005e-06, 'learning_rate/action_model': 5.7551127449468004e-05, 'epoch': 1.21}) +[2026-06-17 23:13:27,145][starVLA.training.train_starvla][INFO] - Step 3317, Loss: {'train/loss': 0.5136747360229492, 'train/grad_norm_pre_clip': 21.427339553833008, 'timing/data': 0.0010904772207140923, 'timing/model': 1.163901132531464, 'learning_rate/qwen_vl_interface': 5.752962264791543e-06, 'learning_rate/action_model': 5.752962264791543e-05, 'epoch': 1.21}) +[2026-06-17 23:13:31,954][starVLA.training.train_starvla][INFO] - Step 3318, Loss: {'train/loss': 0.5766276121139526, 'train/grad_norm_pre_clip': 16.213970184326172, 'timing/data': 0.001196758821606636, 'timing/model': 1.578508042730391, 'learning_rate/qwen_vl_interface': 5.750811680371666e-06, 'learning_rate/action_model': 5.750811680371666e-05, 'epoch': 1.21}) +[2026-06-17 23:13:39,817][starVLA.training.train_starvla][INFO] - Step 3319, Loss: {'train/loss': 0.6544272303581238, 'train/grad_norm_pre_clip': 19.946434020996094, 'timing/data': 0.0011124489828944206, 'timing/model': 1.6664385423064232, 'learning_rate/qwen_vl_interface': 5.7486609921329914e-06, 'learning_rate/action_model': 5.748660992132992e-05, 'epoch': 1.22}) +[2026-06-17 23:13:43,952][starVLA.training.train_starvla][INFO] - Step 3320, Loss: {'train/loss': 0.5799322128295898, 'train/grad_norm_pre_clip': 19.668174743652344, 'timing/data': 0.0009761108085513115, 'timing/model': 1.1347540626302361, 'learning_rate/qwen_vl_interface': 5.746510200521354e-06, 'learning_rate/action_model': 5.746510200521355e-05, 'epoch': 1.22}) +[2026-06-17 23:13:48,950][starVLA.training.train_starvla][INFO] - Step 3321, Loss: {'train/loss': 0.419391930103302, 'train/grad_norm_pre_clip': 13.87524127960205, 'timing/data': 0.0008943676948547363, 'timing/model': 1.1491471296176314, 'learning_rate/qwen_vl_interface': 5.7443593059826206e-06, 'learning_rate/action_model': 5.744359305982621e-05, 'epoch': 1.22}) +[2026-06-17 23:13:53,703][starVLA.training.train_starvla][INFO] - Step 3322, Loss: {'train/loss': 0.7671672701835632, 'train/grad_norm_pre_clip': 22.19841766357422, 'timing/data': 0.001232837326824665, 'timing/model': 1.038216046988964, 'learning_rate/qwen_vl_interface': 5.742208308962669e-06, 'learning_rate/action_model': 5.742208308962669e-05, 'epoch': 1.22}) +[2026-06-17 23:13:58,452][starVLA.training.train_starvla][INFO] - Step 3323, Loss: {'train/loss': 0.8982638716697693, 'train/grad_norm_pre_clip': 24.498821258544922, 'timing/data': 0.0010735774412751198, 'timing/model': 1.1727963872253895, 'learning_rate/qwen_vl_interface': 5.740057209907405e-06, 'learning_rate/action_model': 5.7400572099074046e-05, 'epoch': 1.22}) +[2026-06-17 23:14:07,669][starVLA.training.train_starvla][INFO] - Step 3324, Loss: {'train/loss': 0.5328699350357056, 'train/grad_norm_pre_clip': 17.3751277923584, 'timing/data': 0.001142631284892559, 'timing/model': 4.584576481953263, 'learning_rate/qwen_vl_interface': 5.7379060092627545e-06, 'learning_rate/action_model': 5.7379060092627544e-05, 'epoch': 1.22}) +[2026-06-17 23:14:11,979][starVLA.training.train_starvla][INFO] - Step 3325, Loss: {'train/loss': 0.5674741268157959, 'train/grad_norm_pre_clip': 14.04373836517334, 'timing/data': 0.0014678975567221642, 'timing/model': 1.046727592125535, 'learning_rate/qwen_vl_interface': 5.735754707474663e-06, 'learning_rate/action_model': 5.735754707474663e-05, 'epoch': 1.22}) +[2026-06-17 23:14:16,475][starVLA.training.train_starvla][INFO] - Step 3326, Loss: {'train/loss': 0.8350922465324402, 'train/grad_norm_pre_clip': 19.283687591552734, 'timing/data': 0.0012415200471878052, 'timing/model': 1.0358964027836919, 'learning_rate/qwen_vl_interface': 5.7336033049891e-06, 'learning_rate/action_model': 5.733603304989099e-05, 'epoch': 1.22}) +[2026-06-17 23:14:23,042][starVLA.training.train_starvla][INFO] - Step 3327, Loss: {'train/loss': 0.6024804711341858, 'train/grad_norm_pre_clip': 15.817747116088867, 'timing/data': 0.0009010136127471924, 'timing/model': 3.1154075181111693, 'learning_rate/qwen_vl_interface': 5.731451802252049e-06, 'learning_rate/action_model': 5.731451802252048e-05, 'epoch': 1.22}) +[2026-06-17 23:14:28,435][starVLA.training.train_starvla][INFO] - Step 3328, Loss: {'train/loss': 0.5064290165901184, 'train/grad_norm_pre_clip': 14.903637886047363, 'timing/data': 0.0008877580985426903, 'timing/model': 1.240087983198464, 'learning_rate/qwen_vl_interface': 5.729300199709522e-06, 'learning_rate/action_model': 5.729300199709522e-05, 'epoch': 1.22}) +[2026-06-17 23:14:33,011][starVLA.training.train_starvla][INFO] - Step 3329, Loss: {'train/loss': 0.39421066641807556, 'train/grad_norm_pre_clip': 10.936241149902344, 'timing/data': 0.0013283519074320793, 'timing/model': 1.2686461163684726, 'learning_rate/qwen_vl_interface': 5.7271484978075475e-06, 'learning_rate/action_model': 5.7271484978075475e-05, 'epoch': 1.22}) +[2026-06-17 23:14:38,585][starVLA.training.train_starvla][INFO] - Step 3330, Loss: {'train/loss': 0.4572954475879669, 'train/grad_norm_pre_clip': 13.762036323547363, 'timing/data': 0.0012042969465255737, 'timing/model': 2.5630326261743903, 'learning_rate/qwen_vl_interface': 5.724996696992176e-06, 'learning_rate/action_model': 5.7249966969921764e-05, 'epoch': 1.22}) +[2026-06-17 23:14:46,631][starVLA.training.train_starvla][INFO] - Step 3331, Loss: {'train/loss': 0.4801264703273773, 'train/grad_norm_pre_clip': 16.766355514526367, 'timing/data': 0.0012135840952396393, 'timing/model': 1.2317010182887316, 'learning_rate/qwen_vl_interface': 5.72284479770948e-06, 'learning_rate/action_model': 5.72284479770948e-05, 'epoch': 1.22}) +[2026-06-17 23:14:53,739][starVLA.training.train_starvla][INFO] - Step 3332, Loss: {'train/loss': 0.36959385871887207, 'train/grad_norm_pre_clip': 15.084198951721191, 'timing/data': 0.0014353720471262932, 'timing/model': 1.1249637231230736, 'learning_rate/qwen_vl_interface': 5.720692800405547e-06, 'learning_rate/action_model': 5.720692800405547e-05, 'epoch': 1.22}) +[2026-06-17 23:14:58,660][starVLA.training.train_starvla][INFO] - Step 3333, Loss: {'train/loss': 0.5061643123626709, 'train/grad_norm_pre_clip': 13.933995246887207, 'timing/data': 0.0010425811633467674, 'timing/model': 1.2768823932856321, 'learning_rate/qwen_vl_interface': 5.71854070552649e-06, 'learning_rate/action_model': 5.7185407055264904e-05, 'epoch': 1.22}) +[2026-06-17 23:15:03,533][starVLA.training.train_starvla][INFO] - Step 3334, Loss: {'train/loss': 0.5457189083099365, 'train/grad_norm_pre_clip': 13.583428382873535, 'timing/data': 0.0010153111070394516, 'timing/model': 1.5943671111017466, 'learning_rate/qwen_vl_interface': 5.71638851351844e-06, 'learning_rate/action_model': 5.71638851351844e-05, 'epoch': 1.22}) +[2026-06-17 23:15:13,855][starVLA.training.train_starvla][INFO] - Step 3335, Loss: {'train/loss': 0.3473520278930664, 'train/grad_norm_pre_clip': 11.028011322021484, 'timing/data': 0.0007963338866829872, 'timing/model': 1.1293015405535698, 'learning_rate/qwen_vl_interface': 5.714236224827551e-06, 'learning_rate/action_model': 5.714236224827551e-05, 'epoch': 1.22}) +[2026-06-17 23:15:18,423][starVLA.training.train_starvla][INFO] - Step 3336, Loss: {'train/loss': 0.526771068572998, 'train/grad_norm_pre_clip': 18.25423240661621, 'timing/data': 0.0011625196784734726, 'timing/model': 1.1194576574489474, 'learning_rate/qwen_vl_interface': 5.712083839899991e-06, 'learning_rate/action_model': 5.712083839899991e-05, 'epoch': 1.22}) +[2026-06-17 23:15:22,955][starVLA.training.train_starvla][INFO] - Step 3337, Loss: {'train/loss': 0.497410386800766, 'train/grad_norm_pre_clip': 17.95187759399414, 'timing/data': 0.0014892928302288055, 'timing/model': 1.0603355541825294, 'learning_rate/qwen_vl_interface': 5.709931359181956e-06, 'learning_rate/action_model': 5.709931359181956e-05, 'epoch': 1.22}) +[2026-06-17 23:15:27,599][starVLA.training.train_starvla][INFO] - Step 3338, Loss: {'train/loss': 0.4967195987701416, 'train/grad_norm_pre_clip': 17.062902450561523, 'timing/data': 0.0010739844292402267, 'timing/model': 1.3438669089227915, 'learning_rate/qwen_vl_interface': 5.707778783119654e-06, 'learning_rate/action_model': 5.707778783119654e-05, 'epoch': 1.22}) +[2026-06-17 23:15:34,546][starVLA.training.train_starvla][INFO] - Step 3339, Loss: {'train/loss': 0.8734505772590637, 'train/grad_norm_pre_clip': 19.857284545898438, 'timing/data': 0.0013953782618045807, 'timing/model': 1.400034748017788, 'learning_rate/qwen_vl_interface': 5.705626112159318e-06, 'learning_rate/action_model': 5.7056261121593177e-05, 'epoch': 1.22}) +[2026-06-17 23:15:39,075][starVLA.training.train_starvla][INFO] - Step 3340, Loss: {'train/loss': 0.7020620107650757, 'train/grad_norm_pre_clip': 19.448253631591797, 'timing/data': 0.0009043673053383827, 'timing/model': 1.1706793792545795, 'learning_rate/qwen_vl_interface': 5.703473346747199e-06, 'learning_rate/action_model': 5.703473346747199e-05, 'epoch': 1.22}) +[2026-06-17 23:15:47,380][starVLA.training.train_starvla][INFO] - Step 3341, Loss: {'train/loss': 0.7255867123603821, 'train/grad_norm_pre_clip': 21.643829345703125, 'timing/data': 0.001314215362071991, 'timing/model': 1.1116802599281073, 'learning_rate/qwen_vl_interface': 5.701320487329567e-06, 'learning_rate/action_model': 5.701320487329567e-05, 'epoch': 1.22}) +[2026-06-17 23:15:52,098][starVLA.training.train_starvla][INFO] - Step 3342, Loss: {'train/loss': 0.6003655195236206, 'train/grad_norm_pre_clip': 15.852533340454102, 'timing/data': 0.0012319982051849365, 'timing/model': 1.118380363099277, 'learning_rate/qwen_vl_interface': 5.699167534352712e-06, 'learning_rate/action_model': 5.699167534352712e-05, 'epoch': 1.22}) +[2026-06-17 23:15:56,771][starVLA.training.train_starvla][INFO] - Step 3343, Loss: {'train/loss': 0.7196843028068542, 'train/grad_norm_pre_clip': 17.225788116455078, 'timing/data': 0.0009750034660100937, 'timing/model': 1.1552461693063378, 'learning_rate/qwen_vl_interface': 5.697014488262945e-06, 'learning_rate/action_model': 5.697014488262945e-05, 'epoch': 1.22}) +[2026-06-17 23:16:01,531][starVLA.training.train_starvla][INFO] - Step 3344, Loss: {'train/loss': 0.3487134873867035, 'train/grad_norm_pre_clip': 13.304247856140137, 'timing/data': 0.0008013835176825523, 'timing/model': 1.379789729602635, 'learning_rate/qwen_vl_interface': 5.694861349506591e-06, 'learning_rate/action_model': 5.6948613495065903e-05, 'epoch': 1.22}) +[2026-06-17 23:16:05,983][starVLA.training.train_starvla][INFO] - Step 3345, Loss: {'train/loss': 0.3045874238014221, 'train/grad_norm_pre_clip': 12.628793716430664, 'timing/data': 0.0010895729064941406, 'timing/model': 1.2025203770026565, 'learning_rate/qwen_vl_interface': 5.692708118530003e-06, 'learning_rate/action_model': 5.6927081185300025e-05, 'epoch': 1.22}) +[2026-06-17 23:16:13,874][starVLA.training.train_starvla][INFO] - Step 3346, Loss: {'train/loss': 0.3587700426578522, 'train/grad_norm_pre_clip': 13.764128684997559, 'timing/data': 0.0012702904641628265, 'timing/model': 4.321169789880514, 'learning_rate/qwen_vl_interface': 5.6905547957795446e-06, 'learning_rate/action_model': 5.690554795779545e-05, 'epoch': 1.23}) +[2026-06-17 23:16:18,200][starVLA.training.train_starvla][INFO] - Step 3347, Loss: {'train/loss': 0.28232529759407043, 'train/grad_norm_pre_clip': 9.52027702331543, 'timing/data': 0.0011064084246754646, 'timing/model': 1.0350826736539602, 'learning_rate/qwen_vl_interface': 5.688401381701604e-06, 'learning_rate/action_model': 5.688401381701603e-05, 'epoch': 1.23}) +[2026-06-17 23:16:22,681][starVLA.training.train_starvla][INFO] - Step 3348, Loss: {'train/loss': 0.491578072309494, 'train/grad_norm_pre_clip': 18.67092514038086, 'timing/data': 0.0008939709514379501, 'timing/model': 1.0428647873923182, 'learning_rate/qwen_vl_interface': 5.686247876742583e-06, 'learning_rate/action_model': 5.686247876742583e-05, 'epoch': 1.23}) +[2026-06-17 23:16:27,389][starVLA.training.train_starvla][INFO] - Step 3349, Loss: {'train/loss': 0.7428930997848511, 'train/grad_norm_pre_clip': 17.376237869262695, 'timing/data': 0.001247580163180828, 'timing/model': 1.3566419295966625, 'learning_rate/qwen_vl_interface': 5.684094281348908e-06, 'learning_rate/action_model': 5.684094281348908e-05, 'epoch': 1.23}) +[2026-06-17 23:16:32,591][starVLA.training.train_starvla][INFO] - Step 3350, Loss: {'train/loss': 0.5627080202102661, 'train/grad_norm_pre_clip': 17.49944496154785, 'timing/data': 0.0011005466803908348, 'timing/model': 1.0443905098363757, 'learning_rate/qwen_vl_interface': 5.681940595967023e-06, 'learning_rate/action_model': 5.681940595967024e-05, 'epoch': 1.23}) +[2026-06-17 23:16:37,236][starVLA.training.train_starvla][INFO] - Step 3351, Loss: {'train/loss': 0.30909860134124756, 'train/grad_norm_pre_clip': 13.939672470092773, 'timing/data': 0.0008359542116522789, 'timing/model': 1.09899971075356, 'learning_rate/qwen_vl_interface': 5.679786821043388e-06, 'learning_rate/action_model': 5.679786821043388e-05, 'epoch': 1.23}) +[2026-06-17 23:16:41,770][starVLA.training.train_starvla][INFO] - Step 3352, Loss: {'train/loss': 0.5903454422950745, 'train/grad_norm_pre_clip': 19.48160171508789, 'timing/data': 0.0011231983080506325, 'timing/model': 1.0292185237631202, 'learning_rate/qwen_vl_interface': 5.677632957024482e-06, 'learning_rate/action_model': 5.6776329570244826e-05, 'epoch': 1.23}) +[2026-06-17 23:16:50,776][starVLA.training.train_starvla][INFO] - Step 3353, Loss: {'train/loss': 0.8177729249000549, 'train/grad_norm_pre_clip': 19.22101593017578, 'timing/data': 0.0009514531120657921, 'timing/model': 1.124434976838529, 'learning_rate/qwen_vl_interface': 5.675479004356803e-06, 'learning_rate/action_model': 5.675479004356803e-05, 'epoch': 1.23}) +[2026-06-17 23:16:55,307][starVLA.training.train_starvla][INFO] - Step 3354, Loss: {'train/loss': 0.6204981207847595, 'train/grad_norm_pre_clip': 15.55310344696045, 'timing/data': 0.0010412363335490227, 'timing/model': 1.444943144917488, 'learning_rate/qwen_vl_interface': 5.673324963486871e-06, 'learning_rate/action_model': 5.67332496348687e-05, 'epoch': 1.23}) +[2026-06-17 23:16:59,358][starVLA.training.train_starvla][INFO] - Step 3355, Loss: {'train/loss': 0.502411961555481, 'train/grad_norm_pre_clip': 15.565900802612305, 'timing/data': 0.001183660700917244, 'timing/model': 1.0345425195991993, 'learning_rate/qwen_vl_interface': 5.671170834861215e-06, 'learning_rate/action_model': 5.671170834861215e-05, 'epoch': 1.23}) +[2026-06-17 23:17:04,055][starVLA.training.train_starvla][INFO] - Step 3356, Loss: {'train/loss': 0.4952273666858673, 'train/grad_norm_pre_clip': 15.1902494430542, 'timing/data': 0.0008443919941782951, 'timing/model': 1.1660607689991593, 'learning_rate/qwen_vl_interface': 5.6690166189263925e-06, 'learning_rate/action_model': 5.669016618926393e-05, 'epoch': 1.23}) +[2026-06-17 23:17:08,794][starVLA.training.train_starvla][INFO] - Step 3357, Loss: {'train/loss': 1.028390645980835, 'train/grad_norm_pre_clip': 23.633930206298828, 'timing/data': 0.0010886769741773605, 'timing/model': 1.0856786333024502, 'learning_rate/qwen_vl_interface': 5.666862316128977e-06, 'learning_rate/action_model': 5.666862316128977e-05, 'epoch': 1.23}) +[2026-06-17 23:17:13,333][starVLA.training.train_starvla][INFO] - Step 3358, Loss: {'train/loss': 0.44467929005622864, 'train/grad_norm_pre_clip': 15.757469177246094, 'timing/data': 0.0010476205497980118, 'timing/model': 1.4705386320129037, 'learning_rate/qwen_vl_interface': 5.664707926915551e-06, 'learning_rate/action_model': 5.66470792691555e-05, 'epoch': 1.23}) +[2026-06-17 23:17:17,824][starVLA.training.train_starvla][INFO] - Step 3359, Loss: {'train/loss': 0.9352335333824158, 'train/grad_norm_pre_clip': 23.793109893798828, 'timing/data': 0.0017495350912213326, 'timing/model': 1.4847912788391113, 'learning_rate/qwen_vl_interface': 5.662553451732729e-06, 'learning_rate/action_model': 5.6625534517327285e-05, 'epoch': 1.23}) +[2026-06-17 23:17:21,999][starVLA.training.train_starvla][INFO] - Step 3360, Loss: {'train/loss': 0.2926733195781708, 'train/grad_norm_pre_clip': 14.010797500610352, 'timing/data': 0.0008592698723077774, 'timing/model': 1.0833560889586806, 'learning_rate/qwen_vl_interface': 5.660398891027129e-06, 'learning_rate/action_model': 5.660398891027129e-05, 'epoch': 1.23}) +[2026-06-17 23:17:26,413][starVLA.training.train_starvla][INFO] - Step 3361, Loss: {'train/loss': 0.7120676040649414, 'train/grad_norm_pre_clip': 17.036760330200195, 'timing/data': 0.0010537467896938324, 'timing/model': 1.1142916399985552, 'learning_rate/qwen_vl_interface': 5.658244245245399e-06, 'learning_rate/action_model': 5.658244245245399e-05, 'epoch': 1.23}) +[2026-06-17 23:17:33,771][starVLA.training.train_starvla][INFO] - Step 3362, Loss: {'train/loss': 0.43298113346099854, 'train/grad_norm_pre_clip': 12.721007347106934, 'timing/data': 0.0010942835360765457, 'timing/model': 4.060319083742797, 'learning_rate/qwen_vl_interface': 5.6560895148341985e-06, 'learning_rate/action_model': 5.656089514834199e-05, 'epoch': 1.23}) +[2026-06-17 23:17:46,614][starVLA.training.train_starvla][INFO] - Step 3363, Loss: {'train/loss': 0.4978894591331482, 'train/grad_norm_pre_clip': 37.013633728027344, 'timing/data': 0.0013073096051812172, 'timing/model': 1.1570716528221965, 'learning_rate/qwen_vl_interface': 5.653934700240204e-06, 'learning_rate/action_model': 5.6539347002402034e-05, 'epoch': 1.23}) +[2026-06-17 23:17:51,643][starVLA.training.train_starvla][INFO] - Step 3364, Loss: {'train/loss': 0.5029801726341248, 'train/grad_norm_pre_clip': 18.89131736755371, 'timing/data': 0.0008261632174253464, 'timing/model': 1.4947133669629693, 'learning_rate/qwen_vl_interface': 5.651779801910111e-06, 'learning_rate/action_model': 5.651779801910111e-05, 'epoch': 1.23}) +[2026-06-17 23:17:59,249][starVLA.training.train_starvla][INFO] - Step 3365, Loss: {'train/loss': 0.7631559371948242, 'train/grad_norm_pre_clip': 21.21340560913086, 'timing/data': 0.0015551317483186722, 'timing/model': 1.1289762146770954, 'learning_rate/qwen_vl_interface': 5.649624820290631e-06, 'learning_rate/action_model': 5.649624820290631e-05, 'epoch': 1.23}) +[2026-06-17 23:18:08,429][starVLA.training.train_starvla][INFO] - Step 3366, Loss: {'train/loss': 0.5727739334106445, 'train/grad_norm_pre_clip': 16.72358512878418, 'timing/data': 0.00104602612555027, 'timing/model': 5.6199062410742044, 'learning_rate/qwen_vl_interface': 5.647469755828501e-06, 'learning_rate/action_model': 5.647469755828501e-05, 'epoch': 1.23}) +[2026-06-17 23:18:15,609][starVLA.training.train_starvla][INFO] - Step 3367, Loss: {'train/loss': 0.5226722955703735, 'train/grad_norm_pre_clip': 17.529136657714844, 'timing/data': 0.0011584749445319176, 'timing/model': 1.6539198802784085, 'learning_rate/qwen_vl_interface': 5.645314608970457e-06, 'learning_rate/action_model': 5.6453146089704566e-05, 'epoch': 1.23}) +[2026-06-17 23:18:25,869][starVLA.training.train_starvla][INFO] - Step 3368, Loss: {'train/loss': 0.4354396164417267, 'train/grad_norm_pre_clip': 15.10862922668457, 'timing/data': 0.0011315913870930672, 'timing/model': 1.1312701348215342, 'learning_rate/qwen_vl_interface': 5.643159380163273e-06, 'learning_rate/action_model': 5.643159380163273e-05, 'epoch': 1.23}) +[2026-06-17 23:18:30,724][starVLA.training.train_starvla][INFO] - Step 3369, Loss: {'train/loss': 0.4270866811275482, 'train/grad_norm_pre_clip': 14.760004043579102, 'timing/data': 0.0007868073880672455, 'timing/model': 1.3790932819247246, 'learning_rate/qwen_vl_interface': 5.641004069853725e-06, 'learning_rate/action_model': 5.641004069853725e-05, 'epoch': 1.23}) +[2026-06-17 23:18:34,824][starVLA.training.train_starvla][INFO] - Step 3370, Loss: {'train/loss': 0.30061963200569153, 'train/grad_norm_pre_clip': 11.361618041992188, 'timing/data': 0.000983712263405323, 'timing/model': 1.0676383972167969, 'learning_rate/qwen_vl_interface': 5.638848678488613e-06, 'learning_rate/action_model': 5.638848678488613e-05, 'epoch': 1.23}) +[2026-06-17 23:18:41,625][starVLA.training.train_starvla][INFO] - Step 3371, Loss: {'train/loss': 0.4659384787082672, 'train/grad_norm_pre_clip': 21.782546997070312, 'timing/data': 0.0010595601052045822, 'timing/model': 3.293150845915079, 'learning_rate/qwen_vl_interface': 5.636693206514752e-06, 'learning_rate/action_model': 5.636693206514752e-05, 'epoch': 1.23}) +[2026-06-17 23:18:52,737][starVLA.training.train_starvla][INFO] - Step 3372, Loss: {'train/loss': 0.43667495250701904, 'train/grad_norm_pre_clip': 18.081947326660156, 'timing/data': 0.0032913051545619965, 'timing/model': 5.1331121968105435, 'learning_rate/qwen_vl_interface': 5.634537654378973e-06, 'learning_rate/action_model': 5.634537654378973e-05, 'epoch': 1.23}) +[2026-06-17 23:18:57,216][starVLA.training.train_starvla][INFO] - Step 3373, Loss: {'train/loss': 0.5100326538085938, 'train/grad_norm_pre_clip': 15.171770095825195, 'timing/data': 0.000962194986641407, 'timing/model': 1.076740981079638, 'learning_rate/qwen_vl_interface': 5.632382022528123e-06, 'learning_rate/action_model': 5.632382022528123e-05, 'epoch': 1.24}) +[2026-06-17 23:19:01,624][starVLA.training.train_starvla][INFO] - Step 3374, Loss: {'train/loss': 0.38429683446884155, 'train/grad_norm_pre_clip': 20.663124084472656, 'timing/data': 0.0010334569960832596, 'timing/model': 1.4171702209860086, 'learning_rate/qwen_vl_interface': 5.6302263114090705e-06, 'learning_rate/action_model': 5.63022631140907e-05, 'epoch': 1.24}) +[2026-06-17 23:19:05,719][starVLA.training.train_starvla][INFO] - Step 3375, Loss: {'train/loss': 0.2578577697277069, 'train/grad_norm_pre_clip': 10.197835922241211, 'timing/data': 0.0010966211557388306, 'timing/model': 1.055353743955493, 'learning_rate/qwen_vl_interface': 5.628070521468693e-06, 'learning_rate/action_model': 5.628070521468692e-05, 'epoch': 1.24}) +[2026-06-17 23:19:11,666][starVLA.training.train_starvla][INFO] - Step 3376, Loss: {'train/loss': 0.663735032081604, 'train/grad_norm_pre_clip': 16.972490310668945, 'timing/data': 0.0008784737437963486, 'timing/model': 1.173053439706564, 'learning_rate/qwen_vl_interface': 5.625914653153889e-06, 'learning_rate/action_model': 5.625914653153889e-05, 'epoch': 1.24}) +[2026-06-17 23:19:17,899][starVLA.training.train_starvla][INFO] - Step 3377, Loss: {'train/loss': 0.8904384970664978, 'train/grad_norm_pre_clip': 19.1519775390625, 'timing/data': 0.000911567360162735, 'timing/model': 1.047370694577694, 'learning_rate/qwen_vl_interface': 5.623758706911574e-06, 'learning_rate/action_model': 5.6237587069115737e-05, 'epoch': 1.24}) +[2026-06-17 23:19:22,506][starVLA.training.train_starvla][INFO] - Step 3378, Loss: {'train/loss': 0.5805477499961853, 'train/grad_norm_pre_clip': 26.20761489868164, 'timing/data': 0.0011791568249464035, 'timing/model': 1.0368502596393228, 'learning_rate/qwen_vl_interface': 5.621602683188677e-06, 'learning_rate/action_model': 5.621602683188677e-05, 'epoch': 1.24}) +[2026-06-17 23:19:26,728][starVLA.training.train_starvla][INFO] - Step 3379, Loss: {'train/loss': 0.7739945650100708, 'train/grad_norm_pre_clip': 17.735315322875977, 'timing/data': 0.0015048766508698463, 'timing/model': 1.1038465593010187, 'learning_rate/qwen_vl_interface': 5.619446582432143e-06, 'learning_rate/action_model': 5.619446582432143e-05, 'epoch': 1.24}) +[2026-06-17 23:19:31,227][starVLA.training.train_starvla][INFO] - Step 3380, Loss: {'train/loss': 0.6634761691093445, 'train/grad_norm_pre_clip': 17.955020904541016, 'timing/data': 0.0009775366634130478, 'timing/model': 1.0313946632668376, 'learning_rate/qwen_vl_interface': 5.617290405088936e-06, 'learning_rate/action_model': 5.617290405088936e-05, 'epoch': 1.24}) +[2026-06-17 23:19:35,771][starVLA.training.train_starvla][INFO] - Step 3381, Loss: {'train/loss': 0.7712177634239197, 'train/grad_norm_pre_clip': 24.356952667236328, 'timing/data': 0.0012704506516456604, 'timing/model': 1.0572864888235927, 'learning_rate/qwen_vl_interface': 5.615134151606033e-06, 'learning_rate/action_model': 5.615134151606033e-05, 'epoch': 1.24}) +[2026-06-17 23:19:40,400][starVLA.training.train_starvla][INFO] - Step 3382, Loss: {'train/loss': 0.6137498021125793, 'train/grad_norm_pre_clip': 17.539730072021484, 'timing/data': 0.0007889373227953911, 'timing/model': 1.1664355117827654, 'learning_rate/qwen_vl_interface': 5.612977822430425e-06, 'learning_rate/action_model': 5.612977822430425e-05, 'epoch': 1.24}) +[2026-06-17 23:19:44,960][starVLA.training.train_starvla][INFO] - Step 3383, Loss: {'train/loss': 0.9040598273277283, 'train/grad_norm_pre_clip': 19.136850357055664, 'timing/data': 0.0013134973123669624, 'timing/model': 1.1525497548282146, 'learning_rate/qwen_vl_interface': 5.610821418009127e-06, 'learning_rate/action_model': 5.610821418009127e-05, 'epoch': 1.24}) +[2026-06-17 23:19:49,893][starVLA.training.train_starvla][INFO] - Step 3384, Loss: {'train/loss': 0.6822784543037415, 'train/grad_norm_pre_clip': 23.255783081054688, 'timing/data': 0.0019909245893359184, 'timing/model': 1.4624508386477828, 'learning_rate/qwen_vl_interface': 5.608664938789159e-06, 'learning_rate/action_model': 5.608664938789159e-05, 'epoch': 1.24}) +[2026-06-17 23:19:56,373][starVLA.training.train_starvla][INFO] - Step 3385, Loss: {'train/loss': 0.5833160877227783, 'train/grad_norm_pre_clip': 16.131227493286133, 'timing/data': 0.0009230952709913254, 'timing/model': 1.1458425065502524, 'learning_rate/qwen_vl_interface': 5.606508385217564e-06, 'learning_rate/action_model': 5.6065083852175636e-05, 'epoch': 1.24}) +[2026-06-17 23:20:00,888][starVLA.training.train_starvla][INFO] - Step 3386, Loss: {'train/loss': 0.5317409634590149, 'train/grad_norm_pre_clip': 14.273816108703613, 'timing/data': 0.0014076493680477142, 'timing/model': 1.0452535497024655, 'learning_rate/qwen_vl_interface': 5.604351757741396e-06, 'learning_rate/action_model': 5.6043517577413964e-05, 'epoch': 1.24}) +[2026-06-17 23:20:05,498][starVLA.training.train_starvla][INFO] - Step 3387, Loss: {'train/loss': 0.2648327648639679, 'train/grad_norm_pre_clip': 11.719472885131836, 'timing/data': 0.0008858619257807732, 'timing/model': 1.0743609480559826, 'learning_rate/qwen_vl_interface': 5.602195056807727e-06, 'learning_rate/action_model': 5.602195056807727e-05, 'epoch': 1.24}) +[2026-06-17 23:20:09,972][starVLA.training.train_starvla][INFO] - Step 3388, Loss: {'train/loss': 0.2039669007062912, 'train/grad_norm_pre_clip': 7.858937740325928, 'timing/data': 0.0007895892485976219, 'timing/model': 1.0870647514238954, 'learning_rate/qwen_vl_interface': 5.600038282863647e-06, 'learning_rate/action_model': 5.6000382828636466e-05, 'epoch': 1.24}) +[2026-06-17 23:20:14,401][starVLA.training.train_starvla][INFO] - Step 3389, Loss: {'train/loss': 0.33194682002067566, 'train/grad_norm_pre_clip': 11.989864349365234, 'timing/data': 0.0012101056054234505, 'timing/model': 1.3934090975672007, 'learning_rate/qwen_vl_interface': 5.597881436356252e-06, 'learning_rate/action_model': 5.597881436356252e-05, 'epoch': 1.24}) +[2026-06-17 23:20:18,727][starVLA.training.train_starvla][INFO] - Step 3390, Loss: {'train/loss': 0.4025075137615204, 'train/grad_norm_pre_clip': 14.766827583312988, 'timing/data': 0.0009310208261013031, 'timing/model': 1.0545080406591296, 'learning_rate/qwen_vl_interface': 5.595724517732661e-06, 'learning_rate/action_model': 5.595724517732661e-05, 'epoch': 1.24}) +[2026-06-17 23:20:23,305][starVLA.training.train_starvla][INFO] - Step 3391, Loss: {'train/loss': 0.7440951466560364, 'train/grad_norm_pre_clip': 21.28145980834961, 'timing/data': 0.0012782420963048935, 'timing/model': 1.0487728891894221, 'learning_rate/qwen_vl_interface': 5.5935675274400046e-06, 'learning_rate/action_model': 5.593567527440005e-05, 'epoch': 1.24}) +[2026-06-17 23:20:27,811][starVLA.training.train_starvla][INFO] - Step 3392, Loss: {'train/loss': 0.32636964321136475, 'train/grad_norm_pre_clip': 13.867769241333008, 'timing/data': 0.0010615037754178047, 'timing/model': 1.093428373336792, 'learning_rate/qwen_vl_interface': 5.591410465925431e-06, 'learning_rate/action_model': 5.5914104659254306e-05, 'epoch': 1.24}) +[2026-06-17 23:20:32,457][starVLA.training.train_starvla][INFO] - Step 3393, Loss: {'train/loss': 0.41829630732536316, 'train/grad_norm_pre_clip': 15.490880012512207, 'timing/data': 0.0008800271898508072, 'timing/model': 1.0882130963727832, 'learning_rate/qwen_vl_interface': 5.589253333636101e-06, 'learning_rate/action_model': 5.589253333636101e-05, 'epoch': 1.24}) +[2026-06-17 23:20:38,490][starVLA.training.train_starvla][INFO] - Step 3394, Loss: {'train/loss': 0.29478588700294495, 'train/grad_norm_pre_clip': 9.925541877746582, 'timing/data': 0.001111130230128765, 'timing/model': 2.7009315695613623, 'learning_rate/qwen_vl_interface': 5.587096131019189e-06, 'learning_rate/action_model': 5.5870961310191896e-05, 'epoch': 1.24}) +[2026-06-17 23:20:42,690][starVLA.training.train_starvla][INFO] - Step 3395, Loss: {'train/loss': 0.8667625188827515, 'train/grad_norm_pre_clip': 17.90966796875, 'timing/data': 0.0012591220438480377, 'timing/model': 1.0408882638439536, 'learning_rate/qwen_vl_interface': 5.5849388585218865e-06, 'learning_rate/action_model': 5.5849388585218866e-05, 'epoch': 1.24}) +[2026-06-17 23:20:47,448][starVLA.training.train_starvla][INFO] - Step 3396, Loss: {'train/loss': 0.5373749136924744, 'train/grad_norm_pre_clip': 14.230712890625, 'timing/data': 0.0011102408170700073, 'timing/model': 1.0851567015051842, 'learning_rate/qwen_vl_interface': 5.582781516591398e-06, 'learning_rate/action_model': 5.5827815165913977e-05, 'epoch': 1.24}) +[2026-06-17 23:20:51,909][starVLA.training.train_starvla][INFO] - Step 3397, Loss: {'train/loss': 0.37506774067878723, 'train/grad_norm_pre_clip': 13.81412124633789, 'timing/data': 0.0009552612900733948, 'timing/model': 1.0354820154607296, 'learning_rate/qwen_vl_interface': 5.580624105674945e-06, 'learning_rate/action_model': 5.580624105674944e-05, 'epoch': 1.24}) +[2026-06-17 23:20:56,525][starVLA.training.train_starvla][INFO] - Step 3398, Loss: {'train/loss': 0.6420846581459045, 'train/grad_norm_pre_clip': 17.041349411010742, 'timing/data': 0.0011112997308373451, 'timing/model': 1.097816564142704, 'learning_rate/qwen_vl_interface': 5.578466626219756e-06, 'learning_rate/action_model': 5.578466626219756e-05, 'epoch': 1.24}) +[2026-06-17 23:21:01,179][starVLA.training.train_starvla][INFO] - Step 3399, Loss: {'train/loss': 0.6499062776565552, 'train/grad_norm_pre_clip': 24.6982421875, 'timing/data': 0.0011170711368322372, 'timing/model': 1.5448521226644516, 'learning_rate/qwen_vl_interface': 5.576309078673083e-06, 'learning_rate/action_model': 5.576309078673083e-05, 'epoch': 1.24}) +[2026-06-17 23:24:59,673][starVLA.training.train_starvla][INFO] - Step 3400, Loss: {'train/loss': 0.5842441320419312, 'train/grad_norm_pre_clip': 17.93029022216797, 'eval/loss': 0.6206614971160889, 'eval/latency_2/loss': 0.6206614971160889, 'eval/demon_fire_precision': 0.8722367928062945, 'eval/demon_fire_recall': 0.9740585774058578, 'eval/demon_fire_f1': 0.9203399881399487, 'eval/demon_fire_rate_model': 0.7426265998887034, 'eval/demon_fire_rate_teacher': 0.6649972175848636, 'eval/demon_left_precision': 0.8456591639871383, 'eval/demon_left_recall': 0.8305263157894737, 'eval/demon_left_f1': 0.838024429102496, 'eval/demon_left_rate_model': 0.2595993322203673, 'eval/demon_left_rate_teacher': 0.2643294379521425, 'eval/demon_right_precision': 0.8678090575275398, 'eval/demon_right_recall': 0.8011299435028248, 'eval/demon_right_f1': 0.8331374853113984, 'eval/demon_right_rate_model': 0.2273233166388425, 'eval/demon_right_rate_teacher': 0.24624373956594323, 'eval/demon_move_f1': 0.8355809572069473, 'eval/demon_cc_f1': 0.8931600786129322, 'eval/latency_2/demon_fire_precision': 0.8722367928062945, 'eval/latency_2/demon_fire_recall': 0.9740585774058578, 'eval/latency_2/demon_fire_f1': 0.9203399881399487, 'eval/latency_2/demon_fire_rate_model': 0.7426265998887034, 'eval/latency_2/demon_fire_rate_teacher': 0.6649972175848636, 'eval/latency_2/demon_left_precision': 0.8456591639871383, 'eval/latency_2/demon_left_recall': 0.8305263157894737, 'eval/latency_2/demon_left_f1': 0.838024429102496, 'eval/latency_2/demon_left_rate_model': 0.2595993322203673, 'eval/latency_2/demon_left_rate_teacher': 0.2643294379521425, 'eval/latency_2/demon_right_precision': 0.8678090575275398, 'eval/latency_2/demon_right_recall': 0.8011299435028248, 'eval/latency_2/demon_right_f1': 0.8331374853113984, 'eval/latency_2/demon_right_rate_model': 0.2273233166388425, 'eval/latency_2/demon_right_rate_teacher': 0.24624373956594323, 'eval/latency_2/demon_move_f1': 0.8355809572069473, 'eval/latency_2/demon_cc_f1': 0.8931600786129322, 'timing/data': 0.0026130815967917442, 'timing/model': 1.0780670894309878, 'learning_rate/qwen_vl_interface': 5.5741514634821885e-06, 'learning_rate/action_model': 5.574151463482188e-05, 'epoch': 1.25}) +[2026-06-17 23:24:59,745][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_3400_state +[2026-06-17 23:24:59,753][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 23:25:08,662][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_3400_state/model.safetensors +[2026-06-17 23:25:23,412][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_3400_state/optimizer.bin +[2026-06-17 23:25:23,413][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_3400_state/sampler.bin +[2026-06-17 23:25:23,413][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_3400_state/sampler_1.bin +[2026-06-17 23:25:23,416][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_3400_state/random_states_0.pkl +[2026-06-17 23:25:23,420][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3400 +[2026-06-17 23:25:23,421][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 23:25:23,438][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 23:25:27,586][starVLA.training.train_starvla][INFO] - Step 3401, Loss: {'train/loss': 0.6232869029045105, 'train/grad_norm_pre_clip': 21.746952056884766, 'timing/data': 0.0011114263907074928, 'timing/model': 1.043400821276009, 'learning_rate/qwen_vl_interface': 5.571993781094345e-06, 'learning_rate/action_model': 5.5719937810943444e-05, 'epoch': 1.25}) +[2026-06-17 23:25:32,276][starVLA.training.train_starvla][INFO] - Step 3402, Loss: {'train/loss': 0.36093664169311523, 'train/grad_norm_pre_clip': 15.331525802612305, 'timing/data': 0.0011024698615074158, 'timing/model': 1.0941552799195051, 'learning_rate/qwen_vl_interface': 5.569836031956845e-06, 'learning_rate/action_model': 5.5698360319568454e-05, 'epoch': 1.25}) +[2026-06-17 23:25:36,865][starVLA.training.train_starvla][INFO] - Step 3403, Loss: {'train/loss': 0.5432655215263367, 'train/grad_norm_pre_clip': 27.72136116027832, 'timing/data': 0.0011566299945116043, 'timing/model': 1.1069429032504559, 'learning_rate/qwen_vl_interface': 5.5676782165169915e-06, 'learning_rate/action_model': 5.567678216516991e-05, 'epoch': 1.25}) +[2026-06-17 23:25:41,432][starVLA.training.train_starvla][INFO] - Step 3404, Loss: {'train/loss': 0.35052594542503357, 'train/grad_norm_pre_clip': 13.57892894744873, 'timing/data': 0.000912737101316452, 'timing/model': 1.0538559025153518, 'learning_rate/qwen_vl_interface': 5.565520335222101e-06, 'learning_rate/action_model': 5.565520335222101e-05, 'epoch': 1.25}) +[2026-06-17 23:25:46,162][starVLA.training.train_starvla][INFO] - Step 3405, Loss: {'train/loss': 0.6250693202018738, 'train/grad_norm_pre_clip': 16.35439109802246, 'timing/data': 0.0014635128900408745, 'timing/model': 1.6229264996945858, 'learning_rate/qwen_vl_interface': 5.563362388519507e-06, 'learning_rate/action_model': 5.5633623885195066e-05, 'epoch': 1.25}) +[2026-06-17 23:25:50,531][starVLA.training.train_starvla][INFO] - Step 3406, Loss: {'train/loss': 0.1626005321741104, 'train/grad_norm_pre_clip': 7.549354076385498, 'timing/data': 0.0011099539697170258, 'timing/model': 1.0842368807643652, 'learning_rate/qwen_vl_interface': 5.56120437685655e-06, 'learning_rate/action_model': 5.56120437685655e-05, 'epoch': 1.25}) +[2026-06-17 23:25:55,016][starVLA.training.train_starvla][INFO] - Step 3407, Loss: {'train/loss': 0.5475714802742004, 'train/grad_norm_pre_clip': 19.94013214111328, 'timing/data': 0.0010270094498991966, 'timing/model': 1.071585277095437, 'learning_rate/qwen_vl_interface': 5.559046300680593e-06, 'learning_rate/action_model': 5.5590463006805925e-05, 'epoch': 1.25}) +[2026-06-17 23:25:59,649][starVLA.training.train_starvla][INFO] - Step 3408, Loss: {'train/loss': 0.46084579825401306, 'train/grad_norm_pre_clip': 15.581469535827637, 'timing/data': 0.0014082053676247597, 'timing/model': 1.1441394472494721, 'learning_rate/qwen_vl_interface': 5.556888160439006e-06, 'learning_rate/action_model': 5.556888160439006e-05, 'epoch': 1.25}) +[2026-06-17 23:26:04,222][starVLA.training.train_starvla][INFO] - Step 3409, Loss: {'train/loss': 0.41298139095306396, 'train/grad_norm_pre_clip': 12.717796325683594, 'timing/data': 0.0011639362201094627, 'timing/model': 1.1095132622867823, 'learning_rate/qwen_vl_interface': 5.5547299565791725e-06, 'learning_rate/action_model': 5.5547299565791723e-05, 'epoch': 1.25}) +[2026-06-17 23:26:08,961][starVLA.training.train_starvla][INFO] - Step 3410, Loss: {'train/loss': 0.6142740845680237, 'train/grad_norm_pre_clip': 14.436492919921875, 'timing/data': 0.0012992704287171364, 'timing/model': 1.5839104866608977, 'learning_rate/qwen_vl_interface': 5.55257168954849e-06, 'learning_rate/action_model': 5.5525716895484905e-05, 'epoch': 1.25}) +[2026-06-17 23:26:16,423][starVLA.training.train_starvla][INFO] - Step 3411, Loss: {'train/loss': 0.7819427847862244, 'train/grad_norm_pre_clip': 18.337907791137695, 'timing/data': 0.0013688746839761734, 'timing/model': 4.272026544436812, 'learning_rate/qwen_vl_interface': 5.550413359794371e-06, 'learning_rate/action_model': 5.550413359794371e-05, 'epoch': 1.25}) +[2026-06-17 23:26:20,969][starVLA.training.train_starvla][INFO] - Step 3412, Loss: {'train/loss': 0.5639451146125793, 'train/grad_norm_pre_clip': 18.15673828125, 'timing/data': 0.0011531701311469078, 'timing/model': 1.0888343835249543, 'learning_rate/qwen_vl_interface': 5.548254967764239e-06, 'learning_rate/action_model': 5.5482549677642394e-05, 'epoch': 1.25}) +[2026-06-17 23:26:29,712][starVLA.training.train_starvla][INFO] - Step 3413, Loss: {'train/loss': 0.5611159801483154, 'train/grad_norm_pre_clip': 21.297931671142578, 'timing/data': 0.0010551093146204948, 'timing/model': 3.5671773981302977, 'learning_rate/qwen_vl_interface': 5.546096513905532e-06, 'learning_rate/action_model': 5.546096513905532e-05, 'epoch': 1.25}) +[2026-06-17 23:26:43,012][starVLA.training.train_starvla][INFO] - Step 3414, Loss: {'train/loss': 0.4768659174442291, 'train/grad_norm_pre_clip': 17.173185348510742, 'timing/data': 0.00122850202023983, 'timing/model': 3.558366207405925, 'learning_rate/qwen_vl_interface': 5.5439379986657e-06, 'learning_rate/action_model': 5.543937998665699e-05, 'epoch': 1.25}) +[2026-06-17 23:26:47,299][starVLA.training.train_starvla][INFO] - Step 3415, Loss: {'train/loss': 0.48260313272476196, 'train/grad_norm_pre_clip': 17.525156021118164, 'timing/data': 0.0015419572591781616, 'timing/model': 1.0868791295215487, 'learning_rate/qwen_vl_interface': 5.5417794224922025e-06, 'learning_rate/action_model': 5.5417794224922023e-05, 'epoch': 1.25}) +[2026-06-17 23:26:52,046][starVLA.training.train_starvla][INFO] - Step 3416, Loss: {'train/loss': 0.3459031283855438, 'train/grad_norm_pre_clip': 12.012332916259766, 'timing/data': 0.000900978222489357, 'timing/model': 1.0390608720481396, 'learning_rate/qwen_vl_interface': 5.539620785832517e-06, 'learning_rate/action_model': 5.5396207858325166e-05, 'epoch': 1.25}) +[2026-06-17 23:26:56,711][starVLA.training.train_starvla][INFO] - Step 3417, Loss: {'train/loss': 0.5873227119445801, 'train/grad_norm_pre_clip': 22.241199493408203, 'timing/data': 0.0012353500351309776, 'timing/model': 1.060456725768745, 'learning_rate/qwen_vl_interface': 5.537462089134133e-06, 'learning_rate/action_model': 5.537462089134132e-05, 'epoch': 1.25}) +[2026-06-17 23:27:01,210][starVLA.training.train_starvla][INFO] - Step 3418, Loss: {'train/loss': 0.4902401864528656, 'train/grad_norm_pre_clip': 17.9249324798584, 'timing/data': 0.0013282625004649162, 'timing/model': 1.0463331071659923, 'learning_rate/qwen_vl_interface': 5.535303332844546e-06, 'learning_rate/action_model': 5.5353033328445455e-05, 'epoch': 1.25}) +[2026-06-17 23:27:05,688][starVLA.training.train_starvla][INFO] - Step 3419, Loss: {'train/loss': 0.5340670347213745, 'train/grad_norm_pre_clip': 18.123615264892578, 'timing/data': 0.0009241541847586632, 'timing/model': 1.1126838428899646, 'learning_rate/qwen_vl_interface': 5.533144517411273e-06, 'learning_rate/action_model': 5.533144517411273e-05, 'epoch': 1.25}) +[2026-06-17 23:27:10,216][starVLA.training.train_starvla][INFO] - Step 3420, Loss: {'train/loss': 0.5523690581321716, 'train/grad_norm_pre_clip': 17.611446380615234, 'timing/data': 0.0013556107878684998, 'timing/model': 1.3979834308847785, 'learning_rate/qwen_vl_interface': 5.530985643281834e-06, 'learning_rate/action_model': 5.530985643281834e-05, 'epoch': 1.25}) +[2026-06-17 23:27:14,625][starVLA.training.train_starvla][INFO] - Step 3421, Loss: {'train/loss': 0.6599498391151428, 'train/grad_norm_pre_clip': 24.96527862548828, 'timing/data': 0.0034027667716145515, 'timing/model': 1.0712292520329356, 'learning_rate/qwen_vl_interface': 5.528826710903774e-06, 'learning_rate/action_model': 5.528826710903774e-05, 'epoch': 1.25}) +[2026-06-17 23:27:19,262][starVLA.training.train_starvla][INFO] - Step 3422, Loss: {'train/loss': 0.30141907930374146, 'train/grad_norm_pre_clip': 12.951566696166992, 'timing/data': 0.0011182939633727074, 'timing/model': 1.0718373283743858, 'learning_rate/qwen_vl_interface': 5.526667720724632e-06, 'learning_rate/action_model': 5.526667720724632e-05, 'epoch': 1.25}) +[2026-06-17 23:27:23,895][starVLA.training.train_starvla][INFO] - Step 3423, Loss: {'train/loss': 0.5206497311592102, 'train/grad_norm_pre_clip': 17.828123092651367, 'timing/data': 0.0014443621039390564, 'timing/model': 1.0781262954697013, 'learning_rate/qwen_vl_interface': 5.5245086731919764e-06, 'learning_rate/action_model': 5.524508673191976e-05, 'epoch': 1.25}) +[2026-06-17 23:27:28,481][starVLA.training.train_starvla][INFO] - Step 3424, Loss: {'train/loss': 0.5351437330245972, 'train/grad_norm_pre_clip': 22.244888305664062, 'timing/data': 0.0008777668699622154, 'timing/model': 1.0556468674913049, 'learning_rate/qwen_vl_interface': 5.522349568753376e-06, 'learning_rate/action_model': 5.522349568753377e-05, 'epoch': 1.25}) +[2026-06-17 23:27:32,574][starVLA.training.train_starvla][INFO] - Step 3425, Loss: {'train/loss': 0.36228811740875244, 'train/grad_norm_pre_clip': 13.536145210266113, 'timing/data': 0.0012696152552962303, 'timing/model': 1.0407620444893837, 'learning_rate/qwen_vl_interface': 5.520190407856419e-06, 'learning_rate/action_model': 5.5201904078564185e-05, 'epoch': 1.25}) +[2026-06-17 23:27:37,134][starVLA.training.train_starvla][INFO] - Step 3426, Loss: {'train/loss': 0.34651994705200195, 'train/grad_norm_pre_clip': 12.052248001098633, 'timing/data': 0.0009543737396597862, 'timing/model': 1.0659041842445731, 'learning_rate/qwen_vl_interface': 5.5180311909486974e-06, 'learning_rate/action_model': 5.518031190948697e-05, 'epoch': 1.25}) +[2026-06-17 23:27:42,144][starVLA.training.train_starvla][INFO] - Step 3427, Loss: {'train/loss': 0.357234388589859, 'train/grad_norm_pre_clip': 15.210427284240723, 'timing/data': 0.0012216102331876755, 'timing/model': 1.1415611896663904, 'learning_rate/qwen_vl_interface': 5.515871918477821e-06, 'learning_rate/action_model': 5.515871918477821e-05, 'epoch': 1.25}) +[2026-06-17 23:27:46,567][starVLA.training.train_starvla][INFO] - Step 3428, Loss: {'train/loss': 0.552148699760437, 'train/grad_norm_pre_clip': 15.941652297973633, 'timing/data': 0.0008078282698988914, 'timing/model': 1.03078637085855, 'learning_rate/qwen_vl_interface': 5.513712590891414e-06, 'learning_rate/action_model': 5.513712590891413e-05, 'epoch': 1.26}) +[2026-06-17 23:27:51,363][starVLA.training.train_starvla][INFO] - Step 3429, Loss: {'train/loss': 0.3789685368537903, 'train/grad_norm_pre_clip': 27.418018341064453, 'timing/data': 0.0013298671692609787, 'timing/model': 1.5191673440858722, 'learning_rate/qwen_vl_interface': 5.511553208637098e-06, 'learning_rate/action_model': 5.511553208637098e-05, 'epoch': 1.26}) +[2026-06-17 23:27:57,393][starVLA.training.train_starvla][INFO] - Step 3430, Loss: {'train/loss': 0.6286295056343079, 'train/grad_norm_pre_clip': 17.83468246459961, 'timing/data': 0.0008930880576372147, 'timing/model': 1.4113574279472232, 'learning_rate/qwen_vl_interface': 5.509393772162524e-06, 'learning_rate/action_model': 5.509393772162523e-05, 'epoch': 1.26}) +[2026-06-17 23:28:01,720][starVLA.training.train_starvla][INFO] - Step 3431, Loss: {'train/loss': 0.4880504906177521, 'train/grad_norm_pre_clip': 16.944883346557617, 'timing/data': 0.0008873408660292625, 'timing/model': 1.1216808520257473, 'learning_rate/qwen_vl_interface': 5.50723428191534e-06, 'learning_rate/action_model': 5.50723428191534e-05, 'epoch': 1.26}) +[2026-06-17 23:28:06,326][starVLA.training.train_starvla][INFO] - Step 3432, Loss: {'train/loss': 0.5910765528678894, 'train/grad_norm_pre_clip': 22.203447341918945, 'timing/data': 0.0007984116673469543, 'timing/model': 1.043634812347591, 'learning_rate/qwen_vl_interface': 5.505074738343212e-06, 'learning_rate/action_model': 5.505074738343212e-05, 'epoch': 1.26}) +[2026-06-17 23:28:11,215][starVLA.training.train_starvla][INFO] - Step 3433, Loss: {'train/loss': 0.5325800180435181, 'train/grad_norm_pre_clip': 18.652019500732422, 'timing/data': 0.0009278161451220512, 'timing/model': 1.187430647201836, 'learning_rate/qwen_vl_interface': 5.502915141893819e-06, 'learning_rate/action_model': 5.5029151418938184e-05, 'epoch': 1.26}) +[2026-06-17 23:28:16,261][starVLA.training.train_starvla][INFO] - Step 3434, Loss: {'train/loss': 0.33117473125457764, 'train/grad_norm_pre_clip': 12.661500930786133, 'timing/data': 0.0012729046866297722, 'timing/model': 1.2845355169847608, 'learning_rate/qwen_vl_interface': 5.5007554930148436e-06, 'learning_rate/action_model': 5.5007554930148434e-05, 'epoch': 1.26}) +[2026-06-17 23:28:20,793][starVLA.training.train_starvla][INFO] - Step 3435, Loss: {'train/loss': 0.6083976626396179, 'train/grad_norm_pre_clip': 17.625080108642578, 'timing/data': 0.00148804672062397, 'timing/model': 1.3940385188907385, 'learning_rate/qwen_vl_interface': 5.498595792153985e-06, 'learning_rate/action_model': 5.498595792153984e-05, 'epoch': 1.26}) +[2026-06-17 23:28:25,089][starVLA.training.train_starvla][INFO] - Step 3436, Loss: {'train/loss': 0.2226749211549759, 'train/grad_norm_pre_clip': 9.708686828613281, 'timing/data': 0.0009275535121560097, 'timing/model': 1.2071140930056572, 'learning_rate/qwen_vl_interface': 5.49643603975895e-06, 'learning_rate/action_model': 5.4964360397589495e-05, 'epoch': 1.26}) +[2026-06-17 23:28:30,002][starVLA.training.train_starvla][INFO] - Step 3437, Loss: {'train/loss': 0.5069538950920105, 'train/grad_norm_pre_clip': 20.154075622558594, 'timing/data': 0.0013777147978544235, 'timing/model': 1.1980428146198392, 'learning_rate/qwen_vl_interface': 5.494276236277458e-06, 'learning_rate/action_model': 5.4942762362774584e-05, 'epoch': 1.26}) +[2026-06-17 23:28:34,396][starVLA.training.train_starvla][INFO] - Step 3438, Loss: {'train/loss': 0.4193234145641327, 'train/grad_norm_pre_clip': 12.239946365356445, 'timing/data': 0.000908583402633667, 'timing/model': 1.0456299195066094, 'learning_rate/qwen_vl_interface': 5.492116382157242e-06, 'learning_rate/action_model': 5.492116382157242e-05, 'epoch': 1.26}) +[2026-06-17 23:28:38,793][starVLA.training.train_starvla][INFO] - Step 3439, Loss: {'train/loss': 0.4495733976364136, 'train/grad_norm_pre_clip': 17.634098052978516, 'timing/data': 0.0010601207613945007, 'timing/model': 1.0407465482130647, 'learning_rate/qwen_vl_interface': 5.489956477846038e-06, 'learning_rate/action_model': 5.489956477846038e-05, 'epoch': 1.26}) +[2026-06-17 23:28:43,268][starVLA.training.train_starvla][INFO] - Step 3440, Loss: {'train/loss': 0.4328780174255371, 'train/grad_norm_pre_clip': 80.67207336425781, 'timing/data': 0.0010912185534834862, 'timing/model': 1.369255200959742, 'learning_rate/qwen_vl_interface': 5.4877965237915984e-06, 'learning_rate/action_model': 5.4877965237915986e-05, 'epoch': 1.26}) +[2026-06-17 23:28:47,427][starVLA.training.train_starvla][INFO] - Step 3441, Loss: {'train/loss': 0.4236253798007965, 'train/grad_norm_pre_clip': 16.53207015991211, 'timing/data': 0.0011564139276742935, 'timing/model': 1.0493564242497087, 'learning_rate/qwen_vl_interface': 5.485636520441684e-06, 'learning_rate/action_model': 5.4856365204416836e-05, 'epoch': 1.26}) +[2026-06-17 23:28:52,113][starVLA.training.train_starvla][INFO] - Step 3442, Loss: {'train/loss': 0.42897704243659973, 'train/grad_norm_pre_clip': 13.349733352661133, 'timing/data': 0.0012651728466153145, 'timing/model': 1.0702460100874305, 'learning_rate/qwen_vl_interface': 5.483476468244063e-06, 'learning_rate/action_model': 5.483476468244063e-05, 'epoch': 1.26}) +[2026-06-17 23:28:56,600][starVLA.training.train_starvla][INFO] - Step 3443, Loss: {'train/loss': 0.7125298380851746, 'train/grad_norm_pre_clip': 23.123743057250977, 'timing/data': 0.0010101944208145142, 'timing/model': 1.0810502003878355, 'learning_rate/qwen_vl_interface': 5.481316367646522e-06, 'learning_rate/action_model': 5.4813163676465216e-05, 'epoch': 1.26}) +[2026-06-17 23:29:01,115][starVLA.training.train_starvla][INFO] - Step 3444, Loss: {'train/loss': 0.22492030262947083, 'train/grad_norm_pre_clip': 11.557355880737305, 'timing/data': 0.001067526638507843, 'timing/model': 1.0543400794267654, 'learning_rate/qwen_vl_interface': 5.479156219096847e-06, 'learning_rate/action_model': 5.479156219096847e-05, 'epoch': 1.26}) +[2026-06-17 23:29:12,065][starVLA.training.train_starvla][INFO] - Step 3445, Loss: {'train/loss': 0.3564276397228241, 'train/grad_norm_pre_clip': 12.740047454833984, 'timing/data': 0.0011883173137903214, 'timing/model': 4.564612558111548, 'learning_rate/qwen_vl_interface': 5.476996023042843e-06, 'learning_rate/action_model': 5.476996023042843e-05, 'epoch': 1.26}) +[2026-06-17 23:29:22,093][starVLA.training.train_starvla][INFO] - Step 3446, Loss: {'train/loss': 0.5612971782684326, 'train/grad_norm_pre_clip': 16.15758514404297, 'timing/data': 0.0009550563991069794, 'timing/model': 3.477901490405202, 'learning_rate/qwen_vl_interface': 5.474835779932317e-06, 'learning_rate/action_model': 5.474835779932317e-05, 'epoch': 1.26}) +[2026-06-17 23:29:26,764][starVLA.training.train_starvla][INFO] - Step 3447, Loss: {'train/loss': 0.7797604203224182, 'train/grad_norm_pre_clip': 21.24468994140625, 'timing/data': 0.0011755051091313362, 'timing/model': 1.0769381849095225, 'learning_rate/qwen_vl_interface': 5.472675490213093e-06, 'learning_rate/action_model': 5.472675490213093e-05, 'epoch': 1.26}) +[2026-06-17 23:29:31,317][starVLA.training.train_starvla][INFO] - Step 3448, Loss: {'train/loss': 0.2980872690677643, 'train/grad_norm_pre_clip': 14.327336311340332, 'timing/data': 0.001417430117726326, 'timing/model': 1.0990478666499257, 'learning_rate/qwen_vl_interface': 5.470515154333e-06, 'learning_rate/action_model': 5.470515154333e-05, 'epoch': 1.26}) +[2026-06-17 23:29:38,350][starVLA.training.train_starvla][INFO] - Step 3449, Loss: {'train/loss': 0.476272314786911, 'train/grad_norm_pre_clip': 22.672927856445312, 'timing/data': 0.0009924937039613724, 'timing/model': 1.477283289656043, 'learning_rate/qwen_vl_interface': 5.468354772739879e-06, 'learning_rate/action_model': 5.468354772739879e-05, 'epoch': 1.26}) +[2026-06-17 23:29:43,008][starVLA.training.train_starvla][INFO] - Step 3450, Loss: {'train/loss': 0.5916446447372437, 'train/grad_norm_pre_clip': 20.52146339416504, 'timing/data': 0.0008287206292152405, 'timing/model': 1.4620234351605177, 'learning_rate/qwen_vl_interface': 5.466194345881578e-06, 'learning_rate/action_model': 5.4661943458815776e-05, 'epoch': 1.26}) +[2026-06-17 23:29:47,221][starVLA.training.train_starvla][INFO] - Step 3451, Loss: {'train/loss': 0.39473673701286316, 'train/grad_norm_pre_clip': 15.496973991394043, 'timing/data': 0.0008523901924490929, 'timing/model': 1.0780721679329872, 'learning_rate/qwen_vl_interface': 5.4640338742059564e-06, 'learning_rate/action_model': 5.464033874205956e-05, 'epoch': 1.26}) +[2026-06-17 23:29:51,709][starVLA.training.train_starvla][INFO] - Step 3452, Loss: {'train/loss': 1.2130848169326782, 'train/grad_norm_pre_clip': 23.843183517456055, 'timing/data': 0.0010762996971607208, 'timing/model': 1.0455743819475174, 'learning_rate/qwen_vl_interface': 5.461873358160883e-06, 'learning_rate/action_model': 5.461873358160883e-05, 'epoch': 1.26}) +[2026-06-17 23:29:56,561][starVLA.training.train_starvla][INFO] - Step 3453, Loss: {'train/loss': 0.2770704925060272, 'train/grad_norm_pre_clip': 11.837292671203613, 'timing/data': 0.0013489527627825737, 'timing/model': 1.41500760987401, 'learning_rate/qwen_vl_interface': 5.459712798194234e-06, 'learning_rate/action_model': 5.459712798194233e-05, 'epoch': 1.26}) +[2026-06-17 23:30:02,137][starVLA.training.train_starvla][INFO] - Step 3454, Loss: {'train/loss': 0.7484780550003052, 'train/grad_norm_pre_clip': 20.379514694213867, 'timing/data': 0.0010698512196540833, 'timing/model': 2.0505940513685346, 'learning_rate/qwen_vl_interface': 5.457552194753896e-06, 'learning_rate/action_model': 5.457552194753895e-05, 'epoch': 1.26}) +[2026-06-17 23:30:07,102][starVLA.training.train_starvla][INFO] - Step 3455, Loss: {'train/loss': 0.518960177898407, 'train/grad_norm_pre_clip': 16.65941047668457, 'timing/data': 0.0012820251286029816, 'timing/model': 1.6623731795698404, 'learning_rate/qwen_vl_interface': 5.455391548287764e-06, 'learning_rate/action_model': 5.455391548287764e-05, 'epoch': 1.27}) +[2026-06-17 23:30:11,640][starVLA.training.train_starvla][INFO] - Step 3456, Loss: {'train/loss': 0.4027892053127289, 'train/grad_norm_pre_clip': 12.894946098327637, 'timing/data': 0.0012239236384630203, 'timing/model': 1.0522696720436215, 'learning_rate/qwen_vl_interface': 5.453230859243745e-06, 'learning_rate/action_model': 5.4532308592437444e-05, 'epoch': 1.27}) +[2026-06-17 23:30:16,180][starVLA.training.train_starvla][INFO] - Step 3457, Loss: {'train/loss': 0.6673256754875183, 'train/grad_norm_pre_clip': 18.85860252380371, 'timing/data': 0.0009815646335482597, 'timing/model': 1.0359611650928855, 'learning_rate/qwen_vl_interface': 5.451070128069748e-06, 'learning_rate/action_model': 5.4510701280697485e-05, 'epoch': 1.27}) +[2026-06-17 23:30:20,920][starVLA.training.train_starvla][INFO] - Step 3458, Loss: {'train/loss': 0.5255169868469238, 'train/grad_norm_pre_clip': 20.723207473754883, 'timing/data': 0.0011471938341856003, 'timing/model': 1.1736264619976282, 'learning_rate/qwen_vl_interface': 5.4489093552136986e-06, 'learning_rate/action_model': 5.4489093552136984e-05, 'epoch': 1.27}) +[2026-06-17 23:30:25,685][starVLA.training.train_starvla][INFO] - Step 3459, Loss: {'train/loss': 0.423208087682724, 'train/grad_norm_pre_clip': 18.112157821655273, 'timing/data': 0.0007647862657904625, 'timing/model': 1.0995727134868503, 'learning_rate/qwen_vl_interface': 5.446748541123526e-06, 'learning_rate/action_model': 5.4467485411235254e-05, 'epoch': 1.27}) +[2026-06-17 23:30:30,359][starVLA.training.train_starvla][INFO] - Step 3460, Loss: {'train/loss': 0.636634886264801, 'train/grad_norm_pre_clip': 20.56622886657715, 'timing/data': 0.0011067669838666916, 'timing/model': 1.5475360816344619, 'learning_rate/qwen_vl_interface': 5.444587686247168e-06, 'learning_rate/action_model': 5.4445876862471675e-05, 'epoch': 1.27}) +[2026-06-17 23:30:34,423][starVLA.training.train_starvla][INFO] - Step 3461, Loss: {'train/loss': 0.6439189314842224, 'train/grad_norm_pre_clip': 20.099077224731445, 'timing/data': 0.00110524520277977, 'timing/model': 1.0328333415091038, 'learning_rate/qwen_vl_interface': 5.442426791032574e-06, 'learning_rate/action_model': 5.4424267910325746e-05, 'epoch': 1.27}) +[2026-06-17 23:30:40,811][starVLA.training.train_starvla][INFO] - Step 3462, Loss: {'train/loss': 0.826802670955658, 'train/grad_norm_pre_clip': 19.206649780273438, 'timing/data': 0.0009140297770500183, 'timing/model': 1.0678717019036412, 'learning_rate/qwen_vl_interface': 5.4402658559277e-06, 'learning_rate/action_model': 5.4402658559276995e-05, 'epoch': 1.27}) +[2026-06-17 23:30:47,388][starVLA.training.train_starvla][INFO] - Step 3463, Loss: {'train/loss': 0.439894437789917, 'train/grad_norm_pre_clip': 16.120237350463867, 'timing/data': 0.0010338928550481796, 'timing/model': 1.0481974808499217, 'learning_rate/qwen_vl_interface': 5.438104881380507e-06, 'learning_rate/action_model': 5.438104881380507e-05, 'epoch': 1.27}) +[2026-06-17 23:30:52,048][starVLA.training.train_starvla][INFO] - Step 3464, Loss: {'train/loss': 0.4210231900215149, 'train/grad_norm_pre_clip': 16.94229507446289, 'timing/data': 0.001382860355079174, 'timing/model': 1.0327750416472554, 'learning_rate/qwen_vl_interface': 5.435943867838971e-06, 'learning_rate/action_model': 5.435943867838971e-05, 'epoch': 1.27}) +[2026-06-17 23:30:59,858][starVLA.training.train_starvla][INFO] - Step 3465, Loss: {'train/loss': 0.5857073664665222, 'train/grad_norm_pre_clip': 18.913131713867188, 'timing/data': 0.0009160572662949562, 'timing/model': 4.791496702469885, 'learning_rate/qwen_vl_interface': 5.433782815751072e-06, 'learning_rate/action_model': 5.433782815751072e-05, 'epoch': 1.27}) +[2026-06-17 23:31:04,076][starVLA.training.train_starvla][INFO] - Step 3466, Loss: {'train/loss': 0.69111567735672, 'train/grad_norm_pre_clip': 25.770601272583008, 'timing/data': 0.0009417813271284103, 'timing/model': 1.091304480098188, 'learning_rate/qwen_vl_interface': 5.431621725564797e-06, 'learning_rate/action_model': 5.4316217255647974e-05, 'epoch': 1.27}) +[2026-06-17 23:31:08,563][starVLA.training.train_starvla][INFO] - Step 3467, Loss: {'train/loss': 0.5998610258102417, 'train/grad_norm_pre_clip': 15.49902057647705, 'timing/data': 0.00140279158949852, 'timing/model': 1.105943851172924, 'learning_rate/qwen_vl_interface': 5.429460597728142e-06, 'learning_rate/action_model': 5.429460597728142e-05, 'epoch': 1.27}) +[2026-06-17 23:31:13,117][starVLA.training.train_starvla][INFO] - Step 3468, Loss: {'train/loss': 0.5649881958961487, 'train/grad_norm_pre_clip': 15.855118751525879, 'timing/data': 0.0015888139605522156, 'timing/model': 1.0441538225859404, 'learning_rate/qwen_vl_interface': 5.4272994326891114e-06, 'learning_rate/action_model': 5.427299432689111e-05, 'epoch': 1.27}) +[2026-06-17 23:31:17,849][starVLA.training.train_starvla][INFO] - Step 3469, Loss: {'train/loss': 0.8036512732505798, 'train/grad_norm_pre_clip': 22.129060745239258, 'timing/data': 0.0007708054035902023, 'timing/model': 1.4792224653065205, 'learning_rate/qwen_vl_interface': 5.425138230895719e-06, 'learning_rate/action_model': 5.425138230895719e-05, 'epoch': 1.27}) +[2026-06-17 23:31:22,654][starVLA.training.train_starvla][INFO] - Step 3470, Loss: {'train/loss': 0.4950527846813202, 'train/grad_norm_pre_clip': 16.449138641357422, 'timing/data': 0.0014128396287560463, 'timing/model': 1.5284037021920085, 'learning_rate/qwen_vl_interface': 5.422976992795984e-06, 'learning_rate/action_model': 5.422976992795984e-05, 'epoch': 1.27}) +[2026-06-17 23:31:33,867][starVLA.training.train_starvla][INFO] - Step 3471, Loss: {'train/loss': 0.2687133252620697, 'train/grad_norm_pre_clip': 14.984508514404297, 'timing/data': 0.0013403929769992828, 'timing/model': 1.107370525598526, 'learning_rate/qwen_vl_interface': 5.420815718837929e-06, 'learning_rate/action_model': 5.42081571883793e-05, 'epoch': 1.27}) +[2026-06-17 23:31:38,419][starVLA.training.train_starvla][INFO] - Step 3472, Loss: {'train/loss': 0.3781294524669647, 'train/grad_norm_pre_clip': 14.178231239318848, 'timing/data': 0.000999358482658863, 'timing/model': 1.1330980258062482, 'learning_rate/qwen_vl_interface': 5.418654409469595e-06, 'learning_rate/action_model': 5.4186544094695946e-05, 'epoch': 1.27}) +[2026-06-17 23:31:45,671][starVLA.training.train_starvla][INFO] - Step 3473, Loss: {'train/loss': 0.5175416469573975, 'train/grad_norm_pre_clip': 14.892269134521484, 'timing/data': 0.0009134812280535698, 'timing/model': 1.210765391588211, 'learning_rate/qwen_vl_interface': 5.416493065139019e-06, 'learning_rate/action_model': 5.4164930651390186e-05, 'epoch': 1.27}) +[2026-06-17 23:31:50,323][starVLA.training.train_starvla][INFO] - Step 3474, Loss: {'train/loss': 0.5199170112609863, 'train/grad_norm_pre_clip': 14.017749786376953, 'timing/data': 0.001910671591758728, 'timing/model': 1.1165943602100015, 'learning_rate/qwen_vl_interface': 5.4143316862942516e-06, 'learning_rate/action_model': 5.4143316862942514e-05, 'epoch': 1.27}) +[2026-06-17 23:31:59,078][starVLA.training.train_starvla][INFO] - Step 3475, Loss: {'train/loss': 0.5123847126960754, 'train/grad_norm_pre_clip': 17.617584228515625, 'timing/data': 0.0008727796375751495, 'timing/model': 4.980906217359006, 'learning_rate/qwen_vl_interface': 5.412170273383348e-06, 'learning_rate/action_model': 5.412170273383348e-05, 'epoch': 1.27}) +[2026-06-17 23:32:08,023][starVLA.training.train_starvla][INFO] - Step 3476, Loss: {'train/loss': 0.7405599355697632, 'train/grad_norm_pre_clip': 20.37213134765625, 'timing/data': 0.0007473072037100792, 'timing/model': 5.850326055660844, 'learning_rate/qwen_vl_interface': 5.410008826854374e-06, 'learning_rate/action_model': 5.410008826854373e-05, 'epoch': 1.27}) +[2026-06-17 23:32:15,981][starVLA.training.train_starvla][INFO] - Step 3477, Loss: {'train/loss': 0.40544945001602173, 'train/grad_norm_pre_clip': 16.772964477539062, 'timing/data': 0.0008839322254061699, 'timing/model': 1.163540905341506, 'learning_rate/qwen_vl_interface': 5.407847347155395e-06, 'learning_rate/action_model': 5.407847347155395e-05, 'epoch': 1.27}) +[2026-06-17 23:32:23,848][starVLA.training.train_starvla][INFO] - Step 3478, Loss: {'train/loss': 0.5418637990951538, 'train/grad_norm_pre_clip': 17.076082229614258, 'timing/data': 0.0010825088247656822, 'timing/model': 4.154197629541159, 'learning_rate/qwen_vl_interface': 5.4056858347344934e-06, 'learning_rate/action_model': 5.405685834734493e-05, 'epoch': 1.27}) +[2026-06-17 23:32:40,626][starVLA.training.train_starvla][INFO] - Step 3479, Loss: {'train/loss': 0.4949273467063904, 'train/grad_norm_pre_clip': 16.387182235717773, 'timing/data': 0.0013073300942778587, 'timing/model': 1.4413700448349118, 'learning_rate/qwen_vl_interface': 5.403524290039751e-06, 'learning_rate/action_model': 5.403524290039751e-05, 'epoch': 1.27}) +[2026-06-17 23:32:45,266][starVLA.training.train_starvla][INFO] - Step 3480, Loss: {'train/loss': 0.24451659619808197, 'train/grad_norm_pre_clip': 9.900032043457031, 'timing/data': 0.0010279277339577675, 'timing/model': 1.3651974759995937, 'learning_rate/qwen_vl_interface': 5.401362713519257e-06, 'learning_rate/action_model': 5.4013627135192565e-05, 'epoch': 1.27}) +[2026-06-17 23:32:52,908][starVLA.training.train_starvla][INFO] - Step 3481, Loss: {'train/loss': 0.2816506028175354, 'train/grad_norm_pre_clip': 10.16675090789795, 'timing/data': 0.0007882090285420418, 'timing/model': 1.1177007649093866, 'learning_rate/qwen_vl_interface': 5.399201105621111e-06, 'learning_rate/action_model': 5.39920110562111e-05, 'epoch': 1.27}) +[2026-06-17 23:33:01,126][starVLA.training.train_starvla][INFO] - Step 3482, Loss: {'train/loss': 0.5540459156036377, 'train/grad_norm_pre_clip': 15.138050079345703, 'timing/data': 0.0014166692271828651, 'timing/model': 4.715822024270892, 'learning_rate/qwen_vl_interface': 5.397039466793412e-06, 'learning_rate/action_model': 5.3970394667934123e-05, 'epoch': 1.28}) +[2026-06-17 23:33:07,643][starVLA.training.train_starvla][INFO] - Step 3483, Loss: {'train/loss': 0.6241191029548645, 'train/grad_norm_pre_clip': 15.182435035705566, 'timing/data': 0.0008663451299071312, 'timing/model': 1.0932961590588093, 'learning_rate/qwen_vl_interface': 5.394877797484278e-06, 'learning_rate/action_model': 5.3948777974842777e-05, 'epoch': 1.28}) +[2026-06-17 23:33:13,141][starVLA.training.train_starvla][INFO] - Step 3484, Loss: {'train/loss': 0.5315898656845093, 'train/grad_norm_pre_clip': 16.158519744873047, 'timing/data': 0.0010423865169286728, 'timing/model': 1.1383374324068427, 'learning_rate/qwen_vl_interface': 5.392716098141817e-06, 'learning_rate/action_model': 5.3927160981418165e-05, 'epoch': 1.28}) +[2026-06-17 23:33:17,818][starVLA.training.train_starvla][INFO] - Step 3485, Loss: {'train/loss': 0.7518957257270813, 'train/grad_norm_pre_clip': 21.315948486328125, 'timing/data': 0.003059188835322857, 'timing/model': 1.4890480069443583, 'learning_rate/qwen_vl_interface': 5.390554369214157e-06, 'learning_rate/action_model': 5.390554369214157e-05, 'epoch': 1.28}) +[2026-06-17 23:33:23,676][starVLA.training.train_starvla][INFO] - Step 3486, Loss: {'train/loss': 0.4022770822048187, 'train/grad_norm_pre_clip': 16.82533836364746, 'timing/data': 0.001196705736219883, 'timing/model': 1.058293572627008, 'learning_rate/qwen_vl_interface': 5.388392611149425e-06, 'learning_rate/action_model': 5.388392611149424e-05, 'epoch': 1.28}) +[2026-06-17 23:33:28,772][starVLA.training.train_starvla][INFO] - Step 3487, Loss: {'train/loss': 0.5774721503257751, 'train/grad_norm_pre_clip': 24.864145278930664, 'timing/data': 0.005292551591992378, 'timing/model': 1.163494878448546, 'learning_rate/qwen_vl_interface': 5.386230824395754e-06, 'learning_rate/action_model': 5.386230824395754e-05, 'epoch': 1.28}) +[2026-06-17 23:33:33,824][starVLA.training.train_starvla][INFO] - Step 3488, Loss: {'train/loss': 0.7681565880775452, 'train/grad_norm_pre_clip': 20.691490173339844, 'timing/data': 0.0014618560671806335, 'timing/model': 1.170195440761745, 'learning_rate/qwen_vl_interface': 5.384069009401288e-06, 'learning_rate/action_model': 5.384069009401288e-05, 'epoch': 1.28}) +[2026-06-17 23:33:40,726][starVLA.training.train_starvla][INFO] - Step 3489, Loss: {'train/loss': 0.5683521628379822, 'train/grad_norm_pre_clip': 12.52077579498291, 'timing/data': 0.0012976396828889847, 'timing/model': 1.0699325716122985, 'learning_rate/qwen_vl_interface': 5.381907166614171e-06, 'learning_rate/action_model': 5.3819071666141704e-05, 'epoch': 1.28}) +[2026-06-17 23:33:49,509][starVLA.training.train_starvla][INFO] - Step 3490, Loss: {'train/loss': 0.2934037148952484, 'train/grad_norm_pre_clip': 16.074474334716797, 'timing/data': 0.0008894689381122589, 'timing/model': 1.440563584677875, 'learning_rate/qwen_vl_interface': 5.379745296482559e-06, 'learning_rate/action_model': 5.379745296482559e-05, 'epoch': 1.28}) +[2026-06-17 23:33:53,684][starVLA.training.train_starvla][INFO] - Step 3491, Loss: {'train/loss': 0.40058282017707825, 'train/grad_norm_pre_clip': 15.483297348022461, 'timing/data': 0.0009731464087963104, 'timing/model': 1.0721036167815328, 'learning_rate/qwen_vl_interface': 5.377583399454605e-06, 'learning_rate/action_model': 5.3775833994546044e-05, 'epoch': 1.28}) +[2026-06-17 23:33:58,349][starVLA.training.train_starvla][INFO] - Step 3492, Loss: {'train/loss': 0.4694446921348572, 'train/grad_norm_pre_clip': 15.26489543914795, 'timing/data': 0.0010628541931509972, 'timing/model': 1.0266364514827728, 'learning_rate/qwen_vl_interface': 5.375421475978476e-06, 'learning_rate/action_model': 5.375421475978476e-05, 'epoch': 1.28}) +[2026-06-17 23:34:04,465][starVLA.training.train_starvla][INFO] - Step 3493, Loss: {'train/loss': 0.48310190439224243, 'train/grad_norm_pre_clip': 16.382322311401367, 'timing/data': 0.0010506920516490936, 'timing/model': 1.0486844144761562, 'learning_rate/qwen_vl_interface': 5.37325952650234e-06, 'learning_rate/action_model': 5.37325952650234e-05, 'epoch': 1.28}) +[2026-06-17 23:34:09,629][starVLA.training.train_starvla][INFO] - Step 3494, Loss: {'train/loss': 0.6577933430671692, 'train/grad_norm_pre_clip': 25.123838424682617, 'timing/data': 0.001249442808330059, 'timing/model': 1.0923869274556637, 'learning_rate/qwen_vl_interface': 5.371097551474371e-06, 'learning_rate/action_model': 5.371097551474371e-05, 'epoch': 1.28}) +[2026-06-17 23:34:14,897][starVLA.training.train_starvla][INFO] - Step 3495, Loss: {'train/loss': 0.4660145938396454, 'train/grad_norm_pre_clip': 21.428667068481445, 'timing/data': 0.001065676100552082, 'timing/model': 1.5603258833289146, 'learning_rate/qwen_vl_interface': 5.36893555134275e-06, 'learning_rate/action_model': 5.3689355513427494e-05, 'epoch': 1.28}) +[2026-06-17 23:34:19,738][starVLA.training.train_starvla][INFO] - Step 3496, Loss: {'train/loss': 0.5732462406158447, 'train/grad_norm_pre_clip': 19.79372787475586, 'timing/data': 0.0011052954941987991, 'timing/model': 1.0797059508040547, 'learning_rate/qwen_vl_interface': 5.366773526555661e-06, 'learning_rate/action_model': 5.366773526555661e-05, 'epoch': 1.28}) +[2026-06-17 23:34:26,914][starVLA.training.train_starvla][INFO] - Step 3497, Loss: {'train/loss': 0.33847618103027344, 'train/grad_norm_pre_clip': 14.79775619506836, 'timing/data': 0.0011518895626068115, 'timing/model': 2.0588844371959567, 'learning_rate/qwen_vl_interface': 5.364611477561294e-06, 'learning_rate/action_model': 5.3646114775612945e-05, 'epoch': 1.28}) +[2026-06-17 23:34:37,634][starVLA.training.train_starvla][INFO] - Step 3498, Loss: {'train/loss': 0.5342820882797241, 'train/grad_norm_pre_clip': 18.033376693725586, 'timing/data': 0.0010021505877375603, 'timing/model': 1.2204728834331036, 'learning_rate/qwen_vl_interface': 5.362449404807845e-06, 'learning_rate/action_model': 5.362449404807844e-05, 'epoch': 1.28}) +[2026-06-17 23:34:42,582][starVLA.training.train_starvla][INFO] - Step 3499, Loss: {'train/loss': 0.899307131767273, 'train/grad_norm_pre_clip': 20.449447631835938, 'timing/data': 0.001087399199604988, 'timing/model': 1.0757108284160495, 'learning_rate/qwen_vl_interface': 5.360287308743512e-06, 'learning_rate/action_model': 5.3602873087435126e-05, 'epoch': 1.28}) +[2026-06-17 23:40:03,266][starVLA.training.train_starvla][INFO] - Step 3500, Loss: {'train/loss': 0.6412503123283386, 'train/grad_norm_pre_clip': 18.048538208007812, 'eval/loss': 0.607904851436615, 'eval/latency_2/loss': 0.607904851436615, 'eval/demon_fire_precision': 0.911049062624651, 'eval/demon_fire_recall': 0.9556485355648535, 'eval/demon_fire_f1': 0.9328160098019196, 'eval/demon_fire_rate_model': 0.6975514746800222, 'eval/demon_fire_rate_teacher': 0.6649972175848636, 'eval/demon_left_precision': 0.8902291917973462, 'eval/demon_left_recall': 0.7768421052631579, 'eval/demon_left_f1': 0.8296795952782462, 'eval/demon_left_rate_model': 0.23066221480244853, 'eval/demon_left_rate_teacher': 0.2643294379521425, 'eval/demon_right_precision': 0.861664712778429, 'eval/demon_right_recall': 0.8305084745762712, 'eval/demon_right_f1': 0.8457997698504028, 'eval/demon_right_rate_model': 0.23734001112966055, 'eval/demon_right_rate_teacher': 0.24624373956594323, 'eval/demon_move_f1': 0.8377396825643245, 'eval/demon_cc_f1': 0.9021017475772659, 'eval/latency_2/demon_fire_precision': 0.911049062624651, 'eval/latency_2/demon_fire_recall': 0.9556485355648535, 'eval/latency_2/demon_fire_f1': 0.9328160098019196, 'eval/latency_2/demon_fire_rate_model': 0.6975514746800222, 'eval/latency_2/demon_fire_rate_teacher': 0.6649972175848636, 'eval/latency_2/demon_left_precision': 0.8902291917973462, 'eval/latency_2/demon_left_recall': 0.7768421052631579, 'eval/latency_2/demon_left_f1': 0.8296795952782462, 'eval/latency_2/demon_left_rate_model': 0.23066221480244853, 'eval/latency_2/demon_left_rate_teacher': 0.2643294379521425, 'eval/latency_2/demon_right_precision': 0.861664712778429, 'eval/latency_2/demon_right_recall': 0.8305084745762712, 'eval/latency_2/demon_right_f1': 0.8457997698504028, 'eval/latency_2/demon_right_rate_model': 0.23734001112966055, 'eval/latency_2/demon_right_rate_teacher': 0.24624373956594323, 'eval/latency_2/demon_move_f1': 0.8377396825643245, 'eval/latency_2/demon_cc_f1': 0.9021017475772659, 'timing/data': 0.0008718371391296387, 'timing/model': 1.5266726752743125, 'learning_rate/qwen_vl_interface': 5.358125189816502e-06, 'learning_rate/action_model': 5.358125189816502e-05, 'epoch': 1.28}) +[2026-06-17 23:40:03,298][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_3500_state +[2026-06-17 23:40:03,307][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 23:40:12,448][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_3500_state/model.safetensors +[2026-06-17 23:40:30,313][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_3500_state/optimizer.bin +[2026-06-17 23:40:30,315][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_3500_state/sampler.bin +[2026-06-17 23:40:30,317][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_3500_state/sampler_1.bin +[2026-06-17 23:40:30,320][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_3500_state/random_states_0.pkl +[2026-06-17 23:40:30,326][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3500 +[2026-06-17 23:40:30,327][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 23:40:30,344][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 23:40:35,405][starVLA.training.train_starvla][INFO] - Step 3501, Loss: {'train/loss': 0.4504740238189697, 'train/grad_norm_pre_clip': 14.905228614807129, 'timing/data': 0.0009608594700694084, 'timing/model': 1.4554961267858744, 'learning_rate/qwen_vl_interface': 5.355963048475024e-06, 'learning_rate/action_model': 5.3559630484750244e-05, 'epoch': 1.28}) +[2026-06-17 23:40:39,690][starVLA.training.train_starvla][INFO] - Step 3502, Loss: {'train/loss': 0.4607221782207489, 'train/grad_norm_pre_clip': 14.636811256408691, 'timing/data': 0.001135634258389473, 'timing/model': 1.046384397894144, 'learning_rate/qwen_vl_interface': 5.3538008851672916e-06, 'learning_rate/action_model': 5.3538008851672916e-05, 'epoch': 1.28}) +[2026-06-17 23:40:44,659][starVLA.training.train_starvla][INFO] - Step 3503, Loss: {'train/loss': 0.4779002070426941, 'train/grad_norm_pre_clip': 15.037270545959473, 'timing/data': 0.0010504694655537605, 'timing/model': 1.0422809217125177, 'learning_rate/qwen_vl_interface': 5.351638700341525e-06, 'learning_rate/action_model': 5.351638700341525e-05, 'epoch': 1.28}) +[2026-06-17 23:40:49,268][starVLA.training.train_starvla][INFO] - Step 3504, Loss: {'train/loss': 0.3651209771633148, 'train/grad_norm_pre_clip': 12.651449203491211, 'timing/data': 0.0008940510451793671, 'timing/model': 1.057359466329217, 'learning_rate/qwen_vl_interface': 5.349476494445947e-06, 'learning_rate/action_model': 5.3494764944459466e-05, 'epoch': 1.28}) +[2026-06-17 23:40:53,876][starVLA.training.train_starvla][INFO] - Step 3505, Loss: {'train/loss': 0.5154497623443604, 'train/grad_norm_pre_clip': 16.486543655395508, 'timing/data': 0.001408953219652176, 'timing/model': 1.0887345680966973, 'learning_rate/qwen_vl_interface': 5.3473142679287824e-06, 'learning_rate/action_model': 5.3473142679287826e-05, 'epoch': 1.28}) +[2026-06-17 23:40:58,625][starVLA.training.train_starvla][INFO] - Step 3506, Loss: {'train/loss': 1.0262619256973267, 'train/grad_norm_pre_clip': 21.628286361694336, 'timing/data': 0.0007995562627911568, 'timing/model': 1.4950003726407886, 'learning_rate/qwen_vl_interface': 5.345152021238268e-06, 'learning_rate/action_model': 5.345152021238268e-05, 'epoch': 1.28}) +[2026-06-17 23:41:02,939][starVLA.training.train_starvla][INFO] - Step 3507, Loss: {'train/loss': 0.2829382121562958, 'train/grad_norm_pre_clip': 13.250163078308105, 'timing/data': 0.0009052706882357597, 'timing/model': 1.0735206129029393, 'learning_rate/qwen_vl_interface': 5.342989754822635e-06, 'learning_rate/action_model': 5.342989754822635e-05, 'epoch': 1.28}) +[2026-06-17 23:41:08,128][starVLA.training.train_starvla][INFO] - Step 3508, Loss: {'train/loss': 0.5810277462005615, 'train/grad_norm_pre_clip': 20.12489128112793, 'timing/data': 0.001120251603424549, 'timing/model': 1.2417558282613754, 'learning_rate/qwen_vl_interface': 5.340827469130126e-06, 'learning_rate/action_model': 5.3408274691301255e-05, 'epoch': 1.28}) +[2026-06-17 23:41:12,614][starVLA.training.train_starvla][INFO] - Step 3509, Loss: {'train/loss': 0.456118106842041, 'train/grad_norm_pre_clip': 20.753049850463867, 'timing/data': 0.004033278673887253, 'timing/model': 1.097349121235311, 'learning_rate/qwen_vl_interface': 5.338665164608985e-06, 'learning_rate/action_model': 5.3386651646089845e-05, 'epoch': 1.28}) +[2026-06-17 23:41:17,090][starVLA.training.train_starvla][INFO] - Step 3510, Loss: {'train/loss': 0.4320357143878937, 'train/grad_norm_pre_clip': 14.567573547363281, 'timing/data': 0.0012172972783446312, 'timing/model': 1.037852088920772, 'learning_rate/qwen_vl_interface': 5.336502841707461e-06, 'learning_rate/action_model': 5.33650284170746e-05, 'epoch': 1.29}) +[2026-06-17 23:41:21,871][starVLA.training.train_starvla][INFO] - Step 3511, Loss: {'train/loss': 0.6189189553260803, 'train/grad_norm_pre_clip': 23.230445861816406, 'timing/data': 0.0010916776955127716, 'timing/model': 1.4123503603041172, 'learning_rate/qwen_vl_interface': 5.334340500873803e-06, 'learning_rate/action_model': 5.3343405008738034e-05, 'epoch': 1.29}) +[2026-06-17 23:41:26,537][starVLA.training.train_starvla][INFO] - Step 3512, Loss: {'train/loss': 0.49929729104042053, 'train/grad_norm_pre_clip': 23.21016502380371, 'timing/data': 0.0010471194982528687, 'timing/model': 1.0768984034657478, 'learning_rate/qwen_vl_interface': 5.33217814255627e-06, 'learning_rate/action_model': 5.33217814255627e-05, 'epoch': 1.29}) +[2026-06-17 23:41:31,289][starVLA.training.train_starvla][INFO] - Step 3513, Loss: {'train/loss': 0.6476315855979919, 'train/grad_norm_pre_clip': 21.81831932067871, 'timing/data': 0.0008597755804657936, 'timing/model': 1.0380002213642001, 'learning_rate/qwen_vl_interface': 5.330015767203118e-06, 'learning_rate/action_model': 5.330015767203118e-05, 'epoch': 1.29}) +[2026-06-17 23:41:35,873][starVLA.training.train_starvla][INFO] - Step 3514, Loss: {'train/loss': 0.6926234364509583, 'train/grad_norm_pre_clip': 25.72797966003418, 'timing/data': 0.001428878866136074, 'timing/model': 1.1079418305307627, 'learning_rate/qwen_vl_interface': 5.3278533752626135e-06, 'learning_rate/action_model': 5.3278533752626136e-05, 'epoch': 1.29}) +[2026-06-17 23:41:45,093][starVLA.training.train_starvla][INFO] - Step 3515, Loss: {'train/loss': 0.40320077538490295, 'train/grad_norm_pre_clip': 13.777031898498535, 'timing/data': 0.001370169222354889, 'timing/model': 5.752515654079616, 'learning_rate/qwen_vl_interface': 5.325690967183019e-06, 'learning_rate/action_model': 5.325690967183019e-05, 'epoch': 1.29}) +[2026-06-17 23:41:58,163][starVLA.training.train_starvla][INFO] - Step 3516, Loss: {'train/loss': 0.4885064363479614, 'train/grad_norm_pre_clip': 11.069580078125, 'timing/data': 0.0014728596433997154, 'timing/model': 4.718513820320368, 'learning_rate/qwen_vl_interface': 5.323528543412607e-06, 'learning_rate/action_model': 5.323528543412607e-05, 'epoch': 1.29}) +[2026-06-17 23:42:09,167][starVLA.training.train_starvla][INFO] - Step 3517, Loss: {'train/loss': 0.5315038561820984, 'train/grad_norm_pre_clip': 17.164791107177734, 'timing/data': 0.0008424622938036919, 'timing/model': 1.14211299829185, 'learning_rate/qwen_vl_interface': 5.321366104399648e-06, 'learning_rate/action_model': 5.321366104399648e-05, 'epoch': 1.29}) +[2026-06-17 23:42:22,119][starVLA.training.train_starvla][INFO] - Step 3518, Loss: {'train/loss': 0.5689144730567932, 'train/grad_norm_pre_clip': 19.5346622467041, 'timing/data': 0.0007821805775165558, 'timing/model': 1.4426888320595026, 'learning_rate/qwen_vl_interface': 5.319203650592419e-06, 'learning_rate/action_model': 5.3192036505924194e-05, 'epoch': 1.29}) +[2026-06-17 23:42:34,583][starVLA.training.train_starvla][INFO] - Step 3519, Loss: {'train/loss': 0.2901875972747803, 'train/grad_norm_pre_clip': 16.652767181396484, 'timing/data': 0.0008336743339896202, 'timing/model': 1.3607761915773153, 'learning_rate/qwen_vl_interface': 5.317041182439199e-06, 'learning_rate/action_model': 5.3170411824391985e-05, 'epoch': 1.29}) +[2026-06-17 23:42:47,858][starVLA.training.train_starvla][INFO] - Step 3520, Loss: {'train/loss': 0.5118517875671387, 'train/grad_norm_pre_clip': 19.99229621887207, 'timing/data': 0.0007907962426543236, 'timing/model': 2.214456047862768, 'learning_rate/qwen_vl_interface': 5.314878700388267e-06, 'learning_rate/action_model': 5.314878700388267e-05, 'epoch': 1.29}) +[2026-06-17 23:42:59,019][starVLA.training.train_starvla][INFO] - Step 3521, Loss: {'train/loss': 0.5653424859046936, 'train/grad_norm_pre_clip': 13.018074989318848, 'timing/data': 0.0013237036764621735, 'timing/model': 2.9140856992453337, 'learning_rate/qwen_vl_interface': 5.312716204887914e-06, 'learning_rate/action_model': 5.3127162048879145e-05, 'epoch': 1.29}) +[2026-06-17 23:43:07,674][starVLA.training.train_starvla][INFO] - Step 3522, Loss: {'train/loss': 0.6607814431190491, 'train/grad_norm_pre_clip': 18.017723083496094, 'timing/data': 0.0011580334976315498, 'timing/model': 4.395414820872247, 'learning_rate/qwen_vl_interface': 5.3105536963864226e-06, 'learning_rate/action_model': 5.310553696386422e-05, 'epoch': 1.29}) +[2026-06-17 23:43:12,538][starVLA.training.train_starvla][INFO] - Step 3523, Loss: {'train/loss': 0.5166653990745544, 'train/grad_norm_pre_clip': 28.230546951293945, 'timing/data': 0.0012577548623085022, 'timing/model': 1.1943441880866885, 'learning_rate/qwen_vl_interface': 5.3083911753320875e-06, 'learning_rate/action_model': 5.308391175332087e-05, 'epoch': 1.29}) +[2026-06-17 23:43:23,571][starVLA.training.train_starvla][INFO] - Step 3524, Loss: {'train/loss': 0.521090567111969, 'train/grad_norm_pre_clip': 16.748775482177734, 'timing/data': 0.0008239159360527992, 'timing/model': 5.848989098332822, 'learning_rate/qwen_vl_interface': 5.306228642173194e-06, 'learning_rate/action_model': 5.306228642173194e-05, 'epoch': 1.29}) +[2026-06-17 23:43:35,627][starVLA.training.train_starvla][INFO] - Step 3525, Loss: {'train/loss': 0.5312054753303528, 'train/grad_norm_pre_clip': 16.489471435546875, 'timing/data': 0.00132849533110857, 'timing/model': 1.0826863497495651, 'learning_rate/qwen_vl_interface': 5.304066097358046e-06, 'learning_rate/action_model': 5.304066097358046e-05, 'epoch': 1.29}) +[2026-06-17 23:43:40,220][starVLA.training.train_starvla][INFO] - Step 3526, Loss: {'train/loss': 0.5308021306991577, 'train/grad_norm_pre_clip': 28.611570358276367, 'timing/data': 0.0012604501098394394, 'timing/model': 1.4587054159492254, 'learning_rate/qwen_vl_interface': 5.301903541334937e-06, 'learning_rate/action_model': 5.301903541334937e-05, 'epoch': 1.29}) +[2026-06-17 23:43:44,550][starVLA.training.train_starvla][INFO] - Step 3527, Loss: {'train/loss': 0.4677693545818329, 'train/grad_norm_pre_clip': 19.732818603515625, 'timing/data': 0.0008804406970739365, 'timing/model': 1.0763368364423513, 'learning_rate/qwen_vl_interface': 5.299740974552168e-06, 'learning_rate/action_model': 5.299740974552167e-05, 'epoch': 1.29}) +[2026-06-17 23:43:49,605][starVLA.training.train_starvla][INFO] - Step 3528, Loss: {'train/loss': 0.4110376238822937, 'train/grad_norm_pre_clip': 13.43112564086914, 'timing/data': 0.0018683001399040222, 'timing/model': 1.0888714473694563, 'learning_rate/qwen_vl_interface': 5.297578397458042e-06, 'learning_rate/action_model': 5.297578397458042e-05, 'epoch': 1.29}) +[2026-06-17 23:43:54,178][starVLA.training.train_starvla][INFO] - Step 3529, Loss: {'train/loss': 0.5863603353500366, 'train/grad_norm_pre_clip': 15.382451057434082, 'timing/data': 0.0010801879689097404, 'timing/model': 1.07274850923568, 'learning_rate/qwen_vl_interface': 5.295415810500862e-06, 'learning_rate/action_model': 5.295415810500862e-05, 'epoch': 1.29}) +[2026-06-17 23:43:58,813][starVLA.training.train_starvla][INFO] - Step 3530, Loss: {'train/loss': 0.597667932510376, 'train/grad_norm_pre_clip': 15.888993263244629, 'timing/data': 0.0007828865200281143, 'timing/model': 1.0675468016415834, 'learning_rate/qwen_vl_interface': 5.293253214128938e-06, 'learning_rate/action_model': 5.293253214128937e-05, 'epoch': 1.29}) +[2026-06-17 23:44:03,584][starVLA.training.train_starvla][INFO] - Step 3531, Loss: {'train/loss': 0.5731149911880493, 'train/grad_norm_pre_clip': 19.845739364624023, 'timing/data': 0.0040778350085020065, 'timing/model': 1.575077859684825, 'learning_rate/qwen_vl_interface': 5.291090608790572e-06, 'learning_rate/action_model': 5.291090608790572e-05, 'epoch': 1.29}) +[2026-06-17 23:44:07,689][starVLA.training.train_starvla][INFO] - Step 3532, Loss: {'train/loss': 0.31733402609825134, 'train/grad_norm_pre_clip': 12.69586181640625, 'timing/data': 0.0011755600571632385, 'timing/model': 1.058310846798122, 'learning_rate/qwen_vl_interface': 5.28892799493408e-06, 'learning_rate/action_model': 5.2889279949340796e-05, 'epoch': 1.29}) +[2026-06-17 23:44:12,769][starVLA.training.train_starvla][INFO] - Step 3533, Loss: {'train/loss': 0.6625946760177612, 'train/grad_norm_pre_clip': 19.141708374023438, 'timing/data': 0.0019023744389414787, 'timing/model': 1.3027179557830095, 'learning_rate/qwen_vl_interface': 5.286765373007774e-06, 'learning_rate/action_model': 5.2867653730077735e-05, 'epoch': 1.29}) +[2026-06-17 23:44:17,564][starVLA.training.train_starvla][INFO] - Step 3534, Loss: {'train/loss': 0.4722650647163391, 'train/grad_norm_pre_clip': 14.386811256408691, 'timing/data': 0.001072014681994915, 'timing/model': 1.2744837803766131, 'learning_rate/qwen_vl_interface': 5.284602743459968e-06, 'learning_rate/action_model': 5.284602743459968e-05, 'epoch': 1.29}) +[2026-06-17 23:44:22,312][starVLA.training.train_starvla][INFO] - Step 3535, Loss: {'train/loss': 0.5346226692199707, 'train/grad_norm_pre_clip': 15.252378463745117, 'timing/data': 0.0012540491297841072, 'timing/model': 1.418660861440003, 'learning_rate/qwen_vl_interface': 5.282440106738973e-06, 'learning_rate/action_model': 5.282440106738973e-05, 'epoch': 1.29}) +[2026-06-17 23:44:26,813][starVLA.training.train_starvla][INFO] - Step 3536, Loss: {'train/loss': 0.5042954683303833, 'train/grad_norm_pre_clip': 15.555705070495605, 'timing/data': 0.0008315872400999069, 'timing/model': 1.3909543743357062, 'learning_rate/qwen_vl_interface': 5.280277463293111e-06, 'learning_rate/action_model': 5.280277463293111e-05, 'epoch': 1.29}) +[2026-06-17 23:44:32,790][starVLA.training.train_starvla][INFO] - Step 3537, Loss: {'train/loss': 0.39172783493995667, 'train/grad_norm_pre_clip': 14.083884239196777, 'timing/data': 0.0007968926802277565, 'timing/model': 1.0649689380079508, 'learning_rate/qwen_vl_interface': 5.2781148135706975e-06, 'learning_rate/action_model': 5.278114813570697e-05, 'epoch': 1.3}) +[2026-06-17 23:44:37,386][starVLA.training.train_starvla][INFO] - Step 3538, Loss: {'train/loss': 0.8556725978851318, 'train/grad_norm_pre_clip': 27.279417037963867, 'timing/data': 0.0008612051606178284, 'timing/model': 1.0336969681084156, 'learning_rate/qwen_vl_interface': 5.275952158020054e-06, 'learning_rate/action_model': 5.275952158020053e-05, 'epoch': 1.3}) +[2026-06-17 23:44:42,399][starVLA.training.train_starvla][INFO] - Step 3539, Loss: {'train/loss': 0.3085514307022095, 'train/grad_norm_pre_clip': 16.762557983398438, 'timing/data': 0.001081656664609909, 'timing/model': 1.0375959919765592, 'learning_rate/qwen_vl_interface': 5.273789497089497e-06, 'learning_rate/action_model': 5.2737894970894974e-05, 'epoch': 1.3}) +[2026-06-17 23:44:46,971][starVLA.training.train_starvla][INFO] - Step 3540, Loss: {'train/loss': 0.5249958038330078, 'train/grad_norm_pre_clip': 23.443235397338867, 'timing/data': 0.001279420219361782, 'timing/model': 1.0856446214020252, 'learning_rate/qwen_vl_interface': 5.271626831227355e-06, 'learning_rate/action_model': 5.2716268312273544e-05, 'epoch': 1.3}) +[2026-06-17 23:44:54,897][starVLA.training.train_starvla][INFO] - Step 3541, Loss: {'train/loss': 0.8017398715019226, 'train/grad_norm_pre_clip': 17.757253646850586, 'timing/data': 0.0010257698595523834, 'timing/model': 1.7319503026083112, 'learning_rate/qwen_vl_interface': 5.2694641608819465e-06, 'learning_rate/action_model': 5.269464160881946e-05, 'epoch': 1.3}) +[2026-06-17 23:45:00,930][starVLA.training.train_starvla][INFO] - Step 3542, Loss: {'train/loss': 0.4226556718349457, 'train/grad_norm_pre_clip': 14.34609603881836, 'timing/data': 0.0018281619995832443, 'timing/model': 1.262260776013136, 'learning_rate/qwen_vl_interface': 5.267301486501595e-06, 'learning_rate/action_model': 5.2673014865015945e-05, 'epoch': 1.3}) +[2026-06-17 23:45:06,337][starVLA.training.train_starvla][INFO] - Step 3543, Loss: {'train/loss': 0.5867173671722412, 'train/grad_norm_pre_clip': 21.317148208618164, 'timing/data': 0.0008432716131210327, 'timing/model': 1.08756282646209, 'learning_rate/qwen_vl_interface': 5.265138808534628e-06, 'learning_rate/action_model': 5.265138808534628e-05, 'epoch': 1.3}) +[2026-06-17 23:45:17,810][starVLA.training.train_starvla][INFO] - Step 3544, Loss: {'train/loss': 0.5273066163063049, 'train/grad_norm_pre_clip': 15.454133987426758, 'timing/data': 0.0010989643633365631, 'timing/model': 1.1478475779294968, 'learning_rate/qwen_vl_interface': 5.26297612742937e-06, 'learning_rate/action_model': 5.26297612742937e-05, 'epoch': 1.3}) +[2026-06-17 23:45:22,499][starVLA.training.train_starvla][INFO] - Step 3545, Loss: {'train/loss': 0.5105942487716675, 'train/grad_norm_pre_clip': 19.191150665283203, 'timing/data': 0.0010634632781147957, 'timing/model': 1.0854058787226677, 'learning_rate/qwen_vl_interface': 5.2608134436341476e-06, 'learning_rate/action_model': 5.260813443634147e-05, 'epoch': 1.3}) +[2026-06-17 23:45:30,624][starVLA.training.train_starvla][INFO] - Step 3546, Loss: {'train/loss': 0.6101928353309631, 'train/grad_norm_pre_clip': 19.43437957763672, 'timing/data': 0.0011034691706299782, 'timing/model': 1.5104353921487927, 'learning_rate/qwen_vl_interface': 5.258650757597285e-06, 'learning_rate/action_model': 5.258650757597286e-05, 'epoch': 1.3}) +[2026-06-17 23:45:35,545][starVLA.training.train_starvla][INFO] - Step 3547, Loss: {'train/loss': 0.6013579368591309, 'train/grad_norm_pre_clip': 20.794858932495117, 'timing/data': 0.0013839909806847572, 'timing/model': 1.2574122101068497, 'learning_rate/qwen_vl_interface': 5.256488069767113e-06, 'learning_rate/action_model': 5.256488069767113e-05, 'epoch': 1.3}) +[2026-06-17 23:45:40,693][starVLA.training.train_starvla][INFO] - Step 3548, Loss: {'train/loss': 0.5062162280082703, 'train/grad_norm_pre_clip': 18.06205940246582, 'timing/data': 0.0010721860453486443, 'timing/model': 1.0628862967714667, 'learning_rate/qwen_vl_interface': 5.254325380591955e-06, 'learning_rate/action_model': 5.254325380591955e-05, 'epoch': 1.3}) +[2026-06-17 23:45:45,814][starVLA.training.train_starvla][INFO] - Step 3549, Loss: {'train/loss': 0.2705078125, 'train/grad_norm_pre_clip': 13.341532707214355, 'timing/data': 0.0012650135904550552, 'timing/model': 1.140791380777955, 'learning_rate/qwen_vl_interface': 5.252162690520142e-06, 'learning_rate/action_model': 5.252162690520142e-05, 'epoch': 1.3}) +[2026-06-17 23:45:59,080][starVLA.training.train_starvla][INFO] - Step 3550, Loss: {'train/loss': 0.7603140473365784, 'train/grad_norm_pre_clip': 21.49127769470215, 'timing/data': 0.0007984088733792305, 'timing/model': 1.1706167235970497, 'learning_rate/qwen_vl_interface': 5.2500000000000006e-06, 'learning_rate/action_model': 5.25e-05, 'epoch': 1.3}) +[2026-06-17 23:46:03,913][starVLA.training.train_starvla][INFO] - Step 3551, Loss: {'train/loss': 0.4228035509586334, 'train/grad_norm_pre_clip': 14.9381685256958, 'timing/data': 0.0011925771832466125, 'timing/model': 1.5748594524338841, 'learning_rate/qwen_vl_interface': 5.24783730947986e-06, 'learning_rate/action_model': 5.2478373094798594e-05, 'epoch': 1.3}) +[2026-06-17 23:46:08,432][starVLA.training.train_starvla][INFO] - Step 3552, Loss: {'train/loss': 0.6084578633308411, 'train/grad_norm_pre_clip': 14.799885749816895, 'timing/data': 0.0037990566343069077, 'timing/model': 1.0899024344980717, 'learning_rate/qwen_vl_interface': 5.245674619408045e-06, 'learning_rate/action_model': 5.245674619408045e-05, 'epoch': 1.3}) +[2026-06-17 23:46:17,570][starVLA.training.train_starvla][INFO] - Step 3553, Loss: {'train/loss': 0.44503724575042725, 'train/grad_norm_pre_clip': 12.471135139465332, 'timing/data': 0.0011523207649588585, 'timing/model': 5.570530284196138, 'learning_rate/qwen_vl_interface': 5.243511930232889e-06, 'learning_rate/action_model': 5.243511930232888e-05, 'epoch': 1.3}) +[2026-06-17 23:46:25,677][starVLA.training.train_starvla][INFO] - Step 3554, Loss: {'train/loss': 0.21683092415332794, 'train/grad_norm_pre_clip': 18.786495208740234, 'timing/data': 0.001209089532494545, 'timing/model': 1.1911468813195825, 'learning_rate/qwen_vl_interface': 5.2413492424027165e-06, 'learning_rate/action_model': 5.241349242402717e-05, 'epoch': 1.3}) +[2026-06-17 23:46:30,604][starVLA.training.train_starvla][INFO] - Step 3555, Loss: {'train/loss': 0.6608862280845642, 'train/grad_norm_pre_clip': 23.910362243652344, 'timing/data': 0.0009899633005261421, 'timing/model': 1.4513188442215323, 'learning_rate/qwen_vl_interface': 5.239186556365855e-06, 'learning_rate/action_model': 5.239186556365855e-05, 'epoch': 1.3}) +[2026-06-17 23:46:35,940][starVLA.training.train_starvla][INFO] - Step 3556, Loss: {'train/loss': 0.6241041421890259, 'train/grad_norm_pre_clip': 22.828386306762695, 'timing/data': 0.0009221434593200684, 'timing/model': 1.411053366959095, 'learning_rate/qwen_vl_interface': 5.23702387257063e-06, 'learning_rate/action_model': 5.23702387257063e-05, 'epoch': 1.3}) +[2026-06-17 23:46:44,137][starVLA.training.train_starvla][INFO] - Step 3557, Loss: {'train/loss': 0.5166789889335632, 'train/grad_norm_pre_clip': 15.68290901184082, 'timing/data': 0.0011683311313390732, 'timing/model': 4.9483723770827055, 'learning_rate/qwen_vl_interface': 5.234861191465373e-06, 'learning_rate/action_model': 5.234861191465372e-05, 'epoch': 1.3}) +[2026-06-17 23:46:49,436][starVLA.training.train_starvla][INFO] - Step 3558, Loss: {'train/loss': 0.2835099399089813, 'train/grad_norm_pre_clip': 10.855857849121094, 'timing/data': 0.001077546738088131, 'timing/model': 1.3278170051053166, 'learning_rate/qwen_vl_interface': 5.232698513498407e-06, 'learning_rate/action_model': 5.2326985134984066e-05, 'epoch': 1.3}) +[2026-06-17 23:46:58,768][starVLA.training.train_starvla][INFO] - Step 3559, Loss: {'train/loss': 0.7744854688644409, 'train/grad_norm_pre_clip': 28.585805892944336, 'timing/data': 0.0014842897653579712, 'timing/model': 1.0450460752472281, 'learning_rate/qwen_vl_interface': 5.230535839118057e-06, 'learning_rate/action_model': 5.230535839118057e-05, 'epoch': 1.3}) +[2026-06-17 23:47:11,924][starVLA.training.train_starvla][INFO] - Step 3560, Loss: {'train/loss': 0.5736666321754456, 'train/grad_norm_pre_clip': 42.984962463378906, 'timing/data': 0.0009760092943906784, 'timing/model': 1.1538806455209851, 'learning_rate/qwen_vl_interface': 5.2283731687726455e-06, 'learning_rate/action_model': 5.2283731687726454e-05, 'epoch': 1.3}) +[2026-06-17 23:47:16,997][starVLA.training.train_starvla][INFO] - Step 3561, Loss: {'train/loss': 0.37445947527885437, 'train/grad_norm_pre_clip': 16.434547424316406, 'timing/data': 0.0007664384320378304, 'timing/model': 1.625203519128263, 'learning_rate/qwen_vl_interface': 5.226210502910504e-06, 'learning_rate/action_model': 5.226210502910504e-05, 'epoch': 1.3}) +[2026-06-17 23:47:21,310][starVLA.training.train_starvla][INFO] - Step 3562, Loss: {'train/loss': 0.41034966707229614, 'train/grad_norm_pre_clip': 16.877986907958984, 'timing/data': 0.0008138706907629967, 'timing/model': 1.039578982628882, 'learning_rate/qwen_vl_interface': 5.224047841979949e-06, 'learning_rate/action_model': 5.2240478419799485e-05, 'epoch': 1.3}) +[2026-06-17 23:47:27,926][starVLA.training.train_starvla][INFO] - Step 3563, Loss: {'train/loss': 0.796160101890564, 'train/grad_norm_pre_clip': 18.861957550048828, 'timing/data': 0.0012966329231858253, 'timing/model': 1.1143962880596519, 'learning_rate/qwen_vl_interface': 5.2218851864293044e-06, 'learning_rate/action_model': 5.221885186429304e-05, 'epoch': 1.3}) +[2026-06-17 23:47:32,992][starVLA.training.train_starvla][INFO] - Step 3564, Loss: {'train/loss': 0.5300499796867371, 'train/grad_norm_pre_clip': 16.690458297729492, 'timing/data': 0.0008793631568551064, 'timing/model': 1.0334888575598598, 'learning_rate/qwen_vl_interface': 5.21972253670689e-06, 'learning_rate/action_model': 5.21972253670689e-05, 'epoch': 1.31}) +[2026-06-17 23:47:37,559][starVLA.training.train_starvla][INFO] - Step 3565, Loss: {'train/loss': 0.34120798110961914, 'train/grad_norm_pre_clip': 12.254758834838867, 'timing/data': 0.005960728041827679, 'timing/model': 1.1495140893384814, 'learning_rate/qwen_vl_interface': 5.217559893261028e-06, 'learning_rate/action_model': 5.217559893261028e-05, 'epoch': 1.31}) +[2026-06-17 23:47:42,208][starVLA.training.train_starvla][INFO] - Step 3566, Loss: {'train/loss': 0.41758856177330017, 'train/grad_norm_pre_clip': 15.003694534301758, 'timing/data': 0.0013005435466766357, 'timing/model': 1.5390007477253675, 'learning_rate/qwen_vl_interface': 5.2153972565400355e-06, 'learning_rate/action_model': 5.215397256540035e-05, 'epoch': 1.31}) +[2026-06-17 23:47:46,344][starVLA.training.train_starvla][INFO] - Step 3567, Loss: {'train/loss': 0.4573785960674286, 'train/grad_norm_pre_clip': 16.809309005737305, 'timing/data': 0.0009634159505367279, 'timing/model': 1.0969772338867188, 'learning_rate/qwen_vl_interface': 5.213234626992227e-06, 'learning_rate/action_model': 5.213234626992226e-05, 'epoch': 1.31}) +[2026-06-17 23:47:51,022][starVLA.training.train_starvla][INFO] - Step 3568, Loss: {'train/loss': 0.3384603261947632, 'train/grad_norm_pre_clip': 15.802181243896484, 'timing/data': 0.0011068927124142647, 'timing/model': 1.0935899456962943, 'learning_rate/qwen_vl_interface': 5.21107200506592e-06, 'learning_rate/action_model': 5.21107200506592e-05, 'epoch': 1.31}) +[2026-06-17 23:47:58,722][starVLA.training.train_starvla][INFO] - Step 3569, Loss: {'train/loss': 0.5731334090232849, 'train/grad_norm_pre_clip': 15.132894515991211, 'timing/data': 0.0025503532961010933, 'timing/model': 1.3475959347561002, 'learning_rate/qwen_vl_interface': 5.208909391209429e-06, 'learning_rate/action_model': 5.208909391209429e-05, 'epoch': 1.31}) +[2026-06-17 23:48:07,130][starVLA.training.train_starvla][INFO] - Step 3570, Loss: {'train/loss': 0.3751850724220276, 'train/grad_norm_pre_clip': 17.28737449645996, 'timing/data': 0.0008982783183455467, 'timing/model': 1.0679470375180244, 'learning_rate/qwen_vl_interface': 5.206746785871066e-06, 'learning_rate/action_model': 5.2067467858710664e-05, 'epoch': 1.31}) +[2026-06-17 23:48:11,997][starVLA.training.train_starvla][INFO] - Step 3571, Loss: {'train/loss': 0.643539309501648, 'train/grad_norm_pre_clip': 22.88052749633789, 'timing/data': 0.0010755667462944984, 'timing/model': 1.6265266733244061, 'learning_rate/qwen_vl_interface': 5.204584189499139e-06, 'learning_rate/action_model': 5.2045841894991385e-05, 'epoch': 1.31}) +[2026-06-17 23:48:16,304][starVLA.training.train_starvla][INFO] - Step 3572, Loss: {'train/loss': 0.7319045662879944, 'train/grad_norm_pre_clip': 20.976802825927734, 'timing/data': 0.0010554520413279533, 'timing/model': 1.0409971047192812, 'learning_rate/qwen_vl_interface': 5.202421602541959e-06, 'learning_rate/action_model': 5.2024216025419594e-05, 'epoch': 1.31}) +[2026-06-17 23:48:20,953][starVLA.training.train_starvla][INFO] - Step 3573, Loss: {'train/loss': 0.5304194092750549, 'train/grad_norm_pre_clip': 15.045018196105957, 'timing/data': 0.0011877873912453651, 'timing/model': 1.129286832176149, 'learning_rate/qwen_vl_interface': 5.200259025447835e-06, 'learning_rate/action_model': 5.2002590254478345e-05, 'epoch': 1.31}) +[2026-06-17 23:48:25,779][starVLA.training.train_starvla][INFO] - Step 3574, Loss: {'train/loss': 0.45578670501708984, 'train/grad_norm_pre_clip': 13.289422035217285, 'timing/data': 0.001150502823293209, 'timing/model': 1.2936161840334535, 'learning_rate/qwen_vl_interface': 5.198096458665064e-06, 'learning_rate/action_model': 5.198096458665064e-05, 'epoch': 1.31}) +[2026-06-17 23:48:31,159][starVLA.training.train_starvla][INFO] - Step 3575, Loss: {'train/loss': 0.7629296183586121, 'train/grad_norm_pre_clip': 21.2630672454834, 'timing/data': 0.0010776175186038017, 'timing/model': 1.2888088235631585, 'learning_rate/qwen_vl_interface': 5.195933902641955e-06, 'learning_rate/action_model': 5.195933902641955e-05, 'epoch': 1.31}) +[2026-06-17 23:48:36,109][starVLA.training.train_starvla][INFO] - Step 3576, Loss: {'train/loss': 0.5498439073562622, 'train/grad_norm_pre_clip': 15.22624397277832, 'timing/data': 0.0009657777845859528, 'timing/model': 1.646771646104753, 'learning_rate/qwen_vl_interface': 5.193771357826807e-06, 'learning_rate/action_model': 5.193771357826806e-05, 'epoch': 1.31}) +[2026-06-17 23:48:40,611][starVLA.training.train_starvla][INFO] - Step 3577, Loss: {'train/loss': 0.3245815336704254, 'train/grad_norm_pre_clip': 20.64966583251953, 'timing/data': 0.0008274326100945473, 'timing/model': 1.0473966300487518, 'learning_rate/qwen_vl_interface': 5.191608824667916e-06, 'learning_rate/action_model': 5.1916088246679165e-05, 'epoch': 1.31}) +[2026-06-17 23:48:46,416][starVLA.training.train_starvla][INFO] - Step 3578, Loss: {'train/loss': 0.5560081005096436, 'train/grad_norm_pre_clip': 18.55594253540039, 'timing/data': 0.0015049520879983902, 'timing/model': 2.082015637308359, 'learning_rate/qwen_vl_interface': 5.189446303613578e-06, 'learning_rate/action_model': 5.1894463036135775e-05, 'epoch': 1.31}) +[2026-06-17 23:48:51,230][starVLA.training.train_starvla][INFO] - Step 3579, Loss: {'train/loss': 0.6632436513900757, 'train/grad_norm_pre_clip': 19.79226303100586, 'timing/data': 0.001144547015428543, 'timing/model': 1.3339725751429796, 'learning_rate/qwen_vl_interface': 5.187283795112087e-06, 'learning_rate/action_model': 5.1872837951120866e-05, 'epoch': 1.31}) +[2026-06-17 23:48:56,198][starVLA.training.train_starvla][INFO] - Step 3580, Loss: {'train/loss': 0.686980128288269, 'train/grad_norm_pre_clip': 21.344581604003906, 'timing/data': 0.0010792547836899757, 'timing/model': 1.4312172401696444, 'learning_rate/qwen_vl_interface': 5.185121299611734e-06, 'learning_rate/action_model': 5.1851212996117335e-05, 'epoch': 1.31}) +[2026-06-17 23:49:00,594][starVLA.training.train_starvla][INFO] - Step 3581, Loss: {'train/loss': 0.3787488639354706, 'train/grad_norm_pre_clip': 11.759617805480957, 'timing/data': 0.0011687157675623894, 'timing/model': 1.3757792487740517, 'learning_rate/qwen_vl_interface': 5.182958817560803e-06, 'learning_rate/action_model': 5.1829588175608026e-05, 'epoch': 1.31}) +[2026-06-17 23:49:04,837][starVLA.training.train_starvla][INFO] - Step 3582, Loss: {'train/loss': 0.3485201597213745, 'train/grad_norm_pre_clip': 12.64260196685791, 'timing/data': 0.0008634515106678009, 'timing/model': 1.0850827852264047, 'learning_rate/qwen_vl_interface': 5.180796349407583e-06, 'learning_rate/action_model': 5.180796349407583e-05, 'epoch': 1.31}) +[2026-06-17 23:49:09,284][starVLA.training.train_starvla][INFO] - Step 3583, Loss: {'train/loss': 0.3322177231311798, 'train/grad_norm_pre_clip': 15.116931915283203, 'timing/data': 0.0012513119727373123, 'timing/model': 1.101552713662386, 'learning_rate/qwen_vl_interface': 5.178633895600354e-06, 'learning_rate/action_model': 5.178633895600353e-05, 'epoch': 1.31}) +[2026-06-17 23:49:14,403][starVLA.training.train_starvla][INFO] - Step 3584, Loss: {'train/loss': 0.43333375453948975, 'train/grad_norm_pre_clip': 25.657487869262695, 'timing/data': 0.0013073906302452087, 'timing/model': 1.5311623681336641, 'learning_rate/qwen_vl_interface': 5.176471456587394e-06, 'learning_rate/action_model': 5.176471456587394e-05, 'epoch': 1.31}) +[2026-06-17 23:49:22,439][starVLA.training.train_starvla][INFO] - Step 3585, Loss: {'train/loss': 0.46420150995254517, 'train/grad_norm_pre_clip': 14.638558387756348, 'timing/data': 0.0010553812608122826, 'timing/model': 1.1250474322587252, 'learning_rate/qwen_vl_interface': 5.1743090328169805e-06, 'learning_rate/action_model': 5.1743090328169805e-05, 'epoch': 1.31}) +[2026-06-17 23:49:27,429][starVLA.training.train_starvla][INFO] - Step 3586, Loss: {'train/loss': 0.6371592283248901, 'train/grad_norm_pre_clip': 20.12305450439453, 'timing/data': 0.0012056557461619377, 'timing/model': 1.56754644587636, 'learning_rate/qwen_vl_interface': 5.172146624737388e-06, 'learning_rate/action_model': 5.1721466247373875e-05, 'epoch': 1.31}) +[2026-06-17 23:49:37,898][starVLA.training.train_starvla][INFO] - Step 3587, Loss: {'train/loss': 0.32840272784233093, 'train/grad_norm_pre_clip': 13.512778282165527, 'timing/data': 0.0011868355795741081, 'timing/model': 4.623336154036224, 'learning_rate/qwen_vl_interface': 5.169984232796883e-06, 'learning_rate/action_model': 5.169984232796883e-05, 'epoch': 1.31}) +[2026-06-17 23:49:45,790][starVLA.training.train_starvla][INFO] - Step 3588, Loss: {'train/loss': 0.3891027569770813, 'train/grad_norm_pre_clip': 14.25706672668457, 'timing/data': 0.002105425111949444, 'timing/model': 1.0777106508612633, 'learning_rate/qwen_vl_interface': 5.1678218574437325e-06, 'learning_rate/action_model': 5.167821857443732e-05, 'epoch': 1.31}) +[2026-06-17 23:49:50,641][starVLA.training.train_starvla][INFO] - Step 3589, Loss: {'train/loss': 0.4443622827529907, 'train/grad_norm_pre_clip': 17.193696975708008, 'timing/data': 0.0015865545719861984, 'timing/model': 1.2655208986252546, 'learning_rate/qwen_vl_interface': 5.165659499126198e-06, 'learning_rate/action_model': 5.165659499126198e-05, 'epoch': 1.31}) +[2026-06-17 23:50:09,137][starVLA.training.train_starvla][INFO] - Step 3590, Loss: {'train/loss': 0.39061570167541504, 'train/grad_norm_pre_clip': 15.668476104736328, 'timing/data': 0.0010718852281570435, 'timing/model': 6.226397532038391, 'learning_rate/qwen_vl_interface': 5.1634971582925405e-06, 'learning_rate/action_model': 5.163497158292541e-05, 'epoch': 1.31}) +[2026-06-17 23:50:13,887][starVLA.training.train_starvla][INFO] - Step 3591, Loss: {'train/loss': 0.5305202007293701, 'train/grad_norm_pre_clip': 19.660219192504883, 'timing/data': 0.0013746144250035286, 'timing/model': 1.355188401415944, 'learning_rate/qwen_vl_interface': 5.161334835391016e-06, 'learning_rate/action_model': 5.161334835391016e-05, 'epoch': 1.31}) +[2026-06-17 23:50:26,990][starVLA.training.train_starvla][INFO] - Step 3592, Loss: {'train/loss': 0.26060184836387634, 'train/grad_norm_pre_clip': 12.829237937927246, 'timing/data': 0.0014685066416859627, 'timing/model': 1.4406911926344037, 'learning_rate/qwen_vl_interface': 5.159172530869875e-06, 'learning_rate/action_model': 5.159172530869874e-05, 'epoch': 1.32}) +[2026-06-17 23:50:35,751][starVLA.training.train_starvla][INFO] - Step 3593, Loss: {'train/loss': 0.4435487985610962, 'train/grad_norm_pre_clip': 14.180767059326172, 'timing/data': 0.0008900314569473267, 'timing/model': 5.096843730658293, 'learning_rate/qwen_vl_interface': 5.157010245177366e-06, 'learning_rate/action_model': 5.1570102451773664e-05, 'epoch': 1.32}) +[2026-06-17 23:50:40,420][starVLA.training.train_starvla][INFO] - Step 3594, Loss: {'train/loss': 0.4394305944442749, 'train/grad_norm_pre_clip': 19.088167190551758, 'timing/data': 0.0013583041727542877, 'timing/model': 1.060184725560248, 'learning_rate/qwen_vl_interface': 5.1548479787617346e-06, 'learning_rate/action_model': 5.1548479787617346e-05, 'epoch': 1.32}) +[2026-06-17 23:50:48,619][starVLA.training.train_starvla][INFO] - Step 3595, Loss: {'train/loss': 0.44032296538352966, 'train/grad_norm_pre_clip': 13.718588829040527, 'timing/data': 0.0008777771145105362, 'timing/model': 1.410234966315329, 'learning_rate/qwen_vl_interface': 5.152685732071219e-06, 'learning_rate/action_model': 5.1526857320712185e-05, 'epoch': 1.32}) +[2026-06-17 23:50:53,456][starVLA.training.train_starvla][INFO] - Step 3596, Loss: {'train/loss': 0.6533607244491577, 'train/grad_norm_pre_clip': 20.365659713745117, 'timing/data': 0.0008539799600839615, 'timing/model': 1.5502478079870343, 'learning_rate/qwen_vl_interface': 5.1505235055540545e-06, 'learning_rate/action_model': 5.1505235055540545e-05, 'epoch': 1.32}) +[2026-06-17 23:50:58,137][starVLA.training.train_starvla][INFO] - Step 3597, Loss: {'train/loss': 0.3565140664577484, 'train/grad_norm_pre_clip': 11.149930000305176, 'timing/data': 0.0074851857498288155, 'timing/model': 1.2232410553842783, 'learning_rate/qwen_vl_interface': 5.148361299658477e-06, 'learning_rate/action_model': 5.148361299658476e-05, 'epoch': 1.32}) +[2026-06-17 23:51:02,673][starVLA.training.train_starvla][INFO] - Step 3598, Loss: {'train/loss': 0.5949761271476746, 'train/grad_norm_pre_clip': 19.736160278320312, 'timing/data': 0.0013928059488534927, 'timing/model': 1.050481429323554, 'learning_rate/qwen_vl_interface': 5.1461991148327095e-06, 'learning_rate/action_model': 5.1461991148327095e-05, 'epoch': 1.32}) +[2026-06-17 23:51:07,533][starVLA.training.train_starvla][INFO] - Step 3599, Loss: {'train/loss': 0.29329127073287964, 'train/grad_norm_pre_clip': 12.706175804138184, 'timing/data': 0.0011026421561837196, 'timing/model': 1.1643828582018614, 'learning_rate/qwen_vl_interface': 5.144036951524977e-06, 'learning_rate/action_model': 5.144036951524977e-05, 'epoch': 1.32}) +[2026-06-17 23:55:04,649][starVLA.training.train_starvla][INFO] - Step 3600, Loss: {'train/loss': 0.2773517370223999, 'train/grad_norm_pre_clip': 10.939800262451172, 'eval/loss': 0.5871837735176086, 'eval/latency_2/loss': 0.5871837735176086, 'eval/demon_fire_precision': 0.8896551724137931, 'eval/demon_fire_recall': 0.9715481171548117, 'eval/demon_fire_f1': 0.9288, 'eval/demon_fire_rate_model': 0.7262103505843072, 'eval/demon_fire_rate_teacher': 0.6649972175848636, 'eval/demon_left_precision': 0.8521199586349535, 'eval/demon_left_recall': 0.8673684210526316, 'eval/demon_left_f1': 0.8596765779864372, 'eval/demon_left_rate_model': 0.26905954368391766, 'eval/demon_left_rate_teacher': 0.2643294379521425, 'eval/demon_right_precision': 0.864406779661017, 'eval/demon_right_recall': 0.8067796610169492, 'eval/demon_right_f1': 0.8345996493278786, 'eval/demon_right_rate_model': 0.22982749026154703, 'eval/demon_right_rate_teacher': 0.24624373956594323, 'eval/demon_move_f1': 0.8471381136571579, 'eval/demon_cc_f1': 0.9026947791618228, 'eval/latency_2/demon_fire_precision': 0.8896551724137931, 'eval/latency_2/demon_fire_recall': 0.9715481171548117, 'eval/latency_2/demon_fire_f1': 0.9288, 'eval/latency_2/demon_fire_rate_model': 0.7262103505843072, 'eval/latency_2/demon_fire_rate_teacher': 0.6649972175848636, 'eval/latency_2/demon_left_precision': 0.8521199586349535, 'eval/latency_2/demon_left_recall': 0.8673684210526316, 'eval/latency_2/demon_left_f1': 0.8596765779864372, 'eval/latency_2/demon_left_rate_model': 0.26905954368391766, 'eval/latency_2/demon_left_rate_teacher': 0.2643294379521425, 'eval/latency_2/demon_right_precision': 0.864406779661017, 'eval/latency_2/demon_right_recall': 0.8067796610169492, 'eval/latency_2/demon_right_f1': 0.8345996493278786, 'eval/latency_2/demon_right_rate_model': 0.22982749026154703, 'eval/latency_2/demon_right_rate_teacher': 0.24624373956594323, 'eval/latency_2/demon_move_f1': 0.8471381136571579, 'eval/latency_2/demon_cc_f1': 0.9026947791618228, 'timing/data': 0.001279185526072979, 'timing/model': 1.5222226055338979, 'learning_rate/qwen_vl_interface': 5.141874810183498e-06, 'learning_rate/action_model': 5.141874810183498e-05, 'epoch': 1.32}) +[2026-06-17 23:55:04,654][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_3600_state +[2026-06-17 23:55:04,662][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 23:55:13,019][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_3600_state/model.safetensors +[2026-06-17 23:55:27,813][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_3600_state/optimizer.bin +[2026-06-17 23:55:27,815][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_3600_state/sampler.bin +[2026-06-17 23:55:27,816][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_3600_state/sampler_1.bin +[2026-06-17 23:55:27,818][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_3600_state/random_states_0.pkl +[2026-06-17 23:55:27,824][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3600 +[2026-06-17 23:55:27,824][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 23:55:27,842][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 23:55:32,478][starVLA.training.train_starvla][INFO] - Step 3601, Loss: {'train/loss': 0.6016725897789001, 'train/grad_norm_pre_clip': 20.00942611694336, 'timing/data': 0.0012085922062397003, 'timing/model': 1.1489447243511677, 'learning_rate/qwen_vl_interface': 5.139712691256489e-06, 'learning_rate/action_model': 5.1397126912564885e-05, 'epoch': 1.32}) +[2026-06-17 23:55:40,719][starVLA.training.train_starvla][INFO] - Step 3602, Loss: {'train/loss': 0.6677004098892212, 'train/grad_norm_pre_clip': 20.147846221923828, 'timing/data': 0.0012874435633420944, 'timing/model': 1.617796710692346, 'learning_rate/qwen_vl_interface': 5.137550595192156e-06, 'learning_rate/action_model': 5.137550595192156e-05, 'epoch': 1.32}) +[2026-06-17 23:55:46,862][starVLA.training.train_starvla][INFO] - Step 3603, Loss: {'train/loss': 0.33456313610076904, 'train/grad_norm_pre_clip': 20.332054138183594, 'timing/data': 0.002638031728565693, 'timing/model': 1.2410661214962602, 'learning_rate/qwen_vl_interface': 5.135388522438707e-06, 'learning_rate/action_model': 5.1353885224387066e-05, 'epoch': 1.32}) +[2026-06-17 23:55:53,850][starVLA.training.train_starvla][INFO] - Step 3604, Loss: {'train/loss': 0.6743629574775696, 'train/grad_norm_pre_clip': 25.905670166015625, 'timing/data': 0.02656942792236805, 'timing/model': 1.1138107618317008, 'learning_rate/qwen_vl_interface': 5.133226473444341e-06, 'learning_rate/action_model': 5.13322647344434e-05, 'epoch': 1.32}) +[2026-06-17 23:56:03,977][starVLA.training.train_starvla][INFO] - Step 3605, Loss: {'train/loss': 0.650231122970581, 'train/grad_norm_pre_clip': 18.304508209228516, 'timing/data': 0.022909706458449364, 'timing/model': 4.799218527041376, 'learning_rate/qwen_vl_interface': 5.1310644486572515e-06, 'learning_rate/action_model': 5.131064448657251e-05, 'epoch': 1.32}) +[2026-06-17 23:56:12,716][starVLA.training.train_starvla][INFO] - Step 3606, Loss: {'train/loss': 0.4854981303215027, 'train/grad_norm_pre_clip': 20.134065628051758, 'timing/data': 0.09506694134324789, 'timing/model': 1.4870553649961948, 'learning_rate/qwen_vl_interface': 5.12890244852563e-06, 'learning_rate/action_model': 5.12890244852563e-05, 'epoch': 1.32}) +[2026-06-17 23:56:22,356][starVLA.training.train_starvla][INFO] - Step 3607, Loss: {'train/loss': 0.5503875613212585, 'train/grad_norm_pre_clip': 16.086637496948242, 'timing/data': 0.023164814338088036, 'timing/model': 1.557869660668075, 'learning_rate/qwen_vl_interface': 5.126740473497661e-06, 'learning_rate/action_model': 5.126740473497661e-05, 'epoch': 1.32}) +[2026-06-17 23:56:28,391][starVLA.training.train_starvla][INFO] - Step 3608, Loss: {'train/loss': 0.5942354202270508, 'train/grad_norm_pre_clip': 19.240732192993164, 'timing/data': 0.030356885865330696, 'timing/model': 1.049808606505394, 'learning_rate/qwen_vl_interface': 5.124578524021525e-06, 'learning_rate/action_model': 5.124578524021525e-05, 'epoch': 1.32}) +[2026-06-17 23:56:38,033][starVLA.training.train_starvla][INFO] - Step 3609, Loss: {'train/loss': 0.6122604608535767, 'train/grad_norm_pre_clip': 16.673837661743164, 'timing/data': 0.001222308725118637, 'timing/model': 1.2620367677882314, 'learning_rate/qwen_vl_interface': 5.122416600545396e-06, 'learning_rate/action_model': 5.122416600545397e-05, 'epoch': 1.32}) +[2026-06-17 23:56:48,209][starVLA.training.train_starvla][INFO] - Step 3610, Loss: {'train/loss': 0.5490348935127258, 'train/grad_norm_pre_clip': 18.85462760925293, 'timing/data': 0.0012382874265313148, 'timing/model': 1.4249686319380999, 'learning_rate/qwen_vl_interface': 5.1202547035174436e-06, 'learning_rate/action_model': 5.120254703517443e-05, 'epoch': 1.32}) +[2026-06-17 23:56:56,749][starVLA.training.train_starvla][INFO] - Step 3611, Loss: {'train/loss': 0.539524257183075, 'train/grad_norm_pre_clip': 21.871030807495117, 'timing/data': 0.0010193567723035812, 'timing/model': 1.2480786507949233, 'learning_rate/qwen_vl_interface': 5.118092833385829e-06, 'learning_rate/action_model': 5.1180928333858293e-05, 'epoch': 1.32}) +[2026-06-17 23:57:11,334][starVLA.training.train_starvla][INFO] - Step 3612, Loss: {'train/loss': 0.8915987610816956, 'train/grad_norm_pre_clip': 47.58022689819336, 'timing/data': 0.000921555794775486, 'timing/model': 1.71652574185282, 'learning_rate/qwen_vl_interface': 5.115930990598713e-06, 'learning_rate/action_model': 5.115930990598713e-05, 'epoch': 1.32}) +[2026-06-17 23:57:19,060][starVLA.training.train_starvla][INFO] - Step 3613, Loss: {'train/loss': 0.4728579819202423, 'train/grad_norm_pre_clip': 17.145448684692383, 'timing/data': 0.0011543715372681618, 'timing/model': 1.1491508278995752, 'learning_rate/qwen_vl_interface': 5.113769175604248e-06, 'learning_rate/action_model': 5.113769175604247e-05, 'epoch': 1.32}) +[2026-06-17 23:57:23,977][starVLA.training.train_starvla][INFO] - Step 3614, Loss: {'train/loss': 0.6352201700210571, 'train/grad_norm_pre_clip': 17.603940963745117, 'timing/data': 0.0013748109340667725, 'timing/model': 1.0616800170391798, 'learning_rate/qwen_vl_interface': 5.111607388850576e-06, 'learning_rate/action_model': 5.111607388850576e-05, 'epoch': 1.32}) +[2026-06-17 23:57:28,808][starVLA.training.train_starvla][INFO] - Step 3615, Loss: {'train/loss': 0.40477126836776733, 'train/grad_norm_pre_clip': 16.19249725341797, 'timing/data': 0.001468258909881115, 'timing/model': 1.0618256954476237, 'learning_rate/qwen_vl_interface': 5.1094456307858444e-06, 'learning_rate/action_model': 5.109445630785844e-05, 'epoch': 1.32}) +[2026-06-17 23:57:33,638][starVLA.training.train_starvla][INFO] - Step 3616, Loss: {'train/loss': 0.3224600851535797, 'train/grad_norm_pre_clip': 14.907285690307617, 'timing/data': 0.001035461202263832, 'timing/model': 1.1468126587569714, 'learning_rate/qwen_vl_interface': 5.107283901858184e-06, 'learning_rate/action_model': 5.107283901858184e-05, 'epoch': 1.32}) +[2026-06-17 23:57:38,208][starVLA.training.train_starvla][INFO] - Step 3617, Loss: {'train/loss': 0.3001909852027893, 'train/grad_norm_pre_clip': 15.90799617767334, 'timing/data': 0.001136397011578083, 'timing/model': 1.2363800900056958, 'learning_rate/qwen_vl_interface': 5.105122202515726e-06, 'learning_rate/action_model': 5.1051222025157255e-05, 'epoch': 1.32}) +[2026-06-17 23:57:42,997][starVLA.training.train_starvla][INFO] - Step 3618, Loss: {'train/loss': 0.8181190490722656, 'train/grad_norm_pre_clip': 25.143808364868164, 'timing/data': 0.0007466142997145653, 'timing/model': 1.0530242156237364, 'learning_rate/qwen_vl_interface': 5.102960533206588e-06, 'learning_rate/action_model': 5.1029605332065874e-05, 'epoch': 1.32}) +[2026-06-17 23:57:47,419][starVLA.training.train_starvla][INFO] - Step 3619, Loss: {'train/loss': 0.2935996353626251, 'train/grad_norm_pre_clip': 13.019288063049316, 'timing/data': 0.0011204583570361137, 'timing/model': 1.110298770479858, 'learning_rate/qwen_vl_interface': 5.100798894378891e-06, 'learning_rate/action_model': 5.100798894378891e-05, 'epoch': 1.33}) +[2026-06-17 23:57:52,109][starVLA.training.train_starvla][INFO] - Step 3620, Loss: {'train/loss': 0.46694058179855347, 'train/grad_norm_pre_clip': 20.936054229736328, 'timing/data': 0.0015133563429117203, 'timing/model': 1.073762989602983, 'learning_rate/qwen_vl_interface': 5.098637286480744e-06, 'learning_rate/action_model': 5.098637286480744e-05, 'epoch': 1.33}) +[2026-06-17 23:57:56,476][starVLA.training.train_starvla][INFO] - Step 3621, Loss: {'train/loss': 0.7513269186019897, 'train/grad_norm_pre_clip': 23.193395614624023, 'timing/data': 0.003298328258097172, 'timing/model': 1.0491479588672519, 'learning_rate/qwen_vl_interface': 5.096475709960252e-06, 'learning_rate/action_model': 5.096475709960251e-05, 'epoch': 1.33}) +[2026-06-17 23:58:00,985][starVLA.training.train_starvla][INFO] - Step 3622, Loss: {'train/loss': 0.3845115005970001, 'train/grad_norm_pre_clip': 28.853689193725586, 'timing/data': 0.001289677806198597, 'timing/model': 1.4193767132237554, 'learning_rate/qwen_vl_interface': 5.094314165265506e-06, 'learning_rate/action_model': 5.0943141652655056e-05, 'epoch': 1.33}) +[2026-06-17 23:58:05,097][starVLA.training.train_starvla][INFO] - Step 3623, Loss: {'train/loss': 0.4162089228630066, 'train/grad_norm_pre_clip': 23.369524002075195, 'timing/data': 0.0011167610064148903, 'timing/model': 1.0601530438289046, 'learning_rate/qwen_vl_interface': 5.092152652844606e-06, 'learning_rate/action_model': 5.092152652844606e-05, 'epoch': 1.33}) +[2026-06-17 23:58:09,684][starVLA.training.train_starvla][INFO] - Step 3624, Loss: {'train/loss': 0.5531908273696899, 'train/grad_norm_pre_clip': 21.82771873474121, 'timing/data': 0.0012889960780739784, 'timing/model': 1.0427852170541883, 'learning_rate/qwen_vl_interface': 5.089991173145628e-06, 'learning_rate/action_model': 5.0899911731456285e-05, 'epoch': 1.33}) +[2026-06-17 23:58:14,118][starVLA.training.train_starvla][INFO] - Step 3625, Loss: {'train/loss': 0.4301527440547943, 'train/grad_norm_pre_clip': 17.912540435791016, 'timing/data': 0.0010779285803437233, 'timing/model': 1.0795803247019649, 'learning_rate/qwen_vl_interface': 5.087829726616653e-06, 'learning_rate/action_model': 5.087829726616653e-05, 'epoch': 1.33}) +[2026-06-17 23:58:18,815][starVLA.training.train_starvla][INFO] - Step 3626, Loss: {'train/loss': 0.5750535130500793, 'train/grad_norm_pre_clip': 17.586822509765625, 'timing/data': 0.0011250218376517296, 'timing/model': 1.1481901863589883, 'learning_rate/qwen_vl_interface': 5.08566831370575e-06, 'learning_rate/action_model': 5.0856683137057504e-05, 'epoch': 1.33}) +[2026-06-17 23:58:23,612][starVLA.training.train_starvla][INFO] - Step 3627, Loss: {'train/loss': 0.4899181127548218, 'train/grad_norm_pre_clip': 18.203458786010742, 'timing/data': 0.001307949423789978, 'timing/model': 1.3463694024831057, 'learning_rate/qwen_vl_interface': 5.083506934860982e-06, 'learning_rate/action_model': 5.083506934860982e-05, 'epoch': 1.33}) +[2026-06-17 23:58:30,209][starVLA.training.train_starvla][INFO] - Step 3628, Loss: {'train/loss': 0.591041088104248, 'train/grad_norm_pre_clip': 16.044391632080078, 'timing/data': 0.0011411840096116066, 'timing/model': 1.6534317675977945, 'learning_rate/qwen_vl_interface': 5.081345590530407e-06, 'learning_rate/action_model': 5.081345590530407e-05, 'epoch': 1.33}) +[2026-06-17 23:58:44,744][starVLA.training.train_starvla][INFO] - Step 3629, Loss: {'train/loss': 0.8432159423828125, 'train/grad_norm_pre_clip': 16.862009048461914, 'timing/data': 0.0008298680186271667, 'timing/model': 4.22407113481313, 'learning_rate/qwen_vl_interface': 5.079184281162069e-06, 'learning_rate/action_model': 5.0791842811620694e-05, 'epoch': 1.33}) +[2026-06-17 23:58:53,123][starVLA.training.train_starvla][INFO] - Step 3630, Loss: {'train/loss': 0.6775482296943665, 'train/grad_norm_pre_clip': 18.482654571533203, 'timing/data': 0.000760306604206562, 'timing/model': 1.2518597282469273, 'learning_rate/qwen_vl_interface': 5.077023007204018e-06, 'learning_rate/action_model': 5.077023007204018e-05, 'epoch': 1.33}) +[2026-06-17 23:59:06,911][starVLA.training.train_starvla][INFO] - Step 3631, Loss: {'train/loss': 0.6352798342704773, 'train/grad_norm_pre_clip': 19.830949783325195, 'timing/data': 0.0013445988297462463, 'timing/model': 4.0049417316913605, 'learning_rate/qwen_vl_interface': 5.074861769104281e-06, 'learning_rate/action_model': 5.074861769104281e-05, 'epoch': 1.33}) +[2026-06-17 23:59:11,497][starVLA.training.train_starvla][INFO] - Step 3632, Loss: {'train/loss': 0.6555662751197815, 'train/grad_norm_pre_clip': 21.279674530029297, 'timing/data': 0.0011503593996167183, 'timing/model': 1.4422414181753993, 'learning_rate/qwen_vl_interface': 5.07270056731089e-06, 'learning_rate/action_model': 5.07270056731089e-05, 'epoch': 1.33}) +[2026-06-17 23:59:15,608][starVLA.training.train_starvla][INFO] - Step 3633, Loss: {'train/loss': 0.4101693630218506, 'train/grad_norm_pre_clip': 18.336881637573242, 'timing/data': 0.0010658325627446175, 'timing/model': 1.048392758704722, 'learning_rate/qwen_vl_interface': 5.070539402271859e-06, 'learning_rate/action_model': 5.070539402271859e-05, 'epoch': 1.33}) +[2026-06-17 23:59:20,072][starVLA.training.train_starvla][INFO] - Step 3634, Loss: {'train/loss': 0.7910047173500061, 'train/grad_norm_pre_clip': 23.495094299316406, 'timing/data': 0.0010917959734797478, 'timing/model': 1.0461459830403328, 'learning_rate/qwen_vl_interface': 5.068378274435205e-06, 'learning_rate/action_model': 5.068378274435205e-05, 'epoch': 1.33}) +[2026-06-17 23:59:28,728][starVLA.training.train_starvla][INFO] - Step 3635, Loss: {'train/loss': 0.42915141582489014, 'train/grad_norm_pre_clip': 15.14103889465332, 'timing/data': 0.0011987648904323578, 'timing/model': 1.0981739489361644, 'learning_rate/qwen_vl_interface': 5.066217184248931e-06, 'learning_rate/action_model': 5.0662171842489305e-05, 'epoch': 1.33}) +[2026-06-17 23:59:36,922][starVLA.training.train_starvla][INFO] - Step 3636, Loss: {'train/loss': 0.7310433983802795, 'train/grad_norm_pre_clip': 25.863014221191406, 'timing/data': 0.0011791018769145012, 'timing/model': 1.7255215030163527, 'learning_rate/qwen_vl_interface': 5.0640561321610295e-06, 'learning_rate/action_model': 5.064056132161029e-05, 'epoch': 1.33}) +[2026-06-17 23:59:49,763][starVLA.training.train_starvla][INFO] - Step 3637, Loss: {'train/loss': 0.3589278757572174, 'train/grad_norm_pre_clip': 10.635625839233398, 'timing/data': 0.0009917905554175377, 'timing/model': 1.4899324104189873, 'learning_rate/qwen_vl_interface': 5.061895118619493e-06, 'learning_rate/action_model': 5.061895118619493e-05, 'epoch': 1.33}) +[2026-06-17 23:59:53,876][starVLA.training.train_starvla][INFO] - Step 3638, Loss: {'train/loss': 0.5189486145973206, 'train/grad_norm_pre_clip': 16.578523635864258, 'timing/data': 0.0009653596207499504, 'timing/model': 1.081593749113381, 'learning_rate/qwen_vl_interface': 5.059734144072302e-06, 'learning_rate/action_model': 5.059734144072302e-05, 'epoch': 1.33}) +[2026-06-17 23:59:58,454][starVLA.training.train_starvla][INFO] - Step 3639, Loss: {'train/loss': 0.398636132478714, 'train/grad_norm_pre_clip': 19.017465591430664, 'timing/data': 0.0012298496440052986, 'timing/model': 1.0988306067883968, 'learning_rate/qwen_vl_interface': 5.0575732089674275e-06, 'learning_rate/action_model': 5.057573208967428e-05, 'epoch': 1.33}) +[2026-06-18 00:00:02,867][starVLA.training.train_starvla][INFO] - Step 3640, Loss: {'train/loss': 0.4375130832195282, 'train/grad_norm_pre_clip': 15.835915565490723, 'timing/data': 0.0008965199813246727, 'timing/model': 1.0444350549951196, 'learning_rate/qwen_vl_interface': 5.055412313752832e-06, 'learning_rate/action_model': 5.055412313752832e-05, 'epoch': 1.33}) +[2026-06-18 00:00:07,267][starVLA.training.train_starvla][INFO] - Step 3641, Loss: {'train/loss': 0.48725923895835876, 'train/grad_norm_pre_clip': 34.02188491821289, 'timing/data': 0.0008910149335861206, 'timing/model': 1.0660381615161896, 'learning_rate/qwen_vl_interface': 5.053251458876474e-06, 'learning_rate/action_model': 5.0532514588764744e-05, 'epoch': 1.33}) +[2026-06-18 00:00:12,051][starVLA.training.train_starvla][INFO] - Step 3642, Loss: {'train/loss': 0.27303898334503174, 'train/grad_norm_pre_clip': 11.384862899780273, 'timing/data': 0.0010788198560476303, 'timing/model': 1.3969580680131912, 'learning_rate/qwen_vl_interface': 5.0510906447863025e-06, 'learning_rate/action_model': 5.051090644786303e-05, 'epoch': 1.33}) +[2026-06-18 00:00:16,310][starVLA.training.train_starvla][INFO] - Step 3643, Loss: {'train/loss': 0.6459424495697021, 'train/grad_norm_pre_clip': 15.982614517211914, 'timing/data': 0.0011405646800994873, 'timing/model': 1.0664548389613628, 'learning_rate/qwen_vl_interface': 5.048929871930252e-06, 'learning_rate/action_model': 5.048929871930251e-05, 'epoch': 1.33}) +[2026-06-18 00:00:20,747][starVLA.training.train_starvla][INFO] - Step 3644, Loss: {'train/loss': 0.4255824685096741, 'train/grad_norm_pre_clip': 13.100358963012695, 'timing/data': 0.0010297400876879692, 'timing/model': 1.0266267154365778, 'learning_rate/qwen_vl_interface': 5.046769140756257e-06, 'learning_rate/action_model': 5.046769140756257e-05, 'epoch': 1.33}) +[2026-06-18 00:00:25,301][starVLA.training.train_starvla][INFO] - Step 3645, Loss: {'train/loss': 0.5770345330238342, 'train/grad_norm_pre_clip': 17.425434112548828, 'timing/data': 0.0014445986598730087, 'timing/model': 1.1366709033027291, 'learning_rate/qwen_vl_interface': 5.044608451712237e-06, 'learning_rate/action_model': 5.044608451712237e-05, 'epoch': 1.33}) +[2026-06-18 00:00:30,301][starVLA.training.train_starvla][INFO] - Step 3646, Loss: {'train/loss': 0.44788089394569397, 'train/grad_norm_pre_clip': 17.159317016601562, 'timing/data': 0.0008768606930971146, 'timing/model': 1.4046042747795582, 'learning_rate/qwen_vl_interface': 5.042447805246105e-06, 'learning_rate/action_model': 5.042447805246106e-05, 'epoch': 1.34}) +[2026-06-18 00:00:34,704][starVLA.training.train_starvla][INFO] - Step 3647, Loss: {'train/loss': 0.46539950370788574, 'train/grad_norm_pre_clip': 19.125112533569336, 'timing/data': 0.0029122019186615944, 'timing/model': 1.402199512347579, 'learning_rate/qwen_vl_interface': 5.040287201805767e-06, 'learning_rate/action_model': 5.040287201805767e-05, 'epoch': 1.34}) +[2026-06-18 00:00:52,028][starVLA.training.train_starvla][INFO] - Step 3648, Loss: {'train/loss': 0.7518498301506042, 'train/grad_norm_pre_clip': 19.862972259521484, 'timing/data': 0.0009285835549235344, 'timing/model': 1.1951111759990454, 'learning_rate/qwen_vl_interface': 5.038126641839119e-06, 'learning_rate/action_model': 5.0381266418391184e-05, 'epoch': 1.34}) +[2026-06-18 00:00:56,822][starVLA.training.train_starvla][INFO] - Step 3649, Loss: {'train/loss': 0.8385704159736633, 'train/grad_norm_pre_clip': 20.228628158569336, 'timing/data': 0.0009021218866109848, 'timing/model': 1.0977618359029293, 'learning_rate/qwen_vl_interface': 5.0359661257940455e-06, 'learning_rate/action_model': 5.035966125794045e-05, 'epoch': 1.34}) +[2026-06-18 00:01:02,692][starVLA.training.train_starvla][INFO] - Step 3650, Loss: {'train/loss': 0.4434145987033844, 'train/grad_norm_pre_clip': 19.346553802490234, 'timing/data': 0.0012227818369865417, 'timing/model': 1.0505995592102408, 'learning_rate/qwen_vl_interface': 5.033805654118424e-06, 'learning_rate/action_model': 5.033805654118424e-05, 'epoch': 1.34}) +[2026-06-18 00:01:07,622][starVLA.training.train_starvla][INFO] - Step 3651, Loss: {'train/loss': 0.762902557849884, 'train/grad_norm_pre_clip': 21.604278564453125, 'timing/data': 0.0008879657834768295, 'timing/model': 1.242680854164064, 'learning_rate/qwen_vl_interface': 5.031645227260122e-06, 'learning_rate/action_model': 5.0316452272601224e-05, 'epoch': 1.34}) +[2026-06-18 00:01:12,593][starVLA.training.train_starvla][INFO] - Step 3652, Loss: {'train/loss': 0.5864052772521973, 'train/grad_norm_pre_clip': 21.915922164916992, 'timing/data': 0.0013409415259957314, 'timing/model': 1.5256746895611286, 'learning_rate/qwen_vl_interface': 5.029484845667001e-06, 'learning_rate/action_model': 5.029484845667001e-05, 'epoch': 1.34}) +[2026-06-18 00:01:23,185][starVLA.training.train_starvla][INFO] - Step 3653, Loss: {'train/loss': 0.5958338379859924, 'train/grad_norm_pre_clip': 17.208284378051758, 'timing/data': 0.0011162227019667625, 'timing/model': 4.181623031385243, 'learning_rate/qwen_vl_interface': 5.027324509786908e-06, 'learning_rate/action_model': 5.027324509786908e-05, 'epoch': 1.34}) +[2026-06-18 00:01:29,310][starVLA.training.train_starvla][INFO] - Step 3654, Loss: {'train/loss': 0.6333675980567932, 'train/grad_norm_pre_clip': 17.08305549621582, 'timing/data': 0.0012276703491806984, 'timing/model': 1.1039451174438, 'learning_rate/qwen_vl_interface': 5.025164220067683e-06, 'learning_rate/action_model': 5.025164220067683e-05, 'epoch': 1.34}) +[2026-06-18 00:01:33,793][starVLA.training.train_starvla][INFO] - Step 3655, Loss: {'train/loss': 0.5415199398994446, 'train/grad_norm_pre_clip': 16.554468154907227, 'timing/data': 0.0010400963947176933, 'timing/model': 1.0611676154658198, 'learning_rate/qwen_vl_interface': 5.023003976957159e-06, 'learning_rate/action_model': 5.023003976957159e-05, 'epoch': 1.34}) +[2026-06-18 00:01:38,316][starVLA.training.train_starvla][INFO] - Step 3656, Loss: {'train/loss': 0.6934861540794373, 'train/grad_norm_pre_clip': 18.574939727783203, 'timing/data': 0.0008686631917953491, 'timing/model': 1.0549172507598996, 'learning_rate/qwen_vl_interface': 5.020843780903155e-06, 'learning_rate/action_model': 5.020843780903155e-05, 'epoch': 1.34}) +[2026-06-18 00:01:46,042][starVLA.training.train_starvla][INFO] - Step 3657, Loss: {'train/loss': 0.3556651473045349, 'train/grad_norm_pre_clip': 16.845916748046875, 'timing/data': 0.0010203830897808075, 'timing/model': 1.2594285774976015, 'learning_rate/qwen_vl_interface': 5.01868363235348e-06, 'learning_rate/action_model': 5.01868363235348e-05, 'epoch': 1.34}) +[2026-06-18 00:01:50,893][starVLA.training.train_starvla][INFO] - Step 3658, Loss: {'train/loss': 0.41952353715896606, 'train/grad_norm_pre_clip': 14.996885299682617, 'timing/data': 0.0009442903101444244, 'timing/model': 1.4150179792195559, 'learning_rate/qwen_vl_interface': 5.016523531755937e-06, 'learning_rate/action_model': 5.016523531755937e-05, 'epoch': 1.34}) +[2026-06-18 00:01:59,692][starVLA.training.train_starvla][INFO] - Step 3659, Loss: {'train/loss': 0.35994574427604675, 'train/grad_norm_pre_clip': 15.001701354980469, 'timing/data': 0.0009296601638197899, 'timing/model': 4.026088984683156, 'learning_rate/qwen_vl_interface': 5.014363479558318e-06, 'learning_rate/action_model': 5.014363479558318e-05, 'epoch': 1.34}) +[2026-06-18 00:02:04,495][starVLA.training.train_starvla][INFO] - Step 3660, Loss: {'train/loss': 0.40198224782943726, 'train/grad_norm_pre_clip': 12.017717361450195, 'timing/data': 0.0013914583250880241, 'timing/model': 1.3296119468286633, 'learning_rate/qwen_vl_interface': 5.012203476208403e-06, 'learning_rate/action_model': 5.0122034762084025e-05, 'epoch': 1.34}) +[2026-06-18 00:02:09,121][starVLA.training.train_starvla][INFO] - Step 3661, Loss: {'train/loss': 0.6884733438491821, 'train/grad_norm_pre_clip': 20.01949119567871, 'timing/data': 0.000915108248591423, 'timing/model': 1.0855773650109768, 'learning_rate/qwen_vl_interface': 5.010043522153965e-06, 'learning_rate/action_model': 5.0100435221539646e-05, 'epoch': 1.34}) +[2026-06-18 00:02:15,864][starVLA.training.train_starvla][INFO] - Step 3662, Loss: {'train/loss': 0.5303539037704468, 'train/grad_norm_pre_clip': 17.004655838012695, 'timing/data': 0.0008908472955226898, 'timing/model': 1.5305938497185707, 'learning_rate/qwen_vl_interface': 5.007883617842759e-06, 'learning_rate/action_model': 5.007883617842759e-05, 'epoch': 1.34}) +[2026-06-18 00:02:21,468][starVLA.training.train_starvla][INFO] - Step 3663, Loss: {'train/loss': 0.4084807336330414, 'train/grad_norm_pre_clip': 16.26936912536621, 'timing/data': 0.0010716374963521957, 'timing/model': 2.264850246720016, 'learning_rate/qwen_vl_interface': 5.005723763722543e-06, 'learning_rate/action_model': 5.005723763722543e-05, 'epoch': 1.34}) +[2026-06-18 00:02:30,806][starVLA.training.train_starvla][INFO] - Step 3664, Loss: {'train/loss': 0.413263201713562, 'train/grad_norm_pre_clip': 15.013386726379395, 'timing/data': 0.0008892454206943512, 'timing/model': 1.1485431836917996, 'learning_rate/qwen_vl_interface': 5.003563960241052e-06, 'learning_rate/action_model': 5.003563960241052e-05, 'epoch': 1.34}) +[2026-06-18 00:02:35,396][starVLA.training.train_starvla][INFO] - Step 3665, Loss: {'train/loss': 0.37594443559646606, 'train/grad_norm_pre_clip': 14.58533763885498, 'timing/data': 0.0013525495305657387, 'timing/model': 1.1290114894509315, 'learning_rate/qwen_vl_interface': 5.001404207846017e-06, 'learning_rate/action_model': 5.001404207846017e-05, 'epoch': 1.34}) +[2026-06-18 00:02:39,854][starVLA.training.train_starvla][INFO] - Step 3666, Loss: {'train/loss': 0.5268105864524841, 'train/grad_norm_pre_clip': 20.218305587768555, 'timing/data': 0.001129000447690487, 'timing/model': 1.038040284998715, 'learning_rate/qwen_vl_interface': 4.999244506985158e-06, 'learning_rate/action_model': 4.999244506985158e-05, 'epoch': 1.34}) +[2026-06-18 00:02:44,599][starVLA.training.train_starvla][INFO] - Step 3667, Loss: {'train/loss': 0.5991860032081604, 'train/grad_norm_pre_clip': 17.200626373291016, 'timing/data': 0.0015382478013634682, 'timing/model': 1.3587541244924068, 'learning_rate/qwen_vl_interface': 4.997084858106182e-06, 'learning_rate/action_model': 4.997084858106183e-05, 'epoch': 1.34}) +[2026-06-18 00:02:48,695][starVLA.training.train_starvla][INFO] - Step 3668, Loss: {'train/loss': 0.5626686811447144, 'train/grad_norm_pre_clip': 21.44228172302246, 'timing/data': 0.0010603833943605423, 'timing/model': 1.0364023735746741, 'learning_rate/qwen_vl_interface': 4.994925261656789e-06, 'learning_rate/action_model': 4.994925261656789e-05, 'epoch': 1.34}) +[2026-06-18 00:03:00,386][starVLA.training.train_starvla][INFO] - Step 3669, Loss: {'train/loss': 0.38532471656799316, 'train/grad_norm_pre_clip': 14.879575729370117, 'timing/data': 0.000927257351577282, 'timing/model': 1.2925585145130754, 'learning_rate/qwen_vl_interface': 4.99276571808466e-06, 'learning_rate/action_model': 4.9927657180846605e-05, 'epoch': 1.34}) +[2026-06-18 00:03:06,497][starVLA.training.train_starvla][INFO] - Step 3670, Loss: {'train/loss': 0.26754477620124817, 'train/grad_norm_pre_clip': 16.093765258789062, 'timing/data': 0.0010386928915977478, 'timing/model': 1.072631391696632, 'learning_rate/qwen_vl_interface': 4.9906062278374775e-06, 'learning_rate/action_model': 4.990606227837478e-05, 'epoch': 1.34}) +[2026-06-18 00:03:11,096][starVLA.training.train_starvla][INFO] - Step 3671, Loss: {'train/loss': 0.38895943760871887, 'train/grad_norm_pre_clip': 17.019149780273438, 'timing/data': 0.0010497970506548882, 'timing/model': 1.0485992636531591, 'learning_rate/qwen_vl_interface': 4.988446791362903e-06, 'learning_rate/action_model': 4.988446791362903e-05, 'epoch': 1.34}) +[2026-06-18 00:03:15,750][starVLA.training.train_starvla][INFO] - Step 3672, Loss: {'train/loss': 0.6811266541481018, 'train/grad_norm_pre_clip': 42.19721603393555, 'timing/data': 0.0007919035851955414, 'timing/model': 1.3319167532026768, 'learning_rate/qwen_vl_interface': 4.986287409108589e-06, 'learning_rate/action_model': 4.9862874091085893e-05, 'epoch': 1.34}) +[2026-06-18 00:03:21,665][starVLA.training.train_starvla][INFO] - Step 3673, Loss: {'train/loss': 0.47414177656173706, 'train/grad_norm_pre_clip': 18.392915725708008, 'timing/data': 0.0013983305543661118, 'timing/model': 2.572925252839923, 'learning_rate/qwen_vl_interface': 4.984128081522179e-06, 'learning_rate/action_model': 4.984128081522178e-05, 'epoch': 1.34}) +[2026-06-18 00:03:26,177][starVLA.training.train_starvla][INFO] - Step 3674, Loss: {'train/loss': 0.5117814540863037, 'train/grad_norm_pre_clip': 18.70661163330078, 'timing/data': 0.0008002109825611115, 'timing/model': 1.0202648658305407, 'learning_rate/qwen_vl_interface': 4.981968809051304e-06, 'learning_rate/action_model': 4.981968809051304e-05, 'epoch': 1.35}) +[2026-06-18 00:03:32,745][starVLA.training.train_starvla][INFO] - Step 3675, Loss: {'train/loss': 0.5678908228874207, 'train/grad_norm_pre_clip': 18.692575454711914, 'timing/data': 0.0010630302131175995, 'timing/model': 2.9833788191899657, 'learning_rate/qwen_vl_interface': 4.979809592143583e-06, 'learning_rate/action_model': 4.979809592143583e-05, 'epoch': 1.35}) +[2026-06-18 00:03:37,138][starVLA.training.train_starvla][INFO] - Step 3676, Loss: {'train/loss': 0.5383076071739197, 'train/grad_norm_pre_clip': 20.695226669311523, 'timing/data': 0.0011270157992839813, 'timing/model': 1.3966349205002189, 'learning_rate/qwen_vl_interface': 4.977650431246624e-06, 'learning_rate/action_model': 4.977650431246624e-05, 'epoch': 1.35}) +[2026-06-18 00:03:41,707][starVLA.training.train_starvla][INFO] - Step 3677, Loss: {'train/loss': 0.6930691003799438, 'train/grad_norm_pre_clip': 22.348501205444336, 'timing/data': 0.001098322682082653, 'timing/model': 1.3724654112011194, 'learning_rate/qwen_vl_interface': 4.9754913268080255e-06, 'learning_rate/action_model': 4.975491326808025e-05, 'epoch': 1.35}) +[2026-06-18 00:03:47,142][starVLA.training.train_starvla][INFO] - Step 3678, Loss: {'train/loss': 0.2529248595237732, 'train/grad_norm_pre_clip': 16.99836540222168, 'timing/data': 0.0010745860636234283, 'timing/model': 2.261369344778359, 'learning_rate/qwen_vl_interface': 4.973332279275369e-06, 'learning_rate/action_model': 4.973332279275369e-05, 'epoch': 1.35}) +[2026-06-18 00:03:54,819][starVLA.training.train_starvla][INFO] - Step 3679, Loss: {'train/loss': 0.4054659903049469, 'train/grad_norm_pre_clip': 44.919925689697266, 'timing/data': 0.0013200808316469193, 'timing/model': 1.1693980563431978, 'learning_rate/qwen_vl_interface': 4.971173289096229e-06, 'learning_rate/action_model': 4.9711732890962286e-05, 'epoch': 1.35}) +[2026-06-18 00:03:59,189][starVLA.training.train_starvla][INFO] - Step 3680, Loss: {'train/loss': 0.37780365347862244, 'train/grad_norm_pre_clip': 22.613739013671875, 'timing/data': 0.0016416171565651894, 'timing/model': 1.0418177945539355, 'learning_rate/qwen_vl_interface': 4.969014356718165e-06, 'learning_rate/action_model': 4.969014356718165e-05, 'epoch': 1.35}) +[2026-06-18 00:04:04,271][starVLA.training.train_starvla][INFO] - Step 3681, Loss: {'train/loss': 0.5675057172775269, 'train/grad_norm_pre_clip': 16.878326416015625, 'timing/data': 0.0009114649146795273, 'timing/model': 1.1297521386295557, 'learning_rate/qwen_vl_interface': 4.9668554825887285e-06, 'learning_rate/action_model': 4.966855482588728e-05, 'epoch': 1.35}) +[2026-06-18 00:04:11,955][starVLA.training.train_starvla][INFO] - Step 3682, Loss: {'train/loss': 0.37515968084335327, 'train/grad_norm_pre_clip': 12.606085777282715, 'timing/data': 0.0011188546195626259, 'timing/model': 1.4252031976357102, 'learning_rate/qwen_vl_interface': 4.964696667155455e-06, 'learning_rate/action_model': 4.964696667155454e-05, 'epoch': 1.35}) +[2026-06-18 00:04:16,691][starVLA.training.train_starvla][INFO] - Step 3683, Loss: {'train/loss': 0.6704776883125305, 'train/grad_norm_pre_clip': 20.784914016723633, 'timing/data': 0.001490575261414051, 'timing/model': 1.3235368467867374, 'learning_rate/qwen_vl_interface': 4.96253791086587e-06, 'learning_rate/action_model': 4.96253791086587e-05, 'epoch': 1.35}) +[2026-06-18 00:04:21,547][starVLA.training.train_starvla][INFO] - Step 3684, Loss: {'train/loss': 0.7185002565383911, 'train/grad_norm_pre_clip': 22.15546989440918, 'timing/data': 0.0010671541094779968, 'timing/model': 1.1522806407883763, 'learning_rate/qwen_vl_interface': 4.960379214167483e-06, 'learning_rate/action_model': 4.960379214167483e-05, 'epoch': 1.35}) +[2026-06-18 00:04:30,149][starVLA.training.train_starvla][INFO] - Step 3685, Loss: {'train/loss': 0.2176700085401535, 'train/grad_norm_pre_clip': 10.243697166442871, 'timing/data': 0.0014146370813250542, 'timing/model': 1.1363964723423123, 'learning_rate/qwen_vl_interface': 4.9582205775077994e-06, 'learning_rate/action_model': 4.958220577507799e-05, 'epoch': 1.35}) +[2026-06-18 00:04:36,888][starVLA.training.train_starvla][INFO] - Step 3686, Loss: {'train/loss': 0.1785023957490921, 'train/grad_norm_pre_clip': 7.504225254058838, 'timing/data': 0.0011821836233139038, 'timing/model': 1.0391994807869196, 'learning_rate/qwen_vl_interface': 4.956062001334302e-06, 'learning_rate/action_model': 4.956062001334302e-05, 'epoch': 1.35}) +[2026-06-18 00:04:45,736][starVLA.training.train_starvla][INFO] - Step 3687, Loss: {'train/loss': 0.48534390330314636, 'train/grad_norm_pre_clip': 19.6798038482666, 'timing/data': 0.0010533398017287254, 'timing/model': 5.596134329214692, 'learning_rate/qwen_vl_interface': 4.953903486094469e-06, 'learning_rate/action_model': 4.953903486094469e-05, 'epoch': 1.35}) +[2026-06-18 00:04:53,372][starVLA.training.train_starvla][INFO] - Step 3688, Loss: {'train/loss': 0.6496515870094299, 'train/grad_norm_pre_clip': 22.036056518554688, 'timing/data': 0.0012241676449775696, 'timing/model': 1.2365933628752828, 'learning_rate/qwen_vl_interface': 4.951745032235762e-06, 'learning_rate/action_model': 4.951745032235762e-05, 'epoch': 1.35}) +[2026-06-18 00:04:58,014][starVLA.training.train_starvla][INFO] - Step 3689, Loss: {'train/loss': 0.4885502755641937, 'train/grad_norm_pre_clip': 18.42081642150879, 'timing/data': 0.0009814482182264328, 'timing/model': 1.0698534455150366, 'learning_rate/qwen_vl_interface': 4.949586640205631e-06, 'learning_rate/action_model': 4.94958664020563e-05, 'epoch': 1.35}) +[2026-06-18 00:05:02,544][starVLA.training.train_starvla][INFO] - Step 3690, Loss: {'train/loss': 0.44548657536506653, 'train/grad_norm_pre_clip': 14.95174503326416, 'timing/data': 0.0013141091912984848, 'timing/model': 1.0318097155541182, 'learning_rate/qwen_vl_interface': 4.9474283104515124e-06, 'learning_rate/action_model': 4.9474283104515126e-05, 'epoch': 1.35}) +[2026-06-18 00:05:07,186][starVLA.training.train_starvla][INFO] - Step 3691, Loss: {'train/loss': 0.3166429102420807, 'train/grad_norm_pre_clip': 14.368234634399414, 'timing/data': 0.0010145949199795723, 'timing/model': 1.083784124813974, 'learning_rate/qwen_vl_interface': 4.945270043420829e-06, 'learning_rate/action_model': 4.945270043420828e-05, 'epoch': 1.35}) +[2026-06-18 00:05:11,758][starVLA.training.train_starvla][INFO] - Step 3692, Loss: {'train/loss': 0.4485926032066345, 'train/grad_norm_pre_clip': 17.536466598510742, 'timing/data': 0.0010077925398945808, 'timing/model': 1.4765408262610435, 'learning_rate/qwen_vl_interface': 4.9431118395609954e-06, 'learning_rate/action_model': 4.9431118395609954e-05, 'epoch': 1.35}) +[2026-06-18 00:05:16,128][starVLA.training.train_starvla][INFO] - Step 3693, Loss: {'train/loss': 0.5881767868995667, 'train/grad_norm_pre_clip': 17.658946990966797, 'timing/data': 0.0008086962625384331, 'timing/model': 1.2388234548270702, 'learning_rate/qwen_vl_interface': 4.940953699319407e-06, 'learning_rate/action_model': 4.940953699319407e-05, 'epoch': 1.35}) +[2026-06-18 00:05:20,622][starVLA.training.train_starvla][INFO] - Step 3694, Loss: {'train/loss': 0.6800237894058228, 'train/grad_norm_pre_clip': 22.051788330078125, 'timing/data': 0.0011426378041505814, 'timing/model': 1.0558291794732213, 'learning_rate/qwen_vl_interface': 4.93879562314345e-06, 'learning_rate/action_model': 4.93879562314345e-05, 'epoch': 1.35}) +[2026-06-18 00:05:25,138][starVLA.training.train_starvla][INFO] - Step 3695, Loss: {'train/loss': 0.4715539216995239, 'train/grad_norm_pre_clip': 15.982019424438477, 'timing/data': 0.001101636327803135, 'timing/model': 1.1894485028460622, 'learning_rate/qwen_vl_interface': 4.936637611480495e-06, 'learning_rate/action_model': 4.9366376114804945e-05, 'epoch': 1.35}) +[2026-06-18 00:05:30,911][starVLA.training.train_starvla][INFO] - Step 3696, Loss: {'train/loss': 0.1867893785238266, 'train/grad_norm_pre_clip': 16.792802810668945, 'timing/data': 0.0015152981504797935, 'timing/model': 1.0969212977215648, 'learning_rate/qwen_vl_interface': 4.9344796647779e-06, 'learning_rate/action_model': 4.9344796647778995e-05, 'epoch': 1.35}) +[2026-06-18 00:05:37,537][starVLA.training.train_starvla][INFO] - Step 3697, Loss: {'train/loss': 0.3875211179256439, 'train/grad_norm_pre_clip': 14.390633583068848, 'timing/data': 0.001523081213235855, 'timing/model': 1.647547940723598, 'learning_rate/qwen_vl_interface': 4.9323217834830105e-06, 'learning_rate/action_model': 4.9323217834830105e-05, 'epoch': 1.35}) +[2026-06-18 00:05:41,998][starVLA.training.train_starvla][INFO] - Step 3698, Loss: {'train/loss': 0.55564945936203, 'train/grad_norm_pre_clip': 23.74430274963379, 'timing/data': 0.0007963133975863457, 'timing/model': 1.0636087851598859, 'learning_rate/qwen_vl_interface': 4.930163968043155e-06, 'learning_rate/action_model': 4.930163968043155e-05, 'epoch': 1.35}) +[2026-06-18 00:05:54,887][starVLA.training.train_starvla][INFO] - Step 3699, Loss: {'train/loss': 0.3024454116821289, 'train/grad_norm_pre_clip': 15.727474212646484, 'timing/data': 0.0008768895640969276, 'timing/model': 1.2107110666111112, 'learning_rate/qwen_vl_interface': 4.928006218905656e-06, 'learning_rate/action_model': 4.928006218905656e-05, 'epoch': 1.35}) +[2026-06-18 00:11:01,467][starVLA.training.train_starvla][INFO] - Step 3700, Loss: {'train/loss': 0.3091403841972351, 'train/grad_norm_pre_clip': 12.095626831054688, 'eval/loss': 0.5890346765518188, 'eval/latency_2/loss': 0.5890346765518188, 'eval/demon_fire_precision': 0.9316487158243579, 'eval/demon_fire_recall': 0.9410041841004184, 'eval/demon_fire_f1': 0.9363030807660283, 'eval/demon_fire_rate_model': 0.6716750139120757, 'eval/demon_fire_rate_teacher': 0.6649972175848636, 'eval/demon_left_precision': 0.8718510405257394, 'eval/demon_left_recall': 0.8378947368421052, 'eval/demon_left_f1': 0.8545356951154053, 'eval/demon_left_rate_model': 0.2540345019476906, 'eval/demon_left_rate_teacher': 0.2643294379521425, 'eval/demon_right_precision': 0.8849878934624698, 'eval/demon_right_recall': 0.8259887005649718, 'eval/demon_right_f1': 0.8544710695499708, 'eval/demon_right_rate_model': 0.22982749026154703, 'eval/demon_right_rate_teacher': 0.24624373956594323, 'eval/demon_move_f1': 0.854503382332688, 'eval/demon_cc_f1': 0.9101646719898544, 'eval/latency_2/demon_fire_precision': 0.9316487158243579, 'eval/latency_2/demon_fire_recall': 0.9410041841004184, 'eval/latency_2/demon_fire_f1': 0.9363030807660283, 'eval/latency_2/demon_fire_rate_model': 0.6716750139120757, 'eval/latency_2/demon_fire_rate_teacher': 0.6649972175848636, 'eval/latency_2/demon_left_precision': 0.8718510405257394, 'eval/latency_2/demon_left_recall': 0.8378947368421052, 'eval/latency_2/demon_left_f1': 0.8545356951154053, 'eval/latency_2/demon_left_rate_model': 0.2540345019476906, 'eval/latency_2/demon_left_rate_teacher': 0.2643294379521425, 'eval/latency_2/demon_right_precision': 0.8849878934624698, 'eval/latency_2/demon_right_recall': 0.8259887005649718, 'eval/latency_2/demon_right_f1': 0.8544710695499708, 'eval/latency_2/demon_right_rate_model': 0.22982749026154703, 'eval/latency_2/demon_right_rate_teacher': 0.24624373956594323, 'eval/latency_2/demon_move_f1': 0.854503382332688, 'eval/latency_2/demon_cc_f1': 0.9101646719898544, 'timing/data': 0.001046961173415184, 'timing/model': 1.0760637298226357, 'learning_rate/qwen_vl_interface': 4.9258485365178135e-06, 'learning_rate/action_model': 4.925848536517814e-05, 'epoch': 1.35}) +[2026-06-18 00:11:01,469][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_3700_state +[2026-06-18 00:11:01,477][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 00:11:10,616][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_3700_state/model.safetensors +[2026-06-18 00:11:25,871][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_3700_state/optimizer.bin +[2026-06-18 00:11:25,886][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_3700_state/sampler.bin +[2026-06-18 00:11:25,886][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_3700_state/sampler_1.bin +[2026-06-18 00:11:25,889][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_3700_state/random_states_0.pkl +[2026-06-18 00:11:25,896][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3700 +[2026-06-18 00:11:25,896][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 00:11:25,913][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 00:11:32,305][starVLA.training.train_starvla][INFO] - Step 3701, Loss: {'train/loss': 0.5995903611183167, 'train/grad_norm_pre_clip': 21.842823028564453, 'timing/data': 0.0012767035514116287, 'timing/model': 1.1791309155523777, 'learning_rate/qwen_vl_interface': 4.923690921326918e-06, 'learning_rate/action_model': 4.9236909213269175e-05, 'epoch': 1.36}) +[2026-06-18 00:11:37,766][starVLA.training.train_starvla][INFO] - Step 3702, Loss: {'train/loss': 0.46163254976272583, 'train/grad_norm_pre_clip': 17.866119384765625, 'timing/data': 0.0015459861606359482, 'timing/model': 1.6060851300135255, 'learning_rate/qwen_vl_interface': 4.921533373780244e-06, 'learning_rate/action_model': 4.921533373780244e-05, 'epoch': 1.36}) +[2026-06-18 00:11:42,761][starVLA.training.train_starvla][INFO] - Step 3703, Loss: {'train/loss': 0.6619398593902588, 'train/grad_norm_pre_clip': 26.91019630432129, 'timing/data': 0.0013818610459566116, 'timing/model': 1.5775742279365659, 'learning_rate/qwen_vl_interface': 4.919375894325057e-06, 'learning_rate/action_model': 4.9193758943250566e-05, 'epoch': 1.36}) +[2026-06-18 00:11:47,183][starVLA.training.train_starvla][INFO] - Step 3704, Loss: {'train/loss': 0.5860278010368347, 'train/grad_norm_pre_clip': 20.085580825805664, 'timing/data': 0.0010927943512797356, 'timing/model': 1.0573573634028435, 'learning_rate/qwen_vl_interface': 4.917218483408603e-06, 'learning_rate/action_model': 4.9172184834086035e-05, 'epoch': 1.36}) +[2026-06-18 00:11:58,146][starVLA.training.train_starvla][INFO] - Step 3705, Loss: {'train/loss': 0.44773128628730774, 'train/grad_norm_pre_clip': 19.277095794677734, 'timing/data': 0.0014345990493893623, 'timing/model': 2.8861779142171144, 'learning_rate/qwen_vl_interface': 4.915061141478114e-06, 'learning_rate/action_model': 4.915061141478113e-05, 'epoch': 1.36}) +[2026-06-18 00:12:03,097][starVLA.training.train_starvla][INFO] - Step 3706, Loss: {'train/loss': 0.625717282295227, 'train/grad_norm_pre_clip': 62.20009994506836, 'timing/data': 0.0012471498921513557, 'timing/model': 1.1857451694086194, 'learning_rate/qwen_vl_interface': 4.912903868980812e-06, 'learning_rate/action_model': 4.9129038689808115e-05, 'epoch': 1.36}) +[2026-06-18 00:12:08,402][starVLA.training.train_starvla][INFO] - Step 3707, Loss: {'train/loss': 0.6623215675354004, 'train/grad_norm_pre_clip': 24.200956344604492, 'timing/data': 0.0009573837742209435, 'timing/model': 1.0472743408754468, 'learning_rate/qwen_vl_interface': 4.9107466663639005e-06, 'learning_rate/action_model': 4.910746666363901e-05, 'epoch': 1.36}) +[2026-06-18 00:12:13,418][starVLA.training.train_starvla][INFO] - Step 3708, Loss: {'train/loss': 0.7850378751754761, 'train/grad_norm_pre_clip': 26.182334899902344, 'timing/data': 0.000751810148358345, 'timing/model': 1.5222122324630618, 'learning_rate/qwen_vl_interface': 4.9085895340745705e-06, 'learning_rate/action_model': 4.9085895340745705e-05, 'epoch': 1.36}) +[2026-06-18 00:12:17,640][starVLA.training.train_starvla][INFO] - Step 3709, Loss: {'train/loss': 0.5753258466720581, 'train/grad_norm_pre_clip': 16.917661666870117, 'timing/data': 0.0008598864078521729, 'timing/model': 1.0331038972362876, 'learning_rate/qwen_vl_interface': 4.906432472559996e-06, 'learning_rate/action_model': 4.9064324725599955e-05, 'epoch': 1.36}) +[2026-06-18 00:12:22,995][starVLA.training.train_starvla][INFO] - Step 3710, Loss: {'train/loss': 0.3978399634361267, 'train/grad_norm_pre_clip': 19.16651153564453, 'timing/data': 0.0010319575667381287, 'timing/model': 2.126451979391277, 'learning_rate/qwen_vl_interface': 4.90427548226734e-06, 'learning_rate/action_model': 4.9042754822673396e-05, 'epoch': 1.36}) +[2026-06-18 00:12:30,456][starVLA.training.train_starvla][INFO] - Step 3711, Loss: {'train/loss': 0.9779458045959473, 'train/grad_norm_pre_clip': 26.394182205200195, 'timing/data': 0.001047387719154358, 'timing/model': 1.0100894821807742, 'learning_rate/qwen_vl_interface': 4.902118563643751e-06, 'learning_rate/action_model': 4.902118563643751e-05, 'epoch': 1.36}) +[2026-06-18 00:12:35,069][starVLA.training.train_starvla][INFO] - Step 3712, Loss: {'train/loss': 0.515513002872467, 'train/grad_norm_pre_clip': 16.841529846191406, 'timing/data': 0.006348147988319397, 'timing/model': 1.6550903739407659, 'learning_rate/qwen_vl_interface': 4.899961717136356e-06, 'learning_rate/action_model': 4.899961717136355e-05, 'epoch': 1.36}) +[2026-06-18 00:12:40,061][starVLA.training.train_starvla][INFO] - Step 3713, Loss: {'train/loss': 0.4685188829898834, 'train/grad_norm_pre_clip': 16.298118591308594, 'timing/data': 0.0010363617911934853, 'timing/model': 1.4055300941690803, 'learning_rate/qwen_vl_interface': 4.897804943192273e-06, 'learning_rate/action_model': 4.897804943192273e-05, 'epoch': 1.36}) +[2026-06-18 00:12:45,547][starVLA.training.train_starvla][INFO] - Step 3714, Loss: {'train/loss': 0.39624279737472534, 'train/grad_norm_pre_clip': 18.33563995361328, 'timing/data': 0.0010031666606664658, 'timing/model': 2.3560781655833125, 'learning_rate/qwen_vl_interface': 4.895648242258605e-06, 'learning_rate/action_model': 4.895648242258605e-05, 'epoch': 1.36}) +[2026-06-18 00:12:49,892][starVLA.training.train_starvla][INFO] - Step 3715, Loss: {'train/loss': 0.4551126956939697, 'train/grad_norm_pre_clip': 19.33681297302246, 'timing/data': 0.0010440833866596222, 'timing/model': 1.0975639540702105, 'learning_rate/qwen_vl_interface': 4.893491614782438e-06, 'learning_rate/action_model': 4.8934916147824375e-05, 'epoch': 1.36}) +[2026-06-18 00:12:56,137][starVLA.training.train_starvla][INFO] - Step 3716, Loss: {'train/loss': 0.8405854105949402, 'train/grad_norm_pre_clip': 26.861379623413086, 'timing/data': 0.00568812433630228, 'timing/model': 1.1189409904181957, 'learning_rate/qwen_vl_interface': 4.891335061210841e-06, 'learning_rate/action_model': 4.891335061210841e-05, 'epoch': 1.36}) +[2026-06-18 00:13:00,671][starVLA.training.train_starvla][INFO] - Step 3717, Loss: {'train/loss': 0.4313436448574066, 'train/grad_norm_pre_clip': 17.841123580932617, 'timing/data': 0.0009829988703131676, 'timing/model': 1.0454686796292663, 'learning_rate/qwen_vl_interface': 4.889178581990874e-06, 'learning_rate/action_model': 4.8891785819908734e-05, 'epoch': 1.36}) +[2026-06-18 00:13:05,232][starVLA.training.train_starvla][INFO] - Step 3718, Loss: {'train/loss': 0.5441126227378845, 'train/grad_norm_pre_clip': 19.115148544311523, 'timing/data': 0.0011318791657686234, 'timing/model': 1.37795237172395, 'learning_rate/qwen_vl_interface': 4.887022177569576e-06, 'learning_rate/action_model': 4.8870221775695754e-05, 'epoch': 1.36}) +[2026-06-18 00:13:14,512][starVLA.training.train_starvla][INFO] - Step 3719, Loss: {'train/loss': 0.3398468792438507, 'train/grad_norm_pre_clip': 13.39063835144043, 'timing/data': 0.0008553247898817062, 'timing/model': 1.1559903714805841, 'learning_rate/qwen_vl_interface': 4.88486584839397e-06, 'learning_rate/action_model': 4.8848658483939695e-05, 'epoch': 1.36}) +[2026-06-18 00:13:21,066][starVLA.training.train_starvla][INFO] - Step 3720, Loss: {'train/loss': 0.7320554256439209, 'train/grad_norm_pre_clip': 30.667510986328125, 'timing/data': 0.0013874312862753868, 'timing/model': 1.0746272914111614, 'learning_rate/qwen_vl_interface': 4.882709594911065e-06, 'learning_rate/action_model': 4.8827095949110655e-05, 'epoch': 1.36}) +[2026-06-18 00:13:26,095][starVLA.training.train_starvla][INFO] - Step 3721, Loss: {'train/loss': 0.68375563621521, 'train/grad_norm_pre_clip': 27.66111946105957, 'timing/data': 0.0012323623523116112, 'timing/model': 1.0647544981911778, 'learning_rate/qwen_vl_interface': 4.880553417567858e-06, 'learning_rate/action_model': 4.880553417567858e-05, 'epoch': 1.36}) +[2026-06-18 00:13:30,603][starVLA.training.train_starvla][INFO] - Step 3722, Loss: {'train/loss': 0.5076491236686707, 'train/grad_norm_pre_clip': 16.652435302734375, 'timing/data': 0.0009503401815891266, 'timing/model': 1.070663028396666, 'learning_rate/qwen_vl_interface': 4.878397316811324e-06, 'learning_rate/action_model': 4.878397316811324e-05, 'epoch': 1.36}) +[2026-06-18 00:13:35,645][starVLA.training.train_starvla][INFO] - Step 3723, Loss: {'train/loss': 0.3502242863178253, 'train/grad_norm_pre_clip': 50.33832931518555, 'timing/data': 0.0009777620434761047, 'timing/model': 1.7653352515771985, 'learning_rate/qwen_vl_interface': 4.876241293088428e-06, 'learning_rate/action_model': 4.876241293088428e-05, 'epoch': 1.36}) +[2026-06-18 00:13:40,617][starVLA.training.train_starvla][INFO] - Step 3724, Loss: {'train/loss': 0.5552375912666321, 'train/grad_norm_pre_clip': 22.857837677001953, 'timing/data': 0.0010627172887325287, 'timing/model': 1.401045081205666, 'learning_rate/qwen_vl_interface': 4.874085346846111e-06, 'learning_rate/action_model': 4.87408534684611e-05, 'epoch': 1.36}) +[2026-06-18 00:13:45,220][starVLA.training.train_starvla][INFO] - Step 3725, Loss: {'train/loss': 0.45125582814216614, 'train/grad_norm_pre_clip': 19.497154235839844, 'timing/data': 0.00106826052069664, 'timing/model': 1.0526083055883646, 'learning_rate/qwen_vl_interface': 4.871929478531309e-06, 'learning_rate/action_model': 4.871929478531309e-05, 'epoch': 1.36}) +[2026-06-18 00:13:50,158][starVLA.training.train_starvla][INFO] - Step 3726, Loss: {'train/loss': 0.6772260665893555, 'train/grad_norm_pre_clip': 16.952024459838867, 'timing/data': 0.0017916476354002953, 'timing/model': 1.032053355127573, 'learning_rate/qwen_vl_interface': 4.869773688590932e-06, 'learning_rate/action_model': 4.8697736885909325e-05, 'epoch': 1.36}) +[2026-06-18 00:13:54,618][starVLA.training.train_starvla][INFO] - Step 3727, Loss: {'train/loss': 0.2588595449924469, 'train/grad_norm_pre_clip': 19.319746017456055, 'timing/data': 0.0012140395119786263, 'timing/model': 1.0804774416610599, 'learning_rate/qwen_vl_interface': 4.867617977471878e-06, 'learning_rate/action_model': 4.867617977471878e-05, 'epoch': 1.36}) +[2026-06-18 00:13:58,690][starVLA.training.train_starvla][INFO] - Step 3728, Loss: {'train/loss': 0.4996207356452942, 'train/grad_norm_pre_clip': 19.08626937866211, 'timing/data': 0.0009901812300086021, 'timing/model': 1.0466986149549484, 'learning_rate/qwen_vl_interface': 4.865462345621029e-06, 'learning_rate/action_model': 4.865462345621029e-05, 'epoch': 1.37}) +[2026-06-18 00:14:03,284][starVLA.training.train_starvla][INFO] - Step 3729, Loss: {'train/loss': 0.3831285238265991, 'train/grad_norm_pre_clip': 17.649982452392578, 'timing/data': 0.0009287605062127113, 'timing/model': 1.0447947764769197, 'learning_rate/qwen_vl_interface': 4.86330679348525e-06, 'learning_rate/action_model': 4.863306793485249e-05, 'epoch': 1.37}) +[2026-06-18 00:14:07,914][starVLA.training.train_starvla][INFO] - Step 3730, Loss: {'train/loss': 0.4739338457584381, 'train/grad_norm_pre_clip': 13.501163482666016, 'timing/data': 0.0010746261104941368, 'timing/model': 1.2455955650657415, 'learning_rate/qwen_vl_interface': 4.8611513215113894e-06, 'learning_rate/action_model': 4.861151321511389e-05, 'epoch': 1.37}) +[2026-06-18 00:14:12,454][starVLA.training.train_starvla][INFO] - Step 3731, Loss: {'train/loss': 0.306104451417923, 'train/grad_norm_pre_clip': 15.058579444885254, 'timing/data': 0.0007762052118778229, 'timing/model': 1.046869314275682, 'learning_rate/qwen_vl_interface': 4.858995930146276e-06, 'learning_rate/action_model': 4.858995930146275e-05, 'epoch': 1.37}) +[2026-06-18 00:14:16,881][starVLA.training.train_starvla][INFO] - Step 3732, Loss: {'train/loss': 0.3511047065258026, 'train/grad_norm_pre_clip': 21.465059280395508, 'timing/data': 0.0011682529002428055, 'timing/model': 1.043007081374526, 'learning_rate/qwen_vl_interface': 4.856840619836729e-06, 'learning_rate/action_model': 4.856840619836728e-05, 'epoch': 1.37}) +[2026-06-18 00:14:21,444][starVLA.training.train_starvla][INFO] - Step 3733, Loss: {'train/loss': 0.3407766819000244, 'train/grad_norm_pre_clip': 15.826544761657715, 'timing/data': 0.0011492911726236343, 'timing/model': 1.4437753083184361, 'learning_rate/qwen_vl_interface': 4.854685391029543e-06, 'learning_rate/action_model': 4.854685391029543e-05, 'epoch': 1.37}) +[2026-06-18 00:14:25,735][starVLA.training.train_starvla][INFO] - Step 3734, Loss: {'train/loss': 0.7892604470252991, 'train/grad_norm_pre_clip': 25.54570960998535, 'timing/data': 0.0011558374390006065, 'timing/model': 1.1631545517593622, 'learning_rate/qwen_vl_interface': 4.852530244171502e-06, 'learning_rate/action_model': 4.852530244171502e-05, 'epoch': 1.37}) +[2026-06-18 00:14:31,342][starVLA.training.train_starvla][INFO] - Step 3735, Loss: {'train/loss': 0.8516258597373962, 'train/grad_norm_pre_clip': 25.396650314331055, 'timing/data': 0.000911966897547245, 'timing/model': 1.1535214083269238, 'learning_rate/qwen_vl_interface': 4.850375179709368e-06, 'learning_rate/action_model': 4.8503751797093674e-05, 'epoch': 1.37}) +[2026-06-18 00:14:36,623][starVLA.training.train_starvla][INFO] - Step 3736, Loss: {'train/loss': 0.5188653469085693, 'train/grad_norm_pre_clip': 17.179683685302734, 'timing/data': 0.0007525281980633736, 'timing/model': 1.242228701710701, 'learning_rate/qwen_vl_interface': 4.84822019808989e-06, 'learning_rate/action_model': 4.84822019808989e-05, 'epoch': 1.37}) +[2026-06-18 00:14:41,882][starVLA.training.train_starvla][INFO] - Step 3737, Loss: {'train/loss': 0.862392783164978, 'train/grad_norm_pre_clip': 25.261550903320312, 'timing/data': 0.0008971253409981728, 'timing/model': 1.2290724106132984, 'learning_rate/qwen_vl_interface': 4.846065299759799e-06, 'learning_rate/action_model': 4.846065299759799e-05, 'epoch': 1.37}) +[2026-06-18 00:14:53,663][starVLA.training.train_starvla][INFO] - Step 3738, Loss: {'train/loss': 0.5597401261329651, 'train/grad_norm_pre_clip': 21.73625946044922, 'timing/data': 0.001547299325466156, 'timing/model': 3.488348520360887, 'learning_rate/qwen_vl_interface': 4.843910485165802e-06, 'learning_rate/action_model': 4.843910485165802e-05, 'epoch': 1.37}) +[2026-06-18 00:14:58,931][starVLA.training.train_starvla][INFO] - Step 3739, Loss: {'train/loss': 0.37459105253219604, 'train/grad_norm_pre_clip': 18.491230010986328, 'timing/data': 0.0012017013505101204, 'timing/model': 1.5376478554680943, 'learning_rate/qwen_vl_interface': 4.8417557547546015e-06, 'learning_rate/action_model': 4.8417557547546015e-05, 'epoch': 1.37}) +[2026-06-18 00:15:18,927][starVLA.training.train_starvla][INFO] - Step 3740, Loss: {'train/loss': 0.528961181640625, 'train/grad_norm_pre_clip': 18.15913200378418, 'timing/data': 0.004841343499720097, 'timing/model': 4.800214841961861, 'learning_rate/qwen_vl_interface': 4.839601108972872e-06, 'learning_rate/action_model': 4.839601108972871e-05, 'epoch': 1.37}) +[2026-06-18 00:15:30,703][starVLA.training.train_starvla][INFO] - Step 3741, Loss: {'train/loss': 0.6612072587013245, 'train/grad_norm_pre_clip': 20.09618377685547, 'timing/data': 0.0009245341643691063, 'timing/model': 1.3454255554825068, 'learning_rate/qwen_vl_interface': 4.837446548267275e-06, 'learning_rate/action_model': 4.837446548267274e-05, 'epoch': 1.37}) +[2026-06-18 00:15:37,207][starVLA.training.train_starvla][INFO] - Step 3742, Loss: {'train/loss': 0.7134430408477783, 'train/grad_norm_pre_clip': 17.833492279052734, 'timing/data': 0.0008824532851576805, 'timing/model': 1.8638089429587126, 'learning_rate/qwen_vl_interface': 4.835292073084449e-06, 'learning_rate/action_model': 4.8352920730844486e-05, 'epoch': 1.37}) +[2026-06-18 00:15:48,695][starVLA.training.train_starvla][INFO] - Step 3743, Loss: {'train/loss': 0.5100127458572388, 'train/grad_norm_pre_clip': 18.615909576416016, 'timing/data': 0.0033684587106108665, 'timing/model': 1.6735386066138744, 'learning_rate/qwen_vl_interface': 4.833137683871025e-06, 'learning_rate/action_model': 4.833137683871025e-05, 'epoch': 1.37}) +[2026-06-18 00:15:56,530][starVLA.training.train_starvla][INFO] - Step 3744, Loss: {'train/loss': 0.7105522751808167, 'train/grad_norm_pre_clip': 27.03331184387207, 'timing/data': 0.0008377926424145699, 'timing/model': 2.9730804516002536, 'learning_rate/qwen_vl_interface': 4.830983381073609e-06, 'learning_rate/action_model': 4.830983381073608e-05, 'epoch': 1.37}) +[2026-06-18 00:16:07,133][starVLA.training.train_starvla][INFO] - Step 3745, Loss: {'train/loss': 0.6431128978729248, 'train/grad_norm_pre_clip': 16.939048767089844, 'timing/data': 0.0014558183029294014, 'timing/model': 4.629078389145434, 'learning_rate/qwen_vl_interface': 4.828829165138787e-06, 'learning_rate/action_model': 4.828829165138787e-05, 'epoch': 1.37}) +[2026-06-18 00:16:22,007][starVLA.training.train_starvla][INFO] - Step 3746, Loss: {'train/loss': 0.5779803395271301, 'train/grad_norm_pre_clip': 15.352985382080078, 'timing/data': 0.0010267700999975204, 'timing/model': 6.110728175379336, 'learning_rate/qwen_vl_interface': 4.826675036513132e-06, 'learning_rate/action_model': 4.826675036513132e-05, 'epoch': 1.37}) +[2026-06-18 00:16:35,075][starVLA.training.train_starvla][INFO] - Step 3747, Loss: {'train/loss': 0.7486732006072998, 'train/grad_norm_pre_clip': 24.099401473999023, 'timing/data': 0.0008160462602972984, 'timing/model': 5.128664709627628, 'learning_rate/qwen_vl_interface': 4.824520995643198e-06, 'learning_rate/action_model': 4.8245209956431984e-05, 'epoch': 1.37}) +[2026-06-18 00:16:40,170][starVLA.training.train_starvla][INFO] - Step 3748, Loss: {'train/loss': 0.4107567369937897, 'train/grad_norm_pre_clip': 15.346113204956055, 'timing/data': 0.0012585092335939407, 'timing/model': 1.4392430279403925, 'learning_rate/qwen_vl_interface': 4.82236704297552e-06, 'learning_rate/action_model': 4.82236704297552e-05, 'epoch': 1.37}) +[2026-06-18 00:16:44,352][starVLA.training.train_starvla][INFO] - Step 3749, Loss: {'train/loss': 0.5142568945884705, 'train/grad_norm_pre_clip': 17.946508407592773, 'timing/data': 0.0011759893968701363, 'timing/model': 1.096954665146768, 'learning_rate/qwen_vl_interface': 4.820213178956613e-06, 'learning_rate/action_model': 4.820213178956612e-05, 'epoch': 1.37}) +[2026-06-18 00:16:49,314][starVLA.training.train_starvla][INFO] - Step 3750, Loss: {'train/loss': 0.3785896599292755, 'train/grad_norm_pre_clip': 13.609848976135254, 'timing/data': 0.0008111204952001572, 'timing/model': 1.1076038517057896, 'learning_rate/qwen_vl_interface': 4.818059404032977e-06, 'learning_rate/action_model': 4.818059404032977e-05, 'epoch': 1.37}) +[2026-06-18 00:16:53,898][starVLA.training.train_starvla][INFO] - Step 3751, Loss: {'train/loss': 0.5818954706192017, 'train/grad_norm_pre_clip': 19.828916549682617, 'timing/data': 0.0011813435703516006, 'timing/model': 1.0792178679257631, 'learning_rate/qwen_vl_interface': 4.815905718651092e-06, 'learning_rate/action_model': 4.815905718651092e-05, 'epoch': 1.37}) +[2026-06-18 00:16:58,490][starVLA.training.train_starvla][INFO] - Step 3752, Loss: {'train/loss': 0.454492449760437, 'train/grad_norm_pre_clip': 17.885601043701172, 'timing/data': 0.0012737838551402092, 'timing/model': 1.046484999358654, 'learning_rate/qwen_vl_interface': 4.813752123257419e-06, 'learning_rate/action_model': 4.813752123257418e-05, 'epoch': 1.37}) +[2026-06-18 00:17:03,310][starVLA.training.train_starvla][INFO] - Step 3753, Loss: {'train/loss': 0.4314091205596924, 'train/grad_norm_pre_clip': 19.725202560424805, 'timing/data': 0.003336602821946144, 'timing/model': 1.458877146244049, 'learning_rate/qwen_vl_interface': 4.811598618298398e-06, 'learning_rate/action_model': 4.811598618298398e-05, 'epoch': 1.37}) +[2026-06-18 00:17:07,489][starVLA.training.train_starvla][INFO] - Step 3754, Loss: {'train/loss': 0.48673132061958313, 'train/grad_norm_pre_clip': 15.545575141906738, 'timing/data': 0.0011152476072311401, 'timing/model': 1.0441338876262307, 'learning_rate/qwen_vl_interface': 4.809445204220457e-06, 'learning_rate/action_model': 4.809445204220457e-05, 'epoch': 1.37}) +[2026-06-18 00:17:12,520][starVLA.training.train_starvla][INFO] - Step 3755, Loss: {'train/loss': 0.7789323329925537, 'train/grad_norm_pre_clip': 22.484149932861328, 'timing/data': 0.0012692799791693687, 'timing/model': 1.5295557724311948, 'learning_rate/qwen_vl_interface': 4.8072918814699985e-06, 'learning_rate/action_model': 4.8072918814699986e-05, 'epoch': 1.37}) +[2026-06-18 00:17:21,171][starVLA.training.train_starvla][INFO] - Step 3756, Loss: {'train/loss': 0.5358126163482666, 'train/grad_norm_pre_clip': 22.73851203918457, 'timing/data': 0.0016070082783699036, 'timing/model': 1.3608005233108997, 'learning_rate/qwen_vl_interface': 4.80513865049341e-06, 'learning_rate/action_model': 4.80513865049341e-05, 'epoch': 1.38}) +[2026-06-18 00:17:26,391][starVLA.training.train_starvla][INFO] - Step 3757, Loss: {'train/loss': 0.3050779402256012, 'train/grad_norm_pre_clip': 11.943781852722168, 'timing/data': 0.002104056067764759, 'timing/model': 1.2201173743233085, 'learning_rate/qwen_vl_interface': 4.802985511737057e-06, 'learning_rate/action_model': 4.802985511737057e-05, 'epoch': 1.38}) +[2026-06-18 00:17:33,088][starVLA.training.train_starvla][INFO] - Step 3758, Loss: {'train/loss': 0.40073031187057495, 'train/grad_norm_pre_clip': 14.619542121887207, 'timing/data': 0.0017532194033265114, 'timing/model': 1.2479884512722492, 'learning_rate/qwen_vl_interface': 4.80083246564729e-06, 'learning_rate/action_model': 4.8008324656472894e-05, 'epoch': 1.38}) +[2026-06-18 00:17:38,356][starVLA.training.train_starvla][INFO] - Step 3759, Loss: {'train/loss': 0.5034822821617126, 'train/grad_norm_pre_clip': 22.028711318969727, 'timing/data': 0.0010988060384988785, 'timing/model': 1.230895508080721, 'learning_rate/qwen_vl_interface': 4.798679512670436e-06, 'learning_rate/action_model': 4.798679512670435e-05, 'epoch': 1.38}) +[2026-06-18 00:17:43,110][starVLA.training.train_starvla][INFO] - Step 3760, Loss: {'train/loss': 0.6201791763305664, 'train/grad_norm_pre_clip': 19.486291885375977, 'timing/data': 0.0010685725137591362, 'timing/model': 1.157512211240828, 'learning_rate/qwen_vl_interface': 4.796526653252802e-06, 'learning_rate/action_model': 4.796526653252802e-05, 'epoch': 1.38}) +[2026-06-18 00:17:47,690][starVLA.training.train_starvla][INFO] - Step 3761, Loss: {'train/loss': 0.5579807162284851, 'train/grad_norm_pre_clip': 21.49412727355957, 'timing/data': 0.0008765710517764091, 'timing/model': 1.0590236000716686, 'learning_rate/qwen_vl_interface': 4.794373887840683e-06, 'learning_rate/action_model': 4.794373887840682e-05, 'epoch': 1.38}) +[2026-06-18 00:17:55,889][starVLA.training.train_starvla][INFO] - Step 3762, Loss: {'train/loss': 0.5848937034606934, 'train/grad_norm_pre_clip': 17.377384185791016, 'timing/data': 0.0013270750641822815, 'timing/model': 1.1456707827746868, 'learning_rate/qwen_vl_interface': 4.792221216880347e-06, 'learning_rate/action_model': 4.792221216880347e-05, 'epoch': 1.38}) +[2026-06-18 00:18:00,424][starVLA.training.train_starvla][INFO] - Step 3763, Loss: {'train/loss': 0.34560900926589966, 'train/grad_norm_pre_clip': 11.847764015197754, 'timing/data': 0.0012367982417345047, 'timing/model': 1.446625605225563, 'learning_rate/qwen_vl_interface': 4.790068640818046e-06, 'learning_rate/action_model': 4.790068640818045e-05, 'epoch': 1.38}) +[2026-06-18 00:18:05,748][starVLA.training.train_starvla][INFO] - Step 3764, Loss: {'train/loss': 0.4021293520927429, 'train/grad_norm_pre_clip': 13.887153625488281, 'timing/data': 0.0014408593997359276, 'timing/model': 1.0679230075329542, 'learning_rate/qwen_vl_interface': 4.787916160100008e-06, 'learning_rate/action_model': 4.787916160100008e-05, 'epoch': 1.38}) +[2026-06-18 00:18:12,469][starVLA.training.train_starvla][INFO] - Step 3765, Loss: {'train/loss': 0.4746118187904358, 'train/grad_norm_pre_clip': 15.159649848937988, 'timing/data': 0.001928696408867836, 'timing/model': 2.7189448922872543, 'learning_rate/qwen_vl_interface': 4.78576377517245e-06, 'learning_rate/action_model': 4.7857637751724504e-05, 'epoch': 1.38}) +[2026-06-18 00:18:25,074][starVLA.training.train_starvla][INFO] - Step 3766, Loss: {'train/loss': 0.39271020889282227, 'train/grad_norm_pre_clip': 17.443958282470703, 'timing/data': 0.0009642289951443672, 'timing/model': 2.8169260071590543, 'learning_rate/qwen_vl_interface': 4.783611486481561e-06, 'learning_rate/action_model': 4.7836114864815604e-05, 'epoch': 1.38}) +[2026-06-18 00:18:30,358][starVLA.training.train_starvla][INFO] - Step 3767, Loss: {'train/loss': 0.5354300737380981, 'train/grad_norm_pre_clip': 24.56707763671875, 'timing/data': 0.001098431646823883, 'timing/model': 1.4659842001274228, 'learning_rate/qwen_vl_interface': 4.781459294473511e-06, 'learning_rate/action_model': 4.781459294473511e-05, 'epoch': 1.38}) +[2026-06-18 00:18:38,771][starVLA.training.train_starvla][INFO] - Step 3768, Loss: {'train/loss': 0.44623300433158875, 'train/grad_norm_pre_clip': 18.418210983276367, 'timing/data': 0.0007755700498819351, 'timing/model': 1.412636872380972, 'learning_rate/qwen_vl_interface': 4.779307199594454e-06, 'learning_rate/action_model': 4.779307199594454e-05, 'epoch': 1.38}) +[2026-06-18 00:18:44,146][starVLA.training.train_starvla][INFO] - Step 3769, Loss: {'train/loss': 0.31609129905700684, 'train/grad_norm_pre_clip': 12.35299015045166, 'timing/data': 0.0013738740235567093, 'timing/model': 1.28021907992661, 'learning_rate/qwen_vl_interface': 4.777155202290522e-06, 'learning_rate/action_model': 4.777155202290521e-05, 'epoch': 1.38}) +[2026-06-18 00:18:49,184][starVLA.training.train_starvla][INFO] - Step 3770, Loss: {'train/loss': 0.5057409405708313, 'train/grad_norm_pre_clip': 16.597475051879883, 'timing/data': 0.0012319479137659073, 'timing/model': 1.068273832090199, 'learning_rate/qwen_vl_interface': 4.775003303007824e-06, 'learning_rate/action_model': 4.775003303007824e-05, 'epoch': 1.38}) +[2026-06-18 00:18:58,111][starVLA.training.train_starvla][INFO] - Step 3771, Loss: {'train/loss': 0.5103641748428345, 'train/grad_norm_pre_clip': 20.6485652923584, 'timing/data': 0.0010958127677440643, 'timing/model': 2.9681865544989705, 'learning_rate/qwen_vl_interface': 4.772851502192453e-06, 'learning_rate/action_model': 4.772851502192452e-05, 'epoch': 1.38}) +[2026-06-18 00:19:03,312][starVLA.training.train_starvla][INFO] - Step 3772, Loss: {'train/loss': 0.5656668543815613, 'train/grad_norm_pre_clip': 19.00782012939453, 'timing/data': 0.0009209327399730682, 'timing/model': 1.0837402464821935, 'learning_rate/qwen_vl_interface': 4.770699800290478e-06, 'learning_rate/action_model': 4.770699800290478e-05, 'epoch': 1.38}) +[2026-06-18 00:19:08,454][starVLA.training.train_starvla][INFO] - Step 3773, Loss: {'train/loss': 0.4134839177131653, 'train/grad_norm_pre_clip': 32.91536331176758, 'timing/data': 0.0009544482454657555, 'timing/model': 1.6068831272423267, 'learning_rate/qwen_vl_interface': 4.768548197747953e-06, 'learning_rate/action_model': 4.768548197747953e-05, 'epoch': 1.38}) +[2026-06-18 00:19:13,733][starVLA.training.train_starvla][INFO] - Step 3774, Loss: {'train/loss': 0.7770543694496155, 'train/grad_norm_pre_clip': 23.34818458557129, 'timing/data': 0.001306772232055664, 'timing/model': 1.4181476328521967, 'learning_rate/qwen_vl_interface': 4.766396695010903e-06, 'learning_rate/action_model': 4.766396695010903e-05, 'epoch': 1.38}) +[2026-06-18 00:19:19,528][starVLA.training.train_starvla][INFO] - Step 3775, Loss: {'train/loss': 0.3361362814903259, 'train/grad_norm_pre_clip': 18.06053924560547, 'timing/data': 0.0010441774502396584, 'timing/model': 1.1418724274262786, 'learning_rate/qwen_vl_interface': 4.764245292525337e-06, 'learning_rate/action_model': 4.7642452925253376e-05, 'epoch': 1.38}) +[2026-06-18 00:19:28,148][starVLA.training.train_starvla][INFO] - Step 3776, Loss: {'train/loss': 0.6026390194892883, 'train/grad_norm_pre_clip': 17.699085235595703, 'timing/data': 0.00522194430232048, 'timing/model': 2.819728946313262, 'learning_rate/qwen_vl_interface': 4.762093990737247e-06, 'learning_rate/action_model': 4.762093990737246e-05, 'epoch': 1.38}) +[2026-06-18 00:19:32,888][starVLA.training.train_starvla][INFO] - Step 3777, Loss: {'train/loss': 0.4671747088432312, 'train/grad_norm_pre_clip': 22.528743743896484, 'timing/data': 0.0013559646904468536, 'timing/model': 1.38957582693547, 'learning_rate/qwen_vl_interface': 4.759942790092596e-06, 'learning_rate/action_model': 4.759942790092596e-05, 'epoch': 1.38}) +[2026-06-18 00:19:37,242][starVLA.training.train_starvla][INFO] - Step 3778, Loss: {'train/loss': 0.3591090440750122, 'train/grad_norm_pre_clip': 27.673460006713867, 'timing/data': 0.0010906253010034561, 'timing/model': 1.1144580347463489, 'learning_rate/qwen_vl_interface': 4.757791691037332e-06, 'learning_rate/action_model': 4.757791691037332e-05, 'epoch': 1.38}) +[2026-06-18 00:19:41,769][starVLA.training.train_starvla][INFO] - Step 3779, Loss: {'train/loss': 0.6361258625984192, 'train/grad_norm_pre_clip': 27.43113899230957, 'timing/data': 0.0008938247337937355, 'timing/model': 1.07193299010396, 'learning_rate/qwen_vl_interface': 4.755640694017381e-06, 'learning_rate/action_model': 4.755640694017381e-05, 'epoch': 1.38}) +[2026-06-18 00:19:46,334][starVLA.training.train_starvla][INFO] - Step 3780, Loss: {'train/loss': 0.35291942954063416, 'train/grad_norm_pre_clip': 14.979564666748047, 'timing/data': 0.001489575020968914, 'timing/model': 1.0734614534303546, 'learning_rate/qwen_vl_interface': 4.753489799478647e-06, 'learning_rate/action_model': 4.7534897994786464e-05, 'epoch': 1.38}) +[2026-06-18 00:19:52,758][starVLA.training.train_starvla][INFO] - Step 3781, Loss: {'train/loss': 0.6976386308670044, 'train/grad_norm_pre_clip': 23.174480438232422, 'timing/data': 0.000947142019867897, 'timing/model': 1.1356773246079683, 'learning_rate/qwen_vl_interface': 4.7513390078670105e-06, 'learning_rate/action_model': 4.75133900786701e-05, 'epoch': 1.38}) +[2026-06-18 00:19:58,144][starVLA.training.train_starvla][INFO] - Step 3782, Loss: {'train/loss': 0.3358162045478821, 'train/grad_norm_pre_clip': 16.45068359375, 'timing/data': 0.00106123648583889, 'timing/model': 1.0836372962221503, 'learning_rate/qwen_vl_interface': 4.749188319628333e-06, 'learning_rate/action_model': 4.749188319628333e-05, 'epoch': 1.38}) +[2026-06-18 00:20:03,356][starVLA.training.train_starvla][INFO] - Step 3783, Loss: {'train/loss': 0.6022910475730896, 'train/grad_norm_pre_clip': 20.276824951171875, 'timing/data': 0.0012145964428782463, 'timing/model': 1.595810561440885, 'learning_rate/qwen_vl_interface': 4.747037735208457e-06, 'learning_rate/action_model': 4.7470377352084575e-05, 'epoch': 1.39}) +[2026-06-18 00:20:11,503][starVLA.training.train_starvla][INFO] - Step 3784, Loss: {'train/loss': 0.2344282567501068, 'train/grad_norm_pre_clip': 17.32622528076172, 'timing/data': 0.0009522410109639168, 'timing/model': 4.178515317849815, 'learning_rate/qwen_vl_interface': 4.7448872550532006e-06, 'learning_rate/action_model': 4.7448872550532e-05, 'epoch': 1.39}) +[2026-06-18 00:20:22,755][starVLA.training.train_starvla][INFO] - Step 3785, Loss: {'train/loss': 0.3170784115791321, 'train/grad_norm_pre_clip': 15.400433540344238, 'timing/data': 0.0008906163275241852, 'timing/model': 2.6772392485290766, 'learning_rate/qwen_vl_interface': 4.74273687960836e-06, 'learning_rate/action_model': 4.74273687960836e-05, 'epoch': 1.39}) +[2026-06-18 00:20:27,775][starVLA.training.train_starvla][INFO] - Step 3786, Loss: {'train/loss': 0.5270718336105347, 'train/grad_norm_pre_clip': 22.991907119750977, 'timing/data': 0.0009282007813453674, 'timing/model': 1.163965311832726, 'learning_rate/qwen_vl_interface': 4.7405866093197075e-06, 'learning_rate/action_model': 4.740586609319707e-05, 'epoch': 1.39}) +[2026-06-18 00:20:33,938][starVLA.training.train_starvla][INFO] - Step 3787, Loss: {'train/loss': 0.4749356508255005, 'train/grad_norm_pre_clip': 19.62961196899414, 'timing/data': 0.0007866481319069862, 'timing/model': 2.322222985327244, 'learning_rate/qwen_vl_interface': 4.738436444633002e-06, 'learning_rate/action_model': 4.738436444633001e-05, 'epoch': 1.39}) +[2026-06-18 00:20:43,746][starVLA.training.train_starvla][INFO] - Step 3788, Loss: {'train/loss': 0.4648588001728058, 'train/grad_norm_pre_clip': 15.922754287719727, 'timing/data': 0.001312514767050743, 'timing/model': 5.049310299567878, 'learning_rate/qwen_vl_interface': 4.73628638599397e-06, 'learning_rate/action_model': 4.73628638599397e-05, 'epoch': 1.39}) +[2026-06-18 00:20:48,128][starVLA.training.train_starvla][INFO] - Step 3789, Loss: {'train/loss': 0.4445035755634308, 'train/grad_norm_pre_clip': 18.474748611450195, 'timing/data': 0.0014145486056804657, 'timing/model': 1.2460657488554716, 'learning_rate/qwen_vl_interface': 4.734136433848323e-06, 'learning_rate/action_model': 4.7341364338483224e-05, 'epoch': 1.39}) +[2026-06-18 00:20:53,016][starVLA.training.train_starvla][INFO] - Step 3790, Loss: {'train/loss': 0.2954893112182617, 'train/grad_norm_pre_clip': 24.4222354888916, 'timing/data': 0.0007852921262383461, 'timing/model': 1.1756196161732078, 'learning_rate/qwen_vl_interface': 4.731986588641748e-06, 'learning_rate/action_model': 4.7319865886417484e-05, 'epoch': 1.39}) +[2026-06-18 00:20:57,766][starVLA.training.train_starvla][INFO] - Step 3791, Loss: {'train/loss': 0.6715716123580933, 'train/grad_norm_pre_clip': 20.99884605407715, 'timing/data': 0.0014268234372138977, 'timing/model': 1.1395347025245428, 'learning_rate/qwen_vl_interface': 4.729836850819912e-06, 'learning_rate/action_model': 4.729836850819912e-05, 'epoch': 1.39}) +[2026-06-18 00:21:02,710][starVLA.training.train_starvla][INFO] - Step 3792, Loss: {'train/loss': 0.4537188410758972, 'train/grad_norm_pre_clip': 17.037738800048828, 'timing/data': 0.000874573364853859, 'timing/model': 1.0529653439298272, 'learning_rate/qwen_vl_interface': 4.727687220828458e-06, 'learning_rate/action_model': 4.727687220828458e-05, 'epoch': 1.39}) +[2026-06-18 00:21:07,984][starVLA.training.train_starvla][INFO] - Step 3793, Loss: {'train/loss': 0.7976399064064026, 'train/grad_norm_pre_clip': 23.657257080078125, 'timing/data': 0.000763637013733387, 'timing/model': 1.3602206725627184, 'learning_rate/qwen_vl_interface': 4.7255376991130005e-06, 'learning_rate/action_model': 4.7255376991130005e-05, 'epoch': 1.39}) +[2026-06-18 00:21:15,506][starVLA.training.train_starvla][INFO] - Step 3794, Loss: {'train/loss': 0.4001341760158539, 'train/grad_norm_pre_clip': 15.719792366027832, 'timing/data': 0.04754194617271423, 'timing/model': 4.1006898963823915, 'learning_rate/qwen_vl_interface': 4.723388286119147e-06, 'learning_rate/action_model': 4.723388286119146e-05, 'epoch': 1.39}) +[2026-06-18 00:21:24,591][starVLA.training.train_starvla][INFO] - Step 3795, Loss: {'train/loss': 0.46802589297294617, 'train/grad_norm_pre_clip': 62.87474060058594, 'timing/data': 0.0020024199038743973, 'timing/model': 1.0275570172816515, 'learning_rate/qwen_vl_interface': 4.721238982292467e-06, 'learning_rate/action_model': 4.721238982292466e-05, 'epoch': 1.39}) +[2026-06-18 00:21:36,998][starVLA.training.train_starvla][INFO] - Step 3796, Loss: {'train/loss': 0.2691853940486908, 'train/grad_norm_pre_clip': 12.424102783203125, 'timing/data': 0.001232190988957882, 'timing/model': 1.0878937020897865, 'learning_rate/qwen_vl_interface': 4.719089788078514e-06, 'learning_rate/action_model': 4.719089788078514e-05, 'epoch': 1.39}) +[2026-06-18 00:21:41,604][starVLA.training.train_starvla][INFO] - Step 3797, Loss: {'train/loss': 0.6449469923973083, 'train/grad_norm_pre_clip': 21.988985061645508, 'timing/data': 0.001226847991347313, 'timing/model': 1.0744190216064453, 'learning_rate/qwen_vl_interface': 4.716940703922818e-06, 'learning_rate/action_model': 4.716940703922818e-05, 'epoch': 1.39}) +[2026-06-18 00:21:46,122][starVLA.training.train_starvla][INFO] - Step 3798, Loss: {'train/loss': 0.3682419955730438, 'train/grad_norm_pre_clip': 16.241291046142578, 'timing/data': 0.0010948823764920235, 'timing/model': 1.375449674203992, 'learning_rate/qwen_vl_interface': 4.714791730270889e-06, 'learning_rate/action_model': 4.71479173027089e-05, 'epoch': 1.39}) +[2026-06-18 00:21:50,538][starVLA.training.train_starvla][INFO] - Step 3799, Loss: {'train/loss': 0.5714238882064819, 'train/grad_norm_pre_clip': 16.466218948364258, 'timing/data': 0.0008691735565662384, 'timing/model': 1.219019210897386, 'learning_rate/qwen_vl_interface': 4.712642867568213e-06, 'learning_rate/action_model': 4.712642867568212e-05, 'epoch': 1.39}) +[2026-06-18 00:26:18,845][starVLA.training.train_starvla][INFO] - Step 3800, Loss: {'train/loss': 0.6362658739089966, 'train/grad_norm_pre_clip': 16.69879722595215, 'eval/loss': 0.6105672717094421, 'eval/latency_2/loss': 0.6105672717094421, 'eval/demon_fire_precision': 0.9369712351945855, 'eval/demon_fire_recall': 0.9267782426778243, 'eval/demon_fire_f1': 0.9318468657972234, 'eval/demon_fire_rate_model': 0.657762938230384, 'eval/demon_fire_rate_teacher': 0.6649972175848636, 'eval/demon_left_precision': 0.9356223175965666, 'eval/demon_left_recall': 0.6884210526315789, 'eval/demon_left_f1': 0.7932080048514252, 'eval/demon_left_rate_model': 0.19449081803005008, 'eval/demon_left_rate_teacher': 0.2643294379521425, 'eval/demon_right_precision': 0.8288659793814434, 'eval/demon_right_recall': 0.9084745762711864, 'eval/demon_right_f1': 0.8668463611859839, 'eval/demon_right_rate_model': 0.26989426822481916, 'eval/demon_right_rate_teacher': 0.24624373956594323, 'eval/demon_move_f1': 0.8300271830187045, 'eval/demon_cc_f1': 0.8987711095553833, 'eval/latency_2/demon_fire_precision': 0.9369712351945855, 'eval/latency_2/demon_fire_recall': 0.9267782426778243, 'eval/latency_2/demon_fire_f1': 0.9318468657972234, 'eval/latency_2/demon_fire_rate_model': 0.657762938230384, 'eval/latency_2/demon_fire_rate_teacher': 0.6649972175848636, 'eval/latency_2/demon_left_precision': 0.9356223175965666, 'eval/latency_2/demon_left_recall': 0.6884210526315789, 'eval/latency_2/demon_left_f1': 0.7932080048514252, 'eval/latency_2/demon_left_rate_model': 0.19449081803005008, 'eval/latency_2/demon_left_rate_teacher': 0.2643294379521425, 'eval/latency_2/demon_right_precision': 0.8288659793814434, 'eval/latency_2/demon_right_recall': 0.9084745762711864, 'eval/latency_2/demon_right_f1': 0.8668463611859839, 'eval/latency_2/demon_right_rate_model': 0.26989426822481916, 'eval/latency_2/demon_right_rate_teacher': 0.24624373956594323, 'eval/latency_2/demon_move_f1': 0.8300271830187045, 'eval/latency_2/demon_cc_f1': 0.8987711095553833, 'timing/data': 0.0011489838361740112, 'timing/model': 4.759026821702719, 'learning_rate/qwen_vl_interface': 4.7104941162602436e-06, 'learning_rate/action_model': 4.710494116260244e-05, 'epoch': 1.39}) +[2026-06-18 00:26:18,847][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_3800_state +[2026-06-18 00:26:18,854][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 00:26:27,875][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_3800_state/model.safetensors +[2026-06-18 00:26:43,373][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_3800_state/optimizer.bin +[2026-06-18 00:26:43,375][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_3800_state/sampler.bin +[2026-06-18 00:26:43,376][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_3800_state/sampler_1.bin +[2026-06-18 00:26:43,391][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_3800_state/random_states_0.pkl +[2026-06-18 00:26:43,396][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3800 +[2026-06-18 00:26:43,397][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 00:26:43,421][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 00:26:48,728][starVLA.training.train_starvla][INFO] - Step 3801, Loss: {'train/loss': 0.6867198944091797, 'train/grad_norm_pre_clip': 24.883235931396484, 'timing/data': 0.0011777477338910103, 'timing/model': 1.171201371587813, 'learning_rate/qwen_vl_interface': 4.708345476792428e-06, 'learning_rate/action_model': 4.7083454767924276e-05, 'epoch': 1.39}) +[2026-06-18 00:26:59,418][starVLA.training.train_starvla][INFO] - Step 3802, Loss: {'train/loss': 0.4525177478790283, 'train/grad_norm_pre_clip': 22.86612319946289, 'timing/data': 0.0010698838159441948, 'timing/model': 4.431119477376342, 'learning_rate/qwen_vl_interface': 4.706196949610178e-06, 'learning_rate/action_model': 4.7061969496101776e-05, 'epoch': 1.39}) +[2026-06-18 00:27:07,665][starVLA.training.train_starvla][INFO] - Step 3803, Loss: {'train/loss': 0.4189107418060303, 'train/grad_norm_pre_clip': 20.516727447509766, 'timing/data': 0.000948004424571991, 'timing/model': 1.281962781213224, 'learning_rate/qwen_vl_interface': 4.7040485351588845e-06, 'learning_rate/action_model': 4.7040485351588845e-05, 'epoch': 1.39}) +[2026-06-18 00:27:12,075][starVLA.training.train_starvla][INFO] - Step 3804, Loss: {'train/loss': 0.6708604097366333, 'train/grad_norm_pre_clip': 22.281286239624023, 'timing/data': 0.0010814685374498367, 'timing/model': 1.35103396512568, 'learning_rate/qwen_vl_interface': 4.701900233883916e-06, 'learning_rate/action_model': 4.7019002338839156e-05, 'epoch': 1.39}) +[2026-06-18 00:27:20,676][starVLA.training.train_starvla][INFO] - Step 3805, Loss: {'train/loss': 0.49495065212249756, 'train/grad_norm_pre_clip': 17.586101531982422, 'timing/data': 0.0010741455480456352, 'timing/model': 1.0623847087845206, 'learning_rate/qwen_vl_interface': 4.69975204623062e-06, 'learning_rate/action_model': 4.69975204623062e-05, 'epoch': 1.39}) +[2026-06-18 00:27:27,471][starVLA.training.train_starvla][INFO] - Step 3806, Loss: {'train/loss': 0.38095274567604065, 'train/grad_norm_pre_clip': 16.31493377685547, 'timing/data': 0.000898885540664196, 'timing/model': 1.1357819875702262, 'learning_rate/qwen_vl_interface': 4.697603972644319e-06, 'learning_rate/action_model': 4.697603972644318e-05, 'epoch': 1.39}) +[2026-06-18 00:27:32,247][starVLA.training.train_starvla][INFO] - Step 3807, Loss: {'train/loss': 0.34414488077163696, 'train/grad_norm_pre_clip': 18.241350173950195, 'timing/data': 0.0009189620614051819, 'timing/model': 1.106894033960998, 'learning_rate/qwen_vl_interface': 4.695456013570306e-06, 'learning_rate/action_model': 4.695456013570306e-05, 'epoch': 1.39}) +[2026-06-18 00:27:36,690][starVLA.training.train_starvla][INFO] - Step 3808, Loss: {'train/loss': 0.36179429292678833, 'train/grad_norm_pre_clip': 14.751008987426758, 'timing/data': 0.0010554129257798195, 'timing/model': 1.0297260079532862, 'learning_rate/qwen_vl_interface': 4.693308169453858e-06, 'learning_rate/action_model': 4.693308169453858e-05, 'epoch': 1.39}) +[2026-06-18 00:27:45,091][starVLA.training.train_starvla][INFO] - Step 3809, Loss: {'train/loss': 0.18328185379505157, 'train/grad_norm_pre_clip': 10.755888938903809, 'timing/data': 0.001114601269364357, 'timing/model': 1.4485796736553311, 'learning_rate/qwen_vl_interface': 4.691160440740226e-06, 'learning_rate/action_model': 4.691160440740226e-05, 'epoch': 1.39}) +[2026-06-18 00:27:54,304][starVLA.training.train_starvla][INFO] - Step 3810, Loss: {'train/loss': 0.2127198576927185, 'train/grad_norm_pre_clip': 11.754226684570312, 'timing/data': 0.0007827524095773697, 'timing/model': 5.286402233876288, 'learning_rate/qwen_vl_interface': 4.689012827874637e-06, 'learning_rate/action_model': 4.689012827874637e-05, 'epoch': 1.4}) +[2026-06-18 00:27:59,325][starVLA.training.train_starvla][INFO] - Step 3811, Loss: {'train/loss': 0.46547731757164, 'train/grad_norm_pre_clip': 23.39724349975586, 'timing/data': 0.0009413817897439003, 'timing/model': 1.2315998328849673, 'learning_rate/qwen_vl_interface': 4.6868653313022905e-06, 'learning_rate/action_model': 4.68686533130229e-05, 'epoch': 1.4}) +[2026-06-18 00:28:04,014][starVLA.training.train_starvla][INFO] - Step 3812, Loss: {'train/loss': 0.4386831820011139, 'train/grad_norm_pre_clip': 20.217933654785156, 'timing/data': 0.0011241612955927849, 'timing/model': 1.0697676129639149, 'learning_rate/qwen_vl_interface': 4.684717951468368e-06, 'learning_rate/action_model': 4.6847179514683684e-05, 'epoch': 1.4}) +[2026-06-18 00:28:08,669][starVLA.training.train_starvla][INFO] - Step 3813, Loss: {'train/loss': 0.9241885542869568, 'train/grad_norm_pre_clip': 26.82964515686035, 'timing/data': 0.0008824905380606651, 'timing/model': 1.521741439588368, 'learning_rate/qwen_vl_interface': 4.682570688818024e-06, 'learning_rate/action_model': 4.682570688818024e-05, 'epoch': 1.4}) +[2026-06-18 00:28:21,598][starVLA.training.train_starvla][INFO] - Step 3814, Loss: {'train/loss': 0.35214290022850037, 'train/grad_norm_pre_clip': 13.399941444396973, 'timing/data': 0.004756351932883263, 'timing/model': 1.3920907657593489, 'learning_rate/qwen_vl_interface': 4.680423543796385e-06, 'learning_rate/action_model': 4.680423543796385e-05, 'epoch': 1.4}) +[2026-06-18 00:28:26,982][starVLA.training.train_starvla][INFO] - Step 3815, Loss: {'train/loss': 0.6052278280258179, 'train/grad_norm_pre_clip': 22.051876068115234, 'timing/data': 0.0010933391749858856, 'timing/model': 1.123897423967719, 'learning_rate/qwen_vl_interface': 4.678276516848558e-06, 'learning_rate/action_model': 4.678276516848557e-05, 'epoch': 1.4}) +[2026-06-18 00:28:31,480][starVLA.training.train_starvla][INFO] - Step 3816, Loss: {'train/loss': 0.22117692232131958, 'train/grad_norm_pre_clip': 14.934327125549316, 'timing/data': 0.0011739563196897507, 'timing/model': 1.0698079224675894, 'learning_rate/qwen_vl_interface': 4.676129608419625e-06, 'learning_rate/action_model': 4.676129608419625e-05, 'epoch': 1.4}) +[2026-06-18 00:28:36,657][starVLA.training.train_starvla][INFO] - Step 3817, Loss: {'train/loss': 0.3639806807041168, 'train/grad_norm_pre_clip': 18.066917419433594, 'timing/data': 0.003855222836136818, 'timing/model': 1.639265239238739, 'learning_rate/qwen_vl_interface': 4.6739828189546415e-06, 'learning_rate/action_model': 4.6739828189546417e-05, 'epoch': 1.4}) +[2026-06-18 00:28:41,263][starVLA.training.train_starvla][INFO] - Step 3818, Loss: {'train/loss': 0.3521789014339447, 'train/grad_norm_pre_clip': 17.77794647216797, 'timing/data': 0.0011507123708724976, 'timing/model': 1.1814415603876114, 'learning_rate/qwen_vl_interface': 4.671836148898641e-06, 'learning_rate/action_model': 4.6718361488986404e-05, 'epoch': 1.4}) +[2026-06-18 00:28:51,130][starVLA.training.train_starvla][INFO] - Step 3819, Loss: {'train/loss': 0.2183106690645218, 'train/grad_norm_pre_clip': 10.506129264831543, 'timing/data': 0.001382150687277317, 'timing/model': 3.4712532190606, 'learning_rate/qwen_vl_interface': 4.6696895986966254e-06, 'learning_rate/action_model': 4.669689598696625e-05, 'epoch': 1.4}) +[2026-06-18 00:29:00,309][starVLA.training.train_starvla][INFO] - Step 3820, Loss: {'train/loss': 0.2054024636745453, 'train/grad_norm_pre_clip': 10.124032974243164, 'timing/data': 0.000784287229180336, 'timing/model': 4.754215179011226, 'learning_rate/qwen_vl_interface': 4.667543168793583e-06, 'learning_rate/action_model': 4.667543168793583e-05, 'epoch': 1.4}) +[2026-06-18 00:29:04,896][starVLA.training.train_starvla][INFO] - Step 3821, Loss: {'train/loss': 0.47705623507499695, 'train/grad_norm_pre_clip': 23.126367568969727, 'timing/data': 0.0015486516058444977, 'timing/model': 1.041604932397604, 'learning_rate/qwen_vl_interface': 4.665396859634468e-06, 'learning_rate/action_model': 4.665396859634468e-05, 'epoch': 1.4}) +[2026-06-18 00:29:09,912][starVLA.training.train_starvla][INFO] - Step 3822, Loss: {'train/loss': 0.431535005569458, 'train/grad_norm_pre_clip': 17.57317543029785, 'timing/data': 0.0010684700682759285, 'timing/model': 1.4184040063992143, 'learning_rate/qwen_vl_interface': 4.663250671664212e-06, 'learning_rate/action_model': 4.663250671664211e-05, 'epoch': 1.4}) +[2026-06-18 00:29:18,875][starVLA.training.train_starvla][INFO] - Step 3823, Loss: {'train/loss': 0.6462773680686951, 'train/grad_norm_pre_clip': 31.888214111328125, 'timing/data': 0.0008009970188140869, 'timing/model': 1.8360870042815804, 'learning_rate/qwen_vl_interface': 4.6611046053277245e-06, 'learning_rate/action_model': 4.661104605327724e-05, 'epoch': 1.4}) +[2026-06-18 00:29:23,709][starVLA.training.train_starvla][INFO] - Step 3824, Loss: {'train/loss': 0.44687217473983765, 'train/grad_norm_pre_clip': 19.312580108642578, 'timing/data': 0.0008447263389825821, 'timing/model': 1.592555796727538, 'learning_rate/qwen_vl_interface': 4.658958661069886e-06, 'learning_rate/action_model': 4.6589586610698856e-05, 'epoch': 1.4}) +[2026-06-18 00:29:35,555][starVLA.training.train_starvla][INFO] - Step 3825, Loss: {'train/loss': 0.18150575459003448, 'train/grad_norm_pre_clip': 16.26066780090332, 'timing/data': 0.0009433561936020851, 'timing/model': 4.819425670430064, 'learning_rate/qwen_vl_interface': 4.6568128393355535e-06, 'learning_rate/action_model': 4.656812839335553e-05, 'epoch': 1.4}) +[2026-06-18 00:29:45,648][starVLA.training.train_starvla][INFO] - Step 3826, Loss: {'train/loss': 0.4762234687805176, 'train/grad_norm_pre_clip': 20.570749282836914, 'timing/data': 0.0009270273149013519, 'timing/model': 1.0915471902117133, 'learning_rate/qwen_vl_interface': 4.654667140569555e-06, 'learning_rate/action_model': 4.6546671405695544e-05, 'epoch': 1.4}) +[2026-06-18 00:29:58,207][starVLA.training.train_starvla][INFO] - Step 3827, Loss: {'train/loss': 0.39775094389915466, 'train/grad_norm_pre_clip': 15.526021003723145, 'timing/data': 0.0009778812527656555, 'timing/model': 1.4715988850221038, 'learning_rate/qwen_vl_interface': 4.652521565216701e-06, 'learning_rate/action_model': 4.652521565216701e-05, 'epoch': 1.4}) +[2026-06-18 00:30:06,345][starVLA.training.train_starvla][INFO] - Step 3828, Loss: {'train/loss': 0.5805047750473022, 'train/grad_norm_pre_clip': 24.59239959716797, 'timing/data': 0.0010626781731843948, 'timing/model': 3.980078519321978, 'learning_rate/qwen_vl_interface': 4.65037611372177e-06, 'learning_rate/action_model': 4.650376113721769e-05, 'epoch': 1.4}) +[2026-06-18 00:30:13,408][starVLA.training.train_starvla][INFO] - Step 3829, Loss: {'train/loss': 0.4380834698677063, 'train/grad_norm_pre_clip': 20.763959884643555, 'timing/data': 0.0007800403982400894, 'timing/model': 1.785083582624793, 'learning_rate/qwen_vl_interface': 4.648230786529514e-06, 'learning_rate/action_model': 4.648230786529513e-05, 'epoch': 1.4}) +[2026-06-18 00:30:21,632][starVLA.training.train_starvla][INFO] - Step 3830, Loss: {'train/loss': 0.642894983291626, 'train/grad_norm_pre_clip': 25.551931381225586, 'timing/data': 0.001310541294515133, 'timing/model': 1.1126839499920607, 'learning_rate/qwen_vl_interface': 4.646085584084664e-06, 'learning_rate/action_model': 4.6460855840846637e-05, 'epoch': 1.4}) +[2026-06-18 00:30:33,310][starVLA.training.train_starvla][INFO] - Step 3831, Loss: {'train/loss': 0.43137839436531067, 'train/grad_norm_pre_clip': 16.52754783630371, 'timing/data': 0.0009228633716702461, 'timing/model': 1.1870328253135085, 'learning_rate/qwen_vl_interface': 4.643940506831923e-06, 'learning_rate/action_model': 4.6439405068319226e-05, 'epoch': 1.4}) +[2026-06-18 00:30:44,572][starVLA.training.train_starvla][INFO] - Step 3832, Loss: {'train/loss': 0.5092623829841614, 'train/grad_norm_pre_clip': 18.046430587768555, 'timing/data': 0.0011023906990885735, 'timing/model': 1.1255075354129076, 'learning_rate/qwen_vl_interface': 4.641795555215967e-06, 'learning_rate/action_model': 4.6417955552159676e-05, 'epoch': 1.4}) +[2026-06-18 00:30:56,669][starVLA.training.train_starvla][INFO] - Step 3833, Loss: {'train/loss': 0.6841080784797668, 'train/grad_norm_pre_clip': 19.3149471282959, 'timing/data': 0.003474782221019268, 'timing/model': 1.150523690506816, 'learning_rate/qwen_vl_interface': 4.639650729681445e-06, 'learning_rate/action_model': 4.6396507296814454e-05, 'epoch': 1.4}) +[2026-06-18 00:31:11,027][starVLA.training.train_starvla][INFO] - Step 3834, Loss: {'train/loss': 0.6103862524032593, 'train/grad_norm_pre_clip': 17.57131576538086, 'timing/data': 0.0009845271706581116, 'timing/model': 6.2837272780016065, 'learning_rate/qwen_vl_interface': 4.637506030672987e-06, 'learning_rate/action_model': 4.637506030672987e-05, 'epoch': 1.4}) +[2026-06-18 00:31:15,417][starVLA.training.train_starvla][INFO] - Step 3835, Loss: {'train/loss': 0.34253332018852234, 'train/grad_norm_pre_clip': 16.520551681518555, 'timing/data': 0.0010459423065185547, 'timing/model': 1.0544355725869536, 'learning_rate/qwen_vl_interface': 4.635361458635186e-06, 'learning_rate/action_model': 4.6353614586351864e-05, 'epoch': 1.4}) +[2026-06-18 00:31:19,785][starVLA.training.train_starvla][INFO] - Step 3836, Loss: {'train/loss': 0.7006060481071472, 'train/grad_norm_pre_clip': 26.302461624145508, 'timing/data': 0.0010051913559436798, 'timing/model': 1.0516417417675257, 'learning_rate/qwen_vl_interface': 4.633217014012616e-06, 'learning_rate/action_model': 4.633217014012616e-05, 'epoch': 1.4}) +[2026-06-18 00:31:24,581][starVLA.training.train_starvla][INFO] - Step 3837, Loss: {'train/loss': 0.28479206562042236, 'train/grad_norm_pre_clip': 12.768898010253906, 'timing/data': 0.001064719632267952, 'timing/model': 1.024633607827127, 'learning_rate/qwen_vl_interface': 4.631072697249822e-06, 'learning_rate/action_model': 4.631072697249822e-05, 'epoch': 1.41}) +[2026-06-18 00:31:29,238][starVLA.training.train_starvla][INFO] - Step 3838, Loss: {'train/loss': 0.2679888904094696, 'train/grad_norm_pre_clip': 12.339341163635254, 'timing/data': 0.0008791685104370117, 'timing/model': 1.1607963917776942, 'learning_rate/qwen_vl_interface': 4.628928508791325e-06, 'learning_rate/action_model': 4.628928508791324e-05, 'epoch': 1.41}) +[2026-06-18 00:31:34,069][starVLA.training.train_starvla][INFO] - Step 3839, Loss: {'train/loss': 0.3697115182876587, 'train/grad_norm_pre_clip': 23.10544204711914, 'timing/data': 0.0010948944836854935, 'timing/model': 1.380821148864925, 'learning_rate/qwen_vl_interface': 4.626784449081617e-06, 'learning_rate/action_model': 4.626784449081617e-05, 'epoch': 1.41}) +[2026-06-18 00:31:38,191][starVLA.training.train_starvla][INFO] - Step 3840, Loss: {'train/loss': 0.3627625107765198, 'train/grad_norm_pre_clip': 16.062816619873047, 'timing/data': 0.0013250866904854774, 'timing/model': 1.0401779422536492, 'learning_rate/qwen_vl_interface': 4.62464051856516e-06, 'learning_rate/action_model': 4.6246405185651596e-05, 'epoch': 1.41}) +[2026-06-18 00:31:42,596][starVLA.training.train_starvla][INFO] - Step 3841, Loss: {'train/loss': 0.4508897066116333, 'train/grad_norm_pre_clip': 21.452360153198242, 'timing/data': 0.006604054011404514, 'timing/model': 1.0788145782426, 'learning_rate/qwen_vl_interface': 4.622496717686401e-06, 'learning_rate/action_model': 4.6224967176864005e-05, 'epoch': 1.41}) +[2026-06-18 00:31:47,038][starVLA.training.train_starvla][INFO] - Step 3842, Loss: {'train/loss': 0.5553359985351562, 'train/grad_norm_pre_clip': 15.842595100402832, 'timing/data': 0.0007487768307328224, 'timing/model': 1.1048234151676297, 'learning_rate/qwen_vl_interface': 4.620353046889745e-06, 'learning_rate/action_model': 4.620353046889745e-05, 'epoch': 1.41}) +[2026-06-18 00:31:52,302][starVLA.training.train_starvla][INFO] - Step 3843, Loss: {'train/loss': 0.508256733417511, 'train/grad_norm_pre_clip': 21.4958438873291, 'timing/data': 0.001051102764904499, 'timing/model': 1.702479712665081, 'learning_rate/qwen_vl_interface': 4.6182095066195805e-06, 'learning_rate/action_model': 4.61820950661958e-05, 'epoch': 1.41}) +[2026-06-18 00:31:56,697][starVLA.training.train_starvla][INFO] - Step 3844, Loss: {'train/loss': 0.3683621287345886, 'train/grad_norm_pre_clip': 17.421579360961914, 'timing/data': 0.0024855099618434906, 'timing/model': 1.3871250627562404, 'learning_rate/qwen_vl_interface': 4.616066097320264e-06, 'learning_rate/action_model': 4.6160660973202636e-05, 'epoch': 1.41}) +[2026-06-18 00:32:00,907][starVLA.training.train_starvla][INFO] - Step 3845, Loss: {'train/loss': 0.3193341791629791, 'train/grad_norm_pre_clip': 18.307641983032227, 'timing/data': 0.0007891058921813965, 'timing/model': 1.1694728648290038, 'learning_rate/qwen_vl_interface': 4.613922819436129e-06, 'learning_rate/action_model': 4.613922819436128e-05, 'epoch': 1.41}) +[2026-06-18 00:32:05,616][starVLA.training.train_starvla][INFO] - Step 3846, Loss: {'train/loss': 0.30119723081588745, 'train/grad_norm_pre_clip': 16.367643356323242, 'timing/data': 0.0008115442469716072, 'timing/model': 1.051918943412602, 'learning_rate/qwen_vl_interface': 4.611779673411478e-06, 'learning_rate/action_model': 4.6117796734114785e-05, 'epoch': 1.41}) +[2026-06-18 00:32:10,568][starVLA.training.train_starvla][INFO] - Step 3847, Loss: {'train/loss': 0.6187694668769836, 'train/grad_norm_pre_clip': 21.166847229003906, 'timing/data': 0.0011221850290894508, 'timing/model': 1.1784654865041375, 'learning_rate/qwen_vl_interface': 4.609636659690588e-06, 'learning_rate/action_model': 4.609636659690587e-05, 'epoch': 1.41}) +[2026-06-18 00:32:15,179][starVLA.training.train_starvla][INFO] - Step 3848, Loss: {'train/loss': 0.2250535488128662, 'train/grad_norm_pre_clip': 16.856515884399414, 'timing/data': 0.0010427245870232582, 'timing/model': 1.0905473167076707, 'learning_rate/qwen_vl_interface': 4.607493778717705e-06, 'learning_rate/action_model': 4.6074937787177055e-05, 'epoch': 1.41}) +[2026-06-18 00:32:19,679][starVLA.training.train_starvla][INFO] - Step 3849, Loss: {'train/loss': 0.39477676153182983, 'train/grad_norm_pre_clip': 16.733625411987305, 'timing/data': 0.0008474849164485931, 'timing/model': 1.458553984761238, 'learning_rate/qwen_vl_interface': 4.605351030937054e-06, 'learning_rate/action_model': 4.605351030937055e-05, 'epoch': 1.41}) +[2026-06-18 00:32:28,518][starVLA.training.train_starvla][INFO] - Step 3850, Loss: {'train/loss': 0.8311793804168701, 'train/grad_norm_pre_clip': 27.66475486755371, 'timing/data': 0.0008698226884007454, 'timing/model': 5.591191504150629, 'learning_rate/qwen_vl_interface': 4.60320841679283e-06, 'learning_rate/action_model': 4.60320841679283e-05, 'epoch': 1.41}) +[2026-06-18 00:32:48,268][starVLA.training.train_starvla][INFO] - Step 3851, Loss: {'train/loss': 0.4494670331478119, 'train/grad_norm_pre_clip': 19.78339958190918, 'timing/data': 0.001333429478108883, 'timing/model': 4.633250151760876, 'learning_rate/qwen_vl_interface': 4.6010659367291945e-06, 'learning_rate/action_model': 4.601065936729194e-05, 'epoch': 1.41}) +[2026-06-18 00:32:57,622][starVLA.training.train_starvla][INFO] - Step 3852, Loss: {'train/loss': 0.773293137550354, 'train/grad_norm_pre_clip': 32.1522102355957, 'timing/data': 0.0010867919772863388, 'timing/model': 1.2513029929250479, 'learning_rate/qwen_vl_interface': 4.598923591190289e-06, 'learning_rate/action_model': 4.598923591190289e-05, 'epoch': 1.41}) +[2026-06-18 00:33:12,720][starVLA.training.train_starvla][INFO] - Step 3853, Loss: {'train/loss': 0.5173363089561462, 'train/grad_norm_pre_clip': 17.3983154296875, 'timing/data': 0.0010867659002542496, 'timing/model': 4.6927041271701455, 'learning_rate/qwen_vl_interface': 4.596781380620225e-06, 'learning_rate/action_model': 4.5967813806202245e-05, 'epoch': 1.41}) +[2026-06-18 00:33:26,707][starVLA.training.train_starvla][INFO] - Step 3854, Loss: {'train/loss': 0.3761024475097656, 'train/grad_norm_pre_clip': 21.956546783447266, 'timing/data': 0.0008828341960906982, 'timing/model': 1.7445062538608909, 'learning_rate/qwen_vl_interface': 4.5946393054630845e-06, 'learning_rate/action_model': 4.5946393054630845e-05, 'epoch': 1.41}) +[2026-06-18 00:33:32,083][starVLA.training.train_starvla][INFO] - Step 3855, Loss: {'train/loss': 0.28816846013069153, 'train/grad_norm_pre_clip': 18.438417434692383, 'timing/data': 0.0008985614404082298, 'timing/model': 1.9270530184730887, 'learning_rate/qwen_vl_interface': 4.592497366162919e-06, 'learning_rate/action_model': 4.592497366162918e-05, 'epoch': 1.41}) +[2026-06-18 00:33:46,040][starVLA.training.train_starvla][INFO] - Step 3856, Loss: {'train/loss': 0.4617379605770111, 'train/grad_norm_pre_clip': 24.952247619628906, 'timing/data': 0.001831062138080597, 'timing/model': 1.1959682879969478, 'learning_rate/qwen_vl_interface': 4.590355563163758e-06, 'learning_rate/action_model': 4.5903555631637574e-05, 'epoch': 1.41}) +[2026-06-18 00:33:52,013][starVLA.training.train_starvla][INFO] - Step 3857, Loss: {'train/loss': 0.4278396964073181, 'train/grad_norm_pre_clip': 17.303743362426758, 'timing/data': 0.0009074322879314423, 'timing/model': 1.2882833620533347, 'learning_rate/qwen_vl_interface': 4.588213896909598e-06, 'learning_rate/action_model': 4.588213896909598e-05, 'epoch': 1.41}) +[2026-06-18 00:34:04,080][starVLA.training.train_starvla][INFO] - Step 3858, Loss: {'train/loss': 0.3772786557674408, 'train/grad_norm_pre_clip': 15.988213539123535, 'timing/data': 0.0013262052088975906, 'timing/model': 5.195849166251719, 'learning_rate/qwen_vl_interface': 4.58607236784441e-06, 'learning_rate/action_model': 4.58607236784441e-05, 'epoch': 1.41}) +[2026-06-18 00:34:12,949][starVLA.training.train_starvla][INFO] - Step 3859, Loss: {'train/loss': 0.48563703894615173, 'train/grad_norm_pre_clip': 17.002851486206055, 'timing/data': 0.0012934505939483643, 'timing/model': 1.6140906047075987, 'learning_rate/qwen_vl_interface': 4.583930976412133e-06, 'learning_rate/action_model': 4.583930976412133e-05, 'epoch': 1.41}) +[2026-06-18 00:34:22,499][starVLA.training.train_starvla][INFO] - Step 3860, Loss: {'train/loss': 0.26360026001930237, 'train/grad_norm_pre_clip': 13.136486053466797, 'timing/data': 0.0013205716386437416, 'timing/model': 1.1270874282345176, 'learning_rate/qwen_vl_interface': 4.5817897230566834e-06, 'learning_rate/action_model': 4.581789723056683e-05, 'epoch': 1.41}) +[2026-06-18 00:34:32,975][starVLA.training.train_starvla][INFO] - Step 3861, Loss: {'train/loss': 0.4504593014717102, 'train/grad_norm_pre_clip': 17.22602081298828, 'timing/data': 0.0016188202425837517, 'timing/model': 1.2119969138875604, 'learning_rate/qwen_vl_interface': 4.579648608221943e-06, 'learning_rate/action_model': 4.579648608221943e-05, 'epoch': 1.41}) +[2026-06-18 00:34:37,938][starVLA.training.train_starvla][INFO] - Step 3862, Loss: {'train/loss': 0.48538851737976074, 'train/grad_norm_pre_clip': 18.29435157775879, 'timing/data': 0.0011133123189210892, 'timing/model': 1.532125635072589, 'learning_rate/qwen_vl_interface': 4.577507632351766e-06, 'learning_rate/action_model': 4.5775076323517664e-05, 'epoch': 1.41}) +[2026-06-18 00:34:42,531][starVLA.training.train_starvla][INFO] - Step 3863, Loss: {'train/loss': 0.32908132672309875, 'train/grad_norm_pre_clip': 14.605332374572754, 'timing/data': 0.0013782558962702751, 'timing/model': 1.4104203041642904, 'learning_rate/qwen_vl_interface': 4.575366795889981e-06, 'learning_rate/action_model': 4.575366795889981e-05, 'epoch': 1.41}) +[2026-06-18 00:34:47,483][starVLA.training.train_starvla][INFO] - Step 3864, Loss: {'train/loss': 0.5869465470314026, 'train/grad_norm_pre_clip': 18.78203773498535, 'timing/data': 0.001091853715479374, 'timing/model': 1.4167358679696918, 'learning_rate/qwen_vl_interface': 4.573226099280386e-06, 'learning_rate/action_model': 4.573226099280386e-05, 'epoch': 1.41}) +[2026-06-18 00:34:51,918][starVLA.training.train_starvla][INFO] - Step 3865, Loss: {'train/loss': 0.3026393949985504, 'train/grad_norm_pre_clip': 14.211047172546387, 'timing/data': 0.0009301844984292984, 'timing/model': 1.4123022118583322, 'learning_rate/qwen_vl_interface': 4.571085542966748e-06, 'learning_rate/action_model': 4.5710855429667485e-05, 'epoch': 1.42}) +[2026-06-18 00:34:56,912][starVLA.training.train_starvla][INFO] - Step 3866, Loss: {'train/loss': 0.5300369262695312, 'train/grad_norm_pre_clip': 23.753128051757812, 'timing/data': 0.0008988454937934875, 'timing/model': 1.0568226464092731, 'learning_rate/qwen_vl_interface': 4.568945127392807e-06, 'learning_rate/action_model': 4.5689451273928066e-05, 'epoch': 1.42}) +[2026-06-18 00:35:01,998][starVLA.training.train_starvla][INFO] - Step 3867, Loss: {'train/loss': 0.5979675650596619, 'train/grad_norm_pre_clip': 22.107402801513672, 'timing/data': 0.0013679182156920433, 'timing/model': 1.089627548120916, 'learning_rate/qwen_vl_interface': 4.566804853002275e-06, 'learning_rate/action_model': 4.566804853002275e-05, 'epoch': 1.42}) +[2026-06-18 00:35:06,394][starVLA.training.train_starvla][INFO] - Step 3868, Loss: {'train/loss': 0.34976673126220703, 'train/grad_norm_pre_clip': 22.031909942626953, 'timing/data': 0.00105210579931736, 'timing/model': 1.0379233956336975, 'learning_rate/qwen_vl_interface': 4.564664720238833e-06, 'learning_rate/action_model': 4.564664720238833e-05, 'epoch': 1.42}) +[2026-06-18 00:35:20,018][starVLA.training.train_starvla][INFO] - Step 3869, Loss: {'train/loss': 0.38979649543762207, 'train/grad_norm_pre_clip': 21.224151611328125, 'timing/data': 0.0010496051982045174, 'timing/model': 1.9390814686194062, 'learning_rate/qwen_vl_interface': 4.5625247295461314e-06, 'learning_rate/action_model': 4.562524729546131e-05, 'epoch': 1.42}) +[2026-06-18 00:35:24,913][starVLA.training.train_starvla][INFO] - Step 3870, Loss: {'train/loss': 0.23622070252895355, 'train/grad_norm_pre_clip': 13.890535354614258, 'timing/data': 0.0015991898253560066, 'timing/model': 1.507964065298438, 'learning_rate/qwen_vl_interface': 4.560384881367791e-06, 'learning_rate/action_model': 4.5603848813677904e-05, 'epoch': 1.42}) +[2026-06-18 00:35:31,919][starVLA.training.train_starvla][INFO] - Step 3871, Loss: {'train/loss': 0.36644282937049866, 'train/grad_norm_pre_clip': 14.424585342407227, 'timing/data': 0.0011388882994651794, 'timing/model': 1.084800910204649, 'learning_rate/qwen_vl_interface': 4.5582451761474085e-06, 'learning_rate/action_model': 4.558245176147408e-05, 'epoch': 1.42}) +[2026-06-18 00:35:38,673][starVLA.training.train_starvla][INFO] - Step 3872, Loss: {'train/loss': 0.41892436146736145, 'train/grad_norm_pre_clip': 15.268197059631348, 'timing/data': 0.009463390335440636, 'timing/model': 3.15527709107846, 'learning_rate/qwen_vl_interface': 4.556105614328545e-06, 'learning_rate/action_model': 4.5561056143285455e-05, 'epoch': 1.42}) +[2026-06-18 00:35:43,349][starVLA.training.train_starvla][INFO] - Step 3873, Loss: {'train/loss': 0.2884686291217804, 'train/grad_norm_pre_clip': 13.34084415435791, 'timing/data': 0.0010089864954352379, 'timing/model': 1.02354587148875, 'learning_rate/qwen_vl_interface': 4.553966196354734e-06, 'learning_rate/action_model': 4.553966196354734e-05, 'epoch': 1.42}) +[2026-06-18 00:35:50,313][starVLA.training.train_starvla][INFO] - Step 3874, Loss: {'train/loss': 0.26952627301216125, 'train/grad_norm_pre_clip': 44.304359436035156, 'timing/data': 0.0008444562554359436, 'timing/model': 1.3369066370651126, 'learning_rate/qwen_vl_interface': 4.55182692266948e-06, 'learning_rate/action_model': 4.551826922669479e-05, 'epoch': 1.42}) +[2026-06-18 00:35:56,089][starVLA.training.train_starvla][INFO] - Step 3875, Loss: {'train/loss': 0.500112771987915, 'train/grad_norm_pre_clip': 19.58447265625, 'timing/data': 0.000907394103705883, 'timing/model': 1.1255253814160824, 'learning_rate/qwen_vl_interface': 4.549687793716257e-06, 'learning_rate/action_model': 4.549687793716257e-05, 'epoch': 1.42}) +[2026-06-18 00:36:02,441][starVLA.training.train_starvla][INFO] - Step 3876, Loss: {'train/loss': 0.3247809112071991, 'train/grad_norm_pre_clip': 12.749728202819824, 'timing/data': 0.0010987045243382454, 'timing/model': 1.1766423238441348, 'learning_rate/qwen_vl_interface': 4.547548809938507e-06, 'learning_rate/action_model': 4.5475488099385075e-05, 'epoch': 1.42}) +[2026-06-18 00:36:07,552][starVLA.training.train_starvla][INFO] - Step 3877, Loss: {'train/loss': 0.5058771371841431, 'train/grad_norm_pre_clip': 16.057117462158203, 'timing/data': 0.0012292293831706047, 'timing/model': 1.0595607655122876, 'learning_rate/qwen_vl_interface': 4.545409971779644e-06, 'learning_rate/action_model': 4.545409971779644e-05, 'epoch': 1.42}) +[2026-06-18 00:36:12,473][starVLA.training.train_starvla][INFO] - Step 3878, Loss: {'train/loss': 0.9035131335258484, 'train/grad_norm_pre_clip': 23.964218139648438, 'timing/data': 0.0011013373732566833, 'timing/model': 1.190720614977181, 'learning_rate/qwen_vl_interface': 4.543271279683052e-06, 'learning_rate/action_model': 4.543271279683053e-05, 'epoch': 1.42}) +[2026-06-18 00:36:17,320][starVLA.training.train_starvla][INFO] - Step 3879, Loss: {'train/loss': 0.3960631191730499, 'train/grad_norm_pre_clip': 15.258401870727539, 'timing/data': 0.0017088456079363823, 'timing/model': 1.395750449039042, 'learning_rate/qwen_vl_interface': 4.541132734092086e-06, 'learning_rate/action_model': 4.541132734092086e-05, 'epoch': 1.42}) +[2026-06-18 00:36:21,779][starVLA.training.train_starvla][INFO] - Step 3880, Loss: {'train/loss': 0.22129102051258087, 'train/grad_norm_pre_clip': 12.159629821777344, 'timing/data': 0.0009461194276809692, 'timing/model': 1.2905095955356956, 'learning_rate/qwen_vl_interface': 4.538994335450066e-06, 'learning_rate/action_model': 4.5389943354500664e-05, 'epoch': 1.42}) +[2026-06-18 00:36:26,603][starVLA.training.train_starvla][INFO] - Step 3881, Loss: {'train/loss': 0.3681502342224121, 'train/grad_norm_pre_clip': 14.485383987426758, 'timing/data': 0.0013141855597496033, 'timing/model': 1.1250487500801682, 'learning_rate/qwen_vl_interface': 4.5368560842002836e-06, 'learning_rate/action_model': 4.5368560842002834e-05, 'epoch': 1.42}) +[2026-06-18 00:36:33,745][starVLA.training.train_starvla][INFO] - Step 3882, Loss: {'train/loss': 0.4237077534198761, 'train/grad_norm_pre_clip': 17.150772094726562, 'timing/data': 0.0011748848482966423, 'timing/model': 3.394729462452233, 'learning_rate/qwen_vl_interface': 4.534717980786005e-06, 'learning_rate/action_model': 4.5347179807860046e-05, 'epoch': 1.42}) +[2026-06-18 00:36:41,947][starVLA.training.train_starvla][INFO] - Step 3883, Loss: {'train/loss': 0.2828279435634613, 'train/grad_norm_pre_clip': 12.87734603881836, 'timing/data': 0.000870085321366787, 'timing/model': 1.4341357992962003, 'learning_rate/qwen_vl_interface': 4.532580025650457e-06, 'learning_rate/action_model': 4.5325800256504574e-05, 'epoch': 1.42}) +[2026-06-18 00:36:46,628][starVLA.training.train_starvla][INFO] - Step 3884, Loss: {'train/loss': 0.39950886368751526, 'train/grad_norm_pre_clip': 16.70952796936035, 'timing/data': 0.0008912403136491776, 'timing/model': 1.4098510593175888, 'learning_rate/qwen_vl_interface': 4.530442219236842e-06, 'learning_rate/action_model': 4.530442219236842e-05, 'epoch': 1.42}) +[2026-06-18 00:36:52,210][starVLA.training.train_starvla][INFO] - Step 3885, Loss: {'train/loss': 0.7810061573982239, 'train/grad_norm_pre_clip': 23.200634002685547, 'timing/data': 0.0012931479141116142, 'timing/model': 2.444643451832235, 'learning_rate/qwen_vl_interface': 4.5283045619883285e-06, 'learning_rate/action_model': 4.5283045619883285e-05, 'epoch': 1.42}) +[2026-06-18 00:37:00,812][starVLA.training.train_starvla][INFO] - Step 3886, Loss: {'train/loss': 0.42812904715538025, 'train/grad_norm_pre_clip': 16.99622917175293, 'timing/data': 0.0012383190914988518, 'timing/model': 1.1811060048639774, 'learning_rate/qwen_vl_interface': 4.526167054348057e-06, 'learning_rate/action_model': 4.526167054348057e-05, 'epoch': 1.42}) +[2026-06-18 00:37:05,326][starVLA.training.train_starvla][INFO] - Step 3887, Loss: {'train/loss': 0.532002866268158, 'train/grad_norm_pre_clip': 23.13809585571289, 'timing/data': 0.0011041034013032913, 'timing/model': 1.0591104794293642, 'learning_rate/qwen_vl_interface': 4.524029696759135e-06, 'learning_rate/action_model': 4.5240296967591354e-05, 'epoch': 1.42}) +[2026-06-18 00:37:09,711][starVLA.training.train_starvla][INFO] - Step 3888, Loss: {'train/loss': 0.5927212238311768, 'train/grad_norm_pre_clip': 20.541501998901367, 'timing/data': 0.001092672348022461, 'timing/model': 1.3904154766350985, 'learning_rate/qwen_vl_interface': 4.5218924896646354e-06, 'learning_rate/action_model': 4.521892489664635e-05, 'epoch': 1.42}) +[2026-06-18 00:37:14,809][starVLA.training.train_starvla][INFO] - Step 3889, Loss: {'train/loss': 0.7272108793258667, 'train/grad_norm_pre_clip': 37.463314056396484, 'timing/data': 0.0008761733770370483, 'timing/model': 1.5060413554310799, 'learning_rate/qwen_vl_interface': 4.519755433507609e-06, 'learning_rate/action_model': 4.5197554335076086e-05, 'epoch': 1.42}) +[2026-06-18 00:37:18,937][starVLA.training.train_starvla][INFO] - Step 3890, Loss: {'train/loss': 0.4297485053539276, 'train/grad_norm_pre_clip': 13.52733039855957, 'timing/data': 0.0009019216522574425, 'timing/model': 1.0620124163106084, 'learning_rate/qwen_vl_interface': 4.517618528731066e-06, 'learning_rate/action_model': 4.517618528731065e-05, 'epoch': 1.42}) +[2026-06-18 00:37:25,192][starVLA.training.train_starvla][INFO] - Step 3891, Loss: {'train/loss': 0.679552435874939, 'train/grad_norm_pre_clip': 37.87443161010742, 'timing/data': 0.0012581562623381615, 'timing/model': 1.0254588294774294, 'learning_rate/qwen_vl_interface': 4.515481775777989e-06, 'learning_rate/action_model': 4.515481775777988e-05, 'epoch': 1.42}) +[2026-06-18 00:37:29,832][starVLA.training.train_starvla][INFO] - Step 3892, Loss: {'train/loss': 0.4313361644744873, 'train/grad_norm_pre_clip': 17.582937240600586, 'timing/data': 0.0015713581815361977, 'timing/model': 1.149091717787087, 'learning_rate/qwen_vl_interface': 4.51334517509133e-06, 'learning_rate/action_model': 4.5133451750913305e-05, 'epoch': 1.43}) +[2026-06-18 00:37:34,718][starVLA.training.train_starvla][INFO] - Step 3893, Loss: {'train/loss': 0.2631526589393616, 'train/grad_norm_pre_clip': 16.74932098388672, 'timing/data': 0.0019598910585045815, 'timing/model': 1.1085560880601406, 'learning_rate/qwen_vl_interface': 4.511208727114011e-06, 'learning_rate/action_model': 4.51120872711401e-05, 'epoch': 1.43}) +[2026-06-18 00:37:40,607][starVLA.training.train_starvla][INFO] - Step 3894, Loss: {'train/loss': 0.4209718406200409, 'train/grad_norm_pre_clip': 22.203344345092773, 'timing/data': 0.0008618226274847984, 'timing/model': 1.0470097493380308, 'learning_rate/qwen_vl_interface': 4.5090724322889175e-06, 'learning_rate/action_model': 4.5090724322889174e-05, 'epoch': 1.43}) +[2026-06-18 00:37:45,159][starVLA.training.train_starvla][INFO] - Step 3895, Loss: {'train/loss': 0.6338480710983276, 'train/grad_norm_pre_clip': 20.440820693969727, 'timing/data': 0.0009107356891036034, 'timing/model': 1.037213696166873, 'learning_rate/qwen_vl_interface': 4.506936291058902e-06, 'learning_rate/action_model': 4.506936291058902e-05, 'epoch': 1.43}) +[2026-06-18 00:37:51,083][starVLA.training.train_starvla][INFO] - Step 3896, Loss: {'train/loss': 0.4943186044692993, 'train/grad_norm_pre_clip': 23.955461502075195, 'timing/data': 0.0012107687070965767, 'timing/model': 1.1112557435408235, 'learning_rate/qwen_vl_interface': 4.504800303866798e-06, 'learning_rate/action_model': 4.504800303866798e-05, 'epoch': 1.43}) +[2026-06-18 00:37:55,698][starVLA.training.train_starvla][INFO] - Step 3897, Loss: {'train/loss': 0.7779622673988342, 'train/grad_norm_pre_clip': 30.607934951782227, 'timing/data': 0.0010978477075695992, 'timing/model': 1.1042477516457438, 'learning_rate/qwen_vl_interface': 4.50266447115539e-06, 'learning_rate/action_model': 4.5026644711553904e-05, 'epoch': 1.43}) +[2026-06-18 00:38:01,631][starVLA.training.train_starvla][INFO] - Step 3898, Loss: {'train/loss': 0.49405723810195923, 'train/grad_norm_pre_clip': 16.979597091674805, 'timing/data': 0.0012416243553161621, 'timing/model': 2.575558539479971, 'learning_rate/qwen_vl_interface': 4.500528793367444e-06, 'learning_rate/action_model': 4.5005287933674435e-05, 'epoch': 1.43}) +[2026-06-18 00:38:06,614][starVLA.training.train_starvla][INFO] - Step 3899, Loss: {'train/loss': 0.3751462697982788, 'train/grad_norm_pre_clip': 15.661212921142578, 'timing/data': 0.0011809244751930237, 'timing/model': 1.3634262708947062, 'learning_rate/qwen_vl_interface': 4.498393270945683e-06, 'learning_rate/action_model': 4.498393270945683e-05, 'epoch': 1.43}) +[2026-06-18 00:41:56,684][starVLA.training.train_starvla][INFO] - Step 3900, Loss: {'train/loss': 0.8492481708526611, 'train/grad_norm_pre_clip': 27.34549903869629, 'eval/loss': 0.5910860896110535, 'eval/latency_2/loss': 0.5910860896110535, 'eval/demon_fire_precision': 0.9204683084376262, 'eval/demon_fire_recall': 0.9539748953974896, 'eval/demon_fire_f1': 0.9369221286213273, 'eval/demon_fire_rate_model': 0.6892042292710072, 'eval/demon_fire_rate_teacher': 0.6649972175848636, 'eval/demon_left_precision': 0.8995215311004785, 'eval/demon_left_recall': 0.791578947368421, 'eval/demon_left_f1': 0.8421052631578948, 'eval/demon_left_rate_model': 0.23260990539788537, 'eval/demon_left_rate_teacher': 0.2643294379521425, 'eval/demon_right_precision': 0.8682432432432432, 'eval/demon_right_recall': 0.8711864406779661, 'eval/demon_right_f1': 0.8697123519458545, 'eval/demon_right_rate_model': 0.24707846410684475, 'eval/demon_right_rate_teacher': 0.24624373956594323, 'eval/demon_move_f1': 0.8559088075518746, 'eval/demon_cc_f1': 0.9110523027488382, 'eval/latency_2/demon_fire_precision': 0.9204683084376262, 'eval/latency_2/demon_fire_recall': 0.9539748953974896, 'eval/latency_2/demon_fire_f1': 0.9369221286213273, 'eval/latency_2/demon_fire_rate_model': 0.6892042292710072, 'eval/latency_2/demon_fire_rate_teacher': 0.6649972175848636, 'eval/latency_2/demon_left_precision': 0.8995215311004785, 'eval/latency_2/demon_left_recall': 0.791578947368421, 'eval/latency_2/demon_left_f1': 0.8421052631578948, 'eval/latency_2/demon_left_rate_model': 0.23260990539788537, 'eval/latency_2/demon_left_rate_teacher': 0.2643294379521425, 'eval/latency_2/demon_right_precision': 0.8682432432432432, 'eval/latency_2/demon_right_recall': 0.8711864406779661, 'eval/latency_2/demon_right_f1': 0.8697123519458545, 'eval/latency_2/demon_right_rate_model': 0.24707846410684475, 'eval/latency_2/demon_right_rate_teacher': 0.24624373956594323, 'eval/latency_2/demon_move_f1': 0.8559088075518746, 'eval/latency_2/demon_cc_f1': 0.9110523027488382, 'timing/data': 0.0008952608332037926, 'timing/model': 1.1605474967509508, 'learning_rate/qwen_vl_interface': 4.496257904332806e-06, 'learning_rate/action_model': 4.4962579043328064e-05, 'epoch': 1.43}) +[2026-06-18 00:41:56,687][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_3900_state +[2026-06-18 00:41:56,694][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 00:42:05,603][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_3900_state/model.safetensors +[2026-06-18 00:42:21,567][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_3900_state/optimizer.bin +[2026-06-18 00:42:21,568][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_3900_state/sampler.bin +[2026-06-18 00:42:21,569][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_3900_state/sampler_1.bin +[2026-06-18 00:42:21,571][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_3900_state/random_states_0.pkl +[2026-06-18 00:42:21,576][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3900 +[2026-06-18 00:42:21,577][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 00:42:21,593][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 00:42:28,977][starVLA.training.train_starvla][INFO] - Step 3901, Loss: {'train/loss': 0.6141114234924316, 'train/grad_norm_pre_clip': 23.047685623168945, 'timing/data': 0.0007959557697176933, 'timing/model': 1.0445279870182276, 'learning_rate/qwen_vl_interface': 4.494122693971478e-06, 'learning_rate/action_model': 4.4941226939714785e-05, 'epoch': 1.43}) +[2026-06-18 00:42:33,588][starVLA.training.train_starvla][INFO] - Step 3902, Loss: {'train/loss': 0.5998920202255249, 'train/grad_norm_pre_clip': 20.711408615112305, 'timing/data': 0.000889042392373085, 'timing/model': 1.0583393285050988, 'learning_rate/qwen_vl_interface': 4.491987640304327e-06, 'learning_rate/action_model': 4.4919876403043265e-05, 'epoch': 1.43}) +[2026-06-18 00:42:38,153][starVLA.training.train_starvla][INFO] - Step 3903, Loss: {'train/loss': 0.32511305809020996, 'train/grad_norm_pre_clip': 13.997939109802246, 'timing/data': 0.0010522808879613876, 'timing/model': 1.0733150094747543, 'learning_rate/qwen_vl_interface': 4.489852743773955e-06, 'learning_rate/action_model': 4.489852743773955e-05, 'epoch': 1.43}) +[2026-06-18 00:42:42,685][starVLA.training.train_starvla][INFO] - Step 3904, Loss: {'train/loss': 0.34130510687828064, 'train/grad_norm_pre_clip': 14.981335639953613, 'timing/data': 0.0012190649285912514, 'timing/model': 1.034043432213366, 'learning_rate/qwen_vl_interface': 4.487718004822928e-06, 'learning_rate/action_model': 4.487718004822928e-05, 'epoch': 1.43}) +[2026-06-18 00:42:47,136][starVLA.training.train_starvla][INFO] - Step 3905, Loss: {'train/loss': 0.43960660696029663, 'train/grad_norm_pre_clip': 19.443584442138672, 'timing/data': 0.0011424468830227852, 'timing/model': 1.3770440453663468, 'learning_rate/qwen_vl_interface': 4.485583423893778e-06, 'learning_rate/action_model': 4.4855834238937774e-05, 'epoch': 1.43}) +[2026-06-18 00:42:54,074][starVLA.training.train_starvla][INFO] - Step 3906, Loss: {'train/loss': 0.43338513374328613, 'train/grad_norm_pre_clip': 18.75566291809082, 'timing/data': 0.0011315159499645233, 'timing/model': 3.3173630768433213, 'learning_rate/qwen_vl_interface': 4.483449001429005e-06, 'learning_rate/action_model': 4.483449001429005e-05, 'epoch': 1.43}) +[2026-06-18 00:43:00,022][starVLA.training.train_starvla][INFO] - Step 3907, Loss: {'train/loss': 0.3871738016605377, 'train/grad_norm_pre_clip': 23.744503021240234, 'timing/data': 0.001071135513484478, 'timing/model': 1.0478749424219131, 'learning_rate/qwen_vl_interface': 4.481314737871081e-06, 'learning_rate/action_model': 4.4813147378710804e-05, 'epoch': 1.43}) +[2026-06-18 00:43:06,007][starVLA.training.train_starvla][INFO] - Step 3908, Loss: {'train/loss': 0.48276495933532715, 'train/grad_norm_pre_clip': 16.28141975402832, 'timing/data': 0.0011059138923883438, 'timing/model': 1.0876301229000092, 'learning_rate/qwen_vl_interface': 4.47918063366244e-06, 'learning_rate/action_model': 4.4791806336624395e-05, 'epoch': 1.43}) +[2026-06-18 00:43:10,630][starVLA.training.train_starvla][INFO] - Step 3909, Loss: {'train/loss': 0.4873145520687103, 'train/grad_norm_pre_clip': 24.662519454956055, 'timing/data': 0.0009036818519234657, 'timing/model': 1.1577432351186872, 'learning_rate/qwen_vl_interface': 4.477046689245481e-06, 'learning_rate/action_model': 4.477046689245481e-05, 'epoch': 1.43}) +[2026-06-18 00:43:15,424][starVLA.training.train_starvla][INFO] - Step 3910, Loss: {'train/loss': 0.32071223855018616, 'train/grad_norm_pre_clip': 15.44242000579834, 'timing/data': 0.0008536549285054207, 'timing/model': 1.435865101404488, 'learning_rate/qwen_vl_interface': 4.474912905062575e-06, 'learning_rate/action_model': 4.474912905062575e-05, 'epoch': 1.43}) +[2026-06-18 00:43:19,561][starVLA.training.train_starvla][INFO] - Step 3911, Loss: {'train/loss': 0.3962762951850891, 'train/grad_norm_pre_clip': 18.22639274597168, 'timing/data': 0.001646416261792183, 'timing/model': 1.0286638149991632, 'learning_rate/qwen_vl_interface': 4.472779281556059e-06, 'learning_rate/action_model': 4.4727792815560585e-05, 'epoch': 1.43}) +[2026-06-18 00:43:24,560][starVLA.training.train_starvla][INFO] - Step 3912, Loss: {'train/loss': 0.4599704146385193, 'train/grad_norm_pre_clip': 16.907106399536133, 'timing/data': 0.0009748581796884537, 'timing/model': 1.1557160839438438, 'learning_rate/qwen_vl_interface': 4.470645819168233e-06, 'learning_rate/action_model': 4.4706458191682324e-05, 'epoch': 1.43}) +[2026-06-18 00:43:29,208][starVLA.training.train_starvla][INFO] - Step 3913, Loss: {'train/loss': 0.14283818006515503, 'train/grad_norm_pre_clip': 8.681309700012207, 'timing/data': 0.0010792575776576996, 'timing/model': 1.1165151689201593, 'learning_rate/qwen_vl_interface': 4.468512518341367e-06, 'learning_rate/action_model': 4.4685125183413674e-05, 'epoch': 1.43}) +[2026-06-18 00:43:34,047][starVLA.training.train_starvla][INFO] - Step 3914, Loss: {'train/loss': 0.6155587434768677, 'train/grad_norm_pre_clip': 23.24071502685547, 'timing/data': 0.0010486692190170288, 'timing/model': 1.201006917282939, 'learning_rate/qwen_vl_interface': 4.466379379517699e-06, 'learning_rate/action_model': 4.4663793795176985e-05, 'epoch': 1.43}) +[2026-06-18 00:43:40,311][starVLA.training.train_starvla][INFO] - Step 3915, Loss: {'train/loss': 0.44857552647590637, 'train/grad_norm_pre_clip': 34.41850280761719, 'timing/data': 0.0010340400040149689, 'timing/model': 2.9161219745874405, 'learning_rate/qwen_vl_interface': 4.464246403139429e-06, 'learning_rate/action_model': 4.464246403139428e-05, 'epoch': 1.43}) +[2026-06-18 00:43:45,058][starVLA.training.train_starvla][INFO] - Step 3916, Loss: {'train/loss': 0.299003928899765, 'train/grad_norm_pre_clip': 19.37331771850586, 'timing/data': 0.0014224089682102203, 'timing/model': 1.0902827335521579, 'learning_rate/qwen_vl_interface': 4.462113589648723e-06, 'learning_rate/action_model': 4.462113589648723e-05, 'epoch': 1.43}) +[2026-06-18 00:44:00,642][starVLA.training.train_starvla][INFO] - Step 3917, Loss: {'train/loss': 0.7345923185348511, 'train/grad_norm_pre_clip': 22.13396453857422, 'timing/data': 0.0011180518195033073, 'timing/model': 5.281391621567309, 'learning_rate/qwen_vl_interface': 4.459980939487718e-06, 'learning_rate/action_model': 4.4599809394877176e-05, 'epoch': 1.43}) +[2026-06-18 00:44:05,674][starVLA.training.train_starvla][INFO] - Step 3918, Loss: {'train/loss': 0.42115360498428345, 'train/grad_norm_pre_clip': 16.427593231201172, 'timing/data': 0.0009086821228265762, 'timing/model': 1.1294077634811401, 'learning_rate/qwen_vl_interface': 4.457848453098514e-06, 'learning_rate/action_model': 4.4578484530985144e-05, 'epoch': 1.43}) +[2026-06-18 00:44:10,283][starVLA.training.train_starvla][INFO] - Step 3919, Loss: {'train/loss': 0.2704017758369446, 'train/grad_norm_pre_clip': 15.16994857788086, 'timing/data': 0.00469403900206089, 'timing/model': 1.067298092879355, 'learning_rate/qwen_vl_interface': 4.45571613092318e-06, 'learning_rate/action_model': 4.4557161309231794e-05, 'epoch': 1.44}) +[2026-06-18 00:44:14,550][starVLA.training.train_starvla][INFO] - Step 3920, Loss: {'train/loss': 0.8258283734321594, 'train/grad_norm_pre_clip': 20.914424896240234, 'timing/data': 0.0007347036153078079, 'timing/model': 1.0610492881387472, 'learning_rate/qwen_vl_interface': 4.453583973403746e-06, 'learning_rate/action_model': 4.453583973403746e-05, 'epoch': 1.44}) +[2026-06-18 00:44:24,116][starVLA.training.train_starvla][INFO] - Step 3921, Loss: {'train/loss': 0.4100973904132843, 'train/grad_norm_pre_clip': 17.010894775390625, 'timing/data': 0.0013292348012328148, 'timing/model': 4.450153459794819, 'learning_rate/qwen_vl_interface': 4.451451980982209e-06, 'learning_rate/action_model': 4.4514519809822086e-05, 'epoch': 1.44}) +[2026-06-18 00:44:29,011][starVLA.training.train_starvla][INFO] - Step 3922, Loss: {'train/loss': 0.3751884698867798, 'train/grad_norm_pre_clip': 23.169906616210938, 'timing/data': 0.0011325562372803688, 'timing/model': 1.3023612201213837, 'learning_rate/qwen_vl_interface': 4.4493201541005385e-06, 'learning_rate/action_model': 4.4493201541005384e-05, 'epoch': 1.44}) +[2026-06-18 00:44:33,817][starVLA.training.train_starvla][INFO] - Step 3923, Loss: {'train/loss': 0.36774909496307373, 'train/grad_norm_pre_clip': 15.712175369262695, 'timing/data': 0.0011873329058289528, 'timing/model': 1.3154838420450687, 'learning_rate/qwen_vl_interface': 4.447188493200661e-06, 'learning_rate/action_model': 4.447188493200661e-05, 'epoch': 1.44}) +[2026-06-18 00:44:38,970][starVLA.training.train_starvla][INFO] - Step 3924, Loss: {'train/loss': 0.29655295610427856, 'train/grad_norm_pre_clip': 17.449451446533203, 'timing/data': 0.0013321582227945328, 'timing/model': 1.1166544444859028, 'learning_rate/qwen_vl_interface': 4.445056998724472e-06, 'learning_rate/action_model': 4.4450569987244716e-05, 'epoch': 1.44}) +[2026-06-18 00:44:44,429][starVLA.training.train_starvla][INFO] - Step 3925, Loss: {'train/loss': 0.401848167181015, 'train/grad_norm_pre_clip': 19.828798294067383, 'timing/data': 0.001211082562804222, 'timing/model': 1.4596527321264148, 'learning_rate/qwen_vl_interface': 4.4429256711138335e-06, 'learning_rate/action_model': 4.442925671113833e-05, 'epoch': 1.44}) +[2026-06-18 00:44:50,920][starVLA.training.train_starvla][INFO] - Step 3926, Loss: {'train/loss': 0.435716450214386, 'train/grad_norm_pre_clip': 18.507854461669922, 'timing/data': 0.0019176509231328964, 'timing/model': 1.12776193395257, 'learning_rate/qwen_vl_interface': 4.440794510810572e-06, 'learning_rate/action_model': 4.440794510810572e-05, 'epoch': 1.44}) +[2026-06-18 00:44:55,669][starVLA.training.train_starvla][INFO] - Step 3927, Loss: {'train/loss': 0.3897062838077545, 'train/grad_norm_pre_clip': 20.63623809814453, 'timing/data': 0.012772451154887676, 'timing/model': 1.0586253106594086, 'learning_rate/qwen_vl_interface': 4.438663518256481e-06, 'learning_rate/action_model': 4.438663518256481e-05, 'epoch': 1.44}) +[2026-06-18 00:45:00,161][starVLA.training.train_starvla][INFO] - Step 3928, Loss: {'train/loss': 0.29739952087402344, 'train/grad_norm_pre_clip': 21.86008071899414, 'timing/data': 0.024859627708792686, 'timing/model': 1.067519216798246, 'learning_rate/qwen_vl_interface': 4.436532693893311e-06, 'learning_rate/action_model': 4.436532693893311e-05, 'epoch': 1.44}) +[2026-06-18 00:45:06,318][starVLA.training.train_starvla][INFO] - Step 3929, Loss: {'train/loss': 0.3629373610019684, 'train/grad_norm_pre_clip': 16.941856384277344, 'timing/data': 0.020820624195039272, 'timing/model': 1.2571283150464296, 'learning_rate/qwen_vl_interface': 4.434402038162793e-06, 'learning_rate/action_model': 4.434402038162793e-05, 'epoch': 1.44}) +[2026-06-18 00:45:11,462][starVLA.training.train_starvla][INFO] - Step 3930, Loss: {'train/loss': 0.48372334241867065, 'train/grad_norm_pre_clip': 16.20751190185547, 'timing/data': 0.03448431845754385, 'timing/model': 1.3861596742644906, 'learning_rate/qwen_vl_interface': 4.432271551506609e-06, 'learning_rate/action_model': 4.4322715515066094e-05, 'epoch': 1.44}) +[2026-06-18 00:45:16,408][starVLA.training.train_starvla][INFO] - Step 3931, Loss: {'train/loss': 0.29500067234039307, 'train/grad_norm_pre_clip': 16.74030876159668, 'timing/data': 0.02065837662667036, 'timing/model': 1.0487142195925117, 'learning_rate/qwen_vl_interface': 4.4301412343664125e-06, 'learning_rate/action_model': 4.4301412343664125e-05, 'epoch': 1.44}) +[2026-06-18 00:45:20,891][starVLA.training.train_starvla][INFO] - Step 3932, Loss: {'train/loss': 0.23622828722000122, 'train/grad_norm_pre_clip': 11.844786643981934, 'timing/data': 0.02077876590192318, 'timing/model': 1.1236936952918768, 'learning_rate/qwen_vl_interface': 4.428011087183819e-06, 'learning_rate/action_model': 4.428011087183819e-05, 'epoch': 1.44}) +[2026-06-18 00:45:30,427][starVLA.training.train_starvla][INFO] - Step 3933, Loss: {'train/loss': 0.4214305877685547, 'train/grad_norm_pre_clip': 26.928834915161133, 'timing/data': 0.03087586537003517, 'timing/model': 5.897676362656057, 'learning_rate/qwen_vl_interface': 4.4258811104004125e-06, 'learning_rate/action_model': 4.425881110400413e-05, 'epoch': 1.44}) +[2026-06-18 00:45:35,169][starVLA.training.train_starvla][INFO] - Step 3934, Loss: {'train/loss': 0.3376081883907318, 'train/grad_norm_pre_clip': 18.33509635925293, 'timing/data': 0.02397761680185795, 'timing/model': 1.0685639418661594, 'learning_rate/qwen_vl_interface': 4.4237513044577394e-06, 'learning_rate/action_model': 4.4237513044577386e-05, 'epoch': 1.44}) +[2026-06-18 00:45:40,050][starVLA.training.train_starvla][INFO] - Step 3935, Loss: {'train/loss': 0.5179955959320068, 'train/grad_norm_pre_clip': 20.937976837158203, 'timing/data': 0.023694983683526516, 'timing/model': 1.5799043141305447, 'learning_rate/qwen_vl_interface': 4.421621669797309e-06, 'learning_rate/action_model': 4.421621669797308e-05, 'epoch': 1.44}) +[2026-06-18 00:45:44,675][starVLA.training.train_starvla][INFO] - Step 3936, Loss: {'train/loss': 0.4157220721244812, 'train/grad_norm_pre_clip': 17.473329544067383, 'timing/data': 0.02127001527696848, 'timing/model': 1.038895620033145, 'learning_rate/qwen_vl_interface': 4.419492206860599e-06, 'learning_rate/action_model': 4.419492206860598e-05, 'epoch': 1.44}) +[2026-06-18 00:45:49,432][starVLA.training.train_starvla][INFO] - Step 3937, Loss: {'train/loss': 0.34845414757728577, 'train/grad_norm_pre_clip': 21.468658447265625, 'timing/data': 0.026270785368978977, 'timing/model': 1.348051493987441, 'learning_rate/qwen_vl_interface': 4.417362916089048e-06, 'learning_rate/action_model': 4.417362916089047e-05, 'epoch': 1.44}) +[2026-06-18 00:45:55,118][starVLA.training.train_starvla][INFO] - Step 3938, Loss: {'train/loss': 0.5547336339950562, 'train/grad_norm_pre_clip': 25.77518081665039, 'timing/data': 0.08215810731053352, 'timing/model': 1.6963734403252602, 'learning_rate/qwen_vl_interface': 4.415233797924059e-06, 'learning_rate/action_model': 4.415233797924059e-05, 'epoch': 1.44}) +[2026-06-18 00:46:01,825][starVLA.training.train_starvla][INFO] - Step 3939, Loss: {'train/loss': 0.36308592557907104, 'train/grad_norm_pre_clip': 21.032472610473633, 'timing/data': 0.06583029124885798, 'timing/model': 2.7006825655698776, 'learning_rate/qwen_vl_interface': 4.413104852807002e-06, 'learning_rate/action_model': 4.413104852807001e-05, 'epoch': 1.44}) +[2026-06-18 00:46:14,915][starVLA.training.train_starvla][INFO] - Step 3940, Loss: {'train/loss': 0.6283694505691528, 'train/grad_norm_pre_clip': 19.75181007385254, 'timing/data': 0.029870478436350822, 'timing/model': 6.400752135552466, 'learning_rate/qwen_vl_interface': 4.41097608117921e-06, 'learning_rate/action_model': 4.41097608117921e-05, 'epoch': 1.44}) +[2026-06-18 00:46:20,284][starVLA.training.train_starvla][INFO] - Step 3941, Loss: {'train/loss': 0.4776798486709595, 'train/grad_norm_pre_clip': 21.584829330444336, 'timing/data': 0.03634791634976864, 'timing/model': 1.3150974977761507, 'learning_rate/qwen_vl_interface': 4.40884748348198e-06, 'learning_rate/action_model': 4.40884748348198e-05, 'epoch': 1.44}) +[2026-06-18 00:46:25,021][starVLA.training.train_starvla][INFO] - Step 3942, Loss: {'train/loss': 0.5124818086624146, 'train/grad_norm_pre_clip': 21.53670310974121, 'timing/data': 0.028523094952106476, 'timing/model': 1.0830500070005655, 'learning_rate/qwen_vl_interface': 4.406719060156572e-06, 'learning_rate/action_model': 4.4067190601565716e-05, 'epoch': 1.44}) +[2026-06-18 00:46:31,945][starVLA.training.train_starvla][INFO] - Step 3943, Loss: {'train/loss': 0.7349024415016174, 'train/grad_norm_pre_clip': 31.097625732421875, 'timing/data': 0.020799784921109676, 'timing/model': 1.1047991318628192, 'learning_rate/qwen_vl_interface': 4.404590811644209e-06, 'learning_rate/action_model': 4.4045908116442087e-05, 'epoch': 1.44}) +[2026-06-18 00:46:44,981][starVLA.training.train_starvla][INFO] - Step 3944, Loss: {'train/loss': 0.5983426570892334, 'train/grad_norm_pre_clip': 21.94326400756836, 'timing/data': 0.09584825951606035, 'timing/model': 2.8742831284180284, 'learning_rate/qwen_vl_interface': 4.40246273838608e-06, 'learning_rate/action_model': 4.40246273838608e-05, 'epoch': 1.44}) +[2026-06-18 00:46:59,224][starVLA.training.train_starvla][INFO] - Step 3945, Loss: {'train/loss': 0.4833504259586334, 'train/grad_norm_pre_clip': 22.230051040649414, 'timing/data': 0.08497092220932245, 'timing/model': 3.49132835585624, 'learning_rate/qwen_vl_interface': 4.400334840823339e-06, 'learning_rate/action_model': 4.400334840823339e-05, 'epoch': 1.44}) +[2026-06-18 00:47:07,581][starVLA.training.train_starvla][INFO] - Step 3946, Loss: {'train/loss': 0.48985686898231506, 'train/grad_norm_pre_clip': 19.07956314086914, 'timing/data': 0.022635722532868385, 'timing/model': 1.1225872030481696, 'learning_rate/qwen_vl_interface': 4.398207119397098e-06, 'learning_rate/action_model': 4.398207119397098e-05, 'epoch': 1.44}) +[2026-06-18 00:47:17,425][starVLA.training.train_starvla][INFO] - Step 3947, Loss: {'train/loss': 0.5888810753822327, 'train/grad_norm_pre_clip': 33.85850524902344, 'timing/data': 0.06628238875418901, 'timing/model': 2.4993499675765634, 'learning_rate/qwen_vl_interface': 4.396079574548439e-06, 'learning_rate/action_model': 4.396079574548439e-05, 'epoch': 1.45}) +[2026-06-18 00:47:21,995][starVLA.training.train_starvla][INFO] - Step 3948, Loss: {'train/loss': 0.27882808446884155, 'train/grad_norm_pre_clip': 18.912979125976562, 'timing/data': 0.021384770050644875, 'timing/model': 1.0351464357227087, 'learning_rate/qwen_vl_interface': 4.393952206718403e-06, 'learning_rate/action_model': 4.3939522067184026e-05, 'epoch': 1.45}) +[2026-06-18 00:47:26,489][starVLA.training.train_starvla][INFO] - Step 3949, Loss: {'train/loss': 0.2565257251262665, 'train/grad_norm_pre_clip': 13.431167602539062, 'timing/data': 0.02587536908686161, 'timing/model': 1.1145219160243869, 'learning_rate/qwen_vl_interface': 4.391825016347996e-06, 'learning_rate/action_model': 4.391825016347996e-05, 'epoch': 1.45}) +[2026-06-18 00:47:30,987][starVLA.training.train_starvla][INFO] - Step 3950, Loss: {'train/loss': 0.6123746633529663, 'train/grad_norm_pre_clip': 22.60341453552246, 'timing/data': 0.029530377127230167, 'timing/model': 1.456826038658619, 'learning_rate/qwen_vl_interface': 4.3896980038781844e-06, 'learning_rate/action_model': 4.3896980038781846e-05, 'epoch': 1.45}) +[2026-06-18 00:47:35,178][starVLA.training.train_starvla][INFO] - Step 3951, Loss: {'train/loss': 0.4492437243461609, 'train/grad_norm_pre_clip': 14.579318046569824, 'timing/data': 0.021963694132864475, 'timing/model': 1.0876439427956939, 'learning_rate/qwen_vl_interface': 4.3875711697499054e-06, 'learning_rate/action_model': 4.387571169749905e-05, 'epoch': 1.45}) +[2026-06-18 00:47:39,649][starVLA.training.train_starvla][INFO] - Step 3952, Loss: {'train/loss': 0.3158763349056244, 'train/grad_norm_pre_clip': 18.21701431274414, 'timing/data': 0.022249657660722733, 'timing/model': 1.1665666745975614, 'learning_rate/qwen_vl_interface': 4.38544451440405e-06, 'learning_rate/action_model': 4.3854445144040495e-05, 'epoch': 1.45}) +[2026-06-18 00:47:44,256][starVLA.training.train_starvla][INFO] - Step 3953, Loss: {'train/loss': 0.4245011806488037, 'train/grad_norm_pre_clip': 19.313608169555664, 'timing/data': 0.024471834301948547, 'timing/model': 1.2854054048657417, 'learning_rate/qwen_vl_interface': 4.383318038281475e-06, 'learning_rate/action_model': 4.383318038281475e-05, 'epoch': 1.45}) +[2026-06-18 00:47:49,207][starVLA.training.train_starvla][INFO] - Step 3954, Loss: {'train/loss': 0.3768711984157562, 'train/grad_norm_pre_clip': 14.948295593261719, 'timing/data': 0.02109823003411293, 'timing/model': 1.1267076125368476, 'learning_rate/qwen_vl_interface': 4.381191741823005e-06, 'learning_rate/action_model': 4.3811917418230046e-05, 'epoch': 1.45}) +[2026-06-18 00:47:54,028][starVLA.training.train_starvla][INFO] - Step 3955, Loss: {'train/loss': 0.3848734200000763, 'train/grad_norm_pre_clip': 17.798307418823242, 'timing/data': 0.01925052423030138, 'timing/model': 1.3600540086627007, 'learning_rate/qwen_vl_interface': 4.379065625469422e-06, 'learning_rate/action_model': 4.3790656254694215e-05, 'epoch': 1.45}) +[2026-06-18 00:47:58,080][starVLA.training.train_starvla][INFO] - Step 3956, Loss: {'train/loss': 0.18470336496829987, 'train/grad_norm_pre_clip': 10.260894775390625, 'timing/data': 0.024286523461341858, 'timing/model': 1.0341406799852848, 'learning_rate/qwen_vl_interface': 4.3769396896614724e-06, 'learning_rate/action_model': 4.376939689661472e-05, 'epoch': 1.45}) +[2026-06-18 00:48:02,485][starVLA.training.train_starvla][INFO] - Step 3957, Loss: {'train/loss': 0.36087098717689514, 'train/grad_norm_pre_clip': 15.960343360900879, 'timing/data': 0.02536492608487606, 'timing/model': 1.0279842456802726, 'learning_rate/qwen_vl_interface': 4.374813934839862e-06, 'learning_rate/action_model': 4.374813934839862e-05, 'epoch': 1.45}) +[2026-06-18 00:48:07,335][starVLA.training.train_starvla][INFO] - Step 3958, Loss: {'train/loss': 0.8583725690841675, 'train/grad_norm_pre_clip': 30.61599349975586, 'timing/data': 0.02201608568429947, 'timing/model': 1.3267570044845343, 'learning_rate/qwen_vl_interface': 4.372688361445267e-06, 'learning_rate/action_model': 4.3726883614452665e-05, 'epoch': 1.45}) +[2026-06-18 00:48:12,523][starVLA.training.train_starvla][INFO] - Step 3959, Loss: {'train/loss': 0.36833006143569946, 'train/grad_norm_pre_clip': 15.900422096252441, 'timing/data': 0.03031846508383751, 'timing/model': 1.0649755503982306, 'learning_rate/qwen_vl_interface': 4.370562969918319e-06, 'learning_rate/action_model': 4.370562969918319e-05, 'epoch': 1.45}) +[2026-06-18 00:48:17,259][starVLA.training.train_starvla][INFO] - Step 3960, Loss: {'train/loss': 0.21869197487831116, 'train/grad_norm_pre_clip': 21.017702102661133, 'timing/data': 0.021184386685490608, 'timing/model': 1.3359847189858556, 'learning_rate/qwen_vl_interface': 4.368437760699613e-06, 'learning_rate/action_model': 4.3684377606996124e-05, 'epoch': 1.45}) +[2026-06-18 00:48:21,554][starVLA.training.train_starvla][INFO] - Step 3961, Loss: {'train/loss': 0.4059707522392273, 'train/grad_norm_pre_clip': 16.788928985595703, 'timing/data': 0.023605553433299065, 'timing/model': 1.0381133891642094, 'learning_rate/qwen_vl_interface': 4.366312734229706e-06, 'learning_rate/action_model': 4.366312734229706e-05, 'epoch': 1.45}) +[2026-06-18 00:48:26,115][starVLA.training.train_starvla][INFO] - Step 3962, Loss: {'train/loss': 0.22718557715415955, 'train/grad_norm_pre_clip': 13.524837493896484, 'timing/data': 0.024780471809208393, 'timing/model': 1.1434454191476107, 'learning_rate/qwen_vl_interface': 4.364187890949122e-06, 'learning_rate/action_model': 4.364187890949122e-05, 'epoch': 1.45}) +[2026-06-18 00:48:30,706][starVLA.training.train_starvla][INFO] - Step 3963, Loss: {'train/loss': 0.3870219588279724, 'train/grad_norm_pre_clip': 22.624767303466797, 'timing/data': 0.020815206691622734, 'timing/model': 1.2255446342751384, 'learning_rate/qwen_vl_interface': 4.362063231298342e-06, 'learning_rate/action_model': 4.3620632312983425e-05, 'epoch': 1.45}) +[2026-06-18 00:48:39,885][starVLA.training.train_starvla][INFO] - Step 3964, Loss: {'train/loss': 0.4010727107524872, 'train/grad_norm_pre_clip': 18.995288848876953, 'timing/data': 0.0314135430380702, 'timing/model': 1.448952897451818, 'learning_rate/qwen_vl_interface': 4.359938755717807e-06, 'learning_rate/action_model': 4.359938755717807e-05, 'epoch': 1.45}) +[2026-06-18 00:48:50,599][starVLA.training.train_starvla][INFO] - Step 3965, Loss: {'train/loss': 0.44621437788009644, 'train/grad_norm_pre_clip': 19.17833709716797, 'timing/data': 0.0011667115613818169, 'timing/model': 2.983183161355555, 'learning_rate/qwen_vl_interface': 4.357814464647929e-06, 'learning_rate/action_model': 4.357814464647929e-05, 'epoch': 1.45}) +[2026-06-18 00:49:01,715][starVLA.training.train_starvla][INFO] - Step 3966, Loss: {'train/loss': 0.1708211898803711, 'train/grad_norm_pre_clip': 12.034560203552246, 'timing/data': 0.003264321945607662, 'timing/model': 1.1000576820224524, 'learning_rate/qwen_vl_interface': 4.355690358529071e-06, 'learning_rate/action_model': 4.3556903585290705e-05, 'epoch': 1.45}) +[2026-06-18 00:49:11,701][starVLA.training.train_starvla][INFO] - Step 3967, Loss: {'train/loss': 0.4577227234840393, 'train/grad_norm_pre_clip': 18.72517204284668, 'timing/data': 0.0009637251496315002, 'timing/model': 1.4190168716013432, 'learning_rate/qwen_vl_interface': 4.353566437801564e-06, 'learning_rate/action_model': 4.353566437801563e-05, 'epoch': 1.45}) +[2026-06-18 00:49:22,388][starVLA.training.train_starvla][INFO] - Step 3968, Loss: {'train/loss': 0.5164358019828796, 'train/grad_norm_pre_clip': 17.2469482421875, 'timing/data': 0.0019597262144088745, 'timing/model': 4.147611074149609, 'learning_rate/qwen_vl_interface': 4.351442702905698e-06, 'learning_rate/action_model': 4.3514427029056976e-05, 'epoch': 1.45}) +[2026-06-18 00:49:34,643][starVLA.training.train_starvla][INFO] - Step 3969, Loss: {'train/loss': 0.23496085405349731, 'train/grad_norm_pre_clip': 10.787391662597656, 'timing/data': 0.0009408546611666679, 'timing/model': 1.3310176311060786, 'learning_rate/qwen_vl_interface': 4.349319154281727e-06, 'learning_rate/action_model': 4.349319154281728e-05, 'epoch': 1.45}) +[2026-06-18 00:49:41,011][starVLA.training.train_starvla][INFO] - Step 3970, Loss: {'train/loss': 0.417621374130249, 'train/grad_norm_pre_clip': 17.084657669067383, 'timing/data': 0.006983707658946514, 'timing/model': 1.537190836854279, 'learning_rate/qwen_vl_interface': 4.347195792369867e-06, 'learning_rate/action_model': 4.347195792369867e-05, 'epoch': 1.45}) +[2026-06-18 00:49:46,839][starVLA.training.train_starvla][INFO] - Step 3971, Loss: {'train/loss': 0.4055440127849579, 'train/grad_norm_pre_clip': 25.04513931274414, 'timing/data': 0.001368662342429161, 'timing/model': 1.06992844119668, 'learning_rate/qwen_vl_interface': 4.345072617610288e-06, 'learning_rate/action_model': 4.345072617610288e-05, 'epoch': 1.45}) +[2026-06-18 00:49:53,303][starVLA.training.train_starvla][INFO] - Step 3972, Loss: {'train/loss': 0.5844811201095581, 'train/grad_norm_pre_clip': 27.968463897705078, 'timing/data': 0.001155046746134758, 'timing/model': 1.1250421712175012, 'learning_rate/qwen_vl_interface': 4.342949630443129e-06, 'learning_rate/action_model': 4.342949630443129e-05, 'epoch': 1.45}) +[2026-06-18 00:49:57,893][starVLA.training.train_starvla][INFO] - Step 3973, Loss: {'train/loss': 0.32120078802108765, 'train/grad_norm_pre_clip': 19.198511123657227, 'timing/data': 0.00123574398458004, 'timing/model': 1.0692433482035995, 'learning_rate/qwen_vl_interface': 4.340826831308488e-06, 'learning_rate/action_model': 4.340826831308488e-05, 'epoch': 1.45}) +[2026-06-18 00:50:03,163][starVLA.training.train_starvla][INFO] - Step 3974, Loss: {'train/loss': 0.8132457733154297, 'train/grad_norm_pre_clip': 29.197635650634766, 'timing/data': 0.0012031486257910728, 'timing/model': 1.2121793245896697, 'learning_rate/qwen_vl_interface': 4.338704220646424e-06, 'learning_rate/action_model': 4.338704220646424e-05, 'epoch': 1.46}) +[2026-06-18 00:50:08,234][starVLA.training.train_starvla][INFO] - Step 3975, Loss: {'train/loss': 0.35110801458358765, 'train/grad_norm_pre_clip': 14.677173614501953, 'timing/data': 0.002049149014055729, 'timing/model': 1.657537535764277, 'learning_rate/qwen_vl_interface': 4.336581798896955e-06, 'learning_rate/action_model': 4.336581798896955e-05, 'epoch': 1.46}) +[2026-06-18 00:50:12,557][starVLA.training.train_starvla][INFO] - Step 3976, Loss: {'train/loss': 0.26560425758361816, 'train/grad_norm_pre_clip': 18.73978042602539, 'timing/data': 0.0011970540508627892, 'timing/model': 1.047270193696022, 'learning_rate/qwen_vl_interface': 4.334459566500062e-06, 'learning_rate/action_model': 4.334459566500062e-05, 'epoch': 1.46}) +[2026-06-18 00:50:17,392][starVLA.training.train_starvla][INFO] - Step 3977, Loss: {'train/loss': 0.9293757677078247, 'train/grad_norm_pre_clip': 27.297710418701172, 'timing/data': 0.0008990131318569183, 'timing/model': 1.094766098074615, 'learning_rate/qwen_vl_interface': 4.332337523895688e-06, 'learning_rate/action_model': 4.3323375238956876e-05, 'epoch': 1.46}) +[2026-06-18 00:50:21,810][starVLA.training.train_starvla][INFO] - Step 3978, Loss: {'train/loss': 0.6144795417785645, 'train/grad_norm_pre_clip': 17.849336624145508, 'timing/data': 0.0009369049221277237, 'timing/model': 1.0942601477727294, 'learning_rate/qwen_vl_interface': 4.330215671523731e-06, 'learning_rate/action_model': 4.330215671523731e-05, 'epoch': 1.46}) +[2026-06-18 00:50:26,508][starVLA.training.train_starvla][INFO] - Step 3979, Loss: {'train/loss': 0.45305880904197693, 'train/grad_norm_pre_clip': 26.747800827026367, 'timing/data': 0.0015323860570788383, 'timing/model': 1.3493507653474808, 'learning_rate/qwen_vl_interface': 4.328094009824054e-06, 'learning_rate/action_model': 4.328094009824054e-05, 'epoch': 1.46}) +[2026-06-18 00:50:30,881][starVLA.training.train_starvla][INFO] - Step 3980, Loss: {'train/loss': 0.45361894369125366, 'train/grad_norm_pre_clip': 22.810325622558594, 'timing/data': 0.0011877939105033875, 'timing/model': 1.3234936594963074, 'learning_rate/qwen_vl_interface': 4.3259725392364825e-06, 'learning_rate/action_model': 4.325972539236482e-05, 'epoch': 1.46}) +[2026-06-18 00:50:37,811][starVLA.training.train_starvla][INFO] - Step 3981, Loss: {'train/loss': 0.3062610626220703, 'train/grad_norm_pre_clip': 19.55953598022461, 'timing/data': 0.0009462283924221992, 'timing/model': 1.0612603863701224, 'learning_rate/qwen_vl_interface': 4.323851260200797e-06, 'learning_rate/action_model': 4.323851260200796e-05, 'epoch': 1.46}) +[2026-06-18 00:50:50,052][starVLA.training.train_starvla][INFO] - Step 3982, Loss: {'train/loss': 0.39818882942199707, 'train/grad_norm_pre_clip': 16.302949905395508, 'timing/data': 0.0011610602959990501, 'timing/model': 5.498110039159656, 'learning_rate/qwen_vl_interface': 4.321730173156743e-06, 'learning_rate/action_model': 4.321730173156743e-05, 'epoch': 1.46}) +[2026-06-18 00:50:56,445][starVLA.training.train_starvla][INFO] - Step 3983, Loss: {'train/loss': 0.2156929224729538, 'train/grad_norm_pre_clip': 15.003260612487793, 'timing/data': 0.0013027014210820198, 'timing/model': 1.593805025331676, 'learning_rate/qwen_vl_interface': 4.31960927854402e-06, 'learning_rate/action_model': 4.31960927854402e-05, 'epoch': 1.46}) +[2026-06-18 00:51:05,691][starVLA.training.train_starvla][INFO] - Step 3984, Loss: {'train/loss': 0.5859538912773132, 'train/grad_norm_pre_clip': 19.967206954956055, 'timing/data': 0.0009029526263475418, 'timing/model': 1.0720375897362828, 'learning_rate/qwen_vl_interface': 4.317488576802298e-06, 'learning_rate/action_model': 4.3174885768022986e-05, 'epoch': 1.46}) +[2026-06-18 00:51:10,709][starVLA.training.train_starvla][INFO] - Step 3985, Loss: {'train/loss': 0.21386004984378815, 'train/grad_norm_pre_clip': 13.462677001953125, 'timing/data': 0.0009481478482484818, 'timing/model': 1.8334947051480412, 'learning_rate/qwen_vl_interface': 4.315368068371197e-06, 'learning_rate/action_model': 4.315368068371197e-05, 'epoch': 1.46}) +[2026-06-18 00:51:15,213][starVLA.training.train_starvla][INFO] - Step 3986, Loss: {'train/loss': 0.49586403369903564, 'train/grad_norm_pre_clip': 25.822389602661133, 'timing/data': 0.001133420504629612, 'timing/model': 1.4344948111101985, 'learning_rate/qwen_vl_interface': 4.313247753690298e-06, 'learning_rate/action_model': 4.3132477536902985e-05, 'epoch': 1.46}) +[2026-06-18 00:51:20,474][starVLA.training.train_starvla][INFO] - Step 3987, Loss: {'train/loss': 0.6191734671592712, 'train/grad_norm_pre_clip': 19.003849029541016, 'timing/data': 0.0012333476915955544, 'timing/model': 1.1761481817811728, 'learning_rate/qwen_vl_interface': 4.311127633199151e-06, 'learning_rate/action_model': 4.3111276331991504e-05, 'epoch': 1.46}) +[2026-06-18 00:51:26,003][starVLA.training.train_starvla][INFO] - Step 3988, Loss: {'train/loss': 0.6123104691505432, 'train/grad_norm_pre_clip': 26.39083480834961, 'timing/data': 0.0012777429074048996, 'timing/model': 2.035140964202583, 'learning_rate/qwen_vl_interface': 4.309007707337255e-06, 'learning_rate/action_model': 4.309007707337255e-05, 'epoch': 1.46}) +[2026-06-18 00:51:38,755][starVLA.training.train_starvla][INFO] - Step 3989, Loss: {'train/loss': 0.1224859282374382, 'train/grad_norm_pre_clip': 9.666459083557129, 'timing/data': 0.0011330218985676765, 'timing/model': 4.361162879504263, 'learning_rate/qwen_vl_interface': 4.3068879765440745e-06, 'learning_rate/action_model': 4.306887976544074e-05, 'epoch': 1.46}) +[2026-06-18 00:51:43,393][starVLA.training.train_starvla][INFO] - Step 3990, Loss: {'train/loss': 0.34938177466392517, 'train/grad_norm_pre_clip': 14.575019836425781, 'timing/data': 0.0012030517682433128, 'timing/model': 1.2270980766043067, 'learning_rate/qwen_vl_interface': 4.304768441259029e-06, 'learning_rate/action_model': 4.3047684412590285e-05, 'epoch': 1.46}) +[2026-06-18 00:51:48,001][starVLA.training.train_starvla][INFO] - Step 3991, Loss: {'train/loss': 0.9285422563552856, 'train/grad_norm_pre_clip': 25.620304107666016, 'timing/data': 0.001337866298854351, 'timing/model': 1.0576559118926525, 'learning_rate/qwen_vl_interface': 4.302649101921505e-06, 'learning_rate/action_model': 4.302649101921505e-05, 'epoch': 1.46}) +[2026-06-18 00:51:52,763][starVLA.training.train_starvla][INFO] - Step 3992, Loss: {'train/loss': 0.4133770167827606, 'train/grad_norm_pre_clip': 18.973665237426758, 'timing/data': 0.0012287171557545662, 'timing/model': 1.1966236578300595, 'learning_rate/qwen_vl_interface': 4.300529958970841e-06, 'learning_rate/action_model': 4.300529958970841e-05, 'epoch': 1.46}) +[2026-06-18 00:51:57,512][starVLA.training.train_starvla][INFO] - Step 3993, Loss: {'train/loss': 0.3343648910522461, 'train/grad_norm_pre_clip': 18.38667106628418, 'timing/data': 0.0012075500562787056, 'timing/model': 1.0909999487921596, 'learning_rate/qwen_vl_interface': 4.298411012846338e-06, 'learning_rate/action_model': 4.2984110128463376e-05, 'epoch': 1.46}) +[2026-06-18 00:52:01,945][starVLA.training.train_starvla][INFO] - Step 3994, Loss: {'train/loss': 0.17713408172130585, 'train/grad_norm_pre_clip': 17.106103897094727, 'timing/data': 0.0009006671607494354, 'timing/model': 1.0602431166917086, 'learning_rate/qwen_vl_interface': 4.296292263987255e-06, 'learning_rate/action_model': 4.2962922639872556e-05, 'epoch': 1.46}) +[2026-06-18 00:52:06,233][starVLA.training.train_starvla][INFO] - Step 3995, Loss: {'train/loss': 0.32989731431007385, 'train/grad_norm_pre_clip': 15.181333541870117, 'timing/data': 0.0023280102759599686, 'timing/model': 1.0814142702147365, 'learning_rate/qwen_vl_interface': 4.294173712832813e-06, 'learning_rate/action_model': 4.294173712832813e-05, 'epoch': 1.46}) +[2026-06-18 00:52:10,650][starVLA.training.train_starvla][INFO] - Step 3996, Loss: {'train/loss': 0.44377845525741577, 'train/grad_norm_pre_clip': 17.33855438232422, 'timing/data': 0.0011504888534545898, 'timing/model': 1.0682822652161121, 'learning_rate/qwen_vl_interface': 4.292055359822191e-06, 'learning_rate/action_model': 4.29205535982219e-05, 'epoch': 1.46}) +[2026-06-18 00:52:15,042][starVLA.training.train_starvla][INFO] - Step 3997, Loss: {'train/loss': 0.47083237767219543, 'train/grad_norm_pre_clip': 21.33613395690918, 'timing/data': 0.007049312815070152, 'timing/model': 1.032079664990306, 'learning_rate/qwen_vl_interface': 4.289937205394519e-06, 'learning_rate/action_model': 4.2899372053945185e-05, 'epoch': 1.46}) +[2026-06-18 00:52:19,767][starVLA.training.train_starvla][INFO] - Step 3998, Loss: {'train/loss': 0.3995954692363739, 'train/grad_norm_pre_clip': 17.783065795898438, 'timing/data': 0.0007660100236535072, 'timing/model': 1.064238678663969, 'learning_rate/qwen_vl_interface': 4.287819249988901e-06, 'learning_rate/action_model': 4.2878192499889007e-05, 'epoch': 1.46}) +[2026-06-18 00:52:24,503][starVLA.training.train_starvla][INFO] - Step 3999, Loss: {'train/loss': 0.422232449054718, 'train/grad_norm_pre_clip': 16.74533462524414, 'timing/data': 0.0008703079074621201, 'timing/model': 1.6401312872767448, 'learning_rate/qwen_vl_interface': 4.285701494044387e-06, 'learning_rate/action_model': 4.285701494044387e-05, 'epoch': 1.46}) +[2026-06-18 00:57:11,740][starVLA.training.train_starvla][INFO] - Step 4000, Loss: {'train/loss': 0.39581000804901123, 'train/grad_norm_pre_clip': 19.639516830444336, 'eval/loss': 0.5596145987510681, 'eval/latency_2/loss': 0.5596145987510681, 'eval/demon_fire_precision': 0.920500403551251, 'eval/demon_fire_recall': 0.9543933054393305, 'eval/demon_fire_f1': 0.9371405094494658, 'eval/demon_fire_rate_model': 0.6894824707846411, 'eval/demon_fire_rate_teacher': 0.6649972175848636, 'eval/demon_left_precision': 0.8591101694915254, 'eval/demon_left_recall': 0.8536842105263158, 'eval/demon_left_f1': 0.8563885955649418, 'eval/demon_left_rate_model': 0.2626599888703395, 'eval/demon_left_rate_teacher': 0.2643294379521425, 'eval/demon_right_precision': 0.9054054054054054, 'eval/demon_right_recall': 0.8327683615819209, 'eval/demon_right_f1': 0.8675691583284285, 'eval/demon_right_rate_model': 0.226488592097941, 'eval/demon_right_rate_teacher': 0.24624373956594323, 'eval/demon_move_f1': 0.8619788769466852, 'eval/demon_cc_f1': 0.9132507984376877, 'eval/latency_2/demon_fire_precision': 0.920500403551251, 'eval/latency_2/demon_fire_recall': 0.9543933054393305, 'eval/latency_2/demon_fire_f1': 0.9371405094494658, 'eval/latency_2/demon_fire_rate_model': 0.6894824707846411, 'eval/latency_2/demon_fire_rate_teacher': 0.6649972175848636, 'eval/latency_2/demon_left_precision': 0.8591101694915254, 'eval/latency_2/demon_left_recall': 0.8536842105263158, 'eval/latency_2/demon_left_f1': 0.8563885955649418, 'eval/latency_2/demon_left_rate_model': 0.2626599888703395, 'eval/latency_2/demon_left_rate_teacher': 0.2643294379521425, 'eval/latency_2/demon_right_precision': 0.9054054054054054, 'eval/latency_2/demon_right_recall': 0.8327683615819209, 'eval/latency_2/demon_right_f1': 0.8675691583284285, 'eval/latency_2/demon_right_rate_model': 0.226488592097941, 'eval/latency_2/demon_right_rate_teacher': 0.24624373956594323, 'eval/latency_2/demon_move_f1': 0.8619788769466852, 'eval/latency_2/demon_cc_f1': 0.9132507984376877, 'timing/data': 0.0008704699575901031, 'timing/model': 1.422811179421842, 'learning_rate/qwen_vl_interface': 4.283583937999991e-06, 'learning_rate/action_model': 4.2835839379999906e-05, 'epoch': 1.46}) +[2026-06-18 00:57:11,753][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_4000_state +[2026-06-18 00:57:11,761][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 00:57:21,014][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_4000_state/model.safetensors +[2026-06-18 00:57:36,268][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_4000_state/optimizer.bin +[2026-06-18 00:57:36,269][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_4000_state/sampler.bin +[2026-06-18 00:57:36,270][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_4000_state/sampler_1.bin +[2026-06-18 00:57:36,274][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_4000_state/random_states_0.pkl +[2026-06-18 00:57:36,279][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4000 +[2026-06-18 00:57:36,280][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 00:57:36,297][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 00:57:41,358][starVLA.training.train_starvla][INFO] - Step 4001, Loss: {'train/loss': 0.6304876804351807, 'train/grad_norm_pre_clip': 21.941486358642578, 'timing/data': 0.0012279320508241653, 'timing/model': 1.6186788976192474, 'learning_rate/qwen_vl_interface': 4.2814665822946825e-06, 'learning_rate/action_model': 4.281466582294682e-05, 'epoch': 1.47}) +[2026-06-18 00:57:46,083][starVLA.training.train_starvla][INFO] - Step 4002, Loss: {'train/loss': 0.3063041567802429, 'train/grad_norm_pre_clip': 13.497056007385254, 'timing/data': 0.0012153973802924156, 'timing/model': 1.3153601475059986, 'learning_rate/qwen_vl_interface': 4.279349427367395e-06, 'learning_rate/action_model': 4.279349427367395e-05, 'epoch': 1.47}) +[2026-06-18 00:57:51,816][starVLA.training.train_starvla][INFO] - Step 4003, Loss: {'train/loss': 0.33734387159347534, 'train/grad_norm_pre_clip': 18.612571716308594, 'timing/data': 0.0008821999654173851, 'timing/model': 1.079476767219603, 'learning_rate/qwen_vl_interface': 4.277232473657018e-06, 'learning_rate/action_model': 4.277232473657018e-05, 'epoch': 1.47}) +[2026-06-18 00:57:56,396][starVLA.training.train_starvla][INFO] - Step 4004, Loss: {'train/loss': 0.3154279887676239, 'train/grad_norm_pre_clip': 15.488100051879883, 'timing/data': 0.006341953761875629, 'timing/model': 1.070524075999856, 'learning_rate/qwen_vl_interface': 4.2751157216023934e-06, 'learning_rate/action_model': 4.2751157216023936e-05, 'epoch': 1.47}) +[2026-06-18 00:58:01,308][starVLA.training.train_starvla][INFO] - Step 4005, Loss: {'train/loss': 0.4661400020122528, 'train/grad_norm_pre_clip': 20.408615112304688, 'timing/data': 0.001147470436990261, 'timing/model': 1.0525154722854495, 'learning_rate/qwen_vl_interface': 4.272999171642329e-06, 'learning_rate/action_model': 4.272999171642329e-05, 'epoch': 1.47}) +[2026-06-18 00:58:07,668][starVLA.training.train_starvla][INFO] - Step 4006, Loss: {'train/loss': 0.4323364198207855, 'train/grad_norm_pre_clip': 16.05345916748047, 'timing/data': 0.0008407672867178917, 'timing/model': 3.0215250682085752, 'learning_rate/qwen_vl_interface': 4.270882824215588e-06, 'learning_rate/action_model': 4.270882824215588e-05, 'epoch': 1.47}) +[2026-06-18 00:58:16,115][starVLA.training.train_starvla][INFO] - Step 4007, Loss: {'train/loss': 0.3408740758895874, 'train/grad_norm_pre_clip': 12.14084529876709, 'timing/data': 0.0012476108968257904, 'timing/model': 4.457283799536526, 'learning_rate/qwen_vl_interface': 4.268766679760891e-06, 'learning_rate/action_model': 4.268766679760891e-05, 'epoch': 1.47}) +[2026-06-18 00:58:24,793][starVLA.training.train_starvla][INFO] - Step 4008, Loss: {'train/loss': 0.27084049582481384, 'train/grad_norm_pre_clip': 24.51999855041504, 'timing/data': 0.014291730709373951, 'timing/model': 1.0451849475502968, 'learning_rate/qwen_vl_interface': 4.266650738716916e-06, 'learning_rate/action_model': 4.266650738716916e-05, 'epoch': 1.47}) +[2026-06-18 00:58:29,376][starVLA.training.train_starvla][INFO] - Step 4009, Loss: {'train/loss': 0.44005778431892395, 'train/grad_norm_pre_clip': 19.246755599975586, 'timing/data': 0.031611320562660694, 'timing/model': 1.059451992623508, 'learning_rate/qwen_vl_interface': 4.264535001522302e-06, 'learning_rate/action_model': 4.2645350015223025e-05, 'epoch': 1.47}) +[2026-06-18 00:58:34,322][starVLA.training.train_starvla][INFO] - Step 4010, Loss: {'train/loss': 0.5857511758804321, 'train/grad_norm_pre_clip': 21.099855422973633, 'timing/data': 0.02100671175867319, 'timing/model': 1.2250273944810033, 'learning_rate/qwen_vl_interface': 4.262419468615644e-06, 'learning_rate/action_model': 4.2624194686156435e-05, 'epoch': 1.47}) +[2026-06-18 00:58:41,665][starVLA.training.train_starvla][INFO] - Step 4011, Loss: {'train/loss': 0.24492372572422028, 'train/grad_norm_pre_clip': 18.563196182250977, 'timing/data': 0.03286587633192539, 'timing/model': 2.552158510312438, 'learning_rate/qwen_vl_interface': 4.2603041404354925e-06, 'learning_rate/action_model': 4.260304140435492e-05, 'epoch': 1.47}) +[2026-06-18 00:58:46,488][starVLA.training.train_starvla][INFO] - Step 4012, Loss: {'train/loss': 0.29227861762046814, 'train/grad_norm_pre_clip': 16.988327026367188, 'timing/data': 0.024484089575707912, 'timing/model': 1.3311230605468154, 'learning_rate/qwen_vl_interface': 4.258189017420356e-06, 'learning_rate/action_model': 4.258189017420355e-05, 'epoch': 1.47}) +[2026-06-18 00:58:51,642][starVLA.training.train_starvla][INFO] - Step 4013, Loss: {'train/loss': 0.4871469736099243, 'train/grad_norm_pre_clip': 24.197509765625, 'timing/data': 0.031090390868484974, 'timing/model': 1.1699274647980928, 'learning_rate/qwen_vl_interface': 4.2560741000087035e-06, 'learning_rate/action_model': 4.2560741000087034e-05, 'epoch': 1.47}) +[2026-06-18 00:58:56,428][starVLA.training.train_starvla][INFO] - Step 4014, Loss: {'train/loss': 0.38179269433021545, 'train/grad_norm_pre_clip': 20.59944725036621, 'timing/data': 0.029987918213009834, 'timing/model': 1.1109833009541035, 'learning_rate/qwen_vl_interface': 4.253959388638961e-06, 'learning_rate/action_model': 4.2539593886389604e-05, 'epoch': 1.47}) +[2026-06-18 00:59:01,640][starVLA.training.train_starvla][INFO] - Step 4015, Loss: {'train/loss': 0.39055997133255005, 'train/grad_norm_pre_clip': 22.098630905151367, 'timing/data': 0.03217998240143061, 'timing/model': 1.2048821179196239, 'learning_rate/qwen_vl_interface': 4.251844883749507e-06, 'learning_rate/action_model': 4.2518448837495066e-05, 'epoch': 1.47}) +[2026-06-18 00:59:11,055][starVLA.training.train_starvla][INFO] - Step 4016, Loss: {'train/loss': 0.13017140328884125, 'train/grad_norm_pre_clip': 8.813830375671387, 'timing/data': 0.023914174176752567, 'timing/model': 1.7718584816902876, 'learning_rate/qwen_vl_interface': 4.249730585778684e-06, 'learning_rate/action_model': 4.2497305857786834e-05, 'epoch': 1.47}) +[2026-06-18 00:59:16,234][starVLA.training.train_starvla][INFO] - Step 4017, Loss: {'train/loss': 0.4781106114387512, 'train/grad_norm_pre_clip': 15.818788528442383, 'timing/data': 0.026021553203463554, 'timing/model': 1.055730378255248, 'learning_rate/qwen_vl_interface': 4.247616495164788e-06, 'learning_rate/action_model': 4.247616495164787e-05, 'epoch': 1.47}) +[2026-06-18 00:59:21,357][starVLA.training.train_starvla][INFO] - Step 4018, Loss: {'train/loss': 0.4758804142475128, 'train/grad_norm_pre_clip': 18.7835693359375, 'timing/data': 0.023777450434863567, 'timing/model': 1.2744037369266152, 'learning_rate/qwen_vl_interface': 4.24550261234607e-06, 'learning_rate/action_model': 4.2455026123460694e-05, 'epoch': 1.47}) +[2026-06-18 00:59:28,087][starVLA.training.train_starvla][INFO] - Step 4019, Loss: {'train/loss': 0.3947765827178955, 'train/grad_norm_pre_clip': 23.044300079345703, 'timing/data': 0.023127286694943905, 'timing/model': 1.5269562331959605, 'learning_rate/qwen_vl_interface': 4.2433889377607405e-06, 'learning_rate/action_model': 4.24338893776074e-05, 'epoch': 1.47}) +[2026-06-18 00:59:41,281][starVLA.training.train_starvla][INFO] - Step 4020, Loss: {'train/loss': 0.7291425466537476, 'train/grad_norm_pre_clip': 31.186521530151367, 'timing/data': 0.02493153139948845, 'timing/model': 1.0401434525847435, 'learning_rate/qwen_vl_interface': 4.241275471846968e-06, 'learning_rate/action_model': 4.241275471846968e-05, 'epoch': 1.47}) +[2026-06-18 00:59:46,280][starVLA.training.train_starvla][INFO] - Step 4021, Loss: {'train/loss': 0.46673890948295593, 'train/grad_norm_pre_clip': 20.92971420288086, 'timing/data': 0.02862130757421255, 'timing/model': 1.492512377910316, 'learning_rate/qwen_vl_interface': 4.239162215042876e-06, 'learning_rate/action_model': 4.239162215042876e-05, 'epoch': 1.47}) +[2026-06-18 00:59:50,525][starVLA.training.train_starvla][INFO] - Step 4022, Loss: {'train/loss': 0.3353986144065857, 'train/grad_norm_pre_clip': 15.189139366149902, 'timing/data': 0.020882762037217617, 'timing/model': 1.059524204581976, 'learning_rate/qwen_vl_interface': 4.237049167786544e-06, 'learning_rate/action_model': 4.2370491677865444e-05, 'epoch': 1.47}) +[2026-06-18 00:59:56,438][starVLA.training.train_starvla][INFO] - Step 4023, Loss: {'train/loss': 0.31198108196258545, 'train/grad_norm_pre_clip': 22.990449905395508, 'timing/data': 0.021047220565378666, 'timing/model': 1.1559631442651153, 'learning_rate/qwen_vl_interface': 4.234936330516009e-06, 'learning_rate/action_model': 4.234936330516009e-05, 'epoch': 1.47}) +[2026-06-18 01:00:00,788][starVLA.training.train_starvla][INFO] - Step 4024, Loss: {'train/loss': 0.36814144253730774, 'train/grad_norm_pre_clip': 19.668960571289062, 'timing/data': 0.02109682187438011, 'timing/model': 1.0348498253151774, 'learning_rate/qwen_vl_interface': 4.2328237036692675e-06, 'learning_rate/action_model': 4.232823703669267e-05, 'epoch': 1.47}) +[2026-06-18 01:00:07,031][starVLA.training.train_starvla][INFO] - Step 4025, Loss: {'train/loss': 0.3128606379032135, 'train/grad_norm_pre_clip': 21.471942901611328, 'timing/data': 0.06998689938336611, 'timing/model': 2.5236474527046084, 'learning_rate/qwen_vl_interface': 4.2307112876842645e-06, 'learning_rate/action_model': 4.2307112876842645e-05, 'epoch': 1.47}) +[2026-06-18 01:00:12,144][starVLA.training.train_starvla][INFO] - Step 4026, Loss: {'train/loss': 0.30980339646339417, 'train/grad_norm_pre_clip': 27.34092140197754, 'timing/data': 0.022101351991295815, 'timing/model': 1.5122914612293243, 'learning_rate/qwen_vl_interface': 4.228599082998908e-06, 'learning_rate/action_model': 4.228599082998908e-05, 'epoch': 1.47}) +[2026-06-18 01:00:18,053][starVLA.training.train_starvla][INFO] - Step 4027, Loss: {'train/loss': 0.29359108209609985, 'train/grad_norm_pre_clip': 14.842780113220215, 'timing/data': 0.021307311952114105, 'timing/model': 1.2519888849928975, 'learning_rate/qwen_vl_interface': 4.226487090051062e-06, 'learning_rate/action_model': 4.2264870900510624e-05, 'epoch': 1.47}) +[2026-06-18 01:00:22,621][starVLA.training.train_starvla][INFO] - Step 4028, Loss: {'train/loss': 0.48522958159446716, 'train/grad_norm_pre_clip': 21.44671058654785, 'timing/data': 0.029778397642076015, 'timing/model': 1.176390734501183, 'learning_rate/qwen_vl_interface': 4.224375309278543e-06, 'learning_rate/action_model': 4.2243753092785434e-05, 'epoch': 1.47}) +[2026-06-18 01:00:27,076][starVLA.training.train_starvla][INFO] - Step 4029, Loss: {'train/loss': 0.8783861994743347, 'train/grad_norm_pre_clip': 22.951557159423828, 'timing/data': 0.024293127469718456, 'timing/model': 1.1379231344908476, 'learning_rate/qwen_vl_interface': 4.222263741119127e-06, 'learning_rate/action_model': 4.222263741119126e-05, 'epoch': 1.48}) +[2026-06-18 01:00:31,898][starVLA.training.train_starvla][INFO] - Step 4030, Loss: {'train/loss': 0.44937029480934143, 'train/grad_norm_pre_clip': 18.36113739013672, 'timing/data': 0.02140205819159746, 'timing/model': 1.2161725349724293, 'learning_rate/qwen_vl_interface': 4.220152386010541e-06, 'learning_rate/action_model': 4.2201523860105405e-05, 'epoch': 1.48}) +[2026-06-18 01:00:36,555][starVLA.training.train_starvla][INFO] - Step 4031, Loss: {'train/loss': 0.7214072942733765, 'train/grad_norm_pre_clip': 29.270545959472656, 'timing/data': 0.02001994475722313, 'timing/model': 1.4460711888968945, 'learning_rate/qwen_vl_interface': 4.218041244390476e-06, 'learning_rate/action_model': 4.218041244390475e-05, 'epoch': 1.48}) +[2026-06-18 01:00:40,758][starVLA.training.train_starvla][INFO] - Step 4032, Loss: {'train/loss': 0.4830623269081116, 'train/grad_norm_pre_clip': 31.03228759765625, 'timing/data': 0.0239678043872118, 'timing/model': 1.0721460478380322, 'learning_rate/qwen_vl_interface': 4.215930316696571e-06, 'learning_rate/action_model': 4.215930316696571e-05, 'epoch': 1.48}) +[2026-06-18 01:00:45,158][starVLA.training.train_starvla][INFO] - Step 4033, Loss: {'train/loss': 0.6763601303100586, 'train/grad_norm_pre_clip': 23.267845153808594, 'timing/data': 0.023550556041300297, 'timing/model': 1.0483627477660775, 'learning_rate/qwen_vl_interface': 4.213819603366424e-06, 'learning_rate/action_model': 4.213819603366424e-05, 'epoch': 1.48}) +[2026-06-18 01:00:49,556][starVLA.training.train_starvla][INFO] - Step 4034, Loss: {'train/loss': 0.33089810609817505, 'train/grad_norm_pre_clip': 12.766908645629883, 'timing/data': 0.02031254582107067, 'timing/model': 1.0469876183196902, 'learning_rate/qwen_vl_interface': 4.211709104837587e-06, 'learning_rate/action_model': 4.211709104837587e-05, 'epoch': 1.48}) +[2026-06-18 01:00:54,158][starVLA.training.train_starvla][INFO] - Step 4035, Loss: {'train/loss': 0.3999062478542328, 'train/grad_norm_pre_clip': 22.502788543701172, 'timing/data': 0.030048150569200516, 'timing/model': 1.0398725336417556, 'learning_rate/qwen_vl_interface': 4.2095988215475705e-06, 'learning_rate/action_model': 4.20959882154757e-05, 'epoch': 1.48}) +[2026-06-18 01:00:59,054][starVLA.training.train_starvla][INFO] - Step 4036, Loss: {'train/loss': 0.3955219089984894, 'train/grad_norm_pre_clip': 18.697589874267578, 'timing/data': 0.07156113348901272, 'timing/model': 1.5200080601498485, 'learning_rate/qwen_vl_interface': 4.207488753933838e-06, 'learning_rate/action_model': 4.207488753933838e-05, 'epoch': 1.48}) +[2026-06-18 01:01:07,349][starVLA.training.train_starvla][INFO] - Step 4037, Loss: {'train/loss': 0.6807300448417664, 'train/grad_norm_pre_clip': 21.509830474853516, 'timing/data': 0.02559693343937397, 'timing/model': 1.3864768855273724, 'learning_rate/qwen_vl_interface': 4.205378902433807e-06, 'learning_rate/action_model': 4.205378902433807e-05, 'epoch': 1.48}) +[2026-06-18 01:01:11,889][starVLA.training.train_starvla][INFO] - Step 4038, Loss: {'train/loss': 0.4626932740211487, 'train/grad_norm_pre_clip': 21.775634765625, 'timing/data': 0.021317653357982635, 'timing/model': 1.0466917250305414, 'learning_rate/qwen_vl_interface': 4.203269267484856e-06, 'learning_rate/action_model': 4.203269267484856e-05, 'epoch': 1.48}) +[2026-06-18 01:01:16,456][starVLA.training.train_starvla][INFO] - Step 4039, Loss: {'train/loss': 0.4470655620098114, 'train/grad_norm_pre_clip': 18.534072875976562, 'timing/data': 0.023347601294517517, 'timing/model': 1.041415961459279, 'learning_rate/qwen_vl_interface': 4.201159849524313e-06, 'learning_rate/action_model': 4.201159849524312e-05, 'epoch': 1.48}) +[2026-06-18 01:01:20,953][starVLA.training.train_starvla][INFO] - Step 4040, Loss: {'train/loss': 0.4396692216396332, 'train/grad_norm_pre_clip': 16.890745162963867, 'timing/data': 0.02097894996404648, 'timing/model': 1.3647863380610943, 'learning_rate/qwen_vl_interface': 4.19905064898946e-06, 'learning_rate/action_model': 4.199050648989459e-05, 'epoch': 1.48}) +[2026-06-18 01:01:25,550][starVLA.training.train_starvla][INFO] - Step 4041, Loss: {'train/loss': 0.6323849558830261, 'train/grad_norm_pre_clip': 25.57065773010254, 'timing/data': 0.021855175495147705, 'timing/model': 1.4719619592651725, 'learning_rate/qwen_vl_interface': 4.196941666317537e-06, 'learning_rate/action_model': 4.196941666317537e-05, 'epoch': 1.48}) +[2026-06-18 01:01:33,041][starVLA.training.train_starvla][INFO] - Step 4042, Loss: {'train/loss': 0.3710786700248718, 'train/grad_norm_pre_clip': 19.304882049560547, 'timing/data': 0.025514211505651474, 'timing/model': 1.0840093763545156, 'learning_rate/qwen_vl_interface': 4.194832901945741e-06, 'learning_rate/action_model': 4.194832901945741e-05, 'epoch': 1.48}) +[2026-06-18 01:01:37,908][starVLA.training.train_starvla][INFO] - Step 4043, Loss: {'train/loss': 0.3214668333530426, 'train/grad_norm_pre_clip': 20.436708450317383, 'timing/data': 0.026276621967554092, 'timing/model': 1.1307039624080062, 'learning_rate/qwen_vl_interface': 4.192724356311221e-06, 'learning_rate/action_model': 4.19272435631122e-05, 'epoch': 1.48}) +[2026-06-18 01:01:42,951][starVLA.training.train_starvla][INFO] - Step 4044, Loss: {'train/loss': 0.2641790509223938, 'train/grad_norm_pre_clip': 18.612058639526367, 'timing/data': 0.02871518861502409, 'timing/model': 1.1826616702601314, 'learning_rate/qwen_vl_interface': 4.190616029851079e-06, 'learning_rate/action_model': 4.190616029851079e-05, 'epoch': 1.48}) +[2026-06-18 01:01:47,731][starVLA.training.train_starvla][INFO] - Step 4045, Loss: {'train/loss': 0.3392026126384735, 'train/grad_norm_pre_clip': 21.207237243652344, 'timing/data': 0.035780902951955795, 'timing/model': 1.2526895320042968, 'learning_rate/qwen_vl_interface': 4.1885079230023715e-06, 'learning_rate/action_model': 4.1885079230023715e-05, 'epoch': 1.48}) +[2026-06-18 01:01:52,141][starVLA.training.train_starvla][INFO] - Step 4046, Loss: {'train/loss': 0.4969137907028198, 'train/grad_norm_pre_clip': 24.277368545532227, 'timing/data': 0.01983346790075302, 'timing/model': 1.4371941331773996, 'learning_rate/qwen_vl_interface': 4.186400036202117e-06, 'learning_rate/action_model': 4.1864000362021175e-05, 'epoch': 1.48}) +[2026-06-18 01:01:56,597][starVLA.training.train_starvla][INFO] - Step 4047, Loss: {'train/loss': 0.5602379441261292, 'train/grad_norm_pre_clip': 23.750459671020508, 'timing/data': 0.025021997280418873, 'timing/model': 1.0751817924901843, 'learning_rate/qwen_vl_interface': 4.184292369887279e-06, 'learning_rate/action_model': 4.1842923698872785e-05, 'epoch': 1.48}) +[2026-06-18 01:02:01,203][starVLA.training.train_starvla][INFO] - Step 4048, Loss: {'train/loss': 0.5118407011032104, 'train/grad_norm_pre_clip': 21.571847915649414, 'timing/data': 0.02131829224526882, 'timing/model': 1.0483414800837636, 'learning_rate/qwen_vl_interface': 4.182184924494778e-06, 'learning_rate/action_model': 4.1821849244947773e-05, 'epoch': 1.48}) +[2026-06-18 01:02:06,833][starVLA.training.train_starvla][INFO] - Step 4049, Loss: {'train/loss': 0.4324217736721039, 'train/grad_norm_pre_clip': 16.326231002807617, 'timing/data': 0.032683685421943665, 'timing/model': 1.0802996614947915, 'learning_rate/qwen_vl_interface': 4.1800777004614915e-06, 'learning_rate/action_model': 4.180077700461492e-05, 'epoch': 1.48}) +[2026-06-18 01:02:13,108][starVLA.training.train_starvla][INFO] - Step 4050, Loss: {'train/loss': 0.3362801671028137, 'train/grad_norm_pre_clip': 21.220458984375, 'timing/data': 0.023870990611612797, 'timing/model': 1.5680985292419791, 'learning_rate/qwen_vl_interface': 4.177970698224249e-06, 'learning_rate/action_model': 4.17797069822425e-05, 'epoch': 1.48}) +[2026-06-18 01:02:22,928][starVLA.training.train_starvla][INFO] - Step 4051, Loss: {'train/loss': 0.29807400703430176, 'train/grad_norm_pre_clip': 16.335655212402344, 'timing/data': 0.034686723724007607, 'timing/model': 1.5207074340432882, 'learning_rate/qwen_vl_interface': 4.175863918219836e-06, 'learning_rate/action_model': 4.175863918219836e-05, 'epoch': 1.48}) +[2026-06-18 01:02:32,867][starVLA.training.train_starvla][INFO] - Step 4052, Loss: {'train/loss': 0.39385655522346497, 'train/grad_norm_pre_clip': 19.40715789794922, 'timing/data': 0.07837790437042713, 'timing/model': 2.650648046284914, 'learning_rate/qwen_vl_interface': 4.173757360884986e-06, 'learning_rate/action_model': 4.173757360884986e-05, 'epoch': 1.48}) +[2026-06-18 01:02:45,523][starVLA.training.train_starvla][INFO] - Step 4053, Loss: {'train/loss': 0.6218623518943787, 'train/grad_norm_pre_clip': 18.20462989807129, 'timing/data': 0.022482357919216156, 'timing/model': 1.262380689382553, 'learning_rate/qwen_vl_interface': 4.171651026656396e-06, 'learning_rate/action_model': 4.171651026656395e-05, 'epoch': 1.48}) +[2026-06-18 01:02:54,903][starVLA.training.train_starvla][INFO] - Step 4054, Loss: {'train/loss': 0.4262750744819641, 'train/grad_norm_pre_clip': 20.673526763916016, 'timing/data': 0.028412328101694584, 'timing/model': 5.680728328414261, 'learning_rate/qwen_vl_interface': 4.1695449159707066e-06, 'learning_rate/action_model': 4.169544915970707e-05, 'epoch': 1.48}) +[2026-06-18 01:03:01,454][starVLA.training.train_starvla][INFO] - Step 4055, Loss: {'train/loss': 0.46934249997138977, 'train/grad_norm_pre_clip': 18.151424407958984, 'timing/data': 0.028064321726560593, 'timing/model': 1.2990167001262307, 'learning_rate/qwen_vl_interface': 4.167439029264519e-06, 'learning_rate/action_model': 4.167439029264519e-05, 'epoch': 1.48}) +[2026-06-18 01:03:10,633][starVLA.training.train_starvla][INFO] - Step 4056, Loss: {'train/loss': 0.323012113571167, 'train/grad_norm_pre_clip': 22.31620216369629, 'timing/data': 0.022993236780166626, 'timing/model': 1.3700688006356359, 'learning_rate/qwen_vl_interface': 4.165333366974385e-06, 'learning_rate/action_model': 4.165333366974385e-05, 'epoch': 1.49}) +[2026-06-18 01:03:18,631][starVLA.training.train_starvla][INFO] - Step 4057, Loss: {'train/loss': 0.219946950674057, 'train/grad_norm_pre_clip': 11.683483123779297, 'timing/data': 0.0226550055667758, 'timing/model': 1.255896002985537, 'learning_rate/qwen_vl_interface': 4.163227929536812e-06, 'learning_rate/action_model': 4.163227929536812e-05, 'epoch': 1.49}) +[2026-06-18 01:03:26,535][starVLA.training.train_starvla][INFO] - Step 4058, Loss: {'train/loss': 0.362887442111969, 'train/grad_norm_pre_clip': 14.745152473449707, 'timing/data': 0.02348131686449051, 'timing/model': 1.103158188983798, 'learning_rate/qwen_vl_interface': 4.1611227173882596e-06, 'learning_rate/action_model': 4.1611227173882596e-05, 'epoch': 1.49}) +[2026-06-18 01:03:31,903][starVLA.training.train_starvla][INFO] - Step 4059, Loss: {'train/loss': 0.4757653772830963, 'train/grad_norm_pre_clip': 20.74151039123535, 'timing/data': 0.02287418395280838, 'timing/model': 1.0406101271510124, 'learning_rate/qwen_vl_interface': 4.1590177309651365e-06, 'learning_rate/action_model': 4.159017730965136e-05, 'epoch': 1.49}) +[2026-06-18 01:03:36,725][starVLA.training.train_starvla][INFO] - Step 4060, Loss: {'train/loss': 0.4225671887397766, 'train/grad_norm_pre_clip': 16.798213958740234, 'timing/data': 0.029865475371479988, 'timing/model': 1.1322208167985082, 'learning_rate/qwen_vl_interface': 4.156912970703815e-06, 'learning_rate/action_model': 4.1569129707038155e-05, 'epoch': 1.49}) +[2026-06-18 01:03:41,230][starVLA.training.train_starvla][INFO] - Step 4061, Loss: {'train/loss': 0.4715708792209625, 'train/grad_norm_pre_clip': 17.611515045166016, 'timing/data': 0.020976267755031586, 'timing/model': 1.439346737228334, 'learning_rate/qwen_vl_interface': 4.15480843704061e-06, 'learning_rate/action_model': 4.154808437040609e-05, 'epoch': 1.49}) +[2026-06-18 01:03:45,291][starVLA.training.train_starvla][INFO] - Step 4062, Loss: {'train/loss': 0.3695083260536194, 'train/grad_norm_pre_clip': 19.078384399414062, 'timing/data': 0.02647331077605486, 'timing/model': 1.051060490310192, 'learning_rate/qwen_vl_interface': 4.152704130411794e-06, 'learning_rate/action_model': 4.152704130411793e-05, 'epoch': 1.49}) +[2026-06-18 01:03:49,802][starVLA.training.train_starvla][INFO] - Step 4063, Loss: {'train/loss': 0.3692471385002136, 'train/grad_norm_pre_clip': 13.382027626037598, 'timing/data': 0.023075520992279053, 'timing/model': 1.0789651228114963, 'learning_rate/qwen_vl_interface': 4.150600051253591e-06, 'learning_rate/action_model': 4.150600051253591e-05, 'epoch': 1.49}) +[2026-06-18 01:03:54,350][starVLA.training.train_starvla][INFO] - Step 4064, Loss: {'train/loss': 0.36815503239631653, 'train/grad_norm_pre_clip': 17.799741744995117, 'timing/data': 0.021653744392096996, 'timing/model': 1.078055608086288, 'learning_rate/qwen_vl_interface': 4.148496200002183e-06, 'learning_rate/action_model': 4.148496200002183e-05, 'epoch': 1.49}) +[2026-06-18 01:03:58,875][starVLA.training.train_starvla][INFO] - Step 4065, Loss: {'train/loss': 0.2490968555212021, 'train/grad_norm_pre_clip': 18.389001846313477, 'timing/data': 0.022185806185007095, 'timing/model': 1.0434309113770723, 'learning_rate/qwen_vl_interface': 4.146392577093697e-06, 'learning_rate/action_model': 4.1463925770936974e-05, 'epoch': 1.49}) +[2026-06-18 01:04:03,280][starVLA.training.train_starvla][INFO] - Step 4066, Loss: {'train/loss': 0.40375927090644836, 'train/grad_norm_pre_clip': 18.928552627563477, 'timing/data': 0.019381627440452576, 'timing/model': 1.3764084344729781, 'learning_rate/qwen_vl_interface': 4.144289182964218e-06, 'learning_rate/action_model': 4.144289182964218e-05, 'epoch': 1.49}) +[2026-06-18 01:04:07,460][starVLA.training.train_starvla][INFO] - Step 4067, Loss: {'train/loss': 0.34007155895233154, 'train/grad_norm_pre_clip': 17.688966751098633, 'timing/data': 0.0228998688980937, 'timing/model': 1.0755610661581159, 'learning_rate/qwen_vl_interface': 4.14218601804978e-06, 'learning_rate/action_model': 4.1421860180497796e-05, 'epoch': 1.49}) +[2026-06-18 01:04:12,149][starVLA.training.train_starvla][INFO] - Step 4068, Loss: {'train/loss': 0.1403256207704544, 'train/grad_norm_pre_clip': 11.06124496459961, 'timing/data': 0.026093704625964165, 'timing/model': 1.111790208145976, 'learning_rate/qwen_vl_interface': 4.1400830827863725e-06, 'learning_rate/action_model': 4.1400830827863725e-05, 'epoch': 1.49}) +[2026-06-18 01:04:16,667][starVLA.training.train_starvla][INFO] - Step 4069, Loss: {'train/loss': 0.3793894052505493, 'train/grad_norm_pre_clip': 22.22478485107422, 'timing/data': 0.02316232118755579, 'timing/model': 1.0295676486566663, 'learning_rate/qwen_vl_interface': 4.137980377609937e-06, 'learning_rate/action_model': 4.137980377609937e-05, 'epoch': 1.49}) +[2026-06-18 01:04:21,286][starVLA.training.train_starvla][INFO] - Step 4070, Loss: {'train/loss': 0.2423447072505951, 'train/grad_norm_pre_clip': 11.74959659576416, 'timing/data': 0.02423442993313074, 'timing/model': 1.0393068147823215, 'learning_rate/qwen_vl_interface': 4.135877902956364e-06, 'learning_rate/action_model': 4.1358779029563636e-05, 'epoch': 1.49}) +[2026-06-18 01:04:25,514][starVLA.training.train_starvla][INFO] - Step 4071, Loss: {'train/loss': 0.47395890951156616, 'train/grad_norm_pre_clip': 19.786287307739258, 'timing/data': 0.022033533081412315, 'timing/model': 1.1465618684887886, 'learning_rate/qwen_vl_interface': 4.133775659261503e-06, 'learning_rate/action_model': 4.133775659261502e-05, 'epoch': 1.49}) +[2026-06-18 01:04:30,002][starVLA.training.train_starvla][INFO] - Step 4072, Loss: {'train/loss': 0.18028636276721954, 'train/grad_norm_pre_clip': 12.015436172485352, 'timing/data': 0.021382653154432774, 'timing/model': 1.030333793722093, 'learning_rate/qwen_vl_interface': 4.131673646961149e-06, 'learning_rate/action_model': 4.131673646961149e-05, 'epoch': 1.49}) +[2026-06-18 01:04:34,689][starVLA.training.train_starvla][INFO] - Step 4073, Loss: {'train/loss': 0.545781672000885, 'train/grad_norm_pre_clip': 17.091501235961914, 'timing/data': 0.023266078904271126, 'timing/model': 1.2158730179071426, 'learning_rate/qwen_vl_interface': 4.129571866491051e-06, 'learning_rate/action_model': 4.1295718664910504e-05, 'epoch': 1.49}) +[2026-06-18 01:04:39,607][starVLA.training.train_starvla][INFO] - Step 4074, Loss: {'train/loss': 0.6609887480735779, 'train/grad_norm_pre_clip': 24.396465301513672, 'timing/data': 0.028056911192834377, 'timing/model': 1.238604593090713, 'learning_rate/qwen_vl_interface': 4.127470318286908e-06, 'learning_rate/action_model': 4.127470318286908e-05, 'epoch': 1.49}) +[2026-06-18 01:04:49,212][starVLA.training.train_starvla][INFO] - Step 4075, Loss: {'train/loss': 0.2847009301185608, 'train/grad_norm_pre_clip': 19.14892578125, 'timing/data': 0.024018866941332817, 'timing/model': 1.463661071844399, 'learning_rate/qwen_vl_interface': 4.125369002784378e-06, 'learning_rate/action_model': 4.1253690027843776e-05, 'epoch': 1.49}) +[2026-06-18 01:04:57,318][starVLA.training.train_starvla][INFO] - Step 4076, Loss: {'train/loss': 0.23131969571113586, 'train/grad_norm_pre_clip': 15.168798446655273, 'timing/data': 0.025377577170729637, 'timing/model': 1.5726539893075824, 'learning_rate/qwen_vl_interface': 4.123267920419064e-06, 'learning_rate/action_model': 4.123267920419064e-05, 'epoch': 1.49}) +[2026-06-18 01:05:09,268][starVLA.training.train_starvla][INFO] - Step 4077, Loss: {'train/loss': 0.4281880557537079, 'train/grad_norm_pre_clip': 14.335880279541016, 'timing/data': 0.024981182999908924, 'timing/model': 1.4216907564550638, 'learning_rate/qwen_vl_interface': 4.121167071626522e-06, 'learning_rate/action_model': 4.121167071626522e-05, 'epoch': 1.49}) +[2026-06-18 01:05:15,576][starVLA.training.train_starvla][INFO] - Step 4078, Loss: {'train/loss': 0.3576732575893402, 'train/grad_norm_pre_clip': 19.261579513549805, 'timing/data': 0.0008644331246614456, 'timing/model': 1.231920194812119, 'learning_rate/qwen_vl_interface': 4.119066456842261e-06, 'learning_rate/action_model': 4.119066456842261e-05, 'epoch': 1.49}) +[2026-06-18 01:05:28,209][starVLA.training.train_starvla][INFO] - Step 4079, Loss: {'train/loss': 0.161068856716156, 'train/grad_norm_pre_clip': 10.398021697998047, 'timing/data': 0.001366238109767437, 'timing/model': 4.3588258335366845, 'learning_rate/qwen_vl_interface': 4.1169660765017416e-06, 'learning_rate/action_model': 4.1169660765017414e-05, 'epoch': 1.49}) +[2026-06-18 01:05:35,266][starVLA.training.train_starvla][INFO] - Step 4080, Loss: {'train/loss': 0.47654399275779724, 'train/grad_norm_pre_clip': 20.096384048461914, 'timing/data': 0.0008914321660995483, 'timing/model': 1.024707855656743, 'learning_rate/qwen_vl_interface': 4.1148659310403725e-06, 'learning_rate/action_model': 4.1148659310403723e-05, 'epoch': 1.49}) +[2026-06-18 01:05:45,257][starVLA.training.train_starvla][INFO] - Step 4081, Loss: {'train/loss': 0.22421731054782867, 'train/grad_norm_pre_clip': 17.705612182617188, 'timing/data': 0.0011601978912949562, 'timing/model': 5.080573602579534, 'learning_rate/qwen_vl_interface': 4.112766020893516e-06, 'learning_rate/action_model': 4.1127660208935164e-05, 'epoch': 1.49}) +[2026-06-18 01:05:58,328][starVLA.training.train_starvla][INFO] - Step 4082, Loss: {'train/loss': 0.43549659848213196, 'train/grad_norm_pre_clip': 20.088565826416016, 'timing/data': 0.0012232502922415733, 'timing/model': 2.1798366913571954, 'learning_rate/qwen_vl_interface': 4.11066634649649e-06, 'learning_rate/action_model': 4.1106663464964895e-05, 'epoch': 1.49}) +[2026-06-18 01:06:07,407][starVLA.training.train_starvla][INFO] - Step 4083, Loss: {'train/loss': 0.545587956905365, 'train/grad_norm_pre_clip': 24.5979061126709, 'timing/data': 0.001346701756119728, 'timing/model': 4.584580201655626, 'learning_rate/qwen_vl_interface': 4.1085669082845545e-06, 'learning_rate/action_model': 4.108566908284555e-05, 'epoch': 1.5}) +[2026-06-18 01:06:18,101][starVLA.training.train_starvla][INFO] - Step 4084, Loss: {'train/loss': 0.41855427622795105, 'train/grad_norm_pre_clip': 24.80808448791504, 'timing/data': 0.0009820256382226944, 'timing/model': 1.1087163919582963, 'learning_rate/qwen_vl_interface': 4.10646770669293e-06, 'learning_rate/action_model': 4.106467706692929e-05, 'epoch': 1.5}) +[2026-06-18 01:06:22,686][starVLA.training.train_starvla][INFO] - Step 4085, Loss: {'train/loss': 0.18951945006847382, 'train/grad_norm_pre_clip': 12.657977104187012, 'timing/data': 0.0009159157052636147, 'timing/model': 1.0624712724238634, 'learning_rate/qwen_vl_interface': 4.104368742156774e-06, 'learning_rate/action_model': 4.104368742156774e-05, 'epoch': 1.5}) +[2026-06-18 01:06:27,432][starVLA.training.train_starvla][INFO] - Step 4086, Loss: {'train/loss': 0.3547697961330414, 'train/grad_norm_pre_clip': 18.241989135742188, 'timing/data': 0.0013260142877697945, 'timing/model': 1.323401354253292, 'learning_rate/qwen_vl_interface': 4.102270015111217e-06, 'learning_rate/action_model': 4.102270015111217e-05, 'epoch': 1.5}) +[2026-06-18 01:06:31,680][starVLA.training.train_starvla][INFO] - Step 4087, Loss: {'train/loss': 0.532193124294281, 'train/grad_norm_pre_clip': 21.19795036315918, 'timing/data': 0.0009013377130031586, 'timing/model': 1.0523061230778694, 'learning_rate/qwen_vl_interface': 4.100171525991318e-06, 'learning_rate/action_model': 4.100171525991318e-05, 'epoch': 1.5}) +[2026-06-18 01:06:36,670][starVLA.training.train_starvla][INFO] - Step 4088, Loss: {'train/loss': 0.3130420744419098, 'train/grad_norm_pre_clip': 17.670541763305664, 'timing/data': 0.0010835127905011177, 'timing/model': 1.1204588152468204, 'learning_rate/qwen_vl_interface': 4.098073275232098e-06, 'learning_rate/action_model': 4.098073275232098e-05, 'epoch': 1.5}) +[2026-06-18 01:06:41,215][starVLA.training.train_starvla][INFO] - Step 4089, Loss: {'train/loss': 0.5440306663513184, 'train/grad_norm_pre_clip': 18.446805953979492, 'timing/data': 0.0011198939755558968, 'timing/model': 1.0888588847592473, 'learning_rate/qwen_vl_interface': 4.09597526326853e-06, 'learning_rate/action_model': 4.095975263268529e-05, 'epoch': 1.5}) +[2026-06-18 01:06:46,193][starVLA.training.train_starvla][INFO] - Step 4090, Loss: {'train/loss': 0.23878544569015503, 'train/grad_norm_pre_clip': 15.557401657104492, 'timing/data': 0.0010070623829960823, 'timing/model': 1.0427436353638768, 'learning_rate/qwen_vl_interface': 4.09387749053553e-06, 'learning_rate/action_model': 4.0938774905355296e-05, 'epoch': 1.5}) +[2026-06-18 01:06:50,884][starVLA.training.train_starvla][INFO] - Step 4091, Loss: {'train/loss': 0.43852725625038147, 'train/grad_norm_pre_clip': 16.956422805786133, 'timing/data': 0.0013724267482757568, 'timing/model': 1.493361716158688, 'learning_rate/qwen_vl_interface': 4.0917799574679725e-06, 'learning_rate/action_model': 4.091779957467973e-05, 'epoch': 1.5}) +[2026-06-18 01:06:56,008][starVLA.training.train_starvla][INFO] - Step 4092, Loss: {'train/loss': 0.39197295904159546, 'train/grad_norm_pre_clip': 18.37083625793457, 'timing/data': 0.0011135777458548546, 'timing/model': 1.0521639753133059, 'learning_rate/qwen_vl_interface': 4.0896826645006725e-06, 'learning_rate/action_model': 4.089682664500673e-05, 'epoch': 1.5}) +[2026-06-18 01:07:01,253][starVLA.training.train_starvla][INFO] - Step 4093, Loss: {'train/loss': 0.2535855174064636, 'train/grad_norm_pre_clip': 16.342798233032227, 'timing/data': 0.002051764167845249, 'timing/model': 1.4098345469683409, 'learning_rate/qwen_vl_interface': 4.087585612068408e-06, 'learning_rate/action_model': 4.0875856120684074e-05, 'epoch': 1.5}) +[2026-06-18 01:07:12,131][starVLA.training.train_starvla][INFO] - Step 4094, Loss: {'train/loss': 0.1300705522298813, 'train/grad_norm_pre_clip': 11.434865951538086, 'timing/data': 0.0009040171280503273, 'timing/model': 1.07341437228024, 'learning_rate/qwen_vl_interface': 4.0854888006058955e-06, 'learning_rate/action_model': 4.085488800605896e-05, 'epoch': 1.5}) +[2026-06-18 01:07:17,439][starVLA.training.train_starvla][INFO] - Step 4095, Loss: {'train/loss': 0.5037359595298767, 'train/grad_norm_pre_clip': 25.421293258666992, 'timing/data': 0.003628426231443882, 'timing/model': 1.3917297758162022, 'learning_rate/qwen_vl_interface': 4.083392230547808e-06, 'learning_rate/action_model': 4.083392230547808e-05, 'epoch': 1.5}) +[2026-06-18 01:07:22,285][starVLA.training.train_starvla][INFO] - Step 4096, Loss: {'train/loss': 0.4987831115722656, 'train/grad_norm_pre_clip': 15.569958686828613, 'timing/data': 0.0009973924607038498, 'timing/model': 1.4875462809577584, 'learning_rate/qwen_vl_interface': 4.081295902328765e-06, 'learning_rate/action_model': 4.081295902328765e-05, 'epoch': 1.5}) +[2026-06-18 01:07:29,220][starVLA.training.train_starvla][INFO] - Step 4097, Loss: {'train/loss': 0.37449195981025696, 'train/grad_norm_pre_clip': 26.54262924194336, 'timing/data': 0.0008745035156607628, 'timing/model': 1.0705809202045202, 'learning_rate/qwen_vl_interface': 4.079199816383341e-06, 'learning_rate/action_model': 4.0791998163833404e-05, 'epoch': 1.5}) +[2026-06-18 01:07:33,705][starVLA.training.train_starvla][INFO] - Step 4098, Loss: {'train/loss': 0.2593962848186493, 'train/grad_norm_pre_clip': 15.472439765930176, 'timing/data': 0.0008616084232926369, 'timing/model': 1.0649549840018153, 'learning_rate/qwen_vl_interface': 4.077103973146054e-06, 'learning_rate/action_model': 4.077103973146054e-05, 'epoch': 1.5}) +[2026-06-18 01:07:38,539][starVLA.training.train_starvla][INFO] - Step 4099, Loss: {'train/loss': 0.6523178219795227, 'train/grad_norm_pre_clip': 25.325862884521484, 'timing/data': 0.0007252609357237816, 'timing/model': 1.1879219254478812, 'learning_rate/qwen_vl_interface': 4.075008373051375e-06, 'learning_rate/action_model': 4.075008373051374e-05, 'epoch': 1.5}) +[2026-06-18 01:12:09,083][starVLA.training.train_starvla][INFO] - Step 4100, Loss: {'train/loss': 0.43933436274528503, 'train/grad_norm_pre_clip': 21.072341918945312, 'eval/loss': 0.5853258967399597, 'eval/latency_2/loss': 0.5853258967399597, 'eval/demon_fire_precision': 0.914319714399048, 'eval/demon_fire_recall': 0.9644351464435147, 'eval/demon_fire_f1': 0.9387090205660761, 'eval/demon_fire_rate_model': 0.701446855870896, 'eval/demon_fire_rate_teacher': 0.6649972175848636, 'eval/demon_left_precision': 0.8843995510662177, 'eval/demon_left_recall': 0.8294736842105264, 'eval/demon_left_f1': 0.8560564910374797, 'eval/demon_left_rate_model': 0.24791318864774625, 'eval/demon_left_rate_teacher': 0.2643294379521425, 'eval/demon_right_precision': 0.828838174273859, 'eval/demon_right_recall': 0.9028248587570622, 'eval/demon_right_f1': 0.8642509464575446, 'eval/demon_right_rate_model': 0.26822481914301616, 'eval/demon_right_rate_teacher': 0.24624373956594323, 'eval/demon_move_f1': 0.8601537187475121, 'eval/demon_cc_f1': 0.9136760198570517, 'eval/latency_2/demon_fire_precision': 0.914319714399048, 'eval/latency_2/demon_fire_recall': 0.9644351464435147, 'eval/latency_2/demon_fire_f1': 0.9387090205660761, 'eval/latency_2/demon_fire_rate_model': 0.701446855870896, 'eval/latency_2/demon_fire_rate_teacher': 0.6649972175848636, 'eval/latency_2/demon_left_precision': 0.8843995510662177, 'eval/latency_2/demon_left_recall': 0.8294736842105264, 'eval/latency_2/demon_left_f1': 0.8560564910374797, 'eval/latency_2/demon_left_rate_model': 0.24791318864774625, 'eval/latency_2/demon_left_rate_teacher': 0.2643294379521425, 'eval/latency_2/demon_right_precision': 0.828838174273859, 'eval/latency_2/demon_right_recall': 0.9028248587570622, 'eval/latency_2/demon_right_f1': 0.8642509464575446, 'eval/latency_2/demon_right_rate_model': 0.26822481914301616, 'eval/latency_2/demon_right_rate_teacher': 0.24624373956594323, 'eval/latency_2/demon_move_f1': 0.8601537187475121, 'eval/latency_2/demon_cc_f1': 0.9136760198570517, 'timing/data': 0.001702924259006977, 'timing/model': 1.0614258516579866, 'learning_rate/qwen_vl_interface': 4.072913016533726e-06, 'learning_rate/action_model': 4.072913016533726e-05, 'epoch': 1.5}) +[2026-06-18 01:12:09,105][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_4100_state +[2026-06-18 01:12:09,113][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 01:12:17,786][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_4100_state/model.safetensors +[2026-06-18 01:12:32,962][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_4100_state/optimizer.bin +[2026-06-18 01:12:32,963][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_4100_state/sampler.bin +[2026-06-18 01:12:32,964][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_4100_state/sampler_1.bin +[2026-06-18 01:12:32,967][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_4100_state/random_states_0.pkl +[2026-06-18 01:12:32,971][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4100 +[2026-06-18 01:12:32,972][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 01:12:32,994][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 01:12:37,821][starVLA.training.train_starvla][INFO] - Step 4101, Loss: {'train/loss': 0.6140492558479309, 'train/grad_norm_pre_clip': 31.883907318115234, 'timing/data': 0.0012846775352954865, 'timing/model': 1.085522991605103, 'learning_rate/qwen_vl_interface': 4.070817904027474e-06, 'learning_rate/action_model': 4.070817904027474e-05, 'epoch': 1.5}) +[2026-06-18 01:12:42,953][starVLA.training.train_starvla][INFO] - Step 4102, Loss: {'train/loss': 0.6261213421821594, 'train/grad_norm_pre_clip': 23.520153045654297, 'timing/data': 0.0009060129523277283, 'timing/model': 1.3827290432527661, 'learning_rate/qwen_vl_interface': 4.06872303596694e-06, 'learning_rate/action_model': 4.06872303596694e-05, 'epoch': 1.5}) +[2026-06-18 01:12:53,670][starVLA.training.train_starvla][INFO] - Step 4103, Loss: {'train/loss': 0.38286805152893066, 'train/grad_norm_pre_clip': 17.907926559448242, 'timing/data': 0.0010895291343331337, 'timing/model': 4.451734186150134, 'learning_rate/qwen_vl_interface': 4.066628412786391e-06, 'learning_rate/action_model': 4.066628412786391e-05, 'epoch': 1.5}) +[2026-06-18 01:13:01,866][starVLA.training.train_starvla][INFO] - Step 4104, Loss: {'train/loss': 0.696171760559082, 'train/grad_norm_pre_clip': 21.66855239868164, 'timing/data': 0.0012186989188194275, 'timing/model': 1.0293427305296063, 'learning_rate/qwen_vl_interface': 4.0645340349200466e-06, 'learning_rate/action_model': 4.0645340349200464e-05, 'epoch': 1.5}) +[2026-06-18 01:13:07,116][starVLA.training.train_starvla][INFO] - Step 4105, Loss: {'train/loss': 0.4100206792354584, 'train/grad_norm_pre_clip': 21.387468338012695, 'timing/data': 0.0009374218061566353, 'timing/model': 1.42003442812711, 'learning_rate/qwen_vl_interface': 4.062439902802073e-06, 'learning_rate/action_model': 4.0624399028020725e-05, 'epoch': 1.5}) +[2026-06-18 01:13:20,282][starVLA.training.train_starvla][INFO] - Step 4106, Loss: {'train/loss': 0.7218592762947083, 'train/grad_norm_pre_clip': 25.780471801757812, 'timing/data': 0.001666051335632801, 'timing/model': 1.7301247715950012, 'learning_rate/qwen_vl_interface': 4.060346016866584e-06, 'learning_rate/action_model': 4.060346016866584e-05, 'epoch': 1.5}) +[2026-06-18 01:13:24,879][starVLA.training.train_starvla][INFO] - Step 4107, Loss: {'train/loss': 0.6927080154418945, 'train/grad_norm_pre_clip': 29.68448257446289, 'timing/data': 0.0010897424072027206, 'timing/model': 1.1282775001600385, 'learning_rate/qwen_vl_interface': 4.058252377547645e-06, 'learning_rate/action_model': 4.058252377547645e-05, 'epoch': 1.5}) +[2026-06-18 01:13:29,965][starVLA.training.train_starvla][INFO] - Step 4108, Loss: {'train/loss': 0.4161136746406555, 'train/grad_norm_pre_clip': 22.70669937133789, 'timing/data': 0.0008848151192069054, 'timing/model': 1.392477392219007, 'learning_rate/qwen_vl_interface': 4.056158985279272e-06, 'learning_rate/action_model': 4.0561589852792715e-05, 'epoch': 1.5}) +[2026-06-18 01:13:34,921][starVLA.training.train_starvla][INFO] - Step 4109, Loss: {'train/loss': 0.10510408133268356, 'train/grad_norm_pre_clip': 8.201327323913574, 'timing/data': 0.000862867571413517, 'timing/model': 1.0575837017968297, 'learning_rate/qwen_vl_interface': 4.054065840495425e-06, 'learning_rate/action_model': 4.054065840495425e-05, 'epoch': 1.5}) +[2026-06-18 01:13:43,936][starVLA.training.train_starvla][INFO] - Step 4110, Loss: {'train/loss': 0.4925835430622101, 'train/grad_norm_pre_clip': 30.762540817260742, 'timing/data': 0.0010578623041510582, 'timing/model': 1.2511018067598343, 'learning_rate/qwen_vl_interface': 4.0519729436300165e-06, 'learning_rate/action_model': 4.051972943630016e-05, 'epoch': 1.5}) +[2026-06-18 01:13:48,524][starVLA.training.train_starvla][INFO] - Step 4111, Loss: {'train/loss': 0.5718733668327332, 'train/grad_norm_pre_clip': 21.50701332092285, 'timing/data': 0.0015064384788274765, 'timing/model': 1.3855200549587607, 'learning_rate/qwen_vl_interface': 4.049880295116906e-06, 'learning_rate/action_model': 4.049880295116906e-05, 'epoch': 1.51}) +[2026-06-18 01:13:53,052][starVLA.training.train_starvla][INFO] - Step 4112, Loss: {'train/loss': 0.3375004827976227, 'train/grad_norm_pre_clip': 15.449993133544922, 'timing/data': 0.0009205611422657967, 'timing/model': 1.045687790028751, 'learning_rate/qwen_vl_interface': 4.047787895389902e-06, 'learning_rate/action_model': 4.0477878953899016e-05, 'epoch': 1.51}) +[2026-06-18 01:14:01,290][starVLA.training.train_starvla][INFO] - Step 4113, Loss: {'train/loss': 0.6258141994476318, 'train/grad_norm_pre_clip': 26.236379623413086, 'timing/data': 0.0012878905981779099, 'timing/model': 1.4430034160614014, 'learning_rate/qwen_vl_interface': 4.0456957448827615e-06, 'learning_rate/action_model': 4.045695744882761e-05, 'epoch': 1.51}) +[2026-06-18 01:14:06,291][starVLA.training.train_starvla][INFO] - Step 4114, Loss: {'train/loss': 0.37231528759002686, 'train/grad_norm_pre_clip': 18.818771362304688, 'timing/data': 0.0011099763214588165, 'timing/model': 1.0786033449694514, 'learning_rate/qwen_vl_interface': 4.043603844029188e-06, 'learning_rate/action_model': 4.0436038440291874e-05, 'epoch': 1.51}) +[2026-06-18 01:14:11,502][starVLA.training.train_starvla][INFO] - Step 4115, Loss: {'train/loss': 0.5214719772338867, 'train/grad_norm_pre_clip': 17.30119514465332, 'timing/data': 0.0014192434027791023, 'timing/model': 1.0913260728120804, 'learning_rate/qwen_vl_interface': 4.041512193262836e-06, 'learning_rate/action_model': 4.041512193262836e-05, 'epoch': 1.51}) +[2026-06-18 01:14:16,176][starVLA.training.train_starvla][INFO] - Step 4116, Loss: {'train/loss': 0.48187193274497986, 'train/grad_norm_pre_clip': 18.80261993408203, 'timing/data': 0.0010790154337882996, 'timing/model': 1.1219541747123003, 'learning_rate/qwen_vl_interface': 4.039420793017306e-06, 'learning_rate/action_model': 4.039420793017306e-05, 'epoch': 1.51}) +[2026-06-18 01:14:20,827][starVLA.training.train_starvla][INFO] - Step 4117, Loss: {'train/loss': 0.4155627191066742, 'train/grad_norm_pre_clip': 23.151884078979492, 'timing/data': 0.0019501056522130966, 'timing/model': 1.4011797169223428, 'learning_rate/qwen_vl_interface': 4.03732964372615e-06, 'learning_rate/action_model': 4.037329643726149e-05, 'epoch': 1.51}) +[2026-06-18 01:14:25,010][starVLA.training.train_starvla][INFO] - Step 4118, Loss: {'train/loss': 0.35680270195007324, 'train/grad_norm_pre_clip': 18.749418258666992, 'timing/data': 0.0010364549234509468, 'timing/model': 1.0666048265993595, 'learning_rate/qwen_vl_interface': 4.035238745822861e-06, 'learning_rate/action_model': 4.035238745822861e-05, 'epoch': 1.51}) +[2026-06-18 01:14:33,891][starVLA.training.train_starvla][INFO] - Step 4119, Loss: {'train/loss': 0.26103881001472473, 'train/grad_norm_pre_clip': 14.462563514709473, 'timing/data': 0.0013152975589036942, 'timing/model': 1.0402773646637797, 'learning_rate/qwen_vl_interface': 4.033148099740889e-06, 'learning_rate/action_model': 4.0331480997408885e-05, 'epoch': 1.51}) +[2026-06-18 01:14:38,612][starVLA.training.train_starvla][INFO] - Step 4120, Loss: {'train/loss': 0.4573138654232025, 'train/grad_norm_pre_clip': 22.079675674438477, 'timing/data': 0.0011335024610161781, 'timing/model': 1.086476355791092, 'learning_rate/qwen_vl_interface': 4.031057705913626e-06, 'learning_rate/action_model': 4.031057705913626e-05, 'epoch': 1.51}) +[2026-06-18 01:14:43,873][starVLA.training.train_starvla][INFO] - Step 4121, Loss: {'train/loss': 0.5688793659210205, 'train/grad_norm_pre_clip': 17.931039810180664, 'timing/data': 0.001209024339914322, 'timing/model': 1.0764878066256642, 'learning_rate/qwen_vl_interface': 4.028967564774411e-06, 'learning_rate/action_model': 4.0289675647744106e-05, 'epoch': 1.51}) +[2026-06-18 01:14:48,909][starVLA.training.train_starvla][INFO] - Step 4122, Loss: {'train/loss': 0.5222000479698181, 'train/grad_norm_pre_clip': 21.96371078491211, 'timing/data': 0.0012161852791905403, 'timing/model': 1.5118416510522366, 'learning_rate/qwen_vl_interface': 4.026877676756534e-06, 'learning_rate/action_model': 4.026877676756534e-05, 'epoch': 1.51}) +[2026-06-18 01:14:54,113][starVLA.training.train_starvla][INFO] - Step 4123, Loss: {'train/loss': 0.45256179571151733, 'train/grad_norm_pre_clip': 21.044368743896484, 'timing/data': 0.0013071000576019287, 'timing/model': 1.902153579518199, 'learning_rate/qwen_vl_interface': 4.0247880422932305e-06, 'learning_rate/action_model': 4.02478804229323e-05, 'epoch': 1.51}) +[2026-06-18 01:14:58,850][starVLA.training.train_starvla][INFO] - Step 4124, Loss: {'train/loss': 0.22685925662517548, 'train/grad_norm_pre_clip': 11.583698272705078, 'timing/data': 0.0023202644661068916, 'timing/model': 1.1360778734087944, 'learning_rate/qwen_vl_interface': 4.0226986618176855e-06, 'learning_rate/action_model': 4.022698661817686e-05, 'epoch': 1.51}) +[2026-06-18 01:15:05,193][starVLA.training.train_starvla][INFO] - Step 4125, Loss: {'train/loss': 0.41481339931488037, 'train/grad_norm_pre_clip': 27.379968643188477, 'timing/data': 0.0009895404800772667, 'timing/model': 1.0442482130602002, 'learning_rate/qwen_vl_interface': 4.0206095357630256e-06, 'learning_rate/action_model': 4.020609535763026e-05, 'epoch': 1.51}) +[2026-06-18 01:15:10,280][starVLA.training.train_starvla][INFO] - Step 4126, Loss: {'train/loss': 0.3443279266357422, 'train/grad_norm_pre_clip': 25.400312423706055, 'timing/data': 0.0015535103157162666, 'timing/model': 1.215076089836657, 'learning_rate/qwen_vl_interface': 4.018520664562334e-06, 'learning_rate/action_model': 4.018520664562334e-05, 'epoch': 1.51}) +[2026-06-18 01:15:14,645][starVLA.training.train_starvla][INFO] - Step 4127, Loss: {'train/loss': 0.43319210410118103, 'train/grad_norm_pre_clip': 18.49901008605957, 'timing/data': 0.0014769462868571281, 'timing/model': 1.3200597437098622, 'learning_rate/qwen_vl_interface': 4.016432048648635e-06, 'learning_rate/action_model': 4.016432048648635e-05, 'epoch': 1.51}) +[2026-06-18 01:15:25,354][starVLA.training.train_starvla][INFO] - Step 4128, Loss: {'train/loss': 0.6307844519615173, 'train/grad_norm_pre_clip': 20.349720001220703, 'timing/data': 0.0008535496890544891, 'timing/model': 4.195621241815388, 'learning_rate/qwen_vl_interface': 4.014343688454898e-06, 'learning_rate/action_model': 4.0143436884548976e-05, 'epoch': 1.51}) +[2026-06-18 01:15:30,199][starVLA.training.train_starvla][INFO] - Step 4129, Loss: {'train/loss': 0.4927951693534851, 'train/grad_norm_pre_clip': 25.405860900878906, 'timing/data': 0.0009397128596901894, 'timing/model': 1.1912606982514262, 'learning_rate/qwen_vl_interface': 4.012255584414044e-06, 'learning_rate/action_model': 4.0122555844140436e-05, 'epoch': 1.51}) +[2026-06-18 01:15:34,549][starVLA.training.train_starvla][INFO] - Step 4130, Loss: {'train/loss': 0.34786391258239746, 'train/grad_norm_pre_clip': 16.852039337158203, 'timing/data': 0.0010913722217082977, 'timing/model': 1.0730987265706062, 'learning_rate/qwen_vl_interface': 4.010167736958939e-06, 'learning_rate/action_model': 4.010167736958938e-05, 'epoch': 1.51}) +[2026-06-18 01:15:38,777][starVLA.training.train_starvla][INFO] - Step 4131, Loss: {'train/loss': 0.5576233863830566, 'train/grad_norm_pre_clip': 21.559934616088867, 'timing/data': 0.0009686043485999107, 'timing/model': 1.010640462860465, 'learning_rate/qwen_vl_interface': 4.008080146522397e-06, 'learning_rate/action_model': 4.008080146522396e-05, 'epoch': 1.51}) +[2026-06-18 01:15:43,078][starVLA.training.train_starvla][INFO] - Step 4132, Loss: {'train/loss': 0.2605331838130951, 'train/grad_norm_pre_clip': 17.92105484008789, 'timing/data': 0.0011401940137147903, 'timing/model': 1.3016046723350883, 'learning_rate/qwen_vl_interface': 4.005992813537176e-06, 'learning_rate/action_model': 4.005992813537176e-05, 'epoch': 1.51}) +[2026-06-18 01:15:47,136][starVLA.training.train_starvla][INFO] - Step 4133, Loss: {'train/loss': 0.5001740455627441, 'train/grad_norm_pre_clip': 24.369863510131836, 'timing/data': 0.0010456852614879608, 'timing/model': 1.0375765273347497, 'learning_rate/qwen_vl_interface': 4.003905738435984e-06, 'learning_rate/action_model': 4.003905738435984e-05, 'epoch': 1.51}) +[2026-06-18 01:15:52,785][starVLA.training.train_starvla][INFO] - Step 4134, Loss: {'train/loss': 0.44321390986442566, 'train/grad_norm_pre_clip': 17.41071891784668, 'timing/data': 0.001627642661333084, 'timing/model': 1.0211038775742054, 'learning_rate/qwen_vl_interface': 4.001818921651475e-06, 'learning_rate/action_model': 4.0018189216514746e-05, 'epoch': 1.51}) +[2026-06-18 01:15:57,034][starVLA.training.train_starvla][INFO] - Step 4135, Loss: {'train/loss': 0.5064989328384399, 'train/grad_norm_pre_clip': 17.872861862182617, 'timing/data': 0.0011536134406924248, 'timing/model': 1.0113012744113803, 'learning_rate/qwen_vl_interface': 3.999732363616244e-06, 'learning_rate/action_model': 3.999732363616244e-05, 'epoch': 1.51}) +[2026-06-18 01:16:01,252][starVLA.training.train_starvla][INFO] - Step 4136, Loss: {'train/loss': 0.4601704478263855, 'train/grad_norm_pre_clip': 27.565099716186523, 'timing/data': 0.0011331690475344658, 'timing/model': 1.0230499161407351, 'learning_rate/qwen_vl_interface': 3.9976460647628394e-06, 'learning_rate/action_model': 3.997646064762839e-05, 'epoch': 1.51}) +[2026-06-18 01:16:05,535][starVLA.training.train_starvla][INFO] - Step 4137, Loss: {'train/loss': 0.4734407067298889, 'train/grad_norm_pre_clip': 18.296110153198242, 'timing/data': 0.0009470721706748009, 'timing/model': 1.262229710817337, 'learning_rate/qwen_vl_interface': 3.995560025523755e-06, 'learning_rate/action_model': 3.995560025523754e-05, 'epoch': 1.51}) +[2026-06-18 01:16:09,669][starVLA.training.train_starvla][INFO] - Step 4138, Loss: {'train/loss': 0.7099993228912354, 'train/grad_norm_pre_clip': 33.97693634033203, 'timing/data': 0.0008069602772593498, 'timing/model': 1.0169275384396315, 'learning_rate/qwen_vl_interface': 3.993474246331426e-06, 'learning_rate/action_model': 3.993474246331426e-05, 'epoch': 1.52}) +[2026-06-18 01:16:13,921][starVLA.training.train_starvla][INFO] - Step 4139, Loss: {'train/loss': 0.5924645662307739, 'train/grad_norm_pre_clip': 24.35235023498535, 'timing/data': 0.009073881432414055, 'timing/model': 1.0065530808642507, 'learning_rate/qwen_vl_interface': 3.991388727618237e-06, 'learning_rate/action_model': 3.991388727618237e-05, 'epoch': 1.52}) +[2026-06-18 01:16:19,515][starVLA.training.train_starvla][INFO] - Step 4140, Loss: {'train/loss': 0.4269322454929352, 'train/grad_norm_pre_clip': 18.301528930664062, 'timing/data': 0.0006924644112586975, 'timing/model': 1.0982876056805253, 'learning_rate/qwen_vl_interface': 3.989303469816521e-06, 'learning_rate/action_model': 3.9893034698165205e-05, 'epoch': 1.52}) +[2026-06-18 01:16:23,705][starVLA.training.train_starvla][INFO] - Step 4141, Loss: {'train/loss': 0.3478686213493347, 'train/grad_norm_pre_clip': 15.96515941619873, 'timing/data': 0.0008869804441928864, 'timing/model': 1.0423891451209784, 'learning_rate/qwen_vl_interface': 3.9872184733585526e-06, 'learning_rate/action_model': 3.987218473358552e-05, 'epoch': 1.52}) +[2026-06-18 01:16:27,817][starVLA.training.train_starvla][INFO] - Step 4142, Loss: {'train/loss': 0.4412437081336975, 'train/grad_norm_pre_clip': 21.18807601928711, 'timing/data': 0.0009706765413284302, 'timing/model': 1.0778885846957564, 'learning_rate/qwen_vl_interface': 3.985133738676552e-06, 'learning_rate/action_model': 3.985133738676552e-05, 'epoch': 1.52}) +[2026-06-18 01:16:32,332][starVLA.training.train_starvla][INFO] - Step 4143, Loss: {'train/loss': 0.2425374537706375, 'train/grad_norm_pre_clip': 15.106895446777344, 'timing/data': 0.0012754136696457863, 'timing/model': 1.1812118394300342, 'learning_rate/qwen_vl_interface': 3.983049266202688e-06, 'learning_rate/action_model': 3.983049266202687e-05, 'epoch': 1.52}) +[2026-06-18 01:16:42,875][starVLA.training.train_starvla][INFO] - Step 4144, Loss: {'train/loss': 0.536830723285675, 'train/grad_norm_pre_clip': 19.094051361083984, 'timing/data': 0.0017872266471385956, 'timing/model': 1.0452497834339738, 'learning_rate/qwen_vl_interface': 3.980965056369075e-06, 'learning_rate/action_model': 3.980965056369075e-05, 'epoch': 1.52}) +[2026-06-18 01:16:47,179][starVLA.training.train_starvla][INFO] - Step 4145, Loss: {'train/loss': 0.42232751846313477, 'train/grad_norm_pre_clip': 15.562376022338867, 'timing/data': 0.0011167051270604134, 'timing/model': 1.0543309645727277, 'learning_rate/qwen_vl_interface': 3.978881109607772e-06, 'learning_rate/action_model': 3.978881109607772e-05, 'epoch': 1.52}) +[2026-06-18 01:16:51,496][starVLA.training.train_starvla][INFO] - Step 4146, Loss: {'train/loss': 0.6398690938949585, 'train/grad_norm_pre_clip': 25.48919105529785, 'timing/data': 0.0009794728830456734, 'timing/model': 1.246111554093659, 'learning_rate/qwen_vl_interface': 3.976797426350783e-06, 'learning_rate/action_model': 3.976797426350783e-05, 'epoch': 1.52}) +[2026-06-18 01:16:55,436][starVLA.training.train_starvla][INFO] - Step 4147, Loss: {'train/loss': 0.4052361249923706, 'train/grad_norm_pre_clip': 14.489336967468262, 'timing/data': 0.0009044688194990158, 'timing/model': 1.0030063018202782, 'learning_rate/qwen_vl_interface': 3.974714007030055e-06, 'learning_rate/action_model': 3.974714007030056e-05, 'epoch': 1.52}) +[2026-06-18 01:16:59,642][starVLA.training.train_starvla][INFO] - Step 4148, Loss: {'train/loss': 0.5448622703552246, 'train/grad_norm_pre_clip': 20.10822105407715, 'timing/data': 0.0010307924821972847, 'timing/model': 1.0137160774320364, 'learning_rate/qwen_vl_interface': 3.972630852077488e-06, 'learning_rate/action_model': 3.972630852077488e-05, 'epoch': 1.52}) +[2026-06-18 01:17:04,272][starVLA.training.train_starvla][INFO] - Step 4149, Loss: {'train/loss': 0.4589206874370575, 'train/grad_norm_pre_clip': 17.85983657836914, 'timing/data': 0.0010019689798355103, 'timing/model': 1.4474006360396743, 'learning_rate/qwen_vl_interface': 3.9705479619249195e-06, 'learning_rate/action_model': 3.9705479619249197e-05, 'epoch': 1.52}) +[2026-06-18 01:17:08,643][starVLA.training.train_starvla][INFO] - Step 4150, Loss: {'train/loss': 0.3015475273132324, 'train/grad_norm_pre_clip': 11.919160842895508, 'timing/data': 0.0009261611849069595, 'timing/model': 1.1492338962852955, 'learning_rate/qwen_vl_interface': 3.9684653370041345e-06, 'learning_rate/action_model': 3.968465337004134e-05, 'epoch': 1.52}) +[2026-06-18 01:17:13,847][starVLA.training.train_starvla][INFO] - Step 4151, Loss: {'train/loss': 0.5064772367477417, 'train/grad_norm_pre_clip': 19.37200355529785, 'timing/data': 0.0008967174217104912, 'timing/model': 1.005575024522841, 'learning_rate/qwen_vl_interface': 3.966382977746866e-06, 'learning_rate/action_model': 3.9663829777468656e-05, 'epoch': 1.52}) +[2026-06-18 01:17:18,109][starVLA.training.train_starvla][INFO] - Step 4152, Loss: {'train/loss': 0.35553377866744995, 'train/grad_norm_pre_clip': 50.87156295776367, 'timing/data': 0.0013131406158208847, 'timing/model': 1.289650596678257, 'learning_rate/qwen_vl_interface': 3.9643008845847845e-06, 'learning_rate/action_model': 3.964300884584785e-05, 'epoch': 1.52}) +[2026-06-18 01:17:22,135][starVLA.training.train_starvla][INFO] - Step 4153, Loss: {'train/loss': 0.3353901505470276, 'train/grad_norm_pre_clip': 15.821540832519531, 'timing/data': 0.002007211558520794, 'timing/model': 1.0314972279593349, 'learning_rate/qwen_vl_interface': 3.962219057949517e-06, 'learning_rate/action_model': 3.9622190579495166e-05, 'epoch': 1.52}) +[2026-06-18 01:17:26,260][starVLA.training.train_starvla][INFO] - Step 4154, Loss: {'train/loss': 0.1653052121400833, 'train/grad_norm_pre_clip': 12.891117095947266, 'timing/data': 0.0011218506842851639, 'timing/model': 1.0241362787783146, 'learning_rate/qwen_vl_interface': 3.960137498272619e-06, 'learning_rate/action_model': 3.9601374982726194e-05, 'epoch': 1.52}) +[2026-06-18 01:17:30,433][starVLA.training.train_starvla][INFO] - Step 4155, Loss: {'train/loss': 0.6595286130905151, 'train/grad_norm_pre_clip': 21.09210777282715, 'timing/data': 0.0009537627920508385, 'timing/model': 1.0106985596939921, 'learning_rate/qwen_vl_interface': 3.958056205985611e-06, 'learning_rate/action_model': 3.95805620598561e-05, 'epoch': 1.52}) +[2026-06-18 01:17:41,021][starVLA.training.train_starvla][INFO] - Step 4156, Loss: {'train/loss': 0.47523462772369385, 'train/grad_norm_pre_clip': 19.39124298095703, 'timing/data': 0.001299222931265831, 'timing/model': 1.242074224166572, 'learning_rate/qwen_vl_interface': 3.95597518151994e-06, 'learning_rate/action_model': 3.955975181519939e-05, 'epoch': 1.52}) +[2026-06-18 01:17:47,022][starVLA.training.train_starvla][INFO] - Step 4157, Loss: {'train/loss': 0.37780115008354187, 'train/grad_norm_pre_clip': 17.37302589416504, 'timing/data': 0.0014415653422474861, 'timing/model': 1.3753768699243665, 'learning_rate/qwen_vl_interface': 3.953894425307005e-06, 'learning_rate/action_model': 3.9538944253070055e-05, 'epoch': 1.52}) +[2026-06-18 01:17:53,246][starVLA.training.train_starvla][INFO] - Step 4158, Loss: {'train/loss': 0.3407362103462219, 'train/grad_norm_pre_clip': 19.64162254333496, 'timing/data': 0.0014739083126187325, 'timing/model': 1.339535303413868, 'learning_rate/qwen_vl_interface': 3.951813937778151e-06, 'learning_rate/action_model': 3.95181393777815e-05, 'epoch': 1.52}) +[2026-06-18 01:17:58,649][starVLA.training.train_starvla][INFO] - Step 4159, Loss: {'train/loss': 0.17751270532608032, 'train/grad_norm_pre_clip': 8.9606351852417, 'timing/data': 0.0012524165213108063, 'timing/model': 1.464090690948069, 'learning_rate/qwen_vl_interface': 3.949733719364664e-06, 'learning_rate/action_model': 3.949733719364664e-05, 'epoch': 1.52}) +[2026-06-18 01:18:04,114][starVLA.training.train_starvla][INFO] - Step 4160, Loss: {'train/loss': 0.6811956763267517, 'train/grad_norm_pre_clip': 25.79538917541504, 'timing/data': 0.0010870127007365227, 'timing/model': 1.0616331845521927, 'learning_rate/qwen_vl_interface': 3.947653770497778e-06, 'learning_rate/action_model': 3.947653770497778e-05, 'epoch': 1.52}) +[2026-06-18 01:18:15,214][starVLA.training.train_starvla][INFO] - Step 4161, Loss: {'train/loss': 0.2667223811149597, 'train/grad_norm_pre_clip': 20.597692489624023, 'timing/data': 0.0009376909583806992, 'timing/model': 5.80772412661463, 'learning_rate/qwen_vl_interface': 3.945574091608664e-06, 'learning_rate/action_model': 3.9455740916086636e-05, 'epoch': 1.52}) +[2026-06-18 01:18:29,517][starVLA.training.train_starvla][INFO] - Step 4162, Loss: {'train/loss': 0.18537265062332153, 'train/grad_norm_pre_clip': 12.453646659851074, 'timing/data': 0.0011209165677428246, 'timing/model': 2.0452771158888936, 'learning_rate/qwen_vl_interface': 3.9434946831284465e-06, 'learning_rate/action_model': 3.943494683128447e-05, 'epoch': 1.52}) +[2026-06-18 01:18:43,630][starVLA.training.train_starvla][INFO] - Step 4163, Loss: {'train/loss': 0.59355229139328, 'train/grad_norm_pre_clip': 21.01783561706543, 'timing/data': 0.040371036157011986, 'timing/model': 2.4246798902750015, 'learning_rate/qwen_vl_interface': 3.941415545488187e-06, 'learning_rate/action_model': 3.941415545488187e-05, 'epoch': 1.52}) +[2026-06-18 01:18:52,651][starVLA.training.train_starvla][INFO] - Step 4164, Loss: {'train/loss': 0.4285721480846405, 'train/grad_norm_pre_clip': 16.999980926513672, 'timing/data': 0.0012123174965381622, 'timing/model': 1.3111399542540312, 'learning_rate/qwen_vl_interface': 3.939336679118892e-06, 'learning_rate/action_model': 3.939336679118892e-05, 'epoch': 1.52}) +[2026-06-18 01:18:57,636][starVLA.training.train_starvla][INFO] - Step 4165, Loss: {'train/loss': 0.4684790074825287, 'train/grad_norm_pre_clip': 19.188831329345703, 'timing/data': 0.0010285275056958199, 'timing/model': 1.148836389183998, 'learning_rate/qwen_vl_interface': 3.937258084451514e-06, 'learning_rate/action_model': 3.937258084451514e-05, 'epoch': 1.53}) +[2026-06-18 01:19:06,030][starVLA.training.train_starvla][INFO] - Step 4166, Loss: {'train/loss': 0.232791006565094, 'train/grad_norm_pre_clip': 13.403943061828613, 'timing/data': 0.0016971500590443611, 'timing/model': 1.0861164648085833, 'learning_rate/qwen_vl_interface': 3.9351797619169485e-06, 'learning_rate/action_model': 3.935179761916949e-05, 'epoch': 1.53}) +[2026-06-18 01:19:11,780][starVLA.training.train_starvla][INFO] - Step 4167, Loss: {'train/loss': 0.20230789482593536, 'train/grad_norm_pre_clip': 14.012904167175293, 'timing/data': 0.0013266978785395622, 'timing/model': 1.907251168973744, 'learning_rate/qwen_vl_interface': 3.933101711946034e-06, 'learning_rate/action_model': 3.9331017119460336e-05, 'epoch': 1.53}) +[2026-06-18 01:19:27,101][starVLA.training.train_starvla][INFO] - Step 4168, Loss: {'train/loss': 0.5053473114967346, 'train/grad_norm_pre_clip': 14.42630386352539, 'timing/data': 0.0010929815471172333, 'timing/model': 1.3152168123051524, 'learning_rate/qwen_vl_interface': 3.93102393496955e-06, 'learning_rate/action_model': 3.93102393496955e-05, 'epoch': 1.53}) +[2026-06-18 01:19:38,807][starVLA.training.train_starvla][INFO] - Step 4169, Loss: {'train/loss': 0.3358018398284912, 'train/grad_norm_pre_clip': 19.453353881835938, 'timing/data': 0.0016836877912282944, 'timing/model': 2.6968670981004834, 'learning_rate/qwen_vl_interface': 3.928946431418223e-06, 'learning_rate/action_model': 3.9289464314182225e-05, 'epoch': 1.53}) +[2026-06-18 01:19:48,578][starVLA.training.train_starvla][INFO] - Step 4170, Loss: {'train/loss': 0.39835190773010254, 'train/grad_norm_pre_clip': 22.94133949279785, 'timing/data': 0.0013605840504169464, 'timing/model': 2.6284594545140862, 'learning_rate/qwen_vl_interface': 3.926869201722723e-06, 'learning_rate/action_model': 3.926869201722723e-05, 'epoch': 1.53}) +[2026-06-18 01:19:53,294][starVLA.training.train_starvla][INFO] - Step 4171, Loss: {'train/loss': 0.47084176540374756, 'train/grad_norm_pre_clip': 18.8974666595459, 'timing/data': 0.001222231425344944, 'timing/model': 1.115043317899108, 'learning_rate/qwen_vl_interface': 3.9247922463136615e-06, 'learning_rate/action_model': 3.924792246313662e-05, 'epoch': 1.53}) +[2026-06-18 01:19:57,851][starVLA.training.train_starvla][INFO] - Step 4172, Loss: {'train/loss': 0.4134369194507599, 'train/grad_norm_pre_clip': 21.595388412475586, 'timing/data': 0.001187318004667759, 'timing/model': 1.3884017821401358, 'learning_rate/qwen_vl_interface': 3.922715565621593e-06, 'learning_rate/action_model': 3.9227155656215926e-05, 'epoch': 1.53}) +[2026-06-18 01:20:02,808][starVLA.training.train_starvla][INFO] - Step 4173, Loss: {'train/loss': 0.24685560166835785, 'train/grad_norm_pre_clip': 15.587281227111816, 'timing/data': 0.000973348505795002, 'timing/model': 1.0903326757252216, 'learning_rate/qwen_vl_interface': 3.9206391600770164e-06, 'learning_rate/action_model': 3.920639160077016e-05, 'epoch': 1.53}) +[2026-06-18 01:20:07,727][starVLA.training.train_starvla][INFO] - Step 4174, Loss: {'train/loss': 0.311313271522522, 'train/grad_norm_pre_clip': 16.339208602905273, 'timing/data': 0.0012995684519410133, 'timing/model': 1.3796292524784803, 'learning_rate/qwen_vl_interface': 3.918563030110373e-06, 'learning_rate/action_model': 3.918563030110373e-05, 'epoch': 1.53}) +[2026-06-18 01:20:12,203][starVLA.training.train_starvla][INFO] - Step 4175, Loss: {'train/loss': 0.28077343106269836, 'train/grad_norm_pre_clip': 23.005451202392578, 'timing/data': 0.0011354731395840645, 'timing/model': 1.0342938331887126, 'learning_rate/qwen_vl_interface': 3.916487176152045e-06, 'learning_rate/action_model': 3.916487176152045e-05, 'epoch': 1.53}) +[2026-06-18 01:20:16,817][starVLA.training.train_starvla][INFO] - Step 4176, Loss: {'train/loss': 0.30139458179473877, 'train/grad_norm_pre_clip': 17.68402099609375, 'timing/data': 0.0009875772520899773, 'timing/model': 1.1207492519170046, 'learning_rate/qwen_vl_interface': 3.914411598632359e-06, 'learning_rate/action_model': 3.914411598632359e-05, 'epoch': 1.53}) +[2026-06-18 01:20:21,718][starVLA.training.train_starvla][INFO] - Step 4177, Loss: {'train/loss': 0.2188430279493332, 'train/grad_norm_pre_clip': 11.77446460723877, 'timing/data': 0.0009051728993654251, 'timing/model': 1.0562825920060277, 'learning_rate/qwen_vl_interface': 3.912336297981588e-06, 'learning_rate/action_model': 3.912336297981588e-05, 'epoch': 1.53}) +[2026-06-18 01:20:26,497][starVLA.training.train_starvla][INFO] - Step 4178, Loss: {'train/loss': 0.7356853485107422, 'train/grad_norm_pre_clip': 43.149478912353516, 'timing/data': 0.0010374980047345161, 'timing/model': 1.1527357269078493, 'learning_rate/qwen_vl_interface': 3.910261274629941e-06, 'learning_rate/action_model': 3.910261274629941e-05, 'epoch': 1.53}) +[2026-06-18 01:20:31,492][starVLA.training.train_starvla][INFO] - Step 4179, Loss: {'train/loss': 0.4140259921550751, 'train/grad_norm_pre_clip': 24.182090759277344, 'timing/data': 0.0012243576347827911, 'timing/model': 1.0676223123446107, 'learning_rate/qwen_vl_interface': 3.908186529007576e-06, 'learning_rate/action_model': 3.908186529007575e-05, 'epoch': 1.53}) +[2026-06-18 01:20:36,048][starVLA.training.train_starvla][INFO] - Step 4180, Loss: {'train/loss': 0.49070411920547485, 'train/grad_norm_pre_clip': 23.966154098510742, 'timing/data': 0.0012783138081431389, 'timing/model': 1.0821527177467942, 'learning_rate/qwen_vl_interface': 3.906112061544582e-06, 'learning_rate/action_model': 3.906112061544582e-05, 'epoch': 1.53}) +[2026-06-18 01:20:41,309][starVLA.training.train_starvla][INFO] - Step 4181, Loss: {'train/loss': 0.2724776268005371, 'train/grad_norm_pre_clip': 13.953213691711426, 'timing/data': 0.0009361105039715767, 'timing/model': 1.0545677188783884, 'learning_rate/qwen_vl_interface': 3.904037872671008e-06, 'learning_rate/action_model': 3.904037872671008e-05, 'epoch': 1.53}) +[2026-06-18 01:20:46,105][starVLA.training.train_starvla][INFO] - Step 4182, Loss: {'train/loss': 0.7042051553726196, 'train/grad_norm_pre_clip': 26.80641746520996, 'timing/data': 0.0008970918133854866, 'timing/model': 1.4765405422076583, 'learning_rate/qwen_vl_interface': 3.90196396281683e-06, 'learning_rate/action_model': 3.9019639628168295e-05, 'epoch': 1.53}) +[2026-06-18 01:20:50,952][starVLA.training.train_starvla][INFO] - Step 4183, Loss: {'train/loss': 0.35104602575302124, 'train/grad_norm_pre_clip': 20.380779266357422, 'timing/data': 0.0013137198984622955, 'timing/model': 1.3073521330952644, 'learning_rate/qwen_vl_interface': 3.899890332411972e-06, 'learning_rate/action_model': 3.8998903324119724e-05, 'epoch': 1.53}) +[2026-06-18 01:20:56,271][starVLA.training.train_starvla][INFO] - Step 4184, Loss: {'train/loss': 0.4747002422809601, 'train/grad_norm_pre_clip': 24.652982711791992, 'timing/data': 0.0011328263208270073, 'timing/model': 1.407268832437694, 'learning_rate/qwen_vl_interface': 3.897816981886302e-06, 'learning_rate/action_model': 3.897816981886302e-05, 'epoch': 1.53}) +[2026-06-18 01:21:01,497][starVLA.training.train_starvla][INFO] - Step 4185, Loss: {'train/loss': 0.33293795585632324, 'train/grad_norm_pre_clip': 18.757110595703125, 'timing/data': 0.0007937336340546608, 'timing/model': 1.3367545995861292, 'learning_rate/qwen_vl_interface': 3.895743911669627e-06, 'learning_rate/action_model': 3.895743911669627e-05, 'epoch': 1.53}) +[2026-06-18 01:21:13,426][starVLA.training.train_starvla][INFO] - Step 4186, Loss: {'train/loss': 0.5174600481987, 'train/grad_norm_pre_clip': 19.120901107788086, 'timing/data': 0.0010171588510274887, 'timing/model': 4.734370935708284, 'learning_rate/qwen_vl_interface': 3.893671122191698e-06, 'learning_rate/action_model': 3.8936711221916974e-05, 'epoch': 1.53}) +[2026-06-18 01:21:18,813][starVLA.training.train_starvla][INFO] - Step 4187, Loss: {'train/loss': 0.15598636865615845, 'train/grad_norm_pre_clip': 9.324748992919922, 'timing/data': 0.001924949698150158, 'timing/model': 1.746453712694347, 'learning_rate/qwen_vl_interface': 3.8915986138822e-06, 'learning_rate/action_model': 3.8915986138822e-05, 'epoch': 1.53}) +[2026-06-18 01:21:27,822][starVLA.training.train_starvla][INFO] - Step 4188, Loss: {'train/loss': 0.08891865611076355, 'train/grad_norm_pre_clip': 6.733419895172119, 'timing/data': 0.002283131703734398, 'timing/model': 1.7443507062271237, 'learning_rate/qwen_vl_interface': 3.889526387170775e-06, 'learning_rate/action_model': 3.889526387170775e-05, 'epoch': 1.53}) +[2026-06-18 01:21:39,117][starVLA.training.train_starvla][INFO] - Step 4189, Loss: {'train/loss': 0.7205660343170166, 'train/grad_norm_pre_clip': 31.92811393737793, 'timing/data': 0.0016895439475774765, 'timing/model': 1.7429746482521296, 'learning_rate/qwen_vl_interface': 3.8874544424869924e-06, 'learning_rate/action_model': 3.8874544424869926e-05, 'epoch': 1.53}) +[2026-06-18 01:21:51,615][starVLA.training.train_starvla][INFO] - Step 4190, Loss: {'train/loss': 0.3931572735309601, 'train/grad_norm_pre_clip': 19.465986251831055, 'timing/data': 0.001069243997335434, 'timing/model': 5.06323573179543, 'learning_rate/qwen_vl_interface': 3.885382780260371e-06, 'learning_rate/action_model': 3.885382780260371e-05, 'epoch': 1.53}) +[2026-06-18 01:22:00,606][starVLA.training.train_starvla][INFO] - Step 4191, Loss: {'train/loss': 0.5368870496749878, 'train/grad_norm_pre_clip': 23.5824031829834, 'timing/data': 0.0012269951403141022, 'timing/model': 2.8442105064168572, 'learning_rate/qwen_vl_interface': 3.883311400920366e-06, 'learning_rate/action_model': 3.883311400920366e-05, 'epoch': 1.53}) +[2026-06-18 01:22:11,212][starVLA.training.train_starvla][INFO] - Step 4192, Loss: {'train/loss': 0.29156607389450073, 'train/grad_norm_pre_clip': 17.863435745239258, 'timing/data': 0.007130159996449947, 'timing/model': 3.4574329629540443, 'learning_rate/qwen_vl_interface': 3.88124030489638e-06, 'learning_rate/action_model': 3.88124030489638e-05, 'epoch': 1.54}) +[2026-06-18 01:22:21,417][starVLA.training.train_starvla][INFO] - Step 4193, Loss: {'train/loss': 0.4331512153148651, 'train/grad_norm_pre_clip': 23.571529388427734, 'timing/data': 0.0009801173582673073, 'timing/model': 2.8074223259463906, 'learning_rate/qwen_vl_interface': 3.8791694926177495e-06, 'learning_rate/action_model': 3.87916949261775e-05, 'epoch': 1.54}) +[2026-06-18 01:22:31,106][starVLA.training.train_starvla][INFO] - Step 4194, Loss: {'train/loss': 0.5669266581535339, 'train/grad_norm_pre_clip': 26.771833419799805, 'timing/data': 0.001258830539882183, 'timing/model': 2.1976500814780593, 'learning_rate/qwen_vl_interface': 3.877098964513759e-06, 'learning_rate/action_model': 3.877098964513759e-05, 'epoch': 1.54}) +[2026-06-18 01:22:41,150][starVLA.training.train_starvla][INFO] - Step 4195, Loss: {'train/loss': 0.5146015882492065, 'train/grad_norm_pre_clip': 16.69493865966797, 'timing/data': 0.0009153587743639946, 'timing/model': 1.37629827670753, 'learning_rate/qwen_vl_interface': 3.875028721013633e-06, 'learning_rate/action_model': 3.875028721013632e-05, 'epoch': 1.54}) +[2026-06-18 01:22:56,827][starVLA.training.train_starvla][INFO] - Step 4196, Loss: {'train/loss': 0.3319806158542633, 'train/grad_norm_pre_clip': 16.0410099029541, 'timing/data': 0.003935600630939007, 'timing/model': 3.7846418526023626, 'learning_rate/qwen_vl_interface': 3.87295876254653e-06, 'learning_rate/action_model': 3.8729587625465304e-05, 'epoch': 1.54}) +[2026-06-18 01:23:04,094][starVLA.training.train_starvla][INFO] - Step 4197, Loss: {'train/loss': 0.34194058179855347, 'train/grad_norm_pre_clip': 15.50903034210205, 'timing/data': 0.0011320952326059341, 'timing/model': 1.326330085285008, 'learning_rate/qwen_vl_interface': 3.870889089541558e-06, 'learning_rate/action_model': 3.870889089541558e-05, 'epoch': 1.54}) +[2026-06-18 01:23:08,311][starVLA.training.train_starvla][INFO] - Step 4198, Loss: {'train/loss': 0.5164026021957397, 'train/grad_norm_pre_clip': 17.1929931640625, 'timing/data': 0.0011065397411584854, 'timing/model': 1.0510746510699391, 'learning_rate/qwen_vl_interface': 3.868819702427762e-06, 'learning_rate/action_model': 3.8688197024277616e-05, 'epoch': 1.54}) +[2026-06-18 01:23:12,873][starVLA.training.train_starvla][INFO] - Step 4199, Loss: {'train/loss': 0.35861116647720337, 'train/grad_norm_pre_clip': 18.48203468322754, 'timing/data': 0.0012852465733885765, 'timing/model': 1.106435394845903, 'learning_rate/qwen_vl_interface': 3.866750601634128e-06, 'learning_rate/action_model': 3.8667506016341277e-05, 'epoch': 1.54}) +[2026-06-18 01:27:03,118][starVLA.training.train_starvla][INFO] - Step 4200, Loss: {'train/loss': 0.5960218906402588, 'train/grad_norm_pre_clip': 24.275146484375, 'eval/loss': 0.5544238686561584, 'eval/latency_2/loss': 0.5544238686561584, 'eval/demon_fire_precision': 0.9317531671434409, 'eval/demon_fire_recall': 0.9539748953974896, 'eval/demon_fire_f1': 0.9427330990283234, 'eval/demon_fire_rate_model': 0.6808569838619922, 'eval/demon_fire_rate_teacher': 0.6649972175848636, 'eval/demon_left_precision': 0.8782051282051282, 'eval/demon_left_recall': 0.8652631578947368, 'eval/demon_left_f1': 0.8716861081654294, 'eval/demon_left_rate_model': 0.2604340567612688, 'eval/demon_left_rate_teacher': 0.2643294379521425, 'eval/demon_right_precision': 0.8804960541149943, 'eval/demon_right_recall': 0.8824858757062147, 'eval/demon_right_f1': 0.881489841986456, 'eval/demon_right_rate_model': 0.2468002225932109, 'eval/demon_right_rate_teacher': 0.24624373956594323, 'eval/demon_move_f1': 0.8765879750759427, 'eval/demon_cc_f1': 0.9218646232430682, 'eval/latency_2/demon_fire_precision': 0.9317531671434409, 'eval/latency_2/demon_fire_recall': 0.9539748953974896, 'eval/latency_2/demon_fire_f1': 0.9427330990283234, 'eval/latency_2/demon_fire_rate_model': 0.6808569838619922, 'eval/latency_2/demon_fire_rate_teacher': 0.6649972175848636, 'eval/latency_2/demon_left_precision': 0.8782051282051282, 'eval/latency_2/demon_left_recall': 0.8652631578947368, 'eval/latency_2/demon_left_f1': 0.8716861081654294, 'eval/latency_2/demon_left_rate_model': 0.2604340567612688, 'eval/latency_2/demon_left_rate_teacher': 0.2643294379521425, 'eval/latency_2/demon_right_precision': 0.8804960541149943, 'eval/latency_2/demon_right_recall': 0.8824858757062147, 'eval/latency_2/demon_right_f1': 0.881489841986456, 'eval/latency_2/demon_right_rate_model': 0.2468002225932109, 'eval/latency_2/demon_right_rate_teacher': 0.24624373956594323, 'eval/latency_2/demon_move_f1': 0.8765879750759427, 'eval/latency_2/demon_cc_f1': 0.9218646232430682, 'timing/data': 0.0008198637515306473, 'timing/model': 1.2449333630502224, 'learning_rate/qwen_vl_interface': 3.864681787589581e-06, 'learning_rate/action_model': 3.8646817875895816e-05, 'epoch': 1.54}) +[2026-06-18 01:27:03,141][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_4200_state +[2026-06-18 01:27:03,148][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 01:27:15,561][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_4200_state/model.safetensors +[2026-06-18 01:27:30,976][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_4200_state/optimizer.bin +[2026-06-18 01:27:30,977][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_4200_state/sampler.bin +[2026-06-18 01:27:30,978][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_4200_state/sampler_1.bin +[2026-06-18 01:27:30,981][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_4200_state/random_states_0.pkl +[2026-06-18 01:27:30,986][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4200 +[2026-06-18 01:27:30,986][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 01:27:31,003][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 01:27:35,825][starVLA.training.train_starvla][INFO] - Step 4201, Loss: {'train/loss': 0.5349887609481812, 'train/grad_norm_pre_clip': 22.217077255249023, 'timing/data': 0.0011763516813516617, 'timing/model': 1.3200923521071672, 'learning_rate/qwen_vl_interface': 3.862613260722991e-06, 'learning_rate/action_model': 3.862613260722992e-05, 'epoch': 1.54}) +[2026-06-18 01:27:40,535][starVLA.training.train_starvla][INFO] - Step 4202, Loss: {'train/loss': 0.5075061321258545, 'train/grad_norm_pre_clip': 18.375333786010742, 'timing/data': 0.0010862452909350395, 'timing/model': 1.0695393672212958, 'learning_rate/qwen_vl_interface': 3.860545021463166e-06, 'learning_rate/action_model': 3.860545021463167e-05, 'epoch': 1.54}) +[2026-06-18 01:27:44,929][starVLA.training.train_starvla][INFO] - Step 4203, Loss: {'train/loss': 0.19901373982429504, 'train/grad_norm_pre_clip': 10.320456504821777, 'timing/data': 0.0008819233626127243, 'timing/model': 1.40602233633399, 'learning_rate/qwen_vl_interface': 3.858477070238854e-06, 'learning_rate/action_model': 3.8584770702388534e-05, 'epoch': 1.54}) +[2026-06-18 01:27:52,828][starVLA.training.train_starvla][INFO] - Step 4204, Loss: {'train/loss': 0.7790133357048035, 'train/grad_norm_pre_clip': 21.990751266479492, 'timing/data': 0.001312493346631527, 'timing/model': 1.049344191327691, 'learning_rate/qwen_vl_interface': 3.856409407478741e-06, 'learning_rate/action_model': 3.856409407478741e-05, 'epoch': 1.54}) +[2026-06-18 01:27:57,354][starVLA.training.train_starvla][INFO] - Step 4205, Loss: {'train/loss': 0.3714295029640198, 'train/grad_norm_pre_clip': 31.157424926757812, 'timing/data': 0.0019746720790863037, 'timing/model': 1.0990502275526524, 'learning_rate/qwen_vl_interface': 3.854342033611456e-06, 'learning_rate/action_model': 3.854342033611456e-05, 'epoch': 1.54}) +[2026-06-18 01:28:02,054][starVLA.training.train_starvla][INFO] - Step 4206, Loss: {'train/loss': 0.2917880117893219, 'train/grad_norm_pre_clip': 16.701292037963867, 'timing/data': 0.0014099590480327606, 'timing/model': 1.1432385565713048, 'learning_rate/qwen_vl_interface': 3.852274949065571e-06, 'learning_rate/action_model': 3.852274949065571e-05, 'epoch': 1.54}) +[2026-06-18 01:28:06,898][starVLA.training.train_starvla][INFO] - Step 4207, Loss: {'train/loss': 0.6945812106132507, 'train/grad_norm_pre_clip': 21.171239852905273, 'timing/data': 0.0008958522230386734, 'timing/model': 1.1758147291839123, 'learning_rate/qwen_vl_interface': 3.850208154269593e-06, 'learning_rate/action_model': 3.8502081542695934e-05, 'epoch': 1.54}) +[2026-06-18 01:28:11,883][starVLA.training.train_starvla][INFO] - Step 4208, Loss: {'train/loss': 0.6382046341896057, 'train/grad_norm_pre_clip': 25.017274856567383, 'timing/data': 0.001471678726375103, 'timing/model': 1.365479027852416, 'learning_rate/qwen_vl_interface': 3.8481416496519725e-06, 'learning_rate/action_model': 3.848141649651972e-05, 'epoch': 1.54}) +[2026-06-18 01:28:16,198][starVLA.training.train_starvla][INFO] - Step 4209, Loss: {'train/loss': 0.4821925163269043, 'train/grad_norm_pre_clip': 22.709447860717773, 'timing/data': 0.0010444950312376022, 'timing/model': 1.038706966675818, 'learning_rate/qwen_vl_interface': 3.846075435641093e-06, 'learning_rate/action_model': 3.846075435641093e-05, 'epoch': 1.54}) +[2026-06-18 01:28:20,854][starVLA.training.train_starvla][INFO] - Step 4210, Loss: {'train/loss': 0.47606855630874634, 'train/grad_norm_pre_clip': 18.165803909301758, 'timing/data': 0.0011311862617731094, 'timing/model': 1.1684720143675804, 'learning_rate/qwen_vl_interface': 3.844009512665289e-06, 'learning_rate/action_model': 3.844009512665289e-05, 'epoch': 1.54}) +[2026-06-18 01:28:25,610][starVLA.training.train_starvla][INFO] - Step 4211, Loss: {'train/loss': 0.2597913146018982, 'train/grad_norm_pre_clip': 15.235698699951172, 'timing/data': 0.001214996911585331, 'timing/model': 1.0860797325149179, 'learning_rate/qwen_vl_interface': 3.841943881152826e-06, 'learning_rate/action_model': 3.841943881152826e-05, 'epoch': 1.54}) +[2026-06-18 01:28:30,606][starVLA.training.train_starvla][INFO] - Step 4212, Loss: {'train/loss': 0.5297064781188965, 'train/grad_norm_pre_clip': 19.23285675048828, 'timing/data': 0.0011206800118088722, 'timing/model': 1.2536611035466194, 'learning_rate/qwen_vl_interface': 3.83987854153191e-06, 'learning_rate/action_model': 3.83987854153191e-05, 'epoch': 1.54}) +[2026-06-18 01:28:34,920][starVLA.training.train_starvla][INFO] - Step 4213, Loss: {'train/loss': 0.4337487518787384, 'train/grad_norm_pre_clip': 17.899641036987305, 'timing/data': 0.0011368794366717339, 'timing/model': 1.292241764254868, 'learning_rate/qwen_vl_interface': 3.83781349423069e-06, 'learning_rate/action_model': 3.83781349423069e-05, 'epoch': 1.54}) +[2026-06-18 01:28:41,801][starVLA.training.train_starvla][INFO] - Step 4214, Loss: {'train/loss': 0.4404379725456238, 'train/grad_norm_pre_clip': 20.916486740112305, 'timing/data': 0.0010619061067700386, 'timing/model': 2.9512451756745577, 'learning_rate/qwen_vl_interface': 3.835748739677253e-06, 'learning_rate/action_model': 3.835748739677253e-05, 'epoch': 1.54}) +[2026-06-18 01:28:48,210][starVLA.training.train_starvla][INFO] - Step 4215, Loss: {'train/loss': 0.2767464518547058, 'train/grad_norm_pre_clip': 15.342107772827148, 'timing/data': 0.0012281285598874092, 'timing/model': 1.1045046793296933, 'learning_rate/qwen_vl_interface': 3.8336842782996245e-06, 'learning_rate/action_model': 3.8336842782996244e-05, 'epoch': 1.54}) +[2026-06-18 01:28:53,169][starVLA.training.train_starvla][INFO] - Step 4216, Loss: {'train/loss': 0.589451789855957, 'train/grad_norm_pre_clip': 23.19140625, 'timing/data': 0.0010585850104689598, 'timing/model': 1.0530954655259848, 'learning_rate/qwen_vl_interface': 3.831620110525766e-06, 'learning_rate/action_model': 3.831620110525767e-05, 'epoch': 1.54}) +[2026-06-18 01:28:57,602][starVLA.training.train_starvla][INFO] - Step 4217, Loss: {'train/loss': 0.8257209658622742, 'train/grad_norm_pre_clip': 23.199949264526367, 'timing/data': 0.000794287770986557, 'timing/model': 1.0403070347383618, 'learning_rate/qwen_vl_interface': 3.829556236783588e-06, 'learning_rate/action_model': 3.8295562367835884e-05, 'epoch': 1.54}) +[2026-06-18 01:29:02,452][starVLA.training.train_starvla][INFO] - Step 4218, Loss: {'train/loss': 0.3646164834499359, 'train/grad_norm_pre_clip': 22.28622055053711, 'timing/data': 0.0011360971257090569, 'timing/model': 1.4360504914075136, 'learning_rate/qwen_vl_interface': 3.827492657500931e-06, 'learning_rate/action_model': 3.8274926575009305e-05, 'epoch': 1.54}) +[2026-06-18 01:29:07,254][starVLA.training.train_starvla][INFO] - Step 4219, Loss: {'train/loss': 0.22425764799118042, 'train/grad_norm_pre_clip': 13.224438667297363, 'timing/data': 0.0008514486253261566, 'timing/model': 1.465396260842681, 'learning_rate/qwen_vl_interface': 3.8254293731055754e-06, 'learning_rate/action_model': 3.825429373105576e-05, 'epoch': 1.54}) +[2026-06-18 01:29:12,040][starVLA.training.train_starvla][INFO] - Step 4220, Loss: {'train/loss': 0.13357344269752502, 'train/grad_norm_pre_clip': 10.108451843261719, 'timing/data': 0.0012500248849391937, 'timing/model': 1.0269297249615192, 'learning_rate/qwen_vl_interface': 3.823366384025243e-06, 'learning_rate/action_model': 3.8233663840252426e-05, 'epoch': 1.55}) +[2026-06-18 01:29:16,665][starVLA.training.train_starvla][INFO] - Step 4221, Loss: {'train/loss': 0.24382498860359192, 'train/grad_norm_pre_clip': 25.685447692871094, 'timing/data': 0.0012291399762034416, 'timing/model': 1.1686024265363812, 'learning_rate/qwen_vl_interface': 3.821303690687595e-06, 'learning_rate/action_model': 3.821303690687595e-05, 'epoch': 1.55}) +[2026-06-18 01:29:21,327][starVLA.training.train_starvla][INFO] - Step 4222, Loss: {'train/loss': 0.3566732108592987, 'train/grad_norm_pre_clip': 18.9778995513916, 'timing/data': 0.0009141266345977783, 'timing/model': 1.05890847556293, 'learning_rate/qwen_vl_interface': 3.81924129352023e-06, 'learning_rate/action_model': 3.8192412935202304e-05, 'epoch': 1.55}) +[2026-06-18 01:29:26,325][starVLA.training.train_starvla][INFO] - Step 4223, Loss: {'train/loss': 0.3324241638183594, 'train/grad_norm_pre_clip': 16.49471664428711, 'timing/data': 0.0012256214395165443, 'timing/model': 1.6162791028618813, 'learning_rate/qwen_vl_interface': 3.817179192950682e-06, 'learning_rate/action_model': 3.8171791929506826e-05, 'epoch': 1.55}) +[2026-06-18 01:29:32,124][starVLA.training.train_starvla][INFO] - Step 4224, Loss: {'train/loss': 0.5046924352645874, 'train/grad_norm_pre_clip': 24.457786560058594, 'timing/data': 0.0010752519592642784, 'timing/model': 1.0540707502514124, 'learning_rate/qwen_vl_interface': 3.815117389406432e-06, 'learning_rate/action_model': 3.8151173894064324e-05, 'epoch': 1.55}) +[2026-06-18 01:29:39,611][starVLA.training.train_starvla][INFO] - Step 4225, Loss: {'train/loss': 0.4517945945262909, 'train/grad_norm_pre_clip': 20.84269142150879, 'timing/data': 0.0011802958324551582, 'timing/model': 3.904865621589124, 'learning_rate/qwen_vl_interface': 3.8130558833148906e-06, 'learning_rate/action_model': 3.813055883314891e-05, 'epoch': 1.55}) +[2026-06-18 01:29:52,597][starVLA.training.train_starvla][INFO] - Step 4226, Loss: {'train/loss': 0.5056807398796082, 'train/grad_norm_pre_clip': 28.04339027404785, 'timing/data': 0.0016629882156848907, 'timing/model': 1.1467730151489377, 'learning_rate/qwen_vl_interface': 3.81099467510341e-06, 'learning_rate/action_model': 3.81099467510341e-05, 'epoch': 1.55}) +[2026-06-18 01:29:58,032][starVLA.training.train_starvla][INFO] - Step 4227, Loss: {'train/loss': 0.45667392015457153, 'train/grad_norm_pre_clip': 18.9879150390625, 'timing/data': 0.013217427767813206, 'timing/model': 2.067091471515596, 'learning_rate/qwen_vl_interface': 3.8089337651992815e-06, 'learning_rate/action_model': 3.808933765199281e-05, 'epoch': 1.55}) +[2026-06-18 01:30:02,791][starVLA.training.train_starvla][INFO] - Step 4228, Loss: {'train/loss': 0.256541907787323, 'train/grad_norm_pre_clip': 13.518290519714355, 'timing/data': 0.03408297896385193, 'timing/model': 1.5187928322702646, 'learning_rate/qwen_vl_interface': 3.8068731540297344e-06, 'learning_rate/action_model': 3.806873154029734e-05, 'epoch': 1.55}) +[2026-06-18 01:30:07,128][starVLA.training.train_starvla][INFO] - Step 4229, Loss: {'train/loss': 0.5327631235122681, 'train/grad_norm_pre_clip': 17.730241775512695, 'timing/data': 0.0239030783995986, 'timing/model': 1.0682350611314178, 'learning_rate/qwen_vl_interface': 3.8048128420219366e-06, 'learning_rate/action_model': 3.804812842021936e-05, 'epoch': 1.55}) +[2026-06-18 01:30:11,801][starVLA.training.train_starvla][INFO] - Step 4230, Loss: {'train/loss': 0.684316873550415, 'train/grad_norm_pre_clip': 19.7353458404541, 'timing/data': 0.034567817114293575, 'timing/model': 1.2589259650558233, 'learning_rate/qwen_vl_interface': 3.80275282960299e-06, 'learning_rate/action_model': 3.80275282960299e-05, 'epoch': 1.55}) +[2026-06-18 01:30:16,282][starVLA.training.train_starvla][INFO] - Step 4231, Loss: {'train/loss': 0.4203203320503235, 'train/grad_norm_pre_clip': 19.589296340942383, 'timing/data': 0.02181115932762623, 'timing/model': 1.036292975768447, 'learning_rate/qwen_vl_interface': 3.8006931171999394e-06, 'learning_rate/action_model': 3.800693117199939e-05, 'epoch': 1.55}) +[2026-06-18 01:30:20,818][starVLA.training.train_starvla][INFO] - Step 4232, Loss: {'train/loss': 0.40730032324790955, 'train/grad_norm_pre_clip': 15.89220905303955, 'timing/data': 0.02186335064470768, 'timing/model': 1.0847027879208326, 'learning_rate/qwen_vl_interface': 3.798633705239766e-06, 'learning_rate/action_model': 3.7986337052397655e-05, 'epoch': 1.55}) +[2026-06-18 01:30:25,524][starVLA.training.train_starvla][INFO] - Step 4233, Loss: {'train/loss': 0.4748855233192444, 'train/grad_norm_pre_clip': 27.318918228149414, 'timing/data': 0.027141758240759373, 'timing/model': 1.4628620203584433, 'learning_rate/qwen_vl_interface': 3.796574594149387e-06, 'learning_rate/action_model': 3.796574594149387e-05, 'epoch': 1.55}) +[2026-06-18 01:30:32,042][starVLA.training.train_starvla][INFO] - Step 4234, Loss: {'train/loss': 0.3500663936138153, 'train/grad_norm_pre_clip': 14.246541023254395, 'timing/data': 0.023801546543836594, 'timing/model': 2.8048951122909784, 'learning_rate/qwen_vl_interface': 3.794515784355657e-06, 'learning_rate/action_model': 3.7945157843556575e-05, 'epoch': 1.55}) +[2026-06-18 01:30:38,245][starVLA.training.train_starvla][INFO] - Step 4235, Loss: {'train/loss': 0.369220495223999, 'train/grad_norm_pre_clip': 18.255170822143555, 'timing/data': 0.03085710760205984, 'timing/model': 1.098840931430459, 'learning_rate/qwen_vl_interface': 3.7924572762853727e-06, 'learning_rate/action_model': 3.7924572762853725e-05, 'epoch': 1.55}) +[2026-06-18 01:30:43,454][starVLA.training.train_starvla][INFO] - Step 4236, Loss: {'train/loss': 0.49174633622169495, 'train/grad_norm_pre_clip': 22.648834228515625, 'timing/data': 0.03670296631753445, 'timing/model': 1.266243351623416, 'learning_rate/qwen_vl_interface': 3.7903990703652643e-06, 'learning_rate/action_model': 3.790399070365264e-05, 'epoch': 1.55}) +[2026-06-18 01:30:48,425][starVLA.training.train_starvla][INFO] - Step 4237, Loss: {'train/loss': 0.40115979313850403, 'train/grad_norm_pre_clip': 18.77804946899414, 'timing/data': 0.02108660899102688, 'timing/model': 1.163912638090551, 'learning_rate/qwen_vl_interface': 3.7883411670219976e-06, 'learning_rate/action_model': 3.788341167021998e-05, 'epoch': 1.55}) +[2026-06-18 01:30:53,071][starVLA.training.train_starvla][INFO] - Step 4238, Loss: {'train/loss': 0.2502439022064209, 'train/grad_norm_pre_clip': 17.38772201538086, 'timing/data': 0.025854628533124924, 'timing/model': 1.4093351950868964, 'learning_rate/qwen_vl_interface': 3.7862835666821795e-06, 'learning_rate/action_model': 3.78628356668218e-05, 'epoch': 1.55}) +[2026-06-18 01:31:00,307][starVLA.training.train_starvla][INFO] - Step 4239, Loss: {'train/loss': 0.6686086654663086, 'train/grad_norm_pre_clip': 25.37312126159668, 'timing/data': 0.023518755100667477, 'timing/model': 1.1551462905481458, 'learning_rate/qwen_vl_interface': 3.7842262697723543e-06, 'learning_rate/action_model': 3.784226269772354e-05, 'epoch': 1.55}) +[2026-06-18 01:31:04,817][starVLA.training.train_starvla][INFO] - Step 4240, Loss: {'train/loss': 0.26208052039146423, 'train/grad_norm_pre_clip': 15.79198169708252, 'timing/data': 0.022229409776628017, 'timing/model': 1.0930099347606301, 'learning_rate/qwen_vl_interface': 3.782169276719e-06, 'learning_rate/action_model': 3.782169276719e-05, 'epoch': 1.55}) +[2026-06-18 01:31:13,383][starVLA.training.train_starvla][INFO] - Step 4241, Loss: {'train/loss': 0.5739309191703796, 'train/grad_norm_pre_clip': 28.867326736450195, 'timing/data': 0.02137115504592657, 'timing/model': 1.0353654650971293, 'learning_rate/qwen_vl_interface': 3.7801125879485355e-06, 'learning_rate/action_model': 3.7801125879485356e-05, 'epoch': 1.55}) +[2026-06-18 01:31:17,934][starVLA.training.train_starvla][INFO] - Step 4242, Loss: {'train/loss': 0.4017406702041626, 'train/grad_norm_pre_clip': 24.17892837524414, 'timing/data': 0.021331452764570713, 'timing/model': 1.0491480333730578, 'learning_rate/qwen_vl_interface': 3.77805620388731e-06, 'learning_rate/action_model': 3.7780562038873105e-05, 'epoch': 1.55}) +[2026-06-18 01:31:22,229][starVLA.training.train_starvla][INFO] - Step 4243, Loss: {'train/loss': 0.3245271146297455, 'train/grad_norm_pre_clip': 17.67854881286621, 'timing/data': 0.029067705385386944, 'timing/model': 1.0593361556529999, 'learning_rate/qwen_vl_interface': 3.7760001249616204e-06, 'learning_rate/action_model': 3.77600012496162e-05, 'epoch': 1.55}) +[2026-06-18 01:31:26,747][starVLA.training.train_starvla][INFO] - Step 4244, Loss: {'train/loss': 0.42253777384757996, 'train/grad_norm_pre_clip': 38.55451583862305, 'timing/data': 0.026437525637447834, 'timing/model': 1.1202199682593346, 'learning_rate/qwen_vl_interface': 3.7739443515976898e-06, 'learning_rate/action_model': 3.7739443515976894e-05, 'epoch': 1.55}) +[2026-06-18 01:31:31,281][starVLA.training.train_starvla][INFO] - Step 4245, Loss: {'train/loss': 0.2299860566854477, 'train/grad_norm_pre_clip': 13.337320327758789, 'timing/data': 0.03154757525771856, 'timing/model': 1.0955162486061454, 'learning_rate/qwen_vl_interface': 3.7718888842216822e-06, 'learning_rate/action_model': 3.771888884221682e-05, 'epoch': 1.55}) +[2026-06-18 01:31:38,190][starVLA.training.train_starvla][INFO] - Step 4246, Loss: {'train/loss': 0.17048317193984985, 'train/grad_norm_pre_clip': 10.897957801818848, 'timing/data': 0.021350692957639694, 'timing/model': 3.427685650996864, 'learning_rate/qwen_vl_interface': 3.7698337232597002e-06, 'learning_rate/action_model': 3.7698337232597e-05, 'epoch': 1.55}) +[2026-06-18 01:31:42,851][starVLA.training.train_starvla][INFO] - Step 4247, Loss: {'train/loss': 0.4892424941062927, 'train/grad_norm_pre_clip': 21.233829498291016, 'timing/data': 0.026525555178523064, 'timing/model': 1.3984183510765433, 'learning_rate/qwen_vl_interface': 3.7677788691377814e-06, 'learning_rate/action_model': 3.7677788691377817e-05, 'epoch': 1.56}) +[2026-06-18 01:31:47,467][starVLA.training.train_starvla][INFO] - Step 4248, Loss: {'train/loss': 0.4055686891078949, 'train/grad_norm_pre_clip': 23.91551399230957, 'timing/data': 0.03082690667361021, 'timing/model': 1.3395700380206108, 'learning_rate/qwen_vl_interface': 3.7657243222818978e-06, 'learning_rate/action_model': 3.7657243222818976e-05, 'epoch': 1.56}) +[2026-06-18 01:32:00,589][starVLA.training.train_starvla][INFO] - Step 4249, Loss: {'train/loss': 0.3158101439476013, 'train/grad_norm_pre_clip': 19.195228576660156, 'timing/data': 0.033256216906011105, 'timing/model': 5.487687336280942, 'learning_rate/qwen_vl_interface': 3.7636700831179575e-06, 'learning_rate/action_model': 3.763670083117957e-05, 'epoch': 1.56}) +[2026-06-18 01:32:05,217][starVLA.training.train_starvla][INFO] - Step 4250, Loss: {'train/loss': 0.2206612080335617, 'train/grad_norm_pre_clip': 15.269556045532227, 'timing/data': 0.02622184529900551, 'timing/model': 1.1682011056691408, 'learning_rate/qwen_vl_interface': 3.7616161520718096e-06, 'learning_rate/action_model': 3.76161615207181e-05, 'epoch': 1.56}) +[2026-06-18 01:32:10,225][starVLA.training.train_starvla][INFO] - Step 4251, Loss: {'train/loss': 0.469291090965271, 'train/grad_norm_pre_clip': 20.05602264404297, 'timing/data': 0.03413581382483244, 'timing/model': 1.085858159698546, 'learning_rate/qwen_vl_interface': 3.7595625295692352e-06, 'learning_rate/action_model': 3.759562529569235e-05, 'epoch': 1.56}) +[2026-06-18 01:32:14,908][starVLA.training.train_starvla][INFO] - Step 4252, Loss: {'train/loss': 0.3012436330318451, 'train/grad_norm_pre_clip': 21.78241539001465, 'timing/data': 0.0279406001791358, 'timing/model': 1.0848188363015652, 'learning_rate/qwen_vl_interface': 3.7575092160359523e-06, 'learning_rate/action_model': 3.757509216035952e-05, 'epoch': 1.56}) +[2026-06-18 01:32:19,614][starVLA.training.train_starvla][INFO] - Step 4253, Loss: {'train/loss': 0.5845749378204346, 'train/grad_norm_pre_clip': 29.926898956298828, 'timing/data': 0.02196592278778553, 'timing/model': 1.3874347126111388, 'learning_rate/qwen_vl_interface': 3.7554562118976135e-06, 'learning_rate/action_model': 3.7554562118976136e-05, 'epoch': 1.56}) +[2026-06-18 01:32:30,310][starVLA.training.train_starvla][INFO] - Step 4254, Loss: {'train/loss': 0.43974339962005615, 'train/grad_norm_pre_clip': 20.761981964111328, 'timing/data': 0.06486690789461136, 'timing/model': 2.2949556559324265, 'learning_rate/qwen_vl_interface': 3.7534035175798113e-06, 'learning_rate/action_model': 3.753403517579811e-05, 'epoch': 1.56}) +[2026-06-18 01:32:36,602][starVLA.training.train_starvla][INFO] - Step 4255, Loss: {'train/loss': 0.35888880491256714, 'train/grad_norm_pre_clip': 19.630762100219727, 'timing/data': 0.0013898862525820732, 'timing/model': 1.0800646310672164, 'learning_rate/qwen_vl_interface': 3.7513511335080706e-06, 'learning_rate/action_model': 3.75135113350807e-05, 'epoch': 1.56}) +[2026-06-18 01:32:41,386][starVLA.training.train_starvla][INFO] - Step 4256, Loss: {'train/loss': 0.2762247920036316, 'train/grad_norm_pre_clip': 18.79374885559082, 'timing/data': 0.0014472994953393936, 'timing/model': 1.0541879599913955, 'learning_rate/qwen_vl_interface': 3.74929906010785e-06, 'learning_rate/action_model': 3.74929906010785e-05, 'epoch': 1.56}) +[2026-06-18 01:32:48,801][starVLA.training.train_starvla][INFO] - Step 4257, Loss: {'train/loss': 0.3961077928543091, 'train/grad_norm_pre_clip': 17.073144912719727, 'timing/data': 0.0011624107137322426, 'timing/model': 1.0436161467805505, 'learning_rate/qwen_vl_interface': 3.747247297804553e-06, 'learning_rate/action_model': 3.7472472978045526e-05, 'epoch': 1.56}) +[2026-06-18 01:32:53,789][starVLA.training.train_starvla][INFO] - Step 4258, Loss: {'train/loss': 0.36321741342544556, 'train/grad_norm_pre_clip': 19.001981735229492, 'timing/data': 0.0009774556383490562, 'timing/model': 1.8769618272781372, 'learning_rate/qwen_vl_interface': 3.745195847023506e-06, 'learning_rate/action_model': 3.745195847023506e-05, 'epoch': 1.56}) +[2026-06-18 01:33:04,129][starVLA.training.train_starvla][INFO] - Step 4259, Loss: {'train/loss': 0.20691700279712677, 'train/grad_norm_pre_clip': 15.189644813537598, 'timing/data': 0.0017616897821426392, 'timing/model': 1.0959774078801274, 'learning_rate/qwen_vl_interface': 3.7431447081899795e-06, 'learning_rate/action_model': 3.7431447081899796e-05, 'epoch': 1.56}) +[2026-06-18 01:33:09,350][starVLA.training.train_starvla][INFO] - Step 4260, Loss: {'train/loss': 0.40522947907447815, 'train/grad_norm_pre_clip': 19.432844161987305, 'timing/data': 0.0008762245997786522, 'timing/model': 1.1132191009819508, 'learning_rate/qwen_vl_interface': 3.7410938817291753e-06, 'learning_rate/action_model': 3.741093881729175e-05, 'epoch': 1.56}) +[2026-06-18 01:33:14,094][starVLA.training.train_starvla][INFO] - Step 4261, Loss: {'train/loss': 0.7790281772613525, 'train/grad_norm_pre_clip': 28.856552124023438, 'timing/data': 0.0008664727210998535, 'timing/model': 1.0432047126814723, 'learning_rate/qwen_vl_interface': 3.739043368066234e-06, 'learning_rate/action_model': 3.739043368066234e-05, 'epoch': 1.56}) +[2026-06-18 01:33:24,023][starVLA.training.train_starvla][INFO] - Step 4262, Loss: {'train/loss': 0.5989629030227661, 'train/grad_norm_pre_clip': 24.091896057128906, 'timing/data': 0.0013105999678373337, 'timing/model': 1.1451872270554304, 'learning_rate/qwen_vl_interface': 3.7369931676262304e-06, 'learning_rate/action_model': 3.7369931676262304e-05, 'epoch': 1.56}) +[2026-06-18 01:33:31,694][starVLA.training.train_starvla][INFO] - Step 4263, Loss: {'train/loss': 0.4179895222187042, 'train/grad_norm_pre_clip': 27.677019119262695, 'timing/data': 0.0014354242011904716, 'timing/model': 1.7120960280299187, 'learning_rate/qwen_vl_interface': 3.7349432808341673e-06, 'learning_rate/action_model': 3.7349432808341674e-05, 'epoch': 1.56}) +[2026-06-18 01:33:40,694][starVLA.training.train_starvla][INFO] - Step 4264, Loss: {'train/loss': 0.3628399670124054, 'train/grad_norm_pre_clip': 19.978717803955078, 'timing/data': 0.0012181904166936874, 'timing/model': 1.0453970469534397, 'learning_rate/qwen_vl_interface': 3.7328937081149952e-06, 'learning_rate/action_model': 3.732893708114995e-05, 'epoch': 1.56}) +[2026-06-18 01:33:45,337][starVLA.training.train_starvla][INFO] - Step 4265, Loss: {'train/loss': 0.5463131666183472, 'train/grad_norm_pre_clip': 22.560237884521484, 'timing/data': 0.001049150712788105, 'timing/model': 1.0818596165627241, 'learning_rate/qwen_vl_interface': 3.7308444498935894e-06, 'learning_rate/action_model': 3.730844449893589e-05, 'epoch': 1.56}) +[2026-06-18 01:33:51,018][starVLA.training.train_starvla][INFO] - Step 4266, Loss: {'train/loss': 0.7097487449645996, 'train/grad_norm_pre_clip': 42.9151725769043, 'timing/data': 0.0009242137894034386, 'timing/model': 1.0876026470214128, 'learning_rate/qwen_vl_interface': 3.7287955065947634e-06, 'learning_rate/action_model': 3.728795506594763e-05, 'epoch': 1.56}) +[2026-06-18 01:33:55,685][starVLA.training.train_starvla][INFO] - Step 4267, Loss: {'train/loss': 0.4191701114177704, 'train/grad_norm_pre_clip': 22.173551559448242, 'timing/data': 0.0013605756685137749, 'timing/model': 1.1125056138262153, 'learning_rate/qwen_vl_interface': 3.7267468786432655e-06, 'learning_rate/action_model': 3.7267468786432654e-05, 'epoch': 1.56}) +[2026-06-18 01:33:59,990][starVLA.training.train_starvla][INFO] - Step 4268, Loss: {'train/loss': 0.4914454519748688, 'train/grad_norm_pre_clip': 21.58707618713379, 'timing/data': 0.001027754507958889, 'timing/model': 1.0684990799054503, 'learning_rate/qwen_vl_interface': 3.7246985664637787e-06, 'learning_rate/action_model': 3.724698566463778e-05, 'epoch': 1.56}) +[2026-06-18 01:34:11,140][starVLA.training.train_starvla][INFO] - Step 4269, Loss: {'train/loss': 0.5860537886619568, 'train/grad_norm_pre_clip': 21.394760131835938, 'timing/data': 0.0010217539966106415, 'timing/model': 2.9484345875680447, 'learning_rate/qwen_vl_interface': 3.722650570480922e-06, 'learning_rate/action_model': 3.722650570480922e-05, 'epoch': 1.56}) +[2026-06-18 01:34:15,625][starVLA.training.train_starvla][INFO] - Step 4270, Loss: {'train/loss': 0.3215915560722351, 'train/grad_norm_pre_clip': 13.626440048217773, 'timing/data': 0.0024855881929397583, 'timing/model': 1.0967943351715803, 'learning_rate/qwen_vl_interface': 3.7206028911192445e-06, 'learning_rate/action_model': 3.720602891119244e-05, 'epoch': 1.56}) +[2026-06-18 01:34:20,460][starVLA.training.train_starvla][INFO] - Step 4271, Loss: {'train/loss': 0.1866595298051834, 'train/grad_norm_pre_clip': 16.95136070251465, 'timing/data': 0.0011527985334396362, 'timing/model': 1.2005410604178905, 'learning_rate/qwen_vl_interface': 3.7185555288032315e-06, 'learning_rate/action_model': 3.7185555288032316e-05, 'epoch': 1.56}) +[2026-06-18 01:34:25,438][starVLA.training.train_starvla][INFO] - Step 4272, Loss: {'train/loss': 0.314371258020401, 'train/grad_norm_pre_clip': 18.172666549682617, 'timing/data': 0.002058025449514389, 'timing/model': 1.1812054915353656, 'learning_rate/qwen_vl_interface': 3.716508483957308e-06, 'learning_rate/action_model': 3.716508483957308e-05, 'epoch': 1.56}) +[2026-06-18 01:34:30,324][starVLA.training.train_starvla][INFO] - Step 4273, Loss: {'train/loss': 0.41733306646347046, 'train/grad_norm_pre_clip': 20.803335189819336, 'timing/data': 0.0008983854204416275, 'timing/model': 1.3779550259932876, 'learning_rate/qwen_vl_interface': 3.7144617570058245e-06, 'learning_rate/action_model': 3.7144617570058245e-05, 'epoch': 1.56}) +[2026-06-18 01:34:34,741][starVLA.training.train_starvla][INFO] - Step 4274, Loss: {'train/loss': 0.4727364182472229, 'train/grad_norm_pre_clip': 23.274442672729492, 'timing/data': 0.0008998587727546692, 'timing/model': 1.1844010949134827, 'learning_rate/qwen_vl_interface': 3.7124153483730706e-06, 'learning_rate/action_model': 3.712415348373071e-05, 'epoch': 1.57}) +[2026-06-18 01:34:43,510][starVLA.training.train_starvla][INFO] - Step 4275, Loss: {'train/loss': 0.24715301394462585, 'train/grad_norm_pre_clip': 20.99152946472168, 'timing/data': 0.0028617223724722862, 'timing/model': 1.6567103071138263, 'learning_rate/qwen_vl_interface': 3.71036925848327e-06, 'learning_rate/action_model': 3.71036925848327e-05, 'epoch': 1.57}) +[2026-06-18 01:34:51,615][starVLA.training.train_starvla][INFO] - Step 4276, Loss: {'train/loss': 0.36797356605529785, 'train/grad_norm_pre_clip': 20.5053768157959, 'timing/data': 0.0030072256922721863, 'timing/model': 2.699856836348772, 'learning_rate/qwen_vl_interface': 3.7083234877605804e-06, 'learning_rate/action_model': 3.7083234877605806e-05, 'epoch': 1.57}) +[2026-06-18 01:34:57,980][starVLA.training.train_starvla][INFO] - Step 4277, Loss: {'train/loss': 0.20411553978919983, 'train/grad_norm_pre_clip': 16.581165313720703, 'timing/data': 0.0008116764947772026, 'timing/model': 1.3424539295956492, 'learning_rate/qwen_vl_interface': 3.70627803662909e-06, 'learning_rate/action_model': 3.70627803662909e-05, 'epoch': 1.57}) +[2026-06-18 01:35:07,162][starVLA.training.train_starvla][INFO] - Step 4278, Loss: {'train/loss': 0.7262197732925415, 'train/grad_norm_pre_clip': 32.101070404052734, 'timing/data': 0.001368408091366291, 'timing/model': 5.652206250466406, 'learning_rate/qwen_vl_interface': 3.7042329055128216e-06, 'learning_rate/action_model': 3.704232905512821e-05, 'epoch': 1.57}) +[2026-06-18 01:35:12,410][starVLA.training.train_starvla][INFO] - Step 4279, Loss: {'train/loss': 0.2309153527021408, 'train/grad_norm_pre_clip': 14.828300476074219, 'timing/data': 0.0009163646027445793, 'timing/model': 1.4783722329884768, 'learning_rate/qwen_vl_interface': 3.7021880948357375e-06, 'learning_rate/action_model': 3.702188094835737e-05, 'epoch': 1.57}) +[2026-06-18 01:35:22,242][starVLA.training.train_starvla][INFO] - Step 4280, Loss: {'train/loss': 0.7945662140846252, 'train/grad_norm_pre_clip': 32.754451751708984, 'timing/data': 0.0011365758255124092, 'timing/model': 1.1435201158747077, 'learning_rate/qwen_vl_interface': 3.700143605021726e-06, 'learning_rate/action_model': 3.7001436050217256e-05, 'epoch': 1.57}) +[2026-06-18 01:35:35,349][starVLA.training.train_starvla][INFO] - Step 4281, Loss: {'train/loss': 0.3156060576438904, 'train/grad_norm_pre_clip': 21.76326560974121, 'timing/data': 0.001572970300912857, 'timing/model': 4.571261792443693, 'learning_rate/qwen_vl_interface': 3.6980994364946133e-06, 'learning_rate/action_model': 3.698099436494613e-05, 'epoch': 1.57}) +[2026-06-18 01:35:43,574][starVLA.training.train_starvla][INFO] - Step 4282, Loss: {'train/loss': 0.26076117157936096, 'train/grad_norm_pre_clip': 16.55234718322754, 'timing/data': 0.0021082451567053795, 'timing/model': 1.520418387837708, 'learning_rate/qwen_vl_interface': 3.696055589678156e-06, 'learning_rate/action_model': 3.6960555896781554e-05, 'epoch': 1.57}) +[2026-06-18 01:35:56,018][starVLA.training.train_starvla][INFO] - Step 4283, Loss: {'train/loss': 0.4245619773864746, 'train/grad_norm_pre_clip': 19.08412742614746, 'timing/data': 0.0007849782705307007, 'timing/model': 1.6567544927820563, 'learning_rate/qwen_vl_interface': 3.694012064996049e-06, 'learning_rate/action_model': 3.694012064996049e-05, 'epoch': 1.57}) +[2026-06-18 01:36:05,054][starVLA.training.train_starvla][INFO] - Step 4284, Loss: {'train/loss': 0.2210814207792282, 'train/grad_norm_pre_clip': 16.518714904785156, 'timing/data': 0.0011915825307369232, 'timing/model': 1.388782468624413, 'learning_rate/qwen_vl_interface': 3.6919688628719153e-06, 'learning_rate/action_model': 3.691968862871915e-05, 'epoch': 1.57}) +[2026-06-18 01:36:12,223][starVLA.training.train_starvla][INFO] - Step 4285, Loss: {'train/loss': 0.38297635316848755, 'train/grad_norm_pre_clip': 21.1022891998291, 'timing/data': 0.0007873894646763802, 'timing/model': 1.302571008913219, 'learning_rate/qwen_vl_interface': 3.689925983729311e-06, 'learning_rate/action_model': 3.689925983729311e-05, 'epoch': 1.57}) +[2026-06-18 01:36:16,699][starVLA.training.train_starvla][INFO] - Step 4286, Loss: {'train/loss': 0.36418208479881287, 'train/grad_norm_pre_clip': 16.666980743408203, 'timing/data': 0.0009120544418692589, 'timing/model': 1.0383505010977387, 'learning_rate/qwen_vl_interface': 3.68788342799173e-06, 'learning_rate/action_model': 3.6878834279917294e-05, 'epoch': 1.57}) +[2026-06-18 01:36:21,347][starVLA.training.train_starvla][INFO] - Step 4287, Loss: {'train/loss': 0.2423969954252243, 'train/grad_norm_pre_clip': 18.579925537109375, 'timing/data': 0.0013564350083470345, 'timing/model': 1.2351983115077019, 'learning_rate/qwen_vl_interface': 3.6858411960825953e-06, 'learning_rate/action_model': 3.685841196082595e-05, 'epoch': 1.57}) +[2026-06-18 01:36:25,729][starVLA.training.train_starvla][INFO] - Step 4288, Loss: {'train/loss': 0.18151071667671204, 'train/grad_norm_pre_clip': 14.72285270690918, 'timing/data': 0.0009083542972803116, 'timing/model': 1.3726790407672524, 'learning_rate/qwen_vl_interface': 3.6837992884252636e-06, 'learning_rate/action_model': 3.683799288425264e-05, 'epoch': 1.57}) +[2026-06-18 01:36:30,406][starVLA.training.train_starvla][INFO] - Step 4289, Loss: {'train/loss': 0.5004522800445557, 'train/grad_norm_pre_clip': 30.67583465576172, 'timing/data': 0.0011440422385931015, 'timing/model': 1.0532845696434379, 'learning_rate/qwen_vl_interface': 3.6817577054430243e-06, 'learning_rate/action_model': 3.6817577054430244e-05, 'epoch': 1.57}) +[2026-06-18 01:36:35,413][starVLA.training.train_starvla][INFO] - Step 4290, Loss: {'train/loss': 0.20809175074100494, 'train/grad_norm_pre_clip': 12.978480339050293, 'timing/data': 0.0017608469352126122, 'timing/model': 1.104305885732174, 'learning_rate/qwen_vl_interface': 3.6797164475591003e-06, 'learning_rate/action_model': 3.6797164475591e-05, 'epoch': 1.57}) +[2026-06-18 01:36:40,090][starVLA.training.train_starvla][INFO] - Step 4291, Loss: {'train/loss': 0.13196875154972076, 'train/grad_norm_pre_clip': 11.966302871704102, 'timing/data': 0.0010934537276625633, 'timing/model': 1.0743693048134446, 'learning_rate/qwen_vl_interface': 3.677675515196648e-06, 'learning_rate/action_model': 3.677675515196648e-05, 'epoch': 1.57}) +[2026-06-18 01:36:44,886][starVLA.training.train_starvla][INFO] - Step 4292, Loss: {'train/loss': 0.36893320083618164, 'train/grad_norm_pre_clip': 15.834647178649902, 'timing/data': 0.0010822005569934845, 'timing/model': 1.0323411962017417, 'learning_rate/qwen_vl_interface': 3.6756349087787524e-06, 'learning_rate/action_model': 3.6756349087787524e-05, 'epoch': 1.57}) +[2026-06-18 01:36:49,333][starVLA.training.train_starvla][INFO] - Step 4293, Loss: {'train/loss': 0.2296553999185562, 'train/grad_norm_pre_clip': 18.071571350097656, 'timing/data': 0.0015967795625329018, 'timing/model': 1.4163258923217654, 'learning_rate/qwen_vl_interface': 3.6735946287284325e-06, 'learning_rate/action_model': 3.6735946287284324e-05, 'epoch': 1.57}) +[2026-06-18 01:36:53,816][starVLA.training.train_starvla][INFO] - Step 4294, Loss: {'train/loss': 0.23784279823303223, 'train/grad_norm_pre_clip': 12.844986915588379, 'timing/data': 0.0008917544037103653, 'timing/model': 1.0486687300726771, 'learning_rate/qwen_vl_interface': 3.6715546754686444e-06, 'learning_rate/action_model': 3.6715546754686444e-05, 'epoch': 1.57}) +[2026-06-18 01:36:58,284][starVLA.training.train_starvla][INFO] - Step 4295, Loss: {'train/loss': 0.5039384961128235, 'train/grad_norm_pre_clip': 22.861080169677734, 'timing/data': 0.00115266814827919, 'timing/model': 1.0527271591126919, 'learning_rate/qwen_vl_interface': 3.6695150494222697e-06, 'learning_rate/action_model': 3.6695150494222695e-05, 'epoch': 1.57}) +[2026-06-18 01:37:02,872][starVLA.training.train_starvla][INFO] - Step 4296, Loss: {'train/loss': 0.044183190912008286, 'train/grad_norm_pre_clip': 5.23176383972168, 'timing/data': 0.0011040931567549706, 'timing/model': 1.0459355125203729, 'learning_rate/qwen_vl_interface': 3.6674757510121255e-06, 'learning_rate/action_model': 3.6674757510121255e-05, 'epoch': 1.57}) +[2026-06-18 01:37:07,611][starVLA.training.train_starvla][INFO] - Step 4297, Loss: {'train/loss': 0.296827107667923, 'train/grad_norm_pre_clip': 14.323225975036621, 'timing/data': 0.0008800923824310303, 'timing/model': 1.4113517766818404, 'learning_rate/qwen_vl_interface': 3.6654367806609615e-06, 'learning_rate/action_model': 3.665436780660961e-05, 'epoch': 1.57}) +[2026-06-18 01:37:12,363][starVLA.training.train_starvla][INFO] - Step 4298, Loss: {'train/loss': 0.4215930998325348, 'train/grad_norm_pre_clip': 19.430526733398438, 'timing/data': 0.001123126596212387, 'timing/model': 1.6918828580528498, 'learning_rate/qwen_vl_interface': 3.6633981387914597e-06, 'learning_rate/action_model': 3.66339813879146e-05, 'epoch': 1.57}) +[2026-06-18 01:37:16,569][starVLA.training.train_starvla][INFO] - Step 4299, Loss: {'train/loss': 0.5082482099533081, 'train/grad_norm_pre_clip': 18.24788475036621, 'timing/data': 0.0009799245744943619, 'timing/model': 1.1171415522694588, 'learning_rate/qwen_vl_interface': 3.6613598258262295e-06, 'learning_rate/action_model': 3.6613598258262296e-05, 'epoch': 1.57}) +[2026-06-18 01:44:11,778][starVLA.training.train_starvla][INFO] - Step 4300, Loss: {'train/loss': 0.4982856810092926, 'train/grad_norm_pre_clip': 28.197616577148438, 'eval/loss': 0.5510176420211792, 'eval/latency_2/loss': 0.5510176420211792, 'eval/demon_fire_precision': 0.9154090548054011, 'eval/demon_fire_recall': 0.9644351464435147, 'eval/demon_fire_f1': 0.9392828035859822, 'eval/demon_fire_rate_model': 0.7006121313299944, 'eval/demon_fire_rate_teacher': 0.6649972175848636, 'eval/demon_left_precision': 0.8826040554962646, 'eval/demon_left_recall': 0.8705263157894737, 'eval/demon_left_f1': 0.8765235824059354, 'eval/demon_left_rate_model': 0.2607122982749026, 'eval/demon_left_rate_teacher': 0.2643294379521425, 'eval/demon_right_precision': 0.9191176470588235, 'eval/demon_right_recall': 0.847457627118644, 'eval/demon_right_f1': 0.8818342151675486, 'eval/demon_right_rate_model': 0.2270450751252087, 'eval/demon_right_rate_teacher': 0.24624373956594323, 'eval/demon_move_f1': 0.8791788987867419, 'eval/demon_cc_f1': 0.9204060992065799, 'eval/latency_2/demon_fire_precision': 0.9154090548054011, 'eval/latency_2/demon_fire_recall': 0.9644351464435147, 'eval/latency_2/demon_fire_f1': 0.9392828035859822, 'eval/latency_2/demon_fire_rate_model': 0.7006121313299944, 'eval/latency_2/demon_fire_rate_teacher': 0.6649972175848636, 'eval/latency_2/demon_left_precision': 0.8826040554962646, 'eval/latency_2/demon_left_recall': 0.8705263157894737, 'eval/latency_2/demon_left_f1': 0.8765235824059354, 'eval/latency_2/demon_left_rate_model': 0.2607122982749026, 'eval/latency_2/demon_left_rate_teacher': 0.2643294379521425, 'eval/latency_2/demon_right_precision': 0.9191176470588235, 'eval/latency_2/demon_right_recall': 0.847457627118644, 'eval/latency_2/demon_right_f1': 0.8818342151675486, 'eval/latency_2/demon_right_rate_model': 0.2270450751252087, 'eval/latency_2/demon_right_rate_teacher': 0.24624373956594323, 'eval/latency_2/demon_move_f1': 0.8791788987867419, 'eval/latency_2/demon_cc_f1': 0.9204060992065799, 'timing/data': 0.0008730748668313026, 'timing/model': 4.9107736283913255, 'learning_rate/qwen_vl_interface': 3.6593218421878157e-06, 'learning_rate/action_model': 3.659321842187816e-05, 'epoch': 1.57}) +[2026-06-18 01:44:11,792][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_4300_state +[2026-06-18 01:44:11,800][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 01:44:20,311][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_4300_state/model.safetensors +[2026-06-18 01:44:36,114][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_4300_state/optimizer.bin +[2026-06-18 01:44:36,116][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_4300_state/sampler.bin +[2026-06-18 01:44:36,116][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_4300_state/sampler_1.bin +[2026-06-18 01:44:36,118][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_4300_state/random_states_0.pkl +[2026-06-18 01:44:36,136][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4300 +[2026-06-18 01:44:36,137][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 01:44:36,154][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 01:44:41,486][starVLA.training.train_starvla][INFO] - Step 4301, Loss: {'train/loss': 0.4371320307254791, 'train/grad_norm_pre_clip': 26.04520606994629, 'timing/data': 0.0010025529190897942, 'timing/model': 1.0244703134521842, 'learning_rate/qwen_vl_interface': 3.657284188298697e-06, 'learning_rate/action_model': 3.657284188298697e-05, 'epoch': 1.57}) +[2026-06-18 01:44:54,167][starVLA.training.train_starvla][INFO] - Step 4302, Loss: {'train/loss': 0.3741501569747925, 'train/grad_norm_pre_clip': 17.016340255737305, 'timing/data': 0.0011639194563031197, 'timing/model': 4.9725476717576385, 'learning_rate/qwen_vl_interface': 3.65524686458128e-06, 'learning_rate/action_model': 3.65524686458128e-05, 'epoch': 1.58}) +[2026-06-18 01:45:03,954][starVLA.training.train_starvla][INFO] - Step 4303, Loss: {'train/loss': 0.35184890031814575, 'train/grad_norm_pre_clip': 20.263404846191406, 'timing/data': 0.000960751436650753, 'timing/model': 1.3850120548158884, 'learning_rate/qwen_vl_interface': 3.653209871457904e-06, 'learning_rate/action_model': 3.653209871457904e-05, 'epoch': 1.58}) +[2026-06-18 01:45:08,603][starVLA.training.train_starvla][INFO] - Step 4304, Loss: {'train/loss': 0.4117966890335083, 'train/grad_norm_pre_clip': 19.956674575805664, 'timing/data': 0.0010229609906673431, 'timing/model': 1.3385666776448488, 'learning_rate/qwen_vl_interface': 3.6511732093508376e-06, 'learning_rate/action_model': 3.651173209350838e-05, 'epoch': 1.58}) +[2026-06-18 01:45:13,586][starVLA.training.train_starvla][INFO] - Step 4305, Loss: {'train/loss': 0.2609686851501465, 'train/grad_norm_pre_clip': 15.831461906433105, 'timing/data': 0.0013626497238874435, 'timing/model': 1.3281542444601655, 'learning_rate/qwen_vl_interface': 3.6491368786822885e-06, 'learning_rate/action_model': 3.6491368786822884e-05, 'epoch': 1.58}) +[2026-06-18 01:45:18,116][starVLA.training.train_starvla][INFO] - Step 4306, Loss: {'train/loss': 0.415982723236084, 'train/grad_norm_pre_clip': 21.58070945739746, 'timing/data': 0.0009126793593168259, 'timing/model': 1.0521607175469398, 'learning_rate/qwen_vl_interface': 3.647100879874386e-06, 'learning_rate/action_model': 3.6471008798743855e-05, 'epoch': 1.58}) +[2026-06-18 01:45:22,624][starVLA.training.train_starvla][INFO] - Step 4307, Loss: {'train/loss': 0.5217244625091553, 'train/grad_norm_pre_clip': 25.736948013305664, 'timing/data': 0.0008662296459078789, 'timing/model': 1.113963259384036, 'learning_rate/qwen_vl_interface': 3.645065213349195e-06, 'learning_rate/action_model': 3.645065213349195e-05, 'epoch': 1.58}) +[2026-06-18 01:45:27,607][starVLA.training.train_starvla][INFO] - Step 4308, Loss: {'train/loss': 0.35951414704322815, 'train/grad_norm_pre_clip': 19.51512908935547, 'timing/data': 0.0007775677368044853, 'timing/model': 1.0816046930849552, 'learning_rate/qwen_vl_interface': 3.6430298795287135e-06, 'learning_rate/action_model': 3.643029879528713e-05, 'epoch': 1.58}) +[2026-06-18 01:45:32,077][starVLA.training.train_starvla][INFO] - Step 4309, Loss: {'train/loss': 0.8130205869674683, 'train/grad_norm_pre_clip': 27.97991371154785, 'timing/data': 0.0008936244994401932, 'timing/model': 1.3776048989966512, 'learning_rate/qwen_vl_interface': 3.640994878834866e-06, 'learning_rate/action_model': 3.640994878834866e-05, 'epoch': 1.58}) +[2026-06-18 01:45:36,309][starVLA.training.train_starvla][INFO] - Step 4310, Loss: {'train/loss': 0.3020474910736084, 'train/grad_norm_pre_clip': 18.3008975982666, 'timing/data': 0.0009175324812531471, 'timing/model': 1.135712668299675, 'learning_rate/qwen_vl_interface': 3.6389602116895136e-06, 'learning_rate/action_model': 3.6389602116895134e-05, 'epoch': 1.58}) +[2026-06-18 01:45:42,343][starVLA.training.train_starvla][INFO] - Step 4311, Loss: {'train/loss': 0.39477890729904175, 'train/grad_norm_pre_clip': 26.738664627075195, 'timing/data': 0.0013512028381228447, 'timing/model': 2.5423724660649896, 'learning_rate/qwen_vl_interface': 3.6369258785144395e-06, 'learning_rate/action_model': 3.636925878514439e-05, 'epoch': 1.58}) +[2026-06-18 01:45:46,920][starVLA.training.train_starvla][INFO] - Step 4312, Loss: {'train/loss': 0.3029083013534546, 'train/grad_norm_pre_clip': 18.840105056762695, 'timing/data': 0.0007612220942974091, 'timing/model': 1.0544830625876784, 'learning_rate/qwen_vl_interface': 3.6348918797313693e-06, 'learning_rate/action_model': 3.634891879731369e-05, 'epoch': 1.58}) +[2026-06-18 01:45:51,396][starVLA.training.train_starvla][INFO] - Step 4313, Loss: {'train/loss': 0.34457844495773315, 'train/grad_norm_pre_clip': 18.868648529052734, 'timing/data': 0.000853697769343853, 'timing/model': 1.0510635282844305, 'learning_rate/qwen_vl_interface': 3.6328582157619495e-06, 'learning_rate/action_model': 3.6328582157619496e-05, 'epoch': 1.58}) +[2026-06-18 01:46:01,229][starVLA.training.train_starvla][INFO] - Step 4314, Loss: {'train/loss': 0.2779920995235443, 'train/grad_norm_pre_clip': 19.132652282714844, 'timing/data': 0.0010007340461015701, 'timing/model': 3.246930302120745, 'learning_rate/qwen_vl_interface': 3.6308248870277613e-06, 'learning_rate/action_model': 3.630824887027761e-05, 'epoch': 1.58}) +[2026-06-18 01:46:06,252][starVLA.training.train_starvla][INFO] - Step 4315, Loss: {'train/loss': 0.3537396788597107, 'train/grad_norm_pre_clip': 18.599092483520508, 'timing/data': 0.0010303156450390816, 'timing/model': 1.2016003234311938, 'learning_rate/qwen_vl_interface': 3.628791893950316e-06, 'learning_rate/action_model': 3.628791893950316e-05, 'epoch': 1.58}) +[2026-06-18 01:46:11,377][starVLA.training.train_starvla][INFO] - Step 4316, Loss: {'train/loss': 0.22584226727485657, 'train/grad_norm_pre_clip': 15.651529312133789, 'timing/data': 0.0010493798181414604, 'timing/model': 1.0570724504068494, 'learning_rate/qwen_vl_interface': 3.6267592369510563e-06, 'learning_rate/action_model': 3.6267592369510564e-05, 'epoch': 1.58}) +[2026-06-18 01:46:19,202][starVLA.training.train_starvla][INFO] - Step 4317, Loss: {'train/loss': 0.1991560310125351, 'train/grad_norm_pre_clip': 15.108928680419922, 'timing/data': 0.0007812026888132095, 'timing/model': 1.0814337497577071, 'learning_rate/qwen_vl_interface': 3.624726916451355e-06, 'learning_rate/action_model': 3.624726916451355e-05, 'epoch': 1.58}) +[2026-06-18 01:46:26,128][starVLA.training.train_starvla][INFO] - Step 4318, Loss: {'train/loss': 0.26874616742134094, 'train/grad_norm_pre_clip': 19.35006332397461, 'timing/data': 0.0008753323927521706, 'timing/model': 1.4289848888292909, 'learning_rate/qwen_vl_interface': 3.622694932872509e-06, 'learning_rate/action_model': 3.622694932872509e-05, 'epoch': 1.58}) +[2026-06-18 01:46:34,476][starVLA.training.train_starvla][INFO] - Step 4319, Loss: {'train/loss': 0.630393385887146, 'train/grad_norm_pre_clip': 27.142290115356445, 'timing/data': 0.002987920306622982, 'timing/model': 1.5712699890136719, 'learning_rate/qwen_vl_interface': 3.620663286635758e-06, 'learning_rate/action_model': 3.620663286635758e-05, 'epoch': 1.58}) +[2026-06-18 01:46:38,702][starVLA.training.train_starvla][INFO] - Step 4320, Loss: {'train/loss': 0.4155842065811157, 'train/grad_norm_pre_clip': 19.95366096496582, 'timing/data': 0.0015119249001145363, 'timing/model': 1.0598921505734324, 'learning_rate/qwen_vl_interface': 3.618631978162259e-06, 'learning_rate/action_model': 3.618631978162259e-05, 'epoch': 1.58}) +[2026-06-18 01:46:44,125][starVLA.training.train_starvla][INFO] - Step 4321, Loss: {'train/loss': 0.2904666066169739, 'train/grad_norm_pre_clip': 27.875328063964844, 'timing/data': 0.0009922012686729431, 'timing/model': 1.074015324935317, 'learning_rate/qwen_vl_interface': 3.6166010078731073e-06, 'learning_rate/action_model': 3.6166010078731075e-05, 'epoch': 1.58}) +[2026-06-18 01:46:54,842][starVLA.training.train_starvla][INFO] - Step 4322, Loss: {'train/loss': 0.3371475338935852, 'train/grad_norm_pre_clip': 15.431045532226562, 'timing/data': 0.0009645568206906319, 'timing/model': 2.5492355544120073, 'learning_rate/qwen_vl_interface': 3.614570376189323e-06, 'learning_rate/action_model': 3.614570376189323e-05, 'epoch': 1.58}) +[2026-06-18 01:46:59,635][starVLA.training.train_starvla][INFO] - Step 4323, Loss: {'train/loss': 0.28126102685928345, 'train/grad_norm_pre_clip': 20.11967658996582, 'timing/data': 0.006926191970705986, 'timing/model': 1.1283962950110435, 'learning_rate/qwen_vl_interface': 3.6125400835318604e-06, 'learning_rate/action_model': 3.61254008353186e-05, 'epoch': 1.58}) +[2026-06-18 01:47:04,890][starVLA.training.train_starvla][INFO] - Step 4324, Loss: {'train/loss': 0.2381303608417511, 'train/grad_norm_pre_clip': 18.773378372192383, 'timing/data': 0.001438225619494915, 'timing/model': 1.3361012525856495, 'learning_rate/qwen_vl_interface': 3.6105101303216016e-06, 'learning_rate/action_model': 3.610510130321601e-05, 'epoch': 1.58}) +[2026-06-18 01:47:11,134][starVLA.training.train_starvla][INFO] - Step 4325, Loss: {'train/loss': 0.22271621227264404, 'train/grad_norm_pre_clip': 14.946732521057129, 'timing/data': 0.0011050738394260406, 'timing/model': 2.8816291354596615, 'learning_rate/qwen_vl_interface': 3.608480516979354e-06, 'learning_rate/action_model': 3.608480516979354e-05, 'epoch': 1.58}) +[2026-06-18 01:47:15,902][starVLA.training.train_starvla][INFO] - Step 4326, Loss: {'train/loss': 0.6193338632583618, 'train/grad_norm_pre_clip': 26.109079360961914, 'timing/data': 0.0011869166046380997, 'timing/model': 1.0607916936278343, 'learning_rate/qwen_vl_interface': 3.6064512439258646e-06, 'learning_rate/action_model': 3.606451243925865e-05, 'epoch': 1.58}) +[2026-06-18 01:47:20,511][starVLA.training.train_starvla][INFO] - Step 4327, Loss: {'train/loss': 0.6036733984947205, 'train/grad_norm_pre_clip': 21.658632278442383, 'timing/data': 0.0009048571810126305, 'timing/model': 1.0590077815577388, 'learning_rate/qwen_vl_interface': 3.6044223115818e-06, 'learning_rate/action_model': 3.6044223115818e-05, 'epoch': 1.58}) +[2026-06-18 01:47:25,735][starVLA.training.train_starvla][INFO] - Step 4328, Loss: {'train/loss': 0.12577152252197266, 'train/grad_norm_pre_clip': 9.622694969177246, 'timing/data': 0.0009341947734355927, 'timing/model': 1.653916111215949, 'learning_rate/qwen_vl_interface': 3.6023937203677606e-06, 'learning_rate/action_model': 3.6023937203677604e-05, 'epoch': 1.58}) +[2026-06-18 01:47:30,236][starVLA.training.train_starvla][INFO] - Step 4329, Loss: {'train/loss': 0.4782125949859619, 'train/grad_norm_pre_clip': 19.71375846862793, 'timing/data': 0.0010708915069699287, 'timing/model': 1.4308097269386053, 'learning_rate/qwen_vl_interface': 3.6003654707042746e-06, 'learning_rate/action_model': 3.6003654707042745e-05, 'epoch': 1.59}) +[2026-06-18 01:47:34,534][starVLA.training.train_starvla][INFO] - Step 4330, Loss: {'train/loss': 0.3834066390991211, 'train/grad_norm_pre_clip': 21.314468383789062, 'timing/data': 0.0009524282068014145, 'timing/model': 1.0816073715686798, 'learning_rate/qwen_vl_interface': 3.5983375630118033e-06, 'learning_rate/action_model': 3.598337563011803e-05, 'epoch': 1.59}) +[2026-06-18 01:47:39,120][starVLA.training.train_starvla][INFO] - Step 4331, Loss: {'train/loss': 0.232076495885849, 'train/grad_norm_pre_clip': 18.468969345092773, 'timing/data': 0.002346678636968136, 'timing/model': 1.0777553776279092, 'learning_rate/qwen_vl_interface': 3.596309997710733e-06, 'learning_rate/action_model': 3.596309997710733e-05, 'epoch': 1.59}) +[2026-06-18 01:47:43,816][starVLA.training.train_starvla][INFO] - Step 4332, Loss: {'train/loss': 0.6126355528831482, 'train/grad_norm_pre_clip': 27.73733901977539, 'timing/data': 0.0008512465283274651, 'timing/model': 1.2059881584718823, 'learning_rate/qwen_vl_interface': 3.5942827752213788e-06, 'learning_rate/action_model': 3.594282775221379e-05, 'epoch': 1.59}) +[2026-06-18 01:47:48,587][starVLA.training.train_starvla][INFO] - Step 4333, Loss: {'train/loss': 0.1825079321861267, 'train/grad_norm_pre_clip': 12.3714599609375, 'timing/data': 0.0007751798257231712, 'timing/model': 1.0615818379446864, 'learning_rate/qwen_vl_interface': 3.592255895963986e-06, 'learning_rate/action_model': 3.592255895963986e-05, 'epoch': 1.59}) +[2026-06-18 01:47:54,653][starVLA.training.train_starvla][INFO] - Step 4334, Loss: {'train/loss': 0.7757992744445801, 'train/grad_norm_pre_clip': 26.005950927734375, 'timing/data': 0.0032697105780243874, 'timing/model': 2.912392063997686, 'learning_rate/qwen_vl_interface': 3.590229360358731e-06, 'learning_rate/action_model': 3.590229360358731e-05, 'epoch': 1.59}) +[2026-06-18 01:48:01,210][starVLA.training.train_starvla][INFO] - Step 4335, Loss: {'train/loss': 0.4771631360054016, 'train/grad_norm_pre_clip': 24.18606948852539, 'timing/data': 0.0008803261443972588, 'timing/model': 1.0584772611036897, 'learning_rate/qwen_vl_interface': 3.5882031688257147e-06, 'learning_rate/action_model': 3.588203168825715e-05, 'epoch': 1.59}) +[2026-06-18 01:48:09,078][starVLA.training.train_starvla][INFO] - Step 4336, Loss: {'train/loss': 0.46050989627838135, 'train/grad_norm_pre_clip': 22.446575164794922, 'timing/data': 0.0025790268555283546, 'timing/model': 1.0794646311551332, 'learning_rate/qwen_vl_interface': 3.5861773217849694e-06, 'learning_rate/action_model': 3.5861773217849693e-05, 'epoch': 1.59}) +[2026-06-18 01:48:13,519][starVLA.training.train_starvla][INFO] - Step 4337, Loss: {'train/loss': 0.19167490303516388, 'train/grad_norm_pre_clip': 16.369888305664062, 'timing/data': 0.001929297111928463, 'timing/model': 1.0994077725335956, 'learning_rate/qwen_vl_interface': 3.5841518196564566e-06, 'learning_rate/action_model': 3.584151819656457e-05, 'epoch': 1.59}) +[2026-06-18 01:48:18,832][starVLA.training.train_starvla][INFO] - Step 4338, Loss: {'train/loss': 0.5587064027786255, 'train/grad_norm_pre_clip': 23.222951889038086, 'timing/data': 0.0009688474237918854, 'timing/model': 1.1693544974550605, 'learning_rate/qwen_vl_interface': 3.582126662860065e-06, 'learning_rate/action_model': 3.582126662860065e-05, 'epoch': 1.59}) +[2026-06-18 01:48:23,231][starVLA.training.train_starvla][INFO] - Step 4339, Loss: {'train/loss': 0.6718306541442871, 'train/grad_norm_pre_clip': 28.460771560668945, 'timing/data': 0.0014222217723727226, 'timing/model': 1.326493963599205, 'learning_rate/qwen_vl_interface': 3.5801018518156096e-06, 'learning_rate/action_model': 3.5801018518156096e-05, 'epoch': 1.59}) +[2026-06-18 01:48:29,037][starVLA.training.train_starvla][INFO] - Step 4340, Loss: {'train/loss': 0.47857359051704407, 'train/grad_norm_pre_clip': 25.222734451293945, 'timing/data': 0.0012377481907606125, 'timing/model': 1.1960745630785823, 'learning_rate/qwen_vl_interface': 3.5780773869428366e-06, 'learning_rate/action_model': 3.5780773869428364e-05, 'epoch': 1.59}) +[2026-06-18 01:48:33,521][starVLA.training.train_starvla][INFO] - Step 4341, Loss: {'train/loss': 0.37664884328842163, 'train/grad_norm_pre_clip': 26.42463493347168, 'timing/data': 0.0011821668595075607, 'timing/model': 1.1042576553300023, 'learning_rate/qwen_vl_interface': 3.576053268661422e-06, 'learning_rate/action_model': 3.5760532686614216e-05, 'epoch': 1.59}) +[2026-06-18 01:48:38,367][starVLA.training.train_starvla][INFO] - Step 4342, Loss: {'train/loss': 0.6775060892105103, 'train/grad_norm_pre_clip': 25.50657844543457, 'timing/data': 0.0014373892918229103, 'timing/model': 1.0823547262698412, 'learning_rate/qwen_vl_interface': 3.5740294973909643e-06, 'learning_rate/action_model': 3.5740294973909644e-05, 'epoch': 1.59}) +[2026-06-18 01:48:43,023][starVLA.training.train_starvla][INFO] - Step 4343, Loss: {'train/loss': 0.6483737230300903, 'train/grad_norm_pre_clip': 18.214143753051758, 'timing/data': 0.0008558286353945732, 'timing/model': 1.2366482811048627, 'learning_rate/qwen_vl_interface': 3.5720060735509976e-06, 'learning_rate/action_model': 3.5720060735509975e-05, 'epoch': 1.59}) +[2026-06-18 01:48:48,316][starVLA.training.train_starvla][INFO] - Step 4344, Loss: {'train/loss': 0.346196711063385, 'train/grad_norm_pre_clip': 20.21457862854004, 'timing/data': 0.0008763214573264122, 'timing/model': 1.387232918292284, 'learning_rate/qwen_vl_interface': 3.569982997560974e-06, 'learning_rate/action_model': 3.5699829975609736e-05, 'epoch': 1.59}) +[2026-06-18 01:48:52,377][starVLA.training.train_starvla][INFO] - Step 4345, Loss: {'train/loss': 0.507888674736023, 'train/grad_norm_pre_clip': 27.09551239013672, 'timing/data': 0.0008498262614011765, 'timing/model': 1.0505230678245425, 'learning_rate/qwen_vl_interface': 3.567960269840286e-06, 'learning_rate/action_model': 3.567960269840286e-05, 'epoch': 1.59}) +[2026-06-18 01:49:07,223][starVLA.training.train_starvla][INFO] - Step 4346, Loss: {'train/loss': 0.5348919034004211, 'train/grad_norm_pre_clip': 18.418880462646484, 'timing/data': 0.0013477839529514313, 'timing/model': 5.321251465007663, 'learning_rate/qwen_vl_interface': 3.5659378908082427e-06, 'learning_rate/action_model': 3.5659378908082426e-05, 'epoch': 1.59}) +[2026-06-18 01:49:17,573][starVLA.training.train_starvla][INFO] - Step 4347, Loss: {'train/loss': 0.3789074718952179, 'train/grad_norm_pre_clip': 19.829025268554688, 'timing/data': 0.0012766961008310318, 'timing/model': 1.2901876717805862, 'learning_rate/qwen_vl_interface': 3.5639158608840862e-06, 'learning_rate/action_model': 3.563915860884086e-05, 'epoch': 1.59}) +[2026-06-18 01:49:22,362][starVLA.training.train_starvla][INFO] - Step 4348, Loss: {'train/loss': 0.5079262256622314, 'train/grad_norm_pre_clip': 55.099552154541016, 'timing/data': 0.001134670339524746, 'timing/model': 1.4366452870890498, 'learning_rate/qwen_vl_interface': 3.5618941804869866e-06, 'learning_rate/action_model': 3.561894180486987e-05, 'epoch': 1.59}) +[2026-06-18 01:49:26,856][starVLA.training.train_starvla][INFO] - Step 4349, Loss: {'train/loss': 0.5881465077400208, 'train/grad_norm_pre_clip': 23.465614318847656, 'timing/data': 0.0020338958129286766, 'timing/model': 1.409913819283247, 'learning_rate/qwen_vl_interface': 3.5598728500360395e-06, 'learning_rate/action_model': 3.55987285003604e-05, 'epoch': 1.59}) +[2026-06-18 01:49:31,511][starVLA.training.train_starvla][INFO] - Step 4350, Loss: {'train/loss': 0.3424190580844879, 'train/grad_norm_pre_clip': 23.0864315032959, 'timing/data': 0.001139315776526928, 'timing/model': 1.1651657354086637, 'learning_rate/qwen_vl_interface': 3.557851869950272e-06, 'learning_rate/action_model': 3.557851869950272e-05, 'epoch': 1.59}) +[2026-06-18 01:49:39,148][starVLA.training.train_starvla][INFO] - Step 4351, Loss: {'train/loss': 0.20590613782405853, 'train/grad_norm_pre_clip': 13.418474197387695, 'timing/data': 0.06649747584015131, 'timing/model': 2.4205084117129445, 'learning_rate/qwen_vl_interface': 3.555831240648628e-06, 'learning_rate/action_model': 3.5558312406486274e-05, 'epoch': 1.59}) +[2026-06-18 01:49:43,789][starVLA.training.train_starvla][INFO] - Step 4352, Loss: {'train/loss': 0.24279850721359253, 'train/grad_norm_pre_clip': 16.630239486694336, 'timing/data': 0.00089291762560606, 'timing/model': 1.0478122113272548, 'learning_rate/qwen_vl_interface': 3.553810962549994e-06, 'learning_rate/action_model': 3.553810962549994e-05, 'epoch': 1.59}) +[2026-06-18 01:49:48,651][starVLA.training.train_starvla][INFO] - Step 4353, Loss: {'train/loss': 0.12180343270301819, 'train/grad_norm_pre_clip': 9.003525733947754, 'timing/data': 0.0018759006634354591, 'timing/model': 1.0400538006797433, 'learning_rate/qwen_vl_interface': 3.5517910360731723e-06, 'learning_rate/action_model': 3.551791036073172e-05, 'epoch': 1.59}) +[2026-06-18 01:49:53,318][starVLA.training.train_starvla][INFO] - Step 4354, Loss: {'train/loss': 0.4224790036678314, 'train/grad_norm_pre_clip': 25.387845993041992, 'timing/data': 0.001223396509885788, 'timing/model': 1.5533625958487391, 'learning_rate/qwen_vl_interface': 3.549771461636896e-06, 'learning_rate/action_model': 3.549771461636896e-05, 'epoch': 1.59}) +[2026-06-18 01:50:00,342][starVLA.training.train_starvla][INFO] - Step 4355, Loss: {'train/loss': 0.14151832461357117, 'train/grad_norm_pre_clip': 13.548469543457031, 'timing/data': 0.0009866682812571526, 'timing/model': 2.2920697359368205, 'learning_rate/qwen_vl_interface': 3.5477522396598236e-06, 'learning_rate/action_model': 3.5477522396598235e-05, 'epoch': 1.59}) +[2026-06-18 01:50:07,780][starVLA.training.train_starvla][INFO] - Step 4356, Loss: {'train/loss': 0.2627243101596832, 'train/grad_norm_pre_clip': 12.639459609985352, 'timing/data': 0.0012100804597139359, 'timing/model': 1.1628728993237019, 'learning_rate/qwen_vl_interface': 3.5457333705605456e-06, 'learning_rate/action_model': 3.5457333705605456e-05, 'epoch': 1.6}) +[2026-06-18 01:50:12,641][starVLA.training.train_starvla][INFO] - Step 4357, Loss: {'train/loss': 0.5586504340171814, 'train/grad_norm_pre_clip': 19.471126556396484, 'timing/data': 0.0014318814501166344, 'timing/model': 1.1197218550369143, 'learning_rate/qwen_vl_interface': 3.543714854757574e-06, 'learning_rate/action_model': 3.5437148547575736e-05, 'epoch': 1.6}) +[2026-06-18 01:50:17,537][starVLA.training.train_starvla][INFO] - Step 4358, Loss: {'train/loss': 0.3903617560863495, 'train/grad_norm_pre_clip': 26.316871643066406, 'timing/data': 0.0010188445448875427, 'timing/model': 1.6775327874347568, 'learning_rate/qwen_vl_interface': 3.5416966926693464e-06, 'learning_rate/action_model': 3.5416966926693465e-05, 'epoch': 1.6}) +[2026-06-18 01:50:22,438][starVLA.training.train_starvla][INFO] - Step 4359, Loss: {'train/loss': 0.2564849257469177, 'train/grad_norm_pre_clip': 15.648161888122559, 'timing/data': 0.001114468090236187, 'timing/model': 1.4869641438126564, 'learning_rate/qwen_vl_interface': 3.539678884714235e-06, 'learning_rate/action_model': 3.539678884714235e-05, 'epoch': 1.6}) +[2026-06-18 01:50:31,002][starVLA.training.train_starvla][INFO] - Step 4360, Loss: {'train/loss': 0.29293307662010193, 'train/grad_norm_pre_clip': 13.892216682434082, 'timing/data': 0.001093805767595768, 'timing/model': 5.504807655699551, 'learning_rate/qwen_vl_interface': 3.5376614313105307e-06, 'learning_rate/action_model': 3.53766143131053e-05, 'epoch': 1.6}) +[2026-06-18 01:50:36,841][starVLA.training.train_starvla][INFO] - Step 4361, Loss: {'train/loss': 0.4482150673866272, 'train/grad_norm_pre_clip': 26.68227767944336, 'timing/data': 0.00582304410636425, 'timing/model': 2.021560345776379, 'learning_rate/qwen_vl_interface': 3.5356443328764546e-06, 'learning_rate/action_model': 3.5356443328764546e-05, 'epoch': 1.6}) +[2026-06-18 01:50:47,745][starVLA.training.train_starvla][INFO] - Step 4362, Loss: {'train/loss': 0.31210651993751526, 'train/grad_norm_pre_clip': 15.600314140319824, 'timing/data': 0.0011814478784799576, 'timing/model': 1.0647939061746001, 'learning_rate/qwen_vl_interface': 3.5336275898301515e-06, 'learning_rate/action_model': 3.5336275898301514e-05, 'epoch': 1.6}) +[2026-06-18 01:50:52,410][starVLA.training.train_starvla][INFO] - Step 4363, Loss: {'train/loss': 0.38615208864212036, 'train/grad_norm_pre_clip': 23.062782287597656, 'timing/data': 0.0013321908190846443, 'timing/model': 1.1187994135543704, 'learning_rate/qwen_vl_interface': 3.531611202589697e-06, 'learning_rate/action_model': 3.531611202589697e-05, 'epoch': 1.6}) +[2026-06-18 01:50:56,876][starVLA.training.train_starvla][INFO] - Step 4364, Loss: {'train/loss': 0.29996126890182495, 'train/grad_norm_pre_clip': 63.828975677490234, 'timing/data': 0.001094968058168888, 'timing/model': 1.4183696024119854, 'learning_rate/qwen_vl_interface': 3.5295951715730902e-06, 'learning_rate/action_model': 3.52959517157309e-05, 'epoch': 1.6}) +[2026-06-18 01:51:03,778][starVLA.training.train_starvla][INFO] - Step 4365, Loss: {'train/loss': 0.38861969113349915, 'train/grad_norm_pre_clip': 15.79984188079834, 'timing/data': 0.001109314151108265, 'timing/model': 3.1231820406392217, 'learning_rate/qwen_vl_interface': 3.527579497198255e-06, 'learning_rate/action_model': 3.527579497198255e-05, 'epoch': 1.6}) +[2026-06-18 01:51:08,372][starVLA.training.train_starvla][INFO] - Step 4366, Loss: {'train/loss': 0.2125052809715271, 'train/grad_norm_pre_clip': 14.673830032348633, 'timing/data': 0.001065247692167759, 'timing/model': 1.0748894158750772, 'learning_rate/qwen_vl_interface': 3.525564179883042e-06, 'learning_rate/action_model': 3.525564179883042e-05, 'epoch': 1.6}) +[2026-06-18 01:51:13,237][starVLA.training.train_starvla][INFO] - Step 4367, Loss: {'train/loss': 0.1284741908311844, 'train/grad_norm_pre_clip': 12.686272621154785, 'timing/data': 0.0008895797654986382, 'timing/model': 1.025599175132811, 'learning_rate/qwen_vl_interface': 3.5235492200452308e-06, 'learning_rate/action_model': 3.523549220045231e-05, 'epoch': 1.6}) +[2026-06-18 01:51:19,320][starVLA.training.train_starvla][INFO] - Step 4368, Loss: {'train/loss': 0.404634028673172, 'train/grad_norm_pre_clip': 17.09906578063965, 'timing/data': 0.0008773254230618477, 'timing/model': 1.1323232678696513, 'learning_rate/qwen_vl_interface': 3.521534618102525e-06, 'learning_rate/action_model': 3.5215346181025246e-05, 'epoch': 1.6}) +[2026-06-18 01:51:24,222][starVLA.training.train_starvla][INFO] - Step 4369, Loss: {'train/loss': 0.379731684923172, 'train/grad_norm_pre_clip': 20.88608741760254, 'timing/data': 0.001202072948217392, 'timing/model': 1.6390045089647174, 'learning_rate/qwen_vl_interface': 3.5195203744725516e-06, 'learning_rate/action_model': 3.5195203744725516e-05, 'epoch': 1.6}) +[2026-06-18 01:51:29,214][starVLA.training.train_starvla][INFO] - Step 4370, Loss: {'train/loss': 0.41262295842170715, 'train/grad_norm_pre_clip': 24.870304107666016, 'timing/data': 0.0010380465537309647, 'timing/model': 1.8189636450260878, 'learning_rate/qwen_vl_interface': 3.5175064895728665e-06, 'learning_rate/action_model': 3.5175064895728664e-05, 'epoch': 1.6}) +[2026-06-18 01:51:34,821][starVLA.training.train_starvla][INFO] - Step 4371, Loss: {'train/loss': 0.16616606712341309, 'train/grad_norm_pre_clip': 9.995412826538086, 'timing/data': 0.0010740319266915321, 'timing/model': 1.3586872071027756, 'learning_rate/qwen_vl_interface': 3.515492963820951e-06, 'learning_rate/action_model': 3.515492963820951e-05, 'epoch': 1.6}) +[2026-06-18 01:51:39,381][starVLA.training.train_starvla][INFO] - Step 4372, Loss: {'train/loss': 0.5787919759750366, 'train/grad_norm_pre_clip': 27.485647201538086, 'timing/data': 0.0007687155157327652, 'timing/model': 1.0421054288744926, 'learning_rate/qwen_vl_interface': 3.5134797976342123e-06, 'learning_rate/action_model': 3.513479797634212e-05, 'epoch': 1.6}) +[2026-06-18 01:51:44,006][starVLA.training.train_starvla][INFO] - Step 4373, Loss: {'train/loss': 0.2218448668718338, 'train/grad_norm_pre_clip': 13.188920021057129, 'timing/data': 0.0011220285668969154, 'timing/model': 1.4067186554893851, 'learning_rate/qwen_vl_interface': 3.5114669914299776e-06, 'learning_rate/action_model': 3.511466991429978e-05, 'epoch': 1.6}) +[2026-06-18 01:51:48,295][starVLA.training.train_starvla][INFO] - Step 4374, Loss: {'train/loss': 0.2863039970397949, 'train/grad_norm_pre_clip': 16.062780380249023, 'timing/data': 0.0010994020849466324, 'timing/model': 1.2786138504743576, 'learning_rate/qwen_vl_interface': 3.5094545456255065e-06, 'learning_rate/action_model': 3.509454545625506e-05, 'epoch': 1.6}) +[2026-06-18 01:51:52,710][starVLA.training.train_starvla][INFO] - Step 4375, Loss: {'train/loss': 0.4598614275455475, 'train/grad_norm_pre_clip': 38.9704704284668, 'timing/data': 0.0012497548013925552, 'timing/model': 1.3224271219223738, 'learning_rate/qwen_vl_interface': 3.5074424606379826e-06, 'learning_rate/action_model': 3.5074424606379824e-05, 'epoch': 1.6}) +[2026-06-18 01:51:57,170][starVLA.training.train_starvla][INFO] - Step 4376, Loss: {'train/loss': 0.3750976026058197, 'train/grad_norm_pre_clip': 24.281747817993164, 'timing/data': 0.0011877743527293205, 'timing/model': 1.0374719081446528, 'learning_rate/qwen_vl_interface': 3.505430736884511e-06, 'learning_rate/action_model': 3.5054307368845105e-05, 'epoch': 1.6}) +[2026-06-18 01:52:01,756][starVLA.training.train_starvla][INFO] - Step 4377, Loss: {'train/loss': 0.20356892049312592, 'train/grad_norm_pre_clip': 11.306703567504883, 'timing/data': 0.0007419558241963387, 'timing/model': 1.0399909391999245, 'learning_rate/qwen_vl_interface': 3.503419374782123e-06, 'learning_rate/action_model': 3.503419374782123e-05, 'epoch': 1.6}) +[2026-06-18 01:52:11,508][starVLA.training.train_starvla][INFO] - Step 4378, Loss: {'train/loss': 0.5408077836036682, 'train/grad_norm_pre_clip': 30.577617645263672, 'timing/data': 0.0013070469722151756, 'timing/model': 2.433622038923204, 'learning_rate/qwen_vl_interface': 3.50140837474778e-06, 'learning_rate/action_model': 3.50140837474778e-05, 'epoch': 1.6}) +[2026-06-18 01:52:16,250][starVLA.training.train_starvla][INFO] - Step 4379, Loss: {'train/loss': 0.37951284646987915, 'train/grad_norm_pre_clip': 18.1197509765625, 'timing/data': 0.0008739763870835304, 'timing/model': 1.4527214569970965, 'learning_rate/qwen_vl_interface': 3.4993977371983625e-06, 'learning_rate/action_model': 3.499397737198362e-05, 'epoch': 1.6}) +[2026-06-18 01:52:20,283][starVLA.training.train_starvla][INFO] - Step 4380, Loss: {'train/loss': 0.2912103235721588, 'train/grad_norm_pre_clip': 25.680002212524414, 'timing/data': 0.0012525711208581924, 'timing/model': 1.0291641876101494, 'learning_rate/qwen_vl_interface': 3.497387462550676e-06, 'learning_rate/action_model': 3.4973874625506757e-05, 'epoch': 1.6}) +[2026-06-18 01:52:24,705][starVLA.training.train_starvla][INFO] - Step 4381, Loss: {'train/loss': 0.5714329481124878, 'train/grad_norm_pre_clip': 21.080337524414062, 'timing/data': 0.0011987388134002686, 'timing/model': 1.078523158095777, 'learning_rate/qwen_vl_interface': 3.4953775512214568e-06, 'learning_rate/action_model': 3.495377551221457e-05, 'epoch': 1.6}) +[2026-06-18 01:52:29,333][starVLA.training.train_starvla][INFO] - Step 4382, Loss: {'train/loss': 0.30204227566719055, 'train/grad_norm_pre_clip': 18.481094360351562, 'timing/data': 0.00104597769677639, 'timing/model': 1.0728395963087678, 'learning_rate/qwen_vl_interface': 3.4933680036273583e-06, 'learning_rate/action_model': 3.493368003627358e-05, 'epoch': 1.6}) +[2026-06-18 01:52:33,756][starVLA.training.train_starvla][INFO] - Step 4383, Loss: {'train/loss': 0.39191877841949463, 'train/grad_norm_pre_clip': 16.885967254638672, 'timing/data': 0.0007775276899337769, 'timing/model': 1.0761659247800708, 'learning_rate/qwen_vl_interface': 3.491358820184963e-06, 'learning_rate/action_model': 3.491358820184963e-05, 'epoch': 1.6}) +[2026-06-18 01:52:43,371][starVLA.training.train_starvla][INFO] - Step 4384, Loss: {'train/loss': 0.22861404716968536, 'train/grad_norm_pre_clip': 19.420806884765625, 'timing/data': 0.0011113109067082405, 'timing/model': 5.591770618222654, 'learning_rate/qwen_vl_interface': 3.4893500013107758e-06, 'learning_rate/action_model': 3.4893500013107754e-05, 'epoch': 1.61}) +[2026-06-18 01:52:47,604][starVLA.training.train_starvla][INFO] - Step 4385, Loss: {'train/loss': 0.47557324171066284, 'train/grad_norm_pre_clip': 21.5135440826416, 'timing/data': 0.001161457970738411, 'timing/model': 1.0623725410550833, 'learning_rate/qwen_vl_interface': 3.4873415474212287e-06, 'learning_rate/action_model': 3.487341547421228e-05, 'epoch': 1.61}) +[2026-06-18 01:52:53,564][starVLA.training.train_starvla][INFO] - Step 4386, Loss: {'train/loss': 0.5086069703102112, 'train/grad_norm_pre_clip': 23.603275299072266, 'timing/data': 0.0008683148771524429, 'timing/model': 1.0516921328380704, 'learning_rate/qwen_vl_interface': 3.4853334589326764e-06, 'learning_rate/action_model': 3.4853334589326765e-05, 'epoch': 1.61}) +[2026-06-18 01:52:59,917][starVLA.training.train_starvla][INFO] - Step 4387, Loss: {'train/loss': 0.5211921334266663, 'train/grad_norm_pre_clip': 26.40369415283203, 'timing/data': 0.002452107146382332, 'timing/model': 2.930556159466505, 'learning_rate/qwen_vl_interface': 3.483325736261394e-06, 'learning_rate/action_model': 3.483325736261394e-05, 'epoch': 1.61}) +[2026-06-18 01:53:05,708][starVLA.training.train_starvla][INFO] - Step 4388, Loss: {'train/loss': 0.3552103042602539, 'train/grad_norm_pre_clip': 19.184463500976562, 'timing/data': 0.0009717531502246857, 'timing/model': 1.0919146379455924, 'learning_rate/qwen_vl_interface': 3.48131837982359e-06, 'learning_rate/action_model': 3.48131837982359e-05, 'epoch': 1.61}) +[2026-06-18 01:53:13,018][starVLA.training.train_starvla][INFO] - Step 4389, Loss: {'train/loss': 0.4785970449447632, 'train/grad_norm_pre_clip': 15.221541404724121, 'timing/data': 0.0008828872814774513, 'timing/model': 3.9998982353135943, 'learning_rate/qwen_vl_interface': 3.4793113900353885e-06, 'learning_rate/action_model': 3.479311390035389e-05, 'epoch': 1.61}) +[2026-06-18 01:53:19,724][starVLA.training.train_starvla][INFO] - Step 4390, Loss: {'train/loss': 0.2371326982975006, 'train/grad_norm_pre_clip': 15.742931365966797, 'timing/data': 0.0007927510887384415, 'timing/model': 1.1068440293893218, 'learning_rate/qwen_vl_interface': 3.4773047673128412e-06, 'learning_rate/action_model': 3.477304767312841e-05, 'epoch': 1.61}) +[2026-06-18 01:53:29,214][starVLA.training.train_starvla][INFO] - Step 4391, Loss: {'train/loss': 0.1937866359949112, 'train/grad_norm_pre_clip': 13.542441368103027, 'timing/data': 0.0011609476059675217, 'timing/model': 1.2479474991559982, 'learning_rate/qwen_vl_interface': 3.4752985120719225e-06, 'learning_rate/action_model': 3.475298512071922e-05, 'epoch': 1.61}) +[2026-06-18 01:53:37,331][starVLA.training.train_starvla][INFO] - Step 4392, Loss: {'train/loss': 0.5624270439147949, 'train/grad_norm_pre_clip': 22.05453109741211, 'timing/data': 0.0011346060782670975, 'timing/model': 1.053505108691752, 'learning_rate/qwen_vl_interface': 3.4732926247285324e-06, 'learning_rate/action_model': 3.4732926247285325e-05, 'epoch': 1.61}) +[2026-06-18 01:53:42,226][starVLA.training.train_starvla][INFO] - Step 4393, Loss: {'train/loss': 0.3047560453414917, 'train/grad_norm_pre_clip': 13.294150352478027, 'timing/data': 0.0008721118792891502, 'timing/model': 1.1695447713136673, 'learning_rate/qwen_vl_interface': 3.471287105698494e-06, 'learning_rate/action_model': 3.471287105698494e-05, 'epoch': 1.61}) +[2026-06-18 01:53:53,280][starVLA.training.train_starvla][INFO] - Step 4394, Loss: {'train/loss': 0.6939866542816162, 'train/grad_norm_pre_clip': 29.51156997680664, 'timing/data': 0.0011124443262815475, 'timing/model': 1.4299912536516786, 'learning_rate/qwen_vl_interface': 3.4692819553975514e-06, 'learning_rate/action_model': 3.469281955397551e-05, 'epoch': 1.61}) +[2026-06-18 01:53:57,463][starVLA.training.train_starvla][INFO] - Step 4395, Loss: {'train/loss': 0.48969903588294983, 'train/grad_norm_pre_clip': 16.0660343170166, 'timing/data': 0.0013223066926002502, 'timing/model': 1.027791059575975, 'learning_rate/qwen_vl_interface': 3.4672771742413754e-06, 'learning_rate/action_model': 3.467277174241375e-05, 'epoch': 1.61}) +[2026-06-18 01:54:02,026][starVLA.training.train_starvla][INFO] - Step 4396, Loss: {'train/loss': 0.3744169771671295, 'train/grad_norm_pre_clip': 21.897789001464844, 'timing/data': 0.005262930877506733, 'timing/model': 1.116937750019133, 'learning_rate/qwen_vl_interface': 3.46527276264556e-06, 'learning_rate/action_model': 3.46527276264556e-05, 'epoch': 1.61}) +[2026-06-18 01:54:06,494][starVLA.training.train_starvla][INFO] - Step 4397, Loss: {'train/loss': 0.4275665581226349, 'train/grad_norm_pre_clip': 18.872867584228516, 'timing/data': 0.0012211874127388, 'timing/model': 1.0389463640749454, 'learning_rate/qwen_vl_interface': 3.4632687210256218e-06, 'learning_rate/action_model': 3.463268721025621e-05, 'epoch': 1.61}) +[2026-06-18 01:54:10,852][starVLA.training.train_starvla][INFO] - Step 4398, Loss: {'train/loss': 0.14747121930122375, 'train/grad_norm_pre_clip': 8.27894115447998, 'timing/data': 0.0010941596701741219, 'timing/model': 1.0635952902957797, 'learning_rate/qwen_vl_interface': 3.461265049796999e-06, 'learning_rate/action_model': 3.461265049796999e-05, 'epoch': 1.61}) +[2026-06-18 01:54:15,246][starVLA.training.train_starvla][INFO] - Step 4399, Loss: {'train/loss': 0.22898729145526886, 'train/grad_norm_pre_clip': 15.33182430267334, 'timing/data': 0.0011931415647268295, 'timing/model': 1.3163202926516533, 'learning_rate/qwen_vl_interface': 3.459261749375057e-06, 'learning_rate/action_model': 3.4592617493750567e-05, 'epoch': 1.61}) +[2026-06-18 01:57:32,069][starVLA.training.train_starvla][INFO] - Step 4400, Loss: {'train/loss': 0.34688347578048706, 'train/grad_norm_pre_clip': 19.26847267150879, 'eval/loss': 0.5616344213485718, 'eval/latency_2/loss': 0.5616344213485718, 'eval/demon_fire_precision': 0.9267503035208418, 'eval/demon_fire_recall': 0.9581589958158996, 'eval/demon_fire_f1': 0.9421929644106151, 'eval/demon_fire_rate_model': 0.6875347801892042, 'eval/demon_fire_rate_teacher': 0.6649972175848636, 'eval/demon_left_precision': 0.8597748208802457, 'eval/demon_left_recall': 0.8842105263157894, 'eval/demon_left_f1': 0.8718214841722886, 'eval/demon_left_rate_model': 0.27184195882025597, 'eval/demon_left_rate_teacher': 0.2643294379521425, 'eval/demon_right_precision': 0.8897911832946636, 'eval/demon_right_recall': 0.8666666666666667, 'eval/demon_right_f1': 0.8780767029192902, 'eval/demon_right_rate_model': 0.23984418475236505, 'eval/demon_right_rate_teacher': 0.24624373956594323, 'eval/demon_move_f1': 0.8749490935457893, 'eval/demon_cc_f1': 0.9209589815590683, 'eval/latency_2/demon_fire_precision': 0.9267503035208418, 'eval/latency_2/demon_fire_recall': 0.9581589958158996, 'eval/latency_2/demon_fire_f1': 0.9421929644106151, 'eval/latency_2/demon_fire_rate_model': 0.6875347801892042, 'eval/latency_2/demon_fire_rate_teacher': 0.6649972175848636, 'eval/latency_2/demon_left_precision': 0.8597748208802457, 'eval/latency_2/demon_left_recall': 0.8842105263157894, 'eval/latency_2/demon_left_f1': 0.8718214841722886, 'eval/latency_2/demon_left_rate_model': 0.27184195882025597, 'eval/latency_2/demon_left_rate_teacher': 0.2643294379521425, 'eval/latency_2/demon_right_precision': 0.8897911832946636, 'eval/latency_2/demon_right_recall': 0.8666666666666667, 'eval/latency_2/demon_right_f1': 0.8780767029192902, 'eval/latency_2/demon_right_rate_model': 0.23984418475236505, 'eval/latency_2/demon_right_rate_teacher': 0.24624373956594323, 'eval/latency_2/demon_move_f1': 0.8749490935457893, 'eval/latency_2/demon_cc_f1': 0.9209589815590683, 'timing/data': 0.0011751949787139893, 'timing/model': 1.036523905582726, 'learning_rate/qwen_vl_interface': 3.457258820175082e-06, 'learning_rate/action_model': 3.457258820175082e-05, 'epoch': 1.61}) +[2026-06-18 01:57:32,071][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_4400_state +[2026-06-18 01:57:32,080][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 01:57:40,808][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_4400_state/model.safetensors +[2026-06-18 01:57:55,720][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_4400_state/optimizer.bin +[2026-06-18 01:57:55,721][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_4400_state/sampler.bin +[2026-06-18 01:57:55,722][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_4400_state/sampler_1.bin +[2026-06-18 01:57:55,724][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_4400_state/random_states_0.pkl +[2026-06-18 01:57:55,728][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4400 +[2026-06-18 01:57:55,728][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 01:57:55,745][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 01:58:05,045][starVLA.training.train_starvla][INFO] - Step 4401, Loss: {'train/loss': 0.35655730962753296, 'train/grad_norm_pre_clip': 21.457073211669922, 'timing/data': 0.0009387964382767677, 'timing/model': 5.814046552404761, 'learning_rate/qwen_vl_interface': 3.4552562626122817e-06, 'learning_rate/action_model': 3.455256262612281e-05, 'epoch': 1.61}) +[2026-06-18 01:58:16,232][starVLA.training.train_starvla][INFO] - Step 4402, Loss: {'train/loss': 0.4659983813762665, 'train/grad_norm_pre_clip': 16.46014976501465, 'timing/data': 0.0008875690400600433, 'timing/model': 1.2466849321499467, 'learning_rate/qwen_vl_interface': 3.453254077101789e-06, 'learning_rate/action_model': 3.453254077101789e-05, 'epoch': 1.61}) +[2026-06-18 01:58:20,991][starVLA.training.train_starvla][INFO] - Step 4403, Loss: {'train/loss': 0.6240025162696838, 'train/grad_norm_pre_clip': 27.06946563720703, 'timing/data': 0.0010459553450345993, 'timing/model': 1.0416334299370646, 'learning_rate/qwen_vl_interface': 3.45125226405866e-06, 'learning_rate/action_model': 3.4512522640586595e-05, 'epoch': 1.61}) +[2026-06-18 01:58:25,703][starVLA.training.train_starvla][INFO] - Step 4404, Loss: {'train/loss': 0.3188488185405731, 'train/grad_norm_pre_clip': 13.391989707946777, 'timing/data': 0.0029784180223941803, 'timing/model': 1.0927188945934176, 'learning_rate/qwen_vl_interface': 3.449250823897872e-06, 'learning_rate/action_model': 3.449250823897872e-05, 'epoch': 1.61}) +[2026-06-18 01:58:39,943][starVLA.training.train_starvla][INFO] - Step 4405, Loss: {'train/loss': 0.5221070647239685, 'train/grad_norm_pre_clip': 33.99509811401367, 'timing/data': 0.0010492829605937004, 'timing/model': 1.4937964510172606, 'learning_rate/qwen_vl_interface': 3.447249757034328e-06, 'learning_rate/action_model': 3.4472497570343276e-05, 'epoch': 1.61}) +[2026-06-18 01:58:44,903][starVLA.training.train_starvla][INFO] - Step 4406, Loss: {'train/loss': 0.19791539013385773, 'train/grad_norm_pre_clip': 11.480237007141113, 'timing/data': 0.0007663946598768234, 'timing/model': 1.0351989679038525, 'learning_rate/qwen_vl_interface': 3.445249063882845e-06, 'learning_rate/action_model': 3.445249063882845e-05, 'epoch': 1.61}) +[2026-06-18 01:58:49,619][starVLA.training.train_starvla][INFO] - Step 4407, Loss: {'train/loss': 0.24791625142097473, 'train/grad_norm_pre_clip': 18.106708526611328, 'timing/data': 0.0008875783532857895, 'timing/model': 1.294651542790234, 'learning_rate/qwen_vl_interface': 3.4432487448581762e-06, 'learning_rate/action_model': 3.443248744858176e-05, 'epoch': 1.61}) +[2026-06-18 01:58:59,741][starVLA.training.train_starvla][INFO] - Step 4408, Loss: {'train/loss': 0.4522666335105896, 'train/grad_norm_pre_clip': 20.1296329498291, 'timing/data': 0.0015732496976852417, 'timing/model': 1.3630575546994805, 'learning_rate/qwen_vl_interface': 3.4412488003749867e-06, 'learning_rate/action_model': 3.4412488003749866e-05, 'epoch': 1.61}) +[2026-06-18 01:59:06,843][starVLA.training.train_starvla][INFO] - Step 4409, Loss: {'train/loss': 0.3785293400287628, 'train/grad_norm_pre_clip': 29.783361434936523, 'timing/data': 0.00209187064319849, 'timing/model': 1.1048714518547058, 'learning_rate/qwen_vl_interface': 3.4392492308478657e-06, 'learning_rate/action_model': 3.439249230847866e-05, 'epoch': 1.61}) +[2026-06-18 01:59:12,319][starVLA.training.train_starvla][INFO] - Step 4410, Loss: {'train/loss': 0.24149096012115479, 'train/grad_norm_pre_clip': 21.57709312438965, 'timing/data': 0.0012756912037730217, 'timing/model': 1.6301667457446456, 'learning_rate/qwen_vl_interface': 3.437250036691328e-06, 'learning_rate/action_model': 3.437250036691328e-05, 'epoch': 1.61}) +[2026-06-18 01:59:16,492][starVLA.training.train_starvla][INFO] - Step 4411, Loss: {'train/loss': 0.35927635431289673, 'train/grad_norm_pre_clip': 19.270408630371094, 'timing/data': 0.0012167096138000488, 'timing/model': 1.0580968894064426, 'learning_rate/qwen_vl_interface': 3.4352512183198094e-06, 'learning_rate/action_model': 3.4352512183198095e-05, 'epoch': 1.62}) +[2026-06-18 01:59:21,662][starVLA.training.train_starvla][INFO] - Step 4412, Loss: {'train/loss': 0.3011247217655182, 'train/grad_norm_pre_clip': 15.426661491394043, 'timing/data': 0.0011548614129424095, 'timing/model': 1.127533657476306, 'learning_rate/qwen_vl_interface': 3.4332527761476662e-06, 'learning_rate/action_model': 3.433252776147666e-05, 'epoch': 1.62}) +[2026-06-18 01:59:26,506][starVLA.training.train_starvla][INFO] - Step 4413, Loss: {'train/loss': 0.2563904821872711, 'train/grad_norm_pre_clip': 14.097970008850098, 'timing/data': 0.0016082711517810822, 'timing/model': 1.0494744423776865, 'learning_rate/qwen_vl_interface': 3.431254710589174e-06, 'learning_rate/action_model': 3.431254710589174e-05, 'epoch': 1.62}) +[2026-06-18 01:59:31,135][starVLA.training.train_starvla][INFO] - Step 4414, Loss: {'train/loss': 0.2392093539237976, 'train/grad_norm_pre_clip': 15.418445587158203, 'timing/data': 0.0009210240095853806, 'timing/model': 1.0828975047916174, 'learning_rate/qwen_vl_interface': 3.4292570220585407e-06, 'learning_rate/action_model': 3.4292570220585407e-05, 'epoch': 1.62}) +[2026-06-18 01:59:36,991][starVLA.training.train_starvla][INFO] - Step 4415, Loss: {'train/loss': 0.3144818842411041, 'train/grad_norm_pre_clip': 20.05562973022461, 'timing/data': 0.001205943524837494, 'timing/model': 1.4681436531245708, 'learning_rate/qwen_vl_interface': 3.4272597109698857e-06, 'learning_rate/action_model': 3.427259710969886e-05, 'epoch': 1.62}) +[2026-06-18 01:59:42,126][starVLA.training.train_starvla][INFO] - Step 4416, Loss: {'train/loss': 0.6490344405174255, 'train/grad_norm_pre_clip': 27.8026123046875, 'timing/data': 0.0013188384473323822, 'timing/model': 1.1218637824058533, 'learning_rate/qwen_vl_interface': 3.4252627777372533e-06, 'learning_rate/action_model': 3.425262777737253e-05, 'epoch': 1.62}) +[2026-06-18 01:59:49,203][starVLA.training.train_starvla][INFO] - Step 4417, Loss: {'train/loss': 0.5990326404571533, 'train/grad_norm_pre_clip': 34.98917770385742, 'timing/data': 0.0013285232707858086, 'timing/model': 1.1120201274752617, 'learning_rate/qwen_vl_interface': 3.4232662227746095e-06, 'learning_rate/action_model': 3.423266222774609e-05, 'epoch': 1.62}) +[2026-06-18 01:59:54,614][starVLA.training.train_starvla][INFO] - Step 4418, Loss: {'train/loss': 0.6655411124229431, 'train/grad_norm_pre_clip': 23.300106048583984, 'timing/data': 0.0008597411215305328, 'timing/model': 1.6321931509301066, 'learning_rate/qwen_vl_interface': 3.4212700464958448e-06, 'learning_rate/action_model': 3.421270046495845e-05, 'epoch': 1.62}) +[2026-06-18 01:59:59,214][starVLA.training.train_starvla][INFO] - Step 4419, Loss: {'train/loss': 0.35417047142982483, 'train/grad_norm_pre_clip': 19.42322540283203, 'timing/data': 0.0010356009006500244, 'timing/model': 1.5424664141610265, 'learning_rate/qwen_vl_interface': 3.4192742493147686e-06, 'learning_rate/action_model': 3.419274249314768e-05, 'epoch': 1.62}) +[2026-06-18 02:00:03,752][starVLA.training.train_starvla][INFO] - Step 4420, Loss: {'train/loss': 0.43307843804359436, 'train/grad_norm_pre_clip': 18.141006469726562, 'timing/data': 0.001521882601082325, 'timing/model': 1.186907796189189, 'learning_rate/qwen_vl_interface': 3.4172788316451065e-06, 'learning_rate/action_model': 3.417278831645107e-05, 'epoch': 1.62}) +[2026-06-18 02:00:13,031][starVLA.training.train_starvla][INFO] - Step 4421, Loss: {'train/loss': 0.5957042574882507, 'train/grad_norm_pre_clip': 22.58991241455078, 'timing/data': 0.0039033107459545135, 'timing/model': 5.106624665670097, 'learning_rate/qwen_vl_interface': 3.4152837939005177e-06, 'learning_rate/action_model': 3.415283793900518e-05, 'epoch': 1.62}) +[2026-06-18 02:00:17,505][starVLA.training.train_starvla][INFO] - Step 4422, Loss: {'train/loss': 0.3513883948326111, 'train/grad_norm_pre_clip': 18.258697509765625, 'timing/data': 0.0012955134734511375, 'timing/model': 1.0978039037436247, 'learning_rate/qwen_vl_interface': 3.413289136494572e-06, 'learning_rate/action_model': 3.413289136494572e-05, 'epoch': 1.62}) +[2026-06-18 02:00:24,537][starVLA.training.train_starvla][INFO] - Step 4423, Loss: {'train/loss': 0.5250203609466553, 'train/grad_norm_pre_clip': 21.215240478515625, 'timing/data': 0.0010402360931038857, 'timing/model': 2.8274605683982372, 'learning_rate/qwen_vl_interface': 3.411294859840765e-06, 'learning_rate/action_model': 3.411294859840764e-05, 'epoch': 1.62}) +[2026-06-18 02:00:32,820][starVLA.training.train_starvla][INFO] - Step 4424, Loss: {'train/loss': 0.42189255356788635, 'train/grad_norm_pre_clip': 21.1790771484375, 'timing/data': 0.0009506922215223312, 'timing/model': 1.5860174465924501, 'learning_rate/qwen_vl_interface': 3.409300964352511e-06, 'learning_rate/action_model': 3.409300964352511e-05, 'epoch': 1.62}) +[2026-06-18 02:00:37,546][starVLA.training.train_starvla][INFO] - Step 4425, Loss: {'train/loss': 0.10894625633955002, 'train/grad_norm_pre_clip': 8.922616004943848, 'timing/data': 0.0009536240249872208, 'timing/model': 1.464877074584365, 'learning_rate/qwen_vl_interface': 3.407307450443148e-06, 'learning_rate/action_model': 3.407307450443148e-05, 'epoch': 1.62}) +[2026-06-18 02:00:42,868][starVLA.training.train_starvla][INFO] - Step 4426, Loss: {'train/loss': 0.3201679289340973, 'train/grad_norm_pre_clip': 16.197811126708984, 'timing/data': 0.0010081194341182709, 'timing/model': 1.1883274596184492, 'learning_rate/qwen_vl_interface': 3.4053143185259353e-06, 'learning_rate/action_model': 3.405314318525935e-05, 'epoch': 1.62}) +[2026-06-18 02:00:47,723][starVLA.training.train_starvla][INFO] - Step 4427, Loss: {'train/loss': 0.3356744349002838, 'train/grad_norm_pre_clip': 15.889867782592773, 'timing/data': 0.001269567757844925, 'timing/model': 1.1565564554184675, 'learning_rate/qwen_vl_interface': 3.4033215690140475e-06, 'learning_rate/action_model': 3.4033215690140476e-05, 'epoch': 1.62}) +[2026-06-18 02:00:52,914][starVLA.training.train_starvla][INFO] - Step 4428, Loss: {'train/loss': 0.39436447620391846, 'train/grad_norm_pre_clip': 22.49557876586914, 'timing/data': 0.004693366587162018, 'timing/model': 1.4013449978083372, 'learning_rate/qwen_vl_interface': 3.4013292023205852e-06, 'learning_rate/action_model': 3.401329202320585e-05, 'epoch': 1.62}) +[2026-06-18 02:00:57,808][starVLA.training.train_starvla][INFO] - Step 4429, Loss: {'train/loss': 0.5942831635475159, 'train/grad_norm_pre_clip': 25.94702911376953, 'timing/data': 0.0016659628599882126, 'timing/model': 1.2386999893933535, 'learning_rate/qwen_vl_interface': 3.3993372188585694e-06, 'learning_rate/action_model': 3.3993372188585694e-05, 'epoch': 1.62}) +[2026-06-18 02:01:04,085][starVLA.training.train_starvla][INFO] - Step 4430, Loss: {'train/loss': 0.6062543392181396, 'train/grad_norm_pre_clip': 30.984712600708008, 'timing/data': 0.0017247600480914116, 'timing/model': 1.5436750929802656, 'learning_rate/qwen_vl_interface': 3.3973456190409397e-06, 'learning_rate/action_model': 3.39734561904094e-05, 'epoch': 1.62}) +[2026-06-18 02:01:08,912][starVLA.training.train_starvla][INFO] - Step 4431, Loss: {'train/loss': 0.427991658449173, 'train/grad_norm_pre_clip': 16.49638557434082, 'timing/data': 0.0013126591220498085, 'timing/model': 1.0803928719833493, 'learning_rate/qwen_vl_interface': 3.395354403280555e-06, 'learning_rate/action_model': 3.395354403280555e-05, 'epoch': 1.62}) +[2026-06-18 02:01:20,018][starVLA.training.train_starvla][INFO] - Step 4432, Loss: {'train/loss': 0.3591878414154053, 'train/grad_norm_pre_clip': 16.763805389404297, 'timing/data': 0.055106024257838726, 'timing/model': 5.308110294863582, 'learning_rate/qwen_vl_interface': 3.3933635719902e-06, 'learning_rate/action_model': 3.3933635719901995e-05, 'epoch': 1.62}) +[2026-06-18 02:01:28,949][starVLA.training.train_starvla][INFO] - Step 4433, Loss: {'train/loss': 0.5102276802062988, 'train/grad_norm_pre_clip': 21.650484085083008, 'timing/data': 0.018061595037579536, 'timing/model': 1.3647186597809196, 'learning_rate/qwen_vl_interface': 3.3913731255825754e-06, 'learning_rate/action_model': 3.3913731255825754e-05, 'epoch': 1.62}) +[2026-06-18 02:01:38,482][starVLA.training.train_starvla][INFO] - Step 4434, Loss: {'train/loss': 0.32958918809890747, 'train/grad_norm_pre_clip': 15.40587043762207, 'timing/data': 0.013972800225019455, 'timing/model': 1.24654346331954, 'learning_rate/qwen_vl_interface': 3.3893830644703003e-06, 'learning_rate/action_model': 3.3893830644703e-05, 'epoch': 1.62}) +[2026-06-18 02:01:43,252][starVLA.training.train_starvla][INFO] - Step 4435, Loss: {'train/loss': 0.3312721848487854, 'train/grad_norm_pre_clip': 18.65802574157715, 'timing/data': 0.000887545756995678, 'timing/model': 1.6755692046135664, 'learning_rate/qwen_vl_interface': 3.3873933890659174e-06, 'learning_rate/action_model': 3.3873933890659174e-05, 'epoch': 1.62}) +[2026-06-18 02:01:48,305][starVLA.training.train_starvla][INFO] - Step 4436, Loss: {'train/loss': 0.8694475889205933, 'train/grad_norm_pre_clip': 38.00984191894531, 'timing/data': 0.0014195535331964493, 'timing/model': 1.8711847504600883, 'learning_rate/qwen_vl_interface': 3.3854040997818897e-06, 'learning_rate/action_model': 3.38540409978189e-05, 'epoch': 1.62}) +[2026-06-18 02:01:53,254][starVLA.training.train_starvla][INFO] - Step 4437, Loss: {'train/loss': 0.25338152050971985, 'train/grad_norm_pre_clip': 17.55312728881836, 'timing/data': 0.0011167768388986588, 'timing/model': 1.1577604804188013, 'learning_rate/qwen_vl_interface': 3.383415197030599e-06, 'learning_rate/action_model': 3.383415197030599e-05, 'epoch': 1.62}) +[2026-06-18 02:01:57,682][starVLA.training.train_starvla][INFO] - Step 4438, Loss: {'train/loss': 0.41098618507385254, 'train/grad_norm_pre_clip': 18.445390701293945, 'timing/data': 0.001078820787370205, 'timing/model': 1.056192398071289, 'learning_rate/qwen_vl_interface': 3.3814266812243473e-06, 'learning_rate/action_model': 3.3814266812243474e-05, 'epoch': 1.63}) +[2026-06-18 02:02:02,541][starVLA.training.train_starvla][INFO] - Step 4439, Loss: {'train/loss': 0.41841650009155273, 'train/grad_norm_pre_clip': 19.677385330200195, 'timing/data': 0.001010773703455925, 'timing/model': 1.3531744088977575, 'learning_rate/qwen_vl_interface': 3.3794385527753522e-06, 'learning_rate/action_model': 3.379438552775352e-05, 'epoch': 1.63}) +[2026-06-18 02:02:06,779][starVLA.training.train_starvla][INFO] - Step 4440, Loss: {'train/loss': 0.4861041307449341, 'train/grad_norm_pre_clip': 18.837677001953125, 'timing/data': 0.0011035874485969543, 'timing/model': 1.0696887839585543, 'learning_rate/qwen_vl_interface': 3.3774508120957606e-06, 'learning_rate/action_model': 3.37745081209576e-05, 'epoch': 1.63}) +[2026-06-18 02:02:12,541][starVLA.training.train_starvla][INFO] - Step 4441, Loss: {'train/loss': 0.4176803231239319, 'train/grad_norm_pre_clip': 13.858582496643066, 'timing/data': 0.0007390799000859261, 'timing/model': 2.114485797472298, 'learning_rate/qwen_vl_interface': 3.37546345959763e-06, 'learning_rate/action_model': 3.3754634595976295e-05, 'epoch': 1.63}) +[2026-06-18 02:02:17,218][starVLA.training.train_starvla][INFO] - Step 4442, Loss: {'train/loss': 0.34309178590774536, 'train/grad_norm_pre_clip': 21.71454620361328, 'timing/data': 0.0011137770488858223, 'timing/model': 1.0753214498981833, 'learning_rate/qwen_vl_interface': 3.3734764956929407e-06, 'learning_rate/action_model': 3.37347649569294e-05, 'epoch': 1.63}) +[2026-06-18 02:02:21,647][starVLA.training.train_starvla][INFO] - Step 4443, Loss: {'train/loss': 0.4025906026363373, 'train/grad_norm_pre_clip': 20.26055335998535, 'timing/data': 0.0010014614090323448, 'timing/model': 1.0956947728991508, 'learning_rate/qwen_vl_interface': 3.3714899207935937e-06, 'learning_rate/action_model': 3.371489920793594e-05, 'epoch': 1.63}) +[2026-06-18 02:02:26,106][starVLA.training.train_starvla][INFO] - Step 4444, Loss: {'train/loss': 0.16643913090229034, 'train/grad_norm_pre_clip': 14.745135307312012, 'timing/data': 0.0013922052457928658, 'timing/model': 1.0690875994041562, 'learning_rate/qwen_vl_interface': 3.369503735311407e-06, 'learning_rate/action_model': 3.369503735311407e-05, 'epoch': 1.63}) +[2026-06-18 02:02:30,476][starVLA.training.train_starvla][INFO] - Step 4445, Loss: {'train/loss': 0.19803813099861145, 'train/grad_norm_pre_clip': 26.78886604309082, 'timing/data': 0.0013973908498883247, 'timing/model': 1.1855667298659682, 'learning_rate/qwen_vl_interface': 3.3675179396581208e-06, 'learning_rate/action_model': 3.367517939658121e-05, 'epoch': 1.63}) +[2026-06-18 02:02:38,526][starVLA.training.train_starvla][INFO] - Step 4446, Loss: {'train/loss': 0.3645611107349396, 'train/grad_norm_pre_clip': 22.484790802001953, 'timing/data': 0.0008733682334423065, 'timing/model': 1.0839226050302386, 'learning_rate/qwen_vl_interface': 3.365532534245389e-06, 'learning_rate/action_model': 3.3655325342453894e-05, 'epoch': 1.63}) +[2026-06-18 02:02:44,671][starVLA.training.train_starvla][INFO] - Step 4447, Loss: {'train/loss': 0.2503880560398102, 'train/grad_norm_pre_clip': 13.548218727111816, 'timing/data': 0.000978352501988411, 'timing/model': 1.1239013522863388, 'learning_rate/qwen_vl_interface': 3.3635475194847927e-06, 'learning_rate/action_model': 3.363547519484793e-05, 'epoch': 1.63}) +[2026-06-18 02:02:49,984][starVLA.training.train_starvla][INFO] - Step 4448, Loss: {'train/loss': 0.20624181628227234, 'train/grad_norm_pre_clip': 15.613286972045898, 'timing/data': 0.0007531577721238136, 'timing/model': 1.3689028872177005, 'learning_rate/qwen_vl_interface': 3.361562895787825e-06, 'learning_rate/action_model': 3.361562895787825e-05, 'epoch': 1.63}) +[2026-06-18 02:02:58,795][starVLA.training.train_starvla][INFO] - Step 4449, Loss: {'train/loss': 0.3007062077522278, 'train/grad_norm_pre_clip': 19.840063095092773, 'timing/data': 0.0013898778706789017, 'timing/model': 1.6939651975408196, 'learning_rate/qwen_vl_interface': 3.3595786635659e-06, 'learning_rate/action_model': 3.3595786635659e-05, 'epoch': 1.63}) +[2026-06-18 02:03:03,115][starVLA.training.train_starvla][INFO] - Step 4450, Loss: {'train/loss': 0.6190810203552246, 'train/grad_norm_pre_clip': 29.964754104614258, 'timing/data': 0.0014222487807273865, 'timing/model': 1.1229185815900564, 'learning_rate/qwen_vl_interface': 3.357594823230353e-06, 'learning_rate/action_model': 3.3575948232303525e-05, 'epoch': 1.63}) +[2026-06-18 02:03:11,051][starVLA.training.train_starvla][INFO] - Step 4451, Loss: {'train/loss': 0.3963271677494049, 'train/grad_norm_pre_clip': 24.54275131225586, 'timing/data': 0.0008964324370026588, 'timing/model': 1.1756363939493895, 'learning_rate/qwen_vl_interface': 3.3556113751924357e-06, 'learning_rate/action_model': 3.355611375192436e-05, 'epoch': 1.63}) +[2026-06-18 02:03:15,818][starVLA.training.train_starvla][INFO] - Step 4452, Loss: {'train/loss': 0.3218865394592285, 'train/grad_norm_pre_clip': 18.957855224609375, 'timing/data': 0.0011741314083337784, 'timing/model': 1.1662108032032847, 'learning_rate/qwen_vl_interface': 3.3536283198633183e-06, 'learning_rate/action_model': 3.353628319863318e-05, 'epoch': 1.63}) +[2026-06-18 02:03:25,266][starVLA.training.train_starvla][INFO] - Step 4453, Loss: {'train/loss': 0.3240763545036316, 'train/grad_norm_pre_clip': 18.24398422241211, 'timing/data': 0.005823363550007343, 'timing/model': 4.2473439844325185, 'learning_rate/qwen_vl_interface': 3.3516456576540913e-06, 'learning_rate/action_model': 3.3516456576540914e-05, 'epoch': 1.63}) +[2026-06-18 02:03:36,457][starVLA.training.train_starvla][INFO] - Step 4454, Loss: {'train/loss': 0.2181718498468399, 'train/grad_norm_pre_clip': 18.011138916015625, 'timing/data': 0.0008657239377498627, 'timing/model': 1.0780902188271284, 'learning_rate/qwen_vl_interface': 3.349663388975763e-06, 'learning_rate/action_model': 3.3496633889757625e-05, 'epoch': 1.63}) +[2026-06-18 02:03:41,194][starVLA.training.train_starvla][INFO] - Step 4455, Loss: {'train/loss': 0.14913107454776764, 'train/grad_norm_pre_clip': 10.71877670288086, 'timing/data': 0.0010298630222678185, 'timing/model': 1.3461896320804954, 'learning_rate/qwen_vl_interface': 3.347681514239258e-06, 'learning_rate/action_model': 3.347681514239257e-05, 'epoch': 1.63}) +[2026-06-18 02:03:48,127][starVLA.training.train_starvla][INFO] - Step 4456, Loss: {'train/loss': 0.257568895816803, 'train/grad_norm_pre_clip': 16.2269287109375, 'timing/data': 0.0014053108170628548, 'timing/model': 2.275023940950632, 'learning_rate/qwen_vl_interface': 3.345700033855421e-06, 'learning_rate/action_model': 3.345700033855421e-05, 'epoch': 1.63}) +[2026-06-18 02:04:02,056][starVLA.training.train_starvla][INFO] - Step 4457, Loss: {'train/loss': 0.27680304646492004, 'train/grad_norm_pre_clip': 15.413991928100586, 'timing/data': 0.0008942894637584686, 'timing/model': 5.701812247745693, 'learning_rate/qwen_vl_interface': 3.3437189482350173e-06, 'learning_rate/action_model': 3.343718948235017e-05, 'epoch': 1.63}) +[2026-06-18 02:04:15,083][starVLA.training.train_starvla][INFO] - Step 4458, Loss: {'train/loss': 0.30428600311279297, 'train/grad_norm_pre_clip': 22.139392852783203, 'timing/data': 0.0013078832998871803, 'timing/model': 4.060908675193787, 'learning_rate/qwen_vl_interface': 3.341738257788727e-06, 'learning_rate/action_model': 3.3417382577887266e-05, 'epoch': 1.63}) +[2026-06-18 02:04:20,311][starVLA.training.train_starvla][INFO] - Step 4459, Loss: {'train/loss': 0.2559196949005127, 'train/grad_norm_pre_clip': 13.654064178466797, 'timing/data': 0.0017465660348534584, 'timing/model': 1.0532192084938288, 'learning_rate/qwen_vl_interface': 3.3397579629271487e-06, 'learning_rate/action_model': 3.3397579629271486e-05, 'epoch': 1.63}) +[2026-06-18 02:04:29,126][starVLA.training.train_starvla][INFO] - Step 4460, Loss: {'train/loss': 0.479367733001709, 'train/grad_norm_pre_clip': 27.32196044921875, 'timing/data': 0.001723993569612503, 'timing/model': 2.052410364151001, 'learning_rate/qwen_vl_interface': 3.3377780640607997e-06, 'learning_rate/action_model': 3.3377780640608e-05, 'epoch': 1.63}) +[2026-06-18 02:04:34,119][starVLA.training.train_starvla][INFO] - Step 4461, Loss: {'train/loss': 0.44705185294151306, 'train/grad_norm_pre_clip': 19.698923110961914, 'timing/data': 0.0011171232908964157, 'timing/model': 1.537794473581016, 'learning_rate/qwen_vl_interface': 3.3357985616001166e-06, 'learning_rate/action_model': 3.335798561600116e-05, 'epoch': 1.63}) +[2026-06-18 02:04:39,433][starVLA.training.train_starvla][INFO] - Step 4462, Loss: {'train/loss': 0.3922068476676941, 'train/grad_norm_pre_clip': 22.7334041595459, 'timing/data': 0.0023949528113007545, 'timing/model': 1.5697987843304873, 'learning_rate/qwen_vl_interface': 3.3338194559554524e-06, 'learning_rate/action_model': 3.333819455955453e-05, 'epoch': 1.63}) +[2026-06-18 02:04:44,345][starVLA.training.train_starvla][INFO] - Step 4463, Loss: {'train/loss': 0.16004012525081635, 'train/grad_norm_pre_clip': 15.487088203430176, 'timing/data': 0.0009086532518267632, 'timing/model': 1.121680068783462, 'learning_rate/qwen_vl_interface': 3.3318407475370758e-06, 'learning_rate/action_model': 3.3318407475370755e-05, 'epoch': 1.63}) +[2026-06-18 02:04:50,207][starVLA.training.train_starvla][INFO] - Step 4464, Loss: {'train/loss': 0.3175729513168335, 'train/grad_norm_pre_clip': 17.574748992919922, 'timing/data': 0.0012148339301347733, 'timing/model': 2.210132393054664, 'learning_rate/qwen_vl_interface': 3.329862436755176e-06, 'learning_rate/action_model': 3.329862436755176e-05, 'epoch': 1.63}) +[2026-06-18 02:04:54,772][starVLA.training.train_starvla][INFO] - Step 4465, Loss: {'train/loss': 0.37142515182495117, 'train/grad_norm_pre_clip': 22.137624740600586, 'timing/data': 0.0012770043686032295, 'timing/model': 1.3018446993082762, 'learning_rate/qwen_vl_interface': 3.3278845240198592e-06, 'learning_rate/action_model': 3.3278845240198595e-05, 'epoch': 1.63}) +[2026-06-18 02:04:59,466][starVLA.training.train_starvla][INFO] - Step 4466, Loss: {'train/loss': 0.76557856798172, 'train/grad_norm_pre_clip': 28.527530670166016, 'timing/data': 0.0012324163690209389, 'timing/model': 1.1347876116633415, 'learning_rate/qwen_vl_interface': 3.325907009741149e-06, 'learning_rate/action_model': 3.3259070097411485e-05, 'epoch': 1.64}) +[2026-06-18 02:05:04,254][starVLA.training.train_starvla][INFO] - Step 4467, Loss: {'train/loss': 0.34412962198257446, 'train/grad_norm_pre_clip': 20.003847122192383, 'timing/data': 0.0011904733255505562, 'timing/model': 1.0532821668311954, 'learning_rate/qwen_vl_interface': 3.323929894328987e-06, 'learning_rate/action_model': 3.323929894328987e-05, 'epoch': 1.64}) +[2026-06-18 02:05:08,969][starVLA.training.train_starvla][INFO] - Step 4468, Loss: {'train/loss': 0.16589368879795074, 'train/grad_norm_pre_clip': 18.198347091674805, 'timing/data': 0.0009028837084770203, 'timing/model': 1.1336172400042415, 'learning_rate/qwen_vl_interface': 3.321953178193228e-06, 'learning_rate/action_model': 3.321953178193228e-05, 'epoch': 1.64}) +[2026-06-18 02:05:13,960][starVLA.training.train_starvla][INFO] - Step 4469, Loss: {'train/loss': 0.28117528557777405, 'train/grad_norm_pre_clip': 21.54367446899414, 'timing/data': 0.0008928310126066208, 'timing/model': 1.030895953066647, 'learning_rate/qwen_vl_interface': 3.319976861743653e-06, 'learning_rate/action_model': 3.319976861743653e-05, 'epoch': 1.64}) +[2026-06-18 02:05:18,946][starVLA.training.train_starvla][INFO] - Step 4470, Loss: {'train/loss': 0.23872703313827515, 'train/grad_norm_pre_clip': 17.078718185424805, 'timing/data': 0.001154535450041294, 'timing/model': 1.4812319157645106, 'learning_rate/qwen_vl_interface': 3.3180009453899496e-06, 'learning_rate/action_model': 3.318000945389949e-05, 'epoch': 1.64}) +[2026-06-18 02:05:25,508][starVLA.training.train_starvla][INFO] - Step 4471, Loss: {'train/loss': 0.21000385284423828, 'train/grad_norm_pre_clip': 17.399141311645508, 'timing/data': 0.0013424968346953392, 'timing/model': 1.0570887755602598, 'learning_rate/qwen_vl_interface': 3.3160254295417287e-06, 'learning_rate/action_model': 3.3160254295417285e-05, 'epoch': 1.64}) +[2026-06-18 02:05:33,455][starVLA.training.train_starvla][INFO] - Step 4472, Loss: {'train/loss': 0.3823941946029663, 'train/grad_norm_pre_clip': 18.94805145263672, 'timing/data': 0.0011543482542037964, 'timing/model': 4.47736391518265, 'learning_rate/qwen_vl_interface': 3.314050314608518e-06, 'learning_rate/action_model': 3.314050314608518e-05, 'epoch': 1.64}) +[2026-06-18 02:05:40,166][starVLA.training.train_starvla][INFO] - Step 4473, Loss: {'train/loss': 0.36861467361450195, 'train/grad_norm_pre_clip': 21.26100730895996, 'timing/data': 0.0008564665913581848, 'timing/model': 1.125402582809329, 'learning_rate/qwen_vl_interface': 3.3120756009997606e-06, 'learning_rate/action_model': 3.312075600999761e-05, 'epoch': 1.64}) +[2026-06-18 02:05:45,360][starVLA.training.train_starvla][INFO] - Step 4474, Loss: {'train/loss': 0.2977331578731537, 'train/grad_norm_pre_clip': 18.13793182373047, 'timing/data': 0.0013497425243258476, 'timing/model': 1.2285126596689224, 'learning_rate/qwen_vl_interface': 3.310101289124817e-06, 'learning_rate/action_model': 3.310101289124817e-05, 'epoch': 1.64}) +[2026-06-18 02:05:50,644][starVLA.training.train_starvla][INFO] - Step 4475, Loss: {'train/loss': 0.24322785437107086, 'train/grad_norm_pre_clip': 16.734840393066406, 'timing/data': 0.0020633842796087265, 'timing/model': 1.8850069558247924, 'learning_rate/qwen_vl_interface': 3.3081273793929615e-06, 'learning_rate/action_model': 3.308127379392961e-05, 'epoch': 1.64}) +[2026-06-18 02:05:56,173][starVLA.training.train_starvla][INFO] - Step 4476, Loss: {'train/loss': 0.608684241771698, 'train/grad_norm_pre_clip': 29.861894607543945, 'timing/data': 0.0008989982306957245, 'timing/model': 1.121456685476005, 'learning_rate/qwen_vl_interface': 3.306153872213392e-06, 'learning_rate/action_model': 3.306153872213392e-05, 'epoch': 1.64}) +[2026-06-18 02:06:00,918][starVLA.training.train_starvla][INFO] - Step 4477, Loss: {'train/loss': 0.5121343731880188, 'train/grad_norm_pre_clip': 25.066720962524414, 'timing/data': 0.0009194333106279373, 'timing/model': 1.064118031412363, 'learning_rate/qwen_vl_interface': 3.304180767995215e-06, 'learning_rate/action_model': 3.3041807679952146e-05, 'epoch': 1.64}) +[2026-06-18 02:06:10,836][starVLA.training.train_starvla][INFO] - Step 4478, Loss: {'train/loss': 0.31090405583381653, 'train/grad_norm_pre_clip': 18.42463493347168, 'timing/data': 0.001142597757279873, 'timing/model': 1.323554408736527, 'learning_rate/qwen_vl_interface': 3.3022080671474588e-06, 'learning_rate/action_model': 3.302208067147459e-05, 'epoch': 1.64}) +[2026-06-18 02:06:17,799][starVLA.training.train_starvla][INFO] - Step 4479, Loss: {'train/loss': 0.17118197679519653, 'train/grad_norm_pre_clip': 12.87545108795166, 'timing/data': 0.0014735180884599686, 'timing/model': 3.2300777807831764, 'learning_rate/qwen_vl_interface': 3.300235770079066e-06, 'learning_rate/action_model': 3.300235770079066e-05, 'epoch': 1.64}) +[2026-06-18 02:06:21,982][starVLA.training.train_starvla][INFO] - Step 4480, Loss: {'train/loss': 0.5300269722938538, 'train/grad_norm_pre_clip': 20.56183624267578, 'timing/data': 0.0007853927090764046, 'timing/model': 1.107061198912561, 'learning_rate/qwen_vl_interface': 3.298263877198896e-06, 'learning_rate/action_model': 3.298263877198896e-05, 'epoch': 1.64}) +[2026-06-18 02:06:26,430][starVLA.training.train_starvla][INFO] - Step 4481, Loss: {'train/loss': 0.3197774589061737, 'train/grad_norm_pre_clip': 15.479750633239746, 'timing/data': 0.00139645766466856, 'timing/model': 1.1105417348444462, 'learning_rate/qwen_vl_interface': 3.296292388915726e-06, 'learning_rate/action_model': 3.296292388915726e-05, 'epoch': 1.64}) +[2026-06-18 02:06:31,148][starVLA.training.train_starvla][INFO] - Step 4482, Loss: {'train/loss': 0.24459360539913177, 'train/grad_norm_pre_clip': 17.91008186340332, 'timing/data': 0.0013581179082393646, 'timing/model': 1.1601251848042011, 'learning_rate/qwen_vl_interface': 3.294321305638243e-06, 'learning_rate/action_model': 3.294321305638243e-05, 'epoch': 1.64}) +[2026-06-18 02:06:35,526][starVLA.training.train_starvla][INFO] - Step 4483, Loss: {'train/loss': 0.3155846893787384, 'train/grad_norm_pre_clip': 16.749683380126953, 'timing/data': 0.0009033242240548134, 'timing/model': 1.0284336153417826, 'learning_rate/qwen_vl_interface': 3.292350627775061e-06, 'learning_rate/action_model': 3.292350627775061e-05, 'epoch': 1.64}) +[2026-06-18 02:06:39,985][starVLA.training.train_starvla][INFO] - Step 4484, Loss: {'train/loss': 0.3561024069786072, 'train/grad_norm_pre_clip': 22.062564849853516, 'timing/data': 0.0020241616293787956, 'timing/model': 1.0691563477739692, 'learning_rate/qwen_vl_interface': 3.2903803557346997e-06, 'learning_rate/action_model': 3.2903803557347e-05, 'epoch': 1.64}) +[2026-06-18 02:06:44,375][starVLA.training.train_starvla][INFO] - Step 4485, Loss: {'train/loss': 0.4405328929424286, 'train/grad_norm_pre_clip': 22.779878616333008, 'timing/data': 0.0016145706176757812, 'timing/model': 1.404738220386207, 'learning_rate/qwen_vl_interface': 3.2884104899255986e-06, 'learning_rate/action_model': 3.288410489925599e-05, 'epoch': 1.64}) +[2026-06-18 02:06:48,568][starVLA.training.train_starvla][INFO] - Step 4486, Loss: {'train/loss': 0.6805877685546875, 'train/grad_norm_pre_clip': 23.772693634033203, 'timing/data': 0.0012050149962306023, 'timing/model': 1.0478740017861128, 'learning_rate/qwen_vl_interface': 3.2864410307561124e-06, 'learning_rate/action_model': 3.286441030756112e-05, 'epoch': 1.64}) +[2026-06-18 02:06:53,354][starVLA.training.train_starvla][INFO] - Step 4487, Loss: {'train/loss': 0.8662317395210266, 'train/grad_norm_pre_clip': 27.39835548400879, 'timing/data': 0.0011001471430063248, 'timing/model': 1.1069362815469503, 'learning_rate/qwen_vl_interface': 3.284471978634515e-06, 'learning_rate/action_model': 3.284471978634515e-05, 'epoch': 1.64}) +[2026-06-18 02:07:01,393][starVLA.training.train_starvla][INFO] - Step 4488, Loss: {'train/loss': 0.48934808373451233, 'train/grad_norm_pre_clip': 20.787761688232422, 'timing/data': 0.04672263562679291, 'timing/model': 4.587989518418908, 'learning_rate/qwen_vl_interface': 3.2825033339689926e-06, 'learning_rate/action_model': 3.282503333968993e-05, 'epoch': 1.64}) +[2026-06-18 02:07:13,844][starVLA.training.train_starvla][INFO] - Step 4489, Loss: {'train/loss': 0.6835148334503174, 'train/grad_norm_pre_clip': 28.21022605895996, 'timing/data': 0.0008771270513534546, 'timing/model': 3.6682429499924183, 'learning_rate/qwen_vl_interface': 3.2805350971676447e-06, 'learning_rate/action_model': 3.280535097167645e-05, 'epoch': 1.64}) +[2026-06-18 02:07:22,860][starVLA.training.train_starvla][INFO] - Step 4490, Loss: {'train/loss': 0.24133779108524323, 'train/grad_norm_pre_clip': 12.040908813476562, 'timing/data': 0.0008946266025304794, 'timing/model': 4.261905026622117, 'learning_rate/qwen_vl_interface': 3.2785672686384893e-06, 'learning_rate/action_model': 3.278567268638489e-05, 'epoch': 1.64}) +[2026-06-18 02:07:32,012][starVLA.training.train_starvla][INFO] - Step 4491, Loss: {'train/loss': 0.21453197300434113, 'train/grad_norm_pre_clip': 15.928262710571289, 'timing/data': 0.0008949330076575279, 'timing/model': 1.064834258519113, 'learning_rate/qwen_vl_interface': 3.2765998487894622e-06, 'learning_rate/action_model': 3.276599848789462e-05, 'epoch': 1.64}) +[2026-06-18 02:07:36,632][starVLA.training.train_starvla][INFO] - Step 4492, Loss: {'train/loss': 0.4828825294971466, 'train/grad_norm_pre_clip': 25.397733688354492, 'timing/data': 0.0007782438769936562, 'timing/model': 1.0859173852950335, 'learning_rate/qwen_vl_interface': 3.2746328380284098e-06, 'learning_rate/action_model': 3.27463283802841e-05, 'epoch': 1.64}) +[2026-06-18 02:07:43,891][starVLA.training.train_starvla][INFO] - Step 4493, Loss: {'train/loss': 0.34158775210380554, 'train/grad_norm_pre_clip': 20.158037185668945, 'timing/data': 0.0014040851965546608, 'timing/model': 1.2959271250292659, 'learning_rate/qwen_vl_interface': 3.2726662367630953e-06, 'learning_rate/action_model': 3.272666236763095e-05, 'epoch': 1.65}) +[2026-06-18 02:07:48,962][starVLA.training.train_starvla][INFO] - Step 4494, Loss: {'train/loss': 0.26453274488449097, 'train/grad_norm_pre_clip': 17.23466682434082, 'timing/data': 0.0007413029670715332, 'timing/model': 1.0440344586968422, 'learning_rate/qwen_vl_interface': 3.270700045401197e-06, 'learning_rate/action_model': 3.2707000454011974e-05, 'epoch': 1.65}) +[2026-06-18 02:07:53,611][starVLA.training.train_starvla][INFO] - Step 4495, Loss: {'train/loss': 0.5705544948577881, 'train/grad_norm_pre_clip': 30.80017852783203, 'timing/data': 0.0009748488664627075, 'timing/model': 1.4158839089795947, 'learning_rate/qwen_vl_interface': 3.26873426435031e-06, 'learning_rate/action_model': 3.2687342643503096e-05, 'epoch': 1.65}) +[2026-06-18 02:08:03,065][starVLA.training.train_starvla][INFO] - Step 4496, Loss: {'train/loss': 0.44222432374954224, 'train/grad_norm_pre_clip': 21.54315757751465, 'timing/data': 0.0010403217747807503, 'timing/model': 1.0493124015629292, 'learning_rate/qwen_vl_interface': 3.266768894017942e-06, 'learning_rate/action_model': 3.266768894017942e-05, 'epoch': 1.65}) +[2026-06-18 02:08:15,679][starVLA.training.train_starvla][INFO] - Step 4497, Loss: {'train/loss': 0.3600548207759857, 'train/grad_norm_pre_clip': 16.11466407775879, 'timing/data': 0.0011497559025883675, 'timing/model': 4.295201907865703, 'learning_rate/qwen_vl_interface': 3.264803934811514e-06, 'learning_rate/action_model': 3.264803934811514e-05, 'epoch': 1.65}) +[2026-06-18 02:08:21,023][starVLA.training.train_starvla][INFO] - Step 4498, Loss: {'train/loss': 0.4164666533470154, 'train/grad_norm_pre_clip': 16.969701766967773, 'timing/data': 0.00088447704911232, 'timing/model': 1.3950179275125265, 'learning_rate/qwen_vl_interface': 3.2628393871383706e-06, 'learning_rate/action_model': 3.26283938713837e-05, 'epoch': 1.65}) +[2026-06-18 02:08:26,512][starVLA.training.train_starvla][INFO] - Step 4499, Loss: {'train/loss': 0.6959403157234192, 'train/grad_norm_pre_clip': 27.726913452148438, 'timing/data': 0.0009625265374779701, 'timing/model': 1.3666128646582365, 'learning_rate/qwen_vl_interface': 3.2608752514057585e-06, 'learning_rate/action_model': 3.260875251405758e-05, 'epoch': 1.65}) +[2026-06-18 02:13:28,202][starVLA.training.train_starvla][INFO] - Step 4500, Loss: {'train/loss': 0.5479978919029236, 'train/grad_norm_pre_clip': 27.117509841918945, 'eval/loss': 0.5447846055030823, 'eval/latency_2/loss': 0.5447846055030823, 'eval/demon_fire_precision': 0.9400330851943756, 'eval/demon_fire_recall': 0.9510460251046026, 'eval/demon_fire_f1': 0.9455074875207987, 'eval/demon_fire_rate_model': 0.672787979966611, 'eval/demon_fire_rate_teacher': 0.6649972175848636, 'eval/demon_left_precision': 0.8873538788522848, 'eval/demon_left_recall': 0.8789473684210526, 'eval/demon_left_f1': 0.8831306187202538, 'eval/demon_left_rate_model': 0.261825264329438, 'eval/demon_left_rate_teacher': 0.2643294379521425, 'eval/demon_right_precision': 0.8858757062146893, 'eval/demon_right_recall': 0.8858757062146893, 'eval/demon_right_f1': 0.8858757062146893, 'eval/demon_right_rate_model': 0.24624373956594323, 'eval/demon_right_rate_teacher': 0.24624373956594323, 'eval/demon_move_f1': 0.8845031624674715, 'eval/demon_cc_f1': 0.9263405309984096, 'eval/latency_2/demon_fire_precision': 0.9400330851943756, 'eval/latency_2/demon_fire_recall': 0.9510460251046026, 'eval/latency_2/demon_fire_f1': 0.9455074875207987, 'eval/latency_2/demon_fire_rate_model': 0.672787979966611, 'eval/latency_2/demon_fire_rate_teacher': 0.6649972175848636, 'eval/latency_2/demon_left_precision': 0.8873538788522848, 'eval/latency_2/demon_left_recall': 0.8789473684210526, 'eval/latency_2/demon_left_f1': 0.8831306187202538, 'eval/latency_2/demon_left_rate_model': 0.261825264329438, 'eval/latency_2/demon_left_rate_teacher': 0.2643294379521425, 'eval/latency_2/demon_right_precision': 0.8858757062146893, 'eval/latency_2/demon_right_recall': 0.8858757062146893, 'eval/latency_2/demon_right_f1': 0.8858757062146893, 'eval/latency_2/demon_right_rate_model': 0.24624373956594323, 'eval/latency_2/demon_right_rate_teacher': 0.24624373956594323, 'eval/latency_2/demon_move_f1': 0.8845031624674715, 'eval/latency_2/demon_cc_f1': 0.9263405309984096, 'timing/data': 0.0011771004647016525, 'timing/model': 2.854447959922254, 'learning_rate/qwen_vl_interface': 3.2589115280208473e-06, 'learning_rate/action_model': 3.258911528020847e-05, 'epoch': 1.65}) +[2026-06-18 02:13:28,204][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_4500_state +[2026-06-18 02:13:28,247][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 02:13:36,773][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_4500_state/model.safetensors +[2026-06-18 02:13:52,323][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_4500_state/optimizer.bin +[2026-06-18 02:13:52,326][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_4500_state/sampler.bin +[2026-06-18 02:13:52,326][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_4500_state/sampler_1.bin +[2026-06-18 02:13:52,330][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_4500_state/random_states_0.pkl +[2026-06-18 02:13:52,335][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4500 +[2026-06-18 02:13:52,336][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 02:13:52,355][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 02:14:00,487][starVLA.training.train_starvla][INFO] - Step 4501, Loss: {'train/loss': 0.7678700089454651, 'train/grad_norm_pre_clip': 26.83101463317871, 'timing/data': 0.001089419238269329, 'timing/model': 4.86379552166909, 'learning_rate/qwen_vl_interface': 3.2569482173907164e-06, 'learning_rate/action_model': 3.2569482173907165e-05, 'epoch': 1.65}) +[2026-06-18 02:14:06,213][starVLA.training.train_starvla][INFO] - Step 4502, Loss: {'train/loss': 0.2107425332069397, 'train/grad_norm_pre_clip': 15.312967300415039, 'timing/data': 0.02365822345018387, 'timing/model': 1.5097957076504827, 'learning_rate/qwen_vl_interface': 3.2549853199223668e-06, 'learning_rate/action_model': 3.254985319922367e-05, 'epoch': 1.65}) +[2026-06-18 02:14:15,634][starVLA.training.train_starvla][INFO] - Step 4503, Loss: {'train/loss': 0.569911539554596, 'train/grad_norm_pre_clip': 25.815128326416016, 'timing/data': 0.0010630488395690918, 'timing/model': 1.08953985478729, 'learning_rate/qwen_vl_interface': 3.253022836022708e-06, 'learning_rate/action_model': 3.253022836022708e-05, 'epoch': 1.65}) +[2026-06-18 02:14:20,618][starVLA.training.train_starvla][INFO] - Step 4504, Loss: {'train/loss': 0.771081805229187, 'train/grad_norm_pre_clip': 28.47711944580078, 'timing/data': 0.001309497281908989, 'timing/model': 1.055839796550572, 'learning_rate/qwen_vl_interface': 3.2510607660985627e-06, 'learning_rate/action_model': 3.251060766098563e-05, 'epoch': 1.65}) +[2026-06-18 02:14:26,900][starVLA.training.train_starvla][INFO] - Step 4505, Loss: {'train/loss': 0.3308396339416504, 'train/grad_norm_pre_clip': 17.68250846862793, 'timing/data': 0.0008634962141513824, 'timing/model': 2.692889221943915, 'learning_rate/qwen_vl_interface': 3.249099110556672e-06, 'learning_rate/action_model': 3.249099110556672e-05, 'epoch': 1.65}) +[2026-06-18 02:14:31,853][starVLA.training.train_starvla][INFO] - Step 4506, Loss: {'train/loss': 0.31715860962867737, 'train/grad_norm_pre_clip': 16.156230926513672, 'timing/data': 0.0010047182440757751, 'timing/model': 1.7434386191889644, 'learning_rate/qwen_vl_interface': 3.247137869803689e-06, 'learning_rate/action_model': 3.2471378698036886e-05, 'epoch': 1.65}) +[2026-06-18 02:14:36,528][starVLA.training.train_starvla][INFO] - Step 4507, Loss: {'train/loss': 0.4431651830673218, 'train/grad_norm_pre_clip': 25.17854881286621, 'timing/data': 0.0009804414585232735, 'timing/model': 1.3022124478593469, 'learning_rate/qwen_vl_interface': 3.2451770442461813e-06, 'learning_rate/action_model': 3.245177044246181e-05, 'epoch': 1.65}) +[2026-06-18 02:14:41,165][starVLA.training.train_starvla][INFO] - Step 4508, Loss: {'train/loss': 0.36289504170417786, 'train/grad_norm_pre_clip': 16.79251480102539, 'timing/data': 0.0008772015571594238, 'timing/model': 1.102517700754106, 'learning_rate/qwen_vl_interface': 3.2432166342906286e-06, 'learning_rate/action_model': 3.2432166342906284e-05, 'epoch': 1.65}) +[2026-06-18 02:14:45,745][starVLA.training.train_starvla][INFO] - Step 4509, Loss: {'train/loss': 0.37252435088157654, 'train/grad_norm_pre_clip': 22.084135055541992, 'timing/data': 0.00273866206407547, 'timing/model': 1.126133375801146, 'learning_rate/qwen_vl_interface': 3.241256640343427e-06, 'learning_rate/action_model': 3.241256640343427e-05, 'epoch': 1.65}) +[2026-06-18 02:14:50,204][starVLA.training.train_starvla][INFO] - Step 4510, Loss: {'train/loss': 0.6751479506492615, 'train/grad_norm_pre_clip': 26.726009368896484, 'timing/data': 0.0011184317991137505, 'timing/model': 1.063830615952611, 'learning_rate/qwen_vl_interface': 3.239297062810886e-06, 'learning_rate/action_model': 3.239297062810886e-05, 'epoch': 1.65}) +[2026-06-18 02:14:55,191][starVLA.training.train_starvla][INFO] - Step 4511, Loss: {'train/loss': 0.4545644521713257, 'train/grad_norm_pre_clip': 19.248294830322266, 'timing/data': 0.0011566029861569405, 'timing/model': 1.705070236697793, 'learning_rate/qwen_vl_interface': 3.2373379020992252e-06, 'learning_rate/action_model': 3.237337902099225e-05, 'epoch': 1.65}) +[2026-06-18 02:14:59,285][starVLA.training.train_starvla][INFO] - Step 4512, Loss: {'train/loss': 0.4787102937698364, 'train/grad_norm_pre_clip': 20.319814682006836, 'timing/data': 0.0008880849927663803, 'timing/model': 1.030567206442356, 'learning_rate/qwen_vl_interface': 3.2353791586145855e-06, 'learning_rate/action_model': 3.2353791586145855e-05, 'epoch': 1.65}) +[2026-06-18 02:15:09,438][starVLA.training.train_starvla][INFO] - Step 4513, Loss: {'train/loss': 0.2683798372745514, 'train/grad_norm_pre_clip': 17.612964630126953, 'timing/data': 0.0032270345836877823, 'timing/model': 1.1317382352426648, 'learning_rate/qwen_vl_interface': 3.2334208327630158e-06, 'learning_rate/action_model': 3.233420832763016e-05, 'epoch': 1.65}) +[2026-06-18 02:15:14,112][starVLA.training.train_starvla][INFO] - Step 4514, Loss: {'train/loss': 0.21086454391479492, 'train/grad_norm_pre_clip': 11.704791069030762, 'timing/data': 0.0009420644491910934, 'timing/model': 1.352108959108591, 'learning_rate/qwen_vl_interface': 3.231462924950475e-06, 'learning_rate/action_model': 3.2314629249504746e-05, 'epoch': 1.65}) +[2026-06-18 02:15:18,721][starVLA.training.train_starvla][INFO] - Step 4515, Loss: {'train/loss': 0.6987253427505493, 'train/grad_norm_pre_clip': 23.96541976928711, 'timing/data': 0.0017596641555428505, 'timing/model': 1.0842764414846897, 'learning_rate/qwen_vl_interface': 3.2295054355828396e-06, 'learning_rate/action_model': 3.22950543558284e-05, 'epoch': 1.65}) +[2026-06-18 02:15:23,546][starVLA.training.train_starvla][INFO] - Step 4516, Loss: {'train/loss': 0.29594242572784424, 'train/grad_norm_pre_clip': 15.614218711853027, 'timing/data': 0.0009089745581150055, 'timing/model': 1.4233884047716856, 'learning_rate/qwen_vl_interface': 3.227548365065905e-06, 'learning_rate/action_model': 3.227548365065905e-05, 'epoch': 1.65}) +[2026-06-18 02:15:27,795][starVLA.training.train_starvla][INFO] - Step 4517, Loss: {'train/loss': 0.26606374979019165, 'train/grad_norm_pre_clip': 15.205869674682617, 'timing/data': 0.0010446254163980484, 'timing/model': 1.0356189915910363, 'learning_rate/qwen_vl_interface': 3.2255917138053695e-06, 'learning_rate/action_model': 3.2255917138053696e-05, 'epoch': 1.65}) +[2026-06-18 02:15:32,929][starVLA.training.train_starvla][INFO] - Step 4518, Loss: {'train/loss': 0.2681705951690674, 'train/grad_norm_pre_clip': 14.926003456115723, 'timing/data': 0.0010061189532279968, 'timing/model': 1.0664809178560972, 'learning_rate/qwen_vl_interface': 3.2236354822068515e-06, 'learning_rate/action_model': 3.223635482206852e-05, 'epoch': 1.65}) +[2026-06-18 02:15:42,420][starVLA.training.train_starvla][INFO] - Step 4519, Loss: {'train/loss': 0.23616153001785278, 'train/grad_norm_pre_clip': 15.583907127380371, 'timing/data': 0.0012134946882724762, 'timing/model': 4.330799445509911, 'learning_rate/qwen_vl_interface': 3.2216796706758757e-06, 'learning_rate/action_model': 3.221679670675876e-05, 'epoch': 1.65}) +[2026-06-18 02:15:52,069][starVLA.training.train_starvla][INFO] - Step 4520, Loss: {'train/loss': 0.4365856647491455, 'train/grad_norm_pre_clip': 18.775014877319336, 'timing/data': 0.002365473657846451, 'timing/model': 1.3882748410105705, 'learning_rate/qwen_vl_interface': 3.219724279617888e-06, 'learning_rate/action_model': 3.219724279617888e-05, 'epoch': 1.66}) +[2026-06-18 02:15:56,973][starVLA.training.train_starvla][INFO] - Step 4521, Loss: {'train/loss': 0.27879807353019714, 'train/grad_norm_pre_clip': 10.903806686401367, 'timing/data': 0.00136562529951334, 'timing/model': 1.321076301857829, 'learning_rate/qwen_vl_interface': 3.2177693094382424e-06, 'learning_rate/action_model': 3.2177693094382424e-05, 'epoch': 1.66}) +[2026-06-18 02:16:02,099][starVLA.training.train_starvla][INFO] - Step 4522, Loss: {'train/loss': 0.3798631429672241, 'train/grad_norm_pre_clip': 19.644197463989258, 'timing/data': 0.0008405623957514763, 'timing/model': 1.06994747556746, 'learning_rate/qwen_vl_interface': 3.2158147605422052e-06, 'learning_rate/action_model': 3.215814760542205e-05, 'epoch': 1.66}) +[2026-06-18 02:16:12,869][starVLA.training.train_starvla][INFO] - Step 4523, Loss: {'train/loss': 0.3951168358325958, 'train/grad_norm_pre_clip': 19.673921585083008, 'timing/data': 0.0008728699758648872, 'timing/model': 1.098884086124599, 'learning_rate/qwen_vl_interface': 3.2138606333349576e-06, 'learning_rate/action_model': 3.213860633334957e-05, 'epoch': 1.66}) +[2026-06-18 02:16:17,697][starVLA.training.train_starvla][INFO] - Step 4524, Loss: {'train/loss': 0.08909989148378372, 'train/grad_norm_pre_clip': 7.088405609130859, 'timing/data': 0.0009537665173411369, 'timing/model': 1.250560935586691, 'learning_rate/qwen_vl_interface': 3.2119069282215913e-06, 'learning_rate/action_model': 3.211906928221591e-05, 'epoch': 1.66}) +[2026-06-18 02:16:22,988][starVLA.training.train_starvla][INFO] - Step 4525, Loss: {'train/loss': 0.10022889077663422, 'train/grad_norm_pre_clip': 8.42152214050293, 'timing/data': 0.001077587716281414, 'timing/model': 1.0296430606395006, 'learning_rate/qwen_vl_interface': 3.2099536456071118e-06, 'learning_rate/action_model': 3.209953645607112e-05, 'epoch': 1.66}) +[2026-06-18 02:16:28,407][starVLA.training.train_starvla][INFO] - Step 4526, Loss: {'train/loss': 0.3914717733860016, 'train/grad_norm_pre_clip': 20.785846710205078, 'timing/data': 0.0017212498933076859, 'timing/model': 1.6221212223172188, 'learning_rate/qwen_vl_interface': 3.2080007858964347e-06, 'learning_rate/action_model': 3.2080007858964345e-05, 'epoch': 1.66}) +[2026-06-18 02:16:33,684][starVLA.training.train_starvla][INFO] - Step 4527, Loss: {'train/loss': 0.2785167992115021, 'train/grad_norm_pre_clip': 15.036163330078125, 'timing/data': 0.000998075120151043, 'timing/model': 1.5745950266718864, 'learning_rate/qwen_vl_interface': 3.206048349494396e-06, 'learning_rate/action_model': 3.206048349494396e-05, 'epoch': 1.66}) +[2026-06-18 02:16:38,962][starVLA.training.train_starvla][INFO] - Step 4528, Loss: {'train/loss': 0.34407398104667664, 'train/grad_norm_pre_clip': 18.586633682250977, 'timing/data': 0.0012478278949856758, 'timing/model': 1.296726357191801, 'learning_rate/qwen_vl_interface': 3.2040963368057323e-06, 'learning_rate/action_model': 3.204096336805732e-05, 'epoch': 1.66}) +[2026-06-18 02:16:45,802][starVLA.training.train_starvla][INFO] - Step 4529, Loss: {'train/loss': 0.23888389766216278, 'train/grad_norm_pre_clip': 20.358675003051758, 'timing/data': 0.0030097216367721558, 'timing/model': 1.108826557174325, 'learning_rate/qwen_vl_interface': 3.2021447482350994e-06, 'learning_rate/action_model': 3.202144748235099e-05, 'epoch': 1.66}) +[2026-06-18 02:16:50,444][starVLA.training.train_starvla][INFO] - Step 4530, Loss: {'train/loss': 0.4128720164299011, 'train/grad_norm_pre_clip': 24.377826690673828, 'timing/data': 0.0010308977216482162, 'timing/model': 1.2706680251285434, 'learning_rate/qwen_vl_interface': 3.200193584187061e-06, 'learning_rate/action_model': 3.200193584187061e-05, 'epoch': 1.66}) +[2026-06-18 02:16:58,411][starVLA.training.train_starvla][INFO] - Step 4531, Loss: {'train/loss': 0.29181113839149475, 'train/grad_norm_pre_clip': 18.5733642578125, 'timing/data': 0.0009305151179432869, 'timing/model': 2.0330140674486756, 'learning_rate/qwen_vl_interface': 3.198242845066101e-06, 'learning_rate/action_model': 3.198242845066101e-05, 'epoch': 1.66}) +[2026-06-18 02:17:03,291][starVLA.training.train_starvla][INFO] - Step 4532, Loss: {'train/loss': 0.6923178434371948, 'train/grad_norm_pre_clip': 32.230682373046875, 'timing/data': 0.001041734591126442, 'timing/model': 1.0427839253097773, 'learning_rate/qwen_vl_interface': 3.1962925312766087e-06, 'learning_rate/action_model': 3.196292531276608e-05, 'epoch': 1.66}) +[2026-06-18 02:17:07,918][starVLA.training.train_starvla][INFO] - Step 4533, Loss: {'train/loss': 0.24968232214450836, 'train/grad_norm_pre_clip': 13.810637474060059, 'timing/data': 0.000949525274336338, 'timing/model': 1.0401984369382262, 'learning_rate/qwen_vl_interface': 3.19434264322288e-06, 'learning_rate/action_model': 3.19434264322288e-05, 'epoch': 1.66}) +[2026-06-18 02:17:12,629][starVLA.training.train_starvla][INFO] - Step 4534, Loss: {'train/loss': 0.32655078172683716, 'train/grad_norm_pre_clip': 33.72165298461914, 'timing/data': 0.0009819678962230682, 'timing/model': 1.107346510514617, 'learning_rate/qwen_vl_interface': 3.1923931813091364e-06, 'learning_rate/action_model': 3.192393181309136e-05, 'epoch': 1.66}) +[2026-06-18 02:17:17,099][starVLA.training.train_starvla][INFO] - Step 4535, Loss: {'train/loss': 0.30819910764694214, 'train/grad_norm_pre_clip': 17.039152145385742, 'timing/data': 0.0008842190727591515, 'timing/model': 1.0535093490034342, 'learning_rate/qwen_vl_interface': 3.1904441459395e-06, 'learning_rate/action_model': 3.1904441459395e-05, 'epoch': 1.66}) +[2026-06-18 02:17:21,256][starVLA.training.train_starvla][INFO] - Step 4536, Loss: {'train/loss': 0.22971920669078827, 'train/grad_norm_pre_clip': 17.82453155517578, 'timing/data': 0.0012655220925807953, 'timing/model': 1.0286167869344354, 'learning_rate/qwen_vl_interface': 3.1884955375180088e-06, 'learning_rate/action_model': 3.1884955375180085e-05, 'epoch': 1.66}) +[2026-06-18 02:17:26,313][starVLA.training.train_starvla][INFO] - Step 4537, Loss: {'train/loss': 0.2135203778743744, 'train/grad_norm_pre_clip': 13.549029350280762, 'timing/data': 0.0007807081565260887, 'timing/model': 1.0830090111121535, 'learning_rate/qwen_vl_interface': 3.18654735644861e-06, 'learning_rate/action_model': 3.1865473564486095e-05, 'epoch': 1.66}) +[2026-06-18 02:17:30,947][starVLA.training.train_starvla][INFO] - Step 4538, Loss: {'train/loss': 0.20981353521347046, 'train/grad_norm_pre_clip': 13.378458023071289, 'timing/data': 0.0013927146792411804, 'timing/model': 1.0434903856366873, 'learning_rate/qwen_vl_interface': 3.184599603135165e-06, 'learning_rate/action_model': 3.1845996031351646e-05, 'epoch': 1.66}) +[2026-06-18 02:17:35,884][starVLA.training.train_starvla][INFO] - Step 4539, Loss: {'train/loss': 0.39770635962486267, 'train/grad_norm_pre_clip': 26.167251586914062, 'timing/data': 0.0010935105383396149, 'timing/model': 1.0463514141738415, 'learning_rate/qwen_vl_interface': 3.1826522779814453e-06, 'learning_rate/action_model': 3.182652277981445e-05, 'epoch': 1.66}) +[2026-06-18 02:17:40,356][starVLA.training.train_starvla][INFO] - Step 4540, Loss: {'train/loss': 0.4067252576351166, 'train/grad_norm_pre_clip': 25.986549377441406, 'timing/data': 0.0009735589846968651, 'timing/model': 1.0672061713412404, 'learning_rate/qwen_vl_interface': 3.1807053813911317e-06, 'learning_rate/action_model': 3.180705381391132e-05, 'epoch': 1.66}) +[2026-06-18 02:17:44,798][starVLA.training.train_starvla][INFO] - Step 4541, Loss: {'train/loss': 0.28719937801361084, 'train/grad_norm_pre_clip': 21.626924514770508, 'timing/data': 0.0013646595180034637, 'timing/model': 1.412735995836556, 'learning_rate/qwen_vl_interface': 3.17875891376782e-06, 'learning_rate/action_model': 3.1787589137678196e-05, 'epoch': 1.66}) +[2026-06-18 02:17:54,098][starVLA.training.train_starvla][INFO] - Step 4542, Loss: {'train/loss': 0.30431750416755676, 'train/grad_norm_pre_clip': 30.16422462463379, 'timing/data': 0.0011309636756777763, 'timing/model': 1.1924584629014134, 'learning_rate/qwen_vl_interface': 3.176812875515014e-06, 'learning_rate/action_model': 3.176812875515014e-05, 'epoch': 1.66}) +[2026-06-18 02:17:59,150][starVLA.training.train_starvla][INFO] - Step 4543, Loss: {'train/loss': 0.24267032742500305, 'train/grad_norm_pre_clip': 19.767135620117188, 'timing/data': 0.0009149918332695961, 'timing/model': 1.5861757677048445, 'learning_rate/qwen_vl_interface': 3.1748672670361304e-06, 'learning_rate/action_model': 3.17486726703613e-05, 'epoch': 1.66}) +[2026-06-18 02:18:04,014][starVLA.training.train_starvla][INFO] - Step 4544, Loss: {'train/loss': 0.1341620236635208, 'train/grad_norm_pre_clip': 11.804282188415527, 'timing/data': 0.0027864547446370125, 'timing/model': 1.0962380273267627, 'learning_rate/qwen_vl_interface': 3.172922088734493e-06, 'learning_rate/action_model': 3.172922088734493e-05, 'epoch': 1.66}) +[2026-06-18 02:18:08,690][starVLA.training.train_starvla][INFO] - Step 4545, Loss: {'train/loss': 0.42638927698135376, 'train/grad_norm_pre_clip': 25.87479591369629, 'timing/data': 0.0010629510506987572, 'timing/model': 1.0455837743356824, 'learning_rate/qwen_vl_interface': 3.1709773410133434e-06, 'learning_rate/action_model': 3.170977341013343e-05, 'epoch': 1.66}) +[2026-06-18 02:18:13,890][starVLA.training.train_starvla][INFO] - Step 4546, Loss: {'train/loss': 0.23645883798599243, 'train/grad_norm_pre_clip': 21.41864585876465, 'timing/data': 0.0008703414350748062, 'timing/model': 1.5318047245964408, 'learning_rate/qwen_vl_interface': 3.16903302427583e-06, 'learning_rate/action_model': 3.16903302427583e-05, 'epoch': 1.66}) +[2026-06-18 02:18:26,952][starVLA.training.train_starvla][INFO] - Step 4547, Loss: {'train/loss': 0.47803592681884766, 'train/grad_norm_pre_clip': 25.10303497314453, 'timing/data': 0.0009200219064950943, 'timing/model': 5.651874780654907, 'learning_rate/qwen_vl_interface': 3.1670891389250073e-06, 'learning_rate/action_model': 3.167089138925007e-05, 'epoch': 1.67}) +[2026-06-18 02:18:33,036][starVLA.training.train_starvla][INFO] - Step 4548, Loss: {'train/loss': 0.310212105512619, 'train/grad_norm_pre_clip': 24.491134643554688, 'timing/data': 0.0010935598984360695, 'timing/model': 1.0891654388979077, 'learning_rate/qwen_vl_interface': 3.1651456853638464e-06, 'learning_rate/action_model': 3.165145685363846e-05, 'epoch': 1.67}) +[2026-06-18 02:18:37,953][starVLA.training.train_starvla][INFO] - Step 4549, Loss: {'train/loss': 0.1681990772485733, 'train/grad_norm_pre_clip': 14.887478828430176, 'timing/data': 0.0012023597955703735, 'timing/model': 1.0385046480223536, 'learning_rate/qwen_vl_interface': 3.1632026639952303e-06, 'learning_rate/action_model': 3.16320266399523e-05, 'epoch': 1.67}) +[2026-06-18 02:18:42,738][starVLA.training.train_starvla][INFO] - Step 4550, Loss: {'train/loss': 0.23986271023750305, 'train/grad_norm_pre_clip': 17.997989654541016, 'timing/data': 0.0013239560648798943, 'timing/model': 1.0636132508516312, 'learning_rate/qwen_vl_interface': 3.161260075221947e-06, 'learning_rate/action_model': 3.161260075221947e-05, 'epoch': 1.67}) +[2026-06-18 02:18:49,775][starVLA.training.train_starvla][INFO] - Step 4551, Loss: {'train/loss': 0.7042801976203918, 'train/grad_norm_pre_clip': 31.660932540893555, 'timing/data': 0.0009236037731170654, 'timing/model': 3.5037149153649807, 'learning_rate/qwen_vl_interface': 3.1593179194467e-06, 'learning_rate/action_model': 3.1593179194467e-05, 'epoch': 1.67}) +[2026-06-18 02:18:58,684][starVLA.training.train_starvla][INFO] - Step 4552, Loss: {'train/loss': 0.28800809383392334, 'train/grad_norm_pre_clip': 24.687297821044922, 'timing/data': 0.0013894839212298393, 'timing/model': 1.112389880232513, 'learning_rate/qwen_vl_interface': 3.157376197072094e-06, 'learning_rate/action_model': 3.157376197072094e-05, 'epoch': 1.67}) +[2026-06-18 02:19:03,190][starVLA.training.train_starvla][INFO] - Step 4553, Loss: {'train/loss': 0.4480934143066406, 'train/grad_norm_pre_clip': 17.843353271484375, 'timing/data': 0.0011685527861118317, 'timing/model': 1.0320738358423114, 'learning_rate/qwen_vl_interface': 3.1554349085006557e-06, 'learning_rate/action_model': 3.155434908500656e-05, 'epoch': 1.67}) +[2026-06-18 02:19:07,714][starVLA.training.train_starvla][INFO] - Step 4554, Loss: {'train/loss': 0.33260682225227356, 'train/grad_norm_pre_clip': 25.765398025512695, 'timing/data': 0.0009310264140367508, 'timing/model': 1.1970526278018951, 'learning_rate/qwen_vl_interface': 3.1534940541348152e-06, 'learning_rate/action_model': 3.153494054134815e-05, 'epoch': 1.67}) +[2026-06-18 02:19:11,926][starVLA.training.train_starvla][INFO] - Step 4555, Loss: {'train/loss': 0.2781078517436981, 'train/grad_norm_pre_clip': 24.34600067138672, 'timing/data': 0.0010695448145270348, 'timing/model': 1.0347187304869294, 'learning_rate/qwen_vl_interface': 3.1515536343769133e-06, 'learning_rate/action_model': 3.151553634376913e-05, 'epoch': 1.67}) +[2026-06-18 02:19:16,405][starVLA.training.train_starvla][INFO] - Step 4556, Loss: {'train/loss': 0.5447338819503784, 'train/grad_norm_pre_clip': 20.390024185180664, 'timing/data': 0.0011023478582501411, 'timing/model': 1.4674208713695407, 'learning_rate/qwen_vl_interface': 3.1496136496292e-06, 'learning_rate/action_model': 3.1496136496292e-05, 'epoch': 1.67}) +[2026-06-18 02:19:20,341][starVLA.training.train_starvla][INFO] - Step 4557, Loss: {'train/loss': 0.22228209674358368, 'train/grad_norm_pre_clip': 14.0329008102417, 'timing/data': 0.000785481184720993, 'timing/model': 1.0187898157164454, 'learning_rate/qwen_vl_interface': 3.1476741002938383e-06, 'learning_rate/action_model': 3.147674100293838e-05, 'epoch': 1.67}) +[2026-06-18 02:19:24,529][starVLA.training.train_starvla][INFO] - Step 4558, Loss: {'train/loss': 0.35993292927742004, 'train/grad_norm_pre_clip': 24.32228660583496, 'timing/data': 0.001102750189602375, 'timing/model': 1.000322169624269, 'learning_rate/qwen_vl_interface': 3.145734986772897e-06, 'learning_rate/action_model': 3.1457349867728966e-05, 'epoch': 1.67}) +[2026-06-18 02:19:28,688][starVLA.training.train_starvla][INFO] - Step 4559, Loss: {'train/loss': 0.24527552723884583, 'train/grad_norm_pre_clip': 19.657773971557617, 'timing/data': 0.0021873880177736282, 'timing/model': 1.0200409507378936, 'learning_rate/qwen_vl_interface': 3.1437963094683566e-06, 'learning_rate/action_model': 3.1437963094683564e-05, 'epoch': 1.67}) +[2026-06-18 02:19:33,166][starVLA.training.train_starvla][INFO] - Step 4560, Loss: {'train/loss': 0.28989046812057495, 'train/grad_norm_pre_clip': 19.467844009399414, 'timing/data': 0.0009683705866336823, 'timing/model': 1.0243085250258446, 'learning_rate/qwen_vl_interface': 3.1418580687821103e-06, 'learning_rate/action_model': 3.14185806878211e-05, 'epoch': 1.67}) +[2026-06-18 02:19:37,588][starVLA.training.train_starvla][INFO] - Step 4561, Loss: {'train/loss': 0.5172853469848633, 'train/grad_norm_pre_clip': 23.88193130493164, 'timing/data': 0.0010290499776601791, 'timing/model': 1.2318741772323847, 'learning_rate/qwen_vl_interface': 3.1399202651159532e-06, 'learning_rate/action_model': 3.1399202651159534e-05, 'epoch': 1.67}) +[2026-06-18 02:19:41,640][starVLA.training.train_starvla][INFO] - Step 4562, Loss: {'train/loss': 0.19235506653785706, 'train/grad_norm_pre_clip': 15.370220184326172, 'timing/data': 0.0009755315259099007, 'timing/model': 1.055155442096293, 'learning_rate/qwen_vl_interface': 3.1379828988715955e-06, 'learning_rate/action_model': 3.1379828988715956e-05, 'epoch': 1.67}) +[2026-06-18 02:19:45,748][starVLA.training.train_starvla][INFO] - Step 4563, Loss: {'train/loss': 0.3698943555355072, 'train/grad_norm_pre_clip': 19.539011001586914, 'timing/data': 0.0010292334482073784, 'timing/model': 1.0055917715653777, 'learning_rate/qwen_vl_interface': 3.136045970450654e-06, 'learning_rate/action_model': 3.136045970450654e-05, 'epoch': 1.67}) +[2026-06-18 02:19:49,944][starVLA.training.train_starvla][INFO] - Step 4564, Loss: {'train/loss': 0.3391585946083069, 'train/grad_norm_pre_clip': 22.825551986694336, 'timing/data': 0.000917515717446804, 'timing/model': 1.0069007351994514, 'learning_rate/qwen_vl_interface': 3.1341094802546595e-06, 'learning_rate/action_model': 3.1341094802546595e-05, 'epoch': 1.67}) +[2026-06-18 02:19:54,042][starVLA.training.train_starvla][INFO] - Step 4565, Loss: {'train/loss': 0.7458346486091614, 'train/grad_norm_pre_clip': 37.36360549926758, 'timing/data': 0.0011642985045909882, 'timing/model': 1.011171436868608, 'learning_rate/qwen_vl_interface': 3.132173428685048e-06, 'learning_rate/action_model': 3.132173428685048e-05, 'epoch': 1.67}) +[2026-06-18 02:19:58,252][starVLA.training.train_starvla][INFO] - Step 4566, Loss: {'train/loss': 0.344341903924942, 'train/grad_norm_pre_clip': 25.27572250366211, 'timing/data': 0.0010369354858994484, 'timing/model': 1.2883160002529621, 'learning_rate/qwen_vl_interface': 3.13023781614316e-06, 'learning_rate/action_model': 3.13023781614316e-05, 'epoch': 1.67}) +[2026-06-18 02:20:03,707][starVLA.training.train_starvla][INFO] - Step 4567, Loss: {'train/loss': 0.5163769721984863, 'train/grad_norm_pre_clip': 27.477670669555664, 'timing/data': 0.000797271728515625, 'timing/model': 2.5286298748105764, 'learning_rate/qwen_vl_interface': 3.1283026430302553e-06, 'learning_rate/action_model': 3.128302643030255e-05, 'epoch': 1.67}) +[2026-06-18 02:20:09,289][starVLA.training.train_starvla][INFO] - Step 4568, Loss: {'train/loss': 0.48193109035491943, 'train/grad_norm_pre_clip': 22.687484741210938, 'timing/data': 0.001153741031885147, 'timing/model': 1.096733876504004, 'learning_rate/qwen_vl_interface': 3.1263679097474958e-06, 'learning_rate/action_model': 3.126367909747496e-05, 'epoch': 1.67}) +[2026-06-18 02:20:13,595][starVLA.training.train_starvla][INFO] - Step 4569, Loss: {'train/loss': 0.36750608682632446, 'train/grad_norm_pre_clip': 21.449417114257812, 'timing/data': 0.0008840784430503845, 'timing/model': 1.0358295012265444, 'learning_rate/qwen_vl_interface': 3.124433616695955e-06, 'learning_rate/action_model': 3.1244336166959545e-05, 'epoch': 1.67}) +[2026-06-18 02:20:18,371][starVLA.training.train_starvla][INFO] - Step 4570, Loss: {'train/loss': 0.28704634308815, 'train/grad_norm_pre_clip': 18.481842041015625, 'timing/data': 0.0010586418211460114, 'timing/model': 1.07580850366503, 'learning_rate/qwen_vl_interface': 3.1224997642766115e-06, 'learning_rate/action_model': 3.122499764276611e-05, 'epoch': 1.67}) +[2026-06-18 02:20:26,598][starVLA.training.train_starvla][INFO] - Step 4571, Loss: {'train/loss': 0.6644386053085327, 'train/grad_norm_pre_clip': 24.282686233520508, 'timing/data': 0.0011098859831690788, 'timing/model': 1.4380767941474915, 'learning_rate/qwen_vl_interface': 3.1205663528903563e-06, 'learning_rate/action_model': 3.1205663528903565e-05, 'epoch': 1.67}) +[2026-06-18 02:20:35,119][starVLA.training.train_starvla][INFO] - Step 4572, Loss: {'train/loss': 0.34167736768722534, 'train/grad_norm_pre_clip': 17.149972915649414, 'timing/data': 0.000917789526283741, 'timing/model': 1.0678883660584688, 'learning_rate/qwen_vl_interface': 3.118633382937987e-06, 'learning_rate/action_model': 3.118633382937987e-05, 'epoch': 1.67}) +[2026-06-18 02:20:39,665][starVLA.training.train_starvla][INFO] - Step 4573, Loss: {'train/loss': 0.4717785120010376, 'train/grad_norm_pre_clip': 20.152267456054688, 'timing/data': 0.001169503666460514, 'timing/model': 1.125855934806168, 'learning_rate/qwen_vl_interface': 3.116700854820209e-06, 'learning_rate/action_model': 3.116700854820209e-05, 'epoch': 1.67}) +[2026-06-18 02:20:44,301][starVLA.training.train_starvla][INFO] - Step 4574, Loss: {'train/loss': 0.23536548018455505, 'train/grad_norm_pre_clip': 24.371618270874023, 'timing/data': 0.0011069774627685547, 'timing/model': 1.0467088846489787, 'learning_rate/qwen_vl_interface': 3.1147687689376443e-06, 'learning_rate/action_model': 3.1147687689376444e-05, 'epoch': 1.67}) +[2026-06-18 02:20:50,316][starVLA.training.train_starvla][INFO] - Step 4575, Loss: {'train/loss': 0.5445261001586914, 'train/grad_norm_pre_clip': 19.7840518951416, 'timing/data': 0.001154283992946148, 'timing/model': 1.0713818855583668, 'learning_rate/qwen_vl_interface': 3.1128371256908087e-06, 'learning_rate/action_model': 3.112837125690809e-05, 'epoch': 1.68}) +[2026-06-18 02:20:59,445][starVLA.training.train_starvla][INFO] - Step 4576, Loss: {'train/loss': 0.2877897620201111, 'train/grad_norm_pre_clip': 17.363622665405273, 'timing/data': 0.00113571435213089, 'timing/model': 1.746795273385942, 'learning_rate/qwen_vl_interface': 3.1109059254801368e-06, 'learning_rate/action_model': 3.1109059254801366e-05, 'epoch': 1.68}) +[2026-06-18 02:21:03,782][starVLA.training.train_starvla][INFO] - Step 4577, Loss: {'train/loss': 0.5318838357925415, 'train/grad_norm_pre_clip': 23.370859146118164, 'timing/data': 0.0010159015655517578, 'timing/model': 1.1219276189804077, 'learning_rate/qwen_vl_interface': 3.108975168705966e-06, 'learning_rate/action_model': 3.1089751687059656e-05, 'epoch': 1.68}) +[2026-06-18 02:21:08,693][starVLA.training.train_starvla][INFO] - Step 4578, Loss: {'train/loss': 0.21614713966846466, 'train/grad_norm_pre_clip': 14.701130867004395, 'timing/data': 0.0011706296354532242, 'timing/model': 1.0419168137013912, 'learning_rate/qwen_vl_interface': 3.107044855768549e-06, 'learning_rate/action_model': 3.107044855768549e-05, 'epoch': 1.68}) +[2026-06-18 02:21:19,303][starVLA.training.train_starvla][INFO] - Step 4579, Loss: {'train/loss': 0.3382214605808258, 'train/grad_norm_pre_clip': 18.75950813293457, 'timing/data': 0.0012542558833956718, 'timing/model': 5.419566977769136, 'learning_rate/qwen_vl_interface': 3.1051149870680413e-06, 'learning_rate/action_model': 3.105114987068041e-05, 'epoch': 1.68}) +[2026-06-18 02:21:24,671][starVLA.training.train_starvla][INFO] - Step 4580, Loss: {'train/loss': 0.24442756175994873, 'train/grad_norm_pre_clip': 17.454044342041016, 'timing/data': 0.0009018797427415848, 'timing/model': 1.0503912437707186, 'learning_rate/qwen_vl_interface': 3.103185563004503e-06, 'learning_rate/action_model': 3.103185563004503e-05, 'epoch': 1.68}) +[2026-06-18 02:21:30,880][starVLA.training.train_starvla][INFO] - Step 4581, Loss: {'train/loss': 0.35035237669944763, 'train/grad_norm_pre_clip': 14.16669750213623, 'timing/data': 0.0012968303635716438, 'timing/model': 2.994511364027858, 'learning_rate/qwen_vl_interface': 3.101256583977905e-06, 'learning_rate/action_model': 3.101256583977905e-05, 'epoch': 1.68}) +[2026-06-18 02:21:39,076][starVLA.training.train_starvla][INFO] - Step 4582, Loss: {'train/loss': 0.16498877108097076, 'train/grad_norm_pre_clip': 14.277495384216309, 'timing/data': 0.001143711619079113, 'timing/model': 1.24008097499609, 'learning_rate/qwen_vl_interface': 3.099328050388131e-06, 'learning_rate/action_model': 3.0993280503881305e-05, 'epoch': 1.68}) +[2026-06-18 02:21:47,229][starVLA.training.train_starvla][INFO] - Step 4583, Loss: {'train/loss': 0.5486080646514893, 'train/grad_norm_pre_clip': 22.947500228881836, 'timing/data': 0.0010143481194972992, 'timing/model': 4.5501323929056525, 'learning_rate/qwen_vl_interface': 3.0973999626349667e-06, 'learning_rate/action_model': 3.097399962634966e-05, 'epoch': 1.68}) +[2026-06-18 02:21:52,010][starVLA.training.train_starvla][INFO] - Step 4584, Loss: {'train/loss': 0.323496013879776, 'train/grad_norm_pre_clip': 18.099510192871094, 'timing/data': 0.0011440422385931015, 'timing/model': 1.1125420462340117, 'learning_rate/qwen_vl_interface': 3.095472321118107e-06, 'learning_rate/action_model': 3.095472321118107e-05, 'epoch': 1.68}) +[2026-06-18 02:21:56,931][starVLA.training.train_starvla][INFO] - Step 4585, Loss: {'train/loss': 0.22865647077560425, 'train/grad_norm_pre_clip': 16.006540298461914, 'timing/data': 0.001397610642015934, 'timing/model': 1.4387177526950836, 'learning_rate/qwen_vl_interface': 3.093545126237153e-06, 'learning_rate/action_model': 3.0935451262371527e-05, 'epoch': 1.68}) +[2026-06-18 02:22:03,179][starVLA.training.train_starvla][INFO] - Step 4586, Loss: {'train/loss': 0.6570822596549988, 'train/grad_norm_pre_clip': 23.353670120239258, 'timing/data': 0.000905095599591732, 'timing/model': 2.939755287952721, 'learning_rate/qwen_vl_interface': 3.091618378391615e-06, 'learning_rate/action_model': 3.0916183783916146e-05, 'epoch': 1.68}) +[2026-06-18 02:22:07,759][starVLA.training.train_starvla][INFO] - Step 4587, Loss: {'train/loss': 0.18388307094573975, 'train/grad_norm_pre_clip': 12.158500671386719, 'timing/data': 0.0010809963569045067, 'timing/model': 1.0329367024824023, 'learning_rate/qwen_vl_interface': 3.0896920779809097e-06, 'learning_rate/action_model': 3.08969207798091e-05, 'epoch': 1.68}) +[2026-06-18 02:22:12,552][starVLA.training.train_starvla][INFO] - Step 4588, Loss: {'train/loss': 0.2770927846431732, 'train/grad_norm_pre_clip': 16.622947692871094, 'timing/data': 0.0011930279433727264, 'timing/model': 1.1566753219813108, 'learning_rate/qwen_vl_interface': 3.0877662254043605e-06, 'learning_rate/action_model': 3.0877662254043606e-05, 'epoch': 1.68}) +[2026-06-18 02:22:17,190][starVLA.training.train_starvla][INFO] - Step 4589, Loss: {'train/loss': 0.392299622297287, 'train/grad_norm_pre_clip': 22.099803924560547, 'timing/data': 0.0007969308644533157, 'timing/model': 1.042336548678577, 'learning_rate/qwen_vl_interface': 3.0858408210612e-06, 'learning_rate/action_model': 3.0858408210612e-05, 'epoch': 1.68}) +[2026-06-18 02:22:21,873][starVLA.training.train_starvla][INFO] - Step 4590, Loss: {'train/loss': 0.40118205547332764, 'train/grad_norm_pre_clip': 20.816020965576172, 'timing/data': 0.0009127706289291382, 'timing/model': 1.119049952365458, 'learning_rate/qwen_vl_interface': 3.0839158653505663e-06, 'learning_rate/action_model': 3.0839158653505664e-05, 'epoch': 1.68}) +[2026-06-18 02:22:26,429][starVLA.training.train_starvla][INFO] - Step 4591, Loss: {'train/loss': 0.44793862104415894, 'train/grad_norm_pre_clip': 23.55300521850586, 'timing/data': 0.0011642230674624443, 'timing/model': 1.4873596439138055, 'learning_rate/qwen_vl_interface': 3.081991358671504e-06, 'learning_rate/action_model': 3.081991358671504e-05, 'epoch': 1.68}) +[2026-06-18 02:22:30,716][starVLA.training.train_starvla][INFO] - Step 4592, Loss: {'train/loss': 0.5792865753173828, 'train/grad_norm_pre_clip': 24.582595825195312, 'timing/data': 0.001180524006485939, 'timing/model': 1.149780299514532, 'learning_rate/qwen_vl_interface': 3.080067301422965e-06, 'learning_rate/action_model': 3.080067301422965e-05, 'epoch': 1.68}) +[2026-06-18 02:22:35,233][starVLA.training.train_starvla][INFO] - Step 4593, Loss: {'train/loss': 0.3127189576625824, 'train/grad_norm_pre_clip': 55.2657470703125, 'timing/data': 0.0011520776897668839, 'timing/model': 1.0575975785031915, 'learning_rate/qwen_vl_interface': 3.07814369400381e-06, 'learning_rate/action_model': 3.07814369400381e-05, 'epoch': 1.68}) +[2026-06-18 02:22:41,015][starVLA.training.train_starvla][INFO] - Step 4594, Loss: {'train/loss': 0.313888818025589, 'train/grad_norm_pre_clip': 22.848731994628906, 'timing/data': 0.001370372250676155, 'timing/model': 1.0646259570494294, 'learning_rate/qwen_vl_interface': 3.076220536812807e-06, 'learning_rate/action_model': 3.076220536812807e-05, 'epoch': 1.68}) +[2026-06-18 02:22:45,550][starVLA.training.train_starvla][INFO] - Step 4595, Loss: {'train/loss': 0.2085033357143402, 'train/grad_norm_pre_clip': 16.979698181152344, 'timing/data': 0.0008804481476545334, 'timing/model': 1.379170130006969, 'learning_rate/qwen_vl_interface': 3.0742978302486214e-06, 'learning_rate/action_model': 3.074297830248621e-05, 'epoch': 1.68}) +[2026-06-18 02:22:50,322][starVLA.training.train_starvla][INFO] - Step 4596, Loss: {'train/loss': 0.341769814491272, 'train/grad_norm_pre_clip': 20.91560173034668, 'timing/data': 0.0009551998227834702, 'timing/model': 1.395232398994267, 'learning_rate/qwen_vl_interface': 3.0723755747098387e-06, 'learning_rate/action_model': 3.072375574709839e-05, 'epoch': 1.68}) +[2026-06-18 02:22:58,207][starVLA.training.train_starvla][INFO] - Step 4597, Loss: {'train/loss': 0.3610455095767975, 'train/grad_norm_pre_clip': 19.03093910217285, 'timing/data': 0.0012067360803484917, 'timing/model': 4.676497559063137, 'learning_rate/qwen_vl_interface': 3.070453770594943e-06, 'learning_rate/action_model': 3.070453770594943e-05, 'epoch': 1.68}) +[2026-06-18 02:23:04,052][starVLA.training.train_starvla][INFO] - Step 4598, Loss: {'train/loss': 0.3475610613822937, 'train/grad_norm_pre_clip': 16.130340576171875, 'timing/data': 0.0007829181849956512, 'timing/model': 2.5239152843132615, 'learning_rate/qwen_vl_interface': 3.0685324183023264e-06, 'learning_rate/action_model': 3.068532418302326e-05, 'epoch': 1.68}) +[2026-06-18 02:23:14,134][starVLA.training.train_starvla][INFO] - Step 4599, Loss: {'train/loss': 0.2935403287410736, 'train/grad_norm_pre_clip': 20.58795166015625, 'timing/data': 0.0008591171354055405, 'timing/model': 1.2187146740034223, 'learning_rate/qwen_vl_interface': 3.0666115182302863e-06, 'learning_rate/action_model': 3.0666115182302865e-05, 'epoch': 1.68}) +[2026-06-18 02:28:42,604][starVLA.training.train_starvla][INFO] - Step 4600, Loss: {'train/loss': 0.16147539019584656, 'train/grad_norm_pre_clip': 14.445534706115723, 'eval/loss': 0.5454086065292358, 'eval/latency_2/loss': 0.5454086065292358, 'eval/demon_fire_precision': 0.9436795994993742, 'eval/demon_fire_recall': 0.9464435146443515, 'eval/demon_fire_f1': 0.9450595362439942, 'eval/demon_fire_rate_model': 0.6669449081803005, 'eval/demon_fire_rate_teacher': 0.6649972175848636, 'eval/demon_left_precision': 0.9038901601830663, 'eval/demon_left_recall': 0.8315789473684211, 'eval/demon_left_f1': 0.8662280701754386, 'eval/demon_left_rate_model': 0.24318308291597107, 'eval/demon_left_rate_teacher': 0.2643294379521425, 'eval/demon_right_precision': 0.8648363252375924, 'eval/demon_right_recall': 0.9254237288135593, 'eval/demon_right_f1': 0.8941048034934498, 'eval/demon_right_rate_model': 0.263494713411241, 'eval/demon_right_rate_teacher': 0.24624373956594323, 'eval/demon_move_f1': 0.8801664368344442, 'eval/demon_cc_f1': 0.9246086137226719, 'eval/latency_2/demon_fire_precision': 0.9436795994993742, 'eval/latency_2/demon_fire_recall': 0.9464435146443515, 'eval/latency_2/demon_fire_f1': 0.9450595362439942, 'eval/latency_2/demon_fire_rate_model': 0.6669449081803005, 'eval/latency_2/demon_fire_rate_teacher': 0.6649972175848636, 'eval/latency_2/demon_left_precision': 0.9038901601830663, 'eval/latency_2/demon_left_recall': 0.8315789473684211, 'eval/latency_2/demon_left_f1': 0.8662280701754386, 'eval/latency_2/demon_left_rate_model': 0.24318308291597107, 'eval/latency_2/demon_left_rate_teacher': 0.2643294379521425, 'eval/latency_2/demon_right_precision': 0.8648363252375924, 'eval/latency_2/demon_right_recall': 0.9254237288135593, 'eval/latency_2/demon_right_f1': 0.8941048034934498, 'eval/latency_2/demon_right_rate_model': 0.263494713411241, 'eval/latency_2/demon_right_rate_teacher': 0.24624373956594323, 'eval/latency_2/demon_move_f1': 0.8801664368344442, 'eval/latency_2/demon_cc_f1': 0.9246086137226719, 'timing/data': 0.0015421276912093163, 'timing/model': 3.5924401814118028, 'learning_rate/qwen_vl_interface': 3.064691070777027e-06, 'learning_rate/action_model': 3.064691070777027e-05, 'epoch': 1.68}) +[2026-06-18 02:28:42,607][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_4600_state +[2026-06-18 02:28:42,616][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 02:28:53,547][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_4600_state/model.safetensors +[2026-06-18 02:29:12,253][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_4600_state/optimizer.bin +[2026-06-18 02:29:12,254][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_4600_state/sampler.bin +[2026-06-18 02:29:12,255][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_4600_state/sampler_1.bin +[2026-06-18 02:29:12,257][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_4600_state/random_states_0.pkl +[2026-06-18 02:29:12,264][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4600 +[2026-06-18 02:29:12,264][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 02:29:12,281][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 02:29:17,302][starVLA.training.train_starvla][INFO] - Step 4601, Loss: {'train/loss': 0.922019898891449, 'train/grad_norm_pre_clip': 29.62177085876465, 'timing/data': 0.001623295247554779, 'timing/model': 1.0506422128528357, 'learning_rate/qwen_vl_interface': 3.062771076340662e-06, 'learning_rate/action_model': 3.0627710763406617e-05, 'epoch': 1.68}) +[2026-06-18 02:29:22,446][starVLA.training.train_starvla][INFO] - Step 4602, Loss: {'train/loss': 0.6665727496147156, 'train/grad_norm_pre_clip': 28.300968170166016, 'timing/data': 0.0008681276813149452, 'timing/model': 1.3985469052568078, 'learning_rate/qwen_vl_interface': 3.0608515353192046e-06, 'learning_rate/action_model': 3.060851535319205e-05, 'epoch': 1.69}) +[2026-06-18 02:29:27,145][starVLA.training.train_starvla][INFO] - Step 4603, Loss: {'train/loss': 0.4967469871044159, 'train/grad_norm_pre_clip': 21.451505661010742, 'timing/data': 0.0012758010998368263, 'timing/model': 1.3297565039247274, 'learning_rate/qwen_vl_interface': 3.0589324481105814e-06, 'learning_rate/action_model': 3.058932448110581e-05, 'epoch': 1.69}) +[2026-06-18 02:29:31,760][starVLA.training.train_starvla][INFO] - Step 4604, Loss: {'train/loss': 0.6844285726547241, 'train/grad_norm_pre_clip': 25.888221740722656, 'timing/data': 0.0008087856695055962, 'timing/model': 1.2687374595552683, 'learning_rate/qwen_vl_interface': 3.0570138151126185e-06, 'learning_rate/action_model': 3.057013815112618e-05, 'epoch': 1.69}) +[2026-06-18 02:29:36,509][starVLA.training.train_starvla][INFO] - Step 4605, Loss: {'train/loss': 0.2309909462928772, 'train/grad_norm_pre_clip': 16.900182723999023, 'timing/data': 0.0011923443526029587, 'timing/model': 1.042858175933361, 'learning_rate/qwen_vl_interface': 3.0550956367230506e-06, 'learning_rate/action_model': 3.05509563672305e-05, 'epoch': 1.69}) +[2026-06-18 02:29:41,049][starVLA.training.train_starvla][INFO] - Step 4606, Loss: {'train/loss': 0.4527323544025421, 'train/grad_norm_pre_clip': 21.229623794555664, 'timing/data': 0.0011594202369451523, 'timing/model': 1.220860437490046, 'learning_rate/qwen_vl_interface': 3.0531779133395184e-06, 'learning_rate/action_model': 3.053177913339518e-05, 'epoch': 1.69}) +[2026-06-18 02:29:45,894][starVLA.training.train_starvla][INFO] - Step 4607, Loss: {'train/loss': 0.28482842445373535, 'train/grad_norm_pre_clip': 23.92991065979004, 'timing/data': 0.0007900716736912727, 'timing/model': 1.5938095264136791, 'learning_rate/qwen_vl_interface': 3.0512606453595708e-06, 'learning_rate/action_model': 3.0512606453595704e-05, 'epoch': 1.69}) +[2026-06-18 02:29:53,609][starVLA.training.train_starvla][INFO] - Step 4608, Loss: {'train/loss': 0.40261706709861755, 'train/grad_norm_pre_clip': 21.011354446411133, 'timing/data': 0.0008386233821511269, 'timing/model': 2.291577815078199, 'learning_rate/qwen_vl_interface': 3.049343833180658e-06, 'learning_rate/action_model': 3.049343833180658e-05, 'epoch': 1.69}) +[2026-06-18 02:30:01,149][starVLA.training.train_starvla][INFO] - Step 4609, Loss: {'train/loss': 0.32816410064697266, 'train/grad_norm_pre_clip': 20.085264205932617, 'timing/data': 0.0011360729113221169, 'timing/model': 1.172436786815524, 'learning_rate/qwen_vl_interface': 3.0474274772001355e-06, 'learning_rate/action_model': 3.0474274772001355e-05, 'epoch': 1.69}) +[2026-06-18 02:30:06,221][starVLA.training.train_starvla][INFO] - Step 4610, Loss: {'train/loss': 0.15612229704856873, 'train/grad_norm_pre_clip': 12.197818756103516, 'timing/data': 0.001446225680410862, 'timing/model': 1.4350679693743587, 'learning_rate/qwen_vl_interface': 3.045511577815265e-06, 'learning_rate/action_model': 3.045511577815265e-05, 'epoch': 1.69}) +[2026-06-18 02:30:20,843][starVLA.training.train_starvla][INFO] - Step 4611, Loss: {'train/loss': 0.31377726793289185, 'train/grad_norm_pre_clip': 20.595380783081055, 'timing/data': 0.04196813516318798, 'timing/model': 3.4728044122457504, 'learning_rate/qwen_vl_interface': 3.043596135423218e-06, 'learning_rate/action_model': 3.043596135423218e-05, 'epoch': 1.69}) +[2026-06-18 02:30:25,031][starVLA.training.train_starvla][INFO] - Step 4612, Loss: {'train/loss': 0.40422457456588745, 'train/grad_norm_pre_clip': 23.253034591674805, 'timing/data': 0.035009147599339485, 'timing/model': 1.0832431428134441, 'learning_rate/qwen_vl_interface': 3.041681150421067e-06, 'learning_rate/action_model': 3.041681150421067e-05, 'epoch': 1.69}) +[2026-06-18 02:30:34,936][starVLA.training.train_starvla][INFO] - Step 4613, Loss: {'train/loss': 0.4231036901473999, 'train/grad_norm_pre_clip': 20.689456939697266, 'timing/data': 0.047434852458536625, 'timing/model': 2.799121806398034, 'learning_rate/qwen_vl_interface': 3.039766623205791e-06, 'learning_rate/action_model': 3.039766623205791e-05, 'epoch': 1.69}) +[2026-06-18 02:30:43,521][starVLA.training.train_starvla][INFO] - Step 4614, Loss: {'train/loss': 0.32925015687942505, 'train/grad_norm_pre_clip': 18.283740997314453, 'timing/data': 0.042986901476979256, 'timing/model': 3.3174735214561224, 'learning_rate/qwen_vl_interface': 3.037852554174269e-06, 'learning_rate/action_model': 3.0378525541742692e-05, 'epoch': 1.69}) +[2026-06-18 02:30:49,872][starVLA.training.train_starvla][INFO] - Step 4615, Loss: {'train/loss': 0.45734190940856934, 'train/grad_norm_pre_clip': 24.15581703186035, 'timing/data': 0.04344537016004324, 'timing/model': 1.046293149702251, 'learning_rate/qwen_vl_interface': 3.035938943723295e-06, 'learning_rate/action_model': 3.035938943723295e-05, 'epoch': 1.69}) +[2026-06-18 02:30:54,609][starVLA.training.train_starvla][INFO] - Step 4616, Loss: {'train/loss': 0.22521641850471497, 'train/grad_norm_pre_clip': 15.130328178405762, 'timing/data': 0.046588512137532234, 'timing/model': 1.385528919287026, 'learning_rate/qwen_vl_interface': 3.03402579224956e-06, 'learning_rate/action_model': 3.03402579224956e-05, 'epoch': 1.69}) +[2026-06-18 02:31:00,031][starVLA.training.train_starvla][INFO] - Step 4617, Loss: {'train/loss': 0.2421720325946808, 'train/grad_norm_pre_clip': 17.55889320373535, 'timing/data': 0.04103774484246969, 'timing/model': 1.947359667159617, 'learning_rate/qwen_vl_interface': 3.032113100149663e-06, 'learning_rate/action_model': 3.0321131001496628e-05, 'epoch': 1.69}) +[2026-06-18 02:31:04,206][starVLA.training.train_starvla][INFO] - Step 4618, Loss: {'train/loss': 0.2551197409629822, 'train/grad_norm_pre_clip': 14.967885971069336, 'timing/data': 0.04104012344032526, 'timing/model': 1.0469621336087584, 'learning_rate/qwen_vl_interface': 3.0302008678201067e-06, 'learning_rate/action_model': 3.0302008678201067e-05, 'epoch': 1.69}) +[2026-06-18 02:31:15,816][starVLA.training.train_starvla][INFO] - Step 4619, Loss: {'train/loss': 0.36276328563690186, 'train/grad_norm_pre_clip': 24.417802810668945, 'timing/data': 0.03931126277893782, 'timing/model': 4.314370967447758, 'learning_rate/qwen_vl_interface': 3.028289095657299e-06, 'learning_rate/action_model': 3.028289095657299e-05, 'epoch': 1.69}) +[2026-06-18 02:31:20,389][starVLA.training.train_starvla][INFO] - Step 4620, Loss: {'train/loss': 0.3130902051925659, 'train/grad_norm_pre_clip': 24.70884895324707, 'timing/data': 0.05201713275164366, 'timing/model': 1.0641759186983109, 'learning_rate/qwen_vl_interface': 3.026377784057552e-06, 'learning_rate/action_model': 3.026377784057552e-05, 'epoch': 1.69}) +[2026-06-18 02:31:25,027][starVLA.training.train_starvla][INFO] - Step 4621, Loss: {'train/loss': 0.2704901695251465, 'train/grad_norm_pre_clip': 15.844709396362305, 'timing/data': 0.0384399127215147, 'timing/model': 1.1318206395953894, 'learning_rate/qwen_vl_interface': 3.0244669334170822e-06, 'learning_rate/action_model': 3.0244669334170822e-05, 'epoch': 1.69}) +[2026-06-18 02:31:29,575][starVLA.training.train_starvla][INFO] - Step 4622, Loss: {'train/loss': 0.46497493982315063, 'train/grad_norm_pre_clip': 26.795209884643555, 'timing/data': 0.037778690457344055, 'timing/model': 1.4249666202813387, 'learning_rate/qwen_vl_interface': 3.0225565441320156e-06, 'learning_rate/action_model': 3.0225565441320152e-05, 'epoch': 1.69}) +[2026-06-18 02:31:33,701][starVLA.training.train_starvla][INFO] - Step 4623, Loss: {'train/loss': 0.1875932812690735, 'train/grad_norm_pre_clip': 15.244088172912598, 'timing/data': 0.0417406028136611, 'timing/model': 1.1002243654802442, 'learning_rate/qwen_vl_interface': 3.0206466165983716e-06, 'learning_rate/action_model': 3.0206466165983716e-05, 'epoch': 1.69}) +[2026-06-18 02:31:38,585][starVLA.training.train_starvla][INFO] - Step 4624, Loss: {'train/loss': 0.30072468519210815, 'train/grad_norm_pre_clip': 16.087501525878906, 'timing/data': 0.040279148146510124, 'timing/model': 1.4224244309589267, 'learning_rate/qwen_vl_interface': 3.0187371512120833e-06, 'learning_rate/action_model': 3.0187371512120832e-05, 'epoch': 1.69}) +[2026-06-18 02:31:48,625][starVLA.training.train_starvla][INFO] - Step 4625, Loss: {'train/loss': 0.47141844034194946, 'train/grad_norm_pre_clip': 19.437654495239258, 'timing/data': 0.058201905339956284, 'timing/model': 3.7794422153383493, 'learning_rate/qwen_vl_interface': 3.0168281483689825e-06, 'learning_rate/action_model': 3.0168281483689825e-05, 'epoch': 1.69}) +[2026-06-18 02:31:56,931][starVLA.training.train_starvla][INFO] - Step 4626, Loss: {'train/loss': 0.3077945411205292, 'train/grad_norm_pre_clip': 25.298133850097656, 'timing/data': 0.047497835010290146, 'timing/model': 4.418395820073783, 'learning_rate/qwen_vl_interface': 3.014919608464812e-06, 'learning_rate/action_model': 3.0149196084648117e-05, 'epoch': 1.69}) +[2026-06-18 02:32:01,733][starVLA.training.train_starvla][INFO] - Step 4627, Loss: {'train/loss': 0.5247452259063721, 'train/grad_norm_pre_clip': 19.230876922607422, 'timing/data': 0.04026162251830101, 'timing/model': 1.160608310252428, 'learning_rate/qwen_vl_interface': 3.0130115318952124e-06, 'learning_rate/action_model': 3.0130115318952122e-05, 'epoch': 1.69}) +[2026-06-18 02:32:06,712][starVLA.training.train_starvla][INFO] - Step 4628, Loss: {'train/loss': 0.2564309239387512, 'train/grad_norm_pre_clip': 16.596445083618164, 'timing/data': 0.09185204654932022, 'timing/model': 1.0804950697347522, 'learning_rate/qwen_vl_interface': 3.011103919055726e-06, 'learning_rate/action_model': 3.011103919055726e-05, 'epoch': 1.69}) +[2026-06-18 02:32:11,822][starVLA.training.train_starvla][INFO] - Step 4629, Loss: {'train/loss': 0.3969714641571045, 'train/grad_norm_pre_clip': 20.262969970703125, 'timing/data': 0.052918609231710434, 'timing/model': 1.4465653272345662, 'learning_rate/qwen_vl_interface': 3.009196770341809e-06, 'learning_rate/action_model': 3.009196770341809e-05, 'epoch': 1.7}) +[2026-06-18 02:32:18,636][starVLA.training.train_starvla][INFO] - Step 4630, Loss: {'train/loss': 0.4391728341579437, 'train/grad_norm_pre_clip': 21.68202781677246, 'timing/data': 0.05072708614170551, 'timing/model': 1.1792906792834401, 'learning_rate/qwen_vl_interface': 3.0072900861488125e-06, 'learning_rate/action_model': 3.0072900861488125e-05, 'epoch': 1.7}) +[2026-06-18 02:32:26,553][starVLA.training.train_starvla][INFO] - Step 4631, Loss: {'train/loss': 0.1089157834649086, 'train/grad_norm_pre_clip': 12.681747436523438, 'timing/data': 0.024472909979522228, 'timing/model': 1.0959852626547217, 'learning_rate/qwen_vl_interface': 3.0053838668719944e-06, 'learning_rate/action_model': 3.0053838668719943e-05, 'epoch': 1.7}) +[2026-06-18 02:32:31,615][starVLA.training.train_starvla][INFO] - Step 4632, Loss: {'train/loss': 0.45129382610321045, 'train/grad_norm_pre_clip': 29.140470504760742, 'timing/data': 0.02287087868899107, 'timing/model': 1.9273010268807411, 'learning_rate/qwen_vl_interface': 3.003478112906515e-06, 'learning_rate/action_model': 3.0034781129065148e-05, 'epoch': 1.7}) +[2026-06-18 02:32:43,102][starVLA.training.train_starvla][INFO] - Step 4633, Loss: {'train/loss': 0.3222247064113617, 'train/grad_norm_pre_clip': 17.396162033081055, 'timing/data': 0.025503691285848618, 'timing/model': 1.7182421684265137, 'learning_rate/qwen_vl_interface': 3.0015728246474405e-06, 'learning_rate/action_model': 3.0015728246474405e-05, 'epoch': 1.7}) +[2026-06-18 02:32:47,852][starVLA.training.train_starvla][INFO] - Step 4634, Loss: {'train/loss': 0.2686134874820709, 'train/grad_norm_pre_clip': 16.562185287475586, 'timing/data': 0.022113816812634468, 'timing/model': 1.057219480164349, 'learning_rate/qwen_vl_interface': 2.99966800248974e-06, 'learning_rate/action_model': 2.99966800248974e-05, 'epoch': 1.7}) +[2026-06-18 02:32:53,020][starVLA.training.train_starvla][INFO] - Step 4635, Loss: {'train/loss': 0.5059062838554382, 'train/grad_norm_pre_clip': 22.221778869628906, 'timing/data': 0.027906712144613266, 'timing/model': 1.186406266875565, 'learning_rate/qwen_vl_interface': 2.9977636468282817e-06, 'learning_rate/action_model': 2.9977636468282816e-05, 'epoch': 1.7}) +[2026-06-18 02:33:03,640][starVLA.training.train_starvla][INFO] - Step 4636, Loss: {'train/loss': 0.2553626000881195, 'train/grad_norm_pre_clip': 16.01451873779297, 'timing/data': 0.0011740224435925484, 'timing/model': 1.2402894329279661, 'learning_rate/qwen_vl_interface': 2.995859758057846e-06, 'learning_rate/action_model': 2.9958597580578456e-05, 'epoch': 1.7}) +[2026-06-18 02:33:08,412][starVLA.training.train_starvla][INFO] - Step 4637, Loss: {'train/loss': 0.18432942032814026, 'train/grad_norm_pre_clip': 18.754751205444336, 'timing/data': 0.0010362574830651283, 'timing/model': 1.4933450147509575, 'learning_rate/qwen_vl_interface': 2.993956336573106e-06, 'learning_rate/action_model': 2.9939563365731062e-05, 'epoch': 1.7}) +[2026-06-18 02:33:12,934][starVLA.training.train_starvla][INFO] - Step 4638, Loss: {'train/loss': 0.4408351480960846, 'train/grad_norm_pre_clip': 34.792266845703125, 'timing/data': 0.0012811888009309769, 'timing/model': 1.1580834006890655, 'learning_rate/qwen_vl_interface': 2.9920533827686453e-06, 'learning_rate/action_model': 2.9920533827686453e-05, 'epoch': 1.7}) +[2026-06-18 02:33:22,380][starVLA.training.train_starvla][INFO] - Step 4639, Loss: {'train/loss': 0.35313719511032104, 'train/grad_norm_pre_clip': 28.456518173217773, 'timing/data': 0.0009711161255836487, 'timing/model': 1.1869866270571947, 'learning_rate/qwen_vl_interface': 2.9901508970389477e-06, 'learning_rate/action_model': 2.9901508970389476e-05, 'epoch': 1.7}) +[2026-06-18 02:33:30,785][starVLA.training.train_starvla][INFO] - Step 4640, Loss: {'train/loss': 0.20200102031230927, 'train/grad_norm_pre_clip': 14.785146713256836, 'timing/data': 0.0012665195390582085, 'timing/model': 1.1204501176252961, 'learning_rate/qwen_vl_interface': 2.988248879778401e-06, 'learning_rate/action_model': 2.988248879778401e-05, 'epoch': 1.7}) +[2026-06-18 02:33:37,729][starVLA.training.train_starvla][INFO] - Step 4641, Loss: {'train/loss': 0.3455289304256439, 'train/grad_norm_pre_clip': 18.663002014160156, 'timing/data': 0.0022472012788057327, 'timing/model': 1.2765590865164995, 'learning_rate/qwen_vl_interface': 2.986347331381297e-06, 'learning_rate/action_model': 2.9863473313812967e-05, 'epoch': 1.7}) +[2026-06-18 02:33:42,294][starVLA.training.train_starvla][INFO] - Step 4642, Loss: {'train/loss': 0.22265180945396423, 'train/grad_norm_pre_clip': 17.71695899963379, 'timing/data': 0.001245204359292984, 'timing/model': 1.4357485109940171, 'learning_rate/qwen_vl_interface': 2.984446252241826e-06, 'learning_rate/action_model': 2.984446252241826e-05, 'epoch': 1.7}) +[2026-06-18 02:33:46,741][starVLA.training.train_starvla][INFO] - Step 4643, Loss: {'train/loss': 0.2726154923439026, 'train/grad_norm_pre_clip': 15.44645881652832, 'timing/data': 0.0012070909142494202, 'timing/model': 1.0618501482531428, 'learning_rate/qwen_vl_interface': 2.982545642754083e-06, 'learning_rate/action_model': 2.9825456427540826e-05, 'epoch': 1.7}) +[2026-06-18 02:33:51,860][starVLA.training.train_starvla][INFO] - Step 4644, Loss: {'train/loss': 0.30394890904426575, 'train/grad_norm_pre_clip': 16.910558700561523, 'timing/data': 0.0010513970628380775, 'timing/model': 1.290275325998664, 'learning_rate/qwen_vl_interface': 2.9806455033120697e-06, 'learning_rate/action_model': 2.9806455033120696e-05, 'epoch': 1.7}) +[2026-06-18 02:33:57,066][starVLA.training.train_starvla][INFO] - Step 4645, Loss: {'train/loss': 0.41768476366996765, 'train/grad_norm_pre_clip': 23.219446182250977, 'timing/data': 0.001804092898964882, 'timing/model': 1.026166875846684, 'learning_rate/qwen_vl_interface': 2.978745834309685e-06, 'learning_rate/action_model': 2.9787458343096846e-05, 'epoch': 1.7}) +[2026-06-18 02:34:02,053][starVLA.training.train_starvla][INFO] - Step 4646, Loss: {'train/loss': 0.48317813873291016, 'train/grad_norm_pre_clip': 28.10782814025879, 'timing/data': 0.001150030642747879, 'timing/model': 1.4577659191563725, 'learning_rate/qwen_vl_interface': 2.9768466361407335e-06, 'learning_rate/action_model': 2.9768466361407334e-05, 'epoch': 1.7}) +[2026-06-18 02:34:10,216][starVLA.training.train_starvla][INFO] - Step 4647, Loss: {'train/loss': 0.4067077934741974, 'train/grad_norm_pre_clip': 24.12218475341797, 'timing/data': 0.0015177447348833084, 'timing/model': 4.56097776722163, 'learning_rate/qwen_vl_interface': 2.97494790919892e-06, 'learning_rate/action_model': 2.9749479091989197e-05, 'epoch': 1.7}) +[2026-06-18 02:34:15,447][starVLA.training.train_starvla][INFO] - Step 4648, Loss: {'train/loss': 0.40619993209838867, 'train/grad_norm_pre_clip': 17.86365509033203, 'timing/data': 0.0009822472929954529, 'timing/model': 1.6190985068678856, 'learning_rate/qwen_vl_interface': 2.9730496538778524e-06, 'learning_rate/action_model': 2.9730496538778523e-05, 'epoch': 1.7}) +[2026-06-18 02:34:20,687][starVLA.training.train_starvla][INFO] - Step 4649, Loss: {'train/loss': 0.36896684765815735, 'train/grad_norm_pre_clip': 22.640918731689453, 'timing/data': 0.0010851481929421425, 'timing/model': 1.4141670018434525, 'learning_rate/qwen_vl_interface': 2.9711518705710422e-06, 'learning_rate/action_model': 2.9711518705710424e-05, 'epoch': 1.7}) +[2026-06-18 02:34:25,337][starVLA.training.train_starvla][INFO] - Step 4650, Loss: {'train/loss': 0.33816978335380554, 'train/grad_norm_pre_clip': 20.48892593383789, 'timing/data': 0.0010636737570166588, 'timing/model': 1.1202867990359664, 'learning_rate/qwen_vl_interface': 2.9692545596719014e-06, 'learning_rate/action_model': 2.969254559671901e-05, 'epoch': 1.7}) +[2026-06-18 02:34:38,964][starVLA.training.train_starvla][INFO] - Step 4651, Loss: {'train/loss': 0.2784058749675751, 'train/grad_norm_pre_clip': 18.1532039642334, 'timing/data': 0.0011884011328220367, 'timing/model': 4.487050863914192, 'learning_rate/qwen_vl_interface': 2.9673577215737454e-06, 'learning_rate/action_model': 2.9673577215737452e-05, 'epoch': 1.7}) +[2026-06-18 02:34:44,703][starVLA.training.train_starvla][INFO] - Step 4652, Loss: {'train/loss': 0.22328929603099823, 'train/grad_norm_pre_clip': 18.839258193969727, 'timing/data': 0.0008853813633322716, 'timing/model': 2.0228261509910226, 'learning_rate/qwen_vl_interface': 2.9654613566697886e-06, 'learning_rate/action_model': 2.9654613566697887e-05, 'epoch': 1.7}) +[2026-06-18 02:34:49,312][starVLA.training.train_starvla][INFO] - Step 4653, Loss: {'train/loss': 0.36968180537223816, 'train/grad_norm_pre_clip': 24.385761260986328, 'timing/data': 0.0010907230898737907, 'timing/model': 1.0603953627869487, 'learning_rate/qwen_vl_interface': 2.9635654653531516e-06, 'learning_rate/action_model': 2.9635654653531515e-05, 'epoch': 1.7}) +[2026-06-18 02:35:02,402][starVLA.training.train_starvla][INFO] - Step 4654, Loss: {'train/loss': 0.21075648069381714, 'train/grad_norm_pre_clip': 14.949373245239258, 'timing/data': 0.0010871458798646927, 'timing/model': 1.4185825726017356, 'learning_rate/qwen_vl_interface': 2.961670048016852e-06, 'learning_rate/action_model': 2.9616700480168518e-05, 'epoch': 1.7}) +[2026-06-18 02:35:07,295][starVLA.training.train_starvla][INFO] - Step 4655, Loss: {'train/loss': 0.168196901679039, 'train/grad_norm_pre_clip': 15.129072189331055, 'timing/data': 0.0012240130454301834, 'timing/model': 1.4311278499662876, 'learning_rate/qwen_vl_interface': 2.9597751050538177e-06, 'learning_rate/action_model': 2.9597751050538176e-05, 'epoch': 1.7}) +[2026-06-18 02:35:11,979][starVLA.training.train_starvla][INFO] - Step 4656, Loss: {'train/loss': 0.4704436957836151, 'train/grad_norm_pre_clip': 29.05021095275879, 'timing/data': 0.0008447570726275444, 'timing/model': 1.042353474535048, 'learning_rate/qwen_vl_interface': 2.957880636856866e-06, 'learning_rate/action_model': 2.9578806368568655e-05, 'epoch': 1.7}) +[2026-06-18 02:35:16,707][starVLA.training.train_starvla][INFO] - Step 4657, Loss: {'train/loss': 0.3680698871612549, 'train/grad_norm_pre_clip': 20.85407066345215, 'timing/data': 0.0008966689929366112, 'timing/model': 1.0859473124146461, 'learning_rate/qwen_vl_interface': 2.955986643818723e-06, 'learning_rate/action_model': 2.9559866438187228e-05, 'epoch': 1.71}) +[2026-06-18 02:35:23,434][starVLA.training.train_starvla][INFO] - Step 4658, Loss: {'train/loss': 0.28489404916763306, 'train/grad_norm_pre_clip': 38.72645568847656, 'timing/data': 0.0008864859119057655, 'timing/model': 1.3424782659858465, 'learning_rate/qwen_vl_interface': 2.954093126332019e-06, 'learning_rate/action_model': 2.9540931263320194e-05, 'epoch': 1.71}) +[2026-06-18 02:35:28,695][starVLA.training.train_starvla][INFO] - Step 4659, Loss: {'train/loss': 0.13997048139572144, 'train/grad_norm_pre_clip': 12.896515846252441, 'timing/data': 0.0010628988966345787, 'timing/model': 1.4930557990446687, 'learning_rate/qwen_vl_interface': 2.9522000847892807e-06, 'learning_rate/action_model': 2.9522000847892807e-05, 'epoch': 1.71}) +[2026-06-18 02:35:33,447][starVLA.training.train_starvla][INFO] - Step 4660, Loss: {'train/loss': 0.4421880841255188, 'train/grad_norm_pre_clip': 24.43157196044922, 'timing/data': 0.000809255987405777, 'timing/model': 1.1221143119037151, 'learning_rate/qwen_vl_interface': 2.9503075195829393e-06, 'learning_rate/action_model': 2.9503075195829393e-05, 'epoch': 1.71}) +[2026-06-18 02:35:38,956][starVLA.training.train_starvla][INFO] - Step 4661, Loss: {'train/loss': 0.4520460069179535, 'train/grad_norm_pre_clip': 25.05376625061035, 'timing/data': 0.0012742681428790092, 'timing/model': 1.6500260084867477, 'learning_rate/qwen_vl_interface': 2.9484154311053203e-06, 'learning_rate/action_model': 2.94841543110532e-05, 'epoch': 1.71}) +[2026-06-18 02:35:49,311][starVLA.training.train_starvla][INFO] - Step 4662, Loss: {'train/loss': 0.4925401210784912, 'train/grad_norm_pre_clip': 27.300472259521484, 'timing/data': 0.0013396507129073143, 'timing/model': 1.4925049962475896, 'learning_rate/qwen_vl_interface': 2.94652381974866e-06, 'learning_rate/action_model': 2.94652381974866e-05, 'epoch': 1.71}) +[2026-06-18 02:35:53,545][starVLA.training.train_starvla][INFO] - Step 4663, Loss: {'train/loss': 0.24161694943904877, 'train/grad_norm_pre_clip': 16.47602653503418, 'timing/data': 0.0009924601763486862, 'timing/model': 1.1354433204978704, 'learning_rate/qwen_vl_interface': 2.9446326859050916e-06, 'learning_rate/action_model': 2.9446326859050916e-05, 'epoch': 1.71}) +[2026-06-18 02:35:58,224][starVLA.training.train_starvla][INFO] - Step 4664, Loss: {'train/loss': 0.2117331475019455, 'train/grad_norm_pre_clip': 18.38648796081543, 'timing/data': 0.00094631128013134, 'timing/model': 1.1022315099835396, 'learning_rate/qwen_vl_interface': 2.9427420299666472e-06, 'learning_rate/action_model': 2.942742029966647e-05, 'epoch': 1.71}) +[2026-06-18 02:36:02,715][starVLA.training.train_starvla][INFO] - Step 4665, Loss: {'train/loss': 0.30129876732826233, 'train/grad_norm_pre_clip': 27.714717864990234, 'timing/data': 0.0009758351370692253, 'timing/model': 1.1109928963705897, 'learning_rate/qwen_vl_interface': 2.940851852325263e-06, 'learning_rate/action_model': 2.9408518523252632e-05, 'epoch': 1.71}) +[2026-06-18 02:36:07,820][starVLA.training.train_starvla][INFO] - Step 4666, Loss: {'train/loss': 0.19049863517284393, 'train/grad_norm_pre_clip': 18.459348678588867, 'timing/data': 0.000897631049156189, 'timing/model': 1.25417838152498, 'learning_rate/qwen_vl_interface': 2.9389621533727745e-06, 'learning_rate/action_model': 2.9389621533727747e-05, 'epoch': 1.71}) +[2026-06-18 02:36:11,925][starVLA.training.train_starvla][INFO] - Step 4667, Loss: {'train/loss': 0.37391337752342224, 'train/grad_norm_pre_clip': 21.012374877929688, 'timing/data': 0.0011173048987984657, 'timing/model': 1.0495754797011614, 'learning_rate/qwen_vl_interface': 2.9370729335009185e-06, 'learning_rate/action_model': 2.9370729335009184e-05, 'epoch': 1.71}) +[2026-06-18 02:36:16,447][starVLA.training.train_starvla][INFO] - Step 4668, Loss: {'train/loss': 0.1970096379518509, 'train/grad_norm_pre_clip': 15.217370986938477, 'timing/data': 0.0030280398204922676, 'timing/model': 1.057927569374442, 'learning_rate/qwen_vl_interface': 2.93518419310133e-06, 'learning_rate/action_model': 2.93518419310133e-05, 'epoch': 1.71}) +[2026-06-18 02:36:20,922][starVLA.training.train_starvla][INFO] - Step 4669, Loss: {'train/loss': 0.6050993800163269, 'train/grad_norm_pre_clip': 22.59002685546875, 'timing/data': 0.0009266799315810204, 'timing/model': 1.082024935632944, 'learning_rate/qwen_vl_interface': 2.9332959325655537e-06, 'learning_rate/action_model': 2.9332959325655535e-05, 'epoch': 1.71}) +[2026-06-18 02:36:25,278][starVLA.training.train_starvla][INFO] - Step 4670, Loss: {'train/loss': 0.3303574323654175, 'train/grad_norm_pre_clip': 20.815364837646484, 'timing/data': 0.0007818732410669327, 'timing/model': 1.0353758689016104, 'learning_rate/qwen_vl_interface': 2.9314081522850206e-06, 'learning_rate/action_model': 2.9314081522850205e-05, 'epoch': 1.71}) +[2026-06-18 02:36:30,537][starVLA.training.train_starvla][INFO] - Step 4671, Loss: {'train/loss': 0.4847103953361511, 'train/grad_norm_pre_clip': 28.947540283203125, 'timing/data': 0.000852036289870739, 'timing/model': 1.350024207495153, 'learning_rate/qwen_vl_interface': 2.9295208526510726e-06, 'learning_rate/action_model': 2.9295208526510725e-05, 'epoch': 1.71}) +[2026-06-18 02:36:35,212][starVLA.training.train_starvla][INFO] - Step 4672, Loss: {'train/loss': 0.6426895260810852, 'train/grad_norm_pre_clip': 23.47432518005371, 'timing/data': 0.0018754545599222183, 'timing/model': 1.5497064674273133, 'learning_rate/qwen_vl_interface': 2.9276340340549476e-06, 'learning_rate/action_model': 2.9276340340549473e-05, 'epoch': 1.71}) +[2026-06-18 02:36:39,348][starVLA.training.train_starvla][INFO] - Step 4673, Loss: {'train/loss': 0.3540671169757843, 'train/grad_norm_pre_clip': 90.02110290527344, 'timing/data': 0.0011404920369386673, 'timing/model': 1.0685354471206665, 'learning_rate/qwen_vl_interface': 2.9257476968877875e-06, 'learning_rate/action_model': 2.9257476968877874e-05, 'epoch': 1.71}) +[2026-06-18 02:36:43,846][starVLA.training.train_starvla][INFO] - Step 4674, Loss: {'train/loss': 0.4133668839931488, 'train/grad_norm_pre_clip': 24.4414005279541, 'timing/data': 0.0009770020842552185, 'timing/model': 1.1791518153622746, 'learning_rate/qwen_vl_interface': 2.923861841540631e-06, 'learning_rate/action_model': 2.923861841540631e-05, 'epoch': 1.71}) +[2026-06-18 02:36:48,622][starVLA.training.train_starvla][INFO] - Step 4675, Loss: {'train/loss': 0.15727773308753967, 'train/grad_norm_pre_clip': 17.115360260009766, 'timing/data': 0.05107821151614189, 'timing/model': 1.1190720964223146, 'learning_rate/qwen_vl_interface': 2.9219764684044196e-06, 'learning_rate/action_model': 2.9219764684044193e-05, 'epoch': 1.71}) +[2026-06-18 02:36:53,547][starVLA.training.train_starvla][INFO] - Step 4676, Loss: {'train/loss': 0.1211049035191536, 'train/grad_norm_pre_clip': 10.980621337890625, 'timing/data': 0.0009200870990753174, 'timing/model': 1.1476479405537248, 'learning_rate/qwen_vl_interface': 2.920091577869988e-06, 'learning_rate/action_model': 2.9200915778699877e-05, 'epoch': 1.71}) +[2026-06-18 02:36:58,640][starVLA.training.train_starvla][INFO] - Step 4677, Loss: {'train/loss': 0.4837881624698639, 'train/grad_norm_pre_clip': 37.05256652832031, 'timing/data': 0.0011153770610690117, 'timing/model': 1.3063418557867408, 'learning_rate/qwen_vl_interface': 2.918207170328081e-06, 'learning_rate/action_model': 2.918207170328081e-05, 'epoch': 1.71}) +[2026-06-18 02:37:06,935][starVLA.training.train_starvla][INFO] - Step 4678, Loss: {'train/loss': 0.3447974622249603, 'train/grad_norm_pre_clip': 20.05495834350586, 'timing/data': 0.010556305758655071, 'timing/model': 4.535959420725703, 'learning_rate/qwen_vl_interface': 2.9163232461693373e-06, 'learning_rate/action_model': 2.9163232461693373e-05, 'epoch': 1.71}) +[2026-06-18 02:37:12,126][starVLA.training.train_starvla][INFO] - Step 4679, Loss: {'train/loss': 0.355285108089447, 'train/grad_norm_pre_clip': 19.28802490234375, 'timing/data': 0.0008850060403347015, 'timing/model': 1.0998365385457873, 'learning_rate/qwen_vl_interface': 2.9144398057842948e-06, 'learning_rate/action_model': 2.9144398057842948e-05, 'epoch': 1.71}) +[2026-06-18 02:37:19,976][starVLA.training.train_starvla][INFO] - Step 4680, Loss: {'train/loss': 0.5760630965232849, 'train/grad_norm_pre_clip': 26.962718963623047, 'timing/data': 0.0008405419066548347, 'timing/model': 2.813436420634389, 'learning_rate/qwen_vl_interface': 2.912556849563394e-06, 'learning_rate/action_model': 2.9125568495633943e-05, 'epoch': 1.71}) +[2026-06-18 02:37:31,922][starVLA.training.train_starvla][INFO] - Step 4681, Loss: {'train/loss': 0.391589492559433, 'train/grad_norm_pre_clip': 20.47704315185547, 'timing/data': 0.0007832441478967667, 'timing/model': 1.799749673344195, 'learning_rate/qwen_vl_interface': 2.910674377896974e-06, 'learning_rate/action_model': 2.910674377896974e-05, 'epoch': 1.71}) +[2026-06-18 02:37:38,051][starVLA.training.train_starvla][INFO] - Step 4682, Loss: {'train/loss': 0.708823561668396, 'train/grad_norm_pre_clip': 29.719341278076172, 'timing/data': 0.0010355068370699883, 'timing/model': 2.5429898407310247, 'learning_rate/qwen_vl_interface': 2.9087923911752723e-06, 'learning_rate/action_model': 2.9087923911752725e-05, 'epoch': 1.71}) +[2026-06-18 02:37:42,710][starVLA.training.train_starvla][INFO] - Step 4683, Loss: {'train/loss': 0.2174709439277649, 'train/grad_norm_pre_clip': 30.031787872314453, 'timing/data': 0.0015076985582709312, 'timing/model': 1.0669661248102784, 'learning_rate/qwen_vl_interface': 2.906910889788425e-06, 'learning_rate/action_model': 2.906910889788425e-05, 'epoch': 1.71}) +[2026-06-18 02:37:50,472][starVLA.training.train_starvla][INFO] - Step 4684, Loss: {'train/loss': 0.2316301316022873, 'train/grad_norm_pre_clip': 21.278451919555664, 'timing/data': 0.0013826610520482063, 'timing/model': 1.1269786776974797, 'learning_rate/qwen_vl_interface': 2.905029874126476e-06, 'learning_rate/action_model': 2.905029874126476e-05, 'epoch': 1.72}) +[2026-06-18 02:37:55,428][starVLA.training.train_starvla][INFO] - Step 4685, Loss: {'train/loss': 0.42899608612060547, 'train/grad_norm_pre_clip': 21.306264877319336, 'timing/data': 0.0014194166287779808, 'timing/model': 1.1537975957617164, 'learning_rate/qwen_vl_interface': 2.9031493445793558e-06, 'learning_rate/action_model': 2.9031493445793558e-05, 'epoch': 1.72}) +[2026-06-18 02:37:59,823][starVLA.training.train_starvla][INFO] - Step 4686, Loss: {'train/loss': 0.4999197721481323, 'train/grad_norm_pre_clip': 26.342548370361328, 'timing/data': 0.0011833813041448593, 'timing/model': 1.3613329017534852, 'learning_rate/qwen_vl_interface': 2.9012693015369007e-06, 'learning_rate/action_model': 2.9012693015369008e-05, 'epoch': 1.72}) +[2026-06-18 02:38:11,995][starVLA.training.train_starvla][INFO] - Step 4687, Loss: {'train/loss': 0.25508177280426025, 'train/grad_norm_pre_clip': 19.750791549682617, 'timing/data': 0.0011543137952685356, 'timing/model': 1.3304693223908544, 'learning_rate/qwen_vl_interface': 2.8993897453888454e-06, 'learning_rate/action_model': 2.899389745388845e-05, 'epoch': 1.72}) +[2026-06-18 02:38:16,717][starVLA.training.train_starvla][INFO] - Step 4688, Loss: {'train/loss': 0.25579676032066345, 'train/grad_norm_pre_clip': 34.43977355957031, 'timing/data': 0.0008014645427465439, 'timing/model': 1.517621273174882, 'learning_rate/qwen_vl_interface': 2.8975106765248267e-06, 'learning_rate/action_model': 2.8975106765248267e-05, 'epoch': 1.72}) +[2026-06-18 02:38:21,417][starVLA.training.train_starvla][INFO] - Step 4689, Loss: {'train/loss': 0.4600595235824585, 'train/grad_norm_pre_clip': 30.618043899536133, 'timing/data': 0.0009308569133281708, 'timing/model': 1.11057613696903, 'learning_rate/qwen_vl_interface': 2.895632095334378e-06, 'learning_rate/action_model': 2.895632095334378e-05, 'epoch': 1.72}) +[2026-06-18 02:38:26,915][starVLA.training.train_starvla][INFO] - Step 4690, Loss: {'train/loss': 0.19192685186862946, 'train/grad_norm_pre_clip': 13.263473510742188, 'timing/data': 0.0010504713281989098, 'timing/model': 1.340625929646194, 'learning_rate/qwen_vl_interface': 2.8937540022069267e-06, 'learning_rate/action_model': 2.8937540022069265e-05, 'epoch': 1.72}) +[2026-06-18 02:38:31,893][starVLA.training.train_starvla][INFO] - Step 4691, Loss: {'train/loss': 0.49158570170402527, 'train/grad_norm_pre_clip': 24.637407302856445, 'timing/data': 0.0011672014370560646, 'timing/model': 1.0505113117396832, 'learning_rate/qwen_vl_interface': 2.8918763975318076e-06, 'learning_rate/action_model': 2.8918763975318076e-05, 'epoch': 1.72}) +[2026-06-18 02:38:38,347][starVLA.training.train_starvla][INFO] - Step 4692, Loss: {'train/loss': 0.2711675763130188, 'train/grad_norm_pre_clip': 18.363645553588867, 'timing/data': 0.0013977736234664917, 'timing/model': 1.5884725581854582, 'learning_rate/qwen_vl_interface': 2.889999281698248e-06, 'learning_rate/action_model': 2.889999281698248e-05, 'epoch': 1.72}) +[2026-06-18 02:38:43,198][starVLA.training.train_starvla][INFO] - Step 4693, Loss: {'train/loss': 0.1734396070241928, 'train/grad_norm_pre_clip': 15.627886772155762, 'timing/data': 0.0014076875522732735, 'timing/model': 1.0969111947342753, 'learning_rate/qwen_vl_interface': 2.8881226550953772e-06, 'learning_rate/action_model': 2.888122655095377e-05, 'epoch': 1.72}) +[2026-06-18 02:38:48,241][starVLA.training.train_starvla][INFO] - Step 4694, Loss: {'train/loss': 0.3917054533958435, 'train/grad_norm_pre_clip': 29.393314361572266, 'timing/data': 0.0011233119294047356, 'timing/model': 1.2764744525775313, 'learning_rate/qwen_vl_interface': 2.8862465181122212e-06, 'learning_rate/action_model': 2.886246518112221e-05, 'epoch': 1.72}) +[2026-06-18 02:38:53,715][starVLA.training.train_starvla][INFO] - Step 4695, Loss: {'train/loss': 0.18003101646900177, 'train/grad_norm_pre_clip': 20.91024398803711, 'timing/data': 0.001766006462275982, 'timing/model': 1.0972890090197325, 'learning_rate/qwen_vl_interface': 2.884370871137706e-06, 'learning_rate/action_model': 2.884370871137706e-05, 'epoch': 1.72}) +[2026-06-18 02:39:02,750][starVLA.training.train_starvla][INFO] - Step 4696, Loss: {'train/loss': 0.48403921723365784, 'train/grad_norm_pre_clip': 22.172012329101562, 'timing/data': 0.0009813467040657997, 'timing/model': 1.4909514738246799, 'learning_rate/qwen_vl_interface': 2.8824957145606548e-06, 'learning_rate/action_model': 2.8824957145606548e-05, 'epoch': 1.72}) +[2026-06-18 02:39:07,232][starVLA.training.train_starvla][INFO] - Step 4697, Loss: {'train/loss': 0.39079582691192627, 'train/grad_norm_pre_clip': 27.884004592895508, 'timing/data': 0.0010772105306386948, 'timing/model': 1.3517548758536577, 'learning_rate/qwen_vl_interface': 2.8806210487697866e-06, 'learning_rate/action_model': 2.8806210487697867e-05, 'epoch': 1.72}) +[2026-06-18 02:39:17,811][starVLA.training.train_starvla][INFO] - Step 4698, Loss: {'train/loss': 0.15887556970119476, 'train/grad_norm_pre_clip': 14.221885681152344, 'timing/data': 0.005294473841786385, 'timing/model': 2.596904414705932, 'learning_rate/qwen_vl_interface': 2.87874687415373e-06, 'learning_rate/action_model': 2.87874687415373e-05, 'epoch': 1.72}) +[2026-06-18 02:39:35,409][starVLA.training.train_starvla][INFO] - Step 4699, Loss: {'train/loss': 0.6519319415092468, 'train/grad_norm_pre_clip': 23.160932540893555, 'timing/data': 0.0010071294382214546, 'timing/model': 4.607812994159758, 'learning_rate/qwen_vl_interface': 2.876873191100995e-06, 'learning_rate/action_model': 2.8768731911009948e-05, 'epoch': 1.72}) +[2026-06-18 02:43:45,392][starVLA.training.train_starvla][INFO] - Step 4700, Loss: {'train/loss': 0.3647761344909668, 'train/grad_norm_pre_clip': 21.194746017456055, 'eval/loss': 0.5377618670463562, 'eval/latency_2/loss': 0.5377618670463562, 'eval/demon_fire_precision': 0.9446755407653911, 'eval/demon_fire_recall': 0.9502092050209205, 'eval/demon_fire_f1': 0.9474342928660827, 'eval/demon_fire_rate_model': 0.6688925987757374, 'eval/demon_fire_rate_teacher': 0.6649972175848636, 'eval/demon_left_precision': 0.8966630785791173, 'eval/demon_left_recall': 0.8768421052631579, 'eval/demon_left_f1': 0.886641830761043, 'eval/demon_left_rate_model': 0.2584863661658319, 'eval/demon_left_rate_teacher': 0.2643294379521425, 'eval/demon_right_precision': 0.9102870813397129, 'eval/demon_right_recall': 0.8598870056497175, 'eval/demon_right_f1': 0.884369552585706, 'eval/demon_right_rate_model': 0.23260990539788537, 'eval/demon_right_rate_teacher': 0.24624373956594323, 'eval/demon_move_f1': 0.8855056916733746, 'eval/demon_cc_f1': 0.9279648864309794, 'eval/latency_2/demon_fire_precision': 0.9446755407653911, 'eval/latency_2/demon_fire_recall': 0.9502092050209205, 'eval/latency_2/demon_fire_f1': 0.9474342928660827, 'eval/latency_2/demon_fire_rate_model': 0.6688925987757374, 'eval/latency_2/demon_fire_rate_teacher': 0.6649972175848636, 'eval/latency_2/demon_left_precision': 0.8966630785791173, 'eval/latency_2/demon_left_recall': 0.8768421052631579, 'eval/latency_2/demon_left_f1': 0.886641830761043, 'eval/latency_2/demon_left_rate_model': 0.2584863661658319, 'eval/latency_2/demon_left_rate_teacher': 0.2643294379521425, 'eval/latency_2/demon_right_precision': 0.9102870813397129, 'eval/latency_2/demon_right_recall': 0.8598870056497175, 'eval/latency_2/demon_right_f1': 0.884369552585706, 'eval/latency_2/demon_right_rate_model': 0.23260990539788537, 'eval/latency_2/demon_right_rate_teacher': 0.24624373956594323, 'eval/latency_2/demon_move_f1': 0.8855056916733746, 'eval/latency_2/demon_cc_f1': 0.9279648864309794, 'timing/data': 0.0007762331515550613, 'timing/model': 1.1285615405067801, 'learning_rate/qwen_vl_interface': 2.8750000000000013e-06, 'learning_rate/action_model': 2.875000000000001e-05, 'epoch': 1.72}) +[2026-06-18 02:43:45,418][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_4700_state +[2026-06-18 02:43:45,427][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 02:43:54,009][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_4700_state/model.safetensors +[2026-06-18 02:44:09,173][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_4700_state/optimizer.bin +[2026-06-18 02:44:09,205][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_4700_state/sampler.bin +[2026-06-18 02:44:09,205][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_4700_state/sampler_1.bin +[2026-06-18 02:44:09,209][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_4700_state/random_states_0.pkl +[2026-06-18 02:44:09,218][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4700 +[2026-06-18 02:44:09,219][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 02:44:09,242][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 02:44:14,160][starVLA.training.train_starvla][INFO] - Step 4701, Loss: {'train/loss': 0.20208990573883057, 'train/grad_norm_pre_clip': 12.98967170715332, 'timing/data': 0.0013899654150009155, 'timing/model': 1.0441429214552045, 'learning_rate/qwen_vl_interface': 2.8731273012390614e-06, 'learning_rate/action_model': 2.8731273012390615e-05, 'epoch': 1.72}) +[2026-06-18 02:44:18,613][starVLA.training.train_starvla][INFO] - Step 4702, Loss: {'train/loss': 0.18210004270076752, 'train/grad_norm_pre_clip': 11.455024719238281, 'timing/data': 0.001129964366555214, 'timing/model': 1.0465868078172207, 'learning_rate/qwen_vl_interface': 2.8712550952063912e-06, 'learning_rate/action_model': 2.871255095206391e-05, 'epoch': 1.72}) +[2026-06-18 02:44:28,331][starVLA.training.train_starvla][INFO] - Step 4703, Loss: {'train/loss': 0.22129014134407043, 'train/grad_norm_pre_clip': 21.636754989624023, 'timing/data': 0.011997115798294544, 'timing/model': 5.1353117255494, 'learning_rate/qwen_vl_interface': 2.8693833822901012e-06, 'learning_rate/action_model': 2.8693833822901012e-05, 'epoch': 1.72}) +[2026-06-18 02:44:44,823][starVLA.training.train_starvla][INFO] - Step 4704, Loss: {'train/loss': 0.4878605902194977, 'train/grad_norm_pre_clip': 27.490816116333008, 'timing/data': 0.013474860228598118, 'timing/model': 5.551795491948724, 'learning_rate/qwen_vl_interface': 2.867512162878196e-06, 'learning_rate/action_model': 2.867512162878196e-05, 'epoch': 1.72}) +[2026-06-18 02:44:55,716][starVLA.training.train_starvla][INFO] - Step 4705, Loss: {'train/loss': 0.19382070004940033, 'train/grad_norm_pre_clip': 17.414194107055664, 'timing/data': 0.0169739481061697, 'timing/model': 1.2445673858746886, 'learning_rate/qwen_vl_interface': 2.8656414373585812e-06, 'learning_rate/action_model': 2.865641437358581e-05, 'epoch': 1.72}) +[2026-06-18 02:45:08,660][starVLA.training.train_starvla][INFO] - Step 4706, Loss: {'train/loss': 0.39321160316467285, 'train/grad_norm_pre_clip': 23.21002197265625, 'timing/data': 0.022047302685678005, 'timing/model': 1.142531482502818, 'learning_rate/qwen_vl_interface': 2.8637712061190644e-06, 'learning_rate/action_model': 2.863771206119064e-05, 'epoch': 1.72}) +[2026-06-18 02:45:14,197][starVLA.training.train_starvla][INFO] - Step 4707, Loss: {'train/loss': 0.38858163356781006, 'train/grad_norm_pre_clip': 23.134723663330078, 'timing/data': 0.011746961623430252, 'timing/model': 1.037846276536584, 'learning_rate/qwen_vl_interface': 2.861901469547344e-06, 'learning_rate/action_model': 2.861901469547344e-05, 'epoch': 1.72}) +[2026-06-18 02:45:22,117][starVLA.training.train_starvla][INFO] - Step 4708, Loss: {'train/loss': 0.43344101309776306, 'train/grad_norm_pre_clip': 28.69700050354004, 'timing/data': 0.011326367035508156, 'timing/model': 1.4594290172681212, 'learning_rate/qwen_vl_interface': 2.8600322280310185e-06, 'learning_rate/action_model': 2.8600322280310183e-05, 'epoch': 1.72}) +[2026-06-18 02:45:32,035][starVLA.training.train_starvla][INFO] - Step 4709, Loss: {'train/loss': 0.34492477774620056, 'train/grad_norm_pre_clip': 21.219751358032227, 'timing/data': 0.05075542815029621, 'timing/model': 1.831630503758788, 'learning_rate/qwen_vl_interface': 2.8581634819575843e-06, 'learning_rate/action_model': 2.8581634819575843e-05, 'epoch': 1.72}) +[2026-06-18 02:45:39,977][starVLA.training.train_starvla][INFO] - Step 4710, Loss: {'train/loss': 0.30756905674934387, 'train/grad_norm_pre_clip': 26.630931854248047, 'timing/data': 0.012761279009282589, 'timing/model': 1.139687841758132, 'learning_rate/qwen_vl_interface': 2.856295231714434e-06, 'learning_rate/action_model': 2.8562952317144337e-05, 'epoch': 1.72}) +[2026-06-18 02:45:55,466][starVLA.training.train_starvla][INFO] - Step 4711, Loss: {'train/loss': 0.5235114097595215, 'train/grad_norm_pre_clip': 22.491607666015625, 'timing/data': 0.013762823306024075, 'timing/model': 4.6086468286812305, 'learning_rate/qwen_vl_interface': 2.854427477688857e-06, 'learning_rate/action_model': 2.854427477688857e-05, 'epoch': 1.73}) +[2026-06-18 02:46:13,372][starVLA.training.train_starvla][INFO] - Step 4712, Loss: {'train/loss': 0.3194931745529175, 'train/grad_norm_pre_clip': 22.19137954711914, 'timing/data': 0.05484528001397848, 'timing/model': 3.9682724438607693, 'learning_rate/qwen_vl_interface': 2.8525602202680424e-06, 'learning_rate/action_model': 2.8525602202680424e-05, 'epoch': 1.73}) +[2026-06-18 02:46:20,667][starVLA.training.train_starvla][INFO] - Step 4713, Loss: {'train/loss': 0.28817635774612427, 'train/grad_norm_pre_clip': 16.153287887573242, 'timing/data': 0.012005654163658619, 'timing/model': 1.836532854475081, 'learning_rate/qwen_vl_interface': 2.8506934598390736e-06, 'learning_rate/action_model': 2.8506934598390735e-05, 'epoch': 1.73}) +[2026-06-18 02:46:26,899][starVLA.training.train_starvla][INFO] - Step 4714, Loss: {'train/loss': 0.30041852593421936, 'train/grad_norm_pre_clip': 19.899829864501953, 'timing/data': 0.01074934285134077, 'timing/model': 1.0347878830507398, 'learning_rate/qwen_vl_interface': 2.848827196788932e-06, 'learning_rate/action_model': 2.848827196788932e-05, 'epoch': 1.73}) +[2026-06-18 02:46:31,879][starVLA.training.train_starvla][INFO] - Step 4715, Loss: {'train/loss': 0.2548494040966034, 'train/grad_norm_pre_clip': 16.7408390045166, 'timing/data': 0.013361630029976368, 'timing/model': 1.2919867457821965, 'learning_rate/qwen_vl_interface': 2.8469614315044975e-06, 'learning_rate/action_model': 2.8469614315044975e-05, 'epoch': 1.73}) +[2026-06-18 02:46:36,512][starVLA.training.train_starvla][INFO] - Step 4716, Loss: {'train/loss': 0.19287942349910736, 'train/grad_norm_pre_clip': 13.371499061584473, 'timing/data': 0.012008928693830967, 'timing/model': 1.1187407542020082, 'learning_rate/qwen_vl_interface': 2.8450961643725405e-06, 'learning_rate/action_model': 2.8450961643725405e-05, 'epoch': 1.73}) +[2026-06-18 02:46:41,062][starVLA.training.train_starvla][INFO] - Step 4717, Loss: {'train/loss': 0.3670814633369446, 'train/grad_norm_pre_clip': 26.818607330322266, 'timing/data': 0.011438561603426933, 'timing/model': 1.0653832852840424, 'learning_rate/qwen_vl_interface': 2.8432313957797415e-06, 'learning_rate/action_model': 2.8432313957797414e-05, 'epoch': 1.73}) +[2026-06-18 02:46:45,652][starVLA.training.train_starvla][INFO] - Step 4718, Loss: {'train/loss': 0.40750980377197266, 'train/grad_norm_pre_clip': 29.844341278076172, 'timing/data': 0.011740167625248432, 'timing/model': 1.4314368842169642, 'learning_rate/qwen_vl_interface': 2.8413671261126613e-06, 'learning_rate/action_model': 2.8413671261126613e-05, 'epoch': 1.73}) +[2026-06-18 02:46:49,922][starVLA.training.train_starvla][INFO] - Step 4719, Loss: {'train/loss': 0.30733343958854675, 'train/grad_norm_pre_clip': 25.347440719604492, 'timing/data': 0.01251417025923729, 'timing/model': 1.0558464992791414, 'learning_rate/qwen_vl_interface': 2.8395033557577668e-06, 'learning_rate/action_model': 2.8395033557577668e-05, 'epoch': 1.73}) +[2026-06-18 02:46:54,359][starVLA.training.train_starvla][INFO] - Step 4720, Loss: {'train/loss': 0.33704784512519836, 'train/grad_norm_pre_clip': 20.66728973388672, 'timing/data': 0.020616812631487846, 'timing/model': 1.0368698351085186, 'learning_rate/qwen_vl_interface': 2.8376400851014214e-06, 'learning_rate/action_model': 2.8376400851014214e-05, 'epoch': 1.73}) +[2026-06-18 02:46:59,192][starVLA.training.train_starvla][INFO] - Step 4721, Loss: {'train/loss': 0.2859725058078766, 'train/grad_norm_pre_clip': 17.492473602294922, 'timing/data': 0.0118411211296916, 'timing/model': 1.1378297097980976, 'learning_rate/qwen_vl_interface': 2.8357773145298823e-06, 'learning_rate/action_model': 2.8357773145298823e-05, 'epoch': 1.73}) +[2026-06-18 02:47:04,375][starVLA.training.train_starvla][INFO] - Step 4722, Loss: {'train/loss': 0.2457837462425232, 'train/grad_norm_pre_clip': 15.52898120880127, 'timing/data': 0.010530550964176655, 'timing/model': 1.4267396898940206, 'learning_rate/qwen_vl_interface': 2.8339150444293055e-06, 'learning_rate/action_model': 2.8339150444293055e-05, 'epoch': 1.73}) +[2026-06-18 02:47:16,992][starVLA.training.train_starvla][INFO] - Step 4723, Loss: {'train/loss': 0.3944389522075653, 'train/grad_norm_pre_clip': 20.529640197753906, 'timing/data': 0.013610320165753365, 'timing/model': 1.307313108816743, 'learning_rate/qwen_vl_interface': 2.8320532751857365e-06, 'learning_rate/action_model': 2.8320532751857366e-05, 'epoch': 1.73}) +[2026-06-18 02:47:26,737][starVLA.training.train_starvla][INFO] - Step 4724, Loss: {'train/loss': 0.39181315898895264, 'train/grad_norm_pre_clip': 28.632837295532227, 'timing/data': 0.02193850465118885, 'timing/model': 5.919420655816793, 'learning_rate/qwen_vl_interface': 2.830192007185127e-06, 'learning_rate/action_model': 2.830192007185127e-05, 'epoch': 1.73}) +[2026-06-18 02:47:31,884][starVLA.training.train_starvla][INFO] - Step 4725, Loss: {'train/loss': 0.15585942566394806, 'train/grad_norm_pre_clip': 14.467829704284668, 'timing/data': 0.010093524120748043, 'timing/model': 1.5015513990074396, 'learning_rate/qwen_vl_interface': 2.828331240813318e-06, 'learning_rate/action_model': 2.8283312408133176e-05, 'epoch': 1.73}) +[2026-06-18 02:47:37,052][starVLA.training.train_starvla][INFO] - Step 4726, Loss: {'train/loss': 0.20633167028427124, 'train/grad_norm_pre_clip': 16.46849822998047, 'timing/data': 0.012669120915234089, 'timing/model': 1.0557123105973005, 'learning_rate/qwen_vl_interface': 2.826470976456049e-06, 'learning_rate/action_model': 2.826470976456049e-05, 'epoch': 1.73}) +[2026-06-18 02:47:42,184][starVLA.training.train_starvla][INFO] - Step 4727, Loss: {'train/loss': 0.21777760982513428, 'train/grad_norm_pre_clip': 19.452489852905273, 'timing/data': 0.010764635168015957, 'timing/model': 1.4026431562379003, 'learning_rate/qwen_vl_interface': 2.8246112144989548e-06, 'learning_rate/action_model': 2.8246112144989543e-05, 'epoch': 1.73}) +[2026-06-18 02:47:48,827][starVLA.training.train_starvla][INFO] - Step 4728, Loss: {'train/loss': 0.4057760536670685, 'train/grad_norm_pre_clip': 16.142282485961914, 'timing/data': 0.014360996894538403, 'timing/model': 3.256086952984333, 'learning_rate/qwen_vl_interface': 2.8227519553275644e-06, 'learning_rate/action_model': 2.8227519553275645e-05, 'epoch': 1.73}) +[2026-06-18 02:47:53,506][starVLA.training.train_starvla][INFO] - Step 4729, Loss: {'train/loss': 0.5079408884048462, 'train/grad_norm_pre_clip': 27.461759567260742, 'timing/data': 0.01747952215373516, 'timing/model': 1.0729249911382794, 'learning_rate/qwen_vl_interface': 2.8208931993273066e-06, 'learning_rate/action_model': 2.8208931993273062e-05, 'epoch': 1.73}) +[2026-06-18 02:47:58,732][starVLA.training.train_starvla][INFO] - Step 4730, Loss: {'train/loss': 0.27100637555122375, 'train/grad_norm_pre_clip': 17.060447692871094, 'timing/data': 0.012977457605302334, 'timing/model': 1.1790623320266604, 'learning_rate/qwen_vl_interface': 2.8190349468835006e-06, 'learning_rate/action_model': 2.8190349468835004e-05, 'epoch': 1.73}) +[2026-06-18 02:48:03,531][starVLA.training.train_starvla][INFO] - Step 4731, Loss: {'train/loss': 0.4674522876739502, 'train/grad_norm_pre_clip': 19.630950927734375, 'timing/data': 0.015560444444417953, 'timing/model': 1.102680742740631, 'learning_rate/qwen_vl_interface': 2.817177198381369e-06, 'learning_rate/action_model': 2.817177198381369e-05, 'epoch': 1.73}) +[2026-06-18 02:48:08,205][starVLA.training.train_starvla][INFO] - Step 4732, Loss: {'train/loss': 0.648400068283081, 'train/grad_norm_pre_clip': 26.776004791259766, 'timing/data': 0.011773183941841125, 'timing/model': 1.0707567427307367, 'learning_rate/qwen_vl_interface': 2.8153199542060206e-06, 'learning_rate/action_model': 2.8153199542060206e-05, 'epoch': 1.73}) +[2026-06-18 02:48:13,564][starVLA.training.train_starvla][INFO] - Step 4733, Loss: {'train/loss': 0.24514126777648926, 'train/grad_norm_pre_clip': 15.320340156555176, 'timing/data': 0.01347129512578249, 'timing/model': 1.8858944438397884, 'learning_rate/qwen_vl_interface': 2.813463214742465e-06, 'learning_rate/action_model': 2.813463214742465e-05, 'epoch': 1.73}) +[2026-06-18 02:48:18,180][starVLA.training.train_starvla][INFO] - Step 4734, Loss: {'train/loss': 0.31968069076538086, 'train/grad_norm_pre_clip': 19.865304946899414, 'timing/data': 0.013034502975642681, 'timing/model': 1.265564588829875, 'learning_rate/qwen_vl_interface': 2.8116069803756047e-06, 'learning_rate/action_model': 2.8116069803756047e-05, 'epoch': 1.73}) +[2026-06-18 02:48:23,187][starVLA.training.train_starvla][INFO] - Step 4735, Loss: {'train/loss': 0.3112275302410126, 'train/grad_norm_pre_clip': 19.277034759521484, 'timing/data': 0.016880017705261707, 'timing/model': 1.1421504952013493, 'learning_rate/qwen_vl_interface': 2.809751251490244e-06, 'learning_rate/action_model': 2.809751251490244e-05, 'epoch': 1.73}) +[2026-06-18 02:48:28,064][starVLA.training.train_starvla][INFO] - Step 4736, Loss: {'train/loss': 0.323900043964386, 'train/grad_norm_pre_clip': 19.87360191345215, 'timing/data': 0.012782357633113861, 'timing/model': 1.0801758654415607, 'learning_rate/qwen_vl_interface': 2.807896028471076e-06, 'learning_rate/action_model': 2.807896028471076e-05, 'epoch': 1.73}) +[2026-06-18 02:48:33,992][starVLA.training.train_starvla][INFO] - Step 4737, Loss: {'train/loss': 0.39473187923431396, 'train/grad_norm_pre_clip': 20.261531829833984, 'timing/data': 0.016476266086101532, 'timing/model': 1.043633739463985, 'learning_rate/qwen_vl_interface': 2.806041311702688e-06, 'learning_rate/action_model': 2.806041311702688e-05, 'epoch': 1.73}) +[2026-06-18 02:48:40,189][starVLA.training.train_starvla][INFO] - Step 4738, Loss: {'train/loss': 0.12180152535438538, 'train/grad_norm_pre_clip': 9.169994354248047, 'timing/data': 0.017355268821120262, 'timing/model': 1.3691829582676291, 'learning_rate/qwen_vl_interface': 2.8041871015695626e-06, 'learning_rate/action_model': 2.8041871015695626e-05, 'epoch': 1.73}) +[2026-06-18 02:48:44,415][starVLA.training.train_starvla][INFO] - Step 4739, Loss: {'train/loss': 0.4102047085762024, 'train/grad_norm_pre_clip': 23.75627899169922, 'timing/data': 0.011008557863533497, 'timing/model': 1.0651953080669045, 'learning_rate/qwen_vl_interface': 2.802333398456086e-06, 'learning_rate/action_model': 2.8023333984560856e-05, 'epoch': 1.74}) +[2026-06-18 02:48:48,849][starVLA.training.train_starvla][INFO] - Step 4740, Loss: {'train/loss': 0.5218976140022278, 'train/grad_norm_pre_clip': 28.21609878540039, 'timing/data': 0.011064586229622364, 'timing/model': 1.1041001975536346, 'learning_rate/qwen_vl_interface': 2.8004802027465296e-06, 'learning_rate/action_model': 2.8004802027465293e-05, 'epoch': 1.74}) +[2026-06-18 02:48:56,197][starVLA.training.train_starvla][INFO] - Step 4741, Loss: {'train/loss': 0.28224921226501465, 'train/grad_norm_pre_clip': 27.18657875061035, 'timing/data': 0.009825129993259907, 'timing/model': 1.236055415123701, 'learning_rate/qwen_vl_interface': 2.7986275148250624e-06, 'learning_rate/action_model': 2.7986275148250622e-05, 'epoch': 1.74}) +[2026-06-18 02:49:01,395][starVLA.training.train_starvla][INFO] - Step 4742, Loss: {'train/loss': 0.31687912344932556, 'train/grad_norm_pre_clip': 20.325605392456055, 'timing/data': 0.015265506692230701, 'timing/model': 1.1933717038482428, 'learning_rate/qwen_vl_interface': 2.7967753350757498e-06, 'learning_rate/action_model': 2.7967753350757497e-05, 'epoch': 1.74}) +[2026-06-18 02:49:06,109][starVLA.training.train_starvla][INFO] - Step 4743, Loss: {'train/loss': 0.3391742706298828, 'train/grad_norm_pre_clip': 24.867746353149414, 'timing/data': 0.011536031030118465, 'timing/model': 1.3910262873396277, 'learning_rate/qwen_vl_interface': 2.7949236638825496e-06, 'learning_rate/action_model': 2.7949236638825495e-05, 'epoch': 1.74}) +[2026-06-18 02:49:10,669][starVLA.training.train_starvla][INFO] - Step 4744, Loss: {'train/loss': 0.4346567988395691, 'train/grad_norm_pre_clip': 23.781511306762695, 'timing/data': 0.013541926629841328, 'timing/model': 1.0459616472944617, 'learning_rate/qwen_vl_interface': 2.7930725016293155e-06, 'learning_rate/action_model': 2.793072501629315e-05, 'epoch': 1.74}) +[2026-06-18 02:49:19,992][starVLA.training.train_starvla][INFO] - Step 4745, Loss: {'train/loss': 0.6955279111862183, 'train/grad_norm_pre_clip': 27.49587631225586, 'timing/data': 0.01284477487206459, 'timing/model': 1.9988061906769872, 'learning_rate/qwen_vl_interface': 2.7912218486997954e-06, 'learning_rate/action_model': 2.7912218486997954e-05, 'epoch': 1.74}) +[2026-06-18 02:49:24,790][starVLA.training.train_starvla][INFO] - Step 4746, Loss: {'train/loss': 0.1626029908657074, 'train/grad_norm_pre_clip': 15.20669174194336, 'timing/data': 0.017267702147364616, 'timing/model': 1.1393728023394942, 'learning_rate/qwen_vl_interface': 2.7893717054776327e-06, 'learning_rate/action_model': 2.7893717054776327e-05, 'epoch': 1.74}) +[2026-06-18 02:49:31,255][starVLA.training.train_starvla][INFO] - Step 4747, Loss: {'train/loss': 0.3437657952308655, 'train/grad_norm_pre_clip': 21.172574996948242, 'timing/data': 0.015514852479100227, 'timing/model': 1.041194067336619, 'learning_rate/qwen_vl_interface': 2.7875220723463624e-06, 'learning_rate/action_model': 2.7875220723463625e-05, 'epoch': 1.74}) +[2026-06-18 02:49:42,198][starVLA.training.train_starvla][INFO] - Step 4748, Loss: {'train/loss': 0.47404199838638306, 'train/grad_norm_pre_clip': 29.073204040527344, 'timing/data': 0.01713968999683857, 'timing/model': 3.095393605530262, 'learning_rate/qwen_vl_interface': 2.7856729496894165e-06, 'learning_rate/action_model': 2.7856729496894162e-05, 'epoch': 1.74}) +[2026-06-18 02:49:46,269][starVLA.training.train_starvla][INFO] - Step 4749, Loss: {'train/loss': 0.4604704976081848, 'train/grad_norm_pre_clip': 23.920373916625977, 'timing/data': 0.012493656948208809, 'timing/model': 1.0440925303846598, 'learning_rate/qwen_vl_interface': 2.783824337890118e-06, 'learning_rate/action_model': 2.783824337890118e-05, 'epoch': 1.74}) +[2026-06-18 02:49:50,892][starVLA.training.train_starvla][INFO] - Step 4750, Loss: {'train/loss': 0.40675559639930725, 'train/grad_norm_pre_clip': 20.400314331054688, 'timing/data': 0.012604130432009697, 'timing/model': 1.0941169643774629, 'learning_rate/qwen_vl_interface': 2.7819762373316916e-06, 'learning_rate/action_model': 2.7819762373316917e-05, 'epoch': 1.74}) +[2026-06-18 02:49:55,527][starVLA.training.train_starvla][INFO] - Step 4751, Loss: {'train/loss': 0.2635576128959656, 'train/grad_norm_pre_clip': 21.206802368164062, 'timing/data': 0.012111562304198742, 'timing/model': 1.0815800270065665, 'learning_rate/qwen_vl_interface': 2.7801286483972458e-06, 'learning_rate/action_model': 2.7801286483972457e-05, 'epoch': 1.74}) +[2026-06-18 02:50:00,000][starVLA.training.train_starvla][INFO] - Step 4752, Loss: {'train/loss': 0.312528520822525, 'train/grad_norm_pre_clip': 18.566944122314453, 'timing/data': 0.014855842106044292, 'timing/model': 1.0330982496961951, 'learning_rate/qwen_vl_interface': 2.778281571469786e-06, 'learning_rate/action_model': 2.7782815714697858e-05, 'epoch': 1.74}) +[2026-06-18 02:50:08,085][starVLA.training.train_starvla][INFO] - Step 4753, Loss: {'train/loss': 0.7153795957565308, 'train/grad_norm_pre_clip': 39.00979232788086, 'timing/data': 0.015706857666373253, 'timing/model': 1.6408770391717553, 'learning_rate/qwen_vl_interface': 2.776435006932219e-06, 'learning_rate/action_model': 2.7764350069322187e-05, 'epoch': 1.74}) +[2026-06-18 02:50:13,343][starVLA.training.train_starvla][INFO] - Step 4754, Loss: {'train/loss': 0.5747256278991699, 'train/grad_norm_pre_clip': 25.65418243408203, 'timing/data': 0.055636472068727016, 'timing/model': 1.5415358124300838, 'learning_rate/qwen_vl_interface': 2.7745889551673355e-06, 'learning_rate/action_model': 2.7745889551673353e-05, 'epoch': 1.74}) +[2026-06-18 02:50:24,517][starVLA.training.train_starvla][INFO] - Step 4755, Loss: {'train/loss': 0.26312264800071716, 'train/grad_norm_pre_clip': 18.090606689453125, 'timing/data': 0.01143800187855959, 'timing/model': 2.541968616656959, 'learning_rate/qwen_vl_interface': 2.772743416557826e-06, 'learning_rate/action_model': 2.772743416557826e-05, 'epoch': 1.74}) +[2026-06-18 02:50:33,210][starVLA.training.train_starvla][INFO] - Step 4756, Loss: {'train/loss': 0.2782687544822693, 'train/grad_norm_pre_clip': 21.101436614990234, 'timing/data': 0.01943830866366625, 'timing/model': 1.08912521507591, 'learning_rate/qwen_vl_interface': 2.770898391486272e-06, 'learning_rate/action_model': 2.7708983914862718e-05, 'epoch': 1.74}) +[2026-06-18 02:50:38,634][starVLA.training.train_starvla][INFO] - Step 4757, Loss: {'train/loss': 0.22561916708946228, 'train/grad_norm_pre_clip': 17.652469635009766, 'timing/data': 0.015566625632345676, 'timing/model': 1.5385665176436305, 'learning_rate/qwen_vl_interface': 2.7690538803351484e-06, 'learning_rate/action_model': 2.7690538803351484e-05, 'epoch': 1.74}) +[2026-06-18 02:50:43,372][starVLA.training.train_starvla][INFO] - Step 4758, Loss: {'train/loss': 0.14476297795772552, 'train/grad_norm_pre_clip': 13.076019287109375, 'timing/data': 0.012564890086650848, 'timing/model': 1.5123679619282484, 'learning_rate/qwen_vl_interface': 2.7672098834868253e-06, 'learning_rate/action_model': 2.7672098834868253e-05, 'epoch': 1.74}) +[2026-06-18 02:50:48,245][starVLA.training.train_starvla][INFO] - Step 4759, Loss: {'train/loss': 0.20164141058921814, 'train/grad_norm_pre_clip': 12.557687759399414, 'timing/data': 0.014388802461326122, 'timing/model': 1.284669135697186, 'learning_rate/qwen_vl_interface': 2.7653664013235655e-06, 'learning_rate/action_model': 2.7653664013235653e-05, 'epoch': 1.74}) +[2026-06-18 02:50:52,959][starVLA.training.train_starvla][INFO] - Step 4760, Loss: {'train/loss': 0.1471230536699295, 'train/grad_norm_pre_clip': 13.514606475830078, 'timing/data': 0.018561284057796, 'timing/model': 1.272357259877026, 'learning_rate/qwen_vl_interface': 2.763523434227524e-06, 'learning_rate/action_model': 2.763523434227524e-05, 'epoch': 1.74}) +[2026-06-18 02:51:02,033][starVLA.training.train_starvla][INFO] - Step 4761, Loss: {'train/loss': 0.1781492829322815, 'train/grad_norm_pre_clip': 12.412922859191895, 'timing/data': 0.014599696733057499, 'timing/model': 5.2970491871237755, 'learning_rate/qwen_vl_interface': 2.76168098258075e-06, 'learning_rate/action_model': 2.76168098258075e-05, 'epoch': 1.74}) +[2026-06-18 02:51:08,665][starVLA.training.train_starvla][INFO] - Step 4762, Loss: {'train/loss': 0.3054109811782837, 'train/grad_norm_pre_clip': 18.56930160522461, 'timing/data': 0.015613345429301262, 'timing/model': 1.4799930099397898, 'learning_rate/qwen_vl_interface': 2.7598390467651865e-06, 'learning_rate/action_model': 2.7598390467651863e-05, 'epoch': 1.74}) +[2026-06-18 02:51:14,424][starVLA.training.train_starvla][INFO] - Step 4763, Loss: {'train/loss': 0.20509707927703857, 'train/grad_norm_pre_clip': 12.97067928314209, 'timing/data': 0.013887178152799606, 'timing/model': 2.4548554522916675, 'learning_rate/qwen_vl_interface': 2.757997627162665e-06, 'learning_rate/action_model': 2.7579976271626652e-05, 'epoch': 1.74}) +[2026-06-18 02:51:19,874][starVLA.training.train_starvla][INFO] - Step 4764, Loss: {'train/loss': 0.4747428297996521, 'train/grad_norm_pre_clip': 32.8254508972168, 'timing/data': 0.0014171060174703598, 'timing/model': 1.0397453987970948, 'learning_rate/qwen_vl_interface': 2.7561567241549203e-06, 'learning_rate/action_model': 2.7561567241549202e-05, 'epoch': 1.74}) +[2026-06-18 02:51:24,523][starVLA.training.train_starvla][INFO] - Step 4765, Loss: {'train/loss': 0.3291991353034973, 'train/grad_norm_pre_clip': 23.12000274658203, 'timing/data': 0.0011491794139146805, 'timing/model': 1.055088410153985, 'learning_rate/qwen_vl_interface': 2.7543163381235705e-06, 'learning_rate/action_model': 2.7543163381235702e-05, 'epoch': 1.74}) +[2026-06-18 02:51:33,669][starVLA.training.train_starvla][INFO] - Step 4766, Loss: {'train/loss': 0.20350083708763123, 'train/grad_norm_pre_clip': 15.993966102600098, 'timing/data': 0.0010342616587877274, 'timing/model': 1.2095541674643755, 'learning_rate/qwen_vl_interface': 2.752476469450128e-06, 'learning_rate/action_model': 2.7524764694501283e-05, 'epoch': 1.75}) +[2026-06-18 02:51:40,155][starVLA.training.train_starvla][INFO] - Step 4767, Loss: {'train/loss': 0.43762844800949097, 'train/grad_norm_pre_clip': 26.049285888671875, 'timing/data': 0.0014575980603694916, 'timing/model': 1.2406221171841025, 'learning_rate/qwen_vl_interface': 2.750637118515999e-06, 'learning_rate/action_model': 2.750637118515999e-05, 'epoch': 1.75}) +[2026-06-18 02:51:44,727][starVLA.training.train_starvla][INFO] - Step 4768, Loss: {'train/loss': 0.19594706594944, 'train/grad_norm_pre_clip': 21.34076499938965, 'timing/data': 0.0010773343965411186, 'timing/model': 1.43430844694376, 'learning_rate/qwen_vl_interface': 2.7487982857024875e-06, 'learning_rate/action_model': 2.7487982857024876e-05, 'epoch': 1.75}) +[2026-06-18 02:51:54,138][starVLA.training.train_starvla][INFO] - Step 4769, Loss: {'train/loss': 0.15205034613609314, 'train/grad_norm_pre_clip': 12.680498123168945, 'timing/data': 0.00074776541441679, 'timing/model': 1.1105300271883607, 'learning_rate/qwen_vl_interface': 2.746959971390782e-06, 'learning_rate/action_model': 2.7469599713907817e-05, 'epoch': 1.75}) +[2026-06-18 02:51:58,730][starVLA.training.train_starvla][INFO] - Step 4770, Loss: {'train/loss': 0.20750054717063904, 'train/grad_norm_pre_clip': 16.246625900268555, 'timing/data': 0.0013470742851495743, 'timing/model': 1.0938349524512887, 'learning_rate/qwen_vl_interface': 2.745122175961967e-06, 'learning_rate/action_model': 2.7451221759619672e-05, 'epoch': 1.75}) +[2026-06-18 02:52:04,220][starVLA.training.train_starvla][INFO] - Step 4771, Loss: {'train/loss': 0.4921015202999115, 'train/grad_norm_pre_clip': 28.60874366760254, 'timing/data': 0.0012312596663832664, 'timing/model': 1.2225839449092746, 'learning_rate/qwen_vl_interface': 2.743284899797022e-06, 'learning_rate/action_model': 2.7432848997970218e-05, 'epoch': 1.75}) +[2026-06-18 02:52:08,837][starVLA.training.train_starvla][INFO] - Step 4772, Loss: {'train/loss': 0.4338173568248749, 'train/grad_norm_pre_clip': 21.956968307495117, 'timing/data': 0.0009069759398698807, 'timing/model': 1.1580607127398252, 'learning_rate/qwen_vl_interface': 2.741448143276813e-06, 'learning_rate/action_model': 2.741448143276813e-05, 'epoch': 1.75}) +[2026-06-18 02:52:14,194][starVLA.training.train_starvla][INFO] - Step 4773, Loss: {'train/loss': 0.3737696409225464, 'train/grad_norm_pre_clip': 29.084970474243164, 'timing/data': 0.0008543506264686584, 'timing/model': 1.4844324560835958, 'learning_rate/qwen_vl_interface': 2.7396119067821043e-06, 'learning_rate/action_model': 2.7396119067821043e-05, 'epoch': 1.75}) +[2026-06-18 02:52:18,399][starVLA.training.train_starvla][INFO] - Step 4774, Loss: {'train/loss': 0.34251669049263, 'train/grad_norm_pre_clip': 33.6033821105957, 'timing/data': 0.0012916633859276772, 'timing/model': 1.0828579710796475, 'learning_rate/qwen_vl_interface': 2.737776190693548e-06, 'learning_rate/action_model': 2.7377761906935477e-05, 'epoch': 1.75}) +[2026-06-18 02:52:24,426][starVLA.training.train_starvla][INFO] - Step 4775, Loss: {'train/loss': 0.5181589722633362, 'train/grad_norm_pre_clip': 24.26749038696289, 'timing/data': 0.0010809367522597313, 'timing/model': 1.0507398573681712, 'learning_rate/qwen_vl_interface': 2.7359409953916914e-06, 'learning_rate/action_model': 2.7359409953916913e-05, 'epoch': 1.75}) +[2026-06-18 02:52:28,770][starVLA.training.train_starvla][INFO] - Step 4776, Loss: {'train/loss': 0.4741594195365906, 'train/grad_norm_pre_clip': 20.530771255493164, 'timing/data': 0.0009089754894375801, 'timing/model': 1.0579817574471235, 'learning_rate/qwen_vl_interface': 2.734106321256971e-06, 'learning_rate/action_model': 2.7341063212569705e-05, 'epoch': 1.75}) +[2026-06-18 02:52:33,549][starVLA.training.train_starvla][INFO] - Step 4777, Loss: {'train/loss': 0.39618104696273804, 'train/grad_norm_pre_clip': 22.719074249267578, 'timing/data': 0.0010352293029427528, 'timing/model': 1.071871044114232, 'learning_rate/qwen_vl_interface': 2.732272168669718e-06, 'learning_rate/action_model': 2.732272168669718e-05, 'epoch': 1.75}) +[2026-06-18 02:52:38,115][starVLA.training.train_starvla][INFO] - Step 4778, Loss: {'train/loss': 0.5275834798812866, 'train/grad_norm_pre_clip': 27.661388397216797, 'timing/data': 0.0011672209948301315, 'timing/model': 1.3289560787379742, 'learning_rate/qwen_vl_interface': 2.7304385380101512e-06, 'learning_rate/action_model': 2.730438538010151e-05, 'epoch': 1.75}) +[2026-06-18 02:52:43,757][starVLA.training.train_starvla][INFO] - Step 4779, Loss: {'train/loss': 0.2289309799671173, 'train/grad_norm_pre_clip': 14.886760711669922, 'timing/data': 0.0012370329350233078, 'timing/model': 1.1988319670781493, 'learning_rate/qwen_vl_interface': 2.7286054296583904e-06, 'learning_rate/action_model': 2.72860542965839e-05, 'epoch': 1.75}) +[2026-06-18 02:52:52,630][starVLA.training.train_starvla][INFO] - Step 4780, Loss: {'train/loss': 0.5580347776412964, 'train/grad_norm_pre_clip': 31.4627628326416, 'timing/data': 0.0012173093855381012, 'timing/model': 1.3376296646893024, 'learning_rate/qwen_vl_interface': 2.726772843994435e-06, 'learning_rate/action_model': 2.7267728439944352e-05, 'epoch': 1.75}) +[2026-06-18 02:53:00,423][starVLA.training.train_starvla][INFO] - Step 4781, Loss: {'train/loss': 0.29281654953956604, 'train/grad_norm_pre_clip': 15.52614688873291, 'timing/data': 0.0008992888033390045, 'timing/model': 4.0571705140173435, 'learning_rate/qwen_vl_interface': 2.7249407813981834e-06, 'learning_rate/action_model': 2.724940781398183e-05, 'epoch': 1.75}) +[2026-06-18 02:53:06,885][starVLA.training.train_starvla][INFO] - Step 4782, Loss: {'train/loss': 0.41053879261016846, 'train/grad_norm_pre_clip': 20.951332092285156, 'timing/data': 0.0008813254535198212, 'timing/model': 1.4812072310596704, 'learning_rate/qwen_vl_interface': 2.7231092422494255e-06, 'learning_rate/action_model': 2.7231092422494253e-05, 'epoch': 1.75}) +[2026-06-18 02:53:13,522][starVLA.training.train_starvla][INFO] - Step 4783, Loss: {'train/loss': 0.37659206986427307, 'train/grad_norm_pre_clip': 17.015127182006836, 'timing/data': 0.0011279536411166191, 'timing/model': 1.9805925441905856, 'learning_rate/qwen_vl_interface': 2.7212782269278417e-06, 'learning_rate/action_model': 2.7212782269278413e-05, 'epoch': 1.75}) +[2026-06-18 02:53:21,399][starVLA.training.train_starvla][INFO] - Step 4784, Loss: {'train/loss': 0.2549673914909363, 'train/grad_norm_pre_clip': 19.104719161987305, 'timing/data': 0.0009158644825220108, 'timing/model': 1.0914720380678773, 'learning_rate/qwen_vl_interface': 2.719447735813004e-06, 'learning_rate/action_model': 2.719447735813004e-05, 'epoch': 1.75}) +[2026-06-18 02:53:26,735][starVLA.training.train_starvla][INFO] - Step 4785, Loss: {'train/loss': 0.24356697499752045, 'train/grad_norm_pre_clip': 15.79172134399414, 'timing/data': 0.0013236664235591888, 'timing/model': 1.7735864408314228, 'learning_rate/qwen_vl_interface': 2.7176177692843697e-06, 'learning_rate/action_model': 2.7176177692843697e-05, 'epoch': 1.75}) +[2026-06-18 02:53:31,588][starVLA.training.train_starvla][INFO] - Step 4786, Loss: {'train/loss': 0.5147007703781128, 'train/grad_norm_pre_clip': 30.169736862182617, 'timing/data': 0.001371077261865139, 'timing/model': 1.0611256305128336, 'learning_rate/qwen_vl_interface': 2.7157883277212985e-06, 'learning_rate/action_model': 2.7157883277212985e-05, 'epoch': 1.75}) +[2026-06-18 02:53:36,453][starVLA.training.train_starvla][INFO] - Step 4787, Loss: {'train/loss': 0.4273794889450073, 'train/grad_norm_pre_clip': 16.039512634277344, 'timing/data': 0.0013370076194405556, 'timing/model': 1.084867436438799, 'learning_rate/qwen_vl_interface': 2.7139594115030347e-06, 'learning_rate/action_model': 2.7139594115030348e-05, 'epoch': 1.75}) +[2026-06-18 02:53:41,534][starVLA.training.train_starvla][INFO] - Step 4788, Loss: {'train/loss': 0.30818358063697815, 'train/grad_norm_pre_clip': 19.5419979095459, 'timing/data': 0.0012617884203791618, 'timing/model': 1.8874152693897486, 'learning_rate/qwen_vl_interface': 2.712131021008714e-06, 'learning_rate/action_model': 2.7121310210087137e-05, 'epoch': 1.75}) +[2026-06-18 02:53:46,101][starVLA.training.train_starvla][INFO] - Step 4789, Loss: {'train/loss': 0.4567910134792328, 'train/grad_norm_pre_clip': 42.08625793457031, 'timing/data': 0.0008072694763541222, 'timing/model': 1.0680611627176404, 'learning_rate/qwen_vl_interface': 2.7103031566173622e-06, 'learning_rate/action_model': 2.710303156617362e-05, 'epoch': 1.75}) +[2026-06-18 02:53:50,729][starVLA.training.train_starvla][INFO] - Step 4790, Loss: {'train/loss': 0.40514829754829407, 'train/grad_norm_pre_clip': 23.04199981689453, 'timing/data': 0.0011351276189088821, 'timing/model': 1.091165174730122, 'learning_rate/qwen_vl_interface': 2.7084758187078995e-06, 'learning_rate/action_model': 2.7084758187078997e-05, 'epoch': 1.75}) +[2026-06-18 02:53:55,415][starVLA.training.train_starvla][INFO] - Step 4791, Loss: {'train/loss': 0.3559156656265259, 'train/grad_norm_pre_clip': 28.75295639038086, 'timing/data': 0.0009936261922121048, 'timing/model': 1.0662445165216923, 'learning_rate/qwen_vl_interface': 2.7066490076591344e-06, 'learning_rate/action_model': 2.7066490076591343e-05, 'epoch': 1.75}) +[2026-06-18 02:54:09,710][starVLA.training.train_starvla][INFO] - Step 4792, Loss: {'train/loss': 0.09790971875190735, 'train/grad_norm_pre_clip': 13.299742698669434, 'timing/data': 0.0010379552841186523, 'timing/model': 1.2442455757409334, 'learning_rate/qwen_vl_interface': 2.704822723849765e-06, 'learning_rate/action_model': 2.7048227238497648e-05, 'epoch': 1.75}) +[2026-06-18 02:54:14,578][starVLA.training.train_starvla][INFO] - Step 4793, Loss: {'train/loss': 0.5291633605957031, 'train/grad_norm_pre_clip': 27.890626907348633, 'timing/data': 0.0009608389809727669, 'timing/model': 1.3555631907656789, 'learning_rate/qwen_vl_interface': 2.702996967658388e-06, 'learning_rate/action_model': 2.7029969676583876e-05, 'epoch': 1.76}) +[2026-06-18 02:54:19,400][starVLA.training.train_starvla][INFO] - Step 4794, Loss: {'train/loss': 0.1915564239025116, 'train/grad_norm_pre_clip': 13.634689331054688, 'timing/data': 0.0012428415939211845, 'timing/model': 1.0371269211173058, 'learning_rate/qwen_vl_interface': 2.7011717394634767e-06, 'learning_rate/action_model': 2.701171739463477e-05, 'epoch': 1.76}) +[2026-06-18 02:54:23,979][starVLA.training.train_starvla][INFO] - Step 4795, Loss: {'train/loss': 0.3921158015727997, 'train/grad_norm_pre_clip': 22.820446014404297, 'timing/data': 0.0016007432714104652, 'timing/model': 1.0672203609719872, 'learning_rate/qwen_vl_interface': 2.699347039643408e-06, 'learning_rate/action_model': 2.699347039643408e-05, 'epoch': 1.76}) +[2026-06-18 02:54:32,012][starVLA.training.train_starvla][INFO] - Step 4796, Loss: {'train/loss': 0.262922465801239, 'train/grad_norm_pre_clip': 25.449548721313477, 'timing/data': 0.001352354884147644, 'timing/model': 1.115668911486864, 'learning_rate/qwen_vl_interface': 2.6975228685764398e-06, 'learning_rate/action_model': 2.69752286857644e-05, 'epoch': 1.76}) +[2026-06-18 02:54:37,101][starVLA.training.train_starvla][INFO] - Step 4797, Loss: {'train/loss': 0.18332380056381226, 'train/grad_norm_pre_clip': 20.858251571655273, 'timing/data': 0.0017612893134355545, 'timing/model': 1.1848528683185577, 'learning_rate/qwen_vl_interface': 2.6956992266407307e-06, 'learning_rate/action_model': 2.6956992266407306e-05, 'epoch': 1.76}) +[2026-06-18 02:54:41,986][starVLA.training.train_starvla][INFO] - Step 4798, Loss: {'train/loss': 0.4003291130065918, 'train/grad_norm_pre_clip': 24.924253463745117, 'timing/data': 0.0015916964039206505, 'timing/model': 1.517158605158329, 'learning_rate/qwen_vl_interface': 2.693876114214321e-06, 'learning_rate/action_model': 2.6938761142143208e-05, 'epoch': 1.76}) +[2026-06-18 02:54:46,424][starVLA.training.train_starvla][INFO] - Step 4799, Loss: {'train/loss': 0.3801582455635071, 'train/grad_norm_pre_clip': 29.093900680541992, 'timing/data': 0.0011852728202939034, 'timing/model': 1.0931795500218868, 'learning_rate/qwen_vl_interface': 2.6920535316751417e-06, 'learning_rate/action_model': 2.6920535316751416e-05, 'epoch': 1.76}) +[2026-06-18 02:59:45,392][starVLA.training.train_starvla][INFO] - Step 4800, Loss: {'train/loss': 0.49507465958595276, 'train/grad_norm_pre_clip': 26.137065887451172, 'eval/loss': 0.5338183045387268, 'eval/latency_2/loss': 0.5338183045387268, 'eval/demon_fire_precision': 0.9310624493106245, 'eval/demon_fire_recall': 0.9606694560669456, 'eval/demon_fire_f1': 0.9456342668863261, 'eval/demon_fire_rate_model': 0.6861435726210351, 'eval/demon_fire_rate_teacher': 0.6649972175848636, 'eval/demon_left_precision': 0.8916309012875536, 'eval/demon_left_recall': 0.8747368421052631, 'eval/demon_left_f1': 0.8831030818278426, 'eval/demon_left_rate_model': 0.2593210907067334, 'eval/demon_left_rate_teacher': 0.2643294379521425, 'eval/demon_right_precision': 0.9030023094688222, 'eval/demon_right_recall': 0.8836158192090395, 'eval/demon_right_f1': 0.8932038834951456, 'eval/demon_right_rate_model': 0.2409571508069004, 'eval/demon_right_rate_teacher': 0.24624373956594323, 'eval/demon_move_f1': 0.8881534826614941, 'eval/demon_cc_f1': 0.9276236849594676, 'eval/latency_2/demon_fire_precision': 0.9310624493106245, 'eval/latency_2/demon_fire_recall': 0.9606694560669456, 'eval/latency_2/demon_fire_f1': 0.9456342668863261, 'eval/latency_2/demon_fire_rate_model': 0.6861435726210351, 'eval/latency_2/demon_fire_rate_teacher': 0.6649972175848636, 'eval/latency_2/demon_left_precision': 0.8916309012875536, 'eval/latency_2/demon_left_recall': 0.8747368421052631, 'eval/latency_2/demon_left_f1': 0.8831030818278426, 'eval/latency_2/demon_left_rate_model': 0.2593210907067334, 'eval/latency_2/demon_left_rate_teacher': 0.2643294379521425, 'eval/latency_2/demon_right_precision': 0.9030023094688222, 'eval/latency_2/demon_right_recall': 0.8836158192090395, 'eval/latency_2/demon_right_f1': 0.8932038834951456, 'eval/latency_2/demon_right_rate_model': 0.2409571508069004, 'eval/latency_2/demon_right_rate_teacher': 0.24624373956594323, 'eval/latency_2/demon_move_f1': 0.8881534826614941, 'eval/latency_2/demon_cc_f1': 0.9276236849594676, 'timing/data': 0.0008621932938694954, 'timing/model': 1.154846677556634, 'learning_rate/qwen_vl_interface': 2.6902314794010164e-06, 'learning_rate/action_model': 2.6902314794010165e-05, 'epoch': 1.76}) +[2026-06-18 02:59:45,397][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_4800_state +[2026-06-18 02:59:45,405][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 02:59:55,954][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_4800_state/model.safetensors +[2026-06-18 03:00:13,683][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_4800_state/optimizer.bin +[2026-06-18 03:00:13,684][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_4800_state/sampler.bin +[2026-06-18 03:00:13,685][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_4800_state/sampler_1.bin +[2026-06-18 03:00:13,702][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_4800_state/random_states_0.pkl +[2026-06-18 03:00:13,707][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4800 +[2026-06-18 03:00:13,707][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 03:00:13,724][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 03:00:18,817][starVLA.training.train_starvla][INFO] - Step 4801, Loss: {'train/loss': 0.3069067895412445, 'train/grad_norm_pre_clip': 21.734357833862305, 'timing/data': 0.0009576408192515373, 'timing/model': 1.124239096418023, 'learning_rate/qwen_vl_interface': 2.6884099577696615e-06, 'learning_rate/action_model': 2.6884099577696615e-05, 'epoch': 1.76}) +[2026-06-18 03:00:23,382][starVLA.training.train_starvla][INFO] - Step 4802, Loss: {'train/loss': 0.35163363814353943, 'train/grad_norm_pre_clip': 20.052936553955078, 'timing/data': 0.0008952450007200241, 'timing/model': 1.0507153207436204, 'learning_rate/qwen_vl_interface': 2.686588967158679e-06, 'learning_rate/action_model': 2.6865889671586792e-05, 'epoch': 1.76}) +[2026-06-18 03:00:28,104][starVLA.training.train_starvla][INFO] - Step 4803, Loss: {'train/loss': 0.16936777532100677, 'train/grad_norm_pre_clip': 13.605956077575684, 'timing/data': 0.0011849235743284225, 'timing/model': 1.0460543194785714, 'learning_rate/qwen_vl_interface': 2.6847685079455614e-06, 'learning_rate/action_model': 2.6847685079455613e-05, 'epoch': 1.76}) +[2026-06-18 03:00:32,483][starVLA.training.train_starvla][INFO] - Step 4804, Loss: {'train/loss': 0.14908628165721893, 'train/grad_norm_pre_clip': 12.872234344482422, 'timing/data': 0.0008665043860673904, 'timing/model': 1.3373082168400288, 'learning_rate/qwen_vl_interface': 2.6829485805076927e-06, 'learning_rate/action_model': 2.6829485805076925e-05, 'epoch': 1.76}) +[2026-06-18 03:00:36,592][starVLA.training.train_starvla][INFO] - Step 4805, Loss: {'train/loss': 0.2585018575191498, 'train/grad_norm_pre_clip': 14.03455924987793, 'timing/data': 0.0009492440149188042, 'timing/model': 1.0518037509173155, 'learning_rate/qwen_vl_interface': 2.6811291852223452e-06, 'learning_rate/action_model': 2.6811291852223454e-05, 'epoch': 1.76}) +[2026-06-18 03:00:41,059][starVLA.training.train_starvla][INFO] - Step 4806, Loss: {'train/loss': 0.5440567135810852, 'train/grad_norm_pre_clip': 20.20349884033203, 'timing/data': 0.0014888150617480278, 'timing/model': 1.0565927466377616, 'learning_rate/qwen_vl_interface': 2.679310322466683e-06, 'learning_rate/action_model': 2.6793103224666827e-05, 'epoch': 1.76}) +[2026-06-18 03:00:45,522][starVLA.training.train_starvla][INFO] - Step 4807, Loss: {'train/loss': 0.34985965490341187, 'train/grad_norm_pre_clip': 28.522476196289062, 'timing/data': 0.0010755779221653938, 'timing/model': 1.0562451900914311, 'learning_rate/qwen_vl_interface': 2.6774919926177563e-06, 'learning_rate/action_model': 2.677491992617756e-05, 'epoch': 1.76}) +[2026-06-18 03:00:50,370][starVLA.training.train_starvla][INFO] - Step 4808, Loss: {'train/loss': 0.2942071855068207, 'train/grad_norm_pre_clip': 16.905366897583008, 'timing/data': 0.0007769158110022545, 'timing/model': 1.7902157502248883, 'learning_rate/qwen_vl_interface': 2.675674196052507e-06, 'learning_rate/action_model': 2.6756741960525072e-05, 'epoch': 1.76}) +[2026-06-18 03:00:54,778][starVLA.training.train_starvla][INFO] - Step 4809, Loss: {'train/loss': 0.22122535109519958, 'train/grad_norm_pre_clip': 35.87025451660156, 'timing/data': 0.0012014126405119896, 'timing/model': 1.3240730436518788, 'learning_rate/qwen_vl_interface': 2.6738569331477677e-06, 'learning_rate/action_model': 2.6738569331477674e-05, 'epoch': 1.76}) +[2026-06-18 03:00:58,949][starVLA.training.train_starvla][INFO] - Step 4810, Loss: {'train/loss': 0.31269022822380066, 'train/grad_norm_pre_clip': 19.328245162963867, 'timing/data': 0.0011755311861634254, 'timing/model': 1.0923627438023686, 'learning_rate/qwen_vl_interface': 2.672040204280258e-06, 'learning_rate/action_model': 2.672040204280258e-05, 'epoch': 1.76}) +[2026-06-18 03:01:03,570][starVLA.training.train_starvla][INFO] - Step 4811, Loss: {'train/loss': 0.2507775127887726, 'train/grad_norm_pre_clip': 15.077972412109375, 'timing/data': 0.0010721618309617043, 'timing/model': 1.040266778320074, 'learning_rate/qwen_vl_interface': 2.6702240098265864e-06, 'learning_rate/action_model': 2.6702240098265864e-05, 'epoch': 1.76}) +[2026-06-18 03:01:08,616][starVLA.training.train_starvla][INFO] - Step 4812, Loss: {'train/loss': 0.3415350019931793, 'train/grad_norm_pre_clip': 20.66306495666504, 'timing/data': 0.001132521778345108, 'timing/model': 1.427071575075388, 'learning_rate/qwen_vl_interface': 2.6684083501632573e-06, 'learning_rate/action_model': 2.6684083501632574e-05, 'epoch': 1.76}) +[2026-06-18 03:01:13,166][starVLA.training.train_starvla][INFO] - Step 4813, Loss: {'train/loss': 0.21892184019088745, 'train/grad_norm_pre_clip': 20.453641891479492, 'timing/data': 0.0015959888696670532, 'timing/model': 1.0358614083379507, 'learning_rate/qwen_vl_interface': 2.6665932256666525e-06, 'learning_rate/action_model': 2.6665932256666527e-05, 'epoch': 1.76}) +[2026-06-18 03:01:17,815][starVLA.training.train_starvla][INFO] - Step 4814, Loss: {'train/loss': 0.36719146370887756, 'train/grad_norm_pre_clip': 16.999103546142578, 'timing/data': 0.0012220684438943863, 'timing/model': 1.4980196515098214, 'learning_rate/qwen_vl_interface': 2.6647786367130513e-06, 'learning_rate/action_model': 2.664778636713051e-05, 'epoch': 1.76}) +[2026-06-18 03:01:27,499][starVLA.training.train_starvla][INFO] - Step 4815, Loss: {'train/loss': 0.5132030248641968, 'train/grad_norm_pre_clip': 24.726879119873047, 'timing/data': 0.0008997013792395592, 'timing/model': 4.431947811506689, 'learning_rate/qwen_vl_interface': 2.6629645836786217e-06, 'learning_rate/action_model': 2.6629645836786217e-05, 'epoch': 1.76}) +[2026-06-18 03:01:36,157][starVLA.training.train_starvla][INFO] - Step 4816, Loss: {'train/loss': 0.22956088185310364, 'train/grad_norm_pre_clip': 22.449443817138672, 'timing/data': 0.0012711584568023682, 'timing/model': 1.208272511139512, 'learning_rate/qwen_vl_interface': 2.661151066939417e-06, 'learning_rate/action_model': 2.6611510669394168e-05, 'epoch': 1.76}) +[2026-06-18 03:01:53,018][starVLA.training.train_starvla][INFO] - Step 4817, Loss: {'train/loss': 0.11208886653184891, 'train/grad_norm_pre_clip': 10.313739776611328, 'timing/data': 0.0009115226566791534, 'timing/model': 2.544107179157436, 'learning_rate/qwen_vl_interface': 2.6593380868713843e-06, 'learning_rate/action_model': 2.659338086871384e-05, 'epoch': 1.76}) +[2026-06-18 03:01:58,508][starVLA.training.train_starvla][INFO] - Step 4818, Loss: {'train/loss': 0.5746052861213684, 'train/grad_norm_pre_clip': 21.29866600036621, 'timing/data': 0.0011283624917268753, 'timing/model': 1.3214205531403422, 'learning_rate/qwen_vl_interface': 2.6575256438503507e-06, 'learning_rate/action_model': 2.6575256438503504e-05, 'epoch': 1.76}) +[2026-06-18 03:02:08,827][starVLA.training.train_starvla][INFO] - Step 4819, Loss: {'train/loss': 0.34407538175582886, 'train/grad_norm_pre_clip': 21.95751953125, 'timing/data': 0.0010256171226501465, 'timing/model': 4.885430408641696, 'learning_rate/qwen_vl_interface': 2.6557137382520413e-06, 'learning_rate/action_model': 2.6557137382520412e-05, 'epoch': 1.76}) +[2026-06-18 03:02:19,752][starVLA.training.train_starvla][INFO] - Step 4820, Loss: {'train/loss': 0.21263299882411957, 'train/grad_norm_pre_clip': 16.67474365234375, 'timing/data': 0.0009109638631343842, 'timing/model': 1.1287078745663166, 'learning_rate/qwen_vl_interface': 2.6539023704520666e-06, 'learning_rate/action_model': 2.6539023704520664e-05, 'epoch': 1.76}) +[2026-06-18 03:02:31,888][starVLA.training.train_starvla][INFO] - Step 4821, Loss: {'train/loss': 0.24137456715106964, 'train/grad_norm_pre_clip': 17.42521095275879, 'timing/data': 0.0009777368977665901, 'timing/model': 1.242538477294147, 'learning_rate/qwen_vl_interface': 2.6520915408259234e-06, 'learning_rate/action_model': 2.6520915408259235e-05, 'epoch': 1.77}) +[2026-06-18 03:02:42,990][starVLA.training.train_starvla][INFO] - Step 4822, Loss: {'train/loss': 0.5377857089042664, 'train/grad_norm_pre_clip': 41.09660720825195, 'timing/data': 0.0010065501555800438, 'timing/model': 2.376529752276838, 'learning_rate/qwen_vl_interface': 2.650281249749e-06, 'learning_rate/action_model': 2.6502812497489998e-05, 'epoch': 1.77}) +[2026-06-18 03:02:56,210][starVLA.training.train_starvla][INFO] - Step 4823, Loss: {'train/loss': 0.2822539806365967, 'train/grad_norm_pre_clip': 20.896543502807617, 'timing/data': 0.000906720757484436, 'timing/model': 4.329433846287429, 'learning_rate/qwen_vl_interface': 2.6484714975965704e-06, 'learning_rate/action_model': 2.6484714975965703e-05, 'epoch': 1.77}) +[2026-06-18 03:03:04,647][starVLA.training.train_starvla][INFO] - Step 4824, Loss: {'train/loss': 0.36576151847839355, 'train/grad_norm_pre_clip': 16.539714813232422, 'timing/data': 0.0009121419861912727, 'timing/model': 1.5119368210434914, 'learning_rate/qwen_vl_interface': 2.646662284743799e-06, 'learning_rate/action_model': 2.646662284743799e-05, 'epoch': 1.77}) +[2026-06-18 03:03:14,198][starVLA.training.train_starvla][INFO] - Step 4825, Loss: {'train/loss': 0.43958228826522827, 'train/grad_norm_pre_clip': 23.036062240600586, 'timing/data': 0.001299440860748291, 'timing/model': 1.3728370405733585, 'learning_rate/qwen_vl_interface': 2.6448536115657353e-06, 'learning_rate/action_model': 2.6448536115657353e-05, 'epoch': 1.77}) +[2026-06-18 03:03:18,783][starVLA.training.train_starvla][INFO] - Step 4826, Loss: {'train/loss': 0.36302050948143005, 'train/grad_norm_pre_clip': 22.44133186340332, 'timing/data': 0.0017328476533293724, 'timing/model': 1.0598902190104127, 'learning_rate/qwen_vl_interface': 2.6430454784373248e-06, 'learning_rate/action_model': 2.643045478437325e-05, 'epoch': 1.77}) +[2026-06-18 03:03:23,187][starVLA.training.train_starvla][INFO] - Step 4827, Loss: {'train/loss': 0.4675905108451843, 'train/grad_norm_pre_clip': 29.694047927856445, 'timing/data': 0.0008063269779086113, 'timing/model': 1.050282129086554, 'learning_rate/qwen_vl_interface': 2.6412378857333903e-06, 'learning_rate/action_model': 2.6412378857333903e-05, 'epoch': 1.77}) +[2026-06-18 03:03:27,803][starVLA.training.train_starvla][INFO] - Step 4828, Loss: {'train/loss': 0.24990081787109375, 'train/grad_norm_pre_clip': 26.809864044189453, 'timing/data': 0.0011018645018339157, 'timing/model': 1.6032047625631094, 'learning_rate/qwen_vl_interface': 2.639430833828649e-06, 'learning_rate/action_model': 2.6394308338286488e-05, 'epoch': 1.77}) +[2026-06-18 03:03:32,458][starVLA.training.train_starvla][INFO] - Step 4829, Loss: {'train/loss': 0.3636590838432312, 'train/grad_norm_pre_clip': 35.54640197753906, 'timing/data': 0.0014629969373345375, 'timing/model': 1.4148339303210378, 'learning_rate/qwen_vl_interface': 2.637624323097703e-06, 'learning_rate/action_model': 2.6376243230977026e-05, 'epoch': 1.77}) +[2026-06-18 03:03:37,015][starVLA.training.train_starvla][INFO] - Step 4830, Loss: {'train/loss': 0.3403906524181366, 'train/grad_norm_pre_clip': 17.739307403564453, 'timing/data': 0.0010848846286535263, 'timing/model': 1.0555814178660512, 'learning_rate/qwen_vl_interface': 2.635818353915048e-06, 'learning_rate/action_model': 2.635818353915048e-05, 'epoch': 1.77}) +[2026-06-18 03:03:41,796][starVLA.training.train_starvla][INFO] - Step 4831, Loss: {'train/loss': 0.26039326190948486, 'train/grad_norm_pre_clip': 19.24274253845215, 'timing/data': 0.0013365624472498894, 'timing/model': 1.0509085692465305, 'learning_rate/qwen_vl_interface': 2.634012926655063e-06, 'learning_rate/action_model': 2.6340129266550628e-05, 'epoch': 1.77}) +[2026-06-18 03:03:46,195][starVLA.training.train_starvla][INFO] - Step 4832, Loss: {'train/loss': 0.3033633828163147, 'train/grad_norm_pre_clip': 22.029003143310547, 'timing/data': 0.0008903304114937782, 'timing/model': 1.0663620252162218, 'learning_rate/qwen_vl_interface': 2.632208041692009e-06, 'learning_rate/action_model': 2.6322080416920088e-05, 'epoch': 1.77}) +[2026-06-18 03:03:51,107][starVLA.training.train_starvla][INFO] - Step 4833, Loss: {'train/loss': 0.3880203664302826, 'train/grad_norm_pre_clip': 24.374364852905273, 'timing/data': 0.0012097954750061035, 'timing/model': 1.1959279151633382, 'learning_rate/qwen_vl_interface': 2.630403699400047e-06, 'learning_rate/action_model': 2.6304036994000468e-05, 'epoch': 1.77}) +[2026-06-18 03:03:55,625][starVLA.training.train_starvla][INFO] - Step 4834, Loss: {'train/loss': 0.35808247327804565, 'train/grad_norm_pre_clip': 25.931655883789062, 'timing/data': 0.0008861953392624855, 'timing/model': 1.3724529556930065, 'learning_rate/qwen_vl_interface': 2.628599900153217e-06, 'learning_rate/action_model': 2.628599900153217e-05, 'epoch': 1.77}) +[2026-06-18 03:04:03,464][starVLA.training.train_starvla][INFO] - Step 4835, Loss: {'train/loss': 0.19929035007953644, 'train/grad_norm_pre_clip': 11.265127182006836, 'timing/data': 0.0009066220372915268, 'timing/model': 1.0640719160437584, 'learning_rate/qwen_vl_interface': 2.6267966443254484e-06, 'learning_rate/action_model': 2.6267966443254483e-05, 'epoch': 1.77}) +[2026-06-18 03:04:17,436][starVLA.training.train_starvla][INFO] - Step 4836, Loss: {'train/loss': 0.27241331338882446, 'train/grad_norm_pre_clip': 20.997196197509766, 'timing/data': 0.0008849212899804115, 'timing/model': 4.614615930244327, 'learning_rate/qwen_vl_interface': 2.6249939322905576e-06, 'learning_rate/action_model': 2.6249939322905576e-05, 'epoch': 1.77}) +[2026-06-18 03:04:22,289][starVLA.training.train_starvla][INFO] - Step 4837, Loss: {'train/loss': 0.3299606740474701, 'train/grad_norm_pre_clip': 20.0757999420166, 'timing/data': 0.0010244296863675117, 'timing/model': 1.0273380000144243, 'learning_rate/qwen_vl_interface': 2.6231917644222505e-06, 'learning_rate/action_model': 2.6231917644222504e-05, 'epoch': 1.77}) +[2026-06-18 03:04:28,806][starVLA.training.train_starvla][INFO] - Step 4838, Loss: {'train/loss': 0.30562108755111694, 'train/grad_norm_pre_clip': 18.20123863220215, 'timing/data': 0.0008587483316659927, 'timing/model': 2.2579476218670607, 'learning_rate/qwen_vl_interface': 2.621390141094116e-06, 'learning_rate/action_model': 2.621390141094116e-05, 'epoch': 1.77}) +[2026-06-18 03:04:35,297][starVLA.training.train_starvla][INFO] - Step 4839, Loss: {'train/loss': 0.3782936632633209, 'train/grad_norm_pre_clip': 25.686182022094727, 'timing/data': 0.0009287754073739052, 'timing/model': 1.6461546132341027, 'learning_rate/qwen_vl_interface': 2.6195890626796345e-06, 'learning_rate/action_model': 2.6195890626796345e-05, 'epoch': 1.77}) +[2026-06-18 03:04:39,906][starVLA.training.train_starvla][INFO] - Step 4840, Loss: {'train/loss': 0.36558544635772705, 'train/grad_norm_pre_clip': 28.696643829345703, 'timing/data': 0.001367342658340931, 'timing/model': 1.5124297682195902, 'learning_rate/qwen_vl_interface': 2.6177885295521706e-06, 'learning_rate/action_model': 2.6177885295521702e-05, 'epoch': 1.77}) +[2026-06-18 03:04:44,937][starVLA.training.train_starvla][INFO] - Step 4841, Loss: {'train/loss': 0.4239727854728699, 'train/grad_norm_pre_clip': 21.24622917175293, 'timing/data': 0.0011760704219341278, 'timing/model': 1.2601986099034548, 'learning_rate/qwen_vl_interface': 2.6159885420849766e-06, 'learning_rate/action_model': 2.6159885420849766e-05, 'epoch': 1.77}) +[2026-06-18 03:04:50,540][starVLA.training.train_starvla][INFO] - Step 4842, Loss: {'train/loss': 0.3988634943962097, 'train/grad_norm_pre_clip': 19.34198570251465, 'timing/data': 0.0029801269993185997, 'timing/model': 1.3171513453125954, 'learning_rate/qwen_vl_interface': 2.6141891006511927e-06, 'learning_rate/action_model': 2.6141891006511925e-05, 'epoch': 1.77}) +[2026-06-18 03:04:55,924][starVLA.training.train_starvla][INFO] - Step 4843, Loss: {'train/loss': 0.48218584060668945, 'train/grad_norm_pre_clip': 25.225019454956055, 'timing/data': 0.0011502569541335106, 'timing/model': 1.4080224642530084, 'learning_rate/qwen_vl_interface': 2.612390205623842e-06, 'learning_rate/action_model': 2.612390205623842e-05, 'epoch': 1.77}) +[2026-06-18 03:05:01,151][starVLA.training.train_starvla][INFO] - Step 4844, Loss: {'train/loss': 0.45289409160614014, 'train/grad_norm_pre_clip': 19.277408599853516, 'timing/data': 0.0008897632360458374, 'timing/model': 1.5053033446893096, 'learning_rate/qwen_vl_interface': 2.6105918573758425e-06, 'learning_rate/action_model': 2.6105918573758427e-05, 'epoch': 1.77}) +[2026-06-18 03:05:05,548][starVLA.training.train_starvla][INFO] - Step 4845, Loss: {'train/loss': 0.34138360619544983, 'train/grad_norm_pre_clip': 25.31995391845703, 'timing/data': 0.00164851825684309, 'timing/model': 1.0328694768249989, 'learning_rate/qwen_vl_interface': 2.608794056279991e-06, 'learning_rate/action_model': 2.6087940562799906e-05, 'epoch': 1.77}) +[2026-06-18 03:05:10,187][starVLA.training.train_starvla][INFO] - Step 4846, Loss: {'train/loss': 0.4771180748939514, 'train/grad_norm_pre_clip': 21.975078582763672, 'timing/data': 0.0007796017453074455, 'timing/model': 1.059202360920608, 'learning_rate/qwen_vl_interface': 2.6069968027089744e-06, 'learning_rate/action_model': 2.6069968027089742e-05, 'epoch': 1.77}) +[2026-06-18 03:05:16,658][starVLA.training.train_starvla][INFO] - Step 4847, Loss: {'train/loss': 0.21001054346561432, 'train/grad_norm_pre_clip': 14.471445083618164, 'timing/data': 0.0031113941222429276, 'timing/model': 1.1224066242575645, 'learning_rate/qwen_vl_interface': 2.605200097035362e-06, 'learning_rate/action_model': 2.605200097035362e-05, 'epoch': 1.77}) +[2026-06-18 03:05:21,333][starVLA.training.train_starvla][INFO] - Step 4848, Loss: {'train/loss': 0.47336360812187195, 'train/grad_norm_pre_clip': 27.677717208862305, 'timing/data': 0.0015852591022849083, 'timing/model': 1.6355784041807055, 'learning_rate/qwen_vl_interface': 2.6034039396316167e-06, 'learning_rate/action_model': 2.6034039396316167e-05, 'epoch': 1.78}) +[2026-06-18 03:05:26,704][starVLA.training.train_starvla][INFO] - Step 4849, Loss: {'train/loss': 0.23013921082019806, 'train/grad_norm_pre_clip': 16.05907440185547, 'timing/data': 0.0008801082149147987, 'timing/model': 1.557637450285256, 'learning_rate/qwen_vl_interface': 2.601608330870083e-06, 'learning_rate/action_model': 2.601608330870083e-05, 'epoch': 1.78}) +[2026-06-18 03:05:31,384][starVLA.training.train_starvla][INFO] - Step 4850, Loss: {'train/loss': 0.9859105944633484, 'train/grad_norm_pre_clip': 29.413923263549805, 'timing/data': 0.0020991601049900055, 'timing/model': 1.0498950947076082, 'learning_rate/qwen_vl_interface': 2.5998132711229918e-06, 'learning_rate/action_model': 2.5998132711229915e-05, 'epoch': 1.78}) +[2026-06-18 03:05:35,985][starVLA.training.train_starvla][INFO] - Step 4851, Loss: {'train/loss': 0.07428129017353058, 'train/grad_norm_pre_clip': 7.934892654418945, 'timing/data': 0.0007532723248004913, 'timing/model': 1.064701703377068, 'learning_rate/qwen_vl_interface': 2.598018760762461e-06, 'learning_rate/action_model': 2.598018760762461e-05, 'epoch': 1.78}) +[2026-06-18 03:05:40,953][starVLA.training.train_starvla][INFO] - Step 4852, Loss: {'train/loss': 0.23923787474632263, 'train/grad_norm_pre_clip': 17.294464111328125, 'timing/data': 0.014641948975622654, 'timing/model': 1.1652098335325718, 'learning_rate/qwen_vl_interface': 2.5962248001604958e-06, 'learning_rate/action_model': 2.5962248001604956e-05, 'epoch': 1.78}) +[2026-06-18 03:05:50,753][starVLA.training.train_starvla][INFO] - Step 4853, Loss: {'train/loss': 0.48317694664001465, 'train/grad_norm_pre_clip': 30.140474319458008, 'timing/data': 0.0021174494177103043, 'timing/model': 1.0450464924797416, 'learning_rate/qwen_vl_interface': 2.5944313896889837e-06, 'learning_rate/action_model': 2.594431389688984e-05, 'epoch': 1.78}) +[2026-06-18 03:05:56,368][starVLA.training.train_starvla][INFO] - Step 4854, Loss: {'train/loss': 0.3387078642845154, 'train/grad_norm_pre_clip': 21.1439266204834, 'timing/data': 0.0012051593512296677, 'timing/model': 1.4052980886772275, 'learning_rate/qwen_vl_interface': 2.592638529719703e-06, 'learning_rate/action_model': 2.5926385297197033e-05, 'epoch': 1.78}) +[2026-06-18 03:06:01,314][starVLA.training.train_starvla][INFO] - Step 4855, Loss: {'train/loss': 0.3269011676311493, 'train/grad_norm_pre_clip': 19.805749893188477, 'timing/data': 0.0021295538172125816, 'timing/model': 1.0835879715159535, 'learning_rate/qwen_vl_interface': 2.590846220624314e-06, 'learning_rate/action_model': 2.5908462206243138e-05, 'epoch': 1.78}) +[2026-06-18 03:06:06,410][starVLA.training.train_starvla][INFO] - Step 4856, Loss: {'train/loss': 0.25295290350914, 'train/grad_norm_pre_clip': 17.1212215423584, 'timing/data': 0.0009293351322412491, 'timing/model': 1.318299949169159, 'learning_rate/qwen_vl_interface': 2.5890544627743652e-06, 'learning_rate/action_model': 2.589054462774365e-05, 'epoch': 1.78}) +[2026-06-18 03:06:12,088][starVLA.training.train_starvla][INFO] - Step 4857, Loss: {'train/loss': 0.31905290484428406, 'train/grad_norm_pre_clip': 19.293292999267578, 'timing/data': 0.0009980816394090652, 'timing/model': 1.33355956338346, 'learning_rate/qwen_vl_interface': 2.58726325654129e-06, 'learning_rate/action_model': 2.5872632565412898e-05, 'epoch': 1.78}) +[2026-06-18 03:06:16,948][starVLA.training.train_starvla][INFO] - Step 4858, Loss: {'train/loss': 0.27830106019973755, 'train/grad_norm_pre_clip': 22.9237060546875, 'timing/data': 0.0008981311693787575, 'timing/model': 1.0858841463923454, 'learning_rate/qwen_vl_interface': 2.5854726022964043e-06, 'learning_rate/action_model': 2.5854726022964042e-05, 'epoch': 1.78}) +[2026-06-18 03:06:23,710][starVLA.training.train_starvla][INFO] - Step 4859, Loss: {'train/loss': 0.49477407336235046, 'train/grad_norm_pre_clip': 29.888195037841797, 'timing/data': 0.0007909713312983513, 'timing/model': 3.355444110929966, 'learning_rate/qwen_vl_interface': 2.5836825004109173e-06, 'learning_rate/action_model': 2.5836825004109172e-05, 'epoch': 1.78}) +[2026-06-18 03:06:40,484][starVLA.training.train_starvla][INFO] - Step 4860, Loss: {'train/loss': 0.23146115243434906, 'train/grad_norm_pre_clip': 28.076148986816406, 'timing/data': 0.0012451065704226494, 'timing/model': 4.62221356574446, 'learning_rate/qwen_vl_interface': 2.5818929512559193e-06, 'learning_rate/action_model': 2.5818929512559193e-05, 'epoch': 1.78}) +[2026-06-18 03:06:45,730][starVLA.training.train_starvla][INFO] - Step 4861, Loss: {'train/loss': 0.7296620607376099, 'train/grad_norm_pre_clip': 28.346099853515625, 'timing/data': 0.0018170811235904694, 'timing/model': 1.5428547877818346, 'learning_rate/qwen_vl_interface': 2.580103955202381e-06, 'learning_rate/action_model': 2.580103955202381e-05, 'epoch': 1.78}) +[2026-06-18 03:06:50,576][starVLA.training.train_starvla][INFO] - Step 4862, Loss: {'train/loss': 0.13889794051647186, 'train/grad_norm_pre_clip': 14.264280319213867, 'timing/data': 0.0011645080521702766, 'timing/model': 1.1837162487208843, 'learning_rate/qwen_vl_interface': 2.5783155126211647e-06, 'learning_rate/action_model': 2.5783155126211644e-05, 'epoch': 1.78}) +[2026-06-18 03:06:57,040][starVLA.training.train_starvla][INFO] - Step 4863, Loss: {'train/loss': 0.436701238155365, 'train/grad_norm_pre_clip': 28.48859214782715, 'timing/data': 0.00133524090051651, 'timing/model': 1.1725174374878407, 'learning_rate/qwen_vl_interface': 2.576527623883019e-06, 'learning_rate/action_model': 2.576527623883019e-05, 'epoch': 1.78}) +[2026-06-18 03:07:02,071][starVLA.training.train_starvla][INFO] - Step 4864, Loss: {'train/loss': 0.17692719399929047, 'train/grad_norm_pre_clip': 19.243072509765625, 'timing/data': 0.0011336449533700943, 'timing/model': 1.5301025081425905, 'learning_rate/qwen_vl_interface': 2.5747402893585732e-06, 'learning_rate/action_model': 2.574740289358573e-05, 'epoch': 1.78}) +[2026-06-18 03:07:11,055][starVLA.training.train_starvla][INFO] - Step 4865, Loss: {'train/loss': 0.2500510513782501, 'train/grad_norm_pre_clip': 17.319116592407227, 'timing/data': 0.0012409072369337082, 'timing/model': 5.227022849023342, 'learning_rate/qwen_vl_interface': 2.5729535094183445e-06, 'learning_rate/action_model': 2.5729535094183444e-05, 'epoch': 1.78}) +[2026-06-18 03:07:15,776][starVLA.training.train_starvla][INFO] - Step 4866, Loss: {'train/loss': 0.3011694848537445, 'train/grad_norm_pre_clip': 27.307165145874023, 'timing/data': 0.001591934822499752, 'timing/model': 1.0578554132953286, 'learning_rate/qwen_vl_interface': 2.571167284432734e-06, 'learning_rate/action_model': 2.5711672844327338e-05, 'epoch': 1.78}) +[2026-06-18 03:07:20,407][starVLA.training.train_starvla][INFO] - Step 4867, Loss: {'train/loss': 0.5069040060043335, 'train/grad_norm_pre_clip': 35.209373474121094, 'timing/data': 0.0010282127186655998, 'timing/model': 1.0791505733504891, 'learning_rate/qwen_vl_interface': 2.569381614772028e-06, 'learning_rate/action_model': 2.5693816147720278e-05, 'epoch': 1.78}) +[2026-06-18 03:07:25,008][starVLA.training.train_starvla][INFO] - Step 4868, Loss: {'train/loss': 0.43748757243156433, 'train/grad_norm_pre_clip': 24.627527236938477, 'timing/data': 0.001197698526084423, 'timing/model': 1.450522038154304, 'learning_rate/qwen_vl_interface': 2.5675965008063976e-06, 'learning_rate/action_model': 2.5675965008063974e-05, 'epoch': 1.78}) +[2026-06-18 03:07:30,196][starVLA.training.train_starvla][INFO] - Step 4869, Loss: {'train/loss': 0.5065386295318604, 'train/grad_norm_pre_clip': 27.563796997070312, 'timing/data': 0.000790339894592762, 'timing/model': 1.3931796085089445, 'learning_rate/qwen_vl_interface': 2.565811942905899e-06, 'learning_rate/action_model': 2.5658119429058993e-05, 'epoch': 1.78}) +[2026-06-18 03:07:34,595][starVLA.training.train_starvla][INFO] - Step 4870, Loss: {'train/loss': 0.36831963062286377, 'train/grad_norm_pre_clip': 17.331829071044922, 'timing/data': 0.0010133590549230576, 'timing/model': 1.1476283762603998, 'learning_rate/qwen_vl_interface': 2.5640279414404723e-06, 'learning_rate/action_model': 2.5640279414404722e-05, 'epoch': 1.78}) +[2026-06-18 03:07:42,568][starVLA.training.train_starvla][INFO] - Step 4871, Loss: {'train/loss': 0.3725546598434448, 'train/grad_norm_pre_clip': 26.944969177246094, 'timing/data': 0.0013848347589373589, 'timing/model': 1.1869986653327942, 'learning_rate/qwen_vl_interface': 2.5622444967799444e-06, 'learning_rate/action_model': 2.5622444967799443e-05, 'epoch': 1.78}) +[2026-06-18 03:07:49,242][starVLA.training.train_starvla][INFO] - Step 4872, Loss: {'train/loss': 0.3137841820716858, 'train/grad_norm_pre_clip': 23.246171951293945, 'timing/data': 0.0008600624278187752, 'timing/model': 3.221224936656654, 'learning_rate/qwen_vl_interface': 2.5604616092940245e-06, 'learning_rate/action_model': 2.5604616092940242e-05, 'epoch': 1.78}) +[2026-06-18 03:07:54,581][starVLA.training.train_starvla][INFO] - Step 4873, Loss: {'train/loss': 0.39683735370635986, 'train/grad_norm_pre_clip': 24.63883399963379, 'timing/data': 0.0013104956597089767, 'timing/model': 1.166157496161759, 'learning_rate/qwen_vl_interface': 2.5586792793523052e-06, 'learning_rate/action_model': 2.5586792793523052e-05, 'epoch': 1.78}) +[2026-06-18 03:08:04,056][starVLA.training.train_starvla][INFO] - Step 4874, Loss: {'train/loss': 0.40782371163368225, 'train/grad_norm_pre_clip': 22.39048194885254, 'timing/data': 0.0011670459061861038, 'timing/model': 6.228779464960098, 'learning_rate/qwen_vl_interface': 2.556897507324271e-06, 'learning_rate/action_model': 2.5568975073242705e-05, 'epoch': 1.78}) +[2026-06-18 03:08:09,725][starVLA.training.train_starvla][INFO] - Step 4875, Loss: {'train/loss': 0.2813858985900879, 'train/grad_norm_pre_clip': 22.168182373046875, 'timing/data': 0.0011723153293132782, 'timing/model': 1.344096307642758, 'learning_rate/qwen_vl_interface': 2.5551162935792782e-06, 'learning_rate/action_model': 2.5551162935792784e-05, 'epoch': 1.79}) +[2026-06-18 03:08:14,554][starVLA.training.train_starvla][INFO] - Step 4876, Loss: {'train/loss': 0.39856016635894775, 'train/grad_norm_pre_clip': 26.670772552490234, 'timing/data': 0.0010443702340126038, 'timing/model': 1.1004329761490226, 'learning_rate/qwen_vl_interface': 2.5533356384865777e-06, 'learning_rate/action_model': 2.5533356384865776e-05, 'epoch': 1.79}) +[2026-06-18 03:08:19,706][starVLA.training.train_starvla][INFO] - Step 4877, Loss: {'train/loss': 0.24613262712955475, 'train/grad_norm_pre_clip': 34.72283935546875, 'timing/data': 0.00113687664270401, 'timing/model': 1.2200455600395799, 'learning_rate/qwen_vl_interface': 2.5515555424153017e-06, 'learning_rate/action_model': 2.5515555424153016e-05, 'epoch': 1.79}) +[2026-06-18 03:08:24,608][starVLA.training.train_starvla][INFO] - Step 4878, Loss: {'train/loss': 0.5027800798416138, 'train/grad_norm_pre_clip': 26.82715606689453, 'timing/data': 0.0007797479629516602, 'timing/model': 1.0957559179514647, 'learning_rate/qwen_vl_interface': 2.5497760057344663e-06, 'learning_rate/action_model': 2.549776005734466e-05, 'epoch': 1.79}) +[2026-06-18 03:08:36,712][starVLA.training.train_starvla][INFO] - Step 4879, Loss: {'train/loss': 0.3776397705078125, 'train/grad_norm_pre_clip': 25.25727081298828, 'timing/data': 0.0010730726644396782, 'timing/model': 5.377360260114074, 'learning_rate/qwen_vl_interface': 2.5479970288129714e-06, 'learning_rate/action_model': 2.5479970288129714e-05, 'epoch': 1.79}) +[2026-06-18 03:08:45,930][starVLA.training.train_starvla][INFO] - Step 4880, Loss: {'train/loss': 0.239137202501297, 'train/grad_norm_pre_clip': 33.302528381347656, 'timing/data': 0.0018252693116664886, 'timing/model': 1.4214595519006252, 'learning_rate/qwen_vl_interface': 2.546218612019597e-06, 'learning_rate/action_model': 2.5462186120195967e-05, 'epoch': 1.79}) +[2026-06-18 03:08:51,120][starVLA.training.train_starvla][INFO] - Step 4881, Loss: {'train/loss': 0.36679860949516296, 'train/grad_norm_pre_clip': 23.025257110595703, 'timing/data': 0.0013782521709799767, 'timing/model': 1.5602895487099886, 'learning_rate/qwen_vl_interface': 2.544440755723015e-06, 'learning_rate/action_model': 2.5444407557230147e-05, 'epoch': 1.79}) +[2026-06-18 03:08:55,876][starVLA.training.train_starvla][INFO] - Step 4882, Loss: {'train/loss': 0.495395690202713, 'train/grad_norm_pre_clip': 22.946931838989258, 'timing/data': 0.0009591756388545036, 'timing/model': 1.2236223993822932, 'learning_rate/qwen_vl_interface': 2.5426634602917756e-06, 'learning_rate/action_model': 2.5426634602917755e-05, 'epoch': 1.79}) +[2026-06-18 03:09:04,869][starVLA.training.train_starvla][INFO] - Step 4883, Loss: {'train/loss': 0.31558266282081604, 'train/grad_norm_pre_clip': 20.2298641204834, 'timing/data': 0.0011093243956565857, 'timing/model': 1.1236581355333328, 'learning_rate/qwen_vl_interface': 2.540886726094315e-06, 'learning_rate/action_model': 2.540886726094315e-05, 'epoch': 1.79}) +[2026-06-18 03:09:09,507][starVLA.training.train_starvla][INFO] - Step 4884, Loss: {'train/loss': 0.29742246866226196, 'train/grad_norm_pre_clip': 21.252246856689453, 'timing/data': 0.0012145591899752617, 'timing/model': 1.4600960314273834, 'learning_rate/qwen_vl_interface': 2.5391105534989503e-06, 'learning_rate/action_model': 2.53911055349895e-05, 'epoch': 1.79}) +[2026-06-18 03:09:15,532][starVLA.training.train_starvla][INFO] - Step 4885, Loss: {'train/loss': 0.28789570927619934, 'train/grad_norm_pre_clip': 15.862663269042969, 'timing/data': 0.0010055778548121452, 'timing/model': 2.705528151243925, 'learning_rate/qwen_vl_interface': 2.537334942873885e-06, 'learning_rate/action_model': 2.5373349428738846e-05, 'epoch': 1.79}) +[2026-06-18 03:09:26,829][starVLA.training.train_starvla][INFO] - Step 4886, Loss: {'train/loss': 0.37754547595977783, 'train/grad_norm_pre_clip': 17.137880325317383, 'timing/data': 0.002693583257496357, 'timing/model': 1.1682555358856916, 'learning_rate/qwen_vl_interface': 2.535559894587205e-06, 'learning_rate/action_model': 2.535559894587205e-05, 'epoch': 1.79}) +[2026-06-18 03:09:31,825][starVLA.training.train_starvla][INFO] - Step 4887, Loss: {'train/loss': 0.325068861246109, 'train/grad_norm_pre_clip': 20.848224639892578, 'timing/data': 0.0012444434687495232, 'timing/model': 1.1660103779286146, 'learning_rate/qwen_vl_interface': 2.533785409006878e-06, 'learning_rate/action_model': 2.533785409006878e-05, 'epoch': 1.79}) +[2026-06-18 03:09:40,511][starVLA.training.train_starvla][INFO] - Step 4888, Loss: {'train/loss': 0.3134573996067047, 'train/grad_norm_pre_clip': 18.866657257080078, 'timing/data': 0.001259208656847477, 'timing/model': 1.5987859396263957, 'learning_rate/qwen_vl_interface': 2.532011486500762e-06, 'learning_rate/action_model': 2.532011486500762e-05, 'epoch': 1.79}) +[2026-06-18 03:09:45,235][starVLA.training.train_starvla][INFO] - Step 4889, Loss: {'train/loss': 0.17424209415912628, 'train/grad_norm_pre_clip': 13.954920768737793, 'timing/data': 0.0009185085073113441, 'timing/model': 1.394311798736453, 'learning_rate/qwen_vl_interface': 2.530238127436588e-06, 'learning_rate/action_model': 2.5302381274365882e-05, 'epoch': 1.79}) +[2026-06-18 03:09:50,106][starVLA.training.train_starvla][INFO] - Step 4890, Loss: {'train/loss': 0.23523075878620148, 'train/grad_norm_pre_clip': 15.9573974609375, 'timing/data': 0.001324150711297989, 'timing/model': 1.56620058324188, 'learning_rate/qwen_vl_interface': 2.528465332181977e-06, 'learning_rate/action_model': 2.5284653321819767e-05, 'epoch': 1.79}) +[2026-06-18 03:09:54,685][starVLA.training.train_starvla][INFO] - Step 4891, Loss: {'train/loss': 0.301310658454895, 'train/grad_norm_pre_clip': 20.891422271728516, 'timing/data': 0.0013652555644512177, 'timing/model': 1.033067686483264, 'learning_rate/qwen_vl_interface': 2.526693101104429e-06, 'learning_rate/action_model': 2.526693101104429e-05, 'epoch': 1.79}) +[2026-06-18 03:09:59,116][starVLA.training.train_starvla][INFO] - Step 4892, Loss: {'train/loss': 0.45221373438835144, 'train/grad_norm_pre_clip': 28.572582244873047, 'timing/data': 0.0010879263281822205, 'timing/model': 1.062886162661016, 'learning_rate/qwen_vl_interface': 2.524921434571333e-06, 'learning_rate/action_model': 2.524921434571333e-05, 'epoch': 1.79}) +[2026-06-18 03:10:06,952][starVLA.training.train_starvla][INFO] - Step 4893, Loss: {'train/loss': 0.37442547082901, 'train/grad_norm_pre_clip': 20.731094360351562, 'timing/data': 0.0009629884734749794, 'timing/model': 1.0905832070857286, 'learning_rate/qwen_vl_interface': 2.5231503329499585e-06, 'learning_rate/action_model': 2.5231503329499583e-05, 'epoch': 1.79}) +[2026-06-18 03:10:11,968][starVLA.training.train_starvla][INFO] - Step 4894, Loss: {'train/loss': 0.6622010469436646, 'train/grad_norm_pre_clip': 29.94866180419922, 'timing/data': 0.0016168635338544846, 'timing/model': 1.4235139098018408, 'learning_rate/qwen_vl_interface': 2.5213797966074506e-06, 'learning_rate/action_model': 2.5213797966074505e-05, 'epoch': 1.79}) +[2026-06-18 03:10:20,905][starVLA.training.train_starvla][INFO] - Step 4895, Loss: {'train/loss': 0.26726996898651123, 'train/grad_norm_pre_clip': 17.556962966918945, 'timing/data': 0.0009733987972140312, 'timing/model': 1.62400190718472, 'learning_rate/qwen_vl_interface': 2.519609825910848e-06, 'learning_rate/action_model': 2.5196098259108482e-05, 'epoch': 1.79}) +[2026-06-18 03:10:30,758][starVLA.training.train_starvla][INFO] - Step 4896, Loss: {'train/loss': 0.2936413884162903, 'train/grad_norm_pre_clip': 17.138486862182617, 'timing/data': 0.0010236073285341263, 'timing/model': 4.582500860095024, 'learning_rate/qwen_vl_interface': 2.517840421227067e-06, 'learning_rate/action_model': 2.517840421227067e-05, 'epoch': 1.79}) +[2026-06-18 03:10:35,783][starVLA.training.train_starvla][INFO] - Step 4897, Loss: {'train/loss': 0.30603110790252686, 'train/grad_norm_pre_clip': 15.381790161132812, 'timing/data': 0.0012530097737908363, 'timing/model': 1.0272427508607507, 'learning_rate/qwen_vl_interface': 2.5160715829229067e-06, 'learning_rate/action_model': 2.5160715829229066e-05, 'epoch': 1.79}) +[2026-06-18 03:10:42,392][starVLA.training.train_starvla][INFO] - Step 4898, Loss: {'train/loss': 0.3020200729370117, 'train/grad_norm_pre_clip': 16.128528594970703, 'timing/data': 0.0010661305859684944, 'timing/model': 1.042928227223456, 'learning_rate/qwen_vl_interface': 2.5143033113650484e-06, 'learning_rate/action_model': 2.5143033113650483e-05, 'epoch': 1.79}) +[2026-06-18 03:10:46,802][starVLA.training.train_starvla][INFO] - Step 4899, Loss: {'train/loss': 0.3763377070426941, 'train/grad_norm_pre_clip': 24.402450561523438, 'timing/data': 0.0011097360402345657, 'timing/model': 1.3446524385362864, 'learning_rate/qwen_vl_interface': 2.5125356069200572e-06, 'learning_rate/action_model': 2.5125356069200574e-05, 'epoch': 1.79}) +[2026-06-18 03:15:14,985][starVLA.training.train_starvla][INFO] - Step 4900, Loss: {'train/loss': 0.3260304629802704, 'train/grad_norm_pre_clip': 18.43478775024414, 'eval/loss': 0.5309745073318481, 'eval/latency_2/loss': 0.5309745073318481, 'eval/demon_fire_precision': 0.93480032599837, 'eval/demon_fire_recall': 0.9598326359832636, 'eval/demon_fire_f1': 0.9471511147811728, 'eval/demon_fire_rate_model': 0.6828046744574291, 'eval/demon_fire_rate_teacher': 0.6649972175848636, 'eval/demon_left_precision': 0.8861024033437827, 'eval/demon_left_recall': 0.8926315789473684, 'eval/demon_left_f1': 0.8893550078657577, 'eval/demon_left_rate_model': 0.2662771285475793, 'eval/demon_left_rate_teacher': 0.2643294379521425, 'eval/demon_right_precision': 0.9320987654320988, 'eval/demon_right_recall': 0.8531073446327684, 'eval/demon_right_f1': 0.8908554572271388, 'eval/demon_right_rate_model': 0.22537562604340566, 'eval/demon_right_rate_teacher': 0.24624373956594323, 'eval/demon_move_f1': 0.8901052325464482, 'eval/demon_cc_f1': 0.9292840712829492, 'eval/latency_2/demon_fire_precision': 0.93480032599837, 'eval/latency_2/demon_fire_recall': 0.9598326359832636, 'eval/latency_2/demon_fire_f1': 0.9471511147811728, 'eval/latency_2/demon_fire_rate_model': 0.6828046744574291, 'eval/latency_2/demon_fire_rate_teacher': 0.6649972175848636, 'eval/latency_2/demon_left_precision': 0.8861024033437827, 'eval/latency_2/demon_left_recall': 0.8926315789473684, 'eval/latency_2/demon_left_f1': 0.8893550078657577, 'eval/latency_2/demon_left_rate_model': 0.2662771285475793, 'eval/latency_2/demon_left_rate_teacher': 0.2643294379521425, 'eval/latency_2/demon_right_precision': 0.9320987654320988, 'eval/latency_2/demon_right_recall': 0.8531073446327684, 'eval/latency_2/demon_right_f1': 0.8908554572271388, 'eval/latency_2/demon_right_rate_model': 0.22537562604340566, 'eval/latency_2/demon_right_rate_teacher': 0.24624373956594323, 'eval/latency_2/demon_move_f1': 0.8901052325464482, 'eval/latency_2/demon_cc_f1': 0.9292840712829492, 'timing/data': 0.0007548248395323753, 'timing/model': 1.2747232327237725, 'learning_rate/qwen_vl_interface': 2.5107684699543817e-06, 'learning_rate/action_model': 2.5107684699543816e-05, 'epoch': 1.79}) +[2026-06-18 03:15:14,990][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_4900_state +[2026-06-18 03:15:15,007][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 03:15:24,350][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_4900_state/model.safetensors +[2026-06-18 03:15:41,855][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_4900_state/optimizer.bin +[2026-06-18 03:15:41,856][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_4900_state/sampler.bin +[2026-06-18 03:15:41,857][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_4900_state/sampler_1.bin +[2026-06-18 03:15:41,859][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_4900_state/random_states_0.pkl +[2026-06-18 03:15:41,881][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4900 +[2026-06-18 03:15:41,882][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 03:15:41,898][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 03:15:46,510][starVLA.training.train_starvla][INFO] - Step 4901, Loss: {'train/loss': 0.45169901847839355, 'train/grad_norm_pre_clip': 29.80344581604004, 'timing/data': 0.0011477693915367126, 'timing/model': 1.5458844164386392, 'learning_rate/qwen_vl_interface': 2.509001900834349e-06, 'learning_rate/action_model': 2.5090019008343492e-05, 'epoch': 1.79}) +[2026-06-18 03:15:56,121][starVLA.training.train_starvla][INFO] - Step 4902, Loss: {'train/loss': 0.5073952674865723, 'train/grad_norm_pre_clip': 23.957712173461914, 'timing/data': 0.0012844186276197433, 'timing/model': 1.2453289488330483, 'learning_rate/qwen_vl_interface': 2.507235899926172e-06, 'learning_rate/action_model': 2.5072358999261718e-05, 'epoch': 1.79}) +[2026-06-18 03:16:04,090][starVLA.training.train_starvla][INFO] - Step 4903, Loss: {'train/loss': 0.2676314413547516, 'train/grad_norm_pre_clip': 19.373666763305664, 'timing/data': 0.0010644514113664627, 'timing/model': 4.44950800947845, 'learning_rate/qwen_vl_interface': 2.5054704675959445e-06, 'learning_rate/action_model': 2.5054704675959445e-05, 'epoch': 1.8}) +[2026-06-18 03:16:11,035][starVLA.training.train_starvla][INFO] - Step 4904, Loss: {'train/loss': 0.324892520904541, 'train/grad_norm_pre_clip': 17.664142608642578, 'timing/data': 0.0017914865165948868, 'timing/model': 3.217227980494499, 'learning_rate/qwen_vl_interface': 2.5037056042096418e-06, 'learning_rate/action_model': 2.5037056042096417e-05, 'epoch': 1.8}) +[2026-06-18 03:16:21,552][starVLA.training.train_starvla][INFO] - Step 4905, Loss: {'train/loss': 0.3649452328681946, 'train/grad_norm_pre_clip': 20.136022567749023, 'timing/data': 0.0011881375685334206, 'timing/model': 1.7161902459338307, 'learning_rate/qwen_vl_interface': 2.5019413101331212e-06, 'learning_rate/action_model': 2.5019413101331214e-05, 'epoch': 1.8}) +[2026-06-18 03:16:30,562][starVLA.training.train_starvla][INFO] - Step 4906, Loss: {'train/loss': 0.3430139124393463, 'train/grad_norm_pre_clip': 22.8560733795166, 'timing/data': 0.001058422029018402, 'timing/model': 3.8440084103494883, 'learning_rate/qwen_vl_interface': 2.5001775857321254e-06, 'learning_rate/action_model': 2.5001775857321253e-05, 'epoch': 1.8}) +[2026-06-18 03:16:43,896][starVLA.training.train_starvla][INFO] - Step 4907, Loss: {'train/loss': 0.639629602432251, 'train/grad_norm_pre_clip': 34.342491149902344, 'timing/data': 0.0013004802167415619, 'timing/model': 5.978462650440633, 'learning_rate/qwen_vl_interface': 2.4984144313722765e-06, 'learning_rate/action_model': 2.4984144313722766e-05, 'epoch': 1.8}) +[2026-06-18 03:16:54,561][starVLA.training.train_starvla][INFO] - Step 4908, Loss: {'train/loss': 0.62839674949646, 'train/grad_norm_pre_clip': 28.2620906829834, 'timing/data': 0.003572518937289715, 'timing/model': 2.2222954789176583, 'learning_rate/qwen_vl_interface': 2.496651847419076e-06, 'learning_rate/action_model': 2.496651847419076e-05, 'epoch': 1.8}) +[2026-06-18 03:17:04,623][starVLA.training.train_starvla][INFO] - Step 4909, Loss: {'train/loss': 0.6436957120895386, 'train/grad_norm_pre_clip': 24.831336975097656, 'timing/data': 0.0010214634239673615, 'timing/model': 2.890379642136395, 'learning_rate/qwen_vl_interface': 2.4948898342379076e-06, 'learning_rate/action_model': 2.4948898342379075e-05, 'epoch': 1.8}) +[2026-06-18 03:17:15,036][starVLA.training.train_starvla][INFO] - Step 4910, Loss: {'train/loss': 0.5734534859657288, 'train/grad_norm_pre_clip': 22.757919311523438, 'timing/data': 0.0018432112410664558, 'timing/model': 3.456956125795841, 'learning_rate/qwen_vl_interface': 2.4931283921940434e-06, 'learning_rate/action_model': 2.4931283921940434e-05, 'epoch': 1.8}) +[2026-06-18 03:17:19,728][starVLA.training.train_starvla][INFO] - Step 4911, Loss: {'train/loss': 0.38289138674736023, 'train/grad_norm_pre_clip': 22.47182846069336, 'timing/data': 0.0012283800169825554, 'timing/model': 1.0685620354488492, 'learning_rate/qwen_vl_interface': 2.4913675216526302e-06, 'learning_rate/action_model': 2.4913675216526304e-05, 'epoch': 1.8}) +[2026-06-18 03:17:28,891][starVLA.training.train_starvla][INFO] - Step 4912, Loss: {'train/loss': 0.10446613281965256, 'train/grad_norm_pre_clip': 10.191067695617676, 'timing/data': 0.0011159246787428856, 'timing/model': 1.1404984630644321, 'learning_rate/qwen_vl_interface': 2.4896072229786996e-06, 'learning_rate/action_model': 2.4896072229786995e-05, 'epoch': 1.8}) +[2026-06-18 03:17:34,164][starVLA.training.train_starvla][INFO] - Step 4913, Loss: {'train/loss': 0.2538781464099884, 'train/grad_norm_pre_clip': 17.004180908203125, 'timing/data': 0.013887738808989525, 'timing/model': 1.2766027925536036, 'learning_rate/qwen_vl_interface': 2.4878474965371587e-06, 'learning_rate/action_model': 2.4878474965371586e-05, 'epoch': 1.8}) +[2026-06-18 03:17:38,774][starVLA.training.train_starvla][INFO] - Step 4914, Loss: {'train/loss': 0.40345144271850586, 'train/grad_norm_pre_clip': 21.794940948486328, 'timing/data': 0.0132097527384758, 'timing/model': 1.028402323834598, 'learning_rate/qwen_vl_interface': 2.4860883426928053e-06, 'learning_rate/action_model': 2.4860883426928053e-05, 'epoch': 1.8}) +[2026-06-18 03:17:43,147][starVLA.training.train_starvla][INFO] - Step 4915, Loss: {'train/loss': 0.28285419940948486, 'train/grad_norm_pre_clip': 22.433643341064453, 'timing/data': 0.019917420111596584, 'timing/model': 1.3343957485631108, 'learning_rate/qwen_vl_interface': 2.4843297618103127e-06, 'learning_rate/action_model': 2.4843297618103128e-05, 'epoch': 1.8}) +[2026-06-18 03:17:47,450][starVLA.training.train_starvla][INFO] - Step 4916, Loss: {'train/loss': 0.26779991388320923, 'train/grad_norm_pre_clip': 15.358760833740234, 'timing/data': 0.011489016935229301, 'timing/model': 1.0807761773467064, 'learning_rate/qwen_vl_interface': 2.482571754254237e-06, 'learning_rate/action_model': 2.4825717542542372e-05, 'epoch': 1.8}) +[2026-06-18 03:17:51,953][starVLA.training.train_starvla][INFO] - Step 4917, Loss: {'train/loss': 0.14849913120269775, 'train/grad_norm_pre_clip': 14.1063814163208, 'timing/data': 0.012185810133814812, 'timing/model': 1.067010683938861, 'learning_rate/qwen_vl_interface': 2.4808143203890137e-06, 'learning_rate/action_model': 2.480814320389014e-05, 'epoch': 1.8}) +[2026-06-18 03:17:56,595][starVLA.training.train_starvla][INFO] - Step 4918, Loss: {'train/loss': 0.3599422574043274, 'train/grad_norm_pre_clip': 24.6279239654541, 'timing/data': 0.014687749557197094, 'timing/model': 1.0504845539107919, 'learning_rate/qwen_vl_interface': 2.4790574605789627e-06, 'learning_rate/action_model': 2.4790574605789624e-05, 'epoch': 1.8}) +[2026-06-18 03:18:01,083][starVLA.training.train_starvla][INFO] - Step 4919, Loss: {'train/loss': 0.10077084600925446, 'train/grad_norm_pre_clip': 11.089994430541992, 'timing/data': 0.01207378413528204, 'timing/model': 1.0682580163702369, 'learning_rate/qwen_vl_interface': 2.4773011751882807e-06, 'learning_rate/action_model': 2.4773011751882804e-05, 'epoch': 1.8}) +[2026-06-18 03:18:05,669][starVLA.training.train_starvla][INFO] - Step 4920, Loss: {'train/loss': 0.3047216236591339, 'train/grad_norm_pre_clip': 18.479917526245117, 'timing/data': 0.011841190047562122, 'timing/model': 1.4920720616355538, 'learning_rate/qwen_vl_interface': 2.4755454645810473e-06, 'learning_rate/action_model': 2.4755454645810473e-05, 'epoch': 1.8}) +[2026-06-18 03:18:10,352][starVLA.training.train_starvla][INFO] - Step 4921, Loss: {'train/loss': 0.3448612093925476, 'train/grad_norm_pre_clip': 41.0471305847168, 'timing/data': 0.012334234081208706, 'timing/model': 1.0808578832075, 'learning_rate/qwen_vl_interface': 2.473790329121228e-06, 'learning_rate/action_model': 2.473790329121228e-05, 'epoch': 1.8}) +[2026-06-18 03:18:15,079][starVLA.training.train_starvla][INFO] - Step 4922, Loss: {'train/loss': 0.5216023325920105, 'train/grad_norm_pre_clip': 24.435457229614258, 'timing/data': 0.022600949741899967, 'timing/model': 1.0480058584362268, 'learning_rate/qwen_vl_interface': 2.4720357691726595e-06, 'learning_rate/action_model': 2.4720357691726594e-05, 'epoch': 1.8}) +[2026-06-18 03:18:19,670][starVLA.training.train_starvla][INFO] - Step 4923, Loss: {'train/loss': 0.3033719062805176, 'train/grad_norm_pre_clip': 15.538895606994629, 'timing/data': 0.01628292351961136, 'timing/model': 1.1122085489332676, 'learning_rate/qwen_vl_interface': 2.4702817850990645e-06, 'learning_rate/action_model': 2.470281785099064e-05, 'epoch': 1.8}) +[2026-06-18 03:18:24,191][starVLA.training.train_starvla][INFO] - Step 4924, Loss: {'train/loss': 0.2907326817512512, 'train/grad_norm_pre_clip': 15.841657638549805, 'timing/data': 0.02128264307975769, 'timing/model': 1.0760930282995105, 'learning_rate/qwen_vl_interface': 2.4685283772640452e-06, 'learning_rate/action_model': 2.468528377264045e-05, 'epoch': 1.8}) +[2026-06-18 03:18:28,781][starVLA.training.train_starvla][INFO] - Step 4925, Loss: {'train/loss': 0.22091777622699738, 'train/grad_norm_pre_clip': 17.932531356811523, 'timing/data': 0.011411408893764019, 'timing/model': 1.4363033007830381, 'learning_rate/qwen_vl_interface': 2.466775546031088e-06, 'learning_rate/action_model': 2.4667755460310878e-05, 'epoch': 1.8}) +[2026-06-18 03:18:33,233][starVLA.training.train_starvla][INFO] - Step 4926, Loss: {'train/loss': 0.5507645606994629, 'train/grad_norm_pre_clip': 46.64152908325195, 'timing/data': 0.012892104685306549, 'timing/model': 1.1337323393672705, 'learning_rate/qwen_vl_interface': 2.465023291763555e-06, 'learning_rate/action_model': 2.465023291763555e-05, 'epoch': 1.8}) +[2026-06-18 03:18:38,107][starVLA.training.train_starvla][INFO] - Step 4927, Loss: {'train/loss': 0.40311723947525024, 'train/grad_norm_pre_clip': 26.06749725341797, 'timing/data': 0.01168286707252264, 'timing/model': 1.1400986956432462, 'learning_rate/qwen_vl_interface': 2.4632716148246903e-06, 'learning_rate/action_model': 2.4632716148246903e-05, 'epoch': 1.8}) +[2026-06-18 03:18:45,025][starVLA.training.train_starvla][INFO] - Step 4928, Loss: {'train/loss': 0.36845865845680237, 'train/grad_norm_pre_clip': 27.194915771484375, 'timing/data': 0.01830740924924612, 'timing/model': 1.1813741028308868, 'learning_rate/qwen_vl_interface': 2.4615205155776185e-06, 'learning_rate/action_model': 2.4615205155776188e-05, 'epoch': 1.8}) +[2026-06-18 03:18:55,226][starVLA.training.train_starvla][INFO] - Step 4929, Loss: {'train/loss': 0.27351856231689453, 'train/grad_norm_pre_clip': 18.189485549926758, 'timing/data': 0.04470068123191595, 'timing/model': 1.3396442905068398, 'learning_rate/qwen_vl_interface': 2.4597699943853444e-06, 'learning_rate/action_model': 2.4597699943853442e-05, 'epoch': 1.8}) +[2026-06-18 03:19:01,160][starVLA.training.train_starvla][INFO] - Step 4930, Loss: {'train/loss': 0.2561987638473511, 'train/grad_norm_pre_clip': 19.86012840270996, 'timing/data': 0.04716274980455637, 'timing/model': 1.6382622085511684, 'learning_rate/qwen_vl_interface': 2.458020051610753e-06, 'learning_rate/action_model': 2.4580200516107528e-05, 'epoch': 1.81}) +[2026-06-18 03:19:09,791][starVLA.training.train_starvla][INFO] - Step 4931, Loss: {'train/loss': 0.29947447776794434, 'train/grad_norm_pre_clip': 19.34341049194336, 'timing/data': 0.04450051113963127, 'timing/model': 4.890023761428893, 'learning_rate/qwen_vl_interface': 2.456270687616609e-06, 'learning_rate/action_model': 2.4562706876166087e-05, 'epoch': 1.81}) +[2026-06-18 03:19:18,506][starVLA.training.train_starvla][INFO] - Step 4932, Loss: {'train/loss': 0.19200338423252106, 'train/grad_norm_pre_clip': 15.757875442504883, 'timing/data': 0.06882473267614841, 'timing/model': 1.1625195415690541, 'learning_rate/qwen_vl_interface': 2.4545219027655577e-06, 'learning_rate/action_model': 2.4545219027655574e-05, 'epoch': 1.81}) +[2026-06-18 03:19:24,126][starVLA.training.train_starvla][INFO] - Step 4933, Loss: {'train/loss': 0.26798301935195923, 'train/grad_norm_pre_clip': 30.04437828063965, 'timing/data': 0.09972121100872755, 'timing/model': 1.0825912989675999, 'learning_rate/qwen_vl_interface': 2.452773697420124e-06, 'learning_rate/action_model': 2.452773697420124e-05, 'epoch': 1.81}) +[2026-06-18 03:19:28,775][starVLA.training.train_starvla][INFO] - Step 4934, Loss: {'train/loss': 0.1486218422651291, 'train/grad_norm_pre_clip': 13.33414363861084, 'timing/data': 0.045938221737742424, 'timing/model': 1.062051867134869, 'learning_rate/qwen_vl_interface': 2.4510260719427127e-06, 'learning_rate/action_model': 2.4510260719427127e-05, 'epoch': 1.81}) +[2026-06-18 03:19:33,532][starVLA.training.train_starvla][INFO] - Step 4935, Loss: {'train/loss': 0.5334106087684631, 'train/grad_norm_pre_clip': 25.426225662231445, 'timing/data': 0.01176363229751587, 'timing/model': 1.365052322857082, 'learning_rate/qwen_vl_interface': 2.4492790266956092e-06, 'learning_rate/action_model': 2.449279026695609e-05, 'epoch': 1.81}) +[2026-06-18 03:19:37,927][starVLA.training.train_starvla][INFO] - Step 4936, Loss: {'train/loss': 0.16004343330860138, 'train/grad_norm_pre_clip': 17.520841598510742, 'timing/data': 0.02259436622262001, 'timing/model': 1.2106397561728954, 'learning_rate/qwen_vl_interface': 2.4475325620409765e-06, 'learning_rate/action_model': 2.4475325620409764e-05, 'epoch': 1.81}) +[2026-06-18 03:19:42,534][starVLA.training.train_starvla][INFO] - Step 4937, Loss: {'train/loss': 0.245773047208786, 'train/grad_norm_pre_clip': 29.99348258972168, 'timing/data': 0.016128445975482464, 'timing/model': 1.0644607497379184, 'learning_rate/qwen_vl_interface': 2.445786678340859e-06, 'learning_rate/action_model': 2.4457866783408588e-05, 'epoch': 1.81}) +[2026-06-18 03:19:46,949][starVLA.training.train_starvla][INFO] - Step 4938, Loss: {'train/loss': 0.2671411633491516, 'train/grad_norm_pre_clip': 14.66196060180664, 'timing/data': 0.0164507944136858, 'timing/model': 1.0273530017584562, 'learning_rate/qwen_vl_interface': 2.444041375957179e-06, 'learning_rate/action_model': 2.444041375957179e-05, 'epoch': 1.81}) +[2026-06-18 03:19:51,335][starVLA.training.train_starvla][INFO] - Step 4939, Loss: {'train/loss': 0.5576565265655518, 'train/grad_norm_pre_clip': 26.872055053710938, 'timing/data': 0.012623111717402935, 'timing/model': 1.0817511565983295, 'learning_rate/qwen_vl_interface': 2.4422966552517434e-06, 'learning_rate/action_model': 2.4422966552517433e-05, 'epoch': 1.81}) +[2026-06-18 03:19:56,146][starVLA.training.train_starvla][INFO] - Step 4940, Loss: {'train/loss': 0.404023677110672, 'train/grad_norm_pre_clip': 20.39276885986328, 'timing/data': 0.01181094627827406, 'timing/model': 1.4938926566392183, 'learning_rate/qwen_vl_interface': 2.440552516586232e-06, 'learning_rate/action_model': 2.440552516586232e-05, 'epoch': 1.81}) +[2026-06-18 03:20:00,377][starVLA.training.train_starvla][INFO] - Step 4941, Loss: {'train/loss': 0.5408846735954285, 'train/grad_norm_pre_clip': 25.980558395385742, 'timing/data': 0.014020724222064018, 'timing/model': 1.063923136331141, 'learning_rate/qwen_vl_interface': 2.4388089603222085e-06, 'learning_rate/action_model': 2.4388089603222087e-05, 'epoch': 1.81}) +[2026-06-18 03:20:04,915][starVLA.training.train_starvla][INFO] - Step 4942, Loss: {'train/loss': 0.2811633348464966, 'train/grad_norm_pre_clip': 36.69649887084961, 'timing/data': 0.018649106845259666, 'timing/model': 1.120635182596743, 'learning_rate/qwen_vl_interface': 2.4370659868211095e-06, 'learning_rate/action_model': 2.4370659868211097e-05, 'epoch': 1.81}) +[2026-06-18 03:20:09,714][starVLA.training.train_starvla][INFO] - Step 4943, Loss: {'train/loss': 0.8015106916427612, 'train/grad_norm_pre_clip': 33.96464920043945, 'timing/data': 0.02199250739067793, 'timing/model': 1.0800981102511287, 'learning_rate/qwen_vl_interface': 2.4353235964442597e-06, 'learning_rate/action_model': 2.4353235964442598e-05, 'epoch': 1.81}) +[2026-06-18 03:20:14,287][starVLA.training.train_starvla][INFO] - Step 4944, Loss: {'train/loss': 0.5993945598602295, 'train/grad_norm_pre_clip': 167.8226318359375, 'timing/data': 0.016566084697842598, 'timing/model': 1.0912616439163685, 'learning_rate/qwen_vl_interface': 2.433581789552857e-06, 'learning_rate/action_model': 2.433581789552857e-05, 'epoch': 1.81}) +[2026-06-18 03:20:18,901][starVLA.training.train_starvla][INFO] - Step 4945, Loss: {'train/loss': 0.3889371454715729, 'train/grad_norm_pre_clip': 27.889524459838867, 'timing/data': 0.013546228408813477, 'timing/model': 1.6025191769003868, 'learning_rate/qwen_vl_interface': 2.4318405665079806e-06, 'learning_rate/action_model': 2.4318405665079805e-05, 'epoch': 1.81}) +[2026-06-18 03:20:26,485][starVLA.training.train_starvla][INFO] - Step 4946, Loss: {'train/loss': 0.48793092370033264, 'train/grad_norm_pre_clip': 29.056976318359375, 'timing/data': 0.028972800821065903, 'timing/model': 1.0420148260891438, 'learning_rate/qwen_vl_interface': 2.4300999276705868e-06, 'learning_rate/action_model': 2.4300999276705868e-05, 'epoch': 1.81}) +[2026-06-18 03:20:35,976][starVLA.training.train_starvla][INFO] - Step 4947, Loss: {'train/loss': 0.4082309305667877, 'train/grad_norm_pre_clip': 21.98728370666504, 'timing/data': 0.031123616732656956, 'timing/model': 5.716176534071565, 'learning_rate/qwen_vl_interface': 2.428359873401512e-06, 'learning_rate/action_model': 2.4283598734015124e-05, 'epoch': 1.81}) +[2026-06-18 03:20:51,356][starVLA.training.train_starvla][INFO] - Step 4948, Loss: {'train/loss': 0.4412434995174408, 'train/grad_norm_pre_clip': 26.998445510864258, 'timing/data': 0.03452672529965639, 'timing/model': 1.200188004411757, 'learning_rate/qwen_vl_interface': 2.4266204040614722e-06, 'learning_rate/action_model': 2.4266204040614724e-05, 'epoch': 1.81}) +[2026-06-18 03:20:55,979][starVLA.training.train_starvla][INFO] - Step 4949, Loss: {'train/loss': 0.16160771250724792, 'train/grad_norm_pre_clip': 13.679845809936523, 'timing/data': 0.03705225978046656, 'timing/model': 1.043271386064589, 'learning_rate/qwen_vl_interface': 2.424881520011059e-06, 'learning_rate/action_model': 2.424881520011059e-05, 'epoch': 1.81}) +[2026-06-18 03:21:01,155][starVLA.training.train_starvla][INFO] - Step 4950, Loss: {'train/loss': 0.27000004053115845, 'train/grad_norm_pre_clip': 18.703041076660156, 'timing/data': 0.028276235796511173, 'timing/model': 1.38245302811265, 'learning_rate/qwen_vl_interface': 2.4231432216107494e-06, 'learning_rate/action_model': 2.4231432216107494e-05, 'epoch': 1.81}) +[2026-06-18 03:21:05,315][starVLA.training.train_starvla][INFO] - Step 4951, Loss: {'train/loss': 0.4054641127586365, 'train/grad_norm_pre_clip': 17.915939331054688, 'timing/data': 0.03451239876449108, 'timing/model': 1.0892662229016423, 'learning_rate/qwen_vl_interface': 2.4214055092208907e-06, 'learning_rate/action_model': 2.4214055092208907e-05, 'epoch': 1.81}) +[2026-06-18 03:21:10,118][starVLA.training.train_starvla][INFO] - Step 4952, Loss: {'train/loss': 0.26401448249816895, 'train/grad_norm_pre_clip': 16.439699172973633, 'timing/data': 0.03485955577343702, 'timing/model': 1.3270871406421065, 'learning_rate/qwen_vl_interface': 2.4196683832017134e-06, 'learning_rate/action_model': 2.4196683832017135e-05, 'epoch': 1.81}) +[2026-06-18 03:21:19,448][starVLA.training.train_starvla][INFO] - Step 4953, Loss: {'train/loss': 0.2500312924385071, 'train/grad_norm_pre_clip': 15.133322715759277, 'timing/data': 0.08865062892436981, 'timing/model': 4.922264843247831, 'learning_rate/qwen_vl_interface': 2.4179318439133245e-06, 'learning_rate/action_model': 2.417931843913324e-05, 'epoch': 1.81}) +[2026-06-18 03:21:27,858][starVLA.training.train_starvla][INFO] - Step 4954, Loss: {'train/loss': 0.2723817229270935, 'train/grad_norm_pre_clip': 18.244234085083008, 'timing/data': 0.0118726696819067, 'timing/model': 1.1755284136161208, 'learning_rate/qwen_vl_interface': 2.4161958917157126e-06, 'learning_rate/action_model': 2.4161958917157126e-05, 'epoch': 1.81}) +[2026-06-18 03:21:32,606][starVLA.training.train_starvla][INFO] - Step 4955, Loss: {'train/loss': 0.40440425276756287, 'train/grad_norm_pre_clip': 18.047340393066406, 'timing/data': 0.035043537616729736, 'timing/model': 1.3500008173286915, 'learning_rate/qwen_vl_interface': 2.4144605269687445e-06, 'learning_rate/action_model': 2.4144605269687446e-05, 'epoch': 1.81}) +[2026-06-18 03:21:36,836][starVLA.training.train_starvla][INFO] - Step 4956, Loss: {'train/loss': 0.3358341157436371, 'train/grad_norm_pre_clip': 19.38213348388672, 'timing/data': 0.03948299866169691, 'timing/model': 1.0856738714501262, 'learning_rate/qwen_vl_interface': 2.412725750032156e-06, 'learning_rate/action_model': 2.4127257500321563e-05, 'epoch': 1.81}) +[2026-06-18 03:21:44,449][starVLA.training.train_starvla][INFO] - Step 4957, Loss: {'train/loss': 0.25693103671073914, 'train/grad_norm_pre_clip': 21.95025634765625, 'timing/data': 0.027117611840367317, 'timing/model': 1.0926786242052913, 'learning_rate/qwen_vl_interface': 2.4109915612655745e-06, 'learning_rate/action_model': 2.4109915612655745e-05, 'epoch': 1.82}) +[2026-06-18 03:21:49,078][starVLA.training.train_starvla][INFO] - Step 4958, Loss: {'train/loss': 0.6825165748596191, 'train/grad_norm_pre_clip': 28.629789352416992, 'timing/data': 0.0313401622697711, 'timing/model': 1.135743492282927, 'learning_rate/qwen_vl_interface': 2.409257961028498e-06, 'learning_rate/action_model': 2.4092579610284978e-05, 'epoch': 1.82}) +[2026-06-18 03:21:53,659][starVLA.training.train_starvla][INFO] - Step 4959, Loss: {'train/loss': 0.3151453137397766, 'train/grad_norm_pre_clip': 29.902463912963867, 'timing/data': 0.03220969717949629, 'timing/model': 1.0859062084928155, 'learning_rate/qwen_vl_interface': 2.4075249496803026e-06, 'learning_rate/action_model': 2.4075249496803025e-05, 'epoch': 1.82}) +[2026-06-18 03:21:58,211][starVLA.training.train_starvla][INFO] - Step 4960, Loss: {'train/loss': 0.4093111455440521, 'train/grad_norm_pre_clip': 23.651512145996094, 'timing/data': 0.036945697851479053, 'timing/model': 1.4509930228814483, 'learning_rate/qwen_vl_interface': 2.405792527580243e-06, 'learning_rate/action_model': 2.405792527580243e-05, 'epoch': 1.82}) +[2026-06-18 03:22:02,429][starVLA.training.train_starvla][INFO] - Step 4961, Loss: {'train/loss': 0.14629840850830078, 'train/grad_norm_pre_clip': 12.495444297790527, 'timing/data': 0.03480688389390707, 'timing/model': 1.0826435033231974, 'learning_rate/qwen_vl_interface': 2.404060695087454e-06, 'learning_rate/action_model': 2.4040606950874536e-05, 'epoch': 1.82}) +[2026-06-18 03:22:07,167][starVLA.training.train_starvla][INFO] - Step 4962, Loss: {'train/loss': 0.38022381067276, 'train/grad_norm_pre_clip': 27.41036605834961, 'timing/data': 0.03376708459109068, 'timing/model': 1.183199460618198, 'learning_rate/qwen_vl_interface': 2.4023294525609437e-06, 'learning_rate/action_model': 2.4023294525609437e-05, 'epoch': 1.82}) +[2026-06-18 03:22:11,673][starVLA.training.train_starvla][INFO] - Step 4963, Loss: {'train/loss': 0.18813927471637726, 'train/grad_norm_pre_clip': 20.00543975830078, 'timing/data': 0.029703614301979542, 'timing/model': 1.0610156208276749, 'learning_rate/qwen_vl_interface': 2.400598800359604e-06, 'learning_rate/action_model': 2.4005988003596035e-05, 'epoch': 1.82}) +[2026-06-18 03:22:20,504][starVLA.training.train_starvla][INFO] - Step 4964, Loss: {'train/loss': 0.32336053252220154, 'train/grad_norm_pre_clip': 36.38420486450195, 'timing/data': 0.030035621486604214, 'timing/model': 1.4384522223845124, 'learning_rate/qwen_vl_interface': 2.398868738842198e-06, 'learning_rate/action_model': 2.3988687388421978e-05, 'epoch': 1.82}) +[2026-06-18 03:22:25,110][starVLA.training.train_starvla][INFO] - Step 4965, Loss: {'train/loss': 0.22172723710536957, 'train/grad_norm_pre_clip': 18.957477569580078, 'timing/data': 0.03360519278794527, 'timing/model': 1.4749155128374696, 'learning_rate/qwen_vl_interface': 2.3971392683673705e-06, 'learning_rate/action_model': 2.3971392683673706e-05, 'epoch': 1.82}) +[2026-06-18 03:22:32,500][starVLA.training.train_starvla][INFO] - Step 4966, Loss: {'train/loss': 0.3607061207294464, 'train/grad_norm_pre_clip': 43.95118713378906, 'timing/data': 0.031603108160197735, 'timing/model': 1.0460569076240063, 'learning_rate/qwen_vl_interface': 2.3954103892936425e-06, 'learning_rate/action_model': 2.3954103892936424e-05, 'epoch': 1.82}) +[2026-06-18 03:22:37,115][starVLA.training.train_starvla][INFO] - Step 4967, Loss: {'train/loss': 0.1858265995979309, 'train/grad_norm_pre_clip': 23.044004440307617, 'timing/data': 0.03078068047761917, 'timing/model': 1.104957208968699, 'learning_rate/qwen_vl_interface': 2.393682101979411e-06, 'learning_rate/action_model': 2.3936821019794105e-05, 'epoch': 1.82}) +[2026-06-18 03:22:41,586][starVLA.training.train_starvla][INFO] - Step 4968, Loss: {'train/loss': 0.2585599422454834, 'train/grad_norm_pre_clip': 20.478103637695312, 'timing/data': 0.03232397884130478, 'timing/model': 1.068184838630259, 'learning_rate/qwen_vl_interface': 2.3919544067829546e-06, 'learning_rate/action_model': 2.3919544067829545e-05, 'epoch': 1.82}) +[2026-06-18 03:22:46,377][starVLA.training.train_starvla][INFO] - Step 4969, Loss: {'train/loss': 0.34735366702079773, 'train/grad_norm_pre_clip': 25.226455688476562, 'timing/data': 0.03043318074196577, 'timing/model': 1.0417277598753572, 'learning_rate/qwen_vl_interface': 2.3902273040624267e-06, 'learning_rate/action_model': 2.390227304062427e-05, 'epoch': 1.82}) +[2026-06-18 03:22:50,699][starVLA.training.train_starvla][INFO] - Step 4970, Loss: {'train/loss': 0.16563574969768524, 'train/grad_norm_pre_clip': 13.643452644348145, 'timing/data': 0.029232578352093697, 'timing/model': 1.3470198074355721, 'learning_rate/qwen_vl_interface': 2.388500794175855e-06, 'learning_rate/action_model': 2.3885007941758548e-05, 'epoch': 1.82}) +[2026-06-18 03:22:54,684][starVLA.training.train_starvla][INFO] - Step 4971, Loss: {'train/loss': 0.3063901662826538, 'train/grad_norm_pre_clip': 24.612655639648438, 'timing/data': 0.02447524294257164, 'timing/model': 1.0072448700666428, 'learning_rate/qwen_vl_interface': 2.386774877481145e-06, 'learning_rate/action_model': 2.3867748774811452e-05, 'epoch': 1.82}) +[2026-06-18 03:22:58,855][starVLA.training.train_starvla][INFO] - Step 4972, Loss: {'train/loss': 0.09203076362609863, 'train/grad_norm_pre_clip': 28.510631561279297, 'timing/data': 0.025120782665908337, 'timing/model': 1.0087153147906065, 'learning_rate/qwen_vl_interface': 2.385049554336087e-06, 'learning_rate/action_model': 2.385049554336087e-05, 'epoch': 1.82}) +[2026-06-18 03:23:03,128][starVLA.training.train_starvla][INFO] - Step 4973, Loss: {'train/loss': 0.16837312281131744, 'train/grad_norm_pre_clip': 12.644906997680664, 'timing/data': 0.028238363564014435, 'timing/model': 1.0575996972620487, 'learning_rate/qwen_vl_interface': 2.3833248250983383e-06, 'learning_rate/action_model': 2.3833248250983383e-05, 'epoch': 1.82}) +[2026-06-18 03:23:07,243][starVLA.training.train_starvla][INFO] - Step 4974, Loss: {'train/loss': 0.31668511033058167, 'train/grad_norm_pre_clip': 24.68499755859375, 'timing/data': 0.030535553582012653, 'timing/model': 1.0118101043626666, 'learning_rate/qwen_vl_interface': 2.3816006901254396e-06, 'learning_rate/action_model': 2.3816006901254396e-05, 'epoch': 1.82}) +[2026-06-18 03:23:11,482][starVLA.training.train_starvla][INFO] - Step 4975, Loss: {'train/loss': 0.1611744612455368, 'train/grad_norm_pre_clip': 16.14900016784668, 'timing/data': 0.025413340888917446, 'timing/model': 1.2148522408679128, 'learning_rate/qwen_vl_interface': 2.3798771497748012e-06, 'learning_rate/action_model': 2.379877149774801e-05, 'epoch': 1.82}) +[2026-06-18 03:23:18,845][starVLA.training.train_starvla][INFO] - Step 4976, Loss: {'train/loss': 0.23303689062595367, 'train/grad_norm_pre_clip': 21.885522842407227, 'timing/data': 0.026427377946674824, 'timing/model': 1.0574747920036316, 'learning_rate/qwen_vl_interface': 2.3781542044037195e-06, 'learning_rate/action_model': 2.3781542044037193e-05, 'epoch': 1.82}) +[2026-06-18 03:23:23,057][starVLA.training.train_starvla][INFO] - Step 4977, Loss: {'train/loss': 0.36624056100845337, 'train/grad_norm_pre_clip': 22.928916931152344, 'timing/data': 0.026348174549639225, 'timing/model': 1.0127717349678278, 'learning_rate/qwen_vl_interface': 2.376431854369362e-06, 'learning_rate/action_model': 2.3764318543693618e-05, 'epoch': 1.82}) +[2026-06-18 03:23:27,361][starVLA.training.train_starvla][INFO] - Step 4978, Loss: {'train/loss': 0.3170265853404999, 'train/grad_norm_pre_clip': 21.78697967529297, 'timing/data': 0.03261983580887318, 'timing/model': 1.0466598579660058, 'learning_rate/qwen_vl_interface': 2.374710100028773e-06, 'learning_rate/action_model': 2.3747101000287728e-05, 'epoch': 1.82}) +[2026-06-18 03:23:31,608][starVLA.training.train_starvla][INFO] - Step 4979, Loss: {'train/loss': 0.1443738341331482, 'train/grad_norm_pre_clip': 14.91677188873291, 'timing/data': 0.03097162116318941, 'timing/model': 1.030343876220286, 'learning_rate/qwen_vl_interface': 2.3729889417388736e-06, 'learning_rate/action_model': 2.3729889417388737e-05, 'epoch': 1.82}) +[2026-06-18 03:23:35,584][starVLA.training.train_starvla][INFO] - Step 4980, Loss: {'train/loss': 0.4148074686527252, 'train/grad_norm_pre_clip': 24.74754524230957, 'timing/data': 0.03018909227102995, 'timing/model': 1.046607980504632, 'learning_rate/qwen_vl_interface': 2.3712683798564628e-06, 'learning_rate/action_model': 2.3712683798564625e-05, 'epoch': 1.82}) +[2026-06-18 03:23:40,040][starVLA.training.train_starvla][INFO] - Step 4981, Loss: {'train/loss': 0.23248706758022308, 'train/grad_norm_pre_clip': 31.57286834716797, 'timing/data': 0.03722362779080868, 'timing/model': 1.1109071420505643, 'learning_rate/qwen_vl_interface': 2.3695484147382146e-06, 'learning_rate/action_model': 2.3695484147382147e-05, 'epoch': 1.82}) +[2026-06-18 03:23:44,837][starVLA.training.train_starvla][INFO] - Step 4982, Loss: {'train/loss': 0.33297473192214966, 'train/grad_norm_pre_clip': 19.05254364013672, 'timing/data': 0.03072616085410118, 'timing/model': 1.3432788252830505, 'learning_rate/qwen_vl_interface': 2.367829046740678e-06, 'learning_rate/action_model': 2.3678290467406776e-05, 'epoch': 1.82}) +[2026-06-18 03:23:49,409][starVLA.training.train_starvla][INFO] - Step 4983, Loss: {'train/loss': 0.5536452531814575, 'train/grad_norm_pre_clip': 35.90788650512695, 'timing/data': 0.03191112354397774, 'timing/model': 1.1367309456691146, 'learning_rate/qwen_vl_interface': 2.3661102762202846e-06, 'learning_rate/action_model': 2.3661102762202843e-05, 'epoch': 1.82}) +[2026-06-18 03:23:54,427][starVLA.training.train_starvla][INFO] - Step 4984, Loss: {'train/loss': 0.23187820613384247, 'train/grad_norm_pre_clip': 17.10032844543457, 'timing/data': 0.034510218538343906, 'timing/model': 1.4369717575609684, 'learning_rate/qwen_vl_interface': 2.364392103533332e-06, 'learning_rate/action_model': 2.364392103533332e-05, 'epoch': 1.83}) +[2026-06-18 03:24:03,285][starVLA.training.train_starvla][INFO] - Step 4985, Loss: {'train/loss': 0.23338855803012848, 'train/grad_norm_pre_clip': 22.762046813964844, 'timing/data': 0.038618054240942, 'timing/model': 4.699545616284013, 'learning_rate/qwen_vl_interface': 2.362674529036002e-06, 'learning_rate/action_model': 2.362674529036002e-05, 'epoch': 1.83}) +[2026-06-18 03:24:07,982][starVLA.training.train_starvla][INFO] - Step 4986, Loss: {'train/loss': 0.44027772545814514, 'train/grad_norm_pre_clip': 36.08143615722656, 'timing/data': 0.03048733528703451, 'timing/model': 1.2022738447412848, 'learning_rate/qwen_vl_interface': 2.360957553084347e-06, 'learning_rate/action_model': 2.360957553084347e-05, 'epoch': 1.83}) +[2026-06-18 03:24:14,459][starVLA.training.train_starvla][INFO] - Step 4987, Loss: {'train/loss': 0.2050854116678238, 'train/grad_norm_pre_clip': 24.93232536315918, 'timing/data': 0.03451116383075714, 'timing/model': 2.785973965190351, 'learning_rate/qwen_vl_interface': 2.359241176034302e-06, 'learning_rate/action_model': 2.359241176034302e-05, 'epoch': 1.83}) +[2026-06-18 03:24:19,115][starVLA.training.train_starvla][INFO] - Step 4988, Loss: {'train/loss': 0.5160021185874939, 'train/grad_norm_pre_clip': 20.78263282775879, 'timing/data': 0.04407386854290962, 'timing/model': 1.0804425729438663, 'learning_rate/qwen_vl_interface': 2.3575253982416732e-06, 'learning_rate/action_model': 2.357525398241673e-05, 'epoch': 1.83}) +[2026-06-18 03:24:24,895][starVLA.training.train_starvla][INFO] - Step 4989, Loss: {'train/loss': 0.30330324172973633, 'train/grad_norm_pre_clip': 30.529516220092773, 'timing/data': 0.03086155839264393, 'timing/model': 1.1474404586479068, 'learning_rate/qwen_vl_interface': 2.3558102200621385e-06, 'learning_rate/action_model': 2.3558102200621386e-05, 'epoch': 1.83}) +[2026-06-18 03:24:29,370][starVLA.training.train_starvla][INFO] - Step 4990, Loss: {'train/loss': 0.09995520859956741, 'train/grad_norm_pre_clip': 15.028016090393066, 'timing/data': 0.03378970921039581, 'timing/model': 1.1992872832342982, 'learning_rate/qwen_vl_interface': 2.3540956418512607e-06, 'learning_rate/action_model': 2.354095641851261e-05, 'epoch': 1.83}) +[2026-06-18 03:24:35,013][starVLA.training.train_starvla][INFO] - Step 4991, Loss: {'train/loss': 0.17271840572357178, 'train/grad_norm_pre_clip': 15.107851028442383, 'timing/data': 0.050131361931562424, 'timing/model': 1.2762887254357338, 'learning_rate/qwen_vl_interface': 2.352381663964473e-06, 'learning_rate/action_model': 2.352381663964473e-05, 'epoch': 1.83}) +[2026-06-18 03:24:51,228][starVLA.training.train_starvla][INFO] - Step 4992, Loss: {'train/loss': 0.2120286375284195, 'train/grad_norm_pre_clip': 17.967191696166992, 'timing/data': 0.0619531124830246, 'timing/model': 6.018208471126854, 'learning_rate/qwen_vl_interface': 2.3506682867570846e-06, 'learning_rate/action_model': 2.3506682867570844e-05, 'epoch': 1.83}) +[2026-06-18 03:24:59,774][starVLA.training.train_starvla][INFO] - Step 4993, Loss: {'train/loss': 0.29726675152778625, 'train/grad_norm_pre_clip': 24.230072021484375, 'timing/data': 0.03155171684920788, 'timing/model': 1.0927131893113256, 'learning_rate/qwen_vl_interface': 2.3489555105842796e-06, 'learning_rate/action_model': 2.3489555105842794e-05, 'epoch': 1.83}) +[2026-06-18 03:25:08,853][starVLA.training.train_starvla][INFO] - Step 4994, Loss: {'train/loss': 0.1798618584871292, 'train/grad_norm_pre_clip': 16.02112579345703, 'timing/data': 0.031011120416224003, 'timing/model': 2.159452793188393, 'learning_rate/qwen_vl_interface': 2.347243335801119e-06, 'learning_rate/action_model': 2.347243335801119e-05, 'epoch': 1.83}) +[2026-06-18 03:25:20,873][starVLA.training.train_starvla][INFO] - Step 4995, Loss: {'train/loss': 0.2519263029098511, 'train/grad_norm_pre_clip': 28.913957595825195, 'timing/data': 0.04084122274070978, 'timing/model': 1.5550082046538591, 'learning_rate/qwen_vl_interface': 2.3455317627625385e-06, 'learning_rate/action_model': 2.3455317627625385e-05, 'epoch': 1.83}) +[2026-06-18 03:25:26,400][starVLA.training.train_starvla][INFO] - Step 4996, Loss: {'train/loss': 0.29487475752830505, 'train/grad_norm_pre_clip': 28.462997436523438, 'timing/data': 0.09100024402141571, 'timing/model': 1.968484666198492, 'learning_rate/qwen_vl_interface': 2.3438207918233495e-06, 'learning_rate/action_model': 2.3438207918233494e-05, 'epoch': 1.83}) +[2026-06-18 03:25:31,139][starVLA.training.train_starvla][INFO] - Step 4997, Loss: {'train/loss': 0.13775397837162018, 'train/grad_norm_pre_clip': 16.297033309936523, 'timing/data': 0.0312939053401351, 'timing/model': 1.0387956956401467, 'learning_rate/qwen_vl_interface': 2.3421104233382366e-06, 'learning_rate/action_model': 2.3421104233382364e-05, 'epoch': 1.83}) +[2026-06-18 03:25:36,008][starVLA.training.train_starvla][INFO] - Step 4998, Loss: {'train/loss': 0.10333269834518433, 'train/grad_norm_pre_clip': 11.459750175476074, 'timing/data': 0.034280579537153244, 'timing/model': 1.4225946329534054, 'learning_rate/qwen_vl_interface': 2.340400657661762e-06, 'learning_rate/action_model': 2.340400657661762e-05, 'epoch': 1.83}) +[2026-06-18 03:25:41,614][starVLA.training.train_starvla][INFO] - Step 4999, Loss: {'train/loss': 0.0983869731426239, 'train/grad_norm_pre_clip': 10.69903564453125, 'timing/data': 0.045839449390769005, 'timing/model': 1.0642850268632174, 'learning_rate/qwen_vl_interface': 2.338691495148363e-06, 'learning_rate/action_model': 2.3386914951483626e-05, 'epoch': 1.83}) +[2026-06-18 03:29:41,424][starVLA.training.train_starvla][INFO] - Step 5000, Loss: {'train/loss': 0.22600077092647552, 'train/grad_norm_pre_clip': 15.079581260681152, 'eval/loss': 0.5498151183128357, 'eval/latency_2/loss': 0.5498151183128357, 'eval/demon_fire_precision': 0.9453939141308879, 'eval/demon_fire_recall': 0.9489539748953975, 'eval/demon_fire_f1': 0.9471705992900397, 'eval/demon_fire_rate_model': 0.6675013912075681, 'eval/demon_fire_rate_teacher': 0.6649972175848636, 'eval/demon_left_precision': 0.9035087719298246, 'eval/demon_left_recall': 0.8673684210526316, 'eval/demon_left_f1': 0.8850698174006445, 'eval/demon_left_rate_model': 0.25375626043405675, 'eval/demon_left_rate_teacher': 0.2643294379521425, 'eval/demon_right_precision': 0.8793859649122807, 'eval/demon_right_recall': 0.9062146892655367, 'eval/demon_right_f1': 0.89259877573734, 'eval/demon_right_rate_model': 0.25375626043405675, 'eval/demon_right_rate_teacher': 0.24624373956594323, 'eval/demon_move_f1': 0.8888342965689923, 'eval/demon_cc_f1': 0.9288811965687143, 'eval/latency_2/demon_fire_precision': 0.9453939141308879, 'eval/latency_2/demon_fire_recall': 0.9489539748953975, 'eval/latency_2/demon_fire_f1': 0.9471705992900397, 'eval/latency_2/demon_fire_rate_model': 0.6675013912075681, 'eval/latency_2/demon_fire_rate_teacher': 0.6649972175848636, 'eval/latency_2/demon_left_precision': 0.9035087719298246, 'eval/latency_2/demon_left_recall': 0.8673684210526316, 'eval/latency_2/demon_left_f1': 0.8850698174006445, 'eval/latency_2/demon_left_rate_model': 0.25375626043405675, 'eval/latency_2/demon_left_rate_teacher': 0.2643294379521425, 'eval/latency_2/demon_right_precision': 0.8793859649122807, 'eval/latency_2/demon_right_recall': 0.9062146892655367, 'eval/latency_2/demon_right_f1': 0.89259877573734, 'eval/latency_2/demon_right_rate_model': 0.25375626043405675, 'eval/latency_2/demon_right_rate_teacher': 0.24624373956594323, 'eval/latency_2/demon_move_f1': 0.8888342965689923, 'eval/latency_2/demon_cc_f1': 0.9288811965687143, 'timing/data': 0.045097241178154945, 'timing/model': 1.6876494362950325, 'learning_rate/qwen_vl_interface': 2.336982936152347e-06, 'learning_rate/action_model': 2.3369829361523468e-05, 'epoch': 1.83}) +[2026-06-18 03:29:41,514][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_5000_state +[2026-06-18 03:29:41,522][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 03:29:50,278][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_5000_state/model.safetensors +[2026-06-18 03:30:05,800][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_5000_state/optimizer.bin +[2026-06-18 03:30:05,801][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_5000_state/sampler.bin +[2026-06-18 03:30:05,802][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_5000_state/sampler_1.bin +[2026-06-18 03:30:05,804][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_5000_state/random_states_0.pkl +[2026-06-18 03:30:05,808][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 5000 +[2026-06-18 03:30:05,809][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 03:30:05,825][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 03:30:10,886][starVLA.training.train_starvla][INFO] - Step 5001, Loss: {'train/loss': 0.35999882221221924, 'train/grad_norm_pre_clip': 23.044918060302734, 'timing/data': 0.03074797336012125, 'timing/model': 1.6212514946237206, 'learning_rate/qwen_vl_interface': 2.3352749810279047e-06, 'learning_rate/action_model': 2.3352749810279047e-05, 'epoch': 1.83}) +[2026-06-18 03:30:14,987][starVLA.training.train_starvla][INFO] - Step 5002, Loss: {'train/loss': 0.33763396739959717, 'train/grad_norm_pre_clip': 17.912858963012695, 'timing/data': 0.031176721677184105, 'timing/model': 1.03796233702451, 'learning_rate/qwen_vl_interface': 2.3335676301290972e-06, 'learning_rate/action_model': 2.3335676301290973e-05, 'epoch': 1.83}) +[2026-06-18 03:30:20,062][starVLA.training.train_starvla][INFO] - Step 5003, Loss: {'train/loss': 0.19416606426239014, 'train/grad_norm_pre_clip': 19.246715545654297, 'timing/data': 0.033590592443943024, 'timing/model': 1.1650461545214057, 'learning_rate/qwen_vl_interface': 2.331860883809857e-06, 'learning_rate/action_model': 2.3318608838098565e-05, 'epoch': 1.83}) +[2026-06-18 03:30:24,990][starVLA.training.train_starvla][INFO] - Step 5004, Loss: {'train/loss': 0.2542595863342285, 'train/grad_norm_pre_clip': 30.126588821411133, 'timing/data': 0.03808291535824537, 'timing/model': 1.242103493772447, 'learning_rate/qwen_vl_interface': 2.330154742423993e-06, 'learning_rate/action_model': 2.330154742423993e-05, 'epoch': 1.83}) +[2026-06-18 03:30:29,922][starVLA.training.train_starvla][INFO] - Step 5005, Loss: {'train/loss': 0.6243070960044861, 'train/grad_norm_pre_clip': 33.997467041015625, 'timing/data': 0.03839213214814663, 'timing/model': 1.1838561026379466, 'learning_rate/qwen_vl_interface': 2.328449206325195e-06, 'learning_rate/action_model': 2.3284492063251947e-05, 'epoch': 1.83}) +[2026-06-18 03:30:35,868][starVLA.training.train_starvla][INFO] - Step 5006, Loss: {'train/loss': 0.24088969826698303, 'train/grad_norm_pre_clip': 17.478023529052734, 'timing/data': 0.0327884117141366, 'timing/model': 1.3749058954417706, 'learning_rate/qwen_vl_interface': 2.32674427586702e-06, 'learning_rate/action_model': 2.3267442758670198e-05, 'epoch': 1.83}) +[2026-06-18 03:30:40,226][starVLA.training.train_starvla][INFO] - Step 5007, Loss: {'train/loss': 0.3945327699184418, 'train/grad_norm_pre_clip': 25.195449829101562, 'timing/data': 0.036980184726417065, 'timing/model': 1.0664081368595362, 'learning_rate/qwen_vl_interface': 2.3250399514029018e-06, 'learning_rate/action_model': 2.3250399514029016e-05, 'epoch': 1.83}) +[2026-06-18 03:30:45,394][starVLA.training.train_starvla][INFO] - Step 5008, Loss: {'train/loss': 0.3293241560459137, 'train/grad_norm_pre_clip': 21.708850860595703, 'timing/data': 0.04608994256705046, 'timing/model': 1.801508734934032, 'learning_rate/qwen_vl_interface': 2.32333623328615e-06, 'learning_rate/action_model': 2.3233362332861498e-05, 'epoch': 1.83}) +[2026-06-18 03:30:51,514][starVLA.training.train_starvla][INFO] - Step 5009, Loss: {'train/loss': 0.28903424739837646, 'train/grad_norm_pre_clip': 24.21861457824707, 'timing/data': 0.0366699006408453, 'timing/model': 2.2761265579611063, 'learning_rate/qwen_vl_interface': 2.3216331218699457e-06, 'learning_rate/action_model': 2.3216331218699456e-05, 'epoch': 1.83}) +[2026-06-18 03:30:59,892][starVLA.training.train_starvla][INFO] - Step 5010, Loss: {'train/loss': 0.22414669394493103, 'train/grad_norm_pre_clip': 21.07663345336914, 'timing/data': 0.03913283720612526, 'timing/model': 1.033115427941084, 'learning_rate/qwen_vl_interface': 2.3199306175073465e-06, 'learning_rate/action_model': 2.3199306175073467e-05, 'epoch': 1.83}) +[2026-06-18 03:31:04,924][starVLA.training.train_starvla][INFO] - Step 5011, Loss: {'train/loss': 0.2697417140007019, 'train/grad_norm_pre_clip': 24.896873474121094, 'timing/data': 0.03313526511192322, 'timing/model': 1.3629621863365173, 'learning_rate/qwen_vl_interface': 2.318228720551283e-06, 'learning_rate/action_model': 2.318228720551283e-05, 'epoch': 1.83}) +[2026-06-18 03:31:09,846][starVLA.training.train_starvla][INFO] - Step 5012, Loss: {'train/loss': 0.22983482480049133, 'train/grad_norm_pre_clip': 23.82223892211914, 'timing/data': 0.03391982801258564, 'timing/model': 1.441174958832562, 'learning_rate/qwen_vl_interface': 2.3165274313545606e-06, 'learning_rate/action_model': 2.3165274313545606e-05, 'epoch': 1.84}) +[2026-06-18 03:31:14,730][starVLA.training.train_starvla][INFO] - Step 5013, Loss: {'train/loss': 0.3064720928668976, 'train/grad_norm_pre_clip': 19.872901916503906, 'timing/data': 0.03819427452981472, 'timing/model': 1.0970704117789865, 'learning_rate/qwen_vl_interface': 2.3148267502698584e-06, 'learning_rate/action_model': 2.3148267502698584e-05, 'epoch': 1.84}) +[2026-06-18 03:31:19,585][starVLA.training.train_starvla][INFO] - Step 5014, Loss: {'train/loss': 0.2780952751636505, 'train/grad_norm_pre_clip': 20.186410903930664, 'timing/data': 0.01874916162341833, 'timing/model': 1.1416036151349545, 'learning_rate/qwen_vl_interface': 2.313126677649729e-06, 'learning_rate/action_model': 2.313126677649729e-05, 'epoch': 1.84}) +[2026-06-18 03:31:32,555][starVLA.training.train_starvla][INFO] - Step 5015, Loss: {'train/loss': 0.5423163175582886, 'train/grad_norm_pre_clip': 27.33570671081543, 'timing/data': 0.013448314741253853, 'timing/model': 4.810521947219968, 'learning_rate/qwen_vl_interface': 2.3114272138465975e-06, 'learning_rate/action_model': 2.3114272138465977e-05, 'epoch': 1.84}) +[2026-06-18 03:31:37,214][starVLA.training.train_starvla][INFO] - Step 5016, Loss: {'train/loss': 0.531832218170166, 'train/grad_norm_pre_clip': 26.20602035522461, 'timing/data': 0.013315849006175995, 'timing/model': 1.2662613093852997, 'learning_rate/qwen_vl_interface': 2.3097283592127706e-06, 'learning_rate/action_model': 2.3097283592127706e-05, 'epoch': 1.84}) +[2026-06-18 03:31:47,052][starVLA.training.train_starvla][INFO] - Step 5017, Loss: {'train/loss': 0.10905773937702179, 'train/grad_norm_pre_clip': 11.379626274108887, 'timing/data': 0.013193435966968536, 'timing/model': 1.1916315034031868, 'learning_rate/qwen_vl_interface': 2.3080301141004165e-06, 'learning_rate/action_model': 2.3080301141004165e-05, 'epoch': 1.84}) +[2026-06-18 03:32:03,543][starVLA.training.train_starvla][INFO] - Step 5018, Loss: {'train/loss': 0.2870102822780609, 'train/grad_norm_pre_clip': 19.797204971313477, 'timing/data': 0.01346065104007721, 'timing/model': 5.367813708260655, 'learning_rate/qwen_vl_interface': 2.3063324788615843e-06, 'learning_rate/action_model': 2.306332478861584e-05, 'epoch': 1.84}) +[2026-06-18 03:32:13,310][starVLA.training.train_starvla][INFO] - Step 5019, Loss: {'train/loss': 0.22573894262313843, 'train/grad_norm_pre_clip': 15.429717063903809, 'timing/data': 0.012207338586449623, 'timing/model': 1.3214788241311908, 'learning_rate/qwen_vl_interface': 2.304635453848198e-06, 'learning_rate/action_model': 2.3046354538481983e-05, 'epoch': 1.84}) +[2026-06-18 03:32:23,676][starVLA.training.train_starvla][INFO] - Step 5020, Loss: {'train/loss': 0.08854064345359802, 'train/grad_norm_pre_clip': 7.700130462646484, 'timing/data': 0.014033311977982521, 'timing/model': 4.3335863733664155, 'learning_rate/qwen_vl_interface': 2.3029390394120517e-06, 'learning_rate/action_model': 2.3029390394120513e-05, 'epoch': 1.84}) +[2026-06-18 03:32:36,307][starVLA.training.train_starvla][INFO] - Step 5021, Loss: {'train/loss': 0.4426482319831848, 'train/grad_norm_pre_clip': 27.205190658569336, 'timing/data': 0.01358792558312416, 'timing/model': 1.5970839336514473, 'learning_rate/qwen_vl_interface': 2.301243235904813e-06, 'learning_rate/action_model': 2.3012432359048133e-05, 'epoch': 1.84}) +[2026-06-18 03:32:40,989][starVLA.training.train_starvla][INFO] - Step 5022, Loss: {'train/loss': 0.4758724868297577, 'train/grad_norm_pre_clip': 27.16434669494629, 'timing/data': 0.01878315582871437, 'timing/model': 1.0570792704820633, 'learning_rate/qwen_vl_interface': 2.299548043678026e-06, 'learning_rate/action_model': 2.299548043678026e-05, 'epoch': 1.84}) +[2026-06-18 03:32:51,090][starVLA.training.train_starvla][INFO] - Step 5023, Loss: {'train/loss': 0.08730212599039078, 'train/grad_norm_pre_clip': 8.977458953857422, 'timing/data': 0.016676457598805428, 'timing/model': 1.2490957686677575, 'learning_rate/qwen_vl_interface': 2.297853463083103e-06, 'learning_rate/action_model': 2.297853463083103e-05, 'epoch': 1.84}) +[2026-06-18 03:32:56,702][starVLA.training.train_starvla][INFO] - Step 5024, Loss: {'train/loss': 0.44291526079177856, 'train/grad_norm_pre_clip': 23.170141220092773, 'timing/data': 0.012587134726345539, 'timing/model': 1.0689349612221122, 'learning_rate/qwen_vl_interface': 2.296159494471335e-06, 'learning_rate/action_model': 2.2961594944713347e-05, 'epoch': 1.84}) +[2026-06-18 03:33:01,353][starVLA.training.train_starvla][INFO] - Step 5025, Loss: {'train/loss': 0.22798287868499756, 'train/grad_norm_pre_clip': 18.120031356811523, 'timing/data': 0.013893417082726955, 'timing/model': 1.0956853469833732, 'learning_rate/qwen_vl_interface': 2.294466138193881e-06, 'learning_rate/action_model': 2.2944661381938812e-05, 'epoch': 1.84}) +[2026-06-18 03:33:05,824][starVLA.training.train_starvla][INFO] - Step 5026, Loss: {'train/loss': 0.6332202553749084, 'train/grad_norm_pre_clip': 32.715118408203125, 'timing/data': 0.01112415362149477, 'timing/model': 1.4516652477905154, 'learning_rate/qwen_vl_interface': 2.2927733946017777e-06, 'learning_rate/action_model': 2.2927733946017778e-05, 'epoch': 1.84}) +[2026-06-18 03:33:10,739][starVLA.training.train_starvla][INFO] - Step 5027, Loss: {'train/loss': 0.4016173779964447, 'train/grad_norm_pre_clip': 27.91233253479004, 'timing/data': 0.012618650682270527, 'timing/model': 1.161808312870562, 'learning_rate/qwen_vl_interface': 2.2910812640459314e-06, 'learning_rate/action_model': 2.2910812640459315e-05, 'epoch': 1.84}) +[2026-06-18 03:33:15,163][starVLA.training.train_starvla][INFO] - Step 5028, Loss: {'train/loss': 0.43420571088790894, 'train/grad_norm_pre_clip': 24.499107360839844, 'timing/data': 0.011436352506279945, 'timing/model': 1.0262472406029701, 'learning_rate/qwen_vl_interface': 2.289389746877123e-06, 'learning_rate/action_model': 2.2893897468771228e-05, 'epoch': 1.84}) +[2026-06-18 03:33:19,832][starVLA.training.train_starvla][INFO] - Step 5029, Loss: {'train/loss': 0.40821999311447144, 'train/grad_norm_pre_clip': 26.19116973876953, 'timing/data': 0.01090968493372202, 'timing/model': 1.1368326330557466, 'learning_rate/qwen_vl_interface': 2.2876988434460046e-06, 'learning_rate/action_model': 2.2876988434460046e-05, 'epoch': 1.84}) +[2026-06-18 03:33:24,491][starVLA.training.train_starvla][INFO] - Step 5030, Loss: {'train/loss': 0.314730703830719, 'train/grad_norm_pre_clip': 18.470312118530273, 'timing/data': 0.013354035094380379, 'timing/model': 1.4053145023062825, 'learning_rate/qwen_vl_interface': 2.2860085541031048e-06, 'learning_rate/action_model': 2.2860085541031047e-05, 'epoch': 1.84}) +[2026-06-18 03:33:29,042][starVLA.training.train_starvla][INFO] - Step 5031, Loss: {'train/loss': 0.37920624017715454, 'train/grad_norm_pre_clip': 24.446434020996094, 'timing/data': 0.013242091983556747, 'timing/model': 1.2886614929884672, 'learning_rate/qwen_vl_interface': 2.284318879198823e-06, 'learning_rate/action_model': 2.284318879198823e-05, 'epoch': 1.84}) +[2026-06-18 03:33:33,148][starVLA.training.train_starvla][INFO] - Step 5032, Loss: {'train/loss': 0.17578107118606567, 'train/grad_norm_pre_clip': 14.76491641998291, 'timing/data': 0.014536366797983646, 'timing/model': 1.0933507001027465, 'learning_rate/qwen_vl_interface': 2.2826298190834266e-06, 'learning_rate/action_model': 2.2826298190834266e-05, 'epoch': 1.84}) +[2026-06-18 03:33:37,653][starVLA.training.train_starvla][INFO] - Step 5033, Loss: {'train/loss': 0.21217066049575806, 'train/grad_norm_pre_clip': 13.659310340881348, 'timing/data': 0.016678396612405777, 'timing/model': 1.0223288889974356, 'learning_rate/qwen_vl_interface': 2.2809413741070603e-06, 'learning_rate/action_model': 2.28094137410706e-05, 'epoch': 1.84}) +[2026-06-18 03:33:42,043][starVLA.training.train_starvla][INFO] - Step 5034, Loss: {'train/loss': 0.32069358229637146, 'train/grad_norm_pre_clip': 21.02212142944336, 'timing/data': 0.01392445620149374, 'timing/model': 1.0817499998956919, 'learning_rate/qwen_vl_interface': 2.2792535446197435e-06, 'learning_rate/action_model': 2.2792535446197436e-05, 'epoch': 1.84}) +[2026-06-18 03:33:46,510][starVLA.training.train_starvla][INFO] - Step 5035, Loss: {'train/loss': 0.3738366663455963, 'train/grad_norm_pre_clip': 22.068899154663086, 'timing/data': 0.012736836448311806, 'timing/model': 1.0555135551840067, 'learning_rate/qwen_vl_interface': 2.2775663309713643e-06, 'learning_rate/action_model': 2.2775663309713645e-05, 'epoch': 1.84}) +[2026-06-18 03:33:50,984][starVLA.training.train_starvla][INFO] - Step 5036, Loss: {'train/loss': 0.3290804326534271, 'train/grad_norm_pre_clip': 21.127967834472656, 'timing/data': 0.018521212972700596, 'timing/model': 1.3571156468242407, 'learning_rate/qwen_vl_interface': 2.2758797335116843e-06, 'learning_rate/action_model': 2.2758797335116844e-05, 'epoch': 1.84}) +[2026-06-18 03:33:55,317][starVLA.training.train_starvla][INFO] - Step 5037, Loss: {'train/loss': 0.13529439270496368, 'train/grad_norm_pre_clip': 10.549151420593262, 'timing/data': 0.016136854887008667, 'timing/model': 1.0692774504423141, 'learning_rate/qwen_vl_interface': 2.2741937525903324e-06, 'learning_rate/action_model': 2.2741937525903324e-05, 'epoch': 1.84}) +[2026-06-18 03:33:59,788][starVLA.training.train_starvla][INFO] - Step 5038, Loss: {'train/loss': 0.1284414380788803, 'train/grad_norm_pre_clip': 11.887346267700195, 'timing/data': 0.011275188997387886, 'timing/model': 1.0295201605185866, 'learning_rate/qwen_vl_interface': 2.2725083885568198e-06, 'learning_rate/action_model': 2.2725083885568197e-05, 'epoch': 1.84}) +[2026-06-18 03:34:04,386][starVLA.training.train_starvla][INFO] - Step 5039, Loss: {'train/loss': 0.3104289770126343, 'train/grad_norm_pre_clip': 25.778745651245117, 'timing/data': 0.013027088716626167, 'timing/model': 1.042244028300047, 'learning_rate/qwen_vl_interface': 2.2708236417605217e-06, 'learning_rate/action_model': 2.2708236417605216e-05, 'epoch': 1.85}) +[2026-06-18 03:34:10,092][starVLA.training.train_starvla][INFO] - Step 5040, Loss: {'train/loss': 0.31584224104881287, 'train/grad_norm_pre_clip': 26.998497009277344, 'timing/data': 0.02545664831995964, 'timing/model': 1.1566025409847498, 'learning_rate/qwen_vl_interface': 2.2691395125506885e-06, 'learning_rate/action_model': 2.2691395125506885e-05, 'epoch': 1.85}) +[2026-06-18 03:34:15,864][starVLA.training.train_starvla][INFO] - Step 5041, Loss: {'train/loss': 0.31501349806785583, 'train/grad_norm_pre_clip': 20.98322296142578, 'timing/data': 0.011854217387735844, 'timing/model': 2.092873171903193, 'learning_rate/qwen_vl_interface': 2.2674560012764426e-06, 'learning_rate/action_model': 2.2674560012764423e-05, 'epoch': 1.85}) +[2026-06-18 03:34:25,252][starVLA.training.train_starvla][INFO] - Step 5042, Loss: {'train/loss': 0.5876532196998596, 'train/grad_norm_pre_clip': 23.245468139648438, 'timing/data': 0.011655126698315144, 'timing/model': 1.085125932469964, 'learning_rate/qwen_vl_interface': 2.2657731082867764e-06, 'learning_rate/action_model': 2.265773108286776e-05, 'epoch': 1.85}) +[2026-06-18 03:34:33,373][starVLA.training.train_starvla][INFO] - Step 5043, Loss: {'train/loss': 0.7847214937210083, 'train/grad_norm_pre_clip': 35.70881271362305, 'timing/data': 0.01023116149008274, 'timing/model': 2.5361069412901998, 'learning_rate/qwen_vl_interface': 2.264090833930555e-06, 'learning_rate/action_model': 2.2640908339305553e-05, 'epoch': 1.85}) +[2026-06-18 03:34:38,553][starVLA.training.train_starvla][INFO] - Step 5044, Loss: {'train/loss': 0.17676201462745667, 'train/grad_norm_pre_clip': 12.878997802734375, 'timing/data': 0.013113176450133324, 'timing/model': 2.0168099040165544, 'learning_rate/qwen_vl_interface': 2.2624091785565165e-06, 'learning_rate/action_model': 2.2624091785565167e-05, 'epoch': 1.85}) +[2026-06-18 03:34:42,972][starVLA.training.train_starvla][INFO] - Step 5045, Loss: {'train/loss': 0.27002784609794617, 'train/grad_norm_pre_clip': 17.616065979003906, 'timing/data': 0.01168110966682434, 'timing/model': 1.459870076738298, 'learning_rate/qwen_vl_interface': 2.2607281425132728e-06, 'learning_rate/action_model': 2.260728142513273e-05, 'epoch': 1.85}) +[2026-06-18 03:34:47,264][starVLA.training.train_starvla][INFO] - Step 5046, Loss: {'train/loss': 0.16042380034923553, 'train/grad_norm_pre_clip': 15.032730102539062, 'timing/data': 0.010378596372902393, 'timing/model': 1.2540055625140667, 'learning_rate/qwen_vl_interface': 2.2590477261493e-06, 'learning_rate/action_model': 2.2590477261493e-05, 'epoch': 1.85}) +[2026-06-18 03:34:51,490][starVLA.training.train_starvla][INFO] - Step 5047, Loss: {'train/loss': 0.25423288345336914, 'train/grad_norm_pre_clip': 34.9785041809082, 'timing/data': 0.010310646146535873, 'timing/model': 1.0318809589371085, 'learning_rate/qwen_vl_interface': 2.257367929812953e-06, 'learning_rate/action_model': 2.2573679298129525e-05, 'epoch': 1.85}) +[2026-06-18 03:34:58,424][starVLA.training.train_starvla][INFO] - Step 5048, Loss: {'train/loss': 0.3484457731246948, 'train/grad_norm_pre_clip': 22.68358612060547, 'timing/data': 0.01077335886657238, 'timing/model': 1.0298566985875368, 'learning_rate/qwen_vl_interface': 2.2556887538524514e-06, 'learning_rate/action_model': 2.2556887538524512e-05, 'epoch': 1.85}) +[2026-06-18 03:35:02,629][starVLA.training.train_starvla][INFO] - Step 5049, Loss: {'train/loss': 0.2710826098918915, 'train/grad_norm_pre_clip': 25.659868240356445, 'timing/data': 0.010603410191833973, 'timing/model': 1.003159049898386, 'learning_rate/qwen_vl_interface': 2.2540101986158953e-06, 'learning_rate/action_model': 2.2540101986158953e-05, 'epoch': 1.85}) +[2026-06-18 03:35:06,928][starVLA.training.train_starvla][INFO] - Step 5050, Loss: {'train/loss': 0.45979517698287964, 'train/grad_norm_pre_clip': 28.583555221557617, 'timing/data': 0.010540480725467205, 'timing/model': 1.143192371353507, 'learning_rate/qwen_vl_interface': 2.2523322644512506e-06, 'learning_rate/action_model': 2.2523322644512504e-05, 'epoch': 1.85}) +[2026-06-18 03:35:11,150][starVLA.training.train_starvla][INFO] - Step 5051, Loss: {'train/loss': 0.0542367585003376, 'train/grad_norm_pre_clip': 11.767391204833984, 'timing/data': 0.011267296969890594, 'timing/model': 1.265792883001268, 'learning_rate/qwen_vl_interface': 2.2506549517063497e-06, 'learning_rate/action_model': 2.2506549517063496e-05, 'epoch': 1.85}) +[2026-06-18 03:35:15,090][starVLA.training.train_starvla][INFO] - Step 5052, Loss: {'train/loss': 0.3438751995563507, 'train/grad_norm_pre_clip': 23.028179168701172, 'timing/data': 0.009925205260515213, 'timing/model': 1.0123651893809438, 'learning_rate/qwen_vl_interface': 2.248978260728907e-06, 'learning_rate/action_model': 2.2489782607289067e-05, 'epoch': 1.85}) +[2026-06-18 03:35:19,338][starVLA.training.train_starvla][INFO] - Step 5053, Loss: {'train/loss': 0.0935368686914444, 'train/grad_norm_pre_clip': 23.1113338470459, 'timing/data': 0.009973824955523014, 'timing/model': 1.0392834795638919, 'learning_rate/qwen_vl_interface': 2.2473021918664987e-06, 'learning_rate/action_model': 2.247302191866499e-05, 'epoch': 1.85}) +[2026-06-18 03:35:23,622][starVLA.training.train_starvla][INFO] - Step 5054, Loss: {'train/loss': 0.6492953300476074, 'train/grad_norm_pre_clip': 28.65227508544922, 'timing/data': 0.009444138966500759, 'timing/model': 1.0457407496869564, 'learning_rate/qwen_vl_interface': 2.245626745466578e-06, 'learning_rate/action_model': 2.2456267454665782e-05, 'epoch': 1.85}) +[2026-06-18 03:35:28,686][starVLA.training.train_starvla][INFO] - Step 5055, Loss: {'train/loss': 0.5254141092300415, 'train/grad_norm_pre_clip': 26.38288688659668, 'timing/data': 0.012202736921608448, 'timing/model': 1.1980657000094652, 'learning_rate/qwen_vl_interface': 2.2439519218764657e-06, 'learning_rate/action_model': 2.2439519218764656e-05, 'epoch': 1.85}) +[2026-06-18 03:35:33,683][starVLA.training.train_starvla][INFO] - Step 5056, Loss: {'train/loss': 0.3446862995624542, 'train/grad_norm_pre_clip': 29.728914260864258, 'timing/data': 0.01852362137287855, 'timing/model': 1.4276547627523541, 'learning_rate/qwen_vl_interface': 2.242277721443355e-06, 'learning_rate/action_model': 2.2422777214433545e-05, 'epoch': 1.85}) +[2026-06-18 03:35:45,654][starVLA.training.train_starvla][INFO] - Step 5057, Loss: {'train/loss': 0.16165567934513092, 'train/grad_norm_pre_clip': 17.076128005981445, 'timing/data': 0.013834553770720959, 'timing/model': 1.109895097091794, 'learning_rate/qwen_vl_interface': 2.2406041445143086e-06, 'learning_rate/action_model': 2.2406041445143086e-05, 'epoch': 1.85}) +[2026-06-18 03:35:53,583][starVLA.training.train_starvla][INFO] - Step 5058, Loss: {'train/loss': 0.3581148386001587, 'train/grad_norm_pre_clip': 30.66669464111328, 'timing/data': 0.012745296582579613, 'timing/model': 1.0367097947746515, 'learning_rate/qwen_vl_interface': 2.2389311914362614e-06, 'learning_rate/action_model': 2.2389311914362616e-05, 'epoch': 1.85}) +[2026-06-18 03:36:00,774][starVLA.training.train_starvla][INFO] - Step 5059, Loss: {'train/loss': 0.6138814687728882, 'train/grad_norm_pre_clip': 42.55344009399414, 'timing/data': 0.07097528502345085, 'timing/model': 3.248424280434847, 'learning_rate/qwen_vl_interface': 2.2372588625560185e-06, 'learning_rate/action_model': 2.2372588625560185e-05, 'epoch': 1.85}) +[2026-06-18 03:36:05,429][starVLA.training.train_starvla][INFO] - Step 5060, Loss: {'train/loss': 0.2544563114643097, 'train/grad_norm_pre_clip': 23.06424903869629, 'timing/data': 0.010940860025584698, 'timing/model': 1.054319916293025, 'learning_rate/qwen_vl_interface': 2.2355871582202544e-06, 'learning_rate/action_model': 2.2355871582202544e-05, 'epoch': 1.85}) +[2026-06-18 03:36:10,686][starVLA.training.train_starvla][INFO] - Step 5061, Loss: {'train/loss': 0.2892046272754669, 'train/grad_norm_pre_clip': 18.121109008789062, 'timing/data': 0.011880896054208279, 'timing/model': 1.6178520433604717, 'learning_rate/qwen_vl_interface': 2.2339160787755175e-06, 'learning_rate/action_model': 2.2339160787755174e-05, 'epoch': 1.85}) +[2026-06-18 03:36:16,017][starVLA.training.train_starvla][INFO] - Step 5062, Loss: {'train/loss': 0.2956922948360443, 'train/grad_norm_pre_clip': 36.54673767089844, 'timing/data': 0.01844643708318472, 'timing/model': 1.0924872448667884, 'learning_rate/qwen_vl_interface': 2.23224562456822e-06, 'learning_rate/action_model': 2.23224562456822e-05, 'epoch': 1.85}) +[2026-06-18 03:36:22,246][starVLA.training.train_starvla][INFO] - Step 5063, Loss: {'train/loss': 0.32403162121772766, 'train/grad_norm_pre_clip': 19.705995559692383, 'timing/data': 0.01115848496556282, 'timing/model': 1.125460053794086, 'learning_rate/qwen_vl_interface': 2.2305757959446538e-06, 'learning_rate/action_model': 2.2305757959446537e-05, 'epoch': 1.85}) +[2026-06-18 03:36:27,950][starVLA.training.train_starvla][INFO] - Step 5064, Loss: {'train/loss': 0.3094552755355835, 'train/grad_norm_pre_clip': 19.709436416625977, 'timing/data': 0.016639028675854206, 'timing/model': 1.9397296635434031, 'learning_rate/qwen_vl_interface': 2.228906593250976e-06, 'learning_rate/action_model': 2.228906593250976e-05, 'epoch': 1.85}) +[2026-06-18 03:36:33,381][starVLA.training.train_starvla][INFO] - Step 5065, Loss: {'train/loss': 0.20410852134227753, 'train/grad_norm_pre_clip': 15.194485664367676, 'timing/data': 0.011204282753169537, 'timing/model': 1.512829844839871, 'learning_rate/qwen_vl_interface': 2.2272380168332105e-06, 'learning_rate/action_model': 2.2272380168332103e-05, 'epoch': 1.85}) +[2026-06-18 03:36:39,132][starVLA.training.train_starvla][INFO] - Step 5066, Loss: {'train/loss': 0.20351074635982513, 'train/grad_norm_pre_clip': 17.36365509033203, 'timing/data': 0.013690696097910404, 'timing/model': 1.8986196471378207, 'learning_rate/qwen_vl_interface': 2.225570067037255e-06, 'learning_rate/action_model': 2.2255700670372548e-05, 'epoch': 1.86}) +[2026-06-18 03:36:43,981][starVLA.training.train_starvla][INFO] - Step 5067, Loss: {'train/loss': 0.2371496707201004, 'train/grad_norm_pre_clip': 22.5633602142334, 'timing/data': 0.01311668660491705, 'timing/model': 1.2013689130544662, 'learning_rate/qwen_vl_interface': 2.223902744208881e-06, 'learning_rate/action_model': 2.2239027442088808e-05, 'epoch': 1.86}) +[2026-06-18 03:36:48,806][starVLA.training.train_starvla][INFO] - Step 5068, Loss: {'train/loss': 0.15006494522094727, 'train/grad_norm_pre_clip': 13.650450706481934, 'timing/data': 0.01373500656336546, 'timing/model': 1.2159847849979997, 'learning_rate/qwen_vl_interface': 2.2222360486937238e-06, 'learning_rate/action_model': 2.2222360486937235e-05, 'epoch': 1.86}) +[2026-06-18 03:36:58,292][starVLA.training.train_starvla][INFO] - Step 5069, Loss: {'train/loss': 0.46946126222610474, 'train/grad_norm_pre_clip': 33.3211784362793, 'timing/data': 0.07082099840044975, 'timing/model': 2.0736609855666757, 'learning_rate/qwen_vl_interface': 2.220569980837292e-06, 'learning_rate/action_model': 2.220569980837292e-05, 'epoch': 1.86}) +[2026-06-18 03:37:05,345][starVLA.training.train_starvla][INFO] - Step 5070, Loss: {'train/loss': 0.8636825680732727, 'train/grad_norm_pre_clip': 102.91875457763672, 'timing/data': 0.015529705211520195, 'timing/model': 1.0761572541669011, 'learning_rate/qwen_vl_interface': 2.2189045409849625e-06, 'learning_rate/action_model': 2.2189045409849628e-05, 'epoch': 1.86}) +[2026-06-18 03:37:10,240][starVLA.training.train_starvla][INFO] - Step 5071, Loss: {'train/loss': 0.1705814152956009, 'train/grad_norm_pre_clip': 14.92053508758545, 'timing/data': 0.020673423074185848, 'timing/model': 1.5157772237434983, 'learning_rate/qwen_vl_interface': 2.2172397294819843e-06, 'learning_rate/action_model': 2.2172397294819843e-05, 'epoch': 1.86}) +[2026-06-18 03:37:15,223][starVLA.training.train_starvla][INFO] - Step 5072, Loss: {'train/loss': 0.3089620769023895, 'train/grad_norm_pre_clip': 22.82994270324707, 'timing/data': 0.012969443574547768, 'timing/model': 1.3257625261321664, 'learning_rate/qwen_vl_interface': 2.215575546673473e-06, 'learning_rate/action_model': 2.215575546673473e-05, 'epoch': 1.86}) +[2026-06-18 03:37:21,360][starVLA.training.train_starvla][INFO] - Step 5073, Loss: {'train/loss': 0.13263286650180817, 'train/grad_norm_pre_clip': 11.932061195373535, 'timing/data': 0.01287467498332262, 'timing/model': 2.6016324386000633, 'learning_rate/qwen_vl_interface': 2.213911992904416e-06, 'learning_rate/action_model': 2.213911992904416e-05, 'epoch': 1.86}) +[2026-06-18 03:37:26,554][starVLA.training.train_starvla][INFO] - Step 5074, Loss: {'train/loss': 0.35688111186027527, 'train/grad_norm_pre_clip': 18.132652282714844, 'timing/data': 0.011776190251111984, 'timing/model': 1.2518022069707513, 'learning_rate/qwen_vl_interface': 2.2122490685196703e-06, 'learning_rate/action_model': 2.2122490685196704e-05, 'epoch': 1.86}) +[2026-06-18 03:37:32,196][starVLA.training.train_starvla][INFO] - Step 5075, Loss: {'train/loss': 0.4176553785800934, 'train/grad_norm_pre_clip': 26.75624656677246, 'timing/data': 0.012603161856532097, 'timing/model': 1.255537155084312, 'learning_rate/qwen_vl_interface': 2.210586773863961e-06, 'learning_rate/action_model': 2.210586773863961e-05, 'epoch': 1.86}) +[2026-06-18 03:37:41,517][starVLA.training.train_starvla][INFO] - Step 5076, Loss: {'train/loss': 0.41642385721206665, 'train/grad_norm_pre_clip': 17.31014633178711, 'timing/data': 0.00158002320677042, 'timing/model': 4.292733686044812, 'learning_rate/qwen_vl_interface': 2.208925109281884e-06, 'learning_rate/action_model': 2.208925109281884e-05, 'epoch': 1.86}) +[2026-06-18 03:37:50,535][starVLA.training.train_starvla][INFO] - Step 5077, Loss: {'train/loss': 0.31759753823280334, 'train/grad_norm_pre_clip': 19.126312255859375, 'timing/data': 0.0009202556684613228, 'timing/model': 1.1074389787390828, 'learning_rate/qwen_vl_interface': 2.2072640751179026e-06, 'learning_rate/action_model': 2.2072640751179025e-05, 'epoch': 1.86}) +[2026-06-18 03:37:56,216][starVLA.training.train_starvla][INFO] - Step 5078, Loss: {'train/loss': 0.41432297229766846, 'train/grad_norm_pre_clip': 27.557052612304688, 'timing/data': 0.0010635312646627426, 'timing/model': 1.4464832916855812, 'learning_rate/qwen_vl_interface': 2.205603671716356e-06, 'learning_rate/action_model': 2.2056036717163557e-05, 'epoch': 1.86}) +[2026-06-18 03:38:03,022][starVLA.training.train_starvla][INFO] - Step 5079, Loss: {'train/loss': 0.4891229271888733, 'train/grad_norm_pre_clip': 23.311420440673828, 'timing/data': 0.0011819219216704369, 'timing/model': 1.2489848174154758, 'learning_rate/qwen_vl_interface': 2.2039438994214424e-06, 'learning_rate/action_model': 2.2039438994214424e-05, 'epoch': 1.86}) +[2026-06-18 03:38:08,030][starVLA.training.train_starvla][INFO] - Step 5080, Loss: {'train/loss': 0.1485406905412674, 'train/grad_norm_pre_clip': 14.02936840057373, 'timing/data': 0.0009090965613722801, 'timing/model': 1.1680954555049539, 'learning_rate/qwen_vl_interface': 2.202284758577235e-06, 'learning_rate/action_model': 2.2022847585772347e-05, 'epoch': 1.86}) +[2026-06-18 03:38:13,452][starVLA.training.train_starvla][INFO] - Step 5081, Loss: {'train/loss': 0.6004614233970642, 'train/grad_norm_pre_clip': 34.71075439453125, 'timing/data': 0.0012846197932958603, 'timing/model': 1.4767272481694818, 'learning_rate/qwen_vl_interface': 2.2006262495276774e-06, 'learning_rate/action_model': 2.2006262495276775e-05, 'epoch': 1.86}) +[2026-06-18 03:38:18,531][starVLA.training.train_starvla][INFO] - Step 5082, Loss: {'train/loss': 0.22985851764678955, 'train/grad_norm_pre_clip': 24.877199172973633, 'timing/data': 0.0035979757085442543, 'timing/model': 1.412592189386487, 'learning_rate/qwen_vl_interface': 2.1989683726165795e-06, 'learning_rate/action_model': 2.1989683726165795e-05, 'epoch': 1.86}) +[2026-06-18 03:38:32,440][starVLA.training.train_starvla][INFO] - Step 5083, Loss: {'train/loss': 0.28726598620414734, 'train/grad_norm_pre_clip': 23.82892417907715, 'timing/data': 0.0012196451425552368, 'timing/model': 1.1886480655521154, 'learning_rate/qwen_vl_interface': 2.197311128187623e-06, 'learning_rate/action_model': 2.1973111281876232e-05, 'epoch': 1.86}) +[2026-06-18 03:38:39,286][starVLA.training.train_starvla][INFO] - Step 5084, Loss: {'train/loss': 0.4531060457229614, 'train/grad_norm_pre_clip': 21.809946060180664, 'timing/data': 0.000921458937227726, 'timing/model': 1.3256691675633192, 'learning_rate/qwen_vl_interface': 2.1956545165843517e-06, 'learning_rate/action_model': 2.1956545165843517e-05, 'epoch': 1.86}) +[2026-06-18 03:38:44,332][starVLA.training.train_starvla][INFO] - Step 5085, Loss: {'train/loss': 0.12126384675502777, 'train/grad_norm_pre_clip': 12.975074768066406, 'timing/data': 0.0011496515944600105, 'timing/model': 1.1120225777849555, 'learning_rate/qwen_vl_interface': 2.1939985381501866e-06, 'learning_rate/action_model': 2.1939985381501866e-05, 'epoch': 1.86}) +[2026-06-18 03:38:52,541][starVLA.training.train_starvla][INFO] - Step 5086, Loss: {'train/loss': 0.3398151099681854, 'train/grad_norm_pre_clip': 23.732227325439453, 'timing/data': 0.001340685412287712, 'timing/model': 2.9446033453568816, 'learning_rate/qwen_vl_interface': 2.1923431932284135e-06, 'learning_rate/action_model': 2.1923431932284136e-05, 'epoch': 1.86}) +[2026-06-18 03:39:11,153][starVLA.training.train_starvla][INFO] - Step 5087, Loss: {'train/loss': 0.2936908006668091, 'train/grad_norm_pre_clip': 21.902982711791992, 'timing/data': 0.001044546253979206, 'timing/model': 4.633958422578871, 'learning_rate/qwen_vl_interface': 2.190688482162187e-06, 'learning_rate/action_model': 2.190688482162187e-05, 'epoch': 1.86}) +[2026-06-18 03:39:16,604][starVLA.training.train_starvla][INFO] - Step 5088, Loss: {'train/loss': 0.38503894209861755, 'train/grad_norm_pre_clip': 27.990053176879883, 'timing/data': 0.0012634322047233582, 'timing/model': 1.4160762997344136, 'learning_rate/qwen_vl_interface': 2.189034405294529e-06, 'learning_rate/action_model': 2.189034405294529e-05, 'epoch': 1.86}) +[2026-06-18 03:39:22,600][starVLA.training.train_starvla][INFO] - Step 5089, Loss: {'train/loss': 0.36173921823501587, 'train/grad_norm_pre_clip': 20.506580352783203, 'timing/data': 0.0010398300364613533, 'timing/model': 1.5494406120851636, 'learning_rate/qwen_vl_interface': 2.1873809629683333e-06, 'learning_rate/action_model': 2.1873809629683333e-05, 'epoch': 1.86}) +[2026-06-18 03:39:36,513][starVLA.training.train_starvla][INFO] - Step 5090, Loss: {'train/loss': 0.3037274181842804, 'train/grad_norm_pre_clip': 23.714006423950195, 'timing/data': 0.0008052252233028412, 'timing/model': 1.7693506693467498, 'learning_rate/qwen_vl_interface': 2.185728155526359e-06, 'learning_rate/action_model': 2.1857281555263588e-05, 'epoch': 1.86}) +[2026-06-18 03:39:41,349][starVLA.training.train_starvla][INFO] - Step 5091, Loss: {'train/loss': 0.28166669607162476, 'train/grad_norm_pre_clip': 18.949615478515625, 'timing/data': 0.030014206655323505, 'timing/model': 1.3974397676065564, 'learning_rate/qwen_vl_interface': 2.184075983311234e-06, 'learning_rate/action_model': 2.184075983311234e-05, 'epoch': 1.86}) +[2026-06-18 03:39:46,511][starVLA.training.train_starvla][INFO] - Step 5092, Loss: {'train/loss': 0.47208768129348755, 'train/grad_norm_pre_clip': 21.591474533081055, 'timing/data': 0.03560605924576521, 'timing/model': 1.4145207544788718, 'learning_rate/qwen_vl_interface': 2.18242444666546e-06, 'learning_rate/action_model': 2.18242444666546e-05, 'epoch': 1.86}) +[2026-06-18 03:39:51,447][starVLA.training.train_starvla][INFO] - Step 5093, Loss: {'train/loss': 0.6900525093078613, 'train/grad_norm_pre_clip': 33.19108581542969, 'timing/data': 0.02134573645889759, 'timing/model': 1.0608524223789573, 'learning_rate/qwen_vl_interface': 2.1807735459313974e-06, 'learning_rate/action_model': 2.180773545931397e-05, 'epoch': 1.86}) +[2026-06-18 03:39:56,416][starVLA.training.train_starvla][INFO] - Step 5094, Loss: {'train/loss': 0.24302247166633606, 'train/grad_norm_pre_clip': 14.991031646728516, 'timing/data': 0.02340311650186777, 'timing/model': 1.3442090069875121, 'learning_rate/qwen_vl_interface': 2.1791232814512815e-06, 'learning_rate/action_model': 2.1791232814512815e-05, 'epoch': 1.87}) +[2026-06-18 03:40:05,011][starVLA.training.train_starvla][INFO] - Step 5095, Loss: {'train/loss': 0.22471845149993896, 'train/grad_norm_pre_clip': 16.329126358032227, 'timing/data': 0.02782636322081089, 'timing/model': 1.2593572745099664, 'learning_rate/qwen_vl_interface': 2.1774736535672117e-06, 'learning_rate/action_model': 2.1774736535672115e-05, 'epoch': 1.87}) +[2026-06-18 03:40:18,109][starVLA.training.train_starvla][INFO] - Step 5096, Loss: {'train/loss': 0.19731254875659943, 'train/grad_norm_pre_clip': 21.87142562866211, 'timing/data': 0.02280557621270418, 'timing/model': 1.6116373734548688, 'learning_rate/qwen_vl_interface': 2.1758246626211617e-06, 'learning_rate/action_model': 2.1758246626211617e-05, 'epoch': 1.87}) +[2026-06-18 03:40:22,738][starVLA.training.train_starvla][INFO] - Step 5097, Loss: {'train/loss': 0.23658011853694916, 'train/grad_norm_pre_clip': 17.786027908325195, 'timing/data': 0.06859068665653467, 'timing/model': 1.3616044763475657, 'learning_rate/qwen_vl_interface': 2.174176308954969e-06, 'learning_rate/action_model': 2.174176308954969e-05, 'epoch': 1.87}) +[2026-06-18 03:40:28,928][starVLA.training.train_starvla][INFO] - Step 5098, Loss: {'train/loss': 0.2953948378562927, 'train/grad_norm_pre_clip': 18.508054733276367, 'timing/data': 0.05481620877981186, 'timing/model': 1.0915360124781728, 'learning_rate/qwen_vl_interface': 2.172528592910335e-06, 'learning_rate/action_model': 2.1725285929103348e-05, 'epoch': 1.87}) +[2026-06-18 03:40:33,632][starVLA.training.train_starvla][INFO] - Step 5099, Loss: {'train/loss': 0.12325944751501083, 'train/grad_norm_pre_clip': 18.116207122802734, 'timing/data': 0.044460603035986423, 'timing/model': 1.1611798433586955, 'learning_rate/qwen_vl_interface': 2.170881514828834e-06, 'learning_rate/action_model': 2.170881514828834e-05, 'epoch': 1.87}) +[2026-06-18 03:45:54,431][starVLA.training.train_starvla][INFO] - Step 5100, Loss: {'train/loss': 0.21183620393276215, 'train/grad_norm_pre_clip': 17.490346908569336, 'eval/loss': 0.5290589332580566, 'eval/latency_2/loss': 0.5290589332580566, 'eval/demon_fire_precision': 0.9449047224523612, 'eval/demon_fire_recall': 0.9543933054393305, 'eval/demon_fire_f1': 0.9496253122398002, 'eval/demon_fire_rate_model': 0.6716750139120757, 'eval/demon_fire_rate_teacher': 0.6649972175848636, 'eval/demon_left_precision': 0.9002145922746781, 'eval/demon_left_recall': 0.8831578947368421, 'eval/demon_left_f1': 0.8916046758767268, 'eval/demon_left_rate_model': 0.2593210907067334, 'eval/demon_left_rate_teacher': 0.2643294379521425, 'eval/demon_right_precision': 0.896976483762598, 'eval/demon_right_recall': 0.9050847457627119, 'eval/demon_right_f1': 0.9010123734533183, 'eval/demon_right_rate_model': 0.2484696716750139, 'eval/demon_right_rate_teacher': 0.24624373956594323, 'eval/demon_move_f1': 0.8963085246650225, 'eval/demon_cc_f1': 0.9329759293316723, 'eval/latency_2/demon_fire_precision': 0.9449047224523612, 'eval/latency_2/demon_fire_recall': 0.9543933054393305, 'eval/latency_2/demon_fire_f1': 0.9496253122398002, 'eval/latency_2/demon_fire_rate_model': 0.6716750139120757, 'eval/latency_2/demon_fire_rate_teacher': 0.6649972175848636, 'eval/latency_2/demon_left_precision': 0.9002145922746781, 'eval/latency_2/demon_left_recall': 0.8831578947368421, 'eval/latency_2/demon_left_f1': 0.8916046758767268, 'eval/latency_2/demon_left_rate_model': 0.2593210907067334, 'eval/latency_2/demon_left_rate_teacher': 0.2643294379521425, 'eval/latency_2/demon_right_precision': 0.896976483762598, 'eval/latency_2/demon_right_recall': 0.9050847457627119, 'eval/latency_2/demon_right_f1': 0.9010123734533183, 'eval/latency_2/demon_right_rate_model': 0.2484696716750139, 'eval/latency_2/demon_right_rate_teacher': 0.24624373956594323, 'eval/latency_2/demon_move_f1': 0.8963085246650225, 'eval/latency_2/demon_cc_f1': 0.9329759293316723, 'timing/data': 0.049016281962394714, 'timing/model': 1.4542914228513837, 'learning_rate/qwen_vl_interface': 2.1692350750519093e-06, 'learning_rate/action_model': 2.1692350750519092e-05, 'epoch': 1.87}) +[2026-06-18 03:45:54,455][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_5100_state +[2026-06-18 03:45:54,517][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 03:46:03,995][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_5100_state/model.safetensors +[2026-06-18 03:46:20,135][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_5100_state/optimizer.bin +[2026-06-18 03:46:20,136][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_5100_state/sampler.bin +[2026-06-18 03:46:20,137][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_5100_state/sampler_1.bin +[2026-06-18 03:46:20,141][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_5100_state/random_states_0.pkl +[2026-06-18 03:46:20,147][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 5100 +[2026-06-18 03:46:20,147][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 03:46:20,172][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 03:46:25,117][starVLA.training.train_starvla][INFO] - Step 5101, Loss: {'train/loss': 0.09130208194255829, 'train/grad_norm_pre_clip': 9.859375953674316, 'timing/data': 0.02760067954659462, 'timing/model': 1.0728022316470742, 'learning_rate/qwen_vl_interface': 2.167589273920869e-06, 'learning_rate/action_model': 2.167589273920869e-05, 'epoch': 1.87}) +[2026-06-18 03:46:30,422][starVLA.training.train_starvla][INFO] - Step 5102, Loss: {'train/loss': 0.30065298080444336, 'train/grad_norm_pre_clip': 25.842025756835938, 'timing/data': 0.1083901571109891, 'timing/model': 1.5054512107744813, 'learning_rate/qwen_vl_interface': 2.1659441117768888e-06, 'learning_rate/action_model': 2.1659441117768886e-05, 'epoch': 1.87}) +[2026-06-18 03:46:34,939][starVLA.training.train_starvla][INFO] - Step 5103, Loss: {'train/loss': 0.2521085739135742, 'train/grad_norm_pre_clip': 21.814987182617188, 'timing/data': 0.03482073824852705, 'timing/model': 1.2659876942634583, 'learning_rate/qwen_vl_interface': 2.164299588961013e-06, 'learning_rate/action_model': 2.1642995889610128e-05, 'epoch': 1.87}) +[2026-06-18 03:46:41,638][starVLA.training.train_starvla][INFO] - Step 5104, Loss: {'train/loss': 0.4728977382183075, 'train/grad_norm_pre_clip': 32.55787658691406, 'timing/data': 0.08608530089259148, 'timing/model': 3.103241008706391, 'learning_rate/qwen_vl_interface': 2.162655705814153e-06, 'learning_rate/action_model': 2.1626557058141527e-05, 'epoch': 1.87}) +[2026-06-18 03:46:47,099][starVLA.training.train_starvla][INFO] - Step 5105, Loss: {'train/loss': 0.4513416588306427, 'train/grad_norm_pre_clip': 35.29398727416992, 'timing/data': 0.0015898160636425018, 'timing/model': 1.136123901233077, 'learning_rate/qwen_vl_interface': 2.161012462677087e-06, 'learning_rate/action_model': 2.161012462677087e-05, 'epoch': 1.87}) +[2026-06-18 03:46:56,241][starVLA.training.train_starvla][INFO] - Step 5106, Loss: {'train/loss': 0.11898216605186462, 'train/grad_norm_pre_clip': 13.920942306518555, 'timing/data': 0.0008683623746037483, 'timing/model': 2.0629499116912484, 'learning_rate/qwen_vl_interface': 2.1593698598904612e-06, 'learning_rate/action_model': 2.1593698598904614e-05, 'epoch': 1.87}) +[2026-06-18 03:47:02,883][starVLA.training.train_starvla][INFO] - Step 5107, Loss: {'train/loss': 0.09808972477912903, 'train/grad_norm_pre_clip': 8.204985618591309, 'timing/data': 0.0010938122868537903, 'timing/model': 1.390795100480318, 'learning_rate/qwen_vl_interface': 2.1577278977947896e-06, 'learning_rate/action_model': 2.1577278977947897e-05, 'epoch': 1.87}) +[2026-06-18 03:47:09,025][starVLA.training.train_starvla][INFO] - Step 5108, Loss: {'train/loss': 0.5803099870681763, 'train/grad_norm_pre_clip': 62.820838928222656, 'timing/data': 0.0012309849262237549, 'timing/model': 2.6131285270676017, 'learning_rate/qwen_vl_interface': 2.1560865767304524e-06, 'learning_rate/action_model': 2.156086576730452e-05, 'epoch': 1.87}) +[2026-06-18 03:47:13,690][starVLA.training.train_starvla][INFO] - Step 5109, Loss: {'train/loss': 0.39284655451774597, 'train/grad_norm_pre_clip': 30.96445083618164, 'timing/data': 0.0011713029816746712, 'timing/model': 1.0398074993863702, 'learning_rate/qwen_vl_interface': 2.154445897037696e-06, 'learning_rate/action_model': 2.154445897037696e-05, 'epoch': 1.87}) +[2026-06-18 03:47:21,515][starVLA.training.train_starvla][INFO] - Step 5110, Loss: {'train/loss': 0.5137202143669128, 'train/grad_norm_pre_clip': 26.077640533447266, 'timing/data': 0.0010262317955493927, 'timing/model': 1.0558905843645334, 'learning_rate/qwen_vl_interface': 2.1528058590566353e-06, 'learning_rate/action_model': 2.1528058590566353e-05, 'epoch': 1.87}) +[2026-06-18 03:47:26,045][starVLA.training.train_starvla][INFO] - Step 5111, Loss: {'train/loss': 0.48368749022483826, 'train/grad_norm_pre_clip': 28.79111671447754, 'timing/data': 0.0010839859023690224, 'timing/model': 1.043264544568956, 'learning_rate/qwen_vl_interface': 2.1511664631272543e-06, 'learning_rate/action_model': 2.1511664631272545e-05, 'epoch': 1.87}) +[2026-06-18 03:47:30,316][starVLA.training.train_starvla][INFO] - Step 5112, Loss: {'train/loss': 0.5980954766273499, 'train/grad_norm_pre_clip': 29.020381927490234, 'timing/data': 0.0008387314155697823, 'timing/model': 1.1663520587608218, 'learning_rate/qwen_vl_interface': 2.1495277095894015e-06, 'learning_rate/action_model': 2.1495277095894014e-05, 'epoch': 1.87}) +[2026-06-18 03:47:39,996][starVLA.training.train_starvla][INFO] - Step 5113, Loss: {'train/loss': 0.41394147276878357, 'train/grad_norm_pre_clip': 21.85243034362793, 'timing/data': 0.0010413238778710365, 'timing/model': 1.118783824145794, 'learning_rate/qwen_vl_interface': 2.147889598782787e-06, 'learning_rate/action_model': 2.147889598782787e-05, 'epoch': 1.87}) +[2026-06-18 03:47:44,848][starVLA.training.train_starvla][INFO] - Step 5114, Loss: {'train/loss': 0.13380379974842072, 'train/grad_norm_pre_clip': 15.12761402130127, 'timing/data': 0.001018088310956955, 'timing/model': 1.2650950541719794, 'learning_rate/qwen_vl_interface': 2.1462521310469987e-06, 'learning_rate/action_model': 2.1462521310469987e-05, 'epoch': 1.87}) +[2026-06-18 03:47:50,118][starVLA.training.train_starvla][INFO] - Step 5115, Loss: {'train/loss': 0.33730822801589966, 'train/grad_norm_pre_clip': 20.092981338500977, 'timing/data': 0.0017031878232955933, 'timing/model': 1.361471856944263, 'learning_rate/qwen_vl_interface': 2.144615306721484e-06, 'learning_rate/action_model': 2.144615306721484e-05, 'epoch': 1.87}) +[2026-06-18 03:47:54,983][starVLA.training.train_starvla][INFO] - Step 5116, Loss: {'train/loss': 0.3303331732749939, 'train/grad_norm_pre_clip': 20.464141845703125, 'timing/data': 0.002460375428199768, 'timing/model': 1.0480665769428015, 'learning_rate/qwen_vl_interface': 2.1429791261455575e-06, 'learning_rate/action_model': 2.1429791261455572e-05, 'epoch': 1.87}) +[2026-06-18 03:48:02,864][starVLA.training.train_starvla][INFO] - Step 5117, Loss: {'train/loss': 0.33151355385780334, 'train/grad_norm_pre_clip': 19.66163444519043, 'timing/data': 0.0008670268580317497, 'timing/model': 1.560771280899644, 'learning_rate/qwen_vl_interface': 2.1413435896584006e-06, 'learning_rate/action_model': 2.1413435896584005e-05, 'epoch': 1.87}) +[2026-06-18 03:48:07,962][starVLA.training.train_starvla][INFO] - Step 5118, Loss: {'train/loss': 0.26194238662719727, 'train/grad_norm_pre_clip': 30.196308135986328, 'timing/data': 0.0007727714255452156, 'timing/model': 1.2272015772759914, 'learning_rate/qwen_vl_interface': 2.1397086975990633e-06, 'learning_rate/action_model': 2.139708697599063e-05, 'epoch': 1.87}) +[2026-06-18 03:48:13,552][starVLA.training.train_starvla][INFO] - Step 5119, Loss: {'train/loss': 0.40492182970046997, 'train/grad_norm_pre_clip': 23.871591567993164, 'timing/data': 0.0009218845516443253, 'timing/model': 1.0508466772735119, 'learning_rate/qwen_vl_interface': 2.1380744503064597e-06, 'learning_rate/action_model': 2.1380744503064598e-05, 'epoch': 1.87}) +[2026-06-18 03:48:18,147][starVLA.training.train_starvla][INFO] - Step 5120, Loss: {'train/loss': 0.1832094043493271, 'train/grad_norm_pre_clip': 20.641347885131836, 'timing/data': 0.0011017555370926857, 'timing/model': 1.1042773937806487, 'learning_rate/qwen_vl_interface': 2.1364408481193714e-06, 'learning_rate/action_model': 2.1364408481193712e-05, 'epoch': 1.87}) +[2026-06-18 03:48:22,687][starVLA.training.train_starvla][INFO] - Step 5121, Loss: {'train/loss': 0.1747366487979889, 'train/grad_norm_pre_clip': 19.862272262573242, 'timing/data': 0.0008704913780093193, 'timing/model': 1.0396333672106266, 'learning_rate/qwen_vl_interface': 2.1348078913764454e-06, 'learning_rate/action_model': 2.1348078913764452e-05, 'epoch': 1.88}) +[2026-06-18 03:48:27,737][starVLA.training.train_starvla][INFO] - Step 5122, Loss: {'train/loss': 0.4599204361438751, 'train/grad_norm_pre_clip': 34.48796844482422, 'timing/data': 0.0008960999548435211, 'timing/model': 1.553522339090705, 'learning_rate/qwen_vl_interface': 2.1331755804161957e-06, 'learning_rate/action_model': 2.133175580416196e-05, 'epoch': 1.88}) +[2026-06-18 03:48:32,354][starVLA.training.train_starvla][INFO] - Step 5123, Loss: {'train/loss': 0.33854246139526367, 'train/grad_norm_pre_clip': 21.2359676361084, 'timing/data': 0.03295746073126793, 'timing/model': 1.13654004316777, 'learning_rate/qwen_vl_interface': 2.131543915577002e-06, 'learning_rate/action_model': 2.131543915577002e-05, 'epoch': 1.88}) +[2026-06-18 03:48:37,170][starVLA.training.train_starvla][INFO] - Step 5124, Loss: {'train/loss': 0.34090766310691833, 'train/grad_norm_pre_clip': 23.68634605407715, 'timing/data': 0.03918049018830061, 'timing/model': 1.3058439576998353, 'learning_rate/qwen_vl_interface': 2.129912897197108e-06, 'learning_rate/action_model': 2.1299128971971084e-05, 'epoch': 1.88}) +[2026-06-18 03:48:47,953][starVLA.training.train_starvla][INFO] - Step 5125, Loss: {'train/loss': 0.29242777824401855, 'train/grad_norm_pre_clip': 24.69141387939453, 'timing/data': 0.03536178730428219, 'timing/model': 5.358524133451283, 'learning_rate/qwen_vl_interface': 2.1282825256146305e-06, 'learning_rate/action_model': 2.12828252561463e-05, 'epoch': 1.88}) +[2026-06-18 03:49:00,585][starVLA.training.train_starvla][INFO] - Step 5126, Loss: {'train/loss': 0.5273963212966919, 'train/grad_norm_pre_clip': 31.62062644958496, 'timing/data': 0.04439630638808012, 'timing/model': 4.628409817814827, 'learning_rate/qwen_vl_interface': 2.1266528011675453e-06, 'learning_rate/action_model': 2.126652801167545e-05, 'epoch': 1.88}) +[2026-06-18 03:49:08,582][starVLA.training.train_starvla][INFO] - Step 5127, Loss: {'train/loss': 0.6102241277694702, 'train/grad_norm_pre_clip': 30.51653480529785, 'timing/data': 0.051259925588965416, 'timing/model': 1.150583690032363, 'learning_rate/qwen_vl_interface': 2.125023724193694e-06, 'learning_rate/action_model': 2.1250237241936936e-05, 'epoch': 1.88}) +[2026-06-18 03:49:23,301][starVLA.training.train_starvla][INFO] - Step 5128, Loss: {'train/loss': 0.17152012884616852, 'train/grad_norm_pre_clip': 13.325187683105469, 'timing/data': 0.1007595993578434, 'timing/model': 1.2540654819458723, 'learning_rate/qwen_vl_interface': 2.1233952950307857e-06, 'learning_rate/action_model': 2.1233952950307855e-05, 'epoch': 1.88}) +[2026-06-18 03:49:32,803][starVLA.training.train_starvla][INFO] - Step 5129, Loss: {'train/loss': 0.40113985538482666, 'train/grad_norm_pre_clip': 23.169973373413086, 'timing/data': 0.040394299663603306, 'timing/model': 3.977517473511398, 'learning_rate/qwen_vl_interface': 2.1217675140163984e-06, 'learning_rate/action_model': 2.1217675140163983e-05, 'epoch': 1.88}) +[2026-06-18 03:49:38,069][starVLA.training.train_starvla][INFO] - Step 5130, Loss: {'train/loss': 0.12778091430664062, 'train/grad_norm_pre_clip': 13.269937515258789, 'timing/data': 0.03901228494942188, 'timing/model': 1.2451443104073405, 'learning_rate/qwen_vl_interface': 2.1201403814879715e-06, 'learning_rate/action_model': 2.1201403814879717e-05, 'epoch': 1.88}) +[2026-06-18 03:49:48,037][starVLA.training.train_starvla][INFO] - Step 5131, Loss: {'train/loss': 0.3750342130661011, 'train/grad_norm_pre_clip': 24.827186584472656, 'timing/data': 0.03371379338204861, 'timing/model': 1.6631055790930986, 'learning_rate/qwen_vl_interface': 2.1185138977828122e-06, 'learning_rate/action_model': 2.118513897782812e-05, 'epoch': 1.88}) +[2026-06-18 03:49:52,915][starVLA.training.train_starvla][INFO] - Step 5132, Loss: {'train/loss': 0.22377143800258636, 'train/grad_norm_pre_clip': 24.740711212158203, 'timing/data': 0.04688049666583538, 'timing/model': 1.40049693826586, 'learning_rate/qwen_vl_interface': 2.116888063238091e-06, 'learning_rate/action_model': 2.116888063238091e-05, 'epoch': 1.88}) +[2026-06-18 03:49:57,161][starVLA.training.train_starvla][INFO] - Step 5133, Loss: {'train/loss': 0.2048693150281906, 'train/grad_norm_pre_clip': 17.15116310119629, 'timing/data': 0.031214074231684208, 'timing/model': 1.044117260724306, 'learning_rate/qwen_vl_interface': 2.1152628781908455e-06, 'learning_rate/action_model': 2.1152628781908455e-05, 'epoch': 1.88}) +[2026-06-18 03:50:03,399][starVLA.training.train_starvla][INFO] - Step 5134, Loss: {'train/loss': 0.26491403579711914, 'train/grad_norm_pre_clip': 17.039766311645508, 'timing/data': 0.036752671003341675, 'timing/model': 1.0735132098197937, 'learning_rate/qwen_vl_interface': 2.1136383429779787e-06, 'learning_rate/action_model': 2.1136383429779788e-05, 'epoch': 1.88}) +[2026-06-18 03:50:07,896][starVLA.training.train_starvla][INFO] - Step 5135, Loss: {'train/loss': 0.4159797430038452, 'train/grad_norm_pre_clip': 24.771860122680664, 'timing/data': 0.041144946590065956, 'timing/model': 1.0351450173184276, 'learning_rate/qwen_vl_interface': 2.1120144579362582e-06, 'learning_rate/action_model': 2.1120144579362582e-05, 'epoch': 1.88}) +[2026-06-18 03:50:12,598][starVLA.training.train_starvla][INFO] - Step 5136, Loss: {'train/loss': 0.32332560420036316, 'train/grad_norm_pre_clip': 19.80532455444336, 'timing/data': 0.03274077642709017, 'timing/model': 1.0366588709875941, 'learning_rate/qwen_vl_interface': 2.1103912234023175e-06, 'learning_rate/action_model': 2.1103912234023175e-05, 'epoch': 1.88}) +[2026-06-18 03:50:17,265][starVLA.training.train_starvla][INFO] - Step 5137, Loss: {'train/loss': 0.32146912813186646, 'train/grad_norm_pre_clip': 28.47670555114746, 'timing/data': 0.03273760434240103, 'timing/model': 1.3807603372260928, 'learning_rate/qwen_vl_interface': 2.1087686397126544e-06, 'learning_rate/action_model': 2.1087686397126545e-05, 'epoch': 1.88}) +[2026-06-18 03:50:21,699][starVLA.training.train_starvla][INFO] - Step 5138, Loss: {'train/loss': 0.32354456186294556, 'train/grad_norm_pre_clip': 22.573286056518555, 'timing/data': 0.04209690075367689, 'timing/model': 1.313949592411518, 'learning_rate/qwen_vl_interface': 2.107146707203632e-06, 'learning_rate/action_model': 2.1071467072036323e-05, 'epoch': 1.88}) +[2026-06-18 03:50:26,703][starVLA.training.train_starvla][INFO] - Step 5139, Loss: {'train/loss': 0.3991798460483551, 'train/grad_norm_pre_clip': 43.62501907348633, 'timing/data': 0.05581745319068432, 'timing/model': 1.0464149788022041, 'learning_rate/qwen_vl_interface': 2.105525426211478e-06, 'learning_rate/action_model': 2.1055254262114778e-05, 'epoch': 1.88}) +[2026-06-18 03:50:31,455][starVLA.training.train_starvla][INFO] - Step 5140, Loss: {'train/loss': 0.2202906608581543, 'train/grad_norm_pre_clip': 15.851582527160645, 'timing/data': 0.03255430981516838, 'timing/model': 1.3749269116669893, 'learning_rate/qwen_vl_interface': 2.1039047970722894e-06, 'learning_rate/action_model': 2.1039047970722893e-05, 'epoch': 1.88}) +[2026-06-18 03:50:36,139][starVLA.training.train_starvla][INFO] - Step 5141, Loss: {'train/loss': 0.21863336861133575, 'train/grad_norm_pre_clip': 18.445133209228516, 'timing/data': 0.031366683542728424, 'timing/model': 1.0613426538184285, 'learning_rate/qwen_vl_interface': 2.1022848201220204e-06, 'learning_rate/action_model': 2.1022848201220204e-05, 'epoch': 1.88}) +[2026-06-18 03:50:40,525][starVLA.training.train_starvla][INFO] - Step 5142, Loss: {'train/loss': 0.5333975553512573, 'train/grad_norm_pre_clip': 26.07714080810547, 'timing/data': 0.031906028278172016, 'timing/model': 1.3771814154461026, 'learning_rate/qwen_vl_interface': 2.1006654956964935e-06, 'learning_rate/action_model': 2.1006654956964936e-05, 'epoch': 1.88}) +[2026-06-18 03:50:44,615][starVLA.training.train_starvla][INFO] - Step 5143, Loss: {'train/loss': 0.3291940987110138, 'train/grad_norm_pre_clip': 16.836959838867188, 'timing/data': 0.042542003095149994, 'timing/model': 1.0548525908961892, 'learning_rate/qwen_vl_interface': 2.0990468241314004e-06, 'learning_rate/action_model': 2.0990468241314002e-05, 'epoch': 1.88}) +[2026-06-18 03:50:49,113][starVLA.training.train_starvla][INFO] - Step 5144, Loss: {'train/loss': 0.33223217725753784, 'train/grad_norm_pre_clip': 21.7351016998291, 'timing/data': 0.03805437218397856, 'timing/model': 1.1447482453659177, 'learning_rate/qwen_vl_interface': 2.0974288057622905e-06, 'learning_rate/action_model': 2.0974288057622903e-05, 'epoch': 1.88}) +[2026-06-18 03:50:53,880][starVLA.training.train_starvla][INFO] - Step 5145, Loss: {'train/loss': 0.15809233486652374, 'train/grad_norm_pre_clip': 12.12082576751709, 'timing/data': 0.04004623927175999, 'timing/model': 1.165914305485785, 'learning_rate/qwen_vl_interface': 2.095811440924584e-06, 'learning_rate/action_model': 2.0958114409245835e-05, 'epoch': 1.88}) +[2026-06-18 03:50:59,322][starVLA.training.train_starvla][INFO] - Step 5146, Loss: {'train/loss': 0.2996625006198883, 'train/grad_norm_pre_clip': 20.981910705566406, 'timing/data': 0.040138041600584984, 'timing/model': 1.6226617973297834, 'learning_rate/qwen_vl_interface': 2.0941947299535554e-06, 'learning_rate/action_model': 2.0941947299535552e-05, 'epoch': 1.88}) +[2026-06-18 03:51:03,894][starVLA.training.train_starvla][INFO] - Step 5147, Loss: {'train/loss': 0.2731795012950897, 'train/grad_norm_pre_clip': 18.233213424682617, 'timing/data': 0.03706740867346525, 'timing/model': 1.523051481693983, 'learning_rate/qwen_vl_interface': 2.0925786731843573e-06, 'learning_rate/action_model': 2.0925786731843572e-05, 'epoch': 1.88}) +[2026-06-18 03:51:08,060][starVLA.training.train_starvla][INFO] - Step 5148, Loss: {'train/loss': 0.10724414885044098, 'train/grad_norm_pre_clip': 15.063146591186523, 'timing/data': 0.032429659739136696, 'timing/model': 1.0560268880799413, 'learning_rate/qwen_vl_interface': 2.0909632709519976e-06, 'learning_rate/action_model': 2.090963270951998e-05, 'epoch': 1.89}) +[2026-06-18 03:51:12,853][starVLA.training.train_starvla][INFO] - Step 5149, Loss: {'train/loss': 0.2851221263408661, 'train/grad_norm_pre_clip': 17.5657958984375, 'timing/data': 0.03462514281272888, 'timing/model': 1.0792588638141751, 'learning_rate/qwen_vl_interface': 2.0893485235913514e-06, 'learning_rate/action_model': 2.0893485235913514e-05, 'epoch': 1.89}) +[2026-06-18 03:51:19,474][starVLA.training.train_starvla][INFO] - Step 5150, Loss: {'train/loss': 0.28253060579299927, 'train/grad_norm_pre_clip': 16.242422103881836, 'timing/data': 0.040639858692884445, 'timing/model': 2.7188731115311384, 'learning_rate/qwen_vl_interface': 2.0877344314371575e-06, 'learning_rate/action_model': 2.0877344314371573e-05, 'epoch': 1.89}) +[2026-06-18 03:51:24,049][starVLA.training.train_starvla][INFO] - Step 5151, Loss: {'train/loss': 0.38011476397514343, 'train/grad_norm_pre_clip': 25.197927474975586, 'timing/data': 0.0345555329695344, 'timing/model': 1.043289969675243, 'learning_rate/qwen_vl_interface': 2.086120994824018e-06, 'learning_rate/action_model': 2.0861209948240178e-05, 'epoch': 1.89}) +[2026-06-18 03:51:28,101][starVLA.training.train_starvla][INFO] - Step 5152, Loss: {'train/loss': 0.21329516172409058, 'train/grad_norm_pre_clip': 21.94663429260254, 'timing/data': 0.03129693865776062, 'timing/model': 1.03722582757473, 'learning_rate/qwen_vl_interface': 2.0845082140864012e-06, 'learning_rate/action_model': 2.084508214086401e-05, 'epoch': 1.89}) +[2026-06-18 03:51:32,787][starVLA.training.train_starvla][INFO] - Step 5153, Loss: {'train/loss': 0.33492356538772583, 'train/grad_norm_pre_clip': 20.068214416503906, 'timing/data': 0.03992271330207586, 'timing/model': 1.1755535081028938, 'learning_rate/qwen_vl_interface': 2.0828960895586365e-06, 'learning_rate/action_model': 2.0828960895586364e-05, 'epoch': 1.89}) +[2026-06-18 03:51:37,607][starVLA.training.train_starvla][INFO] - Step 5154, Loss: {'train/loss': 0.4834085702896118, 'train/grad_norm_pre_clip': 32.43181610107422, 'timing/data': 0.04090023320168257, 'timing/model': 1.2622049683704972, 'learning_rate/qwen_vl_interface': 2.0812846215749227e-06, 'learning_rate/action_model': 2.0812846215749227e-05, 'epoch': 1.89}) +[2026-06-18 03:51:45,287][starVLA.training.train_starvla][INFO] - Step 5155, Loss: {'train/loss': 0.26920488476753235, 'train/grad_norm_pre_clip': 22.65349578857422, 'timing/data': 0.16180699411779642, 'timing/model': 2.769350077956915, 'learning_rate/qwen_vl_interface': 2.079673810469316e-06, 'learning_rate/action_model': 2.0796738104693158e-05, 'epoch': 1.89}) +[2026-06-18 03:51:55,658][starVLA.training.train_starvla][INFO] - Step 5156, Loss: {'train/loss': 0.16253045201301575, 'train/grad_norm_pre_clip': 14.720610618591309, 'timing/data': 0.03316565975546837, 'timing/model': 4.204417857341468, 'learning_rate/qwen_vl_interface': 2.0780636565757392e-06, 'learning_rate/action_model': 2.0780636565757392e-05, 'epoch': 1.89}) +[2026-06-18 03:52:06,907][starVLA.training.train_starvla][INFO] - Step 5157, Loss: {'train/loss': 0.15561175346374512, 'train/grad_norm_pre_clip': 15.634872436523438, 'timing/data': 0.032850916497409344, 'timing/model': 4.92745643761009, 'learning_rate/qwen_vl_interface': 2.0764541602279777e-06, 'learning_rate/action_model': 2.0764541602279778e-05, 'epoch': 1.89}) +[2026-06-18 03:52:13,663][starVLA.training.train_starvla][INFO] - Step 5158, Loss: {'train/loss': 0.23592789471149445, 'train/grad_norm_pre_clip': 21.961021423339844, 'timing/data': 0.030494301579892635, 'timing/model': 1.0691006993874907, 'learning_rate/qwen_vl_interface': 2.0748453217596856e-06, 'learning_rate/action_model': 2.0748453217596856e-05, 'epoch': 1.89}) +[2026-06-18 03:52:18,295][starVLA.training.train_starvla][INFO] - Step 5159, Loss: {'train/loss': 0.6913571953773499, 'train/grad_norm_pre_clip': 30.28545379638672, 'timing/data': 0.048881106078624725, 'timing/model': 1.1178362928330898, 'learning_rate/qwen_vl_interface': 2.0732371415043766e-06, 'learning_rate/action_model': 2.0732371415043765e-05, 'epoch': 1.89}) +[2026-06-18 03:52:22,880][starVLA.training.train_starvla][INFO] - Step 5160, Loss: {'train/loss': 0.18556712567806244, 'train/grad_norm_pre_clip': 14.547118186950684, 'timing/data': 0.04115397669374943, 'timing/model': 1.0380374202504754, 'learning_rate/qwen_vl_interface': 2.071629619795425e-06, 'learning_rate/action_model': 2.071629619795425e-05, 'epoch': 1.89}) +[2026-06-18 03:52:27,556][starVLA.training.train_starvla][INFO] - Step 5161, Loss: {'train/loss': 0.1375221461057663, 'train/grad_norm_pre_clip': 18.839982986450195, 'timing/data': 0.03006844874471426, 'timing/model': 1.0550606828182936, 'learning_rate/qwen_vl_interface': 2.070022756966071e-06, 'learning_rate/action_model': 2.0700227569660712e-05, 'epoch': 1.89}) +[2026-06-18 03:52:31,941][starVLA.training.train_starvla][INFO] - Step 5162, Loss: {'train/loss': 0.1405792385339737, 'train/grad_norm_pre_clip': 15.474370956420898, 'timing/data': 0.03322594705969095, 'timing/model': 1.3521338067948818, 'learning_rate/qwen_vl_interface': 2.068416553349424e-06, 'learning_rate/action_model': 2.068416553349424e-05, 'epoch': 1.89}) +[2026-06-18 03:52:36,545][starVLA.training.train_starvla][INFO] - Step 5163, Loss: {'train/loss': 0.3010592460632324, 'train/grad_norm_pre_clip': 16.392316818237305, 'timing/data': 0.04177820589393377, 'timing/model': 1.0984931252896786, 'learning_rate/qwen_vl_interface': 2.0668110092784484e-06, 'learning_rate/action_model': 2.0668110092784484e-05, 'epoch': 1.89}) +[2026-06-18 03:52:40,907][starVLA.training.train_starvla][INFO] - Step 5164, Loss: {'train/loss': 0.2073317915201187, 'train/grad_norm_pre_clip': 13.643085479736328, 'timing/data': 0.03278666827827692, 'timing/model': 1.038868423551321, 'learning_rate/qwen_vl_interface': 2.0652061250859758e-06, 'learning_rate/action_model': 2.0652061250859756e-05, 'epoch': 1.89}) +[2026-06-18 03:52:45,463][starVLA.training.train_starvla][INFO] - Step 5165, Loss: {'train/loss': 0.1115337684750557, 'train/grad_norm_pre_clip': 10.92171573638916, 'timing/data': 0.03504483122378588, 'timing/model': 1.0766995698213577, 'learning_rate/qwen_vl_interface': 2.0636019011047005e-06, 'learning_rate/action_model': 2.0636019011047003e-05, 'epoch': 1.89}) +[2026-06-18 03:52:50,298][starVLA.training.train_starvla][INFO] - Step 5166, Loss: {'train/loss': 0.3070809543132782, 'train/grad_norm_pre_clip': 18.088335037231445, 'timing/data': 0.03505421429872513, 'timing/model': 1.5885932967066765, 'learning_rate/qwen_vl_interface': 2.0619983376671796e-06, 'learning_rate/action_model': 2.0619983376671793e-05, 'epoch': 1.89}) +[2026-06-18 03:52:57,019][starVLA.training.train_starvla][INFO] - Step 5167, Loss: {'train/loss': 0.3913058042526245, 'train/grad_norm_pre_clip': 30.808626174926758, 'timing/data': 0.033362941816449165, 'timing/model': 3.584838878363371, 'learning_rate/qwen_vl_interface': 2.060395435105834e-06, 'learning_rate/action_model': 2.060395435105834e-05, 'epoch': 1.89}) +[2026-06-18 03:53:01,255][starVLA.training.train_starvla][INFO] - Step 5168, Loss: {'train/loss': 0.4083259701728821, 'train/grad_norm_pre_clip': 25.411632537841797, 'timing/data': 0.03430826123803854, 'timing/model': 1.0408910969272256, 'learning_rate/qwen_vl_interface': 2.0587931937529465e-06, 'learning_rate/action_model': 2.0587931937529464e-05, 'epoch': 1.89}) +[2026-06-18 03:53:10,343][starVLA.training.train_starvla][INFO] - Step 5169, Loss: {'train/loss': 0.414655476808548, 'train/grad_norm_pre_clip': 36.719146728515625, 'timing/data': 0.04130479134619236, 'timing/model': 1.0683390060439706, 'learning_rate/qwen_vl_interface': 2.0571916139406648e-06, 'learning_rate/action_model': 2.0571916139406647e-05, 'epoch': 1.89}) +[2026-06-18 03:53:15,573][starVLA.training.train_starvla][INFO] - Step 5170, Loss: {'train/loss': 0.2544119954109192, 'train/grad_norm_pre_clip': 21.445545196533203, 'timing/data': 0.043473455123603344, 'timing/model': 1.069893510080874, 'learning_rate/qwen_vl_interface': 2.055590696000996e-06, 'learning_rate/action_model': 2.0555906960009962e-05, 'epoch': 1.89}) +[2026-06-18 03:53:20,320][starVLA.training.train_starvla][INFO] - Step 5171, Loss: {'train/loss': 0.33410170674324036, 'train/grad_norm_pre_clip': 21.197086334228516, 'timing/data': 0.04767397977411747, 'timing/model': 1.0815685596317053, 'learning_rate/qwen_vl_interface': 2.0539904402658144e-06, 'learning_rate/action_model': 2.0539904402658142e-05, 'epoch': 1.89}) +[2026-06-18 03:53:26,795][starVLA.training.train_starvla][INFO] - Step 5172, Loss: {'train/loss': 0.4814630150794983, 'train/grad_norm_pre_clip': 27.327598571777344, 'timing/data': 0.03537240531295538, 'timing/model': 3.223995323292911, 'learning_rate/qwen_vl_interface': 2.052390847066851e-06, 'learning_rate/action_model': 2.052390847066851e-05, 'epoch': 1.89}) +[2026-06-18 03:53:31,359][starVLA.training.train_starvla][INFO] - Step 5173, Loss: {'train/loss': 0.20940649509429932, 'train/grad_norm_pre_clip': 22.010536193847656, 'timing/data': 0.032801710069179535, 'timing/model': 1.3708746181800961, 'learning_rate/qwen_vl_interface': 2.0507919167357093e-06, 'learning_rate/action_model': 2.050791916735709e-05, 'epoch': 1.89}) +[2026-06-18 03:53:37,591][starVLA.training.train_starvla][INFO] - Step 5174, Loss: {'train/loss': 0.07030797004699707, 'train/grad_norm_pre_clip': 11.039125442504883, 'timing/data': 0.03368669468909502, 'timing/model': 1.3621707670390606, 'learning_rate/qwen_vl_interface': 2.049193649603844e-06, 'learning_rate/action_model': 2.049193649603844e-05, 'epoch': 1.89}) +[2026-06-18 03:53:42,529][starVLA.training.train_starvla][INFO] - Step 5175, Loss: {'train/loss': 0.33384156227111816, 'train/grad_norm_pre_clip': 18.53310203552246, 'timing/data': 0.050600496120750904, 'timing/model': 1.0906766559928656, 'learning_rate/qwen_vl_interface': 2.047596046002578e-06, 'learning_rate/action_model': 2.0475960460025778e-05, 'epoch': 1.89}) +[2026-06-18 03:53:50,547][starVLA.training.train_starvla][INFO] - Step 5176, Loss: {'train/loss': 0.1858285814523697, 'train/grad_norm_pre_clip': 18.881465911865234, 'timing/data': 0.04730184655636549, 'timing/model': 1.442932060919702, 'learning_rate/qwen_vl_interface': 2.0459991062630994e-06, 'learning_rate/action_model': 2.045999106263099e-05, 'epoch': 1.9}) +[2026-06-18 03:53:55,173][starVLA.training.train_starvla][INFO] - Step 5177, Loss: {'train/loss': 0.24482116103172302, 'train/grad_norm_pre_clip': 18.329809188842773, 'timing/data': 0.031757242046296597, 'timing/model': 1.4757144786417484, 'learning_rate/qwen_vl_interface': 2.0444028307164532e-06, 'learning_rate/action_model': 2.0444028307164534e-05, 'epoch': 1.9}) +[2026-06-18 03:53:59,659][starVLA.training.train_starvla][INFO] - Step 5178, Loss: {'train/loss': 0.19854997098445892, 'train/grad_norm_pre_clip': 16.02036476135254, 'timing/data': 0.03273771796375513, 'timing/model': 1.187970919534564, 'learning_rate/qwen_vl_interface': 2.0428072196935516e-06, 'learning_rate/action_model': 2.0428072196935514e-05, 'epoch': 1.9}) +[2026-06-18 03:54:04,512][starVLA.training.train_starvla][INFO] - Step 5179, Loss: {'train/loss': 0.2167951613664627, 'train/grad_norm_pre_clip': 16.203582763671875, 'timing/data': 0.03550780564546585, 'timing/model': 1.2181600704789162, 'learning_rate/qwen_vl_interface': 2.041212273525161e-06, 'learning_rate/action_model': 2.041212273525161e-05, 'epoch': 1.9}) +[2026-06-18 03:54:10,928][starVLA.training.train_starvla][INFO] - Step 5180, Loss: {'train/loss': 0.19173865020275116, 'train/grad_norm_pre_clip': 17.42017936706543, 'timing/data': 0.030560875311493874, 'timing/model': 1.266101443208754, 'learning_rate/qwen_vl_interface': 2.03961799254192e-06, 'learning_rate/action_model': 2.03961799254192e-05, 'epoch': 1.9}) +[2026-06-18 03:54:15,569][starVLA.training.train_starvla][INFO] - Step 5181, Loss: {'train/loss': 0.5775876641273499, 'train/grad_norm_pre_clip': 30.995506286621094, 'timing/data': 0.03259182162582874, 'timing/model': 1.1770281875506043, 'learning_rate/qwen_vl_interface': 2.0380243770743246e-06, 'learning_rate/action_model': 2.0380243770743247e-05, 'epoch': 1.9}) +[2026-06-18 03:54:20,109][starVLA.training.train_starvla][INFO] - Step 5182, Loss: {'train/loss': 0.22901380062103271, 'train/grad_norm_pre_clip': 41.9482421875, 'timing/data': 0.03738164808601141, 'timing/model': 1.431017005816102, 'learning_rate/qwen_vl_interface': 2.0364314274527318e-06, 'learning_rate/action_model': 2.0364314274527314e-05, 'epoch': 1.9}) +[2026-06-18 03:54:24,288][starVLA.training.train_starvla][INFO] - Step 5183, Loss: {'train/loss': 0.3389319181442261, 'train/grad_norm_pre_clip': 21.61229133605957, 'timing/data': 0.03260056860744953, 'timing/model': 1.1180843831971288, 'learning_rate/qwen_vl_interface': 2.034839144007362e-06, 'learning_rate/action_model': 2.034839144007362e-05, 'epoch': 1.9}) +[2026-06-18 03:54:28,821][starVLA.training.train_starvla][INFO] - Step 5184, Loss: {'train/loss': 0.33295977115631104, 'train/grad_norm_pre_clip': 26.02332878112793, 'timing/data': 0.03379426896572113, 'timing/model': 1.0411969786509871, 'learning_rate/qwen_vl_interface': 2.033247527068297e-06, 'learning_rate/action_model': 2.0332475270682972e-05, 'epoch': 1.9}) +[2026-06-18 03:54:34,389][starVLA.training.train_starvla][INFO] - Step 5185, Loss: {'train/loss': 0.39353302121162415, 'train/grad_norm_pre_clip': 30.400331497192383, 'timing/data': 0.03645366057753563, 'timing/model': 2.212661678902805, 'learning_rate/qwen_vl_interface': 2.0316565769654816e-06, 'learning_rate/action_model': 2.0316565769654816e-05, 'epoch': 1.9}) +[2026-06-18 03:54:40,463][starVLA.training.train_starvla][INFO] - Step 5186, Loss: {'train/loss': 0.29013791680336, 'train/grad_norm_pre_clip': 21.97098731994629, 'timing/data': 0.033596158027648926, 'timing/model': 1.0556136202067137, 'learning_rate/qwen_vl_interface': 2.0300662940287193e-06, 'learning_rate/action_model': 2.0300662940287192e-05, 'epoch': 1.9}) +[2026-06-18 03:54:45,068][starVLA.training.train_starvla][INFO] - Step 5187, Loss: {'train/loss': 0.6146103143692017, 'train/grad_norm_pre_clip': 30.012022018432617, 'timing/data': 0.033187237568199635, 'timing/model': 1.4823781317099929, 'learning_rate/qwen_vl_interface': 2.0284766785876812e-06, 'learning_rate/action_model': 2.0284766785876812e-05, 'epoch': 1.9}) +[2026-06-18 03:54:51,216][starVLA.training.train_starvla][INFO] - Step 5188, Loss: {'train/loss': 0.390206515789032, 'train/grad_norm_pre_clip': 24.490022659301758, 'timing/data': 0.03829908836632967, 'timing/model': 1.2667532050982118, 'learning_rate/qwen_vl_interface': 2.026887730971894e-06, 'learning_rate/action_model': 2.0268877309718938e-05, 'epoch': 1.9}) +[2026-06-18 03:54:56,163][starVLA.training.train_starvla][INFO] - Step 5189, Loss: {'train/loss': 0.5086259245872498, 'train/grad_norm_pre_clip': 32.954044342041016, 'timing/data': 0.0504709305241704, 'timing/model': 1.3626990104094148, 'learning_rate/qwen_vl_interface': 2.0252994515107475e-06, 'learning_rate/action_model': 2.0252994515107472e-05, 'epoch': 1.9}) +[2026-06-18 03:55:01,993][starVLA.training.train_starvla][INFO] - Step 5190, Loss: {'train/loss': 0.33647456765174866, 'train/grad_norm_pre_clip': 18.83302879333496, 'timing/data': 0.03827994968742132, 'timing/model': 1.2517460891976953, 'learning_rate/qwen_vl_interface': 2.023711840533493e-06, 'learning_rate/action_model': 2.023711840533493e-05, 'epoch': 1.9}) +[2026-06-18 03:55:06,431][starVLA.training.train_starvla][INFO] - Step 5191, Loss: {'train/loss': 0.0764213502407074, 'train/grad_norm_pre_clip': 9.551700592041016, 'timing/data': 0.032623421400785446, 'timing/model': 1.077048946171999, 'learning_rate/qwen_vl_interface': 2.0221248983692475e-06, 'learning_rate/action_model': 2.0221248983692474e-05, 'epoch': 1.9}) +[2026-06-18 03:55:13,053][starVLA.training.train_starvla][INFO] - Step 5192, Loss: {'train/loss': 0.33310940861701965, 'train/grad_norm_pre_clip': 20.548702239990234, 'timing/data': 0.053727551363408566, 'timing/model': 1.545286943204701, 'learning_rate/qwen_vl_interface': 2.020538625346984e-06, 'learning_rate/action_model': 2.020538625346984e-05, 'epoch': 1.9}) +[2026-06-18 03:55:17,391][starVLA.training.train_starvla][INFO] - Step 5193, Loss: {'train/loss': 0.33142268657684326, 'train/grad_norm_pre_clip': 22.643840789794922, 'timing/data': 0.03521772474050522, 'timing/model': 1.0383615931496024, 'learning_rate/qwen_vl_interface': 2.0189530217955382e-06, 'learning_rate/action_model': 2.018953021795538e-05, 'epoch': 1.9}) +[2026-06-18 03:55:22,529][starVLA.training.train_starvla][INFO] - Step 5194, Loss: {'train/loss': 0.3345785439014435, 'train/grad_norm_pre_clip': 28.438631057739258, 'timing/data': 0.032340931706130505, 'timing/model': 1.1109883775934577, 'learning_rate/qwen_vl_interface': 2.0173680880436076e-06, 'learning_rate/action_model': 2.0173680880436076e-05, 'epoch': 1.9}) +[2026-06-18 03:55:28,067][starVLA.training.train_starvla][INFO] - Step 5195, Loss: {'train/loss': 0.2894541919231415, 'train/grad_norm_pre_clip': 20.327701568603516, 'timing/data': 0.03591443691402674, 'timing/model': 1.5224363626912236, 'learning_rate/qwen_vl_interface': 2.015783824419751e-06, 'learning_rate/action_model': 2.0157838244197508e-05, 'epoch': 1.9}) +[2026-06-18 03:55:35,737][starVLA.training.train_starvla][INFO] - Step 5196, Loss: {'train/loss': 0.3619462549686432, 'train/grad_norm_pre_clip': 22.219568252563477, 'timing/data': 0.03136920556426048, 'timing/model': 1.1434771623462439, 'learning_rate/qwen_vl_interface': 2.0142002312523877e-06, 'learning_rate/action_model': 2.014200231252388e-05, 'epoch': 1.9}) +[2026-06-18 03:55:45,657][starVLA.training.train_starvla][INFO] - Step 5197, Loss: {'train/loss': 0.44296807050704956, 'train/grad_norm_pre_clip': 36.121795654296875, 'timing/data': 0.04734910186380148, 'timing/model': 1.8221748117357492, 'learning_rate/qwen_vl_interface': 2.0126173088697988e-06, 'learning_rate/action_model': 2.0126173088697987e-05, 'epoch': 1.9}) +[2026-06-18 03:55:50,337][starVLA.training.train_starvla][INFO] - Step 5198, Loss: {'train/loss': 0.2947671711444855, 'train/grad_norm_pre_clip': 25.610597610473633, 'timing/data': 0.032780406065285206, 'timing/model': 1.3461890015751123, 'learning_rate/qwen_vl_interface': 2.0110350576001244e-06, 'learning_rate/action_model': 2.0110350576001245e-05, 'epoch': 1.9}) +[2026-06-18 03:55:56,249][starVLA.training.train_starvla][INFO] - Step 5199, Loss: {'train/loss': 0.2872966527938843, 'train/grad_norm_pre_clip': 18.097126007080078, 'timing/data': 0.03266232926398516, 'timing/model': 1.0426447670906782, 'learning_rate/qwen_vl_interface': 2.0094534777713684e-06, 'learning_rate/action_model': 2.0094534777713684e-05, 'epoch': 1.9}) +[2026-06-18 04:00:07,309][starVLA.training.train_starvla][INFO] - Step 5200, Loss: {'train/loss': 0.35007432103157043, 'train/grad_norm_pre_clip': 26.56608772277832, 'eval/loss': 0.5383586883544922, 'eval/latency_2/loss': 0.5383586883544922, 'eval/demon_fire_precision': 0.9405250205086136, 'eval/demon_fire_recall': 0.9594142259414226, 'eval/demon_fire_f1': 0.9498757249378624, 'eval/demon_fire_rate_model': 0.6783528102392877, 'eval/demon_fire_rate_teacher': 0.6649972175848636, 'eval/demon_left_precision': 0.9060773480662984, 'eval/demon_left_recall': 0.8631578947368421, 'eval/demon_left_f1': 0.8840970350404314, 'eval/demon_left_rate_model': 0.25180856983861993, 'eval/demon_left_rate_teacher': 0.2643294379521425, 'eval/demon_right_precision': 0.916083916083916, 'eval/demon_right_recall': 0.888135593220339, 'eval/demon_right_f1': 0.9018932874354562, 'eval/demon_right_rate_model': 0.2387312186978297, 'eval/demon_right_rate_teacher': 0.24624373956594323, 'eval/demon_move_f1': 0.8929951612379438, 'eval/demon_cc_f1': 0.9320649740743955, 'eval/latency_2/demon_fire_precision': 0.9405250205086136, 'eval/latency_2/demon_fire_recall': 0.9594142259414226, 'eval/latency_2/demon_fire_f1': 0.9498757249378624, 'eval/latency_2/demon_fire_rate_model': 0.6783528102392877, 'eval/latency_2/demon_fire_rate_teacher': 0.6649972175848636, 'eval/latency_2/demon_left_precision': 0.9060773480662984, 'eval/latency_2/demon_left_recall': 0.8631578947368421, 'eval/latency_2/demon_left_f1': 0.8840970350404314, 'eval/latency_2/demon_left_rate_model': 0.25180856983861993, 'eval/latency_2/demon_left_rate_teacher': 0.2643294379521425, 'eval/latency_2/demon_right_precision': 0.916083916083916, 'eval/latency_2/demon_right_recall': 0.888135593220339, 'eval/latency_2/demon_right_f1': 0.9018932874354562, 'eval/latency_2/demon_right_rate_model': 0.2387312186978297, 'eval/latency_2/demon_right_rate_teacher': 0.24624373956594323, 'eval/latency_2/demon_move_f1': 0.8929951612379438, 'eval/latency_2/demon_cc_f1': 0.9320649740743955, 'timing/data': 0.03143030684441328, 'timing/model': 1.0709032090380788, 'learning_rate/qwen_vl_interface': 2.007872569711394e-06, 'learning_rate/action_model': 2.007872569711394e-05, 'epoch': 1.9}) +[2026-06-18 04:00:07,358][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_5200_state +[2026-06-18 04:00:07,366][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 04:00:16,586][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_5200_state/model.safetensors +[2026-06-18 04:00:31,954][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_5200_state/optimizer.bin +[2026-06-18 04:00:31,955][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_5200_state/sampler.bin +[2026-06-18 04:00:31,955][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_5200_state/sampler_1.bin +[2026-06-18 04:00:31,957][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_5200_state/random_states_0.pkl +[2026-06-18 04:00:31,995][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 5200 +[2026-06-18 04:00:31,996][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 04:00:32,012][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 04:00:40,325][starVLA.training.train_starvla][INFO] - Step 5201, Loss: {'train/loss': 0.44597724080085754, 'train/grad_norm_pre_clip': 28.072261810302734, 'timing/data': 0.045253499411046505, 'timing/model': 1.2677001571282744, 'learning_rate/qwen_vl_interface': 2.0062923337479223e-06, 'learning_rate/action_model': 2.006292333747922e-05, 'epoch': 1.9}) +[2026-06-18 04:00:45,911][starVLA.training.train_starvla][INFO] - Step 5202, Loss: {'train/loss': 0.4089902937412262, 'train/grad_norm_pre_clip': 26.388713836669922, 'timing/data': 0.0321190832182765, 'timing/model': 1.1566078159958124, 'learning_rate/qwen_vl_interface': 2.0047127702085437e-06, 'learning_rate/action_model': 2.0047127702085437e-05, 'epoch': 1.9}) +[2026-06-18 04:00:52,359][starVLA.training.train_starvla][INFO] - Step 5203, Loss: {'train/loss': 0.22905157506465912, 'train/grad_norm_pre_clip': 17.12073516845703, 'timing/data': 0.040941168554127216, 'timing/model': 1.4122616983950138, 'learning_rate/qwen_vl_interface': 2.003133879420698e-06, 'learning_rate/action_model': 2.003133879420698e-05, 'epoch': 1.91}) +[2026-06-18 04:00:56,677][starVLA.training.train_starvla][INFO] - Step 5204, Loss: {'train/loss': 0.18030551075935364, 'train/grad_norm_pre_clip': 19.14598274230957, 'timing/data': 0.036395334638655186, 'timing/model': 1.0670351078733802, 'learning_rate/qwen_vl_interface': 2.001555661711691e-06, 'learning_rate/action_model': 2.001555661711691e-05, 'epoch': 1.91}) +[2026-06-18 04:01:01,520][starVLA.training.train_starvla][INFO] - Step 5205, Loss: {'train/loss': 0.31738314032554626, 'train/grad_norm_pre_clip': 26.614530563354492, 'timing/data': 0.03175371140241623, 'timing/model': 1.060034823603928, 'learning_rate/qwen_vl_interface': 1.999978117408693e-06, 'learning_rate/action_model': 1.999978117408693e-05, 'epoch': 1.91}) +[2026-06-18 04:01:05,940][starVLA.training.train_starvla][INFO] - Step 5206, Loss: {'train/loss': 0.19283680617809296, 'train/grad_norm_pre_clip': 16.910167694091797, 'timing/data': 0.029382212087512016, 'timing/model': 1.0633745240047574, 'learning_rate/qwen_vl_interface': 1.9984012468387285e-06, 'learning_rate/action_model': 1.9984012468387285e-05, 'epoch': 1.91}) +[2026-06-18 04:01:11,548][starVLA.training.train_starvla][INFO] - Step 5207, Loss: {'train/loss': 0.44478875398635864, 'train/grad_norm_pre_clip': 31.779991149902344, 'timing/data': 0.03490637894719839, 'timing/model': 1.1553660547360778, 'learning_rate/qwen_vl_interface': 1.9968250503286857e-06, 'learning_rate/action_model': 1.996825050328686e-05, 'epoch': 1.91}) +[2026-06-18 04:01:15,623][starVLA.training.train_starvla][INFO] - Step 5208, Loss: {'train/loss': 0.2027602195739746, 'train/grad_norm_pre_clip': 14.706551551818848, 'timing/data': 0.03226195741444826, 'timing/model': 1.029363827779889, 'learning_rate/qwen_vl_interface': 1.9952495282053073e-06, 'learning_rate/action_model': 1.9952495282053073e-05, 'epoch': 1.91}) +[2026-06-18 04:01:20,165][starVLA.training.train_starvla][INFO] - Step 5209, Loss: {'train/loss': 0.08590810745954514, 'train/grad_norm_pre_clip': 8.634076118469238, 'timing/data': 0.037339397706091404, 'timing/model': 1.1229906426742673, 'learning_rate/qwen_vl_interface': 1.993674680795206e-06, 'learning_rate/action_model': 1.9936746807952058e-05, 'epoch': 1.91}) +[2026-06-18 04:01:29,246][starVLA.training.train_starvla][INFO] - Step 5210, Loss: {'train/loss': 0.4952162802219391, 'train/grad_norm_pre_clip': 26.7088680267334, 'timing/data': 0.033004905097186565, 'timing/model': 3.8478084579110146, 'learning_rate/qwen_vl_interface': 1.992100508424847e-06, 'learning_rate/action_model': 1.9921005084248467e-05, 'epoch': 1.91}) +[2026-06-18 04:01:34,067][starVLA.training.train_starvla][INFO] - Step 5211, Loss: {'train/loss': 0.18726854026317596, 'train/grad_norm_pre_clip': 13.589348793029785, 'timing/data': 0.03452088963240385, 'timing/model': 1.1797133041545749, 'learning_rate/qwen_vl_interface': 1.990527011420558e-06, 'learning_rate/action_model': 1.990527011420558e-05, 'epoch': 1.91}) +[2026-06-18 04:01:38,683][starVLA.training.train_starvla][INFO] - Step 5212, Loss: {'train/loss': 0.22730064392089844, 'train/grad_norm_pre_clip': 20.138776779174805, 'timing/data': 0.04082033410668373, 'timing/model': 1.4343629023060203, 'learning_rate/qwen_vl_interface': 1.9889541901085274e-06, 'learning_rate/action_model': 1.9889541901085273e-05, 'epoch': 1.91}) +[2026-06-18 04:01:46,841][starVLA.training.train_starvla][INFO] - Step 5213, Loss: {'train/loss': 0.16589757800102234, 'train/grad_norm_pre_clip': 15.471977233886719, 'timing/data': 0.03106229193508625, 'timing/model': 1.3396307490766048, 'learning_rate/qwen_vl_interface': 1.987382044814802e-06, 'learning_rate/action_model': 1.987382044814802e-05, 'epoch': 1.91}) +[2026-06-18 04:01:51,366][starVLA.training.train_starvla][INFO] - Step 5214, Loss: {'train/loss': 0.14272348582744598, 'train/grad_norm_pre_clip': 15.161980628967285, 'timing/data': 0.03186242934316397, 'timing/model': 1.3943235976621509, 'learning_rate/qwen_vl_interface': 1.9858105758652895e-06, 'learning_rate/action_model': 1.9858105758652892e-05, 'epoch': 1.91}) +[2026-06-18 04:01:57,472][starVLA.training.train_starvla][INFO] - Step 5215, Loss: {'train/loss': 0.38728100061416626, 'train/grad_norm_pre_clip': 14.543695449829102, 'timing/data': 0.03362344205379486, 'timing/model': 1.2161497762426734, 'learning_rate/qwen_vl_interface': 1.9842397835857555e-06, 'learning_rate/action_model': 1.9842397835857554e-05, 'epoch': 1.91}) +[2026-06-18 04:02:07,331][starVLA.training.train_starvla][INFO] - Step 5216, Loss: {'train/loss': 0.4737450182437897, 'train/grad_norm_pre_clip': 30.173786163330078, 'timing/data': 0.036223105154931545, 'timing/model': 2.8386862128973007, 'learning_rate/qwen_vl_interface': 1.982669668301831e-06, 'learning_rate/action_model': 1.982669668301831e-05, 'epoch': 1.91}) +[2026-06-18 04:02:17,658][starVLA.training.train_starvla][INFO] - Step 5217, Loss: {'train/loss': 0.1257053017616272, 'train/grad_norm_pre_clip': 11.384435653686523, 'timing/data': 0.03625272121280432, 'timing/model': 4.749314369633794, 'learning_rate/qwen_vl_interface': 1.9811002303389993e-06, 'learning_rate/action_model': 1.9811002303389993e-05, 'epoch': 1.91}) +[2026-06-18 04:02:22,246][starVLA.training.train_starvla][INFO] - Step 5218, Loss: {'train/loss': 0.24502845108509064, 'train/grad_norm_pre_clip': 23.735576629638672, 'timing/data': 0.03187466785311699, 'timing/model': 1.3354553952813148, 'learning_rate/qwen_vl_interface': 1.9795314700226068e-06, 'learning_rate/action_model': 1.9795314700226066e-05, 'epoch': 1.91}) +[2026-06-18 04:02:26,483][starVLA.training.train_starvla][INFO] - Step 5219, Loss: {'train/loss': 0.42491671442985535, 'train/grad_norm_pre_clip': 44.91144943237305, 'timing/data': 0.04116210713982582, 'timing/model': 1.1132161784917116, 'learning_rate/qwen_vl_interface': 1.9779633876778587e-06, 'learning_rate/action_model': 1.9779633876778586e-05, 'epoch': 1.91}) +[2026-06-18 04:02:31,924][starVLA.training.train_starvla][INFO] - Step 5220, Loss: {'train/loss': 0.0700654610991478, 'train/grad_norm_pre_clip': 8.852304458618164, 'timing/data': 0.02929203025996685, 'timing/model': 1.0713790152221918, 'learning_rate/qwen_vl_interface': 1.9763959836298225e-06, 'learning_rate/action_model': 1.9763959836298225e-05, 'epoch': 1.91}) +[2026-06-18 04:02:36,475][starVLA.training.train_starvla][INFO] - Step 5221, Loss: {'train/loss': 0.46178996562957764, 'train/grad_norm_pre_clip': 39.97187805175781, 'timing/data': 0.03647545725107193, 'timing/model': 1.1296761399134994, 'learning_rate/qwen_vl_interface': 1.9748292582034228e-06, 'learning_rate/action_model': 1.9748292582034228e-05, 'epoch': 1.91}) +[2026-06-18 04:02:41,140][starVLA.training.train_starvla][INFO] - Step 5222, Loss: {'train/loss': 0.4057738482952118, 'train/grad_norm_pre_clip': 32.413002014160156, 'timing/data': 0.03276610653847456, 'timing/model': 1.0518769472837448, 'learning_rate/qwen_vl_interface': 1.9732632117234405e-06, 'learning_rate/action_model': 1.9732632117234404e-05, 'epoch': 1.91}) +[2026-06-18 04:02:45,836][starVLA.training.train_starvla][INFO] - Step 5223, Loss: {'train/loss': 0.1387420892715454, 'train/grad_norm_pre_clip': 12.176562309265137, 'timing/data': 0.03501549921929836, 'timing/model': 1.670254748314619, 'learning_rate/qwen_vl_interface': 1.9716978445145195e-06, 'learning_rate/action_model': 1.9716978445145193e-05, 'epoch': 1.91}) +[2026-06-18 04:02:50,325][starVLA.training.train_starvla][INFO] - Step 5224, Loss: {'train/loss': 0.30550944805145264, 'train/grad_norm_pre_clip': 27.198612213134766, 'timing/data': 0.030273056589066982, 'timing/model': 1.0864337543025613, 'learning_rate/qwen_vl_interface': 1.9701331569011634e-06, 'learning_rate/action_model': 1.9701331569011633e-05, 'epoch': 1.91}) +[2026-06-18 04:02:55,609][starVLA.training.train_starvla][INFO] - Step 5225, Loss: {'train/loss': 0.22468681633472443, 'train/grad_norm_pre_clip': 18.560693740844727, 'timing/data': 0.03383049741387367, 'timing/model': 1.035268671810627, 'learning_rate/qwen_vl_interface': 1.968569149207734e-06, 'learning_rate/action_model': 1.968569149207734e-05, 'epoch': 1.91}) +[2026-06-18 04:03:00,423][starVLA.training.train_starvla][INFO] - Step 5226, Loss: {'train/loss': 0.21297134459018707, 'train/grad_norm_pre_clip': 21.074642181396484, 'timing/data': 0.0345770288258791, 'timing/model': 1.0890109539031982, 'learning_rate/qwen_vl_interface': 1.9670058217584505e-06, 'learning_rate/action_model': 1.9670058217584504e-05, 'epoch': 1.91}) +[2026-06-18 04:03:05,063][starVLA.training.train_starvla][INFO] - Step 5227, Loss: {'train/loss': 0.300654798746109, 'train/grad_norm_pre_clip': 24.62317657470703, 'timing/data': 0.03654692694544792, 'timing/model': 1.071597664617002, 'learning_rate/qwen_vl_interface': 1.965443174877393e-06, 'learning_rate/action_model': 1.965443174877393e-05, 'epoch': 1.91}) +[2026-06-18 04:03:09,631][starVLA.training.train_starvla][INFO] - Step 5228, Loss: {'train/loss': 0.34101366996765137, 'train/grad_norm_pre_clip': 25.20005989074707, 'timing/data': 0.029687967151403427, 'timing/model': 1.4075217079371214, 'learning_rate/qwen_vl_interface': 1.9638812088885003e-06, 'learning_rate/action_model': 1.9638812088885e-05, 'epoch': 1.91}) +[2026-06-18 04:03:15,610][starVLA.training.train_starvla][INFO] - Step 5229, Loss: {'train/loss': 0.4214343726634979, 'train/grad_norm_pre_clip': 31.9598388671875, 'timing/data': 0.032653434202075005, 'timing/model': 2.7907719910144806, 'learning_rate/qwen_vl_interface': 1.9623199241155684e-06, 'learning_rate/action_model': 1.9623199241155683e-05, 'epoch': 1.91}) +[2026-06-18 04:03:20,314][starVLA.training.train_starvla][INFO] - Step 5230, Loss: {'train/loss': 0.3518485426902771, 'train/grad_norm_pre_clip': 23.166086196899414, 'timing/data': 0.0499842893332243, 'timing/model': 1.059965131804347, 'learning_rate/qwen_vl_interface': 1.9607593208822536e-06, 'learning_rate/action_model': 1.9607593208822535e-05, 'epoch': 1.92}) +[2026-06-18 04:03:25,016][starVLA.training.train_starvla][INFO] - Step 5231, Loss: {'train/loss': 0.48951512575149536, 'train/grad_norm_pre_clip': 28.804832458496094, 'timing/data': 0.03503972664475441, 'timing/model': 1.1814415147528052, 'learning_rate/qwen_vl_interface': 1.9591993995120715e-06, 'learning_rate/action_model': 1.9591993995120712e-05, 'epoch': 1.92}) +[2026-06-18 04:03:29,627][starVLA.training.train_starvla][INFO] - Step 5232, Loss: {'train/loss': 0.2678104341030121, 'train/grad_norm_pre_clip': 22.533374786376953, 'timing/data': 0.029012931510806084, 'timing/model': 1.2046519415453076, 'learning_rate/qwen_vl_interface': 1.9576401603283944e-06, 'learning_rate/action_model': 1.9576401603283944e-05, 'epoch': 1.92}) +[2026-06-18 04:03:34,185][starVLA.training.train_starvla][INFO] - Step 5233, Loss: {'train/loss': 0.11640578508377075, 'train/grad_norm_pre_clip': 12.646565437316895, 'timing/data': 0.032447741366922855, 'timing/model': 1.4317680457606912, 'learning_rate/qwen_vl_interface': 1.956081603654455e-06, 'learning_rate/action_model': 1.956081603654455e-05, 'epoch': 1.92}) +[2026-06-18 04:03:38,443][starVLA.training.train_starvla][INFO] - Step 5234, Loss: {'train/loss': 0.059620264917612076, 'train/grad_norm_pre_clip': 7.259487152099609, 'timing/data': 0.029993411153554916, 'timing/model': 1.0602703988552094, 'learning_rate/qwen_vl_interface': 1.9545237298133423e-06, 'learning_rate/action_model': 1.9545237298133424e-05, 'epoch': 1.92}) +[2026-06-18 04:03:43,400][starVLA.training.train_starvla][INFO] - Step 5235, Loss: {'train/loss': 0.1719982922077179, 'train/grad_norm_pre_clip': 14.998490333557129, 'timing/data': 0.031893820501863956, 'timing/model': 1.0858652740716934, 'learning_rate/qwen_vl_interface': 1.9529665391280096e-06, 'learning_rate/action_model': 1.95296653912801e-05, 'epoch': 1.92}) +[2026-06-18 04:03:48,106][starVLA.training.train_starvla][INFO] - Step 5236, Loss: {'train/loss': 0.2729739248752594, 'train/grad_norm_pre_clip': 22.54192543029785, 'timing/data': 0.03487962391227484, 'timing/model': 1.1635664273053408, 'learning_rate/qwen_vl_interface': 1.9514100319212594e-06, 'learning_rate/action_model': 1.9514100319212593e-05, 'epoch': 1.92}) +[2026-06-18 04:03:59,023][starVLA.training.train_starvla][INFO] - Step 5237, Loss: {'train/loss': 0.3181678354740143, 'train/grad_norm_pre_clip': 26.59033203125, 'timing/data': 0.03353380411863327, 'timing/model': 1.1150797391310334, 'learning_rate/qwen_vl_interface': 1.949854208515757e-06, 'learning_rate/action_model': 1.9498542085157573e-05, 'epoch': 1.92}) +[2026-06-18 04:04:05,459][starVLA.training.train_starvla][INFO] - Step 5238, Loss: {'train/loss': 0.20399729907512665, 'train/grad_norm_pre_clip': 21.2060546875, 'timing/data': 0.035573940724134445, 'timing/model': 1.5382078662514687, 'learning_rate/qwen_vl_interface': 1.9482990692340305e-06, 'learning_rate/action_model': 1.9482990692340305e-05, 'epoch': 1.92}) +[2026-06-18 04:04:13,879][starVLA.training.train_starvla][INFO] - Step 5239, Loss: {'train/loss': 0.2266201674938202, 'train/grad_norm_pre_clip': 21.969242095947266, 'timing/data': 0.039114810526371, 'timing/model': 1.3954561557620764, 'learning_rate/qwen_vl_interface': 1.94674461439846e-06, 'learning_rate/action_model': 1.9467446143984598e-05, 'epoch': 1.92}) +[2026-06-18 04:04:27,328][starVLA.training.train_starvla][INFO] - Step 5240, Loss: {'train/loss': 0.12383538484573364, 'train/grad_norm_pre_clip': 17.23370361328125, 'timing/data': 0.03244779910892248, 'timing/model': 3.0400897646322846, 'learning_rate/qwen_vl_interface': 1.9451908443312863e-06, 'learning_rate/action_model': 1.9451908443312865e-05, 'epoch': 1.92}) +[2026-06-18 04:04:35,111][starVLA.training.train_starvla][INFO] - Step 5241, Loss: {'train/loss': 0.2089530974626541, 'train/grad_norm_pre_clip': 19.782180786132812, 'timing/data': 0.04378197621554136, 'timing/model': 2.9629600793123245, 'learning_rate/qwen_vl_interface': 1.943637759354604e-06, 'learning_rate/action_model': 1.9436377593546042e-05, 'epoch': 1.92}) +[2026-06-18 04:04:41,160][starVLA.training.train_starvla][INFO] - Step 5242, Loss: {'train/loss': 0.2480047196149826, 'train/grad_norm_pre_clip': 21.74033546447754, 'timing/data': 0.03083870094269514, 'timing/model': 1.1965162577107549, 'learning_rate/qwen_vl_interface': 1.942085359790373e-06, 'learning_rate/action_model': 1.942085359790373e-05, 'epoch': 1.92}) +[2026-06-18 04:04:51,112][starVLA.training.train_starvla][INFO] - Step 5243, Loss: {'train/loss': 0.2506505250930786, 'train/grad_norm_pre_clip': 21.546865463256836, 'timing/data': 0.04472538735717535, 'timing/model': 1.7926013329997659, 'learning_rate/qwen_vl_interface': 1.940533645960407e-06, 'learning_rate/action_model': 1.940533645960407e-05, 'epoch': 1.92}) +[2026-06-18 04:04:57,031][starVLA.training.train_starvla][INFO] - Step 5244, Loss: {'train/loss': 0.2605980336666107, 'train/grad_norm_pre_clip': 19.329734802246094, 'timing/data': 0.031563072465360165, 'timing/model': 1.650726567953825, 'learning_rate/qwen_vl_interface': 1.9389826181863767e-06, 'learning_rate/action_model': 1.9389826181863767e-05, 'epoch': 1.92}) +[2026-06-18 04:05:01,688][starVLA.training.train_starvla][INFO] - Step 5245, Loss: {'train/loss': 0.16361668705940247, 'train/grad_norm_pre_clip': 19.270139694213867, 'timing/data': 0.03446579352021217, 'timing/model': 1.0441004801541567, 'learning_rate/qwen_vl_interface': 1.937432276789813e-06, 'learning_rate/action_model': 1.9374322767898127e-05, 'epoch': 1.92}) +[2026-06-18 04:05:06,399][starVLA.training.train_starvla][INFO] - Step 5246, Loss: {'train/loss': 0.13362917304039001, 'train/grad_norm_pre_clip': 17.42351531982422, 'timing/data': 0.03968151845037937, 'timing/model': 1.0679029552266002, 'learning_rate/qwen_vl_interface': 1.935882622092103e-06, 'learning_rate/action_model': 1.935882622092103e-05, 'epoch': 1.92}) +[2026-06-18 04:05:10,931][starVLA.training.train_starvla][INFO] - Step 5247, Loss: {'train/loss': 0.6383758187294006, 'train/grad_norm_pre_clip': 35.29661178588867, 'timing/data': 0.038311611860990524, 'timing/model': 1.166854728013277, 'learning_rate/qwen_vl_interface': 1.9343336544144907e-06, 'learning_rate/action_model': 1.9343336544144906e-05, 'epoch': 1.92}) +[2026-06-18 04:05:15,575][starVLA.training.train_starvla][INFO] - Step 5248, Loss: {'train/loss': 0.5271493792533875, 'train/grad_norm_pre_clip': 45.31814193725586, 'timing/data': 0.03301863931119442, 'timing/model': 1.6260416312143207, 'learning_rate/qwen_vl_interface': 1.9327853740780787e-06, 'learning_rate/action_model': 1.932785374078079e-05, 'epoch': 1.92}) +[2026-06-18 04:05:19,841][starVLA.training.train_starvla][INFO] - Step 5249, Loss: {'train/loss': 0.24626000225543976, 'train/grad_norm_pre_clip': 20.4370174407959, 'timing/data': 0.0362952146679163, 'timing/model': 1.094199312850833, 'learning_rate/qwen_vl_interface': 1.9312377814038314e-06, 'learning_rate/action_model': 1.9312377814038312e-05, 'epoch': 1.92}) +[2026-06-18 04:05:24,369][starVLA.training.train_starvla][INFO] - Step 5250, Loss: {'train/loss': 0.21705660223960876, 'train/grad_norm_pre_clip': 20.753273010253906, 'timing/data': 0.03570743650197983, 'timing/model': 1.0427961573004723, 'learning_rate/qwen_vl_interface': 1.9296908767125612e-06, 'learning_rate/action_model': 1.929690876712561e-05, 'epoch': 1.92}) +[2026-06-18 04:05:29,037][starVLA.training.train_starvla][INFO] - Step 5251, Loss: {'train/loss': 0.42053020000457764, 'train/grad_norm_pre_clip': 26.011585235595703, 'timing/data': 0.03505552001297474, 'timing/model': 1.167679000645876, 'learning_rate/qwen_vl_interface': 1.928144660324945e-06, 'learning_rate/action_model': 1.928144660324945e-05, 'epoch': 1.92}) +[2026-06-18 04:05:33,640][starVLA.training.train_starvla][INFO] - Step 5252, Loss: {'train/loss': 0.3791963458061218, 'train/grad_norm_pre_clip': 21.4534969329834, 'timing/data': 0.032678691670298576, 'timing/model': 1.1623482089489698, 'learning_rate/qwen_vl_interface': 1.9265991325615135e-06, 'learning_rate/action_model': 1.9265991325615135e-05, 'epoch': 1.92}) +[2026-06-18 04:05:38,110][starVLA.training.train_starvla][INFO] - Step 5253, Loss: {'train/loss': 0.5266487002372742, 'train/grad_norm_pre_clip': 24.200592041015625, 'timing/data': 0.03113484475761652, 'timing/model': 1.3831243989989161, 'learning_rate/qwen_vl_interface': 1.9250542937426604e-06, 'learning_rate/action_model': 1.9250542937426602e-05, 'epoch': 1.92}) +[2026-06-18 04:05:42,532][starVLA.training.train_starvla][INFO] - Step 5254, Loss: {'train/loss': 0.6306617259979248, 'train/grad_norm_pre_clip': 29.94812774658203, 'timing/data': 0.03845396265387535, 'timing/model': 1.0606929007917643, 'learning_rate/qwen_vl_interface': 1.9235101441886307e-06, 'learning_rate/action_model': 1.923510144188631e-05, 'epoch': 1.92}) +[2026-06-18 04:05:47,126][starVLA.training.train_starvla][INFO] - Step 5255, Loss: {'train/loss': 0.32842567563056946, 'train/grad_norm_pre_clip': 28.653017044067383, 'timing/data': 0.032074655406177044, 'timing/model': 1.17169950902462, 'learning_rate/qwen_vl_interface': 1.921966684219525e-06, 'learning_rate/action_model': 1.921966684219525e-05, 'epoch': 1.92}) +[2026-06-18 04:05:51,783][starVLA.training.train_starvla][INFO] - Step 5256, Loss: {'train/loss': 0.13856159150600433, 'train/grad_norm_pre_clip': 19.5562744140625, 'timing/data': 0.05569877102971077, 'timing/model': 1.0434347735717893, 'learning_rate/qwen_vl_interface': 1.920423914155308e-06, 'learning_rate/action_model': 1.920423914155308e-05, 'epoch': 1.92}) +[2026-06-18 04:05:56,437][starVLA.training.train_starvla][INFO] - Step 5257, Loss: {'train/loss': 0.4029437005519867, 'train/grad_norm_pre_clip': 28.415494918823242, 'timing/data': 0.033967421390116215, 'timing/model': 1.071214433759451, 'learning_rate/qwen_vl_interface': 1.9188818343157962e-06, 'learning_rate/action_model': 1.9188818343157964e-05, 'epoch': 1.92}) +[2026-06-18 04:06:01,225][starVLA.training.train_starvla][INFO] - Step 5258, Loss: {'train/loss': 0.44899752736091614, 'train/grad_norm_pre_clip': 18.878175735473633, 'timing/data': 0.04136061016470194, 'timing/model': 1.372979698702693, 'learning_rate/qwen_vl_interface': 1.9173404450206654e-06, 'learning_rate/action_model': 1.9173404450206654e-05, 'epoch': 1.93}) +[2026-06-18 04:06:05,480][starVLA.training.train_starvla][INFO] - Step 5259, Loss: {'train/loss': 0.40982192754745483, 'train/grad_norm_pre_clip': 32.600914001464844, 'timing/data': 0.03181616310030222, 'timing/model': 1.1170159000903368, 'learning_rate/qwen_vl_interface': 1.9157997465894452e-06, 'learning_rate/action_model': 1.9157997465894454e-05, 'epoch': 1.93}) +[2026-06-18 04:06:10,188][starVLA.training.train_starvla][INFO] - Step 5260, Loss: {'train/loss': 0.4402078688144684, 'train/grad_norm_pre_clip': 26.898691177368164, 'timing/data': 0.03738040570169687, 'timing/model': 1.0648867655545473, 'learning_rate/qwen_vl_interface': 1.9142597393415264e-06, 'learning_rate/action_model': 1.914259739341526e-05, 'epoch': 1.93}) +[2026-06-18 04:06:16,010][starVLA.training.train_starvla][INFO] - Step 5261, Loss: {'train/loss': 0.42052093148231506, 'train/grad_norm_pre_clip': 25.909732818603516, 'timing/data': 0.03313836082816124, 'timing/model': 1.1924971928820014, 'learning_rate/qwen_vl_interface': 1.9127204235961524e-06, 'learning_rate/action_model': 1.9127204235961525e-05, 'epoch': 1.93}) +[2026-06-18 04:06:24,255][starVLA.training.train_starvla][INFO] - Step 5262, Loss: {'train/loss': 0.38465234637260437, 'train/grad_norm_pre_clip': 28.42333221435547, 'timing/data': 0.03412634413689375, 'timing/model': 1.0452143987640738, 'learning_rate/qwen_vl_interface': 1.911181799672426e-06, 'learning_rate/action_model': 1.9111817996724264e-05, 'epoch': 1.93}) +[2026-06-18 04:06:30,842][starVLA.training.train_starvla][INFO] - Step 5263, Loss: {'train/loss': 0.07757066190242767, 'train/grad_norm_pre_clip': 8.990850448608398, 'timing/data': 0.03720751404762268, 'timing/model': 1.6053525125607848, 'learning_rate/qwen_vl_interface': 1.909643867889305e-06, 'learning_rate/action_model': 1.909643867889305e-05, 'epoch': 1.93}) +[2026-06-18 04:06:35,707][starVLA.training.train_starvla][INFO] - Step 5264, Loss: {'train/loss': 0.16022638976573944, 'train/grad_norm_pre_clip': 12.635385513305664, 'timing/data': 0.03230876103043556, 'timing/model': 1.334963801316917, 'learning_rate/qwen_vl_interface': 1.908106628565605e-06, 'learning_rate/action_model': 1.9081066285656047e-05, 'epoch': 1.93}) +[2026-06-18 04:06:49,859][starVLA.training.train_starvla][INFO] - Step 5265, Loss: {'train/loss': 0.4683062732219696, 'train/grad_norm_pre_clip': 35.22015380859375, 'timing/data': 0.03506918251514435, 'timing/model': 4.198281733319163, 'learning_rate/qwen_vl_interface': 1.9065700820199965e-06, 'learning_rate/action_model': 1.9065700820199963e-05, 'epoch': 1.93}) +[2026-06-18 04:06:57,909][starVLA.training.train_starvla][INFO] - Step 5266, Loss: {'train/loss': 0.1534261703491211, 'train/grad_norm_pre_clip': 44.68633270263672, 'timing/data': 0.03357899188995361, 'timing/model': 1.3329006964340806, 'learning_rate/qwen_vl_interface': 1.9050342285710063e-06, 'learning_rate/action_model': 1.9050342285710062e-05, 'epoch': 1.93}) +[2026-06-18 04:07:05,177][starVLA.training.train_starvla][INFO] - Step 5267, Loss: {'train/loss': 0.06443275511264801, 'train/grad_norm_pre_clip': 9.653616905212402, 'timing/data': 0.03246652241796255, 'timing/model': 2.5979285975918174, 'learning_rate/qwen_vl_interface': 1.9034990685370217e-06, 'learning_rate/action_model': 1.9034990685370217e-05, 'epoch': 1.93}) +[2026-06-18 04:07:11,796][starVLA.training.train_starvla][INFO] - Step 5268, Loss: {'train/loss': 0.3223859667778015, 'train/grad_norm_pre_clip': 22.967222213745117, 'timing/data': 0.03074472490698099, 'timing/model': 2.0282254675403237, 'learning_rate/qwen_vl_interface': 1.901964602236282e-06, 'learning_rate/action_model': 1.901964602236282e-05, 'epoch': 1.93}) +[2026-06-18 04:07:16,437][starVLA.training.train_starvla][INFO] - Step 5269, Loss: {'train/loss': 0.44701412320137024, 'train/grad_norm_pre_clip': 20.890972137451172, 'timing/data': 0.031877828761935234, 'timing/model': 1.0520017640665174, 'learning_rate/qwen_vl_interface': 1.900430829986882e-06, 'learning_rate/action_model': 1.900430829986882e-05, 'epoch': 1.93}) +[2026-06-18 04:07:20,980][starVLA.training.train_starvla][INFO] - Step 5270, Loss: {'train/loss': 0.27847686409950256, 'train/grad_norm_pre_clip': 21.95099639892578, 'timing/data': 0.03406665287911892, 'timing/model': 1.0569143006578088, 'learning_rate/qwen_vl_interface': 1.898897752106773e-06, 'learning_rate/action_model': 1.898897752106773e-05, 'epoch': 1.93}) +[2026-06-18 04:07:25,629][starVLA.training.train_starvla][INFO] - Step 5271, Loss: {'train/loss': 0.1678190976381302, 'train/grad_norm_pre_clip': 14.98061466217041, 'timing/data': 0.03496258333325386, 'timing/model': 1.0705846846103668, 'learning_rate/qwen_vl_interface': 1.8973653689137683e-06, 'learning_rate/action_model': 1.8973653689137684e-05, 'epoch': 1.93}) +[2026-06-18 04:07:30,141][starVLA.training.train_starvla][INFO] - Step 5272, Loss: {'train/loss': 0.07103694975376129, 'train/grad_norm_pre_clip': 10.064790725708008, 'timing/data': 0.028129683807492256, 'timing/model': 1.3745792275294662, 'learning_rate/qwen_vl_interface': 1.8958336807255298e-06, 'learning_rate/action_model': 1.8958336807255297e-05, 'epoch': 1.93}) +[2026-06-18 04:07:34,193][starVLA.training.train_starvla][INFO] - Step 5273, Loss: {'train/loss': 0.2652660310268402, 'train/grad_norm_pre_clip': 22.10125160217285, 'timing/data': 0.051288765855133533, 'timing/model': 1.0381456073373556, 'learning_rate/qwen_vl_interface': 1.8943026878595785e-06, 'learning_rate/action_model': 1.8943026878595786e-05, 'epoch': 1.93}) +[2026-06-18 04:07:38,941][starVLA.training.train_starvla][INFO] - Step 5274, Loss: {'train/loss': 0.383553683757782, 'train/grad_norm_pre_clip': 31.71683692932129, 'timing/data': 0.03060877975076437, 'timing/model': 1.038728310726583, 'learning_rate/qwen_vl_interface': 1.8927723906332913e-06, 'learning_rate/action_model': 1.8927723906332912e-05, 'epoch': 1.93}) +[2026-06-18 04:07:43,421][starVLA.training.train_starvla][INFO] - Step 5275, Loss: {'train/loss': 0.531586229801178, 'train/grad_norm_pre_clip': 28.84670639038086, 'timing/data': 0.03663927502930164, 'timing/model': 1.0752399088814855, 'learning_rate/qwen_vl_interface': 1.8912427893638996e-06, 'learning_rate/action_model': 1.8912427893638996e-05, 'epoch': 1.93}) +[2026-06-18 04:07:47,902][starVLA.training.train_starvla][INFO] - Step 5276, Loss: {'train/loss': 0.11561484634876251, 'train/grad_norm_pre_clip': 14.418011665344238, 'timing/data': 0.046847847290337086, 'timing/model': 1.0636398121714592, 'learning_rate/qwen_vl_interface': 1.8897138843684918e-06, 'learning_rate/action_model': 1.8897138843684917e-05, 'epoch': 1.93}) +[2026-06-18 04:07:53,231][starVLA.training.train_starvla][INFO] - Step 5277, Loss: {'train/loss': 0.1690535992383957, 'train/grad_norm_pre_clip': 15.85986328125, 'timing/data': 0.04095934983342886, 'timing/model': 1.4491367135196924, 'learning_rate/qwen_vl_interface': 1.8881856759640122e-06, 'learning_rate/action_model': 1.8881856759640122e-05, 'epoch': 1.93}) +[2026-06-18 04:07:57,587][starVLA.training.train_starvla][INFO] - Step 5278, Loss: {'train/loss': 0.06576739996671677, 'train/grad_norm_pre_clip': 9.763428688049316, 'timing/data': 0.04069275315850973, 'timing/model': 1.3334505259990692, 'learning_rate/qwen_vl_interface': 1.8866581644672589e-06, 'learning_rate/action_model': 1.8866581644672586e-05, 'epoch': 1.93}) +[2026-06-18 04:08:02,011][starVLA.training.train_starvla][INFO] - Step 5279, Loss: {'train/loss': 0.3145790100097656, 'train/grad_norm_pre_clip': 18.124685287475586, 'timing/data': 0.03521724231541157, 'timing/model': 1.2355560781434178, 'learning_rate/qwen_vl_interface': 1.8851313501948872e-06, 'learning_rate/action_model': 1.885131350194887e-05, 'epoch': 1.93}) +[2026-06-18 04:08:06,592][starVLA.training.train_starvla][INFO] - Step 5280, Loss: {'train/loss': 0.6094413995742798, 'train/grad_norm_pre_clip': 40.46664810180664, 'timing/data': 0.04118895251303911, 'timing/model': 1.1470681661739945, 'learning_rate/qwen_vl_interface': 1.8836052334634072e-06, 'learning_rate/action_model': 1.883605233463407e-05, 'epoch': 1.93}) +[2026-06-18 04:08:12,316][starVLA.training.train_starvla][INFO] - Step 5281, Loss: {'train/loss': 0.4513118863105774, 'train/grad_norm_pre_clip': 20.234025955200195, 'timing/data': 0.03764528874307871, 'timing/model': 1.887628749012947, 'learning_rate/qwen_vl_interface': 1.8820798145891834e-06, 'learning_rate/action_model': 1.8820798145891832e-05, 'epoch': 1.93}) +[2026-06-18 04:08:22,007][starVLA.training.train_starvla][INFO] - Step 5282, Loss: {'train/loss': 0.315823495388031, 'train/grad_norm_pre_clip': 29.535932540893555, 'timing/data': 0.03715978376567364, 'timing/model': 1.078789317049086, 'learning_rate/qwen_vl_interface': 1.88055509388844e-06, 'learning_rate/action_model': 1.88055509388844e-05, 'epoch': 1.93}) +[2026-06-18 04:08:27,001][starVLA.training.train_starvla][INFO] - Step 5283, Loss: {'train/loss': 0.6471225619316101, 'train/grad_norm_pre_clip': 25.96651268005371, 'timing/data': 0.0361688444390893, 'timing/model': 1.4904845263808966, 'learning_rate/qwen_vl_interface': 1.8790310716772525e-06, 'learning_rate/action_model': 1.8790310716772524e-05, 'epoch': 1.93}) +[2026-06-18 04:08:34,899][starVLA.training.train_starvla][INFO] - Step 5284, Loss: {'train/loss': 0.5022143125534058, 'train/grad_norm_pre_clip': 28.95079803466797, 'timing/data': 0.037317726761102676, 'timing/model': 3.366500689648092, 'learning_rate/qwen_vl_interface': 1.8775077482715496e-06, 'learning_rate/action_model': 1.8775077482715494e-05, 'epoch': 1.93}) +[2026-06-18 04:08:42,401][starVLA.training.train_starvla][INFO] - Step 5285, Loss: {'train/loss': 0.2760273218154907, 'train/grad_norm_pre_clip': 30.04035186767578, 'timing/data': 0.014595375396311283, 'timing/model': 1.048446542583406, 'learning_rate/qwen_vl_interface': 1.8759851239871176e-06, 'learning_rate/action_model': 1.8759851239871175e-05, 'epoch': 1.94}) +[2026-06-18 04:08:47,264][starVLA.training.train_starvla][INFO] - Step 5286, Loss: {'train/loss': 0.29135558009147644, 'train/grad_norm_pre_clip': 26.620695114135742, 'timing/data': 0.02285081334412098, 'timing/model': 1.2245184928178787, 'learning_rate/qwen_vl_interface': 1.8744631991396023e-06, 'learning_rate/action_model': 1.8744631991396023e-05, 'epoch': 1.94}) +[2026-06-18 04:08:51,689][starVLA.training.train_starvla][INFO] - Step 5287, Loss: {'train/loss': 0.37281104922294617, 'train/grad_norm_pre_clip': 22.038124084472656, 'timing/data': 0.03713513631373644, 'timing/model': 1.0501110330224037, 'learning_rate/qwen_vl_interface': 1.8729419740444977e-06, 'learning_rate/action_model': 1.8729419740444977e-05, 'epoch': 1.94}) +[2026-06-18 04:08:56,627][starVLA.training.train_starvla][INFO] - Step 5288, Loss: {'train/loss': 0.4576866626739502, 'train/grad_norm_pre_clip': 29.148948669433594, 'timing/data': 0.022621457464993, 'timing/model': 1.4519910840317607, 'learning_rate/qwen_vl_interface': 1.8714214490171554e-06, 'learning_rate/action_model': 1.8714214490171552e-05, 'epoch': 1.94}) +[2026-06-18 04:09:03,659][starVLA.training.train_starvla][INFO] - Step 5289, Loss: {'train/loss': 0.24129292368888855, 'train/grad_norm_pre_clip': 19.1801700592041, 'timing/data': 0.03283730149269104, 'timing/model': 3.9135830402374268, 'learning_rate/qwen_vl_interface': 1.8699016243727826e-06, 'learning_rate/action_model': 1.8699016243727827e-05, 'epoch': 1.94}) +[2026-06-18 04:09:08,505][starVLA.training.train_starvla][INFO] - Step 5290, Loss: {'train/loss': 0.17190037667751312, 'train/grad_norm_pre_clip': 12.268876075744629, 'timing/data': 0.02840660698711872, 'timing/model': 1.0534205697476864, 'learning_rate/qwen_vl_interface': 1.8683825004264402e-06, 'learning_rate/action_model': 1.86838250042644e-05, 'epoch': 1.94}) +[2026-06-18 04:09:13,989][starVLA.training.train_starvla][INFO] - Step 5291, Loss: {'train/loss': 0.5148060321807861, 'train/grad_norm_pre_clip': 28.739748001098633, 'timing/data': 0.11412403266876936, 'timing/model': 1.6727919466793537, 'learning_rate/qwen_vl_interface': 1.8668640774930444e-06, 'learning_rate/action_model': 1.8668640774930442e-05, 'epoch': 1.94}) +[2026-06-18 04:09:19,061][starVLA.training.train_starvla][INFO] - Step 5292, Loss: {'train/loss': 0.32618704438209534, 'train/grad_norm_pre_clip': 22.567272186279297, 'timing/data': 0.01640948746353388, 'timing/model': 1.0464265998452902, 'learning_rate/qwen_vl_interface': 1.8653463558873651e-06, 'learning_rate/action_model': 1.8653463558873652e-05, 'epoch': 1.94}) +[2026-06-18 04:09:23,517][starVLA.training.train_starvla][INFO] - Step 5293, Loss: {'train/loss': 0.4885994493961334, 'train/grad_norm_pre_clip': 31.10816764831543, 'timing/data': 0.04644401930272579, 'timing/model': 1.4154298743233085, 'learning_rate/qwen_vl_interface': 1.863829335924029e-06, 'learning_rate/action_model': 1.8638293359240288e-05, 'epoch': 1.94}) +[2026-06-18 04:09:27,923][starVLA.training.train_starvla][INFO] - Step 5294, Loss: {'train/loss': 0.29795557260513306, 'train/grad_norm_pre_clip': 23.395360946655273, 'timing/data': 0.04543277155607939, 'timing/model': 1.3424326376989484, 'learning_rate/qwen_vl_interface': 1.8623130179175148e-06, 'learning_rate/action_model': 1.8623130179175146e-05, 'epoch': 1.94}) +[2026-06-18 04:09:32,468][starVLA.training.train_starvla][INFO] - Step 5295, Loss: {'train/loss': 0.36404725909233093, 'train/grad_norm_pre_clip': 20.265535354614258, 'timing/data': 0.04860469698905945, 'timing/model': 1.1240628138184547, 'learning_rate/qwen_vl_interface': 1.8607974021821573e-06, 'learning_rate/action_model': 1.8607974021821572e-05, 'epoch': 1.94}) +[2026-06-18 04:09:37,769][starVLA.training.train_starvla][INFO] - Step 5296, Loss: {'train/loss': 0.09916029870510101, 'train/grad_norm_pre_clip': 13.145881652832031, 'timing/data': 0.050395491532981396, 'timing/model': 1.1413388783112168, 'learning_rate/qwen_vl_interface': 1.8592824890321443e-06, 'learning_rate/action_model': 1.8592824890321442e-05, 'epoch': 1.94}) +[2026-06-18 04:09:46,759][starVLA.training.train_starvla][INFO] - Step 5297, Loss: {'train/loss': 0.12015651911497116, 'train/grad_norm_pre_clip': 18.873083114624023, 'timing/data': 0.05273760203272104, 'timing/model': 4.458141190931201, 'learning_rate/qwen_vl_interface': 1.8577682787815223e-06, 'learning_rate/action_model': 1.8577682787815225e-05, 'epoch': 1.94}) +[2026-06-18 04:09:56,247][starVLA.training.train_starvla][INFO] - Step 5298, Loss: {'train/loss': 0.3152516484260559, 'train/grad_norm_pre_clip': 25.189359664916992, 'timing/data': 0.050427685491740704, 'timing/model': 1.7433895925059915, 'learning_rate/qwen_vl_interface': 1.856254771744185e-06, 'learning_rate/action_model': 1.856254771744185e-05, 'epoch': 1.94}) +[2026-06-18 04:10:00,499][starVLA.training.train_starvla][INFO] - Step 5299, Loss: {'train/loss': 0.06089464947581291, 'train/grad_norm_pre_clip': 8.099112510681152, 'timing/data': 0.04615804087370634, 'timing/model': 1.0390155455097556, 'learning_rate/qwen_vl_interface': 1.8547419682338838e-06, 'learning_rate/action_model': 1.8547419682338836e-05, 'epoch': 1.94}) +[2026-06-18 04:13:27,139][starVLA.training.train_starvla][INFO] - Step 5300, Loss: {'train/loss': 0.4176467955112457, 'train/grad_norm_pre_clip': 30.71768569946289, 'eval/loss': 0.5401760339736938, 'eval/latency_2/loss': 0.5401760339736938, 'eval/demon_fire_precision': 0.9493355481727574, 'eval/demon_fire_recall': 0.9564853556485355, 'eval/demon_fire_f1': 0.9528970404335139, 'eval/demon_fire_rate_model': 0.6700055648302726, 'eval/demon_fire_rate_teacher': 0.6649972175848636, 'eval/demon_left_precision': 0.8967674661105318, 'eval/demon_left_recall': 0.9052631578947369, 'eval/demon_left_f1': 0.9009952854897852, 'eval/demon_left_rate_model': 0.266833611574847, 'eval/demon_left_rate_teacher': 0.2643294379521425, 'eval/demon_right_precision': 0.8725910064239829, 'eval/demon_right_recall': 0.9209039548022598, 'eval/demon_right_f1': 0.8960967564595932, 'eval/demon_right_rate_model': 0.2598775737340011, 'eval/demon_right_rate_teacher': 0.24624373956594323, 'eval/demon_move_f1': 0.8985460209746892, 'eval/demon_cc_f1': 0.9359136495581954, 'eval/latency_2/demon_fire_precision': 0.9493355481727574, 'eval/latency_2/demon_fire_recall': 0.9564853556485355, 'eval/latency_2/demon_fire_f1': 0.9528970404335139, 'eval/latency_2/demon_fire_rate_model': 0.6700055648302726, 'eval/latency_2/demon_fire_rate_teacher': 0.6649972175848636, 'eval/latency_2/demon_left_precision': 0.8967674661105318, 'eval/latency_2/demon_left_recall': 0.9052631578947369, 'eval/latency_2/demon_left_f1': 0.9009952854897852, 'eval/latency_2/demon_left_rate_model': 0.266833611574847, 'eval/latency_2/demon_left_rate_teacher': 0.2643294379521425, 'eval/latency_2/demon_right_precision': 0.8725910064239829, 'eval/latency_2/demon_right_recall': 0.9209039548022598, 'eval/latency_2/demon_right_f1': 0.8960967564595932, 'eval/latency_2/demon_right_rate_model': 0.2598775737340011, 'eval/latency_2/demon_right_rate_teacher': 0.24624373956594323, 'eval/latency_2/demon_move_f1': 0.8985460209746892, 'eval/latency_2/demon_cc_f1': 0.9359136495581954, 'timing/data': 0.03948611579835415, 'timing/model': 1.0608221450820565, 'learning_rate/qwen_vl_interface': 1.8532298685642275e-06, 'learning_rate/action_model': 1.8532298685642274e-05, 'epoch': 1.94}) +[2026-06-18 04:13:27,164][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_5300_state +[2026-06-18 04:13:27,172][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 04:13:35,896][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_5300_state/model.safetensors +[2026-06-18 04:13:50,862][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_5300_state/optimizer.bin +[2026-06-18 04:13:50,864][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_5300_state/sampler.bin +[2026-06-18 04:13:50,864][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_5300_state/sampler_1.bin +[2026-06-18 04:13:50,867][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_5300_state/random_states_0.pkl +[2026-06-18 04:13:50,874][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 5300 +[2026-06-18 04:13:50,875][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 04:13:50,891][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 04:13:56,333][starVLA.training.train_starvla][INFO] - Step 5301, Loss: {'train/loss': 0.15465661883354187, 'train/grad_norm_pre_clip': 13.414917945861816, 'timing/data': 0.04647800698876381, 'timing/model': 1.1311436612159014, 'learning_rate/qwen_vl_interface': 1.8517184730486745e-06, 'learning_rate/action_model': 1.8517184730486746e-05, 'epoch': 1.94}) +[2026-06-18 04:14:00,941][starVLA.training.train_starvla][INFO] - Step 5302, Loss: {'train/loss': 0.1539081633090973, 'train/grad_norm_pre_clip': 25.803442001342773, 'timing/data': 0.03840945288538933, 'timing/model': 1.114647845737636, 'learning_rate/qwen_vl_interface': 1.8502077820005408e-06, 'learning_rate/action_model': 1.8502077820005407e-05, 'epoch': 1.94}) +[2026-06-18 04:14:05,515][starVLA.training.train_starvla][INFO] - Step 5303, Loss: {'train/loss': 0.5890721082687378, 'train/grad_norm_pre_clip': 40.65834426879883, 'timing/data': 0.04143090732395649, 'timing/model': 1.0623509837314487, 'learning_rate/qwen_vl_interface': 1.8486977957329888e-06, 'learning_rate/action_model': 1.8486977957329886e-05, 'epoch': 1.94}) +[2026-06-18 04:14:10,008][starVLA.training.train_starvla][INFO] - Step 5304, Loss: {'train/loss': 0.20276668667793274, 'train/grad_norm_pre_clip': 25.069143295288086, 'timing/data': 0.05103165656328201, 'timing/model': 1.4126620311290026, 'learning_rate/qwen_vl_interface': 1.847188514559045e-06, 'learning_rate/action_model': 1.847188514559045e-05, 'epoch': 1.94}) +[2026-06-18 04:14:14,089][starVLA.training.train_starvla][INFO] - Step 5305, Loss: {'train/loss': 0.28763139247894287, 'train/grad_norm_pre_clip': 23.131587982177734, 'timing/data': 0.048633867874741554, 'timing/model': 1.0634608324617147, 'learning_rate/qwen_vl_interface': 1.8456799387915839e-06, 'learning_rate/action_model': 1.8456799387915838e-05, 'epoch': 1.94}) +[2026-06-18 04:14:18,487][starVLA.training.train_starvla][INFO] - Step 5306, Loss: {'train/loss': 0.3782918155193329, 'train/grad_norm_pre_clip': 27.160749435424805, 'timing/data': 0.03858174569904804, 'timing/model': 1.0380426673218608, 'learning_rate/qwen_vl_interface': 1.8441720687433345e-06, 'learning_rate/action_model': 1.8441720687433344e-05, 'epoch': 1.94}) +[2026-06-18 04:14:30,381][starVLA.training.train_starvla][INFO] - Step 5307, Loss: {'train/loss': 0.3219734728336334, 'train/grad_norm_pre_clip': 23.5225772857666, 'timing/data': 0.04567293170839548, 'timing/model': 3.8219855688512325, 'learning_rate/qwen_vl_interface': 1.8426649047268808e-06, 'learning_rate/action_model': 1.8426649047268807e-05, 'epoch': 1.94}) +[2026-06-18 04:14:37,118][starVLA.training.train_starvla][INFO] - Step 5308, Loss: {'train/loss': 0.26562654972076416, 'train/grad_norm_pre_clip': 17.250179290771484, 'timing/data': 0.042575216852128506, 'timing/model': 1.0414492972195148, 'learning_rate/qwen_vl_interface': 1.8411584470546582e-06, 'learning_rate/action_model': 1.841158447054658e-05, 'epoch': 1.94}) +[2026-06-18 04:14:41,732][starVLA.training.train_starvla][INFO] - Step 5309, Loss: {'train/loss': 0.6651926636695862, 'train/grad_norm_pre_clip': 41.64170837402344, 'timing/data': 0.042999573051929474, 'timing/model': 1.5557068977504969, 'learning_rate/qwen_vl_interface': 1.8396526960389575e-06, 'learning_rate/action_model': 1.8396526960389575e-05, 'epoch': 1.94}) +[2026-06-18 04:14:45,832][starVLA.training.train_starvla][INFO] - Step 5310, Loss: {'train/loss': 0.20419320464134216, 'train/grad_norm_pre_clip': 22.54051971435547, 'timing/data': 0.037206138484179974, 'timing/model': 1.0786712476983666, 'learning_rate/qwen_vl_interface': 1.8381476519919215e-06, 'learning_rate/action_model': 1.8381476519919212e-05, 'epoch': 1.94}) +[2026-06-18 04:14:50,330][starVLA.training.train_starvla][INFO] - Step 5311, Loss: {'train/loss': 0.2324189692735672, 'train/grad_norm_pre_clip': 14.716876983642578, 'timing/data': 0.041243283078074455, 'timing/model': 1.061210885643959, 'learning_rate/qwen_vl_interface': 1.8366433152255511e-06, 'learning_rate/action_model': 1.8366433152255512e-05, 'epoch': 1.94}) +[2026-06-18 04:14:54,932][starVLA.training.train_starvla][INFO] - Step 5312, Loss: {'train/loss': 0.21042901277542114, 'train/grad_norm_pre_clip': 16.169696807861328, 'timing/data': 0.04153016488999128, 'timing/model': 1.126686998642981, 'learning_rate/qwen_vl_interface': 1.8351396860516931e-06, 'learning_rate/action_model': 1.835139686051693e-05, 'epoch': 1.95}) +[2026-06-18 04:14:59,552][starVLA.training.train_starvla][INFO] - Step 5313, Loss: {'train/loss': 0.3483837842941284, 'train/grad_norm_pre_clip': 16.958744049072266, 'timing/data': 0.04438268393278122, 'timing/model': 1.0872065713629127, 'learning_rate/qwen_vl_interface': 1.8336367647820527e-06, 'learning_rate/action_model': 1.8336367647820528e-05, 'epoch': 1.95}) +[2026-06-18 04:15:04,152][starVLA.training.train_starvla][INFO] - Step 5314, Loss: {'train/loss': 0.2915250360965729, 'train/grad_norm_pre_clip': 40.36344528198242, 'timing/data': 0.04048312082886696, 'timing/model': 1.4011903330683708, 'learning_rate/qwen_vl_interface': 1.832134551728186e-06, 'learning_rate/action_model': 1.832134551728186e-05, 'epoch': 1.95}) +[2026-06-18 04:15:09,328][starVLA.training.train_starvla][INFO] - Step 5315, Loss: {'train/loss': 0.21933935582637787, 'train/grad_norm_pre_clip': 33.48765182495117, 'timing/data': 0.04445941653102636, 'timing/model': 1.1468552695587277, 'learning_rate/qwen_vl_interface': 1.8306330472015051e-06, 'learning_rate/action_model': 1.8306330472015053e-05, 'epoch': 1.95}) +[2026-06-18 04:15:14,666][starVLA.training.train_starvla][INFO] - Step 5316, Loss: {'train/loss': 0.2718420922756195, 'train/grad_norm_pre_clip': 19.038188934326172, 'timing/data': 0.02948853839188814, 'timing/model': 1.077604048885405, 'learning_rate/qwen_vl_interface': 1.8291322515132757e-06, 'learning_rate/action_model': 1.8291322515132756e-05, 'epoch': 1.95}) +[2026-06-18 04:15:19,208][starVLA.training.train_starvla][INFO] - Step 5317, Loss: {'train/loss': 0.21301540732383728, 'train/grad_norm_pre_clip': 16.70259666442871, 'timing/data': 0.03209766186773777, 'timing/model': 1.0585808465257287, 'learning_rate/qwen_vl_interface': 1.8276321649746076e-06, 'learning_rate/action_model': 1.8276321649746076e-05, 'epoch': 1.95}) +[2026-06-18 04:15:23,795][starVLA.training.train_starvla][INFO] - Step 5318, Loss: {'train/loss': 0.6352910399436951, 'train/grad_norm_pre_clip': 37.45551300048828, 'timing/data': 0.036394111812114716, 'timing/model': 1.0790350995957851, 'learning_rate/qwen_vl_interface': 1.826132787896476e-06, 'learning_rate/action_model': 1.826132787896476e-05, 'epoch': 1.95}) +[2026-06-18 04:15:28,422][starVLA.training.train_starvla][INFO] - Step 5319, Loss: {'train/loss': 0.23763516545295715, 'train/grad_norm_pre_clip': 23.621631622314453, 'timing/data': 0.03808958921581507, 'timing/model': 1.375491799786687, 'learning_rate/qwen_vl_interface': 1.8246341205897018e-06, 'learning_rate/action_model': 1.8246341205897017e-05, 'epoch': 1.95}) +[2026-06-18 04:15:32,522][starVLA.training.train_starvla][INFO] - Step 5320, Loss: {'train/loss': 0.26217886805534363, 'train/grad_norm_pre_clip': 14.690699577331543, 'timing/data': 0.03392762877047062, 'timing/model': 1.0487692905589938, 'learning_rate/qwen_vl_interface': 1.8231361633649606e-06, 'learning_rate/action_model': 1.8231361633649606e-05, 'epoch': 1.95}) +[2026-06-18 04:15:37,146][starVLA.training.train_starvla][INFO] - Step 5321, Loss: {'train/loss': 0.3927563726902008, 'train/grad_norm_pre_clip': 32.86198806762695, 'timing/data': 0.04059651866555214, 'timing/model': 1.052584683522582, 'learning_rate/qwen_vl_interface': 1.8216389165327797e-06, 'learning_rate/action_model': 1.8216389165327798e-05, 'epoch': 1.95}) +[2026-06-18 04:15:41,741][starVLA.training.train_starvla][INFO] - Step 5322, Loss: {'train/loss': 0.27691447734832764, 'train/grad_norm_pre_clip': 25.80325698852539, 'timing/data': 0.03663707710802555, 'timing/model': 1.1400886541232467, 'learning_rate/qwen_vl_interface': 1.8201423804035406e-06, 'learning_rate/action_model': 1.8201423804035405e-05, 'epoch': 1.95}) +[2026-06-18 04:15:46,256][starVLA.training.train_starvla][INFO] - Step 5323, Loss: {'train/loss': 0.29431065917015076, 'train/grad_norm_pre_clip': 20.805566787719727, 'timing/data': 0.03570429887622595, 'timing/model': 1.1289665177464485, 'learning_rate/qwen_vl_interface': 1.8186465552874769e-06, 'learning_rate/action_model': 1.8186465552874766e-05, 'epoch': 1.95}) +[2026-06-18 04:15:50,432][starVLA.training.train_starvla][INFO] - Step 5324, Loss: {'train/loss': 0.16121935844421387, 'train/grad_norm_pre_clip': 14.302316665649414, 'timing/data': 0.02844576444476843, 'timing/model': 1.0543544553220272, 'learning_rate/qwen_vl_interface': 1.8171514414946745e-06, 'learning_rate/action_model': 1.8171514414946745e-05, 'epoch': 1.95}) +[2026-06-18 04:15:54,897][starVLA.training.train_starvla][INFO] - Step 5325, Loss: {'train/loss': 0.1059747189283371, 'train/grad_norm_pre_clip': 15.536410331726074, 'timing/data': 0.04625777527689934, 'timing/model': 1.0490652099251747, 'learning_rate/qwen_vl_interface': 1.815657039335072e-06, 'learning_rate/action_model': 1.815657039335072e-05, 'epoch': 1.95}) +[2026-06-18 04:16:00,389][starVLA.training.train_starvla][INFO] - Step 5326, Loss: {'train/loss': 0.5925999283790588, 'train/grad_norm_pre_clip': 28.245086669921875, 'timing/data': 0.029707515612244606, 'timing/model': 1.0305166067555547, 'learning_rate/qwen_vl_interface': 1.8141633491184612e-06, 'learning_rate/action_model': 1.814163349118461e-05, 'epoch': 1.95}) +[2026-06-18 04:16:05,010][starVLA.training.train_starvla][INFO] - Step 5327, Loss: {'train/loss': 0.26548248529434204, 'train/grad_norm_pre_clip': 31.24433135986328, 'timing/data': 0.0325977923348546, 'timing/model': 1.0463739736005664, 'learning_rate/qwen_vl_interface': 1.8126703711544848e-06, 'learning_rate/action_model': 1.812670371154485e-05, 'epoch': 1.95}) +[2026-06-18 04:16:09,598][starVLA.training.train_starvla][INFO] - Step 5328, Loss: {'train/loss': 0.41775813698768616, 'train/grad_norm_pre_clip': 30.96343421936035, 'timing/data': 0.029528705403208733, 'timing/model': 1.40765908267349, 'learning_rate/qwen_vl_interface': 1.8111781057526381e-06, 'learning_rate/action_model': 1.811178105752638e-05, 'epoch': 1.95}) +[2026-06-18 04:16:14,352][starVLA.training.train_starvla][INFO] - Step 5329, Loss: {'train/loss': 0.30585435032844543, 'train/grad_norm_pre_clip': 31.326711654663086, 'timing/data': 0.03487628232687712, 'timing/model': 1.5728862704709172, 'learning_rate/qwen_vl_interface': 1.8096865532222714e-06, 'learning_rate/action_model': 1.8096865532222714e-05, 'epoch': 1.95}) +[2026-06-18 04:16:18,420][starVLA.training.train_starvla][INFO] - Step 5330, Loss: {'train/loss': 0.3654882609844208, 'train/grad_norm_pre_clip': 28.28973388671875, 'timing/data': 0.02777399681508541, 'timing/model': 1.0613915687426925, 'learning_rate/qwen_vl_interface': 1.8081957138725861e-06, 'learning_rate/action_model': 1.808195713872586e-05, 'epoch': 1.95}) +[2026-06-18 04:16:22,801][starVLA.training.train_starvla][INFO] - Step 5331, Loss: {'train/loss': 0.3238639235496521, 'train/grad_norm_pre_clip': 25.85237693786621, 'timing/data': 0.038438986986875534, 'timing/model': 1.0468720756471157, 'learning_rate/qwen_vl_interface': 1.8067055880126311e-06, 'learning_rate/action_model': 1.806705588012631e-05, 'epoch': 1.95}) +[2026-06-18 04:16:27,499][starVLA.training.train_starvla][INFO] - Step 5332, Loss: {'train/loss': 0.18239903450012207, 'train/grad_norm_pre_clip': 15.417539596557617, 'timing/data': 0.029546626843512058, 'timing/model': 1.0816441047936678, 'learning_rate/qwen_vl_interface': 1.8052161759513114e-06, 'learning_rate/action_model': 1.8052161759513115e-05, 'epoch': 1.95}) +[2026-06-18 04:16:32,186][starVLA.training.train_starvla][INFO] - Step 5333, Loss: {'train/loss': 0.20400576293468475, 'train/grad_norm_pre_clip': 23.433738708496094, 'timing/data': 0.029922119341790676, 'timing/model': 1.0864803567528725, 'learning_rate/qwen_vl_interface': 1.8037274779973868e-06, 'learning_rate/action_model': 1.8037274779973867e-05, 'epoch': 1.95}) +[2026-06-18 04:16:36,511][starVLA.training.train_starvla][INFO] - Step 5334, Loss: {'train/loss': 0.29668018221855164, 'train/grad_norm_pre_clip': 22.388633728027344, 'timing/data': 0.029212691821157932, 'timing/model': 1.3298959946259856, 'learning_rate/qwen_vl_interface': 1.8022394944594643e-06, 'learning_rate/action_model': 1.802239494459464e-05, 'epoch': 1.95}) +[2026-06-18 04:16:40,938][starVLA.training.train_starvla][INFO] - Step 5335, Loss: {'train/loss': 0.33916687965393066, 'train/grad_norm_pre_clip': 30.871706008911133, 'timing/data': 0.03214958030730486, 'timing/model': 1.1963502801954746, 'learning_rate/qwen_vl_interface': 1.8007522256460062e-06, 'learning_rate/action_model': 1.8007522256460062e-05, 'epoch': 1.95}) +[2026-06-18 04:16:45,606][starVLA.training.train_starvla][INFO] - Step 5336, Loss: {'train/loss': 0.30222225189208984, 'train/grad_norm_pre_clip': 29.36174964904785, 'timing/data': 0.02789492066949606, 'timing/model': 1.0973758352920413, 'learning_rate/qwen_vl_interface': 1.7992656718653201e-06, 'learning_rate/action_model': 1.79926567186532e-05, 'epoch': 1.95}) +[2026-06-18 04:16:50,140][starVLA.training.train_starvla][INFO] - Step 5337, Loss: {'train/loss': 0.5578614473342896, 'train/grad_norm_pre_clip': 24.73188018798828, 'timing/data': 0.04179361090064049, 'timing/model': 1.0523626897484064, 'learning_rate/qwen_vl_interface': 1.797779833425575e-06, 'learning_rate/action_model': 1.797779833425575e-05, 'epoch': 1.95}) +[2026-06-18 04:16:54,631][starVLA.training.train_starvla][INFO] - Step 5338, Loss: {'train/loss': 0.2730262577533722, 'train/grad_norm_pre_clip': 18.449724197387695, 'timing/data': 0.029668263159692287, 'timing/model': 1.1704685613512993, 'learning_rate/qwen_vl_interface': 1.7962947106347846e-06, 'learning_rate/action_model': 1.7962947106347845e-05, 'epoch': 1.95}) +[2026-06-18 04:16:59,208][starVLA.training.train_starvla][INFO] - Step 5339, Loss: {'train/loss': 0.2289731353521347, 'train/grad_norm_pre_clip': 16.226781845092773, 'timing/data': 0.030947241932153702, 'timing/model': 1.4060818972066045, 'learning_rate/qwen_vl_interface': 1.794810303800818e-06, 'learning_rate/action_model': 1.794810303800818e-05, 'epoch': 1.96}) +[2026-06-18 04:17:03,410][starVLA.training.train_starvla][INFO] - Step 5340, Loss: {'train/loss': 0.513007640838623, 'train/grad_norm_pre_clip': 24.179582595825195, 'timing/data': 0.030093406327068806, 'timing/model': 1.0272407196462154, 'learning_rate/qwen_vl_interface': 1.7933266132313924e-06, 'learning_rate/action_model': 1.7933266132313923e-05, 'epoch': 1.96}) +[2026-06-18 04:17:08,067][starVLA.training.train_starvla][INFO] - Step 5341, Loss: {'train/loss': 0.2962264120578766, 'train/grad_norm_pre_clip': 23.174850463867188, 'timing/data': 0.03318447060883045, 'timing/model': 1.0450711958110332, 'learning_rate/qwen_vl_interface': 1.7918436392340803e-06, 'learning_rate/action_model': 1.7918436392340802e-05, 'epoch': 1.96}) +[2026-06-18 04:17:12,584][starVLA.training.train_starvla][INFO] - Step 5342, Loss: {'train/loss': 0.46568062901496887, 'train/grad_norm_pre_clip': 24.525863647460938, 'timing/data': 0.02915147226303816, 'timing/model': 1.0509719559922814, 'learning_rate/qwen_vl_interface': 1.7903613821163022e-06, 'learning_rate/action_model': 1.790361382116302e-05, 'epoch': 1.96}) +[2026-06-18 04:17:17,156][starVLA.training.train_starvla][INFO] - Step 5343, Loss: {'train/loss': 0.2506610155105591, 'train/grad_norm_pre_clip': 29.320878982543945, 'timing/data': 0.030355030670762062, 'timing/model': 1.4020870393142104, 'learning_rate/qwen_vl_interface': 1.7888798421853313e-06, 'learning_rate/action_model': 1.7888798421853313e-05, 'epoch': 1.96}) +[2026-06-18 04:17:21,597][starVLA.training.train_starvla][INFO] - Step 5344, Loss: {'train/loss': 0.3899974822998047, 'train/grad_norm_pre_clip': 21.015783309936523, 'timing/data': 0.030858236365020275, 'timing/model': 1.3577932445332408, 'learning_rate/qwen_vl_interface': 1.7873990197482962e-06, 'learning_rate/action_model': 1.787399019748296e-05, 'epoch': 1.96}) +[2026-06-18 04:17:25,825][starVLA.training.train_starvla][INFO] - Step 5345, Loss: {'train/loss': 0.1916133612394333, 'train/grad_norm_pre_clip': 26.366939544677734, 'timing/data': 0.03191793151199818, 'timing/model': 1.0603946838527918, 'learning_rate/qwen_vl_interface': 1.785918915112169e-06, 'learning_rate/action_model': 1.785918915112169e-05, 'epoch': 1.96}) +[2026-06-18 04:17:34,847][starVLA.training.train_starvla][INFO] - Step 5346, Loss: {'train/loss': 0.5407610535621643, 'train/grad_norm_pre_clip': 26.256656646728516, 'timing/data': 0.027812033891677856, 'timing/model': 2.1525972336530685, 'learning_rate/qwen_vl_interface': 1.7844395285837773e-06, 'learning_rate/action_model': 1.7844395285837774e-05, 'epoch': 1.96}) +[2026-06-18 04:17:39,509][starVLA.training.train_starvla][INFO] - Step 5347, Loss: {'train/loss': 0.10523030161857605, 'train/grad_norm_pre_clip': 14.399408340454102, 'timing/data': 0.034727000631392, 'timing/model': 1.1143265506252646, 'learning_rate/qwen_vl_interface': 1.7829608604698003e-06, 'learning_rate/action_model': 1.7829608604698e-05, 'epoch': 1.96}) +[2026-06-18 04:17:48,471][starVLA.training.train_starvla][INFO] - Step 5348, Loss: {'train/loss': 0.33867549896240234, 'train/grad_norm_pre_clip': 25.770246505737305, 'timing/data': 0.029410644434392452, 'timing/model': 5.4841529643163085, 'learning_rate/qwen_vl_interface': 1.7814829110767685e-06, 'learning_rate/action_model': 1.7814829110767685e-05, 'epoch': 1.96}) +[2026-06-18 04:17:53,507][starVLA.training.train_starvla][INFO] - Step 5349, Loss: {'train/loss': 0.21991664171218872, 'train/grad_norm_pre_clip': 16.9581298828125, 'timing/data': 0.03708492685109377, 'timing/model': 1.7326213335618377, 'learning_rate/qwen_vl_interface': 1.7800056807110627e-06, 'learning_rate/action_model': 1.7800056807110626e-05, 'epoch': 1.96}) +[2026-06-18 04:18:00,407][starVLA.training.train_starvla][INFO] - Step 5350, Loss: {'train/loss': 0.2378227412700653, 'train/grad_norm_pre_clip': 15.82083511352539, 'timing/data': 0.041561661288142204, 'timing/model': 1.0517460824921727, 'learning_rate/qwen_vl_interface': 1.7785291696789103e-06, 'learning_rate/action_model': 1.7785291696789102e-05, 'epoch': 1.96}) +[2026-06-18 04:18:05,034][starVLA.training.train_starvla][INFO] - Step 5351, Loss: {'train/loss': 0.29186639189720154, 'train/grad_norm_pre_clip': 22.421329498291016, 'timing/data': 0.0317159341648221, 'timing/model': 1.1818706821650267, 'learning_rate/qwen_vl_interface': 1.7770533782863983e-06, 'learning_rate/action_model': 1.7770533782863983e-05, 'epoch': 1.96}) +[2026-06-18 04:18:09,709][starVLA.training.train_starvla][INFO] - Step 5352, Loss: {'train/loss': 0.21093438565731049, 'train/grad_norm_pre_clip': 18.657230377197266, 'timing/data': 0.031366984359920025, 'timing/model': 1.1649878155440092, 'learning_rate/qwen_vl_interface': 1.7755783068394578e-06, 'learning_rate/action_model': 1.7755783068394577e-05, 'epoch': 1.96}) +[2026-06-18 04:18:18,009][starVLA.training.train_starvla][INFO] - Step 5353, Loss: {'train/loss': 0.3564832806587219, 'train/grad_norm_pre_clip': 48.93285369873047, 'timing/data': 0.02991753537207842, 'timing/model': 4.818488718010485, 'learning_rate/qwen_vl_interface': 1.7741039556438733e-06, 'learning_rate/action_model': 1.7741039556438733e-05, 'epoch': 1.96}) +[2026-06-18 04:18:22,668][starVLA.training.train_starvla][INFO] - Step 5354, Loss: {'train/loss': 0.48862424492836, 'train/grad_norm_pre_clip': 33.22171401977539, 'timing/data': 0.036035360768437386, 'timing/model': 1.380614004097879, 'learning_rate/qwen_vl_interface': 1.7726303250052785e-06, 'learning_rate/action_model': 1.7726303250052784e-05, 'epoch': 1.96}) +[2026-06-18 04:18:26,877][starVLA.training.train_starvla][INFO] - Step 5355, Loss: {'train/loss': 0.42381253838539124, 'train/grad_norm_pre_clip': 32.3815803527832, 'timing/data': 0.03528731130063534, 'timing/model': 1.1535209799185395, 'learning_rate/qwen_vl_interface': 1.7711574152291591e-06, 'learning_rate/action_model': 1.771157415229159e-05, 'epoch': 1.96}) +[2026-06-18 04:18:31,431][starVLA.training.train_starvla][INFO] - Step 5356, Loss: {'train/loss': 0.23408746719360352, 'train/grad_norm_pre_clip': 22.446680068969727, 'timing/data': 0.034729829989373684, 'timing/model': 1.079767294228077, 'learning_rate/qwen_vl_interface': 1.7696852266208512e-06, 'learning_rate/action_model': 1.7696852266208512e-05, 'epoch': 1.96}) +[2026-06-18 04:18:35,992][starVLA.training.train_starvla][INFO] - Step 5357, Loss: {'train/loss': 0.484749972820282, 'train/grad_norm_pre_clip': 29.449880599975586, 'timing/data': 0.031585861928761005, 'timing/model': 1.0976824024692178, 'learning_rate/qwen_vl_interface': 1.7682137594855408e-06, 'learning_rate/action_model': 1.768213759485541e-05, 'epoch': 1.96}) +[2026-06-18 04:18:40,452][starVLA.training.train_starvla][INFO] - Step 5358, Loss: {'train/loss': 0.1572190225124359, 'train/grad_norm_pre_clip': 14.837738037109375, 'timing/data': 0.031217588111758232, 'timing/model': 1.0701982285827398, 'learning_rate/qwen_vl_interface': 1.7667430141282649e-06, 'learning_rate/action_model': 1.7667430141282648e-05, 'epoch': 1.96}) +[2026-06-18 04:18:44,945][starVLA.training.train_starvla][INFO] - Step 5359, Loss: {'train/loss': 0.3191134035587311, 'train/grad_norm_pre_clip': 27.18147087097168, 'timing/data': 0.037821936421096325, 'timing/model': 1.4038470797240734, 'learning_rate/qwen_vl_interface': 1.7652729908539101e-06, 'learning_rate/action_model': 1.76527299085391e-05, 'epoch': 1.96}) +[2026-06-18 04:18:49,051][starVLA.training.train_starvla][INFO] - Step 5360, Loss: {'train/loss': 0.4272087514400482, 'train/grad_norm_pre_clip': 21.79641342163086, 'timing/data': 0.0298001691699028, 'timing/model': 1.0627182526513934, 'learning_rate/qwen_vl_interface': 1.7638036899672137e-06, 'learning_rate/action_model': 1.7638036899672136e-05, 'epoch': 1.96}) +[2026-06-18 04:18:53,503][starVLA.training.train_starvla][INFO] - Step 5361, Loss: {'train/loss': 0.30074751377105713, 'train/grad_norm_pre_clip': 16.89737892150879, 'timing/data': 0.03897356614470482, 'timing/model': 1.0545957777649164, 'learning_rate/qwen_vl_interface': 1.7623351117727628e-06, 'learning_rate/action_model': 1.7623351117727626e-05, 'epoch': 1.96}) +[2026-06-18 04:18:58,195][starVLA.training.train_starvla][INFO] - Step 5362, Loss: {'train/loss': 0.294341504573822, 'train/grad_norm_pre_clip': 25.246932983398438, 'timing/data': 0.04405456595122814, 'timing/model': 1.0798041289672256, 'learning_rate/qwen_vl_interface': 1.7608672565749973e-06, 'learning_rate/action_model': 1.760867256574997e-05, 'epoch': 1.96}) +[2026-06-18 04:19:02,653][starVLA.training.train_starvla][INFO] - Step 5363, Loss: {'train/loss': 0.44935229420661926, 'train/grad_norm_pre_clip': 28.855070114135742, 'timing/data': 0.033238197676837444, 'timing/model': 1.0615921281278133, 'learning_rate/qwen_vl_interface': 1.7594001246782039e-06, 'learning_rate/action_model': 1.7594001246782037e-05, 'epoch': 1.96}) +[2026-06-18 04:19:07,268][starVLA.training.train_starvla][INFO] - Step 5364, Loss: {'train/loss': 0.2811135947704315, 'train/grad_norm_pre_clip': 27.593524932861328, 'timing/data': 0.03459795471280813, 'timing/model': 1.4350114008411765, 'learning_rate/qwen_vl_interface': 1.7579337163865221e-06, 'learning_rate/action_model': 1.757933716386522e-05, 'epoch': 1.96}) +[2026-06-18 04:19:11,551][starVLA.training.train_starvla][INFO] - Step 5365, Loss: {'train/loss': 0.09716997295618057, 'train/grad_norm_pre_clip': 11.575389862060547, 'timing/data': 0.030952450819313526, 'timing/model': 1.073217349126935, 'learning_rate/qwen_vl_interface': 1.7564680320039356e-06, 'learning_rate/action_model': 1.7564680320039358e-05, 'epoch': 1.96}) +[2026-06-18 04:19:15,985][starVLA.training.train_starvla][INFO] - Step 5366, Loss: {'train/loss': 0.27829423546791077, 'train/grad_norm_pre_clip': 23.776430130004883, 'timing/data': 0.031869745813310146, 'timing/model': 1.0750015629455447, 'learning_rate/qwen_vl_interface': 1.7550030718342867e-06, 'learning_rate/action_model': 1.7550030718342867e-05, 'epoch': 1.96}) +[2026-06-18 04:19:20,638][starVLA.training.train_starvla][INFO] - Step 5367, Loss: {'train/loss': 0.31372687220573425, 'train/grad_norm_pre_clip': 33.44279861450195, 'timing/data': 0.03165680356323719, 'timing/model': 1.2790005821734667, 'learning_rate/qwen_vl_interface': 1.753538836181262e-06, 'learning_rate/action_model': 1.753538836181262e-05, 'epoch': 1.97}) +[2026-06-18 04:19:25,132][starVLA.training.train_starvla][INFO] - Step 5368, Loss: {'train/loss': 0.11107513308525085, 'train/grad_norm_pre_clip': 14.460744857788086, 'timing/data': 0.03960940148681402, 'timing/model': 1.1018096972256899, 'learning_rate/qwen_vl_interface': 1.7520753253483987e-06, 'learning_rate/action_model': 1.7520753253483986e-05, 'epoch': 1.97}) +[2026-06-18 04:19:36,434][starVLA.training.train_starvla][INFO] - Step 5369, Loss: {'train/loss': 0.2301209568977356, 'train/grad_norm_pre_clip': 22.32332992553711, 'timing/data': 0.03247779328376055, 'timing/model': 1.5011094147339463, 'learning_rate/qwen_vl_interface': 1.7506125396390847e-06, 'learning_rate/action_model': 1.7506125396390847e-05, 'epoch': 1.97}) +[2026-06-18 04:19:40,891][starVLA.training.train_starvla][INFO] - Step 5370, Loss: {'train/loss': 0.2647942900657654, 'train/grad_norm_pre_clip': 23.523468017578125, 'timing/data': 0.03145495057106018, 'timing/model': 1.1958144279196858, 'learning_rate/qwen_vl_interface': 1.7491504793565564e-06, 'learning_rate/action_model': 1.7491504793565562e-05, 'epoch': 1.97}) +[2026-06-18 04:19:45,790][starVLA.training.train_starvla][INFO] - Step 5371, Loss: {'train/loss': 0.6839659810066223, 'train/grad_norm_pre_clip': 41.77201843261719, 'timing/data': 0.02859031967818737, 'timing/model': 1.077710066922009, 'learning_rate/qwen_vl_interface': 1.7476891448039012e-06, 'learning_rate/action_model': 1.7476891448039013e-05, 'epoch': 1.97}) +[2026-06-18 04:19:51,050][starVLA.training.train_starvla][INFO] - Step 5372, Loss: {'train/loss': 0.4009811580181122, 'train/grad_norm_pre_clip': 24.134836196899414, 'timing/data': 0.03261594194918871, 'timing/model': 1.8110656486824155, 'learning_rate/qwen_vl_interface': 1.7462285362840534e-06, 'learning_rate/action_model': 1.746228536284053e-05, 'epoch': 1.97}) +[2026-06-18 04:19:59,142][starVLA.training.train_starvla][INFO] - Step 5373, Loss: {'train/loss': 0.3552759289741516, 'train/grad_norm_pre_clip': 27.042186737060547, 'timing/data': 0.033057513646781445, 'timing/model': 4.461633891798556, 'learning_rate/qwen_vl_interface': 1.7447686540998027e-06, 'learning_rate/action_model': 1.7447686540998026e-05, 'epoch': 1.97}) +[2026-06-18 04:20:11,632][starVLA.training.train_starvla][INFO] - Step 5374, Loss: {'train/loss': 0.23183143138885498, 'train/grad_norm_pre_clip': 26.341955184936523, 'timing/data': 0.029775594361126423, 'timing/model': 4.650058372877538, 'learning_rate/qwen_vl_interface': 1.743309498553781e-06, 'learning_rate/action_model': 1.743309498553781e-05, 'epoch': 1.97}) +[2026-06-18 04:20:22,333][starVLA.training.train_starvla][INFO] - Step 5375, Loss: {'train/loss': 0.10704297572374344, 'train/grad_norm_pre_clip': 20.7316951751709, 'timing/data': 0.03276226110756397, 'timing/model': 1.7464660461992025, 'learning_rate/qwen_vl_interface': 1.7418510699484728e-06, 'learning_rate/action_model': 1.7418510699484727e-05, 'epoch': 1.97}) +[2026-06-18 04:20:33,521][starVLA.training.train_starvla][INFO] - Step 5376, Loss: {'train/loss': 0.1596134752035141, 'train/grad_norm_pre_clip': 17.912546157836914, 'timing/data': 0.030586011707782745, 'timing/model': 1.0753143411129713, 'learning_rate/qwen_vl_interface': 1.7403933685862118e-06, 'learning_rate/action_model': 1.7403933685862118e-05, 'epoch': 1.97}) +[2026-06-18 04:20:41,411][starVLA.training.train_starvla][INFO] - Step 5377, Loss: {'train/loss': 0.18677201867103577, 'train/grad_norm_pre_clip': 26.862104415893555, 'timing/data': 0.03133056964725256, 'timing/model': 1.0428832927718759, 'learning_rate/qwen_vl_interface': 1.7389363947691837e-06, 'learning_rate/action_model': 1.7389363947691838e-05, 'epoch': 1.97}) +[2026-06-18 04:20:45,954][starVLA.training.train_starvla][INFO] - Step 5378, Loss: {'train/loss': 0.3140280544757843, 'train/grad_norm_pre_clip': 19.705055236816406, 'timing/data': 0.0389946810901165, 'timing/model': 1.1040047127753496, 'learning_rate/qwen_vl_interface': 1.7374801487994198e-06, 'learning_rate/action_model': 1.7374801487994198e-05, 'epoch': 1.97}) +[2026-06-18 04:20:50,222][starVLA.training.train_starvla][INFO] - Step 5379, Loss: {'train/loss': 0.1554711014032364, 'train/grad_norm_pre_clip': 15.027897834777832, 'timing/data': 0.03551302012056112, 'timing/model': 1.1488521238788962, 'learning_rate/qwen_vl_interface': 1.7360246309787982e-06, 'learning_rate/action_model': 1.736024630978798e-05, 'epoch': 1.97}) +[2026-06-18 04:20:54,850][starVLA.training.train_starvla][INFO] - Step 5380, Loss: {'train/loss': 0.3460031747817993, 'train/grad_norm_pre_clip': 32.69184112548828, 'timing/data': 0.030227995477616787, 'timing/model': 1.1604729807004333, 'learning_rate/qwen_vl_interface': 1.7345698416090534e-06, 'learning_rate/action_model': 1.7345698416090533e-05, 'epoch': 1.97}) +[2026-06-18 04:20:59,346][starVLA.training.train_starvla][INFO] - Step 5381, Loss: {'train/loss': 0.307404100894928, 'train/grad_norm_pre_clip': 18.569303512573242, 'timing/data': 0.030743532814085484, 'timing/model': 1.104557960294187, 'learning_rate/qwen_vl_interface': 1.7331157809917632e-06, 'learning_rate/action_model': 1.733115780991763e-05, 'epoch': 1.97}) +[2026-06-18 04:21:04,018][starVLA.training.train_starvla][INFO] - Step 5382, Loss: {'train/loss': 0.33617764711380005, 'train/grad_norm_pre_clip': 26.16042709350586, 'timing/data': 0.03983140643686056, 'timing/model': 1.0774688050150871, 'learning_rate/qwen_vl_interface': 1.7316624494283557e-06, 'learning_rate/action_model': 1.7316624494283557e-05, 'epoch': 1.97}) +[2026-06-18 04:21:08,782][starVLA.training.train_starvla][INFO] - Step 5383, Loss: {'train/loss': 0.02587592601776123, 'train/grad_norm_pre_clip': 4.0601277351379395, 'timing/data': 0.03890156652778387, 'timing/model': 1.4682967104017735, 'learning_rate/qwen_vl_interface': 1.7302098472201085e-06, 'learning_rate/action_model': 1.7302098472201084e-05, 'epoch': 1.97}) +[2026-06-18 04:21:13,244][starVLA.training.train_starvla][INFO] - Step 5384, Loss: {'train/loss': 0.4382116198539734, 'train/grad_norm_pre_clip': 30.78017807006836, 'timing/data': 0.02778467256575823, 'timing/model': 1.476621720008552, 'learning_rate/qwen_vl_interface': 1.728757974668147e-06, 'learning_rate/action_model': 1.728757974668147e-05, 'epoch': 1.97}) +[2026-06-18 04:21:17,543][starVLA.training.train_starvla][INFO] - Step 5385, Loss: {'train/loss': 0.28870025277137756, 'train/grad_norm_pre_clip': 21.145889282226562, 'timing/data': 0.029728690162301064, 'timing/model': 1.0626397961750627, 'learning_rate/qwen_vl_interface': 1.727306832073447e-06, 'learning_rate/action_model': 1.7273068320734472e-05, 'epoch': 1.97}) +[2026-06-18 04:21:21,979][starVLA.training.train_starvla][INFO] - Step 5386, Loss: {'train/loss': 0.1789596974849701, 'train/grad_norm_pre_clip': 17.738067626953125, 'timing/data': 0.03438553772866726, 'timing/model': 1.045960501767695, 'learning_rate/qwen_vl_interface': 1.7258564197368308e-06, 'learning_rate/action_model': 1.725856419736831e-05, 'epoch': 1.97}) +[2026-06-18 04:21:26,461][starVLA.training.train_starvla][INFO] - Step 5387, Loss: {'train/loss': 0.3571125864982605, 'train/grad_norm_pre_clip': 29.61119270324707, 'timing/data': 0.03047177940607071, 'timing/model': 1.0679614134132862, 'learning_rate/qwen_vl_interface': 1.7244067379589708e-06, 'learning_rate/action_model': 1.7244067379589707e-05, 'epoch': 1.97}) +[2026-06-18 04:21:31,219][starVLA.training.train_starvla][INFO] - Step 5388, Loss: {'train/loss': 0.3889750838279724, 'train/grad_norm_pre_clip': 25.0977783203125, 'timing/data': 0.030560490675270557, 'timing/model': 1.2422599783167243, 'learning_rate/qwen_vl_interface': 1.7229577870403886e-06, 'learning_rate/action_model': 1.7229577870403886e-05, 'epoch': 1.97}) +[2026-06-18 04:21:38,960][starVLA.training.train_starvla][INFO] - Step 5389, Loss: {'train/loss': 0.3107779622077942, 'train/grad_norm_pre_clip': 30.375362396240234, 'timing/data': 0.04318359401077032, 'timing/model': 3.0986990919336677, 'learning_rate/qwen_vl_interface': 1.7215095672814517e-06, 'learning_rate/action_model': 1.7215095672814516e-05, 'epoch': 1.97}) +[2026-06-18 04:21:43,082][starVLA.training.train_starvla][INFO] - Step 5390, Loss: {'train/loss': 0.17379960417747498, 'train/grad_norm_pre_clip': 17.80824089050293, 'timing/data': 0.03781003598123789, 'timing/model': 1.0621832553297281, 'learning_rate/qwen_vl_interface': 1.7200620789823769e-06, 'learning_rate/action_model': 1.7200620789823767e-05, 'epoch': 1.97}) +[2026-06-18 04:21:50,915][starVLA.training.train_starvla][INFO] - Step 5391, Loss: {'train/loss': 0.45349761843681335, 'train/grad_norm_pre_clip': 29.939023971557617, 'timing/data': 0.031795209273695946, 'timing/model': 4.37555821146816, 'learning_rate/qwen_vl_interface': 1.7186153224432331e-06, 'learning_rate/action_model': 1.718615322443233e-05, 'epoch': 1.97}) +[2026-06-18 04:21:57,328][starVLA.training.train_starvla][INFO] - Step 5392, Loss: {'train/loss': 0.15785081684589386, 'train/grad_norm_pre_clip': 14.138142585754395, 'timing/data': 0.034658001735806465, 'timing/model': 1.1353850979357958, 'learning_rate/qwen_vl_interface': 1.7171692979639342e-06, 'learning_rate/action_model': 1.717169297963934e-05, 'epoch': 1.97}) +[2026-06-18 04:22:01,845][starVLA.training.train_starvla][INFO] - Step 5393, Loss: {'train/loss': 0.3702162206172943, 'train/grad_norm_pre_clip': 32.0844841003418, 'timing/data': 0.038094738498330116, 'timing/model': 1.1592310620471835, 'learning_rate/qwen_vl_interface': 1.7157240058442397e-06, 'learning_rate/action_model': 1.71572400584424e-05, 'epoch': 1.97}) +[2026-06-18 04:22:06,575][starVLA.training.train_starvla][INFO] - Step 5394, Loss: {'train/loss': 0.34271904826164246, 'train/grad_norm_pre_clip': 20.768980026245117, 'timing/data': 0.036223470233380795, 'timing/model': 1.4413813324645162, 'learning_rate/qwen_vl_interface': 1.7142794463837608e-06, 'learning_rate/action_model': 1.714279446383761e-05, 'epoch': 1.98}) +[2026-06-18 04:22:10,820][starVLA.training.train_starvla][INFO] - Step 5395, Loss: {'train/loss': 0.19120851159095764, 'train/grad_norm_pre_clip': 21.792613983154297, 'timing/data': 0.02909725345671177, 'timing/model': 1.0454704035073519, 'learning_rate/qwen_vl_interface': 1.712835619881959e-06, 'learning_rate/action_model': 1.7128356198819588e-05, 'epoch': 1.98}) +[2026-06-18 04:22:16,444][starVLA.training.train_starvla][INFO] - Step 5396, Loss: {'train/loss': 0.16884098947048187, 'train/grad_norm_pre_clip': 13.134989738464355, 'timing/data': 0.036273423582315445, 'timing/model': 1.0599574502557516, 'learning_rate/qwen_vl_interface': 1.7113925266381383e-06, 'learning_rate/action_model': 1.7113925266381383e-05, 'epoch': 1.98}) +[2026-06-18 04:22:23,552][starVLA.training.train_starvla][INFO] - Step 5397, Loss: {'train/loss': 0.36852937936782837, 'train/grad_norm_pre_clip': 25.69713020324707, 'timing/data': 0.030713957734405994, 'timing/model': 1.160043683834374, 'learning_rate/qwen_vl_interface': 1.7099501669514569e-06, 'learning_rate/action_model': 1.7099501669514568e-05, 'epoch': 1.98}) +[2026-06-18 04:22:28,245][starVLA.training.train_starvla][INFO] - Step 5398, Loss: {'train/loss': 0.30219537019729614, 'train/grad_norm_pre_clip': 25.898452758789062, 'timing/data': 0.029451463371515274, 'timing/model': 1.5258547710254788, 'learning_rate/qwen_vl_interface': 1.7085085411209118e-06, 'learning_rate/action_model': 1.7085085411209118e-05, 'epoch': 1.98}) +[2026-06-18 04:22:32,861][starVLA.training.train_starvla][INFO] - Step 5399, Loss: {'train/loss': 0.19501301646232605, 'train/grad_norm_pre_clip': 17.892837524414062, 'timing/data': 0.028594656847417355, 'timing/model': 1.368640661239624, 'learning_rate/qwen_vl_interface': 1.7070676494453587e-06, 'learning_rate/action_model': 1.7070676494453586e-05, 'epoch': 1.98}) +[2026-06-18 04:26:03,570][starVLA.training.train_starvla][INFO] - Step 5400, Loss: {'train/loss': 0.4531080722808838, 'train/grad_norm_pre_clip': 27.770401000976562, 'eval/loss': 0.5396537184715271, 'eval/latency_2/loss': 0.5396537184715271, 'eval/demon_fire_precision': 0.9418032786885245, 'eval/demon_fire_recall': 0.9615062761506277, 'eval/demon_fire_f1': 0.9515527950310559, 'eval/demon_fire_rate_model': 0.6789092932665554, 'eval/demon_fire_rate_teacher': 0.6649972175848636, 'eval/demon_left_precision': 0.8994708994708994, 'eval/demon_left_recall': 0.8947368421052632, 'eval/demon_left_f1': 0.8970976253298153, 'eval/demon_left_rate_model': 0.2629382303839733, 'eval/demon_left_rate_teacher': 0.2643294379521425, 'eval/demon_right_precision': 0.8917127071823204, 'eval/demon_right_recall': 0.911864406779661, 'eval/demon_right_f1': 0.9016759776536313, 'eval/demon_right_rate_model': 0.25180856983861993, 'eval/demon_right_rate_teacher': 0.24624373956594323, 'eval/demon_move_f1': 0.8993868014917232, 'eval/demon_cc_f1': 0.9352784633872253, 'eval/latency_2/demon_fire_precision': 0.9418032786885245, 'eval/latency_2/demon_fire_recall': 0.9615062761506277, 'eval/latency_2/demon_fire_f1': 0.9515527950310559, 'eval/latency_2/demon_fire_rate_model': 0.6789092932665554, 'eval/latency_2/demon_fire_rate_teacher': 0.6649972175848636, 'eval/latency_2/demon_left_precision': 0.8994708994708994, 'eval/latency_2/demon_left_recall': 0.8947368421052632, 'eval/latency_2/demon_left_f1': 0.8970976253298153, 'eval/latency_2/demon_left_rate_model': 0.2629382303839733, 'eval/latency_2/demon_left_rate_teacher': 0.2643294379521425, 'eval/latency_2/demon_right_precision': 0.8917127071823204, 'eval/latency_2/demon_right_recall': 0.911864406779661, 'eval/latency_2/demon_right_f1': 0.9016759776536313, 'eval/latency_2/demon_right_rate_model': 0.25180856983861993, 'eval/latency_2/demon_right_rate_teacher': 0.24624373956594323, 'eval/latency_2/demon_move_f1': 0.8993868014917232, 'eval/latency_2/demon_cc_f1': 0.9352784633872253, 'timing/data': 0.036302090622484684, 'timing/model': 5.3140336237847805, 'learning_rate/qwen_vl_interface': 1.705627492223494e-06, 'learning_rate/action_model': 1.705627492223494e-05, 'epoch': 1.98}) +[2026-06-18 04:26:03,591][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_5400_state +[2026-06-18 04:26:03,598][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 04:26:11,440][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_5400_state/model.safetensors +[2026-06-18 04:26:25,002][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_5400_state/optimizer.bin +[2026-06-18 04:26:25,004][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_5400_state/sampler.bin +[2026-06-18 04:26:25,005][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_5400_state/sampler_1.bin +[2026-06-18 04:26:25,008][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_5400_state/random_states_0.pkl +[2026-06-18 04:26:25,016][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 5400 +[2026-06-18 04:26:25,016][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 04:26:25,033][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 04:26:29,066][starVLA.training.train_starvla][INFO] - Step 5401, Loss: {'train/loss': 0.398229718208313, 'train/grad_norm_pre_clip': 27.6181583404541, 'timing/data': 0.03014926053583622, 'timing/model': 1.040936722420156, 'learning_rate/qwen_vl_interface': 1.7041880697538644e-06, 'learning_rate/action_model': 1.704188069753864e-05, 'epoch': 1.98}) +[2026-06-18 04:26:33,374][starVLA.training.train_starvla][INFO] - Step 5402, Loss: {'train/loss': 0.23773039877414703, 'train/grad_norm_pre_clip': 25.148828506469727, 'timing/data': 0.026854739524424076, 'timing/model': 1.0264955656602979, 'learning_rate/qwen_vl_interface': 1.702749382334863e-06, 'learning_rate/action_model': 1.7027493823348632e-05, 'epoch': 1.98}) +[2026-06-18 04:26:37,689][starVLA.training.train_starvla][INFO] - Step 5403, Loss: {'train/loss': 0.46512776613235474, 'train/grad_norm_pre_clip': 27.905818939208984, 'timing/data': 0.029212288558483124, 'timing/model': 1.0376603212207556, 'learning_rate/qwen_vl_interface': 1.7013114302647318e-06, 'learning_rate/action_model': 1.7013114302647318e-05, 'epoch': 1.98}) +[2026-06-18 04:26:42,009][starVLA.training.train_starvla][INFO] - Step 5404, Loss: {'train/loss': 0.2001655250787735, 'train/grad_norm_pre_clip': 15.361701011657715, 'timing/data': 0.025443682447075844, 'timing/model': 1.0300623439252377, 'learning_rate/qwen_vl_interface': 1.69987421384156e-06, 'learning_rate/action_model': 1.69987421384156e-05, 'epoch': 1.98}) +[2026-06-18 04:26:46,243][starVLA.training.train_starvla][INFO] - Step 5405, Loss: {'train/loss': 0.8085185289382935, 'train/grad_norm_pre_clip': 35.64223098754883, 'timing/data': 0.025867565535008907, 'timing/model': 1.254823294468224, 'learning_rate/qwen_vl_interface': 1.6984377333632815e-06, 'learning_rate/action_model': 1.6984377333632813e-05, 'epoch': 1.98}) +[2026-06-18 04:26:50,293][starVLA.training.train_starvla][INFO] - Step 5406, Loss: {'train/loss': 0.18602392077445984, 'train/grad_norm_pre_clip': 17.92881965637207, 'timing/data': 0.026221847161650658, 'timing/model': 1.0240922765806317, 'learning_rate/qwen_vl_interface': 1.697001989127686e-06, 'learning_rate/action_model': 1.697001989127686e-05, 'epoch': 1.98}) +[2026-06-18 04:26:55,762][starVLA.training.train_starvla][INFO] - Step 5407, Loss: {'train/loss': 0.44340547919273376, 'train/grad_norm_pre_clip': 52.21092987060547, 'timing/data': 0.026822327636182308, 'timing/model': 1.040730464272201, 'learning_rate/qwen_vl_interface': 1.6955669814324003e-06, 'learning_rate/action_model': 1.6955669814324e-05, 'epoch': 1.98}) +[2026-06-18 04:27:00,106][starVLA.training.train_starvla][INFO] - Step 5408, Loss: {'train/loss': 0.33182206749916077, 'train/grad_norm_pre_clip': 31.551986694335938, 'timing/data': 0.03721481375396252, 'timing/model': 1.0516124228015542, 'learning_rate/qwen_vl_interface': 1.6941327105749034e-06, 'learning_rate/action_model': 1.694132710574903e-05, 'epoch': 1.98}) +[2026-06-18 04:27:04,584][starVLA.training.train_starvla][INFO] - Step 5409, Loss: {'train/loss': 0.3544823229312897, 'train/grad_norm_pre_clip': 100.41498565673828, 'timing/data': 0.03158514108508825, 'timing/model': 1.047817231155932, 'learning_rate/qwen_vl_interface': 1.6926991768525205e-06, 'learning_rate/action_model': 1.6926991768525205e-05, 'epoch': 1.98}) +[2026-06-18 04:27:09,178][starVLA.training.train_starvla][INFO] - Step 5410, Loss: {'train/loss': 0.42762067914009094, 'train/grad_norm_pre_clip': 22.196216583251953, 'timing/data': 0.028711194172501564, 'timing/model': 1.4147635120898485, 'learning_rate/qwen_vl_interface': 1.6912663805624278e-06, 'learning_rate/action_model': 1.6912663805624275e-05, 'epoch': 1.98}) +[2026-06-18 04:27:15,009][starVLA.training.train_starvla][INFO] - Step 5411, Loss: {'train/loss': 0.07949342578649521, 'train/grad_norm_pre_clip': 11.634260177612305, 'timing/data': 0.045262450352311134, 'timing/model': 1.0670158304274082, 'learning_rate/qwen_vl_interface': 1.6898343220016446e-06, 'learning_rate/action_model': 1.6898343220016447e-05, 'epoch': 1.98}) +[2026-06-18 04:27:19,682][starVLA.training.train_starvla][INFO] - Step 5412, Loss: {'train/loss': 0.4243212640285492, 'train/grad_norm_pre_clip': 30.100685119628906, 'timing/data': 0.04720819555222988, 'timing/model': 1.1392875984311104, 'learning_rate/qwen_vl_interface': 1.6884030014670353e-06, 'learning_rate/action_model': 1.688403001467035e-05, 'epoch': 1.98}) +[2026-06-18 04:27:24,470][starVLA.training.train_starvla][INFO] - Step 5413, Loss: {'train/loss': 0.17446604371070862, 'train/grad_norm_pre_clip': 19.472225189208984, 'timing/data': 0.03443989437073469, 'timing/model': 1.1423590360209346, 'learning_rate/qwen_vl_interface': 1.6869724192553168e-06, 'learning_rate/action_model': 1.686972419255317e-05, 'epoch': 1.98}) +[2026-06-18 04:27:29,046][starVLA.training.train_starvla][INFO] - Step 5414, Loss: {'train/loss': 0.3012166917324066, 'train/grad_norm_pre_clip': 22.53448486328125, 'timing/data': 0.03287898004055023, 'timing/model': 1.1782469609752297, 'learning_rate/qwen_vl_interface': 1.6855425756630506e-06, 'learning_rate/action_model': 1.6855425756630506e-05, 'epoch': 1.98}) +[2026-06-18 04:27:36,860][starVLA.training.train_starvla][INFO] - Step 5415, Loss: {'train/loss': 0.21030615270137787, 'train/grad_norm_pre_clip': 23.894895553588867, 'timing/data': 0.0380446370691061, 'timing/model': 1.6006923448294401, 'learning_rate/qwen_vl_interface': 1.684113470986644e-06, 'learning_rate/action_model': 1.6841134709866438e-05, 'epoch': 1.98}) +[2026-06-18 04:27:45,776][starVLA.training.train_starvla][INFO] - Step 5416, Loss: {'train/loss': 0.2934468388557434, 'train/grad_norm_pre_clip': 28.97073745727539, 'timing/data': 0.0333533501252532, 'timing/model': 1.091773391701281, 'learning_rate/qwen_vl_interface': 1.682685105522351e-06, 'learning_rate/action_model': 1.682685105522351e-05, 'epoch': 1.98}) +[2026-06-18 04:27:50,595][starVLA.training.train_starvla][INFO] - Step 5417, Loss: {'train/loss': 0.4574194550514221, 'train/grad_norm_pre_clip': 26.901409149169922, 'timing/data': 0.0327953714877367, 'timing/model': 1.2115631643682718, 'learning_rate/qwen_vl_interface': 1.6812574795662751e-06, 'learning_rate/action_model': 1.681257479566275e-05, 'epoch': 1.98}) +[2026-06-18 04:27:55,372][starVLA.training.train_starvla][INFO] - Step 5418, Loss: {'train/loss': 0.22559887170791626, 'train/grad_norm_pre_clip': 18.973007202148438, 'timing/data': 0.04265475366264582, 'timing/model': 1.0537774031981826, 'learning_rate/qwen_vl_interface': 1.6798305934143635e-06, 'learning_rate/action_model': 1.6798305934143633e-05, 'epoch': 1.98}) +[2026-06-18 04:28:00,095][starVLA.training.train_starvla][INFO] - Step 5419, Loss: {'train/loss': 0.17623060941696167, 'train/grad_norm_pre_clip': 41.16128921508789, 'timing/data': 0.03256232105195522, 'timing/model': 1.3059224048629403, 'learning_rate/qwen_vl_interface': 1.6784044473624118e-06, 'learning_rate/action_model': 1.6784044473624118e-05, 'epoch': 1.98}) +[2026-06-18 04:28:07,847][starVLA.training.train_starvla][INFO] - Step 5420, Loss: {'train/loss': 0.42803582549095154, 'train/grad_norm_pre_clip': 29.32512664794922, 'timing/data': 0.03896453231573105, 'timing/model': 4.670981721952558, 'learning_rate/qwen_vl_interface': 1.6769790417060611e-06, 'learning_rate/action_model': 1.676979041706061e-05, 'epoch': 1.98}) +[2026-06-18 04:28:14,729][starVLA.training.train_starvla][INFO] - Step 5421, Loss: {'train/loss': 0.4649006724357605, 'train/grad_norm_pre_clip': 52.51887130737305, 'timing/data': 0.03055308759212494, 'timing/model': 1.2328887665644288, 'learning_rate/qwen_vl_interface': 1.6755543767407998e-06, 'learning_rate/action_model': 1.6755543767407998e-05, 'epoch': 1.99}) +[2026-06-18 04:28:23,620][starVLA.training.train_starvla][INFO] - Step 5422, Loss: {'train/loss': 0.32956093549728394, 'train/grad_norm_pre_clip': 21.613971710205078, 'timing/data': 0.03009956143796444, 'timing/model': 1.0455731144174933, 'learning_rate/qwen_vl_interface': 1.6741304527619624e-06, 'learning_rate/action_model': 1.674130452761962e-05, 'epoch': 1.99}) +[2026-06-18 04:28:28,154][starVLA.training.train_starvla][INFO] - Step 5423, Loss: {'train/loss': 0.2653956711292267, 'train/grad_norm_pre_clip': 25.77810287475586, 'timing/data': 0.03719525132328272, 'timing/model': 1.065364295616746, 'learning_rate/qwen_vl_interface': 1.672707270064729e-06, 'learning_rate/action_model': 1.6727072700647288e-05, 'epoch': 1.99}) +[2026-06-18 04:28:32,821][starVLA.training.train_starvla][INFO] - Step 5424, Loss: {'train/loss': 0.0349682942032814, 'train/grad_norm_pre_clip': 4.461682319641113, 'timing/data': 0.04183868132531643, 'timing/model': 1.0862572891637683, 'learning_rate/qwen_vl_interface': 1.6712848289441288e-06, 'learning_rate/action_model': 1.6712848289441287e-05, 'epoch': 1.99}) +[2026-06-18 04:28:37,714][starVLA.training.train_starvla][INFO] - Step 5425, Loss: {'train/loss': 0.2107122540473938, 'train/grad_norm_pre_clip': 30.62914276123047, 'timing/data': 0.040253378450870514, 'timing/model': 1.669814740307629, 'learning_rate/qwen_vl_interface': 1.669863129695036e-06, 'learning_rate/action_model': 1.669863129695036e-05, 'epoch': 1.99}) +[2026-06-18 04:28:43,751][starVLA.training.train_starvla][INFO] - Step 5426, Loss: {'train/loss': 0.32724881172180176, 'train/grad_norm_pre_clip': 25.6092529296875, 'timing/data': 0.02846135664731264, 'timing/model': 2.873372950591147, 'learning_rate/qwen_vl_interface': 1.6684421726121675e-06, 'learning_rate/action_model': 1.6684421726121674e-05, 'epoch': 1.99}) +[2026-06-18 04:28:50,167][starVLA.training.train_starvla][INFO] - Step 5427, Loss: {'train/loss': 0.15219569206237793, 'train/grad_norm_pre_clip': 14.351569175720215, 'timing/data': 0.03187090251594782, 'timing/model': 1.112084036692977, 'learning_rate/qwen_vl_interface': 1.6670219579900888e-06, 'learning_rate/action_model': 1.6670219579900887e-05, 'epoch': 1.99}) +[2026-06-18 04:28:54,782][starVLA.training.train_starvla][INFO] - Step 5428, Loss: {'train/loss': 0.25432610511779785, 'train/grad_norm_pre_clip': 28.635108947753906, 'timing/data': 0.03266190830618143, 'timing/model': 1.0339224692434072, 'learning_rate/qwen_vl_interface': 1.665602486123215e-06, 'learning_rate/action_model': 1.665602486123215e-05, 'epoch': 1.99}) +[2026-06-18 04:28:59,422][starVLA.training.train_starvla][INFO] - Step 5429, Loss: {'train/loss': 0.25542494654655457, 'train/grad_norm_pre_clip': 20.609386444091797, 'timing/data': 0.03361249715089798, 'timing/model': 1.072956882417202, 'learning_rate/qwen_vl_interface': 1.6641837573058034e-06, 'learning_rate/action_model': 1.6641837573058035e-05, 'epoch': 1.99}) +[2026-06-18 04:29:04,117][starVLA.training.train_starvla][INFO] - Step 5430, Loss: {'train/loss': 0.1648544818162918, 'train/grad_norm_pre_clip': 22.15255355834961, 'timing/data': 0.0908910259604454, 'timing/model': 1.6098436387255788, 'learning_rate/qwen_vl_interface': 1.6627657718319568e-06, 'learning_rate/action_model': 1.6627657718319568e-05, 'epoch': 1.99}) +[2026-06-18 04:29:08,454][starVLA.training.train_starvla][INFO] - Step 5431, Loss: {'train/loss': 0.31682777404785156, 'train/grad_norm_pre_clip': 21.58852767944336, 'timing/data': 0.03300780523568392, 'timing/model': 1.1072730086743832, 'learning_rate/qwen_vl_interface': 1.6613485299956265e-06, 'learning_rate/action_model': 1.6613485299956262e-05, 'epoch': 1.99}) +[2026-06-18 04:29:13,077][starVLA.training.train_starvla][INFO] - Step 5432, Loss: {'train/loss': 0.5269163846969604, 'train/grad_norm_pre_clip': 31.24543571472168, 'timing/data': 0.02925453707575798, 'timing/model': 1.0808116309344769, 'learning_rate/qwen_vl_interface': 1.6599320320906064e-06, 'learning_rate/action_model': 1.6599320320906066e-05, 'epoch': 1.99}) +[2026-06-18 04:29:17,785][starVLA.training.train_starvla][INFO] - Step 5433, Loss: {'train/loss': 0.22805257141590118, 'train/grad_norm_pre_clip': 22.968412399291992, 'timing/data': 0.03290541376918554, 'timing/model': 1.1370961973443627, 'learning_rate/qwen_vl_interface': 1.65851627841054e-06, 'learning_rate/action_model': 1.65851627841054e-05, 'epoch': 1.99}) +[2026-06-18 04:29:22,446][starVLA.training.train_starvla][INFO] - Step 5434, Loss: {'train/loss': 0.2298840582370758, 'train/grad_norm_pre_clip': 15.23806381225586, 'timing/data': 0.030541881918907166, 'timing/model': 1.0489930594339967, 'learning_rate/qwen_vl_interface': 1.657101269248913e-06, 'learning_rate/action_model': 1.657101269248913e-05, 'epoch': 1.99}) +[2026-06-18 04:29:26,615][starVLA.training.train_starvla][INFO] - Step 5435, Loss: {'train/loss': 0.2560316026210785, 'train/grad_norm_pre_clip': 18.685964584350586, 'timing/data': 0.02918393537402153, 'timing/model': 1.0578526807948947, 'learning_rate/qwen_vl_interface': 1.65568700489906e-06, 'learning_rate/action_model': 1.65568700489906e-05, 'epoch': 1.99}) +[2026-06-18 04:29:31,336][starVLA.training.train_starvla][INFO] - Step 5436, Loss: {'train/loss': 0.3006528913974762, 'train/grad_norm_pre_clip': 19.53391456604004, 'timing/data': 0.03056358825415373, 'timing/model': 1.0470232963562012, 'learning_rate/qwen_vl_interface': 1.6542734856541577e-06, 'learning_rate/action_model': 1.6542734856541576e-05, 'epoch': 1.99}) +[2026-06-18 04:29:35,947][starVLA.training.train_starvla][INFO] - Step 5437, Loss: {'train/loss': 0.15835131704807281, 'train/grad_norm_pre_clip': 16.273521423339844, 'timing/data': 0.034678028896450996, 'timing/model': 1.1304036499932408, 'learning_rate/qwen_vl_interface': 1.652860711807231e-06, 'learning_rate/action_model': 1.652860711807231e-05, 'epoch': 1.99}) +[2026-06-18 04:29:40,462][starVLA.training.train_starvla][INFO] - Step 5438, Loss: {'train/loss': 0.13067837059497833, 'train/grad_norm_pre_clip': 17.539722442626953, 'timing/data': 0.033826669678092, 'timing/model': 1.0924313738942146, 'learning_rate/qwen_vl_interface': 1.6514486836511484e-06, 'learning_rate/action_model': 1.6514486836511484e-05, 'epoch': 1.99}) +[2026-06-18 04:29:45,229][starVLA.training.train_starvla][INFO] - Step 5439, Loss: {'train/loss': 0.33381399512290955, 'train/grad_norm_pre_clip': 23.799598693847656, 'timing/data': 0.03431615326553583, 'timing/model': 1.3918987940996885, 'learning_rate/qwen_vl_interface': 1.6500374014786288e-06, 'learning_rate/action_model': 1.650037401478629e-05, 'epoch': 1.99}) +[2026-06-18 04:29:49,668][starVLA.training.train_starvla][INFO] - Step 5440, Loss: {'train/loss': 0.3713095784187317, 'train/grad_norm_pre_clip': 37.37122344970703, 'timing/data': 0.034959519281983376, 'timing/model': 1.3414544221013784, 'learning_rate/qwen_vl_interface': 1.6486268655822277e-06, 'learning_rate/action_model': 1.6486268655822277e-05, 'epoch': 1.99}) +[2026-06-18 04:29:55,118][starVLA.training.train_starvla][INFO] - Step 5441, Loss: {'train/loss': 0.10793078690767288, 'train/grad_norm_pre_clip': 12.992201805114746, 'timing/data': 0.030277306213974953, 'timing/model': 1.2193575343117118, 'learning_rate/qwen_vl_interface': 1.6472170762543515e-06, 'learning_rate/action_model': 1.6472170762543514e-05, 'epoch': 1.99}) +[2026-06-18 04:29:59,683][starVLA.training.train_starvla][INFO] - Step 5442, Loss: {'train/loss': 0.4421839118003845, 'train/grad_norm_pre_clip': 31.931568145751953, 'timing/data': 0.028948507271707058, 'timing/model': 1.0445477403700352, 'learning_rate/qwen_vl_interface': 1.6458080337872534e-06, 'learning_rate/action_model': 1.6458080337872535e-05, 'epoch': 1.99}) +[2026-06-18 04:30:04,223][starVLA.training.train_starvla][INFO] - Step 5443, Loss: {'train/loss': 0.17589204013347626, 'train/grad_norm_pre_clip': 17.17477798461914, 'timing/data': 0.03315012715756893, 'timing/model': 1.1675923131406307, 'learning_rate/qwen_vl_interface': 1.644399738473029e-06, 'learning_rate/action_model': 1.644399738473029e-05, 'epoch': 1.99}) +[2026-06-18 04:30:08,750][starVLA.training.train_starvla][INFO] - Step 5444, Loss: {'train/loss': 0.5741912722587585, 'train/grad_norm_pre_clip': 35.991477966308594, 'timing/data': 0.035669538192451, 'timing/model': 1.0683308308944106, 'learning_rate/qwen_vl_interface': 1.6429921906036175e-06, 'learning_rate/action_model': 1.6429921906036176e-05, 'epoch': 1.99}) +[2026-06-18 04:30:16,691][starVLA.training.train_starvla][INFO] - Step 5445, Loss: {'train/loss': 0.4899117946624756, 'train/grad_norm_pre_clip': 62.98454284667969, 'timing/data': 0.02795260027050972, 'timing/model': 1.4127382347360253, 'learning_rate/qwen_vl_interface': 1.6415853904708068e-06, 'learning_rate/action_model': 1.6415853904708066e-05, 'epoch': 1.99}) +[2026-06-18 04:30:20,937][starVLA.training.train_starvla][INFO] - Step 5446, Loss: {'train/loss': 0.4475051760673523, 'train/grad_norm_pre_clip': 28.496986389160156, 'timing/data': 0.03635384328663349, 'timing/model': 1.067334357649088, 'learning_rate/qwen_vl_interface': 1.6401793383662274e-06, 'learning_rate/action_model': 1.6401793383662273e-05, 'epoch': 1.99}) +[2026-06-18 04:30:25,581][starVLA.training.train_starvla][INFO] - Step 5447, Loss: {'train/loss': 0.2918016314506531, 'train/grad_norm_pre_clip': 28.495798110961914, 'timing/data': 0.028056268580257893, 'timing/model': 1.14017484895885, 'learning_rate/qwen_vl_interface': 1.6387740345813553e-06, 'learning_rate/action_model': 1.6387740345813552e-05, 'epoch': 1.99}) +[2026-06-18 04:30:31,781][starVLA.training.train_starvla][INFO] - Step 5448, Loss: {'train/loss': 0.08050771057605743, 'train/grad_norm_pre_clip': 12.657175064086914, 'timing/data': 0.028530067764222622, 'timing/model': 2.7419163696467876, 'learning_rate/qwen_vl_interface': 1.6373694794075115e-06, 'learning_rate/action_model': 1.6373694794075115e-05, 'epoch': 1.99}) +[2026-06-18 04:30:36,325][starVLA.training.train_starvla][INFO] - Step 5449, Loss: {'train/loss': 0.17925740778446198, 'train/grad_norm_pre_clip': 12.787379264831543, 'timing/data': 0.03379755560308695, 'timing/model': 1.043159807100892, 'learning_rate/qwen_vl_interface': 1.6359656731358626e-06, 'learning_rate/action_model': 1.6359656731358626e-05, 'epoch': 2.0}) +[2026-06-18 04:30:41,136][starVLA.training.train_starvla][INFO] - Step 5450, Loss: {'train/loss': 0.20233528316020966, 'train/grad_norm_pre_clip': 17.535446166992188, 'timing/data': 0.032437462359666824, 'timing/model': 1.4612461440265179, 'learning_rate/qwen_vl_interface': 1.634562616057418e-06, 'learning_rate/action_model': 1.634562616057418e-05, 'epoch': 2.0}) +[2026-06-18 04:30:45,313][starVLA.training.train_starvla][INFO] - Step 5451, Loss: {'train/loss': 0.10160722583532333, 'train/grad_norm_pre_clip': 10.934684753417969, 'timing/data': 0.03531102929264307, 'timing/model': 1.1188650820404291, 'learning_rate/qwen_vl_interface': 1.6331603084630338e-06, 'learning_rate/action_model': 1.6331603084630338e-05, 'epoch': 2.0}) +[2026-06-18 04:30:49,900][starVLA.training.train_starvla][INFO] - Step 5452, Loss: {'train/loss': 0.7187881469726562, 'train/grad_norm_pre_clip': 36.770118713378906, 'timing/data': 0.02820158377289772, 'timing/model': 1.0330316768959165, 'learning_rate/qwen_vl_interface': 1.6317587506434076e-06, 'learning_rate/action_model': 1.6317587506434076e-05, 'epoch': 2.0}) +[2026-06-18 04:30:59,745][starVLA.training.train_starvla][INFO] - Step 5453, Loss: {'train/loss': 0.2641132175922394, 'train/grad_norm_pre_clip': 31.599212646484375, 'timing/data': 0.03118568193167448, 'timing/model': 4.547265263274312, 'learning_rate/qwen_vl_interface': 1.6303579428890881e-06, 'learning_rate/action_model': 1.630357942889088e-05, 'epoch': 2.0}) +[2026-06-18 04:31:04,543][starVLA.training.train_starvla][INFO] - Step 5454, Loss: {'train/loss': 0.2285572588443756, 'train/grad_norm_pre_clip': 16.811203002929688, 'timing/data': 0.030345194973051548, 'timing/model': 1.5730358585715294, 'learning_rate/qwen_vl_interface': 1.6289578854904626e-06, 'learning_rate/action_model': 1.6289578854904625e-05, 'epoch': 2.0}) +[2026-06-18 04:31:17,727][starVLA.training.train_starvla][INFO] - Step 5455, Loss: {'train/loss': 0.20452506840229034, 'train/grad_norm_pre_clip': 27.562345504760742, 'timing/data': 0.03667284827679396, 'timing/model': 5.818937050178647, 'learning_rate/qwen_vl_interface': 1.6275585787377627e-06, 'learning_rate/action_model': 1.6275585787377626e-05, 'epoch': 2.0}) +[2026-06-18 04:31:24,715][starVLA.training.train_starvla][INFO] - Step 5456, Loss: {'train/loss': 0.20880359411239624, 'train/grad_norm_pre_clip': 26.782426834106445, 'timing/data': 0.03806670568883419, 'timing/model': 1.0773581368848681, 'learning_rate/qwen_vl_interface': 1.6261600229210662e-06, 'learning_rate/action_model': 1.626160022921066e-05, 'epoch': 2.0}) +[2026-06-18 04:31:34,485][starVLA.training.train_starvla][INFO] - Step 5457, Loss: {'train/loss': 0.2475637048482895, 'train/grad_norm_pre_clip': 19.903596878051758, 'timing/data': 0.03410954214632511, 'timing/model': 1.1246732529252768, 'learning_rate/qwen_vl_interface': 1.6247622183302972e-06, 'learning_rate/action_model': 1.624762218330297e-05, 'epoch': 2.0}) +[2026-06-18 04:31:39,192][starVLA.training.train_starvla][INFO] - Step 5458, Loss: {'train/loss': 0.2310618907213211, 'train/grad_norm_pre_clip': 17.626379013061523, 'timing/data': 0.03261975944042206, 'timing/model': 1.137221448123455, 'learning_rate/qwen_vl_interface': 1.6233651652552218e-06, 'learning_rate/action_model': 1.6233651652552217e-05, 'epoch': 2.0}) +[2026-06-18 04:31:49,039][starVLA.training.train_starvla][INFO] - Step 5459, Loss: {'train/loss': 0.4610963463783264, 'train/grad_norm_pre_clip': 25.825029373168945, 'timing/data': 0.03898256924003363, 'timing/model': 1.0983705827966332, 'learning_rate/qwen_vl_interface': 1.6219688639854507e-06, 'learning_rate/action_model': 1.6219688639854506e-05, 'epoch': 2.0}) +[2026-06-18 04:31:59,336][starVLA.training.train_starvla][INFO] - Step 5460, Loss: {'train/loss': 0.2672349810600281, 'train/grad_norm_pre_clip': 18.139694213867188, 'timing/data': 0.028951462358236313, 'timing/model': 4.737489738501608, 'learning_rate/qwen_vl_interface': 1.620573314810436e-06, 'learning_rate/action_model': 1.620573314810436e-05, 'epoch': 2.0}) +[2026-06-18 04:32:04,017][starVLA.training.train_starvla][INFO] - Step 5461, Loss: {'train/loss': 0.28740429878234863, 'train/grad_norm_pre_clip': 27.397762298583984, 'timing/data': 0.032431784085929394, 'timing/model': 1.2286386108025908, 'learning_rate/qwen_vl_interface': 1.6191785180194797e-06, 'learning_rate/action_model': 1.6191785180194796e-05, 'epoch': 2.0}) +[2026-06-18 04:32:08,448][starVLA.training.train_starvla][INFO] - Step 5462, Loss: {'train/loss': 0.3079172670841217, 'train/grad_norm_pre_clip': 36.856624603271484, 'timing/data': 0.0010314378887414932, 'timing/model': 0.8803596803918481, 'learning_rate/qwen_vl_interface': 1.6177844739017236e-06, 'learning_rate/action_model': 1.6177844739017236e-05, 'epoch': 2.0}) +[2026-06-18 04:32:13,523][starVLA.training.train_starvla][INFO] - Step 5463, Loss: {'train/loss': 0.36743614077568054, 'train/grad_norm_pre_clip': 24.855039596557617, 'timing/data': 0.013688098639249802, 'timing/model': 1.0637179967015982, 'learning_rate/qwen_vl_interface': 1.6163911827461546e-06, 'learning_rate/action_model': 1.6163911827461546e-05, 'epoch': 2.0}) +[2026-06-18 04:32:18,025][starVLA.training.train_starvla][INFO] - Step 5464, Loss: {'train/loss': 0.46831953525543213, 'train/grad_norm_pre_clip': 29.048959732055664, 'timing/data': 0.013271206989884377, 'timing/model': 1.4290968989953399, 'learning_rate/qwen_vl_interface': 1.6149986448416032e-06, 'learning_rate/action_model': 1.614998644841603e-05, 'epoch': 2.0}) +[2026-06-18 04:32:22,457][starVLA.training.train_starvla][INFO] - Step 5465, Loss: {'train/loss': 0.37717294692993164, 'train/grad_norm_pre_clip': 38.9537467956543, 'timing/data': 0.010649877600371838, 'timing/model': 1.1060655787587166, 'learning_rate/qwen_vl_interface': 1.6136068604767444e-06, 'learning_rate/action_model': 1.6136068604767442e-05, 'epoch': 2.0}) +[2026-06-18 04:32:26,962][starVLA.training.train_starvla][INFO] - Step 5466, Loss: {'train/loss': 0.2928442060947418, 'train/grad_norm_pre_clip': 22.06903839111328, 'timing/data': 0.014023523777723312, 'timing/model': 1.059658644720912, 'learning_rate/qwen_vl_interface': 1.6122158299400962e-06, 'learning_rate/action_model': 1.612215829940096e-05, 'epoch': 2.0}) +[2026-06-18 04:32:31,477][starVLA.training.train_starvla][INFO] - Step 5467, Loss: {'train/loss': 0.04661794379353523, 'train/grad_norm_pre_clip': 7.173714637756348, 'timing/data': 0.010852102190256119, 'timing/model': 1.0935743721202016, 'learning_rate/qwen_vl_interface': 1.6108255535200193e-06, 'learning_rate/action_model': 1.6108255535200193e-05, 'epoch': 2.0}) +[2026-06-18 04:32:36,043][starVLA.training.train_starvla][INFO] - Step 5468, Loss: {'train/loss': 0.22481243312358856, 'train/grad_norm_pre_clip': 22.70793914794922, 'timing/data': 0.01269613765180111, 'timing/model': 1.0780872944742441, 'learning_rate/qwen_vl_interface': 1.6094360315047236e-06, 'learning_rate/action_model': 1.6094360315047237e-05, 'epoch': 2.0}) +[2026-06-18 04:32:40,514][starVLA.training.train_starvla][INFO] - Step 5469, Loss: {'train/loss': 0.17544777691364288, 'train/grad_norm_pre_clip': 11.27872085571289, 'timing/data': 0.013130886480212212, 'timing/model': 1.3803492253646255, 'learning_rate/qwen_vl_interface': 1.608047264182255e-06, 'learning_rate/action_model': 1.608047264182255e-05, 'epoch': 2.0}) +[2026-06-18 04:32:45,071][starVLA.training.train_starvla][INFO] - Step 5470, Loss: {'train/loss': 0.5290889739990234, 'train/grad_norm_pre_clip': 40.496646881103516, 'timing/data': 0.011491373181343079, 'timing/model': 1.502673045732081, 'learning_rate/qwen_vl_interface': 1.6066592518405077e-06, 'learning_rate/action_model': 1.6066592518405076e-05, 'epoch': 2.0}) +[2026-06-18 04:32:49,326][starVLA.training.train_starvla][INFO] - Step 5471, Loss: {'train/loss': 0.149065300822258, 'train/grad_norm_pre_clip': 21.191164016723633, 'timing/data': 0.014101807959377766, 'timing/model': 1.0768281957134604, 'learning_rate/qwen_vl_interface': 1.605271994767216e-06, 'learning_rate/action_model': 1.605271994767216e-05, 'epoch': 2.0}) +[2026-06-18 04:32:53,929][starVLA.training.train_starvla][INFO] - Step 5472, Loss: {'train/loss': 0.1874520629644394, 'train/grad_norm_pre_clip': 17.68720817565918, 'timing/data': 0.010616838000714779, 'timing/model': 1.2391644939780235, 'learning_rate/qwen_vl_interface': 1.6038854932499635e-06, 'learning_rate/action_model': 1.6038854932499633e-05, 'epoch': 2.0}) +[2026-06-18 04:32:58,423][starVLA.training.train_starvla][INFO] - Step 5473, Loss: {'train/loss': 0.13023436069488525, 'train/grad_norm_pre_clip': 15.280756950378418, 'timing/data': 0.01221761666238308, 'timing/model': 1.0670088501647115, 'learning_rate/qwen_vl_interface': 1.6024997475761727e-06, 'learning_rate/action_model': 1.602499747576173e-05, 'epoch': 2.0}) +[2026-06-18 04:33:02,971][starVLA.training.train_starvla][INFO] - Step 5474, Loss: {'train/loss': 0.26895323395729065, 'train/grad_norm_pre_clip': 39.30838394165039, 'timing/data': 0.012410996481776237, 'timing/model': 1.1015833904966712, 'learning_rate/qwen_vl_interface': 1.6011147580331065e-06, 'learning_rate/action_model': 1.6011147580331066e-05, 'epoch': 2.0}) +[2026-06-18 04:33:07,528][starVLA.training.train_starvla][INFO] - Step 5475, Loss: {'train/loss': 0.31468480825424194, 'train/grad_norm_pre_clip': 29.563276290893555, 'timing/data': 0.01142721064388752, 'timing/model': 1.3826934220269322, 'learning_rate/qwen_vl_interface': 1.5997305249078788e-06, 'learning_rate/action_model': 1.5997305249078787e-05, 'epoch': 2.0}) +[2026-06-18 04:33:11,759][starVLA.training.train_starvla][INFO] - Step 5476, Loss: {'train/loss': 0.07429259270429611, 'train/grad_norm_pre_clip': 11.023482322692871, 'timing/data': 0.010801729746162891, 'timing/model': 1.0786275807768106, 'learning_rate/qwen_vl_interface': 1.5983470484874412e-06, 'learning_rate/action_model': 1.598347048487441e-05, 'epoch': 2.01}) +[2026-06-18 04:33:18,043][starVLA.training.train_starvla][INFO] - Step 5477, Loss: {'train/loss': 0.20610764622688293, 'train/grad_norm_pre_clip': 16.330772399902344, 'timing/data': 0.016276631504297256, 'timing/model': 2.810636102221906, 'learning_rate/qwen_vl_interface': 1.5969643290585893e-06, 'learning_rate/action_model': 1.5969643290585892e-05, 'epoch': 2.01}) +[2026-06-18 04:33:22,803][starVLA.training.train_starvla][INFO] - Step 5478, Loss: {'train/loss': 0.30232033133506775, 'train/grad_norm_pre_clip': 24.971025466918945, 'timing/data': 0.010705055668950081, 'timing/model': 1.0630044555291533, 'learning_rate/qwen_vl_interface': 1.5955823669079632e-06, 'learning_rate/action_model': 1.595582366907963e-05, 'epoch': 2.01}) +[2026-06-18 04:33:32,700][starVLA.training.train_starvla][INFO] - Step 5479, Loss: {'train/loss': 0.35617268085479736, 'train/grad_norm_pre_clip': 28.055051803588867, 'timing/data': 0.0165304746478796, 'timing/model': 4.460467068478465, 'learning_rate/qwen_vl_interface': 1.594201162322045e-06, 'learning_rate/action_model': 1.594201162322045e-05, 'epoch': 2.01}) +[2026-06-18 04:33:37,568][starVLA.training.train_starvla][INFO] - Step 5480, Loss: {'train/loss': 0.2452477663755417, 'train/grad_norm_pre_clip': 28.73885726928711, 'timing/data': 0.015390407294034958, 'timing/model': 1.528602591715753, 'learning_rate/qwen_vl_interface': 1.592820715587159e-06, 'learning_rate/action_model': 1.5928207155871593e-05, 'epoch': 2.01}) +[2026-06-18 04:33:41,956][starVLA.training.train_starvla][INFO] - Step 5481, Loss: {'train/loss': 0.19908107817173004, 'train/grad_norm_pre_clip': 16.585878372192383, 'timing/data': 0.011299364268779755, 'timing/model': 1.0913357911631465, 'learning_rate/qwen_vl_interface': 1.5914410269894743e-06, 'learning_rate/action_model': 1.591441026989474e-05, 'epoch': 2.01}) +[2026-06-18 04:33:48,458][starVLA.training.train_starvla][INFO] - Step 5482, Loss: {'train/loss': 0.19279275834560394, 'train/grad_norm_pre_clip': 19.259204864501953, 'timing/data': 0.013709432445466518, 'timing/model': 1.123798318207264, 'learning_rate/qwen_vl_interface': 1.5900620968150015e-06, 'learning_rate/action_model': 1.5900620968150015e-05, 'epoch': 2.01}) +[2026-06-18 04:33:53,881][starVLA.training.train_starvla][INFO] - Step 5483, Loss: {'train/loss': 0.7155276536941528, 'train/grad_norm_pre_clip': 42.047767639160156, 'timing/data': 0.011804493144154549, 'timing/model': 1.0492164986208081, 'learning_rate/qwen_vl_interface': 1.5886839253495944e-06, 'learning_rate/action_model': 1.5886839253495943e-05, 'epoch': 2.01}) +[2026-06-18 04:34:03,371][starVLA.training.train_starvla][INFO] - Step 5484, Loss: {'train/loss': 0.3095778226852417, 'train/grad_norm_pre_clip': 27.18478775024414, 'timing/data': 0.020273350179195404, 'timing/model': 1.346974316984415, 'learning_rate/qwen_vl_interface': 1.5873065128789495e-06, 'learning_rate/action_model': 1.5873065128789493e-05, 'epoch': 2.01}) +[2026-06-18 04:34:08,110][starVLA.training.train_starvla][INFO] - Step 5485, Loss: {'train/loss': 0.2919340431690216, 'train/grad_norm_pre_clip': 14.695290565490723, 'timing/data': 0.010608472861349583, 'timing/model': 1.5621000984683633, 'learning_rate/qwen_vl_interface': 1.5859298596886038e-06, 'learning_rate/action_model': 1.585929859688604e-05, 'epoch': 2.01}) +[2026-06-18 04:34:12,252][starVLA.training.train_starvla][INFO] - Step 5486, Loss: {'train/loss': 0.24026153981685638, 'train/grad_norm_pre_clip': 20.549123764038086, 'timing/data': 0.011283652856945992, 'timing/model': 1.0525872213765979, 'learning_rate/qwen_vl_interface': 1.5845539660639429e-06, 'learning_rate/action_model': 1.5845539660639426e-05, 'epoch': 2.01}) +[2026-06-18 04:34:16,815][starVLA.training.train_starvla][INFO] - Step 5487, Loss: {'train/loss': 0.6877451539039612, 'train/grad_norm_pre_clip': 41.71014404296875, 'timing/data': 0.01453667413443327, 'timing/model': 1.117591081187129, 'learning_rate/qwen_vl_interface': 1.5831788322901901e-06, 'learning_rate/action_model': 1.58317883229019e-05, 'epoch': 2.01}) +[2026-06-18 04:34:21,617][starVLA.training.train_starvla][INFO] - Step 5488, Loss: {'train/loss': 0.15466371178627014, 'train/grad_norm_pre_clip': 16.170358657836914, 'timing/data': 0.013230586424469948, 'timing/model': 1.1460466152057052, 'learning_rate/qwen_vl_interface': 1.5818044586524092e-06, 'learning_rate/action_model': 1.5818044586524092e-05, 'epoch': 2.01}) +[2026-06-18 04:34:26,221][starVLA.training.train_starvla][INFO] - Step 5489, Loss: {'train/loss': 0.3001624047756195, 'train/grad_norm_pre_clip': 21.604496002197266, 'timing/data': 0.016210000030696392, 'timing/model': 1.1483913902193308, 'learning_rate/qwen_vl_interface': 1.5804308454355102e-06, 'learning_rate/action_model': 1.5804308454355102e-05, 'epoch': 2.01}) +[2026-06-18 04:34:30,777][starVLA.training.train_starvla][INFO] - Step 5490, Loss: {'train/loss': 0.38566887378692627, 'train/grad_norm_pre_clip': 29.976106643676758, 'timing/data': 0.010767036117613316, 'timing/model': 1.3638661801815033, 'learning_rate/qwen_vl_interface': 1.5790579929242475e-06, 'learning_rate/action_model': 1.5790579929242475e-05, 'epoch': 2.01}) +[2026-06-18 04:34:35,208][starVLA.training.train_starvla][INFO] - Step 5491, Loss: {'train/loss': 0.30091360211372375, 'train/grad_norm_pre_clip': 23.134994506835938, 'timing/data': 0.013579539023339748, 'timing/model': 1.230767022818327, 'learning_rate/qwen_vl_interface': 1.577685901403212e-06, 'learning_rate/action_model': 1.577685901403212e-05, 'epoch': 2.01}) +[2026-06-18 04:34:39,794][starVLA.training.train_starvla][INFO] - Step 5492, Loss: {'train/loss': 0.3128375709056854, 'train/grad_norm_pre_clip': 24.108049392700195, 'timing/data': 0.015777728520333767, 'timing/model': 1.0656457738950849, 'learning_rate/qwen_vl_interface': 1.5763145711568414e-06, 'learning_rate/action_model': 1.5763145711568413e-05, 'epoch': 2.01}) +[2026-06-18 04:34:44,508][starVLA.training.train_starvla][INFO] - Step 5493, Loss: {'train/loss': 0.4942745864391327, 'train/grad_norm_pre_clip': 33.776851654052734, 'timing/data': 0.010972827672958374, 'timing/model': 1.051200714893639, 'learning_rate/qwen_vl_interface': 1.5749440024694129e-06, 'learning_rate/action_model': 1.574944002469413e-05, 'epoch': 2.01}) +[2026-06-18 04:34:48,975][starVLA.training.train_starvla][INFO] - Step 5494, Loss: {'train/loss': 0.21041005849838257, 'train/grad_norm_pre_clip': 14.247271537780762, 'timing/data': 0.01271724607795477, 'timing/model': 1.0750598488375545, 'learning_rate/qwen_vl_interface': 1.5735741956250476e-06, 'learning_rate/action_model': 1.5735741956250475e-05, 'epoch': 2.01}) +[2026-06-18 04:34:53,313][starVLA.training.train_starvla][INFO] - Step 5495, Loss: {'train/loss': 0.20688532292842865, 'train/grad_norm_pre_clip': 21.937910079956055, 'timing/data': 0.012555324472486973, 'timing/model': 1.0822874223813415, 'learning_rate/qwen_vl_interface': 1.5722051509077068e-06, 'learning_rate/action_model': 1.572205150907707e-05, 'epoch': 2.01}) +[2026-06-18 04:34:58,290][starVLA.training.train_starvla][INFO] - Step 5496, Loss: {'train/loss': 0.5818889141082764, 'train/grad_norm_pre_clip': 37.286895751953125, 'timing/data': 0.010674409568309784, 'timing/model': 1.1090564569458365, 'learning_rate/qwen_vl_interface': 1.570836868601196e-06, 'learning_rate/action_model': 1.570836868601196e-05, 'epoch': 2.01}) +[2026-06-18 04:35:02,919][starVLA.training.train_starvla][INFO] - Step 5497, Loss: {'train/loss': 0.2563696801662445, 'train/grad_norm_pre_clip': 24.28208351135254, 'timing/data': 0.012287861667573452, 'timing/model': 1.2477510822936893, 'learning_rate/qwen_vl_interface': 1.5694693489891616e-06, 'learning_rate/action_model': 1.5694693489891617e-05, 'epoch': 2.01}) +[2026-06-18 04:35:07,545][starVLA.training.train_starvla][INFO] - Step 5498, Loss: {'train/loss': 0.23415395617485046, 'train/grad_norm_pre_clip': 25.718364715576172, 'timing/data': 0.010439456440508366, 'timing/model': 1.056550863198936, 'learning_rate/qwen_vl_interface': 1.5681025923550914e-06, 'learning_rate/action_model': 1.568102592355091e-05, 'epoch': 2.01}) +[2026-06-18 04:35:12,495][starVLA.training.train_starvla][INFO] - Step 5499, Loss: {'train/loss': 0.4443987309932709, 'train/grad_norm_pre_clip': 30.138315200805664, 'timing/data': 0.010344896465539932, 'timing/model': 1.5142815373837948, 'learning_rate/qwen_vl_interface': 1.5667365989823161e-06, 'learning_rate/action_model': 1.566736598982316e-05, 'epoch': 2.01}) +[2026-06-18 04:38:41,158][starVLA.training.train_starvla][INFO] - Step 5500, Loss: {'train/loss': 0.3692679703235626, 'train/grad_norm_pre_clip': 22.395549774169922, 'eval/loss': 0.5366097092628479, 'eval/latency_2/loss': 0.5366097092628479, 'eval/demon_fire_precision': 0.9511278195488722, 'eval/demon_fire_recall': 0.9527196652719665, 'eval/demon_fire_f1': 0.9519230769230769, 'eval/demon_fire_rate_model': 0.666110183639399, 'eval/demon_fire_rate_teacher': 0.6649972175848636, 'eval/demon_left_precision': 0.9091903719912473, 'eval/demon_left_recall': 0.8747368421052631, 'eval/demon_left_f1': 0.8916309012875536, 'eval/demon_left_rate_model': 0.25431274346132443, 'eval/demon_left_rate_teacher': 0.2643294379521425, 'eval/demon_right_precision': 0.8752642706131079, 'eval/demon_right_recall': 0.9355932203389831, 'eval/demon_right_f1': 0.9044238121245222, 'eval/demon_right_rate_model': 0.2632164718976071, 'eval/demon_right_rate_teacher': 0.24624373956594323, 'eval/demon_move_f1': 0.898027356706038, 'eval/demon_cc_f1': 0.9350871109798613, 'eval/latency_2/demon_fire_precision': 0.9511278195488722, 'eval/latency_2/demon_fire_recall': 0.9527196652719665, 'eval/latency_2/demon_fire_f1': 0.9519230769230769, 'eval/latency_2/demon_fire_rate_model': 0.666110183639399, 'eval/latency_2/demon_fire_rate_teacher': 0.6649972175848636, 'eval/latency_2/demon_left_precision': 0.9091903719912473, 'eval/latency_2/demon_left_recall': 0.8747368421052631, 'eval/latency_2/demon_left_f1': 0.8916309012875536, 'eval/latency_2/demon_left_rate_model': 0.25431274346132443, 'eval/latency_2/demon_left_rate_teacher': 0.2643294379521425, 'eval/latency_2/demon_right_precision': 0.8752642706131079, 'eval/latency_2/demon_right_recall': 0.9355932203389831, 'eval/latency_2/demon_right_f1': 0.9044238121245222, 'eval/latency_2/demon_right_rate_model': 0.2632164718976071, 'eval/latency_2/demon_right_rate_teacher': 0.24624373956594323, 'eval/latency_2/demon_move_f1': 0.898027356706038, 'eval/latency_2/demon_cc_f1': 0.9350871109798613, 'timing/data': 0.029072774574160576, 'timing/model': 1.4211959289386868, 'learning_rate/qwen_vl_interface': 1.5653713691540053e-06, 'learning_rate/action_model': 1.5653713691540054e-05, 'epoch': 2.01}) +[2026-06-18 04:38:41,160][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_5500_state +[2026-06-18 04:38:41,182][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 04:38:49,935][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_5500_state/model.safetensors +[2026-06-18 04:39:05,130][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_5500_state/optimizer.bin +[2026-06-18 04:39:05,131][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_5500_state/sampler.bin +[2026-06-18 04:39:05,132][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_5500_state/sampler_1.bin +[2026-06-18 04:39:05,136][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_5500_state/random_states_0.pkl +[2026-06-18 04:39:05,149][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 5500 +[2026-06-18 04:39:05,149][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 04:39:05,166][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 04:39:10,535][starVLA.training.train_starvla][INFO] - Step 5501, Loss: {'train/loss': 0.30116212368011475, 'train/grad_norm_pre_clip': 18.82731056213379, 'timing/data': 0.012457078322768211, 'timing/model': 1.7436188291758299, 'learning_rate/qwen_vl_interface': 1.5640069031531774e-06, 'learning_rate/action_model': 1.5640069031531772e-05, 'epoch': 2.01}) +[2026-06-18 04:39:14,787][starVLA.training.train_starvla][INFO] - Step 5502, Loss: {'train/loss': 0.09580791741609573, 'train/grad_norm_pre_clip': 14.336565971374512, 'timing/data': 0.01167752780020237, 'timing/model': 1.0453225681558251, 'learning_rate/qwen_vl_interface': 1.5626432012626835e-06, 'learning_rate/action_model': 1.5626432012626836e-05, 'epoch': 2.01}) +[2026-06-18 04:39:19,618][starVLA.training.train_starvla][INFO] - Step 5503, Loss: {'train/loss': 0.5800068378448486, 'train/grad_norm_pre_clip': 26.885379791259766, 'timing/data': 0.013007205910980701, 'timing/model': 1.086897431872785, 'learning_rate/qwen_vl_interface': 1.5612802637652202e-06, 'learning_rate/action_model': 1.5612802637652203e-05, 'epoch': 2.02}) +[2026-06-18 04:39:24,302][starVLA.training.train_starvla][INFO] - Step 5504, Loss: {'train/loss': 0.40588435530662537, 'train/grad_norm_pre_clip': 26.178144454956055, 'timing/data': 0.011934545822441578, 'timing/model': 1.060989061370492, 'learning_rate/qwen_vl_interface': 1.5599180909433292e-06, 'learning_rate/action_model': 1.5599180909433292e-05, 'epoch': 2.02}) +[2026-06-18 04:39:28,753][starVLA.training.train_starvla][INFO] - Step 5505, Loss: {'train/loss': 0.20566104352474213, 'train/grad_norm_pre_clip': 20.409318923950195, 'timing/data': 0.016894763335585594, 'timing/model': 1.1124602844938636, 'learning_rate/qwen_vl_interface': 1.5585566830793883e-06, 'learning_rate/action_model': 1.558556683079388e-05, 'epoch': 2.02}) +[2026-06-18 04:39:37,489][starVLA.training.train_starvla][INFO] - Step 5506, Loss: {'train/loss': 0.15516065061092377, 'train/grad_norm_pre_clip': 17.318729400634766, 'timing/data': 0.01421395130455494, 'timing/model': 5.408966606482863, 'learning_rate/qwen_vl_interface': 1.5571960404556204e-06, 'learning_rate/action_model': 1.5571960404556203e-05, 'epoch': 2.02}) +[2026-06-18 04:39:41,705][starVLA.training.train_starvla][INFO] - Step 5507, Loss: {'train/loss': 0.4690238833427429, 'train/grad_norm_pre_clip': 22.593591690063477, 'timing/data': 0.009681623429059982, 'timing/model': 1.0685958275571465, 'learning_rate/qwen_vl_interface': 1.5558361633540848e-06, 'learning_rate/action_model': 1.5558361633540848e-05, 'epoch': 2.02}) +[2026-06-18 04:39:52,197][starVLA.training.train_starvla][INFO] - Step 5508, Loss: {'train/loss': 0.3488921523094177, 'train/grad_norm_pre_clip': 28.68551254272461, 'timing/data': 0.013947831466794014, 'timing/model': 1.0361913023516536, 'learning_rate/qwen_vl_interface': 1.5544770520566884e-06, 'learning_rate/action_model': 1.5544770520566885e-05, 'epoch': 2.02}) +[2026-06-18 04:39:56,682][starVLA.training.train_starvla][INFO] - Step 5509, Loss: {'train/loss': 0.19671645760536194, 'train/grad_norm_pre_clip': 19.55748176574707, 'timing/data': 0.011090779677033424, 'timing/model': 1.055702107027173, 'learning_rate/qwen_vl_interface': 1.553118706845176e-06, 'learning_rate/action_model': 1.5531187068451757e-05, 'epoch': 2.02}) +[2026-06-18 04:40:01,672][starVLA.training.train_starvla][INFO] - Step 5510, Loss: {'train/loss': 0.09442910552024841, 'train/grad_norm_pre_clip': 9.871121406555176, 'timing/data': 0.02052948996424675, 'timing/model': 1.123650767840445, 'learning_rate/qwen_vl_interface': 1.5517611280011333e-06, 'learning_rate/action_model': 1.551761128001133e-05, 'epoch': 2.02}) +[2026-06-18 04:40:06,653][starVLA.training.train_starvla][INFO] - Step 5511, Loss: {'train/loss': 0.49489662051200867, 'train/grad_norm_pre_clip': 26.21414566040039, 'timing/data': 0.01194026693701744, 'timing/model': 1.5514928447082639, 'learning_rate/qwen_vl_interface': 1.5504043158059886e-06, 'learning_rate/action_model': 1.5504043158059885e-05, 'epoch': 2.02}) +[2026-06-18 04:40:11,012][starVLA.training.train_starvla][INFO] - Step 5512, Loss: {'train/loss': 0.27150171995162964, 'train/grad_norm_pre_clip': 35.67551803588867, 'timing/data': 0.011880033649504185, 'timing/model': 1.0632330235093832, 'learning_rate/qwen_vl_interface': 1.54904827054101e-06, 'learning_rate/action_model': 1.54904827054101e-05, 'epoch': 2.02}) +[2026-06-18 04:40:15,921][starVLA.training.train_starvla][INFO] - Step 5513, Loss: {'train/loss': 0.1625284105539322, 'train/grad_norm_pre_clip': 18.150157928466797, 'timing/data': 0.01329571008682251, 'timing/model': 1.148577963002026, 'learning_rate/qwen_vl_interface': 1.5476929924873074e-06, 'learning_rate/action_model': 1.5476929924873073e-05, 'epoch': 2.02}) +[2026-06-18 04:40:20,358][starVLA.training.train_starvla][INFO] - Step 5514, Loss: {'train/loss': 0.5139740705490112, 'train/grad_norm_pre_clip': 38.454044342041016, 'timing/data': 0.012774004600942135, 'timing/model': 1.050632132217288, 'learning_rate/qwen_vl_interface': 1.5463384819258302e-06, 'learning_rate/action_model': 1.54633848192583e-05, 'epoch': 2.02}) +[2026-06-18 04:40:25,030][starVLA.training.train_starvla][INFO] - Step 5515, Loss: {'train/loss': 0.27326518297195435, 'train/grad_norm_pre_clip': 28.633960723876953, 'timing/data': 0.011542522348463535, 'timing/model': 1.6428914768621325, 'learning_rate/qwen_vl_interface': 1.544984739137373e-06, 'learning_rate/action_model': 1.544984739137373e-05, 'epoch': 2.02}) +[2026-06-18 04:40:29,592][starVLA.training.train_starvla][INFO] - Step 5516, Loss: {'train/loss': 0.1613989621400833, 'train/grad_norm_pre_clip': 22.263662338256836, 'timing/data': 0.0139073571190238, 'timing/model': 1.3765779389068484, 'learning_rate/qwen_vl_interface': 1.543631764402566e-06, 'learning_rate/action_model': 1.5436317644025658e-05, 'epoch': 2.02}) +[2026-06-18 04:40:37,543][starVLA.training.train_starvla][INFO] - Step 5517, Loss: {'train/loss': 0.23652082681655884, 'train/grad_norm_pre_clip': 25.9125919342041, 'timing/data': 0.014340725727379322, 'timing/model': 1.0971367908641696, 'learning_rate/qwen_vl_interface': 1.5422795580018813e-06, 'learning_rate/action_model': 1.542279558001881e-05, 'epoch': 2.02}) +[2026-06-18 04:40:42,300][starVLA.training.train_starvla][INFO] - Step 5518, Loss: {'train/loss': 0.22071845829486847, 'train/grad_norm_pre_clip': 25.915267944335938, 'timing/data': 0.01233620010316372, 'timing/model': 1.0603020200505853, 'learning_rate/qwen_vl_interface': 1.5409281202156329e-06, 'learning_rate/action_model': 1.540928120215633e-05, 'epoch': 2.02}) +[2026-06-18 04:40:47,003][starVLA.training.train_starvla][INFO] - Step 5519, Loss: {'train/loss': 0.5041757225990295, 'train/grad_norm_pre_clip': 24.11818504333496, 'timing/data': 0.011173291131854057, 'timing/model': 1.213706512004137, 'learning_rate/qwen_vl_interface': 1.5395774513239777e-06, 'learning_rate/action_model': 1.539577451323978e-05, 'epoch': 2.02}) +[2026-06-18 04:40:53,415][starVLA.training.train_starvla][INFO] - Step 5520, Loss: {'train/loss': 0.25545331835746765, 'train/grad_norm_pre_clip': 19.770727157592773, 'timing/data': 0.021872403100132942, 'timing/model': 2.7939367946237326, 'learning_rate/qwen_vl_interface': 1.5382275516069112e-06, 'learning_rate/action_model': 1.5382275516069113e-05, 'epoch': 2.02}) +[2026-06-18 04:41:01,311][starVLA.training.train_starvla][INFO] - Step 5521, Loss: {'train/loss': 0.21666820347309113, 'train/grad_norm_pre_clip': 24.16815948486328, 'timing/data': 0.015080094337463379, 'timing/model': 1.531780743971467, 'learning_rate/qwen_vl_interface': 1.5368784213442653e-06, 'learning_rate/action_model': 1.5368784213442654e-05, 'epoch': 2.02}) +[2026-06-18 04:41:08,946][starVLA.training.train_starvla][INFO] - Step 5522, Loss: {'train/loss': 0.14959204196929932, 'train/grad_norm_pre_clip': 18.052684783935547, 'timing/data': 0.01262053195387125, 'timing/model': 1.2760544810444117, 'learning_rate/qwen_vl_interface': 1.5355300608157172e-06, 'learning_rate/action_model': 1.5355300608157173e-05, 'epoch': 2.02}) +[2026-06-18 04:41:13,580][starVLA.training.train_starvla][INFO] - Step 5523, Loss: {'train/loss': 0.4233224093914032, 'train/grad_norm_pre_clip': 19.728416442871094, 'timing/data': 0.015742461197078228, 'timing/model': 1.0955874752253294, 'learning_rate/qwen_vl_interface': 1.5341824703007854e-06, 'learning_rate/action_model': 1.5341824703007854e-05, 'epoch': 2.02}) +[2026-06-18 04:41:18,277][starVLA.training.train_starvla][INFO] - Step 5524, Loss: {'train/loss': 0.3459632396697998, 'train/grad_norm_pre_clip': 51.80705261230469, 'timing/data': 0.010938328690826893, 'timing/model': 1.1045389594510198, 'learning_rate/qwen_vl_interface': 1.5328356500788263e-06, 'learning_rate/action_model': 1.5328356500788264e-05, 'epoch': 2.02}) +[2026-06-18 04:41:24,437][starVLA.training.train_starvla][INFO] - Step 5525, Loss: {'train/loss': 0.2960687577724457, 'train/grad_norm_pre_clip': 19.395719528198242, 'timing/data': 0.011397122405469418, 'timing/model': 2.670908296480775, 'learning_rate/qwen_vl_interface': 1.5314896004290366e-06, 'learning_rate/action_model': 1.5314896004290365e-05, 'epoch': 2.02}) +[2026-06-18 04:41:29,343][starVLA.training.train_starvla][INFO] - Step 5526, Loss: {'train/loss': 0.4739517867565155, 'train/grad_norm_pre_clip': 40.37493133544922, 'timing/data': 0.011434683576226234, 'timing/model': 1.5940165659412742, 'learning_rate/qwen_vl_interface': 1.530144321630454e-06, 'learning_rate/action_model': 1.530144321630454e-05, 'epoch': 2.02}) +[2026-06-18 04:41:33,624][starVLA.training.train_starvla][INFO] - Step 5527, Loss: {'train/loss': 0.4342495799064636, 'train/grad_norm_pre_clip': 22.38554573059082, 'timing/data': 0.012002069503068924, 'timing/model': 1.100414127111435, 'learning_rate/qwen_vl_interface': 1.5287998139619564e-06, 'learning_rate/action_model': 1.5287998139619562e-05, 'epoch': 2.02}) +[2026-06-18 04:41:38,247][starVLA.training.train_starvla][INFO] - Step 5528, Loss: {'train/loss': 0.28539860248565674, 'train/grad_norm_pre_clip': 22.09571075439453, 'timing/data': 0.016121329739689827, 'timing/model': 1.1958804717287421, 'learning_rate/qwen_vl_interface': 1.5274560777022613e-06, 'learning_rate/action_model': 1.5274560777022613e-05, 'epoch': 2.02}) +[2026-06-18 04:41:43,485][starVLA.training.train_starvla][INFO] - Step 5529, Loss: {'train/loss': 0.4301815927028656, 'train/grad_norm_pre_clip': 25.76601219177246, 'timing/data': 0.012404270470142365, 'timing/model': 1.220174572430551, 'learning_rate/qwen_vl_interface': 1.5261131131299274e-06, 'learning_rate/action_model': 1.5261131131299272e-05, 'epoch': 2.02}) +[2026-06-18 04:41:48,112][starVLA.training.train_starvla][INFO] - Step 5530, Loss: {'train/loss': 0.3504581153392792, 'train/grad_norm_pre_clip': 22.146438598632812, 'timing/data': 0.010501622222363949, 'timing/model': 1.057733236812055, 'learning_rate/qwen_vl_interface': 1.5247709205233518e-06, 'learning_rate/action_model': 1.5247709205233518e-05, 'epoch': 2.02}) +[2026-06-18 04:42:01,073][starVLA.training.train_starvla][INFO] - Step 5531, Loss: {'train/loss': 0.20271126925945282, 'train/grad_norm_pre_clip': 20.612335205078125, 'timing/data': 0.01245033834129572, 'timing/model': 6.176011296920478, 'learning_rate/qwen_vl_interface': 1.5234295001607732e-06, 'learning_rate/action_model': 1.5234295001607732e-05, 'epoch': 2.03}) +[2026-06-18 04:42:05,362][starVLA.training.train_starvla][INFO] - Step 5532, Loss: {'train/loss': 0.2122081071138382, 'train/grad_norm_pre_clip': 21.4807071685791, 'timing/data': 0.017185050062835217, 'timing/model': 1.1459457715973258, 'learning_rate/qwen_vl_interface': 1.5220888523202687e-06, 'learning_rate/action_model': 1.5220888523202686e-05, 'epoch': 2.03}) +[2026-06-18 04:42:10,300][starVLA.training.train_starvla][INFO] - Step 5533, Loss: {'train/loss': 0.5517175197601318, 'train/grad_norm_pre_clip': 43.02570724487305, 'timing/data': 0.011129630729556084, 'timing/model': 1.1851408332586288, 'learning_rate/qwen_vl_interface': 1.5207489772797557e-06, 'learning_rate/action_model': 1.5207489772797556e-05, 'epoch': 2.03}) +[2026-06-18 04:42:16,689][starVLA.training.train_starvla][INFO] - Step 5534, Loss: {'train/loss': 0.3567356467247009, 'train/grad_norm_pre_clip': 21.368314743041992, 'timing/data': 0.011205791495740414, 'timing/model': 1.0847650924697518, 'learning_rate/qwen_vl_interface': 1.5194098753169933e-06, 'learning_rate/action_model': 1.5194098753169933e-05, 'epoch': 2.03}) +[2026-06-18 04:42:21,396][starVLA.training.train_starvla][INFO] - Step 5535, Loss: {'train/loss': 0.5256853699684143, 'train/grad_norm_pre_clip': 36.30100631713867, 'timing/data': 0.011384788900613785, 'timing/model': 1.1033351700752974, 'learning_rate/qwen_vl_interface': 1.5180715467095799e-06, 'learning_rate/action_model': 1.5180715467095798e-05, 'epoch': 2.03}) +[2026-06-18 04:42:26,085][starVLA.training.train_starvla][INFO] - Step 5536, Loss: {'train/loss': 0.43625274300575256, 'train/grad_norm_pre_clip': 28.499446868896484, 'timing/data': 0.011203210800886154, 'timing/model': 1.4748471155762672, 'learning_rate/qwen_vl_interface': 1.5167339917349482e-06, 'learning_rate/action_model': 1.5167339917349482e-05, 'epoch': 2.03}) +[2026-06-18 04:42:30,431][starVLA.training.train_starvla][INFO] - Step 5537, Loss: {'train/loss': 0.21707729995250702, 'train/grad_norm_pre_clip': 20.668222427368164, 'timing/data': 0.015634281560778618, 'timing/model': 1.223129153251648, 'learning_rate/qwen_vl_interface': 1.5153972106703778e-06, 'learning_rate/action_model': 1.5153972106703779e-05, 'epoch': 2.03}) +[2026-06-18 04:42:36,077][starVLA.training.train_starvla][INFO] - Step 5538, Loss: {'train/loss': 0.33288586139678955, 'train/grad_norm_pre_clip': 19.601966857910156, 'timing/data': 0.013793310150504112, 'timing/model': 1.232273613102734, 'learning_rate/qwen_vl_interface': 1.5140612037929844e-06, 'learning_rate/action_model': 1.5140612037929844e-05, 'epoch': 2.03}) +[2026-06-18 04:42:40,744][starVLA.training.train_starvla][INFO] - Step 5539, Loss: {'train/loss': 0.3424154222011566, 'train/grad_norm_pre_clip': 30.60627555847168, 'timing/data': 0.011549664661288261, 'timing/model': 1.1187300849705935, 'learning_rate/qwen_vl_interface': 1.512725971379724e-06, 'learning_rate/action_model': 1.5127259713797239e-05, 'epoch': 2.03}) +[2026-06-18 04:42:50,674][starVLA.training.train_starvla][INFO] - Step 5540, Loss: {'train/loss': 0.5151575803756714, 'train/grad_norm_pre_clip': 26.67247200012207, 'timing/data': 0.011145494878292084, 'timing/model': 1.1974707106128335, 'learning_rate/qwen_vl_interface': 1.5113915137073905e-06, 'learning_rate/action_model': 1.5113915137073906e-05, 'epoch': 2.03}) +[2026-06-18 04:42:55,048][starVLA.training.train_starvla][INFO] - Step 5541, Loss: {'train/loss': 0.08674031496047974, 'train/grad_norm_pre_clip': 8.339675903320312, 'timing/data': 0.01269047986716032, 'timing/model': 1.151237541809678, 'learning_rate/qwen_vl_interface': 1.5100578310526198e-06, 'learning_rate/action_model': 1.5100578310526197e-05, 'epoch': 2.03}) +[2026-06-18 04:42:59,987][starVLA.training.train_starvla][INFO] - Step 5542, Loss: {'train/loss': 0.29046696424484253, 'train/grad_norm_pre_clip': 24.31784439086914, 'timing/data': 0.012992413714528084, 'timing/model': 1.215627770870924, 'learning_rate/qwen_vl_interface': 1.508724923691885e-06, 'learning_rate/action_model': 1.5087249236918848e-05, 'epoch': 2.03}) +[2026-06-18 04:43:05,681][starVLA.training.train_starvla][INFO] - Step 5543, Loss: {'train/loss': 0.39389970898628235, 'train/grad_norm_pre_clip': 24.713706970214844, 'timing/data': 0.022107252851128578, 'timing/model': 1.0862305257469416, 'learning_rate/qwen_vl_interface': 1.5073927919014986e-06, 'learning_rate/action_model': 1.5073927919014985e-05, 'epoch': 2.03}) +[2026-06-18 04:43:10,149][starVLA.training.train_starvla][INFO] - Step 5544, Loss: {'train/loss': 0.3276422619819641, 'train/grad_norm_pre_clip': 39.5455436706543, 'timing/data': 0.011862361803650856, 'timing/model': 1.0930930394679308, 'learning_rate/qwen_vl_interface': 1.506061435957614e-06, 'learning_rate/action_model': 1.506061435957614e-05, 'epoch': 2.03}) +[2026-06-18 04:43:14,692][starVLA.training.train_starvla][INFO] - Step 5545, Loss: {'train/loss': 0.10536672919988632, 'train/grad_norm_pre_clip': 12.445940971374512, 'timing/data': 0.0124228335916996, 'timing/model': 1.4040572885423899, 'learning_rate/qwen_vl_interface': 1.5047308561362225e-06, 'learning_rate/action_model': 1.5047308561362225e-05, 'epoch': 2.03}) +[2026-06-18 04:43:19,483][starVLA.training.train_starvla][INFO] - Step 5546, Loss: {'train/loss': 0.3445165157318115, 'train/grad_norm_pre_clip': 26.961162567138672, 'timing/data': 0.015017043799161911, 'timing/model': 1.41150953900069, 'learning_rate/qwen_vl_interface': 1.5034010527131548e-06, 'learning_rate/action_model': 1.5034010527131549e-05, 'epoch': 2.03}) +[2026-06-18 04:43:23,682][starVLA.training.train_starvla][INFO] - Step 5547, Loss: {'train/loss': 0.3779231905937195, 'train/grad_norm_pre_clip': 21.699316024780273, 'timing/data': 0.01641952432692051, 'timing/model': 1.0755044221878052, 'learning_rate/qwen_vl_interface': 1.5020720259640792e-06, 'learning_rate/action_model': 1.5020720259640792e-05, 'epoch': 2.03}) +[2026-06-18 04:43:29,591][starVLA.training.train_starvla][INFO] - Step 5548, Loss: {'train/loss': 0.6325114965438843, 'train/grad_norm_pre_clip': 26.985593795776367, 'timing/data': 0.011019780300557613, 'timing/model': 2.2562379399314523, 'learning_rate/qwen_vl_interface': 1.5007437761645064e-06, 'learning_rate/action_model': 1.5007437761645063e-05, 'epoch': 2.03}) +[2026-06-18 04:43:35,337][starVLA.training.train_starvla][INFO] - Step 5549, Loss: {'train/loss': 0.437295526266098, 'train/grad_norm_pre_clip': 24.163551330566406, 'timing/data': 0.01103486493229866, 'timing/model': 1.1362234698608518, 'learning_rate/qwen_vl_interface': 1.4994163035897845e-06, 'learning_rate/action_model': 1.4994163035897846e-05, 'epoch': 2.03}) +[2026-06-18 04:43:40,029][starVLA.training.train_starvla][INFO] - Step 5550, Loss: {'train/loss': 0.11387289315462112, 'train/grad_norm_pre_clip': 14.13603687286377, 'timing/data': 0.012654609978199005, 'timing/model': 1.6205303287133574, 'learning_rate/qwen_vl_interface': 1.4980896085150976e-06, 'learning_rate/action_model': 1.4980896085150975e-05, 'epoch': 2.03}) +[2026-06-18 04:43:44,653][starVLA.training.train_starvla][INFO] - Step 5551, Loss: {'train/loss': 0.22101745009422302, 'train/grad_norm_pre_clip': 19.948135375976562, 'timing/data': 0.013327485881745815, 'timing/model': 1.4683632384985685, 'learning_rate/qwen_vl_interface': 1.4967636912154707e-06, 'learning_rate/action_model': 1.4967636912154706e-05, 'epoch': 2.03}) +[2026-06-18 04:43:48,929][starVLA.training.train_starvla][INFO] - Step 5552, Loss: {'train/loss': 0.23649650812149048, 'train/grad_norm_pre_clip': 18.352096557617188, 'timing/data': 0.011236658319830894, 'timing/model': 1.1531542148441076, 'learning_rate/qwen_vl_interface': 1.4954385519657701e-06, 'learning_rate/action_model': 1.49543855196577e-05, 'epoch': 2.03}) +[2026-06-18 04:43:53,448][starVLA.training.train_starvla][INFO] - Step 5553, Loss: {'train/loss': 0.1217438206076622, 'train/grad_norm_pre_clip': 11.002118110656738, 'timing/data': 0.010957542806863785, 'timing/model': 1.0691797155886889, 'learning_rate/qwen_vl_interface': 1.4941141910406984e-06, 'learning_rate/action_model': 1.4941141910406984e-05, 'epoch': 2.03}) +[2026-06-18 04:43:58,015][starVLA.training.train_starvla][INFO] - Step 5554, Loss: {'train/loss': 0.22871005535125732, 'train/grad_norm_pre_clip': 25.639022827148438, 'timing/data': 0.010906322859227657, 'timing/model': 1.0650378623977304, 'learning_rate/qwen_vl_interface': 1.4927906087147966e-06, 'learning_rate/action_model': 1.4927906087147967e-05, 'epoch': 2.03}) +[2026-06-18 04:44:02,629][starVLA.training.train_starvla][INFO] - Step 5555, Loss: {'train/loss': 0.42765823006629944, 'train/grad_norm_pre_clip': 31.69031524658203, 'timing/data': 0.010900341905653477, 'timing/model': 1.0795303666964173, 'learning_rate/qwen_vl_interface': 1.491467805262444e-06, 'learning_rate/action_model': 1.4914678052624439e-05, 'epoch': 2.03}) +[2026-06-18 04:44:07,221][starVLA.training.train_starvla][INFO] - Step 5556, Loss: {'train/loss': 0.35275089740753174, 'train/grad_norm_pre_clip': 19.214473724365234, 'timing/data': 0.015198465436697006, 'timing/model': 1.3628724832087755, 'learning_rate/qwen_vl_interface': 1.4901457809578598e-06, 'learning_rate/action_model': 1.4901457809578597e-05, 'epoch': 2.03}) +[2026-06-18 04:44:11,331][starVLA.training.train_starvla][INFO] - Step 5557, Loss: {'train/loss': 0.39842909574508667, 'train/grad_norm_pre_clip': 17.943208694458008, 'timing/data': 0.011907396838068962, 'timing/model': 1.0660741943866014, 'learning_rate/qwen_vl_interface': 1.4888245360751014e-06, 'learning_rate/action_model': 1.4888245360751013e-05, 'epoch': 2.03}) +[2026-06-18 04:44:15,736][starVLA.training.train_starvla][INFO] - Step 5558, Loss: {'train/loss': 0.2226247638463974, 'train/grad_norm_pre_clip': 21.944995880126953, 'timing/data': 0.014104106463491917, 'timing/model': 1.0480556096881628, 'learning_rate/qwen_vl_interface': 1.4875040708880631e-06, 'learning_rate/action_model': 1.487504070888063e-05, 'epoch': 2.04}) +[2026-06-18 04:44:21,589][starVLA.training.train_starvla][INFO] - Step 5559, Loss: {'train/loss': 0.49090510606765747, 'train/grad_norm_pre_clip': 29.43210220336914, 'timing/data': 0.012859023176133633, 'timing/model': 1.0896558333188295, 'learning_rate/qwen_vl_interface': 1.4861843856704802e-06, 'learning_rate/action_model': 1.4861843856704801e-05, 'epoch': 2.04}) +[2026-06-18 04:44:26,081][starVLA.training.train_starvla][INFO] - Step 5560, Loss: {'train/loss': 0.5420780181884766, 'train/grad_norm_pre_clip': 34.09223937988281, 'timing/data': 0.010783739387989044, 'timing/model': 1.0638304371386766, 'learning_rate/qwen_vl_interface': 1.4848654806959243e-06, 'learning_rate/action_model': 1.4848654806959242e-05, 'epoch': 2.04}) +[2026-06-18 04:44:30,590][starVLA.training.train_starvla][INFO] - Step 5561, Loss: {'train/loss': 0.44016310572624207, 'train/grad_norm_pre_clip': 26.027938842773438, 'timing/data': 0.01105534564703703, 'timing/model': 1.3849313231185079, 'learning_rate/qwen_vl_interface': 1.483547356237805e-06, 'learning_rate/action_model': 1.4835473562378049e-05, 'epoch': 2.04}) +[2026-06-18 04:44:35,691][starVLA.training.train_starvla][INFO] - Step 5562, Loss: {'train/loss': 0.2193882167339325, 'train/grad_norm_pre_clip': 20.033308029174805, 'timing/data': 0.011045947670936584, 'timing/model': 1.8593471609055996, 'learning_rate/qwen_vl_interface': 1.4822300125693709e-06, 'learning_rate/action_model': 1.4822300125693708e-05, 'epoch': 2.04}) +[2026-06-18 04:44:43,452][starVLA.training.train_starvla][INFO] - Step 5563, Loss: {'train/loss': 0.2787370979785919, 'train/grad_norm_pre_clip': 22.092750549316406, 'timing/data': 0.012189928442239761, 'timing/model': 4.220702264457941, 'learning_rate/qwen_vl_interface': 1.480913449963712e-06, 'learning_rate/action_model': 1.480913449963712e-05, 'epoch': 2.04}) +[2026-06-18 04:44:48,124][starVLA.training.train_starvla][INFO] - Step 5564, Loss: {'train/loss': 0.05774356797337532, 'train/grad_norm_pre_clip': 6.710507392883301, 'timing/data': 0.011088622733950615, 'timing/model': 1.0591206150129437, 'learning_rate/qwen_vl_interface': 1.4795976686937488e-06, 'learning_rate/action_model': 1.4795976686937487e-05, 'epoch': 2.04}) +[2026-06-18 04:44:52,844][starVLA.training.train_starvla][INFO] - Step 5565, Loss: {'train/loss': 0.6516873836517334, 'train/grad_norm_pre_clip': 27.5347843170166, 'timing/data': 0.012891151942312717, 'timing/model': 1.1904796371236444, 'learning_rate/qwen_vl_interface': 1.4782826690322453e-06, 'learning_rate/action_model': 1.4782826690322451e-05, 'epoch': 2.04}) +[2026-06-18 04:44:57,394][starVLA.training.train_starvla][INFO] - Step 5566, Loss: {'train/loss': 0.1698370724916458, 'train/grad_norm_pre_clip': 15.09706974029541, 'timing/data': 0.01140244398266077, 'timing/model': 1.4354361817240715, 'learning_rate/qwen_vl_interface': 1.4769684512518045e-06, 'learning_rate/action_model': 1.4769684512518045e-05, 'epoch': 2.04}) +[2026-06-18 04:45:02,880][starVLA.training.train_starvla][INFO] - Step 5567, Loss: {'train/loss': 0.1445230394601822, 'train/grad_norm_pre_clip': 20.022579193115234, 'timing/data': 0.01238459162414074, 'timing/model': 1.0436809668317437, 'learning_rate/qwen_vl_interface': 1.4756550156248633e-06, 'learning_rate/action_model': 1.4756550156248632e-05, 'epoch': 2.04}) +[2026-06-18 04:45:07,366][starVLA.training.train_starvla][INFO] - Step 5568, Loss: {'train/loss': 0.2032497674226761, 'train/grad_norm_pre_clip': 20.665027618408203, 'timing/data': 0.010591006837785244, 'timing/model': 1.0991754531860352, 'learning_rate/qwen_vl_interface': 1.4743423624237e-06, 'learning_rate/action_model': 1.4743423624236998e-05, 'epoch': 2.04}) +[2026-06-18 04:45:11,989][starVLA.training.train_starvla][INFO] - Step 5569, Loss: {'train/loss': 0.29180851578712463, 'train/grad_norm_pre_clip': 18.327505111694336, 'timing/data': 0.011953259818255901, 'timing/model': 1.1216461015865207, 'learning_rate/qwen_vl_interface': 1.4730304919204253e-06, 'learning_rate/action_model': 1.4730304919204252e-05, 'epoch': 2.04}) +[2026-06-18 04:45:16,505][starVLA.training.train_starvla][INFO] - Step 5570, Loss: {'train/loss': 0.28046631813049316, 'train/grad_norm_pre_clip': 21.040678024291992, 'timing/data': 0.010682633146643639, 'timing/model': 1.0748310592025518, 'learning_rate/qwen_vl_interface': 1.4717194043869953e-06, 'learning_rate/action_model': 1.4717194043869953e-05, 'epoch': 2.04}) +[2026-06-18 04:45:20,949][starVLA.training.train_starvla][INFO] - Step 5571, Loss: {'train/loss': 0.5745077729225159, 'train/grad_norm_pre_clip': 32.37064743041992, 'timing/data': 0.013261360116302967, 'timing/model': 1.3335946388542652, 'learning_rate/qwen_vl_interface': 1.4704091000951982e-06, 'learning_rate/action_model': 1.4704091000951981e-05, 'epoch': 2.04}) +[2026-06-18 04:45:25,149][starVLA.training.train_starvla][INFO] - Step 5572, Loss: {'train/loss': 0.2089250236749649, 'train/grad_norm_pre_clip': 20.053680419921875, 'timing/data': 0.014704614877700806, 'timing/model': 1.0769418431445956, 'learning_rate/qwen_vl_interface': 1.4690995793166615e-06, 'learning_rate/action_model': 1.4690995793166616e-05, 'epoch': 2.04}) +[2026-06-18 04:45:29,817][starVLA.training.train_starvla][INFO] - Step 5573, Loss: {'train/loss': 0.3378746807575226, 'train/grad_norm_pre_clip': 24.697914123535156, 'timing/data': 0.014037669636309147, 'timing/model': 1.0971422791481018, 'learning_rate/qwen_vl_interface': 1.4677908423228506e-06, 'learning_rate/action_model': 1.4677908423228507e-05, 'epoch': 2.04}) +[2026-06-18 04:45:34,262][starVLA.training.train_starvla][INFO] - Step 5574, Loss: {'train/loss': 0.3966936469078064, 'train/grad_norm_pre_clip': 25.812257766723633, 'timing/data': 0.012225335463881493, 'timing/model': 1.0615344969555736, 'learning_rate/qwen_vl_interface': 1.4664828893850677e-06, 'learning_rate/action_model': 1.4664828893850676e-05, 'epoch': 2.04}) +[2026-06-18 04:45:38,832][starVLA.training.train_starvla][INFO] - Step 5575, Loss: {'train/loss': 0.35168972611427307, 'train/grad_norm_pre_clip': 24.513898849487305, 'timing/data': 0.01155147049576044, 'timing/model': 1.0478700092062354, 'learning_rate/qwen_vl_interface': 1.4651757207744539e-06, 'learning_rate/action_model': 1.4651757207744537e-05, 'epoch': 2.04}) +[2026-06-18 04:45:42,856][starVLA.training.train_starvla][INFO] - Step 5576, Loss: {'train/loss': 0.6021009087562561, 'train/grad_norm_pre_clip': 27.180997848510742, 'timing/data': 0.010818194597959518, 'timing/model': 1.0200306549668312, 'learning_rate/qwen_vl_interface': 1.4638693367619837e-06, 'learning_rate/action_model': 1.4638693367619834e-05, 'epoch': 2.04}) +[2026-06-18 04:45:47,385][starVLA.training.train_starvla][INFO] - Step 5577, Loss: {'train/loss': 0.23022358119487762, 'train/grad_norm_pre_clip': 19.568431854248047, 'timing/data': 0.012045050971210003, 'timing/model': 1.2016955334693193, 'learning_rate/qwen_vl_interface': 1.4625637376184768e-06, 'learning_rate/action_model': 1.4625637376184768e-05, 'epoch': 2.04}) +[2026-06-18 04:45:51,893][starVLA.training.train_starvla][INFO] - Step 5578, Loss: {'train/loss': 0.5113831758499146, 'train/grad_norm_pre_clip': 25.872194290161133, 'timing/data': 0.011100184172391891, 'timing/model': 1.0441916910931468, 'learning_rate/qwen_vl_interface': 1.4612589236145811e-06, 'learning_rate/action_model': 1.461258923614581e-05, 'epoch': 2.04}) +[2026-06-18 04:45:56,278][starVLA.training.train_starvla][INFO] - Step 5579, Loss: {'train/loss': 0.2548283338546753, 'train/grad_norm_pre_clip': 20.118446350097656, 'timing/data': 0.012224237434566021, 'timing/model': 1.0436368025839329, 'learning_rate/qwen_vl_interface': 1.4599548950207878e-06, 'learning_rate/action_model': 1.4599548950207878e-05, 'epoch': 2.04}) +[2026-06-18 04:46:00,815][starVLA.training.train_starvla][INFO] - Step 5580, Loss: {'train/loss': 0.35225746035575867, 'train/grad_norm_pre_clip': 24.211153030395508, 'timing/data': 0.014378264546394348, 'timing/model': 1.4190096985548735, 'learning_rate/qwen_vl_interface': 1.4586516521074214e-06, 'learning_rate/action_model': 1.4586516521074214e-05, 'epoch': 2.04}) +[2026-06-18 04:46:05,152][starVLA.training.train_starvla][INFO] - Step 5581, Loss: {'train/loss': 0.1664438396692276, 'train/grad_norm_pre_clip': 19.217172622680664, 'timing/data': 0.010037475265562534, 'timing/model': 1.3391949879005551, 'learning_rate/qwen_vl_interface': 1.457349195144649e-06, 'learning_rate/action_model': 1.457349195144649e-05, 'epoch': 2.04}) +[2026-06-18 04:46:09,414][starVLA.training.train_starvla][INFO] - Step 5582, Loss: {'train/loss': 0.328022837638855, 'train/grad_norm_pre_clip': 23.098169326782227, 'timing/data': 0.011244171299040318, 'timing/model': 1.0760657181963325, 'learning_rate/qwen_vl_interface': 1.4560475244024702e-06, 'learning_rate/action_model': 1.4560475244024702e-05, 'epoch': 2.04}) +[2026-06-18 04:46:14,146][starVLA.training.train_starvla][INFO] - Step 5583, Loss: {'train/loss': 0.12314876168966293, 'train/grad_norm_pre_clip': 14.691995620727539, 'timing/data': 0.012103119865059853, 'timing/model': 1.225328827276826, 'learning_rate/qwen_vl_interface': 1.4547466401507213e-06, 'learning_rate/action_model': 1.4547466401507211e-05, 'epoch': 2.04}) +[2026-06-18 04:46:18,797][starVLA.training.train_starvla][INFO] - Step 5584, Loss: {'train/loss': 0.35373160243034363, 'train/grad_norm_pre_clip': 24.011865615844727, 'timing/data': 0.012096886523067951, 'timing/model': 1.0512190908193588, 'learning_rate/qwen_vl_interface': 1.4534465426590763e-06, 'learning_rate/action_model': 1.4534465426590763e-05, 'epoch': 2.04}) +[2026-06-18 04:46:25,042][starVLA.training.train_starvla][INFO] - Step 5585, Loss: {'train/loss': 0.30048108100891113, 'train/grad_norm_pre_clip': 21.002193450927734, 'timing/data': 0.011534004472196102, 'timing/model': 1.5039320793002844, 'learning_rate/qwen_vl_interface': 1.4521472321970496e-06, 'learning_rate/action_model': 1.4521472321970494e-05, 'epoch': 2.05}) +[2026-06-18 04:46:32,165][starVLA.training.train_starvla][INFO] - Step 5586, Loss: {'train/loss': 0.1913282573223114, 'train/grad_norm_pre_clip': 16.466989517211914, 'timing/data': 0.012166434898972511, 'timing/model': 1.429085317067802, 'learning_rate/qwen_vl_interface': 1.450848709033988e-06, 'learning_rate/action_model': 1.4508487090339878e-05, 'epoch': 2.05}) +[2026-06-18 04:46:46,115][starVLA.training.train_starvla][INFO] - Step 5587, Loss: {'train/loss': 0.21327392756938934, 'train/grad_norm_pre_clip': 17.802692413330078, 'timing/data': 0.017211610451340675, 'timing/model': 4.52534063346684, 'learning_rate/qwen_vl_interface': 1.4495509734390769e-06, 'learning_rate/action_model': 1.4495509734390769e-05, 'epoch': 2.05}) +[2026-06-18 04:46:50,642][starVLA.training.train_starvla][INFO] - Step 5588, Loss: {'train/loss': 0.3997446894645691, 'train/grad_norm_pre_clip': 28.24533462524414, 'timing/data': 0.01640892866998911, 'timing/model': 1.1485106106847525, 'learning_rate/qwen_vl_interface': 1.448254025681339e-06, 'learning_rate/action_model': 1.448254025681339e-05, 'epoch': 2.05}) +[2026-06-18 04:46:57,183][starVLA.training.train_starvla][INFO] - Step 5589, Loss: {'train/loss': 0.2293606400489807, 'train/grad_norm_pre_clip': 17.36412239074707, 'timing/data': 0.01101445872336626, 'timing/model': 3.013624503277242, 'learning_rate/qwen_vl_interface': 1.4469578660296327e-06, 'learning_rate/action_model': 1.4469578660296326e-05, 'epoch': 2.05}) +[2026-06-18 04:47:05,100][starVLA.training.train_starvla][INFO] - Step 5590, Loss: {'train/loss': 0.17029350996017456, 'train/grad_norm_pre_clip': 22.147708892822266, 'timing/data': 0.011137396097183228, 'timing/model': 1.0837633423507214, 'learning_rate/qwen_vl_interface': 1.4456624947526524e-06, 'learning_rate/action_model': 1.4456624947526524e-05, 'epoch': 2.05}) +[2026-06-18 04:47:11,480][starVLA.training.train_starvla][INFO] - Step 5591, Loss: {'train/loss': 0.39713558554649353, 'train/grad_norm_pre_clip': 31.008255004882812, 'timing/data': 0.01281586941331625, 'timing/model': 1.507802010513842, 'learning_rate/qwen_vl_interface': 1.4443679121189306e-06, 'learning_rate/action_model': 1.4443679121189305e-05, 'epoch': 2.05}) +[2026-06-18 04:47:15,791][starVLA.training.train_starvla][INFO] - Step 5592, Loss: {'train/loss': 0.41071817278862, 'train/grad_norm_pre_clip': 23.792224884033203, 'timing/data': 0.010902157984673977, 'timing/model': 1.0700547443702817, 'learning_rate/qwen_vl_interface': 1.4430741183968361e-06, 'learning_rate/action_model': 1.443074118396836e-05, 'epoch': 2.05}) +[2026-06-18 04:47:20,359][starVLA.training.train_starvla][INFO] - Step 5593, Loss: {'train/loss': 0.27570873498916626, 'train/grad_norm_pre_clip': 20.751262664794922, 'timing/data': 0.010857665911316872, 'timing/model': 1.0888976147398353, 'learning_rate/qwen_vl_interface': 1.441781113854573e-06, 'learning_rate/action_model': 1.4417811138545728e-05, 'epoch': 2.05}) +[2026-06-18 04:47:24,919][starVLA.training.train_starvla][INFO] - Step 5594, Loss: {'train/loss': 0.25532591342926025, 'train/grad_norm_pre_clip': 24.192245483398438, 'timing/data': 0.01216008048504591, 'timing/model': 1.060034834779799, 'learning_rate/qwen_vl_interface': 1.4404888987601828e-06, 'learning_rate/action_model': 1.4404888987601828e-05, 'epoch': 2.05}) +[2026-06-18 04:47:29,513][starVLA.training.train_starvla][INFO] - Step 5595, Loss: {'train/loss': 0.45245322585105896, 'train/grad_norm_pre_clip': 23.492908477783203, 'timing/data': 0.020201501436531544, 'timing/model': 1.094531581737101, 'learning_rate/qwen_vl_interface': 1.4391974733815417e-06, 'learning_rate/action_model': 1.4391974733815416e-05, 'epoch': 2.05}) +[2026-06-18 04:47:33,896][starVLA.training.train_starvla][INFO] - Step 5596, Loss: {'train/loss': 0.3254939019680023, 'train/grad_norm_pre_clip': 30.917144775390625, 'timing/data': 0.012195650488138199, 'timing/model': 1.30981301702559, 'learning_rate/qwen_vl_interface': 1.437906837986368e-06, 'learning_rate/action_model': 1.4379068379863677e-05, 'epoch': 2.05}) +[2026-06-18 04:47:38,175][starVLA.training.train_starvla][INFO] - Step 5597, Loss: {'train/loss': 0.15062138438224792, 'train/grad_norm_pre_clip': 25.789844512939453, 'timing/data': 0.010815629735589027, 'timing/model': 1.1059064771980047, 'learning_rate/qwen_vl_interface': 1.436616992842207e-06, 'learning_rate/action_model': 1.4366169928422068e-05, 'epoch': 2.05}) +[2026-06-18 04:47:42,961][starVLA.training.train_starvla][INFO] - Step 5598, Loss: {'train/loss': 0.09968417882919312, 'train/grad_norm_pre_clip': 21.856191635131836, 'timing/data': 0.012056015431880951, 'timing/model': 1.1687238654121757, 'learning_rate/qwen_vl_interface': 1.4353279382164459e-06, 'learning_rate/action_model': 1.4353279382164458e-05, 'epoch': 2.05}) +[2026-06-18 04:47:47,506][starVLA.training.train_starvla][INFO] - Step 5599, Loss: {'train/loss': 0.18378466367721558, 'train/grad_norm_pre_clip': 12.502096176147461, 'timing/data': 0.011308562010526657, 'timing/model': 1.0438596652820706, 'learning_rate/qwen_vl_interface': 1.4340396743763084e-06, 'learning_rate/action_model': 1.4340396743763085e-05, 'epoch': 2.05}) +[2026-06-18 04:50:48,449][starVLA.training.train_starvla][INFO] - Step 5600, Loss: {'train/loss': 0.3276122808456421, 'train/grad_norm_pre_clip': 16.259084701538086, 'eval/loss': 0.5314427614212036, 'eval/latency_2/loss': 0.5314427614212036, 'eval/demon_fire_precision': 0.943163097199341, 'eval/demon_fire_recall': 0.9581589958158996, 'eval/demon_fire_f1': 0.9506019095060191, 'eval/demon_fire_rate_model': 0.6755703951029494, 'eval/demon_fire_rate_teacher': 0.6649972175848636, 'eval/demon_left_precision': 0.8975966562173459, 'eval/demon_left_recall': 0.9042105263157895, 'eval/demon_left_f1': 0.9008914525432616, 'eval/demon_left_rate_model': 0.2662771285475793, 'eval/demon_left_rate_teacher': 0.2643294379521425, 'eval/demon_right_precision': 0.8859934853420195, 'eval/demon_right_recall': 0.9220338983050848, 'eval/demon_right_f1': 0.9036544850498338, 'eval/demon_right_rate_model': 0.25626043405676124, 'eval/demon_right_rate_teacher': 0.24624373956594323, 'eval/demon_move_f1': 0.9022729687965477, 'eval/demon_cc_f1': 0.9355682040225728, 'eval/latency_2/demon_fire_precision': 0.943163097199341, 'eval/latency_2/demon_fire_recall': 0.9581589958158996, 'eval/latency_2/demon_fire_f1': 0.9506019095060191, 'eval/latency_2/demon_fire_rate_model': 0.6755703951029494, 'eval/latency_2/demon_fire_rate_teacher': 0.6649972175848636, 'eval/latency_2/demon_left_precision': 0.8975966562173459, 'eval/latency_2/demon_left_recall': 0.9042105263157895, 'eval/latency_2/demon_left_f1': 0.9008914525432616, 'eval/latency_2/demon_left_rate_model': 0.2662771285475793, 'eval/latency_2/demon_left_rate_teacher': 0.2643294379521425, 'eval/latency_2/demon_right_precision': 0.8859934853420195, 'eval/latency_2/demon_right_recall': 0.9220338983050848, 'eval/latency_2/demon_right_f1': 0.9036544850498338, 'eval/latency_2/demon_right_rate_model': 0.25626043405676124, 'eval/latency_2/demon_right_rate_teacher': 0.24624373956594323, 'eval/latency_2/demon_move_f1': 0.9022729687965477, 'eval/latency_2/demon_cc_f1': 0.9355682040225728, 'timing/data': 0.014863533899188042, 'timing/model': 1.0804432891309261, 'learning_rate/qwen_vl_interface': 1.4327522015888535e-06, 'learning_rate/action_model': 1.4327522015888534e-05, 'epoch': 2.05}) +[2026-06-18 04:50:48,452][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_5600_state +[2026-06-18 04:50:48,469][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 04:50:57,194][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_5600_state/model.safetensors +[2026-06-18 04:51:12,539][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_5600_state/optimizer.bin +[2026-06-18 04:51:12,541][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_5600_state/sampler.bin +[2026-06-18 04:51:12,541][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_5600_state/sampler_1.bin +[2026-06-18 04:51:12,543][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_5600_state/random_states_0.pkl +[2026-06-18 04:51:12,555][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 5600 +[2026-06-18 04:51:12,555][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 04:51:12,571][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 04:51:17,418][starVLA.training.train_starvla][INFO] - Step 5601, Loss: {'train/loss': 0.09833201766014099, 'train/grad_norm_pre_clip': 18.697404861450195, 'timing/data': 0.011904002167284489, 'timing/model': 1.0690794866532087, 'learning_rate/qwen_vl_interface': 1.4314655201209756e-06, 'learning_rate/action_model': 1.4314655201209757e-05, 'epoch': 2.05}) +[2026-06-18 04:51:22,048][starVLA.training.train_starvla][INFO] - Step 5602, Loss: {'train/loss': 0.13642725348472595, 'train/grad_norm_pre_clip': 11.954992294311523, 'timing/data': 0.012924250215291977, 'timing/model': 1.4705442460253835, 'learning_rate/qwen_vl_interface': 1.4301796302394019e-06, 'learning_rate/action_model': 1.4301796302394018e-05, 'epoch': 2.05}) +[2026-06-18 04:51:27,211][starVLA.training.train_starvla][INFO] - Step 5603, Loss: {'train/loss': 0.2722231149673462, 'train/grad_norm_pre_clip': 25.549474716186523, 'timing/data': 0.012078830972313881, 'timing/model': 1.4145692382007837, 'learning_rate/qwen_vl_interface': 1.4288945322107025e-06, 'learning_rate/action_model': 1.4288945322107026e-05, 'epoch': 2.05}) +[2026-06-18 04:51:31,698][starVLA.training.train_starvla][INFO] - Step 5604, Loss: {'train/loss': 0.4641907513141632, 'train/grad_norm_pre_clip': 33.03025436401367, 'timing/data': 0.010769001208245754, 'timing/model': 1.0961125250905752, 'learning_rate/qwen_vl_interface': 1.4276102263012781e-06, 'learning_rate/action_model': 1.427610226301278e-05, 'epoch': 2.05}) +[2026-06-18 04:51:36,428][starVLA.training.train_starvla][INFO] - Step 5605, Loss: {'train/loss': 0.12508971989154816, 'train/grad_norm_pre_clip': 14.549047470092773, 'timing/data': 0.012176594696938992, 'timing/model': 1.100666081532836, 'learning_rate/qwen_vl_interface': 1.426326712777367e-06, 'learning_rate/action_model': 1.4263267127773669e-05, 'epoch': 2.05}) +[2026-06-18 04:51:41,010][starVLA.training.train_starvla][INFO] - Step 5606, Loss: {'train/loss': 0.19477888941764832, 'train/grad_norm_pre_clip': 21.22080421447754, 'timing/data': 0.013777830637991428, 'timing/model': 1.1693684551864862, 'learning_rate/qwen_vl_interface': 1.425043991905043e-06, 'learning_rate/action_model': 1.425043991905043e-05, 'epoch': 2.05}) +[2026-06-18 04:51:45,412][starVLA.training.train_starvla][INFO] - Step 5607, Loss: {'train/loss': 0.3907216787338257, 'train/grad_norm_pre_clip': 25.551576614379883, 'timing/data': 0.011165766976773739, 'timing/model': 1.3974896520376205, 'learning_rate/qwen_vl_interface': 1.4237620639502149e-06, 'learning_rate/action_model': 1.4237620639502147e-05, 'epoch': 2.05}) +[2026-06-18 04:51:50,367][starVLA.training.train_starvla][INFO] - Step 5608, Loss: {'train/loss': 0.7916536927223206, 'train/grad_norm_pre_clip': 31.931547164916992, 'timing/data': 0.015324123203754425, 'timing/model': 1.8756587663665414, 'learning_rate/qwen_vl_interface': 1.4224809291786281e-06, 'learning_rate/action_model': 1.4224809291786282e-05, 'epoch': 2.05}) +[2026-06-18 04:51:54,901][starVLA.training.train_starvla][INFO] - Step 5609, Loss: {'train/loss': 0.2092456817626953, 'train/grad_norm_pre_clip': 18.420167922973633, 'timing/data': 0.011146477423608303, 'timing/model': 1.0428862497210503, 'learning_rate/qwen_vl_interface': 1.4212005878558622e-06, 'learning_rate/action_model': 1.4212005878558622e-05, 'epoch': 2.05}) +[2026-06-18 04:51:59,273][starVLA.training.train_starvla][INFO] - Step 5610, Loss: {'train/loss': 0.0866226851940155, 'train/grad_norm_pre_clip': 8.946402549743652, 'timing/data': 0.011198271997272968, 'timing/model': 1.031095146201551, 'learning_rate/qwen_vl_interface': 1.4199210402473368e-06, 'learning_rate/action_model': 1.4199210402473367e-05, 'epoch': 2.05}) +[2026-06-18 04:52:03,673][starVLA.training.train_starvla][INFO] - Step 5611, Loss: {'train/loss': 0.14019706845283508, 'train/grad_norm_pre_clip': 11.345972061157227, 'timing/data': 0.011812166310846806, 'timing/model': 1.0510451337322593, 'learning_rate/qwen_vl_interface': 1.4186422866183e-06, 'learning_rate/action_model': 1.4186422866182999e-05, 'epoch': 2.05}) +[2026-06-18 04:52:11,195][starVLA.training.train_starvla][INFO] - Step 5612, Loss: {'train/loss': 0.3623322546482086, 'train/grad_norm_pre_clip': 28.576839447021484, 'timing/data': 0.012608209624886513, 'timing/model': 4.36726828571409, 'learning_rate/qwen_vl_interface': 1.41736432723384e-06, 'learning_rate/action_model': 1.41736432723384e-05, 'epoch': 2.05}) +[2026-06-18 04:52:16,730][starVLA.training.train_starvla][INFO] - Step 5613, Loss: {'train/loss': 0.2589648962020874, 'train/grad_norm_pre_clip': 17.966976165771484, 'timing/data': 0.011364011093974113, 'timing/model': 1.0525920372456312, 'learning_rate/qwen_vl_interface': 1.4160871623588774e-06, 'learning_rate/action_model': 1.4160871623588772e-05, 'epoch': 2.06}) +[2026-06-18 04:52:21,663][starVLA.training.train_starvla][INFO] - Step 5614, Loss: {'train/loss': 0.1478392630815506, 'train/grad_norm_pre_clip': 13.766657829284668, 'timing/data': 0.01672749314457178, 'timing/model': 1.1370141291990876, 'learning_rate/qwen_vl_interface': 1.414810792258173e-06, 'learning_rate/action_model': 1.414810792258173e-05, 'epoch': 2.06}) +[2026-06-18 04:52:26,326][starVLA.training.train_starvla][INFO] - Step 5615, Loss: {'train/loss': 0.2435772716999054, 'train/grad_norm_pre_clip': 25.219547271728516, 'timing/data': 0.012760309502482414, 'timing/model': 1.0647193510085344, 'learning_rate/qwen_vl_interface': 1.4135352171963186e-06, 'learning_rate/action_model': 1.4135352171963185e-05, 'epoch': 2.06}) +[2026-06-18 04:52:30,786][starVLA.training.train_starvla][INFO] - Step 5616, Loss: {'train/loss': 0.2725100815296173, 'train/grad_norm_pre_clip': 25.159000396728516, 'timing/data': 0.016594456508755684, 'timing/model': 1.3684216644614935, 'learning_rate/qwen_vl_interface': 1.412260437437741e-06, 'learning_rate/action_model': 1.412260437437741e-05, 'epoch': 2.06}) +[2026-06-18 04:52:35,315][starVLA.training.train_starvla][INFO] - Step 5617, Loss: {'train/loss': 0.16825462877750397, 'train/grad_norm_pre_clip': 16.877403259277344, 'timing/data': 0.012724974192678928, 'timing/model': 1.34715545643121, 'learning_rate/qwen_vl_interface': 1.4109864532467051e-06, 'learning_rate/action_model': 1.4109864532467052e-05, 'epoch': 2.06}) +[2026-06-18 04:52:39,555][starVLA.training.train_starvla][INFO] - Step 5618, Loss: {'train/loss': 0.32693061232566833, 'train/grad_norm_pre_clip': 21.095420837402344, 'timing/data': 0.013300763443112373, 'timing/model': 1.1028761202469468, 'learning_rate/qwen_vl_interface': 1.4097132648873082e-06, 'learning_rate/action_model': 1.4097132648873083e-05, 'epoch': 2.06}) +[2026-06-18 04:52:43,984][starVLA.training.train_starvla][INFO] - Step 5619, Loss: {'train/loss': 0.13817985355854034, 'train/grad_norm_pre_clip': 13.293335914611816, 'timing/data': 0.012774776667356491, 'timing/model': 1.0517314421012998, 'learning_rate/qwen_vl_interface': 1.4084408726234838e-06, 'learning_rate/action_model': 1.4084408726234838e-05, 'epoch': 2.06}) +[2026-06-18 04:52:48,542][starVLA.training.train_starvla][INFO] - Step 5620, Loss: {'train/loss': 0.41890186071395874, 'train/grad_norm_pre_clip': 26.921663284301758, 'timing/data': 0.015139985829591751, 'timing/model': 1.1198461819440126, 'learning_rate/qwen_vl_interface': 1.407169276719e-06, 'learning_rate/action_model': 1.4071692767190001e-05, 'epoch': 2.06}) +[2026-06-18 04:52:53,732][starVLA.training.train_starvla][INFO] - Step 5621, Loss: {'train/loss': 0.11895935237407684, 'train/grad_norm_pre_clip': 16.243709564208984, 'timing/data': 0.013207630254328251, 'timing/model': 1.0717958258464932, 'learning_rate/qwen_vl_interface': 1.4058984774374606e-06, 'learning_rate/action_model': 1.4058984774374607e-05, 'epoch': 2.06}) +[2026-06-18 04:52:58,465][starVLA.training.train_starvla][INFO] - Step 5622, Loss: {'train/loss': 0.1839413046836853, 'train/grad_norm_pre_clip': 14.238889694213867, 'timing/data': 0.010797301307320595, 'timing/model': 1.4622422885149717, 'learning_rate/qwen_vl_interface': 1.404628475042303e-06, 'learning_rate/action_model': 1.404628475042303e-05, 'epoch': 2.06}) +[2026-06-18 04:53:04,015][starVLA.training.train_starvla][INFO] - Step 5623, Loss: {'train/loss': 0.06601957231760025, 'train/grad_norm_pre_clip': 6.474334716796875, 'timing/data': 0.013442790135741234, 'timing/model': 1.1878917450085282, 'learning_rate/qwen_vl_interface': 1.4033592697967998e-06, 'learning_rate/action_model': 1.4033592697967998e-05, 'epoch': 2.06}) +[2026-06-18 04:53:08,708][starVLA.training.train_starvla][INFO] - Step 5624, Loss: {'train/loss': 0.3266817629337311, 'train/grad_norm_pre_clip': 26.055070877075195, 'timing/data': 0.012810732237994671, 'timing/model': 1.0402488689869642, 'learning_rate/qwen_vl_interface': 1.4020908619640602e-06, 'learning_rate/action_model': 1.4020908619640602e-05, 'epoch': 2.06}) +[2026-06-18 04:53:13,212][starVLA.training.train_starvla][INFO] - Step 5625, Loss: {'train/loss': 0.42064371705055237, 'train/grad_norm_pre_clip': 21.145729064941406, 'timing/data': 0.01291146595031023, 'timing/model': 1.0520104598253965, 'learning_rate/qwen_vl_interface': 1.4008232518070248e-06, 'learning_rate/action_model': 1.4008232518070247e-05, 'epoch': 2.06}) +[2026-06-18 04:53:17,688][starVLA.training.train_starvla][INFO] - Step 5626, Loss: {'train/loss': 0.1383880376815796, 'train/grad_norm_pre_clip': 17.82359504699707, 'timing/data': 0.014099057763814926, 'timing/model': 1.0661169532686472, 'learning_rate/qwen_vl_interface': 1.3995564395884712e-06, 'learning_rate/action_model': 1.399556439588471e-05, 'epoch': 2.06}) +[2026-06-18 04:53:22,827][starVLA.training.train_starvla][INFO] - Step 5627, Loss: {'train/loss': 0.2774527370929718, 'train/grad_norm_pre_clip': 19.53911590576172, 'timing/data': 0.02260439284145832, 'timing/model': 1.4360231030732393, 'learning_rate/qwen_vl_interface': 1.3982904255710094e-06, 'learning_rate/action_model': 1.3982904255710094e-05, 'epoch': 2.06}) +[2026-06-18 04:53:27,013][starVLA.training.train_starvla][INFO] - Step 5628, Loss: {'train/loss': 0.07433798909187317, 'train/grad_norm_pre_clip': 14.682368278503418, 'timing/data': 0.013803538866341114, 'timing/model': 1.0825570113956928, 'learning_rate/qwen_vl_interface': 1.3970252100170892e-06, 'learning_rate/action_model': 1.3970252100170892e-05, 'epoch': 2.06}) +[2026-06-18 04:53:31,543][starVLA.training.train_starvla][INFO] - Step 5629, Loss: {'train/loss': 0.19041548669338226, 'train/grad_norm_pre_clip': 15.902445793151855, 'timing/data': 0.01125330664217472, 'timing/model': 1.0625440366566181, 'learning_rate/qwen_vl_interface': 1.3957607931889885e-06, 'learning_rate/action_model': 1.3957607931889884e-05, 'epoch': 2.06}) +[2026-06-18 04:53:36,578][starVLA.training.train_starvla][INFO] - Step 5630, Loss: {'train/loss': 0.2257099151611328, 'train/grad_norm_pre_clip': 27.8824520111084, 'timing/data': 0.011907875537872314, 'timing/model': 1.3010764522477984, 'learning_rate/qwen_vl_interface': 1.3944971753488238e-06, 'learning_rate/action_model': 1.3944971753488236e-05, 'epoch': 2.06}) +[2026-06-18 04:53:41,121][starVLA.training.train_starvla][INFO] - Step 5631, Loss: {'train/loss': 0.46600741147994995, 'train/grad_norm_pre_clip': 19.34729766845703, 'timing/data': 0.013845955021679401, 'timing/model': 1.3993672225624323, 'learning_rate/qwen_vl_interface': 1.3932343567585408e-06, 'learning_rate/action_model': 1.3932343567585407e-05, 'epoch': 2.06}) +[2026-06-18 04:53:45,615][starVLA.training.train_starvla][INFO] - Step 5632, Loss: {'train/loss': 0.2946127653121948, 'train/grad_norm_pre_clip': 15.376972198486328, 'timing/data': 0.011276515200734138, 'timing/model': 1.3684153454378247, 'learning_rate/qwen_vl_interface': 1.3919723376799268e-06, 'learning_rate/action_model': 1.3919723376799269e-05, 'epoch': 2.06}) +[2026-06-18 04:53:49,930][starVLA.training.train_starvla][INFO] - Step 5633, Loss: {'train/loss': 0.07773476094007492, 'train/grad_norm_pre_clip': 10.39554214477539, 'timing/data': 0.01229953020811081, 'timing/model': 1.0681330785155296, 'learning_rate/qwen_vl_interface': 1.3907111183745992e-06, 'learning_rate/action_model': 1.3907111183745991e-05, 'epoch': 2.06}) +[2026-06-18 04:53:54,651][starVLA.training.train_starvla][INFO] - Step 5634, Loss: {'train/loss': 0.43233728408813477, 'train/grad_norm_pre_clip': 40.80606460571289, 'timing/data': 0.01332038827240467, 'timing/model': 1.113213101401925, 'learning_rate/qwen_vl_interface': 1.3894506991040089e-06, 'learning_rate/action_model': 1.3894506991040087e-05, 'epoch': 2.06}) +[2026-06-18 04:53:59,177][starVLA.training.train_starvla][INFO] - Step 5635, Loss: {'train/loss': 0.05535973981022835, 'train/grad_norm_pre_clip': 9.268916130065918, 'timing/data': 0.01238582469522953, 'timing/model': 1.0444301888346672, 'learning_rate/qwen_vl_interface': 1.3881910801294433e-06, 'learning_rate/action_model': 1.3881910801294431e-05, 'epoch': 2.06}) +[2026-06-18 04:54:03,908][starVLA.training.train_starvla][INFO] - Step 5636, Loss: {'train/loss': 0.229990616440773, 'train/grad_norm_pre_clip': 22.192842483520508, 'timing/data': 0.018619739450514317, 'timing/model': 1.0511848498135805, 'learning_rate/qwen_vl_interface': 1.3869322617120223e-06, 'learning_rate/action_model': 1.3869322617120222e-05, 'epoch': 2.06}) +[2026-06-18 04:54:08,434][starVLA.training.train_starvla][INFO] - Step 5637, Loss: {'train/loss': 0.406598299741745, 'train/grad_norm_pre_clip': 25.29145622253418, 'timing/data': 0.014702503569424152, 'timing/model': 1.5220927661284804, 'learning_rate/qwen_vl_interface': 1.3856742441127e-06, 'learning_rate/action_model': 1.3856742441126999e-05, 'epoch': 2.06}) +[2026-06-18 04:54:12,844][starVLA.training.train_starvla][INFO] - Step 5638, Loss: {'train/loss': 0.16373950242996216, 'train/grad_norm_pre_clip': 16.35984992980957, 'timing/data': 0.019459214992821217, 'timing/model': 1.2156949890777469, 'learning_rate/qwen_vl_interface': 1.384417027592264e-06, 'learning_rate/action_model': 1.3844170275922638e-05, 'epoch': 2.06}) +[2026-06-18 04:54:17,542][starVLA.training.train_starvla][INFO] - Step 5639, Loss: {'train/loss': 0.24931170046329498, 'train/grad_norm_pre_clip': 28.449726104736328, 'timing/data': 0.012437926605343819, 'timing/model': 1.2465019365772605, 'learning_rate/qwen_vl_interface': 1.3831606124113397e-06, 'learning_rate/action_model': 1.3831606124113398e-05, 'epoch': 2.06}) +[2026-06-18 04:54:22,001][starVLA.training.train_starvla][INFO] - Step 5640, Loss: {'train/loss': 0.2961864173412323, 'train/grad_norm_pre_clip': 17.648897171020508, 'timing/data': 0.011489770375192165, 'timing/model': 1.0465938374400139, 'learning_rate/qwen_vl_interface': 1.3819049988303809e-06, 'learning_rate/action_model': 1.381904998830381e-05, 'epoch': 2.07}) +[2026-06-18 04:54:26,480][starVLA.training.train_starvla][INFO] - Step 5641, Loss: {'train/loss': 0.1414007991552353, 'train/grad_norm_pre_clip': 16.930850982666016, 'timing/data': 0.013098904862999916, 'timing/model': 1.048087453469634, 'learning_rate/qwen_vl_interface': 1.3806501871096774e-06, 'learning_rate/action_model': 1.3806501871096774e-05, 'epoch': 2.07}) +[2026-06-18 04:54:31,116][starVLA.training.train_starvla][INFO] - Step 5642, Loss: {'train/loss': 0.36281755566596985, 'train/grad_norm_pre_clip': 25.594642639160156, 'timing/data': 0.013373368419706821, 'timing/model': 1.5640607299283147, 'learning_rate/qwen_vl_interface': 1.3793961775093528e-06, 'learning_rate/action_model': 1.3793961775093528e-05, 'epoch': 2.07}) +[2026-06-18 04:54:35,312][starVLA.training.train_starvla][INFO] - Step 5643, Loss: {'train/loss': 0.44879359006881714, 'train/grad_norm_pre_clip': 27.732563018798828, 'timing/data': 0.019626501016318798, 'timing/model': 1.0935229938477278, 'learning_rate/qwen_vl_interface': 1.378142970289367e-06, 'learning_rate/action_model': 1.378142970289367e-05, 'epoch': 2.07}) +[2026-06-18 04:54:40,123][starVLA.training.train_starvla][INFO] - Step 5644, Loss: {'train/loss': 0.16205988824367523, 'train/grad_norm_pre_clip': 15.605656623840332, 'timing/data': 0.012083086185157299, 'timing/model': 1.198848381638527, 'learning_rate/qwen_vl_interface': 1.3768905657095109e-06, 'learning_rate/action_model': 1.3768905657095107e-05, 'epoch': 2.07}) +[2026-06-18 04:54:44,851][starVLA.training.train_starvla][INFO] - Step 5645, Loss: {'train/loss': 0.2676621675491333, 'train/grad_norm_pre_clip': 27.000316619873047, 'timing/data': 0.01135699450969696, 'timing/model': 1.0730739701539278, 'learning_rate/qwen_vl_interface': 1.3756389640294066e-06, 'learning_rate/action_model': 1.3756389640294064e-05, 'epoch': 2.07}) +[2026-06-18 04:54:49,328][starVLA.training.train_starvla][INFO] - Step 5646, Loss: {'train/loss': 0.43990030884742737, 'train/grad_norm_pre_clip': 24.69969367980957, 'timing/data': 0.013328309170901775, 'timing/model': 1.0788619620725513, 'learning_rate/qwen_vl_interface': 1.3743881655085131e-06, 'learning_rate/action_model': 1.3743881655085131e-05, 'epoch': 2.07}) +[2026-06-18 04:54:53,985][starVLA.training.train_starvla][INFO] - Step 5647, Loss: {'train/loss': 0.24504634737968445, 'train/grad_norm_pre_clip': 21.35828399658203, 'timing/data': 0.012246565893292427, 'timing/model': 1.4547377843409777, 'learning_rate/qwen_vl_interface': 1.3731381704061248e-06, 'learning_rate/action_model': 1.3731381704061246e-05, 'epoch': 2.07}) +[2026-06-18 04:54:58,910][starVLA.training.train_starvla][INFO] - Step 5648, Loss: {'train/loss': 0.07672180980443954, 'train/grad_norm_pre_clip': 17.015399932861328, 'timing/data': 0.01383979432284832, 'timing/model': 1.8423173213377595, 'learning_rate/qwen_vl_interface': 1.3718889789813647e-06, 'learning_rate/action_model': 1.3718889789813646e-05, 'epoch': 2.07}) +[2026-06-18 04:55:04,232][starVLA.training.train_starvla][INFO] - Step 5649, Loss: {'train/loss': 0.49369290471076965, 'train/grad_norm_pre_clip': 33.39554214477539, 'timing/data': 0.014840388670563698, 'timing/model': 1.1548641556873918, 'learning_rate/qwen_vl_interface': 1.370640591493193e-06, 'learning_rate/action_model': 1.370640591493193e-05, 'epoch': 2.07}) +[2026-06-18 04:55:13,009][starVLA.training.train_starvla][INFO] - Step 5650, Loss: {'train/loss': 0.11462639272212982, 'train/grad_norm_pre_clip': 11.332304954528809, 'timing/data': 0.012292446568608284, 'timing/model': 1.1085891826078296, 'learning_rate/qwen_vl_interface': 1.3693930082004006e-06, 'learning_rate/action_model': 1.3693930082004006e-05, 'epoch': 2.07}) +[2026-06-18 04:55:17,600][starVLA.training.train_starvla][INFO] - Step 5651, Loss: {'train/loss': 0.44213587045669556, 'train/grad_norm_pre_clip': 28.739343643188477, 'timing/data': 0.016715805046260357, 'timing/model': 1.059967003762722, 'learning_rate/qwen_vl_interface': 1.3681462293616131e-06, 'learning_rate/action_model': 1.3681462293616131e-05, 'epoch': 2.07}) +[2026-06-18 04:55:22,080][starVLA.training.train_starvla][INFO] - Step 5652, Loss: {'train/loss': 0.3165595233440399, 'train/grad_norm_pre_clip': 28.42130470275879, 'timing/data': 0.012145987711846828, 'timing/model': 1.3446579212322831, 'learning_rate/qwen_vl_interface': 1.3669002552352898e-06, 'learning_rate/action_model': 1.3669002552352897e-05, 'epoch': 2.07}) +[2026-06-18 04:55:26,292][starVLA.training.train_starvla][INFO] - Step 5653, Loss: {'train/loss': 0.2544882297515869, 'train/grad_norm_pre_clip': 20.02223777770996, 'timing/data': 0.01105163712054491, 'timing/model': 1.0343977976590395, 'learning_rate/qwen_vl_interface': 1.3656550860797223e-06, 'learning_rate/action_model': 1.3656550860797223e-05, 'epoch': 2.07}) +[2026-06-18 04:55:30,705][starVLA.training.train_starvla][INFO] - Step 5654, Loss: {'train/loss': 0.14839932322502136, 'train/grad_norm_pre_clip': 12.055702209472656, 'timing/data': 0.0148336011916399, 'timing/model': 1.04857571516186, 'learning_rate/qwen_vl_interface': 1.3644107221530344e-06, 'learning_rate/action_model': 1.3644107221530345e-05, 'epoch': 2.07}) +[2026-06-18 04:55:35,426][starVLA.training.train_starvla][INFO] - Step 5655, Loss: {'train/loss': 0.10174393653869629, 'train/grad_norm_pre_clip': 10.627052307128906, 'timing/data': 0.011162837967276573, 'timing/model': 1.1323352297767997, 'learning_rate/qwen_vl_interface': 1.3631671637131856e-06, 'learning_rate/action_model': 1.3631671637131855e-05, 'epoch': 2.07}) +[2026-06-18 04:55:40,070][starVLA.training.train_starvla][INFO] - Step 5656, Loss: {'train/loss': 0.2850484848022461, 'train/grad_norm_pre_clip': 25.498750686645508, 'timing/data': 0.012062066234648228, 'timing/model': 1.0603139596059918, 'learning_rate/qwen_vl_interface': 1.3619244110179655e-06, 'learning_rate/action_model': 1.3619244110179654e-05, 'epoch': 2.07}) +[2026-06-18 04:55:44,232][starVLA.training.train_starvla][INFO] - Step 5657, Loss: {'train/loss': 0.3107934594154358, 'train/grad_norm_pre_clip': 21.95868682861328, 'timing/data': 0.0198777811601758, 'timing/model': 1.1086137294769287, 'learning_rate/qwen_vl_interface': 1.360682464324998e-06, 'learning_rate/action_model': 1.360682464324998e-05, 'epoch': 2.07}) +[2026-06-18 04:55:48,989][starVLA.training.train_starvla][INFO] - Step 5658, Loss: {'train/loss': 0.05702261999249458, 'train/grad_norm_pre_clip': 11.205528259277344, 'timing/data': 0.0188651904463768, 'timing/model': 1.3120069792494178, 'learning_rate/qwen_vl_interface': 1.3594413238917425e-06, 'learning_rate/action_model': 1.3594413238917425e-05, 'epoch': 2.07}) +[2026-06-18 04:55:53,534][starVLA.training.train_starvla][INFO] - Step 5659, Loss: {'train/loss': 0.2940613031387329, 'train/grad_norm_pre_clip': 16.94506072998047, 'timing/data': 0.011904601939022541, 'timing/model': 1.0586842652410269, 'learning_rate/qwen_vl_interface': 1.3582009899754864e-06, 'learning_rate/action_model': 1.3582009899754864e-05, 'epoch': 2.07}) +[2026-06-18 04:56:00,555][starVLA.training.train_starvla][INFO] - Step 5660, Loss: {'train/loss': 0.2736857235431671, 'train/grad_norm_pre_clip': 17.711225509643555, 'timing/data': 0.013539022766053677, 'timing/model': 1.0846543069928885, 'learning_rate/qwen_vl_interface': 1.3569614628333515e-06, 'learning_rate/action_model': 1.3569614628333516e-05, 'epoch': 2.07}) +[2026-06-18 04:56:05,160][starVLA.training.train_starvla][INFO] - Step 5661, Loss: {'train/loss': 0.19343630969524384, 'train/grad_norm_pre_clip': 16.825881958007812, 'timing/data': 0.012338994070887566, 'timing/model': 1.4869819479063153, 'learning_rate/qwen_vl_interface': 1.3557227427222954e-06, 'learning_rate/action_model': 1.3557227427222954e-05, 'epoch': 2.07}) +[2026-06-18 04:56:09,648][starVLA.training.train_starvla][INFO] - Step 5662, Loss: {'train/loss': 0.3193238377571106, 'train/grad_norm_pre_clip': 27.233509063720703, 'timing/data': 0.01286524347960949, 'timing/model': 1.329356295056641, 'learning_rate/qwen_vl_interface': 1.3544848298991056e-06, 'learning_rate/action_model': 1.3544848298991056e-05, 'epoch': 2.07}) +[2026-06-18 04:56:15,176][starVLA.training.train_starvla][INFO] - Step 5663, Loss: {'train/loss': 0.12165593355894089, 'train/grad_norm_pre_clip': 12.253055572509766, 'timing/data': 0.012464740313589573, 'timing/model': 1.0521834259852767, 'learning_rate/qwen_vl_interface': 1.3532477246204037e-06, 'learning_rate/action_model': 1.3532477246204036e-05, 'epoch': 2.07}) +[2026-06-18 04:56:19,725][starVLA.training.train_starvla][INFO] - Step 5664, Loss: {'train/loss': 0.11034556478261948, 'train/grad_norm_pre_clip': 12.276549339294434, 'timing/data': 0.014471207745373249, 'timing/model': 1.0645009949803352, 'learning_rate/qwen_vl_interface': 1.3520114271426382e-06, 'learning_rate/action_model': 1.3520114271426381e-05, 'epoch': 2.07}) +[2026-06-18 04:56:24,315][starVLA.training.train_starvla][INFO] - Step 5665, Loss: {'train/loss': 0.24870330095291138, 'train/grad_norm_pre_clip': 18.74346160888672, 'timing/data': 0.011555108241736889, 'timing/model': 1.0562489842996001, 'learning_rate/qwen_vl_interface': 1.3507759377221001e-06, 'learning_rate/action_model': 1.3507759377221003e-05, 'epoch': 2.07}) +[2026-06-18 04:56:28,737][starVLA.training.train_starvla][INFO] - Step 5666, Loss: {'train/loss': 0.42769768834114075, 'train/grad_norm_pre_clip': 26.568002700805664, 'timing/data': 0.01708568073809147, 'timing/model': 1.1153731914237142, 'learning_rate/qwen_vl_interface': 1.3495412566149058e-06, 'learning_rate/action_model': 1.3495412566149057e-05, 'epoch': 2.07}) +[2026-06-18 04:56:33,557][starVLA.training.train_starvla][INFO] - Step 5667, Loss: {'train/loss': 0.2817176282405853, 'train/grad_norm_pre_clip': 32.504981994628906, 'timing/data': 0.024822969920933247, 'timing/model': 1.5081715481355786, 'learning_rate/qwen_vl_interface': 1.3483073840770067e-06, 'learning_rate/action_model': 1.3483073840770066e-05, 'epoch': 2.08}) +[2026-06-18 04:56:38,495][starVLA.training.train_starvla][INFO] - Step 5668, Loss: {'train/loss': 0.31343579292297363, 'train/grad_norm_pre_clip': 39.0980110168457, 'timing/data': 0.012152965180575848, 'timing/model': 1.1145397732034326, 'learning_rate/qwen_vl_interface': 1.3470743203641846e-06, 'learning_rate/action_model': 1.3470743203641844e-05, 'epoch': 2.08}) +[2026-06-18 04:56:45,753][starVLA.training.train_starvla][INFO] - Step 5669, Loss: {'train/loss': 0.556612491607666, 'train/grad_norm_pre_clip': 38.596771240234375, 'timing/data': 0.017378369346261024, 'timing/model': 1.7396223293617368, 'learning_rate/qwen_vl_interface': 1.3458420657320555e-06, 'learning_rate/action_model': 1.3458420657320555e-05, 'epoch': 2.08}) +[2026-06-18 04:56:50,716][starVLA.training.train_starvla][INFO] - Step 5670, Loss: {'train/loss': 0.4531715214252472, 'train/grad_norm_pre_clip': 31.50634765625, 'timing/data': 0.047184606082737446, 'timing/model': 1.0973565764725208, 'learning_rate/qwen_vl_interface': 1.344610620436067e-06, 'learning_rate/action_model': 1.344610620436067e-05, 'epoch': 2.08}) +[2026-06-18 04:56:56,938][starVLA.training.train_starvla][INFO] - Step 5671, Loss: {'train/loss': 0.08978834003210068, 'train/grad_norm_pre_clip': 12.136147499084473, 'timing/data': 0.015080327168107033, 'timing/model': 1.1087903687730432, 'learning_rate/qwen_vl_interface': 1.3433799847314989e-06, 'learning_rate/action_model': 1.3433799847314988e-05, 'epoch': 2.08}) +[2026-06-18 04:57:05,447][starVLA.training.train_starvla][INFO] - Step 5672, Loss: {'train/loss': 0.5497170090675354, 'train/grad_norm_pre_clip': 26.474653244018555, 'timing/data': 0.014580869115889072, 'timing/model': 1.4687502654269338, 'learning_rate/qwen_vl_interface': 1.3421501588734655e-06, 'learning_rate/action_model': 1.3421501588734656e-05, 'epoch': 2.08}) +[2026-06-18 04:57:09,776][starVLA.training.train_starvla][INFO] - Step 5673, Loss: {'train/loss': 0.19278757274150848, 'train/grad_norm_pre_clip': 16.335182189941406, 'timing/data': 0.010834509506821632, 'timing/model': 1.1067751552909613, 'learning_rate/qwen_vl_interface': 1.3409211431169084e-06, 'learning_rate/action_model': 1.3409211431169083e-05, 'epoch': 2.08}) +[2026-06-18 04:57:14,532][starVLA.training.train_starvla][INFO] - Step 5674, Loss: {'train/loss': 0.11837293207645416, 'train/grad_norm_pre_clip': 12.381911277770996, 'timing/data': 0.017726046964526176, 'timing/model': 1.2472374876961112, 'learning_rate/qwen_vl_interface': 1.3396929377166042e-06, 'learning_rate/action_model': 1.3396929377166043e-05, 'epoch': 2.08}) +[2026-06-18 04:57:24,288][starVLA.training.train_starvla][INFO] - Step 5675, Loss: {'train/loss': 0.271056205034256, 'train/grad_norm_pre_clip': 22.537315368652344, 'timing/data': 0.01294382568448782, 'timing/model': 2.6451258156448603, 'learning_rate/qwen_vl_interface': 1.338465542927161e-06, 'learning_rate/action_model': 1.3384655429271608e-05, 'epoch': 2.08}) +[2026-06-18 04:57:32,457][starVLA.training.train_starvla][INFO] - Step 5676, Loss: {'train/loss': 0.7445017099380493, 'train/grad_norm_pre_clip': 31.089263916015625, 'timing/data': 0.01359628140926361, 'timing/model': 1.4976178193464875, 'learning_rate/qwen_vl_interface': 1.3372389590030204e-06, 'learning_rate/action_model': 1.3372389590030204e-05, 'epoch': 2.08}) +[2026-06-18 04:57:37,024][starVLA.training.train_starvla][INFO] - Step 5677, Loss: {'train/loss': 0.17887361347675323, 'train/grad_norm_pre_clip': 21.698652267456055, 'timing/data': 0.01130268257111311, 'timing/model': 1.4272311376407743, 'learning_rate/qwen_vl_interface': 1.3360131861984551e-06, 'learning_rate/action_model': 1.3360131861984552e-05, 'epoch': 2.08}) +[2026-06-18 04:57:41,221][starVLA.training.train_starvla][INFO] - Step 5678, Loss: {'train/loss': 0.39879006147384644, 'train/grad_norm_pre_clip': 24.47909927368164, 'timing/data': 0.012733379378914833, 'timing/model': 1.0889891795814037, 'learning_rate/qwen_vl_interface': 1.3347882247675662e-06, 'learning_rate/action_model': 1.3347882247675661e-05, 'epoch': 2.08}) +[2026-06-18 04:57:45,653][starVLA.training.train_starvla][INFO] - Step 5679, Loss: {'train/loss': 0.2456490695476532, 'train/grad_norm_pre_clip': 24.04121208190918, 'timing/data': 0.011564130894839764, 'timing/model': 1.085525649599731, 'learning_rate/qwen_vl_interface': 1.3335640749642923e-06, 'learning_rate/action_model': 1.3335640749642923e-05, 'epoch': 2.08}) +[2026-06-18 04:57:50,228][starVLA.training.train_starvla][INFO] - Step 5680, Loss: {'train/loss': 0.07963692396879196, 'train/grad_norm_pre_clip': 11.993913650512695, 'timing/data': 0.01210144255310297, 'timing/model': 1.090775934047997, 'learning_rate/qwen_vl_interface': 1.3323407370423993e-06, 'learning_rate/action_model': 1.3323407370423994e-05, 'epoch': 2.08}) +[2026-06-18 04:57:54,589][starVLA.training.train_starvla][INFO] - Step 5681, Loss: {'train/loss': 0.1676129847764969, 'train/grad_norm_pre_clip': 21.186471939086914, 'timing/data': 0.012279191054403782, 'timing/model': 1.053010955452919, 'learning_rate/qwen_vl_interface': 1.3311182112554877e-06, 'learning_rate/action_model': 1.3311182112554876e-05, 'epoch': 2.08}) +[2026-06-18 04:57:59,034][starVLA.training.train_starvla][INFO] - Step 5682, Loss: {'train/loss': 0.4259461760520935, 'train/grad_norm_pre_clip': 20.675729751586914, 'timing/data': 0.013517328538000584, 'timing/model': 1.4377856459468603, 'learning_rate/qwen_vl_interface': 1.3298964978569873e-06, 'learning_rate/action_model': 1.3298964978569872e-05, 'epoch': 2.08}) +[2026-06-18 04:58:03,136][starVLA.training.train_starvla][INFO] - Step 5683, Loss: {'train/loss': 0.37439730763435364, 'train/grad_norm_pre_clip': 23.631446838378906, 'timing/data': 0.01147681288421154, 'timing/model': 1.0511989668011665, 'learning_rate/qwen_vl_interface': 1.3286755971001618e-06, 'learning_rate/action_model': 1.3286755971001618e-05, 'epoch': 2.08}) +[2026-06-18 04:58:07,580][starVLA.training.train_starvla][INFO] - Step 5684, Loss: {'train/loss': 0.269097238779068, 'train/grad_norm_pre_clip': 26.28804588317871, 'timing/data': 0.014062808826565742, 'timing/model': 1.070208851248026, 'learning_rate/qwen_vl_interface': 1.327455509238104e-06, 'learning_rate/action_model': 1.3274555092381039e-05, 'epoch': 2.08}) +[2026-06-18 04:58:12,036][starVLA.training.train_starvla][INFO] - Step 5685, Loss: {'train/loss': 0.1629456728696823, 'train/grad_norm_pre_clip': 16.390260696411133, 'timing/data': 0.012377677485346794, 'timing/model': 1.0279869409278035, 'learning_rate/qwen_vl_interface': 1.32623623452374e-06, 'learning_rate/action_model': 1.32623623452374e-05, 'epoch': 2.08}) +[2026-06-18 04:58:16,639][starVLA.training.train_starvla][INFO] - Step 5686, Loss: {'train/loss': 0.3648916482925415, 'train/grad_norm_pre_clip': 23.411802291870117, 'timing/data': 0.013204622082412243, 'timing/model': 1.3787728184834123, 'learning_rate/qwen_vl_interface': 1.325017773209827e-06, 'learning_rate/action_model': 1.325017773209827e-05, 'epoch': 2.08}) +[2026-06-18 04:58:21,343][starVLA.training.train_starvla][INFO] - Step 5687, Loss: {'train/loss': 0.24430033564567566, 'train/grad_norm_pre_clip': 30.509296417236328, 'timing/data': 0.022085342556238174, 'timing/model': 1.489723295904696, 'learning_rate/qwen_vl_interface': 1.323800125548953e-06, 'learning_rate/action_model': 1.323800125548953e-05, 'epoch': 2.08}) +[2026-06-18 04:58:25,458][starVLA.training.train_starvla][INFO] - Step 5688, Loss: {'train/loss': 0.22183531522750854, 'train/grad_norm_pre_clip': 27.49203872680664, 'timing/data': 0.016928871162235737, 'timing/model': 1.0463160080835223, 'learning_rate/qwen_vl_interface': 1.3225832917935376e-06, 'learning_rate/action_model': 1.3225832917935377e-05, 'epoch': 2.08}) +[2026-06-18 04:58:29,986][starVLA.training.train_starvla][INFO] - Step 5689, Loss: {'train/loss': 0.20937491953372955, 'train/grad_norm_pre_clip': 26.490272521972656, 'timing/data': 0.02114214189350605, 'timing/model': 1.0656065717339516, 'learning_rate/qwen_vl_interface': 1.3213672721958309e-06, 'learning_rate/action_model': 1.321367272195831e-05, 'epoch': 2.08}) +[2026-06-18 04:58:34,448][starVLA.training.train_starvla][INFO] - Step 5690, Loss: {'train/loss': 0.45448654890060425, 'train/grad_norm_pre_clip': 28.006973266601562, 'timing/data': 0.01302315667271614, 'timing/model': 1.0486398180946708, 'learning_rate/qwen_vl_interface': 1.3201520670079172e-06, 'learning_rate/action_model': 1.3201520670079171e-05, 'epoch': 2.08}) +[2026-06-18 04:58:39,116][starVLA.training.train_starvla][INFO] - Step 5691, Loss: {'train/loss': 0.53440922498703, 'train/grad_norm_pre_clip': 31.885616302490234, 'timing/data': 0.011118646711111069, 'timing/model': 1.0751606160774827, 'learning_rate/qwen_vl_interface': 1.31893767648171e-06, 'learning_rate/action_model': 1.3189376764817099e-05, 'epoch': 2.08}) +[2026-06-18 04:58:43,972][starVLA.training.train_starvla][INFO] - Step 5692, Loss: {'train/loss': 0.601673424243927, 'train/grad_norm_pre_clip': 38.201744079589844, 'timing/data': 0.01342842448502779, 'timing/model': 1.6860055141150951, 'learning_rate/qwen_vl_interface': 1.3177241008689513e-06, 'learning_rate/action_model': 1.3177241008689511e-05, 'epoch': 2.08}) +[2026-06-18 04:58:49,058][starVLA.training.train_starvla][INFO] - Step 5693, Loss: {'train/loss': 0.10547585785388947, 'train/grad_norm_pre_clip': 10.2172212600708, 'timing/data': 0.02500071655958891, 'timing/model': 1.1329067051410675, 'learning_rate/qwen_vl_interface': 1.316511340421216e-06, 'learning_rate/action_model': 1.316511340421216e-05, 'epoch': 2.08}) +[2026-06-18 04:58:54,342][starVLA.training.train_starvla][INFO] - Step 5694, Loss: {'train/loss': 0.3242666721343994, 'train/grad_norm_pre_clip': 33.43448257446289, 'timing/data': 0.012322797439992428, 'timing/model': 1.2681918758898973, 'learning_rate/qwen_vl_interface': 1.315299395389914e-06, 'learning_rate/action_model': 1.315299395389914e-05, 'epoch': 2.09}) +[2026-06-18 04:59:02,750][starVLA.training.train_starvla][INFO] - Step 5695, Loss: {'train/loss': 0.11720947921276093, 'train/grad_norm_pre_clip': 14.3167724609375, 'timing/data': 0.014413957484066486, 'timing/model': 1.3876464236527681, 'learning_rate/qwen_vl_interface': 1.3140882660262818e-06, 'learning_rate/action_model': 1.3140882660262818e-05, 'epoch': 2.09}) +[2026-06-18 04:59:12,817][starVLA.training.train_starvla][INFO] - Step 5696, Loss: {'train/loss': 0.3154681622982025, 'train/grad_norm_pre_clip': 29.939350128173828, 'timing/data': 0.01577630825340748, 'timing/model': 1.0846649380400777, 'learning_rate/qwen_vl_interface': 1.3128779525813865e-06, 'learning_rate/action_model': 1.3128779525813864e-05, 'epoch': 2.09}) +[2026-06-18 04:59:18,808][starVLA.training.train_starvla][INFO] - Step 5697, Loss: {'train/loss': 0.13447952270507812, 'train/grad_norm_pre_clip': 15.279009819030762, 'timing/data': 0.014956260100007057, 'timing/model': 2.6474505197256804, 'learning_rate/qwen_vl_interface': 1.3116684553061288e-06, 'learning_rate/action_model': 1.3116684553061287e-05, 'epoch': 2.09}) +[2026-06-18 04:59:23,402][starVLA.training.train_starvla][INFO] - Step 5698, Loss: {'train/loss': 0.26536768674850464, 'train/grad_norm_pre_clip': 28.907703399658203, 'timing/data': 0.06788739189505577, 'timing/model': 1.0782946925610304, 'learning_rate/qwen_vl_interface': 1.310459774451238e-06, 'learning_rate/action_model': 1.3104597744512379e-05, 'epoch': 2.09}) +[2026-06-18 04:59:28,733][starVLA.training.train_starvla][INFO] - Step 5699, Loss: {'train/loss': 0.09819915145635605, 'train/grad_norm_pre_clip': 13.66259479522705, 'timing/data': 0.014987453818321228, 'timing/model': 1.3919483814388514, 'learning_rate/qwen_vl_interface': 1.3092519102672744e-06, 'learning_rate/action_model': 1.3092519102672742e-05, 'epoch': 2.09}) +[2026-06-18 05:03:59,558][starVLA.training.train_starvla][INFO] - Step 5700, Loss: {'train/loss': 0.5886321663856506, 'train/grad_norm_pre_clip': 28.614648818969727, 'eval/loss': 0.5363407135009766, 'eval/latency_2/loss': 0.5363407135009766, 'eval/demon_fire_precision': 0.9550231189575452, 'eval/demon_fire_recall': 0.9506276150627615, 'eval/demon_fire_f1': 0.952820297756343, 'eval/demon_fire_rate_model': 0.6619365609348915, 'eval/demon_fire_rate_teacher': 0.6649972175848636, 'eval/demon_left_precision': 0.9047120418848168, 'eval/demon_left_recall': 0.9094736842105263, 'eval/demon_left_f1': 0.9070866141732284, 'eval/demon_left_rate_model': 0.2657206455203116, 'eval/demon_left_rate_teacher': 0.2643294379521425, 'eval/demon_right_precision': 0.9212962962962963, 'eval/demon_right_recall': 0.8994350282485876, 'eval/demon_right_f1': 0.910234419668382, 'eval/demon_right_rate_model': 0.24040066777963273, 'eval/demon_right_rate_teacher': 0.24624373956594323, 'eval/demon_move_f1': 0.9086605169208053, 'eval/demon_cc_f1': 0.939128156588387, 'eval/latency_2/demon_fire_precision': 0.9550231189575452, 'eval/latency_2/demon_fire_recall': 0.9506276150627615, 'eval/latency_2/demon_fire_f1': 0.952820297756343, 'eval/latency_2/demon_fire_rate_model': 0.6619365609348915, 'eval/latency_2/demon_fire_rate_teacher': 0.6649972175848636, 'eval/latency_2/demon_left_precision': 0.9047120418848168, 'eval/latency_2/demon_left_recall': 0.9094736842105263, 'eval/latency_2/demon_left_f1': 0.9070866141732284, 'eval/latency_2/demon_left_rate_model': 0.2657206455203116, 'eval/latency_2/demon_left_rate_teacher': 0.2643294379521425, 'eval/latency_2/demon_right_precision': 0.9212962962962963, 'eval/latency_2/demon_right_recall': 0.8994350282485876, 'eval/latency_2/demon_right_f1': 0.910234419668382, 'eval/latency_2/demon_right_rate_model': 0.24040066777963273, 'eval/latency_2/demon_right_rate_teacher': 0.24624373956594323, 'eval/latency_2/demon_move_f1': 0.9086605169208053, 'eval/latency_2/demon_cc_f1': 0.939128156588387, 'timing/data': 0.012583434581756592, 'timing/model': 2.8984760250896215, 'learning_rate/qwen_vl_interface': 1.3080448630046312e-06, 'learning_rate/action_model': 1.308044863004631e-05, 'epoch': 2.09}) +[2026-06-18 05:03:59,560][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_5700_state +[2026-06-18 05:03:59,571][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 05:04:09,828][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_5700_state/model.safetensors +[2026-06-18 05:04:25,514][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_5700_state/optimizer.bin +[2026-06-18 05:04:25,529][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_5700_state/sampler.bin +[2026-06-18 05:04:25,529][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_5700_state/sampler_1.bin +[2026-06-18 05:04:25,533][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_5700_state/random_states_0.pkl +[2026-06-18 05:04:25,545][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 5700 +[2026-06-18 05:04:25,545][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 05:04:25,564][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 05:04:30,607][starVLA.training.train_starvla][INFO] - Step 5701, Loss: {'train/loss': 0.16100795567035675, 'train/grad_norm_pre_clip': 17.630970001220703, 'timing/data': 0.013347423635423183, 'timing/model': 1.122753051109612, 'learning_rate/qwen_vl_interface': 1.3068386329135287e-06, 'learning_rate/action_model': 1.3068386329135287e-05, 'epoch': 2.09}) +[2026-06-18 05:04:38,410][starVLA.training.train_starvla][INFO] - Step 5702, Loss: {'train/loss': 0.354627788066864, 'train/grad_norm_pre_clip': 30.64838981628418, 'timing/data': 0.014614996500313282, 'timing/model': 1.7036616411060095, 'learning_rate/qwen_vl_interface': 1.3056332202440211e-06, 'learning_rate/action_model': 1.305633220244021e-05, 'epoch': 2.09}) +[2026-06-18 05:04:43,309][starVLA.training.train_starvla][INFO] - Step 5703, Loss: {'train/loss': 0.18370412290096283, 'train/grad_norm_pre_clip': 20.69678497314453, 'timing/data': 0.01295658852905035, 'timing/model': 1.3667084965854883, 'learning_rate/qwen_vl_interface': 1.3044286252459914e-06, 'learning_rate/action_model': 1.3044286252459915e-05, 'epoch': 2.09}) +[2026-06-18 05:04:59,414][starVLA.training.train_starvla][INFO] - Step 5704, Loss: {'train/loss': 0.13668343424797058, 'train/grad_norm_pre_clip': 14.694363594055176, 'timing/data': 0.01371088344603777, 'timing/model': 5.452643531374633, 'learning_rate/qwen_vl_interface': 1.3032248481691523e-06, 'learning_rate/action_model': 1.3032248481691523e-05, 'epoch': 2.09}) +[2026-06-18 05:05:09,107][starVLA.training.train_starvla][INFO] - Step 5705, Loss: {'train/loss': 0.09816932678222656, 'train/grad_norm_pre_clip': 13.589273452758789, 'timing/data': 0.013975176028907299, 'timing/model': 1.1556119658052921, 'learning_rate/qwen_vl_interface': 1.302021889263051e-06, 'learning_rate/action_model': 1.3020218892630508e-05, 'epoch': 2.09}) +[2026-06-18 05:05:14,101][starVLA.training.train_starvla][INFO] - Step 5706, Loss: {'train/loss': 0.40971997380256653, 'train/grad_norm_pre_clip': 25.482892990112305, 'timing/data': 0.0154645387083292, 'timing/model': 1.129861299879849, 'learning_rate/qwen_vl_interface': 1.3008197487770593e-06, 'learning_rate/action_model': 1.3008197487770593e-05, 'epoch': 2.09}) +[2026-06-18 05:05:19,209][starVLA.training.train_starvla][INFO] - Step 5707, Loss: {'train/loss': 0.6126868724822998, 'train/grad_norm_pre_clip': 22.80674171447754, 'timing/data': 0.013261014595627785, 'timing/model': 1.437147406861186, 'learning_rate/qwen_vl_interface': 1.2996184269603835e-06, 'learning_rate/action_model': 1.2996184269603833e-05, 'epoch': 2.09}) +[2026-06-18 05:05:24,707][starVLA.training.train_starvla][INFO] - Step 5708, Loss: {'train/loss': 0.1697310507297516, 'train/grad_norm_pre_clip': 22.435550689697266, 'timing/data': 0.09138475265353918, 'timing/model': 1.4594058906659484, 'learning_rate/qwen_vl_interface': 1.2984179240620567e-06, 'learning_rate/action_model': 1.2984179240620567e-05, 'epoch': 2.09}) +[2026-06-18 05:05:29,722][starVLA.training.train_starvla][INFO] - Step 5709, Loss: {'train/loss': 0.18715620040893555, 'train/grad_norm_pre_clip': 18.80183982849121, 'timing/data': 0.04443185497075319, 'timing/model': 1.7349508292973042, 'learning_rate/qwen_vl_interface': 1.2972182403309474e-06, 'learning_rate/action_model': 1.2972182403309474e-05, 'epoch': 2.09}) +[2026-06-18 05:05:34,368][starVLA.training.train_starvla][INFO] - Step 5710, Loss: {'train/loss': 0.6126258969306946, 'train/grad_norm_pre_clip': 31.206087112426758, 'timing/data': 0.04218084458261728, 'timing/model': 1.2009302424266934, 'learning_rate/qwen_vl_interface': 1.2960193760157507e-06, 'learning_rate/action_model': 1.2960193760157507e-05, 'epoch': 2.09}) +[2026-06-18 05:05:44,361][starVLA.training.train_starvla][INFO] - Step 5711, Loss: {'train/loss': 0.6616334915161133, 'train/grad_norm_pre_clip': 47.72758102416992, 'timing/data': 0.0425988445058465, 'timing/model': 1.6822358751669526, 'learning_rate/qwen_vl_interface': 1.294821331364992e-06, 'learning_rate/action_model': 1.294821331364992e-05, 'epoch': 2.09}) +[2026-06-18 05:05:52,614][starVLA.training.train_starvla][INFO] - Step 5712, Loss: {'train/loss': 0.3366050720214844, 'train/grad_norm_pre_clip': 23.638362884521484, 'timing/data': 0.04411944467574358, 'timing/model': 1.1071498887613416, 'learning_rate/qwen_vl_interface': 1.2936241066270278e-06, 'learning_rate/action_model': 1.2936241066270278e-05, 'epoch': 2.09}) +[2026-06-18 05:05:57,249][starVLA.training.train_starvla][INFO] - Step 5713, Loss: {'train/loss': 0.11782845109701157, 'train/grad_norm_pre_clip': 15.221534729003906, 'timing/data': 0.05937416013330221, 'timing/model': 1.3058203244581819, 'learning_rate/qwen_vl_interface': 1.2924277020500427e-06, 'learning_rate/action_model': 1.2924277020500428e-05, 'epoch': 2.09}) +[2026-06-18 05:06:01,572][starVLA.training.train_starvla][INFO] - Step 5714, Loss: {'train/loss': 0.2014201283454895, 'train/grad_norm_pre_clip': 15.883004188537598, 'timing/data': 0.05643365625292063, 'timing/model': 1.1599106444045901, 'learning_rate/qwen_vl_interface': 1.291232117882054e-06, 'learning_rate/action_model': 1.2912321178820538e-05, 'epoch': 2.09}) +[2026-06-18 05:06:06,909][starVLA.training.train_starvla][INFO] - Step 5715, Loss: {'train/loss': 0.08315715938806534, 'train/grad_norm_pre_clip': 49.02769470214844, 'timing/data': 0.04511143919080496, 'timing/model': 1.1966729275882244, 'learning_rate/qwen_vl_interface': 1.2900373543709065e-06, 'learning_rate/action_model': 1.2900373543709065e-05, 'epoch': 2.09}) +[2026-06-18 05:06:13,118][starVLA.training.train_starvla][INFO] - Step 5716, Loss: {'train/loss': 0.27497297525405884, 'train/grad_norm_pre_clip': 21.92544174194336, 'timing/data': 0.042214502580463886, 'timing/model': 1.0490494957193732, 'learning_rate/qwen_vl_interface': 1.288843411764277e-06, 'learning_rate/action_model': 1.288843411764277e-05, 'epoch': 2.09}) +[2026-06-18 05:06:20,822][starVLA.training.train_starvla][INFO] - Step 5717, Loss: {'train/loss': 0.527122437953949, 'train/grad_norm_pre_clip': 35.815555572509766, 'timing/data': 0.05462402757257223, 'timing/model': 1.0809245025739074, 'learning_rate/qwen_vl_interface': 1.2876502903096709e-06, 'learning_rate/action_model': 1.2876502903096708e-05, 'epoch': 2.09}) +[2026-06-18 05:06:27,807][starVLA.training.train_starvla][INFO] - Step 5718, Loss: {'train/loss': 0.4971643388271332, 'train/grad_norm_pre_clip': 42.73076248168945, 'timing/data': 0.046266159042716026, 'timing/model': 3.858227335847914, 'learning_rate/qwen_vl_interface': 1.286457990254423e-06, 'learning_rate/action_model': 1.286457990254423e-05, 'epoch': 2.09}) +[2026-06-18 05:06:42,448][starVLA.training.train_starvla][INFO] - Step 5719, Loss: {'train/loss': 0.22064240276813507, 'train/grad_norm_pre_clip': 20.6445255279541, 'timing/data': 0.05310465022921562, 'timing/model': 5.718149276450276, 'learning_rate/qwen_vl_interface': 1.2852665118456971e-06, 'learning_rate/action_model': 1.2852665118456972e-05, 'epoch': 2.09}) +[2026-06-18 05:06:47,562][starVLA.training.train_starvla][INFO] - Step 5720, Loss: {'train/loss': 0.1272163838148117, 'train/grad_norm_pre_clip': 19.794384002685547, 'timing/data': 0.041834247298538685, 'timing/model': 1.207881703041494, 'learning_rate/qwen_vl_interface': 1.2840758553304927e-06, 'learning_rate/action_model': 1.2840758553304927e-05, 'epoch': 2.09}) +[2026-06-18 05:06:52,532][starVLA.training.train_starvla][INFO] - Step 5721, Loss: {'train/loss': 0.19170640408992767, 'train/grad_norm_pre_clip': 29.987552642822266, 'timing/data': 0.10287999175488949, 'timing/model': 1.0869374256581068, 'learning_rate/qwen_vl_interface': 1.2828860209556293e-06, 'learning_rate/action_model': 1.2828860209556292e-05, 'epoch': 2.09}) +[2026-06-18 05:06:59,888][starVLA.training.train_starvla][INFO] - Step 5722, Loss: {'train/loss': 0.468659907579422, 'train/grad_norm_pre_clip': 31.221200942993164, 'timing/data': 0.04279271233826876, 'timing/model': 1.1371885892003775, 'learning_rate/qwen_vl_interface': 1.2816970089677617e-06, 'learning_rate/action_model': 1.2816970089677615e-05, 'epoch': 2.1}) +[2026-06-18 05:07:04,675][starVLA.training.train_starvla][INFO] - Step 5723, Loss: {'train/loss': 0.2867448627948761, 'train/grad_norm_pre_clip': 26.856502532958984, 'timing/data': 0.05014995206147432, 'timing/model': 1.513062808662653, 'learning_rate/qwen_vl_interface': 1.280508819613375e-06, 'learning_rate/action_model': 1.280508819613375e-05, 'epoch': 2.1}) +[2026-06-18 05:07:15,579][starVLA.training.train_starvla][INFO] - Step 5724, Loss: {'train/loss': 0.24423418939113617, 'train/grad_norm_pre_clip': 23.171846389770508, 'timing/data': 0.0502538587898016, 'timing/model': 1.0510945841670036, 'learning_rate/qwen_vl_interface': 1.2793214531387812e-06, 'learning_rate/action_model': 1.2793214531387812e-05, 'epoch': 2.1}) +[2026-06-18 05:07:20,217][starVLA.training.train_starvla][INFO] - Step 5725, Loss: {'train/loss': 0.436527818441391, 'train/grad_norm_pre_clip': 39.28288269042969, 'timing/data': 0.044819824397563934, 'timing/model': 1.1242042547091842, 'learning_rate/qwen_vl_interface': 1.2781349097901246e-06, 'learning_rate/action_model': 1.2781349097901247e-05, 'epoch': 2.1}) +[2026-06-18 05:07:25,075][starVLA.training.train_starvla][INFO] - Step 5726, Loss: {'train/loss': 0.16532331705093384, 'train/grad_norm_pre_clip': 15.120664596557617, 'timing/data': 0.04901240952312946, 'timing/model': 1.1834459509700537, 'learning_rate/qwen_vl_interface': 1.2769491898133723e-06, 'learning_rate/action_model': 1.2769491898133722e-05, 'epoch': 2.1}) +[2026-06-18 05:07:29,968][starVLA.training.train_starvla][INFO] - Step 5727, Loss: {'train/loss': 0.6542550921440125, 'train/grad_norm_pre_clip': 44.04488754272461, 'timing/data': 0.11908189207315445, 'timing/model': 1.4167370544746518, 'learning_rate/qwen_vl_interface': 1.2757642934543293e-06, 'learning_rate/action_model': 1.2757642934543293e-05, 'epoch': 2.1}) +[2026-06-18 05:07:37,033][starVLA.training.train_starvla][INFO] - Step 5728, Loss: {'train/loss': 0.1735161542892456, 'train/grad_norm_pre_clip': 21.04149627685547, 'timing/data': 0.04964625462889671, 'timing/model': 1.3827908597886562, 'learning_rate/qwen_vl_interface': 1.274580220958625e-06, 'learning_rate/action_model': 1.274580220958625e-05, 'epoch': 2.1}) +[2026-06-18 05:07:41,524][starVLA.training.train_starvla][INFO] - Step 5729, Loss: {'train/loss': 0.4446091949939728, 'train/grad_norm_pre_clip': 35.348636627197266, 'timing/data': 0.052524980157613754, 'timing/model': 1.2646519551053643, 'learning_rate/qwen_vl_interface': 1.2733969725717184e-06, 'learning_rate/action_model': 1.2733969725717184e-05, 'epoch': 2.1}) +[2026-06-18 05:07:55,641][starVLA.training.train_starvla][INFO] - Step 5730, Loss: {'train/loss': 0.19646252691745758, 'train/grad_norm_pre_clip': 18.51588249206543, 'timing/data': 0.05016287602484226, 'timing/model': 4.490765077061951, 'learning_rate/qwen_vl_interface': 1.272214548538899e-06, 'learning_rate/action_model': 1.2722145485388989e-05, 'epoch': 2.1}) +[2026-06-18 05:08:04,094][starVLA.training.train_starvla][INFO] - Step 5731, Loss: {'train/loss': 0.2715013921260834, 'train/grad_norm_pre_clip': 22.402048110961914, 'timing/data': 0.04640589375048876, 'timing/model': 1.1633446319028735, 'learning_rate/qwen_vl_interface': 1.2710329491052834e-06, 'learning_rate/action_model': 1.2710329491052833e-05, 'epoch': 2.1}) +[2026-06-18 05:08:09,411][starVLA.training.train_starvla][INFO] - Step 5732, Loss: {'train/loss': 0.3939763009548187, 'train/grad_norm_pre_clip': 29.666574478149414, 'timing/data': 0.045933037996292114, 'timing/model': 1.6307606361806393, 'learning_rate/qwen_vl_interface': 1.2698521745158188e-06, 'learning_rate/action_model': 1.2698521745158187e-05, 'epoch': 2.1}) +[2026-06-18 05:08:14,126][starVLA.training.train_starvla][INFO] - Step 5733, Loss: {'train/loss': 0.20858079195022583, 'train/grad_norm_pre_clip': 17.064393997192383, 'timing/data': 0.0525430403649807, 'timing/model': 1.4081033132970333, 'learning_rate/qwen_vl_interface': 1.2686722250152804e-06, 'learning_rate/action_model': 1.2686722250152806e-05, 'epoch': 2.1}) +[2026-06-18 05:08:19,555][starVLA.training.train_starvla][INFO] - Step 5734, Loss: {'train/loss': 0.12751321494579315, 'train/grad_norm_pre_clip': 16.475887298583984, 'timing/data': 0.039285910315811634, 'timing/model': 1.8618737533688545, 'learning_rate/qwen_vl_interface': 1.2674931008482763e-06, 'learning_rate/action_model': 1.2674931008482763e-05, 'epoch': 2.1}) +[2026-06-18 05:08:28,952][starVLA.training.train_starvla][INFO] - Step 5735, Loss: {'train/loss': 0.29851680994033813, 'train/grad_norm_pre_clip': 24.15575408935547, 'timing/data': 0.045738253742456436, 'timing/model': 1.1911193057894707, 'learning_rate/qwen_vl_interface': 1.2663148022592365e-06, 'learning_rate/action_model': 1.2663148022592363e-05, 'epoch': 2.1}) +[2026-06-18 05:08:33,992][starVLA.training.train_starvla][INFO] - Step 5736, Loss: {'train/loss': 0.227988138794899, 'train/grad_norm_pre_clip': 16.304994583129883, 'timing/data': 0.05301829054951668, 'timing/model': 1.1794468713924289, 'learning_rate/qwen_vl_interface': 1.2651373294924249e-06, 'learning_rate/action_model': 1.2651373294924248e-05, 'epoch': 2.1}) +[2026-06-18 05:08:42,035][starVLA.training.train_starvla][INFO] - Step 5737, Loss: {'train/loss': 0.22782017290592194, 'train/grad_norm_pre_clip': 13.988642692565918, 'timing/data': 0.05133412405848503, 'timing/model': 1.2782808933407068, 'learning_rate/qwen_vl_interface': 1.2639606827919313e-06, 'learning_rate/action_model': 1.2639606827919311e-05, 'epoch': 2.1}) +[2026-06-18 05:08:47,039][starVLA.training.train_starvla][INFO] - Step 5738, Loss: {'train/loss': 0.25245821475982666, 'train/grad_norm_pre_clip': 28.21726417541504, 'timing/data': 0.04424579069018364, 'timing/model': 1.530813948251307, 'learning_rate/qwen_vl_interface': 1.2627848624016788e-06, 'learning_rate/action_model': 1.2627848624016788e-05, 'epoch': 2.1}) +[2026-06-18 05:08:51,375][starVLA.training.train_starvla][INFO] - Step 5739, Loss: {'train/loss': 0.5414779186248779, 'train/grad_norm_pre_clip': 25.160696029663086, 'timing/data': 0.04456177446991205, 'timing/model': 1.193505541421473, 'learning_rate/qwen_vl_interface': 1.261609868565416e-06, 'learning_rate/action_model': 1.2616098685654159e-05, 'epoch': 2.1}) +[2026-06-18 05:08:56,006][starVLA.training.train_starvla][INFO] - Step 5740, Loss: {'train/loss': 0.12404853105545044, 'train/grad_norm_pre_clip': 10.798949241638184, 'timing/data': 0.04853004589676857, 'timing/model': 1.0477593196555972, 'learning_rate/qwen_vl_interface': 1.2604357015267165e-06, 'learning_rate/action_model': 1.2604357015267166e-05, 'epoch': 2.1}) +[2026-06-18 05:09:02,296][starVLA.training.train_starvla][INFO] - Step 5741, Loss: {'train/loss': 0.45911967754364014, 'train/grad_norm_pre_clip': 36.682498931884766, 'timing/data': 0.05724285915493965, 'timing/model': 1.07836152613163, 'learning_rate/qwen_vl_interface': 1.2592623615289906e-06, 'learning_rate/action_model': 1.2592623615289906e-05, 'epoch': 2.1}) +[2026-06-18 05:09:06,987][starVLA.training.train_starvla][INFO] - Step 5742, Loss: {'train/loss': 0.5879374146461487, 'train/grad_norm_pre_clip': 31.379138946533203, 'timing/data': 0.04553293902426958, 'timing/model': 1.2077080123126507, 'learning_rate/qwen_vl_interface': 1.258089848815471e-06, 'learning_rate/action_model': 1.2580898488154708e-05, 'epoch': 2.1}) +[2026-06-18 05:09:12,042][starVLA.training.train_starvla][INFO] - Step 5743, Loss: {'train/loss': 0.33353278040885925, 'train/grad_norm_pre_clip': 23.848100662231445, 'timing/data': 0.04905517306178808, 'timing/model': 1.4956760061904788, 'learning_rate/qwen_vl_interface': 1.256918163629221e-06, 'learning_rate/action_model': 1.256918163629221e-05, 'epoch': 2.1}) +[2026-06-18 05:09:21,489][starVLA.training.train_starvla][INFO] - Step 5744, Loss: {'train/loss': 0.17947252094745636, 'train/grad_norm_pre_clip': 19.160001754760742, 'timing/data': 0.05072258971631527, 'timing/model': 4.637326108291745, 'learning_rate/qwen_vl_interface': 1.2557473062131332e-06, 'learning_rate/action_model': 1.2557473062131331e-05, 'epoch': 2.1}) +[2026-06-18 05:09:29,559][starVLA.training.train_starvla][INFO] - Step 5745, Loss: {'train/loss': 0.20410597324371338, 'train/grad_norm_pre_clip': 29.95069694519043, 'timing/data': 0.04294787161052227, 'timing/model': 1.042395486496389, 'learning_rate/qwen_vl_interface': 1.2545772768099263e-06, 'learning_rate/action_model': 1.2545772768099263e-05, 'epoch': 2.1}) +[2026-06-18 05:09:35,822][starVLA.training.train_starvla][INFO] - Step 5746, Loss: {'train/loss': 0.25794631242752075, 'train/grad_norm_pre_clip': 22.11892318725586, 'timing/data': 0.0441006300970912, 'timing/model': 1.1181544186547399, 'learning_rate/qwen_vl_interface': 1.2534080756621488e-06, 'learning_rate/action_model': 1.2534080756621488e-05, 'epoch': 2.1}) +[2026-06-18 05:09:40,735][starVLA.training.train_starvla][INFO] - Step 5747, Loss: {'train/loss': 0.18186891078948975, 'train/grad_norm_pre_clip': 15.409564971923828, 'timing/data': 0.044623871333897114, 'timing/model': 1.2434813287109137, 'learning_rate/qwen_vl_interface': 1.2522397030121782e-06, 'learning_rate/action_model': 1.2522397030121783e-05, 'epoch': 2.1}) +[2026-06-18 05:09:45,391][starVLA.training.train_starvla][INFO] - Step 5748, Loss: {'train/loss': 0.3065907657146454, 'train/grad_norm_pre_clip': 28.479278564453125, 'timing/data': 0.04349797964096069, 'timing/model': 1.4800189305096865, 'learning_rate/qwen_vl_interface': 1.2510721591022194e-06, 'learning_rate/action_model': 1.2510721591022192e-05, 'epoch': 2.1}) +[2026-06-18 05:09:50,406][starVLA.training.train_starvla][INFO] - Step 5749, Loss: {'train/loss': 0.19816745817661285, 'train/grad_norm_pre_clip': 25.055994033813477, 'timing/data': 0.043706550262868404, 'timing/model': 1.3194317957386374, 'learning_rate/qwen_vl_interface': 1.2499054441743047e-06, 'learning_rate/action_model': 1.2499054441743047e-05, 'epoch': 2.11}) +[2026-06-18 05:09:58,106][starVLA.training.train_starvla][INFO] - Step 5750, Loss: {'train/loss': 0.4119113087654114, 'train/grad_norm_pre_clip': 27.330541610717773, 'timing/data': 0.04820533096790314, 'timing/model': 1.1976089105010033, 'learning_rate/qwen_vl_interface': 1.2487395584702955e-06, 'learning_rate/action_model': 1.2487395584702954e-05, 'epoch': 2.11}) +[2026-06-18 05:10:04,641][starVLA.training.train_starvla][INFO] - Step 5751, Loss: {'train/loss': 0.2417382299900055, 'train/grad_norm_pre_clip': 18.463876724243164, 'timing/data': 0.057330974377691746, 'timing/model': 1.0889698965474963, 'learning_rate/qwen_vl_interface': 1.2475745022318803e-06, 'learning_rate/action_model': 1.2475745022318803e-05, 'epoch': 2.11}) +[2026-06-18 05:10:09,750][starVLA.training.train_starvla][INFO] - Step 5752, Loss: {'train/loss': 0.38562002778053284, 'train/grad_norm_pre_clip': 23.536252975463867, 'timing/data': 0.04471075348556042, 'timing/model': 1.2475336212664843, 'learning_rate/qwen_vl_interface': 1.2464102757005788e-06, 'learning_rate/action_model': 1.2464102757005788e-05, 'epoch': 2.11}) +[2026-06-18 05:10:15,043][starVLA.training.train_starvla][INFO] - Step 5753, Loss: {'train/loss': 0.41350847482681274, 'train/grad_norm_pre_clip': 27.666675567626953, 'timing/data': 0.05209790822118521, 'timing/model': 1.4305639369413257, 'learning_rate/qwen_vl_interface': 1.2452468791177357e-06, 'learning_rate/action_model': 1.2452468791177357e-05, 'epoch': 2.11}) +[2026-06-18 05:10:23,618][starVLA.training.train_starvla][INFO] - Step 5754, Loss: {'train/loss': 0.26893526315689087, 'train/grad_norm_pre_clip': 20.60811424255371, 'timing/data': 0.011983063071966171, 'timing/model': 2.348620223812759, 'learning_rate/qwen_vl_interface': 1.2440843127245235e-06, 'learning_rate/action_model': 1.2440843127245235e-05, 'epoch': 2.11}) +[2026-06-18 05:10:29,069][starVLA.training.train_starvla][INFO] - Step 5755, Loss: {'train/loss': 0.10230409353971481, 'train/grad_norm_pre_clip': 16.74700355529785, 'timing/data': 0.02619867492467165, 'timing/model': 1.067174375988543, 'learning_rate/qwen_vl_interface': 1.2429225767619427e-06, 'learning_rate/action_model': 1.2429225767619427e-05, 'epoch': 2.11}) +[2026-06-18 05:10:40,093][starVLA.training.train_starvla][INFO] - Step 5756, Loss: {'train/loss': 0.5476065874099731, 'train/grad_norm_pre_clip': 40.76001739501953, 'timing/data': 0.013709746301174164, 'timing/model': 1.0459616538137197, 'learning_rate/qwen_vl_interface': 1.2417616714708246e-06, 'learning_rate/action_model': 1.2417616714708247e-05, 'epoch': 2.11}) +[2026-06-18 05:10:44,712][starVLA.training.train_starvla][INFO] - Step 5757, Loss: {'train/loss': 0.4202095568180084, 'train/grad_norm_pre_clip': 25.376916885375977, 'timing/data': 0.016890322789549828, 'timing/model': 1.0908688381314278, 'learning_rate/qwen_vl_interface': 1.2406015970918247e-06, 'learning_rate/action_model': 1.2406015970918247e-05, 'epoch': 2.11}) +[2026-06-18 05:10:50,318][starVLA.training.train_starvla][INFO] - Step 5758, Loss: {'train/loss': 0.19986476004123688, 'train/grad_norm_pre_clip': 18.759382247924805, 'timing/data': 0.018933428451418877, 'timing/model': 2.490525061264634, 'learning_rate/qwen_vl_interface': 1.2394423538654296e-06, 'learning_rate/action_model': 1.2394423538654296e-05, 'epoch': 2.11}) +[2026-06-18 05:10:54,594][starVLA.training.train_starvla][INFO] - Step 5759, Loss: {'train/loss': 0.3288310468196869, 'train/grad_norm_pre_clip': 36.90330505371094, 'timing/data': 0.015113696455955505, 'timing/model': 1.0299880737438798, 'learning_rate/qwen_vl_interface': 1.2382839420319478e-06, 'learning_rate/action_model': 1.2382839420319478e-05, 'epoch': 2.11}) +[2026-06-18 05:11:02,640][starVLA.training.train_starvla][INFO] - Step 5760, Loss: {'train/loss': 0.14494435489177704, 'train/grad_norm_pre_clip': 13.629988670349121, 'timing/data': 0.026287715882062912, 'timing/model': 1.3648223839700222, 'learning_rate/qwen_vl_interface': 1.2371263618315225e-06, 'learning_rate/action_model': 1.2371263618315224e-05, 'epoch': 2.11}) +[2026-06-18 05:11:07,287][starVLA.training.train_starvla][INFO] - Step 5761, Loss: {'train/loss': 0.39042216539382935, 'train/grad_norm_pre_clip': 26.669294357299805, 'timing/data': 0.012806902639567852, 'timing/model': 1.0453843465074897, 'learning_rate/qwen_vl_interface': 1.2359696135041202e-06, 'learning_rate/action_model': 1.2359696135041201e-05, 'epoch': 2.11}) +[2026-06-18 05:11:12,462][starVLA.training.train_starvla][INFO] - Step 5762, Loss: {'train/loss': 0.16611552238464355, 'train/grad_norm_pre_clip': 13.787632942199707, 'timing/data': 0.018184597603976727, 'timing/model': 1.5524339089170098, 'learning_rate/qwen_vl_interface': 1.2348136972895369e-06, 'learning_rate/action_model': 1.2348136972895368e-05, 'epoch': 2.11}) +[2026-06-18 05:11:17,512][starVLA.training.train_starvla][INFO] - Step 5763, Loss: {'train/loss': 0.18794965744018555, 'train/grad_norm_pre_clip': 19.232364654541016, 'timing/data': 0.012202048674225807, 'timing/model': 1.4436423694714904, 'learning_rate/qwen_vl_interface': 1.2336586134273945e-06, 'learning_rate/action_model': 1.2336586134273945e-05, 'epoch': 2.11}) +[2026-06-18 05:11:30,294][starVLA.training.train_starvla][INFO] - Step 5764, Loss: {'train/loss': 0.3622041642665863, 'train/grad_norm_pre_clip': 21.4943790435791, 'timing/data': 0.014391948468983173, 'timing/model': 1.201187701895833, 'learning_rate/qwen_vl_interface': 1.2325043621571428e-06, 'learning_rate/action_model': 1.2325043621571427e-05, 'epoch': 2.11}) +[2026-06-18 05:11:35,529][starVLA.training.train_starvla][INFO] - Step 5765, Loss: {'train/loss': 0.1542011797428131, 'train/grad_norm_pre_clip': 16.780181884765625, 'timing/data': 0.017812099307775497, 'timing/model': 1.1972038382664323, 'learning_rate/qwen_vl_interface': 1.2313509437180603e-06, 'learning_rate/action_model': 1.2313509437180604e-05, 'epoch': 2.11}) +[2026-06-18 05:11:40,299][starVLA.training.train_starvla][INFO] - Step 5766, Loss: {'train/loss': 0.43587562441825867, 'train/grad_norm_pre_clip': 28.990386962890625, 'timing/data': 0.012081007473170757, 'timing/model': 1.226634381338954, 'learning_rate/qwen_vl_interface': 1.2301983583492502e-06, 'learning_rate/action_model': 1.2301983583492502e-05, 'epoch': 2.11}) +[2026-06-18 05:11:45,165][starVLA.training.train_starvla][INFO] - Step 5767, Loss: {'train/loss': 0.051497407257556915, 'train/grad_norm_pre_clip': 6.8813934326171875, 'timing/data': 0.012904287315905094, 'timing/model': 1.4970315992832184, 'learning_rate/qwen_vl_interface': 1.229046606289648e-06, 'learning_rate/action_model': 1.229046606289648e-05, 'epoch': 2.11}) +[2026-06-18 05:11:50,573][starVLA.training.train_starvla][INFO] - Step 5768, Loss: {'train/loss': 0.22383424639701843, 'train/grad_norm_pre_clip': 14.039892196655273, 'timing/data': 0.02277915272861719, 'timing/model': 1.3873183820396662, 'learning_rate/qwen_vl_interface': 1.2278956877780102e-06, 'learning_rate/action_model': 1.2278956877780101e-05, 'epoch': 2.11}) +[2026-06-18 05:11:55,006][starVLA.training.train_starvla][INFO] - Step 5769, Loss: {'train/loss': 0.5399600863456726, 'train/grad_norm_pre_clip': 35.789955139160156, 'timing/data': 0.012970445677638054, 'timing/model': 1.0500547559931874, 'learning_rate/qwen_vl_interface': 1.2267456030529245e-06, 'learning_rate/action_model': 1.2267456030529245e-05, 'epoch': 2.11}) +[2026-06-18 05:12:01,118][starVLA.training.train_starvla][INFO] - Step 5770, Loss: {'train/loss': 0.19983547925949097, 'train/grad_norm_pre_clip': 24.375635147094727, 'timing/data': 0.01436858531087637, 'timing/model': 1.138656203635037, 'learning_rate/qwen_vl_interface': 1.225596352352803e-06, 'learning_rate/action_model': 1.225596352352803e-05, 'epoch': 2.11}) +[2026-06-18 05:12:06,023][starVLA.training.train_starvla][INFO] - Step 5771, Loss: {'train/loss': 0.143285870552063, 'train/grad_norm_pre_clip': 18.08724021911621, 'timing/data': 0.014069559052586555, 'timing/model': 1.2054880252107978, 'learning_rate/qwen_vl_interface': 1.2244479359158895e-06, 'learning_rate/action_model': 1.2244479359158895e-05, 'epoch': 2.11}) +[2026-06-18 05:12:10,606][starVLA.training.train_starvla][INFO] - Step 5772, Loss: {'train/loss': 0.2238890528678894, 'train/grad_norm_pre_clip': 25.34168815612793, 'timing/data': 0.012343907728791237, 'timing/model': 1.379953001625836, 'learning_rate/qwen_vl_interface': 1.2233003539802512e-06, 'learning_rate/action_model': 1.2233003539802513e-05, 'epoch': 2.11}) +[2026-06-18 05:12:15,307][starVLA.training.train_starvla][INFO] - Step 5773, Loss: {'train/loss': 0.4275989830493927, 'train/grad_norm_pre_clip': 25.216543197631836, 'timing/data': 0.01987847127020359, 'timing/model': 1.6593317175284028, 'learning_rate/qwen_vl_interface': 1.2221536067837807e-06, 'learning_rate/action_model': 1.2221536067837808e-05, 'epoch': 2.11}) +[2026-06-18 05:12:19,597][starVLA.training.train_starvla][INFO] - Step 5774, Loss: {'train/loss': 0.14162351191043854, 'train/grad_norm_pre_clip': 17.573762893676758, 'timing/data': 0.012872273102402687, 'timing/model': 1.0458285259082913, 'learning_rate/qwen_vl_interface': 1.2210076945642025e-06, 'learning_rate/action_model': 1.2210076945642024e-05, 'epoch': 2.11}) +[2026-06-18 05:12:24,299][starVLA.training.train_starvla][INFO] - Step 5775, Loss: {'train/loss': 0.7929074764251709, 'train/grad_norm_pre_clip': 37.183563232421875, 'timing/data': 0.014918684028089046, 'timing/model': 1.3199944384396076, 'learning_rate/qwen_vl_interface': 1.219862617559065e-06, 'learning_rate/action_model': 1.219862617559065e-05, 'epoch': 2.11}) +[2026-06-18 05:12:30,336][starVLA.training.train_starvla][INFO] - Step 5776, Loss: {'train/loss': 0.1476564109325409, 'train/grad_norm_pre_clip': 16.316293716430664, 'timing/data': 0.000882997177541256, 'timing/model': 1.1156053338199854, 'learning_rate/qwen_vl_interface': 1.2187183760057431e-06, 'learning_rate/action_model': 1.218718376005743e-05, 'epoch': 2.12}) +[2026-06-18 05:12:34,700][starVLA.training.train_starvla][INFO] - Step 5777, Loss: {'train/loss': 0.2499205321073532, 'train/grad_norm_pre_clip': 23.764068603515625, 'timing/data': 0.0008925767615437508, 'timing/model': 1.0652542989701033, 'learning_rate/qwen_vl_interface': 1.2175749701414393e-06, 'learning_rate/action_model': 1.2175749701414394e-05, 'epoch': 2.12}) +[2026-06-18 05:12:39,560][starVLA.training.train_starvla][INFO] - Step 5778, Loss: {'train/loss': 0.13538505136966705, 'train/grad_norm_pre_clip': 14.624558448791504, 'timing/data': 0.0011304663494229317, 'timing/model': 1.4497595382854342, 'learning_rate/qwen_vl_interface': 1.216432400203184e-06, 'learning_rate/action_model': 1.2164324002031841e-05, 'epoch': 2.12}) +[2026-06-18 05:12:47,922][starVLA.training.train_starvla][INFO] - Step 5779, Loss: {'train/loss': 0.15290740132331848, 'train/grad_norm_pre_clip': 15.262785911560059, 'timing/data': 0.0009230896830558777, 'timing/model': 4.9276438523083925, 'learning_rate/qwen_vl_interface': 1.215290666427832e-06, 'learning_rate/action_model': 1.2152906664278318e-05, 'epoch': 2.12}) +[2026-06-18 05:12:53,320][starVLA.training.train_starvla][INFO] - Step 5780, Loss: {'train/loss': 0.24021409451961517, 'train/grad_norm_pre_clip': 28.29620361328125, 'timing/data': 0.0011108564212918282, 'timing/model': 1.3152484577149153, 'learning_rate/qwen_vl_interface': 1.214149769052067e-06, 'learning_rate/action_model': 1.214149769052067e-05, 'epoch': 2.12}) +[2026-06-18 05:13:00,207][starVLA.training.train_starvla][INFO] - Step 5781, Loss: {'train/loss': 0.4528718590736389, 'train/grad_norm_pre_clip': 79.23736572265625, 'timing/data': 0.000901336781680584, 'timing/model': 1.1883743721991777, 'learning_rate/qwen_vl_interface': 1.2130097083123971e-06, 'learning_rate/action_model': 1.213009708312397e-05, 'epoch': 2.12}) +[2026-06-18 05:13:09,712][starVLA.training.train_starvla][INFO] - Step 5782, Loss: {'train/loss': 0.07156812399625778, 'train/grad_norm_pre_clip': 10.401384353637695, 'timing/data': 0.0024945633485913277, 'timing/model': 5.360404037870467, 'learning_rate/qwen_vl_interface': 1.2118704844451587e-06, 'learning_rate/action_model': 1.2118704844451586e-05, 'epoch': 2.12}) +[2026-06-18 05:13:20,297][starVLA.training.train_starvla][INFO] - Step 5783, Loss: {'train/loss': 0.34383052587509155, 'train/grad_norm_pre_clip': 23.132173538208008, 'timing/data': 0.0010615633800625801, 'timing/model': 2.9153409404680133, 'learning_rate/qwen_vl_interface': 1.2107320976865138e-06, 'learning_rate/action_model': 1.2107320976865139e-05, 'epoch': 2.12}) +[2026-06-18 05:13:30,962][starVLA.training.train_starvla][INFO] - Step 5784, Loss: {'train/loss': 0.05573540925979614, 'train/grad_norm_pre_clip': 7.710741996765137, 'timing/data': 0.0007803719490766525, 'timing/model': 1.3294897340238094, 'learning_rate/qwen_vl_interface': 1.209594548272451e-06, 'learning_rate/action_model': 1.209594548272451e-05, 'epoch': 2.12}) +[2026-06-18 05:13:44,519][starVLA.training.train_starvla][INFO] - Step 5785, Loss: {'train/loss': 0.21065877377986908, 'train/grad_norm_pre_clip': 17.915128707885742, 'timing/data': 0.0011332407593727112, 'timing/model': 1.1580957546830177, 'learning_rate/qwen_vl_interface': 1.2084578364387869e-06, 'learning_rate/action_model': 1.208457836438787e-05, 'epoch': 2.12}) +[2026-06-18 05:13:55,197][starVLA.training.train_starvla][INFO] - Step 5786, Loss: {'train/loss': 0.37042680382728577, 'train/grad_norm_pre_clip': 31.176815032958984, 'timing/data': 0.0011140843853354454, 'timing/model': 1.400793807581067, 'learning_rate/qwen_vl_interface': 1.2073219624211635e-06, 'learning_rate/action_model': 1.2073219624211635e-05, 'epoch': 2.12}) +[2026-06-18 05:14:00,817][starVLA.training.train_starvla][INFO] - Step 5787, Loss: {'train/loss': 0.3590952754020691, 'train/grad_norm_pre_clip': 29.060598373413086, 'timing/data': 0.001305634155869484, 'timing/model': 1.2129255747422576, 'learning_rate/qwen_vl_interface': 1.2061869264550458e-06, 'learning_rate/action_model': 1.2061869264550457e-05, 'epoch': 2.12}) +[2026-06-18 05:14:13,235][starVLA.training.train_starvla][INFO] - Step 5788, Loss: {'train/loss': 0.21959441900253296, 'train/grad_norm_pre_clip': 23.16242790222168, 'timing/data': 0.0011571059003472328, 'timing/model': 1.5765590788796544, 'learning_rate/qwen_vl_interface': 1.2050527287757288e-06, 'learning_rate/action_model': 1.2050527287757289e-05, 'epoch': 2.12}) +[2026-06-18 05:14:18,807][starVLA.training.train_starvla][INFO] - Step 5789, Loss: {'train/loss': 0.056129369884729385, 'train/grad_norm_pre_clip': 7.164241790771484, 'timing/data': 0.0011305250227451324, 'timing/model': 1.1454730741679668, 'learning_rate/qwen_vl_interface': 1.203919369618335e-06, 'learning_rate/action_model': 1.2039193696183348e-05, 'epoch': 2.12}) +[2026-06-18 05:14:23,521][starVLA.training.train_starvla][INFO] - Step 5790, Loss: {'train/loss': 0.1810990869998932, 'train/grad_norm_pre_clip': 18.09461784362793, 'timing/data': 0.0010664425790309906, 'timing/model': 1.0647428538650274, 'learning_rate/qwen_vl_interface': 1.202786849217809e-06, 'learning_rate/action_model': 1.202786849217809e-05, 'epoch': 2.12}) +[2026-06-18 05:14:27,982][starVLA.training.train_starvla][INFO] - Step 5791, Loss: {'train/loss': 0.2903287410736084, 'train/grad_norm_pre_clip': 24.880178451538086, 'timing/data': 0.0010798471048474312, 'timing/model': 1.0576712042093277, 'learning_rate/qwen_vl_interface': 1.2016551678089247e-06, 'learning_rate/action_model': 1.2016551678089246e-05, 'epoch': 2.12}) +[2026-06-18 05:14:32,947][starVLA.training.train_starvla][INFO] - Step 5792, Loss: {'train/loss': 0.26037758588790894, 'train/grad_norm_pre_clip': 25.883962631225586, 'timing/data': 0.0013481508940458298, 'timing/model': 1.5891994070261717, 'learning_rate/qwen_vl_interface': 1.2005243256262803e-06, 'learning_rate/action_model': 1.2005243256262802e-05, 'epoch': 2.12}) +[2026-06-18 05:14:37,412][starVLA.training.train_starvla][INFO] - Step 5793, Loss: {'train/loss': 0.15537115931510925, 'train/grad_norm_pre_clip': 20.274396896362305, 'timing/data': 0.0014994386583566666, 'timing/model': 1.394479038193822, 'learning_rate/qwen_vl_interface': 1.1993943229043003e-06, 'learning_rate/action_model': 1.1993943229043003e-05, 'epoch': 2.12}) +[2026-06-18 05:14:41,893][starVLA.training.train_starvla][INFO] - Step 5794, Loss: {'train/loss': 0.3041175603866577, 'train/grad_norm_pre_clip': 20.038232803344727, 'timing/data': 0.00123603455722332, 'timing/model': 1.043412164784968, 'learning_rate/qwen_vl_interface': 1.1982651598772358e-06, 'learning_rate/action_model': 1.1982651598772356e-05, 'epoch': 2.12}) +[2026-06-18 05:14:47,325][starVLA.training.train_starvla][INFO] - Step 5795, Loss: {'train/loss': 0.40295788645744324, 'train/grad_norm_pre_clip': 64.10148620605469, 'timing/data': 0.0009761080145835876, 'timing/model': 1.3354201707988977, 'learning_rate/qwen_vl_interface': 1.197136836779164e-06, 'learning_rate/action_model': 1.1971368367791639e-05, 'epoch': 2.12}) +[2026-06-18 05:14:52,186][starVLA.training.train_starvla][INFO] - Step 5796, Loss: {'train/loss': 0.7376887202262878, 'train/grad_norm_pre_clip': 29.81354522705078, 'timing/data': 0.0010234732180833817, 'timing/model': 1.0828911364078522, 'learning_rate/qwen_vl_interface': 1.1960093538439866e-06, 'learning_rate/action_model': 1.1960093538439864e-05, 'epoch': 2.12}) +[2026-06-18 05:14:56,993][starVLA.training.train_starvla][INFO] - Step 5797, Loss: {'train/loss': 0.3054484724998474, 'train/grad_norm_pre_clip': 32.19261169433594, 'timing/data': 0.0013859691098332405, 'timing/model': 1.0353056769818068, 'learning_rate/qwen_vl_interface': 1.1948827113054323e-06, 'learning_rate/action_model': 1.1948827113054323e-05, 'epoch': 2.12}) +[2026-06-18 05:15:02,218][starVLA.training.train_starvla][INFO] - Step 5798, Loss: {'train/loss': 0.15400591492652893, 'train/grad_norm_pre_clip': 14.765663146972656, 'timing/data': 0.0012726075947284698, 'timing/model': 1.4474501572549343, 'learning_rate/qwen_vl_interface': 1.1937569093970558e-06, 'learning_rate/action_model': 1.1937569093970558e-05, 'epoch': 2.12}) +[2026-06-18 05:15:06,566][starVLA.training.train_starvla][INFO] - Step 5799, Loss: {'train/loss': 0.3603980541229248, 'train/grad_norm_pre_clip': 25.555124282836914, 'timing/data': 0.0010105408728122711, 'timing/model': 1.1805908987298608, 'learning_rate/qwen_vl_interface': 1.1926319483522353e-06, 'learning_rate/action_model': 1.1926319483522353e-05, 'epoch': 2.12}) +[2026-06-18 05:20:02,989][starVLA.training.train_starvla][INFO] - Step 5800, Loss: {'train/loss': 0.1613653302192688, 'train/grad_norm_pre_clip': 28.074588775634766, 'eval/loss': 0.5378016829490662, 'eval/latency_2/loss': 0.5378016829490662, 'eval/demon_fire_precision': 0.9437140509449465, 'eval/demon_fire_recall': 0.9610878661087866, 'eval/demon_fire_f1': 0.9523217247097844, 'eval/demon_fire_rate_model': 0.6772398441847524, 'eval/demon_fire_rate_teacher': 0.6649972175848636, 'eval/demon_left_precision': 0.9156756756756756, 'eval/demon_left_recall': 0.891578947368421, 'eval/demon_left_f1': 0.9034666666666666, 'eval/demon_left_rate_model': 0.2573734001112966, 'eval/demon_left_rate_teacher': 0.2643294379521425, 'eval/demon_right_precision': 0.8900979325353645, 'eval/demon_right_recall': 0.9242937853107345, 'eval/demon_right_f1': 0.9068736141906872, 'eval/demon_right_rate_model': 0.2557039510294936, 'eval/demon_right_rate_teacher': 0.24624373956594323, 'eval/demon_move_f1': 0.9051701404286769, 'eval/demon_cc_f1': 0.937668385650416, 'eval/latency_2/demon_fire_precision': 0.9437140509449465, 'eval/latency_2/demon_fire_recall': 0.9610878661087866, 'eval/latency_2/demon_fire_f1': 0.9523217247097844, 'eval/latency_2/demon_fire_rate_model': 0.6772398441847524, 'eval/latency_2/demon_fire_rate_teacher': 0.6649972175848636, 'eval/latency_2/demon_left_precision': 0.9156756756756756, 'eval/latency_2/demon_left_recall': 0.891578947368421, 'eval/latency_2/demon_left_f1': 0.9034666666666666, 'eval/latency_2/demon_left_rate_model': 0.2573734001112966, 'eval/latency_2/demon_left_rate_teacher': 0.2643294379521425, 'eval/latency_2/demon_right_precision': 0.8900979325353645, 'eval/latency_2/demon_right_recall': 0.9242937853107345, 'eval/latency_2/demon_right_f1': 0.9068736141906872, 'eval/latency_2/demon_right_rate_model': 0.2557039510294936, 'eval/latency_2/demon_right_rate_teacher': 0.24624373956594323, 'eval/latency_2/demon_move_f1': 0.9051701404286769, 'eval/latency_2/demon_cc_f1': 0.937668385650416, 'timing/data': 0.0008582565933465958, 'timing/model': 1.0852974439039826, 'learning_rate/qwen_vl_interface': 1.1915078284041793e-06, 'learning_rate/action_model': 1.1915078284041793e-05, 'epoch': 2.12}) +[2026-06-18 05:20:02,991][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_5800_state +[2026-06-18 05:20:02,998][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 05:20:12,634][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_5800_state/model.safetensors +[2026-06-18 05:20:27,936][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_5800_state/optimizer.bin +[2026-06-18 05:20:27,937][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_5800_state/sampler.bin +[2026-06-18 05:20:27,938][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_5800_state/sampler_1.bin +[2026-06-18 05:20:27,941][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_5800_state/random_states_0.pkl +[2026-06-18 05:20:27,969][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 5800 +[2026-06-18 05:20:27,969][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 05:20:27,989][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 05:20:38,116][starVLA.training.train_starvla][INFO] - Step 5801, Loss: {'train/loss': 0.31627291440963745, 'train/grad_norm_pre_clip': 28.458988189697266, 'timing/data': 0.001098734326660633, 'timing/model': 1.6187288733199239, 'learning_rate/qwen_vl_interface': 1.1903845497859185e-06, 'learning_rate/action_model': 1.1903845497859184e-05, 'epoch': 2.12}) +[2026-06-18 05:20:42,880][starVLA.training.train_starvla][INFO] - Step 5802, Loss: {'train/loss': 0.29584673047065735, 'train/grad_norm_pre_clip': 23.619592666625977, 'timing/data': 0.0012135868892073631, 'timing/model': 1.143932175822556, 'learning_rate/qwen_vl_interface': 1.1892621127303067e-06, 'learning_rate/action_model': 1.1892621127303066e-05, 'epoch': 2.12}) +[2026-06-18 05:20:54,557][starVLA.training.train_starvla][INFO] - Step 5803, Loss: {'train/loss': 0.2830274701118469, 'train/grad_norm_pre_clip': 41.101924896240234, 'timing/data': 0.0011178674176335335, 'timing/model': 4.173256960697472, 'learning_rate/qwen_vl_interface': 1.1881405174700292e-06, 'learning_rate/action_model': 1.1881405174700291e-05, 'epoch': 2.12}) +[2026-06-18 05:21:03,690][starVLA.training.train_starvla][INFO] - Step 5804, Loss: {'train/loss': 0.3200016915798187, 'train/grad_norm_pre_clip': 21.531705856323242, 'timing/data': 0.0009071948006749153, 'timing/model': 1.4624152462929487, 'learning_rate/qwen_vl_interface': 1.1870197642375935e-06, 'learning_rate/action_model': 1.1870197642375935e-05, 'epoch': 2.13}) +[2026-06-18 05:21:08,464][starVLA.training.train_starvla][INFO] - Step 5805, Loss: {'train/loss': 0.49964064359664917, 'train/grad_norm_pre_clip': 40.360321044921875, 'timing/data': 0.0009180223569273949, 'timing/model': 1.0667339013889432, 'learning_rate/qwen_vl_interface': 1.1858998532653326e-06, 'learning_rate/action_model': 1.1858998532653325e-05, 'epoch': 2.13}) +[2026-06-18 05:21:13,380][starVLA.training.train_starvla][INFO] - Step 5806, Loss: {'train/loss': 0.2695014476776123, 'train/grad_norm_pre_clip': 26.36583709716797, 'timing/data': 0.0009810635820031166, 'timing/model': 1.1145401066169143, 'learning_rate/qwen_vl_interface': 1.1847807847854054e-06, 'learning_rate/action_model': 1.1847807847854054e-05, 'epoch': 2.13}) +[2026-06-18 05:21:19,232][starVLA.training.train_starvla][INFO] - Step 5807, Loss: {'train/loss': 0.19272981584072113, 'train/grad_norm_pre_clip': 23.468549728393555, 'timing/data': 0.0008637774735689163, 'timing/model': 2.220826425589621, 'learning_rate/qwen_vl_interface': 1.1836625590297959e-06, 'learning_rate/action_model': 1.1836625590297958e-05, 'epoch': 2.13}) +[2026-06-18 05:21:29,120][starVLA.training.train_starvla][INFO] - Step 5808, Loss: {'train/loss': 0.3225488066673279, 'train/grad_norm_pre_clip': 28.063688278198242, 'timing/data': 0.0011206259950995445, 'timing/model': 1.7449862798675895, 'learning_rate/qwen_vl_interface': 1.1825451762303134e-06, 'learning_rate/action_model': 1.1825451762303133e-05, 'epoch': 2.13}) +[2026-06-18 05:21:33,413][starVLA.training.train_starvla][INFO] - Step 5809, Loss: {'train/loss': 0.2810760736465454, 'train/grad_norm_pre_clip': 30.50082015991211, 'timing/data': 0.0007646046578884125, 'timing/model': 1.1078040730208158, 'learning_rate/qwen_vl_interface': 1.1814286366185923e-06, 'learning_rate/action_model': 1.1814286366185922e-05, 'epoch': 2.13}) +[2026-06-18 05:21:38,396][starVLA.training.train_starvla][INFO] - Step 5810, Loss: {'train/loss': 0.3357006311416626, 'train/grad_norm_pre_clip': 31.109554290771484, 'timing/data': 0.0013832636177539825, 'timing/model': 1.0447139497846365, 'learning_rate/qwen_vl_interface': 1.1803129404260932e-06, 'learning_rate/action_model': 1.1803129404260931e-05, 'epoch': 2.13}) +[2026-06-18 05:21:43,098][starVLA.training.train_starvla][INFO] - Step 5811, Loss: {'train/loss': 0.08520432561635971, 'train/grad_norm_pre_clip': 8.704355239868164, 'timing/data': 0.0012645255774259567, 'timing/model': 1.071749484166503, 'learning_rate/qwen_vl_interface': 1.1791980878841008e-06, 'learning_rate/action_model': 1.1791980878841008e-05, 'epoch': 2.13}) +[2026-06-18 05:21:48,341][starVLA.training.train_starvla][INFO] - Step 5812, Loss: {'train/loss': 0.4101293683052063, 'train/grad_norm_pre_clip': 23.487693786621094, 'timing/data': 0.0008299769833683968, 'timing/model': 1.243363686837256, 'learning_rate/qwen_vl_interface': 1.1780840792237246e-06, 'learning_rate/action_model': 1.1780840792237246e-05, 'epoch': 2.13}) +[2026-06-18 05:21:57,103][starVLA.training.train_starvla][INFO] - Step 5813, Loss: {'train/loss': 0.17445433139801025, 'train/grad_norm_pre_clip': 21.877206802368164, 'timing/data': 0.0010813120752573013, 'timing/model': 1.0569812068715692, 'learning_rate/qwen_vl_interface': 1.1769709146759002e-06, 'learning_rate/action_model': 1.1769709146759e-05, 'epoch': 2.13}) +[2026-06-18 05:22:01,881][starVLA.training.train_starvla][INFO] - Step 5814, Loss: {'train/loss': 0.3547455072402954, 'train/grad_norm_pre_clip': 27.445283889770508, 'timing/data': 0.0010997075587511063, 'timing/model': 1.4775999318808317, 'learning_rate/qwen_vl_interface': 1.175858594471388e-06, 'learning_rate/action_model': 1.175858594471388e-05, 'epoch': 2.13}) +[2026-06-18 05:22:06,485][starVLA.training.train_starvla][INFO] - Step 5815, Loss: {'train/loss': 0.0643901601433754, 'train/grad_norm_pre_clip': 8.379236221313477, 'timing/data': 0.0011584898456931114, 'timing/model': 1.050170574337244, 'learning_rate/qwen_vl_interface': 1.174747118840775e-06, 'learning_rate/action_model': 1.1747471188407749e-05, 'epoch': 2.13}) +[2026-06-18 05:22:11,052][starVLA.training.train_starvla][INFO] - Step 5816, Loss: {'train/loss': 0.19300618767738342, 'train/grad_norm_pre_clip': 31.30611801147461, 'timing/data': 0.0009555257856845856, 'timing/model': 1.137168281711638, 'learning_rate/qwen_vl_interface': 1.173636488014468e-06, 'learning_rate/action_model': 1.173636488014468e-05, 'epoch': 2.13}) +[2026-06-18 05:22:15,484][starVLA.training.train_starvla][INFO] - Step 5817, Loss: {'train/loss': 0.17383264005184174, 'train/grad_norm_pre_clip': 17.375301361083984, 'timing/data': 0.0008036158978939056, 'timing/model': 1.072583880275488, 'learning_rate/qwen_vl_interface': 1.1725267022227024e-06, 'learning_rate/action_model': 1.1725267022227025e-05, 'epoch': 2.13}) +[2026-06-18 05:22:22,920][starVLA.training.train_starvla][INFO] - Step 5818, Loss: {'train/loss': 0.10432767122983932, 'train/grad_norm_pre_clip': 13.734517097473145, 'timing/data': 0.001222989521920681, 'timing/model': 1.3680628892034292, 'learning_rate/qwen_vl_interface': 1.17141776169554e-06, 'learning_rate/action_model': 1.17141776169554e-05, 'epoch': 2.13}) +[2026-06-18 05:22:27,178][starVLA.training.train_starvla][INFO] - Step 5819, Loss: {'train/loss': 0.3468276560306549, 'train/grad_norm_pre_clip': 16.275684356689453, 'timing/data': 0.0008944598957896233, 'timing/model': 1.0478773284703493, 'learning_rate/qwen_vl_interface': 1.170309666662864e-06, 'learning_rate/action_model': 1.170309666662864e-05, 'epoch': 2.13}) +[2026-06-18 05:22:33,377][starVLA.training.train_starvla][INFO] - Step 5820, Loss: {'train/loss': 0.27294474840164185, 'train/grad_norm_pre_clip': 21.710718154907227, 'timing/data': 0.0014174515381455421, 'timing/model': 1.0589954862371087, 'learning_rate/qwen_vl_interface': 1.1692024173543858e-06, 'learning_rate/action_model': 1.1692024173543859e-05, 'epoch': 2.13}) +[2026-06-18 05:22:38,351][starVLA.training.train_starvla][INFO] - Step 5821, Loss: {'train/loss': 0.36924272775650024, 'train/grad_norm_pre_clip': 37.688899993896484, 'timing/data': 0.0011357776820659637, 'timing/model': 1.1140793869271874, 'learning_rate/qwen_vl_interface': 1.168096013999635e-06, 'learning_rate/action_model': 1.168096013999635e-05, 'epoch': 2.13}) +[2026-06-18 05:22:44,915][starVLA.training.train_starvla][INFO] - Step 5822, Loss: {'train/loss': 0.38870394229888916, 'train/grad_norm_pre_clip': 26.160205841064453, 'timing/data': 0.0011118678376078606, 'timing/model': 3.041942804120481, 'learning_rate/qwen_vl_interface': 1.1669904568279742e-06, 'learning_rate/action_model': 1.1669904568279741e-05, 'epoch': 2.13}) +[2026-06-18 05:22:53,650][starVLA.training.train_starvla][INFO] - Step 5823, Loss: {'train/loss': 0.6677607297897339, 'train/grad_norm_pre_clip': 39.65503692626953, 'timing/data': 0.0008938740938901901, 'timing/model': 1.1952038882300258, 'learning_rate/qwen_vl_interface': 1.165885746068585e-06, 'learning_rate/action_model': 1.1658857460685848e-05, 'epoch': 2.13}) +[2026-06-18 05:22:59,091][starVLA.training.train_starvla][INFO] - Step 5824, Loss: {'train/loss': 0.13163159787654877, 'train/grad_norm_pre_clip': 14.468523979187012, 'timing/data': 0.001268451102077961, 'timing/model': 1.414830812253058, 'learning_rate/qwen_vl_interface': 1.1647818819504754e-06, 'learning_rate/action_model': 1.1647818819504756e-05, 'epoch': 2.13}) +[2026-06-18 05:23:05,046][starVLA.training.train_starvla][INFO] - Step 5825, Loss: {'train/loss': 0.2541574537754059, 'train/grad_norm_pre_clip': 20.217853546142578, 'timing/data': 0.001255120150744915, 'timing/model': 1.3809749456122518, 'learning_rate/qwen_vl_interface': 1.1636788647024776e-06, 'learning_rate/action_model': 1.1636788647024776e-05, 'epoch': 2.13}) +[2026-06-18 05:23:09,783][starVLA.training.train_starvla][INFO] - Step 5826, Loss: {'train/loss': 0.16484275460243225, 'train/grad_norm_pre_clip': 18.535707473754883, 'timing/data': 0.0008799713104963303, 'timing/model': 1.0324446149170399, 'learning_rate/qwen_vl_interface': 1.1625766945532485e-06, 'learning_rate/action_model': 1.1625766945532484e-05, 'epoch': 2.13}) +[2026-06-18 05:23:15,020][starVLA.training.train_starvla][INFO] - Step 5827, Loss: {'train/loss': 0.14024655520915985, 'train/grad_norm_pre_clip': 21.57353973388672, 'timing/data': 0.0009104674682021141, 'timing/model': 1.675182462669909, 'learning_rate/qwen_vl_interface': 1.1614753717312683e-06, 'learning_rate/action_model': 1.1614753717312681e-05, 'epoch': 2.13}) +[2026-06-18 05:23:19,633][starVLA.training.train_starvla][INFO] - Step 5828, Loss: {'train/loss': 0.12941518425941467, 'train/grad_norm_pre_clip': 14.469642639160156, 'timing/data': 0.001041923649609089, 'timing/model': 1.539502477273345, 'learning_rate/qwen_vl_interface': 1.1603748964648423e-06, 'learning_rate/action_model': 1.1603748964648422e-05, 'epoch': 2.13}) +[2026-06-18 05:23:27,107][starVLA.training.train_starvla][INFO] - Step 5829, Loss: {'train/loss': 0.3599645793437958, 'train/grad_norm_pre_clip': 29.0686092376709, 'timing/data': 0.0008614137768745422, 'timing/model': 4.474195433780551, 'learning_rate/qwen_vl_interface': 1.1592752689821024e-06, 'learning_rate/action_model': 1.1592752689821024e-05, 'epoch': 2.13}) +[2026-06-18 05:23:35,123][starVLA.training.train_starvla][INFO] - Step 5830, Loss: {'train/loss': 0.3358430862426758, 'train/grad_norm_pre_clip': 22.395030975341797, 'timing/data': 0.0012892447412014008, 'timing/model': 1.0597218815237284, 'learning_rate/qwen_vl_interface': 1.1581764895110007e-06, 'learning_rate/action_model': 1.1581764895110007e-05, 'epoch': 2.13}) +[2026-06-18 05:23:39,664][starVLA.training.train_starvla][INFO] - Step 5831, Loss: {'train/loss': 0.18314167857170105, 'train/grad_norm_pre_clip': 23.560468673706055, 'timing/data': 0.0012609688565135002, 'timing/model': 1.168216610327363, 'learning_rate/qwen_vl_interface': 1.1570785582793153e-06, 'learning_rate/action_model': 1.1570785582793154e-05, 'epoch': 2.14}) +[2026-06-18 05:23:47,678][starVLA.training.train_starvla][INFO] - Step 5832, Loss: {'train/loss': 0.2840401232242584, 'train/grad_norm_pre_clip': 19.026622772216797, 'timing/data': 0.0013867812231183052, 'timing/model': 1.1946225371211767, 'learning_rate/qwen_vl_interface': 1.1559814755146484e-06, 'learning_rate/action_model': 1.1559814755146483e-05, 'epoch': 2.14}) +[2026-06-18 05:23:52,488][starVLA.training.train_starvla][INFO] - Step 5833, Loss: {'train/loss': 0.07134342938661575, 'train/grad_norm_pre_clip': 8.918837547302246, 'timing/data': 0.001337558962404728, 'timing/model': 1.0440831650048494, 'learning_rate/qwen_vl_interface': 1.154885241444428e-06, 'learning_rate/action_model': 1.1548852414444277e-05, 'epoch': 2.14}) +[2026-06-18 05:23:57,406][starVLA.training.train_starvla][INFO] - Step 5834, Loss: {'train/loss': 0.26774123311042786, 'train/grad_norm_pre_clip': 21.887392044067383, 'timing/data': 0.0010559968650341034, 'timing/model': 1.4354071328416467, 'learning_rate/qwen_vl_interface': 1.1537898562959039e-06, 'learning_rate/action_model': 1.1537898562959037e-05, 'epoch': 2.14}) +[2026-06-18 05:24:02,146][starVLA.training.train_starvla][INFO] - Step 5835, Loss: {'train/loss': 0.07032231986522675, 'train/grad_norm_pre_clip': 10.849766731262207, 'timing/data': 0.0008525252342224121, 'timing/model': 1.512263610959053, 'learning_rate/qwen_vl_interface': 1.152695320296149e-06, 'learning_rate/action_model': 1.1526953202961488e-05, 'epoch': 2.14}) +[2026-06-18 05:24:06,767][starVLA.training.train_starvla][INFO] - Step 5836, Loss: {'train/loss': 0.09912386536598206, 'train/grad_norm_pre_clip': 14.45862865447998, 'timing/data': 0.001081574708223343, 'timing/model': 1.1661697179079056, 'learning_rate/qwen_vl_interface': 1.1516016336720635e-06, 'learning_rate/action_model': 1.1516016336720633e-05, 'epoch': 2.14}) +[2026-06-18 05:24:11,443][starVLA.training.train_starvla][INFO] - Step 5837, Loss: {'train/loss': 0.35181012749671936, 'train/grad_norm_pre_clip': 35.15943908691406, 'timing/data': 0.001501518301665783, 'timing/model': 1.1572919022291899, 'learning_rate/qwen_vl_interface': 1.1505087966503697e-06, 'learning_rate/action_model': 1.1505087966503697e-05, 'epoch': 2.14}) +[2026-06-18 05:24:16,032][starVLA.training.train_starvla][INFO] - Step 5838, Loss: {'train/loss': 0.18262995779514313, 'train/grad_norm_pre_clip': 21.40972137451172, 'timing/data': 0.0008780583739280701, 'timing/model': 1.095269507728517, 'learning_rate/qwen_vl_interface': 1.1494168094576138e-06, 'learning_rate/action_model': 1.1494168094576138e-05, 'epoch': 2.14}) +[2026-06-18 05:24:26,118][starVLA.training.train_starvla][INFO] - Step 5839, Loss: {'train/loss': 0.28767919540405273, 'train/grad_norm_pre_clip': 24.62631607055664, 'timing/data': 0.0008622650057077408, 'timing/model': 5.1683116145431995, 'learning_rate/qwen_vl_interface': 1.1483256723201656e-06, 'learning_rate/action_model': 1.1483256723201655e-05, 'epoch': 2.14}) +[2026-06-18 05:24:30,581][starVLA.training.train_starvla][INFO] - Step 5840, Loss: {'train/loss': 0.08374767750501633, 'train/grad_norm_pre_clip': 11.19505500793457, 'timing/data': 0.0009287390857934952, 'timing/model': 1.123296894133091, 'learning_rate/qwen_vl_interface': 1.1472353854642187e-06, 'learning_rate/action_model': 1.1472353854642187e-05, 'epoch': 2.14}) +[2026-06-18 05:24:35,462][starVLA.training.train_starvla][INFO] - Step 5841, Loss: {'train/loss': 0.1965368092060089, 'train/grad_norm_pre_clip': 18.4417781829834, 'timing/data': 0.0011005178093910217, 'timing/model': 1.1388435531407595, 'learning_rate/qwen_vl_interface': 1.1461459491157914e-06, 'learning_rate/action_model': 1.1461459491157914e-05, 'epoch': 2.14}) +[2026-06-18 05:24:42,009][starVLA.training.train_starvla][INFO] - Step 5842, Loss: {'train/loss': 0.2683185935020447, 'train/grad_norm_pre_clip': 35.79977798461914, 'timing/data': 0.0009111072868108749, 'timing/model': 1.1036024373024702, 'learning_rate/qwen_vl_interface': 1.145057363500725e-06, 'learning_rate/action_model': 1.1450573635007251e-05, 'epoch': 2.14}) +[2026-06-18 05:24:46,960][starVLA.training.train_starvla][INFO] - Step 5843, Loss: {'train/loss': 0.4277961552143097, 'train/grad_norm_pre_clip': 26.858959197998047, 'timing/data': 0.0008151354268193245, 'timing/model': 1.3071940671652555, 'learning_rate/qwen_vl_interface': 1.1439696288446845e-06, 'learning_rate/action_model': 1.1439696288446844e-05, 'epoch': 2.14}) +[2026-06-18 05:24:54,965][starVLA.training.train_starvla][INFO] - Step 5844, Loss: {'train/loss': 0.056725434958934784, 'train/grad_norm_pre_clip': 8.178091049194336, 'timing/data': 0.0017374763265252113, 'timing/model': 1.4902151683345437, 'learning_rate/qwen_vl_interface': 1.1428827453731575e-06, 'learning_rate/action_model': 1.1428827453731574e-05, 'epoch': 2.14}) +[2026-06-18 05:24:59,612][starVLA.training.train_starvla][INFO] - Step 5845, Loss: {'train/loss': 0.4123155176639557, 'train/grad_norm_pre_clip': 17.089094161987305, 'timing/data': 0.001484244130551815, 'timing/model': 1.136404860764742, 'learning_rate/qwen_vl_interface': 1.1417967133114575e-06, 'learning_rate/action_model': 1.1417967133114573e-05, 'epoch': 2.14}) +[2026-06-18 05:25:10,465][starVLA.training.train_starvla][INFO] - Step 5846, Loss: {'train/loss': 0.08932433277368546, 'train/grad_norm_pre_clip': 15.785989761352539, 'timing/data': 0.001179543323814869, 'timing/model': 3.673642374575138, 'learning_rate/qwen_vl_interface': 1.140711532884718e-06, 'learning_rate/action_model': 1.140711532884718e-05, 'epoch': 2.14}) +[2026-06-18 05:25:15,224][starVLA.training.train_starvla][INFO] - Step 5847, Loss: {'train/loss': 0.2658321261405945, 'train/grad_norm_pre_clip': 20.77288818359375, 'timing/data': 0.0011672992259263992, 'timing/model': 1.1968839364126325, 'learning_rate/qwen_vl_interface': 1.1396272043179004e-06, 'learning_rate/action_model': 1.1396272043179005e-05, 'epoch': 2.14}) +[2026-06-18 05:25:20,857][starVLA.training.train_starvla][INFO] - Step 5848, Loss: {'train/loss': 0.1663096845149994, 'train/grad_norm_pre_clip': 39.47974395751953, 'timing/data': 0.0010595601052045822, 'timing/model': 1.134198123589158, 'learning_rate/qwen_vl_interface': 1.138543727835788e-06, 'learning_rate/action_model': 1.138543727835788e-05, 'epoch': 2.14}) +[2026-06-18 05:25:25,413][starVLA.training.train_starvla][INFO] - Step 5849, Loss: {'train/loss': 0.2902345061302185, 'train/grad_norm_pre_clip': 30.199207305908203, 'timing/data': 0.0009891251102089882, 'timing/model': 1.2201490513980389, 'learning_rate/qwen_vl_interface': 1.1374611036629832e-06, 'learning_rate/action_model': 1.1374611036629831e-05, 'epoch': 2.14}) +[2026-06-18 05:25:31,078][starVLA.training.train_starvla][INFO] - Step 5850, Loss: {'train/loss': 0.07806716859340668, 'train/grad_norm_pre_clip': 22.352519989013672, 'timing/data': 0.0009163804352283478, 'timing/model': 2.131469054147601, 'learning_rate/qwen_vl_interface': 1.136379332023916e-06, 'learning_rate/action_model': 1.1363793320239161e-05, 'epoch': 2.14}) +[2026-06-18 05:25:36,072][starVLA.training.train_starvla][INFO] - Step 5851, Loss: {'train/loss': 0.32722848653793335, 'train/grad_norm_pre_clip': 25.707761764526367, 'timing/data': 0.0009076222777366638, 'timing/model': 1.057992552407086, 'learning_rate/qwen_vl_interface': 1.1352984131428413e-06, 'learning_rate/action_model': 1.1352984131428413e-05, 'epoch': 2.14}) +[2026-06-18 05:25:41,285][starVLA.training.train_starvla][INFO] - Step 5852, Loss: {'train/loss': 0.24960212409496307, 'train/grad_norm_pre_clip': 23.965852737426758, 'timing/data': 0.0008925534784793854, 'timing/model': 1.0564361419528723, 'learning_rate/qwen_vl_interface': 1.1342183472438329e-06, 'learning_rate/action_model': 1.1342183472438328e-05, 'epoch': 2.14}) +[2026-06-18 05:25:46,244][starVLA.training.train_starvla][INFO] - Step 5853, Loss: {'train/loss': 0.12134801596403122, 'train/grad_norm_pre_clip': 18.08025550842285, 'timing/data': 0.0013082800433039665, 'timing/model': 1.2366284457966685, 'learning_rate/qwen_vl_interface': 1.1331391345507905e-06, 'learning_rate/action_model': 1.1331391345507905e-05, 'epoch': 2.14}) +[2026-06-18 05:25:51,661][starVLA.training.train_starvla][INFO] - Step 5854, Loss: {'train/loss': 0.24682602286338806, 'train/grad_norm_pre_clip': 24.497509002685547, 'timing/data': 0.0009472733363509178, 'timing/model': 1.6659201253205538, 'learning_rate/qwen_vl_interface': 1.1320607752874341e-06, 'learning_rate/action_model': 1.1320607752874341e-05, 'epoch': 2.14}) +[2026-06-18 05:26:03,438][starVLA.training.train_starvla][INFO] - Step 5855, Loss: {'train/loss': 0.11028733849525452, 'train/grad_norm_pre_clip': 15.97512435913086, 'timing/data': 0.0015797875821590424, 'timing/model': 1.9764912724494934, 'learning_rate/qwen_vl_interface': 1.1309832696773104e-06, 'learning_rate/action_model': 1.1309832696773104e-05, 'epoch': 2.14}) +[2026-06-18 05:26:11,757][starVLA.training.train_starvla][INFO] - Step 5856, Loss: {'train/loss': 0.23911844193935394, 'train/grad_norm_pre_clip': 20.50959587097168, 'timing/data': 0.0010622385889291763, 'timing/model': 1.151914651505649, 'learning_rate/qwen_vl_interface': 1.1299066179437876e-06, 'learning_rate/action_model': 1.1299066179437876e-05, 'epoch': 2.14}) +[2026-06-18 05:26:17,077][starVLA.training.train_starvla][INFO] - Step 5857, Loss: {'train/loss': 0.11232256889343262, 'train/grad_norm_pre_clip': 34.99166488647461, 'timing/data': 0.004116908647119999, 'timing/model': 1.826298982836306, 'learning_rate/qwen_vl_interface': 1.1288308203100565e-06, 'learning_rate/action_model': 1.1288308203100564e-05, 'epoch': 2.14}) +[2026-06-18 05:26:25,361][starVLA.training.train_starvla][INFO] - Step 5858, Loss: {'train/loss': 0.22341765463352203, 'train/grad_norm_pre_clip': 18.589988708496094, 'timing/data': 0.0009136442095041275, 'timing/model': 1.3842610158026218, 'learning_rate/qwen_vl_interface': 1.12775587699913e-06, 'learning_rate/action_model': 1.12775587699913e-05, 'epoch': 2.15}) +[2026-06-18 05:26:35,008][starVLA.training.train_starvla][INFO] - Step 5859, Loss: {'train/loss': 0.16299457848072052, 'train/grad_norm_pre_clip': 17.944353103637695, 'timing/data': 0.0010642334818840027, 'timing/model': 2.856444153934717, 'learning_rate/qwen_vl_interface': 1.126681788233847e-06, 'learning_rate/action_model': 1.1266817882338468e-05, 'epoch': 2.15}) +[2026-06-18 05:26:49,657][starVLA.training.train_starvla][INFO] - Step 5860, Loss: {'train/loss': 0.19783495366573334, 'train/grad_norm_pre_clip': 17.182064056396484, 'timing/data': 0.0008168676868081093, 'timing/model': 1.1512646954506636, 'learning_rate/qwen_vl_interface': 1.1256085542368655e-06, 'learning_rate/action_model': 1.1256085542368655e-05, 'epoch': 2.15}) +[2026-06-18 05:26:54,611][starVLA.training.train_starvla][INFO] - Step 5861, Loss: {'train/loss': 0.4321558475494385, 'train/grad_norm_pre_clip': 25.888389587402344, 'timing/data': 0.0009046504274010658, 'timing/model': 1.1509073534980416, 'learning_rate/qwen_vl_interface': 1.1245361752306676e-06, 'learning_rate/action_model': 1.1245361752306677e-05, 'epoch': 2.15}) +[2026-06-18 05:27:03,055][starVLA.training.train_starvla][INFO] - Step 5862, Loss: {'train/loss': 0.17724044620990753, 'train/grad_norm_pre_clip': 20.223636627197266, 'timing/data': 0.0009629130363464355, 'timing/model': 1.0979639990255237, 'learning_rate/qwen_vl_interface': 1.1234646514375625e-06, 'learning_rate/action_model': 1.1234646514375624e-05, 'epoch': 2.15}) +[2026-06-18 05:27:08,425][starVLA.training.train_starvla][INFO] - Step 5863, Loss: {'train/loss': 0.2610875368118286, 'train/grad_norm_pre_clip': 23.8969669342041, 'timing/data': 0.0011068470776081085, 'timing/model': 1.296603280119598, 'learning_rate/qwen_vl_interface': 1.122393983079674e-06, 'learning_rate/action_model': 1.122393983079674e-05, 'epoch': 2.15}) +[2026-06-18 05:27:18,476][starVLA.training.train_starvla][INFO] - Step 5864, Loss: {'train/loss': 0.22874252498149872, 'train/grad_norm_pre_clip': 25.455289840698242, 'timing/data': 0.0015081092715263367, 'timing/model': 2.087947890162468, 'learning_rate/qwen_vl_interface': 1.1213241703789534e-06, 'learning_rate/action_model': 1.1213241703789534e-05, 'epoch': 2.15}) +[2026-06-18 05:27:23,600][starVLA.training.train_starvla][INFO] - Step 5865, Loss: {'train/loss': 0.3002684712409973, 'train/grad_norm_pre_clip': 23.963687896728516, 'timing/data': 0.001213032752275467, 'timing/model': 1.0516662327572703, 'learning_rate/qwen_vl_interface': 1.1202552135571759e-06, 'learning_rate/action_model': 1.1202552135571759e-05, 'epoch': 2.15}) +[2026-06-18 05:27:28,792][starVLA.training.train_starvla][INFO] - Step 5866, Loss: {'train/loss': 0.14951911568641663, 'train/grad_norm_pre_clip': 19.438854217529297, 'timing/data': 0.0007836222648620605, 'timing/model': 1.7013355875387788, 'learning_rate/qwen_vl_interface': 1.1191871128359367e-06, 'learning_rate/action_model': 1.1191871128359365e-05, 'epoch': 2.15}) +[2026-06-18 05:27:33,591][starVLA.training.train_starvla][INFO] - Step 5867, Loss: {'train/loss': 0.5205708146095276, 'train/grad_norm_pre_clip': 45.13603210449219, 'timing/data': 0.0012887585908174515, 'timing/model': 1.2162119839340448, 'learning_rate/qwen_vl_interface': 1.1181198684366548e-06, 'learning_rate/action_model': 1.1181198684366548e-05, 'epoch': 2.15}) +[2026-06-18 05:27:38,739][starVLA.training.train_starvla][INFO] - Step 5868, Loss: {'train/loss': 0.28232961893081665, 'train/grad_norm_pre_clip': 24.466819763183594, 'timing/data': 0.0014574220404028893, 'timing/model': 1.7520331675186753, 'learning_rate/qwen_vl_interface': 1.117053480580568e-06, 'learning_rate/action_model': 1.1170534805805679e-05, 'epoch': 2.15}) +[2026-06-18 05:27:43,880][starVLA.training.train_starvla][INFO] - Step 5869, Loss: {'train/loss': 0.08162470161914825, 'train/grad_norm_pre_clip': 14.668071746826172, 'timing/data': 0.0011326158419251442, 'timing/model': 1.427665558643639, 'learning_rate/qwen_vl_interface': 1.1159879494887429e-06, 'learning_rate/action_model': 1.115987949488743e-05, 'epoch': 2.15}) +[2026-06-18 05:27:53,264][starVLA.training.train_starvla][INFO] - Step 5870, Loss: {'train/loss': 0.05203814432024956, 'train/grad_norm_pre_clip': 16.3491268157959, 'timing/data': 0.0012712972238659859, 'timing/model': 1.209752220660448, 'learning_rate/qwen_vl_interface': 1.1149232753820645e-06, 'learning_rate/action_model': 1.1149232753820647e-05, 'epoch': 2.15}) +[2026-06-18 05:27:59,207][starVLA.training.train_starvla][INFO] - Step 5871, Loss: {'train/loss': 0.22575074434280396, 'train/grad_norm_pre_clip': 18.84040069580078, 'timing/data': 0.0008308086544275284, 'timing/model': 1.9997169617563486, 'learning_rate/qwen_vl_interface': 1.1138594584812406e-06, 'learning_rate/action_model': 1.1138594584812406e-05, 'epoch': 2.15}) +[2026-06-18 05:28:04,993][starVLA.training.train_starvla][INFO] - Step 5872, Loss: {'train/loss': 0.31288060545921326, 'train/grad_norm_pre_clip': 34.69012451171875, 'timing/data': 0.0010806890204548836, 'timing/model': 1.1406831974163651, 'learning_rate/qwen_vl_interface': 1.1127964990068017e-06, 'learning_rate/action_model': 1.1127964990068017e-05, 'epoch': 2.15}) +[2026-06-18 05:28:09,967][starVLA.training.train_starvla][INFO] - Step 5873, Loss: {'train/loss': 0.0828198492527008, 'train/grad_norm_pre_clip': 13.630985260009766, 'timing/data': 0.0011285748332738876, 'timing/model': 1.2361723557114601, 'learning_rate/qwen_vl_interface': 1.1117343971791005e-06, 'learning_rate/action_model': 1.1117343971791006e-05, 'epoch': 2.15}) +[2026-06-18 05:28:14,313][starVLA.training.train_starvla][INFO] - Step 5874, Loss: {'train/loss': 0.3849267065525055, 'train/grad_norm_pre_clip': 27.2960147857666, 'timing/data': 0.0011014509946107864, 'timing/model': 1.063327600248158, 'learning_rate/qwen_vl_interface': 1.1106731532183114e-06, 'learning_rate/action_model': 1.1106731532183114e-05, 'epoch': 2.15}) +[2026-06-18 05:28:19,450][starVLA.training.train_starvla][INFO] - Step 5875, Loss: {'train/loss': 0.1853870153427124, 'train/grad_norm_pre_clip': 19.03400230407715, 'timing/data': 0.0008957330137491226, 'timing/model': 1.2212316086515784, 'learning_rate/qwen_vl_interface': 1.1096127673444309e-06, 'learning_rate/action_model': 1.109612767344431e-05, 'epoch': 2.15}) +[2026-06-18 05:28:24,259][starVLA.training.train_starvla][INFO] - Step 5876, Loss: {'train/loss': 0.5253819227218628, 'train/grad_norm_pre_clip': 35.21147537231445, 'timing/data': 0.0011942293494939804, 'timing/model': 1.1113157998770475, 'learning_rate/qwen_vl_interface': 1.108553239777281e-06, 'learning_rate/action_model': 1.108553239777281e-05, 'epoch': 2.15}) +[2026-06-18 05:28:29,287][starVLA.training.train_starvla][INFO] - Step 5877, Loss: {'train/loss': 0.1470467448234558, 'train/grad_norm_pre_clip': 20.075668334960938, 'timing/data': 0.0010589305311441422, 'timing/model': 1.5318078873679042, 'learning_rate/qwen_vl_interface': 1.1074945707365004e-06, 'learning_rate/action_model': 1.1074945707365003e-05, 'epoch': 2.15}) +[2026-06-18 05:28:34,082][starVLA.training.train_starvla][INFO] - Step 5878, Loss: {'train/loss': 0.22893476486206055, 'train/grad_norm_pre_clip': 19.68857192993164, 'timing/data': 0.0011943131685256958, 'timing/model': 1.0966084636747837, 'learning_rate/qwen_vl_interface': 1.1064367604415522e-06, 'learning_rate/action_model': 1.1064367604415521e-05, 'epoch': 2.15}) +[2026-06-18 05:28:40,116][starVLA.training.train_starvla][INFO] - Step 5879, Loss: {'train/loss': 0.6310192942619324, 'train/grad_norm_pre_clip': 40.02327346801758, 'timing/data': 0.001092381775379181, 'timing/model': 1.333708880469203, 'learning_rate/qwen_vl_interface': 1.105379809111721e-06, 'learning_rate/action_model': 1.105379809111721e-05, 'epoch': 2.15}) +[2026-06-18 05:28:45,226][starVLA.training.train_starvla][INFO] - Step 5880, Loss: {'train/loss': 0.26697200536727905, 'train/grad_norm_pre_clip': 22.059213638305664, 'timing/data': 0.0011836271733045578, 'timing/model': 1.0552595322951674, 'learning_rate/qwen_vl_interface': 1.1043237169661167e-06, 'learning_rate/action_model': 1.1043237169661166e-05, 'epoch': 2.15}) +[2026-06-18 05:28:53,409][starVLA.training.train_starvla][INFO] - Step 5881, Loss: {'train/loss': 0.35185444355010986, 'train/grad_norm_pre_clip': 29.711933135986328, 'timing/data': 0.001544162631034851, 'timing/model': 4.450942114926875, 'learning_rate/qwen_vl_interface': 1.103268484223667e-06, 'learning_rate/action_model': 1.1032684842236669e-05, 'epoch': 2.15}) +[2026-06-18 05:29:04,341][starVLA.training.train_starvla][INFO] - Step 5882, Loss: {'train/loss': 0.536075234413147, 'train/grad_norm_pre_clip': 38.14944076538086, 'timing/data': 0.001132916659116745, 'timing/model': 1.7821154100820422, 'learning_rate/qwen_vl_interface': 1.102214111103123e-06, 'learning_rate/action_model': 1.102214111103123e-05, 'epoch': 2.15}) +[2026-06-18 05:29:08,976][starVLA.training.train_starvla][INFO] - Step 5883, Loss: {'train/loss': 0.1928517371416092, 'train/grad_norm_pre_clip': 19.896907806396484, 'timing/data': 0.0013090474531054497, 'timing/model': 1.54715939424932, 'learning_rate/qwen_vl_interface': 1.1011605978230552e-06, 'learning_rate/action_model': 1.1011605978230551e-05, 'epoch': 2.15}) +[2026-06-18 05:29:12,982][starVLA.training.train_starvla][INFO] - Step 5884, Loss: {'train/loss': 0.07846939563751221, 'train/grad_norm_pre_clip': 15.0254487991333, 'timing/data': 0.0012059863656759262, 'timing/model': 1.011457834392786, 'learning_rate/qwen_vl_interface': 1.100107944601861e-06, 'learning_rate/action_model': 1.100107944601861e-05, 'epoch': 2.15}) +[2026-06-18 05:29:17,236][starVLA.training.train_starvla][INFO] - Step 5885, Loss: {'train/loss': 0.25717949867248535, 'train/grad_norm_pre_clip': 25.504169464111328, 'timing/data': 0.0009332941845059395, 'timing/model': 1.0022071432322264, 'learning_rate/qwen_vl_interface': 1.0990561516577552e-06, 'learning_rate/action_model': 1.0990561516577552e-05, 'epoch': 2.15}) +[2026-06-18 05:29:23,597][starVLA.training.train_starvla][INFO] - Step 5886, Loss: {'train/loss': 0.2749980092048645, 'train/grad_norm_pre_clip': 29.269386291503906, 'timing/data': 0.0009026797488331795, 'timing/model': 1.0251198653131723, 'learning_rate/qwen_vl_interface': 1.0980052192087753e-06, 'learning_rate/action_model': 1.0980052192087754e-05, 'epoch': 2.16}) +[2026-06-18 05:29:27,826][starVLA.training.train_starvla][INFO] - Step 5887, Loss: {'train/loss': 0.4636159837245941, 'train/grad_norm_pre_clip': 28.53581428527832, 'timing/data': 0.0010730838403105736, 'timing/model': 1.0141123700886965, 'learning_rate/qwen_vl_interface': 1.096955147472782e-06, 'learning_rate/action_model': 1.0969551474727821e-05, 'epoch': 2.16}) +[2026-06-18 05:29:32,028][starVLA.training.train_starvla][INFO] - Step 5888, Loss: {'train/loss': 0.4155680239200592, 'train/grad_norm_pre_clip': 44.517581939697266, 'timing/data': 0.0010362137109041214, 'timing/model': 1.036843752488494, 'learning_rate/qwen_vl_interface': 1.095905936667455e-06, 'learning_rate/action_model': 1.0959059366674552e-05, 'epoch': 2.16}) +[2026-06-18 05:29:36,124][starVLA.training.train_starvla][INFO] - Step 5889, Loss: {'train/loss': 0.3217630982398987, 'train/grad_norm_pre_clip': 28.690982818603516, 'timing/data': 0.0009499341249465942, 'timing/model': 1.1690682237967849, 'learning_rate/qwen_vl_interface': 1.0948575870102977e-06, 'learning_rate/action_model': 1.0948575870102978e-05, 'epoch': 2.16}) +[2026-06-18 05:29:44,379][starVLA.training.train_starvla][INFO] - Step 5890, Loss: {'train/loss': 0.12447091192007065, 'train/grad_norm_pre_clip': 12.660326957702637, 'timing/data': 0.0009160032495856285, 'timing/model': 1.0699285631999373, 'learning_rate/qwen_vl_interface': 1.0938100987186325e-06, 'learning_rate/action_model': 1.0938100987186323e-05, 'epoch': 2.16}) +[2026-06-18 05:29:48,669][starVLA.training.train_starvla][INFO] - Step 5891, Loss: {'train/loss': 0.08984940499067307, 'train/grad_norm_pre_clip': 20.2423038482666, 'timing/data': 0.0011372687295079231, 'timing/model': 1.01916712988168, 'learning_rate/qwen_vl_interface': 1.092763472009608e-06, 'learning_rate/action_model': 1.092763472009608e-05, 'epoch': 2.16}) +[2026-06-18 05:29:56,433][starVLA.training.train_starvla][INFO] - Step 5892, Loss: {'train/loss': 0.20720872282981873, 'train/grad_norm_pre_clip': 16.88269805908203, 'timing/data': 0.0008425479754805565, 'timing/model': 1.9878466399386525, 'learning_rate/qwen_vl_interface': 1.0917177071001884e-06, 'learning_rate/action_model': 1.0917177071001884e-05, 'epoch': 2.16}) +[2026-06-18 05:30:05,478][starVLA.training.train_starvla][INFO] - Step 5893, Loss: {'train/loss': 0.11954817920923233, 'train/grad_norm_pre_clip': 16.449716567993164, 'timing/data': 0.0021068742498755455, 'timing/model': 5.668540775775909, 'learning_rate/qwen_vl_interface': 1.0906728042071627e-06, 'learning_rate/action_model': 1.0906728042071627e-05, 'epoch': 2.16}) +[2026-06-18 05:30:10,923][starVLA.training.train_starvla][INFO] - Step 5894, Loss: {'train/loss': 0.16425897181034088, 'train/grad_norm_pre_clip': 16.796297073364258, 'timing/data': 0.0013296809047460556, 'timing/model': 1.790341499261558, 'learning_rate/qwen_vl_interface': 1.089628763547139e-06, 'learning_rate/action_model': 1.0896287635471389e-05, 'epoch': 2.16}) +[2026-06-18 05:30:20,220][starVLA.training.train_starvla][INFO] - Step 5895, Loss: {'train/loss': 0.29812654852867126, 'train/grad_norm_pre_clip': 31.18526840209961, 'timing/data': 0.0011093737557530403, 'timing/model': 1.218226721510291, 'learning_rate/qwen_vl_interface': 1.0885855853365505e-06, 'learning_rate/action_model': 1.0885855853365503e-05, 'epoch': 2.16}) +[2026-06-18 05:30:34,752][starVLA.training.train_starvla][INFO] - Step 5896, Loss: {'train/loss': 0.33328142762184143, 'train/grad_norm_pre_clip': 33.773033142089844, 'timing/data': 0.0012762602418661118, 'timing/model': 1.2016892591491342, 'learning_rate/qwen_vl_interface': 1.0875432697916492e-06, 'learning_rate/action_model': 1.087543269791649e-05, 'epoch': 2.16}) +[2026-06-18 05:30:52,664][starVLA.training.train_starvla][INFO] - Step 5897, Loss: {'train/loss': 0.42376965284347534, 'train/grad_norm_pre_clip': 29.4844970703125, 'timing/data': 0.0008704392239451408, 'timing/model': 4.4448955077677965, 'learning_rate/qwen_vl_interface': 1.0865018171285047e-06, 'learning_rate/action_model': 1.0865018171285047e-05, 'epoch': 2.16}) +[2026-06-18 05:31:03,367][starVLA.training.train_starvla][INFO] - Step 5898, Loss: {'train/loss': 0.260403573513031, 'train/grad_norm_pre_clip': 20.882875442504883, 'timing/data': 0.0037581901997327805, 'timing/model': 1.4198758825659752, 'learning_rate/qwen_vl_interface': 1.0854612275630146e-06, 'learning_rate/action_model': 1.0854612275630146e-05, 'epoch': 2.16}) +[2026-06-18 05:31:17,836][starVLA.training.train_starvla][INFO] - Step 5899, Loss: {'train/loss': 0.687920868396759, 'train/grad_norm_pre_clip': 46.402099609375, 'timing/data': 0.00189150869846344, 'timing/model': 1.8719509160146117, 'learning_rate/qwen_vl_interface': 1.0844215013108936e-06, 'learning_rate/action_model': 1.0844215013108936e-05, 'epoch': 2.16}) +[2026-06-18 05:35:19,290][starVLA.training.train_starvla][INFO] - Step 5900, Loss: {'train/loss': 0.43802958726882935, 'train/grad_norm_pre_clip': 27.122705459594727, 'eval/loss': 0.5387254953384399, 'eval/latency_2/loss': 0.5387254953384399, 'eval/demon_fire_precision': 0.9427168576104746, 'eval/demon_fire_recall': 0.9640167364016736, 'eval/demon_fire_f1': 0.9532478278858088, 'eval/demon_fire_rate_model': 0.6800222593210907, 'eval/demon_fire_rate_teacher': 0.6649972175848636, 'eval/demon_left_precision': 0.9091880341880342, 'eval/demon_left_recall': 0.8957894736842106, 'eval/demon_left_f1': 0.9024390243902439, 'eval/demon_left_rate_model': 0.2604340567612688, 'eval/demon_left_rate_teacher': 0.2643294379521425, 'eval/demon_right_precision': 0.9011111111111111, 'eval/demon_right_recall': 0.9163841807909604, 'eval/demon_right_f1': 0.9086834733893556, 'eval/demon_right_rate_model': 0.25041736227045075, 'eval/demon_right_rate_teacher': 0.24624373956594323, 'eval/demon_move_f1': 0.9055612488897997, 'eval/demon_cc_f1': 0.9384227118046, 'eval/latency_2/demon_fire_precision': 0.9427168576104746, 'eval/latency_2/demon_fire_recall': 0.9640167364016736, 'eval/latency_2/demon_fire_f1': 0.9532478278858088, 'eval/latency_2/demon_fire_rate_model': 0.6800222593210907, 'eval/latency_2/demon_fire_rate_teacher': 0.6649972175848636, 'eval/latency_2/demon_left_precision': 0.9091880341880342, 'eval/latency_2/demon_left_recall': 0.8957894736842106, 'eval/latency_2/demon_left_f1': 0.9024390243902439, 'eval/latency_2/demon_left_rate_model': 0.2604340567612688, 'eval/latency_2/demon_left_rate_teacher': 0.2643294379521425, 'eval/latency_2/demon_right_precision': 0.9011111111111111, 'eval/latency_2/demon_right_recall': 0.9163841807909604, 'eval/latency_2/demon_right_f1': 0.9086834733893556, 'eval/latency_2/demon_right_rate_model': 0.25041736227045075, 'eval/latency_2/demon_right_rate_teacher': 0.24624373956594323, 'eval/latency_2/demon_move_f1': 0.9055612488897997, 'eval/latency_2/demon_cc_f1': 0.9384227118046, 'timing/data': 0.0009396178647875786, 'timing/model': 1.0954065704718232, 'learning_rate/qwen_vl_interface': 1.0833826385876772e-06, 'learning_rate/action_model': 1.0833826385876771e-05, 'epoch': 2.16}) +[2026-06-18 05:35:19,294][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_5900_state +[2026-06-18 05:35:19,302][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 05:35:28,984][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_5900_state/model.safetensors +[2026-06-18 05:35:44,914][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_5900_state/optimizer.bin +[2026-06-18 05:35:44,915][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_5900_state/sampler.bin +[2026-06-18 05:35:44,915][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_5900_state/sampler_1.bin +[2026-06-18 05:35:44,918][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_5900_state/random_states_0.pkl +[2026-06-18 05:35:44,941][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 5900 +[2026-06-18 05:35:44,942][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 05:35:44,960][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 05:35:49,349][starVLA.training.train_starvla][INFO] - Step 5901, Loss: {'train/loss': 0.4493955373764038, 'train/grad_norm_pre_clip': 32.41933059692383, 'timing/data': 0.0025815777480602264, 'timing/model': 1.1136253969743848, 'learning_rate/qwen_vl_interface': 1.0823446396087239e-06, 'learning_rate/action_model': 1.0823446396087239e-05, 'epoch': 2.16}) +[2026-06-18 05:35:58,938][starVLA.training.train_starvla][INFO] - Step 5902, Loss: {'train/loss': 0.1995452344417572, 'train/grad_norm_pre_clip': 17.709556579589844, 'timing/data': 0.0008797179907560349, 'timing/model': 1.2937456062063575, 'learning_rate/qwen_vl_interface': 1.0813075045892116e-06, 'learning_rate/action_model': 1.0813075045892116e-05, 'epoch': 2.16}) +[2026-06-18 05:36:03,672][starVLA.training.train_starvla][INFO] - Step 5903, Loss: {'train/loss': 0.24683524668216705, 'train/grad_norm_pre_clip': 27.064876556396484, 'timing/data': 0.0012880945578217506, 'timing/model': 1.0686966693028808, 'learning_rate/qwen_vl_interface': 1.0802712337441383e-06, 'learning_rate/action_model': 1.0802712337441384e-05, 'epoch': 2.16}) +[2026-06-18 05:36:08,708][starVLA.training.train_starvla][INFO] - Step 5904, Loss: {'train/loss': 0.43727466464042664, 'train/grad_norm_pre_clip': 33.21694564819336, 'timing/data': 0.0010284557938575745, 'timing/model': 1.1937721651047468, 'learning_rate/qwen_vl_interface': 1.0792358272883262e-06, 'learning_rate/action_model': 1.0792358272883262e-05, 'epoch': 2.16}) +[2026-06-18 05:36:15,306][starVLA.training.train_starvla][INFO] - Step 5905, Loss: {'train/loss': 0.5936453342437744, 'train/grad_norm_pre_clip': 31.547718048095703, 'timing/data': 0.0009600156918168068, 'timing/model': 1.4737250758334994, 'learning_rate/qwen_vl_interface': 1.0782012854364145e-06, 'learning_rate/action_model': 1.0782012854364145e-05, 'epoch': 2.16}) +[2026-06-18 05:36:19,762][starVLA.training.train_starvla][INFO] - Step 5906, Loss: {'train/loss': 0.1495450735092163, 'train/grad_norm_pre_clip': 14.988433837890625, 'timing/data': 0.0007562711834907532, 'timing/model': 1.2187994727864861, 'learning_rate/qwen_vl_interface': 1.0771676084028652e-06, 'learning_rate/action_model': 1.0771676084028652e-05, 'epoch': 2.16}) +[2026-06-18 05:36:24,395][starVLA.training.train_starvla][INFO] - Step 5907, Loss: {'train/loss': 0.10600896179676056, 'train/grad_norm_pre_clip': 11.237716674804688, 'timing/data': 0.0013345805928111076, 'timing/model': 1.0814683614298701, 'learning_rate/qwen_vl_interface': 1.0761347964019611e-06, 'learning_rate/action_model': 1.076134796401961e-05, 'epoch': 2.16}) +[2026-06-18 05:36:35,123][starVLA.training.train_starvla][INFO] - Step 5908, Loss: {'train/loss': 0.07256967574357986, 'train/grad_norm_pre_clip': 14.842073440551758, 'timing/data': 0.001107371412217617, 'timing/model': 5.7991795893758535, 'learning_rate/qwen_vl_interface': 1.075102849647804e-06, 'learning_rate/action_model': 1.0751028496478038e-05, 'epoch': 2.16}) +[2026-06-18 05:36:40,036][starVLA.training.train_starvla][INFO] - Step 5909, Loss: {'train/loss': 0.2205979824066162, 'train/grad_norm_pre_clip': 25.587543487548828, 'timing/data': 0.0007732026278972626, 'timing/model': 1.4464237838983536, 'learning_rate/qwen_vl_interface': 1.0740717683543193e-06, 'learning_rate/action_model': 1.0740717683543191e-05, 'epoch': 2.16}) +[2026-06-18 05:36:45,039][starVLA.training.train_starvla][INFO] - Step 5910, Loss: {'train/loss': 0.0839603915810585, 'train/grad_norm_pre_clip': 24.978601455688477, 'timing/data': 0.0009260112419724464, 'timing/model': 1.4875238239765167, 'learning_rate/qwen_vl_interface': 1.0730415527352518e-06, 'learning_rate/action_model': 1.0730415527352517e-05, 'epoch': 2.16}) +[2026-06-18 05:36:49,405][starVLA.training.train_starvla][INFO] - Step 5911, Loss: {'train/loss': 0.3650103807449341, 'train/grad_norm_pre_clip': 24.879314422607422, 'timing/data': 0.0015404429286718369, 'timing/model': 1.12860909383744, 'learning_rate/qwen_vl_interface': 1.0720122030041638e-06, 'learning_rate/action_model': 1.0720122030041637e-05, 'epoch': 2.16}) +[2026-06-18 05:36:54,785][starVLA.training.train_starvla][INFO] - Step 5912, Loss: {'train/loss': 0.10177426785230637, 'train/grad_norm_pre_clip': 14.376005172729492, 'timing/data': 0.0009456314146518707, 'timing/model': 1.311469929292798, 'learning_rate/qwen_vl_interface': 1.0709837193744412e-06, 'learning_rate/action_model': 1.070983719374441e-05, 'epoch': 2.16}) +[2026-06-18 05:36:59,256][starVLA.training.train_starvla][INFO] - Step 5913, Loss: {'train/loss': 0.5789754986763, 'train/grad_norm_pre_clip': 33.88555908203125, 'timing/data': 0.0010896893218159676, 'timing/model': 1.0696822768077254, 'learning_rate/qwen_vl_interface': 1.069956102059291e-06, 'learning_rate/action_model': 1.069956102059291e-05, 'epoch': 2.17}) +[2026-06-18 05:37:04,414][starVLA.training.train_starvla][INFO] - Step 5914, Loss: {'train/loss': 0.2088288515806198, 'train/grad_norm_pre_clip': 27.922950744628906, 'timing/data': 0.0015914766117930412, 'timing/model': 1.1842769598588347, 'learning_rate/qwen_vl_interface': 1.0689293512717387e-06, 'learning_rate/action_model': 1.0689293512717385e-05, 'epoch': 2.17}) +[2026-06-18 05:37:09,732][starVLA.training.train_starvla][INFO] - Step 5915, Loss: {'train/loss': 0.23445524275302887, 'train/grad_norm_pre_clip': 23.538570404052734, 'timing/data': 0.0007859142497181892, 'timing/model': 1.2470382787287235, 'learning_rate/qwen_vl_interface': 1.0679034672246322e-06, 'learning_rate/action_model': 1.0679034672246322e-05, 'epoch': 2.17}) +[2026-06-18 05:37:14,449][starVLA.training.train_starvla][INFO] - Step 5916, Loss: {'train/loss': 0.18512071669101715, 'train/grad_norm_pre_clip': 16.5882625579834, 'timing/data': 0.0009027943015098572, 'timing/model': 1.1111891129985452, 'learning_rate/qwen_vl_interface': 1.0668784501306348e-06, 'learning_rate/action_model': 1.0668784501306348e-05, 'epoch': 2.17}) +[2026-06-18 05:37:22,911][starVLA.training.train_starvla][INFO] - Step 5917, Loss: {'train/loss': 0.2709487974643707, 'train/grad_norm_pre_clip': 23.037961959838867, 'timing/data': 0.0008703740313649178, 'timing/model': 1.1033697202801704, 'learning_rate/qwen_vl_interface': 1.0658543002022377e-06, 'learning_rate/action_model': 1.0658543002022375e-05, 'epoch': 2.17}) +[2026-06-18 05:37:27,293][starVLA.training.train_starvla][INFO] - Step 5918, Loss: {'train/loss': 0.2914854884147644, 'train/grad_norm_pre_clip': 60.136898040771484, 'timing/data': 0.001104608178138733, 'timing/model': 1.03849495947361, 'learning_rate/qwen_vl_interface': 1.0648310176517454e-06, 'learning_rate/action_model': 1.0648310176517453e-05, 'epoch': 2.17}) +[2026-06-18 05:37:36,176][starVLA.training.train_starvla][INFO] - Step 5919, Loss: {'train/loss': 0.21124206483364105, 'train/grad_norm_pre_clip': 15.911157608032227, 'timing/data': 0.0008634738624095917, 'timing/model': 1.5518216453492641, 'learning_rate/qwen_vl_interface': 1.0638086026912873e-06, 'learning_rate/action_model': 1.0638086026912872e-05, 'epoch': 2.17}) +[2026-06-18 05:37:41,213][starVLA.training.train_starvla][INFO] - Step 5920, Loss: {'train/loss': 0.033605583012104034, 'train/grad_norm_pre_clip': 5.158095359802246, 'timing/data': 0.00125208031386137, 'timing/model': 1.3511272948235273, 'learning_rate/qwen_vl_interface': 1.0627870555328102e-06, 'learning_rate/action_model': 1.06278705553281e-05, 'epoch': 2.17}) +[2026-06-18 05:37:45,714][starVLA.training.train_starvla][INFO] - Step 5921, Loss: {'train/loss': 0.2691096067428589, 'train/grad_norm_pre_clip': 24.707103729248047, 'timing/data': 0.001160324551165104, 'timing/model': 1.3371629137545824, 'learning_rate/qwen_vl_interface': 1.061766376388082e-06, 'learning_rate/action_model': 1.0617663763880821e-05, 'epoch': 2.17}) +[2026-06-18 05:37:50,716][starVLA.training.train_starvla][INFO] - Step 5922, Loss: {'train/loss': 0.2507205009460449, 'train/grad_norm_pre_clip': 22.928451538085938, 'timing/data': 0.0007548974826931953, 'timing/model': 1.3609898192808032, 'learning_rate/qwen_vl_interface': 1.0607465654686908e-06, 'learning_rate/action_model': 1.0607465654686907e-05, 'epoch': 2.17}) +[2026-06-18 05:37:59,476][starVLA.training.train_starvla][INFO] - Step 5923, Loss: {'train/loss': 0.23921428620815277, 'train/grad_norm_pre_clip': 23.35007667541504, 'timing/data': 0.001340152695775032, 'timing/model': 1.1803473886102438, 'learning_rate/qwen_vl_interface': 1.0597276229860434e-06, 'learning_rate/action_model': 1.0597276229860434e-05, 'epoch': 2.17}) +[2026-06-18 05:38:04,725][starVLA.training.train_starvla][INFO] - Step 5924, Loss: {'train/loss': 0.3223722279071808, 'train/grad_norm_pre_clip': 23.152515411376953, 'timing/data': 0.00100737065076828, 'timing/model': 1.9348900625482202, 'learning_rate/qwen_vl_interface': 1.0587095491513709e-06, 'learning_rate/action_model': 1.0587095491513709e-05, 'epoch': 2.17}) +[2026-06-18 05:38:12,896][starVLA.training.train_starvla][INFO] - Step 5925, Loss: {'train/loss': 0.031468287110328674, 'train/grad_norm_pre_clip': 5.060885906219482, 'timing/data': 0.0008997498080134392, 'timing/model': 1.5463619576767087, 'learning_rate/qwen_vl_interface': 1.0576923441757174e-06, 'learning_rate/action_model': 1.0576923441757176e-05, 'epoch': 2.17}) +[2026-06-18 05:38:22,207][starVLA.training.train_starvla][INFO] - Step 5926, Loss: {'train/loss': 0.4634411931037903, 'train/grad_norm_pre_clip': 33.64118194580078, 'timing/data': 0.0013751629739999771, 'timing/model': 5.983407166786492, 'learning_rate/qwen_vl_interface': 1.0566760082699515e-06, 'learning_rate/action_model': 1.0566760082699514e-05, 'epoch': 2.17}) +[2026-06-18 05:38:31,735][starVLA.training.train_starvla][INFO] - Step 5927, Loss: {'train/loss': 0.2722856104373932, 'train/grad_norm_pre_clip': 32.23851776123047, 'timing/data': 0.0012896982952952385, 'timing/model': 3.712864251807332, 'learning_rate/qwen_vl_interface': 1.0556605416447626e-06, 'learning_rate/action_model': 1.0556605416447625e-05, 'epoch': 2.17}) +[2026-06-18 05:38:38,018][starVLA.training.train_starvla][INFO] - Step 5928, Loss: {'train/loss': 0.3047150671482086, 'train/grad_norm_pre_clip': 28.31709861755371, 'timing/data': 0.000897970050573349, 'timing/model': 1.100344288162887, 'learning_rate/qwen_vl_interface': 1.054645944510656e-06, 'learning_rate/action_model': 1.054645944510656e-05, 'epoch': 2.17}) +[2026-06-18 05:38:45,964][starVLA.training.train_starvla][INFO] - Step 5929, Loss: {'train/loss': 0.22787372767925262, 'train/grad_norm_pre_clip': 28.653888702392578, 'timing/data': 0.0018706582486629486, 'timing/model': 4.166648594662547, 'learning_rate/qwen_vl_interface': 1.0536322170779605e-06, 'learning_rate/action_model': 1.0536322170779606e-05, 'epoch': 2.17}) +[2026-06-18 05:38:55,141][starVLA.training.train_starvla][INFO] - Step 5930, Loss: {'train/loss': 0.5380223393440247, 'train/grad_norm_pre_clip': 36.15243911743164, 'timing/data': 0.001131589524447918, 'timing/model': 1.6662888415157795, 'learning_rate/qwen_vl_interface': 1.0526193595568203e-06, 'learning_rate/action_model': 1.0526193595568203e-05, 'epoch': 2.17}) +[2026-06-18 05:39:00,051][starVLA.training.train_starvla][INFO] - Step 5931, Loss: {'train/loss': 0.17402851581573486, 'train/grad_norm_pre_clip': 21.8713321685791, 'timing/data': 0.0032611945644021034, 'timing/model': 1.2938742935657501, 'learning_rate/qwen_vl_interface': 1.0516073721572043e-06, 'learning_rate/action_model': 1.0516073721572044e-05, 'epoch': 2.17}) +[2026-06-18 05:39:05,140][starVLA.training.train_starvla][INFO] - Step 5932, Loss: {'train/loss': 0.30764469504356384, 'train/grad_norm_pre_clip': 24.342037200927734, 'timing/data': 0.00096152164041996, 'timing/model': 1.2702968204393983, 'learning_rate/qwen_vl_interface': 1.0505962550888982e-06, 'learning_rate/action_model': 1.0505962550888982e-05, 'epoch': 2.17}) +[2026-06-18 05:39:10,291][starVLA.training.train_starvla][INFO] - Step 5933, Loss: {'train/loss': 0.24118392169475555, 'train/grad_norm_pre_clip': 23.870019912719727, 'timing/data': 0.001154591329395771, 'timing/model': 1.2303003501147032, 'learning_rate/qwen_vl_interface': 1.0495860085615065e-06, 'learning_rate/action_model': 1.0495860085615064e-05, 'epoch': 2.17}) +[2026-06-18 05:39:15,323][starVLA.training.train_starvla][INFO] - Step 5934, Loss: {'train/loss': 0.3337899446487427, 'train/grad_norm_pre_clip': 24.802268981933594, 'timing/data': 0.0007409555837512016, 'timing/model': 1.5198832275345922, 'learning_rate/qwen_vl_interface': 1.0485766327844558e-06, 'learning_rate/action_model': 1.0485766327844557e-05, 'epoch': 2.17}) +[2026-06-18 05:39:19,940][starVLA.training.train_starvla][INFO] - Step 5935, Loss: {'train/loss': 0.5457342863082886, 'train/grad_norm_pre_clip': 39.12244415283203, 'timing/data': 0.0008831312879920006, 'timing/model': 1.392841531895101, 'learning_rate/qwen_vl_interface': 1.0475681279669892e-06, 'learning_rate/action_model': 1.0475681279669891e-05, 'epoch': 2.17}) +[2026-06-18 05:39:24,529][starVLA.training.train_starvla][INFO] - Step 5936, Loss: {'train/loss': 0.08834241330623627, 'train/grad_norm_pre_clip': 6.96882438659668, 'timing/data': 0.0007731374353170395, 'timing/model': 1.2774690678343177, 'learning_rate/qwen_vl_interface': 1.0465604943181717e-06, 'learning_rate/action_model': 1.0465604943181717e-05, 'epoch': 2.17}) +[2026-06-18 05:39:30,443][starVLA.training.train_starvla][INFO] - Step 5937, Loss: {'train/loss': 0.0835992768406868, 'train/grad_norm_pre_clip': 14.283975601196289, 'timing/data': 0.0031887954100966454, 'timing/model': 1.3048414988443255, 'learning_rate/qwen_vl_interface': 1.0455537320468861e-06, 'learning_rate/action_model': 1.045553732046886e-05, 'epoch': 2.17}) +[2026-06-18 05:39:40,318][starVLA.training.train_starvla][INFO] - Step 5938, Loss: {'train/loss': 0.27195048332214355, 'train/grad_norm_pre_clip': 15.144667625427246, 'timing/data': 0.0014488762244582176, 'timing/model': 4.729716164059937, 'learning_rate/qwen_vl_interface': 1.0445478413618382e-06, 'learning_rate/action_model': 1.0445478413618382e-05, 'epoch': 2.17}) +[2026-06-18 05:39:45,402][starVLA.training.train_starvla][INFO] - Step 5939, Loss: {'train/loss': 0.09785279631614685, 'train/grad_norm_pre_clip': 10.971773147583008, 'timing/data': 0.0010845465585589409, 'timing/model': 1.0619086557999253, 'learning_rate/qwen_vl_interface': 1.0435428224715472e-06, 'learning_rate/action_model': 1.0435428224715471e-05, 'epoch': 2.17}) +[2026-06-18 05:39:52,160][starVLA.training.train_starvla][INFO] - Step 5940, Loss: {'train/loss': 0.4405197501182556, 'train/grad_norm_pre_clip': 30.733585357666016, 'timing/data': 0.010682561434805393, 'timing/model': 3.329888703301549, 'learning_rate/qwen_vl_interface': 1.0425386755843547e-06, 'learning_rate/action_model': 1.0425386755843547e-05, 'epoch': 2.18}) +[2026-06-18 05:39:57,082][starVLA.training.train_starvla][INFO] - Step 5941, Loss: {'train/loss': 0.2770138084888458, 'train/grad_norm_pre_clip': 24.762969970703125, 'timing/data': 0.0009201355278491974, 'timing/model': 1.2348154401406646, 'learning_rate/qwen_vl_interface': 1.0415354009084219e-06, 'learning_rate/action_model': 1.041535400908422e-05, 'epoch': 2.18}) +[2026-06-18 05:40:01,879][starVLA.training.train_starvla][INFO] - Step 5942, Loss: {'train/loss': 0.20657120645046234, 'train/grad_norm_pre_clip': 28.852231979370117, 'timing/data': 0.001128142699599266, 'timing/model': 1.1505766762420535, 'learning_rate/qwen_vl_interface': 1.0405329986517301e-06, 'learning_rate/action_model': 1.0405329986517301e-05, 'epoch': 2.18}) +[2026-06-18 05:40:10,419][starVLA.training.train_starvla][INFO] - Step 5943, Loss: {'train/loss': 0.012207749299705029, 'train/grad_norm_pre_clip': 2.452702760696411, 'timing/data': 0.0010721823200583458, 'timing/model': 4.120185655541718, 'learning_rate/qwen_vl_interface': 1.0395314690220776e-06, 'learning_rate/action_model': 1.0395314690220776e-05, 'epoch': 2.18}) +[2026-06-18 05:40:15,649][starVLA.training.train_starvla][INFO] - Step 5944, Loss: {'train/loss': 0.4182901978492737, 'train/grad_norm_pre_clip': 31.142589569091797, 'timing/data': 0.0011712778359651566, 'timing/model': 1.4895923817530274, 'learning_rate/qwen_vl_interface': 1.0385308122270815e-06, 'learning_rate/action_model': 1.0385308122270814e-05, 'epoch': 2.18}) +[2026-06-18 05:40:20,471][starVLA.training.train_starvla][INFO] - Step 5945, Loss: {'train/loss': 0.5948814153671265, 'train/grad_norm_pre_clip': 34.5493278503418, 'timing/data': 0.000778423622250557, 'timing/model': 1.429690659046173, 'learning_rate/qwen_vl_interface': 1.0375310284741787e-06, 'learning_rate/action_model': 1.0375310284741788e-05, 'epoch': 2.18}) +[2026-06-18 05:40:26,488][starVLA.training.train_starvla][INFO] - Step 5946, Loss: {'train/loss': 0.5103457570075989, 'train/grad_norm_pre_clip': 27.022926330566406, 'timing/data': 0.0009535020217299461, 'timing/model': 2.8524298844859004, 'learning_rate/qwen_vl_interface': 1.0365321179706276e-06, 'learning_rate/action_model': 1.0365321179706276e-05, 'epoch': 2.18}) +[2026-06-18 05:40:33,788][starVLA.training.train_starvla][INFO] - Step 5947, Loss: {'train/loss': 0.2578674852848053, 'train/grad_norm_pre_clip': 70.1259536743164, 'timing/data': 0.001148042269051075, 'timing/model': 1.481590353883803, 'learning_rate/qwen_vl_interface': 1.035534080923503e-06, 'learning_rate/action_model': 1.0355340809235029e-05, 'epoch': 2.18}) +[2026-06-18 05:40:38,888][starVLA.training.train_starvla][INFO] - Step 5948, Loss: {'train/loss': 0.47341978549957275, 'train/grad_norm_pre_clip': 44.742366790771484, 'timing/data': 0.0009768027812242508, 'timing/model': 1.3161635687574744, 'learning_rate/qwen_vl_interface': 1.0345369175396972e-06, 'learning_rate/action_model': 1.0345369175396972e-05, 'epoch': 2.18}) +[2026-06-18 05:40:43,659][starVLA.training.train_starvla][INFO] - Step 5949, Loss: {'train/loss': 0.5512211322784424, 'train/grad_norm_pre_clip': 33.2335205078125, 'timing/data': 0.0008468981832265854, 'timing/model': 1.2617564229294658, 'learning_rate/qwen_vl_interface': 1.0335406280259255e-06, 'learning_rate/action_model': 1.0335406280259254e-05, 'epoch': 2.18}) +[2026-06-18 05:40:51,273][starVLA.training.train_starvla][INFO] - Step 5950, Loss: {'train/loss': 0.0902676209807396, 'train/grad_norm_pre_clip': 16.67238998413086, 'timing/data': 0.0009659919887781143, 'timing/model': 4.431356416083872, 'learning_rate/qwen_vl_interface': 1.0325452125887177e-06, 'learning_rate/action_model': 1.0325452125887178e-05, 'epoch': 2.18}) +[2026-06-18 05:40:57,977][starVLA.training.train_starvla][INFO] - Step 5951, Loss: {'train/loss': 0.6967471837997437, 'train/grad_norm_pre_clip': 36.98517990112305, 'timing/data': 0.0012012133374810219, 'timing/model': 1.073712213896215, 'learning_rate/qwen_vl_interface': 1.031550671434426e-06, 'learning_rate/action_model': 1.031550671434426e-05, 'epoch': 2.18}) +[2026-06-18 05:41:02,977][starVLA.training.train_starvla][INFO] - Step 5952, Loss: {'train/loss': 0.2759016156196594, 'train/grad_norm_pre_clip': 21.42926597595215, 'timing/data': 0.0011385781690478325, 'timing/model': 1.2785401623696089, 'learning_rate/qwen_vl_interface': 1.0305570047692197e-06, 'learning_rate/action_model': 1.0305570047692195e-05, 'epoch': 2.18}) +[2026-06-18 05:41:07,561][starVLA.training.train_starvla][INFO] - Step 5953, Loss: {'train/loss': 0.3632866442203522, 'train/grad_norm_pre_clip': 19.730350494384766, 'timing/data': 0.001064559444785118, 'timing/model': 1.1170627642422915, 'learning_rate/qwen_vl_interface': 1.0295642127990863e-06, 'learning_rate/action_model': 1.0295642127990863e-05, 'epoch': 2.18}) +[2026-06-18 05:41:12,725][starVLA.training.train_starvla][INFO] - Step 5954, Loss: {'train/loss': 0.22690029442310333, 'train/grad_norm_pre_clip': 20.61081886291504, 'timing/data': 0.0009700162336230278, 'timing/model': 1.1927969995886087, 'learning_rate/qwen_vl_interface': 1.0285722957298326e-06, 'learning_rate/action_model': 1.0285722957298326e-05, 'epoch': 2.18}) +[2026-06-18 05:41:18,236][starVLA.training.train_starvla][INFO] - Step 5955, Loss: {'train/loss': 0.25548285245895386, 'train/grad_norm_pre_clip': 35.49311828613281, 'timing/data': 0.0015319185331463814, 'timing/model': 1.4086182937026024, 'learning_rate/qwen_vl_interface': 1.027581253767085e-06, 'learning_rate/action_model': 1.027581253767085e-05, 'epoch': 2.18}) +[2026-06-18 05:41:22,295][starVLA.training.train_starvla][INFO] - Step 5956, Loss: {'train/loss': 0.3355542719364166, 'train/grad_norm_pre_clip': 38.356201171875, 'timing/data': 0.0008671926334500313, 'timing/model': 1.0424298634752631, 'learning_rate/qwen_vl_interface': 1.0265910871162858e-06, 'learning_rate/action_model': 1.0265910871162858e-05, 'epoch': 2.18}) +[2026-06-18 05:41:30,433][starVLA.training.train_starvla][INFO] - Step 5957, Loss: {'train/loss': 0.4484197795391083, 'train/grad_norm_pre_clip': 28.34568214416504, 'timing/data': 0.001238485798239708, 'timing/model': 1.1670054635033011, 'learning_rate/qwen_vl_interface': 1.0256017959827008e-06, 'learning_rate/action_model': 1.0256017959827008e-05, 'epoch': 2.18}) +[2026-06-18 05:41:35,394][starVLA.training.train_starvla][INFO] - Step 5958, Loss: {'train/loss': 0.15597347915172577, 'train/grad_norm_pre_clip': 18.807947158813477, 'timing/data': 0.0010771816596388817, 'timing/model': 1.1807297552004457, 'learning_rate/qwen_vl_interface': 1.0246133805714084e-06, 'learning_rate/action_model': 1.0246133805714084e-05, 'epoch': 2.18}) +[2026-06-18 05:41:41,081][starVLA.training.train_starvla][INFO] - Step 5959, Loss: {'train/loss': 0.1983780860900879, 'train/grad_norm_pre_clip': 12.692639350891113, 'timing/data': 0.006886279210448265, 'timing/model': 1.570438771508634, 'learning_rate/qwen_vl_interface': 1.0236258410873076e-06, 'learning_rate/action_model': 1.0236258410873076e-05, 'epoch': 2.18}) +[2026-06-18 05:41:47,058][starVLA.training.train_starvla][INFO] - Step 5960, Loss: {'train/loss': 0.21999067068099976, 'train/grad_norm_pre_clip': 17.894027709960938, 'timing/data': 0.0008925106376409531, 'timing/model': 1.6400525718927383, 'learning_rate/qwen_vl_interface': 1.0226391777351188e-06, 'learning_rate/action_model': 1.0226391777351189e-05, 'epoch': 2.18}) +[2026-06-18 05:41:51,844][starVLA.training.train_starvla][INFO] - Step 5961, Loss: {'train/loss': 0.43767932057380676, 'train/grad_norm_pre_clip': 25.6414852142334, 'timing/data': 0.001304873265326023, 'timing/model': 1.2620195336639881, 'learning_rate/qwen_vl_interface': 1.0216533907193777e-06, 'learning_rate/action_model': 1.0216533907193776e-05, 'epoch': 2.18}) +[2026-06-18 05:41:56,942][starVLA.training.train_starvla][INFO] - Step 5962, Loss: {'train/loss': 0.2340751737356186, 'train/grad_norm_pre_clip': 25.13435173034668, 'timing/data': 0.0007396731525659561, 'timing/model': 1.0227377144619823, 'learning_rate/qwen_vl_interface': 1.020668480244438e-06, 'learning_rate/action_model': 1.020668480244438e-05, 'epoch': 2.18}) +[2026-06-18 05:42:01,749][starVLA.training.train_starvla][INFO] - Step 5963, Loss: {'train/loss': 0.19697099924087524, 'train/grad_norm_pre_clip': 33.83747100830078, 'timing/data': 0.0010471921414136887, 'timing/model': 1.2319527994841337, 'learning_rate/qwen_vl_interface': 1.0196844465144735e-06, 'learning_rate/action_model': 1.0196844465144734e-05, 'epoch': 2.18}) +[2026-06-18 05:42:06,227][starVLA.training.train_starvla][INFO] - Step 5964, Loss: {'train/loss': 0.018594227731227875, 'train/grad_norm_pre_clip': 2.791107177734375, 'timing/data': 0.0011141970753669739, 'timing/model': 1.0417561959475279, 'learning_rate/qwen_vl_interface': 1.0187012897334755e-06, 'learning_rate/action_model': 1.0187012897334755e-05, 'epoch': 2.18}) +[2026-06-18 05:42:11,322][starVLA.training.train_starvla][INFO] - Step 5965, Loss: {'train/loss': 0.29933640360832214, 'train/grad_norm_pre_clip': 32.96369552612305, 'timing/data': 0.0008698860183358192, 'timing/model': 1.6337745198979974, 'learning_rate/qwen_vl_interface': 1.017719010105253e-06, 'learning_rate/action_model': 1.017719010105253e-05, 'epoch': 2.18}) +[2026-06-18 05:42:15,738][starVLA.training.train_starvla][INFO] - Step 5966, Loss: {'train/loss': 0.09042654931545258, 'train/grad_norm_pre_clip': 25.58297348022461, 'timing/data': 0.0009147534146904945, 'timing/model': 1.2639289135113358, 'learning_rate/qwen_vl_interface': 1.0167376078334337e-06, 'learning_rate/action_model': 1.0167376078334337e-05, 'epoch': 2.18}) +[2026-06-18 05:42:20,604][starVLA.training.train_starvla][INFO] - Step 5967, Loss: {'train/loss': 0.11170130223035812, 'train/grad_norm_pre_clip': 15.792305946350098, 'timing/data': 0.0012578992173075676, 'timing/model': 1.4761300794780254, 'learning_rate/qwen_vl_interface': 1.0157570831214634e-06, 'learning_rate/action_model': 1.0157570831214634e-05, 'epoch': 2.18}) +[2026-06-18 05:42:25,280][starVLA.training.train_starvla][INFO] - Step 5968, Loss: {'train/loss': 0.3792497217655182, 'train/grad_norm_pre_clip': 25.421783447265625, 'timing/data': 0.0009365584701299667, 'timing/model': 1.0392867233604193, 'learning_rate/qwen_vl_interface': 1.0147774361726066e-06, 'learning_rate/action_model': 1.0147774361726064e-05, 'epoch': 2.19}) +[2026-06-18 05:42:30,907][starVLA.training.train_starvla][INFO] - Step 5969, Loss: {'train/loss': 0.4673178791999817, 'train/grad_norm_pre_clip': 32.16499328613281, 'timing/data': 0.0007675820961594582, 'timing/model': 1.1286505553871393, 'learning_rate/qwen_vl_interface': 1.013798667189944e-06, 'learning_rate/action_model': 1.013798667189944e-05, 'epoch': 2.19}) +[2026-06-18 05:42:37,525][starVLA.training.train_starvla][INFO] - Step 5970, Loss: {'train/loss': 0.4019565284252167, 'train/grad_norm_pre_clip': 24.603652954101562, 'timing/data': 0.0009589958935976028, 'timing/model': 2.872440475039184, 'learning_rate/qwen_vl_interface': 1.0128207763763755e-06, 'learning_rate/action_model': 1.0128207763763755e-05, 'epoch': 2.19}) +[2026-06-18 05:42:43,212][starVLA.training.train_starvla][INFO] - Step 5971, Loss: {'train/loss': 0.3310503363609314, 'train/grad_norm_pre_clip': 24.867521286010742, 'timing/data': 0.0009348858147859573, 'timing/model': 1.0693486770614982, 'learning_rate/qwen_vl_interface': 1.0118437639346204e-06, 'learning_rate/action_model': 1.0118437639346204e-05, 'epoch': 2.19}) +[2026-06-18 05:42:48,455][starVLA.training.train_starvla][INFO] - Step 5972, Loss: {'train/loss': 0.1986236870288849, 'train/grad_norm_pre_clip': 26.15273094177246, 'timing/data': 0.0011475514620542526, 'timing/model': 1.3850568188354373, 'learning_rate/qwen_vl_interface': 1.0108676300672144e-06, 'learning_rate/action_model': 1.0108676300672144e-05, 'epoch': 2.19}) +[2026-06-18 05:42:59,008][starVLA.training.train_starvla][INFO] - Step 5973, Loss: {'train/loss': 0.24635040760040283, 'train/grad_norm_pre_clip': 22.789508819580078, 'timing/data': 0.0011118361726403236, 'timing/model': 2.9276363914832473, 'learning_rate/qwen_vl_interface': 1.0098923749765086e-06, 'learning_rate/action_model': 1.0098923749765086e-05, 'epoch': 2.19}) +[2026-06-18 05:43:09,897][starVLA.training.train_starvla][INFO] - Step 5974, Loss: {'train/loss': 0.15258583426475525, 'train/grad_norm_pre_clip': 17.052148818969727, 'timing/data': 0.0017480794340372086, 'timing/model': 1.228373495861888, 'learning_rate/qwen_vl_interface': 1.008917998864676e-06, 'learning_rate/action_model': 1.008917998864676e-05, 'epoch': 2.19}) +[2026-06-18 05:43:15,593][starVLA.training.train_starvla][INFO] - Step 5975, Loss: {'train/loss': 0.3157673478126526, 'train/grad_norm_pre_clip': 31.73162078857422, 'timing/data': 0.0011293943971395493, 'timing/model': 1.3336329208686948, 'learning_rate/qwen_vl_interface': 1.0079445019337063e-06, 'learning_rate/action_model': 1.0079445019337063e-05, 'epoch': 2.19}) +[2026-06-18 05:43:20,009][starVLA.training.train_starvla][INFO] - Step 5976, Loss: {'train/loss': 0.3875191807746887, 'train/grad_norm_pre_clip': 38.22721481323242, 'timing/data': 0.0014078253880143166, 'timing/model': 1.0629625171422958, 'learning_rate/qwen_vl_interface': 1.0069718843854058e-06, 'learning_rate/action_model': 1.0069718843854058e-05, 'epoch': 2.19}) +[2026-06-18 05:43:25,030][starVLA.training.train_starvla][INFO] - Step 5977, Loss: {'train/loss': 0.09243589639663696, 'train/grad_norm_pre_clip': 18.43206024169922, 'timing/data': 0.001028396189212799, 'timing/model': 1.4312847973778844, 'learning_rate/qwen_vl_interface': 1.0060001464214007e-06, 'learning_rate/action_model': 1.0060001464214006e-05, 'epoch': 2.19}) +[2026-06-18 05:43:29,822][starVLA.training.train_starvla][INFO] - Step 5978, Loss: {'train/loss': 0.4457552134990692, 'train/grad_norm_pre_clip': 34.24673080444336, 'timing/data': 0.0009368415921926498, 'timing/model': 1.063900456763804, 'learning_rate/qwen_vl_interface': 1.00502928824313e-06, 'learning_rate/action_model': 1.00502928824313e-05, 'epoch': 2.19}) +[2026-06-18 05:43:34,713][starVLA.training.train_starvla][INFO] - Step 5979, Loss: {'train/loss': 0.05847739800810814, 'train/grad_norm_pre_clip': 13.802457809448242, 'timing/data': 0.0011056652292609215, 'timing/model': 1.154041651636362, 'learning_rate/qwen_vl_interface': 1.0040593100518567e-06, 'learning_rate/action_model': 1.0040593100518566e-05, 'epoch': 2.19}) +[2026-06-18 05:43:40,094][starVLA.training.train_starvla][INFO] - Step 5980, Loss: {'train/loss': 0.3196559548377991, 'train/grad_norm_pre_clip': 29.2812557220459, 'timing/data': 0.0014594579115509987, 'timing/model': 1.8111211936920881, 'learning_rate/qwen_vl_interface': 1.003090212048657e-06, 'learning_rate/action_model': 1.003090212048657e-05, 'epoch': 2.19}) +[2026-06-18 05:43:46,213][starVLA.training.train_starvla][INFO] - Step 5981, Loss: {'train/loss': 0.3149455189704895, 'train/grad_norm_pre_clip': 26.76806640625, 'timing/data': 0.0007501933723688126, 'timing/model': 2.562879982404411, 'learning_rate/qwen_vl_interface': 1.0021219944344263e-06, 'learning_rate/action_model': 1.0021219944344262e-05, 'epoch': 2.19}) +[2026-06-18 05:43:57,611][starVLA.training.train_starvla][INFO] - Step 5982, Loss: {'train/loss': 0.38577625155448914, 'train/grad_norm_pre_clip': 27.64312171936035, 'timing/data': 0.001134198158979416, 'timing/model': 4.713911564089358, 'learning_rate/qwen_vl_interface': 1.0011546574098773e-06, 'learning_rate/action_model': 1.0011546574098772e-05, 'epoch': 2.19}) +[2026-06-18 05:44:07,740][starVLA.training.train_starvla][INFO] - Step 5983, Loss: {'train/loss': 0.047763314098119736, 'train/grad_norm_pre_clip': 8.246088027954102, 'timing/data': 0.0013445699587464333, 'timing/model': 1.2898953398689628, 'learning_rate/qwen_vl_interface': 1.0001882011755396e-06, 'learning_rate/action_model': 1.0001882011755395e-05, 'epoch': 2.19}) +[2026-06-18 05:44:13,059][starVLA.training.train_starvla][INFO] - Step 5984, Loss: {'train/loss': 0.4309116005897522, 'train/grad_norm_pre_clip': 47.90707778930664, 'timing/data': 0.0013459250330924988, 'timing/model': 1.108042510226369, 'learning_rate/qwen_vl_interface': 9.992226259317605e-07, 'learning_rate/action_model': 9.992226259317605e-06, 'epoch': 2.19}) +[2026-06-18 05:44:17,200][starVLA.training.train_starvla][INFO] - Step 5985, Loss: {'train/loss': 0.11138197779655457, 'train/grad_norm_pre_clip': 13.948659896850586, 'timing/data': 0.0009380802512168884, 'timing/model': 1.076048239134252, 'learning_rate/qwen_vl_interface': 9.982579318787044e-07, 'learning_rate/action_model': 9.982579318787044e-06, 'epoch': 2.19}) +[2026-06-18 05:44:21,772][starVLA.training.train_starvla][INFO] - Step 5986, Loss: {'train/loss': 0.38999995589256287, 'train/grad_norm_pre_clip': 37.793392181396484, 'timing/data': 0.0009035114198923111, 'timing/model': 1.0462286164984107, 'learning_rate/qwen_vl_interface': 9.97294119216356e-07, 'learning_rate/action_model': 9.97294119216356e-06, 'epoch': 2.19}) +[2026-06-18 05:44:28,237][starVLA.training.train_starvla][INFO] - Step 5987, Loss: {'train/loss': 0.3102203905582428, 'train/grad_norm_pre_clip': 24.528425216674805, 'timing/data': 0.0012968070805072784, 'timing/model': 1.1186999157071114, 'learning_rate/qwen_vl_interface': 9.96331188144512e-07, 'learning_rate/action_model': 9.96331188144512e-06, 'epoch': 2.19}) +[2026-06-18 05:44:33,472][starVLA.training.train_starvla][INFO] - Step 5988, Loss: {'train/loss': 0.2656678855419159, 'train/grad_norm_pre_clip': 18.495731353759766, 'timing/data': 0.0009931698441505432, 'timing/model': 1.3256121575832367, 'learning_rate/qwen_vl_interface': 9.953691388627887e-07, 'learning_rate/action_model': 9.953691388627886e-06, 'epoch': 2.19}) +[2026-06-18 05:44:40,037][starVLA.training.train_starvla][INFO] - Step 5989, Loss: {'train/loss': 0.42972221970558167, 'train/grad_norm_pre_clip': 30.110898971557617, 'timing/data': 0.0015238923951983452, 'timing/model': 1.3432887513190508, 'learning_rate/qwen_vl_interface': 9.94407971570622e-07, 'learning_rate/action_model': 9.94407971570622e-06, 'epoch': 2.19}) +[2026-06-18 05:44:44,491][starVLA.training.train_starvla][INFO] - Step 5990, Loss: {'train/loss': 0.18969908356666565, 'train/grad_norm_pre_clip': 20.550926208496094, 'timing/data': 0.0008627623319625854, 'timing/model': 1.3423265079036355, 'learning_rate/qwen_vl_interface': 9.934476864672618e-07, 'learning_rate/action_model': 9.934476864672617e-06, 'epoch': 2.19}) +[2026-06-18 05:44:52,422][starVLA.training.train_starvla][INFO] - Step 5991, Loss: {'train/loss': 0.4797159731388092, 'train/grad_norm_pre_clip': 37.988155364990234, 'timing/data': 0.0011131148785352707, 'timing/model': 4.642525655217469, 'learning_rate/qwen_vl_interface': 9.924882837517781e-07, 'learning_rate/action_model': 9.924882837517782e-06, 'epoch': 2.19}) +[2026-06-18 05:44:57,607][starVLA.training.train_starvla][INFO] - Step 5992, Loss: {'train/loss': 0.42284250259399414, 'train/grad_norm_pre_clip': 32.76914978027344, 'timing/data': 0.0009608082473278046, 'timing/model': 1.0607049288228154, 'learning_rate/qwen_vl_interface': 9.91529763623052e-07, 'learning_rate/action_model': 9.91529763623052e-06, 'epoch': 2.19}) +[2026-06-18 05:45:03,140][starVLA.training.train_starvla][INFO] - Step 5993, Loss: {'train/loss': 0.30106425285339355, 'train/grad_norm_pre_clip': 22.08647918701172, 'timing/data': 0.0012922575697302818, 'timing/model': 1.1089917048811913, 'learning_rate/qwen_vl_interface': 9.905721262797894e-07, 'learning_rate/action_model': 9.905721262797892e-06, 'epoch': 2.19}) +[2026-06-18 05:45:07,678][starVLA.training.train_starvla][INFO] - Step 5994, Loss: {'train/loss': 0.4560416340827942, 'train/grad_norm_pre_clip': 32.99982833862305, 'timing/data': 0.0009843409061431885, 'timing/model': 1.166491680778563, 'learning_rate/qwen_vl_interface': 9.896153719205082e-07, 'learning_rate/action_model': 9.89615371920508e-06, 'epoch': 2.19}) +[2026-06-18 05:45:12,775][starVLA.training.train_starvla][INFO] - Step 5995, Loss: {'train/loss': 0.18007193505764008, 'train/grad_norm_pre_clip': 16.155120849609375, 'timing/data': 0.0007595745846629143, 'timing/model': 1.7009769706055522, 'learning_rate/qwen_vl_interface': 9.886595007435448e-07, 'learning_rate/action_model': 9.886595007435447e-06, 'epoch': 2.2}) +[2026-06-18 05:45:17,560][starVLA.training.train_starvla][INFO] - Step 5996, Loss: {'train/loss': 0.22949841618537903, 'train/grad_norm_pre_clip': 26.86989402770996, 'timing/data': 0.0011856192722916603, 'timing/model': 1.1005152594298124, 'learning_rate/qwen_vl_interface': 9.877045129470524e-07, 'learning_rate/action_model': 9.877045129470524e-06, 'epoch': 2.2}) +[2026-06-18 05:45:28,047][starVLA.training.train_starvla][INFO] - Step 5997, Loss: {'train/loss': 0.4919951260089874, 'train/grad_norm_pre_clip': 24.24396324157715, 'timing/data': 0.0010862303897738457, 'timing/model': 2.702231948263943, 'learning_rate/qwen_vl_interface': 9.86750408729001e-07, 'learning_rate/action_model': 9.867504087290009e-06, 'epoch': 2.2}) +[2026-06-18 05:45:32,870][starVLA.training.train_starvla][INFO] - Step 5998, Loss: {'train/loss': 0.09365316480398178, 'train/grad_norm_pre_clip': 14.805174827575684, 'timing/data': 0.0007750131189823151, 'timing/model': 1.068999052979052, 'learning_rate/qwen_vl_interface': 9.857971882871773e-07, 'learning_rate/action_model': 9.857971882871772e-06, 'epoch': 2.2}) +[2026-06-18 05:45:46,476][starVLA.training.train_starvla][INFO] - Step 5999, Loss: {'train/loss': 0.27264609932899475, 'train/grad_norm_pre_clip': 23.572826385498047, 'timing/data': 0.0009197406470775604, 'timing/model': 1.219792403280735, 'learning_rate/qwen_vl_interface': 9.848448518191838e-07, 'learning_rate/action_model': 9.848448518191836e-06, 'epoch': 2.2}) +[2026-06-18 05:51:32,566][starVLA.training.train_starvla][INFO] - Step 6000, Loss: {'train/loss': 0.07666201144456863, 'train/grad_norm_pre_clip': 13.799388885498047, 'eval/loss': 0.5361967086791992, 'eval/latency_2/loss': 0.5361967086791992, 'eval/demon_fire_precision': 0.9503722084367245, 'eval/demon_fire_recall': 0.9615062761506277, 'eval/demon_fire_f1': 0.9559068219633944, 'eval/demon_fire_rate_model': 0.672787979966611, 'eval/demon_fire_rate_teacher': 0.6649972175848636, 'eval/demon_left_precision': 0.907427341227126, 'eval/demon_left_recall': 0.8873684210526316, 'eval/demon_left_f1': 0.8972857903139967, 'eval/demon_left_rate_model': 0.2584863661658319, 'eval/demon_left_rate_teacher': 0.2643294379521425, 'eval/demon_right_precision': 0.9176201372997712, 'eval/demon_right_recall': 0.9062146892655367, 'eval/demon_right_f1': 0.9118817509948834, 'eval/demon_right_rate_model': 0.24318308291597107, 'eval/demon_right_rate_teacher': 0.24624373956594323, 'eval/demon_move_f1': 0.9045837706544401, 'eval/demon_cc_f1': 0.9399086421273581, 'eval/latency_2/demon_fire_precision': 0.9503722084367245, 'eval/latency_2/demon_fire_recall': 0.9615062761506277, 'eval/latency_2/demon_fire_f1': 0.9559068219633944, 'eval/latency_2/demon_fire_rate_model': 0.672787979966611, 'eval/latency_2/demon_fire_rate_teacher': 0.6649972175848636, 'eval/latency_2/demon_left_precision': 0.907427341227126, 'eval/latency_2/demon_left_recall': 0.8873684210526316, 'eval/latency_2/demon_left_f1': 0.8972857903139967, 'eval/latency_2/demon_left_rate_model': 0.2584863661658319, 'eval/latency_2/demon_left_rate_teacher': 0.2643294379521425, 'eval/latency_2/demon_right_precision': 0.9176201372997712, 'eval/latency_2/demon_right_recall': 0.9062146892655367, 'eval/latency_2/demon_right_f1': 0.9118817509948834, 'eval/latency_2/demon_right_rate_model': 0.24318308291597107, 'eval/latency_2/demon_right_rate_teacher': 0.24624373956594323, 'eval/latency_2/demon_move_f1': 0.9045837706544401, 'eval/latency_2/demon_cc_f1': 0.9399086421273581, 'timing/data': 0.002262430265545845, 'timing/model': 1.5091661512851715, 'learning_rate/qwen_vl_interface': 9.838933995224439e-07, 'learning_rate/action_model': 9.83893399522444e-06, 'epoch': 2.2}) +[2026-06-18 05:51:32,568][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_6000_state +[2026-06-18 05:51:32,575][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 05:51:41,047][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_6000_state/model.safetensors +[2026-06-18 05:51:54,409][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_6000_state/optimizer.bin +[2026-06-18 05:51:54,410][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_6000_state/sampler.bin +[2026-06-18 05:51:54,410][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_6000_state/sampler_1.bin +[2026-06-18 05:51:54,413][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_6000_state/random_states_0.pkl +[2026-06-18 05:51:54,417][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 6000 +[2026-06-18 05:51:54,422][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 05:51:54,444][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 05:51:58,710][starVLA.training.train_starvla][INFO] - Step 6001, Loss: {'train/loss': 0.14170634746551514, 'train/grad_norm_pre_clip': 14.634276390075684, 'timing/data': 0.0010447073727846146, 'timing/model': 1.3157483087852597, 'learning_rate/qwen_vl_interface': 9.829428315941921e-07, 'learning_rate/action_model': 9.829428315941922e-06, 'epoch': 2.2}) +[2026-06-18 05:52:02,692][starVLA.training.train_starvla][INFO] - Step 6002, Loss: {'train/loss': 0.16119742393493652, 'train/grad_norm_pre_clip': 25.450759887695312, 'timing/data': 0.0010164054110646248, 'timing/model': 1.034451873973012, 'learning_rate/qwen_vl_interface': 9.819931482314823e-07, 'learning_rate/action_model': 9.819931482314823e-06, 'epoch': 2.2}) +[2026-06-18 05:52:06,940][starVLA.training.train_starvla][INFO] - Step 6003, Loss: {'train/loss': 0.32138994336128235, 'train/grad_norm_pre_clip': 21.605154037475586, 'timing/data': 0.0010989932343363762, 'timing/model': 1.0269699078053236, 'learning_rate/qwen_vl_interface': 9.81044349631185e-07, 'learning_rate/action_model': 9.81044349631185e-06, 'epoch': 2.2}) +[2026-06-18 05:52:11,217][starVLA.training.train_starvla][INFO] - Step 6004, Loss: {'train/loss': 0.14399510622024536, 'train/grad_norm_pre_clip': 18.44942855834961, 'timing/data': 0.0010302038863301277, 'timing/model': 1.0507146045565605, 'learning_rate/qwen_vl_interface': 9.800964359899886e-07, 'learning_rate/action_model': 9.800964359899886e-06, 'epoch': 2.2}) +[2026-06-18 05:52:15,442][starVLA.training.train_starvla][INFO] - Step 6005, Loss: {'train/loss': 0.15009352564811707, 'train/grad_norm_pre_clip': 12.622529983520508, 'timing/data': 0.0010627182200551033, 'timing/model': 1.0355136366561055, 'learning_rate/qwen_vl_interface': 9.791494075043963e-07, 'learning_rate/action_model': 9.791494075043963e-06, 'epoch': 2.2}) +[2026-06-18 05:52:21,007][starVLA.training.train_starvla][INFO] - Step 6006, Loss: {'train/loss': 0.1571250706911087, 'train/grad_norm_pre_clip': 14.946967124938965, 'timing/data': 0.0009185709059238434, 'timing/model': 1.058048433624208, 'learning_rate/qwen_vl_interface': 9.78203264370727e-07, 'learning_rate/action_model': 9.78203264370727e-06, 'epoch': 2.2}) +[2026-06-18 05:52:32,662][starVLA.training.train_starvla][INFO] - Step 6007, Loss: {'train/loss': 0.4093996286392212, 'train/grad_norm_pre_clip': 67.35582733154297, 'timing/data': 0.0008131992071866989, 'timing/model': 1.6528553012758493, 'learning_rate/qwen_vl_interface': 9.772580067851162e-07, 'learning_rate/action_model': 9.772580067851161e-06, 'epoch': 2.2}) +[2026-06-18 05:52:38,077][starVLA.training.train_starvla][INFO] - Step 6008, Loss: {'train/loss': 0.39406830072402954, 'train/grad_norm_pre_clip': 36.530635833740234, 'timing/data': 0.0011282814666628838, 'timing/model': 1.144249969162047, 'learning_rate/qwen_vl_interface': 9.763136349435195e-07, 'learning_rate/action_model': 9.763136349435194e-06, 'epoch': 2.2}) +[2026-06-18 05:52:43,514][starVLA.training.train_starvla][INFO] - Step 6009, Loss: {'train/loss': 0.3100033402442932, 'train/grad_norm_pre_clip': 21.604631423950195, 'timing/data': 0.0009900564327836037, 'timing/model': 1.1070935567840934, 'learning_rate/qwen_vl_interface': 9.753701490417048e-07, 'learning_rate/action_model': 9.753701490417047e-06, 'epoch': 2.2}) +[2026-06-18 05:52:53,230][starVLA.training.train_starvla][INFO] - Step 6010, Loss: {'train/loss': 0.21820250153541565, 'train/grad_norm_pre_clip': 17.885250091552734, 'timing/data': 0.000805901363492012, 'timing/model': 1.539208997040987, 'learning_rate/qwen_vl_interface': 9.744275492752584e-07, 'learning_rate/action_model': 9.744275492752583e-06, 'epoch': 2.2}) +[2026-06-18 05:53:06,602][starVLA.training.train_starvla][INFO] - Step 6011, Loss: {'train/loss': 0.12743011116981506, 'train/grad_norm_pre_clip': 12.806268692016602, 'timing/data': 0.0009754635393619537, 'timing/model': 5.574749529361725, 'learning_rate/qwen_vl_interface': 9.734858358395808e-07, 'learning_rate/action_model': 9.734858358395807e-06, 'epoch': 2.2}) +[2026-06-18 05:53:17,754][starVLA.training.train_starvla][INFO] - Step 6012, Loss: {'train/loss': 0.18781673908233643, 'train/grad_norm_pre_clip': 18.727807998657227, 'timing/data': 0.0010379143059253693, 'timing/model': 1.458257731050253, 'learning_rate/qwen_vl_interface': 9.72545008929892e-07, 'learning_rate/action_model': 9.725450089298921e-06, 'epoch': 2.2}) +[2026-06-18 05:53:22,964][starVLA.training.train_starvla][INFO] - Step 6013, Loss: {'train/loss': 0.18978942930698395, 'train/grad_norm_pre_clip': 17.058792114257812, 'timing/data': 0.00129729975014925, 'timing/model': 1.3237052615731955, 'learning_rate/qwen_vl_interface': 9.71605068741225e-07, 'learning_rate/action_model': 9.716050687412251e-06, 'epoch': 2.2}) +[2026-06-18 05:53:28,007][starVLA.training.train_starvla][INFO] - Step 6014, Loss: {'train/loss': 0.14819270372390747, 'train/grad_norm_pre_clip': 13.54053783416748, 'timing/data': 0.0014172066003084183, 'timing/model': 1.3206520713865757, 'learning_rate/qwen_vl_interface': 9.706660154684312e-07, 'learning_rate/action_model': 9.706660154684312e-06, 'epoch': 2.2}) +[2026-06-18 05:53:32,827][starVLA.training.train_starvla][INFO] - Step 6015, Loss: {'train/loss': 0.11650826036930084, 'train/grad_norm_pre_clip': 15.129914283752441, 'timing/data': 0.0010632863268256187, 'timing/model': 1.0947288339957595, 'learning_rate/qwen_vl_interface': 9.697278493061773e-07, 'learning_rate/action_model': 9.697278493061773e-06, 'epoch': 2.2}) +[2026-06-18 05:53:44,720][starVLA.training.train_starvla][INFO] - Step 6016, Loss: {'train/loss': 0.3912592828273773, 'train/grad_norm_pre_clip': 25.101388931274414, 'timing/data': 0.0011592507362365723, 'timing/model': 1.7462859340012074, 'learning_rate/qwen_vl_interface': 9.687905704489461e-07, 'learning_rate/action_model': 9.687905704489462e-06, 'epoch': 2.2}) +[2026-06-18 05:53:53,201][starVLA.training.train_starvla][INFO] - Step 6017, Loss: {'train/loss': 0.4069927930831909, 'train/grad_norm_pre_clip': 26.37995147705078, 'timing/data': 0.0011861836537718773, 'timing/model': 1.2772975377738476, 'learning_rate/qwen_vl_interface': 9.678541790910361e-07, 'learning_rate/action_model': 9.678541790910361e-06, 'epoch': 2.2}) +[2026-06-18 05:54:00,480][starVLA.training.train_starvla][INFO] - Step 6018, Loss: {'train/loss': 0.07012699544429779, 'train/grad_norm_pre_clip': 13.116250038146973, 'timing/data': 0.0012619048357009888, 'timing/model': 3.294505652040243, 'learning_rate/qwen_vl_interface': 9.669186754265619e-07, 'learning_rate/action_model': 9.669186754265618e-06, 'epoch': 2.2}) +[2026-06-18 05:54:13,144][starVLA.training.train_starvla][INFO] - Step 6019, Loss: {'train/loss': 0.2231026440858841, 'train/grad_norm_pre_clip': 25.757144927978516, 'timing/data': 0.0009069200605154037, 'timing/model': 1.1900918493047357, 'learning_rate/qwen_vl_interface': 9.65984059649457e-07, 'learning_rate/action_model': 9.65984059649457e-06, 'epoch': 2.2}) +[2026-06-18 05:54:17,958][starVLA.training.train_starvla][INFO] - Step 6020, Loss: {'train/loss': 0.27910706400871277, 'train/grad_norm_pre_clip': 37.26551818847656, 'timing/data': 0.0010410230606794357, 'timing/model': 1.223596009425819, 'learning_rate/qwen_vl_interface': 9.650503319534652e-07, 'learning_rate/action_model': 9.650503319534651e-06, 'epoch': 2.2}) +[2026-06-18 05:54:27,011][starVLA.training.train_starvla][INFO] - Step 6021, Loss: {'train/loss': 0.030217383056879044, 'train/grad_norm_pre_clip': 7.328804969787598, 'timing/data': 0.0012026391923427582, 'timing/model': 1.6125910561531782, 'learning_rate/qwen_vl_interface': 9.641174925321494e-07, 'learning_rate/action_model': 9.641174925321494e-06, 'epoch': 2.2}) +[2026-06-18 05:54:32,799][starVLA.training.train_starvla][INFO] - Step 6022, Loss: {'train/loss': 0.05570605769753456, 'train/grad_norm_pre_clip': 10.523628234863281, 'timing/data': 0.0008720140904188156, 'timing/model': 1.1402360368520021, 'learning_rate/qwen_vl_interface': 9.631855415788904e-07, 'learning_rate/action_model': 9.631855415788904e-06, 'epoch': 2.21}) +[2026-06-18 05:54:38,376][starVLA.training.train_starvla][INFO] - Step 6023, Loss: {'train/loss': 0.40302637219429016, 'train/grad_norm_pre_clip': 23.312641143798828, 'timing/data': 0.0012144437059760094, 'timing/model': 1.2113261753693223, 'learning_rate/qwen_vl_interface': 9.622544792868812e-07, 'learning_rate/action_model': 9.622544792868812e-06, 'epoch': 2.21}) +[2026-06-18 05:54:49,712][starVLA.training.train_starvla][INFO] - Step 6024, Loss: {'train/loss': 0.2394746094942093, 'train/grad_norm_pre_clip': 16.703393936157227, 'timing/data': 0.004619717597961426, 'timing/model': 6.0820984514430165, 'learning_rate/qwen_vl_interface': 9.613243058491334e-07, 'learning_rate/action_model': 9.613243058491334e-06, 'epoch': 2.21}) +[2026-06-18 05:54:56,241][starVLA.training.train_starvla][INFO] - Step 6025, Loss: {'train/loss': 0.28723782300949097, 'train/grad_norm_pre_clip': 23.7769718170166, 'timing/data': 0.000760955736041069, 'timing/model': 1.0933905122801661, 'learning_rate/qwen_vl_interface': 9.603950214584695e-07, 'learning_rate/action_model': 9.603950214584696e-06, 'epoch': 2.21}) +[2026-06-18 05:55:03,353][starVLA.training.train_starvla][INFO] - Step 6026, Loss: {'train/loss': 0.2583079934120178, 'train/grad_norm_pre_clip': 21.060890197753906, 'timing/data': 0.00453584361821413, 'timing/model': 1.4466751795262098, 'learning_rate/qwen_vl_interface': 9.59466626307535e-07, 'learning_rate/action_model': 9.594666263075349e-06, 'epoch': 2.21}) +[2026-06-18 05:55:07,782][starVLA.training.train_starvla][INFO] - Step 6027, Loss: {'train/loss': 0.2320718616247177, 'train/grad_norm_pre_clip': 17.944570541381836, 'timing/data': 0.0009422926232218742, 'timing/model': 1.053292284719646, 'learning_rate/qwen_vl_interface': 9.585391205887851e-07, 'learning_rate/action_model': 9.585391205887851e-06, 'epoch': 2.21}) +[2026-06-18 05:55:13,173][starVLA.training.train_starvla][INFO] - Step 6028, Loss: {'train/loss': 0.6411396861076355, 'train/grad_norm_pre_clip': 39.889896392822266, 'timing/data': 0.0010164184495806694, 'timing/model': 1.058777323924005, 'learning_rate/qwen_vl_interface': 9.57612504494494e-07, 'learning_rate/action_model': 9.57612504494494e-06, 'epoch': 2.21}) +[2026-06-18 05:55:21,448][starVLA.training.train_starvla][INFO] - Step 6029, Loss: {'train/loss': 0.26551544666290283, 'train/grad_norm_pre_clip': 31.593276977539062, 'timing/data': 0.001312514767050743, 'timing/model': 1.1123841060325503, 'learning_rate/qwen_vl_interface': 9.566867782167494e-07, 'learning_rate/action_model': 9.566867782167494e-06, 'epoch': 2.21}) +[2026-06-18 05:55:31,313][starVLA.training.train_starvla][INFO] - Step 6030, Loss: {'train/loss': 0.21426698565483093, 'train/grad_norm_pre_clip': 26.649311065673828, 'timing/data': 0.0030659446492791176, 'timing/model': 1.8969640946015716, 'learning_rate/qwen_vl_interface': 9.55761941947455e-07, 'learning_rate/action_model': 9.557619419474551e-06, 'epoch': 2.21}) +[2026-06-18 05:55:36,428][starVLA.training.train_starvla][INFO] - Step 6031, Loss: {'train/loss': 0.26901084184646606, 'train/grad_norm_pre_clip': 31.630817413330078, 'timing/data': 0.0008974727243185043, 'timing/model': 1.3736696289852262, 'learning_rate/qwen_vl_interface': 9.548379958783308e-07, 'learning_rate/action_model': 9.548379958783308e-06, 'epoch': 2.21}) +[2026-06-18 05:55:42,530][starVLA.training.train_starvla][INFO] - Step 6032, Loss: {'train/loss': 0.3734247088432312, 'train/grad_norm_pre_clip': 46.7161979675293, 'timing/data': 0.0011294037103652954, 'timing/model': 1.1036159750074148, 'learning_rate/qwen_vl_interface': 9.539149402009114e-07, 'learning_rate/action_model': 9.539149402009115e-06, 'epoch': 2.21}) +[2026-06-18 05:55:59,490][starVLA.training.train_starvla][INFO] - Step 6033, Loss: {'train/loss': 0.3823527693748474, 'train/grad_norm_pre_clip': 34.688499450683594, 'timing/data': 0.0011271405965089798, 'timing/model': 2.57027699239552, 'learning_rate/qwen_vl_interface': 9.529927751065493e-07, 'learning_rate/action_model': 9.529927751065492e-06, 'epoch': 2.21}) +[2026-06-18 05:56:05,115][starVLA.training.train_starvla][INFO] - Step 6034, Loss: {'train/loss': 0.4932073652744293, 'train/grad_norm_pre_clip': 30.592954635620117, 'timing/data': 0.0011897776275873184, 'timing/model': 1.1947622234001756, 'learning_rate/qwen_vl_interface': 9.520715007864082e-07, 'learning_rate/action_model': 9.52071500786408e-06, 'epoch': 2.21}) +[2026-06-18 05:56:20,153][starVLA.training.train_starvla][INFO] - Step 6035, Loss: {'train/loss': 0.6058032512664795, 'train/grad_norm_pre_clip': 37.80854415893555, 'timing/data': 0.0012319842353463173, 'timing/model': 1.891687142662704, 'learning_rate/qwen_vl_interface': 9.511511174314689e-07, 'learning_rate/action_model': 9.511511174314688e-06, 'epoch': 2.21}) +[2026-06-18 05:56:25,211][starVLA.training.train_starvla][INFO] - Step 6036, Loss: {'train/loss': 0.190015509724617, 'train/grad_norm_pre_clip': 14.679423332214355, 'timing/data': 0.0012212768197059631, 'timing/model': 1.6587392706423998, 'learning_rate/qwen_vl_interface': 9.502316252325283e-07, 'learning_rate/action_model': 9.502316252325283e-06, 'epoch': 2.21}) +[2026-06-18 05:56:30,199][starVLA.training.train_starvla][INFO] - Step 6037, Loss: {'train/loss': 0.5796462297439575, 'train/grad_norm_pre_clip': 41.81230926513672, 'timing/data': 0.0013834433630108833, 'timing/model': 1.395561957731843, 'learning_rate/qwen_vl_interface': 9.493130243801993e-07, 'learning_rate/action_model': 9.493130243801992e-06, 'epoch': 2.21}) +[2026-06-18 05:56:35,174][starVLA.training.train_starvla][INFO] - Step 6038, Loss: {'train/loss': 0.09491904824972153, 'train/grad_norm_pre_clip': 11.991730690002441, 'timing/data': 0.0011837240308523178, 'timing/model': 1.196091577410698, 'learning_rate/qwen_vl_interface': 9.483953150649091e-07, 'learning_rate/action_model': 9.483953150649091e-06, 'epoch': 2.21}) +[2026-06-18 05:56:53,285][starVLA.training.train_starvla][INFO] - Step 6039, Loss: {'train/loss': 0.20411992073059082, 'train/grad_norm_pre_clip': 24.368295669555664, 'timing/data': 0.0009331302717328072, 'timing/model': 3.7109247660264373, 'learning_rate/qwen_vl_interface': 9.474784974768982e-07, 'learning_rate/action_model': 9.474784974768982e-06, 'epoch': 2.21}) +[2026-06-18 05:57:01,932][starVLA.training.train_starvla][INFO] - Step 6040, Loss: {'train/loss': 0.2747550308704376, 'train/grad_norm_pre_clip': 23.96147918701172, 'timing/data': 0.00102300476282835, 'timing/model': 1.3327745208516717, 'learning_rate/qwen_vl_interface': 9.465625718062236e-07, 'learning_rate/action_model': 9.465625718062235e-06, 'epoch': 2.21}) +[2026-06-18 05:57:06,944][starVLA.training.train_starvla][INFO] - Step 6041, Loss: {'train/loss': 0.2810315191745758, 'train/grad_norm_pre_clip': 45.72376251220703, 'timing/data': 0.001596192829310894, 'timing/model': 1.5181029615923762, 'learning_rate/qwen_vl_interface': 9.456475382427591e-07, 'learning_rate/action_model': 9.45647538242759e-06, 'epoch': 2.21}) +[2026-06-18 05:57:14,309][starVLA.training.train_starvla][INFO] - Step 6042, Loss: {'train/loss': 0.16875578463077545, 'train/grad_norm_pre_clip': 16.229015350341797, 'timing/data': 0.0014858469367027283, 'timing/model': 1.161172416061163, 'learning_rate/qwen_vl_interface': 9.447333969761915e-07, 'learning_rate/action_model': 9.447333969761915e-06, 'epoch': 2.21}) +[2026-06-18 05:57:20,701][starVLA.training.train_starvla][INFO] - Step 6043, Loss: {'train/loss': 0.16647139191627502, 'train/grad_norm_pre_clip': 17.932918548583984, 'timing/data': 0.014101275242865086, 'timing/model': 1.32555883936584, 'learning_rate/qwen_vl_interface': 9.438201481960235e-07, 'learning_rate/action_model': 9.438201481960234e-06, 'epoch': 2.21}) +[2026-06-18 05:57:25,749][starVLA.training.train_starvla][INFO] - Step 6044, Loss: {'train/loss': 0.37469199299812317, 'train/grad_norm_pre_clip': 38.08396530151367, 'timing/data': 0.02641421929001808, 'timing/model': 1.1002143863588572, 'learning_rate/qwen_vl_interface': 9.429077920915724e-07, 'learning_rate/action_model': 9.429077920915724e-06, 'epoch': 2.21}) +[2026-06-18 05:57:37,989][starVLA.training.train_starvla][INFO] - Step 6045, Loss: {'train/loss': 0.020985644310712814, 'train/grad_norm_pre_clip': 3.7072436809539795, 'timing/data': 0.02702120877802372, 'timing/model': 2.64960684068501, 'learning_rate/qwen_vl_interface': 9.419963288519711e-07, 'learning_rate/action_model': 9.41996328851971e-06, 'epoch': 2.21}) +[2026-06-18 05:57:43,326][starVLA.training.train_starvla][INFO] - Step 6046, Loss: {'train/loss': 0.4148086607456207, 'train/grad_norm_pre_clip': 29.13062858581543, 'timing/data': 0.023721757344901562, 'timing/model': 1.5835439581423998, 'learning_rate/qwen_vl_interface': 9.410857586661659e-07, 'learning_rate/action_model': 9.410857586661659e-06, 'epoch': 2.21}) +[2026-06-18 05:57:49,274][starVLA.training.train_starvla][INFO] - Step 6047, Loss: {'train/loss': 0.18367384374141693, 'train/grad_norm_pre_clip': 15.349838256835938, 'timing/data': 0.026422697119414806, 'timing/model': 1.0697678215801716, 'learning_rate/qwen_vl_interface': 9.401760817229197e-07, 'learning_rate/action_model': 9.401760817229196e-06, 'epoch': 2.21}) +[2026-06-18 05:57:54,623][starVLA.training.train_starvla][INFO] - Step 6048, Loss: {'train/loss': 0.31294021010398865, 'train/grad_norm_pre_clip': 28.58485984802246, 'timing/data': 0.024627188220620155, 'timing/model': 1.1140184178948402, 'learning_rate/qwen_vl_interface': 9.392672982108091e-07, 'learning_rate/action_model': 9.392672982108091e-06, 'epoch': 2.21}) +[2026-06-18 05:58:09,820][starVLA.training.train_starvla][INFO] - Step 6049, Loss: {'train/loss': 0.21681921184062958, 'train/grad_norm_pre_clip': 17.57366943359375, 'timing/data': 0.09947999380528927, 'timing/model': 1.2818792657926679, 'learning_rate/qwen_vl_interface': 9.383594083182265e-07, 'learning_rate/action_model': 9.383594083182265e-06, 'epoch': 2.22}) +[2026-06-18 05:58:16,526][starVLA.training.train_starvla][INFO] - Step 6050, Loss: {'train/loss': 0.4556865096092224, 'train/grad_norm_pre_clip': 27.842906951904297, 'timing/data': 0.02164173312485218, 'timing/model': 2.0435735462233424, 'learning_rate/qwen_vl_interface': 9.374524122333769e-07, 'learning_rate/action_model': 9.374524122333769e-06, 'epoch': 2.22}) +[2026-06-18 05:58:21,715][starVLA.training.train_starvla][INFO] - Step 6051, Loss: {'train/loss': 0.14436985552310944, 'train/grad_norm_pre_clip': 15.206758499145508, 'timing/data': 0.1033123666420579, 'timing/model': 1.997257080860436, 'learning_rate/qwen_vl_interface': 9.365463101442842e-07, 'learning_rate/action_model': 9.365463101442842e-06, 'epoch': 2.22}) +[2026-06-18 05:58:27,308][starVLA.training.train_starvla][INFO] - Step 6052, Loss: {'train/loss': 0.07245903462171555, 'train/grad_norm_pre_clip': 9.560270309448242, 'timing/data': 0.027290997095406055, 'timing/model': 1.6290603512898088, 'learning_rate/qwen_vl_interface': 9.356411022387823e-07, 'learning_rate/action_model': 9.356411022387824e-06, 'epoch': 2.22}) +[2026-06-18 05:58:32,612][starVLA.training.train_starvla][INFO] - Step 6053, Loss: {'train/loss': 0.3644360303878784, 'train/grad_norm_pre_clip': 40.06446075439453, 'timing/data': 0.042685664258897305, 'timing/model': 1.590275103226304, 'learning_rate/qwen_vl_interface': 9.347367887045236e-07, 'learning_rate/action_model': 9.347367887045236e-06, 'epoch': 2.22}) +[2026-06-18 05:58:39,260][starVLA.training.train_starvla][INFO] - Step 6054, Loss: {'train/loss': 0.5747895240783691, 'train/grad_norm_pre_clip': 35.3759880065918, 'timing/data': 0.04727782495319843, 'timing/model': 1.1511303158476949, 'learning_rate/qwen_vl_interface': 9.338333697289697e-07, 'learning_rate/action_model': 9.338333697289696e-06, 'epoch': 2.22}) +[2026-06-18 05:58:45,391][starVLA.training.train_starvla][INFO] - Step 6055, Loss: {'train/loss': 0.2873009741306305, 'train/grad_norm_pre_clip': 21.229019165039062, 'timing/data': 0.045961509458720684, 'timing/model': 2.80009950324893, 'learning_rate/qwen_vl_interface': 9.329308454994044e-07, 'learning_rate/action_model': 9.329308454994043e-06, 'epoch': 2.22}) +[2026-06-18 05:58:50,985][starVLA.training.train_starvla][INFO] - Step 6056, Loss: {'train/loss': 0.20289131999015808, 'train/grad_norm_pre_clip': 17.50971031188965, 'timing/data': 0.051186783239245415, 'timing/model': 1.5168076679110527, 'learning_rate/qwen_vl_interface': 9.320292162029192e-07, 'learning_rate/action_model': 9.320292162029191e-06, 'epoch': 2.22}) +[2026-06-18 05:58:55,534][starVLA.training.train_starvla][INFO] - Step 6057, Loss: {'train/loss': 0.3390311598777771, 'train/grad_norm_pre_clip': 22.84039306640625, 'timing/data': 0.04500693082809448, 'timing/model': 1.2027341285720468, 'learning_rate/qwen_vl_interface': 9.311284820264245e-07, 'learning_rate/action_model': 9.311284820264244e-06, 'epoch': 2.22}) +[2026-06-18 05:59:01,269][starVLA.training.train_starvla][INFO] - Step 6058, Loss: {'train/loss': 0.3783651888370514, 'train/grad_norm_pre_clip': 27.019535064697266, 'timing/data': 0.044822956435382366, 'timing/model': 1.192141824401915, 'learning_rate/qwen_vl_interface': 9.302286431566428e-07, 'learning_rate/action_model': 9.302286431566428e-06, 'epoch': 2.22}) +[2026-06-18 05:59:06,709][starVLA.training.train_starvla][INFO] - Step 6059, Loss: {'train/loss': 0.30160582065582275, 'train/grad_norm_pre_clip': 20.93998146057129, 'timing/data': 0.0487864026799798, 'timing/model': 1.186583075672388, 'learning_rate/qwen_vl_interface': 9.293296997801113e-07, 'learning_rate/action_model': 9.293296997801112e-06, 'epoch': 2.22}) +[2026-06-18 05:59:11,341][starVLA.training.train_starvla][INFO] - Step 6060, Loss: {'train/loss': 0.15564513206481934, 'train/grad_norm_pre_clip': 18.79141616821289, 'timing/data': 0.043882064521312714, 'timing/model': 1.253790003247559, 'learning_rate/qwen_vl_interface': 9.284316520831825e-07, 'learning_rate/action_model': 9.284316520831825e-06, 'epoch': 2.22}) +[2026-06-18 05:59:15,967][starVLA.training.train_starvla][INFO] - Step 6061, Loss: {'train/loss': 0.15661649405956268, 'train/grad_norm_pre_clip': 16.097999572753906, 'timing/data': 0.046667784452438354, 'timing/model': 1.4111182456836104, 'learning_rate/qwen_vl_interface': 9.275345002520225e-07, 'learning_rate/action_model': 9.275345002520225e-06, 'epoch': 2.22}) +[2026-06-18 05:59:21,121][starVLA.training.train_starvla][INFO] - Step 6062, Loss: {'train/loss': 0.1309397667646408, 'train/grad_norm_pre_clip': 22.96828269958496, 'timing/data': 0.046394494362175465, 'timing/model': 1.3381667211651802, 'learning_rate/qwen_vl_interface': 9.266382444726117e-07, 'learning_rate/action_model': 9.266382444726117e-06, 'epoch': 2.22}) +[2026-06-18 05:59:36,677][starVLA.training.train_starvla][INFO] - Step 6063, Loss: {'train/loss': 0.09504444152116776, 'train/grad_norm_pre_clip': 15.056097984313965, 'timing/data': 0.048401436768472195, 'timing/model': 5.888632673770189, 'learning_rate/qwen_vl_interface': 9.257428849307446e-07, 'learning_rate/action_model': 9.257428849307445e-06, 'epoch': 2.22}) +[2026-06-18 05:59:46,003][starVLA.training.train_starvla][INFO] - Step 6064, Loss: {'train/loss': 0.22862550616264343, 'train/grad_norm_pre_clip': 22.651151657104492, 'timing/data': 0.06372969597578049, 'timing/model': 1.0638112230226398, 'learning_rate/qwen_vl_interface': 9.248484218120307e-07, 'learning_rate/action_model': 9.248484218120307e-06, 'epoch': 2.22}) +[2026-06-18 05:59:50,889][starVLA.training.train_starvla][INFO] - Step 6065, Loss: {'train/loss': 0.35226112604141235, 'train/grad_norm_pre_clip': 24.424596786499023, 'timing/data': 0.06029456201940775, 'timing/model': 1.2269801190122962, 'learning_rate/qwen_vl_interface': 9.239548553018925e-07, 'learning_rate/action_model': 9.239548553018924e-06, 'epoch': 2.22}) +[2026-06-18 05:59:56,176][starVLA.training.train_starvla][INFO] - Step 6066, Loss: {'train/loss': 0.1084592267870903, 'train/grad_norm_pre_clip': 19.511629104614258, 'timing/data': 0.03881950955837965, 'timing/model': 1.8854585494846106, 'learning_rate/qwen_vl_interface': 9.230621855855687e-07, 'learning_rate/action_model': 9.230621855855687e-06, 'epoch': 2.22}) +[2026-06-18 06:00:00,747][starVLA.training.train_starvla][INFO] - Step 6067, Loss: {'train/loss': 0.38438379764556885, 'train/grad_norm_pre_clip': 25.636945724487305, 'timing/data': 0.04306078236550093, 'timing/model': 1.1358882077038288, 'learning_rate/qwen_vl_interface': 9.221704128481107e-07, 'learning_rate/action_model': 9.221704128481106e-06, 'epoch': 2.22}) +[2026-06-18 06:00:12,608][starVLA.training.train_starvla][INFO] - Step 6068, Loss: {'train/loss': 0.12412001937627792, 'train/grad_norm_pre_clip': 18.36030387878418, 'timing/data': 0.05122469086199999, 'timing/model': 1.3298874283209443, 'learning_rate/qwen_vl_interface': 9.212795372743815e-07, 'learning_rate/action_model': 9.212795372743813e-06, 'epoch': 2.22}) +[2026-06-18 06:00:19,532][starVLA.training.train_starvla][INFO] - Step 6069, Loss: {'train/loss': 0.23402924835681915, 'train/grad_norm_pre_clip': 29.32935905456543, 'timing/data': 0.053853060118854046, 'timing/model': 3.151122828014195, 'learning_rate/qwen_vl_interface': 9.203895590490617e-07, 'learning_rate/action_model': 9.203895590490616e-06, 'epoch': 2.22}) +[2026-06-18 06:00:36,810][starVLA.training.train_starvla][INFO] - Step 6070, Loss: {'train/loss': 0.5597277879714966, 'train/grad_norm_pre_clip': 24.609560012817383, 'timing/data': 0.04574587941169739, 'timing/model': 1.1900365110486746, 'learning_rate/qwen_vl_interface': 9.195004783566464e-07, 'learning_rate/action_model': 9.195004783566463e-06, 'epoch': 2.22}) +[2026-06-18 06:00:42,764][starVLA.training.train_starvla][INFO] - Step 6071, Loss: {'train/loss': 0.47871747612953186, 'train/grad_norm_pre_clip': 27.592411041259766, 'timing/data': 0.049237536266446114, 'timing/model': 1.4525363249704242, 'learning_rate/qwen_vl_interface': 9.186122953814414e-07, 'learning_rate/action_model': 9.186122953814413e-06, 'epoch': 2.22}) +[2026-06-18 06:00:47,775][starVLA.training.train_starvla][INFO] - Step 6072, Loss: {'train/loss': 0.6854494214057922, 'train/grad_norm_pre_clip': 42.04790496826172, 'timing/data': 0.09929895028471947, 'timing/model': 1.5209421310573816, 'learning_rate/qwen_vl_interface': 9.177250103075686e-07, 'learning_rate/action_model': 9.177250103075686e-06, 'epoch': 2.22}) +[2026-06-18 06:00:54,472][starVLA.training.train_starvla][INFO] - Step 6073, Loss: {'train/loss': 0.20267868041992188, 'train/grad_norm_pre_clip': 21.08642578125, 'timing/data': 0.04482006933540106, 'timing/model': 1.4976234529167414, 'learning_rate/qwen_vl_interface': 9.168386233189627e-07, 'learning_rate/action_model': 9.168386233189627e-06, 'epoch': 2.22}) +[2026-06-18 06:00:59,604][starVLA.training.train_starvla][INFO] - Step 6074, Loss: {'train/loss': 0.36464786529541016, 'train/grad_norm_pre_clip': 24.61043357849121, 'timing/data': 0.04571147542446852, 'timing/model': 1.2394254449754953, 'learning_rate/qwen_vl_interface': 9.159531345993726e-07, 'learning_rate/action_model': 9.159531345993725e-06, 'epoch': 2.22}) +[2026-06-18 06:01:08,847][starVLA.training.train_starvla][INFO] - Step 6075, Loss: {'train/loss': 0.09399602562189102, 'train/grad_norm_pre_clip': 11.472158432006836, 'timing/data': 0.04201794881373644, 'timing/model': 1.0915065538138151, 'learning_rate/qwen_vl_interface': 9.150685443323619e-07, 'learning_rate/action_model': 9.150685443323618e-06, 'epoch': 2.22}) +[2026-06-18 06:01:13,812][starVLA.training.train_starvla][INFO] - Step 6076, Loss: {'train/loss': 0.10686780512332916, 'train/grad_norm_pre_clip': 11.873505592346191, 'timing/data': 0.09431507717818022, 'timing/model': 1.6576670790091157, 'learning_rate/qwen_vl_interface': 9.141848527013059e-07, 'learning_rate/action_model': 9.141848527013059e-06, 'epoch': 2.22}) +[2026-06-18 06:01:18,308][starVLA.training.train_starvla][INFO] - Step 6077, Loss: {'train/loss': 0.0966852679848671, 'train/grad_norm_pre_clip': 13.295965194702148, 'timing/data': 0.048811059445142746, 'timing/model': 1.1839839490130544, 'learning_rate/qwen_vl_interface': 9.13302059889396e-07, 'learning_rate/action_model': 9.13302059889396e-06, 'epoch': 2.23}) +[2026-06-18 06:01:26,752][starVLA.training.train_starvla][INFO] - Step 6078, Loss: {'train/loss': 0.07170595228672028, 'train/grad_norm_pre_clip': 8.87631607055664, 'timing/data': 0.05340353399515152, 'timing/model': 4.86575570423156, 'learning_rate/qwen_vl_interface': 9.12420166079635e-07, 'learning_rate/action_model': 9.12420166079635e-06, 'epoch': 2.23}) +[2026-06-18 06:01:31,694][starVLA.training.train_starvla][INFO] - Step 6079, Loss: {'train/loss': 0.42520496249198914, 'train/grad_norm_pre_clip': 30.596040725708008, 'timing/data': 0.05198060628026724, 'timing/model': 1.2960215518251061, 'learning_rate/qwen_vl_interface': 9.115391714548418e-07, 'learning_rate/action_model': 9.115391714548418e-06, 'epoch': 2.23}) +[2026-06-18 06:01:36,279][starVLA.training.train_starvla][INFO] - Step 6080, Loss: {'train/loss': 0.4115520417690277, 'train/grad_norm_pre_clip': 27.154861450195312, 'timing/data': 0.05688950885087252, 'timing/model': 1.0674749910831451, 'learning_rate/qwen_vl_interface': 9.106590761976454e-07, 'learning_rate/action_model': 9.106590761976455e-06, 'epoch': 2.23}) +[2026-06-18 06:01:44,711][starVLA.training.train_starvla][INFO] - Step 6081, Loss: {'train/loss': 0.4341471195220947, 'train/grad_norm_pre_clip': 25.157072067260742, 'timing/data': 0.041207519359886646, 'timing/model': 1.5205276254564524, 'learning_rate/qwen_vl_interface': 9.097798804904934e-07, 'learning_rate/action_model': 9.097798804904933e-06, 'epoch': 2.23}) +[2026-06-18 06:01:51,037][starVLA.training.train_starvla][INFO] - Step 6082, Loss: {'train/loss': 0.08914946764707565, 'train/grad_norm_pre_clip': 21.72164535522461, 'timing/data': 0.04742607846856117, 'timing/model': 3.260179183445871, 'learning_rate/qwen_vl_interface': 9.089015845156422e-07, 'learning_rate/action_model': 9.089015845156422e-06, 'epoch': 2.23}) +[2026-06-18 06:01:55,595][starVLA.training.train_starvla][INFO] - Step 6083, Loss: {'train/loss': 0.23711800575256348, 'train/grad_norm_pre_clip': 24.30803680419922, 'timing/data': 0.04882537201046944, 'timing/model': 1.0875294264405966, 'learning_rate/qwen_vl_interface': 9.080241884551627e-07, 'learning_rate/action_model': 9.080241884551627e-06, 'epoch': 2.23}) +[2026-06-18 06:02:00,178][starVLA.training.train_starvla][INFO] - Step 6084, Loss: {'train/loss': 0.4589531123638153, 'train/grad_norm_pre_clip': 21.42695426940918, 'timing/data': 0.0402221642434597, 'timing/model': 1.1032593045383692, 'learning_rate/qwen_vl_interface': 9.071476924909417e-07, 'learning_rate/action_model': 9.071476924909416e-06, 'epoch': 2.23}) +[2026-06-18 06:02:05,004][starVLA.training.train_starvla][INFO] - Step 6085, Loss: {'train/loss': 0.314022958278656, 'train/grad_norm_pre_clip': 24.951017379760742, 'timing/data': 0.05094822868704796, 'timing/model': 1.1354620717465878, 'learning_rate/qwen_vl_interface': 9.062720968046773e-07, 'learning_rate/action_model': 9.062720968046773e-06, 'epoch': 2.23}) +[2026-06-18 06:02:10,503][starVLA.training.train_starvla][INFO] - Step 6086, Loss: {'train/loss': 0.26234835386276245, 'train/grad_norm_pre_clip': 19.78476333618164, 'timing/data': 0.0689036687836051, 'timing/model': 1.447360297665, 'learning_rate/qwen_vl_interface': 9.053974015778817e-07, 'learning_rate/action_model': 9.053974015778817e-06, 'epoch': 2.23}) +[2026-06-18 06:02:17,210][starVLA.training.train_starvla][INFO] - Step 6087, Loss: {'train/loss': 0.30124595761299133, 'train/grad_norm_pre_clip': 27.674659729003906, 'timing/data': 0.07145513407886028, 'timing/model': 1.1010646326467395, 'learning_rate/qwen_vl_interface': 9.045236069918788e-07, 'learning_rate/action_model': 9.045236069918788e-06, 'epoch': 2.23}) +[2026-06-18 06:02:22,626][starVLA.training.train_starvla][INFO] - Step 6088, Loss: {'train/loss': 0.5685074925422668, 'train/grad_norm_pre_clip': 36.511653900146484, 'timing/data': 0.050707973539829254, 'timing/model': 1.0898832324892282, 'learning_rate/qwen_vl_interface': 9.036507132278085e-07, 'learning_rate/action_model': 9.036507132278086e-06, 'epoch': 2.23}) +[2026-06-18 06:02:27,618][starVLA.training.train_starvla][INFO] - Step 6089, Loss: {'train/loss': 0.2803778350353241, 'train/grad_norm_pre_clip': 41.245174407958984, 'timing/data': 0.05972052179276943, 'timing/model': 1.0779252396896482, 'learning_rate/qwen_vl_interface': 9.02778720466622e-07, 'learning_rate/action_model': 9.02778720466622e-06, 'epoch': 2.23}) +[2026-06-18 06:02:32,635][starVLA.training.train_starvla][INFO] - Step 6090, Loss: {'train/loss': 0.4125787615776062, 'train/grad_norm_pre_clip': 26.752704620361328, 'timing/data': 0.04077292326837778, 'timing/model': 1.4687707480043173, 'learning_rate/qwen_vl_interface': 9.019076288890852e-07, 'learning_rate/action_model': 9.019076288890852e-06, 'epoch': 2.23}) +[2026-06-18 06:02:37,195][starVLA.training.train_starvla][INFO] - Step 6091, Loss: {'train/loss': 0.10778678953647614, 'train/grad_norm_pre_clip': 18.170421600341797, 'timing/data': 0.05721462517976761, 'timing/model': 1.4836228573694825, 'learning_rate/qwen_vl_interface': 9.010374386757755e-07, 'learning_rate/action_model': 9.010374386757754e-06, 'epoch': 2.23}) +[2026-06-18 06:02:50,362][starVLA.training.train_starvla][INFO] - Step 6092, Loss: {'train/loss': 0.36287549138069153, 'train/grad_norm_pre_clip': 28.128433227539062, 'timing/data': 0.046556838788092136, 'timing/model': 5.856298812665045, 'learning_rate/qwen_vl_interface': 9.001681500070842e-07, 'learning_rate/action_model': 9.001681500070842e-06, 'epoch': 2.23}) +[2026-06-18 06:02:55,795][starVLA.training.train_starvla][INFO] - Step 6093, Loss: {'train/loss': 0.11093628406524658, 'train/grad_norm_pre_clip': 11.135242462158203, 'timing/data': 0.04733919631689787, 'timing/model': 1.2950435299426317, 'learning_rate/qwen_vl_interface': 8.992997630632161e-07, 'learning_rate/action_model': 8.99299763063216e-06, 'epoch': 2.23}) +[2026-06-18 06:03:02,408][starVLA.training.train_starvla][INFO] - Step 6094, Loss: {'train/loss': 0.17217445373535156, 'train/grad_norm_pre_clip': 19.15478515625, 'timing/data': 0.10947002097964287, 'timing/model': 1.15141212195158, 'learning_rate/qwen_vl_interface': 8.984322780241879e-07, 'learning_rate/action_model': 8.984322780241878e-06, 'epoch': 2.23}) +[2026-06-18 06:03:10,509][starVLA.training.train_starvla][INFO] - Step 6095, Loss: {'train/loss': 0.14135032892227173, 'train/grad_norm_pre_clip': 28.336233139038086, 'timing/data': 0.01248264405876398, 'timing/model': 1.3054077625274658, 'learning_rate/qwen_vl_interface': 8.975656950698325e-07, 'learning_rate/action_model': 8.975656950698324e-06, 'epoch': 2.23}) +[2026-06-18 06:03:15,541][starVLA.training.train_starvla][INFO] - Step 6096, Loss: {'train/loss': 0.35119447112083435, 'train/grad_norm_pre_clip': 21.29684829711914, 'timing/data': 0.04771739523857832, 'timing/model': 1.457484939135611, 'learning_rate/qwen_vl_interface': 8.967000143797913e-07, 'learning_rate/action_model': 8.967000143797913e-06, 'epoch': 2.23}) +[2026-06-18 06:03:20,058][starVLA.training.train_starvla][INFO] - Step 6097, Loss: {'train/loss': 0.1285715252161026, 'train/grad_norm_pre_clip': 16.363861083984375, 'timing/data': 0.04493934754282236, 'timing/model': 1.2350519839674234, 'learning_rate/qwen_vl_interface': 8.958352361335214e-07, 'learning_rate/action_model': 8.958352361335213e-06, 'epoch': 2.23}) +[2026-06-18 06:03:24,807][starVLA.training.train_starvla][INFO] - Step 6098, Loss: {'train/loss': 0.2778334319591522, 'train/grad_norm_pre_clip': 25.396699905395508, 'timing/data': 0.061483354307711124, 'timing/model': 1.0883418060839176, 'learning_rate/qwen_vl_interface': 8.949713605102917e-07, 'learning_rate/action_model': 8.949713605102916e-06, 'epoch': 2.23}) +[2026-06-18 06:03:29,758][starVLA.training.train_starvla][INFO] - Step 6099, Loss: {'train/loss': 0.25655221939086914, 'train/grad_norm_pre_clip': 20.61538314819336, 'timing/data': 0.013194486498832703, 'timing/model': 1.125223284587264, 'learning_rate/qwen_vl_interface': 8.941083876891857e-07, 'learning_rate/action_model': 8.941083876891856e-06, 'epoch': 2.23}) +[2026-06-18 06:07:42,337][starVLA.training.train_starvla][INFO] - Step 6100, Loss: {'train/loss': 0.1795961856842041, 'train/grad_norm_pre_clip': 16.99225425720215, 'eval/loss': 0.5307633280754089, 'eval/latency_2/loss': 0.5307633280754089, 'eval/demon_fire_precision': 0.958771560790913, 'eval/demon_fire_recall': 0.9535564853556485, 'eval/demon_fire_f1': 0.9561569121040487, 'eval/demon_fire_rate_model': 0.6613800779076238, 'eval/demon_fire_rate_teacher': 0.6649972175848636, 'eval/demon_left_precision': 0.9066232356134636, 'eval/demon_left_recall': 0.8789473684210526, 'eval/demon_left_f1': 0.8925708177445216, 'eval/demon_left_rate_model': 0.25626043405676124, 'eval/demon_left_rate_teacher': 0.2643294379521425, 'eval/demon_right_precision': 0.9184597961494904, 'eval/demon_right_recall': 0.9163841807909604, 'eval/demon_right_f1': 0.917420814479638, 'eval/demon_right_rate_model': 0.24568725653867557, 'eval/demon_right_rate_teacher': 0.24624373956594323, 'eval/demon_move_f1': 0.9049958161120798, 'eval/demon_cc_f1': 0.94021134353578, 'eval/latency_2/demon_fire_precision': 0.958771560790913, 'eval/latency_2/demon_fire_recall': 0.9535564853556485, 'eval/latency_2/demon_fire_f1': 0.9561569121040487, 'eval/latency_2/demon_fire_rate_model': 0.6613800779076238, 'eval/latency_2/demon_fire_rate_teacher': 0.6649972175848636, 'eval/latency_2/demon_left_precision': 0.9066232356134636, 'eval/latency_2/demon_left_recall': 0.8789473684210526, 'eval/latency_2/demon_left_f1': 0.8925708177445216, 'eval/latency_2/demon_left_rate_model': 0.25626043405676124, 'eval/latency_2/demon_left_rate_teacher': 0.2643294379521425, 'eval/latency_2/demon_right_precision': 0.9184597961494904, 'eval/latency_2/demon_right_recall': 0.9163841807909604, 'eval/latency_2/demon_right_f1': 0.917420814479638, 'eval/latency_2/demon_right_rate_model': 0.24568725653867557, 'eval/latency_2/demon_right_rate_teacher': 0.24624373956594323, 'eval/latency_2/demon_move_f1': 0.9049958161120798, 'eval/latency_2/demon_cc_f1': 0.94021134353578, 'timing/data': 0.010801241733133793, 'timing/model': 1.0968052698299289, 'learning_rate/qwen_vl_interface': 8.932463178490991e-07, 'learning_rate/action_model': 8.93246317849099e-06, 'epoch': 2.23}) +[2026-06-18 06:07:42,371][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_6100_state +[2026-06-18 06:07:42,378][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 06:07:51,490][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_6100_state/model.safetensors +[2026-06-18 06:08:08,251][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_6100_state/optimizer.bin +[2026-06-18 06:08:08,252][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_6100_state/sampler.bin +[2026-06-18 06:08:08,253][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_6100_state/sampler_1.bin +[2026-06-18 06:08:08,256][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_6100_state/random_states_0.pkl +[2026-06-18 06:08:08,310][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 6100 +[2026-06-18 06:08:08,311][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 06:08:08,332][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 06:08:28,916][starVLA.training.train_starvla][INFO] - Step 6101, Loss: {'train/loss': 0.19843219220638275, 'train/grad_norm_pre_clip': 19.72181510925293, 'timing/data': 0.0018560858443379402, 'timing/model': 5.94611821975559, 'learning_rate/qwen_vl_interface': 8.923851511687379e-07, 'learning_rate/action_model': 8.923851511687378e-06, 'epoch': 2.23}) +[2026-06-18 06:08:39,318][starVLA.training.train_starvla][INFO] - Step 6102, Loss: {'train/loss': 0.1574597805738449, 'train/grad_norm_pre_clip': 16.418346405029297, 'timing/data': 0.0011440711095929146, 'timing/model': 2.222627205774188, 'learning_rate/qwen_vl_interface': 8.915248878266228e-07, 'learning_rate/action_model': 8.915248878266227e-06, 'epoch': 2.23}) +[2026-06-18 06:08:48,582][starVLA.training.train_starvla][INFO] - Step 6103, Loss: {'train/loss': 0.19383101165294647, 'train/grad_norm_pre_clip': 20.730010986328125, 'timing/data': 0.0009042629972100258, 'timing/model': 1.1264852872118354, 'learning_rate/qwen_vl_interface': 8.906655280010895e-07, 'learning_rate/action_model': 8.906655280010896e-06, 'epoch': 2.23}) +[2026-06-18 06:08:54,709][starVLA.training.train_starvla][INFO] - Step 6104, Loss: {'train/loss': 0.15455232560634613, 'train/grad_norm_pre_clip': 17.742218017578125, 'timing/data': 0.0008706878870725632, 'timing/model': 2.286067735403776, 'learning_rate/qwen_vl_interface': 8.898070718702831e-07, 'learning_rate/action_model': 8.898070718702831e-06, 'epoch': 2.24}) +[2026-06-18 06:09:05,026][starVLA.training.train_starvla][INFO] - Step 6105, Loss: {'train/loss': 0.1651708036661148, 'train/grad_norm_pre_clip': 25.02393913269043, 'timing/data': 0.0012041423469781876, 'timing/model': 5.82960997056216, 'learning_rate/qwen_vl_interface': 8.889495196121628e-07, 'learning_rate/action_model': 8.889495196121627e-06, 'epoch': 2.24}) +[2026-06-18 06:09:19,159][starVLA.training.train_starvla][INFO] - Step 6106, Loss: {'train/loss': 0.027549175545573235, 'train/grad_norm_pre_clip': 4.423222541809082, 'timing/data': 0.0009255725890398026, 'timing/model': 4.493539999239147, 'learning_rate/qwen_vl_interface': 8.880928714044994e-07, 'learning_rate/action_model': 8.880928714044993e-06, 'epoch': 2.24}) +[2026-06-18 06:09:32,423][starVLA.training.train_starvla][INFO] - Step 6107, Loss: {'train/loss': 0.3524697721004486, 'train/grad_norm_pre_clip': 25.636489868164062, 'timing/data': 0.005632403306663036, 'timing/model': 3.288505868986249, 'learning_rate/qwen_vl_interface': 8.872371274248773e-07, 'learning_rate/action_model': 8.872371274248772e-06, 'epoch': 2.24}) +[2026-06-18 06:09:46,238][starVLA.training.train_starvla][INFO] - Step 6108, Loss: {'train/loss': 0.677116870880127, 'train/grad_norm_pre_clip': 43.9768180847168, 'timing/data': 0.0009865853935480118, 'timing/model': 1.1365208765491843, 'learning_rate/qwen_vl_interface': 8.863822878506931e-07, 'learning_rate/action_model': 8.86382287850693e-06, 'epoch': 2.24}) +[2026-06-18 06:09:54,158][starVLA.training.train_starvla][INFO] - Step 6109, Loss: {'train/loss': 0.1672705113887787, 'train/grad_norm_pre_clip': 18.619380950927734, 'timing/data': 0.0008962266147136688, 'timing/model': 2.299413342960179, 'learning_rate/qwen_vl_interface': 8.855283528591566e-07, 'learning_rate/action_model': 8.855283528591565e-06, 'epoch': 2.24}) +[2026-06-18 06:10:00,037][starVLA.training.train_starvla][INFO] - Step 6110, Loss: {'train/loss': 0.4098935127258301, 'train/grad_norm_pre_clip': 29.237104415893555, 'timing/data': 0.001044667325913906, 'timing/model': 1.4416192406788468, 'learning_rate/qwen_vl_interface': 8.846753226272884e-07, 'learning_rate/action_model': 8.846753226272883e-06, 'epoch': 2.24}) +[2026-06-18 06:10:04,693][starVLA.training.train_starvla][INFO] - Step 6111, Loss: {'train/loss': 0.30729272961616516, 'train/grad_norm_pre_clip': 34.01676940917969, 'timing/data': 0.0010894397273659706, 'timing/model': 1.0359944170340896, 'learning_rate/qwen_vl_interface': 8.838231973319235e-07, 'learning_rate/action_model': 8.838231973319235e-06, 'epoch': 2.24}) +[2026-06-18 06:10:09,517][starVLA.training.train_starvla][INFO] - Step 6112, Loss: {'train/loss': 0.23606212437152863, 'train/grad_norm_pre_clip': 22.176528930664062, 'timing/data': 0.0010987082496285439, 'timing/model': 1.7268123980611563, 'learning_rate/qwen_vl_interface': 8.829719771497072e-07, 'learning_rate/action_model': 8.829719771497071e-06, 'epoch': 2.24}) +[2026-06-18 06:10:14,031][starVLA.training.train_starvla][INFO] - Step 6113, Loss: {'train/loss': 0.32691460847854614, 'train/grad_norm_pre_clip': 21.992605209350586, 'timing/data': 0.0009838342666625977, 'timing/model': 1.4239944200962782, 'learning_rate/qwen_vl_interface': 8.821216622570999e-07, 'learning_rate/action_model': 8.821216622570998e-06, 'epoch': 2.24}) +[2026-06-18 06:10:18,873][starVLA.training.train_starvla][INFO] - Step 6114, Loss: {'train/loss': 0.4649171233177185, 'train/grad_norm_pre_clip': 29.09187889099121, 'timing/data': 0.0011200634762644768, 'timing/model': 1.0367581145837903, 'learning_rate/qwen_vl_interface': 8.812722528303727e-07, 'learning_rate/action_model': 8.812722528303727e-06, 'epoch': 2.24}) +[2026-06-18 06:10:23,323][starVLA.training.train_starvla][INFO] - Step 6115, Loss: {'train/loss': 0.19766926765441895, 'train/grad_norm_pre_clip': 21.307592391967773, 'timing/data': 0.00129670649766922, 'timing/model': 1.0740955341607332, 'learning_rate/qwen_vl_interface': 8.804237490456076e-07, 'learning_rate/action_model': 8.804237490456075e-06, 'epoch': 2.24}) +[2026-06-18 06:10:28,312][starVLA.training.train_starvla][INFO] - Step 6116, Loss: {'train/loss': 0.3732927143573761, 'train/grad_norm_pre_clip': 29.00886344909668, 'timing/data': 0.0011255443096160889, 'timing/model': 1.0641592741012573, 'learning_rate/qwen_vl_interface': 8.795761510787007e-07, 'learning_rate/action_model': 8.795761510787006e-06, 'epoch': 2.24}) +[2026-06-18 06:10:32,735][starVLA.training.train_starvla][INFO] - Step 6117, Loss: {'train/loss': 0.27571845054626465, 'train/grad_norm_pre_clip': 22.863319396972656, 'timing/data': 0.0010792510583996773, 'timing/model': 1.389064715243876, 'learning_rate/qwen_vl_interface': 8.78729459105361e-07, 'learning_rate/action_model': 8.78729459105361e-06, 'epoch': 2.24}) +[2026-06-18 06:10:36,982][starVLA.training.train_starvla][INFO] - Step 6118, Loss: {'train/loss': 0.08956146985292435, 'train/grad_norm_pre_clip': 9.17663288116455, 'timing/data': 0.0008836966007947922, 'timing/model': 1.0562555147334933, 'learning_rate/qwen_vl_interface': 8.778836733011079e-07, 'learning_rate/action_model': 8.778836733011079e-06, 'epoch': 2.24}) +[2026-06-18 06:10:45,419][starVLA.training.train_starvla][INFO] - Step 6119, Loss: {'train/loss': 0.03356707841157913, 'train/grad_norm_pre_clip': 4.549919605255127, 'timing/data': 0.0023982329294085503, 'timing/model': 1.1857988657429814, 'learning_rate/qwen_vl_interface': 8.770387938412753e-07, 'learning_rate/action_model': 8.770387938412753e-06, 'epoch': 2.24}) +[2026-06-18 06:10:50,331][starVLA.training.train_starvla][INFO] - Step 6120, Loss: {'train/loss': 0.10708059370517731, 'train/grad_norm_pre_clip': 11.713348388671875, 'timing/data': 0.0012542903423309326, 'timing/model': 1.0752457519993186, 'learning_rate/qwen_vl_interface': 8.76194820901004e-07, 'learning_rate/action_model': 8.76194820901004e-06, 'epoch': 2.24}) +[2026-06-18 06:10:55,070][starVLA.training.train_starvla][INFO] - Step 6121, Loss: {'train/loss': 0.25807008147239685, 'train/grad_norm_pre_clip': 29.715070724487305, 'timing/data': 0.0010365108028054237, 'timing/model': 1.1696500899270177, 'learning_rate/qwen_vl_interface': 8.753517546552545e-07, 'learning_rate/action_model': 8.753517546552544e-06, 'epoch': 2.24}) +[2026-06-18 06:10:59,606][starVLA.training.train_starvla][INFO] - Step 6122, Loss: {'train/loss': 0.38662755489349365, 'train/grad_norm_pre_clip': 44.057579040527344, 'timing/data': 0.0011050887405872345, 'timing/model': 1.3340474516153336, 'learning_rate/qwen_vl_interface': 8.74509595278793e-07, 'learning_rate/action_model': 8.74509595278793e-06, 'epoch': 2.24}) +[2026-06-18 06:11:05,349][starVLA.training.train_starvla][INFO] - Step 6123, Loss: {'train/loss': 0.1855868399143219, 'train/grad_norm_pre_clip': 27.137205123901367, 'timing/data': 0.0010983524844050407, 'timing/model': 1.854024738073349, 'learning_rate/qwen_vl_interface': 8.736683429462003e-07, 'learning_rate/action_model': 8.736683429462004e-06, 'epoch': 2.24}) +[2026-06-18 06:11:10,419][starVLA.training.train_starvla][INFO] - Step 6124, Loss: {'train/loss': 0.25572967529296875, 'train/grad_norm_pre_clip': 33.96518325805664, 'timing/data': 0.0010990677401423454, 'timing/model': 1.072322585619986, 'learning_rate/qwen_vl_interface': 8.728279978318696e-07, 'learning_rate/action_model': 8.728279978318695e-06, 'epoch': 2.24}) +[2026-06-18 06:11:15,010][starVLA.training.train_starvla][INFO] - Step 6125, Loss: {'train/loss': 0.0768468901515007, 'train/grad_norm_pre_clip': 10.886756896972656, 'timing/data': 0.001364157535135746, 'timing/model': 1.0653104418888688, 'learning_rate/qwen_vl_interface': 8.719885601100052e-07, 'learning_rate/action_model': 8.719885601100051e-06, 'epoch': 2.24}) +[2026-06-18 06:11:20,129][starVLA.training.train_starvla][INFO] - Step 6126, Loss: {'train/loss': 0.5391836762428284, 'train/grad_norm_pre_clip': 37.96208572387695, 'timing/data': 0.0008629290387034416, 'timing/model': 1.5590572841465473, 'learning_rate/qwen_vl_interface': 8.71150029954623e-07, 'learning_rate/action_model': 8.71150029954623e-06, 'epoch': 2.24}) +[2026-06-18 06:11:33,546][starVLA.training.train_starvla][INFO] - Step 6127, Loss: {'train/loss': 0.4381437301635742, 'train/grad_norm_pre_clip': 31.51651954650879, 'timing/data': 0.0011529792100191116, 'timing/model': 1.5448702340945601, 'learning_rate/qwen_vl_interface': 8.703124075395504e-07, 'learning_rate/action_model': 8.703124075395504e-06, 'epoch': 2.24}) +[2026-06-18 06:11:41,308][starVLA.training.train_starvla][INFO] - Step 6128, Loss: {'train/loss': 0.09874773770570755, 'train/grad_norm_pre_clip': 12.914536476135254, 'timing/data': 0.001272382214665413, 'timing/model': 1.059120051562786, 'learning_rate/qwen_vl_interface': 8.694756930384305e-07, 'learning_rate/action_model': 8.694756930384306e-06, 'epoch': 2.24}) +[2026-06-18 06:11:46,107][starVLA.training.train_starvla][INFO] - Step 6129, Loss: {'train/loss': 0.3123049736022949, 'train/grad_norm_pre_clip': 30.28882598876953, 'timing/data': 0.001026148907840252, 'timing/model': 1.061384717002511, 'learning_rate/qwen_vl_interface': 8.686398866247121e-07, 'learning_rate/action_model': 8.68639886624712e-06, 'epoch': 2.24}) +[2026-06-18 06:11:50,770][starVLA.training.train_starvla][INFO] - Step 6130, Loss: {'train/loss': 0.3072211742401123, 'train/grad_norm_pre_clip': 26.4312801361084, 'timing/data': 0.0010139672085642815, 'timing/model': 1.1916686780750751, 'learning_rate/qwen_vl_interface': 8.6780498847166e-07, 'learning_rate/action_model': 8.678049884716599e-06, 'epoch': 2.24}) +[2026-06-18 06:12:00,668][starVLA.training.train_starvla][INFO] - Step 6131, Loss: {'train/loss': 0.3028099536895752, 'train/grad_norm_pre_clip': 34.123046875, 'timing/data': 0.0007514981552958488, 'timing/model': 5.747809669934213, 'learning_rate/qwen_vl_interface': 8.669709987523479e-07, 'learning_rate/action_model': 8.669709987523477e-06, 'epoch': 2.25}) +[2026-06-18 06:12:09,994][starVLA.training.train_starvla][INFO] - Step 6132, Loss: {'train/loss': 0.21443679928779602, 'train/grad_norm_pre_clip': 28.829906463623047, 'timing/data': 0.0007530041038990021, 'timing/model': 1.4691521637141705, 'learning_rate/qwen_vl_interface': 8.661379176396651e-07, 'learning_rate/action_model': 8.66137917639665e-06, 'epoch': 2.25}) +[2026-06-18 06:12:18,916][starVLA.training.train_starvla][INFO] - Step 6133, Loss: {'train/loss': 0.12648677825927734, 'train/grad_norm_pre_clip': 16.193248748779297, 'timing/data': 0.0007799332961440086, 'timing/model': 5.4671322563663125, 'learning_rate/qwen_vl_interface': 8.653057453063092e-07, 'learning_rate/action_model': 8.653057453063091e-06, 'epoch': 2.25}) +[2026-06-18 06:12:25,346][starVLA.training.train_starvla][INFO] - Step 6134, Loss: {'train/loss': 0.4470476508140564, 'train/grad_norm_pre_clip': 30.46141242980957, 'timing/data': 0.00222957506775856, 'timing/model': 1.186575590632856, 'learning_rate/qwen_vl_interface': 8.644744819247908e-07, 'learning_rate/action_model': 8.644744819247908e-06, 'epoch': 2.25}) +[2026-06-18 06:12:30,080][starVLA.training.train_starvla][INFO] - Step 6135, Loss: {'train/loss': 0.13669775426387787, 'train/grad_norm_pre_clip': 18.15083122253418, 'timing/data': 0.0009044753387570381, 'timing/model': 1.064354750327766, 'learning_rate/qwen_vl_interface': 8.636441276674308e-07, 'learning_rate/action_model': 8.636441276674308e-06, 'epoch': 2.25}) +[2026-06-18 06:12:39,904][starVLA.training.train_starvla][INFO] - Step 6136, Loss: {'train/loss': 0.29561907052993774, 'train/grad_norm_pre_clip': 23.06279945373535, 'timing/data': 0.001232760027050972, 'timing/model': 1.054014140740037, 'learning_rate/qwen_vl_interface': 8.628146827063632e-07, 'learning_rate/action_model': 8.628146827063633e-06, 'epoch': 2.25}) +[2026-06-18 06:12:52,012][starVLA.training.train_starvla][INFO] - Step 6137, Loss: {'train/loss': 0.017550310119986534, 'train/grad_norm_pre_clip': 3.362086772918701, 'timing/data': 0.004149692133069038, 'timing/model': 6.090363913215697, 'learning_rate/qwen_vl_interface': 8.619861472135332e-07, 'learning_rate/action_model': 8.619861472135332e-06, 'epoch': 2.25}) +[2026-06-18 06:13:03,453][starVLA.training.train_starvla][INFO] - Step 6138, Loss: {'train/loss': 0.18674269318580627, 'train/grad_norm_pre_clip': 21.696006774902344, 'timing/data': 0.00097730103880167, 'timing/model': 4.537577774375677, 'learning_rate/qwen_vl_interface': 8.611585213606964e-07, 'learning_rate/action_model': 8.611585213606964e-06, 'epoch': 2.25}) +[2026-06-18 06:13:08,293][starVLA.training.train_starvla][INFO] - Step 6139, Loss: {'train/loss': 0.1550265997648239, 'train/grad_norm_pre_clip': 18.374500274658203, 'timing/data': 0.001270008273422718, 'timing/model': 1.1252545975148678, 'learning_rate/qwen_vl_interface': 8.603318053194206e-07, 'learning_rate/action_model': 8.603318053194205e-06, 'epoch': 2.25}) +[2026-06-18 06:13:14,894][starVLA.training.train_starvla][INFO] - Step 6140, Loss: {'train/loss': 0.12446372210979462, 'train/grad_norm_pre_clip': 16.95439338684082, 'timing/data': 0.000914527103304863, 'timing/model': 1.0450888704508543, 'learning_rate/qwen_vl_interface': 8.595059992610846e-07, 'learning_rate/action_model': 8.595059992610846e-06, 'epoch': 2.25}) +[2026-06-18 06:13:21,538][starVLA.training.train_starvla][INFO] - Step 6141, Loss: {'train/loss': 0.3372149169445038, 'train/grad_norm_pre_clip': 28.066091537475586, 'timing/data': 0.001209920272231102, 'timing/model': 1.0808220924809575, 'learning_rate/qwen_vl_interface': 8.586811033568796e-07, 'learning_rate/action_model': 8.586811033568797e-06, 'epoch': 2.25}) +[2026-06-18 06:13:26,440][starVLA.training.train_starvla][INFO] - Step 6142, Loss: {'train/loss': 0.888139545917511, 'train/grad_norm_pre_clip': 44.95871353149414, 'timing/data': 0.0011236732825636864, 'timing/model': 1.4862277880311012, 'learning_rate/qwen_vl_interface': 8.578571177778076e-07, 'learning_rate/action_model': 8.578571177778075e-06, 'epoch': 2.25}) +[2026-06-18 06:13:30,612][starVLA.training.train_starvla][INFO] - Step 6143, Loss: {'train/loss': 0.36800649762153625, 'train/grad_norm_pre_clip': 33.476036071777344, 'timing/data': 0.000894324854016304, 'timing/model': 1.13308313395828, 'learning_rate/qwen_vl_interface': 8.570340426946805e-07, 'learning_rate/action_model': 8.570340426946804e-06, 'epoch': 2.25}) +[2026-06-18 06:13:35,632][starVLA.training.train_starvla][INFO] - Step 6144, Loss: {'train/loss': 0.3391459882259369, 'train/grad_norm_pre_clip': 35.352317810058594, 'timing/data': 0.0011259755119681358, 'timing/model': 1.3250104002654552, 'learning_rate/qwen_vl_interface': 8.562118782781241e-07, 'learning_rate/action_model': 8.56211878278124e-06, 'epoch': 2.25}) +[2026-06-18 06:13:42,703][starVLA.training.train_starvla][INFO] - Step 6145, Loss: {'train/loss': 0.3428111970424652, 'train/grad_norm_pre_clip': 44.7840461730957, 'timing/data': 0.0009949728846549988, 'timing/model': 1.0880050156265497, 'learning_rate/qwen_vl_interface': 8.553906246985722e-07, 'learning_rate/action_model': 8.553906246985721e-06, 'epoch': 2.25}) +[2026-06-18 06:13:48,306][starVLA.training.train_starvla][INFO] - Step 6146, Loss: {'train/loss': 0.07208728790283203, 'train/grad_norm_pre_clip': 10.440512657165527, 'timing/data': 0.05335881467908621, 'timing/model': 1.3951734602451324, 'learning_rate/qwen_vl_interface': 8.545702821262733e-07, 'learning_rate/action_model': 8.545702821262733e-06, 'epoch': 2.25}) +[2026-06-18 06:13:53,088][starVLA.training.train_starvla][INFO] - Step 6147, Loss: {'train/loss': 0.13655824959278107, 'train/grad_norm_pre_clip': 18.8162784576416, 'timing/data': 0.001092781312763691, 'timing/model': 1.134512503631413, 'learning_rate/qwen_vl_interface': 8.537508507312844e-07, 'learning_rate/action_model': 8.537508507312844e-06, 'epoch': 2.25}) +[2026-06-18 06:13:58,341][starVLA.training.train_starvla][INFO] - Step 6148, Loss: {'train/loss': 0.2593702971935272, 'train/grad_norm_pre_clip': 26.27223014831543, 'timing/data': 0.0009751636534929276, 'timing/model': 1.1379074109718204, 'learning_rate/qwen_vl_interface': 8.52932330683476e-07, 'learning_rate/action_model': 8.529323306834759e-06, 'epoch': 2.25}) +[2026-06-18 06:14:03,143][starVLA.training.train_starvla][INFO] - Step 6149, Loss: {'train/loss': 0.24137139320373535, 'train/grad_norm_pre_clip': 24.04928207397461, 'timing/data': 0.0008772788569331169, 'timing/model': 1.3748449916020036, 'learning_rate/qwen_vl_interface': 8.52114722152525e-07, 'learning_rate/action_model': 8.521147221525249e-06, 'epoch': 2.25}) +[2026-06-18 06:14:09,819][starVLA.training.train_starvla][INFO] - Step 6150, Loss: {'train/loss': 0.18593829870224, 'train/grad_norm_pre_clip': 26.275829315185547, 'timing/data': 0.0016524028033018112, 'timing/model': 1.5622915206477046, 'learning_rate/qwen_vl_interface': 8.51298025307925e-07, 'learning_rate/action_model': 8.51298025307925e-06, 'epoch': 2.25}) +[2026-06-18 06:14:14,898][starVLA.training.train_starvla][INFO] - Step 6151, Loss: {'train/loss': 0.17473198473453522, 'train/grad_norm_pre_clip': 22.679931640625, 'timing/data': 0.000946911983191967, 'timing/model': 1.1900852527469397, 'learning_rate/qwen_vl_interface': 8.504822403189783e-07, 'learning_rate/action_model': 8.504822403189782e-06, 'epoch': 2.25}) +[2026-06-18 06:14:19,850][starVLA.training.train_starvla][INFO] - Step 6152, Loss: {'train/loss': 0.39998969435691833, 'train/grad_norm_pre_clip': 28.016300201416016, 'timing/data': 0.0017957575619220734, 'timing/model': 1.3462223075330257, 'learning_rate/qwen_vl_interface': 8.496673673547967e-07, 'learning_rate/action_model': 8.496673673547968e-06, 'epoch': 2.25}) +[2026-06-18 06:14:24,380][starVLA.training.train_starvla][INFO] - Step 6153, Loss: {'train/loss': 0.1430797278881073, 'train/grad_norm_pre_clip': 16.891529083251953, 'timing/data': 0.001254342496395111, 'timing/model': 1.1184118511155248, 'learning_rate/qwen_vl_interface': 8.488534065843046e-07, 'learning_rate/action_model': 8.488534065843045e-06, 'epoch': 2.25}) +[2026-06-18 06:14:29,411][starVLA.training.train_starvla][INFO] - Step 6154, Loss: {'train/loss': 0.2741084396839142, 'train/grad_norm_pre_clip': 25.3566837310791, 'timing/data': 0.0007788529619574547, 'timing/model': 1.052455622702837, 'learning_rate/qwen_vl_interface': 8.480403581762378e-07, 'learning_rate/action_model': 8.480403581762378e-06, 'epoch': 2.25}) +[2026-06-18 06:14:34,187][starVLA.training.train_starvla][INFO] - Step 6155, Loss: {'train/loss': 0.37090015411376953, 'train/grad_norm_pre_clip': 29.292020797729492, 'timing/data': 0.001324341632425785, 'timing/model': 1.226430238224566, 'learning_rate/qwen_vl_interface': 8.472282222991411e-07, 'learning_rate/action_model': 8.47228222299141e-06, 'epoch': 2.25}) +[2026-06-18 06:14:39,013][starVLA.training.train_starvla][INFO] - Step 6156, Loss: {'train/loss': 0.15579330921173096, 'train/grad_norm_pre_clip': 26.354263305664062, 'timing/data': 0.0013915197923779488, 'timing/model': 1.138668674044311, 'learning_rate/qwen_vl_interface': 8.464169991213705e-07, 'learning_rate/action_model': 8.464169991213706e-06, 'epoch': 2.25}) +[2026-06-18 06:14:43,498][starVLA.training.train_starvla][INFO] - Step 6157, Loss: {'train/loss': 0.10857212543487549, 'train/grad_norm_pre_clip': 12.270118713378906, 'timing/data': 0.0008794469758868217, 'timing/model': 1.4312864216044545, 'learning_rate/qwen_vl_interface': 8.456066888110964e-07, 'learning_rate/action_model': 8.456066888110964e-06, 'epoch': 2.25}) +[2026-06-18 06:14:48,290][starVLA.training.train_starvla][INFO] - Step 6158, Loss: {'train/loss': 0.06719648838043213, 'train/grad_norm_pre_clip': 11.015710830688477, 'timing/data': 0.001145191490650177, 'timing/model': 1.0268751503899693, 'learning_rate/qwen_vl_interface': 8.447972915362941e-07, 'learning_rate/action_model': 8.44797291536294e-06, 'epoch': 2.25}) +[2026-06-18 06:15:00,549][starVLA.training.train_starvla][INFO] - Step 6159, Loss: {'train/loss': 0.3032793402671814, 'train/grad_norm_pre_clip': 26.153024673461914, 'timing/data': 0.0011009639129042625, 'timing/model': 1.078927137888968, 'learning_rate/qwen_vl_interface': 8.439888074647536e-07, 'learning_rate/action_model': 8.439888074647536e-06, 'epoch': 2.26}) +[2026-06-18 06:15:05,467][starVLA.training.train_starvla][INFO] - Step 6160, Loss: {'train/loss': 0.147198885679245, 'train/grad_norm_pre_clip': 15.868736267089844, 'timing/data': 0.001887694001197815, 'timing/model': 1.3889439906924963, 'learning_rate/qwen_vl_interface': 8.431812367640732e-07, 'learning_rate/action_model': 8.431812367640731e-06, 'epoch': 2.26}) +[2026-06-18 06:15:11,815][starVLA.training.train_starvla][INFO] - Step 6161, Loss: {'train/loss': 0.2869425117969513, 'train/grad_norm_pre_clip': 26.664705276489258, 'timing/data': 0.0013481201604008675, 'timing/model': 1.0578726520761847, 'learning_rate/qwen_vl_interface': 8.423745796016659e-07, 'learning_rate/action_model': 8.42374579601666e-06, 'epoch': 2.26}) +[2026-06-18 06:15:16,353][starVLA.training.train_starvla][INFO] - Step 6162, Loss: {'train/loss': 0.31248563528060913, 'train/grad_norm_pre_clip': 23.265592575073242, 'timing/data': 0.0007556267082691193, 'timing/model': 1.4650891283527017, 'learning_rate/qwen_vl_interface': 8.41568836144751e-07, 'learning_rate/action_model': 8.41568836144751e-06, 'epoch': 2.26}) +[2026-06-18 06:15:24,256][starVLA.training.train_starvla][INFO] - Step 6163, Loss: {'train/loss': 0.12427490204572678, 'train/grad_norm_pre_clip': 14.208782196044922, 'timing/data': 0.0011813314631581306, 'timing/model': 1.1133758518844843, 'learning_rate/qwen_vl_interface': 8.407640065603601e-07, 'learning_rate/action_model': 8.4076400656036e-06, 'epoch': 2.26}) +[2026-06-18 06:15:29,355][starVLA.training.train_starvla][INFO] - Step 6164, Loss: {'train/loss': 0.19663825631141663, 'train/grad_norm_pre_clip': 19.17524528503418, 'timing/data': 0.0011791810393333435, 'timing/model': 1.2274442501366138, 'learning_rate/qwen_vl_interface': 8.399600910153336e-07, 'learning_rate/action_model': 8.399600910153336e-06, 'epoch': 2.26}) +[2026-06-18 06:15:34,014][starVLA.training.train_starvla][INFO] - Step 6165, Loss: {'train/loss': 0.2548799514770508, 'train/grad_norm_pre_clip': 19.53656768798828, 'timing/data': 0.000906023196876049, 'timing/model': 1.0456299427896738, 'learning_rate/qwen_vl_interface': 8.391570896763266e-07, 'learning_rate/action_model': 8.391570896763267e-06, 'epoch': 2.26}) +[2026-06-18 06:15:38,970][starVLA.training.train_starvla][INFO] - Step 6166, Loss: {'train/loss': 0.25274187326431274, 'train/grad_norm_pre_clip': 18.321805953979492, 'timing/data': 0.0038176393136382103, 'timing/model': 1.3642884893342853, 'learning_rate/qwen_vl_interface': 8.38355002709802e-07, 'learning_rate/action_model': 8.38355002709802e-06, 'epoch': 2.26}) +[2026-06-18 06:15:46,002][starVLA.training.train_starvla][INFO] - Step 6167, Loss: {'train/loss': 0.1763489544391632, 'train/grad_norm_pre_clip': 55.369712829589844, 'timing/data': 0.0012356285005807877, 'timing/model': 3.8001390444114804, 'learning_rate/qwen_vl_interface': 8.375538302820314e-07, 'learning_rate/action_model': 8.375538302820315e-06, 'epoch': 2.26}) +[2026-06-18 06:15:54,263][starVLA.training.train_starvla][INFO] - Step 6168, Loss: {'train/loss': 0.1622312217950821, 'train/grad_norm_pre_clip': 24.863719940185547, 'timing/data': 0.0013863816857337952, 'timing/model': 1.39118058886379, 'learning_rate/qwen_vl_interface': 8.367535725591003e-07, 'learning_rate/action_model': 8.367535725591003e-06, 'epoch': 2.26}) +[2026-06-18 06:16:00,918][starVLA.training.train_starvla][INFO] - Step 6169, Loss: {'train/loss': 0.1553458422422409, 'train/grad_norm_pre_clip': 15.182032585144043, 'timing/data': 0.0009744903072714806, 'timing/model': 1.1574104279279709, 'learning_rate/qwen_vl_interface': 8.359542297069024e-07, 'learning_rate/action_model': 8.359542297069024e-06, 'epoch': 2.26}) +[2026-06-18 06:16:07,104][starVLA.training.train_starvla][INFO] - Step 6170, Loss: {'train/loss': 0.18109558522701263, 'train/grad_norm_pre_clip': 20.35853385925293, 'timing/data': 0.0010468894615769386, 'timing/model': 1.3072959948331118, 'learning_rate/qwen_vl_interface': 8.351558018911423e-07, 'learning_rate/action_model': 8.351558018911423e-06, 'epoch': 2.26}) +[2026-06-18 06:16:18,226][starVLA.training.train_starvla][INFO] - Step 6171, Loss: {'train/loss': 0.32497647404670715, 'train/grad_norm_pre_clip': 25.660524368286133, 'timing/data': 0.0026066452264785767, 'timing/model': 1.068853889591992, 'learning_rate/qwen_vl_interface': 8.343582892773352e-07, 'learning_rate/action_model': 8.343582892773351e-06, 'epoch': 2.26}) +[2026-06-18 06:16:22,950][starVLA.training.train_starvla][INFO] - Step 6172, Loss: {'train/loss': 0.21604765951633453, 'train/grad_norm_pre_clip': 18.83745002746582, 'timing/data': 0.0010447166860103607, 'timing/model': 1.4564708899706602, 'learning_rate/qwen_vl_interface': 8.335616920308061e-07, 'learning_rate/action_model': 8.33561692030806e-06, 'epoch': 2.26}) +[2026-06-18 06:16:27,879][starVLA.training.train_starvla][INFO] - Step 6173, Loss: {'train/loss': 0.1492503136396408, 'train/grad_norm_pre_clip': 12.392791748046875, 'timing/data': 0.0012150956317782402, 'timing/model': 1.1905061826109886, 'learning_rate/qwen_vl_interface': 8.327660103166907e-07, 'learning_rate/action_model': 8.327660103166906e-06, 'epoch': 2.26}) +[2026-06-18 06:16:32,788][starVLA.training.train_starvla][INFO] - Step 6174, Loss: {'train/loss': 0.319541871547699, 'train/grad_norm_pre_clip': 26.259990692138672, 'timing/data': 0.0008643483743071556, 'timing/model': 1.0349638666957617, 'learning_rate/qwen_vl_interface': 8.319712442999335e-07, 'learning_rate/action_model': 8.319712442999335e-06, 'epoch': 2.26}) +[2026-06-18 06:16:37,847][starVLA.training.train_starvla][INFO] - Step 6175, Loss: {'train/loss': 0.048084307461977005, 'train/grad_norm_pre_clip': 8.681133270263672, 'timing/data': 0.0012186523526906967, 'timing/model': 1.1843579793348908, 'learning_rate/qwen_vl_interface': 8.311773941452924e-07, 'learning_rate/action_model': 8.311773941452925e-06, 'epoch': 2.26}) +[2026-06-18 06:16:43,123][starVLA.training.train_starvla][INFO] - Step 6176, Loss: {'train/loss': 0.2320351004600525, 'train/grad_norm_pre_clip': 21.81698989868164, 'timing/data': 0.0011294232681393623, 'timing/model': 1.335710871964693, 'learning_rate/qwen_vl_interface': 8.303844600173326e-07, 'learning_rate/action_model': 8.303844600173325e-06, 'epoch': 2.26}) +[2026-06-18 06:16:49,003][starVLA.training.train_starvla][INFO] - Step 6177, Loss: {'train/loss': 0.15663127601146698, 'train/grad_norm_pre_clip': 18.97468376159668, 'timing/data': 0.0007719239220023155, 'timing/model': 2.1183513198047876, 'learning_rate/qwen_vl_interface': 8.295924420804297e-07, 'learning_rate/action_model': 8.295924420804296e-06, 'epoch': 2.26}) +[2026-06-18 06:16:57,444][starVLA.training.train_starvla][INFO] - Step 6178, Loss: {'train/loss': 0.45827716588974, 'train/grad_norm_pre_clip': 44.27118682861328, 'timing/data': 0.0012505268678069115, 'timing/model': 1.0808001896366477, 'learning_rate/qwen_vl_interface': 8.288013404987691e-07, 'learning_rate/action_model': 8.28801340498769e-06, 'epoch': 2.26}) +[2026-06-18 06:17:03,889][starVLA.training.train_starvla][INFO] - Step 6179, Loss: {'train/loss': 0.06261234730482101, 'train/grad_norm_pre_clip': 10.443422317504883, 'timing/data': 0.0020744483917951584, 'timing/model': 2.6510997088626027, 'learning_rate/qwen_vl_interface': 8.28011155436349e-07, 'learning_rate/action_model': 8.28011155436349e-06, 'epoch': 2.26}) +[2026-06-18 06:17:08,794][starVLA.training.train_starvla][INFO] - Step 6180, Loss: {'train/loss': 0.3550347089767456, 'train/grad_norm_pre_clip': 25.261432647705078, 'timing/data': 0.0012627467513084412, 'timing/model': 1.0444193799048662, 'learning_rate/qwen_vl_interface': 8.272218870569748e-07, 'learning_rate/action_model': 8.272218870569748e-06, 'epoch': 2.26}) +[2026-06-18 06:17:13,471][starVLA.training.train_starvla][INFO] - Step 6181, Loss: {'train/loss': 0.19558095932006836, 'train/grad_norm_pre_clip': 23.435182571411133, 'timing/data': 0.0015478003770112991, 'timing/model': 1.189203747548163, 'learning_rate/qwen_vl_interface': 8.264335355242625e-07, 'learning_rate/action_model': 8.264335355242625e-06, 'epoch': 2.26}) +[2026-06-18 06:17:18,509][starVLA.training.train_starvla][INFO] - Step 6182, Loss: {'train/loss': 0.1397964209318161, 'train/grad_norm_pre_clip': 51.60418701171875, 'timing/data': 0.0008721537888050079, 'timing/model': 1.4680213024839759, 'learning_rate/qwen_vl_interface': 8.256461010016376e-07, 'learning_rate/action_model': 8.256461010016376e-06, 'epoch': 2.26}) +[2026-06-18 06:17:23,288][starVLA.training.train_starvla][INFO] - Step 6183, Loss: {'train/loss': 0.8782185912132263, 'train/grad_norm_pre_clip': 39.19605255126953, 'timing/data': 0.0008613374084234238, 'timing/model': 1.0637089051306248, 'learning_rate/qwen_vl_interface': 8.248595836523368e-07, 'learning_rate/action_model': 8.248595836523368e-06, 'epoch': 2.26}) +[2026-06-18 06:17:28,504][starVLA.training.train_starvla][INFO] - Step 6184, Loss: {'train/loss': 0.0657699927687645, 'train/grad_norm_pre_clip': 9.489222526550293, 'timing/data': 0.0009270887821912766, 'timing/model': 1.141050928272307, 'learning_rate/qwen_vl_interface': 8.240739836394068e-07, 'learning_rate/action_model': 8.240739836394068e-06, 'epoch': 2.26}) +[2026-06-18 06:17:33,642][starVLA.training.train_starvla][INFO] - Step 6185, Loss: {'train/loss': 0.3845633268356323, 'train/grad_norm_pre_clip': 30.120994567871094, 'timing/data': 0.0011950256302952766, 'timing/model': 1.0810550451278687, 'learning_rate/qwen_vl_interface': 8.232893011257019e-07, 'learning_rate/action_model': 8.232893011257018e-06, 'epoch': 2.26}) +[2026-06-18 06:17:38,286][starVLA.training.train_starvla][INFO] - Step 6186, Loss: {'train/loss': 0.2074204385280609, 'train/grad_norm_pre_clip': 20.577953338623047, 'timing/data': 0.0008727423846721649, 'timing/model': 1.1963648786768317, 'learning_rate/qwen_vl_interface': 8.225055362738886e-07, 'learning_rate/action_model': 8.225055362738885e-06, 'epoch': 2.27}) +[2026-06-18 06:17:44,850][starVLA.training.train_starvla][INFO] - Step 6187, Loss: {'train/loss': 0.22106456756591797, 'train/grad_norm_pre_clip': 14.793957710266113, 'timing/data': 0.001062050461769104, 'timing/model': 1.408715914003551, 'learning_rate/qwen_vl_interface': 8.217226892464411e-07, 'learning_rate/action_model': 8.21722689246441e-06, 'epoch': 2.27}) +[2026-06-18 06:17:49,290][starVLA.training.train_starvla][INFO] - Step 6188, Loss: {'train/loss': 0.19727694988250732, 'train/grad_norm_pre_clip': 24.386329650878906, 'timing/data': 0.0011412762105464935, 'timing/model': 1.04943688493222, 'learning_rate/qwen_vl_interface': 8.209407602056457e-07, 'learning_rate/action_model': 8.209407602056457e-06, 'epoch': 2.27}) +[2026-06-18 06:17:56,921][starVLA.training.train_starvla][INFO] - Step 6189, Loss: {'train/loss': 0.18862254917621613, 'train/grad_norm_pre_clip': 25.044647216796875, 'timing/data': 0.0008554989472031593, 'timing/model': 1.5209731115028262, 'learning_rate/qwen_vl_interface': 8.201597493135962e-07, 'learning_rate/action_model': 8.201597493135961e-06, 'epoch': 2.27}) +[2026-06-18 06:18:07,830][starVLA.training.train_starvla][INFO] - Step 6190, Loss: {'train/loss': 0.19817249476909637, 'train/grad_norm_pre_clip': 31.97115135192871, 'timing/data': 0.0009739277884364128, 'timing/model': 1.2416645595803857, 'learning_rate/qwen_vl_interface': 8.193796567321982e-07, 'learning_rate/action_model': 8.193796567321982e-06, 'epoch': 2.27}) +[2026-06-18 06:18:19,081][starVLA.training.train_starvla][INFO] - Step 6191, Loss: {'train/loss': 0.3038175404071808, 'train/grad_norm_pre_clip': 35.61611557006836, 'timing/data': 0.00092313252389431, 'timing/model': 1.4856627387925982, 'learning_rate/qwen_vl_interface': 8.186004826231644e-07, 'learning_rate/action_model': 8.186004826231644e-06, 'epoch': 2.27}) +[2026-06-18 06:18:29,957][starVLA.training.train_starvla][INFO] - Step 6192, Loss: {'train/loss': 0.18586644530296326, 'train/grad_norm_pre_clip': 24.52888298034668, 'timing/data': 0.012810759246349335, 'timing/model': 1.277638839557767, 'learning_rate/qwen_vl_interface': 8.178222271480192e-07, 'learning_rate/action_model': 8.178222271480191e-06, 'epoch': 2.27}) +[2026-06-18 06:18:35,227][starVLA.training.train_starvla][INFO] - Step 6193, Loss: {'train/loss': 0.4575832188129425, 'train/grad_norm_pre_clip': 31.82265853881836, 'timing/data': 0.012061039917171001, 'timing/model': 1.2743516145274043, 'learning_rate/qwen_vl_interface': 8.170448904680943e-07, 'learning_rate/action_model': 8.170448904680943e-06, 'epoch': 2.27}) +[2026-06-18 06:18:41,622][starVLA.training.train_starvla][INFO] - Step 6194, Loss: {'train/loss': 0.20702381432056427, 'train/grad_norm_pre_clip': 21.17253303527832, 'timing/data': 0.010423317551612854, 'timing/model': 1.064772816374898, 'learning_rate/qwen_vl_interface': 8.162684727445347e-07, 'learning_rate/action_model': 8.162684727445346e-06, 'epoch': 2.27}) +[2026-06-18 06:18:46,672][starVLA.training.train_starvla][INFO] - Step 6195, Loss: {'train/loss': 0.13394102454185486, 'train/grad_norm_pre_clip': 17.104944229125977, 'timing/data': 0.014634269289672375, 'timing/model': 1.1386394817382097, 'learning_rate/qwen_vl_interface': 8.154929741382922e-07, 'learning_rate/action_model': 8.154929741382922e-06, 'epoch': 2.27}) +[2026-06-18 06:18:51,846][starVLA.training.train_starvla][INFO] - Step 6196, Loss: {'train/loss': 0.20272669196128845, 'train/grad_norm_pre_clip': 21.966073989868164, 'timing/data': 0.010112207382917404, 'timing/model': 1.467154179699719, 'learning_rate/qwen_vl_interface': 8.147183948101266e-07, 'learning_rate/action_model': 8.147183948101265e-06, 'epoch': 2.27}) +[2026-06-18 06:18:57,382][starVLA.training.train_starvla][INFO] - Step 6197, Loss: {'train/loss': 0.35188019275665283, 'train/grad_norm_pre_clip': 31.63422966003418, 'timing/data': 0.027498772367835045, 'timing/model': 2.1802050871774554, 'learning_rate/qwen_vl_interface': 8.139447349206111e-07, 'learning_rate/action_model': 8.139447349206111e-06, 'epoch': 2.27}) +[2026-06-18 06:19:08,327][starVLA.training.train_starvla][INFO] - Step 6198, Loss: {'train/loss': 0.2224026322364807, 'train/grad_norm_pre_clip': 14.198236465454102, 'timing/data': 0.010584783740341663, 'timing/model': 5.946793983690441, 'learning_rate/qwen_vl_interface': 8.131719946301255e-07, 'learning_rate/action_model': 8.131719946301255e-06, 'epoch': 2.27}) +[2026-06-18 06:19:17,625][starVLA.training.train_starvla][INFO] - Step 6199, Loss: {'train/loss': 0.19372643530368805, 'train/grad_norm_pre_clip': 48.527767181396484, 'timing/data': 0.012276316061615944, 'timing/model': 1.148969671688974, 'learning_rate/qwen_vl_interface': 8.124001740988595e-07, 'learning_rate/action_model': 8.124001740988595e-06, 'epoch': 2.27}) +[2026-06-18 06:26:04,380][starVLA.training.train_starvla][INFO] - Step 6200, Loss: {'train/loss': 0.3867524564266205, 'train/grad_norm_pre_clip': 25.318395614624023, 'eval/loss': 0.5317621827125549, 'eval/latency_2/loss': 0.5317621827125549, 'eval/demon_fire_precision': 0.9543189368770764, 'eval/demon_fire_recall': 0.9615062761506277, 'eval/demon_fire_f1': 0.9578991246352647, 'eval/demon_fire_rate_model': 0.6700055648302726, 'eval/demon_fire_rate_teacher': 0.6649972175848636, 'eval/demon_left_precision': 0.9124324324324324, 'eval/demon_left_recall': 0.888421052631579, 'eval/demon_left_f1': 0.9002666666666668, 'eval/demon_left_rate_model': 0.2573734001112966, 'eval/demon_left_rate_teacher': 0.2643294379521425, 'eval/demon_right_precision': 0.9037610619469026, 'eval/demon_right_recall': 0.9231638418079096, 'eval/demon_right_f1': 0.9133594186696479, 'eval/demon_right_rate_model': 0.25153032832498606, 'eval/demon_right_rate_teacher': 0.24624373956594323, 'eval/demon_move_f1': 0.9068130426681573, 'eval/demon_cc_f1': 0.9419789693523085, 'eval/latency_2/demon_fire_precision': 0.9543189368770764, 'eval/latency_2/demon_fire_recall': 0.9615062761506277, 'eval/latency_2/demon_fire_f1': 0.9578991246352647, 'eval/latency_2/demon_fire_rate_model': 0.6700055648302726, 'eval/latency_2/demon_fire_rate_teacher': 0.6649972175848636, 'eval/latency_2/demon_left_precision': 0.9124324324324324, 'eval/latency_2/demon_left_recall': 0.888421052631579, 'eval/latency_2/demon_left_f1': 0.9002666666666668, 'eval/latency_2/demon_left_rate_model': 0.2573734001112966, 'eval/latency_2/demon_left_rate_teacher': 0.2643294379521425, 'eval/latency_2/demon_right_precision': 0.9037610619469026, 'eval/latency_2/demon_right_recall': 0.9231638418079096, 'eval/latency_2/demon_right_f1': 0.9133594186696479, 'eval/latency_2/demon_right_rate_model': 0.25153032832498606, 'eval/latency_2/demon_right_rate_teacher': 0.24624373956594323, 'eval/latency_2/demon_move_f1': 0.9068130426681573, 'eval/latency_2/demon_cc_f1': 0.9419789693523085, 'timing/data': 0.01411230955272913, 'timing/model': 1.19819073099643, 'learning_rate/qwen_vl_interface': 8.116292734868126e-07, 'learning_rate/action_model': 8.116292734868125e-06, 'epoch': 2.27}) +[2026-06-18 06:26:04,395][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_6200_state +[2026-06-18 06:26:04,402][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 06:26:13,809][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_6200_state/model.safetensors +[2026-06-18 06:26:28,558][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_6200_state/optimizer.bin +[2026-06-18 06:26:28,559][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_6200_state/sampler.bin +[2026-06-18 06:26:28,560][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_6200_state/sampler_1.bin +[2026-06-18 06:26:28,563][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_6200_state/random_states_0.pkl +[2026-06-18 06:26:28,570][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 6200 +[2026-06-18 06:26:28,570][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 06:26:28,588][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 06:26:33,609][starVLA.training.train_starvla][INFO] - Step 6201, Loss: {'train/loss': 0.20801837742328644, 'train/grad_norm_pre_clip': 21.180803298950195, 'timing/data': 0.010912717320024967, 'timing/model': 1.0328557752072811, 'learning_rate/qwen_vl_interface': 8.108592929537928e-07, 'learning_rate/action_model': 8.108592929537928e-06, 'epoch': 2.27}) +[2026-06-18 06:26:38,292][starVLA.training.train_starvla][INFO] - Step 6202, Loss: {'train/loss': 0.08633945137262344, 'train/grad_norm_pre_clip': 14.844099998474121, 'timing/data': 0.01260884664952755, 'timing/model': 1.0320703005418181, 'learning_rate/qwen_vl_interface': 8.100902326594186e-07, 'learning_rate/action_model': 8.100902326594186e-06, 'epoch': 2.27}) +[2026-06-18 06:26:43,036][starVLA.training.train_starvla][INFO] - Step 6203, Loss: {'train/loss': 0.23344235122203827, 'train/grad_norm_pre_clip': 21.182872772216797, 'timing/data': 0.015028919093310833, 'timing/model': 1.461910368874669, 'learning_rate/qwen_vl_interface': 8.093220927631171e-07, 'learning_rate/action_model': 8.09322092763117e-06, 'epoch': 2.27}) +[2026-06-18 06:26:47,115][starVLA.training.train_starvla][INFO] - Step 6204, Loss: {'train/loss': 0.24673005938529968, 'train/grad_norm_pre_clip': 20.572277069091797, 'timing/data': 0.013410712592303753, 'timing/model': 1.0449577067047358, 'learning_rate/qwen_vl_interface': 8.085548734241233e-07, 'learning_rate/action_model': 8.085548734241232e-06, 'epoch': 2.27}) +[2026-06-18 06:26:51,480][starVLA.training.train_starvla][INFO] - Step 6205, Loss: {'train/loss': 0.3584366738796234, 'train/grad_norm_pre_clip': 27.893648147583008, 'timing/data': 0.011210820637643337, 'timing/model': 1.0363432373851538, 'learning_rate/qwen_vl_interface': 8.077885748014843e-07, 'learning_rate/action_model': 8.077885748014841e-06, 'epoch': 2.27}) +[2026-06-18 06:26:56,297][starVLA.training.train_starvla][INFO] - Step 6206, Loss: {'train/loss': 0.24041900038719177, 'train/grad_norm_pre_clip': 19.33336067199707, 'timing/data': 0.015049324370920658, 'timing/model': 1.2168919499963522, 'learning_rate/qwen_vl_interface': 8.070231970540532e-07, 'learning_rate/action_model': 8.070231970540532e-06, 'epoch': 2.27}) +[2026-06-18 06:27:00,900][starVLA.training.train_starvla][INFO] - Step 6207, Loss: {'train/loss': 0.3372887074947357, 'train/grad_norm_pre_clip': 22.833301544189453, 'timing/data': 0.011255428194999695, 'timing/model': 1.1193222170695662, 'learning_rate/qwen_vl_interface': 8.062587403404934e-07, 'learning_rate/action_model': 8.062587403404934e-06, 'epoch': 2.27}) +[2026-06-18 06:27:05,446][starVLA.training.train_starvla][INFO] - Step 6208, Loss: {'train/loss': 0.1375303417444229, 'train/grad_norm_pre_clip': 18.804771423339844, 'timing/data': 0.013600388541817665, 'timing/model': 1.5096393879503012, 'learning_rate/qwen_vl_interface': 8.054952048192793e-07, 'learning_rate/action_model': 8.054952048192794e-06, 'epoch': 2.27}) +[2026-06-18 06:27:09,990][starVLA.training.train_starvla][INFO] - Step 6209, Loss: {'train/loss': 0.19869564473628998, 'train/grad_norm_pre_clip': 21.60734748840332, 'timing/data': 0.01097158994525671, 'timing/model': 1.0885563706979156, 'learning_rate/qwen_vl_interface': 8.047325906486926e-07, 'learning_rate/action_model': 8.047325906486925e-06, 'epoch': 2.27}) +[2026-06-18 06:27:14,924][starVLA.training.train_starvla][INFO] - Step 6210, Loss: {'train/loss': 0.25427159667015076, 'train/grad_norm_pre_clip': 25.876792907714844, 'timing/data': 0.018293902277946472, 'timing/model': 1.2065665572881699, 'learning_rate/qwen_vl_interface': 8.039708979868229e-07, 'learning_rate/action_model': 8.039708979868227e-06, 'epoch': 2.27}) +[2026-06-18 06:27:22,053][starVLA.training.train_starvla][INFO] - Step 6211, Loss: {'train/loss': 0.050676263868808746, 'train/grad_norm_pre_clip': 7.67872953414917, 'timing/data': 0.011249789036810398, 'timing/model': 1.2122678635641932, 'learning_rate/qwen_vl_interface': 8.032101269915687e-07, 'learning_rate/action_model': 8.032101269915687e-06, 'epoch': 2.27}) +[2026-06-18 06:27:31,463][starVLA.training.train_starvla][INFO] - Step 6212, Loss: {'train/loss': 0.05804206803441048, 'train/grad_norm_pre_clip': 7.23072624206543, 'timing/data': 0.0243909927085042, 'timing/model': 4.630652994848788, 'learning_rate/qwen_vl_interface': 8.024502778206411e-07, 'learning_rate/action_model': 8.02450277820641e-06, 'epoch': 2.27}) +[2026-06-18 06:27:40,739][starVLA.training.train_starvla][INFO] - Step 6213, Loss: {'train/loss': 0.42088809609413147, 'train/grad_norm_pre_clip': 28.801559448242188, 'timing/data': 0.011419483460485935, 'timing/model': 2.283707207068801, 'learning_rate/qwen_vl_interface': 8.016913506315566e-07, 'learning_rate/action_model': 8.016913506315565e-06, 'epoch': 2.28}) +[2026-06-18 06:27:49,089][starVLA.training.train_starvla][INFO] - Step 6214, Loss: {'train/loss': 0.45630744099617004, 'train/grad_norm_pre_clip': 23.847408294677734, 'timing/data': 0.024389775469899178, 'timing/model': 1.3651034170761704, 'learning_rate/qwen_vl_interface': 8.009333455816416e-07, 'learning_rate/action_model': 8.009333455816417e-06, 'epoch': 2.28}) +[2026-06-18 06:27:54,545][starVLA.training.train_starvla][INFO] - Step 6215, Loss: {'train/loss': 0.30603542923927307, 'train/grad_norm_pre_clip': 33.37309265136719, 'timing/data': 0.015845504589378834, 'timing/model': 1.160109425894916, 'learning_rate/qwen_vl_interface': 8.001762628280318e-07, 'learning_rate/action_model': 8.001762628280319e-06, 'epoch': 2.28}) +[2026-06-18 06:28:02,326][starVLA.training.train_starvla][INFO] - Step 6216, Loss: {'train/loss': 0.23372896015644073, 'train/grad_norm_pre_clip': 22.928180694580078, 'timing/data': 0.02620429452508688, 'timing/model': 2.187634589150548, 'learning_rate/qwen_vl_interface': 7.994201025276708e-07, 'learning_rate/action_model': 7.994201025276708e-06, 'epoch': 2.28}) +[2026-06-18 06:28:11,463][starVLA.training.train_starvla][INFO] - Step 6217, Loss: {'train/loss': 0.3011860251426697, 'train/grad_norm_pre_clip': 25.076078414916992, 'timing/data': 0.0015151109546422958, 'timing/model': 1.2354816105216742, 'learning_rate/qwen_vl_interface': 7.986648648373115e-07, 'learning_rate/action_model': 7.986648648373115e-06, 'epoch': 2.28}) +[2026-06-18 06:28:20,941][starVLA.training.train_starvla][INFO] - Step 6218, Loss: {'train/loss': 0.20603598654270172, 'train/grad_norm_pre_clip': 22.74677848815918, 'timing/data': 0.0010892488062381744, 'timing/model': 1.795498832128942, 'learning_rate/qwen_vl_interface': 7.979105499135157e-07, 'learning_rate/action_model': 7.979105499135155e-06, 'epoch': 2.28}) +[2026-06-18 06:28:30,206][starVLA.training.train_starvla][INFO] - Step 6219, Loss: {'train/loss': 0.297835111618042, 'train/grad_norm_pre_clip': 22.081865310668945, 'timing/data': 0.011435551568865776, 'timing/model': 4.809149172157049, 'learning_rate/qwen_vl_interface': 7.971571579126533e-07, 'learning_rate/action_model': 7.971571579126534e-06, 'epoch': 2.28}) +[2026-06-18 06:28:40,361][starVLA.training.train_starvla][INFO] - Step 6220, Loss: {'train/loss': 0.41997259855270386, 'train/grad_norm_pre_clip': 24.08281707763672, 'timing/data': 0.012396208010613918, 'timing/model': 1.239185349084437, 'learning_rate/qwen_vl_interface': 7.964046889909038e-07, 'learning_rate/action_model': 7.964046889909038e-06, 'epoch': 2.28}) +[2026-06-18 06:28:50,216][starVLA.training.train_starvla][INFO] - Step 6221, Loss: {'train/loss': 0.2035052478313446, 'train/grad_norm_pre_clip': 23.49686050415039, 'timing/data': 0.013028821907937527, 'timing/model': 1.3133357390761375, 'learning_rate/qwen_vl_interface': 7.956531433042543e-07, 'learning_rate/action_model': 7.956531433042542e-06, 'epoch': 2.28}) +[2026-06-18 06:28:54,761][starVLA.training.train_starvla][INFO] - Step 6222, Loss: {'train/loss': 0.16593188047409058, 'train/grad_norm_pre_clip': 15.092734336853027, 'timing/data': 0.01414206437766552, 'timing/model': 1.1273341365158558, 'learning_rate/qwen_vl_interface': 7.94902521008501e-07, 'learning_rate/action_model': 7.94902521008501e-06, 'epoch': 2.28}) +[2026-06-18 06:28:59,653][starVLA.training.train_starvla][INFO] - Step 6223, Loss: {'train/loss': 0.26925379037857056, 'train/grad_norm_pre_clip': 23.58985137939453, 'timing/data': 0.016268680803477764, 'timing/model': 1.4240073813125491, 'learning_rate/qwen_vl_interface': 7.941528222592501e-07, 'learning_rate/action_model': 7.941528222592501e-06, 'epoch': 2.28}) +[2026-06-18 06:29:04,066][starVLA.training.train_starvla][INFO] - Step 6224, Loss: {'train/loss': 0.10858496278524399, 'train/grad_norm_pre_clip': 13.901899337768555, 'timing/data': 0.013012566603720188, 'timing/model': 1.168387508019805, 'learning_rate/qwen_vl_interface': 7.934040472119132e-07, 'learning_rate/action_model': 7.934040472119133e-06, 'epoch': 2.28}) +[2026-06-18 06:29:08,842][starVLA.training.train_starvla][INFO] - Step 6225, Loss: {'train/loss': 0.4185228645801544, 'train/grad_norm_pre_clip': 29.986007690429688, 'timing/data': 0.016557546332478523, 'timing/model': 1.0826830565929413, 'learning_rate/qwen_vl_interface': 7.926561960217127e-07, 'learning_rate/action_model': 7.926561960217126e-06, 'epoch': 2.28}) +[2026-06-18 06:29:13,502][starVLA.training.train_starvla][INFO] - Step 6226, Loss: {'train/loss': 0.09526363760232925, 'train/grad_norm_pre_clip': 42.112648010253906, 'timing/data': 0.011069493368268013, 'timing/model': 1.0356894759461284, 'learning_rate/qwen_vl_interface': 7.919092688436783e-07, 'learning_rate/action_model': 7.919092688436783e-06, 'epoch': 2.28}) +[2026-06-18 06:29:18,390][starVLA.training.train_starvla][INFO] - Step 6227, Loss: {'train/loss': 0.16909164190292358, 'train/grad_norm_pre_clip': 18.137165069580078, 'timing/data': 0.013585856184363365, 'timing/model': 1.0460379989817739, 'learning_rate/qwen_vl_interface': 7.911632658326505e-07, 'learning_rate/action_model': 7.911632658326505e-06, 'epoch': 2.28}) +[2026-06-18 06:29:23,788][starVLA.training.train_starvla][INFO] - Step 6228, Loss: {'train/loss': 0.09393393248319626, 'train/grad_norm_pre_clip': 14.68455696105957, 'timing/data': 0.011422611773014069, 'timing/model': 1.4913899693638086, 'learning_rate/qwen_vl_interface': 7.904181871432757e-07, 'learning_rate/action_model': 7.904181871432757e-06, 'epoch': 2.28}) +[2026-06-18 06:29:33,708][starVLA.training.train_starvla][INFO] - Step 6229, Loss: {'train/loss': 0.35339170694351196, 'train/grad_norm_pre_clip': 30.025419235229492, 'timing/data': 0.01699643023312092, 'timing/model': 1.1541561000049114, 'learning_rate/qwen_vl_interface': 7.896740329300091e-07, 'learning_rate/action_model': 7.89674032930009e-06, 'epoch': 2.28}) +[2026-06-18 06:29:38,661][starVLA.training.train_starvla][INFO] - Step 6230, Loss: {'train/loss': 0.5598035454750061, 'train/grad_norm_pre_clip': 36.742713928222656, 'timing/data': 0.012273602187633514, 'timing/model': 1.1426701080054045, 'learning_rate/qwen_vl_interface': 7.889308033471156e-07, 'learning_rate/action_model': 7.889308033471155e-06, 'epoch': 2.28}) +[2026-06-18 06:29:43,108][starVLA.training.train_starvla][INFO] - Step 6231, Loss: {'train/loss': 0.2401013970375061, 'train/grad_norm_pre_clip': 17.9118709564209, 'timing/data': 0.010600348003208637, 'timing/model': 1.059663413092494, 'learning_rate/qwen_vl_interface': 7.881884985486662e-07, 'learning_rate/action_model': 7.881884985486662e-06, 'epoch': 2.28}) +[2026-06-18 06:29:47,980][starVLA.training.train_starvla][INFO] - Step 6232, Loss: {'train/loss': 0.1795508712530136, 'train/grad_norm_pre_clip': 16.414749145507812, 'timing/data': 0.012591127306222916, 'timing/model': 1.143990045413375, 'learning_rate/qwen_vl_interface': 7.874471186885427e-07, 'learning_rate/action_model': 7.874471186885427e-06, 'epoch': 2.28}) +[2026-06-18 06:29:52,818][starVLA.training.train_starvla][INFO] - Step 6233, Loss: {'train/loss': 0.17389965057373047, 'train/grad_norm_pre_clip': 13.771693229675293, 'timing/data': 0.019526749849319458, 'timing/model': 1.4940835116431117, 'learning_rate/qwen_vl_interface': 7.867066639204326e-07, 'learning_rate/action_model': 7.867066639204327e-06, 'epoch': 2.28}) +[2026-06-18 06:29:57,240][starVLA.training.train_starvla][INFO] - Step 6234, Loss: {'train/loss': 0.13931222259998322, 'train/grad_norm_pre_clip': 21.926122665405273, 'timing/data': 0.010783021338284016, 'timing/model': 1.2469655144959688, 'learning_rate/qwen_vl_interface': 7.859671343978348e-07, 'learning_rate/action_model': 7.859671343978348e-06, 'epoch': 2.28}) +[2026-06-18 06:30:04,815][starVLA.training.train_starvla][INFO] - Step 6235, Loss: {'train/loss': 0.15691334009170532, 'train/grad_norm_pre_clip': 18.858068466186523, 'timing/data': 0.04382998123764992, 'timing/model': 4.101795211434364, 'learning_rate/qwen_vl_interface': 7.852285302740528e-07, 'learning_rate/action_model': 7.852285302740527e-06, 'epoch': 2.28}) +[2026-06-18 06:30:12,946][starVLA.training.train_starvla][INFO] - Step 6236, Loss: {'train/loss': 0.5574695467948914, 'train/grad_norm_pre_clip': 53.068199157714844, 'timing/data': 0.010994955897331238, 'timing/model': 1.0775819038972259, 'learning_rate/qwen_vl_interface': 7.844908517022005e-07, 'learning_rate/action_model': 7.844908517022005e-06, 'epoch': 2.28}) +[2026-06-18 06:30:18,098][starVLA.training.train_starvla][INFO] - Step 6237, Loss: {'train/loss': 0.47102802991867065, 'train/grad_norm_pre_clip': 34.32554244995117, 'timing/data': 0.0125071220099926, 'timing/model': 1.1497331531718373, 'learning_rate/qwen_vl_interface': 7.837540988352001e-07, 'learning_rate/action_model': 7.837540988352e-06, 'epoch': 2.28}) +[2026-06-18 06:30:30,471][starVLA.training.train_starvla][INFO] - Step 6238, Loss: {'train/loss': 0.1476999819278717, 'train/grad_norm_pre_clip': 17.630388259887695, 'timing/data': 0.014720886945724487, 'timing/model': 4.63911631796509, 'learning_rate/qwen_vl_interface': 7.83018271825782e-07, 'learning_rate/action_model': 7.83018271825782e-06, 'epoch': 2.28}) +[2026-06-18 06:30:35,410][starVLA.training.train_starvla][INFO] - Step 6239, Loss: {'train/loss': 0.191511332988739, 'train/grad_norm_pre_clip': 17.332778930664062, 'timing/data': 0.011416235938668251, 'timing/model': 1.240671779960394, 'learning_rate/qwen_vl_interface': 7.822833708264821e-07, 'learning_rate/action_model': 7.822833708264821e-06, 'epoch': 2.28}) +[2026-06-18 06:30:40,350][starVLA.training.train_starvla][INFO] - Step 6240, Loss: {'train/loss': 0.10945770889520645, 'train/grad_norm_pre_clip': 26.75865936279297, 'timing/data': 0.01124880276620388, 'timing/model': 1.2757188137620687, 'learning_rate/qwen_vl_interface': 7.81549395989646e-07, 'learning_rate/action_model': 7.81549395989646e-06, 'epoch': 2.28}) +[2026-06-18 06:30:44,701][starVLA.training.train_starvla][INFO] - Step 6241, Loss: {'train/loss': 0.2818308472633362, 'train/grad_norm_pre_clip': 19.136993408203125, 'timing/data': 0.01185159757733345, 'timing/model': 1.0485616913065314, 'learning_rate/qwen_vl_interface': 7.808163474674291e-07, 'learning_rate/action_model': 7.80816347467429e-06, 'epoch': 2.29}) +[2026-06-18 06:30:49,336][starVLA.training.train_starvla][INFO] - Step 6242, Loss: {'train/loss': 0.16777403652668, 'train/grad_norm_pre_clip': 23.907421112060547, 'timing/data': 0.011953615583479404, 'timing/model': 1.3958488795906305, 'learning_rate/qwen_vl_interface': 7.800842254117922e-07, 'learning_rate/action_model': 7.80084225411792e-06, 'epoch': 2.29}) +[2026-06-18 06:30:53,939][starVLA.training.train_starvla][INFO] - Step 6243, Loss: {'train/loss': 0.20781853795051575, 'train/grad_norm_pre_clip': 32.30975341796875, 'timing/data': 0.012568690814077854, 'timing/model': 1.4749241257086396, 'learning_rate/qwen_vl_interface': 7.793530299745061e-07, 'learning_rate/action_model': 7.793530299745061e-06, 'epoch': 2.29}) +[2026-06-18 06:30:58,531][starVLA.training.train_starvla][INFO] - Step 6244, Loss: {'train/loss': 0.1942712366580963, 'train/grad_norm_pre_clip': 17.63900375366211, 'timing/data': 0.014494603499770164, 'timing/model': 1.266477950848639, 'learning_rate/qwen_vl_interface': 7.786227613071464e-07, 'learning_rate/action_model': 7.786227613071464e-06, 'epoch': 2.29}) +[2026-06-18 06:31:03,314][starVLA.training.train_starvla][INFO] - Step 6245, Loss: {'train/loss': 0.18637724220752716, 'train/grad_norm_pre_clip': 24.12932586669922, 'timing/data': 0.010802055709064007, 'timing/model': 1.1274836659431458, 'learning_rate/qwen_vl_interface': 7.778934195610998e-07, 'learning_rate/action_model': 7.778934195610998e-06, 'epoch': 2.29}) +[2026-06-18 06:31:08,103][starVLA.training.train_starvla][INFO] - Step 6246, Loss: {'train/loss': 0.16890603303909302, 'train/grad_norm_pre_clip': 18.5352725982666, 'timing/data': 0.016959759406745434, 'timing/model': 1.305067516863346, 'learning_rate/qwen_vl_interface': 7.771650048875598e-07, 'learning_rate/action_model': 7.771650048875597e-06, 'epoch': 2.29}) +[2026-06-18 06:31:14,447][starVLA.training.train_starvla][INFO] - Step 6247, Loss: {'train/loss': 0.32598382234573364, 'train/grad_norm_pre_clip': 20.82418441772461, 'timing/data': 0.013429536484181881, 'timing/model': 1.093323027715087, 'learning_rate/qwen_vl_interface': 7.764375174375276e-07, 'learning_rate/action_model': 7.764375174375276e-06, 'epoch': 2.29}) +[2026-06-18 06:31:19,320][starVLA.training.train_starvla][INFO] - Step 6248, Loss: {'train/loss': 0.2570050060749054, 'train/grad_norm_pre_clip': 19.700021743774414, 'timing/data': 0.0012834817171096802, 'timing/model': 1.4661562349647284, 'learning_rate/qwen_vl_interface': 7.757109573618116e-07, 'learning_rate/action_model': 7.757109573618115e-06, 'epoch': 2.29}) +[2026-06-18 06:31:23,674][starVLA.training.train_starvla][INFO] - Step 6249, Loss: {'train/loss': 0.23751482367515564, 'train/grad_norm_pre_clip': 20.826736450195312, 'timing/data': 0.0009782733395695686, 'timing/model': 1.034524280577898, 'learning_rate/qwen_vl_interface': 7.749853248110287e-07, 'learning_rate/action_model': 7.749853248110287e-06, 'epoch': 2.29}) +[2026-06-18 06:31:28,380][starVLA.training.train_starvla][INFO] - Step 6250, Loss: {'train/loss': 0.11172584444284439, 'train/grad_norm_pre_clip': 18.923357009887695, 'timing/data': 0.001058950088918209, 'timing/model': 1.032226295210421, 'learning_rate/qwen_vl_interface': 7.742606199356032e-07, 'learning_rate/action_model': 7.74260619935603e-06, 'epoch': 2.29}) +[2026-06-18 06:31:35,104][starVLA.training.train_starvla][INFO] - Step 6251, Loss: {'train/loss': 0.24626412987709045, 'train/grad_norm_pre_clip': 22.9166316986084, 'timing/data': 0.0011724857613444328, 'timing/model': 3.0600064657628536, 'learning_rate/qwen_vl_interface': 7.735368428857665e-07, 'learning_rate/action_model': 7.735368428857665e-06, 'epoch': 2.29}) +[2026-06-18 06:31:42,915][starVLA.training.train_starvla][INFO] - Step 6252, Loss: {'train/loss': 0.3574289083480835, 'train/grad_norm_pre_clip': 19.88987922668457, 'timing/data': 0.000891086645424366, 'timing/model': 1.0443765092641115, 'learning_rate/qwen_vl_interface': 7.728139938115605e-07, 'learning_rate/action_model': 7.728139938115604e-06, 'epoch': 2.29}) +[2026-06-18 06:31:47,614][starVLA.training.train_starvla][INFO] - Step 6253, Loss: {'train/loss': 0.4149057865142822, 'train/grad_norm_pre_clip': 27.756298065185547, 'timing/data': 0.0020009269937872887, 'timing/model': 1.5112091666087508, 'learning_rate/qwen_vl_interface': 7.720920728628306e-07, 'learning_rate/action_model': 7.720920728628306e-06, 'epoch': 2.29}) +[2026-06-18 06:31:52,213][starVLA.training.train_starvla][INFO] - Step 6254, Loss: {'train/loss': 0.35553452372550964, 'train/grad_norm_pre_clip': 32.47403335571289, 'timing/data': 0.004536482505500317, 'timing/model': 1.0842329245060682, 'learning_rate/qwen_vl_interface': 7.713710801892327e-07, 'learning_rate/action_model': 7.713710801892326e-06, 'epoch': 2.29}) +[2026-06-18 06:32:01,694][starVLA.training.train_starvla][INFO] - Step 6255, Loss: {'train/loss': 0.37614843249320984, 'train/grad_norm_pre_clip': 24.381591796875, 'timing/data': 0.0014779279008507729, 'timing/model': 1.0419348506256938, 'learning_rate/qwen_vl_interface': 7.70651015940228e-07, 'learning_rate/action_model': 7.70651015940228e-06, 'epoch': 2.29}) +[2026-06-18 06:32:06,343][starVLA.training.train_starvla][INFO] - Step 6256, Loss: {'train/loss': 0.2583567202091217, 'train/grad_norm_pre_clip': 15.42729377746582, 'timing/data': 0.0009414767846465111, 'timing/model': 1.1453522918745875, 'learning_rate/qwen_vl_interface': 7.699318802650891e-07, 'learning_rate/action_model': 7.69931880265089e-06, 'epoch': 2.29}) +[2026-06-18 06:32:10,945][starVLA.training.train_starvla][INFO] - Step 6257, Loss: {'train/loss': 0.2069689780473709, 'train/grad_norm_pre_clip': 22.080698013305664, 'timing/data': 0.0010463651269674301, 'timing/model': 1.3577910894528031, 'learning_rate/qwen_vl_interface': 7.692136733128918e-07, 'learning_rate/action_model': 7.692136733128917e-06, 'epoch': 2.29}) +[2026-06-18 06:32:15,409][starVLA.training.train_starvla][INFO] - Step 6258, Loss: {'train/loss': 0.16393738985061646, 'train/grad_norm_pre_clip': 17.38339614868164, 'timing/data': 0.0012101326137781143, 'timing/model': 1.1907449988648295, 'learning_rate/qwen_vl_interface': 7.684963952325207e-07, 'learning_rate/action_model': 7.684963952325206e-06, 'epoch': 2.29}) +[2026-06-18 06:32:34,545][starVLA.training.train_starvla][INFO] - Step 6259, Loss: {'train/loss': 0.10502325743436813, 'train/grad_norm_pre_clip': 15.135159492492676, 'timing/data': 0.0016978541389107704, 'timing/model': 5.306799019686878, 'learning_rate/qwen_vl_interface': 7.6778004617267e-07, 'learning_rate/action_model': 7.6778004617267e-06, 'epoch': 2.29}) +[2026-06-18 06:32:39,671][starVLA.training.train_starvla][INFO] - Step 6260, Loss: {'train/loss': 0.1885778307914734, 'train/grad_norm_pre_clip': 19.391611099243164, 'timing/data': 0.0012673120945692062, 'timing/model': 1.0489216232672334, 'learning_rate/qwen_vl_interface': 7.670646262818387e-07, 'learning_rate/action_model': 7.670646262818387e-06, 'epoch': 2.29}) +[2026-06-18 06:32:44,074][starVLA.training.train_starvla][INFO] - Step 6261, Loss: {'train/loss': 0.3599133789539337, 'train/grad_norm_pre_clip': 26.350521087646484, 'timing/data': 0.000977807678282261, 'timing/model': 1.0213137054815888, 'learning_rate/qwen_vl_interface': 7.663501357083342e-07, 'learning_rate/action_model': 7.66350135708334e-06, 'epoch': 2.29}) +[2026-06-18 06:32:48,819][starVLA.training.train_starvla][INFO] - Step 6262, Loss: {'train/loss': 0.16611775755882263, 'train/grad_norm_pre_clip': 15.62065601348877, 'timing/data': 0.0011050719767808914, 'timing/model': 1.5899674780666828, 'learning_rate/qwen_vl_interface': 7.656365746002715e-07, 'learning_rate/action_model': 7.656365746002716e-06, 'epoch': 2.29}) +[2026-06-18 06:32:52,935][starVLA.training.train_starvla][INFO] - Step 6263, Loss: {'train/loss': 0.07059551775455475, 'train/grad_norm_pre_clip': 11.466421127319336, 'timing/data': 0.0010408982634544373, 'timing/model': 1.2220565704628825, 'learning_rate/qwen_vl_interface': 7.649239431055718e-07, 'learning_rate/action_model': 7.649239431055717e-06, 'epoch': 2.29}) +[2026-06-18 06:32:58,346][starVLA.training.train_starvla][INFO] - Step 6264, Loss: {'train/loss': 0.361198365688324, 'train/grad_norm_pre_clip': 23.248655319213867, 'timing/data': 0.0010227598249912262, 'timing/model': 1.0451340163126588, 'learning_rate/qwen_vl_interface': 7.642122413719652e-07, 'learning_rate/action_model': 7.642122413719651e-06, 'epoch': 2.29}) +[2026-06-18 06:33:02,571][starVLA.training.train_starvla][INFO] - Step 6265, Loss: {'train/loss': 0.09911900758743286, 'train/grad_norm_pre_clip': 12.945804595947266, 'timing/data': 0.0013073142617940903, 'timing/model': 1.0350282527506351, 'learning_rate/qwen_vl_interface': 7.635014695469877e-07, 'learning_rate/action_model': 7.635014695469878e-06, 'epoch': 2.29}) +[2026-06-18 06:33:06,806][starVLA.training.train_starvla][INFO] - Step 6266, Loss: {'train/loss': 0.41417190432548523, 'train/grad_norm_pre_clip': 30.363039016723633, 'timing/data': 0.0009004678577184677, 'timing/model': 1.0024420991539955, 'learning_rate/qwen_vl_interface': 7.627916277779832e-07, 'learning_rate/action_model': 7.627916277779831e-06, 'epoch': 2.29}) +[2026-06-18 06:33:11,072][starVLA.training.train_starvla][INFO] - Step 6267, Loss: {'train/loss': 0.17054180800914764, 'train/grad_norm_pre_clip': 18.338600158691406, 'timing/data': 0.0010942574590444565, 'timing/model': 1.0099501078948379, 'learning_rate/qwen_vl_interface': 7.62082716212103e-07, 'learning_rate/action_model': 7.62082716212103e-06, 'epoch': 2.29}) +[2026-06-18 06:33:15,114][starVLA.training.train_starvla][INFO] - Step 6268, Loss: {'train/loss': 0.05829105153679848, 'train/grad_norm_pre_clip': 8.226649284362793, 'timing/data': 0.0008308421820402145, 'timing/model': 1.0178763614967465, 'learning_rate/qwen_vl_interface': 7.613747349963048e-07, 'learning_rate/action_model': 7.613747349963048e-06, 'epoch': 2.3}) +[2026-06-18 06:33:19,341][starVLA.training.train_starvla][INFO] - Step 6269, Loss: {'train/loss': 0.26283591985702515, 'train/grad_norm_pre_clip': 21.01516342163086, 'timing/data': 0.0010547377169132233, 'timing/model': 1.047829832881689, 'learning_rate/qwen_vl_interface': 7.606676842773531e-07, 'learning_rate/action_model': 7.60667684277353e-06, 'epoch': 2.3}) +[2026-06-18 06:33:24,484][starVLA.training.train_starvla][INFO] - Step 6270, Loss: {'train/loss': 0.132804274559021, 'train/grad_norm_pre_clip': 16.545059204101562, 'timing/data': 0.0010900404304265976, 'timing/model': 1.0666157752275467, 'learning_rate/qwen_vl_interface': 7.599615642018218e-07, 'learning_rate/action_model': 7.599615642018218e-06, 'epoch': 2.3}) +[2026-06-18 06:33:28,636][starVLA.training.train_starvla][INFO] - Step 6271, Loss: {'train/loss': 0.24405606091022491, 'train/grad_norm_pre_clip': 21.413766860961914, 'timing/data': 0.0009065093472599983, 'timing/model': 1.0208872556686401, 'learning_rate/qwen_vl_interface': 7.592563749160908e-07, 'learning_rate/action_model': 7.592563749160908e-06, 'epoch': 2.3}) +[2026-06-18 06:33:32,822][starVLA.training.train_starvla][INFO] - Step 6272, Loss: {'train/loss': 0.5664753913879395, 'train/grad_norm_pre_clip': 25.809236526489258, 'timing/data': 0.0015228884294629097, 'timing/model': 1.016454097814858, 'learning_rate/qwen_vl_interface': 7.585521165663445e-07, 'learning_rate/action_model': 7.585521165663445e-06, 'epoch': 2.3}) +[2026-06-18 06:33:37,471][starVLA.training.train_starvla][INFO] - Step 6273, Loss: {'train/loss': 0.15144839882850647, 'train/grad_norm_pre_clip': 22.745746612548828, 'timing/data': 0.0010910574346780777, 'timing/model': 1.6427167290821671, 'learning_rate/qwen_vl_interface': 7.578487892985769e-07, 'learning_rate/action_model': 7.578487892985769e-06, 'epoch': 2.3}) +[2026-06-18 06:33:45,318][starVLA.training.train_starvla][INFO] - Step 6274, Loss: {'train/loss': 0.3190018832683563, 'train/grad_norm_pre_clip': 24.47049903869629, 'timing/data': 0.0012034624814987183, 'timing/model': 1.1641893945634365, 'learning_rate/qwen_vl_interface': 7.571463932585901e-07, 'learning_rate/action_model': 7.571463932585901e-06, 'epoch': 2.3}) +[2026-06-18 06:33:50,109][starVLA.training.train_starvla][INFO] - Step 6275, Loss: {'train/loss': 0.10996586084365845, 'train/grad_norm_pre_clip': 16.75368309020996, 'timing/data': 0.0011328086256980896, 'timing/model': 1.0849497681483626, 'learning_rate/qwen_vl_interface': 7.564449285919897e-07, 'learning_rate/action_model': 7.564449285919897e-06, 'epoch': 2.3}) +[2026-06-18 06:33:54,652][starVLA.training.train_starvla][INFO] - Step 6276, Loss: {'train/loss': 0.27827849984169006, 'train/grad_norm_pre_clip': 22.11615562438965, 'timing/data': 0.0016122879460453987, 'timing/model': 1.0618148567155004, 'learning_rate/qwen_vl_interface': 7.557443954441924e-07, 'learning_rate/action_model': 7.557443954441923e-06, 'epoch': 2.3}) +[2026-06-18 06:34:03,303][starVLA.training.train_starvla][INFO] - Step 6277, Loss: {'train/loss': 0.26361826062202454, 'train/grad_norm_pre_clip': 29.111291885375977, 'timing/data': 0.001012805849313736, 'timing/model': 4.1465041702613235, 'learning_rate/qwen_vl_interface': 7.550447939604161e-07, 'learning_rate/action_model': 7.550447939604161e-06, 'epoch': 2.3}) +[2026-06-18 06:34:08,047][starVLA.training.train_starvla][INFO] - Step 6278, Loss: {'train/loss': 0.24200254678726196, 'train/grad_norm_pre_clip': 28.091854095458984, 'timing/data': 0.0009873425588011742, 'timing/model': 1.4263111520558596, 'learning_rate/qwen_vl_interface': 7.543461242856916e-07, 'learning_rate/action_model': 7.543461242856916e-06, 'epoch': 2.3}) +[2026-06-18 06:34:15,759][starVLA.training.train_starvla][INFO] - Step 6279, Loss: {'train/loss': 0.23092086613178253, 'train/grad_norm_pre_clip': 25.85159683227539, 'timing/data': 0.00113692507147789, 'timing/model': 4.196080030873418, 'learning_rate/qwen_vl_interface': 7.536483865648531e-07, 'learning_rate/action_model': 7.536483865648531e-06, 'epoch': 2.3}) +[2026-06-18 06:34:21,387][starVLA.training.train_starvla][INFO] - Step 6280, Loss: {'train/loss': 0.1891954243183136, 'train/grad_norm_pre_clip': 22.057880401611328, 'timing/data': 0.0008882787078619003, 'timing/model': 1.2009957851842046, 'learning_rate/qwen_vl_interface': 7.52951580942542e-07, 'learning_rate/action_model': 7.5295158094254195e-06, 'epoch': 2.3}) +[2026-06-18 06:34:25,900][starVLA.training.train_starvla][INFO] - Step 6281, Loss: {'train/loss': 0.2748721241950989, 'train/grad_norm_pre_clip': 25.695871353149414, 'timing/data': 0.001283041201531887, 'timing/model': 1.0935233607888222, 'learning_rate/qwen_vl_interface': 7.522557075632078e-07, 'learning_rate/action_model': 7.522557075632078e-06, 'epoch': 2.3}) +[2026-06-18 06:34:36,214][starVLA.training.train_starvla][INFO] - Step 6282, Loss: {'train/loss': 0.2116832584142685, 'train/grad_norm_pre_clip': 17.546228408813477, 'timing/data': 0.0014754803851246834, 'timing/model': 3.154987281188369, 'learning_rate/qwen_vl_interface': 7.515607665711051e-07, 'learning_rate/action_model': 7.515607665711051e-06, 'epoch': 2.3}) +[2026-06-18 06:34:41,839][starVLA.training.train_starvla][INFO] - Step 6283, Loss: {'train/loss': 0.45612815022468567, 'train/grad_norm_pre_clip': 31.30549430847168, 'timing/data': 0.0011077383533120155, 'timing/model': 1.0383459711447358, 'learning_rate/qwen_vl_interface': 7.508667581102959e-07, 'learning_rate/action_model': 7.508667581102958e-06, 'epoch': 2.3}) +[2026-06-18 06:34:46,318][starVLA.training.train_starvla][INFO] - Step 6284, Loss: {'train/loss': 0.12081944942474365, 'train/grad_norm_pre_clip': 18.613378524780273, 'timing/data': 0.000891372561454773, 'timing/model': 1.0593613293021917, 'learning_rate/qwen_vl_interface': 7.50173682324648e-07, 'learning_rate/action_model': 7.501736823246481e-06, 'epoch': 2.3}) +[2026-06-18 06:34:50,888][starVLA.training.train_starvla][INFO] - Step 6285, Loss: {'train/loss': 0.03399890288710594, 'train/grad_norm_pre_clip': 5.020908832550049, 'timing/data': 0.00166363175958395, 'timing/model': 1.1189052667468786, 'learning_rate/qwen_vl_interface': 7.494815393578394e-07, 'learning_rate/action_model': 7.494815393578393e-06, 'epoch': 2.3}) +[2026-06-18 06:34:55,756][starVLA.training.train_starvla][INFO] - Step 6286, Loss: {'train/loss': 0.42315423488616943, 'train/grad_norm_pre_clip': 28.831636428833008, 'timing/data': 0.001312139444053173, 'timing/model': 1.159461627714336, 'learning_rate/qwen_vl_interface': 7.487903293533501e-07, 'learning_rate/action_model': 7.4879032935335e-06, 'epoch': 2.3}) +[2026-06-18 06:35:00,260][starVLA.training.train_starvla][INFO] - Step 6287, Loss: {'train/loss': 0.22265276312828064, 'train/grad_norm_pre_clip': 16.805755615234375, 'timing/data': 0.0011933241039514542, 'timing/model': 1.4177919048815966, 'learning_rate/qwen_vl_interface': 7.481000524544686e-07, 'learning_rate/action_model': 7.481000524544686e-06, 'epoch': 2.3}) +[2026-06-18 06:35:04,510][starVLA.training.train_starvla][INFO] - Step 6288, Loss: {'train/loss': 0.22670748829841614, 'train/grad_norm_pre_clip': 21.07657241821289, 'timing/data': 0.0007852204144001007, 'timing/model': 1.2024692697450519, 'learning_rate/qwen_vl_interface': 7.474107088042898e-07, 'learning_rate/action_model': 7.474107088042898e-06, 'epoch': 2.3}) +[2026-06-18 06:35:09,445][starVLA.training.train_starvla][INFO] - Step 6289, Loss: {'train/loss': 0.17738573253154755, 'train/grad_norm_pre_clip': 21.598072052001953, 'timing/data': 0.0017024967819452286, 'timing/model': 1.0975211430341005, 'learning_rate/qwen_vl_interface': 7.467222985457167e-07, 'learning_rate/action_model': 7.4672229854571675e-06, 'epoch': 2.3}) +[2026-06-18 06:35:14,421][starVLA.training.train_starvla][INFO] - Step 6290, Loss: {'train/loss': 0.4793570041656494, 'train/grad_norm_pre_clip': 44.650054931640625, 'timing/data': 0.0010569188743829727, 'timing/model': 1.048533613793552, 'learning_rate/qwen_vl_interface': 7.46034821821458e-07, 'learning_rate/action_model': 7.46034821821458e-06, 'epoch': 2.3}) +[2026-06-18 06:35:19,196][starVLA.training.train_starvla][INFO] - Step 6291, Loss: {'train/loss': 0.25860562920570374, 'train/grad_norm_pre_clip': 22.52033042907715, 'timing/data': 0.0009318478405475616, 'timing/model': 1.0538639156147838, 'learning_rate/qwen_vl_interface': 7.453482787740255e-07, 'learning_rate/action_model': 7.4534827877402545e-06, 'epoch': 2.3}) +[2026-06-18 06:35:24,101][starVLA.training.train_starvla][INFO] - Step 6292, Loss: {'train/loss': 0.153517484664917, 'train/grad_norm_pre_clip': 24.684329986572266, 'timing/data': 0.0008882982656359673, 'timing/model': 1.4062160747125745, 'learning_rate/qwen_vl_interface': 7.446626695457427e-07, 'learning_rate/action_model': 7.446626695457427e-06, 'epoch': 2.3}) +[2026-06-18 06:35:36,250][starVLA.training.train_starvla][INFO] - Step 6293, Loss: {'train/loss': 0.19986934959888458, 'train/grad_norm_pre_clip': 16.049549102783203, 'timing/data': 0.002283194102346897, 'timing/model': 1.6797166354954243, 'learning_rate/qwen_vl_interface': 7.439779942787367e-07, 'learning_rate/action_model': 7.439779942787367e-06, 'epoch': 2.3}) +[2026-06-18 06:35:40,663][starVLA.training.train_starvla][INFO] - Step 6294, Loss: {'train/loss': 0.46090295910835266, 'train/grad_norm_pre_clip': 26.388700485229492, 'timing/data': 0.002241663634777069, 'timing/model': 1.1056481655687094, 'learning_rate/qwen_vl_interface': 7.432942531149409e-07, 'learning_rate/action_model': 7.432942531149409e-06, 'epoch': 2.3}) +[2026-06-18 06:35:49,520][starVLA.training.train_starvla][INFO] - Step 6295, Loss: {'train/loss': 0.4070976674556732, 'train/grad_norm_pre_clip': 25.198686599731445, 'timing/data': 0.0012183943763375282, 'timing/model': 5.340783250518143, 'learning_rate/qwen_vl_interface': 7.42611446196096e-07, 'learning_rate/action_model': 7.426114461960959e-06, 'epoch': 2.31}) +[2026-06-18 06:35:58,228][starVLA.training.train_starvla][INFO] - Step 6296, Loss: {'train/loss': 0.1172410175204277, 'train/grad_norm_pre_clip': 13.660321235656738, 'timing/data': 0.0010814480483531952, 'timing/model': 4.664087704382837, 'learning_rate/qwen_vl_interface': 7.419295736637483e-07, 'learning_rate/action_model': 7.419295736637483e-06, 'epoch': 2.31}) +[2026-06-18 06:36:04,700][starVLA.training.train_starvla][INFO] - Step 6297, Loss: {'train/loss': 0.11829739063978195, 'train/grad_norm_pre_clip': 15.144177436828613, 'timing/data': 0.002699289470911026, 'timing/model': 2.566143387928605, 'learning_rate/qwen_vl_interface': 7.412486356592512e-07, 'learning_rate/action_model': 7.412486356592511e-06, 'epoch': 2.31}) +[2026-06-18 06:36:09,370][starVLA.training.train_starvla][INFO] - Step 6298, Loss: {'train/loss': 0.21217618882656097, 'train/grad_norm_pre_clip': 19.31071662902832, 'timing/data': 0.0009972164407372475, 'timing/model': 1.4269201569259167, 'learning_rate/qwen_vl_interface': 7.405686323237623e-07, 'learning_rate/action_model': 7.405686323237623e-06, 'epoch': 2.31}) +[2026-06-18 06:36:13,781][starVLA.training.train_starvla][INFO] - Step 6299, Loss: {'train/loss': 0.1893928050994873, 'train/grad_norm_pre_clip': 17.43444061279297, 'timing/data': 0.0022038957104086876, 'timing/model': 1.055129243992269, 'learning_rate/qwen_vl_interface': 7.398895637982504e-07, 'learning_rate/action_model': 7.398895637982503e-06, 'epoch': 2.31}) +[2026-06-18 06:41:01,093][starVLA.training.train_starvla][INFO] - Step 6300, Loss: {'train/loss': 0.3168521821498871, 'train/grad_norm_pre_clip': 33.63425064086914, 'eval/loss': 0.5349156856536865, 'eval/latency_2/loss': 0.5349156856536865, 'eval/demon_fire_precision': 0.9607760438633488, 'eval/demon_fire_recall': 0.9531380753138076, 'eval/demon_fire_f1': 0.9569418189455998, 'eval/demon_fire_rate_model': 0.6597106288258208, 'eval/demon_fire_rate_teacher': 0.6649972175848636, 'eval/demon_left_precision': 0.8974093264248705, 'eval/demon_left_recall': 0.911578947368421, 'eval/demon_left_f1': 0.9044386422976501, 'eval/demon_left_rate_model': 0.26850306065664997, 'eval/demon_left_rate_teacher': 0.2643294379521425, 'eval/demon_right_precision': 0.9064587973273942, 'eval/demon_right_recall': 0.919774011299435, 'eval/demon_right_f1': 0.9130678631519911, 'eval/demon_right_rate_model': 0.2498608792431831, 'eval/demon_right_rate_teacher': 0.24624373956594323, 'eval/demon_move_f1': 0.9087532527248205, 'eval/demon_cc_f1': 0.9419570387526405, 'eval/latency_2/demon_fire_precision': 0.9607760438633488, 'eval/latency_2/demon_fire_recall': 0.9531380753138076, 'eval/latency_2/demon_fire_f1': 0.9569418189455998, 'eval/latency_2/demon_fire_rate_model': 0.6597106288258208, 'eval/latency_2/demon_fire_rate_teacher': 0.6649972175848636, 'eval/latency_2/demon_left_precision': 0.8974093264248705, 'eval/latency_2/demon_left_recall': 0.911578947368421, 'eval/latency_2/demon_left_f1': 0.9044386422976501, 'eval/latency_2/demon_left_rate_model': 0.26850306065664997, 'eval/latency_2/demon_left_rate_teacher': 0.2643294379521425, 'eval/latency_2/demon_right_precision': 0.9064587973273942, 'eval/latency_2/demon_right_recall': 0.919774011299435, 'eval/latency_2/demon_right_f1': 0.9130678631519911, 'eval/latency_2/demon_right_rate_model': 0.2498608792431831, 'eval/latency_2/demon_right_rate_teacher': 0.24624373956594323, 'eval/latency_2/demon_move_f1': 0.9087532527248205, 'eval/latency_2/demon_cc_f1': 0.9419570387526405, 'timing/data': 0.0010711755603551865, 'timing/model': 1.2088901996612549, 'learning_rate/qwen_vl_interface': 7.392114302234838e-07, 'learning_rate/action_model': 7.3921143022348385e-06, 'epoch': 2.31}) +[2026-06-18 06:41:01,116][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_6300_state +[2026-06-18 06:41:01,123][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 06:41:11,042][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_6300_state/model.safetensors +[2026-06-18 06:41:29,651][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_6300_state/optimizer.bin +[2026-06-18 06:41:29,652][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_6300_state/sampler.bin +[2026-06-18 06:41:29,653][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_6300_state/sampler_1.bin +[2026-06-18 06:41:29,655][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_6300_state/random_states_0.pkl +[2026-06-18 06:41:29,661][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 6300 +[2026-06-18 06:41:29,661][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 06:41:29,678][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 06:41:44,560][starVLA.training.train_starvla][INFO] - Step 6301, Loss: {'train/loss': 0.10986577719449997, 'train/grad_norm_pre_clip': 10.267680168151855, 'timing/data': 0.0010605743154883385, 'timing/model': 1.2089153183624148, 'learning_rate/qwen_vl_interface': 7.385342317400418e-07, 'learning_rate/action_model': 7.385342317400417e-06, 'epoch': 2.31}) +[2026-06-18 06:41:57,784][starVLA.training.train_starvla][INFO] - Step 6302, Loss: {'train/loss': 0.42979180812835693, 'train/grad_norm_pre_clip': 31.57355499267578, 'timing/data': 0.0009318804368376732, 'timing/model': 1.2268235087394714, 'learning_rate/qwen_vl_interface': 7.378579684883069e-07, 'learning_rate/action_model': 7.378579684883069e-06, 'epoch': 2.31}) +[2026-06-18 06:42:11,272][starVLA.training.train_starvla][INFO] - Step 6303, Loss: {'train/loss': 0.31228312849998474, 'train/grad_norm_pre_clip': 23.770227432250977, 'timing/data': 0.0012868987396359444, 'timing/model': 1.2562174182385206, 'learning_rate/qwen_vl_interface': 7.371826406084714e-07, 'learning_rate/action_model': 7.371826406084714e-06, 'epoch': 2.31}) +[2026-06-18 06:42:17,428][starVLA.training.train_starvla][INFO] - Step 6304, Loss: {'train/loss': 0.1989758461713791, 'train/grad_norm_pre_clip': 21.52044105529785, 'timing/data': 0.001329159364104271, 'timing/model': 1.7180436477065086, 'learning_rate/qwen_vl_interface': 7.365082482405315e-07, 'learning_rate/action_model': 7.365082482405315e-06, 'epoch': 2.31}) +[2026-06-18 06:42:34,888][starVLA.training.train_starvla][INFO] - Step 6305, Loss: {'train/loss': 0.31610724329948425, 'train/grad_norm_pre_clip': 24.964168548583984, 'timing/data': 0.0008614864200353622, 'timing/model': 5.031749790534377, 'learning_rate/qwen_vl_interface': 7.358347915242877e-07, 'learning_rate/action_model': 7.358347915242876e-06, 'epoch': 2.31}) +[2026-06-18 06:42:48,722][starVLA.training.train_starvla][INFO] - Step 6306, Loss: {'train/loss': 0.4441946744918823, 'train/grad_norm_pre_clip': 34.378456115722656, 'timing/data': 0.000795646570622921, 'timing/model': 2.3120828103274107, 'learning_rate/qwen_vl_interface': 7.351622705993481e-07, 'learning_rate/action_model': 7.351622705993481e-06, 'epoch': 2.31}) +[2026-06-18 06:42:58,026][starVLA.training.train_starvla][INFO] - Step 6307, Loss: {'train/loss': 0.45845362544059753, 'train/grad_norm_pre_clip': 39.52523422241211, 'timing/data': 0.0008740173652768135, 'timing/model': 1.341929603368044, 'learning_rate/qwen_vl_interface': 7.344906856051288e-07, 'learning_rate/action_model': 7.344906856051288e-06, 'epoch': 2.31}) +[2026-06-18 06:43:02,383][starVLA.training.train_starvla][INFO] - Step 6308, Loss: {'train/loss': 0.06291569024324417, 'train/grad_norm_pre_clip': 10.521650314331055, 'timing/data': 0.0013163331896066666, 'timing/model': 1.3577300161123276, 'learning_rate/qwen_vl_interface': 7.338200366808493e-07, 'learning_rate/action_model': 7.338200366808492e-06, 'epoch': 2.31}) +[2026-06-18 06:43:07,122][starVLA.training.train_starvla][INFO] - Step 6309, Loss: {'train/loss': 0.3132537305355072, 'train/grad_norm_pre_clip': 29.794979095458984, 'timing/data': 0.0009119892492890358, 'timing/model': 1.3918182086199522, 'learning_rate/qwen_vl_interface': 7.331503239655361e-07, 'learning_rate/action_model': 7.33150323965536e-06, 'epoch': 2.31}) +[2026-06-18 06:43:11,592][starVLA.training.train_starvla][INFO] - Step 6310, Loss: {'train/loss': 0.3919437527656555, 'train/grad_norm_pre_clip': 30.94061279296875, 'timing/data': 0.0008966317400336266, 'timing/model': 1.0451571410521865, 'learning_rate/qwen_vl_interface': 7.324815475980207e-07, 'learning_rate/action_model': 7.324815475980207e-06, 'epoch': 2.31}) +[2026-06-18 06:43:16,090][starVLA.training.train_starvla][INFO] - Step 6311, Loss: {'train/loss': 0.37672969698905945, 'train/grad_norm_pre_clip': 27.64519691467285, 'timing/data': 0.0012037698179483414, 'timing/model': 1.0401951121166348, 'learning_rate/qwen_vl_interface': 7.318137077169416e-07, 'learning_rate/action_model': 7.318137077169415e-06, 'epoch': 2.31}) +[2026-06-18 06:43:21,130][starVLA.training.train_starvla][INFO] - Step 6312, Loss: {'train/loss': 0.19912755489349365, 'train/grad_norm_pre_clip': 13.328981399536133, 'timing/data': 0.0013514403253793716, 'timing/model': 1.2333140382543206, 'learning_rate/qwen_vl_interface': 7.311468044607425e-07, 'learning_rate/action_model': 7.311468044607425e-06, 'epoch': 2.31}) +[2026-06-18 06:43:25,509][starVLA.training.train_starvla][INFO] - Step 6313, Loss: {'train/loss': 0.16696509718894958, 'train/grad_norm_pre_clip': 21.43939971923828, 'timing/data': 0.0009746523573994637, 'timing/model': 1.0346933845430613, 'learning_rate/qwen_vl_interface': 7.304808379676739e-07, 'learning_rate/action_model': 7.3048083796767394e-06, 'epoch': 2.31}) +[2026-06-18 06:43:29,930][starVLA.training.train_starvla][INFO] - Step 6314, Loss: {'train/loss': 0.19554001092910767, 'train/grad_norm_pre_clip': 21.240217208862305, 'timing/data': 0.0010443730279803276, 'timing/model': 1.3721093507483602, 'learning_rate/qwen_vl_interface': 7.298158083757903e-07, 'learning_rate/action_model': 7.298158083757903e-06, 'epoch': 2.31}) +[2026-06-18 06:43:34,156][starVLA.training.train_starvla][INFO] - Step 6315, Loss: {'train/loss': 0.18602782487869263, 'train/grad_norm_pre_clip': 30.436567306518555, 'timing/data': 0.0009091496467590332, 'timing/model': 1.0376389287412167, 'learning_rate/qwen_vl_interface': 7.291517158229536e-07, 'learning_rate/action_model': 7.291517158229535e-06, 'epoch': 2.31}) +[2026-06-18 06:43:39,167][starVLA.training.train_starvla][INFO] - Step 6316, Loss: {'train/loss': 0.16843774914741516, 'train/grad_norm_pre_clip': 15.837775230407715, 'timing/data': 0.0007656710222363472, 'timing/model': 1.0643802750855684, 'learning_rate/qwen_vl_interface': 7.284885604468312e-07, 'learning_rate/action_model': 7.284885604468311e-06, 'epoch': 2.31}) +[2026-06-18 06:43:43,882][starVLA.training.train_starvla][INFO] - Step 6317, Loss: {'train/loss': 0.4365523159503937, 'train/grad_norm_pre_clip': 32.60578536987305, 'timing/data': 0.009083041921257973, 'timing/model': 1.0672054719179869, 'learning_rate/qwen_vl_interface': 7.278263423848948e-07, 'learning_rate/action_model': 7.278263423848948e-06, 'epoch': 2.31}) +[2026-06-18 06:43:48,548][starVLA.training.train_starvla][INFO] - Step 6318, Loss: {'train/loss': 0.21800567209720612, 'train/grad_norm_pre_clip': 21.98223114013672, 'timing/data': 0.0008399691432714462, 'timing/model': 1.4210764188319445, 'learning_rate/qwen_vl_interface': 7.271650617744239e-07, 'learning_rate/action_model': 7.271650617744238e-06, 'epoch': 2.31}) +[2026-06-18 06:43:53,081][starVLA.training.train_starvla][INFO] - Step 6319, Loss: {'train/loss': 0.07958546280860901, 'train/grad_norm_pre_clip': 9.762993812561035, 'timing/data': 0.001111694611608982, 'timing/model': 1.3592953458428383, 'learning_rate/qwen_vl_interface': 7.265047187525033e-07, 'learning_rate/action_model': 7.265047187525033e-06, 'epoch': 2.31}) +[2026-06-18 06:43:57,174][starVLA.training.train_starvla][INFO] - Step 6320, Loss: {'train/loss': 0.21651768684387207, 'train/grad_norm_pre_clip': 22.41629409790039, 'timing/data': 0.001324441283941269, 'timing/model': 1.0346623789519072, 'learning_rate/qwen_vl_interface': 7.258453134560212e-07, 'learning_rate/action_model': 7.258453134560211e-06, 'epoch': 2.31}) +[2026-06-18 06:44:01,723][starVLA.training.train_starvla][INFO] - Step 6321, Loss: {'train/loss': 0.1810203641653061, 'train/grad_norm_pre_clip': 13.26941204071045, 'timing/data': 0.0008494546636939049, 'timing/model': 1.094546870328486, 'learning_rate/qwen_vl_interface': 7.251868460216743e-07, 'learning_rate/action_model': 7.251868460216743e-06, 'epoch': 2.31}) +[2026-06-18 06:44:06,722][starVLA.training.train_starvla][INFO] - Step 6322, Loss: {'train/loss': 0.20922207832336426, 'train/grad_norm_pre_clip': 35.15936279296875, 'timing/data': 0.001305924728512764, 'timing/model': 1.0757407760247588, 'learning_rate/qwen_vl_interface': 7.245293165859629e-07, 'learning_rate/action_model': 7.245293165859629e-06, 'epoch': 2.31}) +[2026-06-18 06:44:14,907][starVLA.training.train_starvla][INFO] - Step 6323, Loss: {'train/loss': 0.0766192153096199, 'train/grad_norm_pre_clip': 11.060315132141113, 'timing/data': 0.0013425061479210854, 'timing/model': 1.2123885219916701, 'learning_rate/qwen_vl_interface': 7.238727252851935e-07, 'learning_rate/action_model': 7.238727252851935e-06, 'epoch': 2.32}) +[2026-06-18 06:44:25,190][starVLA.training.train_starvla][INFO] - Step 6324, Loss: {'train/loss': 0.1505233645439148, 'train/grad_norm_pre_clip': 20.447107315063477, 'timing/data': 0.0014382721856236458, 'timing/model': 1.0542082758620381, 'learning_rate/qwen_vl_interface': 7.23217072255479e-07, 'learning_rate/action_model': 7.23217072255479e-06, 'epoch': 2.32}) +[2026-06-18 06:44:45,210][starVLA.training.train_starvla][INFO] - Step 6325, Loss: {'train/loss': 0.13790056109428406, 'train/grad_norm_pre_clip': 15.897660255432129, 'timing/data': 0.0012998394668102264, 'timing/model': 4.652889867313206, 'learning_rate/qwen_vl_interface': 7.225623576327361e-07, 'learning_rate/action_model': 7.22562357632736e-06, 'epoch': 2.32}) +[2026-06-18 06:44:53,629][starVLA.training.train_starvla][INFO] - Step 6326, Loss: {'train/loss': 0.30862829089164734, 'train/grad_norm_pre_clip': 27.03233528137207, 'timing/data': 0.0011050058528780937, 'timing/model': 1.0833992585539818, 'learning_rate/qwen_vl_interface': 7.219085815526887e-07, 'learning_rate/action_model': 7.219085815526887e-06, 'epoch': 2.32}) +[2026-06-18 06:45:00,806][starVLA.training.train_starvla][INFO] - Step 6327, Loss: {'train/loss': 0.2448977380990982, 'train/grad_norm_pre_clip': 18.58155632019043, 'timing/data': 0.0007358519360423088, 'timing/model': 2.097128354012966, 'learning_rate/qwen_vl_interface': 7.212557441508641e-07, 'learning_rate/action_model': 7.212557441508641e-06, 'epoch': 2.32}) +[2026-06-18 06:45:14,697][starVLA.training.train_starvla][INFO] - Step 6328, Loss: {'train/loss': 0.1731114685535431, 'train/grad_norm_pre_clip': 14.17028522491455, 'timing/data': 0.001182032749056816, 'timing/model': 1.0479298923164606, 'learning_rate/qwen_vl_interface': 7.206038455625972e-07, 'learning_rate/action_model': 7.206038455625971e-06, 'epoch': 2.32}) +[2026-06-18 06:45:23,185][starVLA.training.train_starvla][INFO] - Step 6329, Loss: {'train/loss': 0.4348284900188446, 'train/grad_norm_pre_clip': 46.18724060058594, 'timing/data': 0.001206420361995697, 'timing/model': 1.6211505364626646, 'learning_rate/qwen_vl_interface': 7.199528859230271e-07, 'learning_rate/action_model': 7.199528859230272e-06, 'epoch': 2.32}) +[2026-06-18 06:45:33,170][starVLA.training.train_starvla][INFO] - Step 6330, Loss: {'train/loss': 0.14663679897785187, 'train/grad_norm_pre_clip': 17.36003875732422, 'timing/data': 0.0013410365208983421, 'timing/model': 4.41076179407537, 'learning_rate/qwen_vl_interface': 7.19302865367098e-07, 'learning_rate/action_model': 7.19302865367098e-06, 'epoch': 2.32}) +[2026-06-18 06:45:46,147][starVLA.training.train_starvla][INFO] - Step 6331, Loss: {'train/loss': 0.08132526278495789, 'train/grad_norm_pre_clip': 8.836655616760254, 'timing/data': 0.0008814139291644096, 'timing/model': 2.2946963123977184, 'learning_rate/qwen_vl_interface': 7.186537840295599e-07, 'learning_rate/action_model': 7.186537840295598e-06, 'epoch': 2.32}) +[2026-06-18 06:45:53,466][starVLA.training.train_starvla][INFO] - Step 6332, Loss: {'train/loss': 0.296383261680603, 'train/grad_norm_pre_clip': 18.469282150268555, 'timing/data': 0.014607477001845837, 'timing/model': 2.3327374355867505, 'learning_rate/qwen_vl_interface': 7.18005642044969e-07, 'learning_rate/action_model': 7.180056420449689e-06, 'epoch': 2.32}) +[2026-06-18 06:46:00,208][starVLA.training.train_starvla][INFO] - Step 6333, Loss: {'train/loss': 0.6012031435966492, 'train/grad_norm_pre_clip': 37.405452728271484, 'timing/data': 0.014006654731929302, 'timing/model': 1.0543237486854196, 'learning_rate/qwen_vl_interface': 7.173584395476852e-07, 'learning_rate/action_model': 7.173584395476851e-06, 'epoch': 2.32}) +[2026-06-18 06:46:04,615][starVLA.training.train_starvla][INFO] - Step 6334, Loss: {'train/loss': 0.10163579136133194, 'train/grad_norm_pre_clip': 20.10960578918457, 'timing/data': 0.01187259890139103, 'timing/model': 1.3724030647426844, 'learning_rate/qwen_vl_interface': 7.16712176671874e-07, 'learning_rate/action_model': 7.16712176671874e-06, 'epoch': 2.32}) +[2026-06-18 06:46:08,681][starVLA.training.train_starvla][INFO] - Step 6335, Loss: {'train/loss': 0.3654891550540924, 'train/grad_norm_pre_clip': 43.289493560791016, 'timing/data': 0.013150501996278763, 'timing/model': 1.058588683605194, 'learning_rate/qwen_vl_interface': 7.160668535515059e-07, 'learning_rate/action_model': 7.1606685355150585e-06, 'epoch': 2.32}) +[2026-06-18 06:46:13,598][starVLA.training.train_starvla][INFO] - Step 6336, Loss: {'train/loss': 0.30347779393196106, 'train/grad_norm_pre_clip': 27.604045867919922, 'timing/data': 0.014970259740948677, 'timing/model': 1.0431513795629144, 'learning_rate/qwen_vl_interface': 7.154224703203584e-07, 'learning_rate/action_model': 7.154224703203584e-06, 'epoch': 2.32}) +[2026-06-18 06:46:18,148][starVLA.training.train_starvla][INFO] - Step 6337, Loss: {'train/loss': 0.26836124062538147, 'train/grad_norm_pre_clip': 22.541259765625, 'timing/data': 0.010683858767151833, 'timing/model': 1.0284361178055406, 'learning_rate/qwen_vl_interface': 7.147790271120122e-07, 'learning_rate/action_model': 7.147790271120122e-06, 'epoch': 2.32}) +[2026-06-18 06:46:22,882][starVLA.training.train_starvla][INFO] - Step 6338, Loss: {'train/loss': 0.26134026050567627, 'train/grad_norm_pre_clip': 22.518346786499023, 'timing/data': 0.018383663147687912, 'timing/model': 1.0531256077811122, 'learning_rate/qwen_vl_interface': 7.141365240598545e-07, 'learning_rate/action_model': 7.141365240598545e-06, 'epoch': 2.32}) +[2026-06-18 06:46:27,836][starVLA.training.train_starvla][INFO] - Step 6339, Loss: {'train/loss': 0.26982811093330383, 'train/grad_norm_pre_clip': 29.71287727355957, 'timing/data': 0.01258779689669609, 'timing/model': 1.6402489598840475, 'learning_rate/qwen_vl_interface': 7.134949612970749e-07, 'learning_rate/action_model': 7.134949612970748e-06, 'epoch': 2.32}) +[2026-06-18 06:46:32,335][starVLA.training.train_starvla][INFO] - Step 6340, Loss: {'train/loss': 0.34544891119003296, 'train/grad_norm_pre_clip': 31.020383834838867, 'timing/data': 0.023339408449828625, 'timing/model': 1.1018600184470415, 'learning_rate/qwen_vl_interface': 7.128543389566718e-07, 'learning_rate/action_model': 7.128543389566718e-06, 'epoch': 2.32}) +[2026-06-18 06:46:37,442][starVLA.training.train_starvla][INFO] - Step 6341, Loss: {'train/loss': 0.3022386431694031, 'train/grad_norm_pre_clip': 26.510122299194336, 'timing/data': 0.012866235338151455, 'timing/model': 1.4174008769914508, 'learning_rate/qwen_vl_interface': 7.122146571714464e-07, 'learning_rate/action_model': 7.122146571714463e-06, 'epoch': 2.32}) +[2026-06-18 06:46:42,634][starVLA.training.train_starvla][INFO] - Step 6342, Loss: {'train/loss': 0.24669267237186432, 'train/grad_norm_pre_clip': 17.110246658325195, 'timing/data': 0.01867341808974743, 'timing/model': 1.2591780368238688, 'learning_rate/qwen_vl_interface': 7.115759160740058e-07, 'learning_rate/action_model': 7.115759160740057e-06, 'epoch': 2.32}) +[2026-06-18 06:46:47,396][starVLA.training.train_starvla][INFO] - Step 6343, Loss: {'train/loss': 0.4585734009742737, 'train/grad_norm_pre_clip': 30.17438507080078, 'timing/data': 0.012690236791968346, 'timing/model': 1.040393185801804, 'learning_rate/qwen_vl_interface': 7.109381157967607e-07, 'learning_rate/action_model': 7.109381157967607e-06, 'epoch': 2.32}) +[2026-06-18 06:46:57,642][starVLA.training.train_starvla][INFO] - Step 6344, Loss: {'train/loss': 0.22123728692531586, 'train/grad_norm_pre_clip': 16.80243492126465, 'timing/data': 0.013157357461750507, 'timing/model': 3.8985297987237573, 'learning_rate/qwen_vl_interface': 7.103012564719289e-07, 'learning_rate/action_model': 7.103012564719288e-06, 'epoch': 2.32}) +[2026-06-18 06:47:09,978][starVLA.training.train_starvla][INFO] - Step 6345, Loss: {'train/loss': 0.4472738802433014, 'train/grad_norm_pre_clip': 42.244197845458984, 'timing/data': 0.012095725163817406, 'timing/model': 1.222662060521543, 'learning_rate/qwen_vl_interface': 7.096653382315315e-07, 'learning_rate/action_model': 7.096653382315315e-06, 'epoch': 2.32}) +[2026-06-18 06:47:15,047][starVLA.training.train_starvla][INFO] - Step 6346, Loss: {'train/loss': 0.31892621517181396, 'train/grad_norm_pre_clip': 25.06880760192871, 'timing/data': 0.011994938366115093, 'timing/model': 1.0952862538397312, 'learning_rate/qwen_vl_interface': 7.090303612073946e-07, 'learning_rate/action_model': 7.090303612073946e-06, 'epoch': 2.32}) +[2026-06-18 06:47:24,802][starVLA.training.train_starvla][INFO] - Step 6347, Loss: {'train/loss': 0.14599360525608063, 'train/grad_norm_pre_clip': 17.78007698059082, 'timing/data': 0.02006168942898512, 'timing/model': 1.8437679065391421, 'learning_rate/qwen_vl_interface': 7.083963255311508e-07, 'learning_rate/action_model': 7.083963255311509e-06, 'epoch': 2.32}) +[2026-06-18 06:47:32,078][starVLA.training.train_starvla][INFO] - Step 6348, Loss: {'train/loss': 0.054454416036605835, 'train/grad_norm_pre_clip': 11.375225067138672, 'timing/data': 0.01496865414083004, 'timing/model': 2.267406632192433, 'learning_rate/qwen_vl_interface': 7.07763231334236e-07, 'learning_rate/action_model': 7.077632313342359e-06, 'epoch': 2.32}) +[2026-06-18 06:47:36,871][starVLA.training.train_starvla][INFO] - Step 6349, Loss: {'train/loss': 0.3219851851463318, 'train/grad_norm_pre_clip': 36.12587356567383, 'timing/data': 0.013735690154135227, 'timing/model': 1.337850202806294, 'learning_rate/qwen_vl_interface': 7.071310787478902e-07, 'learning_rate/action_model': 7.071310787478902e-06, 'epoch': 2.32}) +[2026-06-18 06:47:41,661][starVLA.training.train_starvla][INFO] - Step 6350, Loss: {'train/loss': 0.1456780582666397, 'train/grad_norm_pre_clip': 19.84098243713379, 'timing/data': 0.01156493742018938, 'timing/model': 1.2573475474491715, 'learning_rate/qwen_vl_interface': 7.0649986790316e-07, 'learning_rate/action_model': 7.0649986790315995e-06, 'epoch': 2.33}) +[2026-06-18 06:47:47,086][starVLA.training.train_starvla][INFO] - Step 6351, Loss: {'train/loss': 0.21922723948955536, 'train/grad_norm_pre_clip': 30.128385543823242, 'timing/data': 0.016509228385984898, 'timing/model': 1.066257786937058, 'learning_rate/qwen_vl_interface': 7.058695989308965e-07, 'learning_rate/action_model': 7.058695989308965e-06, 'epoch': 2.33}) +[2026-06-18 06:47:56,645][starVLA.training.train_starvla][INFO] - Step 6352, Loss: {'train/loss': 0.38504546880722046, 'train/grad_norm_pre_clip': 33.853946685791016, 'timing/data': 0.013060417026281357, 'timing/model': 5.872236181050539, 'learning_rate/qwen_vl_interface': 7.052402719617554e-07, 'learning_rate/action_model': 7.052402719617554e-06, 'epoch': 2.33}) +[2026-06-18 06:48:01,386][starVLA.training.train_starvla][INFO] - Step 6353, Loss: {'train/loss': 0.24072794616222382, 'train/grad_norm_pre_clip': 23.293567657470703, 'timing/data': 0.015148043632507324, 'timing/model': 1.0520571898669004, 'learning_rate/qwen_vl_interface': 7.046118871261952e-07, 'learning_rate/action_model': 7.0461188712619514e-06, 'epoch': 2.33}) +[2026-06-18 06:48:06,871][starVLA.training.train_starvla][INFO] - Step 6354, Loss: {'train/loss': 0.13970693945884705, 'train/grad_norm_pre_clip': 18.541902542114258, 'timing/data': 0.0644105477258563, 'timing/model': 1.9552657473832369, 'learning_rate/qwen_vl_interface': 7.039844445544828e-07, 'learning_rate/action_model': 7.039844445544827e-06, 'epoch': 2.33}) +[2026-06-18 06:48:11,879][starVLA.training.train_starvla][INFO] - Step 6355, Loss: {'train/loss': 0.18276193737983704, 'train/grad_norm_pre_clip': 24.91896629333496, 'timing/data': 0.0108222346752882, 'timing/model': 1.045809332281351, 'learning_rate/qwen_vl_interface': 7.033579443766864e-07, 'learning_rate/action_model': 7.033579443766864e-06, 'epoch': 2.33}) +[2026-06-18 06:48:20,032][starVLA.training.train_starvla][INFO] - Step 6356, Loss: {'train/loss': 0.11813036352396011, 'train/grad_norm_pre_clip': 22.861900329589844, 'timing/data': 0.012682987377047539, 'timing/model': 4.479040036909282, 'learning_rate/qwen_vl_interface': 7.027323867226807e-07, 'learning_rate/action_model': 7.0273238672268065e-06, 'epoch': 2.33}) +[2026-06-18 06:48:30,414][starVLA.training.train_starvla][INFO] - Step 6357, Loss: {'train/loss': 0.33381760120391846, 'train/grad_norm_pre_clip': 17.69514274597168, 'timing/data': 0.017604827880859375, 'timing/model': 1.4091811180114746, 'learning_rate/qwen_vl_interface': 7.021077717221441e-07, 'learning_rate/action_model': 7.0210777172214415e-06, 'epoch': 2.33}) +[2026-06-18 06:48:35,284][starVLA.training.train_starvla][INFO] - Step 6358, Loss: {'train/loss': 0.18655362725257874, 'train/grad_norm_pre_clip': 21.475107192993164, 'timing/data': 0.024127701297402382, 'timing/model': 1.053180648945272, 'learning_rate/qwen_vl_interface': 7.014840995045602e-07, 'learning_rate/action_model': 7.014840995045602e-06, 'epoch': 2.33}) +[2026-06-18 06:48:40,324][starVLA.training.train_starvla][INFO] - Step 6359, Loss: {'train/loss': 0.30132898688316345, 'train/grad_norm_pre_clip': 15.69924545288086, 'timing/data': 0.012012806721031666, 'timing/model': 1.4794943667948246, 'learning_rate/qwen_vl_interface': 7.008613701992172e-07, 'learning_rate/action_model': 7.008613701992171e-06, 'epoch': 2.33}) +[2026-06-18 06:48:44,634][starVLA.training.train_starvla][INFO] - Step 6360, Loss: {'train/loss': 0.7197523713111877, 'train/grad_norm_pre_clip': 31.691884994506836, 'timing/data': 0.013192052952945232, 'timing/model': 1.170592070557177, 'learning_rate/qwen_vl_interface': 7.002395839352063e-07, 'learning_rate/action_model': 7.002395839352063e-06, 'epoch': 2.33}) +[2026-06-18 06:48:49,449][starVLA.training.train_starvla][INFO] - Step 6361, Loss: {'train/loss': 0.29964280128479004, 'train/grad_norm_pre_clip': 23.456647872924805, 'timing/data': 0.013510177843272686, 'timing/model': 1.274537768214941, 'learning_rate/qwen_vl_interface': 6.996187408414268e-07, 'learning_rate/action_model': 6.996187408414268e-06, 'epoch': 2.33}) +[2026-06-18 06:48:56,682][starVLA.training.train_starvla][INFO] - Step 6362, Loss: {'train/loss': 0.4710150957107544, 'train/grad_norm_pre_clip': 28.294530868530273, 'timing/data': 0.012321204878389835, 'timing/model': 1.2041876753792167, 'learning_rate/qwen_vl_interface': 6.98998841046578e-07, 'learning_rate/action_model': 6.989988410465779e-06, 'epoch': 2.33}) +[2026-06-18 06:49:01,664][starVLA.training.train_starvla][INFO] - Step 6363, Loss: {'train/loss': 0.48874104022979736, 'train/grad_norm_pre_clip': 32.24651336669922, 'timing/data': 0.01926109753549099, 'timing/model': 1.2510592136532068, 'learning_rate/qwen_vl_interface': 6.983798846791667e-07, 'learning_rate/action_model': 6.983798846791667e-06, 'epoch': 2.33}) +[2026-06-18 06:49:06,019][starVLA.training.train_starvla][INFO] - Step 6364, Loss: {'train/loss': 0.12433947622776031, 'train/grad_norm_pre_clip': 20.05623435974121, 'timing/data': 0.012116613797843456, 'timing/model': 1.3810559445992112, 'learning_rate/qwen_vl_interface': 6.977618718675023e-07, 'learning_rate/action_model': 6.977618718675024e-06, 'epoch': 2.33}) +[2026-06-18 06:49:10,694][starVLA.training.train_starvla][INFO] - Step 6365, Loss: {'train/loss': 0.43397408723831177, 'train/grad_norm_pre_clip': 33.07330322265625, 'timing/data': 0.015343882143497467, 'timing/model': 1.4246171871200204, 'learning_rate/qwen_vl_interface': 6.971448027397007e-07, 'learning_rate/action_model': 6.971448027397007e-06, 'epoch': 2.33}) +[2026-06-18 06:49:15,602][starVLA.training.train_starvla][INFO] - Step 6366, Loss: {'train/loss': 0.15714310109615326, 'train/grad_norm_pre_clip': 15.705846786499023, 'timing/data': 0.01079883798956871, 'timing/model': 1.0483035640791059, 'learning_rate/qwen_vl_interface': 6.965286774236814e-07, 'learning_rate/action_model': 6.965286774236813e-06, 'epoch': 2.33}) +[2026-06-18 06:49:26,582][starVLA.training.train_starvla][INFO] - Step 6367, Loss: {'train/loss': 0.20195987820625305, 'train/grad_norm_pre_clip': 25.945560455322266, 'timing/data': 0.015971804037690163, 'timing/model': 1.4990730779245496, 'learning_rate/qwen_vl_interface': 6.959134960471664e-07, 'learning_rate/action_model': 6.959134960471664e-06, 'epoch': 2.33}) +[2026-06-18 06:49:31,411][starVLA.training.train_starvla][INFO] - Step 6368, Loss: {'train/loss': 0.13704511523246765, 'train/grad_norm_pre_clip': 13.909042358398438, 'timing/data': 0.011653524823486805, 'timing/model': 1.063939593732357, 'learning_rate/qwen_vl_interface': 6.952992587376836e-07, 'learning_rate/action_model': 6.952992587376836e-06, 'epoch': 2.33}) +[2026-06-18 06:49:37,001][starVLA.training.train_starvla][INFO] - Step 6369, Loss: {'train/loss': 0.32409557700157166, 'train/grad_norm_pre_clip': 30.485858917236328, 'timing/data': 0.013097943738102913, 'timing/model': 1.392197041772306, 'learning_rate/qwen_vl_interface': 6.946859656225662e-07, 'learning_rate/action_model': 6.946859656225661e-06, 'epoch': 2.33}) +[2026-06-18 06:49:44,900][starVLA.training.train_starvla][INFO] - Step 6370, Loss: {'train/loss': 0.17500603199005127, 'train/grad_norm_pre_clip': 20.242717742919922, 'timing/data': 0.017378981225192547, 'timing/model': 1.0706067671999335, 'learning_rate/qwen_vl_interface': 6.940736168289497e-07, 'learning_rate/action_model': 6.940736168289497e-06, 'epoch': 2.33}) +[2026-06-18 06:49:50,246][starVLA.training.train_starvla][INFO] - Step 6371, Loss: {'train/loss': 0.14122317731380463, 'train/grad_norm_pre_clip': 17.532859802246094, 'timing/data': 0.016348847188055515, 'timing/model': 1.7051945328712463, 'learning_rate/qwen_vl_interface': 6.934622124837751e-07, 'learning_rate/action_model': 6.93462212483775e-06, 'epoch': 2.33}) +[2026-06-18 06:49:57,716][starVLA.training.train_starvla][INFO] - Step 6372, Loss: {'train/loss': 0.2228853553533554, 'train/grad_norm_pre_clip': 22.61389923095703, 'timing/data': 0.0010515926405787468, 'timing/model': 1.2553048813715577, 'learning_rate/qwen_vl_interface': 6.928517527137864e-07, 'learning_rate/action_model': 6.9285175271378635e-06, 'epoch': 2.33}) +[2026-06-18 06:50:04,427][starVLA.training.train_starvla][INFO] - Step 6373, Loss: {'train/loss': 0.032066524028778076, 'train/grad_norm_pre_clip': 5.119893550872803, 'timing/data': 0.003875437192618847, 'timing/model': 1.3138279747217894, 'learning_rate/qwen_vl_interface': 6.922422376455333e-07, 'learning_rate/action_model': 6.9224223764553326e-06, 'epoch': 2.33}) +[2026-06-18 06:50:10,618][starVLA.training.train_starvla][INFO] - Step 6374, Loss: {'train/loss': 0.5102235078811646, 'train/grad_norm_pre_clip': 24.909826278686523, 'timing/data': 0.0008978703990578651, 'timing/model': 1.4724797224625945, 'learning_rate/qwen_vl_interface': 6.916336674053688e-07, 'learning_rate/action_model': 6.9163366740536875e-06, 'epoch': 2.33}) +[2026-06-18 06:50:17,843][starVLA.training.train_starvla][INFO] - Step 6375, Loss: {'train/loss': 0.15127819776535034, 'train/grad_norm_pre_clip': 16.46125602722168, 'timing/data': 0.000912749208509922, 'timing/model': 1.101384892128408, 'learning_rate/qwen_vl_interface': 6.9102604211945e-07, 'learning_rate/action_model': 6.910260421194501e-06, 'epoch': 2.33}) +[2026-06-18 06:50:22,587][starVLA.training.train_starvla][INFO] - Step 6376, Loss: {'train/loss': 0.30694282054901123, 'train/grad_norm_pre_clip': 22.918941497802734, 'timing/data': 0.001904417760670185, 'timing/model': 1.075778415426612, 'learning_rate/qwen_vl_interface': 6.904193619137381e-07, 'learning_rate/action_model': 6.904193619137381e-06, 'epoch': 2.33}) +[2026-06-18 06:50:27,690][starVLA.training.train_starvla][INFO] - Step 6377, Loss: {'train/loss': 0.2438201904296875, 'train/grad_norm_pre_clip': 32.59926223754883, 'timing/data': 0.0020882217213511467, 'timing/model': 1.463396112434566, 'learning_rate/qwen_vl_interface': 6.898136269139989e-07, 'learning_rate/action_model': 6.8981362691399885e-06, 'epoch': 2.34}) +[2026-06-18 06:50:34,483][starVLA.training.train_starvla][INFO] - Step 6378, Loss: {'train/loss': 0.09061003476381302, 'train/grad_norm_pre_clip': 12.109922409057617, 'timing/data': 0.0009325835853815079, 'timing/model': 2.961585145443678, 'learning_rate/qwen_vl_interface': 6.89208837245802e-07, 'learning_rate/action_model': 6.892088372458019e-06, 'epoch': 2.34}) +[2026-06-18 06:50:41,415][starVLA.training.train_starvla][INFO] - Step 6379, Loss: {'train/loss': 0.4907337427139282, 'train/grad_norm_pre_clip': 29.758039474487305, 'timing/data': 0.0012687081471085548, 'timing/model': 1.4131008004769683, 'learning_rate/qwen_vl_interface': 6.886049930345204e-07, 'learning_rate/action_model': 6.8860499303452035e-06, 'epoch': 2.34}) +[2026-06-18 06:50:53,036][starVLA.training.train_starvla][INFO] - Step 6380, Loss: {'train/loss': 0.5471644997596741, 'train/grad_norm_pre_clip': 36.24890899658203, 'timing/data': 0.0022932803258299828, 'timing/model': 4.430107885971665, 'learning_rate/qwen_vl_interface': 6.880020944053326e-07, 'learning_rate/action_model': 6.880020944053325e-06, 'epoch': 2.34}) +[2026-06-18 06:50:57,803][starVLA.training.train_starvla][INFO] - Step 6381, Loss: {'train/loss': 0.17053373157978058, 'train/grad_norm_pre_clip': 19.3819637298584, 'timing/data': 0.0008707847446203232, 'timing/model': 1.3017324786633253, 'learning_rate/qwen_vl_interface': 6.87400141483219e-07, 'learning_rate/action_model': 6.87400141483219e-06, 'epoch': 2.34}) +[2026-06-18 06:51:03,012][starVLA.training.train_starvla][INFO] - Step 6382, Loss: {'train/loss': 0.5491873025894165, 'train/grad_norm_pre_clip': 34.513755798339844, 'timing/data': 0.002184689976274967, 'timing/model': 1.148901203647256, 'learning_rate/qwen_vl_interface': 6.867991343929655e-07, 'learning_rate/action_model': 6.867991343929656e-06, 'epoch': 2.34}) +[2026-06-18 06:51:09,308][starVLA.training.train_starvla][INFO] - Step 6383, Loss: {'train/loss': 0.3269768953323364, 'train/grad_norm_pre_clip': 25.630922317504883, 'timing/data': 0.0011293841525912285, 'timing/model': 1.0600597308948636, 'learning_rate/qwen_vl_interface': 6.861990732591622e-07, 'learning_rate/action_model': 6.861990732591622e-06, 'epoch': 2.34}) +[2026-06-18 06:51:14,160][starVLA.training.train_starvla][INFO] - Step 6384, Loss: {'train/loss': 0.34942054748535156, 'train/grad_norm_pre_clip': 30.87896156311035, 'timing/data': 0.0011241864413022995, 'timing/model': 1.508236726745963, 'learning_rate/qwen_vl_interface': 6.855999582062017e-07, 'learning_rate/action_model': 6.855999582062017e-06, 'epoch': 2.34}) +[2026-06-18 06:51:18,420][starVLA.training.train_starvla][INFO] - Step 6385, Loss: {'train/loss': 0.36653798818588257, 'train/grad_norm_pre_clip': 19.830293655395508, 'timing/data': 0.001161874271929264, 'timing/model': 1.1671056328341365, 'learning_rate/qwen_vl_interface': 6.850017893582818e-07, 'learning_rate/action_model': 6.850017893582818e-06, 'epoch': 2.34}) +[2026-06-18 06:51:23,671][starVLA.training.train_starvla][INFO] - Step 6386, Loss: {'train/loss': 0.22575251758098602, 'train/grad_norm_pre_clip': 20.27178382873535, 'timing/data': 0.0012031495571136475, 'timing/model': 1.370127771049738, 'learning_rate/qwen_vl_interface': 6.844045668394024e-07, 'learning_rate/action_model': 6.844045668394024e-06, 'epoch': 2.34}) +[2026-06-18 06:51:28,411][starVLA.training.train_starvla][INFO] - Step 6387, Loss: {'train/loss': 0.5391337871551514, 'train/grad_norm_pre_clip': 41.341957092285156, 'timing/data': 0.0009912438690662384, 'timing/model': 1.0882852654904127, 'learning_rate/qwen_vl_interface': 6.838082907733694e-07, 'learning_rate/action_model': 6.838082907733694e-06, 'epoch': 2.34}) +[2026-06-18 06:51:36,748][starVLA.training.train_starvla][INFO] - Step 6388, Loss: {'train/loss': 0.15168209373950958, 'train/grad_norm_pre_clip': 25.01840591430664, 'timing/data': 0.0014679376035928726, 'timing/model': 1.3630034625530243, 'learning_rate/qwen_vl_interface': 6.832129612837912e-07, 'learning_rate/action_model': 6.832129612837912e-06, 'epoch': 2.34}) +[2026-06-18 06:51:42,141][starVLA.training.train_starvla][INFO] - Step 6389, Loss: {'train/loss': 0.1830473095178604, 'train/grad_norm_pre_clip': 19.5430965423584, 'timing/data': 0.0008675064891576767, 'timing/model': 1.6457985183224082, 'learning_rate/qwen_vl_interface': 6.826185784940801e-07, 'learning_rate/action_model': 6.826185784940801e-06, 'epoch': 2.34}) +[2026-06-18 06:51:46,382][starVLA.training.train_starvla][INFO] - Step 6390, Loss: {'train/loss': 0.12691591680049896, 'train/grad_norm_pre_clip': 16.748079299926758, 'timing/data': 0.0007884623482823372, 'timing/model': 1.074828114360571, 'learning_rate/qwen_vl_interface': 6.820251425274526e-07, 'learning_rate/action_model': 6.820251425274527e-06, 'epoch': 2.34}) +[2026-06-18 06:51:54,501][starVLA.training.train_starvla][INFO] - Step 6391, Loss: {'train/loss': 0.17730331420898438, 'train/grad_norm_pre_clip': 13.271327018737793, 'timing/data': 0.0007862355560064316, 'timing/model': 1.1353211449459195, 'learning_rate/qwen_vl_interface': 6.81432653506928e-07, 'learning_rate/action_model': 6.814326535069279e-06, 'epoch': 2.34}) +[2026-06-18 06:52:00,019][starVLA.training.train_starvla][INFO] - Step 6392, Loss: {'train/loss': 0.3711022138595581, 'train/grad_norm_pre_clip': 31.566516876220703, 'timing/data': 0.0016513196751475334, 'timing/model': 1.5594193181023002, 'learning_rate/qwen_vl_interface': 6.808411115553305e-07, 'learning_rate/action_model': 6.808411115553305e-06, 'epoch': 2.34}) +[2026-06-18 06:52:06,094][starVLA.training.train_starvla][INFO] - Step 6393, Loss: {'train/loss': 0.2789795696735382, 'train/grad_norm_pre_clip': 33.875083923339844, 'timing/data': 0.001184580847620964, 'timing/model': 1.0323262652382255, 'learning_rate/qwen_vl_interface': 6.802505167952871e-07, 'learning_rate/action_model': 6.80250516795287e-06, 'epoch': 2.34}) +[2026-06-18 06:52:11,314][starVLA.training.train_starvla][INFO] - Step 6394, Loss: {'train/loss': 0.1536596417427063, 'train/grad_norm_pre_clip': 14.375061988830566, 'timing/data': 0.0012050420045852661, 'timing/model': 1.8429003963246942, 'learning_rate/qwen_vl_interface': 6.796608693492293e-07, 'learning_rate/action_model': 6.7966086934922925e-06, 'epoch': 2.34}) +[2026-06-18 06:52:22,807][starVLA.training.train_starvla][INFO] - Step 6395, Loss: {'train/loss': 0.2668623626232147, 'train/grad_norm_pre_clip': 18.487133026123047, 'timing/data': 0.0016973940655589104, 'timing/model': 5.01775526907295, 'learning_rate/qwen_vl_interface': 6.790721693393912e-07, 'learning_rate/action_model': 6.790721693393912e-06, 'epoch': 2.34}) +[2026-06-18 06:52:29,487][starVLA.training.train_starvla][INFO] - Step 6396, Loss: {'train/loss': 0.1114860326051712, 'train/grad_norm_pre_clip': 13.908321380615234, 'timing/data': 0.0007690805941820145, 'timing/model': 1.1674842359498143, 'learning_rate/qwen_vl_interface': 6.784844168878108e-07, 'learning_rate/action_model': 6.784844168878107e-06, 'epoch': 2.34}) +[2026-06-18 06:52:37,915][starVLA.training.train_starvla][INFO] - Step 6397, Loss: {'train/loss': 0.1924789994955063, 'train/grad_norm_pre_clip': 24.735212326049805, 'timing/data': 0.0012159598991274834, 'timing/model': 1.27899613045156, 'learning_rate/qwen_vl_interface': 6.778976121163289e-07, 'learning_rate/action_model': 6.778976121163288e-06, 'epoch': 2.34}) +[2026-06-18 06:52:42,485][starVLA.training.train_starvla][INFO] - Step 6398, Loss: {'train/loss': 0.31338921189308167, 'train/grad_norm_pre_clip': 25.664562225341797, 'timing/data': 0.0009016422554850578, 'timing/model': 1.0532885631546378, 'learning_rate/qwen_vl_interface': 6.773117551465927e-07, 'learning_rate/action_model': 6.773117551465927e-06, 'epoch': 2.34}) +[2026-06-18 06:52:48,320][starVLA.training.train_starvla][INFO] - Step 6399, Loss: {'train/loss': 0.3346267640590668, 'train/grad_norm_pre_clip': 29.06857681274414, 'timing/data': 0.0008828090503811836, 'timing/model': 1.4953073021024466, 'learning_rate/qwen_vl_interface': 6.767268461000498e-07, 'learning_rate/action_model': 6.7672684610004975e-06, 'epoch': 2.34}) +[2026-06-18 06:57:26,540][starVLA.training.train_starvla][INFO] - Step 6400, Loss: {'train/loss': 0.18548613786697388, 'train/grad_norm_pre_clip': 18.454200744628906, 'eval/loss': 0.5248108506202698, 'eval/latency_2/loss': 0.5248108506202698, 'eval/demon_fire_precision': 0.9526578073089701, 'eval/demon_fire_recall': 0.9598326359832636, 'eval/demon_fire_f1': 0.9562317632346813, 'eval/demon_fire_rate_model': 0.6700055648302726, 'eval/demon_fire_rate_teacher': 0.6649972175848636, 'eval/demon_left_precision': 0.9100642398286938, 'eval/demon_left_recall': 0.8947368421052632, 'eval/demon_left_f1': 0.9023354564755839, 'eval/demon_left_rate_model': 0.2598775737340011, 'eval/demon_left_rate_teacher': 0.2643294379521425, 'eval/demon_right_precision': 0.9091928251121076, 'eval/demon_right_recall': 0.9163841807909604, 'eval/demon_right_f1': 0.9127743387732132, 'eval/demon_right_rate_model': 0.24819143016138007, 'eval/demon_right_rate_teacher': 0.24624373956594323, 'eval/demon_move_f1': 0.9075548976243986, 'eval/demon_cc_f1': 0.9410891193972316, 'eval/latency_2/demon_fire_precision': 0.9526578073089701, 'eval/latency_2/demon_fire_recall': 0.9598326359832636, 'eval/latency_2/demon_fire_f1': 0.9562317632346813, 'eval/latency_2/demon_fire_rate_model': 0.6700055648302726, 'eval/latency_2/demon_fire_rate_teacher': 0.6649972175848636, 'eval/latency_2/demon_left_precision': 0.9100642398286938, 'eval/latency_2/demon_left_recall': 0.8947368421052632, 'eval/latency_2/demon_left_f1': 0.9023354564755839, 'eval/latency_2/demon_left_rate_model': 0.2598775737340011, 'eval/latency_2/demon_left_rate_teacher': 0.2643294379521425, 'eval/latency_2/demon_right_precision': 0.9091928251121076, 'eval/latency_2/demon_right_recall': 0.9163841807909604, 'eval/latency_2/demon_right_f1': 0.9127743387732132, 'eval/latency_2/demon_right_rate_model': 0.24819143016138007, 'eval/latency_2/demon_right_rate_teacher': 0.24624373956594323, 'eval/latency_2/demon_move_f1': 0.9075548976243986, 'eval/latency_2/demon_cc_f1': 0.9410891193972316, 'timing/data': 0.0008048843592405319, 'timing/model': 2.6839619614183903, 'learning_rate/qwen_vl_interface': 6.761428850979542e-07, 'learning_rate/action_model': 6.7614288509795415e-06, 'epoch': 2.34}) +[2026-06-18 06:57:26,555][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_6400_state +[2026-06-18 06:57:26,565][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 06:57:36,477][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_6400_state/model.safetensors +[2026-06-18 06:57:51,684][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_6400_state/optimizer.bin +[2026-06-18 06:57:51,686][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_6400_state/sampler.bin +[2026-06-18 06:57:51,687][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_6400_state/sampler_1.bin +[2026-06-18 06:57:51,689][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_6400_state/random_states_0.pkl +[2026-06-18 06:57:51,694][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 6400 +[2026-06-18 06:57:51,694][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 06:57:51,711][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 06:57:55,882][starVLA.training.train_starvla][INFO] - Step 6401, Loss: {'train/loss': 0.06494764983654022, 'train/grad_norm_pre_clip': 10.960664749145508, 'timing/data': 0.0008798250928521156, 'timing/model': 1.028081689029932, 'learning_rate/qwen_vl_interface': 6.755598722613579e-07, 'learning_rate/action_model': 6.755598722613579e-06, 'epoch': 2.34}) +[2026-06-18 06:58:00,723][starVLA.training.train_starvla][INFO] - Step 6402, Loss: {'train/loss': 0.20243845880031586, 'train/grad_norm_pre_clip': 18.557209014892578, 'timing/data': 0.0017285048961639404, 'timing/model': 1.1382237328216434, 'learning_rate/qwen_vl_interface': 6.749778077111241e-07, 'learning_rate/action_model': 6.749778077111241e-06, 'epoch': 2.34}) +[2026-06-18 06:58:11,709][starVLA.training.train_starvla][INFO] - Step 6403, Loss: {'train/loss': 0.5711154937744141, 'train/grad_norm_pre_clip': 37.73255157470703, 'timing/data': 0.0010980870574712753, 'timing/model': 1.247135897167027, 'learning_rate/qwen_vl_interface': 6.743966915679128e-07, 'learning_rate/action_model': 6.743966915679129e-06, 'epoch': 2.34}) +[2026-06-18 06:58:16,374][starVLA.training.train_starvla][INFO] - Step 6404, Loss: {'train/loss': 0.2510184347629547, 'train/grad_norm_pre_clip': 19.105730056762695, 'timing/data': 0.0009920429438352585, 'timing/model': 1.0524325361475348, 'learning_rate/qwen_vl_interface': 6.738165239521909e-07, 'learning_rate/action_model': 6.7381652395219094e-06, 'epoch': 2.34}) +[2026-06-18 06:58:21,370][starVLA.training.train_starvla][INFO] - Step 6405, Loss: {'train/loss': 0.2438739836215973, 'train/grad_norm_pre_clip': 17.612018585205078, 'timing/data': 0.0012861331924796104, 'timing/model': 1.585748522542417, 'learning_rate/qwen_vl_interface': 6.732373049842284e-07, 'learning_rate/action_model': 6.7323730498422846e-06, 'epoch': 2.35}) +[2026-06-18 06:58:25,706][starVLA.training.train_starvla][INFO] - Step 6406, Loss: {'train/loss': 0.30806005001068115, 'train/grad_norm_pre_clip': 28.64442253112793, 'timing/data': 0.0007474925369024277, 'timing/model': 1.0868675550445914, 'learning_rate/qwen_vl_interface': 6.726590347840972e-07, 'learning_rate/action_model': 6.726590347840972e-06, 'epoch': 2.35}) +[2026-06-18 06:58:30,591][starVLA.training.train_starvla][INFO] - Step 6407, Loss: {'train/loss': 0.29930442571640015, 'train/grad_norm_pre_clip': 23.671735763549805, 'timing/data': 0.0008797850459814072, 'timing/model': 1.025228082202375, 'learning_rate/qwen_vl_interface': 6.720817134716736e-07, 'learning_rate/action_model': 6.720817134716736e-06, 'epoch': 2.35}) +[2026-06-18 06:58:38,895][starVLA.training.train_starvla][INFO] - Step 6408, Loss: {'train/loss': 0.5027791261672974, 'train/grad_norm_pre_clip': 39.50389862060547, 'timing/data': 0.0013730600476264954, 'timing/model': 1.083501549437642, 'learning_rate/qwen_vl_interface': 6.71505341166636e-07, 'learning_rate/action_model': 6.71505341166636e-06, 'epoch': 2.35}) +[2026-06-18 06:58:43,918][starVLA.training.train_starvla][INFO] - Step 6409, Loss: {'train/loss': 0.11776372045278549, 'train/grad_norm_pre_clip': 16.35841941833496, 'timing/data': 0.0008719433099031448, 'timing/model': 1.0813217163085938, 'learning_rate/qwen_vl_interface': 6.709299179884689e-07, 'learning_rate/action_model': 6.709299179884689e-06, 'epoch': 2.35}) +[2026-06-18 06:58:48,291][starVLA.training.train_starvla][INFO] - Step 6410, Loss: {'train/loss': 0.39539292454719543, 'train/grad_norm_pre_clip': 26.478776931762695, 'timing/data': 0.0012523746117949486, 'timing/model': 1.3694481570273638, 'learning_rate/qwen_vl_interface': 6.703554440564569e-07, 'learning_rate/action_model': 6.703554440564569e-06, 'epoch': 2.35}) +[2026-06-18 06:58:59,345][starVLA.training.train_starvla][INFO] - Step 6411, Loss: {'train/loss': 0.2203591763973236, 'train/grad_norm_pre_clip': 29.591806411743164, 'timing/data': 0.0010495753958821297, 'timing/model': 2.7136298278346658, 'learning_rate/qwen_vl_interface': 6.697819194896889e-07, 'learning_rate/action_model': 6.697819194896889e-06, 'epoch': 2.35}) +[2026-06-18 06:59:09,056][starVLA.training.train_starvla][INFO] - Step 6412, Loss: {'train/loss': 0.29682815074920654, 'train/grad_norm_pre_clip': 25.576871871948242, 'timing/data': 0.003723287023603916, 'timing/model': 2.4327713577076793, 'learning_rate/qwen_vl_interface': 6.692093444070571e-07, 'learning_rate/action_model': 6.692093444070572e-06, 'epoch': 2.35}) +[2026-06-18 06:59:19,258][starVLA.training.train_starvla][INFO] - Step 6413, Loss: {'train/loss': 0.2455918937921524, 'train/grad_norm_pre_clip': 18.913015365600586, 'timing/data': 0.0010812748223543167, 'timing/model': 5.048445610329509, 'learning_rate/qwen_vl_interface': 6.686377189272584e-07, 'learning_rate/action_model': 6.686377189272583e-06, 'epoch': 2.35}) +[2026-06-18 06:59:31,214][starVLA.training.train_starvla][INFO] - Step 6414, Loss: {'train/loss': 0.48828914761543274, 'train/grad_norm_pre_clip': 30.142248153686523, 'timing/data': 0.0020027579739689827, 'timing/model': 5.060869857668877, 'learning_rate/qwen_vl_interface': 6.680670431687898e-07, 'learning_rate/action_model': 6.680670431687898e-06, 'epoch': 2.35}) +[2026-06-18 06:59:37,221][starVLA.training.train_starvla][INFO] - Step 6415, Loss: {'train/loss': 0.4716498851776123, 'train/grad_norm_pre_clip': 30.43393898010254, 'timing/data': 0.000898454338312149, 'timing/model': 2.71426185593009, 'learning_rate/qwen_vl_interface': 6.674973172499536e-07, 'learning_rate/action_model': 6.674973172499535e-06, 'epoch': 2.35}) +[2026-06-18 06:59:41,817][starVLA.training.train_starvla][INFO] - Step 6416, Loss: {'train/loss': 0.0350305549800396, 'train/grad_norm_pre_clip': 4.759254455566406, 'timing/data': 0.0009700357913970947, 'timing/model': 1.0767343174666166, 'learning_rate/qwen_vl_interface': 6.669285412888544e-07, 'learning_rate/action_model': 6.669285412888544e-06, 'epoch': 2.35}) +[2026-06-18 06:59:50,083][starVLA.training.train_starvla][INFO] - Step 6417, Loss: {'train/loss': 0.03371540829539299, 'train/grad_norm_pre_clip': 6.200414180755615, 'timing/data': 0.001177680678665638, 'timing/model': 4.566863814368844, 'learning_rate/qwen_vl_interface': 6.663607154034003e-07, 'learning_rate/action_model': 6.663607154034003e-06, 'epoch': 2.35}) +[2026-06-18 06:59:58,636][starVLA.training.train_starvla][INFO] - Step 6418, Loss: {'train/loss': 0.1322789043188095, 'train/grad_norm_pre_clip': 17.094741821289062, 'timing/data': 0.001203242689371109, 'timing/model': 1.8486713636666536, 'learning_rate/qwen_vl_interface': 6.657938397113019e-07, 'learning_rate/action_model': 6.657938397113018e-06, 'epoch': 2.35}) +[2026-06-18 07:00:06,888][starVLA.training.train_starvla][INFO] - Step 6419, Loss: {'train/loss': 0.28201407194137573, 'train/grad_norm_pre_clip': 29.00405502319336, 'timing/data': 0.0012504290789365768, 'timing/model': 1.2332933815196157, 'learning_rate/qwen_vl_interface': 6.652279143300734e-07, 'learning_rate/action_model': 6.652279143300734e-06, 'epoch': 2.35}) +[2026-06-18 07:00:17,475][starVLA.training.train_starvla][INFO] - Step 6420, Loss: {'train/loss': 0.19436562061309814, 'train/grad_norm_pre_clip': 25.056659698486328, 'timing/data': 0.0018137693405151367, 'timing/model': 1.6524101989343762, 'learning_rate/qwen_vl_interface': 6.64662939377032e-07, 'learning_rate/action_model': 6.646629393770321e-06, 'epoch': 2.35}) +[2026-06-18 07:00:25,378][starVLA.training.train_starvla][INFO] - Step 6421, Loss: {'train/loss': 0.2969829738140106, 'train/grad_norm_pre_clip': 26.47412872314453, 'timing/data': 0.001019875518977642, 'timing/model': 1.0592022528871894, 'learning_rate/qwen_vl_interface': 6.64098914969297e-07, 'learning_rate/action_model': 6.6409891496929695e-06, 'epoch': 2.35}) +[2026-06-18 07:00:29,902][starVLA.training.train_starvla][INFO] - Step 6422, Loss: {'train/loss': 0.11958792060613632, 'train/grad_norm_pre_clip': 13.512481689453125, 'timing/data': 0.0009164623916149139, 'timing/model': 1.0855190614238381, 'learning_rate/qwen_vl_interface': 6.635358412237912e-07, 'learning_rate/action_model': 6.635358412237912e-06, 'epoch': 2.35}) +[2026-06-18 07:00:34,730][starVLA.training.train_starvla][INFO] - Step 6423, Loss: {'train/loss': 0.31776413321495056, 'train/grad_norm_pre_clip': 23.39145278930664, 'timing/data': 0.001391134224832058, 'timing/model': 1.1270075403153896, 'learning_rate/qwen_vl_interface': 6.629737182572418e-07, 'learning_rate/action_model': 6.629737182572418e-06, 'epoch': 2.35}) +[2026-06-18 07:00:39,281][starVLA.training.train_starvla][INFO] - Step 6424, Loss: {'train/loss': 0.2944942116737366, 'train/grad_norm_pre_clip': 25.3697566986084, 'timing/data': 0.0011739116162061691, 'timing/model': 1.0412457026541233, 'learning_rate/qwen_vl_interface': 6.62412546186176e-07, 'learning_rate/action_model': 6.62412546186176e-06, 'epoch': 2.35}) +[2026-06-18 07:00:43,833][starVLA.training.train_starvla][INFO] - Step 6425, Loss: {'train/loss': 0.419801265001297, 'train/grad_norm_pre_clip': 30.04054069519043, 'timing/data': 0.0007687294855713844, 'timing/model': 1.3067783107981086, 'learning_rate/qwen_vl_interface': 6.61852325126926e-07, 'learning_rate/action_model': 6.61852325126926e-06, 'epoch': 2.35}) +[2026-06-18 07:00:48,305][starVLA.training.train_starvla][INFO] - Step 6426, Loss: {'train/loss': 0.21947436034679413, 'train/grad_norm_pre_clip': 20.309951782226562, 'timing/data': 0.0010273605585098267, 'timing/model': 1.058379971422255, 'learning_rate/qwen_vl_interface': 6.61293055195625e-07, 'learning_rate/action_model': 6.6129305519562504e-06, 'epoch': 2.35}) +[2026-06-18 07:00:52,800][starVLA.training.train_starvla][INFO] - Step 6427, Loss: {'train/loss': 0.06577424705028534, 'train/grad_norm_pre_clip': 11.653641700744629, 'timing/data': 0.0008562915027141571, 'timing/model': 1.0512706898152828, 'learning_rate/qwen_vl_interface': 6.607347365082127e-07, 'learning_rate/action_model': 6.607347365082127e-06, 'epoch': 2.35}) +[2026-06-18 07:00:57,528][starVLA.training.train_starvla][INFO] - Step 6428, Loss: {'train/loss': 0.1674206405878067, 'train/grad_norm_pre_clip': 19.00470542907715, 'timing/data': 0.0012503666803240776, 'timing/model': 1.1125832265242934, 'learning_rate/qwen_vl_interface': 6.601773691804273e-07, 'learning_rate/action_model': 6.6017736918042735e-06, 'epoch': 2.35}) +[2026-06-18 07:01:02,294][starVLA.training.train_starvla][INFO] - Step 6429, Loss: {'train/loss': 0.2669379413127899, 'train/grad_norm_pre_clip': 27.02927017211914, 'timing/data': 0.0010448088869452477, 'timing/model': 1.0660588014870882, 'learning_rate/qwen_vl_interface': 6.596209533278122e-07, 'learning_rate/action_model': 6.596209533278122e-06, 'epoch': 2.35}) +[2026-06-18 07:01:06,521][starVLA.training.train_starvla][INFO] - Step 6430, Loss: {'train/loss': 0.11246828734874725, 'train/grad_norm_pre_clip': 15.530472755432129, 'timing/data': 0.0011474108323454857, 'timing/model': 1.1099395044147968, 'learning_rate/qwen_vl_interface': 6.590654890657125e-07, 'learning_rate/action_model': 6.590654890657125e-06, 'epoch': 2.35}) +[2026-06-18 07:01:10,997][starVLA.training.train_starvla][INFO] - Step 6431, Loss: {'train/loss': 0.1091846153140068, 'train/grad_norm_pre_clip': 10.542081832885742, 'timing/data': 0.001015237532556057, 'timing/model': 1.0510037178173661, 'learning_rate/qwen_vl_interface': 6.585109765092772e-07, 'learning_rate/action_model': 6.585109765092772e-06, 'epoch': 2.35}) +[2026-06-18 07:01:15,556][starVLA.training.train_starvla][INFO] - Step 6432, Loss: {'train/loss': 0.433019757270813, 'train/grad_norm_pre_clip': 43.70532989501953, 'timing/data': 0.002100869081914425, 'timing/model': 1.0380479078739882, 'learning_rate/qwen_vl_interface': 6.579574157734575e-07, 'learning_rate/action_model': 6.579574157734575e-06, 'epoch': 2.36}) +[2026-06-18 07:01:19,982][starVLA.training.train_starvla][INFO] - Step 6433, Loss: {'train/loss': 0.3177127242088318, 'train/grad_norm_pre_clip': 51.67463684082031, 'timing/data': 0.0012000314891338348, 'timing/model': 1.0221540620550513, 'learning_rate/qwen_vl_interface': 6.574048069730062e-07, 'learning_rate/action_model': 6.574048069730063e-06, 'epoch': 2.36}) +[2026-06-18 07:01:24,448][starVLA.training.train_starvla][INFO] - Step 6434, Loss: {'train/loss': 0.5370810627937317, 'train/grad_norm_pre_clip': 31.562829971313477, 'timing/data': 0.0009673340246081352, 'timing/model': 1.1737572718411684, 'learning_rate/qwen_vl_interface': 6.568531502224808e-07, 'learning_rate/action_model': 6.568531502224808e-06, 'epoch': 2.36}) +[2026-06-18 07:01:30,082][starVLA.training.train_starvla][INFO] - Step 6435, Loss: {'train/loss': 0.3405672609806061, 'train/grad_norm_pre_clip': 27.90431022644043, 'timing/data': 0.0012644603848457336, 'timing/model': 2.349586428143084, 'learning_rate/qwen_vl_interface': 6.563024456362395e-07, 'learning_rate/action_model': 6.563024456362395e-06, 'epoch': 2.36}) +[2026-06-18 07:01:34,708][starVLA.training.train_starvla][INFO] - Step 6436, Loss: {'train/loss': 0.1471582055091858, 'train/grad_norm_pre_clip': 19.613449096679688, 'timing/data': 0.004012857563793659, 'timing/model': 1.234165208414197, 'learning_rate/qwen_vl_interface': 6.557526933284443e-07, 'learning_rate/action_model': 6.557526933284443e-06, 'epoch': 2.36}) +[2026-06-18 07:01:41,011][starVLA.training.train_starvla][INFO] - Step 6437, Loss: {'train/loss': 0.3519645929336548, 'train/grad_norm_pre_clip': 41.91379928588867, 'timing/data': 0.0008516088128089905, 'timing/model': 2.505541962571442, 'learning_rate/qwen_vl_interface': 6.552038934130595e-07, 'learning_rate/action_model': 6.552038934130594e-06, 'epoch': 2.36}) +[2026-06-18 07:01:49,405][starVLA.training.train_starvla][INFO] - Step 6438, Loss: {'train/loss': 0.2281792014837265, 'train/grad_norm_pre_clip': 24.22193717956543, 'timing/data': 0.05409812834113836, 'timing/model': 1.302800533361733, 'learning_rate/qwen_vl_interface': 6.546560460038514e-07, 'learning_rate/action_model': 6.546560460038514e-06, 'epoch': 2.36}) +[2026-06-18 07:01:56,577][starVLA.training.train_starvla][INFO] - Step 6439, Loss: {'train/loss': 0.4563429653644562, 'train/grad_norm_pre_clip': 31.05107307434082, 'timing/data': 0.0014104172587394714, 'timing/model': 1.309591663070023, 'learning_rate/qwen_vl_interface': 6.5410915121439e-07, 'learning_rate/action_model': 6.5410915121439e-06, 'epoch': 2.36}) +[2026-06-18 07:02:04,073][starVLA.training.train_starvla][INFO] - Step 6440, Loss: {'train/loss': 0.30932095646858215, 'train/grad_norm_pre_clip': 27.339981079101562, 'timing/data': 0.0009580543264746666, 'timing/model': 3.0849750712513924, 'learning_rate/qwen_vl_interface': 6.535632091580474e-07, 'learning_rate/action_model': 6.535632091580474e-06, 'epoch': 2.36}) +[2026-06-18 07:02:13,291][starVLA.training.train_starvla][INFO] - Step 6441, Loss: {'train/loss': 0.19647778570652008, 'train/grad_norm_pre_clip': 23.191450119018555, 'timing/data': 0.0010337969288229942, 'timing/model': 1.416906108148396, 'learning_rate/qwen_vl_interface': 6.530182199479965e-07, 'learning_rate/action_model': 6.5301821994799645e-06, 'epoch': 2.36}) +[2026-06-18 07:02:27,844][starVLA.training.train_starvla][INFO] - Step 6442, Loss: {'train/loss': 0.603783369064331, 'train/grad_norm_pre_clip': 39.356719970703125, 'timing/data': 0.0011560572311282158, 'timing/model': 1.1044800346717238, 'learning_rate/qwen_vl_interface': 6.524741836972162e-07, 'learning_rate/action_model': 6.524741836972162e-06, 'epoch': 2.36}) +[2026-06-18 07:02:34,391][starVLA.training.train_starvla][INFO] - Step 6443, Loss: {'train/loss': 0.2723740041255951, 'train/grad_norm_pre_clip': 19.364545822143555, 'timing/data': 0.000941702164709568, 'timing/model': 2.584984085522592, 'learning_rate/qwen_vl_interface': 6.519311005184848e-07, 'learning_rate/action_model': 6.5193110051848476e-06, 'epoch': 2.36}) +[2026-06-18 07:02:47,466][starVLA.training.train_starvla][INFO] - Step 6444, Loss: {'train/loss': 0.13847723603248596, 'train/grad_norm_pre_clip': 13.52110767364502, 'timing/data': 0.0011149318888783455, 'timing/model': 5.668357027694583, 'learning_rate/qwen_vl_interface': 6.51388970524383e-07, 'learning_rate/action_model': 6.513889705243829e-06, 'epoch': 2.36}) +[2026-06-18 07:02:52,545][starVLA.training.train_starvla][INFO] - Step 6445, Loss: {'train/loss': 0.12181905657052994, 'train/grad_norm_pre_clip': 21.06749153137207, 'timing/data': 0.000895393081009388, 'timing/model': 1.3803391447290778, 'learning_rate/qwen_vl_interface': 6.508477938272963e-07, 'learning_rate/action_model': 6.508477938272962e-06, 'epoch': 2.36}) +[2026-06-18 07:02:56,862][starVLA.training.train_starvla][INFO] - Step 6446, Loss: {'train/loss': 0.28124865889549255, 'train/grad_norm_pre_clip': 28.055530548095703, 'timing/data': 0.0009114043787121773, 'timing/model': 1.0359188122674823, 'learning_rate/qwen_vl_interface': 6.503075705394113e-07, 'learning_rate/action_model': 6.5030757053941125e-06, 'epoch': 2.36}) +[2026-06-18 07:03:01,387][starVLA.training.train_starvla][INFO] - Step 6447, Loss: {'train/loss': 0.3395631015300751, 'train/grad_norm_pre_clip': 26.786495208740234, 'timing/data': 0.0010810717940330505, 'timing/model': 1.034760874696076, 'learning_rate/qwen_vl_interface': 6.497683007727166e-07, 'learning_rate/action_model': 6.497683007727166e-06, 'epoch': 2.36}) +[2026-06-18 07:03:05,878][starVLA.training.train_starvla][INFO] - Step 6448, Loss: {'train/loss': 0.1354856938123703, 'train/grad_norm_pre_clip': 31.965192794799805, 'timing/data': 0.0011101216077804565, 'timing/model': 1.040569657459855, 'learning_rate/qwen_vl_interface': 6.492299846390024e-07, 'learning_rate/action_model': 6.492299846390023e-06, 'epoch': 2.36}) +[2026-06-18 07:03:10,541][starVLA.training.train_starvla][INFO] - Step 6449, Loss: {'train/loss': 0.3344891667366028, 'train/grad_norm_pre_clip': 26.2253475189209, 'timing/data': 0.0008534062653779984, 'timing/model': 1.0560509348288178, 'learning_rate/qwen_vl_interface': 6.486926222498633e-07, 'learning_rate/action_model': 6.486926222498633e-06, 'epoch': 2.36}) +[2026-06-18 07:03:15,339][starVLA.training.train_starvla][INFO] - Step 6450, Loss: {'train/loss': 0.1280706375837326, 'train/grad_norm_pre_clip': 14.802023887634277, 'timing/data': 0.0010063042864203453, 'timing/model': 1.4515112871304154, 'learning_rate/qwen_vl_interface': 6.481562137166952e-07, 'learning_rate/action_model': 6.481562137166952e-06, 'epoch': 2.36}) +[2026-06-18 07:03:19,547][starVLA.training.train_starvla][INFO] - Step 6451, Loss: {'train/loss': 0.19780492782592773, 'train/grad_norm_pre_clip': 18.62278938293457, 'timing/data': 0.0011129118502140045, 'timing/model': 1.1330545078963041, 'learning_rate/qwen_vl_interface': 6.476207591506954e-07, 'learning_rate/action_model': 6.476207591506954e-06, 'epoch': 2.36}) +[2026-06-18 07:03:24,034][starVLA.training.train_starvla][INFO] - Step 6452, Loss: {'train/loss': 0.3436800241470337, 'train/grad_norm_pre_clip': 28.719066619873047, 'timing/data': 0.0008904198184609413, 'timing/model': 1.064585791900754, 'learning_rate/qwen_vl_interface': 6.470862586628645e-07, 'learning_rate/action_model': 6.470862586628646e-06, 'epoch': 2.36}) +[2026-06-18 07:03:28,668][starVLA.training.train_starvla][INFO] - Step 6453, Loss: {'train/loss': 0.1946219950914383, 'train/grad_norm_pre_clip': 19.916223526000977, 'timing/data': 0.0014182711020112038, 'timing/model': 1.1417854856699705, 'learning_rate/qwen_vl_interface': 6.465527123640056e-07, 'learning_rate/action_model': 6.465527123640056e-06, 'epoch': 2.36}) +[2026-06-18 07:03:33,401][starVLA.training.train_starvla][INFO] - Step 6454, Loss: {'train/loss': 0.1901998072862625, 'train/grad_norm_pre_clip': 25.462703704833984, 'timing/data': 0.0008641406893730164, 'timing/model': 1.0838623195886612, 'learning_rate/qwen_vl_interface': 6.460201203647229e-07, 'learning_rate/action_model': 6.460201203647229e-06, 'epoch': 2.36}) +[2026-06-18 07:03:38,020][starVLA.training.train_starvla][INFO] - Step 6455, Loss: {'train/loss': 0.22616752982139587, 'train/grad_norm_pre_clip': 26.325042724609375, 'timing/data': 0.0009015509858727455, 'timing/model': 1.5069213146343827, 'learning_rate/qwen_vl_interface': 6.454884827754224e-07, 'learning_rate/action_model': 6.4548848277542235e-06, 'epoch': 2.36}) +[2026-06-18 07:03:46,956][starVLA.training.train_starvla][INFO] - Step 6456, Loss: {'train/loss': 0.17187342047691345, 'train/grad_norm_pre_clip': 20.140722274780273, 'timing/data': 0.001156313344836235, 'timing/model': 1.1708049699664116, 'learning_rate/qwen_vl_interface': 6.449577997063156e-07, 'learning_rate/action_model': 6.449577997063155e-06, 'epoch': 2.36}) +[2026-06-18 07:03:51,923][starVLA.training.train_starvla][INFO] - Step 6457, Loss: {'train/loss': 0.5674946904182434, 'train/grad_norm_pre_clip': 22.65689468383789, 'timing/data': 0.0014975136145949364, 'timing/model': 1.0643427651375532, 'learning_rate/qwen_vl_interface': 6.444280712674112e-07, 'learning_rate/action_model': 6.444280712674111e-06, 'epoch': 2.36}) +[2026-06-18 07:04:02,846][starVLA.training.train_starvla][INFO] - Step 6458, Loss: {'train/loss': 0.03903378173708916, 'train/grad_norm_pre_clip': 10.170466423034668, 'timing/data': 0.0009238375350832939, 'timing/model': 4.245751367881894, 'learning_rate/qwen_vl_interface': 6.438992975685237e-07, 'learning_rate/action_model': 6.438992975685237e-06, 'epoch': 2.36}) +[2026-06-18 07:04:07,847][starVLA.training.train_starvla][INFO] - Step 6459, Loss: {'train/loss': 0.1600450873374939, 'train/grad_norm_pre_clip': 16.574121475219727, 'timing/data': 0.0008874256163835526, 'timing/model': 1.1138154258951545, 'learning_rate/qwen_vl_interface': 6.433714787192677e-07, 'learning_rate/action_model': 6.4337147871926766e-06, 'epoch': 2.37}) +[2026-06-18 07:04:12,579][starVLA.training.train_starvla][INFO] - Step 6460, Loss: {'train/loss': 0.15861432254314423, 'train/grad_norm_pre_clip': 14.313173294067383, 'timing/data': 0.0015051178634166718, 'timing/model': 1.6723187826573849, 'learning_rate/qwen_vl_interface': 6.428446148290615e-07, 'learning_rate/action_model': 6.428446148290615e-06, 'epoch': 2.37}) +[2026-06-18 07:04:18,708][starVLA.training.train_starvla][INFO] - Step 6461, Loss: {'train/loss': 0.29713666439056396, 'train/grad_norm_pre_clip': 20.926401138305664, 'timing/data': 0.0013656225055456161, 'timing/model': 2.3521511806175113, 'learning_rate/qwen_vl_interface': 6.423187060071248e-07, 'learning_rate/action_model': 6.423187060071248e-06, 'epoch': 2.37}) +[2026-06-18 07:04:28,917][starVLA.training.train_starvla][INFO] - Step 6462, Loss: {'train/loss': 0.19702933728694916, 'train/grad_norm_pre_clip': 20.87875747680664, 'timing/data': 0.0008609909564256668, 'timing/model': 4.768569645471871, 'learning_rate/qwen_vl_interface': 6.417937523624778e-07, 'learning_rate/action_model': 6.417937523624778e-06, 'epoch': 2.37}) +[2026-06-18 07:04:39,590][starVLA.training.train_starvla][INFO] - Step 6463, Loss: {'train/loss': 0.23339010775089264, 'train/grad_norm_pre_clip': 31.381868362426758, 'timing/data': 0.0008697984740138054, 'timing/model': 1.053666627034545, 'learning_rate/qwen_vl_interface': 6.412697540039442e-07, 'learning_rate/action_model': 6.412697540039442e-06, 'epoch': 2.37}) +[2026-06-18 07:04:44,315][starVLA.training.train_starvla][INFO] - Step 6464, Loss: {'train/loss': 0.09627452492713928, 'train/grad_norm_pre_clip': 13.43234634399414, 'timing/data': 0.0012503471225500107, 'timing/model': 1.084120904095471, 'learning_rate/qwen_vl_interface': 6.407467110401505e-07, 'learning_rate/action_model': 6.407467110401505e-06, 'epoch': 2.37}) +[2026-06-18 07:04:51,852][starVLA.training.train_starvla][INFO] - Step 6465, Loss: {'train/loss': 0.3973012864589691, 'train/grad_norm_pre_clip': 33.097068786621094, 'timing/data': 0.0016777608543634415, 'timing/model': 1.59296918194741, 'learning_rate/qwen_vl_interface': 6.402246235795228e-07, 'learning_rate/action_model': 6.402246235795227e-06, 'epoch': 2.37}) +[2026-06-18 07:04:56,023][starVLA.training.train_starvla][INFO] - Step 6466, Loss: {'train/loss': 0.3626674711704254, 'train/grad_norm_pre_clip': 25.96297836303711, 'timing/data': 0.0011840565130114555, 'timing/model': 1.1176267880946398, 'learning_rate/qwen_vl_interface': 6.397034917302913e-07, 'learning_rate/action_model': 6.397034917302913e-06, 'epoch': 2.37}) +[2026-06-18 07:05:05,360][starVLA.training.train_starvla][INFO] - Step 6467, Loss: {'train/loss': 0.25850531458854675, 'train/grad_norm_pre_clip': 25.227155685424805, 'timing/data': 0.0008422983810305595, 'timing/model': 5.861702633090317, 'learning_rate/qwen_vl_interface': 6.391833156004871e-07, 'learning_rate/action_model': 6.39183315600487e-06, 'epoch': 2.37}) +[2026-06-18 07:05:20,777][starVLA.training.train_starvla][INFO] - Step 6468, Loss: {'train/loss': 0.8007364869117737, 'train/grad_norm_pre_clip': 36.871192932128906, 'timing/data': 0.0010370006784796715, 'timing/model': 4.042290876619518, 'learning_rate/qwen_vl_interface': 6.386640952979427e-07, 'learning_rate/action_model': 6.386640952979426e-06, 'epoch': 2.37}) +[2026-06-18 07:05:25,723][starVLA.training.train_starvla][INFO] - Step 6469, Loss: {'train/loss': 0.34864529967308044, 'train/grad_norm_pre_clip': 28.237823486328125, 'timing/data': 0.0011995593085885048, 'timing/model': 1.6368887918069959, 'learning_rate/qwen_vl_interface': 6.381458309302937e-07, 'learning_rate/action_model': 6.381458309302937e-06, 'epoch': 2.37}) +[2026-06-18 07:05:32,208][starVLA.training.train_starvla][INFO] - Step 6470, Loss: {'train/loss': 0.1512145698070526, 'train/grad_norm_pre_clip': 23.095407485961914, 'timing/data': 0.0008849333971738815, 'timing/model': 3.0982304997742176, 'learning_rate/qwen_vl_interface': 6.376285226049762e-07, 'learning_rate/action_model': 6.376285226049762e-06, 'epoch': 2.37}) +[2026-06-18 07:05:38,514][starVLA.training.train_starvla][INFO] - Step 6471, Loss: {'train/loss': 0.3364831805229187, 'train/grad_norm_pre_clip': 23.246797561645508, 'timing/data': 0.0008417600765824318, 'timing/model': 2.777690140530467, 'learning_rate/qwen_vl_interface': 6.371121704292292e-07, 'learning_rate/action_model': 6.371121704292292e-06, 'epoch': 2.37}) +[2026-06-18 07:05:48,207][starVLA.training.train_starvla][INFO] - Step 6472, Loss: {'train/loss': 0.5357183814048767, 'train/grad_norm_pre_clip': 32.501190185546875, 'timing/data': 0.0009923316538333893, 'timing/model': 1.0450021093711257, 'learning_rate/qwen_vl_interface': 6.365967745100926e-07, 'learning_rate/action_model': 6.3659677451009256e-06, 'epoch': 2.37}) +[2026-06-18 07:05:54,594][starVLA.training.train_starvla][INFO] - Step 6473, Loss: {'train/loss': 0.3422227203845978, 'train/grad_norm_pre_clip': 28.535295486450195, 'timing/data': 0.0011144643649458885, 'timing/model': 1.1492787636816502, 'learning_rate/qwen_vl_interface': 6.360823349544092e-07, 'learning_rate/action_model': 6.360823349544092e-06, 'epoch': 2.37}) +[2026-06-18 07:05:59,425][starVLA.training.train_starvla][INFO] - Step 6474, Loss: {'train/loss': 0.32938528060913086, 'train/grad_norm_pre_clip': 24.03323745727539, 'timing/data': 0.0012093307450413704, 'timing/model': 1.0923248091712594, 'learning_rate/qwen_vl_interface': 6.355688518688222e-07, 'learning_rate/action_model': 6.355688518688221e-06, 'epoch': 2.37}) +[2026-06-18 07:06:03,836][starVLA.training.train_starvla][INFO] - Step 6475, Loss: {'train/loss': 0.6020224690437317, 'train/grad_norm_pre_clip': 40.27295684814453, 'timing/data': 0.0010160021483898163, 'timing/model': 1.3382372222840786, 'learning_rate/qwen_vl_interface': 6.350563253597781e-07, 'learning_rate/action_model': 6.350563253597781e-06, 'epoch': 2.37}) +[2026-06-18 07:06:13,014][starVLA.training.train_starvla][INFO] - Step 6476, Loss: {'train/loss': 0.30027836561203003, 'train/grad_norm_pre_clip': 26.45271110534668, 'timing/data': 0.0012576756998896599, 'timing/model': 5.700212007388473, 'learning_rate/qwen_vl_interface': 6.345447555335236e-07, 'learning_rate/action_model': 6.3454475553352345e-06, 'epoch': 2.37}) +[2026-06-18 07:06:18,147][starVLA.training.train_starvla][INFO] - Step 6477, Loss: {'train/loss': 0.2556503415107727, 'train/grad_norm_pre_clip': 20.444913864135742, 'timing/data': 0.0007794154807925224, 'timing/model': 1.1002858076244593, 'learning_rate/qwen_vl_interface': 6.340341424961072e-07, 'learning_rate/action_model': 6.340341424961071e-06, 'epoch': 2.37}) +[2026-06-18 07:06:26,021][starVLA.training.train_starvla][INFO] - Step 6478, Loss: {'train/loss': 0.7365237474441528, 'train/grad_norm_pre_clip': 43.91731643676758, 'timing/data': 0.0011592097580432892, 'timing/model': 2.8562078438699245, 'learning_rate/qwen_vl_interface': 6.335244863533802e-07, 'learning_rate/action_model': 6.3352448635338016e-06, 'epoch': 2.37}) +[2026-06-18 07:06:35,261][starVLA.training.train_starvla][INFO] - Step 6479, Loss: {'train/loss': 0.12409616261720657, 'train/grad_norm_pre_clip': 11.797406196594238, 'timing/data': 0.0010357014834880829, 'timing/model': 1.0437012501060963, 'learning_rate/qwen_vl_interface': 6.330157872109953e-07, 'learning_rate/action_model': 6.330157872109953e-06, 'epoch': 2.37}) +[2026-06-18 07:06:41,254][starVLA.training.train_starvla][INFO] - Step 6480, Loss: {'train/loss': 0.17091728746891022, 'train/grad_norm_pre_clip': 15.359049797058105, 'timing/data': 0.0010847821831703186, 'timing/model': 2.705039866268635, 'learning_rate/qwen_vl_interface': 6.325080451744053e-07, 'learning_rate/action_model': 6.325080451744053e-06, 'epoch': 2.37}) +[2026-06-18 07:06:46,594][starVLA.training.train_starvla][INFO] - Step 6481, Loss: {'train/loss': 0.16518756747245789, 'train/grad_norm_pre_clip': 12.773422241210938, 'timing/data': 0.000766347162425518, 'timing/model': 1.2133577419444919, 'learning_rate/qwen_vl_interface': 6.320012603488663e-07, 'learning_rate/action_model': 6.320012603488662e-06, 'epoch': 2.37}) +[2026-06-18 07:06:51,800][starVLA.training.train_starvla][INFO] - Step 6482, Loss: {'train/loss': 0.27080798149108887, 'train/grad_norm_pre_clip': 35.761634826660156, 'timing/data': 0.0012931348755955696, 'timing/model': 1.3435500366613269, 'learning_rate/qwen_vl_interface': 6.314954328394351e-07, 'learning_rate/action_model': 6.3149543283943515e-06, 'epoch': 2.37}) +[2026-06-18 07:06:57,203][starVLA.training.train_starvla][INFO] - Step 6483, Loss: {'train/loss': 0.0756153091788292, 'train/grad_norm_pre_clip': 13.271984100341797, 'timing/data': 0.002205589786171913, 'timing/model': 1.0825335858389735, 'learning_rate/qwen_vl_interface': 6.309905627509707e-07, 'learning_rate/action_model': 6.309905627509706e-06, 'epoch': 2.37}) +[2026-06-18 07:07:09,086][starVLA.training.train_starvla][INFO] - Step 6484, Loss: {'train/loss': 0.210643008351326, 'train/grad_norm_pre_clip': 30.193143844604492, 'timing/data': 0.0009148251265287399, 'timing/model': 1.0845916951075196, 'learning_rate/qwen_vl_interface': 6.304866501881327e-07, 'learning_rate/action_model': 6.304866501881327e-06, 'epoch': 2.37}) +[2026-06-18 07:07:14,182][starVLA.training.train_starvla][INFO] - Step 6485, Loss: {'train/loss': 0.19395367801189423, 'train/grad_norm_pre_clip': 22.778024673461914, 'timing/data': 0.0011445963755249977, 'timing/model': 1.3220721753314137, 'learning_rate/qwen_vl_interface': 6.299836952553831e-07, 'learning_rate/action_model': 6.299836952553831e-06, 'epoch': 2.37}) +[2026-06-18 07:07:18,925][starVLA.training.train_starvla][INFO] - Step 6486, Loss: {'train/loss': 0.3491900861263275, 'train/grad_norm_pre_clip': 31.721099853515625, 'timing/data': 0.0010412251576781273, 'timing/model': 1.260105723515153, 'learning_rate/qwen_vl_interface': 6.294816980569845e-07, 'learning_rate/action_model': 6.294816980569845e-06, 'epoch': 2.38}) +[2026-06-18 07:07:23,516][starVLA.training.train_starvla][INFO] - Step 6487, Loss: {'train/loss': 0.3316870629787445, 'train/grad_norm_pre_clip': 25.332664489746094, 'timing/data': 0.005904427729547024, 'timing/model': 1.0945463916286826, 'learning_rate/qwen_vl_interface': 6.289806586970021e-07, 'learning_rate/action_model': 6.289806586970021e-06, 'epoch': 2.38}) +[2026-06-18 07:07:28,256][starVLA.training.train_starvla][INFO] - Step 6488, Loss: {'train/loss': 0.11536344140768051, 'train/grad_norm_pre_clip': 11.724532127380371, 'timing/data': 0.0009126607328653336, 'timing/model': 1.1416687984019518, 'learning_rate/qwen_vl_interface': 6.284805772793011e-07, 'learning_rate/action_model': 6.284805772793011e-06, 'epoch': 2.38}) +[2026-06-18 07:07:34,912][starVLA.training.train_starvla][INFO] - Step 6489, Loss: {'train/loss': 0.14877453446388245, 'train/grad_norm_pre_clip': 20.590682983398438, 'timing/data': 0.0012558866292238235, 'timing/model': 1.1409081099554896, 'learning_rate/qwen_vl_interface': 6.279814539075499e-07, 'learning_rate/action_model': 6.2798145390754985e-06, 'epoch': 2.38}) +[2026-06-18 07:07:41,652][starVLA.training.train_starvla][INFO] - Step 6490, Loss: {'train/loss': 0.21704065799713135, 'train/grad_norm_pre_clip': 33.59572219848633, 'timing/data': 0.0012311656028032303, 'timing/model': 1.4414582066237926, 'learning_rate/qwen_vl_interface': 6.274832886852172e-07, 'learning_rate/action_model': 6.2748328868521724e-06, 'epoch': 2.38}) +[2026-06-18 07:07:48,972][starVLA.training.train_starvla][INFO] - Step 6491, Loss: {'train/loss': 0.4944663643836975, 'train/grad_norm_pre_clip': 34.636436462402344, 'timing/data': 0.0012261783704161644, 'timing/model': 1.075227478519082, 'learning_rate/qwen_vl_interface': 6.269860817155721e-07, 'learning_rate/action_model': 6.269860817155721e-06, 'epoch': 2.38}) +[2026-06-18 07:07:56,743][starVLA.training.train_starvla][INFO] - Step 6492, Loss: {'train/loss': 0.3046802282333374, 'train/grad_norm_pre_clip': 34.5963134765625, 'timing/data': 0.0009180782362818718, 'timing/model': 1.198043018579483, 'learning_rate/qwen_vl_interface': 6.264898331016865e-07, 'learning_rate/action_model': 6.264898331016865e-06, 'epoch': 2.38}) +[2026-06-18 07:08:02,540][starVLA.training.train_starvla][INFO] - Step 6493, Loss: {'train/loss': 0.36083051562309265, 'train/grad_norm_pre_clip': 34.043060302734375, 'timing/data': 0.03934160713106394, 'timing/model': 2.0383174987509847, 'learning_rate/qwen_vl_interface': 6.259945429464343e-07, 'learning_rate/action_model': 6.259945429464342e-06, 'epoch': 2.38}) +[2026-06-18 07:08:15,914][starVLA.training.train_starvla][INFO] - Step 6494, Loss: {'train/loss': 0.07542981952428818, 'train/grad_norm_pre_clip': 9.449995994567871, 'timing/data': 0.0010013310238718987, 'timing/model': 2.659531098790467, 'learning_rate/qwen_vl_interface': 6.255002113524885e-07, 'learning_rate/action_model': 6.255002113524886e-06, 'epoch': 2.38}) +[2026-06-18 07:08:20,325][starVLA.training.train_starvla][INFO] - Step 6495, Loss: {'train/loss': 0.19808316230773926, 'train/grad_norm_pre_clip': 21.836963653564453, 'timing/data': 0.0012334510684013367, 'timing/model': 1.0760457003489137, 'learning_rate/qwen_vl_interface': 6.250068384223255e-07, 'learning_rate/action_model': 6.250068384223255e-06, 'epoch': 2.38}) +[2026-06-18 07:08:25,324][starVLA.training.train_starvla][INFO] - Step 6496, Loss: {'train/loss': 0.3996078670024872, 'train/grad_norm_pre_clip': 37.15884017944336, 'timing/data': 0.0016084546223282814, 'timing/model': 1.0803662249818444, 'learning_rate/qwen_vl_interface': 6.245144242582214e-07, 'learning_rate/action_model': 6.245144242582214e-06, 'epoch': 2.38}) +[2026-06-18 07:08:30,030][starVLA.training.train_starvla][INFO] - Step 6497, Loss: {'train/loss': 0.09883172810077667, 'train/grad_norm_pre_clip': 16.458341598510742, 'timing/data': 0.0008780239149928093, 'timing/model': 1.0738393776118755, 'learning_rate/qwen_vl_interface': 6.240229689622546e-07, 'learning_rate/action_model': 6.240229689622545e-06, 'epoch': 2.38}) +[2026-06-18 07:08:39,228][starVLA.training.train_starvla][INFO] - Step 6498, Loss: {'train/loss': 0.4231656789779663, 'train/grad_norm_pre_clip': 41.175132751464844, 'timing/data': 0.0010301247239112854, 'timing/model': 1.0899991346523166, 'learning_rate/qwen_vl_interface': 6.235324726363042e-07, 'learning_rate/action_model': 6.235324726363041e-06, 'epoch': 2.38}) +[2026-06-18 07:08:43,864][starVLA.training.train_starvla][INFO] - Step 6499, Loss: {'train/loss': 0.17615266144275665, 'train/grad_norm_pre_clip': 18.986082077026367, 'timing/data': 0.0011054053902626038, 'timing/model': 1.426739314571023, 'learning_rate/qwen_vl_interface': 6.230429353820503e-07, 'learning_rate/action_model': 6.230429353820503e-06, 'epoch': 2.38}) +[2026-06-18 07:13:23,338][starVLA.training.train_starvla][INFO] - Step 6500, Loss: {'train/loss': 0.1720941811800003, 'train/grad_norm_pre_clip': 12.680456161499023, 'eval/loss': 0.5272500514984131, 'eval/latency_2/loss': 0.5272500514984131, 'eval/demon_fire_precision': 0.9628535246939637, 'eval/demon_fire_recall': 0.9543933054393305, 'eval/demon_fire_f1': 0.9586047488968271, 'eval/demon_fire_rate_model': 0.6591541457985531, 'eval/demon_fire_rate_teacher': 0.6649972175848636, 'eval/demon_left_precision': 0.9063829787234042, 'eval/demon_left_recall': 0.8968421052631579, 'eval/demon_left_f1': 0.9015873015873016, 'eval/demon_left_rate_model': 0.2615470228158041, 'eval/demon_left_rate_teacher': 0.2643294379521425, 'eval/demon_right_precision': 0.9121245828698554, 'eval/demon_right_recall': 0.9265536723163842, 'eval/demon_right_f1': 0.9192825112107624, 'eval/demon_right_rate_model': 0.25013912075681694, 'eval/demon_right_rate_teacher': 0.24624373956594323, 'eval/demon_move_f1': 0.910434906399032, 'eval/demon_cc_f1': 0.9436269526697825, 'eval/latency_2/demon_fire_precision': 0.9628535246939637, 'eval/latency_2/demon_fire_recall': 0.9543933054393305, 'eval/latency_2/demon_fire_f1': 0.9586047488968271, 'eval/latency_2/demon_fire_rate_model': 0.6591541457985531, 'eval/latency_2/demon_fire_rate_teacher': 0.6649972175848636, 'eval/latency_2/demon_left_precision': 0.9063829787234042, 'eval/latency_2/demon_left_recall': 0.8968421052631579, 'eval/latency_2/demon_left_f1': 0.9015873015873016, 'eval/latency_2/demon_left_rate_model': 0.2615470228158041, 'eval/latency_2/demon_left_rate_teacher': 0.2643294379521425, 'eval/latency_2/demon_right_precision': 0.9121245828698554, 'eval/latency_2/demon_right_recall': 0.9265536723163842, 'eval/latency_2/demon_right_f1': 0.9192825112107624, 'eval/latency_2/demon_right_rate_model': 0.25013912075681694, 'eval/latency_2/demon_right_rate_teacher': 0.24624373956594323, 'eval/latency_2/demon_move_f1': 0.910434906399032, 'eval/latency_2/demon_cc_f1': 0.9436269526697825, 'timing/data': 0.0010748347267508507, 'timing/model': 1.3286985382437706, 'learning_rate/qwen_vl_interface': 6.225543573009747e-07, 'learning_rate/action_model': 6.225543573009748e-06, 'epoch': 2.38}) +[2026-06-18 07:13:23,340][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_6500_state +[2026-06-18 07:13:23,365][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 07:13:32,961][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_6500_state/model.safetensors +[2026-06-18 07:13:48,291][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_6500_state/optimizer.bin +[2026-06-18 07:13:48,292][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_6500_state/sampler.bin +[2026-06-18 07:13:48,293][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_6500_state/sampler_1.bin +[2026-06-18 07:13:48,296][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_6500_state/random_states_0.pkl +[2026-06-18 07:13:48,300][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 6500 +[2026-06-18 07:13:48,301][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 07:13:48,318][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 07:13:53,011][starVLA.training.train_starvla][INFO] - Step 6501, Loss: {'train/loss': 0.33085235953330994, 'train/grad_norm_pre_clip': 23.429149627685547, 'timing/data': 0.0015002796426415443, 'timing/model': 1.4894235832616687, 'learning_rate/qwen_vl_interface': 6.2206673849436e-07, 'learning_rate/action_model': 6.2206673849436e-06, 'epoch': 2.38}) +[2026-06-18 07:14:04,145][starVLA.training.train_starvla][INFO] - Step 6502, Loss: {'train/loss': 0.14197281002998352, 'train/grad_norm_pre_clip': 12.194063186645508, 'timing/data': 0.0008452245965600014, 'timing/model': 4.604506875388324, 'learning_rate/qwen_vl_interface': 6.215800790632905e-07, 'learning_rate/action_model': 6.215800790632904e-06, 'epoch': 2.38}) +[2026-06-18 07:14:12,151][starVLA.training.train_starvla][INFO] - Step 6503, Loss: {'train/loss': 0.301284521818161, 'train/grad_norm_pre_clip': 22.927928924560547, 'timing/data': 0.0014474131166934967, 'timing/model': 1.0864275908097625, 'learning_rate/qwen_vl_interface': 6.210943791086506e-07, 'learning_rate/action_model': 6.210943791086506e-06, 'epoch': 2.38}) +[2026-06-18 07:14:20,321][starVLA.training.train_starvla][INFO] - Step 6504, Loss: {'train/loss': 0.2808486223220825, 'train/grad_norm_pre_clip': 26.285070419311523, 'timing/data': 0.0009426968172192574, 'timing/model': 4.612137953750789, 'learning_rate/qwen_vl_interface': 6.206096387311278e-07, 'learning_rate/action_model': 6.206096387311277e-06, 'epoch': 2.38}) +[2026-06-18 07:14:30,746][starVLA.training.train_starvla][INFO] - Step 6505, Loss: {'train/loss': 0.5190053582191467, 'train/grad_norm_pre_clip': 28.76534652709961, 'timing/data': 0.003265450708568096, 'timing/model': 1.5326605923473835, 'learning_rate/qwen_vl_interface': 6.201258580312078e-07, 'learning_rate/action_model': 6.201258580312077e-06, 'epoch': 2.38}) +[2026-06-18 07:14:35,512][starVLA.training.train_starvla][INFO] - Step 6506, Loss: {'train/loss': 0.19351989030838013, 'train/grad_norm_pre_clip': 22.399999618530273, 'timing/data': 0.003382212482392788, 'timing/model': 1.3617789736017585, 'learning_rate/qwen_vl_interface': 6.196430371091792e-07, 'learning_rate/action_model': 6.196430371091792e-06, 'epoch': 2.38}) +[2026-06-18 07:14:40,419][starVLA.training.train_starvla][INFO] - Step 6507, Loss: {'train/loss': 0.494743287563324, 'train/grad_norm_pre_clip': 28.497112274169922, 'timing/data': 0.0007809540256857872, 'timing/model': 1.5216081431135535, 'learning_rate/qwen_vl_interface': 6.191611760651312e-07, 'learning_rate/action_model': 6.191611760651312e-06, 'epoch': 2.38}) +[2026-06-18 07:14:45,181][starVLA.training.train_starvla][INFO] - Step 6508, Loss: {'train/loss': 0.23235741257667542, 'train/grad_norm_pre_clip': 13.735060691833496, 'timing/data': 0.0013961205258965492, 'timing/model': 1.063182845711708, 'learning_rate/qwen_vl_interface': 6.18680274998955e-07, 'learning_rate/action_model': 6.186802749989549e-06, 'epoch': 2.38}) +[2026-06-18 07:14:57,008][starVLA.training.train_starvla][INFO] - Step 6509, Loss: {'train/loss': 0.26137569546699524, 'train/grad_norm_pre_clip': 34.990516662597656, 'timing/data': 0.0011118464171886444, 'timing/model': 3.0204194942489266, 'learning_rate/qwen_vl_interface': 6.182003340103413e-07, 'learning_rate/action_model': 6.1820033401034135e-06, 'epoch': 2.38}) +[2026-06-18 07:15:01,892][starVLA.training.train_starvla][INFO] - Step 6510, Loss: {'train/loss': 0.36620938777923584, 'train/grad_norm_pre_clip': 42.38960647583008, 'timing/data': 0.0014339424669742584, 'timing/model': 1.321423213928938, 'learning_rate/qwen_vl_interface': 6.177213531987819e-07, 'learning_rate/action_model': 6.177213531987819e-06, 'epoch': 2.38}) +[2026-06-18 07:15:13,626][starVLA.training.train_starvla][INFO] - Step 6511, Loss: {'train/loss': 0.19988225400447845, 'train/grad_norm_pre_clip': 20.52505874633789, 'timing/data': 0.0008600754663348198, 'timing/model': 5.175086909905076, 'learning_rate/qwen_vl_interface': 6.172433326635714e-07, 'learning_rate/action_model': 6.172433326635713e-06, 'epoch': 2.38}) +[2026-06-18 07:15:19,920][starVLA.training.train_starvla][INFO] - Step 6512, Loss: {'train/loss': 0.24550482630729675, 'train/grad_norm_pre_clip': 19.199390411376953, 'timing/data': 0.001273476518690586, 'timing/model': 1.536661446094513, 'learning_rate/qwen_vl_interface': 6.167662725038025e-07, 'learning_rate/action_model': 6.1676627250380255e-06, 'epoch': 2.38}) +[2026-06-18 07:15:24,461][starVLA.training.train_starvla][INFO] - Step 6513, Loss: {'train/loss': 0.19000351428985596, 'train/grad_norm_pre_clip': 22.007450103759766, 'timing/data': 0.0013754293322563171, 'timing/model': 1.075797775760293, 'learning_rate/qwen_vl_interface': 6.16290172818371e-07, 'learning_rate/action_model': 6.16290172818371e-06, 'epoch': 2.38}) +[2026-06-18 07:15:29,807][starVLA.training.train_starvla][INFO] - Step 6514, Loss: {'train/loss': 0.12615716457366943, 'train/grad_norm_pre_clip': 22.79262351989746, 'timing/data': 0.001125928945839405, 'timing/model': 1.2449568435549736, 'learning_rate/qwen_vl_interface': 6.158150337059725e-07, 'learning_rate/action_model': 6.158150337059725e-06, 'epoch': 2.39}) +[2026-06-18 07:15:34,952][starVLA.training.train_starvla][INFO] - Step 6515, Loss: {'train/loss': 0.05199259892106056, 'train/grad_norm_pre_clip': 10.45648193359375, 'timing/data': 0.0009521795436739922, 'timing/model': 1.3711456265300512, 'learning_rate/qwen_vl_interface': 6.153408552651045e-07, 'learning_rate/action_model': 6.153408552651045e-06, 'epoch': 2.39}) +[2026-06-18 07:15:50,910][starVLA.training.train_starvla][INFO] - Step 6516, Loss: {'train/loss': 0.06236618757247925, 'train/grad_norm_pre_clip': 10.027813911437988, 'timing/data': 0.0008982326835393906, 'timing/model': 5.542275057174265, 'learning_rate/qwen_vl_interface': 6.14867637594064e-07, 'learning_rate/action_model': 6.14867637594064e-06, 'epoch': 2.39}) +[2026-06-18 07:15:58,336][starVLA.training.train_starvla][INFO] - Step 6517, Loss: {'train/loss': 0.1961246132850647, 'train/grad_norm_pre_clip': 21.92152214050293, 'timing/data': 0.0012914873659610748, 'timing/model': 1.102141061797738, 'learning_rate/qwen_vl_interface': 6.143953807909497e-07, 'learning_rate/action_model': 6.143953807909496e-06, 'epoch': 2.39}) +[2026-06-18 07:16:03,633][starVLA.training.train_starvla][INFO] - Step 6518, Loss: {'train/loss': 0.21811753511428833, 'train/grad_norm_pre_clip': 24.20357322692871, 'timing/data': 0.0011821519583463669, 'timing/model': 1.2184980073943734, 'learning_rate/qwen_vl_interface': 6.139240849536616e-07, 'learning_rate/action_model': 6.1392408495366156e-06, 'epoch': 2.39}) +[2026-06-18 07:16:09,256][starVLA.training.train_starvla][INFO] - Step 6519, Loss: {'train/loss': 0.184450164437294, 'train/grad_norm_pre_clip': 14.914583206176758, 'timing/data': 0.0007695099338889122, 'timing/model': 1.4044223129749298, 'learning_rate/qwen_vl_interface': 6.134537501798993e-07, 'learning_rate/action_model': 6.1345375017989925e-06, 'epoch': 2.39}) +[2026-06-18 07:16:17,169][starVLA.training.train_starvla][INFO] - Step 6520, Loss: {'train/loss': 0.139746755361557, 'train/grad_norm_pre_clip': 15.030844688415527, 'timing/data': 0.001120489090681076, 'timing/model': 1.1357670361176133, 'learning_rate/qwen_vl_interface': 6.129843765671634e-07, 'learning_rate/action_model': 6.129843765671633e-06, 'epoch': 2.39}) +[2026-06-18 07:16:23,769][starVLA.training.train_starvla][INFO] - Step 6521, Loss: {'train/loss': 0.19356383383274078, 'train/grad_norm_pre_clip': 20.360485076904297, 'timing/data': 0.0010653529316186905, 'timing/model': 1.1085084304213524, 'learning_rate/qwen_vl_interface': 6.125159642127559e-07, 'learning_rate/action_model': 6.125159642127558e-06, 'epoch': 2.39}) +[2026-06-18 07:16:28,624][starVLA.training.train_starvla][INFO] - Step 6522, Loss: {'train/loss': 0.5361614227294922, 'train/grad_norm_pre_clip': 33.6019401550293, 'timing/data': 0.001973832957446575, 'timing/model': 1.65098890196532, 'learning_rate/qwen_vl_interface': 6.120485132137793e-07, 'learning_rate/action_model': 6.120485132137794e-06, 'epoch': 2.39}) +[2026-06-18 07:16:36,730][starVLA.training.train_starvla][INFO] - Step 6523, Loss: {'train/loss': 0.25222083926200867, 'train/grad_norm_pre_clip': 25.897480010986328, 'timing/data': 0.0010505877435207367, 'timing/model': 1.1103368336334825, 'learning_rate/qwen_vl_interface': 6.115820236671369e-07, 'learning_rate/action_model': 6.115820236671369e-06, 'epoch': 2.39}) +[2026-06-18 07:16:49,712][starVLA.training.train_starvla][INFO] - Step 6524, Loss: {'train/loss': 0.36524444818496704, 'train/grad_norm_pre_clip': 27.85869598388672, 'timing/data': 0.0011153938248753548, 'timing/model': 1.9152848636731505, 'learning_rate/qwen_vl_interface': 6.111164956695316e-07, 'learning_rate/action_model': 6.111164956695316e-06, 'epoch': 2.39}) +[2026-06-18 07:16:57,280][starVLA.training.train_starvla][INFO] - Step 6525, Loss: {'train/loss': 0.21930424869060516, 'train/grad_norm_pre_clip': 18.942338943481445, 'timing/data': 0.0014920607209205627, 'timing/model': 2.725645932368934, 'learning_rate/qwen_vl_interface': 6.106519293174684e-07, 'learning_rate/action_model': 6.106519293174684e-06, 'epoch': 2.39}) +[2026-06-18 07:17:07,998][starVLA.training.train_starvla][INFO] - Step 6526, Loss: {'train/loss': 0.36520105600357056, 'train/grad_norm_pre_clip': 34.70436096191406, 'timing/data': 0.0009125098586082458, 'timing/model': 4.861313482746482, 'learning_rate/qwen_vl_interface': 6.101883247072525e-07, 'learning_rate/action_model': 6.1018832470725245e-06, 'epoch': 2.39}) +[2026-06-18 07:17:18,004][starVLA.training.train_starvla][INFO] - Step 6527, Loss: {'train/loss': 0.07115763425827026, 'train/grad_norm_pre_clip': 11.661178588867188, 'timing/data': 0.0010849563404917717, 'timing/model': 6.4939926927909255, 'learning_rate/qwen_vl_interface': 6.097256819349899e-07, 'learning_rate/action_model': 6.0972568193498985e-06, 'epoch': 2.39}) +[2026-06-18 07:17:24,948][starVLA.training.train_starvla][INFO] - Step 6528, Loss: {'train/loss': 0.29919350147247314, 'train/grad_norm_pre_clip': 22.478496551513672, 'timing/data': 0.0033020274713635445, 'timing/model': 3.6951540391892195, 'learning_rate/qwen_vl_interface': 6.092640010965864e-07, 'learning_rate/action_model': 6.092640010965864e-06, 'epoch': 2.39}) +[2026-06-18 07:17:32,854][starVLA.training.train_starvla][INFO] - Step 6529, Loss: {'train/loss': 0.2038165181875229, 'train/grad_norm_pre_clip': 23.687938690185547, 'timing/data': 0.001185677945613861, 'timing/model': 1.9095703335478902, 'learning_rate/qwen_vl_interface': 6.088032822877488e-07, 'learning_rate/action_model': 6.088032822877488e-06, 'epoch': 2.39}) +[2026-06-18 07:17:44,225][starVLA.training.train_starvla][INFO] - Step 6530, Loss: {'train/loss': 0.2986689805984497, 'train/grad_norm_pre_clip': 22.747587203979492, 'timing/data': 0.014223537407815456, 'timing/model': 5.568068199791014, 'learning_rate/qwen_vl_interface': 6.083435256039854e-07, 'learning_rate/action_model': 6.083435256039854e-06, 'epoch': 2.39}) +[2026-06-18 07:17:52,938][starVLA.training.train_starvla][INFO] - Step 6531, Loss: {'train/loss': 0.28188812732696533, 'train/grad_norm_pre_clip': 16.57988166809082, 'timing/data': 0.015134069137275219, 'timing/model': 1.2686169631779194, 'learning_rate/qwen_vl_interface': 6.078847311406033e-07, 'learning_rate/action_model': 6.078847311406032e-06, 'epoch': 2.39}) +[2026-06-18 07:17:59,655][starVLA.training.train_starvla][INFO] - Step 6532, Loss: {'train/loss': 0.3480653762817383, 'train/grad_norm_pre_clip': 31.303234100341797, 'timing/data': 0.013012001290917397, 'timing/model': 1.5810640882700682, 'learning_rate/qwen_vl_interface': 6.07426898992712e-07, 'learning_rate/action_model': 6.074268989927119e-06, 'epoch': 2.39}) +[2026-06-18 07:18:04,080][starVLA.training.train_starvla][INFO] - Step 6533, Loss: {'train/loss': 0.12281093746423721, 'train/grad_norm_pre_clip': 18.76703453063965, 'timing/data': 0.011492478661239147, 'timing/model': 1.313912925310433, 'learning_rate/qwen_vl_interface': 6.069700292552202e-07, 'learning_rate/action_model': 6.069700292552202e-06, 'epoch': 2.39}) +[2026-06-18 07:18:09,458][starVLA.training.train_starvla][INFO] - Step 6534, Loss: {'train/loss': 0.16365179419517517, 'train/grad_norm_pre_clip': 15.240315437316895, 'timing/data': 0.012234248220920563, 'timing/model': 1.0537631455808878, 'learning_rate/qwen_vl_interface': 6.065141220228372e-07, 'learning_rate/action_model': 6.065141220228372e-06, 'epoch': 2.39}) +[2026-06-18 07:18:14,376][starVLA.training.train_starvla][INFO] - Step 6535, Loss: {'train/loss': 0.42124611139297485, 'train/grad_norm_pre_clip': 30.525775909423828, 'timing/data': 0.01789642684161663, 'timing/model': 1.0653496524319053, 'learning_rate/qwen_vl_interface': 6.060591773900732e-07, 'learning_rate/action_model': 6.060591773900733e-06, 'epoch': 2.39}) +[2026-06-18 07:18:19,309][starVLA.training.train_starvla][INFO] - Step 6536, Loss: {'train/loss': 0.1380884349346161, 'train/grad_norm_pre_clip': 17.867944717407227, 'timing/data': 0.01331536378711462, 'timing/model': 1.5186243886128068, 'learning_rate/qwen_vl_interface': 6.056051954512388e-07, 'learning_rate/action_model': 6.0560519545123875e-06, 'epoch': 2.39}) +[2026-06-18 07:18:23,615][starVLA.training.train_starvla][INFO] - Step 6537, Loss: {'train/loss': 0.17126041650772095, 'train/grad_norm_pre_clip': 21.944665908813477, 'timing/data': 0.011556967161595821, 'timing/model': 1.2841902496293187, 'learning_rate/qwen_vl_interface': 6.051521763004455e-07, 'learning_rate/action_model': 6.051521763004456e-06, 'epoch': 2.39}) +[2026-06-18 07:18:27,839][starVLA.training.train_starvla][INFO] - Step 6538, Loss: {'train/loss': 0.08935464173555374, 'train/grad_norm_pre_clip': 15.2063570022583, 'timing/data': 0.013578539714217186, 'timing/model': 1.0858234697952867, 'learning_rate/qwen_vl_interface': 6.047001200316038e-07, 'learning_rate/action_model': 6.047001200316037e-06, 'epoch': 2.39}) +[2026-06-18 07:18:32,417][starVLA.training.train_starvla][INFO] - Step 6539, Loss: {'train/loss': 0.18219426274299622, 'train/grad_norm_pre_clip': 38.43207931518555, 'timing/data': 0.010576185770332813, 'timing/model': 1.0543063906952739, 'learning_rate/qwen_vl_interface': 6.042490267384255e-07, 'learning_rate/action_model': 6.0424902673842545e-06, 'epoch': 2.39}) +[2026-06-18 07:18:37,218][starVLA.training.train_starvla][INFO] - Step 6540, Loss: {'train/loss': 0.02919742837548256, 'train/grad_norm_pre_clip': 6.61201810836792, 'timing/data': 0.013690975494682789, 'timing/model': 1.2411743234843016, 'learning_rate/qwen_vl_interface': 6.03798896514423e-07, 'learning_rate/action_model': 6.0379889651442295e-06, 'epoch': 2.39}) +[2026-06-18 07:18:42,010][starVLA.training.train_starvla][INFO] - Step 6541, Loss: {'train/loss': 0.0580347403883934, 'train/grad_norm_pre_clip': 8.126554489135742, 'timing/data': 0.011382718570530415, 'timing/model': 1.0477630710229278, 'learning_rate/qwen_vl_interface': 6.033497294529093e-07, 'learning_rate/action_model': 6.033497294529093e-06, 'epoch': 2.4}) +[2026-06-18 07:18:47,016][starVLA.training.train_starvla][INFO] - Step 6542, Loss: {'train/loss': 0.1837826520204544, 'train/grad_norm_pre_clip': 25.648820877075195, 'timing/data': 0.015573940239846706, 'timing/model': 1.6267317105084658, 'learning_rate/qwen_vl_interface': 6.029015256469967e-07, 'learning_rate/action_model': 6.0290152564699666e-06, 'epoch': 2.4}) +[2026-06-18 07:18:51,454][starVLA.training.train_starvla][INFO] - Step 6543, Loss: {'train/loss': 0.30832746624946594, 'train/grad_norm_pre_clip': 105.11524200439453, 'timing/data': 0.012730716727674007, 'timing/model': 1.0216865977272391, 'learning_rate/qwen_vl_interface': 6.024542851895976e-07, 'learning_rate/action_model': 6.0245428518959756e-06, 'epoch': 2.4}) +[2026-06-18 07:18:56,088][starVLA.training.train_starvla][INFO] - Step 6544, Loss: {'train/loss': 0.6814870834350586, 'train/grad_norm_pre_clip': 47.92702865600586, 'timing/data': 0.013710752129554749, 'timing/model': 1.0345408227294683, 'learning_rate/qwen_vl_interface': 6.020080081734267e-07, 'learning_rate/action_model': 6.020080081734266e-06, 'epoch': 2.4}) +[2026-06-18 07:19:00,587][starVLA.training.train_starvla][INFO] - Step 6545, Loss: {'train/loss': 0.19570665061473846, 'train/grad_norm_pre_clip': 16.600658416748047, 'timing/data': 0.01513687614351511, 'timing/model': 1.0530865360051394, 'learning_rate/qwen_vl_interface': 6.015626946909974e-07, 'learning_rate/action_model': 6.015626946909973e-06, 'epoch': 2.4}) +[2026-06-18 07:19:05,874][starVLA.training.train_starvla][INFO] - Step 6546, Loss: {'train/loss': 0.23892003297805786, 'train/grad_norm_pre_clip': 23.854034423828125, 'timing/data': 0.01286306045949459, 'timing/model': 1.975692960433662, 'learning_rate/qwen_vl_interface': 6.011183448346233e-07, 'learning_rate/action_model': 6.011183448346233e-06, 'epoch': 2.4}) +[2026-06-18 07:19:11,133][starVLA.training.train_starvla][INFO] - Step 6547, Loss: {'train/loss': 0.0931912213563919, 'train/grad_norm_pre_clip': 19.878515243530273, 'timing/data': 0.010771899484097958, 'timing/model': 1.8390573235228658, 'learning_rate/qwen_vl_interface': 6.006749586964183e-07, 'learning_rate/action_model': 6.006749586964183e-06, 'epoch': 2.4}) +[2026-06-18 07:19:20,820][starVLA.training.train_starvla][INFO] - Step 6548, Loss: {'train/loss': 0.2090083211660385, 'train/grad_norm_pre_clip': 26.171342849731445, 'timing/data': 0.014146357774734497, 'timing/model': 1.144707614555955, 'learning_rate/qwen_vl_interface': 6.002325363682974e-07, 'learning_rate/action_model': 6.0023253636829736e-06, 'epoch': 2.4}) +[2026-06-18 07:19:34,504][starVLA.training.train_starvla][INFO] - Step 6549, Loss: {'train/loss': 0.16067232191562653, 'train/grad_norm_pre_clip': 27.907730102539062, 'timing/data': 0.013773842714726925, 'timing/model': 1.06399055570364, 'learning_rate/qwen_vl_interface': 5.997910779419756e-07, 'learning_rate/action_model': 5.9979107794197565e-06, 'epoch': 2.4}) +[2026-06-18 07:19:45,644][starVLA.training.train_starvla][INFO] - Step 6550, Loss: {'train/loss': 0.09673721343278885, 'train/grad_norm_pre_clip': 13.501651763916016, 'timing/data': 0.024541174061596394, 'timing/model': 2.606666929088533, 'learning_rate/qwen_vl_interface': 5.993505835089661e-07, 'learning_rate/action_model': 5.993505835089661e-06, 'epoch': 2.4}) +[2026-06-18 07:19:54,487][starVLA.training.train_starvla][INFO] - Step 6551, Loss: {'train/loss': 0.0610695444047451, 'train/grad_norm_pre_clip': 12.352760314941406, 'timing/data': 0.012523981742560863, 'timing/model': 1.205066661350429, 'learning_rate/qwen_vl_interface': 5.98911053160586e-07, 'learning_rate/action_model': 5.989110531605859e-06, 'epoch': 2.4}) +[2026-06-18 07:20:14,899][starVLA.training.train_starvla][INFO] - Step 6552, Loss: {'train/loss': 0.10038590431213379, 'train/grad_norm_pre_clip': 17.80228042602539, 'timing/data': 0.012575970031321049, 'timing/model': 6.213620847091079, 'learning_rate/qwen_vl_interface': 5.984724869879494e-07, 'learning_rate/action_model': 5.984724869879493e-06, 'epoch': 2.4}) +[2026-06-18 07:20:23,901][starVLA.training.train_starvla][INFO] - Step 6553, Loss: {'train/loss': 0.25454363226890564, 'train/grad_norm_pre_clip': 29.800325393676758, 'timing/data': 0.0153680220246315, 'timing/model': 5.177277895621955, 'learning_rate/qwen_vl_interface': 5.980348850819711e-07, 'learning_rate/action_model': 5.98034885081971e-06, 'epoch': 2.4}) +[2026-06-18 07:20:36,207][starVLA.training.train_starvla][INFO] - Step 6554, Loss: {'train/loss': 0.3520456850528717, 'train/grad_norm_pre_clip': 24.68538475036621, 'timing/data': 0.0121769392862916, 'timing/model': 1.2046568039804697, 'learning_rate/qwen_vl_interface': 5.975982475333664e-07, 'learning_rate/action_model': 5.975982475333663e-06, 'epoch': 2.4}) +[2026-06-18 07:20:45,437][starVLA.training.train_starvla][INFO] - Step 6555, Loss: {'train/loss': 0.20902450382709503, 'train/grad_norm_pre_clip': 15.0227632522583, 'timing/data': 0.01899889949709177, 'timing/model': 1.5108249392360449, 'learning_rate/qwen_vl_interface': 5.971625744326516e-07, 'learning_rate/action_model': 5.971625744326515e-06, 'epoch': 2.4}) +[2026-06-18 07:20:53,416][starVLA.training.train_starvla][INFO] - Step 6556, Loss: {'train/loss': 0.06800718605518341, 'train/grad_norm_pre_clip': 13.388510704040527, 'timing/data': 0.014860042370855808, 'timing/model': 1.5840326799079776, 'learning_rate/qwen_vl_interface': 5.967278658701421e-07, 'learning_rate/action_model': 5.967278658701421e-06, 'epoch': 2.4}) +[2026-06-18 07:21:00,469][starVLA.training.train_starvla][INFO] - Step 6557, Loss: {'train/loss': 0.19010300934314728, 'train/grad_norm_pre_clip': 35.145259857177734, 'timing/data': 0.020332480780780315, 'timing/model': 1.644513969309628, 'learning_rate/qwen_vl_interface': 5.962941219359527e-07, 'learning_rate/action_model': 5.962941219359527e-06, 'epoch': 2.4}) +[2026-06-18 07:21:07,576][starVLA.training.train_starvla][INFO] - Step 6558, Loss: {'train/loss': 0.1973513513803482, 'train/grad_norm_pre_clip': 22.1791934967041, 'timing/data': 0.013673373498022556, 'timing/model': 1.0483660474419594, 'learning_rate/qwen_vl_interface': 5.958613427199998e-07, 'learning_rate/action_model': 5.958613427199997e-06, 'epoch': 2.4}) +[2026-06-18 07:21:12,675][starVLA.training.train_starvla][INFO] - Step 6559, Loss: {'train/loss': 0.3061090111732483, 'train/grad_norm_pre_clip': 23.179885864257812, 'timing/data': 0.020556061528623104, 'timing/model': 1.0567065421491861, 'learning_rate/qwen_vl_interface': 5.954295283119985e-07, 'learning_rate/action_model': 5.954295283119985e-06, 'epoch': 2.4}) +[2026-06-18 07:21:17,336][starVLA.training.train_starvla][INFO] - Step 6560, Loss: {'train/loss': 0.17786656320095062, 'train/grad_norm_pre_clip': 17.88097381591797, 'timing/data': 0.020227576605975628, 'timing/model': 1.0921321483328938, 'learning_rate/qwen_vl_interface': 5.949986788014643e-07, 'learning_rate/action_model': 5.949986788014643e-06, 'epoch': 2.4}) +[2026-06-18 07:21:21,897][starVLA.training.train_starvla][INFO] - Step 6561, Loss: {'train/loss': 0.39559659361839294, 'train/grad_norm_pre_clip': 30.59119987487793, 'timing/data': 0.01055485662072897, 'timing/model': 1.106398779898882, 'learning_rate/qwen_vl_interface': 5.94568794277713e-07, 'learning_rate/action_model': 5.94568794277713e-06, 'epoch': 2.4}) +[2026-06-18 07:21:26,663][starVLA.training.train_starvla][INFO] - Step 6562, Loss: {'train/loss': 0.06904106587171555, 'train/grad_norm_pre_clip': 10.822017669677734, 'timing/data': 0.015293866395950317, 'timing/model': 1.5943003259599209, 'learning_rate/qwen_vl_interface': 5.941398748298603e-07, 'learning_rate/action_model': 5.941398748298603e-06, 'epoch': 2.4}) +[2026-06-18 07:21:30,900][starVLA.training.train_starvla][INFO] - Step 6563, Loss: {'train/loss': 0.19920693337917328, 'train/grad_norm_pre_clip': 20.48785972595215, 'timing/data': 0.011333698406815529, 'timing/model': 1.122005625627935, 'learning_rate/qwen_vl_interface': 5.937119205468217e-07, 'learning_rate/action_model': 5.9371192054682165e-06, 'epoch': 2.4}) +[2026-06-18 07:21:35,390][starVLA.training.train_starvla][INFO] - Step 6564, Loss: {'train/loss': 0.419596403837204, 'train/grad_norm_pre_clip': 25.428815841674805, 'timing/data': 0.022750074043869972, 'timing/model': 1.0514953043311834, 'learning_rate/qwen_vl_interface': 5.93284931517312e-07, 'learning_rate/action_model': 5.93284931517312e-06, 'epoch': 2.4}) +[2026-06-18 07:21:39,946][starVLA.training.train_starvla][INFO] - Step 6565, Loss: {'train/loss': 0.4035104513168335, 'train/grad_norm_pre_clip': 25.99606704711914, 'timing/data': 0.012819238938391209, 'timing/model': 1.1517532132565975, 'learning_rate/qwen_vl_interface': 5.928589078298468e-07, 'learning_rate/action_model': 5.928589078298468e-06, 'epoch': 2.4}) +[2026-06-18 07:21:44,404][starVLA.training.train_starvla][INFO] - Step 6566, Loss: {'train/loss': 0.2568218410015106, 'train/grad_norm_pre_clip': 22.11322784423828, 'timing/data': 0.013728611171245575, 'timing/model': 1.341908228583634, 'learning_rate/qwen_vl_interface': 5.924338495727416e-07, 'learning_rate/action_model': 5.9243384957274155e-06, 'epoch': 2.4}) +[2026-06-18 07:21:49,462][starVLA.training.train_starvla][INFO] - Step 6567, Loss: {'train/loss': 0.20158696174621582, 'train/grad_norm_pre_clip': 17.85931968688965, 'timing/data': 0.012759415432810783, 'timing/model': 1.5641414085403085, 'learning_rate/qwen_vl_interface': 5.92009756834111e-07, 'learning_rate/action_model': 5.9200975683411094e-06, 'epoch': 2.4}) +[2026-06-18 07:21:53,704][starVLA.training.train_starvla][INFO] - Step 6568, Loss: {'train/loss': 0.06322164833545685, 'train/grad_norm_pre_clip': 9.169120788574219, 'timing/data': 0.01226572971791029, 'timing/model': 1.1185713857412338, 'learning_rate/qwen_vl_interface': 5.915866297018697e-07, 'learning_rate/action_model': 5.915866297018696e-06, 'epoch': 2.41}) +[2026-06-18 07:21:58,517][starVLA.training.train_starvla][INFO] - Step 6569, Loss: {'train/loss': 0.09935495257377625, 'train/grad_norm_pre_clip': 15.896856307983398, 'timing/data': 0.01582159474492073, 'timing/model': 1.2368698548525572, 'learning_rate/qwen_vl_interface': 5.911644682637329e-07, 'learning_rate/action_model': 5.911644682637329e-06, 'epoch': 2.41}) +[2026-06-18 07:22:03,565][starVLA.training.train_starvla][INFO] - Step 6570, Loss: {'train/loss': 0.06627975404262543, 'train/grad_norm_pre_clip': 11.234800338745117, 'timing/data': 0.01274915412068367, 'timing/model': 1.1809354200959206, 'learning_rate/qwen_vl_interface': 5.907432726072149e-07, 'learning_rate/action_model': 5.9074327260721485e-06, 'epoch': 2.41}) +[2026-06-18 07:22:08,328][starVLA.training.train_starvla][INFO] - Step 6571, Loss: {'train/loss': 0.12128742784261703, 'train/grad_norm_pre_clip': 23.59564781188965, 'timing/data': 0.013501769863069057, 'timing/model': 1.3279774188995361, 'learning_rate/qwen_vl_interface': 5.903230428196305e-07, 'learning_rate/action_model': 5.9032304281963045e-06, 'epoch': 2.41}) +[2026-06-18 07:22:19,578][starVLA.training.train_starvla][INFO] - Step 6572, Loss: {'train/loss': 0.2172265201807022, 'train/grad_norm_pre_clip': 22.161428451538086, 'timing/data': 0.010911207646131516, 'timing/model': 2.4143433636054397, 'learning_rate/qwen_vl_interface': 5.899037789880929e-07, 'learning_rate/action_model': 5.899037789880929e-06, 'epoch': 2.41}) +[2026-06-18 07:22:26,039][starVLA.training.train_starvla][INFO] - Step 6573, Loss: {'train/loss': 0.29009002447128296, 'train/grad_norm_pre_clip': 26.147605895996094, 'timing/data': 0.01326626818627119, 'timing/model': 1.1831428641453385, 'learning_rate/qwen_vl_interface': 5.894854811995166e-07, 'learning_rate/action_model': 5.894854811995165e-06, 'epoch': 2.41}) +[2026-06-18 07:22:31,004][starVLA.training.train_starvla][INFO] - Step 6574, Loss: {'train/loss': 0.20800408720970154, 'train/grad_norm_pre_clip': 29.51154899597168, 'timing/data': 0.012937656603753567, 'timing/model': 1.0349680222570896, 'learning_rate/qwen_vl_interface': 5.890681495406152e-07, 'learning_rate/action_model': 5.890681495406151e-06, 'epoch': 2.41}) +[2026-06-18 07:22:35,896][starVLA.training.train_starvla][INFO] - Step 6575, Loss: {'train/loss': 0.08821816742420197, 'train/grad_norm_pre_clip': 11.711150169372559, 'timing/data': 0.013673006556928158, 'timing/model': 1.4041270269080997, 'learning_rate/qwen_vl_interface': 5.886517840979015e-07, 'learning_rate/action_model': 5.886517840979016e-06, 'epoch': 2.41}) +[2026-06-18 07:22:42,849][starVLA.training.train_starvla][INFO] - Step 6576, Loss: {'train/loss': 0.11948508769273758, 'train/grad_norm_pre_clip': 16.965572357177734, 'timing/data': 0.013235209509730339, 'timing/model': 1.4573811069130898, 'learning_rate/qwen_vl_interface': 5.882363849576889e-07, 'learning_rate/action_model': 5.882363849576889e-06, 'epoch': 2.41}) +[2026-06-18 07:22:55,230][starVLA.training.train_starvla][INFO] - Step 6577, Loss: {'train/loss': 0.05303328484296799, 'train/grad_norm_pre_clip': 7.190310955047607, 'timing/data': 0.012865080498158932, 'timing/model': 1.3637267891317606, 'learning_rate/qwen_vl_interface': 5.8782195220609e-07, 'learning_rate/action_model': 5.878219522060899e-06, 'epoch': 2.41}) +[2026-06-18 07:22:59,298][starVLA.training.train_starvla][INFO] - Step 6578, Loss: {'train/loss': 0.3146190047264099, 'train/grad_norm_pre_clip': 24.716655731201172, 'timing/data': 0.012161039747297764, 'timing/model': 1.0448208786547184, 'learning_rate/qwen_vl_interface': 5.874084859290172e-07, 'learning_rate/action_model': 5.874084859290171e-06, 'epoch': 2.41}) +[2026-06-18 07:23:04,199][starVLA.training.train_starvla][INFO] - Step 6579, Loss: {'train/loss': 0.062426429241895676, 'train/grad_norm_pre_clip': 23.184558868408203, 'timing/data': 0.01632886752486229, 'timing/model': 1.0563413109630346, 'learning_rate/qwen_vl_interface': 5.869959862121822e-07, 'learning_rate/action_model': 5.869959862121822e-06, 'epoch': 2.41}) +[2026-06-18 07:23:08,889][starVLA.training.train_starvla][INFO] - Step 6580, Loss: {'train/loss': 0.2777756154537201, 'train/grad_norm_pre_clip': 30.10776710510254, 'timing/data': 0.013293939642608166, 'timing/model': 1.1984408125281334, 'learning_rate/qwen_vl_interface': 5.86584453141097e-07, 'learning_rate/action_model': 5.86584453141097e-06, 'epoch': 2.41}) +[2026-06-18 07:23:13,493][starVLA.training.train_starvla][INFO] - Step 6581, Loss: {'train/loss': 0.47414645552635193, 'train/grad_norm_pre_clip': 33.32936477661133, 'timing/data': 0.0116328289732337, 'timing/model': 1.0449183359742165, 'learning_rate/qwen_vl_interface': 5.861738868010723e-07, 'learning_rate/action_model': 5.861738868010723e-06, 'epoch': 2.41}) +[2026-06-18 07:23:19,772][starVLA.training.train_starvla][INFO] - Step 6582, Loss: {'train/loss': 0.2203165590763092, 'train/grad_norm_pre_clip': 23.059465408325195, 'timing/data': 0.011677633970975876, 'timing/model': 1.6989470105618238, 'learning_rate/qwen_vl_interface': 5.857642872772198e-07, 'learning_rate/action_model': 5.8576428727721986e-06, 'epoch': 2.41}) +[2026-06-18 07:23:30,004][starVLA.training.train_starvla][INFO] - Step 6583, Loss: {'train/loss': 0.5787637829780579, 'train/grad_norm_pre_clip': 46.082618713378906, 'timing/data': 0.011259946972131729, 'timing/model': 1.3211308540776372, 'learning_rate/qwen_vl_interface': 5.853556546544491e-07, 'learning_rate/action_model': 5.8535565465444904e-06, 'epoch': 2.41}) +[2026-06-18 07:23:34,815][starVLA.training.train_starvla][INFO] - Step 6584, Loss: {'train/loss': 0.45255354046821594, 'train/grad_norm_pre_clip': 40.770267486572266, 'timing/data': 0.011361261829733849, 'timing/model': 1.1568689718842506, 'learning_rate/qwen_vl_interface': 5.849479890174708e-07, 'learning_rate/action_model': 5.8494798901747075e-06, 'epoch': 2.41}) +[2026-06-18 07:23:39,200][starVLA.training.train_starvla][INFO] - Step 6585, Loss: {'train/loss': 0.29210203886032104, 'train/grad_norm_pre_clip': 34.21987533569336, 'timing/data': 0.012361103668808937, 'timing/model': 1.0512059545144439, 'learning_rate/qwen_vl_interface': 5.845412904507948e-07, 'learning_rate/action_model': 5.845412904507948e-06, 'epoch': 2.41}) +[2026-06-18 07:23:43,942][starVLA.training.train_starvla][INFO] - Step 6586, Loss: {'train/loss': 0.09062331169843674, 'train/grad_norm_pre_clip': 11.187515258789062, 'timing/data': 0.018681525252759457, 'timing/model': 1.415674558840692, 'learning_rate/qwen_vl_interface': 5.841355590387291e-07, 'learning_rate/action_model': 5.841355590387291e-06, 'epoch': 2.41}) +[2026-06-18 07:23:48,452][starVLA.training.train_starvla][INFO] - Step 6587, Loss: {'train/loss': 0.2036074995994568, 'train/grad_norm_pre_clip': 17.842100143432617, 'timing/data': 0.011297889985144138, 'timing/model': 1.4608178297057748, 'learning_rate/qwen_vl_interface': 5.837307948653821e-07, 'learning_rate/action_model': 5.837307948653821e-06, 'epoch': 2.41}) +[2026-06-18 07:23:52,555][starVLA.training.train_starvla][INFO] - Step 6588, Loss: {'train/loss': 0.23132282495498657, 'train/grad_norm_pre_clip': 21.578384399414062, 'timing/data': 0.0121762715280056, 'timing/model': 1.0311256274580956, 'learning_rate/qwen_vl_interface': 5.833269980146631e-07, 'learning_rate/action_model': 5.83326998014663e-06, 'epoch': 2.41}) +[2026-06-18 07:23:58,564][starVLA.training.train_starvla][INFO] - Step 6589, Loss: {'train/loss': 0.1512836068868637, 'train/grad_norm_pre_clip': 18.677776336669922, 'timing/data': 0.012770002707839012, 'timing/model': 1.126295691356063, 'learning_rate/qwen_vl_interface': 5.829241685702791e-07, 'learning_rate/action_model': 5.82924168570279e-06, 'epoch': 2.41}) +[2026-06-18 07:24:06,580][starVLA.training.train_starvla][INFO] - Step 6590, Loss: {'train/loss': 0.33553653955459595, 'train/grad_norm_pre_clip': 26.193401336669922, 'timing/data': 0.012066860683262348, 'timing/model': 4.231419401243329, 'learning_rate/qwen_vl_interface': 5.825223066157365e-07, 'learning_rate/action_model': 5.8252230661573645e-06, 'epoch': 2.41}) +[2026-06-18 07:24:15,138][starVLA.training.train_starvla][INFO] - Step 6591, Loss: {'train/loss': 0.09491384774446487, 'train/grad_norm_pre_clip': 15.064543724060059, 'timing/data': 0.013206013478338718, 'timing/model': 1.2507314775139093, 'learning_rate/qwen_vl_interface': 5.821214122343423e-07, 'learning_rate/action_model': 5.821214122343422e-06, 'epoch': 2.41}) +[2026-06-18 07:24:20,395][starVLA.training.train_starvla][INFO] - Step 6592, Loss: {'train/loss': 0.24587436020374298, 'train/grad_norm_pre_clip': 19.61455535888672, 'timing/data': 0.016816851682960987, 'timing/model': 1.6194728258997202, 'learning_rate/qwen_vl_interface': 5.817214855092024e-07, 'learning_rate/action_model': 5.817214855092024e-06, 'epoch': 2.41}) +[2026-06-18 07:24:24,572][starVLA.training.train_starvla][INFO] - Step 6593, Loss: {'train/loss': 0.19272612035274506, 'train/grad_norm_pre_clip': 18.997041702270508, 'timing/data': 0.013201458379626274, 'timing/model': 1.0418182713910937, 'learning_rate/qwen_vl_interface': 5.813225265232216e-07, 'learning_rate/action_model': 5.813225265232215e-06, 'epoch': 2.41}) +[2026-06-18 07:24:29,168][starVLA.training.train_starvla][INFO] - Step 6594, Loss: {'train/loss': 0.15063050389289856, 'train/grad_norm_pre_clip': 21.88483428955078, 'timing/data': 0.012496128678321838, 'timing/model': 1.0358371008187532, 'learning_rate/qwen_vl_interface': 5.809245353591048e-07, 'learning_rate/action_model': 5.8092453535910485e-06, 'epoch': 2.41}) +[2026-06-18 07:24:37,558][starVLA.training.train_starvla][INFO] - Step 6595, Loss: {'train/loss': 0.4546886682510376, 'train/grad_norm_pre_clip': 29.77103614807129, 'timing/data': 0.01314895786345005, 'timing/model': 1.1645050281658769, 'learning_rate/qwen_vl_interface': 5.805275120993561e-07, 'learning_rate/action_model': 5.805275120993561e-06, 'epoch': 2.41}) +[2026-06-18 07:24:42,742][starVLA.training.train_starvla][INFO] - Step 6596, Loss: {'train/loss': 0.07661806046962738, 'train/grad_norm_pre_clip': 11.63866138458252, 'timing/data': 0.012554298155009747, 'timing/model': 1.6652497136965394, 'learning_rate/qwen_vl_interface': 5.801314568262787e-07, 'learning_rate/action_model': 5.801314568262787e-06, 'epoch': 2.42}) +[2026-06-18 07:24:51,572][starVLA.training.train_starvla][INFO] - Step 6597, Loss: {'train/loss': 0.22632773220539093, 'train/grad_norm_pre_clip': 24.366474151611328, 'timing/data': 0.015945514664053917, 'timing/model': 1.4125939039513469, 'learning_rate/qwen_vl_interface': 5.797363696219751e-07, 'learning_rate/action_model': 5.797363696219751e-06, 'epoch': 2.42}) +[2026-06-18 07:24:56,126][starVLA.training.train_starvla][INFO] - Step 6598, Loss: {'train/loss': 0.3274289667606354, 'train/grad_norm_pre_clip': 30.108104705810547, 'timing/data': 0.012844346463680267, 'timing/model': 1.1498436648398638, 'learning_rate/qwen_vl_interface': 5.793422505683474e-07, 'learning_rate/action_model': 5.793422505683474e-06, 'epoch': 2.42}) +[2026-06-18 07:25:00,812][starVLA.training.train_starvla][INFO] - Step 6599, Loss: {'train/loss': 0.3226470351219177, 'train/grad_norm_pre_clip': 25.395414352416992, 'timing/data': 0.01406331080943346, 'timing/model': 1.077261614613235, 'learning_rate/qwen_vl_interface': 5.789490997470978e-07, 'learning_rate/action_model': 5.789490997470978e-06, 'epoch': 2.42}) +[2026-06-18 07:29:33,925][starVLA.training.train_starvla][INFO] - Step 6600, Loss: {'train/loss': 0.3551087975502014, 'train/grad_norm_pre_clip': 34.48691940307617, 'eval/loss': 0.5458158850669861, 'eval/latency_2/loss': 0.5458158850669861, 'eval/demon_fire_precision': 0.9564853556485355, 'eval/demon_fire_recall': 0.9564853556485355, 'eval/demon_fire_f1': 0.9564853556485355, 'eval/demon_fire_rate_model': 0.6649972175848636, 'eval/demon_fire_rate_teacher': 0.6649972175848636, 'eval/demon_left_precision': 0.9075451647183846, 'eval/demon_left_recall': 0.8989473684210526, 'eval/demon_left_f1': 0.9032258064516129, 'eval/demon_left_rate_model': 0.261825264329438, 'eval/demon_left_rate_teacher': 0.2643294379521425, 'eval/demon_right_precision': 0.9108138238573021, 'eval/demon_right_recall': 0.9231638418079096, 'eval/demon_right_f1': 0.9169472502805835, 'eval/demon_right_rate_model': 0.24958263772954925, 'eval/demon_right_rate_teacher': 0.24624373956594323, 'eval/demon_move_f1': 0.9100865283660982, 'eval/demon_cc_f1': 0.9420764263769, 'eval/latency_2/demon_fire_precision': 0.9564853556485355, 'eval/latency_2/demon_fire_recall': 0.9564853556485355, 'eval/latency_2/demon_fire_f1': 0.9564853556485355, 'eval/latency_2/demon_fire_rate_model': 0.6649972175848636, 'eval/latency_2/demon_fire_rate_teacher': 0.6649972175848636, 'eval/latency_2/demon_left_precision': 0.9075451647183846, 'eval/latency_2/demon_left_recall': 0.8989473684210526, 'eval/latency_2/demon_left_f1': 0.9032258064516129, 'eval/latency_2/demon_left_rate_model': 0.261825264329438, 'eval/latency_2/demon_left_rate_teacher': 0.2643294379521425, 'eval/latency_2/demon_right_precision': 0.9108138238573021, 'eval/latency_2/demon_right_recall': 0.9231638418079096, 'eval/latency_2/demon_right_f1': 0.9169472502805835, 'eval/latency_2/demon_right_rate_model': 0.24958263772954925, 'eval/latency_2/demon_right_rate_teacher': 0.24624373956594323, 'eval/latency_2/demon_move_f1': 0.9100865283660982, 'eval/latency_2/demon_cc_f1': 0.9420764263769, 'timing/data': 0.010738487355411053, 'timing/model': 1.0362260770052671, 'learning_rate/qwen_vl_interface': 5.785569172397254e-07, 'learning_rate/action_model': 5.785569172397254e-06, 'epoch': 2.42}) +[2026-06-18 07:29:33,927][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_6600_state +[2026-06-18 07:29:33,934][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 07:29:43,685][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_6600_state/model.safetensors +[2026-06-18 07:29:59,666][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_6600_state/optimizer.bin +[2026-06-18 07:29:59,668][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_6600_state/sampler.bin +[2026-06-18 07:29:59,668][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_6600_state/sampler_1.bin +[2026-06-18 07:29:59,671][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_6600_state/random_states_0.pkl +[2026-06-18 07:29:59,676][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 6600 +[2026-06-18 07:29:59,676][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 07:29:59,692][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 07:30:04,459][starVLA.training.train_starvla][INFO] - Step 6601, Loss: {'train/loss': 0.11221904307603836, 'train/grad_norm_pre_clip': 12.331632614135742, 'timing/data': 0.01603955402970314, 'timing/model': 1.0383083270862699, 'learning_rate/qwen_vl_interface': 5.781657031275306e-07, 'learning_rate/action_model': 5.781657031275306e-06, 'epoch': 2.42}) +[2026-06-18 07:30:09,696][starVLA.training.train_starvla][INFO] - Step 6602, Loss: {'train/loss': 0.08839041739702225, 'train/grad_norm_pre_clip': 15.482572555541992, 'timing/data': 0.020868901163339615, 'timing/model': 1.258227951824665, 'learning_rate/qwen_vl_interface': 5.777754574916129e-07, 'learning_rate/action_model': 5.7777545749161285e-06, 'epoch': 2.42}) +[2026-06-18 07:30:15,127][starVLA.training.train_starvla][INFO] - Step 6603, Loss: {'train/loss': 0.2349131852388382, 'train/grad_norm_pre_clip': 13.729853630065918, 'timing/data': 0.011583138257265091, 'timing/model': 1.9624219527468085, 'learning_rate/qwen_vl_interface': 5.773861804128701e-07, 'learning_rate/action_model': 5.773861804128701e-06, 'epoch': 2.42}) +[2026-06-18 07:30:19,664][starVLA.training.train_starvla][INFO] - Step 6604, Loss: {'train/loss': 0.12001322209835052, 'train/grad_norm_pre_clip': 14.436783790588379, 'timing/data': 0.01228802464902401, 'timing/model': 1.1105450112372637, 'learning_rate/qwen_vl_interface': 5.769978719720011e-07, 'learning_rate/action_model': 5.76997871972001e-06, 'epoch': 2.42}) +[2026-06-18 07:30:32,917][starVLA.training.train_starvla][INFO] - Step 6605, Loss: {'train/loss': 0.12996014952659607, 'train/grad_norm_pre_clip': 18.77996063232422, 'timing/data': 0.011634942144155502, 'timing/model': 4.613206314854324, 'learning_rate/qwen_vl_interface': 5.766105322495002e-07, 'learning_rate/action_model': 5.766105322495002e-06, 'epoch': 2.42}) +[2026-06-18 07:30:41,052][starVLA.training.train_starvla][INFO] - Step 6606, Loss: {'train/loss': 0.21733953058719635, 'train/grad_norm_pre_clip': 24.45868492126465, 'timing/data': 0.013045662082731724, 'timing/model': 3.2544805789366364, 'learning_rate/qwen_vl_interface': 5.762241613256654e-07, 'learning_rate/action_model': 5.762241613256654e-06, 'epoch': 2.42}) +[2026-06-18 07:30:45,931][starVLA.training.train_starvla][INFO] - Step 6607, Loss: {'train/loss': 0.09306352585554123, 'train/grad_norm_pre_clip': 12.005427360534668, 'timing/data': 0.011520038358867168, 'timing/model': 1.098740343004465, 'learning_rate/qwen_vl_interface': 5.758387592805916e-07, 'learning_rate/action_model': 5.758387592805915e-06, 'epoch': 2.42}) +[2026-06-18 07:30:50,814][starVLA.training.train_starvla][INFO] - Step 6608, Loss: {'train/loss': 0.29055365920066833, 'train/grad_norm_pre_clip': 23.96983528137207, 'timing/data': 0.016464552842080593, 'timing/model': 1.3996112700551748, 'learning_rate/qwen_vl_interface': 5.754543261941719e-07, 'learning_rate/action_model': 5.754543261941718e-06, 'epoch': 2.42}) +[2026-06-18 07:30:58,160][starVLA.training.train_starvla][INFO] - Step 6609, Loss: {'train/loss': 0.7002360820770264, 'train/grad_norm_pre_clip': 31.867029190063477, 'timing/data': 0.014749986119568348, 'timing/model': 1.1952668447047472, 'learning_rate/qwen_vl_interface': 5.750708621461011e-07, 'learning_rate/action_model': 5.75070862146101e-06, 'epoch': 2.42}) +[2026-06-18 07:31:06,660][starVLA.training.train_starvla][INFO] - Step 6610, Loss: {'train/loss': 0.1450297087430954, 'train/grad_norm_pre_clip': 23.675256729125977, 'timing/data': 0.011746016331017017, 'timing/model': 1.3094589617103338, 'learning_rate/qwen_vl_interface': 5.746883672158705e-07, 'learning_rate/action_model': 5.746883672158705e-06, 'epoch': 2.42}) +[2026-06-18 07:31:11,898][starVLA.training.train_starvla][INFO] - Step 6611, Loss: {'train/loss': 0.06198282539844513, 'train/grad_norm_pre_clip': 13.861352920532227, 'timing/data': 0.011698948219418526, 'timing/model': 1.0940224528312683, 'learning_rate/qwen_vl_interface': 5.743068414827726e-07, 'learning_rate/action_model': 5.743068414827726e-06, 'epoch': 2.42}) +[2026-06-18 07:31:16,609][starVLA.training.train_starvla][INFO] - Step 6612, Loss: {'train/loss': 0.15451885759830475, 'train/grad_norm_pre_clip': 21.046527862548828, 'timing/data': 0.011228936724364758, 'timing/model': 1.050375172868371, 'learning_rate/qwen_vl_interface': 5.739262850258968e-07, 'learning_rate/action_model': 5.7392628502589684e-06, 'epoch': 2.42}) +[2026-06-18 07:31:21,280][starVLA.training.train_starvla][INFO] - Step 6613, Loss: {'train/loss': 0.11514884978532791, 'train/grad_norm_pre_clip': 15.707892417907715, 'timing/data': 0.012240682728588581, 'timing/model': 1.6147251287475228, 'learning_rate/qwen_vl_interface': 5.735466979241352e-07, 'learning_rate/action_model': 5.735466979241352e-06, 'epoch': 2.42}) +[2026-06-18 07:31:25,569][starVLA.training.train_starvla][INFO] - Step 6614, Loss: {'train/loss': 0.15122760832309723, 'train/grad_norm_pre_clip': 13.163542747497559, 'timing/data': 0.010818129405379295, 'timing/model': 1.0750676738098264, 'learning_rate/qwen_vl_interface': 5.731680802561744e-07, 'learning_rate/action_model': 5.731680802561745e-06, 'epoch': 2.42}) +[2026-06-18 07:31:30,274][starVLA.training.train_starvla][INFO] - Step 6615, Loss: {'train/loss': 0.2695894241333008, 'train/grad_norm_pre_clip': 23.86930274963379, 'timing/data': 0.010892468504607677, 'timing/model': 1.0311564365401864, 'learning_rate/qwen_vl_interface': 5.727904321005031e-07, 'learning_rate/action_model': 5.727904321005031e-06, 'epoch': 2.42}) +[2026-06-18 07:31:36,404][starVLA.training.train_starvla][INFO] - Step 6616, Loss: {'train/loss': 0.326029896736145, 'train/grad_norm_pre_clip': 25.85712242126465, 'timing/data': 0.011535119265317917, 'timing/model': 1.215598339214921, 'learning_rate/qwen_vl_interface': 5.724137535354077e-07, 'learning_rate/action_model': 5.724137535354077e-06, 'epoch': 2.42}) +[2026-06-18 07:31:42,878][starVLA.training.train_starvla][INFO] - Step 6617, Loss: {'train/loss': 0.2742723524570465, 'train/grad_norm_pre_clip': 18.236392974853516, 'timing/data': 0.013645889237523079, 'timing/model': 1.1304490473121405, 'learning_rate/qwen_vl_interface': 5.720380446389748e-07, 'learning_rate/action_model': 5.720380446389748e-06, 'epoch': 2.42}) +[2026-06-18 07:31:47,498][starVLA.training.train_starvla][INFO] - Step 6618, Loss: {'train/loss': 0.26209500432014465, 'train/grad_norm_pre_clip': 20.63538360595703, 'timing/data': 0.011396391317248344, 'timing/model': 1.364013722166419, 'learning_rate/qwen_vl_interface': 5.716633054890897e-07, 'learning_rate/action_model': 5.716633054890896e-06, 'epoch': 2.42}) +[2026-06-18 07:31:55,579][starVLA.training.train_starvla][INFO] - Step 6619, Loss: {'train/loss': 0.25279027223587036, 'train/grad_norm_pre_clip': 29.843048095703125, 'timing/data': 0.014615467749536037, 'timing/model': 1.0711168944835663, 'learning_rate/qwen_vl_interface': 5.712895361634345e-07, 'learning_rate/action_model': 5.712895361634345e-06, 'epoch': 2.42}) +[2026-06-18 07:32:00,591][starVLA.training.train_starvla][INFO] - Step 6620, Loss: {'train/loss': 0.5531674027442932, 'train/grad_norm_pre_clip': 46.5641975402832, 'timing/data': 0.0010777981951832771, 'timing/model': 1.0223226565867662, 'learning_rate/qwen_vl_interface': 5.709167367394936e-07, 'learning_rate/action_model': 5.709167367394936e-06, 'epoch': 2.42}) +[2026-06-18 07:32:05,659][starVLA.training.train_starvla][INFO] - Step 6621, Loss: {'train/loss': 0.21537217497825623, 'train/grad_norm_pre_clip': 31.67783546447754, 'timing/data': 0.0012459103018045425, 'timing/model': 1.0940328715369105, 'learning_rate/qwen_vl_interface': 5.705449072945479e-07, 'learning_rate/action_model': 5.705449072945479e-06, 'epoch': 2.42}) +[2026-06-18 07:32:14,208][starVLA.training.train_starvla][INFO] - Step 6622, Loss: {'train/loss': 0.46024399995803833, 'train/grad_norm_pre_clip': 57.31608581542969, 'timing/data': 0.01675291731953621, 'timing/model': 1.1967498855665326, 'learning_rate/qwen_vl_interface': 5.70174047905678e-07, 'learning_rate/action_model': 5.701740479056779e-06, 'epoch': 2.42}) +[2026-06-18 07:32:19,651][starVLA.training.train_starvla][INFO] - Step 6623, Loss: {'train/loss': 0.2245696485042572, 'train/grad_norm_pre_clip': 27.90420150756836, 'timing/data': 0.013721559196710587, 'timing/model': 1.392683039419353, 'learning_rate/qwen_vl_interface': 5.69804158649764e-07, 'learning_rate/action_model': 5.6980415864976405e-06, 'epoch': 2.43}) +[2026-06-18 07:32:24,369][starVLA.training.train_starvla][INFO] - Step 6624, Loss: {'train/loss': 0.08707208931446075, 'train/grad_norm_pre_clip': 16.040170669555664, 'timing/data': 0.015856506302952766, 'timing/model': 1.3589488919824362, 'learning_rate/qwen_vl_interface': 5.69435239603484e-07, 'learning_rate/action_model': 5.69435239603484e-06, 'epoch': 2.43}) +[2026-06-18 07:32:29,507][starVLA.training.train_starvla][INFO] - Step 6625, Loss: {'train/loss': 0.20134374499320984, 'train/grad_norm_pre_clip': 29.939523696899414, 'timing/data': 0.017490681260824203, 'timing/model': 1.430345194414258, 'learning_rate/qwen_vl_interface': 5.690672908433149e-07, 'learning_rate/action_model': 5.690672908433149e-06, 'epoch': 2.43}) +[2026-06-18 07:32:35,355][starVLA.training.train_starvla][INFO] - Step 6626, Loss: {'train/loss': 0.3570972681045532, 'train/grad_norm_pre_clip': 27.08775520324707, 'timing/data': 0.012483453378081322, 'timing/model': 2.387777667492628, 'learning_rate/qwen_vl_interface': 5.68700312445534e-07, 'learning_rate/action_model': 5.687003124455339e-06, 'epoch': 2.43}) +[2026-06-18 07:32:45,907][starVLA.training.train_starvla][INFO] - Step 6627, Loss: {'train/loss': 0.17129823565483093, 'train/grad_norm_pre_clip': 22.96487045288086, 'timing/data': 0.012190851382911205, 'timing/model': 1.4163822615519166, 'learning_rate/qwen_vl_interface': 5.683343044862151e-07, 'learning_rate/action_model': 5.68334304486215e-06, 'epoch': 2.43}) +[2026-06-18 07:32:53,871][starVLA.training.train_starvla][INFO] - Step 6628, Loss: {'train/loss': 0.5610613226890564, 'train/grad_norm_pre_clip': 43.36857604980469, 'timing/data': 0.010951794683933258, 'timing/model': 4.790511140599847, 'learning_rate/qwen_vl_interface': 5.679692670412324e-07, 'learning_rate/action_model': 5.679692670412324e-06, 'epoch': 2.43}) +[2026-06-18 07:33:06,788][starVLA.training.train_starvla][INFO] - Step 6629, Loss: {'train/loss': 0.3011176884174347, 'train/grad_norm_pre_clip': 33.66551971435547, 'timing/data': 0.012876426801085472, 'timing/model': 1.0541589381173253, 'learning_rate/qwen_vl_interface': 5.67605200186259e-07, 'learning_rate/action_model': 5.67605200186259e-06, 'epoch': 2.43}) +[2026-06-18 07:33:11,850][starVLA.training.train_starvla][INFO] - Step 6630, Loss: {'train/loss': 0.1812707930803299, 'train/grad_norm_pre_clip': 16.812969207763672, 'timing/data': 0.01587001606822014, 'timing/model': 1.465831377543509, 'learning_rate/qwen_vl_interface': 5.67242103996766e-07, 'learning_rate/action_model': 5.67242103996766e-06, 'epoch': 2.43}) +[2026-06-18 07:33:17,011][starVLA.training.train_starvla][INFO] - Step 6631, Loss: {'train/loss': 0.5300900340080261, 'train/grad_norm_pre_clip': 30.934587478637695, 'timing/data': 0.011621004901826382, 'timing/model': 1.055418542586267, 'learning_rate/qwen_vl_interface': 5.668799785480242e-07, 'learning_rate/action_model': 5.668799785480242e-06, 'epoch': 2.43}) +[2026-06-18 07:33:23,307][starVLA.training.train_starvla][INFO] - Step 6632, Loss: {'train/loss': 0.27366533875465393, 'train/grad_norm_pre_clip': 18.788400650024414, 'timing/data': 0.059559340588748455, 'timing/model': 2.786472236737609, 'learning_rate/qwen_vl_interface': 5.665188239151019e-07, 'learning_rate/action_model': 5.665188239151018e-06, 'epoch': 2.43}) +[2026-06-18 07:33:31,735][starVLA.training.train_starvla][INFO] - Step 6633, Loss: {'train/loss': 0.4662434160709381, 'train/grad_norm_pre_clip': 34.347293853759766, 'timing/data': 0.013332112692296505, 'timing/model': 1.483116190880537, 'learning_rate/qwen_vl_interface': 5.661586401728671e-07, 'learning_rate/action_model': 5.6615864017286705e-06, 'epoch': 2.43}) +[2026-06-18 07:33:36,315][starVLA.training.train_starvla][INFO] - Step 6634, Loss: {'train/loss': 0.09324927628040314, 'train/grad_norm_pre_clip': 11.19369888305664, 'timing/data': 0.012227575294673443, 'timing/model': 1.0586171494796872, 'learning_rate/qwen_vl_interface': 5.657994273959857e-07, 'learning_rate/action_model': 5.657994273959856e-06, 'epoch': 2.43}) +[2026-06-18 07:33:41,207][starVLA.training.train_starvla][INFO] - Step 6635, Loss: {'train/loss': 0.1288924366235733, 'train/grad_norm_pre_clip': 13.900506019592285, 'timing/data': 0.013210808858275414, 'timing/model': 1.4271398233249784, 'learning_rate/qwen_vl_interface': 5.654411856589233e-07, 'learning_rate/action_model': 5.654411856589233e-06, 'epoch': 2.43}) +[2026-06-18 07:33:49,114][starVLA.training.train_starvla][INFO] - Step 6636, Loss: {'train/loss': 0.18877792358398438, 'train/grad_norm_pre_clip': 53.27739715576172, 'timing/data': 0.015852794982492924, 'timing/model': 1.2771602189168334, 'learning_rate/qwen_vl_interface': 5.650839150359444e-07, 'learning_rate/action_model': 5.650839150359444e-06, 'epoch': 2.43}) +[2026-06-18 07:34:02,672][starVLA.training.train_starvla][INFO] - Step 6637, Loss: {'train/loss': 0.11538331210613251, 'train/grad_norm_pre_clip': 18.550174713134766, 'timing/data': 0.02667534351348877, 'timing/model': 3.145063945092261, 'learning_rate/qwen_vl_interface': 5.647276156011114e-07, 'learning_rate/action_model': 5.6472761560111135e-06, 'epoch': 2.43}) +[2026-06-18 07:34:19,683][starVLA.training.train_starvla][INFO] - Step 6638, Loss: {'train/loss': 0.14860470592975616, 'train/grad_norm_pre_clip': 22.489225387573242, 'timing/data': 0.01440493855625391, 'timing/model': 5.53190434910357, 'learning_rate/qwen_vl_interface': 5.643722874282846e-07, 'learning_rate/action_model': 5.643722874282846e-06, 'epoch': 2.43}) +[2026-06-18 07:34:27,901][starVLA.training.train_starvla][INFO] - Step 6639, Loss: {'train/loss': 0.29881224036216736, 'train/grad_norm_pre_clip': 20.778221130371094, 'timing/data': 0.010889745317399502, 'timing/model': 2.706142846494913, 'learning_rate/qwen_vl_interface': 5.640179305911253e-07, 'learning_rate/action_model': 5.640179305911253e-06, 'epoch': 2.43}) +[2026-06-18 07:34:38,505][starVLA.training.train_starvla][INFO] - Step 6640, Loss: {'train/loss': 0.23278526961803436, 'train/grad_norm_pre_clip': 28.3098087310791, 'timing/data': 0.02120336890220642, 'timing/model': 1.1801287680864334, 'learning_rate/qwen_vl_interface': 5.636645451630908e-07, 'learning_rate/action_model': 5.6366454516309084e-06, 'epoch': 2.43}) +[2026-06-18 07:34:49,712][starVLA.training.train_starvla][INFO] - Step 6641, Loss: {'train/loss': 0.139864981174469, 'train/grad_norm_pre_clip': 22.089906692504883, 'timing/data': 0.02056444250047207, 'timing/model': 2.0266514336690307, 'learning_rate/qwen_vl_interface': 5.63312131217439e-07, 'learning_rate/action_model': 5.633121312174389e-06, 'epoch': 2.43}) +[2026-06-18 07:35:03,425][starVLA.training.train_starvla][INFO] - Step 6642, Loss: {'train/loss': 0.05835729092359543, 'train/grad_norm_pre_clip': 9.032182693481445, 'timing/data': 0.018436172977089882, 'timing/model': 4.403710421174765, 'learning_rate/qwen_vl_interface': 5.629606888272253e-07, 'learning_rate/action_model': 5.629606888272252e-06, 'epoch': 2.43}) +[2026-06-18 07:35:12,136][starVLA.training.train_starvla][INFO] - Step 6643, Loss: {'train/loss': 0.2263851761817932, 'train/grad_norm_pre_clip': 20.711442947387695, 'timing/data': 0.014191288501024246, 'timing/model': 1.4779973486438394, 'learning_rate/qwen_vl_interface': 5.626102180653046e-07, 'learning_rate/action_model': 5.626102180653045e-06, 'epoch': 2.43}) +[2026-06-18 07:35:24,532][starVLA.training.train_starvla][INFO] - Step 6644, Loss: {'train/loss': 0.02430887706577778, 'train/grad_norm_pre_clip': 3.8782548904418945, 'timing/data': 0.0136719960719347, 'timing/model': 5.552130185998976, 'learning_rate/qwen_vl_interface': 5.62260719004329e-07, 'learning_rate/action_model': 5.6226071900432895e-06, 'epoch': 2.43}) +[2026-06-18 07:35:29,759][starVLA.training.train_starvla][INFO] - Step 6645, Loss: {'train/loss': 0.31965306401252747, 'train/grad_norm_pre_clip': 31.51020050048828, 'timing/data': 0.011699933558702469, 'timing/model': 1.0984226651489735, 'learning_rate/qwen_vl_interface': 5.619121917167504e-07, 'learning_rate/action_model': 5.619121917167505e-06, 'epoch': 2.43}) +[2026-06-18 07:35:34,730][starVLA.training.train_starvla][INFO] - Step 6646, Loss: {'train/loss': 0.18974338471889496, 'train/grad_norm_pre_clip': 23.839513778686523, 'timing/data': 0.016044548712670803, 'timing/model': 1.08913354575634, 'learning_rate/qwen_vl_interface': 5.615646362748199e-07, 'learning_rate/action_model': 5.6156463627482e-06, 'epoch': 2.43}) +[2026-06-18 07:35:39,352][starVLA.training.train_starvla][INFO] - Step 6647, Loss: {'train/loss': 0.09707218408584595, 'train/grad_norm_pre_clip': 17.531631469726562, 'timing/data': 0.01309511624276638, 'timing/model': 1.2748385444283485, 'learning_rate/qwen_vl_interface': 5.612180527505854e-07, 'learning_rate/action_model': 5.612180527505854e-06, 'epoch': 2.43}) +[2026-06-18 07:35:43,876][starVLA.training.train_starvla][INFO] - Step 6648, Loss: {'train/loss': 0.4366280734539032, 'train/grad_norm_pre_clip': 34.632816314697266, 'timing/data': 0.016331533901393414, 'timing/model': 1.379331223666668, 'learning_rate/qwen_vl_interface': 5.608724412158931e-07, 'learning_rate/action_model': 5.608724412158931e-06, 'epoch': 2.43}) +[2026-06-18 07:35:48,592][starVLA.training.train_starvla][INFO] - Step 6649, Loss: {'train/loss': 0.2568737268447876, 'train/grad_norm_pre_clip': 19.84632110595703, 'timing/data': 0.014361120760440826, 'timing/model': 1.0456371298059821, 'learning_rate/qwen_vl_interface': 5.605278017423897e-07, 'learning_rate/action_model': 5.605278017423897e-06, 'epoch': 2.43}) +[2026-06-18 07:35:53,682][starVLA.training.train_starvla][INFO] - Step 6650, Loss: {'train/loss': 0.07005023956298828, 'train/grad_norm_pre_clip': 11.075074195861816, 'timing/data': 0.012547994963824749, 'timing/model': 1.030705682002008, 'learning_rate/qwen_vl_interface': 5.601841344015193e-07, 'learning_rate/action_model': 5.601841344015193e-06, 'epoch': 2.44}) +[2026-06-18 07:35:58,273][starVLA.training.train_starvla][INFO] - Step 6651, Loss: {'train/loss': 0.5980923771858215, 'train/grad_norm_pre_clip': 34.23520278930664, 'timing/data': 0.01654778979718685, 'timing/model': 1.1122247679159045, 'learning_rate/qwen_vl_interface': 5.598414392645239e-07, 'learning_rate/action_model': 5.598414392645239e-06, 'epoch': 2.44}) +[2026-06-18 07:36:02,812][starVLA.training.train_starvla][INFO] - Step 6652, Loss: {'train/loss': 0.2020435929298401, 'train/grad_norm_pre_clip': 24.16020965576172, 'timing/data': 0.013074848800897598, 'timing/model': 1.3566640364006162, 'learning_rate/qwen_vl_interface': 5.594997164024454e-07, 'learning_rate/action_model': 5.594997164024454e-06, 'epoch': 2.44}) +[2026-06-18 07:36:07,466][starVLA.training.train_starvla][INFO] - Step 6653, Loss: {'train/loss': 0.29535728693008423, 'train/grad_norm_pre_clip': 30.03290367126465, 'timing/data': 0.011511106975376606, 'timing/model': 1.353267994709313, 'learning_rate/qwen_vl_interface': 5.591589658861228e-07, 'learning_rate/action_model': 5.591589658861228e-06, 'epoch': 2.44}) +[2026-06-18 07:36:11,911][starVLA.training.train_starvla][INFO] - Step 6654, Loss: {'train/loss': 0.2285858392715454, 'train/grad_norm_pre_clip': 34.52897262573242, 'timing/data': 0.01192588359117508, 'timing/model': 1.0460026357322931, 'learning_rate/qwen_vl_interface': 5.588191877861945e-07, 'learning_rate/action_model': 5.588191877861944e-06, 'epoch': 2.44}) +[2026-06-18 07:36:16,279][starVLA.training.train_starvla][INFO] - Step 6655, Loss: {'train/loss': 0.0440535694360733, 'train/grad_norm_pre_clip': 5.491833686828613, 'timing/data': 0.012225069105625153, 'timing/model': 1.0236164843663573, 'learning_rate/qwen_vl_interface': 5.584803821730961e-07, 'learning_rate/action_model': 5.58480382173096e-06, 'epoch': 2.44}) +[2026-06-18 07:36:20,579][starVLA.training.train_starvla][INFO] - Step 6656, Loss: {'train/loss': 0.20557716488838196, 'train/grad_norm_pre_clip': 27.518447875976562, 'timing/data': 0.010623431764543056, 'timing/model': 1.011602597311139, 'learning_rate/qwen_vl_interface': 5.581425491170627e-07, 'learning_rate/action_model': 5.581425491170626e-06, 'epoch': 2.44}) +[2026-06-18 07:36:24,725][starVLA.training.train_starvla][INFO] - Step 6657, Loss: {'train/loss': 0.15680797398090363, 'train/grad_norm_pre_clip': 17.431867599487305, 'timing/data': 0.010352324694395065, 'timing/model': 1.0345679661259055, 'learning_rate/qwen_vl_interface': 5.578056886881274e-07, 'learning_rate/action_model': 5.578056886881274e-06, 'epoch': 2.44}) +[2026-06-18 07:36:28,912][starVLA.training.train_starvla][INFO] - Step 6658, Loss: {'train/loss': 0.13801459968090057, 'train/grad_norm_pre_clip': 33.28789138793945, 'timing/data': 0.010581957176327705, 'timing/model': 1.2306473264470696, 'learning_rate/qwen_vl_interface': 5.574698009561223e-07, 'learning_rate/action_model': 5.574698009561223e-06, 'epoch': 2.44}) +[2026-06-18 07:36:32,936][starVLA.training.train_starvla][INFO] - Step 6659, Loss: {'train/loss': 0.08245570212602615, 'train/grad_norm_pre_clip': 18.354042053222656, 'timing/data': 0.016864686273038387, 'timing/model': 1.0173867605626583, 'learning_rate/qwen_vl_interface': 5.571348859906763e-07, 'learning_rate/action_model': 5.571348859906764e-06, 'epoch': 2.44}) +[2026-06-18 07:36:37,075][starVLA.training.train_starvla][INFO] - Step 6660, Loss: {'train/loss': 0.1747923642396927, 'train/grad_norm_pre_clip': 25.277555465698242, 'timing/data': 0.010251611471176147, 'timing/model': 1.0108952419832349, 'learning_rate/qwen_vl_interface': 5.568009438612186e-07, 'learning_rate/action_model': 5.568009438612186e-06, 'epoch': 2.44}) +[2026-06-18 07:36:41,549][starVLA.training.train_starvla][INFO] - Step 6661, Loss: {'train/loss': 0.21937866508960724, 'train/grad_norm_pre_clip': 23.205360412597656, 'timing/data': 0.025816419161856174, 'timing/model': 1.2148884972557425, 'learning_rate/qwen_vl_interface': 5.564679746369755e-07, 'learning_rate/action_model': 5.5646797463697556e-06, 'epoch': 2.44}) +[2026-06-18 07:36:46,478][starVLA.training.train_starvla][INFO] - Step 6662, Loss: {'train/loss': 0.14784950017929077, 'train/grad_norm_pre_clip': 19.638233184814453, 'timing/data': 0.00992675218731165, 'timing/model': 1.0345565006136894, 'learning_rate/qwen_vl_interface': 5.561359783869712e-07, 'learning_rate/action_model': 5.5613597838697116e-06, 'epoch': 2.44}) +[2026-06-18 07:36:50,851][starVLA.training.train_starvla][INFO] - Step 6663, Loss: {'train/loss': 0.07578616589307785, 'train/grad_norm_pre_clip': 12.656444549560547, 'timing/data': 0.010165730491280556, 'timing/model': 1.0709515688940883, 'learning_rate/qwen_vl_interface': 5.558049551800291e-07, 'learning_rate/action_model': 5.558049551800292e-06, 'epoch': 2.44}) +[2026-06-18 07:36:56,720][starVLA.training.train_starvla][INFO] - Step 6664, Loss: {'train/loss': 0.1197548359632492, 'train/grad_norm_pre_clip': 14.571188926696777, 'timing/data': 0.010409889742732048, 'timing/model': 1.426620502024889, 'learning_rate/qwen_vl_interface': 5.554749050847718e-07, 'learning_rate/action_model': 5.5547490508477176e-06, 'epoch': 2.44}) +[2026-06-18 07:37:02,249][starVLA.training.train_starvla][INFO] - Step 6665, Loss: {'train/loss': 0.10137280076742172, 'train/grad_norm_pre_clip': 17.866863250732422, 'timing/data': 0.011047129519283772, 'timing/model': 1.0143946977332234, 'learning_rate/qwen_vl_interface': 5.55145828169618e-07, 'learning_rate/action_model': 5.5514582816961805e-06, 'epoch': 2.44}) +[2026-06-18 07:37:06,639][starVLA.training.train_starvla][INFO] - Step 6666, Loss: {'train/loss': 0.4600316882133484, 'train/grad_norm_pre_clip': 40.57685470581055, 'timing/data': 0.012716934084892273, 'timing/model': 1.0879615852609277, 'learning_rate/qwen_vl_interface': 5.548177245027859e-07, 'learning_rate/action_model': 5.548177245027859e-06, 'epoch': 2.44}) +[2026-06-18 07:37:15,956][starVLA.training.train_starvla][INFO] - Step 6667, Loss: {'train/loss': 0.2631741464138031, 'train/grad_norm_pre_clip': 35.275177001953125, 'timing/data': 0.012162206694483757, 'timing/model': 4.653178098611534, 'learning_rate/qwen_vl_interface': 5.544905941522913e-07, 'learning_rate/action_model': 5.544905941522913e-06, 'epoch': 2.44}) +[2026-06-18 07:37:29,067][starVLA.training.train_starvla][INFO] - Step 6668, Loss: {'train/loss': 0.12547779083251953, 'train/grad_norm_pre_clip': 18.813806533813477, 'timing/data': 0.024326328188180923, 'timing/model': 6.467193583957851, 'learning_rate/qwen_vl_interface': 5.541644371859498e-07, 'learning_rate/action_model': 5.541644371859498e-06, 'epoch': 2.44}) +[2026-06-18 07:37:37,232][starVLA.training.train_starvla][INFO] - Step 6669, Loss: {'train/loss': 0.30511873960494995, 'train/grad_norm_pre_clip': 33.968505859375, 'timing/data': 0.012715901248157024, 'timing/model': 1.6264749960973859, 'learning_rate/qwen_vl_interface': 5.53839253671373e-07, 'learning_rate/action_model': 5.5383925367137305e-06, 'epoch': 2.44}) +[2026-06-18 07:37:42,015][starVLA.training.train_starvla][INFO] - Step 6670, Loss: {'train/loss': 0.09283123165369034, 'train/grad_norm_pre_clip': 20.65365982055664, 'timing/data': 0.01153884269297123, 'timing/model': 1.1521579967811704, 'learning_rate/qwen_vl_interface': 5.535150436759725e-07, 'learning_rate/action_model': 5.535150436759725e-06, 'epoch': 2.44}) +[2026-06-18 07:37:46,562][starVLA.training.train_starvla][INFO] - Step 6671, Loss: {'train/loss': 0.48550307750701904, 'train/grad_norm_pre_clip': 33.48324966430664, 'timing/data': 0.01377533096820116, 'timing/model': 1.0374071188271046, 'learning_rate/qwen_vl_interface': 5.531918072669566e-07, 'learning_rate/action_model': 5.531918072669565e-06, 'epoch': 2.44}) +[2026-06-18 07:37:51,655][starVLA.training.train_starvla][INFO] - Step 6672, Loss: {'train/loss': 0.3633699119091034, 'train/grad_norm_pre_clip': 29.051530838012695, 'timing/data': 0.011554872617125511, 'timing/model': 1.3879381651058793, 'learning_rate/qwen_vl_interface': 5.528695445113336e-07, 'learning_rate/action_model': 5.528695445113335e-06, 'epoch': 2.44}) +[2026-06-18 07:37:55,772][starVLA.training.train_starvla][INFO] - Step 6673, Loss: {'train/loss': 0.22808818519115448, 'train/grad_norm_pre_clip': 23.872047424316406, 'timing/data': 0.014437892474234104, 'timing/model': 1.041336603462696, 'learning_rate/qwen_vl_interface': 5.525482554759079e-07, 'learning_rate/action_model': 5.5254825547590795e-06, 'epoch': 2.44}) +[2026-06-18 07:38:00,673][starVLA.training.train_starvla][INFO] - Step 6674, Loss: {'train/loss': 0.3173748254776001, 'train/grad_norm_pre_clip': 31.801137924194336, 'timing/data': 0.013593717478215694, 'timing/model': 1.038610584102571, 'learning_rate/qwen_vl_interface': 5.522279402272834e-07, 'learning_rate/action_model': 5.522279402272834e-06, 'epoch': 2.44}) +[2026-06-18 07:38:05,188][starVLA.training.train_starvla][INFO] - Step 6675, Loss: {'train/loss': 0.40897899866104126, 'train/grad_norm_pre_clip': 21.635408401489258, 'timing/data': 0.01810863707214594, 'timing/model': 1.1492668371647596, 'learning_rate/qwen_vl_interface': 5.519085988318627e-07, 'learning_rate/action_model': 5.5190859883186265e-06, 'epoch': 2.44}) +[2026-06-18 07:38:09,793][starVLA.training.train_starvla][INFO] - Step 6676, Loss: {'train/loss': 0.1887785643339157, 'train/grad_norm_pre_clip': 28.70134735107422, 'timing/data': 0.0134240398183465, 'timing/model': 1.0436849556863308, 'learning_rate/qwen_vl_interface': 5.515902313558443e-07, 'learning_rate/action_model': 5.515902313558444e-06, 'epoch': 2.44}) +[2026-06-18 07:38:19,517][starVLA.training.train_starvla][INFO] - Step 6677, Loss: {'train/loss': 0.24127274751663208, 'train/grad_norm_pre_clip': 24.558515548706055, 'timing/data': 0.015101083554327488, 'timing/model': 5.692045551724732, 'learning_rate/qwen_vl_interface': 5.512728378652266e-07, 'learning_rate/action_model': 5.512728378652266e-06, 'epoch': 2.44}) +[2026-06-18 07:38:26,131][starVLA.training.train_starvla][INFO] - Step 6678, Loss: {'train/loss': 0.20504333078861237, 'train/grad_norm_pre_clip': 20.86074447631836, 'timing/data': 0.012144968844950199, 'timing/model': 1.4970220727846026, 'learning_rate/qwen_vl_interface': 5.509564184258055e-07, 'learning_rate/action_model': 5.509564184258055e-06, 'epoch': 2.45}) +[2026-06-18 07:38:30,810][starVLA.training.train_starvla][INFO] - Step 6679, Loss: {'train/loss': 0.07490569353103638, 'train/grad_norm_pre_clip': 13.949718475341797, 'timing/data': 0.01386680081486702, 'timing/model': 1.051502937451005, 'learning_rate/qwen_vl_interface': 5.506409731031754e-07, 'learning_rate/action_model': 5.506409731031755e-06, 'epoch': 2.45}) +[2026-06-18 07:38:36,010][starVLA.training.train_starvla][INFO] - Step 6680, Loss: {'train/loss': 0.04965142905712128, 'train/grad_norm_pre_clip': 11.94775104522705, 'timing/data': 0.013499555177986622, 'timing/model': 1.0417066784575582, 'learning_rate/qwen_vl_interface': 5.503265019627289e-07, 'learning_rate/action_model': 5.503265019627289e-06, 'epoch': 2.45}) +[2026-06-18 07:38:40,806][starVLA.training.train_starvla][INFO] - Step 6681, Loss: {'train/loss': 0.35791832208633423, 'train/grad_norm_pre_clip': 27.331724166870117, 'timing/data': 0.012369120493531227, 'timing/model': 1.1434473870322108, 'learning_rate/qwen_vl_interface': 5.500130050696548e-07, 'learning_rate/action_model': 5.500130050696548e-06, 'epoch': 2.45}) +[2026-06-18 07:38:45,582][starVLA.training.train_starvla][INFO] - Step 6682, Loss: {'train/loss': 0.32871657609939575, 'train/grad_norm_pre_clip': 21.679330825805664, 'timing/data': 0.013256913051009178, 'timing/model': 1.4648285815492272, 'learning_rate/qwen_vl_interface': 5.497004824889424e-07, 'learning_rate/action_model': 5.497004824889423e-06, 'epoch': 2.45}) +[2026-06-18 07:38:51,897][starVLA.training.train_starvla][INFO] - Step 6683, Loss: {'train/loss': 0.10647008568048477, 'train/grad_norm_pre_clip': 27.348907470703125, 'timing/data': 0.04661516938358545, 'timing/model': 1.5392952589318156, 'learning_rate/qwen_vl_interface': 5.493889342853779e-07, 'learning_rate/action_model': 5.493889342853778e-06, 'epoch': 2.45}) +[2026-06-18 07:38:56,664][starVLA.training.train_starvla][INFO] - Step 6684, Loss: {'train/loss': 0.20246197283267975, 'train/grad_norm_pre_clip': 26.2861328125, 'timing/data': 0.037470332346856594, 'timing/model': 1.0466279396787286, 'learning_rate/qwen_vl_interface': 5.490783605235451e-07, 'learning_rate/action_model': 5.49078360523545e-06, 'epoch': 2.45}) +[2026-06-18 07:39:01,311][starVLA.training.train_starvla][INFO] - Step 6685, Loss: {'train/loss': 0.08853991329669952, 'train/grad_norm_pre_clip': 10.837928771972656, 'timing/data': 0.04990651737898588, 'timing/model': 1.054342808201909, 'learning_rate/qwen_vl_interface': 5.487687612678265e-07, 'learning_rate/action_model': 5.4876876126782645e-06, 'epoch': 2.45}) +[2026-06-18 07:39:06,286][starVLA.training.train_starvla][INFO] - Step 6686, Loss: {'train/loss': 0.20014601945877075, 'train/grad_norm_pre_clip': 21.295921325683594, 'timing/data': 0.046743317507207394, 'timing/model': 1.04905338678509, 'learning_rate/qwen_vl_interface': 5.484601365824032e-07, 'learning_rate/action_model': 5.4846013658240315e-06, 'epoch': 2.45}) +[2026-06-18 07:39:11,516][starVLA.training.train_starvla][INFO] - Step 6687, Loss: {'train/loss': 0.06196148321032524, 'train/grad_norm_pre_clip': 15.002143859863281, 'timing/data': 0.047607540152966976, 'timing/model': 1.4230071045458317, 'learning_rate/qwen_vl_interface': 5.481524865312518e-07, 'learning_rate/action_model': 5.481524865312518e-06, 'epoch': 2.45}) +[2026-06-18 07:39:22,644][starVLA.training.train_starvla][INFO] - Step 6688, Loss: {'train/loss': 0.12449397891759872, 'train/grad_norm_pre_clip': 16.609954833984375, 'timing/data': 0.048988266848027706, 'timing/model': 5.050357023254037, 'learning_rate/qwen_vl_interface': 5.478458111781496e-07, 'learning_rate/action_model': 5.478458111781496e-06, 'epoch': 2.45}) +[2026-06-18 07:39:26,988][starVLA.training.train_starvla][INFO] - Step 6689, Loss: {'train/loss': 0.7439446449279785, 'train/grad_norm_pre_clip': 47.16080856323242, 'timing/data': 0.05882324278354645, 'timing/model': 1.074120563454926, 'learning_rate/qwen_vl_interface': 5.475401105866712e-07, 'learning_rate/action_model': 5.4754011058667114e-06, 'epoch': 2.45}) +[2026-06-18 07:39:32,240][starVLA.training.train_starvla][INFO] - Step 6690, Loss: {'train/loss': 0.12963417172431946, 'train/grad_norm_pre_clip': 17.1011905670166, 'timing/data': 0.044764104299247265, 'timing/model': 1.1747507508844137, 'learning_rate/qwen_vl_interface': 5.47235384820187e-07, 'learning_rate/action_model': 5.47235384820187e-06, 'epoch': 2.45}) +[2026-06-18 07:39:37,371][starVLA.training.train_starvla][INFO] - Step 6691, Loss: {'train/loss': 0.03960162401199341, 'train/grad_norm_pre_clip': 9.03875732421875, 'timing/data': 0.0454587135463953, 'timing/model': 1.06137784011662, 'learning_rate/qwen_vl_interface': 5.469316339418681e-07, 'learning_rate/action_model': 5.469316339418681e-06, 'epoch': 2.45}) +[2026-06-18 07:39:44,653][starVLA.training.train_starvla][INFO] - Step 6692, Loss: {'train/loss': 0.17125634849071503, 'train/grad_norm_pre_clip': 18.99673843383789, 'timing/data': 0.046561301685869694, 'timing/model': 1.3699405016377568, 'learning_rate/qwen_vl_interface': 5.466288580146825e-07, 'learning_rate/action_model': 5.466288580146825e-06, 'epoch': 2.45}) +[2026-06-18 07:39:49,088][starVLA.training.train_starvla][INFO] - Step 6693, Loss: {'train/loss': 0.06860654801130295, 'train/grad_norm_pre_clip': 13.656645774841309, 'timing/data': 0.04679190367460251, 'timing/model': 1.4097948111593723, 'learning_rate/qwen_vl_interface': 5.463270571013954e-07, 'learning_rate/action_model': 5.463270571013954e-06, 'epoch': 2.45}) +[2026-06-18 07:39:53,307][starVLA.training.train_starvla][INFO] - Step 6694, Loss: {'train/loss': 0.10465152561664581, 'train/grad_norm_pre_clip': 16.797090530395508, 'timing/data': 0.041360337287187576, 'timing/model': 1.146380883641541, 'learning_rate/qwen_vl_interface': 5.460262312645709e-07, 'learning_rate/action_model': 5.460262312645709e-06, 'epoch': 2.45}) +[2026-06-18 07:39:58,060][starVLA.training.train_starvla][INFO] - Step 6695, Loss: {'train/loss': 0.13811548054218292, 'train/grad_norm_pre_clip': 20.001054763793945, 'timing/data': 0.039059135131537914, 'timing/model': 1.2640830846503377, 'learning_rate/qwen_vl_interface': 5.457263805665707e-07, 'learning_rate/action_model': 5.457263805665707e-06, 'epoch': 2.45}) +[2026-06-18 07:40:03,140][starVLA.training.train_starvla][INFO] - Step 6696, Loss: {'train/loss': 0.2874465584754944, 'train/grad_norm_pre_clip': 27.152877807617188, 'timing/data': 0.04297685157507658, 'timing/model': 1.068928218446672, 'learning_rate/qwen_vl_interface': 5.454275050695529e-07, 'learning_rate/action_model': 5.454275050695529e-06, 'epoch': 2.45}) +[2026-06-18 07:40:07,744][starVLA.training.train_starvla][INFO] - Step 6697, Loss: {'train/loss': 0.05165589600801468, 'train/grad_norm_pre_clip': 8.571732521057129, 'timing/data': 0.043117718771100044, 'timing/model': 1.1421207627281547, 'learning_rate/qwen_vl_interface': 5.451296048354764e-07, 'learning_rate/action_model': 5.4512960483547635e-06, 'epoch': 2.45}) +[2026-06-18 07:40:13,832][starVLA.training.train_starvla][INFO] - Step 6698, Loss: {'train/loss': 0.2592098116874695, 'train/grad_norm_pre_clip': 14.220254898071289, 'timing/data': 0.03955733776092529, 'timing/model': 1.4464936489239335, 'learning_rate/qwen_vl_interface': 5.448326799260951e-07, 'learning_rate/action_model': 5.448326799260952e-06, 'epoch': 2.45}) +[2026-06-18 07:40:18,181][starVLA.training.train_starvla][INFO] - Step 6699, Loss: {'train/loss': 0.26525962352752686, 'train/grad_norm_pre_clip': 20.11225700378418, 'timing/data': 0.04160068929195404, 'timing/model': 1.1257355893030763, 'learning_rate/qwen_vl_interface': 5.445367304029627e-07, 'learning_rate/action_model': 5.445367304029626e-06, 'epoch': 2.45}) +[2026-06-18 07:45:09,114][starVLA.training.train_starvla][INFO] - Step 6700, Loss: {'train/loss': 0.08891398459672928, 'train/grad_norm_pre_clip': 15.5011625289917, 'eval/loss': 0.5546364784240723, 'eval/latency_2/loss': 0.5546364784240723, 'eval/demon_fire_precision': 0.9474116680361545, 'eval/demon_fire_recall': 0.9648535564853556, 'eval/demon_fire_f1': 0.9560530679933665, 'eval/demon_fire_rate_model': 0.6772398441847524, 'eval/demon_fire_rate_teacher': 0.6649972175848636, 'eval/demon_left_precision': 0.9158469945355191, 'eval/demon_left_recall': 0.8821052631578947, 'eval/demon_left_f1': 0.8986595174262735, 'eval/demon_left_rate_model': 0.25459098497495825, 'eval/demon_left_rate_teacher': 0.2643294379521425, 'eval/demon_right_precision': 0.9236111111111112, 'eval/demon_right_recall': 0.9016949152542373, 'eval/demon_right_f1': 0.9125214408233276, 'eval/demon_right_rate_model': 0.24040066777963273, 'eval/demon_right_rate_teacher': 0.24624373956594323, 'eval/demon_move_f1': 0.9055904791248006, 'eval/demon_cc_f1': 0.940333491602087, 'eval/latency_2/demon_fire_precision': 0.9474116680361545, 'eval/latency_2/demon_fire_recall': 0.9648535564853556, 'eval/latency_2/demon_fire_f1': 0.9560530679933665, 'eval/latency_2/demon_fire_rate_model': 0.6772398441847524, 'eval/latency_2/demon_fire_rate_teacher': 0.6649972175848636, 'eval/latency_2/demon_left_precision': 0.9158469945355191, 'eval/latency_2/demon_left_recall': 0.8821052631578947, 'eval/latency_2/demon_left_f1': 0.8986595174262735, 'eval/latency_2/demon_left_rate_model': 0.25459098497495825, 'eval/latency_2/demon_left_rate_teacher': 0.2643294379521425, 'eval/latency_2/demon_right_precision': 0.9236111111111112, 'eval/latency_2/demon_right_recall': 0.9016949152542373, 'eval/latency_2/demon_right_f1': 0.9125214408233276, 'eval/latency_2/demon_right_rate_model': 0.24040066777963273, 'eval/latency_2/demon_right_rate_teacher': 0.24624373956594323, 'eval/latency_2/demon_move_f1': 0.9055904791248006, 'eval/latency_2/demon_cc_f1': 0.940333491602087, 'timing/data': 0.055404363200068474, 'timing/model': 1.291525992564857, 'learning_rate/qwen_vl_interface': 5.442417563274289e-07, 'learning_rate/action_model': 5.442417563274289e-06, 'epoch': 2.45}) +[2026-06-18 07:45:09,118][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_6700_state +[2026-06-18 07:45:09,125][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 07:45:18,882][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_6700_state/model.safetensors +[2026-06-18 07:45:35,052][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_6700_state/optimizer.bin +[2026-06-18 07:45:35,054][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_6700_state/sampler.bin +[2026-06-18 07:45:35,054][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_6700_state/sampler_1.bin +[2026-06-18 07:45:35,085][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_6700_state/random_states_0.pkl +[2026-06-18 07:45:35,092][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 6700 +[2026-06-18 07:45:35,092][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 07:45:35,110][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 07:45:40,467][starVLA.training.train_starvla][INFO] - Step 6701, Loss: {'train/loss': 0.27078136801719666, 'train/grad_norm_pre_clip': 32.00203323364258, 'timing/data': 0.047468746080994606, 'timing/model': 1.1125611439347267, 'learning_rate/qwen_vl_interface': 5.439477577606429e-07, 'learning_rate/action_model': 5.4394775776064295e-06, 'epoch': 2.45}) +[2026-06-18 07:45:45,039][starVLA.training.train_starvla][INFO] - Step 6702, Loss: {'train/loss': 0.2858172059059143, 'train/grad_norm_pre_clip': 27.949432373046875, 'timing/data': 0.037766363471746445, 'timing/model': 1.0340098617598414, 'learning_rate/qwen_vl_interface': 5.436547347635506e-07, 'learning_rate/action_model': 5.436547347635506e-06, 'epoch': 2.45}) +[2026-06-18 07:45:50,114][starVLA.training.train_starvla][INFO] - Step 6703, Loss: {'train/loss': 0.16426827013492584, 'train/grad_norm_pre_clip': 23.527713775634766, 'timing/data': 0.04759763088077307, 'timing/model': 1.295300593599677, 'learning_rate/qwen_vl_interface': 5.433626873968962e-07, 'learning_rate/action_model': 5.433626873968962e-06, 'epoch': 2.45}) +[2026-06-18 07:45:54,979][starVLA.training.train_starvla][INFO] - Step 6704, Loss: {'train/loss': 0.1325223594903946, 'train/grad_norm_pre_clip': 27.96278953552246, 'timing/data': 0.04749571159482002, 'timing/model': 1.5041366377845407, 'learning_rate/qwen_vl_interface': 5.43071615721221e-07, 'learning_rate/action_model': 5.43071615721221e-06, 'epoch': 2.45}) +[2026-06-18 07:45:59,613][starVLA.training.train_starvla][INFO] - Step 6705, Loss: {'train/loss': 0.44261816143989563, 'train/grad_norm_pre_clip': 36.24681854248047, 'timing/data': 0.05055273976176977, 'timing/model': 1.0915429331362247, 'learning_rate/qwen_vl_interface': 5.427815197968647e-07, 'learning_rate/action_model': 5.4278151979686465e-06, 'epoch': 2.46}) +[2026-06-18 07:46:13,907][starVLA.training.train_starvla][INFO] - Step 6706, Loss: {'train/loss': 0.11892993003129959, 'train/grad_norm_pre_clip': 14.069976806640625, 'timing/data': 0.031237132847309113, 'timing/model': 5.722718988545239, 'learning_rate/qwen_vl_interface': 5.424923996839647e-07, 'learning_rate/action_model': 5.4249239968396466e-06, 'epoch': 2.46}) +[2026-06-18 07:46:23,950][starVLA.training.train_starvla][INFO] - Step 6707, Loss: {'train/loss': 0.10725795477628708, 'train/grad_norm_pre_clip': 15.597390174865723, 'timing/data': 0.03689911030232906, 'timing/model': 1.3843359360471368, 'learning_rate/qwen_vl_interface': 5.422042554424557e-07, 'learning_rate/action_model': 5.4220425544245565e-06, 'epoch': 2.46}) +[2026-06-18 07:46:29,028][starVLA.training.train_starvla][INFO] - Step 6708, Loss: {'train/loss': 0.46912676095962524, 'train/grad_norm_pre_clip': 32.45680618286133, 'timing/data': 0.052878876216709614, 'timing/model': 1.0754255698993802, 'learning_rate/qwen_vl_interface': 5.419170871320705e-07, 'learning_rate/action_model': 5.419170871320706e-06, 'epoch': 2.46}) +[2026-06-18 07:46:38,468][starVLA.training.train_starvla][INFO] - Step 6709, Loss: {'train/loss': 0.2168276607990265, 'train/grad_norm_pre_clip': 24.757408142089844, 'timing/data': 0.049958204850554466, 'timing/model': 1.6347565995529294, 'learning_rate/qwen_vl_interface': 5.41630894812339e-07, 'learning_rate/action_model': 5.41630894812339e-06, 'epoch': 2.46}) +[2026-06-18 07:46:55,693][starVLA.training.train_starvla][INFO] - Step 6710, Loss: {'train/loss': 0.2059665322303772, 'train/grad_norm_pre_clip': 20.653825759887695, 'timing/data': 0.03104026708751917, 'timing/model': 1.2130955327302217, 'learning_rate/qwen_vl_interface': 5.413456785425896e-07, 'learning_rate/action_model': 5.413456785425896e-06, 'epoch': 2.46}) +[2026-06-18 07:47:04,342][starVLA.training.train_starvla][INFO] - Step 6711, Loss: {'train/loss': 0.23291826248168945, 'train/grad_norm_pre_clip': 28.838382720947266, 'timing/data': 0.0316669000312686, 'timing/model': 1.3607934573665261, 'learning_rate/qwen_vl_interface': 5.410614383819474e-07, 'learning_rate/action_model': 5.4106143838194736e-06, 'epoch': 2.46}) +[2026-06-18 07:47:09,658][starVLA.training.train_starvla][INFO] - Step 6712, Loss: {'train/loss': 0.07727580517530441, 'train/grad_norm_pre_clip': 14.310202598571777, 'timing/data': 0.042021626606583595, 'timing/model': 1.1924213422462344, 'learning_rate/qwen_vl_interface': 5.407781743893364e-07, 'learning_rate/action_model': 5.407781743893364e-06, 'epoch': 2.46}) +[2026-06-18 07:47:15,509][starVLA.training.train_starvla][INFO] - Step 6713, Loss: {'train/loss': 0.11619563400745392, 'train/grad_norm_pre_clip': 14.017924308776855, 'timing/data': 0.034638055600225925, 'timing/model': 1.2552765645086765, 'learning_rate/qwen_vl_interface': 5.404958866234772e-07, 'learning_rate/action_model': 5.404958866234772e-06, 'epoch': 2.46}) +[2026-06-18 07:47:21,006][starVLA.training.train_starvla][INFO] - Step 6714, Loss: {'train/loss': 0.045166015625, 'train/grad_norm_pre_clip': 12.356732368469238, 'timing/data': 0.038191597908735275, 'timing/model': 1.855140428058803, 'learning_rate/qwen_vl_interface': 5.402145751428884e-07, 'learning_rate/action_model': 5.402145751428884e-06, 'epoch': 2.46}) +[2026-06-18 07:47:27,173][starVLA.training.train_starvla][INFO] - Step 6715, Loss: {'train/loss': 0.5093144178390503, 'train/grad_norm_pre_clip': 36.612789154052734, 'timing/data': 0.018103504553437233, 'timing/model': 1.6088404143229127, 'learning_rate/qwen_vl_interface': 5.399342400058856e-07, 'learning_rate/action_model': 5.399342400058856e-06, 'epoch': 2.46}) +[2026-06-18 07:47:31,922][starVLA.training.train_starvla][INFO] - Step 6716, Loss: {'train/loss': 0.11128334701061249, 'train/grad_norm_pre_clip': 16.972700119018555, 'timing/data': 0.001266377978026867, 'timing/model': 1.1369126802310348, 'learning_rate/qwen_vl_interface': 5.396548812705838e-07, 'learning_rate/action_model': 5.396548812705839e-06, 'epoch': 2.46}) +[2026-06-18 07:47:36,449][starVLA.training.train_starvla][INFO] - Step 6717, Loss: {'train/loss': 0.2610892057418823, 'train/grad_norm_pre_clip': 23.601545333862305, 'timing/data': 0.0013234494253993034, 'timing/model': 1.0969645278528333, 'learning_rate/qwen_vl_interface': 5.393764989948933e-07, 'learning_rate/action_model': 5.393764989948933e-06, 'epoch': 2.46}) +[2026-06-18 07:47:52,671][starVLA.training.train_starvla][INFO] - Step 6718, Loss: {'train/loss': 0.3688616156578064, 'train/grad_norm_pre_clip': 21.648616790771484, 'timing/data': 0.0012443363666534424, 'timing/model': 1.2108799647539854, 'learning_rate/qwen_vl_interface': 5.390990932365234e-07, 'learning_rate/action_model': 5.390990932365233e-06, 'epoch': 2.46}) +[2026-06-18 07:47:57,434][starVLA.training.train_starvla][INFO] - Step 6719, Loss: {'train/loss': 0.3557664752006531, 'train/grad_norm_pre_clip': 35.95904541015625, 'timing/data': 0.0007794303819537163, 'timing/model': 1.4144663121551275, 'learning_rate/qwen_vl_interface': 5.388226640529808e-07, 'learning_rate/action_model': 5.388226640529808e-06, 'epoch': 2.46}) +[2026-06-18 07:48:02,140][starVLA.training.train_starvla][INFO] - Step 6720, Loss: {'train/loss': 0.056909192353487015, 'train/grad_norm_pre_clip': 13.139159202575684, 'timing/data': 0.0009101415053009987, 'timing/model': 1.240855791606009, 'learning_rate/qwen_vl_interface': 5.385472115015697e-07, 'learning_rate/action_model': 5.385472115015696e-06, 'epoch': 2.46}) +[2026-06-18 07:48:07,516][starVLA.training.train_starvla][INFO] - Step 6721, Loss: {'train/loss': 0.20713388919830322, 'train/grad_norm_pre_clip': 41.10091781616211, 'timing/data': 0.0013633565977215767, 'timing/model': 1.280165120959282, 'learning_rate/qwen_vl_interface': 5.382727356393909e-07, 'learning_rate/action_model': 5.382727356393909e-06, 'epoch': 2.46}) +[2026-06-18 07:48:12,628][starVLA.training.train_starvla][INFO] - Step 6722, Loss: {'train/loss': 0.9327301383018494, 'train/grad_norm_pre_clip': 52.348697662353516, 'timing/data': 0.0009143995121121407, 'timing/model': 1.0867139725014567, 'learning_rate/qwen_vl_interface': 5.379992365233442e-07, 'learning_rate/action_model': 5.379992365233442e-06, 'epoch': 2.46}) +[2026-06-18 07:48:17,507][starVLA.training.train_starvla][INFO] - Step 6723, Loss: {'train/loss': 0.2960321307182312, 'train/grad_norm_pre_clip': 26.99430274963379, 'timing/data': 0.0011590011417865753, 'timing/model': 1.3383790040388703, 'learning_rate/qwen_vl_interface': 5.377267142101261e-07, 'learning_rate/action_model': 5.37726714210126e-06, 'epoch': 2.46}) +[2026-06-18 07:48:25,433][starVLA.training.train_starvla][INFO] - Step 6724, Loss: {'train/loss': 0.08432704210281372, 'train/grad_norm_pre_clip': 20.202423095703125, 'timing/data': 0.0007813768461346626, 'timing/model': 4.782431825064123, 'learning_rate/qwen_vl_interface': 5.374551687562306e-07, 'learning_rate/action_model': 5.374551687562306e-06, 'epoch': 2.46}) +[2026-06-18 07:48:30,302][starVLA.training.train_starvla][INFO] - Step 6725, Loss: {'train/loss': 0.12326132506132126, 'train/grad_norm_pre_clip': 20.383943557739258, 'timing/data': 0.0013253670185804367, 'timing/model': 1.0553349340334535, 'learning_rate/qwen_vl_interface': 5.371846002179494e-07, 'learning_rate/action_model': 5.371846002179493e-06, 'epoch': 2.46}) +[2026-06-18 07:48:34,891][starVLA.training.train_starvla][INFO] - Step 6726, Loss: {'train/loss': 0.24897131323814392, 'train/grad_norm_pre_clip': 24.74737548828125, 'timing/data': 0.0011660614982247353, 'timing/model': 1.0420945342630148, 'learning_rate/qwen_vl_interface': 5.36915008651372e-07, 'learning_rate/action_model': 5.369150086513719e-06, 'epoch': 2.46}) +[2026-06-18 07:48:40,211][starVLA.training.train_starvla][INFO] - Step 6727, Loss: {'train/loss': 0.2885342240333557, 'train/grad_norm_pre_clip': 44.82522201538086, 'timing/data': 0.0015550730749964714, 'timing/model': 1.6365106077864766, 'learning_rate/qwen_vl_interface': 5.366463941123853e-07, 'learning_rate/action_model': 5.366463941123853e-06, 'epoch': 2.46}) +[2026-06-18 07:48:44,965][starVLA.training.train_starvla][INFO] - Step 6728, Loss: {'train/loss': 0.164298877120018, 'train/grad_norm_pre_clip': 158.1358642578125, 'timing/data': 0.0012179771438241005, 'timing/model': 1.050671599805355, 'learning_rate/qwen_vl_interface': 5.36378756656672e-07, 'learning_rate/action_model': 5.36378756656672e-06, 'epoch': 2.46}) +[2026-06-18 07:48:53,552][starVLA.training.train_starvla][INFO] - Step 6729, Loss: {'train/loss': 0.2722926437854767, 'train/grad_norm_pre_clip': 23.949562072753906, 'timing/data': 0.006905989721417427, 'timing/model': 3.3090104162693024, 'learning_rate/qwen_vl_interface': 5.361120963397151e-07, 'learning_rate/action_model': 5.361120963397152e-06, 'epoch': 2.46}) +[2026-06-18 07:49:07,067][starVLA.training.train_starvla][INFO] - Step 6730, Loss: {'train/loss': 0.16992177069187164, 'train/grad_norm_pre_clip': 18.248382568359375, 'timing/data': 0.0009245201945304871, 'timing/model': 1.244327763095498, 'learning_rate/qwen_vl_interface': 5.358464132167927e-07, 'learning_rate/action_model': 5.358464132167927e-06, 'epoch': 2.46}) +[2026-06-18 07:49:19,809][starVLA.training.train_starvla][INFO] - Step 6731, Loss: {'train/loss': 0.25757595896720886, 'train/grad_norm_pre_clip': 30.092361450195312, 'timing/data': 0.0012128949165344238, 'timing/model': 5.656771930865943, 'learning_rate/qwen_vl_interface': 5.355817073429818e-07, 'learning_rate/action_model': 5.355817073429818e-06, 'epoch': 2.46}) +[2026-06-18 07:49:33,811][starVLA.training.train_starvla][INFO] - Step 6732, Loss: {'train/loss': 0.4949107766151428, 'train/grad_norm_pre_clip': 37.03155517578125, 'timing/data': 0.0008669737726449966, 'timing/model': 1.635496348142624, 'learning_rate/qwen_vl_interface': 5.353179787731557e-07, 'learning_rate/action_model': 5.353179787731557e-06, 'epoch': 2.47}) +[2026-06-18 07:49:40,633][starVLA.training.train_starvla][INFO] - Step 6733, Loss: {'train/loss': 0.0661468580365181, 'train/grad_norm_pre_clip': 10.807333946228027, 'timing/data': 0.0011865459382534027, 'timing/model': 1.2546725571155548, 'learning_rate/qwen_vl_interface': 5.350552275619857e-07, 'learning_rate/action_model': 5.350552275619857e-06, 'epoch': 2.47}) +[2026-06-18 07:49:48,416][starVLA.training.train_starvla][INFO] - Step 6734, Loss: {'train/loss': 0.10090767592191696, 'train/grad_norm_pre_clip': 15.01800537109375, 'timing/data': 0.001112295314669609, 'timing/model': 1.377656789496541, 'learning_rate/qwen_vl_interface': 5.347934537639405e-07, 'learning_rate/action_model': 5.347934537639405e-06, 'epoch': 2.47}) +[2026-06-18 07:49:54,029][starVLA.training.train_starvla][INFO] - Step 6735, Loss: {'train/loss': 0.14419618248939514, 'train/grad_norm_pre_clip': 19.057403564453125, 'timing/data': 0.001079385168850422, 'timing/model': 1.4578384580090642, 'learning_rate/qwen_vl_interface': 5.345326574332865e-07, 'learning_rate/action_model': 5.345326574332865e-06, 'epoch': 2.47}) +[2026-06-18 07:50:02,212][starVLA.training.train_starvla][INFO] - Step 6736, Loss: {'train/loss': 0.32027384638786316, 'train/grad_norm_pre_clip': 19.99405288696289, 'timing/data': 0.0007843123748898506, 'timing/model': 4.84266281593591, 'learning_rate/qwen_vl_interface': 5.342728386240867e-07, 'learning_rate/action_model': 5.3427283862408665e-06, 'epoch': 2.47}) +[2026-06-18 07:50:14,920][starVLA.training.train_starvla][INFO] - Step 6737, Loss: {'train/loss': 0.2000647932291031, 'train/grad_norm_pre_clip': 16.70511817932129, 'timing/data': 0.0016354424878954887, 'timing/model': 2.939989590086043, 'learning_rate/qwen_vl_interface': 5.340139973902013e-07, 'learning_rate/action_model': 5.340139973902013e-06, 'epoch': 2.47}) +[2026-06-18 07:50:22,210][starVLA.training.train_starvla][INFO] - Step 6738, Loss: {'train/loss': 0.15804165601730347, 'train/grad_norm_pre_clip': 11.191420555114746, 'timing/data': 0.0013654837384819984, 'timing/model': 1.0994596919044852, 'learning_rate/qwen_vl_interface': 5.337561337852892e-07, 'learning_rate/action_model': 5.337561337852892e-06, 'epoch': 2.47}) +[2026-06-18 07:50:26,857][starVLA.training.train_starvla][INFO] - Step 6739, Loss: {'train/loss': 0.2818842828273773, 'train/grad_norm_pre_clip': 47.28960418701172, 'timing/data': 0.0026080263778567314, 'timing/model': 1.471268474124372, 'learning_rate/qwen_vl_interface': 5.334992478628054e-07, 'learning_rate/action_model': 5.334992478628054e-06, 'epoch': 2.47}) +[2026-06-18 07:50:31,943][starVLA.training.train_starvla][INFO] - Step 6740, Loss: {'train/loss': 0.20172052085399628, 'train/grad_norm_pre_clip': 24.687150955200195, 'timing/data': 0.00747350137680769, 'timing/model': 1.129278775304556, 'learning_rate/qwen_vl_interface': 5.332433396760027e-07, 'learning_rate/action_model': 5.332433396760028e-06, 'epoch': 2.47}) +[2026-06-18 07:50:36,703][starVLA.training.train_starvla][INFO] - Step 6741, Loss: {'train/loss': 0.2212461233139038, 'train/grad_norm_pre_clip': 21.466283798217773, 'timing/data': 0.0011136522516608238, 'timing/model': 1.083565159700811, 'learning_rate/qwen_vl_interface': 5.329884092779312e-07, 'learning_rate/action_model': 5.329884092779312e-06, 'epoch': 2.47}) +[2026-06-18 07:50:41,783][starVLA.training.train_starvla][INFO] - Step 6742, Loss: {'train/loss': 0.12949354946613312, 'train/grad_norm_pre_clip': 25.257484436035156, 'timing/data': 0.001513402909040451, 'timing/model': 1.035651314072311, 'learning_rate/qwen_vl_interface': 5.327344567214386e-07, 'learning_rate/action_model': 5.327344567214385e-06, 'epoch': 2.47}) +[2026-06-18 07:50:50,232][starVLA.training.train_starvla][INFO] - Step 6743, Loss: {'train/loss': 0.09197234362363815, 'train/grad_norm_pre_clip': 37.803619384765625, 'timing/data': 0.004049096256494522, 'timing/model': 4.624324733391404, 'learning_rate/qwen_vl_interface': 5.324814820591683e-07, 'learning_rate/action_model': 5.324814820591682e-06, 'epoch': 2.47}) +[2026-06-18 07:50:54,764][starVLA.training.train_starvla][INFO] - Step 6744, Loss: {'train/loss': 0.21255768835544586, 'train/grad_norm_pre_clip': 21.999711990356445, 'timing/data': 0.0009686192497611046, 'timing/model': 1.3394858064129949, 'learning_rate/qwen_vl_interface': 5.322294853435631e-07, 'learning_rate/action_model': 5.322294853435631e-06, 'epoch': 2.47}) +[2026-06-18 07:51:01,714][starVLA.training.train_starvla][INFO] - Step 6745, Loss: {'train/loss': 0.1957583874464035, 'train/grad_norm_pre_clip': 24.952489852905273, 'timing/data': 0.0011180713772773743, 'timing/model': 1.1438684910535812, 'learning_rate/qwen_vl_interface': 5.319784666268626e-07, 'learning_rate/action_model': 5.319784666268626e-06, 'epoch': 2.47}) +[2026-06-18 07:51:11,792][starVLA.training.train_starvla][INFO] - Step 6746, Loss: {'train/loss': 0.19246219098567963, 'train/grad_norm_pre_clip': 30.295900344848633, 'timing/data': 0.000891772098839283, 'timing/model': 1.0705677662044764, 'learning_rate/qwen_vl_interface': 5.317284259611024e-07, 'learning_rate/action_model': 5.317284259611024e-06, 'epoch': 2.47}) +[2026-06-18 07:51:18,626][starVLA.training.train_starvla][INFO] - Step 6747, Loss: {'train/loss': 0.0300952959805727, 'train/grad_norm_pre_clip': 5.502569675445557, 'timing/data': 0.001415933482348919, 'timing/model': 1.267965765669942, 'learning_rate/qwen_vl_interface': 5.314793633981169e-07, 'learning_rate/action_model': 5.314793633981169e-06, 'epoch': 2.47}) +[2026-06-18 07:51:27,476][starVLA.training.train_starvla][INFO] - Step 6748, Loss: {'train/loss': 0.17937448620796204, 'train/grad_norm_pre_clip': 17.39855194091797, 'timing/data': 0.0009260829538106918, 'timing/model': 1.1499427137896419, 'learning_rate/qwen_vl_interface': 5.312312789895369e-07, 'learning_rate/action_model': 5.3123127898953685e-06, 'epoch': 2.47}) +[2026-06-18 07:51:32,637][starVLA.training.train_starvla][INFO] - Step 6749, Loss: {'train/loss': 0.23250052332878113, 'train/grad_norm_pre_clip': 16.441316604614258, 'timing/data': 0.001056247390806675, 'timing/model': 1.5814084960147738, 'learning_rate/qwen_vl_interface': 5.3098417278679e-07, 'learning_rate/action_model': 5.3098417278679e-06, 'epoch': 2.47}) +[2026-06-18 07:51:48,778][starVLA.training.train_starvla][INFO] - Step 6750, Loss: {'train/loss': 0.13490253686904907, 'train/grad_norm_pre_clip': 20.107139587402344, 'timing/data': 0.0010807057842612267, 'timing/model': 5.619258787482977, 'learning_rate/qwen_vl_interface': 5.307380448411022e-07, 'learning_rate/action_model': 5.307380448411022e-06, 'epoch': 2.47}) +[2026-06-18 07:51:58,389][starVLA.training.train_starvla][INFO] - Step 6751, Loss: {'train/loss': 0.2788706123828888, 'train/grad_norm_pre_clip': 29.649005889892578, 'timing/data': 0.0008758250623941422, 'timing/model': 4.843261321075261, 'learning_rate/qwen_vl_interface': 5.304928952034955e-07, 'learning_rate/action_model': 5.304928952034956e-06, 'epoch': 2.47}) +[2026-06-18 07:52:07,845][starVLA.training.train_starvla][INFO] - Step 6752, Loss: {'train/loss': 0.41989731788635254, 'train/grad_norm_pre_clip': 34.60123825073242, 'timing/data': 0.0011830022558569908, 'timing/model': 4.725447537377477, 'learning_rate/qwen_vl_interface': 5.302487239247905e-07, 'learning_rate/action_model': 5.302487239247904e-06, 'epoch': 2.47}) +[2026-06-18 07:52:18,943][starVLA.training.train_starvla][INFO] - Step 6753, Loss: {'train/loss': 0.3356573283672333, 'train/grad_norm_pre_clip': 21.694080352783203, 'timing/data': 0.0032322974875569344, 'timing/model': 2.1078971633687615, 'learning_rate/qwen_vl_interface': 5.300055310556034e-07, 'learning_rate/action_model': 5.300055310556034e-06, 'epoch': 2.47}) +[2026-06-18 07:52:27,082][starVLA.training.train_starvla][INFO] - Step 6754, Loss: {'train/loss': 0.12995867431163788, 'train/grad_norm_pre_clip': 23.23748779296875, 'timing/data': 0.0017113657668232918, 'timing/model': 1.8670905232429504, 'learning_rate/qwen_vl_interface': 5.297633166463482e-07, 'learning_rate/action_model': 5.297633166463481e-06, 'epoch': 2.47}) +[2026-06-18 07:52:35,807][starVLA.training.train_starvla][INFO] - Step 6755, Loss: {'train/loss': 0.13311001658439636, 'train/grad_norm_pre_clip': 16.15924644470215, 'timing/data': 0.0012969765812158585, 'timing/model': 4.5652103731408715, 'learning_rate/qwen_vl_interface': 5.295220807472373e-07, 'learning_rate/action_model': 5.295220807472374e-06, 'epoch': 2.47}) +[2026-06-18 07:52:42,560][starVLA.training.train_starvla][INFO] - Step 6756, Loss: {'train/loss': 0.2307325303554535, 'train/grad_norm_pre_clip': 26.450830459594727, 'timing/data': 0.0013770442456007004, 'timing/model': 1.0561287496238947, 'learning_rate/qwen_vl_interface': 5.292818234082786e-07, 'learning_rate/action_model': 5.292818234082787e-06, 'epoch': 2.47}) +[2026-06-18 07:52:47,204][starVLA.training.train_starvla][INFO] - Step 6757, Loss: {'train/loss': 0.13155579566955566, 'train/grad_norm_pre_clip': 16.77106285095215, 'timing/data': 0.0010964302346110344, 'timing/model': 1.1165287969633937, 'learning_rate/qwen_vl_interface': 5.290425446792772e-07, 'learning_rate/action_model': 5.2904254467927716e-06, 'epoch': 2.47}) +[2026-06-18 07:52:51,650][starVLA.training.train_starvla][INFO] - Step 6758, Loss: {'train/loss': 0.22820493578910828, 'train/grad_norm_pre_clip': 22.01534080505371, 'timing/data': 0.0010110726580023766, 'timing/model': 1.4260922502726316, 'learning_rate/qwen_vl_interface': 5.288042446098364e-07, 'learning_rate/action_model': 5.2880424460983634e-06, 'epoch': 2.47}) +[2026-06-18 07:52:56,067][starVLA.training.train_starvla][INFO] - Step 6759, Loss: {'train/loss': 0.4005189836025238, 'train/grad_norm_pre_clip': 40.73204040527344, 'timing/data': 0.0010785264894366264, 'timing/model': 1.321027901954949, 'learning_rate/qwen_vl_interface': 5.28566923249356e-07, 'learning_rate/action_model': 5.28566923249356e-06, 'epoch': 2.47}) +[2026-06-18 07:53:00,220][starVLA.training.train_starvla][INFO] - Step 6760, Loss: {'train/loss': 0.572402834892273, 'train/grad_norm_pre_clip': 33.677852630615234, 'timing/data': 0.001455073244869709, 'timing/model': 1.1034456389024854, 'learning_rate/qwen_vl_interface': 5.283305806470329e-07, 'learning_rate/action_model': 5.2833058064703286e-06, 'epoch': 2.48}) +[2026-06-18 07:53:05,041][starVLA.training.train_starvla][INFO] - Step 6761, Loss: {'train/loss': 0.17169411480426788, 'train/grad_norm_pre_clip': 16.126832962036133, 'timing/data': 0.000897006131708622, 'timing/model': 1.2644663387909532, 'learning_rate/qwen_vl_interface': 5.280952168518612e-07, 'learning_rate/action_model': 5.280952168518612e-06, 'epoch': 2.48}) +[2026-06-18 07:53:09,831][starVLA.training.train_starvla][INFO] - Step 6762, Loss: {'train/loss': 0.32016241550445557, 'train/grad_norm_pre_clip': 35.97606658935547, 'timing/data': 0.001090281642973423, 'timing/model': 1.255342960357666, 'learning_rate/qwen_vl_interface': 5.278608319126318e-07, 'learning_rate/action_model': 5.278608319126318e-06, 'epoch': 2.48}) +[2026-06-18 07:53:14,543][starVLA.training.train_starvla][INFO] - Step 6763, Loss: {'train/loss': 0.27249014377593994, 'train/grad_norm_pre_clip': 25.43752670288086, 'timing/data': 0.0014063827693462372, 'timing/model': 1.271186912432313, 'learning_rate/qwen_vl_interface': 5.276274258779332e-07, 'learning_rate/action_model': 5.2762742587793315e-06, 'epoch': 2.48}) +[2026-06-18 07:53:19,234][starVLA.training.train_starvla][INFO] - Step 6764, Loss: {'train/loss': 0.4302639663219452, 'train/grad_norm_pre_clip': 28.721242904663086, 'timing/data': 0.0010754922404885292, 'timing/model': 1.4193140380084515, 'learning_rate/qwen_vl_interface': 5.273949987961508e-07, 'learning_rate/action_model': 5.273949987961508e-06, 'epoch': 2.48}) +[2026-06-18 07:53:23,381][starVLA.training.train_starvla][INFO] - Step 6765, Loss: {'train/loss': 0.20272839069366455, 'train/grad_norm_pre_clip': 18.626422882080078, 'timing/data': 0.0010573193430900574, 'timing/model': 1.0222255643457174, 'learning_rate/qwen_vl_interface': 5.271635507154671e-07, 'learning_rate/action_model': 5.2716355071546706e-06, 'epoch': 2.48}) +[2026-06-18 07:53:28,195][starVLA.training.train_starvla][INFO] - Step 6766, Loss: {'train/loss': 0.13089224696159363, 'train/grad_norm_pre_clip': 10.224552154541016, 'timing/data': 0.003827848471701145, 'timing/model': 1.0806646076962352, 'learning_rate/qwen_vl_interface': 5.269330816838614e-07, 'learning_rate/action_model': 5.269330816838613e-06, 'epoch': 2.48}) +[2026-06-18 07:53:32,795][starVLA.training.train_starvla][INFO] - Step 6767, Loss: {'train/loss': 0.16083016991615295, 'train/grad_norm_pre_clip': 20.835777282714844, 'timing/data': 0.001006104052066803, 'timing/model': 1.1051662983372808, 'learning_rate/qwen_vl_interface': 5.267035917491095e-07, 'learning_rate/action_model': 5.267035917491094e-06, 'epoch': 2.48}) +[2026-06-18 07:53:37,328][starVLA.training.train_starvla][INFO] - Step 6768, Loss: {'train/loss': 0.4885824918746948, 'train/grad_norm_pre_clip': 36.36216735839844, 'timing/data': 0.0009295688942074776, 'timing/model': 1.1254064366221428, 'learning_rate/qwen_vl_interface': 5.264750809587862e-07, 'learning_rate/action_model': 5.264750809587863e-06, 'epoch': 2.48}) +[2026-06-18 07:53:42,015][starVLA.training.train_starvla][INFO] - Step 6769, Loss: {'train/loss': 0.2402683049440384, 'train/grad_norm_pre_clip': 19.3621768951416, 'timing/data': 0.0008789161220192909, 'timing/model': 1.326317060738802, 'learning_rate/qwen_vl_interface': 5.262475493602609e-07, 'learning_rate/action_model': 5.26247549360261e-06, 'epoch': 2.48}) +[2026-06-18 07:53:46,458][starVLA.training.train_starvla][INFO] - Step 6770, Loss: {'train/loss': 0.23149296641349792, 'train/grad_norm_pre_clip': 23.950143814086914, 'timing/data': 0.0009050918743014336, 'timing/model': 1.1550123263150454, 'learning_rate/qwen_vl_interface': 5.260209970007018e-07, 'learning_rate/action_model': 5.260209970007018e-06, 'epoch': 2.48}) +[2026-06-18 07:53:54,039][starVLA.training.train_starvla][INFO] - Step 6771, Loss: {'train/loss': 0.2002989649772644, 'train/grad_norm_pre_clip': 24.65201187133789, 'timing/data': 0.011029321700334549, 'timing/model': 3.923215572722256, 'learning_rate/qwen_vl_interface': 5.257954239270732e-07, 'learning_rate/action_model': 5.257954239270732e-06, 'epoch': 2.48}) +[2026-06-18 07:54:02,509][starVLA.training.train_starvla][INFO] - Step 6772, Loss: {'train/loss': 0.3417983055114746, 'train/grad_norm_pre_clip': 38.438140869140625, 'timing/data': 0.0011708950623869896, 'timing/model': 1.2988255638629198, 'learning_rate/qwen_vl_interface': 5.255708301861367e-07, 'learning_rate/action_model': 5.255708301861367e-06, 'epoch': 2.48}) +[2026-06-18 07:54:09,581][starVLA.training.train_starvla][INFO] - Step 6773, Loss: {'train/loss': 0.30407702922821045, 'train/grad_norm_pre_clip': 34.86716079711914, 'timing/data': 0.010702625848352909, 'timing/model': 3.083948416635394, 'learning_rate/qwen_vl_interface': 5.253472158244506e-07, 'learning_rate/action_model': 5.253472158244505e-06, 'epoch': 2.48}) +[2026-06-18 07:54:20,462][starVLA.training.train_starvla][INFO] - Step 6774, Loss: {'train/loss': 0.09234905242919922, 'train/grad_norm_pre_clip': 15.412483215332031, 'timing/data': 0.04172671213746071, 'timing/model': 6.043625960126519, 'learning_rate/qwen_vl_interface': 5.251245808883707e-07, 'learning_rate/action_model': 5.251245808883707e-06, 'epoch': 2.48}) +[2026-06-18 07:54:28,616][starVLA.training.train_starvla][INFO] - Step 6775, Loss: {'train/loss': 0.12616099417209625, 'train/grad_norm_pre_clip': 15.1495361328125, 'timing/data': 0.031023958697915077, 'timing/model': 4.260515891946852, 'learning_rate/qwen_vl_interface': 5.249029254240495e-07, 'learning_rate/action_model': 5.249029254240495e-06, 'epoch': 2.48}) +[2026-06-18 07:54:33,527][starVLA.training.train_starvla][INFO] - Step 6776, Loss: {'train/loss': 0.11896158754825592, 'train/grad_norm_pre_clip': 17.324264526367188, 'timing/data': 0.0010935654863715172, 'timing/model': 1.1490958631038666, 'learning_rate/qwen_vl_interface': 5.246822494774364e-07, 'learning_rate/action_model': 5.246822494774363e-06, 'epoch': 2.48}) +[2026-06-18 07:54:41,743][starVLA.training.train_starvla][INFO] - Step 6777, Loss: {'train/loss': 0.29740074276924133, 'train/grad_norm_pre_clip': 40.46990203857422, 'timing/data': 0.0018608896061778069, 'timing/model': 1.4539003055542707, 'learning_rate/qwen_vl_interface': 5.244625530942774e-07, 'learning_rate/action_model': 5.244625530942774e-06, 'epoch': 2.48}) +[2026-06-18 07:54:47,709][starVLA.training.train_starvla][INFO] - Step 6778, Loss: {'train/loss': 0.37131834030151367, 'train/grad_norm_pre_clip': 24.753755569458008, 'timing/data': 0.0009802104905247688, 'timing/model': 1.4786673439666629, 'learning_rate/qwen_vl_interface': 5.242438363201161e-07, 'learning_rate/action_model': 5.2424383632011605e-06, 'epoch': 2.48}) +[2026-06-18 07:54:57,932][starVLA.training.train_starvla][INFO] - Step 6779, Loss: {'train/loss': 0.27471810579299927, 'train/grad_norm_pre_clip': 32.211002349853516, 'timing/data': 0.0019012847915291786, 'timing/model': 1.5178911546245217, 'learning_rate/qwen_vl_interface': 5.240260992002928e-07, 'learning_rate/action_model': 5.240260992002928e-06, 'epoch': 2.48}) +[2026-06-18 07:55:06,336][starVLA.training.train_starvla][INFO] - Step 6780, Loss: {'train/loss': 0.26065322756767273, 'train/grad_norm_pre_clip': 29.945274353027344, 'timing/data': 0.0009361570701003075, 'timing/model': 1.0641495361924171, 'learning_rate/qwen_vl_interface': 5.238093417799446e-07, 'learning_rate/action_model': 5.238093417799446e-06, 'epoch': 2.48}) +[2026-06-18 07:55:11,117][starVLA.training.train_starvla][INFO] - Step 6781, Loss: {'train/loss': 0.06226370856165886, 'train/grad_norm_pre_clip': 9.851216316223145, 'timing/data': 0.0009220056235790253, 'timing/model': 1.1489462740719318, 'learning_rate/qwen_vl_interface': 5.235935641040048e-07, 'learning_rate/action_model': 5.235935641040048e-06, 'epoch': 2.48}) +[2026-06-18 07:55:15,843][starVLA.training.train_starvla][INFO] - Step 6782, Loss: {'train/loss': 0.11347067356109619, 'train/grad_norm_pre_clip': 18.19953727722168, 'timing/data': 0.0009748348966240883, 'timing/model': 1.1092638308182359, 'learning_rate/qwen_vl_interface': 5.233787662172059e-07, 'learning_rate/action_model': 5.23378766217206e-06, 'epoch': 2.48}) +[2026-06-18 07:55:20,617][starVLA.training.train_starvla][INFO] - Step 6783, Loss: {'train/loss': 0.16930118203163147, 'train/grad_norm_pre_clip': 23.481399536132812, 'timing/data': 0.0008813301101326942, 'timing/model': 1.330696016550064, 'learning_rate/qwen_vl_interface': 5.231649481640741e-07, 'learning_rate/action_model': 5.231649481640741e-06, 'epoch': 2.48}) +[2026-06-18 07:55:25,299][starVLA.training.train_starvla][INFO] - Step 6784, Loss: {'train/loss': 0.3794959783554077, 'train/grad_norm_pre_clip': 22.292131423950195, 'timing/data': 0.0010516494512557983, 'timing/model': 1.4264984605833888, 'learning_rate/qwen_vl_interface': 5.229521099889351e-07, 'learning_rate/action_model': 5.229521099889352e-06, 'epoch': 2.48}) +[2026-06-18 07:55:29,990][starVLA.training.train_starvla][INFO] - Step 6785, Loss: {'train/loss': 0.2269233912229538, 'train/grad_norm_pre_clip': 22.258098602294922, 'timing/data': 0.0014492329210042953, 'timing/model': 1.074995112605393, 'learning_rate/qwen_vl_interface': 5.227402517359105e-07, 'learning_rate/action_model': 5.2274025173591045e-06, 'epoch': 2.48}) +[2026-06-18 07:55:34,386][starVLA.training.train_starvla][INFO] - Step 6786, Loss: {'train/loss': 0.25339990854263306, 'train/grad_norm_pre_clip': 36.425628662109375, 'timing/data': 0.0011378219351172447, 'timing/model': 1.036360690370202, 'learning_rate/qwen_vl_interface': 5.225293734489182e-07, 'learning_rate/action_model': 5.225293734489181e-06, 'epoch': 2.48}) +[2026-06-18 07:55:38,996][starVLA.training.train_starvla][INFO] - Step 6787, Loss: {'train/loss': 0.16032323241233826, 'train/grad_norm_pre_clip': 22.194595336914062, 'timing/data': 0.0016951421275734901, 'timing/model': 1.2005851920694113, 'learning_rate/qwen_vl_interface': 5.223194751716733e-07, 'learning_rate/action_model': 5.223194751716733e-06, 'epoch': 2.49}) +[2026-06-18 07:55:46,858][starVLA.training.train_starvla][INFO] - Step 6788, Loss: {'train/loss': 0.08017902821302414, 'train/grad_norm_pre_clip': 14.616080284118652, 'timing/data': 0.0012564146891236305, 'timing/model': 4.152258779853582, 'learning_rate/qwen_vl_interface': 5.221105569476893e-07, 'learning_rate/action_model': 5.221105569476892e-06, 'epoch': 2.49}) +[2026-06-18 07:55:57,105][starVLA.training.train_starvla][INFO] - Step 6789, Loss: {'train/loss': 0.25754618644714355, 'train/grad_norm_pre_clip': 22.03768539428711, 'timing/data': 0.0009037405252456665, 'timing/model': 1.4093326218426228, 'learning_rate/qwen_vl_interface': 5.21902618820274e-07, 'learning_rate/action_model': 5.21902618820274e-06, 'epoch': 2.49}) +[2026-06-18 07:56:01,379][starVLA.training.train_starvla][INFO] - Step 6790, Loss: {'train/loss': 0.2153099626302719, 'train/grad_norm_pre_clip': 23.610496520996094, 'timing/data': 0.0011043278500437737, 'timing/model': 1.1550442390143871, 'learning_rate/qwen_vl_interface': 5.216956608325331e-07, 'learning_rate/action_model': 5.2169566083253304e-06, 'epoch': 2.49}) +[2026-06-18 07:56:06,468][starVLA.training.train_starvla][INFO] - Step 6791, Loss: {'train/loss': 0.24976643919944763, 'train/grad_norm_pre_clip': 21.080656051635742, 'timing/data': 0.0016776947304606438, 'timing/model': 1.0545907281339169, 'learning_rate/qwen_vl_interface': 5.2148968302737e-07, 'learning_rate/action_model': 5.2148968302737005e-06, 'epoch': 2.49}) +[2026-06-18 07:56:12,418][starVLA.training.train_starvla][INFO] - Step 6792, Loss: {'train/loss': 0.25945764780044556, 'train/grad_norm_pre_clip': 14.218234062194824, 'timing/data': 0.0008815452456474304, 'timing/model': 1.1280606966465712, 'learning_rate/qwen_vl_interface': 5.212846854474834e-07, 'learning_rate/action_model': 5.212846854474833e-06, 'epoch': 2.49}) +[2026-06-18 07:56:17,078][starVLA.training.train_starvla][INFO] - Step 6793, Loss: {'train/loss': 0.2358095347881317, 'train/grad_norm_pre_clip': 31.203439712524414, 'timing/data': 0.0012475205585360527, 'timing/model': 1.471034443937242, 'learning_rate/qwen_vl_interface': 5.2108066813537e-07, 'learning_rate/action_model': 5.2108066813537e-06, 'epoch': 2.49}) +[2026-06-18 07:56:21,597][starVLA.training.train_starvla][INFO] - Step 6794, Loss: {'train/loss': 0.19888050854206085, 'train/grad_norm_pre_clip': 17.743898391723633, 'timing/data': 0.0010811183601617813, 'timing/model': 1.3199515054002404, 'learning_rate/qwen_vl_interface': 5.208776311333223e-07, 'learning_rate/action_model': 5.208776311333223e-06, 'epoch': 2.49}) +[2026-06-18 07:56:28,031][starVLA.training.train_starvla][INFO] - Step 6795, Loss: {'train/loss': 0.2152777910232544, 'train/grad_norm_pre_clip': 24.761987686157227, 'timing/data': 0.001380053348839283, 'timing/model': 1.1032958701252937, 'learning_rate/qwen_vl_interface': 5.206755744834312e-07, 'learning_rate/action_model': 5.2067557448343125e-06, 'epoch': 2.49}) +[2026-06-18 07:56:33,102][starVLA.training.train_starvla][INFO] - Step 6796, Loss: {'train/loss': 0.049271684139966965, 'train/grad_norm_pre_clip': 8.223000526428223, 'timing/data': 0.0025397613644599915, 'timing/model': 1.3380558071658015, 'learning_rate/qwen_vl_interface': 5.20474498227582e-07, 'learning_rate/action_model': 5.2047449822758196e-06, 'epoch': 2.49}) +[2026-06-18 07:56:40,612][starVLA.training.train_starvla][INFO] - Step 6797, Loss: {'train/loss': 0.28453192114830017, 'train/grad_norm_pre_clip': 28.076929092407227, 'timing/data': 0.0033569559454917908, 'timing/model': 4.073627430945635, 'learning_rate/qwen_vl_interface': 5.202744024074583e-07, 'learning_rate/action_model': 5.202744024074584e-06, 'epoch': 2.49}) +[2026-06-18 07:56:49,705][starVLA.training.train_starvla][INFO] - Step 6798, Loss: {'train/loss': 0.11767789721488953, 'train/grad_norm_pre_clip': 16.55986976623535, 'timing/data': 0.0007384056225419044, 'timing/model': 1.2323704827576876, 'learning_rate/qwen_vl_interface': 5.200752870645407e-07, 'learning_rate/action_model': 5.200752870645406e-06, 'epoch': 2.49}) +[2026-06-18 07:56:55,896][starVLA.training.train_starvla][INFO] - Step 6799, Loss: {'train/loss': 0.02299434132874012, 'train/grad_norm_pre_clip': 4.070487022399902, 'timing/data': 0.0010356782004237175, 'timing/model': 2.94309513643384, 'learning_rate/qwen_vl_interface': 5.198771522401054e-07, 'learning_rate/action_model': 5.198771522401054e-06, 'epoch': 2.49}) +[2026-06-18 08:01:19,912][starVLA.training.train_starvla][INFO] - Step 6800, Loss: {'train/loss': 0.16462093591690063, 'train/grad_norm_pre_clip': 28.830825805664062, 'eval/loss': 0.5446638464927673, 'eval/latency_2/loss': 0.5446638464927673, 'eval/demon_fire_precision': 0.9496076001652209, 'eval/demon_fire_recall': 0.9619246861924686, 'eval/demon_fire_f1': 0.9557264601953855, 'eval/demon_fire_rate_model': 0.6736227045075125, 'eval/demon_fire_rate_teacher': 0.6649972175848636, 'eval/demon_left_precision': 0.9130901287553648, 'eval/demon_left_recall': 0.8957894736842106, 'eval/demon_left_f1': 0.9043570669500532, 'eval/demon_left_rate_model': 0.2593210907067334, 'eval/demon_left_rate_teacher': 0.2643294379521425, 'eval/demon_right_precision': 0.9180887372013652, 'eval/demon_right_recall': 0.911864406779661, 'eval/demon_right_f1': 0.9149659863945577, 'eval/demon_right_rate_model': 0.24457429048414023, 'eval/demon_right_rate_teacher': 0.24624373956594323, 'eval/demon_move_f1': 0.9096615266723055, 'eval/demon_cc_f1': 0.941424441355471, 'eval/latency_2/demon_fire_precision': 0.9496076001652209, 'eval/latency_2/demon_fire_recall': 0.9619246861924686, 'eval/latency_2/demon_fire_f1': 0.9557264601953855, 'eval/latency_2/demon_fire_rate_model': 0.6736227045075125, 'eval/latency_2/demon_fire_rate_teacher': 0.6649972175848636, 'eval/latency_2/demon_left_precision': 0.9130901287553648, 'eval/latency_2/demon_left_recall': 0.8957894736842106, 'eval/latency_2/demon_left_f1': 0.9043570669500532, 'eval/latency_2/demon_left_rate_model': 0.2593210907067334, 'eval/latency_2/demon_left_rate_teacher': 0.2643294379521425, 'eval/latency_2/demon_right_precision': 0.9180887372013652, 'eval/latency_2/demon_right_recall': 0.911864406779661, 'eval/latency_2/demon_right_f1': 0.9149659863945577, 'eval/latency_2/demon_right_rate_model': 0.24457429048414023, 'eval/latency_2/demon_right_rate_teacher': 0.24624373956594323, 'eval/latency_2/demon_move_f1': 0.9096615266723055, 'eval/latency_2/demon_cc_f1': 0.941424441355471, 'timing/data': 0.000800631009042263, 'timing/model': 4.250217561610043, 'learning_rate/qwen_vl_interface': 5.196799979752265e-07, 'learning_rate/action_model': 5.196799979752264e-06, 'epoch': 2.49}) +[2026-06-18 08:01:19,914][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_6800_state +[2026-06-18 08:01:19,923][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 08:01:29,389][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_6800_state/model.safetensors +[2026-06-18 08:01:44,580][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_6800_state/optimizer.bin +[2026-06-18 08:01:44,582][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_6800_state/sampler.bin +[2026-06-18 08:01:44,582][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_6800_state/sampler_1.bin +[2026-06-18 08:01:44,586][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_6800_state/random_states_0.pkl +[2026-06-18 08:01:44,625][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 6800 +[2026-06-18 08:01:44,626][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 08:01:44,644][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 08:01:51,879][starVLA.training.train_starvla][INFO] - Step 6801, Loss: {'train/loss': 0.04021742194890976, 'train/grad_norm_pre_clip': 6.723947525024414, 'timing/data': 0.0018346328288316727, 'timing/model': 1.1048624124377966, 'learning_rate/qwen_vl_interface': 5.194838243107742e-07, 'learning_rate/action_model': 5.194838243107742e-06, 'epoch': 2.49}) +[2026-06-18 08:01:56,520][starVLA.training.train_starvla][INFO] - Step 6802, Loss: {'train/loss': 0.10289129614830017, 'train/grad_norm_pre_clip': 15.562314987182617, 'timing/data': 0.0011253664270043373, 'timing/model': 1.0648498041555285, 'learning_rate/qwen_vl_interface': 5.192886312874148e-07, 'learning_rate/action_model': 5.1928863128741475e-06, 'epoch': 2.49}) +[2026-06-18 08:02:01,123][starVLA.training.train_starvla][INFO] - Step 6803, Loss: {'train/loss': 0.04017218202352524, 'train/grad_norm_pre_clip': 8.898043632507324, 'timing/data': 0.0009023891761898994, 'timing/model': 1.1237100707367063, 'learning_rate/qwen_vl_interface': 5.190944189456133e-07, 'learning_rate/action_model': 5.190944189456132e-06, 'epoch': 2.49}) +[2026-06-18 08:02:05,533][starVLA.training.train_starvla][INFO] - Step 6804, Loss: {'train/loss': 0.4423135221004486, 'train/grad_norm_pre_clip': 31.681711196899414, 'timing/data': 0.0009013060480356216, 'timing/model': 1.0843223631381989, 'learning_rate/qwen_vl_interface': 5.189011873256295e-07, 'learning_rate/action_model': 5.189011873256295e-06, 'epoch': 2.49}) +[2026-06-18 08:02:12,395][starVLA.training.train_starvla][INFO] - Step 6805, Loss: {'train/loss': 0.3786764144897461, 'train/grad_norm_pre_clip': 30.1307430267334, 'timing/data': 0.0021027131006121635, 'timing/model': 2.111589202657342, 'learning_rate/qwen_vl_interface': 5.187089364675195e-07, 'learning_rate/action_model': 5.187089364675194e-06, 'epoch': 2.49}) +[2026-06-18 08:02:20,107][starVLA.training.train_starvla][INFO] - Step 6806, Loss: {'train/loss': 0.06970327347517014, 'train/grad_norm_pre_clip': 10.435566902160645, 'timing/data': 0.0011117886751890182, 'timing/model': 1.3471525227651, 'learning_rate/qwen_vl_interface': 5.18517666411139e-07, 'learning_rate/action_model': 5.18517666411139e-06, 'epoch': 2.49}) +[2026-06-18 08:02:24,412][starVLA.training.train_starvla][INFO] - Step 6807, Loss: {'train/loss': 0.026623215526342392, 'train/grad_norm_pre_clip': 5.9499125480651855, 'timing/data': 0.0013371100649237633, 'timing/model': 1.0893877847120166, 'learning_rate/qwen_vl_interface': 5.18327377196137e-07, 'learning_rate/action_model': 5.18327377196137e-06, 'epoch': 2.49}) +[2026-06-18 08:02:29,151][starVLA.training.train_starvla][INFO] - Step 6808, Loss: {'train/loss': 0.27724483609199524, 'train/grad_norm_pre_clip': 15.644762992858887, 'timing/data': 0.0011178050190210342, 'timing/model': 1.0546229742467403, 'learning_rate/qwen_vl_interface': 5.181380688619617e-07, 'learning_rate/action_model': 5.181380688619617e-06, 'epoch': 2.49}) +[2026-06-18 08:02:33,790][starVLA.training.train_starvla][INFO] - Step 6809, Loss: {'train/loss': 0.20913104712963104, 'train/grad_norm_pre_clip': 24.171907424926758, 'timing/data': 0.0009005032479763031, 'timing/model': 1.1235233433544636, 'learning_rate/qwen_vl_interface': 5.179497414478561e-07, 'learning_rate/action_model': 5.179497414478562e-06, 'epoch': 2.49}) +[2026-06-18 08:02:38,430][starVLA.training.train_starvla][INFO] - Step 6810, Loss: {'train/loss': 0.06253927946090698, 'train/grad_norm_pre_clip': 16.487104415893555, 'timing/data': 0.0013730842620134354, 'timing/model': 1.1373522840440273, 'learning_rate/qwen_vl_interface': 5.177623949928608e-07, 'learning_rate/action_model': 5.177623949928608e-06, 'epoch': 2.49}) +[2026-06-18 08:02:42,758][starVLA.training.train_starvla][INFO] - Step 6811, Loss: {'train/loss': 0.12027338892221451, 'train/grad_norm_pre_clip': 16.956377029418945, 'timing/data': 0.0017627766355872154, 'timing/model': 1.0514398096129298, 'learning_rate/qwen_vl_interface': 5.175760295358137e-07, 'learning_rate/action_model': 5.175760295358136e-06, 'epoch': 2.49}) +[2026-06-18 08:02:47,479][starVLA.training.train_starvla][INFO] - Step 6812, Loss: {'train/loss': 0.038909174501895905, 'train/grad_norm_pre_clip': 6.310312747955322, 'timing/data': 0.0010511940345168114, 'timing/model': 1.2016432425007224, 'learning_rate/qwen_vl_interface': 5.173906451153477e-07, 'learning_rate/action_model': 5.173906451153477e-06, 'epoch': 2.49}) +[2026-06-18 08:02:52,488][starVLA.training.train_starvla][INFO] - Step 6813, Loss: {'train/loss': 0.22266839444637299, 'train/grad_norm_pre_clip': 29.672170639038086, 'timing/data': 0.001574547030031681, 'timing/model': 1.068058917298913, 'learning_rate/qwen_vl_interface': 5.172062417698931e-07, 'learning_rate/action_model': 5.1720624176989304e-06, 'epoch': 2.49}) +[2026-06-18 08:02:57,423][starVLA.training.train_starvla][INFO] - Step 6814, Loss: {'train/loss': 0.22561120986938477, 'train/grad_norm_pre_clip': 23.847209930419922, 'timing/data': 0.001398014836013317, 'timing/model': 1.0836711907759309, 'learning_rate/qwen_vl_interface': 5.170228195376778e-07, 'learning_rate/action_model': 5.170228195376778e-06, 'epoch': 2.5}) +[2026-06-18 08:03:02,301][starVLA.training.train_starvla][INFO] - Step 6815, Loss: {'train/loss': 0.18455582857131958, 'train/grad_norm_pre_clip': 17.540231704711914, 'timing/data': 0.0008322903886437416, 'timing/model': 1.3014220539480448, 'learning_rate/qwen_vl_interface': 5.168403784567247e-07, 'learning_rate/action_model': 5.1684037845672475e-06, 'epoch': 2.5}) +[2026-06-18 08:03:06,865][starVLA.training.train_starvla][INFO] - Step 6816, Loss: {'train/loss': 0.0611821711063385, 'train/grad_norm_pre_clip': 13.065847396850586, 'timing/data': 0.0008825520053505898, 'timing/model': 1.4756881957873702, 'learning_rate/qwen_vl_interface': 5.166589185648543e-07, 'learning_rate/action_model': 5.1665891856485425e-06, 'epoch': 2.5}) +[2026-06-18 08:03:12,334][starVLA.training.train_starvla][INFO] - Step 6817, Loss: {'train/loss': 0.1998406946659088, 'train/grad_norm_pre_clip': 18.105093002319336, 'timing/data': 0.0009211739525198936, 'timing/model': 2.354758933186531, 'learning_rate/qwen_vl_interface': 5.164784398996836e-07, 'learning_rate/action_model': 5.164784398996836e-06, 'epoch': 2.5}) +[2026-06-18 08:03:20,318][starVLA.training.train_starvla][INFO] - Step 6818, Loss: {'train/loss': 0.34066566824913025, 'train/grad_norm_pre_clip': 30.396909713745117, 'timing/data': 0.0009422870352864265, 'timing/model': 1.0941740926355124, 'learning_rate/qwen_vl_interface': 5.162989424986253e-07, 'learning_rate/action_model': 5.162989424986253e-06, 'epoch': 2.5}) +[2026-06-18 08:03:24,959][starVLA.training.train_starvla][INFO] - Step 6819, Loss: {'train/loss': 0.16111767292022705, 'train/grad_norm_pre_clip': 18.40007972717285, 'timing/data': 0.0014739446341991425, 'timing/model': 1.13220324087888, 'learning_rate/qwen_vl_interface': 5.161204263988904e-07, 'learning_rate/action_model': 5.161204263988904e-06, 'epoch': 2.5}) +[2026-06-18 08:03:29,502][starVLA.training.train_starvla][INFO] - Step 6820, Loss: {'train/loss': 0.2729463577270508, 'train/grad_norm_pre_clip': 23.264400482177734, 'timing/data': 0.0023682070896029472, 'timing/model': 1.446113764308393, 'learning_rate/qwen_vl_interface': 5.159428916374845e-07, 'learning_rate/action_model': 5.159428916374845e-06, 'epoch': 2.5}) +[2026-06-18 08:03:34,333][starVLA.training.train_starvla][INFO] - Step 6821, Loss: {'train/loss': 0.11596060544252396, 'train/grad_norm_pre_clip': 23.53501319885254, 'timing/data': 0.00119088776409626, 'timing/model': 1.6472482765093446, 'learning_rate/qwen_vl_interface': 5.157663382512118e-07, 'learning_rate/action_model': 5.157663382512119e-06, 'epoch': 2.5}) +[2026-06-18 08:03:39,031][starVLA.training.train_starvla][INFO] - Step 6822, Loss: {'train/loss': 0.22807632386684418, 'train/grad_norm_pre_clip': 21.8851318359375, 'timing/data': 0.0011006398126482964, 'timing/model': 1.1728736888617277, 'learning_rate/qwen_vl_interface': 5.155907662766711e-07, 'learning_rate/action_model': 5.155907662766711e-06, 'epoch': 2.5}) +[2026-06-18 08:03:43,662][starVLA.training.train_starvla][INFO] - Step 6823, Loss: {'train/loss': 0.21493647992610931, 'train/grad_norm_pre_clip': 27.39263916015625, 'timing/data': 0.0010845940560102463, 'timing/model': 1.1959972372278571, 'learning_rate/qwen_vl_interface': 5.15416175750259e-07, 'learning_rate/action_model': 5.154161757502589e-06, 'epoch': 2.5}) +[2026-06-18 08:03:48,508][starVLA.training.train_starvla][INFO] - Step 6824, Loss: {'train/loss': 0.09887132793664932, 'train/grad_norm_pre_clip': 19.293893814086914, 'timing/data': 0.0013145208358764648, 'timing/model': 1.1081803133711219, 'learning_rate/qwen_vl_interface': 5.15242566708168e-07, 'learning_rate/action_model': 5.15242566708168e-06, 'epoch': 2.5}) +[2026-06-18 08:03:53,552][starVLA.training.train_starvla][INFO] - Step 6825, Loss: {'train/loss': 0.02020854875445366, 'train/grad_norm_pre_clip': 5.934633731842041, 'timing/data': 0.0009530819952487946, 'timing/model': 1.0556107629090548, 'learning_rate/qwen_vl_interface': 5.150699391863884e-07, 'learning_rate/action_model': 5.150699391863884e-06, 'epoch': 2.5}) +[2026-06-18 08:03:59,443][starVLA.training.train_starvla][INFO] - Step 6826, Loss: {'train/loss': 0.12815411388874054, 'train/grad_norm_pre_clip': 15.926365852355957, 'timing/data': 0.0014604190364480019, 'timing/model': 1.4828622983768582, 'learning_rate/qwen_vl_interface': 5.148982932207052e-07, 'learning_rate/action_model': 5.148982932207052e-06, 'epoch': 2.5}) +[2026-06-18 08:04:03,637][starVLA.training.train_starvla][INFO] - Step 6827, Loss: {'train/loss': 0.33157217502593994, 'train/grad_norm_pre_clip': 23.839920043945312, 'timing/data': 0.0013100486248731613, 'timing/model': 1.0882592415437102, 'learning_rate/qwen_vl_interface': 5.147276288467008e-07, 'learning_rate/action_model': 5.147276288467008e-06, 'epoch': 2.5}) +[2026-06-18 08:04:08,542][starVLA.training.train_starvla][INFO] - Step 6828, Loss: {'train/loss': 0.1327652484178543, 'train/grad_norm_pre_clip': 22.269454956054688, 'timing/data': 0.0009590061381459236, 'timing/model': 1.0772404354065657, 'learning_rate/qwen_vl_interface': 5.145579460997544e-07, 'learning_rate/action_model': 5.145579460997544e-06, 'epoch': 2.5}) +[2026-06-18 08:04:13,181][starVLA.training.train_starvla][INFO] - Step 6829, Loss: {'train/loss': 0.07704666256904602, 'train/grad_norm_pre_clip': 17.350631713867188, 'timing/data': 0.0008786562830209732, 'timing/model': 1.0837596859782934, 'learning_rate/qwen_vl_interface': 5.143892450150411e-07, 'learning_rate/action_model': 5.143892450150411e-06, 'epoch': 2.5}) +[2026-06-18 08:04:17,924][starVLA.training.train_starvla][INFO] - Step 6830, Loss: {'train/loss': 0.39026373624801636, 'train/grad_norm_pre_clip': 32.107879638671875, 'timing/data': 0.0009396234527230263, 'timing/model': 1.0689713768661022, 'learning_rate/qwen_vl_interface': 5.142215256275338e-07, 'learning_rate/action_model': 5.142215256275338e-06, 'epoch': 2.5}) +[2026-06-18 08:04:27,235][starVLA.training.train_starvla][INFO] - Step 6831, Loss: {'train/loss': 0.32670971751213074, 'train/grad_norm_pre_clip': 43.05323791503906, 'timing/data': 0.0009280238300561905, 'timing/model': 1.6635833140462637, 'learning_rate/qwen_vl_interface': 5.140547879719993e-07, 'learning_rate/action_model': 5.140547879719993e-06, 'epoch': 2.5}) +[2026-06-18 08:04:32,909][starVLA.training.train_starvla][INFO] - Step 6832, Loss: {'train/loss': 0.2999347448348999, 'train/grad_norm_pre_clip': 26.144372940063477, 'timing/data': 0.0011237440630793571, 'timing/model': 1.0646287938579917, 'learning_rate/qwen_vl_interface': 5.138890320830038e-07, 'learning_rate/action_model': 5.138890320830038e-06, 'epoch': 2.5}) +[2026-06-18 08:04:42,587][starVLA.training.train_starvla][INFO] - Step 6833, Loss: {'train/loss': 0.25080934166908264, 'train/grad_norm_pre_clip': 30.469816207885742, 'timing/data': 0.001276293769478798, 'timing/model': 1.1206787042319775, 'learning_rate/qwen_vl_interface': 5.137242579949081e-07, 'learning_rate/action_model': 5.137242579949081e-06, 'epoch': 2.5}) +[2026-06-18 08:04:47,649][starVLA.training.train_starvla][INFO] - Step 6834, Loss: {'train/loss': 0.14955304563045502, 'train/grad_norm_pre_clip': 19.853164672851562, 'timing/data': 0.0010436782613396645, 'timing/model': 1.1113426061347127, 'learning_rate/qwen_vl_interface': 5.135604657418702e-07, 'learning_rate/action_model': 5.135604657418702e-06, 'epoch': 2.5}) +[2026-06-18 08:04:52,381][starVLA.training.train_starvla][INFO] - Step 6835, Loss: {'train/loss': 0.4532737731933594, 'train/grad_norm_pre_clip': 32.772064208984375, 'timing/data': 0.0012709125876426697, 'timing/model': 1.0885167429223657, 'learning_rate/qwen_vl_interface': 5.133976553578443e-07, 'learning_rate/action_model': 5.1339765535784425e-06, 'epoch': 2.5}) +[2026-06-18 08:04:57,147][starVLA.training.train_starvla][INFO] - Step 6836, Loss: {'train/loss': 0.14744630455970764, 'train/grad_norm_pre_clip': 18.586214065551758, 'timing/data': 0.001186678186058998, 'timing/model': 1.419111673720181, 'learning_rate/qwen_vl_interface': 5.132358268765813e-07, 'learning_rate/action_model': 5.132358268765813e-06, 'epoch': 2.5}) +[2026-06-18 08:05:01,438][starVLA.training.train_starvla][INFO] - Step 6837, Loss: {'train/loss': 0.5915015935897827, 'train/grad_norm_pre_clip': 33.2791633605957, 'timing/data': 0.0010323980823159218, 'timing/model': 1.0588419623672962, 'learning_rate/qwen_vl_interface': 5.130749803316284e-07, 'learning_rate/action_model': 5.130749803316284e-06, 'epoch': 2.5}) +[2026-06-18 08:05:07,116][starVLA.training.train_starvla][INFO] - Step 6838, Loss: {'train/loss': 0.34852465987205505, 'train/grad_norm_pre_clip': 23.744680404663086, 'timing/data': 0.0008882805705070496, 'timing/model': 1.071644832380116, 'learning_rate/qwen_vl_interface': 5.129151157563294e-07, 'learning_rate/action_model': 5.129151157563293e-06, 'epoch': 2.5}) +[2026-06-18 08:05:11,938][starVLA.training.train_starvla][INFO] - Step 6839, Loss: {'train/loss': 0.21785776317119598, 'train/grad_norm_pre_clip': 25.788896560668945, 'timing/data': 0.0012131854891777039, 'timing/model': 1.1070347931236029, 'learning_rate/qwen_vl_interface': 5.127562331838241e-07, 'learning_rate/action_model': 5.12756233183824e-06, 'epoch': 2.5}) +[2026-06-18 08:05:16,759][starVLA.training.train_starvla][INFO] - Step 6840, Loss: {'train/loss': 0.021405449137091637, 'train/grad_norm_pre_clip': 5.916555404663086, 'timing/data': 0.001150759868323803, 'timing/model': 1.6160799330100417, 'learning_rate/qwen_vl_interface': 5.125983326470492e-07, 'learning_rate/action_model': 5.125983326470492e-06, 'epoch': 2.5}) +[2026-06-18 08:05:29,096][starVLA.training.train_starvla][INFO] - Step 6841, Loss: {'train/loss': 0.10776635259389877, 'train/grad_norm_pre_clip': 16.1375732421875, 'timing/data': 0.0009131236001849174, 'timing/model': 6.308728280477226, 'learning_rate/qwen_vl_interface': 5.124414141787372e-07, 'learning_rate/action_model': 5.124414141787372e-06, 'epoch': 2.51}) +[2026-06-18 08:05:33,813][starVLA.training.train_starvla][INFO] - Step 6842, Loss: {'train/loss': 0.26598602533340454, 'train/grad_norm_pre_clip': 25.626676559448242, 'timing/data': 0.001416538842022419, 'timing/model': 1.0356743270531297, 'learning_rate/qwen_vl_interface': 5.122854778114185e-07, 'learning_rate/action_model': 5.122854778114185e-06, 'epoch': 2.51}) +[2026-06-18 08:05:38,828][starVLA.training.train_starvla][INFO] - Step 6843, Loss: {'train/loss': 0.1824011206626892, 'train/grad_norm_pre_clip': 25.74155616760254, 'timing/data': 0.0014200005680322647, 'timing/model': 1.089301435276866, 'learning_rate/qwen_vl_interface': 5.12130523577418e-07, 'learning_rate/action_model': 5.12130523577418e-06, 'epoch': 2.51}) +[2026-06-18 08:05:43,812][starVLA.training.train_starvla][INFO] - Step 6844, Loss: {'train/loss': 0.2823147475719452, 'train/grad_norm_pre_clip': 36.48689270019531, 'timing/data': 0.0011704936623573303, 'timing/model': 1.1198004018515348, 'learning_rate/qwen_vl_interface': 5.119765515088579e-07, 'learning_rate/action_model': 5.1197655150885785e-06, 'epoch': 2.51}) +[2026-06-18 08:05:48,472][starVLA.training.train_starvla][INFO] - Step 6845, Loss: {'train/loss': 0.2807435989379883, 'train/grad_norm_pre_clip': 29.84760284423828, 'timing/data': 0.0010580429807305336, 'timing/model': 1.0555076310411096, 'learning_rate/qwen_vl_interface': 5.11823561637657e-07, 'learning_rate/action_model': 5.1182356163765705e-06, 'epoch': 2.51}) +[2026-06-18 08:05:53,201][starVLA.training.train_starvla][INFO] - Step 6846, Loss: {'train/loss': 0.27435171604156494, 'train/grad_norm_pre_clip': 21.34330940246582, 'timing/data': 0.0007529063150286674, 'timing/model': 1.473535556346178, 'learning_rate/qwen_vl_interface': 5.116715539955306e-07, 'learning_rate/action_model': 5.116715539955305e-06, 'epoch': 2.51}) +[2026-06-18 08:06:01,528][starVLA.training.train_starvla][INFO] - Step 6847, Loss: {'train/loss': 0.2247413992881775, 'train/grad_norm_pre_clip': 25.707683563232422, 'timing/data': 0.0011351332068443298, 'timing/model': 4.472519584931433, 'learning_rate/qwen_vl_interface': 5.115205286139895e-07, 'learning_rate/action_model': 5.115205286139895e-06, 'epoch': 2.51}) +[2026-06-18 08:06:09,646][starVLA.training.train_starvla][INFO] - Step 6848, Loss: {'train/loss': 0.2735646665096283, 'train/grad_norm_pre_clip': 25.389999389648438, 'timing/data': 0.0012550214305520058, 'timing/model': 3.8200371647253633, 'learning_rate/qwen_vl_interface': 5.11370485524342e-07, 'learning_rate/action_model': 5.1137048552434205e-06, 'epoch': 2.51}) +[2026-06-18 08:06:14,510][starVLA.training.train_starvla][INFO] - Step 6849, Loss: {'train/loss': 0.08027943968772888, 'train/grad_norm_pre_clip': 13.25452709197998, 'timing/data': 0.0009457292035222054, 'timing/model': 1.2217707773670554, 'learning_rate/qwen_vl_interface': 5.112214247576915e-07, 'learning_rate/action_model': 5.1122142475769155e-06, 'epoch': 2.51}) +[2026-06-18 08:06:19,357][starVLA.training.train_starvla][INFO] - Step 6850, Loss: {'train/loss': 0.16067016124725342, 'train/grad_norm_pre_clip': 60.89640808105469, 'timing/data': 0.0011614644899964333, 'timing/model': 1.1544634457677603, 'learning_rate/qwen_vl_interface': 5.110733463449384e-07, 'learning_rate/action_model': 5.110733463449385e-06, 'epoch': 2.51}) +[2026-06-18 08:06:24,570][starVLA.training.train_starvla][INFO] - Step 6851, Loss: {'train/loss': 0.22359339892864227, 'train/grad_norm_pre_clip': 19.05203628540039, 'timing/data': 0.0011620549485087395, 'timing/model': 1.4311133017763495, 'learning_rate/qwen_vl_interface': 5.109262503167808e-07, 'learning_rate/action_model': 5.109262503167808e-06, 'epoch': 2.51}) +[2026-06-18 08:06:28,919][starVLA.training.train_starvla][INFO] - Step 6852, Loss: {'train/loss': 0.3872319459915161, 'train/grad_norm_pre_clip': 31.404088973999023, 'timing/data': 0.0018268069252371788, 'timing/model': 1.1506660971790552, 'learning_rate/qwen_vl_interface': 5.107801367037106e-07, 'learning_rate/action_model': 5.107801367037107e-06, 'epoch': 2.51}) +[2026-06-18 08:06:39,007][starVLA.training.train_starvla][INFO] - Step 6853, Loss: {'train/loss': 0.41052815318107605, 'train/grad_norm_pre_clip': 34.718502044677734, 'timing/data': 0.0010530594736337662, 'timing/model': 5.909061526879668, 'learning_rate/qwen_vl_interface': 5.106350055360173e-07, 'learning_rate/action_model': 5.106350055360173e-06, 'epoch': 2.51}) +[2026-06-18 08:06:57,681][starVLA.training.train_starvla][INFO] - Step 6854, Loss: {'train/loss': 0.22951827943325043, 'train/grad_norm_pre_clip': 20.154041290283203, 'timing/data': 0.0012644417583942413, 'timing/model': 4.502635918557644, 'learning_rate/qwen_vl_interface': 5.104908568437875e-07, 'learning_rate/action_model': 5.104908568437875e-06, 'epoch': 2.51}) +[2026-06-18 08:07:04,434][starVLA.training.train_starvla][INFO] - Step 6855, Loss: {'train/loss': 0.05865863710641861, 'train/grad_norm_pre_clip': 8.202417373657227, 'timing/data': 0.0008872579783201218, 'timing/model': 1.1927662016823888, 'learning_rate/qwen_vl_interface': 5.103476906569034e-07, 'learning_rate/action_model': 5.103476906569033e-06, 'epoch': 2.51}) +[2026-06-18 08:07:14,483][starVLA.training.train_starvla][INFO] - Step 6856, Loss: {'train/loss': 0.3280327022075653, 'train/grad_norm_pre_clip': 32.05990219116211, 'timing/data': 0.0007948381826281548, 'timing/model': 1.466582639142871, 'learning_rate/qwen_vl_interface': 5.102055070050428e-07, 'learning_rate/action_model': 5.102055070050427e-06, 'epoch': 2.51}) +[2026-06-18 08:07:24,105][starVLA.training.train_starvla][INFO] - Step 6857, Loss: {'train/loss': 0.09645478427410126, 'train/grad_norm_pre_clip': 12.711339950561523, 'timing/data': 0.0013629822060465813, 'timing/model': 3.1600555274635553, 'learning_rate/qwen_vl_interface': 5.100643059176803e-07, 'learning_rate/action_model': 5.100643059176802e-06, 'epoch': 2.51}) +[2026-06-18 08:07:32,508][starVLA.training.train_starvla][INFO] - Step 6858, Loss: {'train/loss': 0.21259255707263947, 'train/grad_norm_pre_clip': 24.437644958496094, 'timing/data': 0.0010475153103470802, 'timing/model': 2.8060809383168817, 'learning_rate/qwen_vl_interface': 5.099240874240883e-07, 'learning_rate/action_model': 5.099240874240883e-06, 'epoch': 2.51}) +[2026-06-18 08:07:42,649][starVLA.training.train_starvla][INFO] - Step 6859, Loss: {'train/loss': 0.5663414001464844, 'train/grad_norm_pre_clip': 40.053348541259766, 'timing/data': 0.0011381125077605247, 'timing/model': 1.1417981563135982, 'learning_rate/qwen_vl_interface': 5.097848515533331e-07, 'learning_rate/action_model': 5.0978485155333305e-06, 'epoch': 2.51}) +[2026-06-18 08:07:51,436][starVLA.training.train_starvla][INFO] - Step 6860, Loss: {'train/loss': 0.24946175515651703, 'train/grad_norm_pre_clip': 20.985002517700195, 'timing/data': 0.0011589247733354568, 'timing/model': 1.9273457899689674, 'learning_rate/qwen_vl_interface': 5.096465983342791e-07, 'learning_rate/action_model': 5.096465983342791e-06, 'epoch': 2.51}) +[2026-06-18 08:07:56,268][starVLA.training.train_starvla][INFO] - Step 6861, Loss: {'train/loss': 0.2754601836204529, 'train/grad_norm_pre_clip': 28.31354522705078, 'timing/data': 0.0009175511077046394, 'timing/model': 1.5007730219513178, 'learning_rate/qwen_vl_interface': 5.095093277955859e-07, 'learning_rate/action_model': 5.095093277955859e-06, 'epoch': 2.51}) +[2026-06-18 08:08:02,257][starVLA.training.train_starvla][INFO] - Step 6862, Loss: {'train/loss': 0.2712244987487793, 'train/grad_norm_pre_clip': 28.197235107421875, 'timing/data': 0.0010403348132967949, 'timing/model': 1.1724725281819701, 'learning_rate/qwen_vl_interface': 5.093730399657101e-07, 'learning_rate/action_model': 5.0937303996571e-06, 'epoch': 2.51}) +[2026-06-18 08:08:12,673][starVLA.training.train_starvla][INFO] - Step 6863, Loss: {'train/loss': 0.07641807943582535, 'train/grad_norm_pre_clip': 15.344443321228027, 'timing/data': 0.0010752668604254723, 'timing/model': 1.093122229911387, 'learning_rate/qwen_vl_interface': 5.092377348729043e-07, 'learning_rate/action_model': 5.092377348729043e-06, 'epoch': 2.51}) +[2026-06-18 08:08:17,366][starVLA.training.train_starvla][INFO] - Step 6864, Loss: {'train/loss': 0.31117677688598633, 'train/grad_norm_pre_clip': 28.60904312133789, 'timing/data': 0.001292875036597252, 'timing/model': 1.0576342828571796, 'learning_rate/qwen_vl_interface': 5.091034125452171e-07, 'learning_rate/action_model': 5.091034125452171e-06, 'epoch': 2.51}) +[2026-06-18 08:08:22,024][starVLA.training.train_starvla][INFO] - Step 6865, Loss: {'train/loss': 0.2962774932384491, 'train/grad_norm_pre_clip': 23.725828170776367, 'timing/data': 0.0011926349252462387, 'timing/model': 1.034664100036025, 'learning_rate/qwen_vl_interface': 5.089700730104944e-07, 'learning_rate/action_model': 5.089700730104944e-06, 'epoch': 2.51}) +[2026-06-18 08:08:26,587][starVLA.training.train_starvla][INFO] - Step 6866, Loss: {'train/loss': 0.31908029317855835, 'train/grad_norm_pre_clip': 20.10430335998535, 'timing/data': 0.0009953910484910011, 'timing/model': 1.4594223005697131, 'learning_rate/qwen_vl_interface': 5.08837716296377e-07, 'learning_rate/action_model': 5.0883771629637696e-06, 'epoch': 2.51}) +[2026-06-18 08:08:30,856][starVLA.training.train_starvla][INFO] - Step 6867, Loss: {'train/loss': 0.2485838383436203, 'train/grad_norm_pre_clip': 27.316558837890625, 'timing/data': 0.0011807726696133614, 'timing/model': 1.1904281694442034, 'learning_rate/qwen_vl_interface': 5.087063424303026e-07, 'learning_rate/action_model': 5.087063424303026e-06, 'epoch': 2.51}) +[2026-06-18 08:08:35,641][starVLA.training.train_starvla][INFO] - Step 6868, Loss: {'train/loss': 0.3264407217502594, 'train/grad_norm_pre_clip': 27.911426544189453, 'timing/data': 0.001168183982372284, 'timing/model': 1.0690973987802863, 'learning_rate/qwen_vl_interface': 5.085759514395052e-07, 'learning_rate/action_model': 5.085759514395051e-06, 'epoch': 2.51}) +[2026-06-18 08:08:40,482][starVLA.training.train_starvla][INFO] - Step 6869, Loss: {'train/loss': 0.5512834787368774, 'train/grad_norm_pre_clip': 30.066179275512695, 'timing/data': 0.0013867001980543137, 'timing/model': 1.0660160295665264, 'learning_rate/qwen_vl_interface': 5.084465433510151e-07, 'learning_rate/action_model': 5.0844654335101515e-06, 'epoch': 2.52}) +[2026-06-18 08:08:45,042][starVLA.training.train_starvla][INFO] - Step 6870, Loss: {'train/loss': 0.2010091096162796, 'train/grad_norm_pre_clip': 19.654605865478516, 'timing/data': 0.0012154150754213333, 'timing/model': 1.1408674577251077, 'learning_rate/qwen_vl_interface': 5.083181181916588e-07, 'learning_rate/action_model': 5.083181181916588e-06, 'epoch': 2.52}) +[2026-06-18 08:08:49,928][starVLA.training.train_starvla][INFO] - Step 6871, Loss: {'train/loss': 0.25750887393951416, 'train/grad_norm_pre_clip': 20.964134216308594, 'timing/data': 0.0012823045253753662, 'timing/model': 1.6569310342893004, 'learning_rate/qwen_vl_interface': 5.081906759880585e-07, 'learning_rate/action_model': 5.081906759880585e-06, 'epoch': 2.52}) +[2026-06-18 08:08:54,560][starVLA.training.train_starvla][INFO] - Step 6872, Loss: {'train/loss': 0.19954751431941986, 'train/grad_norm_pre_clip': 21.982524871826172, 'timing/data': 0.0009851614013314247, 'timing/model': 1.1820533778518438, 'learning_rate/qwen_vl_interface': 5.080642167666338e-07, 'learning_rate/action_model': 5.080642167666337e-06, 'epoch': 2.52}) +[2026-06-18 08:08:59,180][starVLA.training.train_starvla][INFO] - Step 6873, Loss: {'train/loss': 0.21804755926132202, 'train/grad_norm_pre_clip': 26.648605346679688, 'timing/data': 0.0011550039052963257, 'timing/model': 1.0555311795324087, 'learning_rate/qwen_vl_interface': 5.079387405535995e-07, 'learning_rate/action_model': 5.079387405535995e-06, 'epoch': 2.52}) +[2026-06-18 08:09:03,746][starVLA.training.train_starvla][INFO] - Step 6874, Loss: {'train/loss': 0.07458045333623886, 'train/grad_norm_pre_clip': 8.841506004333496, 'timing/data': 0.0011994708329439163, 'timing/model': 1.0366856148466468, 'learning_rate/qwen_vl_interface': 5.07814247374967e-07, 'learning_rate/action_model': 5.078142473749671e-06, 'epoch': 2.52}) +[2026-06-18 08:09:08,585][starVLA.training.train_starvla][INFO] - Step 6875, Loss: {'train/loss': 0.33171331882476807, 'train/grad_norm_pre_clip': 20.036705017089844, 'timing/data': 0.0010675108060240746, 'timing/model': 1.125391316600144, 'learning_rate/qwen_vl_interface': 5.07690737256544e-07, 'learning_rate/action_model': 5.07690737256544e-06, 'epoch': 2.52}) +[2026-06-18 08:09:13,045][starVLA.training.train_starvla][INFO] - Step 6876, Loss: {'train/loss': 0.274580180644989, 'train/grad_norm_pre_clip': 18.17990493774414, 'timing/data': 0.0008987477049231529, 'timing/model': 1.393996479921043, 'learning_rate/qwen_vl_interface': 5.075682102239336e-07, 'learning_rate/action_model': 5.075682102239335e-06, 'epoch': 2.52}) +[2026-06-18 08:09:17,158][starVLA.training.train_starvla][INFO] - Step 6877, Loss: {'train/loss': 0.3351772725582123, 'train/grad_norm_pre_clip': 36.89809036254883, 'timing/data': 0.00111492071300745, 'timing/model': 1.026791414245963, 'learning_rate/qwen_vl_interface': 5.074466663025369e-07, 'learning_rate/action_model': 5.074466663025369e-06, 'epoch': 2.52}) +[2026-06-18 08:09:21,589][starVLA.training.train_starvla][INFO] - Step 6878, Loss: {'train/loss': 0.22133302688598633, 'train/grad_norm_pre_clip': 21.364322662353516, 'timing/data': 0.0008356086909770966, 'timing/model': 1.014230778440833, 'learning_rate/qwen_vl_interface': 5.07326105517549e-07, 'learning_rate/action_model': 5.07326105517549e-06, 'epoch': 2.52}) +[2026-06-18 08:09:26,154][starVLA.training.train_starvla][INFO] - Step 6879, Loss: {'train/loss': 0.3205064535140991, 'train/grad_norm_pre_clip': 27.095060348510742, 'timing/data': 0.0013590110465884209, 'timing/model': 1.0507579632103443, 'learning_rate/qwen_vl_interface': 5.072065278939628e-07, 'learning_rate/action_model': 5.072065278939627e-06, 'epoch': 2.52}) +[2026-06-18 08:09:31,005][starVLA.training.train_starvla][INFO] - Step 6880, Loss: {'train/loss': 0.40546005964279175, 'train/grad_norm_pre_clip': 35.371639251708984, 'timing/data': 0.0011304095387458801, 'timing/model': 1.8043833253905177, 'learning_rate/qwen_vl_interface': 5.070879334565671e-07, 'learning_rate/action_model': 5.070879334565671e-06, 'epoch': 2.52}) +[2026-06-18 08:09:37,635][starVLA.training.train_starvla][INFO] - Step 6881, Loss: {'train/loss': 0.24518759548664093, 'train/grad_norm_pre_clip': 21.411787033081055, 'timing/data': 0.0013485262170433998, 'timing/model': 1.4644681261852384, 'learning_rate/qwen_vl_interface': 5.069703222299464e-07, 'learning_rate/action_model': 5.069703222299463e-06, 'epoch': 2.52}) +[2026-06-18 08:09:41,939][starVLA.training.train_starvla][INFO] - Step 6882, Loss: {'train/loss': 0.3943537771701813, 'train/grad_norm_pre_clip': 37.911590576171875, 'timing/data': 0.0011969087645411491, 'timing/model': 1.2745159408077598, 'learning_rate/qwen_vl_interface': 5.068536942384813e-07, 'learning_rate/action_model': 5.068536942384812e-06, 'epoch': 2.52}) +[2026-06-18 08:09:46,535][starVLA.training.train_starvla][INFO] - Step 6883, Loss: {'train/loss': 0.2733626961708069, 'train/grad_norm_pre_clip': 31.206954956054688, 'timing/data': 0.0014065466821193695, 'timing/model': 1.0545336147770286, 'learning_rate/qwen_vl_interface': 5.067380495063494e-07, 'learning_rate/action_model': 5.067380495063494e-06, 'epoch': 2.52}) +[2026-06-18 08:09:51,057][starVLA.training.train_starvla][INFO] - Step 6884, Loss: {'train/loss': 0.23662354052066803, 'train/grad_norm_pre_clip': 29.88699722290039, 'timing/data': 0.0030157091096043587, 'timing/model': 1.0553376050665975, 'learning_rate/qwen_vl_interface': 5.066233880575239e-07, 'learning_rate/action_model': 5.0662338805752395e-06, 'epoch': 2.52}) +[2026-06-18 08:09:55,498][starVLA.training.train_starvla][INFO] - Step 6885, Loss: {'train/loss': 0.029364369809627533, 'train/grad_norm_pre_clip': 6.762351036071777, 'timing/data': 0.0012972485274076462, 'timing/model': 1.0627451902255416, 'learning_rate/qwen_vl_interface': 5.065097099157743e-07, 'learning_rate/action_model': 5.065097099157743e-06, 'epoch': 2.52}) +[2026-06-18 08:10:00,014][starVLA.training.train_starvla][INFO] - Step 6886, Loss: {'train/loss': 0.4675448536872864, 'train/grad_norm_pre_clip': 29.551090240478516, 'timing/data': 0.0013016574084758759, 'timing/model': 1.5274880267679691, 'learning_rate/qwen_vl_interface': 5.063970151046654e-07, 'learning_rate/action_model': 5.0639701510466545e-06, 'epoch': 2.52}) +[2026-06-18 08:10:04,154][starVLA.training.train_starvla][INFO] - Step 6887, Loss: {'train/loss': 0.36554354429244995, 'train/grad_norm_pre_clip': 44.2451286315918, 'timing/data': 0.0014235302805900574, 'timing/model': 1.052497137337923, 'learning_rate/qwen_vl_interface': 5.062853036475604e-07, 'learning_rate/action_model': 5.062853036475603e-06, 'epoch': 2.52}) +[2026-06-18 08:10:09,060][starVLA.training.train_starvla][INFO] - Step 6888, Loss: {'train/loss': 0.26250606775283813, 'train/grad_norm_pre_clip': 23.419795989990234, 'timing/data': 0.0017134882509708405, 'timing/model': 1.4034015024080873, 'learning_rate/qwen_vl_interface': 5.061745755676161e-07, 'learning_rate/action_model': 5.061745755676161e-06, 'epoch': 2.52}) +[2026-06-18 08:10:13,898][starVLA.training.train_starvla][INFO] - Step 6889, Loss: {'train/loss': 0.22894877195358276, 'train/grad_norm_pre_clip': 26.832942962646484, 'timing/data': 0.008024638518691063, 'timing/model': 1.083351481705904, 'learning_rate/qwen_vl_interface': 5.060648308877866e-07, 'learning_rate/action_model': 5.0606483088778656e-06, 'epoch': 2.52}) +[2026-06-18 08:10:18,806][starVLA.training.train_starvla][INFO] - Step 6890, Loss: {'train/loss': 0.05325542762875557, 'train/grad_norm_pre_clip': 8.227987289428711, 'timing/data': 0.0011739088222384453, 'timing/model': 1.0737581988796592, 'learning_rate/qwen_vl_interface': 5.059560696308229e-07, 'learning_rate/action_model': 5.059560696308228e-06, 'epoch': 2.52}) +[2026-06-18 08:10:23,965][starVLA.training.train_starvla][INFO] - Step 6891, Loss: {'train/loss': 0.3392319977283478, 'train/grad_norm_pre_clip': 40.43090057373047, 'timing/data': 0.0011249855160713196, 'timing/model': 1.573835777118802, 'learning_rate/qwen_vl_interface': 5.0584829181927e-07, 'learning_rate/action_model': 5.058482918192699e-06, 'epoch': 2.52}) +[2026-06-18 08:10:28,304][starVLA.training.train_starvla][INFO] - Step 6892, Loss: {'train/loss': 0.030847208574414253, 'train/grad_norm_pre_clip': 8.451455116271973, 'timing/data': 0.0017682434991002083, 'timing/model': 1.1203712653368711, 'learning_rate/qwen_vl_interface': 5.05741497475472e-07, 'learning_rate/action_model': 5.05741497475472e-06, 'epoch': 2.52}) +[2026-06-18 08:10:33,056][starVLA.training.train_starvla][INFO] - Step 6893, Loss: {'train/loss': 0.049168892204761505, 'train/grad_norm_pre_clip': 7.811830043792725, 'timing/data': 0.0015288600698113441, 'timing/model': 1.0479112640023232, 'learning_rate/qwen_vl_interface': 5.056356866215663e-07, 'learning_rate/action_model': 5.056356866215662e-06, 'epoch': 2.52}) +[2026-06-18 08:10:37,616][starVLA.training.train_starvla][INFO] - Step 6894, Loss: {'train/loss': 0.4614044725894928, 'train/grad_norm_pre_clip': 36.55025100708008, 'timing/data': 0.0019061695784330368, 'timing/model': 1.078801665455103, 'learning_rate/qwen_vl_interface': 5.055308592794876e-07, 'learning_rate/action_model': 5.0553085927948756e-06, 'epoch': 2.52}) +[2026-06-18 08:10:42,310][starVLA.training.train_starvla][INFO] - Step 6895, Loss: {'train/loss': 0.39621850848197937, 'train/grad_norm_pre_clip': 38.07394027709961, 'timing/data': 0.001040884293615818, 'timing/model': 1.0782358041033149, 'learning_rate/qwen_vl_interface': 5.054270154709674e-07, 'learning_rate/action_model': 5.054270154709673e-06, 'epoch': 2.52}) +[2026-06-18 08:10:47,693][starVLA.training.train_starvla][INFO] - Step 6896, Loss: {'train/loss': 0.33815619349479675, 'train/grad_norm_pre_clip': 33.81414794921875, 'timing/data': 0.0010948516428470612, 'timing/model': 1.5714141353964806, 'learning_rate/qwen_vl_interface': 5.053241552175322e-07, 'learning_rate/action_model': 5.053241552175321e-06, 'epoch': 2.53}) +[2026-06-18 08:10:51,903][starVLA.training.train_starvla][INFO] - Step 6897, Loss: {'train/loss': 0.05976162850856781, 'train/grad_norm_pre_clip': 8.12954044342041, 'timing/data': 0.0013186344876885414, 'timing/model': 1.0935689369216561, 'learning_rate/qwen_vl_interface': 5.05222278540505e-07, 'learning_rate/action_model': 5.05222278540505e-06, 'epoch': 2.53}) +[2026-06-18 08:10:59,074][starVLA.training.train_starvla][INFO] - Step 6898, Loss: {'train/loss': 0.026760593056678772, 'train/grad_norm_pre_clip': 6.146215915679932, 'timing/data': 0.003017776645720005, 'timing/model': 1.218173511326313, 'learning_rate/qwen_vl_interface': 5.051213854610051e-07, 'learning_rate/action_model': 5.0512138546100514e-06, 'epoch': 2.53}) +[2026-06-18 08:11:03,710][starVLA.training.train_starvla][INFO] - Step 6899, Loss: {'train/loss': 0.3633796274662018, 'train/grad_norm_pre_clip': 17.37786293029785, 'timing/data': 0.0011746492236852646, 'timing/model': 1.0479254890233278, 'learning_rate/qwen_vl_interface': 5.050214759999478e-07, 'learning_rate/action_model': 5.050214759999478e-06, 'epoch': 2.53}) +[2026-06-18 08:13:47,145][starVLA.training.train_starvla][INFO] - Step 6900, Loss: {'train/loss': 0.19507387280464172, 'train/grad_norm_pre_clip': 22.940673828125, 'eval/loss': 0.5562907457351685, 'eval/latency_2/loss': 0.5562907457351685, 'eval/demon_fire_precision': 0.9458794587945879, 'eval/demon_fire_recall': 0.9652719665271966, 'eval/demon_fire_f1': 0.9554773244978256, 'eval/demon_fire_rate_model': 0.6786310517529215, 'eval/demon_fire_rate_teacher': 0.6649972175848636, 'eval/demon_left_precision': 0.9127155172413793, 'eval/demon_left_recall': 0.891578947368421, 'eval/demon_left_f1': 0.9020234291799786, 'eval/demon_left_rate_model': 0.2582081246521981, 'eval/demon_left_rate_teacher': 0.2643294379521425, 'eval/demon_right_precision': 0.904550499445061, 'eval/demon_right_recall': 0.9209039548022598, 'eval/demon_right_f1': 0.9126539753639418, 'eval/demon_right_rate_model': 0.25069560378408456, 'eval/demon_right_rate_teacher': 0.24624373956594323, 'eval/demon_move_f1': 0.9073387022719601, 'eval/demon_cc_f1': 0.9405078043677236, 'eval/latency_2/demon_fire_precision': 0.9458794587945879, 'eval/latency_2/demon_fire_recall': 0.9652719665271966, 'eval/latency_2/demon_fire_f1': 0.9554773244978256, 'eval/latency_2/demon_fire_rate_model': 0.6786310517529215, 'eval/latency_2/demon_fire_rate_teacher': 0.6649972175848636, 'eval/latency_2/demon_left_precision': 0.9127155172413793, 'eval/latency_2/demon_left_recall': 0.891578947368421, 'eval/latency_2/demon_left_f1': 0.9020234291799786, 'eval/latency_2/demon_left_rate_model': 0.2582081246521981, 'eval/latency_2/demon_left_rate_teacher': 0.2643294379521425, 'eval/latency_2/demon_right_precision': 0.904550499445061, 'eval/latency_2/demon_right_recall': 0.9209039548022598, 'eval/latency_2/demon_right_f1': 0.9126539753639418, 'eval/latency_2/demon_right_rate_model': 0.25069560378408456, 'eval/latency_2/demon_right_rate_teacher': 0.24624373956594323, 'eval/latency_2/demon_move_f1': 0.9073387022719601, 'eval/latency_2/demon_cc_f1': 0.9405078043677236, 'timing/data': 0.0009758472442626953, 'timing/model': 1.52767323423177, 'learning_rate/qwen_vl_interface': 5.049225501780441e-07, 'learning_rate/action_model': 5.049225501780441e-06, 'epoch': 2.53}) +[2026-06-18 08:13:47,206][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_6900_state +[2026-06-18 08:13:47,216][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 08:13:56,304][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_6900_state/model.safetensors +[2026-06-18 08:14:11,512][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_6900_state/optimizer.bin +[2026-06-18 08:14:11,514][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_6900_state/sampler.bin +[2026-06-18 08:14:11,514][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_6900_state/sampler_1.bin +[2026-06-18 08:14:11,517][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_6900_state/random_states_0.pkl +[2026-06-18 08:14:11,534][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 6900 +[2026-06-18 08:14:11,535][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 08:14:11,552][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 08:14:16,220][starVLA.training.train_starvla][INFO] - Step 6901, Loss: {'train/loss': 0.19884534180164337, 'train/grad_norm_pre_clip': 27.152559280395508, 'timing/data': 0.0012757787480950356, 'timing/model': 1.0326029863208532, 'learning_rate/qwen_vl_interface': 5.048246080158016e-07, 'learning_rate/action_model': 5.048246080158016e-06, 'epoch': 2.53}) +[2026-06-18 08:14:20,575][starVLA.training.train_starvla][INFO] - Step 6902, Loss: {'train/loss': 0.15485870838165283, 'train/grad_norm_pre_clip': 16.579845428466797, 'timing/data': 0.0011490043252706528, 'timing/model': 1.3221768932417035, 'learning_rate/qwen_vl_interface': 5.047276495335243e-07, 'learning_rate/action_model': 5.047276495335242e-06, 'epoch': 2.53}) +[2026-06-18 08:14:24,635][starVLA.training.train_starvla][INFO] - Step 6903, Loss: {'train/loss': 0.29073286056518555, 'train/grad_norm_pre_clip': 31.90064811706543, 'timing/data': 0.0009656958281993866, 'timing/model': 1.0521897822618484, 'learning_rate/qwen_vl_interface': 5.046316747513115e-07, 'learning_rate/action_model': 5.046316747513115e-06, 'epoch': 2.53}) +[2026-06-18 08:14:29,030][starVLA.training.train_starvla][INFO] - Step 6904, Loss: {'train/loss': 0.06045101583003998, 'train/grad_norm_pre_clip': 12.081584930419922, 'timing/data': 0.0010821335017681122, 'timing/model': 1.0561630446463823, 'learning_rate/qwen_vl_interface': 5.045366836890585e-07, 'learning_rate/action_model': 5.045366836890584e-06, 'epoch': 2.53}) +[2026-06-18 08:14:33,505][starVLA.training.train_starvla][INFO] - Step 6905, Loss: {'train/loss': 0.11798892915248871, 'train/grad_norm_pre_clip': 22.643577575683594, 'timing/data': 0.0008923355489969254, 'timing/model': 1.0298415934666991, 'learning_rate/qwen_vl_interface': 5.044426763664572e-07, 'learning_rate/action_model': 5.044426763664571e-06, 'epoch': 2.53}) +[2026-06-18 08:14:37,875][starVLA.training.train_starvla][INFO] - Step 6906, Loss: {'train/loss': 0.09663811326026917, 'train/grad_norm_pre_clip': 14.885921478271484, 'timing/data': 0.0013245726004242897, 'timing/model': 1.36433894187212, 'learning_rate/qwen_vl_interface': 5.043496528029954e-07, 'learning_rate/action_model': 5.043496528029954e-06, 'epoch': 2.53}) +[2026-06-18 08:14:42,195][starVLA.training.train_starvla][INFO] - Step 6907, Loss: {'train/loss': 0.1878533959388733, 'train/grad_norm_pre_clip': 22.92790412902832, 'timing/data': 0.0011362573131918907, 'timing/model': 1.3082472514361143, 'learning_rate/qwen_vl_interface': 5.042576130179573e-07, 'learning_rate/action_model': 5.042576130179572e-06, 'epoch': 2.53}) +[2026-06-18 08:14:50,017][starVLA.training.train_starvla][INFO] - Step 6908, Loss: {'train/loss': 0.08274427056312561, 'train/grad_norm_pre_clip': 10.14518928527832, 'timing/data': 0.0008840328082442284, 'timing/model': 4.305573063902557, 'learning_rate/qwen_vl_interface': 5.041665570304227e-07, 'learning_rate/action_model': 5.041665570304226e-06, 'epoch': 2.53}) +[2026-06-18 08:14:56,609][starVLA.training.train_starvla][INFO] - Step 6909, Loss: {'train/loss': 0.3854371905326843, 'train/grad_norm_pre_clip': 27.86618995666504, 'timing/data': 0.0011380324140191078, 'timing/model': 1.2313851285725832, 'learning_rate/qwen_vl_interface': 5.040764848592675e-07, 'learning_rate/action_model': 5.040764848592675e-06, 'epoch': 2.53}) +[2026-06-18 08:15:01,141][starVLA.training.train_starvla][INFO] - Step 6910, Loss: {'train/loss': 0.20053939521312714, 'train/grad_norm_pre_clip': 28.26317024230957, 'timing/data': 0.001087958924472332, 'timing/model': 1.15241829585284, 'learning_rate/qwen_vl_interface': 5.039873965231644e-07, 'learning_rate/action_model': 5.039873965231643e-06, 'epoch': 2.53}) +[2026-06-18 08:15:05,522][starVLA.training.train_starvla][INFO] - Step 6911, Loss: {'train/loss': 0.3891964852809906, 'train/grad_norm_pre_clip': 58.694984436035156, 'timing/data': 0.0009901933372020721, 'timing/model': 1.0314590875059366, 'learning_rate/qwen_vl_interface': 5.0389929204058e-07, 'learning_rate/action_model': 5.0389929204058e-06, 'epoch': 2.53}) +[2026-06-18 08:15:10,027][starVLA.training.train_starvla][INFO] - Step 6912, Loss: {'train/loss': 0.10531789064407349, 'train/grad_norm_pre_clip': 13.902952194213867, 'timing/data': 0.0009718006476759911, 'timing/model': 1.3690084973350167, 'learning_rate/qwen_vl_interface': 5.038121714297802e-07, 'learning_rate/action_model': 5.038121714297802e-06, 'epoch': 2.53}) +[2026-06-18 08:15:14,221][starVLA.training.train_starvla][INFO] - Step 6913, Loss: {'train/loss': 0.46863865852355957, 'train/grad_norm_pre_clip': 49.85062026977539, 'timing/data': 0.0007590921595692635, 'timing/model': 1.0221915002912283, 'learning_rate/qwen_vl_interface': 5.037260347088236e-07, 'learning_rate/action_model': 5.037260347088236e-06, 'epoch': 2.53}) +[2026-06-18 08:15:18,659][starVLA.training.train_starvla][INFO] - Step 6914, Loss: {'train/loss': 0.0496823750436306, 'train/grad_norm_pre_clip': 10.282228469848633, 'timing/data': 0.001072453334927559, 'timing/model': 1.0360557716339827, 'learning_rate/qwen_vl_interface': 5.03640881895568e-07, 'learning_rate/action_model': 5.03640881895568e-06, 'epoch': 2.53}) +[2026-06-18 08:15:23,183][starVLA.training.train_starvla][INFO] - Step 6915, Loss: {'train/loss': 0.19452717900276184, 'train/grad_norm_pre_clip': 33.00189971923828, 'timing/data': 0.0010293414816260338, 'timing/model': 1.061361413449049, 'learning_rate/qwen_vl_interface': 5.035567130076642e-07, 'learning_rate/action_model': 5.035567130076642e-06, 'epoch': 2.53}) +[2026-06-18 08:15:27,567][starVLA.training.train_starvla][INFO] - Step 6916, Loss: {'train/loss': 0.15721754729747772, 'train/grad_norm_pre_clip': 22.493629455566406, 'timing/data': 0.0010449662804603577, 'timing/model': 1.0265269186347723, 'learning_rate/qwen_vl_interface': 5.034735280625617e-07, 'learning_rate/action_model': 5.034735280625616e-06, 'epoch': 2.53}) +[2026-06-18 08:15:32,009][starVLA.training.train_starvla][INFO] - Step 6917, Loss: {'train/loss': 0.25709426403045654, 'train/grad_norm_pre_clip': 49.16200256347656, 'timing/data': 0.000887138769030571, 'timing/model': 1.3647022899240255, 'learning_rate/qwen_vl_interface': 5.033913270775039e-07, 'learning_rate/action_model': 5.033913270775039e-06, 'epoch': 2.53}) +[2026-06-18 08:15:36,136][starVLA.training.train_starvla][INFO] - Step 6918, Loss: {'train/loss': 0.02580864354968071, 'train/grad_norm_pre_clip': 4.030684947967529, 'timing/data': 0.0010047461837530136, 'timing/model': 1.1336793461814523, 'learning_rate/qwen_vl_interface': 5.03310110069532e-07, 'learning_rate/action_model': 5.03310110069532e-06, 'epoch': 2.53}) +[2026-06-18 08:15:40,831][starVLA.training.train_starvla][INFO] - Step 6919, Loss: {'train/loss': 0.008663156069815159, 'train/grad_norm_pre_clip': 1.5044440031051636, 'timing/data': 0.0008177794516086578, 'timing/model': 1.0182387856766582, 'learning_rate/qwen_vl_interface': 5.032298770554818e-07, 'learning_rate/action_model': 5.032298770554818e-06, 'epoch': 2.53}) +[2026-06-18 08:15:45,293][starVLA.training.train_starvla][INFO] - Step 6920, Loss: {'train/loss': 0.42418423295021057, 'train/grad_norm_pre_clip': 34.19599151611328, 'timing/data': 0.001078200526535511, 'timing/model': 1.0803770031780005, 'learning_rate/qwen_vl_interface': 5.031506280519855e-07, 'learning_rate/action_model': 5.031506280519855e-06, 'epoch': 2.53}) +[2026-06-18 08:15:49,771][starVLA.training.train_starvla][INFO] - Step 6921, Loss: {'train/loss': 0.10305812209844589, 'train/grad_norm_pre_clip': 18.85354995727539, 'timing/data': 0.001475743018090725, 'timing/model': 1.3939994908869267, 'learning_rate/qwen_vl_interface': 5.030723630754723e-07, 'learning_rate/action_model': 5.030723630754723e-06, 'epoch': 2.53}) +[2026-06-18 08:15:54,866][starVLA.training.train_starvla][INFO] - Step 6922, Loss: {'train/loss': 0.2761678397655487, 'train/grad_norm_pre_clip': 27.752763748168945, 'timing/data': 0.0011351564899086952, 'timing/model': 2.105458250269294, 'learning_rate/qwen_vl_interface': 5.02995082142166e-07, 'learning_rate/action_model': 5.029950821421659e-06, 'epoch': 2.53}) +[2026-06-18 08:15:59,005][starVLA.training.train_starvla][INFO] - Step 6923, Loss: {'train/loss': 0.04517989605665207, 'train/grad_norm_pre_clip': 10.95655345916748, 'timing/data': 0.0012145796790719032, 'timing/model': 1.0862714294344187, 'learning_rate/qwen_vl_interface': 5.029187852680869e-07, 'learning_rate/action_model': 5.029187852680868e-06, 'epoch': 2.54}) +[2026-06-18 08:16:03,348][starVLA.training.train_starvla][INFO] - Step 6924, Loss: {'train/loss': 0.22598549723625183, 'train/grad_norm_pre_clip': 28.8196964263916, 'timing/data': 0.0013193190097808838, 'timing/model': 1.0335743790492415, 'learning_rate/qwen_vl_interface': 5.028434724690521e-07, 'learning_rate/action_model': 5.028434724690521e-06, 'epoch': 2.54}) +[2026-06-18 08:16:07,713][starVLA.training.train_starvla][INFO] - Step 6925, Loss: {'train/loss': 0.3723108470439911, 'train/grad_norm_pre_clip': 47.65290451049805, 'timing/data': 0.0010334746912121773, 'timing/model': 1.06394396815449, 'learning_rate/qwen_vl_interface': 5.027691437606734e-07, 'learning_rate/action_model': 5.027691437606733e-06, 'epoch': 2.54}) +[2026-06-18 08:16:12,315][starVLA.training.train_starvla][INFO] - Step 6926, Loss: {'train/loss': 0.33887678384780884, 'train/grad_norm_pre_clip': 32.07986831665039, 'timing/data': 0.0012297267094254494, 'timing/model': 1.1545498427003622, 'learning_rate/qwen_vl_interface': 5.026957991583597e-07, 'learning_rate/action_model': 5.0269579915835965e-06, 'epoch': 2.54}) +[2026-06-18 08:16:16,750][starVLA.training.train_starvla][INFO] - Step 6927, Loss: {'train/loss': 0.06798002868890762, 'train/grad_norm_pre_clip': 15.01242733001709, 'timing/data': 0.0008810451254248619, 'timing/model': 1.3432052535936236, 'learning_rate/qwen_vl_interface': 5.026234386773145e-07, 'learning_rate/action_model': 5.026234386773145e-06, 'epoch': 2.54}) +[2026-06-18 08:16:20,790][starVLA.training.train_starvla][INFO] - Step 6928, Loss: {'train/loss': 0.12319525331258774, 'train/grad_norm_pre_clip': 26.761789321899414, 'timing/data': 0.0021573519334197044, 'timing/model': 1.035100587643683, 'learning_rate/qwen_vl_interface': 5.025520623325398e-07, 'learning_rate/action_model': 5.025520623325397e-06, 'epoch': 2.54}) +[2026-06-18 08:16:25,142][starVLA.training.train_starvla][INFO] - Step 6929, Loss: {'train/loss': 0.3036257028579712, 'train/grad_norm_pre_clip': 23.537006378173828, 'timing/data': 0.0008597858250141144, 'timing/model': 1.0802002642303705, 'learning_rate/qwen_vl_interface': 5.024816701388302e-07, 'learning_rate/action_model': 5.024816701388302e-06, 'epoch': 2.54}) +[2026-06-18 08:16:31,520][starVLA.training.train_starvla][INFO] - Step 6930, Loss: {'train/loss': 0.1700555682182312, 'train/grad_norm_pre_clip': 17.580801010131836, 'timing/data': 0.001056961715221405, 'timing/model': 1.9802330378443003, 'learning_rate/qwen_vl_interface': 5.024122621107795e-07, 'learning_rate/action_model': 5.024122621107795e-06, 'epoch': 2.54}) +[2026-06-18 08:16:35,924][starVLA.training.train_starvla][INFO] - Step 6931, Loss: {'train/loss': 0.5119947195053101, 'train/grad_norm_pre_clip': 50.896305084228516, 'timing/data': 0.0008723894134163857, 'timing/model': 1.033121439628303, 'learning_rate/qwen_vl_interface': 5.023438382627749e-07, 'learning_rate/action_model': 5.0234383826277486e-06, 'epoch': 2.54}) +[2026-06-18 08:16:40,291][starVLA.training.train_starvla][INFO] - Step 6932, Loss: {'train/loss': 0.2424466460943222, 'train/grad_norm_pre_clip': 27.052486419677734, 'timing/data': 0.0011170823127031326, 'timing/model': 1.289222993887961, 'learning_rate/qwen_vl_interface': 5.022763986090018e-07, 'learning_rate/action_model': 5.022763986090018e-06, 'epoch': 2.54}) +[2026-06-18 08:16:44,503][starVLA.training.train_starvla][INFO] - Step 6933, Loss: {'train/loss': 0.20451579988002777, 'train/grad_norm_pre_clip': 23.617284774780273, 'timing/data': 0.0011220350861549377, 'timing/model': 1.063612355850637, 'learning_rate/qwen_vl_interface': 5.022099431634398e-07, 'learning_rate/action_model': 5.022099431634398e-06, 'epoch': 2.54}) +[2026-06-18 08:16:48,977][starVLA.training.train_starvla][INFO] - Step 6934, Loss: {'train/loss': 0.06254678219556808, 'train/grad_norm_pre_clip': 19.14671516418457, 'timing/data': 0.0008591078221797943, 'timing/model': 1.0369812538847327, 'learning_rate/qwen_vl_interface': 5.021444719398656e-07, 'learning_rate/action_model': 5.021444719398656e-06, 'epoch': 2.54}) +[2026-06-18 08:16:53,338][starVLA.training.train_starvla][INFO] - Step 6935, Loss: {'train/loss': 0.5669541954994202, 'train/grad_norm_pre_clip': 25.073932647705078, 'timing/data': 0.000884215347468853, 'timing/model': 1.0233011590316892, 'learning_rate/qwen_vl_interface': 5.020799849518513e-07, 'learning_rate/action_model': 5.020799849518513e-06, 'epoch': 2.54}) +[2026-06-18 08:16:57,744][starVLA.training.train_starvla][INFO] - Step 6936, Loss: {'train/loss': 0.1663365364074707, 'train/grad_norm_pre_clip': 18.372220993041992, 'timing/data': 0.0011046500876545906, 'timing/model': 1.0553052043542266, 'learning_rate/qwen_vl_interface': 5.020164822127652e-07, 'learning_rate/action_model': 5.020164822127652e-06, 'epoch': 2.54}) +[2026-06-18 08:17:02,293][starVLA.training.train_starvla][INFO] - Step 6937, Loss: {'train/loss': 0.1212477758526802, 'train/grad_norm_pre_clip': 20.375625610351562, 'timing/data': 0.001202855259180069, 'timing/model': 1.4988221237435937, 'learning_rate/qwen_vl_interface': 5.019539637357715e-07, 'learning_rate/action_model': 5.019539637357714e-06, 'epoch': 2.54}) +[2026-06-18 08:17:07,319][starVLA.training.train_starvla][INFO] - Step 6938, Loss: {'train/loss': 0.37143242359161377, 'train/grad_norm_pre_clip': 29.326982498168945, 'timing/data': 0.000986507162451744, 'timing/model': 1.0839307773858309, 'learning_rate/qwen_vl_interface': 5.0189242953383e-07, 'learning_rate/action_model': 5.018924295338299e-06, 'epoch': 2.54}) +[2026-06-18 08:17:11,964][starVLA.training.train_starvla][INFO] - Step 6939, Loss: {'train/loss': 0.5251280665397644, 'train/grad_norm_pre_clip': 35.108158111572266, 'timing/data': 0.001215747557580471, 'timing/model': 1.0839807447046041, 'learning_rate/qwen_vl_interface': 5.018318796196967e-07, 'learning_rate/action_model': 5.018318796196967e-06, 'epoch': 2.54}) +[2026-06-18 08:17:16,341][starVLA.training.train_starvla][INFO] - Step 6940, Loss: {'train/loss': 0.28737112879753113, 'train/grad_norm_pre_clip': 26.2434139251709, 'timing/data': 0.0008788714185357094, 'timing/model': 1.0277955885976553, 'learning_rate/qwen_vl_interface': 5.017723140059243e-07, 'learning_rate/action_model': 5.017723140059242e-06, 'epoch': 2.54}) +[2026-06-18 08:17:20,899][starVLA.training.train_starvla][INFO] - Step 6941, Loss: {'train/loss': 0.15181304514408112, 'train/grad_norm_pre_clip': 15.816424369812012, 'timing/data': 0.001040630042552948, 'timing/model': 1.3994161086156964, 'learning_rate/qwen_vl_interface': 5.017137327048606e-07, 'learning_rate/action_model': 5.017137327048606e-06, 'epoch': 2.54}) +[2026-06-18 08:17:25,215][starVLA.training.train_starvla][INFO] - Step 6942, Loss: {'train/loss': 0.0842537209391594, 'train/grad_norm_pre_clip': 16.016082763671875, 'timing/data': 0.000967380590736866, 'timing/model': 1.3395781069993973, 'learning_rate/qwen_vl_interface': 5.016561357286498e-07, 'learning_rate/action_model': 5.016561357286498e-06, 'epoch': 2.54}) +[2026-06-18 08:17:29,397][starVLA.training.train_starvla][INFO] - Step 6943, Loss: {'train/loss': 0.27285099029541016, 'train/grad_norm_pre_clip': 26.606224060058594, 'timing/data': 0.000928044319152832, 'timing/model': 1.0409539211541414, 'learning_rate/qwen_vl_interface': 5.015995230892312e-07, 'learning_rate/action_model': 5.015995230892311e-06, 'epoch': 2.54}) +[2026-06-18 08:17:34,131][starVLA.training.train_starvla][INFO] - Step 6944, Loss: {'train/loss': 0.18645095825195312, 'train/grad_norm_pre_clip': 20.953304290771484, 'timing/data': 0.0011865859851241112, 'timing/model': 1.2161331502720714, 'learning_rate/qwen_vl_interface': 5.015438947983409e-07, 'learning_rate/action_model': 5.015438947983409e-06, 'epoch': 2.54}) +[2026-06-18 08:17:38,969][starVLA.training.train_starvla][INFO] - Step 6945, Loss: {'train/loss': 0.1763334423303604, 'train/grad_norm_pre_clip': 18.669103622436523, 'timing/data': 0.001365690492093563, 'timing/model': 1.5461406419053674, 'learning_rate/qwen_vl_interface': 5.014892508675108e-07, 'learning_rate/action_model': 5.014892508675107e-06, 'epoch': 2.54}) +[2026-06-18 08:17:45,890][starVLA.training.train_starvla][INFO] - Step 6946, Loss: {'train/loss': 0.1215851902961731, 'train/grad_norm_pre_clip': 15.468598365783691, 'timing/data': 0.0012027313932776451, 'timing/model': 1.8544414564967155, 'learning_rate/qwen_vl_interface': 5.014355913080683e-07, 'learning_rate/action_model': 5.014355913080683e-06, 'epoch': 2.54}) +[2026-06-18 08:17:50,350][starVLA.training.train_starvla][INFO] - Step 6947, Loss: {'train/loss': 0.4625769853591919, 'train/grad_norm_pre_clip': 29.54344367980957, 'timing/data': 0.001917487010359764, 'timing/model': 1.4957886654883623, 'learning_rate/qwen_vl_interface': 5.013829161311377e-07, 'learning_rate/action_model': 5.013829161311377e-06, 'epoch': 2.54}) +[2026-06-18 08:17:54,409][starVLA.training.train_starvla][INFO] - Step 6948, Loss: {'train/loss': 0.7838209867477417, 'train/grad_norm_pre_clip': 36.05615997314453, 'timing/data': 0.0012441184371709824, 'timing/model': 1.0532486690208316, 'learning_rate/qwen_vl_interface': 5.013312253476383e-07, 'learning_rate/action_model': 5.0133122534763826e-06, 'epoch': 2.54}) +[2026-06-18 08:18:00,275][starVLA.training.train_starvla][INFO] - Step 6949, Loss: {'train/loss': 0.1430957317352295, 'train/grad_norm_pre_clip': 24.271154403686523, 'timing/data': 0.0016226693987846375, 'timing/model': 2.5659031635150313, 'learning_rate/qwen_vl_interface': 5.012805189682861e-07, 'learning_rate/action_model': 5.01280518968286e-06, 'epoch': 2.54}) +[2026-06-18 08:18:07,828][starVLA.training.train_starvla][INFO] - Step 6950, Loss: {'train/loss': 0.04695196822285652, 'train/grad_norm_pre_clip': 8.23481559753418, 'timing/data': 0.0012365970760583878, 'timing/model': 1.2621092814952135, 'learning_rate/qwen_vl_interface': 5.012307970035915e-07, 'learning_rate/action_model': 5.012307970035915e-06, 'epoch': 2.54}) +[2026-06-18 08:18:12,323][starVLA.training.train_starvla][INFO] - Step 6951, Loss: {'train/loss': 0.02403971739113331, 'train/grad_norm_pre_clip': 4.687572002410889, 'timing/data': 0.0011032652109861374, 'timing/model': 1.1210353607311845, 'learning_rate/qwen_vl_interface': 5.011820594638626e-07, 'learning_rate/action_model': 5.0118205946386256e-06, 'epoch': 2.55}) +[2026-06-18 08:18:17,190][starVLA.training.train_starvla][INFO] - Step 6952, Loss: {'train/loss': 0.08150220662355423, 'train/grad_norm_pre_clip': 20.634119033813477, 'timing/data': 0.001134985126554966, 'timing/model': 1.902118663303554, 'learning_rate/qwen_vl_interface': 5.011343063592032e-07, 'learning_rate/action_model': 5.011343063592032e-06, 'epoch': 2.55}) +[2026-06-18 08:18:21,486][starVLA.training.train_starvla][INFO] - Step 6953, Loss: {'train/loss': 0.23062805831432343, 'train/grad_norm_pre_clip': 42.52245330810547, 'timing/data': 0.0012678885832428932, 'timing/model': 1.1853587497025728, 'learning_rate/qwen_vl_interface': 5.010875376995121e-07, 'learning_rate/action_model': 5.01087537699512e-06, 'epoch': 2.55}) +[2026-06-18 08:18:25,958][starVLA.training.train_starvla][INFO] - Step 6954, Loss: {'train/loss': 0.16062059998512268, 'train/grad_norm_pre_clip': 19.13267707824707, 'timing/data': 0.001071479171514511, 'timing/model': 1.0571621246635914, 'learning_rate/qwen_vl_interface': 5.010417534944839e-07, 'learning_rate/action_model': 5.010417534944838e-06, 'epoch': 2.55}) +[2026-06-18 08:18:30,335][starVLA.training.train_starvla][INFO] - Step 6955, Loss: {'train/loss': 0.2977001667022705, 'train/grad_norm_pre_clip': 31.663976669311523, 'timing/data': 0.0013936441391706467, 'timing/model': 1.0447731111198664, 'learning_rate/qwen_vl_interface': 5.009969537536106e-07, 'learning_rate/action_model': 5.009969537536106e-06, 'epoch': 2.55}) +[2026-06-18 08:18:34,877][starVLA.training.train_starvla][INFO] - Step 6956, Loss: {'train/loss': 0.17078059911727905, 'train/grad_norm_pre_clip': 20.966468811035156, 'timing/data': 0.0010271910578012466, 'timing/model': 1.03119592461735, 'learning_rate/qwen_vl_interface': 5.00953138486179e-07, 'learning_rate/action_model': 5.00953138486179e-06, 'epoch': 2.55}) +[2026-06-18 08:18:39,291][starVLA.training.train_starvla][INFO] - Step 6957, Loss: {'train/loss': 0.3183358609676361, 'train/grad_norm_pre_clip': 40.316925048828125, 'timing/data': 0.0011044861748814583, 'timing/model': 1.446356987580657, 'learning_rate/qwen_vl_interface': 5.009103077012723e-07, 'learning_rate/action_model': 5.009103077012723e-06, 'epoch': 2.55}) +[2026-06-18 08:18:43,443][starVLA.training.train_starvla][INFO] - Step 6958, Loss: {'train/loss': 0.21812556684017181, 'train/grad_norm_pre_clip': 22.49842643737793, 'timing/data': 0.0011688768863677979, 'timing/model': 1.1050720559433103, 'learning_rate/qwen_vl_interface': 5.008684614077684e-07, 'learning_rate/action_model': 5.008684614077684e-06, 'epoch': 2.55}) +[2026-06-18 08:18:47,826][starVLA.training.train_starvla][INFO] - Step 6959, Loss: {'train/loss': 0.12472435086965561, 'train/grad_norm_pre_clip': 23.321556091308594, 'timing/data': 0.0009609675034880638, 'timing/model': 1.031116814352572, 'learning_rate/qwen_vl_interface': 5.008275996143429e-07, 'learning_rate/action_model': 5.0082759961434295e-06, 'epoch': 2.55}) +[2026-06-18 08:18:52,169][starVLA.training.train_starvla][INFO] - Step 6960, Loss: {'train/loss': 0.39169222116470337, 'train/grad_norm_pre_clip': 29.676298141479492, 'timing/data': 0.0010829176753759384, 'timing/model': 1.0522564677521586, 'learning_rate/qwen_vl_interface': 5.007877223294668e-07, 'learning_rate/action_model': 5.0078772232946675e-06, 'epoch': 2.55}) +[2026-06-18 08:18:56,869][starVLA.training.train_starvla][INFO] - Step 6961, Loss: {'train/loss': 0.49147942662239075, 'train/grad_norm_pre_clip': 36.67348098754883, 'timing/data': 0.0013405727222561836, 'timing/model': 1.4100859612226486, 'learning_rate/qwen_vl_interface': 5.007488295614058e-07, 'learning_rate/action_model': 5.007488295614058e-06, 'epoch': 2.55}) +[2026-06-18 08:19:01,225][starVLA.training.train_starvla][INFO] - Step 6962, Loss: {'train/loss': 0.17063042521476746, 'train/grad_norm_pre_clip': 17.050533294677734, 'timing/data': 0.0012431545183062553, 'timing/model': 1.3192001823335886, 'learning_rate/qwen_vl_interface': 5.007109213182231e-07, 'learning_rate/action_model': 5.007109213182231e-06, 'epoch': 2.55}) +[2026-06-18 08:19:05,310][starVLA.training.train_starvla][INFO] - Step 6963, Loss: {'train/loss': 0.14909856021404266, 'train/grad_norm_pre_clip': 19.892677307128906, 'timing/data': 0.0008611083030700684, 'timing/model': 1.0499514658004045, 'learning_rate/qwen_vl_interface': 5.006739976077769e-07, 'learning_rate/action_model': 5.0067399760777695e-06, 'epoch': 2.55}) +[2026-06-18 08:19:10,030][starVLA.training.train_starvla][INFO] - Step 6964, Loss: {'train/loss': 0.33976510167121887, 'train/grad_norm_pre_clip': 43.89663314819336, 'timing/data': 0.0011610016226768494, 'timing/model': 1.1430051624774933, 'learning_rate/qwen_vl_interface': 5.006380584377213e-07, 'learning_rate/action_model': 5.006380584377213e-06, 'epoch': 2.55}) +[2026-06-18 08:19:14,347][starVLA.training.train_starvla][INFO] - Step 6965, Loss: {'train/loss': 0.20185884833335876, 'train/grad_norm_pre_clip': 20.21074104309082, 'timing/data': 0.001086619682610035, 'timing/model': 1.0350506082177162, 'learning_rate/qwen_vl_interface': 5.006031038155068e-07, 'learning_rate/action_model': 5.006031038155067e-06, 'epoch': 2.55}) +[2026-06-18 08:19:18,780][starVLA.training.train_starvla][INFO] - Step 6966, Loss: {'train/loss': 0.22218036651611328, 'train/grad_norm_pre_clip': 25.427656173706055, 'timing/data': 0.000910620205104351, 'timing/model': 1.0365204326808453, 'learning_rate/qwen_vl_interface': 5.00569133748379e-07, 'learning_rate/action_model': 5.00569133748379e-06, 'epoch': 2.55}) +[2026-06-18 08:19:23,259][starVLA.training.train_starvla][INFO] - Step 6967, Loss: {'train/loss': 0.3284381926059723, 'train/grad_norm_pre_clip': 24.841100692749023, 'timing/data': 0.001342201605439186, 'timing/model': 1.4366199215874076, 'learning_rate/qwen_vl_interface': 5.005361482433811e-07, 'learning_rate/action_model': 5.005361482433811e-06, 'epoch': 2.55}) +[2026-06-18 08:19:27,557][starVLA.training.train_starvla][INFO] - Step 6968, Loss: {'train/loss': 0.23883874714374542, 'train/grad_norm_pre_clip': 22.559354782104492, 'timing/data': 0.001034722663462162, 'timing/model': 1.0802257573232055, 'learning_rate/qwen_vl_interface': 5.005041473073501e-07, 'learning_rate/action_model': 5.005041473073501e-06, 'epoch': 2.55}) +[2026-06-18 08:19:32,127][starVLA.training.train_starvla][INFO] - Step 6969, Loss: {'train/loss': 0.17298199236392975, 'train/grad_norm_pre_clip': 27.007766723632812, 'timing/data': 0.0010756030678749084, 'timing/model': 1.0708388602361083, 'learning_rate/qwen_vl_interface': 5.004731309469201e-07, 'learning_rate/action_model': 5.004731309469201e-06, 'epoch': 2.55}) +[2026-06-18 08:19:36,512][starVLA.training.train_starvla][INFO] - Step 6970, Loss: {'train/loss': 0.2293117642402649, 'train/grad_norm_pre_clip': 29.678165435791016, 'timing/data': 0.0009956853464245796, 'timing/model': 1.0627669896930456, 'learning_rate/qwen_vl_interface': 5.004430991685201e-07, 'learning_rate/action_model': 5.0044309916852016e-06, 'epoch': 2.55}) +[2026-06-18 08:19:40,951][starVLA.training.train_starvla][INFO] - Step 6971, Loss: {'train/loss': 0.06693854928016663, 'train/grad_norm_pre_clip': 21.681596755981445, 'timing/data': 0.0009979987516999245, 'timing/model': 1.0530157601460814, 'learning_rate/qwen_vl_interface': 5.004140519783773e-07, 'learning_rate/action_model': 5.0041405197837725e-06, 'epoch': 2.55}) +[2026-06-18 08:19:45,200][starVLA.training.train_starvla][INFO] - Step 6972, Loss: {'train/loss': 0.19045574963092804, 'train/grad_norm_pre_clip': 20.34063720703125, 'timing/data': 0.0009104348719120026, 'timing/model': 1.2774851135909557, 'learning_rate/qwen_vl_interface': 5.003859893825117e-07, 'learning_rate/action_model': 5.003859893825117e-06, 'epoch': 2.55}) +[2026-06-18 08:19:49,349][starVLA.training.train_starvla][INFO] - Step 6973, Loss: {'train/loss': 0.18463200330734253, 'train/grad_norm_pre_clip': 30.868488311767578, 'timing/data': 0.0010850448161363602, 'timing/model': 1.04037656635046, 'learning_rate/qwen_vl_interface': 5.003589113867419e-07, 'learning_rate/action_model': 5.003589113867419e-06, 'epoch': 2.55}) +[2026-06-18 08:19:53,802][starVLA.training.train_starvla][INFO] - Step 6974, Loss: {'train/loss': 0.22782255709171295, 'train/grad_norm_pre_clip': 45.35702133178711, 'timing/data': 0.0017937961965799332, 'timing/model': 1.0917493412271142, 'learning_rate/qwen_vl_interface': 5.003328179966799e-07, 'learning_rate/action_model': 5.003328179966798e-06, 'epoch': 2.55}) +[2026-06-18 08:19:58,304][starVLA.training.train_starvla][INFO] - Step 6975, Loss: {'train/loss': 0.3511011302471161, 'train/grad_norm_pre_clip': 48.414642333984375, 'timing/data': 0.0010872641578316689, 'timing/model': 1.0788401700556278, 'learning_rate/qwen_vl_interface': 5.003077092177363e-07, 'learning_rate/action_model': 5.003077092177363e-06, 'epoch': 2.55}) +[2026-06-18 08:20:02,912][starVLA.training.train_starvla][INFO] - Step 6976, Loss: {'train/loss': 0.07165759056806564, 'train/grad_norm_pre_clip': 9.57798957824707, 'timing/data': 0.0010835081338882446, 'timing/model': 1.024908261373639, 'learning_rate/qwen_vl_interface': 5.002835850551146e-07, 'learning_rate/action_model': 5.002835850551146e-06, 'epoch': 2.55}) +[2026-06-18 08:20:07,269][starVLA.training.train_starvla][INFO] - Step 6977, Loss: {'train/loss': 0.10081049799919128, 'train/grad_norm_pre_clip': 16.213035583496094, 'timing/data': 0.0012011714279651642, 'timing/model': 1.367531206458807, 'learning_rate/qwen_vl_interface': 5.002604455138174e-07, 'learning_rate/action_model': 5.002604455138173e-06, 'epoch': 2.55}) +[2026-06-18 08:20:11,431][starVLA.training.train_starvla][INFO] - Step 6978, Loss: {'train/loss': 0.3260093331336975, 'train/grad_norm_pre_clip': 226.8781280517578, 'timing/data': 0.0011404752731323242, 'timing/model': 1.0477747581899166, 'learning_rate/qwen_vl_interface': 5.002382905986407e-07, 'learning_rate/action_model': 5.002382905986407e-06, 'epoch': 2.56}) +[2026-06-18 08:20:15,822][starVLA.training.train_starvla][INFO] - Step 6979, Loss: {'train/loss': 0.18093065917491913, 'train/grad_norm_pre_clip': 22.829973220825195, 'timing/data': 0.0010816175490617752, 'timing/model': 1.0268610808998346, 'learning_rate/qwen_vl_interface': 5.002171203141771e-07, 'learning_rate/action_model': 5.0021712031417704e-06, 'epoch': 2.56}) +[2026-06-18 08:20:20,384][starVLA.training.train_starvla][INFO] - Step 6980, Loss: {'train/loss': 0.1713852435350418, 'train/grad_norm_pre_clip': 25.900297164916992, 'timing/data': 0.0012463247403502464, 'timing/model': 1.0445222705602646, 'learning_rate/qwen_vl_interface': 5.001969346648154e-07, 'learning_rate/action_model': 5.001969346648154e-06, 'epoch': 2.56}) +[2026-06-18 08:20:24,808][starVLA.training.train_starvla][INFO] - Step 6981, Loss: {'train/loss': 0.09711743146181107, 'train/grad_norm_pre_clip': 18.32933235168457, 'timing/data': 0.0008944477885961533, 'timing/model': 1.4342044470831752, 'learning_rate/qwen_vl_interface': 5.0017773365474e-07, 'learning_rate/action_model': 5.001777336547399e-06, 'epoch': 2.56}) +[2026-06-18 08:20:29,125][starVLA.training.train_starvla][INFO] - Step 6982, Loss: {'train/loss': 0.23320835828781128, 'train/grad_norm_pre_clip': 24.030378341674805, 'timing/data': 0.001138460822403431, 'timing/model': 1.343606524169445, 'learning_rate/qwen_vl_interface': 5.001595172879317e-07, 'learning_rate/action_model': 5.001595172879316e-06, 'epoch': 2.56}) +[2026-06-18 08:20:33,139][starVLA.training.train_starvla][INFO] - Step 6983, Loss: {'train/loss': 0.12147755175828934, 'train/grad_norm_pre_clip': 20.21504783630371, 'timing/data': 0.0030566323548555374, 'timing/model': 1.0274261869490147, 'learning_rate/qwen_vl_interface': 5.001422855681669e-07, 'learning_rate/action_model': 5.001422855681669e-06, 'epoch': 2.56}) +[2026-06-18 08:20:37,671][starVLA.training.train_starvla][INFO] - Step 6984, Loss: {'train/loss': 0.4176506996154785, 'train/grad_norm_pre_clip': 35.2392578125, 'timing/data': 0.0012426050379872322, 'timing/model': 1.133451683446765, 'learning_rate/qwen_vl_interface': 5.001260384990169e-07, 'learning_rate/action_model': 5.001260384990169e-06, 'epoch': 2.56}) +[2026-06-18 08:20:42,515][starVLA.training.train_starvla][INFO] - Step 6985, Loss: {'train/loss': 0.23665393888950348, 'train/grad_norm_pre_clip': 23.15025520324707, 'timing/data': 0.00132724829018116, 'timing/model': 1.3315785406157374, 'learning_rate/qwen_vl_interface': 5.001107760838504e-07, 'learning_rate/action_model': 5.001107760838504e-06, 'epoch': 2.56}) +[2026-06-18 08:20:47,026][starVLA.training.train_starvla][INFO] - Step 6986, Loss: {'train/loss': 0.1894158124923706, 'train/grad_norm_pre_clip': 21.60993003845215, 'timing/data': 0.0008945968002080917, 'timing/model': 1.031247777864337, 'learning_rate/qwen_vl_interface': 5.000964983258306e-07, 'learning_rate/action_model': 5.000964983258306e-06, 'epoch': 2.56}) +[2026-06-18 08:20:52,413][starVLA.training.train_starvla][INFO] - Step 6987, Loss: {'train/loss': 0.05748318135738373, 'train/grad_norm_pre_clip': 8.48563003540039, 'timing/data': 0.0009069368243217468, 'timing/model': 1.6887886663898826, 'learning_rate/qwen_vl_interface': 5.000832052279185e-07, 'learning_rate/action_model': 5.000832052279185e-06, 'epoch': 2.56}) +[2026-06-18 08:20:56,454][starVLA.training.train_starvla][INFO] - Step 6988, Loss: {'train/loss': 0.2576662003993988, 'train/grad_norm_pre_clip': 39.312950134277344, 'timing/data': 0.0007709572091698647, 'timing/model': 1.0307426042854786, 'learning_rate/qwen_vl_interface': 5.000708967928686e-07, 'learning_rate/action_model': 5.000708967928685e-06, 'epoch': 2.56}) +[2026-06-18 08:21:01,343][starVLA.training.train_starvla][INFO] - Step 6989, Loss: {'train/loss': 0.15650713443756104, 'train/grad_norm_pre_clip': 22.706161499023438, 'timing/data': 0.0009422143921256065, 'timing/model': 1.0796193182468414, 'learning_rate/qwen_vl_interface': 5.00059573023233e-07, 'learning_rate/action_model': 5.0005957302323305e-06, 'epoch': 2.56}) +[2026-06-18 08:21:07,007][starVLA.training.train_starvla][INFO] - Step 6990, Loss: {'train/loss': 0.03701705113053322, 'train/grad_norm_pre_clip': 5.01486873626709, 'timing/data': 0.0010961377993226051, 'timing/model': 1.2635411554947495, 'learning_rate/qwen_vl_interface': 5.000492339213594e-07, 'learning_rate/action_model': 5.000492339213594e-06, 'epoch': 2.56}) +[2026-06-18 08:21:11,630][starVLA.training.train_starvla][INFO] - Step 6991, Loss: {'train/loss': 0.22386598587036133, 'train/grad_norm_pre_clip': 20.824922561645508, 'timing/data': 0.0009530745446681976, 'timing/model': 1.0723106563091278, 'learning_rate/qwen_vl_interface': 5.000398794893907e-07, 'learning_rate/action_model': 5.000398794893906e-06, 'epoch': 2.56}) +[2026-06-18 08:21:16,194][starVLA.training.train_starvla][INFO] - Step 6992, Loss: {'train/loss': 0.41134604811668396, 'train/grad_norm_pre_clip': 32.39986801147461, 'timing/data': 0.0010429667308926582, 'timing/model': 1.2757713347673416, 'learning_rate/qwen_vl_interface': 5.000315097292659e-07, 'learning_rate/action_model': 5.000315097292659e-06, 'epoch': 2.56}) +[2026-06-18 08:21:20,326][starVLA.training.train_starvla][INFO] - Step 6993, Loss: {'train/loss': 0.32667309045791626, 'train/grad_norm_pre_clip': 29.962621688842773, 'timing/data': 0.0011361604556441307, 'timing/model': 1.05526611674577, 'learning_rate/qwen_vl_interface': 5.000241246427207e-07, 'learning_rate/action_model': 5.000241246427207e-06, 'epoch': 2.56}) +[2026-06-18 08:21:24,864][starVLA.training.train_starvla][INFO] - Step 6994, Loss: {'train/loss': 0.15300804376602173, 'train/grad_norm_pre_clip': 24.61732292175293, 'timing/data': 0.0011390633881092072, 'timing/model': 1.0552350264042616, 'learning_rate/qwen_vl_interface': 5.000177242312855e-07, 'learning_rate/action_model': 5.000177242312855e-06, 'epoch': 2.56}) +[2026-06-18 08:21:29,427][starVLA.training.train_starvla][INFO] - Step 6995, Loss: {'train/loss': 0.28567618131637573, 'train/grad_norm_pre_clip': 47.412940979003906, 'timing/data': 0.0013108663260936737, 'timing/model': 1.0380364144220948, 'learning_rate/qwen_vl_interface': 5.00012308496287e-07, 'learning_rate/action_model': 5.000123084962871e-06, 'epoch': 2.56}) +[2026-06-18 08:21:33,786][starVLA.training.train_starvla][INFO] - Step 6996, Loss: {'train/loss': 0.2810525894165039, 'train/grad_norm_pre_clip': 43.73307418823242, 'timing/data': 0.0010905805975198746, 'timing/model': 1.0359244924038649, 'learning_rate/qwen_vl_interface': 5.000078774388487e-07, 'learning_rate/action_model': 5.000078774388487e-06, 'epoch': 2.56}) +[2026-06-18 08:21:38,254][starVLA.training.train_starvla][INFO] - Step 6997, Loss: {'train/loss': 0.5587499141693115, 'train/grad_norm_pre_clip': 59.8240966796875, 'timing/data': 0.001221977174282074, 'timing/model': 1.416851757094264, 'learning_rate/qwen_vl_interface': 5.000044310598879e-07, 'learning_rate/action_model': 5.00004431059888e-06, 'epoch': 2.56}) +[2026-06-18 08:21:42,344][starVLA.training.train_starvla][INFO] - Step 6998, Loss: {'train/loss': 0.24492168426513672, 'train/grad_norm_pre_clip': 26.95585823059082, 'timing/data': 0.00104471854865551, 'timing/model': 1.0196983283385634, 'learning_rate/qwen_vl_interface': 5.000019693601204e-07, 'learning_rate/action_model': 5.000019693601204e-06, 'epoch': 2.56}) +[2026-06-18 08:21:46,876][starVLA.training.train_starvla][INFO] - Step 6999, Loss: {'train/loss': 0.29204756021499634, 'train/grad_norm_pre_clip': 29.109050750732422, 'timing/data': 0.0013328716158866882, 'timing/model': 1.1575434440746903, 'learning_rate/qwen_vl_interface': 5.000004923400556e-07, 'learning_rate/action_model': 5.0000049234005565e-06, 'epoch': 2.56}) +[2026-06-18 08:24:26,788][starVLA.training.train_starvla][INFO] - Step 7000, Loss: {'train/loss': 0.5004007816314697, 'train/grad_norm_pre_clip': 25.304582595825195, 'eval/loss': 0.5513215065002441, 'eval/latency_2/loss': 0.5513215065002441, 'eval/demon_fire_precision': 0.9538653366583542, 'eval/demon_fire_recall': 0.9602510460251046, 'eval/demon_fire_f1': 0.957047539616347, 'eval/demon_fire_rate_model': 0.669449081803005, 'eval/demon_fire_rate_teacher': 0.6649972175848636, 'eval/demon_left_precision': 0.9115426105717368, 'eval/demon_left_recall': 0.8894736842105263, 'eval/demon_left_f1': 0.9003729355354289, 'eval/demon_left_rate_model': 0.2579298831385643, 'eval/demon_left_rate_teacher': 0.2643294379521425, 'eval/demon_right_precision': 0.9134831460674158, 'eval/demon_right_recall': 0.9186440677966101, 'eval/demon_right_f1': 0.916056338028169, 'eval/demon_right_rate_model': 0.2476349471341124, 'eval/demon_right_rate_teacher': 0.24624373956594323, 'eval/demon_move_f1': 0.9082146367817989, 'eval/demon_cc_f1': 0.9418550354761545, 'eval/latency_2/demon_fire_precision': 0.9538653366583542, 'eval/latency_2/demon_fire_recall': 0.9602510460251046, 'eval/latency_2/demon_fire_f1': 0.957047539616347, 'eval/latency_2/demon_fire_rate_model': 0.669449081803005, 'eval/latency_2/demon_fire_rate_teacher': 0.6649972175848636, 'eval/latency_2/demon_left_precision': 0.9115426105717368, 'eval/latency_2/demon_left_recall': 0.8894736842105263, 'eval/latency_2/demon_left_f1': 0.9003729355354289, 'eval/latency_2/demon_left_rate_model': 0.2579298831385643, 'eval/latency_2/demon_left_rate_teacher': 0.2643294379521425, 'eval/latency_2/demon_right_precision': 0.9134831460674158, 'eval/latency_2/demon_right_recall': 0.9186440677966101, 'eval/latency_2/demon_right_f1': 0.916056338028169, 'eval/latency_2/demon_right_rate_model': 0.2476349471341124, 'eval/latency_2/demon_right_rate_teacher': 0.24624373956594323, 'eval/latency_2/demon_move_f1': 0.9082146367817989, 'eval/latency_2/demon_cc_f1': 0.9418550354761545, 'timing/data': 0.001116698607802391, 'timing/model': 1.075097749941051, 'learning_rate/qwen_vl_interface': 5e-07, 'learning_rate/action_model': 4.9999999999999996e-06, 'epoch': 2.56}) +[2026-06-18 08:24:26,807][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_7000_state +[2026-06-18 08:24:26,814][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 08:24:35,793][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_7000_state/model.safetensors +[2026-06-18 08:24:50,684][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_7000_state/optimizer.bin +[2026-06-18 08:24:50,685][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_7000_state/sampler.bin +[2026-06-18 08:24:50,686][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_7000_state/sampler_1.bin +[2026-06-18 08:24:50,689][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints/steps_7000_state/random_states_0.pkl +[2026-06-18 08:24:50,697][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 7000 +[2026-06-18 08:24:50,697][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 08:24:50,715][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 08:24:50,716][starVLA.training.train_starvla][INFO] - Training complete. Final training state is saved under /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_2_200ep/checkpoints +[2026-06-18 08:24:51,251][starVLA.training.train_starvla][INFO] - ... and that's all, folks! diff --git a/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_2_200ep/summary.jsonl b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_2_200ep/summary.jsonl new file mode 100644 index 0000000000000000000000000000000000000000..67f4020ac3871f475553f86f0c342379cb905502 --- /dev/null +++ b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_2_200ep/summary.jsonl @@ -0,0 +1,70 @@ +{"steps": 100} +{"steps": 200} +{"steps": 300} +{"steps": 400} +{"steps": 500} +{"steps": 600} +{"steps": 700} +{"steps": 800} +{"steps": 900} +{"steps": 1000} +{"steps": 1100} +{"steps": 1200} +{"steps": 1300} +{"steps": 1400} +{"steps": 1500} +{"steps": 1600} +{"steps": 1700} +{"steps": 1800} +{"steps": 1900} +{"steps": 2000} +{"steps": 2100} +{"steps": 2200} +{"steps": 2300} +{"steps": 2400} +{"steps": 2500} +{"steps": 2600} +{"steps": 2700} +{"steps": 2800} +{"steps": 2900} +{"steps": 3000} +{"steps": 3100} +{"steps": 3200} +{"steps": 3300} +{"steps": 3400} +{"steps": 3500} +{"steps": 3600} +{"steps": 3700} +{"steps": 3800} +{"steps": 3900} +{"steps": 4000} +{"steps": 4100} +{"steps": 4200} +{"steps": 4300} +{"steps": 4400} +{"steps": 4500} +{"steps": 4600} +{"steps": 4700} +{"steps": 4800} +{"steps": 4900} +{"steps": 5000} +{"steps": 5100} +{"steps": 5200} +{"steps": 5300} +{"steps": 5400} +{"steps": 5500} +{"steps": 5600} +{"steps": 5700} +{"steps": 5800} +{"steps": 5900} +{"steps": 6000} +{"steps": 6100} +{"steps": 6200} +{"steps": 6300} +{"steps": 6400} +{"steps": 6500} +{"steps": 6600} +{"steps": 6700} +{"steps": 6800} +{"steps": 6900} +{"steps": 7000} diff --git a/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_4_200ep/checkpoints/steps_7000_model.safetensors b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_4_200ep/checkpoints/steps_7000_model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..ffeaf26b9c97dffa7383631710e1601c743c3963 --- /dev/null +++ b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_4_200ep/checkpoints/steps_7000_model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:62e9b2a2b38f2b52cedc248c6e135b2593b212d1965fc3efae42e8130b57b373 +size 9138230516 diff --git a/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_4_200ep/config.full.yaml b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_4_200ep/config.full.yaml new file mode 100644 index 0000000000000000000000000000000000000000..da78ad59d695a03d7e3cb7db8e269e13afccbcba --- /dev/null +++ b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_4_200ep/config.full.yaml @@ -0,0 +1,209 @@ +framework: + name: QwenOFT + qwenvl: + base_vlm: /workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + attn_implementation: flash_attention_2 + enable_gradient_checkpointing: true + action_model: + action_model_type: MLP + action_dim: 7 + action_hidden_dim: 2560 + future_action_window_size: 0 + past_action_window_size: 0 + loss_type: discrete_ce + state_dim: 7 + action_horizon: 1 + action_env_dim: 6 +datasets: + vla_data: + dataset_py: lerobot_datasets + include_state: true + data_root_dir: /workspace/starVLA/data/demon_attack_fix_latency_4_200ep + data_mix: demon_attack_train__bridge + eval_data_mix: demon_attack_train__bridge__val + custom_mixtures_path: null + action_type: discrete + sequential_step_sampling: false + shuffle: true + action_balance: + enabled: false + strategy: balanced_epoch + action_key: action_id + target_flap_fraction: 0.3 + noop_id: 0 + flap_id: 1 + latency_curriculum: + enabled: false + strategy: exclusive + latencies: null + phase_steps: null + per_device_batch_size: 32 + load_all_data_for_training: true + obs_image_size: null + video_backend: torchvision_av +dataset: + source_hf: '' + config_name: null + source_subdir: null + converted_name: demon_attack_train + single_source_hf: '' + mixed_source_hf: '' + single_converted_name: demon_attack_train + mixed_converted_name: demon_attack_mixed_latency_train + single_latency_filter: null + mixed_latency_filter: null + force_download: false + setup_force: false + skip_verification: false + verify_rows: 200 + max_episodes: null + episodes_per_latency: null + latency_filter: null + debug_subset: + enabled: false + max_episodes: 5 + suffix: debug +base_model: + repo_id: Qwen/Qwen3-VL-4B-Instruct +initialization: + checkpoint_local_dir: playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_hf_repo_id: StarVLA/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_filename: checkpoints/steps_5000_pytorch_model.pt +trainer: + max_train_steps: 7000 + num_warmup_steps: 100 + save_interval: 100 + eval_interval: 100 + eval_num_batches: 100 + per_latency_eval_num_batches: null + eval_action_classification: true + cc_f1_tolerance: 1 + learning_rate: + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + action_model: 0.0001 + lr_scheduler_type: cosine_with_min_lr + scheduler_specific_kwargs: + min_lr: 1.0e-06 + freeze_modules: '' + loss_scale: + vla: 1.0 + vlm: 0.1 + max_grad_norm: 1.0 + weight_decay: 0.0 + logging_frequency: 1 + gradient_clipping: 1.0 + gradient_accumulation_steps: 4 + distributed_backend: none + is_resume: false + pretrained_checkpoint: null + resume_step: 0 + reload_modules: null + optimizer: + name: AdamW + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + weight_decay: 1.0e-08 + fused: false + save_format: pt +workspace_dir: WORKSPACE_DIR +run_root_dir: /workspace/starVLA/results/Checkpoints +seed: 42 +wandb_entity: dongqianyu99-zhejiang-university +wandb_project: starVLA_rl_games +auth: + env_file: null + hf_token_env: HF_TOKEN + wandb_api_key_env: WANDB_API_KEY +paths: + run_root_dir: results/Checkpoints + dataset_local_dir: playground/Datasets/rl_games + dataset_cache_dir: null + base_model_dir: playground/Pretrained_models/Qwen3-VL-4B-Instruct + accelerate_config: starVLA/config/deepseeds/deepspeed_zero2.yaml +launch: + use_accelerate: true + gpus: null + num_processes: 1 + dry_run: false +conda: + enabled: true + env_name: null +rl_games: + model_alias: openvla + env_eval: + image_size: 224 + frameskip: 4 + seed: 42 + fixed_episode_seeds: true + latency_seed_stride: 0 + task_seed_stride: 0 + task_description: 'You are playing Demon Attack from a single game image. Choose + exactly one action from: NOOP, FIRE, RIGHT, LEFT, RIGHTFIRE, LEFTFIRE.' + demon_attack: + noop_max: 30 + enabled: true + distributed_mode: none + vectorized: + enabled: false + batch_size: 1 + latency: + prompt_map_path: /workspace/starVLA/data/demon_attack_fix_latency_4_200ep/demon_attack_train__bridge/latency_prompt_map.json + mode: single + values: + - 0 + mid_train: + enabled: false + interval_steps: 100 + latencies: + - 0 + num_episodes: 5 + max_steps_per_episode: 3600 + post_train: + enabled: false + latencies: + - 0 + - 1 + - 2 + - 3 + - 4 + num_episodes: 5 + max_steps_per_episode: 3600 + task: demon_attack + initialization_mode: bridge + action_carrier: bridge +model: openvla +env: demon_attack +init: bridge +bridge_base_model: + repo_id: + openvla: Qwen/Qwen3-VL-4B-Instruct + pi0: StarVLA/Qwen2.5-VL-3B-Instruct-Action + pi05: Qwen/Qwen3-VL-4B-Instruct + gr00t: Qwen/Qwen3-VL-4B-Instruct + local_dir: + openvla: playground/Pretrained_models/Qwen3-VL-4B-Instruct + pi0: playground/Pretrained_models/Qwen2.5-VL-3B-Instruct-Action + pi05: playground/Pretrained_models/Qwen3-VL-4B-Instruct + gr00t: playground/Pretrained_models/Qwen3-VL-4B-Instruct +mode: single +checkpoint: + load: auto + hf_repo_id: null + save_best_model: false + save_pt_file: false + local: + keep_last_n: 1 + sync: + enabled: false + repo_id: null + keep_last_n: 0 + sync_every_n_checkpoints: 1 + resume_policy: local_latest +run_id: demon_attack_fix_latency_4_200ep +output_dir: /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_4_200ep +config_yaml: null +is_debug: false +version_id: '0.21' diff --git a/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_4_200ep/config.yaml b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_4_200ep/config.yaml new file mode 100644 index 0000000000000000000000000000000000000000..5e738dd9ea8fcee323575fc90b19b68b35d85d60 --- /dev/null +++ b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_4_200ep/config.yaml @@ -0,0 +1,75 @@ +checkpoint: + local: + keep_last_n: 1 + save_best_model: false + save_pt_file: false + sync: + enabled: false + keep_last_n: 0 + repo_id: null +datasets: + vla_data: + data_mix: demon_attack_train__bridge + dataset_py: lerobot_datasets + eval_data_mix: demon_attack_train__bridge__val + latency_curriculum: + enabled: false + obs_image_size: null + per_device_batch_size: 32 +framework: + action_model: + action_dim: 7 + action_env_dim: 6 + action_hidden_dim: 2560 + action_horizon: 1 + action_model_type: MLP + loss_type: discrete_ce + name: QwenOFT + qwenvl: + attn_implementation: flash_attention_2 + base_vlm: /workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + enable_gradient_checkpointing: true +output_dir: /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_2_200ep +rl_games: + env_eval: + enabled: true + mid_train: + enabled: false + interval_steps: 100 + task: demon_attack +run_id: demon_attack_fix_latency_4_200ep +run_root_dir: /workspace/starVLA/results/Checkpoints +seed: 42 +trainer: + cc_f1_tolerance: 1 + distributed_backend: none + eval_action_classification: true + eval_interval: 100 + eval_num_batches: 100 + freeze_modules: '' + gradient_accumulation_steps: 4 + gradient_clipping: 1.0 + is_resume: false + learning_rate: + action_model: 0.0001 + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + logging_frequency: 1 + lr_scheduler_type: cosine_with_min_lr + max_train_steps: 7000 + num_warmup_steps: 100 + optimizer: + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + fused: false + weight_decay: 1.0e-08 + per_latency_eval_num_batches: null + pretrained_checkpoint: null + reload_modules: null + save_interval: 100 + scheduler_specific_kwargs: + min_lr: 1.0e-06 +wandb_entity: dongqianyu99-zhejiang-university +wandb_project: starVLA_rl_games diff --git a/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_4_200ep/dataset_statistics.json b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_4_200ep/dataset_statistics.json new file mode 100644 index 0000000000000000000000000000000000000000..261675b64c8142ef8d22f5c3ce90024b0dcbd2e6 --- /dev/null +++ b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_4_200ep/dataset_statistics.json @@ -0,0 +1,127 @@ +{ + "new_embodiment": { + "action": { + "mean": [ + 0.02783012203872204, + 0.6656418442726135, + 0.008087566122412682, + 0.061184193938970566, + 0.1591275930404663, + 0.07812871038913727, + 0.0 + ], + "std": [ + 0.16479234397411346, + 0.47177085280418396, + 0.0895964726805687, + 0.23992209136486053, + 0.36527156829833984, + 0.26814261078834534, + 0.0 + ], + "max": [ + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 1.0, + 1.0, + 0.0, + 1.0, + 1.0, + 1.0, + 0.0 + ], + "mask": [ + true, + true, + true, + true, + true, + true, + true + ] + }, + "state": { + "mean": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ] + }, + "num_transitions": 332733, + "num_trajectories": 180 + } +} \ No newline at end of file diff --git a/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_4_200ep/dataset_statistics_eval.json b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_4_200ep/dataset_statistics_eval.json new file mode 100644 index 0000000000000000000000000000000000000000..3202b6b8cbd4fc14f79744b31e837f99e97f1cab --- /dev/null +++ b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_4_200ep/dataset_statistics_eval.json @@ -0,0 +1,127 @@ +{ + "new_embodiment": { + "action": { + "mean": [ + 0.028564849868416786, + 0.675146758556366, + 0.00799223780632019, + 0.06372412294149399, + 0.15450015664100647, + 0.07007186114788055, + 0.0 + ], + "std": [ + 0.1666077822446823, + 0.46833404898643494, + 0.08902717381715775, + 0.24433407187461853, + 0.36147764325141907, + 0.2552240490913391, + 0.0 + ], + "max": [ + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 1.0, + 1.0, + 0.0, + 1.0, + 1.0, + 1.0, + 0.0 + ], + "mask": [ + true, + true, + true, + true, + true, + true, + true + ] + }, + "state": { + "mean": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ] + }, + "num_transitions": 60809, + "num_trajectories": 20 + } +} \ No newline at end of file diff --git a/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_4_200ep/hydra/.hydra/config.yaml b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_4_200ep/hydra/.hydra/config.yaml new file mode 100644 index 0000000000000000000000000000000000000000..aaf43b1c47c7132422055ae8964cf0feefec72de --- /dev/null +++ b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_4_200ep/hydra/.hydra/config.yaml @@ -0,0 +1,207 @@ +framework: + qwenvl: + base_vlm: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/playground/Pretrained_models/Qwen3-VL-4B-Instruct + attn_implementation: flash_attention_2 + enable_gradient_checkpointing: true + action_model: + state_dim: 7 + loss_type: discrete_ce + action_horizon: 1 + future_action_window_size: 0 + past_action_window_size: 0 + action_dim: 7 + action_env_dim: 6 + name: QwenOFT +datasets: + vla_data: + dataset_py: lerobot_datasets + include_state: true + data_root_dir: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/data/demon_attack_fix_latency_4_200ep + data_mix: demon_attack_train__bridge + eval_data_mix: demon_attack_train__bridge__val + custom_mixtures_path: null + action_type: discrete + sequential_step_sampling: false + shuffle: true + action_balance: + enabled: false + strategy: balanced_epoch + action_key: action_id + target_flap_fraction: 0.3 + noop_id: 0 + flap_id: 1 + latency_curriculum: + enabled: false + strategy: exclusive + latencies: null + phase_steps: null + per_device_batch_size: 32 + load_all_data_for_training: true + obs_image_size: null + video_backend: torchvision_av +dataset: + source_hf: ${dataset.single_source_hf} + config_name: null + source_subdir: null + converted_name: ${dataset.single_converted_name} + single_source_hf: '' + mixed_source_hf: '' + single_converted_name: demon_attack_train + mixed_converted_name: demon_attack_mixed_latency_train + single_latency_filter: null + mixed_latency_filter: null + force_download: false + setup_force: false + skip_verification: false + verify_rows: 200 + max_episodes: null + episodes_per_latency: null + latency_filter: ${dataset.single_latency_filter} + debug_subset: + enabled: false + max_episodes: 5 + suffix: debug +base_model: + repo_id: ${bridge_base_model.repo_id.${model}} +initialization: + checkpoint_local_dir: playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_hf_repo_id: StarVLA/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_filename: checkpoints/steps_5000_pytorch_model.pt +trainer: + max_train_steps: 7000 + num_warmup_steps: 100 + save_interval: 100 + eval_interval: 100 + eval_num_batches: 100 + per_latency_eval_num_batches: null + eval_action_classification: true + cc_f1_tolerance: 1 + learning_rate: + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + action_model: 0.0001 + lr_scheduler_type: cosine_with_min_lr + scheduler_specific_kwargs: + min_lr: 1.0e-06 + freeze_modules: '' + loss_scale: + vla: 1.0 + vlm: 0.1 + max_grad_norm: 1.0 + weight_decay: 0.0 + logging_frequency: 1 + gradient_clipping: 1.0 + gradient_accumulation_steps: 4 + distributed_backend: none + is_resume: false + pretrained_checkpoint: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt + resume_step: 0 + reload_modules: null + optimizer: + name: AdamW + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + weight_decay: 1.0e-08 + fused: false + save_format: pt +workspace_dir: WORKSPACE_DIR +run_root_dir: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints +seed: 42 +wandb_entity: dongqianyu99-zhejiang-university +wandb_project: starVLA_rl_games +auth: + env_file: null + hf_token_env: HF_TOKEN + wandb_api_key_env: WANDB_API_KEY +paths: + run_root_dir: results/Checkpoints + dataset_local_dir: playground/Datasets/rl_games + dataset_cache_dir: null + base_model_dir: ${bridge_base_model.local_dir.${model}} + accelerate_config: starVLA/config/deepseeds/deepspeed_zero2.yaml +launch: + use_accelerate: true + gpus: null + num_processes: 1 + dry_run: false +conda: + enabled: true + env_name: null +rl_games: + model_alias: openvla + env_eval: + image_size: 224 + frameskip: 4 + seed: 42 + fixed_episode_seeds: true + latency_seed_stride: 0 + task_seed_stride: 0 + task_description: 'You are playing Demon Attack from a single game image. Choose + exactly one action from: NOOP, FIRE, RIGHT, LEFT, RIGHTFIRE, LEFTFIRE.' + demon_attack: + noop_max: 30 + enabled: true + distributed_mode: none + vectorized: + enabled: false + batch_size: 1 + latency: + prompt_map_path: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/data/demon_attack_fix_latency_4_200ep/demon_attack_train__bridge/latency_prompt_map.json + mode: single + values: + - 0 + mid_train: + enabled: false + interval_steps: 100 + latencies: + - 0 + num_episodes: 5 + max_steps_per_episode: 3600 + post_train: + enabled: false + latencies: + - 0 + - 1 + - 2 + - 3 + - 4 + num_episodes: 5 + max_steps_per_episode: 3600 + task: demon_attack + initialization_mode: bridge + action_carrier: bridge +model: openvla +env: demon_attack +init: bridge +bridge_base_model: + repo_id: + openvla: Qwen/Qwen3-VL-4B-Instruct + pi0: StarVLA/Qwen2.5-VL-3B-Instruct-Action + pi05: Qwen/Qwen3-VL-4B-Instruct + gr00t: Qwen/Qwen3-VL-4B-Instruct + local_dir: + openvla: playground/Pretrained_models/Qwen3-VL-4B-Instruct + pi0: playground/Pretrained_models/Qwen2.5-VL-3B-Instruct-Action + pi05: playground/Pretrained_models/Qwen3-VL-4B-Instruct + gr00t: playground/Pretrained_models/Qwen3-VL-4B-Instruct +mode: single +checkpoint: + load: auto + hf_repo_id: null + save_best_model: false + save_pt_file: false + local: + keep_last_n: 1 + sync: + enabled: false + repo_id: null + keep_last_n: 0 + sync_every_n_checkpoints: 1 + resume_policy: local_latest +run_id: demon_attack_fix_latency_4_200ep +output_dir: null +config_yaml: null +is_debug: false +version_id: '0.21' diff --git a/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_4_200ep/hydra/.hydra/hydra.yaml b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_4_200ep/hydra/.hydra/hydra.yaml new file mode 100644 index 0000000000000000000000000000000000000000..f1860f9e57938eb16ba16370299625c4bd3dbbad --- /dev/null +++ b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_4_200ep/hydra/.hydra/hydra.yaml @@ -0,0 +1,259 @@ +hydra: + run: + dir: ${run_root_dir}/${run_id}/hydra + sweep: + dir: multirun/${now:%Y-%m-%d}/${now:%H-%M-%S} + subdir: ${hydra.job.num} + launcher: + _target_: hydra._internal.core_plugins.basic_launcher.BasicLauncher + sweeper: + _target_: hydra._internal.core_plugins.basic_sweeper.BasicSweeper + max_batch_size: null + params: null + help: + app_name: ${hydra.job.name} + header: '${hydra.help.app_name} is powered by Hydra. + + ' + footer: 'Powered by Hydra (https://hydra.cc) + + Use --hydra-help to view Hydra specific help + + ' + template: '${hydra.help.header} + + == Configuration groups == + + Compose your configuration from those groups (group=option) + + + $APP_CONFIG_GROUPS + + + == Config == + + Override anything in the config (foo.bar=value) + + + $CONFIG + + + ${hydra.help.footer} + + ' + hydra_help: + template: 'Hydra (${hydra.runtime.version}) + + See https://hydra.cc for more info. + + + == Flags == + + $FLAGS_HELP + + + == Configuration groups == + + Compose your configuration from those groups (For example, append hydra/job_logging=disabled + to command line) + + + $HYDRA_CONFIG_GROUPS + + + Use ''--cfg hydra'' to Show the Hydra config. + + ' + hydra_help: ??? + hydra_logging: + version: 1 + formatters: + simple: + format: '[%(asctime)s][HYDRA] %(message)s' + handlers: + console: + class: logging.StreamHandler + formatter: simple + stream: ext://sys.stdout + root: + level: INFO + handlers: + - console + loggers: + logging_example: + level: DEBUG + disable_existing_loggers: false + job_logging: + version: 1 + formatters: + simple: + format: '[%(asctime)s][%(name)s][%(levelname)s] - %(message)s' + handlers: + console: + class: logging.StreamHandler + formatter: simple + stream: ext://sys.stdout + file: + class: logging.FileHandler + formatter: simple + filename: ${hydra.runtime.output_dir}/${hydra.job.name}.log + root: + level: INFO + handlers: + - console + - file + disable_existing_loggers: false + env: {} + mode: RUN + searchpath: [] + callbacks: {} + output_subdir: .hydra + overrides: + hydra: + - hydra.mode=RUN + task: + - model=openvla + - env=demon_attack + - init=bridge + - mode=single + - run_id=demon_attack_fix_latency_4_200ep + - run_root_dir=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints + - seed=42 + - wandb_entity=dongqianyu99-zhejiang-university + - wandb_project=starVLA_rl_games + - rl_games.env_eval.enabled=true + - checkpoint.sync.enabled=false + - checkpoint.sync.keep_last_n=0 + - checkpoint.local.keep_last_n=1 + - checkpoint.save_best_model=false + - checkpoint.save_pt_file=false + - trainer.is_resume=false + - trainer.pretrained_checkpoint=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt + - trainer.resume_step=0 + - trainer.max_train_steps=7000 + - trainer.num_warmup_steps=100 + - trainer.save_interval=100 + - trainer.eval_interval=100 + - trainer.eval_num_batches=100 + - trainer.logging_frequency=1 + - trainer.gradient_accumulation_steps=4 + - trainer.distributed_backend=none + - trainer.learning_rate.base=2e-05 + - trainer.learning_rate.qwen_vl_interface=1e-05 + - trainer.learning_rate.action_model=0.0001 + - trainer.lr_scheduler_type=cosine_with_min_lr + - trainer.scheduler_specific_kwargs.min_lr=1e-06 + - trainer.loss_scale.vla=1.0 + - trainer.loss_scale.vlm=0.1 + - trainer.max_grad_norm=1.0 + - trainer.weight_decay=0.0 + - trainer.gradient_clipping=1.0 + - trainer.optimizer.name=AdamW + - trainer.optimizer.betas=[0.9,0.95] + - trainer.optimizer.eps=1e-08 + - trainer.optimizer.weight_decay=1e-08 + - trainer.optimizer.fused=false + - trainer.save_format=pt + - framework.name=QwenOFT + - framework.qwenvl.attn_implementation=flash_attention_2 + - framework.qwenvl.enable_gradient_checkpointing=true + - framework.action_model.action_dim=7 + - framework.action_model.action_env_dim=6 + - framework.action_model.state_dim=7 + - framework.action_model.loss_type=discrete_ce + - framework.action_model.action_horizon=1 + - framework.action_model.future_action_window_size=0 + - framework.action_model.past_action_window_size=0 + - datasets.vla_data.include_state=true + - datasets.vla_data.action_type=discrete + - datasets.vla_data.sequential_step_sampling=false + - datasets.vla_data.shuffle=true + - datasets.vla_data.action_balance.enabled=false + - datasets.vla_data.action_balance.strategy=balanced_epoch + - datasets.vla_data.action_balance.action_key=action_id + - datasets.vla_data.action_balance.target_flap_fraction=0.3 + - datasets.vla_data.action_balance.noop_id=0 + - datasets.vla_data.action_balance.flap_id=1 + - datasets.vla_data.latency_curriculum.enabled=false + - datasets.vla_data.latency_curriculum.strategy=exclusive + - datasets.vla_data.per_device_batch_size=32 + - datasets.vla_data.load_all_data_for_training=true + - datasets.vla_data.video_backend=torchvision_av + - datasets.vla_data.data_root_dir=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/data/demon_attack_fix_latency_4_200ep + - datasets.vla_data.data_mix=demon_attack_train__bridge + - datasets.vla_data.eval_data_mix=demon_attack_train__bridge__val + - framework.qwenvl.base_vlm=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/playground/Pretrained_models/Qwen3-VL-4B-Instruct + - rl_games.task=demon_attack + - rl_games.model_alias=openvla + - rl_games.initialization_mode=bridge + - rl_games.action_carrier=bridge + - rl_games.env_eval.distributed_mode=none + - rl_games.env_eval.latency.mode=single + - rl_games.env_eval.frameskip=4 + - rl_games.env_eval.image_size=224 + - rl_games.env_eval.seed=42 + - rl_games.env_eval.fixed_episode_seeds=true + - rl_games.env_eval.latency_seed_stride=0 + - rl_games.env_eval.task_seed_stride=0 + - 'rl_games.env_eval.task_description=''You are playing Demon Attack from a single + game image. Choose exactly one action from: NOOP, FIRE, RIGHT, LEFT, RIGHTFIRE, + LEFTFIRE.''' + - rl_games.env_eval.latency.values=[0] + - rl_games.env_eval.latency.prompt_map_path=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/data/demon_attack_fix_latency_4_200ep/demon_attack_train__bridge/latency_prompt_map.json + - rl_games.env_eval.mid_train.enabled=false + - rl_games.env_eval.mid_train.interval_steps=100 + - rl_games.env_eval.mid_train.num_episodes=5 + - rl_games.env_eval.mid_train.max_steps_per_episode=3600 + - rl_games.env_eval.mid_train.latencies=[0] + - rl_games.env_eval.post_train.enabled=false + - rl_games.env_eval.post_train.num_episodes=5 + - rl_games.env_eval.post_train.max_steps_per_episode=3600 + - rl_games.env_eval.post_train.latencies=[0,1,2,3,4] + job: + name: train_starvla_hydra + chdir: false + override_dirname: 'checkpoint.local.keep_last_n=1,checkpoint.save_best_model=false,checkpoint.save_pt_file=false,checkpoint.sync.enabled=false,checkpoint.sync.keep_last_n=0,datasets.vla_data.action_balance.action_key=action_id,datasets.vla_data.action_balance.enabled=false,datasets.vla_data.action_balance.flap_id=1,datasets.vla_data.action_balance.noop_id=0,datasets.vla_data.action_balance.strategy=balanced_epoch,datasets.vla_data.action_balance.target_flap_fraction=0.3,datasets.vla_data.action_type=discrete,datasets.vla_data.data_mix=demon_attack_train__bridge,datasets.vla_data.data_root_dir=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/data/demon_attack_fix_latency_4_200ep,datasets.vla_data.eval_data_mix=demon_attack_train__bridge__val,datasets.vla_data.include_state=true,datasets.vla_data.latency_curriculum.enabled=false,datasets.vla_data.latency_curriculum.strategy=exclusive,datasets.vla_data.load_all_data_for_training=true,datasets.vla_data.per_device_batch_size=32,datasets.vla_data.sequential_step_sampling=false,datasets.vla_data.shuffle=true,datasets.vla_data.video_backend=torchvision_av,env=demon_attack,framework.action_model.action_dim=7,framework.action_model.action_env_dim=6,framework.action_model.action_horizon=1,framework.action_model.future_action_window_size=0,framework.action_model.loss_type=discrete_ce,framework.action_model.past_action_window_size=0,framework.action_model.state_dim=7,framework.name=QwenOFT,framework.qwenvl.attn_implementation=flash_attention_2,framework.qwenvl.base_vlm=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/playground/Pretrained_models/Qwen3-VL-4B-Instruct,framework.qwenvl.enable_gradient_checkpointing=true,init=bridge,mode=single,model=openvla,rl_games.action_carrier=bridge,rl_games.env_eval.distributed_mode=none,rl_games.env_eval.enabled=true,rl_games.env_eval.fixed_episode_seeds=true,rl_games.env_eval.frameskip=4,rl_games.env_eval.image_size=224,rl_games.env_eval.latency.mode=single,rl_games.env_eval.latency.prompt_map_path=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/data/demon_attack_fix_latency_4_200ep/demon_attack_train__bridge/latency_prompt_map.json,rl_games.env_eval.latency.values=[0],rl_games.env_eval.latency_seed_stride=0,rl_games.env_eval.mid_train.enabled=false,rl_games.env_eval.mid_train.interval_steps=100,rl_games.env_eval.mid_train.latencies=[0],rl_games.env_eval.mid_train.max_steps_per_episode=3600,rl_games.env_eval.mid_train.num_episodes=5,rl_games.env_eval.post_train.enabled=false,rl_games.env_eval.post_train.latencies=[0,1,2,3,4],rl_games.env_eval.post_train.max_steps_per_episode=3600,rl_games.env_eval.post_train.num_episodes=5,rl_games.env_eval.seed=42,rl_games.env_eval.task_description=''You + are playing Demon Attack from a single game image. Choose exactly one action + from: NOOP, FIRE, RIGHT, LEFT, RIGHTFIRE, LEFTFIRE.'',rl_games.env_eval.task_seed_stride=0,rl_games.initialization_mode=bridge,rl_games.model_alias=openvla,rl_games.task=demon_attack,run_id=demon_attack_fix_latency_4_200ep,run_root_dir=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints,seed=42,trainer.distributed_backend=none,trainer.eval_interval=100,trainer.eval_num_batches=100,trainer.gradient_accumulation_steps=4,trainer.gradient_clipping=1.0,trainer.is_resume=false,trainer.learning_rate.action_model=0.0001,trainer.learning_rate.base=2e-05,trainer.learning_rate.qwen_vl_interface=1e-05,trainer.logging_frequency=1,trainer.loss_scale.vla=1.0,trainer.loss_scale.vlm=0.1,trainer.lr_scheduler_type=cosine_with_min_lr,trainer.max_grad_norm=1.0,trainer.max_train_steps=7000,trainer.num_warmup_steps=100,trainer.optimizer.betas=[0.9,0.95],trainer.optimizer.eps=1e-08,trainer.optimizer.fused=false,trainer.optimizer.name=AdamW,trainer.optimizer.weight_decay=1e-08,trainer.pretrained_checkpoint=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt,trainer.resume_step=0,trainer.save_format=pt,trainer.save_interval=100,trainer.scheduler_specific_kwargs.min_lr=1e-06,trainer.weight_decay=0.0,wandb_entity=dongqianyu99-zhejiang-university,wandb_project=starVLA_rl_games' + id: ??? + num: ??? + config_name: train + env_set: {} + env_copy: [] + config: + override_dirname: + kv_sep: '=' + item_sep: ',' + exclude_keys: [] + runtime: + version: 1.3.2 + version_base: '1.1' + cwd: /inspire/hdd/project/spatialintelligence/public/lzj/starVLA + config_sources: + - path: hydra.conf + schema: pkg + provider: hydra + - path: /inspire/hdd/project/spatialintelligence/public/lzj/starVLA/examples/rl_games/config + schema: file + provider: main + - path: '' + schema: structured + provider: schema + output_dir: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/hydra + choices: + cross_task_setup: null + checkpoint: default + mode: single + init: bridge + env: demon_attack + model: openvla + hydra/env: default + hydra/callbacks: null + hydra/job_logging: default + hydra/hydra_logging: default + hydra/hydra_help: default + hydra/help: default + hydra/sweeper: basic + hydra/launcher: basic + hydra/output: default + verbose: false diff --git a/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_4_200ep/hydra/.hydra/overrides.yaml b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_4_200ep/hydra/.hydra/overrides.yaml new file mode 100644 index 0000000000000000000000000000000000000000..4a6e1a7694f93083b2f6a2838cf29440ba6d8e15 --- /dev/null +++ b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_4_200ep/hydra/.hydra/overrides.yaml @@ -0,0 +1,97 @@ +- model=openvla +- env=demon_attack +- init=bridge +- mode=single +- run_id=demon_attack_fix_latency_4_200ep +- run_root_dir=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints +- seed=42 +- wandb_entity=dongqianyu99-zhejiang-university +- wandb_project=starVLA_rl_games +- rl_games.env_eval.enabled=true +- checkpoint.sync.enabled=false +- checkpoint.sync.keep_last_n=0 +- checkpoint.local.keep_last_n=1 +- checkpoint.save_best_model=false +- checkpoint.save_pt_file=false +- trainer.is_resume=false +- trainer.pretrained_checkpoint=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt +- trainer.resume_step=0 +- trainer.max_train_steps=7000 +- trainer.num_warmup_steps=100 +- trainer.save_interval=100 +- trainer.eval_interval=100 +- trainer.eval_num_batches=100 +- trainer.logging_frequency=1 +- trainer.gradient_accumulation_steps=4 +- trainer.distributed_backend=none +- trainer.learning_rate.base=2e-05 +- trainer.learning_rate.qwen_vl_interface=1e-05 +- trainer.learning_rate.action_model=0.0001 +- trainer.lr_scheduler_type=cosine_with_min_lr +- trainer.scheduler_specific_kwargs.min_lr=1e-06 +- trainer.loss_scale.vla=1.0 +- trainer.loss_scale.vlm=0.1 +- trainer.max_grad_norm=1.0 +- trainer.weight_decay=0.0 +- trainer.gradient_clipping=1.0 +- trainer.optimizer.name=AdamW +- trainer.optimizer.betas=[0.9,0.95] +- trainer.optimizer.eps=1e-08 +- trainer.optimizer.weight_decay=1e-08 +- trainer.optimizer.fused=false +- trainer.save_format=pt +- framework.name=QwenOFT +- framework.qwenvl.attn_implementation=flash_attention_2 +- framework.qwenvl.enable_gradient_checkpointing=true +- framework.action_model.action_dim=7 +- framework.action_model.action_env_dim=6 +- framework.action_model.state_dim=7 +- framework.action_model.loss_type=discrete_ce +- framework.action_model.action_horizon=1 +- framework.action_model.future_action_window_size=0 +- framework.action_model.past_action_window_size=0 +- datasets.vla_data.include_state=true +- datasets.vla_data.action_type=discrete +- datasets.vla_data.sequential_step_sampling=false +- datasets.vla_data.shuffle=true +- datasets.vla_data.action_balance.enabled=false +- datasets.vla_data.action_balance.strategy=balanced_epoch +- datasets.vla_data.action_balance.action_key=action_id +- datasets.vla_data.action_balance.target_flap_fraction=0.3 +- datasets.vla_data.action_balance.noop_id=0 +- datasets.vla_data.action_balance.flap_id=1 +- datasets.vla_data.latency_curriculum.enabled=false +- datasets.vla_data.latency_curriculum.strategy=exclusive +- datasets.vla_data.per_device_batch_size=32 +- datasets.vla_data.load_all_data_for_training=true +- datasets.vla_data.video_backend=torchvision_av +- datasets.vla_data.data_root_dir=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/data/demon_attack_fix_latency_4_200ep +- datasets.vla_data.data_mix=demon_attack_train__bridge +- datasets.vla_data.eval_data_mix=demon_attack_train__bridge__val +- framework.qwenvl.base_vlm=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/playground/Pretrained_models/Qwen3-VL-4B-Instruct +- rl_games.task=demon_attack +- rl_games.model_alias=openvla +- rl_games.initialization_mode=bridge +- rl_games.action_carrier=bridge +- rl_games.env_eval.distributed_mode=none +- rl_games.env_eval.latency.mode=single +- rl_games.env_eval.frameskip=4 +- rl_games.env_eval.image_size=224 +- rl_games.env_eval.seed=42 +- rl_games.env_eval.fixed_episode_seeds=true +- rl_games.env_eval.latency_seed_stride=0 +- rl_games.env_eval.task_seed_stride=0 +- 'rl_games.env_eval.task_description=''You are playing Demon Attack from a single + game image. Choose exactly one action from: NOOP, FIRE, RIGHT, LEFT, RIGHTFIRE, + LEFTFIRE.''' +- rl_games.env_eval.latency.values=[0] +- rl_games.env_eval.latency.prompt_map_path=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/data/demon_attack_fix_latency_4_200ep/demon_attack_train__bridge/latency_prompt_map.json +- rl_games.env_eval.mid_train.enabled=false +- rl_games.env_eval.mid_train.interval_steps=100 +- rl_games.env_eval.mid_train.num_episodes=5 +- rl_games.env_eval.mid_train.max_steps_per_episode=3600 +- rl_games.env_eval.mid_train.latencies=[0] +- rl_games.env_eval.post_train.enabled=false +- rl_games.env_eval.post_train.num_episodes=5 +- rl_games.env_eval.post_train.max_steps_per_episode=3600 +- rl_games.env_eval.post_train.latencies=[0,1,2,3,4] diff --git a/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_4_200ep/hydra/train_starvla_hydra.log b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_4_200ep/hydra/train_starvla_hydra.log new file mode 100644 index 0000000000000000000000000000000000000000..f5bcd0bbcdf9d7d9a7347ac6df5bbe247a71ee35 --- /dev/null +++ b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_4_200ep/hydra/train_starvla_hydra.log @@ -0,0 +1,7752 @@ +[2026-06-17 16:18:20,893][starVLA.training.train_starvla][INFO] - VLA Training :: Warming Up +[2026-06-17 16:18:20,915][starVLA.training.train_starvla][INFO] - ✅ Configuration wrapped for access tracking +[2026-06-17 16:19:10,118][starVLA.training.train_starvla][INFO] - Preloaded model checkpoint before Accelerator/DeepSpeed initialization: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt +[2026-06-17 16:19:10,332][starVLA.training.train_starvla][INFO] - Creating VLA Dataset with Mixture `demon_attack_train__bridge` +[2026-06-17 16:19:20,520][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Behavior (data_config): ['BEHAVIOR_challenge'] +[2026-06-17 16:19:20,521][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from DOMINO: ['robotwin'] +[2026-06-17 16:19:20,521][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from DOMINO (data_config): [] +[2026-06-17 16:19:20,521][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from DOMINO (data_config): ['domino', 'domino_clean', 'domino_random', 'domino_cotrain'] +[2026-06-17 16:19:20,521][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Franka: ['custom_robot_config', 'demo_sim_franka_delta_joints', 'SO101'] +[2026-06-17 16:19:20,521][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Franka (data_config): [] +[2026-06-17 16:19:20,521][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Franka (data_config): ['custom_dataset', 'custom_dataset_2', 'demo_sim_pick_place', 'SO101_pick'] +[2026-06-17 16:19:20,521][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from LIBERO: ['libero_franka'] +[2026-06-17 16:19:20,521][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from LIBERO (data_config): [] +[2026-06-17 16:19:20,522][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from LIBERO (data_config): ['libero_all', 'libero_goal', 'multi_robot'] +[2026-06-17 16:19:20,522][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from RoboChallenge_table30v2: ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-17 16:19:20,522][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from RoboChallenge_table30v2 (data_config): ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-17 16:19:20,522][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from RoboChallenge_table30v2 (data_config): ['robochallenge_table30v2_shred_paper', 'robochallenge_table30v2_ur5_all', 'robochallenge_table30v2_arx5_all', 'robochallenge_table30v2_dosw1_all'] +[2026-06-17 16:19:20,522][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_365: ['panda_omron_robocasa365'] +[2026-06-17 16:19:20,522][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_365 (data_config): [] +[2026-06-17 16:19:20,522][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_365 (data_config): ['robocasa365_open_drawer_target_human', 'robocasa365_atomic_target_human_all', 'robocasa365_composite_target_human_all', 'robocasa365_target_human_all'] +[2026-06-17 16:19:20,523][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_tabletop: ['fourier_gr1_arms_waist'] +[2026-06-17 16:19:20,523][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_tabletop (data_config): [] +[2026-06-17 16:19:20,523][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_tabletop (data_config): ['fourier_gr1_unified_1000'] +[2026-06-17 16:19:20,550][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robotwin: ['robotwin', 'robotwin50', 'arx_x5'] +[2026-06-17 16:19:20,550][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robotwin (data_config): [] +[2026-06-17 16:19:20,550][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robotwin (data_config): ['robotwin_all', 'robotwin_all_50', 'robotwin', 'robotwin_task1', 'robotwin_task2', 'arx_x5'] +[2026-06-17 16:19:20,551][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from SimplerEnv: ['oxe_droid', 'oxe_bridge', 'oxe_rt1'] +[2026-06-17 16:19:20,551][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from SimplerEnv (data_config): [] +[2026-06-17 16:19:20,551][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from SimplerEnv (data_config): ['bridge', 'bridge_rt_1'] +[2026-06-17 16:19:20,551][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from VLA-Arena: ['vla_arena_franka'] +[2026-06-17 16:19:20,551][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from VLA-Arena (data_config): [] +[2026-06-17 16:19:20,551][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from VLA-Arena (data_config): ['vla_arena_L0_S', 'vla_arena_L0_M', 'vla_arena_L0_L'] +[2026-06-17 16:19:20,551][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from rl_games: ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-17 16:19:20,551][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from rl_games (data_config): ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-17 16:19:20,551][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from rl_games (data_config): ['flappy_train', 'flappy_train__bridge', 'flappy_mixed_latency_train', 'flappy_mixed_latency_train__bridge', 'demon_attack_train', 'demon_attack_train__bridge', 'demon_attack_mixed_latency_train', 'demon_attack_mixed_latency_train__bridge', 'deadly_corridor_train', 'deadly_corridor_train__bridge', 'deadly_corridor_mixed_latency_train', 'deadly_corridor_mixed_latency_train__bridge'] +[2026-06-17 16:19:20,690][starVLA.training.train_starvla][INFO] - Creating VLA Eval Dataset with Mixture `demon_attack_train__bridge__val` +[2026-06-17 16:19:20,742][starVLA.training.train_starvla][INFO] - 📝 Full config saved at /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/config.full.yaml +[2026-06-17 16:19:20,754][starVLA.training.train_starvla][INFO] - 📊 Accessed config snapshot saved at /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/config.yaml +[2026-06-17 16:19:20,755][starVLA.training.train_starvla][INFO] - No pretrained checkpoint provided. Starting training from scratch. +[2026-06-17 16:19:25,638][starVLA.training.train_starvla][INFO] - ***** Training Configuration ***** +[2026-06-17 16:19:25,638][starVLA.training.train_starvla][INFO] - Total optimization steps = 7000 +[2026-06-17 16:19:25,639][starVLA.training.train_starvla][INFO] - Per device batch size = 32 +[2026-06-17 16:19:25,639][starVLA.training.train_starvla][INFO] - Gradient accumulation steps = 4 +[2026-06-17 16:19:25,640][starVLA.training.train_starvla][INFO] - Total batch size = 128 +[2026-06-17 16:20:38,628][starVLA.training.train_starvla][INFO] - Step 1, Loss: {'train/loss': 1.7575832605361938, 'train/grad_norm_pre_clip': 8.648223876953125, 'timing/data': 0.04244851507246494, 'timing/model': 0.8770550554618239, 'learning_rate/qwen_vl_interface': 1.0000000000000001e-07, 'learning_rate/action_model': 1.0000000000000002e-06, 'epoch': 0.0}) +[2026-06-17 16:20:41,888][starVLA.training.train_starvla][INFO] - Step 2, Loss: {'train/loss': 1.771561861038208, 'train/grad_norm_pre_clip': 8.821575164794922, 'timing/data': 0.039215123280882835, 'timing/model': 0.8267916198819876, 'learning_rate/qwen_vl_interface': 2.0000000000000002e-07, 'learning_rate/action_model': 2.0000000000000003e-06, 'epoch': 0.0}) +[2026-06-17 16:20:45,085][starVLA.training.train_starvla][INFO] - Step 3, Loss: {'train/loss': 1.7610609531402588, 'train/grad_norm_pre_clip': 9.310937881469727, 'timing/data': 0.03401107061654329, 'timing/model': 0.8645133608952165, 'learning_rate/qwen_vl_interface': 3.0000000000000004e-07, 'learning_rate/action_model': 3e-06, 'epoch': 0.0}) +[2026-06-17 16:20:48,477][starVLA.training.train_starvla][INFO] - Step 4, Loss: {'train/loss': 1.75812828540802, 'train/grad_norm_pre_clip': 8.618371963500977, 'timing/data': 0.03875360544770956, 'timing/model': 0.9110704278573394, 'learning_rate/qwen_vl_interface': 4.0000000000000003e-07, 'learning_rate/action_model': 4.000000000000001e-06, 'epoch': 0.0}) +[2026-06-17 16:20:51,661][starVLA.training.train_starvla][INFO] - Step 5, Loss: {'train/loss': 1.706113576889038, 'train/grad_norm_pre_clip': 8.896660804748535, 'timing/data': 0.04194411542266607, 'timing/model': 0.8422575946897268, 'learning_rate/qwen_vl_interface': 5.000000000000001e-07, 'learning_rate/action_model': 5e-06, 'epoch': 0.0}) +[2026-06-17 16:20:55,383][starVLA.training.train_starvla][INFO] - Step 6, Loss: {'train/loss': 1.6612062454223633, 'train/grad_norm_pre_clip': 9.778078079223633, 'timing/data': 0.034458622336387634, 'timing/model': 1.1456358367577195, 'learning_rate/qwen_vl_interface': 6.000000000000001e-07, 'learning_rate/action_model': 6e-06, 'epoch': 0.0}) +[2026-06-17 16:21:03,121][starVLA.training.train_starvla][INFO] - Step 7, Loss: {'train/loss': 1.6342629194259644, 'train/grad_norm_pre_clip': 7.802255630493164, 'timing/data': 0.04513193015009165, 'timing/model': 0.8761656135320663, 'learning_rate/qwen_vl_interface': 7.000000000000001e-07, 'learning_rate/action_model': 7.000000000000001e-06, 'epoch': 0.0}) +[2026-06-17 16:21:06,524][starVLA.training.train_starvla][INFO] - Step 8, Loss: {'train/loss': 1.5601885318756104, 'train/grad_norm_pre_clip': 8.575005531311035, 'timing/data': 0.039686514995992184, 'timing/model': 0.8765090489760041, 'learning_rate/qwen_vl_interface': 8.000000000000001e-07, 'learning_rate/action_model': 8.000000000000001e-06, 'epoch': 0.0}) +[2026-06-17 16:21:09,754][starVLA.training.train_starvla][INFO] - Step 9, Loss: {'train/loss': 1.4695920944213867, 'train/grad_norm_pre_clip': 9.156634330749512, 'timing/data': 0.030523018911480904, 'timing/model': 0.8961183791980147, 'learning_rate/qwen_vl_interface': 9.000000000000001e-07, 'learning_rate/action_model': 9e-06, 'epoch': 0.0}) +[2026-06-17 16:21:12,958][starVLA.training.train_starvla][INFO] - Step 10, Loss: {'train/loss': 1.479962944984436, 'train/grad_norm_pre_clip': 6.175558090209961, 'timing/data': 0.033771343529224396, 'timing/model': 0.8153701759874821, 'learning_rate/qwen_vl_interface': 1.0000000000000002e-06, 'learning_rate/action_model': 1e-05, 'epoch': 0.0}) +[2026-06-17 16:21:16,179][starVLA.training.train_starvla][INFO] - Step 11, Loss: {'train/loss': 1.3757140636444092, 'train/grad_norm_pre_clip': 6.505069255828857, 'timing/data': 0.03545033931732178, 'timing/model': 0.8427243921905756, 'learning_rate/qwen_vl_interface': 1.1e-06, 'learning_rate/action_model': 1.1000000000000001e-05, 'epoch': 0.0}) +[2026-06-17 16:21:19,978][starVLA.training.train_starvla][INFO] - Step 12, Loss: {'train/loss': 1.4005690813064575, 'train/grad_norm_pre_clip': 4.463327884674072, 'timing/data': 0.0380361070856452, 'timing/model': 0.8198019973933697, 'learning_rate/qwen_vl_interface': 1.2000000000000002e-06, 'learning_rate/action_model': 1.2e-05, 'epoch': 0.0}) +[2026-06-17 16:21:23,238][starVLA.training.train_starvla][INFO] - Step 13, Loss: {'train/loss': 1.327957272529602, 'train/grad_norm_pre_clip': 4.0933685302734375, 'timing/data': 0.038499703630805016, 'timing/model': 0.8214955208823085, 'learning_rate/qwen_vl_interface': 1.3e-06, 'learning_rate/action_model': 1.3000000000000001e-05, 'epoch': 0.01}) +[2026-06-17 16:21:26,529][starVLA.training.train_starvla][INFO] - Step 14, Loss: {'train/loss': 1.244111180305481, 'train/grad_norm_pre_clip': 3.5165319442749023, 'timing/data': 0.03826550394296646, 'timing/model': 0.8489387044683099, 'learning_rate/qwen_vl_interface': 1.4000000000000001e-06, 'learning_rate/action_model': 1.4000000000000001e-05, 'epoch': 0.01}) +[2026-06-17 16:21:29,763][starVLA.training.train_starvla][INFO] - Step 15, Loss: {'train/loss': 1.5188205242156982, 'train/grad_norm_pre_clip': 3.4714291095733643, 'timing/data': 0.03383414354175329, 'timing/model': 0.895749126560986, 'learning_rate/qwen_vl_interface': 1.5e-06, 'learning_rate/action_model': 1.5e-05, 'epoch': 0.01}) +[2026-06-17 16:21:33,000][starVLA.training.train_starvla][INFO] - Step 16, Loss: {'train/loss': 1.0928707122802734, 'train/grad_norm_pre_clip': 3.0049142837524414, 'timing/data': 0.0341269550845027, 'timing/model': 0.832219397649169, 'learning_rate/qwen_vl_interface': 1.6000000000000001e-06, 'learning_rate/action_model': 1.6000000000000003e-05, 'epoch': 0.01}) +[2026-06-17 16:21:36,204][starVLA.training.train_starvla][INFO] - Step 17, Loss: {'train/loss': 0.9924074411392212, 'train/grad_norm_pre_clip': 2.168043851852417, 'timing/data': 0.03435525484383106, 'timing/model': 0.8213974228128791, 'learning_rate/qwen_vl_interface': 1.7000000000000002e-06, 'learning_rate/action_model': 1.7000000000000003e-05, 'epoch': 0.01}) +[2026-06-17 16:21:39,498][starVLA.training.train_starvla][INFO] - Step 18, Loss: {'train/loss': 1.0182018280029297, 'train/grad_norm_pre_clip': 1.5664985179901123, 'timing/data': 0.04227401968091726, 'timing/model': 0.8580375630408525, 'learning_rate/qwen_vl_interface': 1.8000000000000001e-06, 'learning_rate/action_model': 1.8e-05, 'epoch': 0.01}) +[2026-06-17 16:21:42,611][starVLA.training.train_starvla][INFO] - Step 19, Loss: {'train/loss': 1.2109824419021606, 'train/grad_norm_pre_clip': 2.609020471572876, 'timing/data': 0.03466993570327759, 'timing/model': 0.815283834002912, 'learning_rate/qwen_vl_interface': 1.9000000000000002e-06, 'learning_rate/action_model': 1.9e-05, 'epoch': 0.01}) +[2026-06-17 16:21:45,850][starVLA.training.train_starvla][INFO] - Step 20, Loss: {'train/loss': 1.103690266609192, 'train/grad_norm_pre_clip': 2.466728448867798, 'timing/data': 0.03749989718198776, 'timing/model': 0.8514417270198464, 'learning_rate/qwen_vl_interface': 2.0000000000000003e-06, 'learning_rate/action_model': 2e-05, 'epoch': 0.01}) +[2026-06-17 16:21:49,076][starVLA.training.train_starvla][INFO] - Step 21, Loss: {'train/loss': 1.5918042659759521, 'train/grad_norm_pre_clip': 7.818161487579346, 'timing/data': 0.049408067017793655, 'timing/model': 0.8356786528602242, 'learning_rate/qwen_vl_interface': 2.1000000000000002e-06, 'learning_rate/action_model': 2.1e-05, 'epoch': 0.01}) +[2026-06-17 16:21:52,419][starVLA.training.train_starvla][INFO] - Step 22, Loss: {'train/loss': 1.1230030059814453, 'train/grad_norm_pre_clip': 2.24954891204834, 'timing/data': 0.035665931180119514, 'timing/model': 0.8762732148170471, 'learning_rate/qwen_vl_interface': 2.2e-06, 'learning_rate/action_model': 2.2000000000000003e-05, 'epoch': 0.01}) +[2026-06-17 16:22:00,783][starVLA.training.train_starvla][INFO] - Step 23, Loss: {'train/loss': 1.5068449974060059, 'train/grad_norm_pre_clip': 7.04041862487793, 'timing/data': 0.03973513003438711, 'timing/model': 2.211630191653967, 'learning_rate/qwen_vl_interface': 2.3000000000000004e-06, 'learning_rate/action_model': 2.3000000000000003e-05, 'epoch': 0.01}) +[2026-06-17 16:22:06,525][starVLA.training.train_starvla][INFO] - Step 24, Loss: {'train/loss': 1.4075632095336914, 'train/grad_norm_pre_clip': 4.347489833831787, 'timing/data': 0.08090422954410315, 'timing/model': 0.8104102406650782, 'learning_rate/qwen_vl_interface': 2.4000000000000003e-06, 'learning_rate/action_model': 2.4e-05, 'epoch': 0.01}) +[2026-06-17 16:22:09,734][starVLA.training.train_starvla][INFO] - Step 25, Loss: {'train/loss': 1.1584676504135132, 'train/grad_norm_pre_clip': 2.424802541732788, 'timing/data': 0.03462090902030468, 'timing/model': 0.8214995767921209, 'learning_rate/qwen_vl_interface': 2.5e-06, 'learning_rate/action_model': 2.5e-05, 'epoch': 0.01}) +[2026-06-17 16:22:12,899][starVLA.training.train_starvla][INFO] - Step 26, Loss: {'train/loss': 1.005659818649292, 'train/grad_norm_pre_clip': 2.1122124195098877, 'timing/data': 0.03685944154858589, 'timing/model': 0.8032708885148168, 'learning_rate/qwen_vl_interface': 2.6e-06, 'learning_rate/action_model': 2.6000000000000002e-05, 'epoch': 0.01}) +[2026-06-17 16:22:16,624][starVLA.training.train_starvla][INFO] - Step 27, Loss: {'train/loss': 1.3368737697601318, 'train/grad_norm_pre_clip': 3.8549904823303223, 'timing/data': 0.03366440534591675, 'timing/model': 0.9234375320374966, 'learning_rate/qwen_vl_interface': 2.7000000000000004e-06, 'learning_rate/action_model': 2.7000000000000002e-05, 'epoch': 0.01}) +[2026-06-17 16:22:19,871][starVLA.training.train_starvla][INFO] - Step 28, Loss: {'train/loss': 1.074057936668396, 'train/grad_norm_pre_clip': 3.3700952529907227, 'timing/data': 0.033829606138169765, 'timing/model': 0.8436201214790344, 'learning_rate/qwen_vl_interface': 2.8000000000000003e-06, 'learning_rate/action_model': 2.8000000000000003e-05, 'epoch': 0.01}) +[2026-06-17 16:22:25,084][starVLA.training.train_starvla][INFO] - Step 29, Loss: {'train/loss': 1.2635595798492432, 'train/grad_norm_pre_clip': 2.9230802059173584, 'timing/data': 0.03571972995996475, 'timing/model': 2.3677304228767753, 'learning_rate/qwen_vl_interface': 2.9e-06, 'learning_rate/action_model': 2.9e-05, 'epoch': 0.01}) +[2026-06-17 16:22:30,384][starVLA.training.train_starvla][INFO] - Step 30, Loss: {'train/loss': 1.2900046110153198, 'train/grad_norm_pre_clip': 2.4543464183807373, 'timing/data': 0.043783060275018215, 'timing/model': 0.9481671825051308, 'learning_rate/qwen_vl_interface': 3e-06, 'learning_rate/action_model': 3e-05, 'epoch': 0.01}) +[2026-06-17 16:22:33,673][starVLA.training.train_starvla][INFO] - Step 31, Loss: {'train/loss': 0.9274886846542358, 'train/grad_norm_pre_clip': 3.769167184829712, 'timing/data': 0.038389758206903934, 'timing/model': 0.8519249893724918, 'learning_rate/qwen_vl_interface': 3.1000000000000004e-06, 'learning_rate/action_model': 3.1e-05, 'epoch': 0.01}) +[2026-06-17 16:22:36,938][starVLA.training.train_starvla][INFO] - Step 32, Loss: {'train/loss': 1.1819627285003662, 'train/grad_norm_pre_clip': 3.7530198097229004, 'timing/data': 0.03160673566162586, 'timing/model': 0.815059500746429, 'learning_rate/qwen_vl_interface': 3.2000000000000003e-06, 'learning_rate/action_model': 3.2000000000000005e-05, 'epoch': 0.01}) +[2026-06-17 16:22:40,264][starVLA.training.train_starvla][INFO] - Step 33, Loss: {'train/loss': 1.1837750673294067, 'train/grad_norm_pre_clip': 1.9018782377243042, 'timing/data': 0.035481647588312626, 'timing/model': 0.8639767458662391, 'learning_rate/qwen_vl_interface': 3.3000000000000006e-06, 'learning_rate/action_model': 3.3e-05, 'epoch': 0.01}) +[2026-06-17 16:22:44,897][starVLA.training.train_starvla][INFO] - Step 34, Loss: {'train/loss': 1.2030794620513916, 'train/grad_norm_pre_clip': 1.765981912612915, 'timing/data': 0.03522441070526838, 'timing/model': 0.9082671226933599, 'learning_rate/qwen_vl_interface': 3.4000000000000005e-06, 'learning_rate/action_model': 3.4000000000000007e-05, 'epoch': 0.01}) +[2026-06-17 16:22:48,183][starVLA.training.train_starvla][INFO] - Step 35, Loss: {'train/loss': 0.8953263759613037, 'train/grad_norm_pre_clip': 2.4545178413391113, 'timing/data': 0.05287067871540785, 'timing/model': 0.847823609597981, 'learning_rate/qwen_vl_interface': 3.5e-06, 'learning_rate/action_model': 3.5e-05, 'epoch': 0.01}) +[2026-06-17 16:22:51,355][starVLA.training.train_starvla][INFO] - Step 36, Loss: {'train/loss': 1.4005600214004517, 'train/grad_norm_pre_clip': 3.9341847896575928, 'timing/data': 0.03218868840485811, 'timing/model': 0.859401686117053, 'learning_rate/qwen_vl_interface': 3.6000000000000003e-06, 'learning_rate/action_model': 3.6e-05, 'epoch': 0.01}) +[2026-06-17 16:22:54,527][starVLA.training.train_starvla][INFO] - Step 37, Loss: {'train/loss': 0.9309097528457642, 'train/grad_norm_pre_clip': 1.8279547691345215, 'timing/data': 0.038362630642950535, 'timing/model': 0.8194469045847654, 'learning_rate/qwen_vl_interface': 3.7e-06, 'learning_rate/action_model': 3.7e-05, 'epoch': 0.01}) +[2026-06-17 16:22:57,851][starVLA.training.train_starvla][INFO] - Step 38, Loss: {'train/loss': 1.0877821445465088, 'train/grad_norm_pre_clip': 2.846165180206299, 'timing/data': 0.034870970994234085, 'timing/model': 0.8469114871695638, 'learning_rate/qwen_vl_interface': 3.8000000000000005e-06, 'learning_rate/action_model': 3.8e-05, 'epoch': 0.01}) +[2026-06-17 16:23:01,187][starVLA.training.train_starvla][INFO] - Step 39, Loss: {'train/loss': 1.2308294773101807, 'train/grad_norm_pre_clip': 3.791943073272705, 'timing/data': 0.03856768552213907, 'timing/model': 0.8795984163880348, 'learning_rate/qwen_vl_interface': 3.900000000000001e-06, 'learning_rate/action_model': 3.9000000000000006e-05, 'epoch': 0.02}) +[2026-06-17 16:23:04,594][starVLA.training.train_starvla][INFO] - Step 40, Loss: {'train/loss': 1.2095574140548706, 'train/grad_norm_pre_clip': 2.5670909881591797, 'timing/data': 0.03259336855262518, 'timing/model': 0.8975831298157573, 'learning_rate/qwen_vl_interface': 4.000000000000001e-06, 'learning_rate/action_model': 4e-05, 'epoch': 0.02}) +[2026-06-17 16:23:10,681][starVLA.training.train_starvla][INFO] - Step 41, Loss: {'train/loss': 1.0414401292800903, 'train/grad_norm_pre_clip': 1.414860486984253, 'timing/data': 0.025518957525491714, 'timing/model': 2.567931923083961, 'learning_rate/qwen_vl_interface': 4.1e-06, 'learning_rate/action_model': 4.1e-05, 'epoch': 0.02}) +[2026-06-17 16:23:13,979][starVLA.training.train_starvla][INFO] - Step 42, Loss: {'train/loss': 0.8782193660736084, 'train/grad_norm_pre_clip': 1.902760624885559, 'timing/data': 0.02854766510426998, 'timing/model': 0.8463669680058956, 'learning_rate/qwen_vl_interface': 4.2000000000000004e-06, 'learning_rate/action_model': 4.2e-05, 'epoch': 0.02}) +[2026-06-17 16:23:17,208][starVLA.training.train_starvla][INFO] - Step 43, Loss: {'train/loss': 1.571236491203308, 'train/grad_norm_pre_clip': 4.749803066253662, 'timing/data': 0.0263296440243721, 'timing/model': 0.8542622616514564, 'learning_rate/qwen_vl_interface': 4.3e-06, 'learning_rate/action_model': 4.3e-05, 'epoch': 0.02}) +[2026-06-17 16:23:20,450][starVLA.training.train_starvla][INFO] - Step 44, Loss: {'train/loss': 1.01870596408844, 'train/grad_norm_pre_clip': 2.0373775959014893, 'timing/data': 0.027148072607815266, 'timing/model': 0.8088739486411214, 'learning_rate/qwen_vl_interface': 4.4e-06, 'learning_rate/action_model': 4.4000000000000006e-05, 'epoch': 0.02}) +[2026-06-17 16:23:23,716][starVLA.training.train_starvla][INFO] - Step 45, Loss: {'train/loss': 0.9371678233146667, 'train/grad_norm_pre_clip': 2.0073273181915283, 'timing/data': 0.024247518740594387, 'timing/model': 0.8410047637298703, 'learning_rate/qwen_vl_interface': 4.5e-06, 'learning_rate/action_model': 4.5e-05, 'epoch': 0.02}) +[2026-06-17 16:23:26,925][starVLA.training.train_starvla][INFO] - Step 46, Loss: {'train/loss': 0.7530011534690857, 'train/grad_norm_pre_clip': 3.7817981243133545, 'timing/data': 0.025115245021879673, 'timing/model': 0.8018983136862516, 'learning_rate/qwen_vl_interface': 4.600000000000001e-06, 'learning_rate/action_model': 4.600000000000001e-05, 'epoch': 0.02}) +[2026-06-17 16:23:30,103][starVLA.training.train_starvla][INFO] - Step 47, Loss: {'train/loss': 0.9308484196662903, 'train/grad_norm_pre_clip': 2.088493824005127, 'timing/data': 0.026101041585206985, 'timing/model': 0.8279818370938301, 'learning_rate/qwen_vl_interface': 4.7e-06, 'learning_rate/action_model': 4.7e-05, 'epoch': 0.02}) +[2026-06-17 16:23:33,380][starVLA.training.train_starvla][INFO] - Step 48, Loss: {'train/loss': 1.0021648406982422, 'train/grad_norm_pre_clip': 2.0198071002960205, 'timing/data': 0.030651040375232697, 'timing/model': 0.834973780438304, 'learning_rate/qwen_vl_interface': 4.800000000000001e-06, 'learning_rate/action_model': 4.8e-05, 'epoch': 0.02}) +[2026-06-17 16:23:36,567][starVLA.training.train_starvla][INFO] - Step 49, Loss: {'train/loss': 0.8373925089836121, 'train/grad_norm_pre_clip': 2.538214921951294, 'timing/data': 0.02514268085360527, 'timing/model': 0.8248742148280144, 'learning_rate/qwen_vl_interface': 4.9000000000000005e-06, 'learning_rate/action_model': 4.9e-05, 'epoch': 0.02}) +[2026-06-17 16:23:39,746][starVLA.training.train_starvla][INFO] - Step 50, Loss: {'train/loss': 1.145048975944519, 'train/grad_norm_pre_clip': 2.3201379776000977, 'timing/data': 0.024463161826133728, 'timing/model': 0.817630928941071, 'learning_rate/qwen_vl_interface': 5e-06, 'learning_rate/action_model': 5e-05, 'epoch': 0.02}) +[2026-06-17 16:23:42,866][starVLA.training.train_starvla][INFO] - Step 51, Loss: {'train/loss': 0.8340986371040344, 'train/grad_norm_pre_clip': 2.1128127574920654, 'timing/data': 0.023366852663457394, 'timing/model': 0.8063266342505813, 'learning_rate/qwen_vl_interface': 5.1e-06, 'learning_rate/action_model': 5.1000000000000006e-05, 'epoch': 0.02}) +[2026-06-17 16:23:46,148][starVLA.training.train_starvla][INFO] - Step 52, Loss: {'train/loss': 1.077864646911621, 'train/grad_norm_pre_clip': 3.145813465118408, 'timing/data': 0.029661783948540688, 'timing/model': 0.8904298190027475, 'learning_rate/qwen_vl_interface': 5.2e-06, 'learning_rate/action_model': 5.2000000000000004e-05, 'epoch': 0.02}) +[2026-06-17 16:23:49,358][starVLA.training.train_starvla][INFO] - Step 53, Loss: {'train/loss': 1.1601028442382812, 'train/grad_norm_pre_clip': 3.1160337924957275, 'timing/data': 0.023368324153125286, 'timing/model': 0.854515134356916, 'learning_rate/qwen_vl_interface': 5.300000000000001e-06, 'learning_rate/action_model': 5.300000000000001e-05, 'epoch': 0.02}) +[2026-06-17 16:23:52,860][starVLA.training.train_starvla][INFO] - Step 54, Loss: {'train/loss': 1.4609466791152954, 'train/grad_norm_pre_clip': 5.608553409576416, 'timing/data': 0.02651333622634411, 'timing/model': 0.9801161894574761, 'learning_rate/qwen_vl_interface': 5.400000000000001e-06, 'learning_rate/action_model': 5.4000000000000005e-05, 'epoch': 0.02}) +[2026-06-17 16:23:56,464][starVLA.training.train_starvla][INFO] - Step 55, Loss: {'train/loss': 1.186181902885437, 'train/grad_norm_pre_clip': 3.32392954826355, 'timing/data': 0.024268467910587788, 'timing/model': 0.8057179898023605, 'learning_rate/qwen_vl_interface': 5.500000000000001e-06, 'learning_rate/action_model': 5.500000000000001e-05, 'epoch': 0.02}) +[2026-06-17 16:23:59,628][starVLA.training.train_starvla][INFO] - Step 56, Loss: {'train/loss': 1.0649737119674683, 'train/grad_norm_pre_clip': 4.064742565155029, 'timing/data': 0.022799765691161156, 'timing/model': 0.8241953179240227, 'learning_rate/qwen_vl_interface': 5.600000000000001e-06, 'learning_rate/action_model': 5.6000000000000006e-05, 'epoch': 0.02}) +[2026-06-17 16:24:02,837][starVLA.training.train_starvla][INFO] - Step 57, Loss: {'train/loss': 1.0308384895324707, 'train/grad_norm_pre_clip': 2.449374198913574, 'timing/data': 0.03169647790491581, 'timing/model': 0.8270085193216801, 'learning_rate/qwen_vl_interface': 5.7e-06, 'learning_rate/action_model': 5.6999999999999996e-05, 'epoch': 0.02}) +[2026-06-17 16:24:06,080][starVLA.training.train_starvla][INFO] - Step 58, Loss: {'train/loss': 0.9886202812194824, 'train/grad_norm_pre_clip': 2.901650905609131, 'timing/data': 0.023093611001968384, 'timing/model': 0.8223625961691141, 'learning_rate/qwen_vl_interface': 5.8e-06, 'learning_rate/action_model': 5.8e-05, 'epoch': 0.02}) +[2026-06-17 16:24:09,533][starVLA.training.train_starvla][INFO] - Step 59, Loss: {'train/loss': 1.3267370462417603, 'train/grad_norm_pre_clip': 4.39871883392334, 'timing/data': 0.024556096643209457, 'timing/model': 1.0350495362654328, 'learning_rate/qwen_vl_interface': 5.9e-06, 'learning_rate/action_model': 5.9e-05, 'epoch': 0.02}) +[2026-06-17 16:24:12,783][starVLA.training.train_starvla][INFO] - Step 60, Loss: {'train/loss': 0.9803406000137329, 'train/grad_norm_pre_clip': 3.9713761806488037, 'timing/data': 0.032618009485304356, 'timing/model': 0.8250365257263184, 'learning_rate/qwen_vl_interface': 6e-06, 'learning_rate/action_model': 6e-05, 'epoch': 0.02}) +[2026-06-17 16:24:16,026][starVLA.training.train_starvla][INFO] - Step 61, Loss: {'train/loss': 1.214407205581665, 'train/grad_norm_pre_clip': 4.62202262878418, 'timing/data': 0.02333261538296938, 'timing/model': 0.8050186540931463, 'learning_rate/qwen_vl_interface': 6.1e-06, 'learning_rate/action_model': 6.1e-05, 'epoch': 0.02}) +[2026-06-17 16:24:19,214][starVLA.training.train_starvla][INFO] - Step 62, Loss: {'train/loss': 1.069327712059021, 'train/grad_norm_pre_clip': 3.354534864425659, 'timing/data': 0.02843478601425886, 'timing/model': 0.8133870447054505, 'learning_rate/qwen_vl_interface': 6.200000000000001e-06, 'learning_rate/action_model': 6.2e-05, 'epoch': 0.02}) +[2026-06-17 16:24:22,368][starVLA.training.train_starvla][INFO] - Step 63, Loss: {'train/loss': 0.7857930660247803, 'train/grad_norm_pre_clip': 4.857115745544434, 'timing/data': 0.024548781104385853, 'timing/model': 0.8409353876486421, 'learning_rate/qwen_vl_interface': 6.300000000000001e-06, 'learning_rate/action_model': 6.3e-05, 'epoch': 0.02}) +[2026-06-17 16:24:25,573][starVLA.training.train_starvla][INFO] - Step 64, Loss: {'train/loss': 0.9646056890487671, 'train/grad_norm_pre_clip': 2.008730411529541, 'timing/data': 0.027873589657247066, 'timing/model': 0.8104878403246403, 'learning_rate/qwen_vl_interface': 6.4000000000000006e-06, 'learning_rate/action_model': 6.400000000000001e-05, 'epoch': 0.02}) +[2026-06-17 16:24:28,878][starVLA.training.train_starvla][INFO] - Step 65, Loss: {'train/loss': 1.0543429851531982, 'train/grad_norm_pre_clip': 1.930863380432129, 'timing/data': 0.023900938220322132, 'timing/model': 0.8860622858628631, 'learning_rate/qwen_vl_interface': 6.5000000000000004e-06, 'learning_rate/action_model': 6.500000000000001e-05, 'epoch': 0.03}) +[2026-06-17 16:24:32,137][starVLA.training.train_starvla][INFO] - Step 66, Loss: {'train/loss': 0.7435235977172852, 'train/grad_norm_pre_clip': 0.9751908779144287, 'timing/data': 0.022631983272731304, 'timing/model': 0.8058291291818023, 'learning_rate/qwen_vl_interface': 6.600000000000001e-06, 'learning_rate/action_model': 6.6e-05, 'epoch': 0.03}) +[2026-06-17 16:24:35,300][starVLA.training.train_starvla][INFO] - Step 67, Loss: {'train/loss': 0.9029316306114197, 'train/grad_norm_pre_clip': 2.4027955532073975, 'timing/data': 0.02513911109417677, 'timing/model': 0.8249770076945424, 'learning_rate/qwen_vl_interface': 6.700000000000001e-06, 'learning_rate/action_model': 6.7e-05, 'epoch': 0.03}) +[2026-06-17 16:24:38,494][starVLA.training.train_starvla][INFO] - Step 68, Loss: {'train/loss': 1.1813466548919678, 'train/grad_norm_pre_clip': 4.040131568908691, 'timing/data': 0.028013337403535843, 'timing/model': 0.8475811462849379, 'learning_rate/qwen_vl_interface': 6.800000000000001e-06, 'learning_rate/action_model': 6.800000000000001e-05, 'epoch': 0.03}) +[2026-06-17 16:24:41,622][starVLA.training.train_starvla][INFO] - Step 69, Loss: {'train/loss': 1.1059659719467163, 'train/grad_norm_pre_clip': 3.4509763717651367, 'timing/data': 0.023347308859229088, 'timing/model': 0.8015133673325181, 'learning_rate/qwen_vl_interface': 6.9e-06, 'learning_rate/action_model': 6.9e-05, 'epoch': 0.03}) +[2026-06-17 16:24:44,820][starVLA.training.train_starvla][INFO] - Step 70, Loss: {'train/loss': 1.0750327110290527, 'train/grad_norm_pre_clip': 1.6414942741394043, 'timing/data': 0.030111762695014477, 'timing/model': 0.8083723466843367, 'learning_rate/qwen_vl_interface': 7e-06, 'learning_rate/action_model': 7e-05, 'epoch': 0.03}) +[2026-06-17 16:24:47,960][starVLA.training.train_starvla][INFO] - Step 71, Loss: {'train/loss': 1.1667667627334595, 'train/grad_norm_pre_clip': 2.53330659866333, 'timing/data': 0.022781281732022762, 'timing/model': 0.7979774456471205, 'learning_rate/qwen_vl_interface': 7.100000000000001e-06, 'learning_rate/action_model': 7.1e-05, 'epoch': 0.03}) +[2026-06-17 16:24:51,150][starVLA.training.train_starvla][INFO] - Step 72, Loss: {'train/loss': 1.3080544471740723, 'train/grad_norm_pre_clip': 3.00270938873291, 'timing/data': 0.03288899175822735, 'timing/model': 0.8079084437340498, 'learning_rate/qwen_vl_interface': 7.2000000000000005e-06, 'learning_rate/action_model': 7.2e-05, 'epoch': 0.03}) +[2026-06-17 16:24:55,470][starVLA.training.train_starvla][INFO] - Step 73, Loss: {'train/loss': 1.3518249988555908, 'train/grad_norm_pre_clip': 2.9531543254852295, 'timing/data': 0.023259573616087437, 'timing/model': 2.0059261219576, 'learning_rate/qwen_vl_interface': 7.3e-06, 'learning_rate/action_model': 7.3e-05, 'epoch': 0.03}) +[2026-06-17 16:24:58,799][starVLA.training.train_starvla][INFO] - Step 74, Loss: {'train/loss': 1.2992403507232666, 'train/grad_norm_pre_clip': 4.768779754638672, 'timing/data': 0.028722818940877914, 'timing/model': 0.8583306362852454, 'learning_rate/qwen_vl_interface': 7.4e-06, 'learning_rate/action_model': 7.4e-05, 'epoch': 0.03}) +[2026-06-17 16:25:02,283][starVLA.training.train_starvla][INFO] - Step 75, Loss: {'train/loss': 0.9432625770568848, 'train/grad_norm_pre_clip': 4.418501853942871, 'timing/data': 0.031017609871923923, 'timing/model': 0.8474089046940207, 'learning_rate/qwen_vl_interface': 7.500000000000001e-06, 'learning_rate/action_model': 7.500000000000001e-05, 'epoch': 0.03}) +[2026-06-17 16:25:05,453][starVLA.training.train_starvla][INFO] - Step 76, Loss: {'train/loss': 1.124688744544983, 'train/grad_norm_pre_clip': 1.7559548616409302, 'timing/data': 0.022516475059092045, 'timing/model': 0.8385624578222632, 'learning_rate/qwen_vl_interface': 7.600000000000001e-06, 'learning_rate/action_model': 7.6e-05, 'epoch': 0.03}) +[2026-06-17 16:25:09,160][starVLA.training.train_starvla][INFO] - Step 77, Loss: {'train/loss': 1.1073194742202759, 'train/grad_norm_pre_clip': 3.5053322315216064, 'timing/data': 0.023282391019165516, 'timing/model': 0.8236825736239552, 'learning_rate/qwen_vl_interface': 7.7e-06, 'learning_rate/action_model': 7.7e-05, 'epoch': 0.03}) +[2026-06-17 16:25:12,794][starVLA.training.train_starvla][INFO] - Step 78, Loss: {'train/loss': 0.8947027921676636, 'train/grad_norm_pre_clip': 2.08463454246521, 'timing/data': 0.02295618411153555, 'timing/model': 0.8633439494296908, 'learning_rate/qwen_vl_interface': 7.800000000000002e-06, 'learning_rate/action_model': 7.800000000000001e-05, 'epoch': 0.03}) +[2026-06-17 16:25:16,221][starVLA.training.train_starvla][INFO] - Step 79, Loss: {'train/loss': 1.4823904037475586, 'train/grad_norm_pre_clip': 6.253221035003662, 'timing/data': 0.02576049044728279, 'timing/model': 0.8566319942474365, 'learning_rate/qwen_vl_interface': 7.9e-06, 'learning_rate/action_model': 7.900000000000001e-05, 'epoch': 0.03}) +[2026-06-17 16:25:19,598][starVLA.training.train_starvla][INFO] - Step 80, Loss: {'train/loss': 1.1716996431350708, 'train/grad_norm_pre_clip': 3.871633768081665, 'timing/data': 0.015838243067264557, 'timing/model': 0.8147793095558882, 'learning_rate/qwen_vl_interface': 8.000000000000001e-06, 'learning_rate/action_model': 8e-05, 'epoch': 0.03}) +[2026-06-17 16:25:23,499][starVLA.training.train_starvla][INFO] - Step 81, Loss: {'train/loss': 1.029255986213684, 'train/grad_norm_pre_clip': 2.9269673824310303, 'timing/data': 0.014951041899621487, 'timing/model': 0.8951889891177416, 'learning_rate/qwen_vl_interface': 8.1e-06, 'learning_rate/action_model': 8.1e-05, 'epoch': 0.03}) +[2026-06-17 16:25:28,201][starVLA.training.train_starvla][INFO] - Step 82, Loss: {'train/loss': 1.0028437376022339, 'train/grad_norm_pre_clip': 1.533387541770935, 'timing/data': 0.01349264569580555, 'timing/model': 0.8541998211294413, 'learning_rate/qwen_vl_interface': 8.2e-06, 'learning_rate/action_model': 8.2e-05, 'epoch': 0.03}) +[2026-06-17 16:25:31,662][starVLA.training.train_starvla][INFO] - Step 83, Loss: {'train/loss': 1.086920976638794, 'train/grad_norm_pre_clip': 3.766538381576538, 'timing/data': 0.01347308699041605, 'timing/model': 0.9193128561601043, 'learning_rate/qwen_vl_interface': 8.3e-06, 'learning_rate/action_model': 8.3e-05, 'epoch': 0.03}) +[2026-06-17 16:25:35,197][starVLA.training.train_starvla][INFO] - Step 84, Loss: {'train/loss': 1.127474069595337, 'train/grad_norm_pre_clip': 3.596298933029175, 'timing/data': 0.015145448967814445, 'timing/model': 0.8037465764209628, 'learning_rate/qwen_vl_interface': 8.400000000000001e-06, 'learning_rate/action_model': 8.4e-05, 'epoch': 0.03}) +[2026-06-17 16:25:38,479][starVLA.training.train_starvla][INFO] - Step 85, Loss: {'train/loss': 1.2944154739379883, 'train/grad_norm_pre_clip': 3.7810778617858887, 'timing/data': 0.014603846706449986, 'timing/model': 0.8962207864969969, 'learning_rate/qwen_vl_interface': 8.5e-06, 'learning_rate/action_model': 8.5e-05, 'epoch': 0.03}) +[2026-06-17 16:25:41,663][starVLA.training.train_starvla][INFO] - Step 86, Loss: {'train/loss': 1.1928069591522217, 'train/grad_norm_pre_clip': 3.831382989883423, 'timing/data': 0.012915488332509995, 'timing/model': 0.8337944252416492, 'learning_rate/qwen_vl_interface': 8.6e-06, 'learning_rate/action_model': 8.6e-05, 'epoch': 0.03}) +[2026-06-17 16:25:44,778][starVLA.training.train_starvla][INFO] - Step 87, Loss: {'train/loss': 1.2580978870391846, 'train/grad_norm_pre_clip': 1.895842432975769, 'timing/data': 0.01171769667416811, 'timing/model': 0.8041408462449908, 'learning_rate/qwen_vl_interface': 8.700000000000001e-06, 'learning_rate/action_model': 8.7e-05, 'epoch': 0.03}) +[2026-06-17 16:25:48,077][starVLA.training.train_starvla][INFO] - Step 88, Loss: {'train/loss': 1.0472831726074219, 'train/grad_norm_pre_clip': 1.8316885232925415, 'timing/data': 0.01738276146352291, 'timing/model': 0.8266118736937642, 'learning_rate/qwen_vl_interface': 8.8e-06, 'learning_rate/action_model': 8.800000000000001e-05, 'epoch': 0.03}) +[2026-06-17 16:25:51,283][starVLA.training.train_starvla][INFO] - Step 89, Loss: {'train/loss': 0.8874430656433105, 'train/grad_norm_pre_clip': 2.838747024536133, 'timing/data': 0.014510528184473515, 'timing/model': 0.8686184482648969, 'learning_rate/qwen_vl_interface': 8.900000000000001e-06, 'learning_rate/action_model': 8.900000000000001e-05, 'epoch': 0.03}) +[2026-06-17 16:25:54,453][starVLA.training.train_starvla][INFO] - Step 90, Loss: {'train/loss': 1.1570087671279907, 'train/grad_norm_pre_clip': 3.249016284942627, 'timing/data': 0.018324757926166058, 'timing/model': 0.8690743548795581, 'learning_rate/qwen_vl_interface': 9e-06, 'learning_rate/action_model': 9e-05, 'epoch': 0.03}) +[2026-06-17 16:25:57,758][starVLA.training.train_starvla][INFO] - Step 91, Loss: {'train/loss': 1.1111117601394653, 'train/grad_norm_pre_clip': 5.0005927085876465, 'timing/data': 0.012812968343496323, 'timing/model': 0.8113083988428116, 'learning_rate/qwen_vl_interface': 9.100000000000001e-06, 'learning_rate/action_model': 9.1e-05, 'epoch': 0.04}) +[2026-06-17 16:26:00,977][starVLA.training.train_starvla][INFO] - Step 92, Loss: {'train/loss': 0.7822999954223633, 'train/grad_norm_pre_clip': 1.759605884552002, 'timing/data': 0.015601199120283127, 'timing/model': 0.804766601882875, 'learning_rate/qwen_vl_interface': 9.200000000000002e-06, 'learning_rate/action_model': 9.200000000000001e-05, 'epoch': 0.04}) +[2026-06-17 16:26:04,219][starVLA.training.train_starvla][INFO] - Step 93, Loss: {'train/loss': 0.9664233326911926, 'train/grad_norm_pre_clip': 2.2716572284698486, 'timing/data': 0.011383474804461002, 'timing/model': 0.9234676584601402, 'learning_rate/qwen_vl_interface': 9.3e-06, 'learning_rate/action_model': 9.300000000000001e-05, 'epoch': 0.04}) +[2026-06-17 16:26:07,521][starVLA.training.train_starvla][INFO] - Step 94, Loss: {'train/loss': 1.0463160276412964, 'train/grad_norm_pre_clip': 3.6665096282958984, 'timing/data': 0.0156629029661417, 'timing/model': 0.8200420467182994, 'learning_rate/qwen_vl_interface': 9.4e-06, 'learning_rate/action_model': 9.4e-05, 'epoch': 0.04}) +[2026-06-17 16:26:10,808][starVLA.training.train_starvla][INFO] - Step 95, Loss: {'train/loss': 1.13018000125885, 'train/grad_norm_pre_clip': 2.29055118560791, 'timing/data': 0.011805969290435314, 'timing/model': 0.8574851276353002, 'learning_rate/qwen_vl_interface': 9.5e-06, 'learning_rate/action_model': 9.5e-05, 'epoch': 0.04}) +[2026-06-17 16:26:14,065][starVLA.training.train_starvla][INFO] - Step 96, Loss: {'train/loss': 1.482440710067749, 'train/grad_norm_pre_clip': 4.68086051940918, 'timing/data': 0.014945216476917267, 'timing/model': 0.8144885580986738, 'learning_rate/qwen_vl_interface': 9.600000000000001e-06, 'learning_rate/action_model': 9.6e-05, 'epoch': 0.04}) +[2026-06-17 16:26:17,439][starVLA.training.train_starvla][INFO] - Step 97, Loss: {'train/loss': 1.3539788722991943, 'train/grad_norm_pre_clip': 3.212711811065674, 'timing/data': 0.013221442699432373, 'timing/model': 0.8649200750514865, 'learning_rate/qwen_vl_interface': 9.7e-06, 'learning_rate/action_model': 9.7e-05, 'epoch': 0.04}) +[2026-06-17 16:26:20,638][starVLA.training.train_starvla][INFO] - Step 98, Loss: {'train/loss': 1.087705373764038, 'train/grad_norm_pre_clip': 8.80473518371582, 'timing/data': 0.013255443423986435, 'timing/model': 0.8597143469378352, 'learning_rate/qwen_vl_interface': 9.800000000000001e-06, 'learning_rate/action_model': 9.8e-05, 'epoch': 0.04}) +[2026-06-17 16:26:23,861][starVLA.training.train_starvla][INFO] - Step 99, Loss: {'train/loss': 1.195339560508728, 'train/grad_norm_pre_clip': 5.853241443634033, 'timing/data': 0.011605976149439812, 'timing/model': 0.8304829532280564, 'learning_rate/qwen_vl_interface': 9.9e-06, 'learning_rate/action_model': 9.900000000000001e-05, 'epoch': 0.04}) +[2026-06-17 16:28:07,120][root][INFO] - gcc -pthread -B /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/include -fPIC -O2 -isystem /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/include -fPIC -c /tmp/tmp1e2nscjc/test.c -o /tmp/tmp1e2nscjc/test.o +[2026-06-17 16:28:07,141][root][INFO] - gcc -pthread -B /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/compiler_compat /tmp/tmp1e2nscjc/test.o -laio -o /tmp/tmp1e2nscjc/a.out +[2026-06-17 16:28:07,774][root][INFO] - gcc -pthread -B /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/include -fPIC -O2 -isystem /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/include -fPIC -c /tmp/tmpozz7taro/test.c -o /tmp/tmpozz7taro/test.o +[2026-06-17 16:28:07,794][root][INFO] - gcc -pthread -B /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/compiler_compat /tmp/tmpozz7taro/test.o -L/usr/local/cuda -L/usr/local/cuda/lib64 -lcufile -o /tmp/tmpozz7taro/a.out +[2026-06-17 16:28:07,849][root][INFO] - gcc -pthread -B /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/include -fPIC -O2 -isystem /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/include -fPIC -c /tmp/tmp5kczq8rk/test.c -o /tmp/tmp5kczq8rk/test.o +[2026-06-17 16:28:07,868][root][INFO] - gcc -pthread -B /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/compiler_compat /tmp/tmp5kczq8rk/test.o -laio -o /tmp/tmp5kczq8rk/a.out +[2026-06-17 16:29:52,939][starVLA.training.train_starvla][INFO] - Step 100, Loss: {'train/loss': 1.0188014507293701, 'train/grad_norm_pre_clip': 3.271052122116089, 'eval/loss': 1.0362234115600586, 'eval/latency_4/loss': 1.0362234115600586, 'eval/demon_fire_precision': 0.9106773144040224, 'eval/demon_fire_recall': 1.0, 'eval/demon_fire_f1': 0.953250773993808, 'eval/demon_fire_rate_model': 0.9929515418502203, 'eval/demon_fire_rate_teacher': 0.9042584434654919, 'eval/demon_left_precision': 0.0, 'eval/demon_left_recall': 0.0, 'eval/demon_left_f1': 0.0, 'eval/demon_left_rate_model': 0.0, 'eval/demon_left_rate_teacher': 0.12863436123348018, 'eval/demon_right_precision': 0.9705882352941176, 'eval/demon_right_recall': 0.05956678700361011, 'eval/demon_right_f1': 0.11224489795918366, 'eval/demon_right_rate_model': 0.00998531571218796, 'eval/demon_right_rate_teacher': 0.16270190895741557, 'eval/demon_move_f1': 0.05612244897959183, 'eval/demon_cc_f1': 0.16447953099367607, 'eval/latency_4/demon_fire_precision': 0.9106773144040224, 'eval/latency_4/demon_fire_recall': 1.0, 'eval/latency_4/demon_fire_f1': 0.953250773993808, 'eval/latency_4/demon_fire_rate_model': 0.9929515418502203, 'eval/latency_4/demon_fire_rate_teacher': 0.9042584434654919, 'eval/latency_4/demon_left_precision': 0.0, 'eval/latency_4/demon_left_recall': 0.0, 'eval/latency_4/demon_left_f1': 0.0, 'eval/latency_4/demon_left_rate_model': 0.0, 'eval/latency_4/demon_left_rate_teacher': 0.12863436123348018, 'eval/latency_4/demon_right_precision': 0.9705882352941176, 'eval/latency_4/demon_right_recall': 0.05956678700361011, 'eval/latency_4/demon_right_f1': 0.11224489795918366, 'eval/latency_4/demon_right_rate_model': 0.00998531571218796, 'eval/latency_4/demon_right_rate_teacher': 0.16270190895741557, 'eval/latency_4/demon_move_f1': 0.05612244897959183, 'eval/latency_4/demon_cc_f1': 0.16447953099367607, 'timing/data': 0.015210217796266079, 'timing/model': 0.8193160742521286, 'learning_rate/qwen_vl_interface': 1e-05, 'learning_rate/action_model': 0.0001, 'epoch': 0.04}) +[2026-06-17 16:29:52,943][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_100_state +[2026-06-17 16:29:52,951][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 16:30:01,969][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_100_state/model.safetensors +[2026-06-17 16:30:24,316][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_100_state/optimizer.bin +[2026-06-17 16:30:24,318][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_100_state/sampler.bin +[2026-06-17 16:30:24,319][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_100_state/sampler_1.bin +[2026-06-17 16:30:24,333][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_100_state/random_states_0.pkl +[2026-06-17 16:30:24,341][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 100 +[2026-06-17 16:30:24,341][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 16:30:24,360][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 16:30:29,475][starVLA.training.train_starvla][INFO] - Step 101, Loss: {'train/loss': 1.0300631523132324, 'train/grad_norm_pre_clip': 2.7070724964141846, 'timing/data': 0.011206508614122868, 'timing/model': 1.6536225713789463, 'learning_rate/qwen_vl_interface': 9.999999507659945e-06, 'learning_rate/action_model': 9.999999507659945e-05, 'epoch': 0.04}) +[2026-06-17 16:30:34,068][starVLA.training.train_starvla][INFO] - Step 102, Loss: {'train/loss': 1.3890721797943115, 'train/grad_norm_pre_clip': 4.469397068023682, 'timing/data': 0.017773528583347797, 'timing/model': 1.412815441377461, 'learning_rate/qwen_vl_interface': 9.99999803063988e-06, 'learning_rate/action_model': 9.99999803063988e-05, 'epoch': 0.04}) +[2026-06-17 16:30:38,293][starVLA.training.train_starvla][INFO] - Step 103, Loss: {'train/loss': 0.7509788870811462, 'train/grad_norm_pre_clip': 2.5291404724121094, 'timing/data': 0.011072221212089062, 'timing/model': 1.077124496921897, 'learning_rate/qwen_vl_interface': 9.999995568940113e-06, 'learning_rate/action_model': 9.999995568940113e-05, 'epoch': 0.04}) +[2026-06-17 16:30:42,933][starVLA.training.train_starvla][INFO] - Step 104, Loss: {'train/loss': 1.1881812810897827, 'train/grad_norm_pre_clip': 4.151135444641113, 'timing/data': 0.01041903905570507, 'timing/model': 1.0758885452523828, 'learning_rate/qwen_vl_interface': 9.999992122561152e-06, 'learning_rate/action_model': 9.999992122561152e-05, 'epoch': 0.04}) +[2026-06-17 16:30:47,828][starVLA.training.train_starvla][INFO] - Step 105, Loss: {'train/loss': 1.3631596565246582, 'train/grad_norm_pre_clip': 7.049410343170166, 'timing/data': 0.013684128411114216, 'timing/model': 1.2285854239016771, 'learning_rate/qwen_vl_interface': 9.999987691503714e-06, 'learning_rate/action_model': 9.999987691503714e-05, 'epoch': 0.04}) +[2026-06-17 16:30:52,403][starVLA.training.train_starvla][INFO] - Step 106, Loss: {'train/loss': 1.0053396224975586, 'train/grad_norm_pre_clip': 1.7503031492233276, 'timing/data': 0.010676247999072075, 'timing/model': 1.0768429087474942, 'learning_rate/qwen_vl_interface': 9.999982275768716e-06, 'learning_rate/action_model': 9.999982275768715e-05, 'epoch': 0.04}) +[2026-06-17 16:30:57,031][starVLA.training.train_starvla][INFO] - Step 107, Loss: {'train/loss': 1.1325228214263916, 'train/grad_norm_pre_clip': 5.088133335113525, 'timing/data': 0.01409755740314722, 'timing/model': 1.427366259507835, 'learning_rate/qwen_vl_interface': 9.99997587535728e-06, 'learning_rate/action_model': 9.99997587535728e-05, 'epoch': 0.04}) +[2026-06-17 16:31:01,255][starVLA.training.train_starvla][INFO] - Step 108, Loss: {'train/loss': 1.026840329170227, 'train/grad_norm_pre_clip': 3.2782487869262695, 'timing/data': 0.012554697692394257, 'timing/model': 1.0680540800094604, 'learning_rate/qwen_vl_interface': 9.999968490270735e-06, 'learning_rate/action_model': 9.999968490270735e-05, 'epoch': 0.04}) +[2026-06-17 16:31:05,866][starVLA.training.train_starvla][INFO] - Step 109, Loss: {'train/loss': 0.8675819635391235, 'train/grad_norm_pre_clip': 11.385590553283691, 'timing/data': 0.011896573938429356, 'timing/model': 1.088974510319531, 'learning_rate/qwen_vl_interface': 9.99996012051061e-06, 'learning_rate/action_model': 9.99996012051061e-05, 'epoch': 0.04}) +[2026-06-17 16:31:10,469][starVLA.training.train_starvla][INFO] - Step 110, Loss: {'train/loss': 1.115728735923767, 'train/grad_norm_pre_clip': 5.946002960205078, 'timing/data': 0.010917270556092262, 'timing/model': 1.077411700040102, 'learning_rate/qwen_vl_interface': 9.999950766078641e-06, 'learning_rate/action_model': 9.99995076607864e-05, 'epoch': 0.04}) +[2026-06-17 16:31:14,933][starVLA.training.train_starvla][INFO] - Step 111, Loss: {'train/loss': 1.1660913228988647, 'train/grad_norm_pre_clip': 2.5819714069366455, 'timing/data': 0.011976142413914204, 'timing/model': 1.086023467592895, 'learning_rate/qwen_vl_interface': 9.999940426976768e-06, 'learning_rate/action_model': 9.999940426976768e-05, 'epoch': 0.04}) +[2026-06-17 16:31:19,072][starVLA.training.train_starvla][INFO] - Step 112, Loss: {'train/loss': 1.005719780921936, 'train/grad_norm_pre_clip': 4.533588409423828, 'timing/data': 0.010495245456695557, 'timing/model': 1.0907731782644987, 'learning_rate/qwen_vl_interface': 9.999929103207133e-06, 'learning_rate/action_model': 9.999929103207132e-05, 'epoch': 0.04}) +[2026-06-17 16:31:23,779][starVLA.training.train_starvla][INFO] - Step 113, Loss: {'train/loss': 1.002850890159607, 'train/grad_norm_pre_clip': 4.255741596221924, 'timing/data': 0.012936243787407875, 'timing/model': 1.0660658413544297, 'learning_rate/qwen_vl_interface': 9.999916794772083e-06, 'learning_rate/action_model': 9.999916794772082e-05, 'epoch': 0.04}) +[2026-06-17 16:31:28,352][starVLA.training.train_starvla][INFO] - Step 114, Loss: {'train/loss': 1.1874699592590332, 'train/grad_norm_pre_clip': 3.959261894226074, 'timing/data': 0.01062515564262867, 'timing/model': 1.0915609821677208, 'learning_rate/qwen_vl_interface': 9.99990350167417e-06, 'learning_rate/action_model': 9.999903501674169e-05, 'epoch': 0.04}) +[2026-06-17 16:31:32,838][starVLA.training.train_starvla][INFO] - Step 115, Loss: {'train/loss': 0.6349169611930847, 'train/grad_norm_pre_clip': 2.4812324047088623, 'timing/data': 0.011569405905902386, 'timing/model': 1.059834792278707, 'learning_rate/qwen_vl_interface': 9.99988922391615e-06, 'learning_rate/action_model': 9.99988922391615e-05, 'epoch': 0.04}) +[2026-06-17 16:31:37,313][starVLA.training.train_starvla][INFO] - Step 116, Loss: {'train/loss': 0.9431407451629639, 'train/grad_norm_pre_clip': 5.233458042144775, 'timing/data': 0.011989588849246502, 'timing/model': 1.0699935648590326, 'learning_rate/qwen_vl_interface': 9.999873961500985e-06, 'learning_rate/action_model': 9.999873961500984e-05, 'epoch': 0.04}) +[2026-06-17 16:31:41,917][starVLA.training.train_starvla][INFO] - Step 117, Loss: {'train/loss': 0.7059390544891357, 'train/grad_norm_pre_clip': 2.1368236541748047, 'timing/data': 0.013824615627527237, 'timing/model': 1.4289910681545734, 'learning_rate/qwen_vl_interface': 9.999857714431835e-06, 'learning_rate/action_model': 9.999857714431834e-05, 'epoch': 0.05}) +[2026-06-17 16:31:46,174][starVLA.training.train_starvla][INFO] - Step 118, Loss: {'train/loss': 1.3067809343338013, 'train/grad_norm_pre_clip': 6.720343112945557, 'timing/data': 0.01442328654229641, 'timing/model': 1.1262048743665218, 'learning_rate/qwen_vl_interface': 9.999840482712069e-06, 'learning_rate/action_model': 9.999840482712068e-05, 'epoch': 0.05}) +[2026-06-17 16:31:51,852][starVLA.training.train_starvla][INFO] - Step 119, Loss: {'train/loss': 1.2901853322982788, 'train/grad_norm_pre_clip': 4.348841667175293, 'timing/data': 0.018567107617855072, 'timing/model': 1.12063476908952, 'learning_rate/qwen_vl_interface': 9.99982226634526e-06, 'learning_rate/action_model': 9.99982226634526e-05, 'epoch': 0.05}) +[2026-06-17 16:31:56,481][starVLA.training.train_starvla][INFO] - Step 120, Loss: {'train/loss': 0.9834229350090027, 'train/grad_norm_pre_clip': 3.9359426498413086, 'timing/data': 0.011226836591959, 'timing/model': 1.073467554524541, 'learning_rate/qwen_vl_interface': 9.999803065335185e-06, 'learning_rate/action_model': 9.999803065335185e-05, 'epoch': 0.05}) +[2026-06-17 16:32:01,028][starVLA.training.train_starvla][INFO] - Step 121, Loss: {'train/loss': 1.1323093175888062, 'train/grad_norm_pre_clip': 3.5136849880218506, 'timing/data': 0.01214711181819439, 'timing/model': 1.074986140243709, 'learning_rate/qwen_vl_interface': 9.999782879685825e-06, 'learning_rate/action_model': 9.999782879685824e-05, 'epoch': 0.05}) +[2026-06-17 16:32:05,430][starVLA.training.train_starvla][INFO] - Step 122, Loss: {'train/loss': 0.943282425403595, 'train/grad_norm_pre_clip': 3.4403603076934814, 'timing/data': 0.011685912497341633, 'timing/model': 1.1768908947706223, 'learning_rate/qwen_vl_interface': 9.99976170940136e-06, 'learning_rate/action_model': 9.99976170940136e-05, 'epoch': 0.05}) +[2026-06-17 16:32:10,013][starVLA.training.train_starvla][INFO] - Step 123, Loss: {'train/loss': 1.1874651908874512, 'train/grad_norm_pre_clip': 3.0146305561065674, 'timing/data': 0.012151681818068027, 'timing/model': 1.085967842489481, 'learning_rate/qwen_vl_interface': 9.999739554486184e-06, 'learning_rate/action_model': 9.999739554486182e-05, 'epoch': 0.05}) +[2026-06-17 16:32:14,780][starVLA.training.train_starvla][INFO] - Step 124, Loss: {'train/loss': 1.0970476865768433, 'train/grad_norm_pre_clip': 8.391766548156738, 'timing/data': 0.01055796630680561, 'timing/model': 1.158581125549972, 'learning_rate/qwen_vl_interface': 9.999716414944887e-06, 'learning_rate/action_model': 9.999716414944886e-05, 'epoch': 0.05}) +[2026-06-17 16:32:21,249][starVLA.training.train_starvla][INFO] - Step 125, Loss: {'train/loss': 1.044297456741333, 'train/grad_norm_pre_clip': 6.552198886871338, 'timing/data': 0.0160626620054245, 'timing/model': 1.2692725397646427, 'learning_rate/qwen_vl_interface': 9.999692290782264e-06, 'learning_rate/action_model': 9.999692290782263e-05, 'epoch': 0.05}) +[2026-06-17 16:32:25,832][starVLA.training.train_starvla][INFO] - Step 126, Loss: {'train/loss': 1.2044901847839355, 'train/grad_norm_pre_clip': 3.831205368041992, 'timing/data': 0.01069697830826044, 'timing/model': 1.3951806463301182, 'learning_rate/qwen_vl_interface': 9.999667182003321e-06, 'learning_rate/action_model': 9.999667182003321e-05, 'epoch': 0.05}) +[2026-06-17 16:32:31,823][starVLA.training.train_starvla][INFO] - Step 127, Loss: {'train/loss': 0.8311891555786133, 'train/grad_norm_pre_clip': 3.5536117553710938, 'timing/data': 0.013414343819022179, 'timing/model': 2.8561570150777698, 'learning_rate/qwen_vl_interface': 9.99964108861326e-06, 'learning_rate/action_model': 9.99964108861326e-05, 'epoch': 0.05}) +[2026-06-17 16:32:36,137][starVLA.training.train_starvla][INFO] - Step 128, Loss: {'train/loss': 0.7144098281860352, 'train/grad_norm_pre_clip': 6.970891952514648, 'timing/data': 0.011743354611098766, 'timing/model': 1.1983080664649606, 'learning_rate/qwen_vl_interface': 9.99961401061749e-06, 'learning_rate/action_model': 9.99961401061749e-05, 'epoch': 0.05}) +[2026-06-17 16:32:40,853][starVLA.training.train_starvla][INFO] - Step 129, Loss: {'train/loss': 0.9293928146362305, 'train/grad_norm_pre_clip': 4.451779365539551, 'timing/data': 0.016147246584296227, 'timing/model': 1.103252929635346, 'learning_rate/qwen_vl_interface': 9.999585948021623e-06, 'learning_rate/action_model': 9.999585948021624e-05, 'epoch': 0.05}) +[2026-06-17 16:32:45,373][starVLA.training.train_starvla][INFO] - Step 130, Loss: {'train/loss': 0.9718278050422668, 'train/grad_norm_pre_clip': 10.98017406463623, 'timing/data': 0.010728464461863041, 'timing/model': 1.0728160087019205, 'learning_rate/qwen_vl_interface': 9.99955690083148e-06, 'learning_rate/action_model': 9.99955690083148e-05, 'epoch': 0.05}) +[2026-06-17 16:32:49,996][starVLA.training.train_starvla][INFO] - Step 131, Loss: {'train/loss': 0.7061062455177307, 'train/grad_norm_pre_clip': 2.698651075363159, 'timing/data': 0.0120389424264431, 'timing/model': 1.0762833030894399, 'learning_rate/qwen_vl_interface': 9.99952686905308e-06, 'learning_rate/action_model': 9.999526869053081e-05, 'epoch': 0.05}) +[2026-06-17 16:32:54,507][starVLA.training.train_starvla][INFO] - Step 132, Loss: {'train/loss': 0.7772436141967773, 'train/grad_norm_pre_clip': 2.3504886627197266, 'timing/data': 0.010376466438174248, 'timing/model': 1.403086450882256, 'learning_rate/qwen_vl_interface': 9.99949585269265e-06, 'learning_rate/action_model': 9.99949585269265e-05, 'epoch': 0.05}) +[2026-06-17 16:32:58,698][starVLA.training.train_starvla][INFO] - Step 133, Loss: {'train/loss': 1.4446543455123901, 'train/grad_norm_pre_clip': 5.775286674499512, 'timing/data': 0.012769543565809727, 'timing/model': 1.0777198681607842, 'learning_rate/qwen_vl_interface': 9.99946385175662e-06, 'learning_rate/action_model': 9.999463851756619e-05, 'epoch': 0.05}) +[2026-06-17 16:33:03,300][starVLA.training.train_starvla][INFO] - Step 134, Loss: {'train/loss': 0.832023561000824, 'train/grad_norm_pre_clip': 4.338196754455566, 'timing/data': 0.016597554087638855, 'timing/model': 1.0681916689500213, 'learning_rate/qwen_vl_interface': 9.999430866251623e-06, 'learning_rate/action_model': 9.999430866251622e-05, 'epoch': 0.05}) +[2026-06-17 16:33:08,549][starVLA.training.train_starvla][INFO] - Step 135, Loss: {'train/loss': 1.0712432861328125, 'train/grad_norm_pre_clip': 5.8870954513549805, 'timing/data': 0.011373516172170639, 'timing/model': 1.0704827709123492, 'learning_rate/qwen_vl_interface': 9.999396896184494e-06, 'learning_rate/action_model': 9.999396896184494e-05, 'epoch': 0.05}) +[2026-06-17 16:33:13,221][starVLA.training.train_starvla][INFO] - Step 136, Loss: {'train/loss': 1.0559797286987305, 'train/grad_norm_pre_clip': 7.467884063720703, 'timing/data': 0.011292580515146255, 'timing/model': 1.0760280769318342, 'learning_rate/qwen_vl_interface': 9.99936194156228e-06, 'learning_rate/action_model': 9.99936194156228e-05, 'epoch': 0.05}) +[2026-06-17 16:33:17,868][starVLA.training.train_starvla][INFO] - Step 137, Loss: {'train/loss': 0.951216995716095, 'train/grad_norm_pre_clip': 4.869441032409668, 'timing/data': 0.010961711406707764, 'timing/model': 1.5715782549232244, 'learning_rate/qwen_vl_interface': 9.999326002392225e-06, 'learning_rate/action_model': 9.999326002392224e-05, 'epoch': 0.05}) +[2026-06-17 16:33:22,049][starVLA.training.train_starvla][INFO] - Step 138, Loss: {'train/loss': 1.0854943990707397, 'train/grad_norm_pre_clip': 4.45547342300415, 'timing/data': 0.012936867773532867, 'timing/model': 1.085763887502253, 'learning_rate/qwen_vl_interface': 9.999289078681779e-06, 'learning_rate/action_model': 9.999289078681778e-05, 'epoch': 0.05}) +[2026-06-17 16:33:26,726][starVLA.training.train_starvla][INFO] - Step 139, Loss: {'train/loss': 1.1576534509658813, 'train/grad_norm_pre_clip': 4.182802200317383, 'timing/data': 0.010091956704854965, 'timing/model': 1.0675928108394146, 'learning_rate/qwen_vl_interface': 9.999251170438594e-06, 'learning_rate/action_model': 9.999251170438594e-05, 'epoch': 0.05}) +[2026-06-17 16:33:31,472][starVLA.training.train_starvla][INFO] - Step 140, Loss: {'train/loss': 0.8817929029464722, 'train/grad_norm_pre_clip': 3.188838243484497, 'timing/data': 0.011607847176492214, 'timing/model': 1.0675431545823812, 'learning_rate/qwen_vl_interface': 9.999212277670533e-06, 'learning_rate/action_model': 9.999212277670533e-05, 'epoch': 0.05}) +[2026-06-17 16:33:36,071][starVLA.training.train_starvla][INFO] - Step 141, Loss: {'train/loss': 1.0915296077728271, 'train/grad_norm_pre_clip': 3.6634743213653564, 'timing/data': 0.011152699589729309, 'timing/model': 1.174987856298685, 'learning_rate/qwen_vl_interface': 9.999172400385658e-06, 'learning_rate/action_model': 9.999172400385658e-05, 'epoch': 0.05}) +[2026-06-17 16:33:40,694][starVLA.training.train_starvla][INFO] - Step 142, Loss: {'train/loss': 0.743683397769928, 'train/grad_norm_pre_clip': 5.71626091003418, 'timing/data': 0.01224660873413086, 'timing/model': 1.4789734799414873, 'learning_rate/qwen_vl_interface': 9.999131538592232e-06, 'learning_rate/action_model': 9.999131538592232e-05, 'epoch': 0.05}) +[2026-06-17 16:33:44,992][starVLA.training.train_starvla][INFO] - Step 143, Loss: {'train/loss': 0.8427615761756897, 'train/grad_norm_pre_clip': 2.4254374504089355, 'timing/data': 0.01638936437666416, 'timing/model': 1.1406540982425213, 'learning_rate/qwen_vl_interface': 9.999089692298729e-06, 'learning_rate/action_model': 9.99908969229873e-05, 'epoch': 0.06}) +[2026-06-17 16:33:49,988][starVLA.training.train_starvla][INFO] - Step 144, Loss: {'train/loss': 0.8404865264892578, 'train/grad_norm_pre_clip': 3.2086386680603027, 'timing/data': 0.010723226703703403, 'timing/model': 1.0699427193030715, 'learning_rate/qwen_vl_interface': 9.999046861513822e-06, 'learning_rate/action_model': 9.999046861513821e-05, 'epoch': 0.06}) +[2026-06-17 16:33:54,617][starVLA.training.train_starvla][INFO] - Step 145, Loss: {'train/loss': 1.17562735080719, 'train/grad_norm_pre_clip': 3.4962387084960938, 'timing/data': 0.010383332148194313, 'timing/model': 1.1025514546781778, 'learning_rate/qwen_vl_interface': 9.99900304624639e-06, 'learning_rate/action_model': 9.99900304624639e-05, 'epoch': 0.06}) +[2026-06-17 16:33:59,223][starVLA.training.train_starvla][INFO] - Step 146, Loss: {'train/loss': 1.011065125465393, 'train/grad_norm_pre_clip': 3.092712163925171, 'timing/data': 0.013580070808529854, 'timing/model': 1.1526386216282845, 'learning_rate/qwen_vl_interface': 9.998958246505518e-06, 'learning_rate/action_model': 9.998958246505516e-05, 'epoch': 0.06}) +[2026-06-17 16:34:03,818][starVLA.training.train_starvla][INFO] - Step 147, Loss: {'train/loss': 1.4368945360183716, 'train/grad_norm_pre_clip': 7.4638824462890625, 'timing/data': 0.013138862326741219, 'timing/model': 1.4283930212259293, 'learning_rate/qwen_vl_interface': 9.998912462300489e-06, 'learning_rate/action_model': 9.998912462300488e-05, 'epoch': 0.06}) +[2026-06-17 16:34:08,993][starVLA.training.train_starvla][INFO] - Step 148, Loss: {'train/loss': 1.023432970046997, 'train/grad_norm_pre_clip': 4.201242923736572, 'timing/data': 0.017797320149838924, 'timing/model': 1.0788902128115296, 'learning_rate/qwen_vl_interface': 9.998865693640797e-06, 'learning_rate/action_model': 9.998865693640796e-05, 'epoch': 0.06}) +[2026-06-17 16:34:15,491][starVLA.training.train_starvla][INFO] - Step 149, Loss: {'train/loss': 1.0028223991394043, 'train/grad_norm_pre_clip': 4.893470287322998, 'timing/data': 0.016327211633324623, 'timing/model': 1.0857205465435982, 'learning_rate/qwen_vl_interface': 9.998817940536138e-06, 'learning_rate/action_model': 9.998817940536137e-05, 'epoch': 0.06}) +[2026-06-17 16:34:20,085][starVLA.training.train_starvla][INFO] - Step 150, Loss: {'train/loss': 0.8861450552940369, 'train/grad_norm_pre_clip': 8.019522666931152, 'timing/data': 0.011985751800239086, 'timing/model': 1.070465026423335, 'learning_rate/qwen_vl_interface': 9.99876920299641e-06, 'learning_rate/action_model': 9.998769202996408e-05, 'epoch': 0.06}) +[2026-06-17 16:34:24,720][starVLA.training.train_starvla][INFO] - Step 151, Loss: {'train/loss': 1.047117829322815, 'train/grad_norm_pre_clip': 9.884783744812012, 'timing/data': 0.010278655216097832, 'timing/model': 1.1247891802340746, 'learning_rate/qwen_vl_interface': 9.998719481031715e-06, 'learning_rate/action_model': 9.998719481031715e-05, 'epoch': 0.06}) +[2026-06-17 16:34:29,330][starVLA.training.train_starvla][INFO] - Step 152, Loss: {'train/loss': 1.244120717048645, 'train/grad_norm_pre_clip': 11.61606216430664, 'timing/data': 0.012977996841073036, 'timing/model': 1.5228114649653435, 'learning_rate/qwen_vl_interface': 9.998668774652363e-06, 'learning_rate/action_model': 9.998668774652362e-05, 'epoch': 0.06}) +[2026-06-17 16:34:33,617][starVLA.training.train_starvla][INFO] - Step 153, Loss: {'train/loss': 0.9742523431777954, 'train/grad_norm_pre_clip': 6.699670791625977, 'timing/data': 0.013352484442293644, 'timing/model': 1.1259789168834686, 'learning_rate/qwen_vl_interface': 9.998617083868863e-06, 'learning_rate/action_model': 9.998617083868863e-05, 'epoch': 0.06}) +[2026-06-17 16:34:38,165][starVLA.training.train_starvla][INFO] - Step 154, Loss: {'train/loss': 1.0172139406204224, 'train/grad_norm_pre_clip': 13.232314109802246, 'timing/data': 0.018288658000528812, 'timing/model': 1.0874957675114274, 'learning_rate/qwen_vl_interface': 9.998564408691933e-06, 'learning_rate/action_model': 9.998564408691932e-05, 'epoch': 0.06}) +[2026-06-17 16:34:42,721][starVLA.training.train_starvla][INFO] - Step 155, Loss: {'train/loss': 1.345048427581787, 'train/grad_norm_pre_clip': 9.066437721252441, 'timing/data': 0.010613029822707176, 'timing/model': 1.0654134880751371, 'learning_rate/qwen_vl_interface': 9.998510749132491e-06, 'learning_rate/action_model': 9.998510749132491e-05, 'epoch': 0.06}) +[2026-06-17 16:34:47,311][starVLA.training.train_starvla][INFO] - Step 156, Loss: {'train/loss': 1.1770284175872803, 'train/grad_norm_pre_clip': 10.884309768676758, 'timing/data': 0.011117655783891678, 'timing/model': 1.1180780278518796, 'learning_rate/qwen_vl_interface': 9.998456105201659e-06, 'learning_rate/action_model': 9.99845610520166e-05, 'epoch': 0.06}) +[2026-06-17 16:34:52,181][starVLA.training.train_starvla][INFO] - Step 157, Loss: {'train/loss': 0.9092298150062561, 'train/grad_norm_pre_clip': 3.8813300132751465, 'timing/data': 0.011754630133509636, 'timing/model': 1.4660528898239136, 'learning_rate/qwen_vl_interface': 9.99840047691077e-06, 'learning_rate/action_model': 9.99840047691077e-05, 'epoch': 0.06}) +[2026-06-17 16:34:56,389][starVLA.training.train_starvla][INFO] - Step 158, Loss: {'train/loss': 0.7589980363845825, 'train/grad_norm_pre_clip': 4.4464921951293945, 'timing/data': 0.011051582172513008, 'timing/model': 1.0865583885461092, 'learning_rate/qwen_vl_interface': 9.998343864271352e-06, 'learning_rate/action_model': 9.998343864271351e-05, 'epoch': 0.06}) +[2026-06-17 16:35:01,036][starVLA.training.train_starvla][INFO] - Step 159, Loss: {'train/loss': 0.9220606684684753, 'train/grad_norm_pre_clip': 5.57442569732666, 'timing/data': 0.014001060277223587, 'timing/model': 1.0745411701500416, 'learning_rate/qwen_vl_interface': 9.99828626729514e-06, 'learning_rate/action_model': 9.998286267295139e-05, 'epoch': 0.06}) +[2026-06-17 16:35:05,583][starVLA.training.train_starvla][INFO] - Step 160, Loss: {'train/loss': 1.107994794845581, 'train/grad_norm_pre_clip': 4.299795150756836, 'timing/data': 0.009987820871174335, 'timing/model': 1.0718546053394675, 'learning_rate/qwen_vl_interface': 9.998227685994077e-06, 'learning_rate/action_model': 9.998227685994075e-05, 'epoch': 0.06}) +[2026-06-17 16:35:10,324][starVLA.training.train_starvla][INFO] - Step 161, Loss: {'train/loss': 0.9123132824897766, 'train/grad_norm_pre_clip': 2.9011495113372803, 'timing/data': 0.014126935042440891, 'timing/model': 1.1266343919560313, 'learning_rate/qwen_vl_interface': 9.998168120380303e-06, 'learning_rate/action_model': 9.998168120380303e-05, 'epoch': 0.06}) +[2026-06-17 16:35:15,096][starVLA.training.train_starvla][INFO] - Step 162, Loss: {'train/loss': 1.415859341621399, 'train/grad_norm_pre_clip': 4.598431587219238, 'timing/data': 0.012238580733537674, 'timing/model': 1.5403762198984623, 'learning_rate/qwen_vl_interface': 9.99810757046617e-06, 'learning_rate/action_model': 9.99810757046617e-05, 'epoch': 0.06}) +[2026-06-17 16:35:19,772][starVLA.training.train_starvla][INFO] - Step 163, Loss: {'train/loss': 0.8438358902931213, 'train/grad_norm_pre_clip': 7.280831336975098, 'timing/data': 0.012870822101831436, 'timing/model': 1.3876456515863538, 'learning_rate/qwen_vl_interface': 9.998046036264229e-06, 'learning_rate/action_model': 9.99804603626423e-05, 'epoch': 0.06}) +[2026-06-17 16:35:24,868][starVLA.training.train_starvla][INFO] - Step 164, Loss: {'train/loss': 0.88211989402771, 'train/grad_norm_pre_clip': 7.4847283363342285, 'timing/data': 0.011362282559275627, 'timing/model': 1.1074566710740328, 'learning_rate/qwen_vl_interface': 9.997983517787236e-06, 'learning_rate/action_model': 9.997983517787236e-05, 'epoch': 0.06}) +[2026-06-17 16:35:29,491][starVLA.training.train_starvla][INFO] - Step 165, Loss: {'train/loss': 1.002625823020935, 'train/grad_norm_pre_clip': 6.23567533493042, 'timing/data': 0.014904242008924484, 'timing/model': 1.0688083237037063, 'learning_rate/qwen_vl_interface': 9.99792001504815e-06, 'learning_rate/action_model': 9.997920015048149e-05, 'epoch': 0.06}) +[2026-06-17 16:35:34,067][starVLA.training.train_starvla][INFO] - Step 166, Loss: {'train/loss': 0.9842133522033691, 'train/grad_norm_pre_clip': 4.806310176849365, 'timing/data': 0.015203014947474003, 'timing/model': 1.0913786264136434, 'learning_rate/qwen_vl_interface': 9.997855528060136e-06, 'learning_rate/action_model': 9.997855528060135e-05, 'epoch': 0.06}) +[2026-06-17 16:35:38,629][starVLA.training.train_starvla][INFO] - Step 167, Loss: {'train/loss': 1.14923894405365, 'train/grad_norm_pre_clip': 3.6826932430267334, 'timing/data': 0.011290276423096657, 'timing/model': 1.3912693075835705, 'learning_rate/qwen_vl_interface': 9.997790056836562e-06, 'learning_rate/action_model': 9.997790056836562e-05, 'epoch': 0.06}) +[2026-06-17 16:35:42,889][starVLA.training.train_starvla][INFO] - Step 168, Loss: {'train/loss': 1.3064239025115967, 'train/grad_norm_pre_clip': 5.140592575073242, 'timing/data': 0.01069695595651865, 'timing/model': 1.0656047817319632, 'learning_rate/qwen_vl_interface': 9.997723601390998e-06, 'learning_rate/action_model': 9.997723601390998e-05, 'epoch': 0.06}) +[2026-06-17 16:35:48,774][starVLA.training.train_starvla][INFO] - Step 169, Loss: {'train/loss': 1.0396803617477417, 'train/grad_norm_pre_clip': 8.729095458984375, 'timing/data': 0.01145661249756813, 'timing/model': 1.545573285780847, 'learning_rate/qwen_vl_interface': 9.997656161737227e-06, 'learning_rate/action_model': 9.997656161737226e-05, 'epoch': 0.07}) +[2026-06-17 16:35:53,622][starVLA.training.train_starvla][INFO] - Step 170, Loss: {'train/loss': 0.6971237063407898, 'train/grad_norm_pre_clip': 3.664698600769043, 'timing/data': 0.010631339624524117, 'timing/model': 1.3080905629321933, 'learning_rate/qwen_vl_interface': 9.997587737889222e-06, 'learning_rate/action_model': 9.997587737889221e-05, 'epoch': 0.07}) +[2026-06-17 16:35:58,907][starVLA.training.train_starvla][INFO] - Step 171, Loss: {'train/loss': 0.6494559049606323, 'train/grad_norm_pre_clip': 3.9433419704437256, 'timing/data': 0.012773528695106506, 'timing/model': 1.7229865668341517, 'learning_rate/qwen_vl_interface': 9.997518329861172e-06, 'learning_rate/action_model': 9.997518329861171e-05, 'epoch': 0.07}) +[2026-06-17 16:36:03,409][starVLA.training.train_starvla][INFO] - Step 172, Loss: {'train/loss': 0.8182448744773865, 'train/grad_norm_pre_clip': 5.411736488342285, 'timing/data': 0.01200704462826252, 'timing/model': 1.3900039531290531, 'learning_rate/qwen_vl_interface': 9.997447937667462e-06, 'learning_rate/action_model': 9.997447937667462e-05, 'epoch': 0.07}) +[2026-06-17 16:36:08,765][starVLA.training.train_starvla][INFO] - Step 173, Loss: {'train/loss': 0.9594770669937134, 'train/grad_norm_pre_clip': 12.745850563049316, 'timing/data': 0.015838325954973698, 'timing/model': 1.4294019127264619, 'learning_rate/qwen_vl_interface': 9.997376561322686e-06, 'learning_rate/action_model': 9.997376561322686e-05, 'epoch': 0.07}) +[2026-06-17 16:36:13,579][starVLA.training.train_starvla][INFO] - Step 174, Loss: {'train/loss': 0.7870934009552002, 'train/grad_norm_pre_clip': 12.781222343444824, 'timing/data': 0.011304794810712337, 'timing/model': 1.0816901428624988, 'learning_rate/qwen_vl_interface': 9.997304200841642e-06, 'learning_rate/action_model': 9.997304200841641e-05, 'epoch': 0.07}) +[2026-06-17 16:36:18,294][starVLA.training.train_starvla][INFO] - Step 175, Loss: {'train/loss': 0.7472814321517944, 'train/grad_norm_pre_clip': 6.776250839233398, 'timing/data': 0.012784396298229694, 'timing/model': 1.1841225996613503, 'learning_rate/qwen_vl_interface': 9.997230856239327e-06, 'learning_rate/action_model': 9.997230856239327e-05, 'epoch': 0.07}) +[2026-06-17 16:36:22,935][starVLA.training.train_starvla][INFO] - Step 176, Loss: {'train/loss': 1.174208402633667, 'train/grad_norm_pre_clip': 5.095629692077637, 'timing/data': 0.011792100965976715, 'timing/model': 1.1241083471104503, 'learning_rate/qwen_vl_interface': 9.997156527530948e-06, 'learning_rate/action_model': 9.997156527530948e-05, 'epoch': 0.07}) +[2026-06-17 16:36:27,557][starVLA.training.train_starvla][INFO] - Step 177, Loss: {'train/loss': 0.9826245903968811, 'train/grad_norm_pre_clip': 3.644099473953247, 'timing/data': 0.010985025204718113, 'timing/model': 1.4365269280970097, 'learning_rate/qwen_vl_interface': 9.997081214731915e-06, 'learning_rate/action_model': 9.997081214731914e-05, 'epoch': 0.07}) +[2026-06-17 16:36:31,892][starVLA.training.train_starvla][INFO] - Step 178, Loss: {'train/loss': 0.7978557348251343, 'train/grad_norm_pre_clip': 3.5888850688934326, 'timing/data': 0.010775201953947544, 'timing/model': 1.1282743699848652, 'learning_rate/qwen_vl_interface': 9.997004917857834e-06, 'learning_rate/action_model': 9.997004917857834e-05, 'epoch': 0.07}) +[2026-06-17 16:36:36,469][starVLA.training.train_starvla][INFO] - Step 179, Loss: {'train/loss': 0.6616004109382629, 'train/grad_norm_pre_clip': 3.7146105766296387, 'timing/data': 0.010815546847879887, 'timing/model': 1.1254568845033646, 'learning_rate/qwen_vl_interface': 9.996927636924529e-06, 'learning_rate/action_model': 9.996927636924528e-05, 'epoch': 0.07}) +[2026-06-17 16:36:40,993][starVLA.training.train_starvla][INFO] - Step 180, Loss: {'train/loss': 1.168986201286316, 'train/grad_norm_pre_clip': 4.330134868621826, 'timing/data': 0.013435974717140198, 'timing/model': 1.0646594259887934, 'learning_rate/qwen_vl_interface': 9.996849371948016e-06, 'learning_rate/action_model': 9.996849371948016e-05, 'epoch': 0.07}) +[2026-06-17 16:36:45,510][starVLA.training.train_starvla][INFO] - Step 181, Loss: {'train/loss': 0.6941206455230713, 'train/grad_norm_pre_clip': 4.072828769683838, 'timing/data': 0.010987953282892704, 'timing/model': 1.0889117792248726, 'learning_rate/qwen_vl_interface': 9.996770122944518e-06, 'learning_rate/action_model': 9.996770122944518e-05, 'epoch': 0.07}) +[2026-06-17 16:36:50,141][starVLA.training.train_starvla][INFO] - Step 182, Loss: {'train/loss': 0.684899628162384, 'train/grad_norm_pre_clip': 5.6607747077941895, 'timing/data': 0.013189130462706089, 'timing/model': 1.4372756481170654, 'learning_rate/qwen_vl_interface': 9.99668988993047e-06, 'learning_rate/action_model': 9.99668988993047e-05, 'epoch': 0.07}) +[2026-06-17 16:36:54,299][starVLA.training.train_starvla][INFO] - Step 183, Loss: {'train/loss': 0.7567936778068542, 'train/grad_norm_pre_clip': 4.794442176818848, 'timing/data': 0.011070878244936466, 'timing/model': 1.0674962373450398, 'learning_rate/qwen_vl_interface': 9.996608672922496e-06, 'learning_rate/action_model': 9.996608672922496e-05, 'epoch': 0.07}) +[2026-06-17 16:36:58,932][starVLA.training.train_starvla][INFO] - Step 184, Loss: {'train/loss': 0.8143054842948914, 'train/grad_norm_pre_clip': 3.3031067848205566, 'timing/data': 0.011630835942924023, 'timing/model': 1.1119118686765432, 'learning_rate/qwen_vl_interface': 9.996526471937438e-06, 'learning_rate/action_model': 9.996526471937439e-05, 'epoch': 0.07}) +[2026-06-17 16:37:03,657][starVLA.training.train_starvla][INFO] - Step 185, Loss: {'train/loss': 0.5324039459228516, 'train/grad_norm_pre_clip': 2.9192399978637695, 'timing/data': 0.01145102083683014, 'timing/model': 1.0655099125579, 'learning_rate/qwen_vl_interface': 9.996443286992337e-06, 'learning_rate/action_model': 9.996443286992336e-05, 'epoch': 0.07}) +[2026-06-17 16:37:08,241][starVLA.training.train_starvla][INFO] - Step 186, Loss: {'train/loss': 1.0920125246047974, 'train/grad_norm_pre_clip': 5.99272346496582, 'timing/data': 0.01067386381328106, 'timing/model': 1.077722104266286, 'learning_rate/qwen_vl_interface': 9.996359118104432e-06, 'learning_rate/action_model': 9.996359118104432e-05, 'epoch': 0.07}) +[2026-06-17 16:37:12,929][starVLA.training.train_starvla][INFO] - Step 187, Loss: {'train/loss': 0.8627322316169739, 'train/grad_norm_pre_clip': 5.6966705322265625, 'timing/data': 0.014841219410300255, 'timing/model': 1.417346472851932, 'learning_rate/qwen_vl_interface': 9.996273965291177e-06, 'learning_rate/action_model': 9.996273965291176e-05, 'epoch': 0.07}) +[2026-06-17 16:37:17,273][starVLA.training.train_starvla][INFO] - Step 188, Loss: {'train/loss': 1.2300522327423096, 'train/grad_norm_pre_clip': 17.9871768951416, 'timing/data': 0.011566285975277424, 'timing/model': 1.190312722697854, 'learning_rate/qwen_vl_interface': 9.996187828570221e-06, 'learning_rate/action_model': 9.996187828570221e-05, 'epoch': 0.07}) +[2026-06-17 16:37:23,408][starVLA.training.train_starvla][INFO] - Step 189, Loss: {'train/loss': 0.8287484049797058, 'train/grad_norm_pre_clip': 4.221988201141357, 'timing/data': 0.011090604588389397, 'timing/model': 1.170533923432231, 'learning_rate/qwen_vl_interface': 9.996100707959422e-06, 'learning_rate/action_model': 9.996100707959421e-05, 'epoch': 0.07}) +[2026-06-17 16:37:29,185][starVLA.training.train_starvla][INFO] - Step 190, Loss: {'train/loss': 1.0128037929534912, 'train/grad_norm_pre_clip': 2.8997573852539062, 'timing/data': 0.013466428965330124, 'timing/model': 1.2041391748934984, 'learning_rate/qwen_vl_interface': 9.996012603476836e-06, 'learning_rate/action_model': 9.996012603476837e-05, 'epoch': 0.07}) +[2026-06-17 16:37:33,771][starVLA.training.train_starvla][INFO] - Step 191, Loss: {'train/loss': 0.8476590514183044, 'train/grad_norm_pre_clip': 6.7175703048706055, 'timing/data': 0.010398034937679768, 'timing/model': 1.0839679297059774, 'learning_rate/qwen_vl_interface': 9.995923515140734e-06, 'learning_rate/action_model': 9.995923515140732e-05, 'epoch': 0.07}) +[2026-06-17 16:37:40,513][starVLA.training.train_starvla][INFO] - Step 192, Loss: {'train/loss': 0.8880115747451782, 'train/grad_norm_pre_clip': 10.318643569946289, 'timing/data': 0.01816361490637064, 'timing/model': 2.4058126816526055, 'learning_rate/qwen_vl_interface': 9.995833442969578e-06, 'learning_rate/action_model': 9.995833442969578e-05, 'epoch': 0.07}) +[2026-06-17 16:37:45,056][starVLA.training.train_starvla][INFO] - Step 193, Loss: {'train/loss': 1.0697075128555298, 'train/grad_norm_pre_clip': 4.775543212890625, 'timing/data': 0.012712450698018074, 'timing/model': 1.1942890379577875, 'learning_rate/qwen_vl_interface': 9.995742386982045e-06, 'learning_rate/action_model': 9.995742386982045e-05, 'epoch': 0.07}) +[2026-06-17 16:37:49,815][starVLA.training.train_starvla][INFO] - Step 194, Loss: {'train/loss': 0.6188533902168274, 'train/grad_norm_pre_clip': 5.879185676574707, 'timing/data': 0.011702440679073334, 'timing/model': 1.0759842982515693, 'learning_rate/qwen_vl_interface': 9.995650347197006e-06, 'learning_rate/action_model': 9.995650347197005e-05, 'epoch': 0.07}) +[2026-06-17 16:37:56,574][starVLA.training.train_starvla][INFO] - Step 195, Loss: {'train/loss': 0.763529896736145, 'train/grad_norm_pre_clip': 4.540480136871338, 'timing/data': 0.011235657148063183, 'timing/model': 1.9672206174582243, 'learning_rate/qwen_vl_interface': 9.995557323633544e-06, 'learning_rate/action_model': 9.995557323633543e-05, 'epoch': 0.08}) +[2026-06-17 16:38:01,362][starVLA.training.train_starvla][INFO] - Step 196, Loss: {'train/loss': 0.8993850350379944, 'train/grad_norm_pre_clip': 3.7801666259765625, 'timing/data': 0.011837470345199108, 'timing/model': 1.073596052825451, 'learning_rate/qwen_vl_interface': 9.995463316310942e-06, 'learning_rate/action_model': 9.995463316310943e-05, 'epoch': 0.08}) +[2026-06-17 16:38:05,976][starVLA.training.train_starvla][INFO] - Step 197, Loss: {'train/loss': 1.1126408576965332, 'train/grad_norm_pre_clip': 4.93461799621582, 'timing/data': 0.011121656745672226, 'timing/model': 1.5024203220382333, 'learning_rate/qwen_vl_interface': 9.995368325248688e-06, 'learning_rate/action_model': 9.995368325248688e-05, 'epoch': 0.08}) +[2026-06-17 16:38:10,219][starVLA.training.train_starvla][INFO] - Step 198, Loss: {'train/loss': 0.8835973143577576, 'train/grad_norm_pre_clip': 5.345019340515137, 'timing/data': 0.01784365065395832, 'timing/model': 1.1069120485335588, 'learning_rate/qwen_vl_interface': 9.995272350466476e-06, 'learning_rate/action_model': 9.995272350466476e-05, 'epoch': 0.08}) +[2026-06-17 16:38:14,770][starVLA.training.train_starvla][INFO] - Step 199, Loss: {'train/loss': 0.783322811126709, 'train/grad_norm_pre_clip': 5.743014812469482, 'timing/data': 0.01612106990069151, 'timing/model': 1.0863513620570302, 'learning_rate/qwen_vl_interface': 9.995175391984199e-06, 'learning_rate/action_model': 9.995175391984198e-05, 'epoch': 0.08}) +[2026-06-17 16:40:39,547][starVLA.training.train_starvla][INFO] - Step 200, Loss: {'train/loss': 0.8010333776473999, 'train/grad_norm_pre_clip': 7.110709190368652, 'eval/loss': 0.8833144903182983, 'eval/latency_4/loss': 0.8833144903182983, 'eval/demon_fire_precision': 0.9114860864416815, 'eval/demon_fire_recall': 1.0, 'eval/demon_fire_f1': 0.9536936657890662, 'eval/demon_fire_rate_model': 0.9920704845814978, 'eval/demon_fire_rate_teacher': 0.9042584434654919, 'eval/demon_left_precision': 0.6666666666666666, 'eval/demon_left_recall': 0.0045662100456621, 'eval/demon_left_f1': 0.009070294784580497, 'eval/demon_left_rate_model': 0.000881057268722467, 'eval/demon_left_rate_teacher': 0.12863436123348018, 'eval/demon_right_precision': 0.905982905982906, 'eval/demon_right_recall': 0.19133574007220217, 'eval/demon_right_f1': 0.31594634873323396, 'eval/demon_right_rate_model': 0.03436123348017621, 'eval/demon_right_rate_teacher': 0.16270190895741557, 'eval/demon_move_f1': 0.16250832175890723, 'eval/demon_cc_f1': 0.3875897459664601, 'eval/latency_4/demon_fire_precision': 0.9114860864416815, 'eval/latency_4/demon_fire_recall': 1.0, 'eval/latency_4/demon_fire_f1': 0.9536936657890662, 'eval/latency_4/demon_fire_rate_model': 0.9920704845814978, 'eval/latency_4/demon_fire_rate_teacher': 0.9042584434654919, 'eval/latency_4/demon_left_precision': 0.6666666666666666, 'eval/latency_4/demon_left_recall': 0.0045662100456621, 'eval/latency_4/demon_left_f1': 0.009070294784580497, 'eval/latency_4/demon_left_rate_model': 0.000881057268722467, 'eval/latency_4/demon_left_rate_teacher': 0.12863436123348018, 'eval/latency_4/demon_right_precision': 0.905982905982906, 'eval/latency_4/demon_right_recall': 0.19133574007220217, 'eval/latency_4/demon_right_f1': 0.31594634873323396, 'eval/latency_4/demon_right_rate_model': 0.03436123348017621, 'eval/latency_4/demon_right_rate_teacher': 0.16270190895741557, 'eval/latency_4/demon_move_f1': 0.16250832175890723, 'eval/latency_4/demon_cc_f1': 0.3875897459664601, 'timing/data': 0.011102661490440369, 'timing/model': 1.0685019819065928, 'learning_rate/qwen_vl_interface': 9.995077449821957e-06, 'learning_rate/action_model': 9.995077449821956e-05, 'epoch': 0.08}) +[2026-06-17 16:40:39,553][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_200_state +[2026-06-17 16:40:39,562][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 16:40:49,256][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_200_state/model.safetensors +[2026-06-17 16:41:10,909][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_200_state/optimizer.bin +[2026-06-17 16:41:10,911][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_200_state/sampler.bin +[2026-06-17 16:41:10,912][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_200_state/sampler_1.bin +[2026-06-17 16:41:10,914][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_200_state/random_states_0.pkl +[2026-06-17 16:41:11,016][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 200 +[2026-06-17 16:41:11,017][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 16:41:11,035][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 16:41:16,006][starVLA.training.train_starvla][INFO] - Step 201, Loss: {'train/loss': 0.8297780752182007, 'train/grad_norm_pre_clip': 5.4900665283203125, 'timing/data': 0.02333439327776432, 'timing/model': 1.2081434885039926, 'learning_rate/qwen_vl_interface': 9.994978524000053e-06, 'learning_rate/action_model': 9.994978524000052e-05, 'epoch': 0.08}) +[2026-06-17 16:41:20,746][starVLA.training.train_starvla][INFO] - Step 202, Loss: {'train/loss': 1.270453691482544, 'train/grad_norm_pre_clip': 7.103745937347412, 'timing/data': 0.012969130650162697, 'timing/model': 1.422259054146707, 'learning_rate/qwen_vl_interface': 9.994878614538996e-06, 'learning_rate/action_model': 9.994878614538996e-05, 'epoch': 0.08}) +[2026-06-17 16:41:25,193][starVLA.training.train_starvla][INFO] - Step 203, Loss: {'train/loss': 0.7870923280715942, 'train/grad_norm_pre_clip': 7.445056915283203, 'timing/data': 0.013762309215962887, 'timing/model': 1.1361437048763037, 'learning_rate/qwen_vl_interface': 9.994777721459496e-06, 'learning_rate/action_model': 9.994777721459497e-05, 'epoch': 0.08}) +[2026-06-17 16:41:29,971][starVLA.training.train_starvla][INFO] - Step 204, Loss: {'train/loss': 0.7652762532234192, 'train/grad_norm_pre_clip': 7.754420280456543, 'timing/data': 0.019671195186674595, 'timing/model': 1.1509327786043286, 'learning_rate/qwen_vl_interface': 9.994675844782468e-06, 'learning_rate/action_model': 9.994675844782468e-05, 'epoch': 0.08}) +[2026-06-17 16:41:34,662][starVLA.training.train_starvla][INFO] - Step 205, Loss: {'train/loss': 1.096868634223938, 'train/grad_norm_pre_clip': 5.6338372230529785, 'timing/data': 0.012552719563245773, 'timing/model': 1.0919847348704934, 'learning_rate/qwen_vl_interface': 9.994572984529033e-06, 'learning_rate/action_model': 9.994572984529034e-05, 'epoch': 0.08}) +[2026-06-17 16:41:39,736][starVLA.training.train_starvla][INFO] - Step 206, Loss: {'train/loss': 1.0944072008132935, 'train/grad_norm_pre_clip': 3.9296603202819824, 'timing/data': 0.013219390995800495, 'timing/model': 1.1850110553205013, 'learning_rate/qwen_vl_interface': 9.994469140720513e-06, 'learning_rate/action_model': 9.994469140720512e-05, 'epoch': 0.08}) +[2026-06-17 16:41:44,281][starVLA.training.train_starvla][INFO] - Step 207, Loss: {'train/loss': 0.9439321160316467, 'train/grad_norm_pre_clip': 6.746558666229248, 'timing/data': 0.014195667579770088, 'timing/model': 1.3674267046153545, 'learning_rate/qwen_vl_interface': 9.994364313378435e-06, 'learning_rate/action_model': 9.994364313378433e-05, 'epoch': 0.08}) +[2026-06-17 16:41:48,560][starVLA.training.train_starvla][INFO] - Step 208, Loss: {'train/loss': 1.2140228748321533, 'train/grad_norm_pre_clip': 8.056325912475586, 'timing/data': 0.011331991292536259, 'timing/model': 1.0960377482697368, 'learning_rate/qwen_vl_interface': 9.99425850252453e-06, 'learning_rate/action_model': 9.994258502524529e-05, 'epoch': 0.08}) +[2026-06-17 16:41:53,137][starVLA.training.train_starvla][INFO] - Step 209, Loss: {'train/loss': 1.1058465242385864, 'train/grad_norm_pre_clip': 8.036086082458496, 'timing/data': 0.010779676027595997, 'timing/model': 1.1413672296330333, 'learning_rate/qwen_vl_interface': 9.994151708180731e-06, 'learning_rate/action_model': 9.99415170818073e-05, 'epoch': 0.08}) +[2026-06-17 16:41:57,839][starVLA.training.train_starvla][INFO] - Step 210, Loss: {'train/loss': 1.0114095211029053, 'train/grad_norm_pre_clip': 6.9423651695251465, 'timing/data': 0.011562061496078968, 'timing/model': 1.0914497105404735, 'learning_rate/qwen_vl_interface': 9.994043930369179e-06, 'learning_rate/action_model': 9.994043930369178e-05, 'epoch': 0.08}) +[2026-06-17 16:42:02,555][starVLA.training.train_starvla][INFO] - Step 211, Loss: {'train/loss': 0.8901981711387634, 'train/grad_norm_pre_clip': 6.619754314422607, 'timing/data': 0.009920870885252953, 'timing/model': 1.1517804432660341, 'learning_rate/qwen_vl_interface': 9.993935169112213e-06, 'learning_rate/action_model': 9.993935169112213e-05, 'epoch': 0.08}) +[2026-06-17 16:42:07,608][starVLA.training.train_starvla][INFO] - Step 212, Loss: {'train/loss': 1.034954309463501, 'train/grad_norm_pre_clip': 6.39982795715332, 'timing/data': 0.013551987707614899, 'timing/model': 1.7376108458265662, 'learning_rate/qwen_vl_interface': 9.993825424432385e-06, 'learning_rate/action_model': 9.993825424432385e-05, 'epoch': 0.08}) +[2026-06-17 16:42:12,174][starVLA.training.train_starvla][INFO] - Step 213, Loss: {'train/loss': 1.2018779516220093, 'train/grad_norm_pre_clip': 7.319307804107666, 'timing/data': 0.012742164544761181, 'timing/model': 1.061423464678228, 'learning_rate/qwen_vl_interface': 9.99371469635244e-06, 'learning_rate/action_model': 9.99371469635244e-05, 'epoch': 0.08}) +[2026-06-17 16:42:16,847][starVLA.training.train_starvla][INFO] - Step 214, Loss: {'train/loss': 0.9100761413574219, 'train/grad_norm_pre_clip': 4.5177507400512695, 'timing/data': 0.010936600156128407, 'timing/model': 1.1199242509901524, 'learning_rate/qwen_vl_interface': 9.993602984895336e-06, 'learning_rate/action_model': 9.993602984895335e-05, 'epoch': 0.08}) +[2026-06-17 16:42:21,511][starVLA.training.train_starvla][INFO] - Step 215, Loss: {'train/loss': 0.7572615146636963, 'train/grad_norm_pre_clip': 3.5402603149414062, 'timing/data': 0.014371002092957497, 'timing/model': 1.0646602399647236, 'learning_rate/qwen_vl_interface': 9.993490290084225e-06, 'learning_rate/action_model': 9.993490290084225e-05, 'epoch': 0.08}) +[2026-06-17 16:42:26,011][starVLA.training.train_starvla][INFO] - Step 216, Loss: {'train/loss': 0.8191946148872375, 'train/grad_norm_pre_clip': 3.573939323425293, 'timing/data': 0.011317480355501175, 'timing/model': 1.3974915659055114, 'learning_rate/qwen_vl_interface': 9.993376611942477e-06, 'learning_rate/action_model': 9.993376611942477e-05, 'epoch': 0.08}) +[2026-06-17 16:42:31,991][starVLA.training.train_starvla][INFO] - Step 217, Loss: {'train/loss': 1.2048895359039307, 'train/grad_norm_pre_clip': 4.857536315917969, 'timing/data': 0.010337751358747482, 'timing/model': 1.486693429760635, 'learning_rate/qwen_vl_interface': 9.99326195049365e-06, 'learning_rate/action_model': 9.99326195049365e-05, 'epoch': 0.08}) +[2026-06-17 16:42:38,405][starVLA.training.train_starvla][INFO] - Step 218, Loss: {'train/loss': 0.731050431728363, 'train/grad_norm_pre_clip': 3.7427687644958496, 'timing/data': 0.021136106923222542, 'timing/model': 1.4973217230290174, 'learning_rate/qwen_vl_interface': 9.99314630576152e-06, 'learning_rate/action_model': 9.993146305761519e-05, 'epoch': 0.08}) +[2026-06-17 16:42:42,987][starVLA.training.train_starvla][INFO] - Step 219, Loss: {'train/loss': 1.0241883993148804, 'train/grad_norm_pre_clip': 5.236337184906006, 'timing/data': 0.014640746638178825, 'timing/model': 1.072997865267098, 'learning_rate/qwen_vl_interface': 9.993029677770055e-06, 'learning_rate/action_model': 9.993029677770055e-05, 'epoch': 0.08}) +[2026-06-17 16:42:47,678][starVLA.training.train_starvla][INFO] - Step 220, Loss: {'train/loss': 0.8332566022872925, 'train/grad_norm_pre_clip': 7.821256637573242, 'timing/data': 0.011627396568655968, 'timing/model': 1.1353667508810759, 'learning_rate/qwen_vl_interface': 9.992912066543433e-06, 'learning_rate/action_model': 9.992912066543433e-05, 'epoch': 0.08}) +[2026-06-17 16:42:52,309][starVLA.training.train_starvla][INFO] - Step 221, Loss: {'train/loss': 0.7658007740974426, 'train/grad_norm_pre_clip': 2.922722816467285, 'timing/data': 0.011650275439023972, 'timing/model': 1.142976745031774, 'learning_rate/qwen_vl_interface': 9.992793472106039e-06, 'learning_rate/action_model': 9.992793472106038e-05, 'epoch': 0.09}) +[2026-06-17 16:42:56,982][starVLA.training.train_starvla][INFO] - Step 222, Loss: {'train/loss': 0.6616278290748596, 'train/grad_norm_pre_clip': 3.5350756645202637, 'timing/data': 0.011426953598856926, 'timing/model': 1.4485406056046486, 'learning_rate/qwen_vl_interface': 9.992673894482452e-06, 'learning_rate/action_model': 9.992673894482451e-05, 'epoch': 0.09}) +[2026-06-17 16:43:01,238][starVLA.training.train_starvla][INFO] - Step 223, Loss: {'train/loss': 0.8024823665618896, 'train/grad_norm_pre_clip': 5.845354080200195, 'timing/data': 0.012137428857386112, 'timing/model': 1.0671935584396124, 'learning_rate/qwen_vl_interface': 9.992553333697464e-06, 'learning_rate/action_model': 9.992553333697463e-05, 'epoch': 0.09}) +[2026-06-17 16:43:07,476][starVLA.training.train_starvla][INFO] - Step 224, Loss: {'train/loss': 0.8213391304016113, 'train/grad_norm_pre_clip': 4.327073574066162, 'timing/data': 0.011337460950016975, 'timing/model': 2.3167821634560823, 'learning_rate/qwen_vl_interface': 9.992431789776067e-06, 'learning_rate/action_model': 9.992431789776066e-05, 'epoch': 0.09}) +[2026-06-17 16:43:12,200][starVLA.training.train_starvla][INFO] - Step 225, Loss: {'train/loss': 1.0695656538009644, 'train/grad_norm_pre_clip': 5.281611919403076, 'timing/data': 0.011721075512468815, 'timing/model': 1.2141786348074675, 'learning_rate/qwen_vl_interface': 9.992309262743458e-06, 'learning_rate/action_model': 9.992309262743457e-05, 'epoch': 0.09}) +[2026-06-17 16:43:16,864][starVLA.training.train_starvla][INFO] - Step 226, Loss: {'train/loss': 0.9779719710350037, 'train/grad_norm_pre_clip': 3.8130874633789062, 'timing/data': 0.011698145419359207, 'timing/model': 1.0742827439680696, 'learning_rate/qwen_vl_interface': 9.992185752625035e-06, 'learning_rate/action_model': 9.992185752625033e-05, 'epoch': 0.09}) +[2026-06-17 16:43:21,411][starVLA.training.train_starvla][INFO] - Step 227, Loss: {'train/loss': 0.7977818846702576, 'train/grad_norm_pre_clip': 3.195996046066284, 'timing/data': 0.01374484971165657, 'timing/model': 1.4201367553323507, 'learning_rate/qwen_vl_interface': 9.992061259446402e-06, 'learning_rate/action_model': 9.992061259446401e-05, 'epoch': 0.09}) +[2026-06-17 16:43:25,580][starVLA.training.train_starvla][INFO] - Step 228, Loss: {'train/loss': 0.9249550700187683, 'train/grad_norm_pre_clip': 3.7360455989837646, 'timing/data': 0.011833916418254375, 'timing/model': 1.0572948958724737, 'learning_rate/qwen_vl_interface': 9.991935783233368e-06, 'learning_rate/action_model': 9.991935783233367e-05, 'epoch': 0.09}) +[2026-06-17 16:43:30,224][starVLA.training.train_starvla][INFO] - Step 229, Loss: {'train/loss': 0.7686160802841187, 'train/grad_norm_pre_clip': 4.998872756958008, 'timing/data': 0.011389427818357944, 'timing/model': 1.1756592588499188, 'learning_rate/qwen_vl_interface': 9.991809324011943e-06, 'learning_rate/action_model': 9.991809324011943e-05, 'epoch': 0.09}) +[2026-06-17 16:43:34,692][starVLA.training.train_starvla][INFO] - Step 230, Loss: {'train/loss': 1.1487466096878052, 'train/grad_norm_pre_clip': 3.6068997383117676, 'timing/data': 0.012111769989132881, 'timing/model': 1.0765345161780715, 'learning_rate/qwen_vl_interface': 9.991681881808343e-06, 'learning_rate/action_model': 9.991681881808343e-05, 'epoch': 0.09}) +[2026-06-17 16:43:39,324][starVLA.training.train_starvla][INFO] - Step 231, Loss: {'train/loss': 1.239833116531372, 'train/grad_norm_pre_clip': 6.634476184844971, 'timing/data': 0.011685713194310665, 'timing/model': 1.0730776190757751, 'learning_rate/qwen_vl_interface': 9.991553456648986e-06, 'learning_rate/action_model': 9.991553456648985e-05, 'epoch': 0.09}) +[2026-06-17 16:43:44,145][starVLA.training.train_starvla][INFO] - Step 232, Loss: {'train/loss': 0.8394277691841125, 'train/grad_norm_pre_clip': 5.648953914642334, 'timing/data': 0.014582201838493347, 'timing/model': 1.5153760006651282, 'learning_rate/qwen_vl_interface': 9.991424048560497e-06, 'learning_rate/action_model': 9.991424048560496e-05, 'epoch': 0.09}) +[2026-06-17 16:43:49,403][starVLA.training.train_starvla][INFO] - Step 233, Loss: {'train/loss': 1.3875350952148438, 'train/grad_norm_pre_clip': 10.28446102142334, 'timing/data': 0.010856952518224716, 'timing/model': 1.2290886072441936, 'learning_rate/qwen_vl_interface': 9.991293657569698e-06, 'learning_rate/action_model': 9.991293657569698e-05, 'epoch': 0.09}) +[2026-06-17 16:43:55,616][starVLA.training.train_starvla][INFO] - Step 234, Loss: {'train/loss': 0.973518431186676, 'train/grad_norm_pre_clip': 6.78671407699585, 'timing/data': 0.018213381059467793, 'timing/model': 1.0959274182096124, 'learning_rate/qwen_vl_interface': 9.991162283703624e-06, 'learning_rate/action_model': 9.991162283703624e-05, 'epoch': 0.09}) +[2026-06-17 16:44:00,241][starVLA.training.train_starvla][INFO] - Step 235, Loss: {'train/loss': 0.66424959897995, 'train/grad_norm_pre_clip': 5.481321334838867, 'timing/data': 0.012695081532001495, 'timing/model': 1.1400568019598722, 'learning_rate/qwen_vl_interface': 9.991029926989506e-06, 'learning_rate/action_model': 9.991029926989506e-05, 'epoch': 0.09}) +[2026-06-17 16:44:04,908][starVLA.training.train_starvla][INFO] - Step 236, Loss: {'train/loss': 1.1107572317123413, 'train/grad_norm_pre_clip': 6.046741962432861, 'timing/data': 0.011130144819617271, 'timing/model': 1.0761752212420106, 'learning_rate/qwen_vl_interface': 9.990896587454783e-06, 'learning_rate/action_model': 9.990896587454783e-05, 'epoch': 0.09}) +[2026-06-17 16:44:11,016][starVLA.training.train_starvla][INFO] - Step 237, Loss: {'train/loss': 1.2538447380065918, 'train/grad_norm_pre_clip': 8.993484497070312, 'timing/data': 0.01300359982997179, 'timing/model': 1.4990773163735867, 'learning_rate/qwen_vl_interface': 9.990762265127096e-06, 'learning_rate/action_model': 9.990762265127096e-05, 'epoch': 0.09}) +[2026-06-17 16:44:15,268][starVLA.training.train_starvla][INFO] - Step 238, Loss: {'train/loss': 1.1124573945999146, 'train/grad_norm_pre_clip': 6.034200668334961, 'timing/data': 0.011812910437583923, 'timing/model': 1.0712086707353592, 'learning_rate/qwen_vl_interface': 9.990626960034291e-06, 'learning_rate/action_model': 9.990626960034291e-05, 'epoch': 0.09}) +[2026-06-17 16:44:20,024][starVLA.training.train_starvla][INFO] - Step 239, Loss: {'train/loss': 0.7729935050010681, 'train/grad_norm_pre_clip': 4.532880783081055, 'timing/data': 0.01345010381191969, 'timing/model': 1.1274828873574734, 'learning_rate/qwen_vl_interface': 9.990490672204417e-06, 'learning_rate/action_model': 9.990490672204415e-05, 'epoch': 0.09}) +[2026-06-17 16:44:24,582][starVLA.training.train_starvla][INFO] - Step 240, Loss: {'train/loss': 0.8113449215888977, 'train/grad_norm_pre_clip': 4.217892169952393, 'timing/data': 0.0140584921464324, 'timing/model': 1.0596117042005062, 'learning_rate/qwen_vl_interface': 9.99035340166572e-06, 'learning_rate/action_model': 9.990353401665721e-05, 'epoch': 0.09}) +[2026-06-17 16:44:30,345][starVLA.training.train_starvla][INFO] - Step 241, Loss: {'train/loss': 0.7949649095535278, 'train/grad_norm_pre_clip': 8.617364883422852, 'timing/data': 0.01757956203073263, 'timing/model': 1.1613645181059837, 'learning_rate/qwen_vl_interface': 9.990215148446667e-06, 'learning_rate/action_model': 9.990215148446666e-05, 'epoch': 0.09}) +[2026-06-17 16:44:35,052][starVLA.training.train_starvla][INFO] - Step 242, Loss: {'train/loss': 0.9243668913841248, 'train/grad_norm_pre_clip': 5.361505031585693, 'timing/data': 0.012604928575456142, 'timing/model': 1.517158300615847, 'learning_rate/qwen_vl_interface': 9.990075912575912e-06, 'learning_rate/action_model': 9.990075912575911e-05, 'epoch': 0.09}) +[2026-06-17 16:44:39,437][starVLA.training.train_starvla][INFO] - Step 243, Loss: {'train/loss': 0.8130911588668823, 'train/grad_norm_pre_clip': 14.302745819091797, 'timing/data': 0.013797447085380554, 'timing/model': 1.1628538547083735, 'learning_rate/qwen_vl_interface': 9.98993569408232e-06, 'learning_rate/action_model': 9.989935694082321e-05, 'epoch': 0.09}) +[2026-06-17 16:44:43,993][starVLA.training.train_starvla][INFO] - Step 244, Loss: {'train/loss': 0.8650860786437988, 'train/grad_norm_pre_clip': 4.462236404418945, 'timing/data': 0.011095959693193436, 'timing/model': 1.0752369817346334, 'learning_rate/qwen_vl_interface': 9.989794492994958e-06, 'learning_rate/action_model': 9.989794492994958e-05, 'epoch': 0.09}) +[2026-06-17 16:44:49,792][starVLA.training.train_starvla][INFO] - Step 245, Loss: {'train/loss': 0.7920752167701721, 'train/grad_norm_pre_clip': 4.861967086791992, 'timing/data': 0.0123404236510396, 'timing/model': 1.1134876878932118, 'learning_rate/qwen_vl_interface': 9.989652309343097e-06, 'learning_rate/action_model': 9.989652309343098e-05, 'epoch': 0.09}) +[2026-06-17 16:44:54,446][starVLA.training.train_starvla][INFO] - Step 246, Loss: {'train/loss': 1.028282880783081, 'train/grad_norm_pre_clip': 4.8851847648620605, 'timing/data': 0.016519557684659958, 'timing/model': 1.0789150670170784, 'learning_rate/qwen_vl_interface': 9.989509143156213e-06, 'learning_rate/action_model': 9.989509143156213e-05, 'epoch': 0.09}) +[2026-06-17 16:44:59,172][starVLA.training.train_starvla][INFO] - Step 247, Loss: {'train/loss': 1.0210739374160767, 'train/grad_norm_pre_clip': 5.268438816070557, 'timing/data': 0.017478386871516705, 'timing/model': 1.4827697351574898, 'learning_rate/qwen_vl_interface': 9.989364994463983e-06, 'learning_rate/action_model': 9.989364994463983e-05, 'epoch': 0.1}) +[2026-06-17 16:45:03,349][starVLA.training.train_starvla][INFO] - Step 248, Loss: {'train/loss': 1.2725423574447632, 'train/grad_norm_pre_clip': 5.098412036895752, 'timing/data': 0.010421982035040855, 'timing/model': 1.0636678356677294, 'learning_rate/qwen_vl_interface': 9.989219863296291e-06, 'learning_rate/action_model': 9.989219863296291e-05, 'epoch': 0.1}) +[2026-06-17 16:45:08,025][starVLA.training.train_starvla][INFO] - Step 249, Loss: {'train/loss': 1.0668554306030273, 'train/grad_norm_pre_clip': 5.255367279052734, 'timing/data': 0.020414240658283234, 'timing/model': 1.0954609969630837, 'learning_rate/qwen_vl_interface': 9.98907374968322e-06, 'learning_rate/action_model': 9.989073749683219e-05, 'epoch': 0.1}) +[2026-06-17 16:45:12,755][starVLA.training.train_starvla][INFO] - Step 250, Loss: {'train/loss': 1.0262621641159058, 'train/grad_norm_pre_clip': 6.677448749542236, 'timing/data': 0.017298413440585136, 'timing/model': 1.169144719839096, 'learning_rate/qwen_vl_interface': 9.988926653655062e-06, 'learning_rate/action_model': 9.988926653655062e-05, 'epoch': 0.1}) +[2026-06-17 16:45:17,330][starVLA.training.train_starvla][INFO] - Step 251, Loss: {'train/loss': 1.011883020401001, 'train/grad_norm_pre_clip': 5.279397487640381, 'timing/data': 0.015458343550562859, 'timing/model': 1.1321386182680726, 'learning_rate/qwen_vl_interface': 9.988778575242309e-06, 'learning_rate/action_model': 9.988778575242309e-05, 'epoch': 0.1}) +[2026-06-17 16:45:21,918][starVLA.training.train_starvla][INFO] - Step 252, Loss: {'train/loss': 1.0201189517974854, 'train/grad_norm_pre_clip': 4.864012718200684, 'timing/data': 0.011164096184074879, 'timing/model': 1.4573062993586063, 'learning_rate/qwen_vl_interface': 9.98862951447566e-06, 'learning_rate/action_model': 9.988629514475659e-05, 'epoch': 0.1}) +[2026-06-17 16:45:26,135][starVLA.training.train_starvla][INFO] - Step 253, Loss: {'train/loss': 0.6953170895576477, 'train/grad_norm_pre_clip': 5.440064430236816, 'timing/data': 0.012242080643773079, 'timing/model': 1.0792247671633959, 'learning_rate/qwen_vl_interface': 9.98847947138601e-06, 'learning_rate/action_model': 9.98847947138601e-05, 'epoch': 0.1}) +[2026-06-17 16:45:30,746][starVLA.training.train_starvla][INFO] - Step 254, Loss: {'train/loss': 0.7794857025146484, 'train/grad_norm_pre_clip': 2.4033291339874268, 'timing/data': 0.010647401213645935, 'timing/model': 1.167778443545103, 'learning_rate/qwen_vl_interface': 9.98832844600447e-06, 'learning_rate/action_model': 9.98832844600447e-05, 'epoch': 0.1}) +[2026-06-17 16:45:35,602][starVLA.training.train_starvla][INFO] - Step 255, Loss: {'train/loss': 0.9377986192703247, 'train/grad_norm_pre_clip': 3.883505344390869, 'timing/data': 0.012984732165932655, 'timing/model': 1.1395549783483148, 'learning_rate/qwen_vl_interface': 9.988176438362344e-06, 'learning_rate/action_model': 9.988176438362343e-05, 'epoch': 0.1}) +[2026-06-17 16:45:40,739][starVLA.training.train_starvla][INFO] - Step 256, Loss: {'train/loss': 0.728362500667572, 'train/grad_norm_pre_clip': 3.123703956604004, 'timing/data': 0.012323951348662376, 'timing/model': 1.1092673614621162, 'learning_rate/qwen_vl_interface': 9.988023448491143e-06, 'learning_rate/action_model': 9.988023448491143e-05, 'epoch': 0.1}) +[2026-06-17 16:45:45,275][starVLA.training.train_starvla][INFO] - Step 257, Loss: {'train/loss': 0.9175889492034912, 'train/grad_norm_pre_clip': 5.060039043426514, 'timing/data': 0.0134171973913908, 'timing/model': 1.4041837751865387, 'learning_rate/qwen_vl_interface': 9.987869476422582e-06, 'learning_rate/action_model': 9.987869476422583e-05, 'epoch': 0.1}) +[2026-06-17 16:45:49,621][starVLA.training.train_starvla][INFO] - Step 258, Loss: {'train/loss': 1.1463993787765503, 'train/grad_norm_pre_clip': 3.6256937980651855, 'timing/data': 0.011595504358410835, 'timing/model': 1.0843082023784518, 'learning_rate/qwen_vl_interface': 9.987714522188583e-06, 'learning_rate/action_model': 9.987714522188582e-05, 'epoch': 0.1}) +[2026-06-17 16:45:54,270][starVLA.training.train_starvla][INFO] - Step 259, Loss: {'train/loss': 0.7391652464866638, 'train/grad_norm_pre_clip': 3.0799920558929443, 'timing/data': 0.013005084358155727, 'timing/model': 1.0713782859966159, 'learning_rate/qwen_vl_interface': 9.987558585821264e-06, 'learning_rate/action_model': 9.987558585821263e-05, 'epoch': 0.1}) +[2026-06-17 16:45:58,951][starVLA.training.train_starvla][INFO] - Step 260, Loss: {'train/loss': 0.9433417916297913, 'train/grad_norm_pre_clip': 6.328248023986816, 'timing/data': 0.013175326399505138, 'timing/model': 1.0730199562385678, 'learning_rate/qwen_vl_interface': 9.987401667352952e-06, 'learning_rate/action_model': 9.987401667352951e-05, 'epoch': 0.1}) +[2026-06-17 16:46:04,050][starVLA.training.train_starvla][INFO] - Step 261, Loss: {'train/loss': 0.9246513843536377, 'train/grad_norm_pre_clip': 5.013181686401367, 'timing/data': 0.01196098979562521, 'timing/model': 1.1163387335836887, 'learning_rate/qwen_vl_interface': 9.987243766816175e-06, 'learning_rate/action_model': 9.987243766816176e-05, 'epoch': 0.1}) +[2026-06-17 16:46:08,753][starVLA.training.train_starvla][INFO] - Step 262, Loss: {'train/loss': 1.0154762268066406, 'train/grad_norm_pre_clip': 6.097379684448242, 'timing/data': 0.01473554503172636, 'timing/model': 1.4642269993200898, 'learning_rate/qwen_vl_interface': 9.987084884243671e-06, 'learning_rate/action_model': 9.987084884243672e-05, 'epoch': 0.1}) +[2026-06-17 16:46:13,055][starVLA.training.train_starvla][INFO] - Step 263, Loss: {'train/loss': 1.0952621698379517, 'train/grad_norm_pre_clip': 4.787456512451172, 'timing/data': 0.01326052751392126, 'timing/model': 1.0924996957182884, 'learning_rate/qwen_vl_interface': 9.986925019668374e-06, 'learning_rate/action_model': 9.986925019668373e-05, 'epoch': 0.1}) +[2026-06-17 16:46:17,861][starVLA.training.train_starvla][INFO] - Step 264, Loss: {'train/loss': 0.9906283020973206, 'train/grad_norm_pre_clip': 7.073967456817627, 'timing/data': 0.016738396137952805, 'timing/model': 1.0980123141780496, 'learning_rate/qwen_vl_interface': 9.98676417312342e-06, 'learning_rate/action_model': 9.986764173123419e-05, 'epoch': 0.1}) +[2026-06-17 16:46:22,534][starVLA.training.train_starvla][INFO] - Step 265, Loss: {'train/loss': 0.8943383097648621, 'train/grad_norm_pre_clip': 6.758089542388916, 'timing/data': 0.012655885890126228, 'timing/model': 1.162138756364584, 'learning_rate/qwen_vl_interface': 9.986602344642157e-06, 'learning_rate/action_model': 9.986602344642156e-05, 'epoch': 0.1}) +[2026-06-17 16:46:28,895][starVLA.training.train_starvla][INFO] - Step 266, Loss: {'train/loss': 1.0444995164871216, 'train/grad_norm_pre_clip': 11.543886184692383, 'timing/data': 0.012931290082633495, 'timing/model': 2.827005097642541, 'learning_rate/qwen_vl_interface': 9.986439534258132e-06, 'learning_rate/action_model': 9.986439534258131e-05, 'epoch': 0.1}) +[2026-06-17 16:46:34,933][starVLA.training.train_starvla][INFO] - Step 267, Loss: {'train/loss': 1.0373914241790771, 'train/grad_norm_pre_clip': 6.235805988311768, 'timing/data': 0.02181662619113922, 'timing/model': 1.481447852216661, 'learning_rate/qwen_vl_interface': 9.986275742005093e-06, 'learning_rate/action_model': 9.986275742005093e-05, 'epoch': 0.1}) +[2026-06-17 16:46:39,754][starVLA.training.train_starvla][INFO] - Step 268, Loss: {'train/loss': 0.8377564549446106, 'train/grad_norm_pre_clip': 5.29772424697876, 'timing/data': 0.01877487637102604, 'timing/model': 1.1474778847768903, 'learning_rate/qwen_vl_interface': 9.986110967916997e-06, 'learning_rate/action_model': 9.986110967916996e-05, 'epoch': 0.1}) +[2026-06-17 16:46:44,527][starVLA.training.train_starvla][INFO] - Step 269, Loss: {'train/loss': 0.6262129545211792, 'train/grad_norm_pre_clip': 5.644097805023193, 'timing/data': 0.010796531103551388, 'timing/model': 1.100184042006731, 'learning_rate/qwen_vl_interface': 9.985945212028002e-06, 'learning_rate/action_model': 9.985945212028001e-05, 'epoch': 0.1}) +[2026-06-17 16:46:49,376][starVLA.training.train_starvla][INFO] - Step 270, Loss: {'train/loss': 0.972694993019104, 'train/grad_norm_pre_clip': 7.288150310516357, 'timing/data': 0.011309108696877956, 'timing/model': 1.0705957412719727, 'learning_rate/qwen_vl_interface': 9.985778474372467e-06, 'learning_rate/action_model': 9.985778474372466e-05, 'epoch': 0.1}) +[2026-06-17 16:46:53,991][starVLA.training.train_starvla][INFO] - Step 271, Loss: {'train/loss': 0.8783066272735596, 'train/grad_norm_pre_clip': 3.8845667839050293, 'timing/data': 0.014150640927255154, 'timing/model': 1.069662880152464, 'learning_rate/qwen_vl_interface': 9.98561075498496e-06, 'learning_rate/action_model': 9.98561075498496e-05, 'epoch': 0.1}) +[2026-06-17 16:46:58,683][starVLA.training.train_starvla][INFO] - Step 272, Loss: {'train/loss': 0.6560519337654114, 'train/grad_norm_pre_clip': 2.9311933517456055, 'timing/data': 0.018206694163382053, 'timing/model': 1.4938425496220589, 'learning_rate/qwen_vl_interface': 9.985442053900246e-06, 'learning_rate/action_model': 9.985442053900246e-05, 'epoch': 0.1}) +[2026-06-17 16:47:02,898][starVLA.training.train_starvla][INFO] - Step 273, Loss: {'train/loss': 1.1323237419128418, 'train/grad_norm_pre_clip': 7.729162216186523, 'timing/data': 0.01197837758809328, 'timing/model': 1.0808716462925076, 'learning_rate/qwen_vl_interface': 9.9852723711533e-06, 'learning_rate/action_model': 9.9852723711533e-05, 'epoch': 0.11}) +[2026-06-17 16:47:07,382][starVLA.training.train_starvla][INFO] - Step 274, Loss: {'train/loss': 0.9869439005851746, 'train/grad_norm_pre_clip': 6.607097148895264, 'timing/data': 0.012444320134818554, 'timing/model': 1.0644782558083534, 'learning_rate/qwen_vl_interface': 9.985101706779297e-06, 'learning_rate/action_model': 9.985101706779296e-05, 'epoch': 0.11}) +[2026-06-17 16:47:11,960][starVLA.training.train_starvla][INFO] - Step 275, Loss: {'train/loss': 1.0493444204330444, 'train/grad_norm_pre_clip': 4.276364326477051, 'timing/data': 0.011257277801632881, 'timing/model': 1.0747316759079695, 'learning_rate/qwen_vl_interface': 9.984930060813613e-06, 'learning_rate/action_model': 9.984930060813613e-05, 'epoch': 0.11}) +[2026-06-17 16:47:16,527][starVLA.training.train_starvla][INFO] - Step 276, Loss: {'train/loss': 0.7458998560905457, 'train/grad_norm_pre_clip': 4.424931526184082, 'timing/data': 0.012622367590665817, 'timing/model': 1.1076676659286022, 'learning_rate/qwen_vl_interface': 9.984757433291832e-06, 'learning_rate/action_model': 9.984757433291832e-05, 'epoch': 0.11}) +[2026-06-17 16:47:21,196][starVLA.training.train_starvla][INFO] - Step 277, Loss: {'train/loss': 0.8634069561958313, 'train/grad_norm_pre_clip': 5.12622594833374, 'timing/data': 0.0140380235388875, 'timing/model': 1.4669728083536029, 'learning_rate/qwen_vl_interface': 9.984583824249742e-06, 'learning_rate/action_model': 9.984583824249742e-05, 'epoch': 0.11}) +[2026-06-17 16:47:25,440][starVLA.training.train_starvla][INFO] - Step 278, Loss: {'train/loss': 1.0064998865127563, 'train/grad_norm_pre_clip': 4.149361610412598, 'timing/data': 0.012414061464369297, 'timing/model': 1.0939340041950345, 'learning_rate/qwen_vl_interface': 9.98440923372333e-06, 'learning_rate/action_model': 9.984409233723328e-05, 'epoch': 0.11}) +[2026-06-17 16:47:29,908][starVLA.training.train_starvla][INFO] - Step 279, Loss: {'train/loss': 0.6010297536849976, 'train/grad_norm_pre_clip': 4.633713722229004, 'timing/data': 0.010682781226933002, 'timing/model': 1.0722127379849553, 'learning_rate/qwen_vl_interface': 9.984233661748788e-06, 'learning_rate/action_model': 9.984233661748788e-05, 'epoch': 0.11}) +[2026-06-17 16:47:34,373][starVLA.training.train_starvla][INFO] - Step 280, Loss: {'train/loss': 0.8375999331474304, 'train/grad_norm_pre_clip': 6.731870651245117, 'timing/data': 0.01010813470929861, 'timing/model': 1.0644578346982598, 'learning_rate/qwen_vl_interface': 9.984057108362516e-06, 'learning_rate/action_model': 9.984057108362515e-05, 'epoch': 0.11}) +[2026-06-17 16:47:39,317][starVLA.training.train_starvla][INFO] - Step 281, Loss: {'train/loss': 1.0493531227111816, 'train/grad_norm_pre_clip': 11.105584144592285, 'timing/data': 0.013146176002919674, 'timing/model': 1.1622019251808524, 'learning_rate/qwen_vl_interface': 9.98387957360111e-06, 'learning_rate/action_model': 9.98387957360111e-05, 'epoch': 0.11}) +[2026-06-17 16:47:43,834][starVLA.training.train_starvla][INFO] - Step 282, Loss: {'train/loss': 0.8419176936149597, 'train/grad_norm_pre_clip': 5.332195281982422, 'timing/data': 0.010786411352455616, 'timing/model': 1.4169385321438313, 'learning_rate/qwen_vl_interface': 9.983701057501375e-06, 'learning_rate/action_model': 9.983701057501375e-05, 'epoch': 0.11}) +[2026-06-17 16:47:48,090][starVLA.training.train_starvla][INFO] - Step 283, Loss: {'train/loss': 1.0029164552688599, 'train/grad_norm_pre_clip': 7.215226173400879, 'timing/data': 0.010962137021124363, 'timing/model': 1.0712140537798405, 'learning_rate/qwen_vl_interface': 9.983521560100317e-06, 'learning_rate/action_model': 9.983521560100316e-05, 'epoch': 0.11}) +[2026-06-17 16:47:52,705][starVLA.training.train_starvla][INFO] - Step 284, Loss: {'train/loss': 1.0877110958099365, 'train/grad_norm_pre_clip': 7.221944332122803, 'timing/data': 0.011439557187259197, 'timing/model': 1.0719626052305102, 'learning_rate/qwen_vl_interface': 9.983341081435145e-06, 'learning_rate/action_model': 9.983341081435145e-05, 'epoch': 0.11}) +[2026-06-17 16:47:57,472][starVLA.training.train_starvla][INFO] - Step 285, Loss: {'train/loss': 1.0934162139892578, 'train/grad_norm_pre_clip': 5.941559791564941, 'timing/data': 0.015961583703756332, 'timing/model': 1.1162620829418302, 'learning_rate/qwen_vl_interface': 9.983159621543275e-06, 'learning_rate/action_model': 9.983159621543275e-05, 'epoch': 0.11}) +[2026-06-17 16:48:02,154][starVLA.training.train_starvla][INFO] - Step 286, Loss: {'train/loss': 0.7517309188842773, 'train/grad_norm_pre_clip': 6.787220478057861, 'timing/data': 0.010523217730224133, 'timing/model': 1.1090459898114204, 'learning_rate/qwen_vl_interface': 9.982977180462324e-06, 'learning_rate/action_model': 9.982977180462323e-05, 'epoch': 0.11}) +[2026-06-17 16:48:07,339][starVLA.training.train_starvla][INFO] - Step 287, Loss: {'train/loss': 0.7551657557487488, 'train/grad_norm_pre_clip': 11.473045349121094, 'timing/data': 0.013089356943964958, 'timing/model': 1.4763352712616324, 'learning_rate/qwen_vl_interface': 9.982793758230107e-06, 'learning_rate/action_model': 9.982793758230108e-05, 'epoch': 0.11}) +[2026-06-17 16:48:13,275][starVLA.training.train_starvla][INFO] - Step 288, Loss: {'train/loss': 0.8195544481277466, 'train/grad_norm_pre_clip': 4.204925537109375, 'timing/data': 0.010839535854756832, 'timing/model': 1.0963920122012496, 'learning_rate/qwen_vl_interface': 9.982609354884653e-06, 'learning_rate/action_model': 9.982609354884653e-05, 'epoch': 0.11}) +[2026-06-17 16:48:19,368][starVLA.training.train_starvla][INFO] - Step 289, Loss: {'train/loss': 0.8176867961883545, 'train/grad_norm_pre_clip': 5.7809648513793945, 'timing/data': 0.011730711907148361, 'timing/model': 1.07909897249192, 'learning_rate/qwen_vl_interface': 9.982423970464187e-06, 'learning_rate/action_model': 9.982423970464186e-05, 'epoch': 0.11}) +[2026-06-17 16:48:24,625][starVLA.training.train_starvla][INFO] - Step 290, Loss: {'train/loss': 0.8891103267669678, 'train/grad_norm_pre_clip': 8.13759994506836, 'timing/data': 0.010125095024704933, 'timing/model': 1.0851087169721723, 'learning_rate/qwen_vl_interface': 9.98223760500714e-06, 'learning_rate/action_model': 9.982237605007139e-05, 'epoch': 0.11}) +[2026-06-17 16:48:31,195][starVLA.training.train_starvla][INFO] - Step 291, Loss: {'train/loss': 0.7379264831542969, 'train/grad_norm_pre_clip': 4.634956359863281, 'timing/data': 0.01653748471289873, 'timing/model': 1.162055341526866, 'learning_rate/qwen_vl_interface': 9.982050258552146e-06, 'learning_rate/action_model': 9.982050258552145e-05, 'epoch': 0.11}) +[2026-06-17 16:48:36,372][starVLA.training.train_starvla][INFO] - Step 292, Loss: {'train/loss': 0.9107229709625244, 'train/grad_norm_pre_clip': 10.179591178894043, 'timing/data': 0.010192478075623512, 'timing/model': 1.3578386465087533, 'learning_rate/qwen_vl_interface': 9.981861931138039e-06, 'learning_rate/action_model': 9.981861931138039e-05, 'epoch': 0.11}) +[2026-06-17 16:48:43,862][starVLA.training.train_starvla][INFO] - Step 293, Loss: {'train/loss': 0.7938336730003357, 'train/grad_norm_pre_clip': 8.880834579467773, 'timing/data': 0.019829041324555874, 'timing/model': 2.159331959672272, 'learning_rate/qwen_vl_interface': 9.981672622803863e-06, 'learning_rate/action_model': 9.981672622803862e-05, 'epoch': 0.11}) +[2026-06-17 16:48:48,623][starVLA.training.train_starvla][INFO] - Step 294, Loss: {'train/loss': 0.8520776629447937, 'train/grad_norm_pre_clip': 6.333413124084473, 'timing/data': 0.010899120941758156, 'timing/model': 1.1968454718589783, 'learning_rate/qwen_vl_interface': 9.981482333588863e-06, 'learning_rate/action_model': 9.981482333588862e-05, 'epoch': 0.11}) +[2026-06-17 16:48:53,198][starVLA.training.train_starvla][INFO] - Step 295, Loss: {'train/loss': 1.0341380834579468, 'train/grad_norm_pre_clip': 6.588998317718506, 'timing/data': 0.01649581268429756, 'timing/model': 1.0733464658260345, 'learning_rate/qwen_vl_interface': 9.981291063532481e-06, 'learning_rate/action_model': 9.981291063532481e-05, 'epoch': 0.11}) +[2026-06-17 16:48:57,714][starVLA.training.train_starvla][INFO] - Step 296, Loss: {'train/loss': 0.9506987929344177, 'train/grad_norm_pre_clip': 5.541721820831299, 'timing/data': 0.012548843398690224, 'timing/model': 1.0719210105016828, 'learning_rate/qwen_vl_interface': 9.981098812674372e-06, 'learning_rate/action_model': 9.981098812674371e-05, 'epoch': 0.11}) +[2026-06-17 16:49:02,393][starVLA.training.train_starvla][INFO] - Step 297, Loss: {'train/loss': 0.8760098218917847, 'train/grad_norm_pre_clip': 5.881033420562744, 'timing/data': 0.011831670068204403, 'timing/model': 1.4994159955531359, 'learning_rate/qwen_vl_interface': 9.980905581054388e-06, 'learning_rate/action_model': 9.980905581054387e-05, 'epoch': 0.11}) +[2026-06-17 16:49:06,779][starVLA.training.train_starvla][INFO] - Step 298, Loss: {'train/loss': 0.6844325065612793, 'train/grad_norm_pre_clip': 4.760311126708984, 'timing/data': 0.010179834440350533, 'timing/model': 1.0720821246504784, 'learning_rate/qwen_vl_interface': 9.980711368712586e-06, 'learning_rate/action_model': 9.980711368712585e-05, 'epoch': 0.11}) +[2026-06-17 16:49:11,559][starVLA.training.train_starvla][INFO] - Step 299, Loss: {'train/loss': 0.9195054769515991, 'train/grad_norm_pre_clip': 8.675580978393555, 'timing/data': 0.012639256194233894, 'timing/model': 1.0767465606331825, 'learning_rate/qwen_vl_interface': 9.980516175689228e-06, 'learning_rate/action_model': 9.980516175689227e-05, 'epoch': 0.12}) +[2026-06-17 16:51:44,810][starVLA.training.train_starvla][INFO] - Step 300, Loss: {'train/loss': 0.8305049538612366, 'train/grad_norm_pre_clip': 4.8435211181640625, 'eval/loss': 0.8482404351234436, 'eval/latency_4/loss': 0.8482404351234436, 'eval/demon_fire_precision': 0.9114860864416815, 'eval/demon_fire_recall': 1.0, 'eval/demon_fire_f1': 0.9536936657890662, 'eval/demon_fire_rate_model': 0.9920704845814978, 'eval/demon_fire_rate_teacher': 0.9042584434654919, 'eval/demon_left_precision': 1.0, 'eval/demon_left_recall': 0.00228310502283105, 'eval/demon_left_f1': 0.004555808656036446, 'eval/demon_left_rate_model': 0.0002936857562408223, 'eval/demon_left_rate_teacher': 0.12863436123348018, 'eval/demon_right_precision': 0.9894736842105263, 'eval/demon_right_recall': 0.16967509025270758, 'eval/demon_right_f1': 0.2896764252696456, 'eval/demon_right_rate_model': 0.027900146842878122, 'eval/demon_right_rate_teacher': 0.16270190895741557, 'eval/demon_move_f1': 0.14711611696284102, 'eval/demon_cc_f1': 0.3605950599374656, 'eval/latency_4/demon_fire_precision': 0.9114860864416815, 'eval/latency_4/demon_fire_recall': 1.0, 'eval/latency_4/demon_fire_f1': 0.9536936657890662, 'eval/latency_4/demon_fire_rate_model': 0.9920704845814978, 'eval/latency_4/demon_fire_rate_teacher': 0.9042584434654919, 'eval/latency_4/demon_left_precision': 1.0, 'eval/latency_4/demon_left_recall': 0.00228310502283105, 'eval/latency_4/demon_left_f1': 0.004555808656036446, 'eval/latency_4/demon_left_rate_model': 0.0002936857562408223, 'eval/latency_4/demon_left_rate_teacher': 0.12863436123348018, 'eval/latency_4/demon_right_precision': 0.9894736842105263, 'eval/latency_4/demon_right_recall': 0.16967509025270758, 'eval/latency_4/demon_right_f1': 0.2896764252696456, 'eval/latency_4/demon_right_rate_model': 0.027900146842878122, 'eval/latency_4/demon_right_rate_teacher': 0.16270190895741557, 'eval/latency_4/demon_move_f1': 0.14711611696284102, 'eval/latency_4/demon_cc_f1': 0.3605950599374656, 'timing/data': 0.01995895616710186, 'timing/model': 1.0799881936982274, 'learning_rate/qwen_vl_interface': 9.980320002024774e-06, 'learning_rate/action_model': 9.980320002024774e-05, 'epoch': 0.12}) +[2026-06-17 16:51:44,812][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_300_state +[2026-06-17 16:51:44,819][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 16:51:53,370][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_300_state/model.safetensors +[2026-06-17 16:52:15,251][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_300_state/optimizer.bin +[2026-06-17 16:52:15,254][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_300_state/sampler.bin +[2026-06-17 16:52:15,255][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_300_state/sampler_1.bin +[2026-06-17 16:52:15,258][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_300_state/random_states_0.pkl +[2026-06-17 16:52:15,263][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 300 +[2026-06-17 16:52:15,263][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 16:52:15,280][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 16:52:19,976][starVLA.training.train_starvla][INFO] - Step 301, Loss: {'train/loss': 0.7895995378494263, 'train/grad_norm_pre_clip': 13.721135139465332, 'timing/data': 0.017226618714630604, 'timing/model': 1.0824112044647336, 'learning_rate/qwen_vl_interface': 9.980122847759896e-06, 'learning_rate/action_model': 9.980122847759896e-05, 'epoch': 0.12}) +[2026-06-17 16:52:24,605][starVLA.training.train_starvla][INFO] - Step 302, Loss: {'train/loss': 0.9034211039543152, 'train/grad_norm_pre_clip': 7.2313666343688965, 'timing/data': 0.013886398635804653, 'timing/model': 1.4405780555680394, 'learning_rate/qwen_vl_interface': 9.97992471293546e-06, 'learning_rate/action_model': 9.97992471293546e-05, 'epoch': 0.12}) +[2026-06-17 16:52:29,014][starVLA.training.train_starvla][INFO] - Step 303, Loss: {'train/loss': 0.766185462474823, 'train/grad_norm_pre_clip': 6.260125637054443, 'timing/data': 0.012361791916191578, 'timing/model': 1.1378345619887114, 'learning_rate/qwen_vl_interface': 9.979725597592543e-06, 'learning_rate/action_model': 9.979725597592541e-05, 'epoch': 0.12}) +[2026-06-17 16:52:34,636][starVLA.training.train_starvla][INFO] - Step 304, Loss: {'train/loss': 0.762520968914032, 'train/grad_norm_pre_clip': 6.230747222900391, 'timing/data': 0.010653025470674038, 'timing/model': 1.117788327857852, 'learning_rate/qwen_vl_interface': 9.979525501772418e-06, 'learning_rate/action_model': 9.979525501772417e-05, 'epoch': 0.12}) +[2026-06-17 16:52:39,328][starVLA.training.train_starvla][INFO] - Step 305, Loss: {'train/loss': 0.9022511839866638, 'train/grad_norm_pre_clip': 8.207027435302734, 'timing/data': 0.014357728883624077, 'timing/model': 1.1294853705912828, 'learning_rate/qwen_vl_interface': 9.97932442551657e-06, 'learning_rate/action_model': 9.979324425516569e-05, 'epoch': 0.12}) +[2026-06-17 16:52:43,839][starVLA.training.train_starvla][INFO] - Step 306, Loss: {'train/loss': 0.8548976182937622, 'train/grad_norm_pre_clip': 10.482362747192383, 'timing/data': 0.010825756005942822, 'timing/model': 1.1008758805692196, 'learning_rate/qwen_vl_interface': 9.979122368866678e-06, 'learning_rate/action_model': 9.979122368866679e-05, 'epoch': 0.12}) +[2026-06-17 16:52:48,628][starVLA.training.train_starvla][INFO] - Step 307, Loss: {'train/loss': 1.1891822814941406, 'train/grad_norm_pre_clip': 13.263270378112793, 'timing/data': 0.018261989578604698, 'timing/model': 1.5721020633354783, 'learning_rate/qwen_vl_interface': 9.978919331864631e-06, 'learning_rate/action_model': 9.97891933186463e-05, 'epoch': 0.12}) +[2026-06-17 16:52:52,913][starVLA.training.train_starvla][INFO] - Step 308, Loss: {'train/loss': 1.142893671989441, 'train/grad_norm_pre_clip': 9.331328392028809, 'timing/data': 0.011432121507823467, 'timing/model': 1.1643093889579177, 'learning_rate/qwen_vl_interface': 9.978715314552518e-06, 'learning_rate/action_model': 9.978715314552518e-05, 'epoch': 0.12}) +[2026-06-17 16:52:57,544][starVLA.training.train_starvla][INFO] - Step 309, Loss: {'train/loss': 0.7172571420669556, 'train/grad_norm_pre_clip': 9.22152328491211, 'timing/data': 0.013867669738829136, 'timing/model': 1.0637707877904177, 'learning_rate/qwen_vl_interface': 9.978510316972631e-06, 'learning_rate/action_model': 9.978510316972631e-05, 'epoch': 0.12}) +[2026-06-17 16:53:02,111][starVLA.training.train_starvla][INFO] - Step 310, Loss: {'train/loss': 0.8084849119186401, 'train/grad_norm_pre_clip': 4.621730804443359, 'timing/data': 0.010530131869018078, 'timing/model': 1.0600049076601863, 'learning_rate/qwen_vl_interface': 9.978304339167467e-06, 'learning_rate/action_model': 9.978304339167467e-05, 'epoch': 0.12}) +[2026-06-17 16:53:06,580][starVLA.training.train_starvla][INFO] - Step 311, Loss: {'train/loss': 0.6199123859405518, 'train/grad_norm_pre_clip': 3.621506929397583, 'timing/data': 0.01246300432831049, 'timing/model': 1.0757208233699203, 'learning_rate/qwen_vl_interface': 9.978097381179726e-06, 'learning_rate/action_model': 9.978097381179726e-05, 'epoch': 0.12}) +[2026-06-17 16:53:11,045][starVLA.training.train_starvla][INFO] - Step 312, Loss: {'train/loss': 0.8415930867195129, 'train/grad_norm_pre_clip': 5.242180824279785, 'timing/data': 0.011351856403052807, 'timing/model': 1.1478552836924791, 'learning_rate/qwen_vl_interface': 9.977889443052312e-06, 'learning_rate/action_model': 9.977889443052311e-05, 'epoch': 0.12}) +[2026-06-17 16:53:15,633][starVLA.training.train_starvla][INFO] - Step 313, Loss: {'train/loss': 1.0761414766311646, 'train/grad_norm_pre_clip': 6.174630165100098, 'timing/data': 0.011238329112529755, 'timing/model': 1.0784167619422078, 'learning_rate/qwen_vl_interface': 9.977680524828327e-06, 'learning_rate/action_model': 9.977680524828327e-05, 'epoch': 0.12}) +[2026-06-17 16:53:20,259][starVLA.training.train_starvla][INFO] - Step 314, Loss: {'train/loss': 0.9943806529045105, 'train/grad_norm_pre_clip': 7.496921539306641, 'timing/data': 0.011497984640300274, 'timing/model': 1.1206220649182796, 'learning_rate/qwen_vl_interface': 9.977470626551083e-06, 'learning_rate/action_model': 9.977470626551083e-05, 'epoch': 0.12}) +[2026-06-17 16:53:24,872][starVLA.training.train_starvla][INFO] - Step 315, Loss: {'train/loss': 1.4388923645019531, 'train/grad_norm_pre_clip': 9.456914901733398, 'timing/data': 0.01598930172622204, 'timing/model': 1.071138635277748, 'learning_rate/qwen_vl_interface': 9.977259748264089e-06, 'learning_rate/action_model': 9.97725974826409e-05, 'epoch': 0.12}) +[2026-06-17 16:53:29,567][starVLA.training.train_starvla][INFO] - Step 316, Loss: {'train/loss': 1.2177631855010986, 'train/grad_norm_pre_clip': 7.512405872344971, 'timing/data': 0.01162655744701624, 'timing/model': 1.527927285991609, 'learning_rate/qwen_vl_interface': 9.977047890011065e-06, 'learning_rate/action_model': 9.977047890011065e-05, 'epoch': 0.12}) +[2026-06-17 16:53:34,125][starVLA.training.train_starvla][INFO] - Step 317, Loss: {'train/loss': 0.6661608815193176, 'train/grad_norm_pre_clip': 3.0795388221740723, 'timing/data': 0.014490289613604546, 'timing/model': 1.3469090331345797, 'learning_rate/qwen_vl_interface': 9.976835051835928e-06, 'learning_rate/action_model': 9.976835051835926e-05, 'epoch': 0.12}) +[2026-06-17 16:53:38,329][starVLA.training.train_starvla][INFO] - Step 318, Loss: {'train/loss': 0.7713433504104614, 'train/grad_norm_pre_clip': 4.502242565155029, 'timing/data': 0.012090318836271763, 'timing/model': 1.0803503999486566, 'learning_rate/qwen_vl_interface': 9.976621233782795e-06, 'learning_rate/action_model': 9.976621233782795e-05, 'epoch': 0.12}) +[2026-06-17 16:53:45,819][starVLA.training.train_starvla][INFO] - Step 319, Loss: {'train/loss': 0.952086329460144, 'train/grad_norm_pre_clip': 5.200545787811279, 'timing/data': 0.016425679437816143, 'timing/model': 2.645971448160708, 'learning_rate/qwen_vl_interface': 9.976406435895995e-06, 'learning_rate/action_model': 9.976406435895995e-05, 'epoch': 0.12}) +[2026-06-17 16:53:50,559][starVLA.training.train_starvla][INFO] - Step 320, Loss: {'train/loss': 0.8694404363632202, 'train/grad_norm_pre_clip': 6.583250045776367, 'timing/data': 0.012217297218739986, 'timing/model': 1.1011237343773246, 'learning_rate/qwen_vl_interface': 9.976190658220056e-06, 'learning_rate/action_model': 9.976190658220056e-05, 'epoch': 0.12}) +[2026-06-17 16:53:55,252][starVLA.training.train_starvla][INFO] - Step 321, Loss: {'train/loss': 0.7452970147132874, 'train/grad_norm_pre_clip': 5.125586986541748, 'timing/data': 0.011376782320439816, 'timing/model': 1.15792100969702, 'learning_rate/qwen_vl_interface': 9.975973900799709e-06, 'learning_rate/action_model': 9.975973900799708e-05, 'epoch': 0.12}) +[2026-06-17 16:53:59,882][starVLA.training.train_starvla][INFO] - Step 322, Loss: {'train/loss': 0.7403272986412048, 'train/grad_norm_pre_clip': 6.046120643615723, 'timing/data': 0.011759287677705288, 'timing/model': 1.4532211311161518, 'learning_rate/qwen_vl_interface': 9.975756163679884e-06, 'learning_rate/action_model': 9.975756163679885e-05, 'epoch': 0.12}) +[2026-06-17 16:54:04,072][starVLA.training.train_starvla][INFO] - Step 323, Loss: {'train/loss': 1.0064935684204102, 'train/grad_norm_pre_clip': 3.8146793842315674, 'timing/data': 0.011647740378975868, 'timing/model': 1.079276985488832, 'learning_rate/qwen_vl_interface': 9.975537446905722e-06, 'learning_rate/action_model': 9.975537446905723e-05, 'epoch': 0.12}) +[2026-06-17 16:54:08,638][starVLA.training.train_starvla][INFO] - Step 324, Loss: {'train/loss': 0.6640347838401794, 'train/grad_norm_pre_clip': 6.947854518890381, 'timing/data': 0.011964594945311546, 'timing/model': 1.0881120404228568, 'learning_rate/qwen_vl_interface': 9.975317750522565e-06, 'learning_rate/action_model': 9.975317750522564e-05, 'epoch': 0.12}) +[2026-06-17 16:54:13,346][starVLA.training.train_starvla][INFO] - Step 325, Loss: {'train/loss': 0.756123423576355, 'train/grad_norm_pre_clip': 9.679237365722656, 'timing/data': 0.012175827287137508, 'timing/model': 1.2409232705831528, 'learning_rate/qwen_vl_interface': 9.97509707457595e-06, 'learning_rate/action_model': 9.97509707457595e-05, 'epoch': 0.13}) +[2026-06-17 16:54:19,308][starVLA.training.train_starvla][INFO] - Step 326, Loss: {'train/loss': 0.9474503993988037, 'train/grad_norm_pre_clip': 8.630509376525879, 'timing/data': 0.011238866485655308, 'timing/model': 2.575614138506353, 'learning_rate/qwen_vl_interface': 9.974875419111631e-06, 'learning_rate/action_model': 9.97487541911163e-05, 'epoch': 0.13}) +[2026-06-17 16:54:23,939][starVLA.training.train_starvla][INFO] - Step 327, Loss: {'train/loss': 0.6700257658958435, 'train/grad_norm_pre_clip': 7.151867389678955, 'timing/data': 0.015320131555199623, 'timing/model': 1.4313000859692693, 'learning_rate/qwen_vl_interface': 9.974652784175552e-06, 'learning_rate/action_model': 9.974652784175551e-05, 'epoch': 0.13}) +[2026-06-17 16:54:29,666][starVLA.training.train_starvla][INFO] - Step 328, Loss: {'train/loss': 0.8567866086959839, 'train/grad_norm_pre_clip': 4.477929592132568, 'timing/data': 0.013163184747099876, 'timing/model': 1.2461180426180363, 'learning_rate/qwen_vl_interface': 9.974429169813865e-06, 'learning_rate/action_model': 9.974429169813864e-05, 'epoch': 0.13}) +[2026-06-17 16:54:34,351][starVLA.training.train_starvla][INFO] - Step 329, Loss: {'train/loss': 0.7237485647201538, 'train/grad_norm_pre_clip': 5.3863205909729, 'timing/data': 0.013168388977646828, 'timing/model': 1.0822978438809514, 'learning_rate/qwen_vl_interface': 9.974204576072928e-06, 'learning_rate/action_model': 9.974204576072928e-05, 'epoch': 0.13}) +[2026-06-17 16:54:39,062][starVLA.training.train_starvla][INFO] - Step 330, Loss: {'train/loss': 0.7571982741355896, 'train/grad_norm_pre_clip': 5.075780868530273, 'timing/data': 0.013649649918079376, 'timing/model': 1.1507255928590894, 'learning_rate/qwen_vl_interface': 9.973979002999299e-06, 'learning_rate/action_model': 9.973979002999299e-05, 'epoch': 0.13}) +[2026-06-17 16:54:43,673][starVLA.training.train_starvla][INFO] - Step 331, Loss: {'train/loss': 0.8567222356796265, 'train/grad_norm_pre_clip': 5.536384105682373, 'timing/data': 0.014053892344236374, 'timing/model': 1.0769683048129082, 'learning_rate/qwen_vl_interface': 9.97375245063974e-06, 'learning_rate/action_model': 9.97375245063974e-05, 'epoch': 0.13}) +[2026-06-17 16:54:48,212][starVLA.training.train_starvla][INFO] - Step 332, Loss: {'train/loss': 0.7878903746604919, 'train/grad_norm_pre_clip': 4.6470866203308105, 'timing/data': 0.011298560537397861, 'timing/model': 1.4073307141661644, 'learning_rate/qwen_vl_interface': 9.973524919041215e-06, 'learning_rate/action_model': 9.973524919041215e-05, 'epoch': 0.13}) +[2026-06-17 16:54:53,905][starVLA.training.train_starvla][INFO] - Step 333, Loss: {'train/loss': 0.544549822807312, 'train/grad_norm_pre_clip': 4.7606658935546875, 'timing/data': 0.0130341537296772, 'timing/model': 1.1193899232894182, 'learning_rate/qwen_vl_interface': 9.97329640825089e-06, 'learning_rate/action_model': 9.973296408250891e-05, 'epoch': 0.13}) +[2026-06-17 16:54:58,664][starVLA.training.train_starvla][INFO] - Step 334, Loss: {'train/loss': 0.7508956789970398, 'train/grad_norm_pre_clip': 7.207926273345947, 'timing/data': 0.011511529795825481, 'timing/model': 1.1775564011186361, 'learning_rate/qwen_vl_interface': 9.97306691831614e-06, 'learning_rate/action_model': 9.97306691831614e-05, 'epoch': 0.13}) +[2026-06-17 16:55:03,228][starVLA.training.train_starvla][INFO] - Step 335, Loss: {'train/loss': 0.7539361715316772, 'train/grad_norm_pre_clip': 6.311657905578613, 'timing/data': 0.012297223322093487, 'timing/model': 1.0786437187343836, 'learning_rate/qwen_vl_interface': 9.972836449284533e-06, 'learning_rate/action_model': 9.972836449284532e-05, 'epoch': 0.13}) +[2026-06-17 16:55:07,854][starVLA.training.train_starvla][INFO] - Step 336, Loss: {'train/loss': 0.713159441947937, 'train/grad_norm_pre_clip': 5.368681907653809, 'timing/data': 0.01211397536098957, 'timing/model': 1.126629495061934, 'learning_rate/qwen_vl_interface': 9.97260500120385e-06, 'learning_rate/action_model': 9.97260500120385e-05, 'epoch': 0.13}) +[2026-06-17 16:55:12,448][starVLA.training.train_starvla][INFO] - Step 337, Loss: {'train/loss': 0.5885767936706543, 'train/grad_norm_pre_clip': 3.8710367679595947, 'timing/data': 0.015305811539292336, 'timing/model': 1.4191420171409845, 'learning_rate/qwen_vl_interface': 9.972372574122067e-06, 'learning_rate/action_model': 9.972372574122068e-05, 'epoch': 0.13}) +[2026-06-17 16:55:16,729][starVLA.training.train_starvla][INFO] - Step 338, Loss: {'train/loss': 0.6315290927886963, 'train/grad_norm_pre_clip': 5.392354965209961, 'timing/data': 0.010641664266586304, 'timing/model': 1.1504617594182491, 'learning_rate/qwen_vl_interface': 9.972139168087368e-06, 'learning_rate/action_model': 9.972139168087369e-05, 'epoch': 0.13}) +[2026-06-17 16:55:21,269][starVLA.training.train_starvla][INFO] - Step 339, Loss: {'train/loss': 0.6812741756439209, 'train/grad_norm_pre_clip': 6.905471324920654, 'timing/data': 0.0148219158872962, 'timing/model': 1.0968002825975418, 'learning_rate/qwen_vl_interface': 9.97190478314814e-06, 'learning_rate/action_model': 9.97190478314814e-05, 'epoch': 0.13}) +[2026-06-17 16:55:26,177][starVLA.training.train_starvla][INFO] - Step 340, Loss: {'train/loss': 0.8542462587356567, 'train/grad_norm_pre_clip': 6.09728479385376, 'timing/data': 0.012134560383856297, 'timing/model': 1.0701499367132783, 'learning_rate/qwen_vl_interface': 9.971669419352968e-06, 'learning_rate/action_model': 9.971669419352966e-05, 'epoch': 0.13}) +[2026-06-17 16:55:30,666][starVLA.training.train_starvla][INFO] - Step 341, Loss: {'train/loss': 0.8387255668640137, 'train/grad_norm_pre_clip': 5.101222991943359, 'timing/data': 0.01062746625393629, 'timing/model': 1.0792242903262377, 'learning_rate/qwen_vl_interface': 9.971433076750644e-06, 'learning_rate/action_model': 9.971433076750643e-05, 'epoch': 0.13}) +[2026-06-17 16:55:35,219][starVLA.training.train_starvla][INFO] - Step 342, Loss: {'train/loss': 0.9713298082351685, 'train/grad_norm_pre_clip': 65.02928924560547, 'timing/data': 0.01593735720962286, 'timing/model': 1.4661504998803139, 'learning_rate/qwen_vl_interface': 9.971195755390166e-06, 'learning_rate/action_model': 9.971195755390164e-05, 'epoch': 0.13}) +[2026-06-17 16:55:39,431][starVLA.training.train_starvla][INFO] - Step 343, Loss: {'train/loss': 0.8337004780769348, 'train/grad_norm_pre_clip': 22.52045440673828, 'timing/data': 0.010257354937493801, 'timing/model': 1.0558616276830435, 'learning_rate/qwen_vl_interface': 9.970957455320723e-06, 'learning_rate/action_model': 9.970957455320723e-05, 'epoch': 0.13}) +[2026-06-17 16:55:44,035][starVLA.training.train_starvla][INFO] - Step 344, Loss: {'train/loss': 0.7293623685836792, 'train/grad_norm_pre_clip': 8.830704689025879, 'timing/data': 0.010254832915961742, 'timing/model': 1.0505464021116495, 'learning_rate/qwen_vl_interface': 9.970718176591723e-06, 'learning_rate/action_model': 9.970718176591722e-05, 'epoch': 0.13}) +[2026-06-17 16:55:48,258][starVLA.training.train_starvla][INFO] - Step 345, Loss: {'train/loss': 0.9263321757316589, 'train/grad_norm_pre_clip': 9.286686897277832, 'timing/data': 0.011067356914281845, 'timing/model': 1.0375372972339392, 'learning_rate/qwen_vl_interface': 9.970477919252763e-06, 'learning_rate/action_model': 9.970477919252763e-05, 'epoch': 0.13}) +[2026-06-17 16:55:52,410][starVLA.training.train_starvla][INFO] - Step 346, Loss: {'train/loss': 0.818823516368866, 'train/grad_norm_pre_clip': 5.814953804016113, 'timing/data': 0.009959185495972633, 'timing/model': 1.041317761875689, 'learning_rate/qwen_vl_interface': 9.970236683353652e-06, 'learning_rate/action_model': 9.970236683353652e-05, 'epoch': 0.13}) +[2026-06-17 16:55:56,673][starVLA.training.train_starvla][INFO] - Step 347, Loss: {'train/loss': 0.7197779417037964, 'train/grad_norm_pre_clip': 37.1962776184082, 'timing/data': 0.009095152840018272, 'timing/model': 1.3040861999616027, 'learning_rate/qwen_vl_interface': 9.969994468944399e-06, 'learning_rate/action_model': 9.969994468944398e-05, 'epoch': 0.13}) +[2026-06-17 16:56:00,710][starVLA.training.train_starvla][INFO] - Step 348, Loss: {'train/loss': 0.6597217917442322, 'train/grad_norm_pre_clip': 4.666164875030518, 'timing/data': 0.010632983408868313, 'timing/model': 1.0282148569822311, 'learning_rate/qwen_vl_interface': 9.969751276075212e-06, 'learning_rate/action_model': 9.969751276075212e-05, 'epoch': 0.13}) +[2026-06-17 16:56:04,984][starVLA.training.train_starvla][INFO] - Step 349, Loss: {'train/loss': 0.9374630451202393, 'train/grad_norm_pre_clip': 5.727560520172119, 'timing/data': 0.01017451286315918, 'timing/model': 1.0441156271845102, 'learning_rate/qwen_vl_interface': 9.969507104796506e-06, 'learning_rate/action_model': 9.969507104796505e-05, 'epoch': 0.13}) +[2026-06-17 16:56:09,277][starVLA.training.train_starvla][INFO] - Step 350, Loss: {'train/loss': 0.8240549564361572, 'train/grad_norm_pre_clip': 7.921024322509766, 'timing/data': 0.009056376293301582, 'timing/model': 1.0349657880142331, 'learning_rate/qwen_vl_interface': 9.969261955158898e-06, 'learning_rate/action_model': 9.969261955158898e-05, 'epoch': 0.13}) +[2026-06-17 16:56:13,496][starVLA.training.train_starvla][INFO] - Step 351, Loss: {'train/loss': 0.7377271056175232, 'train/grad_norm_pre_clip': 4.884880065917969, 'timing/data': 0.011852581053972244, 'timing/model': 1.0391843682155013, 'learning_rate/qwen_vl_interface': 9.96901582721321e-06, 'learning_rate/action_model': 9.96901582721321e-05, 'epoch': 0.14}) +[2026-06-17 16:56:17,650][starVLA.training.train_starvla][INFO] - Step 352, Loss: {'train/loss': 1.0135738849639893, 'train/grad_norm_pre_clip': 6.73974084854126, 'timing/data': 0.009150800295174122, 'timing/model': 1.192608779296279, 'learning_rate/qwen_vl_interface': 9.968768721010463e-06, 'learning_rate/action_model': 9.968768721010463e-05, 'epoch': 0.14}) +[2026-06-17 16:56:21,652][starVLA.training.train_starvla][INFO] - Step 353, Loss: {'train/loss': 1.0814318656921387, 'train/grad_norm_pre_clip': 8.142619132995605, 'timing/data': 0.009139182046055794, 'timing/model': 1.0303373793140054, 'learning_rate/qwen_vl_interface': 9.968520636601883e-06, 'learning_rate/action_model': 9.968520636601883e-05, 'epoch': 0.14}) +[2026-06-17 16:56:25,843][starVLA.training.train_starvla][INFO] - Step 354, Loss: {'train/loss': 0.7577674388885498, 'train/grad_norm_pre_clip': 5.018499851226807, 'timing/data': 0.009035429917275906, 'timing/model': 1.0434190519154072, 'learning_rate/qwen_vl_interface': 9.968271574038899e-06, 'learning_rate/action_model': 9.968271574038898e-05, 'epoch': 0.14}) +[2026-06-17 16:56:30,610][starVLA.training.train_starvla][INFO] - Step 355, Loss: {'train/loss': 0.9674438238143921, 'train/grad_norm_pre_clip': 7.53695821762085, 'timing/data': 0.010466942563652992, 'timing/model': 1.2303540110588074, 'learning_rate/qwen_vl_interface': 9.968021533373138e-06, 'learning_rate/action_model': 9.968021533373138e-05, 'epoch': 0.14}) +[2026-06-17 16:56:35,127][starVLA.training.train_starvla][INFO] - Step 356, Loss: {'train/loss': 0.8884931206703186, 'train/grad_norm_pre_clip': 4.100969314575195, 'timing/data': 0.009984452277421951, 'timing/model': 1.1552523896098137, 'learning_rate/qwen_vl_interface': 9.967770514656439e-06, 'learning_rate/action_model': 9.967770514656438e-05, 'epoch': 0.14}) +[2026-06-17 16:56:39,808][starVLA.training.train_starvla][INFO] - Step 357, Loss: {'train/loss': 0.6439478397369385, 'train/grad_norm_pre_clip': 7.304736614227295, 'timing/data': 0.01521286554634571, 'timing/model': 1.4463982647284865, 'learning_rate/qwen_vl_interface': 9.967518517940832e-06, 'learning_rate/action_model': 9.967518517940832e-05, 'epoch': 0.14}) +[2026-06-17 16:56:44,023][starVLA.training.train_starvla][INFO] - Step 358, Loss: {'train/loss': 0.6390354633331299, 'train/grad_norm_pre_clip': 6.1362128257751465, 'timing/data': 0.014874056912958622, 'timing/model': 1.1248003542423248, 'learning_rate/qwen_vl_interface': 9.967265543278563e-06, 'learning_rate/action_model': 9.967265543278562e-05, 'epoch': 0.14}) +[2026-06-17 16:56:48,604][starVLA.training.train_starvla][INFO] - Step 359, Loss: {'train/loss': 0.6717547178268433, 'train/grad_norm_pre_clip': 4.564572334289551, 'timing/data': 0.013168799690902233, 'timing/model': 1.1009263778105378, 'learning_rate/qwen_vl_interface': 9.96701159072207e-06, 'learning_rate/action_model': 9.96701159072207e-05, 'epoch': 0.14}) +[2026-06-17 16:56:53,408][starVLA.training.train_starvla][INFO] - Step 360, Loss: {'train/loss': 0.7121518850326538, 'train/grad_norm_pre_clip': 7.159807205200195, 'timing/data': 0.012961048632860184, 'timing/model': 1.1645574606955051, 'learning_rate/qwen_vl_interface': 9.966756660323998e-06, 'learning_rate/action_model': 9.966756660323999e-05, 'epoch': 0.14}) +[2026-06-17 16:56:59,233][starVLA.training.train_starvla][INFO] - Step 361, Loss: {'train/loss': 0.6588356494903564, 'train/grad_norm_pre_clip': 5.064120292663574, 'timing/data': 0.011180070228874683, 'timing/model': 1.0705772014334798, 'learning_rate/qwen_vl_interface': 9.966500752137195e-06, 'learning_rate/action_model': 9.966500752137196e-05, 'epoch': 0.14}) +[2026-06-17 16:57:05,752][starVLA.training.train_starvla][INFO] - Step 362, Loss: {'train/loss': 1.1078866720199585, 'train/grad_norm_pre_clip': 6.420952320098877, 'timing/data': 0.014488219283521175, 'timing/model': 1.4314352665096521, 'learning_rate/qwen_vl_interface': 9.966243866214712e-06, 'learning_rate/action_model': 9.966243866214712e-05, 'epoch': 0.14}) +[2026-06-17 16:57:11,206][starVLA.training.train_starvla][INFO] - Step 363, Loss: {'train/loss': 0.7325795292854309, 'train/grad_norm_pre_clip': 9.060296058654785, 'timing/data': 0.015867628157138824, 'timing/model': 1.076603202149272, 'learning_rate/qwen_vl_interface': 9.965986002609799e-06, 'learning_rate/action_model': 9.965986002609799e-05, 'epoch': 0.14}) +[2026-06-17 16:57:15,972][starVLA.training.train_starvla][INFO] - Step 364, Loss: {'train/loss': 0.960045576095581, 'train/grad_norm_pre_clip': 8.985408782958984, 'timing/data': 0.01457649003714323, 'timing/model': 1.120032660663128, 'learning_rate/qwen_vl_interface': 9.965727161375915e-06, 'learning_rate/action_model': 9.965727161375914e-05, 'epoch': 0.14}) +[2026-06-17 16:57:21,734][starVLA.training.train_starvla][INFO] - Step 365, Loss: {'train/loss': 0.6864843964576721, 'train/grad_norm_pre_clip': 10.355925559997559, 'timing/data': 0.01238609105348587, 'timing/model': 1.0746709052473307, 'learning_rate/qwen_vl_interface': 9.965467342566715e-06, 'learning_rate/action_model': 9.965467342566715e-05, 'epoch': 0.14}) +[2026-06-17 16:57:27,089][starVLA.training.train_starvla][INFO] - Step 366, Loss: {'train/loss': 0.7420873641967773, 'train/grad_norm_pre_clip': 9.685741424560547, 'timing/data': 0.01033196970820427, 'timing/model': 1.1242844127118587, 'learning_rate/qwen_vl_interface': 9.965206546236061e-06, 'learning_rate/action_model': 9.96520654623606e-05, 'epoch': 0.14}) +[2026-06-17 16:57:31,737][starVLA.training.train_starvla][INFO] - Step 367, Loss: {'train/loss': 0.6919366717338562, 'train/grad_norm_pre_clip': 6.1552252769470215, 'timing/data': 0.013544381596148014, 'timing/model': 1.4139921376481652, 'learning_rate/qwen_vl_interface': 9.964944772438014e-06, 'learning_rate/action_model': 9.964944772438013e-05, 'epoch': 0.14}) +[2026-06-17 16:57:35,950][starVLA.training.train_starvla][INFO] - Step 368, Loss: {'train/loss': 0.583614706993103, 'train/grad_norm_pre_clip': 8.914793968200684, 'timing/data': 0.01080032903701067, 'timing/model': 1.0754014290869236, 'learning_rate/qwen_vl_interface': 9.964682021226844e-06, 'learning_rate/action_model': 9.964682021226844e-05, 'epoch': 0.14}) +[2026-06-17 16:57:40,883][starVLA.training.train_starvla][INFO] - Step 369, Loss: {'train/loss': 0.8761038184165955, 'train/grad_norm_pre_clip': 7.787483215332031, 'timing/data': 0.012650285847485065, 'timing/model': 1.0869183465838432, 'learning_rate/qwen_vl_interface': 9.964418292657019e-06, 'learning_rate/action_model': 9.964418292657018e-05, 'epoch': 0.14}) +[2026-06-17 16:57:45,552][starVLA.training.train_starvla][INFO] - Step 370, Loss: {'train/loss': 0.6627538800239563, 'train/grad_norm_pre_clip': 13.230918884277344, 'timing/data': 0.011105664074420929, 'timing/model': 1.0687026688829064, 'learning_rate/qwen_vl_interface': 9.96415358678321e-06, 'learning_rate/action_model': 9.964153586783209e-05, 'epoch': 0.14}) +[2026-06-17 16:57:50,367][starVLA.training.train_starvla][INFO] - Step 371, Loss: {'train/loss': 0.7407515645027161, 'train/grad_norm_pre_clip': 7.757720470428467, 'timing/data': 0.016535053960978985, 'timing/model': 1.1807571044191718, 'learning_rate/qwen_vl_interface': 9.963887903660285e-06, 'learning_rate/action_model': 9.963887903660285e-05, 'epoch': 0.14}) +[2026-06-17 16:57:55,083][starVLA.training.train_starvla][INFO] - Step 372, Loss: {'train/loss': 0.809504508972168, 'train/grad_norm_pre_clip': 7.637071132659912, 'timing/data': 0.014555093832314014, 'timing/model': 1.424175539985299, 'learning_rate/qwen_vl_interface': 9.96362124334333e-06, 'learning_rate/action_model': 9.963621243343329e-05, 'epoch': 0.14}) +[2026-06-17 16:57:59,316][starVLA.training.train_starvla][INFO] - Step 373, Loss: {'train/loss': 0.4097931683063507, 'train/grad_norm_pre_clip': 12.99285888671875, 'timing/data': 0.01321528758853674, 'timing/model': 1.0838687922805548, 'learning_rate/qwen_vl_interface': 9.963353605887617e-06, 'learning_rate/action_model': 9.963353605887616e-05, 'epoch': 0.14}) +[2026-06-17 16:58:04,104][starVLA.training.train_starvla][INFO] - Step 374, Loss: {'train/loss': 0.750271737575531, 'train/grad_norm_pre_clip': 8.573586463928223, 'timing/data': 0.011824899353086948, 'timing/model': 1.1209017736837268, 'learning_rate/qwen_vl_interface': 9.963084991348628e-06, 'learning_rate/action_model': 9.963084991348628e-05, 'epoch': 0.14}) +[2026-06-17 16:58:08,769][starVLA.training.train_starvla][INFO] - Step 375, Loss: {'train/loss': 0.649959921836853, 'train/grad_norm_pre_clip': 8.366368293762207, 'timing/data': 0.018561571836471558, 'timing/model': 1.1585783511400223, 'learning_rate/qwen_vl_interface': 9.962815399782051e-06, 'learning_rate/action_model': 9.96281539978205e-05, 'epoch': 0.14}) +[2026-06-17 16:58:13,363][starVLA.training.train_starvla][INFO] - Step 376, Loss: {'train/loss': 0.9518682360649109, 'train/grad_norm_pre_clip': 8.973740577697754, 'timing/data': 0.01101908553391695, 'timing/model': 1.1111119175329804, 'learning_rate/qwen_vl_interface': 9.96254483124377e-06, 'learning_rate/action_model': 9.96254483124377e-05, 'epoch': 0.14}) +[2026-06-17 16:58:18,132][starVLA.training.train_starvla][INFO] - Step 377, Loss: {'train/loss': 0.9051256775856018, 'train/grad_norm_pre_clip': 14.865260124206543, 'timing/data': 0.012186236679553986, 'timing/model': 1.5044567650184035, 'learning_rate/qwen_vl_interface': 9.962273285789876e-06, 'learning_rate/action_model': 9.962273285789875e-05, 'epoch': 0.15}) +[2026-06-17 16:58:22,439][starVLA.training.train_starvla][INFO] - Step 378, Loss: {'train/loss': 1.0563366413116455, 'train/grad_norm_pre_clip': 9.302461624145508, 'timing/data': 0.011129158549010754, 'timing/model': 1.153921696357429, 'learning_rate/qwen_vl_interface': 9.962000763476658e-06, 'learning_rate/action_model': 9.962000763476658e-05, 'epoch': 0.15}) +[2026-06-17 16:58:27,139][starVLA.training.train_starvla][INFO] - Step 379, Loss: {'train/loss': 0.7630104422569275, 'train/grad_norm_pre_clip': 6.906381130218506, 'timing/data': 0.016100922599434853, 'timing/model': 1.1135970763862133, 'learning_rate/qwen_vl_interface': 9.96172726436061e-06, 'learning_rate/action_model': 9.96172726436061e-05, 'epoch': 0.15}) +[2026-06-17 16:58:31,642][starVLA.training.train_starvla][INFO] - Step 380, Loss: {'train/loss': 0.8885148763656616, 'train/grad_norm_pre_clip': 6.247840881347656, 'timing/data': 0.01129772700369358, 'timing/model': 1.0964112132787704, 'learning_rate/qwen_vl_interface': 9.961452788498432e-06, 'learning_rate/action_model': 9.961452788498433e-05, 'epoch': 0.15}) +[2026-06-17 16:58:36,381][starVLA.training.train_starvla][INFO] - Step 381, Loss: {'train/loss': 0.7669980525970459, 'train/grad_norm_pre_clip': 6.309559345245361, 'timing/data': 0.011086336337029934, 'timing/model': 1.077861095778644, 'learning_rate/qwen_vl_interface': 9.96117733594702e-06, 'learning_rate/action_model': 9.96117733594702e-05, 'epoch': 0.15}) +[2026-06-17 16:58:43,214][starVLA.training.train_starvla][INFO] - Step 382, Loss: {'train/loss': 0.6921284198760986, 'train/grad_norm_pre_clip': 9.790112495422363, 'timing/data': 0.020019163377583027, 'timing/model': 1.5741608496755362, 'learning_rate/qwen_vl_interface': 9.960900906763478e-06, 'learning_rate/action_model': 9.960900906763477e-05, 'epoch': 0.15}) +[2026-06-17 16:58:49,346][starVLA.training.train_starvla][INFO] - Step 383, Loss: {'train/loss': 0.903128981590271, 'train/grad_norm_pre_clip': 5.813108444213867, 'timing/data': 0.012444517575204372, 'timing/model': 1.1316312430426478, 'learning_rate/qwen_vl_interface': 9.960623501005108e-06, 'learning_rate/action_model': 9.960623501005107e-05, 'epoch': 0.15}) +[2026-06-17 16:58:55,015][starVLA.training.train_starvla][INFO] - Step 384, Loss: {'train/loss': 0.7073560953140259, 'train/grad_norm_pre_clip': 7.306790351867676, 'timing/data': 0.012686081230640411, 'timing/model': 1.220746187493205, 'learning_rate/qwen_vl_interface': 9.960345118729416e-06, 'learning_rate/action_model': 9.960345118729417e-05, 'epoch': 0.15}) +[2026-06-17 16:58:59,603][starVLA.training.train_starvla][INFO] - Step 385, Loss: {'train/loss': 0.9289526343345642, 'train/grad_norm_pre_clip': 6.324265003204346, 'timing/data': 0.012819602154195309, 'timing/model': 1.0825740778818727, 'learning_rate/qwen_vl_interface': 9.960065759994116e-06, 'learning_rate/action_model': 9.960065759994114e-05, 'epoch': 0.15}) +[2026-06-17 16:59:04,364][starVLA.training.train_starvla][INFO] - Step 386, Loss: {'train/loss': 1.1907283067703247, 'train/grad_norm_pre_clip': 8.342960357666016, 'timing/data': 0.01648261770606041, 'timing/model': 1.1677616145461798, 'learning_rate/qwen_vl_interface': 9.959785424857113e-06, 'learning_rate/action_model': 9.959785424857112e-05, 'epoch': 0.15}) +[2026-06-17 16:59:09,445][starVLA.training.train_starvla][INFO] - Step 387, Loss: {'train/loss': 0.6293027400970459, 'train/grad_norm_pre_clip': 5.0625, 'timing/data': 0.013350944966077805, 'timing/model': 1.5200454788282514, 'learning_rate/qwen_vl_interface': 9.959504113376523e-06, 'learning_rate/action_model': 9.959504113376523e-05, 'epoch': 0.15}) +[2026-06-17 16:59:13,775][starVLA.training.train_starvla][INFO] - Step 388, Loss: {'train/loss': 1.024780511856079, 'train/grad_norm_pre_clip': 6.897103786468506, 'timing/data': 0.0128641901537776, 'timing/model': 1.0942440731450915, 'learning_rate/qwen_vl_interface': 9.959221825610664e-06, 'learning_rate/action_model': 9.959221825610663e-05, 'epoch': 0.15}) +[2026-06-17 16:59:18,702][starVLA.training.train_starvla][INFO] - Step 389, Loss: {'train/loss': 0.6413803100585938, 'train/grad_norm_pre_clip': 8.61959171295166, 'timing/data': 0.011255728080868721, 'timing/model': 1.1525534624233842, 'learning_rate/qwen_vl_interface': 9.958938561618053e-06, 'learning_rate/action_model': 9.958938561618053e-05, 'epoch': 0.15}) +[2026-06-17 16:59:23,379][starVLA.training.train_starvla][INFO] - Step 390, Loss: {'train/loss': 0.9978068470954895, 'train/grad_norm_pre_clip': 5.628598690032959, 'timing/data': 0.01080822292715311, 'timing/model': 1.114416828379035, 'learning_rate/qwen_vl_interface': 9.958654321457412e-06, 'learning_rate/action_model': 9.958654321457412e-05, 'epoch': 0.15}) +[2026-06-17 16:59:28,146][starVLA.training.train_starvla][INFO] - Step 391, Loss: {'train/loss': 0.7102017402648926, 'train/grad_norm_pre_clip': 3.899384021759033, 'timing/data': 0.01081199012696743, 'timing/model': 1.080974797718227, 'learning_rate/qwen_vl_interface': 9.958369105187661e-06, 'learning_rate/action_model': 9.958369105187661e-05, 'epoch': 0.15}) +[2026-06-17 16:59:32,899][starVLA.training.train_starvla][INFO] - Step 392, Loss: {'train/loss': 0.7702648639678955, 'train/grad_norm_pre_clip': 7.200927257537842, 'timing/data': 0.01236195582896471, 'timing/model': 1.4684233264997602, 'learning_rate/qwen_vl_interface': 9.958082912867932e-06, 'learning_rate/action_model': 9.958082912867931e-05, 'epoch': 0.15}) +[2026-06-17 16:59:37,337][starVLA.training.train_starvla][INFO] - Step 393, Loss: {'train/loss': 0.8038672804832458, 'train/grad_norm_pre_clip': 6.309657096862793, 'timing/data': 0.014179890044033527, 'timing/model': 1.1800933787599206, 'learning_rate/qwen_vl_interface': 9.957795744557545e-06, 'learning_rate/action_model': 9.957795744557546e-05, 'epoch': 0.15}) +[2026-06-17 16:59:42,066][starVLA.training.train_starvla][INFO] - Step 394, Loss: {'train/loss': 0.8813778758049011, 'train/grad_norm_pre_clip': 7.961936950683594, 'timing/data': 0.014149664901196957, 'timing/model': 1.1760899713262916, 'learning_rate/qwen_vl_interface': 9.957507600316036e-06, 'learning_rate/action_model': 9.957507600316036e-05, 'epoch': 0.15}) +[2026-06-17 16:59:48,415][starVLA.training.train_starvla][INFO] - Step 395, Loss: {'train/loss': 0.668722927570343, 'train/grad_norm_pre_clip': 9.44173812866211, 'timing/data': 0.011897529475390911, 'timing/model': 1.0965570993721485, 'learning_rate/qwen_vl_interface': 9.957218480203137e-06, 'learning_rate/action_model': 9.957218480203136e-05, 'epoch': 0.15}) +[2026-06-17 16:59:53,136][starVLA.training.train_starvla][INFO] - Step 396, Loss: {'train/loss': 0.9166054725646973, 'train/grad_norm_pre_clip': 10.29389762878418, 'timing/data': 0.012916496954858303, 'timing/model': 1.2018105247989297, 'learning_rate/qwen_vl_interface': 9.95692838427878e-06, 'learning_rate/action_model': 9.956928384278778e-05, 'epoch': 0.15}) +[2026-06-17 17:00:00,507][starVLA.training.train_starvla][INFO] - Step 397, Loss: {'train/loss': 0.7073128819465637, 'train/grad_norm_pre_clip': 8.247756958007812, 'timing/data': 0.01795890461653471, 'timing/model': 1.7727382276207209, 'learning_rate/qwen_vl_interface': 9.956637312603105e-06, 'learning_rate/action_model': 9.956637312603104e-05, 'epoch': 0.15}) +[2026-06-17 17:00:04,987][starVLA.training.train_starvla][INFO] - Step 398, Loss: {'train/loss': 0.8059131503105164, 'train/grad_norm_pre_clip': 4.578461647033691, 'timing/data': 0.01062171533703804, 'timing/model': 1.1835233382880688, 'learning_rate/qwen_vl_interface': 9.95634526523645e-06, 'learning_rate/action_model': 9.956345265236449e-05, 'epoch': 0.15}) +[2026-06-17 17:00:15,233][starVLA.training.train_starvla][INFO] - Step 399, Loss: {'train/loss': 1.1449521780014038, 'train/grad_norm_pre_clip': 12.992323875427246, 'timing/data': 0.015145515091717243, 'timing/model': 3.922328452579677, 'learning_rate/qwen_vl_interface': 9.956052242239358e-06, 'learning_rate/action_model': 9.956052242239358e-05, 'epoch': 0.15}) +[2026-06-17 17:03:31,089][starVLA.training.train_starvla][INFO] - Step 400, Loss: {'train/loss': 0.7645134925842285, 'train/grad_norm_pre_clip': 5.86122465133667, 'eval/loss': 0.8085877299308777, 'eval/latency_4/loss': 0.8085877299308777, 'eval/demon_fire_precision': 0.9269254093389934, 'eval/demon_fire_recall': 0.9928548229944787, 'eval/demon_fire_f1': 0.9587580366943704, 'eval/demon_fire_rate_model': 0.968575624082232, 'eval/demon_fire_rate_teacher': 0.9042584434654919, 'eval/demon_left_precision': 0.896551724137931, 'eval/demon_left_recall': 0.1187214611872146, 'eval/demon_left_f1': 0.20967741935483875, 'eval/demon_left_rate_model': 0.017033773861967694, 'eval/demon_left_rate_teacher': 0.12863436123348018, 'eval/demon_right_precision': 0.6540404040404041, 'eval/demon_right_recall': 0.4675090252707581, 'eval/demon_right_f1': 0.5452631578947369, 'eval/demon_right_rate_model': 0.11629955947136564, 'eval/demon_right_rate_teacher': 0.16270190895741557, 'eval/demon_move_f1': 0.3774702886247878, 'eval/demon_cc_f1': 0.6557911317489526, 'eval/latency_4/demon_fire_precision': 0.9269254093389934, 'eval/latency_4/demon_fire_recall': 0.9928548229944787, 'eval/latency_4/demon_fire_f1': 0.9587580366943704, 'eval/latency_4/demon_fire_rate_model': 0.968575624082232, 'eval/latency_4/demon_fire_rate_teacher': 0.9042584434654919, 'eval/latency_4/demon_left_precision': 0.896551724137931, 'eval/latency_4/demon_left_recall': 0.1187214611872146, 'eval/latency_4/demon_left_f1': 0.20967741935483875, 'eval/latency_4/demon_left_rate_model': 0.017033773861967694, 'eval/latency_4/demon_left_rate_teacher': 0.12863436123348018, 'eval/latency_4/demon_right_precision': 0.6540404040404041, 'eval/latency_4/demon_right_recall': 0.4675090252707581, 'eval/latency_4/demon_right_f1': 0.5452631578947369, 'eval/latency_4/demon_right_rate_model': 0.11629955947136564, 'eval/latency_4/demon_right_rate_teacher': 0.16270190895741557, 'eval/latency_4/demon_move_f1': 0.3774702886247878, 'eval/latency_4/demon_cc_f1': 0.6557911317489526, 'timing/data': 0.011545340530574322, 'timing/model': 1.1172120943665504, 'learning_rate/qwen_vl_interface': 9.955758243672572e-06, 'learning_rate/action_model': 9.955758243672572e-05, 'epoch': 0.15}) +[2026-06-17 17:03:31,091][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_400_state +[2026-06-17 17:03:31,099][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 17:03:39,813][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_400_state/model.safetensors +[2026-06-17 17:03:54,607][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_400_state/optimizer.bin +[2026-06-17 17:03:54,610][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_400_state/sampler.bin +[2026-06-17 17:03:54,610][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_400_state/sampler_1.bin +[2026-06-17 17:03:54,613][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_400_state/random_states_0.pkl +[2026-06-17 17:03:54,620][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 400 +[2026-06-17 17:03:54,620][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 17:03:54,638][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 17:03:59,426][starVLA.training.train_starvla][INFO] - Step 401, Loss: {'train/loss': 0.8091632127761841, 'train/grad_norm_pre_clip': 8.099766731262207, 'timing/data': 0.01858001109212637, 'timing/model': 1.1467121178284287, 'learning_rate/qwen_vl_interface': 9.955463269597038e-06, 'learning_rate/action_model': 9.955463269597038e-05, 'epoch': 0.15}) +[2026-06-17 17:04:05,912][starVLA.training.train_starvla][INFO] - Step 402, Loss: {'train/loss': 0.9879324436187744, 'train/grad_norm_pre_clip': 10.015565872192383, 'timing/data': 0.012562451884150505, 'timing/model': 3.314309012144804, 'learning_rate/qwen_vl_interface': 9.955167320073905e-06, 'learning_rate/action_model': 9.955167320073905e-05, 'epoch': 0.15}) +[2026-06-17 17:04:10,309][starVLA.training.train_starvla][INFO] - Step 403, Loss: {'train/loss': 0.7424731850624084, 'train/grad_norm_pre_clip': 5.873526096343994, 'timing/data': 0.012239937670528889, 'timing/model': 1.0848270887508988, 'learning_rate/qwen_vl_interface': 9.954870395164524e-06, 'learning_rate/action_model': 9.954870395164524e-05, 'epoch': 0.16}) +[2026-06-17 17:04:14,977][starVLA.training.train_starvla][INFO] - Step 404, Loss: {'train/loss': 0.6403638124465942, 'train/grad_norm_pre_clip': 5.830399036407471, 'timing/data': 0.011398636735975742, 'timing/model': 1.1467527141794562, 'learning_rate/qwen_vl_interface': 9.954572494930448e-06, 'learning_rate/action_model': 9.954572494930448e-05, 'epoch': 0.16}) +[2026-06-17 17:04:19,769][starVLA.training.train_starvla][INFO] - Step 405, Loss: {'train/loss': 1.1054679155349731, 'train/grad_norm_pre_clip': 5.990403652191162, 'timing/data': 0.013355479575693607, 'timing/model': 1.0985764395445585, 'learning_rate/qwen_vl_interface': 9.95427361943343e-06, 'learning_rate/action_model': 9.954273619433431e-05, 'epoch': 0.16}) +[2026-06-17 17:04:25,899][starVLA.training.train_starvla][INFO] - Step 406, Loss: {'train/loss': 0.7324576377868652, 'train/grad_norm_pre_clip': 7.5049920082092285, 'timing/data': 0.011661974713206291, 'timing/model': 2.6029964797198772, 'learning_rate/qwen_vl_interface': 9.95397376873543e-06, 'learning_rate/action_model': 9.95397376873543e-05, 'epoch': 0.16}) +[2026-06-17 17:04:30,572][starVLA.training.train_starvla][INFO] - Step 407, Loss: {'train/loss': 0.7123912572860718, 'train/grad_norm_pre_clip': 12.149510383605957, 'timing/data': 0.019770782440900803, 'timing/model': 1.493016080930829, 'learning_rate/qwen_vl_interface': 9.953672942898606e-06, 'learning_rate/action_model': 9.953672942898605e-05, 'epoch': 0.16}) +[2026-06-17 17:04:34,711][starVLA.training.train_starvla][INFO] - Step 408, Loss: {'train/loss': 0.9906588196754456, 'train/grad_norm_pre_clip': 7.294236660003662, 'timing/data': 0.0113257747143507, 'timing/model': 1.0571905123069882, 'learning_rate/qwen_vl_interface': 9.953371141985318e-06, 'learning_rate/action_model': 9.953371141985319e-05, 'epoch': 0.16}) +[2026-06-17 17:04:39,261][starVLA.training.train_starvla][INFO] - Step 409, Loss: {'train/loss': 0.7399798035621643, 'train/grad_norm_pre_clip': 6.93442964553833, 'timing/data': 0.014299517497420311, 'timing/model': 1.0785863110795617, 'learning_rate/qwen_vl_interface': 9.953068366058133e-06, 'learning_rate/action_model': 9.953068366058133e-05, 'epoch': 0.16}) +[2026-06-17 17:04:43,845][starVLA.training.train_starvla][INFO] - Step 410, Loss: {'train/loss': 0.5488879084587097, 'train/grad_norm_pre_clip': 6.965993881225586, 'timing/data': 0.011715068481862545, 'timing/model': 1.0635394118726254, 'learning_rate/qwen_vl_interface': 9.952764615179813e-06, 'learning_rate/action_model': 9.952764615179812e-05, 'epoch': 0.16}) +[2026-06-17 17:04:48,433][starVLA.training.train_starvla][INFO] - Step 411, Loss: {'train/loss': 0.7354949712753296, 'train/grad_norm_pre_clip': 8.663551330566406, 'timing/data': 0.012662822380661964, 'timing/model': 1.090051400475204, 'learning_rate/qwen_vl_interface': 9.95245988941333e-06, 'learning_rate/action_model': 9.95245988941333e-05, 'epoch': 0.16}) +[2026-06-17 17:04:53,085][starVLA.training.train_starvla][INFO] - Step 412, Loss: {'train/loss': 0.5902643203735352, 'train/grad_norm_pre_clip': 7.212611198425293, 'timing/data': 0.02968232985585928, 'timing/model': 1.483606144785881, 'learning_rate/qwen_vl_interface': 9.952154188821852e-06, 'learning_rate/action_model': 9.952154188821851e-05, 'epoch': 0.16}) +[2026-06-17 17:04:57,518][starVLA.training.train_starvla][INFO] - Step 413, Loss: {'train/loss': 0.8521721363067627, 'train/grad_norm_pre_clip': 6.678890228271484, 'timing/data': 0.011593797244131565, 'timing/model': 1.1333787087351084, 'learning_rate/qwen_vl_interface': 9.95184751346875e-06, 'learning_rate/action_model': 9.95184751346875e-05, 'epoch': 0.16}) +[2026-06-17 17:05:02,116][starVLA.training.train_starvla][INFO] - Step 414, Loss: {'train/loss': 0.47864416241645813, 'train/grad_norm_pre_clip': 5.1452717781066895, 'timing/data': 0.011442085728049278, 'timing/model': 1.0698835784569383, 'learning_rate/qwen_vl_interface': 9.951539863417598e-06, 'learning_rate/action_model': 9.951539863417599e-05, 'epoch': 0.16}) +[2026-06-17 17:05:06,948][starVLA.training.train_starvla][INFO] - Step 415, Loss: {'train/loss': 0.690510630607605, 'train/grad_norm_pre_clip': 6.512946605682373, 'timing/data': 0.012978010810911655, 'timing/model': 1.1492088688537478, 'learning_rate/qwen_vl_interface': 9.951231238732174e-06, 'learning_rate/action_model': 9.951231238732173e-05, 'epoch': 0.16}) +[2026-06-17 17:05:11,743][starVLA.training.train_starvla][INFO] - Step 416, Loss: {'train/loss': 0.8459861278533936, 'train/grad_norm_pre_clip': 10.308232307434082, 'timing/data': 0.012171352282166481, 'timing/model': 1.0996367176994681, 'learning_rate/qwen_vl_interface': 9.950921639476455e-06, 'learning_rate/action_model': 9.950921639476455e-05, 'epoch': 0.16}) +[2026-06-17 17:05:16,226][starVLA.training.train_starvla][INFO] - Step 417, Loss: {'train/loss': 0.6463121175765991, 'train/grad_norm_pre_clip': 6.381840705871582, 'timing/data': 0.012296034954488277, 'timing/model': 1.3538306839764118, 'learning_rate/qwen_vl_interface': 9.950611065714623e-06, 'learning_rate/action_model': 9.950611065714623e-05, 'epoch': 0.16}) +[2026-06-17 17:05:21,523][starVLA.training.train_starvla][INFO] - Step 418, Loss: {'train/loss': 0.8849398493766785, 'train/grad_norm_pre_clip': 7.515785217285156, 'timing/data': 0.019092338159680367, 'timing/model': 2.1422713585197926, 'learning_rate/qwen_vl_interface': 9.950299517511058e-06, 'learning_rate/action_model': 9.950299517511058e-05, 'epoch': 0.16}) +[2026-06-17 17:05:26,144][starVLA.training.train_starvla][INFO] - Step 419, Loss: {'train/loss': 0.4990246295928955, 'train/grad_norm_pre_clip': 6.934450149536133, 'timing/data': 0.015633911825716496, 'timing/model': 1.1150835044682026, 'learning_rate/qwen_vl_interface': 9.949986994930346e-06, 'learning_rate/action_model': 9.949986994930346e-05, 'epoch': 0.16}) +[2026-06-17 17:05:30,694][starVLA.training.train_starvla][INFO] - Step 420, Loss: {'train/loss': 0.8950783014297485, 'train/grad_norm_pre_clip': 10.872337341308594, 'timing/data': 0.012850826606154442, 'timing/model': 1.0702038472518325, 'learning_rate/qwen_vl_interface': 9.949673498037272e-06, 'learning_rate/action_model': 9.949673498037272e-05, 'epoch': 0.16}) +[2026-06-17 17:05:35,442][starVLA.training.train_starvla][INFO] - Step 421, Loss: {'train/loss': 0.562373161315918, 'train/grad_norm_pre_clip': 6.3586907386779785, 'timing/data': 0.012706984765827656, 'timing/model': 1.545396857894957, 'learning_rate/qwen_vl_interface': 9.949359026896826e-06, 'learning_rate/action_model': 9.949359026896826e-05, 'epoch': 0.16}) +[2026-06-17 17:05:39,693][starVLA.training.train_starvla][INFO] - Step 422, Loss: {'train/loss': 0.7101369500160217, 'train/grad_norm_pre_clip': 6.691257953643799, 'timing/data': 0.014618667773902416, 'timing/model': 1.072688490152359, 'learning_rate/qwen_vl_interface': 9.949043581574196e-06, 'learning_rate/action_model': 9.949043581574196e-05, 'epoch': 0.16}) +[2026-06-17 17:05:44,449][starVLA.training.train_starvla][INFO] - Step 423, Loss: {'train/loss': 0.4719283878803253, 'train/grad_norm_pre_clip': 6.907154560089111, 'timing/data': 0.011110024526715279, 'timing/model': 1.0796451568603516, 'learning_rate/qwen_vl_interface': 9.948727162134775e-06, 'learning_rate/action_model': 9.948727162134775e-05, 'epoch': 0.16}) +[2026-06-17 17:05:49,052][starVLA.training.train_starvla][INFO] - Step 424, Loss: {'train/loss': 0.7318530678749084, 'train/grad_norm_pre_clip': 6.897974491119385, 'timing/data': 0.011379861272871494, 'timing/model': 1.0763084357604384, 'learning_rate/qwen_vl_interface': 9.948409768644157e-06, 'learning_rate/action_model': 9.948409768644156e-05, 'epoch': 0.16}) +[2026-06-17 17:05:54,903][starVLA.training.train_starvla][INFO] - Step 425, Loss: {'train/loss': 0.3999481797218323, 'train/grad_norm_pre_clip': 7.108658313751221, 'timing/data': 0.012618248350918293, 'timing/model': 1.1249070763587952, 'learning_rate/qwen_vl_interface': 9.948091401168139e-06, 'learning_rate/action_model': 9.948091401168138e-05, 'epoch': 0.16}) +[2026-06-17 17:05:59,598][starVLA.training.train_starvla][INFO] - Step 426, Loss: {'train/loss': 1.1773864030838013, 'train/grad_norm_pre_clip': 13.808237075805664, 'timing/data': 0.01161622442305088, 'timing/model': 1.5136400116607547, 'learning_rate/qwen_vl_interface': 9.947772059772718e-06, 'learning_rate/action_model': 9.947772059772717e-05, 'epoch': 0.16}) +[2026-06-17 17:06:04,177][starVLA.training.train_starvla][INFO] - Step 427, Loss: {'train/loss': 1.0725644826889038, 'train/grad_norm_pre_clip': 9.893050193786621, 'timing/data': 0.013010825961828232, 'timing/model': 1.4067352134734392, 'learning_rate/qwen_vl_interface': 9.947451744524094e-06, 'learning_rate/action_model': 9.947451744524093e-05, 'epoch': 0.16}) +[2026-06-17 17:06:08,500][starVLA.training.train_starvla][INFO] - Step 428, Loss: {'train/loss': 0.9011202454566956, 'train/grad_norm_pre_clip': 8.709869384765625, 'timing/data': 0.014048121869564056, 'timing/model': 1.0956481201574206, 'learning_rate/qwen_vl_interface': 9.947130455488668e-06, 'learning_rate/action_model': 9.947130455488666e-05, 'epoch': 0.16}) +[2026-06-17 17:06:13,099][starVLA.training.train_starvla][INFO] - Step 429, Loss: {'train/loss': 0.7573571801185608, 'train/grad_norm_pre_clip': 13.27852725982666, 'timing/data': 0.012146937660872936, 'timing/model': 1.115310586988926, 'learning_rate/qwen_vl_interface': 9.946808192733045e-06, 'learning_rate/action_model': 9.946808192733044e-05, 'epoch': 0.17}) +[2026-06-17 17:06:17,834][starVLA.training.train_starvla][INFO] - Step 430, Loss: {'train/loss': 0.9840074777603149, 'train/grad_norm_pre_clip': 10.424714088439941, 'timing/data': 0.011223356239497662, 'timing/model': 1.1359117589890957, 'learning_rate/qwen_vl_interface': 9.946484956324028e-06, 'learning_rate/action_model': 9.946484956324028e-05, 'epoch': 0.17}) +[2026-06-17 17:06:22,641][starVLA.training.train_starvla][INFO] - Step 431, Loss: {'train/loss': 0.6573261022567749, 'train/grad_norm_pre_clip': 8.947357177734375, 'timing/data': 0.014125163666903973, 'timing/model': 1.0987713150680065, 'learning_rate/qwen_vl_interface': 9.946160746328628e-06, 'learning_rate/action_model': 9.946160746328627e-05, 'epoch': 0.17}) +[2026-06-17 17:06:27,303][starVLA.training.train_starvla][INFO] - Step 432, Loss: {'train/loss': 0.6807630062103271, 'train/grad_norm_pre_clip': 10.524039268493652, 'timing/data': 0.013181529939174652, 'timing/model': 1.397678586654365, 'learning_rate/qwen_vl_interface': 9.945835562814052e-06, 'learning_rate/action_model': 9.945835562814051e-05, 'epoch': 0.17}) +[2026-06-17 17:06:31,654][starVLA.training.train_starvla][INFO] - Step 433, Loss: {'train/loss': 1.0574886798858643, 'train/grad_norm_pre_clip': 12.907651901245117, 'timing/data': 0.013786748982965946, 'timing/model': 1.1519158789888024, 'learning_rate/qwen_vl_interface': 9.94550940584771e-06, 'learning_rate/action_model': 9.945509405847709e-05, 'epoch': 0.17}) +[2026-06-17 17:06:36,312][starVLA.training.train_starvla][INFO] - Step 434, Loss: {'train/loss': 0.9813334345817566, 'train/grad_norm_pre_clip': 10.264719009399414, 'timing/data': 0.013485663570463657, 'timing/model': 1.0914561338722706, 'learning_rate/qwen_vl_interface': 9.945182275497215e-06, 'learning_rate/action_model': 9.945182275497215e-05, 'epoch': 0.17}) +[2026-06-17 17:06:41,182][starVLA.training.train_starvla][INFO] - Step 435, Loss: {'train/loss': 0.9068309664726257, 'train/grad_norm_pre_clip': 8.104419708251953, 'timing/data': 0.013076494447886944, 'timing/model': 1.092589576728642, 'learning_rate/qwen_vl_interface': 9.944854171830382e-06, 'learning_rate/action_model': 9.944854171830383e-05, 'epoch': 0.17}) +[2026-06-17 17:06:45,789][starVLA.training.train_starvla][INFO] - Step 436, Loss: {'train/loss': 0.7976364493370056, 'train/grad_norm_pre_clip': 7.679595470428467, 'timing/data': 0.015254213474690914, 'timing/model': 1.0852690767496824, 'learning_rate/qwen_vl_interface': 9.94452509491523e-06, 'learning_rate/action_model': 9.94452509491523e-05, 'epoch': 0.17}) +[2026-06-17 17:06:50,590][starVLA.training.train_starvla][INFO] - Step 437, Loss: {'train/loss': 0.7916442155838013, 'train/grad_norm_pre_clip': 16.741674423217773, 'timing/data': 0.01272489968687296, 'timing/model': 1.5105349710211158, 'learning_rate/qwen_vl_interface': 9.944195044819973e-06, 'learning_rate/action_model': 9.944195044819972e-05, 'epoch': 0.17}) +[2026-06-17 17:06:55,381][starVLA.training.train_starvla][INFO] - Step 438, Loss: {'train/loss': 0.8832641243934631, 'train/grad_norm_pre_clip': 7.945284843444824, 'timing/data': 0.012133766897022724, 'timing/model': 1.0927308462560177, 'learning_rate/qwen_vl_interface': 9.94386402161303e-06, 'learning_rate/action_model': 9.943864021613029e-05, 'epoch': 0.17}) +[2026-06-17 17:07:00,018][starVLA.training.train_starvla][INFO] - Step 439, Loss: {'train/loss': 0.9767321348190308, 'train/grad_norm_pre_clip': 11.583629608154297, 'timing/data': 0.013472050428390503, 'timing/model': 1.1307225292548537, 'learning_rate/qwen_vl_interface': 9.943532025363026e-06, 'learning_rate/action_model': 9.943532025363025e-05, 'epoch': 0.17}) +[2026-06-17 17:07:04,696][starVLA.training.train_starvla][INFO] - Step 440, Loss: {'train/loss': 0.741624116897583, 'train/grad_norm_pre_clip': 6.510015964508057, 'timing/data': 0.021779256872832775, 'timing/model': 1.0983632281422615, 'learning_rate/qwen_vl_interface': 9.943199056138781e-06, 'learning_rate/action_model': 9.943199056138782e-05, 'epoch': 0.17}) +[2026-06-17 17:07:09,635][starVLA.training.train_starvla][INFO] - Step 441, Loss: {'train/loss': 0.6748288869857788, 'train/grad_norm_pre_clip': 4.925070762634277, 'timing/data': 0.013350115157663822, 'timing/model': 1.1305138543248177, 'learning_rate/qwen_vl_interface': 9.942865114009325e-06, 'learning_rate/action_model': 9.942865114009324e-05, 'epoch': 0.17}) +[2026-06-17 17:07:14,414][starVLA.training.train_starvla][INFO] - Step 442, Loss: {'train/loss': 0.6543954610824585, 'train/grad_norm_pre_clip': 5.646193504333496, 'timing/data': 0.0120059410110116, 'timing/model': 1.5576840043067932, 'learning_rate/qwen_vl_interface': 9.94253019904388e-06, 'learning_rate/action_model': 9.942530199043879e-05, 'epoch': 0.17}) +[2026-06-17 17:07:18,700][starVLA.training.train_starvla][INFO] - Step 443, Loss: {'train/loss': 0.9448226690292358, 'train/grad_norm_pre_clip': 7.0664591789245605, 'timing/data': 0.016732526943087578, 'timing/model': 1.1435659741982818, 'learning_rate/qwen_vl_interface': 9.942194311311874e-06, 'learning_rate/action_model': 9.942194311311874e-05, 'epoch': 0.17}) +[2026-06-17 17:07:23,497][starVLA.training.train_starvla][INFO] - Step 444, Loss: {'train/loss': 0.7894587516784668, 'train/grad_norm_pre_clip': 7.063201427459717, 'timing/data': 0.016553444787859917, 'timing/model': 1.1437259018421173, 'learning_rate/qwen_vl_interface': 9.941857450882939e-06, 'learning_rate/action_model': 9.941857450882937e-05, 'epoch': 0.17}) +[2026-06-17 17:07:28,228][starVLA.training.train_starvla][INFO] - Step 445, Loss: {'train/loss': 0.7868587374687195, 'train/grad_norm_pre_clip': 5.983524322509766, 'timing/data': 0.01354063581675291, 'timing/model': 1.068322179839015, 'learning_rate/qwen_vl_interface': 9.941519617826906e-06, 'learning_rate/action_model': 9.941519617826906e-05, 'epoch': 0.17}) +[2026-06-17 17:07:32,760][starVLA.training.train_starvla][INFO] - Step 446, Loss: {'train/loss': 0.6363700032234192, 'train/grad_norm_pre_clip': 5.938126564025879, 'timing/data': 0.013656161725521088, 'timing/model': 1.156259330920875, 'learning_rate/qwen_vl_interface': 9.941180812213807e-06, 'learning_rate/action_model': 9.941180812213806e-05, 'epoch': 0.17}) +[2026-06-17 17:07:37,428][starVLA.training.train_starvla][INFO] - Step 447, Loss: {'train/loss': 0.8737533092498779, 'train/grad_norm_pre_clip': 7.9882683753967285, 'timing/data': 0.015234208665788174, 'timing/model': 1.5187874203547835, 'learning_rate/qwen_vl_interface': 9.940841034113878e-06, 'learning_rate/action_model': 9.940841034113877e-05, 'epoch': 0.17}) +[2026-06-17 17:07:42,892][starVLA.training.train_starvla][INFO] - Step 448, Loss: {'train/loss': 0.4617786109447479, 'train/grad_norm_pre_clip': 4.358555793762207, 'timing/data': 0.013673413544893265, 'timing/model': 1.4791619144380093, 'learning_rate/qwen_vl_interface': 9.940500283597557e-06, 'learning_rate/action_model': 9.940500283597556e-05, 'epoch': 0.17}) +[2026-06-17 17:07:47,420][starVLA.training.train_starvla][INFO] - Step 449, Loss: {'train/loss': 0.6433159112930298, 'train/grad_norm_pre_clip': 6.409801959991455, 'timing/data': 0.017177319154143333, 'timing/model': 1.0821764832362533, 'learning_rate/qwen_vl_interface': 9.940158560735478e-06, 'learning_rate/action_model': 9.940158560735477e-05, 'epoch': 0.17}) +[2026-06-17 17:07:52,807][starVLA.training.train_starvla][INFO] - Step 450, Loss: {'train/loss': 0.43340450525283813, 'train/grad_norm_pre_clip': 3.430128335952759, 'timing/data': 0.012601461261510849, 'timing/model': 1.119378225877881, 'learning_rate/qwen_vl_interface': 9.939815865598482e-06, 'learning_rate/action_model': 9.93981586559848e-05, 'epoch': 0.17}) +[2026-06-17 17:07:57,517][starVLA.training.train_starvla][INFO] - Step 451, Loss: {'train/loss': 0.5081261992454529, 'train/grad_norm_pre_clip': 8.335775375366211, 'timing/data': 0.014597822912037373, 'timing/model': 1.1016725096851587, 'learning_rate/qwen_vl_interface': 9.939472198257611e-06, 'learning_rate/action_model': 9.93947219825761e-05, 'epoch': 0.17}) +[2026-06-17 17:08:02,166][starVLA.training.train_starvla][INFO] - Step 452, Loss: {'train/loss': 0.6952805519104004, 'train/grad_norm_pre_clip': 7.704648017883301, 'timing/data': 0.01204037107527256, 'timing/model': 1.4572201492264867, 'learning_rate/qwen_vl_interface': 9.939127558784108e-06, 'learning_rate/action_model': 9.939127558784107e-05, 'epoch': 0.17}) +[2026-06-17 17:08:07,790][starVLA.training.train_starvla][INFO] - Step 453, Loss: {'train/loss': 0.8721998929977417, 'train/grad_norm_pre_clip': 10.45557689666748, 'timing/data': 0.013820143416523933, 'timing/model': 1.098425904288888, 'learning_rate/qwen_vl_interface': 9.938781947249417e-06, 'learning_rate/action_model': 9.938781947249416e-05, 'epoch': 0.17}) +[2026-06-17 17:08:12,472][starVLA.training.train_starvla][INFO] - Step 454, Loss: {'train/loss': 0.6458197236061096, 'train/grad_norm_pre_clip': 7.512472629547119, 'timing/data': 0.013985487632453442, 'timing/model': 1.0773762548342347, 'learning_rate/qwen_vl_interface': 9.93843536372518e-06, 'learning_rate/action_model': 9.93843536372518e-05, 'epoch': 0.17}) +[2026-06-17 17:08:19,285][starVLA.training.train_starvla][INFO] - Step 455, Loss: {'train/loss': 0.6050135493278503, 'train/grad_norm_pre_clip': 5.409582614898682, 'timing/data': 0.013485094532370567, 'timing/model': 1.0601251758635044, 'learning_rate/qwen_vl_interface': 9.938087808283248e-06, 'learning_rate/action_model': 9.938087808283248e-05, 'epoch': 0.18}) +[2026-06-17 17:08:25,455][starVLA.training.train_starvla][INFO] - Step 456, Loss: {'train/loss': 0.7539260983467102, 'train/grad_norm_pre_clip': 8.850607872009277, 'timing/data': 0.013717486523091793, 'timing/model': 1.0803978312760592, 'learning_rate/qwen_vl_interface': 9.937739280995672e-06, 'learning_rate/action_model': 9.937739280995672e-05, 'epoch': 0.18}) +[2026-06-17 17:08:30,244][starVLA.training.train_starvla][INFO] - Step 457, Loss: {'train/loss': 1.241365909576416, 'train/grad_norm_pre_clip': 22.24140167236328, 'timing/data': 0.017290188930928707, 'timing/model': 1.5502942223101854, 'learning_rate/qwen_vl_interface': 9.937389781934697e-06, 'learning_rate/action_model': 9.937389781934696e-05, 'epoch': 0.18}) +[2026-06-17 17:08:38,005][starVLA.training.train_starvla][INFO] - Step 458, Loss: {'train/loss': 0.5951414108276367, 'train/grad_norm_pre_clip': 8.734784126281738, 'timing/data': 0.013671849854290485, 'timing/model': 1.0813512615859509, 'learning_rate/qwen_vl_interface': 9.937039311172776e-06, 'learning_rate/action_model': 9.937039311172776e-05, 'epoch': 0.18}) +[2026-06-17 17:08:43,235][starVLA.training.train_starvla][INFO] - Step 459, Loss: {'train/loss': 0.6931760311126709, 'train/grad_norm_pre_clip': 9.896501541137695, 'timing/data': 0.017645719461143017, 'timing/model': 1.214246422983706, 'learning_rate/qwen_vl_interface': 9.936687868782562e-06, 'learning_rate/action_model': 9.936687868782561e-05, 'epoch': 0.18}) +[2026-06-17 17:08:48,031][starVLA.training.train_starvla][INFO] - Step 460, Loss: {'train/loss': 0.6809157729148865, 'train/grad_norm_pre_clip': 8.787720680236816, 'timing/data': 0.01262896228581667, 'timing/model': 1.1769428672268987, 'learning_rate/qwen_vl_interface': 9.93633545483691e-06, 'learning_rate/action_model': 9.93633545483691e-05, 'epoch': 0.18}) +[2026-06-17 17:08:54,172][starVLA.training.train_starvla][INFO] - Step 461, Loss: {'train/loss': 0.7007470726966858, 'train/grad_norm_pre_clip': 9.21008014678955, 'timing/data': 0.01429464016109705, 'timing/model': 1.0781695358455181, 'learning_rate/qwen_vl_interface': 9.935982069408877e-06, 'learning_rate/action_model': 9.935982069408877e-05, 'epoch': 0.18}) +[2026-06-17 17:08:58,793][starVLA.training.train_starvla][INFO] - Step 462, Loss: {'train/loss': 0.7242118716239929, 'train/grad_norm_pre_clip': 6.871227741241455, 'timing/data': 0.01707781944423914, 'timing/model': 1.4780403478071094, 'learning_rate/qwen_vl_interface': 9.935627712571716e-06, 'learning_rate/action_model': 9.935627712571715e-05, 'epoch': 0.18}) +[2026-06-17 17:09:03,008][starVLA.training.train_starvla][INFO] - Step 463, Loss: {'train/loss': 0.967582106590271, 'train/grad_norm_pre_clip': 7.414393424987793, 'timing/data': 0.013890864327549934, 'timing/model': 1.0726015958935022, 'learning_rate/qwen_vl_interface': 9.93527238439889e-06, 'learning_rate/action_model': 9.935272384398889e-05, 'epoch': 0.18}) +[2026-06-17 17:09:07,732][starVLA.training.train_starvla][INFO] - Step 464, Loss: {'train/loss': 0.720095157623291, 'train/grad_norm_pre_clip': 8.6123628616333, 'timing/data': 0.013924327678978443, 'timing/model': 1.1134223584085703, 'learning_rate/qwen_vl_interface': 9.934916084964056e-06, 'learning_rate/action_model': 9.934916084964056e-05, 'epoch': 0.18}) +[2026-06-17 17:09:12,361][starVLA.training.train_starvla][INFO] - Step 465, Loss: {'train/loss': 0.648597776889801, 'train/grad_norm_pre_clip': 10.40048599243164, 'timing/data': 0.012230307795107365, 'timing/model': 1.1233754120767117, 'learning_rate/qwen_vl_interface': 9.934558814341077e-06, 'learning_rate/action_model': 9.934558814341077e-05, 'epoch': 0.18}) +[2026-06-17 17:09:17,124][starVLA.training.train_starvla][INFO] - Step 466, Loss: {'train/loss': 0.6010805368423462, 'train/grad_norm_pre_clip': 7.972537994384766, 'timing/data': 0.011998615227639675, 'timing/model': 1.0687905382364988, 'learning_rate/qwen_vl_interface': 9.934200572604015e-06, 'learning_rate/action_model': 9.934200572604015e-05, 'epoch': 0.18}) +[2026-06-17 17:09:21,742][starVLA.training.train_starvla][INFO] - Step 467, Loss: {'train/loss': 0.41606342792510986, 'train/grad_norm_pre_clip': 6.270270347595215, 'timing/data': 0.011730115860700607, 'timing/model': 1.491840704344213, 'learning_rate/qwen_vl_interface': 9.933841359827133e-06, 'learning_rate/action_model': 9.933841359827133e-05, 'epoch': 0.18}) +[2026-06-17 17:09:26,242][starVLA.training.train_starvla][INFO] - Step 468, Loss: {'train/loss': 0.9423431754112244, 'train/grad_norm_pre_clip': 15.064533233642578, 'timing/data': 0.01186329685151577, 'timing/model': 1.2561833038926125, 'learning_rate/qwen_vl_interface': 9.933481176084899e-06, 'learning_rate/action_model': 9.933481176084899e-05, 'epoch': 0.18}) +[2026-06-17 17:09:31,114][starVLA.training.train_starvla][INFO] - Step 469, Loss: {'train/loss': 1.048563838005066, 'train/grad_norm_pre_clip': 8.453134536743164, 'timing/data': 0.001614733599126339, 'timing/model': 1.0702084004878998, 'learning_rate/qwen_vl_interface': 9.933120021451976e-06, 'learning_rate/action_model': 9.933120021451976e-05, 'epoch': 0.18}) +[2026-06-17 17:09:35,733][starVLA.training.train_starvla][INFO] - Step 470, Loss: {'train/loss': 0.6764561533927917, 'train/grad_norm_pre_clip': 5.138073921203613, 'timing/data': 0.0013673314824700356, 'timing/model': 1.0895082661882043, 'learning_rate/qwen_vl_interface': 9.932757896003235e-06, 'learning_rate/action_model': 9.932757896003235e-05, 'epoch': 0.18}) +[2026-06-17 17:09:40,314][starVLA.training.train_starvla][INFO] - Step 471, Loss: {'train/loss': 0.7552300691604614, 'train/grad_norm_pre_clip': 6.542697906494141, 'timing/data': 0.0014209747314453125, 'timing/model': 1.105262540280819, 'learning_rate/qwen_vl_interface': 9.932394799813742e-06, 'learning_rate/action_model': 9.932394799813741e-05, 'epoch': 0.18}) +[2026-06-17 17:09:44,554][starVLA.training.train_starvla][INFO] - Step 472, Loss: {'train/loss': 0.7437868714332581, 'train/grad_norm_pre_clip': 5.693140983581543, 'timing/data': 0.0011247536167502403, 'timing/model': 1.0732428450137377, 'learning_rate/qwen_vl_interface': 9.932030732958768e-06, 'learning_rate/action_model': 9.932030732958769e-05, 'epoch': 0.18}) +[2026-06-17 17:09:49,148][starVLA.training.train_starvla][INFO] - Step 473, Loss: {'train/loss': 0.5825066566467285, 'train/grad_norm_pre_clip': 7.168248176574707, 'timing/data': 0.0012519070878624916, 'timing/model': 1.0971413599327207, 'learning_rate/qwen_vl_interface': 9.931665695513786e-06, 'learning_rate/action_model': 9.931665695513786e-05, 'epoch': 0.18}) +[2026-06-17 17:09:53,733][starVLA.training.train_starvla][INFO] - Step 474, Loss: {'train/loss': 1.071628451347351, 'train/grad_norm_pre_clip': 7.365584373474121, 'timing/data': 0.001134202815592289, 'timing/model': 1.0658449735492468, 'learning_rate/qwen_vl_interface': 9.931299687554466e-06, 'learning_rate/action_model': 9.931299687554466e-05, 'epoch': 0.18}) +[2026-06-17 17:09:58,369][starVLA.training.train_starvla][INFO] - Step 475, Loss: {'train/loss': 0.9033335447311401, 'train/grad_norm_pre_clip': 7.05517053604126, 'timing/data': 0.0010729068890213966, 'timing/model': 1.177482133731246, 'learning_rate/qwen_vl_interface': 9.930932709156686e-06, 'learning_rate/action_model': 9.930932709156685e-05, 'epoch': 0.18}) +[2026-06-17 17:10:03,849][starVLA.training.train_starvla][INFO] - Step 476, Loss: {'train/loss': 0.8379897475242615, 'train/grad_norm_pre_clip': 6.3379292488098145, 'timing/data': 0.000893658958375454, 'timing/model': 2.032231225632131, 'learning_rate/qwen_vl_interface': 9.930564760396518e-06, 'learning_rate/action_model': 9.930564760396518e-05, 'epoch': 0.18}) +[2026-06-17 17:10:11,102][starVLA.training.train_starvla][INFO] - Step 477, Loss: {'train/loss': 0.7108252048492432, 'train/grad_norm_pre_clip': 5.839650630950928, 'timing/data': 0.0010861363261938095, 'timing/model': 1.0610033236443996, 'learning_rate/qwen_vl_interface': 9.930195841350237e-06, 'learning_rate/action_model': 9.930195841350236e-05, 'epoch': 0.18}) +[2026-06-17 17:10:16,518][starVLA.training.train_starvla][INFO] - Step 478, Loss: {'train/loss': 0.6108108162879944, 'train/grad_norm_pre_clip': 6.399008750915527, 'timing/data': 0.0021067848429083824, 'timing/model': 1.1981585789471865, 'learning_rate/qwen_vl_interface': 9.929825952094323e-06, 'learning_rate/action_model': 9.929825952094323e-05, 'epoch': 0.18}) +[2026-06-17 17:10:23,242][starVLA.training.train_starvla][INFO] - Step 479, Loss: {'train/loss': 0.860571026802063, 'train/grad_norm_pre_clip': 5.893457412719727, 'timing/data': 0.0011207805946469307, 'timing/model': 3.158311835490167, 'learning_rate/qwen_vl_interface': 9.929455092705453e-06, 'learning_rate/action_model': 9.929455092705453e-05, 'epoch': 0.18}) +[2026-06-17 17:10:29,454][starVLA.training.train_starvla][INFO] - Step 480, Loss: {'train/loss': 0.9008824229240417, 'train/grad_norm_pre_clip': 8.134841918945312, 'timing/data': 0.0012049069628119469, 'timing/model': 1.0799761023372412, 'learning_rate/qwen_vl_interface': 9.929083263260508e-06, 'learning_rate/action_model': 9.929083263260507e-05, 'epoch': 0.18}) +[2026-06-17 17:10:34,762][starVLA.training.train_starvla][INFO] - Step 481, Loss: {'train/loss': 0.7906978130340576, 'train/grad_norm_pre_clip': 8.78363037109375, 'timing/data': 0.001374823972582817, 'timing/model': 1.0819502472877502, 'learning_rate/qwen_vl_interface': 9.928710463836566e-06, 'learning_rate/action_model': 9.928710463836565e-05, 'epoch': 0.19}) +[2026-06-17 17:10:40,403][starVLA.training.train_starvla][INFO] - Step 482, Loss: {'train/loss': 1.062679648399353, 'train/grad_norm_pre_clip': 7.921644687652588, 'timing/data': 0.0009623048827052116, 'timing/model': 1.438135671429336, 'learning_rate/qwen_vl_interface': 9.928336694510912e-06, 'learning_rate/action_model': 9.928336694510911e-05, 'epoch': 0.19}) +[2026-06-17 17:10:44,919][starVLA.training.train_starvla][INFO] - Step 483, Loss: {'train/loss': 0.7032026052474976, 'train/grad_norm_pre_clip': 8.717166900634766, 'timing/data': 0.0014269538223743439, 'timing/model': 1.135894755832851, 'learning_rate/qwen_vl_interface': 9.927961955361025e-06, 'learning_rate/action_model': 9.927961955361025e-05, 'epoch': 0.19}) +[2026-06-17 17:10:49,613][starVLA.training.train_starvla][INFO] - Step 484, Loss: {'train/loss': 0.6535489559173584, 'train/grad_norm_pre_clip': 8.329265594482422, 'timing/data': 0.00156385637819767, 'timing/model': 1.070701202377677, 'learning_rate/qwen_vl_interface': 9.927586246464593e-06, 'learning_rate/action_model': 9.927586246464593e-05, 'epoch': 0.19}) +[2026-06-17 17:10:54,210][starVLA.training.train_starvla][INFO] - Step 485, Loss: {'train/loss': 0.9810606241226196, 'train/grad_norm_pre_clip': 10.839190483093262, 'timing/data': 0.0009176386520266533, 'timing/model': 1.0835252264514565, 'learning_rate/qwen_vl_interface': 9.927209567899499e-06, 'learning_rate/action_model': 9.927209567899498e-05, 'epoch': 0.19}) +[2026-06-17 17:10:58,916][starVLA.training.train_starvla][INFO] - Step 486, Loss: {'train/loss': 0.8634269833564758, 'train/grad_norm_pre_clip': 6.715250015258789, 'timing/data': 0.0013524973765015602, 'timing/model': 1.0995013769716024, 'learning_rate/qwen_vl_interface': 9.926831919743827e-06, 'learning_rate/action_model': 9.926831919743827e-05, 'epoch': 0.19}) +[2026-06-17 17:11:03,537][starVLA.training.train_starvla][INFO] - Step 487, Loss: {'train/loss': 1.069627285003662, 'train/grad_norm_pre_clip': 8.260723114013672, 'timing/data': 0.0015795091167092323, 'timing/model': 1.483989579603076, 'learning_rate/qwen_vl_interface': 9.926453302075866e-06, 'learning_rate/action_model': 9.926453302075867e-05, 'epoch': 0.19}) +[2026-06-17 17:11:07,811][starVLA.training.train_starvla][INFO] - Step 488, Loss: {'train/loss': 0.5873214602470398, 'train/grad_norm_pre_clip': 5.010100841522217, 'timing/data': 0.0011835815384984016, 'timing/model': 1.0960388546809554, 'learning_rate/qwen_vl_interface': 9.926073714974103e-06, 'learning_rate/action_model': 9.926073714974104e-05, 'epoch': 0.19}) +[2026-06-17 17:11:12,598][starVLA.training.train_starvla][INFO] - Step 489, Loss: {'train/loss': 0.7242729663848877, 'train/grad_norm_pre_clip': 5.60869026184082, 'timing/data': 0.0011277534067630768, 'timing/model': 1.1744285132735968, 'learning_rate/qwen_vl_interface': 9.925693158517228e-06, 'learning_rate/action_model': 9.925693158517228e-05, 'epoch': 0.19}) +[2026-06-17 17:11:17,680][starVLA.training.train_starvla][INFO] - Step 490, Loss: {'train/loss': 1.0737037658691406, 'train/grad_norm_pre_clip': 7.478569030761719, 'timing/data': 0.001813972368836403, 'timing/model': 1.2584382593631744, 'learning_rate/qwen_vl_interface': 9.92531163278413e-06, 'learning_rate/action_model': 9.92531163278413e-05, 'epoch': 0.19}) +[2026-06-17 17:11:22,445][starVLA.training.train_starvla][INFO] - Step 491, Loss: {'train/loss': 0.8062282800674438, 'train/grad_norm_pre_clip': 8.655226707458496, 'timing/data': 0.0011797323822975159, 'timing/model': 1.4751579202711582, 'learning_rate/qwen_vl_interface': 9.924929137853899e-06, 'learning_rate/action_model': 9.924929137853898e-05, 'epoch': 0.19}) +[2026-06-17 17:11:27,103][starVLA.training.train_starvla][INFO] - Step 492, Loss: {'train/loss': 1.0544790029525757, 'train/grad_norm_pre_clip': 7.081881523132324, 'timing/data': 0.0011868281289935112, 'timing/model': 1.4072505459189415, 'learning_rate/qwen_vl_interface': 9.92454567380583e-06, 'learning_rate/action_model': 9.924545673805829e-05, 'epoch': 0.19}) +[2026-06-17 17:11:31,587][starVLA.training.train_starvla][INFO] - Step 493, Loss: {'train/loss': 0.732175886631012, 'train/grad_norm_pre_clip': 8.56189250946045, 'timing/data': 0.0010953200981020927, 'timing/model': 1.2501009041443467, 'learning_rate/qwen_vl_interface': 9.924161240719409e-06, 'learning_rate/action_model': 9.924161240719408e-05, 'epoch': 0.19}) +[2026-06-17 17:11:36,342][starVLA.training.train_starvla][INFO] - Step 494, Loss: {'train/loss': 0.6912162899971008, 'train/grad_norm_pre_clip': 8.563549995422363, 'timing/data': 0.0014465488493442535, 'timing/model': 1.1752527998760343, 'learning_rate/qwen_vl_interface': 9.923775838674336e-06, 'learning_rate/action_model': 9.923775838674335e-05, 'epoch': 0.19}) +[2026-06-17 17:11:41,030][starVLA.training.train_starvla][INFO] - Step 495, Loss: {'train/loss': 0.6723496913909912, 'train/grad_norm_pre_clip': 5.74692440032959, 'timing/data': 0.0014767013490200043, 'timing/model': 1.07591955922544, 'learning_rate/qwen_vl_interface': 9.9233894677505e-06, 'learning_rate/action_model': 9.9233894677505e-05, 'epoch': 0.19}) +[2026-06-17 17:11:45,917][starVLA.training.train_starvla][INFO] - Step 496, Loss: {'train/loss': 0.6423646211624146, 'train/grad_norm_pre_clip': 23.51245880126953, 'timing/data': 0.0011210665106773376, 'timing/model': 1.4320735353976488, 'learning_rate/qwen_vl_interface': 9.923002128028e-06, 'learning_rate/action_model': 9.923002128028e-05, 'epoch': 0.19}) +[2026-06-17 17:11:50,584][starVLA.training.train_starvla][INFO] - Step 497, Loss: {'train/loss': 0.8553363680839539, 'train/grad_norm_pre_clip': 8.94643783569336, 'timing/data': 0.0011696461588144302, 'timing/model': 1.3843842279165983, 'learning_rate/qwen_vl_interface': 9.922613819587131e-06, 'learning_rate/action_model': 9.92261381958713e-05, 'epoch': 0.19}) +[2026-06-17 17:11:55,287][starVLA.training.train_starvla][INFO] - Step 498, Loss: {'train/loss': 0.8149783611297607, 'train/grad_norm_pre_clip': 7.094069480895996, 'timing/data': 0.001405150629580021, 'timing/model': 1.4763140510767698, 'learning_rate/qwen_vl_interface': 9.922224542508389e-06, 'learning_rate/action_model': 9.922224542508388e-05, 'epoch': 0.19}) +[2026-06-17 17:11:59,809][starVLA.training.train_starvla][INFO] - Step 499, Loss: {'train/loss': 1.1527115106582642, 'train/grad_norm_pre_clip': 11.418906211853027, 'timing/data': 0.0012572836130857468, 'timing/model': 1.0953172389417887, 'learning_rate/qwen_vl_interface': 9.92183429687247e-06, 'learning_rate/action_model': 9.92183429687247e-05, 'epoch': 0.19}) +[2026-06-17 17:15:36,904][starVLA.training.train_starvla][INFO] - Step 500, Loss: {'train/loss': 0.7162574529647827, 'train/grad_norm_pre_clip': 5.228918552398682, 'eval/loss': 0.724399745464325, 'eval/latency_4/loss': 0.724399745464325, 'eval/demon_fire_precision': 0.920863309352518, 'eval/demon_fire_recall': 0.9977265345891523, 'eval/demon_fire_f1': 0.957755261106781, 'eval/demon_fire_rate_model': 0.9797356828193833, 'eval/demon_fire_rate_teacher': 0.9042584434654919, 'eval/demon_left_precision': 0.7534246575342466, 'eval/demon_left_recall': 0.12557077625570776, 'eval/demon_left_f1': 0.21526418786692755, 'eval/demon_left_rate_model': 0.021439060205580028, 'eval/demon_left_rate_teacher': 0.12863436123348018, 'eval/demon_right_precision': 0.8081632653061225, 'eval/demon_right_recall': 0.3574007220216607, 'eval/demon_right_f1': 0.4956195244055069, 'eval/demon_right_rate_model': 0.07195301027900147, 'eval/demon_right_rate_teacher': 0.16270190895741557, 'eval/demon_move_f1': 0.3554418561362172, 'eval/demon_cc_f1': 0.6349627434719401, 'eval/latency_4/demon_fire_precision': 0.920863309352518, 'eval/latency_4/demon_fire_recall': 0.9977265345891523, 'eval/latency_4/demon_fire_f1': 0.957755261106781, 'eval/latency_4/demon_fire_rate_model': 0.9797356828193833, 'eval/latency_4/demon_fire_rate_teacher': 0.9042584434654919, 'eval/latency_4/demon_left_precision': 0.7534246575342466, 'eval/latency_4/demon_left_recall': 0.12557077625570776, 'eval/latency_4/demon_left_f1': 0.21526418786692755, 'eval/latency_4/demon_left_rate_model': 0.021439060205580028, 'eval/latency_4/demon_left_rate_teacher': 0.12863436123348018, 'eval/latency_4/demon_right_precision': 0.8081632653061225, 'eval/latency_4/demon_right_recall': 0.3574007220216607, 'eval/latency_4/demon_right_f1': 0.4956195244055069, 'eval/latency_4/demon_right_rate_model': 0.07195301027900147, 'eval/latency_4/demon_right_rate_teacher': 0.16270190895741557, 'eval/latency_4/demon_move_f1': 0.3554418561362172, 'eval/latency_4/demon_cc_f1': 0.6349627434719401, 'timing/data': 0.001681688241660595, 'timing/model': 1.0918515343219042, 'learning_rate/qwen_vl_interface': 9.921443082760276e-06, 'learning_rate/action_model': 9.921443082760276e-05, 'epoch': 0.19}) +[2026-06-17 17:15:36,906][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_500_state +[2026-06-17 17:15:36,914][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 17:15:45,936][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_500_state/model.safetensors +[2026-06-17 17:16:08,250][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_500_state/optimizer.bin +[2026-06-17 17:16:08,253][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_500_state/sampler.bin +[2026-06-17 17:16:08,253][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_500_state/sampler_1.bin +[2026-06-17 17:16:08,256][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_500_state/random_states_0.pkl +[2026-06-17 17:16:08,261][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 500 +[2026-06-17 17:16:08,261][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 17:16:08,279][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 17:16:13,543][starVLA.training.train_starvla][INFO] - Step 501, Loss: {'train/loss': 0.5232425332069397, 'train/grad_norm_pre_clip': 5.446854114532471, 'timing/data': 0.0009600771591067314, 'timing/model': 1.5617029080167413, 'learning_rate/qwen_vl_interface': 9.921050900252904e-06, 'learning_rate/action_model': 9.921050900252903e-05, 'epoch': 0.19}) +[2026-06-17 17:16:21,104][starVLA.training.train_starvla][INFO] - Step 502, Loss: {'train/loss': 0.8180534839630127, 'train/grad_norm_pre_clip': 7.537105083465576, 'timing/data': 0.0020919302478432655, 'timing/model': 1.081138590350747, 'learning_rate/qwen_vl_interface': 9.920657749431654e-06, 'learning_rate/action_model': 9.920657749431654e-05, 'epoch': 0.19}) +[2026-06-17 17:16:26,014][starVLA.training.train_starvla][INFO] - Step 503, Loss: {'train/loss': 0.4853627383708954, 'train/grad_norm_pre_clip': 4.774901390075684, 'timing/data': 0.0011174408718943596, 'timing/model': 1.137572281062603, 'learning_rate/qwen_vl_interface': 9.920263630378026e-06, 'learning_rate/action_model': 9.920263630378024e-05, 'epoch': 0.19}) +[2026-06-17 17:16:30,735][starVLA.training.train_starvla][INFO] - Step 504, Loss: {'train/loss': 0.7112621665000916, 'train/grad_norm_pre_clip': 9.372954368591309, 'timing/data': 0.0016546407714486122, 'timing/model': 1.0991746531799436, 'learning_rate/qwen_vl_interface': 9.919868543173722e-06, 'learning_rate/action_model': 9.919868543173721e-05, 'epoch': 0.19}) +[2026-06-17 17:16:35,572][starVLA.training.train_starvla][INFO] - Step 505, Loss: {'train/loss': 0.5437443256378174, 'train/grad_norm_pre_clip': 8.5460844039917, 'timing/data': 0.0011828932911157608, 'timing/model': 1.124000558629632, 'learning_rate/qwen_vl_interface': 9.919472487900645e-06, 'learning_rate/action_model': 9.919472487900644e-05, 'epoch': 0.19}) +[2026-06-17 17:16:41,157][starVLA.training.train_starvla][INFO] - Step 506, Loss: {'train/loss': 0.8384838700294495, 'train/grad_norm_pre_clip': 7.805671215057373, 'timing/data': 0.001149042509496212, 'timing/model': 1.3729910776019096, 'learning_rate/qwen_vl_interface': 9.919075464640895e-06, 'learning_rate/action_model': 9.919075464640895e-05, 'epoch': 0.19}) +[2026-06-17 17:16:46,215][starVLA.training.train_starvla][INFO] - Step 507, Loss: {'train/loss': 0.8597913384437561, 'train/grad_norm_pre_clip': 8.177327156066895, 'timing/data': 0.001110202632844448, 'timing/model': 1.47375591378659, 'learning_rate/qwen_vl_interface': 9.91867747347678e-06, 'learning_rate/action_model': 9.918677473476779e-05, 'epoch': 0.2}) +[2026-06-17 17:16:50,923][starVLA.training.train_starvla][INFO] - Step 508, Loss: {'train/loss': 1.0013196468353271, 'train/grad_norm_pre_clip': 7.630039215087891, 'timing/data': 0.0011113611981272697, 'timing/model': 1.2372186873108149, 'learning_rate/qwen_vl_interface': 9.918278514490798e-06, 'learning_rate/action_model': 9.918278514490799e-05, 'epoch': 0.2}) +[2026-06-17 17:16:55,946][starVLA.training.train_starvla][INFO] - Step 509, Loss: {'train/loss': 0.586479663848877, 'train/grad_norm_pre_clip': 6.3461151123046875, 'timing/data': 0.0018450012430548668, 'timing/model': 1.2723941635340452, 'learning_rate/qwen_vl_interface': 9.917878587765658e-06, 'learning_rate/action_model': 9.917878587765658e-05, 'epoch': 0.2}) +[2026-06-17 17:17:04,029][starVLA.training.train_starvla][INFO] - Step 510, Loss: {'train/loss': 0.5693456530570984, 'train/grad_norm_pre_clip': 5.819746017456055, 'timing/data': 0.0009678713977336884, 'timing/model': 1.1295758271589875, 'learning_rate/qwen_vl_interface': 9.917477693384264e-06, 'learning_rate/action_model': 9.917477693384264e-05, 'epoch': 0.2}) +[2026-06-17 17:17:08,745][starVLA.training.train_starvla][INFO] - Step 511, Loss: {'train/loss': 0.5874708294868469, 'train/grad_norm_pre_clip': 5.736019611358643, 'timing/data': 0.0013492610305547714, 'timing/model': 1.1076368438079953, 'learning_rate/qwen_vl_interface': 9.917075831429721e-06, 'learning_rate/action_model': 9.917075831429722e-05, 'epoch': 0.2}) +[2026-06-17 17:17:13,507][starVLA.training.train_starvla][INFO] - Step 512, Loss: {'train/loss': 1.1410839557647705, 'train/grad_norm_pre_clip': 14.34563159942627, 'timing/data': 0.0010938355699181557, 'timing/model': 1.4798839800059795, 'learning_rate/qwen_vl_interface': 9.916673001985339e-06, 'learning_rate/action_model': 9.916673001985338e-05, 'epoch': 0.2}) +[2026-06-17 17:17:17,826][starVLA.training.train_starvla][INFO] - Step 513, Loss: {'train/loss': 0.9084111452102661, 'train/grad_norm_pre_clip': 6.250807762145996, 'timing/data': 0.0011728722602128983, 'timing/model': 1.0779092395678163, 'learning_rate/qwen_vl_interface': 9.91626920513462e-06, 'learning_rate/action_model': 9.916269205134619e-05, 'epoch': 0.2}) +[2026-06-17 17:17:22,635][starVLA.training.train_starvla][INFO] - Step 514, Loss: {'train/loss': 0.6801998019218445, 'train/grad_norm_pre_clip': 5.5780720710754395, 'timing/data': 0.0013155490159988403, 'timing/model': 1.175085368566215, 'learning_rate/qwen_vl_interface': 9.915864440961272e-06, 'learning_rate/action_model': 9.915864440961272e-05, 'epoch': 0.2}) +[2026-06-17 17:17:27,840][starVLA.training.train_starvla][INFO] - Step 515, Loss: {'train/loss': 0.7894797325134277, 'train/grad_norm_pre_clip': 7.962822437286377, 'timing/data': 0.0012268973514437675, 'timing/model': 1.4395621996372938, 'learning_rate/qwen_vl_interface': 9.915458709549205e-06, 'learning_rate/action_model': 9.915458709549206e-05, 'epoch': 0.2}) +[2026-06-17 17:17:34,506][starVLA.training.train_starvla][INFO] - Step 516, Loss: {'train/loss': 0.5828539133071899, 'train/grad_norm_pre_clip': 6.302844047546387, 'timing/data': 0.0017532240599393845, 'timing/model': 1.5748654212802649, 'learning_rate/qwen_vl_interface': 9.91505201098253e-06, 'learning_rate/action_model': 9.91505201098253e-05, 'epoch': 0.2}) +[2026-06-17 17:17:39,433][starVLA.training.train_starvla][INFO] - Step 517, Loss: {'train/loss': 0.6680856943130493, 'train/grad_norm_pre_clip': 6.72130823135376, 'timing/data': 0.0011651599779725075, 'timing/model': 1.4204807644709945, 'learning_rate/qwen_vl_interface': 9.914644345345551e-06, 'learning_rate/action_model': 9.914644345345552e-05, 'epoch': 0.2}) +[2026-06-17 17:17:43,699][starVLA.training.train_starvla][INFO] - Step 518, Loss: {'train/loss': 0.9181272387504578, 'train/grad_norm_pre_clip': 8.155723571777344, 'timing/data': 0.0011137370020151138, 'timing/model': 1.0717397704720497, 'learning_rate/qwen_vl_interface': 9.914235712722782e-06, 'learning_rate/action_model': 9.914235712722781e-05, 'epoch': 0.2}) +[2026-06-17 17:17:48,127][starVLA.training.train_starvla][INFO] - Step 519, Loss: {'train/loss': 0.7837318181991577, 'train/grad_norm_pre_clip': 7.368549346923828, 'timing/data': 0.0011482583358883858, 'timing/model': 1.0636184988543391, 'learning_rate/qwen_vl_interface': 9.913826113198927e-06, 'learning_rate/action_model': 9.913826113198927e-05, 'epoch': 0.2}) +[2026-06-17 17:17:52,890][starVLA.training.train_starvla][INFO] - Step 520, Loss: {'train/loss': 0.7626475095748901, 'train/grad_norm_pre_clip': 10.617695808410645, 'timing/data': 0.0013334471732378006, 'timing/model': 1.1912408228963614, 'learning_rate/qwen_vl_interface': 9.913415546858904e-06, 'learning_rate/action_model': 9.913415546858904e-05, 'epoch': 0.2}) +[2026-06-17 17:17:57,845][starVLA.training.train_starvla][INFO] - Step 521, Loss: {'train/loss': 0.7035449147224426, 'train/grad_norm_pre_clip': 8.563968658447266, 'timing/data': 0.0009849919006228447, 'timing/model': 1.4661933640018106, 'learning_rate/qwen_vl_interface': 9.913004013787818e-06, 'learning_rate/action_model': 9.913004013787819e-05, 'epoch': 0.2}) +[2026-06-17 17:18:03,097][starVLA.training.train_starvla][INFO] - Step 522, Loss: {'train/loss': 0.7590946555137634, 'train/grad_norm_pre_clip': 8.680404663085938, 'timing/data': 0.001400255598127842, 'timing/model': 1.6577250827103853, 'learning_rate/qwen_vl_interface': 9.912591514070983e-06, 'learning_rate/action_model': 9.912591514070983e-05, 'epoch': 0.2}) +[2026-06-17 17:18:07,557][starVLA.training.train_starvla][INFO] - Step 523, Loss: {'train/loss': 0.5793420076370239, 'train/grad_norm_pre_clip': 6.31788444519043, 'timing/data': 0.00124281644821167, 'timing/model': 1.1099062971770763, 'learning_rate/qwen_vl_interface': 9.912178047793912e-06, 'learning_rate/action_model': 9.91217804779391e-05, 'epoch': 0.2}) +[2026-06-17 17:18:18,980][starVLA.training.train_starvla][INFO] - Step 524, Loss: {'train/loss': 0.694378674030304, 'train/grad_norm_pre_clip': 6.212255954742432, 'timing/data': 0.00123573187738657, 'timing/model': 1.3265373334288597, 'learning_rate/qwen_vl_interface': 9.911763615042311e-06, 'learning_rate/action_model': 9.91176361504231e-05, 'epoch': 0.2}) +[2026-06-17 17:18:24,332][starVLA.training.train_starvla][INFO] - Step 525, Loss: {'train/loss': 0.8493850827217102, 'train/grad_norm_pre_clip': 7.511268615722656, 'timing/data': 0.0012371493503451347, 'timing/model': 1.2524956753477454, 'learning_rate/qwen_vl_interface': 9.911348215902099e-06, 'learning_rate/action_model': 9.911348215902099e-05, 'epoch': 0.2}) +[2026-06-17 17:18:29,640][starVLA.training.train_starvla][INFO] - Step 526, Loss: {'train/loss': 0.8690280914306641, 'train/grad_norm_pre_clip': 8.184442520141602, 'timing/data': 0.0012489650398492813, 'timing/model': 1.6407781029120088, 'learning_rate/qwen_vl_interface': 9.910931850459385e-06, 'learning_rate/action_model': 9.910931850459385e-05, 'epoch': 0.2}) +[2026-06-17 17:18:35,555][starVLA.training.train_starvla][INFO] - Step 527, Loss: {'train/loss': 0.6239244341850281, 'train/grad_norm_pre_clip': 8.966612815856934, 'timing/data': 0.0012098168954253197, 'timing/model': 1.5759081104770303, 'learning_rate/qwen_vl_interface': 9.910514518800485e-06, 'learning_rate/action_model': 9.910514518800484e-05, 'epoch': 0.2}) +[2026-06-17 17:18:40,367][starVLA.training.train_starvla][INFO] - Step 528, Loss: {'train/loss': 1.0736937522888184, 'train/grad_norm_pre_clip': 6.663690567016602, 'timing/data': 0.0014063892886042595, 'timing/model': 1.1041968166828156, 'learning_rate/qwen_vl_interface': 9.910096221011908e-06, 'learning_rate/action_model': 9.910096221011907e-05, 'epoch': 0.2}) +[2026-06-17 17:18:46,514][starVLA.training.train_starvla][INFO] - Step 529, Loss: {'train/loss': 0.867857038974762, 'train/grad_norm_pre_clip': 10.305978775024414, 'timing/data': 0.0011986661702394485, 'timing/model': 1.0788542227819562, 'learning_rate/qwen_vl_interface': 9.90967695718037e-06, 'learning_rate/action_model': 9.90967695718037e-05, 'epoch': 0.2}) +[2026-06-17 17:18:53,119][starVLA.training.train_starvla][INFO] - Step 530, Loss: {'train/loss': 0.7592126131057739, 'train/grad_norm_pre_clip': 7.7166266441345215, 'timing/data': 0.0011317236348986626, 'timing/model': 1.3433889588341117, 'learning_rate/qwen_vl_interface': 9.909256727392786e-06, 'learning_rate/action_model': 9.909256727392784e-05, 'epoch': 0.2}) +[2026-06-17 17:18:58,081][starVLA.training.train_starvla][INFO] - Step 531, Loss: {'train/loss': 0.8071184158325195, 'train/grad_norm_pre_clip': 19.600534439086914, 'timing/data': 0.0016847113147377968, 'timing/model': 1.1469201426953077, 'learning_rate/qwen_vl_interface': 9.908835531736269e-06, 'learning_rate/action_model': 9.908835531736267e-05, 'epoch': 0.2}) +[2026-06-17 17:19:02,883][starVLA.training.train_starvla][INFO] - Step 532, Loss: {'train/loss': 0.5311938524246216, 'train/grad_norm_pre_clip': 7.07825231552124, 'timing/data': 0.0012886673212051392, 'timing/model': 1.4121693475171924, 'learning_rate/qwen_vl_interface': 9.908413370298131e-06, 'learning_rate/action_model': 9.90841337029813e-05, 'epoch': 0.2}) +[2026-06-17 17:19:08,519][starVLA.training.train_starvla][INFO] - Step 533, Loss: {'train/loss': 0.7495124936103821, 'train/grad_norm_pre_clip': 9.28459358215332, 'timing/data': 0.001070304773747921, 'timing/model': 2.379970599897206, 'learning_rate/qwen_vl_interface': 9.90799024316589e-06, 'learning_rate/action_model': 9.90799024316589e-05, 'epoch': 0.21}) +[2026-06-17 17:19:14,851][starVLA.training.train_starvla][INFO] - Step 534, Loss: {'train/loss': 0.7478132843971252, 'train/grad_norm_pre_clip': 8.649328231811523, 'timing/data': 0.0011031553149223328, 'timing/model': 1.1471236115321517, 'learning_rate/qwen_vl_interface': 9.90756615042726e-06, 'learning_rate/action_model': 9.907566150427259e-05, 'epoch': 0.21}) +[2026-06-17 17:19:21,244][starVLA.training.train_starvla][INFO] - Step 535, Loss: {'train/loss': 0.8199766874313354, 'train/grad_norm_pre_clip': 29.157779693603516, 'timing/data': 0.0011504106223583221, 'timing/model': 1.1360460566356778, 'learning_rate/qwen_vl_interface': 9.907141092170154e-06, 'learning_rate/action_model': 9.907141092170153e-05, 'epoch': 0.21}) +[2026-06-17 17:19:25,820][starVLA.training.train_starvla][INFO] - Step 536, Loss: {'train/loss': 0.4965658485889435, 'train/grad_norm_pre_clip': 7.2599382400512695, 'timing/data': 0.0008614938706159592, 'timing/model': 1.4359918292611837, 'learning_rate/qwen_vl_interface': 9.906715068482689e-06, 'learning_rate/action_model': 9.90671506848269e-05, 'epoch': 0.21}) +[2026-06-17 17:19:30,323][starVLA.training.train_starvla][INFO] - Step 537, Loss: {'train/loss': 0.9801694750785828, 'train/grad_norm_pre_clip': 8.896217346191406, 'timing/data': 0.0008345469832420349, 'timing/model': 1.3895182758569717, 'learning_rate/qwen_vl_interface': 9.90628807945318e-06, 'learning_rate/action_model': 9.90628807945318e-05, 'epoch': 0.21}) +[2026-06-17 17:19:34,647][starVLA.training.train_starvla][INFO] - Step 538, Loss: {'train/loss': 0.46878376603126526, 'train/grad_norm_pre_clip': 6.472054958343506, 'timing/data': 0.0013026446104049683, 'timing/model': 1.0791746955364943, 'learning_rate/qwen_vl_interface': 9.90586012517014e-06, 'learning_rate/action_model': 9.90586012517014e-05, 'epoch': 0.21}) +[2026-06-17 17:19:39,136][starVLA.training.train_starvla][INFO] - Step 539, Loss: {'train/loss': 0.5529382824897766, 'train/grad_norm_pre_clip': 6.789633274078369, 'timing/data': 0.0011522863060235977, 'timing/model': 1.080682361498475, 'learning_rate/qwen_vl_interface': 9.905431205722288e-06, 'learning_rate/action_model': 9.905431205722288e-05, 'epoch': 0.21}) +[2026-06-17 17:19:43,673][starVLA.training.train_starvla][INFO] - Step 540, Loss: {'train/loss': 1.012758493423462, 'train/grad_norm_pre_clip': 7.734312057495117, 'timing/data': 0.0011562397703528404, 'timing/model': 1.0627594413235784, 'learning_rate/qwen_vl_interface': 9.905001321198536e-06, 'learning_rate/action_model': 9.905001321198536e-05, 'epoch': 0.21}) +[2026-06-17 17:19:48,340][starVLA.training.train_starvla][INFO] - Step 541, Loss: {'train/loss': 0.5390551686286926, 'train/grad_norm_pre_clip': 6.908435821533203, 'timing/data': 0.0011750739067792892, 'timing/model': 1.0788772460073233, 'learning_rate/qwen_vl_interface': 9.904570471688003e-06, 'learning_rate/action_model': 9.904570471688002e-05, 'epoch': 0.21}) +[2026-06-17 17:19:52,852][starVLA.training.train_starvla][INFO] - Step 542, Loss: {'train/loss': 0.6843416690826416, 'train/grad_norm_pre_clip': 5.874638557434082, 'timing/data': 0.0011060405522584915, 'timing/model': 1.419816792011261, 'learning_rate/qwen_vl_interface': 9.90413865728e-06, 'learning_rate/action_model': 9.90413865728e-05, 'epoch': 0.21}) +[2026-06-17 17:19:57,435][starVLA.training.train_starvla][INFO] - Step 543, Loss: {'train/loss': 0.7461613416671753, 'train/grad_norm_pre_clip': 8.213756561279297, 'timing/data': 0.0011682268232107162, 'timing/model': 1.1815646337345243, 'learning_rate/qwen_vl_interface': 9.903705878064047e-06, 'learning_rate/action_model': 9.903705878064046e-05, 'epoch': 0.21}) +[2026-06-17 17:20:02,301][starVLA.training.train_starvla][INFO] - Step 544, Loss: {'train/loss': 0.8348800539970398, 'train/grad_norm_pre_clip': 7.129577159881592, 'timing/data': 0.002378617413341999, 'timing/model': 1.1624248083680868, 'learning_rate/qwen_vl_interface': 9.903272134129858e-06, 'learning_rate/action_model': 9.903272134129858e-05, 'epoch': 0.21}) +[2026-06-17 17:20:06,979][starVLA.training.train_starvla][INFO] - Step 545, Loss: {'train/loss': 1.1737819910049438, 'train/grad_norm_pre_clip': 14.284260749816895, 'timing/data': 0.0010991841554641724, 'timing/model': 1.1154240481555462, 'learning_rate/qwen_vl_interface': 9.90283742556735e-06, 'learning_rate/action_model': 9.902837425567349e-05, 'epoch': 0.21}) +[2026-06-17 17:20:11,856][starVLA.training.train_starvla][INFO] - Step 546, Loss: {'train/loss': 0.4553092122077942, 'train/grad_norm_pre_clip': 5.74735164642334, 'timing/data': 0.0013392344117164612, 'timing/model': 1.415553079918027, 'learning_rate/qwen_vl_interface': 9.902401752466635e-06, 'learning_rate/action_model': 9.902401752466634e-05, 'epoch': 0.21}) +[2026-06-17 17:20:16,665][starVLA.training.train_starvla][INFO] - Step 547, Loss: {'train/loss': 0.6119365692138672, 'train/grad_norm_pre_clip': 7.6964030265808105, 'timing/data': 0.000916249118745327, 'timing/model': 1.4019664619117975, 'learning_rate/qwen_vl_interface': 9.90196511491803e-06, 'learning_rate/action_model': 9.901965114918029e-05, 'epoch': 0.21}) +[2026-06-17 17:20:24,015][starVLA.training.train_starvla][INFO] - Step 548, Loss: {'train/loss': 0.836683988571167, 'train/grad_norm_pre_clip': 8.112998962402344, 'timing/data': 0.0013683121651411057, 'timing/model': 1.07844815030694, 'learning_rate/qwen_vl_interface': 9.901527513012053e-06, 'learning_rate/action_model': 9.901527513012052e-05, 'epoch': 0.21}) +[2026-06-17 17:20:37,192][starVLA.training.train_starvla][INFO] - Step 549, Loss: {'train/loss': 0.5322166085243225, 'train/grad_norm_pre_clip': 5.128423690795898, 'timing/data': 0.0012554526329040527, 'timing/model': 4.244658598676324, 'learning_rate/qwen_vl_interface': 9.901088946839415e-06, 'learning_rate/action_model': 9.901088946839414e-05, 'epoch': 0.21}) +[2026-06-17 17:20:46,512][starVLA.training.train_starvla][INFO] - Step 550, Loss: {'train/loss': 0.7489823698997498, 'train/grad_norm_pre_clip': 6.616180419921875, 'timing/data': 0.0012875311076641083, 'timing/model': 1.0870287669822574, 'learning_rate/qwen_vl_interface': 9.900649416491035e-06, 'learning_rate/action_model': 9.900649416491035e-05, 'epoch': 0.21}) +[2026-06-17 17:20:53,084][starVLA.training.train_starvla][INFO] - Step 551, Loss: {'train/loss': 0.7927590608596802, 'train/grad_norm_pre_clip': 7.6010966300964355, 'timing/data': 0.0009216908365488052, 'timing/model': 1.0696698315441608, 'learning_rate/qwen_vl_interface': 9.900208922058026e-06, 'learning_rate/action_model': 9.900208922058025e-05, 'epoch': 0.21}) +[2026-06-17 17:21:01,468][starVLA.training.train_starvla][INFO] - Step 552, Loss: {'train/loss': 0.7330001592636108, 'train/grad_norm_pre_clip': 8.210989952087402, 'timing/data': 0.0013172607868909836, 'timing/model': 1.7753152213990688, 'learning_rate/qwen_vl_interface': 9.899767463631704e-06, 'learning_rate/action_model': 9.899767463631704e-05, 'epoch': 0.21}) +[2026-06-17 17:21:11,378][starVLA.training.train_starvla][INFO] - Step 553, Loss: {'train/loss': 0.9737299680709839, 'train/grad_norm_pre_clip': 10.149412155151367, 'timing/data': 0.0013453159481287003, 'timing/model': 1.133035533130169, 'learning_rate/qwen_vl_interface': 9.899325041303582e-06, 'learning_rate/action_model': 9.899325041303583e-05, 'epoch': 0.21}) +[2026-06-17 17:21:17,775][starVLA.training.train_starvla][INFO] - Step 554, Loss: {'train/loss': 0.6237303018569946, 'train/grad_norm_pre_clip': 7.3767242431640625, 'timing/data': 0.001096123829483986, 'timing/model': 2.6551265455782413, 'learning_rate/qwen_vl_interface': 9.89888165516538e-06, 'learning_rate/action_model': 9.898881655165379e-05, 'epoch': 0.21}) +[2026-06-17 17:21:23,808][starVLA.training.train_starvla][INFO] - Step 555, Loss: {'train/loss': 0.9759656190872192, 'train/grad_norm_pre_clip': 11.067988395690918, 'timing/data': 0.001341220922768116, 'timing/model': 1.7223162166774273, 'learning_rate/qwen_vl_interface': 9.898437305309004e-06, 'learning_rate/action_model': 9.898437305309003e-05, 'epoch': 0.21}) +[2026-06-17 17:21:28,585][starVLA.training.train_starvla][INFO] - Step 556, Loss: {'train/loss': 0.6728031039237976, 'train/grad_norm_pre_clip': 6.846717357635498, 'timing/data': 0.0013838205486536026, 'timing/model': 1.4701625956222415, 'learning_rate/qwen_vl_interface': 9.897991991826575e-06, 'learning_rate/action_model': 9.897991991826574e-05, 'epoch': 0.21}) +[2026-06-17 17:21:33,458][starVLA.training.train_starvla][INFO] - Step 557, Loss: {'train/loss': 0.9669480323791504, 'train/grad_norm_pre_clip': 9.975956916809082, 'timing/data': 0.0022175386548042297, 'timing/model': 1.4045383417978883, 'learning_rate/qwen_vl_interface': 9.897545714810404e-06, 'learning_rate/action_model': 9.897545714810403e-05, 'epoch': 0.21}) +[2026-06-17 17:21:38,311][starVLA.training.train_starvla][INFO] - Step 558, Loss: {'train/loss': 1.161097526550293, 'train/grad_norm_pre_clip': 11.346793174743652, 'timing/data': 0.0012381169945001602, 'timing/model': 1.3055257583037019, 'learning_rate/qwen_vl_interface': 9.897098474353004e-06, 'learning_rate/action_model': 9.897098474353004e-05, 'epoch': 0.21}) +[2026-06-17 17:21:43,078][starVLA.training.train_starvla][INFO] - Step 559, Loss: {'train/loss': 0.68631911277771, 'train/grad_norm_pre_clip': 18.504764556884766, 'timing/data': 0.0011460520327091217, 'timing/model': 1.1512292921543121, 'learning_rate/qwen_vl_interface': 9.896650270547092e-06, 'learning_rate/action_model': 9.89665027054709e-05, 'epoch': 0.22}) +[2026-06-17 17:21:47,851][starVLA.training.train_starvla][INFO] - Step 560, Loss: {'train/loss': 0.8955492973327637, 'train/grad_norm_pre_clip': 7.810861587524414, 'timing/data': 0.0009490028023719788, 'timing/model': 1.179971195757389, 'learning_rate/qwen_vl_interface': 9.896201103485577e-06, 'learning_rate/action_model': 9.896201103485577e-05, 'epoch': 0.22}) +[2026-06-17 17:21:52,681][starVLA.training.train_starvla][INFO] - Step 561, Loss: {'train/loss': 1.1649129390716553, 'train/grad_norm_pre_clip': 15.504873275756836, 'timing/data': 0.0010998332872986794, 'timing/model': 1.1665339935570955, 'learning_rate/qwen_vl_interface': 9.895750973261576e-06, 'learning_rate/action_model': 9.895750973261574e-05, 'epoch': 0.22}) +[2026-06-17 17:21:57,697][starVLA.training.train_starvla][INFO] - Step 562, Loss: {'train/loss': 0.5962474942207336, 'train/grad_norm_pre_clip': 10.180155754089355, 'timing/data': 0.0013574771583080292, 'timing/model': 1.4719484029337764, 'learning_rate/qwen_vl_interface': 9.895299879968398e-06, 'learning_rate/action_model': 9.895299879968398e-05, 'epoch': 0.22}) +[2026-06-17 17:22:01,997][starVLA.training.train_starvla][INFO] - Step 563, Loss: {'train/loss': 0.6951168179512024, 'train/grad_norm_pre_clip': 6.724837779998779, 'timing/data': 0.0011986680328845978, 'timing/model': 1.0811420315876603, 'learning_rate/qwen_vl_interface': 9.894847823699555e-06, 'learning_rate/action_model': 9.894847823699555e-05, 'epoch': 0.22}) +[2026-06-17 17:22:06,651][starVLA.training.train_starvla][INFO] - Step 564, Loss: {'train/loss': 0.6868123412132263, 'train/grad_norm_pre_clip': 5.689421653747559, 'timing/data': 0.0014722300693392754, 'timing/model': 1.092947960831225, 'learning_rate/qwen_vl_interface': 9.894394804548763e-06, 'learning_rate/action_model': 9.894394804548762e-05, 'epoch': 0.22}) +[2026-06-17 17:22:11,478][starVLA.training.train_starvla][INFO] - Step 565, Loss: {'train/loss': 0.6470766067504883, 'train/grad_norm_pre_clip': 7.024612903594971, 'timing/data': 0.0012046759948134422, 'timing/model': 1.2311999276280403, 'learning_rate/qwen_vl_interface': 9.893940822609928e-06, 'learning_rate/action_model': 9.893940822609927e-05, 'epoch': 0.22}) +[2026-06-17 17:22:16,155][starVLA.training.train_starvla][INFO] - Step 566, Loss: {'train/loss': 1.1156432628631592, 'train/grad_norm_pre_clip': 14.905193328857422, 'timing/data': 0.0011182865127921104, 'timing/model': 1.4076034342870116, 'learning_rate/qwen_vl_interface': 9.893485877977163e-06, 'learning_rate/action_model': 9.893485877977163e-05, 'epoch': 0.22}) +[2026-06-17 17:22:20,627][starVLA.training.train_starvla][INFO] - Step 567, Loss: {'train/loss': 0.738548994064331, 'train/grad_norm_pre_clip': 5.204103469848633, 'timing/data': 0.0012163110077381134, 'timing/model': 1.363959084264934, 'learning_rate/qwen_vl_interface': 9.89302997074478e-06, 'learning_rate/action_model': 9.89302997074478e-05, 'epoch': 0.22}) +[2026-06-17 17:22:24,928][starVLA.training.train_starvla][INFO] - Step 568, Loss: {'train/loss': 0.7689584493637085, 'train/grad_norm_pre_clip': 7.149752616882324, 'timing/data': 0.0008295634761452675, 'timing/model': 1.084252349101007, 'learning_rate/qwen_vl_interface': 9.892573101007288e-06, 'learning_rate/action_model': 9.892573101007289e-05, 'epoch': 0.22}) +[2026-06-17 17:22:29,525][starVLA.training.train_starvla][INFO] - Step 569, Loss: {'train/loss': 1.1764060258865356, 'train/grad_norm_pre_clip': 13.483332633972168, 'timing/data': 0.0011760164052248001, 'timing/model': 1.11598698515445, 'learning_rate/qwen_vl_interface': 9.892115268859398e-06, 'learning_rate/action_model': 9.892115268859397e-05, 'epoch': 0.22}) +[2026-06-17 17:22:40,636][starVLA.training.train_starvla][INFO] - Step 570, Loss: {'train/loss': 0.6693464517593384, 'train/grad_norm_pre_clip': 6.719275951385498, 'timing/data': 0.0012749852612614632, 'timing/model': 5.544245311059058, 'learning_rate/qwen_vl_interface': 9.891656474396015e-06, 'learning_rate/action_model': 9.891656474396015e-05, 'epoch': 0.22}) +[2026-06-17 17:22:47,432][starVLA.training.train_starvla][INFO] - Step 571, Loss: {'train/loss': 0.9745526909828186, 'train/grad_norm_pre_clip': 11.11583137512207, 'timing/data': 0.02269687596708536, 'timing/model': 1.2063095169141889, 'learning_rate/qwen_vl_interface': 9.891196717712253e-06, 'learning_rate/action_model': 9.891196717712252e-05, 'epoch': 0.22}) +[2026-06-17 17:22:53,542][starVLA.training.train_starvla][INFO] - Step 572, Loss: {'train/loss': 0.8983180522918701, 'train/grad_norm_pre_clip': 7.964481830596924, 'timing/data': 0.01949450746178627, 'timing/model': 1.4707203144207597, 'learning_rate/qwen_vl_interface': 9.890735998903414e-06, 'learning_rate/action_model': 9.890735998903414e-05, 'epoch': 0.22}) +[2026-06-17 17:23:08,100][starVLA.training.train_starvla][INFO] - Step 573, Loss: {'train/loss': 1.0625003576278687, 'train/grad_norm_pre_clip': 7.1614670753479, 'timing/data': 0.021947751753032207, 'timing/model': 4.355457233265042, 'learning_rate/qwen_vl_interface': 9.890274318065012e-06, 'learning_rate/action_model': 9.890274318065011e-05, 'epoch': 0.22}) +[2026-06-17 17:23:22,066][starVLA.training.train_starvla][INFO] - Step 574, Loss: {'train/loss': 0.8757986426353455, 'train/grad_norm_pre_clip': 6.922802925109863, 'timing/data': 0.022031637839972973, 'timing/model': 4.966507437638938, 'learning_rate/qwen_vl_interface': 9.889811675292748e-06, 'learning_rate/action_model': 9.889811675292749e-05, 'epoch': 0.22}) +[2026-06-17 17:23:40,343][starVLA.training.train_starvla][INFO] - Step 575, Loss: {'train/loss': 0.7491977214813232, 'train/grad_norm_pre_clip': 5.053128719329834, 'timing/data': 0.022153040394186974, 'timing/model': 5.288141781464219, 'learning_rate/qwen_vl_interface': 9.889348070682533e-06, 'learning_rate/action_model': 9.889348070682533e-05, 'epoch': 0.22}) +[2026-06-17 17:23:52,906][starVLA.training.train_starvla][INFO] - Step 576, Loss: {'train/loss': 0.7387268543243408, 'train/grad_norm_pre_clip': 6.420742511749268, 'timing/data': 0.0380009189248085, 'timing/model': 4.668890924192965, 'learning_rate/qwen_vl_interface': 9.888883504330469e-06, 'learning_rate/action_model': 9.888883504330468e-05, 'epoch': 0.22}) +[2026-06-17 17:23:59,642][starVLA.training.train_starvla][INFO] - Step 577, Loss: {'train/loss': 0.714123547077179, 'train/grad_norm_pre_clip': 4.654113292694092, 'timing/data': 0.029052620753645897, 'timing/model': 1.9147961363196373, 'learning_rate/qwen_vl_interface': 9.888417976332865e-06, 'learning_rate/action_model': 9.888417976332865e-05, 'epoch': 0.22}) +[2026-06-17 17:24:04,271][starVLA.training.train_starvla][INFO] - Step 578, Loss: {'train/loss': 0.49892887473106384, 'train/grad_norm_pre_clip': 7.402191638946533, 'timing/data': 0.023531094193458557, 'timing/model': 1.3745101122185588, 'learning_rate/qwen_vl_interface': 9.887951486786222e-06, 'learning_rate/action_model': 9.887951486786221e-05, 'epoch': 0.22}) +[2026-06-17 17:24:09,491][starVLA.training.train_starvla][INFO] - Step 579, Loss: {'train/loss': 0.6806455254554749, 'train/grad_norm_pre_clip': 12.502257347106934, 'timing/data': 0.036213082261383533, 'timing/model': 1.3386464836075902, 'learning_rate/qwen_vl_interface': 9.887484035787246e-06, 'learning_rate/action_model': 9.887484035787246e-05, 'epoch': 0.22}) +[2026-06-17 17:24:14,816][starVLA.training.train_starvla][INFO] - Step 580, Loss: {'train/loss': 1.0071508884429932, 'train/grad_norm_pre_clip': 7.5452775955200195, 'timing/data': 0.025203709490597248, 'timing/model': 1.3286991063505411, 'learning_rate/qwen_vl_interface': 9.887015623432837e-06, 'learning_rate/action_model': 9.887015623432837e-05, 'epoch': 0.22}) +[2026-06-17 17:24:19,944][starVLA.training.train_starvla][INFO] - Step 581, Loss: {'train/loss': 0.7223568558692932, 'train/grad_norm_pre_clip': 7.744753360748291, 'timing/data': 0.02272997424006462, 'timing/model': 1.2861123764887452, 'learning_rate/qwen_vl_interface': 9.886546249820101e-06, 'learning_rate/action_model': 9.886546249820102e-05, 'epoch': 0.22}) +[2026-06-17 17:24:24,599][starVLA.training.train_starvla][INFO] - Step 582, Loss: {'train/loss': 0.9106394052505493, 'train/grad_norm_pre_clip': 6.502991199493408, 'timing/data': 0.02170910406857729, 'timing/model': 1.4427596256136894, 'learning_rate/qwen_vl_interface': 9.886075915046339e-06, 'learning_rate/action_model': 9.886075915046338e-05, 'epoch': 0.22}) +[2026-06-17 17:24:29,120][starVLA.training.train_starvla][INFO] - Step 583, Loss: {'train/loss': 0.7987818717956543, 'train/grad_norm_pre_clip': 8.071556091308594, 'timing/data': 0.020513678900897503, 'timing/model': 1.2416173024103045, 'learning_rate/qwen_vl_interface': 9.885604619209052e-06, 'learning_rate/action_model': 9.885604619209051e-05, 'epoch': 0.22}) +[2026-06-17 17:24:38,188][starVLA.training.train_starvla][INFO] - Step 584, Loss: {'train/loss': 0.6583800911903381, 'train/grad_norm_pre_clip': 5.010773181915283, 'timing/data': 0.023760302923619747, 'timing/model': 5.467690444551408, 'learning_rate/qwen_vl_interface': 9.885132362405935e-06, 'learning_rate/action_model': 9.885132362405936e-05, 'epoch': 0.22}) +[2026-06-17 17:24:43,977][starVLA.training.train_starvla][INFO] - Step 585, Loss: {'train/loss': 0.7304143905639648, 'train/grad_norm_pre_clip': 16.00806427001953, 'timing/data': 0.020099462941288948, 'timing/model': 1.1250226506963372, 'learning_rate/qwen_vl_interface': 9.884659144734896e-06, 'learning_rate/action_model': 9.884659144734896e-05, 'epoch': 0.23}) +[2026-06-17 17:25:00,059][starVLA.training.train_starvla][INFO] - Step 586, Loss: {'train/loss': 0.5513424277305603, 'train/grad_norm_pre_clip': 7.854249954223633, 'timing/data': 0.02285223826766014, 'timing/model': 1.1883658524602652, 'learning_rate/qwen_vl_interface': 9.884184966294028e-06, 'learning_rate/action_model': 9.884184966294028e-05, 'epoch': 0.23}) +[2026-06-17 17:25:04,998][starVLA.training.train_starvla][INFO] - Step 587, Loss: {'train/loss': 0.7349247932434082, 'train/grad_norm_pre_clip': 7.839886665344238, 'timing/data': 0.02529391460120678, 'timing/model': 1.5099929412826896, 'learning_rate/qwen_vl_interface': 9.883709827181632e-06, 'learning_rate/action_model': 9.88370982718163e-05, 'epoch': 0.23}) +[2026-06-17 17:25:09,776][starVLA.training.train_starvla][INFO] - Step 588, Loss: {'train/loss': 0.892346978187561, 'train/grad_norm_pre_clip': 9.860353469848633, 'timing/data': 0.037537889555096626, 'timing/model': 1.2760997973382473, 'learning_rate/qwen_vl_interface': 9.8832337274962e-06, 'learning_rate/action_model': 9.883233727496198e-05, 'epoch': 0.23}) +[2026-06-17 17:25:19,068][starVLA.training.train_starvla][INFO] - Step 589, Loss: {'train/loss': 0.7318708300590515, 'train/grad_norm_pre_clip': 8.596677780151367, 'timing/data': 0.021682842634618282, 'timing/model': 1.3592012859880924, 'learning_rate/qwen_vl_interface': 9.88275666733643e-06, 'learning_rate/action_model': 9.882756667336429e-05, 'epoch': 0.23}) +[2026-06-17 17:25:27,795][starVLA.training.train_starvla][INFO] - Step 590, Loss: {'train/loss': 0.6787338256835938, 'train/grad_norm_pre_clip': 5.894908428192139, 'timing/data': 0.022221820428967476, 'timing/model': 1.318303200416267, 'learning_rate/qwen_vl_interface': 9.88227864680122e-06, 'learning_rate/action_model': 9.88227864680122e-05, 'epoch': 0.23}) +[2026-06-17 17:25:32,763][starVLA.training.train_starvla][INFO] - Step 591, Loss: {'train/loss': 0.7314244508743286, 'train/grad_norm_pre_clip': 6.837931156158447, 'timing/data': 0.024424677714705467, 'timing/model': 1.1062205862253904, 'learning_rate/qwen_vl_interface': 9.88179966598966e-06, 'learning_rate/action_model': 9.88179966598966e-05, 'epoch': 0.23}) +[2026-06-17 17:25:39,103][starVLA.training.train_starvla][INFO] - Step 592, Loss: {'train/loss': 0.7800613045692444, 'train/grad_norm_pre_clip': 9.397436141967773, 'timing/data': 0.022711114026606083, 'timing/model': 1.5842440277338028, 'learning_rate/qwen_vl_interface': 9.881319725001046e-06, 'learning_rate/action_model': 9.881319725001046e-05, 'epoch': 0.23}) +[2026-06-17 17:25:43,712][starVLA.training.train_starvla][INFO] - Step 593, Loss: {'train/loss': 0.6815464496612549, 'train/grad_norm_pre_clip': 8.133036613464355, 'timing/data': 0.020955950021743774, 'timing/model': 1.089819529093802, 'learning_rate/qwen_vl_interface': 9.88083882393487e-06, 'learning_rate/action_model': 9.880838823934869e-05, 'epoch': 0.23}) +[2026-06-17 17:25:48,593][starVLA.training.train_starvla][INFO] - Step 594, Loss: {'train/loss': 0.7673183083534241, 'train/grad_norm_pre_clip': 5.77830696105957, 'timing/data': 0.0012125177308917046, 'timing/model': 1.1614255849272013, 'learning_rate/qwen_vl_interface': 9.880356962890823e-06, 'learning_rate/action_model': 9.880356962890822e-05, 'epoch': 0.23}) +[2026-06-17 17:25:53,708][starVLA.training.train_starvla][INFO] - Step 595, Loss: {'train/loss': 0.5733508467674255, 'train/grad_norm_pre_clip': 7.035711288452148, 'timing/data': 0.0030854959040880203, 'timing/model': 1.1532454788684845, 'learning_rate/qwen_vl_interface': 9.879874141968792e-06, 'learning_rate/action_model': 9.879874141968792e-05, 'epoch': 0.23}) +[2026-06-17 17:25:58,273][starVLA.training.train_starvla][INFO] - Step 596, Loss: {'train/loss': 0.6660922169685364, 'train/grad_norm_pre_clip': 5.856125354766846, 'timing/data': 0.0013387510553002357, 'timing/model': 1.1021431228145957, 'learning_rate/qwen_vl_interface': 9.879390361268873e-06, 'learning_rate/action_model': 9.879390361268873e-05, 'epoch': 0.23}) +[2026-06-17 17:26:03,013][starVLA.training.train_starvla][INFO] - Step 597, Loss: {'train/loss': 0.5279538631439209, 'train/grad_norm_pre_clip': 5.514115333557129, 'timing/data': 0.0014484301209449768, 'timing/model': 1.4685633480548859, 'learning_rate/qwen_vl_interface': 9.878905620891349e-06, 'learning_rate/action_model': 9.878905620891349e-05, 'epoch': 0.23}) +[2026-06-17 17:26:07,458][starVLA.training.train_starvla][INFO] - Step 598, Loss: {'train/loss': 0.783392071723938, 'train/grad_norm_pre_clip': 7.156315803527832, 'timing/data': 0.0014987168833613396, 'timing/model': 1.1053402591496706, 'learning_rate/qwen_vl_interface': 9.87841992093671e-06, 'learning_rate/action_model': 9.87841992093671e-05, 'epoch': 0.23}) +[2026-06-17 17:26:12,094][starVLA.training.train_starvla][INFO] - Step 599, Loss: {'train/loss': 0.8472829461097717, 'train/grad_norm_pre_clip': 6.054590225219727, 'timing/data': 0.001172768883407116, 'timing/model': 1.0680724009871483, 'learning_rate/qwen_vl_interface': 9.877933261505642e-06, 'learning_rate/action_model': 9.877933261505641e-05, 'epoch': 0.23}) +[2026-06-17 17:29:40,912][starVLA.training.train_starvla][INFO] - Step 600, Loss: {'train/loss': 0.5460842847824097, 'train/grad_norm_pre_clip': 12.535909652709961, 'eval/loss': 0.6971498727798462, 'eval/latency_4/loss': 0.6971498727798462, 'eval/demon_fire_precision': 0.950788643533123, 'eval/demon_fire_recall': 0.9788892497564144, 'eval/demon_fire_f1': 0.9646343414946391, 'eval/demon_fire_rate_model': 0.9309838472834068, 'eval/demon_fire_rate_teacher': 0.9042584434654919, 'eval/demon_left_precision': 0.6531986531986532, 'eval/demon_left_recall': 0.4429223744292237, 'eval/demon_left_f1': 0.527891156462585, 'eval/demon_left_rate_model': 0.08722466960352424, 'eval/demon_left_rate_teacher': 0.12863436123348018, 'eval/demon_right_precision': 0.8430493273542601, 'eval/demon_right_recall': 0.33935018050541516, 'eval/demon_right_f1': 0.48391248391248387, 'eval/demon_right_rate_model': 0.06549192364170338, 'eval/demon_right_rate_teacher': 0.16270190895741557, 'eval/demon_move_f1': 0.5059018201875345, 'eval/demon_cc_f1': 0.7583432893910552, 'eval/latency_4/demon_fire_precision': 0.950788643533123, 'eval/latency_4/demon_fire_recall': 0.9788892497564144, 'eval/latency_4/demon_fire_f1': 0.9646343414946391, 'eval/latency_4/demon_fire_rate_model': 0.9309838472834068, 'eval/latency_4/demon_fire_rate_teacher': 0.9042584434654919, 'eval/latency_4/demon_left_precision': 0.6531986531986532, 'eval/latency_4/demon_left_recall': 0.4429223744292237, 'eval/latency_4/demon_left_f1': 0.527891156462585, 'eval/latency_4/demon_left_rate_model': 0.08722466960352424, 'eval/latency_4/demon_left_rate_teacher': 0.12863436123348018, 'eval/latency_4/demon_right_precision': 0.8430493273542601, 'eval/latency_4/demon_right_recall': 0.33935018050541516, 'eval/latency_4/demon_right_f1': 0.48391248391248387, 'eval/latency_4/demon_right_rate_model': 0.06549192364170338, 'eval/latency_4/demon_right_rate_teacher': 0.16270190895741557, 'eval/latency_4/demon_move_f1': 0.5059018201875345, 'eval/latency_4/demon_cc_f1': 0.7583432893910552, 'timing/data': 0.0012351097539067268, 'timing/model': 1.0781510528177023, 'learning_rate/qwen_vl_interface': 9.877445642699026e-06, 'learning_rate/action_model': 9.877445642699026e-05, 'epoch': 0.23}) +[2026-06-17 17:29:40,914][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_600_state +[2026-06-17 17:29:40,922][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 17:29:49,880][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_600_state/model.safetensors +[2026-06-17 17:30:05,003][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_600_state/optimizer.bin +[2026-06-17 17:30:05,005][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_600_state/sampler.bin +[2026-06-17 17:30:05,005][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_600_state/sampler_1.bin +[2026-06-17 17:30:05,007][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_600_state/random_states_0.pkl +[2026-06-17 17:30:05,013][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 600 +[2026-06-17 17:30:05,014][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 17:30:05,031][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 17:30:10,100][starVLA.training.train_starvla][INFO] - Step 601, Loss: {'train/loss': 0.33918678760528564, 'train/grad_norm_pre_clip': 4.653416633605957, 'timing/data': 0.0012774830684065819, 'timing/model': 1.7964303474873304, 'learning_rate/qwen_vl_interface': 9.876957064617951e-06, 'learning_rate/action_model': 9.876957064617951e-05, 'epoch': 0.23}) +[2026-06-17 17:30:14,502][starVLA.training.train_starvla][INFO] - Step 602, Loss: {'train/loss': 0.5998074412345886, 'train/grad_norm_pre_clip': 5.665029048919678, 'timing/data': 0.0012396657839417458, 'timing/model': 1.0747538935393095, 'learning_rate/qwen_vl_interface': 9.876467527363696e-06, 'learning_rate/action_model': 9.876467527363697e-05, 'epoch': 0.23}) +[2026-06-17 17:30:19,513][starVLA.training.train_starvla][INFO] - Step 603, Loss: {'train/loss': 0.6847964525222778, 'train/grad_norm_pre_clip': 9.717577934265137, 'timing/data': 0.001780693419277668, 'timing/model': 1.0776859177276492, 'learning_rate/qwen_vl_interface': 9.875977031037746e-06, 'learning_rate/action_model': 9.875977031037745e-05, 'epoch': 0.23}) +[2026-06-17 17:30:24,403][starVLA.training.train_starvla][INFO] - Step 604, Loss: {'train/loss': 0.8601392507553101, 'train/grad_norm_pre_clip': 6.026032447814941, 'timing/data': 0.0010858550667762756, 'timing/model': 1.071593152359128, 'learning_rate/qwen_vl_interface': 9.875485575741779e-06, 'learning_rate/action_model': 9.875485575741779e-05, 'epoch': 0.23}) +[2026-06-17 17:30:29,282][starVLA.training.train_starvla][INFO] - Step 605, Loss: {'train/loss': 0.7369330525398254, 'train/grad_norm_pre_clip': 11.34415054321289, 'timing/data': 0.0009706132113933563, 'timing/model': 1.1428549280390143, 'learning_rate/qwen_vl_interface': 9.874993161577674e-06, 'learning_rate/action_model': 9.874993161577675e-05, 'epoch': 0.23}) +[2026-06-17 17:30:35,167][starVLA.training.train_starvla][INFO] - Step 606, Loss: {'train/loss': 0.6739900708198547, 'train/grad_norm_pre_clip': 14.134760856628418, 'timing/data': 0.001523778773844242, 'timing/model': 1.9181456286460161, 'learning_rate/qwen_vl_interface': 9.874499788647512e-06, 'learning_rate/action_model': 9.874499788647513e-05, 'epoch': 0.23}) +[2026-06-17 17:30:40,167][starVLA.training.train_starvla][INFO] - Step 607, Loss: {'train/loss': 0.6307492852210999, 'train/grad_norm_pre_clip': 7.473176002502441, 'timing/data': 0.0011176364496350288, 'timing/model': 1.476954696699977, 'learning_rate/qwen_vl_interface': 9.874005457053565e-06, 'learning_rate/action_model': 9.874005457053565e-05, 'epoch': 0.23}) +[2026-06-17 17:30:44,566][starVLA.training.train_starvla][INFO] - Step 608, Loss: {'train/loss': 0.8332240581512451, 'train/grad_norm_pre_clip': 7.06917667388916, 'timing/data': 0.0011878442019224167, 'timing/model': 1.0894945105537772, 'learning_rate/qwen_vl_interface': 9.873510166898315e-06, 'learning_rate/action_model': 9.873510166898314e-05, 'epoch': 0.23}) +[2026-06-17 17:30:49,537][starVLA.training.train_starvla][INFO] - Step 609, Loss: {'train/loss': 0.8376331329345703, 'train/grad_norm_pre_clip': 9.919326782226562, 'timing/data': 0.0012165233492851257, 'timing/model': 1.2682748446241021, 'learning_rate/qwen_vl_interface': 9.87301391828443e-06, 'learning_rate/action_model': 9.873013918284428e-05, 'epoch': 0.23}) +[2026-06-17 17:31:03,796][starVLA.training.train_starvla][INFO] - Step 610, Loss: {'train/loss': 0.8064548969268799, 'train/grad_norm_pre_clip': 9.010030746459961, 'timing/data': 0.0009700097143650055, 'timing/model': 1.5677491938695312, 'learning_rate/qwen_vl_interface': 9.872516711314784e-06, 'learning_rate/action_model': 9.872516711314784e-05, 'epoch': 0.23}) +[2026-06-17 17:31:09,082][starVLA.training.train_starvla][INFO] - Step 611, Loss: {'train/loss': 0.8095687627792358, 'train/grad_norm_pre_clip': 10.185019493103027, 'timing/data': 0.001341518945991993, 'timing/model': 1.2059581726789474, 'learning_rate/qwen_vl_interface': 9.872018546092452e-06, 'learning_rate/action_model': 9.872018546092451e-05, 'epoch': 0.24}) +[2026-06-17 17:31:14,074][starVLA.training.train_starvla][INFO] - Step 612, Loss: {'train/loss': 0.744631826877594, 'train/grad_norm_pre_clip': 5.631694316864014, 'timing/data': 0.001363488845527172, 'timing/model': 1.484731557779014, 'learning_rate/qwen_vl_interface': 9.871519422720699e-06, 'learning_rate/action_model': 9.871519422720698e-05, 'epoch': 0.24}) +[2026-06-17 17:31:18,614][starVLA.training.train_starvla][INFO] - Step 613, Loss: {'train/loss': 0.7562205791473389, 'train/grad_norm_pre_clip': 6.537240505218506, 'timing/data': 0.0008723167702555656, 'timing/model': 1.0642460510134697, 'learning_rate/qwen_vl_interface': 9.871019341303e-06, 'learning_rate/action_model': 9.871019341303e-05, 'epoch': 0.24}) +[2026-06-17 17:31:23,242][starVLA.training.train_starvla][INFO] - Step 614, Loss: {'train/loss': 0.6104472875595093, 'train/grad_norm_pre_clip': 9.944883346557617, 'timing/data': 0.0008693179115653038, 'timing/model': 1.1034480473026633, 'learning_rate/qwen_vl_interface': 9.870518301943018e-06, 'learning_rate/action_model': 9.870518301943018e-05, 'epoch': 0.24}) +[2026-06-17 17:31:28,252][starVLA.training.train_starvla][INFO] - Step 615, Loss: {'train/loss': 0.5350400805473328, 'train/grad_norm_pre_clip': 8.805628776550293, 'timing/data': 0.0011553717777132988, 'timing/model': 1.1879554465413094, 'learning_rate/qwen_vl_interface': 9.870016304744617e-06, 'learning_rate/action_model': 9.870016304744616e-05, 'epoch': 0.24}) +[2026-06-17 17:31:32,964][starVLA.training.train_starvla][INFO] - Step 616, Loss: {'train/loss': 0.6333713531494141, 'train/grad_norm_pre_clip': 6.385093688964844, 'timing/data': 0.001336689107120037, 'timing/model': 1.1934227058663964, 'learning_rate/qwen_vl_interface': 9.869513349811868e-06, 'learning_rate/action_model': 9.869513349811868e-05, 'epoch': 0.24}) +[2026-06-17 17:31:37,427][starVLA.training.train_starvla][INFO] - Step 617, Loss: {'train/loss': 0.7771221995353699, 'train/grad_norm_pre_clip': 10.159205436706543, 'timing/data': 0.0011661592870950699, 'timing/model': 1.1281104013323784, 'learning_rate/qwen_vl_interface': 9.86900943724903e-06, 'learning_rate/action_model': 9.869009437249031e-05, 'epoch': 0.24}) +[2026-06-17 17:31:42,292][starVLA.training.train_starvla][INFO] - Step 618, Loss: {'train/loss': 0.8440121412277222, 'train/grad_norm_pre_clip': 11.113286972045898, 'timing/data': 0.0011519333347678185, 'timing/model': 1.2324065919965506, 'learning_rate/qwen_vl_interface': 9.868504567160566e-06, 'learning_rate/action_model': 9.868504567160565e-05, 'epoch': 0.24}) +[2026-06-17 17:31:47,268][starVLA.training.train_starvla][INFO] - Step 619, Loss: {'train/loss': 0.40762802958488464, 'train/grad_norm_pre_clip': 5.626038074493408, 'timing/data': 0.0010360842570662498, 'timing/model': 1.1499950904399157, 'learning_rate/qwen_vl_interface': 9.867998739651134e-06, 'learning_rate/action_model': 9.867998739651134e-05, 'epoch': 0.24}) +[2026-06-17 17:31:52,003][starVLA.training.train_starvla][INFO] - Step 620, Loss: {'train/loss': 0.8209741115570068, 'train/grad_norm_pre_clip': 11.010672569274902, 'timing/data': 0.0012328028678894043, 'timing/model': 1.1408871449530125, 'learning_rate/qwen_vl_interface': 9.867491954825595e-06, 'learning_rate/action_model': 9.867491954825595e-05, 'epoch': 0.24}) +[2026-06-17 17:32:00,628][starVLA.training.train_starvla][INFO] - Step 621, Loss: {'train/loss': 0.8286721706390381, 'train/grad_norm_pre_clip': 8.192201614379883, 'timing/data': 0.0012521510943770409, 'timing/model': 1.1026202570647001, 'learning_rate/qwen_vl_interface': 9.866984212789007e-06, 'learning_rate/action_model': 9.866984212789006e-05, 'epoch': 0.24}) +[2026-06-17 17:32:05,408][starVLA.training.train_starvla][INFO] - Step 622, Loss: {'train/loss': 0.7845149040222168, 'train/grad_norm_pre_clip': 7.636502742767334, 'timing/data': 0.001210012473165989, 'timing/model': 1.6317522330209613, 'learning_rate/qwen_vl_interface': 9.86647551364662e-06, 'learning_rate/action_model': 9.86647551364662e-05, 'epoch': 0.24}) +[2026-06-17 17:32:10,750][starVLA.training.train_starvla][INFO] - Step 623, Loss: {'train/loss': 0.6368141770362854, 'train/grad_norm_pre_clip': 6.956721782684326, 'timing/data': 0.001185101456940174, 'timing/model': 1.8561867605894804, 'learning_rate/qwen_vl_interface': 9.865965857503893e-06, 'learning_rate/action_model': 9.865965857503894e-05, 'epoch': 0.24}) +[2026-06-17 17:32:16,057][starVLA.training.train_starvla][INFO] - Step 624, Loss: {'train/loss': 0.7348265647888184, 'train/grad_norm_pre_clip': 7.669505596160889, 'timing/data': 0.0012704329565167427, 'timing/model': 1.300186817534268, 'learning_rate/qwen_vl_interface': 9.865455244466478e-06, 'learning_rate/action_model': 9.865455244466477e-05, 'epoch': 0.24}) +[2026-06-17 17:32:25,660][starVLA.training.train_starvla][INFO] - Step 625, Loss: {'train/loss': 0.7443555593490601, 'train/grad_norm_pre_clip': 8.264183044433594, 'timing/data': 0.0024433229118585587, 'timing/model': 2.5659763338044286, 'learning_rate/qwen_vl_interface': 9.864943674640223e-06, 'learning_rate/action_model': 9.864943674640222e-05, 'epoch': 0.24}) +[2026-06-17 17:32:31,365][starVLA.training.train_starvla][INFO] - Step 626, Loss: {'train/loss': 1.1431097984313965, 'train/grad_norm_pre_clip': 8.132094383239746, 'timing/data': 0.0011958787217736244, 'timing/model': 1.4150669099763036, 'learning_rate/qwen_vl_interface': 9.864431148131178e-06, 'learning_rate/action_model': 9.864431148131178e-05, 'epoch': 0.24}) +[2026-06-17 17:32:36,220][starVLA.training.train_starvla][INFO] - Step 627, Loss: {'train/loss': 0.9454618096351624, 'train/grad_norm_pre_clip': 11.900691986083984, 'timing/data': 0.0011864285916090012, 'timing/model': 1.5016672601923347, 'learning_rate/qwen_vl_interface': 9.863917665045591e-06, 'learning_rate/action_model': 9.863917665045592e-05, 'epoch': 0.24}) +[2026-06-17 17:32:43,390][starVLA.training.train_starvla][INFO] - Step 628, Loss: {'train/loss': 0.9338282346725464, 'train/grad_norm_pre_clip': 9.56816291809082, 'timing/data': 0.0012009544298052788, 'timing/model': 2.8275222713127732, 'learning_rate/qwen_vl_interface': 9.863403225489909e-06, 'learning_rate/action_model': 9.863403225489909e-05, 'epoch': 0.24}) +[2026-06-17 17:32:48,234][starVLA.training.train_starvla][INFO] - Step 629, Loss: {'train/loss': 0.8876628875732422, 'train/grad_norm_pre_clip': 7.573631763458252, 'timing/data': 0.0012525701895356178, 'timing/model': 1.1287244129925966, 'learning_rate/qwen_vl_interface': 9.862887829570771e-06, 'learning_rate/action_model': 9.862887829570772e-05, 'epoch': 0.24}) +[2026-06-17 17:32:57,382][starVLA.training.train_starvla][INFO] - Step 630, Loss: {'train/loss': 0.9169740676879883, 'train/grad_norm_pre_clip': 6.318690299987793, 'timing/data': 0.001151735894382, 'timing/model': 1.4046188779175282, 'learning_rate/qwen_vl_interface': 9.862371477395024e-06, 'learning_rate/action_model': 9.862371477395025e-05, 'epoch': 0.24}) +[2026-06-17 17:33:05,234][starVLA.training.train_starvla][INFO] - Step 631, Loss: {'train/loss': 0.9069738388061523, 'train/grad_norm_pre_clip': 5.888552188873291, 'timing/data': 0.0013761380687355995, 'timing/model': 1.1168093513697386, 'learning_rate/qwen_vl_interface': 9.861854169069708e-06, 'learning_rate/action_model': 9.861854169069707e-05, 'epoch': 0.24}) +[2026-06-17 17:33:10,172][starVLA.training.train_starvla][INFO] - Step 632, Loss: {'train/loss': 0.9652579426765442, 'train/grad_norm_pre_clip': 6.6134867668151855, 'timing/data': 0.0011029057204723358, 'timing/model': 1.4339843764901161, 'learning_rate/qwen_vl_interface': 9.861335904702058e-06, 'learning_rate/action_model': 9.861335904702058e-05, 'epoch': 0.24}) +[2026-06-17 17:33:14,439][starVLA.training.train_starvla][INFO] - Step 633, Loss: {'train/loss': 0.769440770149231, 'train/grad_norm_pre_clip': 7.302474021911621, 'timing/data': 0.0011590803042054176, 'timing/model': 1.081214090809226, 'learning_rate/qwen_vl_interface': 9.860816684399514e-06, 'learning_rate/action_model': 9.860816684399514e-05, 'epoch': 0.24}) +[2026-06-17 17:33:19,063][starVLA.training.train_starvla][INFO] - Step 634, Loss: {'train/loss': 0.7506968975067139, 'train/grad_norm_pre_clip': 7.212962627410889, 'timing/data': 0.0011202562600374222, 'timing/model': 1.2882759822532535, 'learning_rate/qwen_vl_interface': 9.86029650826971e-06, 'learning_rate/action_model': 9.86029650826971e-05, 'epoch': 0.24}) +[2026-06-17 17:33:23,897][starVLA.training.train_starvla][INFO] - Step 635, Loss: {'train/loss': 0.5067384839057922, 'train/grad_norm_pre_clip': 4.323378086090088, 'timing/data': 0.001388981007039547, 'timing/model': 1.0676682749763131, 'learning_rate/qwen_vl_interface': 9.859775376420478e-06, 'learning_rate/action_model': 9.859775376420478e-05, 'epoch': 0.24}) +[2026-06-17 17:33:28,906][starVLA.training.train_starvla][INFO] - Step 636, Loss: {'train/loss': 0.9149178266525269, 'train/grad_norm_pre_clip': 8.831673622131348, 'timing/data': 0.0009306231513619423, 'timing/model': 1.2024534726515412, 'learning_rate/qwen_vl_interface': 9.85925328895985e-06, 'learning_rate/action_model': 9.85925328895985e-05, 'epoch': 0.24}) +[2026-06-17 17:33:33,845][starVLA.training.train_starvla][INFO] - Step 637, Loss: {'train/loss': 1.1679232120513916, 'train/grad_norm_pre_clip': 7.782026767730713, 'timing/data': 0.0015922943130135536, 'timing/model': 1.5458941897377372, 'learning_rate/qwen_vl_interface': 9.858730245996057e-06, 'learning_rate/action_model': 9.858730245996057e-05, 'epoch': 0.25}) +[2026-06-17 17:33:42,005][starVLA.training.train_starvla][INFO] - Step 638, Loss: {'train/loss': 0.62140291929245, 'train/grad_norm_pre_clip': 4.126392841339111, 'timing/data': 0.0008007977157831192, 'timing/model': 1.4362712446600199, 'learning_rate/qwen_vl_interface': 9.858206247637522e-06, 'learning_rate/action_model': 9.858206247637523e-05, 'epoch': 0.25}) +[2026-06-17 17:33:47,144][starVLA.training.train_starvla][INFO] - Step 639, Loss: {'train/loss': 0.6633888483047485, 'train/grad_norm_pre_clip': 4.910614967346191, 'timing/data': 0.0012047616764903069, 'timing/model': 1.2611726252362132, 'learning_rate/qwen_vl_interface': 9.857681293992876e-06, 'learning_rate/action_model': 9.857681293992877e-05, 'epoch': 0.25}) +[2026-06-17 17:33:52,346][starVLA.training.train_starvla][INFO] - Step 640, Loss: {'train/loss': 0.7525733113288879, 'train/grad_norm_pre_clip': 6.825132846832275, 'timing/data': 0.0010733185335993767, 'timing/model': 1.484218880534172, 'learning_rate/qwen_vl_interface': 9.85715538517094e-06, 'learning_rate/action_model': 9.857155385170938e-05, 'epoch': 0.25}) +[2026-06-17 17:34:00,147][starVLA.training.train_starvla][INFO] - Step 641, Loss: {'train/loss': 0.6841551661491394, 'train/grad_norm_pre_clip': 4.645129680633545, 'timing/data': 0.001221473328769207, 'timing/model': 4.2998853493481874, 'learning_rate/qwen_vl_interface': 9.856628521280733e-06, 'learning_rate/action_model': 9.856628521280733e-05, 'epoch': 0.25}) +[2026-06-17 17:34:08,184][starVLA.training.train_starvla][INFO] - Step 642, Loss: {'train/loss': 0.700485348701477, 'train/grad_norm_pre_clip': 8.281442642211914, 'timing/data': 0.0011617299169301987, 'timing/model': 4.6016553565859795, 'learning_rate/qwen_vl_interface': 9.856100702431477e-06, 'learning_rate/action_model': 9.856100702431477e-05, 'epoch': 0.25}) +[2026-06-17 17:34:16,919][starVLA.training.train_starvla][INFO] - Step 643, Loss: {'train/loss': 0.7554746270179749, 'train/grad_norm_pre_clip': 4.940661907196045, 'timing/data': 0.0012925611808896065, 'timing/model': 1.3492332436144352, 'learning_rate/qwen_vl_interface': 9.85557192873259e-06, 'learning_rate/action_model': 9.85557192873259e-05, 'epoch': 0.25}) +[2026-06-17 17:34:23,867][starVLA.training.train_starvla][INFO] - Step 644, Loss: {'train/loss': 0.8089832663536072, 'train/grad_norm_pre_clip': 5.728267669677734, 'timing/data': 0.0011403635144233704, 'timing/model': 1.2665665345266461, 'learning_rate/qwen_vl_interface': 9.855042200293684e-06, 'learning_rate/action_model': 9.855042200293685e-05, 'epoch': 0.25}) +[2026-06-17 17:34:39,721][starVLA.training.train_starvla][INFO] - Step 645, Loss: {'train/loss': 0.6600490212440491, 'train/grad_norm_pre_clip': 5.203063011169434, 'timing/data': 0.001710117794573307, 'timing/model': 2.1499988567084074, 'learning_rate/qwen_vl_interface': 9.854511517224578e-06, 'learning_rate/action_model': 9.854511517224578e-05, 'epoch': 0.25}) +[2026-06-17 17:34:46,238][starVLA.training.train_starvla][INFO] - Step 646, Loss: {'train/loss': 0.523219108581543, 'train/grad_norm_pre_clip': 6.603754997253418, 'timing/data': 0.0009120162576436996, 'timing/model': 2.870638867840171, 'learning_rate/qwen_vl_interface': 9.853979879635278e-06, 'learning_rate/action_model': 9.853979879635278e-05, 'epoch': 0.25}) +[2026-06-17 17:34:51,048][starVLA.training.train_starvla][INFO] - Step 647, Loss: {'train/loss': 0.669628381729126, 'train/grad_norm_pre_clip': 8.173088073730469, 'timing/data': 0.001200036145746708, 'timing/model': 1.4409065004438162, 'learning_rate/qwen_vl_interface': 9.853447287635995e-06, 'learning_rate/action_model': 9.853447287635995e-05, 'epoch': 0.25}) +[2026-06-17 17:34:57,880][starVLA.training.train_starvla][INFO] - Step 648, Loss: {'train/loss': 0.7742818593978882, 'train/grad_norm_pre_clip': 6.5262956619262695, 'timing/data': 0.0010948795825242996, 'timing/model': 1.0564196649938822, 'learning_rate/qwen_vl_interface': 9.852913741337136e-06, 'learning_rate/action_model': 9.852913741337136e-05, 'epoch': 0.25}) +[2026-06-17 17:35:02,397][starVLA.training.train_starvla][INFO] - Step 649, Loss: {'train/loss': 0.5144185423851013, 'train/grad_norm_pre_clip': 6.102718830108643, 'timing/data': 0.0009145596995949745, 'timing/model': 1.061848970130086, 'learning_rate/qwen_vl_interface': 9.852379240849306e-06, 'learning_rate/action_model': 9.852379240849305e-05, 'epoch': 0.25}) +[2026-06-17 17:35:07,017][starVLA.training.train_starvla][INFO] - Step 650, Loss: {'train/loss': 0.5004538297653198, 'train/grad_norm_pre_clip': 14.452281951904297, 'timing/data': 0.0008832253515720367, 'timing/model': 1.0634913025423884, 'learning_rate/qwen_vl_interface': 9.851843786283306e-06, 'learning_rate/action_model': 9.851843786283306e-05, 'epoch': 0.25}) +[2026-06-17 17:35:11,637][starVLA.training.train_starvla][INFO] - Step 651, Loss: {'train/loss': 0.8386367559432983, 'train/grad_norm_pre_clip': 7.533234596252441, 'timing/data': 0.0031702760607004166, 'timing/model': 1.0859485305845737, 'learning_rate/qwen_vl_interface': 9.851307377750137e-06, 'learning_rate/action_model': 9.851307377750137e-05, 'epoch': 0.25}) +[2026-06-17 17:35:16,296][starVLA.training.train_starvla][INFO] - Step 652, Loss: {'train/loss': 0.7660812139511108, 'train/grad_norm_pre_clip': 5.607500076293945, 'timing/data': 0.0012248409911990166, 'timing/model': 1.4383504735305905, 'learning_rate/qwen_vl_interface': 9.850770015360999e-06, 'learning_rate/action_model': 9.850770015360999e-05, 'epoch': 0.25}) +[2026-06-17 17:35:20,613][starVLA.training.train_starvla][INFO] - Step 653, Loss: {'train/loss': 0.580539882183075, 'train/grad_norm_pre_clip': 5.992076873779297, 'timing/data': 0.0015610307455062866, 'timing/model': 1.1467032181099057, 'learning_rate/qwen_vl_interface': 9.850231699227284e-06, 'learning_rate/action_model': 9.850231699227284e-05, 'epoch': 0.25}) +[2026-06-17 17:35:25,594][starVLA.training.train_starvla][INFO] - Step 654, Loss: {'train/loss': 0.8264142870903015, 'train/grad_norm_pre_clip': 10.487569808959961, 'timing/data': 0.001444488763809204, 'timing/model': 1.1784828510135412, 'learning_rate/qwen_vl_interface': 9.84969242946059e-06, 'learning_rate/action_model': 9.84969242946059e-05, 'epoch': 0.25}) +[2026-06-17 17:35:30,190][starVLA.training.train_starvla][INFO] - Step 655, Loss: {'train/loss': 0.7271349430084229, 'train/grad_norm_pre_clip': 7.333932876586914, 'timing/data': 0.0011136680841445923, 'timing/model': 1.0671237921342254, 'learning_rate/qwen_vl_interface': 9.849152206172704e-06, 'learning_rate/action_model': 9.849152206172704e-05, 'epoch': 0.25}) +[2026-06-17 17:35:34,790][starVLA.training.train_starvla][INFO] - Step 656, Loss: {'train/loss': 0.9639070630073547, 'train/grad_norm_pre_clip': 11.562673568725586, 'timing/data': 0.0007715141400694847, 'timing/model': 1.0796166388317943, 'learning_rate/qwen_vl_interface': 9.84861102947562e-06, 'learning_rate/action_model': 9.848611029475619e-05, 'epoch': 0.25}) +[2026-06-17 17:35:39,336][starVLA.training.train_starvla][INFO] - Step 657, Loss: {'train/loss': 0.8343703746795654, 'train/grad_norm_pre_clip': 10.603217124938965, 'timing/data': 0.0013301931321620941, 'timing/model': 1.4371622456237674, 'learning_rate/qwen_vl_interface': 9.848068899481516e-06, 'learning_rate/action_model': 9.848068899481516e-05, 'epoch': 0.25}) +[2026-06-17 17:35:43,574][starVLA.training.train_starvla][INFO] - Step 658, Loss: {'train/loss': 0.3487716615200043, 'train/grad_norm_pre_clip': 5.104213237762451, 'timing/data': 0.0013501467183232307, 'timing/model': 1.07708705868572, 'learning_rate/qwen_vl_interface': 9.847525816302784e-06, 'learning_rate/action_model': 9.847525816302784e-05, 'epoch': 0.25}) +[2026-06-17 17:35:48,835][starVLA.training.train_starvla][INFO] - Step 659, Loss: {'train/loss': 0.6735888123512268, 'train/grad_norm_pre_clip': 6.326278209686279, 'timing/data': 0.0013303933665156364, 'timing/model': 1.1730071492493153, 'learning_rate/qwen_vl_interface': 9.846981780052005e-06, 'learning_rate/action_model': 9.846981780052003e-05, 'epoch': 0.25}) +[2026-06-17 17:35:53,451][starVLA.training.train_starvla][INFO] - Step 660, Loss: {'train/loss': 0.7403250336647034, 'train/grad_norm_pre_clip': 4.685258865356445, 'timing/data': 0.0011600982397794724, 'timing/model': 1.0776891866698861, 'learning_rate/qwen_vl_interface': 9.846436790841955e-06, 'learning_rate/action_model': 9.846436790841955e-05, 'epoch': 0.25}) +[2026-06-17 17:35:57,908][starVLA.training.train_starvla][INFO] - Step 661, Loss: {'train/loss': 0.5210055708885193, 'train/grad_norm_pre_clip': 7.347040176391602, 'timing/data': 0.0013635093346238136, 'timing/model': 1.0842938721179962, 'learning_rate/qwen_vl_interface': 9.84589084878561e-06, 'learning_rate/action_model': 9.845890848785611e-05, 'epoch': 0.25}) +[2026-06-17 17:36:02,385][starVLA.training.train_starvla][INFO] - Step 662, Loss: {'train/loss': 0.8347724676132202, 'train/grad_norm_pre_clip': 6.333577632904053, 'timing/data': 0.001209976151585579, 'timing/model': 1.4095097472891212, 'learning_rate/qwen_vl_interface': 9.84534395399615e-06, 'learning_rate/action_model': 9.84534395399615e-05, 'epoch': 0.25}) +[2026-06-17 17:36:09,980][starVLA.training.train_starvla][INFO] - Step 663, Loss: {'train/loss': 0.3864610493183136, 'train/grad_norm_pre_clip': 3.538093328475952, 'timing/data': 0.001104511320590973, 'timing/model': 4.324551725760102, 'learning_rate/qwen_vl_interface': 9.844796106586942e-06, 'learning_rate/action_model': 9.844796106586941e-05, 'epoch': 0.26}) +[2026-06-17 17:36:20,881][starVLA.training.train_starvla][INFO] - Step 664, Loss: {'train/loss': 0.6847114562988281, 'train/grad_norm_pre_clip': 6.708979606628418, 'timing/data': 0.0013999994844198227, 'timing/model': 1.3619617279618979, 'learning_rate/qwen_vl_interface': 9.844247306671557e-06, 'learning_rate/action_model': 9.844247306671555e-05, 'epoch': 0.26}) +[2026-06-17 17:36:29,609][starVLA.training.train_starvla][INFO] - Step 665, Loss: {'train/loss': 0.7466737031936646, 'train/grad_norm_pre_clip': 6.302065372467041, 'timing/data': 0.0042486125603318214, 'timing/model': 1.253916242159903, 'learning_rate/qwen_vl_interface': 9.843697554363762e-06, 'learning_rate/action_model': 9.843697554363762e-05, 'epoch': 0.26}) +[2026-06-17 17:36:34,692][starVLA.training.train_starvla][INFO] - Step 666, Loss: {'train/loss': 0.7180001735687256, 'train/grad_norm_pre_clip': 4.351414680480957, 'timing/data': 0.0012171696871519089, 'timing/model': 1.1530644074082375, 'learning_rate/qwen_vl_interface': 9.84314684977752e-06, 'learning_rate/action_model': 9.84314684977752e-05, 'epoch': 0.26}) +[2026-06-17 17:36:40,235][starVLA.training.train_starvla][INFO] - Step 667, Loss: {'train/loss': 0.9006209969520569, 'train/grad_norm_pre_clip': 6.860239028930664, 'timing/data': 0.0010909251868724823, 'timing/model': 1.5916434526443481, 'learning_rate/qwen_vl_interface': 9.842595193026993e-06, 'learning_rate/action_model': 9.842595193026993e-05, 'epoch': 0.26}) +[2026-06-17 17:36:48,170][starVLA.training.train_starvla][INFO] - Step 668, Loss: {'train/loss': 0.5997358560562134, 'train/grad_norm_pre_clip': 5.136195659637451, 'timing/data': 0.0013255076482892036, 'timing/model': 4.315556629560888, 'learning_rate/qwen_vl_interface': 9.842042584226545e-06, 'learning_rate/action_model': 9.842042584226544e-05, 'epoch': 0.26}) +[2026-06-17 17:36:52,772][starVLA.training.train_starvla][INFO] - Step 669, Loss: {'train/loss': 0.38856977224349976, 'train/grad_norm_pre_clip': 3.963616132736206, 'timing/data': 0.0013239756226539612, 'timing/model': 1.1115090884268284, 'learning_rate/qwen_vl_interface': 9.841489023490723e-06, 'learning_rate/action_model': 9.841489023490723e-05, 'epoch': 0.26}) +[2026-06-17 17:36:58,462][starVLA.training.train_starvla][INFO] - Step 670, Loss: {'train/loss': 0.451700359582901, 'train/grad_norm_pre_clip': 6.770754337310791, 'timing/data': 0.0012548388913273811, 'timing/model': 1.1014626771211624, 'learning_rate/qwen_vl_interface': 9.840934510934288e-06, 'learning_rate/action_model': 9.840934510934288e-05, 'epoch': 0.26}) +[2026-06-17 17:37:06,582][starVLA.training.train_starvla][INFO] - Step 671, Loss: {'train/loss': 0.7215419411659241, 'train/grad_norm_pre_clip': 9.141334533691406, 'timing/data': 0.0012591751292347908, 'timing/model': 4.637393091805279, 'learning_rate/qwen_vl_interface': 9.840379046672189e-06, 'learning_rate/action_model': 9.840379046672189e-05, 'epoch': 0.26}) +[2026-06-17 17:37:11,480][starVLA.training.train_starvla][INFO] - Step 672, Loss: {'train/loss': 1.1743866205215454, 'train/grad_norm_pre_clip': 16.991939544677734, 'timing/data': 0.001081574708223343, 'timing/model': 1.3675525952130556, 'learning_rate/qwen_vl_interface': 9.839822630819573e-06, 'learning_rate/action_model': 9.839822630819572e-05, 'epoch': 0.26}) +[2026-06-17 17:37:15,829][starVLA.training.train_starvla][INFO] - Step 673, Loss: {'train/loss': 0.6467791795730591, 'train/grad_norm_pre_clip': 13.199092864990234, 'timing/data': 0.0012685898691415787, 'timing/model': 1.0879025552421808, 'learning_rate/qwen_vl_interface': 9.839265263491789e-06, 'learning_rate/action_model': 9.839265263491789e-05, 'epoch': 0.26}) +[2026-06-17 17:37:20,899][starVLA.training.train_starvla][INFO] - Step 674, Loss: {'train/loss': 0.45127367973327637, 'train/grad_norm_pre_clip': 7.965175628662109, 'timing/data': 0.0013097422197461128, 'timing/model': 1.0951456893235445, 'learning_rate/qwen_vl_interface': 9.838706944804376e-06, 'learning_rate/action_model': 9.838706944804374e-05, 'epoch': 0.26}) +[2026-06-17 17:37:26,043][starVLA.training.train_starvla][INFO] - Step 675, Loss: {'train/loss': 0.623557984828949, 'train/grad_norm_pre_clip': 10.020401954650879, 'timing/data': 0.0014779437333345413, 'timing/model': 1.2474242635071278, 'learning_rate/qwen_vl_interface': 9.838147674873076e-06, 'learning_rate/action_model': 9.838147674873075e-05, 'epoch': 0.26}) +[2026-06-17 17:37:30,880][starVLA.training.train_starvla][INFO] - Step 676, Loss: {'train/loss': 0.28435593843460083, 'train/grad_norm_pre_clip': 5.6837992668151855, 'timing/data': 0.0011457791551947594, 'timing/model': 1.1020603515207767, 'learning_rate/qwen_vl_interface': 9.837587453813826e-06, 'learning_rate/action_model': 9.837587453813824e-05, 'epoch': 0.26}) +[2026-06-17 17:37:35,427][starVLA.training.train_starvla][INFO] - Step 677, Loss: {'train/loss': 0.4160076081752777, 'train/grad_norm_pre_clip': 12.914689064025879, 'timing/data': 0.001200217753648758, 'timing/model': 1.0750975236296654, 'learning_rate/qwen_vl_interface': 9.83702628174276e-06, 'learning_rate/action_model': 9.837026281742759e-05, 'epoch': 0.26}) +[2026-06-17 17:37:40,562][starVLA.training.train_starvla][INFO] - Step 678, Loss: {'train/loss': 0.6779778003692627, 'train/grad_norm_pre_clip': 9.966252326965332, 'timing/data': 0.001401502639055252, 'timing/model': 1.3722984679043293, 'learning_rate/qwen_vl_interface': 9.83646415877621e-06, 'learning_rate/action_model': 9.83646415877621e-05, 'epoch': 0.26}) +[2026-06-17 17:37:45,511][starVLA.training.train_starvla][INFO] - Step 679, Loss: {'train/loss': 0.7053162455558777, 'train/grad_norm_pre_clip': 12.112639427185059, 'timing/data': 0.0011851787567138672, 'timing/model': 1.0809008376672864, 'learning_rate/qwen_vl_interface': 9.835901085030703e-06, 'learning_rate/action_model': 9.835901085030703e-05, 'epoch': 0.26}) +[2026-06-17 17:37:53,615][starVLA.training.train_starvla][INFO] - Step 680, Loss: {'train/loss': 0.6816073656082153, 'train/grad_norm_pre_clip': 9.3572998046875, 'timing/data': 0.0014628367498517036, 'timing/model': 1.0858871107921004, 'learning_rate/qwen_vl_interface': 9.83533706062297e-06, 'learning_rate/action_model': 9.835337060622969e-05, 'epoch': 0.26}) +[2026-06-17 17:38:01,723][starVLA.training.train_starvla][INFO] - Step 681, Loss: {'train/loss': 1.1661180257797241, 'train/grad_norm_pre_clip': 10.828275680541992, 'timing/data': 0.0016236798837780952, 'timing/model': 1.7561327703297138, 'learning_rate/qwen_vl_interface': 9.834772085669927e-06, 'learning_rate/action_model': 9.834772085669927e-05, 'epoch': 0.26}) +[2026-06-17 17:38:06,990][starVLA.training.train_starvla][INFO] - Step 682, Loss: {'train/loss': 0.7256942987442017, 'train/grad_norm_pre_clip': 10.54991340637207, 'timing/data': 0.0010867910459637642, 'timing/model': 1.7221720069646835, 'learning_rate/qwen_vl_interface': 9.834206160288698e-06, 'learning_rate/action_model': 9.834206160288698e-05, 'epoch': 0.26}) +[2026-06-17 17:38:19,954][starVLA.training.train_starvla][INFO] - Step 683, Loss: {'train/loss': 1.0778768062591553, 'train/grad_norm_pre_clip': 8.781086921691895, 'timing/data': 0.0010475069284439087, 'timing/model': 1.25947190169245, 'learning_rate/qwen_vl_interface': 9.8336392845966e-06, 'learning_rate/action_model': 9.833639284596599e-05, 'epoch': 0.26}) +[2026-06-17 17:38:24,954][starVLA.training.train_starvla][INFO] - Step 684, Loss: {'train/loss': 0.6125866174697876, 'train/grad_norm_pre_clip': 7.209470748901367, 'timing/data': 0.002016894519329071, 'timing/model': 1.1487935464829206, 'learning_rate/qwen_vl_interface': 9.833071458711146e-06, 'learning_rate/action_model': 9.833071458711146e-05, 'epoch': 0.26}) +[2026-06-17 17:38:29,960][starVLA.training.train_starvla][INFO] - Step 685, Loss: {'train/loss': 0.9900728464126587, 'train/grad_norm_pre_clip': 11.398698806762695, 'timing/data': 0.0011619823053479195, 'timing/model': 1.1792009864002466, 'learning_rate/qwen_vl_interface': 9.832502682750048e-06, 'learning_rate/action_model': 9.832502682750047e-05, 'epoch': 0.26}) +[2026-06-17 17:38:35,013][starVLA.training.train_starvla][INFO] - Step 686, Loss: {'train/loss': 0.9295437335968018, 'train/grad_norm_pre_clip': 8.404393196105957, 'timing/data': 0.001164003275334835, 'timing/model': 1.611626978032291, 'learning_rate/qwen_vl_interface': 9.831932956831212e-06, 'learning_rate/action_model': 9.831932956831211e-05, 'epoch': 0.26}) +[2026-06-17 17:38:44,391][starVLA.training.train_starvla][INFO] - Step 687, Loss: {'train/loss': 0.8654699325561523, 'train/grad_norm_pre_clip': 8.470902442932129, 'timing/data': 0.0015233317390084267, 'timing/model': 1.5860610827803612, 'learning_rate/qwen_vl_interface': 9.831362281072743e-06, 'learning_rate/action_model': 9.831362281072742e-05, 'epoch': 0.26}) +[2026-06-17 17:38:49,209][starVLA.training.train_starvla][INFO] - Step 688, Loss: {'train/loss': 0.7874296307563782, 'train/grad_norm_pre_clip': 7.1800150871276855, 'timing/data': 0.001288251020014286, 'timing/model': 1.317680923268199, 'learning_rate/qwen_vl_interface': 9.830790655592944e-06, 'learning_rate/action_model': 9.830790655592943e-05, 'epoch': 0.26}) +[2026-06-17 17:38:54,048][starVLA.training.train_starvla][INFO] - Step 689, Loss: {'train/loss': 0.8005664944648743, 'train/grad_norm_pre_clip': 10.33035945892334, 'timing/data': 0.0013474524021148682, 'timing/model': 1.1751659261062741, 'learning_rate/qwen_vl_interface': 9.830218080510313e-06, 'learning_rate/action_model': 9.830218080510312e-05, 'epoch': 0.27}) +[2026-06-17 17:39:02,800][starVLA.training.train_starvla][INFO] - Step 690, Loss: {'train/loss': 0.785358726978302, 'train/grad_norm_pre_clip': 13.894824028015137, 'timing/data': 0.0012906147167086601, 'timing/model': 1.0770380171015859, 'learning_rate/qwen_vl_interface': 9.829644555943545e-06, 'learning_rate/action_model': 9.829644555943545e-05, 'epoch': 0.27}) +[2026-06-17 17:39:08,021][starVLA.training.train_starvla][INFO] - Step 691, Loss: {'train/loss': 0.9066546559333801, 'train/grad_norm_pre_clip': 9.167203903198242, 'timing/data': 0.0008160918951034546, 'timing/model': 1.4502123575657606, 'learning_rate/qwen_vl_interface': 9.82907008201153e-06, 'learning_rate/action_model': 9.829070082011531e-05, 'epoch': 0.27}) +[2026-06-17 17:39:13,433][starVLA.training.train_starvla][INFO] - Step 692, Loss: {'train/loss': 0.6929513812065125, 'train/grad_norm_pre_clip': 8.43895149230957, 'timing/data': 0.0009292140603065491, 'timing/model': 1.1686397520825267, 'learning_rate/qwen_vl_interface': 9.828494658833366e-06, 'learning_rate/action_model': 9.828494658833365e-05, 'epoch': 0.27}) +[2026-06-17 17:39:18,224][starVLA.training.train_starvla][INFO] - Step 693, Loss: {'train/loss': 0.6566983461380005, 'train/grad_norm_pre_clip': 6.418970108032227, 'timing/data': 0.0013651782646775246, 'timing/model': 1.231138507835567, 'learning_rate/qwen_vl_interface': 9.82791828652833e-06, 'learning_rate/action_model': 9.827918286528329e-05, 'epoch': 0.27}) +[2026-06-17 17:39:22,785][starVLA.training.train_starvla][INFO] - Step 694, Loss: {'train/loss': 0.6904628276824951, 'train/grad_norm_pre_clip': 7.831206798553467, 'timing/data': 0.001316106878221035, 'timing/model': 1.138927474617958, 'learning_rate/qwen_vl_interface': 9.827340965215903e-06, 'learning_rate/action_model': 9.827340965215904e-05, 'epoch': 0.27}) +[2026-06-17 17:39:30,940][starVLA.training.train_starvla][INFO] - Step 695, Loss: {'train/loss': 0.824712336063385, 'train/grad_norm_pre_clip': 4.568272113800049, 'timing/data': 0.0016656555235385895, 'timing/model': 1.1396152442321181, 'learning_rate/qwen_vl_interface': 9.826762695015773e-06, 'learning_rate/action_model': 9.826762695015772e-05, 'epoch': 0.27}) +[2026-06-17 17:39:38,732][starVLA.training.train_starvla][INFO] - Step 696, Loss: {'train/loss': 0.48691847920417786, 'train/grad_norm_pre_clip': 4.29478645324707, 'timing/data': 0.0016180230304598808, 'timing/model': 1.60310984775424, 'learning_rate/qwen_vl_interface': 9.82618347604781e-06, 'learning_rate/action_model': 9.82618347604781e-05, 'epoch': 0.27}) +[2026-06-17 17:39:43,288][starVLA.training.train_starvla][INFO] - Step 697, Loss: {'train/loss': 0.5654178857803345, 'train/grad_norm_pre_clip': 3.928288459777832, 'timing/data': 0.0011177053675055504, 'timing/model': 1.4105857824906707, 'learning_rate/qwen_vl_interface': 9.825603308432089e-06, 'learning_rate/action_model': 9.825603308432088e-05, 'epoch': 0.27}) +[2026-06-17 17:39:47,519][starVLA.training.train_starvla][INFO] - Step 698, Loss: {'train/loss': 0.9952552914619446, 'train/grad_norm_pre_clip': 8.867603302001953, 'timing/data': 0.0014425721019506454, 'timing/model': 1.0507669020444155, 'learning_rate/qwen_vl_interface': 9.825022192288878e-06, 'learning_rate/action_model': 9.825022192288877e-05, 'epoch': 0.27}) +[2026-06-17 17:39:52,134][starVLA.training.train_starvla][INFO] - Step 699, Loss: {'train/loss': 0.7260399460792542, 'train/grad_norm_pre_clip': 7.57065486907959, 'timing/data': 0.0008798465132713318, 'timing/model': 1.0797769306227565, 'learning_rate/qwen_vl_interface': 9.824440127738643e-06, 'learning_rate/action_model': 9.824440127738642e-05, 'epoch': 0.27}) +[2026-06-17 17:43:06,316][starVLA.training.train_starvla][INFO] - Step 700, Loss: {'train/loss': 1.013500452041626, 'train/grad_norm_pre_clip': 8.231782913208008, 'eval/loss': 0.7354665398597717, 'eval/latency_4/loss': 0.7354665398597717, 'eval/demon_fire_precision': 0.9316317228805834, 'eval/demon_fire_recall': 0.9957778499512829, 'eval/demon_fire_f1': 0.9626373626373627, 'eval/demon_fire_rate_model': 0.9665198237885463, 'eval/demon_fire_rate_teacher': 0.9042584434654919, 'eval/demon_left_precision': 0.8372093023255814, 'eval/demon_left_recall': 0.1643835616438356, 'eval/demon_left_f1': 0.2748091603053435, 'eval/demon_left_rate_model': 0.02525697503671072, 'eval/demon_left_rate_teacher': 0.12863436123348018, 'eval/demon_right_precision': 0.7122302158273381, 'eval/demon_right_recall': 0.5361010830324909, 'eval/demon_right_f1': 0.6117404737384139, 'eval/demon_right_rate_model': 0.12246696035242291, 'eval/demon_right_rate_teacher': 0.16270190895741557, 'eval/demon_move_f1': 0.4432748170218787, 'eval/demon_cc_f1': 0.7122760738631173, 'eval/latency_4/demon_fire_precision': 0.9316317228805834, 'eval/latency_4/demon_fire_recall': 0.9957778499512829, 'eval/latency_4/demon_fire_f1': 0.9626373626373627, 'eval/latency_4/demon_fire_rate_model': 0.9665198237885463, 'eval/latency_4/demon_fire_rate_teacher': 0.9042584434654919, 'eval/latency_4/demon_left_precision': 0.8372093023255814, 'eval/latency_4/demon_left_recall': 0.1643835616438356, 'eval/latency_4/demon_left_f1': 0.2748091603053435, 'eval/latency_4/demon_left_rate_model': 0.02525697503671072, 'eval/latency_4/demon_left_rate_teacher': 0.12863436123348018, 'eval/latency_4/demon_right_precision': 0.7122302158273381, 'eval/latency_4/demon_right_recall': 0.5361010830324909, 'eval/latency_4/demon_right_f1': 0.6117404737384139, 'eval/latency_4/demon_right_rate_model': 0.12246696035242291, 'eval/latency_4/demon_right_rate_teacher': 0.16270190895741557, 'eval/latency_4/demon_move_f1': 0.4432748170218787, 'eval/latency_4/demon_cc_f1': 0.7122760738631173, 'timing/data': 0.0007948931306600571, 'timing/model': 1.1143201077356935, 'learning_rate/qwen_vl_interface': 9.823857114902047e-06, 'learning_rate/action_model': 9.823857114902047e-05, 'epoch': 0.27}) +[2026-06-17 17:43:06,318][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_700_state +[2026-06-17 17:43:06,325][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 17:43:15,811][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_700_state/model.safetensors +[2026-06-17 17:43:38,262][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_700_state/optimizer.bin +[2026-06-17 17:43:38,264][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_700_state/sampler.bin +[2026-06-17 17:43:38,265][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_700_state/sampler_1.bin +[2026-06-17 17:43:38,268][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_700_state/random_states_0.pkl +[2026-06-17 17:43:38,274][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 700 +[2026-06-17 17:43:38,275][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 17:43:38,292][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 17:43:48,640][starVLA.training.train_starvla][INFO] - Step 701, Loss: {'train/loss': 1.0907747745513916, 'train/grad_norm_pre_clip': 13.097407341003418, 'timing/data': 0.0014427928254008293, 'timing/model': 1.3750182138755918, 'learning_rate/qwen_vl_interface': 9.823273153899951e-06, 'learning_rate/action_model': 9.82327315389995e-05, 'epoch': 0.27}) +[2026-06-17 17:43:53,419][starVLA.training.train_starvla][INFO] - Step 702, Loss: {'train/loss': 0.8017857670783997, 'train/grad_norm_pre_clip': 6.382943153381348, 'timing/data': 0.0017386442050337791, 'timing/model': 1.4976637484505773, 'learning_rate/qwen_vl_interface': 9.822688244853407e-06, 'learning_rate/action_model': 9.822688244853408e-05, 'epoch': 0.27}) +[2026-06-17 17:44:05,147][starVLA.training.train_starvla][INFO] - Step 703, Loss: {'train/loss': 0.6606091260910034, 'train/grad_norm_pre_clip': 6.002195358276367, 'timing/data': 0.0013035750016570091, 'timing/model': 1.236366305500269, 'learning_rate/qwen_vl_interface': 9.822102387883672e-06, 'learning_rate/action_model': 9.822102387883671e-05, 'epoch': 0.27}) +[2026-06-17 17:44:10,652][starVLA.training.train_starvla][INFO] - Step 704, Loss: {'train/loss': 0.7072222232818604, 'train/grad_norm_pre_clip': 4.831257343292236, 'timing/data': 0.0010186918079853058, 'timing/model': 1.324375320225954, 'learning_rate/qwen_vl_interface': 9.82151558311219e-06, 'learning_rate/action_model': 9.821515583112191e-05, 'epoch': 0.27}) +[2026-06-17 17:44:15,595][starVLA.training.train_starvla][INFO] - Step 705, Loss: {'train/loss': 0.738386869430542, 'train/grad_norm_pre_clip': 7.554230690002441, 'timing/data': 0.0012568915262818336, 'timing/model': 1.1866264212876558, 'learning_rate/qwen_vl_interface': 9.82092783066061e-06, 'learning_rate/action_model': 9.82092783066061e-05, 'epoch': 0.27}) +[2026-06-17 17:44:20,569][starVLA.training.train_starvla][INFO] - Step 706, Loss: {'train/loss': 0.38986748456954956, 'train/grad_norm_pre_clip': 5.0386247634887695, 'timing/data': 0.03300496097654104, 'timing/model': 1.5784491319209337, 'learning_rate/qwen_vl_interface': 9.820339130650771e-06, 'learning_rate/action_model': 9.820339130650771e-05, 'epoch': 0.27}) +[2026-06-17 17:44:25,855][starVLA.training.train_starvla][INFO] - Step 707, Loss: {'train/loss': 0.6231816411018372, 'train/grad_norm_pre_clip': 5.243691444396973, 'timing/data': 0.03472308628261089, 'timing/model': 1.5806945441290736, 'learning_rate/qwen_vl_interface': 9.819749483204715e-06, 'learning_rate/action_model': 9.819749483204715e-05, 'epoch': 0.27}) +[2026-06-17 17:44:30,520][starVLA.training.train_starvla][INFO] - Step 708, Loss: {'train/loss': 0.6657730340957642, 'train/grad_norm_pre_clip': 6.3697052001953125, 'timing/data': 0.03229838702827692, 'timing/model': 1.078431055881083, 'learning_rate/qwen_vl_interface': 9.81915888844467e-06, 'learning_rate/action_model': 9.81915888844467e-05, 'epoch': 0.27}) +[2026-06-17 17:44:35,678][starVLA.training.train_starvla][INFO] - Step 709, Loss: {'train/loss': 0.5763801336288452, 'train/grad_norm_pre_clip': 4.878749847412109, 'timing/data': 0.03419941570609808, 'timing/model': 1.39893988519907, 'learning_rate/qwen_vl_interface': 9.818567346493072e-06, 'learning_rate/action_model': 9.818567346493071e-05, 'epoch': 0.27}) +[2026-06-17 17:44:47,698][starVLA.training.train_starvla][INFO] - Step 710, Loss: {'train/loss': 0.5866876840591431, 'train/grad_norm_pre_clip': 5.883514881134033, 'timing/data': 0.031334465369582176, 'timing/model': 1.0714838486164808, 'learning_rate/qwen_vl_interface': 9.817974857472549e-06, 'learning_rate/action_model': 9.817974857472548e-05, 'epoch': 0.27}) +[2026-06-17 17:44:54,086][starVLA.training.train_starvla][INFO] - Step 711, Loss: {'train/loss': 0.5469833612442017, 'train/grad_norm_pre_clip': 7.504527568817139, 'timing/data': 0.03287270665168762, 'timing/model': 1.099839923903346, 'learning_rate/qwen_vl_interface': 9.817381421505922e-06, 'learning_rate/action_model': 9.817381421505921e-05, 'epoch': 0.27}) +[2026-06-17 17:44:58,813][starVLA.training.train_starvla][INFO] - Step 712, Loss: {'train/loss': 0.5477069020271301, 'train/grad_norm_pre_clip': 7.681894779205322, 'timing/data': 0.029918063431978226, 'timing/model': 1.4057914316654205, 'learning_rate/qwen_vl_interface': 9.81678703871621e-06, 'learning_rate/action_model': 9.81678703871621e-05, 'epoch': 0.27}) +[2026-06-17 17:45:04,024][starVLA.training.train_starvla][INFO] - Step 713, Loss: {'train/loss': 0.8508478999137878, 'train/grad_norm_pre_clip': 6.43865966796875, 'timing/data': 0.03609240986406803, 'timing/model': 1.233409340493381, 'learning_rate/qwen_vl_interface': 9.816191709226631e-06, 'learning_rate/action_model': 9.816191709226632e-05, 'epoch': 0.27}) +[2026-06-17 17:45:08,596][starVLA.training.train_starvla][INFO] - Step 714, Loss: {'train/loss': 0.9811233878135681, 'train/grad_norm_pre_clip': 7.843461513519287, 'timing/data': 0.04005460627377033, 'timing/model': 1.1815932365134358, 'learning_rate/qwen_vl_interface': 9.815595433160599e-06, 'learning_rate/action_model': 9.815595433160598e-05, 'epoch': 0.27}) +[2026-06-17 17:45:13,261][starVLA.training.train_starvla][INFO] - Step 715, Loss: {'train/loss': 0.9883396029472351, 'train/grad_norm_pre_clip': 10.642558097839355, 'timing/data': 0.03068523481488228, 'timing/model': 1.0755472583696246, 'learning_rate/qwen_vl_interface': 9.81499821064172e-06, 'learning_rate/action_model': 9.81499821064172e-05, 'epoch': 0.28}) +[2026-06-17 17:45:22,766][starVLA.training.train_starvla][INFO] - Step 716, Loss: {'train/loss': 0.6100305914878845, 'train/grad_norm_pre_clip': 6.838109016418457, 'timing/data': 0.03155565354973078, 'timing/model': 5.579768438823521, 'learning_rate/qwen_vl_interface': 9.814400041793799e-06, 'learning_rate/action_model': 9.814400041793798e-05, 'epoch': 0.28}) +[2026-06-17 17:45:31,503][starVLA.training.train_starvla][INFO] - Step 717, Loss: {'train/loss': 0.773939311504364, 'train/grad_norm_pre_clip': 7.397876262664795, 'timing/data': 0.04344727750867605, 'timing/model': 1.8541429853066802, 'learning_rate/qwen_vl_interface': 9.813800926740839e-06, 'learning_rate/action_model': 9.813800926740838e-05, 'epoch': 0.28}) +[2026-06-17 17:45:35,978][starVLA.training.train_starvla][INFO] - Step 718, Loss: {'train/loss': 0.7398108839988708, 'train/grad_norm_pre_clip': 8.236483573913574, 'timing/data': 0.034026104025542736, 'timing/model': 1.112653523683548, 'learning_rate/qwen_vl_interface': 9.813200865607034e-06, 'learning_rate/action_model': 9.813200865607034e-05, 'epoch': 0.28}) +[2026-06-17 17:45:43,385][starVLA.training.train_starvla][INFO] - Step 719, Loss: {'train/loss': 0.5740742087364197, 'train/grad_norm_pre_clip': 6.982445240020752, 'timing/data': 0.030118901282548904, 'timing/model': 1.4143792083486915, 'learning_rate/qwen_vl_interface': 9.812599858516782e-06, 'learning_rate/action_model': 9.812599858516782e-05, 'epoch': 0.28}) +[2026-06-17 17:45:48,252][starVLA.training.train_starvla][INFO] - Step 720, Loss: {'train/loss': 0.7831931710243225, 'train/grad_norm_pre_clip': 7.452160835266113, 'timing/data': 0.043706211261451244, 'timing/model': 1.163616606965661, 'learning_rate/qwen_vl_interface': 9.81199790559467e-06, 'learning_rate/action_model': 9.811997905594668e-05, 'epoch': 0.28}) +[2026-06-17 17:45:53,561][starVLA.training.train_starvla][INFO] - Step 721, Loss: {'train/loss': 0.6014871597290039, 'train/grad_norm_pre_clip': 5.737395286560059, 'timing/data': 0.03301449678838253, 'timing/model': 1.2565495856106281, 'learning_rate/qwen_vl_interface': 9.811395006965481e-06, 'learning_rate/action_model': 9.81139500696548e-05, 'epoch': 0.28}) +[2026-06-17 17:45:58,624][starVLA.training.train_starvla][INFO] - Step 722, Loss: {'train/loss': 0.6393216848373413, 'train/grad_norm_pre_clip': 5.956714153289795, 'timing/data': 0.0359635092318058, 'timing/model': 1.5447961883619428, 'learning_rate/qwen_vl_interface': 9.810791162754199e-06, 'learning_rate/action_model': 9.810791162754198e-05, 'epoch': 0.28}) +[2026-06-17 17:46:03,828][starVLA.training.train_starvla][INFO] - Step 723, Loss: {'train/loss': 0.7657453417778015, 'train/grad_norm_pre_clip': 6.105935096740723, 'timing/data': 0.035713931545615196, 'timing/model': 1.189969609491527, 'learning_rate/qwen_vl_interface': 9.810186373086002e-06, 'learning_rate/action_model': 9.810186373086002e-05, 'epoch': 0.28}) +[2026-06-17 17:46:11,400][starVLA.training.train_starvla][INFO] - Step 724, Loss: {'train/loss': 0.8859434127807617, 'train/grad_norm_pre_clip': 8.388453483581543, 'timing/data': 0.033549060113728046, 'timing/model': 1.0828396333381534, 'learning_rate/qwen_vl_interface': 9.809580638086264e-06, 'learning_rate/action_model': 9.809580638086263e-05, 'epoch': 0.28}) +[2026-06-17 17:46:16,220][starVLA.training.train_starvla][INFO] - Step 725, Loss: {'train/loss': 0.7005800604820251, 'train/grad_norm_pre_clip': 10.528158187866211, 'timing/data': 0.031073108315467834, 'timing/model': 1.0682567870244384, 'learning_rate/qwen_vl_interface': 9.80897395788055e-06, 'learning_rate/action_model': 9.808973957880551e-05, 'epoch': 0.28}) +[2026-06-17 17:46:20,909][starVLA.training.train_starvla][INFO] - Step 726, Loss: {'train/loss': 0.570511519908905, 'train/grad_norm_pre_clip': 6.785536289215088, 'timing/data': 0.03244565427303314, 'timing/model': 1.1276657087728381, 'learning_rate/qwen_vl_interface': 9.808366332594632e-06, 'learning_rate/action_model': 9.808366332594632e-05, 'epoch': 0.28}) +[2026-06-17 17:46:25,799][starVLA.training.train_starvla][INFO] - Step 727, Loss: {'train/loss': 0.6270915865898132, 'train/grad_norm_pre_clip': 5.805953502655029, 'timing/data': 0.07005682773888111, 'timing/model': 1.467630879022181, 'learning_rate/qwen_vl_interface': 9.807757762354467e-06, 'learning_rate/action_model': 9.807757762354467e-05, 'epoch': 0.28}) +[2026-06-17 17:46:30,325][starVLA.training.train_starvla][INFO] - Step 728, Loss: {'train/loss': 0.5721203684806824, 'train/grad_norm_pre_clip': 4.595840930938721, 'timing/data': 0.03739572875201702, 'timing/model': 1.0871068062260747, 'learning_rate/qwen_vl_interface': 9.807148247286215e-06, 'learning_rate/action_model': 9.807148247286215e-05, 'epoch': 0.28}) +[2026-06-17 17:46:35,147][starVLA.training.train_starvla][INFO] - Step 729, Loss: {'train/loss': 0.44123509526252747, 'train/grad_norm_pre_clip': 6.977632522583008, 'timing/data': 0.031041822396218777, 'timing/model': 1.0879713017493486, 'learning_rate/qwen_vl_interface': 9.806537787516226e-06, 'learning_rate/action_model': 9.806537787516226e-05, 'epoch': 0.28}) +[2026-06-17 17:46:40,013][starVLA.training.train_starvla][INFO] - Step 730, Loss: {'train/loss': 0.8639811873435974, 'train/grad_norm_pre_clip': 10.974409103393555, 'timing/data': 0.0338311493396759, 'timing/model': 1.0857644220814109, 'learning_rate/qwen_vl_interface': 9.80592638317105e-06, 'learning_rate/action_model': 9.80592638317105e-05, 'epoch': 0.28}) +[2026-06-17 17:46:45,056][starVLA.training.train_starvla][INFO] - Step 731, Loss: {'train/loss': 0.7112300395965576, 'train/grad_norm_pre_clip': 11.61965560913086, 'timing/data': 0.042400117963552475, 'timing/model': 1.121343377046287, 'learning_rate/qwen_vl_interface': 9.805314034377434e-06, 'learning_rate/action_model': 9.805314034377434e-05, 'epoch': 0.28}) +[2026-06-17 17:46:51,110][starVLA.training.train_starvla][INFO] - Step 732, Loss: {'train/loss': 0.7097612619400024, 'train/grad_norm_pre_clip': 8.21818733215332, 'timing/data': 0.03366871830075979, 'timing/model': 2.8018318340182304, 'learning_rate/qwen_vl_interface': 9.804700741262318e-06, 'learning_rate/action_model': 9.804700741262317e-05, 'epoch': 0.28}) +[2026-06-17 17:46:56,011][starVLA.training.train_starvla][INFO] - Step 733, Loss: {'train/loss': 0.4015919268131256, 'train/grad_norm_pre_clip': 4.434615612030029, 'timing/data': 0.05033020582050085, 'timing/model': 1.081847807392478, 'learning_rate/qwen_vl_interface': 9.804086503952835e-06, 'learning_rate/action_model': 9.804086503952835e-05, 'epoch': 0.28}) +[2026-06-17 17:47:03,267][starVLA.training.train_starvla][INFO] - Step 734, Loss: {'train/loss': 0.7449535131454468, 'train/grad_norm_pre_clip': 6.798038482666016, 'timing/data': 0.03370937891304493, 'timing/model': 1.3120072400197387, 'learning_rate/qwen_vl_interface': 9.803471322576319e-06, 'learning_rate/action_model': 9.803471322576319e-05, 'epoch': 0.28}) +[2026-06-17 17:47:12,518][starVLA.training.train_starvla][INFO] - Step 735, Loss: {'train/loss': 0.7291616797447205, 'train/grad_norm_pre_clip': 7.640810012817383, 'timing/data': 0.032396684400737286, 'timing/model': 2.586710867471993, 'learning_rate/qwen_vl_interface': 9.8028551972603e-06, 'learning_rate/action_model': 9.802855197260299e-05, 'epoch': 0.28}) +[2026-06-17 17:47:19,075][starVLA.training.train_starvla][INFO] - Step 736, Loss: {'train/loss': 0.696200430393219, 'train/grad_norm_pre_clip': 8.3728666305542, 'timing/data': 0.04149444308131933, 'timing/model': 1.5839196182787418, 'learning_rate/qwen_vl_interface': 9.802238128132498e-06, 'learning_rate/action_model': 9.802238128132498e-05, 'epoch': 0.28}) +[2026-06-17 17:47:25,171][starVLA.training.train_starvla][INFO] - Step 737, Loss: {'train/loss': 0.6150409579277039, 'train/grad_norm_pre_clip': 9.016836166381836, 'timing/data': 0.031570833176374435, 'timing/model': 1.384152714163065, 'learning_rate/qwen_vl_interface': 9.801620115320836e-06, 'learning_rate/action_model': 9.801620115320835e-05, 'epoch': 0.28}) +[2026-06-17 17:47:29,882][starVLA.training.train_starvla][INFO] - Step 738, Loss: {'train/loss': 1.0415798425674438, 'train/grad_norm_pre_clip': 7.8244171142578125, 'timing/data': 0.034859780222177505, 'timing/model': 1.4084582598879933, 'learning_rate/qwen_vl_interface': 9.801001158953425e-06, 'learning_rate/action_model': 9.801001158953423e-05, 'epoch': 0.28}) +[2026-06-17 17:47:45,310][starVLA.training.train_starvla][INFO] - Step 739, Loss: {'train/loss': 0.5141213536262512, 'train/grad_norm_pre_clip': 8.761186599731445, 'timing/data': 0.02985504176467657, 'timing/model': 1.1695426469668746, 'learning_rate/qwen_vl_interface': 9.800381259158575e-06, 'learning_rate/action_model': 9.800381259158574e-05, 'epoch': 0.28}) +[2026-06-17 17:47:53,870][starVLA.training.train_starvla][INFO] - Step 740, Loss: {'train/loss': 0.758992075920105, 'train/grad_norm_pre_clip': 8.847806930541992, 'timing/data': 0.03859999403357506, 'timing/model': 4.749591409228742, 'learning_rate/qwen_vl_interface': 9.799760416064794e-06, 'learning_rate/action_model': 9.799760416064793e-05, 'epoch': 0.28}) +[2026-06-17 17:48:02,255][starVLA.training.train_starvla][INFO] - Step 741, Loss: {'train/loss': 0.9547867774963379, 'train/grad_norm_pre_clip': 11.383749008178711, 'timing/data': 0.038846878334879875, 'timing/model': 1.178278866223991, 'learning_rate/qwen_vl_interface': 9.799138629800785e-06, 'learning_rate/action_model': 9.799138629800785e-05, 'epoch': 0.29}) +[2026-06-17 17:48:07,221][starVLA.training.train_starvla][INFO] - Step 742, Loss: {'train/loss': 0.9406172633171082, 'train/grad_norm_pre_clip': 11.587828636169434, 'timing/data': 0.04735789727419615, 'timing/model': 1.6445955298841, 'learning_rate/qwen_vl_interface': 9.798515900495441e-06, 'learning_rate/action_model': 9.79851590049544e-05, 'epoch': 0.29}) +[2026-06-17 17:48:11,699][starVLA.training.train_starvla][INFO] - Step 743, Loss: {'train/loss': 0.5419028997421265, 'train/grad_norm_pre_clip': 4.716996192932129, 'timing/data': 0.032156312838196754, 'timing/model': 1.06808069255203, 'learning_rate/qwen_vl_interface': 9.797892228277857e-06, 'learning_rate/action_model': 9.797892228277857e-05, 'epoch': 0.29}) +[2026-06-17 17:48:16,501][starVLA.training.train_starvla][INFO] - Step 744, Loss: {'train/loss': 0.9640257358551025, 'train/grad_norm_pre_clip': 8.257530212402344, 'timing/data': 0.03983692359179258, 'timing/model': 1.166343916207552, 'learning_rate/qwen_vl_interface': 9.797267613277319e-06, 'learning_rate/action_model': 9.797267613277319e-05, 'epoch': 0.29}) +[2026-06-17 17:48:21,539][starVLA.training.train_starvla][INFO] - Step 745, Loss: {'train/loss': 0.5274255275726318, 'train/grad_norm_pre_clip': 4.981997013092041, 'timing/data': 0.033949773758649826, 'timing/model': 1.1260560816153884, 'learning_rate/qwen_vl_interface': 9.796642055623314e-06, 'learning_rate/action_model': 9.796642055623314e-05, 'epoch': 0.29}) +[2026-06-17 17:48:26,225][starVLA.training.train_starvla][INFO] - Step 746, Loss: {'train/loss': 0.4953560531139374, 'train/grad_norm_pre_clip': 5.916871070861816, 'timing/data': 0.03152591176331043, 'timing/model': 1.189781978726387, 'learning_rate/qwen_vl_interface': 9.796015555445518e-06, 'learning_rate/action_model': 9.796015555445517e-05, 'epoch': 0.29}) +[2026-06-17 17:48:31,098][starVLA.training.train_starvla][INFO] - Step 747, Loss: {'train/loss': 0.6900236010551453, 'train/grad_norm_pre_clip': 6.337297439575195, 'timing/data': 0.04110502079129219, 'timing/model': 1.675808809697628, 'learning_rate/qwen_vl_interface': 9.795388112873806e-06, 'learning_rate/action_model': 9.795388112873806e-05, 'epoch': 0.29}) +[2026-06-17 17:48:35,526][starVLA.training.train_starvla][INFO] - Step 748, Loss: {'train/loss': 0.5533215999603271, 'train/grad_norm_pre_clip': 7.735203742980957, 'timing/data': 0.033700739964842796, 'timing/model': 1.0789867648854852, 'learning_rate/qwen_vl_interface': 9.794759728038246e-06, 'learning_rate/action_model': 9.794759728038246e-05, 'epoch': 0.29}) +[2026-06-17 17:48:40,624][starVLA.training.train_starvla][INFO] - Step 749, Loss: {'train/loss': 0.6878243684768677, 'train/grad_norm_pre_clip': 6.64269495010376, 'timing/data': 0.03425676189363003, 'timing/model': 1.1704983217641711, 'learning_rate/qwen_vl_interface': 9.794130401069105e-06, 'learning_rate/action_model': 9.794130401069104e-05, 'epoch': 0.29}) +[2026-06-17 17:48:45,410][starVLA.training.train_starvla][INFO] - Step 750, Loss: {'train/loss': 0.7270338535308838, 'train/grad_norm_pre_clip': 6.32037878036499, 'timing/data': 0.03225817158818245, 'timing/model': 1.2187631065025926, 'learning_rate/qwen_vl_interface': 9.79350013209684e-06, 'learning_rate/action_model': 9.79350013209684e-05, 'epoch': 0.29}) +[2026-06-17 17:48:50,361][starVLA.training.train_starvla][INFO] - Step 751, Loss: {'train/loss': 0.34556448459625244, 'train/grad_norm_pre_clip': 5.982320308685303, 'timing/data': 0.03213039040565491, 'timing/model': 1.2854799600318074, 'learning_rate/qwen_vl_interface': 9.792868921252112e-06, 'learning_rate/action_model': 9.792868921252111e-05, 'epoch': 0.29}) +[2026-06-17 17:48:55,161][starVLA.training.train_starvla][INFO] - Step 752, Loss: {'train/loss': 0.5287168622016907, 'train/grad_norm_pre_clip': 4.8004937171936035, 'timing/data': 0.03983395453542471, 'timing/model': 1.4235228169709444, 'learning_rate/qwen_vl_interface': 9.792236768665765e-06, 'learning_rate/action_model': 9.792236768665764e-05, 'epoch': 0.29}) +[2026-06-17 17:48:59,633][starVLA.training.train_starvla][INFO] - Step 753, Loss: {'train/loss': 1.013249158859253, 'train/grad_norm_pre_clip': 11.1891450881958, 'timing/data': 0.03397096507251263, 'timing/model': 1.083311103284359, 'learning_rate/qwen_vl_interface': 9.79160367446885e-06, 'learning_rate/action_model': 9.79160367446885e-05, 'epoch': 0.29}) +[2026-06-17 17:49:04,466][starVLA.training.train_starvla][INFO] - Step 754, Loss: {'train/loss': 0.7119148969650269, 'train/grad_norm_pre_clip': 9.7845458984375, 'timing/data': 0.03968801349401474, 'timing/model': 1.2653912007808685, 'learning_rate/qwen_vl_interface': 9.790969638792606e-06, 'learning_rate/action_model': 9.790969638792606e-05, 'epoch': 0.29}) +[2026-06-17 17:49:09,471][starVLA.training.train_starvla][INFO] - Step 755, Loss: {'train/loss': 0.7415923476219177, 'train/grad_norm_pre_clip': 15.75678539276123, 'timing/data': 0.028885641135275364, 'timing/model': 1.2679409375414252, 'learning_rate/qwen_vl_interface': 9.790334661768471e-06, 'learning_rate/action_model': 9.79033466176847e-05, 'epoch': 0.29}) +[2026-06-17 17:49:22,000][starVLA.training.train_starvla][INFO] - Step 756, Loss: {'train/loss': 0.6551311016082764, 'train/grad_norm_pre_clip': 7.459217548370361, 'timing/data': 0.03565372992306948, 'timing/model': 2.84636995755136, 'learning_rate/qwen_vl_interface': 9.78969874352807e-06, 'learning_rate/action_model': 9.78969874352807e-05, 'epoch': 0.29}) +[2026-06-17 17:49:27,324][starVLA.training.train_starvla][INFO] - Step 757, Loss: {'train/loss': 0.34668785333633423, 'train/grad_norm_pre_clip': 7.346678256988525, 'timing/data': 0.03126715403050184, 'timing/model': 1.5144472168758512, 'learning_rate/qwen_vl_interface': 9.78906188420324e-06, 'learning_rate/action_model': 9.78906188420324e-05, 'epoch': 0.29}) +[2026-06-17 17:49:32,241][starVLA.training.train_starvla][INFO] - Step 758, Loss: {'train/loss': 0.6103430986404419, 'train/grad_norm_pre_clip': 9.247871398925781, 'timing/data': 0.040352839045226574, 'timing/model': 1.7141681080684066, 'learning_rate/qwen_vl_interface': 9.788424083925996e-06, 'learning_rate/action_model': 9.788424083925995e-05, 'epoch': 0.29}) +[2026-06-17 17:49:37,231][starVLA.training.train_starvla][INFO] - Step 759, Loss: {'train/loss': 0.5217959880828857, 'train/grad_norm_pre_clip': 10.122237205505371, 'timing/data': 0.033654206432402134, 'timing/model': 1.1812647990882397, 'learning_rate/qwen_vl_interface': 9.787785342828554e-06, 'learning_rate/action_model': 9.787785342828553e-05, 'epoch': 0.29}) +[2026-06-17 17:49:42,336][starVLA.training.train_starvla][INFO] - Step 760, Loss: {'train/loss': 0.7927712798118591, 'train/grad_norm_pre_clip': 12.083355903625488, 'timing/data': 0.03286805655807257, 'timing/model': 1.0973217617720366, 'learning_rate/qwen_vl_interface': 9.78714566104333e-06, 'learning_rate/action_model': 9.787145661043329e-05, 'epoch': 0.29}) +[2026-06-17 17:49:48,908][starVLA.training.train_starvla][INFO] - Step 761, Loss: {'train/loss': 0.7509185075759888, 'train/grad_norm_pre_clip': 11.060307502746582, 'timing/data': 0.055930071510374546, 'timing/model': 2.434277154505253, 'learning_rate/qwen_vl_interface': 9.786505038702926e-06, 'learning_rate/action_model': 9.786505038702925e-05, 'epoch': 0.29}) +[2026-06-17 17:49:57,481][starVLA.training.train_starvla][INFO] - Step 762, Loss: {'train/loss': 0.8492574095726013, 'train/grad_norm_pre_clip': 14.219810485839844, 'timing/data': 0.03339292760938406, 'timing/model': 1.9082641452550888, 'learning_rate/qwen_vl_interface': 9.785863475940147e-06, 'learning_rate/action_model': 9.785863475940145e-05, 'epoch': 0.29}) +[2026-06-17 17:50:12,561][starVLA.training.train_starvla][INFO] - Step 763, Loss: {'train/loss': 0.43044790625572205, 'train/grad_norm_pre_clip': 11.455745697021484, 'timing/data': 0.03183966502547264, 'timing/model': 4.2440793458372355, 'learning_rate/qwen_vl_interface': 9.785220972887988e-06, 'learning_rate/action_model': 9.785220972887989e-05, 'epoch': 0.29}) +[2026-06-17 17:50:17,445][starVLA.training.train_starvla][INFO] - Step 764, Loss: {'train/loss': 0.3792669177055359, 'train/grad_norm_pre_clip': 6.266643047332764, 'timing/data': 0.04107226338237524, 'timing/model': 1.1057693976908922, 'learning_rate/qwen_vl_interface': 9.784577529679643e-06, 'learning_rate/action_model': 9.784577529679643e-05, 'epoch': 0.29}) +[2026-06-17 17:50:22,231][starVLA.training.train_starvla][INFO] - Step 765, Loss: {'train/loss': 0.9641739130020142, 'train/grad_norm_pre_clip': 14.131584167480469, 'timing/data': 0.030929187312722206, 'timing/model': 1.163684906437993, 'learning_rate/qwen_vl_interface': 9.783933146448495e-06, 'learning_rate/action_model': 9.783933146448495e-05, 'epoch': 0.29}) +[2026-06-17 17:50:27,148][starVLA.training.train_starvla][INFO] - Step 766, Loss: {'train/loss': 0.5757789015769958, 'train/grad_norm_pre_clip': 6.706051826477051, 'timing/data': 0.03321823105216026, 'timing/model': 1.2751331822946668, 'learning_rate/qwen_vl_interface': 9.783287823328127e-06, 'learning_rate/action_model': 9.783287823328127e-05, 'epoch': 0.29}) +[2026-06-17 17:50:31,730][starVLA.training.train_starvla][INFO] - Step 767, Loss: {'train/loss': 0.7341786623001099, 'train/grad_norm_pre_clip': 12.546788215637207, 'timing/data': 0.029830762185156345, 'timing/model': 1.0870207529515028, 'learning_rate/qwen_vl_interface': 9.782641560452317e-06, 'learning_rate/action_model': 9.782641560452316e-05, 'epoch': 0.3}) +[2026-06-17 17:50:36,557][starVLA.training.train_starvla][INFO] - Step 768, Loss: {'train/loss': 0.3646389842033386, 'train/grad_norm_pre_clip': 6.135972023010254, 'timing/data': 0.03984817024320364, 'timing/model': 1.1711338674649596, 'learning_rate/qwen_vl_interface': 9.781994357955032e-06, 'learning_rate/action_model': 9.781994357955032e-05, 'epoch': 0.3}) +[2026-06-17 17:50:41,308][starVLA.training.train_starvla][INFO] - Step 769, Loss: {'train/loss': 0.402281790971756, 'train/grad_norm_pre_clip': 2.862185478210449, 'timing/data': 0.02244400791823864, 'timing/model': 1.0792856579646468, 'learning_rate/qwen_vl_interface': 9.78134621597044e-06, 'learning_rate/action_model': 9.781346215970441e-05, 'epoch': 0.3}) +[2026-06-17 17:50:46,206][starVLA.training.train_starvla][INFO] - Step 770, Loss: {'train/loss': 0.5121309757232666, 'train/grad_norm_pre_clip': 8.189448356628418, 'timing/data': 0.028104616329073906, 'timing/model': 1.0907319262623787, 'learning_rate/qwen_vl_interface': 9.780697134632903e-06, 'learning_rate/action_model': 9.780697134632902e-05, 'epoch': 0.3}) +[2026-06-17 17:50:51,128][starVLA.training.train_starvla][INFO] - Step 771, Loss: {'train/loss': 0.317256361246109, 'train/grad_norm_pre_clip': 5.375133037567139, 'timing/data': 0.03036775439977646, 'timing/model': 1.6596107669174671, 'learning_rate/qwen_vl_interface': 9.780047114076974e-06, 'learning_rate/action_model': 9.780047114076973e-05, 'epoch': 0.3}) +[2026-06-17 17:51:00,700][starVLA.training.train_starvla][INFO] - Step 772, Loss: {'train/loss': 1.338804006576538, 'train/grad_norm_pre_clip': 11.684671401977539, 'timing/data': 0.0238991966471076, 'timing/model': 2.760783616453409, 'learning_rate/qwen_vl_interface': 9.779396154437404e-06, 'learning_rate/action_model': 9.779396154437403e-05, 'epoch': 0.3}) +[2026-06-17 17:51:05,629][starVLA.training.train_starvla][INFO] - Step 773, Loss: {'train/loss': 0.67575603723526, 'train/grad_norm_pre_clip': 13.081088066101074, 'timing/data': 0.020435061305761337, 'timing/model': 1.0683140400797129, 'learning_rate/qwen_vl_interface': 9.778744255849138e-06, 'learning_rate/action_model': 9.778744255849138e-05, 'epoch': 0.3}) +[2026-06-17 17:51:10,434][starVLA.training.train_starvla][INFO] - Step 774, Loss: {'train/loss': 0.5451629161834717, 'train/grad_norm_pre_clip': 9.792092323303223, 'timing/data': 0.024362480267882347, 'timing/model': 1.262675454840064, 'learning_rate/qwen_vl_interface': 9.778091418447313e-06, 'learning_rate/action_model': 9.778091418447312e-05, 'epoch': 0.3}) +[2026-06-17 17:51:17,990][starVLA.training.train_starvla][INFO] - Step 775, Loss: {'train/loss': 0.9927921891212463, 'train/grad_norm_pre_clip': 12.443122863769531, 'timing/data': 0.0221093762665987, 'timing/model': 1.0759227983653545, 'learning_rate/qwen_vl_interface': 9.777437642367265e-06, 'learning_rate/action_model': 9.777437642367264e-05, 'epoch': 0.3}) +[2026-06-17 17:51:24,124][starVLA.training.train_starvla][INFO] - Step 776, Loss: {'train/loss': 0.7899581789970398, 'train/grad_norm_pre_clip': 9.769227027893066, 'timing/data': 0.02483492624014616, 'timing/model': 1.0847317771986127, 'learning_rate/qwen_vl_interface': 9.776782927744523e-06, 'learning_rate/action_model': 9.776782927744522e-05, 'epoch': 0.3}) +[2026-06-17 17:51:28,977][starVLA.training.train_starvla][INFO] - Step 777, Loss: {'train/loss': 0.7094220519065857, 'train/grad_norm_pre_clip': 7.444020748138428, 'timing/data': 0.026013505645096302, 'timing/model': 1.658529557287693, 'learning_rate/qwen_vl_interface': 9.776127274714808e-06, 'learning_rate/action_model': 9.776127274714807e-05, 'epoch': 0.3}) +[2026-06-17 17:51:33,333][starVLA.training.train_starvla][INFO] - Step 778, Loss: {'train/loss': 0.8108769655227661, 'train/grad_norm_pre_clip': 6.960175514221191, 'timing/data': 0.025914276018738747, 'timing/model': 1.0832111136987805, 'learning_rate/qwen_vl_interface': 9.775470683414038e-06, 'learning_rate/action_model': 9.775470683414038e-05, 'epoch': 0.3}) +[2026-06-17 17:51:37,914][starVLA.training.train_starvla][INFO] - Step 779, Loss: {'train/loss': 0.5534404516220093, 'train/grad_norm_pre_clip': 7.085450172424316, 'timing/data': 0.02140045166015625, 'timing/model': 1.0869105150923133, 'learning_rate/qwen_vl_interface': 9.774813153978327e-06, 'learning_rate/action_model': 9.774813153978327e-05, 'epoch': 0.3}) +[2026-06-17 17:51:44,950][starVLA.training.train_starvla][INFO] - Step 780, Loss: {'train/loss': 1.1072502136230469, 'train/grad_norm_pre_clip': 12.724611282348633, 'timing/data': 0.027234665118157864, 'timing/model': 1.172603121958673, 'learning_rate/qwen_vl_interface': 9.77415468654398e-06, 'learning_rate/action_model': 9.77415468654398e-05, 'epoch': 0.3}) +[2026-06-17 17:51:49,657][starVLA.training.train_starvla][INFO] - Step 781, Loss: {'train/loss': 0.7397933602333069, 'train/grad_norm_pre_clip': 8.584301948547363, 'timing/data': 0.02104975190013647, 'timing/model': 1.1588047621771693, 'learning_rate/qwen_vl_interface': 9.773495281247497e-06, 'learning_rate/action_model': 9.773495281247497e-05, 'epoch': 0.3}) +[2026-06-17 17:51:54,387][starVLA.training.train_starvla][INFO] - Step 782, Loss: {'train/loss': 0.7123590111732483, 'train/grad_norm_pre_clip': 6.24699068069458, 'timing/data': 0.026385053992271423, 'timing/model': 1.4378267582505941, 'learning_rate/qwen_vl_interface': 9.772834938225577e-06, 'learning_rate/action_model': 9.772834938225576e-05, 'epoch': 0.3}) +[2026-06-17 17:51:58,715][starVLA.training.train_starvla][INFO] - Step 783, Loss: {'train/loss': 0.7239261269569397, 'train/grad_norm_pre_clip': 12.304279327392578, 'timing/data': 0.028033221140503883, 'timing/model': 1.0808121347799897, 'learning_rate/qwen_vl_interface': 9.772173657615107e-06, 'learning_rate/action_model': 9.772173657615106e-05, 'epoch': 0.3}) +[2026-06-17 17:52:03,902][starVLA.training.train_starvla][INFO] - Step 784, Loss: {'train/loss': 0.4320283830165863, 'train/grad_norm_pre_clip': 5.530242919921875, 'timing/data': 0.025286097079515457, 'timing/model': 1.3107193019241095, 'learning_rate/qwen_vl_interface': 9.77151143955317e-06, 'learning_rate/action_model': 9.77151143955317e-05, 'epoch': 0.3}) +[2026-06-17 17:52:09,359][starVLA.training.train_starvla][INFO] - Step 785, Loss: {'train/loss': 0.49041953682899475, 'train/grad_norm_pre_clip': 5.953609466552734, 'timing/data': 0.03224568534642458, 'timing/model': 1.1090653911232948, 'learning_rate/qwen_vl_interface': 9.770848284177048e-06, 'learning_rate/action_model': 9.770848284177047e-05, 'epoch': 0.3}) +[2026-06-17 17:52:14,176][starVLA.training.train_starvla][INFO] - Step 786, Loss: {'train/loss': 0.7147811055183411, 'train/grad_norm_pre_clip': 9.765175819396973, 'timing/data': 0.020199217833578587, 'timing/model': 1.1888242680579424, 'learning_rate/qwen_vl_interface': 9.77018419162421e-06, 'learning_rate/action_model': 9.77018419162421e-05, 'epoch': 0.3}) +[2026-06-17 17:52:18,943][starVLA.training.train_starvla][INFO] - Step 787, Loss: {'train/loss': 0.6748232841491699, 'train/grad_norm_pre_clip': 5.374780654907227, 'timing/data': 0.020714331418275833, 'timing/model': 1.3928639832884073, 'learning_rate/qwen_vl_interface': 9.769519162032328e-06, 'learning_rate/action_model': 9.769519162032327e-05, 'epoch': 0.3}) +[2026-06-17 17:52:23,450][starVLA.training.train_starvla][INFO] - Step 788, Loss: {'train/loss': 0.6742116212844849, 'train/grad_norm_pre_clip': 7.594797611236572, 'timing/data': 0.026571919210255146, 'timing/model': 1.1427988493815064, 'learning_rate/qwen_vl_interface': 9.768853195539259e-06, 'learning_rate/action_model': 9.768853195539258e-05, 'epoch': 0.3}) +[2026-06-17 17:52:28,289][starVLA.training.train_starvla][INFO] - Step 789, Loss: {'train/loss': 0.43742814660072327, 'train/grad_norm_pre_clip': 5.345695495605469, 'timing/data': 0.025606470182538033, 'timing/model': 1.1954060103744268, 'learning_rate/qwen_vl_interface': 9.76818629228306e-06, 'learning_rate/action_model': 9.768186292283059e-05, 'epoch': 0.3}) +[2026-06-17 17:52:33,016][starVLA.training.train_starvla][INFO] - Step 790, Loss: {'train/loss': 0.6363821625709534, 'train/grad_norm_pre_clip': 6.55987548828125, 'timing/data': 0.033959721215069294, 'timing/model': 1.159778124652803, 'learning_rate/qwen_vl_interface': 9.767518452401981e-06, 'learning_rate/action_model': 9.76751845240198e-05, 'epoch': 0.3}) +[2026-06-17 17:52:37,561][starVLA.training.train_starvla][INFO] - Step 791, Loss: {'train/loss': 0.5313319563865662, 'train/grad_norm_pre_clip': 6.86891508102417, 'timing/data': 0.022586142644286156, 'timing/model': 1.1030775625258684, 'learning_rate/qwen_vl_interface': 9.766849676034465e-06, 'learning_rate/action_model': 9.766849676034464e-05, 'epoch': 0.3}) +[2026-06-17 17:52:42,371][starVLA.training.train_starvla][INFO] - Step 792, Loss: {'train/loss': 0.8370544910430908, 'train/grad_norm_pre_clip': 10.13199520111084, 'timing/data': 0.023700186982750893, 'timing/model': 1.4171970253810287, 'learning_rate/qwen_vl_interface': 9.766179963319152e-06, 'learning_rate/action_model': 9.76617996331915e-05, 'epoch': 0.3}) +[2026-06-17 17:52:46,787][starVLA.training.train_starvla][INFO] - Step 793, Loss: {'train/loss': 0.5066648125648499, 'train/grad_norm_pre_clip': 8.41005802154541, 'timing/data': 0.0216502845287323, 'timing/model': 1.1202601362019777, 'learning_rate/qwen_vl_interface': 9.765509314394872e-06, 'learning_rate/action_model': 9.765509314394872e-05, 'epoch': 0.31}) +[2026-06-17 17:52:51,409][starVLA.training.train_starvla][INFO] - Step 794, Loss: {'train/loss': 0.6059560775756836, 'train/grad_norm_pre_clip': 6.875948905944824, 'timing/data': 0.023267265409231186, 'timing/model': 1.0899731507524848, 'learning_rate/qwen_vl_interface': 9.764837729400652e-06, 'learning_rate/action_model': 9.764837729400653e-05, 'epoch': 0.31}) +[2026-06-17 17:52:59,280][starVLA.training.train_starvla][INFO] - Step 795, Loss: {'train/loss': 0.676618754863739, 'train/grad_norm_pre_clip': 6.829074859619141, 'timing/data': 0.02986667212098837, 'timing/model': 1.137677513062954, 'learning_rate/qwen_vl_interface': 9.764165208475714e-06, 'learning_rate/action_model': 9.764165208475714e-05, 'epoch': 0.31}) +[2026-06-17 17:53:04,047][starVLA.training.train_starvla][INFO] - Step 796, Loss: {'train/loss': 0.7334861159324646, 'train/grad_norm_pre_clip': 13.62918758392334, 'timing/data': 0.03005264326930046, 'timing/model': 1.0904180463403463, 'learning_rate/qwen_vl_interface': 9.76349175175947e-06, 'learning_rate/action_model': 9.763491751759469e-05, 'epoch': 0.31}) +[2026-06-17 17:53:09,103][starVLA.training.train_starvla][INFO] - Step 797, Loss: {'train/loss': 0.4725630581378937, 'train/grad_norm_pre_clip': 5.958674907684326, 'timing/data': 0.02297316212207079, 'timing/model': 1.482440765015781, 'learning_rate/qwen_vl_interface': 9.76281735939153e-06, 'learning_rate/action_model': 9.762817359391529e-05, 'epoch': 0.31}) +[2026-06-17 17:53:14,336][starVLA.training.train_starvla][INFO] - Step 798, Loss: {'train/loss': 0.31510862708091736, 'train/grad_norm_pre_clip': 8.177482604980469, 'timing/data': 0.02244628593325615, 'timing/model': 1.1500665517523885, 'learning_rate/qwen_vl_interface': 9.762142031511695e-06, 'learning_rate/action_model': 9.762142031511695e-05, 'epoch': 0.31}) +[2026-06-17 17:53:23,228][starVLA.training.train_starvla][INFO] - Step 799, Loss: {'train/loss': 0.5442636013031006, 'train/grad_norm_pre_clip': 8.545309066772461, 'timing/data': 0.022575887851417065, 'timing/model': 2.5869196811690927, 'learning_rate/qwen_vl_interface': 9.76146576825996e-06, 'learning_rate/action_model': 9.761465768259959e-05, 'epoch': 0.31}) +[2026-06-17 17:56:48,865][starVLA.training.train_starvla][INFO] - Step 800, Loss: {'train/loss': 0.5718985795974731, 'train/grad_norm_pre_clip': 15.107931137084961, 'eval/loss': 0.6822866797447205, 'eval/latency_4/loss': 0.6822866797447205, 'eval/demon_fire_precision': 0.9301270417422868, 'eval/demon_fire_recall': 0.9987008769080871, 'eval/demon_fire_f1': 0.9631949882537197, 'eval/demon_fire_rate_model': 0.9709251101321585, 'eval/demon_fire_rate_teacher': 0.9042584434654919, 'eval/demon_left_precision': 0.7453703703703703, 'eval/demon_left_recall': 0.3675799086757991, 'eval/demon_left_f1': 0.49235474006116214, 'eval/demon_left_rate_model': 0.06343612334801763, 'eval/demon_left_rate_teacher': 0.12863436123348018, 'eval/demon_right_precision': 0.7268722466960352, 'eval/demon_right_recall': 0.5956678700361011, 'eval/demon_right_f1': 0.6547619047619048, 'eval/demon_right_rate_model': 0.13333333333333333, 'eval/demon_right_rate_teacher': 0.16270190895741557, 'eval/demon_move_f1': 0.5735583224115335, 'eval/demon_cc_f1': 0.8001289651300831, 'eval/latency_4/demon_fire_precision': 0.9301270417422868, 'eval/latency_4/demon_fire_recall': 0.9987008769080871, 'eval/latency_4/demon_fire_f1': 0.9631949882537197, 'eval/latency_4/demon_fire_rate_model': 0.9709251101321585, 'eval/latency_4/demon_fire_rate_teacher': 0.9042584434654919, 'eval/latency_4/demon_left_precision': 0.7453703703703703, 'eval/latency_4/demon_left_recall': 0.3675799086757991, 'eval/latency_4/demon_left_f1': 0.49235474006116214, 'eval/latency_4/demon_left_rate_model': 0.06343612334801763, 'eval/latency_4/demon_left_rate_teacher': 0.12863436123348018, 'eval/latency_4/demon_right_precision': 0.7268722466960352, 'eval/latency_4/demon_right_recall': 0.5956678700361011, 'eval/latency_4/demon_right_f1': 0.6547619047619048, 'eval/latency_4/demon_right_rate_model': 0.13333333333333333, 'eval/latency_4/demon_right_rate_teacher': 0.16270190895741557, 'eval/latency_4/demon_move_f1': 0.5735583224115335, 'eval/latency_4/demon_cc_f1': 0.8001289651300831, 'timing/data': 0.023884257301688194, 'timing/model': 1.095254661515355, 'learning_rate/qwen_vl_interface': 9.760788569776518e-06, 'learning_rate/action_model': 9.760788569776517e-05, 'epoch': 0.31}) +[2026-06-17 17:56:48,867][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_800_state +[2026-06-17 17:56:48,875][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 17:56:57,553][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_800_state/model.safetensors +[2026-06-17 17:57:19,504][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_800_state/optimizer.bin +[2026-06-17 17:57:19,506][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_800_state/sampler.bin +[2026-06-17 17:57:19,507][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_800_state/sampler_1.bin +[2026-06-17 17:57:19,509][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_800_state/random_states_0.pkl +[2026-06-17 17:57:19,516][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 800 +[2026-06-17 17:57:19,517][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 17:57:19,535][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 17:57:26,214][starVLA.training.train_starvla][INFO] - Step 801, Loss: {'train/loss': 0.6399685740470886, 'train/grad_norm_pre_clip': 6.411799907684326, 'timing/data': 0.03202989883720875, 'timing/model': 1.105933473445475, 'learning_rate/qwen_vl_interface': 9.760110436201752e-06, 'learning_rate/action_model': 9.76011043620175e-05, 'epoch': 0.31}) +[2026-06-17 17:57:30,809][starVLA.training.train_starvla][INFO] - Step 802, Loss: {'train/loss': 0.8829273581504822, 'train/grad_norm_pre_clip': 24.262638092041016, 'timing/data': 0.020067845471203327, 'timing/model': 1.403778137639165, 'learning_rate/qwen_vl_interface': 9.759431367676238e-06, 'learning_rate/action_model': 9.759431367676237e-05, 'epoch': 0.31}) +[2026-06-17 17:57:34,990][starVLA.training.train_starvla][INFO] - Step 803, Loss: {'train/loss': 0.5842077136039734, 'train/grad_norm_pre_clip': 15.194710731506348, 'timing/data': 0.02210143394768238, 'timing/model': 1.0688571110367775, 'learning_rate/qwen_vl_interface': 9.758751364340751e-06, 'learning_rate/action_model': 9.75875136434075e-05, 'epoch': 0.31}) +[2026-06-17 17:57:39,813][starVLA.training.train_starvla][INFO] - Step 804, Loss: {'train/loss': 0.42900416254997253, 'train/grad_norm_pre_clip': 9.058419227600098, 'timing/data': 0.028351791203022003, 'timing/model': 1.1010500192642212, 'learning_rate/qwen_vl_interface': 9.758070426336252e-06, 'learning_rate/action_model': 9.758070426336252e-05, 'epoch': 0.31}) +[2026-06-17 17:57:44,468][starVLA.training.train_starvla][INFO] - Step 805, Loss: {'train/loss': 1.1033581495285034, 'train/grad_norm_pre_clip': 9.78394889831543, 'timing/data': 0.023308156058192253, 'timing/model': 1.0566472681239247, 'learning_rate/qwen_vl_interface': 9.757388553803905e-06, 'learning_rate/action_model': 9.757388553803905e-05, 'epoch': 0.31}) +[2026-06-17 17:57:48,883][starVLA.training.train_starvla][INFO] - Step 806, Loss: {'train/loss': 1.123287320137024, 'train/grad_norm_pre_clip': 14.872325897216797, 'timing/data': 0.018681902438402176, 'timing/model': 1.0318468855693936, 'learning_rate/qwen_vl_interface': 9.75670574688506e-06, 'learning_rate/action_model': 9.75670574688506e-05, 'epoch': 0.31}) +[2026-06-17 17:57:53,193][starVLA.training.train_starvla][INFO] - Step 807, Loss: {'train/loss': 0.9251108169555664, 'train/grad_norm_pre_clip': 9.51021957397461, 'timing/data': 0.018578422255814075, 'timing/model': 1.2854621512815356, 'learning_rate/qwen_vl_interface': 9.756022005721265e-06, 'learning_rate/action_model': 9.756022005721265e-05, 'epoch': 0.31}) +[2026-06-17 17:57:57,330][starVLA.training.train_starvla][INFO] - Step 808, Loss: {'train/loss': 0.957146167755127, 'train/grad_norm_pre_clip': 11.028730392456055, 'timing/data': 0.021742315031588078, 'timing/model': 1.0549707990139723, 'learning_rate/qwen_vl_interface': 9.755337330454257e-06, 'learning_rate/action_model': 9.755337330454257e-05, 'epoch': 0.31}) +[2026-06-17 17:58:01,865][starVLA.training.train_starvla][INFO] - Step 809, Loss: {'train/loss': 0.7466985583305359, 'train/grad_norm_pre_clip': 7.957074165344238, 'timing/data': 0.032508065924048424, 'timing/model': 1.0878184596076608, 'learning_rate/qwen_vl_interface': 9.754651721225975e-06, 'learning_rate/action_model': 9.754651721225975e-05, 'epoch': 0.31}) +[2026-06-17 17:58:07,062][starVLA.training.train_starvla][INFO] - Step 810, Loss: {'train/loss': 0.6207554340362549, 'train/grad_norm_pre_clip': 9.525853157043457, 'timing/data': 0.019807886332273483, 'timing/model': 1.9738700101152062, 'learning_rate/qwen_vl_interface': 9.753965178178543e-06, 'learning_rate/action_model': 9.753965178178543e-05, 'epoch': 0.31}) +[2026-06-17 17:58:11,436][starVLA.training.train_starvla][INFO] - Step 811, Loss: {'train/loss': 0.4977436661720276, 'train/grad_norm_pre_clip': 5.280289649963379, 'timing/data': 0.018967649899423122, 'timing/model': 1.044075551442802, 'learning_rate/qwen_vl_interface': 9.753277701454285e-06, 'learning_rate/action_model': 9.753277701454284e-05, 'epoch': 0.31}) +[2026-06-17 17:58:15,739][starVLA.training.train_starvla][INFO] - Step 812, Loss: {'train/loss': 0.7830345630645752, 'train/grad_norm_pre_clip': 7.187257766723633, 'timing/data': 0.03312098141759634, 'timing/model': 1.2917566057294607, 'learning_rate/qwen_vl_interface': 9.752589291195711e-06, 'learning_rate/action_model': 9.75258929119571e-05, 'epoch': 0.31}) +[2026-06-17 17:58:19,790][starVLA.training.train_starvla][INFO] - Step 813, Loss: {'train/loss': 0.7341424226760864, 'train/grad_norm_pre_clip': 10.352538108825684, 'timing/data': 0.020074496045708656, 'timing/model': 1.0444610668346286, 'learning_rate/qwen_vl_interface': 9.751899947545534e-06, 'learning_rate/action_model': 9.751899947545533e-05, 'epoch': 0.31}) +[2026-06-17 17:58:24,090][starVLA.training.train_starvla][INFO] - Step 814, Loss: {'train/loss': 0.7642132043838501, 'train/grad_norm_pre_clip': 8.27896499633789, 'timing/data': 0.02159311342984438, 'timing/model': 1.0461886562407017, 'learning_rate/qwen_vl_interface': 9.751209670646653e-06, 'learning_rate/action_model': 9.751209670646652e-05, 'epoch': 0.31}) +[2026-06-17 17:58:28,454][starVLA.training.train_starvla][INFO] - Step 815, Loss: {'train/loss': 0.6439335942268372, 'train/grad_norm_pre_clip': 40.98155212402344, 'timing/data': 0.02462756074965, 'timing/model': 1.0623763771727681, 'learning_rate/qwen_vl_interface': 9.750518460642162e-06, 'learning_rate/action_model': 9.75051846064216e-05, 'epoch': 0.31}) +[2026-06-17 17:58:32,967][starVLA.training.train_starvla][INFO] - Step 816, Loss: {'train/loss': 0.7578650712966919, 'train/grad_norm_pre_clip': 6.0297698974609375, 'timing/data': 0.032649606466293335, 'timing/model': 1.1097420919686556, 'learning_rate/qwen_vl_interface': 9.749826317675352e-06, 'learning_rate/action_model': 9.749826317675352e-05, 'epoch': 0.31}) +[2026-06-17 17:58:37,690][starVLA.training.train_starvla][INFO] - Step 817, Loss: {'train/loss': 0.7902257442474365, 'train/grad_norm_pre_clip': 9.870477676391602, 'timing/data': 0.020191136747598648, 'timing/model': 1.4035209426656365, 'learning_rate/qwen_vl_interface': 9.749133241889706e-06, 'learning_rate/action_model': 9.749133241889705e-05, 'epoch': 0.31}) +[2026-06-17 17:58:42,089][starVLA.training.train_starvla][INFO] - Step 818, Loss: {'train/loss': 0.6489297151565552, 'train/grad_norm_pre_clip': 6.289400100708008, 'timing/data': 0.031677547842264175, 'timing/model': 1.118788638152182, 'learning_rate/qwen_vl_interface': 9.748439233428897e-06, 'learning_rate/action_model': 9.748439233428896e-05, 'epoch': 0.31}) +[2026-06-17 17:58:48,604][starVLA.training.train_starvla][INFO] - Step 819, Loss: {'train/loss': 0.4532466530799866, 'train/grad_norm_pre_clip': 15.158621788024902, 'timing/data': 0.022389797493815422, 'timing/model': 1.066186748445034, 'learning_rate/qwen_vl_interface': 9.747744292436794e-06, 'learning_rate/action_model': 9.747744292436793e-05, 'epoch': 0.32}) +[2026-06-17 17:58:53,283][starVLA.training.train_starvla][INFO] - Step 820, Loss: {'train/loss': 0.7035443186759949, 'train/grad_norm_pre_clip': 5.516571998596191, 'timing/data': 0.024634568952023983, 'timing/model': 1.1318780444562435, 'learning_rate/qwen_vl_interface': 9.747048419057458e-06, 'learning_rate/action_model': 9.747048419057458e-05, 'epoch': 0.32}) +[2026-06-17 17:58:57,981][starVLA.training.train_starvla][INFO] - Step 821, Loss: {'train/loss': 0.702182948589325, 'train/grad_norm_pre_clip': 6.062291622161865, 'timing/data': 0.031134119257330894, 'timing/model': 1.1228159554302692, 'learning_rate/qwen_vl_interface': 9.746351613435148e-06, 'learning_rate/action_model': 9.746351613435147e-05, 'epoch': 0.32}) +[2026-06-17 17:59:02,847][starVLA.training.train_starvla][INFO] - Step 822, Loss: {'train/loss': 0.7774845361709595, 'train/grad_norm_pre_clip': 5.723268985748291, 'timing/data': 0.02234900090843439, 'timing/model': 1.5008229641243815, 'learning_rate/qwen_vl_interface': 9.74565387571431e-06, 'learning_rate/action_model': 9.745653875714309e-05, 'epoch': 0.32}) +[2026-06-17 17:59:07,156][starVLA.training.train_starvla][INFO] - Step 823, Loss: {'train/loss': 0.7432774305343628, 'train/grad_norm_pre_clip': 6.528942584991455, 'timing/data': 0.020531403832137585, 'timing/model': 1.1110945828258991, 'learning_rate/qwen_vl_interface': 9.744955206039585e-06, 'learning_rate/action_model': 9.744955206039585e-05, 'epoch': 0.32}) +[2026-06-17 17:59:11,953][starVLA.training.train_starvla][INFO] - Step 824, Loss: {'train/loss': 0.6132854223251343, 'train/grad_norm_pre_clip': 7.417403697967529, 'timing/data': 0.03546044696122408, 'timing/model': 1.1249897684901953, 'learning_rate/qwen_vl_interface': 9.74425560455581e-06, 'learning_rate/action_model': 9.74425560455581e-05, 'epoch': 0.32}) +[2026-06-17 17:59:16,710][starVLA.training.train_starvla][INFO] - Step 825, Loss: {'train/loss': 0.6542482972145081, 'train/grad_norm_pre_clip': 6.430919647216797, 'timing/data': 0.02610183972865343, 'timing/model': 1.094460729509592, 'learning_rate/qwen_vl_interface': 9.743555071408011e-06, 'learning_rate/action_model': 9.743555071408011e-05, 'epoch': 0.32}) +[2026-06-17 17:59:23,216][starVLA.training.train_starvla][INFO] - Step 826, Loss: {'train/loss': 0.592078447341919, 'train/grad_norm_pre_clip': 5.724738121032715, 'timing/data': 0.022152029909193516, 'timing/model': 2.863820591941476, 'learning_rate/qwen_vl_interface': 9.742853606741412e-06, 'learning_rate/action_model': 9.742853606741411e-05, 'epoch': 0.32}) +[2026-06-17 17:59:28,099][starVLA.training.train_starvla][INFO] - Step 827, Loss: {'train/loss': 0.4986264109611511, 'train/grad_norm_pre_clip': 9.959245681762695, 'timing/data': 0.02316203061491251, 'timing/model': 1.3805113080888987, 'learning_rate/qwen_vl_interface': 9.742151210701423e-06, 'learning_rate/action_model': 9.742151210701424e-05, 'epoch': 0.32}) +[2026-06-17 17:59:32,393][starVLA.training.train_starvla][INFO] - Step 828, Loss: {'train/loss': 0.6875967979431152, 'train/grad_norm_pre_clip': 7.508326530456543, 'timing/data': 0.021405724808573723, 'timing/model': 1.1211123168468475, 'learning_rate/qwen_vl_interface': 9.741447883433656e-06, 'learning_rate/action_model': 9.741447883433656e-05, 'epoch': 0.32}) +[2026-06-17 17:59:37,865][starVLA.training.train_starvla][INFO] - Step 829, Loss: {'train/loss': 0.4311898946762085, 'train/grad_norm_pre_clip': 5.638940811157227, 'timing/data': 0.0279453843832016, 'timing/model': 1.1010735388845205, 'learning_rate/qwen_vl_interface': 9.74074362508391e-06, 'learning_rate/action_model': 9.740743625083911e-05, 'epoch': 0.32}) +[2026-06-17 17:59:46,844][starVLA.training.train_starvla][INFO] - Step 830, Loss: {'train/loss': 0.3518103361129761, 'train/grad_norm_pre_clip': 4.152699947357178, 'timing/data': 0.02306464035063982, 'timing/model': 1.7538626128807664, 'learning_rate/qwen_vl_interface': 9.74003843579818e-06, 'learning_rate/action_model': 9.740038435798178e-05, 'epoch': 0.32}) +[2026-06-17 17:59:52,922][starVLA.training.train_starvla][INFO] - Step 831, Loss: {'train/loss': 0.7695751190185547, 'train/grad_norm_pre_clip': 7.002887725830078, 'timing/data': 0.025625369511544704, 'timing/model': 1.748946820385754, 'learning_rate/qwen_vl_interface': 9.739332315722647e-06, 'learning_rate/action_model': 9.739332315722646e-05, 'epoch': 0.32}) +[2026-06-17 18:00:06,810][starVLA.training.train_starvla][INFO] - Step 832, Loss: {'train/loss': 0.6279260516166687, 'train/grad_norm_pre_clip': 6.893983364105225, 'timing/data': 0.021659716963768005, 'timing/model': 5.741201183758676, 'learning_rate/qwen_vl_interface': 9.738625265003695e-06, 'learning_rate/action_model': 9.738625265003695e-05, 'epoch': 0.32}) +[2026-06-17 18:00:13,133][starVLA.training.train_starvla][INFO] - Step 833, Loss: {'train/loss': 0.6713226437568665, 'train/grad_norm_pre_clip': 8.212615013122559, 'timing/data': 0.025518913753330708, 'timing/model': 1.2409977735951543, 'learning_rate/qwen_vl_interface': 9.7379172837879e-06, 'learning_rate/action_model': 9.737917283787898e-05, 'epoch': 0.32}) +[2026-06-17 18:00:17,823][starVLA.training.train_starvla][INFO] - Step 834, Loss: {'train/loss': 0.8874143958091736, 'train/grad_norm_pre_clip': 5.82181453704834, 'timing/data': 0.02727824728935957, 'timing/model': 1.085414806380868, 'learning_rate/qwen_vl_interface': 9.737208372222018e-06, 'learning_rate/action_model': 9.737208372222017e-05, 'epoch': 0.32}) +[2026-06-17 18:00:22,615][starVLA.training.train_starvla][INFO] - Step 835, Loss: {'train/loss': 0.6577526330947876, 'train/grad_norm_pre_clip': 7.037896633148193, 'timing/data': 0.025339284911751747, 'timing/model': 1.0747335879132152, 'learning_rate/qwen_vl_interface': 9.736498530453015e-06, 'learning_rate/action_model': 9.736498530453014e-05, 'epoch': 0.32}) +[2026-06-17 18:00:27,486][starVLA.training.train_starvla][INFO] - Step 836, Loss: {'train/loss': 0.4558415412902832, 'train/grad_norm_pre_clip': 7.099072456359863, 'timing/data': 0.021806524135172367, 'timing/model': 1.1946854013949633, 'learning_rate/qwen_vl_interface': 9.735787758628036e-06, 'learning_rate/action_model': 9.735787758628035e-05, 'epoch': 0.32}) +[2026-06-17 18:00:32,541][starVLA.training.train_starvla][INFO] - Step 837, Loss: {'train/loss': 0.34510815143585205, 'train/grad_norm_pre_clip': 5.019652843475342, 'timing/data': 0.0208259429782629, 'timing/model': 1.6544451331719756, 'learning_rate/qwen_vl_interface': 9.73507605689443e-06, 'learning_rate/action_model': 9.735076056894429e-05, 'epoch': 0.32}) +[2026-06-17 18:00:37,068][starVLA.training.train_starvla][INFO] - Step 838, Loss: {'train/loss': 0.6029143929481506, 'train/grad_norm_pre_clip': 43.437782287597656, 'timing/data': 0.028185829520225525, 'timing/model': 1.1210803613066673, 'learning_rate/qwen_vl_interface': 9.73436342539973e-06, 'learning_rate/action_model': 9.734363425399728e-05, 'epoch': 0.32}) +[2026-06-17 18:00:41,932][starVLA.training.train_starvla][INFO] - Step 839, Loss: {'train/loss': 1.0827502012252808, 'train/grad_norm_pre_clip': 14.042383193969727, 'timing/data': 0.024668757803738117, 'timing/model': 1.1481452081352472, 'learning_rate/qwen_vl_interface': 9.733649864291667e-06, 'learning_rate/action_model': 9.733649864291666e-05, 'epoch': 0.32}) +[2026-06-17 18:00:46,888][starVLA.training.train_starvla][INFO] - Step 840, Loss: {'train/loss': 0.6861620545387268, 'train/grad_norm_pre_clip': 9.33924388885498, 'timing/data': 0.02175798360258341, 'timing/model': 1.0795131819322705, 'learning_rate/qwen_vl_interface': 9.732935373718162e-06, 'learning_rate/action_model': 9.732935373718161e-05, 'epoch': 0.32}) +[2026-06-17 18:00:52,264][starVLA.training.train_starvla][INFO] - Step 841, Loss: {'train/loss': 0.6445237398147583, 'train/grad_norm_pre_clip': 11.839127540588379, 'timing/data': 0.02066747657954693, 'timing/model': 1.3862117705866694, 'learning_rate/qwen_vl_interface': 9.73221995382733e-06, 'learning_rate/action_model': 9.73221995382733e-05, 'epoch': 0.32}) +[2026-06-17 18:00:56,960][starVLA.training.train_starvla][INFO] - Step 842, Loss: {'train/loss': 0.5282796621322632, 'train/grad_norm_pre_clip': 7.50485372543335, 'timing/data': 0.02250627987086773, 'timing/model': 1.463639079593122, 'learning_rate/qwen_vl_interface': 9.73150360476748e-06, 'learning_rate/action_model': 9.73150360476748e-05, 'epoch': 0.32}) +[2026-06-17 18:01:01,466][starVLA.training.train_starvla][INFO] - Step 843, Loss: {'train/loss': 0.730218231678009, 'train/grad_norm_pre_clip': 7.765782356262207, 'timing/data': 0.02795402705669403, 'timing/model': 1.1746629774570465, 'learning_rate/qwen_vl_interface': 9.73078632668711e-06, 'learning_rate/action_model': 9.73078632668711e-05, 'epoch': 0.32}) +[2026-06-17 18:01:06,095][starVLA.training.train_starvla][INFO] - Step 844, Loss: {'train/loss': 0.5443635582923889, 'train/grad_norm_pre_clip': 5.1180925369262695, 'timing/data': 0.01918622199445963, 'timing/model': 1.0869166990742087, 'learning_rate/qwen_vl_interface': 9.730068119734912e-06, 'learning_rate/action_model': 9.730068119734912e-05, 'epoch': 0.32}) +[2026-06-17 18:01:10,751][starVLA.training.train_starvla][INFO] - Step 845, Loss: {'train/loss': 0.503183901309967, 'train/grad_norm_pre_clip': 7.251347541809082, 'timing/data': 0.021926009096205235, 'timing/model': 1.064770364202559, 'learning_rate/qwen_vl_interface': 9.729348984059773e-06, 'learning_rate/action_model': 9.729348984059774e-05, 'epoch': 0.33}) +[2026-06-17 18:01:15,385][starVLA.training.train_starvla][INFO] - Step 846, Loss: {'train/loss': 0.5368965268135071, 'train/grad_norm_pre_clip': 8.869590759277344, 'timing/data': 0.021938406862318516, 'timing/model': 1.0759885367006063, 'learning_rate/qwen_vl_interface': 9.728628919810769e-06, 'learning_rate/action_model': 9.728628919810768e-05, 'epoch': 0.33}) +[2026-06-17 18:01:20,099][starVLA.training.train_starvla][INFO] - Step 847, Loss: {'train/loss': 0.4229585528373718, 'train/grad_norm_pre_clip': 7.128658294677734, 'timing/data': 0.020106293261051178, 'timing/model': 1.4595573600381613, 'learning_rate/qwen_vl_interface': 9.727907927137171e-06, 'learning_rate/action_model': 9.727907927137171e-05, 'epoch': 0.33}) +[2026-06-17 18:01:24,639][starVLA.training.train_starvla][INFO] - Step 848, Loss: {'train/loss': 0.7074492573738098, 'train/grad_norm_pre_clip': 12.599166870117188, 'timing/data': 0.027283254079520702, 'timing/model': 1.1386343846097589, 'learning_rate/qwen_vl_interface': 9.72718600618844e-06, 'learning_rate/action_model': 9.727186006188441e-05, 'epoch': 0.33}) +[2026-06-17 18:01:29,313][starVLA.training.train_starvla][INFO] - Step 849, Loss: {'train/loss': 0.45760345458984375, 'train/grad_norm_pre_clip': 7.502320289611816, 'timing/data': 0.0347856841981411, 'timing/model': 1.214534055441618, 'learning_rate/qwen_vl_interface': 9.726463157114233e-06, 'learning_rate/action_model': 9.726463157114234e-05, 'epoch': 0.33}) +[2026-06-17 18:01:34,186][starVLA.training.train_starvla][INFO] - Step 850, Loss: {'train/loss': 0.8015880584716797, 'train/grad_norm_pre_clip': 8.146820068359375, 'timing/data': 0.027823914773762226, 'timing/model': 1.1404603570699692, 'learning_rate/qwen_vl_interface': 9.725739380064399e-06, 'learning_rate/action_model': 9.725739380064398e-05, 'epoch': 0.33}) +[2026-06-17 18:01:38,897][starVLA.training.train_starvla][INFO] - Step 851, Loss: {'train/loss': 0.6509061455726624, 'train/grad_norm_pre_clip': 7.192780494689941, 'timing/data': 0.019610359333455563, 'timing/model': 1.0964200887829065, 'learning_rate/qwen_vl_interface': 9.725014675188972e-06, 'learning_rate/action_model': 9.725014675188973e-05, 'epoch': 0.33}) +[2026-06-17 18:01:43,532][starVLA.training.train_starvla][INFO] - Step 852, Loss: {'train/loss': 0.8372798562049866, 'train/grad_norm_pre_clip': 10.320866584777832, 'timing/data': 0.020087272860109806, 'timing/model': 1.0704440800473094, 'learning_rate/qwen_vl_interface': 9.72428904263819e-06, 'learning_rate/action_model': 9.72428904263819e-05, 'epoch': 0.33}) +[2026-06-17 18:01:48,296][starVLA.training.train_starvla][INFO] - Step 853, Loss: {'train/loss': 0.7650392055511475, 'train/grad_norm_pre_clip': 9.237468719482422, 'timing/data': 0.022461633197963238, 'timing/model': 1.0728855226188898, 'learning_rate/qwen_vl_interface': 9.723562482562474e-06, 'learning_rate/action_model': 9.723562482562473e-05, 'epoch': 0.33}) +[2026-06-17 18:01:53,269][starVLA.training.train_starvla][INFO] - Step 854, Loss: {'train/loss': 0.5153301358222961, 'train/grad_norm_pre_clip': 8.090110778808594, 'timing/data': 0.023646293208003044, 'timing/model': 1.1589754819869995, 'learning_rate/qwen_vl_interface': 9.72283499511244e-06, 'learning_rate/action_model': 9.722834995112439e-05, 'epoch': 0.33}) +[2026-06-17 18:01:57,995][starVLA.training.train_starvla][INFO] - Step 855, Loss: {'train/loss': 0.8023538589477539, 'train/grad_norm_pre_clip': 6.882640361785889, 'timing/data': 0.02462573442608118, 'timing/model': 1.1867205956950784, 'learning_rate/qwen_vl_interface': 9.722106580438901e-06, 'learning_rate/action_model': 9.7221065804389e-05, 'epoch': 0.33}) +[2026-06-17 18:02:02,747][starVLA.training.train_starvla][INFO] - Step 856, Loss: {'train/loss': 0.43353354930877686, 'train/grad_norm_pre_clip': 4.9187541007995605, 'timing/data': 0.022678296081721783, 'timing/model': 1.580506093800068, 'learning_rate/qwen_vl_interface': 9.721377238692855e-06, 'learning_rate/action_model': 9.721377238692854e-05, 'epoch': 0.33}) +[2026-06-17 18:02:07,286][starVLA.training.train_starvla][INFO] - Step 857, Loss: {'train/loss': 0.5910806059837341, 'train/grad_norm_pre_clip': 7.62232780456543, 'timing/data': 0.022131488658487797, 'timing/model': 1.365463281981647, 'learning_rate/qwen_vl_interface': 9.720646970025494e-06, 'learning_rate/action_model': 9.720646970025494e-05, 'epoch': 0.33}) +[2026-06-17 18:02:11,953][starVLA.training.train_starvla][INFO] - Step 858, Loss: {'train/loss': 0.6560582518577576, 'train/grad_norm_pre_clip': 7.062966823577881, 'timing/data': 0.025456909090280533, 'timing/model': 1.4650766998529434, 'learning_rate/qwen_vl_interface': 9.719915774588209e-06, 'learning_rate/action_model': 9.719915774588209e-05, 'epoch': 0.33}) +[2026-06-17 18:02:16,714][starVLA.training.train_starvla][INFO] - Step 859, Loss: {'train/loss': 0.38391387462615967, 'train/grad_norm_pre_clip': 9.008206367492676, 'timing/data': 0.02692585252225399, 'timing/model': 1.0985425468534231, 'learning_rate/qwen_vl_interface': 9.719183652532572e-06, 'learning_rate/action_model': 9.719183652532572e-05, 'epoch': 0.33}) +[2026-06-17 18:02:21,761][starVLA.training.train_starvla][INFO] - Step 860, Loss: {'train/loss': 0.5128323435783386, 'train/grad_norm_pre_clip': 6.896451473236084, 'timing/data': 0.022034049034118652, 'timing/model': 1.2635240210220218, 'learning_rate/qwen_vl_interface': 9.718450604010355e-06, 'learning_rate/action_model': 9.718450604010355e-05, 'epoch': 0.33}) +[2026-06-17 18:02:26,792][starVLA.training.train_starvla][INFO] - Step 861, Loss: {'train/loss': 0.7029898762702942, 'train/grad_norm_pre_clip': 8.094583511352539, 'timing/data': 0.021228952333331108, 'timing/model': 1.0776578830555081, 'learning_rate/qwen_vl_interface': 9.717716629173521e-06, 'learning_rate/action_model': 9.71771662917352e-05, 'epoch': 0.33}) +[2026-06-17 18:02:31,630][starVLA.training.train_starvla][INFO] - Step 862, Loss: {'train/loss': 0.6060183048248291, 'train/grad_norm_pre_clip': 8.012186050415039, 'timing/data': 0.02651576977223158, 'timing/model': 1.4400609508156776, 'learning_rate/qwen_vl_interface': 9.71698172817422e-06, 'learning_rate/action_model': 9.716981728174219e-05, 'epoch': 0.33}) +[2026-06-17 18:02:38,499][starVLA.training.train_starvla][INFO] - Step 863, Loss: {'train/loss': 1.591604471206665, 'train/grad_norm_pre_clip': 12.510734558105469, 'timing/data': 0.026527502574026585, 'timing/model': 1.2009915364906192, 'learning_rate/qwen_vl_interface': 9.7162459011648e-06, 'learning_rate/action_model': 9.7162459011648e-05, 'epoch': 0.33}) +[2026-06-17 18:02:43,564][starVLA.training.train_starvla][INFO] - Step 864, Loss: {'train/loss': 0.407832533121109, 'train/grad_norm_pre_clip': 6.0014214515686035, 'timing/data': 0.022000263445079327, 'timing/model': 1.1210049539804459, 'learning_rate/qwen_vl_interface': 9.7155091482978e-06, 'learning_rate/action_model': 9.7155091482978e-05, 'epoch': 0.33}) +[2026-06-17 18:02:48,564][starVLA.training.train_starvla][INFO] - Step 865, Loss: {'train/loss': 0.8270939588546753, 'train/grad_norm_pre_clip': 6.894428253173828, 'timing/data': 0.02020424697548151, 'timing/model': 1.1424588542431593, 'learning_rate/qwen_vl_interface': 9.714771469725947e-06, 'learning_rate/action_model': 9.714771469725947e-05, 'epoch': 0.33}) +[2026-06-17 18:02:53,195][starVLA.training.train_starvla][INFO] - Step 866, Loss: {'train/loss': 1.0881922245025635, 'train/grad_norm_pre_clip': 11.881509780883789, 'timing/data': 0.02255272399634123, 'timing/model': 1.0930100539699197, 'learning_rate/qwen_vl_interface': 9.714032865602167e-06, 'learning_rate/action_model': 9.714032865602166e-05, 'epoch': 0.33}) +[2026-06-17 18:02:59,453][starVLA.training.train_starvla][INFO] - Step 867, Loss: {'train/loss': 0.6858280897140503, 'train/grad_norm_pre_clip': 6.698771953582764, 'timing/data': 0.029661295004189014, 'timing/model': 2.527370320633054, 'learning_rate/qwen_vl_interface': 9.713293336079568e-06, 'learning_rate/action_model': 9.713293336079568e-05, 'epoch': 0.33}) +[2026-06-17 18:03:03,947][starVLA.training.train_starvla][INFO] - Step 868, Loss: {'train/loss': 0.7765710353851318, 'train/grad_norm_pre_clip': 8.59457778930664, 'timing/data': 0.034233054146170616, 'timing/model': 1.2194902747869492, 'learning_rate/qwen_vl_interface': 9.712552881311459e-06, 'learning_rate/action_model': 9.712552881311458e-05, 'epoch': 0.33}) +[2026-06-17 18:03:08,722][starVLA.training.train_starvla][INFO] - Step 869, Loss: {'train/loss': 0.5030353665351868, 'train/grad_norm_pre_clip': 6.578168869018555, 'timing/data': 0.02330698072910309, 'timing/model': 1.1068034674972296, 'learning_rate/qwen_vl_interface': 9.711811501451336e-06, 'learning_rate/action_model': 9.711811501451335e-05, 'epoch': 0.33}) +[2026-06-17 18:03:13,764][starVLA.training.train_starvla][INFO] - Step 870, Loss: {'train/loss': 0.6097075343132019, 'train/grad_norm_pre_clip': 11.63031005859375, 'timing/data': 0.027135753072798252, 'timing/model': 1.1386624174192548, 'learning_rate/qwen_vl_interface': 9.711069196652886e-06, 'learning_rate/action_model': 9.711069196652886e-05, 'epoch': 0.33}) +[2026-06-17 18:03:21,901][starVLA.training.train_starvla][INFO] - Step 871, Loss: {'train/loss': 0.6198546290397644, 'train/grad_norm_pre_clip': 5.6235222816467285, 'timing/data': 0.03185935132205486, 'timing/model': 1.2186275161802769, 'learning_rate/qwen_vl_interface': 9.710325967069992e-06, 'learning_rate/action_model': 9.710325967069991e-05, 'epoch': 0.34}) +[2026-06-17 18:03:27,279][starVLA.training.train_starvla][INFO] - Step 872, Loss: {'train/loss': 0.7858980298042297, 'train/grad_norm_pre_clip': 7.3237690925598145, 'timing/data': 0.022291500121355057, 'timing/model': 1.5049727335572243, 'learning_rate/qwen_vl_interface': 9.709581812856725e-06, 'learning_rate/action_model': 9.709581812856726e-05, 'epoch': 0.34}) +[2026-06-17 18:03:31,652][starVLA.training.train_starvla][INFO] - Step 873, Loss: {'train/loss': 0.7166945934295654, 'train/grad_norm_pre_clip': 5.154289245605469, 'timing/data': 0.02705542277544737, 'timing/model': 1.1838096352294087, 'learning_rate/qwen_vl_interface': 9.708836734167351e-06, 'learning_rate/action_model': 9.708836734167351e-05, 'epoch': 0.34}) +[2026-06-17 18:03:36,430][starVLA.training.train_starvla][INFO] - Step 874, Loss: {'train/loss': 0.8735966086387634, 'train/grad_norm_pre_clip': 8.473376274108887, 'timing/data': 0.021995147690176964, 'timing/model': 1.0832021636888385, 'learning_rate/qwen_vl_interface': 9.708090731156321e-06, 'learning_rate/action_model': 9.708090731156321e-05, 'epoch': 0.34}) +[2026-06-17 18:03:41,198][starVLA.training.train_starvla][INFO] - Step 875, Loss: {'train/loss': 0.6608964204788208, 'train/grad_norm_pre_clip': 5.16496467590332, 'timing/data': 0.02178053278476, 'timing/model': 1.1591564128175378, 'learning_rate/qwen_vl_interface': 9.707343803978288e-06, 'learning_rate/action_model': 9.707343803978288e-05, 'epoch': 0.34}) +[2026-06-17 18:03:46,939][starVLA.training.train_starvla][INFO] - Step 876, Loss: {'train/loss': 0.7912740111351013, 'train/grad_norm_pre_clip': 6.734005928039551, 'timing/data': 0.027192603796720505, 'timing/model': 1.1373617788776755, 'learning_rate/qwen_vl_interface': 9.706595952788088e-06, 'learning_rate/action_model': 9.706595952788088e-05, 'epoch': 0.34}) +[2026-06-17 18:03:53,648][starVLA.training.train_starvla][INFO] - Step 877, Loss: {'train/loss': 0.7206787467002869, 'train/grad_norm_pre_clip': 7.093291759490967, 'timing/data': 0.023188459686934948, 'timing/model': 2.1516711860895157, 'learning_rate/qwen_vl_interface': 9.705847177740751e-06, 'learning_rate/action_model': 9.705847177740751e-05, 'epoch': 0.34}) +[2026-06-17 18:03:58,142][starVLA.training.train_starvla][INFO] - Step 878, Loss: {'train/loss': 0.8633698225021362, 'train/grad_norm_pre_clip': 7.369277477264404, 'timing/data': 0.02514104824513197, 'timing/model': 1.102360642515123, 'learning_rate/qwen_vl_interface': 9.7050974789915e-06, 'learning_rate/action_model': 9.7050974789915e-05, 'epoch': 0.34}) +[2026-06-17 18:04:02,931][starVLA.training.train_starvla][INFO] - Step 879, Loss: {'train/loss': 1.0359225273132324, 'train/grad_norm_pre_clip': 9.263317108154297, 'timing/data': 0.030561555176973343, 'timing/model': 1.1155376015231013, 'learning_rate/qwen_vl_interface': 9.704346856695747e-06, 'learning_rate/action_model': 9.704346856695746e-05, 'epoch': 0.34}) +[2026-06-17 18:04:07,961][starVLA.training.train_starvla][INFO] - Step 880, Loss: {'train/loss': 0.740653932094574, 'train/grad_norm_pre_clip': 6.194695472717285, 'timing/data': 0.022121391259133816, 'timing/model': 1.0838249223306775, 'learning_rate/qwen_vl_interface': 9.703595311009098e-06, 'learning_rate/action_model': 9.703595311009097e-05, 'epoch': 0.34}) +[2026-06-17 18:04:12,670][starVLA.training.train_starvla][INFO] - Step 881, Loss: {'train/loss': 0.5605044960975647, 'train/grad_norm_pre_clip': 5.745354175567627, 'timing/data': 0.03574364818632603, 'timing/model': 1.1055793054401875, 'learning_rate/qwen_vl_interface': 9.702842842087348e-06, 'learning_rate/action_model': 9.702842842087348e-05, 'epoch': 0.34}) +[2026-06-17 18:04:17,338][starVLA.training.train_starvla][INFO] - Step 882, Loss: {'train/loss': 0.771239697933197, 'train/grad_norm_pre_clip': 8.653360366821289, 'timing/data': 0.022562552243471146, 'timing/model': 1.455156766809523, 'learning_rate/qwen_vl_interface': 9.702089450086486e-06, 'learning_rate/action_model': 9.702089450086486e-05, 'epoch': 0.34}) +[2026-06-17 18:04:21,798][starVLA.training.train_starvla][INFO] - Step 883, Loss: {'train/loss': 0.4185011386871338, 'train/grad_norm_pre_clip': 6.09004545211792, 'timing/data': 0.022095889784395695, 'timing/model': 1.172712005674839, 'learning_rate/qwen_vl_interface': 9.70133513516269e-06, 'learning_rate/action_model': 9.701335135162688e-05, 'epoch': 0.34}) +[2026-06-17 18:04:26,732][starVLA.training.train_starvla][INFO] - Step 884, Loss: {'train/loss': 0.4271373152732849, 'train/grad_norm_pre_clip': 5.766938209533691, 'timing/data': 0.020243949256837368, 'timing/model': 1.0747716445475817, 'learning_rate/qwen_vl_interface': 9.70057989747233e-06, 'learning_rate/action_model': 9.700579897472329e-05, 'epoch': 0.34}) +[2026-06-17 18:04:31,601][starVLA.training.train_starvla][INFO] - Step 885, Loss: {'train/loss': 0.7653040885925293, 'train/grad_norm_pre_clip': 6.359534740447998, 'timing/data': 0.03057364746928215, 'timing/model': 1.146145193837583, 'learning_rate/qwen_vl_interface': 9.699823737171969e-06, 'learning_rate/action_model': 9.699823737171969e-05, 'epoch': 0.34}) +[2026-06-17 18:04:36,491][starVLA.training.train_starvla][INFO] - Step 886, Loss: {'train/loss': 0.5778555870056152, 'train/grad_norm_pre_clip': 7.029323577880859, 'timing/data': 0.04462319519370794, 'timing/model': 1.1823989637196064, 'learning_rate/qwen_vl_interface': 9.699066654418358e-06, 'learning_rate/action_model': 9.699066654418358e-05, 'epoch': 0.34}) +[2026-06-17 18:04:41,466][starVLA.training.train_starvla][INFO] - Step 887, Loss: {'train/loss': 0.5608863830566406, 'train/grad_norm_pre_clip': 6.208065032958984, 'timing/data': 0.021965418942272663, 'timing/model': 1.5995433209463954, 'learning_rate/qwen_vl_interface': 9.698308649368444e-06, 'learning_rate/action_model': 9.698308649368444e-05, 'epoch': 0.34}) +[2026-06-17 18:04:46,942][starVLA.training.train_starvla][INFO] - Step 888, Loss: {'train/loss': 0.7530273199081421, 'train/grad_norm_pre_clip': 10.27103042602539, 'timing/data': 0.023237308487296104, 'timing/model': 2.2689284840598702, 'learning_rate/qwen_vl_interface': 9.69754972217936e-06, 'learning_rate/action_model': 9.69754972217936e-05, 'epoch': 0.34}) +[2026-06-17 18:04:51,684][starVLA.training.train_starvla][INFO] - Step 889, Loss: {'train/loss': 1.0567679405212402, 'train/grad_norm_pre_clip': 9.941481590270996, 'timing/data': 0.024768458679318428, 'timing/model': 1.0889360383152962, 'learning_rate/qwen_vl_interface': 9.696789873008432e-06, 'learning_rate/action_model': 9.696789873008433e-05, 'epoch': 0.34}) +[2026-06-17 18:04:56,305][starVLA.training.train_starvla][INFO] - Step 890, Loss: {'train/loss': 0.653289794921875, 'train/grad_norm_pre_clip': 7.728049278259277, 'timing/data': 0.03519985917955637, 'timing/model': 1.1027432158589363, 'learning_rate/qwen_vl_interface': 9.69602910201318e-06, 'learning_rate/action_model': 9.696029102013179e-05, 'epoch': 0.34}) +[2026-06-17 18:05:01,115][starVLA.training.train_starvla][INFO] - Step 891, Loss: {'train/loss': 0.604166567325592, 'train/grad_norm_pre_clip': 8.291202545166016, 'timing/data': 0.024585030041635036, 'timing/model': 1.0970696182921529, 'learning_rate/qwen_vl_interface': 9.695267409351308e-06, 'learning_rate/action_model': 9.695267409351307e-05, 'epoch': 0.34}) +[2026-06-17 18:05:06,017][starVLA.training.train_starvla][INFO] - Step 892, Loss: {'train/loss': 0.5830290913581848, 'train/grad_norm_pre_clip': 6.434820652008057, 'timing/data': 0.024386268109083176, 'timing/model': 1.6669527608901262, 'learning_rate/qwen_vl_interface': 9.694504795180722e-06, 'learning_rate/action_model': 9.694504795180721e-05, 'epoch': 0.34}) +[2026-06-17 18:05:11,347][starVLA.training.train_starvla][INFO] - Step 893, Loss: {'train/loss': 0.3281744718551636, 'train/grad_norm_pre_clip': 4.2982964515686035, 'timing/data': 0.024442806839942932, 'timing/model': 2.1260239286348224, 'learning_rate/qwen_vl_interface': 9.693741259659507e-06, 'learning_rate/action_model': 9.693741259659507e-05, 'epoch': 0.34}) +[2026-06-17 18:05:16,444][starVLA.training.train_starvla][INFO] - Step 894, Loss: {'train/loss': 0.8703708052635193, 'train/grad_norm_pre_clip': 9.289084434509277, 'timing/data': 0.027489916421473026, 'timing/model': 1.081941788084805, 'learning_rate/qwen_vl_interface': 9.692976802945947e-06, 'learning_rate/action_model': 9.692976802945948e-05, 'epoch': 0.34}) +[2026-06-17 18:05:28,401][starVLA.training.train_starvla][INFO] - Step 895, Loss: {'train/loss': 0.6517179608345032, 'train/grad_norm_pre_clip': 7.419715404510498, 'timing/data': 0.02366001531481743, 'timing/model': 5.349651541560888, 'learning_rate/qwen_vl_interface': 9.692211425198517e-06, 'learning_rate/action_model': 9.692211425198517e-05, 'epoch': 0.34}) +[2026-06-17 18:05:36,844][starVLA.training.train_starvla][INFO] - Step 896, Loss: {'train/loss': 0.6556012034416199, 'train/grad_norm_pre_clip': 6.775177955627441, 'timing/data': 0.02756260149180889, 'timing/model': 4.637874905951321, 'learning_rate/qwen_vl_interface': 9.691445126575879e-06, 'learning_rate/action_model': 9.691445126575877e-05, 'epoch': 0.34}) +[2026-06-17 18:05:41,947][starVLA.training.train_starvla][INFO] - Step 897, Loss: {'train/loss': 0.5481480956077576, 'train/grad_norm_pre_clip': 7.148726940155029, 'timing/data': 0.029447216540575027, 'timing/model': 1.753715816885233, 'learning_rate/qwen_vl_interface': 9.690677907236884e-06, 'learning_rate/action_model': 9.690677907236884e-05, 'epoch': 0.35}) +[2026-06-17 18:05:46,515][starVLA.training.train_starvla][INFO] - Step 898, Loss: {'train/loss': 0.7174297571182251, 'train/grad_norm_pre_clip': 8.083273887634277, 'timing/data': 0.025806167162954807, 'timing/model': 1.1018829494714737, 'learning_rate/qwen_vl_interface': 9.689909767340582e-06, 'learning_rate/action_model': 9.689909767340583e-05, 'epoch': 0.35}) +[2026-06-17 18:05:51,725][starVLA.training.train_starvla][INFO] - Step 899, Loss: {'train/loss': 0.431939959526062, 'train/grad_norm_pre_clip': 4.985462665557861, 'timing/data': 0.0013382770121097565, 'timing/model': 1.1226393934339285, 'learning_rate/qwen_vl_interface': 9.689140707046209e-06, 'learning_rate/action_model': 9.689140707046209e-05, 'epoch': 0.35}) +[2026-06-17 18:08:47,407][starVLA.training.train_starvla][INFO] - Step 900, Loss: {'train/loss': 0.6999604105949402, 'train/grad_norm_pre_clip': 7.7095255851745605, 'eval/loss': 0.682277500629425, 'eval/latency_4/loss': 0.682277500629425, 'eval/demon_fire_precision': 0.9672346002621232, 'eval/demon_fire_recall': 0.9587528418317636, 'eval/demon_fire_f1': 0.9629750448540205, 'eval/demon_fire_rate_model': 0.8963289280469897, 'eval/demon_fire_rate_teacher': 0.9042584434654919, 'eval/demon_left_precision': 0.5925155925155925, 'eval/demon_left_recall': 0.6506849315068494, 'eval/demon_left_f1': 0.6202393906420022, 'eval/demon_left_rate_model': 0.14126284875183553, 'eval/demon_left_rate_teacher': 0.12863436123348018, 'eval/demon_right_precision': 0.7236580516898609, 'eval/demon_right_recall': 0.6570397111913358, 'eval/demon_right_f1': 0.6887417218543047, 'eval/demon_right_rate_model': 0.14772393538913361, 'eval/demon_right_rate_teacher': 0.16270190895741557, 'eval/demon_move_f1': 0.6544905562481534, 'eval/demon_cc_f1': 0.8436784234493435, 'eval/latency_4/demon_fire_precision': 0.9672346002621232, 'eval/latency_4/demon_fire_recall': 0.9587528418317636, 'eval/latency_4/demon_fire_f1': 0.9629750448540205, 'eval/latency_4/demon_fire_rate_model': 0.8963289280469897, 'eval/latency_4/demon_fire_rate_teacher': 0.9042584434654919, 'eval/latency_4/demon_left_precision': 0.5925155925155925, 'eval/latency_4/demon_left_recall': 0.6506849315068494, 'eval/latency_4/demon_left_f1': 0.6202393906420022, 'eval/latency_4/demon_left_rate_model': 0.14126284875183553, 'eval/latency_4/demon_left_rate_teacher': 0.12863436123348018, 'eval/latency_4/demon_right_precision': 0.7236580516898609, 'eval/latency_4/demon_right_recall': 0.6570397111913358, 'eval/latency_4/demon_right_f1': 0.6887417218543047, 'eval/latency_4/demon_right_rate_model': 0.14772393538913361, 'eval/latency_4/demon_right_rate_teacher': 0.16270190895741557, 'eval/latency_4/demon_move_f1': 0.6544905562481534, 'eval/latency_4/demon_cc_f1': 0.8436784234493435, 'timing/data': 0.0012323753908276558, 'timing/model': 1.06845320481807, 'learning_rate/qwen_vl_interface': 9.68837072651319e-06, 'learning_rate/action_model': 9.688370726513188e-05, 'epoch': 0.35}) +[2026-06-17 18:08:47,409][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_900_state +[2026-06-17 18:08:47,420][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 18:08:56,614][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_900_state/model.safetensors +[2026-06-17 18:09:18,171][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_900_state/optimizer.bin +[2026-06-17 18:09:18,186][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_900_state/sampler.bin +[2026-06-17 18:09:18,187][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_900_state/sampler_1.bin +[2026-06-17 18:09:18,190][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_900_state/random_states_0.pkl +[2026-06-17 18:09:18,196][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 900 +[2026-06-17 18:09:18,197][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 18:09:18,219][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 18:09:22,938][starVLA.training.train_starvla][INFO] - Step 901, Loss: {'train/loss': 0.6624740958213806, 'train/grad_norm_pre_clip': 8.156001091003418, 'timing/data': 0.0013658888638019562, 'timing/model': 1.0888224337249994, 'learning_rate/qwen_vl_interface': 9.687599825901141e-06, 'learning_rate/action_model': 9.687599825901142e-05, 'epoch': 0.35}) +[2026-06-17 18:09:27,565][starVLA.training.train_starvla][INFO] - Step 902, Loss: {'train/loss': 0.5302305817604065, 'train/grad_norm_pre_clip': 6.732598304748535, 'timing/data': 0.0011756522580981255, 'timing/model': 1.3986158426851034, 'learning_rate/qwen_vl_interface': 9.686828005369876e-06, 'learning_rate/action_model': 9.686828005369875e-05, 'epoch': 0.35}) +[2026-06-17 18:09:33,173][starVLA.training.train_starvla][INFO] - Step 903, Loss: {'train/loss': 1.002421259880066, 'train/grad_norm_pre_clip': 6.464686393737793, 'timing/data': 0.0011513475328683853, 'timing/model': 1.2641611164435744, 'learning_rate/qwen_vl_interface': 9.68605526507939e-06, 'learning_rate/action_model': 9.686055265079389e-05, 'epoch': 0.35}) +[2026-06-17 18:09:37,811][starVLA.training.train_starvla][INFO] - Step 904, Loss: {'train/loss': 0.6756957769393921, 'train/grad_norm_pre_clip': 8.547125816345215, 'timing/data': 0.001429559662938118, 'timing/model': 1.0959202852100134, 'learning_rate/qwen_vl_interface': 9.685281605189874e-06, 'learning_rate/action_model': 9.685281605189874e-05, 'epoch': 0.35}) +[2026-06-17 18:09:42,616][starVLA.training.train_starvla][INFO] - Step 905, Loss: {'train/loss': 0.5544089078903198, 'train/grad_norm_pre_clip': 6.3325581550598145, 'timing/data': 0.0012783482670783997, 'timing/model': 1.2111116107553244, 'learning_rate/qwen_vl_interface': 9.68450702586171e-06, 'learning_rate/action_model': 9.684507025861709e-05, 'epoch': 0.35}) +[2026-06-17 18:09:48,080][starVLA.training.train_starvla][INFO] - Step 906, Loss: {'train/loss': 0.855059802532196, 'train/grad_norm_pre_clip': 9.011006355285645, 'timing/data': 0.0013044504448771477, 'timing/model': 1.3998879306018353, 'learning_rate/qwen_vl_interface': 9.683731527255467e-06, 'learning_rate/action_model': 9.683731527255467e-05, 'epoch': 0.35}) +[2026-06-17 18:09:53,503][starVLA.training.train_starvla][INFO] - Step 907, Loss: {'train/loss': 0.5466113686561584, 'train/grad_norm_pre_clip': 8.476513862609863, 'timing/data': 0.001013440079987049, 'timing/model': 1.6261120364069939, 'learning_rate/qwen_vl_interface': 9.682955109531906e-06, 'learning_rate/action_model': 9.682955109531905e-05, 'epoch': 0.35}) +[2026-06-17 18:09:57,933][starVLA.training.train_starvla][INFO] - Step 908, Loss: {'train/loss': 0.5786304473876953, 'train/grad_norm_pre_clip': 6.226262092590332, 'timing/data': 0.0011220593005418777, 'timing/model': 1.114491323940456, 'learning_rate/qwen_vl_interface': 9.682177772851983e-06, 'learning_rate/action_model': 9.682177772851982e-05, 'epoch': 0.35}) +[2026-06-17 18:10:02,914][starVLA.training.train_starvla][INFO] - Step 909, Loss: {'train/loss': 0.7035274505615234, 'train/grad_norm_pre_clip': 6.502653121948242, 'timing/data': 0.0011751484125852585, 'timing/model': 1.3229055246338248, 'learning_rate/qwen_vl_interface': 9.681399517376837e-06, 'learning_rate/action_model': 9.681399517376837e-05, 'epoch': 0.35}) +[2026-06-17 18:10:07,669][starVLA.training.train_starvla][INFO] - Step 910, Loss: {'train/loss': 0.9062986969947815, 'train/grad_norm_pre_clip': 8.09768009185791, 'timing/data': 0.0011970466002821922, 'timing/model': 1.0912966784089804, 'learning_rate/qwen_vl_interface': 9.680620343267803e-06, 'learning_rate/action_model': 9.680620343267802e-05, 'epoch': 0.35}) +[2026-06-17 18:10:13,645][starVLA.training.train_starvla][INFO] - Step 911, Loss: {'train/loss': 0.7096648216247559, 'train/grad_norm_pre_clip': 6.468626976013184, 'timing/data': 0.0012051304802298546, 'timing/model': 1.3338274070993066, 'learning_rate/qwen_vl_interface': 9.679840250686403e-06, 'learning_rate/action_model': 9.679840250686404e-05, 'epoch': 0.35}) +[2026-06-17 18:10:18,029][starVLA.training.train_starvla][INFO] - Step 912, Loss: {'train/loss': 0.7047273516654968, 'train/grad_norm_pre_clip': 9.562883377075195, 'timing/data': 0.0018129069358110428, 'timing/model': 1.1586087606847286, 'learning_rate/qwen_vl_interface': 9.679059239794355e-06, 'learning_rate/action_model': 9.679059239794355e-05, 'epoch': 0.35}) +[2026-06-17 18:10:27,577][starVLA.training.train_starvla][INFO] - Step 913, Loss: {'train/loss': 0.736219048500061, 'train/grad_norm_pre_clip': 5.262621879577637, 'timing/data': 0.0011363271623849869, 'timing/model': 2.5329901706427336, 'learning_rate/qwen_vl_interface': 9.67827731075356e-06, 'learning_rate/action_model': 9.67827731075356e-05, 'epoch': 0.35}) +[2026-06-17 18:10:33,287][starVLA.training.train_starvla][INFO] - Step 914, Loss: {'train/loss': 0.9015666842460632, 'train/grad_norm_pre_clip': 11.5214204788208, 'timing/data': 0.001229877583682537, 'timing/model': 1.1662193285301328, 'learning_rate/qwen_vl_interface': 9.677494463726112e-06, 'learning_rate/action_model': 9.677494463726112e-05, 'epoch': 0.35}) +[2026-06-17 18:10:39,427][starVLA.training.train_starvla][INFO] - Step 915, Loss: {'train/loss': 0.540367603302002, 'train/grad_norm_pre_clip': 6.190460205078125, 'timing/data': 0.0011391527950763702, 'timing/model': 1.3665766110643744, 'learning_rate/qwen_vl_interface': 9.676710698874299e-06, 'learning_rate/action_model': 9.676710698874298e-05, 'epoch': 0.35}) +[2026-06-17 18:10:44,342][starVLA.training.train_starvla][INFO] - Step 916, Loss: {'train/loss': 0.6140714287757874, 'train/grad_norm_pre_clip': 8.384736061096191, 'timing/data': 0.0010861195623874664, 'timing/model': 1.4369232496246696, 'learning_rate/qwen_vl_interface': 9.675926016360595e-06, 'learning_rate/action_model': 9.675926016360595e-05, 'epoch': 0.35}) +[2026-06-17 18:10:49,517][starVLA.training.train_starvla][INFO] - Step 917, Loss: {'train/loss': 0.5947328805923462, 'train/grad_norm_pre_clip': 7.599573135375977, 'timing/data': 0.0011170478537678719, 'timing/model': 1.9316304745152593, 'learning_rate/qwen_vl_interface': 9.675140416347664e-06, 'learning_rate/action_model': 9.675140416347663e-05, 'epoch': 0.35}) +[2026-06-17 18:10:54,992][starVLA.training.train_starvla][INFO] - Step 918, Loss: {'train/loss': 0.7731103301048279, 'train/grad_norm_pre_clip': 5.321933746337891, 'timing/data': 0.0011296505108475685, 'timing/model': 1.2034898260608315, 'learning_rate/qwen_vl_interface': 9.674353898998362e-06, 'learning_rate/action_model': 9.674353898998363e-05, 'epoch': 0.35}) +[2026-06-17 18:11:04,850][starVLA.training.train_starvla][INFO] - Step 919, Loss: {'train/loss': 0.4119444787502289, 'train/grad_norm_pre_clip': 5.205009460449219, 'timing/data': 0.0011416338384151459, 'timing/model': 1.1276506809517741, 'learning_rate/qwen_vl_interface': 9.673566464475738e-06, 'learning_rate/action_model': 9.673566464475739e-05, 'epoch': 0.35}) +[2026-06-17 18:11:13,031][starVLA.training.train_starvla][INFO] - Step 920, Loss: {'train/loss': 0.4426038861274719, 'train/grad_norm_pre_clip': 8.134276390075684, 'timing/data': 0.0016352199018001556, 'timing/model': 1.3504639957100153, 'learning_rate/qwen_vl_interface': 9.672778112943026e-06, 'learning_rate/action_model': 9.672778112943026e-05, 'epoch': 0.35}) +[2026-06-17 18:11:22,321][starVLA.training.train_starvla][INFO] - Step 921, Loss: {'train/loss': 0.5119479894638062, 'train/grad_norm_pre_clip': 6.007041931152344, 'timing/data': 0.0009513981640338898, 'timing/model': 5.531628851778805, 'learning_rate/qwen_vl_interface': 9.671988844563652e-06, 'learning_rate/action_model': 9.671988844563652e-05, 'epoch': 0.35}) +[2026-06-17 18:11:27,090][starVLA.training.train_starvla][INFO] - Step 922, Loss: {'train/loss': 0.7429613471031189, 'train/grad_norm_pre_clip': 11.007092475891113, 'timing/data': 0.0011280588805675507, 'timing/model': 1.6208303440362215, 'learning_rate/qwen_vl_interface': 9.671198659501233e-06, 'learning_rate/action_model': 9.671198659501231e-05, 'epoch': 0.35}) +[2026-06-17 18:11:33,832][starVLA.training.train_starvla][INFO] - Step 923, Loss: {'train/loss': 0.8255192637443542, 'train/grad_norm_pre_clip': 8.34237003326416, 'timing/data': 0.0011007050052285194, 'timing/model': 1.0943141868337989, 'learning_rate/qwen_vl_interface': 9.670407557919573e-06, 'learning_rate/action_model': 9.670407557919572e-05, 'epoch': 0.36}) +[2026-06-17 18:11:38,545][starVLA.training.train_starvla][INFO] - Step 924, Loss: {'train/loss': 0.4459148943424225, 'train/grad_norm_pre_clip': 7.532253265380859, 'timing/data': 0.001216706819832325, 'timing/model': 1.135351274162531, 'learning_rate/qwen_vl_interface': 9.669615539982668e-06, 'learning_rate/action_model': 9.669615539982668e-05, 'epoch': 0.36}) +[2026-06-17 18:11:43,166][starVLA.training.train_starvla][INFO] - Step 925, Loss: {'train/loss': 0.7472286820411682, 'train/grad_norm_pre_clip': 7.631350994110107, 'timing/data': 0.0013477951288223267, 'timing/model': 1.0746115483343601, 'learning_rate/qwen_vl_interface': 9.668822605854708e-06, 'learning_rate/action_model': 9.668822605854706e-05, 'epoch': 0.36}) +[2026-06-17 18:11:47,706][starVLA.training.train_starvla][INFO] - Step 926, Loss: {'train/loss': 0.8137065172195435, 'train/grad_norm_pre_clip': 10.590506553649902, 'timing/data': 0.0012749675661325455, 'timing/model': 1.079846566542983, 'learning_rate/qwen_vl_interface': 9.668028755700067e-06, 'learning_rate/action_model': 9.668028755700067e-05, 'epoch': 0.36}) +[2026-06-17 18:11:52,215][starVLA.training.train_starvla][INFO] - Step 927, Loss: {'train/loss': 0.6459701657295227, 'train/grad_norm_pre_clip': 8.026874542236328, 'timing/data': 0.001123594120144844, 'timing/model': 1.3823269046843052, 'learning_rate/qwen_vl_interface': 9.66723398968331e-06, 'learning_rate/action_model': 9.66723398968331e-05, 'epoch': 0.36}) +[2026-06-17 18:11:56,462][starVLA.training.train_starvla][INFO] - Step 928, Loss: {'train/loss': 0.5686743259429932, 'train/grad_norm_pre_clip': 8.45145320892334, 'timing/data': 0.0010665450245141983, 'timing/model': 1.0967577258124948, 'learning_rate/qwen_vl_interface': 9.666438307969194e-06, 'learning_rate/action_model': 9.666438307969195e-05, 'epoch': 0.36}) +[2026-06-17 18:12:01,154][starVLA.training.train_starvla][INFO] - Step 929, Loss: {'train/loss': 0.8788666725158691, 'train/grad_norm_pre_clip': 10.443453788757324, 'timing/data': 0.000935843214392662, 'timing/model': 1.1690243808552623, 'learning_rate/qwen_vl_interface': 9.665641710722666e-06, 'learning_rate/action_model': 9.665641710722665e-05, 'epoch': 0.36}) +[2026-06-17 18:12:05,602][starVLA.training.train_starvla][INFO] - Step 930, Loss: {'train/loss': 0.8323975801467896, 'train/grad_norm_pre_clip': 5.962311267852783, 'timing/data': 0.001483728177845478, 'timing/model': 1.0749780731275678, 'learning_rate/qwen_vl_interface': 9.664844198108858e-06, 'learning_rate/action_model': 9.664844198108858e-05, 'epoch': 0.36}) +[2026-06-17 18:12:10,511][starVLA.training.train_starvla][INFO] - Step 931, Loss: {'train/loss': 0.6671561598777771, 'train/grad_norm_pre_clip': 8.818560600280762, 'timing/data': 0.000894937664270401, 'timing/model': 1.207035981118679, 'learning_rate/qwen_vl_interface': 9.6640457702931e-06, 'learning_rate/action_model': 9.664045770293099e-05, 'epoch': 0.36}) +[2026-06-17 18:12:15,138][starVLA.training.train_starvla][INFO] - Step 932, Loss: {'train/loss': 0.40861350297927856, 'train/grad_norm_pre_clip': 7.479591369628906, 'timing/data': 0.0011377036571502686, 'timing/model': 1.1218394879251719, 'learning_rate/qwen_vl_interface': 9.663246427440901e-06, 'learning_rate/action_model': 9.6632464274409e-05, 'epoch': 0.36}) +[2026-06-17 18:12:19,849][starVLA.training.train_starvla][INFO] - Step 933, Loss: {'train/loss': 0.9894775748252869, 'train/grad_norm_pre_clip': 10.922837257385254, 'timing/data': 0.0011326223611831665, 'timing/model': 1.1438141344115138, 'learning_rate/qwen_vl_interface': 9.662446169717969e-06, 'learning_rate/action_model': 9.662446169717969e-05, 'epoch': 0.36}) +[2026-06-17 18:12:24,757][starVLA.training.train_starvla][INFO] - Step 934, Loss: {'train/loss': 0.6825920343399048, 'train/grad_norm_pre_clip': 7.227729797363281, 'timing/data': 0.0014639860019087791, 'timing/model': 1.2765022981911898, 'learning_rate/qwen_vl_interface': 9.6616449972902e-06, 'learning_rate/action_model': 9.6616449972902e-05, 'epoch': 0.36}) +[2026-06-17 18:12:29,544][starVLA.training.train_starvla][INFO] - Step 935, Loss: {'train/loss': 0.6341066956520081, 'train/grad_norm_pre_clip': 7.537021636962891, 'timing/data': 0.001297871582210064, 'timing/model': 1.251807271502912, 'learning_rate/qwen_vl_interface': 9.660842910323675e-06, 'learning_rate/action_model': 9.660842910323674e-05, 'epoch': 0.36}) +[2026-06-17 18:12:34,826][starVLA.training.train_starvla][INFO] - Step 936, Loss: {'train/loss': 0.6551423668861389, 'train/grad_norm_pre_clip': 9.549982070922852, 'timing/data': 0.0007979217916727066, 'timing/model': 1.737382280640304, 'learning_rate/qwen_vl_interface': 9.660039908984667e-06, 'learning_rate/action_model': 9.660039908984666e-05, 'epoch': 0.36}) +[2026-06-17 18:12:39,200][starVLA.training.train_starvla][INFO] - Step 937, Loss: {'train/loss': 0.5401288270950317, 'train/grad_norm_pre_clip': 5.3181939125061035, 'timing/data': 0.0009602699428796768, 'timing/model': 1.1685580620542169, 'learning_rate/qwen_vl_interface': 9.659235993439643e-06, 'learning_rate/action_model': 9.659235993439643e-05, 'epoch': 0.36}) +[2026-06-17 18:12:45,310][starVLA.training.train_starvla][INFO] - Step 938, Loss: {'train/loss': 0.8281605839729309, 'train/grad_norm_pre_clip': 8.459945678710938, 'timing/data': 0.0011510038748383522, 'timing/model': 1.0925810923799872, 'learning_rate/qwen_vl_interface': 9.658431163855251e-06, 'learning_rate/action_model': 9.658431163855251e-05, 'epoch': 0.36}) +[2026-06-17 18:13:01,128][starVLA.training.train_starvla][INFO] - Step 939, Loss: {'train/loss': 0.6953288912773132, 'train/grad_norm_pre_clip': 8.339146614074707, 'timing/data': 0.0010789372026920319, 'timing/model': 2.8790086172521114, 'learning_rate/qwen_vl_interface': 9.657625420398334e-06, 'learning_rate/action_model': 9.657625420398333e-05, 'epoch': 0.36}) +[2026-06-17 18:13:09,747][starVLA.training.train_starvla][INFO] - Step 940, Loss: {'train/loss': 0.947664737701416, 'train/grad_norm_pre_clip': 8.309453964233398, 'timing/data': 0.0018461523577570915, 'timing/model': 1.5928304763510823, 'learning_rate/qwen_vl_interface': 9.656818763235927e-06, 'learning_rate/action_model': 9.656818763235926e-05, 'epoch': 0.36}) +[2026-06-17 18:13:15,171][starVLA.training.train_starvla][INFO] - Step 941, Loss: {'train/loss': 0.8589718341827393, 'train/grad_norm_pre_clip': 14.471360206604004, 'timing/data': 0.0011949026957154274, 'timing/model': 1.2125847581773996, 'learning_rate/qwen_vl_interface': 9.656011192535248e-06, 'learning_rate/action_model': 9.656011192535247e-05, 'epoch': 0.36}) +[2026-06-17 18:13:23,110][starVLA.training.train_starvla][INFO] - Step 942, Loss: {'train/loss': 1.0898460149765015, 'train/grad_norm_pre_clip': 15.567204475402832, 'timing/data': 0.001139441505074501, 'timing/model': 2.689656462520361, 'learning_rate/qwen_vl_interface': 9.655202708463706e-06, 'learning_rate/action_model': 9.655202708463707e-05, 'epoch': 0.36}) +[2026-06-17 18:13:29,408][starVLA.training.train_starvla][INFO] - Step 943, Loss: {'train/loss': 0.734676718711853, 'train/grad_norm_pre_clip': 5.396584987640381, 'timing/data': 0.00128129031509161, 'timing/model': 1.1903821481391788, 'learning_rate/qwen_vl_interface': 9.654393311188905e-06, 'learning_rate/action_model': 9.654393311188905e-05, 'epoch': 0.36}) +[2026-06-17 18:13:38,242][starVLA.training.train_starvla][INFO] - Step 944, Loss: {'train/loss': 0.5506030917167664, 'train/grad_norm_pre_clip': 5.592942237854004, 'timing/data': 0.0010978495702147484, 'timing/model': 1.1904523726552725, 'learning_rate/qwen_vl_interface': 9.653583000878631e-06, 'learning_rate/action_model': 9.65358300087863e-05, 'epoch': 0.36}) +[2026-06-17 18:13:51,745][starVLA.training.train_starvla][INFO] - Step 945, Loss: {'train/loss': 0.6964737176895142, 'train/grad_norm_pre_clip': 7.270692348480225, 'timing/data': 0.0015713972970843315, 'timing/model': 2.7807806124910712, 'learning_rate/qwen_vl_interface': 9.65277177770086e-06, 'learning_rate/action_model': 9.65277177770086e-05, 'epoch': 0.36}) +[2026-06-17 18:13:59,556][starVLA.training.train_starvla][INFO] - Step 946, Loss: {'train/loss': 0.7119094729423523, 'train/grad_norm_pre_clip': 10.770479202270508, 'timing/data': 0.0012050187215209007, 'timing/model': 1.1152954353019595, 'learning_rate/qwen_vl_interface': 9.651959641823764e-06, 'learning_rate/action_model': 9.651959641823763e-05, 'epoch': 0.36}) +[2026-06-17 18:14:04,286][starVLA.training.train_starvla][INFO] - Step 947, Loss: {'train/loss': 0.6979174613952637, 'train/grad_norm_pre_clip': 8.672566413879395, 'timing/data': 0.0009638238698244095, 'timing/model': 1.5329924123361707, 'learning_rate/qwen_vl_interface': 9.651146593415696e-06, 'learning_rate/action_model': 9.651146593415695e-05, 'epoch': 0.36}) +[2026-06-17 18:14:08,692][starVLA.training.train_starvla][INFO] - Step 948, Loss: {'train/loss': 0.5199296474456787, 'train/grad_norm_pre_clip': 4.776651859283447, 'timing/data': 0.0011374754831194878, 'timing/model': 1.1232514511793852, 'learning_rate/qwen_vl_interface': 9.650332632645206e-06, 'learning_rate/action_model': 9.650332632645205e-05, 'epoch': 0.36}) +[2026-06-17 18:14:13,603][starVLA.training.train_starvla][INFO] - Step 949, Loss: {'train/loss': 0.6249595284461975, 'train/grad_norm_pre_clip': 7.331045627593994, 'timing/data': 0.0012025237083435059, 'timing/model': 1.1303505850955844, 'learning_rate/qwen_vl_interface': 9.649517759681022e-06, 'learning_rate/action_model': 9.649517759681023e-05, 'epoch': 0.37}) +[2026-06-17 18:14:18,621][starVLA.training.train_starvla][INFO] - Step 950, Loss: {'train/loss': 0.4794416129589081, 'train/grad_norm_pre_clip': 5.511746883392334, 'timing/data': 0.0011884458363056183, 'timing/model': 1.3251563841477036, 'learning_rate/qwen_vl_interface': 9.648701974692075e-06, 'learning_rate/action_model': 9.648701974692075e-05, 'epoch': 0.37}) +[2026-06-17 18:14:23,717][starVLA.training.train_starvla][INFO] - Step 951, Loss: {'train/loss': 0.4758119285106659, 'train/grad_norm_pre_clip': 9.546031951904297, 'timing/data': 0.001367814838886261, 'timing/model': 1.177107979543507, 'learning_rate/qwen_vl_interface': 9.647885277847476e-06, 'learning_rate/action_model': 9.647885277847476e-05, 'epoch': 0.37}) +[2026-06-17 18:14:30,866][starVLA.training.train_starvla][INFO] - Step 952, Loss: {'train/loss': 0.61294025182724, 'train/grad_norm_pre_clip': 8.007944107055664, 'timing/data': 0.0011597657576203346, 'timing/model': 1.587489471770823, 'learning_rate/qwen_vl_interface': 9.647067669316525e-06, 'learning_rate/action_model': 9.647067669316525e-05, 'epoch': 0.37}) +[2026-06-17 18:14:37,670][starVLA.training.train_starvla][INFO] - Step 953, Loss: {'train/loss': 0.8128729462623596, 'train/grad_norm_pre_clip': 7.10050106048584, 'timing/data': 0.0011828569695353508, 'timing/model': 2.8773674089461565, 'learning_rate/qwen_vl_interface': 9.646249149268716e-06, 'learning_rate/action_model': 9.646249149268715e-05, 'epoch': 0.37}) +[2026-06-17 18:14:42,848][starVLA.training.train_starvla][INFO] - Step 954, Loss: {'train/loss': 0.6162001490592957, 'train/grad_norm_pre_clip': 9.270885467529297, 'timing/data': 0.0011657290160655975, 'timing/model': 1.097429996356368, 'learning_rate/qwen_vl_interface': 9.645429717873728e-06, 'learning_rate/action_model': 9.645429717873727e-05, 'epoch': 0.37}) +[2026-06-17 18:14:47,808][starVLA.training.train_starvla][INFO] - Step 955, Loss: {'train/loss': 0.6151928901672363, 'train/grad_norm_pre_clip': 6.037512302398682, 'timing/data': 0.0023516034707427025, 'timing/model': 1.1742193847894669, 'learning_rate/qwen_vl_interface': 9.644609375301429e-06, 'learning_rate/action_model': 9.64460937530143e-05, 'epoch': 0.37}) +[2026-06-17 18:14:54,224][starVLA.training.train_starvla][INFO] - Step 956, Loss: {'train/loss': 0.8915524482727051, 'train/grad_norm_pre_clip': 37.24724578857422, 'timing/data': 0.0012821955606341362, 'timing/model': 1.122677793726325, 'learning_rate/qwen_vl_interface': 9.643788121721877e-06, 'learning_rate/action_model': 9.643788121721876e-05, 'epoch': 0.37}) +[2026-06-17 18:14:59,087][starVLA.training.train_starvla][INFO] - Step 957, Loss: {'train/loss': 0.9239149689674377, 'train/grad_norm_pre_clip': 11.264098167419434, 'timing/data': 0.0012132208794355392, 'timing/model': 1.4968268107622862, 'learning_rate/qwen_vl_interface': 9.64296595730532e-06, 'learning_rate/action_model': 9.64296595730532e-05, 'epoch': 0.37}) +[2026-06-17 18:15:03,550][starVLA.training.train_starvla][INFO] - Step 958, Loss: {'train/loss': 0.5350750088691711, 'train/grad_norm_pre_clip': 6.642348766326904, 'timing/data': 0.001355496235191822, 'timing/model': 1.1028986284509301, 'learning_rate/qwen_vl_interface': 9.642142882222193e-06, 'learning_rate/action_model': 9.642142882222193e-05, 'epoch': 0.37}) +[2026-06-17 18:15:08,852][starVLA.training.train_starvla][INFO] - Step 959, Loss: {'train/loss': 0.7297618985176086, 'train/grad_norm_pre_clip': 8.806053161621094, 'timing/data': 0.0026006773114204407, 'timing/model': 1.4466703003272414, 'learning_rate/qwen_vl_interface': 9.641318896643123e-06, 'learning_rate/action_model': 9.641318896643122e-05, 'epoch': 0.37}) +[2026-06-17 18:15:13,838][starVLA.training.train_starvla][INFO] - Step 960, Loss: {'train/loss': 0.5165160894393921, 'train/grad_norm_pre_clip': 5.9868292808532715, 'timing/data': 0.0013094665482640266, 'timing/model': 1.1124942554160953, 'learning_rate/qwen_vl_interface': 9.640494000738917e-06, 'learning_rate/action_model': 9.640494000738917e-05, 'epoch': 0.37}) +[2026-06-17 18:15:18,590][starVLA.training.train_starvla][INFO] - Step 961, Loss: {'train/loss': 0.7467456459999084, 'train/grad_norm_pre_clip': 9.80774211883545, 'timing/data': 0.001327785663306713, 'timing/model': 1.23271747212857, 'learning_rate/qwen_vl_interface': 9.63966819468058e-06, 'learning_rate/action_model': 9.63966819468058e-05, 'epoch': 0.37}) +[2026-06-17 18:15:32,130][starVLA.training.train_starvla][INFO] - Step 962, Loss: {'train/loss': 0.8643311262130737, 'train/grad_norm_pre_clip': 6.9069294929504395, 'timing/data': 0.0011570164933800697, 'timing/model': 5.943597794510424, 'learning_rate/qwen_vl_interface': 9.638841478639306e-06, 'learning_rate/action_model': 9.638841478639306e-05, 'epoch': 0.37}) +[2026-06-17 18:15:36,853][starVLA.training.train_starvla][INFO] - Step 963, Loss: {'train/loss': 0.9654062986373901, 'train/grad_norm_pre_clip': 9.81434440612793, 'timing/data': 0.0016889255493879318, 'timing/model': 1.2642363235354424, 'learning_rate/qwen_vl_interface': 9.638013852786468e-06, 'learning_rate/action_model': 9.638013852786467e-05, 'epoch': 0.37}) +[2026-06-17 18:15:43,520][starVLA.training.train_starvla][INFO] - Step 964, Loss: {'train/loss': 0.9257365465164185, 'train/grad_norm_pre_clip': 9.126019477844238, 'timing/data': 0.0016355831176042557, 'timing/model': 1.339435562491417, 'learning_rate/qwen_vl_interface': 9.637185317293638e-06, 'learning_rate/action_model': 9.637185317293638e-05, 'epoch': 0.37}) +[2026-06-17 18:15:48,234][starVLA.training.train_starvla][INFO] - Step 965, Loss: {'train/loss': 0.6664307713508606, 'train/grad_norm_pre_clip': 6.010780334472656, 'timing/data': 0.0010405508801341057, 'timing/model': 1.1542565608397126, 'learning_rate/qwen_vl_interface': 9.63635587233257e-06, 'learning_rate/action_model': 9.63635587233257e-05, 'epoch': 0.37}) +[2026-06-17 18:15:53,286][starVLA.training.train_starvla][INFO] - Step 966, Loss: {'train/loss': 0.37486720085144043, 'train/grad_norm_pre_clip': 4.892449378967285, 'timing/data': 0.00096889678388834, 'timing/model': 1.304063650779426, 'learning_rate/qwen_vl_interface': 9.63552551807521e-06, 'learning_rate/action_model': 9.63552551807521e-05, 'epoch': 0.37}) +[2026-06-17 18:16:04,477][starVLA.training.train_starvla][INFO] - Step 967, Loss: {'train/loss': 0.5671734809875488, 'train/grad_norm_pre_clip': 5.379199981689453, 'timing/data': 0.0010906998068094254, 'timing/model': 1.7538704965263605, 'learning_rate/qwen_vl_interface': 9.634694254693691e-06, 'learning_rate/action_model': 9.634694254693692e-05, 'epoch': 0.37}) +[2026-06-17 18:16:09,374][starVLA.training.train_starvla][INFO] - Step 968, Loss: {'train/loss': 0.7332016825675964, 'train/grad_norm_pre_clip': 6.5797929763793945, 'timing/data': 0.001134631223976612, 'timing/model': 1.194353544153273, 'learning_rate/qwen_vl_interface': 9.633862082360336e-06, 'learning_rate/action_model': 9.633862082360335e-05, 'epoch': 0.37}) +[2026-06-17 18:16:14,414][starVLA.training.train_starvla][INFO] - Step 969, Loss: {'train/loss': 0.8004233241081238, 'train/grad_norm_pre_clip': 8.609862327575684, 'timing/data': 0.0012841997668147087, 'timing/model': 1.1051834635436535, 'learning_rate/qwen_vl_interface': 9.633029001247654e-06, 'learning_rate/action_model': 9.633029001247653e-05, 'epoch': 0.37}) +[2026-06-17 18:16:19,808][starVLA.training.train_starvla][INFO] - Step 970, Loss: {'train/loss': 0.40769702196121216, 'train/grad_norm_pre_clip': 3.948174238204956, 'timing/data': 0.0011397302150726318, 'timing/model': 1.183883867226541, 'learning_rate/qwen_vl_interface': 9.63219501152834e-06, 'learning_rate/action_model': 9.632195011528341e-05, 'epoch': 0.37}) +[2026-06-17 18:16:24,813][starVLA.training.train_starvla][INFO] - Step 971, Loss: {'train/loss': 0.6965833306312561, 'train/grad_norm_pre_clip': 5.607763290405273, 'timing/data': 0.0011056102812290192, 'timing/model': 1.1595921162515879, 'learning_rate/qwen_vl_interface': 9.63136011337529e-06, 'learning_rate/action_model': 9.63136011337529e-05, 'epoch': 0.37}) +[2026-06-17 18:16:29,823][starVLA.training.train_starvla][INFO] - Step 972, Loss: {'train/loss': 0.5754087567329407, 'train/grad_norm_pre_clip': 7.049887180328369, 'timing/data': 0.001153486780822277, 'timing/model': 1.4380078101530671, 'learning_rate/qwen_vl_interface': 9.63052430696157e-06, 'learning_rate/action_model': 9.63052430696157e-05, 'epoch': 0.37}) +[2026-06-17 18:16:34,813][starVLA.training.train_starvla][INFO] - Step 973, Loss: {'train/loss': 0.8144211173057556, 'train/grad_norm_pre_clip': 8.970213890075684, 'timing/data': 0.0011740876361727715, 'timing/model': 1.1612043092027307, 'learning_rate/qwen_vl_interface': 9.629687592460449e-06, 'learning_rate/action_model': 9.62968759246045e-05, 'epoch': 0.37}) +[2026-06-17 18:16:43,568][starVLA.training.train_starvla][INFO] - Step 974, Loss: {'train/loss': 1.0655121803283691, 'train/grad_norm_pre_clip': 11.017139434814453, 'timing/data': 0.0012499969452619553, 'timing/model': 1.1119662523269653, 'learning_rate/qwen_vl_interface': 9.628849970045378e-06, 'learning_rate/action_model': 9.628849970045378e-05, 'epoch': 0.37}) +[2026-06-17 18:16:48,376][starVLA.training.train_starvla][INFO] - Step 975, Loss: {'train/loss': 0.5764665603637695, 'train/grad_norm_pre_clip': 9.297301292419434, 'timing/data': 0.01505678053945303, 'timing/model': 1.201040718704462, 'learning_rate/qwen_vl_interface': 9.628011439889997e-06, 'learning_rate/action_model': 9.628011439889996e-05, 'epoch': 0.38}) +[2026-06-17 18:16:53,257][starVLA.training.train_starvla][INFO] - Step 976, Loss: {'train/loss': 0.6125088930130005, 'train/grad_norm_pre_clip': 7.352110385894775, 'timing/data': 0.01540434081107378, 'timing/model': 1.079333613626659, 'learning_rate/qwen_vl_interface': 9.62717200216813e-06, 'learning_rate/action_model': 9.62717200216813e-05, 'epoch': 0.38}) +[2026-06-17 18:16:58,068][starVLA.training.train_starvla][INFO] - Step 977, Loss: {'train/loss': 0.5687515735626221, 'train/grad_norm_pre_clip': 8.930277824401855, 'timing/data': 0.011755241081118584, 'timing/model': 1.3953127469867468, 'learning_rate/qwen_vl_interface': 9.6263316570538e-06, 'learning_rate/action_model': 9.6263316570538e-05, 'epoch': 0.38}) +[2026-06-17 18:17:02,759][starVLA.training.train_starvla][INFO] - Step 978, Loss: {'train/loss': 0.4318920373916626, 'train/grad_norm_pre_clip': 7.967752456665039, 'timing/data': 0.011755201034247875, 'timing/model': 1.1080909287557006, 'learning_rate/qwen_vl_interface': 9.625490404721208e-06, 'learning_rate/action_model': 9.625490404721208e-05, 'epoch': 0.38}) +[2026-06-17 18:17:08,127][starVLA.training.train_starvla][INFO] - Step 979, Loss: {'train/loss': 0.7746783494949341, 'train/grad_norm_pre_clip': 8.665778160095215, 'timing/data': 0.024780621752142906, 'timing/model': 1.31380776129663, 'learning_rate/qwen_vl_interface': 9.624648245344747e-06, 'learning_rate/action_model': 9.624648245344746e-05, 'epoch': 0.38}) +[2026-06-17 18:17:13,283][starVLA.training.train_starvla][INFO] - Step 980, Loss: {'train/loss': 0.8657872676849365, 'train/grad_norm_pre_clip': 5.6595139503479, 'timing/data': 0.015749900601804256, 'timing/model': 1.3954081162810326, 'learning_rate/qwen_vl_interface': 9.623805179098996e-06, 'learning_rate/action_model': 9.623805179098996e-05, 'epoch': 0.38}) +[2026-06-17 18:17:19,170][starVLA.training.train_starvla][INFO] - Step 981, Loss: {'train/loss': 0.5581275224685669, 'train/grad_norm_pre_clip': 10.899727821350098, 'timing/data': 0.01135596726089716, 'timing/model': 1.4500062633305788, 'learning_rate/qwen_vl_interface': 9.622961206158726e-06, 'learning_rate/action_model': 9.622961206158726e-05, 'epoch': 0.38}) +[2026-06-17 18:17:25,692][starVLA.training.train_starvla][INFO] - Step 982, Loss: {'train/loss': 0.5049234628677368, 'train/grad_norm_pre_clip': 8.121404647827148, 'timing/data': 0.019213490188121796, 'timing/model': 2.9941162765026093, 'learning_rate/qwen_vl_interface': 9.622116326698894e-06, 'learning_rate/action_model': 9.622116326698893e-05, 'epoch': 0.38}) +[2026-06-17 18:17:37,908][starVLA.training.train_starvla][INFO] - Step 983, Loss: {'train/loss': 1.1018824577331543, 'train/grad_norm_pre_clip': 7.606765270233154, 'timing/data': 0.012020144611597061, 'timing/model': 1.1148948958143592, 'learning_rate/qwen_vl_interface': 9.621270540894639e-06, 'learning_rate/action_model': 9.621270540894639e-05, 'epoch': 0.38}) +[2026-06-17 18:17:42,894][starVLA.training.train_starvla][INFO] - Step 984, Loss: {'train/loss': 0.8106216192245483, 'train/grad_norm_pre_clip': 10.509757041931152, 'timing/data': 0.012065301649272442, 'timing/model': 1.131586715579033, 'learning_rate/qwen_vl_interface': 9.6204238489213e-06, 'learning_rate/action_model': 9.6204238489213e-05, 'epoch': 0.38}) +[2026-06-17 18:17:47,689][starVLA.training.train_starvla][INFO] - Step 985, Loss: {'train/loss': 0.7636857032775879, 'train/grad_norm_pre_clip': 9.73604679107666, 'timing/data': 0.026665409095585346, 'timing/model': 1.1679410878568888, 'learning_rate/qwen_vl_interface': 9.619576250954393e-06, 'learning_rate/action_model': 9.619576250954393e-05, 'epoch': 0.38}) +[2026-06-17 18:17:52,923][starVLA.training.train_starvla][INFO] - Step 986, Loss: {'train/loss': 0.49814268946647644, 'train/grad_norm_pre_clip': 7.373072624206543, 'timing/data': 0.01614762842655182, 'timing/model': 1.3401676481589675, 'learning_rate/qwen_vl_interface': 9.618727747169628e-06, 'learning_rate/action_model': 9.618727747169627e-05, 'epoch': 0.38}) +[2026-06-17 18:17:57,756][starVLA.training.train_starvla][INFO] - Step 987, Loss: {'train/loss': 0.752183198928833, 'train/grad_norm_pre_clip': 8.781270027160645, 'timing/data': 0.013200501911342144, 'timing/model': 1.5368348192423582, 'learning_rate/qwen_vl_interface': 9.617878337742902e-06, 'learning_rate/action_model': 9.617878337742902e-05, 'epoch': 0.38}) +[2026-06-17 18:18:08,231][starVLA.training.train_starvla][INFO] - Step 988, Loss: {'train/loss': 0.48602232336997986, 'train/grad_norm_pre_clip': 6.000054359436035, 'timing/data': 0.012333889491856098, 'timing/model': 1.2246865928173065, 'learning_rate/qwen_vl_interface': 9.617028022850294e-06, 'learning_rate/action_model': 9.617028022850293e-05, 'epoch': 0.38}) +[2026-06-17 18:18:13,391][starVLA.training.train_starvla][INFO] - Step 989, Loss: {'train/loss': 0.5859026908874512, 'train/grad_norm_pre_clip': 15.361150741577148, 'timing/data': 0.011118171736598015, 'timing/model': 1.2976273726671934, 'learning_rate/qwen_vl_interface': 9.616176802668078e-06, 'learning_rate/action_model': 9.616176802668078e-05, 'epoch': 0.38}) +[2026-06-17 18:18:21,789][starVLA.training.train_starvla][INFO] - Step 990, Loss: {'train/loss': 0.678450345993042, 'train/grad_norm_pre_clip': 8.10362720489502, 'timing/data': 0.011077149771153927, 'timing/model': 1.3124325079843402, 'learning_rate/qwen_vl_interface': 9.615324677372712e-06, 'learning_rate/action_model': 9.615324677372713e-05, 'epoch': 0.38}) +[2026-06-17 18:18:27,215][starVLA.training.train_starvla][INFO] - Step 991, Loss: {'train/loss': 0.6492809653282166, 'train/grad_norm_pre_clip': 10.351099967956543, 'timing/data': 0.012023949064314365, 'timing/model': 1.6434480883181095, 'learning_rate/qwen_vl_interface': 9.614471647140844e-06, 'learning_rate/action_model': 9.614471647140844e-05, 'epoch': 0.38}) +[2026-06-17 18:18:35,282][starVLA.training.train_starvla][INFO] - Step 992, Loss: {'train/loss': 0.6194961667060852, 'train/grad_norm_pre_clip': 7.248254776000977, 'timing/data': 0.013401699252426624, 'timing/model': 4.506781260482967, 'learning_rate/qwen_vl_interface': 9.613617712149309e-06, 'learning_rate/action_model': 9.613617712149308e-05, 'epoch': 0.38}) +[2026-06-17 18:18:39,628][starVLA.training.train_starvla][INFO] - Step 993, Loss: {'train/loss': 0.8080936074256897, 'train/grad_norm_pre_clip': 11.498332977294922, 'timing/data': 0.012372130528092384, 'timing/model': 1.073192236945033, 'learning_rate/qwen_vl_interface': 9.612762872575124e-06, 'learning_rate/action_model': 9.612762872575124e-05, 'epoch': 0.38}) +[2026-06-17 18:18:44,196][starVLA.training.train_starvla][INFO] - Step 994, Loss: {'train/loss': 0.6342941522598267, 'train/grad_norm_pre_clip': 7.556636810302734, 'timing/data': 0.015338181518018246, 'timing/model': 1.0892562959343195, 'learning_rate/qwen_vl_interface': 9.611907128595503e-06, 'learning_rate/action_model': 9.611907128595502e-05, 'epoch': 0.38}) +[2026-06-17 18:18:48,978][starVLA.training.train_starvla][INFO] - Step 995, Loss: {'train/loss': 0.7946251630783081, 'train/grad_norm_pre_clip': 7.776127338409424, 'timing/data': 0.02754660602658987, 'timing/model': 1.2002734560519457, 'learning_rate/qwen_vl_interface': 9.61105048038784e-06, 'learning_rate/action_model': 9.611050480387839e-05, 'epoch': 0.38}) +[2026-06-17 18:18:53,477][starVLA.training.train_starvla][INFO] - Step 996, Loss: {'train/loss': 0.5284910202026367, 'train/grad_norm_pre_clip': 8.156413078308105, 'timing/data': 0.011825334280729294, 'timing/model': 1.067443910986185, 'learning_rate/qwen_vl_interface': 9.610192928129716e-06, 'learning_rate/action_model': 9.610192928129717e-05, 'epoch': 0.38}) +[2026-06-17 18:19:02,602][starVLA.training.train_starvla][INFO] - Step 997, Loss: {'train/loss': 0.6130519509315491, 'train/grad_norm_pre_clip': 6.442831516265869, 'timing/data': 0.012002008035779, 'timing/model': 5.917741847224534, 'learning_rate/qwen_vl_interface': 9.609334471998911e-06, 'learning_rate/action_model': 9.609334471998912e-05, 'epoch': 0.38}) +[2026-06-17 18:19:07,800][starVLA.training.train_starvla][INFO] - Step 998, Loss: {'train/loss': 0.5719708204269409, 'train/grad_norm_pre_clip': 5.710515975952148, 'timing/data': 0.013090698979794979, 'timing/model': 1.1852065036073327, 'learning_rate/qwen_vl_interface': 9.608475112173379e-06, 'learning_rate/action_model': 9.608475112173379e-05, 'epoch': 0.38}) +[2026-06-17 18:19:13,076][starVLA.training.train_starvla][INFO] - Step 999, Loss: {'train/loss': 0.5935871005058289, 'train/grad_norm_pre_clip': 6.600350379943848, 'timing/data': 0.011391480453312397, 'timing/model': 1.1466710614040494, 'learning_rate/qwen_vl_interface': 9.607614848831264e-06, 'learning_rate/action_model': 9.607614848831264e-05, 'epoch': 0.38}) +[2026-06-17 18:23:10,293][starVLA.training.train_starvla][INFO] - Step 1000, Loss: {'train/loss': 0.5397909879684448, 'train/grad_norm_pre_clip': 5.115655422210693, 'eval/loss': 0.640949547290802, 'eval/latency_4/loss': 0.640949547290802, 'eval/demon_fire_precision': 0.959079283887468, 'eval/demon_fire_recall': 0.974342318934719, 'eval/demon_fire_f1': 0.9666505558240696, 'eval/demon_fire_rate_model': 0.9186490455212922, 'eval/demon_fire_rate_teacher': 0.9042584434654919, 'eval/demon_left_precision': 0.6490765171503958, 'eval/demon_left_recall': 0.5616438356164384, 'eval/demon_left_f1': 0.602203182374541, 'eval/demon_left_rate_model': 0.11130690161527165, 'eval/demon_left_rate_teacher': 0.12863436123348018, 'eval/demon_right_precision': 0.7017208413001912, 'eval/demon_right_recall': 0.6624548736462094, 'eval/demon_right_f1': 0.681522748375116, 'eval/demon_right_rate_model': 0.15359765051395008, 'eval/demon_right_rate_teacher': 0.16270190895741557, 'eval/demon_move_f1': 0.6418629653748285, 'eval/demon_cc_f1': 0.8392504098019632, 'eval/latency_4/demon_fire_precision': 0.959079283887468, 'eval/latency_4/demon_fire_recall': 0.974342318934719, 'eval/latency_4/demon_fire_f1': 0.9666505558240696, 'eval/latency_4/demon_fire_rate_model': 0.9186490455212922, 'eval/latency_4/demon_fire_rate_teacher': 0.9042584434654919, 'eval/latency_4/demon_left_precision': 0.6490765171503958, 'eval/latency_4/demon_left_recall': 0.5616438356164384, 'eval/latency_4/demon_left_f1': 0.602203182374541, 'eval/latency_4/demon_left_rate_model': 0.11130690161527165, 'eval/latency_4/demon_left_rate_teacher': 0.12863436123348018, 'eval/latency_4/demon_right_precision': 0.7017208413001912, 'eval/latency_4/demon_right_recall': 0.6624548736462094, 'eval/latency_4/demon_right_f1': 0.681522748375116, 'eval/latency_4/demon_right_rate_model': 0.15359765051395008, 'eval/latency_4/demon_right_rate_teacher': 0.16270190895741557, 'eval/latency_4/demon_move_f1': 0.6418629653748285, 'eval/latency_4/demon_cc_f1': 0.8392504098019632, 'timing/data': 0.011680470779538155, 'timing/model': 1.3309743823483586, 'learning_rate/qwen_vl_interface': 9.606753682150904e-06, 'learning_rate/action_model': 9.606753682150903e-05, 'epoch': 0.38}) +[2026-06-17 18:23:10,296][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_1000_state +[2026-06-17 18:23:10,305][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 18:23:19,412][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_1000_state/model.safetensors +[2026-06-17 18:23:41,810][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_1000_state/optimizer.bin +[2026-06-17 18:23:41,812][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_1000_state/sampler.bin +[2026-06-17 18:23:41,812][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_1000_state/sampler_1.bin +[2026-06-17 18:23:41,815][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_1000_state/random_states_0.pkl +[2026-06-17 18:23:41,820][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1000 +[2026-06-17 18:23:41,820][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 18:23:41,837][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 18:23:46,554][starVLA.training.train_starvla][INFO] - Step 1001, Loss: {'train/loss': 0.6344307065010071, 'train/grad_norm_pre_clip': 5.45216178894043, 'timing/data': 0.013102545402944088, 'timing/model': 1.178202378563583, 'learning_rate/qwen_vl_interface': 9.605891612310815e-06, 'learning_rate/action_model': 9.605891612310814e-05, 'epoch': 0.39}) +[2026-06-17 18:23:51,009][starVLA.training.train_starvla][INFO] - Step 1002, Loss: {'train/loss': 0.6381576061248779, 'train/grad_norm_pre_clip': 6.986920356750488, 'timing/data': 0.010608198121190071, 'timing/model': 1.380974331870675, 'learning_rate/qwen_vl_interface': 9.60502863948971e-06, 'learning_rate/action_model': 9.60502863948971e-05, 'epoch': 0.39}) +[2026-06-17 18:23:55,314][starVLA.training.train_starvla][INFO] - Step 1003, Loss: {'train/loss': 0.4642742872238159, 'train/grad_norm_pre_clip': 8.020008087158203, 'timing/data': 0.01198651734739542, 'timing/model': 1.0794842736795545, 'learning_rate/qwen_vl_interface': 9.604164763866479e-06, 'learning_rate/action_model': 9.60416476386648e-05, 'epoch': 0.39}) +[2026-06-17 18:24:03,296][starVLA.training.train_starvla][INFO] - Step 1004, Loss: {'train/loss': 0.5742278695106506, 'train/grad_norm_pre_clip': 8.982733726501465, 'timing/data': 0.012269836850464344, 'timing/model': 3.714884527027607, 'learning_rate/qwen_vl_interface': 9.60329998562021e-06, 'learning_rate/action_model': 9.60329998562021e-05, 'epoch': 0.39}) +[2026-06-17 18:24:08,122][starVLA.training.train_starvla][INFO] - Step 1005, Loss: {'train/loss': 0.47218114137649536, 'train/grad_norm_pre_clip': 7.077021598815918, 'timing/data': 0.01299351267516613, 'timing/model': 1.1640815706923604, 'learning_rate/qwen_vl_interface': 9.602434304930168e-06, 'learning_rate/action_model': 9.602434304930168e-05, 'epoch': 0.39}) +[2026-06-17 18:24:17,830][starVLA.training.train_starvla][INFO] - Step 1006, Loss: {'train/loss': 0.5502521395683289, 'train/grad_norm_pre_clip': 7.53270149230957, 'timing/data': 0.01583554409444332, 'timing/model': 4.7946535320952535, 'learning_rate/qwen_vl_interface': 9.601567721975811e-06, 'learning_rate/action_model': 9.601567721975812e-05, 'epoch': 0.39}) +[2026-06-17 18:24:25,140][starVLA.training.train_starvla][INFO] - Step 1007, Loss: {'train/loss': 0.4013945162296295, 'train/grad_norm_pre_clip': 6.1965837478637695, 'timing/data': 0.01121472381055355, 'timing/model': 1.4888617657124996, 'learning_rate/qwen_vl_interface': 9.600700236936785e-06, 'learning_rate/action_model': 9.600700236936785e-05, 'epoch': 0.39}) +[2026-06-17 18:24:31,649][starVLA.training.train_starvla][INFO] - Step 1008, Loss: {'train/loss': 0.4793494939804077, 'train/grad_norm_pre_clip': 8.471710205078125, 'timing/data': 0.012987122870981693, 'timing/model': 1.1335409581661224, 'learning_rate/qwen_vl_interface': 9.599831849992916e-06, 'learning_rate/action_model': 9.599831849992916e-05, 'epoch': 0.39}) +[2026-06-17 18:24:41,134][starVLA.training.train_starvla][INFO] - Step 1009, Loss: {'train/loss': 0.44086939096450806, 'train/grad_norm_pre_clip': 7.851612091064453, 'timing/data': 0.010602844879031181, 'timing/model': 2.4936793707311153, 'learning_rate/qwen_vl_interface': 9.598962561324225e-06, 'learning_rate/action_model': 9.598962561324225e-05, 'epoch': 0.39}) +[2026-06-17 18:24:46,478][starVLA.training.train_starvla][INFO] - Step 1010, Loss: {'train/loss': 0.7907339930534363, 'train/grad_norm_pre_clip': 9.436415672302246, 'timing/data': 0.013262256979942322, 'timing/model': 1.1295560393482447, 'learning_rate/qwen_vl_interface': 9.598092371110916e-06, 'learning_rate/action_model': 9.598092371110916e-05, 'epoch': 0.39}) +[2026-06-17 18:24:54,347][starVLA.training.train_starvla][INFO] - Step 1011, Loss: {'train/loss': 0.5368844866752625, 'train/grad_norm_pre_clip': 14.692099571228027, 'timing/data': 0.014560453593730927, 'timing/model': 3.6075317841023207, 'learning_rate/qwen_vl_interface': 9.59722127953338e-06, 'learning_rate/action_model': 9.597221279533379e-05, 'epoch': 0.39}) +[2026-06-17 18:25:00,904][starVLA.training.train_starvla][INFO] - Step 1012, Loss: {'train/loss': 0.6101450324058533, 'train/grad_norm_pre_clip': 12.935964584350586, 'timing/data': 0.01165954489260912, 'timing/model': 1.463558322750032, 'learning_rate/qwen_vl_interface': 9.596349286772194e-06, 'learning_rate/action_model': 9.596349286772193e-05, 'epoch': 0.39}) +[2026-06-17 18:25:05,407][starVLA.training.train_starvla][INFO] - Step 1013, Loss: {'train/loss': 0.6563089489936829, 'train/grad_norm_pre_clip': 9.080286979675293, 'timing/data': 0.011043469421565533, 'timing/model': 1.0938444891944528, 'learning_rate/qwen_vl_interface': 9.595476393008123e-06, 'learning_rate/action_model': 9.595476393008123e-05, 'epoch': 0.39}) +[2026-06-17 18:25:10,281][starVLA.training.train_starvla][INFO] - Step 1014, Loss: {'train/loss': 0.739304780960083, 'train/grad_norm_pre_clip': 11.560145378112793, 'timing/data': 0.012529406696557999, 'timing/model': 1.1187272118404508, 'learning_rate/qwen_vl_interface': 9.59460259842212e-06, 'learning_rate/action_model': 9.59460259842212e-05, 'epoch': 0.39}) +[2026-06-17 18:25:15,117][starVLA.training.train_starvla][INFO] - Step 1015, Loss: {'train/loss': 0.5986677408218384, 'train/grad_norm_pre_clip': 8.520790100097656, 'timing/data': 0.010636367835104465, 'timing/model': 1.1012154230847955, 'learning_rate/qwen_vl_interface': 9.593727903195323e-06, 'learning_rate/action_model': 9.593727903195324e-05, 'epoch': 0.39}) +[2026-06-17 18:25:19,951][starVLA.training.train_starvla][INFO] - Step 1016, Loss: {'train/loss': 0.4591262638568878, 'train/grad_norm_pre_clip': 9.360437393188477, 'timing/data': 0.01099860668182373, 'timing/model': 1.083109780214727, 'learning_rate/qwen_vl_interface': 9.59285230750906e-06, 'learning_rate/action_model': 9.592852307509059e-05, 'epoch': 0.39}) +[2026-06-17 18:25:24,849][starVLA.training.train_starvla][INFO] - Step 1017, Loss: {'train/loss': 0.8148298263549805, 'train/grad_norm_pre_clip': 13.950276374816895, 'timing/data': 0.020185822620987892, 'timing/model': 1.5267934100702405, 'learning_rate/qwen_vl_interface': 9.591975811544839e-06, 'learning_rate/action_model': 9.591975811544839e-05, 'epoch': 0.39}) +[2026-06-17 18:25:29,250][starVLA.training.train_starvla][INFO] - Step 1018, Loss: {'train/loss': 0.7279303669929504, 'train/grad_norm_pre_clip': 8.256220817565918, 'timing/data': 0.013554333709180355, 'timing/model': 1.218448824249208, 'learning_rate/qwen_vl_interface': 9.59109841548436e-06, 'learning_rate/action_model': 9.591098415484359e-05, 'epoch': 0.39}) +[2026-06-17 18:25:34,129][starVLA.training.train_starvla][INFO] - Step 1019, Loss: {'train/loss': 0.6658692955970764, 'train/grad_norm_pre_clip': 6.828184604644775, 'timing/data': 0.010571195743978024, 'timing/model': 1.3395055597648025, 'learning_rate/qwen_vl_interface': 9.590220119509508e-06, 'learning_rate/action_model': 9.590220119509508e-05, 'epoch': 0.39}) +[2026-06-17 18:25:38,920][starVLA.training.train_starvla][INFO] - Step 1020, Loss: {'train/loss': 0.7735174298286438, 'train/grad_norm_pre_clip': 10.981858253479004, 'timing/data': 0.012519952841103077, 'timing/model': 1.0872538741678, 'learning_rate/qwen_vl_interface': 9.589340923802355e-06, 'learning_rate/action_model': 9.589340923802354e-05, 'epoch': 0.39}) +[2026-06-17 18:25:43,610][starVLA.training.train_starvla][INFO] - Step 1021, Loss: {'train/loss': 0.8174148201942444, 'train/grad_norm_pre_clip': 11.228955268859863, 'timing/data': 0.011401692405343056, 'timing/model': 1.5184232126921415, 'learning_rate/qwen_vl_interface': 9.58846082854516e-06, 'learning_rate/action_model': 9.58846082854516e-05, 'epoch': 0.39}) +[2026-06-17 18:25:48,483][starVLA.training.train_starvla][INFO] - Step 1022, Loss: {'train/loss': 0.8211666941642761, 'train/grad_norm_pre_clip': 8.084985733032227, 'timing/data': 0.011556894518435001, 'timing/model': 1.4714968223124743, 'learning_rate/qwen_vl_interface': 9.587579833920366e-06, 'learning_rate/action_model': 9.587579833920365e-05, 'epoch': 0.39}) +[2026-06-17 18:25:52,836][starVLA.training.train_starvla][INFO] - Step 1023, Loss: {'train/loss': 0.4901782274246216, 'train/grad_norm_pre_clip': 6.626978397369385, 'timing/data': 0.012740553356707096, 'timing/model': 1.1045314511284232, 'learning_rate/qwen_vl_interface': 9.586697940110605e-06, 'learning_rate/action_model': 9.586697940110605e-05, 'epoch': 0.39}) +[2026-06-17 18:25:58,003][starVLA.training.train_starvla][INFO] - Step 1024, Loss: {'train/loss': 0.47535791993141174, 'train/grad_norm_pre_clip': 9.11539363861084, 'timing/data': 0.011240484192967415, 'timing/model': 1.0727206114679575, 'learning_rate/qwen_vl_interface': 9.585815147298696e-06, 'learning_rate/action_model': 9.585815147298696e-05, 'epoch': 0.39}) +[2026-06-17 18:26:02,699][starVLA.training.train_starvla][INFO] - Step 1025, Loss: {'train/loss': 0.7663823366165161, 'train/grad_norm_pre_clip': 6.5066986083984375, 'timing/data': 0.011353466659784317, 'timing/model': 1.0745916943997145, 'learning_rate/qwen_vl_interface': 9.58493145566764e-06, 'learning_rate/action_model': 9.58493145566764e-05, 'epoch': 0.39}) +[2026-06-17 18:26:07,628][starVLA.training.train_starvla][INFO] - Step 1026, Loss: {'train/loss': 0.6809972524642944, 'train/grad_norm_pre_clip': 5.77929162979126, 'timing/data': 0.010382475331425667, 'timing/model': 1.1536443587392569, 'learning_rate/qwen_vl_interface': 9.584046865400627e-06, 'learning_rate/action_model': 9.584046865400627e-05, 'epoch': 0.39}) +[2026-06-17 18:26:13,069][starVLA.training.train_starvla][INFO] - Step 1027, Loss: {'train/loss': 0.6814751625061035, 'train/grad_norm_pre_clip': 6.145144462585449, 'timing/data': 0.013012472540140152, 'timing/model': 1.7480264706537127, 'learning_rate/qwen_vl_interface': 9.583161376681039e-06, 'learning_rate/action_model': 9.583161376681038e-05, 'epoch': 0.4}) +[2026-06-17 18:26:23,617][starVLA.training.train_starvla][INFO] - Step 1028, Loss: {'train/loss': 0.706661581993103, 'train/grad_norm_pre_clip': 8.635322570800781, 'timing/data': 0.02010706253349781, 'timing/model': 1.3220203602686524, 'learning_rate/qwen_vl_interface': 9.582274989692432e-06, 'learning_rate/action_model': 9.582274989692432e-05, 'epoch': 0.4}) +[2026-06-17 18:26:32,161][starVLA.training.train_starvla][INFO] - Step 1029, Loss: {'train/loss': 0.8517388701438904, 'train/grad_norm_pre_clip': 7.923851013183594, 'timing/data': 0.010854116640985012, 'timing/model': 2.716544024646282, 'learning_rate/qwen_vl_interface': 9.58138770461856e-06, 'learning_rate/action_model': 9.58138770461856e-05, 'epoch': 0.4}) +[2026-06-17 18:26:45,705][starVLA.training.train_starvla][INFO] - Step 1030, Loss: {'train/loss': 0.5702354311943054, 'train/grad_norm_pre_clip': 8.771078109741211, 'timing/data': 0.01084301620721817, 'timing/model': 1.4033774472773075, 'learning_rate/qwen_vl_interface': 9.580499521643355e-06, 'learning_rate/action_model': 9.580499521643355e-05, 'epoch': 0.4}) +[2026-06-17 18:26:53,938][starVLA.training.train_starvla][INFO] - Step 1031, Loss: {'train/loss': 0.739721417427063, 'train/grad_norm_pre_clip': 8.479931831359863, 'timing/data': 0.013640900142490864, 'timing/model': 1.2073546163737774, 'learning_rate/qwen_vl_interface': 9.579610440950939e-06, 'learning_rate/action_model': 9.579610440950937e-05, 'epoch': 0.4}) +[2026-06-17 18:27:01,562][starVLA.training.train_starvla][INFO] - Step 1032, Loss: {'train/loss': 0.6620903611183167, 'train/grad_norm_pre_clip': 7.423482418060303, 'timing/data': 0.015498983673751354, 'timing/model': 1.4478572439402342, 'learning_rate/qwen_vl_interface': 9.57872046272562e-06, 'learning_rate/action_model': 9.57872046272562e-05, 'epoch': 0.4}) +[2026-06-17 18:27:10,232][starVLA.training.train_starvla][INFO] - Step 1033, Loss: {'train/loss': 0.6435162425041199, 'train/grad_norm_pre_clip': 8.795005798339844, 'timing/data': 0.011568568646907806, 'timing/model': 2.859253345988691, 'learning_rate/qwen_vl_interface': 9.577829587151892e-06, 'learning_rate/action_model': 9.577829587151891e-05, 'epoch': 0.4}) +[2026-06-17 18:27:23,806][starVLA.training.train_starvla][INFO] - Step 1034, Loss: {'train/loss': 0.7956319451332092, 'train/grad_norm_pre_clip': 7.042773246765137, 'timing/data': 0.011197769083082676, 'timing/model': 4.231045340187848, 'learning_rate/qwen_vl_interface': 9.576937814414433e-06, 'learning_rate/action_model': 9.576937814414432e-05, 'epoch': 0.4}) +[2026-06-17 18:27:41,283][starVLA.training.train_starvla][INFO] - Step 1035, Loss: {'train/loss': 0.49941036105155945, 'train/grad_norm_pre_clip': 5.744216442108154, 'timing/data': 0.022479398176074028, 'timing/model': 4.5901006972417235, 'learning_rate/qwen_vl_interface': 9.576045144698109e-06, 'learning_rate/action_model': 9.576045144698108e-05, 'epoch': 0.4}) +[2026-06-17 18:27:47,037][starVLA.training.train_starvla][INFO] - Step 1036, Loss: {'train/loss': 0.5364043712615967, 'train/grad_norm_pre_clip': 6.767968654632568, 'timing/data': 0.010945947840809822, 'timing/model': 1.175316235050559, 'learning_rate/qwen_vl_interface': 9.57515157818797e-06, 'learning_rate/action_model': 9.57515157818797e-05, 'epoch': 0.4}) +[2026-06-17 18:27:52,004][starVLA.training.train_starvla][INFO] - Step 1037, Loss: {'train/loss': 0.7921776175498962, 'train/grad_norm_pre_clip': 7.162604331970215, 'timing/data': 0.019763078540563583, 'timing/model': 1.6266026627272367, 'learning_rate/qwen_vl_interface': 9.574257115069256e-06, 'learning_rate/action_model': 9.574257115069256e-05, 'epoch': 0.4}) +[2026-06-17 18:27:56,507][starVLA.training.train_starvla][INFO] - Step 1038, Loss: {'train/loss': 0.6445093750953674, 'train/grad_norm_pre_clip': 6.519508361816406, 'timing/data': 0.011453545652329922, 'timing/model': 1.0679532093927264, 'learning_rate/qwen_vl_interface': 9.573361755527389e-06, 'learning_rate/action_model': 9.573361755527389e-05, 'epoch': 0.4}) +[2026-06-17 18:28:01,047][starVLA.training.train_starvla][INFO] - Step 1039, Loss: {'train/loss': 0.611622154712677, 'train/grad_norm_pre_clip': 6.304576873779297, 'timing/data': 0.012832647189497948, 'timing/model': 1.0817064261063933, 'learning_rate/qwen_vl_interface': 9.572465499747978e-06, 'learning_rate/action_model': 9.572465499747978e-05, 'epoch': 0.4}) +[2026-06-17 18:28:05,819][starVLA.training.train_starvla][INFO] - Step 1040, Loss: {'train/loss': 0.5477952361106873, 'train/grad_norm_pre_clip': 5.869062900543213, 'timing/data': 0.012249303981661797, 'timing/model': 1.0722776176407933, 'learning_rate/qwen_vl_interface': 9.571568347916819e-06, 'learning_rate/action_model': 9.571568347916819e-05, 'epoch': 0.4}) +[2026-06-17 18:28:12,101][starVLA.training.train_starvla][INFO] - Step 1041, Loss: {'train/loss': 0.6288712024688721, 'train/grad_norm_pre_clip': 7.040038108825684, 'timing/data': 0.01031422894448042, 'timing/model': 2.5566179901361465, 'learning_rate/qwen_vl_interface': 9.570670300219889e-06, 'learning_rate/action_model': 9.570670300219889e-05, 'epoch': 0.4}) +[2026-06-17 18:28:17,195][starVLA.training.train_starvla][INFO] - Step 1042, Loss: {'train/loss': 0.7488119006156921, 'train/grad_norm_pre_clip': 7.094275951385498, 'timing/data': 0.018056713975965977, 'timing/model': 1.4470246164128184, 'learning_rate/qwen_vl_interface': 9.56977135684336e-06, 'learning_rate/action_model': 9.569771356843359e-05, 'epoch': 0.4}) +[2026-06-17 18:28:22,001][starVLA.training.train_starvla][INFO] - Step 1043, Loss: {'train/loss': 0.8704031705856323, 'train/grad_norm_pre_clip': 8.776504516601562, 'timing/data': 0.01506454311311245, 'timing/model': 1.1880070278421044, 'learning_rate/qwen_vl_interface': 9.568871517973576e-06, 'learning_rate/action_model': 9.568871517973577e-05, 'epoch': 0.4}) +[2026-06-17 18:28:34,372][starVLA.training.train_starvla][INFO] - Step 1044, Loss: {'train/loss': 0.8459386229515076, 'train/grad_norm_pre_clip': 6.9532904624938965, 'timing/data': 0.01429041475057602, 'timing/model': 4.090329873375595, 'learning_rate/qwen_vl_interface': 9.567970783797081e-06, 'learning_rate/action_model': 9.567970783797082e-05, 'epoch': 0.4}) +[2026-06-17 18:28:39,738][starVLA.training.train_starvla][INFO] - Step 1045, Loss: {'train/loss': 0.7810350656509399, 'train/grad_norm_pre_clip': 8.775280952453613, 'timing/data': 0.011051762849092484, 'timing/model': 1.252829018048942, 'learning_rate/qwen_vl_interface': 9.567069154500597e-06, 'learning_rate/action_model': 9.567069154500596e-05, 'epoch': 0.4}) +[2026-06-17 18:28:46,648][starVLA.training.train_starvla][INFO] - Step 1046, Loss: {'train/loss': 0.4581170082092285, 'train/grad_norm_pre_clip': 6.855525493621826, 'timing/data': 0.01277520414441824, 'timing/model': 2.974511749111116, 'learning_rate/qwen_vl_interface': 9.56616663027103e-06, 'learning_rate/action_model': 9.566166630271029e-05, 'epoch': 0.4}) +[2026-06-17 18:28:50,989][starVLA.training.train_starvla][INFO] - Step 1047, Loss: {'train/loss': 0.6501606702804565, 'train/grad_norm_pre_clip': 7.629181385040283, 'timing/data': 0.020213698036968708, 'timing/model': 1.138083011843264, 'learning_rate/qwen_vl_interface': 9.56526321129548e-06, 'learning_rate/action_model': 9.565263211295479e-05, 'epoch': 0.4}) +[2026-06-17 18:28:55,278][starVLA.training.train_starvla][INFO] - Step 1048, Loss: {'train/loss': 0.8936353325843811, 'train/grad_norm_pre_clip': 8.43492317199707, 'timing/data': 0.009361134842038155, 'timing/model': 1.0371624929830432, 'learning_rate/qwen_vl_interface': 9.56435889776122e-06, 'learning_rate/action_model': 9.564358897761218e-05, 'epoch': 0.4}) +[2026-06-17 18:28:59,572][starVLA.training.train_starvla][INFO] - Step 1049, Loss: {'train/loss': 0.4375067353248596, 'train/grad_norm_pre_clip': 4.776138782501221, 'timing/data': 0.0110106710344553, 'timing/model': 1.0512141091749072, 'learning_rate/qwen_vl_interface': 9.563453689855717e-06, 'learning_rate/action_model': 9.563453689855716e-05, 'epoch': 0.4}) +[2026-06-17 18:29:03,819][starVLA.training.train_starvla][INFO] - Step 1050, Loss: {'train/loss': 0.8858456611633301, 'train/grad_norm_pre_clip': 6.978402137756348, 'timing/data': 0.009569639340043068, 'timing/model': 1.0409827018156648, 'learning_rate/qwen_vl_interface': 9.562547587766624e-06, 'learning_rate/action_model': 9.562547587766623e-05, 'epoch': 0.4}) +[2026-06-17 18:29:08,307][starVLA.training.train_starvla][INFO] - Step 1051, Loss: {'train/loss': 0.49477440118789673, 'train/grad_norm_pre_clip': 6.388482093811035, 'timing/data': 0.011458088643848896, 'timing/model': 1.4141857912763953, 'learning_rate/qwen_vl_interface': 9.561640591681775e-06, 'learning_rate/action_model': 9.561640591681774e-05, 'epoch': 0.4}) +[2026-06-17 18:29:12,830][starVLA.training.train_starvla][INFO] - Step 1052, Loss: {'train/loss': 0.6604200601577759, 'train/grad_norm_pre_clip': 8.23923110961914, 'timing/data': 0.010012718848884106, 'timing/model': 1.2460020976141095, 'learning_rate/qwen_vl_interface': 9.560732701789193e-06, 'learning_rate/action_model': 9.560732701789193e-05, 'epoch': 0.4}) +[2026-06-17 18:29:16,886][starVLA.training.train_starvla][INFO] - Step 1053, Loss: {'train/loss': 0.43422555923461914, 'train/grad_norm_pre_clip': 4.9401116371154785, 'timing/data': 0.011466702446341515, 'timing/model': 1.0445148581638932, 'learning_rate/qwen_vl_interface': 9.55982391827708e-06, 'learning_rate/action_model': 9.559823918277081e-05, 'epoch': 0.41}) +[2026-06-17 18:29:21,133][starVLA.training.train_starvla][INFO] - Step 1054, Loss: {'train/loss': 0.3332032859325409, 'train/grad_norm_pre_clip': 4.522992134094238, 'timing/data': 0.009317781776189804, 'timing/model': 1.0435706749558449, 'learning_rate/qwen_vl_interface': 9.558914241333835e-06, 'learning_rate/action_model': 9.558914241333835e-05, 'epoch': 0.41}) +[2026-06-17 18:29:25,362][starVLA.training.train_starvla][INFO] - Step 1055, Loss: {'train/loss': 0.7896407246589661, 'train/grad_norm_pre_clip': 9.390758514404297, 'timing/data': 0.009495849721133709, 'timing/model': 1.0349292140454054, 'learning_rate/qwen_vl_interface': 9.55800367114803e-06, 'learning_rate/action_model': 9.55800367114803e-05, 'epoch': 0.41}) +[2026-06-17 18:29:29,640][starVLA.training.train_starvla][INFO] - Step 1056, Loss: {'train/loss': 0.7914896607398987, 'train/grad_norm_pre_clip': 13.247727394104004, 'timing/data': 0.010249932296574116, 'timing/model': 1.0587324481457472, 'learning_rate/qwen_vl_interface': 9.557092207908428e-06, 'learning_rate/action_model': 9.557092207908428e-05, 'epoch': 0.41}) +[2026-06-17 18:29:33,852][starVLA.training.train_starvla][INFO] - Step 1057, Loss: {'train/loss': 0.6661474108695984, 'train/grad_norm_pre_clip': 7.385007858276367, 'timing/data': 0.009095017798244953, 'timing/model': 1.2246834384277463, 'learning_rate/qwen_vl_interface': 9.556179851803978e-06, 'learning_rate/action_model': 9.556179851803978e-05, 'epoch': 0.41}) +[2026-06-17 18:29:38,676][starVLA.training.train_starvla][INFO] - Step 1058, Loss: {'train/loss': 0.6534567475318909, 'train/grad_norm_pre_clip': 7.938417911529541, 'timing/data': 0.009350940585136414, 'timing/model': 1.0469848969951272, 'learning_rate/qwen_vl_interface': 9.555266603023809e-06, 'learning_rate/action_model': 9.555266603023808e-05, 'epoch': 0.41}) +[2026-06-17 18:29:43,094][starVLA.training.train_starvla][INFO] - Step 1059, Loss: {'train/loss': 0.5670627951622009, 'train/grad_norm_pre_clip': 10.771045684814453, 'timing/data': 0.009284009225666523, 'timing/model': 1.077943479642272, 'learning_rate/qwen_vl_interface': 9.554352461757243e-06, 'learning_rate/action_model': 9.554352461757242e-05, 'epoch': 0.41}) +[2026-06-17 18:29:51,366][starVLA.training.train_starvla][INFO] - Step 1060, Loss: {'train/loss': 0.9576247334480286, 'train/grad_norm_pre_clip': 9.62996768951416, 'timing/data': 0.01132814958691597, 'timing/model': 1.1062930347397923, 'learning_rate/qwen_vl_interface': 9.553437428193778e-06, 'learning_rate/action_model': 9.553437428193778e-05, 'epoch': 0.41}) +[2026-06-17 18:29:55,973][starVLA.training.train_starvla][INFO] - Step 1061, Loss: {'train/loss': 0.4707597494125366, 'train/grad_norm_pre_clip': 5.985938549041748, 'timing/data': 0.01100615318864584, 'timing/model': 1.0914530605077744, 'learning_rate/qwen_vl_interface': 9.552521502523104e-06, 'learning_rate/action_model': 9.552521502523102e-05, 'epoch': 0.41}) +[2026-06-17 18:30:02,943][starVLA.training.train_starvla][INFO] - Step 1062, Loss: {'train/loss': 0.5645387172698975, 'train/grad_norm_pre_clip': 5.834814548492432, 'timing/data': 0.011125565506517887, 'timing/model': 3.755398136563599, 'learning_rate/qwen_vl_interface': 9.551604684935093e-06, 'learning_rate/action_model': 9.551604684935092e-05, 'epoch': 0.41}) +[2026-06-17 18:30:07,422][starVLA.training.train_starvla][INFO] - Step 1063, Loss: {'train/loss': 0.5392230153083801, 'train/grad_norm_pre_clip': 5.49293327331543, 'timing/data': 0.011592689901590347, 'timing/model': 1.0999650657176971, 'learning_rate/qwen_vl_interface': 9.550686975619803e-06, 'learning_rate/action_model': 9.550686975619802e-05, 'epoch': 0.41}) +[2026-06-17 18:30:12,247][starVLA.training.train_starvla][INFO] - Step 1064, Loss: {'train/loss': 0.5145567655563354, 'train/grad_norm_pre_clip': 8.117114067077637, 'timing/data': 0.014796563424170017, 'timing/model': 1.1839683800935745, 'learning_rate/qwen_vl_interface': 9.549768374767473e-06, 'learning_rate/action_model': 9.549768374767473e-05, 'epoch': 0.41}) +[2026-06-17 18:30:17,213][starVLA.training.train_starvla][INFO] - Step 1065, Loss: {'train/loss': 0.739690899848938, 'train/grad_norm_pre_clip': 6.325065612792969, 'timing/data': 0.0111931674182415, 'timing/model': 1.1484949458390474, 'learning_rate/qwen_vl_interface': 9.548848882568533e-06, 'learning_rate/action_model': 9.548848882568532e-05, 'epoch': 0.41}) +[2026-06-17 18:30:21,986][starVLA.training.train_starvla][INFO] - Step 1066, Loss: {'train/loss': 0.24137219786643982, 'train/grad_norm_pre_clip': 4.334833145141602, 'timing/data': 0.011611984111368656, 'timing/model': 1.1032182415947318, 'learning_rate/qwen_vl_interface': 9.547928499213593e-06, 'learning_rate/action_model': 9.547928499213593e-05, 'epoch': 0.41}) +[2026-06-17 18:30:28,651][starVLA.training.train_starvla][INFO] - Step 1067, Loss: {'train/loss': 0.42217206954956055, 'train/grad_norm_pre_clip': 8.331536293029785, 'timing/data': 0.012359593994915485, 'timing/model': 3.418452956713736, 'learning_rate/qwen_vl_interface': 9.547007224893452e-06, 'learning_rate/action_model': 9.547007224893452e-05, 'epoch': 0.41}) +[2026-06-17 18:30:33,261][starVLA.training.train_starvla][INFO] - Step 1068, Loss: {'train/loss': 0.5458937883377075, 'train/grad_norm_pre_clip': 6.637343883514404, 'timing/data': 0.012598596513271332, 'timing/model': 1.2417464284226298, 'learning_rate/qwen_vl_interface': 9.546085059799088e-06, 'learning_rate/action_model': 9.546085059799088e-05, 'epoch': 0.41}) +[2026-06-17 18:30:38,158][starVLA.training.train_starvla][INFO] - Step 1069, Loss: {'train/loss': 0.7454882264137268, 'train/grad_norm_pre_clip': 7.28493070602417, 'timing/data': 0.017450264655053616, 'timing/model': 1.158751959912479, 'learning_rate/qwen_vl_interface': 9.545162004121672e-06, 'learning_rate/action_model': 9.54516200412167e-05, 'epoch': 0.41}) +[2026-06-17 18:30:43,070][starVLA.training.train_starvla][INFO] - Step 1070, Loss: {'train/loss': 0.3552994728088379, 'train/grad_norm_pre_clip': 5.653204917907715, 'timing/data': 0.013156753033399582, 'timing/model': 1.1297469353303313, 'learning_rate/qwen_vl_interface': 9.544238058052546e-06, 'learning_rate/action_model': 9.544238058052545e-05, 'epoch': 0.41}) +[2026-06-17 18:30:48,015][starVLA.training.train_starvla][INFO] - Step 1071, Loss: {'train/loss': 0.4747295677661896, 'train/grad_norm_pre_clip': 7.29633092880249, 'timing/data': 0.012645472772419453, 'timing/model': 1.2611040016636252, 'learning_rate/qwen_vl_interface': 9.54331322178325e-06, 'learning_rate/action_model': 9.543313221783251e-05, 'epoch': 0.41}) +[2026-06-17 18:30:56,878][starVLA.training.train_starvla][INFO] - Step 1072, Loss: {'train/loss': 0.5405827760696411, 'train/grad_norm_pre_clip': 3.836280584335327, 'timing/data': 0.011033156886696815, 'timing/model': 1.789303851313889, 'learning_rate/qwen_vl_interface': 9.542387495505505e-06, 'learning_rate/action_model': 9.542387495505506e-05, 'epoch': 0.41}) +[2026-06-17 18:31:02,080][starVLA.training.train_starvla][INFO] - Step 1073, Loss: {'train/loss': 0.5545523762702942, 'train/grad_norm_pre_clip': 10.755664825439453, 'timing/data': 0.015411133877933025, 'timing/model': 1.2285852814093232, 'learning_rate/qwen_vl_interface': 9.541460879411214e-06, 'learning_rate/action_model': 9.541460879411215e-05, 'epoch': 0.41}) +[2026-06-17 18:31:08,107][starVLA.training.train_starvla][INFO] - Step 1074, Loss: {'train/loss': 0.538659930229187, 'train/grad_norm_pre_clip': 8.05401611328125, 'timing/data': 0.014701636508107185, 'timing/model': 1.1655449224635959, 'learning_rate/qwen_vl_interface': 9.540533373692466e-06, 'learning_rate/action_model': 9.540533373692466e-05, 'epoch': 0.41}) +[2026-06-17 18:31:18,297][starVLA.training.train_starvla][INFO] - Step 1075, Loss: {'train/loss': 0.5074322819709778, 'train/grad_norm_pre_clip': 8.843012809753418, 'timing/data': 0.022877697832882404, 'timing/model': 4.598042831756175, 'learning_rate/qwen_vl_interface': 9.539604978541531e-06, 'learning_rate/action_model': 9.539604978541531e-05, 'epoch': 0.41}) +[2026-06-17 18:31:27,953][starVLA.training.train_starvla][INFO] - Step 1076, Loss: {'train/loss': 0.34660759568214417, 'train/grad_norm_pre_clip': 9.01015567779541, 'timing/data': 0.01296385657042265, 'timing/model': 1.0883198212832212, 'learning_rate/qwen_vl_interface': 9.538675694150868e-06, 'learning_rate/action_model': 9.538675694150868e-05, 'epoch': 0.41}) +[2026-06-17 18:31:36,213][starVLA.training.train_starvla][INFO] - Step 1077, Loss: {'train/loss': 0.4186897575855255, 'train/grad_norm_pre_clip': 10.092885971069336, 'timing/data': 0.012747837230563164, 'timing/model': 1.5816460773348808, 'learning_rate/qwen_vl_interface': 9.53774552071312e-06, 'learning_rate/action_model': 9.537745520713119e-05, 'epoch': 0.41}) +[2026-06-17 18:31:44,210][starVLA.training.train_starvla][INFO] - Step 1078, Loss: {'train/loss': 0.6837762594223022, 'train/grad_norm_pre_clip': 9.04025650024414, 'timing/data': 0.016453332267701626, 'timing/model': 2.5806963611394167, 'learning_rate/qwen_vl_interface': 9.53681445842111e-06, 'learning_rate/action_model': 9.53681445842111e-05, 'epoch': 0.41}) +[2026-06-17 18:31:53,204][starVLA.training.train_starvla][INFO] - Step 1079, Loss: {'train/loss': 0.5566554665565491, 'train/grad_norm_pre_clip': 8.343786239624023, 'timing/data': 0.012480871751904488, 'timing/model': 1.2438810197636485, 'learning_rate/qwen_vl_interface': 9.53588250746785e-06, 'learning_rate/action_model': 9.53588250746785e-05, 'epoch': 0.42}) +[2026-06-17 18:32:02,530][starVLA.training.train_starvla][INFO] - Step 1080, Loss: {'train/loss': 0.7953627109527588, 'train/grad_norm_pre_clip': 14.724300384521484, 'timing/data': 0.012362012639641762, 'timing/model': 5.5521567622199655, 'learning_rate/qwen_vl_interface': 9.534949668046537e-06, 'learning_rate/action_model': 9.534949668046536e-05, 'epoch': 0.42}) +[2026-06-17 18:32:10,241][starVLA.training.train_starvla][INFO] - Step 1081, Loss: {'train/loss': 1.1853270530700684, 'train/grad_norm_pre_clip': 15.659464836120605, 'timing/data': 0.03387950174510479, 'timing/model': 1.4961393037810922, 'learning_rate/qwen_vl_interface': 9.534015940350545e-06, 'learning_rate/action_model': 9.534015940350544e-05, 'epoch': 0.42}) +[2026-06-17 18:32:14,665][starVLA.training.train_starvla][INFO] - Step 1082, Loss: {'train/loss': 0.615962564945221, 'train/grad_norm_pre_clip': 7.72102689743042, 'timing/data': 0.013148046098649502, 'timing/model': 1.1377316946163774, 'learning_rate/qwen_vl_interface': 9.533081324573437e-06, 'learning_rate/action_model': 9.533081324573437e-05, 'epoch': 0.42}) +[2026-06-17 18:32:22,308][starVLA.training.train_starvla][INFO] - Step 1083, Loss: {'train/loss': 0.4433760344982147, 'train/grad_norm_pre_clip': 9.999942779541016, 'timing/data': 0.014147993177175522, 'timing/model': 1.1908498918637633, 'learning_rate/qwen_vl_interface': 9.532145820908965e-06, 'learning_rate/action_model': 9.532145820908964e-05, 'epoch': 0.42}) +[2026-06-17 18:32:30,097][starVLA.training.train_starvla][INFO] - Step 1084, Loss: {'train/loss': 0.6097357869148254, 'train/grad_norm_pre_clip': 6.385408878326416, 'timing/data': 0.011411457322537899, 'timing/model': 4.060974576510489, 'learning_rate/qwen_vl_interface': 9.531209429551056e-06, 'learning_rate/action_model': 9.531209429551054e-05, 'epoch': 0.42}) +[2026-06-17 18:32:34,750][starVLA.training.train_starvla][INFO] - Step 1085, Loss: {'train/loss': 1.1144745349884033, 'train/grad_norm_pre_clip': 8.595006942749023, 'timing/data': 0.011899899691343307, 'timing/model': 1.0860460195690393, 'learning_rate/qwen_vl_interface': 9.530272150693823e-06, 'learning_rate/action_model': 9.530272150693824e-05, 'epoch': 0.42}) +[2026-06-17 18:32:42,520][starVLA.training.train_starvla][INFO] - Step 1086, Loss: {'train/loss': 0.6550374031066895, 'train/grad_norm_pre_clip': 7.494945526123047, 'timing/data': 0.015085462480783463, 'timing/model': 1.7950935596600175, 'learning_rate/qwen_vl_interface': 9.52933398453157e-06, 'learning_rate/action_model': 9.52933398453157e-05, 'epoch': 0.42}) +[2026-06-17 18:32:47,088][starVLA.training.train_starvla][INFO] - Step 1087, Loss: {'train/loss': 0.706204354763031, 'train/grad_norm_pre_clip': 6.264886856079102, 'timing/data': 0.015599164180457592, 'timing/model': 1.345721465535462, 'learning_rate/qwen_vl_interface': 9.528394931258777e-06, 'learning_rate/action_model': 9.528394931258776e-05, 'epoch': 0.42}) +[2026-06-17 18:32:51,545][starVLA.training.train_starvla][INFO] - Step 1088, Loss: {'train/loss': 0.4606225788593292, 'train/grad_norm_pre_clip': 7.985325813293457, 'timing/data': 0.012370791286230087, 'timing/model': 1.086410210467875, 'learning_rate/qwen_vl_interface': 9.527454991070108e-06, 'learning_rate/action_model': 9.527454991070108e-05, 'epoch': 0.42}) +[2026-06-17 18:32:56,184][starVLA.training.train_starvla][INFO] - Step 1089, Loss: {'train/loss': 0.6963974833488464, 'train/grad_norm_pre_clip': 10.31590461730957, 'timing/data': 0.020239712670445442, 'timing/model': 1.0934401797130704, 'learning_rate/qwen_vl_interface': 9.52651416416042e-06, 'learning_rate/action_model': 9.526514164160421e-05, 'epoch': 0.42}) +[2026-06-17 18:33:00,892][starVLA.training.train_starvla][INFO] - Step 1090, Loss: {'train/loss': 1.1311815977096558, 'train/grad_norm_pre_clip': 9.427282333374023, 'timing/data': 0.010833046399056911, 'timing/model': 1.1708902241662145, 'learning_rate/qwen_vl_interface': 9.525572450724743e-06, 'learning_rate/action_model': 9.525572450724743e-05, 'epoch': 0.42}) +[2026-06-17 18:33:06,483][starVLA.training.train_starvla][INFO] - Step 1091, Loss: {'train/loss': 0.6619450449943542, 'train/grad_norm_pre_clip': 8.114171028137207, 'timing/data': 0.012557970359921455, 'timing/model': 2.0106711629778147, 'learning_rate/qwen_vl_interface': 9.524629850958297e-06, 'learning_rate/action_model': 9.524629850958297e-05, 'epoch': 0.42}) +[2026-06-17 18:33:15,703][starVLA.training.train_starvla][INFO] - Step 1092, Loss: {'train/loss': 0.5788419246673584, 'train/grad_norm_pre_clip': 4.100623607635498, 'timing/data': 0.010705280117690563, 'timing/model': 3.3144214898347855, 'learning_rate/qwen_vl_interface': 9.523686365056482e-06, 'learning_rate/action_model': 9.523686365056481e-05, 'epoch': 0.42}) +[2026-06-17 18:33:20,051][starVLA.training.train_starvla][INFO] - Step 1093, Loss: {'train/loss': 0.4887583553791046, 'train/grad_norm_pre_clip': 4.428664684295654, 'timing/data': 0.010903089307248592, 'timing/model': 1.193537856452167, 'learning_rate/qwen_vl_interface': 9.522741993214886e-06, 'learning_rate/action_model': 9.522741993214886e-05, 'epoch': 0.42}) +[2026-06-17 18:33:24,542][starVLA.training.train_starvla][INFO] - Step 1094, Loss: {'train/loss': 0.6852519512176514, 'train/grad_norm_pre_clip': 4.906423568725586, 'timing/data': 0.012248950079083443, 'timing/model': 1.0708171194419265, 'learning_rate/qwen_vl_interface': 9.521796735629275e-06, 'learning_rate/action_model': 9.521796735629275e-05, 'epoch': 0.42}) +[2026-06-17 18:33:29,134][starVLA.training.train_starvla][INFO] - Step 1095, Loss: {'train/loss': 0.6146484613418579, 'train/grad_norm_pre_clip': 5.437389850616455, 'timing/data': 0.014258336275815964, 'timing/model': 1.1673033135011792, 'learning_rate/qwen_vl_interface': 9.520850592495603e-06, 'learning_rate/action_model': 9.520850592495604e-05, 'epoch': 0.42}) +[2026-06-17 18:33:33,782][starVLA.training.train_starvla][INFO] - Step 1096, Loss: {'train/loss': 0.5718796253204346, 'train/grad_norm_pre_clip': 8.034147262573242, 'timing/data': 0.012494469992816448, 'timing/model': 1.0915421359241009, 'learning_rate/qwen_vl_interface': 9.519903564010012e-06, 'learning_rate/action_model': 9.51990356401001e-05, 'epoch': 0.42}) +[2026-06-17 18:33:39,068][starVLA.training.train_starvla][INFO] - Step 1097, Loss: {'train/loss': 0.9203730821609497, 'train/grad_norm_pre_clip': 6.711265563964844, 'timing/data': 0.01256656739860773, 'timing/model': 1.6957256449386477, 'learning_rate/qwen_vl_interface': 9.518955650368816e-06, 'learning_rate/action_model': 9.518955650368814e-05, 'epoch': 0.42}) +[2026-06-17 18:33:57,298][starVLA.training.train_starvla][INFO] - Step 1098, Loss: {'train/loss': 0.5258504748344421, 'train/grad_norm_pre_clip': 6.861949920654297, 'timing/data': 0.01275292132049799, 'timing/model': 5.497269897721708, 'learning_rate/qwen_vl_interface': 9.51800685176852e-06, 'learning_rate/action_model': 9.518006851768518e-05, 'epoch': 0.42}) +[2026-06-17 18:34:06,160][starVLA.training.train_starvla][INFO] - Step 1099, Loss: {'train/loss': 0.755175769329071, 'train/grad_norm_pre_clip': 7.994923114776611, 'timing/data': 0.020562170073390007, 'timing/model': 1.3767189783975482, 'learning_rate/qwen_vl_interface': 9.51705716840581e-06, 'learning_rate/action_model': 9.517057168405808e-05, 'epoch': 0.42}) +[2026-06-17 18:37:30,923][starVLA.training.train_starvla][INFO] - Step 1100, Loss: {'train/loss': 0.5549768805503845, 'train/grad_norm_pre_clip': 14.026005744934082, 'eval/loss': 0.6242296099662781, 'eval/latency_4/loss': 0.6242296099662781, 'eval/demon_fire_precision': 0.9422661315220747, 'eval/demon_fire_recall': 0.9912309191295875, 'eval/demon_fire_f1': 0.9661285216840773, 'eval/demon_fire_rate_model': 0.9512481644640235, 'eval/demon_fire_rate_teacher': 0.9042584434654919, 'eval/demon_left_precision': 0.7649402390438247, 'eval/demon_left_recall': 0.4383561643835616, 'eval/demon_left_f1': 0.5573294629898403, 'eval/demon_left_rate_model': 0.0737151248164464, 'eval/demon_left_rate_teacher': 0.12863436123348018, 'eval/demon_right_precision': 0.7386138613861386, 'eval/demon_right_recall': 0.6732851985559567, 'eval/demon_right_f1': 0.7044381491973559, 'eval/demon_right_rate_model': 0.14831130690161526, 'eval/demon_right_rate_teacher': 0.16270190895741557, 'eval/demon_move_f1': 0.6308838060935982, 'eval/demon_cc_f1': 0.833288429764056, 'eval/latency_4/demon_fire_precision': 0.9422661315220747, 'eval/latency_4/demon_fire_recall': 0.9912309191295875, 'eval/latency_4/demon_fire_f1': 0.9661285216840773, 'eval/latency_4/demon_fire_rate_model': 0.9512481644640235, 'eval/latency_4/demon_fire_rate_teacher': 0.9042584434654919, 'eval/latency_4/demon_left_precision': 0.7649402390438247, 'eval/latency_4/demon_left_recall': 0.4383561643835616, 'eval/latency_4/demon_left_f1': 0.5573294629898403, 'eval/latency_4/demon_left_rate_model': 0.0737151248164464, 'eval/latency_4/demon_left_rate_teacher': 0.12863436123348018, 'eval/latency_4/demon_right_precision': 0.7386138613861386, 'eval/latency_4/demon_right_recall': 0.6732851985559567, 'eval/latency_4/demon_right_f1': 0.7044381491973559, 'eval/latency_4/demon_right_rate_model': 0.14831130690161526, 'eval/latency_4/demon_right_rate_teacher': 0.16270190895741557, 'eval/latency_4/demon_move_f1': 0.6308838060935982, 'eval/latency_4/demon_cc_f1': 0.833288429764056, 'timing/data': 0.012750492431223392, 'timing/model': 1.23438482824713, 'learning_rate/qwen_vl_interface': 9.516106600477557e-06, 'learning_rate/action_model': 9.516106600477557e-05, 'epoch': 0.42}) +[2026-06-17 18:37:30,963][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_1100_state +[2026-06-17 18:37:30,971][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 18:37:39,629][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_1100_state/model.safetensors +[2026-06-17 18:38:01,913][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_1100_state/optimizer.bin +[2026-06-17 18:38:01,914][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_1100_state/sampler.bin +[2026-06-17 18:38:01,914][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_1100_state/sampler_1.bin +[2026-06-17 18:38:01,917][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_1100_state/random_states_0.pkl +[2026-06-17 18:38:01,922][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1100 +[2026-06-17 18:38:01,922][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 18:38:01,940][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 18:38:13,511][starVLA.training.train_starvla][INFO] - Step 1101, Loss: {'train/loss': 0.341962069272995, 'train/grad_norm_pre_clip': 3.832892417907715, 'timing/data': 0.012152564711868763, 'timing/model': 2.7093522623181343, 'learning_rate/qwen_vl_interface': 9.515155148180816e-06, 'learning_rate/action_model': 9.515155148180816e-05, 'epoch': 0.42}) +[2026-06-17 18:38:18,647][starVLA.training.train_starvla][INFO] - Step 1102, Loss: {'train/loss': 0.6176291704177856, 'train/grad_norm_pre_clip': 5.884969711303711, 'timing/data': 0.011275166645646095, 'timing/model': 1.0825340505689383, 'learning_rate/qwen_vl_interface': 9.514202811712825e-06, 'learning_rate/action_model': 9.514202811712825e-05, 'epoch': 0.42}) +[2026-06-17 18:38:23,733][starVLA.training.train_starvla][INFO] - Step 1103, Loss: {'train/loss': 0.578559935092926, 'train/grad_norm_pre_clip': 6.785403728485107, 'timing/data': 0.012286623008549213, 'timing/model': 1.1075428165495396, 'learning_rate/qwen_vl_interface': 9.513249591271e-06, 'learning_rate/action_model': 9.513249591271e-05, 'epoch': 0.42}) +[2026-06-17 18:38:29,731][starVLA.training.train_starvla][INFO] - Step 1104, Loss: {'train/loss': 0.6836461424827576, 'train/grad_norm_pre_clip': 5.6631035804748535, 'timing/data': 0.014366169460117817, 'timing/model': 1.1129243709146976, 'learning_rate/qwen_vl_interface': 9.512295487052949e-06, 'learning_rate/action_model': 9.512295487052949e-05, 'epoch': 0.42}) +[2026-06-17 18:38:34,465][starVLA.training.train_starvla][INFO] - Step 1105, Loss: {'train/loss': 0.5186769962310791, 'train/grad_norm_pre_clip': 5.684183597564697, 'timing/data': 0.013697012327611446, 'timing/model': 1.1023290762677789, 'learning_rate/qwen_vl_interface': 9.511340499256455e-06, 'learning_rate/action_model': 9.511340499256455e-05, 'epoch': 0.43}) +[2026-06-17 18:38:39,380][starVLA.training.train_starvla][INFO] - Step 1106, Loss: {'train/loss': 0.4105464518070221, 'train/grad_norm_pre_clip': 5.439547538757324, 'timing/data': 0.01373222190886736, 'timing/model': 1.2138836979866028, 'learning_rate/qwen_vl_interface': 9.510384628079493e-06, 'learning_rate/action_model': 9.510384628079493e-05, 'epoch': 0.43}) +[2026-06-17 18:38:44,043][starVLA.training.train_starvla][INFO] - Step 1107, Loss: {'train/loss': 0.5752785801887512, 'train/grad_norm_pre_clip': 7.714157581329346, 'timing/data': 0.013111749663949013, 'timing/model': 1.391015744768083, 'learning_rate/qwen_vl_interface': 9.50942787372021e-06, 'learning_rate/action_model': 9.50942787372021e-05, 'epoch': 0.43}) +[2026-06-17 18:38:49,214][starVLA.training.train_starvla][INFO] - Step 1108, Loss: {'train/loss': 0.6572842597961426, 'train/grad_norm_pre_clip': 8.49495792388916, 'timing/data': 0.011687946505844593, 'timing/model': 1.2312883045524359, 'learning_rate/qwen_vl_interface': 9.508470236376948e-06, 'learning_rate/action_model': 9.508470236376948e-05, 'epoch': 0.43}) +[2026-06-17 18:38:53,895][starVLA.training.train_starvla][INFO] - Step 1109, Loss: {'train/loss': 0.7400968670845032, 'train/grad_norm_pre_clip': 7.983773708343506, 'timing/data': 0.012725982815027237, 'timing/model': 1.0782871106639504, 'learning_rate/qwen_vl_interface': 9.507511716248224e-06, 'learning_rate/action_model': 9.507511716248224e-05, 'epoch': 0.43}) +[2026-06-17 18:38:58,717][starVLA.training.train_starvla][INFO] - Step 1110, Loss: {'train/loss': 0.4041862189769745, 'train/grad_norm_pre_clip': 4.181163787841797, 'timing/data': 0.014299419708549976, 'timing/model': 1.081283618696034, 'learning_rate/qwen_vl_interface': 9.506552313532739e-06, 'learning_rate/action_model': 9.506552313532739e-05, 'epoch': 0.43}) +[2026-06-17 18:39:03,530][starVLA.training.train_starvla][INFO] - Step 1111, Loss: {'train/loss': 0.7996609210968018, 'train/grad_norm_pre_clip': 7.786909103393555, 'timing/data': 0.011491373181343079, 'timing/model': 1.1266488442197442, 'learning_rate/qwen_vl_interface': 9.505592028429377e-06, 'learning_rate/action_model': 9.505592028429377e-05, 'epoch': 0.43}) +[2026-06-17 18:39:07,788][starVLA.training.train_starvla][INFO] - Step 1112, Loss: {'train/loss': 0.740003764629364, 'train/grad_norm_pre_clip': 9.426977157592773, 'timing/data': 0.010914443992078304, 'timing/model': 1.0860619526356459, 'learning_rate/qwen_vl_interface': 9.504630861137212e-06, 'learning_rate/action_model': 9.504630861137212e-05, 'epoch': 0.43}) +[2026-06-17 18:39:12,392][starVLA.training.train_starvla][INFO] - Step 1113, Loss: {'train/loss': 0.7259926199913025, 'train/grad_norm_pre_clip': 11.60402774810791, 'timing/data': 0.015710880979895592, 'timing/model': 1.0955435885116458, 'learning_rate/qwen_vl_interface': 9.503668811855489e-06, 'learning_rate/action_model': 9.503668811855488e-05, 'epoch': 0.43}) +[2026-06-17 18:39:16,969][starVLA.training.train_starvla][INFO] - Step 1114, Loss: {'train/loss': 0.5335854887962341, 'train/grad_norm_pre_clip': 9.20093059539795, 'timing/data': 0.011756882071495056, 'timing/model': 1.0811217576265335, 'learning_rate/qwen_vl_interface': 9.502705880783644e-06, 'learning_rate/action_model': 9.502705880783645e-05, 'epoch': 0.43}) +[2026-06-17 18:39:21,607][starVLA.training.train_starvla][INFO] - Step 1115, Loss: {'train/loss': 0.5646790862083435, 'train/grad_norm_pre_clip': 9.219888687133789, 'timing/data': 0.011360067874193192, 'timing/model': 1.0735383462160826, 'learning_rate/qwen_vl_interface': 9.501742068121296e-06, 'learning_rate/action_model': 9.501742068121296e-05, 'epoch': 0.43}) +[2026-06-17 18:39:26,392][starVLA.training.train_starvla][INFO] - Step 1116, Loss: {'train/loss': 0.5319070816040039, 'train/grad_norm_pre_clip': 5.761951923370361, 'timing/data': 0.014842127449810505, 'timing/model': 1.579724540002644, 'learning_rate/qwen_vl_interface': 9.500777374068241e-06, 'learning_rate/action_model': 9.50077737406824e-05, 'epoch': 0.43}) +[2026-06-17 18:39:30,907][starVLA.training.train_starvla][INFO] - Step 1117, Loss: {'train/loss': 0.6664423942565918, 'train/grad_norm_pre_clip': 6.948282241821289, 'timing/data': 0.01181276235729456, 'timing/model': 1.3469435758888721, 'learning_rate/qwen_vl_interface': 9.49981179882446e-06, 'learning_rate/action_model': 9.499811798824461e-05, 'epoch': 0.43}) +[2026-06-17 18:39:35,090][starVLA.training.train_starvla][INFO] - Step 1118, Loss: {'train/loss': 0.37851062417030334, 'train/grad_norm_pre_clip': 9.338520050048828, 'timing/data': 0.013858701102435589, 'timing/model': 1.0860954727977514, 'learning_rate/qwen_vl_interface': 9.498845342590124e-06, 'learning_rate/action_model': 9.498845342590124e-05, 'epoch': 0.43}) +[2026-06-17 18:39:42,060][starVLA.training.train_starvla][INFO] - Step 1119, Loss: {'train/loss': 0.627724289894104, 'train/grad_norm_pre_clip': 7.386776924133301, 'timing/data': 0.011588189750909805, 'timing/model': 1.1153587372973561, 'learning_rate/qwen_vl_interface': 9.497878005565575e-06, 'learning_rate/action_model': 9.497878005565574e-05, 'epoch': 0.43}) +[2026-06-17 18:39:49,188][starVLA.training.train_starvla][INFO] - Step 1120, Loss: {'train/loss': 0.9381581544876099, 'train/grad_norm_pre_clip': 10.150731086730957, 'timing/data': 0.013271701522171497, 'timing/model': 1.208116652444005, 'learning_rate/qwen_vl_interface': 9.496909787951344e-06, 'learning_rate/action_model': 9.496909787951343e-05, 'epoch': 0.43}) +[2026-06-17 18:39:54,209][starVLA.training.train_starvla][INFO] - Step 1121, Loss: {'train/loss': 0.621600866317749, 'train/grad_norm_pre_clip': 7.025304317474365, 'timing/data': 0.013617434538900852, 'timing/model': 1.153869204223156, 'learning_rate/qwen_vl_interface': 9.495940689948144e-06, 'learning_rate/action_model': 9.495940689948144e-05, 'epoch': 0.43}) +[2026-06-17 18:40:00,226][starVLA.training.train_starvla][INFO] - Step 1122, Loss: {'train/loss': 0.6317599415779114, 'train/grad_norm_pre_clip': 9.204845428466797, 'timing/data': 0.014900311827659607, 'timing/model': 1.4701711582019925, 'learning_rate/qwen_vl_interface': 9.49497071175687e-06, 'learning_rate/action_model': 9.49497071175687e-05, 'epoch': 0.43}) +[2026-06-17 18:40:07,108][starVLA.training.train_starvla][INFO] - Step 1123, Loss: {'train/loss': 0.7835632562637329, 'train/grad_norm_pre_clip': 6.786667823791504, 'timing/data': 0.011250958777964115, 'timing/model': 1.097831537015736, 'learning_rate/qwen_vl_interface': 9.493999853578601e-06, 'learning_rate/action_model': 9.493999853578602e-05, 'epoch': 0.43}) +[2026-06-17 18:40:15,945][starVLA.training.train_starvla][INFO] - Step 1124, Loss: {'train/loss': 0.31610944867134094, 'train/grad_norm_pre_clip': 5.021737575531006, 'timing/data': 0.014279443770647049, 'timing/model': 2.840421081520617, 'learning_rate/qwen_vl_interface': 9.493028115614595e-06, 'learning_rate/action_model': 9.493028115614595e-05, 'epoch': 0.43}) +[2026-06-17 18:40:21,884][starVLA.training.train_starvla][INFO] - Step 1125, Loss: {'train/loss': 0.8349100947380066, 'train/grad_norm_pre_clip': 8.771855354309082, 'timing/data': 0.01490176934748888, 'timing/model': 1.1073295325040817, 'learning_rate/qwen_vl_interface': 9.492055498066295e-06, 'learning_rate/action_model': 9.492055498066294e-05, 'epoch': 0.43}) +[2026-06-17 18:40:26,690][starVLA.training.train_starvla][INFO] - Step 1126, Loss: {'train/loss': 0.41436004638671875, 'train/grad_norm_pre_clip': 8.384307861328125, 'timing/data': 0.01357007771730423, 'timing/model': 1.09507529810071, 'learning_rate/qwen_vl_interface': 9.491082001135326e-06, 'learning_rate/action_model': 9.491082001135325e-05, 'epoch': 0.43}) +[2026-06-17 18:40:31,345][starVLA.training.train_starvla][INFO] - Step 1127, Loss: {'train/loss': 0.47278499603271484, 'train/grad_norm_pre_clip': 9.4630126953125, 'timing/data': 0.012553556822240353, 'timing/model': 1.429703801870346, 'learning_rate/qwen_vl_interface': 9.490107625023492e-06, 'learning_rate/action_model': 9.490107625023492e-05, 'epoch': 0.43}) +[2026-06-17 18:40:35,637][starVLA.training.train_starvla][INFO] - Step 1128, Loss: {'train/loss': 0.577691912651062, 'train/grad_norm_pre_clip': 5.348672389984131, 'timing/data': 0.012157549150288105, 'timing/model': 1.1011418225243688, 'learning_rate/qwen_vl_interface': 9.489132369932788e-06, 'learning_rate/action_model': 9.489132369932788e-05, 'epoch': 0.43}) +[2026-06-17 18:40:40,804][starVLA.training.train_starvla][INFO] - Step 1129, Loss: {'train/loss': 0.7687106728553772, 'train/grad_norm_pre_clip': 7.187137126922607, 'timing/data': 0.01333639770746231, 'timing/model': 1.2379919113591313, 'learning_rate/qwen_vl_interface': 9.488156236065378e-06, 'learning_rate/action_model': 9.488156236065379e-05, 'epoch': 0.43}) +[2026-06-17 18:40:45,882][starVLA.training.train_starvla][INFO] - Step 1130, Loss: {'train/loss': 0.6442385911941528, 'train/grad_norm_pre_clip': 8.18288516998291, 'timing/data': 0.012887071818113327, 'timing/model': 1.1609975770115852, 'learning_rate/qwen_vl_interface': 9.487179223623626e-06, 'learning_rate/action_model': 9.487179223623624e-05, 'epoch': 0.43}) +[2026-06-17 18:40:51,551][starVLA.training.train_starvla][INFO] - Step 1131, Loss: {'train/loss': 0.7571231722831726, 'train/grad_norm_pre_clip': 10.52307415008545, 'timing/data': 0.013524304144084454, 'timing/model': 1.8386619947850704, 'learning_rate/qwen_vl_interface': 9.486201332810058e-06, 'learning_rate/action_model': 9.486201332810057e-05, 'epoch': 0.44}) +[2026-06-17 18:40:56,394][starVLA.training.train_starvla][INFO] - Step 1132, Loss: {'train/loss': 0.34048110246658325, 'train/grad_norm_pre_clip': 5.350935935974121, 'timing/data': 0.012113139033317566, 'timing/model': 1.4680746970698237, 'learning_rate/qwen_vl_interface': 9.485222563827395e-06, 'learning_rate/action_model': 9.485222563827395e-05, 'epoch': 0.44}) +[2026-06-17 18:41:00,991][starVLA.training.train_starvla][INFO] - Step 1133, Loss: {'train/loss': 0.5587116479873657, 'train/grad_norm_pre_clip': 9.30787467956543, 'timing/data': 0.012299938127398491, 'timing/model': 1.2298346543684602, 'learning_rate/qwen_vl_interface': 9.484242916878537e-06, 'learning_rate/action_model': 9.484242916878536e-05, 'epoch': 0.44}) +[2026-06-17 18:41:05,702][starVLA.training.train_starvla][INFO] - Step 1134, Loss: {'train/loss': 0.5966514348983765, 'train/grad_norm_pre_clip': 6.697657585144043, 'timing/data': 0.015704265795648098, 'timing/model': 1.098341609351337, 'learning_rate/qwen_vl_interface': 9.483262392166567e-06, 'learning_rate/action_model': 9.483262392166566e-05, 'epoch': 0.44}) +[2026-06-17 18:41:11,334][starVLA.training.train_starvla][INFO] - Step 1135, Loss: {'train/loss': 0.7187029123306274, 'train/grad_norm_pre_clip': 7.247952461242676, 'timing/data': 0.01171476487070322, 'timing/model': 1.34912970662117, 'learning_rate/qwen_vl_interface': 9.482280989894748e-06, 'learning_rate/action_model': 9.482280989894748e-05, 'epoch': 0.44}) +[2026-06-17 18:41:15,979][starVLA.training.train_starvla][INFO] - Step 1136, Loss: {'train/loss': 0.450556218624115, 'train/grad_norm_pre_clip': 6.886334419250488, 'timing/data': 0.01924868766218424, 'timing/model': 1.2102868426591158, 'learning_rate/qwen_vl_interface': 9.481298710266525e-06, 'learning_rate/action_model': 9.481298710266526e-05, 'epoch': 0.44}) +[2026-06-17 18:41:20,580][starVLA.training.train_starvla][INFO] - Step 1137, Loss: {'train/loss': 0.8164011836051941, 'train/grad_norm_pre_clip': 16.085468292236328, 'timing/data': 0.012507443316280842, 'timing/model': 1.3989580059424043, 'learning_rate/qwen_vl_interface': 9.480315553485527e-06, 'learning_rate/action_model': 9.480315553485527e-05, 'epoch': 0.44}) +[2026-06-17 18:41:25,832][starVLA.training.train_starvla][INFO] - Step 1138, Loss: {'train/loss': 0.9994580745697021, 'train/grad_norm_pre_clip': 8.225630760192871, 'timing/data': 0.013541032560169697, 'timing/model': 1.1094619520008564, 'learning_rate/qwen_vl_interface': 9.479331519755563e-06, 'learning_rate/action_model': 9.479331519755562e-05, 'epoch': 0.44}) +[2026-06-17 18:41:30,514][starVLA.training.train_starvla][INFO] - Step 1139, Loss: {'train/loss': 0.9078220129013062, 'train/grad_norm_pre_clip': 9.223694801330566, 'timing/data': 0.011201285757124424, 'timing/model': 1.2254127748310566, 'learning_rate/qwen_vl_interface': 9.478346609280622e-06, 'learning_rate/action_model': 9.478346609280623e-05, 'epoch': 0.44}) +[2026-06-17 18:41:35,309][starVLA.training.train_starvla][INFO] - Step 1140, Loss: {'train/loss': 0.7792199850082397, 'train/grad_norm_pre_clip': 6.950463771820068, 'timing/data': 0.014445936307311058, 'timing/model': 1.0737028159201145, 'learning_rate/qwen_vl_interface': 9.477360822264883e-06, 'learning_rate/action_model': 9.477360822264882e-05, 'epoch': 0.44}) +[2026-06-17 18:41:39,997][starVLA.training.train_starvla][INFO] - Step 1141, Loss: {'train/loss': 0.3540290296077728, 'train/grad_norm_pre_clip': 5.450395584106445, 'timing/data': 0.01216735690832138, 'timing/model': 1.4404493812471628, 'learning_rate/qwen_vl_interface': 9.476374158912692e-06, 'learning_rate/action_model': 9.476374158912691e-05, 'epoch': 0.44}) +[2026-06-17 18:41:44,612][starVLA.training.train_starvla][INFO] - Step 1142, Loss: {'train/loss': 0.5091819167137146, 'train/grad_norm_pre_clip': 5.11637020111084, 'timing/data': 0.019744892604649067, 'timing/model': 1.4095090553164482, 'learning_rate/qwen_vl_interface': 9.475386619428593e-06, 'learning_rate/action_model': 9.475386619428594e-05, 'epoch': 0.44}) +[2026-06-17 18:41:48,780][starVLA.training.train_starvla][INFO] - Step 1143, Loss: {'train/loss': 0.9960201978683472, 'train/grad_norm_pre_clip': 6.332829475402832, 'timing/data': 0.01185677107423544, 'timing/model': 1.0771282464265823, 'learning_rate/qwen_vl_interface': 9.4743982040173e-06, 'learning_rate/action_model': 9.4743982040173e-05, 'epoch': 0.44}) +[2026-06-17 18:41:53,306][starVLA.training.train_starvla][INFO] - Step 1144, Loss: {'train/loss': 0.879326343536377, 'train/grad_norm_pre_clip': 9.3633451461792, 'timing/data': 0.01142041850835085, 'timing/model': 1.0959103824570775, 'learning_rate/qwen_vl_interface': 9.473408912883714e-06, 'learning_rate/action_model': 9.473408912883714e-05, 'epoch': 0.44}) +[2026-06-17 18:41:57,940][starVLA.training.train_starvla][INFO] - Step 1145, Loss: {'train/loss': 0.5984663963317871, 'train/grad_norm_pre_clip': 5.177375793457031, 'timing/data': 0.010228609666228294, 'timing/model': 1.091373990289867, 'learning_rate/qwen_vl_interface': 9.472418746232916e-06, 'learning_rate/action_model': 9.472418746232916e-05, 'epoch': 0.44}) +[2026-06-17 18:42:02,525][starVLA.training.train_starvla][INFO] - Step 1146, Loss: {'train/loss': 0.43310970067977905, 'train/grad_norm_pre_clip': 6.020083427429199, 'timing/data': 0.011746315285563469, 'timing/model': 1.0749339563772082, 'learning_rate/qwen_vl_interface': 9.471427704270168e-06, 'learning_rate/action_model': 9.471427704270169e-05, 'epoch': 0.44}) +[2026-06-17 18:42:08,823][starVLA.training.train_starvla][INFO] - Step 1147, Loss: {'train/loss': 0.3281271755695343, 'train/grad_norm_pre_clip': 4.659449100494385, 'timing/data': 0.012939438223838806, 'timing/model': 1.4996744189411402, 'learning_rate/qwen_vl_interface': 9.470435787200914e-06, 'learning_rate/action_model': 9.470435787200915e-05, 'epoch': 0.44}) +[2026-06-17 18:42:13,357][starVLA.training.train_starvla][INFO] - Step 1148, Loss: {'train/loss': 0.7456772327423096, 'train/grad_norm_pre_clip': 9.649398803710938, 'timing/data': 0.012698800303041935, 'timing/model': 1.2885541263967752, 'learning_rate/qwen_vl_interface': 9.469442995230781e-06, 'learning_rate/action_model': 9.469442995230782e-05, 'epoch': 0.44}) +[2026-06-17 18:42:21,222][starVLA.training.train_starvla][INFO] - Step 1149, Loss: {'train/loss': 0.711686372756958, 'train/grad_norm_pre_clip': 8.052372932434082, 'timing/data': 0.021505442447960377, 'timing/model': 1.1783161060884595, 'learning_rate/qwen_vl_interface': 9.468449328565574e-06, 'learning_rate/action_model': 9.468449328565573e-05, 'epoch': 0.44}) +[2026-06-17 18:42:26,368][starVLA.training.train_starvla][INFO] - Step 1150, Loss: {'train/loss': 0.5204783082008362, 'train/grad_norm_pre_clip': 5.886702060699463, 'timing/data': 0.012177696451544762, 'timing/model': 1.3620559088885784, 'learning_rate/qwen_vl_interface': 9.467454787411282e-06, 'learning_rate/action_model': 9.467454787411281e-05, 'epoch': 0.44}) +[2026-06-17 18:42:34,570][starVLA.training.train_starvla][INFO] - Step 1151, Loss: {'train/loss': 0.6422215700149536, 'train/grad_norm_pre_clip': 5.940345287322998, 'timing/data': 0.014238324016332626, 'timing/model': 1.1071460489183664, 'learning_rate/qwen_vl_interface': 9.466459371974076e-06, 'learning_rate/action_model': 9.466459371974076e-05, 'epoch': 0.44}) +[2026-06-17 18:42:39,898][starVLA.training.train_starvla][INFO] - Step 1152, Loss: {'train/loss': 0.5495525002479553, 'train/grad_norm_pre_clip': 7.617327690124512, 'timing/data': 0.010459071956574917, 'timing/model': 1.6125189373269677, 'learning_rate/qwen_vl_interface': 9.465463082460304e-06, 'learning_rate/action_model': 9.465463082460304e-05, 'epoch': 0.44}) +[2026-06-17 18:42:50,850][starVLA.training.train_starvla][INFO] - Step 1153, Loss: {'train/loss': 0.3988204300403595, 'train/grad_norm_pre_clip': 5.449570655822754, 'timing/data': 0.01660686358809471, 'timing/model': 1.6334069101139903, 'learning_rate/qwen_vl_interface': 9.4644659190765e-06, 'learning_rate/action_model': 9.464465919076499e-05, 'epoch': 0.44}) +[2026-06-17 18:42:55,680][starVLA.training.train_starvla][INFO] - Step 1154, Loss: {'train/loss': 0.7791342735290527, 'train/grad_norm_pre_clip': 9.154224395751953, 'timing/data': 0.01339933555573225, 'timing/model': 1.0899687400087714, 'learning_rate/qwen_vl_interface': 9.463467882029374e-06, 'learning_rate/action_model': 9.463467882029373e-05, 'epoch': 0.44}) +[2026-06-17 18:43:00,675][starVLA.training.train_starvla][INFO] - Step 1155, Loss: {'train/loss': 0.46933794021606445, 'train/grad_norm_pre_clip': 6.99783992767334, 'timing/data': 0.013331740163266659, 'timing/model': 1.373192135244608, 'learning_rate/qwen_vl_interface': 9.462468971525822e-06, 'learning_rate/action_model': 9.462468971525821e-05, 'epoch': 0.44}) +[2026-06-17 18:43:05,449][starVLA.training.train_starvla][INFO] - Step 1156, Loss: {'train/loss': 0.46743202209472656, 'train/grad_norm_pre_clip': 5.242835998535156, 'timing/data': 0.013763703405857086, 'timing/model': 1.0984405670315027, 'learning_rate/qwen_vl_interface': 9.461469187772921e-06, 'learning_rate/action_model': 9.46146918777292e-05, 'epoch': 0.44}) +[2026-06-17 18:43:10,337][starVLA.training.train_starvla][INFO] - Step 1157, Loss: {'train/loss': 0.672291100025177, 'train/grad_norm_pre_clip': 8.137673377990723, 'timing/data': 0.01315706130117178, 'timing/model': 1.5428122514858842, 'learning_rate/qwen_vl_interface': 9.460468530977924e-06, 'learning_rate/action_model': 9.460468530977924e-05, 'epoch': 0.45}) +[2026-06-17 18:43:14,668][starVLA.training.train_starvla][INFO] - Step 1158, Loss: {'train/loss': 0.4670919179916382, 'train/grad_norm_pre_clip': 6.641620635986328, 'timing/data': 0.013146478682756424, 'timing/model': 1.1558591146022081, 'learning_rate/qwen_vl_interface': 9.459467001348271e-06, 'learning_rate/action_model': 9.45946700134827e-05, 'epoch': 0.45}) +[2026-06-17 18:43:20,758][starVLA.training.train_starvla][INFO] - Step 1159, Loss: {'train/loss': 0.5940799713134766, 'train/grad_norm_pre_clip': 7.663870811462402, 'timing/data': 0.014799928292632103, 'timing/model': 1.363102063536644, 'learning_rate/qwen_vl_interface': 9.458464599091578e-06, 'learning_rate/action_model': 9.458464599091578e-05, 'epoch': 0.45}) +[2026-06-17 18:43:25,792][starVLA.training.train_starvla][INFO] - Step 1160, Loss: {'train/loss': 0.2387283891439438, 'train/grad_norm_pre_clip': 4.068839073181152, 'timing/data': 0.010364196263253689, 'timing/model': 1.1977313878014684, 'learning_rate/qwen_vl_interface': 9.457461324415647e-06, 'learning_rate/action_model': 9.457461324415646e-05, 'epoch': 0.45}) +[2026-06-17 18:43:34,424][starVLA.training.train_starvla][INFO] - Step 1161, Loss: {'train/loss': 0.558138370513916, 'train/grad_norm_pre_clip': 9.33533763885498, 'timing/data': 0.013348390348255634, 'timing/model': 1.2087544342502952, 'learning_rate/qwen_vl_interface': 9.456457177528454e-06, 'learning_rate/action_model': 9.456457177528453e-05, 'epoch': 0.45}) +[2026-06-17 18:43:39,557][starVLA.training.train_starvla][INFO] - Step 1162, Loss: {'train/loss': 0.5853140354156494, 'train/grad_norm_pre_clip': 7.108277797698975, 'timing/data': 0.01327636931091547, 'timing/model': 1.6286343391984701, 'learning_rate/qwen_vl_interface': 9.455452158638163e-06, 'learning_rate/action_model': 9.455452158638163e-05, 'epoch': 0.45}) +[2026-06-17 18:43:44,055][starVLA.training.train_starvla][INFO] - Step 1163, Loss: {'train/loss': 0.792836606502533, 'train/grad_norm_pre_clip': 8.502925872802734, 'timing/data': 0.012201626785099506, 'timing/model': 1.182153950445354, 'learning_rate/qwen_vl_interface': 9.454446267953114e-06, 'learning_rate/action_model': 9.454446267953114e-05, 'epoch': 0.45}) +[2026-06-17 18:43:52,723][starVLA.training.train_starvla][INFO] - Step 1164, Loss: {'train/loss': 0.8710387349128723, 'train/grad_norm_pre_clip': 9.670433044433594, 'timing/data': 0.015025070868432522, 'timing/model': 4.759415762498975, 'learning_rate/qwen_vl_interface': 9.453439505681829e-06, 'learning_rate/action_model': 9.453439505681829e-05, 'epoch': 0.45}) +[2026-06-17 18:43:59,838][starVLA.training.train_starvla][INFO] - Step 1165, Loss: {'train/loss': 0.48895183205604553, 'train/grad_norm_pre_clip': 6.586193561553955, 'timing/data': 0.013204718008637428, 'timing/model': 2.6346019003540277, 'learning_rate/qwen_vl_interface': 9.452431872033012e-06, 'learning_rate/action_model': 9.452431872033012e-05, 'epoch': 0.45}) +[2026-06-17 18:44:04,724][starVLA.training.train_starvla][INFO] - Step 1166, Loss: {'train/loss': 0.36400312185287476, 'train/grad_norm_pre_clip': 6.795571327209473, 'timing/data': 0.0147247901186347, 'timing/model': 1.1163410060107708, 'learning_rate/qwen_vl_interface': 9.451423367215546e-06, 'learning_rate/action_model': 9.451423367215545e-05, 'epoch': 0.45}) +[2026-06-17 18:44:09,197][starVLA.training.train_starvla][INFO] - Step 1167, Loss: {'train/loss': 0.39152148365974426, 'train/grad_norm_pre_clip': 5.877120494842529, 'timing/data': 0.014588923193514347, 'timing/model': 1.1155562279745936, 'learning_rate/qwen_vl_interface': 9.450413991438494e-06, 'learning_rate/action_model': 9.450413991438494e-05, 'epoch': 0.45}) +[2026-06-17 18:44:14,181][starVLA.training.train_starvla][INFO] - Step 1168, Loss: {'train/loss': 0.7717544436454773, 'train/grad_norm_pre_clip': 13.132340431213379, 'timing/data': 0.014916040003299713, 'timing/model': 1.0825999518856406, 'learning_rate/qwen_vl_interface': 9.449403744911104e-06, 'learning_rate/action_model': 9.449403744911103e-05, 'epoch': 0.45}) +[2026-06-17 18:44:19,040][starVLA.training.train_starvla][INFO] - Step 1169, Loss: {'train/loss': 0.5268280506134033, 'train/grad_norm_pre_clip': 7.855398178100586, 'timing/data': 0.011204302310943604, 'timing/model': 1.2620650241151452, 'learning_rate/qwen_vl_interface': 9.448392627842797e-06, 'learning_rate/action_model': 9.448392627842795e-05, 'epoch': 0.45}) +[2026-06-17 18:44:23,819][starVLA.training.train_starvla][INFO] - Step 1170, Loss: {'train/loss': 0.5150704383850098, 'train/grad_norm_pre_clip': 8.367121696472168, 'timing/data': 0.01573199313133955, 'timing/model': 1.096932609565556, 'learning_rate/qwen_vl_interface': 9.44738064044318e-06, 'learning_rate/action_model': 9.44738064044318e-05, 'epoch': 0.45}) +[2026-06-17 18:44:28,918][starVLA.training.train_starvla][INFO] - Step 1171, Loss: {'train/loss': 0.5090776085853577, 'train/grad_norm_pre_clip': 7.211300849914551, 'timing/data': 0.011397638358175755, 'timing/model': 1.617484424263239, 'learning_rate/qwen_vl_interface': 9.44636778292204e-06, 'learning_rate/action_model': 9.44636778292204e-05, 'epoch': 0.45}) +[2026-06-17 18:44:33,895][starVLA.training.train_starvla][INFO] - Step 1172, Loss: {'train/loss': 0.724645733833313, 'train/grad_norm_pre_clip': 6.513301372528076, 'timing/data': 0.018459449522197247, 'timing/model': 1.441761402413249, 'learning_rate/qwen_vl_interface': 9.445354055489345e-06, 'learning_rate/action_model': 9.445354055489345e-05, 'epoch': 0.45}) +[2026-06-17 18:44:38,571][starVLA.training.train_starvla][INFO] - Step 1173, Loss: {'train/loss': 0.5065886378288269, 'train/grad_norm_pre_clip': 6.2949113845825195, 'timing/data': 0.012960397638380527, 'timing/model': 1.3887845305725932, 'learning_rate/qwen_vl_interface': 9.444339458355239e-06, 'learning_rate/action_model': 9.444339458355239e-05, 'epoch': 0.45}) +[2026-06-17 18:44:43,614][starVLA.training.train_starvla][INFO] - Step 1174, Loss: {'train/loss': 0.6250756978988647, 'train/grad_norm_pre_clip': 5.859904766082764, 'timing/data': 0.010965283028781414, 'timing/model': 1.1855429159477353, 'learning_rate/qwen_vl_interface': 9.443323991730049e-06, 'learning_rate/action_model': 9.443323991730049e-05, 'epoch': 0.45}) +[2026-06-17 18:44:48,799][starVLA.training.train_starvla][INFO] - Step 1175, Loss: {'train/loss': 0.6202192306518555, 'train/grad_norm_pre_clip': 9.276092529296875, 'timing/data': 0.013130340725183487, 'timing/model': 1.0894661387428641, 'learning_rate/qwen_vl_interface': 9.442307655824284e-06, 'learning_rate/action_model': 9.442307655824285e-05, 'epoch': 0.45}) +[2026-06-17 18:44:53,919][starVLA.training.train_starvla][INFO] - Step 1176, Loss: {'train/loss': 0.7997599244117737, 'train/grad_norm_pre_clip': 8.421846389770508, 'timing/data': 0.013738781213760376, 'timing/model': 1.1199080664664507, 'learning_rate/qwen_vl_interface': 9.441290450848631e-06, 'learning_rate/action_model': 9.44129045084863e-05, 'epoch': 0.45}) +[2026-06-17 18:44:58,746][starVLA.training.train_starvla][INFO] - Step 1177, Loss: {'train/loss': 0.5750458240509033, 'train/grad_norm_pre_clip': 10.072357177734375, 'timing/data': 0.012622040696442127, 'timing/model': 1.4055615542456508, 'learning_rate/qwen_vl_interface': 9.440272377013957e-06, 'learning_rate/action_model': 9.440272377013956e-05, 'epoch': 0.45}) +[2026-06-17 18:45:03,370][starVLA.training.train_starvla][INFO] - Step 1178, Loss: {'train/loss': 0.7427041530609131, 'train/grad_norm_pre_clip': 12.800829887390137, 'timing/data': 0.016328523866832256, 'timing/model': 1.206615786999464, 'learning_rate/qwen_vl_interface': 9.43925343453131e-06, 'learning_rate/action_model': 9.43925343453131e-05, 'epoch': 0.45}) +[2026-06-17 18:45:11,457][starVLA.training.train_starvla][INFO] - Step 1179, Loss: {'train/loss': 0.7229030728340149, 'train/grad_norm_pre_clip': 12.297330856323242, 'timing/data': 0.012145414017140865, 'timing/model': 4.468631428666413, 'learning_rate/qwen_vl_interface': 9.438233623611918e-06, 'learning_rate/action_model': 9.438233623611918e-05, 'epoch': 0.45}) +[2026-06-17 18:45:19,454][starVLA.training.train_starvla][INFO] - Step 1180, Loss: {'train/loss': 0.9124541878700256, 'train/grad_norm_pre_clip': 11.554621696472168, 'timing/data': 0.012750794179737568, 'timing/model': 1.162566357292235, 'learning_rate/qwen_vl_interface': 9.43721294446719e-06, 'learning_rate/action_model': 9.437212944467191e-05, 'epoch': 0.45}) +[2026-06-17 18:45:24,086][starVLA.training.train_starvla][INFO] - Step 1181, Loss: {'train/loss': 0.4609382748603821, 'train/grad_norm_pre_clip': 7.344397068023682, 'timing/data': 0.011545870453119278, 'timing/model': 1.1626530541107059, 'learning_rate/qwen_vl_interface': 9.436191397308713e-06, 'learning_rate/action_model': 9.436191397308713e-05, 'epoch': 0.45}) +[2026-06-17 18:45:29,039][starVLA.training.train_starvla][INFO] - Step 1182, Loss: {'train/loss': 0.6636401414871216, 'train/grad_norm_pre_clip': 8.409266471862793, 'timing/data': 0.012482278048992157, 'timing/model': 1.4069040743634105, 'learning_rate/qwen_vl_interface': 9.435168982348254e-06, 'learning_rate/action_model': 9.435168982348255e-05, 'epoch': 0.45}) +[2026-06-17 18:45:38,342][starVLA.training.train_starvla][INFO] - Step 1183, Loss: {'train/loss': 1.0019372701644897, 'train/grad_norm_pre_clip': 10.790468215942383, 'timing/data': 0.014289040118455887, 'timing/model': 1.5385896265506744, 'learning_rate/qwen_vl_interface': 9.434145699797763e-06, 'learning_rate/action_model': 9.434145699797764e-05, 'epoch': 0.46}) +[2026-06-17 18:45:42,911][starVLA.training.train_starvla][INFO] - Step 1184, Loss: {'train/loss': 0.8311780691146851, 'train/grad_norm_pre_clip': 8.653148651123047, 'timing/data': 0.010508217848837376, 'timing/model': 1.121981636621058, 'learning_rate/qwen_vl_interface': 9.433121549869365e-06, 'learning_rate/action_model': 9.433121549869365e-05, 'epoch': 0.46}) +[2026-06-17 18:45:50,995][starVLA.training.train_starvla][INFO] - Step 1185, Loss: {'train/loss': 0.8922990560531616, 'train/grad_norm_pre_clip': 8.10561466217041, 'timing/data': 0.01317632757127285, 'timing/model': 1.0735438177362084, 'learning_rate/qwen_vl_interface': 9.43209653277537e-06, 'learning_rate/action_model': 9.432096532775369e-05, 'epoch': 0.46}) +[2026-06-17 18:45:56,002][starVLA.training.train_starvla][INFO] - Step 1186, Loss: {'train/loss': 0.6786877512931824, 'train/grad_norm_pre_clip': 11.497334480285645, 'timing/data': 0.012101884000003338, 'timing/model': 1.5895341914147139, 'learning_rate/qwen_vl_interface': 9.431070648728262e-06, 'learning_rate/action_model': 9.431070648728261e-05, 'epoch': 0.46}) +[2026-06-17 18:46:08,970][starVLA.training.train_starvla][INFO] - Step 1187, Loss: {'train/loss': 0.5216777920722961, 'train/grad_norm_pre_clip': 7.537672996520996, 'timing/data': 0.014023381285369396, 'timing/model': 1.4254570016637444, 'learning_rate/qwen_vl_interface': 9.430043897940709e-06, 'learning_rate/action_model': 9.43004389794071e-05, 'epoch': 0.46}) +[2026-06-17 18:46:13,421][starVLA.training.train_starvla][INFO] - Step 1188, Loss: {'train/loss': 0.39656010270118713, 'train/grad_norm_pre_clip': 8.523700714111328, 'timing/data': 0.012341072782874107, 'timing/model': 1.0937952101230621, 'learning_rate/qwen_vl_interface': 9.429016280625559e-06, 'learning_rate/action_model': 9.429016280625559e-05, 'epoch': 0.46}) +[2026-06-17 18:46:18,313][starVLA.training.train_starvla][INFO] - Step 1189, Loss: {'train/loss': 0.6213213801383972, 'train/grad_norm_pre_clip': 8.169869422912598, 'timing/data': 0.01591378916054964, 'timing/model': 1.0852470295503736, 'learning_rate/qwen_vl_interface': 9.427987796995838e-06, 'learning_rate/action_model': 9.427987796995837e-05, 'epoch': 0.46}) +[2026-06-17 18:46:23,236][starVLA.training.train_starvla][INFO] - Step 1190, Loss: {'train/loss': 0.6865094900131226, 'train/grad_norm_pre_clip': 7.171679496765137, 'timing/data': 0.012055015191435814, 'timing/model': 1.1401350796222687, 'learning_rate/qwen_vl_interface': 9.42695844726475e-06, 'learning_rate/action_model': 9.426958447264749e-05, 'epoch': 0.46}) +[2026-06-17 18:46:32,457][starVLA.training.train_starvla][INFO] - Step 1191, Loss: {'train/loss': 0.2774219810962677, 'train/grad_norm_pre_clip': 5.286283493041992, 'timing/data': 0.01533141266554594, 'timing/model': 5.522054524160922, 'learning_rate/qwen_vl_interface': 9.42592823164568e-06, 'learning_rate/action_model': 9.42592823164568e-05, 'epoch': 0.46}) +[2026-06-17 18:46:41,560][starVLA.training.train_starvla][INFO] - Step 1192, Loss: {'train/loss': 0.5575509071350098, 'train/grad_norm_pre_clip': 5.908555507659912, 'timing/data': 0.012895644642412663, 'timing/model': 5.524411810562015, 'learning_rate/qwen_vl_interface': 9.424897150352197e-06, 'learning_rate/action_model': 9.424897150352197e-05, 'epoch': 0.46}) +[2026-06-17 18:46:46,015][starVLA.training.train_starvla][INFO] - Step 1193, Loss: {'train/loss': 0.7904669046401978, 'train/grad_norm_pre_clip': 10.905622482299805, 'timing/data': 0.011856642551720142, 'timing/model': 1.0738755706697702, 'learning_rate/qwen_vl_interface': 9.42386520359804e-06, 'learning_rate/action_model': 9.42386520359804e-05, 'epoch': 0.46}) +[2026-06-17 18:46:53,394][starVLA.training.train_starvla][INFO] - Step 1194, Loss: {'train/loss': 0.5722953081130981, 'train/grad_norm_pre_clip': 6.2048773765563965, 'timing/data': 0.01292184367775917, 'timing/model': 1.1116164792329073, 'learning_rate/qwen_vl_interface': 9.422832391597136e-06, 'learning_rate/action_model': 9.422832391597136e-05, 'epoch': 0.46}) +[2026-06-17 18:46:59,696][starVLA.training.train_starvla][INFO] - Step 1195, Loss: {'train/loss': 0.7618853449821472, 'train/grad_norm_pre_clip': 11.24315357208252, 'timing/data': 0.02155811432749033, 'timing/model': 1.0747270258143544, 'learning_rate/qwen_vl_interface': 9.421798714563586e-06, 'learning_rate/action_model': 9.421798714563586e-05, 'epoch': 0.46}) +[2026-06-17 18:47:07,726][starVLA.training.train_starvla][INFO] - Step 1196, Loss: {'train/loss': 0.8195365071296692, 'train/grad_norm_pre_clip': 12.69567584991455, 'timing/data': 0.011869083158671856, 'timing/model': 1.0927332723513246, 'learning_rate/qwen_vl_interface': 9.420764172711676e-06, 'learning_rate/action_model': 9.420764172711676e-05, 'epoch': 0.46}) +[2026-06-17 18:47:12,325][starVLA.training.train_starvla][INFO] - Step 1197, Loss: {'train/loss': 0.5492038130760193, 'train/grad_norm_pre_clip': 7.912779331207275, 'timing/data': 0.011674962006509304, 'timing/model': 1.3864311845973134, 'learning_rate/qwen_vl_interface': 9.419728766255863e-06, 'learning_rate/action_model': 9.419728766255862e-05, 'epoch': 0.46}) +[2026-06-17 18:47:16,651][starVLA.training.train_starvla][INFO] - Step 1198, Loss: {'train/loss': 0.5748847723007202, 'train/grad_norm_pre_clip': 5.786643981933594, 'timing/data': 0.012738918885588646, 'timing/model': 1.1060356078669429, 'learning_rate/qwen_vl_interface': 9.418692495410789e-06, 'learning_rate/action_model': 9.418692495410789e-05, 'epoch': 0.46}) +[2026-06-17 18:47:21,764][starVLA.training.train_starvla][INFO] - Step 1199, Loss: {'train/loss': 0.5863447785377502, 'train/grad_norm_pre_clip': 6.883009433746338, 'timing/data': 0.010758140124380589, 'timing/model': 1.2650066912174225, 'learning_rate/qwen_vl_interface': 9.417655360391277e-06, 'learning_rate/action_model': 9.417655360391277e-05, 'epoch': 0.46}) +[2026-06-17 18:51:14,778][starVLA.training.train_starvla][INFO] - Step 1200, Loss: {'train/loss': 0.7547833919525146, 'train/grad_norm_pre_clip': 8.19936752319336, 'eval/loss': 0.6315951347351074, 'eval/latency_4/loss': 0.6315951347351074, 'eval/demon_fire_precision': 0.9429716399506781, 'eval/demon_fire_recall': 0.9935043845404352, 'eval/demon_fire_f1': 0.9675786810058515, 'eval/demon_fire_rate_model': 0.9527165932452276, 'eval/demon_fire_rate_teacher': 0.9042584434654919, 'eval/demon_left_precision': 0.6383495145631068, 'eval/demon_left_recall': 0.6004566210045662, 'eval/demon_left_f1': 0.6188235294117648, 'eval/demon_left_rate_model': 0.1209985315712188, 'eval/demon_left_rate_teacher': 0.12863436123348018, 'eval/demon_right_precision': 0.816711590296496, 'eval/demon_right_recall': 0.5469314079422383, 'eval/demon_right_f1': 0.6551351351351351, 'eval/demon_right_rate_model': 0.10895741556534508, 'eval/demon_right_rate_teacher': 0.16270190895741557, 'eval/demon_move_f1': 0.6369793322734499, 'eval/demon_cc_f1': 0.8372206378504444, 'eval/latency_4/demon_fire_precision': 0.9429716399506781, 'eval/latency_4/demon_fire_recall': 0.9935043845404352, 'eval/latency_4/demon_fire_f1': 0.9675786810058515, 'eval/latency_4/demon_fire_rate_model': 0.9527165932452276, 'eval/latency_4/demon_fire_rate_teacher': 0.9042584434654919, 'eval/latency_4/demon_left_precision': 0.6383495145631068, 'eval/latency_4/demon_left_recall': 0.6004566210045662, 'eval/latency_4/demon_left_f1': 0.6188235294117648, 'eval/latency_4/demon_left_rate_model': 0.1209985315712188, 'eval/latency_4/demon_left_rate_teacher': 0.12863436123348018, 'eval/latency_4/demon_right_precision': 0.816711590296496, 'eval/latency_4/demon_right_recall': 0.5469314079422383, 'eval/latency_4/demon_right_f1': 0.6551351351351351, 'eval/latency_4/demon_right_rate_model': 0.10895741556534508, 'eval/latency_4/demon_right_rate_teacher': 0.16270190895741557, 'eval/latency_4/demon_move_f1': 0.6369793322734499, 'eval/latency_4/demon_cc_f1': 0.8372206378504444, 'timing/data': 0.01547390315681696, 'timing/model': 1.256312252022326, 'learning_rate/qwen_vl_interface': 9.416617361412325e-06, 'learning_rate/action_model': 9.416617361412325e-05, 'epoch': 0.46}) +[2026-06-17 18:51:14,800][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_1200_state +[2026-06-17 18:51:14,808][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 18:51:23,770][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_1200_state/model.safetensors +[2026-06-17 18:51:45,792][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_1200_state/optimizer.bin +[2026-06-17 18:51:45,793][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_1200_state/sampler.bin +[2026-06-17 18:51:45,794][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_1200_state/sampler_1.bin +[2026-06-17 18:51:45,797][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_1200_state/random_states_0.pkl +[2026-06-17 18:51:45,872][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1200 +[2026-06-17 18:51:45,872][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 18:51:45,891][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 18:51:50,471][starVLA.training.train_starvla][INFO] - Step 1201, Loss: {'train/loss': 0.7216119766235352, 'train/grad_norm_pre_clip': 7.807584762573242, 'timing/data': 0.011531366035342216, 'timing/model': 1.062897378578782, 'learning_rate/qwen_vl_interface': 9.415578498689106e-06, 'learning_rate/action_model': 9.415578498689106e-05, 'epoch': 0.46}) +[2026-06-17 18:51:54,955][starVLA.training.train_starvla][INFO] - Step 1202, Loss: {'train/loss': 0.44062724709510803, 'train/grad_norm_pre_clip': 7.69755744934082, 'timing/data': 0.011364758014678955, 'timing/model': 1.3656070800498128, 'learning_rate/qwen_vl_interface': 9.414538772436987e-06, 'learning_rate/action_model': 9.414538772436986e-05, 'epoch': 0.46}) +[2026-06-17 18:51:59,298][starVLA.training.train_starvla][INFO] - Step 1203, Loss: {'train/loss': 0.577555239200592, 'train/grad_norm_pre_clip': 7.077802658081055, 'timing/data': 0.01401794608682394, 'timing/model': 1.0810502050444484, 'learning_rate/qwen_vl_interface': 9.413498182871496e-06, 'learning_rate/action_model': 9.413498182871496e-05, 'epoch': 0.46}) +[2026-06-17 18:52:04,083][starVLA.training.train_starvla][INFO] - Step 1204, Loss: {'train/loss': 0.5123233199119568, 'train/grad_norm_pre_clip': 6.98738956451416, 'timing/data': 0.012527799233794212, 'timing/model': 1.1534437472000718, 'learning_rate/qwen_vl_interface': 9.412456730208352e-06, 'learning_rate/action_model': 9.412456730208352e-05, 'epoch': 0.46}) +[2026-06-17 18:52:08,811][starVLA.training.train_starvla][INFO] - Step 1205, Loss: {'train/loss': 0.5136757493019104, 'train/grad_norm_pre_clip': 7.197671890258789, 'timing/data': 0.014798390679061413, 'timing/model': 1.1674918867647648, 'learning_rate/qwen_vl_interface': 9.41141441466345e-06, 'learning_rate/action_model': 9.41141441466345e-05, 'epoch': 0.46}) +[2026-06-17 18:52:13,397][starVLA.training.train_starvla][INFO] - Step 1206, Loss: {'train/loss': 0.6976801156997681, 'train/grad_norm_pre_clip': 8.005841255187988, 'timing/data': 0.016017111018300056, 'timing/model': 1.446535455994308, 'learning_rate/qwen_vl_interface': 9.41037123645286e-06, 'learning_rate/action_model': 9.41037123645286e-05, 'epoch': 0.46}) +[2026-06-17 18:52:17,739][starVLA.training.train_starvla][INFO] - Step 1207, Loss: {'train/loss': 0.4643115997314453, 'train/grad_norm_pre_clip': 5.445949077606201, 'timing/data': 0.01161189004778862, 'timing/model': 1.1318381419405341, 'learning_rate/qwen_vl_interface': 9.409327195792838e-06, 'learning_rate/action_model': 9.409327195792838e-05, 'epoch': 0.46}) +[2026-06-17 18:52:22,602][starVLA.training.train_starvla][INFO] - Step 1208, Loss: {'train/loss': 0.5674233436584473, 'train/grad_norm_pre_clip': 6.258009910583496, 'timing/data': 0.011933980509638786, 'timing/model': 1.2202881444245577, 'learning_rate/qwen_vl_interface': 9.408282292899812e-06, 'learning_rate/action_model': 9.408282292899812e-05, 'epoch': 0.46}) +[2026-06-17 18:52:27,358][starVLA.training.train_starvla][INFO] - Step 1209, Loss: {'train/loss': 0.32625097036361694, 'train/grad_norm_pre_clip': 5.447458267211914, 'timing/data': 0.013776655308902264, 'timing/model': 1.0679153893142939, 'learning_rate/qwen_vl_interface': 9.407236527990394e-06, 'learning_rate/action_model': 9.407236527990393e-05, 'epoch': 0.47}) +[2026-06-17 18:52:33,431][starVLA.training.train_starvla][INFO] - Step 1210, Loss: {'train/loss': 0.586270272731781, 'train/grad_norm_pre_clip': 10.091076850891113, 'timing/data': 0.011396857909858227, 'timing/model': 2.4860476013273, 'learning_rate/qwen_vl_interface': 9.406189901281369e-06, 'learning_rate/action_model': 9.406189901281368e-05, 'epoch': 0.47}) +[2026-06-17 18:52:39,442][starVLA.training.train_starvla][INFO] - Step 1211, Loss: {'train/loss': 0.8636982440948486, 'train/grad_norm_pre_clip': 7.3914008140563965, 'timing/data': 0.012714642100036144, 'timing/model': 1.5773611655458808, 'learning_rate/qwen_vl_interface': 9.405142412989705e-06, 'learning_rate/action_model': 9.405142412989704e-05, 'epoch': 0.47}) +[2026-06-17 18:52:47,991][starVLA.training.train_starvla][INFO] - Step 1212, Loss: {'train/loss': 0.8171229958534241, 'train/grad_norm_pre_clip': 10.005463600158691, 'timing/data': 0.016207881271839142, 'timing/model': 1.5977903213351965, 'learning_rate/qwen_vl_interface': 9.404094063332546e-06, 'learning_rate/action_model': 9.404094063332546e-05, 'epoch': 0.47}) +[2026-06-17 18:52:52,344][starVLA.training.train_starvla][INFO] - Step 1213, Loss: {'train/loss': 0.5644659399986267, 'train/grad_norm_pre_clip': 9.050060272216797, 'timing/data': 0.013331795111298561, 'timing/model': 1.0764187285676599, 'learning_rate/qwen_vl_interface': 9.40304485252722e-06, 'learning_rate/action_model': 9.403044852527219e-05, 'epoch': 0.47}) +[2026-06-17 18:53:01,756][starVLA.training.train_starvla][INFO] - Step 1214, Loss: {'train/loss': 0.3632197380065918, 'train/grad_norm_pre_clip': 9.2871732711792, 'timing/data': 0.01179181132465601, 'timing/model': 4.156202496029437, 'learning_rate/qwen_vl_interface': 9.401994780791226e-06, 'learning_rate/action_model': 9.401994780791226e-05, 'epoch': 0.47}) +[2026-06-17 18:53:07,355][starVLA.training.train_starvla][INFO] - Step 1215, Loss: {'train/loss': 0.47867798805236816, 'train/grad_norm_pre_clip': 8.86440658569336, 'timing/data': 0.012884057126939297, 'timing/model': 1.567049560137093, 'learning_rate/qwen_vl_interface': 9.400943848342246e-06, 'learning_rate/action_model': 9.400943848342246e-05, 'epoch': 0.47}) +[2026-06-17 18:53:13,925][starVLA.training.train_starvla][INFO] - Step 1216, Loss: {'train/loss': 0.6424351930618286, 'train/grad_norm_pre_clip': 7.376143932342529, 'timing/data': 0.01131223700940609, 'timing/model': 2.3462752643972635, 'learning_rate/qwen_vl_interface': 9.39989205539814e-06, 'learning_rate/action_model': 9.39989205539814e-05, 'epoch': 0.47}) +[2026-06-17 18:53:18,485][starVLA.training.train_starvla][INFO] - Step 1217, Loss: {'train/loss': 0.8388234376907349, 'train/grad_norm_pre_clip': 8.768622398376465, 'timing/data': 0.011594109237194061, 'timing/model': 1.3689398374408484, 'learning_rate/qwen_vl_interface': 9.398839402176946e-06, 'learning_rate/action_model': 9.398839402176946e-05, 'epoch': 0.47}) +[2026-06-17 18:53:22,685][starVLA.training.train_starvla][INFO] - Step 1218, Loss: {'train/loss': 0.47146040201187134, 'train/grad_norm_pre_clip': 5.395936489105225, 'timing/data': 0.01833588071167469, 'timing/model': 1.0813848255202174, 'learning_rate/qwen_vl_interface': 9.39778588889688e-06, 'learning_rate/action_model': 9.39778588889688e-05, 'epoch': 0.47}) +[2026-06-17 18:53:27,183][starVLA.training.train_starvla][INFO] - Step 1219, Loss: {'train/loss': 0.5311731100082397, 'train/grad_norm_pre_clip': 6.291974067687988, 'timing/data': 0.011871100403368473, 'timing/model': 1.0704493941739202, 'learning_rate/qwen_vl_interface': 9.396731515776335e-06, 'learning_rate/action_model': 9.396731515776333e-05, 'epoch': 0.47}) +[2026-06-17 18:53:31,881][starVLA.training.train_starvla][INFO] - Step 1220, Loss: {'train/loss': 0.7712579965591431, 'train/grad_norm_pre_clip': 9.070822715759277, 'timing/data': 0.015132579952478409, 'timing/model': 1.0900240745395422, 'learning_rate/qwen_vl_interface': 9.395676283033884e-06, 'learning_rate/action_model': 9.395676283033885e-05, 'epoch': 0.47}) +[2026-06-17 18:53:36,675][starVLA.training.train_starvla][INFO] - Step 1221, Loss: {'train/loss': 0.7476876974105835, 'train/grad_norm_pre_clip': 8.155619621276855, 'timing/data': 0.012652531266212463, 'timing/model': 1.1875644102692604, 'learning_rate/qwen_vl_interface': 9.39462019088828e-06, 'learning_rate/action_model': 9.394620190888279e-05, 'epoch': 0.47}) +[2026-06-17 18:53:41,451][starVLA.training.train_starvla][INFO] - Step 1222, Loss: {'train/loss': 0.6098542213439941, 'train/grad_norm_pre_clip': 6.13977575302124, 'timing/data': 0.012903069145977497, 'timing/model': 1.5268055014312267, 'learning_rate/qwen_vl_interface': 9.393563239558449e-06, 'learning_rate/action_model': 9.39356323955845e-05, 'epoch': 0.47}) +[2026-06-17 18:53:45,698][starVLA.training.train_starvla][INFO] - Step 1223, Loss: {'train/loss': 0.6983740329742432, 'train/grad_norm_pre_clip': 8.935240745544434, 'timing/data': 0.01164046861231327, 'timing/model': 1.093335366807878, 'learning_rate/qwen_vl_interface': 9.392505429263501e-06, 'learning_rate/action_model': 9.392505429263501e-05, 'epoch': 0.47}) +[2026-06-17 18:53:50,187][starVLA.training.train_starvla][INFO] - Step 1224, Loss: {'train/loss': 0.733352541923523, 'train/grad_norm_pre_clip': 19.589115142822266, 'timing/data': 0.013177850283682346, 'timing/model': 1.0749803865328431, 'learning_rate/qwen_vl_interface': 9.39144676022272e-06, 'learning_rate/action_model': 9.39144676022272e-05, 'epoch': 0.47}) +[2026-06-17 18:53:54,998][starVLA.training.train_starvla][INFO] - Step 1225, Loss: {'train/loss': 0.5402567982673645, 'train/grad_norm_pre_clip': 5.795485019683838, 'timing/data': 0.015258788131177425, 'timing/model': 1.0876232963055372, 'learning_rate/qwen_vl_interface': 9.390387232655568e-06, 'learning_rate/action_model': 9.390387232655569e-05, 'epoch': 0.47}) +[2026-06-17 18:53:59,647][starVLA.training.train_starvla][INFO] - Step 1226, Loss: {'train/loss': 0.624225378036499, 'train/grad_norm_pre_clip': 5.877843856811523, 'timing/data': 0.010943610221147537, 'timing/model': 1.140358597971499, 'learning_rate/qwen_vl_interface': 9.38932684678169e-06, 'learning_rate/action_model': 9.38932684678169e-05, 'epoch': 0.47}) +[2026-06-17 18:54:04,265][starVLA.training.train_starvla][INFO] - Step 1227, Loss: {'train/loss': 0.5938639044761658, 'train/grad_norm_pre_clip': 9.935067176818848, 'timing/data': 0.011427024379372597, 'timing/model': 1.3824387919157743, 'learning_rate/qwen_vl_interface': 9.388265602820901e-06, 'learning_rate/action_model': 9.3882656028209e-05, 'epoch': 0.47}) +[2026-06-17 18:54:08,548][starVLA.training.train_starvla][INFO] - Step 1228, Loss: {'train/loss': 0.7601989507675171, 'train/grad_norm_pre_clip': 10.214713096618652, 'timing/data': 0.014009561389684677, 'timing/model': 1.0964423073455691, 'learning_rate/qwen_vl_interface': 9.387203500993199e-06, 'learning_rate/action_model': 9.387203500993199e-05, 'epoch': 0.47}) +[2026-06-17 18:54:13,196][starVLA.training.train_starvla][INFO] - Step 1229, Loss: {'train/loss': 0.5128355622291565, 'train/grad_norm_pre_clip': 6.280709743499756, 'timing/data': 0.01241274829953909, 'timing/model': 1.097308874130249, 'learning_rate/qwen_vl_interface': 9.38614054151876e-06, 'learning_rate/action_model': 9.38614054151876e-05, 'epoch': 0.47}) +[2026-06-17 18:54:17,825][starVLA.training.train_starvla][INFO] - Step 1230, Loss: {'train/loss': 0.4980059564113617, 'train/grad_norm_pre_clip': 7.846078872680664, 'timing/data': 0.015957637690007687, 'timing/model': 1.110969988629222, 'learning_rate/qwen_vl_interface': 9.385076724617937e-06, 'learning_rate/action_model': 9.385076724617935e-05, 'epoch': 0.47}) +[2026-06-17 18:54:22,461][starVLA.training.train_starvla][INFO] - Step 1231, Loss: {'train/loss': 0.4407609701156616, 'train/grad_norm_pre_clip': 8.234619140625, 'timing/data': 0.01121652964502573, 'timing/model': 1.0723425252363086, 'learning_rate/qwen_vl_interface': 9.384012050511258e-06, 'learning_rate/action_model': 9.384012050511258e-05, 'epoch': 0.47}) +[2026-06-17 18:54:27,405][starVLA.training.train_starvla][INFO] - Step 1232, Loss: {'train/loss': 0.5748576521873474, 'train/grad_norm_pre_clip': 5.6326775550842285, 'timing/data': 0.025998626835644245, 'timing/model': 1.5001331493258476, 'learning_rate/qwen_vl_interface': 9.382946519419433e-06, 'learning_rate/action_model': 9.382946519419434e-05, 'epoch': 0.47}) +[2026-06-17 18:54:31,831][starVLA.training.train_starvla][INFO] - Step 1233, Loss: {'train/loss': 0.7102271914482117, 'train/grad_norm_pre_clip': 6.814610481262207, 'timing/data': 0.01782249193638563, 'timing/model': 1.194590331055224, 'learning_rate/qwen_vl_interface': 9.381880131563347e-06, 'learning_rate/action_model': 9.381880131563346e-05, 'epoch': 0.47}) +[2026-06-17 18:54:36,611][starVLA.training.train_starvla][INFO] - Step 1234, Loss: {'train/loss': 0.7460442781448364, 'train/grad_norm_pre_clip': 9.55941104888916, 'timing/data': 0.013818924315273762, 'timing/model': 1.1802481980994344, 'learning_rate/qwen_vl_interface': 9.380812887164065e-06, 'learning_rate/action_model': 9.380812887164065e-05, 'epoch': 0.47}) +[2026-06-17 18:54:41,467][starVLA.training.train_starvla][INFO] - Step 1235, Loss: {'train/loss': 0.8182049989700317, 'train/grad_norm_pre_clip': 14.14419937133789, 'timing/data': 0.020080897957086563, 'timing/model': 1.1071744319051504, 'learning_rate/qwen_vl_interface': 9.379744786442824e-06, 'learning_rate/action_model': 9.379744786442824e-05, 'epoch': 0.48}) +[2026-06-17 18:54:46,030][starVLA.training.train_starvla][INFO] - Step 1236, Loss: {'train/loss': 0.45851340889930725, 'train/grad_norm_pre_clip': 4.570812702178955, 'timing/data': 0.011526026763021946, 'timing/model': 1.077348261140287, 'learning_rate/qwen_vl_interface': 9.378675829621047e-06, 'learning_rate/action_model': 9.378675829621046e-05, 'epoch': 0.48}) +[2026-06-17 18:54:50,874][starVLA.training.train_starvla][INFO] - Step 1237, Loss: {'train/loss': 0.6660171151161194, 'train/grad_norm_pre_clip': 9.437697410583496, 'timing/data': 0.012266626581549644, 'timing/model': 1.532597349025309, 'learning_rate/qwen_vl_interface': 9.377606016920328e-06, 'learning_rate/action_model': 9.377606016920327e-05, 'epoch': 0.48}) +[2026-06-17 18:54:55,161][starVLA.training.train_starvla][INFO] - Step 1238, Loss: {'train/loss': 0.5418466925621033, 'train/grad_norm_pre_clip': 9.013221740722656, 'timing/data': 0.012286442331969738, 'timing/model': 1.1402100287377834, 'learning_rate/qwen_vl_interface': 9.376535348562438e-06, 'learning_rate/action_model': 9.376535348562439e-05, 'epoch': 0.48}) +[2026-06-17 18:55:00,243][starVLA.training.train_starvla][INFO] - Step 1239, Loss: {'train/loss': 0.8508819341659546, 'train/grad_norm_pre_clip': 11.313882827758789, 'timing/data': 0.016840477474033833, 'timing/model': 1.1522625871002674, 'learning_rate/qwen_vl_interface': 9.375463824769332e-06, 'learning_rate/action_model': 9.375463824769333e-05, 'epoch': 0.48}) +[2026-06-17 18:55:04,942][starVLA.training.train_starvla][INFO] - Step 1240, Loss: {'train/loss': 0.986028254032135, 'train/grad_norm_pre_clip': 8.745477676391602, 'timing/data': 0.012091184966266155, 'timing/model': 1.154137283563614, 'learning_rate/qwen_vl_interface': 9.374391445763135e-06, 'learning_rate/action_model': 9.374391445763135e-05, 'epoch': 0.48}) +[2026-06-17 18:55:11,587][starVLA.training.train_starvla][INFO] - Step 1241, Loss: {'train/loss': 0.5124837160110474, 'train/grad_norm_pre_clip': 7.239007472991943, 'timing/data': 0.01203838735818863, 'timing/model': 1.0708712609484792, 'learning_rate/qwen_vl_interface': 9.373318211766155e-06, 'learning_rate/action_model': 9.373318211766155e-05, 'epoch': 0.48}) +[2026-06-17 18:55:16,337][starVLA.training.train_starvla][INFO] - Step 1242, Loss: {'train/loss': 0.3700081408023834, 'train/grad_norm_pre_clip': 5.401633262634277, 'timing/data': 0.012489700689911842, 'timing/model': 1.4105391222983599, 'learning_rate/qwen_vl_interface': 9.37224412300087e-06, 'learning_rate/action_model': 9.37224412300087e-05, 'epoch': 0.48}) +[2026-06-17 18:55:20,699][starVLA.training.train_starvla][INFO] - Step 1243, Loss: {'train/loss': 0.5176275372505188, 'train/grad_norm_pre_clip': 6.642997741699219, 'timing/data': 0.011753324419260025, 'timing/model': 1.2216461142525077, 'learning_rate/qwen_vl_interface': 9.371169179689945e-06, 'learning_rate/action_model': 9.371169179689945e-05, 'epoch': 0.48}) +[2026-06-17 18:55:25,381][starVLA.training.train_starvla][INFO] - Step 1244, Loss: {'train/loss': 0.7228865623474121, 'train/grad_norm_pre_clip': 6.4287848472595215, 'timing/data': 0.018550983630120754, 'timing/model': 1.1417540665715933, 'learning_rate/qwen_vl_interface': 9.370093382056213e-06, 'learning_rate/action_model': 9.370093382056214e-05, 'epoch': 0.48}) +[2026-06-17 18:55:29,940][starVLA.training.train_starvla][INFO] - Step 1245, Loss: {'train/loss': 0.6658468246459961, 'train/grad_norm_pre_clip': 7.12742280960083, 'timing/data': 0.011718172580003738, 'timing/model': 1.0802775342017412, 'learning_rate/qwen_vl_interface': 9.369016730322692e-06, 'learning_rate/action_model': 9.36901673032269e-05, 'epoch': 0.48}) +[2026-06-17 18:55:34,604][starVLA.training.train_starvla][INFO] - Step 1246, Loss: {'train/loss': 0.5417943596839905, 'train/grad_norm_pre_clip': 13.29116153717041, 'timing/data': 0.018536224961280823, 'timing/model': 1.1442724419757724, 'learning_rate/qwen_vl_interface': 9.367939224712566e-06, 'learning_rate/action_model': 9.367939224712566e-05, 'epoch': 0.48}) +[2026-06-17 18:55:44,974][starVLA.training.train_starvla][INFO] - Step 1247, Loss: {'train/loss': 0.7402918338775635, 'train/grad_norm_pre_clip': 8.419585227966309, 'timing/data': 0.014303446747362614, 'timing/model': 1.5880921622738242, 'learning_rate/qwen_vl_interface': 9.366860865449213e-06, 'learning_rate/action_model': 9.366860865449212e-05, 'epoch': 0.48}) +[2026-06-17 18:55:49,276][starVLA.training.train_starvla][INFO] - Step 1248, Loss: {'train/loss': 0.3514828681945801, 'train/grad_norm_pre_clip': 6.120975494384766, 'timing/data': 0.01380463968962431, 'timing/model': 1.110478120855987, 'learning_rate/qwen_vl_interface': 9.365781652756168e-06, 'learning_rate/action_model': 9.365781652756168e-05, 'epoch': 0.48}) +[2026-06-17 18:55:54,129][starVLA.training.train_starvla][INFO] - Step 1249, Loss: {'train/loss': 0.4455350339412689, 'train/grad_norm_pre_clip': 6.0835771560668945, 'timing/data': 0.013801886700093746, 'timing/model': 1.0786878298968077, 'learning_rate/qwen_vl_interface': 9.36470158685716e-06, 'learning_rate/action_model': 9.36470158685716e-05, 'epoch': 0.48}) +[2026-06-17 18:55:58,717][starVLA.training.train_starvla][INFO] - Step 1250, Loss: {'train/loss': 0.8988906145095825, 'train/grad_norm_pre_clip': 8.990198135375977, 'timing/data': 0.011933423578739166, 'timing/model': 1.072545071132481, 'learning_rate/qwen_vl_interface': 9.363620667976085e-06, 'learning_rate/action_model': 9.363620667976085e-05, 'epoch': 0.48}) +[2026-06-17 18:56:03,466][starVLA.training.train_starvla][INFO] - Step 1251, Loss: {'train/loss': 0.6906290054321289, 'train/grad_norm_pre_clip': 7.08174991607666, 'timing/data': 0.015415091998875141, 'timing/model': 1.1398201817646623, 'learning_rate/qwen_vl_interface': 9.362538896337019e-06, 'learning_rate/action_model': 9.362538896337019e-05, 'epoch': 0.48}) +[2026-06-17 18:56:08,177][starVLA.training.train_starvla][INFO] - Step 1252, Loss: {'train/loss': 0.4506426155567169, 'train/grad_norm_pre_clip': 4.11005973815918, 'timing/data': 0.01276425737887621, 'timing/model': 1.4745486667379737, 'learning_rate/qwen_vl_interface': 9.361456272164215e-06, 'learning_rate/action_model': 9.361456272164214e-05, 'epoch': 0.48}) +[2026-06-17 18:56:12,958][starVLA.training.train_starvla][INFO] - Step 1253, Loss: {'train/loss': 0.5259438753128052, 'train/grad_norm_pre_clip': 5.141210556030273, 'timing/data': 0.01161565538495779, 'timing/model': 1.6155056804418564, 'learning_rate/qwen_vl_interface': 9.3603727956821e-06, 'learning_rate/action_model': 9.3603727956821e-05, 'epoch': 0.48}) +[2026-06-17 18:56:17,692][starVLA.training.train_starvla][INFO] - Step 1254, Loss: {'train/loss': 0.8127859830856323, 'train/grad_norm_pre_clip': 7.294066429138184, 'timing/data': 0.016598698683083057, 'timing/model': 1.1939800838008523, 'learning_rate/qwen_vl_interface': 9.359288467115283e-06, 'learning_rate/action_model': 9.359288467115283e-05, 'epoch': 0.48}) +[2026-06-17 18:56:22,696][starVLA.training.train_starvla][INFO] - Step 1255, Loss: {'train/loss': 0.7279096245765686, 'train/grad_norm_pre_clip': 9.498612403869629, 'timing/data': 0.011802216060459614, 'timing/model': 1.217312103137374, 'learning_rate/qwen_vl_interface': 9.358203286688544e-06, 'learning_rate/action_model': 9.358203286688544e-05, 'epoch': 0.48}) +[2026-06-17 18:56:27,593][starVLA.training.train_starvla][INFO] - Step 1256, Loss: {'train/loss': 0.5555992722511292, 'train/grad_norm_pre_clip': 5.538628101348877, 'timing/data': 0.015292716212570667, 'timing/model': 1.0758647723123431, 'learning_rate/qwen_vl_interface': 9.357117254626844e-06, 'learning_rate/action_model': 9.357117254626844e-05, 'epoch': 0.48}) +[2026-06-17 18:56:32,464][starVLA.training.train_starvla][INFO] - Step 1257, Loss: {'train/loss': 0.6231383681297302, 'train/grad_norm_pre_clip': 7.429609775543213, 'timing/data': 0.014054526574909687, 'timing/model': 1.5349253034219146, 'learning_rate/qwen_vl_interface': 9.356030371155316e-06, 'learning_rate/action_model': 9.356030371155316e-05, 'epoch': 0.48}) +[2026-06-17 18:56:36,864][starVLA.training.train_starvla][INFO] - Step 1258, Loss: {'train/loss': 0.7309159636497498, 'train/grad_norm_pre_clip': 6.854716777801514, 'timing/data': 0.015755468048155308, 'timing/model': 1.1138449031859636, 'learning_rate/qwen_vl_interface': 9.354942636499275e-06, 'learning_rate/action_model': 9.354942636499276e-05, 'epoch': 0.48}) +[2026-06-17 18:56:41,636][starVLA.training.train_starvla][INFO] - Step 1259, Loss: {'train/loss': 0.4086284935474396, 'train/grad_norm_pre_clip': 5.532961845397949, 'timing/data': 0.01427516806870699, 'timing/model': 1.0962223019450903, 'learning_rate/qwen_vl_interface': 9.35385405088421e-06, 'learning_rate/action_model': 9.353854050884208e-05, 'epoch': 0.48}) +[2026-06-17 18:56:46,425][starVLA.training.train_starvla][INFO] - Step 1260, Loss: {'train/loss': 0.7201300859451294, 'train/grad_norm_pre_clip': 6.004535675048828, 'timing/data': 0.011247429996728897, 'timing/model': 1.0690556857734919, 'learning_rate/qwen_vl_interface': 9.352764614535782e-06, 'learning_rate/action_model': 9.352764614535781e-05, 'epoch': 0.48}) +[2026-06-17 18:56:51,192][starVLA.training.train_starvla][INFO] - Step 1261, Loss: {'train/loss': 0.2296944558620453, 'train/grad_norm_pre_clip': 3.302215337753296, 'timing/data': 0.012805548496544361, 'timing/model': 1.1158573189750314, 'learning_rate/qwen_vl_interface': 9.351674327679837e-06, 'learning_rate/action_model': 9.351674327679836e-05, 'epoch': 0.49}) +[2026-06-17 18:57:00,033][starVLA.training.train_starvla][INFO] - Step 1262, Loss: {'train/loss': 0.4262937307357788, 'train/grad_norm_pre_clip': 6.4542365074157715, 'timing/data': 0.010591013357043266, 'timing/model': 1.4316288344562054, 'learning_rate/qwen_vl_interface': 9.350583190542387e-06, 'learning_rate/action_model': 9.350583190542387e-05, 'epoch': 0.49}) +[2026-06-17 18:57:04,315][starVLA.training.train_starvla][INFO] - Step 1263, Loss: {'train/loss': 0.3037591576576233, 'train/grad_norm_pre_clip': 5.9456586837768555, 'timing/data': 0.014359655790030956, 'timing/model': 1.0890300367027521, 'learning_rate/qwen_vl_interface': 9.34949120334963e-06, 'learning_rate/action_model': 9.34949120334963e-05, 'epoch': 0.49}) +[2026-06-17 18:57:09,040][starVLA.training.train_starvla][INFO] - Step 1264, Loss: {'train/loss': 0.6274573802947998, 'train/grad_norm_pre_clip': 5.173881530761719, 'timing/data': 0.014251396991312504, 'timing/model': 1.1208360567688942, 'learning_rate/qwen_vl_interface': 9.348398366327938e-06, 'learning_rate/action_model': 9.348398366327937e-05, 'epoch': 0.49}) +[2026-06-17 18:57:13,690][starVLA.training.train_starvla][INFO] - Step 1265, Loss: {'train/loss': 0.40395069122314453, 'train/grad_norm_pre_clip': 5.07777738571167, 'timing/data': 0.011110282503068447, 'timing/model': 1.081528295762837, 'learning_rate/qwen_vl_interface': 9.347304679703852e-06, 'learning_rate/action_model': 9.347304679703851e-05, 'epoch': 0.49}) +[2026-06-17 18:57:18,560][starVLA.training.train_starvla][INFO] - Step 1266, Loss: {'train/loss': 0.7578689455986023, 'train/grad_norm_pre_clip': 8.441985130310059, 'timing/data': 0.01152842864394188, 'timing/model': 1.2123198108747602, 'learning_rate/qwen_vl_interface': 9.346210143704099e-06, 'learning_rate/action_model': 9.346210143704098e-05, 'epoch': 0.49}) +[2026-06-17 18:57:23,324][starVLA.training.train_starvla][INFO] - Step 1267, Loss: {'train/loss': 0.40240705013275146, 'train/grad_norm_pre_clip': 6.829497337341309, 'timing/data': 0.018031137995421886, 'timing/model': 1.531296537257731, 'learning_rate/qwen_vl_interface': 9.345114758555573e-06, 'learning_rate/action_model': 9.345114758555573e-05, 'epoch': 0.49}) +[2026-06-17 18:57:27,745][starVLA.training.train_starvla][INFO] - Step 1268, Loss: {'train/loss': 0.6604411005973816, 'train/grad_norm_pre_clip': 6.598249435424805, 'timing/data': 0.01124486606568098, 'timing/model': 1.1592339389026165, 'learning_rate/qwen_vl_interface': 9.344018524485352e-06, 'learning_rate/action_model': 9.344018524485351e-05, 'epoch': 0.49}) +[2026-06-17 18:57:32,347][starVLA.training.train_starvla][INFO] - Step 1269, Loss: {'train/loss': 0.5913330912590027, 'train/grad_norm_pre_clip': 6.195855617523193, 'timing/data': 0.019816934131085873, 'timing/model': 1.107499299570918, 'learning_rate/qwen_vl_interface': 9.342921441720686e-06, 'learning_rate/action_model': 9.342921441720685e-05, 'epoch': 0.49}) +[2026-06-17 18:57:37,193][starVLA.training.train_starvla][INFO] - Step 1270, Loss: {'train/loss': 0.603325605392456, 'train/grad_norm_pre_clip': 7.8846235275268555, 'timing/data': 0.011071253567934036, 'timing/model': 1.1491850325837731, 'learning_rate/qwen_vl_interface': 9.341823510489e-06, 'learning_rate/action_model': 9.341823510489e-05, 'epoch': 0.49}) +[2026-06-17 18:57:43,558][starVLA.training.train_starvla][INFO] - Step 1271, Loss: {'train/loss': 0.689579427242279, 'train/grad_norm_pre_clip': 9.172945976257324, 'timing/data': 0.011875680647790432, 'timing/model': 1.0672786114737391, 'learning_rate/qwen_vl_interface': 9.340724731017898e-06, 'learning_rate/action_model': 9.340724731017898e-05, 'epoch': 0.49}) +[2026-06-17 18:57:48,370][starVLA.training.train_starvla][INFO] - Step 1272, Loss: {'train/loss': 1.2584022283554077, 'train/grad_norm_pre_clip': 8.465917587280273, 'timing/data': 0.011662868782877922, 'timing/model': 1.5185816837474704, 'learning_rate/qwen_vl_interface': 9.339625103535158e-06, 'learning_rate/action_model': 9.339625103535156e-05, 'epoch': 0.49}) +[2026-06-17 18:57:52,776][starVLA.training.train_starvla][INFO] - Step 1273, Loss: {'train/loss': 0.7847467660903931, 'train/grad_norm_pre_clip': 9.21939468383789, 'timing/data': 0.014955800957977772, 'timing/model': 1.1986730778589845, 'learning_rate/qwen_vl_interface': 9.338524628268734e-06, 'learning_rate/action_model': 9.338524628268733e-05, 'epoch': 0.49}) +[2026-06-17 18:57:57,511][starVLA.training.train_starvla][INFO] - Step 1274, Loss: {'train/loss': 0.44695204496383667, 'train/grad_norm_pre_clip': 5.1471943855285645, 'timing/data': 0.015563447028398514, 'timing/model': 1.1447246987372637, 'learning_rate/qwen_vl_interface': 9.337423305446753e-06, 'learning_rate/action_model': 9.337423305446753e-05, 'epoch': 0.49}) +[2026-06-17 18:58:02,249][starVLA.training.train_starvla][INFO] - Step 1275, Loss: {'train/loss': 0.7672021389007568, 'train/grad_norm_pre_clip': 10.461003303527832, 'timing/data': 0.013471562415361404, 'timing/model': 1.0939093688502908, 'learning_rate/qwen_vl_interface': 9.336321135297525e-06, 'learning_rate/action_model': 9.336321135297525e-05, 'epoch': 0.49}) +[2026-06-17 18:58:06,998][starVLA.training.train_starvla][INFO] - Step 1276, Loss: {'train/loss': 0.46327483654022217, 'train/grad_norm_pre_clip': 5.070184230804443, 'timing/data': 0.01233257818967104, 'timing/model': 1.0783351939171553, 'learning_rate/qwen_vl_interface': 9.335218118049526e-06, 'learning_rate/action_model': 9.335218118049526e-05, 'epoch': 0.49}) +[2026-06-17 18:58:11,889][starVLA.training.train_starvla][INFO] - Step 1277, Loss: {'train/loss': 0.6782186031341553, 'train/grad_norm_pre_clip': 5.760763168334961, 'timing/data': 0.012382833287119865, 'timing/model': 1.4798646410927176, 'learning_rate/qwen_vl_interface': 9.334114253931417e-06, 'learning_rate/action_model': 9.334114253931416e-05, 'epoch': 0.49}) +[2026-06-17 18:58:16,224][starVLA.training.train_starvla][INFO] - Step 1278, Loss: {'train/loss': 0.6296796798706055, 'train/grad_norm_pre_clip': 7.524243354797363, 'timing/data': 0.013950810767710209, 'timing/model': 1.1260406654328108, 'learning_rate/qwen_vl_interface': 9.333009543172027e-06, 'learning_rate/action_model': 9.333009543172027e-05, 'epoch': 0.49}) +[2026-06-17 18:58:21,075][starVLA.training.train_starvla][INFO] - Step 1279, Loss: {'train/loss': 0.8782808184623718, 'train/grad_norm_pre_clip': 8.796749114990234, 'timing/data': 0.012443380430340767, 'timing/model': 1.1845278618857265, 'learning_rate/qwen_vl_interface': 9.331903986000365e-06, 'learning_rate/action_model': 9.331903986000365e-05, 'epoch': 0.49}) +[2026-06-17 18:58:25,937][starVLA.training.train_starvla][INFO] - Step 1280, Loss: {'train/loss': 0.7227172255516052, 'train/grad_norm_pre_clip': 9.28602123260498, 'timing/data': 0.011363480240106583, 'timing/model': 1.1015496803447604, 'learning_rate/qwen_vl_interface': 9.330797582645615e-06, 'learning_rate/action_model': 9.330797582645615e-05, 'epoch': 0.49}) +[2026-06-17 18:58:30,670][starVLA.training.train_starvla][INFO] - Step 1281, Loss: {'train/loss': 0.5351050496101379, 'train/grad_norm_pre_clip': 6.159273624420166, 'timing/data': 0.014835628680884838, 'timing/model': 1.0885843383148313, 'learning_rate/qwen_vl_interface': 9.329690333337137e-06, 'learning_rate/action_model': 9.329690333337137e-05, 'epoch': 0.49}) +[2026-06-17 18:58:38,884][starVLA.training.train_starvla][INFO] - Step 1282, Loss: {'train/loss': 0.5941593647003174, 'train/grad_norm_pre_clip': 8.780238151550293, 'timing/data': 0.011375563219189644, 'timing/model': 2.9979660706594586, 'learning_rate/qwen_vl_interface': 9.328582238304461e-06, 'learning_rate/action_model': 9.328582238304461e-05, 'epoch': 0.49}) +[2026-06-17 18:58:43,907][starVLA.training.train_starvla][INFO] - Step 1283, Loss: {'train/loss': 0.4253104329109192, 'train/grad_norm_pre_clip': 5.1870832443237305, 'timing/data': 0.015774961560964584, 'timing/model': 1.0905811879783869, 'learning_rate/qwen_vl_interface': 9.327473297777297e-06, 'learning_rate/action_model': 9.327473297777297e-05, 'epoch': 0.49}) +[2026-06-17 18:58:48,856][starVLA.training.train_starvla][INFO] - Step 1284, Loss: {'train/loss': 0.5733820796012878, 'train/grad_norm_pre_clip': 6.58274507522583, 'timing/data': 0.01206963136792183, 'timing/model': 1.3510792423039675, 'learning_rate/qwen_vl_interface': 9.326363511985533e-06, 'learning_rate/action_model': 9.326363511985533e-05, 'epoch': 0.49}) +[2026-06-17 18:58:55,536][starVLA.training.train_starvla][INFO] - Step 1285, Loss: {'train/loss': 0.5643430352210999, 'train/grad_norm_pre_clip': 6.450506210327148, 'timing/data': 0.011717968620359898, 'timing/model': 1.1200141170993447, 'learning_rate/qwen_vl_interface': 9.325252881159226e-06, 'learning_rate/action_model': 9.325252881159226e-05, 'epoch': 0.49}) +[2026-06-17 18:59:00,261][starVLA.training.train_starvla][INFO] - Step 1286, Loss: {'train/loss': 0.5034986734390259, 'train/grad_norm_pre_clip': 5.032071113586426, 'timing/data': 0.013381445780396461, 'timing/model': 1.1794172273948789, 'learning_rate/qwen_vl_interface': 9.324141405528612e-06, 'learning_rate/action_model': 9.324141405528612e-05, 'epoch': 0.49}) +[2026-06-17 18:59:05,259][starVLA.training.train_starvla][INFO] - Step 1287, Loss: {'train/loss': 1.0155237913131714, 'train/grad_norm_pre_clip': 12.947040557861328, 'timing/data': 0.011966634541749954, 'timing/model': 1.7021334394812584, 'learning_rate/qwen_vl_interface': 9.3230290853241e-06, 'learning_rate/action_model': 9.3230290853241e-05, 'epoch': 0.5}) +[2026-06-17 18:59:11,402][starVLA.training.train_starvla][INFO] - Step 1288, Loss: {'train/loss': 0.5944715142250061, 'train/grad_norm_pre_clip': 7.791405200958252, 'timing/data': 0.01305715274065733, 'timing/model': 1.0955258002504706, 'learning_rate/qwen_vl_interface': 9.321915920776278e-06, 'learning_rate/action_model': 9.321915920776278e-05, 'epoch': 0.5}) +[2026-06-17 18:59:16,268][starVLA.training.train_starvla][INFO] - Step 1289, Loss: {'train/loss': 0.6048415303230286, 'train/grad_norm_pre_clip': 5.4073405265808105, 'timing/data': 0.013094797730445862, 'timing/model': 1.132883356884122, 'learning_rate/qwen_vl_interface': 9.320801912115901e-06, 'learning_rate/action_model': 9.320801912115901e-05, 'epoch': 0.5}) +[2026-06-17 18:59:20,978][starVLA.training.train_starvla][INFO] - Step 1290, Loss: {'train/loss': 0.6190483570098877, 'train/grad_norm_pre_clip': 8.577528953552246, 'timing/data': 0.011800257489085197, 'timing/model': 1.0747538581490517, 'learning_rate/qwen_vl_interface': 9.319687059573909e-06, 'learning_rate/action_model': 9.319687059573909e-05, 'epoch': 0.5}) +[2026-06-17 18:59:25,717][starVLA.training.train_starvla][INFO] - Step 1291, Loss: {'train/loss': 0.8682080507278442, 'train/grad_norm_pre_clip': 13.875701904296875, 'timing/data': 0.012220543809235096, 'timing/model': 1.1716920528560877, 'learning_rate/qwen_vl_interface': 9.318571363381408e-06, 'learning_rate/action_model': 9.318571363381408e-05, 'epoch': 0.5}) +[2026-06-17 18:59:31,652][starVLA.training.train_starvla][INFO] - Step 1292, Loss: {'train/loss': 0.7785354256629944, 'train/grad_norm_pre_clip': 9.721460342407227, 'timing/data': 0.014049874618649483, 'timing/model': 1.6151239862665534, 'learning_rate/qwen_vl_interface': 9.317454823769686e-06, 'learning_rate/action_model': 9.317454823769687e-05, 'epoch': 0.5}) +[2026-06-17 18:59:36,087][starVLA.training.train_starvla][INFO] - Step 1293, Loss: {'train/loss': 0.768730103969574, 'train/grad_norm_pre_clip': 10.848370552062988, 'timing/data': 0.011262410320341587, 'timing/model': 1.1273885825648904, 'learning_rate/qwen_vl_interface': 9.316337440970206e-06, 'learning_rate/action_model': 9.316337440970206e-05, 'epoch': 0.5}) +[2026-06-17 18:59:40,966][starVLA.training.train_starvla][INFO] - Step 1294, Loss: {'train/loss': 0.5518128275871277, 'train/grad_norm_pre_clip': 4.784773826599121, 'timing/data': 0.012722332961857319, 'timing/model': 1.183272578753531, 'learning_rate/qwen_vl_interface': 9.315219215214596e-06, 'learning_rate/action_model': 9.315219215214595e-05, 'epoch': 0.5}) +[2026-06-17 18:59:45,923][starVLA.training.train_starvla][INFO] - Step 1295, Loss: {'train/loss': 0.5455390214920044, 'train/grad_norm_pre_clip': 7.994703769683838, 'timing/data': 0.014907140284776688, 'timing/model': 1.099883595481515, 'learning_rate/qwen_vl_interface': 9.314100146734668e-06, 'learning_rate/action_model': 9.314100146734668e-05, 'epoch': 0.5}) +[2026-06-17 18:59:50,887][starVLA.training.train_starvla][INFO] - Step 1296, Loss: {'train/loss': 0.47367167472839355, 'train/grad_norm_pre_clip': 6.395683288574219, 'timing/data': 0.011616921983659267, 'timing/model': 1.0896893171593547, 'learning_rate/qwen_vl_interface': 9.312980235762408e-06, 'learning_rate/action_model': 9.312980235762408e-05, 'epoch': 0.5}) +[2026-06-17 18:59:55,868][starVLA.training.train_starvla][INFO] - Step 1297, Loss: {'train/loss': 0.5082778334617615, 'train/grad_norm_pre_clip': 5.8801398277282715, 'timing/data': 0.011452876962721348, 'timing/model': 1.747490898706019, 'learning_rate/qwen_vl_interface': 9.311859482529971e-06, 'learning_rate/action_model': 9.31185948252997e-05, 'epoch': 0.5}) +[2026-06-17 19:00:00,116][starVLA.training.train_starvla][INFO] - Step 1298, Loss: {'train/loss': 0.6870688796043396, 'train/grad_norm_pre_clip': 7.211328029632568, 'timing/data': 0.011025884188711643, 'timing/model': 1.0828644065186381, 'learning_rate/qwen_vl_interface': 9.310737887269694e-06, 'learning_rate/action_model': 9.310737887269694e-05, 'epoch': 0.5}) +[2026-06-17 19:00:04,620][starVLA.training.train_starvla][INFO] - Step 1299, Loss: {'train/loss': 1.1039751768112183, 'train/grad_norm_pre_clip': 9.337532997131348, 'timing/data': 0.012349925935268402, 'timing/model': 1.0447176350280643, 'learning_rate/qwen_vl_interface': 9.309615450214084e-06, 'learning_rate/action_model': 9.309615450214084e-05, 'epoch': 0.5}) +[2026-06-17 19:02:31,639][starVLA.training.train_starvla][INFO] - Step 1300, Loss: {'train/loss': 0.6276218891143799, 'train/grad_norm_pre_clip': 9.441130638122559, 'eval/loss': 0.6598249077796936, 'eval/latency_4/loss': 0.6598249077796936, 'eval/demon_fire_precision': 0.9415564441710244, 'eval/demon_fire_recall': 0.9941539460863917, 'eval/demon_fire_f1': 0.967140600315956, 'eval/demon_fire_rate_model': 0.9547723935389134, 'eval/demon_fire_rate_teacher': 0.9042584434654919, 'eval/demon_left_precision': 0.7584541062801933, 'eval/demon_left_recall': 0.3584474885844749, 'eval/demon_left_f1': 0.48682170542635667, 'eval/demon_left_rate_model': 0.06079295154185022, 'eval/demon_left_rate_teacher': 0.12863436123348018, 'eval/demon_right_precision': 0.5971128608923885, 'eval/demon_right_recall': 0.8212996389891697, 'eval/demon_right_f1': 0.6914893617021276, 'eval/demon_right_rate_model': 0.22378854625550662, 'eval/demon_right_rate_teacher': 0.16270190895741557, 'eval/demon_move_f1': 0.5891555335642421, 'eval/demon_cc_f1': 0.8110389038623691, 'eval/latency_4/demon_fire_precision': 0.9415564441710244, 'eval/latency_4/demon_fire_recall': 0.9941539460863917, 'eval/latency_4/demon_fire_f1': 0.967140600315956, 'eval/latency_4/demon_fire_rate_model': 0.9547723935389134, 'eval/latency_4/demon_fire_rate_teacher': 0.9042584434654919, 'eval/latency_4/demon_left_precision': 0.7584541062801933, 'eval/latency_4/demon_left_recall': 0.3584474885844749, 'eval/latency_4/demon_left_f1': 0.48682170542635667, 'eval/latency_4/demon_left_rate_model': 0.06079295154185022, 'eval/latency_4/demon_left_rate_teacher': 0.12863436123348018, 'eval/latency_4/demon_right_precision': 0.5971128608923885, 'eval/latency_4/demon_right_recall': 0.8212996389891697, 'eval/latency_4/demon_right_f1': 0.6914893617021276, 'eval/latency_4/demon_right_rate_model': 0.22378854625550662, 'eval/latency_4/demon_right_rate_teacher': 0.16270190895741557, 'eval/latency_4/demon_move_f1': 0.5891555335642421, 'eval/latency_4/demon_cc_f1': 0.8110389038623691, 'timing/data': 0.011338555254042149, 'timing/model': 1.0543644353747368, 'learning_rate/qwen_vl_interface': 9.308492171595821e-06, 'learning_rate/action_model': 9.30849217159582e-05, 'epoch': 0.5}) +[2026-06-17 19:02:31,689][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_1300_state +[2026-06-17 19:02:31,697][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 19:02:40,308][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_1300_state/model.safetensors +[2026-06-17 19:03:02,479][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_1300_state/optimizer.bin +[2026-06-17 19:03:02,482][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_1300_state/sampler.bin +[2026-06-17 19:03:02,482][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_1300_state/sampler_1.bin +[2026-06-17 19:03:02,485][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_1300_state/random_states_0.pkl +[2026-06-17 19:03:02,492][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1300 +[2026-06-17 19:03:02,492][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 19:03:02,511][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 19:03:07,254][starVLA.training.train_starvla][INFO] - Step 1301, Loss: {'train/loss': 0.6841801404953003, 'train/grad_norm_pre_clip': 8.508257865905762, 'timing/data': 0.011624964885413647, 'timing/model': 1.5500812381505966, 'learning_rate/qwen_vl_interface': 9.307368051647765e-06, 'learning_rate/action_model': 9.307368051647765e-05, 'epoch': 0.5}) +[2026-06-17 19:03:12,227][starVLA.training.train_starvla][INFO] - Step 1302, Loss: {'train/loss': 0.47020968794822693, 'train/grad_norm_pre_clip': 6.127381324768066, 'timing/data': 0.012351345270872116, 'timing/model': 1.5789702627807856, 'learning_rate/qwen_vl_interface': 9.306243090602945e-06, 'learning_rate/action_model': 9.306243090602946e-05, 'epoch': 0.5}) +[2026-06-17 19:03:16,765][starVLA.training.train_starvla][INFO] - Step 1303, Loss: {'train/loss': 0.29511991143226624, 'train/grad_norm_pre_clip': 4.931942462921143, 'timing/data': 0.012271616607904434, 'timing/model': 1.116945443674922, 'learning_rate/qwen_vl_interface': 9.30511728869457e-06, 'learning_rate/action_model': 9.305117288694569e-05, 'epoch': 0.5}) +[2026-06-17 19:03:21,480][starVLA.training.train_starvla][INFO] - Step 1304, Loss: {'train/loss': 0.8045198321342468, 'train/grad_norm_pre_clip': 8.663768768310547, 'timing/data': 0.011929763481020927, 'timing/model': 1.1459863670170307, 'learning_rate/qwen_vl_interface': 9.303990646156015e-06, 'learning_rate/action_model': 9.303990646156014e-05, 'epoch': 0.5}) +[2026-06-17 19:03:28,394][starVLA.training.train_starvla][INFO] - Step 1305, Loss: {'train/loss': 0.39641639590263367, 'train/grad_norm_pre_clip': 5.052367210388184, 'timing/data': 0.013719387352466583, 'timing/model': 1.1214665975421667, 'learning_rate/qwen_vl_interface': 9.302863163220837e-06, 'learning_rate/action_model': 9.302863163220837e-05, 'epoch': 0.5}) +[2026-06-17 19:03:33,322][starVLA.training.train_starvla][INFO] - Step 1306, Loss: {'train/loss': 0.5921461582183838, 'train/grad_norm_pre_clip': 10.242713928222656, 'timing/data': 0.013862481340765953, 'timing/model': 1.1633462104946375, 'learning_rate/qwen_vl_interface': 9.301734840122765e-06, 'learning_rate/action_model': 9.301734840122766e-05, 'epoch': 0.5}) +[2026-06-17 19:03:38,288][starVLA.training.train_starvla][INFO] - Step 1307, Loss: {'train/loss': 0.2902523875236511, 'train/grad_norm_pre_clip': 4.091306209564209, 'timing/data': 0.014007325284183025, 'timing/model': 1.5568100092932582, 'learning_rate/qwen_vl_interface': 9.300605677095701e-06, 'learning_rate/action_model': 9.3006056770957e-05, 'epoch': 0.5}) +[2026-06-17 19:03:42,776][starVLA.training.train_starvla][INFO] - Step 1308, Loss: {'train/loss': 0.39337217807769775, 'train/grad_norm_pre_clip': 6.386719226837158, 'timing/data': 0.013749727979302406, 'timing/model': 1.1610828721895814, 'learning_rate/qwen_vl_interface': 9.299475674373722e-06, 'learning_rate/action_model': 9.299475674373721e-05, 'epoch': 0.5}) +[2026-06-17 19:03:47,861][starVLA.training.train_starvla][INFO] - Step 1309, Loss: {'train/loss': 0.5144569277763367, 'train/grad_norm_pre_clip': 7.466784954071045, 'timing/data': 0.012495632283389568, 'timing/model': 1.1378404684364796, 'learning_rate/qwen_vl_interface': 9.298344832191076e-06, 'learning_rate/action_model': 9.298344832191076e-05, 'epoch': 0.5}) +[2026-06-17 19:03:52,972][starVLA.training.train_starvla][INFO] - Step 1310, Loss: {'train/loss': 0.3582214117050171, 'train/grad_norm_pre_clip': 10.54133129119873, 'timing/data': 0.013066146522760391, 'timing/model': 1.149626448750496, 'learning_rate/qwen_vl_interface': 9.297213150782192e-06, 'learning_rate/action_model': 9.297213150782192e-05, 'epoch': 0.5}) +[2026-06-17 19:03:57,747][starVLA.training.train_starvla][INFO] - Step 1311, Loss: {'train/loss': 0.7872603535652161, 'train/grad_norm_pre_clip': 6.796628475189209, 'timing/data': 0.030311807058751583, 'timing/model': 1.1034846687689424, 'learning_rate/qwen_vl_interface': 9.296080630381666e-06, 'learning_rate/action_model': 9.296080630381666e-05, 'epoch': 0.5}) +[2026-06-17 19:04:02,542][starVLA.training.train_starvla][INFO] - Step 1312, Loss: {'train/loss': 0.7391799092292786, 'train/grad_norm_pre_clip': 14.192488670349121, 'timing/data': 0.018130884505808353, 'timing/model': 1.5828179698437452, 'learning_rate/qwen_vl_interface': 9.294947271224271e-06, 'learning_rate/action_model': 9.294947271224272e-05, 'epoch': 0.5}) +[2026-06-17 19:04:06,989][starVLA.training.train_starvla][INFO] - Step 1313, Loss: {'train/loss': 0.718527615070343, 'train/grad_norm_pre_clip': 7.808061599731445, 'timing/data': 0.0112706134095788, 'timing/model': 1.0880708480253816, 'learning_rate/qwen_vl_interface': 9.293813073544957e-06, 'learning_rate/action_model': 9.293813073544956e-05, 'epoch': 0.51}) +[2026-06-17 19:04:11,931][starVLA.training.train_starvla][INFO] - Step 1314, Loss: {'train/loss': 0.6229219436645508, 'train/grad_norm_pre_clip': 28.181337356567383, 'timing/data': 0.013119526207447052, 'timing/model': 1.1580677898600698, 'learning_rate/qwen_vl_interface': 9.292678037578838e-06, 'learning_rate/action_model': 9.292678037578837e-05, 'epoch': 0.51}) +[2026-06-17 19:04:16,911][starVLA.training.train_starvla][INFO] - Step 1315, Loss: {'train/loss': 0.6067870259284973, 'train/grad_norm_pre_clip': 6.443450450897217, 'timing/data': 0.0134343346580863, 'timing/model': 1.109664160758257, 'learning_rate/qwen_vl_interface': 9.291542163561214e-06, 'learning_rate/action_model': 9.291542163561213e-05, 'epoch': 0.51}) +[2026-06-17 19:04:21,769][starVLA.training.train_starvla][INFO] - Step 1316, Loss: {'train/loss': 0.701936662197113, 'train/grad_norm_pre_clip': 8.029131889343262, 'timing/data': 0.011551336385309696, 'timing/model': 1.1899175150319934, 'learning_rate/qwen_vl_interface': 9.29040545172755e-06, 'learning_rate/action_model': 9.290405451727549e-05, 'epoch': 0.51}) +[2026-06-17 19:04:26,123][starVLA.training.train_starvla][INFO] - Step 1317, Loss: {'train/loss': 0.4058963656425476, 'train/grad_norm_pre_clip': 7.397790431976318, 'timing/data': 0.013427408412098885, 'timing/model': 1.0900502633303404, 'learning_rate/qwen_vl_interface': 9.289267902313486e-06, 'learning_rate/action_model': 9.289267902313487e-05, 'epoch': 0.51}) +[2026-06-17 19:04:31,871][starVLA.training.train_starvla][INFO] - Step 1318, Loss: {'train/loss': 0.42145636677742004, 'train/grad_norm_pre_clip': 10.669011116027832, 'timing/data': 0.01308223232626915, 'timing/model': 1.105645440518856, 'learning_rate/qwen_vl_interface': 9.288129515554843e-06, 'learning_rate/action_model': 9.288129515554843e-05, 'epoch': 0.51}) +[2026-06-17 19:04:36,738][starVLA.training.train_starvla][INFO] - Step 1319, Loss: {'train/loss': 0.4244822561740875, 'train/grad_norm_pre_clip': 7.645445823669434, 'timing/data': 0.013022537343204021, 'timing/model': 1.1286215716972947, 'learning_rate/qwen_vl_interface': 9.286990291687605e-06, 'learning_rate/action_model': 9.286990291687604e-05, 'epoch': 0.51}) +[2026-06-17 19:04:41,708][starVLA.training.train_starvla][INFO] - Step 1320, Loss: {'train/loss': 0.855119526386261, 'train/grad_norm_pre_clip': 14.09846019744873, 'timing/data': 0.012710601091384888, 'timing/model': 1.1629938781261444, 'learning_rate/qwen_vl_interface': 9.285850230947934e-06, 'learning_rate/action_model': 9.285850230947934e-05, 'epoch': 0.51}) +[2026-06-17 19:04:47,626][starVLA.training.train_starvla][INFO] - Step 1321, Loss: {'train/loss': 0.5646265745162964, 'train/grad_norm_pre_clip': 9.314851760864258, 'timing/data': 0.012918893247842789, 'timing/model': 2.7278823228552938, 'learning_rate/qwen_vl_interface': 9.28470933357217e-06, 'learning_rate/action_model': 9.28470933357217e-05, 'epoch': 0.51}) +[2026-06-17 19:04:52,505][starVLA.training.train_starvla][INFO] - Step 1322, Loss: {'train/loss': 0.5798495411872864, 'train/grad_norm_pre_clip': 8.419185638427734, 'timing/data': 0.010910899378359318, 'timing/model': 1.6455067517235875, 'learning_rate/qwen_vl_interface': 9.283567599796819e-06, 'learning_rate/action_model': 9.283567599796818e-05, 'epoch': 0.51}) +[2026-06-17 19:04:56,915][starVLA.training.train_starvla][INFO] - Step 1323, Loss: {'train/loss': 0.7575280070304871, 'train/grad_norm_pre_clip': 6.4787774085998535, 'timing/data': 0.011545085348188877, 'timing/model': 1.1211647111922503, 'learning_rate/qwen_vl_interface': 9.282425029858561e-06, 'learning_rate/action_model': 9.282425029858561e-05, 'epoch': 0.51}) +[2026-06-17 19:05:04,150][starVLA.training.train_starvla][INFO] - Step 1324, Loss: {'train/loss': 0.5528448224067688, 'train/grad_norm_pre_clip': 8.122776985168457, 'timing/data': 0.012789996340870857, 'timing/model': 1.088136164471507, 'learning_rate/qwen_vl_interface': 9.281281623994258e-06, 'learning_rate/action_model': 9.281281623994258e-05, 'epoch': 0.51}) +[2026-06-17 19:05:10,725][starVLA.training.train_starvla][INFO] - Step 1325, Loss: {'train/loss': 0.508490800857544, 'train/grad_norm_pre_clip': 9.54784107208252, 'timing/data': 0.012541285715997219, 'timing/model': 1.0877273892983794, 'learning_rate/qwen_vl_interface': 9.280137382440936e-06, 'learning_rate/action_model': 9.280137382440937e-05, 'epoch': 0.51}) +[2026-06-17 19:05:17,481][starVLA.training.train_starvla][INFO] - Step 1326, Loss: {'train/loss': 0.7268239855766296, 'train/grad_norm_pre_clip': 9.26333999633789, 'timing/data': 0.011387323029339314, 'timing/model': 1.0959346173331141, 'learning_rate/qwen_vl_interface': 9.278992305435798e-06, 'learning_rate/action_model': 9.278992305435797e-05, 'epoch': 0.51}) +[2026-06-17 19:05:22,488][starVLA.training.train_starvla][INFO] - Step 1327, Loss: {'train/loss': 0.8001845479011536, 'train/grad_norm_pre_clip': 6.912156105041504, 'timing/data': 0.011600236408412457, 'timing/model': 1.5421804981306195, 'learning_rate/qwen_vl_interface': 9.27784639321622e-06, 'learning_rate/action_model': 9.27784639321622e-05, 'epoch': 0.51}) +[2026-06-17 19:05:28,619][starVLA.training.train_starvla][INFO] - Step 1328, Loss: {'train/loss': 0.5921321511268616, 'train/grad_norm_pre_clip': 7.966729164123535, 'timing/data': 0.012861323542892933, 'timing/model': 1.7138461591675878, 'learning_rate/qwen_vl_interface': 9.27669964601975e-06, 'learning_rate/action_model': 9.27669964601975e-05, 'epoch': 0.51}) +[2026-06-17 19:05:33,534][starVLA.training.train_starvla][INFO] - Step 1329, Loss: {'train/loss': 0.5336981415748596, 'train/grad_norm_pre_clip': 6.0318098068237305, 'timing/data': 0.011350039392709732, 'timing/model': 1.0948499711230397, 'learning_rate/qwen_vl_interface': 9.27555206408411e-06, 'learning_rate/action_model': 9.275552064084111e-05, 'epoch': 0.51}) +[2026-06-17 19:05:38,393][starVLA.training.train_starvla][INFO] - Step 1330, Loss: {'train/loss': 0.6910914182662964, 'train/grad_norm_pre_clip': 6.809901714324951, 'timing/data': 0.015639829449355602, 'timing/model': 1.1605282174423337, 'learning_rate/qwen_vl_interface': 9.274403647647197e-06, 'learning_rate/action_model': 9.274403647647197e-05, 'epoch': 0.51}) +[2026-06-17 19:05:44,773][starVLA.training.train_starvla][INFO] - Step 1331, Loss: {'train/loss': 0.9440140128135681, 'train/grad_norm_pre_clip': 7.7785258293151855, 'timing/data': 0.010779624804854393, 'timing/model': 1.1068365601822734, 'learning_rate/qwen_vl_interface': 9.273254396947079e-06, 'learning_rate/action_model': 9.273254396947078e-05, 'epoch': 0.51}) +[2026-06-17 19:05:49,761][starVLA.training.train_starvla][INFO] - Step 1332, Loss: {'train/loss': 0.44386717677116394, 'train/grad_norm_pre_clip': 6.363814830780029, 'timing/data': 0.013621375896036625, 'timing/model': 1.7229613568633795, 'learning_rate/qwen_vl_interface': 9.27210431222199e-06, 'learning_rate/action_model': 9.27210431222199e-05, 'epoch': 0.51}) +[2026-06-17 19:05:54,156][starVLA.training.train_starvla][INFO] - Step 1333, Loss: {'train/loss': 0.39924225211143494, 'train/grad_norm_pre_clip': 9.252123832702637, 'timing/data': 0.015130186453461647, 'timing/model': 1.1338348621502519, 'learning_rate/qwen_vl_interface': 9.270953393710353e-06, 'learning_rate/action_model': 9.270953393710352e-05, 'epoch': 0.51}) +[2026-06-17 19:06:00,431][starVLA.training.train_starvla][INFO] - Step 1334, Loss: {'train/loss': 0.7200495600700378, 'train/grad_norm_pre_clip': 8.277538299560547, 'timing/data': 0.0124029740691185, 'timing/model': 1.123988608829677, 'learning_rate/qwen_vl_interface': 9.269801641650751e-06, 'learning_rate/action_model': 9.26980164165075e-05, 'epoch': 0.51}) +[2026-06-17 19:06:04,942][starVLA.training.train_starvla][INFO] - Step 1335, Loss: {'train/loss': 0.3910740613937378, 'train/grad_norm_pre_clip': 5.364421844482422, 'timing/data': 0.012564060278236866, 'timing/model': 1.079462943598628, 'learning_rate/qwen_vl_interface': 9.268649056281941e-06, 'learning_rate/action_model': 9.268649056281941e-05, 'epoch': 0.51}) +[2026-06-17 19:06:09,704][starVLA.training.train_starvla][INFO] - Step 1336, Loss: {'train/loss': 0.6907416582107544, 'train/grad_norm_pre_clip': 7.609375476837158, 'timing/data': 0.01208738423883915, 'timing/model': 1.1719453241676092, 'learning_rate/qwen_vl_interface': 9.267495637842859e-06, 'learning_rate/action_model': 9.267495637842857e-05, 'epoch': 0.51}) +[2026-06-17 19:06:14,461][starVLA.training.train_starvla][INFO] - Step 1337, Loss: {'train/loss': 0.3476967215538025, 'train/grad_norm_pre_clip': 367.8100891113281, 'timing/data': 0.012294990941882133, 'timing/model': 1.478059472516179, 'learning_rate/qwen_vl_interface': 9.266341386572607e-06, 'learning_rate/action_model': 9.266341386572605e-05, 'epoch': 0.51}) +[2026-06-17 19:06:20,272][starVLA.training.train_starvla][INFO] - Step 1338, Loss: {'train/loss': 0.523735761642456, 'train/grad_norm_pre_clip': 8.328949928283691, 'timing/data': 0.01293925940990448, 'timing/model': 2.353526948019862, 'learning_rate/qwen_vl_interface': 9.265186302710465e-06, 'learning_rate/action_model': 9.265186302710465e-05, 'epoch': 0.51}) +[2026-06-17 19:06:25,231][starVLA.training.train_starvla][INFO] - Step 1339, Loss: {'train/loss': 0.42101848125457764, 'train/grad_norm_pre_clip': 5.131523609161377, 'timing/data': 0.011076750233769417, 'timing/model': 1.1438661934807897, 'learning_rate/qwen_vl_interface': 9.26403038649588e-06, 'learning_rate/action_model': 9.26403038649588e-05, 'epoch': 0.52}) +[2026-06-17 19:06:30,021][starVLA.training.train_starvla][INFO] - Step 1340, Loss: {'train/loss': 0.8144128322601318, 'train/grad_norm_pre_clip': 9.500640869140625, 'timing/data': 0.012359067797660828, 'timing/model': 1.0801542699337006, 'learning_rate/qwen_vl_interface': 9.262873638168478e-06, 'learning_rate/action_model': 9.262873638168478e-05, 'epoch': 0.52}) +[2026-06-17 19:06:34,936][starVLA.training.train_starvla][INFO] - Step 1341, Loss: {'train/loss': 0.6268983483314514, 'train/grad_norm_pre_clip': 7.325516223907471, 'timing/data': 0.014029301702976227, 'timing/model': 1.1375067364424467, 'learning_rate/qwen_vl_interface': 9.261716057968054e-06, 'learning_rate/action_model': 9.261716057968053e-05, 'epoch': 0.52}) +[2026-06-17 19:06:39,727][starVLA.training.train_starvla][INFO] - Step 1342, Loss: {'train/loss': 0.695063054561615, 'train/grad_norm_pre_clip': 6.918735027313232, 'timing/data': 0.012944776564836502, 'timing/model': 1.5673943012952805, 'learning_rate/qwen_vl_interface': 9.260557646134572e-06, 'learning_rate/action_model': 9.260557646134572e-05, 'epoch': 0.52}) +[2026-06-17 19:06:44,109][starVLA.training.train_starvla][INFO] - Step 1343, Loss: {'train/loss': 0.6061152219772339, 'train/grad_norm_pre_clip': 9.439746856689453, 'timing/data': 0.015121274627745152, 'timing/model': 1.0976298665627837, 'learning_rate/qwen_vl_interface': 9.259398402908175e-06, 'learning_rate/action_model': 9.259398402908174e-05, 'epoch': 0.52}) +[2026-06-17 19:06:48,637][starVLA.training.train_starvla][INFO] - Step 1344, Loss: {'train/loss': 0.7072158455848694, 'train/grad_norm_pre_clip': 10.285725593566895, 'timing/data': 0.010233649984002113, 'timing/model': 1.077155233360827, 'learning_rate/qwen_vl_interface': 9.258238328529177e-06, 'learning_rate/action_model': 9.258238328529176e-05, 'epoch': 0.52}) +[2026-06-17 19:06:53,341][starVLA.training.train_starvla][INFO] - Step 1345, Loss: {'train/loss': 0.5127639770507812, 'train/grad_norm_pre_clip': 6.531553745269775, 'timing/data': 0.011137955822050571, 'timing/model': 1.0692186830565333, 'learning_rate/qwen_vl_interface': 9.257077423238058e-06, 'learning_rate/action_model': 9.257077423238058e-05, 'epoch': 0.52}) +[2026-06-17 19:06:58,133][starVLA.training.train_starvla][INFO] - Step 1346, Loss: {'train/loss': 0.5693984627723694, 'train/grad_norm_pre_clip': 8.885712623596191, 'timing/data': 0.017024429515004158, 'timing/model': 1.0855676624923944, 'learning_rate/qwen_vl_interface': 9.25591568727548e-06, 'learning_rate/action_model': 9.255915687275478e-05, 'epoch': 0.52}) +[2026-06-17 19:07:02,776][starVLA.training.train_starvla][INFO] - Step 1347, Loss: {'train/loss': 0.5588223934173584, 'train/grad_norm_pre_clip': 6.080527305603027, 'timing/data': 0.011743603274226189, 'timing/model': 1.3800432067364454, 'learning_rate/qwen_vl_interface': 9.254753120882265e-06, 'learning_rate/action_model': 9.254753120882266e-05, 'epoch': 0.52}) +[2026-06-17 19:07:07,245][starVLA.training.train_starvla][INFO] - Step 1348, Loss: {'train/loss': 0.45830604434013367, 'train/grad_norm_pre_clip': 5.718464374542236, 'timing/data': 0.01125876884907484, 'timing/model': 1.1379503086209297, 'learning_rate/qwen_vl_interface': 9.25358972429942e-06, 'learning_rate/action_model': 9.253589724299421e-05, 'epoch': 0.52}) +[2026-06-17 19:07:12,085][starVLA.training.train_starvla][INFO] - Step 1349, Loss: {'train/loss': 0.6418561935424805, 'train/grad_norm_pre_clip': 10.396903991699219, 'timing/data': 0.013106605969369411, 'timing/model': 1.1011552084237337, 'learning_rate/qwen_vl_interface': 9.25242549776812e-06, 'learning_rate/action_model': 9.252425497768119e-05, 'epoch': 0.52}) +[2026-06-17 19:07:16,639][starVLA.training.train_starvla][INFO] - Step 1350, Loss: {'train/loss': 0.8804283142089844, 'train/grad_norm_pre_clip': 10.95661735534668, 'timing/data': 0.01341680996119976, 'timing/model': 1.0978022012859583, 'learning_rate/qwen_vl_interface': 9.251260441529706e-06, 'learning_rate/action_model': 9.251260441529705e-05, 'epoch': 0.52}) +[2026-06-17 19:07:21,325][starVLA.training.train_starvla][INFO] - Step 1351, Loss: {'train/loss': 0.7372916340827942, 'train/grad_norm_pre_clip': 14.890093803405762, 'timing/data': 0.012025776319205761, 'timing/model': 1.1055464399978518, 'learning_rate/qwen_vl_interface': 9.250094555825697e-06, 'learning_rate/action_model': 9.250094555825696e-05, 'epoch': 0.52}) +[2026-06-17 19:07:25,896][starVLA.training.train_starvla][INFO] - Step 1352, Loss: {'train/loss': 0.5385767221450806, 'train/grad_norm_pre_clip': 6.3538689613342285, 'timing/data': 0.011644955724477768, 'timing/model': 1.4495082180947065, 'learning_rate/qwen_vl_interface': 9.248927840897782e-06, 'learning_rate/action_model': 9.248927840897781e-05, 'epoch': 0.52}) +[2026-06-17 19:07:30,201][starVLA.training.train_starvla][INFO] - Step 1353, Loss: {'train/loss': 0.8810745477676392, 'train/grad_norm_pre_clip': 15.679337501525879, 'timing/data': 0.013233713805675507, 'timing/model': 1.1272613313049078, 'learning_rate/qwen_vl_interface': 9.247760296987822e-06, 'learning_rate/action_model': 9.247760296987821e-05, 'epoch': 0.52}) +[2026-06-17 19:07:34,796][starVLA.training.train_starvla][INFO] - Step 1354, Loss: {'train/loss': 0.3175995945930481, 'train/grad_norm_pre_clip': 5.084665298461914, 'timing/data': 0.010865346528589725, 'timing/model': 1.1778651429340243, 'learning_rate/qwen_vl_interface': 9.246591924337852e-06, 'learning_rate/action_model': 9.246591924337852e-05, 'epoch': 0.52}) +[2026-06-17 19:07:39,487][starVLA.training.train_starvla][INFO] - Step 1355, Loss: {'train/loss': 0.7541965842247009, 'train/grad_norm_pre_clip': 8.267645835876465, 'timing/data': 0.016363156959414482, 'timing/model': 1.113994436338544, 'learning_rate/qwen_vl_interface': 9.245422723190076e-06, 'learning_rate/action_model': 9.245422723190074e-05, 'epoch': 0.52}) +[2026-06-17 19:07:44,180][starVLA.training.train_starvla][INFO] - Step 1356, Loss: {'train/loss': 0.6696326732635498, 'train/grad_norm_pre_clip': 7.347282886505127, 'timing/data': 0.011127011850476265, 'timing/model': 1.2097708312794566, 'learning_rate/qwen_vl_interface': 9.244252693786868e-06, 'learning_rate/action_model': 9.244252693786869e-05, 'epoch': 0.52}) +[2026-06-17 19:07:48,979][starVLA.training.train_starvla][INFO] - Step 1357, Loss: {'train/loss': 0.49436983466148376, 'train/grad_norm_pre_clip': 5.687870979309082, 'timing/data': 0.013497929088771343, 'timing/model': 1.508213410153985, 'learning_rate/qwen_vl_interface': 9.24308183637078e-06, 'learning_rate/action_model': 9.24308183637078e-05, 'epoch': 0.52}) +[2026-06-17 19:07:53,365][starVLA.training.train_starvla][INFO] - Step 1358, Loss: {'train/loss': 0.30252769589424133, 'train/grad_norm_pre_clip': 5.413684844970703, 'timing/data': 0.013576108030974865, 'timing/model': 1.1464598262682557, 'learning_rate/qwen_vl_interface': 9.24191015118453e-06, 'learning_rate/action_model': 9.24191015118453e-05, 'epoch': 0.52}) +[2026-06-17 19:08:00,190][starVLA.training.train_starvla][INFO] - Step 1359, Loss: {'train/loss': 0.5313495993614197, 'train/grad_norm_pre_clip': 7.011152744293213, 'timing/data': 0.012540880590677261, 'timing/model': 1.1614709859713912, 'learning_rate/qwen_vl_interface': 9.24073763847101e-06, 'learning_rate/action_model': 9.24073763847101e-05, 'epoch': 0.52}) +[2026-06-17 19:08:05,057][starVLA.training.train_starvla][INFO] - Step 1360, Loss: {'train/loss': 0.5241135358810425, 'train/grad_norm_pre_clip': 7.807741165161133, 'timing/data': 0.013856373727321625, 'timing/model': 1.0877070352435112, 'learning_rate/qwen_vl_interface': 9.239564298473284e-06, 'learning_rate/action_model': 9.239564298473284e-05, 'epoch': 0.52}) +[2026-06-17 19:08:09,793][starVLA.training.train_starvla][INFO] - Step 1361, Loss: {'train/loss': 0.49413928389549255, 'train/grad_norm_pre_clip': 9.275788307189941, 'timing/data': 0.011192130856215954, 'timing/model': 1.1446036398410797, 'learning_rate/qwen_vl_interface': 9.238390131434586e-06, 'learning_rate/action_model': 9.238390131434586e-05, 'epoch': 0.52}) +[2026-06-17 19:08:17,218][starVLA.training.train_starvla][INFO] - Step 1362, Loss: {'train/loss': 0.9437286257743835, 'train/grad_norm_pre_clip': 9.281978607177734, 'timing/data': 0.011870319955050945, 'timing/model': 1.4959493419155478, 'learning_rate/qwen_vl_interface': 9.237215137598322e-06, 'learning_rate/action_model': 9.237215137598322e-05, 'epoch': 0.52}) +[2026-06-17 19:08:21,851][starVLA.training.train_starvla][INFO] - Step 1363, Loss: {'train/loss': 0.7382844090461731, 'train/grad_norm_pre_clip': 6.915082931518555, 'timing/data': 0.017959301359951496, 'timing/model': 1.2465682281181216, 'learning_rate/qwen_vl_interface': 9.236039317208068e-06, 'learning_rate/action_model': 9.236039317208068e-05, 'epoch': 0.52}) +[2026-06-17 19:08:26,554][starVLA.training.train_starvla][INFO] - Step 1364, Loss: {'train/loss': 0.84088134765625, 'train/grad_norm_pre_clip': 7.388839244842529, 'timing/data': 0.017011286690831184, 'timing/model': 1.1855362746864557, 'learning_rate/qwen_vl_interface': 9.234862670507577e-06, 'learning_rate/action_model': 9.234862670507576e-05, 'epoch': 0.52}) +[2026-06-17 19:08:31,457][starVLA.training.train_starvla][INFO] - Step 1365, Loss: {'train/loss': 0.2616586685180664, 'train/grad_norm_pre_clip': 4.186104774475098, 'timing/data': 0.014408156275749207, 'timing/model': 1.1302420636639, 'learning_rate/qwen_vl_interface': 9.233685197740765e-06, 'learning_rate/action_model': 9.233685197740764e-05, 'epoch': 0.53}) +[2026-06-17 19:08:36,127][starVLA.training.train_starvla][INFO] - Step 1366, Loss: {'train/loss': 0.33103811740875244, 'train/grad_norm_pre_clip': 6.503146648406982, 'timing/data': 0.011276115663349628, 'timing/model': 1.0688137700781226, 'learning_rate/qwen_vl_interface': 9.232506899151725e-06, 'learning_rate/action_model': 9.232506899151725e-05, 'epoch': 0.53}) +[2026-06-17 19:08:40,873][starVLA.training.train_starvla][INFO] - Step 1367, Loss: {'train/loss': 0.5725383758544922, 'train/grad_norm_pre_clip': 6.032171249389648, 'timing/data': 0.01384369470179081, 'timing/model': 1.417116948403418, 'learning_rate/qwen_vl_interface': 9.23132777498472e-06, 'learning_rate/action_model': 9.231327774984719e-05, 'epoch': 0.53}) +[2026-06-17 19:08:45,228][starVLA.training.train_starvla][INFO] - Step 1368, Loss: {'train/loss': 1.0230212211608887, 'train/grad_norm_pre_clip': 11.526152610778809, 'timing/data': 0.013401512987911701, 'timing/model': 1.072914899326861, 'learning_rate/qwen_vl_interface': 9.230147825484182e-06, 'learning_rate/action_model': 9.230147825484183e-05, 'epoch': 0.53}) +[2026-06-17 19:08:49,744][starVLA.training.train_starvla][INFO] - Step 1369, Loss: {'train/loss': 0.5147428512573242, 'train/grad_norm_pre_clip': 9.261585235595703, 'timing/data': 0.013731148093938828, 'timing/model': 1.0716505274176598, 'learning_rate/qwen_vl_interface': 9.228967050894717e-06, 'learning_rate/action_model': 9.228967050894716e-05, 'epoch': 0.53}) +[2026-06-17 19:08:54,344][starVLA.training.train_starvla][INFO] - Step 1370, Loss: {'train/loss': 0.8115258812904358, 'train/grad_norm_pre_clip': 10.175410270690918, 'timing/data': 0.014323691837489605, 'timing/model': 1.1226935079321265, 'learning_rate/qwen_vl_interface': 9.227785451461103e-06, 'learning_rate/action_model': 9.227785451461102e-05, 'epoch': 0.53}) +[2026-06-17 19:08:59,354][starVLA.training.train_starvla][INFO] - Step 1371, Loss: {'train/loss': 0.4000924229621887, 'train/grad_norm_pre_clip': 6.501712799072266, 'timing/data': 0.012114935554564, 'timing/model': 1.2297264207154512, 'learning_rate/qwen_vl_interface': 9.226603027428281e-06, 'learning_rate/action_model': 9.226603027428281e-05, 'epoch': 0.53}) +[2026-06-17 19:09:04,135][starVLA.training.train_starvla][INFO] - Step 1372, Loss: {'train/loss': 0.5092350244522095, 'train/grad_norm_pre_clip': 8.766271591186523, 'timing/data': 0.010788192972540855, 'timing/model': 1.547367854975164, 'learning_rate/qwen_vl_interface': 9.225419779041376e-06, 'learning_rate/action_model': 9.225419779041376e-05, 'epoch': 0.53}) +[2026-06-17 19:09:08,400][starVLA.training.train_starvla][INFO] - Step 1373, Loss: {'train/loss': 0.6463620662689209, 'train/grad_norm_pre_clip': 6.760365962982178, 'timing/data': 0.0113868098706007, 'timing/model': 1.08634281065315, 'learning_rate/qwen_vl_interface': 9.224235706545672e-06, 'learning_rate/action_model': 9.224235706545671e-05, 'epoch': 0.53}) +[2026-06-17 19:09:15,194][starVLA.training.train_starvla][INFO] - Step 1374, Loss: {'train/loss': 0.6160488724708557, 'train/grad_norm_pre_clip': 6.090738773345947, 'timing/data': 0.01256694458425045, 'timing/model': 1.2052196664735675, 'learning_rate/qwen_vl_interface': 9.22305081018663e-06, 'learning_rate/action_model': 9.223050810186629e-05, 'epoch': 0.53}) +[2026-06-17 19:09:19,802][starVLA.training.train_starvla][INFO] - Step 1375, Loss: {'train/loss': 0.35012173652648926, 'train/grad_norm_pre_clip': 9.736720085144043, 'timing/data': 0.011864867992699146, 'timing/model': 1.069659006781876, 'learning_rate/qwen_vl_interface': 9.221865090209878e-06, 'learning_rate/action_model': 9.221865090209878e-05, 'epoch': 0.53}) +[2026-06-17 19:09:24,576][starVLA.training.train_starvla][INFO] - Step 1376, Loss: {'train/loss': 0.7093712687492371, 'train/grad_norm_pre_clip': 8.954609870910645, 'timing/data': 0.012775170616805553, 'timing/model': 1.1444515157490969, 'learning_rate/qwen_vl_interface': 9.22067854686122e-06, 'learning_rate/action_model': 9.22067854686122e-05, 'epoch': 0.53}) +[2026-06-17 19:09:31,081][starVLA.training.train_starvla][INFO] - Step 1377, Loss: {'train/loss': 0.7347147464752197, 'train/grad_norm_pre_clip': 11.849583625793457, 'timing/data': 0.011452979408204556, 'timing/model': 1.3815580336377025, 'learning_rate/qwen_vl_interface': 9.219491180386626e-06, 'learning_rate/action_model': 9.219491180386626e-05, 'epoch': 0.53}) +[2026-06-17 19:09:35,527][starVLA.training.train_starvla][INFO] - Step 1378, Loss: {'train/loss': 0.49461814761161804, 'train/grad_norm_pre_clip': 7.1287102699279785, 'timing/data': 0.018662932328879833, 'timing/model': 1.0882595432922244, 'learning_rate/qwen_vl_interface': 9.21830299103224e-06, 'learning_rate/action_model': 9.21830299103224e-05, 'epoch': 0.53}) +[2026-06-17 19:09:40,347][starVLA.training.train_starvla][INFO] - Step 1379, Loss: {'train/loss': 0.5512117147445679, 'train/grad_norm_pre_clip': 7.682737827301025, 'timing/data': 0.01153079979121685, 'timing/model': 1.1210154285654426, 'learning_rate/qwen_vl_interface': 9.217113979044373e-06, 'learning_rate/action_model': 9.217113979044373e-05, 'epoch': 0.53}) +[2026-06-17 19:09:45,071][starVLA.training.train_starvla][INFO] - Step 1380, Loss: {'train/loss': 0.5305910110473633, 'train/grad_norm_pre_clip': 6.4656982421875, 'timing/data': 0.012477641925215721, 'timing/model': 1.1976632429286838, 'learning_rate/qwen_vl_interface': 9.215924144669509e-06, 'learning_rate/action_model': 9.215924144669509e-05, 'epoch': 0.53}) +[2026-06-17 19:09:49,817][starVLA.training.train_starvla][INFO] - Step 1381, Loss: {'train/loss': 0.6725062131881714, 'train/grad_norm_pre_clip': 7.366610527038574, 'timing/data': 0.011600323021411896, 'timing/model': 1.1065255850553513, 'learning_rate/qwen_vl_interface': 9.214733488154304e-06, 'learning_rate/action_model': 9.214733488154304e-05, 'epoch': 0.53}) +[2026-06-17 19:09:54,527][starVLA.training.train_starvla][INFO] - Step 1382, Loss: {'train/loss': 0.5181318521499634, 'train/grad_norm_pre_clip': 8.856511116027832, 'timing/data': 0.01327043678611517, 'timing/model': 1.414676021784544, 'learning_rate/qwen_vl_interface': 9.213542009745579e-06, 'learning_rate/action_model': 9.213542009745578e-05, 'epoch': 0.53}) +[2026-06-17 19:09:58,876][starVLA.training.train_starvla][INFO] - Step 1383, Loss: {'train/loss': 0.40493378043174744, 'train/grad_norm_pre_clip': 7.080763339996338, 'timing/data': 0.017189601436257362, 'timing/model': 1.1668971544131637, 'learning_rate/qwen_vl_interface': 9.21234970969033e-06, 'learning_rate/action_model': 9.21234970969033e-05, 'epoch': 0.53}) +[2026-06-17 19:10:03,623][starVLA.training.train_starvla][INFO] - Step 1384, Loss: {'train/loss': 0.695522129535675, 'train/grad_norm_pre_clip': 8.108552932739258, 'timing/data': 0.011889957822859287, 'timing/model': 1.1377757610753179, 'learning_rate/qwen_vl_interface': 9.211156588235724e-06, 'learning_rate/action_model': 9.211156588235725e-05, 'epoch': 0.53}) +[2026-06-17 19:10:08,362][starVLA.training.train_starvla][INFO] - Step 1385, Loss: {'train/loss': 0.5771239995956421, 'train/grad_norm_pre_clip': 6.96627950668335, 'timing/data': 0.012653330340981483, 'timing/model': 1.136188367381692, 'learning_rate/qwen_vl_interface': 9.209962645629095e-06, 'learning_rate/action_model': 9.209962645629095e-05, 'epoch': 0.53}) +[2026-06-17 19:10:13,004][starVLA.training.train_starvla][INFO] - Step 1386, Loss: {'train/loss': 0.4169386625289917, 'train/grad_norm_pre_clip': 5.355342864990234, 'timing/data': 0.01176756713539362, 'timing/model': 1.0854046875610948, 'learning_rate/qwen_vl_interface': 9.208767882117947e-06, 'learning_rate/action_model': 9.208767882117947e-05, 'epoch': 0.53}) +[2026-06-17 19:10:17,744][starVLA.training.train_starvla][INFO] - Step 1387, Loss: {'train/loss': 0.5194580554962158, 'train/grad_norm_pre_clip': 5.70192289352417, 'timing/data': 0.012223714031279087, 'timing/model': 1.548660115338862, 'learning_rate/qwen_vl_interface': 9.207572297949958e-06, 'learning_rate/action_model': 9.207572297949958e-05, 'epoch': 0.53}) +[2026-06-17 19:10:22,089][starVLA.training.train_starvla][INFO] - Step 1388, Loss: {'train/loss': 0.32244545221328735, 'train/grad_norm_pre_clip': 6.354922294616699, 'timing/data': 0.015523471869528294, 'timing/model': 1.1322982218116522, 'learning_rate/qwen_vl_interface': 9.206375893372973e-06, 'learning_rate/action_model': 9.206375893372973e-05, 'epoch': 0.53}) +[2026-06-17 19:10:27,828][starVLA.training.train_starvla][INFO] - Step 1389, Loss: {'train/loss': 0.4286295175552368, 'train/grad_norm_pre_clip': 5.939134120941162, 'timing/data': 0.012516296468675137, 'timing/model': 1.1145900851115584, 'learning_rate/qwen_vl_interface': 9.20517866863501e-06, 'learning_rate/action_model': 9.205178668635009e-05, 'epoch': 0.53}) +[2026-06-17 19:10:32,432][starVLA.training.train_starvla][INFO] - Step 1390, Loss: {'train/loss': 0.8269827961921692, 'train/grad_norm_pre_clip': 10.889657020568848, 'timing/data': 0.01254157442599535, 'timing/model': 1.0881458511576056, 'learning_rate/qwen_vl_interface': 9.20398062398425e-06, 'learning_rate/action_model': 9.20398062398425e-05, 'epoch': 0.53}) +[2026-06-17 19:10:39,719][starVLA.training.train_starvla][INFO] - Step 1391, Loss: {'train/loss': 1.0038095712661743, 'train/grad_norm_pre_clip': 11.266901969909668, 'timing/data': 0.012896998785436153, 'timing/model': 1.1231834888458252, 'learning_rate/qwen_vl_interface': 9.202781759669053e-06, 'learning_rate/action_model': 9.202781759669052e-05, 'epoch': 0.54}) +[2026-06-17 19:10:44,760][starVLA.training.train_starvla][INFO] - Step 1392, Loss: {'train/loss': 0.31653857231140137, 'train/grad_norm_pre_clip': 5.044044494628906, 'timing/data': 0.01325805764645338, 'timing/model': 1.4779869550839067, 'learning_rate/qwen_vl_interface': 9.201582075937943e-06, 'learning_rate/action_model': 9.201582075937944e-05, 'epoch': 0.54}) +[2026-06-17 19:10:49,087][starVLA.training.train_starvla][INFO] - Step 1393, Loss: {'train/loss': 1.0463706254959106, 'train/grad_norm_pre_clip': 14.981457710266113, 'timing/data': 0.012494510039687157, 'timing/model': 1.1001954060047865, 'learning_rate/qwen_vl_interface': 9.200381573039619e-06, 'learning_rate/action_model': 9.200381573039619e-05, 'epoch': 0.54}) +[2026-06-17 19:10:53,856][starVLA.training.train_starvla][INFO] - Step 1394, Loss: {'train/loss': 0.38798093795776367, 'train/grad_norm_pre_clip': 6.562918663024902, 'timing/data': 0.011564644053578377, 'timing/model': 1.0756963286548853, 'learning_rate/qwen_vl_interface': 9.199180251222942e-06, 'learning_rate/action_model': 9.199180251222942e-05, 'epoch': 0.54}) +[2026-06-17 19:10:58,475][starVLA.training.train_starvla][INFO] - Step 1395, Loss: {'train/loss': 0.5595471858978271, 'train/grad_norm_pre_clip': 8.500879287719727, 'timing/data': 0.012750213034451008, 'timing/model': 1.0814391877502203, 'learning_rate/qwen_vl_interface': 9.197978110736951e-06, 'learning_rate/action_model': 9.19797811073695e-05, 'epoch': 0.54}) +[2026-06-17 19:11:02,988][starVLA.training.train_starvla][INFO] - Step 1396, Loss: {'train/loss': 0.2980419993400574, 'train/grad_norm_pre_clip': 5.366885662078857, 'timing/data': 0.011546721681952477, 'timing/model': 1.0720940930768847, 'learning_rate/qwen_vl_interface': 9.196775151830849e-06, 'learning_rate/action_model': 9.196775151830849e-05, 'epoch': 0.54}) +[2026-06-17 19:11:07,746][starVLA.training.train_starvla][INFO] - Step 1397, Loss: {'train/loss': 0.5242218375205994, 'train/grad_norm_pre_clip': 6.446330547332764, 'timing/data': 0.012876691296696663, 'timing/model': 1.433178186416626, 'learning_rate/qwen_vl_interface': 9.19557137475401e-06, 'learning_rate/action_model': 9.19557137475401e-05, 'epoch': 0.54}) +[2026-06-17 19:11:12,268][starVLA.training.train_starvla][INFO] - Step 1398, Loss: {'train/loss': 0.6567580103874207, 'train/grad_norm_pre_clip': 8.965721130371094, 'timing/data': 0.013790732249617577, 'timing/model': 1.1374329403042793, 'learning_rate/qwen_vl_interface': 9.19436677975598e-06, 'learning_rate/action_model': 9.19436677975598e-05, 'epoch': 0.54}) +[2026-06-17 19:11:16,976][starVLA.training.train_starvla][INFO] - Step 1399, Loss: {'train/loss': 0.6213505864143372, 'train/grad_norm_pre_clip': 10.337366104125977, 'timing/data': 0.012453613802790642, 'timing/model': 1.089122025296092, 'learning_rate/qwen_vl_interface': 9.193161367086472e-06, 'learning_rate/action_model': 9.193161367086472e-05, 'epoch': 0.54}) +[2026-06-17 19:13:48,889][starVLA.training.train_starvla][INFO] - Step 1400, Loss: {'train/loss': 0.5811259150505066, 'train/grad_norm_pre_clip': 10.114307403564453, 'eval/loss': 0.6061790585517883, 'eval/latency_4/loss': 0.6061790585517883, 'eval/demon_fire_precision': 0.9508760951188986, 'eval/demon_fire_recall': 0.9870087690808704, 'eval/demon_fire_f1': 0.9686055776892429, 'eval/demon_fire_rate_model': 0.9386196769456682, 'eval/demon_fire_rate_teacher': 0.9042584434654919, 'eval/demon_left_precision': 0.7526132404181185, 'eval/demon_left_recall': 0.4931506849315068, 'eval/demon_left_f1': 0.5958620689655173, 'eval/demon_left_rate_model': 0.084287812041116, 'eval/demon_left_rate_teacher': 0.12863436123348018, 'eval/demon_right_precision': 0.7112810707456979, 'eval/demon_right_recall': 0.6714801444043321, 'eval/demon_right_f1': 0.6908077994428969, 'eval/demon_right_rate_model': 0.15359765051395008, 'eval/demon_right_rate_teacher': 0.16270190895741557, 'eval/demon_move_f1': 0.6433349342042072, 'eval/demon_cc_f1': 0.8410369037391937, 'eval/latency_4/demon_fire_precision': 0.9508760951188986, 'eval/latency_4/demon_fire_recall': 0.9870087690808704, 'eval/latency_4/demon_fire_f1': 0.9686055776892429, 'eval/latency_4/demon_fire_rate_model': 0.9386196769456682, 'eval/latency_4/demon_fire_rate_teacher': 0.9042584434654919, 'eval/latency_4/demon_left_precision': 0.7526132404181185, 'eval/latency_4/demon_left_recall': 0.4931506849315068, 'eval/latency_4/demon_left_f1': 0.5958620689655173, 'eval/latency_4/demon_left_rate_model': 0.084287812041116, 'eval/latency_4/demon_left_rate_teacher': 0.12863436123348018, 'eval/latency_4/demon_right_precision': 0.7112810707456979, 'eval/latency_4/demon_right_recall': 0.6714801444043321, 'eval/latency_4/demon_right_f1': 0.6908077994428969, 'eval/latency_4/demon_right_rate_model': 0.15359765051395008, 'eval/latency_4/demon_right_rate_teacher': 0.16270190895741557, 'eval/latency_4/demon_move_f1': 0.6433349342042072, 'eval/latency_4/demon_cc_f1': 0.8410369037391937, 'timing/data': 0.013460428453981876, 'timing/model': 1.1424736948683858, 'learning_rate/qwen_vl_interface': 9.191955136995371e-06, 'learning_rate/action_model': 9.191955136995371e-05, 'epoch': 0.54}) +[2026-06-17 19:13:48,913][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_1400_state +[2026-06-17 19:13:48,921][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 19:13:57,873][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_1400_state/model.safetensors +[2026-06-17 19:14:12,968][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_1400_state/optimizer.bin +[2026-06-17 19:14:12,993][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_1400_state/sampler.bin +[2026-06-17 19:14:12,994][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_1400_state/sampler_1.bin +[2026-06-17 19:14:12,996][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_1400_state/random_states_0.pkl +[2026-06-17 19:14:13,002][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1400 +[2026-06-17 19:14:13,003][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 19:14:13,021][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 19:14:22,004][starVLA.training.train_starvla][INFO] - Step 1401, Loss: {'train/loss': 0.5343899130821228, 'train/grad_norm_pre_clip': 6.515172481536865, 'timing/data': 0.015530915930867195, 'timing/model': 4.463937263935804, 'learning_rate/qwen_vl_interface': 9.190748089732727e-06, 'learning_rate/action_model': 9.190748089732728e-05, 'epoch': 0.54}) +[2026-06-17 19:14:26,346][starVLA.training.train_starvla][INFO] - Step 1402, Loss: {'train/loss': 0.4133569300174713, 'train/grad_norm_pre_clip': 6.832847595214844, 'timing/data': 0.015450787730515003, 'timing/model': 1.1510888934135437, 'learning_rate/qwen_vl_interface': 9.189540225548763e-06, 'learning_rate/action_model': 9.189540225548764e-05, 'epoch': 0.54}) +[2026-06-17 19:14:31,458][starVLA.training.train_starvla][INFO] - Step 1403, Loss: {'train/loss': 0.6912562847137451, 'train/grad_norm_pre_clip': 8.906015396118164, 'timing/data': 0.012374510988593102, 'timing/model': 1.2386593399569392, 'learning_rate/qwen_vl_interface': 9.188331544693872e-06, 'learning_rate/action_model': 9.188331544693873e-05, 'epoch': 0.54}) +[2026-06-17 19:14:36,357][starVLA.training.train_starvla][INFO] - Step 1404, Loss: {'train/loss': 0.4753977954387665, 'train/grad_norm_pre_clip': 9.012114524841309, 'timing/data': 0.012046249583363533, 'timing/model': 1.2220936920493841, 'learning_rate/qwen_vl_interface': 9.187122047418614e-06, 'learning_rate/action_model': 9.187122047418614e-05, 'epoch': 0.54}) +[2026-06-17 19:14:41,006][starVLA.training.train_starvla][INFO] - Step 1405, Loss: {'train/loss': 0.6793303489685059, 'train/grad_norm_pre_clip': 10.891294479370117, 'timing/data': 0.010686111636459827, 'timing/model': 1.0842777537181973, 'learning_rate/qwen_vl_interface': 9.18591173397372e-06, 'learning_rate/action_model': 9.185911733973719e-05, 'epoch': 0.54}) +[2026-06-17 19:14:45,975][starVLA.training.train_starvla][INFO] - Step 1406, Loss: {'train/loss': 0.812842845916748, 'train/grad_norm_pre_clip': 10.475250244140625, 'timing/data': 0.04198158532381058, 'timing/model': 1.2366400705650449, 'learning_rate/qwen_vl_interface': 9.184700604610087e-06, 'learning_rate/action_model': 9.184700604610085e-05, 'epoch': 0.54}) +[2026-06-17 19:14:50,691][starVLA.training.train_starvla][INFO] - Step 1407, Loss: {'train/loss': 0.799826979637146, 'train/grad_norm_pre_clip': 7.762750625610352, 'timing/data': 0.013208878226578236, 'timing/model': 1.4922911981120706, 'learning_rate/qwen_vl_interface': 9.183488659578786e-06, 'learning_rate/action_model': 9.183488659578785e-05, 'epoch': 0.54}) +[2026-06-17 19:14:55,008][starVLA.training.train_starvla][INFO] - Step 1408, Loss: {'train/loss': 0.25901633501052856, 'train/grad_norm_pre_clip': 3.414754867553711, 'timing/data': 0.013573938049376011, 'timing/model': 1.079208035953343, 'learning_rate/qwen_vl_interface': 9.18227589913105e-06, 'learning_rate/action_model': 9.18227589913105e-05, 'epoch': 0.54}) +[2026-06-17 19:14:59,855][starVLA.training.train_starvla][INFO] - Step 1409, Loss: {'train/loss': 0.8662942051887512, 'train/grad_norm_pre_clip': 10.51795768737793, 'timing/data': 0.012013217434287071, 'timing/model': 1.09136757068336, 'learning_rate/qwen_vl_interface': 9.181062323518293e-06, 'learning_rate/action_model': 9.181062323518292e-05, 'epoch': 0.54}) +[2026-06-17 19:15:04,663][starVLA.training.train_starvla][INFO] - Step 1410, Loss: {'train/loss': 0.43052130937576294, 'train/grad_norm_pre_clip': 7.552095413208008, 'timing/data': 0.013614162802696228, 'timing/model': 1.0687759229913354, 'learning_rate/qwen_vl_interface': 9.179847932992084e-06, 'learning_rate/action_model': 9.179847932992083e-05, 'epoch': 0.54}) +[2026-06-17 19:15:09,429][starVLA.training.train_starvla][INFO] - Step 1411, Loss: {'train/loss': 0.4876307547092438, 'train/grad_norm_pre_clip': 7.142054080963135, 'timing/data': 0.014842637814581394, 'timing/model': 1.177178479731083, 'learning_rate/qwen_vl_interface': 9.17863272780417e-06, 'learning_rate/action_model': 9.178632727804168e-05, 'epoch': 0.54}) +[2026-06-17 19:15:14,161][starVLA.training.train_starvla][INFO] - Step 1412, Loss: {'train/loss': 0.2648288905620575, 'train/grad_norm_pre_clip': 4.387040615081787, 'timing/data': 0.015804781578481197, 'timing/model': 1.4912741528823972, 'learning_rate/qwen_vl_interface': 9.177416708206463e-06, 'learning_rate/action_model': 9.177416708206463e-05, 'epoch': 0.54}) +[2026-06-17 19:15:24,453][starVLA.training.train_starvla][INFO] - Step 1413, Loss: {'train/loss': 0.7414684295654297, 'train/grad_norm_pre_clip': 9.60681438446045, 'timing/data': 0.0167276319116354, 'timing/model': 1.3003353336825967, 'learning_rate/qwen_vl_interface': 9.176199874451049e-06, 'learning_rate/action_model': 9.176199874451048e-05, 'epoch': 0.54}) +[2026-06-17 19:15:29,113][starVLA.training.train_starvla][INFO] - Step 1414, Loss: {'train/loss': 0.6598329544067383, 'train/grad_norm_pre_clip': 7.6015520095825195, 'timing/data': 0.01272798702120781, 'timing/model': 1.0943297613412142, 'learning_rate/qwen_vl_interface': 9.174982226790175e-06, 'learning_rate/action_model': 9.174982226790175e-05, 'epoch': 0.54}) +[2026-06-17 19:15:33,767][starVLA.training.train_starvla][INFO] - Step 1415, Loss: {'train/loss': 0.4147647023200989, 'train/grad_norm_pre_clip': 6.4881415367126465, 'timing/data': 0.017050391994416714, 'timing/model': 1.0565440375357866, 'learning_rate/qwen_vl_interface': 9.173763765476262e-06, 'learning_rate/action_model': 9.17376376547626e-05, 'epoch': 0.54}) +[2026-06-17 19:15:38,360][starVLA.training.train_starvla][INFO] - Step 1416, Loss: {'train/loss': 0.45118841528892517, 'train/grad_norm_pre_clip': 7.556483745574951, 'timing/data': 0.011715568602085114, 'timing/model': 1.0718239163979888, 'learning_rate/qwen_vl_interface': 9.172544490761897e-06, 'learning_rate/action_model': 9.172544490761897e-05, 'epoch': 0.54}) +[2026-06-17 19:15:42,597][starVLA.training.train_starvla][INFO] - Step 1417, Loss: {'train/loss': 0.743193507194519, 'train/grad_norm_pre_clip': 10.436383247375488, 'timing/data': 0.013251133263111115, 'timing/model': 1.0733378138393164, 'learning_rate/qwen_vl_interface': 9.171324402899841e-06, 'learning_rate/action_model': 9.17132440289984e-05, 'epoch': 0.55}) +[2026-06-17 19:15:47,462][starVLA.training.train_starvla][INFO] - Step 1418, Loss: {'train/loss': 0.5665028095245361, 'train/grad_norm_pre_clip': 10.493967056274414, 'timing/data': 0.013918491080403328, 'timing/model': 1.273707939311862, 'learning_rate/qwen_vl_interface': 9.170103502143014e-06, 'learning_rate/action_model': 9.170103502143013e-05, 'epoch': 0.55}) +[2026-06-17 19:15:52,026][starVLA.training.train_starvla][INFO] - Step 1419, Loss: {'train/loss': 0.5563821792602539, 'train/grad_norm_pre_clip': 9.468755722045898, 'timing/data': 0.011824989691376686, 'timing/model': 1.1137538393959403, 'learning_rate/qwen_vl_interface': 9.168881788744513e-06, 'learning_rate/action_model': 9.168881788744513e-05, 'epoch': 0.55}) +[2026-06-17 19:15:56,826][starVLA.training.train_starvla][INFO] - Step 1420, Loss: {'train/loss': 0.6036268472671509, 'train/grad_norm_pre_clip': 6.471362113952637, 'timing/data': 0.01335088536143303, 'timing/model': 1.078292878344655, 'learning_rate/qwen_vl_interface': 9.1676592629576e-06, 'learning_rate/action_model': 9.167659262957601e-05, 'epoch': 0.55}) +[2026-06-17 19:16:01,689][starVLA.training.train_starvla][INFO] - Step 1421, Loss: {'train/loss': 0.5642667412757874, 'train/grad_norm_pre_clip': 6.263047218322754, 'timing/data': 0.011679954826831818, 'timing/model': 1.5007802471518517, 'learning_rate/qwen_vl_interface': 9.16643592503571e-06, 'learning_rate/action_model': 9.166435925035709e-05, 'epoch': 0.55}) +[2026-06-17 19:16:06,444][starVLA.training.train_starvla][INFO] - Step 1422, Loss: {'train/loss': 0.5947996377944946, 'train/grad_norm_pre_clip': 8.223623275756836, 'timing/data': 0.015859761275351048, 'timing/model': 1.528090718202293, 'learning_rate/qwen_vl_interface': 9.165211775232436e-06, 'learning_rate/action_model': 9.165211775232435e-05, 'epoch': 0.55}) +[2026-06-17 19:16:10,769][starVLA.training.train_starvla][INFO] - Step 1423, Loss: {'train/loss': 0.3955853283405304, 'train/grad_norm_pre_clip': 7.550525665283203, 'timing/data': 0.01234210655093193, 'timing/model': 1.0889451308175921, 'learning_rate/qwen_vl_interface': 9.163986813801548e-06, 'learning_rate/action_model': 9.163986813801548e-05, 'epoch': 0.55}) +[2026-06-17 19:16:15,454][starVLA.training.train_starvla][INFO] - Step 1424, Loss: {'train/loss': 0.4674777090549469, 'train/grad_norm_pre_clip': 7.666658401489258, 'timing/data': 0.012103408575057983, 'timing/model': 1.1879605576395988, 'learning_rate/qwen_vl_interface': 9.162761040996979e-06, 'learning_rate/action_model': 9.162761040996978e-05, 'epoch': 0.55}) +[2026-06-17 19:16:20,364][starVLA.training.train_starvla][INFO] - Step 1425, Loss: {'train/loss': 0.4319794178009033, 'train/grad_norm_pre_clip': 7.2265625, 'timing/data': 0.013742947019636631, 'timing/model': 1.085901734419167, 'learning_rate/qwen_vl_interface': 9.161534457072839e-06, 'learning_rate/action_model': 9.161534457072839e-05, 'epoch': 0.55}) +[2026-06-17 19:16:24,963][starVLA.training.train_starvla][INFO] - Step 1426, Loss: {'train/loss': 0.7175350189208984, 'train/grad_norm_pre_clip': 10.426929473876953, 'timing/data': 0.012672481127083302, 'timing/model': 1.095920380204916, 'learning_rate/qwen_vl_interface': 9.160307062283397e-06, 'learning_rate/action_model': 9.160307062283397e-05, 'epoch': 0.55}) +[2026-06-17 19:16:29,826][starVLA.training.train_starvla][INFO] - Step 1427, Loss: {'train/loss': 0.4881618618965149, 'train/grad_norm_pre_clip': 7.226873874664307, 'timing/data': 0.01095860730856657, 'timing/model': 1.5635630758479238, 'learning_rate/qwen_vl_interface': 9.159078856883093e-06, 'learning_rate/action_model': 9.159078856883093e-05, 'epoch': 0.55}) +[2026-06-17 19:16:34,070][starVLA.training.train_starvla][INFO] - Step 1428, Loss: {'train/loss': 0.7658511996269226, 'train/grad_norm_pre_clip': 9.884069442749023, 'timing/data': 0.011782856658101082, 'timing/model': 1.0879260636866093, 'learning_rate/qwen_vl_interface': 9.157849841126535e-06, 'learning_rate/action_model': 9.157849841126535e-05, 'epoch': 0.55}) +[2026-06-17 19:16:40,502][starVLA.training.train_starvla][INFO] - Step 1429, Loss: {'train/loss': 0.4650513529777527, 'train/grad_norm_pre_clip': 6.558887958526611, 'timing/data': 0.018560167402029037, 'timing/model': 1.1782422373071313, 'learning_rate/qwen_vl_interface': 9.156620015268502e-06, 'learning_rate/action_model': 9.1566200152685e-05, 'epoch': 0.55}) +[2026-06-17 19:16:45,201][starVLA.training.train_starvla][INFO] - Step 1430, Loss: {'train/loss': 0.37733420729637146, 'train/grad_norm_pre_clip': 6.596944808959961, 'timing/data': 0.011442876420915127, 'timing/model': 1.120658260770142, 'learning_rate/qwen_vl_interface': 9.155389379563934e-06, 'learning_rate/action_model': 9.155389379563933e-05, 'epoch': 0.55}) +[2026-06-17 19:16:49,769][starVLA.training.train_starvla][INFO] - Step 1431, Loss: {'train/loss': 0.41336286067962646, 'train/grad_norm_pre_clip': 8.060736656188965, 'timing/data': 0.010800899006426334, 'timing/model': 1.0739402705803514, 'learning_rate/qwen_vl_interface': 9.154157934267946e-06, 'learning_rate/action_model': 9.154157934267944e-05, 'epoch': 0.55}) +[2026-06-17 19:16:54,366][starVLA.training.train_starvla][INFO] - Step 1432, Loss: {'train/loss': 0.6991517543792725, 'train/grad_norm_pre_clip': 9.905169486999512, 'timing/data': 0.012275275774300098, 'timing/model': 1.41207124479115, 'learning_rate/qwen_vl_interface': 9.152925679635817e-06, 'learning_rate/action_model': 9.152925679635817e-05, 'epoch': 0.55}) +[2026-06-17 19:16:58,855][starVLA.training.train_starvla][INFO] - Step 1433, Loss: {'train/loss': 0.5421286225318909, 'train/grad_norm_pre_clip': 7.008180618286133, 'timing/data': 0.01197572611272335, 'timing/model': 1.2552309669554234, 'learning_rate/qwen_vl_interface': 9.151692615922995e-06, 'learning_rate/action_model': 9.151692615922994e-05, 'epoch': 0.55}) +[2026-06-17 19:17:03,590][starVLA.training.train_starvla][INFO] - Step 1434, Loss: {'train/loss': 0.3716394007205963, 'train/grad_norm_pre_clip': 5.292535781860352, 'timing/data': 0.013302912935614586, 'timing/model': 1.0766222542151809, 'learning_rate/qwen_vl_interface': 9.150458743385095e-06, 'learning_rate/action_model': 9.150458743385094e-05, 'epoch': 0.55}) +[2026-06-17 19:17:11,823][starVLA.training.train_starvla][INFO] - Step 1435, Loss: {'train/loss': 0.511634886264801, 'train/grad_norm_pre_clip': 6.195682525634766, 'timing/data': 0.015186641365289688, 'timing/model': 3.3526611728593707, 'learning_rate/qwen_vl_interface': 9.1492240622779e-06, 'learning_rate/action_model': 9.1492240622779e-05, 'epoch': 0.55}) +[2026-06-17 19:17:16,450][starVLA.training.train_starvla][INFO] - Step 1436, Loss: {'train/loss': 0.572887659072876, 'train/grad_norm_pre_clip': 7.284160614013672, 'timing/data': 0.011286864057183266, 'timing/model': 1.0871188007295132, 'learning_rate/qwen_vl_interface': 9.147988572857362e-06, 'learning_rate/action_model': 9.147988572857362e-05, 'epoch': 0.55}) +[2026-06-17 19:17:24,779][starVLA.training.train_starvla][INFO] - Step 1437, Loss: {'train/loss': 0.6558599472045898, 'train/grad_norm_pre_clip': 8.134347915649414, 'timing/data': 0.016055690124630928, 'timing/model': 4.039060038514435, 'learning_rate/qwen_vl_interface': 9.146752275379599e-06, 'learning_rate/action_model': 9.146752275379598e-05, 'epoch': 0.55}) +[2026-06-17 19:17:29,112][starVLA.training.train_starvla][INFO] - Step 1438, Loss: {'train/loss': 0.48297128081321716, 'train/grad_norm_pre_clip': 7.850893020629883, 'timing/data': 0.012226441875100136, 'timing/model': 1.1288237133994699, 'learning_rate/qwen_vl_interface': 9.145515170100894e-06, 'learning_rate/action_model': 9.145515170100894e-05, 'epoch': 0.55}) +[2026-06-17 19:17:33,861][starVLA.training.train_starvla][INFO] - Step 1439, Loss: {'train/loss': 0.4251899719238281, 'train/grad_norm_pre_clip': 12.588825225830078, 'timing/data': 0.01250442024320364, 'timing/model': 1.126232652924955, 'learning_rate/qwen_vl_interface': 9.144277257277705e-06, 'learning_rate/action_model': 9.144277257277703e-05, 'epoch': 0.55}) +[2026-06-17 19:17:38,586][starVLA.training.train_starvla][INFO] - Step 1440, Loss: {'train/loss': 0.5375844836235046, 'train/grad_norm_pre_clip': 9.23192310333252, 'timing/data': 0.011648518033325672, 'timing/model': 1.2094268174842, 'learning_rate/qwen_vl_interface': 9.14303853716665e-06, 'learning_rate/action_model': 9.143038537166649e-05, 'epoch': 0.55}) +[2026-06-17 19:17:43,250][starVLA.training.train_starvla][INFO] - Step 1441, Loss: {'train/loss': 0.5690770745277405, 'train/grad_norm_pre_clip': 8.307211875915527, 'timing/data': 0.014272592961788177, 'timing/model': 1.0652707098051906, 'learning_rate/qwen_vl_interface': 9.141799010024516e-06, 'learning_rate/action_model': 9.141799010024515e-05, 'epoch': 0.55}) +[2026-06-17 19:17:47,839][starVLA.training.train_starvla][INFO] - Step 1442, Loss: {'train/loss': 0.5648683309555054, 'train/grad_norm_pre_clip': 8.089380264282227, 'timing/data': 0.012351672165095806, 'timing/model': 1.3969605779275298, 'learning_rate/qwen_vl_interface': 9.140558676108258e-06, 'learning_rate/action_model': 9.140558676108259e-05, 'epoch': 0.55}) +[2026-06-17 19:17:52,090][starVLA.training.train_starvla][INFO] - Step 1443, Loss: {'train/loss': 0.5179274082183838, 'train/grad_norm_pre_clip': 8.608680725097656, 'timing/data': 0.011597195640206337, 'timing/model': 1.0651943199336529, 'learning_rate/qwen_vl_interface': 9.139317535675002e-06, 'learning_rate/action_model': 9.139317535675002e-05, 'epoch': 0.56}) +[2026-06-17 19:17:56,953][starVLA.training.train_starvla][INFO] - Step 1444, Loss: {'train/loss': 0.8559888005256653, 'train/grad_norm_pre_clip': 11.682662010192871, 'timing/data': 0.012160496786236763, 'timing/model': 1.0859517119824886, 'learning_rate/qwen_vl_interface': 9.138075588982035e-06, 'learning_rate/action_model': 9.138075588982036e-05, 'epoch': 0.56}) +[2026-06-17 19:18:01,492][starVLA.training.train_starvla][INFO] - Step 1445, Loss: {'train/loss': 0.8358203768730164, 'train/grad_norm_pre_clip': 9.206136703491211, 'timing/data': 0.011547275818884373, 'timing/model': 1.1338910376653075, 'learning_rate/qwen_vl_interface': 9.136832836286816e-06, 'learning_rate/action_model': 9.136832836286816e-05, 'epoch': 0.56}) +[2026-06-17 19:18:06,108][starVLA.training.train_starvla][INFO] - Step 1446, Loss: {'train/loss': 0.7090938687324524, 'train/grad_norm_pre_clip': 9.536288261413574, 'timing/data': 0.010986343026161194, 'timing/model': 1.0917471321299672, 'learning_rate/qwen_vl_interface': 9.135589277846967e-06, 'learning_rate/action_model': 9.135589277846966e-05, 'epoch': 0.56}) +[2026-06-17 19:18:10,843][starVLA.training.train_starvla][INFO] - Step 1447, Loss: {'train/loss': 0.6470459699630737, 'train/grad_norm_pre_clip': 6.904921531677246, 'timing/data': 0.013238567858934402, 'timing/model': 1.5107785575091839, 'learning_rate/qwen_vl_interface': 9.134344913920278e-06, 'learning_rate/action_model': 9.134344913920278e-05, 'epoch': 0.56}) +[2026-06-17 19:18:15,143][starVLA.training.train_starvla][INFO] - Step 1448, Loss: {'train/loss': 0.5201696157455444, 'train/grad_norm_pre_clip': 8.575536727905273, 'timing/data': 0.013200510293245316, 'timing/model': 1.1444574389606714, 'learning_rate/qwen_vl_interface': 9.133099744764711e-06, 'learning_rate/action_model': 9.133099744764711e-05, 'epoch': 0.56}) +[2026-06-17 19:18:19,855][starVLA.training.train_starvla][INFO] - Step 1449, Loss: {'train/loss': 0.42078763246536255, 'train/grad_norm_pre_clip': 6.020438194274902, 'timing/data': 0.013039784505963326, 'timing/model': 1.1155203534290195, 'learning_rate/qwen_vl_interface': 9.131853770638386e-06, 'learning_rate/action_model': 9.131853770638386e-05, 'epoch': 0.56}) +[2026-06-17 19:18:27,810][starVLA.training.train_starvla][INFO] - Step 1450, Loss: {'train/loss': 0.5779252052307129, 'train/grad_norm_pre_clip': 9.464463233947754, 'timing/data': 0.011898553930222988, 'timing/model': 1.068684047088027, 'learning_rate/qwen_vl_interface': 9.130606991799599e-06, 'learning_rate/action_model': 9.130606991799599e-05, 'epoch': 0.56}) +[2026-06-17 19:18:32,480][starVLA.training.train_starvla][INFO] - Step 1451, Loss: {'train/loss': 0.41512545943260193, 'train/grad_norm_pre_clip': 6.210628032684326, 'timing/data': 0.02044757455587387, 'timing/model': 1.1433117985725403, 'learning_rate/qwen_vl_interface': 9.129359408506808e-06, 'learning_rate/action_model': 9.129359408506808e-05, 'epoch': 0.56}) +[2026-06-17 19:18:37,332][starVLA.training.train_starvla][INFO] - Step 1452, Loss: {'train/loss': 0.5999597311019897, 'train/grad_norm_pre_clip': 7.737216472625732, 'timing/data': 0.010851315222680569, 'timing/model': 1.642749764956534, 'learning_rate/qwen_vl_interface': 9.128111021018636e-06, 'learning_rate/action_model': 9.128111021018636e-05, 'epoch': 0.56}) +[2026-06-17 19:18:42,022][starVLA.training.train_starvla][INFO] - Step 1453, Loss: {'train/loss': 0.814122200012207, 'train/grad_norm_pre_clip': 8.106218338012695, 'timing/data': 0.02230720967054367, 'timing/model': 1.1531490562483668, 'learning_rate/qwen_vl_interface': 9.126861829593876e-06, 'learning_rate/action_model': 9.126861829593876e-05, 'epoch': 0.56}) +[2026-06-17 19:18:46,693][starVLA.training.train_starvla][INFO] - Step 1454, Loss: {'train/loss': 0.63019859790802, 'train/grad_norm_pre_clip': 10.22640609741211, 'timing/data': 0.013708515092730522, 'timing/model': 1.1183398831635714, 'learning_rate/qwen_vl_interface': 9.125611834491487e-06, 'learning_rate/action_model': 9.125611834491486e-05, 'epoch': 0.56}) +[2026-06-17 19:18:51,253][starVLA.training.train_starvla][INFO] - Step 1455, Loss: {'train/loss': 0.6176106333732605, 'train/grad_norm_pre_clip': 9.704357147216797, 'timing/data': 0.011842417530715466, 'timing/model': 1.0669918805360794, 'learning_rate/qwen_vl_interface': 9.124361035970594e-06, 'learning_rate/action_model': 9.124361035970594e-05, 'epoch': 0.56}) +[2026-06-17 19:18:56,934][starVLA.training.train_starvla][INFO] - Step 1456, Loss: {'train/loss': 0.35974758863449097, 'train/grad_norm_pre_clip': 5.283785343170166, 'timing/data': 0.013519945554435253, 'timing/model': 2.0824601789936423, 'learning_rate/qwen_vl_interface': 9.123109434290491e-06, 'learning_rate/action_model': 9.123109434290492e-05, 'epoch': 0.56}) +[2026-06-17 19:19:01,673][starVLA.training.train_starvla][INFO] - Step 1457, Loss: {'train/loss': 0.6623581051826477, 'train/grad_norm_pre_clip': 9.165385246276855, 'timing/data': 0.01194941345602274, 'timing/model': 1.4528941856697202, 'learning_rate/qwen_vl_interface': 9.121857029710632e-06, 'learning_rate/action_model': 9.121857029710633e-05, 'epoch': 0.56}) +[2026-06-17 19:19:08,321][starVLA.training.train_starvla][INFO] - Step 1458, Loss: {'train/loss': 0.4469168782234192, 'train/grad_norm_pre_clip': 5.32361364364624, 'timing/data': 0.018348868004977703, 'timing/model': 1.1195427123457193, 'learning_rate/qwen_vl_interface': 9.120603822490647e-06, 'learning_rate/action_model': 9.120603822490647e-05, 'epoch': 0.56}) +[2026-06-17 19:19:14,256][starVLA.training.train_starvla][INFO] - Step 1459, Loss: {'train/loss': 0.8588316440582275, 'train/grad_norm_pre_clip': 11.438008308410645, 'timing/data': 0.01194710936397314, 'timing/model': 1.0829884726554155, 'learning_rate/qwen_vl_interface': 9.119349812890323e-06, 'learning_rate/action_model': 9.119349812890323e-05, 'epoch': 0.56}) +[2026-06-17 19:19:18,933][starVLA.training.train_starvla][INFO] - Step 1460, Loss: {'train/loss': 0.5947669744491577, 'train/grad_norm_pre_clip': 6.358133316040039, 'timing/data': 0.012944526970386505, 'timing/model': 1.1407127194106579, 'learning_rate/qwen_vl_interface': 9.11809500116962e-06, 'learning_rate/action_model': 9.11809500116962e-05, 'epoch': 0.56}) +[2026-06-17 19:19:23,393][starVLA.training.train_starvla][INFO] - Step 1461, Loss: {'train/loss': 0.5972380042076111, 'train/grad_norm_pre_clip': 6.494372844696045, 'timing/data': 0.011482035741209984, 'timing/model': 1.0751492083072662, 'learning_rate/qwen_vl_interface': 9.116839387588662e-06, 'learning_rate/action_model': 9.116839387588661e-05, 'epoch': 0.56}) +[2026-06-17 19:19:28,032][starVLA.training.train_starvla][INFO] - Step 1462, Loss: {'train/loss': 0.6899412274360657, 'train/grad_norm_pre_clip': 6.601815223693848, 'timing/data': 0.01433853805065155, 'timing/model': 1.4766836185008287, 'learning_rate/qwen_vl_interface': 9.115582972407735e-06, 'learning_rate/action_model': 9.115582972407736e-05, 'epoch': 0.56}) +[2026-06-17 19:19:33,592][starVLA.training.train_starvla][INFO] - Step 1463, Loss: {'train/loss': 0.4501914083957672, 'train/grad_norm_pre_clip': 7.923220634460449, 'timing/data': 0.013715721666812897, 'timing/model': 1.1152856508269906, 'learning_rate/qwen_vl_interface': 9.114325755887301e-06, 'learning_rate/action_model': 9.114325755887301e-05, 'epoch': 0.56}) +[2026-06-17 19:19:39,855][starVLA.training.train_starvla][INFO] - Step 1464, Loss: {'train/loss': 0.6232378482818604, 'train/grad_norm_pre_clip': 6.454967498779297, 'timing/data': 0.012819459661841393, 'timing/model': 1.1280536660924554, 'learning_rate/qwen_vl_interface': 9.113067738287978e-06, 'learning_rate/action_model': 9.113067738287979e-05, 'epoch': 0.56}) +[2026-06-17 19:19:44,945][starVLA.training.train_starvla][INFO] - Step 1465, Loss: {'train/loss': 0.7253103852272034, 'train/grad_norm_pre_clip': 6.8875651359558105, 'timing/data': 0.011938192881643772, 'timing/model': 1.5648063765838742, 'learning_rate/qwen_vl_interface': 9.111808919870557e-06, 'learning_rate/action_model': 9.111808919870557e-05, 'epoch': 0.56}) +[2026-06-17 19:19:49,761][starVLA.training.train_starvla][INFO] - Step 1466, Loss: {'train/loss': 0.434916615486145, 'train/grad_norm_pre_clip': 5.0809221267700195, 'timing/data': 0.011346114799380302, 'timing/model': 1.0642850939184427, 'learning_rate/qwen_vl_interface': 9.11054930089599e-06, 'learning_rate/action_model': 9.110549300895991e-05, 'epoch': 0.56}) +[2026-06-17 19:19:54,334][starVLA.training.train_starvla][INFO] - Step 1467, Loss: {'train/loss': 0.6031894683837891, 'train/grad_norm_pre_clip': 6.2477593421936035, 'timing/data': 0.019248710945248604, 'timing/model': 1.4311474291607738, 'learning_rate/qwen_vl_interface': 9.109288881625402e-06, 'learning_rate/action_model': 9.109288881625402e-05, 'epoch': 0.56}) +[2026-06-17 19:19:58,663][starVLA.training.train_starvla][INFO] - Step 1468, Loss: {'train/loss': 0.740366518497467, 'train/grad_norm_pre_clip': 9.494053840637207, 'timing/data': 0.016265593469142914, 'timing/model': 1.0780606707558036, 'learning_rate/qwen_vl_interface': 9.108027662320073e-06, 'learning_rate/action_model': 9.108027662320073e-05, 'epoch': 0.56}) +[2026-06-17 19:20:03,358][starVLA.training.train_starvla][INFO] - Step 1469, Loss: {'train/loss': 0.7239282131195068, 'train/grad_norm_pre_clip': 10.972959518432617, 'timing/data': 0.011024686507880688, 'timing/model': 1.0632225768640637, 'learning_rate/qwen_vl_interface': 9.10676564324146e-06, 'learning_rate/action_model': 9.10676564324146e-05, 'epoch': 0.57}) +[2026-06-17 19:20:07,903][starVLA.training.train_starvla][INFO] - Step 1470, Loss: {'train/loss': 0.5336514711380005, 'train/grad_norm_pre_clip': 7.0638203620910645, 'timing/data': 0.012440224178135395, 'timing/model': 1.0868275444954634, 'learning_rate/qwen_vl_interface': 9.105502824651179e-06, 'learning_rate/action_model': 9.105502824651178e-05, 'epoch': 0.57}) +[2026-06-17 19:20:14,942][starVLA.training.train_starvla][INFO] - Step 1471, Loss: {'train/loss': 0.35398685932159424, 'train/grad_norm_pre_clip': 22.079593658447266, 'timing/data': 0.011999625712633133, 'timing/model': 3.538929576985538, 'learning_rate/qwen_vl_interface': 9.104239206811012e-06, 'learning_rate/action_model': 9.104239206811011e-05, 'epoch': 0.57}) +[2026-06-17 19:20:19,674][starVLA.training.train_starvla][INFO] - Step 1472, Loss: {'train/loss': 0.9925426244735718, 'train/grad_norm_pre_clip': 8.802938461303711, 'timing/data': 0.010791092179715633, 'timing/model': 1.4677824405953288, 'learning_rate/qwen_vl_interface': 9.102974789982911e-06, 'learning_rate/action_model': 9.102974789982912e-05, 'epoch': 0.57}) +[2026-06-17 19:20:23,909][starVLA.training.train_starvla][INFO] - Step 1473, Loss: {'train/loss': 0.7133750915527344, 'train/grad_norm_pre_clip': 9.473508834838867, 'timing/data': 0.012434291653335094, 'timing/model': 1.0616285083815455, 'learning_rate/qwen_vl_interface': 9.10170957442899e-06, 'learning_rate/action_model': 9.101709574428991e-05, 'epoch': 0.57}) +[2026-06-17 19:20:28,540][starVLA.training.train_starvla][INFO] - Step 1474, Loss: {'train/loss': 0.6282762289047241, 'train/grad_norm_pre_clip': 7.361161708831787, 'timing/data': 0.011068793013691902, 'timing/model': 1.0567945996299386, 'learning_rate/qwen_vl_interface': 9.10044356041153e-06, 'learning_rate/action_model': 9.10044356041153e-05, 'epoch': 0.57}) +[2026-06-17 19:20:33,179][starVLA.training.train_starvla][INFO] - Step 1475, Loss: {'train/loss': 0.5472376942634583, 'train/grad_norm_pre_clip': 5.600241184234619, 'timing/data': 0.018978344276547432, 'timing/model': 1.1203176649287343, 'learning_rate/qwen_vl_interface': 9.099176748192976e-06, 'learning_rate/action_model': 9.099176748192976e-05, 'epoch': 0.57}) +[2026-06-17 19:20:40,558][starVLA.training.train_starvla][INFO] - Step 1476, Loss: {'train/loss': 0.640002429485321, 'train/grad_norm_pre_clip': 9.46513557434082, 'timing/data': 0.013656245544552803, 'timing/model': 1.0692124059423804, 'learning_rate/qwen_vl_interface': 9.097909138035943e-06, 'learning_rate/action_model': 9.097909138035941e-05, 'epoch': 0.57}) +[2026-06-17 19:20:45,259][starVLA.training.train_starvla][INFO] - Step 1477, Loss: {'train/loss': 0.841301441192627, 'train/grad_norm_pre_clip': 9.943588256835938, 'timing/data': 0.011514495126903057, 'timing/model': 1.455918911844492, 'learning_rate/qwen_vl_interface': 9.096640730203202e-06, 'learning_rate/action_model': 9.096640730203202e-05, 'epoch': 0.57}) +[2026-06-17 19:20:49,426][starVLA.training.train_starvla][INFO] - Step 1478, Loss: {'train/loss': 0.6159250736236572, 'train/grad_norm_pre_clip': 7.125080108642578, 'timing/data': 0.013090978376567364, 'timing/model': 1.0977010363712907, 'learning_rate/qwen_vl_interface': 9.095371524957698e-06, 'learning_rate/action_model': 9.095371524957698e-05, 'epoch': 0.57}) +[2026-06-17 19:20:53,956][starVLA.training.train_starvla][INFO] - Step 1479, Loss: {'train/loss': 0.8830966353416443, 'train/grad_norm_pre_clip': 15.972596168518066, 'timing/data': 0.011925769038498402, 'timing/model': 1.0570760918781161, 'learning_rate/qwen_vl_interface': 9.094101522562541e-06, 'learning_rate/action_model': 9.09410152256254e-05, 'epoch': 0.57}) +[2026-06-17 19:20:58,444][starVLA.training.train_starvla][INFO] - Step 1480, Loss: {'train/loss': 0.6024139523506165, 'train/grad_norm_pre_clip': 10.91199016571045, 'timing/data': 0.012101437896490097, 'timing/model': 1.0769163239747286, 'learning_rate/qwen_vl_interface': 9.092830723281e-06, 'learning_rate/action_model': 9.092830723281e-05, 'epoch': 0.57}) +[2026-06-17 19:21:03,216][starVLA.training.train_starvla][INFO] - Step 1481, Loss: {'train/loss': 0.9326831698417664, 'train/grad_norm_pre_clip': 12.421435356140137, 'timing/data': 0.011203973554074764, 'timing/model': 1.127810226753354, 'learning_rate/qwen_vl_interface': 9.091559127376519e-06, 'learning_rate/action_model': 9.091559127376517e-05, 'epoch': 0.57}) +[2026-06-17 19:21:08,013][starVLA.training.train_starvla][INFO] - Step 1482, Loss: {'train/loss': 0.7028058767318726, 'train/grad_norm_pre_clip': 11.2113037109375, 'timing/data': 0.01230289414525032, 'timing/model': 1.54304473195225, 'learning_rate/qwen_vl_interface': 9.090286735112693e-06, 'learning_rate/action_model': 9.090286735112693e-05, 'epoch': 0.57}) +[2026-06-17 19:21:12,551][starVLA.training.train_starvla][INFO] - Step 1483, Loss: {'train/loss': 0.3929213881492615, 'train/grad_norm_pre_clip': 6.14021635055542, 'timing/data': 0.01291684526950121, 'timing/model': 1.2383290929719806, 'learning_rate/qwen_vl_interface': 9.089013546753296e-06, 'learning_rate/action_model': 9.089013546753295e-05, 'epoch': 0.57}) +[2026-06-17 19:21:17,099][starVLA.training.train_starvla][INFO] - Step 1484, Loss: {'train/loss': 0.6843807697296143, 'train/grad_norm_pre_clip': 7.918624401092529, 'timing/data': 0.01268829870969057, 'timing/model': 1.1113252639770508, 'learning_rate/qwen_vl_interface': 9.08773956256226e-06, 'learning_rate/action_model': 9.087739562562259e-05, 'epoch': 0.57}) +[2026-06-17 19:21:21,663][starVLA.training.train_starvla][INFO] - Step 1485, Loss: {'train/loss': 0.63346928358078, 'train/grad_norm_pre_clip': 8.287952423095703, 'timing/data': 0.013016636483371258, 'timing/model': 1.095397101715207, 'learning_rate/qwen_vl_interface': 9.086464782803682e-06, 'learning_rate/action_model': 9.086464782803682e-05, 'epoch': 0.57}) +[2026-06-17 19:21:26,436][starVLA.training.train_starvla][INFO] - Step 1486, Loss: {'train/loss': 0.4110022783279419, 'train/grad_norm_pre_clip': 5.927463531494141, 'timing/data': 0.011439220979809761, 'timing/model': 1.1128026153892279, 'learning_rate/qwen_vl_interface': 9.085189207741827e-06, 'learning_rate/action_model': 9.085189207741827e-05, 'epoch': 0.57}) +[2026-06-17 19:21:34,209][starVLA.training.train_starvla][INFO] - Step 1487, Loss: {'train/loss': 0.49578502774238586, 'train/grad_norm_pre_clip': 5.229997158050537, 'timing/data': 0.01146265771239996, 'timing/model': 4.584163763560355, 'learning_rate/qwen_vl_interface': 9.083912837641123e-06, 'learning_rate/action_model': 9.083912837641123e-05, 'epoch': 0.57}) +[2026-06-17 19:21:38,597][starVLA.training.train_starvla][INFO] - Step 1488, Loss: {'train/loss': 0.49829915165901184, 'train/grad_norm_pre_clip': 8.945036888122559, 'timing/data': 0.013675995171070099, 'timing/model': 1.1732342755421996, 'learning_rate/qwen_vl_interface': 9.082635672766162e-06, 'learning_rate/action_model': 9.082635672766161e-05, 'epoch': 0.57}) +[2026-06-17 19:21:43,132][starVLA.training.train_starvla][INFO] - Step 1489, Loss: {'train/loss': 0.5924174189567566, 'train/grad_norm_pre_clip': 7.402153015136719, 'timing/data': 0.012436476536095142, 'timing/model': 1.0686141671612859, 'learning_rate/qwen_vl_interface': 9.081357713381702e-06, 'learning_rate/action_model': 9.081357713381701e-05, 'epoch': 0.57}) +[2026-06-17 19:21:47,683][starVLA.training.train_starvla][INFO] - Step 1490, Loss: {'train/loss': 0.8099464178085327, 'train/grad_norm_pre_clip': 9.036421775817871, 'timing/data': 0.012768194079399109, 'timing/model': 1.084124262444675, 'learning_rate/qwen_vl_interface': 9.080078959752664e-06, 'learning_rate/action_model': 9.080078959752664e-05, 'epoch': 0.57}) +[2026-06-17 19:21:52,259][starVLA.training.train_starvla][INFO] - Step 1491, Loss: {'train/loss': 0.7228524684906006, 'train/grad_norm_pre_clip': 10.521851539611816, 'timing/data': 0.012944994494318962, 'timing/model': 1.0642300974577665, 'learning_rate/qwen_vl_interface': 9.078799412144137e-06, 'learning_rate/action_model': 9.078799412144137e-05, 'epoch': 0.57}) +[2026-06-17 19:21:56,986][starVLA.training.train_starvla][INFO] - Step 1492, Loss: {'train/loss': 0.7244673371315002, 'train/grad_norm_pre_clip': 11.361642837524414, 'timing/data': 0.011787290684878826, 'timing/model': 1.499169884249568, 'learning_rate/qwen_vl_interface': 9.077519070821373e-06, 'learning_rate/action_model': 9.077519070821373e-05, 'epoch': 0.57}) +[2026-06-17 19:22:01,172][starVLA.training.train_starvla][INFO] - Step 1493, Loss: {'train/loss': 0.7588801383972168, 'train/grad_norm_pre_clip': 8.762362480163574, 'timing/data': 0.01159476675093174, 'timing/model': 1.0646378276869655, 'learning_rate/qwen_vl_interface': 9.076237936049787e-06, 'learning_rate/action_model': 9.076237936049786e-05, 'epoch': 0.57}) +[2026-06-17 19:22:05,839][starVLA.training.train_starvla][INFO] - Step 1494, Loss: {'train/loss': 0.44842395186424255, 'train/grad_norm_pre_clip': 6.264894485473633, 'timing/data': 0.012656140141189098, 'timing/model': 1.0880684154108167, 'learning_rate/qwen_vl_interface': 9.074956008094958e-06, 'learning_rate/action_model': 9.074956008094958e-05, 'epoch': 0.57}) +[2026-06-17 19:22:10,694][starVLA.training.train_starvla][INFO] - Step 1495, Loss: {'train/loss': 0.5016108155250549, 'train/grad_norm_pre_clip': 7.221200942993164, 'timing/data': 0.01109075453132391, 'timing/model': 1.124326091259718, 'learning_rate/qwen_vl_interface': 9.073673287222634e-06, 'learning_rate/action_model': 9.073673287222633e-05, 'epoch': 0.58}) +[2026-06-17 19:22:15,392][starVLA.training.train_starvla][INFO] - Step 1496, Loss: {'train/loss': 0.5920820236206055, 'train/grad_norm_pre_clip': 7.8924665451049805, 'timing/data': 0.014378279447555542, 'timing/model': 1.0744997458532453, 'learning_rate/qwen_vl_interface': 9.072389773698723e-06, 'learning_rate/action_model': 9.072389773698723e-05, 'epoch': 0.58}) +[2026-06-17 19:22:20,126][starVLA.training.train_starvla][INFO] - Step 1497, Loss: {'train/loss': 0.6044760346412659, 'train/grad_norm_pre_clip': 7.819275379180908, 'timing/data': 0.011979778297245502, 'timing/model': 1.5424451110884547, 'learning_rate/qwen_vl_interface': 9.071105467789299e-06, 'learning_rate/action_model': 9.071105467789298e-05, 'epoch': 0.58}) +[2026-06-17 19:22:24,458][starVLA.training.train_starvla][INFO] - Step 1498, Loss: {'train/loss': 0.4194454550743103, 'train/grad_norm_pre_clip': 6.487653732299805, 'timing/data': 0.016367007046937943, 'timing/model': 1.1507408348843455, 'learning_rate/qwen_vl_interface': 9.0698203697606e-06, 'learning_rate/action_model': 9.0698203697606e-05, 'epoch': 0.58}) +[2026-06-17 19:22:29,231][starVLA.training.train_starvla][INFO] - Step 1499, Loss: {'train/loss': 0.4632417559623718, 'train/grad_norm_pre_clip': 12.876334190368652, 'timing/data': 0.012865356169641018, 'timing/model': 1.0865289280191064, 'learning_rate/qwen_vl_interface': 9.068534479879026e-06, 'learning_rate/action_model': 9.068534479879024e-05, 'epoch': 0.58}) +[2026-06-17 19:25:36,454][starVLA.training.train_starvla][INFO] - Step 1500, Loss: {'train/loss': 0.8625784516334534, 'train/grad_norm_pre_clip': 9.302940368652344, 'eval/loss': 0.5702466368675232, 'eval/latency_4/loss': 0.5702466368675232, 'eval/demon_fire_precision': 0.9465224111282844, 'eval/demon_fire_recall': 0.9944787268593699, 'eval/demon_fire_f1': 0.9699081406398481, 'eval/demon_fire_rate_model': 0.9500734214390602, 'eval/demon_fire_rate_teacher': 0.9042584434654919, 'eval/demon_left_precision': 0.8471074380165289, 'eval/demon_left_recall': 0.4680365296803653, 'eval/demon_left_f1': 0.6029411764705882, 'eval/demon_left_rate_model': 0.07107195301027901, 'eval/demon_left_rate_teacher': 0.12863436123348018, 'eval/demon_right_precision': 0.7343453510436433, 'eval/demon_right_recall': 0.6985559566787004, 'eval/demon_right_f1': 0.7160037002775208, 'eval/demon_right_rate_model': 0.15477239353891337, 'eval/demon_right_rate_teacher': 0.16270190895741557, 'eval/demon_move_f1': 0.6594724383740544, 'eval/demon_cc_f1': 0.8498870717067494, 'eval/latency_4/demon_fire_precision': 0.9465224111282844, 'eval/latency_4/demon_fire_recall': 0.9944787268593699, 'eval/latency_4/demon_fire_f1': 0.9699081406398481, 'eval/latency_4/demon_fire_rate_model': 0.9500734214390602, 'eval/latency_4/demon_fire_rate_teacher': 0.9042584434654919, 'eval/latency_4/demon_left_precision': 0.8471074380165289, 'eval/latency_4/demon_left_recall': 0.4680365296803653, 'eval/latency_4/demon_left_f1': 0.6029411764705882, 'eval/latency_4/demon_left_rate_model': 0.07107195301027901, 'eval/latency_4/demon_left_rate_teacher': 0.12863436123348018, 'eval/latency_4/demon_right_precision': 0.7343453510436433, 'eval/latency_4/demon_right_recall': 0.6985559566787004, 'eval/latency_4/demon_right_f1': 0.7160037002775208, 'eval/latency_4/demon_right_rate_model': 0.15477239353891337, 'eval/latency_4/demon_right_rate_teacher': 0.16270190895741557, 'eval/latency_4/demon_move_f1': 0.6594724383740544, 'eval/latency_4/demon_cc_f1': 0.8498870717067494, 'timing/data': 0.011120660230517387, 'timing/model': 1.094914955087006, 'learning_rate/qwen_vl_interface': 9.067247798411147e-06, 'learning_rate/action_model': 9.067247798411147e-05, 'epoch': 0.58}) +[2026-06-17 19:25:36,456][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_1500_state +[2026-06-17 19:25:36,463][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 19:25:45,343][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_1500_state/model.safetensors +[2026-06-17 19:26:08,197][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_1500_state/optimizer.bin +[2026-06-17 19:26:08,198][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_1500_state/sampler.bin +[2026-06-17 19:26:08,199][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_1500_state/sampler_1.bin +[2026-06-17 19:26:08,202][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_1500_state/random_states_0.pkl +[2026-06-17 19:26:08,207][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1500 +[2026-06-17 19:26:08,208][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 19:26:08,226][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 19:26:13,169][starVLA.training.train_starvla][INFO] - Step 1501, Loss: {'train/loss': 0.6534186005592346, 'train/grad_norm_pre_clip': 6.423663139343262, 'timing/data': 0.01868177019059658, 'timing/model': 1.12215375341475, 'learning_rate/qwen_vl_interface': 9.065960325623692e-06, 'learning_rate/action_model': 9.065960325623692e-05, 'epoch': 0.58}) +[2026-06-17 19:26:17,920][starVLA.training.train_starvla][INFO] - Step 1502, Loss: {'train/loss': 0.4028618037700653, 'train/grad_norm_pre_clip': 6.338484764099121, 'timing/data': 0.013297775760293007, 'timing/model': 1.5878291195258498, 'learning_rate/qwen_vl_interface': 9.064672061783555e-06, 'learning_rate/action_model': 9.064672061783554e-05, 'epoch': 0.58}) +[2026-06-17 19:26:22,262][starVLA.training.train_starvla][INFO] - Step 1503, Loss: {'train/loss': 0.6890053153038025, 'train/grad_norm_pre_clip': 6.981248378753662, 'timing/data': 0.012256769463419914, 'timing/model': 1.1025265045464039, 'learning_rate/qwen_vl_interface': 9.063383007157795e-06, 'learning_rate/action_model': 9.063383007157796e-05, 'epoch': 0.58}) +[2026-06-17 19:26:26,897][starVLA.training.train_starvla][INFO] - Step 1504, Loss: {'train/loss': 0.6616501212120056, 'train/grad_norm_pre_clip': 10.124420166015625, 'timing/data': 0.013294646516442299, 'timing/model': 1.100842196494341, 'learning_rate/qwen_vl_interface': 9.062093162013635e-06, 'learning_rate/action_model': 9.062093162013634e-05, 'epoch': 0.58}) +[2026-06-17 19:26:31,529][starVLA.training.train_starvla][INFO] - Step 1505, Loss: {'train/loss': 0.4883824586868286, 'train/grad_norm_pre_clip': 7.518085956573486, 'timing/data': 0.011319863609969616, 'timing/model': 1.077071095816791, 'learning_rate/qwen_vl_interface': 9.06080252661846e-06, 'learning_rate/action_model': 9.060802526618458e-05, 'epoch': 0.58}) +[2026-06-17 19:26:40,723][starVLA.training.train_starvla][INFO] - Step 1506, Loss: {'train/loss': 1.0253785848617554, 'train/grad_norm_pre_clip': 11.787925720214844, 'timing/data': 0.011487716808915138, 'timing/model': 1.211063134483993, 'learning_rate/qwen_vl_interface': 9.059511101239818e-06, 'learning_rate/action_model': 9.059511101239818e-05, 'epoch': 0.58}) +[2026-06-17 19:26:45,202][starVLA.training.train_starvla][INFO] - Step 1507, Loss: {'train/loss': 0.4829471707344055, 'train/grad_norm_pre_clip': 6.868488788604736, 'timing/data': 0.011696211993694305, 'timing/model': 1.201121248304844, 'learning_rate/qwen_vl_interface': 9.058218886145427e-06, 'learning_rate/action_model': 9.058218886145427e-05, 'epoch': 0.58}) +[2026-06-17 19:26:50,208][starVLA.training.train_starvla][INFO] - Step 1508, Loss: {'train/loss': 0.58880215883255, 'train/grad_norm_pre_clip': 7.939931869506836, 'timing/data': 0.011262723244726658, 'timing/model': 1.1941939070820808, 'learning_rate/qwen_vl_interface': 9.056925881603165e-06, 'learning_rate/action_model': 9.056925881603166e-05, 'epoch': 0.58}) +[2026-06-17 19:26:57,737][starVLA.training.train_starvla][INFO] - Step 1509, Loss: {'train/loss': 0.5773752927780151, 'train/grad_norm_pre_clip': 5.76976203918457, 'timing/data': 0.014558198861777782, 'timing/model': 1.131497903726995, 'learning_rate/qwen_vl_interface': 9.05563208788107e-06, 'learning_rate/action_model': 9.05563208788107e-05, 'epoch': 0.58}) +[2026-06-17 19:27:03,027][starVLA.training.train_starvla][INFO] - Step 1510, Loss: {'train/loss': 0.9029351472854614, 'train/grad_norm_pre_clip': 11.954805374145508, 'timing/data': 0.011979957111179829, 'timing/model': 1.1084123719483614, 'learning_rate/qwen_vl_interface': 9.054337505247349e-06, 'learning_rate/action_model': 9.054337505247348e-05, 'epoch': 0.58}) +[2026-06-17 19:27:07,695][starVLA.training.train_starvla][INFO] - Step 1511, Loss: {'train/loss': 0.460367351770401, 'train/grad_norm_pre_clip': 6.4202446937561035, 'timing/data': 0.011608581058681011, 'timing/model': 1.537426060065627, 'learning_rate/qwen_vl_interface': 9.05304213397037e-06, 'learning_rate/action_model': 9.053042133970369e-05, 'epoch': 0.58}) +[2026-06-17 19:27:12,505][starVLA.training.train_starvla][INFO] - Step 1512, Loss: {'train/loss': 0.6395890712738037, 'train/grad_norm_pre_clip': 7.946982383728027, 'timing/data': 0.012576458975672722, 'timing/model': 1.4202394420281053, 'learning_rate/qwen_vl_interface': 9.051745974318663e-06, 'learning_rate/action_model': 9.051745974318662e-05, 'epoch': 0.58}) +[2026-06-17 19:27:19,613][starVLA.training.train_starvla][INFO] - Step 1513, Loss: {'train/loss': 0.5525608062744141, 'train/grad_norm_pre_clip': 5.792171955108643, 'timing/data': 0.011679883114993572, 'timing/model': 1.2403792152181268, 'learning_rate/qwen_vl_interface': 9.050449026560923e-06, 'learning_rate/action_model': 9.050449026560924e-05, 'epoch': 0.58}) +[2026-06-17 19:27:24,293][starVLA.training.train_starvla][INFO] - Step 1514, Loss: {'train/loss': 0.30872392654418945, 'train/grad_norm_pre_clip': 7.485041618347168, 'timing/data': 0.012842914089560509, 'timing/model': 1.1973100770264864, 'learning_rate/qwen_vl_interface': 9.049151290966013e-06, 'learning_rate/action_model': 9.049151290966012e-05, 'epoch': 0.58}) +[2026-06-17 19:27:28,852][starVLA.training.train_starvla][INFO] - Step 1515, Loss: {'train/loss': 0.6497048139572144, 'train/grad_norm_pre_clip': 8.08598804473877, 'timing/data': 0.010747038759291172, 'timing/model': 1.1021302314475179, 'learning_rate/qwen_vl_interface': 9.04785276780295e-06, 'learning_rate/action_model': 9.04785276780295e-05, 'epoch': 0.58}) +[2026-06-17 19:27:33,544][starVLA.training.train_starvla][INFO] - Step 1516, Loss: {'train/loss': 0.545993447303772, 'train/grad_norm_pre_clip': 7.028480052947998, 'timing/data': 0.011522234417498112, 'timing/model': 1.0874652778729796, 'learning_rate/qwen_vl_interface': 9.046553457340924e-06, 'learning_rate/action_model': 9.046553457340924e-05, 'epoch': 0.58}) +[2026-06-17 19:27:38,194][starVLA.training.train_starvla][INFO] - Step 1517, Loss: {'train/loss': 0.6461360454559326, 'train/grad_norm_pre_clip': 8.99704647064209, 'timing/data': 0.011040718294680119, 'timing/model': 1.463973382487893, 'learning_rate/qwen_vl_interface': 9.04525335984928e-06, 'learning_rate/action_model': 9.04525335984928e-05, 'epoch': 0.58}) +[2026-06-17 19:27:42,937][starVLA.training.train_starvla][INFO] - Step 1518, Loss: {'train/loss': 0.6419062614440918, 'train/grad_norm_pre_clip': 8.457921981811523, 'timing/data': 0.011574465781450272, 'timing/model': 1.1190521288663149, 'learning_rate/qwen_vl_interface': 9.043952475597532e-06, 'learning_rate/action_model': 9.043952475597532e-05, 'epoch': 0.58}) +[2026-06-17 19:27:49,642][starVLA.training.train_starvla][INFO] - Step 1519, Loss: {'train/loss': 0.5274463295936584, 'train/grad_norm_pre_clip': 10.07153034210205, 'timing/data': 0.012306463904678822, 'timing/model': 1.1243838462978601, 'learning_rate/qwen_vl_interface': 9.04265080485535e-06, 'learning_rate/action_model': 9.042650804855351e-05, 'epoch': 0.58}) +[2026-06-17 19:27:54,504][starVLA.training.train_starvla][INFO] - Step 1520, Loss: {'train/loss': 0.6212337017059326, 'train/grad_norm_pre_clip': 7.879696846008301, 'timing/data': 0.011356683447957039, 'timing/model': 1.0826138062402606, 'learning_rate/qwen_vl_interface': 9.041348347892578e-06, 'learning_rate/action_model': 9.041348347892579e-05, 'epoch': 0.58}) +[2026-06-17 19:27:59,367][starVLA.training.train_starvla][INFO] - Step 1521, Loss: {'train/loss': 0.34256038069725037, 'train/grad_norm_pre_clip': 5.053216457366943, 'timing/data': 0.014604411087930202, 'timing/model': 1.1031204219907522, 'learning_rate/qwen_vl_interface': 9.040045104979213e-06, 'learning_rate/action_model': 9.040045104979214e-05, 'epoch': 0.59}) +[2026-06-17 19:28:04,160][starVLA.training.train_starvla][INFO] - Step 1522, Loss: {'train/loss': 0.5880849361419678, 'train/grad_norm_pre_clip': 6.538029670715332, 'timing/data': 0.01197739690542221, 'timing/model': 1.5460942704230547, 'learning_rate/qwen_vl_interface': 9.03874107638542e-06, 'learning_rate/action_model': 9.038741076385419e-05, 'epoch': 0.59}) +[2026-06-17 19:28:08,500][starVLA.training.train_starvla][INFO] - Step 1523, Loss: {'train/loss': 0.6523849964141846, 'train/grad_norm_pre_clip': 8.601339340209961, 'timing/data': 0.011889315210282803, 'timing/model': 1.092402871698141, 'learning_rate/qwen_vl_interface': 9.037436262381525e-06, 'learning_rate/action_model': 9.037436262381525e-05, 'epoch': 0.59}) +[2026-06-17 19:28:13,298][starVLA.training.train_starvla][INFO] - Step 1524, Loss: {'train/loss': 0.8141306042671204, 'train/grad_norm_pre_clip': 46.5704345703125, 'timing/data': 0.01353248581290245, 'timing/model': 1.0886749532073736, 'learning_rate/qwen_vl_interface': 9.036130663238018e-06, 'learning_rate/action_model': 9.036130663238017e-05, 'epoch': 0.59}) +[2026-06-17 19:28:18,008][starVLA.training.train_starvla][INFO] - Step 1525, Loss: {'train/loss': 0.8927243947982788, 'train/grad_norm_pre_clip': 6.741554260253906, 'timing/data': 0.013926473446190357, 'timing/model': 1.0751883164048195, 'learning_rate/qwen_vl_interface': 9.034824279225547e-06, 'learning_rate/action_model': 9.034824279225547e-05, 'epoch': 0.59}) +[2026-06-17 19:28:22,637][starVLA.training.train_starvla][INFO] - Step 1526, Loss: {'train/loss': 0.4125927686691284, 'train/grad_norm_pre_clip': 5.634405136108398, 'timing/data': 0.012156575918197632, 'timing/model': 1.0664050867781043, 'learning_rate/qwen_vl_interface': 9.033517110614933e-06, 'learning_rate/action_model': 9.033517110614933e-05, 'epoch': 0.59}) +[2026-06-17 19:28:27,519][starVLA.training.train_starvla][INFO] - Step 1527, Loss: {'train/loss': 0.6722619533538818, 'train/grad_norm_pre_clip': 8.447916984558105, 'timing/data': 0.014205730520188808, 'timing/model': 1.522812052629888, 'learning_rate/qwen_vl_interface': 9.032209157677151e-06, 'learning_rate/action_model': 9.032209157677151e-05, 'epoch': 0.59}) +[2026-06-17 19:28:32,645][starVLA.training.train_starvla][INFO] - Step 1528, Loss: {'train/loss': 0.4572875201702118, 'train/grad_norm_pre_clip': 5.638359546661377, 'timing/data': 0.012645391747355461, 'timing/model': 1.070287236943841, 'learning_rate/qwen_vl_interface': 9.03090042068334e-06, 'learning_rate/action_model': 9.030900420683339e-05, 'epoch': 0.59}) +[2026-06-17 19:28:37,322][starVLA.training.train_starvla][INFO] - Step 1529, Loss: {'train/loss': 0.5328934788703918, 'train/grad_norm_pre_clip': 9.86668586730957, 'timing/data': 0.013591843657195568, 'timing/model': 1.0805133981630206, 'learning_rate/qwen_vl_interface': 9.029590899904803e-06, 'learning_rate/action_model': 9.029590899904803e-05, 'epoch': 0.59}) +[2026-06-17 19:28:42,047][starVLA.training.train_starvla][INFO] - Step 1530, Loss: {'train/loss': 0.2839016616344452, 'train/grad_norm_pre_clip': 4.716017723083496, 'timing/data': 0.011815016157925129, 'timing/model': 1.098809354007244, 'learning_rate/qwen_vl_interface': 9.028280595613006e-06, 'learning_rate/action_model': 9.028280595613005e-05, 'epoch': 0.59}) +[2026-06-17 19:28:46,755][starVLA.training.train_starvla][INFO] - Step 1531, Loss: {'train/loss': 0.23269213736057281, 'train/grad_norm_pre_clip': 5.478999137878418, 'timing/data': 0.012844443321228027, 'timing/model': 1.194256971590221, 'learning_rate/qwen_vl_interface': 9.026969508079575e-06, 'learning_rate/action_model': 9.026969508079575e-05, 'epoch': 0.59}) +[2026-06-17 19:28:52,705][starVLA.training.train_starvla][INFO] - Step 1532, Loss: {'train/loss': 0.6265789270401001, 'train/grad_norm_pre_clip': 7.04703426361084, 'timing/data': 0.018645198084414005, 'timing/model': 1.525664772838354, 'learning_rate/qwen_vl_interface': 9.025657637576302e-06, 'learning_rate/action_model': 9.025657637576302e-05, 'epoch': 0.59}) +[2026-06-17 19:29:03,787][starVLA.training.train_starvla][INFO] - Step 1533, Loss: {'train/loss': 0.3733736276626587, 'train/grad_norm_pre_clip': 6.783360958099365, 'timing/data': 0.014366365037858486, 'timing/model': 5.097764409147203, 'learning_rate/qwen_vl_interface': 9.024344984375138e-06, 'learning_rate/action_model': 9.024344984375137e-05, 'epoch': 0.59}) +[2026-06-17 19:29:08,575][starVLA.training.train_starvla][INFO] - Step 1534, Loss: {'train/loss': 0.5122492909431458, 'train/grad_norm_pre_clip': 5.8600263595581055, 'timing/data': 0.01253887265920639, 'timing/model': 1.1068376926705241, 'learning_rate/qwen_vl_interface': 9.023031548748197e-06, 'learning_rate/action_model': 9.023031548748196e-05, 'epoch': 0.59}) +[2026-06-17 19:29:13,567][starVLA.training.train_starvla][INFO] - Step 1535, Loss: {'train/loss': 0.5851102471351624, 'train/grad_norm_pre_clip': 7.724804401397705, 'timing/data': 0.01209171675145626, 'timing/model': 1.0698687713593245, 'learning_rate/qwen_vl_interface': 9.021717330967755e-06, 'learning_rate/action_model': 9.021717330967754e-05, 'epoch': 0.59}) +[2026-06-17 19:29:22,458][starVLA.training.train_starvla][INFO] - Step 1536, Loss: {'train/loss': 0.6618058085441589, 'train/grad_norm_pre_clip': 12.633994102478027, 'timing/data': 0.0141531303524971, 'timing/model': 1.0929792812094092, 'learning_rate/qwen_vl_interface': 9.020402331306254e-06, 'learning_rate/action_model': 9.020402331306253e-05, 'epoch': 0.59}) +[2026-06-17 19:29:27,214][starVLA.training.train_starvla][INFO] - Step 1537, Loss: {'train/loss': 0.635919451713562, 'train/grad_norm_pre_clip': 12.858397483825684, 'timing/data': 0.012019939720630646, 'timing/model': 1.410630440339446, 'learning_rate/qwen_vl_interface': 9.01908655003629e-06, 'learning_rate/action_model': 9.01908655003629e-05, 'epoch': 0.59}) +[2026-06-17 19:29:31,545][starVLA.training.train_starvla][INFO] - Step 1538, Loss: {'train/loss': 0.831712007522583, 'train/grad_norm_pre_clip': 11.7002534866333, 'timing/data': 0.013420884497463703, 'timing/model': 1.1178225222975016, 'learning_rate/qwen_vl_interface': 9.017769987430629e-06, 'learning_rate/action_model': 9.017769987430628e-05, 'epoch': 0.59}) +[2026-06-17 19:29:36,399][starVLA.training.train_starvla][INFO] - Step 1539, Loss: {'train/loss': 0.4355621337890625, 'train/grad_norm_pre_clip': 6.742751598358154, 'timing/data': 0.014852810651063919, 'timing/model': 1.182716996408999, 'learning_rate/qwen_vl_interface': 9.016452643762197e-06, 'learning_rate/action_model': 9.016452643762197e-05, 'epoch': 0.59}) +[2026-06-17 19:29:41,240][starVLA.training.train_starvla][INFO] - Step 1540, Loss: {'train/loss': 0.5625818967819214, 'train/grad_norm_pre_clip': 9.508259773254395, 'timing/data': 0.014935338869690895, 'timing/model': 1.1037146719172597, 'learning_rate/qwen_vl_interface': 9.015134519304078e-06, 'learning_rate/action_model': 9.015134519304078e-05, 'epoch': 0.59}) +[2026-06-17 19:29:45,976][starVLA.training.train_starvla][INFO] - Step 1541, Loss: {'train/loss': 0.6705636978149414, 'train/grad_norm_pre_clip': 7.631291389465332, 'timing/data': 0.011640919372439384, 'timing/model': 1.0847589410841465, 'learning_rate/qwen_vl_interface': 9.013815614329521e-06, 'learning_rate/action_model': 9.013815614329522e-05, 'epoch': 0.59}) +[2026-06-17 19:29:52,298][starVLA.training.train_starvla][INFO] - Step 1542, Loss: {'train/loss': 0.49227434396743774, 'train/grad_norm_pre_clip': 6.730690002441406, 'timing/data': 0.010738885961472988, 'timing/model': 1.4804677898064256, 'learning_rate/qwen_vl_interface': 9.01249592911194e-06, 'learning_rate/action_model': 9.012495929111938e-05, 'epoch': 0.59}) +[2026-06-17 19:29:56,870][starVLA.training.train_starvla][INFO] - Step 1543, Loss: {'train/loss': 0.7099299430847168, 'train/grad_norm_pre_clip': 7.4827423095703125, 'timing/data': 0.01947026327252388, 'timing/model': 1.094181784428656, 'learning_rate/qwen_vl_interface': 9.0111754639249e-06, 'learning_rate/action_model': 9.0111754639249e-05, 'epoch': 0.59}) +[2026-06-17 19:30:01,542][starVLA.training.train_starvla][INFO] - Step 1544, Loss: {'train/loss': 0.478593111038208, 'train/grad_norm_pre_clip': 8.671075820922852, 'timing/data': 0.011817934922873974, 'timing/model': 1.0736132692545652, 'learning_rate/qwen_vl_interface': 9.009854219042142e-06, 'learning_rate/action_model': 9.009854219042142e-05, 'epoch': 0.59}) +[2026-06-17 19:30:06,047][starVLA.training.train_starvla][INFO] - Step 1545, Loss: {'train/loss': 0.5312386155128479, 'train/grad_norm_pre_clip': 9.791548728942871, 'timing/data': 0.012565778568387032, 'timing/model': 1.0838878089562058, 'learning_rate/qwen_vl_interface': 9.008532194737556e-06, 'learning_rate/action_model': 9.008532194737557e-05, 'epoch': 0.59}) +[2026-06-17 19:30:10,729][starVLA.training.train_starvla][INFO] - Step 1546, Loss: {'train/loss': 0.806010901927948, 'train/grad_norm_pre_clip': 10.695633888244629, 'timing/data': 0.01144405733793974, 'timing/model': 1.1294295098632574, 'learning_rate/qwen_vl_interface': 9.007209391285205e-06, 'learning_rate/action_model': 9.007209391285205e-05, 'epoch': 0.59}) +[2026-06-17 19:30:15,427][starVLA.training.train_starvla][INFO] - Step 1547, Loss: {'train/loss': 0.8496047258377075, 'train/grad_norm_pre_clip': 12.009517669677734, 'timing/data': 0.011547715403139591, 'timing/model': 1.4324089363217354, 'learning_rate/qwen_vl_interface': 9.005885808959301e-06, 'learning_rate/action_model': 9.005885808959302e-05, 'epoch': 0.6}) +[2026-06-17 19:30:19,675][starVLA.training.train_starvla][INFO] - Step 1548, Loss: {'train/loss': 0.541557252407074, 'train/grad_norm_pre_clip': 8.785746574401855, 'timing/data': 0.011446245014667511, 'timing/model': 1.134128039702773, 'learning_rate/qwen_vl_interface': 9.00456144803423e-06, 'learning_rate/action_model': 9.00456144803423e-05, 'epoch': 0.6}) +[2026-06-17 19:30:24,456][starVLA.training.train_starvla][INFO] - Step 1549, Loss: {'train/loss': 0.6262245774269104, 'train/grad_norm_pre_clip': 7.302807331085205, 'timing/data': 0.01275794766843319, 'timing/model': 1.0969820357859135, 'learning_rate/qwen_vl_interface': 9.00323630878453e-06, 'learning_rate/action_model': 9.00323630878453e-05, 'epoch': 0.6}) +[2026-06-17 19:30:29,061][starVLA.training.train_starvla][INFO] - Step 1550, Loss: {'train/loss': 0.5845047235488892, 'train/grad_norm_pre_clip': 8.21720027923584, 'timing/data': 0.011947494000196457, 'timing/model': 1.0808460861444473, 'learning_rate/qwen_vl_interface': 9.001910391484903e-06, 'learning_rate/action_model': 9.001910391484904e-05, 'epoch': 0.6}) +[2026-06-17 19:30:33,627][starVLA.training.train_starvla][INFO] - Step 1551, Loss: {'train/loss': 0.457330584526062, 'train/grad_norm_pre_clip': 6.60530424118042, 'timing/data': 0.01125340722501278, 'timing/model': 1.128451376222074, 'learning_rate/qwen_vl_interface': 9.000583696410216e-06, 'learning_rate/action_model': 9.000583696410217e-05, 'epoch': 0.6}) +[2026-06-17 19:30:38,535][starVLA.training.train_starvla][INFO] - Step 1552, Loss: {'train/loss': 0.6675018668174744, 'train/grad_norm_pre_clip': 9.215428352355957, 'timing/data': 0.013981552794575691, 'timing/model': 1.6224784003570676, 'learning_rate/qwen_vl_interface': 8.999256223835493e-06, 'learning_rate/action_model': 8.999256223835494e-05, 'epoch': 0.6}) +[2026-06-17 19:30:45,564][starVLA.training.train_starvla][INFO] - Step 1553, Loss: {'train/loss': 0.8419968485832214, 'train/grad_norm_pre_clip': 6.7089314460754395, 'timing/data': 0.011609761975705624, 'timing/model': 1.07342994119972, 'learning_rate/qwen_vl_interface': 8.997927974035922e-06, 'learning_rate/action_model': 8.997927974035921e-05, 'epoch': 0.6}) +[2026-06-17 19:30:55,824][starVLA.training.train_starvla][INFO] - Step 1554, Loss: {'train/loss': 0.6986957788467407, 'train/grad_norm_pre_clip': 8.850955963134766, 'timing/data': 0.014585177414119244, 'timing/model': 3.0488130440935493, 'learning_rate/qwen_vl_interface': 8.996598947286846e-06, 'learning_rate/action_model': 8.996598947286846e-05, 'epoch': 0.6}) +[2026-06-17 19:31:00,763][starVLA.training.train_starvla][INFO] - Step 1555, Loss: {'train/loss': 0.3213443160057068, 'train/grad_norm_pre_clip': 6.565278053283691, 'timing/data': 0.01170049887150526, 'timing/model': 1.1163045996800065, 'learning_rate/qwen_vl_interface': 8.99526914386378e-06, 'learning_rate/action_model': 8.995269143863779e-05, 'epoch': 0.6}) +[2026-06-17 19:31:08,151][starVLA.training.train_starvla][INFO] - Step 1556, Loss: {'train/loss': 0.6835176944732666, 'train/grad_norm_pre_clip': 7.207050800323486, 'timing/data': 0.01229871902614832, 'timing/model': 1.172365234233439, 'learning_rate/qwen_vl_interface': 8.993938564042386e-06, 'learning_rate/action_model': 8.993938564042387e-05, 'epoch': 0.6}) +[2026-06-17 19:31:12,887][starVLA.training.train_starvla][INFO] - Step 1557, Loss: {'train/loss': 0.9518597722053528, 'train/grad_norm_pre_clip': 10.857760429382324, 'timing/data': 0.012749997898936272, 'timing/model': 1.4238286716863513, 'learning_rate/qwen_vl_interface': 8.992607208098502e-06, 'learning_rate/action_model': 8.992607208098503e-05, 'epoch': 0.6}) +[2026-06-17 19:31:17,092][starVLA.training.train_starvla][INFO] - Step 1558, Loss: {'train/loss': 0.22226622700691223, 'train/grad_norm_pre_clip': 3.762540102005005, 'timing/data': 0.011385909281671047, 'timing/model': 1.1014602659270167, 'learning_rate/qwen_vl_interface': 8.991275076308117e-06, 'learning_rate/action_model': 8.991275076308117e-05, 'epoch': 0.6}) +[2026-06-17 19:31:21,642][starVLA.training.train_starvla][INFO] - Step 1559, Loss: {'train/loss': 0.38201433420181274, 'train/grad_norm_pre_clip': 5.2162251472473145, 'timing/data': 0.01186458021402359, 'timing/model': 1.0686235129833221, 'learning_rate/qwen_vl_interface': 8.989942168947382e-06, 'learning_rate/action_model': 8.989942168947381e-05, 'epoch': 0.6}) +[2026-06-17 19:31:26,273][starVLA.training.train_starvla][INFO] - Step 1560, Loss: {'train/loss': 0.6944326162338257, 'train/grad_norm_pre_clip': 7.859866619110107, 'timing/data': 0.013126550242304802, 'timing/model': 1.0811577085405588, 'learning_rate/qwen_vl_interface': 8.98860848629261e-06, 'learning_rate/action_model': 8.988608486292611e-05, 'epoch': 0.6}) +[2026-06-17 19:31:30,941][starVLA.training.train_starvla][INFO] - Step 1561, Loss: {'train/loss': 0.4632694125175476, 'train/grad_norm_pre_clip': 5.263168811798096, 'timing/data': 0.015508530661463737, 'timing/model': 1.2150292694568634, 'learning_rate/qwen_vl_interface': 8.987274028620276e-06, 'learning_rate/action_model': 8.987274028620276e-05, 'epoch': 0.6}) +[2026-06-17 19:31:36,339][starVLA.training.train_starvla][INFO] - Step 1562, Loss: {'train/loss': 0.5164080858230591, 'train/grad_norm_pre_clip': 8.598130226135254, 'timing/data': 0.014755184762179852, 'timing/model': 1.5313115632161498, 'learning_rate/qwen_vl_interface': 8.985938796207017e-06, 'learning_rate/action_model': 8.985938796207017e-05, 'epoch': 0.6}) +[2026-06-17 19:31:40,628][starVLA.training.train_starvla][INFO] - Step 1563, Loss: {'train/loss': 0.7169725894927979, 'train/grad_norm_pre_clip': 9.104921340942383, 'timing/data': 0.011871895752847195, 'timing/model': 1.0885343570262194, 'learning_rate/qwen_vl_interface': 8.984602789329623e-06, 'learning_rate/action_model': 8.984602789329623e-05, 'epoch': 0.6}) +[2026-06-17 19:31:45,179][starVLA.training.train_starvla][INFO] - Step 1564, Loss: {'train/loss': 0.6759227514266968, 'train/grad_norm_pre_clip': 7.8902740478515625, 'timing/data': 0.01257906761020422, 'timing/model': 1.1130212843418121, 'learning_rate/qwen_vl_interface': 8.983266008265052e-06, 'learning_rate/action_model': 8.983266008265052e-05, 'epoch': 0.6}) +[2026-06-17 19:31:49,854][starVLA.training.train_starvla][INFO] - Step 1565, Loss: {'train/loss': 0.5095037221908569, 'train/grad_norm_pre_clip': 7.798882961273193, 'timing/data': 0.011890570633113384, 'timing/model': 1.147714041173458, 'learning_rate/qwen_vl_interface': 8.981928453290421e-06, 'learning_rate/action_model': 8.981928453290421e-05, 'epoch': 0.6}) +[2026-06-17 19:31:54,450][starVLA.training.train_starvla][INFO] - Step 1566, Loss: {'train/loss': 0.5938271284103394, 'train/grad_norm_pre_clip': 12.364977836608887, 'timing/data': 0.011779787950217724, 'timing/model': 1.1119837751612067, 'learning_rate/qwen_vl_interface': 8.980590124683007e-06, 'learning_rate/action_model': 8.980590124683007e-05, 'epoch': 0.6}) +[2026-06-17 19:31:59,056][starVLA.training.train_starvla][INFO] - Step 1567, Loss: {'train/loss': 0.765654444694519, 'train/grad_norm_pre_clip': 9.887097358703613, 'timing/data': 0.011262191459536552, 'timing/model': 1.4541904088109732, 'learning_rate/qwen_vl_interface': 8.979251022720244e-06, 'learning_rate/action_model': 8.979251022720244e-05, 'epoch': 0.6}) +[2026-06-17 19:32:03,325][starVLA.training.train_starvla][INFO] - Step 1568, Loss: {'train/loss': 0.3151428997516632, 'train/grad_norm_pre_clip': 10.166504859924316, 'timing/data': 0.013005626387894154, 'timing/model': 1.1024671411141753, 'learning_rate/qwen_vl_interface': 8.977911147679733e-06, 'learning_rate/action_model': 8.977911147679733e-05, 'epoch': 0.6}) +[2026-06-17 19:32:09,783][starVLA.training.train_starvla][INFO] - Step 1569, Loss: {'train/loss': 0.5915570855140686, 'train/grad_norm_pre_clip': 16.277475357055664, 'timing/data': 0.011812588199973106, 'timing/model': 1.1914347130805254, 'learning_rate/qwen_vl_interface': 8.976570499839229e-06, 'learning_rate/action_model': 8.976570499839228e-05, 'epoch': 0.6}) +[2026-06-17 19:32:14,555][starVLA.training.train_starvla][INFO] - Step 1570, Loss: {'train/loss': 0.5321940779685974, 'train/grad_norm_pre_clip': 8.051668167114258, 'timing/data': 0.017850800417363644, 'timing/model': 1.111816386692226, 'learning_rate/qwen_vl_interface': 8.97522907947665e-06, 'learning_rate/action_model': 8.975229079476649e-05, 'epoch': 0.6}) +[2026-06-17 19:32:19,033][starVLA.training.train_starvla][INFO] - Step 1571, Loss: {'train/loss': 0.4476762115955353, 'train/grad_norm_pre_clip': 15.596395492553711, 'timing/data': 0.012718094512820244, 'timing/model': 1.083190070465207, 'learning_rate/qwen_vl_interface': 8.973886886870075e-06, 'learning_rate/action_model': 8.973886886870074e-05, 'epoch': 0.6}) +[2026-06-17 19:32:23,707][starVLA.training.train_starvla][INFO] - Step 1572, Loss: {'train/loss': 0.7046092748641968, 'train/grad_norm_pre_clip': 8.452906608581543, 'timing/data': 0.011418212205171585, 'timing/model': 1.4565322063863277, 'learning_rate/qwen_vl_interface': 8.972543922297739e-06, 'learning_rate/action_model': 8.972543922297739e-05, 'epoch': 0.6}) +[2026-06-17 19:32:28,101][starVLA.training.train_starvla][INFO] - Step 1573, Loss: {'train/loss': 0.5328376293182373, 'train/grad_norm_pre_clip': 7.4707417488098145, 'timing/data': 0.011857717297971249, 'timing/model': 1.0912940166890621, 'learning_rate/qwen_vl_interface': 8.971200186038046e-06, 'learning_rate/action_model': 8.971200186038045e-05, 'epoch': 0.61}) +[2026-06-17 19:32:33,801][starVLA.training.train_starvla][INFO] - Step 1574, Loss: {'train/loss': 0.31175944209098816, 'train/grad_norm_pre_clip': 14.652365684509277, 'timing/data': 0.012285005301237106, 'timing/model': 1.0719591928645968, 'learning_rate/qwen_vl_interface': 8.969855678369547e-06, 'learning_rate/action_model': 8.969855678369546e-05, 'epoch': 0.61}) +[2026-06-17 19:32:38,663][starVLA.training.train_starvla][INFO] - Step 1575, Loss: {'train/loss': 0.42812788486480713, 'train/grad_norm_pre_clip': 6.128484725952148, 'timing/data': 0.017065842635929585, 'timing/model': 1.1688255509361625, 'learning_rate/qwen_vl_interface': 8.968510399570964e-06, 'learning_rate/action_model': 8.968510399570965e-05, 'epoch': 0.61}) +[2026-06-17 19:32:43,386][starVLA.training.train_starvla][INFO] - Step 1576, Loss: {'train/loss': 0.4464568793773651, 'train/grad_norm_pre_clip': 8.600188255310059, 'timing/data': 0.023393802344799042, 'timing/model': 1.1910337498411536, 'learning_rate/qwen_vl_interface': 8.967164349921173e-06, 'learning_rate/action_model': 8.967164349921174e-05, 'epoch': 0.61}) +[2026-06-17 19:32:48,061][starVLA.training.train_starvla][INFO] - Step 1577, Loss: {'train/loss': 0.6901983618736267, 'train/grad_norm_pre_clip': 7.68897819519043, 'timing/data': 0.011518602259457111, 'timing/model': 1.4246700312942266, 'learning_rate/qwen_vl_interface': 8.965817529699216e-06, 'learning_rate/action_model': 8.965817529699215e-05, 'epoch': 0.61}) +[2026-06-17 19:32:55,519][starVLA.training.train_starvla][INFO] - Step 1578, Loss: {'train/loss': 0.42878997325897217, 'train/grad_norm_pre_clip': 7.30612325668335, 'timing/data': 0.012336828745901585, 'timing/model': 1.1223919447511435, 'learning_rate/qwen_vl_interface': 8.964469939184283e-06, 'learning_rate/action_model': 8.964469939184282e-05, 'epoch': 0.61}) +[2026-06-17 19:33:00,215][starVLA.training.train_starvla][INFO] - Step 1579, Loss: {'train/loss': 0.5563990473747253, 'train/grad_norm_pre_clip': 52.69822311401367, 'timing/data': 0.011570679023861885, 'timing/model': 1.1695926804095507, 'learning_rate/qwen_vl_interface': 8.963121578655736e-06, 'learning_rate/action_model': 8.963121578655736e-05, 'epoch': 0.61}) +[2026-06-17 19:33:07,932][starVLA.training.train_starvla][INFO] - Step 1580, Loss: {'train/loss': 0.8959439396858215, 'train/grad_norm_pre_clip': 8.188072204589844, 'timing/data': 0.01546663697808981, 'timing/model': 1.1243381761014462, 'learning_rate/qwen_vl_interface': 8.96177244839309e-06, 'learning_rate/action_model': 8.961772448393092e-05, 'epoch': 0.61}) +[2026-06-17 19:33:12,795][starVLA.training.train_starvla][INFO] - Step 1581, Loss: {'train/loss': 0.38898634910583496, 'train/grad_norm_pre_clip': 5.679961681365967, 'timing/data': 0.01161169447004795, 'timing/model': 1.217588963918388, 'learning_rate/qwen_vl_interface': 8.960422548676024e-06, 'learning_rate/action_model': 8.960422548676022e-05, 'epoch': 0.61}) +[2026-06-17 19:33:19,608][starVLA.training.train_starvla][INFO] - Step 1582, Loss: {'train/loss': 0.3379572331905365, 'train/grad_norm_pre_clip': 4.583681583404541, 'timing/data': 0.012631265446543694, 'timing/model': 1.459521196782589, 'learning_rate/qwen_vl_interface': 8.959071879784368e-06, 'learning_rate/action_model': 8.959071879784368e-05, 'epoch': 0.61}) +[2026-06-17 19:33:23,816][starVLA.training.train_starvla][INFO] - Step 1583, Loss: {'train/loss': 0.4715169668197632, 'train/grad_norm_pre_clip': 7.231911659240723, 'timing/data': 0.012830321677029133, 'timing/model': 1.0863539865240455, 'learning_rate/qwen_vl_interface': 8.95772044199812e-06, 'learning_rate/action_model': 8.95772044199812e-05, 'epoch': 0.61}) +[2026-06-17 19:33:31,821][starVLA.training.train_starvla][INFO] - Step 1584, Loss: {'train/loss': 0.5885934829711914, 'train/grad_norm_pre_clip': 7.897254943847656, 'timing/data': 0.02192941028624773, 'timing/model': 1.1256148405373096, 'learning_rate/qwen_vl_interface': 8.956368235597437e-06, 'learning_rate/action_model': 8.956368235597436e-05, 'epoch': 0.61}) +[2026-06-17 19:33:36,597][starVLA.training.train_starvla][INFO] - Step 1585, Loss: {'train/loss': 0.41408365964889526, 'train/grad_norm_pre_clip': 6.607750415802002, 'timing/data': 0.013820773921906948, 'timing/model': 1.2000828515738249, 'learning_rate/qwen_vl_interface': 8.955015260862627e-06, 'learning_rate/action_model': 8.955015260862627e-05, 'epoch': 0.61}) +[2026-06-17 19:33:41,214][starVLA.training.train_starvla][INFO] - Step 1586, Loss: {'train/loss': 0.5992754697799683, 'train/grad_norm_pre_clip': 6.744997978210449, 'timing/data': 0.011596321128308773, 'timing/model': 1.1167707834392786, 'learning_rate/qwen_vl_interface': 8.95366151807417e-06, 'learning_rate/action_model': 8.95366151807417e-05, 'epoch': 0.61}) +[2026-06-17 19:33:45,770][starVLA.training.train_starvla][INFO] - Step 1587, Loss: {'train/loss': 0.4781322479248047, 'train/grad_norm_pre_clip': 8.536251068115234, 'timing/data': 0.01146643701940775, 'timing/model': 1.385418951511383, 'learning_rate/qwen_vl_interface': 8.952307007512694e-06, 'learning_rate/action_model': 8.952307007512694e-05, 'epoch': 0.61}) +[2026-06-17 19:33:50,185][starVLA.training.train_starvla][INFO] - Step 1588, Loss: {'train/loss': 0.7765296697616577, 'train/grad_norm_pre_clip': 9.39930248260498, 'timing/data': 0.012654239311814308, 'timing/model': 1.1283161314204335, 'learning_rate/qwen_vl_interface': 8.950951729458991e-06, 'learning_rate/action_model': 8.950951729458991e-05, 'epoch': 0.61}) +[2026-06-17 19:33:54,702][starVLA.training.train_starvla][INFO] - Step 1589, Loss: {'train/loss': 0.510191798210144, 'train/grad_norm_pre_clip': 7.357193470001221, 'timing/data': 0.015948399901390076, 'timing/model': 1.0698248762637377, 'learning_rate/qwen_vl_interface': 8.949595684194012e-06, 'learning_rate/action_model': 8.949595684194012e-05, 'epoch': 0.61}) +[2026-06-17 19:34:00,148][starVLA.training.train_starvla][INFO] - Step 1590, Loss: {'train/loss': 0.603253185749054, 'train/grad_norm_pre_clip': 7.204811096191406, 'timing/data': 0.011190202087163925, 'timing/model': 1.123938910663128, 'learning_rate/qwen_vl_interface': 8.948238871998866e-06, 'learning_rate/action_model': 8.948238871998866e-05, 'epoch': 0.61}) +[2026-06-17 19:34:06,576][starVLA.training.train_starvla][INFO] - Step 1591, Loss: {'train/loss': 0.4405057430267334, 'train/grad_norm_pre_clip': 9.153108596801758, 'timing/data': 0.011359465308487415, 'timing/model': 1.0929125789552927, 'learning_rate/qwen_vl_interface': 8.946881293154825e-06, 'learning_rate/action_model': 8.946881293154825e-05, 'epoch': 0.61}) +[2026-06-17 19:34:11,265][starVLA.training.train_starvla][INFO] - Step 1592, Loss: {'train/loss': 0.4377824664115906, 'train/grad_norm_pre_clip': 5.498100280761719, 'timing/data': 0.012656298466026783, 'timing/model': 1.517818153835833, 'learning_rate/qwen_vl_interface': 8.945522947943312e-06, 'learning_rate/action_model': 8.945522947943313e-05, 'epoch': 0.61}) +[2026-06-17 19:34:15,623][starVLA.training.train_starvla][INFO] - Step 1593, Loss: {'train/loss': 0.4070107936859131, 'train/grad_norm_pre_clip': 6.977297306060791, 'timing/data': 0.013280627317726612, 'timing/model': 1.1302222833037376, 'learning_rate/qwen_vl_interface': 8.944163836645916e-06, 'learning_rate/action_model': 8.944163836645915e-05, 'epoch': 0.61}) +[2026-06-17 19:34:20,407][starVLA.training.train_starvla][INFO] - Step 1594, Loss: {'train/loss': 0.7181561589241028, 'train/grad_norm_pre_clip': 10.490884780883789, 'timing/data': 0.012975368648767471, 'timing/model': 1.0983211444690824, 'learning_rate/qwen_vl_interface': 8.942803959544381e-06, 'learning_rate/action_model': 8.942803959544381e-05, 'epoch': 0.61}) +[2026-06-17 19:34:25,191][starVLA.training.train_starvla][INFO] - Step 1595, Loss: {'train/loss': 0.6158161163330078, 'train/grad_norm_pre_clip': 10.533507347106934, 'timing/data': 0.020987817086279392, 'timing/model': 1.128183207474649, 'learning_rate/qwen_vl_interface': 8.941443316920613e-06, 'learning_rate/action_model': 8.941443316920612e-05, 'epoch': 0.61}) +[2026-06-17 19:34:29,846][starVLA.training.train_starvla][INFO] - Step 1596, Loss: {'train/loss': 0.6052523851394653, 'train/grad_norm_pre_clip': 7.675922870635986, 'timing/data': 0.011320618912577629, 'timing/model': 1.102991284802556, 'learning_rate/qwen_vl_interface': 8.940081909056671e-06, 'learning_rate/action_model': 8.940081909056671e-05, 'epoch': 0.61}) +[2026-06-17 19:34:34,436][starVLA.training.train_starvla][INFO] - Step 1597, Loss: {'train/loss': 0.7884350419044495, 'train/grad_norm_pre_clip': 12.280867576599121, 'timing/data': 0.01114062499254942, 'timing/model': 1.4767005145549774, 'learning_rate/qwen_vl_interface': 8.93871973623478e-06, 'learning_rate/action_model': 8.93871973623478e-05, 'epoch': 0.61}) +[2026-06-17 19:34:38,787][starVLA.training.train_starvla][INFO] - Step 1598, Loss: {'train/loss': 0.5269076824188232, 'train/grad_norm_pre_clip': 5.8184919357299805, 'timing/data': 0.014058195054531097, 'timing/model': 1.1158344335854053, 'learning_rate/qwen_vl_interface': 8.937356798737317e-06, 'learning_rate/action_model': 8.937356798737317e-05, 'epoch': 0.61}) +[2026-06-17 19:34:43,661][starVLA.training.train_starvla][INFO] - Step 1599, Loss: {'train/loss': 0.5676131844520569, 'train/grad_norm_pre_clip': 8.3856201171875, 'timing/data': 0.014245177619159222, 'timing/model': 1.2584966560825706, 'learning_rate/qwen_vl_interface': 8.935993096846825e-06, 'learning_rate/action_model': 8.935993096846825e-05, 'epoch': 0.62}) +[2026-06-17 19:37:53,880][starVLA.training.train_starvla][INFO] - Step 1600, Loss: {'train/loss': 0.5542068481445312, 'train/grad_norm_pre_clip': 7.56726598739624, 'eval/loss': 0.5850298404693604, 'eval/latency_4/loss': 0.5850298404693604, 'eval/demon_fire_precision': 0.948405253283302, 'eval/demon_fire_recall': 0.985060084443001, 'eval/demon_fire_f1': 0.9663852158674526, 'eval/demon_fire_rate_model': 0.9392070484581497, 'eval/demon_fire_rate_teacher': 0.9042584434654919, 'eval/demon_left_precision': 0.7851239669421488, 'eval/demon_left_recall': 0.4337899543378995, 'eval/demon_left_f1': 0.5588235294117646, 'eval/demon_left_rate_model': 0.07107195301027901, 'eval/demon_left_rate_teacher': 0.12863436123348018, 'eval/demon_right_precision': 0.8563049853372434, 'eval/demon_right_recall': 0.5270758122743683, 'eval/demon_right_f1': 0.652513966480447, 'eval/demon_right_rate_model': 0.10014684287812041, 'eval/demon_right_rate_teacher': 0.16270190895741557, 'eval/demon_move_f1': 0.6056687479461058, 'eval/demon_cc_f1': 0.8198945196568818, 'eval/latency_4/demon_fire_precision': 0.948405253283302, 'eval/latency_4/demon_fire_recall': 0.985060084443001, 'eval/latency_4/demon_fire_f1': 0.9663852158674526, 'eval/latency_4/demon_fire_rate_model': 0.9392070484581497, 'eval/latency_4/demon_fire_rate_teacher': 0.9042584434654919, 'eval/latency_4/demon_left_precision': 0.7851239669421488, 'eval/latency_4/demon_left_recall': 0.4337899543378995, 'eval/latency_4/demon_left_f1': 0.5588235294117646, 'eval/latency_4/demon_left_rate_model': 0.07107195301027901, 'eval/latency_4/demon_left_rate_teacher': 0.12863436123348018, 'eval/latency_4/demon_right_precision': 0.8563049853372434, 'eval/latency_4/demon_right_recall': 0.5270758122743683, 'eval/latency_4/demon_right_f1': 0.652513966480447, 'eval/latency_4/demon_right_rate_model': 0.10014684287812041, 'eval/latency_4/demon_right_rate_teacher': 0.16270190895741557, 'eval/latency_4/demon_move_f1': 0.6056687479461058, 'eval/latency_4/demon_cc_f1': 0.8198945196568818, 'timing/data': 0.011920344084501266, 'timing/model': 1.8792779184877872, 'learning_rate/qwen_vl_interface': 8.934628630845994e-06, 'learning_rate/action_model': 8.934628630845995e-05, 'epoch': 0.62}) +[2026-06-17 19:37:53,881][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_1600_state +[2026-06-17 19:37:53,889][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 19:38:02,894][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_1600_state/model.safetensors +[2026-06-17 19:38:25,270][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_1600_state/optimizer.bin +[2026-06-17 19:38:25,273][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_1600_state/sampler.bin +[2026-06-17 19:38:25,273][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_1600_state/sampler_1.bin +[2026-06-17 19:38:25,317][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_1600_state/random_states_0.pkl +[2026-06-17 19:38:25,338][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1600 +[2026-06-17 19:38:25,338][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 19:38:25,356][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 19:38:30,253][starVLA.training.train_starvla][INFO] - Step 1601, Loss: {'train/loss': 0.6714824438095093, 'train/grad_norm_pre_clip': 9.59029483795166, 'timing/data': 0.013898644596338272, 'timing/model': 1.115604349412024, 'learning_rate/qwen_vl_interface': 8.933263401017685e-06, 'learning_rate/action_model': 8.933263401017684e-05, 'epoch': 0.62}) +[2026-06-17 19:38:35,046][starVLA.training.train_starvla][INFO] - Step 1602, Loss: {'train/loss': 0.9247544407844543, 'train/grad_norm_pre_clip': 13.48032283782959, 'timing/data': 0.013831290416419506, 'timing/model': 1.5310387862846255, 'learning_rate/qwen_vl_interface': 8.93189740764491e-06, 'learning_rate/action_model': 8.931897407644909e-05, 'epoch': 0.62}) +[2026-06-17 19:38:39,442][starVLA.training.train_starvla][INFO] - Step 1603, Loss: {'train/loss': 0.38798364996910095, 'train/grad_norm_pre_clip': 5.601204872131348, 'timing/data': 0.015406269580125809, 'timing/model': 1.1298937136307359, 'learning_rate/qwen_vl_interface': 8.93053065101084e-06, 'learning_rate/action_model': 8.93053065101084e-05, 'epoch': 0.62}) +[2026-06-17 19:38:44,664][starVLA.training.train_starvla][INFO] - Step 1604, Loss: {'train/loss': 0.7444983124732971, 'train/grad_norm_pre_clip': 9.195725440979004, 'timing/data': 0.016878092661499977, 'timing/model': 1.482226082123816, 'learning_rate/qwen_vl_interface': 8.929163131398804e-06, 'learning_rate/action_model': 8.929163131398805e-05, 'epoch': 0.62}) +[2026-06-17 19:38:49,311][starVLA.training.train_starvla][INFO] - Step 1605, Loss: {'train/loss': 0.4383071959018707, 'train/grad_norm_pre_clip': 5.044400215148926, 'timing/data': 0.012609073892235756, 'timing/model': 1.0684205023571849, 'learning_rate/qwen_vl_interface': 8.927794849092294e-06, 'learning_rate/action_model': 8.927794849092293e-05, 'epoch': 0.62}) +[2026-06-17 19:38:54,171][starVLA.training.train_starvla][INFO] - Step 1606, Loss: {'train/loss': 0.6142145395278931, 'train/grad_norm_pre_clip': 6.542531490325928, 'timing/data': 0.013733988627791405, 'timing/model': 1.1189822209998965, 'learning_rate/qwen_vl_interface': 8.926425804374953e-06, 'learning_rate/action_model': 8.926425804374954e-05, 'epoch': 0.62}) +[2026-06-17 19:38:58,657][starVLA.training.train_starvla][INFO] - Step 1607, Loss: {'train/loss': 0.528033971786499, 'train/grad_norm_pre_clip': 13.110267639160156, 'timing/data': 0.012583010829985142, 'timing/model': 1.118133551441133, 'learning_rate/qwen_vl_interface': 8.925055997530587e-06, 'learning_rate/action_model': 8.925055997530587e-05, 'epoch': 0.62}) +[2026-06-17 19:39:03,445][starVLA.training.train_starvla][INFO] - Step 1608, Loss: {'train/loss': 0.37152212858200073, 'train/grad_norm_pre_clip': 5.8879923820495605, 'timing/data': 0.01558750681579113, 'timing/model': 1.0888585485517979, 'learning_rate/qwen_vl_interface': 8.92368542884316e-06, 'learning_rate/action_model': 8.923685428843159e-05, 'epoch': 0.62}) +[2026-06-17 19:39:09,662][starVLA.training.train_starvla][INFO] - Step 1609, Loss: {'train/loss': 0.648556649684906, 'train/grad_norm_pre_clip': 5.775754928588867, 'timing/data': 0.011794806458055973, 'timing/model': 1.1698202630504966, 'learning_rate/qwen_vl_interface': 8.922314098596789e-06, 'learning_rate/action_model': 8.922314098596789e-05, 'epoch': 0.62}) +[2026-06-17 19:39:14,380][starVLA.training.train_starvla][INFO] - Step 1610, Loss: {'train/loss': 0.6507195830345154, 'train/grad_norm_pre_clip': 16.72694206237793, 'timing/data': 0.0248889634385705, 'timing/model': 1.1203035339713097, 'learning_rate/qwen_vl_interface': 8.920942007075752e-06, 'learning_rate/action_model': 8.920942007075752e-05, 'epoch': 0.62}) +[2026-06-17 19:39:19,100][starVLA.training.train_starvla][INFO] - Step 1611, Loss: {'train/loss': 0.752231240272522, 'train/grad_norm_pre_clip': 8.116271018981934, 'timing/data': 0.01840564701706171, 'timing/model': 1.5560972383245826, 'learning_rate/qwen_vl_interface': 8.91956915456449e-06, 'learning_rate/action_model': 8.91956915456449e-05, 'epoch': 0.62}) +[2026-06-17 19:39:25,986][starVLA.training.train_starvla][INFO] - Step 1612, Loss: {'train/loss': 0.695962131023407, 'train/grad_norm_pre_clip': 8.27238941192627, 'timing/data': 0.011961610056459904, 'timing/model': 1.6006383737549186, 'learning_rate/qwen_vl_interface': 8.918195541347593e-06, 'learning_rate/action_model': 8.918195541347593e-05, 'epoch': 0.62}) +[2026-06-17 19:39:30,311][starVLA.training.train_starvla][INFO] - Step 1613, Loss: {'train/loss': 0.6315962672233582, 'train/grad_norm_pre_clip': 7.748058319091797, 'timing/data': 0.014830470085144043, 'timing/model': 1.0998755916953087, 'learning_rate/qwen_vl_interface': 8.916821167709811e-06, 'learning_rate/action_model': 8.916821167709812e-05, 'epoch': 0.62}) +[2026-06-17 19:39:34,974][starVLA.training.train_starvla][INFO] - Step 1614, Loss: {'train/loss': 0.4183202385902405, 'train/grad_norm_pre_clip': 5.407803058624268, 'timing/data': 0.011620559729635715, 'timing/model': 1.07753886282444, 'learning_rate/qwen_vl_interface': 8.915446033936057e-06, 'learning_rate/action_model': 8.915446033936056e-05, 'epoch': 0.62}) +[2026-06-17 19:39:39,560][starVLA.training.train_starvla][INFO] - Step 1615, Loss: {'train/loss': 0.3735849857330322, 'train/grad_norm_pre_clip': 6.9376420974731445, 'timing/data': 0.02918749302625656, 'timing/model': 1.097108367830515, 'learning_rate/qwen_vl_interface': 8.914070140311395e-06, 'learning_rate/action_model': 8.914070140311395e-05, 'epoch': 0.62}) +[2026-06-17 19:39:44,408][starVLA.training.train_starvla][INFO] - Step 1616, Loss: {'train/loss': 0.9224916100502014, 'train/grad_norm_pre_clip': 10.431755065917969, 'timing/data': 0.012812417931854725, 'timing/model': 1.1925509376451373, 'learning_rate/qwen_vl_interface': 8.912693487121053e-06, 'learning_rate/action_model': 8.912693487121052e-05, 'epoch': 0.62}) +[2026-06-17 19:39:51,891][starVLA.training.train_starvla][INFO] - Step 1617, Loss: {'train/loss': 0.5263071656227112, 'train/grad_norm_pre_clip': 6.598260402679443, 'timing/data': 0.015912740491330624, 'timing/model': 2.0662891240790486, 'learning_rate/qwen_vl_interface': 8.911316074650407e-06, 'learning_rate/action_model': 8.911316074650406e-05, 'epoch': 0.62}) +[2026-06-17 19:39:56,372][starVLA.training.train_starvla][INFO] - Step 1618, Loss: {'train/loss': 0.36561280488967896, 'train/grad_norm_pre_clip': 5.40064001083374, 'timing/data': 0.016797023825347424, 'timing/model': 1.0727660888805985, 'learning_rate/qwen_vl_interface': 8.909937903185e-06, 'learning_rate/action_model': 8.909937903185e-05, 'epoch': 0.62}) +[2026-06-17 19:40:04,284][starVLA.training.train_starvla][INFO] - Step 1619, Loss: {'train/loss': 0.5894021391868591, 'train/grad_norm_pre_clip': 6.499917030334473, 'timing/data': 0.011422964744269848, 'timing/model': 1.2925391038879752, 'learning_rate/qwen_vl_interface': 8.908558973010528e-06, 'learning_rate/action_model': 8.908558973010526e-05, 'epoch': 0.62}) +[2026-06-17 19:40:11,242][starVLA.training.train_starvla][INFO] - Step 1620, Loss: {'train/loss': 0.7552499771118164, 'train/grad_norm_pre_clip': 8.731117248535156, 'timing/data': 0.01224228460341692, 'timing/model': 3.2308499682694674, 'learning_rate/qwen_vl_interface': 8.907179284412841e-06, 'learning_rate/action_model': 8.90717928441284e-05, 'epoch': 0.62}) +[2026-06-17 19:40:15,865][starVLA.training.train_starvla][INFO] - Step 1621, Loss: {'train/loss': 0.49140092730522156, 'train/grad_norm_pre_clip': 5.603271007537842, 'timing/data': 0.013650157488882542, 'timing/model': 1.0948741901665926, 'learning_rate/qwen_vl_interface': 8.905798837677956e-06, 'learning_rate/action_model': 8.905798837677956e-05, 'epoch': 0.62}) +[2026-06-17 19:40:20,728][starVLA.training.train_starvla][INFO] - Step 1622, Loss: {'train/loss': 0.6320763826370239, 'train/grad_norm_pre_clip': 9.181286811828613, 'timing/data': 0.011204502545297146, 'timing/model': 1.615073997527361, 'learning_rate/qwen_vl_interface': 8.904417633092037e-06, 'learning_rate/action_model': 8.904417633092037e-05, 'epoch': 0.62}) +[2026-06-17 19:40:25,085][starVLA.training.train_starvla][INFO] - Step 1623, Loss: {'train/loss': 0.6458603739738464, 'train/grad_norm_pre_clip': 9.334076881408691, 'timing/data': 0.013493032194674015, 'timing/model': 1.1432028310373425, 'learning_rate/qwen_vl_interface': 8.903035670941411e-06, 'learning_rate/action_model': 8.903035670941412e-05, 'epoch': 0.62}) +[2026-06-17 19:40:29,731][starVLA.training.train_starvla][INFO] - Step 1624, Loss: {'train/loss': 0.7139051556587219, 'train/grad_norm_pre_clip': 23.705848693847656, 'timing/data': 0.012018217705190182, 'timing/model': 1.1371343089267612, 'learning_rate/qwen_vl_interface': 8.901652951512559e-06, 'learning_rate/action_model': 8.90165295151256e-05, 'epoch': 0.62}) +[2026-06-17 19:40:34,383][starVLA.training.train_starvla][INFO] - Step 1625, Loss: {'train/loss': 0.6593157649040222, 'train/grad_norm_pre_clip': 6.928936004638672, 'timing/data': 0.012506937608122826, 'timing/model': 1.1386149236932397, 'learning_rate/qwen_vl_interface': 8.900269475092122e-06, 'learning_rate/action_model': 8.900269475092122e-05, 'epoch': 0.63}) +[2026-06-17 19:40:39,077][starVLA.training.train_starvla][INFO] - Step 1626, Loss: {'train/loss': 0.4472506642341614, 'train/grad_norm_pre_clip': 7.070297718048096, 'timing/data': 0.01348189078271389, 'timing/model': 1.0925070010125637, 'learning_rate/qwen_vl_interface': 8.898885241966894e-06, 'learning_rate/action_model': 8.898885241966893e-05, 'epoch': 0.63}) +[2026-06-17 19:40:43,882][starVLA.training.train_starvla][INFO] - Step 1627, Loss: {'train/loss': 0.30109408497810364, 'train/grad_norm_pre_clip': 4.417644023895264, 'timing/data': 0.019680672325193882, 'timing/model': 1.55293099116534, 'learning_rate/qwen_vl_interface': 8.89750025242383e-06, 'learning_rate/action_model': 8.89750025242383e-05, 'epoch': 0.63}) +[2026-06-17 19:40:48,352][starVLA.training.train_starvla][INFO] - Step 1628, Loss: {'train/loss': 0.5394414663314819, 'train/grad_norm_pre_clip': 8.811939239501953, 'timing/data': 0.011460223235189915, 'timing/model': 1.1487574586644769, 'learning_rate/qwen_vl_interface': 8.896114506750037e-06, 'learning_rate/action_model': 8.896114506750036e-05, 'epoch': 0.63}) +[2026-06-17 19:40:52,963][starVLA.training.train_starvla][INFO] - Step 1629, Loss: {'train/loss': 0.8316143155097961, 'train/grad_norm_pre_clip': 10.18173885345459, 'timing/data': 0.01282366830855608, 'timing/model': 1.1273501943796873, 'learning_rate/qwen_vl_interface': 8.894728005232785e-06, 'learning_rate/action_model': 8.894728005232785e-05, 'epoch': 0.63}) +[2026-06-17 19:40:57,908][starVLA.training.train_starvla][INFO] - Step 1630, Loss: {'train/loss': 0.6989571452140808, 'train/grad_norm_pre_clip': 8.080828666687012, 'timing/data': 0.015868590213358402, 'timing/model': 1.1581986732780933, 'learning_rate/qwen_vl_interface': 8.893340748159494e-06, 'learning_rate/action_model': 8.893340748159493e-05, 'epoch': 0.63}) +[2026-06-17 19:41:02,573][starVLA.training.train_starvla][INFO] - Step 1631, Loss: {'train/loss': 0.49699875712394714, 'train/grad_norm_pre_clip': 6.359292030334473, 'timing/data': 0.011287890374660492, 'timing/model': 1.2016468569636345, 'learning_rate/qwen_vl_interface': 8.891952735817746e-06, 'learning_rate/action_model': 8.891952735817745e-05, 'epoch': 0.63}) +[2026-06-17 19:41:07,292][starVLA.training.train_starvla][INFO] - Step 1632, Loss: {'train/loss': 0.4102395176887512, 'train/grad_norm_pre_clip': 7.5049004554748535, 'timing/data': 0.01185029000043869, 'timing/model': 1.4929422568529844, 'learning_rate/qwen_vl_interface': 8.890563968495278e-06, 'learning_rate/action_model': 8.890563968495278e-05, 'epoch': 0.63}) +[2026-06-17 19:41:11,702][starVLA.training.train_starvla][INFO] - Step 1633, Loss: {'train/loss': 0.37904149293899536, 'train/grad_norm_pre_clip': 4.256283760070801, 'timing/data': 0.021092827431857586, 'timing/model': 1.1189108928665519, 'learning_rate/qwen_vl_interface': 8.88917444647998e-06, 'learning_rate/action_model': 8.889174446479981e-05, 'epoch': 0.63}) +[2026-06-17 19:41:21,161][starVLA.training.train_starvla][INFO] - Step 1634, Loss: {'train/loss': 0.47164490818977356, 'train/grad_norm_pre_clip': 7.834030628204346, 'timing/data': 0.011991455219686031, 'timing/model': 1.08895668014884, 'learning_rate/qwen_vl_interface': 8.887784170059905e-06, 'learning_rate/action_model': 8.887784170059905e-05, 'epoch': 0.63}) +[2026-06-17 19:41:25,883][starVLA.training.train_starvla][INFO] - Step 1635, Loss: {'train/loss': 1.148350477218628, 'train/grad_norm_pre_clip': 9.143616676330566, 'timing/data': 0.015711828134953976, 'timing/model': 1.1150644663721323, 'learning_rate/qwen_vl_interface': 8.886393139523258e-06, 'learning_rate/action_model': 8.886393139523257e-05, 'epoch': 0.63}) +[2026-06-17 19:41:34,329][starVLA.training.train_starvla][INFO] - Step 1636, Loss: {'train/loss': 0.6555549502372742, 'train/grad_norm_pre_clip': 11.027543067932129, 'timing/data': 0.012260526418685913, 'timing/model': 1.108186922967434, 'learning_rate/qwen_vl_interface': 8.885001355158397e-06, 'learning_rate/action_model': 8.885001355158397e-05, 'epoch': 0.63}) +[2026-06-17 19:41:39,269][starVLA.training.train_starvla][INFO] - Step 1637, Loss: {'train/loss': 0.7193595767021179, 'train/grad_norm_pre_clip': 13.413644790649414, 'timing/data': 0.018095151521265507, 'timing/model': 1.495715861208737, 'learning_rate/qwen_vl_interface': 8.883608817253847e-06, 'learning_rate/action_model': 8.883608817253847e-05, 'epoch': 0.63}) +[2026-06-17 19:41:44,478][starVLA.training.train_starvla][INFO] - Step 1638, Loss: {'train/loss': 0.5514928698539734, 'train/grad_norm_pre_clip': 6.803822040557861, 'timing/data': 0.012289583683013916, 'timing/model': 1.07478285767138, 'learning_rate/qwen_vl_interface': 8.882215526098276e-06, 'learning_rate/action_model': 8.882215526098277e-05, 'epoch': 0.63}) +[2026-06-17 19:41:51,187][starVLA.training.train_starvla][INFO] - Step 1639, Loss: {'train/loss': 0.5188993215560913, 'train/grad_norm_pre_clip': 5.9859089851379395, 'timing/data': 0.014017459005117416, 'timing/model': 1.1337561523541808, 'learning_rate/qwen_vl_interface': 8.880821481980521e-06, 'learning_rate/action_model': 8.880821481980521e-05, 'epoch': 0.63}) +[2026-06-17 19:42:00,709][starVLA.training.train_starvla][INFO] - Step 1640, Loss: {'train/loss': 0.3853735029697418, 'train/grad_norm_pre_clip': 8.401865005493164, 'timing/data': 0.017216289415955544, 'timing/model': 3.4863530853763223, 'learning_rate/qwen_vl_interface': 8.879426685189565e-06, 'learning_rate/action_model': 8.879426685189565e-05, 'epoch': 0.63}) +[2026-06-17 19:42:05,370][starVLA.training.train_starvla][INFO] - Step 1641, Loss: {'train/loss': 0.37010812759399414, 'train/grad_norm_pre_clip': 6.080718994140625, 'timing/data': 0.012610156089067459, 'timing/model': 1.162698975764215, 'learning_rate/qwen_vl_interface': 8.878031136014552e-06, 'learning_rate/action_model': 8.878031136014551e-05, 'epoch': 0.63}) +[2026-06-17 19:42:10,031][starVLA.training.train_starvla][INFO] - Step 1642, Loss: {'train/loss': 0.5989086627960205, 'train/grad_norm_pre_clip': 7.368249893188477, 'timing/data': 0.012060959823429585, 'timing/model': 1.5067799696698785, 'learning_rate/qwen_vl_interface': 8.87663483474478e-06, 'learning_rate/action_model': 8.87663483474478e-05, 'epoch': 0.63}) +[2026-06-17 19:42:14,476][starVLA.training.train_starvla][INFO] - Step 1643, Loss: {'train/loss': 0.7989389896392822, 'train/grad_norm_pre_clip': 7.391462326049805, 'timing/data': 0.011523909866809845, 'timing/model': 1.1773070637136698, 'learning_rate/qwen_vl_interface': 8.875237781669704e-06, 'learning_rate/action_model': 8.875237781669703e-05, 'epoch': 0.63}) +[2026-06-17 19:42:19,336][starVLA.training.train_starvla][INFO] - Step 1644, Loss: {'train/loss': 0.8569590449333191, 'train/grad_norm_pre_clip': 8.097455024719238, 'timing/data': 0.011842446401715279, 'timing/model': 1.1515638874843717, 'learning_rate/qwen_vl_interface': 8.873839977078933e-06, 'learning_rate/action_model': 8.873839977078934e-05, 'epoch': 0.63}) +[2026-06-17 19:42:24,171][starVLA.training.train_starvla][INFO] - Step 1645, Loss: {'train/loss': 0.3839455544948578, 'train/grad_norm_pre_clip': 5.394680976867676, 'timing/data': 0.014057389460504055, 'timing/model': 1.1331277200952172, 'learning_rate/qwen_vl_interface': 8.87244142126224e-06, 'learning_rate/action_model': 8.87244142126224e-05, 'epoch': 0.63}) +[2026-06-17 19:42:28,842][starVLA.training.train_starvla][INFO] - Step 1646, Loss: {'train/loss': 0.4687337875366211, 'train/grad_norm_pre_clip': 5.03980016708374, 'timing/data': 0.01207058411091566, 'timing/model': 1.1655087620019913, 'learning_rate/qwen_vl_interface': 8.871042114509539e-06, 'learning_rate/action_model': 8.87104211450954e-05, 'epoch': 0.63}) +[2026-06-17 19:42:33,490][starVLA.training.train_starvla][INFO] - Step 1647, Loss: {'train/loss': 0.6228846311569214, 'train/grad_norm_pre_clip': 6.634808540344238, 'timing/data': 0.010839836671948433, 'timing/model': 1.4436686662957072, 'learning_rate/qwen_vl_interface': 8.869642057110912e-06, 'learning_rate/action_model': 8.869642057110911e-05, 'epoch': 0.63}) +[2026-06-17 19:42:37,959][starVLA.training.train_starvla][INFO] - Step 1648, Loss: {'train/loss': 0.4782126247882843, 'train/grad_norm_pre_clip': 6.826226234436035, 'timing/data': 0.014967188239097595, 'timing/model': 1.207539259456098, 'learning_rate/qwen_vl_interface': 8.868241249356592e-06, 'learning_rate/action_model': 8.868241249356592e-05, 'epoch': 0.63}) +[2026-06-17 19:42:42,781][starVLA.training.train_starvla][INFO] - Step 1649, Loss: {'train/loss': 0.46314242482185364, 'train/grad_norm_pre_clip': 5.783671855926514, 'timing/data': 0.01302388682961464, 'timing/model': 1.0974990306422114, 'learning_rate/qwen_vl_interface': 8.866839691536968e-06, 'learning_rate/action_model': 8.866839691536968e-05, 'epoch': 0.63}) +[2026-06-17 19:42:47,805][starVLA.training.train_starvla][INFO] - Step 1650, Loss: {'train/loss': 0.3868858516216278, 'train/grad_norm_pre_clip': 5.81790018081665, 'timing/data': 0.011525018140673637, 'timing/model': 1.2990814838558435, 'learning_rate/qwen_vl_interface': 8.865437383942584e-06, 'learning_rate/action_model': 8.865437383942583e-05, 'epoch': 0.63}) +[2026-06-17 19:42:52,513][starVLA.training.train_starvla][INFO] - Step 1651, Loss: {'train/loss': 0.3061256408691406, 'train/grad_norm_pre_clip': 5.006444931030273, 'timing/data': 0.012440218590199947, 'timing/model': 1.1606269301846623, 'learning_rate/qwen_vl_interface': 8.86403432686414e-06, 'learning_rate/action_model': 8.86403432686414e-05, 'epoch': 0.64}) +[2026-06-17 19:42:57,563][starVLA.training.train_starvla][INFO] - Step 1652, Loss: {'train/loss': 0.5896598696708679, 'train/grad_norm_pre_clip': 8.908804893493652, 'timing/data': 0.0681407880038023, 'timing/model': 1.6496982667595148, 'learning_rate/qwen_vl_interface': 8.86263052059249e-06, 'learning_rate/action_model': 8.862630520592489e-05, 'epoch': 0.64}) +[2026-06-17 19:43:01,947][starVLA.training.train_starvla][INFO] - Step 1653, Loss: {'train/loss': 0.3484691381454468, 'train/grad_norm_pre_clip': 6.346061706542969, 'timing/data': 0.011870298534631729, 'timing/model': 1.2167191691696644, 'learning_rate/qwen_vl_interface': 8.861225965418647e-06, 'learning_rate/action_model': 8.861225965418646e-05, 'epoch': 0.64}) +[2026-06-17 19:43:07,957][starVLA.training.train_starvla][INFO] - Step 1654, Loss: {'train/loss': 0.3208825886249542, 'train/grad_norm_pre_clip': 5.6166253089904785, 'timing/data': 0.021226363256573677, 'timing/model': 1.1147017078474164, 'learning_rate/qwen_vl_interface': 8.859820661633775e-06, 'learning_rate/action_model': 8.859820661633775e-05, 'epoch': 0.64}) +[2026-06-17 19:43:15,197][starVLA.training.train_starvla][INFO] - Step 1655, Loss: {'train/loss': 0.8165789842605591, 'train/grad_norm_pre_clip': 9.828385353088379, 'timing/data': 0.01242340262979269, 'timing/model': 1.212870362214744, 'learning_rate/qwen_vl_interface': 8.858414609529193e-06, 'learning_rate/action_model': 8.858414609529193e-05, 'epoch': 0.64}) +[2026-06-17 19:43:19,925][starVLA.training.train_starvla][INFO] - Step 1656, Loss: {'train/loss': 0.5378336310386658, 'train/grad_norm_pre_clip': 8.017998695373535, 'timing/data': 0.014497412368655205, 'timing/model': 1.1152943689376116, 'learning_rate/qwen_vl_interface': 8.857007809396384e-06, 'learning_rate/action_model': 8.857007809396383e-05, 'epoch': 0.64}) +[2026-06-17 19:43:24,587][starVLA.training.train_starvla][INFO] - Step 1657, Loss: {'train/loss': 0.5028579831123352, 'train/grad_norm_pre_clip': 8.47717571258545, 'timing/data': 0.012766064144670963, 'timing/model': 1.4561379002407193, 'learning_rate/qwen_vl_interface': 8.855600261526973e-06, 'learning_rate/action_model': 8.855600261526973e-05, 'epoch': 0.64}) +[2026-06-17 19:43:31,102][starVLA.training.train_starvla][INFO] - Step 1658, Loss: {'train/loss': 0.7294842004776001, 'train/grad_norm_pre_clip': 8.166823387145996, 'timing/data': 0.012690690346062183, 'timing/model': 1.1050442131236196, 'learning_rate/qwen_vl_interface': 8.854191966212747e-06, 'learning_rate/action_model': 8.854191966212748e-05, 'epoch': 0.64}) +[2026-06-17 19:43:39,088][starVLA.training.train_starvla][INFO] - Step 1659, Loss: {'train/loss': 0.6911196112632751, 'train/grad_norm_pre_clip': 9.046831130981445, 'timing/data': 0.01351245678961277, 'timing/model': 3.1523100063204765, 'learning_rate/qwen_vl_interface': 8.852782923745648e-06, 'learning_rate/action_model': 8.852782923745648e-05, 'epoch': 0.64}) +[2026-06-17 19:43:43,936][starVLA.training.train_starvla][INFO] - Step 1660, Loss: {'train/loss': 0.5112605690956116, 'train/grad_norm_pre_clip': 5.2638750076293945, 'timing/data': 0.017597805708646774, 'timing/model': 1.2160989129915833, 'learning_rate/qwen_vl_interface': 8.851373134417773e-06, 'learning_rate/action_model': 8.851373134417773e-05, 'epoch': 0.64}) +[2026-06-17 19:43:48,662][starVLA.training.train_starvla][INFO] - Step 1661, Loss: {'train/loss': 0.8089820742607117, 'train/grad_norm_pre_clip': 10.450631141662598, 'timing/data': 0.013677173294126987, 'timing/model': 1.1682517947629094, 'learning_rate/qwen_vl_interface': 8.849962598521373e-06, 'learning_rate/action_model': 8.849962598521373e-05, 'epoch': 0.64}) +[2026-06-17 19:43:53,640][starVLA.training.train_starvla][INFO] - Step 1662, Loss: {'train/loss': 0.5501742959022522, 'train/grad_norm_pre_clip': 9.268239974975586, 'timing/data': 0.013165963813662529, 'timing/model': 1.5783074777573347, 'learning_rate/qwen_vl_interface': 8.848551316348852e-06, 'learning_rate/action_model': 8.848551316348853e-05, 'epoch': 0.64}) +[2026-06-17 19:43:59,161][starVLA.training.train_starvla][INFO] - Step 1663, Loss: {'train/loss': 0.5952946543693542, 'train/grad_norm_pre_clip': 9.249100685119629, 'timing/data': 0.019807524047791958, 'timing/model': 1.1546091493219137, 'learning_rate/qwen_vl_interface': 8.847139288192771e-06, 'learning_rate/action_model': 8.84713928819277e-05, 'epoch': 0.64}) +[2026-06-17 19:44:06,635][starVLA.training.train_starvla][INFO] - Step 1664, Loss: {'train/loss': 0.7908050417900085, 'train/grad_norm_pre_clip': 12.195535659790039, 'timing/data': 0.012695161625742912, 'timing/model': 1.1376589816063643, 'learning_rate/qwen_vl_interface': 8.845726514345845e-06, 'learning_rate/action_model': 8.845726514345843e-05, 'epoch': 0.64}) +[2026-06-17 19:44:11,556][starVLA.training.train_starvla][INFO] - Step 1665, Loss: {'train/loss': 0.4615522623062134, 'train/grad_norm_pre_clip': 5.917097091674805, 'timing/data': 0.012272610329091549, 'timing/model': 1.1831595869734883, 'learning_rate/qwen_vl_interface': 8.844312995100943e-06, 'learning_rate/action_model': 8.844312995100942e-05, 'epoch': 0.64}) +[2026-06-17 19:44:16,473][starVLA.training.train_starvla][INFO] - Step 1666, Loss: {'train/loss': 0.6824490427970886, 'train/grad_norm_pre_clip': 8.901351928710938, 'timing/data': 0.012170450761914253, 'timing/model': 1.1231382442638278, 'learning_rate/qwen_vl_interface': 8.842898730751088e-06, 'learning_rate/action_model': 8.842898730751087e-05, 'epoch': 0.64}) +[2026-06-17 19:44:21,769][starVLA.training.train_starvla][INFO] - Step 1667, Loss: {'train/loss': 0.5458546876907349, 'train/grad_norm_pre_clip': 6.934974193572998, 'timing/data': 0.011245453730225563, 'timing/model': 1.419154742732644, 'learning_rate/qwen_vl_interface': 8.841483721589462e-06, 'learning_rate/action_model': 8.84148372158946e-05, 'epoch': 0.64}) +[2026-06-17 19:44:26,250][starVLA.training.train_starvla][INFO] - Step 1668, Loss: {'train/loss': 0.42371562123298645, 'train/grad_norm_pre_clip': 10.415520668029785, 'timing/data': 0.014391317032277584, 'timing/model': 1.0918395705521107, 'learning_rate/qwen_vl_interface': 8.840067967909396e-06, 'learning_rate/action_model': 8.840067967909394e-05, 'epoch': 0.64}) +[2026-06-17 19:44:31,164][starVLA.training.train_starvla][INFO] - Step 1669, Loss: {'train/loss': 0.48855850100517273, 'train/grad_norm_pre_clip': 8.526015281677246, 'timing/data': 0.012359416112303734, 'timing/model': 1.1768096908926964, 'learning_rate/qwen_vl_interface': 8.838651470004376e-06, 'learning_rate/action_model': 8.838651470004375e-05, 'epoch': 0.64}) +[2026-06-17 19:44:35,865][starVLA.training.train_starvla][INFO] - Step 1670, Loss: {'train/loss': 0.3718150854110718, 'train/grad_norm_pre_clip': 7.261629581451416, 'timing/data': 0.011054715141654015, 'timing/model': 1.1018644021824002, 'learning_rate/qwen_vl_interface': 8.837234228168045e-06, 'learning_rate/action_model': 8.837234228168044e-05, 'epoch': 0.64}) +[2026-06-17 19:44:42,432][starVLA.training.train_starvla][INFO] - Step 1671, Loss: {'train/loss': 0.48269596695899963, 'train/grad_norm_pre_clip': 6.0992631912231445, 'timing/data': 0.01088854018598795, 'timing/model': 1.0768128195777535, 'learning_rate/qwen_vl_interface': 8.835816242694197e-06, 'learning_rate/action_model': 8.835816242694197e-05, 'epoch': 0.64}) +[2026-06-17 19:44:46,966][starVLA.training.train_starvla][INFO] - Step 1672, Loss: {'train/loss': 0.5024811625480652, 'train/grad_norm_pre_clip': 10.026437759399414, 'timing/data': 0.012596547603607178, 'timing/model': 1.408398224040866, 'learning_rate/qwen_vl_interface': 8.834397513876786e-06, 'learning_rate/action_model': 8.834397513876785e-05, 'epoch': 0.64}) +[2026-06-17 19:44:51,312][starVLA.training.train_starvla][INFO] - Step 1673, Loss: {'train/loss': 0.4764144718647003, 'train/grad_norm_pre_clip': 6.180002689361572, 'timing/data': 0.01503699366003275, 'timing/model': 1.2056139037013054, 'learning_rate/qwen_vl_interface': 8.832978042009913e-06, 'learning_rate/action_model': 8.832978042009912e-05, 'epoch': 0.64}) +[2026-06-17 19:44:56,058][starVLA.training.train_starvla][INFO] - Step 1674, Loss: {'train/loss': 0.8041319847106934, 'train/grad_norm_pre_clip': 11.722475051879883, 'timing/data': 0.011725956574082375, 'timing/model': 1.147005569189787, 'learning_rate/qwen_vl_interface': 8.831557827387835e-06, 'learning_rate/action_model': 8.831557827387834e-05, 'epoch': 0.64}) +[2026-06-17 19:45:00,804][starVLA.training.train_starvla][INFO] - Step 1675, Loss: {'train/loss': 0.8147211074829102, 'train/grad_norm_pre_clip': 10.339597702026367, 'timing/data': 0.011304729618132114, 'timing/model': 1.0759814474731684, 'learning_rate/qwen_vl_interface': 8.830136870304966e-06, 'learning_rate/action_model': 8.830136870304965e-05, 'epoch': 0.64}) +[2026-06-17 19:45:07,174][starVLA.training.train_starvla][INFO] - Step 1676, Loss: {'train/loss': 0.4975621700286865, 'train/grad_norm_pre_clip': 15.778258323669434, 'timing/data': 0.011601616628468037, 'timing/model': 1.1197117296978831, 'learning_rate/qwen_vl_interface': 8.828715171055871e-06, 'learning_rate/action_model': 8.828715171055871e-05, 'epoch': 0.64}) +[2026-06-17 19:45:11,896][starVLA.training.train_starvla][INFO] - Step 1677, Loss: {'train/loss': 0.5255117416381836, 'train/grad_norm_pre_clip': 7.91392183303833, 'timing/data': 0.011276323348283768, 'timing/model': 1.3793522901833057, 'learning_rate/qwen_vl_interface': 8.827292729935271e-06, 'learning_rate/action_model': 8.827292729935271e-05, 'epoch': 0.65}) +[2026-06-17 19:45:16,304][starVLA.training.train_starvla][INFO] - Step 1678, Loss: {'train/loss': 0.6295683979988098, 'train/grad_norm_pre_clip': 10.060158729553223, 'timing/data': 0.010512481443583965, 'timing/model': 1.2091709738597274, 'learning_rate/qwen_vl_interface': 8.825869547238038e-06, 'learning_rate/action_model': 8.825869547238038e-05, 'epoch': 0.65}) +[2026-06-17 19:45:23,045][starVLA.training.train_starvla][INFO] - Step 1679, Loss: {'train/loss': 0.5758703351020813, 'train/grad_norm_pre_clip': 10.923942565917969, 'timing/data': 0.013673661276698112, 'timing/model': 3.147468385286629, 'learning_rate/qwen_vl_interface': 8.824445623259202e-06, 'learning_rate/action_model': 8.8244456232592e-05, 'epoch': 0.65}) +[2026-06-17 19:45:27,970][starVLA.training.train_starvla][INFO] - Step 1680, Loss: {'train/loss': 0.7745459079742432, 'train/grad_norm_pre_clip': 10.018383979797363, 'timing/data': 0.01193283125758171, 'timing/model': 1.103356972336769, 'learning_rate/qwen_vl_interface': 8.82302095829394e-06, 'learning_rate/action_model': 8.82302095829394e-05, 'epoch': 0.65}) +[2026-06-17 19:45:32,874][starVLA.training.train_starvla][INFO] - Step 1681, Loss: {'train/loss': 0.5807380080223083, 'train/grad_norm_pre_clip': 13.21006965637207, 'timing/data': 0.012522841803729534, 'timing/model': 1.1912948321551085, 'learning_rate/qwen_vl_interface': 8.82159555263759e-06, 'learning_rate/action_model': 8.821595552637589e-05, 'epoch': 0.65}) +[2026-06-17 19:45:39,915][starVLA.training.train_starvla][INFO] - Step 1682, Loss: {'train/loss': 0.2817927300930023, 'train/grad_norm_pre_clip': 6.9969682693481445, 'timing/data': 0.011234060861170292, 'timing/model': 1.4621262941509485, 'learning_rate/qwen_vl_interface': 8.820169406585638e-06, 'learning_rate/action_model': 8.820169406585637e-05, 'epoch': 0.65}) +[2026-06-17 19:45:44,270][starVLA.training.train_starvla][INFO] - Step 1683, Loss: {'train/loss': 0.2572169899940491, 'train/grad_norm_pre_clip': 5.667392730712891, 'timing/data': 0.014836017973721027, 'timing/model': 1.118873443454504, 'learning_rate/qwen_vl_interface': 8.818742520433727e-06, 'learning_rate/action_model': 8.818742520433726e-05, 'epoch': 0.65}) +[2026-06-17 19:45:49,017][starVLA.training.train_starvla][INFO] - Step 1684, Loss: {'train/loss': 0.7326655387878418, 'train/grad_norm_pre_clip': 6.93480110168457, 'timing/data': 0.014204996638000011, 'timing/model': 1.1034584073349833, 'learning_rate/qwen_vl_interface': 8.81731489447765e-06, 'learning_rate/action_model': 8.81731489447765e-05, 'epoch': 0.65}) +[2026-06-17 19:45:53,743][starVLA.training.train_starvla][INFO] - Step 1685, Loss: {'train/loss': 0.4646235406398773, 'train/grad_norm_pre_clip': 6.916011810302734, 'timing/data': 0.011651381850242615, 'timing/model': 1.1258868426084518, 'learning_rate/qwen_vl_interface': 8.815886529013356e-06, 'learning_rate/action_model': 8.815886529013356e-05, 'epoch': 0.65}) +[2026-06-17 19:45:58,269][starVLA.training.train_starvla][INFO] - Step 1686, Loss: {'train/loss': 0.48882779479026794, 'train/grad_norm_pre_clip': 7.438896656036377, 'timing/data': 0.011813848279416561, 'timing/model': 1.0648161880671978, 'learning_rate/qwen_vl_interface': 8.814457424336951e-06, 'learning_rate/action_model': 8.81445742433695e-05, 'epoch': 0.65}) +[2026-06-17 19:46:02,984][starVLA.training.train_starvla][INFO] - Step 1687, Loss: {'train/loss': 0.5824940800666809, 'train/grad_norm_pre_clip': 13.46390151977539, 'timing/data': 0.01508356723934412, 'timing/model': 1.5130244493484497, 'learning_rate/qwen_vl_interface': 8.813027580744685e-06, 'learning_rate/action_model': 8.813027580744684e-05, 'epoch': 0.65}) +[2026-06-17 19:46:07,496][starVLA.training.train_starvla][INFO] - Step 1688, Loss: {'train/loss': 0.5018190741539001, 'train/grad_norm_pre_clip': 7.466568470001221, 'timing/data': 0.013302989304065704, 'timing/model': 1.0759916119277477, 'learning_rate/qwen_vl_interface': 8.811596998532965e-06, 'learning_rate/action_model': 8.811596998532965e-05, 'epoch': 0.65}) +[2026-06-17 19:46:14,867][starVLA.training.train_starvla][INFO] - Step 1689, Loss: {'train/loss': 0.5243136882781982, 'train/grad_norm_pre_clip': 8.614657402038574, 'timing/data': 0.011127407662570477, 'timing/model': 3.5461619151756167, 'learning_rate/qwen_vl_interface': 8.810165677998356e-06, 'learning_rate/action_model': 8.810165677998356e-05, 'epoch': 0.65}) +[2026-06-17 19:46:19,566][starVLA.training.train_starvla][INFO] - Step 1690, Loss: {'train/loss': 0.3839512765407562, 'train/grad_norm_pre_clip': 5.145091533660889, 'timing/data': 0.011277926154434681, 'timing/model': 1.1083874786272645, 'learning_rate/qwen_vl_interface': 8.808733619437572e-06, 'learning_rate/action_model': 8.808733619437572e-05, 'epoch': 0.65}) +[2026-06-17 19:46:29,066][starVLA.training.train_starvla][INFO] - Step 1691, Loss: {'train/loss': 0.690247654914856, 'train/grad_norm_pre_clip': 8.333353996276855, 'timing/data': 0.011137791909277439, 'timing/model': 1.1549521582201123, 'learning_rate/qwen_vl_interface': 8.80730082314748e-06, 'learning_rate/action_model': 8.80730082314748e-05, 'epoch': 0.65}) +[2026-06-17 19:46:37,007][starVLA.training.train_starvla][INFO] - Step 1692, Loss: {'train/loss': 0.3862321078777313, 'train/grad_norm_pre_clip': 5.348722457885742, 'timing/data': 0.017225509509444237, 'timing/model': 1.6172044463455677, 'learning_rate/qwen_vl_interface': 8.805867289425098e-06, 'learning_rate/action_model': 8.805867289425098e-05, 'epoch': 0.65}) +[2026-06-17 19:46:41,468][starVLA.training.train_starvla][INFO] - Step 1693, Loss: {'train/loss': 0.6686798930168152, 'train/grad_norm_pre_clip': 11.671618461608887, 'timing/data': 0.013744251802563667, 'timing/model': 1.0752938902005553, 'learning_rate/qwen_vl_interface': 8.804433018567601e-06, 'learning_rate/action_model': 8.8044330185676e-05, 'epoch': 0.65}) +[2026-06-17 19:46:46,384][starVLA.training.train_starvla][INFO] - Step 1694, Loss: {'train/loss': 0.5582987666130066, 'train/grad_norm_pre_clip': 7.35921049118042, 'timing/data': 0.012081802822649479, 'timing/model': 1.3102270886301994, 'learning_rate/qwen_vl_interface': 8.802998010872316e-06, 'learning_rate/action_model': 8.802998010872316e-05, 'epoch': 0.65}) +[2026-06-17 19:46:52,828][starVLA.training.train_starvla][INFO] - Step 1695, Loss: {'train/loss': 0.6314796805381775, 'train/grad_norm_pre_clip': 6.817811965942383, 'timing/data': 0.011840090155601501, 'timing/model': 1.0824342835694551, 'learning_rate/qwen_vl_interface': 8.801562266636719e-06, 'learning_rate/action_model': 8.801562266636719e-05, 'epoch': 0.65}) +[2026-06-17 19:46:57,744][starVLA.training.train_starvla][INFO] - Step 1696, Loss: {'train/loss': 0.4903596043586731, 'train/grad_norm_pre_clip': 7.995787620544434, 'timing/data': 0.01350769866257906, 'timing/model': 1.2809835318475962, 'learning_rate/qwen_vl_interface': 8.800125786158442e-06, 'learning_rate/action_model': 8.800125786158442e-05, 'epoch': 0.65}) +[2026-06-17 19:47:04,307][starVLA.training.train_starvla][INFO] - Step 1697, Loss: {'train/loss': 0.37139058113098145, 'train/grad_norm_pre_clip': 6.339028835296631, 'timing/data': 0.015303079970180988, 'timing/model': 1.5184465134516358, 'learning_rate/qwen_vl_interface': 8.798688569735268e-06, 'learning_rate/action_model': 8.798688569735268e-05, 'epoch': 0.65}) +[2026-06-17 19:47:08,739][starVLA.training.train_starvla][INFO] - Step 1698, Loss: {'train/loss': 0.42138999700546265, 'train/grad_norm_pre_clip': 8.302556991577148, 'timing/data': 0.01348997000604868, 'timing/model': 1.1104568215087056, 'learning_rate/qwen_vl_interface': 8.797250617665139e-06, 'learning_rate/action_model': 8.797250617665138e-05, 'epoch': 0.65}) +[2026-06-17 19:47:13,555][starVLA.training.train_starvla][INFO] - Step 1699, Loss: {'train/loss': 0.45245593786239624, 'train/grad_norm_pre_clip': 7.0555195808410645, 'timing/data': 0.012807603925466537, 'timing/model': 1.1145584033802152, 'learning_rate/qwen_vl_interface': 8.795811930246136e-06, 'learning_rate/action_model': 8.795811930246136e-05, 'epoch': 0.65}) +[2026-06-17 19:49:41,905][starVLA.training.train_starvla][INFO] - Step 1700, Loss: {'train/loss': 0.3358452022075653, 'train/grad_norm_pre_clip': 8.475774765014648, 'eval/loss': 0.5762653350830078, 'eval/latency_4/loss': 0.5762653350830078, 'eval/demon_fire_precision': 0.9554020100502513, 'eval/demon_fire_recall': 0.9879831113998051, 'eval/demon_fire_f1': 0.9714194475490978, 'eval/demon_fire_rate_model': 0.9350954478707783, 'eval/demon_fire_rate_teacher': 0.9042584434654919, 'eval/demon_left_precision': 0.6599496221662469, 'eval/demon_left_recall': 0.5981735159817352, 'eval/demon_left_f1': 0.6275449101796408, 'eval/demon_left_rate_model': 0.11659324522760646, 'eval/demon_left_rate_teacher': 0.12863436123348018, 'eval/demon_right_precision': 0.8023255813953488, 'eval/demon_right_recall': 0.6227436823104693, 'eval/demon_right_f1': 0.7012195121951219, 'eval/demon_right_rate_model': 0.1262848751835536, 'eval/demon_right_rate_teacher': 0.16270190895741557, 'eval/demon_move_f1': 0.6643822111873814, 'eval/demon_cc_f1': 0.8531387264287807, 'eval/latency_4/demon_fire_precision': 0.9554020100502513, 'eval/latency_4/demon_fire_recall': 0.9879831113998051, 'eval/latency_4/demon_fire_f1': 0.9714194475490978, 'eval/latency_4/demon_fire_rate_model': 0.9350954478707783, 'eval/latency_4/demon_fire_rate_teacher': 0.9042584434654919, 'eval/latency_4/demon_left_precision': 0.6599496221662469, 'eval/latency_4/demon_left_recall': 0.5981735159817352, 'eval/latency_4/demon_left_f1': 0.6275449101796408, 'eval/latency_4/demon_left_rate_model': 0.11659324522760646, 'eval/latency_4/demon_left_rate_teacher': 0.12863436123348018, 'eval/latency_4/demon_right_precision': 0.8023255813953488, 'eval/latency_4/demon_right_recall': 0.6227436823104693, 'eval/latency_4/demon_right_f1': 0.7012195121951219, 'eval/latency_4/demon_right_rate_model': 0.1262848751835536, 'eval/latency_4/demon_right_rate_teacher': 0.16270190895741557, 'eval/latency_4/demon_move_f1': 0.6643822111873814, 'eval/latency_4/demon_cc_f1': 0.8531387264287807, 'timing/data': 0.015606746077537537, 'timing/model': 1.130719373933971, 'learning_rate/qwen_vl_interface': 8.794372507776508e-06, 'learning_rate/action_model': 8.794372507776508e-05, 'epoch': 0.65}) +[2026-06-17 19:49:41,922][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_1700_state +[2026-06-17 19:49:41,933][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 19:49:50,874][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_1700_state/model.safetensors +[2026-06-17 19:50:13,133][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_1700_state/optimizer.bin +[2026-06-17 19:50:13,134][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_1700_state/sampler.bin +[2026-06-17 19:50:13,135][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_1700_state/sampler_1.bin +[2026-06-17 19:50:13,137][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_1700_state/random_states_0.pkl +[2026-06-17 19:50:13,142][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1700 +[2026-06-17 19:50:13,143][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 19:50:13,196][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 19:50:23,188][starVLA.training.train_starvla][INFO] - Step 1701, Loss: {'train/loss': 0.2828415334224701, 'train/grad_norm_pre_clip': 5.396603584289551, 'timing/data': 0.011684817261993885, 'timing/model': 1.5685506025329232, 'learning_rate/qwen_vl_interface': 8.792932350554642e-06, 'learning_rate/action_model': 8.792932350554643e-05, 'epoch': 0.65}) +[2026-06-17 19:50:28,218][starVLA.training.train_starvla][INFO] - Step 1702, Loss: {'train/loss': 0.6394736766815186, 'train/grad_norm_pre_clip': 12.386876106262207, 'timing/data': 0.016502664424479008, 'timing/model': 1.5128854727372527, 'learning_rate/qwen_vl_interface': 8.791491458879089e-06, 'learning_rate/action_model': 8.791491458879089e-05, 'epoch': 0.65}) +[2026-06-17 19:50:32,712][starVLA.training.train_starvla][INFO] - Step 1703, Loss: {'train/loss': 0.6539420485496521, 'train/grad_norm_pre_clip': 7.704674243927002, 'timing/data': 0.01204504445195198, 'timing/model': 1.0959233278408647, 'learning_rate/qwen_vl_interface': 8.790049833048545e-06, 'learning_rate/action_model': 8.790049833048545e-05, 'epoch': 0.66}) +[2026-06-17 19:50:37,452][starVLA.training.train_starvla][INFO] - Step 1704, Loss: {'train/loss': 0.4240664839744568, 'train/grad_norm_pre_clip': 6.232337474822998, 'timing/data': 0.013504826463758945, 'timing/model': 1.1589889246970415, 'learning_rate/qwen_vl_interface': 8.788607473361862e-06, 'learning_rate/action_model': 8.788607473361862e-05, 'epoch': 0.66}) +[2026-06-17 19:50:42,191][starVLA.training.train_starvla][INFO] - Step 1705, Loss: {'train/loss': 0.8694555163383484, 'train/grad_norm_pre_clip': 13.047165870666504, 'timing/data': 0.015045478008687496, 'timing/model': 1.1547052087262273, 'learning_rate/qwen_vl_interface': 8.787164380118044e-06, 'learning_rate/action_model': 8.787164380118043e-05, 'epoch': 0.66}) +[2026-06-17 19:50:46,858][starVLA.training.train_starvla][INFO] - Step 1706, Loss: {'train/loss': 0.39347031712532043, 'train/grad_norm_pre_clip': 5.313139915466309, 'timing/data': 0.02861777786165476, 'timing/model': 1.135371790267527, 'learning_rate/qwen_vl_interface': 8.785720553616241e-06, 'learning_rate/action_model': 8.78572055361624e-05, 'epoch': 0.66}) +[2026-06-17 19:50:51,531][starVLA.training.train_starvla][INFO] - Step 1707, Loss: {'train/loss': 0.4105739891529083, 'train/grad_norm_pre_clip': 6.5584716796875, 'timing/data': 0.011808610521256924, 'timing/model': 1.5368237309157848, 'learning_rate/qwen_vl_interface': 8.784275994155762e-06, 'learning_rate/action_model': 8.784275994155762e-05, 'epoch': 0.66}) +[2026-06-17 19:50:58,198][starVLA.training.train_starvla][INFO] - Step 1708, Loss: {'train/loss': 0.6061173677444458, 'train/grad_norm_pre_clip': 12.013459205627441, 'timing/data': 0.01276194304227829, 'timing/model': 1.0790955359116197, 'learning_rate/qwen_vl_interface': 8.782830702036067e-06, 'learning_rate/action_model': 8.782830702036067e-05, 'epoch': 0.66}) +[2026-06-17 19:51:03,085][starVLA.training.train_starvla][INFO] - Step 1709, Loss: {'train/loss': 0.49508121609687805, 'train/grad_norm_pre_clip': 8.166858673095703, 'timing/data': 0.015990241430699825, 'timing/model': 1.1394019592553377, 'learning_rate/qwen_vl_interface': 8.781384677556767e-06, 'learning_rate/action_model': 8.781384677556767e-05, 'epoch': 0.66}) +[2026-06-17 19:51:07,881][starVLA.training.train_starvla][INFO] - Step 1710, Loss: {'train/loss': 0.46141886711120605, 'train/grad_norm_pre_clip': 8.527750015258789, 'timing/data': 0.012072096578776836, 'timing/model': 1.178107101470232, 'learning_rate/qwen_vl_interface': 8.779937921017624e-06, 'learning_rate/action_model': 8.779937921017623e-05, 'epoch': 0.66}) +[2026-06-17 19:51:12,997][starVLA.training.train_starvla][INFO] - Step 1711, Loss: {'train/loss': 0.5588166117668152, 'train/grad_norm_pre_clip': 12.054683685302734, 'timing/data': 0.012993735261261463, 'timing/model': 1.227865464054048, 'learning_rate/qwen_vl_interface': 8.77849043271855e-06, 'learning_rate/action_model': 8.778490432718551e-05, 'epoch': 0.66}) +[2026-06-17 19:51:18,964][starVLA.training.train_starvla][INFO] - Step 1712, Loss: {'train/loss': 0.5918099284172058, 'train/grad_norm_pre_clip': 11.726082801818848, 'timing/data': 0.013643506914377213, 'timing/model': 1.4297497887164354, 'learning_rate/qwen_vl_interface': 8.777042212959615e-06, 'learning_rate/action_model': 8.777042212959614e-05, 'epoch': 0.66}) +[2026-06-17 19:51:23,216][starVLA.training.train_starvla][INFO] - Step 1713, Loss: {'train/loss': 0.748201310634613, 'train/grad_norm_pre_clip': 6.75609016418457, 'timing/data': 0.015610004775226116, 'timing/model': 1.068212536163628, 'learning_rate/qwen_vl_interface': 8.775593262041032e-06, 'learning_rate/action_model': 8.77559326204103e-05, 'epoch': 0.66}) +[2026-06-17 19:51:27,982][starVLA.training.train_starvla][INFO] - Step 1714, Loss: {'train/loss': 0.5957993865013123, 'train/grad_norm_pre_clip': 9.000651359558105, 'timing/data': 0.012950621545314789, 'timing/model': 1.0857880655676126, 'learning_rate/qwen_vl_interface': 8.774143580263172e-06, 'learning_rate/action_model': 8.77414358026317e-05, 'epoch': 0.66}) +[2026-06-17 19:51:32,675][starVLA.training.train_starvla][INFO] - Step 1715, Loss: {'train/loss': 0.7217798829078674, 'train/grad_norm_pre_clip': 10.566701889038086, 'timing/data': 0.012469742447137833, 'timing/model': 1.0751543221995234, 'learning_rate/qwen_vl_interface': 8.772693167926554e-06, 'learning_rate/action_model': 8.772693167926555e-05, 'epoch': 0.66}) +[2026-06-17 19:51:38,546][starVLA.training.train_starvla][INFO] - Step 1716, Loss: {'train/loss': 0.6362073421478271, 'train/grad_norm_pre_clip': 13.689717292785645, 'timing/data': 0.012707212939858437, 'timing/model': 1.1171132950112224, 'learning_rate/qwen_vl_interface': 8.771242025331855e-06, 'learning_rate/action_model': 8.771242025331854e-05, 'epoch': 0.66}) +[2026-06-17 19:51:43,975][starVLA.training.train_starvla][INFO] - Step 1717, Loss: {'train/loss': 0.4147496223449707, 'train/grad_norm_pre_clip': 10.238020896911621, 'timing/data': 0.01978949923068285, 'timing/model': 1.1993864746764302, 'learning_rate/qwen_vl_interface': 8.769790152779893e-06, 'learning_rate/action_model': 8.769790152779892e-05, 'epoch': 0.66}) +[2026-06-17 19:51:48,665][starVLA.training.train_starvla][INFO] - Step 1718, Loss: {'train/loss': 0.5975855588912964, 'train/grad_norm_pre_clip': 8.819009780883789, 'timing/data': 0.012456543743610382, 'timing/model': 1.0866710795089602, 'learning_rate/qwen_vl_interface': 8.768337550571644e-06, 'learning_rate/action_model': 8.768337550571645e-05, 'epoch': 0.66}) +[2026-06-17 19:51:55,464][starVLA.training.train_starvla][INFO] - Step 1719, Loss: {'train/loss': 0.7903744578361511, 'train/grad_norm_pre_clip': 11.462841987609863, 'timing/data': 0.011690285056829453, 'timing/model': 1.074119378812611, 'learning_rate/qwen_vl_interface': 8.766884219008238e-06, 'learning_rate/action_model': 8.766884219008238e-05, 'epoch': 0.66}) +[2026-06-17 19:52:00,188][starVLA.training.train_starvla][INFO] - Step 1720, Loss: {'train/loss': 0.6972383260726929, 'train/grad_norm_pre_clip': 10.344285011291504, 'timing/data': 0.01707602757960558, 'timing/model': 1.1114487424492836, 'learning_rate/qwen_vl_interface': 8.765430158390948e-06, 'learning_rate/action_model': 8.765430158390947e-05, 'epoch': 0.66}) +[2026-06-17 19:52:04,951][starVLA.training.train_starvla][INFO] - Step 1721, Loss: {'train/loss': 0.4684077203273773, 'train/grad_norm_pre_clip': 6.948488712310791, 'timing/data': 0.012350596487522125, 'timing/model': 1.4947053827345371, 'learning_rate/qwen_vl_interface': 8.763975369021202e-06, 'learning_rate/action_model': 8.763975369021203e-05, 'epoch': 0.66}) +[2026-06-17 19:52:09,886][starVLA.training.train_starvla][INFO] - Step 1722, Loss: {'train/loss': 0.40028131008148193, 'train/grad_norm_pre_clip': 6.59293270111084, 'timing/data': 0.01761933695524931, 'timing/model': 1.4613683046773076, 'learning_rate/qwen_vl_interface': 8.762519851200581e-06, 'learning_rate/action_model': 8.762519851200581e-05, 'epoch': 0.66}) +[2026-06-17 19:52:14,588][starVLA.training.train_starvla][INFO] - Step 1723, Loss: {'train/loss': 0.2883405089378357, 'train/grad_norm_pre_clip': 5.789316177368164, 'timing/data': 0.01415279507637024, 'timing/model': 1.2624787269160151, 'learning_rate/qwen_vl_interface': 8.761063605230818e-06, 'learning_rate/action_model': 8.761063605230818e-05, 'epoch': 0.66}) +[2026-06-17 19:52:19,163][starVLA.training.train_starvla][INFO] - Step 1724, Loss: {'train/loss': 0.5409997701644897, 'train/grad_norm_pre_clip': 9.141944885253906, 'timing/data': 0.010689359158277512, 'timing/model': 1.0930550312623382, 'learning_rate/qwen_vl_interface': 8.759606631413788e-06, 'learning_rate/action_model': 8.759606631413788e-05, 'epoch': 0.66}) +[2026-06-17 19:52:24,172][starVLA.training.train_starvla][INFO] - Step 1725, Loss: {'train/loss': 0.5260115265846252, 'train/grad_norm_pre_clip': 9.660216331481934, 'timing/data': 0.011488813906908035, 'timing/model': 1.1673953430727124, 'learning_rate/qwen_vl_interface': 8.75814893005153e-06, 'learning_rate/action_model': 8.75814893005153e-05, 'epoch': 0.66}) +[2026-06-17 19:52:29,027][starVLA.training.train_starvla][INFO] - Step 1726, Loss: {'train/loss': 0.8406148552894592, 'train/grad_norm_pre_clip': 10.319393157958984, 'timing/data': 0.01453014463186264, 'timing/model': 1.0883909333497286, 'learning_rate/qwen_vl_interface': 8.756690501446222e-06, 'learning_rate/action_model': 8.756690501446221e-05, 'epoch': 0.66}) +[2026-06-17 19:52:33,904][starVLA.training.train_starvla][INFO] - Step 1727, Loss: {'train/loss': 0.8476836681365967, 'train/grad_norm_pre_clip': 15.339479446411133, 'timing/data': 0.013545641675591469, 'timing/model': 1.4622969534248114, 'learning_rate/qwen_vl_interface': 8.755231345900199e-06, 'learning_rate/action_model': 8.755231345900198e-05, 'epoch': 0.66}) +[2026-06-17 19:52:38,445][starVLA.training.train_starvla][INFO] - Step 1728, Loss: {'train/loss': 0.5092191100120544, 'train/grad_norm_pre_clip': 8.610870361328125, 'timing/data': 0.011071741580963135, 'timing/model': 1.0781954899430275, 'learning_rate/qwen_vl_interface': 8.753771463715947e-06, 'learning_rate/action_model': 8.753771463715947e-05, 'epoch': 0.66}) +[2026-06-17 19:52:43,277][starVLA.training.train_starvla][INFO] - Step 1729, Loss: {'train/loss': 0.8059954047203064, 'train/grad_norm_pre_clip': 7.5675740242004395, 'timing/data': 0.01147348154336214, 'timing/model': 1.1541202263906598, 'learning_rate/qwen_vl_interface': 8.7523108551961e-06, 'learning_rate/action_model': 8.7523108551961e-05, 'epoch': 0.67}) +[2026-06-17 19:52:48,185][starVLA.training.train_starvla][INFO] - Step 1730, Loss: {'train/loss': 0.5013414025306702, 'train/grad_norm_pre_clip': 8.337953567504883, 'timing/data': 0.0114529337733984, 'timing/model': 1.0735825393348932, 'learning_rate/qwen_vl_interface': 8.750849520643445e-06, 'learning_rate/action_model': 8.750849520643444e-05, 'epoch': 0.67}) +[2026-06-17 19:52:52,991][starVLA.training.train_starvla][INFO] - Step 1731, Loss: {'train/loss': 0.37441086769104004, 'train/grad_norm_pre_clip': 5.593825340270996, 'timing/data': 0.0164578128606081, 'timing/model': 1.096249450929463, 'learning_rate/qwen_vl_interface': 8.749387460360917e-06, 'learning_rate/action_model': 8.749387460360916e-05, 'epoch': 0.67}) +[2026-06-17 19:52:58,990][starVLA.training.train_starvla][INFO] - Step 1732, Loss: {'train/loss': 0.6765965223312378, 'train/grad_norm_pre_clip': 20.80316162109375, 'timing/data': 0.012226046994328499, 'timing/model': 1.561483707278967, 'learning_rate/qwen_vl_interface': 8.747924674651602e-06, 'learning_rate/action_model': 8.747924674651602e-05, 'epoch': 0.67}) +[2026-06-17 19:53:03,434][starVLA.training.train_starvla][INFO] - Step 1733, Loss: {'train/loss': 0.34738433361053467, 'train/grad_norm_pre_clip': 7.896588325500488, 'timing/data': 0.019105462357401848, 'timing/model': 1.0863142935559154, 'learning_rate/qwen_vl_interface': 8.74646116381874e-06, 'learning_rate/action_model': 8.74646116381874e-05, 'epoch': 0.67}) +[2026-06-17 19:53:08,289][starVLA.training.train_starvla][INFO] - Step 1734, Loss: {'train/loss': 0.46638840436935425, 'train/grad_norm_pre_clip': 6.8443684577941895, 'timing/data': 0.011638615280389786, 'timing/model': 1.2784401681274176, 'learning_rate/qwen_vl_interface': 8.744996928165714e-06, 'learning_rate/action_model': 8.744996928165714e-05, 'epoch': 0.67}) +[2026-06-17 19:53:13,456][starVLA.training.train_starvla][INFO] - Step 1735, Loss: {'train/loss': 0.6113168597221375, 'train/grad_norm_pre_clip': 9.148082733154297, 'timing/data': 0.011989693157374859, 'timing/model': 1.1083999574184418, 'learning_rate/qwen_vl_interface': 8.743531967996066e-06, 'learning_rate/action_model': 8.743531967996065e-05, 'epoch': 0.67}) +[2026-06-17 19:53:21,037][starVLA.training.train_starvla][INFO] - Step 1736, Loss: {'train/loss': 0.40747734904289246, 'train/grad_norm_pre_clip': 5.577141284942627, 'timing/data': 0.013108393177390099, 'timing/model': 3.9603566164150834, 'learning_rate/qwen_vl_interface': 8.742066283613481e-06, 'learning_rate/action_model': 8.74206628361348e-05, 'epoch': 0.67}) +[2026-06-17 19:53:26,076][starVLA.training.train_starvla][INFO] - Step 1737, Loss: {'train/loss': 0.7543286681175232, 'train/grad_norm_pre_clip': 8.848713874816895, 'timing/data': 0.011419301852583885, 'timing/model': 1.4625640017911792, 'learning_rate/qwen_vl_interface': 8.740599875321797e-06, 'learning_rate/action_model': 8.740599875321797e-05, 'epoch': 0.67}) +[2026-06-17 19:53:30,657][starVLA.training.train_starvla][INFO] - Step 1738, Loss: {'train/loss': 0.3374316096305847, 'train/grad_norm_pre_clip': 5.121859073638916, 'timing/data': 0.01203862950205803, 'timing/model': 1.0846175393089652, 'learning_rate/qwen_vl_interface': 8.739132743425004e-06, 'learning_rate/action_model': 8.739132743425002e-05, 'epoch': 0.67}) +[2026-06-17 19:53:35,386][starVLA.training.train_starvla][INFO] - Step 1739, Loss: {'train/loss': 0.5236712694168091, 'train/grad_norm_pre_clip': 6.297178268432617, 'timing/data': 0.012234226800501347, 'timing/model': 1.0930501241236925, 'learning_rate/qwen_vl_interface': 8.737664888227239e-06, 'learning_rate/action_model': 8.737664888227239e-05, 'epoch': 0.67}) +[2026-06-17 19:53:40,486][starVLA.training.train_starvla][INFO] - Step 1740, Loss: {'train/loss': 0.46316465735435486, 'train/grad_norm_pre_clip': 6.414477825164795, 'timing/data': 0.012180457822978497, 'timing/model': 1.2959461836144328, 'learning_rate/qwen_vl_interface': 8.736196310032787e-06, 'learning_rate/action_model': 8.736196310032787e-05, 'epoch': 0.67}) +[2026-06-17 19:53:45,104][starVLA.training.train_starvla][INFO] - Step 1741, Loss: {'train/loss': 0.5934870839118958, 'train/grad_norm_pre_clip': 7.354665756225586, 'timing/data': 0.013091601431369781, 'timing/model': 1.0693359402939677, 'learning_rate/qwen_vl_interface': 8.734727009146092e-06, 'learning_rate/action_model': 8.734727009146092e-05, 'epoch': 0.67}) +[2026-06-17 19:53:49,903][starVLA.training.train_starvla][INFO] - Step 1742, Loss: {'train/loss': 0.42737674713134766, 'train/grad_norm_pre_clip': 8.692082405090332, 'timing/data': 0.013722443953156471, 'timing/model': 1.5753812361508608, 'learning_rate/qwen_vl_interface': 8.733256985871737e-06, 'learning_rate/action_model': 8.733256985871737e-05, 'epoch': 0.67}) +[2026-06-17 19:53:54,266][starVLA.training.train_starvla][INFO] - Step 1743, Loss: {'train/loss': 0.37797269225120544, 'train/grad_norm_pre_clip': 6.524034023284912, 'timing/data': 0.0132722994312644, 'timing/model': 1.1646690592169762, 'learning_rate/qwen_vl_interface': 8.73178624051446e-06, 'learning_rate/action_model': 8.73178624051446e-05, 'epoch': 0.67}) +[2026-06-17 19:54:00,326][starVLA.training.train_starvla][INFO] - Step 1744, Loss: {'train/loss': 0.6956444382667542, 'train/grad_norm_pre_clip': 8.90422534942627, 'timing/data': 0.01990850828588009, 'timing/model': 2.556139573454857, 'learning_rate/qwen_vl_interface': 8.73031477337915e-06, 'learning_rate/action_model': 8.73031477337915e-05, 'epoch': 0.67}) +[2026-06-17 19:54:05,354][starVLA.training.train_starvla][INFO] - Step 1745, Loss: {'train/loss': 0.2508779764175415, 'train/grad_norm_pre_clip': 7.150975227355957, 'timing/data': 0.013326234184205532, 'timing/model': 1.1255680890753865, 'learning_rate/qwen_vl_interface': 8.728842584770842e-06, 'learning_rate/action_model': 8.728842584770842e-05, 'epoch': 0.67}) +[2026-06-17 19:54:10,255][starVLA.training.train_starvla][INFO] - Step 1746, Loss: {'train/loss': 0.6754621863365173, 'train/grad_norm_pre_clip': 10.809806823730469, 'timing/data': 0.01353437826037407, 'timing/model': 1.187676296569407, 'learning_rate/qwen_vl_interface': 8.727369674994722e-06, 'learning_rate/action_model': 8.727369674994722e-05, 'epoch': 0.67}) +[2026-06-17 19:54:15,077][starVLA.training.train_starvla][INFO] - Step 1747, Loss: {'train/loss': 0.27927958965301514, 'train/grad_norm_pre_clip': 5.213651657104492, 'timing/data': 0.01139078103005886, 'timing/model': 1.3678314089775085, 'learning_rate/qwen_vl_interface': 8.725896044356128e-06, 'learning_rate/action_model': 8.725896044356128e-05, 'epoch': 0.67}) +[2026-06-17 19:54:20,257][starVLA.training.train_starvla][INFO] - Step 1748, Loss: {'train/loss': 0.770521879196167, 'train/grad_norm_pre_clip': 10.886402130126953, 'timing/data': 0.011675387620925903, 'timing/model': 1.8366733323782682, 'learning_rate/qwen_vl_interface': 8.724421693160544e-06, 'learning_rate/action_model': 8.724421693160543e-05, 'epoch': 0.67}) +[2026-06-17 19:54:26,923][starVLA.training.train_starvla][INFO] - Step 1749, Loss: {'train/loss': 0.48439595103263855, 'train/grad_norm_pre_clip': 11.791176795959473, 'timing/data': 0.013931182213127613, 'timing/model': 1.1186790820211172, 'learning_rate/qwen_vl_interface': 8.722946621713603e-06, 'learning_rate/action_model': 8.722946621713602e-05, 'epoch': 0.67}) +[2026-06-17 19:54:33,209][starVLA.training.train_starvla][INFO] - Step 1750, Loss: {'train/loss': 0.5410114526748657, 'train/grad_norm_pre_clip': 8.514129638671875, 'timing/data': 0.011612391099333763, 'timing/model': 1.0836715633049607, 'learning_rate/qwen_vl_interface': 8.721470830321091e-06, 'learning_rate/action_model': 8.72147083032109e-05, 'epoch': 0.67}) +[2026-06-17 19:54:38,144][starVLA.training.train_starvla][INFO] - Step 1751, Loss: {'train/loss': 0.8435848951339722, 'train/grad_norm_pre_clip': 9.867924690246582, 'timing/data': 0.013510314747691154, 'timing/model': 1.3335596416145563, 'learning_rate/qwen_vl_interface': 8.719994319288939e-06, 'learning_rate/action_model': 8.71999431928894e-05, 'epoch': 0.67}) +[2026-06-17 19:54:42,785][starVLA.training.train_starvla][INFO] - Step 1752, Loss: {'train/loss': 0.5500805974006653, 'train/grad_norm_pre_clip': 8.02246379852295, 'timing/data': 0.011388975195586681, 'timing/model': 1.4737232374027371, 'learning_rate/qwen_vl_interface': 8.718517088923233e-06, 'learning_rate/action_model': 8.718517088923233e-05, 'epoch': 0.67}) +[2026-06-17 19:54:47,206][starVLA.training.train_starvla][INFO] - Step 1753, Loss: {'train/loss': 0.4353758990764618, 'train/grad_norm_pre_clip': 7.849350929260254, 'timing/data': 0.011040879413485527, 'timing/model': 1.0898709846660495, 'learning_rate/qwen_vl_interface': 8.7170391395302e-06, 'learning_rate/action_model': 8.7170391395302e-05, 'epoch': 0.67}) +[2026-06-17 19:54:52,156][starVLA.training.train_starvla][INFO] - Step 1754, Loss: {'train/loss': 0.4121304750442505, 'train/grad_norm_pre_clip': 7.7716474533081055, 'timing/data': 0.013189438730478287, 'timing/model': 1.1572252782061696, 'learning_rate/qwen_vl_interface': 8.715560471416223e-06, 'learning_rate/action_model': 8.715560471416224e-05, 'epoch': 0.67}) +[2026-06-17 19:55:01,988][starVLA.training.train_starvla][INFO] - Step 1755, Loss: {'train/loss': 0.8417397141456604, 'train/grad_norm_pre_clip': 11.561129570007324, 'timing/data': 0.01085241325199604, 'timing/model': 1.1809995118528605, 'learning_rate/qwen_vl_interface': 8.714081084887833e-06, 'learning_rate/action_model': 8.714081084887833e-05, 'epoch': 0.68}) +[2026-06-17 19:55:06,939][starVLA.training.train_starvla][INFO] - Step 1756, Loss: {'train/loss': 0.44675540924072266, 'train/grad_norm_pre_clip': 8.749503135681152, 'timing/data': 0.014552506618201733, 'timing/model': 1.101206528954208, 'learning_rate/qwen_vl_interface': 8.712600980251706e-06, 'learning_rate/action_model': 8.712600980251705e-05, 'epoch': 0.68}) +[2026-06-17 19:55:12,181][starVLA.training.train_starvla][INFO] - Step 1757, Loss: {'train/loss': 0.7216551303863525, 'train/grad_norm_pre_clip': 10.826543807983398, 'timing/data': 0.013515924103558064, 'timing/model': 1.6632344257086515, 'learning_rate/qwen_vl_interface': 8.71112015781467e-06, 'learning_rate/action_model': 8.71112015781467e-05, 'epoch': 0.68}) +[2026-06-17 19:55:16,545][starVLA.training.train_starvla][INFO] - Step 1758, Loss: {'train/loss': 0.5180928707122803, 'train/grad_norm_pre_clip': 9.930112838745117, 'timing/data': 0.014396139420568943, 'timing/model': 1.1127555267885327, 'learning_rate/qwen_vl_interface': 8.709638617883699e-06, 'learning_rate/action_model': 8.709638617883699e-05, 'epoch': 0.68}) +[2026-06-17 19:55:23,555][starVLA.training.train_starvla][INFO] - Step 1759, Loss: {'train/loss': 0.4399573504924774, 'train/grad_norm_pre_clip': 6.443676471710205, 'timing/data': 0.011615623719990253, 'timing/model': 3.285537547431886, 'learning_rate/qwen_vl_interface': 8.708156360765922e-06, 'learning_rate/action_model': 8.708156360765922e-05, 'epoch': 0.68}) +[2026-06-17 19:55:28,633][starVLA.training.train_starvla][INFO] - Step 1760, Loss: {'train/loss': 0.5777763724327087, 'train/grad_norm_pre_clip': 45.00450134277344, 'timing/data': 0.01153576746582985, 'timing/model': 1.1293377820402384, 'learning_rate/qwen_vl_interface': 8.706673386768609e-06, 'learning_rate/action_model': 8.706673386768609e-05, 'epoch': 0.68}) +[2026-06-17 19:55:33,236][starVLA.training.train_starvla][INFO] - Step 1761, Loss: {'train/loss': 0.5309082865715027, 'train/grad_norm_pre_clip': 15.257511138916016, 'timing/data': 0.013085558079183102, 'timing/model': 1.1143297823145986, 'learning_rate/qwen_vl_interface': 8.705189696199184e-06, 'learning_rate/action_model': 8.705189696199183e-05, 'epoch': 0.68}) +[2026-06-17 19:55:37,966][starVLA.training.train_starvla][INFO] - Step 1762, Loss: {'train/loss': 0.450672447681427, 'train/grad_norm_pre_clip': 7.757657051086426, 'timing/data': 0.012083233334124088, 'timing/model': 1.4506474062800407, 'learning_rate/qwen_vl_interface': 8.703705289365216e-06, 'learning_rate/action_model': 8.703705289365216e-05, 'epoch': 0.68}) +[2026-06-17 19:55:42,541][starVLA.training.train_starvla][INFO] - Step 1763, Loss: {'train/loss': 0.6550630331039429, 'train/grad_norm_pre_clip': 8.875144004821777, 'timing/data': 0.015774276107549667, 'timing/model': 1.1322683226317167, 'learning_rate/qwen_vl_interface': 8.702220166574426e-06, 'learning_rate/action_model': 8.702220166574425e-05, 'epoch': 0.68}) +[2026-06-17 19:55:48,177][starVLA.training.train_starvla][INFO] - Step 1764, Loss: {'train/loss': 0.4438166618347168, 'train/grad_norm_pre_clip': 7.052288055419922, 'timing/data': 0.012791736051440239, 'timing/model': 1.687873755581677, 'learning_rate/qwen_vl_interface': 8.700734328134681e-06, 'learning_rate/action_model': 8.70073432813468e-05, 'epoch': 0.68}) +[2026-06-17 19:55:55,329][starVLA.training.train_starvla][INFO] - Step 1765, Loss: {'train/loss': 0.5047101378440857, 'train/grad_norm_pre_clip': 7.294745445251465, 'timing/data': 0.01120719313621521, 'timing/model': 1.066600394435227, 'learning_rate/qwen_vl_interface': 8.699247774353997e-06, 'learning_rate/action_model': 8.699247774353996e-05, 'epoch': 0.68}) +[2026-06-17 19:56:00,368][starVLA.training.train_starvla][INFO] - Step 1766, Loss: {'train/loss': 0.40301695466041565, 'train/grad_norm_pre_clip': 5.644693851470947, 'timing/data': 0.011607527732849121, 'timing/model': 1.1026927093043923, 'learning_rate/qwen_vl_interface': 8.697760505540536e-06, 'learning_rate/action_model': 8.697760505540536e-05, 'epoch': 0.68}) +[2026-06-17 19:56:05,583][starVLA.training.train_starvla][INFO] - Step 1767, Loss: {'train/loss': 0.450840562582016, 'train/grad_norm_pre_clip': 9.055397033691406, 'timing/data': 0.020258117467164993, 'timing/model': 1.7238282933831215, 'learning_rate/qwen_vl_interface': 8.696272522002614e-06, 'learning_rate/action_model': 8.696272522002615e-05, 'epoch': 0.68}) +[2026-06-17 19:56:10,041][starVLA.training.train_starvla][INFO] - Step 1768, Loss: {'train/loss': 0.5484852194786072, 'train/grad_norm_pre_clip': 11.333447456359863, 'timing/data': 0.012813818641006947, 'timing/model': 1.0878123585134745, 'learning_rate/qwen_vl_interface': 8.69478382404869e-06, 'learning_rate/action_model': 8.69478382404869e-05, 'epoch': 0.68}) +[2026-06-17 19:56:18,892][starVLA.training.train_starvla][INFO] - Step 1769, Loss: {'train/loss': 0.6577850580215454, 'train/grad_norm_pre_clip': 8.551412582397461, 'timing/data': 0.014104509726166725, 'timing/model': 1.21283068228513, 'learning_rate/qwen_vl_interface': 8.693294411987371e-06, 'learning_rate/action_model': 8.693294411987371e-05, 'epoch': 0.68}) +[2026-06-17 19:56:24,484][starVLA.training.train_starvla][INFO] - Step 1770, Loss: {'train/loss': 0.5599659085273743, 'train/grad_norm_pre_clip': 9.333075523376465, 'timing/data': 0.015809569507837296, 'timing/model': 1.3283948777243495, 'learning_rate/qwen_vl_interface': 8.691804286127416e-06, 'learning_rate/action_model': 8.691804286127415e-05, 'epoch': 0.68}) +[2026-06-17 19:56:29,146][starVLA.training.train_starvla][INFO] - Step 1771, Loss: {'train/loss': 0.49819305539131165, 'train/grad_norm_pre_clip': 7.901528835296631, 'timing/data': 0.011547969654202461, 'timing/model': 1.0769155267626047, 'learning_rate/qwen_vl_interface': 8.690313446777729e-06, 'learning_rate/action_model': 8.690313446777728e-05, 'epoch': 0.68}) +[2026-06-17 19:56:34,032][starVLA.training.train_starvla][INFO] - Step 1772, Loss: {'train/loss': 0.6191400289535522, 'train/grad_norm_pre_clip': 9.24026870727539, 'timing/data': 0.016503330320119858, 'timing/model': 1.6021826714277267, 'learning_rate/qwen_vl_interface': 8.688821894247363e-06, 'learning_rate/action_model': 8.688821894247362e-05, 'epoch': 0.68}) +[2026-06-17 19:56:40,239][starVLA.training.train_starvla][INFO] - Step 1773, Loss: {'train/loss': 0.23184822499752045, 'train/grad_norm_pre_clip': 5.843424320220947, 'timing/data': 0.013517023995518684, 'timing/model': 1.2638895530253649, 'learning_rate/qwen_vl_interface': 8.687329628845517e-06, 'learning_rate/action_model': 8.687329628845516e-05, 'epoch': 0.68}) +[2026-06-17 19:56:44,897][starVLA.training.train_starvla][INFO] - Step 1774, Loss: {'train/loss': 0.29140838980674744, 'train/grad_norm_pre_clip': 5.154912948608398, 'timing/data': 0.011487504467368126, 'timing/model': 1.1035549882799387, 'learning_rate/qwen_vl_interface': 8.68583665088154e-06, 'learning_rate/action_model': 8.68583665088154e-05, 'epoch': 0.68}) +[2026-06-17 19:56:51,951][starVLA.training.train_starvla][INFO] - Step 1775, Loss: {'train/loss': 0.33614665269851685, 'train/grad_norm_pre_clip': 7.859996318817139, 'timing/data': 0.014516064897179604, 'timing/model': 1.097484029829502, 'learning_rate/qwen_vl_interface': 8.684342960664928e-06, 'learning_rate/action_model': 8.684342960664928e-05, 'epoch': 0.68}) +[2026-06-17 19:56:56,833][starVLA.training.train_starvla][INFO] - Step 1776, Loss: {'train/loss': 0.7551742196083069, 'train/grad_norm_pre_clip': 19.482961654663086, 'timing/data': 0.013545343652367592, 'timing/model': 1.0996400974690914, 'learning_rate/qwen_vl_interface': 8.682848558505327e-06, 'learning_rate/action_model': 8.682848558505326e-05, 'epoch': 0.68}) +[2026-06-17 19:57:02,137][starVLA.training.train_starvla][INFO] - Step 1777, Loss: {'train/loss': 0.5354229211807251, 'train/grad_norm_pre_clip': 9.1978120803833, 'timing/data': 0.01224017608910799, 'timing/model': 1.7064177906140685, 'learning_rate/qwen_vl_interface': 8.681353444712525e-06, 'learning_rate/action_model': 8.681353444712524e-05, 'epoch': 0.68}) +[2026-06-17 19:57:12,845][starVLA.training.train_starvla][INFO] - Step 1778, Loss: {'train/loss': 0.7933303713798523, 'train/grad_norm_pre_clip': 11.661686897277832, 'timing/data': 0.04197759833186865, 'timing/model': 1.1849123928695917, 'learning_rate/qwen_vl_interface': 8.67985761959646e-06, 'learning_rate/action_model': 8.67985761959646e-05, 'epoch': 0.68}) +[2026-06-17 19:57:17,942][starVLA.training.train_starvla][INFO] - Step 1779, Loss: {'train/loss': 0.7814410924911499, 'train/grad_norm_pre_clip': 10.126903533935547, 'timing/data': 0.013037468306720257, 'timing/model': 1.0843138471245766, 'learning_rate/qwen_vl_interface': 8.678361083467221e-06, 'learning_rate/action_model': 8.67836108346722e-05, 'epoch': 0.68}) +[2026-06-17 19:57:27,260][starVLA.training.train_starvla][INFO] - Step 1780, Loss: {'train/loss': 0.3541010022163391, 'train/grad_norm_pre_clip': 6.020819187164307, 'timing/data': 0.014041521586477757, 'timing/model': 3.472681488841772, 'learning_rate/qwen_vl_interface': 8.67686383663504e-06, 'learning_rate/action_model': 8.67686383663504e-05, 'epoch': 0.68}) +[2026-06-17 19:57:32,287][starVLA.training.train_starvla][INFO] - Step 1781, Loss: {'train/loss': 0.3354089558124542, 'train/grad_norm_pre_clip': 10.236499786376953, 'timing/data': 0.014097368344664574, 'timing/model': 1.1328126173466444, 'learning_rate/qwen_vl_interface': 8.6753658794103e-06, 'learning_rate/action_model': 8.675365879410298e-05, 'epoch': 0.69}) +[2026-06-17 19:57:37,114][starVLA.training.train_starvla][INFO] - Step 1782, Loss: {'train/loss': 0.6976850032806396, 'train/grad_norm_pre_clip': 8.577275276184082, 'timing/data': 0.014831120148301125, 'timing/model': 1.5506705790758133, 'learning_rate/qwen_vl_interface': 8.673867212103525e-06, 'learning_rate/action_model': 8.673867212103524e-05, 'epoch': 0.69}) +[2026-06-17 19:57:45,333][starVLA.training.train_starvla][INFO] - Step 1783, Loss: {'train/loss': 0.4721752405166626, 'train/grad_norm_pre_clip': 8.37237548828125, 'timing/data': 0.012655290775001049, 'timing/model': 1.1282513849437237, 'learning_rate/qwen_vl_interface': 8.672367835025395e-06, 'learning_rate/action_model': 8.672367835025393e-05, 'epoch': 0.69}) +[2026-06-17 19:57:50,437][starVLA.training.train_starvla][INFO] - Step 1784, Loss: {'train/loss': 0.6888788938522339, 'train/grad_norm_pre_clip': 8.135621070861816, 'timing/data': 0.013342095538973808, 'timing/model': 1.2274534218013287, 'learning_rate/qwen_vl_interface': 8.670867748486727e-06, 'learning_rate/action_model': 8.670867748486726e-05, 'epoch': 0.69}) +[2026-06-17 19:58:01,831][starVLA.training.train_starvla][INFO] - Step 1785, Loss: {'train/loss': 0.46944159269332886, 'train/grad_norm_pre_clip': 8.057304382324219, 'timing/data': 0.010722932405769825, 'timing/model': 4.167646931484342, 'learning_rate/qwen_vl_interface': 8.669366952798494e-06, 'learning_rate/action_model': 8.669366952798494e-05, 'epoch': 0.69}) +[2026-06-17 19:58:07,261][starVLA.training.train_starvla][INFO] - Step 1786, Loss: {'train/loss': 0.46521326899528503, 'train/grad_norm_pre_clip': 6.756068706512451, 'timing/data': 0.016028999350965023, 'timing/model': 1.129797900095582, 'learning_rate/qwen_vl_interface': 8.667865448271814e-06, 'learning_rate/action_model': 8.667865448271814e-05, 'epoch': 0.69}) +[2026-06-17 19:58:14,023][starVLA.training.train_starvla][INFO] - Step 1787, Loss: {'train/loss': 0.4517996311187744, 'train/grad_norm_pre_clip': 6.242893695831299, 'timing/data': 0.01421732921153307, 'timing/model': 3.3233868293464184, 'learning_rate/qwen_vl_interface': 8.666363235217949e-06, 'learning_rate/action_model': 8.66636323521795e-05, 'epoch': 0.69}) +[2026-06-17 19:58:21,429][starVLA.training.train_starvla][INFO] - Step 1788, Loss: {'train/loss': 0.619896411895752, 'train/grad_norm_pre_clip': 9.455105781555176, 'timing/data': 0.01258294377475977, 'timing/model': 1.112433728761971, 'learning_rate/qwen_vl_interface': 8.66486031394831e-06, 'learning_rate/action_model': 8.664860313948309e-05, 'epoch': 0.69}) +[2026-06-17 19:58:26,185][starVLA.training.train_starvla][INFO] - Step 1789, Loss: {'train/loss': 0.39932841062545776, 'train/grad_norm_pre_clip': 5.361551761627197, 'timing/data': 0.015386367216706276, 'timing/model': 1.201588187366724, 'learning_rate/qwen_vl_interface': 8.66335668477445e-06, 'learning_rate/action_model': 8.66335668477445e-05, 'epoch': 0.69}) +[2026-06-17 19:58:31,116][starVLA.training.train_starvla][INFO] - Step 1790, Loss: {'train/loss': 0.5708800554275513, 'train/grad_norm_pre_clip': 9.178869247436523, 'timing/data': 0.034041169099509716, 'timing/model': 1.0995598994195461, 'learning_rate/qwen_vl_interface': 8.661852348008079e-06, 'learning_rate/action_model': 8.661852348008078e-05, 'epoch': 0.69}) +[2026-06-17 19:58:36,004][starVLA.training.train_starvla][INFO] - Step 1791, Loss: {'train/loss': 0.47385773062705994, 'train/grad_norm_pre_clip': 6.325483322143555, 'timing/data': 0.012540105730295181, 'timing/model': 1.1574056232348084, 'learning_rate/qwen_vl_interface': 8.660347303961044e-06, 'learning_rate/action_model': 8.660347303961044e-05, 'epoch': 0.69}) +[2026-06-17 19:58:40,918][starVLA.training.train_starvla][INFO] - Step 1792, Loss: {'train/loss': 0.7333717942237854, 'train/grad_norm_pre_clip': 7.692622661590576, 'timing/data': 0.012086069211363792, 'timing/model': 1.4847201397642493, 'learning_rate/qwen_vl_interface': 8.658841552945342e-06, 'learning_rate/action_model': 8.658841552945342e-05, 'epoch': 0.69}) +[2026-06-17 19:58:45,443][starVLA.training.train_starvla][INFO] - Step 1793, Loss: {'train/loss': 0.5229372382164001, 'train/grad_norm_pre_clip': 9.261736869812012, 'timing/data': 0.01626559253782034, 'timing/model': 1.2033403581008315, 'learning_rate/qwen_vl_interface': 8.65733509527312e-06, 'learning_rate/action_model': 8.65733509527312e-05, 'epoch': 0.69}) +[2026-06-17 19:58:50,404][starVLA.training.train_starvla][INFO] - Step 1794, Loss: {'train/loss': 0.36575907468795776, 'train/grad_norm_pre_clip': 6.668811798095703, 'timing/data': 0.013892770744860172, 'timing/model': 1.1879852646961808, 'learning_rate/qwen_vl_interface': 8.655827931256667e-06, 'learning_rate/action_model': 8.655827931256667e-05, 'epoch': 0.69}) +[2026-06-17 19:58:55,165][starVLA.training.train_starvla][INFO] - Step 1795, Loss: {'train/loss': 0.5457926988601685, 'train/grad_norm_pre_clip': 9.06233024597168, 'timing/data': 0.012094410136342049, 'timing/model': 1.1183619881048799, 'learning_rate/qwen_vl_interface': 8.654320061208418e-06, 'learning_rate/action_model': 8.654320061208417e-05, 'epoch': 0.69}) +[2026-06-17 19:59:00,562][starVLA.training.train_starvla][INFO] - Step 1796, Loss: {'train/loss': 0.5827216506004333, 'train/grad_norm_pre_clip': 8.384135246276855, 'timing/data': 0.016712467186152935, 'timing/model': 1.293259366415441, 'learning_rate/qwen_vl_interface': 8.652811485440956e-06, 'learning_rate/action_model': 8.652811485440956e-05, 'epoch': 0.69}) +[2026-06-17 19:59:05,614][starVLA.training.train_starvla][INFO] - Step 1797, Loss: {'train/loss': 0.37401723861694336, 'train/grad_norm_pre_clip': 8.564384460449219, 'timing/data': 0.013837792910635471, 'timing/model': 1.5436090528964996, 'learning_rate/qwen_vl_interface': 8.651302204267013e-06, 'learning_rate/action_model': 8.651302204267013e-05, 'epoch': 0.69}) +[2026-06-17 19:59:09,969][starVLA.training.train_starvla][INFO] - Step 1798, Loss: {'train/loss': 0.47455301880836487, 'train/grad_norm_pre_clip': 9.851119995117188, 'timing/data': 0.015216301195323467, 'timing/model': 1.202015195041895, 'learning_rate/qwen_vl_interface': 8.649792217999461e-06, 'learning_rate/action_model': 8.649792217999462e-05, 'epoch': 0.69}) +[2026-06-17 19:59:14,791][starVLA.training.train_starvla][INFO] - Step 1799, Loss: {'train/loss': 0.4845948815345764, 'train/grad_norm_pre_clip': 7.964655876159668, 'timing/data': 0.013125767931342125, 'timing/model': 1.1195920528844, 'learning_rate/qwen_vl_interface': 8.648281526951327e-06, 'learning_rate/action_model': 8.648281526951327e-05, 'epoch': 0.69}) +[2026-06-17 20:02:46,864][starVLA.training.train_starvla][INFO] - Step 1800, Loss: {'train/loss': 0.8113278746604919, 'train/grad_norm_pre_clip': 22.997976303100586, 'eval/loss': 0.5524510741233826, 'eval/latency_4/loss': 0.5524510741233826, 'eval/demon_fire_precision': 0.954244241085516, 'eval/demon_fire_recall': 0.9821370574861968, 'eval/demon_fire_f1': 0.9679897567221512, 'eval/demon_fire_rate_model': 0.9306901615271659, 'eval/demon_fire_rate_teacher': 0.9042584434654919, 'eval/demon_left_precision': 0.713903743315508, 'eval/demon_left_recall': 0.6095890410958904, 'eval/demon_left_f1': 0.6576354679802955, 'eval/demon_left_rate_model': 0.10983847283406754, 'eval/demon_left_rate_teacher': 0.12863436123348018, 'eval/demon_right_precision': 0.769857433808554, 'eval/demon_right_recall': 0.6823104693140795, 'eval/demon_right_f1': 0.7234449760765551, 'eval/demon_right_rate_model': 0.14419970631424375, 'eval/demon_right_rate_teacher': 0.16270190895741557, 'eval/demon_move_f1': 0.6905402220284254, 'eval/demon_cc_f1': 0.8638632552282707, 'eval/latency_4/demon_fire_precision': 0.954244241085516, 'eval/latency_4/demon_fire_recall': 0.9821370574861968, 'eval/latency_4/demon_fire_f1': 0.9679897567221512, 'eval/latency_4/demon_fire_rate_model': 0.9306901615271659, 'eval/latency_4/demon_fire_rate_teacher': 0.9042584434654919, 'eval/latency_4/demon_left_precision': 0.713903743315508, 'eval/latency_4/demon_left_recall': 0.6095890410958904, 'eval/latency_4/demon_left_f1': 0.6576354679802955, 'eval/latency_4/demon_left_rate_model': 0.10983847283406754, 'eval/latency_4/demon_left_rate_teacher': 0.12863436123348018, 'eval/latency_4/demon_right_precision': 0.769857433808554, 'eval/latency_4/demon_right_recall': 0.6823104693140795, 'eval/latency_4/demon_right_f1': 0.7234449760765551, 'eval/latency_4/demon_right_rate_model': 0.14419970631424375, 'eval/latency_4/demon_right_rate_teacher': 0.16270190895741557, 'eval/latency_4/demon_move_f1': 0.6905402220284254, 'eval/latency_4/demon_cc_f1': 0.8638632552282707, 'timing/data': 0.015285848639905453, 'timing/model': 1.1229039737954736, 'learning_rate/qwen_vl_interface': 8.646770131435774e-06, 'learning_rate/action_model': 8.646770131435773e-05, 'epoch': 0.69}) +[2026-06-17 20:02:46,878][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_1800_state +[2026-06-17 20:02:46,885][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 20:02:54,712][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_1800_state/model.safetensors +[2026-06-17 20:03:19,191][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_1800_state/optimizer.bin +[2026-06-17 20:03:19,192][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_1800_state/sampler.bin +[2026-06-17 20:03:19,193][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_1800_state/sampler_1.bin +[2026-06-17 20:03:19,195][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_1800_state/random_states_0.pkl +[2026-06-17 20:03:19,200][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1800 +[2026-06-17 20:03:19,200][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 20:03:19,218][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 20:03:24,034][starVLA.training.train_starvla][INFO] - Step 1801, Loss: {'train/loss': 0.5535246729850769, 'train/grad_norm_pre_clip': 7.436598300933838, 'timing/data': 0.01211525872349739, 'timing/model': 1.5007855575531721, 'learning_rate/qwen_vl_interface': 8.645258031766116e-06, 'learning_rate/action_model': 8.645258031766116e-05, 'epoch': 0.69}) +[2026-06-17 20:03:28,644][starVLA.training.train_starvla][INFO] - Step 1802, Loss: {'train/loss': 0.4535256624221802, 'train/grad_norm_pre_clip': 6.830501556396484, 'timing/data': 0.014596480876207352, 'timing/model': 1.388585016131401, 'learning_rate/qwen_vl_interface': 8.643745228255816e-06, 'learning_rate/action_model': 8.643745228255817e-05, 'epoch': 0.69}) +[2026-06-17 20:03:37,840][starVLA.training.train_starvla][INFO] - Step 1803, Loss: {'train/loss': 0.5667619109153748, 'train/grad_norm_pre_clip': 22.76799774169922, 'timing/data': 0.014959496445953846, 'timing/model': 1.087957501411438, 'learning_rate/qwen_vl_interface': 8.64223172121848e-06, 'learning_rate/action_model': 8.642231721218479e-05, 'epoch': 0.69}) +[2026-06-17 20:03:42,609][starVLA.training.train_starvla][INFO] - Step 1804, Loss: {'train/loss': 0.3085833787918091, 'train/grad_norm_pre_clip': 6.005810260772705, 'timing/data': 0.012585080228745937, 'timing/model': 1.1383884083479643, 'learning_rate/qwen_vl_interface': 8.640717510967857e-06, 'learning_rate/action_model': 8.640717510967856e-05, 'epoch': 0.69}) +[2026-06-17 20:03:48,101][starVLA.training.train_starvla][INFO] - Step 1805, Loss: {'train/loss': 0.3320210576057434, 'train/grad_norm_pre_clip': 7.704136848449707, 'timing/data': 0.014445707201957703, 'timing/model': 1.3195236781612039, 'learning_rate/qwen_vl_interface': 8.639202597817844e-06, 'learning_rate/action_model': 8.639202597817844e-05, 'epoch': 0.69}) +[2026-06-17 20:03:53,265][starVLA.training.train_starvla][INFO] - Step 1806, Loss: {'train/loss': 0.416019082069397, 'train/grad_norm_pre_clip': 14.155031204223633, 'timing/data': 0.018767972476780415, 'timing/model': 1.2839660411700606, 'learning_rate/qwen_vl_interface': 8.637686982082487e-06, 'learning_rate/action_model': 8.637686982082487e-05, 'epoch': 0.69}) +[2026-06-17 20:03:58,132][starVLA.training.train_starvla][INFO] - Step 1807, Loss: {'train/loss': 0.546877920627594, 'train/grad_norm_pre_clip': 14.114706993103027, 'timing/data': 0.012773833237588406, 'timing/model': 1.472678237594664, 'learning_rate/qwen_vl_interface': 8.636170664075972e-06, 'learning_rate/action_model': 8.636170664075972e-05, 'epoch': 0.7}) +[2026-06-17 20:04:05,053][starVLA.training.train_starvla][INFO] - Step 1808, Loss: {'train/loss': 0.8709122538566589, 'train/grad_norm_pre_clip': 11.950305938720703, 'timing/data': 0.012893331237137318, 'timing/model': 1.3435346176847816, 'learning_rate/qwen_vl_interface': 8.634653644112636e-06, 'learning_rate/action_model': 8.634653644112635e-05, 'epoch': 0.7}) +[2026-06-17 20:04:09,745][starVLA.training.train_starvla][INFO] - Step 1809, Loss: {'train/loss': 0.5298029780387878, 'train/grad_norm_pre_clip': 8.928940773010254, 'timing/data': 0.013643425889313221, 'timing/model': 1.1375444326549768, 'learning_rate/qwen_vl_interface': 8.633135922506957e-06, 'learning_rate/action_model': 8.633135922506957e-05, 'epoch': 0.7}) +[2026-06-17 20:04:14,525][starVLA.training.train_starvla][INFO] - Step 1810, Loss: {'train/loss': 0.5096703171730042, 'train/grad_norm_pre_clip': 12.056897163391113, 'timing/data': 0.010441703721880913, 'timing/model': 1.0733332112431526, 'learning_rate/qwen_vl_interface': 8.63161749957356e-06, 'learning_rate/action_model': 8.63161749957356e-05, 'epoch': 0.7}) +[2026-06-17 20:04:19,417][starVLA.training.train_starvla][INFO] - Step 1811, Loss: {'train/loss': 0.827664852142334, 'train/grad_norm_pre_clip': 9.879240989685059, 'timing/data': 0.011425546370446682, 'timing/model': 1.0753899523988366, 'learning_rate/qwen_vl_interface': 8.630098375627216e-06, 'learning_rate/action_model': 8.630098375627217e-05, 'epoch': 0.7}) +[2026-06-17 20:04:24,146][starVLA.training.train_starvla][INFO] - Step 1812, Loss: {'train/loss': 0.39794185757637024, 'train/grad_norm_pre_clip': 9.525978088378906, 'timing/data': 0.011041375808417797, 'timing/model': 1.5498182913288474, 'learning_rate/qwen_vl_interface': 8.628578550982846e-06, 'learning_rate/action_model': 8.628578550982845e-05, 'epoch': 0.7}) +[2026-06-17 20:04:33,363][starVLA.training.train_starvla][INFO] - Step 1813, Loss: {'train/loss': 0.4956067204475403, 'train/grad_norm_pre_clip': 7.7347846031188965, 'timing/data': 0.013183439150452614, 'timing/model': 5.05745599232614, 'learning_rate/qwen_vl_interface': 8.627058025955504e-06, 'learning_rate/action_model': 8.627058025955503e-05, 'epoch': 0.7}) +[2026-06-17 20:04:37,946][starVLA.training.train_starvla][INFO] - Step 1814, Loss: {'train/loss': 0.5875213146209717, 'train/grad_norm_pre_clip': 13.34964370727539, 'timing/data': 0.012874206528067589, 'timing/model': 1.104640201665461, 'learning_rate/qwen_vl_interface': 8.6255368008604e-06, 'learning_rate/action_model': 8.625536800860398e-05, 'epoch': 0.7}) +[2026-06-17 20:04:42,881][starVLA.training.train_starvla][INFO] - Step 1815, Loss: {'train/loss': 0.3121733069419861, 'train/grad_norm_pre_clip': 30.261613845825195, 'timing/data': 0.01924354862421751, 'timing/model': 1.1695287739858031, 'learning_rate/qwen_vl_interface': 8.624014876012883e-06, 'learning_rate/action_model': 8.624014876012882e-05, 'epoch': 0.7}) +[2026-06-17 20:04:47,612][starVLA.training.train_starvla][INFO] - Step 1816, Loss: {'train/loss': 0.6489211320877075, 'train/grad_norm_pre_clip': 11.423490524291992, 'timing/data': 0.012703019194304943, 'timing/model': 1.0853347005322576, 'learning_rate/qwen_vl_interface': 8.622492251728453e-06, 'learning_rate/action_model': 8.622492251728452e-05, 'epoch': 0.7}) +[2026-06-17 20:04:52,120][starVLA.training.train_starvla][INFO] - Step 1817, Loss: {'train/loss': 0.25596165657043457, 'train/grad_norm_pre_clip': 10.52927017211914, 'timing/data': 0.012224621139466763, 'timing/model': 1.2504333425313234, 'learning_rate/qwen_vl_interface': 8.62096892832275e-06, 'learning_rate/action_model': 8.620968928322749e-05, 'epoch': 0.7}) +[2026-06-17 20:05:00,215][starVLA.training.train_starvla][INFO] - Step 1818, Loss: {'train/loss': 0.44170907139778137, 'train/grad_norm_pre_clip': 9.931746482849121, 'timing/data': 0.01819617673754692, 'timing/model': 1.1439280798658729, 'learning_rate/qwen_vl_interface': 8.619444906111562e-06, 'learning_rate/action_model': 8.619444906111561e-05, 'epoch': 0.7}) +[2026-06-17 20:05:04,952][starVLA.training.train_starvla][INFO] - Step 1819, Loss: {'train/loss': 0.4018479585647583, 'train/grad_norm_pre_clip': 6.496758937835693, 'timing/data': 0.011329013854265213, 'timing/model': 1.1151518048718572, 'learning_rate/qwen_vl_interface': 8.617920185410817e-06, 'learning_rate/action_model': 8.617920185410817e-05, 'epoch': 0.7}) +[2026-06-17 20:05:10,029][starVLA.training.train_starvla][INFO] - Step 1820, Loss: {'train/loss': 0.6880256533622742, 'train/grad_norm_pre_clip': 6.961076259613037, 'timing/data': 0.012575935572385788, 'timing/model': 1.2251164559274912, 'learning_rate/qwen_vl_interface': 8.616394766536594e-06, 'learning_rate/action_model': 8.616394766536593e-05, 'epoch': 0.7}) +[2026-06-17 20:05:14,680][starVLA.training.train_starvla][INFO] - Step 1821, Loss: {'train/loss': 1.067837119102478, 'train/grad_norm_pre_clip': 14.329306602478027, 'timing/data': 0.010684913955628872, 'timing/model': 1.5290901577100158, 'learning_rate/qwen_vl_interface': 8.614868649805115e-06, 'learning_rate/action_model': 8.614868649805114e-05, 'epoch': 0.7}) +[2026-06-17 20:05:20,259][starVLA.training.train_starvla][INFO] - Step 1822, Loss: {'train/loss': 0.33351150155067444, 'train/grad_norm_pre_clip': 6.901658535003662, 'timing/data': 0.0121946781873703, 'timing/model': 1.4869197634980083, 'learning_rate/qwen_vl_interface': 8.613341835532743e-06, 'learning_rate/action_model': 8.613341835532743e-05, 'epoch': 0.7}) +[2026-06-17 20:05:24,529][starVLA.training.train_starvla][INFO] - Step 1823, Loss: {'train/loss': 0.5702545642852783, 'train/grad_norm_pre_clip': 11.980365753173828, 'timing/data': 0.012583056464791298, 'timing/model': 1.0942438812926412, 'learning_rate/qwen_vl_interface': 8.611814324035989e-06, 'learning_rate/action_model': 8.61181432403599e-05, 'epoch': 0.7}) +[2026-06-17 20:05:29,419][starVLA.training.train_starvla][INFO] - Step 1824, Loss: {'train/loss': 0.18578915297985077, 'train/grad_norm_pre_clip': 4.363123893737793, 'timing/data': 0.011257664300501347, 'timing/model': 1.1241082223132253, 'learning_rate/qwen_vl_interface': 8.610286115631508e-06, 'learning_rate/action_model': 8.610286115631509e-05, 'epoch': 0.7}) +[2026-06-17 20:05:34,337][starVLA.training.train_starvla][INFO] - Step 1825, Loss: {'train/loss': 0.705426037311554, 'train/grad_norm_pre_clip': 16.131053924560547, 'timing/data': 0.012344308197498322, 'timing/model': 1.1095689414069057, 'learning_rate/qwen_vl_interface': 8.608757210636101e-06, 'learning_rate/action_model': 8.608757210636101e-05, 'epoch': 0.7}) +[2026-06-17 20:05:41,390][starVLA.training.train_starvla][INFO] - Step 1826, Loss: {'train/loss': 0.5235735177993774, 'train/grad_norm_pre_clip': 11.300209045410156, 'timing/data': 0.020386281423270702, 'timing/model': 1.318453649058938, 'learning_rate/qwen_vl_interface': 8.60722760936671e-06, 'learning_rate/action_model': 8.60722760936671e-05, 'epoch': 0.7}) +[2026-06-17 20:05:46,131][starVLA.training.train_starvla][INFO] - Step 1827, Loss: {'train/loss': 0.4311192035675049, 'train/grad_norm_pre_clip': 6.6022138595581055, 'timing/data': 0.011801505461335182, 'timing/model': 1.5307653630152345, 'learning_rate/qwen_vl_interface': 8.605697312140421e-06, 'learning_rate/action_model': 8.605697312140422e-05, 'epoch': 0.7}) +[2026-06-17 20:05:50,508][starVLA.training.train_starvla][INFO] - Step 1828, Loss: {'train/loss': 0.5843584537506104, 'train/grad_norm_pre_clip': 9.107200622558594, 'timing/data': 0.011314892210066319, 'timing/model': 1.0863883420825005, 'learning_rate/qwen_vl_interface': 8.604166319274472e-06, 'learning_rate/action_model': 8.604166319274471e-05, 'epoch': 0.7}) +[2026-06-17 20:05:56,559][starVLA.training.train_starvla][INFO] - Step 1829, Loss: {'train/loss': 0.5372401475906372, 'train/grad_norm_pre_clip': 10.712172508239746, 'timing/data': 0.019104369916021824, 'timing/model': 1.1408867053687572, 'learning_rate/qwen_vl_interface': 8.602634631086233e-06, 'learning_rate/action_model': 8.602634631086233e-05, 'epoch': 0.7}) +[2026-06-17 20:06:06,572][starVLA.training.train_starvla][INFO] - Step 1830, Loss: {'train/loss': 0.5375402569770813, 'train/grad_norm_pre_clip': 9.576303482055664, 'timing/data': 0.011896231211721897, 'timing/model': 1.1058217119425535, 'learning_rate/qwen_vl_interface': 8.601102247893227e-06, 'learning_rate/action_model': 8.601102247893226e-05, 'epoch': 0.7}) +[2026-06-17 20:06:11,479][starVLA.training.train_starvla][INFO] - Step 1831, Loss: {'train/loss': 0.6074313521385193, 'train/grad_norm_pre_clip': 9.417435646057129, 'timing/data': 0.011708213947713375, 'timing/model': 1.1265106247738004, 'learning_rate/qwen_vl_interface': 8.599569170013121e-06, 'learning_rate/action_model': 8.59956917001312e-05, 'epoch': 0.7}) +[2026-06-17 20:06:16,170][starVLA.training.train_starvla][INFO] - Step 1832, Loss: {'train/loss': 0.35742858052253723, 'train/grad_norm_pre_clip': 6.848351001739502, 'timing/data': 0.016158117912709713, 'timing/model': 1.4501809049397707, 'learning_rate/qwen_vl_interface': 8.59803539776372e-06, 'learning_rate/action_model': 8.59803539776372e-05, 'epoch': 0.7}) +[2026-06-17 20:06:20,525][starVLA.training.train_starvla][INFO] - Step 1833, Loss: {'train/loss': 0.9508106112480164, 'train/grad_norm_pre_clip': 12.835951805114746, 'timing/data': 0.010796267539262772, 'timing/model': 1.092812587507069, 'learning_rate/qwen_vl_interface': 8.596500931462979e-06, 'learning_rate/action_model': 8.596500931462979e-05, 'epoch': 0.71}) +[2026-06-17 20:06:25,306][starVLA.training.train_starvla][INFO] - Step 1834, Loss: {'train/loss': 0.7666585445404053, 'train/grad_norm_pre_clip': 19.680023193359375, 'timing/data': 0.019337395206093788, 'timing/model': 1.1740450616925955, 'learning_rate/qwen_vl_interface': 8.594965771428993e-06, 'learning_rate/action_model': 8.594965771428993e-05, 'epoch': 0.71}) +[2026-06-17 20:06:30,068][starVLA.training.train_starvla][INFO] - Step 1835, Loss: {'train/loss': 0.4271155893802643, 'train/grad_norm_pre_clip': 8.244369506835938, 'timing/data': 0.011744685471057892, 'timing/model': 1.1373703200370073, 'learning_rate/qwen_vl_interface': 8.593429917980006e-06, 'learning_rate/action_model': 8.593429917980006e-05, 'epoch': 0.71}) +[2026-06-17 20:06:34,896][starVLA.training.train_starvla][INFO] - Step 1836, Loss: {'train/loss': 0.7305604815483093, 'train/grad_norm_pre_clip': 9.10830307006836, 'timing/data': 0.011234528385102749, 'timing/model': 1.063371048308909, 'learning_rate/qwen_vl_interface': 8.591893371434397e-06, 'learning_rate/action_model': 8.591893371434396e-05, 'epoch': 0.71}) +[2026-06-17 20:06:39,872][starVLA.training.train_starvla][INFO] - Step 1837, Loss: {'train/loss': 0.3042561411857605, 'train/grad_norm_pre_clip': 5.458536624908447, 'timing/data': 0.013001786544919014, 'timing/model': 1.5570864276960492, 'learning_rate/qwen_vl_interface': 8.590356132110695e-06, 'learning_rate/action_model': 8.590356132110695e-05, 'epoch': 0.71}) +[2026-06-17 20:06:44,415][starVLA.training.train_starvla][INFO] - Step 1838, Loss: {'train/loss': 0.41628384590148926, 'train/grad_norm_pre_clip': 9.545101165771484, 'timing/data': 0.011905108578503132, 'timing/model': 1.1040804283693433, 'learning_rate/qwen_vl_interface': 8.588818200327575e-06, 'learning_rate/action_model': 8.588818200327575e-05, 'epoch': 0.71}) +[2026-06-17 20:06:49,119][starVLA.training.train_starvla][INFO] - Step 1839, Loss: {'train/loss': 0.4057333171367645, 'train/grad_norm_pre_clip': 7.907263278961182, 'timing/data': 0.013667832128703594, 'timing/model': 1.1456568706780672, 'learning_rate/qwen_vl_interface': 8.587279576403848e-06, 'learning_rate/action_model': 8.587279576403847e-05, 'epoch': 0.71}) +[2026-06-17 20:06:53,725][starVLA.training.train_starvla][INFO] - Step 1840, Loss: {'train/loss': 0.44905105233192444, 'train/grad_norm_pre_clip': 8.074516296386719, 'timing/data': 0.01757163368165493, 'timing/model': 1.0817664256319404, 'learning_rate/qwen_vl_interface': 8.585740260658475e-06, 'learning_rate/action_model': 8.585740260658475e-05, 'epoch': 0.71}) +[2026-06-17 20:06:58,316][starVLA.training.train_starvla][INFO] - Step 1841, Loss: {'train/loss': 0.45294541120529175, 'train/grad_norm_pre_clip': 8.882132530212402, 'timing/data': 0.012081431224942207, 'timing/model': 1.1012991899624467, 'learning_rate/qwen_vl_interface': 8.584200253410557e-06, 'learning_rate/action_model': 8.584200253410556e-05, 'epoch': 0.71}) +[2026-06-17 20:07:03,302][starVLA.training.train_starvla][INFO] - Step 1842, Loss: {'train/loss': 0.7980971336364746, 'train/grad_norm_pre_clip': 13.202359199523926, 'timing/data': 0.012525012716650963, 'timing/model': 1.5709519106894732, 'learning_rate/qwen_vl_interface': 8.582659554979336e-06, 'learning_rate/action_model': 8.582659554979336e-05, 'epoch': 0.71}) +[2026-06-17 20:07:11,792][starVLA.training.train_starvla][INFO] - Step 1843, Loss: {'train/loss': 0.3821799159049988, 'train/grad_norm_pre_clip': 11.999661445617676, 'timing/data': 0.02490155678242445, 'timing/model': 3.144940354861319, 'learning_rate/qwen_vl_interface': 8.581118165684204e-06, 'learning_rate/action_model': 8.581118165684204e-05, 'epoch': 0.71}) +[2026-06-17 20:07:17,675][starVLA.training.train_starvla][INFO] - Step 1844, Loss: {'train/loss': 0.6220603585243225, 'train/grad_norm_pre_clip': 11.097370147705078, 'timing/data': 0.011272125877439976, 'timing/model': 1.1082265162840486, 'learning_rate/qwen_vl_interface': 8.579576085844695e-06, 'learning_rate/action_model': 8.579576085844694e-05, 'epoch': 0.71}) +[2026-06-17 20:07:22,470][starVLA.training.train_starvla][INFO] - Step 1845, Loss: {'train/loss': 0.7963021397590637, 'train/grad_norm_pre_clip': 8.719159126281738, 'timing/data': 0.012308536097407341, 'timing/model': 1.1258697463199496, 'learning_rate/qwen_vl_interface': 8.578033315780474e-06, 'learning_rate/action_model': 8.578033315780475e-05, 'epoch': 0.71}) +[2026-06-17 20:07:27,210][starVLA.training.train_starvla][INFO] - Step 1846, Loss: {'train/loss': 0.6342382431030273, 'train/grad_norm_pre_clip': 8.354438781738281, 'timing/data': 0.01114228181540966, 'timing/model': 1.2148476196452975, 'learning_rate/qwen_vl_interface': 8.576489855811371e-06, 'learning_rate/action_model': 8.57648985581137e-05, 'epoch': 0.71}) +[2026-06-17 20:07:31,827][starVLA.training.train_starvla][INFO] - Step 1847, Loss: {'train/loss': 0.46008798480033875, 'train/grad_norm_pre_clip': 10.806619644165039, 'timing/data': 0.011607280932366848, 'timing/model': 1.3504993561655283, 'learning_rate/qwen_vl_interface': 8.57494570625734e-06, 'learning_rate/action_model': 8.57494570625734e-05, 'epoch': 0.71}) +[2026-06-17 20:07:36,469][starVLA.training.train_starvla][INFO] - Step 1848, Loss: {'train/loss': 0.8460186123847961, 'train/grad_norm_pre_clip': 10.10036563873291, 'timing/data': 0.01201704517006874, 'timing/model': 1.2124317465350032, 'learning_rate/qwen_vl_interface': 8.573400867438486e-06, 'learning_rate/action_model': 8.573400867438486e-05, 'epoch': 0.71}) +[2026-06-17 20:07:41,701][starVLA.training.train_starvla][INFO] - Step 1849, Loss: {'train/loss': 0.4494777321815491, 'train/grad_norm_pre_clip': 6.509730339050293, 'timing/data': 0.01679378468543291, 'timing/model': 1.164991807192564, 'learning_rate/qwen_vl_interface': 8.571855339675057e-06, 'learning_rate/action_model': 8.571855339675057e-05, 'epoch': 0.71}) +[2026-06-17 20:07:46,524][starVLA.training.train_starvla][INFO] - Step 1850, Loss: {'train/loss': 0.7287387847900391, 'train/grad_norm_pre_clip': 8.049044609069824, 'timing/data': 0.013884427025914192, 'timing/model': 1.1231005704030395, 'learning_rate/qwen_vl_interface': 8.57030912328744e-06, 'learning_rate/action_model': 8.57030912328744e-05, 'epoch': 0.71}) +[2026-06-17 20:07:51,258][starVLA.training.train_starvla][INFO] - Step 1851, Loss: {'train/loss': 0.709360659122467, 'train/grad_norm_pre_clip': 8.015663146972656, 'timing/data': 0.012740470468997955, 'timing/model': 1.1146845007315278, 'learning_rate/qwen_vl_interface': 8.568762218596172e-06, 'learning_rate/action_model': 8.568762218596171e-05, 'epoch': 0.71}) +[2026-06-17 20:07:56,329][starVLA.training.train_starvla][INFO] - Step 1852, Loss: {'train/loss': 0.7948085069656372, 'train/grad_norm_pre_clip': 10.143964767456055, 'timing/data': 0.011814371682703495, 'timing/model': 1.6397206988185644, 'learning_rate/qwen_vl_interface': 8.567214625921922e-06, 'learning_rate/action_model': 8.567214625921922e-05, 'epoch': 0.71}) +[2026-06-17 20:08:00,844][starVLA.training.train_starvla][INFO] - Step 1853, Loss: {'train/loss': 0.7041462659835815, 'train/grad_norm_pre_clip': 8.70520305633545, 'timing/data': 0.019508826546370983, 'timing/model': 1.1648906776681542, 'learning_rate/qwen_vl_interface': 8.565666345585511e-06, 'learning_rate/action_model': 8.56566634558551e-05, 'epoch': 0.71}) +[2026-06-17 20:08:05,547][starVLA.training.train_starvla][INFO] - Step 1854, Loss: {'train/loss': 0.4671984016895294, 'train/grad_norm_pre_clip': 6.7057390213012695, 'timing/data': 0.011438053101301193, 'timing/model': 1.1621564403176308, 'learning_rate/qwen_vl_interface': 8.564117377907898e-06, 'learning_rate/action_model': 8.564117377907898e-05, 'epoch': 0.71}) +[2026-06-17 20:08:10,336][starVLA.training.train_starvla][INFO] - Step 1855, Loss: {'train/loss': 0.8617857694625854, 'train/grad_norm_pre_clip': 13.678766250610352, 'timing/data': 0.011953365057706833, 'timing/model': 1.1317495396360755, 'learning_rate/qwen_vl_interface': 8.56256772321019e-06, 'learning_rate/action_model': 8.562567723210188e-05, 'epoch': 0.71}) +[2026-06-17 20:08:15,224][starVLA.training.train_starvla][INFO] - Step 1856, Loss: {'train/loss': 0.6832274198532104, 'train/grad_norm_pre_clip': 8.589974403381348, 'timing/data': 0.012444907799363136, 'timing/model': 1.415514025837183, 'learning_rate/qwen_vl_interface': 8.561017381813624e-06, 'learning_rate/action_model': 8.561017381813624e-05, 'epoch': 0.71}) +[2026-06-17 20:08:19,791][starVLA.training.train_starvla][INFO] - Step 1857, Loss: {'train/loss': 0.3534613251686096, 'train/grad_norm_pre_clip': 8.168776512145996, 'timing/data': 0.018498506397008896, 'timing/model': 1.4222060292959213, 'learning_rate/qwen_vl_interface': 8.559466354039594e-06, 'learning_rate/action_model': 8.559466354039594e-05, 'epoch': 0.71}) +[2026-06-17 20:08:24,243][starVLA.training.train_starvla][INFO] - Step 1858, Loss: {'train/loss': 0.5875610113143921, 'train/grad_norm_pre_clip': 7.209347248077393, 'timing/data': 0.018635776825249195, 'timing/model': 1.1265275664627552, 'learning_rate/qwen_vl_interface': 8.557914640209628e-06, 'learning_rate/action_model': 8.557914640209627e-05, 'epoch': 0.71}) +[2026-06-17 20:08:28,987][starVLA.training.train_starvla][INFO] - Step 1859, Loss: {'train/loss': 0.6523515582084656, 'train/grad_norm_pre_clip': 7.732518196105957, 'timing/data': 0.0136417793110013, 'timing/model': 1.2867717426270247, 'learning_rate/qwen_vl_interface': 8.556362240645397e-06, 'learning_rate/action_model': 8.556362240645397e-05, 'epoch': 0.72}) +[2026-06-17 20:08:33,672][starVLA.training.train_starvla][INFO] - Step 1860, Loss: {'train/loss': 0.4941370189189911, 'train/grad_norm_pre_clip': 6.772757053375244, 'timing/data': 0.01203186810016632, 'timing/model': 1.1609256900846958, 'learning_rate/qwen_vl_interface': 8.554809155668716e-06, 'learning_rate/action_model': 8.554809155668716e-05, 'epoch': 0.72}) +[2026-06-17 20:08:38,616][starVLA.training.train_starvla][INFO] - Step 1861, Loss: {'train/loss': 0.7094195485115051, 'train/grad_norm_pre_clip': 10.34989070892334, 'timing/data': 0.013929652981460094, 'timing/model': 1.1772102629765868, 'learning_rate/qwen_vl_interface': 8.553255385601541e-06, 'learning_rate/action_model': 8.55325538560154e-05, 'epoch': 0.72}) +[2026-06-17 20:08:43,443][starVLA.training.train_starvla][INFO] - Step 1862, Loss: {'train/loss': 0.7159666419029236, 'train/grad_norm_pre_clip': 13.018077850341797, 'timing/data': 0.011516472324728966, 'timing/model': 1.3801291156560183, 'learning_rate/qwen_vl_interface': 8.551700930765969e-06, 'learning_rate/action_model': 8.551700930765969e-05, 'epoch': 0.72}) +[2026-06-17 20:08:47,934][starVLA.training.train_starvla][INFO] - Step 1863, Loss: {'train/loss': 0.6474072933197021, 'train/grad_norm_pre_clip': 11.729270935058594, 'timing/data': 0.010868556797504425, 'timing/model': 1.0754375718533993, 'learning_rate/qwen_vl_interface': 8.550145791484243e-06, 'learning_rate/action_model': 8.550145791484242e-05, 'epoch': 0.72}) +[2026-06-17 20:08:58,779][starVLA.training.train_starvla][INFO] - Step 1864, Loss: {'train/loss': 0.5447772741317749, 'train/grad_norm_pre_clip': 7.54215145111084, 'timing/data': 0.015236222185194492, 'timing/model': 4.290285425260663, 'learning_rate/qwen_vl_interface': 8.548589968078743e-06, 'learning_rate/action_model': 8.548589968078743e-05, 'epoch': 0.72}) +[2026-06-17 20:09:12,158][starVLA.training.train_starvla][INFO] - Step 1865, Loss: {'train/loss': 0.4700200855731964, 'train/grad_norm_pre_clip': 36.33964920043945, 'timing/data': 0.06774871237576008, 'timing/model': 1.1534519344568253, 'learning_rate/qwen_vl_interface': 8.547033460871992e-06, 'learning_rate/action_model': 8.547033460871992e-05, 'epoch': 0.72}) +[2026-06-17 20:09:17,054][starVLA.training.train_starvla][INFO] - Step 1866, Loss: {'train/loss': 0.41615086793899536, 'train/grad_norm_pre_clip': 6.684080123901367, 'timing/data': 0.011860913597047329, 'timing/model': 1.1389571148902178, 'learning_rate/qwen_vl_interface': 8.545476270186658e-06, 'learning_rate/action_model': 8.545476270186659e-05, 'epoch': 0.72}) +[2026-06-17 20:09:21,829][starVLA.training.train_starvla][INFO] - Step 1867, Loss: {'train/loss': 0.5439127087593079, 'train/grad_norm_pre_clip': 7.1958441734313965, 'timing/data': 0.012702219188213348, 'timing/model': 1.408963417634368, 'learning_rate/qwen_vl_interface': 8.543918396345546e-06, 'learning_rate/action_model': 8.543918396345546e-05, 'epoch': 0.72}) +[2026-06-17 20:09:26,391][starVLA.training.train_starvla][INFO] - Step 1868, Loss: {'train/loss': 0.7143151164054871, 'train/grad_norm_pre_clip': 8.486257553100586, 'timing/data': 0.011735823005437851, 'timing/model': 1.2768416237086058, 'learning_rate/qwen_vl_interface': 8.542359839671607e-06, 'learning_rate/action_model': 8.542359839671606e-05, 'epoch': 0.72}) +[2026-06-17 20:09:32,250][starVLA.training.train_starvla][INFO] - Step 1869, Loss: {'train/loss': 0.8120572566986084, 'train/grad_norm_pre_clip': 12.196744918823242, 'timing/data': 0.011360137723386288, 'timing/model': 1.1469940058887005, 'learning_rate/qwen_vl_interface': 8.540800600487931e-06, 'learning_rate/action_model': 8.54080060048793e-05, 'epoch': 0.72}) +[2026-06-17 20:09:36,795][starVLA.training.train_starvla][INFO] - Step 1870, Loss: {'train/loss': 0.4921133518218994, 'train/grad_norm_pre_clip': 8.379852294921875, 'timing/data': 0.01703838724642992, 'timing/model': 1.0834232801571488, 'learning_rate/qwen_vl_interface': 8.539240679117748e-06, 'learning_rate/action_model': 8.539240679117748e-05, 'epoch': 0.72}) +[2026-06-17 20:09:42,765][starVLA.training.train_starvla][INFO] - Step 1871, Loss: {'train/loss': 0.4361959397792816, 'train/grad_norm_pre_clip': 5.501613616943359, 'timing/data': 0.013316608034074306, 'timing/model': 1.159838200546801, 'learning_rate/qwen_vl_interface': 8.537680075884433e-06, 'learning_rate/action_model': 8.537680075884433e-05, 'epoch': 0.72}) +[2026-06-17 20:09:47,554][starVLA.training.train_starvla][INFO] - Step 1872, Loss: {'train/loss': 0.8051154613494873, 'train/grad_norm_pre_clip': 6.785646915435791, 'timing/data': 0.014923341572284698, 'timing/model': 1.5848710453137755, 'learning_rate/qwen_vl_interface': 8.5361187911115e-06, 'learning_rate/action_model': 8.5361187911115e-05, 'epoch': 0.72}) +[2026-06-17 20:09:51,878][starVLA.training.train_starvla][INFO] - Step 1873, Loss: {'train/loss': 0.442403644323349, 'train/grad_norm_pre_clip': 5.890995502471924, 'timing/data': 0.011731283739209175, 'timing/model': 1.1763692488893867, 'learning_rate/qwen_vl_interface': 8.534556825122608e-06, 'learning_rate/action_model': 8.534556825122607e-05, 'epoch': 0.72}) +[2026-06-17 20:09:56,592][starVLA.training.train_starvla][INFO] - Step 1874, Loss: {'train/loss': 0.6022398471832275, 'train/grad_norm_pre_clip': 7.8236083984375, 'timing/data': 0.011456618085503578, 'timing/model': 1.0924837347120047, 'learning_rate/qwen_vl_interface': 8.532994178241551e-06, 'learning_rate/action_model': 8.53299417824155e-05, 'epoch': 0.72}) +[2026-06-17 20:10:01,371][starVLA.training.train_starvla][INFO] - Step 1875, Loss: {'train/loss': 0.5639674067497253, 'train/grad_norm_pre_clip': 8.85744571685791, 'timing/data': 0.010806260630488396, 'timing/model': 1.0816756458953023, 'learning_rate/qwen_vl_interface': 8.531430850792267e-06, 'learning_rate/action_model': 8.531430850792266e-05, 'epoch': 0.72}) +[2026-06-17 20:10:07,862][starVLA.training.train_starvla][INFO] - Step 1876, Loss: {'train/loss': 0.8045833110809326, 'train/grad_norm_pre_clip': 8.79813289642334, 'timing/data': 0.01610078103840351, 'timing/model': 1.2285095676779747, 'learning_rate/qwen_vl_interface': 8.529866843098838e-06, 'learning_rate/action_model': 8.529866843098837e-05, 'epoch': 0.72}) +[2026-06-17 20:10:13,123][starVLA.training.train_starvla][INFO] - Step 1877, Loss: {'train/loss': 0.41978511214256287, 'train/grad_norm_pre_clip': 7.445926666259766, 'timing/data': 0.010746066458523273, 'timing/model': 2.0110654029995203, 'learning_rate/qwen_vl_interface': 8.528302155485482e-06, 'learning_rate/action_model': 8.528302155485481e-05, 'epoch': 0.72}) +[2026-06-17 20:10:17,479][starVLA.training.train_starvla][INFO] - Step 1878, Loss: {'train/loss': 0.2473696619272232, 'train/grad_norm_pre_clip': 3.978980779647827, 'timing/data': 0.014177395030856133, 'timing/model': 1.1193284774199128, 'learning_rate/qwen_vl_interface': 8.526736788276563e-06, 'learning_rate/action_model': 8.526736788276563e-05, 'epoch': 0.72}) +[2026-06-17 20:10:23,654][starVLA.training.train_starvla][INFO] - Step 1879, Loss: {'train/loss': 0.47451817989349365, 'train/grad_norm_pre_clip': 6.782088756561279, 'timing/data': 0.012475728057324886, 'timing/model': 2.2069905856624246, 'learning_rate/qwen_vl_interface': 8.525170741796578e-06, 'learning_rate/action_model': 8.525170741796579e-05, 'epoch': 0.72}) +[2026-06-17 20:10:28,622][starVLA.training.train_starvla][INFO] - Step 1880, Loss: {'train/loss': 0.3999510407447815, 'train/grad_norm_pre_clip': 5.406577110290527, 'timing/data': 0.01144432369619608, 'timing/model': 1.1751889437437057, 'learning_rate/qwen_vl_interface': 8.523604016370179e-06, 'learning_rate/action_model': 8.523604016370179e-05, 'epoch': 0.72}) +[2026-06-17 20:10:33,465][starVLA.training.train_starvla][INFO] - Step 1881, Loss: {'train/loss': 0.3217194378376007, 'train/grad_norm_pre_clip': 25.835153579711914, 'timing/data': 0.011305825784802437, 'timing/model': 1.063482522033155, 'learning_rate/qwen_vl_interface': 8.522036612322143e-06, 'learning_rate/action_model': 8.522036612322142e-05, 'epoch': 0.72}) +[2026-06-17 20:10:38,178][starVLA.training.train_starvla][INFO] - Step 1882, Loss: {'train/loss': 0.4919547438621521, 'train/grad_norm_pre_clip': 8.864949226379395, 'timing/data': 0.012199558317661285, 'timing/model': 1.486118832603097, 'learning_rate/qwen_vl_interface': 8.520468529977395e-06, 'learning_rate/action_model': 8.520468529977395e-05, 'epoch': 0.72}) +[2026-06-17 20:10:42,669][starVLA.training.train_starvla][INFO] - Step 1883, Loss: {'train/loss': 0.4249502420425415, 'train/grad_norm_pre_clip': 9.985457420349121, 'timing/data': 0.01184314489364624, 'timing/model': 1.1012574918568134, 'learning_rate/qwen_vl_interface': 8.518899769661002e-06, 'learning_rate/action_model': 8.518899769661002e-05, 'epoch': 0.72}) +[2026-06-17 20:10:47,416][starVLA.training.train_starvla][INFO] - Step 1884, Loss: {'train/loss': 0.7999574542045593, 'train/grad_norm_pre_clip': 11.524360656738281, 'timing/data': 0.012377177365124226, 'timing/model': 1.2136223996058106, 'learning_rate/qwen_vl_interface': 8.517330331698171e-06, 'learning_rate/action_model': 8.51733033169817e-05, 'epoch': 0.72}) +[2026-06-17 20:10:52,353][starVLA.training.train_starvla][INFO] - Step 1885, Loss: {'train/loss': 0.6584240794181824, 'train/grad_norm_pre_clip': 7.027801990509033, 'timing/data': 0.011152273043990135, 'timing/model': 1.1788305528461933, 'learning_rate/qwen_vl_interface': 8.515760216414244e-06, 'learning_rate/action_model': 8.515760216414244e-05, 'epoch': 0.73}) +[2026-06-17 20:10:57,401][starVLA.training.train_starvla][INFO] - Step 1886, Loss: {'train/loss': 0.37157392501831055, 'train/grad_norm_pre_clip': 5.306651592254639, 'timing/data': 0.01214512251317501, 'timing/model': 1.3088705725967884, 'learning_rate/qwen_vl_interface': 8.514189424134712e-06, 'learning_rate/action_model': 8.514189424134713e-05, 'epoch': 0.73}) +[2026-06-17 20:11:02,186][starVLA.training.train_starvla][INFO] - Step 1887, Loss: {'train/loss': 0.7515087127685547, 'train/grad_norm_pre_clip': 11.402692794799805, 'timing/data': 0.011779741384088993, 'timing/model': 1.4143925420939922, 'learning_rate/qwen_vl_interface': 8.512617955185199e-06, 'learning_rate/action_model': 8.5126179551852e-05, 'epoch': 0.73}) +[2026-06-17 20:11:06,499][starVLA.training.train_starvla][INFO] - Step 1888, Loss: {'train/loss': 0.26848867535591125, 'train/grad_norm_pre_clip': 7.718338489532471, 'timing/data': 0.012560134753584862, 'timing/model': 1.0631154347211123, 'learning_rate/qwen_vl_interface': 8.511045809891472e-06, 'learning_rate/action_model': 8.511045809891472e-05, 'epoch': 0.73}) +[2026-06-17 20:11:11,586][starVLA.training.train_starvla][INFO] - Step 1889, Loss: {'train/loss': 0.30803367495536804, 'train/grad_norm_pre_clip': 8.883843421936035, 'timing/data': 0.01311657764017582, 'timing/model': 1.3273520516231656, 'learning_rate/qwen_vl_interface': 8.509472988579443e-06, 'learning_rate/action_model': 8.509472988579443e-05, 'epoch': 0.73}) +[2026-06-17 20:11:16,652][starVLA.training.train_starvla][INFO] - Step 1890, Loss: {'train/loss': 0.5320782661437988, 'train/grad_norm_pre_clip': 10.025362968444824, 'timing/data': 0.01339911948889494, 'timing/model': 1.3341610096395016, 'learning_rate/qwen_vl_interface': 8.507899491575153e-06, 'learning_rate/action_model': 8.507899491575154e-05, 'epoch': 0.73}) +[2026-06-17 20:11:21,580][starVLA.training.train_starvla][INFO] - Step 1891, Loss: {'train/loss': 0.5415151715278625, 'train/grad_norm_pre_clip': 32.420928955078125, 'timing/data': 0.013190147466957569, 'timing/model': 1.0912497900426388, 'learning_rate/qwen_vl_interface': 8.506325319204795e-06, 'learning_rate/action_model': 8.506325319204796e-05, 'epoch': 0.73}) +[2026-06-17 20:11:26,443][starVLA.training.train_starvla][INFO] - Step 1892, Loss: {'train/loss': 0.7906723618507385, 'train/grad_norm_pre_clip': 11.476263046264648, 'timing/data': 0.01135145965963602, 'timing/model': 1.4109444748610258, 'learning_rate/qwen_vl_interface': 8.504750471794694e-06, 'learning_rate/action_model': 8.504750471794693e-05, 'epoch': 0.73}) +[2026-06-17 20:11:30,782][starVLA.training.train_starvla][INFO] - Step 1893, Loss: {'train/loss': 0.5598545074462891, 'train/grad_norm_pre_clip': 8.138211250305176, 'timing/data': 0.013952480629086494, 'timing/model': 1.0625365413725376, 'learning_rate/qwen_vl_interface': 8.503174949671315e-06, 'learning_rate/action_model': 8.503174949671315e-05, 'epoch': 0.73}) +[2026-06-17 20:11:35,314][starVLA.training.train_starvla][INFO] - Step 1894, Loss: {'train/loss': 0.7112075090408325, 'train/grad_norm_pre_clip': 11.419878005981445, 'timing/data': 0.011146677657961845, 'timing/model': 1.0753906928002834, 'learning_rate/qwen_vl_interface': 8.501598753161273e-06, 'learning_rate/action_model': 8.501598753161272e-05, 'epoch': 0.73}) +[2026-06-17 20:11:40,335][starVLA.training.train_starvla][INFO] - Step 1895, Loss: {'train/loss': 0.6091889142990112, 'train/grad_norm_pre_clip': 7.696547031402588, 'timing/data': 0.011249464005231857, 'timing/model': 1.1360555179417133, 'learning_rate/qwen_vl_interface': 8.500021882591308e-06, 'learning_rate/action_model': 8.500021882591308e-05, 'epoch': 0.73}) +[2026-06-17 20:11:50,248][starVLA.training.train_starvla][INFO] - Step 1896, Loss: {'train/loss': 0.5532318353652954, 'train/grad_norm_pre_clip': 8.55082893371582, 'timing/data': 0.02122167870402336, 'timing/model': 4.295746662653983, 'learning_rate/qwen_vl_interface': 8.498444338288308e-06, 'learning_rate/action_model': 8.498444338288309e-05, 'epoch': 0.73}) +[2026-06-17 20:12:05,372][starVLA.training.train_starvla][INFO] - Step 1897, Loss: {'train/loss': 0.7885638475418091, 'train/grad_norm_pre_clip': 11.747407913208008, 'timing/data': 0.012856611981987953, 'timing/model': 5.29485927708447, 'learning_rate/qwen_vl_interface': 8.496866120579304e-06, 'learning_rate/action_model': 8.496866120579303e-05, 'epoch': 0.73}) +[2026-06-17 20:12:12,412][starVLA.training.train_starvla][INFO] - Step 1898, Loss: {'train/loss': 0.4655659794807434, 'train/grad_norm_pre_clip': 7.751579761505127, 'timing/data': 0.012257516384124756, 'timing/model': 1.093211148865521, 'learning_rate/qwen_vl_interface': 8.495287229791459e-06, 'learning_rate/action_model': 8.495287229791459e-05, 'epoch': 0.73}) +[2026-06-17 20:12:21,338][starVLA.training.train_starvla][INFO] - Step 1899, Loss: {'train/loss': 0.6945474743843079, 'train/grad_norm_pre_clip': 6.475635528564453, 'timing/data': 0.015559539198875427, 'timing/model': 1.1804546862840652, 'learning_rate/qwen_vl_interface': 8.493707666252078e-06, 'learning_rate/action_model': 8.493707666252078e-05, 'epoch': 0.73}) +[2026-06-17 20:15:37,824][starVLA.training.train_starvla][INFO] - Step 1900, Loss: {'train/loss': 0.510094940662384, 'train/grad_norm_pre_clip': 6.992099761962891, 'eval/loss': 0.5711391568183899, 'eval/latency_4/loss': 0.5711391568183899, 'eval/demon_fire_precision': 0.9524851516098781, 'eval/demon_fire_recall': 0.9896070152646963, 'eval/demon_fire_f1': 0.9706913029627271, 'eval/demon_fire_rate_model': 0.9395007342143906, 'eval/demon_fire_rate_teacher': 0.9042584434654919, 'eval/demon_left_precision': 0.750733137829912, 'eval/demon_left_recall': 0.5844748858447488, 'eval/demon_left_f1': 0.6572528883183568, 'eval/demon_left_rate_model': 0.10014684287812041, 'eval/demon_left_rate_teacher': 0.12863436123348018, 'eval/demon_right_precision': 0.6300813008130082, 'eval/demon_right_recall': 0.8393501805054152, 'eval/demon_right_f1': 0.7198142414860681, 'eval/demon_right_rate_model': 0.21674008810572687, 'eval/demon_right_rate_teacher': 0.16270190895741557, 'eval/demon_move_f1': 0.6885335649022124, 'eval/demon_cc_f1': 0.864420674480523, 'eval/latency_4/demon_fire_precision': 0.9524851516098781, 'eval/latency_4/demon_fire_recall': 0.9896070152646963, 'eval/latency_4/demon_fire_f1': 0.9706913029627271, 'eval/latency_4/demon_fire_rate_model': 0.9395007342143906, 'eval/latency_4/demon_fire_rate_teacher': 0.9042584434654919, 'eval/latency_4/demon_left_precision': 0.750733137829912, 'eval/latency_4/demon_left_recall': 0.5844748858447488, 'eval/latency_4/demon_left_f1': 0.6572528883183568, 'eval/latency_4/demon_left_rate_model': 0.10014684287812041, 'eval/latency_4/demon_left_rate_teacher': 0.12863436123348018, 'eval/latency_4/demon_right_precision': 0.6300813008130082, 'eval/latency_4/demon_right_recall': 0.8393501805054152, 'eval/latency_4/demon_right_f1': 0.7198142414860681, 'eval/latency_4/demon_right_rate_model': 0.21674008810572687, 'eval/latency_4/demon_right_rate_teacher': 0.16270190895741557, 'eval/latency_4/demon_move_f1': 0.6885335649022124, 'eval/latency_4/demon_cc_f1': 0.864420674480523, 'timing/data': 0.011461571790277958, 'timing/model': 5.013156971894205, 'learning_rate/qwen_vl_interface': 8.492127430288609e-06, 'learning_rate/action_model': 8.492127430288608e-05, 'epoch': 0.73}) +[2026-06-17 20:15:37,838][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_1900_state +[2026-06-17 20:15:37,847][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 20:15:46,671][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_1900_state/model.safetensors +[2026-06-17 20:16:09,149][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_1900_state/optimizer.bin +[2026-06-17 20:16:09,152][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_1900_state/sampler.bin +[2026-06-17 20:16:09,153][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_1900_state/sampler_1.bin +[2026-06-17 20:16:09,155][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_1900_state/random_states_0.pkl +[2026-06-17 20:16:09,160][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1900 +[2026-06-17 20:16:09,160][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 20:16:09,178][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 20:16:14,016][starVLA.training.train_starvla][INFO] - Step 1901, Loss: {'train/loss': 0.4480433464050293, 'train/grad_norm_pre_clip': 7.729280948638916, 'timing/data': 0.0116628622636199, 'timing/model': 1.6087740594521165, 'learning_rate/qwen_vl_interface': 8.490546522228633e-06, 'learning_rate/action_model': 8.490546522228633e-05, 'epoch': 0.73}) +[2026-06-17 20:16:21,683][starVLA.training.train_starvla][INFO] - Step 1902, Loss: {'train/loss': 0.3329530656337738, 'train/grad_norm_pre_clip': 4.9693450927734375, 'timing/data': 0.011610597372055054, 'timing/model': 4.419431020505726, 'learning_rate/qwen_vl_interface': 8.488964942399876e-06, 'learning_rate/action_model': 8.488964942399877e-05, 'epoch': 0.73}) +[2026-06-17 20:16:26,415][starVLA.training.train_starvla][INFO] - Step 1903, Loss: {'train/loss': 0.5835828185081482, 'train/grad_norm_pre_clip': 8.42431354522705, 'timing/data': 0.014879933558404446, 'timing/model': 1.2004151782020926, 'learning_rate/qwen_vl_interface': 8.487382691130203e-06, 'learning_rate/action_model': 8.487382691130202e-05, 'epoch': 0.73}) +[2026-06-17 20:16:31,108][starVLA.training.train_starvla][INFO] - Step 1904, Loss: {'train/loss': 0.41581636667251587, 'train/grad_norm_pre_clip': 6.252800464630127, 'timing/data': 0.012755844742059708, 'timing/model': 1.1213003043085337, 'learning_rate/qwen_vl_interface': 8.485799768747614e-06, 'learning_rate/action_model': 8.485799768747614e-05, 'epoch': 0.73}) +[2026-06-17 20:16:41,199][starVLA.training.train_starvla][INFO] - Step 1905, Loss: {'train/loss': 0.5384307503700256, 'train/grad_norm_pre_clip': 9.94869613647461, 'timing/data': 0.01227900292724371, 'timing/model': 3.911813059821725, 'learning_rate/qwen_vl_interface': 8.484216175580248e-06, 'learning_rate/action_model': 8.484216175580249e-05, 'epoch': 0.73}) +[2026-06-17 20:16:46,262][starVLA.training.train_starvla][INFO] - Step 1906, Loss: {'train/loss': 0.5045837759971619, 'train/grad_norm_pre_clip': 8.888935089111328, 'timing/data': 0.013285510241985321, 'timing/model': 1.2777154110372066, 'learning_rate/qwen_vl_interface': 8.482631911956395e-06, 'learning_rate/action_model': 8.482631911956395e-05, 'epoch': 0.73}) +[2026-06-17 20:16:51,584][starVLA.training.train_starvla][INFO] - Step 1907, Loss: {'train/loss': 0.6047602891921997, 'train/grad_norm_pre_clip': 9.001120567321777, 'timing/data': 0.01138350274413824, 'timing/model': 1.6661651944741607, 'learning_rate/qwen_vl_interface': 8.481046978204462e-06, 'learning_rate/action_model': 8.481046978204463e-05, 'epoch': 0.73}) +[2026-06-17 20:16:57,929][starVLA.training.train_starvla][INFO] - Step 1908, Loss: {'train/loss': 0.4725024104118347, 'train/grad_norm_pre_clip': 7.078158855438232, 'timing/data': 0.015676552429795265, 'timing/model': 1.2385869175195694, 'learning_rate/qwen_vl_interface': 8.479461374653017e-06, 'learning_rate/action_model': 8.479461374653018e-05, 'epoch': 0.73}) +[2026-06-17 20:17:02,714][starVLA.training.train_starvla][INFO] - Step 1909, Loss: {'train/loss': 0.8221837878227234, 'train/grad_norm_pre_clip': 12.04664421081543, 'timing/data': 0.013745452277362347, 'timing/model': 1.0886325184255838, 'learning_rate/qwen_vl_interface': 8.477875101630752e-06, 'learning_rate/action_model': 8.477875101630753e-05, 'epoch': 0.73}) +[2026-06-17 20:17:07,741][starVLA.training.train_starvla][INFO] - Step 1910, Loss: {'train/loss': 0.29323217272758484, 'train/grad_norm_pre_clip': 4.398926258087158, 'timing/data': 0.01311495341360569, 'timing/model': 1.1845292383804917, 'learning_rate/qwen_vl_interface': 8.476288159466508e-06, 'learning_rate/action_model': 8.476288159466508e-05, 'epoch': 0.73}) +[2026-06-17 20:17:16,646][starVLA.training.train_starvla][INFO] - Step 1911, Loss: {'train/loss': 0.40478190779685974, 'train/grad_norm_pre_clip': 7.523417949676514, 'timing/data': 0.020177658647298813, 'timing/model': 4.624306321144104, 'learning_rate/qwen_vl_interface': 8.474700548489255e-06, 'learning_rate/action_model': 8.474700548489255e-05, 'epoch': 0.74}) +[2026-06-17 20:17:26,883][starVLA.training.train_starvla][INFO] - Step 1912, Loss: {'train/loss': 0.7247912883758545, 'train/grad_norm_pre_clip': 10.5831298828125, 'timing/data': 0.01292090117931366, 'timing/model': 4.491061289794743, 'learning_rate/qwen_vl_interface': 8.473112269028108e-06, 'learning_rate/action_model': 8.473112269028107e-05, 'epoch': 0.74}) +[2026-06-17 20:17:31,546][starVLA.training.train_starvla][INFO] - Step 1913, Loss: {'train/loss': 0.7804433703422546, 'train/grad_norm_pre_clip': 9.723548889160156, 'timing/data': 0.011591702699661255, 'timing/model': 1.0877263424918056, 'learning_rate/qwen_vl_interface': 8.47152332141232e-06, 'learning_rate/action_model': 8.47152332141232e-05, 'epoch': 0.74}) +[2026-06-17 20:17:36,557][starVLA.training.train_starvla][INFO] - Step 1914, Loss: {'train/loss': 0.3711903691291809, 'train/grad_norm_pre_clip': 7.884764671325684, 'timing/data': 0.012001613155007362, 'timing/model': 1.1646465444937348, 'learning_rate/qwen_vl_interface': 8.469933705971282e-06, 'learning_rate/action_model': 8.469933705971282e-05, 'epoch': 0.74}) +[2026-06-17 20:17:41,607][starVLA.training.train_starvla][INFO] - Step 1915, Loss: {'train/loss': 0.9874204993247986, 'train/grad_norm_pre_clip': 11.617074012756348, 'timing/data': 0.011578363366425037, 'timing/model': 1.2057383647188544, 'learning_rate/qwen_vl_interface': 8.468343423034521e-06, 'learning_rate/action_model': 8.46834342303452e-05, 'epoch': 0.74}) +[2026-06-17 20:17:46,318][starVLA.training.train_starvla][INFO] - Step 1916, Loss: {'train/loss': 0.3495030105113983, 'train/grad_norm_pre_clip': 5.295301914215088, 'timing/data': 0.012013510800898075, 'timing/model': 1.0692717405036092, 'learning_rate/qwen_vl_interface': 8.466752472931704e-06, 'learning_rate/action_model': 8.466752472931704e-05, 'epoch': 0.74}) +[2026-06-17 20:17:50,851][starVLA.training.train_starvla][INFO] - Step 1917, Loss: {'train/loss': 0.2737460136413574, 'train/grad_norm_pre_clip': 7.281381130218506, 'timing/data': 0.013100383803248405, 'timing/model': 1.0849101208150387, 'learning_rate/qwen_vl_interface': 8.465160855992639e-06, 'learning_rate/action_model': 8.465160855992638e-05, 'epoch': 0.74}) +[2026-06-17 20:17:55,923][starVLA.training.train_starvla][INFO] - Step 1918, Loss: {'train/loss': 0.507645308971405, 'train/grad_norm_pre_clip': 9.526594161987305, 'timing/data': 0.012752904556691647, 'timing/model': 1.3039071504026651, 'learning_rate/qwen_vl_interface': 8.46356857254727e-06, 'learning_rate/action_model': 8.46356857254727e-05, 'epoch': 0.74}) +[2026-06-17 20:18:00,681][starVLA.training.train_starvla][INFO] - Step 1919, Loss: {'train/loss': 0.21461641788482666, 'train/grad_norm_pre_clip': 4.486361980438232, 'timing/data': 0.01475523877888918, 'timing/model': 1.1556009016931057, 'learning_rate/qwen_vl_interface': 8.461975622925676e-06, 'learning_rate/action_model': 8.461975622925676e-05, 'epoch': 0.74}) +[2026-06-17 20:18:05,759][starVLA.training.train_starvla][INFO] - Step 1920, Loss: {'train/loss': 0.5989916324615479, 'train/grad_norm_pre_clip': 7.468070983886719, 'timing/data': 0.015576515346765518, 'timing/model': 1.2783340672031045, 'learning_rate/qwen_vl_interface': 8.460382007458081e-06, 'learning_rate/action_model': 8.460382007458081e-05, 'epoch': 0.74}) +[2026-06-17 20:18:10,623][starVLA.training.train_starvla][INFO] - Step 1921, Loss: {'train/loss': 0.41113966703414917, 'train/grad_norm_pre_clip': 5.876315116882324, 'timing/data': 0.016183954663574696, 'timing/model': 1.491847800090909, 'learning_rate/qwen_vl_interface': 8.45878772647484e-06, 'learning_rate/action_model': 8.458787726474839e-05, 'epoch': 0.74}) +[2026-06-17 20:18:17,351][starVLA.training.train_starvla][INFO] - Step 1922, Loss: {'train/loss': 0.4839652478694916, 'train/grad_norm_pre_clip': 7.936248302459717, 'timing/data': 0.01231073122471571, 'timing/model': 1.4374223565682769, 'learning_rate/qwen_vl_interface': 8.457192780306452e-06, 'learning_rate/action_model': 8.457192780306451e-05, 'epoch': 0.74}) +[2026-06-17 20:18:21,857][starVLA.training.train_starvla][INFO] - Step 1923, Loss: {'train/loss': 0.6837880611419678, 'train/grad_norm_pre_clip': 9.205451965332031, 'timing/data': 0.014264324679970741, 'timing/model': 1.1255968119949102, 'learning_rate/qwen_vl_interface': 8.455597169283549e-06, 'learning_rate/action_model': 8.455597169283548e-05, 'epoch': 0.74}) +[2026-06-17 20:18:26,895][starVLA.training.train_starvla][INFO] - Step 1924, Loss: {'train/loss': 0.4402933418750763, 'train/grad_norm_pre_clip': 9.015189170837402, 'timing/data': 0.014434580691158772, 'timing/model': 1.3087859358638525, 'learning_rate/qwen_vl_interface': 8.454000893736901e-06, 'learning_rate/action_model': 8.4540008937369e-05, 'epoch': 0.74}) +[2026-06-17 20:18:33,925][starVLA.training.train_starvla][INFO] - Step 1925, Loss: {'train/loss': 0.7784429788589478, 'train/grad_norm_pre_clip': 8.473453521728516, 'timing/data': 0.014433753676712513, 'timing/model': 1.3274979926645756, 'learning_rate/qwen_vl_interface': 8.452403953997422e-06, 'learning_rate/action_model': 8.452403953997422e-05, 'epoch': 0.74}) +[2026-06-17 20:18:38,874][starVLA.training.train_starvla][INFO] - Step 1926, Loss: {'train/loss': 0.3153972029685974, 'train/grad_norm_pre_clip': 5.698570251464844, 'timing/data': 0.012897259555757046, 'timing/model': 1.1503130057826638, 'learning_rate/qwen_vl_interface': 8.450806350396158e-06, 'learning_rate/action_model': 8.450806350396158e-05, 'epoch': 0.74}) +[2026-06-17 20:18:43,707][starVLA.training.train_starvla][INFO] - Step 1927, Loss: {'train/loss': 0.8291165828704834, 'train/grad_norm_pre_clip': 8.838205337524414, 'timing/data': 0.014502448961138725, 'timing/model': 1.5506535470485687, 'learning_rate/qwen_vl_interface': 8.449208083264291e-06, 'learning_rate/action_model': 8.449208083264292e-05, 'epoch': 0.74}) +[2026-06-17 20:18:53,339][starVLA.training.train_starvla][INFO] - Step 1928, Loss: {'train/loss': 0.8153195381164551, 'train/grad_norm_pre_clip': 12.55144214630127, 'timing/data': 0.016245068050920963, 'timing/model': 1.108774121850729, 'learning_rate/qwen_vl_interface': 8.447609152933149e-06, 'learning_rate/action_model': 8.447609152933149e-05, 'epoch': 0.74}) +[2026-06-17 20:18:58,043][starVLA.training.train_starvla][INFO] - Step 1929, Loss: {'train/loss': 0.4274410009384155, 'train/grad_norm_pre_clip': 7.397317409515381, 'timing/data': 0.012462442740797997, 'timing/model': 1.12154287006706, 'learning_rate/qwen_vl_interface': 8.44600955973419e-06, 'learning_rate/action_model': 8.446009559734188e-05, 'epoch': 0.74}) +[2026-06-17 20:19:02,844][starVLA.training.train_starvla][INFO] - Step 1930, Loss: {'train/loss': 0.5368013978004456, 'train/grad_norm_pre_clip': 7.9153642654418945, 'timing/data': 0.019774905405938625, 'timing/model': 1.2134946379810572, 'learning_rate/qwen_vl_interface': 8.444409303999005e-06, 'learning_rate/action_model': 8.444409303999005e-05, 'epoch': 0.74}) +[2026-06-17 20:19:07,573][starVLA.training.train_starvla][INFO] - Step 1931, Loss: {'train/loss': 0.3032350242137909, 'train/grad_norm_pre_clip': 6.011130332946777, 'timing/data': 0.013000013306736946, 'timing/model': 1.1727881664410233, 'learning_rate/qwen_vl_interface': 8.442808386059337e-06, 'learning_rate/action_model': 8.442808386059337e-05, 'epoch': 0.74}) +[2026-06-17 20:19:12,239][starVLA.training.train_starvla][INFO] - Step 1932, Loss: {'train/loss': 0.36112919449806213, 'train/grad_norm_pre_clip': 9.720519065856934, 'timing/data': 0.01430888008326292, 'timing/model': 1.4325243951752782, 'learning_rate/qwen_vl_interface': 8.441206806247055e-06, 'learning_rate/action_model': 8.441206806247054e-05, 'epoch': 0.74}) +[2026-06-17 20:19:18,331][starVLA.training.train_starvla][INFO] - Step 1933, Loss: {'train/loss': 0.5076681971549988, 'train/grad_norm_pre_clip': 8.607721328735352, 'timing/data': 0.013025599531829357, 'timing/model': 1.2283972520381212, 'learning_rate/qwen_vl_interface': 8.439604564894167e-06, 'learning_rate/action_model': 8.439604564894167e-05, 'epoch': 0.74}) +[2026-06-17 20:19:23,270][starVLA.training.train_starvla][INFO] - Step 1934, Loss: {'train/loss': 0.5108685493469238, 'train/grad_norm_pre_clip': 7.331393241882324, 'timing/data': 0.014015406370162964, 'timing/model': 1.0897167026996613, 'learning_rate/qwen_vl_interface': 8.438001662332821e-06, 'learning_rate/action_model': 8.43800166233282e-05, 'epoch': 0.74}) +[2026-06-17 20:19:28,240][starVLA.training.train_starvla][INFO] - Step 1935, Loss: {'train/loss': 0.4855423867702484, 'train/grad_norm_pre_clip': 8.291181564331055, 'timing/data': 0.012284908443689346, 'timing/model': 1.091887858696282, 'learning_rate/qwen_vl_interface': 8.4363980988953e-06, 'learning_rate/action_model': 8.436398098895301e-05, 'epoch': 0.74}) +[2026-06-17 20:19:33,012][starVLA.training.train_starvla][INFO] - Step 1936, Loss: {'train/loss': 0.82697993516922, 'train/grad_norm_pre_clip': 15.35585880279541, 'timing/data': 0.012318022549152374, 'timing/model': 1.0779885202646255, 'learning_rate/qwen_vl_interface': 8.434793874914026e-06, 'learning_rate/action_model': 8.434793874914026e-05, 'epoch': 0.74}) +[2026-06-17 20:19:40,498][starVLA.training.train_starvla][INFO] - Step 1937, Loss: {'train/loss': 0.43908676505088806, 'train/grad_norm_pre_clip': 6.906094074249268, 'timing/data': 0.014165055006742477, 'timing/model': 1.4508344763889909, 'learning_rate/qwen_vl_interface': 8.433188990721553e-06, 'learning_rate/action_model': 8.433188990721552e-05, 'epoch': 0.75}) +[2026-06-17 20:19:44,980][starVLA.training.train_starvla][INFO] - Step 1938, Loss: {'train/loss': 0.2892262935638428, 'train/grad_norm_pre_clip': 5.535455703735352, 'timing/data': 0.013096065260469913, 'timing/model': 1.1231855917721987, 'learning_rate/qwen_vl_interface': 8.431583446650577e-06, 'learning_rate/action_model': 8.431583446650578e-05, 'epoch': 0.75}) +[2026-06-17 20:19:50,043][starVLA.training.train_starvla][INFO] - Step 1939, Loss: {'train/loss': 0.6884941458702087, 'train/grad_norm_pre_clip': 29.879926681518555, 'timing/data': 0.011319190263748169, 'timing/model': 1.1623276034370065, 'learning_rate/qwen_vl_interface': 8.429977243033928e-06, 'learning_rate/action_model': 8.429977243033928e-05, 'epoch': 0.75}) +[2026-06-17 20:19:55,028][starVLA.training.train_starvla][INFO] - Step 1940, Loss: {'train/loss': 0.8067368865013123, 'train/grad_norm_pre_clip': 10.084861755371094, 'timing/data': 0.021667277440428734, 'timing/model': 1.198367410339415, 'learning_rate/qwen_vl_interface': 8.428370380204577e-06, 'learning_rate/action_model': 8.428370380204577e-05, 'epoch': 0.75}) +[2026-06-17 20:19:59,869][starVLA.training.train_starvla][INFO] - Step 1941, Loss: {'train/loss': 0.596635103225708, 'train/grad_norm_pre_clip': 9.209199905395508, 'timing/data': 0.017804211005568504, 'timing/model': 1.1583973346278071, 'learning_rate/qwen_vl_interface': 8.426762858495625e-06, 'learning_rate/action_model': 8.426762858495625e-05, 'epoch': 0.75}) +[2026-06-17 20:20:04,745][starVLA.training.train_starvla][INFO] - Step 1942, Loss: {'train/loss': 0.3494984209537506, 'train/grad_norm_pre_clip': 9.2785062789917, 'timing/data': 0.015990477986633778, 'timing/model': 1.392873122356832, 'learning_rate/qwen_vl_interface': 8.425154678240315e-06, 'learning_rate/action_model': 8.425154678240314e-05, 'epoch': 0.75}) +[2026-06-17 20:20:22,058][starVLA.training.train_starvla][INFO] - Step 1943, Loss: {'train/loss': 0.6400882005691528, 'train/grad_norm_pre_clip': 10.011402130126953, 'timing/data': 0.013363075442612171, 'timing/model': 5.278261313214898, 'learning_rate/qwen_vl_interface': 8.423545839772021e-06, 'learning_rate/action_model': 8.423545839772021e-05, 'epoch': 0.75}) +[2026-06-17 20:20:26,976][starVLA.training.train_starvla][INFO] - Step 1944, Loss: {'train/loss': 0.30081838369369507, 'train/grad_norm_pre_clip': 4.4600677490234375, 'timing/data': 0.011559690348803997, 'timing/model': 1.152136566117406, 'learning_rate/qwen_vl_interface': 8.421936343424264e-06, 'learning_rate/action_model': 8.421936343424262e-05, 'epoch': 0.75}) +[2026-06-17 20:20:31,744][starVLA.training.train_starvla][INFO] - Step 1945, Loss: {'train/loss': 0.6648479700088501, 'train/grad_norm_pre_clip': 12.081127166748047, 'timing/data': 0.015933061949908733, 'timing/model': 1.0845454866066575, 'learning_rate/qwen_vl_interface': 8.420326189530685e-06, 'learning_rate/action_model': 8.420326189530685e-05, 'epoch': 0.75}) +[2026-06-17 20:20:36,539][starVLA.training.train_starvla][INFO] - Step 1946, Loss: {'train/loss': 0.7983052730560303, 'train/grad_norm_pre_clip': 13.650150299072266, 'timing/data': 0.011602463200688362, 'timing/model': 1.223449352197349, 'learning_rate/qwen_vl_interface': 8.418715378425078e-06, 'learning_rate/action_model': 8.418715378425078e-05, 'epoch': 0.75}) +[2026-06-17 20:20:41,574][starVLA.training.train_starvla][INFO] - Step 1947, Loss: {'train/loss': 0.34870314598083496, 'train/grad_norm_pre_clip': 6.696439743041992, 'timing/data': 0.014021316543221474, 'timing/model': 1.551345077343285, 'learning_rate/qwen_vl_interface': 8.417103910441364e-06, 'learning_rate/action_model': 8.417103910441363e-05, 'epoch': 0.75}) +[2026-06-17 20:20:46,967][starVLA.training.train_starvla][INFO] - Step 1948, Loss: {'train/loss': 0.7703242301940918, 'train/grad_norm_pre_clip': 9.829742431640625, 'timing/data': 0.01424874272197485, 'timing/model': 2.061489342711866, 'learning_rate/qwen_vl_interface': 8.415491785913599e-06, 'learning_rate/action_model': 8.4154917859136e-05, 'epoch': 0.75}) +[2026-06-17 20:20:51,604][starVLA.training.train_starvla][INFO] - Step 1949, Loss: {'train/loss': 0.7741551399230957, 'train/grad_norm_pre_clip': 10.685264587402344, 'timing/data': 0.012054416351020336, 'timing/model': 1.08282998111099, 'learning_rate/qwen_vl_interface': 8.413879005175983e-06, 'learning_rate/action_model': 8.413879005175983e-05, 'epoch': 0.75}) +[2026-06-17 20:20:56,447][starVLA.training.train_starvla][INFO] - Step 1950, Loss: {'train/loss': 0.6175352931022644, 'train/grad_norm_pre_clip': 6.52786111831665, 'timing/data': 0.012422525323927402, 'timing/model': 1.1733451765030622, 'learning_rate/qwen_vl_interface': 8.412265568562843e-06, 'learning_rate/action_model': 8.412265568562842e-05, 'epoch': 0.75}) +[2026-06-17 20:21:01,097][starVLA.training.train_starvla][INFO] - Step 1951, Loss: {'train/loss': 0.7810317873954773, 'train/grad_norm_pre_clip': 7.8544206619262695, 'timing/data': 0.014271264895796776, 'timing/model': 1.174835597164929, 'learning_rate/qwen_vl_interface': 8.41065147640865e-06, 'learning_rate/action_model': 8.41065147640865e-05, 'epoch': 0.75}) +[2026-06-17 20:21:09,244][starVLA.training.train_starvla][INFO] - Step 1952, Loss: {'train/loss': 0.3343415856361389, 'train/grad_norm_pre_clip': 4.831442356109619, 'timing/data': 0.012397228740155697, 'timing/model': 4.843669490888715, 'learning_rate/qwen_vl_interface': 8.409036729048002e-06, 'learning_rate/action_model': 8.409036729048002e-05, 'epoch': 0.75}) +[2026-06-17 20:21:14,118][starVLA.training.train_starvla][INFO] - Step 1953, Loss: {'train/loss': 0.4969015121459961, 'train/grad_norm_pre_clip': 9.694079399108887, 'timing/data': 0.012095068581402302, 'timing/model': 1.3537817047908902, 'learning_rate/qwen_vl_interface': 8.407421326815644e-06, 'learning_rate/action_model': 8.407421326815644e-05, 'epoch': 0.75}) +[2026-06-17 20:21:24,810][starVLA.training.train_starvla][INFO] - Step 1954, Loss: {'train/loss': 0.6158897280693054, 'train/grad_norm_pre_clip': 7.303534507751465, 'timing/data': 0.018494032323360443, 'timing/model': 3.9006116399541497, 'learning_rate/qwen_vl_interface': 8.405805270046445e-06, 'learning_rate/action_model': 8.405805270046445e-05, 'epoch': 0.75}) +[2026-06-17 20:21:29,962][starVLA.training.train_starvla][INFO] - Step 1955, Loss: {'train/loss': 0.24463604390621185, 'train/grad_norm_pre_clip': 9.210390090942383, 'timing/data': 0.01322239264845848, 'timing/model': 1.0898280953988433, 'learning_rate/qwen_vl_interface': 8.404188559075419e-06, 'learning_rate/action_model': 8.404188559075419e-05, 'epoch': 0.75}) +[2026-06-17 20:21:34,982][starVLA.training.train_starvla][INFO] - Step 1956, Loss: {'train/loss': 0.5736700296401978, 'train/grad_norm_pre_clip': 12.651470184326172, 'timing/data': 0.012630614452064037, 'timing/model': 1.0848243236541748, 'learning_rate/qwen_vl_interface': 8.40257119423771e-06, 'learning_rate/action_model': 8.402571194237709e-05, 'epoch': 0.75}) +[2026-06-17 20:21:39,679][starVLA.training.train_starvla][INFO] - Step 1957, Loss: {'train/loss': 0.607518196105957, 'train/grad_norm_pre_clip': 8.74048900604248, 'timing/data': 0.01286831684410572, 'timing/model': 1.4359072521328926, 'learning_rate/qwen_vl_interface': 8.4009531758686e-06, 'learning_rate/action_model': 8.4009531758686e-05, 'epoch': 0.75}) +[2026-06-17 20:21:45,811][starVLA.training.train_starvla][INFO] - Step 1958, Loss: {'train/loss': 0.2896403670310974, 'train/grad_norm_pre_clip': 7.069911479949951, 'timing/data': 0.012057337909936905, 'timing/model': 1.110887998715043, 'learning_rate/qwen_vl_interface': 8.399334504303507e-06, 'learning_rate/action_model': 8.399334504303507e-05, 'epoch': 0.75}) +[2026-06-17 20:21:54,544][starVLA.training.train_starvla][INFO] - Step 1959, Loss: {'train/loss': 0.5963709950447083, 'train/grad_norm_pre_clip': 9.22524642944336, 'timing/data': 0.013964680954813957, 'timing/model': 3.2489667600020766, 'learning_rate/qwen_vl_interface': 8.397715179877982e-06, 'learning_rate/action_model': 8.397715179877982e-05, 'epoch': 0.75}) +[2026-06-17 20:22:01,156][starVLA.training.train_starvla][INFO] - Step 1960, Loss: {'train/loss': 0.5174689888954163, 'train/grad_norm_pre_clip': 9.38746452331543, 'timing/data': 0.012300245463848114, 'timing/model': 1.1213711705058813, 'learning_rate/qwen_vl_interface': 8.396095202927712e-06, 'learning_rate/action_model': 8.396095202927712e-05, 'epoch': 0.75}) +[2026-06-17 20:22:06,138][starVLA.training.train_starvla][INFO] - Step 1961, Loss: {'train/loss': 0.6360897421836853, 'train/grad_norm_pre_clip': 12.418310165405273, 'timing/data': 0.015740486793220043, 'timing/model': 1.36330974008888, 'learning_rate/qwen_vl_interface': 8.394474573788522e-06, 'learning_rate/action_model': 8.394474573788522e-05, 'epoch': 0.75}) +[2026-06-17 20:22:11,250][starVLA.training.train_starvla][INFO] - Step 1962, Loss: {'train/loss': 0.37009134888648987, 'train/grad_norm_pre_clip': 8.191545486450195, 'timing/data': 0.013957947492599487, 'timing/model': 1.6799232941120863, 'learning_rate/qwen_vl_interface': 8.39285329279637e-06, 'learning_rate/action_model': 8.39285329279637e-05, 'epoch': 0.75}) +[2026-06-17 20:22:15,871][starVLA.training.train_starvla][INFO] - Step 1963, Loss: {'train/loss': 0.37375712394714355, 'train/grad_norm_pre_clip': 6.417374134063721, 'timing/data': 0.014155952259898186, 'timing/model': 1.1800338001921773, 'learning_rate/qwen_vl_interface': 8.391231360287348e-06, 'learning_rate/action_model': 8.391231360287347e-05, 'epoch': 0.76}) +[2026-06-17 20:22:20,947][starVLA.training.train_starvla][INFO] - Step 1964, Loss: {'train/loss': 0.749180793762207, 'train/grad_norm_pre_clip': 10.550772666931152, 'timing/data': 0.013846395537257195, 'timing/model': 1.1449482832103968, 'learning_rate/qwen_vl_interface': 8.389608776597684e-06, 'learning_rate/action_model': 8.389608776597684e-05, 'epoch': 0.76}) +[2026-06-17 20:22:26,039][starVLA.training.train_starvla][INFO] - Step 1965, Loss: {'train/loss': 0.6485049724578857, 'train/grad_norm_pre_clip': 8.363068580627441, 'timing/data': 0.014306452125310898, 'timing/model': 1.347496708855033, 'learning_rate/qwen_vl_interface': 8.387985542063743e-06, 'learning_rate/action_model': 8.387985542063743e-05, 'epoch': 0.76}) +[2026-06-17 20:22:30,944][starVLA.training.train_starvla][INFO] - Step 1966, Loss: {'train/loss': 0.44771242141723633, 'train/grad_norm_pre_clip': 10.155677795410156, 'timing/data': 0.013191736303269863, 'timing/model': 1.2740618884563446, 'learning_rate/qwen_vl_interface': 8.386361657022024e-06, 'learning_rate/action_model': 8.386361657022023e-05, 'epoch': 0.76}) +[2026-06-17 20:22:40,068][starVLA.training.train_starvla][INFO] - Step 1967, Loss: {'train/loss': 0.2965407371520996, 'train/grad_norm_pre_clip': 5.38007116317749, 'timing/data': 0.019665073603391647, 'timing/model': 1.6885517034679651, 'learning_rate/qwen_vl_interface': 8.384737121809156e-06, 'learning_rate/action_model': 8.384737121809156e-05, 'epoch': 0.76}) +[2026-06-17 20:22:51,321][starVLA.training.train_starvla][INFO] - Step 1968, Loss: {'train/loss': 0.683593213558197, 'train/grad_norm_pre_clip': 9.24608039855957, 'timing/data': 0.03267286904156208, 'timing/model': 1.5119337905198336, 'learning_rate/qwen_vl_interface': 8.38311193676191e-06, 'learning_rate/action_model': 8.383111936761909e-05, 'epoch': 0.76}) +[2026-06-17 20:23:03,531][starVLA.training.train_starvla][INFO] - Step 1969, Loss: {'train/loss': 0.569222629070282, 'train/grad_norm_pre_clip': 13.553289413452148, 'timing/data': 0.011534121818840504, 'timing/model': 2.210750015452504, 'learning_rate/qwen_vl_interface': 8.381486102217188e-06, 'learning_rate/action_model': 8.381486102217188e-05, 'epoch': 0.76}) +[2026-06-17 20:23:18,908][starVLA.training.train_starvla][INFO] - Step 1970, Loss: {'train/loss': 0.610952615737915, 'train/grad_norm_pre_clip': 16.131149291992188, 'timing/data': 0.011946268379688263, 'timing/model': 1.2581578334793448, 'learning_rate/qwen_vl_interface': 8.379859618512029e-06, 'learning_rate/action_model': 8.379859618512029e-05, 'epoch': 0.76}) +[2026-06-17 20:23:23,911][starVLA.training.train_starvla][INFO] - Step 1971, Loss: {'train/loss': 0.5626742839813232, 'train/grad_norm_pre_clip': 8.527745246887207, 'timing/data': 0.012868971563875675, 'timing/model': 1.3023156803101301, 'learning_rate/qwen_vl_interface': 8.378232485983603e-06, 'learning_rate/action_model': 8.378232485983603e-05, 'epoch': 0.76}) +[2026-06-17 20:23:41,860][starVLA.training.train_starvla][INFO] - Step 1972, Loss: {'train/loss': 0.5932306051254272, 'train/grad_norm_pre_clip': 7.732630729675293, 'timing/data': 0.014246223494410515, 'timing/model': 5.616907589137554, 'learning_rate/qwen_vl_interface': 8.376604704969216e-06, 'learning_rate/action_model': 8.376604704969216e-05, 'epoch': 0.76}) +[2026-06-17 20:23:46,583][starVLA.training.train_starvla][INFO] - Step 1973, Loss: {'train/loss': 0.4829427897930145, 'train/grad_norm_pre_clip': 8.485590934753418, 'timing/data': 0.021670651622116566, 'timing/model': 1.2641714680939913, 'learning_rate/qwen_vl_interface': 8.374976275806307e-06, 'learning_rate/action_model': 8.374976275806308e-05, 'epoch': 0.76}) +[2026-06-17 20:23:51,860][starVLA.training.train_starvla][INFO] - Step 1974, Loss: {'train/loss': 0.4845750331878662, 'train/grad_norm_pre_clip': 16.802810668945312, 'timing/data': 0.01672982331365347, 'timing/model': 1.371483089402318, 'learning_rate/qwen_vl_interface': 8.373347198832457e-06, 'learning_rate/action_model': 8.373347198832457e-05, 'epoch': 0.76}) +[2026-06-17 20:23:56,628][starVLA.training.train_starvla][INFO] - Step 1975, Loss: {'train/loss': 0.5033227205276489, 'train/grad_norm_pre_clip': 10.43227481842041, 'timing/data': 0.011707376688718796, 'timing/model': 1.0702766412869096, 'learning_rate/qwen_vl_interface': 8.37171747438537e-06, 'learning_rate/action_model': 8.37171747438537e-05, 'epoch': 0.76}) +[2026-06-17 20:24:01,506][starVLA.training.train_starvla][INFO] - Step 1976, Loss: {'train/loss': 0.5027855038642883, 'train/grad_norm_pre_clip': 9.063477516174316, 'timing/data': 0.0162313561886549, 'timing/model': 1.2284756368026137, 'learning_rate/qwen_vl_interface': 8.370087102802892e-06, 'learning_rate/action_model': 8.370087102802892e-05, 'epoch': 0.76}) +[2026-06-17 20:24:06,585][starVLA.training.train_starvla][INFO] - Step 1977, Loss: {'train/loss': 0.5461220741271973, 'train/grad_norm_pre_clip': 10.553780555725098, 'timing/data': 0.014043220318853855, 'timing/model': 1.7998125310987234, 'learning_rate/qwen_vl_interface': 8.368456084423e-06, 'learning_rate/action_model': 8.368456084422999e-05, 'epoch': 0.76}) +[2026-06-17 20:24:11,462][starVLA.training.train_starvla][INFO] - Step 1978, Loss: {'train/loss': 0.22498004138469696, 'train/grad_norm_pre_clip': 23.69219207763672, 'timing/data': 0.012203390710055828, 'timing/model': 1.2912903614342213, 'learning_rate/qwen_vl_interface': 8.366824419583806e-06, 'learning_rate/action_model': 8.366824419583806e-05, 'epoch': 0.76}) +[2026-06-17 20:24:16,142][starVLA.training.train_starvla][INFO] - Step 1979, Loss: {'train/loss': 0.8910015225410461, 'train/grad_norm_pre_clip': 10.35102367401123, 'timing/data': 0.016861719079315662, 'timing/model': 1.0898717679083347, 'learning_rate/qwen_vl_interface': 8.365192108623555e-06, 'learning_rate/action_model': 8.365192108623555e-05, 'epoch': 0.76}) +[2026-06-17 20:24:20,937][starVLA.training.train_starvla][INFO] - Step 1980, Loss: {'train/loss': 0.5574886202812195, 'train/grad_norm_pre_clip': 8.755488395690918, 'timing/data': 0.015907198190689087, 'timing/model': 1.179855303838849, 'learning_rate/qwen_vl_interface': 8.363559151880631e-06, 'learning_rate/action_model': 8.36355915188063e-05, 'epoch': 0.76}) +[2026-06-17 20:24:25,916][starVLA.training.train_starvla][INFO] - Step 1981, Loss: {'train/loss': 0.27632787823677063, 'train/grad_norm_pre_clip': 5.833693504333496, 'timing/data': 0.021762876771390438, 'timing/model': 1.4412244288250804, 'learning_rate/qwen_vl_interface': 8.361925549693541e-06, 'learning_rate/action_model': 8.36192554969354e-05, 'epoch': 0.76}) +[2026-06-17 20:24:30,781][starVLA.training.train_starvla][INFO] - Step 1982, Loss: {'train/loss': 0.8321988582611084, 'train/grad_norm_pre_clip': 10.349682807922363, 'timing/data': 0.012775578536093235, 'timing/model': 1.564298060722649, 'learning_rate/qwen_vl_interface': 8.360291302400937e-06, 'learning_rate/action_model': 8.360291302400937e-05, 'epoch': 0.76}) +[2026-06-17 20:24:35,204][starVLA.training.train_starvla][INFO] - Step 1983, Loss: {'train/loss': 0.40776389837265015, 'train/grad_norm_pre_clip': 8.21562385559082, 'timing/data': 0.01881830021739006, 'timing/model': 1.1986092869192362, 'learning_rate/qwen_vl_interface': 8.3586564103416e-06, 'learning_rate/action_model': 8.358656410341599e-05, 'epoch': 0.76}) +[2026-06-17 20:24:39,794][starVLA.training.train_starvla][INFO] - Step 1984, Loss: {'train/loss': 0.4753216803073883, 'train/grad_norm_pre_clip': 6.975966930389404, 'timing/data': 0.011734032072126865, 'timing/model': 1.0667493995279074, 'learning_rate/qwen_vl_interface': 8.357020873854445e-06, 'learning_rate/action_model': 8.357020873854444e-05, 'epoch': 0.76}) +[2026-06-17 20:24:44,705][starVLA.training.train_starvla][INFO] - Step 1985, Loss: {'train/loss': 0.5608844757080078, 'train/grad_norm_pre_clip': 7.930985927581787, 'timing/data': 0.012589161284267902, 'timing/model': 1.3262564800679684, 'learning_rate/qwen_vl_interface': 8.355384693278516e-06, 'learning_rate/action_model': 8.355384693278517e-05, 'epoch': 0.76}) +[2026-06-17 20:24:49,444][starVLA.training.train_starvla][INFO] - Step 1986, Loss: {'train/loss': 0.3913307785987854, 'train/grad_norm_pre_clip': 5.155281066894531, 'timing/data': 0.012298268266022205, 'timing/model': 1.1068858793005347, 'learning_rate/qwen_vl_interface': 8.353747868953002e-06, 'learning_rate/action_model': 8.353747868953003e-05, 'epoch': 0.76}) +[2026-06-17 20:24:54,222][starVLA.training.train_starvla][INFO] - Step 1987, Loss: {'train/loss': 0.4036113917827606, 'train/grad_norm_pre_clip': 4.988646984100342, 'timing/data': 0.013762025162577629, 'timing/model': 1.4402041686698794, 'learning_rate/qwen_vl_interface': 8.352110401217213e-06, 'learning_rate/action_model': 8.352110401217213e-05, 'epoch': 0.76}) +[2026-06-17 20:25:01,339][starVLA.training.train_starvla][INFO] - Step 1988, Loss: {'train/loss': 0.6214268207550049, 'train/grad_norm_pre_clip': 8.415157318115234, 'timing/data': 0.01202982384711504, 'timing/model': 2.669223388656974, 'learning_rate/qwen_vl_interface': 8.350472290410602e-06, 'learning_rate/action_model': 8.350472290410602e-05, 'epoch': 0.76}) +[2026-06-17 20:25:08,403][starVLA.training.train_starvla][INFO] - Step 1989, Loss: {'train/loss': 0.6978294849395752, 'train/grad_norm_pre_clip': 9.398755073547363, 'timing/data': 0.017508641816675663, 'timing/model': 1.4311305461451411, 'learning_rate/qwen_vl_interface': 8.348833536872747e-06, 'learning_rate/action_model': 8.348833536872747e-05, 'epoch': 0.77}) +[2026-06-17 20:25:16,653][starVLA.training.train_starvla][INFO] - Step 1990, Loss: {'train/loss': 0.6188610196113586, 'train/grad_norm_pre_clip': 7.389952659606934, 'timing/data': 0.016012626700103283, 'timing/model': 1.561622099019587, 'learning_rate/qwen_vl_interface': 8.347194140943366e-06, 'learning_rate/action_model': 8.347194140943366e-05, 'epoch': 0.77}) +[2026-06-17 20:25:29,313][starVLA.training.train_starvla][INFO] - Step 1991, Loss: {'train/loss': 0.6349951028823853, 'train/grad_norm_pre_clip': 9.892035484313965, 'timing/data': 0.014483187347650528, 'timing/model': 1.46641028765589, 'learning_rate/qwen_vl_interface': 8.345554102962306e-06, 'learning_rate/action_model': 8.345554102962305e-05, 'epoch': 0.77}) +[2026-06-17 20:25:34,503][starVLA.training.train_starvla][INFO] - Step 1992, Loss: {'train/loss': 0.3344044089317322, 'train/grad_norm_pre_clip': 7.729795932769775, 'timing/data': 0.014089291915297508, 'timing/model': 1.5362894013524055, 'learning_rate/qwen_vl_interface': 8.34391342326955e-06, 'learning_rate/action_model': 8.343913423269549e-05, 'epoch': 0.77}) +[2026-06-17 20:25:46,147][starVLA.training.train_starvla][INFO] - Step 1993, Loss: {'train/loss': 0.3979277014732361, 'train/grad_norm_pre_clip': 7.700469970703125, 'timing/data': 0.0012206826359033585, 'timing/model': 1.837733373977244, 'learning_rate/qwen_vl_interface': 8.342272102205213e-06, 'learning_rate/action_model': 8.342272102205212e-05, 'epoch': 0.77}) +[2026-06-17 20:25:59,500][starVLA.training.train_starvla][INFO] - Step 1994, Loss: {'train/loss': 0.6865715384483337, 'train/grad_norm_pre_clip': 15.232747077941895, 'timing/data': 0.01391123328357935, 'timing/model': 1.4412966025993228, 'learning_rate/qwen_vl_interface': 8.340630140109539e-06, 'learning_rate/action_model': 8.340630140109539e-05, 'epoch': 0.77}) +[2026-06-17 20:26:05,750][starVLA.training.train_starvla][INFO] - Step 1995, Loss: {'train/loss': 0.5953344702720642, 'train/grad_norm_pre_clip': 10.783361434936523, 'timing/data': 0.014514305628836155, 'timing/model': 2.3608101652935147, 'learning_rate/qwen_vl_interface': 8.338987537322916e-06, 'learning_rate/action_model': 8.338987537322916e-05, 'epoch': 0.77}) +[2026-06-17 20:26:16,909][starVLA.training.train_starvla][INFO] - Step 1996, Loss: {'train/loss': 0.5795034170150757, 'train/grad_norm_pre_clip': 7.876189231872559, 'timing/data': 0.013498971238732338, 'timing/model': 1.4696030467748642, 'learning_rate/qwen_vl_interface': 8.337344294185848e-06, 'learning_rate/action_model': 8.337344294185849e-05, 'epoch': 0.77}) +[2026-06-17 20:26:21,222][starVLA.training.train_starvla][INFO] - Step 1997, Loss: {'train/loss': 0.9072121381759644, 'train/grad_norm_pre_clip': 9.221640586853027, 'timing/data': 0.011441245675086975, 'timing/model': 1.1210123421624303, 'learning_rate/qwen_vl_interface': 8.335700411038988e-06, 'learning_rate/action_model': 8.335700411038988e-05, 'epoch': 0.77}) +[2026-06-17 20:26:26,330][starVLA.training.train_starvla][INFO] - Step 1998, Loss: {'train/loss': 0.5496031641960144, 'train/grad_norm_pre_clip': 7.722846984863281, 'timing/data': 0.011954900808632374, 'timing/model': 1.3389252396300435, 'learning_rate/qwen_vl_interface': 8.334055888223111e-06, 'learning_rate/action_model': 8.334055888223111e-05, 'epoch': 0.77}) +[2026-06-17 20:26:31,389][starVLA.training.train_starvla][INFO] - Step 1999, Loss: {'train/loss': 0.6070945858955383, 'train/grad_norm_pre_clip': 8.941752433776855, 'timing/data': 0.014814374968409538, 'timing/model': 1.1835632594302297, 'learning_rate/qwen_vl_interface': 8.332410726079132e-06, 'learning_rate/action_model': 8.332410726079132e-05, 'epoch': 0.77}) +[2026-06-17 20:30:18,107][starVLA.training.train_starvla][INFO] - Step 2000, Loss: {'train/loss': 0.8861544132232666, 'train/grad_norm_pre_clip': 8.990437507629395, 'eval/loss': 0.5481570959091187, 'eval/latency_4/loss': 0.5481570959091187, 'eval/demon_fire_precision': 0.9533646322378717, 'eval/demon_fire_recall': 0.989282234491718, 'eval/demon_fire_f1': 0.9709913930506854, 'eval/demon_fire_rate_model': 0.9383259911894273, 'eval/demon_fire_rate_teacher': 0.9042584434654919, 'eval/demon_left_precision': 0.7054263565891473, 'eval/demon_left_recall': 0.6232876712328768, 'eval/demon_left_f1': 0.661818181818182, 'eval/demon_left_rate_model': 0.11365638766519824, 'eval/demon_left_rate_teacher': 0.12863436123348018, 'eval/demon_right_precision': 0.7561436672967864, 'eval/demon_right_recall': 0.7220216606498195, 'eval/demon_right_f1': 0.7386888273314867, 'eval/demon_right_rate_model': 0.155359765051395, 'eval/demon_right_rate_teacher': 0.16270190895741557, 'eval/demon_move_f1': 0.7002535045748344, 'eval/demon_cc_f1': 0.8700729236800135, 'eval/latency_4/demon_fire_precision': 0.9533646322378717, 'eval/latency_4/demon_fire_recall': 0.989282234491718, 'eval/latency_4/demon_fire_f1': 0.9709913930506854, 'eval/latency_4/demon_fire_rate_model': 0.9383259911894273, 'eval/latency_4/demon_fire_rate_teacher': 0.9042584434654919, 'eval/latency_4/demon_left_precision': 0.7054263565891473, 'eval/latency_4/demon_left_recall': 0.6232876712328768, 'eval/latency_4/demon_left_f1': 0.661818181818182, 'eval/latency_4/demon_left_rate_model': 0.11365638766519824, 'eval/latency_4/demon_left_rate_teacher': 0.12863436123348018, 'eval/latency_4/demon_right_precision': 0.7561436672967864, 'eval/latency_4/demon_right_recall': 0.7220216606498195, 'eval/latency_4/demon_right_f1': 0.7386888273314867, 'eval/latency_4/demon_right_rate_model': 0.155359765051395, 'eval/latency_4/demon_right_rate_teacher': 0.16270190895741557, 'eval/latency_4/demon_move_f1': 0.7002535045748344, 'eval/latency_4/demon_cc_f1': 0.8700729236800135, 'timing/data': 0.01141476258635521, 'timing/model': 1.1975203193724155, 'learning_rate/qwen_vl_interface': 8.330764924948093e-06, 'learning_rate/action_model': 8.330764924948093e-05, 'epoch': 0.77}) +[2026-06-17 20:30:18,133][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_2000_state +[2026-06-17 20:30:18,140][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 20:30:26,673][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_2000_state/model.safetensors +[2026-06-17 20:30:49,029][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_2000_state/optimizer.bin +[2026-06-17 20:30:49,031][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_2000_state/sampler.bin +[2026-06-17 20:30:49,031][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_2000_state/sampler_1.bin +[2026-06-17 20:30:49,034][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_2000_state/random_states_0.pkl +[2026-06-17 20:30:49,038][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2000 +[2026-06-17 20:30:49,038][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 20:30:49,055][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 20:30:54,224][starVLA.training.train_starvla][INFO] - Step 2001, Loss: {'train/loss': 0.34640583395957947, 'train/grad_norm_pre_clip': 6.3954176902771, 'timing/data': 0.011685460805892944, 'timing/model': 1.7067176941782236, 'learning_rate/qwen_vl_interface': 8.329118485171167e-06, 'learning_rate/action_model': 8.329118485171167e-05, 'epoch': 0.77}) +[2026-06-17 20:31:02,414][starVLA.training.train_starvla][INFO] - Step 2002, Loss: {'train/loss': 0.5350351929664612, 'train/grad_norm_pre_clip': 6.637075901031494, 'timing/data': 0.010411204770207405, 'timing/model': 1.4180900566279888, 'learning_rate/qwen_vl_interface': 8.327471407089667e-06, 'learning_rate/action_model': 8.327471407089668e-05, 'epoch': 0.77}) +[2026-06-17 20:31:07,793][starVLA.training.train_starvla][INFO] - Step 2003, Loss: {'train/loss': 0.9867076873779297, 'train/grad_norm_pre_clip': 26.001699447631836, 'timing/data': 0.011444581672549248, 'timing/model': 1.3404800854623318, 'learning_rate/qwen_vl_interface': 8.325823691045032e-06, 'learning_rate/action_model': 8.325823691045033e-05, 'epoch': 0.77}) +[2026-06-17 20:31:17,457][starVLA.training.train_starvla][INFO] - Step 2004, Loss: {'train/loss': 0.6341461539268494, 'train/grad_norm_pre_clip': 9.803888320922852, 'timing/data': 0.014348424971103668, 'timing/model': 1.263953355140984, 'learning_rate/qwen_vl_interface': 8.32417533737884e-06, 'learning_rate/action_model': 8.324175337378839e-05, 'epoch': 0.77}) +[2026-06-17 20:31:22,329][starVLA.training.train_starvla][INFO] - Step 2005, Loss: {'train/loss': 0.4195922911167145, 'train/grad_norm_pre_clip': 7.85297966003418, 'timing/data': 0.011163296177983284, 'timing/model': 1.0741718355566263, 'learning_rate/qwen_vl_interface': 8.322526346432789e-06, 'learning_rate/action_model': 8.322526346432789e-05, 'epoch': 0.77}) +[2026-06-17 20:31:32,666][starVLA.training.train_starvla][INFO] - Step 2006, Loss: {'train/loss': 0.9151108860969543, 'train/grad_norm_pre_clip': 10.41138744354248, 'timing/data': 0.010884732007980347, 'timing/model': 1.1726912194862962, 'learning_rate/qwen_vl_interface': 8.32087671854872e-06, 'learning_rate/action_model': 8.32087671854872e-05, 'epoch': 0.77}) +[2026-06-17 20:31:37,905][starVLA.training.train_starvla][INFO] - Step 2007, Loss: {'train/loss': 0.4508648216724396, 'train/grad_norm_pre_clip': 6.521083831787109, 'timing/data': 0.01176297664642334, 'timing/model': 1.8357651084661484, 'learning_rate/qwen_vl_interface': 8.319226454068604e-06, 'learning_rate/action_model': 8.319226454068604e-05, 'epoch': 0.77}) +[2026-06-17 20:31:42,764][starVLA.training.train_starvla][INFO] - Step 2008, Loss: {'train/loss': 0.8176841735839844, 'train/grad_norm_pre_clip': 11.569071769714355, 'timing/data': 0.011460226029157639, 'timing/model': 1.1847295360639691, 'learning_rate/qwen_vl_interface': 8.317575553334542e-06, 'learning_rate/action_model': 8.317575553334542e-05, 'epoch': 0.77}) +[2026-06-17 20:31:48,882][starVLA.training.train_starvla][INFO] - Step 2009, Loss: {'train/loss': 0.6481130719184875, 'train/grad_norm_pre_clip': 10.29520320892334, 'timing/data': 0.012072251178324223, 'timing/model': 2.266454574652016, 'learning_rate/qwen_vl_interface': 8.315924016688767e-06, 'learning_rate/action_model': 8.315924016688766e-05, 'epoch': 0.77}) +[2026-06-17 20:31:54,199][starVLA.training.train_starvla][INFO] - Step 2010, Loss: {'train/loss': 0.48345908522605896, 'train/grad_norm_pre_clip': 8.337225914001465, 'timing/data': 0.011812694370746613, 'timing/model': 1.319531500339508, 'learning_rate/qwen_vl_interface': 8.314271844473644e-06, 'learning_rate/action_model': 8.314271844473644e-05, 'epoch': 0.77}) +[2026-06-17 20:31:59,131][starVLA.training.train_starvla][INFO] - Step 2011, Loss: {'train/loss': 0.6106703877449036, 'train/grad_norm_pre_clip': 8.833023071289062, 'timing/data': 0.01134413480758667, 'timing/model': 1.0723898960277438, 'learning_rate/qwen_vl_interface': 8.312619037031668e-06, 'learning_rate/action_model': 8.312619037031668e-05, 'epoch': 0.77}) +[2026-06-17 20:32:04,605][starVLA.training.train_starvla][INFO] - Step 2012, Loss: {'train/loss': 0.28795701265335083, 'train/grad_norm_pre_clip': 5.350541114807129, 'timing/data': 0.011094938963651657, 'timing/model': 1.4346690541133285, 'learning_rate/qwen_vl_interface': 8.310965594705473e-06, 'learning_rate/action_model': 8.310965594705472e-05, 'epoch': 0.77}) +[2026-06-17 20:32:08,952][starVLA.training.train_starvla][INFO] - Step 2013, Loss: {'train/loss': 0.512565016746521, 'train/grad_norm_pre_clip': 6.753851890563965, 'timing/data': 0.013417080976068974, 'timing/model': 1.134992990642786, 'learning_rate/qwen_vl_interface': 8.309311517837814e-06, 'learning_rate/action_model': 8.309311517837813e-05, 'epoch': 0.77}) +[2026-06-17 20:32:13,731][starVLA.training.train_starvla][INFO] - Step 2014, Loss: {'train/loss': 0.8811278343200684, 'train/grad_norm_pre_clip': 9.5801420211792, 'timing/data': 0.014086716808378696, 'timing/model': 1.0889346953481436, 'learning_rate/qwen_vl_interface': 8.307656806771587e-06, 'learning_rate/action_model': 8.307656806771588e-05, 'epoch': 0.77}) +[2026-06-17 20:32:19,620][starVLA.training.train_starvla][INFO] - Step 2015, Loss: {'train/loss': 0.5879655480384827, 'train/grad_norm_pre_clip': 8.503386497497559, 'timing/data': 0.010935601778328419, 'timing/model': 2.153505638241768, 'learning_rate/qwen_vl_interface': 8.306001461849814e-06, 'learning_rate/action_model': 8.306001461849814e-05, 'epoch': 0.78}) +[2026-06-17 20:32:34,828][starVLA.training.train_starvla][INFO] - Step 2016, Loss: {'train/loss': 0.500626802444458, 'train/grad_norm_pre_clip': 7.604133605957031, 'timing/data': 0.012397371232509613, 'timing/model': 3.2901016185060143, 'learning_rate/qwen_vl_interface': 8.30434548341565e-06, 'learning_rate/action_model': 8.304345483415649e-05, 'epoch': 0.78}) +[2026-06-17 20:32:40,082][starVLA.training.train_starvla][INFO] - Step 2017, Loss: {'train/loss': 0.48083165287971497, 'train/grad_norm_pre_clip': 10.316184997558594, 'timing/data': 0.011237121187150478, 'timing/model': 1.543688385747373, 'learning_rate/qwen_vl_interface': 8.30268887181238e-06, 'learning_rate/action_model': 8.302688871812378e-05, 'epoch': 0.78}) +[2026-06-17 20:32:44,724][starVLA.training.train_starvla][INFO] - Step 2018, Loss: {'train/loss': 0.5129169821739197, 'train/grad_norm_pre_clip': 8.145881652832031, 'timing/data': 0.023232427425682545, 'timing/model': 1.1639831252396107, 'learning_rate/qwen_vl_interface': 8.301031627383422e-06, 'learning_rate/action_model': 8.301031627383421e-05, 'epoch': 0.78}) +[2026-06-17 20:32:49,532][starVLA.training.train_starvla][INFO] - Step 2019, Loss: {'train/loss': 0.6701465845108032, 'train/grad_norm_pre_clip': 8.17356014251709, 'timing/data': 0.02165285497903824, 'timing/model': 1.115143376402557, 'learning_rate/qwen_vl_interface': 8.299373750472323e-06, 'learning_rate/action_model': 8.299373750472324e-05, 'epoch': 0.78}) +[2026-06-17 20:32:54,413][starVLA.training.train_starvla][INFO] - Step 2020, Loss: {'train/loss': 0.5049963593482971, 'train/grad_norm_pre_clip': 7.99583625793457, 'timing/data': 0.021157730370759964, 'timing/model': 1.108043328858912, 'learning_rate/qwen_vl_interface': 8.297715241422765e-06, 'learning_rate/action_model': 8.297715241422766e-05, 'epoch': 0.78}) +[2026-06-17 20:32:59,596][starVLA.training.train_starvla][INFO] - Step 2021, Loss: {'train/loss': 0.8511751294136047, 'train/grad_norm_pre_clip': 11.65548324584961, 'timing/data': 0.024015909992158413, 'timing/model': 1.1448624692857265, 'learning_rate/qwen_vl_interface': 8.29605610057856e-06, 'learning_rate/action_model': 8.29605610057856e-05, 'epoch': 0.78}) +[2026-06-17 20:33:04,852][starVLA.training.train_starvla][INFO] - Step 2022, Loss: {'train/loss': 0.47737911343574524, 'train/grad_norm_pre_clip': 9.7044038772583, 'timing/data': 0.012491509318351746, 'timing/model': 1.677201041020453, 'learning_rate/qwen_vl_interface': 8.294396328283647e-06, 'learning_rate/action_model': 8.294396328283645e-05, 'epoch': 0.78}) +[2026-06-17 20:33:09,278][starVLA.training.train_starvla][INFO] - Step 2023, Loss: {'train/loss': 0.557948648929596, 'train/grad_norm_pre_clip': 9.35485553741455, 'timing/data': 0.011959862895309925, 'timing/model': 1.1918646488338709, 'learning_rate/qwen_vl_interface': 8.292735924882097e-06, 'learning_rate/action_model': 8.292735924882097e-05, 'epoch': 0.78}) +[2026-06-17 20:33:14,325][starVLA.training.train_starvla][INFO] - Step 2024, Loss: {'train/loss': 0.6797708868980408, 'train/grad_norm_pre_clip': 11.844183921813965, 'timing/data': 0.011092551052570343, 'timing/model': 1.225723173469305, 'learning_rate/qwen_vl_interface': 8.291074890718117e-06, 'learning_rate/action_model': 8.291074890718117e-05, 'epoch': 0.78}) +[2026-06-17 20:33:19,008][starVLA.training.train_starvla][INFO] - Step 2025, Loss: {'train/loss': 0.9905694723129272, 'train/grad_norm_pre_clip': 9.141191482543945, 'timing/data': 0.011605443432927132, 'timing/model': 1.100337597541511, 'learning_rate/qwen_vl_interface': 8.28941322613604e-06, 'learning_rate/action_model': 8.289413226136039e-05, 'epoch': 0.78}) +[2026-06-17 20:33:23,855][starVLA.training.train_starvla][INFO] - Step 2026, Loss: {'train/loss': 0.9837198257446289, 'train/grad_norm_pre_clip': 10.176450729370117, 'timing/data': 0.010487237945199013, 'timing/model': 1.2674221359193325, 'learning_rate/qwen_vl_interface': 8.287750931480332e-06, 'learning_rate/action_model': 8.287750931480332e-05, 'epoch': 0.78}) +[2026-06-17 20:33:33,290][starVLA.training.train_starvla][INFO] - Step 2027, Loss: {'train/loss': 0.5241144299507141, 'train/grad_norm_pre_clip': 5.9213643074035645, 'timing/data': 0.011679037474095821, 'timing/model': 1.6914938846603036, 'learning_rate/qwen_vl_interface': 8.286088007095584e-06, 'learning_rate/action_model': 8.286088007095584e-05, 'epoch': 0.78}) +[2026-06-17 20:33:37,873][starVLA.training.train_starvla][INFO] - Step 2028, Loss: {'train/loss': 0.7458400726318359, 'train/grad_norm_pre_clip': 11.369662284851074, 'timing/data': 0.012661714106798172, 'timing/model': 1.179081222973764, 'learning_rate/qwen_vl_interface': 8.284424453326528e-06, 'learning_rate/action_model': 8.284424453326528e-05, 'epoch': 0.78}) +[2026-06-17 20:33:42,763][starVLA.training.train_starvla][INFO] - Step 2029, Loss: {'train/loss': 0.3810356557369232, 'train/grad_norm_pre_clip': 6.7719268798828125, 'timing/data': 0.0111969830468297, 'timing/model': 1.0721898535266519, 'learning_rate/qwen_vl_interface': 8.282760270518017e-06, 'learning_rate/action_model': 8.282760270518017e-05, 'epoch': 0.78}) +[2026-06-17 20:33:47,431][starVLA.training.train_starvla][INFO] - Step 2030, Loss: {'train/loss': 0.42214447259902954, 'train/grad_norm_pre_clip': 11.182357788085938, 'timing/data': 0.013142337091267109, 'timing/model': 1.1162677789106965, 'learning_rate/qwen_vl_interface': 8.281095459015037e-06, 'learning_rate/action_model': 8.281095459015036e-05, 'epoch': 0.78}) +[2026-06-17 20:33:52,323][starVLA.training.train_starvla][INFO] - Step 2031, Loss: {'train/loss': 0.385286808013916, 'train/grad_norm_pre_clip': 5.826391696929932, 'timing/data': 0.012966775335371494, 'timing/model': 1.393945712596178, 'learning_rate/qwen_vl_interface': 8.279430019162709e-06, 'learning_rate/action_model': 8.279430019162708e-05, 'epoch': 0.78}) +[2026-06-17 20:33:57,131][starVLA.training.train_starvla][INFO] - Step 2032, Loss: {'train/loss': 0.7183865904808044, 'train/grad_norm_pre_clip': 8.87122631072998, 'timing/data': 0.011676391586661339, 'timing/model': 1.5493168635293841, 'learning_rate/qwen_vl_interface': 8.277763951306277e-06, 'learning_rate/action_model': 8.277763951306277e-05, 'epoch': 0.78}) +[2026-06-17 20:34:02,577][starVLA.training.train_starvla][INFO] - Step 2033, Loss: {'train/loss': 0.6384892463684082, 'train/grad_norm_pre_clip': 11.801924705505371, 'timing/data': 0.01774314045906067, 'timing/model': 1.2598533127456903, 'learning_rate/qwen_vl_interface': 8.27609725579112e-06, 'learning_rate/action_model': 8.27609725579112e-05, 'epoch': 0.78}) +[2026-06-17 20:34:07,707][starVLA.training.train_starvla][INFO] - Step 2034, Loss: {'train/loss': 0.5183061957359314, 'train/grad_norm_pre_clip': 10.008891105651855, 'timing/data': 0.01207483559846878, 'timing/model': 1.0918471338227391, 'learning_rate/qwen_vl_interface': 8.274429932962745e-06, 'learning_rate/action_model': 8.274429932962746e-05, 'epoch': 0.78}) +[2026-06-17 20:34:12,918][starVLA.training.train_starvla][INFO] - Step 2035, Loss: {'train/loss': 0.34522828459739685, 'train/grad_norm_pre_clip': 7.141136169433594, 'timing/data': 0.010776554234325886, 'timing/model': 1.0793865453451872, 'learning_rate/qwen_vl_interface': 8.272761983166793e-06, 'learning_rate/action_model': 8.272761983166792e-05, 'epoch': 0.78}) +[2026-06-17 20:34:22,089][starVLA.training.train_starvla][INFO] - Step 2036, Loss: {'train/loss': 0.5218222737312317, 'train/grad_norm_pre_clip': 7.169755458831787, 'timing/data': 0.011363882571458817, 'timing/model': 1.1738616479560733, 'learning_rate/qwen_vl_interface': 8.271093406749025e-06, 'learning_rate/action_model': 8.271093406749025e-05, 'epoch': 0.78}) +[2026-06-17 20:34:27,923][starVLA.training.train_starvla][INFO] - Step 2037, Loss: {'train/loss': 0.4216400682926178, 'train/grad_norm_pre_clip': 7.207734107971191, 'timing/data': 0.013325074687600136, 'timing/model': 1.5238193031400442, 'learning_rate/qwen_vl_interface': 8.269424204055346e-06, 'learning_rate/action_model': 8.269424204055346e-05, 'epoch': 0.78}) +[2026-06-17 20:34:36,871][starVLA.training.train_starvla][INFO] - Step 2038, Loss: {'train/loss': 0.4403710663318634, 'train/grad_norm_pre_clip': 6.2294464111328125, 'timing/data': 0.017067063599824905, 'timing/model': 1.1317333867773414, 'learning_rate/qwen_vl_interface': 8.26775437543178e-06, 'learning_rate/action_model': 8.26775437543178e-05, 'epoch': 0.78}) +[2026-06-17 20:34:41,820][starVLA.training.train_starvla][INFO] - Step 2039, Loss: {'train/loss': 0.3543301522731781, 'train/grad_norm_pre_clip': 6.580453395843506, 'timing/data': 0.014471134170889854, 'timing/model': 1.1148299407213926, 'learning_rate/qwen_vl_interface': 8.266083921224484e-06, 'learning_rate/action_model': 8.266083921224485e-05, 'epoch': 0.78}) +[2026-06-17 20:34:47,581][starVLA.training.train_starvla][INFO] - Step 2040, Loss: {'train/loss': 0.3981228470802307, 'train/grad_norm_pre_clip': 6.466924667358398, 'timing/data': 0.012622019276022911, 'timing/model': 1.2893290529027581, 'learning_rate/qwen_vl_interface': 8.264412841779747e-06, 'learning_rate/action_model': 8.264412841779746e-05, 'epoch': 0.78}) +[2026-06-17 20:34:59,244][starVLA.training.train_starvla][INFO] - Step 2041, Loss: {'train/loss': 0.4799913167953491, 'train/grad_norm_pre_clip': 6.634981632232666, 'timing/data': 0.01196228712797165, 'timing/model': 1.5332304490730166, 'learning_rate/qwen_vl_interface': 8.262741137443984e-06, 'learning_rate/action_model': 8.262741137443985e-05, 'epoch': 0.79}) +[2026-06-17 20:35:04,195][starVLA.training.train_starvla][INFO] - Step 2042, Loss: {'train/loss': 0.5623634457588196, 'train/grad_norm_pre_clip': 7.496932029724121, 'timing/data': 0.013792206533253193, 'timing/model': 1.3674132460728288, 'learning_rate/qwen_vl_interface': 8.26106880856374e-06, 'learning_rate/action_model': 8.26106880856374e-05, 'epoch': 0.79}) +[2026-06-17 20:35:08,775][starVLA.training.train_starvla][INFO] - Step 2043, Loss: {'train/loss': 0.31515857577323914, 'train/grad_norm_pre_clip': 7.067351341247559, 'timing/data': 0.014580468647181988, 'timing/model': 1.2237896686419845, 'learning_rate/qwen_vl_interface': 8.259395855485693e-06, 'learning_rate/action_model': 8.259395855485694e-05, 'epoch': 0.79}) +[2026-06-17 20:35:13,811][starVLA.training.train_starvla][INFO] - Step 2044, Loss: {'train/loss': 0.6530131101608276, 'train/grad_norm_pre_clip': 12.05180835723877, 'timing/data': 0.012604148127138615, 'timing/model': 1.0747032882645726, 'learning_rate/qwen_vl_interface': 8.257722278556647e-06, 'learning_rate/action_model': 8.257722278556646e-05, 'epoch': 0.79}) +[2026-06-17 20:35:22,649][starVLA.training.train_starvla][INFO] - Step 2045, Loss: {'train/loss': 0.44399723410606384, 'train/grad_norm_pre_clip': 5.617978096008301, 'timing/data': 0.012272559106349945, 'timing/model': 1.090213923715055, 'learning_rate/qwen_vl_interface': 8.256048078123536e-06, 'learning_rate/action_model': 8.256048078123535e-05, 'epoch': 0.79}) +[2026-06-17 20:35:29,248][starVLA.training.train_starvla][INFO] - Step 2046, Loss: {'train/loss': 0.8211739659309387, 'train/grad_norm_pre_clip': 14.212074279785156, 'timing/data': 0.011287004686892033, 'timing/model': 1.3658154057338834, 'learning_rate/qwen_vl_interface': 8.254373254533424e-06, 'learning_rate/action_model': 8.254373254533423e-05, 'epoch': 0.79}) +[2026-06-17 20:35:35,093][starVLA.training.train_starvla][INFO] - Step 2047, Loss: {'train/loss': 0.5751574635505676, 'train/grad_norm_pre_clip': 9.398475646972656, 'timing/data': 0.015934200957417488, 'timing/model': 1.704086302779615, 'learning_rate/qwen_vl_interface': 8.252697808133501e-06, 'learning_rate/action_model': 8.252697808133501e-05, 'epoch': 0.79}) +[2026-06-17 20:35:39,548][starVLA.training.train_starvla][INFO] - Step 2048, Loss: {'train/loss': 0.5440986156463623, 'train/grad_norm_pre_clip': 8.904518127441406, 'timing/data': 0.013938657008111477, 'timing/model': 1.1055981302633882, 'learning_rate/qwen_vl_interface': 8.251021739271095e-06, 'learning_rate/action_model': 8.251021739271094e-05, 'epoch': 0.79}) +[2026-06-17 20:35:44,766][starVLA.training.train_starvla][INFO] - Step 2049, Loss: {'train/loss': 0.32124537229537964, 'train/grad_norm_pre_clip': 7.243020057678223, 'timing/data': 0.012937265448272228, 'timing/model': 1.3722499692812562, 'learning_rate/qwen_vl_interface': 8.249345048293652e-06, 'learning_rate/action_model': 8.24934504829365e-05, 'epoch': 0.79}) +[2026-06-17 20:35:49,584][starVLA.training.train_starvla][INFO] - Step 2050, Loss: {'train/loss': 0.4818357527256012, 'train/grad_norm_pre_clip': 7.760210990905762, 'timing/data': 0.010970016941428185, 'timing/model': 1.124121400527656, 'learning_rate/qwen_vl_interface': 8.247667735548753e-06, 'learning_rate/action_model': 8.247667735548752e-05, 'epoch': 0.79}) +[2026-06-17 20:35:54,576][starVLA.training.train_starvla][INFO] - Step 2051, Loss: {'train/loss': 0.43807926774024963, 'train/grad_norm_pre_clip': 7.341152667999268, 'timing/data': 0.011538369581103325, 'timing/model': 1.6042284965515137, 'learning_rate/qwen_vl_interface': 8.245989801384104e-06, 'learning_rate/action_model': 8.245989801384104e-05, 'epoch': 0.79}) +[2026-06-17 20:35:59,344][starVLA.training.train_starvla][INFO] - Step 2052, Loss: {'train/loss': 0.5306115746498108, 'train/grad_norm_pre_clip': 8.433234214782715, 'timing/data': 0.011104047298431396, 'timing/model': 1.419489729218185, 'learning_rate/qwen_vl_interface': 8.24431124614755e-06, 'learning_rate/action_model': 8.244311246147549e-05, 'epoch': 0.79}) +[2026-06-17 20:36:04,223][starVLA.training.train_starvla][INFO] - Step 2053, Loss: {'train/loss': 0.2238072007894516, 'train/grad_norm_pre_clip': 8.034173011779785, 'timing/data': 0.01220270898193121, 'timing/model': 1.2612114259973168, 'learning_rate/qwen_vl_interface': 8.24263207018705e-06, 'learning_rate/action_model': 8.24263207018705e-05, 'epoch': 0.79}) +[2026-06-17 20:36:09,469][starVLA.training.train_starvla][INFO] - Step 2054, Loss: {'train/loss': 0.46156546473503113, 'train/grad_norm_pre_clip': 9.157987594604492, 'timing/data': 0.016965053044259548, 'timing/model': 1.397903222590685, 'learning_rate/qwen_vl_interface': 8.2409522738507e-06, 'learning_rate/action_model': 8.240952273850701e-05, 'epoch': 0.79}) +[2026-06-17 20:36:14,342][starVLA.training.train_starvla][INFO] - Step 2055, Loss: {'train/loss': 0.6320776343345642, 'train/grad_norm_pre_clip': 8.587606430053711, 'timing/data': 0.014451555907726288, 'timing/model': 1.1444103689864278, 'learning_rate/qwen_vl_interface': 8.23927185748673e-06, 'learning_rate/action_model': 8.239271857486729e-05, 'epoch': 0.79}) +[2026-06-17 20:36:21,621][starVLA.training.train_starvla][INFO] - Step 2056, Loss: {'train/loss': 0.5093780159950256, 'train/grad_norm_pre_clip': 12.125749588012695, 'timing/data': 0.011241307482123375, 'timing/model': 1.8908166727051139, 'learning_rate/qwen_vl_interface': 8.237590821443484e-06, 'learning_rate/action_model': 8.237590821443483e-05, 'epoch': 0.79}) +[2026-06-17 20:36:31,771][starVLA.training.train_starvla][INFO] - Step 2057, Loss: {'train/loss': 0.34502387046813965, 'train/grad_norm_pre_clip': 6.633650779724121, 'timing/data': 0.015055470168590546, 'timing/model': 5.331770402379334, 'learning_rate/qwen_vl_interface': 8.235909166069447e-06, 'learning_rate/action_model': 8.235909166069447e-05, 'epoch': 0.79}) +[2026-06-17 20:36:47,333][starVLA.training.train_starvla][INFO] - Step 2058, Loss: {'train/loss': 0.6357517242431641, 'train/grad_norm_pre_clip': 8.968242645263672, 'timing/data': 0.011414261534810066, 'timing/model': 4.3509077513590455, 'learning_rate/qwen_vl_interface': 8.234226891713225e-06, 'learning_rate/action_model': 8.234226891713225e-05, 'epoch': 0.79}) +[2026-06-17 20:36:52,240][starVLA.training.train_starvla][INFO] - Step 2059, Loss: {'train/loss': 0.4905511140823364, 'train/grad_norm_pre_clip': 7.125848770141602, 'timing/data': 0.01410119328647852, 'timing/model': 1.091634695418179, 'learning_rate/qwen_vl_interface': 8.232543998723559e-06, 'learning_rate/action_model': 8.232543998723559e-05, 'epoch': 0.79}) +[2026-06-17 20:37:02,407][starVLA.training.train_starvla][INFO] - Step 2060, Loss: {'train/loss': 0.20602211356163025, 'train/grad_norm_pre_clip': 4.348930358886719, 'timing/data': 0.013643335551023483, 'timing/model': 1.3303218930959702, 'learning_rate/qwen_vl_interface': 8.230860487449314e-06, 'learning_rate/action_model': 8.230860487449313e-05, 'epoch': 0.79}) +[2026-06-17 20:37:17,410][starVLA.training.train_starvla][INFO] - Step 2061, Loss: {'train/loss': 0.6258521676063538, 'train/grad_norm_pre_clip': 10.414944648742676, 'timing/data': 0.014396640472114086, 'timing/model': 1.8642494063824415, 'learning_rate/qwen_vl_interface': 8.22917635823948e-06, 'learning_rate/action_model': 8.229176358239478e-05, 'epoch': 0.79}) +[2026-06-17 20:37:25,123][starVLA.training.train_starvla][INFO] - Step 2062, Loss: {'train/loss': 0.3000924587249756, 'train/grad_norm_pre_clip': 7.012938976287842, 'timing/data': 0.017210391350090504, 'timing/model': 1.4098045192658901, 'learning_rate/qwen_vl_interface': 8.227491611443182e-06, 'learning_rate/action_model': 8.227491611443182e-05, 'epoch': 0.79}) +[2026-06-17 20:37:29,744][starVLA.training.train_starvla][INFO] - Step 2063, Loss: {'train/loss': 0.5881537795066833, 'train/grad_norm_pre_clip': 13.848896980285645, 'timing/data': 0.01187441311776638, 'timing/model': 1.088545604608953, 'learning_rate/qwen_vl_interface': 8.225806247409667e-06, 'learning_rate/action_model': 8.225806247409668e-05, 'epoch': 0.79}) +[2026-06-17 20:37:34,761][starVLA.training.train_starvla][INFO] - Step 2064, Loss: {'train/loss': 0.4436836242675781, 'train/grad_norm_pre_clip': 6.484488010406494, 'timing/data': 0.01986851915717125, 'timing/model': 1.1905942726880312, 'learning_rate/qwen_vl_interface': 8.224120266488318e-06, 'learning_rate/action_model': 8.224120266488318e-05, 'epoch': 0.79}) +[2026-06-17 20:37:39,725][starVLA.training.train_starvla][INFO] - Step 2065, Loss: {'train/loss': 0.46444177627563477, 'train/grad_norm_pre_clip': 8.743640899658203, 'timing/data': 0.012313554994761944, 'timing/model': 1.2085265861824155, 'learning_rate/qwen_vl_interface': 8.222433669028636e-06, 'learning_rate/action_model': 8.222433669028635e-05, 'epoch': 0.79}) +[2026-06-17 20:37:44,550][starVLA.training.train_starvla][INFO] - Step 2066, Loss: {'train/loss': 0.6254366636276245, 'train/grad_norm_pre_clip': 11.386128425598145, 'timing/data': 0.012020905502140522, 'timing/model': 1.2667668405920267, 'learning_rate/qwen_vl_interface': 8.220746455380257e-06, 'learning_rate/action_model': 8.220746455380257e-05, 'epoch': 0.79}) +[2026-06-17 20:37:49,417][starVLA.training.train_starvla][INFO] - Step 2067, Loss: {'train/loss': 0.364702969789505, 'train/grad_norm_pre_clip': 7.858005523681641, 'timing/data': 0.012267661280930042, 'timing/model': 1.6271746791899204, 'learning_rate/qwen_vl_interface': 8.219058625892939e-06, 'learning_rate/action_model': 8.21905862589294e-05, 'epoch': 0.8}) +[2026-06-17 20:37:53,610][starVLA.training.train_starvla][INFO] - Step 2068, Loss: {'train/loss': 0.45017918944358826, 'train/grad_norm_pre_clip': 7.867743015289307, 'timing/data': 0.010976150631904602, 'timing/model': 1.0667512072250247, 'learning_rate/qwen_vl_interface': 8.217370180916574e-06, 'learning_rate/action_model': 8.217370180916575e-05, 'epoch': 0.8}) +[2026-06-17 20:37:58,629][starVLA.training.train_starvla][INFO] - Step 2069, Loss: {'train/loss': 0.34837621450424194, 'train/grad_norm_pre_clip': 8.540075302124023, 'timing/data': 0.013802609406411648, 'timing/model': 1.2506809830665588, 'learning_rate/qwen_vl_interface': 8.215681120801179e-06, 'learning_rate/action_model': 8.215681120801179e-05, 'epoch': 0.8}) +[2026-06-17 20:38:03,288][starVLA.training.train_starvla][INFO] - Step 2070, Loss: {'train/loss': 0.2951033413410187, 'train/grad_norm_pre_clip': 6.204377174377441, 'timing/data': 0.01154117751866579, 'timing/model': 1.1904657259583473, 'learning_rate/qwen_vl_interface': 8.213991445896895e-06, 'learning_rate/action_model': 8.213991445896895e-05, 'epoch': 0.8}) +[2026-06-17 20:38:08,257][starVLA.training.train_starvla][INFO] - Step 2071, Loss: {'train/loss': 0.8786373734474182, 'train/grad_norm_pre_clip': 13.70569133758545, 'timing/data': 0.015713391825556755, 'timing/model': 1.644648396410048, 'learning_rate/qwen_vl_interface': 8.212301156553995e-06, 'learning_rate/action_model': 8.212301156553995e-05, 'epoch': 0.8}) +[2026-06-17 20:38:13,213][starVLA.training.train_starvla][INFO] - Step 2072, Loss: {'train/loss': 0.9971617460250854, 'train/grad_norm_pre_clip': 17.86736297607422, 'timing/data': 0.01251591369509697, 'timing/model': 1.4046386843547225, 'learning_rate/qwen_vl_interface': 8.210610253122879e-06, 'learning_rate/action_model': 8.210610253122879e-05, 'epoch': 0.8}) +[2026-06-17 20:38:17,538][starVLA.training.train_starvla][INFO] - Step 2073, Loss: {'train/loss': 0.4683656394481659, 'train/grad_norm_pre_clip': 6.28384256362915, 'timing/data': 0.013143426738679409, 'timing/model': 1.1802130602300167, 'learning_rate/qwen_vl_interface': 8.20891873595407e-06, 'learning_rate/action_model': 8.20891873595407e-05, 'epoch': 0.8}) +[2026-06-17 20:38:22,525][starVLA.training.train_starvla][INFO] - Step 2074, Loss: {'train/loss': 0.4401598572731018, 'train/grad_norm_pre_clip': 6.340285301208496, 'timing/data': 0.012399851344525814, 'timing/model': 1.0892947353422642, 'learning_rate/qwen_vl_interface': 8.207226605398225e-06, 'learning_rate/action_model': 8.207226605398224e-05, 'epoch': 0.8}) +[2026-06-17 20:38:27,465][starVLA.training.train_starvla][INFO] - Step 2075, Loss: {'train/loss': 0.4414528012275696, 'train/grad_norm_pre_clip': 9.068408966064453, 'timing/data': 0.011998722329735756, 'timing/model': 1.1863231230527163, 'learning_rate/qwen_vl_interface': 8.20553386180612e-06, 'learning_rate/action_model': 8.205533861806119e-05, 'epoch': 0.8}) +[2026-06-17 20:38:32,542][starVLA.training.train_starvla][INFO] - Step 2076, Loss: {'train/loss': 0.5049185156822205, 'train/grad_norm_pre_clip': 8.18803596496582, 'timing/data': 0.023685774765908718, 'timing/model': 1.2438053581863642, 'learning_rate/qwen_vl_interface': 8.203840505528665e-06, 'learning_rate/action_model': 8.203840505528665e-05, 'epoch': 0.8}) +[2026-06-17 20:38:41,705][starVLA.training.train_starvla][INFO] - Step 2077, Loss: {'train/loss': 0.47274458408355713, 'train/grad_norm_pre_clip': 10.587530136108398, 'timing/data': 0.011200763285160065, 'timing/model': 4.4124490255489945, 'learning_rate/qwen_vl_interface': 8.202146536916899e-06, 'learning_rate/action_model': 8.202146536916899e-05, 'epoch': 0.8}) +[2026-06-17 20:38:47,425][starVLA.training.train_starvla][INFO] - Step 2078, Loss: {'train/loss': 0.54160076379776, 'train/grad_norm_pre_clip': 10.519272804260254, 'timing/data': 0.011856265366077423, 'timing/model': 1.5600933842360973, 'learning_rate/qwen_vl_interface': 8.200451956321974e-06, 'learning_rate/action_model': 8.200451956321974e-05, 'epoch': 0.8}) +[2026-06-17 20:39:01,429][starVLA.training.train_starvla][INFO] - Step 2079, Loss: {'train/loss': 0.2903149425983429, 'train/grad_norm_pre_clip': 7.587736129760742, 'timing/data': 0.01265642885118723, 'timing/model': 5.085110508836806, 'learning_rate/qwen_vl_interface': 8.198756764095187e-06, 'learning_rate/action_model': 8.198756764095187e-05, 'epoch': 0.8}) +[2026-06-17 20:39:10,425][starVLA.training.train_starvla][INFO] - Step 2080, Loss: {'train/loss': 1.0417797565460205, 'train/grad_norm_pre_clip': 11.310831069946289, 'timing/data': 0.029898984357714653, 'timing/model': 1.712779146619141, 'learning_rate/qwen_vl_interface': 8.19706096058795e-06, 'learning_rate/action_model': 8.197060960587949e-05, 'epoch': 0.8}) +[2026-06-17 20:39:15,808][starVLA.training.train_starvla][INFO] - Step 2081, Loss: {'train/loss': 1.0570769309997559, 'train/grad_norm_pre_clip': 12.336371421813965, 'timing/data': 0.012366244569420815, 'timing/model': 1.8509729336947203, 'learning_rate/qwen_vl_interface': 8.195364546151802e-06, 'learning_rate/action_model': 8.195364546151801e-05, 'epoch': 0.8}) +[2026-06-17 20:39:26,824][starVLA.training.train_starvla][INFO] - Step 2082, Loss: {'train/loss': 0.379328191280365, 'train/grad_norm_pre_clip': 9.550870895385742, 'timing/data': 0.01500999927520752, 'timing/model': 6.133287278003991, 'learning_rate/qwen_vl_interface': 8.193667521138416e-06, 'learning_rate/action_model': 8.193667521138416e-05, 'epoch': 0.8}) +[2026-06-17 20:39:33,578][starVLA.training.train_starvla][INFO] - Step 2083, Loss: {'train/loss': 0.735677182674408, 'train/grad_norm_pre_clip': 11.175646781921387, 'timing/data': 0.012767321430146694, 'timing/model': 1.4291599141433835, 'learning_rate/qwen_vl_interface': 8.191969885899585e-06, 'learning_rate/action_model': 8.191969885899585e-05, 'epoch': 0.8}) +[2026-06-17 20:39:39,232][starVLA.training.train_starvla][INFO] - Step 2084, Loss: {'train/loss': 0.3843432366847992, 'train/grad_norm_pre_clip': 10.01549243927002, 'timing/data': 0.06784698832780123, 'timing/model': 1.515780963934958, 'learning_rate/qwen_vl_interface': 8.190271640787232e-06, 'learning_rate/action_model': 8.190271640787232e-05, 'epoch': 0.8}) +[2026-06-17 20:39:44,321][starVLA.training.train_starvla][INFO] - Step 2085, Loss: {'train/loss': 0.5949745178222656, 'train/grad_norm_pre_clip': 8.253612518310547, 'timing/data': 0.012903042137622833, 'timing/model': 1.0829364024102688, 'learning_rate/qwen_vl_interface': 8.188572786153403e-06, 'learning_rate/action_model': 8.188572786153402e-05, 'epoch': 0.8}) +[2026-06-17 20:39:49,124][starVLA.training.train_starvla][INFO] - Step 2086, Loss: {'train/loss': 0.4112662076950073, 'train/grad_norm_pre_clip': 6.602636814117432, 'timing/data': 0.011754177510738373, 'timing/model': 1.1443355539813638, 'learning_rate/qwen_vl_interface': 8.186873322350274e-06, 'learning_rate/action_model': 8.186873322350273e-05, 'epoch': 0.8}) +[2026-06-17 20:39:53,674][starVLA.training.train_starvla][INFO] - Step 2087, Loss: {'train/loss': 0.3250803053379059, 'train/grad_norm_pre_clip': 6.01364803314209, 'timing/data': 0.011341716162860394, 'timing/model': 1.4364614114165306, 'learning_rate/qwen_vl_interface': 8.185173249730143e-06, 'learning_rate/action_model': 8.185173249730142e-05, 'epoch': 0.8}) +[2026-06-17 20:39:58,711][starVLA.training.train_starvla][INFO] - Step 2088, Loss: {'train/loss': 0.4862441122531891, 'train/grad_norm_pre_clip': 8.060135841369629, 'timing/data': 0.02129403594881296, 'timing/model': 1.0907418448477983, 'learning_rate/qwen_vl_interface': 8.183472568645441e-06, 'learning_rate/action_model': 8.18347256864544e-05, 'epoch': 0.8}) +[2026-06-17 20:40:03,801][starVLA.training.train_starvla][INFO] - Step 2089, Loss: {'train/loss': 0.2982688844203949, 'train/grad_norm_pre_clip': 6.958232402801514, 'timing/data': 0.014402607455849648, 'timing/model': 1.3918928122147918, 'learning_rate/qwen_vl_interface': 8.181771279448718e-06, 'learning_rate/action_model': 8.181771279448718e-05, 'epoch': 0.8}) +[2026-06-17 20:40:08,680][starVLA.training.train_starvla][INFO] - Step 2090, Loss: {'train/loss': 0.36374709010124207, 'train/grad_norm_pre_clip': 7.381313323974609, 'timing/data': 0.011783702298998833, 'timing/model': 1.1941883163526654, 'learning_rate/qwen_vl_interface': 8.180069382492655e-06, 'learning_rate/action_model': 8.180069382492654e-05, 'epoch': 0.8}) +[2026-06-17 20:40:13,580][starVLA.training.train_starvla][INFO] - Step 2091, Loss: {'train/loss': 0.5846970677375793, 'train/grad_norm_pre_clip': 9.763282775878906, 'timing/data': 0.012026929296553135, 'timing/model': 1.592065842822194, 'learning_rate/qwen_vl_interface': 8.178366878130056e-06, 'learning_rate/action_model': 8.178366878130056e-05, 'epoch': 0.8}) +[2026-06-17 20:40:19,189][starVLA.training.train_starvla][INFO] - Step 2092, Loss: {'train/loss': 0.8330325484275818, 'train/grad_norm_pre_clip': 10.45030403137207, 'timing/data': 0.013947207480669022, 'timing/model': 1.4553579855710268, 'learning_rate/qwen_vl_interface': 8.176663766713852e-06, 'learning_rate/action_model': 8.176663766713852e-05, 'epoch': 0.8}) +[2026-06-17 20:40:31,029][starVLA.training.train_starvla][INFO] - Step 2093, Loss: {'train/loss': 0.902311384677887, 'train/grad_norm_pre_clip': 13.28979206085205, 'timing/data': 0.016540479846298695, 'timing/model': 1.098922927863896, 'learning_rate/qwen_vl_interface': 8.174960048597099e-06, 'learning_rate/action_model': 8.174960048597099e-05, 'epoch': 0.81}) +[2026-06-17 20:40:36,218][starVLA.training.train_starvla][INFO] - Step 2094, Loss: {'train/loss': 0.626642107963562, 'train/grad_norm_pre_clip': 10.089249610900879, 'timing/data': 0.01805195678025484, 'timing/model': 1.4219330111518502, 'learning_rate/qwen_vl_interface': 8.173255724132982e-06, 'learning_rate/action_model': 8.17325572413298e-05, 'epoch': 0.81}) +[2026-06-17 20:40:40,901][starVLA.training.train_starvla][INFO] - Step 2095, Loss: {'train/loss': 0.5810946226119995, 'train/grad_norm_pre_clip': 6.93402624130249, 'timing/data': 0.014397762715816498, 'timing/model': 1.0870231240987778, 'learning_rate/qwen_vl_interface': 8.171550793674806e-06, 'learning_rate/action_model': 8.171550793674805e-05, 'epoch': 0.81}) +[2026-06-17 20:40:45,626][starVLA.training.train_starvla][INFO] - Step 2096, Loss: {'train/loss': 0.20577377080917358, 'train/grad_norm_pre_clip': 5.157343864440918, 'timing/data': 0.01368835661560297, 'timing/model': 1.1817453689873219, 'learning_rate/qwen_vl_interface': 8.169845257576008e-06, 'learning_rate/action_model': 8.169845257576007e-05, 'epoch': 0.81}) +[2026-06-17 20:40:50,520][starVLA.training.train_starvla][INFO] - Step 2097, Loss: {'train/loss': 0.5214648246765137, 'train/grad_norm_pre_clip': 6.353389739990234, 'timing/data': 0.015377725474536419, 'timing/model': 1.5108142476528883, 'learning_rate/qwen_vl_interface': 8.168139116190146e-06, 'learning_rate/action_model': 8.168139116190146e-05, 'epoch': 0.81}) +[2026-06-17 20:40:58,125][starVLA.training.train_starvla][INFO] - Step 2098, Loss: {'train/loss': 0.7923218011856079, 'train/grad_norm_pre_clip': 10.797622680664062, 'timing/data': 0.011808842420578003, 'timing/model': 1.0896170372143388, 'learning_rate/qwen_vl_interface': 8.166432369870905e-06, 'learning_rate/action_model': 8.166432369870904e-05, 'epoch': 0.81}) +[2026-06-17 20:41:03,237][starVLA.training.train_starvla][INFO] - Step 2099, Loss: {'train/loss': 0.7142546772956848, 'train/grad_norm_pre_clip': 8.714611053466797, 'timing/data': 0.013902631588280201, 'timing/model': 1.4530828995630145, 'learning_rate/qwen_vl_interface': 8.164725018972096e-06, 'learning_rate/action_model': 8.164725018972096e-05, 'epoch': 0.81}) +[2026-06-17 20:44:30,339][starVLA.training.train_starvla][INFO] - Step 2100, Loss: {'train/loss': 0.496656209230423, 'train/grad_norm_pre_clip': 8.376485824584961, 'eval/loss': 0.5556282997131348, 'eval/latency_4/loss': 0.5556282997131348, 'eval/demon_fire_precision': 0.9426810477657935, 'eval/demon_fire_recall': 0.9935043845404352, 'eval/demon_fire_f1': 0.9674256799493992, 'eval/demon_fire_rate_model': 0.9530102790014684, 'eval/demon_fire_rate_teacher': 0.9042584434654919, 'eval/demon_left_precision': 0.8225806451612904, 'eval/demon_left_recall': 0.3493150684931507, 'eval/demon_left_f1': 0.4903846153846155, 'eval/demon_left_rate_model': 0.05462555066079295, 'eval/demon_left_rate_teacher': 0.12863436123348018, 'eval/demon_right_precision': 0.6607669616519174, 'eval/demon_right_recall': 0.8086642599277978, 'eval/demon_right_f1': 0.7272727272727272, 'eval/demon_right_rate_model': 0.19911894273127753, 'eval/demon_right_rate_teacher': 0.16270190895741557, 'eval/demon_move_f1': 0.6088286713286714, 'eval/demon_cc_f1': 0.8221525683818205, 'eval/latency_4/demon_fire_precision': 0.9426810477657935, 'eval/latency_4/demon_fire_recall': 0.9935043845404352, 'eval/latency_4/demon_fire_f1': 0.9674256799493992, 'eval/latency_4/demon_fire_rate_model': 0.9530102790014684, 'eval/latency_4/demon_fire_rate_teacher': 0.9042584434654919, 'eval/latency_4/demon_left_precision': 0.8225806451612904, 'eval/latency_4/demon_left_recall': 0.3493150684931507, 'eval/latency_4/demon_left_f1': 0.4903846153846155, 'eval/latency_4/demon_left_rate_model': 0.05462555066079295, 'eval/latency_4/demon_left_rate_teacher': 0.12863436123348018, 'eval/latency_4/demon_right_precision': 0.6607669616519174, 'eval/latency_4/demon_right_recall': 0.8086642599277978, 'eval/latency_4/demon_right_f1': 0.7272727272727272, 'eval/latency_4/demon_right_rate_model': 0.19911894273127753, 'eval/latency_4/demon_right_rate_teacher': 0.16270190895741557, 'eval/latency_4/demon_move_f1': 0.6088286713286714, 'eval/latency_4/demon_cc_f1': 0.8221525683818205, 'timing/data': 0.014366247691214085, 'timing/model': 1.1928979167714715, 'learning_rate/qwen_vl_interface': 8.163017063847654e-06, 'learning_rate/action_model': 8.163017063847654e-05, 'epoch': 0.81}) +[2026-06-17 20:44:30,340][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_2100_state +[2026-06-17 20:44:30,347][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 20:44:38,321][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_2100_state/model.safetensors +[2026-06-17 20:45:03,275][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_2100_state/optimizer.bin +[2026-06-17 20:45:03,277][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_2100_state/sampler.bin +[2026-06-17 20:45:03,278][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_2100_state/sampler_1.bin +[2026-06-17 20:45:03,280][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_2100_state/random_states_0.pkl +[2026-06-17 20:45:03,286][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2100 +[2026-06-17 20:45:03,286][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 20:45:03,304][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 20:45:08,547][starVLA.training.train_starvla][INFO] - Step 2101, Loss: {'train/loss': 0.6465628147125244, 'train/grad_norm_pre_clip': 7.250955104827881, 'timing/data': 0.014879425056278706, 'timing/model': 1.3495822008699179, 'learning_rate/qwen_vl_interface': 8.16130850485164e-06, 'learning_rate/action_model': 8.161308504851639e-05, 'epoch': 0.81}) +[2026-06-17 20:45:16,942][starVLA.training.train_starvla][INFO] - Step 2102, Loss: {'train/loss': 0.5324621796607971, 'train/grad_norm_pre_clip': 8.807061195373535, 'timing/data': 0.029156322591006756, 'timing/model': 1.6412958269938827, 'learning_rate/qwen_vl_interface': 8.15959934233824e-06, 'learning_rate/action_model': 8.15959934233824e-05, 'epoch': 0.81}) +[2026-06-17 20:45:22,185][starVLA.training.train_starvla][INFO] - Step 2103, Loss: {'train/loss': 0.49955490231513977, 'train/grad_norm_pre_clip': 7.025611400604248, 'timing/data': 0.012314178980886936, 'timing/model': 1.4092753138393164, 'learning_rate/qwen_vl_interface': 8.157889576661765e-06, 'learning_rate/action_model': 8.157889576661764e-05, 'epoch': 0.81}) +[2026-06-17 20:45:28,011][starVLA.training.train_starvla][INFO] - Step 2104, Loss: {'train/loss': 0.5775049328804016, 'train/grad_norm_pre_clip': 7.600554466247559, 'timing/data': 0.020600934512913227, 'timing/model': 1.3605031156912446, 'learning_rate/qwen_vl_interface': 8.156179208176652e-06, 'learning_rate/action_model': 8.156179208176652e-05, 'epoch': 0.81}) +[2026-06-17 20:45:33,413][starVLA.training.train_starvla][INFO] - Step 2105, Loss: {'train/loss': 0.477107435464859, 'train/grad_norm_pre_clip': 7.901850700378418, 'timing/data': 0.01847406104207039, 'timing/model': 1.0854809060692787, 'learning_rate/qwen_vl_interface': 8.154468237237463e-06, 'learning_rate/action_model': 8.154468237237462e-05, 'epoch': 0.81}) +[2026-06-17 20:45:38,622][starVLA.training.train_starvla][INFO] - Step 2106, Loss: {'train/loss': 0.38802462816238403, 'train/grad_norm_pre_clip': 10.037758827209473, 'timing/data': 0.01805596984922886, 'timing/model': 1.0911671863868833, 'learning_rate/qwen_vl_interface': 8.152756664198882e-06, 'learning_rate/action_model': 8.152756664198882e-05, 'epoch': 0.81}) +[2026-06-17 20:45:43,504][starVLA.training.train_starvla][INFO] - Step 2107, Loss: {'train/loss': 0.5084813833236694, 'train/grad_norm_pre_clip': 5.595035552978516, 'timing/data': 0.012304666452109814, 'timing/model': 1.5416455268859863, 'learning_rate/qwen_vl_interface': 8.151044489415722e-06, 'learning_rate/action_model': 8.151044489415722e-05, 'epoch': 0.81}) +[2026-06-17 20:45:50,413][starVLA.training.train_starvla][INFO] - Step 2108, Loss: {'train/loss': 0.6518458127975464, 'train/grad_norm_pre_clip': 8.788641929626465, 'timing/data': 0.011866183020174503, 'timing/model': 1.3484376426786184, 'learning_rate/qwen_vl_interface': 8.149331713242916e-06, 'learning_rate/action_model': 8.149331713242916e-05, 'epoch': 0.81}) +[2026-06-17 20:46:04,637][starVLA.training.train_starvla][INFO] - Step 2109, Loss: {'train/loss': 0.6319363713264465, 'train/grad_norm_pre_clip': 6.508394718170166, 'timing/data': 0.016243789345026016, 'timing/model': 3.061146770603955, 'learning_rate/qwen_vl_interface': 8.147618336035526e-06, 'learning_rate/action_model': 8.147618336035526e-05, 'epoch': 0.81}) +[2026-06-17 20:46:12,478][starVLA.training.train_starvla][INFO] - Step 2110, Loss: {'train/loss': 0.6736183762550354, 'train/grad_norm_pre_clip': 11.327506065368652, 'timing/data': 0.02342251595109701, 'timing/model': 1.5838097045198083, 'learning_rate/qwen_vl_interface': 8.14590435814874e-06, 'learning_rate/action_model': 8.14590435814874e-05, 'epoch': 0.81}) +[2026-06-17 20:46:26,646][starVLA.training.train_starvla][INFO] - Step 2111, Loss: {'train/loss': 0.7119235396385193, 'train/grad_norm_pre_clip': 7.815996170043945, 'timing/data': 0.0023596882820129395, 'timing/model': 1.1400770265609026, 'learning_rate/qwen_vl_interface': 8.144189779937862e-06, 'learning_rate/action_model': 8.144189779937862e-05, 'epoch': 0.81}) +[2026-06-17 20:46:36,966][starVLA.training.train_starvla][INFO] - Step 2112, Loss: {'train/loss': 0.47922080755233765, 'train/grad_norm_pre_clip': 7.091773509979248, 'timing/data': 0.0014403685927391052, 'timing/model': 1.8856504615396261, 'learning_rate/qwen_vl_interface': 8.142474601758331e-06, 'learning_rate/action_model': 8.142474601758329e-05, 'epoch': 0.81}) +[2026-06-17 20:46:42,101][starVLA.training.train_starvla][INFO] - Step 2113, Loss: {'train/loss': 0.702665388584137, 'train/grad_norm_pre_clip': 7.0876994132995605, 'timing/data': 0.001019096001982689, 'timing/model': 1.4118498945608735, 'learning_rate/qwen_vl_interface': 8.1407588239657e-06, 'learning_rate/action_model': 8.140758823965698e-05, 'epoch': 0.81}) +[2026-06-17 20:46:47,556][starVLA.training.train_starvla][INFO] - Step 2114, Loss: {'train/loss': 0.3430936932563782, 'train/grad_norm_pre_clip': 4.976504325866699, 'timing/data': 0.004485325887799263, 'timing/model': 1.6334068346768618, 'learning_rate/qwen_vl_interface': 8.139042446915653e-06, 'learning_rate/action_model': 8.139042446915654e-05, 'epoch': 0.81}) +[2026-06-17 20:46:53,541][starVLA.training.train_starvla][INFO] - Step 2115, Loss: {'train/loss': 0.3301767408847809, 'train/grad_norm_pre_clip': 10.984932899475098, 'timing/data': 0.0012345686554908752, 'timing/model': 1.2901383927091956, 'learning_rate/qwen_vl_interface': 8.137325470964001e-06, 'learning_rate/action_model': 8.137325470964e-05, 'epoch': 0.81}) +[2026-06-17 20:46:59,144][starVLA.training.train_starvla][INFO] - Step 2116, Loss: {'train/loss': 0.5423939228057861, 'train/grad_norm_pre_clip': 6.911818027496338, 'timing/data': 0.0021855225786566734, 'timing/model': 1.5892584817484021, 'learning_rate/qwen_vl_interface': 8.135607896466669e-06, 'learning_rate/action_model': 8.13560789646667e-05, 'epoch': 0.81}) +[2026-06-17 20:47:09,737][starVLA.training.train_starvla][INFO] - Step 2117, Loss: {'train/loss': 0.6657010912895203, 'train/grad_norm_pre_clip': 6.984287738800049, 'timing/data': 0.0010233987122774124, 'timing/model': 1.957762992940843, 'learning_rate/qwen_vl_interface': 8.133889723779718e-06, 'learning_rate/action_model': 8.133889723779717e-05, 'epoch': 0.81}) +[2026-06-17 20:47:14,596][starVLA.training.train_starvla][INFO] - Step 2118, Loss: {'train/loss': 0.35510683059692383, 'train/grad_norm_pre_clip': 6.517223834991455, 'timing/data': 0.0012655854225158691, 'timing/model': 1.2776457741856575, 'learning_rate/qwen_vl_interface': 8.132170953259321e-06, 'learning_rate/action_model': 8.132170953259322e-05, 'epoch': 0.81}) +[2026-06-17 20:47:19,857][starVLA.training.train_starvla][INFO] - Step 2119, Loss: {'train/loss': 0.4382966458797455, 'train/grad_norm_pre_clip': 8.819416046142578, 'timing/data': 0.0010257074609398842, 'timing/model': 1.4529614830389619, 'learning_rate/qwen_vl_interface': 8.130451585261786e-06, 'learning_rate/action_model': 8.130451585261786e-05, 'epoch': 0.82}) +[2026-06-17 20:47:26,898][starVLA.training.train_starvla][INFO] - Step 2120, Loss: {'train/loss': 0.4196225702762604, 'train/grad_norm_pre_clip': 9.08751106262207, 'timing/data': 0.0027724802494049072, 'timing/model': 2.688651505857706, 'learning_rate/qwen_vl_interface': 8.128731620143538e-06, 'learning_rate/action_model': 8.128731620143538e-05, 'epoch': 0.82}) +[2026-06-17 20:47:38,446][starVLA.training.train_starvla][INFO] - Step 2121, Loss: {'train/loss': 0.922444760799408, 'train/grad_norm_pre_clip': 11.572559356689453, 'timing/data': 0.0012840721756219864, 'timing/model': 1.3924592360854149, 'learning_rate/qwen_vl_interface': 8.127011058261129e-06, 'learning_rate/action_model': 8.127011058261128e-05, 'epoch': 0.82}) +[2026-06-17 20:47:43,479][starVLA.training.train_starvla][INFO] - Step 2122, Loss: {'train/loss': 0.469568133354187, 'train/grad_norm_pre_clip': 9.588600158691406, 'timing/data': 0.0013443520292639732, 'timing/model': 1.2374846786260605, 'learning_rate/qwen_vl_interface': 8.125289899971229e-06, 'learning_rate/action_model': 8.125289899971228e-05, 'epoch': 0.82}) +[2026-06-17 20:47:50,433][starVLA.training.train_starvla][INFO] - Step 2123, Loss: {'train/loss': 0.7270400524139404, 'train/grad_norm_pre_clip': 9.6555814743042, 'timing/data': 0.00140360277146101, 'timing/model': 1.2156863389536738, 'learning_rate/qwen_vl_interface': 8.123568145630641e-06, 'learning_rate/action_model': 8.12356814563064e-05, 'epoch': 0.82}) +[2026-06-17 20:47:59,285][starVLA.training.train_starvla][INFO] - Step 2124, Loss: {'train/loss': 0.5371881127357483, 'train/grad_norm_pre_clip': 9.620656967163086, 'timing/data': 0.0011885566636919975, 'timing/model': 1.2223956249654293, 'learning_rate/qwen_vl_interface': 8.121845795596282e-06, 'learning_rate/action_model': 8.121845795596282e-05, 'epoch': 0.82}) +[2026-06-17 20:48:05,329][starVLA.training.train_starvla][INFO] - Step 2125, Loss: {'train/loss': 0.8207418322563171, 'train/grad_norm_pre_clip': 13.237861633300781, 'timing/data': 0.0011993711814284325, 'timing/model': 1.5408911919221282, 'learning_rate/qwen_vl_interface': 8.1201228502252e-06, 'learning_rate/action_model': 8.120122850225199e-05, 'epoch': 0.82}) +[2026-06-17 20:48:18,142][starVLA.training.train_starvla][INFO] - Step 2126, Loss: {'train/loss': 0.44591590762138367, 'train/grad_norm_pre_clip': 7.852816581726074, 'timing/data': 0.0010876143351197243, 'timing/model': 5.652296295389533, 'learning_rate/qwen_vl_interface': 8.118399309874563e-06, 'learning_rate/action_model': 8.118399309874563e-05, 'epoch': 0.82}) +[2026-06-17 20:48:23,535][starVLA.training.train_starvla][INFO] - Step 2127, Loss: {'train/loss': 0.36834996938705444, 'train/grad_norm_pre_clip': 6.925703525543213, 'timing/data': 0.0012434255331754684, 'timing/model': 1.4811425926163793, 'learning_rate/qwen_vl_interface': 8.116675174901663e-06, 'learning_rate/action_model': 8.116675174901662e-05, 'epoch': 0.82}) +[2026-06-17 20:48:28,014][starVLA.training.train_starvla][INFO] - Step 2128, Loss: {'train/loss': 0.4658900201320648, 'train/grad_norm_pre_clip': 6.619154930114746, 'timing/data': 0.0011337501928210258, 'timing/model': 1.0772585133090615, 'learning_rate/qwen_vl_interface': 8.114950445663915e-06, 'learning_rate/action_model': 8.114950445663913e-05, 'epoch': 0.82}) +[2026-06-17 20:48:33,045][starVLA.training.train_starvla][INFO] - Step 2129, Loss: {'train/loss': 0.42256659269332886, 'train/grad_norm_pre_clip': 8.322479248046875, 'timing/data': 0.001181144267320633, 'timing/model': 1.1109079429879785, 'learning_rate/qwen_vl_interface': 8.113225122518854e-06, 'learning_rate/action_model': 8.113225122518854e-05, 'epoch': 0.82}) +[2026-06-17 20:48:39,261][starVLA.training.train_starvla][INFO] - Step 2130, Loss: {'train/loss': 0.21658213436603546, 'train/grad_norm_pre_clip': 6.517366409301758, 'timing/data': 0.0011627469211816788, 'timing/model': 1.1532789636403322, 'learning_rate/qwen_vl_interface': 8.111499205824148e-06, 'learning_rate/action_model': 8.111499205824147e-05, 'epoch': 0.82}) +[2026-06-17 20:48:45,501][starVLA.training.train_starvla][INFO] - Step 2131, Loss: {'train/loss': 0.49057161808013916, 'train/grad_norm_pre_clip': 8.9144287109375, 'timing/data': 0.0012590466067194939, 'timing/model': 1.317405161447823, 'learning_rate/qwen_vl_interface': 8.109772695937575e-06, 'learning_rate/action_model': 8.109772695937574e-05, 'epoch': 0.82}) +[2026-06-17 20:48:50,519][starVLA.training.train_starvla][INFO] - Step 2132, Loss: {'train/loss': 0.4598729610443115, 'train/grad_norm_pre_clip': 7.674750804901123, 'timing/data': 0.0015053637325763702, 'timing/model': 1.4674800857901573, 'learning_rate/qwen_vl_interface': 8.108045593217045e-06, 'learning_rate/action_model': 8.108045593217045e-05, 'epoch': 0.82}) +[2026-06-17 20:48:56,740][starVLA.training.train_starvla][INFO] - Step 2133, Loss: {'train/loss': 0.6679346561431885, 'train/grad_norm_pre_clip': 11.090401649475098, 'timing/data': 0.0015296367928385735, 'timing/model': 1.2602933831512928, 'learning_rate/qwen_vl_interface': 8.106317898020589e-06, 'learning_rate/action_model': 8.106317898020589e-05, 'epoch': 0.82}) +[2026-06-17 20:49:01,873][starVLA.training.train_starvla][INFO] - Step 2134, Loss: {'train/loss': 0.5961713790893555, 'train/grad_norm_pre_clip': 11.66183090209961, 'timing/data': 0.0012863660231232643, 'timing/model': 1.216404733248055, 'learning_rate/qwen_vl_interface': 8.104589610706359e-06, 'learning_rate/action_model': 8.104589610706359e-05, 'epoch': 0.82}) +[2026-06-17 20:49:09,183][starVLA.training.train_starvla][INFO] - Step 2135, Loss: {'train/loss': 0.5050110220909119, 'train/grad_norm_pre_clip': 8.17977523803711, 'timing/data': 0.000994691625237465, 'timing/model': 1.7992197945713997, 'learning_rate/qwen_vl_interface': 8.102860731632632e-06, 'learning_rate/action_model': 8.102860731632631e-05, 'epoch': 0.82}) +[2026-06-17 20:49:17,943][starVLA.training.train_starvla][INFO] - Step 2136, Loss: {'train/loss': 0.5329054594039917, 'train/grad_norm_pre_clip': 7.048564434051514, 'timing/data': 0.001163196749985218, 'timing/model': 1.2469997946172953, 'learning_rate/qwen_vl_interface': 8.101131261157802e-06, 'learning_rate/action_model': 8.101131261157802e-05, 'epoch': 0.82}) +[2026-06-17 20:49:22,923][starVLA.training.train_starvla][INFO] - Step 2137, Loss: {'train/loss': 0.697285532951355, 'train/grad_norm_pre_clip': 8.409733772277832, 'timing/data': 0.0016692252829670906, 'timing/model': 1.7453607879579067, 'learning_rate/qwen_vl_interface': 8.099401199640397e-06, 'learning_rate/action_model': 8.099401199640397e-05, 'epoch': 0.82}) +[2026-06-17 20:49:34,883][starVLA.training.train_starvla][INFO] - Step 2138, Loss: {'train/loss': 0.7034832835197449, 'train/grad_norm_pre_clip': 10.640250205993652, 'timing/data': 0.0013601798564195633, 'timing/model': 1.1443495424464345, 'learning_rate/qwen_vl_interface': 8.097670547439057e-06, 'learning_rate/action_model': 8.097670547439058e-05, 'epoch': 0.82}) +[2026-06-17 20:49:39,782][starVLA.training.train_starvla][INFO] - Step 2139, Loss: {'train/loss': 0.4854448139667511, 'train/grad_norm_pre_clip': 7.5983452796936035, 'timing/data': 0.0013798661530017853, 'timing/model': 1.1886084228754044, 'learning_rate/qwen_vl_interface': 8.095939304912549e-06, 'learning_rate/action_model': 8.095939304912548e-05, 'epoch': 0.82}) +[2026-06-17 20:49:45,500][starVLA.training.train_starvla][INFO] - Step 2140, Loss: {'train/loss': 0.43085598945617676, 'train/grad_norm_pre_clip': 7.418162822723389, 'timing/data': 0.0012172060087323189, 'timing/model': 1.6384409936144948, 'learning_rate/qwen_vl_interface': 8.094207472419759e-06, 'learning_rate/action_model': 8.094207472419759e-05, 'epoch': 0.82}) +[2026-06-17 20:49:53,330][starVLA.training.train_starvla][INFO] - Step 2141, Loss: {'train/loss': 0.43013995885849, 'train/grad_norm_pre_clip': 7.062049865722656, 'timing/data': 0.0010643098503351212, 'timing/model': 1.1628104150295258, 'learning_rate/qwen_vl_interface': 8.092475050319699e-06, 'learning_rate/action_model': 8.092475050319698e-05, 'epoch': 0.82}) +[2026-06-17 20:49:57,915][starVLA.training.train_starvla][INFO] - Step 2142, Loss: {'train/loss': 0.28707510232925415, 'train/grad_norm_pre_clip': 7.283459663391113, 'timing/data': 0.001197807490825653, 'timing/model': 1.0765347434207797, 'learning_rate/qwen_vl_interface': 8.090742038971503e-06, 'learning_rate/action_model': 8.090742038971503e-05, 'epoch': 0.82}) +[2026-06-17 20:50:03,037][starVLA.training.train_starvla][INFO] - Step 2143, Loss: {'train/loss': 0.24848392605781555, 'train/grad_norm_pre_clip': 4.60862922668457, 'timing/data': 0.0012404480949044228, 'timing/model': 1.3526082439348102, 'learning_rate/qwen_vl_interface': 8.089008438734427e-06, 'learning_rate/action_model': 8.089008438734427e-05, 'epoch': 0.82}) +[2026-06-17 20:50:08,624][starVLA.training.train_starvla][INFO] - Step 2144, Loss: {'train/loss': 0.6697216629981995, 'train/grad_norm_pre_clip': 9.428705215454102, 'timing/data': 0.0011814404278993607, 'timing/model': 1.4835055796429515, 'learning_rate/qwen_vl_interface': 8.087274249967844e-06, 'learning_rate/action_model': 8.087274249967845e-05, 'epoch': 0.82}) +[2026-06-17 20:50:13,728][starVLA.training.train_starvla][INFO] - Step 2145, Loss: {'train/loss': 0.45023879408836365, 'train/grad_norm_pre_clip': 9.247791290283203, 'timing/data': 0.001326550729572773, 'timing/model': 1.438472799025476, 'learning_rate/qwen_vl_interface': 8.08553947303126e-06, 'learning_rate/action_model': 8.085539473031259e-05, 'epoch': 0.83}) +[2026-06-17 20:50:18,977][starVLA.training.train_starvla][INFO] - Step 2146, Loss: {'train/loss': 0.8336777687072754, 'train/grad_norm_pre_clip': 13.045969009399414, 'timing/data': 0.0009194985032081604, 'timing/model': 1.8091266313567758, 'learning_rate/qwen_vl_interface': 8.083804108284288e-06, 'learning_rate/action_model': 8.083804108284287e-05, 'epoch': 0.83}) +[2026-06-17 20:50:26,229][starVLA.training.train_starvla][INFO] - Step 2147, Loss: {'train/loss': 0.6674500703811646, 'train/grad_norm_pre_clip': 9.115487098693848, 'timing/data': 0.0011942191049456596, 'timing/model': 1.5836797077208757, 'learning_rate/qwen_vl_interface': 8.082068156086675e-06, 'learning_rate/action_model': 8.082068156086675e-05, 'epoch': 0.83}) +[2026-06-17 20:50:34,243][starVLA.training.train_starvla][INFO] - Step 2148, Loss: {'train/loss': 0.31681251525878906, 'train/grad_norm_pre_clip': 6.427856922149658, 'timing/data': 0.0013224100694060326, 'timing/model': 4.437658751383424, 'learning_rate/qwen_vl_interface': 8.080331616798289e-06, 'learning_rate/action_model': 8.080331616798287e-05, 'epoch': 0.83}) +[2026-06-17 20:50:44,315][starVLA.training.train_starvla][INFO] - Step 2149, Loss: {'train/loss': 0.46615973114967346, 'train/grad_norm_pre_clip': 7.1720194816589355, 'timing/data': 0.0011920584365725517, 'timing/model': 3.817460708320141, 'learning_rate/qwen_vl_interface': 8.078594490779112e-06, 'learning_rate/action_model': 8.078594490779111e-05, 'epoch': 0.83}) +[2026-06-17 20:50:55,869][starVLA.training.train_starvla][INFO] - Step 2150, Loss: {'train/loss': 0.6383561491966248, 'train/grad_norm_pre_clip': 8.991933822631836, 'timing/data': 0.0011671008542180061, 'timing/model': 2.3854365088045597, 'learning_rate/qwen_vl_interface': 8.076856778389253e-06, 'learning_rate/action_model': 8.076856778389252e-05, 'epoch': 0.83}) +[2026-06-17 20:51:01,222][starVLA.training.train_starvla][INFO] - Step 2151, Loss: {'train/loss': 0.6064362525939941, 'train/grad_norm_pre_clip': 9.571614265441895, 'timing/data': 0.0013342369347810745, 'timing/model': 1.250756736844778, 'learning_rate/qwen_vl_interface': 8.075118479988942e-06, 'learning_rate/action_model': 8.075118479988941e-05, 'epoch': 0.83}) +[2026-06-17 20:51:07,318][starVLA.training.train_starvla][INFO] - Step 2152, Loss: {'train/loss': 0.7754102945327759, 'train/grad_norm_pre_clip': 8.36206340789795, 'timing/data': 0.0011257706210017204, 'timing/model': 1.2479840405285358, 'learning_rate/qwen_vl_interface': 8.07337959593853e-06, 'learning_rate/action_model': 8.073379595938529e-05, 'epoch': 0.83}) +[2026-06-17 20:51:12,185][starVLA.training.train_starvla][INFO] - Step 2153, Loss: {'train/loss': 0.30279022455215454, 'train/grad_norm_pre_clip': 6.065469264984131, 'timing/data': 0.0010543009266257286, 'timing/model': 1.1256624991074204, 'learning_rate/qwen_vl_interface': 8.071640126598488e-06, 'learning_rate/action_model': 8.071640126598488e-05, 'epoch': 0.83}) +[2026-06-17 20:51:17,301][starVLA.training.train_starvla][INFO] - Step 2154, Loss: {'train/loss': 0.8337363600730896, 'train/grad_norm_pre_clip': 8.596659660339355, 'timing/data': 0.0011799558997154236, 'timing/model': 1.1539882980287075, 'learning_rate/qwen_vl_interface': 8.069900072329414e-06, 'learning_rate/action_model': 8.069900072329415e-05, 'epoch': 0.83}) +[2026-06-17 20:51:22,255][starVLA.training.train_starvla][INFO] - Step 2155, Loss: {'train/loss': 0.3559441864490509, 'train/grad_norm_pre_clip': 7.058782577514648, 'timing/data': 0.0014105560258030891, 'timing/model': 1.1904233554378152, 'learning_rate/qwen_vl_interface': 8.068159433492021e-06, 'learning_rate/action_model': 8.068159433492021e-05, 'epoch': 0.83}) +[2026-06-17 20:51:27,022][starVLA.training.train_starvla][INFO] - Step 2156, Loss: {'train/loss': 0.9130879044532776, 'train/grad_norm_pre_clip': 13.180436134338379, 'timing/data': 0.0009729666635394096, 'timing/model': 1.5482577132061124, 'learning_rate/qwen_vl_interface': 8.066418210447144e-06, 'learning_rate/action_model': 8.066418210447143e-05, 'epoch': 0.83}) +[2026-06-17 20:51:31,581][starVLA.training.train_starvla][INFO] - Step 2157, Loss: {'train/loss': 0.625027060508728, 'train/grad_norm_pre_clip': 10.49142074584961, 'timing/data': 0.0009677289053797722, 'timing/model': 1.4348839363083243, 'learning_rate/qwen_vl_interface': 8.064676403555741e-06, 'learning_rate/action_model': 8.064676403555741e-05, 'epoch': 0.83}) +[2026-06-17 20:51:36,320][starVLA.training.train_starvla][INFO] - Step 2158, Loss: {'train/loss': 0.6667194366455078, 'train/grad_norm_pre_clip': 9.757834434509277, 'timing/data': 0.0011864015832543373, 'timing/model': 1.0849765045568347, 'learning_rate/qwen_vl_interface': 8.06293401317889e-06, 'learning_rate/action_model': 8.062934013178891e-05, 'epoch': 0.83}) +[2026-06-17 20:51:41,228][starVLA.training.train_starvla][INFO] - Step 2159, Loss: {'train/loss': 0.7669550776481628, 'train/grad_norm_pre_clip': 9.69857406616211, 'timing/data': 0.0008088340982794762, 'timing/model': 1.2347958330065012, 'learning_rate/qwen_vl_interface': 8.061191039677793e-06, 'learning_rate/action_model': 8.061191039677793e-05, 'epoch': 0.83}) +[2026-06-17 20:51:46,139][starVLA.training.train_starvla][INFO] - Step 2160, Loss: {'train/loss': 0.708135187625885, 'train/grad_norm_pre_clip': 9.38190746307373, 'timing/data': 0.001069740392267704, 'timing/model': 1.127028807066381, 'learning_rate/qwen_vl_interface': 8.059447483413768e-06, 'learning_rate/action_model': 8.059447483413767e-05, 'epoch': 0.83}) +[2026-06-17 20:51:50,831][starVLA.training.train_starvla][INFO] - Step 2161, Loss: {'train/loss': 0.4108744263648987, 'train/grad_norm_pre_clip': 8.92462158203125, 'timing/data': 0.0014091599732637405, 'timing/model': 1.0847585629671812, 'learning_rate/qwen_vl_interface': 8.057703344748258e-06, 'learning_rate/action_model': 8.057703344748257e-05, 'epoch': 0.83}) +[2026-06-17 20:51:55,526][starVLA.training.train_starvla][INFO] - Step 2162, Loss: {'train/loss': 0.4759935736656189, 'train/grad_norm_pre_clip': 6.478546619415283, 'timing/data': 0.0013220049440860748, 'timing/model': 1.1258914778009057, 'learning_rate/qwen_vl_interface': 8.05595862404282e-06, 'learning_rate/action_model': 8.05595862404282e-05, 'epoch': 0.83}) +[2026-06-17 20:52:00,532][starVLA.training.train_starvla][INFO] - Step 2163, Loss: {'train/loss': 0.5664733052253723, 'train/grad_norm_pre_clip': 7.349049091339111, 'timing/data': 0.0009611630812287331, 'timing/model': 1.24996402207762, 'learning_rate/qwen_vl_interface': 8.054213321659143e-06, 'learning_rate/action_model': 8.054213321659142e-05, 'epoch': 0.83}) +[2026-06-17 20:52:05,395][starVLA.training.train_starvla][INFO] - Step 2164, Loss: {'train/loss': 0.31917956471443176, 'train/grad_norm_pre_clip': 6.524020671844482, 'timing/data': 0.0013245772570371628, 'timing/model': 1.162221934646368, 'learning_rate/qwen_vl_interface': 8.052467437959025e-06, 'learning_rate/action_model': 8.052467437959025e-05, 'epoch': 0.83}) +[2026-06-17 20:52:10,547][starVLA.training.train_starvla][INFO] - Step 2165, Loss: {'train/loss': 0.5468301773071289, 'train/grad_norm_pre_clip': 7.208756923675537, 'timing/data': 0.002013077959418297, 'timing/model': 1.2404520194977522, 'learning_rate/qwen_vl_interface': 8.050720973304393e-06, 'learning_rate/action_model': 8.050720973304391e-05, 'epoch': 0.83}) +[2026-06-17 20:52:15,273][starVLA.training.train_starvla][INFO] - Step 2166, Loss: {'train/loss': 0.2977721691131592, 'train/grad_norm_pre_clip': 6.194407939910889, 'timing/data': 0.0011286772787570953, 'timing/model': 1.1350854225456715, 'learning_rate/qwen_vl_interface': 8.048973928057288e-06, 'learning_rate/action_model': 8.048973928057288e-05, 'epoch': 0.83}) +[2026-06-17 20:52:19,960][starVLA.training.train_starvla][INFO] - Step 2167, Loss: {'train/loss': 0.4883843660354614, 'train/grad_norm_pre_clip': 8.797796249389648, 'timing/data': 0.0009134458377957344, 'timing/model': 1.482860879972577, 'learning_rate/qwen_vl_interface': 8.047226302579877e-06, 'learning_rate/action_model': 8.047226302579877e-05, 'epoch': 0.83}) +[2026-06-17 20:52:26,252][starVLA.training.train_starvla][INFO] - Step 2168, Loss: {'train/loss': 0.5403999090194702, 'train/grad_norm_pre_clip': 9.856005668640137, 'timing/data': 0.0010146070271730423, 'timing/model': 2.9574136789888144, 'learning_rate/qwen_vl_interface': 8.045478097234443e-06, 'learning_rate/action_model': 8.045478097234444e-05, 'epoch': 0.83}) +[2026-06-17 20:52:33,964][starVLA.training.train_starvla][INFO] - Step 2169, Loss: {'train/loss': 0.5528265237808228, 'train/grad_norm_pre_clip': 7.732611179351807, 'timing/data': 0.0014067050069570541, 'timing/model': 1.0986533490940928, 'learning_rate/qwen_vl_interface': 8.043729312383393e-06, 'learning_rate/action_model': 8.043729312383392e-05, 'epoch': 0.83}) +[2026-06-17 20:52:41,970][starVLA.training.train_starvla][INFO] - Step 2170, Loss: {'train/loss': 0.5734376907348633, 'train/grad_norm_pre_clip': 9.04190731048584, 'timing/data': 0.0011868299916386604, 'timing/model': 2.2569254972040653, 'learning_rate/qwen_vl_interface': 8.041979948389248e-06, 'learning_rate/action_model': 8.041979948389248e-05, 'epoch': 0.83}) +[2026-06-17 20:52:48,844][starVLA.training.train_starvla][INFO] - Step 2171, Loss: {'train/loss': 0.4053308367729187, 'train/grad_norm_pre_clip': 7.651564121246338, 'timing/data': 0.0015603294596076012, 'timing/model': 3.0704902689903975, 'learning_rate/qwen_vl_interface': 8.040230005614656e-06, 'learning_rate/action_model': 8.040230005614656e-05, 'epoch': 0.84}) +[2026-06-17 20:53:01,585][starVLA.training.train_starvla][INFO] - Step 2172, Loss: {'train/loss': 0.8830463290214539, 'train/grad_norm_pre_clip': 12.136636734008789, 'timing/data': 0.0009456900879740715, 'timing/model': 3.622081716544926, 'learning_rate/qwen_vl_interface': 8.038479484422383e-06, 'learning_rate/action_model': 8.038479484422383e-05, 'epoch': 0.84}) +[2026-06-17 20:53:12,946][starVLA.training.train_starvla][INFO] - Step 2173, Loss: {'train/loss': 0.7441113591194153, 'train/grad_norm_pre_clip': 11.107563018798828, 'timing/data': 0.0008514914661645889, 'timing/model': 4.489147322252393, 'learning_rate/qwen_vl_interface': 8.03672838517531e-06, 'learning_rate/action_model': 8.036728385175311e-05, 'epoch': 0.84}) +[2026-06-17 20:53:23,342][starVLA.training.train_starvla][INFO] - Step 2174, Loss: {'train/loss': 0.6450837850570679, 'train/grad_norm_pre_clip': 7.8084282875061035, 'timing/data': 0.0011829240247607231, 'timing/model': 1.6829942027106881, 'learning_rate/qwen_vl_interface': 8.034976708236446e-06, 'learning_rate/action_model': 8.034976708236445e-05, 'epoch': 0.84}) +[2026-06-17 20:53:29,708][starVLA.training.train_starvla][INFO] - Step 2175, Loss: {'train/loss': 0.5518574714660645, 'train/grad_norm_pre_clip': 10.871517181396484, 'timing/data': 0.0009698905050754547, 'timing/model': 2.2173649752512574, 'learning_rate/qwen_vl_interface': 8.033224453968913e-06, 'learning_rate/action_model': 8.033224453968913e-05, 'epoch': 0.84}) +[2026-06-17 20:53:35,251][starVLA.training.train_starvla][INFO] - Step 2176, Loss: {'train/loss': 0.4527069330215454, 'train/grad_norm_pre_clip': 8.577735900878906, 'timing/data': 0.0028883181512355804, 'timing/model': 1.2452043117955327, 'learning_rate/qwen_vl_interface': 8.031471622735956e-06, 'learning_rate/action_model': 8.031471622735956e-05, 'epoch': 0.84}) +[2026-06-17 20:53:40,326][starVLA.training.train_starvla][INFO] - Step 2177, Loss: {'train/loss': 0.43356382846832275, 'train/grad_norm_pre_clip': 6.0149827003479, 'timing/data': 0.0012101158499717712, 'timing/model': 1.5828088521957397, 'learning_rate/qwen_vl_interface': 8.029718214900937e-06, 'learning_rate/action_model': 8.029718214900937e-05, 'epoch': 0.84}) +[2026-06-17 20:53:44,705][starVLA.training.train_starvla][INFO] - Step 2178, Loss: {'train/loss': 0.49444547295570374, 'train/grad_norm_pre_clip': 8.804744720458984, 'timing/data': 0.0009867073968052864, 'timing/model': 1.091534225270152, 'learning_rate/qwen_vl_interface': 8.027964230827342e-06, 'learning_rate/action_model': 8.027964230827342e-05, 'epoch': 0.84}) +[2026-06-17 20:53:49,750][starVLA.training.train_starvla][INFO] - Step 2179, Loss: {'train/loss': 0.7069408893585205, 'train/grad_norm_pre_clip': 11.14992904663086, 'timing/data': 0.0011542821303009987, 'timing/model': 1.2605149373412132, 'learning_rate/qwen_vl_interface': 8.026209670878776e-06, 'learning_rate/action_model': 8.026209670878775e-05, 'epoch': 0.84}) +[2026-06-17 20:53:54,672][starVLA.training.train_starvla][INFO] - Step 2180, Loss: {'train/loss': 0.5657914280891418, 'train/grad_norm_pre_clip': 7.015567302703857, 'timing/data': 0.001239071600139141, 'timing/model': 1.2765725459903479, 'learning_rate/qwen_vl_interface': 8.024454535418952e-06, 'learning_rate/action_model': 8.024454535418951e-05, 'epoch': 0.84}) +[2026-06-17 20:53:59,880][starVLA.training.train_starvla][INFO] - Step 2181, Loss: {'train/loss': 0.3448849320411682, 'train/grad_norm_pre_clip': 6.493546009063721, 'timing/data': 0.0018306737765669823, 'timing/model': 1.385773898102343, 'learning_rate/qwen_vl_interface': 8.02269882481172e-06, 'learning_rate/action_model': 8.022698824811721e-05, 'epoch': 0.84}) +[2026-06-17 20:54:04,860][starVLA.training.train_starvla][INFO] - Step 2182, Loss: {'train/loss': 0.37646186351776123, 'train/grad_norm_pre_clip': 8.153594017028809, 'timing/data': 0.0011921534314751625, 'timing/model': 1.227080118842423, 'learning_rate/qwen_vl_interface': 8.02094253942104e-06, 'learning_rate/action_model': 8.02094253942104e-05, 'epoch': 0.84}) +[2026-06-17 20:54:11,927][starVLA.training.train_starvla][INFO] - Step 2183, Loss: {'train/loss': 0.6132839322090149, 'train/grad_norm_pre_clip': 8.918275833129883, 'timing/data': 0.001440003514289856, 'timing/model': 1.0968686109408736, 'learning_rate/qwen_vl_interface': 8.019185679610988e-06, 'learning_rate/action_model': 8.019185679610988e-05, 'epoch': 0.84}) +[2026-06-17 20:54:16,528][starVLA.training.train_starvla][INFO] - Step 2184, Loss: {'train/loss': 0.5667828917503357, 'train/grad_norm_pre_clip': 8.003290176391602, 'timing/data': 0.0009593823924660683, 'timing/model': 1.074161727912724, 'learning_rate/qwen_vl_interface': 8.017428245745765e-06, 'learning_rate/action_model': 8.017428245745763e-05, 'epoch': 0.84}) +[2026-06-17 20:54:21,139][starVLA.training.train_starvla][INFO] - Step 2185, Loss: {'train/loss': 0.5549149513244629, 'train/grad_norm_pre_clip': 7.642599582672119, 'timing/data': 0.0009721685200929642, 'timing/model': 1.1081987228244543, 'learning_rate/qwen_vl_interface': 8.015670238189688e-06, 'learning_rate/action_model': 8.015670238189688e-05, 'epoch': 0.84}) +[2026-06-17 20:54:26,019][starVLA.training.train_starvla][INFO] - Step 2186, Loss: {'train/loss': 0.44822967052459717, 'train/grad_norm_pre_clip': 9.635963439941406, 'timing/data': 0.0032169120386242867, 'timing/model': 1.0931980079039931, 'learning_rate/qwen_vl_interface': 8.013911657307196e-06, 'learning_rate/action_model': 8.013911657307196e-05, 'epoch': 0.84}) +[2026-06-17 20:54:30,716][starVLA.training.train_starvla][INFO] - Step 2187, Loss: {'train/loss': 0.6386741399765015, 'train/grad_norm_pre_clip': 13.9027738571167, 'timing/data': 0.0011997586116194725, 'timing/model': 1.4332686457782984, 'learning_rate/qwen_vl_interface': 8.012152503462843e-06, 'learning_rate/action_model': 8.012152503462842e-05, 'epoch': 0.84}) +[2026-06-17 20:54:35,535][starVLA.training.train_starvla][INFO] - Step 2188, Loss: {'train/loss': 0.40126699209213257, 'train/grad_norm_pre_clip': 6.660190582275391, 'timing/data': 0.0013567442074418068, 'timing/model': 1.2033901363611221, 'learning_rate/qwen_vl_interface': 8.010392777021304e-06, 'learning_rate/action_model': 8.010392777021303e-05, 'epoch': 0.84}) +[2026-06-17 20:54:43,194][starVLA.training.train_starvla][INFO] - Step 2189, Loss: {'train/loss': 0.4725804328918457, 'train/grad_norm_pre_clip': 7.861132621765137, 'timing/data': 0.0011361874639987946, 'timing/model': 1.0824029743671417, 'learning_rate/qwen_vl_interface': 8.008632478347372e-06, 'learning_rate/action_model': 8.008632478347371e-05, 'epoch': 0.84}) +[2026-06-17 20:54:59,340][starVLA.training.train_starvla][INFO] - Step 2190, Loss: {'train/loss': 0.3645927309989929, 'train/grad_norm_pre_clip': 7.377368450164795, 'timing/data': 0.00237210001796484, 'timing/model': 1.1352519411593676, 'learning_rate/qwen_vl_interface': 8.006871607805957e-06, 'learning_rate/action_model': 8.006871607805957e-05, 'epoch': 0.84}) +[2026-06-17 20:55:04,154][starVLA.training.train_starvla][INFO] - Step 2191, Loss: {'train/loss': 0.4475305378437042, 'train/grad_norm_pre_clip': 7.542264938354492, 'timing/data': 0.0011817263439297676, 'timing/model': 1.1626816932111979, 'learning_rate/qwen_vl_interface': 8.005110165762093e-06, 'learning_rate/action_model': 8.005110165762093e-05, 'epoch': 0.84}) +[2026-06-17 20:55:08,803][starVLA.training.train_starvla][INFO] - Step 2192, Loss: {'train/loss': 0.39947327971458435, 'train/grad_norm_pre_clip': 6.1622161865234375, 'timing/data': 0.0022080987691879272, 'timing/model': 1.1045458428561687, 'learning_rate/qwen_vl_interface': 8.003348152580927e-06, 'learning_rate/action_model': 8.003348152580927e-05, 'epoch': 0.84}) +[2026-06-17 20:55:13,866][starVLA.training.train_starvla][INFO] - Step 2193, Loss: {'train/loss': 0.5267956256866455, 'train/grad_norm_pre_clip': 7.201656341552734, 'timing/data': 0.0012281909584999084, 'timing/model': 1.1003772486001253, 'learning_rate/qwen_vl_interface': 8.001585568627724e-06, 'learning_rate/action_model': 8.001585568627725e-05, 'epoch': 0.84}) +[2026-06-17 20:55:20,499][starVLA.training.train_starvla][INFO] - Step 2194, Loss: {'train/loss': 0.347019225358963, 'train/grad_norm_pre_clip': 9.669695854187012, 'timing/data': 0.00109134241938591, 'timing/model': 1.149000633507967, 'learning_rate/qwen_vl_interface': 7.999822414267874e-06, 'learning_rate/action_model': 7.999822414267874e-05, 'epoch': 0.84}) +[2026-06-17 20:55:28,509][starVLA.training.train_starvla][INFO] - Step 2195, Loss: {'train/loss': 0.60507732629776, 'train/grad_norm_pre_clip': 9.493643760681152, 'timing/data': 0.001406804658472538, 'timing/model': 1.096911865286529, 'learning_rate/qwen_vl_interface': 7.998058689866879e-06, 'learning_rate/action_model': 7.998058689866879e-05, 'epoch': 0.84}) +[2026-06-17 20:55:33,681][starVLA.training.train_starvla][INFO] - Step 2196, Loss: {'train/loss': 0.5881791710853577, 'train/grad_norm_pre_clip': 8.107375144958496, 'timing/data': 0.001300162635743618, 'timing/model': 1.2082435665652156, 'learning_rate/qwen_vl_interface': 7.99629439579036e-06, 'learning_rate/action_model': 7.99629439579036e-05, 'epoch': 0.84}) +[2026-06-17 20:55:38,485][starVLA.training.train_starvla][INFO] - Step 2197, Loss: {'train/loss': 0.47716420888900757, 'train/grad_norm_pre_clip': 8.163930892944336, 'timing/data': 0.0008424753323197365, 'timing/model': 1.432449440471828, 'learning_rate/qwen_vl_interface': 7.994529532404059e-06, 'learning_rate/action_model': 7.994529532404058e-05, 'epoch': 0.85}) +[2026-06-17 20:55:47,904][starVLA.training.train_starvla][INFO] - Step 2198, Loss: {'train/loss': 0.3960265815258026, 'train/grad_norm_pre_clip': 7.766133785247803, 'timing/data': 0.0019158869981765747, 'timing/model': 1.1445951601490378, 'learning_rate/qwen_vl_interface': 7.99276410007383e-06, 'learning_rate/action_model': 7.99276410007383e-05, 'epoch': 0.85}) +[2026-06-17 20:55:52,662][starVLA.training.train_starvla][INFO] - Step 2199, Loss: {'train/loss': 0.4324604272842407, 'train/grad_norm_pre_clip': 6.080982685089111, 'timing/data': 0.0013817744329571724, 'timing/model': 1.12074950709939, 'learning_rate/qwen_vl_interface': 7.990998099165652e-06, 'learning_rate/action_model': 7.990998099165652e-05, 'epoch': 0.85}) +[2026-06-17 20:59:31,285][starVLA.training.train_starvla][INFO] - Step 2200, Loss: {'train/loss': 0.4041174650192261, 'train/grad_norm_pre_clip': 6.583171844482422, 'eval/loss': 0.5429030060768127, 'eval/latency_4/loss': 0.5429030060768127, 'eval/demon_fire_precision': 0.9498442367601246, 'eval/demon_fire_recall': 0.9902565768106528, 'eval/demon_fire_f1': 0.9696295118460804, 'eval/demon_fire_rate_model': 0.9427312775330396, 'eval/demon_fire_rate_teacher': 0.9042584434654919, 'eval/demon_left_precision': 0.6971153846153846, 'eval/demon_left_recall': 0.6621004566210046, 'eval/demon_left_f1': 0.6791569086651054, 'eval/demon_left_rate_model': 0.12217327459618209, 'eval/demon_left_rate_teacher': 0.12863436123348018, 'eval/demon_right_precision': 0.8309859154929577, 'eval/demon_right_recall': 0.6389891696750902, 'eval/demon_right_f1': 0.7224489795918367, 'eval/demon_right_rate_model': 0.12511013215859032, 'eval/demon_right_rate_teacher': 0.16270190895741557, 'eval/demon_move_f1': 0.7008029441284711, 'eval/demon_cc_f1': 0.869560971848469, 'eval/latency_4/demon_fire_precision': 0.9498442367601246, 'eval/latency_4/demon_fire_recall': 0.9902565768106528, 'eval/latency_4/demon_fire_f1': 0.9696295118460804, 'eval/latency_4/demon_fire_rate_model': 0.9427312775330396, 'eval/latency_4/demon_fire_rate_teacher': 0.9042584434654919, 'eval/latency_4/demon_left_precision': 0.6971153846153846, 'eval/latency_4/demon_left_recall': 0.6621004566210046, 'eval/latency_4/demon_left_f1': 0.6791569086651054, 'eval/latency_4/demon_left_rate_model': 0.12217327459618209, 'eval/latency_4/demon_left_rate_teacher': 0.12863436123348018, 'eval/latency_4/demon_right_precision': 0.8309859154929577, 'eval/latency_4/demon_right_recall': 0.6389891696750902, 'eval/latency_4/demon_right_f1': 0.7224489795918367, 'eval/latency_4/demon_right_rate_model': 0.12511013215859032, 'eval/latency_4/demon_right_rate_teacher': 0.16270190895741557, 'eval/latency_4/demon_move_f1': 0.7008029441284711, 'eval/latency_4/demon_cc_f1': 0.869560971848469, 'timing/data': 0.0015135332942008972, 'timing/model': 1.0715654082596302, 'learning_rate/qwen_vl_interface': 7.98923153004562e-06, 'learning_rate/action_model': 7.98923153004562e-05, 'epoch': 0.85}) +[2026-06-17 20:59:31,287][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_2200_state +[2026-06-17 20:59:31,295][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 20:59:39,902][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_2200_state/model.safetensors +[2026-06-17 21:00:01,996][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_2200_state/optimizer.bin +[2026-06-17 21:00:02,013][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_2200_state/sampler.bin +[2026-06-17 21:00:02,013][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_2200_state/sampler_1.bin +[2026-06-17 21:00:02,017][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_2200_state/random_states_0.pkl +[2026-06-17 21:00:02,024][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2200 +[2026-06-17 21:00:02,025][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 21:00:02,058][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 21:00:10,761][starVLA.training.train_starvla][INFO] - Step 2201, Loss: {'train/loss': 0.5146670937538147, 'train/grad_norm_pre_clip': 7.848415851593018, 'timing/data': 0.0014142189174890518, 'timing/model': 4.357167174108326, 'learning_rate/qwen_vl_interface': 7.987464393079943e-06, 'learning_rate/action_model': 7.987464393079943e-05, 'epoch': 0.85}) +[2026-06-17 21:00:17,579][starVLA.training.train_starvla][INFO] - Step 2202, Loss: {'train/loss': 0.5093821883201599, 'train/grad_norm_pre_clip': 8.831339836120605, 'timing/data': 0.0008327467367053032, 'timing/model': 2.4134369688108563, 'learning_rate/qwen_vl_interface': 7.985696688634953e-06, 'learning_rate/action_model': 7.985696688634952e-05, 'epoch': 0.85}) +[2026-06-17 21:00:22,502][starVLA.training.train_starvla][INFO] - Step 2203, Loss: {'train/loss': 0.6695992946624756, 'train/grad_norm_pre_clip': 11.711784362792969, 'timing/data': 0.0012581031769514084, 'timing/model': 1.0784905515611172, 'learning_rate/qwen_vl_interface': 7.983928417077095e-06, 'learning_rate/action_model': 7.983928417077095e-05, 'epoch': 0.85}) +[2026-06-17 21:00:30,347][starVLA.training.train_starvla][INFO] - Step 2204, Loss: {'train/loss': 0.4992131292819977, 'train/grad_norm_pre_clip': 10.781231880187988, 'timing/data': 0.001017056405544281, 'timing/model': 1.1315820962190628, 'learning_rate/qwen_vl_interface': 7.982159578772935e-06, 'learning_rate/action_model': 7.982159578772934e-05, 'epoch': 0.85}) +[2026-06-17 21:00:37,269][starVLA.training.train_starvla][INFO] - Step 2205, Loss: {'train/loss': 0.5315539836883545, 'train/grad_norm_pre_clip': 8.957983016967773, 'timing/data': 0.0010001715272665024, 'timing/model': 3.0497559662908316, 'learning_rate/qwen_vl_interface': 7.980390174089154e-06, 'learning_rate/action_model': 7.980390174089153e-05, 'epoch': 0.85}) +[2026-06-17 21:00:41,923][starVLA.training.train_starvla][INFO] - Step 2206, Loss: {'train/loss': 0.4127335548400879, 'train/grad_norm_pre_clip': 9.125319480895996, 'timing/data': 0.0010947370901703835, 'timing/model': 1.0701599894091487, 'learning_rate/qwen_vl_interface': 7.978620203392551e-06, 'learning_rate/action_model': 7.97862020339255e-05, 'epoch': 0.85}) +[2026-06-17 21:00:46,961][starVLA.training.train_starvla][INFO] - Step 2207, Loss: {'train/loss': 0.6059141755104065, 'train/grad_norm_pre_clip': 11.975507736206055, 'timing/data': 0.0013026036322116852, 'timing/model': 1.4806144516915083, 'learning_rate/qwen_vl_interface': 7.976849667050044e-06, 'learning_rate/action_model': 7.976849667050044e-05, 'epoch': 0.85}) +[2026-06-17 21:01:01,198][starVLA.training.train_starvla][INFO] - Step 2208, Loss: {'train/loss': 0.37952566146850586, 'train/grad_norm_pre_clip': 14.47230052947998, 'timing/data': 0.0012111589312553406, 'timing/model': 4.296211215667427, 'learning_rate/qwen_vl_interface': 7.975078565428667e-06, 'learning_rate/action_model': 7.975078565428668e-05, 'epoch': 0.85}) +[2026-06-17 21:01:19,029][starVLA.training.train_starvla][INFO] - Step 2209, Loss: {'train/loss': 0.40022408962249756, 'train/grad_norm_pre_clip': 8.566390991210938, 'timing/data': 0.0013783704489469528, 'timing/model': 4.651969066821039, 'learning_rate/qwen_vl_interface': 7.97330689889557e-06, 'learning_rate/action_model': 7.97330689889557e-05, 'epoch': 0.85}) +[2026-06-17 21:01:24,266][starVLA.training.train_starvla][INFO] - Step 2210, Loss: {'train/loss': 0.5119268298149109, 'train/grad_norm_pre_clip': 10.22204303741455, 'timing/data': 0.0013379883021116257, 'timing/model': 1.2909844676032662, 'learning_rate/qwen_vl_interface': 7.971534667818026e-06, 'learning_rate/action_model': 7.971534667818025e-05, 'epoch': 0.85}) +[2026-06-17 21:01:32,280][starVLA.training.train_starvla][INFO] - Step 2211, Loss: {'train/loss': 0.6566486954689026, 'train/grad_norm_pre_clip': 10.043230056762695, 'timing/data': 0.0011762259528040886, 'timing/model': 4.1832054844126105, 'learning_rate/qwen_vl_interface': 7.969761872563413e-06, 'learning_rate/action_model': 7.969761872563413e-05, 'epoch': 0.85}) +[2026-06-17 21:01:44,357][starVLA.training.train_starvla][INFO] - Step 2212, Loss: {'train/loss': 0.42701902985572815, 'train/grad_norm_pre_clip': 7.675614833831787, 'timing/data': 0.0011661238968372345, 'timing/model': 1.4527365323156118, 'learning_rate/qwen_vl_interface': 7.96798851349924e-06, 'learning_rate/action_model': 7.96798851349924e-05, 'epoch': 0.85}) +[2026-06-17 21:01:56,039][starVLA.training.train_starvla][INFO] - Step 2213, Loss: {'train/loss': 0.6984401941299438, 'train/grad_norm_pre_clip': 8.694393157958984, 'timing/data': 0.001324477605521679, 'timing/model': 1.2390927504748106, 'learning_rate/qwen_vl_interface': 7.966214590993122e-06, 'learning_rate/action_model': 7.966214590993122e-05, 'epoch': 0.85}) +[2026-06-17 21:02:05,193][starVLA.training.train_starvla][INFO] - Step 2214, Loss: {'train/loss': 0.570134162902832, 'train/grad_norm_pre_clip': 10.614200592041016, 'timing/data': 0.0009406497702002525, 'timing/model': 1.227187612093985, 'learning_rate/qwen_vl_interface': 7.964440105412797e-06, 'learning_rate/action_model': 7.964440105412796e-05, 'epoch': 0.85}) +[2026-06-17 21:02:10,011][starVLA.training.train_starvla][INFO] - Step 2215, Loss: {'train/loss': 0.6041013598442078, 'train/grad_norm_pre_clip': 9.912576675415039, 'timing/data': 0.0024220356717705727, 'timing/model': 1.0879677338525653, 'learning_rate/qwen_vl_interface': 7.962665057126116e-06, 'learning_rate/action_model': 7.962665057126117e-05, 'epoch': 0.85}) +[2026-06-17 21:02:14,688][starVLA.training.train_starvla][INFO] - Step 2216, Loss: {'train/loss': 0.35709214210510254, 'train/grad_norm_pre_clip': 5.341444492340088, 'timing/data': 0.00187753327190876, 'timing/model': 1.1342080691829324, 'learning_rate/qwen_vl_interface': 7.960889446501052e-06, 'learning_rate/action_model': 7.960889446501052e-05, 'epoch': 0.85}) +[2026-06-17 21:02:23,890][starVLA.training.train_starvla][INFO] - Step 2217, Loss: {'train/loss': 0.7849813103675842, 'train/grad_norm_pre_clip': 11.29755973815918, 'timing/data': 0.0009101433679461479, 'timing/model': 1.4792001452296972, 'learning_rate/qwen_vl_interface': 7.959113273905688e-06, 'learning_rate/action_model': 7.959113273905688e-05, 'epoch': 0.85}) +[2026-06-17 21:02:28,542][starVLA.training.train_starvla][INFO] - Step 2218, Loss: {'train/loss': 0.772500216960907, 'train/grad_norm_pre_clip': 9.973054885864258, 'timing/data': 0.0009742416441440582, 'timing/model': 1.3721276186406612, 'learning_rate/qwen_vl_interface': 7.957336539708226e-06, 'learning_rate/action_model': 7.957336539708225e-05, 'epoch': 0.85}) +[2026-06-17 21:02:33,829][starVLA.training.train_starvla][INFO] - Step 2219, Loss: {'train/loss': 0.24738603830337524, 'train/grad_norm_pre_clip': 5.456048965454102, 'timing/data': 0.0010986290872097015, 'timing/model': 1.1207404863089323, 'learning_rate/qwen_vl_interface': 7.955559244276987e-06, 'learning_rate/action_model': 7.955559244276987e-05, 'epoch': 0.85}) +[2026-06-17 21:02:39,208][starVLA.training.train_starvla][INFO] - Step 2220, Loss: {'train/loss': 0.6915441751480103, 'train/grad_norm_pre_clip': 8.765225410461426, 'timing/data': 0.0014652814716100693, 'timing/model': 1.4427091600373387, 'learning_rate/qwen_vl_interface': 7.953781387980404e-06, 'learning_rate/action_model': 7.953781387980404e-05, 'epoch': 0.85}) +[2026-06-17 21:02:44,347][starVLA.training.train_starvla][INFO] - Step 2221, Loss: {'train/loss': 0.3887927234172821, 'train/grad_norm_pre_clip': 7.637292861938477, 'timing/data': 0.001183866523206234, 'timing/model': 1.1129386434331536, 'learning_rate/qwen_vl_interface': 7.952002971187032e-06, 'learning_rate/action_model': 7.952002971187032e-05, 'epoch': 0.85}) +[2026-06-17 21:02:49,315][starVLA.training.train_starvla][INFO] - Step 2222, Loss: {'train/loss': 0.5237248539924622, 'train/grad_norm_pre_clip': 11.113761901855469, 'timing/data': 0.003085620701313019, 'timing/model': 1.5435049496591091, 'learning_rate/qwen_vl_interface': 7.950223994265534e-06, 'learning_rate/action_model': 7.950223994265533e-05, 'epoch': 0.85}) +[2026-06-17 21:02:53,791][starVLA.training.train_starvla][INFO] - Step 2223, Loss: {'train/loss': 0.5085035562515259, 'train/grad_norm_pre_clip': 6.187657356262207, 'timing/data': 0.0013360287994146347, 'timing/model': 1.0806818781420588, 'learning_rate/qwen_vl_interface': 7.9484444575847e-06, 'learning_rate/action_model': 7.948444457584699e-05, 'epoch': 0.86}) +[2026-06-17 21:03:00,226][starVLA.training.train_starvla][INFO] - Step 2224, Loss: {'train/loss': 0.48014503717422485, 'train/grad_norm_pre_clip': 9.038033485412598, 'timing/data': 0.0013364320620894432, 'timing/model': 1.1220194855704904, 'learning_rate/qwen_vl_interface': 7.946664361513423e-06, 'learning_rate/action_model': 7.946664361513421e-05, 'epoch': 0.86}) +[2026-06-17 21:03:04,812][starVLA.training.train_starvla][INFO] - Step 2225, Loss: {'train/loss': 0.5161062479019165, 'train/grad_norm_pre_clip': 9.155466079711914, 'timing/data': 0.0011808983981609344, 'timing/model': 1.0827341387048364, 'learning_rate/qwen_vl_interface': 7.944883706420724e-06, 'learning_rate/action_model': 7.944883706420724e-05, 'epoch': 0.86}) +[2026-06-17 21:03:09,734][starVLA.training.train_starvla][INFO] - Step 2226, Loss: {'train/loss': 0.3051568567752838, 'train/grad_norm_pre_clip': 8.719069480895996, 'timing/data': 0.0013203686103224754, 'timing/model': 1.074043294414878, 'learning_rate/qwen_vl_interface': 7.943102492675732e-06, 'learning_rate/action_model': 7.943102492675732e-05, 'epoch': 0.86}) +[2026-06-17 21:03:14,510][starVLA.training.train_starvla][INFO] - Step 2227, Loss: {'train/loss': 0.32886791229248047, 'train/grad_norm_pre_clip': 6.476786136627197, 'timing/data': 0.0009083263576030731, 'timing/model': 1.5334346108138561, 'learning_rate/qwen_vl_interface': 7.941320720647694e-06, 'learning_rate/action_model': 7.941320720647695e-05, 'epoch': 0.86}) +[2026-06-17 21:03:23,684][starVLA.training.train_starvla][INFO] - Step 2228, Loss: {'train/loss': 0.41699689626693726, 'train/grad_norm_pre_clip': 10.449443817138672, 'timing/data': 0.000956876203417778, 'timing/model': 1.2067925473675132, 'learning_rate/qwen_vl_interface': 7.939538390705976e-06, 'learning_rate/action_model': 7.939538390705976e-05, 'epoch': 0.86}) +[2026-06-17 21:03:28,377][starVLA.training.train_starvla][INFO] - Step 2229, Loss: {'train/loss': 0.5615492463111877, 'train/grad_norm_pre_clip': 11.869073867797852, 'timing/data': 0.001359580084681511, 'timing/model': 1.1767310034483671, 'learning_rate/qwen_vl_interface': 7.937755503220057e-06, 'learning_rate/action_model': 7.937755503220056e-05, 'epoch': 0.86}) +[2026-06-17 21:03:33,287][starVLA.training.train_starvla][INFO] - Step 2230, Loss: {'train/loss': 0.6898974180221558, 'train/grad_norm_pre_clip': 11.516921997070312, 'timing/data': 0.0018684491515159607, 'timing/model': 1.2956942832097411, 'learning_rate/qwen_vl_interface': 7.93597205855953e-06, 'learning_rate/action_model': 7.935972058559529e-05, 'epoch': 0.86}) +[2026-06-17 21:03:46,587][starVLA.training.train_starvla][INFO] - Step 2231, Loss: {'train/loss': 0.85394287109375, 'train/grad_norm_pre_clip': 10.030211448669434, 'timing/data': 0.0009568985551595688, 'timing/model': 1.3664253847673535, 'learning_rate/qwen_vl_interface': 7.934188057094103e-06, 'learning_rate/action_model': 7.934188057094103e-05, 'epoch': 0.86}) +[2026-06-17 21:03:52,175][starVLA.training.train_starvla][INFO] - Step 2232, Loss: {'train/loss': 0.5175046920776367, 'train/grad_norm_pre_clip': 8.044585227966309, 'timing/data': 0.0012133875861763954, 'timing/model': 1.7801854303106666, 'learning_rate/qwen_vl_interface': 7.932403499193604e-06, 'learning_rate/action_model': 7.932403499193605e-05, 'epoch': 0.86}) +[2026-06-17 21:03:57,124][starVLA.training.train_starvla][INFO] - Step 2233, Loss: {'train/loss': 0.5722052454948425, 'train/grad_norm_pre_clip': 8.07541561126709, 'timing/data': 0.001176244579255581, 'timing/model': 1.5688150087371469, 'learning_rate/qwen_vl_interface': 7.930618385227975e-06, 'learning_rate/action_model': 7.930618385227973e-05, 'epoch': 0.86}) +[2026-06-17 21:04:02,052][starVLA.training.train_starvla][INFO] - Step 2234, Loss: {'train/loss': 0.5185425877571106, 'train/grad_norm_pre_clip': 7.5356669425964355, 'timing/data': 0.0013782931491732597, 'timing/model': 1.1758792931213975, 'learning_rate/qwen_vl_interface': 7.928832715567268e-06, 'learning_rate/action_model': 7.928832715567267e-05, 'epoch': 0.86}) +[2026-06-17 21:04:07,331][starVLA.training.train_starvla][INFO] - Step 2235, Loss: {'train/loss': 0.8198574781417847, 'train/grad_norm_pre_clip': 8.518210411071777, 'timing/data': 0.0012597115710377693, 'timing/model': 1.3257661694660783, 'learning_rate/qwen_vl_interface': 7.927046490581655e-06, 'learning_rate/action_model': 7.927046490581655e-05, 'epoch': 0.86}) +[2026-06-17 21:04:15,323][starVLA.training.train_starvla][INFO] - Step 2236, Loss: {'train/loss': 0.3676946461200714, 'train/grad_norm_pre_clip': 8.033066749572754, 'timing/data': 0.0012372871860861778, 'timing/model': 1.117111605592072, 'learning_rate/qwen_vl_interface': 7.925259710641428e-06, 'learning_rate/action_model': 7.925259710641427e-05, 'epoch': 0.86}) +[2026-06-17 21:04:20,450][starVLA.training.train_starvla][INFO] - Step 2237, Loss: {'train/loss': 0.49526840448379517, 'train/grad_norm_pre_clip': 9.03557300567627, 'timing/data': 0.0013952786102890968, 'timing/model': 1.7939176969230175, 'learning_rate/qwen_vl_interface': 7.923472376116982e-06, 'learning_rate/action_model': 7.923472376116982e-05, 'epoch': 0.86}) +[2026-06-17 21:04:31,771][starVLA.training.train_starvla][INFO] - Step 2238, Loss: {'train/loss': 0.5337762832641602, 'train/grad_norm_pre_clip': 10.118643760681152, 'timing/data': 0.001236533746123314, 'timing/model': 4.353435496799648, 'learning_rate/qwen_vl_interface': 7.921684487378836e-06, 'learning_rate/action_model': 7.921684487378837e-05, 'epoch': 0.86}) +[2026-06-17 21:04:38,480][starVLA.training.train_starvla][INFO] - Step 2239, Loss: {'train/loss': 0.5525219440460205, 'train/grad_norm_pre_clip': 8.9285888671875, 'timing/data': 0.0011426275596022606, 'timing/model': 1.1938965190201998, 'learning_rate/qwen_vl_interface': 7.919896044797622e-06, 'learning_rate/action_model': 7.91989604479762e-05, 'epoch': 0.86}) +[2026-06-17 21:04:42,936][starVLA.training.train_starvla][INFO] - Step 2240, Loss: {'train/loss': 0.19895291328430176, 'train/grad_norm_pre_clip': 3.759826183319092, 'timing/data': 0.0011194050312042236, 'timing/model': 1.0533794527873397, 'learning_rate/qwen_vl_interface': 7.918107048744083e-06, 'learning_rate/action_model': 7.918107048744083e-05, 'epoch': 0.86}) +[2026-06-17 21:04:47,295][starVLA.training.train_starvla][INFO] - Step 2241, Loss: {'train/loss': 0.7134525775909424, 'train/grad_norm_pre_clip': 11.476179122924805, 'timing/data': 0.0010659871622920036, 'timing/model': 1.1050869850441813, 'learning_rate/qwen_vl_interface': 7.916317499589083e-06, 'learning_rate/action_model': 7.916317499589083e-05, 'epoch': 0.86}) +[2026-06-17 21:04:52,009][starVLA.training.train_starvla][INFO] - Step 2242, Loss: {'train/loss': 0.5993262529373169, 'train/grad_norm_pre_clip': 10.020509719848633, 'timing/data': 0.0010566497221589088, 'timing/model': 1.2685255836695433, 'learning_rate/qwen_vl_interface': 7.914527397703596e-06, 'learning_rate/action_model': 7.914527397703595e-05, 'epoch': 0.86}) +[2026-06-17 21:04:59,317][starVLA.training.train_starvla][INFO] - Step 2243, Loss: {'train/loss': 0.49726447463035583, 'train/grad_norm_pre_clip': 7.2218194007873535, 'timing/data': 0.0013429922983050346, 'timing/model': 2.53615490719676, 'learning_rate/qwen_vl_interface': 7.912736743458712e-06, 'learning_rate/action_model': 7.912736743458711e-05, 'epoch': 0.86}) +[2026-06-17 21:05:03,627][starVLA.training.train_starvla][INFO] - Step 2244, Loss: {'train/loss': 0.5741741061210632, 'train/grad_norm_pre_clip': 7.497354984283447, 'timing/data': 0.0010425271466374397, 'timing/model': 1.0355694890022278, 'learning_rate/qwen_vl_interface': 7.910945537225638e-06, 'learning_rate/action_model': 7.910945537225637e-05, 'epoch': 0.86}) +[2026-06-17 21:05:07,873][starVLA.training.train_starvla][INFO] - Step 2245, Loss: {'train/loss': 0.5833169221878052, 'train/grad_norm_pre_clip': 7.426580905914307, 'timing/data': 0.0009829811751842499, 'timing/model': 1.039261774159968, 'learning_rate/qwen_vl_interface': 7.909153779375688e-06, 'learning_rate/action_model': 7.909153779375688e-05, 'epoch': 0.86}) +[2026-06-17 21:05:12,113][starVLA.training.train_starvla][INFO] - Step 2246, Loss: {'train/loss': 0.2421930730342865, 'train/grad_norm_pre_clip': 4.109988689422607, 'timing/data': 0.0009667398408055305, 'timing/model': 1.0397031838074327, 'learning_rate/qwen_vl_interface': 7.907361470280299e-06, 'learning_rate/action_model': 7.907361470280299e-05, 'epoch': 0.86}) +[2026-06-17 21:05:16,364][starVLA.training.train_starvla][INFO] - Step 2247, Loss: {'train/loss': 0.6109264492988586, 'train/grad_norm_pre_clip': 8.50147819519043, 'timing/data': 0.0011697877198457718, 'timing/model': 1.2540978649631143, 'learning_rate/qwen_vl_interface': 7.905568610311018e-06, 'learning_rate/action_model': 7.905568610311017e-05, 'epoch': 0.86}) +[2026-06-17 21:05:20,433][starVLA.training.train_starvla][INFO] - Step 2248, Loss: {'train/loss': 0.38816866278648376, 'train/grad_norm_pre_clip': 7.137967586517334, 'timing/data': 0.0010620281100273132, 'timing/model': 1.0553942304104567, 'learning_rate/qwen_vl_interface': 7.903775199839506e-06, 'learning_rate/action_model': 7.903775199839505e-05, 'epoch': 0.86}) +[2026-06-17 21:05:24,692][starVLA.training.train_starvla][INFO] - Step 2249, Loss: {'train/loss': 0.7061878442764282, 'train/grad_norm_pre_clip': 8.630221366882324, 'timing/data': 0.0010155308991670609, 'timing/model': 1.0489297276362777, 'learning_rate/qwen_vl_interface': 7.901981239237539e-06, 'learning_rate/action_model': 7.901981239237539e-05, 'epoch': 0.87}) +[2026-06-17 21:05:29,148][starVLA.training.train_starvla][INFO] - Step 2250, Loss: {'train/loss': 0.4183192849159241, 'train/grad_norm_pre_clip': 7.800562381744385, 'timing/data': 0.001164780929684639, 'timing/model': 1.0709437252953649, 'learning_rate/qwen_vl_interface': 7.900186728877009e-06, 'learning_rate/action_model': 7.90018672887701e-05, 'epoch': 0.87}) +[2026-06-17 21:05:33,953][starVLA.training.train_starvla][INFO] - Step 2251, Loss: {'train/loss': 0.5822800397872925, 'train/grad_norm_pre_clip': 7.996401786804199, 'timing/data': 0.0012872125953435898, 'timing/model': 1.1145350756123662, 'learning_rate/qwen_vl_interface': 7.898391669129919e-06, 'learning_rate/action_model': 7.898391669129918e-05, 'epoch': 0.87}) +[2026-06-17 21:05:38,891][starVLA.training.train_starvla][INFO] - Step 2252, Loss: {'train/loss': 0.6409890651702881, 'train/grad_norm_pre_clip': 8.22740650177002, 'timing/data': 0.0012106513604521751, 'timing/model': 1.4380043810233474, 'learning_rate/qwen_vl_interface': 7.896596060368384e-06, 'learning_rate/action_model': 7.896596060368384e-05, 'epoch': 0.87}) +[2026-06-17 21:05:43,567][starVLA.training.train_starvla][INFO] - Step 2253, Loss: {'train/loss': 0.3687672019004822, 'train/grad_norm_pre_clip': 5.992802143096924, 'timing/data': 0.0011575240641832352, 'timing/model': 1.1277562035247684, 'learning_rate/qwen_vl_interface': 7.89479990296464e-06, 'learning_rate/action_model': 7.894799902964639e-05, 'epoch': 0.87}) +[2026-06-17 21:05:48,511][starVLA.training.train_starvla][INFO] - Step 2254, Loss: {'train/loss': 0.5097135305404663, 'train/grad_norm_pre_clip': 7.539535045623779, 'timing/data': 0.0012028468772768974, 'timing/model': 1.2018305426463485, 'learning_rate/qwen_vl_interface': 7.893003197291028e-06, 'learning_rate/action_model': 7.893003197291027e-05, 'epoch': 0.87}) +[2026-06-17 21:05:53,486][starVLA.training.train_starvla][INFO] - Step 2255, Loss: {'train/loss': 0.438839316368103, 'train/grad_norm_pre_clip': 6.900986194610596, 'timing/data': 0.001262221485376358, 'timing/model': 1.1765020629391074, 'learning_rate/qwen_vl_interface': 7.89120594372001e-06, 'learning_rate/action_model': 7.89120594372001e-05, 'epoch': 0.87}) +[2026-06-17 21:05:58,106][starVLA.training.train_starvla][INFO] - Step 2256, Loss: {'train/loss': 0.3320382833480835, 'train/grad_norm_pre_clip': 6.233504295349121, 'timing/data': 0.0011251922696828842, 'timing/model': 1.1069283997640014, 'learning_rate/qwen_vl_interface': 7.88940814262416e-06, 'learning_rate/action_model': 7.889408142624158e-05, 'epoch': 0.87}) +[2026-06-17 21:06:02,964][starVLA.training.train_starvla][INFO] - Step 2257, Loss: {'train/loss': 0.3073171377182007, 'train/grad_norm_pre_clip': 7.380682945251465, 'timing/data': 0.0008415328338742256, 'timing/model': 1.5920144282281399, 'learning_rate/qwen_vl_interface': 7.887609794376158e-06, 'learning_rate/action_model': 7.887609794376158e-05, 'epoch': 0.87}) +[2026-06-17 21:06:07,749][starVLA.training.train_starvla][INFO] - Step 2258, Loss: {'train/loss': 0.4488951861858368, 'train/grad_norm_pre_clip': 7.836310386657715, 'timing/data': 0.0009379535913467407, 'timing/model': 1.1548834890127182, 'learning_rate/qwen_vl_interface': 7.88581089934881e-06, 'learning_rate/action_model': 7.88581089934881e-05, 'epoch': 0.87}) +[2026-06-17 21:06:14,221][starVLA.training.train_starvla][INFO] - Step 2259, Loss: {'train/loss': 0.6618039011955261, 'train/grad_norm_pre_clip': 10.928454399108887, 'timing/data': 0.0011209147050976753, 'timing/model': 2.427444189786911, 'learning_rate/qwen_vl_interface': 7.884011457915025e-06, 'learning_rate/action_model': 7.884011457915024e-05, 'epoch': 0.87}) +[2026-06-17 21:06:21,389][starVLA.training.train_starvla][INFO] - Step 2260, Loss: {'train/loss': 0.6712128520011902, 'train/grad_norm_pre_clip': 9.60892391204834, 'timing/data': 0.0009462470188736916, 'timing/model': 1.154487619176507, 'learning_rate/qwen_vl_interface': 7.882211470447832e-06, 'learning_rate/action_model': 7.882211470447831e-05, 'epoch': 0.87}) +[2026-06-17 21:06:31,198][starVLA.training.train_starvla][INFO] - Step 2261, Loss: {'train/loss': 0.2903458774089813, 'train/grad_norm_pre_clip': 6.609025955200195, 'timing/data': 0.001267600804567337, 'timing/model': 1.3677843157202005, 'learning_rate/qwen_vl_interface': 7.880410937320367e-06, 'learning_rate/action_model': 7.880410937320368e-05, 'epoch': 0.87}) +[2026-06-17 21:06:43,514][starVLA.training.train_starvla][INFO] - Step 2262, Loss: {'train/loss': 0.41619282960891724, 'train/grad_norm_pre_clip': 7.89219856262207, 'timing/data': 0.0011348044499754906, 'timing/model': 5.842470501549542, 'learning_rate/qwen_vl_interface': 7.878609858905885e-06, 'learning_rate/action_model': 7.878609858905886e-05, 'epoch': 0.87}) +[2026-06-17 21:06:49,418][starVLA.training.train_starvla][INFO] - Step 2263, Loss: {'train/loss': 0.6171618700027466, 'train/grad_norm_pre_clip': 10.529291152954102, 'timing/data': 0.001078854314982891, 'timing/model': 1.2416197638958693, 'learning_rate/qwen_vl_interface': 7.876808235577751e-06, 'learning_rate/action_model': 7.87680823557775e-05, 'epoch': 0.87}) +[2026-06-17 21:07:00,027][starVLA.training.train_starvla][INFO] - Step 2264, Loss: {'train/loss': 0.5012474656105042, 'train/grad_norm_pre_clip': 9.160850524902344, 'timing/data': 0.001652671955525875, 'timing/model': 1.251507773064077, 'learning_rate/qwen_vl_interface': 7.875006067709444e-06, 'learning_rate/action_model': 7.875006067709443e-05, 'epoch': 0.87}) +[2026-06-17 21:07:06,349][starVLA.training.train_starvla][INFO] - Step 2265, Loss: {'train/loss': 0.25937214493751526, 'train/grad_norm_pre_clip': 7.12387752532959, 'timing/data': 0.0010749287903308868, 'timing/model': 1.235897397622466, 'learning_rate/qwen_vl_interface': 7.873203355674553e-06, 'learning_rate/action_model': 7.873203355674552e-05, 'epoch': 0.87}) +[2026-06-17 21:07:18,919][starVLA.training.train_starvla][INFO] - Step 2266, Loss: {'train/loss': 0.5923982858657837, 'train/grad_norm_pre_clip': 10.328521728515625, 'timing/data': 0.0008934009820222855, 'timing/model': 4.133297925814986, 'learning_rate/qwen_vl_interface': 7.871400099846784e-06, 'learning_rate/action_model': 7.871400099846783e-05, 'epoch': 0.87}) +[2026-06-17 21:07:24,274][starVLA.training.train_starvla][INFO] - Step 2267, Loss: {'train/loss': 0.34941622614860535, 'train/grad_norm_pre_clip': 8.591594696044922, 'timing/data': 0.0012465566396713257, 'timing/model': 1.6845484655350447, 'learning_rate/qwen_vl_interface': 7.869596300599954e-06, 'learning_rate/action_model': 7.869596300599953e-05, 'epoch': 0.87}) +[2026-06-17 21:07:28,998][starVLA.training.train_starvla][INFO] - Step 2268, Loss: {'train/loss': 0.604525625705719, 'train/grad_norm_pre_clip': 9.331774711608887, 'timing/data': 0.00132750254124403, 'timing/model': 1.2153641851618886, 'learning_rate/qwen_vl_interface': 7.867791958307992e-06, 'learning_rate/action_model': 7.867791958307992e-05, 'epoch': 0.87}) +[2026-06-17 21:07:34,262][starVLA.training.train_starvla][INFO] - Step 2269, Loss: {'train/loss': 0.3487962484359741, 'train/grad_norm_pre_clip': 8.172109603881836, 'timing/data': 0.0012838738039135933, 'timing/model': 1.1770907258614898, 'learning_rate/qwen_vl_interface': 7.865987073344942e-06, 'learning_rate/action_model': 7.865987073344942e-05, 'epoch': 0.87}) +[2026-06-17 21:07:39,304][starVLA.training.train_starvla][INFO] - Step 2270, Loss: {'train/loss': 0.4576854407787323, 'train/grad_norm_pre_clip': 7.814359188079834, 'timing/data': 0.002247534692287445, 'timing/model': 1.1852882904931903, 'learning_rate/qwen_vl_interface': 7.864181646084953e-06, 'learning_rate/action_model': 7.864181646084953e-05, 'epoch': 0.87}) +[2026-06-17 21:07:44,076][starVLA.training.train_starvla][INFO] - Step 2271, Loss: {'train/loss': 0.5506125092506409, 'train/grad_norm_pre_clip': 9.013014793395996, 'timing/data': 0.001137150451540947, 'timing/model': 1.136379949748516, 'learning_rate/qwen_vl_interface': 7.862375676902298e-06, 'learning_rate/action_model': 7.862375676902297e-05, 'epoch': 0.87}) +[2026-06-17 21:07:49,091][starVLA.training.train_starvla][INFO] - Step 2272, Loss: {'train/loss': 0.2970295548439026, 'train/grad_norm_pre_clip': 8.80194091796875, 'timing/data': 0.0010245097801089287, 'timing/model': 1.4710011566057801, 'learning_rate/qwen_vl_interface': 7.860569166171353e-06, 'learning_rate/action_model': 7.860569166171354e-05, 'epoch': 0.87}) +[2026-06-17 21:07:53,725][starVLA.training.train_starvla][INFO] - Step 2273, Loss: {'train/loss': 0.359763503074646, 'train/grad_norm_pre_clip': 5.389678478240967, 'timing/data': 0.001260385848581791, 'timing/model': 1.229622295126319, 'learning_rate/qwen_vl_interface': 7.858762114266613e-06, 'learning_rate/action_model': 7.858762114266612e-05, 'epoch': 0.87}) +[2026-06-17 21:08:03,201][starVLA.training.train_starvla][INFO] - Step 2274, Loss: {'train/loss': 0.48427852988243103, 'train/grad_norm_pre_clip': 7.494003772735596, 'timing/data': 0.001555287279188633, 'timing/model': 1.220655707642436, 'learning_rate/qwen_vl_interface': 7.856954521562677e-06, 'learning_rate/action_model': 7.856954521562677e-05, 'epoch': 0.87}) +[2026-06-17 21:08:08,554][starVLA.training.train_starvla][INFO] - Step 2275, Loss: {'train/loss': 0.6461656093597412, 'train/grad_norm_pre_clip': 10.925186157226562, 'timing/data': 0.0009283926337957382, 'timing/model': 1.3850622214376926, 'learning_rate/qwen_vl_interface': 7.855146388434264e-06, 'learning_rate/action_model': 7.855146388434264e-05, 'epoch': 0.88}) +[2026-06-17 21:08:13,398][starVLA.training.train_starvla][INFO] - Step 2276, Loss: {'train/loss': 0.3369638919830322, 'train/grad_norm_pre_clip': 6.2395830154418945, 'timing/data': 0.0013048015534877777, 'timing/model': 1.144613959826529, 'learning_rate/qwen_vl_interface': 7.853337715256202e-06, 'learning_rate/action_model': 7.853337715256203e-05, 'epoch': 0.88}) +[2026-06-17 21:08:18,335][starVLA.training.train_starvla][INFO] - Step 2277, Loss: {'train/loss': 0.2642245292663574, 'train/grad_norm_pre_clip': 5.819525241851807, 'timing/data': 0.0011809878051280975, 'timing/model': 1.5642841709777713, 'learning_rate/qwen_vl_interface': 7.851528502403431e-06, 'learning_rate/action_model': 7.851528502403431e-05, 'epoch': 0.88}) +[2026-06-17 21:08:23,014][starVLA.training.train_starvla][INFO] - Step 2278, Loss: {'train/loss': 0.5348381996154785, 'train/grad_norm_pre_clip': 13.00685977935791, 'timing/data': 0.0011357497423887253, 'timing/model': 1.5226624673232436, 'learning_rate/qwen_vl_interface': 7.849718750251001e-06, 'learning_rate/action_model': 7.849718750251002e-05, 'epoch': 0.88}) +[2026-06-17 21:08:28,619][starVLA.training.train_starvla][INFO] - Step 2279, Loss: {'train/loss': 0.43291959166526794, 'train/grad_norm_pre_clip': 10.166309356689453, 'timing/data': 0.0014968467876315117, 'timing/model': 1.604515552520752, 'learning_rate/qwen_vl_interface': 7.847908459174077e-06, 'learning_rate/action_model': 7.847908459174078e-05, 'epoch': 0.88}) +[2026-06-17 21:08:34,159][starVLA.training.train_starvla][INFO] - Step 2280, Loss: {'train/loss': 0.6706732511520386, 'train/grad_norm_pre_clip': 8.821478843688965, 'timing/data': 0.0009285500273108482, 'timing/model': 1.1165121775120497, 'learning_rate/qwen_vl_interface': 7.846097629547934e-06, 'learning_rate/action_model': 7.846097629547933e-05, 'epoch': 0.88}) +[2026-06-17 21:08:42,294][starVLA.training.train_starvla][INFO] - Step 2281, Loss: {'train/loss': 0.5219611525535583, 'train/grad_norm_pre_clip': 7.854763984680176, 'timing/data': 0.0011673578992486, 'timing/model': 1.4152206014841795, 'learning_rate/qwen_vl_interface': 7.844286261747959e-06, 'learning_rate/action_model': 7.844286261747958e-05, 'epoch': 0.88}) +[2026-06-17 21:08:47,390][starVLA.training.train_starvla][INFO] - Step 2282, Loss: {'train/loss': 0.23654675483703613, 'train/grad_norm_pre_clip': 7.656221866607666, 'timing/data': 0.0010437490418553352, 'timing/model': 1.5317391138523817, 'learning_rate/qwen_vl_interface': 7.84247435614965e-06, 'learning_rate/action_model': 7.84247435614965e-05, 'epoch': 0.88}) +[2026-06-17 21:08:52,448][starVLA.training.train_starvla][INFO] - Step 2283, Loss: {'train/loss': 0.6800571084022522, 'train/grad_norm_pre_clip': 8.23737907409668, 'timing/data': 0.0012954799458384514, 'timing/model': 1.1042457623407245, 'learning_rate/qwen_vl_interface': 7.840661913128618e-06, 'learning_rate/action_model': 7.840661913128619e-05, 'epoch': 0.88}) +[2026-06-17 21:08:57,542][starVLA.training.train_starvla][INFO] - Step 2284, Loss: {'train/loss': 0.4547538161277771, 'train/grad_norm_pre_clip': 6.740610599517822, 'timing/data': 0.002117685042321682, 'timing/model': 1.3711735252290964, 'learning_rate/qwen_vl_interface': 7.838848933060583e-06, 'learning_rate/action_model': 7.838848933060583e-05, 'epoch': 0.88}) +[2026-06-17 21:09:02,687][starVLA.training.train_starvla][INFO] - Step 2285, Loss: {'train/loss': 0.5201621651649475, 'train/grad_norm_pre_clip': 8.45296859741211, 'timing/data': 0.0012770891189575195, 'timing/model': 1.1968932067975402, 'learning_rate/qwen_vl_interface': 7.83703541632138e-06, 'learning_rate/action_model': 7.83703541632138e-05, 'epoch': 0.88}) +[2026-06-17 21:09:09,620][starVLA.training.train_starvla][INFO] - Step 2286, Loss: {'train/loss': 0.7930518984794617, 'train/grad_norm_pre_clip': 8.94923210144043, 'timing/data': 0.0015792110934853554, 'timing/model': 1.4049143623560667, 'learning_rate/qwen_vl_interface': 7.835221363286949e-06, 'learning_rate/action_model': 7.835221363286949e-05, 'epoch': 0.88}) +[2026-06-17 21:09:18,028][starVLA.training.train_starvla][INFO] - Step 2287, Loss: {'train/loss': 0.42169249057769775, 'train/grad_norm_pre_clip': 9.699827194213867, 'timing/data': 0.0014188531786203384, 'timing/model': 4.427019313909113, 'learning_rate/qwen_vl_interface': 7.83340677433335e-06, 'learning_rate/action_model': 7.83340677433335e-05, 'epoch': 0.88}) +[2026-06-17 21:09:23,312][starVLA.training.train_starvla][INFO] - Step 2288, Loss: {'train/loss': 0.4474119544029236, 'train/grad_norm_pre_clip': 9.336987495422363, 'timing/data': 0.0012126723304390907, 'timing/model': 1.2196167083457112, 'learning_rate/qwen_vl_interface': 7.831591649836745e-06, 'learning_rate/action_model': 7.831591649836744e-05, 'epoch': 0.88}) +[2026-06-17 21:09:28,638][starVLA.training.train_starvla][INFO] - Step 2289, Loss: {'train/loss': 0.44260478019714355, 'train/grad_norm_pre_clip': 8.655138969421387, 'timing/data': 0.000996498391032219, 'timing/model': 1.2573208445683122, 'learning_rate/qwen_vl_interface': 7.829775990173413e-06, 'learning_rate/action_model': 7.829775990173413e-05, 'epoch': 0.88}) +[2026-06-17 21:09:33,291][starVLA.training.train_starvla][INFO] - Step 2290, Loss: {'train/loss': 0.3123791813850403, 'train/grad_norm_pre_clip': 7.162339687347412, 'timing/data': 0.0009249728173017502, 'timing/model': 1.1563480515033007, 'learning_rate/qwen_vl_interface': 7.827959795719744e-06, 'learning_rate/action_model': 7.827959795719744e-05, 'epoch': 0.88}) +[2026-06-17 21:09:41,182][starVLA.training.train_starvla][INFO] - Step 2291, Loss: {'train/loss': 0.5670052170753479, 'train/grad_norm_pre_clip': 10.522994995117188, 'timing/data': 0.001146884635090828, 'timing/model': 4.0202633291482925, 'learning_rate/qwen_vl_interface': 7.826143066852233e-06, 'learning_rate/action_model': 7.826143066852233e-05, 'epoch': 0.88}) +[2026-06-17 21:09:52,132][starVLA.training.train_starvla][INFO] - Step 2292, Loss: {'train/loss': 0.5689653754234314, 'train/grad_norm_pre_clip': 7.930947780609131, 'timing/data': 0.0012269802391529083, 'timing/model': 1.393522177822888, 'learning_rate/qwen_vl_interface': 7.824325803947494e-06, 'learning_rate/action_model': 7.824325803947495e-05, 'epoch': 0.88}) +[2026-06-17 21:09:56,629][starVLA.training.train_starvla][INFO] - Step 2293, Loss: {'train/loss': 0.3918687701225281, 'train/grad_norm_pre_clip': 8.183511734008789, 'timing/data': 0.0015459069982171059, 'timing/model': 1.1074207182973623, 'learning_rate/qwen_vl_interface': 7.822508007382246e-06, 'learning_rate/action_model': 7.822508007382245e-05, 'epoch': 0.88}) +[2026-06-17 21:10:01,734][starVLA.training.train_starvla][INFO] - Step 2294, Loss: {'train/loss': 0.5373351573944092, 'train/grad_norm_pre_clip': 16.45330810546875, 'timing/data': 0.0012002214789390564, 'timing/model': 1.2587708979845047, 'learning_rate/qwen_vl_interface': 7.820689677533319e-06, 'learning_rate/action_model': 7.82068967753332e-05, 'epoch': 0.88}) +[2026-06-17 21:10:06,874][starVLA.training.train_starvla][INFO] - Step 2295, Loss: {'train/loss': 0.3695746660232544, 'train/grad_norm_pre_clip': 5.946263790130615, 'timing/data': 0.0012275511398911476, 'timing/model': 1.2735943300649524, 'learning_rate/qwen_vl_interface': 7.818870814777655e-06, 'learning_rate/action_model': 7.818870814777655e-05, 'epoch': 0.88}) +[2026-06-17 21:10:12,872][starVLA.training.train_starvla][INFO] - Step 2296, Loss: {'train/loss': 0.4508579671382904, 'train/grad_norm_pre_clip': 20.456066131591797, 'timing/data': 0.001197713427245617, 'timing/model': 2.099248031154275, 'learning_rate/qwen_vl_interface': 7.81705141949231e-06, 'learning_rate/action_model': 7.817051419492308e-05, 'epoch': 0.88}) +[2026-06-17 21:10:18,802][starVLA.training.train_starvla][INFO] - Step 2297, Loss: {'train/loss': 0.5952382683753967, 'train/grad_norm_pre_clip': 11.6209716796875, 'timing/data': 0.0012530554085969925, 'timing/model': 2.614512545056641, 'learning_rate/qwen_vl_interface': 7.81523149205444e-06, 'learning_rate/action_model': 7.81523149205444e-05, 'epoch': 0.88}) +[2026-06-17 21:10:28,139][starVLA.training.train_starvla][INFO] - Step 2298, Loss: {'train/loss': 0.5032673478126526, 'train/grad_norm_pre_clip': 8.606507301330566, 'timing/data': 0.001383662223815918, 'timing/model': 1.1208684099838138, 'learning_rate/qwen_vl_interface': 7.813411032841323e-06, 'learning_rate/action_model': 7.813411032841323e-05, 'epoch': 0.88}) +[2026-06-17 21:10:40,857][starVLA.training.train_starvla][INFO] - Step 2299, Loss: {'train/loss': 0.613060712814331, 'train/grad_norm_pre_clip': 11.67162799835205, 'timing/data': 0.0012671193107962608, 'timing/model': 4.551550608128309, 'learning_rate/qwen_vl_interface': 7.81159004223034e-06, 'learning_rate/action_model': 7.81159004223034e-05, 'epoch': 0.88}) +[2026-06-17 21:14:13,449][starVLA.training.train_starvla][INFO] - Step 2300, Loss: {'train/loss': 0.5582379698753357, 'train/grad_norm_pre_clip': 9.863304138183594, 'eval/loss': 0.5269126892089844, 'eval/latency_4/loss': 0.5269126892089844, 'eval/demon_fire_precision': 0.9599618684461392, 'eval/demon_fire_recall': 0.9811627151672621, 'eval/demon_fire_f1': 0.9704465146161259, 'eval/demon_fire_rate_model': 0.9242290748898678, 'eval/demon_fire_rate_teacher': 0.9042584434654919, 'eval/demon_left_precision': 0.7486187845303868, 'eval/demon_left_recall': 0.6187214611872146, 'eval/demon_left_f1': 0.6775, 'eval/demon_left_rate_model': 0.10631424375917768, 'eval/demon_left_rate_teacher': 0.12863436123348018, 'eval/demon_right_precision': 0.7542213883677298, 'eval/demon_right_recall': 0.7256317689530686, 'eval/demon_right_f1': 0.7396504139834406, 'eval/demon_right_rate_model': 0.1565345080763583, 'eval/demon_right_rate_teacher': 0.16270190895741557, 'eval/demon_move_f1': 0.7085752069917203, 'eval/demon_cc_f1': 0.8735895911819325, 'eval/latency_4/demon_fire_precision': 0.9599618684461392, 'eval/latency_4/demon_fire_recall': 0.9811627151672621, 'eval/latency_4/demon_fire_f1': 0.9704465146161259, 'eval/latency_4/demon_fire_rate_model': 0.9242290748898678, 'eval/latency_4/demon_fire_rate_teacher': 0.9042584434654919, 'eval/latency_4/demon_left_precision': 0.7486187845303868, 'eval/latency_4/demon_left_recall': 0.6187214611872146, 'eval/latency_4/demon_left_f1': 0.6775, 'eval/latency_4/demon_left_rate_model': 0.10631424375917768, 'eval/latency_4/demon_left_rate_teacher': 0.12863436123348018, 'eval/latency_4/demon_right_precision': 0.7542213883677298, 'eval/latency_4/demon_right_recall': 0.7256317689530686, 'eval/latency_4/demon_right_f1': 0.7396504139834406, 'eval/latency_4/demon_right_rate_model': 0.1565345080763583, 'eval/latency_4/demon_right_rate_teacher': 0.16270190895741557, 'eval/latency_4/demon_move_f1': 0.7085752069917203, 'eval/latency_4/demon_cc_f1': 0.8735895911819325, 'timing/data': 0.0012284833937883377, 'timing/model': 1.284355933777988, 'learning_rate/qwen_vl_interface': 7.809768520598986e-06, 'learning_rate/action_model': 7.809768520598984e-05, 'epoch': 0.88}) +[2026-06-17 21:14:13,451][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_2300_state +[2026-06-17 21:14:13,467][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 21:14:22,327][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_2300_state/model.safetensors +[2026-06-17 21:14:44,846][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_2300_state/optimizer.bin +[2026-06-17 21:14:44,847][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_2300_state/sampler.bin +[2026-06-17 21:14:44,847][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_2300_state/sampler_1.bin +[2026-06-17 21:14:44,850][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_2300_state/random_states_0.pkl +[2026-06-17 21:14:44,869][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2300 +[2026-06-17 21:14:44,869][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 21:14:44,886][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 21:14:49,731][starVLA.training.train_starvla][INFO] - Step 2301, Loss: {'train/loss': 0.5790979266166687, 'train/grad_norm_pre_clip': 9.992810249328613, 'timing/data': 0.0011409353464841843, 'timing/model': 1.0844316575676203, 'learning_rate/qwen_vl_interface': 7.80794646832486e-06, 'learning_rate/action_model': 7.80794646832486e-05, 'epoch': 0.89}) +[2026-06-17 21:14:54,607][starVLA.training.train_starvla][INFO] - Step 2302, Loss: {'train/loss': 0.6523913145065308, 'train/grad_norm_pre_clip': 11.58095932006836, 'timing/data': 0.0011286363005638123, 'timing/model': 1.44697112031281, 'learning_rate/qwen_vl_interface': 7.80612388578568e-06, 'learning_rate/action_model': 7.80612388578568e-05, 'epoch': 0.89}) +[2026-06-17 21:14:59,059][starVLA.training.train_starvla][INFO] - Step 2303, Loss: {'train/loss': 0.41115158796310425, 'train/grad_norm_pre_clip': 13.26572036743164, 'timing/data': 0.0015481077134609222, 'timing/model': 1.1856944244354963, 'learning_rate/qwen_vl_interface': 7.80430077335927e-06, 'learning_rate/action_model': 7.80430077335927e-05, 'epoch': 0.89}) +[2026-06-17 21:15:03,859][starVLA.training.train_starvla][INFO] - Step 2304, Loss: {'train/loss': 0.5081908106803894, 'train/grad_norm_pre_clip': 9.766072273254395, 'timing/data': 0.0012065311893820763, 'timing/model': 1.1916068829596043, 'learning_rate/qwen_vl_interface': 7.802477131423561e-06, 'learning_rate/action_model': 7.80247713142356e-05, 'epoch': 0.89}) +[2026-06-17 21:15:08,862][starVLA.training.train_starvla][INFO] - Step 2305, Loss: {'train/loss': 0.3110406994819641, 'train/grad_norm_pre_clip': 7.109095096588135, 'timing/data': 0.0011595264077186584, 'timing/model': 1.238217999227345, 'learning_rate/qwen_vl_interface': 7.800652960356594e-06, 'learning_rate/action_model': 7.800652960356594e-05, 'epoch': 0.89}) +[2026-06-17 21:15:14,996][starVLA.training.train_starvla][INFO] - Step 2306, Loss: {'train/loss': 0.5387402176856995, 'train/grad_norm_pre_clip': 9.900898933410645, 'timing/data': 0.0009299926459789276, 'timing/model': 1.224340407177806, 'learning_rate/qwen_vl_interface': 7.798828260536524e-06, 'learning_rate/action_model': 7.798828260536524e-05, 'epoch': 0.89}) +[2026-06-17 21:15:20,078][starVLA.training.train_starvla][INFO] - Step 2307, Loss: {'train/loss': 0.662723183631897, 'train/grad_norm_pre_clip': 9.707645416259766, 'timing/data': 0.0008343644440174103, 'timing/model': 1.5818918701261282, 'learning_rate/qwen_vl_interface': 7.797003032341615e-06, 'learning_rate/action_model': 7.797003032341614e-05, 'epoch': 0.89}) +[2026-06-17 21:15:31,083][starVLA.training.train_starvla][INFO] - Step 2308, Loss: {'train/loss': 0.4167759120464325, 'train/grad_norm_pre_clip': 12.740933418273926, 'timing/data': 0.00151812843978405, 'timing/model': 1.0728169148787856, 'learning_rate/qwen_vl_interface': 7.795177276150234e-06, 'learning_rate/action_model': 7.795177276150234e-05, 'epoch': 0.89}) +[2026-06-17 21:15:35,714][starVLA.training.train_starvla][INFO] - Step 2309, Loss: {'train/loss': 0.7188752889633179, 'train/grad_norm_pre_clip': 12.523178100585938, 'timing/data': 0.0009977715089917183, 'timing/model': 1.087235358543694, 'learning_rate/qwen_vl_interface': 7.793350992340867e-06, 'learning_rate/action_model': 7.793350992340867e-05, 'epoch': 0.89}) +[2026-06-17 21:15:40,800][starVLA.training.train_starvla][INFO] - Step 2310, Loss: {'train/loss': 0.5326520800590515, 'train/grad_norm_pre_clip': 7.774203300476074, 'timing/data': 0.0012632273137569427, 'timing/model': 1.2112404201179743, 'learning_rate/qwen_vl_interface': 7.791524181292102e-06, 'learning_rate/action_model': 7.791524181292101e-05, 'epoch': 0.89}) +[2026-06-17 21:15:45,389][starVLA.training.train_starvla][INFO] - Step 2311, Loss: {'train/loss': 0.5466148853302002, 'train/grad_norm_pre_clip': 7.994499206542969, 'timing/data': 0.0011820225045084953, 'timing/model': 1.0771742379292846, 'learning_rate/qwen_vl_interface': 7.789696843382638e-06, 'learning_rate/action_model': 7.789696843382639e-05, 'epoch': 0.89}) +[2026-06-17 21:15:50,132][starVLA.training.train_starvla][INFO] - Step 2312, Loss: {'train/loss': 0.9290964603424072, 'train/grad_norm_pre_clip': 10.978944778442383, 'timing/data': 0.00120452418923378, 'timing/model': 1.524460719898343, 'learning_rate/qwen_vl_interface': 7.787868978991288e-06, 'learning_rate/action_model': 7.787868978991287e-05, 'epoch': 0.89}) +[2026-06-17 21:15:54,818][starVLA.training.train_starvla][INFO] - Step 2313, Loss: {'train/loss': 0.8848070502281189, 'train/grad_norm_pre_clip': 12.034242630004883, 'timing/data': 0.0012173978611826897, 'timing/model': 1.431683680973947, 'learning_rate/qwen_vl_interface': 7.786040588496966e-06, 'learning_rate/action_model': 7.786040588496965e-05, 'epoch': 0.89}) +[2026-06-17 21:16:03,090][starVLA.training.train_starvla][INFO] - Step 2314, Loss: {'train/loss': 0.651157021522522, 'train/grad_norm_pre_clip': 9.833222389221191, 'timing/data': 0.000916886143386364, 'timing/model': 1.1388354385271668, 'learning_rate/qwen_vl_interface': 7.784211672278703e-06, 'learning_rate/action_model': 7.784211672278703e-05, 'epoch': 0.89}) +[2026-06-17 21:16:08,282][starVLA.training.train_starvla][INFO] - Step 2315, Loss: {'train/loss': 0.31755226850509644, 'train/grad_norm_pre_clip': 6.728030204772949, 'timing/data': 0.0010955939069390297, 'timing/model': 1.4084664825350046, 'learning_rate/qwen_vl_interface': 7.78238223071563e-06, 'learning_rate/action_model': 7.78238223071563e-05, 'epoch': 0.89}) +[2026-06-17 21:16:16,338][starVLA.training.train_starvla][INFO] - Step 2316, Loss: {'train/loss': 0.7689973711967468, 'train/grad_norm_pre_clip': 13.269082069396973, 'timing/data': 0.0011219661682844162, 'timing/model': 1.0775561993941665, 'learning_rate/qwen_vl_interface': 7.780552264187e-06, 'learning_rate/action_model': 7.780552264186999e-05, 'epoch': 0.89}) +[2026-06-17 21:16:22,920][starVLA.training.train_starvla][INFO] - Step 2317, Loss: {'train/loss': 0.5180041193962097, 'train/grad_norm_pre_clip': 7.177798271179199, 'timing/data': 0.0009204559028148651, 'timing/model': 1.4815556816756725, 'learning_rate/qwen_vl_interface': 7.77872177307216e-06, 'learning_rate/action_model': 7.778721773072159e-05, 'epoch': 0.89}) +[2026-06-17 21:16:27,287][starVLA.training.train_starvla][INFO] - Step 2318, Loss: {'train/loss': 0.36417055130004883, 'train/grad_norm_pre_clip': 6.417911529541016, 'timing/data': 0.0012717647477984428, 'timing/model': 1.078760669566691, 'learning_rate/qwen_vl_interface': 7.776890757750575e-06, 'learning_rate/action_model': 7.776890757750575e-05, 'epoch': 0.89}) +[2026-06-17 21:16:32,260][starVLA.training.train_starvla][INFO] - Step 2319, Loss: {'train/loss': 0.5325567722320557, 'train/grad_norm_pre_clip': 10.6246337890625, 'timing/data': 0.0014888765290379524, 'timing/model': 1.2068810481578112, 'learning_rate/qwen_vl_interface': 7.775059218601816e-06, 'learning_rate/action_model': 7.775059218601816e-05, 'epoch': 0.89}) +[2026-06-17 21:16:37,248][starVLA.training.train_starvla][INFO] - Step 2320, Loss: {'train/loss': 0.36603131890296936, 'train/grad_norm_pre_clip': 8.037019729614258, 'timing/data': 0.0012449678033590317, 'timing/model': 1.2481479868292809, 'learning_rate/qwen_vl_interface': 7.773227156005567e-06, 'learning_rate/action_model': 7.773227156005567e-05, 'epoch': 0.89}) +[2026-06-17 21:16:45,341][starVLA.training.train_starvla][INFO] - Step 2321, Loss: {'train/loss': 0.4642336666584015, 'train/grad_norm_pre_clip': 9.030372619628906, 'timing/data': 0.0013296594843268394, 'timing/model': 4.4675571359694, 'learning_rate/qwen_vl_interface': 7.771394570341612e-06, 'learning_rate/action_model': 7.771394570341612e-05, 'epoch': 0.89}) +[2026-06-17 21:16:50,011][starVLA.training.train_starvla][INFO] - Step 2322, Loss: {'train/loss': 0.4806329011917114, 'train/grad_norm_pre_clip': 7.073591232299805, 'timing/data': 0.0015145903453230858, 'timing/model': 1.4727506870403886, 'learning_rate/qwen_vl_interface': 7.769561461989848e-06, 'learning_rate/action_model': 7.769561461989848e-05, 'epoch': 0.89}) +[2026-06-17 21:16:54,878][starVLA.training.train_starvla][INFO] - Step 2323, Loss: {'train/loss': 0.37731584906578064, 'train/grad_norm_pre_clip': 7.61778450012207, 'timing/data': 0.0018857484683394432, 'timing/model': 1.5824097078293562, 'learning_rate/qwen_vl_interface': 7.767727831330285e-06, 'learning_rate/action_model': 7.767727831330285e-05, 'epoch': 0.89}) +[2026-06-17 21:16:59,929][starVLA.training.train_starvla][INFO] - Step 2324, Loss: {'train/loss': 0.6108770966529846, 'train/grad_norm_pre_clip': 7.084339141845703, 'timing/data': 0.0010384274646639824, 'timing/model': 1.0719799594953656, 'learning_rate/qwen_vl_interface': 7.765893678743029e-06, 'learning_rate/action_model': 7.76589367874303e-05, 'epoch': 0.89}) +[2026-06-17 21:17:07,793][starVLA.training.train_starvla][INFO] - Step 2325, Loss: {'train/loss': 0.6528974771499634, 'train/grad_norm_pre_clip': 8.318903923034668, 'timing/data': 0.0010384973138570786, 'timing/model': 1.1630634795874357, 'learning_rate/qwen_vl_interface': 7.76405900460831e-06, 'learning_rate/action_model': 7.76405900460831e-05, 'epoch': 0.89}) +[2026-06-17 21:17:12,841][starVLA.training.train_starvla][INFO] - Step 2326, Loss: {'train/loss': 0.5531032681465149, 'train/grad_norm_pre_clip': 7.6518754959106445, 'timing/data': 0.001179116778075695, 'timing/model': 1.3496844246983528, 'learning_rate/qwen_vl_interface': 7.762223809306453e-06, 'learning_rate/action_model': 7.762223809306452e-05, 'epoch': 0.89}) +[2026-06-17 21:17:17,577][starVLA.training.train_starvla][INFO] - Step 2327, Loss: {'train/loss': 0.6121692061424255, 'train/grad_norm_pre_clip': 8.253170013427734, 'timing/data': 0.0011338666081428528, 'timing/model': 1.5010637966915965, 'learning_rate/qwen_vl_interface': 7.760388093217897e-06, 'learning_rate/action_model': 7.760388093217898e-05, 'epoch': 0.9}) +[2026-06-17 21:17:23,067][starVLA.training.train_starvla][INFO] - Step 2328, Loss: {'train/loss': 0.4657897651195526, 'train/grad_norm_pre_clip': 6.1811089515686035, 'timing/data': 0.002493619918823242, 'timing/model': 1.1247010435909033, 'learning_rate/qwen_vl_interface': 7.758551856723187e-06, 'learning_rate/action_model': 7.758551856723188e-05, 'epoch': 0.9}) +[2026-06-17 21:17:28,505][starVLA.training.train_starvla][INFO] - Step 2329, Loss: {'train/loss': 0.46587902307510376, 'train/grad_norm_pre_clip': 8.851353645324707, 'timing/data': 0.0009749568998813629, 'timing/model': 1.236002174206078, 'learning_rate/qwen_vl_interface': 7.756715100202979e-06, 'learning_rate/action_model': 7.75671510020298e-05, 'epoch': 0.9}) +[2026-06-17 21:17:34,530][starVLA.training.train_starvla][INFO] - Step 2330, Loss: {'train/loss': 0.34634748101234436, 'train/grad_norm_pre_clip': 5.368038654327393, 'timing/data': 0.00111297145485878, 'timing/model': 1.0953282229602337, 'learning_rate/qwen_vl_interface': 7.754877824038035e-06, 'learning_rate/action_model': 7.754877824038035e-05, 'epoch': 0.9}) +[2026-06-17 21:17:43,665][starVLA.training.train_starvla][INFO] - Step 2331, Loss: {'train/loss': 0.5557742714881897, 'train/grad_norm_pre_clip': 18.920902252197266, 'timing/data': 0.001305699348449707, 'timing/model': 1.0844150464981794, 'learning_rate/qwen_vl_interface': 7.75304002860922e-06, 'learning_rate/action_model': 7.75304002860922e-05, 'epoch': 0.9}) +[2026-06-17 21:17:50,344][starVLA.training.train_starvla][INFO] - Step 2332, Loss: {'train/loss': 0.786648690700531, 'train/grad_norm_pre_clip': 11.19362735748291, 'timing/data': 0.0009572282433509827, 'timing/model': 1.5100588370114565, 'learning_rate/qwen_vl_interface': 7.751201714297513e-06, 'learning_rate/action_model': 7.751201714297512e-05, 'epoch': 0.9}) +[2026-06-17 21:17:57,518][starVLA.training.train_starvla][INFO] - Step 2333, Loss: {'train/loss': 0.2662668824195862, 'train/grad_norm_pre_clip': 4.313509464263916, 'timing/data': 0.001296069473028183, 'timing/model': 1.1124347550794482, 'learning_rate/qwen_vl_interface': 7.749362881484002e-06, 'learning_rate/action_model': 7.749362881484001e-05, 'epoch': 0.9}) +[2026-06-17 21:18:04,141][starVLA.training.train_starvla][INFO] - Step 2334, Loss: {'train/loss': 0.5780677795410156, 'train/grad_norm_pre_clip': 13.603338241577148, 'timing/data': 0.002686607651412487, 'timing/model': 1.2255369080230594, 'learning_rate/qwen_vl_interface': 7.747523530549874e-06, 'learning_rate/action_model': 7.747523530549875e-05, 'epoch': 0.9}) +[2026-06-17 21:18:09,238][starVLA.training.train_starvla][INFO] - Step 2335, Loss: {'train/loss': 0.5194029211997986, 'train/grad_norm_pre_clip': 9.924211502075195, 'timing/data': 0.0015670862048864365, 'timing/model': 1.2430144483223557, 'learning_rate/qwen_vl_interface': 7.745683661876432e-06, 'learning_rate/action_model': 7.745683661876432e-05, 'epoch': 0.9}) +[2026-06-17 21:18:14,447][starVLA.training.train_starvla][INFO] - Step 2336, Loss: {'train/loss': 0.3416398763656616, 'train/grad_norm_pre_clip': 5.821260452270508, 'timing/data': 0.0011876346543431282, 'timing/model': 1.347839280962944, 'learning_rate/qwen_vl_interface': 7.743843275845081e-06, 'learning_rate/action_model': 7.74384327584508e-05, 'epoch': 0.9}) +[2026-06-17 21:18:19,448][starVLA.training.train_starvla][INFO] - Step 2337, Loss: {'train/loss': 0.5722664594650269, 'train/grad_norm_pre_clip': 7.557100296020508, 'timing/data': 0.001016249880194664, 'timing/model': 1.5851880963891745, 'learning_rate/qwen_vl_interface': 7.742002372837336e-06, 'learning_rate/action_model': 7.742002372837335e-05, 'epoch': 0.9}) +[2026-06-17 21:18:24,056][starVLA.training.train_starvla][INFO] - Step 2338, Loss: {'train/loss': 0.36513179540634155, 'train/grad_norm_pre_clip': 5.698680400848389, 'timing/data': 0.0011906763538718224, 'timing/model': 1.1671103071421385, 'learning_rate/qwen_vl_interface': 7.740160953234815e-06, 'learning_rate/action_model': 7.740160953234815e-05, 'epoch': 0.9}) +[2026-06-17 21:18:29,245][starVLA.training.train_starvla][INFO] - Step 2339, Loss: {'train/loss': 0.3819393217563629, 'train/grad_norm_pre_clip': 8.20623779296875, 'timing/data': 0.0012285178527235985, 'timing/model': 1.3731062291190028, 'learning_rate/qwen_vl_interface': 7.738319017419252e-06, 'learning_rate/action_model': 7.738319017419252e-05, 'epoch': 0.9}) +[2026-06-17 21:18:34,386][starVLA.training.train_starvla][INFO] - Step 2340, Loss: {'train/loss': 0.6810622215270996, 'train/grad_norm_pre_clip': 14.229928016662598, 'timing/data': 0.0011222222819924355, 'timing/model': 1.2001255052164197, 'learning_rate/qwen_vl_interface': 7.736476565772478e-06, 'learning_rate/action_model': 7.736476565772478e-05, 'epoch': 0.9}) +[2026-06-17 21:18:39,248][starVLA.training.train_starvla][INFO] - Step 2341, Loss: {'train/loss': 0.5387986898422241, 'train/grad_norm_pre_clip': 7.499976634979248, 'timing/data': 0.001262403093278408, 'timing/model': 1.213314302265644, 'learning_rate/qwen_vl_interface': 7.734633598676435e-06, 'learning_rate/action_model': 7.734633598676434e-05, 'epoch': 0.9}) +[2026-06-17 21:18:44,200][starVLA.training.train_starvla][INFO] - Step 2342, Loss: {'train/loss': 0.7036200761795044, 'train/grad_norm_pre_clip': 9.11224365234375, 'timing/data': 0.0009060045704245567, 'timing/model': 1.1337267914786935, 'learning_rate/qwen_vl_interface': 7.732790116513176e-06, 'learning_rate/action_model': 7.732790116513176e-05, 'epoch': 0.9}) +[2026-06-17 21:18:49,180][starVLA.training.train_starvla][INFO] - Step 2343, Loss: {'train/loss': 0.42140719294548035, 'train/grad_norm_pre_clip': 11.72222900390625, 'timing/data': 0.0017690621316432953, 'timing/model': 1.1840242696925998, 'learning_rate/qwen_vl_interface': 7.730946119664852e-06, 'learning_rate/action_model': 7.730946119664853e-05, 'epoch': 0.9}) +[2026-06-17 21:18:54,067][starVLA.training.train_starvla][INFO] - Step 2344, Loss: {'train/loss': 0.312695175409317, 'train/grad_norm_pre_clip': 7.922905921936035, 'timing/data': 0.001011814922094345, 'timing/model': 1.1903583267703652, 'learning_rate/qwen_vl_interface': 7.729101608513729e-06, 'learning_rate/action_model': 7.72910160851373e-05, 'epoch': 0.9}) +[2026-06-17 21:18:58,725][starVLA.training.train_starvla][INFO] - Step 2345, Loss: {'train/loss': 0.29912447929382324, 'train/grad_norm_pre_clip': 6.608072757720947, 'timing/data': 0.0012482861056923866, 'timing/model': 1.0789634846150875, 'learning_rate/qwen_vl_interface': 7.727256583442175e-06, 'learning_rate/action_model': 7.727256583442175e-05, 'epoch': 0.9}) +[2026-06-17 21:19:03,432][starVLA.training.train_starvla][INFO] - Step 2346, Loss: {'train/loss': 0.32418790459632874, 'train/grad_norm_pre_clip': 4.606380939483643, 'timing/data': 0.001292189583182335, 'timing/model': 1.09920359775424, 'learning_rate/qwen_vl_interface': 7.725411044832665e-06, 'learning_rate/action_model': 7.725411044832665e-05, 'epoch': 0.9}) +[2026-06-17 21:19:08,436][starVLA.training.train_starvla][INFO] - Step 2347, Loss: {'train/loss': 0.7181903123855591, 'train/grad_norm_pre_clip': 14.34858512878418, 'timing/data': 0.0010887477546930313, 'timing/model': 1.6556193511933088, 'learning_rate/qwen_vl_interface': 7.723564993067783e-06, 'learning_rate/action_model': 7.723564993067783e-05, 'epoch': 0.9}) +[2026-06-17 21:19:13,031][starVLA.training.train_starvla][INFO] - Step 2348, Loss: {'train/loss': 0.3073311746120453, 'train/grad_norm_pre_clip': 6.634171485900879, 'timing/data': 0.0012244908139109612, 'timing/model': 1.2559215631335974, 'learning_rate/qwen_vl_interface': 7.721718428530214e-06, 'learning_rate/action_model': 7.721718428530214e-05, 'epoch': 0.9}) +[2026-06-17 21:19:17,834][starVLA.training.train_starvla][INFO] - Step 2349, Loss: {'train/loss': 0.5312243700027466, 'train/grad_norm_pre_clip': 8.17824649810791, 'timing/data': 0.0011114496737718582, 'timing/model': 1.0864487504586577, 'learning_rate/qwen_vl_interface': 7.719871351602757e-06, 'learning_rate/action_model': 7.719871351602757e-05, 'epoch': 0.9}) +[2026-06-17 21:19:23,256][starVLA.training.train_starvla][INFO] - Step 2350, Loss: {'train/loss': 0.4526764452457428, 'train/grad_norm_pre_clip': 7.010700702667236, 'timing/data': 0.0009713564068078995, 'timing/model': 1.7096793064847589, 'learning_rate/qwen_vl_interface': 7.718023762668309e-06, 'learning_rate/action_model': 7.718023762668309e-05, 'epoch': 0.9}) +[2026-06-17 21:19:34,079][starVLA.training.train_starvla][INFO] - Step 2351, Loss: {'train/loss': 0.3423544466495514, 'train/grad_norm_pre_clip': 7.729765892028809, 'timing/data': 0.0012846970930695534, 'timing/model': 1.1598397865891457, 'learning_rate/qwen_vl_interface': 7.716175662109882e-06, 'learning_rate/action_model': 7.716175662109881e-05, 'epoch': 0.9}) +[2026-06-17 21:19:45,021][starVLA.training.train_starvla][INFO] - Step 2352, Loss: {'train/loss': 0.39740636944770813, 'train/grad_norm_pre_clip': 13.054421424865723, 'timing/data': 0.0014395928010344505, 'timing/model': 1.482359983958304, 'learning_rate/qwen_vl_interface': 7.714327050310586e-06, 'learning_rate/action_model': 7.714327050310585e-05, 'epoch': 0.9}) +[2026-06-17 21:19:51,598][starVLA.training.train_starvla][INFO] - Step 2353, Loss: {'train/loss': 0.4815283715724945, 'train/grad_norm_pre_clip': 8.927513122558594, 'timing/data': 0.0012373421341180801, 'timing/model': 1.172560359351337, 'learning_rate/qwen_vl_interface': 7.71247792765364e-06, 'learning_rate/action_model': 7.712477927653638e-05, 'epoch': 0.91}) +[2026-06-17 21:20:07,373][starVLA.training.train_starvla][INFO] - Step 2354, Loss: {'train/loss': 0.6171534061431885, 'train/grad_norm_pre_clip': 12.371752738952637, 'timing/data': 0.0019637523218989372, 'timing/model': 2.92165460344404, 'learning_rate/qwen_vl_interface': 7.71062829452237e-06, 'learning_rate/action_model': 7.710628294522369e-05, 'epoch': 0.91}) +[2026-06-17 21:20:21,014][starVLA.training.train_starvla][INFO] - Step 2355, Loss: {'train/loss': 0.5283776521682739, 'train/grad_norm_pre_clip': 7.845637321472168, 'timing/data': 0.000993901863694191, 'timing/model': 4.176608447916806, 'learning_rate/qwen_vl_interface': 7.708778151300206e-06, 'learning_rate/action_model': 7.708778151300206e-05, 'epoch': 0.91}) +[2026-06-17 21:20:32,775][starVLA.training.train_starvla][INFO] - Step 2356, Loss: {'train/loss': 0.6664970517158508, 'train/grad_norm_pre_clip': 27.02615737915039, 'timing/data': 0.0011520357802510262, 'timing/model': 1.1201055711135268, 'learning_rate/qwen_vl_interface': 7.706927498370687e-06, 'learning_rate/action_model': 7.706927498370686e-05, 'epoch': 0.91}) +[2026-06-17 21:20:42,062][starVLA.training.train_starvla][INFO] - Step 2357, Loss: {'train/loss': 0.5943456888198853, 'train/grad_norm_pre_clip': 10.879775047302246, 'timing/data': 0.001230323687195778, 'timing/model': 4.1874881172552705, 'learning_rate/qwen_vl_interface': 7.705076336117452e-06, 'learning_rate/action_model': 7.705076336117451e-05, 'epoch': 0.91}) +[2026-06-17 21:20:46,704][starVLA.training.train_starvla][INFO] - Step 2358, Loss: {'train/loss': 0.6691620349884033, 'train/grad_norm_pre_clip': 9.15383243560791, 'timing/data': 0.0010011065751314163, 'timing/model': 1.0794906429946423, 'learning_rate/qwen_vl_interface': 7.703224664924252e-06, 'learning_rate/action_model': 7.703224664924251e-05, 'epoch': 0.91}) +[2026-06-17 21:20:52,546][starVLA.training.train_starvla][INFO] - Step 2359, Loss: {'train/loss': 0.41828835010528564, 'train/grad_norm_pre_clip': 10.029027938842773, 'timing/data': 0.0010912371799349785, 'timing/model': 1.2655130634084344, 'learning_rate/qwen_vl_interface': 7.70137248517494e-06, 'learning_rate/action_model': 7.701372485174939e-05, 'epoch': 0.91}) +[2026-06-17 21:20:57,409][starVLA.training.train_starvla][INFO] - Step 2360, Loss: {'train/loss': 0.40794846415519714, 'train/grad_norm_pre_clip': 7.01186990737915, 'timing/data': 0.0009608957916498184, 'timing/model': 1.1027339911088347, 'learning_rate/qwen_vl_interface': 7.699519797253471e-06, 'learning_rate/action_model': 7.699519797253472e-05, 'epoch': 0.91}) +[2026-06-17 21:21:02,098][starVLA.training.train_starvla][INFO] - Step 2361, Loss: {'train/loss': 0.4290928840637207, 'train/grad_norm_pre_clip': 8.429500579833984, 'timing/data': 0.0014752056449651718, 'timing/model': 1.1635602386668324, 'learning_rate/qwen_vl_interface': 7.697666601543915e-06, 'learning_rate/action_model': 7.697666601543915e-05, 'epoch': 0.91}) +[2026-06-17 21:21:06,908][starVLA.training.train_starvla][INFO] - Step 2362, Loss: {'train/loss': 0.425567626953125, 'train/grad_norm_pre_clip': 7.972314834594727, 'timing/data': 0.0014535319060087204, 'timing/model': 1.130017939954996, 'learning_rate/qwen_vl_interface': 7.695812898430438e-06, 'learning_rate/action_model': 7.695812898430437e-05, 'epoch': 0.91}) +[2026-06-17 21:21:14,949][starVLA.training.train_starvla][INFO] - Step 2363, Loss: {'train/loss': 0.6008375883102417, 'train/grad_norm_pre_clip': 10.428580284118652, 'timing/data': 0.001238144002854824, 'timing/model': 4.068122335709631, 'learning_rate/qwen_vl_interface': 7.693958688297316e-06, 'learning_rate/action_model': 7.693958688297314e-05, 'epoch': 0.91}) +[2026-06-17 21:21:30,026][starVLA.training.train_starvla][INFO] - Step 2364, Loss: {'train/loss': 0.5725732445716858, 'train/grad_norm_pre_clip': 11.356306076049805, 'timing/data': 0.001157977618277073, 'timing/model': 1.1076514963060617, 'learning_rate/qwen_vl_interface': 7.692103971528925e-06, 'learning_rate/action_model': 7.692103971528926e-05, 'epoch': 0.91}) +[2026-06-17 21:21:35,363][starVLA.training.train_starvla][INFO] - Step 2365, Loss: {'train/loss': 0.3516906499862671, 'train/grad_norm_pre_clip': 7.420144081115723, 'timing/data': 0.0010932395234704018, 'timing/model': 1.241770075634122, 'learning_rate/qwen_vl_interface': 7.690248748509757e-06, 'learning_rate/action_model': 7.690248748509757e-05, 'epoch': 0.91}) +[2026-06-17 21:21:40,256][starVLA.training.train_starvla][INFO] - Step 2366, Loss: {'train/loss': 0.729621410369873, 'train/grad_norm_pre_clip': 21.65349578857422, 'timing/data': 0.0013260291889309883, 'timing/model': 1.1528067784383893, 'learning_rate/qwen_vl_interface': 7.688393019624396e-06, 'learning_rate/action_model': 7.688393019624395e-05, 'epoch': 0.91}) +[2026-06-17 21:21:48,314][starVLA.training.train_starvla][INFO] - Step 2367, Loss: {'train/loss': 1.0959956645965576, 'train/grad_norm_pre_clip': 18.21309471130371, 'timing/data': 0.0014653094112873077, 'timing/model': 1.4543015584349632, 'learning_rate/qwen_vl_interface': 7.686536785257539e-06, 'learning_rate/action_model': 7.686536785257538e-05, 'epoch': 0.91}) +[2026-06-17 21:21:52,932][starVLA.training.train_starvla][INFO] - Step 2368, Loss: {'train/loss': 0.8628005385398865, 'train/grad_norm_pre_clip': 8.205910682678223, 'timing/data': 0.0021288860589265823, 'timing/model': 1.1295280121266842, 'learning_rate/qwen_vl_interface': 7.684680045793982e-06, 'learning_rate/action_model': 7.684680045793983e-05, 'epoch': 0.91}) +[2026-06-17 21:21:57,728][starVLA.training.train_starvla][INFO] - Step 2369, Loss: {'train/loss': 0.6136645674705505, 'train/grad_norm_pre_clip': 8.479935646057129, 'timing/data': 0.0011753728613257408, 'timing/model': 1.101027213037014, 'learning_rate/qwen_vl_interface': 7.682822801618634e-06, 'learning_rate/action_model': 7.682822801618633e-05, 'epoch': 0.91}) +[2026-06-17 21:22:02,688][starVLA.training.train_starvla][INFO] - Step 2370, Loss: {'train/loss': 0.518999457359314, 'train/grad_norm_pre_clip': 6.594295024871826, 'timing/data': 0.0012374883517622948, 'timing/model': 1.3120366167277098, 'learning_rate/qwen_vl_interface': 7.680965053116498e-06, 'learning_rate/action_model': 7.680965053116499e-05, 'epoch': 0.91}) +[2026-06-17 21:22:09,218][starVLA.training.train_starvla][INFO] - Step 2371, Loss: {'train/loss': 0.35814428329467773, 'train/grad_norm_pre_clip': 6.954507350921631, 'timing/data': 0.000907248817384243, 'timing/model': 1.288045832887292, 'learning_rate/qwen_vl_interface': 7.679106800672695e-06, 'learning_rate/action_model': 7.679106800672695e-05, 'epoch': 0.91}) +[2026-06-17 21:22:13,678][starVLA.training.train_starvla][INFO] - Step 2372, Loss: {'train/loss': 0.33408331871032715, 'train/grad_norm_pre_clip': 6.048761367797852, 'timing/data': 0.0011782580986618996, 'timing/model': 1.114538257010281, 'learning_rate/qwen_vl_interface': 7.677248044672438e-06, 'learning_rate/action_model': 7.677248044672437e-05, 'epoch': 0.91}) +[2026-06-17 21:22:18,615][starVLA.training.train_starvla][INFO] - Step 2373, Loss: {'train/loss': 0.48945364356040955, 'train/grad_norm_pre_clip': 9.247504234313965, 'timing/data': 0.0011048996821045876, 'timing/model': 1.0737468507140875, 'learning_rate/qwen_vl_interface': 7.675388785501048e-06, 'learning_rate/action_model': 7.675388785501047e-05, 'epoch': 0.91}) +[2026-06-17 21:22:27,020][starVLA.training.train_starvla][INFO] - Step 2374, Loss: {'train/loss': 0.45738399028778076, 'train/grad_norm_pre_clip': 7.495805263519287, 'timing/data': 0.0012389104813337326, 'timing/model': 1.183986003510654, 'learning_rate/qwen_vl_interface': 7.673529023543951e-06, 'learning_rate/action_model': 7.673529023543951e-05, 'epoch': 0.91}) +[2026-06-17 21:22:39,631][starVLA.training.train_starvla][INFO] - Step 2375, Loss: {'train/loss': 0.1993623673915863, 'train/grad_norm_pre_clip': 6.279899597167969, 'timing/data': 0.0013093175366520882, 'timing/model': 1.0987144988030195, 'learning_rate/qwen_vl_interface': 7.671668759186683e-06, 'learning_rate/action_model': 7.671668759186683e-05, 'epoch': 0.91}) +[2026-06-17 21:22:44,674][starVLA.training.train_starvla][INFO] - Step 2376, Loss: {'train/loss': 0.6143481135368347, 'train/grad_norm_pre_clip': 7.960227966308594, 'timing/data': 0.0020806454122066498, 'timing/model': 1.2041852679103613, 'learning_rate/qwen_vl_interface': 7.669807992814875e-06, 'learning_rate/action_model': 7.669807992814875e-05, 'epoch': 0.91}) +[2026-06-17 21:22:54,949][starVLA.training.train_starvla][INFO] - Step 2377, Loss: {'train/loss': 0.26911628246307373, 'train/grad_norm_pre_clip': 6.043102741241455, 'timing/data': 0.0010457057505846024, 'timing/model': 1.6683179792016745, 'learning_rate/qwen_vl_interface': 7.667946724814264e-06, 'learning_rate/action_model': 7.667946724814265e-05, 'epoch': 0.91}) +[2026-06-17 21:23:07,102][starVLA.training.train_starvla][INFO] - Step 2378, Loss: {'train/loss': 0.709308385848999, 'train/grad_norm_pre_clip': 13.544275283813477, 'timing/data': 0.0012417370453476906, 'timing/model': 4.695369217544794, 'learning_rate/qwen_vl_interface': 7.666084955570698e-06, 'learning_rate/action_model': 7.666084955570698e-05, 'epoch': 0.91}) +[2026-06-17 21:23:12,728][starVLA.training.train_starvla][INFO] - Step 2379, Loss: {'train/loss': 0.6277512311935425, 'train/grad_norm_pre_clip': 12.717944145202637, 'timing/data': 0.0010669194161891937, 'timing/model': 1.6357909692451358, 'learning_rate/qwen_vl_interface': 7.664222685470119e-06, 'learning_rate/action_model': 7.664222685470118e-05, 'epoch': 0.92}) +[2026-06-17 21:23:20,853][starVLA.training.train_starvla][INFO] - Step 2380, Loss: {'train/loss': 0.4520684778690338, 'train/grad_norm_pre_clip': 10.577469825744629, 'timing/data': 0.000939621590077877, 'timing/model': 4.515977299772203, 'learning_rate/qwen_vl_interface': 7.66235991489858e-06, 'learning_rate/action_model': 7.66235991489858e-05, 'epoch': 0.92}) +[2026-06-17 21:23:26,339][starVLA.training.train_starvla][INFO] - Step 2381, Loss: {'train/loss': 0.6879075765609741, 'train/grad_norm_pre_clip': 9.640055656433105, 'timing/data': 0.0012458758428692818, 'timing/model': 1.3621831107884645, 'learning_rate/qwen_vl_interface': 7.660496644242233e-06, 'learning_rate/action_model': 7.660496644242233e-05, 'epoch': 0.92}) +[2026-06-17 21:23:32,316][starVLA.training.train_starvla][INFO] - Step 2382, Loss: {'train/loss': 0.33282017707824707, 'train/grad_norm_pre_clip': 9.544931411743164, 'timing/data': 0.0013459408655762672, 'timing/model': 1.0885496847331524, 'learning_rate/qwen_vl_interface': 7.658632873887342e-06, 'learning_rate/action_model': 7.658632873887342e-05, 'epoch': 0.92}) +[2026-06-17 21:23:37,930][starVLA.training.train_starvla][INFO] - Step 2383, Loss: {'train/loss': 0.49795976281166077, 'train/grad_norm_pre_clip': 9.28307056427002, 'timing/data': 0.0008704671636223793, 'timing/model': 1.0706507163122296, 'learning_rate/qwen_vl_interface': 7.656768604220262e-06, 'learning_rate/action_model': 7.656768604220261e-05, 'epoch': 0.92}) +[2026-06-17 21:23:43,004][starVLA.training.train_starvla][INFO] - Step 2384, Loss: {'train/loss': 0.2544375956058502, 'train/grad_norm_pre_clip': 6.386267185211182, 'timing/data': 0.0015877941623330116, 'timing/model': 1.1741043822839856, 'learning_rate/qwen_vl_interface': 7.654903835627459e-06, 'learning_rate/action_model': 7.654903835627458e-05, 'epoch': 0.92}) +[2026-06-17 21:23:47,933][starVLA.training.train_starvla][INFO] - Step 2385, Loss: {'train/loss': 0.6596428751945496, 'train/grad_norm_pre_clip': 11.408960342407227, 'timing/data': 0.0009338520467281342, 'timing/model': 1.0720975575968623, 'learning_rate/qwen_vl_interface': 7.653038568495505e-06, 'learning_rate/action_model': 7.653038568495506e-05, 'epoch': 0.92}) +[2026-06-17 21:23:52,885][starVLA.training.train_starvla][INFO] - Step 2386, Loss: {'train/loss': 0.8352051973342896, 'train/grad_norm_pre_clip': 13.874944686889648, 'timing/data': 0.001283048652112484, 'timing/model': 1.0744954301044345, 'learning_rate/qwen_vl_interface': 7.651172803211068e-06, 'learning_rate/action_model': 7.651172803211068e-05, 'epoch': 0.92}) +[2026-06-17 21:23:57,925][starVLA.training.train_starvla][INFO] - Step 2387, Loss: {'train/loss': 0.4304576516151428, 'train/grad_norm_pre_clip': 7.416922569274902, 'timing/data': 0.0011192606762051582, 'timing/model': 1.5298664588481188, 'learning_rate/qwen_vl_interface': 7.649306540160928e-06, 'learning_rate/action_model': 7.649306540160927e-05, 'epoch': 0.92}) +[2026-06-17 21:24:07,344][starVLA.training.train_starvla][INFO] - Step 2388, Loss: {'train/loss': 0.7624794840812683, 'train/grad_norm_pre_clip': 10.330589294433594, 'timing/data': 0.0008965376764535904, 'timing/model': 1.1773964827880263, 'learning_rate/qwen_vl_interface': 7.647439779731959e-06, 'learning_rate/action_model': 7.647439779731958e-05, 'epoch': 0.92}) +[2026-06-17 21:24:12,075][starVLA.training.train_starvla][INFO] - Step 2389, Loss: {'train/loss': 0.41431671380996704, 'train/grad_norm_pre_clip': 7.239593505859375, 'timing/data': 0.0013176044449210167, 'timing/model': 1.1513373274356127, 'learning_rate/qwen_vl_interface': 7.645572522311144e-06, 'learning_rate/action_model': 7.645572522311144e-05, 'epoch': 0.92}) +[2026-06-17 21:24:17,068][starVLA.training.train_starvla][INFO] - Step 2390, Loss: {'train/loss': 0.4693981409072876, 'train/grad_norm_pre_clip': 8.582283973693848, 'timing/data': 0.0009493222460150719, 'timing/model': 1.0998111767694354, 'learning_rate/qwen_vl_interface': 7.643704768285568e-06, 'learning_rate/action_model': 7.643704768285568e-05, 'epoch': 0.92}) +[2026-06-17 21:24:26,452][starVLA.training.train_starvla][INFO] - Step 2391, Loss: {'train/loss': 0.5677151679992676, 'train/grad_norm_pre_clip': 10.645249366760254, 'timing/data': 0.001253741793334484, 'timing/model': 1.1868459023535252, 'learning_rate/qwen_vl_interface': 7.641836518042417e-06, 'learning_rate/action_model': 7.641836518042416e-05, 'epoch': 0.92}) +[2026-06-17 21:24:35,350][starVLA.training.train_starvla][INFO] - Step 2392, Loss: {'train/loss': 0.6105310320854187, 'train/grad_norm_pre_clip': 13.68373966217041, 'timing/data': 0.0011658156290650368, 'timing/model': 1.1235513649880886, 'learning_rate/qwen_vl_interface': 7.639967771968983e-06, 'learning_rate/action_model': 7.639967771968983e-05, 'epoch': 0.92}) +[2026-06-17 21:24:40,233][starVLA.training.train_starvla][INFO] - Step 2393, Loss: {'train/loss': 0.4111197292804718, 'train/grad_norm_pre_clip': 10.670308113098145, 'timing/data': 0.001126476563513279, 'timing/model': 1.1288951253518462, 'learning_rate/qwen_vl_interface': 7.638098530452657e-06, 'learning_rate/action_model': 7.638098530452658e-05, 'epoch': 0.92}) +[2026-06-17 21:24:44,937][starVLA.training.train_starvla][INFO] - Step 2394, Loss: {'train/loss': 0.4597235918045044, 'train/grad_norm_pre_clip': 11.984106063842773, 'timing/data': 0.0013369722291827202, 'timing/model': 1.2470759432762861, 'learning_rate/qwen_vl_interface': 7.636228793880935e-06, 'learning_rate/action_model': 7.636228793880935e-05, 'epoch': 0.92}) +[2026-06-17 21:24:52,728][starVLA.training.train_starvla][INFO] - Step 2395, Loss: {'train/loss': 0.4678424298763275, 'train/grad_norm_pre_clip': 9.01152515411377, 'timing/data': 0.0011591017246246338, 'timing/model': 4.04867790825665, 'learning_rate/qwen_vl_interface': 7.63435856264142e-06, 'learning_rate/action_model': 7.63435856264142e-05, 'epoch': 0.92}) +[2026-06-17 21:25:00,943][starVLA.training.train_starvla][INFO] - Step 2396, Loss: {'train/loss': 0.5816420316696167, 'train/grad_norm_pre_clip': 9.427762031555176, 'timing/data': 0.04595664795488119, 'timing/model': 1.2414161749184132, 'learning_rate/qwen_vl_interface': 7.632487837121806e-06, 'learning_rate/action_model': 7.632487837121806e-05, 'epoch': 0.92}) +[2026-06-17 21:25:05,754][starVLA.training.train_starvla][INFO] - Step 2397, Loss: {'train/loss': 0.5860974788665771, 'train/grad_norm_pre_clip': 12.01286506652832, 'timing/data': 0.0007950235158205032, 'timing/model': 1.5652698203921318, 'learning_rate/qwen_vl_interface': 7.630616617709902e-06, 'learning_rate/action_model': 7.630616617709902e-05, 'epoch': 0.92}) +[2026-06-17 21:25:10,311][starVLA.training.train_starvla][INFO] - Step 2398, Loss: {'train/loss': 0.41825565695762634, 'train/grad_norm_pre_clip': 7.578214645385742, 'timing/data': 0.0012531150132417679, 'timing/model': 1.0711909625679255, 'learning_rate/qwen_vl_interface': 7.6287449047936095e-06, 'learning_rate/action_model': 7.628744904793609e-05, 'epoch': 0.92}) +[2026-06-17 21:25:15,070][starVLA.training.train_starvla][INFO] - Step 2399, Loss: {'train/loss': 0.5911135673522949, 'train/grad_norm_pre_clip': 10.815179824829102, 'timing/data': 0.0013487283140420914, 'timing/model': 1.1727416645735502, 'learning_rate/qwen_vl_interface': 7.6268726987609384e-06, 'learning_rate/action_model': 7.626872698760939e-05, 'epoch': 0.92}) +[2026-06-17 21:29:23,637][starVLA.training.train_starvla][INFO] - Step 2400, Loss: {'train/loss': 0.9067716002464294, 'train/grad_norm_pre_clip': 13.068632125854492, 'eval/loss': 0.5191028714179993, 'eval/latency_4/loss': 0.5191028714179993, 'eval/demon_fire_precision': 0.9535175879396985, 'eval/demon_fire_recall': 0.9860344267619356, 'eval/demon_fire_f1': 0.969503432859652, 'eval/demon_fire_rate_model': 0.9350954478707783, 'eval/demon_fire_rate_teacher': 0.9042584434654919, 'eval/demon_left_precision': 0.7027027027027027, 'eval/demon_left_recall': 0.593607305936073, 'eval/demon_left_f1': 0.6435643564356436, 'eval/demon_left_rate_model': 0.10866372980910426, 'eval/demon_left_rate_teacher': 0.12863436123348018, 'eval/demon_right_precision': 0.7801724137931034, 'eval/demon_right_recall': 0.6534296028880866, 'eval/demon_right_f1': 0.7111984282907662, 'eval/demon_right_rate_model': 0.13627019089574155, 'eval/demon_right_rate_teacher': 0.16270190895741557, 'eval/demon_move_f1': 0.6773813923632048, 'eval/demon_cc_f1': 0.8584419992444209, 'eval/latency_4/demon_fire_precision': 0.9535175879396985, 'eval/latency_4/demon_fire_recall': 0.9860344267619356, 'eval/latency_4/demon_fire_f1': 0.969503432859652, 'eval/latency_4/demon_fire_rate_model': 0.9350954478707783, 'eval/latency_4/demon_fire_rate_teacher': 0.9042584434654919, 'eval/latency_4/demon_left_precision': 0.7027027027027027, 'eval/latency_4/demon_left_recall': 0.593607305936073, 'eval/latency_4/demon_left_f1': 0.6435643564356436, 'eval/latency_4/demon_left_rate_model': 0.10866372980910426, 'eval/latency_4/demon_left_rate_teacher': 0.12863436123348018, 'eval/latency_4/demon_right_precision': 0.7801724137931034, 'eval/latency_4/demon_right_recall': 0.6534296028880866, 'eval/latency_4/demon_right_f1': 0.7111984282907662, 'eval/latency_4/demon_right_rate_model': 0.13627019089574155, 'eval/latency_4/demon_right_rate_teacher': 0.16270190895741557, 'eval/latency_4/demon_move_f1': 0.6773813923632048, 'eval/latency_4/demon_cc_f1': 0.8584419992444209, 'timing/data': 0.001087152399122715, 'timing/model': 1.8920717518776655, 'learning_rate/qwen_vl_interface': 7.625e-06, 'learning_rate/action_model': 7.625e-05, 'epoch': 0.92}) +[2026-06-17 21:29:23,651][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_2400_state +[2026-06-17 21:29:23,660][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 21:29:32,640][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_2400_state/model.safetensors +[2026-06-17 21:29:54,926][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_2400_state/optimizer.bin +[2026-06-17 21:29:54,927][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_2400_state/sampler.bin +[2026-06-17 21:29:54,927][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_2400_state/sampler_1.bin +[2026-06-17 21:29:54,930][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_2400_state/random_states_0.pkl +[2026-06-17 21:29:54,937][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2400 +[2026-06-17 21:29:54,937][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 21:29:54,956][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 21:30:00,035][starVLA.training.train_starvla][INFO] - Step 2401, Loss: {'train/loss': 0.24607929587364197, 'train/grad_norm_pre_clip': 5.433911323547363, 'timing/data': 0.0011332528665661812, 'timing/model': 1.2348365001380444, 'learning_rate/qwen_vl_interface': 7.6231268088990075e-06, 'learning_rate/action_model': 7.623126808899008e-05, 'epoch': 0.92}) +[2026-06-17 21:30:05,920][starVLA.training.train_starvla][INFO] - Step 2402, Loss: {'train/loss': 0.4636019170284271, 'train/grad_norm_pre_clip': 10.955425262451172, 'timing/data': 0.0012862887233495712, 'timing/model': 1.517685311846435, 'learning_rate/qwen_vl_interface': 7.621253125846273e-06, 'learning_rate/action_model': 7.621253125846273e-05, 'epoch': 0.92}) +[2026-06-17 21:30:10,315][starVLA.training.train_starvla][INFO] - Step 2403, Loss: {'train/loss': 0.37622132897377014, 'train/grad_norm_pre_clip': 10.74953556060791, 'timing/data': 0.0011782553046941757, 'timing/model': 1.0841473303735256, 'learning_rate/qwen_vl_interface': 7.619378951230213e-06, 'learning_rate/action_model': 7.619378951230213e-05, 'epoch': 0.92}) +[2026-06-17 21:30:15,136][starVLA.training.train_starvla][INFO] - Step 2404, Loss: {'train/loss': 0.28376102447509766, 'train/grad_norm_pre_clip': 5.891241550445557, 'timing/data': 0.0011515682563185692, 'timing/model': 1.234356002882123, 'learning_rate/qwen_vl_interface': 7.617504285439346e-06, 'learning_rate/action_model': 7.617504285439346e-05, 'epoch': 0.92}) +[2026-06-17 21:30:22,775][starVLA.training.train_starvla][INFO] - Step 2405, Loss: {'train/loss': 0.7808217406272888, 'train/grad_norm_pre_clip': 12.69630241394043, 'timing/data': 0.0011634379625320435, 'timing/model': 4.064642392098904, 'learning_rate/qwen_vl_interface': 7.615629128862295e-06, 'learning_rate/action_model': 7.615629128862295e-05, 'epoch': 0.93}) +[2026-06-17 21:30:27,701][starVLA.training.train_starvla][INFO] - Step 2406, Loss: {'train/loss': 0.6576436161994934, 'train/grad_norm_pre_clip': 12.1289644241333, 'timing/data': 0.0014393161982297897, 'timing/model': 1.231355207040906, 'learning_rate/qwen_vl_interface': 7.61375348188778e-06, 'learning_rate/action_model': 7.61375348188778e-05, 'epoch': 0.93}) +[2026-06-17 21:30:32,355][starVLA.training.train_starvla][INFO] - Step 2407, Loss: {'train/loss': 0.7722082138061523, 'train/grad_norm_pre_clip': 12.92892837524414, 'timing/data': 0.0011461256071925163, 'timing/model': 1.4503728449344635, 'learning_rate/qwen_vl_interface': 7.611877344904624e-06, 'learning_rate/action_model': 7.611877344904624e-05, 'epoch': 0.93}) +[2026-06-17 21:30:39,633][starVLA.training.train_starvla][INFO] - Step 2408, Loss: {'train/loss': 0.2772299647331238, 'train/grad_norm_pre_clip': 7.5720391273498535, 'timing/data': 0.0012887082993984222, 'timing/model': 3.851381318643689, 'learning_rate/qwen_vl_interface': 7.610000718301753e-06, 'learning_rate/action_model': 7.610000718301753e-05, 'epoch': 0.93}) +[2026-06-17 21:30:46,083][starVLA.training.train_starvla][INFO] - Step 2409, Loss: {'train/loss': 0.5433983206748962, 'train/grad_norm_pre_clip': 9.14310073852539, 'timing/data': 0.00103798508644104, 'timing/model': 1.0877522500231862, 'learning_rate/qwen_vl_interface': 7.608123602468195e-06, 'learning_rate/action_model': 7.608123602468194e-05, 'epoch': 0.93}) +[2026-06-17 21:30:50,922][starVLA.training.train_starvla][INFO] - Step 2410, Loss: {'train/loss': 0.4938579797744751, 'train/grad_norm_pre_clip': 9.000752449035645, 'timing/data': 0.0012421263381838799, 'timing/model': 1.112623275257647, 'learning_rate/qwen_vl_interface': 7.606245997793074e-06, 'learning_rate/action_model': 7.606245997793075e-05, 'epoch': 0.93}) +[2026-06-17 21:30:56,329][starVLA.training.train_starvla][INFO] - Step 2411, Loss: {'train/loss': 0.7523273825645447, 'train/grad_norm_pre_clip': 10.199464797973633, 'timing/data': 0.0009569618850946426, 'timing/model': 1.284864755347371, 'learning_rate/qwen_vl_interface': 7.604367904665625e-06, 'learning_rate/action_model': 7.604367904665624e-05, 'epoch': 0.93}) +[2026-06-17 21:31:04,981][starVLA.training.train_starvla][INFO] - Step 2412, Loss: {'train/loss': 1.0900518894195557, 'train/grad_norm_pre_clip': 14.772329330444336, 'timing/data': 0.0017593279480934143, 'timing/model': 1.5926933521404862, 'learning_rate/qwen_vl_interface': 7.602489323475174e-06, 'learning_rate/action_model': 7.602489323475174e-05, 'epoch': 0.93}) +[2026-06-17 21:31:09,628][starVLA.training.train_starvla][INFO] - Step 2413, Loss: {'train/loss': 0.33382198214530945, 'train/grad_norm_pre_clip': 9.09687614440918, 'timing/data': 0.0012027816846966743, 'timing/model': 1.2529214294627309, 'learning_rate/qwen_vl_interface': 7.600610254611155e-06, 'learning_rate/action_model': 7.600610254611155e-05, 'epoch': 0.93}) +[2026-06-17 21:31:15,421][starVLA.training.train_starvla][INFO] - Step 2414, Loss: {'train/loss': 0.5215259194374084, 'train/grad_norm_pre_clip': 10.28553295135498, 'timing/data': 0.0015011224895715714, 'timing/model': 1.0826694117859006, 'learning_rate/qwen_vl_interface': 7.5987306984631025e-06, 'learning_rate/action_model': 7.598730698463102e-05, 'epoch': 0.93}) +[2026-06-17 21:31:20,240][starVLA.training.train_starvla][INFO] - Step 2415, Loss: {'train/loss': 0.52690190076828, 'train/grad_norm_pre_clip': 9.916096687316895, 'timing/data': 0.0014381147921085358, 'timing/model': 1.090208563953638, 'learning_rate/qwen_vl_interface': 7.596850655420647e-06, 'learning_rate/action_model': 7.596850655420646e-05, 'epoch': 0.93}) +[2026-06-17 21:31:25,108][starVLA.training.train_starvla][INFO] - Step 2416, Loss: {'train/loss': 0.6302534937858582, 'train/grad_norm_pre_clip': 10.759893417358398, 'timing/data': 0.0011015646159648895, 'timing/model': 1.0982877481728792, 'learning_rate/qwen_vl_interface': 7.5949701258735265e-06, 'learning_rate/action_model': 7.594970125873526e-05, 'epoch': 0.93}) +[2026-06-17 21:31:30,730][starVLA.training.train_starvla][INFO] - Step 2417, Loss: {'train/loss': 0.6732305884361267, 'train/grad_norm_pre_clip': 11.917476654052734, 'timing/data': 0.0009229900315403938, 'timing/model': 1.1956096310168505, 'learning_rate/qwen_vl_interface': 7.5930891102115735e-06, 'learning_rate/action_model': 7.593089110211573e-05, 'epoch': 0.93}) +[2026-06-17 21:31:39,474][starVLA.training.train_starvla][INFO] - Step 2418, Loss: {'train/loss': 0.8867565393447876, 'train/grad_norm_pre_clip': 11.144524574279785, 'timing/data': 0.0015038251876831055, 'timing/model': 1.3647051779553294, 'learning_rate/qwen_vl_interface': 7.59120760882473e-06, 'learning_rate/action_model': 7.59120760882473e-05, 'epoch': 0.93}) +[2026-06-17 21:31:46,452][starVLA.training.train_starvla][INFO] - Step 2419, Loss: {'train/loss': 0.5145534873008728, 'train/grad_norm_pre_clip': 9.499713897705078, 'timing/data': 0.0024332329630851746, 'timing/model': 1.1142734559252858, 'learning_rate/qwen_vl_interface': 7.5893256221030275e-06, 'learning_rate/action_model': 7.589325622103027e-05, 'epoch': 0.93}) +[2026-06-17 21:31:55,419][starVLA.training.train_starvla][INFO] - Step 2420, Loss: {'train/loss': 0.5456286072731018, 'train/grad_norm_pre_clip': 9.711941719055176, 'timing/data': 0.0012974860146641731, 'timing/model': 1.4018244659528136, 'learning_rate/qwen_vl_interface': 7.587443150436608e-06, 'learning_rate/action_model': 7.587443150436608e-05, 'epoch': 0.93}) +[2026-06-17 21:32:11,446][starVLA.training.train_starvla][INFO] - Step 2421, Loss: {'train/loss': 0.4408573806285858, 'train/grad_norm_pre_clip': 10.60805606842041, 'timing/data': 0.0010925345122814178, 'timing/model': 4.124272440560162, 'learning_rate/qwen_vl_interface': 7.585560194215706e-06, 'learning_rate/action_model': 7.585560194215706e-05, 'epoch': 0.93}) +[2026-06-17 21:32:16,207][starVLA.training.train_starvla][INFO] - Step 2422, Loss: {'train/loss': 0.6654005646705627, 'train/grad_norm_pre_clip': 11.641236305236816, 'timing/data': 0.0016428297385573387, 'timing/model': 1.5334553215652704, 'learning_rate/qwen_vl_interface': 7.583676753830665e-06, 'learning_rate/action_model': 7.583676753830664e-05, 'epoch': 0.93}) +[2026-06-17 21:32:30,261][starVLA.training.train_starvla][INFO] - Step 2423, Loss: {'train/loss': 0.3370106518268585, 'train/grad_norm_pre_clip': 6.3132781982421875, 'timing/data': 0.0016713803634047508, 'timing/model': 4.712049673311412, 'learning_rate/qwen_vl_interface': 7.58179282967192e-06, 'learning_rate/action_model': 7.58179282967192e-05, 'epoch': 0.93}) +[2026-06-17 21:32:45,232][starVLA.training.train_starvla][INFO] - Step 2424, Loss: {'train/loss': 0.5792210698127747, 'train/grad_norm_pre_clip': 10.568387031555176, 'timing/data': 0.0012894310057163239, 'timing/model': 1.6979580903425813, 'learning_rate/qwen_vl_interface': 7.579908422130013e-06, 'learning_rate/action_model': 7.579908422130014e-05, 'epoch': 0.93}) +[2026-06-17 21:32:54,616][starVLA.training.train_starvla][INFO] - Step 2425, Loss: {'train/loss': 0.592158854007721, 'train/grad_norm_pre_clip': 8.127114295959473, 'timing/data': 0.00145723856985569, 'timing/model': 1.4558103727176785, 'learning_rate/qwen_vl_interface': 7.578023531595583e-06, 'learning_rate/action_model': 7.578023531595583e-05, 'epoch': 0.93}) +[2026-06-17 21:32:59,186][starVLA.training.train_starvla][INFO] - Step 2426, Loss: {'train/loss': 0.6370210647583008, 'train/grad_norm_pre_clip': 9.516477584838867, 'timing/data': 0.0011711493134498596, 'timing/model': 1.0759720718488097, 'learning_rate/qwen_vl_interface': 7.576138158459369e-06, 'learning_rate/action_model': 7.576138158459368e-05, 'epoch': 0.93}) +[2026-06-17 21:33:03,893][starVLA.training.train_starvla][INFO] - Step 2427, Loss: {'train/loss': 0.49917447566986084, 'train/grad_norm_pre_clip': 9.181963920593262, 'timing/data': 0.0008879508823156357, 'timing/model': 1.0924186436459422, 'learning_rate/qwen_vl_interface': 7.574252303112213e-06, 'learning_rate/action_model': 7.574252303112212e-05, 'epoch': 0.93}) +[2026-06-17 21:33:08,928][starVLA.training.train_starvla][INFO] - Step 2428, Loss: {'train/loss': 0.2682941257953644, 'train/grad_norm_pre_clip': 6.099678993225098, 'timing/data': 0.0019176872447133064, 'timing/model': 1.161024209111929, 'learning_rate/qwen_vl_interface': 7.572365965945053e-06, 'learning_rate/action_model': 7.572365965945052e-05, 'epoch': 0.93}) +[2026-06-17 21:33:13,677][starVLA.training.train_starvla][INFO] - Step 2429, Loss: {'train/loss': 0.46776989102363586, 'train/grad_norm_pre_clip': 8.121089935302734, 'timing/data': 0.001144135370850563, 'timing/model': 1.1429724479094148, 'learning_rate/qwen_vl_interface': 7.570479147348929e-06, 'learning_rate/action_model': 7.570479147348929e-05, 'epoch': 0.93}) +[2026-06-17 21:33:18,417][starVLA.training.train_starvla][INFO] - Step 2430, Loss: {'train/loss': 0.6019809246063232, 'train/grad_norm_pre_clip': 8.348179817199707, 'timing/data': 0.0007666479796171188, 'timing/model': 1.0705822054296732, 'learning_rate/qwen_vl_interface': 7.568591847714983e-06, 'learning_rate/action_model': 7.568591847714982e-05, 'epoch': 0.93}) +[2026-06-17 21:33:23,086][starVLA.training.train_starvla][INFO] - Step 2431, Loss: {'train/loss': 0.7242377400398254, 'train/grad_norm_pre_clip': 12.857867240905762, 'timing/data': 0.001038726419210434, 'timing/model': 1.078006012365222, 'learning_rate/qwen_vl_interface': 7.5667040674344495e-06, 'learning_rate/action_model': 7.566704067434449e-05, 'epoch': 0.94}) +[2026-06-17 21:33:28,176][starVLA.training.train_starvla][INFO] - Step 2432, Loss: {'train/loss': 0.6207306385040283, 'train/grad_norm_pre_clip': 12.030074119567871, 'timing/data': 0.0011647185310721397, 'timing/model': 1.4948402428999543, 'learning_rate/qwen_vl_interface': 7.564815806898669e-06, 'learning_rate/action_model': 7.564815806898668e-05, 'epoch': 0.94}) +[2026-06-17 21:33:32,732][starVLA.training.train_starvla][INFO] - Step 2433, Loss: {'train/loss': 0.5208908915519714, 'train/grad_norm_pre_clip': 11.552942276000977, 'timing/data': 0.0013465993106365204, 'timing/model': 1.0865230420604348, 'learning_rate/qwen_vl_interface': 7.562927066499084e-06, 'learning_rate/action_model': 7.562927066499083e-05, 'epoch': 0.94}) +[2026-06-17 21:33:37,492][starVLA.training.train_starvla][INFO] - Step 2434, Loss: {'train/loss': 0.56208336353302, 'train/grad_norm_pre_clip': 8.564382553100586, 'timing/data': 0.0018687695264816284, 'timing/model': 1.0771927013993263, 'learning_rate/qwen_vl_interface': 7.5610378466272274e-06, 'learning_rate/action_model': 7.561037846627227e-05, 'epoch': 0.94}) +[2026-06-17 21:33:42,102][starVLA.training.train_starvla][INFO] - Step 2435, Loss: {'train/loss': 0.3965151607990265, 'train/grad_norm_pre_clip': 9.14494514465332, 'timing/data': 0.0011129556223750114, 'timing/model': 1.0639501251280308, 'learning_rate/qwen_vl_interface': 7.5591481476747384e-06, 'learning_rate/action_model': 7.559148147674739e-05, 'epoch': 0.94}) +[2026-06-17 21:33:46,800][starVLA.training.train_starvla][INFO] - Step 2436, Loss: {'train/loss': 0.9634405374526978, 'train/grad_norm_pre_clip': 13.363201141357422, 'timing/data': 0.0014296714216470718, 'timing/model': 1.062156275846064, 'learning_rate/qwen_vl_interface': 7.557257970033355e-06, 'learning_rate/action_model': 7.557257970033354e-05, 'epoch': 0.94}) +[2026-06-17 21:33:51,690][starVLA.training.train_starvla][INFO] - Step 2437, Loss: {'train/loss': 0.334570050239563, 'train/grad_norm_pre_clip': 7.51995849609375, 'timing/data': 0.0013642143458127975, 'timing/model': 1.1721134223043919, 'learning_rate/qwen_vl_interface': 7.55536731409491e-06, 'learning_rate/action_model': 7.55536731409491e-05, 'epoch': 0.94}) +[2026-06-17 21:33:56,754][starVLA.training.train_starvla][INFO] - Step 2438, Loss: {'train/loss': 0.538217306137085, 'train/grad_norm_pre_clip': 10.669075012207031, 'timing/data': 0.0012262845411896706, 'timing/model': 1.2973651504144073, 'learning_rate/qwen_vl_interface': 7.553476180251341e-06, 'learning_rate/action_model': 7.55347618025134e-05, 'epoch': 0.94}) +[2026-06-17 21:34:01,991][starVLA.training.train_starvla][INFO] - Step 2439, Loss: {'train/loss': 0.5757595300674438, 'train/grad_norm_pre_clip': 10.803011894226074, 'timing/data': 0.0016663484275341034, 'timing/model': 1.3996687540784478, 'learning_rate/qwen_vl_interface': 7.551584568894681e-06, 'learning_rate/action_model': 7.551584568894681e-05, 'epoch': 0.94}) +[2026-06-17 21:34:11,188][starVLA.training.train_starvla][INFO] - Step 2440, Loss: {'train/loss': 0.4145018756389618, 'train/grad_norm_pre_clip': 8.87413501739502, 'timing/data': 0.0011765286326408386, 'timing/model': 1.616489002481103, 'learning_rate/qwen_vl_interface': 7.5496924804170635e-06, 'learning_rate/action_model': 7.549692480417063e-05, 'epoch': 0.94}) +[2026-06-17 21:34:18,661][starVLA.training.train_starvla][INFO] - Step 2441, Loss: {'train/loss': 0.5197795033454895, 'train/grad_norm_pre_clip': 10.010141372680664, 'timing/data': 0.0011161956936120987, 'timing/model': 2.550932912155986, 'learning_rate/qwen_vl_interface': 7.5477999152107195e-06, 'learning_rate/action_model': 7.54779991521072e-05, 'epoch': 0.94}) +[2026-06-17 21:34:23,741][starVLA.training.train_starvla][INFO] - Step 2442, Loss: {'train/loss': 0.48102036118507385, 'train/grad_norm_pre_clip': 10.211454391479492, 'timing/data': 0.001948668621480465, 'timing/model': 1.7475755652412772, 'learning_rate/qwen_vl_interface': 7.545906873667982e-06, 'learning_rate/action_model': 7.545906873667981e-05, 'epoch': 0.94}) +[2026-06-17 21:34:32,863][starVLA.training.train_starvla][INFO] - Step 2443, Loss: {'train/loss': 0.3899175822734833, 'train/grad_norm_pre_clip': 8.165121078491211, 'timing/data': 0.001484706997871399, 'timing/model': 1.1424328070133924, 'learning_rate/qwen_vl_interface': 7.544013356181277e-06, 'learning_rate/action_model': 7.544013356181277e-05, 'epoch': 0.94}) +[2026-06-17 21:34:44,219][starVLA.training.train_starvla][INFO] - Step 2444, Loss: {'train/loss': 0.7114120721817017, 'train/grad_norm_pre_clip': 15.685234069824219, 'timing/data': 0.0017469888553023338, 'timing/model': 1.173558907583356, 'learning_rate/qwen_vl_interface': 7.542119363143135e-06, 'learning_rate/action_model': 7.542119363143136e-05, 'epoch': 0.94}) +[2026-06-17 21:34:56,995][starVLA.training.train_starvla][INFO] - Step 2445, Loss: {'train/loss': 0.5577065348625183, 'train/grad_norm_pre_clip': 8.662240028381348, 'timing/data': 0.0021765809506177902, 'timing/model': 3.9410424008965492, 'learning_rate/qwen_vl_interface': 7.540224894946184e-06, 'learning_rate/action_model': 7.540224894946184e-05, 'epoch': 0.94}) +[2026-06-17 21:35:05,517][starVLA.training.train_starvla][INFO] - Step 2446, Loss: {'train/loss': 0.5525936484336853, 'train/grad_norm_pre_clip': 8.771025657653809, 'timing/data': 0.0018842406570911407, 'timing/model': 1.3030027952045202, 'learning_rate/qwen_vl_interface': 7.538329951983148e-06, 'learning_rate/action_model': 7.538329951983147e-05, 'epoch': 0.94}) +[2026-06-17 21:35:14,967][starVLA.training.train_starvla][INFO] - Step 2447, Loss: {'train/loss': 0.2792944014072418, 'train/grad_norm_pre_clip': 5.740303039550781, 'timing/data': 0.0018376549705863, 'timing/model': 4.280847926624119, 'learning_rate/qwen_vl_interface': 7.536434534646851e-06, 'learning_rate/action_model': 7.53643453464685e-05, 'epoch': 0.94}) +[2026-06-17 21:35:20,491][starVLA.training.train_starvla][INFO] - Step 2448, Loss: {'train/loss': 0.9275201559066772, 'train/grad_norm_pre_clip': 12.906702041625977, 'timing/data': 0.0014148689806461334, 'timing/model': 1.5006505530327559, 'learning_rate/qwen_vl_interface': 7.534538643330213e-06, 'learning_rate/action_model': 7.534538643330214e-05, 'epoch': 0.94}) +[2026-06-17 21:35:25,699][starVLA.training.train_starvla][INFO] - Step 2449, Loss: {'train/loss': 0.25675901770591736, 'train/grad_norm_pre_clip': 4.9243950843811035, 'timing/data': 0.001333148218691349, 'timing/model': 1.4311102982610464, 'learning_rate/qwen_vl_interface': 7.5326422784262566e-06, 'learning_rate/action_model': 7.532642278426256e-05, 'epoch': 0.94}) +[2026-06-17 21:35:31,062][starVLA.training.train_starvla][INFO] - Step 2450, Loss: {'train/loss': 0.4874531924724579, 'train/grad_norm_pre_clip': 9.321351051330566, 'timing/data': 0.0012057311832904816, 'timing/model': 1.2659288896247745, 'learning_rate/qwen_vl_interface': 7.5307454403281e-06, 'learning_rate/action_model': 7.5307454403281e-05, 'epoch': 0.94}) +[2026-06-17 21:35:35,705][starVLA.training.train_starvla][INFO] - Step 2451, Loss: {'train/loss': 0.8458274602890015, 'train/grad_norm_pre_clip': 13.7579345703125, 'timing/data': 0.000845971517264843, 'timing/model': 1.07818956207484, 'learning_rate/qwen_vl_interface': 7.528848129428958e-06, 'learning_rate/action_model': 7.528848129428958e-05, 'epoch': 0.94}) +[2026-06-17 21:35:40,884][starVLA.training.train_starvla][INFO] - Step 2452, Loss: {'train/loss': 0.47708436846733093, 'train/grad_norm_pre_clip': 5.863950252532959, 'timing/data': 0.0012017842382192612, 'timing/model': 1.4794742306694388, 'learning_rate/qwen_vl_interface': 7.526950346122149e-06, 'learning_rate/action_model': 7.526950346122148e-05, 'epoch': 0.94}) +[2026-06-17 21:35:45,732][starVLA.training.train_starvla][INFO] - Step 2453, Loss: {'train/loss': 0.6514931321144104, 'train/grad_norm_pre_clip': 10.312533378601074, 'timing/data': 0.0014294730499386787, 'timing/model': 1.2507751109078526, 'learning_rate/qwen_vl_interface': 7.525052090801083e-06, 'learning_rate/action_model': 7.525052090801083e-05, 'epoch': 0.94}) +[2026-06-17 21:35:50,473][starVLA.training.train_starvla][INFO] - Step 2454, Loss: {'train/loss': 0.44672632217407227, 'train/grad_norm_pre_clip': 8.108258247375488, 'timing/data': 0.0017082393169403076, 'timing/model': 1.1163751306012273, 'learning_rate/qwen_vl_interface': 7.523153363859269e-06, 'learning_rate/action_model': 7.523153363859269e-05, 'epoch': 0.94}) +[2026-06-17 21:35:55,319][starVLA.training.train_starvla][INFO] - Step 2455, Loss: {'train/loss': 0.527855634689331, 'train/grad_norm_pre_clip': 9.589234352111816, 'timing/data': 0.001223921775817871, 'timing/model': 1.2729944186285138, 'learning_rate/qwen_vl_interface': 7.521254165690317e-06, 'learning_rate/action_model': 7.521254165690317e-05, 'epoch': 0.94}) +[2026-06-17 21:36:00,211][starVLA.training.train_starvla][INFO] - Step 2456, Loss: {'train/loss': 0.45020994544029236, 'train/grad_norm_pre_clip': 9.511070251464844, 'timing/data': 0.001600249670445919, 'timing/model': 1.1705528944730759, 'learning_rate/qwen_vl_interface': 7.519354496687933e-06, 'learning_rate/action_model': 7.519354496687933e-05, 'epoch': 0.94}) +[2026-06-17 21:36:05,880][starVLA.training.train_starvla][INFO] - Step 2457, Loss: {'train/loss': 0.6358985304832458, 'train/grad_norm_pre_clip': 11.847908973693848, 'timing/data': 0.0012571178376674652, 'timing/model': 2.499359809793532, 'learning_rate/qwen_vl_interface': 7.517454357245918e-06, 'learning_rate/action_model': 7.517454357245917e-05, 'epoch': 0.95}) +[2026-06-17 21:36:11,028][starVLA.training.train_starvla][INFO] - Step 2458, Loss: {'train/loss': 0.7590280771255493, 'train/grad_norm_pre_clip': 10.281861305236816, 'timing/data': 0.0022506816312670708, 'timing/model': 1.0861635096371174, 'learning_rate/qwen_vl_interface': 7.515553747758176e-06, 'learning_rate/action_model': 7.515553747758176e-05, 'epoch': 0.95}) +[2026-06-17 21:36:16,062][starVLA.training.train_starvla][INFO] - Step 2459, Loss: {'train/loss': 0.3920227885246277, 'train/grad_norm_pre_clip': 8.876029014587402, 'timing/data': 0.0014151204377412796, 'timing/model': 1.1465907450765371, 'learning_rate/qwen_vl_interface': 7.513652668618705e-06, 'learning_rate/action_model': 7.513652668618705e-05, 'epoch': 0.95}) +[2026-06-17 21:36:20,743][starVLA.training.train_starvla][INFO] - Step 2460, Loss: {'train/loss': 0.5986502170562744, 'train/grad_norm_pre_clip': 9.390674591064453, 'timing/data': 0.0011614691466093063, 'timing/model': 1.103466940112412, 'learning_rate/qwen_vl_interface': 7.5117511202215994e-06, 'learning_rate/action_model': 7.5117511202216e-05, 'epoch': 0.95}) +[2026-06-17 21:36:25,536][starVLA.training.train_starvla][INFO] - Step 2461, Loss: {'train/loss': 0.37666991353034973, 'train/grad_norm_pre_clip': 7.1370768547058105, 'timing/data': 0.002403414808213711, 'timing/model': 1.1226192442700267, 'learning_rate/qwen_vl_interface': 7.509849102961053e-06, 'learning_rate/action_model': 7.509849102961053e-05, 'epoch': 0.95}) +[2026-06-17 21:36:30,266][starVLA.training.train_starvla][INFO] - Step 2462, Loss: {'train/loss': 0.6306059956550598, 'train/grad_norm_pre_clip': 7.918281555175781, 'timing/data': 0.0013568764552474022, 'timing/model': 1.424886324442923, 'learning_rate/qwen_vl_interface': 7.507946617231357e-06, 'learning_rate/action_model': 7.507946617231356e-05, 'epoch': 0.95}) +[2026-06-17 21:36:34,745][starVLA.training.train_starvla][INFO] - Step 2463, Loss: {'train/loss': 0.3413489758968353, 'train/grad_norm_pre_clip': 7.339207172393799, 'timing/data': 0.0011526308953762054, 'timing/model': 1.095395976677537, 'learning_rate/qwen_vl_interface': 7.5060436634268965e-06, 'learning_rate/action_model': 7.506043663426896e-05, 'epoch': 0.95}) +[2026-06-17 21:36:41,256][starVLA.training.train_starvla][INFO] - Step 2464, Loss: {'train/loss': 0.43413978815078735, 'train/grad_norm_pre_clip': 8.396909713745117, 'timing/data': 0.0009539909660816193, 'timing/model': 1.374639680609107, 'learning_rate/qwen_vl_interface': 7.504140241942157e-06, 'learning_rate/action_model': 7.504140241942157e-05, 'epoch': 0.95}) +[2026-06-17 21:36:49,587][starVLA.training.train_starvla][INFO] - Step 2465, Loss: {'train/loss': 0.5552070140838623, 'train/grad_norm_pre_clip': 8.242161750793457, 'timing/data': 0.0011904966086149216, 'timing/model': 1.0620863046497107, 'learning_rate/qwen_vl_interface': 7.502236353171718e-06, 'learning_rate/action_model': 7.502236353171718e-05, 'epoch': 0.95}) +[2026-06-17 21:36:54,506][starVLA.training.train_starvla][INFO] - Step 2466, Loss: {'train/loss': 0.5218989253044128, 'train/grad_norm_pre_clip': 7.913517475128174, 'timing/data': 0.001250135712325573, 'timing/model': 1.1955055268481374, 'learning_rate/qwen_vl_interface': 7.500331997510262e-06, 'learning_rate/action_model': 7.500331997510262e-05, 'epoch': 0.95}) +[2026-06-17 21:36:58,828][starVLA.training.train_starvla][INFO] - Step 2467, Loss: {'train/loss': 0.4447058141231537, 'train/grad_norm_pre_clip': 7.376157283782959, 'timing/data': 0.0011744890362024307, 'timing/model': 1.0875839879736304, 'learning_rate/qwen_vl_interface': 7.498427175352561e-06, 'learning_rate/action_model': 7.49842717535256e-05, 'epoch': 0.95}) +[2026-06-17 21:37:03,650][starVLA.training.train_starvla][INFO] - Step 2468, Loss: {'train/loss': 0.6908324360847473, 'train/grad_norm_pre_clip': 24.29439353942871, 'timing/data': 0.0012127365916967392, 'timing/model': 1.1012191511690617, 'learning_rate/qwen_vl_interface': 7.496521887093487e-06, 'learning_rate/action_model': 7.496521887093486e-05, 'epoch': 0.95}) +[2026-06-17 21:37:15,376][starVLA.training.train_starvla][INFO] - Step 2469, Loss: {'train/loss': 0.47544538974761963, 'train/grad_norm_pre_clip': 10.512835502624512, 'timing/data': 0.0013281917199492455, 'timing/model': 1.3724327320232987, 'learning_rate/qwen_vl_interface': 7.494616133128008e-06, 'learning_rate/action_model': 7.494616133128008e-05, 'epoch': 0.95}) +[2026-06-17 21:37:20,300][starVLA.training.train_starvla][INFO] - Step 2470, Loss: {'train/loss': 0.4614167809486389, 'train/grad_norm_pre_clip': 9.76039981842041, 'timing/data': 0.0014244569465517998, 'timing/model': 1.1823267135769129, 'learning_rate/qwen_vl_interface': 7.4927099138511895e-06, 'learning_rate/action_model': 7.492709913851189e-05, 'epoch': 0.95}) +[2026-06-17 21:37:25,156][starVLA.training.train_starvla][INFO] - Step 2471, Loss: {'train/loss': 0.6318881511688232, 'train/grad_norm_pre_clip': 11.107074737548828, 'timing/data': 0.0009489478543400764, 'timing/model': 1.1963617596775293, 'learning_rate/qwen_vl_interface': 7.490803229658192e-06, 'learning_rate/action_model': 7.490803229658192e-05, 'epoch': 0.95}) +[2026-06-17 21:37:30,046][starVLA.training.train_starvla][INFO] - Step 2472, Loss: {'train/loss': 0.4105146527290344, 'train/grad_norm_pre_clip': 6.891342639923096, 'timing/data': 0.0013086963444948196, 'timing/model': 1.5411805296316743, 'learning_rate/qwen_vl_interface': 7.488896080944274e-06, 'learning_rate/action_model': 7.488896080944273e-05, 'epoch': 0.95}) +[2026-06-17 21:37:38,609][starVLA.training.train_starvla][INFO] - Step 2473, Loss: {'train/loss': 0.47611990571022034, 'train/grad_norm_pre_clip': 8.856276512145996, 'timing/data': 0.0011743465438485146, 'timing/model': 1.5064260130748153, 'learning_rate/qwen_vl_interface': 7.4869884681047896e-06, 'learning_rate/action_model': 7.486988468104789e-05, 'epoch': 0.95}) +[2026-06-17 21:37:43,548][starVLA.training.train_starvla][INFO] - Step 2474, Loss: {'train/loss': 0.36652565002441406, 'train/grad_norm_pre_clip': 8.06240463256836, 'timing/data': 0.0010797139257192612, 'timing/model': 1.1294738100841641, 'learning_rate/qwen_vl_interface': 7.485080391535189e-06, 'learning_rate/action_model': 7.485080391535188e-05, 'epoch': 0.95}) +[2026-06-17 21:37:49,517][starVLA.training.train_starvla][INFO] - Step 2475, Loss: {'train/loss': 0.5867341160774231, 'train/grad_norm_pre_clip': 8.61451244354248, 'timing/data': 0.0010670898482203484, 'timing/model': 1.41660744138062, 'learning_rate/qwen_vl_interface': 7.483171851631018e-06, 'learning_rate/action_model': 7.483171851631017e-05, 'epoch': 0.95}) +[2026-06-17 21:37:57,202][starVLA.training.train_starvla][INFO] - Step 2476, Loss: {'train/loss': 0.25334465503692627, 'train/grad_norm_pre_clip': 5.5006022453308105, 'timing/data': 0.0011457912623882294, 'timing/model': 1.1155792707577348, 'learning_rate/qwen_vl_interface': 7.481262848787919e-06, 'learning_rate/action_model': 7.481262848787919e-05, 'epoch': 0.95}) +[2026-06-17 21:38:02,905][starVLA.training.train_starvla][INFO] - Step 2477, Loss: {'train/loss': 0.44762325286865234, 'train/grad_norm_pre_clip': 12.48564338684082, 'timing/data': 0.0011272197589278221, 'timing/model': 1.1796886455267668, 'learning_rate/qwen_vl_interface': 7.479353383401631e-06, 'learning_rate/action_model': 7.47935338340163e-05, 'epoch': 0.95}) +[2026-06-17 21:38:08,024][starVLA.training.train_starvla][INFO] - Step 2478, Loss: {'train/loss': 0.6388171315193176, 'train/grad_norm_pre_clip': 10.899539947509766, 'timing/data': 0.0013472391292452812, 'timing/model': 1.1968145305290818, 'learning_rate/qwen_vl_interface': 7.477443455867987e-06, 'learning_rate/action_model': 7.477443455867988e-05, 'epoch': 0.95}) +[2026-06-17 21:38:17,475][starVLA.training.train_starvla][INFO] - Step 2479, Loss: {'train/loss': 0.2934243977069855, 'train/grad_norm_pre_clip': 6.673226356506348, 'timing/data': 0.0014143222942948341, 'timing/model': 1.3785667577758431, 'learning_rate/qwen_vl_interface': 7.475533066582917e-06, 'learning_rate/action_model': 7.475533066582918e-05, 'epoch': 0.95}) +[2026-06-17 21:38:22,420][starVLA.training.train_starvla][INFO] - Step 2480, Loss: {'train/loss': 0.5658715963363647, 'train/grad_norm_pre_clip': 9.158145904541016, 'timing/data': 0.000888272188603878, 'timing/model': 1.0838638413697481, 'learning_rate/qwen_vl_interface': 7.47362221594245e-06, 'learning_rate/action_model': 7.47362221594245e-05, 'epoch': 0.95}) +[2026-06-17 21:38:27,142][starVLA.training.train_starvla][INFO] - Step 2481, Loss: {'train/loss': 0.2769966721534729, 'train/grad_norm_pre_clip': 8.812872886657715, 'timing/data': 0.0012708231806755066, 'timing/model': 1.1274766447022557, 'learning_rate/qwen_vl_interface': 7.4717109043427025e-06, 'learning_rate/action_model': 7.471710904342702e-05, 'epoch': 0.95}) +[2026-06-17 21:38:39,215][starVLA.training.train_starvla][INFO] - Step 2482, Loss: {'train/loss': 0.300202876329422, 'train/grad_norm_pre_clip': 9.311716079711914, 'timing/data': 0.0014436151832342148, 'timing/model': 4.451990423724055, 'learning_rate/qwen_vl_interface': 7.469799132179894e-06, 'learning_rate/action_model': 7.469799132179895e-05, 'epoch': 0.95}) +[2026-06-17 21:38:43,902][starVLA.training.train_starvla][INFO] - Step 2483, Loss: {'train/loss': 0.5479496121406555, 'train/grad_norm_pre_clip': 10.322827339172363, 'timing/data': 0.0011351760476827621, 'timing/model': 1.0728267030790448, 'learning_rate/qwen_vl_interface': 7.467886899850338e-06, 'learning_rate/action_model': 7.467886899850338e-05, 'epoch': 0.96}) +[2026-06-17 21:38:49,779][starVLA.training.train_starvla][INFO] - Step 2484, Loss: {'train/loss': 0.811941921710968, 'train/grad_norm_pre_clip': 12.900729179382324, 'timing/data': 0.001088772900402546, 'timing/model': 2.1081470930948853, 'learning_rate/qwen_vl_interface': 7.465974207750442e-06, 'learning_rate/action_model': 7.465974207750442e-05, 'epoch': 0.96}) +[2026-06-17 21:38:54,762][starVLA.training.train_starvla][INFO] - Step 2485, Loss: {'train/loss': 0.5211580395698547, 'train/grad_norm_pre_clip': 12.729065895080566, 'timing/data': 0.0012058056890964508, 'timing/model': 1.2631579339504242, 'learning_rate/qwen_vl_interface': 7.464061056276707e-06, 'learning_rate/action_model': 7.464061056276707e-05, 'epoch': 0.96}) +[2026-06-17 21:38:59,845][starVLA.training.train_starvla][INFO] - Step 2486, Loss: {'train/loss': 0.4088614583015442, 'train/grad_norm_pre_clip': 8.676077842712402, 'timing/data': 0.0011376040056347847, 'timing/model': 1.2310649575665593, 'learning_rate/qwen_vl_interface': 7.462147445825733e-06, 'learning_rate/action_model': 7.462147445825733e-05, 'epoch': 0.96}) +[2026-06-17 21:39:07,855][starVLA.training.train_starvla][INFO] - Step 2487, Loss: {'train/loss': 0.35970619320869446, 'train/grad_norm_pre_clip': 12.499662399291992, 'timing/data': 0.0008818069472908974, 'timing/model': 1.6262842752039433, 'learning_rate/qwen_vl_interface': 7.460233376794211e-06, 'learning_rate/action_model': 7.46023337679421e-05, 'epoch': 0.96}) +[2026-06-17 21:39:12,638][starVLA.training.train_starvla][INFO] - Step 2488, Loss: {'train/loss': 0.7197471857070923, 'train/grad_norm_pre_clip': 11.636125564575195, 'timing/data': 0.0011886842548847198, 'timing/model': 1.2620740896090865, 'learning_rate/qwen_vl_interface': 7.458318849578934e-06, 'learning_rate/action_model': 7.458318849578933e-05, 'epoch': 0.96}) +[2026-06-17 21:39:19,361][starVLA.training.train_starvla][INFO] - Step 2489, Loss: {'train/loss': 0.6069860458374023, 'train/grad_norm_pre_clip': 22.306211471557617, 'timing/data': 0.0012449240311980247, 'timing/model': 1.1964316852390766, 'learning_rate/qwen_vl_interface': 7.456403864576782e-06, 'learning_rate/action_model': 7.456403864576782e-05, 'epoch': 0.96}) +[2026-06-17 21:39:28,884][starVLA.training.train_starvla][INFO] - Step 2490, Loss: {'train/loss': 0.3407154679298401, 'train/grad_norm_pre_clip': 7.555917263031006, 'timing/data': 0.0015373043715953827, 'timing/model': 2.969242942519486, 'learning_rate/qwen_vl_interface': 7.4544884221847354e-06, 'learning_rate/action_model': 7.454488422184735e-05, 'epoch': 0.96}) +[2026-06-17 21:39:33,522][starVLA.training.train_starvla][INFO] - Step 2491, Loss: {'train/loss': 0.40789005160331726, 'train/grad_norm_pre_clip': 9.103955268859863, 'timing/data': 0.001416631042957306, 'timing/model': 1.0993088753893971, 'learning_rate/qwen_vl_interface': 7.452572522799867e-06, 'learning_rate/action_model': 7.452572522799866e-05, 'epoch': 0.96}) +[2026-06-17 21:39:39,251][starVLA.training.train_starvla][INFO] - Step 2492, Loss: {'train/loss': 0.5050264596939087, 'train/grad_norm_pre_clip': 9.221360206604004, 'timing/data': 0.001154344528913498, 'timing/model': 2.47240346763283, 'learning_rate/qwen_vl_interface': 7.450656166819344e-06, 'learning_rate/action_model': 7.450656166819344e-05, 'epoch': 0.96}) +[2026-06-17 21:39:47,726][starVLA.training.train_starvla][INFO] - Step 2493, Loss: {'train/loss': 0.4334782660007477, 'train/grad_norm_pre_clip': 7.018252372741699, 'timing/data': 0.001519273966550827, 'timing/model': 1.098887731321156, 'learning_rate/qwen_vl_interface': 7.4487393546404295e-06, 'learning_rate/action_model': 7.44873935464043e-05, 'epoch': 0.96}) +[2026-06-17 21:39:52,459][starVLA.training.train_starvla][INFO] - Step 2494, Loss: {'train/loss': 0.5077570676803589, 'train/grad_norm_pre_clip': 10.858186721801758, 'timing/data': 0.0009448127821087837, 'timing/model': 1.1138456240296364, 'learning_rate/qwen_vl_interface': 7.446822086660481e-06, 'learning_rate/action_model': 7.446822086660481e-05, 'epoch': 0.96}) +[2026-06-17 21:39:59,409][starVLA.training.train_starvla][INFO] - Step 2495, Loss: {'train/loss': 0.6270408034324646, 'train/grad_norm_pre_clip': 9.959336280822754, 'timing/data': 0.001147543080151081, 'timing/model': 2.9422716265544295, 'learning_rate/qwen_vl_interface': 7.444904363276951e-06, 'learning_rate/action_model': 7.444904363276951e-05, 'epoch': 0.96}) +[2026-06-17 21:40:04,311][starVLA.training.train_starvla][INFO] - Step 2496, Loss: {'train/loss': 0.4951900243759155, 'train/grad_norm_pre_clip': 14.292864799499512, 'timing/data': 0.0012558717280626297, 'timing/model': 1.1512689692899585, 'learning_rate/qwen_vl_interface': 7.442986184887383e-06, 'learning_rate/action_model': 7.442986184887383e-05, 'epoch': 0.96}) +[2026-06-17 21:40:09,185][starVLA.training.train_starvla][INFO] - Step 2497, Loss: {'train/loss': 0.3648263216018677, 'train/grad_norm_pre_clip': 10.034650802612305, 'timing/data': 0.0010846620425581932, 'timing/model': 1.5871051102876663, 'learning_rate/qwen_vl_interface': 7.4410675518894205e-06, 'learning_rate/action_model': 7.44106755188942e-05, 'epoch': 0.96}) +[2026-06-17 21:40:13,899][starVLA.training.train_starvla][INFO] - Step 2498, Loss: {'train/loss': 0.508094847202301, 'train/grad_norm_pre_clip': 10.488394737243652, 'timing/data': 0.0010767169296741486, 'timing/model': 1.155872406437993, 'learning_rate/qwen_vl_interface': 7.439148464680795e-06, 'learning_rate/action_model': 7.439148464680796e-05, 'epoch': 0.96}) +[2026-06-17 21:40:18,840][starVLA.training.train_starvla][INFO] - Step 2499, Loss: {'train/loss': 0.6032065749168396, 'train/grad_norm_pre_clip': 11.136570930480957, 'timing/data': 0.0009774351492524147, 'timing/model': 1.1384072415530682, 'learning_rate/qwen_vl_interface': 7.4372289236593395e-06, 'learning_rate/action_model': 7.43722892365934e-05, 'epoch': 0.96}) +[2026-06-17 21:43:57,303][starVLA.training.train_starvla][INFO] - Step 2500, Loss: {'train/loss': 0.6545981168746948, 'train/grad_norm_pre_clip': 10.769131660461426, 'eval/loss': 0.5396966338157654, 'eval/latency_4/loss': 0.5396966338157654, 'eval/demon_fire_precision': 0.9705298013245033, 'eval/demon_fire_recall': 0.9519324455992205, 'eval/demon_fire_f1': 0.9611411706837186, 'eval/demon_fire_rate_model': 0.8869309838472834, 'eval/demon_fire_rate_teacher': 0.9042584434654919, 'eval/demon_left_precision': 0.7017543859649122, 'eval/demon_left_recall': 0.639269406392694, 'eval/demon_left_f1': 0.6690561529271206, 'eval/demon_left_rate_model': 0.1171806167400881, 'eval/demon_left_rate_teacher': 0.12863436123348018, 'eval/demon_right_precision': 0.764065335753176, 'eval/demon_right_recall': 0.759927797833935, 'eval/demon_right_f1': 0.7619909502262444, 'eval/demon_right_rate_model': 0.1618208516886931, 'eval/demon_right_rate_teacher': 0.16270190895741557, 'eval/demon_move_f1': 0.7155235515766825, 'eval/demon_cc_f1': 0.8714032477883941, 'eval/latency_4/demon_fire_precision': 0.9705298013245033, 'eval/latency_4/demon_fire_recall': 0.9519324455992205, 'eval/latency_4/demon_fire_f1': 0.9611411706837186, 'eval/latency_4/demon_fire_rate_model': 0.8869309838472834, 'eval/latency_4/demon_fire_rate_teacher': 0.9042584434654919, 'eval/latency_4/demon_left_precision': 0.7017543859649122, 'eval/latency_4/demon_left_recall': 0.639269406392694, 'eval/latency_4/demon_left_f1': 0.6690561529271206, 'eval/latency_4/demon_left_rate_model': 0.1171806167400881, 'eval/latency_4/demon_left_rate_teacher': 0.12863436123348018, 'eval/latency_4/demon_right_precision': 0.764065335753176, 'eval/latency_4/demon_right_recall': 0.759927797833935, 'eval/latency_4/demon_right_f1': 0.7619909502262444, 'eval/latency_4/demon_right_rate_model': 0.1618208516886931, 'eval/latency_4/demon_right_rate_teacher': 0.16270190895741557, 'eval/latency_4/demon_move_f1': 0.7155235515766825, 'eval/latency_4/demon_cc_f1': 0.8714032477883941, 'timing/data': 0.0011714613065123558, 'timing/model': 1.100379434414208, 'learning_rate/qwen_vl_interface': 7.435308929222973e-06, 'learning_rate/action_model': 7.435308929222972e-05, 'epoch': 0.96}) +[2026-06-17 21:43:57,321][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_2500_state +[2026-06-17 21:43:57,329][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 21:44:06,107][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_2500_state/model.safetensors +[2026-06-17 21:44:28,537][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_2500_state/optimizer.bin +[2026-06-17 21:44:28,554][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_2500_state/sampler.bin +[2026-06-17 21:44:28,555][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_2500_state/sampler_1.bin +[2026-06-17 21:44:28,557][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_2500_state/random_states_0.pkl +[2026-06-17 21:44:28,562][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2500 +[2026-06-17 21:44:28,562][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 21:44:28,580][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 21:44:33,666][starVLA.training.train_starvla][INFO] - Step 2501, Loss: {'train/loss': 0.5079816579818726, 'train/grad_norm_pre_clip': 14.305188179016113, 'timing/data': 0.0011914260685443878, 'timing/model': 1.1642548711970448, 'learning_rate/qwen_vl_interface': 7.433388481769716e-06, 'learning_rate/action_model': 7.433388481769715e-05, 'epoch': 0.96}) +[2026-06-17 21:44:39,523][starVLA.training.train_starvla][INFO] - Step 2502, Loss: {'train/loss': 0.5829187035560608, 'train/grad_norm_pre_clip': 11.492575645446777, 'timing/data': 0.0013350537046790123, 'timing/model': 1.5034333169460297, 'learning_rate/qwen_vl_interface': 7.431467581697675e-06, 'learning_rate/action_model': 7.431467581697674e-05, 'epoch': 0.96}) +[2026-06-17 21:44:51,201][starVLA.training.train_starvla][INFO] - Step 2503, Loss: {'train/loss': 0.46307143568992615, 'train/grad_norm_pre_clip': 9.903450965881348, 'timing/data': 0.00324950460344553, 'timing/model': 1.1759071610867977, 'learning_rate/qwen_vl_interface': 7.429546229405057e-06, 'learning_rate/action_model': 7.429546229405057e-05, 'epoch': 0.96}) +[2026-06-17 21:44:57,152][starVLA.training.train_starvla][INFO] - Step 2504, Loss: {'train/loss': 0.6978685855865479, 'train/grad_norm_pre_clip': 13.049370765686035, 'timing/data': 0.0014985334128141403, 'timing/model': 1.5574599420651793, 'learning_rate/qwen_vl_interface': 7.427624425290162e-06, 'learning_rate/action_model': 7.427624425290161e-05, 'epoch': 0.96}) +[2026-06-17 21:45:05,566][starVLA.training.train_starvla][INFO] - Step 2505, Loss: {'train/loss': 0.6753007769584656, 'train/grad_norm_pre_clip': 12.425436973571777, 'timing/data': 0.0011591538786888123, 'timing/model': 1.2263417337089777, 'learning_rate/qwen_vl_interface': 7.425702169751379e-06, 'learning_rate/action_model': 7.425702169751378e-05, 'epoch': 0.96}) +[2026-06-17 21:45:11,105][starVLA.training.train_starvla][INFO] - Step 2506, Loss: {'train/loss': 0.6004239320755005, 'train/grad_norm_pre_clip': 9.791949272155762, 'timing/data': 0.0012542977929115295, 'timing/model': 1.08087498228997, 'learning_rate/qwen_vl_interface': 7.4237794631871965e-06, 'learning_rate/action_model': 7.423779463187196e-05, 'epoch': 0.96}) +[2026-06-17 21:45:15,425][starVLA.training.train_starvla][INFO] - Step 2507, Loss: {'train/loss': 0.4917628765106201, 'train/grad_norm_pre_clip': 7.9876604080200195, 'timing/data': 0.000944674015045166, 'timing/model': 1.075753914192319, 'learning_rate/qwen_vl_interface': 7.42185630599619e-06, 'learning_rate/action_model': 7.42185630599619e-05, 'epoch': 0.96}) +[2026-06-17 21:45:20,227][starVLA.training.train_starvla][INFO] - Step 2508, Loss: {'train/loss': 0.6258635520935059, 'train/grad_norm_pre_clip': 11.371042251586914, 'timing/data': 0.0013432912528514862, 'timing/model': 1.1494182161986828, 'learning_rate/qwen_vl_interface': 7.419932698577036e-06, 'learning_rate/action_model': 7.419932698577035e-05, 'epoch': 0.96}) +[2026-06-17 21:45:25,196][starVLA.training.train_starvla][INFO] - Step 2509, Loss: {'train/loss': 0.447426438331604, 'train/grad_norm_pre_clip': 7.805009365081787, 'timing/data': 0.0011554993689060211, 'timing/model': 1.1433732248842716, 'learning_rate/qwen_vl_interface': 7.418008641328498e-06, 'learning_rate/action_model': 7.418008641328497e-05, 'epoch': 0.97}) +[2026-06-17 21:45:35,981][starVLA.training.train_starvla][INFO] - Step 2510, Loss: {'train/loss': 0.5080734491348267, 'train/grad_norm_pre_clip': 7.52043342590332, 'timing/data': 0.0014392118901014328, 'timing/model': 1.349730497226119, 'learning_rate/qwen_vl_interface': 7.416084134649435e-06, 'learning_rate/action_model': 7.416084134649435e-05, 'epoch': 0.97}) +[2026-06-17 21:45:45,118][starVLA.training.train_starvla][INFO] - Step 2511, Loss: {'train/loss': 0.4973217844963074, 'train/grad_norm_pre_clip': 10.550196647644043, 'timing/data': 0.003911987878382206, 'timing/model': 4.179164775647223, 'learning_rate/qwen_vl_interface': 7.414159178938801e-06, 'learning_rate/action_model': 7.414159178938801e-05, 'epoch': 0.97}) +[2026-06-17 21:45:51,817][starVLA.training.train_starvla][INFO] - Step 2512, Loss: {'train/loss': 0.45317381620407104, 'train/grad_norm_pre_clip': 10.953015327453613, 'timing/data': 0.0011289305984973907, 'timing/model': 1.568173998966813, 'learning_rate/qwen_vl_interface': 7.412233774595639e-06, 'learning_rate/action_model': 7.41223377459564e-05, 'epoch': 0.97}) +[2026-06-17 21:46:00,551][starVLA.training.train_starvla][INFO] - Step 2513, Loss: {'train/loss': 0.6520957350730896, 'train/grad_norm_pre_clip': 10.55490779876709, 'timing/data': 0.0010981177911162376, 'timing/model': 2.211199887096882, 'learning_rate/qwen_vl_interface': 7.410307922019091e-06, 'learning_rate/action_model': 7.41030792201909e-05, 'epoch': 0.97}) +[2026-06-17 21:46:08,743][starVLA.training.train_starvla][INFO] - Step 2514, Loss: {'train/loss': 0.45302703976631165, 'train/grad_norm_pre_clip': 7.005169868469238, 'timing/data': 0.0017113881185650826, 'timing/model': 1.3657762855291367, 'learning_rate/qwen_vl_interface': 7.408381621608387e-06, 'learning_rate/action_model': 7.408381621608387e-05, 'epoch': 0.97}) +[2026-06-17 21:46:15,476][starVLA.training.train_starvla][INFO] - Step 2515, Loss: {'train/loss': 0.5508754253387451, 'train/grad_norm_pre_clip': 8.473003387451172, 'timing/data': 0.0008402802050113678, 'timing/model': 1.1588656147941947, 'learning_rate/qwen_vl_interface': 7.406454873762848e-06, 'learning_rate/action_model': 7.406454873762848e-05, 'epoch': 0.97}) +[2026-06-17 21:46:20,577][starVLA.training.train_starvla][INFO] - Step 2516, Loss: {'train/loss': 0.199677512049675, 'train/grad_norm_pre_clip': 6.851348876953125, 'timing/data': 0.0009243432432413101, 'timing/model': 1.1166113950312138, 'learning_rate/qwen_vl_interface': 7.404527678881894e-06, 'learning_rate/action_model': 7.404527678881894e-05, 'epoch': 0.97}) +[2026-06-17 21:46:25,805][starVLA.training.train_starvla][INFO] - Step 2517, Loss: {'train/loss': 0.4669483006000519, 'train/grad_norm_pre_clip': 9.737997055053711, 'timing/data': 0.0010831616818904877, 'timing/model': 1.627216256223619, 'learning_rate/qwen_vl_interface': 7.402600037365034e-06, 'learning_rate/action_model': 7.402600037365034e-05, 'epoch': 0.97}) +[2026-06-17 21:46:30,212][starVLA.training.train_starvla][INFO] - Step 2518, Loss: {'train/loss': 0.5156353116035461, 'train/grad_norm_pre_clip': 10.156152725219727, 'timing/data': 0.0017598606646060944, 'timing/model': 1.1346341464668512, 'learning_rate/qwen_vl_interface': 7.400671949611869e-06, 'learning_rate/action_model': 7.40067194961187e-05, 'epoch': 0.97}) +[2026-06-17 21:46:35,139][starVLA.training.train_starvla][INFO] - Step 2519, Loss: {'train/loss': 0.35157695412635803, 'train/grad_norm_pre_clip': 7.449916362762451, 'timing/data': 0.0011463901028037071, 'timing/model': 1.2563855228945613, 'learning_rate/qwen_vl_interface': 7.398743416022095e-06, 'learning_rate/action_model': 7.398743416022095e-05, 'epoch': 0.97}) +[2026-06-17 21:46:40,078][starVLA.training.train_starvla][INFO] - Step 2520, Loss: {'train/loss': 0.5469762682914734, 'train/grad_norm_pre_clip': 10.516443252563477, 'timing/data': 0.0016202675178647041, 'timing/model': 1.1603631414473057, 'learning_rate/qwen_vl_interface': 7.3968144369954995e-06, 'learning_rate/action_model': 7.396814436995499e-05, 'epoch': 0.97}) +[2026-06-17 21:46:45,074][starVLA.training.train_starvla][INFO] - Step 2521, Loss: {'train/loss': 0.6254515647888184, 'train/grad_norm_pre_clip': 7.913120746612549, 'timing/data': 0.0007622111588716507, 'timing/model': 1.2979701207950711, 'learning_rate/qwen_vl_interface': 7.39488501293196e-06, 'learning_rate/action_model': 7.39488501293196e-05, 'epoch': 0.97}) +[2026-06-17 21:46:49,757][starVLA.training.train_starvla][INFO] - Step 2522, Loss: {'train/loss': 0.38239890336990356, 'train/grad_norm_pre_clip': 6.35687780380249, 'timing/data': 0.0011082608252763748, 'timing/model': 1.4197961147874594, 'learning_rate/qwen_vl_interface': 7.392955144231452e-06, 'learning_rate/action_model': 7.392955144231452e-05, 'epoch': 0.97}) +[2026-06-17 21:46:54,088][starVLA.training.train_starvla][INFO] - Step 2523, Loss: {'train/loss': 0.2632219195365906, 'train/grad_norm_pre_clip': 9.843953132629395, 'timing/data': 0.001415361650288105, 'timing/model': 1.132233327254653, 'learning_rate/qwen_vl_interface': 7.391024831294034e-06, 'learning_rate/action_model': 7.391024831294033e-05, 'epoch': 0.97}) +[2026-06-17 21:46:59,462][starVLA.training.train_starvla][INFO] - Step 2524, Loss: {'train/loss': 0.3720405101776123, 'train/grad_norm_pre_clip': 6.38340425491333, 'timing/data': 0.001460384577512741, 'timing/model': 1.1559815974906087, 'learning_rate/qwen_vl_interface': 7.389094074519866e-06, 'learning_rate/action_model': 7.389094074519866e-05, 'epoch': 0.97}) +[2026-06-17 21:47:04,473][starVLA.training.train_starvla][INFO] - Step 2525, Loss: {'train/loss': 0.49182453751564026, 'train/grad_norm_pre_clip': 7.737542629241943, 'timing/data': 0.0010385410860180855, 'timing/model': 1.1527630984783173, 'learning_rate/qwen_vl_interface': 7.387162874309194e-06, 'learning_rate/action_model': 7.387162874309193e-05, 'epoch': 0.97}) +[2026-06-17 21:47:09,852][starVLA.training.train_starvla][INFO] - Step 2526, Loss: {'train/loss': 0.401659220457077, 'train/grad_norm_pre_clip': 7.468902111053467, 'timing/data': 0.0011095721274614334, 'timing/model': 1.668986001983285, 'learning_rate/qwen_vl_interface': 7.385231231062359e-06, 'learning_rate/action_model': 7.385231231062358e-05, 'epoch': 0.97}) +[2026-06-17 21:47:14,595][starVLA.training.train_starvla][INFO] - Step 2527, Loss: {'train/loss': 0.17629607021808624, 'train/grad_norm_pre_clip': 3.3964309692382812, 'timing/data': 0.001221480779349804, 'timing/model': 1.294729471206665, 'learning_rate/qwen_vl_interface': 7.383299145179789e-06, 'learning_rate/action_model': 7.383299145179789e-05, 'epoch': 0.97}) +[2026-06-17 21:47:19,328][starVLA.training.train_starvla][INFO] - Step 2528, Loss: {'train/loss': 0.2383825033903122, 'train/grad_norm_pre_clip': 5.926798343658447, 'timing/data': 0.0010778307914733887, 'timing/model': 1.0764154009521008, 'learning_rate/qwen_vl_interface': 7.381366617062014e-06, 'learning_rate/action_model': 7.381366617062014e-05, 'epoch': 0.97}) +[2026-06-17 21:47:23,928][starVLA.training.train_starvla][INFO] - Step 2529, Loss: {'train/loss': 0.42134401202201843, 'train/grad_norm_pre_clip': 7.309817790985107, 'timing/data': 0.0012015625834465027, 'timing/model': 1.0698407609015703, 'learning_rate/qwen_vl_interface': 7.379433647109645e-06, 'learning_rate/action_model': 7.379433647109645e-05, 'epoch': 0.97}) +[2026-06-17 21:47:28,802][starVLA.training.train_starvla][INFO] - Step 2530, Loss: {'train/loss': 0.32220807671546936, 'train/grad_norm_pre_clip': 5.936426162719727, 'timing/data': 0.0009074443951249123, 'timing/model': 1.1653435500338674, 'learning_rate/qwen_vl_interface': 7.37750023572339e-06, 'learning_rate/action_model': 7.37750023572339e-05, 'epoch': 0.97}) +[2026-06-17 21:47:41,525][starVLA.training.train_starvla][INFO] - Step 2531, Loss: {'train/loss': 0.3241516947746277, 'train/grad_norm_pre_clip': 7.684916019439697, 'timing/data': 0.0009643146768212318, 'timing/model': 2.721432394348085, 'learning_rate/qwen_vl_interface': 7.375566383304047e-06, 'learning_rate/action_model': 7.375566383304047e-05, 'epoch': 0.97}) +[2026-06-17 21:47:52,112][starVLA.training.train_starvla][INFO] - Step 2532, Loss: {'train/loss': 0.31426528096199036, 'train/grad_norm_pre_clip': 8.200056076049805, 'timing/data': 0.0010033119469881058, 'timing/model': 4.285180046223104, 'learning_rate/qwen_vl_interface': 7.373632090252505e-06, 'learning_rate/action_model': 7.373632090252505e-05, 'epoch': 0.97}) +[2026-06-17 21:47:58,301][starVLA.training.train_starvla][INFO] - Step 2533, Loss: {'train/loss': 0.5905876159667969, 'train/grad_norm_pre_clip': 10.44214916229248, 'timing/data': 0.0011982731521129608, 'timing/model': 2.1625539883971214, 'learning_rate/qwen_vl_interface': 7.371697356969746e-06, 'learning_rate/action_model': 7.371697356969746e-05, 'epoch': 0.97}) +[2026-06-17 21:48:03,547][starVLA.training.train_starvla][INFO] - Step 2534, Loss: {'train/loss': 0.5799426436424255, 'train/grad_norm_pre_clip': 9.56119441986084, 'timing/data': 0.0008839312940835953, 'timing/model': 1.3849940365180373, 'learning_rate/qwen_vl_interface': 7.36976218385684e-06, 'learning_rate/action_model': 7.36976218385684e-05, 'epoch': 0.97}) +[2026-06-17 21:48:12,138][starVLA.training.train_starvla][INFO] - Step 2535, Loss: {'train/loss': 0.4502283036708832, 'train/grad_norm_pre_clip': 9.325172424316406, 'timing/data': 0.0014981236308813095, 'timing/model': 1.8243662640452385, 'learning_rate/qwen_vl_interface': 7.367826571314955e-06, 'learning_rate/action_model': 7.367826571314954e-05, 'epoch': 0.98}) +[2026-06-17 21:48:19,087][starVLA.training.train_starvla][INFO] - Step 2536, Loss: {'train/loss': 0.4248593747615814, 'train/grad_norm_pre_clip': 8.086273193359375, 'timing/data': 0.00130364578217268, 'timing/model': 1.7014806428924203, 'learning_rate/qwen_vl_interface': 7.36589051974534e-06, 'learning_rate/action_model': 7.36589051974534e-05, 'epoch': 0.98}) +[2026-06-17 21:48:33,472][starVLA.training.train_starvla][INFO] - Step 2537, Loss: {'train/loss': 0.42131319642066956, 'train/grad_norm_pre_clip': 7.080686569213867, 'timing/data': 0.001001330092549324, 'timing/model': 5.209305623546243, 'learning_rate/qwen_vl_interface': 7.363954029549348e-06, 'learning_rate/action_model': 7.363954029549348e-05, 'epoch': 0.98}) +[2026-06-17 21:48:43,492][starVLA.training.train_starvla][INFO] - Step 2538, Loss: {'train/loss': 0.3411567509174347, 'train/grad_norm_pre_clip': 7.750672340393066, 'timing/data': 0.0010264143347740173, 'timing/model': 1.398250276222825, 'learning_rate/qwen_vl_interface': 7.362017101128406e-06, 'learning_rate/action_model': 7.362017101128406e-05, 'epoch': 0.98}) +[2026-06-17 21:48:48,302][starVLA.training.train_starvla][INFO] - Step 2539, Loss: {'train/loss': 0.5411584973335266, 'train/grad_norm_pre_clip': 7.461716651916504, 'timing/data': 0.0012464579194784164, 'timing/model': 1.1633572429418564, 'learning_rate/qwen_vl_interface': 7.360079734884049e-06, 'learning_rate/action_model': 7.360079734884049e-05, 'epoch': 0.98}) +[2026-06-17 21:48:52,878][starVLA.training.train_starvla][INFO] - Step 2540, Loss: {'train/loss': 0.3333456218242645, 'train/grad_norm_pre_clip': 7.102633953094482, 'timing/data': 0.0014711758121848106, 'timing/model': 1.121848190203309, 'learning_rate/qwen_vl_interface': 7.358141931217893e-06, 'learning_rate/action_model': 7.358141931217892e-05, 'epoch': 0.98}) +[2026-06-17 21:48:57,631][starVLA.training.train_starvla][INFO] - Step 2541, Loss: {'train/loss': 0.4994736313819885, 'train/grad_norm_pre_clip': 8.445341110229492, 'timing/data': 0.00132081750780344, 'timing/model': 1.1991516780108213, 'learning_rate/qwen_vl_interface': 7.356203690531643e-06, 'learning_rate/action_model': 7.356203690531643e-05, 'epoch': 0.98}) +[2026-06-17 21:49:03,263][starVLA.training.train_starvla][INFO] - Step 2542, Loss: {'train/loss': 0.6339550614356995, 'train/grad_norm_pre_clip': 10.365923881530762, 'timing/data': 0.0013165203854441643, 'timing/model': 1.6799175264313817, 'learning_rate/qwen_vl_interface': 7.354265013227104e-06, 'learning_rate/action_model': 7.354265013227104e-05, 'epoch': 0.98}) +[2026-06-17 21:49:08,101][starVLA.training.train_starvla][INFO] - Step 2543, Loss: {'train/loss': 0.8753466606140137, 'train/grad_norm_pre_clip': 11.78201675415039, 'timing/data': 0.0015138257294893265, 'timing/model': 1.1057699639350176, 'learning_rate/qwen_vl_interface': 7.352325899706163e-06, 'learning_rate/action_model': 7.352325899706162e-05, 'epoch': 0.98}) +[2026-06-17 21:49:13,046][starVLA.training.train_starvla][INFO] - Step 2544, Loss: {'train/loss': 0.44158682227134705, 'train/grad_norm_pre_clip': 8.275452613830566, 'timing/data': 0.0014530951157212257, 'timing/model': 1.2303296234458685, 'learning_rate/qwen_vl_interface': 7.350386350370802e-06, 'learning_rate/action_model': 7.350386350370801e-05, 'epoch': 0.98}) +[2026-06-17 21:49:18,052][starVLA.training.train_starvla][INFO] - Step 2545, Loss: {'train/loss': 0.42358556389808655, 'train/grad_norm_pre_clip': 8.097169876098633, 'timing/data': 0.0014188233762979507, 'timing/model': 1.1698523992672563, 'learning_rate/qwen_vl_interface': 7.348446365623089e-06, 'learning_rate/action_model': 7.348446365623088e-05, 'epoch': 0.98}) +[2026-06-17 21:49:27,497][starVLA.training.train_starvla][INFO] - Step 2546, Loss: {'train/loss': 0.4066123366355896, 'train/grad_norm_pre_clip': 11.40420150756836, 'timing/data': 0.0012603895738720894, 'timing/model': 1.1386024011299014, 'learning_rate/qwen_vl_interface': 7.346505945865186e-06, 'learning_rate/action_model': 7.346505945865186e-05, 'epoch': 0.98}) +[2026-06-17 21:49:32,672][starVLA.training.train_starvla][INFO] - Step 2547, Loss: {'train/loss': 0.4812460243701935, 'train/grad_norm_pre_clip': 7.448348522186279, 'timing/data': 0.001389743760228157, 'timing/model': 1.7950576478615403, 'learning_rate/qwen_vl_interface': 7.344565091499345e-06, 'learning_rate/action_model': 7.344565091499344e-05, 'epoch': 0.98}) +[2026-06-17 21:49:52,719][starVLA.training.train_starvla][INFO] - Step 2548, Loss: {'train/loss': 0.6017274856567383, 'train/grad_norm_pre_clip': 8.433708190917969, 'timing/data': 0.001127387396991253, 'timing/model': 5.5800010319799185, 'learning_rate/qwen_vl_interface': 7.342623802927907e-06, 'learning_rate/action_model': 7.342623802927907e-05, 'epoch': 0.98}) +[2026-06-17 21:50:03,155][starVLA.training.train_starvla][INFO] - Step 2549, Loss: {'train/loss': 0.7488616704940796, 'train/grad_norm_pre_clip': 17.45528793334961, 'timing/data': 0.0009116530418395996, 'timing/model': 5.374642366543412, 'learning_rate/qwen_vl_interface': 7.340682080553303e-06, 'learning_rate/action_model': 7.340682080553303e-05, 'epoch': 0.98}) +[2026-06-17 21:50:08,040][starVLA.training.train_starvla][INFO] - Step 2550, Loss: {'train/loss': 0.3996294438838959, 'train/grad_norm_pre_clip': 10.39974594116211, 'timing/data': 0.0011372528970241547, 'timing/model': 1.0912997853010893, 'learning_rate/qwen_vl_interface': 7.338739924778053e-06, 'learning_rate/action_model': 7.338739924778053e-05, 'epoch': 0.98}) +[2026-06-17 21:50:13,026][starVLA.training.train_starvla][INFO] - Step 2551, Loss: {'train/loss': 0.4379934072494507, 'train/grad_norm_pre_clip': 10.057332992553711, 'timing/data': 0.0007920628413558006, 'timing/model': 1.142550066113472, 'learning_rate/qwen_vl_interface': 7.336797336004771e-06, 'learning_rate/action_model': 7.33679733600477e-05, 'epoch': 0.98}) +[2026-06-17 21:50:18,024][starVLA.training.train_starvla][INFO] - Step 2552, Loss: {'train/loss': 0.31530094146728516, 'train/grad_norm_pre_clip': 7.335195541381836, 'timing/data': 0.0011773845180869102, 'timing/model': 1.3747478742152452, 'learning_rate/qwen_vl_interface': 7.334854314636153e-06, 'learning_rate/action_model': 7.334854314636152e-05, 'epoch': 0.98}) +[2026-06-17 21:50:22,690][starVLA.training.train_starvla][INFO] - Step 2553, Loss: {'train/loss': 0.5141826868057251, 'train/grad_norm_pre_clip': 9.904973983764648, 'timing/data': 0.0008896952494978905, 'timing/model': 1.0776373092085123, 'learning_rate/qwen_vl_interface': 7.3329108610749955e-06, 'learning_rate/action_model': 7.332910861074995e-05, 'epoch': 0.98}) +[2026-06-17 21:50:27,502][starVLA.training.train_starvla][INFO] - Step 2554, Loss: {'train/loss': 0.37065285444259644, 'train/grad_norm_pre_clip': 12.216103553771973, 'timing/data': 0.0008944971486926079, 'timing/model': 1.0942588988691568, 'learning_rate/qwen_vl_interface': 7.330966975724173e-06, 'learning_rate/action_model': 7.330966975724173e-05, 'epoch': 0.98}) +[2026-06-17 21:50:32,586][starVLA.training.train_starvla][INFO] - Step 2555, Loss: {'train/loss': 0.37285730242729187, 'train/grad_norm_pre_clip': 13.295125007629395, 'timing/data': 0.001240001991391182, 'timing/model': 1.3275930089876056, 'learning_rate/qwen_vl_interface': 7.329022658986657e-06, 'learning_rate/action_model': 7.329022658986658e-05, 'epoch': 0.98}) +[2026-06-17 21:50:37,301][starVLA.training.train_starvla][INFO] - Step 2556, Loss: {'train/loss': 0.34125807881355286, 'train/grad_norm_pre_clip': 6.864286422729492, 'timing/data': 0.0012271283194422722, 'timing/model': 1.1467213751748204, 'learning_rate/qwen_vl_interface': 7.327077911265508e-06, 'learning_rate/action_model': 7.327077911265507e-05, 'epoch': 0.98}) +[2026-06-17 21:50:42,023][starVLA.training.train_starvla][INFO] - Step 2557, Loss: {'train/loss': 0.6006531119346619, 'train/grad_norm_pre_clip': 9.98330307006836, 'timing/data': 0.0010907091200351715, 'timing/model': 1.4001840455457568, 'learning_rate/qwen_vl_interface': 7.325132732963872e-06, 'learning_rate/action_model': 7.325132732963872e-05, 'epoch': 0.98}) +[2026-06-17 21:50:46,447][starVLA.training.train_starvla][INFO] - Step 2558, Loss: {'train/loss': 0.610881507396698, 'train/grad_norm_pre_clip': 9.707499504089355, 'timing/data': 0.0010767718777060509, 'timing/model': 1.244696468114853, 'learning_rate/qwen_vl_interface': 7.323187124484988e-06, 'learning_rate/action_model': 7.323187124484988e-05, 'epoch': 0.98}) +[2026-06-17 21:50:54,213][starVLA.training.train_starvla][INFO] - Step 2559, Loss: {'train/loss': 0.6644535660743713, 'train/grad_norm_pre_clip': 10.38298511505127, 'timing/data': 0.0012660836800932884, 'timing/model': 1.9277865700423717, 'learning_rate/qwen_vl_interface': 7.3212410862321815e-06, 'learning_rate/action_model': 7.321241086232182e-05, 'epoch': 0.98}) +[2026-06-17 21:51:01,221][starVLA.training.train_starvla][INFO] - Step 2560, Loss: {'train/loss': 0.7508639693260193, 'train/grad_norm_pre_clip': 10.23734188079834, 'timing/data': 0.0011659199371933937, 'timing/model': 1.112696380354464, 'learning_rate/qwen_vl_interface': 7.319294618608869e-06, 'learning_rate/action_model': 7.319294618608868e-05, 'epoch': 0.98}) +[2026-06-17 21:51:06,348][starVLA.training.train_starvla][INFO] - Step 2561, Loss: {'train/loss': 0.6391123533248901, 'train/grad_norm_pre_clip': 10.55601978302002, 'timing/data': 0.0010665198788046837, 'timing/model': 1.1335783321410418, 'learning_rate/qwen_vl_interface': 7.317347722018557e-06, 'learning_rate/action_model': 7.317347722018557e-05, 'epoch': 0.99}) +[2026-06-17 21:51:16,648][starVLA.training.train_starvla][INFO] - Step 2562, Loss: {'train/loss': 0.649604320526123, 'train/grad_norm_pre_clip': 9.493753433227539, 'timing/data': 0.001108604483306408, 'timing/model': 1.5367058729752898, 'learning_rate/qwen_vl_interface': 7.3154003968648355e-06, 'learning_rate/action_model': 7.315400396864835e-05, 'epoch': 0.99}) +[2026-06-17 21:51:21,385][starVLA.training.train_starvla][INFO] - Step 2563, Loss: {'train/loss': 0.3746627867221832, 'train/grad_norm_pre_clip': 7.788583278656006, 'timing/data': 0.0013770535588264465, 'timing/model': 1.1137573393061757, 'learning_rate/qwen_vl_interface': 7.313452643551392e-06, 'learning_rate/action_model': 7.313452643551392e-05, 'epoch': 0.99}) +[2026-06-17 21:51:26,459][starVLA.training.train_starvla][INFO] - Step 2564, Loss: {'train/loss': 0.7475138306617737, 'train/grad_norm_pre_clip': 11.70171070098877, 'timing/data': 0.0011843815445899963, 'timing/model': 1.3633217606693506, 'learning_rate/qwen_vl_interface': 7.311504462481993e-06, 'learning_rate/action_model': 7.311504462481993e-05, 'epoch': 0.99}) +[2026-06-17 21:51:31,582][starVLA.training.train_starvla][INFO] - Step 2565, Loss: {'train/loss': 0.47122326493263245, 'train/grad_norm_pre_clip': 9.134988784790039, 'timing/data': 0.003124503418803215, 'timing/model': 1.189335827715695, 'learning_rate/qwen_vl_interface': 7.3095558540605e-06, 'learning_rate/action_model': 7.309555854060499e-05, 'epoch': 0.99}) +[2026-06-17 21:51:36,642][starVLA.training.train_starvla][INFO] - Step 2566, Loss: {'train/loss': 0.6068528890609741, 'train/grad_norm_pre_clip': 14.071122169494629, 'timing/data': 0.0013983482494950294, 'timing/model': 1.3576482655480504, 'learning_rate/qwen_vl_interface': 7.307606818690865e-06, 'learning_rate/action_model': 7.307606818690864e-05, 'epoch': 0.99}) +[2026-06-17 21:51:41,672][starVLA.training.train_starvla][INFO] - Step 2567, Loss: {'train/loss': 0.498122900724411, 'train/grad_norm_pre_clip': 7.670241355895996, 'timing/data': 0.001174519769847393, 'timing/model': 1.6922760708257556, 'learning_rate/qwen_vl_interface': 7.30565735677712e-06, 'learning_rate/action_model': 7.305657356777119e-05, 'epoch': 0.99}) +[2026-06-17 21:51:49,216][starVLA.training.train_starvla][INFO] - Step 2568, Loss: {'train/loss': 0.5061889290809631, 'train/grad_norm_pre_clip': 9.23484992980957, 'timing/data': 0.0010273726657032967, 'timing/model': 4.0317531786859035, 'learning_rate/qwen_vl_interface': 7.303707468723395e-06, 'learning_rate/action_model': 7.303707468723395e-05, 'epoch': 0.99}) +[2026-06-17 21:51:57,408][starVLA.training.train_starvla][INFO] - Step 2569, Loss: {'train/loss': 0.4011772871017456, 'train/grad_norm_pre_clip': 5.159534931182861, 'timing/data': 0.0016830749809741974, 'timing/model': 1.3179201846942306, 'learning_rate/qwen_vl_interface': 7.301757154933899e-06, 'learning_rate/action_model': 7.3017571549339e-05, 'epoch': 0.99}) +[2026-06-17 21:52:02,358][starVLA.training.train_starvla][INFO] - Step 2570, Loss: {'train/loss': 0.409099817276001, 'train/grad_norm_pre_clip': 6.615879535675049, 'timing/data': 0.0012493794783949852, 'timing/model': 1.1885343249887228, 'learning_rate/qwen_vl_interface': 7.29980641581294e-06, 'learning_rate/action_model': 7.299806415812939e-05, 'epoch': 0.99}) +[2026-06-17 21:52:11,247][starVLA.training.train_starvla][INFO] - Step 2571, Loss: {'train/loss': 0.40988749265670776, 'train/grad_norm_pre_clip': 7.126307487487793, 'timing/data': 0.001241643913090229, 'timing/model': 1.3557171113789082, 'learning_rate/qwen_vl_interface': 7.297855251764903e-06, 'learning_rate/action_model': 7.297855251764903e-05, 'epoch': 0.99}) +[2026-06-17 21:52:16,255][starVLA.training.train_starvla][INFO] - Step 2572, Loss: {'train/loss': 0.6330365538597107, 'train/grad_norm_pre_clip': 13.537405967712402, 'timing/data': 0.0010979566723108292, 'timing/model': 1.441539477556944, 'learning_rate/qwen_vl_interface': 7.295903663194269e-06, 'learning_rate/action_model': 7.29590366319427e-05, 'epoch': 0.99}) +[2026-06-17 21:52:22,563][starVLA.training.train_starvla][INFO] - Step 2573, Loss: {'train/loss': 0.398581862449646, 'train/grad_norm_pre_clip': 8.364212036132812, 'timing/data': 0.000996938906610012, 'timing/model': 2.9690694445744157, 'learning_rate/qwen_vl_interface': 7.293951650505606e-06, 'learning_rate/action_model': 7.293951650505607e-05, 'epoch': 0.99}) +[2026-06-17 21:52:27,586][starVLA.training.train_starvla][INFO] - Step 2574, Loss: {'train/loss': 0.42530325055122375, 'train/grad_norm_pre_clip': 7.705081939697266, 'timing/data': 0.0011122319847345352, 'timing/model': 1.1965575432404876, 'learning_rate/qwen_vl_interface': 7.291999214103565e-06, 'learning_rate/action_model': 7.291999214103564e-05, 'epoch': 0.99}) +[2026-06-17 21:52:32,449][starVLA.training.train_starvla][INFO] - Step 2575, Loss: {'train/loss': 0.556155264377594, 'train/grad_norm_pre_clip': 8.77442455291748, 'timing/data': 0.0011780792847275734, 'timing/model': 1.1376563403755426, 'learning_rate/qwen_vl_interface': 7.29004635439289e-06, 'learning_rate/action_model': 7.29004635439289e-05, 'epoch': 0.99}) +[2026-06-17 21:52:38,542][starVLA.training.train_starvla][INFO] - Step 2576, Loss: {'train/loss': 0.3182087540626526, 'train/grad_norm_pre_clip': 5.757761001586914, 'timing/data': 0.0011160755529999733, 'timing/model': 1.2063880460336804, 'learning_rate/qwen_vl_interface': 7.28809307177841e-06, 'learning_rate/action_model': 7.28809307177841e-05, 'epoch': 0.99}) +[2026-06-17 21:52:43,497][starVLA.training.train_starvla][INFO] - Step 2577, Loss: {'train/loss': 0.4959968328475952, 'train/grad_norm_pre_clip': 11.028328895568848, 'timing/data': 0.0011744620278477669, 'timing/model': 1.4682771852239966, 'learning_rate/qwen_vl_interface': 7.286139366665044e-06, 'learning_rate/action_model': 7.286139366665043e-05, 'epoch': 0.99}) +[2026-06-17 21:52:47,871][starVLA.training.train_starvla][INFO] - Step 2578, Loss: {'train/loss': 0.3746635317802429, 'train/grad_norm_pre_clip': 8.151766777038574, 'timing/data': 0.0011589443311095238, 'timing/model': 1.1605227580294013, 'learning_rate/qwen_vl_interface': 7.284185239457796e-06, 'learning_rate/action_model': 7.284185239457795e-05, 'epoch': 0.99}) +[2026-06-17 21:52:53,251][starVLA.training.train_starvla][INFO] - Step 2579, Loss: {'train/loss': 0.7005346417427063, 'train/grad_norm_pre_clip': 10.113362312316895, 'timing/data': 0.0011044163256883621, 'timing/model': 1.2157406453043222, 'learning_rate/qwen_vl_interface': 7.282230690561759e-06, 'learning_rate/action_model': 7.28223069056176e-05, 'epoch': 0.99}) +[2026-06-17 21:52:59,752][starVLA.training.train_starvla][INFO] - Step 2580, Loss: {'train/loss': 0.31838130950927734, 'train/grad_norm_pre_clip': 9.06469440460205, 'timing/data': 0.0011146748438477516, 'timing/model': 1.1669262377545238, 'learning_rate/qwen_vl_interface': 7.280275720382113e-06, 'learning_rate/action_model': 7.280275720382113e-05, 'epoch': 0.99}) +[2026-06-17 21:53:04,752][starVLA.training.train_starvla][INFO] - Step 2581, Loss: {'train/loss': 0.42732512950897217, 'train/grad_norm_pre_clip': 8.046188354492188, 'timing/data': 0.002107814885675907, 'timing/model': 1.129106585867703, 'learning_rate/qwen_vl_interface': 7.2783203293241246e-06, 'learning_rate/action_model': 7.278320329324125e-05, 'epoch': 0.99}) +[2026-06-17 21:53:17,184][starVLA.training.train_starvla][INFO] - Step 2582, Loss: {'train/loss': 0.5989651083946228, 'train/grad_norm_pre_clip': 10.57052230834961, 'timing/data': 0.0015597986057400703, 'timing/model': 5.385790716856718, 'learning_rate/qwen_vl_interface': 7.276364517793152e-06, 'learning_rate/action_model': 7.276364517793151e-05, 'epoch': 0.99}) +[2026-06-17 21:53:26,179][starVLA.training.train_starvla][INFO] - Step 2583, Loss: {'train/loss': 0.518337070941925, 'train/grad_norm_pre_clip': 11.266142845153809, 'timing/data': 0.0011026961728930473, 'timing/model': 1.373597539961338, 'learning_rate/qwen_vl_interface': 7.274408286194631e-06, 'learning_rate/action_model': 7.27440828619463e-05, 'epoch': 0.99}) +[2026-06-17 21:53:31,233][starVLA.training.train_starvla][INFO] - Step 2584, Loss: {'train/loss': 0.8258975744247437, 'train/grad_norm_pre_clip': 10.324986457824707, 'timing/data': 0.0013094134628772736, 'timing/model': 1.1459940867498517, 'learning_rate/qwen_vl_interface': 7.272451634934096e-06, 'learning_rate/action_model': 7.272451634934096e-05, 'epoch': 0.99}) +[2026-06-17 21:53:43,375][starVLA.training.train_starvla][INFO] - Step 2585, Loss: {'train/loss': 0.2735983431339264, 'train/grad_norm_pre_clip': 5.498037815093994, 'timing/data': 0.0009781615808606148, 'timing/model': 3.497126596979797, 'learning_rate/qwen_vl_interface': 7.270494564417161e-06, 'learning_rate/action_model': 7.270494564417161e-05, 'epoch': 0.99}) +[2026-06-17 21:53:48,361][starVLA.training.train_starvla][INFO] - Step 2586, Loss: {'train/loss': 0.5463632345199585, 'train/grad_norm_pre_clip': 12.179137229919434, 'timing/data': 0.0014251787215471268, 'timing/model': 1.1215700590983033, 'learning_rate/qwen_vl_interface': 7.268537075049527e-06, 'learning_rate/action_model': 7.268537075049526e-05, 'epoch': 0.99}) +[2026-06-17 21:53:53,679][starVLA.training.train_starvla][INFO] - Step 2587, Loss: {'train/loss': 0.3944205343723297, 'train/grad_norm_pre_clip': 8.724625587463379, 'timing/data': 0.0011806515976786613, 'timing/model': 1.4358517248183489, 'learning_rate/qwen_vl_interface': 7.2665791672369875e-06, 'learning_rate/action_model': 7.266579167236988e-05, 'epoch': 1.0}) +[2026-06-17 21:53:59,128][starVLA.training.train_starvla][INFO] - Step 2588, Loss: {'train/loss': 0.49530237913131714, 'train/grad_norm_pre_clip': 11.61401081085205, 'timing/data': 0.0013342704623937607, 'timing/model': 1.1602280903607607, 'learning_rate/qwen_vl_interface': 7.264620841385415e-06, 'learning_rate/action_model': 7.264620841385415e-05, 'epoch': 1.0}) +[2026-06-17 21:54:03,959][starVLA.training.train_starvla][INFO] - Step 2589, Loss: {'train/loss': 0.4310087561607361, 'train/grad_norm_pre_clip': 10.393672943115234, 'timing/data': 0.002070845104753971, 'timing/model': 1.144798269495368, 'learning_rate/qwen_vl_interface': 7.262662097900775e-06, 'learning_rate/action_model': 7.262662097900774e-05, 'epoch': 1.0}) +[2026-06-17 21:54:09,636][starVLA.training.train_starvla][INFO] - Step 2590, Loss: {'train/loss': 0.5011512637138367, 'train/grad_norm_pre_clip': 8.534582138061523, 'timing/data': 0.001156885176897049, 'timing/model': 1.2527986289933324, 'learning_rate/qwen_vl_interface': 7.260702937189116e-06, 'learning_rate/action_model': 7.260702937189115e-05, 'epoch': 1.0}) +[2026-06-17 21:54:14,512][starVLA.training.train_starvla][INFO] - Step 2591, Loss: {'train/loss': 0.3819703459739685, 'train/grad_norm_pre_clip': 7.5916523933410645, 'timing/data': 0.0014164112508296967, 'timing/model': 1.0885032126680017, 'learning_rate/qwen_vl_interface': 7.258743359656574e-06, 'learning_rate/action_model': 7.258743359656573e-05, 'epoch': 1.0}) +[2026-06-17 21:54:19,576][starVLA.training.train_starvla][INFO] - Step 2592, Loss: {'train/loss': 0.6821322441101074, 'train/grad_norm_pre_clip': 16.00214195251465, 'timing/data': 0.001284860074520111, 'timing/model': 1.4395210416987538, 'learning_rate/qwen_vl_interface': 7.2567833657093725e-06, 'learning_rate/action_model': 7.256783365709371e-05, 'epoch': 1.0}) +[2026-06-17 21:54:32,646][starVLA.training.train_starvla][INFO] - Step 2593, Loss: {'train/loss': 0.582740843296051, 'train/grad_norm_pre_clip': 10.801013946533203, 'timing/data': 0.0016372483223676682, 'timing/model': 5.462963999249041, 'learning_rate/qwen_vl_interface': 7.254822955753821e-06, 'learning_rate/action_model': 7.25482295575382e-05, 'epoch': 1.0}) +[2026-06-17 21:54:37,421][starVLA.training.train_starvla][INFO] - Step 2594, Loss: {'train/loss': 0.4634246230125427, 'train/grad_norm_pre_clip': 9.870539665222168, 'timing/data': 0.0010885298252105713, 'timing/model': 1.1372337844222784, 'learning_rate/qwen_vl_interface': 7.2528621301963135e-06, 'learning_rate/action_model': 7.252862130196313e-05, 'epoch': 1.0}) +[2026-06-17 21:54:43,959][starVLA.training.train_starvla][INFO] - Step 2595, Loss: {'train/loss': 0.6455627679824829, 'train/grad_norm_pre_clip': 13.650694847106934, 'timing/data': 0.0008164690807461739, 'timing/model': 1.2328467527404428, 'learning_rate/qwen_vl_interface': 7.25090088944333e-06, 'learning_rate/action_model': 7.250900889443329e-05, 'epoch': 1.0}) +[2026-06-17 21:54:53,507][starVLA.training.train_starvla][INFO] - Step 2596, Loss: {'train/loss': 0.47581997513771057, 'train/grad_norm_pre_clip': 10.75051212310791, 'timing/data': 0.001406751573085785, 'timing/model': 1.4483333248645067, 'learning_rate/qwen_vl_interface': 7.248939233901438e-06, 'learning_rate/action_model': 7.248939233901439e-05, 'epoch': 1.0}) +[2026-06-17 21:54:58,080][starVLA.training.train_starvla][INFO] - Step 2597, Loss: {'train/loss': 0.5341308116912842, 'train/grad_norm_pre_clip': 8.765311241149902, 'timing/data': 0.0009618382900953293, 'timing/model': 1.4394166059792042, 'learning_rate/qwen_vl_interface': 7.2469771639772944e-06, 'learning_rate/action_model': 7.246977163977294e-05, 'epoch': 1.0}) +[2026-06-17 21:55:02,379][starVLA.training.train_starvla][INFO] - Step 2598, Loss: {'train/loss': 0.49759185314178467, 'train/grad_norm_pre_clip': 7.9465861320495605, 'timing/data': 0.0010718181729316711, 'timing/model': 1.0980936149135232, 'learning_rate/qwen_vl_interface': 7.245014680077633e-06, 'learning_rate/action_model': 7.245014680077633e-05, 'epoch': 1.0}) +[2026-06-17 21:55:07,177][starVLA.training.train_starvla][INFO] - Step 2599, Loss: {'train/loss': 0.6123910546302795, 'train/grad_norm_pre_clip': 12.911410331726074, 'timing/data': 0.0012937337160110474, 'timing/model': 1.1069431332871318, 'learning_rate/qwen_vl_interface': 7.243051782609285e-06, 'learning_rate/action_model': 7.243051782609285e-05, 'epoch': 1.0}) +[2026-06-17 21:58:51,780][starVLA.training.train_starvla][INFO] - Step 2600, Loss: {'train/loss': 0.2104666382074356, 'train/grad_norm_pre_clip': 6.527577877044678, 'eval/loss': 0.5075297951698303, 'eval/latency_4/loss': 0.5075297951698303, 'eval/demon_fire_precision': 0.9566446748350612, 'eval/demon_fire_recall': 0.9889574537187399, 'eval/demon_fire_f1': 0.9725327371446822, 'eval/demon_fire_rate_model': 0.9348017621145375, 'eval/demon_fire_rate_teacher': 0.9042584434654919, 'eval/demon_left_precision': 0.803030303030303, 'eval/demon_left_recall': 0.6050228310502284, 'eval/demon_left_f1': 0.6901041666666667, 'eval/demon_left_rate_model': 0.09691629955947137, 'eval/demon_left_rate_teacher': 0.12863436123348018, 'eval/demon_right_precision': 0.8568075117370892, 'eval/demon_right_recall': 0.6588447653429603, 'eval/demon_right_f1': 0.7448979591836735, 'eval/demon_right_rate_model': 0.12511013215859032, 'eval/demon_right_rate_teacher': 0.16270190895741557, 'eval/demon_move_f1': 0.7175010629251701, 'eval/demon_cc_f1': 0.8788211432211345, 'eval/latency_4/demon_fire_precision': 0.9566446748350612, 'eval/latency_4/demon_fire_recall': 0.9889574537187399, 'eval/latency_4/demon_fire_f1': 0.9725327371446822, 'eval/latency_4/demon_fire_rate_model': 0.9348017621145375, 'eval/latency_4/demon_fire_rate_teacher': 0.9042584434654919, 'eval/latency_4/demon_left_precision': 0.803030303030303, 'eval/latency_4/demon_left_recall': 0.6050228310502284, 'eval/latency_4/demon_left_f1': 0.6901041666666667, 'eval/latency_4/demon_left_rate_model': 0.09691629955947137, 'eval/latency_4/demon_left_rate_teacher': 0.12863436123348018, 'eval/latency_4/demon_right_precision': 0.8568075117370892, 'eval/latency_4/demon_right_recall': 0.6588447653429603, 'eval/latency_4/demon_right_f1': 0.7448979591836735, 'eval/latency_4/demon_right_rate_model': 0.12511013215859032, 'eval/latency_4/demon_right_rate_teacher': 0.16270190895741557, 'eval/latency_4/demon_move_f1': 0.7175010629251701, 'eval/latency_4/demon_cc_f1': 0.8788211432211345, 'timing/data': 0.0009091990068554878, 'timing/model': 1.1296915523707867, 'learning_rate/qwen_vl_interface': 7.241088471979155e-06, 'learning_rate/action_model': 7.241088471979155e-05, 'epoch': 1.0}) +[2026-06-17 21:58:51,799][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_2600_state +[2026-06-17 21:58:51,809][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 21:59:00,679][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_2600_state/model.safetensors +[2026-06-17 21:59:23,775][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_2600_state/optimizer.bin +[2026-06-17 21:59:23,776][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_2600_state/sampler.bin +[2026-06-17 21:59:23,777][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_2600_state/sampler_1.bin +[2026-06-17 21:59:23,779][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_2600_state/random_states_0.pkl +[2026-06-17 21:59:23,786][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2600 +[2026-06-17 21:59:23,786][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 21:59:23,803][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 21:59:33,833][starVLA.training.train_starvla][INFO] - Step 2601, Loss: {'train/loss': 0.6472464799880981, 'train/grad_norm_pre_clip': 8.58060073852539, 'timing/data': 0.0011927159503102303, 'timing/model': 2.8972227582708, 'learning_rate/qwen_vl_interface': 7.239124748594244e-06, 'learning_rate/action_model': 7.239124748594244e-05, 'epoch': 1.0}) +[2026-06-17 21:59:49,344][starVLA.training.train_starvla][INFO] - Step 2602, Loss: {'train/loss': 0.5122830271720886, 'train/grad_norm_pre_clip': 9.338226318359375, 'timing/data': 0.0011063981801271439, 'timing/model': 1.583771213889122, 'learning_rate/qwen_vl_interface': 7.237160612861632e-06, 'learning_rate/action_model': 7.237160612861632e-05, 'epoch': 1.0}) +[2026-06-17 21:59:57,209][starVLA.training.train_starvla][INFO] - Step 2603, Loss: {'train/loss': 0.35123640298843384, 'train/grad_norm_pre_clip': 14.2896089553833, 'timing/data': 0.0008095521479845047, 'timing/model': 1.1571926632896066, 'learning_rate/qwen_vl_interface': 7.235196065188485e-06, 'learning_rate/action_model': 7.235196065188485e-05, 'epoch': 1.0}) +[2026-06-17 22:00:10,644][starVLA.training.train_starvla][INFO] - Step 2604, Loss: {'train/loss': 0.3086651861667633, 'train/grad_norm_pre_clip': 6.039185047149658, 'timing/data': 0.0011849133297801018, 'timing/model': 4.30458359234035, 'learning_rate/qwen_vl_interface': 7.23323110598206e-06, 'learning_rate/action_model': 7.23323110598206e-05, 'epoch': 1.0}) +[2026-06-17 22:00:19,095][starVLA.training.train_starvla][INFO] - Step 2605, Loss: {'train/loss': 0.4952879548072815, 'train/grad_norm_pre_clip': 6.597811222076416, 'timing/data': 0.001186266541481018, 'timing/model': 1.1491163242608309, 'learning_rate/qwen_vl_interface': 7.231265735649691e-06, 'learning_rate/action_model': 7.231265735649691e-05, 'epoch': 1.0}) +[2026-06-17 22:00:23,854][starVLA.training.train_starvla][INFO] - Step 2606, Loss: {'train/loss': 0.7213452458381653, 'train/grad_norm_pre_clip': 11.58738899230957, 'timing/data': 0.00128850806504488, 'timing/model': 1.2522290172055364, 'learning_rate/qwen_vl_interface': 7.229299954598805e-06, 'learning_rate/action_model': 7.229299954598804e-05, 'epoch': 1.0}) +[2026-06-17 22:00:28,880][starVLA.training.train_starvla][INFO] - Step 2607, Loss: {'train/loss': 0.3748895525932312, 'train/grad_norm_pre_clip': 8.144718170166016, 'timing/data': 0.0018212059512734413, 'timing/model': 1.2939700987190008, 'learning_rate/qwen_vl_interface': 7.227333763236907e-06, 'learning_rate/action_model': 7.227333763236906e-05, 'epoch': 1.0}) +[2026-06-17 22:00:34,435][starVLA.training.train_starvla][INFO] - Step 2608, Loss: {'train/loss': 0.5697404146194458, 'train/grad_norm_pre_clip': 9.3045015335083, 'timing/data': 0.001190907321870327, 'timing/model': 1.4620601553469896, 'learning_rate/qwen_vl_interface': 7.225367161971592e-06, 'learning_rate/action_model': 7.225367161971591e-05, 'epoch': 1.0}) +[2026-06-17 22:00:39,337][starVLA.training.train_starvla][INFO] - Step 2609, Loss: {'train/loss': 0.2570682168006897, 'train/grad_norm_pre_clip': 6.055659770965576, 'timing/data': 0.0009291945025324821, 'timing/model': 1.1712318127974868, 'learning_rate/qwen_vl_interface': 7.223400151210539e-06, 'learning_rate/action_model': 7.223400151210538e-05, 'epoch': 1.0}) +[2026-06-17 22:00:44,164][starVLA.training.train_starvla][INFO] - Step 2610, Loss: {'train/loss': 0.41421371698379517, 'train/grad_norm_pre_clip': 11.984886169433594, 'timing/data': 0.0012469738721847534, 'timing/model': 1.0873471358790994, 'learning_rate/qwen_vl_interface': 7.2214327313615105e-06, 'learning_rate/action_model': 7.22143273136151e-05, 'epoch': 1.0}) +[2026-06-17 22:00:48,967][starVLA.training.train_starvla][INFO] - Step 2611, Loss: {'train/loss': 0.57309889793396, 'train/grad_norm_pre_clip': 10.236639022827148, 'timing/data': 0.0012016193941235542, 'timing/model': 1.414610073901713, 'learning_rate/qwen_vl_interface': 7.219464902832357e-06, 'learning_rate/action_model': 7.219464902832356e-05, 'epoch': 1.0}) +[2026-06-17 22:00:53,696][starVLA.training.train_starvla][INFO] - Step 2612, Loss: {'train/loss': 0.5739300847053528, 'train/grad_norm_pre_clip': 10.780701637268066, 'timing/data': 0.0011674435809254646, 'timing/model': 1.3693430433049798, 'learning_rate/qwen_vl_interface': 7.2174966660310085e-06, 'learning_rate/action_model': 7.217496666031008e-05, 'epoch': 1.0}) +[2026-06-17 22:00:58,207][starVLA.training.train_starvla][INFO] - Step 2613, Loss: {'train/loss': 0.5273558497428894, 'train/grad_norm_pre_clip': 9.10898494720459, 'timing/data': 0.0014472352340817451, 'timing/model': 1.0904158474877477, 'learning_rate/qwen_vl_interface': 7.215528021365487e-06, 'learning_rate/action_model': 7.215528021365487e-05, 'epoch': 1.01}) +[2026-06-17 22:01:03,198][starVLA.training.train_starvla][INFO] - Step 2614, Loss: {'train/loss': 0.32949286699295044, 'train/grad_norm_pre_clip': 7.477908134460449, 'timing/data': 0.0009411638602614403, 'timing/model': 1.2334971278905869, 'learning_rate/qwen_vl_interface': 7.213558969243888e-06, 'learning_rate/action_model': 7.213558969243888e-05, 'epoch': 1.01}) +[2026-06-17 22:01:08,036][starVLA.training.train_starvla][INFO] - Step 2615, Loss: {'train/loss': 0.4771655201911926, 'train/grad_norm_pre_clip': 5.973881721496582, 'timing/data': 0.001168442890048027, 'timing/model': 1.1742805829271674, 'learning_rate/qwen_vl_interface': 7.211589510074404e-06, 'learning_rate/action_model': 7.211589510074404e-05, 'epoch': 1.01}) +[2026-06-17 22:01:13,169][starVLA.training.train_starvla][INFO] - Step 2616, Loss: {'train/loss': 0.31222906708717346, 'train/grad_norm_pre_clip': 9.526615142822266, 'timing/data': 0.0009728614240884781, 'timing/model': 1.203058465383947, 'learning_rate/qwen_vl_interface': 7.209619644265302e-06, 'learning_rate/action_model': 7.209619644265302e-05, 'epoch': 1.01}) +[2026-06-17 22:01:18,185][starVLA.training.train_starvla][INFO] - Step 2617, Loss: {'train/loss': 0.4753250181674957, 'train/grad_norm_pre_clip': 7.53707218170166, 'timing/data': 0.001138690859079361, 'timing/model': 1.5848613837733865, 'learning_rate/qwen_vl_interface': 7.20764937222494e-06, 'learning_rate/action_model': 7.20764937222494e-05, 'epoch': 1.01}) +[2026-06-17 22:01:22,733][starVLA.training.train_starvla][INFO] - Step 2618, Loss: {'train/loss': 0.4219417870044708, 'train/grad_norm_pre_clip': 10.720039367675781, 'timing/data': 0.0011290153488516808, 'timing/model': 1.0758279040455818, 'learning_rate/qwen_vl_interface': 7.205678694361757e-06, 'learning_rate/action_model': 7.205678694361757e-05, 'epoch': 1.01}) +[2026-06-17 22:01:27,545][starVLA.training.train_starvla][INFO] - Step 2619, Loss: {'train/loss': 0.4198492467403412, 'train/grad_norm_pre_clip': 10.281499862670898, 'timing/data': 0.0017945710569620132, 'timing/model': 1.0978237632662058, 'learning_rate/qwen_vl_interface': 7.203707611084276e-06, 'learning_rate/action_model': 7.203707611084275e-05, 'epoch': 1.01}) +[2026-06-17 22:01:37,637][starVLA.training.train_starvla][INFO] - Step 2620, Loss: {'train/loss': 0.6061505079269409, 'train/grad_norm_pre_clip': 12.433247566223145, 'timing/data': 0.0010271081700921059, 'timing/model': 2.8304731901735067, 'learning_rate/qwen_vl_interface': 7.2017361228011055e-06, 'learning_rate/action_model': 7.201736122801104e-05, 'epoch': 1.01}) +[2026-06-17 22:01:45,818][starVLA.training.train_starvla][INFO] - Step 2621, Loss: {'train/loss': 0.3126693367958069, 'train/grad_norm_pre_clip': 7.314639091491699, 'timing/data': 0.0012890594080090523, 'timing/model': 4.323446829803288, 'learning_rate/qwen_vl_interface': 7.199764229920935e-06, 'learning_rate/action_model': 7.199764229920934e-05, 'epoch': 1.01}) +[2026-06-17 22:01:59,072][starVLA.training.train_starvla][INFO] - Step 2622, Loss: {'train/loss': 0.4364824593067169, 'train/grad_norm_pre_clip': 8.69165325164795, 'timing/data': 0.0014223884791135788, 'timing/model': 5.858018079772592, 'learning_rate/qwen_vl_interface': 7.197791932852543e-06, 'learning_rate/action_model': 7.197791932852542e-05, 'epoch': 1.01}) +[2026-06-17 22:02:09,459][starVLA.training.train_starvla][INFO] - Step 2623, Loss: {'train/loss': 0.38749611377716064, 'train/grad_norm_pre_clip': 8.173599243164062, 'timing/data': 0.0011510718613862991, 'timing/model': 2.314051560126245, 'learning_rate/qwen_vl_interface': 7.195819232004787e-06, 'learning_rate/action_model': 7.195819232004787e-05, 'epoch': 1.01}) +[2026-06-17 22:02:15,964][starVLA.training.train_starvla][INFO] - Step 2624, Loss: {'train/loss': 0.3686090111732483, 'train/grad_norm_pre_clip': 7.428614139556885, 'timing/data': 0.020208400674164295, 'timing/model': 1.248083857819438, 'learning_rate/qwen_vl_interface': 7.1938461277866105e-06, 'learning_rate/action_model': 7.19384612778661e-05, 'epoch': 1.01}) +[2026-06-17 22:02:25,933][starVLA.training.train_starvla][INFO] - Step 2625, Loss: {'train/loss': 0.42666229605674744, 'train/grad_norm_pre_clip': 6.359691619873047, 'timing/data': 0.02136443555355072, 'timing/model': 5.632843050174415, 'learning_rate/qwen_vl_interface': 7.19187262060704e-06, 'learning_rate/action_model': 7.19187262060704e-05, 'epoch': 1.01}) +[2026-06-17 22:02:36,746][starVLA.training.train_starvla][INFO] - Step 2626, Loss: {'train/loss': 0.6049458384513855, 'train/grad_norm_pre_clip': 8.825989723205566, 'timing/data': 0.02556903474032879, 'timing/model': 2.746866420842707, 'learning_rate/qwen_vl_interface': 7.189898710875185e-06, 'learning_rate/action_model': 7.189898710875184e-05, 'epoch': 1.01}) +[2026-06-17 22:02:49,866][starVLA.training.train_starvla][INFO] - Step 2627, Loss: {'train/loss': 0.4531777799129486, 'train/grad_norm_pre_clip': 8.80905532836914, 'timing/data': 0.031710060313344, 'timing/model': 1.902446354739368, 'learning_rate/qwen_vl_interface': 7.187924399000241e-06, 'learning_rate/action_model': 7.18792439900024e-05, 'epoch': 1.01}) +[2026-06-17 22:02:57,323][starVLA.training.train_starvla][INFO] - Step 2628, Loss: {'train/loss': 0.35390934348106384, 'train/grad_norm_pre_clip': 15.13988208770752, 'timing/data': 0.02123651374131441, 'timing/model': 1.1212768545374274, 'learning_rate/qwen_vl_interface': 7.185949685391483e-06, 'learning_rate/action_model': 7.185949685391483e-05, 'epoch': 1.01}) +[2026-06-17 22:03:02,102][starVLA.training.train_starvla][INFO] - Step 2629, Loss: {'train/loss': 0.5219599604606628, 'train/grad_norm_pre_clip': 16.026578903198242, 'timing/data': 0.019056173041462898, 'timing/model': 1.0681080566719174, 'learning_rate/qwen_vl_interface': 7.183974570458272e-06, 'learning_rate/action_model': 7.183974570458272e-05, 'epoch': 1.01}) +[2026-06-17 22:03:07,150][starVLA.training.train_starvla][INFO] - Step 2630, Loss: {'train/loss': 0.6254990100860596, 'train/grad_norm_pre_clip': 13.90522575378418, 'timing/data': 0.023974628187716007, 'timing/model': 1.1885901000350714, 'learning_rate/qwen_vl_interface': 7.181999054610053e-06, 'learning_rate/action_model': 7.181999054610052e-05, 'epoch': 1.01}) +[2026-06-17 22:03:12,252][starVLA.training.train_starvla][INFO] - Step 2631, Loss: {'train/loss': 0.4010215699672699, 'train/grad_norm_pre_clip': 8.22802734375, 'timing/data': 0.022806868888437748, 'timing/model': 1.1049227556213737, 'learning_rate/qwen_vl_interface': 7.180023138256349e-06, 'learning_rate/action_model': 7.180023138256348e-05, 'epoch': 1.01}) +[2026-06-17 22:03:17,335][starVLA.training.train_starvla][INFO] - Step 2632, Loss: {'train/loss': 0.1980275809764862, 'train/grad_norm_pre_clip': 4.884043216705322, 'timing/data': 0.023942044004797935, 'timing/model': 1.830607844516635, 'learning_rate/qwen_vl_interface': 7.178046821806771e-06, 'learning_rate/action_model': 7.17804682180677e-05, 'epoch': 1.01}) +[2026-06-17 22:03:21,736][starVLA.training.train_starvla][INFO] - Step 2633, Loss: {'train/loss': 0.5498179197311401, 'train/grad_norm_pre_clip': 10.975555419921875, 'timing/data': 0.022354843094944954, 'timing/model': 1.1160861002281308, 'learning_rate/qwen_vl_interface': 7.176070105671016e-06, 'learning_rate/action_model': 7.176070105671015e-05, 'epoch': 1.01}) +[2026-06-17 22:03:27,037][starVLA.training.train_starvla][INFO] - Step 2634, Loss: {'train/loss': 0.6047397255897522, 'train/grad_norm_pre_clip': 12.383389472961426, 'timing/data': 0.01998700201511383, 'timing/model': 1.356029405258596, 'learning_rate/qwen_vl_interface': 7.174092990258852e-06, 'learning_rate/action_model': 7.174092990258851e-05, 'epoch': 1.01}) +[2026-06-17 22:03:31,813][starVLA.training.train_starvla][INFO] - Step 2635, Loss: {'train/loss': 0.6587561368942261, 'train/grad_norm_pre_clip': 9.60971736907959, 'timing/data': 0.021229577250778675, 'timing/model': 1.263569132424891, 'learning_rate/qwen_vl_interface': 7.172115475980142e-06, 'learning_rate/action_model': 7.172115475980142e-05, 'epoch': 1.01}) +[2026-06-17 22:03:36,852][starVLA.training.train_starvla][INFO] - Step 2636, Loss: {'train/loss': 0.37743836641311646, 'train/grad_norm_pre_clip': 9.485276222229004, 'timing/data': 0.024448849260807037, 'timing/model': 1.6259809890761971, 'learning_rate/qwen_vl_interface': 7.170137563244825e-06, 'learning_rate/action_model': 7.170137563244824e-05, 'epoch': 1.01}) +[2026-06-17 22:03:45,203][starVLA.training.train_starvla][INFO] - Step 2637, Loss: {'train/loss': 0.31204596161842346, 'train/grad_norm_pre_clip': 7.65964412689209, 'timing/data': 0.023494799621403217, 'timing/model': 1.7826424455270171, 'learning_rate/qwen_vl_interface': 7.168159252462927e-06, 'learning_rate/action_model': 7.168159252462926e-05, 'epoch': 1.01}) +[2026-06-17 22:03:49,851][starVLA.training.train_starvla][INFO] - Step 2638, Loss: {'train/loss': 0.2847365140914917, 'train/grad_norm_pre_clip': 6.103153228759766, 'timing/data': 0.02869264781475067, 'timing/model': 1.3562715025618672, 'learning_rate/qwen_vl_interface': 7.166180544044549e-06, 'learning_rate/action_model': 7.166180544044549e-05, 'epoch': 1.01}) +[2026-06-17 22:03:54,974][starVLA.training.train_starvla][INFO] - Step 2639, Loss: {'train/loss': 0.5124244093894958, 'train/grad_norm_pre_clip': 11.780855178833008, 'timing/data': 0.027547703124582767, 'timing/model': 1.293115503154695, 'learning_rate/qwen_vl_interface': 7.164201438399884e-06, 'learning_rate/action_model': 7.164201438399884e-05, 'epoch': 1.02}) +[2026-06-17 22:04:03,163][starVLA.training.train_starvla][INFO] - Step 2640, Loss: {'train/loss': 0.5040748119354248, 'train/grad_norm_pre_clip': 12.236603736877441, 'timing/data': 0.02441626973450184, 'timing/model': 4.367087377235293, 'learning_rate/qwen_vl_interface': 7.1622219359392e-06, 'learning_rate/action_model': 7.162221935939199e-05, 'epoch': 1.02}) +[2026-06-17 22:04:08,153][starVLA.training.train_starvla][INFO] - Step 2641, Loss: {'train/loss': 0.750680685043335, 'train/grad_norm_pre_clip': 10.026070594787598, 'timing/data': 0.023926055058836937, 'timing/model': 1.1243454720824957, 'learning_rate/qwen_vl_interface': 7.160242037072852e-06, 'learning_rate/action_model': 7.160242037072851e-05, 'epoch': 1.02}) +[2026-06-17 22:04:13,661][starVLA.training.train_starvla][INFO] - Step 2642, Loss: {'train/loss': 0.8166754841804504, 'train/grad_norm_pre_clip': 15.425027847290039, 'timing/data': 0.02272356953471899, 'timing/model': 2.2444822508841753, 'learning_rate/qwen_vl_interface': 7.158261742211275e-06, 'learning_rate/action_model': 7.158261742211274e-05, 'epoch': 1.02}) +[2026-06-17 22:04:18,207][starVLA.training.train_starvla][INFO] - Step 2643, Loss: {'train/loss': 0.6588870882987976, 'train/grad_norm_pre_clip': 13.90156364440918, 'timing/data': 0.020735652185976505, 'timing/model': 1.3091486366465688, 'learning_rate/qwen_vl_interface': 7.156281051764983e-06, 'learning_rate/action_model': 7.156281051764983e-05, 'epoch': 1.02}) +[2026-06-17 22:04:23,148][starVLA.training.train_starvla][INFO] - Step 2644, Loss: {'train/loss': 0.2963765859603882, 'train/grad_norm_pre_clip': 9.161286354064941, 'timing/data': 0.02129727229475975, 'timing/model': 1.0963263977319002, 'learning_rate/qwen_vl_interface': 7.15429996614458e-06, 'learning_rate/action_model': 7.15429996614458e-05, 'epoch': 1.02}) +[2026-06-17 22:04:27,937][starVLA.training.train_starvla][INFO] - Step 2645, Loss: {'train/loss': 0.29146808385849, 'train/grad_norm_pre_clip': 7.046273231506348, 'timing/data': 0.02732142060995102, 'timing/model': 1.220540308393538, 'learning_rate/qwen_vl_interface': 7.152318485760743e-06, 'learning_rate/action_model': 7.152318485760742e-05, 'epoch': 1.02}) +[2026-06-17 22:04:32,851][starVLA.training.train_starvla][INFO] - Step 2646, Loss: {'train/loss': 0.2598531246185303, 'train/grad_norm_pre_clip': 8.74045467376709, 'timing/data': 0.026709550991654396, 'timing/model': 1.6547521455213428, 'learning_rate/qwen_vl_interface': 7.150336611024239e-06, 'learning_rate/action_model': 7.150336611024239e-05, 'epoch': 1.02}) +[2026-06-17 22:04:39,166][starVLA.training.train_starvla][INFO] - Step 2647, Loss: {'train/loss': 0.3673591613769531, 'train/grad_norm_pre_clip': 8.963691711425781, 'timing/data': 0.02283411007374525, 'timing/model': 1.7643141690641642, 'learning_rate/qwen_vl_interface': 7.1483543423459086e-06, 'learning_rate/action_model': 7.148354342345908e-05, 'epoch': 1.02}) +[2026-06-17 22:04:44,089][starVLA.training.train_starvla][INFO] - Step 2648, Loss: {'train/loss': 0.2104896605014801, 'train/grad_norm_pre_clip': 6.825043678283691, 'timing/data': 0.01921028085052967, 'timing/model': 1.1822018418461084, 'learning_rate/qwen_vl_interface': 7.146371680136682e-06, 'learning_rate/action_model': 7.146371680136681e-05, 'epoch': 1.02}) +[2026-06-17 22:04:49,115][starVLA.training.train_starvla][INFO] - Step 2649, Loss: {'train/loss': 0.4459700584411621, 'train/grad_norm_pre_clip': 19.175722122192383, 'timing/data': 0.023704118095338345, 'timing/model': 1.0997189600020647, 'learning_rate/qwen_vl_interface': 7.144388624807565e-06, 'learning_rate/action_model': 7.144388624807566e-05, 'epoch': 1.02}) +[2026-06-17 22:04:54,683][starVLA.training.train_starvla][INFO] - Step 2650, Loss: {'train/loss': 0.4773878753185272, 'train/grad_norm_pre_clip': 9.868474006652832, 'timing/data': 0.0210382342338562, 'timing/model': 1.2258572597056627, 'learning_rate/qwen_vl_interface': 7.142405176769648e-06, 'learning_rate/action_model': 7.142405176769648e-05, 'epoch': 1.02}) +[2026-06-17 22:04:59,641][starVLA.training.train_starvla][INFO] - Step 2651, Loss: {'train/loss': 0.68988037109375, 'train/grad_norm_pre_clip': 8.358495712280273, 'timing/data': 0.024483672343194485, 'timing/model': 1.2773010535165668, 'learning_rate/qwen_vl_interface': 7.140421336434102e-06, 'learning_rate/action_model': 7.140421336434102e-05, 'epoch': 1.02}) +[2026-06-17 22:05:04,650][starVLA.training.train_starvla][INFO] - Step 2652, Loss: {'train/loss': 0.7126986980438232, 'train/grad_norm_pre_clip': 12.322091102600098, 'timing/data': 0.02960359863936901, 'timing/model': 1.6719755213707685, 'learning_rate/qwen_vl_interface': 7.138437104212176e-06, 'learning_rate/action_model': 7.138437104212176e-05, 'epoch': 1.02}) +[2026-06-17 22:05:09,111][starVLA.training.train_starvla][INFO] - Step 2653, Loss: {'train/loss': 0.3619929254055023, 'train/grad_norm_pre_clip': 11.31468391418457, 'timing/data': 0.020591868087649345, 'timing/model': 1.1439976133406162, 'learning_rate/qwen_vl_interface': 7.136452480515208e-06, 'learning_rate/action_model': 7.136452480515208e-05, 'epoch': 1.02}) +[2026-06-17 22:05:16,591][starVLA.training.train_starvla][INFO] - Step 2654, Loss: {'train/loss': 0.39550402760505676, 'train/grad_norm_pre_clip': 6.111557960510254, 'timing/data': 0.022963661700487137, 'timing/model': 1.1592892073094845, 'learning_rate/qwen_vl_interface': 7.134467465754612e-06, 'learning_rate/action_model': 7.134467465754612e-05, 'epoch': 1.02}) +[2026-06-17 22:05:21,708][starVLA.training.train_starvla][INFO] - Step 2655, Loss: {'train/loss': 0.6022298336029053, 'train/grad_norm_pre_clip': 13.668547630310059, 'timing/data': 0.020600113086402416, 'timing/model': 1.075092631392181, 'learning_rate/qwen_vl_interface': 7.132482060341881e-06, 'learning_rate/action_model': 7.13248206034188e-05, 'epoch': 1.02}) +[2026-06-17 22:05:26,439][starVLA.training.train_starvla][INFO] - Step 2656, Loss: {'train/loss': 0.7680826187133789, 'train/grad_norm_pre_clip': 11.84183406829834, 'timing/data': 0.020575590431690216, 'timing/model': 1.4361837450414896, 'learning_rate/qwen_vl_interface': 7.1304962646885935e-06, 'learning_rate/action_model': 7.130496264688593e-05, 'epoch': 1.02}) +[2026-06-17 22:05:33,016][starVLA.training.train_starvla][INFO] - Step 2657, Loss: {'train/loss': 0.23168449103832245, 'train/grad_norm_pre_clip': 8.421992301940918, 'timing/data': 0.03397082071751356, 'timing/model': 3.0441323313862085, 'learning_rate/qwen_vl_interface': 7.128510079206408e-06, 'learning_rate/action_model': 7.128510079206407e-05, 'epoch': 1.02}) +[2026-06-17 22:05:37,608][starVLA.training.train_starvla][INFO] - Step 2658, Loss: {'train/loss': 0.3773329555988312, 'train/grad_norm_pre_clip': 7.01515531539917, 'timing/data': 0.019672931171953678, 'timing/model': 1.2587148612365127, 'learning_rate/qwen_vl_interface': 7.126523504307061e-06, 'learning_rate/action_model': 7.126523504307061e-05, 'epoch': 1.02}) +[2026-06-17 22:05:42,153][starVLA.training.train_starvla][INFO] - Step 2659, Loss: {'train/loss': 0.19329938292503357, 'train/grad_norm_pre_clip': 25.275148391723633, 'timing/data': 0.020746875554323196, 'timing/model': 1.0984447347000241, 'learning_rate/qwen_vl_interface': 7.124536540402371e-06, 'learning_rate/action_model': 7.124536540402372e-05, 'epoch': 1.02}) +[2026-06-17 22:05:47,538][starVLA.training.train_starvla][INFO] - Step 2660, Loss: {'train/loss': 0.4175474941730499, 'train/grad_norm_pre_clip': 8.250707626342773, 'timing/data': 0.020029829815030098, 'timing/model': 1.2582912938669324, 'learning_rate/qwen_vl_interface': 7.1225491879042405e-06, 'learning_rate/action_model': 7.12254918790424e-05, 'epoch': 1.02}) +[2026-06-17 22:05:53,617][starVLA.training.train_starvla][INFO] - Step 2661, Loss: {'train/loss': 0.39379310607910156, 'train/grad_norm_pre_clip': 7.5873284339904785, 'timing/data': 0.025091182440519333, 'timing/model': 1.2471770122647285, 'learning_rate/qwen_vl_interface': 7.120561447224649e-06, 'learning_rate/action_model': 7.120561447224649e-05, 'epoch': 1.02}) +[2026-06-17 22:05:59,183][starVLA.training.train_starvla][INFO] - Step 2662, Loss: {'train/loss': 0.6291475892066956, 'train/grad_norm_pre_clip': 11.908628463745117, 'timing/data': 0.020505595952272415, 'timing/model': 1.5551842385903, 'learning_rate/qwen_vl_interface': 7.118573318775656e-06, 'learning_rate/action_model': 7.118573318775655e-05, 'epoch': 1.02}) +[2026-06-17 22:06:03,616][starVLA.training.train_starvla][INFO] - Step 2663, Loss: {'train/loss': 0.31443968415260315, 'train/grad_norm_pre_clip': 6.4685797691345215, 'timing/data': 0.021238546818494797, 'timing/model': 1.0899665765464306, 'learning_rate/qwen_vl_interface': 7.116584802969403e-06, 'learning_rate/action_model': 7.116584802969402e-05, 'epoch': 1.02}) +[2026-06-17 22:06:08,731][starVLA.training.train_starvla][INFO] - Step 2664, Loss: {'train/loss': 0.4670273959636688, 'train/grad_norm_pre_clip': 16.460046768188477, 'timing/data': 0.02624857146292925, 'timing/model': 1.3538385573774576, 'learning_rate/qwen_vl_interface': 7.114595900218112e-06, 'learning_rate/action_model': 7.114595900218112e-05, 'epoch': 1.02}) +[2026-06-17 22:06:13,853][starVLA.training.train_starvla][INFO] - Step 2665, Loss: {'train/loss': 0.8680171966552734, 'train/grad_norm_pre_clip': 19.317670822143555, 'timing/data': 0.02472539898008108, 'timing/model': 1.153910607099533, 'learning_rate/qwen_vl_interface': 7.112606610934084e-06, 'learning_rate/action_model': 7.112606610934084e-05, 'epoch': 1.03}) +[2026-06-17 22:06:18,715][starVLA.training.train_starvla][INFO] - Step 2666, Loss: {'train/loss': 0.5322073698043823, 'train/grad_norm_pre_clip': 10.88264274597168, 'timing/data': 0.020192768424749374, 'timing/model': 1.6268180347979069, 'learning_rate/qwen_vl_interface': 7.110616935529702e-06, 'learning_rate/action_model': 7.110616935529702e-05, 'epoch': 1.03}) +[2026-06-17 22:06:23,440][starVLA.training.train_starvla][INFO] - Step 2667, Loss: {'train/loss': 0.5002624988555908, 'train/grad_norm_pre_clip': 7.869185924530029, 'timing/data': 0.026401489973068237, 'timing/model': 1.4308619778603315, 'learning_rate/qwen_vl_interface': 7.108626874417426e-06, 'learning_rate/action_model': 7.108626874417426e-05, 'epoch': 1.03}) +[2026-06-17 22:06:27,760][starVLA.training.train_starvla][INFO] - Step 2668, Loss: {'train/loss': 0.38807323575019836, 'train/grad_norm_pre_clip': 8.941522598266602, 'timing/data': 0.02245443407446146, 'timing/model': 1.184282036498189, 'learning_rate/qwen_vl_interface': 7.106636428009801e-06, 'learning_rate/action_model': 7.1066364280098e-05, 'epoch': 1.03}) +[2026-06-17 22:06:32,882][starVLA.training.train_starvla][INFO] - Step 2669, Loss: {'train/loss': 0.37145644426345825, 'train/grad_norm_pre_clip': 8.461074829101562, 'timing/data': 0.029589404352009296, 'timing/model': 1.2783458437770605, 'learning_rate/qwen_vl_interface': 7.104645596719445e-06, 'learning_rate/action_model': 7.104645596719445e-05, 'epoch': 1.03}) +[2026-06-17 22:06:39,257][starVLA.training.train_starvla][INFO] - Step 2670, Loss: {'train/loss': 0.4533587098121643, 'train/grad_norm_pre_clip': 8.052349090576172, 'timing/data': 0.019222446717321873, 'timing/model': 1.1032290430739522, 'learning_rate/qwen_vl_interface': 7.102654380959063e-06, 'learning_rate/action_model': 7.102654380959063e-05, 'epoch': 1.03}) +[2026-06-17 22:06:44,183][starVLA.training.train_starvla][INFO] - Step 2671, Loss: {'train/loss': 0.5805069208145142, 'train/grad_norm_pre_clip': 10.016294479370117, 'timing/data': 0.0284148957580328, 'timing/model': 1.0648867981508374, 'learning_rate/qwen_vl_interface': 7.100662781141432e-06, 'learning_rate/action_model': 7.100662781141432e-05, 'epoch': 1.03}) +[2026-06-17 22:06:48,954][starVLA.training.train_starvla][INFO] - Step 2672, Loss: {'train/loss': 0.35512083768844604, 'train/grad_norm_pre_clip': 8.153851509094238, 'timing/data': 0.011081679724156857, 'timing/model': 1.3870990769937634, 'learning_rate/qwen_vl_interface': 7.0986707976794155e-06, 'learning_rate/action_model': 7.098670797679415e-05, 'epoch': 1.03}) +[2026-06-17 22:06:54,766][starVLA.training.train_starvla][INFO] - Step 2673, Loss: {'train/loss': 0.5208148956298828, 'train/grad_norm_pre_clip': 6.881781578063965, 'timing/data': 0.010805684141814709, 'timing/model': 1.0659981723874807, 'learning_rate/qwen_vl_interface': 7.096678430985955e-06, 'learning_rate/action_model': 7.096678430985954e-05, 'epoch': 1.03}) +[2026-06-17 22:06:59,096][starVLA.training.train_starvla][INFO] - Step 2674, Loss: {'train/loss': 0.41128483414649963, 'train/grad_norm_pre_clip': 7.296501636505127, 'timing/data': 0.010933468118309975, 'timing/model': 1.0914973076432943, 'learning_rate/qwen_vl_interface': 7.094685681474066e-06, 'learning_rate/action_model': 7.094685681474065e-05, 'epoch': 1.03}) +[2026-06-17 22:07:03,366][starVLA.training.train_starvla][INFO] - Step 2675, Loss: {'train/loss': 0.5175393223762512, 'train/grad_norm_pre_clip': 9.15712833404541, 'timing/data': 0.010488836094737053, 'timing/model': 1.0391680411994457, 'learning_rate/qwen_vl_interface': 7.092692549556853e-06, 'learning_rate/action_model': 7.092692549556853e-05, 'epoch': 1.03}) +[2026-06-17 22:07:07,822][starVLA.training.train_starvla][INFO] - Step 2676, Loss: {'train/loss': 0.38653600215911865, 'train/grad_norm_pre_clip': 8.763322830200195, 'timing/data': 0.014989150688052177, 'timing/model': 1.3784851506352425, 'learning_rate/qwen_vl_interface': 7.0906990356474895e-06, 'learning_rate/action_model': 7.090699035647489e-05, 'epoch': 1.03}) +[2026-06-17 22:07:12,175][starVLA.training.train_starvla][INFO] - Step 2677, Loss: {'train/loss': 0.20847460627555847, 'train/grad_norm_pre_clip': 5.335050582885742, 'timing/data': 0.009478050284087658, 'timing/model': 1.2663112431764603, 'learning_rate/qwen_vl_interface': 7.088705140159238e-06, 'learning_rate/action_model': 7.088705140159237e-05, 'epoch': 1.03}) +[2026-06-17 22:07:16,259][starVLA.training.train_starvla][INFO] - Step 2678, Loss: {'train/loss': 0.607919454574585, 'train/grad_norm_pre_clip': 11.483071327209473, 'timing/data': 0.009339137002825737, 'timing/model': 1.0315758883953094, 'learning_rate/qwen_vl_interface': 7.086710863505428e-06, 'learning_rate/action_model': 7.086710863505428e-05, 'epoch': 1.03}) +[2026-06-17 22:07:20,464][starVLA.training.train_starvla][INFO] - Step 2679, Loss: {'train/loss': 0.5126445889472961, 'train/grad_norm_pre_clip': 8.719155311584473, 'timing/data': 0.00937571283429861, 'timing/model': 1.0441273245960474, 'learning_rate/qwen_vl_interface': 7.084716206099485e-06, 'learning_rate/action_model': 7.084716206099484e-05, 'epoch': 1.03}) +[2026-06-17 22:07:24,743][starVLA.training.train_starvla][INFO] - Step 2680, Loss: {'train/loss': 0.37892064452171326, 'train/grad_norm_pre_clip': 7.789432525634766, 'timing/data': 0.011917348019778728, 'timing/model': 1.0464708367362618, 'learning_rate/qwen_vl_interface': 7.082721168354894e-06, 'learning_rate/action_model': 7.082721168354894e-05, 'epoch': 1.03}) +[2026-06-17 22:07:28,946][starVLA.training.train_starvla][INFO] - Step 2681, Loss: {'train/loss': 0.21538779139518738, 'train/grad_norm_pre_clip': 5.654221534729004, 'timing/data': 0.009222857654094696, 'timing/model': 1.0419264240190387, 'learning_rate/qwen_vl_interface': 7.080725750685235e-06, 'learning_rate/action_model': 7.080725750685234e-05, 'epoch': 1.03}) +[2026-06-17 22:07:33,199][starVLA.training.train_starvla][INFO] - Step 2682, Loss: {'train/loss': 0.7606934309005737, 'train/grad_norm_pre_clip': 9.646095275878906, 'timing/data': 0.009404787793755531, 'timing/model': 1.2793063316494226, 'learning_rate/qwen_vl_interface': 7.078729953504157e-06, 'learning_rate/action_model': 7.078729953504156e-05, 'epoch': 1.03}) +[2026-06-17 22:07:37,247][starVLA.training.train_starvla][INFO] - Step 2683, Loss: {'train/loss': 0.5228740572929382, 'train/grad_norm_pre_clip': 9.650274276733398, 'timing/data': 0.009973502717912197, 'timing/model': 1.0473756147548556, 'learning_rate/qwen_vl_interface': 7.076733777225391e-06, 'learning_rate/action_model': 7.07673377722539e-05, 'epoch': 1.03}) +[2026-06-17 22:07:42,014][starVLA.training.train_starvla][INFO] - Step 2684, Loss: {'train/loss': 0.5497288107872009, 'train/grad_norm_pre_clip': 8.872047424316406, 'timing/data': 0.011396193876862526, 'timing/model': 1.131449643522501, 'learning_rate/qwen_vl_interface': 7.074737222262749e-06, 'learning_rate/action_model': 7.074737222262748e-05, 'epoch': 1.03}) +[2026-06-17 22:07:48,381][starVLA.training.train_starvla][INFO] - Step 2685, Loss: {'train/loss': 0.5135157108306885, 'train/grad_norm_pre_clip': 9.03055191040039, 'timing/data': 0.012263461947441101, 'timing/model': 1.1960233692079782, 'learning_rate/qwen_vl_interface': 7.072740289030115e-06, 'learning_rate/action_model': 7.072740289030115e-05, 'epoch': 1.03}) +[2026-06-17 22:07:53,546][starVLA.training.train_starvla][INFO] - Step 2686, Loss: {'train/loss': 0.5957098603248596, 'train/grad_norm_pre_clip': 10.058893203735352, 'timing/data': 0.010663052089512348, 'timing/model': 1.4774883594363928, 'learning_rate/qwen_vl_interface': 7.07074297794146e-06, 'learning_rate/action_model': 7.07074297794146e-05, 'epoch': 1.03}) +[2026-06-17 22:07:58,582][starVLA.training.train_starvla][INFO] - Step 2687, Loss: {'train/loss': 0.6061983704566956, 'train/grad_norm_pre_clip': 11.309076309204102, 'timing/data': 0.019755087792873383, 'timing/model': 1.6273560710251331, 'learning_rate/qwen_vl_interface': 7.068745289410826e-06, 'learning_rate/action_model': 7.068745289410827e-05, 'epoch': 1.03}) +[2026-06-17 22:08:09,707][starVLA.training.train_starvla][INFO] - Step 2688, Loss: {'train/loss': 0.6944382786750793, 'train/grad_norm_pre_clip': 11.595412254333496, 'timing/data': 0.011081616394221783, 'timing/model': 4.644071505405009, 'learning_rate/qwen_vl_interface': 7.066747223852337e-06, 'learning_rate/action_model': 7.066747223852337e-05, 'epoch': 1.03}) +[2026-06-17 22:08:16,948][starVLA.training.train_starvla][INFO] - Step 2689, Loss: {'train/loss': 0.37802526354789734, 'train/grad_norm_pre_clip': 8.653428077697754, 'timing/data': 0.011034088209271431, 'timing/model': 1.1826646076515317, 'learning_rate/qwen_vl_interface': 7.064748781680192e-06, 'learning_rate/action_model': 7.064748781680192e-05, 'epoch': 1.03}) +[2026-06-17 22:08:21,995][starVLA.training.train_starvla][INFO] - Step 2690, Loss: {'train/loss': 0.5622760653495789, 'train/grad_norm_pre_clip': 9.324138641357422, 'timing/data': 0.011965651996433735, 'timing/model': 1.2967969616875052, 'learning_rate/qwen_vl_interface': 7.062749963308673e-06, 'learning_rate/action_model': 7.062749963308672e-05, 'epoch': 1.03}) +[2026-06-17 22:08:26,615][starVLA.training.train_starvla][INFO] - Step 2691, Loss: {'train/loss': 0.21210063993930817, 'train/grad_norm_pre_clip': 7.298324108123779, 'timing/data': 0.011448321864008904, 'timing/model': 1.0845988616347313, 'learning_rate/qwen_vl_interface': 7.060750769152135e-06, 'learning_rate/action_model': 7.060750769152135e-05, 'epoch': 1.04}) +[2026-06-17 22:08:31,568][starVLA.training.train_starvla][INFO] - Step 2692, Loss: {'train/loss': 0.5691904425621033, 'train/grad_norm_pre_clip': 10.130069732666016, 'timing/data': 0.01180325634777546, 'timing/model': 1.5882105603814125, 'learning_rate/qwen_vl_interface': 7.058751199625014e-06, 'learning_rate/action_model': 7.058751199625015e-05, 'epoch': 1.04}) +[2026-06-17 22:08:36,441][starVLA.training.train_starvla][INFO] - Step 2693, Loss: {'train/loss': 0.6072810888290405, 'train/grad_norm_pre_clip': 9.485618591308594, 'timing/data': 0.011137868277728558, 'timing/model': 1.2387793846428394, 'learning_rate/qwen_vl_interface': 7.056751255141824e-06, 'learning_rate/action_model': 7.056751255141824e-05, 'epoch': 1.04}) +[2026-06-17 22:08:41,567][starVLA.training.train_starvla][INFO] - Step 2694, Loss: {'train/loss': 0.39636892080307007, 'train/grad_norm_pre_clip': 7.792229175567627, 'timing/data': 0.012202112004160881, 'timing/model': 1.1335222441703081, 'learning_rate/qwen_vl_interface': 7.054750936117155e-06, 'learning_rate/action_model': 7.054750936117156e-05, 'epoch': 1.04}) +[2026-06-17 22:08:46,342][starVLA.training.train_starvla][INFO] - Step 2695, Loss: {'train/loss': 0.5309743881225586, 'train/grad_norm_pre_clip': 9.806447982788086, 'timing/data': 0.01086671743541956, 'timing/model': 1.080668698064983, 'learning_rate/qwen_vl_interface': 7.0527502429656745e-06, 'learning_rate/action_model': 7.052750242965674e-05, 'epoch': 1.04}) +[2026-06-17 22:08:59,033][starVLA.training.train_starvla][INFO] - Step 2696, Loss: {'train/loss': 0.5104264616966248, 'train/grad_norm_pre_clip': 9.049286842346191, 'timing/data': 0.0131291588768363, 'timing/model': 2.070475463755429, 'learning_rate/qwen_vl_interface': 7.050749176102128e-06, 'learning_rate/action_model': 7.050749176102129e-05, 'epoch': 1.04}) +[2026-06-17 22:09:04,041][starVLA.training.train_starvla][INFO] - Step 2697, Loss: {'train/loss': 0.488095760345459, 'train/grad_norm_pre_clip': 10.83263111114502, 'timing/data': 0.010859907604753971, 'timing/model': 1.4903503963723779, 'learning_rate/qwen_vl_interface': 7.048747735941342e-06, 'learning_rate/action_model': 7.048747735941341e-05, 'epoch': 1.04}) +[2026-06-17 22:09:15,038][starVLA.training.train_starvla][INFO] - Step 2698, Loss: {'train/loss': 0.49262407422065735, 'train/grad_norm_pre_clip': 9.397034645080566, 'timing/data': 0.011705373413860798, 'timing/model': 1.422526960261166, 'learning_rate/qwen_vl_interface': 7.046745922898212e-06, 'learning_rate/action_model': 7.046745922898211e-05, 'epoch': 1.04}) +[2026-06-17 22:09:21,494][starVLA.training.train_starvla][INFO] - Step 2699, Loss: {'train/loss': 0.38234439492225647, 'train/grad_norm_pre_clip': 6.770119667053223, 'timing/data': 0.010894553735852242, 'timing/model': 1.6431962447240949, 'learning_rate/qwen_vl_interface': 7.04474373738772e-06, 'learning_rate/action_model': 7.04474373738772e-05, 'epoch': 1.04}) +[2026-06-17 22:13:43,335][starVLA.training.train_starvla][INFO] - Step 2700, Loss: {'train/loss': 0.47927361726760864, 'train/grad_norm_pre_clip': 8.59151554107666, 'eval/loss': 0.49730923771858215, 'eval/latency_4/loss': 0.49730923771858215, 'eval/demon_fire_precision': 0.9593059936908518, 'eval/demon_fire_recall': 0.9876583306268268, 'eval/demon_fire_f1': 0.9732757241158586, 'eval/demon_fire_rate_model': 0.9309838472834068, 'eval/demon_fire_rate_teacher': 0.9042584434654919, 'eval/demon_left_precision': 0.7603305785123967, 'eval/demon_left_recall': 0.6301369863013698, 'eval/demon_left_f1': 0.6891385767790262, 'eval/demon_left_rate_model': 0.1066079295154185, 'eval/demon_left_rate_teacher': 0.12863436123348018, 'eval/demon_right_precision': 0.8526570048309179, 'eval/demon_right_recall': 0.6371841155234657, 'eval/demon_right_f1': 0.7293388429752066, 'eval/demon_right_rate_model': 0.12158590308370044, 'eval/demon_right_rate_teacher': 0.16270190895741557, 'eval/demon_move_f1': 0.7092387098771165, 'eval/demon_cc_f1': 0.8754962034862339, 'eval/latency_4/demon_fire_precision': 0.9593059936908518, 'eval/latency_4/demon_fire_recall': 0.9876583306268268, 'eval/latency_4/demon_fire_f1': 0.9732757241158586, 'eval/latency_4/demon_fire_rate_model': 0.9309838472834068, 'eval/latency_4/demon_fire_rate_teacher': 0.9042584434654919, 'eval/latency_4/demon_left_precision': 0.7603305785123967, 'eval/latency_4/demon_left_recall': 0.6301369863013698, 'eval/latency_4/demon_left_f1': 0.6891385767790262, 'eval/latency_4/demon_left_rate_model': 0.1066079295154185, 'eval/latency_4/demon_left_rate_teacher': 0.12863436123348018, 'eval/latency_4/demon_right_precision': 0.8526570048309179, 'eval/latency_4/demon_right_recall': 0.6371841155234657, 'eval/latency_4/demon_right_f1': 0.7293388429752066, 'eval/latency_4/demon_right_rate_model': 0.12158590308370044, 'eval/latency_4/demon_right_rate_teacher': 0.16270190895741557, 'eval/latency_4/demon_move_f1': 0.7092387098771165, 'eval/latency_4/demon_cc_f1': 0.8754962034862339, 'timing/data': 0.011631138622760773, 'timing/model': 1.2306918343529105, 'learning_rate/qwen_vl_interface': 7.042741179824919e-06, 'learning_rate/action_model': 7.042741179824918e-05, 'epoch': 1.04}) +[2026-06-17 22:13:43,360][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_2700_state +[2026-06-17 22:13:43,368][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 22:13:52,251][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_2700_state/model.safetensors +[2026-06-17 22:14:14,352][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_2700_state/optimizer.bin +[2026-06-17 22:14:14,355][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_2700_state/sampler.bin +[2026-06-17 22:14:14,356][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_2700_state/sampler_1.bin +[2026-06-17 22:14:14,358][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_2700_state/random_states_0.pkl +[2026-06-17 22:14:14,363][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2700 +[2026-06-17 22:14:14,364][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 22:14:14,407][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 22:14:19,477][starVLA.training.train_starvla][INFO] - Step 2701, Loss: {'train/loss': 0.70835280418396, 'train/grad_norm_pre_clip': 14.672572135925293, 'timing/data': 0.011749664321541786, 'timing/model': 1.6747486675158143, 'learning_rate/qwen_vl_interface': 7.040738250624944e-06, 'learning_rate/action_model': 7.040738250624944e-05, 'epoch': 1.04}) +[2026-06-17 22:14:24,294][starVLA.training.train_starvla][INFO] - Step 2702, Loss: {'train/loss': 0.3861221671104431, 'train/grad_norm_pre_clip': 10.346553802490234, 'timing/data': 0.017765440978109837, 'timing/model': 1.5289425710216165, 'learning_rate/qwen_vl_interface': 7.038734950203001e-06, 'learning_rate/action_model': 7.038734950203001e-05, 'epoch': 1.04}) +[2026-06-17 22:14:28,773][starVLA.training.train_starvla][INFO] - Step 2703, Loss: {'train/loss': 0.7395547032356262, 'train/grad_norm_pre_clip': 10.94840145111084, 'timing/data': 0.013974477536976337, 'timing/model': 1.2155252918601036, 'learning_rate/qwen_vl_interface': 7.036731278974379e-06, 'learning_rate/action_model': 7.036731278974378e-05, 'epoch': 1.04}) +[2026-06-17 22:14:33,604][starVLA.training.train_starvla][INFO] - Step 2704, Loss: {'train/loss': 0.46258318424224854, 'train/grad_norm_pre_clip': 9.580008506774902, 'timing/data': 0.01439242996275425, 'timing/model': 1.0873306049034, 'learning_rate/qwen_vl_interface': 7.034727237354442e-06, 'learning_rate/action_model': 7.034727237354441e-05, 'epoch': 1.04}) +[2026-06-17 22:14:38,337][starVLA.training.train_starvla][INFO] - Step 2705, Loss: {'train/loss': 0.5856015086174011, 'train/grad_norm_pre_clip': 11.0755615234375, 'timing/data': 0.011622775346040726, 'timing/model': 1.0894418749958277, 'learning_rate/qwen_vl_interface': 7.0327228257586244e-06, 'learning_rate/action_model': 7.032722825758625e-05, 'epoch': 1.04}) +[2026-06-17 22:14:43,177][starVLA.training.train_starvla][INFO] - Step 2706, Loss: {'train/loss': 0.3857097625732422, 'train/grad_norm_pre_clip': 8.20059871673584, 'timing/data': 0.01747590024024248, 'timing/model': 1.337762645445764, 'learning_rate/qwen_vl_interface': 7.030718044602451e-06, 'learning_rate/action_model': 7.03071804460245e-05, 'epoch': 1.04}) +[2026-06-17 22:14:47,995][starVLA.training.train_starvla][INFO] - Step 2707, Loss: {'train/loss': 0.48842737078666687, 'train/grad_norm_pre_clip': 7.625247955322266, 'timing/data': 0.01208254974335432, 'timing/model': 1.4636768568307161, 'learning_rate/qwen_vl_interface': 7.028712894301507e-06, 'learning_rate/action_model': 7.028712894301506e-05, 'epoch': 1.04}) +[2026-06-17 22:14:52,482][starVLA.training.train_starvla][INFO] - Step 2708, Loss: {'train/loss': 0.5565676093101501, 'train/grad_norm_pre_clip': 9.616377830505371, 'timing/data': 0.012252507731318474, 'timing/model': 1.1584600545465946, 'learning_rate/qwen_vl_interface': 7.026707375271469e-06, 'learning_rate/action_model': 7.026707375271468e-05, 'epoch': 1.04}) +[2026-06-17 22:14:57,758][starVLA.training.train_starvla][INFO] - Step 2709, Loss: {'train/loss': 0.534665584564209, 'train/grad_norm_pre_clip': 8.155921936035156, 'timing/data': 0.011818356812000275, 'timing/model': 1.121527386829257, 'learning_rate/qwen_vl_interface': 7.024701487928079e-06, 'learning_rate/action_model': 7.024701487928078e-05, 'epoch': 1.04}) +[2026-06-17 22:15:02,727][starVLA.training.train_starvla][INFO] - Step 2710, Loss: {'train/loss': 0.6428921222686768, 'train/grad_norm_pre_clip': 11.541608810424805, 'timing/data': 0.013929411768913269, 'timing/model': 1.3163244854658842, 'learning_rate/qwen_vl_interface': 7.02269523268716e-06, 'learning_rate/action_model': 7.022695232687159e-05, 'epoch': 1.04}) +[2026-06-17 22:15:07,903][starVLA.training.train_starvla][INFO] - Step 2711, Loss: {'train/loss': 0.5586956739425659, 'train/grad_norm_pre_clip': 10.674849510192871, 'timing/data': 0.015627512708306313, 'timing/model': 1.1262355018407106, 'learning_rate/qwen_vl_interface': 7.020688609964612e-06, 'learning_rate/action_model': 7.020688609964612e-05, 'epoch': 1.04}) +[2026-06-17 22:15:17,428][starVLA.training.train_starvla][INFO] - Step 2712, Loss: {'train/loss': 0.3208405673503876, 'train/grad_norm_pre_clip': 9.076672554016113, 'timing/data': 0.012192978523671627, 'timing/model': 1.4762168638408184, 'learning_rate/qwen_vl_interface': 7.018681620176412e-06, 'learning_rate/action_model': 7.018681620176412e-05, 'epoch': 1.04}) +[2026-06-17 22:15:22,422][starVLA.training.train_starvla][INFO] - Step 2713, Loss: {'train/loss': 0.4968613386154175, 'train/grad_norm_pre_clip': 7.8892998695373535, 'timing/data': 0.01719429437071085, 'timing/model': 1.2374537540599704, 'learning_rate/qwen_vl_interface': 7.016674263738606e-06, 'learning_rate/action_model': 7.016674263738607e-05, 'epoch': 1.04}) +[2026-06-17 22:15:29,198][starVLA.training.train_starvla][INFO] - Step 2714, Loss: {'train/loss': 0.24400709569454193, 'train/grad_norm_pre_clip': 4.761714458465576, 'timing/data': 0.012782735750079155, 'timing/model': 1.095468801446259, 'learning_rate/qwen_vl_interface': 7.014666541067325e-06, 'learning_rate/action_model': 7.014666541067325e-05, 'epoch': 1.04}) +[2026-06-17 22:15:34,520][starVLA.training.train_starvla][INFO] - Step 2715, Loss: {'train/loss': 0.6815904378890991, 'train/grad_norm_pre_clip': 10.968829154968262, 'timing/data': 0.01137438602745533, 'timing/model': 1.1525585800409317, 'learning_rate/qwen_vl_interface': 7.012658452578774e-06, 'learning_rate/action_model': 7.012658452578773e-05, 'epoch': 1.04}) +[2026-06-17 22:15:44,731][starVLA.training.train_starvla][INFO] - Step 2716, Loss: {'train/loss': 0.43974751234054565, 'train/grad_norm_pre_clip': 12.394858360290527, 'timing/data': 0.011810469441115856, 'timing/model': 1.3480936093255877, 'learning_rate/qwen_vl_interface': 7.010649998689225e-06, 'learning_rate/action_model': 7.010649998689224e-05, 'epoch': 1.04}) +[2026-06-17 22:15:51,806][starVLA.training.train_starvla][INFO] - Step 2717, Loss: {'train/loss': 0.5224501490592957, 'train/grad_norm_pre_clip': 12.722439765930176, 'timing/data': 0.01282177772372961, 'timing/model': 3.8510393537580967, 'learning_rate/qwen_vl_interface': 7.0086411798150385e-06, 'learning_rate/action_model': 7.008641179815038e-05, 'epoch': 1.05}) +[2026-06-17 22:16:04,991][starVLA.training.train_starvla][INFO] - Step 2718, Loss: {'train/loss': 0.7188605070114136, 'train/grad_norm_pre_clip': 10.117409706115723, 'timing/data': 0.011564554646611214, 'timing/model': 5.151984505355358, 'learning_rate/qwen_vl_interface': 7.006631996372642e-06, 'learning_rate/action_model': 7.006631996372643e-05, 'epoch': 1.05}) +[2026-06-17 22:16:17,440][starVLA.training.train_starvla][INFO] - Step 2719, Loss: {'train/loss': 0.42465606331825256, 'train/grad_norm_pre_clip': 8.517274856567383, 'timing/data': 0.013091601431369781, 'timing/model': 1.2325448086485267, 'learning_rate/qwen_vl_interface': 7.004622448778545e-06, 'learning_rate/action_model': 7.004622448778544e-05, 'epoch': 1.05}) +[2026-06-17 22:16:22,711][starVLA.training.train_starvla][INFO] - Step 2720, Loss: {'train/loss': 0.5876383781433105, 'train/grad_norm_pre_clip': 14.506485939025879, 'timing/data': 0.012676483020186424, 'timing/model': 1.0904891267418861, 'learning_rate/qwen_vl_interface': 7.0026125374493244e-06, 'learning_rate/action_model': 7.002612537449324e-05, 'epoch': 1.05}) +[2026-06-17 22:16:27,479][starVLA.training.train_starvla][INFO] - Step 2721, Loss: {'train/loss': 0.5562028884887695, 'train/grad_norm_pre_clip': 12.309809684753418, 'timing/data': 0.0013316599652171135, 'timing/model': 1.4362277779728174, 'learning_rate/qwen_vl_interface': 7.000602262801639e-06, 'learning_rate/action_model': 7.000602262801638e-05, 'epoch': 1.05}) +[2026-06-17 22:16:32,747][starVLA.training.train_starvla][INFO] - Step 2722, Loss: {'train/loss': 0.9058089852333069, 'train/grad_norm_pre_clip': 14.731053352355957, 'timing/data': 0.0011998899281024933, 'timing/model': 1.776130735874176, 'learning_rate/qwen_vl_interface': 6.9985916252522225e-06, 'learning_rate/action_model': 6.998591625252222e-05, 'epoch': 1.05}) +[2026-06-17 22:16:37,372][starVLA.training.train_starvla][INFO] - Step 2723, Loss: {'train/loss': 0.4969243109226227, 'train/grad_norm_pre_clip': 11.82155704498291, 'timing/data': 0.001448887400329113, 'timing/model': 1.1251023588702083, 'learning_rate/qwen_vl_interface': 6.9965806252178785e-06, 'learning_rate/action_model': 6.996580625217877e-05, 'epoch': 1.05}) +[2026-06-17 22:16:42,199][starVLA.training.train_starvla][INFO] - Step 2724, Loss: {'train/loss': 0.49786660075187683, 'train/grad_norm_pre_clip': 8.057941436767578, 'timing/data': 0.0011382149532437325, 'timing/model': 1.0839535240083933, 'learning_rate/qwen_vl_interface': 6.994569263115491e-06, 'learning_rate/action_model': 6.994569263115491e-05, 'epoch': 1.05}) +[2026-06-17 22:16:47,088][starVLA.training.train_starvla][INFO] - Step 2725, Loss: {'train/loss': 0.27700504660606384, 'train/grad_norm_pre_clip': 6.5778727531433105, 'timing/data': 0.0011372100561857224, 'timing/model': 1.0647066589444876, 'learning_rate/qwen_vl_interface': 6.992557539362019e-06, 'learning_rate/action_model': 6.99255753936202e-05, 'epoch': 1.05}) +[2026-06-17 22:16:55,717][starVLA.training.train_starvla][INFO] - Step 2726, Loss: {'train/loss': 0.4339502155780792, 'train/grad_norm_pre_clip': 8.229560852050781, 'timing/data': 0.0011026132851839066, 'timing/model': 1.102582967840135, 'learning_rate/qwen_vl_interface': 6.990545454374495e-06, 'learning_rate/action_model': 6.990545454374494e-05, 'epoch': 1.05}) +[2026-06-17 22:17:00,857][starVLA.training.train_starvla][INFO] - Step 2727, Loss: {'train/loss': 0.31370192766189575, 'train/grad_norm_pre_clip': 6.107093334197998, 'timing/data': 0.0016675777733325958, 'timing/model': 1.606124528683722, 'learning_rate/qwen_vl_interface': 6.988533008570024e-06, 'learning_rate/action_model': 6.988533008570024e-05, 'epoch': 1.05}) +[2026-06-17 22:17:05,589][starVLA.training.train_starvla][INFO] - Step 2728, Loss: {'train/loss': 0.6926025748252869, 'train/grad_norm_pre_clip': 9.787989616394043, 'timing/data': 0.0009047994390130043, 'timing/model': 1.2361875977367163, 'learning_rate/qwen_vl_interface': 6.98652020236579e-06, 'learning_rate/action_model': 6.986520202365789e-05, 'epoch': 1.05}) +[2026-06-17 22:17:12,278][starVLA.training.train_starvla][INFO] - Step 2729, Loss: {'train/loss': 0.6233052015304565, 'train/grad_norm_pre_clip': 13.125340461730957, 'timing/data': 0.0009176330640912056, 'timing/model': 2.766511824913323, 'learning_rate/qwen_vl_interface': 6.98450703617905e-06, 'learning_rate/action_model': 6.98450703617905e-05, 'epoch': 1.05}) +[2026-06-17 22:17:16,854][starVLA.training.train_starvla][INFO] - Step 2730, Loss: {'train/loss': 0.47661447525024414, 'train/grad_norm_pre_clip': 8.167563438415527, 'timing/data': 0.0013878075405955315, 'timing/model': 1.1781000336632133, 'learning_rate/qwen_vl_interface': 6.982493510427134e-06, 'learning_rate/action_model': 6.982493510427133e-05, 'epoch': 1.05}) +[2026-06-17 22:17:21,631][starVLA.training.train_starvla][INFO] - Step 2731, Loss: {'train/loss': 0.40779849886894226, 'train/grad_norm_pre_clip': 7.995419025421143, 'timing/data': 0.0022355830296874046, 'timing/model': 1.1435657879337668, 'learning_rate/qwen_vl_interface': 6.980479625527449e-06, 'learning_rate/action_model': 6.980479625527448e-05, 'epoch': 1.05}) +[2026-06-17 22:17:26,544][starVLA.training.train_starvla][INFO] - Step 2732, Loss: {'train/loss': 0.43411603569984436, 'train/grad_norm_pre_clip': 8.818453788757324, 'timing/data': 0.0009248564019799232, 'timing/model': 1.571450942195952, 'learning_rate/qwen_vl_interface': 6.978465381897477e-06, 'learning_rate/action_model': 6.978465381897477e-05, 'epoch': 1.05}) +[2026-06-17 22:17:38,213][starVLA.training.train_starvla][INFO] - Step 2733, Loss: {'train/loss': 0.282621830701828, 'train/grad_norm_pre_clip': 6.783302307128906, 'timing/data': 0.001310477964580059, 'timing/model': 1.4116996601223946, 'learning_rate/qwen_vl_interface': 6.97645077995477e-06, 'learning_rate/action_model': 6.97645077995477e-05, 'epoch': 1.05}) +[2026-06-17 22:17:43,476][starVLA.training.train_starvla][INFO] - Step 2734, Loss: {'train/loss': 0.34498727321624756, 'train/grad_norm_pre_clip': 9.230676651000977, 'timing/data': 0.0013405783101916313, 'timing/model': 1.4788601957261562, 'learning_rate/qwen_vl_interface': 6.97443582011696e-06, 'learning_rate/action_model': 6.974435820116959e-05, 'epoch': 1.05}) +[2026-06-17 22:17:48,549][starVLA.training.train_starvla][INFO] - Step 2735, Loss: {'train/loss': 0.6051895022392273, 'train/grad_norm_pre_clip': 12.600847244262695, 'timing/data': 0.001171208918094635, 'timing/model': 1.226804208010435, 'learning_rate/qwen_vl_interface': 6.972420502801748e-06, 'learning_rate/action_model': 6.972420502801747e-05, 'epoch': 1.05}) +[2026-06-17 22:17:53,822][starVLA.training.train_starvla][INFO] - Step 2736, Loss: {'train/loss': 0.48226577043533325, 'train/grad_norm_pre_clip': 8.018861770629883, 'timing/data': 0.0012858724221587181, 'timing/model': 1.427292611449957, 'learning_rate/qwen_vl_interface': 6.970404828426911e-06, 'learning_rate/action_model': 6.97040482842691e-05, 'epoch': 1.05}) +[2026-06-17 22:17:59,560][starVLA.training.train_starvla][INFO] - Step 2737, Loss: {'train/loss': 0.386379599571228, 'train/grad_norm_pre_clip': 10.216136932373047, 'timing/data': 0.001238599419593811, 'timing/model': 1.7927065081894398, 'learning_rate/qwen_vl_interface': 6.9683887974103044e-06, 'learning_rate/action_model': 6.968388797410305e-05, 'epoch': 1.05}) +[2026-06-17 22:18:11,329][starVLA.training.train_starvla][INFO] - Step 2738, Loss: {'train/loss': 0.46323058009147644, 'train/grad_norm_pre_clip': 10.315683364868164, 'timing/data': 0.0010529998689889908, 'timing/model': 1.3266067877411842, 'learning_rate/qwen_vl_interface': 6.966372410169848e-06, 'learning_rate/action_model': 6.966372410169848e-05, 'epoch': 1.05}) +[2026-06-17 22:18:16,119][starVLA.training.train_starvla][INFO] - Step 2739, Loss: {'train/loss': 0.5234965682029724, 'train/grad_norm_pre_clip': 8.38194751739502, 'timing/data': 0.0011050961911678314, 'timing/model': 1.0780187351629138, 'learning_rate/qwen_vl_interface': 6.964355667123548e-06, 'learning_rate/action_model': 6.964355667123547e-05, 'epoch': 1.05}) +[2026-06-17 22:18:21,232][starVLA.training.train_starvla][INFO] - Step 2740, Loss: {'train/loss': 0.4286941885948181, 'train/grad_norm_pre_clip': 7.146193027496338, 'timing/data': 0.0010639345273375511, 'timing/model': 1.0793564692139626, 'learning_rate/qwen_vl_interface': 6.96233856868947e-06, 'learning_rate/action_model': 6.962338568689469e-05, 'epoch': 1.05}) +[2026-06-17 22:18:26,715][starVLA.training.train_starvla][INFO] - Step 2741, Loss: {'train/loss': 0.683026909828186, 'train/grad_norm_pre_clip': 9.276055335998535, 'timing/data': 0.0009135724976658821, 'timing/model': 1.4825198156759143, 'learning_rate/qwen_vl_interface': 6.960321115285767e-06, 'learning_rate/action_model': 6.960321115285767e-05, 'epoch': 1.05}) +[2026-06-17 22:18:43,424][starVLA.training.train_starvla][INFO] - Step 2742, Loss: {'train/loss': 0.5905953645706177, 'train/grad_norm_pre_clip': 25.778884887695312, 'timing/data': 0.0012957742437720299, 'timing/model': 4.947054283693433, 'learning_rate/qwen_vl_interface': 6.958303307330655e-06, 'learning_rate/action_model': 6.958303307330655e-05, 'epoch': 1.05}) +[2026-06-17 22:18:48,294][starVLA.training.train_starvla][INFO] - Step 2743, Loss: {'train/loss': 0.6872500777244568, 'train/grad_norm_pre_clip': 10.46868896484375, 'timing/data': 0.0009710192680358887, 'timing/model': 1.2901096157729626, 'learning_rate/qwen_vl_interface': 6.956285145242429e-06, 'learning_rate/action_model': 6.956285145242428e-05, 'epoch': 1.06}) +[2026-06-17 22:18:54,493][starVLA.training.train_starvla][INFO] - Step 2744, Loss: {'train/loss': 0.3372552990913391, 'train/grad_norm_pre_clip': 6.1162848472595215, 'timing/data': 0.0011420929804444313, 'timing/model': 2.7221393659710884, 'learning_rate/qwen_vl_interface': 6.954266629439456e-06, 'learning_rate/action_model': 6.954266629439457e-05, 'epoch': 1.06}) +[2026-06-17 22:19:02,716][starVLA.training.train_starvla][INFO] - Step 2745, Loss: {'train/loss': 0.536373496055603, 'train/grad_norm_pre_clip': 7.756333827972412, 'timing/data': 0.0023644128814339638, 'timing/model': 1.2192175583913922, 'learning_rate/qwen_vl_interface': 6.952247760340176e-06, 'learning_rate/action_model': 6.952247760340176e-05, 'epoch': 1.06}) +[2026-06-17 22:19:08,031][starVLA.training.train_starvla][INFO] - Step 2746, Loss: {'train/loss': 0.5019528865814209, 'train/grad_norm_pre_clip': 8.918597221374512, 'timing/data': 0.001251116394996643, 'timing/model': 1.2299480717629194, 'learning_rate/qwen_vl_interface': 6.950228538363108e-06, 'learning_rate/action_model': 6.950228538363107e-05, 'epoch': 1.06}) +[2026-06-17 22:19:13,054][starVLA.training.train_starvla][INFO] - Step 2747, Loss: {'train/loss': 0.483925998210907, 'train/grad_norm_pre_clip': 9.550775527954102, 'timing/data': 0.0011034328490495682, 'timing/model': 1.5707744201645255, 'learning_rate/qwen_vl_interface': 6.948208963926829e-06, 'learning_rate/action_model': 6.94820896392683e-05, 'epoch': 1.06}) +[2026-06-17 22:19:17,508][starVLA.training.train_starvla][INFO] - Step 2748, Loss: {'train/loss': 0.4206019639968872, 'train/grad_norm_pre_clip': 7.0207839012146, 'timing/data': 0.00180909875780344, 'timing/model': 1.1079933270812035, 'learning_rate/qwen_vl_interface': 6.946189037450007e-06, 'learning_rate/action_model': 6.946189037450007e-05, 'epoch': 1.06}) +[2026-06-17 22:19:22,768][starVLA.training.train_starvla][INFO] - Step 2749, Loss: {'train/loss': 0.6026645302772522, 'train/grad_norm_pre_clip': 10.049162864685059, 'timing/data': 0.0012070583179593086, 'timing/model': 1.104402132332325, 'learning_rate/qwen_vl_interface': 6.9441687593513715e-06, 'learning_rate/action_model': 6.944168759351372e-05, 'epoch': 1.06}) +[2026-06-17 22:19:27,332][starVLA.training.train_starvla][INFO] - Step 2750, Loss: {'train/loss': 0.48270928859710693, 'train/grad_norm_pre_clip': 10.090023040771484, 'timing/data': 0.0008483733981847763, 'timing/model': 1.0744628654792905, 'learning_rate/qwen_vl_interface': 6.942148130049732e-06, 'learning_rate/action_model': 6.942148130049731e-05, 'epoch': 1.06}) +[2026-06-17 22:19:32,194][starVLA.training.train_starvla][INFO] - Step 2751, Loss: {'train/loss': 0.5459117293357849, 'train/grad_norm_pre_clip': 9.403030395507812, 'timing/data': 0.0011484846472740173, 'timing/model': 1.0836989125236869, 'learning_rate/qwen_vl_interface': 6.940127149963961e-06, 'learning_rate/action_model': 6.94012714996396e-05, 'epoch': 1.06}) +[2026-06-17 22:19:36,996][starVLA.training.train_starvla][INFO] - Step 2752, Loss: {'train/loss': 0.5227871537208557, 'train/grad_norm_pre_clip': 8.654702186584473, 'timing/data': 0.0010009026154875755, 'timing/model': 1.6167521048337221, 'learning_rate/qwen_vl_interface': 6.9381058195130146e-06, 'learning_rate/action_model': 6.938105819513014e-05, 'epoch': 1.06}) +[2026-06-17 22:19:41,421][starVLA.training.train_starvla][INFO] - Step 2753, Loss: {'train/loss': 0.3633801341056824, 'train/grad_norm_pre_clip': 6.214017391204834, 'timing/data': 0.0009346595034003258, 'timing/model': 1.0808285400271416, 'learning_rate/qwen_vl_interface': 6.936084139115915e-06, 'learning_rate/action_model': 6.936084139115915e-05, 'epoch': 1.06}) +[2026-06-17 22:19:46,481][starVLA.training.train_starvla][INFO] - Step 2754, Loss: {'train/loss': 0.5741464495658875, 'train/grad_norm_pre_clip': 27.43985939025879, 'timing/data': 0.0011848602443933487, 'timing/model': 1.1131675355136395, 'learning_rate/qwen_vl_interface': 6.934062109191759e-06, 'learning_rate/action_model': 6.934062109191759e-05, 'epoch': 1.06}) +[2026-06-17 22:19:51,217][starVLA.training.train_starvla][INFO] - Step 2755, Loss: {'train/loss': 0.4571293592453003, 'train/grad_norm_pre_clip': 7.971060752868652, 'timing/data': 0.0012044981122016907, 'timing/model': 1.2471468411386013, 'learning_rate/qwen_vl_interface': 6.9320397301597155e-06, 'learning_rate/action_model': 6.932039730159715e-05, 'epoch': 1.06}) +[2026-06-17 22:19:57,693][starVLA.training.train_starvla][INFO] - Step 2756, Loss: {'train/loss': 0.7121626734733582, 'train/grad_norm_pre_clip': 13.87289810180664, 'timing/data': 0.0010987771674990654, 'timing/model': 2.7468027798458934, 'learning_rate/qwen_vl_interface': 6.930017002439027e-06, 'learning_rate/action_model': 6.930017002439027e-05, 'epoch': 1.06}) +[2026-06-17 22:20:02,293][starVLA.training.train_starvla][INFO] - Step 2757, Loss: {'train/loss': 0.8246733546257019, 'train/grad_norm_pre_clip': 11.333909034729004, 'timing/data': 0.0014364654198288918, 'timing/model': 1.4104751218110323, 'learning_rate/qwen_vl_interface': 6.927993926449005e-06, 'learning_rate/action_model': 6.927993926449005e-05, 'epoch': 1.06}) +[2026-06-17 22:20:08,335][starVLA.training.train_starvla][INFO] - Step 2758, Loss: {'train/loss': 0.3633359372615814, 'train/grad_norm_pre_clip': 14.790599822998047, 'timing/data': 0.0009219804778695107, 'timing/model': 2.75978297367692, 'learning_rate/qwen_vl_interface': 6.925970502609036e-06, 'learning_rate/action_model': 6.925970502609036e-05, 'epoch': 1.06}) +[2026-06-17 22:20:13,072][starVLA.training.train_starvla][INFO] - Step 2759, Loss: {'train/loss': 0.6996234655380249, 'train/grad_norm_pre_clip': 10.660737037658691, 'timing/data': 0.001211589202284813, 'timing/model': 1.1350230108946562, 'learning_rate/qwen_vl_interface': 6.92394673133858e-06, 'learning_rate/action_model': 6.92394673133858e-05, 'epoch': 1.06}) +[2026-06-17 22:20:17,848][starVLA.training.train_starvla][INFO] - Step 2760, Loss: {'train/loss': 0.554469108581543, 'train/grad_norm_pre_clip': 8.408674240112305, 'timing/data': 0.001231381669640541, 'timing/model': 1.2487929444760084, 'learning_rate/qwen_vl_interface': 6.921922613057165e-06, 'learning_rate/action_model': 6.921922613057164e-05, 'epoch': 1.06}) +[2026-06-17 22:20:26,413][starVLA.training.train_starvla][INFO] - Step 2761, Loss: {'train/loss': 0.5215047597885132, 'train/grad_norm_pre_clip': 6.692943096160889, 'timing/data': 0.0013306532055139542, 'timing/model': 4.623315006494522, 'learning_rate/qwen_vl_interface': 6.9198981481843915e-06, 'learning_rate/action_model': 6.919898148184392e-05, 'epoch': 1.06}) +[2026-06-17 22:20:31,599][starVLA.training.train_starvla][INFO] - Step 2762, Loss: {'train/loss': 0.5598410964012146, 'train/grad_norm_pre_clip': 8.494970321655273, 'timing/data': 0.0011094938963651657, 'timing/model': 1.446332149207592, 'learning_rate/qwen_vl_interface': 6.917873337139936e-06, 'learning_rate/action_model': 6.917873337139937e-05, 'epoch': 1.06}) +[2026-06-17 22:20:40,032][starVLA.training.train_starvla][INFO] - Step 2763, Loss: {'train/loss': 0.6668445467948914, 'train/grad_norm_pre_clip': 14.124502182006836, 'timing/data': 0.000803828239440918, 'timing/model': 1.3032360495999455, 'learning_rate/qwen_vl_interface': 6.915848180343545e-06, 'learning_rate/action_model': 6.915848180343545e-05, 'epoch': 1.06}) +[2026-06-17 22:20:44,903][starVLA.training.train_starvla][INFO] - Step 2764, Loss: {'train/loss': 0.4671940207481384, 'train/grad_norm_pre_clip': 8.113919258117676, 'timing/data': 0.0011515039950609207, 'timing/model': 1.3489228682592511, 'learning_rate/qwen_vl_interface': 6.913822678215031e-06, 'learning_rate/action_model': 6.913822678215031e-05, 'epoch': 1.06}) +[2026-06-17 22:20:49,436][starVLA.training.train_starvla][INFO] - Step 2765, Loss: {'train/loss': 0.4603728652000427, 'train/grad_norm_pre_clip': 7.925034046173096, 'timing/data': 0.00122116319835186, 'timing/model': 1.0782634755596519, 'learning_rate/qwen_vl_interface': 6.911796831174287e-06, 'learning_rate/action_model': 6.911796831174286e-05, 'epoch': 1.06}) +[2026-06-17 22:20:54,579][starVLA.training.train_starvla][INFO] - Step 2766, Loss: {'train/loss': 0.3917606770992279, 'train/grad_norm_pre_clip': 6.608345031738281, 'timing/data': 0.0011611822992563248, 'timing/model': 1.2047850145027041, 'learning_rate/qwen_vl_interface': 6.909770639641271e-06, 'learning_rate/action_model': 6.909770639641271e-05, 'epoch': 1.06}) +[2026-06-17 22:20:59,361][starVLA.training.train_starvla][INFO] - Step 2767, Loss: {'train/loss': 0.6863272786140442, 'train/grad_norm_pre_clip': 9.016707420349121, 'timing/data': 0.0012262724339962006, 'timing/model': 1.4938712837174535, 'learning_rate/qwen_vl_interface': 6.907744104036015e-06, 'learning_rate/action_model': 6.907744104036015e-05, 'epoch': 1.06}) +[2026-06-17 22:21:04,118][starVLA.training.train_starvla][INFO] - Step 2768, Loss: {'train/loss': 0.5743536949157715, 'train/grad_norm_pre_clip': 10.722416877746582, 'timing/data': 0.0010188044980168343, 'timing/model': 1.246536921709776, 'learning_rate/qwen_vl_interface': 6.905717224778623e-06, 'learning_rate/action_model': 6.905717224778622e-05, 'epoch': 1.06}) +[2026-06-17 22:21:09,144][starVLA.training.train_starvla][INFO] - Step 2769, Loss: {'train/loss': 0.3508288562297821, 'train/grad_norm_pre_clip': 6.619131565093994, 'timing/data': 0.0011311378329992294, 'timing/model': 1.1298071537166834, 'learning_rate/qwen_vl_interface': 6.903690002289268e-06, 'learning_rate/action_model': 6.903690002289267e-05, 'epoch': 1.07}) +[2026-06-17 22:21:17,468][starVLA.training.train_starvla][INFO] - Step 2770, Loss: {'train/loss': 0.44482579827308655, 'train/grad_norm_pre_clip': 9.130030632019043, 'timing/data': 0.001243390142917633, 'timing/model': 1.3609528131783009, 'learning_rate/qwen_vl_interface': 6.901662436988199e-06, 'learning_rate/action_model': 6.901662436988199e-05, 'epoch': 1.07}) +[2026-06-17 22:21:22,588][starVLA.training.train_starvla][INFO] - Step 2771, Loss: {'train/loss': 0.5900173187255859, 'train/grad_norm_pre_clip': 12.938949584960938, 'timing/data': 0.001707250252366066, 'timing/model': 1.1570706842467189, 'learning_rate/qwen_vl_interface': 6.899634529295726e-06, 'learning_rate/action_model': 6.899634529295726e-05, 'epoch': 1.07}) +[2026-06-17 22:21:27,753][starVLA.training.train_starvla][INFO] - Step 2772, Loss: {'train/loss': 0.6739691495895386, 'train/grad_norm_pre_clip': 10.392468452453613, 'timing/data': 0.0017520906403660774, 'timing/model': 1.6472638957202435, 'learning_rate/qwen_vl_interface': 6.897606279632241e-06, 'learning_rate/action_model': 6.897606279632241e-05, 'epoch': 1.07}) +[2026-06-17 22:21:36,745][starVLA.training.train_starvla][INFO] - Step 2773, Loss: {'train/loss': 0.6460043787956238, 'train/grad_norm_pre_clip': 12.075719833374023, 'timing/data': 0.0012104455381631851, 'timing/model': 1.2642071601003408, 'learning_rate/qwen_vl_interface': 6.895577688418202e-06, 'learning_rate/action_model': 6.895577688418202e-05, 'epoch': 1.07}) +[2026-06-17 22:21:49,175][starVLA.training.train_starvla][INFO] - Step 2774, Loss: {'train/loss': 0.6249944567680359, 'train/grad_norm_pre_clip': 12.398366928100586, 'timing/data': 0.0015259897336363792, 'timing/model': 1.292666700668633, 'learning_rate/qwen_vl_interface': 6.893548756074136e-06, 'learning_rate/action_model': 6.893548756074136e-05, 'epoch': 1.07}) +[2026-06-17 22:21:54,272][starVLA.training.train_starvla][INFO] - Step 2775, Loss: {'train/loss': 0.5076296925544739, 'train/grad_norm_pre_clip': 8.7503662109375, 'timing/data': 0.001367153599858284, 'timing/model': 1.3836910063400865, 'learning_rate/qwen_vl_interface': 6.891519483020647e-06, 'learning_rate/action_model': 6.891519483020647e-05, 'epoch': 1.07}) +[2026-06-17 22:21:59,293][starVLA.training.train_starvla][INFO] - Step 2776, Loss: {'train/loss': 0.47459694743156433, 'train/grad_norm_pre_clip': 9.240015029907227, 'timing/data': 0.0014884043484926224, 'timing/model': 1.0887142671272159, 'learning_rate/qwen_vl_interface': 6.8894898696784e-06, 'learning_rate/action_model': 6.889489869678399e-05, 'epoch': 1.07}) +[2026-06-17 22:22:04,346][starVLA.training.train_starvla][INFO] - Step 2777, Loss: {'train/loss': 0.36944466829299927, 'train/grad_norm_pre_clip': 7.528585433959961, 'timing/data': 0.0009119315072894096, 'timing/model': 1.4725278411060572, 'learning_rate/qwen_vl_interface': 6.887459916468141e-06, 'learning_rate/action_model': 6.88745991646814e-05, 'epoch': 1.07}) +[2026-06-17 22:22:18,694][starVLA.training.train_starvla][INFO] - Step 2778, Loss: {'train/loss': 0.3557738959789276, 'train/grad_norm_pre_clip': 6.462818145751953, 'timing/data': 0.0011709872633218765, 'timing/model': 1.3295046556741, 'learning_rate/qwen_vl_interface': 6.885429623810678e-06, 'learning_rate/action_model': 6.885429623810677e-05, 'epoch': 1.07}) +[2026-06-17 22:22:23,639][starVLA.training.train_starvla][INFO] - Step 2779, Loss: {'train/loss': 0.43509042263031006, 'train/grad_norm_pre_clip': 7.181778907775879, 'timing/data': 0.0012456215918064117, 'timing/model': 1.0866835061460733, 'learning_rate/qwen_vl_interface': 6.883398992126894e-06, 'learning_rate/action_model': 6.883398992126894e-05, 'epoch': 1.07}) +[2026-06-17 22:22:28,581][starVLA.training.train_starvla][INFO] - Step 2780, Loss: {'train/loss': 0.38224077224731445, 'train/grad_norm_pre_clip': 9.671457290649414, 'timing/data': 0.0012970389798283577, 'timing/model': 1.3019688287749887, 'learning_rate/qwen_vl_interface': 6.8813680218377425e-06, 'learning_rate/action_model': 6.881368021837741e-05, 'epoch': 1.07}) +[2026-06-17 22:22:42,057][starVLA.training.train_starvla][INFO] - Step 2781, Loss: {'train/loss': 0.30345284938812256, 'train/grad_norm_pre_clip': 6.967467784881592, 'timing/data': 0.002004767768085003, 'timing/model': 5.6465524500235915, 'learning_rate/qwen_vl_interface': 6.8793367133642445e-06, 'learning_rate/action_model': 6.879336713364243e-05, 'epoch': 1.07}) +[2026-06-17 22:22:47,072][starVLA.training.train_starvla][INFO] - Step 2782, Loss: {'train/loss': 0.6833816766738892, 'train/grad_norm_pre_clip': 8.775193214416504, 'timing/data': 0.0009166896343231201, 'timing/model': 1.4686615504324436, 'learning_rate/qwen_vl_interface': 6.877305067127492e-06, 'learning_rate/action_model': 6.877305067127492e-05, 'epoch': 1.07}) +[2026-06-17 22:22:51,376][starVLA.training.train_starvla][INFO] - Step 2783, Loss: {'train/loss': 0.9045602083206177, 'train/grad_norm_pre_clip': 13.314065933227539, 'timing/data': 0.0011554434895515442, 'timing/model': 1.1150243347510695, 'learning_rate/qwen_vl_interface': 6.875273083548648e-06, 'learning_rate/action_model': 6.875273083548648e-05, 'epoch': 1.07}) +[2026-06-17 22:22:56,427][starVLA.training.train_starvla][INFO] - Step 2784, Loss: {'train/loss': 0.5163760781288147, 'train/grad_norm_pre_clip': 7.409963607788086, 'timing/data': 0.0012008938938379288, 'timing/model': 1.1592999519780278, 'learning_rate/qwen_vl_interface': 6.873240763048945e-06, 'learning_rate/action_model': 6.873240763048945e-05, 'epoch': 1.07}) +[2026-06-17 22:23:01,951][starVLA.training.train_starvla][INFO] - Step 2785, Loss: {'train/loss': 0.3538535237312317, 'train/grad_norm_pre_clip': 5.942999362945557, 'timing/data': 0.0012873578816652298, 'timing/model': 1.1279936600476503, 'learning_rate/qwen_vl_interface': 6.871208106049685e-06, 'learning_rate/action_model': 6.871208106049685e-05, 'epoch': 1.07}) +[2026-06-17 22:23:12,486][starVLA.training.train_starvla][INFO] - Step 2786, Loss: {'train/loss': 0.5689687132835388, 'train/grad_norm_pre_clip': 9.427967071533203, 'timing/data': 0.0010944893583655357, 'timing/model': 1.1293810475617647, 'learning_rate/qwen_vl_interface': 6.869175112972239e-06, 'learning_rate/action_model': 6.86917511297224e-05, 'epoch': 1.07}) +[2026-06-17 22:23:17,374][starVLA.training.train_starvla][INFO] - Step 2787, Loss: {'train/loss': 0.3683934807777405, 'train/grad_norm_pre_clip': 9.188990592956543, 'timing/data': 0.0032593710348010063, 'timing/model': 1.4783902447670698, 'learning_rate/qwen_vl_interface': 6.867141784238052e-06, 'learning_rate/action_model': 6.867141784238051e-05, 'epoch': 1.07}) +[2026-06-17 22:23:28,246][starVLA.training.train_starvla][INFO] - Step 2788, Loss: {'train/loss': 0.5440230965614319, 'train/grad_norm_pre_clip': 12.113350868225098, 'timing/data': 0.0009432593360543251, 'timing/model': 1.3911795215681195, 'learning_rate/qwen_vl_interface': 6.865108120268632e-06, 'learning_rate/action_model': 6.865108120268632e-05, 'epoch': 1.07}) +[2026-06-17 22:23:33,145][starVLA.training.train_starvla][INFO] - Step 2789, Loss: {'train/loss': 0.4891771376132965, 'train/grad_norm_pre_clip': 9.177752494812012, 'timing/data': 0.0012938184663653374, 'timing/model': 1.0785710597410798, 'learning_rate/qwen_vl_interface': 6.863074121485561e-06, 'learning_rate/action_model': 6.86307412148556e-05, 'epoch': 1.07}) +[2026-06-17 22:23:38,308][starVLA.training.train_starvla][INFO] - Step 2790, Loss: {'train/loss': 0.408526211977005, 'train/grad_norm_pre_clip': 10.50594425201416, 'timing/data': 0.0012912824749946594, 'timing/model': 1.3372176103293896, 'learning_rate/qwen_vl_interface': 6.861039788310489e-06, 'learning_rate/action_model': 6.861039788310489e-05, 'epoch': 1.07}) +[2026-06-17 22:23:43,431][starVLA.training.train_starvla][INFO] - Step 2791, Loss: {'train/loss': 0.41489315032958984, 'train/grad_norm_pre_clip': 6.735101222991943, 'timing/data': 0.001136726699769497, 'timing/model': 1.5333845475688577, 'learning_rate/qwen_vl_interface': 6.859005121165135e-06, 'learning_rate/action_model': 6.859005121165135e-05, 'epoch': 1.07}) +[2026-06-17 22:23:47,880][starVLA.training.train_starvla][INFO] - Step 2792, Loss: {'train/loss': 0.5634368658065796, 'train/grad_norm_pre_clip': 9.348379135131836, 'timing/data': 0.001318659633398056, 'timing/model': 1.138487628661096, 'learning_rate/qwen_vl_interface': 6.856970120471289e-06, 'learning_rate/action_model': 6.856970120471289e-05, 'epoch': 1.07}) +[2026-06-17 22:23:52,726][starVLA.training.train_starvla][INFO] - Step 2793, Loss: {'train/loss': 0.2801138460636139, 'train/grad_norm_pre_clip': 5.224998950958252, 'timing/data': 0.0009042331948876381, 'timing/model': 1.1215666001662612, 'learning_rate/qwen_vl_interface': 6.854934786650805e-06, 'learning_rate/action_model': 6.854934786650806e-05, 'epoch': 1.07}) +[2026-06-17 22:24:00,653][starVLA.training.train_starvla][INFO] - Step 2794, Loss: {'train/loss': 0.5014766454696655, 'train/grad_norm_pre_clip': 8.908348083496094, 'timing/data': 0.0013235248625278473, 'timing/model': 4.2007458955049515, 'learning_rate/qwen_vl_interface': 6.852899120125615e-06, 'learning_rate/action_model': 6.852899120125616e-05, 'epoch': 1.07}) +[2026-06-17 22:24:06,131][starVLA.training.train_starvla][INFO] - Step 2795, Loss: {'train/loss': 0.6381166577339172, 'train/grad_norm_pre_clip': 10.892485618591309, 'timing/data': 0.0013636937364935875, 'timing/model': 1.1064627235755324, 'learning_rate/qwen_vl_interface': 6.850863121317712e-06, 'learning_rate/action_model': 6.850863121317712e-05, 'epoch': 1.08}) +[2026-06-17 22:24:13,980][starVLA.training.train_starvla][INFO] - Step 2796, Loss: {'train/loss': 0.4673965275287628, 'train/grad_norm_pre_clip': 9.106383323669434, 'timing/data': 0.0011623445898294449, 'timing/model': 1.3132819375023246, 'learning_rate/qwen_vl_interface': 6.848826790649162e-06, 'learning_rate/action_model': 6.848826790649163e-05, 'epoch': 1.08}) +[2026-06-17 22:24:19,028][starVLA.training.train_starvla][INFO] - Step 2797, Loss: {'train/loss': 0.733363687992096, 'train/grad_norm_pre_clip': 9.011537551879883, 'timing/data': 0.0011660102754831314, 'timing/model': 1.4718017131090164, 'learning_rate/qwen_vl_interface': 6.846790128542098e-06, 'learning_rate/action_model': 6.846790128542098e-05, 'epoch': 1.08}) +[2026-06-17 22:24:24,081][starVLA.training.train_starvla][INFO] - Step 2798, Loss: {'train/loss': 0.7868739366531372, 'train/grad_norm_pre_clip': 12.89750862121582, 'timing/data': 0.0015470115467905998, 'timing/model': 1.2788752876222134, 'learning_rate/qwen_vl_interface': 6.8447531354187214e-06, 'learning_rate/action_model': 6.844753135418722e-05, 'epoch': 1.08}) +[2026-06-17 22:24:29,284][starVLA.training.train_starvla][INFO] - Step 2799, Loss: {'train/loss': 0.6811909675598145, 'train/grad_norm_pre_clip': 8.32028579711914, 'timing/data': 0.001183181069791317, 'timing/model': 1.1204235414043069, 'learning_rate/qwen_vl_interface': 6.842715811701305e-06, 'learning_rate/action_model': 6.842715811701304e-05, 'epoch': 1.08}) +[2026-06-17 22:27:49,942][starVLA.training.train_starvla][INFO] - Step 2800, Loss: {'train/loss': 0.34063005447387695, 'train/grad_norm_pre_clip': 6.655013561248779, 'eval/loss': 0.5041114091873169, 'eval/latency_4/loss': 0.5041114091873169, 'eval/demon_fire_precision': 0.9669554058389477, 'eval/demon_fire_recall': 0.9788892497564144, 'eval/demon_fire_f1': 0.972885732730794, 'eval/demon_fire_rate_model': 0.9154185022026432, 'eval/demon_fire_rate_teacher': 0.9042584434654919, 'eval/demon_left_precision': 0.7524038461538461, 'eval/demon_left_recall': 0.7146118721461188, 'eval/demon_left_f1': 0.7330210772833725, 'eval/demon_left_rate_model': 0.12217327459618209, 'eval/demon_left_rate_teacher': 0.12863436123348018, 'eval/demon_right_precision': 0.8169934640522876, 'eval/demon_right_recall': 0.6768953068592057, 'eval/demon_right_f1': 0.7403751233958538, 'eval/demon_right_rate_model': 0.13480176211453745, 'eval/demon_right_rate_teacher': 0.16270190895741557, 'eval/demon_move_f1': 0.7366981003396131, 'eval/demon_cc_f1': 0.8875229705577216, 'eval/latency_4/demon_fire_precision': 0.9669554058389477, 'eval/latency_4/demon_fire_recall': 0.9788892497564144, 'eval/latency_4/demon_fire_f1': 0.972885732730794, 'eval/latency_4/demon_fire_rate_model': 0.9154185022026432, 'eval/latency_4/demon_fire_rate_teacher': 0.9042584434654919, 'eval/latency_4/demon_left_precision': 0.7524038461538461, 'eval/latency_4/demon_left_recall': 0.7146118721461188, 'eval/latency_4/demon_left_f1': 0.7330210772833725, 'eval/latency_4/demon_left_rate_model': 0.12217327459618209, 'eval/latency_4/demon_left_rate_teacher': 0.12863436123348018, 'eval/latency_4/demon_right_precision': 0.8169934640522876, 'eval/latency_4/demon_right_recall': 0.6768953068592057, 'eval/latency_4/demon_right_f1': 0.7403751233958538, 'eval/latency_4/demon_right_rate_model': 0.13480176211453745, 'eval/latency_4/demon_right_rate_teacher': 0.16270190895741557, 'eval/latency_4/demon_move_f1': 0.7366981003396131, 'eval/latency_4/demon_cc_f1': 0.8875229705577216, 'timing/data': 0.0011108284816145897, 'timing/model': 1.3088848246261477, 'learning_rate/qwen_vl_interface': 6.840678157812184e-06, 'learning_rate/action_model': 6.840678157812184e-05, 'epoch': 1.08}) +[2026-06-17 22:27:49,943][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_2800_state +[2026-06-17 22:27:49,960][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 22:27:58,469][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_2800_state/model.safetensors +[2026-06-17 22:28:21,189][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_2800_state/optimizer.bin +[2026-06-17 22:28:21,191][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_2800_state/sampler.bin +[2026-06-17 22:28:21,191][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_2800_state/sampler_1.bin +[2026-06-17 22:28:21,195][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_2800_state/random_states_0.pkl +[2026-06-17 22:28:21,205][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2800 +[2026-06-17 22:28:21,205][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 22:28:21,224][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 22:28:33,609][starVLA.training.train_starvla][INFO] - Step 2801, Loss: {'train/loss': 0.6134293675422668, 'train/grad_norm_pre_clip': 9.263605117797852, 'timing/data': 0.0012395652011036873, 'timing/model': 4.822014535777271, 'learning_rate/qwen_vl_interface': 6.8386401741737725e-06, 'learning_rate/action_model': 6.838640174173773e-05, 'epoch': 1.08}) +[2026-06-17 22:28:40,813][starVLA.training.train_starvla][INFO] - Step 2802, Loss: {'train/loss': 0.6839306950569153, 'train/grad_norm_pre_clip': 10.313276290893555, 'timing/data': 0.0013874536380171776, 'timing/model': 1.6242993511259556, 'learning_rate/qwen_vl_interface': 6.836601861208542e-06, 'learning_rate/action_model': 6.836601861208541e-05, 'epoch': 1.08}) +[2026-06-17 22:28:48,307][starVLA.training.train_starvla][INFO] - Step 2803, Loss: {'train/loss': 0.4206884503364563, 'train/grad_norm_pre_clip': 11.925942420959473, 'timing/data': 0.0011297333985567093, 'timing/model': 1.1541935252025723, 'learning_rate/qwen_vl_interface': 6.834563219339039e-06, 'learning_rate/action_model': 6.834563219339038e-05, 'epoch': 1.08}) +[2026-06-17 22:28:59,161][starVLA.training.train_starvla][INFO] - Step 2804, Loss: {'train/loss': 0.5097195506095886, 'train/grad_norm_pre_clip': 8.8475923538208, 'timing/data': 0.0009573902934789658, 'timing/model': 2.3383373375982046, 'learning_rate/qwen_vl_interface': 6.832524248987874e-06, 'learning_rate/action_model': 6.832524248987874e-05, 'epoch': 1.08}) +[2026-06-17 22:29:12,126][starVLA.training.train_starvla][INFO] - Step 2805, Loss: {'train/loss': 0.359573632478714, 'train/grad_norm_pre_clip': 9.68326187133789, 'timing/data': 0.001498684287071228, 'timing/model': 5.446792362257838, 'learning_rate/qwen_vl_interface': 6.8304849505777314e-06, 'learning_rate/action_model': 6.830484950577732e-05, 'epoch': 1.08}) +[2026-06-17 22:29:18,049][starVLA.training.train_starvla][INFO] - Step 2806, Loss: {'train/loss': 0.8793834447860718, 'train/grad_norm_pre_clip': 11.387660026550293, 'timing/data': 0.0012181466445326805, 'timing/model': 1.3229564437642694, 'learning_rate/qwen_vl_interface': 6.828445324531357e-06, 'learning_rate/action_model': 6.828445324531357e-05, 'epoch': 1.08}) +[2026-06-17 22:29:29,308][starVLA.training.train_starvla][INFO] - Step 2807, Loss: {'train/loss': 0.46796128153800964, 'train/grad_norm_pre_clip': 6.855042457580566, 'timing/data': 0.0010367780923843384, 'timing/model': 1.6788330553099513, 'learning_rate/qwen_vl_interface': 6.826405371271567e-06, 'learning_rate/action_model': 6.826405371271567e-05, 'epoch': 1.08}) +[2026-06-17 22:29:39,512][starVLA.training.train_starvla][INFO] - Step 2808, Loss: {'train/loss': 0.42992115020751953, 'train/grad_norm_pre_clip': 8.512240409851074, 'timing/data': 0.001324547454714775, 'timing/model': 2.5054721450433135, 'learning_rate/qwen_vl_interface': 6.82436509122125e-06, 'learning_rate/action_model': 6.82436509122125e-05, 'epoch': 1.08}) +[2026-06-17 22:29:51,722][starVLA.training.train_starvla][INFO] - Step 2809, Loss: {'train/loss': 0.42188119888305664, 'train/grad_norm_pre_clip': 10.711268424987793, 'timing/data': 0.0022338367998600006, 'timing/model': 1.090663556009531, 'learning_rate/qwen_vl_interface': 6.822324484803353e-06, 'learning_rate/action_model': 6.822324484803353e-05, 'epoch': 1.08}) +[2026-06-17 22:29:56,666][starVLA.training.train_starvla][INFO] - Step 2810, Loss: {'train/loss': 0.35591650009155273, 'train/grad_norm_pre_clip': 8.877266883850098, 'timing/data': 0.0011112866923213005, 'timing/model': 1.156981742940843, 'learning_rate/qwen_vl_interface': 6.8202835524409e-06, 'learning_rate/action_model': 6.8202835524409e-05, 'epoch': 1.08}) +[2026-06-17 22:30:01,872][starVLA.training.train_starvla][INFO] - Step 2811, Loss: {'train/loss': 0.4512885510921478, 'train/grad_norm_pre_clip': 11.763711929321289, 'timing/data': 0.0012003406882286072, 'timing/model': 1.1906833741813898, 'learning_rate/qwen_vl_interface': 6.818242294556977e-06, 'learning_rate/action_model': 6.818242294556977e-05, 'epoch': 1.08}) +[2026-06-17 22:30:06,973][starVLA.training.train_starvla][INFO] - Step 2812, Loss: {'train/loss': 0.3540100157260895, 'train/grad_norm_pre_clip': 7.626098155975342, 'timing/data': 0.0013791611418128014, 'timing/model': 1.6329954378306866, 'learning_rate/qwen_vl_interface': 6.816200711574738e-06, 'learning_rate/action_model': 6.816200711574738e-05, 'epoch': 1.08}) +[2026-06-17 22:30:11,859][starVLA.training.train_starvla][INFO] - Step 2813, Loss: {'train/loss': 0.5803514122962952, 'train/grad_norm_pre_clip': 10.33847427368164, 'timing/data': 0.0013582902029156685, 'timing/model': 1.268805094063282, 'learning_rate/qwen_vl_interface': 6.814158803917406e-06, 'learning_rate/action_model': 6.814158803917406e-05, 'epoch': 1.08}) +[2026-06-17 22:30:16,641][starVLA.training.train_starvla][INFO] - Step 2814, Loss: {'train/loss': 0.4699920415878296, 'train/grad_norm_pre_clip': 9.035700798034668, 'timing/data': 0.001202213577926159, 'timing/model': 1.0974002527073026, 'learning_rate/qwen_vl_interface': 6.812116572008271e-06, 'learning_rate/action_model': 6.812116572008271e-05, 'epoch': 1.08}) +[2026-06-17 22:30:21,442][starVLA.training.train_starvla][INFO] - Step 2815, Loss: {'train/loss': 0.5179919004440308, 'train/grad_norm_pre_clip': 9.62524127960205, 'timing/data': 0.0011876411736011505, 'timing/model': 1.0687450487166643, 'learning_rate/qwen_vl_interface': 6.8100740162706906e-06, 'learning_rate/action_model': 6.81007401627069e-05, 'epoch': 1.08}) +[2026-06-17 22:30:26,576][starVLA.training.train_starvla][INFO] - Step 2816, Loss: {'train/loss': 0.34476152062416077, 'train/grad_norm_pre_clip': 5.999329090118408, 'timing/data': 0.0012068245559930801, 'timing/model': 1.141600590199232, 'learning_rate/qwen_vl_interface': 6.808031137128087e-06, 'learning_rate/action_model': 6.808031137128086e-05, 'epoch': 1.08}) +[2026-06-17 22:30:35,708][starVLA.training.train_starvla][INFO] - Step 2817, Loss: {'train/loss': 0.3752005994319916, 'train/grad_norm_pre_clip': 8.652661323547363, 'timing/data': 0.0011622551828622818, 'timing/model': 5.558421680703759, 'learning_rate/qwen_vl_interface': 6.805987935003951e-06, 'learning_rate/action_model': 6.805987935003951e-05, 'epoch': 1.08}) +[2026-06-17 22:30:40,876][starVLA.training.train_starvla][INFO] - Step 2818, Loss: {'train/loss': 0.4097083508968353, 'train/grad_norm_pre_clip': 9.308540344238281, 'timing/data': 0.000913526862859726, 'timing/model': 1.1065565459430218, 'learning_rate/qwen_vl_interface': 6.803944410321843e-06, 'learning_rate/action_model': 6.803944410321844e-05, 'epoch': 1.08}) +[2026-06-17 22:30:45,758][starVLA.training.train_starvla][INFO] - Step 2819, Loss: {'train/loss': 0.31235501170158386, 'train/grad_norm_pre_clip': 10.16519546508789, 'timing/data': 0.0011544404551386833, 'timing/model': 1.1171845076605678, 'learning_rate/qwen_vl_interface': 6.801900563505388e-06, 'learning_rate/action_model': 6.801900563505388e-05, 'epoch': 1.08}) +[2026-06-17 22:30:50,862][starVLA.training.train_starvla][INFO] - Step 2820, Loss: {'train/loss': 0.548331618309021, 'train/grad_norm_pre_clip': 10.011674880981445, 'timing/data': 0.0011472059413790703, 'timing/model': 1.1250075669959188, 'learning_rate/qwen_vl_interface': 6.7998563949782755e-06, 'learning_rate/action_model': 6.799856394978276e-05, 'epoch': 1.08}) +[2026-06-17 22:30:55,859][starVLA.training.train_starvla][INFO] - Step 2821, Loss: {'train/loss': 0.42345041036605835, 'train/grad_norm_pre_clip': 9.419861793518066, 'timing/data': 0.0009349510073661804, 'timing/model': 1.19883578363806, 'learning_rate/qwen_vl_interface': 6.797811905164263e-06, 'learning_rate/action_model': 6.797811905164263e-05, 'epoch': 1.09}) +[2026-06-17 22:31:00,765][starVLA.training.train_starvla][INFO] - Step 2822, Loss: {'train/loss': 0.4886971116065979, 'train/grad_norm_pre_clip': 9.42230224609375, 'timing/data': 0.0011380910873413086, 'timing/model': 1.4901181673631072, 'learning_rate/qwen_vl_interface': 6.795767094487179e-06, 'learning_rate/action_model': 6.795767094487179e-05, 'epoch': 1.09}) +[2026-06-17 22:31:06,932][starVLA.training.train_starvla][INFO] - Step 2823, Loss: {'train/loss': 0.3489235043525696, 'train/grad_norm_pre_clip': 13.397483825683594, 'timing/data': 0.0011699367314577103, 'timing/model': 2.367167813703418, 'learning_rate/qwen_vl_interface': 6.793721963370913e-06, 'learning_rate/action_model': 6.793721963370913e-05, 'epoch': 1.09}) +[2026-06-17 22:31:12,045][starVLA.training.train_starvla][INFO] - Step 2824, Loss: {'train/loss': 0.3812617063522339, 'train/grad_norm_pre_clip': 9.612804412841797, 'timing/data': 0.0014000581577420235, 'timing/model': 1.0871508680284023, 'learning_rate/qwen_vl_interface': 6.7916765122394215e-06, 'learning_rate/action_model': 6.791676512239421e-05, 'epoch': 1.09}) +[2026-06-17 22:31:17,070][starVLA.training.train_starvla][INFO] - Step 2825, Loss: {'train/loss': 0.3224273920059204, 'train/grad_norm_pre_clip': 5.118534088134766, 'timing/data': 0.0009653670713305473, 'timing/model': 1.2974237129092216, 'learning_rate/qwen_vl_interface': 6.7896307415167305e-06, 'learning_rate/action_model': 6.78963074151673e-05, 'epoch': 1.09}) +[2026-06-17 22:31:21,758][starVLA.training.train_starvla][INFO] - Step 2826, Loss: {'train/loss': 0.36641788482666016, 'train/grad_norm_pre_clip': 10.7506103515625, 'timing/data': 0.0014699054881930351, 'timing/model': 1.19853344745934, 'learning_rate/qwen_vl_interface': 6.787584651626929e-06, 'learning_rate/action_model': 6.78758465162693e-05, 'epoch': 1.09}) +[2026-06-17 22:31:32,466][starVLA.training.train_starvla][INFO] - Step 2827, Loss: {'train/loss': 0.3756594955921173, 'train/grad_norm_pre_clip': 8.193070411682129, 'timing/data': 0.0011562854051589966, 'timing/model': 1.2723474651575089, 'learning_rate/qwen_vl_interface': 6.785538242994177e-06, 'learning_rate/action_model': 6.785538242994177e-05, 'epoch': 1.09}) +[2026-06-17 22:31:37,545][starVLA.training.train_starvla][INFO] - Step 2828, Loss: {'train/loss': 0.6125476360321045, 'train/grad_norm_pre_clip': 9.693058967590332, 'timing/data': 0.0014681201428174973, 'timing/model': 1.17725052498281, 'learning_rate/qwen_vl_interface': 6.783491516042694e-06, 'learning_rate/action_model': 6.783491516042693e-05, 'epoch': 1.09}) +[2026-06-17 22:31:43,147][starVLA.training.train_starvla][INFO] - Step 2829, Loss: {'train/loss': 0.6966461539268494, 'train/grad_norm_pre_clip': 14.057329177856445, 'timing/data': 0.0009076548740267754, 'timing/model': 1.2167639341205359, 'learning_rate/qwen_vl_interface': 6.781444471196768e-06, 'learning_rate/action_model': 6.781444471196769e-05, 'epoch': 1.09}) +[2026-06-17 22:31:48,697][starVLA.training.train_starvla][INFO] - Step 2830, Loss: {'train/loss': 0.5711153149604797, 'train/grad_norm_pre_clip': 13.487606048583984, 'timing/data': 0.0009835781529545784, 'timing/model': 1.417966578155756, 'learning_rate/qwen_vl_interface': 6.7793971088807575e-06, 'learning_rate/action_model': 6.779397108880757e-05, 'epoch': 1.09}) +[2026-06-17 22:31:56,463][starVLA.training.train_starvla][INFO] - Step 2831, Loss: {'train/loss': 0.3686641454696655, 'train/grad_norm_pre_clip': 6.488109111785889, 'timing/data': 0.0009256880730390549, 'timing/model': 4.191407258622348, 'learning_rate/qwen_vl_interface': 6.77734942951908e-06, 'learning_rate/action_model': 6.77734942951908e-05, 'epoch': 1.09}) +[2026-06-17 22:32:07,381][starVLA.training.train_starvla][INFO] - Step 2832, Loss: {'train/loss': 0.30402421951293945, 'train/grad_norm_pre_clip': 6.151796340942383, 'timing/data': 0.0009748693555593491, 'timing/model': 5.84333555214107, 'learning_rate/qwen_vl_interface': 6.7753014335362224e-06, 'learning_rate/action_model': 6.775301433536222e-05, 'epoch': 1.09}) +[2026-06-17 22:32:11,979][starVLA.training.train_starvla][INFO] - Step 2833, Loss: {'train/loss': 0.6005650162696838, 'train/grad_norm_pre_clip': 10.409218788146973, 'timing/data': 0.0012500584125518799, 'timing/model': 1.1358344564214349, 'learning_rate/qwen_vl_interface': 6.773253121356735e-06, 'learning_rate/action_model': 6.773253121356735e-05, 'epoch': 1.09}) +[2026-06-17 22:32:16,584][starVLA.training.train_starvla][INFO] - Step 2834, Loss: {'train/loss': 0.7451845407485962, 'train/grad_norm_pre_clip': 12.508670806884766, 'timing/data': 0.0014136703684926033, 'timing/model': 1.1544484933838248, 'learning_rate/qwen_vl_interface': 6.771204493405239e-06, 'learning_rate/action_model': 6.771204493405239e-05, 'epoch': 1.09}) +[2026-06-17 22:32:21,296][starVLA.training.train_starvla][INFO] - Step 2835, Loss: {'train/loss': 0.42642009258270264, 'train/grad_norm_pre_clip': 5.967884540557861, 'timing/data': 0.0011133607476949692, 'timing/model': 1.1433441722765565, 'learning_rate/qwen_vl_interface': 6.769155550106413e-06, 'learning_rate/action_model': 6.769155550106413e-05, 'epoch': 1.09}) +[2026-06-17 22:32:25,990][starVLA.training.train_starvla][INFO] - Step 2836, Loss: {'train/loss': 0.26919475197792053, 'train/grad_norm_pre_clip': 7.055950164794922, 'timing/data': 0.0014947839081287384, 'timing/model': 1.076952856965363, 'learning_rate/qwen_vl_interface': 6.767106291885006e-06, 'learning_rate/action_model': 6.767106291885006e-05, 'epoch': 1.09}) +[2026-06-17 22:32:34,502][starVLA.training.train_starvla][INFO] - Step 2837, Loss: {'train/loss': 0.28833404183387756, 'train/grad_norm_pre_clip': 6.625860691070557, 'timing/data': 0.0014016088098287582, 'timing/model': 1.8869449961930513, 'learning_rate/qwen_vl_interface': 6.765056719165833e-06, 'learning_rate/action_model': 6.765056719165834e-05, 'epoch': 1.09}) +[2026-06-17 22:32:39,788][starVLA.training.train_starvla][INFO] - Step 2838, Loss: {'train/loss': 0.4944908022880554, 'train/grad_norm_pre_clip': 9.936310768127441, 'timing/data': 0.0011745858937501907, 'timing/model': 1.28824283182621, 'learning_rate/qwen_vl_interface': 6.763006832373772e-06, 'learning_rate/action_model': 6.763006832373771e-05, 'epoch': 1.09}) +[2026-06-17 22:32:44,432][starVLA.training.train_starvla][INFO] - Step 2839, Loss: {'train/loss': 0.7023932337760925, 'train/grad_norm_pre_clip': 8.95704460144043, 'timing/data': 0.0014229128137230873, 'timing/model': 1.0891613094136119, 'learning_rate/qwen_vl_interface': 6.760956631933766e-06, 'learning_rate/action_model': 6.760956631933765e-05, 'epoch': 1.09}) +[2026-06-17 22:32:51,136][starVLA.training.train_starvla][INFO] - Step 2840, Loss: {'train/loss': 0.4494737684726715, 'train/grad_norm_pre_clip': 9.7800931930542, 'timing/data': 0.0011633187532424927, 'timing/model': 1.2377243032678962, 'learning_rate/qwen_vl_interface': 6.758906118270826e-06, 'learning_rate/action_model': 6.758906118270825e-05, 'epoch': 1.09}) +[2026-06-17 22:32:59,363][starVLA.training.train_starvla][INFO] - Step 2841, Loss: {'train/loss': 0.3517279326915741, 'train/grad_norm_pre_clip': 7.20066499710083, 'timing/data': 0.0009689200669527054, 'timing/model': 1.3712193882092834, 'learning_rate/qwen_vl_interface': 6.756855291810024e-06, 'learning_rate/action_model': 6.756855291810024e-05, 'epoch': 1.09}) +[2026-06-17 22:33:04,516][starVLA.training.train_starvla][INFO] - Step 2842, Loss: {'train/loss': 0.5558578372001648, 'train/grad_norm_pre_clip': 10.60988712310791, 'timing/data': 0.0011520013213157654, 'timing/model': 1.7280989168211818, 'learning_rate/qwen_vl_interface': 6.754804152976495e-06, 'learning_rate/action_model': 6.754804152976495e-05, 'epoch': 1.09}) +[2026-06-17 22:33:15,102][starVLA.training.train_starvla][INFO] - Step 2843, Loss: {'train/loss': 0.6015893816947937, 'train/grad_norm_pre_clip': 10.062934875488281, 'timing/data': 0.0011358661577105522, 'timing/model': 2.7406661799177527, 'learning_rate/qwen_vl_interface': 6.752752702195448e-06, 'learning_rate/action_model': 6.752752702195448e-05, 'epoch': 1.09}) +[2026-06-17 22:33:20,277][starVLA.training.train_starvla][INFO] - Step 2844, Loss: {'train/loss': 0.18181562423706055, 'train/grad_norm_pre_clip': 4.571841239929199, 'timing/data': 0.0012453757226467133, 'timing/model': 1.3801823183894157, 'learning_rate/qwen_vl_interface': 6.750700939892151e-06, 'learning_rate/action_model': 6.75070093989215e-05, 'epoch': 1.09}) +[2026-06-17 22:33:25,097][starVLA.training.train_starvla][INFO] - Step 2845, Loss: {'train/loss': 0.5053064227104187, 'train/grad_norm_pre_clip': 7.6668477058410645, 'timing/data': 0.0011240523308515549, 'timing/model': 1.1008360097184777, 'learning_rate/qwen_vl_interface': 6.748648866491931e-06, 'learning_rate/action_model': 6.748648866491932e-05, 'epoch': 1.09}) +[2026-06-17 22:33:30,371][starVLA.training.train_starvla][INFO] - Step 2846, Loss: {'train/loss': 0.2412796914577484, 'train/grad_norm_pre_clip': 5.600264549255371, 'timing/data': 0.0012131249532103539, 'timing/model': 1.4128306796774268, 'learning_rate/qwen_vl_interface': 6.7465964824201915e-06, 'learning_rate/action_model': 6.746596482420192e-05, 'epoch': 1.09}) +[2026-06-17 22:33:40,129][starVLA.training.train_starvla][INFO] - Step 2847, Loss: {'train/loss': 0.38307029008865356, 'train/grad_norm_pre_clip': 8.071499824523926, 'timing/data': 0.0012891283258795738, 'timing/model': 2.9200543630868196, 'learning_rate/qwen_vl_interface': 6.744543788102388e-06, 'learning_rate/action_model': 6.744543788102388e-05, 'epoch': 1.1}) +[2026-06-17 22:33:45,175][starVLA.training.train_starvla][INFO] - Step 2848, Loss: {'train/loss': 0.27864205837249756, 'train/grad_norm_pre_clip': 6.960561752319336, 'timing/data': 0.0012432113289833069, 'timing/model': 1.1355533553287387, 'learning_rate/qwen_vl_interface': 6.742490783964051e-06, 'learning_rate/action_model': 6.742490783964051e-05, 'epoch': 1.1}) +[2026-06-17 22:33:49,940][starVLA.training.train_starvla][INFO] - Step 2849, Loss: {'train/loss': 0.3877480924129486, 'train/grad_norm_pre_clip': 8.16893482208252, 'timing/data': 0.0014383727684617043, 'timing/model': 1.101059547625482, 'learning_rate/qwen_vl_interface': 6.740437470430766e-06, 'learning_rate/action_model': 6.740437470430766e-05, 'epoch': 1.1}) +[2026-06-17 22:33:55,165][starVLA.training.train_starvla][INFO] - Step 2850, Loss: {'train/loss': 0.5013530850410461, 'train/grad_norm_pre_clip': 8.469504356384277, 'timing/data': 0.0013434672728180885, 'timing/model': 1.1611815420910716, 'learning_rate/qwen_vl_interface': 6.7383838479281915e-06, 'learning_rate/action_model': 6.738383847928191e-05, 'epoch': 1.1}) +[2026-06-17 22:34:00,519][starVLA.training.train_starvla][INFO] - Step 2851, Loss: {'train/loss': 0.4078575372695923, 'train/grad_norm_pre_clip': 6.3068528175354, 'timing/data': 0.0011231554672122002, 'timing/model': 1.14029788877815, 'learning_rate/qwen_vl_interface': 6.736329916882044e-06, 'learning_rate/action_model': 6.736329916882044e-05, 'epoch': 1.1}) +[2026-06-17 22:34:10,978][starVLA.training.train_starvla][INFO] - Step 2852, Loss: {'train/loss': 0.7612632513046265, 'train/grad_norm_pre_clip': 11.19637680053711, 'timing/data': 0.0014836909249424934, 'timing/model': 1.4639993030577898, 'learning_rate/qwen_vl_interface': 6.734275677718105e-06, 'learning_rate/action_model': 6.734275677718105e-05, 'epoch': 1.1}) +[2026-06-17 22:34:21,830][starVLA.training.train_starvla][INFO] - Step 2853, Loss: {'train/loss': 0.6084339022636414, 'train/grad_norm_pre_clip': 11.221868515014648, 'timing/data': 0.0014403155073523521, 'timing/model': 1.2839335016906261, 'learning_rate/qwen_vl_interface': 6.73222113086222e-06, 'learning_rate/action_model': 6.73222113086222e-05, 'epoch': 1.1}) +[2026-06-17 22:34:26,596][starVLA.training.train_starvla][INFO] - Step 2854, Loss: {'train/loss': 0.361592173576355, 'train/grad_norm_pre_clip': 7.790590286254883, 'timing/data': 0.0011104419827461243, 'timing/model': 1.0732118664309382, 'learning_rate/qwen_vl_interface': 6.7301662767403e-06, 'learning_rate/action_model': 6.7301662767403e-05, 'epoch': 1.1}) +[2026-06-17 22:34:31,366][starVLA.training.train_starvla][INFO] - Step 2855, Loss: {'train/loss': 0.24691292643547058, 'train/grad_norm_pre_clip': 6.451178073883057, 'timing/data': 0.0015094084665179253, 'timing/model': 1.1136201145127416, 'learning_rate/qwen_vl_interface': 6.728111115778319e-06, 'learning_rate/action_model': 6.728111115778318e-05, 'epoch': 1.1}) +[2026-06-17 22:34:36,573][starVLA.training.train_starvla][INFO] - Step 2856, Loss: {'train/loss': 0.4875752329826355, 'train/grad_norm_pre_clip': 6.756204605102539, 'timing/data': 0.0013047317042946815, 'timing/model': 1.2658426612615585, 'learning_rate/qwen_vl_interface': 6.7260556484023135e-06, 'learning_rate/action_model': 6.726055648402313e-05, 'epoch': 1.1}) +[2026-06-17 22:34:41,186][starVLA.training.train_starvla][INFO] - Step 2857, Loss: {'train/loss': 0.36976221203804016, 'train/grad_norm_pre_clip': 7.656350135803223, 'timing/data': 0.0011251149699091911, 'timing/model': 1.07829075679183, 'learning_rate/qwen_vl_interface': 6.723999875038381e-06, 'learning_rate/action_model': 6.72399987503838e-05, 'epoch': 1.1}) +[2026-06-17 22:34:46,079][starVLA.training.train_starvla][INFO] - Step 2858, Loss: {'train/loss': 0.48408135771751404, 'train/grad_norm_pre_clip': 12.150219917297363, 'timing/data': 0.0009512566030025482, 'timing/model': 1.1966909943148494, 'learning_rate/qwen_vl_interface': 6.721943796112691e-06, 'learning_rate/action_model': 6.721943796112691e-05, 'epoch': 1.1}) +[2026-06-17 22:34:54,304][starVLA.training.train_starvla][INFO] - Step 2859, Loss: {'train/loss': 0.26185086369514465, 'train/grad_norm_pre_clip': 7.0280561447143555, 'timing/data': 0.0012118583545088768, 'timing/model': 1.0914452020078897, 'learning_rate/qwen_vl_interface': 6.719887412051468e-06, 'learning_rate/action_model': 6.719887412051468e-05, 'epoch': 1.1}) +[2026-06-17 22:34:59,229][starVLA.training.train_starvla][INFO] - Step 2860, Loss: {'train/loss': 0.6207850575447083, 'train/grad_norm_pre_clip': 12.254274368286133, 'timing/data': 0.0011783670634031296, 'timing/model': 1.1566297691315413, 'learning_rate/qwen_vl_interface': 6.7178307232810015e-06, 'learning_rate/action_model': 6.717830723281e-05, 'epoch': 1.1}) +[2026-06-17 22:35:04,554][starVLA.training.train_starvla][INFO] - Step 2861, Loss: {'train/loss': 0.6348273158073425, 'train/grad_norm_pre_clip': 10.51888370513916, 'timing/data': 0.0012064948678016663, 'timing/model': 1.0833266396075487, 'learning_rate/qwen_vl_interface': 6.715773730227646e-06, 'learning_rate/action_model': 6.715773730227646e-05, 'epoch': 1.1}) +[2026-06-17 22:35:09,475][starVLA.training.train_starvla][INFO] - Step 2862, Loss: {'train/loss': 0.4020751118659973, 'train/grad_norm_pre_clip': 8.60995864868164, 'timing/data': 0.0011398708447813988, 'timing/model': 1.4893135111778975, 'learning_rate/qwen_vl_interface': 6.7137164333178216e-06, 'learning_rate/action_model': 6.713716433317821e-05, 'epoch': 1.1}) +[2026-06-17 22:35:14,013][starVLA.training.train_starvla][INFO] - Step 2863, Loss: {'train/loss': 0.14862976968288422, 'train/grad_norm_pre_clip': 3.6105234622955322, 'timing/data': 0.0009086132049560547, 'timing/model': 1.0907177850604057, 'learning_rate/qwen_vl_interface': 6.711658832978004e-06, 'learning_rate/action_model': 6.711658832978003e-05, 'epoch': 1.1}) +[2026-06-17 22:35:20,693][starVLA.training.train_starvla][INFO] - Step 2864, Loss: {'train/loss': 0.4689897298812866, 'train/grad_norm_pre_clip': 8.660590171813965, 'timing/data': 0.0018912926316261292, 'timing/model': 2.713459744118154, 'learning_rate/qwen_vl_interface': 6.709600929634737e-06, 'learning_rate/action_model': 6.709600929634738e-05, 'epoch': 1.1}) +[2026-06-17 22:35:29,325][starVLA.training.train_starvla][INFO] - Step 2865, Loss: {'train/loss': 0.5246294140815735, 'train/grad_norm_pre_clip': 8.461233139038086, 'timing/data': 0.0011455239728093147, 'timing/model': 1.258799278177321, 'learning_rate/qwen_vl_interface': 6.707542723714629e-06, 'learning_rate/action_model': 6.707542723714629e-05, 'epoch': 1.1}) +[2026-06-17 22:35:34,138][starVLA.training.train_starvla][INFO] - Step 2866, Loss: {'train/loss': 0.3872893750667572, 'train/grad_norm_pre_clip': 9.744568824768066, 'timing/data': 0.0011546928435564041, 'timing/model': 1.1303408965468407, 'learning_rate/qwen_vl_interface': 6.7054842156443435e-06, 'learning_rate/action_model': 6.705484215644343e-05, 'epoch': 1.1}) +[2026-06-17 22:35:43,431][starVLA.training.train_starvla][INFO] - Step 2867, Loss: {'train/loss': 0.3414691686630249, 'train/grad_norm_pre_clip': 9.485836029052734, 'timing/data': 0.001254018396139145, 'timing/model': 1.1543238488957286, 'learning_rate/qwen_vl_interface': 6.703425405850614e-06, 'learning_rate/action_model': 6.703425405850613e-05, 'epoch': 1.1}) +[2026-06-17 22:35:50,782][starVLA.training.train_starvla][INFO] - Step 2868, Loss: {'train/loss': 0.5695250034332275, 'train/grad_norm_pre_clip': 10.390192031860352, 'timing/data': 0.0007920870557427406, 'timing/model': 3.3053611340001225, 'learning_rate/qwen_vl_interface': 6.701366294760236e-06, 'learning_rate/action_model': 6.701366294760235e-05, 'epoch': 1.1}) +[2026-06-17 22:35:55,920][starVLA.training.train_starvla][INFO] - Step 2869, Loss: {'train/loss': 0.22059893608093262, 'train/grad_norm_pre_clip': 9.471333503723145, 'timing/data': 0.0017349757254123688, 'timing/model': 1.0926639307290316, 'learning_rate/qwen_vl_interface': 6.6993068828000605e-06, 'learning_rate/action_model': 6.69930688280006e-05, 'epoch': 1.1}) +[2026-06-17 22:36:05,641][starVLA.training.train_starvla][INFO] - Step 2870, Loss: {'train/loss': 0.5936658978462219, 'train/grad_norm_pre_clip': 11.222004890441895, 'timing/data': 0.0021370379254221916, 'timing/model': 5.341229948215187, 'learning_rate/qwen_vl_interface': 6.697247170397011e-06, 'learning_rate/action_model': 6.697247170397011e-05, 'epoch': 1.1}) +[2026-06-17 22:36:13,085][starVLA.training.train_starvla][INFO] - Step 2871, Loss: {'train/loss': 0.6644088625907898, 'train/grad_norm_pre_clip': 51.64632797241211, 'timing/data': 0.0008135717362165451, 'timing/model': 3.2561892131343484, 'learning_rate/qwen_vl_interface': 6.6951871579780645e-06, 'learning_rate/action_model': 6.695187157978064e-05, 'epoch': 1.1}) +[2026-06-17 22:36:21,003][starVLA.training.train_starvla][INFO] - Step 2872, Loss: {'train/loss': 0.4784338176250458, 'train/grad_norm_pre_clip': 8.050683975219727, 'timing/data': 0.0011564437299966812, 'timing/model': 3.1680372413247824, 'learning_rate/qwen_vl_interface': 6.693126845970267e-06, 'learning_rate/action_model': 6.693126845970267e-05, 'epoch': 1.1}) +[2026-06-17 22:36:25,425][starVLA.training.train_starvla][INFO] - Step 2873, Loss: {'train/loss': 0.5210180878639221, 'train/grad_norm_pre_clip': 8.755210876464844, 'timing/data': 0.0010933103039860725, 'timing/model': 1.0803570123389363, 'learning_rate/qwen_vl_interface': 6.691066234800719e-06, 'learning_rate/action_model': 6.691066234800719e-05, 'epoch': 1.11}) +[2026-06-17 22:36:30,190][starVLA.training.train_starvla][INFO] - Step 2874, Loss: {'train/loss': 0.37754127383232117, 'train/grad_norm_pre_clip': 10.949975967407227, 'timing/data': 0.001937716268002987, 'timing/model': 1.1831776797771454, 'learning_rate/qwen_vl_interface': 6.689005324896591e-06, 'learning_rate/action_model': 6.689005324896591e-05, 'epoch': 1.11}) +[2026-06-17 22:36:35,329][starVLA.training.train_starvla][INFO] - Step 2875, Loss: {'train/loss': 0.6421446800231934, 'train/grad_norm_pre_clip': 13.179975509643555, 'timing/data': 0.0009380020201206207, 'timing/model': 1.2562218345701694, 'learning_rate/qwen_vl_interface': 6.686944116685111e-06, 'learning_rate/action_model': 6.686944116685111e-05, 'epoch': 1.11}) +[2026-06-17 22:36:44,765][starVLA.training.train_starvla][INFO] - Step 2876, Loss: {'train/loss': 0.2998627722263336, 'train/grad_norm_pre_clip': 12.096537590026855, 'timing/data': 0.0012658853083848953, 'timing/model': 1.1397457448765635, 'learning_rate/qwen_vl_interface': 6.684882610593568e-06, 'learning_rate/action_model': 6.684882610593567e-05, 'epoch': 1.11}) +[2026-06-17 22:36:49,662][starVLA.training.train_starvla][INFO] - Step 2877, Loss: {'train/loss': 0.5021986365318298, 'train/grad_norm_pre_clip': 14.297972679138184, 'timing/data': 0.0018789218738675117, 'timing/model': 1.3583474522456527, 'learning_rate/qwen_vl_interface': 6.682820807049317e-06, 'learning_rate/action_model': 6.682820807049317e-05, 'epoch': 1.11}) +[2026-06-17 22:36:54,665][starVLA.training.train_starvla][INFO] - Step 2878, Loss: {'train/loss': 0.3434496521949768, 'train/grad_norm_pre_clip': 6.252625942230225, 'timing/data': 0.0011525247246026993, 'timing/model': 1.1971833072602749, 'learning_rate/qwen_vl_interface': 6.680758706479772e-06, 'learning_rate/action_model': 6.680758706479772e-05, 'epoch': 1.11}) +[2026-06-17 22:37:02,710][starVLA.training.train_starvla][INFO] - Step 2879, Loss: {'train/loss': 0.4732265770435333, 'train/grad_norm_pre_clip': 9.337007522583008, 'timing/data': 0.0011224020272493362, 'timing/model': 1.0830483911558986, 'learning_rate/qwen_vl_interface': 6.678696309312405e-06, 'learning_rate/action_model': 6.678696309312406e-05, 'epoch': 1.11}) +[2026-06-17 22:37:07,339][starVLA.training.train_starvla][INFO] - Step 2880, Loss: {'train/loss': 0.19937720894813538, 'train/grad_norm_pre_clip': 7.062325954437256, 'timing/data': 0.0014042994007468224, 'timing/model': 1.099198117852211, 'learning_rate/qwen_vl_interface': 6.676633615974758e-06, 'learning_rate/action_model': 6.676633615974757e-05, 'epoch': 1.11}) +[2026-06-17 22:37:15,061][starVLA.training.train_starvla][INFO] - Step 2881, Loss: {'train/loss': 0.31772947311401367, 'train/grad_norm_pre_clip': 10.11198902130127, 'timing/data': 0.0015343250706791878, 'timing/model': 3.6861454378813505, 'learning_rate/qwen_vl_interface': 6.6745706268944265e-06, 'learning_rate/action_model': 6.674570626894426e-05, 'epoch': 1.11}) +[2026-06-17 22:37:20,007][starVLA.training.train_starvla][INFO] - Step 2882, Loss: {'train/loss': 0.6403862833976746, 'train/grad_norm_pre_clip': 9.419143676757812, 'timing/data': 0.0010162796825170517, 'timing/model': 1.4063935307785869, 'learning_rate/qwen_vl_interface': 6.672507342499071e-06, 'learning_rate/action_model': 6.67250734249907e-05, 'epoch': 1.11}) +[2026-06-17 22:37:24,305][starVLA.training.train_starvla][INFO] - Step 2883, Loss: {'train/loss': 0.9966038465499878, 'train/grad_norm_pre_clip': 16.616165161132812, 'timing/data': 0.0009254198521375656, 'timing/model': 1.064088678918779, 'learning_rate/qwen_vl_interface': 6.670443763216412e-06, 'learning_rate/action_model': 6.670443763216411e-05, 'epoch': 1.11}) +[2026-06-17 22:37:28,905][starVLA.training.train_starvla][INFO] - Step 2884, Loss: {'train/loss': 0.525747537612915, 'train/grad_norm_pre_clip': 11.351192474365234, 'timing/data': 0.0009482484310865402, 'timing/model': 1.1529839904978871, 'learning_rate/qwen_vl_interface': 6.668379889474233e-06, 'learning_rate/action_model': 6.668379889474234e-05, 'epoch': 1.11}) +[2026-06-17 22:37:33,824][starVLA.training.train_starvla][INFO] - Step 2885, Loss: {'train/loss': 0.6301940679550171, 'train/grad_norm_pre_clip': 13.031290054321289, 'timing/data': 0.001261894591152668, 'timing/model': 1.2294602394104004, 'learning_rate/qwen_vl_interface': 6.666315721700378e-06, 'learning_rate/action_model': 6.666315721700377e-05, 'epoch': 1.11}) +[2026-06-17 22:37:38,738][starVLA.training.train_starvla][INFO] - Step 2886, Loss: {'train/loss': 0.45122388005256653, 'train/grad_norm_pre_clip': 14.584776878356934, 'timing/data': 0.0014778776094317436, 'timing/model': 1.0999806020408869, 'learning_rate/qwen_vl_interface': 6.664251260322748e-06, 'learning_rate/action_model': 6.664251260322749e-05, 'epoch': 1.11}) +[2026-06-17 22:37:47,980][starVLA.training.train_starvla][INFO] - Step 2887, Loss: {'train/loss': 0.35065388679504395, 'train/grad_norm_pre_clip': 9.608702659606934, 'timing/data': 0.0012638512998819351, 'timing/model': 1.2461562287062407, 'learning_rate/qwen_vl_interface': 6.66218650576931e-06, 'learning_rate/action_model': 6.66218650576931e-05, 'epoch': 1.11}) +[2026-06-17 22:37:52,746][starVLA.training.train_starvla][INFO] - Step 2888, Loss: {'train/loss': 0.3957687020301819, 'train/grad_norm_pre_clip': 11.371925354003906, 'timing/data': 0.0014087716117501259, 'timing/model': 1.0893639763817191, 'learning_rate/qwen_vl_interface': 6.66012145846809e-06, 'learning_rate/action_model': 6.660121458468091e-05, 'epoch': 1.11}) +[2026-06-17 22:37:57,700][starVLA.training.train_starvla][INFO] - Step 2889, Loss: {'train/loss': 0.24921217560768127, 'train/grad_norm_pre_clip': 5.616518974304199, 'timing/data': 0.0018497537821531296, 'timing/model': 1.0778238233178854, 'learning_rate/qwen_vl_interface': 6.658056118847176e-06, 'learning_rate/action_model': 6.658056118847175e-05, 'epoch': 1.11}) +[2026-06-17 22:38:05,352][starVLA.training.train_starvla][INFO] - Step 2890, Loss: {'train/loss': 0.6975468993186951, 'train/grad_norm_pre_clip': 10.534218788146973, 'timing/data': 0.000837305560708046, 'timing/model': 3.9182774126529694, 'learning_rate/qwen_vl_interface': 6.655990487334713e-06, 'learning_rate/action_model': 6.655990487334712e-05, 'epoch': 1.11}) +[2026-06-17 22:38:14,637][starVLA.training.train_starvla][INFO] - Step 2891, Loss: {'train/loss': 0.6910794973373413, 'train/grad_norm_pre_clip': 14.131768226623535, 'timing/data': 0.0008944813162088394, 'timing/model': 1.298686289228499, 'learning_rate/qwen_vl_interface': 6.653924564358907e-06, 'learning_rate/action_model': 6.653924564358907e-05, 'epoch': 1.11}) +[2026-06-17 22:38:19,357][starVLA.training.train_starvla][INFO] - Step 2892, Loss: {'train/loss': 0.7148781418800354, 'train/grad_norm_pre_clip': 15.317243576049805, 'timing/data': 0.002332910895347595, 'timing/model': 1.4042990962043405, 'learning_rate/qwen_vl_interface': 6.6518583503480294e-06, 'learning_rate/action_model': 6.651858350348029e-05, 'epoch': 1.11}) +[2026-06-17 22:38:23,901][starVLA.training.train_starvla][INFO] - Step 2893, Loss: {'train/loss': 0.20012013614177704, 'train/grad_norm_pre_clip': 6.820121765136719, 'timing/data': 0.0010714521631598473, 'timing/model': 1.1885352535173297, 'learning_rate/qwen_vl_interface': 6.649791845730408e-06, 'learning_rate/action_model': 6.649791845730408e-05, 'epoch': 1.11}) +[2026-06-17 22:38:30,464][starVLA.training.train_starvla][INFO] - Step 2894, Loss: {'train/loss': 0.5377330183982849, 'train/grad_norm_pre_clip': 10.266705513000488, 'timing/data': 0.0010976409539580345, 'timing/model': 1.3804014045745134, 'learning_rate/qwen_vl_interface': 6.64772505093443e-06, 'learning_rate/action_model': 6.647725050934429e-05, 'epoch': 1.11}) +[2026-06-17 22:38:39,778][starVLA.training.train_starvla][INFO] - Step 2895, Loss: {'train/loss': 0.4485529959201813, 'train/grad_norm_pre_clip': 9.673553466796875, 'timing/data': 0.0014001140370965004, 'timing/model': 1.4267581133171916, 'learning_rate/qwen_vl_interface': 6.645657966388543e-06, 'learning_rate/action_model': 6.645657966388543e-05, 'epoch': 1.11}) +[2026-06-17 22:38:44,822][starVLA.training.train_starvla][INFO] - Step 2896, Loss: {'train/loss': 0.7397588491439819, 'train/grad_norm_pre_clip': 12.939642906188965, 'timing/data': 0.001092725433409214, 'timing/model': 1.2786069251596928, 'learning_rate/qwen_vl_interface': 6.643590592521261e-06, 'learning_rate/action_model': 6.643590592521261e-05, 'epoch': 1.11}) +[2026-06-17 22:38:51,967][starVLA.training.train_starvla][INFO] - Step 2897, Loss: {'train/loss': 0.49844008684158325, 'train/grad_norm_pre_clip': 10.557354927062988, 'timing/data': 0.0013209162279963493, 'timing/model': 3.7984601529315114, 'learning_rate/qwen_vl_interface': 6.641522929761148e-06, 'learning_rate/action_model': 6.641522929761148e-05, 'epoch': 1.11}) +[2026-06-17 22:38:57,141][starVLA.training.train_starvla][INFO] - Step 2898, Loss: {'train/loss': 0.6951258778572083, 'train/grad_norm_pre_clip': 14.41527271270752, 'timing/data': 0.0011722669005393982, 'timing/model': 1.169030132703483, 'learning_rate/qwen_vl_interface': 6.639454978536835e-06, 'learning_rate/action_model': 6.639454978536834e-05, 'epoch': 1.11}) +[2026-06-17 22:39:02,008][starVLA.training.train_starvla][INFO] - Step 2899, Loss: {'train/loss': 0.6920823454856873, 'train/grad_norm_pre_clip': 12.738707542419434, 'timing/data': 0.001167861744761467, 'timing/model': 1.178849502466619, 'learning_rate/qwen_vl_interface': 6.63738673927701e-06, 'learning_rate/action_model': 6.63738673927701e-05, 'epoch': 1.12}) +[2026-06-17 22:43:01,607][starVLA.training.train_starvla][INFO] - Step 2900, Loss: {'train/loss': 0.3333573639392853, 'train/grad_norm_pre_clip': 11.995063781738281, 'eval/loss': 0.47602519392967224, 'eval/latency_4/loss': 0.47602519392967224, 'eval/demon_fire_precision': 0.9596953348143447, 'eval/demon_fire_recall': 0.9821370574861968, 'eval/demon_fire_f1': 0.9707865168539326, 'eval/demon_fire_rate_model': 0.9254038179148312, 'eval/demon_fire_rate_teacher': 0.9042584434654919, 'eval/demon_left_precision': 0.7643835616438356, 'eval/demon_left_recall': 0.636986301369863, 'eval/demon_left_f1': 0.6948941469489415, 'eval/demon_left_rate_model': 0.10719530102790015, 'eval/demon_left_rate_teacher': 0.12863436123348018, 'eval/demon_right_precision': 0.7862595419847328, 'eval/demon_right_recall': 0.7436823104693141, 'eval/demon_right_f1': 0.7643784786641928, 'eval/demon_right_rate_model': 0.1538913362701909, 'eval/demon_right_rate_teacher': 0.16270190895741557, 'eval/demon_move_f1': 0.7296363128065672, 'eval/demon_cc_f1': 0.8832138970116324, 'eval/latency_4/demon_fire_precision': 0.9596953348143447, 'eval/latency_4/demon_fire_recall': 0.9821370574861968, 'eval/latency_4/demon_fire_f1': 0.9707865168539326, 'eval/latency_4/demon_fire_rate_model': 0.9254038179148312, 'eval/latency_4/demon_fire_rate_teacher': 0.9042584434654919, 'eval/latency_4/demon_left_precision': 0.7643835616438356, 'eval/latency_4/demon_left_recall': 0.636986301369863, 'eval/latency_4/demon_left_f1': 0.6948941469489415, 'eval/latency_4/demon_left_rate_model': 0.10719530102790015, 'eval/latency_4/demon_left_rate_teacher': 0.12863436123348018, 'eval/latency_4/demon_right_precision': 0.7862595419847328, 'eval/latency_4/demon_right_recall': 0.7436823104693141, 'eval/latency_4/demon_right_f1': 0.7643784786641928, 'eval/latency_4/demon_right_rate_model': 0.1538913362701909, 'eval/latency_4/demon_right_rate_teacher': 0.16270190895741557, 'eval/latency_4/demon_move_f1': 0.7296363128065672, 'eval/latency_4/demon_cc_f1': 0.8832138970116324, 'timing/data': 0.006927649490535259, 'timing/model': 1.251121873036027, 'learning_rate/qwen_vl_interface': 6.63531821241042e-06, 'learning_rate/action_model': 6.63531821241042e-05, 'epoch': 1.12}) +[2026-06-17 22:43:01,609][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_2900_state +[2026-06-17 22:43:01,617][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 22:43:12,250][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_2900_state/model.safetensors +[2026-06-17 22:43:36,356][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_2900_state/optimizer.bin +[2026-06-17 22:43:36,357][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_2900_state/sampler.bin +[2026-06-17 22:43:36,358][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_2900_state/sampler_1.bin +[2026-06-17 22:43:36,361][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_2900_state/random_states_0.pkl +[2026-06-17 22:43:36,376][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2900 +[2026-06-17 22:43:36,376][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 22:43:36,394][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 22:43:41,484][starVLA.training.train_starvla][INFO] - Step 2901, Loss: {'train/loss': 0.40399080514907837, 'train/grad_norm_pre_clip': 8.629130363464355, 'timing/data': 0.0012220386415719986, 'timing/model': 1.615435608662665, 'learning_rate/qwen_vl_interface': 6.633249398365874e-06, 'learning_rate/action_model': 6.633249398365873e-05, 'epoch': 1.12}) +[2026-06-17 22:43:45,988][starVLA.training.train_starvla][INFO] - Step 2902, Loss: {'train/loss': 0.2747373878955841, 'train/grad_norm_pre_clip': 7.442472457885742, 'timing/data': 0.0013031736016273499, 'timing/model': 1.2041263785213232, 'learning_rate/qwen_vl_interface': 6.6311802975722395e-06, 'learning_rate/action_model': 6.63118029757224e-05, 'epoch': 1.12}) +[2026-06-17 22:43:51,108][starVLA.training.train_starvla][INFO] - Step 2903, Loss: {'train/loss': 0.3664340376853943, 'train/grad_norm_pre_clip': 9.391133308410645, 'timing/data': 0.0009593088179826736, 'timing/model': 1.310702527873218, 'learning_rate/qwen_vl_interface': 6.629110910458444e-06, 'learning_rate/action_model': 6.629110910458444e-05, 'epoch': 1.12}) +[2026-06-17 22:43:56,039][starVLA.training.train_starvla][INFO] - Step 2904, Loss: {'train/loss': 0.3908393681049347, 'train/grad_norm_pre_clip': 7.943083763122559, 'timing/data': 0.0015675080940127373, 'timing/model': 1.1847054287791252, 'learning_rate/qwen_vl_interface': 6.627041237453471e-06, 'learning_rate/action_model': 6.627041237453471e-05, 'epoch': 1.12}) +[2026-06-17 22:44:04,309][starVLA.training.train_starvla][INFO] - Step 2905, Loss: {'train/loss': 0.3449659049510956, 'train/grad_norm_pre_clip': 6.100162982940674, 'timing/data': 0.0011275839060544968, 'timing/model': 4.140153529122472, 'learning_rate/qwen_vl_interface': 6.624971278986369e-06, 'learning_rate/action_model': 6.62497127898637e-05, 'epoch': 1.12}) +[2026-06-17 22:44:14,908][starVLA.training.train_starvla][INFO] - Step 2906, Loss: {'train/loss': 0.32657983899116516, 'train/grad_norm_pre_clip': 10.834590911865234, 'timing/data': 0.0014492552727460861, 'timing/model': 1.0782951219007373, 'learning_rate/qwen_vl_interface': 6.622901035486241e-06, 'learning_rate/action_model': 6.622901035486241e-05, 'epoch': 1.12}) +[2026-06-17 22:44:19,772][starVLA.training.train_starvla][INFO] - Step 2907, Loss: {'train/loss': 0.2734092175960541, 'train/grad_norm_pre_clip': 6.409523010253906, 'timing/data': 0.0010833200067281723, 'timing/model': 1.595072048716247, 'learning_rate/qwen_vl_interface': 6.620830507382251e-06, 'learning_rate/action_model': 6.620830507382251e-05, 'epoch': 1.12}) +[2026-06-17 22:44:24,386][starVLA.training.train_starvla][INFO] - Step 2908, Loss: {'train/loss': 0.5644547343254089, 'train/grad_norm_pre_clip': 9.684415817260742, 'timing/data': 0.0012336662039160728, 'timing/model': 1.1523839309811592, 'learning_rate/qwen_vl_interface': 6.618759695103623e-06, 'learning_rate/action_model': 6.618759695103622e-05, 'epoch': 1.12}) +[2026-06-17 22:44:31,618][starVLA.training.train_starvla][INFO] - Step 2909, Loss: {'train/loss': 0.5220893025398254, 'train/grad_norm_pre_clip': 10.227800369262695, 'timing/data': 0.0009206710383296013, 'timing/model': 1.1313750091940165, 'learning_rate/qwen_vl_interface': 6.616688599079635e-06, 'learning_rate/action_model': 6.616688599079634e-05, 'epoch': 1.12}) +[2026-06-17 22:44:36,591][starVLA.training.train_starvla][INFO] - Step 2910, Loss: {'train/loss': 0.48843443393707275, 'train/grad_norm_pre_clip': 8.607544898986816, 'timing/data': 0.0012973658740520477, 'timing/model': 1.0617278553545475, 'learning_rate/qwen_vl_interface': 6.614617219739632e-06, 'learning_rate/action_model': 6.614617219739632e-05, 'epoch': 1.12}) +[2026-06-17 22:44:41,532][starVLA.training.train_starvla][INFO] - Step 2911, Loss: {'train/loss': 0.394653856754303, 'train/grad_norm_pre_clip': 8.298382759094238, 'timing/data': 0.0013142693787813187, 'timing/model': 1.2508895490318537, 'learning_rate/qwen_vl_interface': 6.612545557513008e-06, 'learning_rate/action_model': 6.612545557513008e-05, 'epoch': 1.12}) +[2026-06-17 22:44:46,060][starVLA.training.train_starvla][INFO] - Step 2912, Loss: {'train/loss': 0.5697304606437683, 'train/grad_norm_pre_clip': 9.232900619506836, 'timing/data': 0.0011852523311972618, 'timing/model': 1.2880998197942972, 'learning_rate/qwen_vl_interface': 6.610473612829227e-06, 'learning_rate/action_model': 6.610473612829226e-05, 'epoch': 1.12}) +[2026-06-17 22:44:51,161][starVLA.training.train_starvla][INFO] - Step 2913, Loss: {'train/loss': 0.6967052221298218, 'train/grad_norm_pre_clip': 10.321211814880371, 'timing/data': 0.001124827191233635, 'timing/model': 1.2069923439994454, 'learning_rate/qwen_vl_interface': 6.608401386117801e-06, 'learning_rate/action_model': 6.6084013861178e-05, 'epoch': 1.12}) +[2026-06-17 22:44:56,598][starVLA.training.train_starvla][INFO] - Step 2914, Loss: {'train/loss': 0.20174339413642883, 'train/grad_norm_pre_clip': 7.3963398933410645, 'timing/data': 0.000866670161485672, 'timing/model': 1.1842336878180504, 'learning_rate/qwen_vl_interface': 6.606328877808306e-06, 'learning_rate/action_model': 6.606328877808305e-05, 'epoch': 1.12}) +[2026-06-17 22:45:01,552][starVLA.training.train_starvla][INFO] - Step 2915, Loss: {'train/loss': 0.3840583860874176, 'train/grad_norm_pre_clip': 7.983837127685547, 'timing/data': 0.0011485060676932335, 'timing/model': 1.2684824047610164, 'learning_rate/qwen_vl_interface': 6.604256088330374e-06, 'learning_rate/action_model': 6.604256088330374e-05, 'epoch': 1.12}) +[2026-06-17 22:45:12,860][starVLA.training.train_starvla][INFO] - Step 2916, Loss: {'train/loss': 0.22218303382396698, 'train/grad_norm_pre_clip': 7.257081031799316, 'timing/data': 0.0010984735563397408, 'timing/model': 1.1838864479213953, 'learning_rate/qwen_vl_interface': 6.602183018113698e-06, 'learning_rate/action_model': 6.602183018113698e-05, 'epoch': 1.12}) +[2026-06-17 22:45:18,151][starVLA.training.train_starvla][INFO] - Step 2917, Loss: {'train/loss': 0.5410794019699097, 'train/grad_norm_pre_clip': 10.739933967590332, 'timing/data': 0.001085236668586731, 'timing/model': 1.38921039365232, 'learning_rate/qwen_vl_interface': 6.600109667588029e-06, 'learning_rate/action_model': 6.60010966758803e-05, 'epoch': 1.12}) +[2026-06-17 22:45:22,846][starVLA.training.train_starvla][INFO] - Step 2918, Loss: {'train/loss': 0.3320615887641907, 'train/grad_norm_pre_clip': 6.826925754547119, 'timing/data': 0.001131516881287098, 'timing/model': 1.0719084329903126, 'learning_rate/qwen_vl_interface': 6.598036037183172e-06, 'learning_rate/action_model': 6.598036037183172e-05, 'epoch': 1.12}) +[2026-06-17 22:45:32,543][starVLA.training.train_starvla][INFO] - Step 2919, Loss: {'train/loss': 0.27922460436820984, 'train/grad_norm_pre_clip': 6.932262420654297, 'timing/data': 0.00113615021109581, 'timing/model': 1.3725802227854729, 'learning_rate/qwen_vl_interface': 6.595962127328993e-06, 'learning_rate/action_model': 6.595962127328993e-05, 'epoch': 1.12}) +[2026-06-17 22:45:42,126][starVLA.training.train_starvla][INFO] - Step 2920, Loss: {'train/loss': 0.6072621941566467, 'train/grad_norm_pre_clip': 9.413137435913086, 'timing/data': 0.001171213574707508, 'timing/model': 1.1120259203016758, 'learning_rate/qwen_vl_interface': 6.593887938455418e-06, 'learning_rate/action_model': 6.593887938455418e-05, 'epoch': 1.12}) +[2026-06-17 22:45:46,815][starVLA.training.train_starvla][INFO] - Step 2921, Loss: {'train/loss': 0.4285805821418762, 'train/grad_norm_pre_clip': 8.621965408325195, 'timing/data': 0.001207665540277958, 'timing/model': 1.0741949072107673, 'learning_rate/qwen_vl_interface': 6.591813470992428e-06, 'learning_rate/action_model': 6.591813470992427e-05, 'epoch': 1.12}) +[2026-06-17 22:45:51,200][starVLA.training.train_starvla][INFO] - Step 2922, Loss: {'train/loss': 0.3019048571586609, 'train/grad_norm_pre_clip': 10.126181602478027, 'timing/data': 0.0011522108688950539, 'timing/model': 1.063780389726162, 'learning_rate/qwen_vl_interface': 6.58973872537006e-06, 'learning_rate/action_model': 6.589738725370059e-05, 'epoch': 1.12}) +[2026-06-17 22:45:56,184][starVLA.training.train_starvla][INFO] - Step 2923, Loss: {'train/loss': 0.2731702923774719, 'train/grad_norm_pre_clip': 5.219520568847656, 'timing/data': 0.0009868154302239418, 'timing/model': 1.213260912336409, 'learning_rate/qwen_vl_interface': 6.587663702018413e-06, 'learning_rate/action_model': 6.587663702018412e-05, 'epoch': 1.12}) +[2026-06-17 22:46:01,067][starVLA.training.train_starvla][INFO] - Step 2924, Loss: {'train/loss': 0.747353732585907, 'train/grad_norm_pre_clip': 13.953224182128906, 'timing/data': 0.0014727972447872162, 'timing/model': 1.1875849794596434, 'learning_rate/qwen_vl_interface': 6.585588401367641e-06, 'learning_rate/action_model': 6.58558840136764e-05, 'epoch': 1.12}) +[2026-06-17 22:46:05,847][starVLA.training.train_starvla][INFO] - Step 2925, Loss: {'train/loss': 0.4611685574054718, 'train/grad_norm_pre_clip': 7.11543083190918, 'timing/data': 0.001149756833910942, 'timing/model': 1.1443566279485822, 'learning_rate/qwen_vl_interface': 6.583512823847957e-06, 'learning_rate/action_model': 6.583512823847958e-05, 'epoch': 1.13}) +[2026-06-17 22:46:10,740][starVLA.training.train_starvla][INFO] - Step 2926, Loss: {'train/loss': 0.6628575921058655, 'train/grad_norm_pre_clip': 11.852725982666016, 'timing/data': 0.0009837988764047623, 'timing/model': 1.0853670174255967, 'learning_rate/qwen_vl_interface': 6.5814369698896285e-06, 'learning_rate/action_model': 6.581436969889628e-05, 'epoch': 1.13}) +[2026-06-17 22:46:18,949][starVLA.training.train_starvla][INFO] - Step 2927, Loss: {'train/loss': 0.4837530851364136, 'train/grad_norm_pre_clip': 11.11319637298584, 'timing/data': 0.0014637289568781853, 'timing/model': 4.669785875827074, 'learning_rate/qwen_vl_interface': 6.5793608399229855e-06, 'learning_rate/action_model': 6.579360839922985e-05, 'epoch': 1.13}) +[2026-06-17 22:46:23,437][starVLA.training.train_starvla][INFO] - Step 2928, Loss: {'train/loss': 0.46572691202163696, 'train/grad_norm_pre_clip': 12.682642936706543, 'timing/data': 0.0012621916830539703, 'timing/model': 1.0950954230502248, 'learning_rate/qwen_vl_interface': 6.577284434378408e-06, 'learning_rate/action_model': 6.577284434378408e-05, 'epoch': 1.13}) +[2026-06-17 22:46:28,436][starVLA.training.train_starvla][INFO] - Step 2929, Loss: {'train/loss': 0.422172874212265, 'train/grad_norm_pre_clip': 9.266398429870605, 'timing/data': 0.0011646002531051636, 'timing/model': 1.1361619681119919, 'learning_rate/qwen_vl_interface': 6.5752077536863404e-06, 'learning_rate/action_model': 6.57520775368634e-05, 'epoch': 1.13}) +[2026-06-17 22:46:34,899][starVLA.training.train_starvla][INFO] - Step 2930, Loss: {'train/loss': 0.5149843096733093, 'train/grad_norm_pre_clip': 10.46779727935791, 'timing/data': 0.001218230463564396, 'timing/model': 2.7324509462341666, 'learning_rate/qwen_vl_interface': 6.573130798277277e-06, 'learning_rate/action_model': 6.573130798277277e-05, 'epoch': 1.13}) +[2026-06-17 22:46:40,156][starVLA.training.train_starvla][INFO] - Step 2931, Loss: {'train/loss': 0.434623122215271, 'train/grad_norm_pre_clip': 9.354246139526367, 'timing/data': 0.0012344792485237122, 'timing/model': 1.1967131765559316, 'learning_rate/qwen_vl_interface': 6.571053568581778e-06, 'learning_rate/action_model': 6.571053568581777e-05, 'epoch': 1.13}) +[2026-06-17 22:46:44,994][starVLA.training.train_starvla][INFO] - Step 2932, Loss: {'train/loss': 0.4719317853450775, 'train/grad_norm_pre_clip': 11.660271644592285, 'timing/data': 0.0011358344927430153, 'timing/model': 1.0618306528776884, 'learning_rate/qwen_vl_interface': 6.568976065030453e-06, 'learning_rate/action_model': 6.568976065030453e-05, 'epoch': 1.13}) +[2026-06-17 22:46:51,387][starVLA.training.train_starvla][INFO] - Step 2933, Loss: {'train/loss': 0.4180072247982025, 'train/grad_norm_pre_clip': 11.531601905822754, 'timing/data': 0.001896444708108902, 'timing/model': 1.1952811693772674, 'learning_rate/qwen_vl_interface': 6.566898288053968e-06, 'learning_rate/action_model': 6.566898288053968e-05, 'epoch': 1.13}) +[2026-06-17 22:46:57,520][starVLA.training.train_starvla][INFO] - Step 2934, Loss: {'train/loss': 0.2921030819416046, 'train/grad_norm_pre_clip': 9.322908401489258, 'timing/data': 0.0013283882290124893, 'timing/model': 1.0748305190354586, 'learning_rate/qwen_vl_interface': 6.564820238083053e-06, 'learning_rate/action_model': 6.564820238083052e-05, 'epoch': 1.13}) +[2026-06-17 22:47:02,257][starVLA.training.train_starvla][INFO] - Step 2935, Loss: {'train/loss': 0.35802868008613586, 'train/grad_norm_pre_clip': 9.67929458618164, 'timing/data': 0.0011616023257374763, 'timing/model': 1.2900345576927066, 'learning_rate/qwen_vl_interface': 6.562741915548486e-06, 'learning_rate/action_model': 6.562741915548486e-05, 'epoch': 1.13}) +[2026-06-17 22:47:10,759][starVLA.training.train_starvla][INFO] - Step 2936, Loss: {'train/loss': 0.3531591594219208, 'train/grad_norm_pre_clip': 8.541923522949219, 'timing/data': 0.0010372130200266838, 'timing/model': 4.667923699133098, 'learning_rate/qwen_vl_interface': 6.5606633208811096e-06, 'learning_rate/action_model': 6.560663320881109e-05, 'epoch': 1.13}) +[2026-06-17 22:47:23,376][starVLA.training.train_starvla][INFO] - Step 2937, Loss: {'train/loss': 0.5591035485267639, 'train/grad_norm_pre_clip': 10.969736099243164, 'timing/data': 0.0014609452337026596, 'timing/model': 5.864382463507354, 'learning_rate/qwen_vl_interface': 6.558584454511814e-06, 'learning_rate/action_model': 6.558584454511814e-05, 'epoch': 1.13}) +[2026-06-17 22:47:28,791][starVLA.training.train_starvla][INFO] - Step 2938, Loss: {'train/loss': 0.3951587975025177, 'train/grad_norm_pre_clip': 8.496307373046875, 'timing/data': 0.0011426489800214767, 'timing/model': 1.496746027842164, 'learning_rate/qwen_vl_interface': 6.556505316871555e-06, 'learning_rate/action_model': 6.556505316871554e-05, 'epoch': 1.13}) +[2026-06-17 22:47:34,037][starVLA.training.train_starvla][INFO] - Step 2939, Loss: {'train/loss': 0.7759860754013062, 'train/grad_norm_pre_clip': 11.912145614624023, 'timing/data': 0.0011548399925231934, 'timing/model': 1.2206547651439905, 'learning_rate/qwen_vl_interface': 6.554425908391337e-06, 'learning_rate/action_model': 6.554425908391336e-05, 'epoch': 1.13}) +[2026-06-17 22:47:38,825][starVLA.training.train_starvla][INFO] - Step 2940, Loss: {'train/loss': 0.4740997850894928, 'train/grad_norm_pre_clip': 13.041605949401855, 'timing/data': 0.001387670636177063, 'timing/model': 1.0926151620224118, 'learning_rate/qwen_vl_interface': 6.552346229502224e-06, 'learning_rate/action_model': 6.552346229502224e-05, 'epoch': 1.13}) +[2026-06-17 22:47:43,701][starVLA.training.train_starvla][INFO] - Step 2941, Loss: {'train/loss': 0.3152056634426117, 'train/grad_norm_pre_clip': 6.824498653411865, 'timing/data': 0.0012773992493748665, 'timing/model': 1.1024369839578867, 'learning_rate/qwen_vl_interface': 6.550266280635337e-06, 'learning_rate/action_model': 6.550266280635336e-05, 'epoch': 1.13}) +[2026-06-17 22:47:48,800][starVLA.training.train_starvla][INFO] - Step 2942, Loss: {'train/loss': 0.6919881701469421, 'train/grad_norm_pre_clip': 16.149898529052734, 'timing/data': 0.002222331240773201, 'timing/model': 1.2513341484591365, 'learning_rate/qwen_vl_interface': 6.54818606222185e-06, 'learning_rate/action_model': 6.548186062221849e-05, 'epoch': 1.13}) +[2026-06-17 22:47:54,568][starVLA.training.train_starvla][INFO] - Step 2943, Loss: {'train/loss': 0.48516952991485596, 'train/grad_norm_pre_clip': 12.50917911529541, 'timing/data': 0.0011587776243686676, 'timing/model': 1.3900355314835906, 'learning_rate/qwen_vl_interface': 6.546105574692997e-06, 'learning_rate/action_model': 6.546105574692997e-05, 'epoch': 1.13}) +[2026-06-17 22:48:02,587][starVLA.training.train_starvla][INFO] - Step 2944, Loss: {'train/loss': 0.440592497587204, 'train/grad_norm_pre_clip': 8.93485164642334, 'timing/data': 0.0010075327008962631, 'timing/model': 1.0748484125360847, 'learning_rate/qwen_vl_interface': 6.544024818480061e-06, 'learning_rate/action_model': 6.544024818480062e-05, 'epoch': 1.13}) +[2026-06-17 22:48:07,613][starVLA.training.train_starvla][INFO] - Step 2945, Loss: {'train/loss': 0.35960203409194946, 'train/grad_norm_pre_clip': 7.09317684173584, 'timing/data': 0.0013868417590856552, 'timing/model': 1.1304341340437531, 'learning_rate/qwen_vl_interface': 6.5419437940143905e-06, 'learning_rate/action_model': 6.54194379401439e-05, 'epoch': 1.13}) +[2026-06-17 22:48:15,640][starVLA.training.train_starvla][INFO] - Step 2946, Loss: {'train/loss': 0.4347495436668396, 'train/grad_norm_pre_clip': 7.622096061706543, 'timing/data': 0.0011974014341831207, 'timing/model': 4.3898142939433455, 'learning_rate/qwen_vl_interface': 6.5398625017273795e-06, 'learning_rate/action_model': 6.539862501727379e-05, 'epoch': 1.13}) +[2026-06-17 22:48:20,488][starVLA.training.train_starvla][INFO] - Step 2947, Loss: {'train/loss': 0.3689616918563843, 'train/grad_norm_pre_clip': 7.896643161773682, 'timing/data': 0.00112835131585598, 'timing/model': 1.7037383299320936, 'learning_rate/qwen_vl_interface': 6.537780942050486e-06, 'learning_rate/action_model': 6.537780942050486e-05, 'epoch': 1.13}) +[2026-06-17 22:48:25,128][starVLA.training.train_starvla][INFO] - Step 2948, Loss: {'train/loss': 0.5960493087768555, 'train/grad_norm_pre_clip': 11.59746265411377, 'timing/data': 0.0009213080629706383, 'timing/model': 1.2040063822641969, 'learning_rate/qwen_vl_interface': 6.535699115415216e-06, 'learning_rate/action_model': 6.535699115415216e-05, 'epoch': 1.13}) +[2026-06-17 22:48:31,400][starVLA.training.train_starvla][INFO] - Step 2949, Loss: {'train/loss': 0.289949506521225, 'train/grad_norm_pre_clip': 6.939723968505859, 'timing/data': 0.0015450697392225266, 'timing/model': 1.1097469152882695, 'learning_rate/qwen_vl_interface': 6.533617022253137e-06, 'learning_rate/action_model': 6.533617022253137e-05, 'epoch': 1.13}) +[2026-06-17 22:48:36,091][starVLA.training.train_starvla][INFO] - Step 2950, Loss: {'train/loss': 0.261215478181839, 'train/grad_norm_pre_clip': 6.25858736038208, 'timing/data': 0.0013200575485825539, 'timing/model': 1.109387150965631, 'learning_rate/qwen_vl_interface': 6.531534662995866e-06, 'learning_rate/action_model': 6.531534662995866e-05, 'epoch': 1.13}) +[2026-06-17 22:48:47,313][starVLA.training.train_starvla][INFO] - Step 2951, Loss: {'train/loss': 0.32714104652404785, 'train/grad_norm_pre_clip': 8.261083602905273, 'timing/data': 0.0011369260028004646, 'timing/model': 1.4703680174425244, 'learning_rate/qwen_vl_interface': 6.529452038075082e-06, 'learning_rate/action_model': 6.529452038075081e-05, 'epoch': 1.14}) +[2026-06-17 22:48:52,036][starVLA.training.train_starvla][INFO] - Step 2952, Loss: {'train/loss': 0.7290112972259521, 'train/grad_norm_pre_clip': 13.451925277709961, 'timing/data': 0.0013571828603744507, 'timing/model': 1.1687235264107585, 'learning_rate/qwen_vl_interface': 6.5273691479225124e-06, 'learning_rate/action_model': 6.527369147922511e-05, 'epoch': 1.14}) +[2026-06-17 22:49:02,237][starVLA.training.train_starvla][INFO] - Step 2953, Loss: {'train/loss': 0.49920496344566345, 'train/grad_norm_pre_clip': 8.761446952819824, 'timing/data': 0.0011504869908094406, 'timing/model': 1.258638909086585, 'learning_rate/qwen_vl_interface': 6.525285992969945e-06, 'learning_rate/action_model': 6.525285992969945e-05, 'epoch': 1.14}) +[2026-06-17 22:49:07,645][starVLA.training.train_starvla][INFO] - Step 2954, Loss: {'train/loss': 0.3036760985851288, 'train/grad_norm_pre_clip': 8.505157470703125, 'timing/data': 0.0012065749615430832, 'timing/model': 1.0792528446763754, 'learning_rate/qwen_vl_interface': 6.523202573649219e-06, 'learning_rate/action_model': 6.523202573649219e-05, 'epoch': 1.14}) +[2026-06-17 22:49:12,825][starVLA.training.train_starvla][INFO] - Step 2955, Loss: {'train/loss': 0.32472777366638184, 'train/grad_norm_pre_clip': 8.822639465332031, 'timing/data': 0.001355564221739769, 'timing/model': 1.247476422227919, 'learning_rate/qwen_vl_interface': 6.5211188903922294e-06, 'learning_rate/action_model': 6.52111889039223e-05, 'epoch': 1.14}) +[2026-06-17 22:49:17,666][starVLA.training.train_starvla][INFO] - Step 2956, Loss: {'train/loss': 0.314250648021698, 'train/grad_norm_pre_clip': 6.930123805999756, 'timing/data': 0.0012148162350058556, 'timing/model': 1.2761511243879795, 'learning_rate/qwen_vl_interface': 6.519034943630926e-06, 'learning_rate/action_model': 6.519034943630927e-05, 'epoch': 1.14}) +[2026-06-17 22:49:26,566][starVLA.training.train_starvla][INFO] - Step 2957, Loss: {'train/loss': 0.6163544654846191, 'train/grad_norm_pre_clip': 10.05606460571289, 'timing/data': 0.0011441279202699661, 'timing/model': 5.706621324643493, 'learning_rate/qwen_vl_interface': 6.5169507337973125e-06, 'learning_rate/action_model': 6.516950733797312e-05, 'epoch': 1.14}) +[2026-06-17 22:49:31,408][starVLA.training.train_starvla][INFO] - Step 2958, Loss: {'train/loss': 0.8199678063392639, 'train/grad_norm_pre_clip': 13.02624797821045, 'timing/data': 0.0011312365531921387, 'timing/model': 1.2255437895655632, 'learning_rate/qwen_vl_interface': 6.51486626132345e-06, 'learning_rate/action_model': 6.51486626132345e-05, 'epoch': 1.14}) +[2026-06-17 22:49:38,396][starVLA.training.train_starvla][INFO] - Step 2959, Loss: {'train/loss': 0.4067337214946747, 'train/grad_norm_pre_clip': 9.153889656066895, 'timing/data': 0.00082387775182724, 'timing/model': 1.3328020088374615, 'learning_rate/qwen_vl_interface': 6.512781526641449e-06, 'learning_rate/action_model': 6.512781526641449e-05, 'epoch': 1.14}) +[2026-06-17 22:49:43,143][starVLA.training.train_starvla][INFO] - Step 2960, Loss: {'train/loss': 0.21560591459274292, 'train/grad_norm_pre_clip': 6.479280948638916, 'timing/data': 0.001273825764656067, 'timing/model': 1.09221145324409, 'learning_rate/qwen_vl_interface': 6.51069653018348e-06, 'learning_rate/action_model': 6.51069653018348e-05, 'epoch': 1.14}) +[2026-06-17 22:49:47,993][starVLA.training.train_starvla][INFO] - Step 2961, Loss: {'train/loss': 0.33528855443000793, 'train/grad_norm_pre_clip': 9.210094451904297, 'timing/data': 0.000931951217353344, 'timing/model': 1.0796831250190735, 'learning_rate/qwen_vl_interface': 6.508611272381763e-06, 'learning_rate/action_model': 6.508611272381763e-05, 'epoch': 1.14}) +[2026-06-17 22:49:52,374][starVLA.training.train_starvla][INFO] - Step 2962, Loss: {'train/loss': 0.46285930275917053, 'train/grad_norm_pre_clip': 13.303529739379883, 'timing/data': 0.001137295737862587, 'timing/model': 1.2051473408937454, 'learning_rate/qwen_vl_interface': 6.506525753668576e-06, 'learning_rate/action_model': 6.506525753668576e-05, 'epoch': 1.14}) +[2026-06-17 22:49:57,091][starVLA.training.train_starvla][INFO] - Step 2963, Loss: {'train/loss': 0.6246224045753479, 'train/grad_norm_pre_clip': 14.975373268127441, 'timing/data': 0.000775706022977829, 'timing/model': 1.1228838665410876, 'learning_rate/qwen_vl_interface': 6.504439974476246e-06, 'learning_rate/action_model': 6.504439974476246e-05, 'epoch': 1.14}) +[2026-06-17 22:50:09,788][starVLA.training.train_starvla][INFO] - Step 2964, Loss: {'train/loss': 0.43396109342575073, 'train/grad_norm_pre_clip': 23.517574310302734, 'timing/data': 0.0014347406104207039, 'timing/model': 1.5185035094618797, 'learning_rate/qwen_vl_interface': 6.50235393523716e-06, 'learning_rate/action_model': 6.502353935237159e-05, 'epoch': 1.14}) +[2026-06-17 22:50:14,770][starVLA.training.train_starvla][INFO] - Step 2965, Loss: {'train/loss': 0.8417869806289673, 'train/grad_norm_pre_clip': 12.4725923538208, 'timing/data': 0.0011904053390026093, 'timing/model': 1.377800015732646, 'learning_rate/qwen_vl_interface': 6.500267636383758e-06, 'learning_rate/action_model': 6.500267636383758e-05, 'epoch': 1.14}) +[2026-06-17 22:50:31,060][starVLA.training.train_starvla][INFO] - Step 2966, Loss: {'train/loss': 0.3604786694049835, 'train/grad_norm_pre_clip': 11.902810096740723, 'timing/data': 0.0011676037684082985, 'timing/model': 1.294299803674221, 'learning_rate/qwen_vl_interface': 6.498181078348527e-06, 'learning_rate/action_model': 6.498181078348527e-05, 'epoch': 1.14}) +[2026-06-17 22:50:36,111][starVLA.training.train_starvla][INFO] - Step 2967, Loss: {'train/loss': 0.46511954069137573, 'train/grad_norm_pre_clip': 9.173821449279785, 'timing/data': 0.0011671129614114761, 'timing/model': 1.6244521420449018, 'learning_rate/qwen_vl_interface': 6.496094261564018e-06, 'learning_rate/action_model': 6.496094261564018e-05, 'epoch': 1.14}) +[2026-06-17 22:50:40,309][starVLA.training.train_starvla][INFO] - Step 2968, Loss: {'train/loss': 0.34024786949157715, 'train/grad_norm_pre_clip': 8.044332504272461, 'timing/data': 0.0016230149194598198, 'timing/model': 1.1053350763395429, 'learning_rate/qwen_vl_interface': 6.494007186462824e-06, 'learning_rate/action_model': 6.494007186462823e-05, 'epoch': 1.14}) +[2026-06-17 22:50:44,862][starVLA.training.train_starvla][INFO] - Step 2969, Loss: {'train/loss': 0.31421905755996704, 'train/grad_norm_pre_clip': 6.9586381912231445, 'timing/data': 0.0011503444984555244, 'timing/model': 1.1029810328036547, 'learning_rate/qwen_vl_interface': 6.491919853477604e-06, 'learning_rate/action_model': 6.491919853477604e-05, 'epoch': 1.14}) +[2026-06-17 22:50:51,427][starVLA.training.train_starvla][INFO] - Step 2970, Loss: {'train/loss': 0.6380734443664551, 'train/grad_norm_pre_clip': 12.02167797088623, 'timing/data': 0.001132383942604065, 'timing/model': 1.0738973068073392, 'learning_rate/qwen_vl_interface': 6.4898322630410615e-06, 'learning_rate/action_model': 6.489832263041062e-05, 'epoch': 1.14}) +[2026-06-17 22:50:56,155][starVLA.training.train_starvla][INFO] - Step 2971, Loss: {'train/loss': 0.17924125492572784, 'train/grad_norm_pre_clip': 7.0467448234558105, 'timing/data': 0.0009370669722557068, 'timing/model': 1.1032446585595608, 'learning_rate/qwen_vl_interface': 6.487744415585958e-06, 'learning_rate/action_model': 6.487744415585958e-05, 'epoch': 1.14}) +[2026-06-17 22:51:00,782][starVLA.training.train_starvla][INFO] - Step 2972, Loss: {'train/loss': 0.4023604989051819, 'train/grad_norm_pre_clip': 10.750171661376953, 'timing/data': 0.001243196427822113, 'timing/model': 1.1882401816546917, 'learning_rate/qwen_vl_interface': 6.485656311545103e-06, 'learning_rate/action_model': 6.485656311545103e-05, 'epoch': 1.14}) +[2026-06-17 22:51:09,046][starVLA.training.train_starvla][INFO] - Step 2973, Loss: {'train/loss': 0.20052742958068848, 'train/grad_norm_pre_clip': 6.889756202697754, 'timing/data': 0.0015427758917212486, 'timing/model': 4.63834095094353, 'learning_rate/qwen_vl_interface': 6.483567951351366e-06, 'learning_rate/action_model': 6.483567951351365e-05, 'epoch': 1.14}) +[2026-06-17 22:51:14,356][starVLA.training.train_starvla][INFO] - Step 2974, Loss: {'train/loss': 0.4385052025318146, 'train/grad_norm_pre_clip': 6.4696431159973145, 'timing/data': 0.0013289107009768486, 'timing/model': 1.3706461554393172, 'learning_rate/qwen_vl_interface': 6.481479335437667e-06, 'learning_rate/action_model': 6.481479335437667e-05, 'epoch': 1.14}) +[2026-06-17 22:51:19,614][starVLA.training.train_starvla][INFO] - Step 2975, Loss: {'train/loss': 0.40792328119277954, 'train/grad_norm_pre_clip': 10.514674186706543, 'timing/data': 0.0014957934617996216, 'timing/model': 1.0925186024978757, 'learning_rate/qwen_vl_interface': 6.479390464236974e-06, 'learning_rate/action_model': 6.479390464236974e-05, 'epoch': 1.14}) +[2026-06-17 22:51:24,855][starVLA.training.train_starvla][INFO] - Step 2976, Loss: {'train/loss': 0.3649303615093231, 'train/grad_norm_pre_clip': 10.5443696975708, 'timing/data': 0.0008842283859848976, 'timing/model': 1.1715382542461157, 'learning_rate/qwen_vl_interface': 6.477301338182317e-06, 'learning_rate/action_model': 6.477301338182317e-05, 'epoch': 1.14}) +[2026-06-17 22:51:30,077][starVLA.training.train_starvla][INFO] - Step 2977, Loss: {'train/loss': 0.4263289272785187, 'train/grad_norm_pre_clip': 8.8250093460083, 'timing/data': 0.001316639594733715, 'timing/model': 1.7098017409443855, 'learning_rate/qwen_vl_interface': 6.47521195770677e-06, 'learning_rate/action_model': 6.47521195770677e-05, 'epoch': 1.15}) +[2026-06-17 22:51:34,494][starVLA.training.train_starvla][INFO] - Step 2978, Loss: {'train/loss': 0.5552293062210083, 'train/grad_norm_pre_clip': 8.237494468688965, 'timing/data': 0.0011251606047153473, 'timing/model': 1.0745710264891386, 'learning_rate/qwen_vl_interface': 6.473122323243467e-06, 'learning_rate/action_model': 6.473122323243467e-05, 'epoch': 1.15}) +[2026-06-17 22:51:41,187][starVLA.training.train_starvla][INFO] - Step 2979, Loss: {'train/loss': 0.40811029076576233, 'train/grad_norm_pre_clip': 7.625417232513428, 'timing/data': 0.0014100996777415276, 'timing/model': 3.1823367876932025, 'learning_rate/qwen_vl_interface': 6.471032435225591e-06, 'learning_rate/action_model': 6.47103243522559e-05, 'epoch': 1.15}) +[2026-06-17 22:51:46,101][starVLA.training.train_starvla][INFO] - Step 2980, Loss: {'train/loss': 0.3555522561073303, 'train/grad_norm_pre_clip': 10.779467582702637, 'timing/data': 0.000970982015132904, 'timing/model': 1.0728183975443244, 'learning_rate/qwen_vl_interface': 6.468942294086376e-06, 'learning_rate/action_model': 6.468942294086376e-05, 'epoch': 1.15}) +[2026-06-17 22:51:55,413][starVLA.training.train_starvla][INFO] - Step 2981, Loss: {'train/loss': 0.40645071864128113, 'train/grad_norm_pre_clip': 7.678371906280518, 'timing/data': 0.0010021375492215157, 'timing/model': 5.701404634863138, 'learning_rate/qwen_vl_interface': 6.466851900259113e-06, 'learning_rate/action_model': 6.466851900259112e-05, 'epoch': 1.15}) +[2026-06-17 22:52:09,835][starVLA.training.train_starvla][INFO] - Step 2982, Loss: {'train/loss': 0.67514967918396, 'train/grad_norm_pre_clip': 15.626775741577148, 'timing/data': 0.0010418733581900597, 'timing/model': 1.1969685470685363, 'learning_rate/qwen_vl_interface': 6.464761254177139e-06, 'learning_rate/action_model': 6.464761254177138e-05, 'epoch': 1.15}) +[2026-06-17 22:52:15,045][starVLA.training.train_starvla][INFO] - Step 2983, Loss: {'train/loss': 0.21046166121959686, 'train/grad_norm_pre_clip': 6.884722709655762, 'timing/data': 0.0012244600802659988, 'timing/model': 1.0976520515978336, 'learning_rate/qwen_vl_interface': 6.462670356273852e-06, 'learning_rate/action_model': 6.462670356273852e-05, 'epoch': 1.15}) +[2026-06-17 22:52:20,411][starVLA.training.train_starvla][INFO] - Step 2984, Loss: {'train/loss': 0.45079097151756287, 'train/grad_norm_pre_clip': 9.148321151733398, 'timing/data': 0.0011851638555526733, 'timing/model': 1.079885745421052, 'learning_rate/qwen_vl_interface': 6.460579206982695e-06, 'learning_rate/action_model': 6.460579206982696e-05, 'epoch': 1.15}) +[2026-06-17 22:52:25,176][starVLA.training.train_starvla][INFO] - Step 2985, Loss: {'train/loss': 0.4453776180744171, 'train/grad_norm_pre_clip': 10.677515983581543, 'timing/data': 0.00111472699791193, 'timing/model': 1.1119919857010245, 'learning_rate/qwen_vl_interface': 6.458487806737166e-06, 'learning_rate/action_model': 6.458487806737165e-05, 'epoch': 1.15}) +[2026-06-17 22:52:29,923][starVLA.training.train_starvla][INFO] - Step 2986, Loss: {'train/loss': 0.6501078009605408, 'train/grad_norm_pre_clip': 14.081283569335938, 'timing/data': 0.0014845840632915497, 'timing/model': 1.0989487944170833, 'learning_rate/qwen_vl_interface': 6.456396155970814e-06, 'learning_rate/action_model': 6.456396155970814e-05, 'epoch': 1.15}) +[2026-06-17 22:52:36,991][starVLA.training.train_starvla][INFO] - Step 2987, Loss: {'train/loss': 0.6911870241165161, 'train/grad_norm_pre_clip': 10.353472709655762, 'timing/data': 0.0012774309143424034, 'timing/model': 2.542527454905212, 'learning_rate/qwen_vl_interface': 6.4543042551172404e-06, 'learning_rate/action_model': 6.45430425511724e-05, 'epoch': 1.15}) +[2026-06-17 22:52:41,374][starVLA.training.train_starvla][INFO] - Step 2988, Loss: {'train/loss': 0.3696494996547699, 'train/grad_norm_pre_clip': 6.4172868728637695, 'timing/data': 0.0011615464463829994, 'timing/model': 1.1572050591930747, 'learning_rate/qwen_vl_interface': 6.4522121046101e-06, 'learning_rate/action_model': 6.452212104610099e-05, 'epoch': 1.15}) +[2026-06-17 22:52:46,325][starVLA.training.train_starvla][INFO] - Step 2989, Loss: {'train/loss': 0.4178842008113861, 'train/grad_norm_pre_clip': 10.470122337341309, 'timing/data': 0.0014148876070976257, 'timing/model': 1.0796026596799493, 'learning_rate/qwen_vl_interface': 6.450119704883095e-06, 'learning_rate/action_model': 6.450119704883096e-05, 'epoch': 1.15}) +[2026-06-17 22:52:51,419][starVLA.training.train_starvla][INFO] - Step 2990, Loss: {'train/loss': 0.406954824924469, 'train/grad_norm_pre_clip': 11.316104888916016, 'timing/data': 0.0012348676100373268, 'timing/model': 1.211603864096105, 'learning_rate/qwen_vl_interface': 6.448027056369983e-06, 'learning_rate/action_model': 6.448027056369983e-05, 'epoch': 1.15}) +[2026-06-17 22:52:56,186][starVLA.training.train_starvla][INFO] - Step 2991, Loss: {'train/loss': 0.32907694578170776, 'train/grad_norm_pre_clip': 7.620418071746826, 'timing/data': 0.0012992415577173233, 'timing/model': 1.1831730548292398, 'learning_rate/qwen_vl_interface': 6.445934159504576e-06, 'learning_rate/action_model': 6.445934159504576e-05, 'epoch': 1.15}) +[2026-06-17 22:53:00,834][starVLA.training.train_starvla][INFO] - Step 2992, Loss: {'train/loss': 0.41585594415664673, 'train/grad_norm_pre_clip': 9.674281120300293, 'timing/data': 0.0012716799974441528, 'timing/model': 1.129441517405212, 'learning_rate/qwen_vl_interface': 6.44384101472073e-06, 'learning_rate/action_model': 6.443841014720729e-05, 'epoch': 1.15}) +[2026-06-17 22:53:05,599][starVLA.training.train_starvla][INFO] - Step 2993, Loss: {'train/loss': 0.6117128133773804, 'train/grad_norm_pre_clip': 14.662225723266602, 'timing/data': 0.0016459403559565544, 'timing/model': 1.2219664035364985, 'learning_rate/qwen_vl_interface': 6.441747622452357e-06, 'learning_rate/action_model': 6.441747622452356e-05, 'epoch': 1.15}) +[2026-06-17 22:53:18,453][starVLA.training.train_starvla][INFO] - Step 2994, Loss: {'train/loss': 0.3282254636287689, 'train/grad_norm_pre_clip': 8.006176948547363, 'timing/data': 0.0015013916417956352, 'timing/model': 1.4588534301146865, 'learning_rate/qwen_vl_interface': 6.4396539831334195e-06, 'learning_rate/action_model': 6.439653983133419e-05, 'epoch': 1.15}) +[2026-06-17 22:53:23,639][starVLA.training.train_starvla][INFO] - Step 2995, Loss: {'train/loss': 0.42082178592681885, 'train/grad_norm_pre_clip': 10.094649314880371, 'timing/data': 0.0008953306823968887, 'timing/model': 1.1032213475555182, 'learning_rate/qwen_vl_interface': 6.437560097197928e-06, 'learning_rate/action_model': 6.437560097197927e-05, 'epoch': 1.15}) +[2026-06-17 22:53:36,762][starVLA.training.train_starvla][INFO] - Step 2996, Loss: {'train/loss': 0.4246256351470947, 'train/grad_norm_pre_clip': 12.240019798278809, 'timing/data': 0.0011543622240424156, 'timing/model': 1.4114224128425121, 'learning_rate/qwen_vl_interface': 6.4354659650799546e-06, 'learning_rate/action_model': 6.435465965079955e-05, 'epoch': 1.15}) +[2026-06-17 22:53:43,486][starVLA.training.train_starvla][INFO] - Step 2997, Loss: {'train/loss': 0.5725749135017395, 'train/grad_norm_pre_clip': 13.488531112670898, 'timing/data': 0.0014469986781477928, 'timing/model': 3.0785419810563326, 'learning_rate/qwen_vl_interface': 6.433371587213609e-06, 'learning_rate/action_model': 6.433371587213607e-05, 'epoch': 1.15}) +[2026-06-17 22:53:49,636][starVLA.training.train_starvla][INFO] - Step 2998, Loss: {'train/loss': 0.29114773869514465, 'train/grad_norm_pre_clip': 9.258466720581055, 'timing/data': 0.0012246258556842804, 'timing/model': 1.223767594434321, 'learning_rate/qwen_vl_interface': 6.431276964033062e-06, 'learning_rate/action_model': 6.431276964033062e-05, 'epoch': 1.15}) +[2026-06-17 22:53:57,302][starVLA.training.train_starvla][INFO] - Step 2999, Loss: {'train/loss': 0.3633861839771271, 'train/grad_norm_pre_clip': 8.059042930603027, 'timing/data': 0.0008495580404996872, 'timing/model': 3.32613015267998, 'learning_rate/qwen_vl_interface': 6.429182095972526e-06, 'learning_rate/action_model': 6.429182095972526e-05, 'epoch': 1.15}) +[2026-06-17 22:58:37,803][starVLA.training.train_starvla][INFO] - Step 3000, Loss: {'train/loss': 0.6260039806365967, 'train/grad_norm_pre_clip': 11.297232627868652, 'eval/loss': 0.5044286847114563, 'eval/latency_4/loss': 0.5044286847114563, 'eval/demon_fire_precision': 0.9570263488080301, 'eval/demon_fire_recall': 0.9909061383566092, 'eval/demon_fire_f1': 0.9736716132120632, 'eval/demon_fire_rate_model': 0.9362701908957416, 'eval/demon_fire_rate_teacher': 0.9042584434654919, 'eval/demon_left_precision': 0.8149466192170819, 'eval/demon_left_recall': 0.5228310502283106, 'eval/demon_left_f1': 0.6369958275382476, 'eval/demon_left_rate_model': 0.08252569750367107, 'eval/demon_left_rate_teacher': 0.12863436123348018, 'eval/demon_right_precision': 0.746218487394958, 'eval/demon_right_recall': 0.8014440433212996, 'eval/demon_right_f1': 0.772845953002611, 'eval/demon_right_rate_model': 0.17474302496328928, 'eval/demon_right_rate_teacher': 0.16270190895741557, 'eval/demon_move_f1': 0.7049208902704294, 'eval/demon_cc_f1': 0.873737969250583, 'eval/latency_4/demon_fire_precision': 0.9570263488080301, 'eval/latency_4/demon_fire_recall': 0.9909061383566092, 'eval/latency_4/demon_fire_f1': 0.9736716132120632, 'eval/latency_4/demon_fire_rate_model': 0.9362701908957416, 'eval/latency_4/demon_fire_rate_teacher': 0.9042584434654919, 'eval/latency_4/demon_left_precision': 0.8149466192170819, 'eval/latency_4/demon_left_recall': 0.5228310502283106, 'eval/latency_4/demon_left_f1': 0.6369958275382476, 'eval/latency_4/demon_left_rate_model': 0.08252569750367107, 'eval/latency_4/demon_left_rate_teacher': 0.12863436123348018, 'eval/latency_4/demon_right_precision': 0.746218487394958, 'eval/latency_4/demon_right_recall': 0.8014440433212996, 'eval/latency_4/demon_right_f1': 0.772845953002611, 'eval/latency_4/demon_right_rate_model': 0.17474302496328928, 'eval/latency_4/demon_right_rate_teacher': 0.16270190895741557, 'eval/latency_4/demon_move_f1': 0.7049208902704294, 'eval/latency_4/demon_cc_f1': 0.873737969250583, 'timing/data': 0.001459154300391674, 'timing/model': 1.3727752324193716, 'learning_rate/qwen_vl_interface': 6.427086983466275e-06, 'learning_rate/action_model': 6.427086983466275e-05, 'epoch': 1.15}) +[2026-06-17 22:58:37,827][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_3000_state +[2026-06-17 22:58:37,836][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 22:58:46,983][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_3000_state/model.safetensors +[2026-06-17 22:59:07,143][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_3000_state/optimizer.bin +[2026-06-17 22:59:07,144][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_3000_state/sampler.bin +[2026-06-17 22:59:07,222][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_3000_state/sampler_1.bin +[2026-06-17 22:59:07,224][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_3000_state/random_states_0.pkl +[2026-06-17 22:59:07,229][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3000 +[2026-06-17 22:59:07,230][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 22:59:07,247][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 22:59:11,811][starVLA.training.train_starvla][INFO] - Step 3001, Loss: {'train/loss': 0.6972932815551758, 'train/grad_norm_pre_clip': 18.337982177734375, 'timing/data': 0.0012143878266215324, 'timing/model': 1.0386962983757257, 'learning_rate/qwen_vl_interface': 6.4249916269486255e-06, 'learning_rate/action_model': 6.424991626948626e-05, 'epoch': 1.15}) +[2026-06-17 22:59:16,166][starVLA.training.train_starvla][INFO] - Step 3002, Loss: {'train/loss': 0.6370294094085693, 'train/grad_norm_pre_clip': 14.320171356201172, 'timing/data': 0.0011872882023453712, 'timing/model': 1.3565038666129112, 'learning_rate/qwen_vl_interface': 6.422896026853948e-06, 'learning_rate/action_model': 6.422896026853947e-05, 'epoch': 1.15}) +[2026-06-17 22:59:20,341][starVLA.training.train_starvla][INFO] - Step 3003, Loss: {'train/loss': 0.9062581658363342, 'train/grad_norm_pre_clip': 15.161455154418945, 'timing/data': 0.0010670851916074753, 'timing/model': 1.0570508847013116, 'learning_rate/qwen_vl_interface': 6.420800183616661e-06, 'learning_rate/action_model': 6.420800183616661e-05, 'epoch': 1.16}) +[2026-06-17 22:59:24,696][starVLA.training.train_starvla][INFO] - Step 3004, Loss: {'train/loss': 0.5864464044570923, 'train/grad_norm_pre_clip': 15.146814346313477, 'timing/data': 0.0010852990671992302, 'timing/model': 1.048693984746933, 'learning_rate/qwen_vl_interface': 6.4187040976712355e-06, 'learning_rate/action_model': 6.418704097671235e-05, 'epoch': 1.16}) +[2026-06-17 22:59:29,183][starVLA.training.train_starvla][INFO] - Step 3005, Loss: {'train/loss': 0.3799372911453247, 'train/grad_norm_pre_clip': 12.187771797180176, 'timing/data': 0.0011143861338496208, 'timing/model': 1.0902998456731439, 'learning_rate/qwen_vl_interface': 6.416607769452195e-06, 'learning_rate/action_model': 6.416607769452194e-05, 'epoch': 1.16}) +[2026-06-17 22:59:33,835][starVLA.training.train_starvla][INFO] - Step 3006, Loss: {'train/loss': 0.6474485993385315, 'train/grad_norm_pre_clip': 14.165548324584961, 'timing/data': 0.0012271981686353683, 'timing/model': 1.3100580852478743, 'learning_rate/qwen_vl_interface': 6.414511199394106e-06, 'learning_rate/action_model': 6.414511199394105e-05, 'epoch': 1.16}) +[2026-06-17 22:59:38,141][starVLA.training.train_starvla][INFO] - Step 3007, Loss: {'train/loss': 0.29800674319267273, 'train/grad_norm_pre_clip': 7.956894874572754, 'timing/data': 0.0011701863259077072, 'timing/model': 1.3206287184730172, 'learning_rate/qwen_vl_interface': 6.4124143879315935e-06, 'learning_rate/action_model': 6.412414387931593e-05, 'epoch': 1.16}) +[2026-06-17 22:59:42,196][starVLA.training.train_starvla][INFO] - Step 3008, Loss: {'train/loss': 0.38247454166412354, 'train/grad_norm_pre_clip': 7.850623607635498, 'timing/data': 0.0013179592788219452, 'timing/model': 1.0388988759368658, 'learning_rate/qwen_vl_interface': 6.410317335499328e-06, 'learning_rate/action_model': 6.410317335499328e-05, 'epoch': 1.16}) +[2026-06-17 22:59:46,552][starVLA.training.train_starvla][INFO] - Step 3009, Loss: {'train/loss': 0.6745245456695557, 'train/grad_norm_pre_clip': 12.403253555297852, 'timing/data': 0.0011657532304525375, 'timing/model': 1.0479825735092163, 'learning_rate/qwen_vl_interface': 6.40822004253203e-06, 'learning_rate/action_model': 6.40822004253203e-05, 'epoch': 1.16}) +[2026-06-17 22:59:55,230][starVLA.training.train_starvla][INFO] - Step 3010, Loss: {'train/loss': 0.4605753421783447, 'train/grad_norm_pre_clip': 8.924535751342773, 'timing/data': 0.0015389062464237213, 'timing/model': 1.7851735036820173, 'learning_rate/qwen_vl_interface': 6.406122509464471e-06, 'learning_rate/action_model': 6.40612250946447e-05, 'epoch': 1.16}) +[2026-06-17 23:00:00,687][starVLA.training.train_starvla][INFO] - Step 3011, Loss: {'train/loss': 0.2429424375295639, 'train/grad_norm_pre_clip': 6.7467145919799805, 'timing/data': 0.001475743018090725, 'timing/model': 1.2209016792476177, 'learning_rate/qwen_vl_interface': 6.404024736731472e-06, 'learning_rate/action_model': 6.404024736731472e-05, 'epoch': 1.16}) +[2026-06-17 23:00:06,077][starVLA.training.train_starvla][INFO] - Step 3012, Loss: {'train/loss': 0.4924757778644562, 'train/grad_norm_pre_clip': 12.70364761352539, 'timing/data': 0.0011951066553592682, 'timing/model': 1.738698442466557, 'learning_rate/qwen_vl_interface': 6.401926724767903e-06, 'learning_rate/action_model': 6.401926724767902e-05, 'epoch': 1.16}) +[2026-06-17 23:00:11,710][starVLA.training.train_starvla][INFO] - Step 3013, Loss: {'train/loss': 0.4741067886352539, 'train/grad_norm_pre_clip': 8.7914400100708, 'timing/data': 0.0011754631996154785, 'timing/model': 1.2293686643242836, 'learning_rate/qwen_vl_interface': 6.399828474008683e-06, 'learning_rate/action_model': 6.399828474008682e-05, 'epoch': 1.16}) +[2026-06-17 23:00:16,725][starVLA.training.train_starvla][INFO] - Step 3014, Loss: {'train/loss': 0.19296519458293915, 'train/grad_norm_pre_clip': 7.360339641571045, 'timing/data': 0.0012146783992648125, 'timing/model': 1.2631083028391004, 'learning_rate/qwen_vl_interface': 6.397729984888785e-06, 'learning_rate/action_model': 6.397729984888785e-05, 'epoch': 1.16}) +[2026-06-17 23:00:22,029][starVLA.training.train_starvla][INFO] - Step 3015, Loss: {'train/loss': 0.44080305099487305, 'train/grad_norm_pre_clip': 8.289250373840332, 'timing/data': 0.0013403641059994698, 'timing/model': 1.1017216062173247, 'learning_rate/qwen_vl_interface': 6.395631257843224e-06, 'learning_rate/action_model': 6.395631257843224e-05, 'epoch': 1.16}) +[2026-06-17 23:00:27,040][starVLA.training.train_starvla][INFO] - Step 3016, Loss: {'train/loss': 0.44785410165786743, 'train/grad_norm_pre_clip': 11.414637565612793, 'timing/data': 0.0013438072055578232, 'timing/model': 1.596064711920917, 'learning_rate/qwen_vl_interface': 6.393532293307073e-06, 'learning_rate/action_model': 6.393532293307072e-05, 'epoch': 1.16}) +[2026-06-17 23:00:35,702][starVLA.training.train_starvla][INFO] - Step 3017, Loss: {'train/loss': 0.7366512417793274, 'train/grad_norm_pre_clip': 12.081318855285645, 'timing/data': 0.0010739108547568321, 'timing/model': 3.1475197449326515, 'learning_rate/qwen_vl_interface': 6.391433091715447e-06, 'learning_rate/action_model': 6.391433091715446e-05, 'epoch': 1.16}) +[2026-06-17 23:00:43,203][starVLA.training.train_starvla][INFO] - Step 3018, Loss: {'train/loss': 0.36540430784225464, 'train/grad_norm_pre_clip': 7.492925643920898, 'timing/data': 0.0014249598607420921, 'timing/model': 2.697294524870813, 'learning_rate/qwen_vl_interface': 6.389333653503511e-06, 'learning_rate/action_model': 6.38933365350351e-05, 'epoch': 1.16}) +[2026-06-17 23:00:49,352][starVLA.training.train_starvla][INFO] - Step 3019, Loss: {'train/loss': 0.4659760892391205, 'train/grad_norm_pre_clip': 12.380315780639648, 'timing/data': 0.0015013013035058975, 'timing/model': 1.1215589800849557, 'learning_rate/qwen_vl_interface': 6.387233979106484e-06, 'learning_rate/action_model': 6.387233979106483e-05, 'epoch': 1.16}) +[2026-06-17 23:00:55,499][starVLA.training.train_starvla][INFO] - Step 3020, Loss: {'train/loss': 0.31465399265289307, 'train/grad_norm_pre_clip': 8.280333518981934, 'timing/data': 0.012607601471245289, 'timing/model': 1.1787786232307553, 'learning_rate/qwen_vl_interface': 6.385134068959629e-06, 'learning_rate/action_model': 6.385134068959628e-05, 'epoch': 1.16}) +[2026-06-17 23:01:03,417][starVLA.training.train_starvla][INFO] - Step 3021, Loss: {'train/loss': 0.40574824810028076, 'train/grad_norm_pre_clip': 15.625710487365723, 'timing/data': 0.017388811334967613, 'timing/model': 1.2439209930598736, 'learning_rate/qwen_vl_interface': 6.3830339234982596e-06, 'learning_rate/action_model': 6.38303392349826e-05, 'epoch': 1.16}) +[2026-06-17 23:01:09,178][starVLA.training.train_starvla][INFO] - Step 3022, Loss: {'train/loss': 0.49347835779190063, 'train/grad_norm_pre_clip': 10.19224739074707, 'timing/data': 0.01894790679216385, 'timing/model': 1.9228072501718998, 'learning_rate/qwen_vl_interface': 6.38093354315774e-06, 'learning_rate/action_model': 6.38093354315774e-05, 'epoch': 1.16}) +[2026-06-17 23:01:14,220][starVLA.training.train_starvla][INFO] - Step 3023, Loss: {'train/loss': 0.41622981429100037, 'train/grad_norm_pre_clip': 12.185347557067871, 'timing/data': 0.013896637596189976, 'timing/model': 1.2996839433908463, 'learning_rate/qwen_vl_interface': 6.378832928373479e-06, 'learning_rate/action_model': 6.378832928373478e-05, 'epoch': 1.16}) +[2026-06-17 23:01:22,521][starVLA.training.train_starvla][INFO] - Step 3024, Loss: {'train/loss': 0.28176525235176086, 'train/grad_norm_pre_clip': 6.989371299743652, 'timing/data': 0.012567685917019844, 'timing/model': 1.2755326367914677, 'learning_rate/qwen_vl_interface': 6.376732079580936e-06, 'learning_rate/action_model': 6.376732079580936e-05, 'epoch': 1.16}) +[2026-06-17 23:01:29,458][starVLA.training.train_starvla][INFO] - Step 3025, Loss: {'train/loss': 0.43298470973968506, 'train/grad_norm_pre_clip': 10.517176628112793, 'timing/data': 0.012827864848077297, 'timing/model': 1.1217290116474032, 'learning_rate/qwen_vl_interface': 6.374630997215623e-06, 'learning_rate/action_model': 6.374630997215623e-05, 'epoch': 1.16}) +[2026-06-17 23:01:34,911][starVLA.training.train_starvla][INFO] - Step 3026, Loss: {'train/loss': 0.5444229245185852, 'train/grad_norm_pre_clip': 11.428556442260742, 'timing/data': 0.013387826271355152, 'timing/model': 1.5806019939482212, 'learning_rate/qwen_vl_interface': 6.372529681713093e-06, 'learning_rate/action_model': 6.372529681713094e-05, 'epoch': 1.16}) +[2026-06-17 23:01:40,425][starVLA.training.train_starvla][INFO] - Step 3027, Loss: {'train/loss': 0.462677001953125, 'train/grad_norm_pre_clip': 13.50485897064209, 'timing/data': 0.012252448126673698, 'timing/model': 1.8607285441830754, 'learning_rate/qwen_vl_interface': 6.370428133508953e-06, 'learning_rate/action_model': 6.370428133508953e-05, 'epoch': 1.16}) +[2026-06-17 23:01:45,336][starVLA.training.train_starvla][INFO] - Step 3028, Loss: {'train/loss': 0.5152285695075989, 'train/grad_norm_pre_clip': 12.043692588806152, 'timing/data': 0.011901541613042355, 'timing/model': 1.1012500831857324, 'learning_rate/qwen_vl_interface': 6.368326353038852e-06, 'learning_rate/action_model': 6.368326353038853e-05, 'epoch': 1.16}) +[2026-06-17 23:02:02,009][starVLA.training.train_starvla][INFO] - Step 3029, Loss: {'train/loss': 0.5137982964515686, 'train/grad_norm_pre_clip': 11.27457046508789, 'timing/data': 0.061241498216986656, 'timing/model': 4.7040538573637605, 'learning_rate/qwen_vl_interface': 6.366224340738499e-06, 'learning_rate/action_model': 6.366224340738499e-05, 'epoch': 1.17}) +[2026-06-17 23:02:07,500][starVLA.training.train_starvla][INFO] - Step 3030, Loss: {'train/loss': 0.29167184233665466, 'train/grad_norm_pre_clip': 6.517003059387207, 'timing/data': 0.012556743808090687, 'timing/model': 1.220548728480935, 'learning_rate/qwen_vl_interface': 6.3641220970436355e-06, 'learning_rate/action_model': 6.364122097043635e-05, 'epoch': 1.17}) +[2026-06-17 23:02:13,044][starVLA.training.train_starvla][INFO] - Step 3031, Loss: {'train/loss': 0.24442796409130096, 'train/grad_norm_pre_clip': 9.157550811767578, 'timing/data': 0.09038771875202656, 'timing/model': 1.2167111337184906, 'learning_rate/qwen_vl_interface': 6.362019622390065e-06, 'learning_rate/action_model': 6.362019622390066e-05, 'epoch': 1.17}) +[2026-06-17 23:02:23,883][starVLA.training.train_starvla][INFO] - Step 3032, Loss: {'train/loss': 0.6580246090888977, 'train/grad_norm_pre_clip': 13.488868713378906, 'timing/data': 0.014719386585056782, 'timing/model': 1.8595891958102584, 'learning_rate/qwen_vl_interface': 6.359916917213629e-06, 'learning_rate/action_model': 6.359916917213629e-05, 'epoch': 1.17}) +[2026-06-17 23:02:28,767][starVLA.training.train_starvla][INFO] - Step 3033, Loss: {'train/loss': 0.24075770378112793, 'train/grad_norm_pre_clip': 7.0475287437438965, 'timing/data': 0.011881728656589985, 'timing/model': 1.1789868976920843, 'learning_rate/qwen_vl_interface': 6.357813981950223e-06, 'learning_rate/action_model': 6.357813981950223e-05, 'epoch': 1.17}) +[2026-06-17 23:02:37,658][starVLA.training.train_starvla][INFO] - Step 3034, Loss: {'train/loss': 0.525921106338501, 'train/grad_norm_pre_clip': 9.910842895507812, 'timing/data': 0.0180073706433177, 'timing/model': 1.2277462352067232, 'learning_rate/qwen_vl_interface': 6.355710817035784e-06, 'learning_rate/action_model': 6.355710817035783e-05, 'epoch': 1.17}) +[2026-06-17 23:02:43,255][starVLA.training.train_starvla][INFO] - Step 3035, Loss: {'train/loss': 0.4730795919895172, 'train/grad_norm_pre_clip': 10.716005325317383, 'timing/data': 0.014541112817823887, 'timing/model': 1.4644853984937072, 'learning_rate/qwen_vl_interface': 6.353607422906304e-06, 'learning_rate/action_model': 6.353607422906304e-05, 'epoch': 1.17}) +[2026-06-17 23:02:52,066][starVLA.training.train_starvla][INFO] - Step 3036, Loss: {'train/loss': 0.5161805152893066, 'train/grad_norm_pre_clip': 10.745430946350098, 'timing/data': 0.011996994726359844, 'timing/model': 1.794266282580793, 'learning_rate/qwen_vl_interface': 6.351503799997819e-06, 'learning_rate/action_model': 6.351503799997818e-05, 'epoch': 1.17}) +[2026-06-17 23:02:57,359][starVLA.training.train_starvla][INFO] - Step 3037, Loss: {'train/loss': 0.4078952968120575, 'train/grad_norm_pre_clip': 8.80984115600586, 'timing/data': 0.011499215848743916, 'timing/model': 1.623846479691565, 'learning_rate/qwen_vl_interface': 6.349399948746409e-06, 'learning_rate/action_model': 6.349399948746409e-05, 'epoch': 1.17}) +[2026-06-17 23:03:04,167][starVLA.training.train_starvla][INFO] - Step 3038, Loss: {'train/loss': 0.3200024664402008, 'train/grad_norm_pre_clip': 11.565450668334961, 'timing/data': 0.012901433743536472, 'timing/model': 2.0756396017968655, 'learning_rate/qwen_vl_interface': 6.347295869588208e-06, 'learning_rate/action_model': 6.347295869588208e-05, 'epoch': 1.17}) +[2026-06-17 23:03:10,115][starVLA.training.train_starvla][INFO] - Step 3039, Loss: {'train/loss': 0.30189216136932373, 'train/grad_norm_pre_clip': 6.161222457885742, 'timing/data': 0.02325083129107952, 'timing/model': 1.1276682168245316, 'learning_rate/qwen_vl_interface': 6.345191562959392e-06, 'learning_rate/action_model': 6.345191562959391e-05, 'epoch': 1.17}) +[2026-06-17 23:03:15,532][starVLA.training.train_starvla][INFO] - Step 3040, Loss: {'train/loss': 0.557121217250824, 'train/grad_norm_pre_clip': 11.592503547668457, 'timing/data': 0.013457768596708775, 'timing/model': 1.5496367700397968, 'learning_rate/qwen_vl_interface': 6.343087029296187e-06, 'learning_rate/action_model': 6.343087029296186e-05, 'epoch': 1.17}) +[2026-06-17 23:03:20,197][starVLA.training.train_starvla][INFO] - Step 3041, Loss: {'train/loss': 0.4337485134601593, 'train/grad_norm_pre_clip': 10.67430591583252, 'timing/data': 0.011421188712120056, 'timing/model': 1.0803354540839791, 'learning_rate/qwen_vl_interface': 6.340982269034863e-06, 'learning_rate/action_model': 6.340982269034863e-05, 'epoch': 1.17}) +[2026-06-17 23:03:25,385][starVLA.training.train_starvla][INFO] - Step 3042, Loss: {'train/loss': 0.3202778995037079, 'train/grad_norm_pre_clip': 8.622326850891113, 'timing/data': 0.012723381631076336, 'timing/model': 1.4430573470890522, 'learning_rate/qwen_vl_interface': 6.3388772826117415e-06, 'learning_rate/action_model': 6.338877282611742e-05, 'epoch': 1.17}) +[2026-06-17 23:03:32,351][starVLA.training.train_starvla][INFO] - Step 3043, Loss: {'train/loss': 0.33477461338043213, 'train/grad_norm_pre_clip': 5.697251796722412, 'timing/data': 0.05976180452853441, 'timing/model': 3.2475433237850666, 'learning_rate/qwen_vl_interface': 6.3367720704631895e-06, 'learning_rate/action_model': 6.33677207046319e-05, 'epoch': 1.17}) +[2026-06-17 23:03:37,442][starVLA.training.train_starvla][INFO] - Step 3044, Loss: {'train/loss': 0.47116225957870483, 'train/grad_norm_pre_clip': 10.792929649353027, 'timing/data': 0.012527333572506905, 'timing/model': 1.117835035547614, 'learning_rate/qwen_vl_interface': 6.334666633025616e-06, 'learning_rate/action_model': 6.334666633025615e-05, 'epoch': 1.17}) +[2026-06-17 23:03:42,866][starVLA.training.train_starvla][INFO] - Step 3045, Loss: {'train/loss': 0.29965126514434814, 'train/grad_norm_pre_clip': 7.988534927368164, 'timing/data': 0.015322579070925713, 'timing/model': 1.3598717553541064, 'learning_rate/qwen_vl_interface': 6.332560970735482e-06, 'learning_rate/action_model': 6.332560970735482e-05, 'epoch': 1.17}) +[2026-06-17 23:03:55,117][starVLA.training.train_starvla][INFO] - Step 3046, Loss: {'train/loss': 0.8712384104728699, 'train/grad_norm_pre_clip': 15.458029747009277, 'timing/data': 0.013023726642131805, 'timing/model': 4.346310475841165, 'learning_rate/qwen_vl_interface': 6.3304550840292945e-06, 'learning_rate/action_model': 6.330455084029294e-05, 'epoch': 1.17}) +[2026-06-17 23:04:00,170][starVLA.training.train_starvla][INFO] - Step 3047, Loss: {'train/loss': 0.28298434615135193, 'train/grad_norm_pre_clip': 6.639747619628906, 'timing/data': 0.013021268881857395, 'timing/model': 1.5206319764256477, 'learning_rate/qwen_vl_interface': 6.328348973343605e-06, 'learning_rate/action_model': 6.328348973343605e-05, 'epoch': 1.17}) +[2026-06-17 23:04:04,757][starVLA.training.train_starvla][INFO] - Step 3048, Loss: {'train/loss': 0.3961799144744873, 'train/grad_norm_pre_clip': 6.925195217132568, 'timing/data': 0.01145361177623272, 'timing/model': 1.2367478925734758, 'learning_rate/qwen_vl_interface': 6.326242639115014e-06, 'learning_rate/action_model': 6.326242639115015e-05, 'epoch': 1.17}) +[2026-06-17 23:04:09,693][starVLA.training.train_starvla][INFO] - Step 3049, Loss: {'train/loss': 0.4756019413471222, 'train/grad_norm_pre_clip': 9.27834701538086, 'timing/data': 0.0117828119546175, 'timing/model': 1.2214429127052426, 'learning_rate/qwen_vl_interface': 6.324136081780166e-06, 'learning_rate/action_model': 6.324136081780165e-05, 'epoch': 1.17}) +[2026-06-17 23:04:14,439][starVLA.training.train_starvla][INFO] - Step 3050, Loss: {'train/loss': 0.40666794776916504, 'train/grad_norm_pre_clip': 11.331130027770996, 'timing/data': 0.013448469340801239, 'timing/model': 1.0922865215688944, 'learning_rate/qwen_vl_interface': 6.322029301775751e-06, 'learning_rate/action_model': 6.32202930177575e-05, 'epoch': 1.17}) +[2026-06-17 23:04:19,312][starVLA.training.train_starvla][INFO] - Step 3051, Loss: {'train/loss': 0.6613180041313171, 'train/grad_norm_pre_clip': 14.037384033203125, 'timing/data': 0.01517556793987751, 'timing/model': 1.071575584821403, 'learning_rate/qwen_vl_interface': 6.3199222995385105e-06, 'learning_rate/action_model': 6.31992229953851e-05, 'epoch': 1.17}) +[2026-06-17 23:04:24,347][starVLA.training.train_starvla][INFO] - Step 3052, Loss: {'train/loss': 0.5946407318115234, 'train/grad_norm_pre_clip': 9.807901382446289, 'timing/data': 0.0122707299888134, 'timing/model': 1.676960028707981, 'learning_rate/qwen_vl_interface': 6.317815075505223e-06, 'learning_rate/action_model': 6.317815075505222e-05, 'epoch': 1.17}) +[2026-06-17 23:04:28,905][starVLA.training.train_starvla][INFO] - Step 3053, Loss: {'train/loss': 0.4429584741592407, 'train/grad_norm_pre_clip': 12.414776802062988, 'timing/data': 0.013343972153961658, 'timing/model': 1.228216540068388, 'learning_rate/qwen_vl_interface': 6.315707630112724e-06, 'learning_rate/action_model': 6.315707630112723e-05, 'epoch': 1.17}) +[2026-06-17 23:04:34,243][starVLA.training.train_starvla][INFO] - Step 3054, Loss: {'train/loss': 0.21042734384536743, 'train/grad_norm_pre_clip': 5.793015956878662, 'timing/data': 0.01408468745648861, 'timing/model': 1.1001145979389548, 'learning_rate/qwen_vl_interface': 6.313599963797884e-06, 'learning_rate/action_model': 6.313599963797884e-05, 'epoch': 1.17}) +[2026-06-17 23:04:38,922][starVLA.training.train_starvla][INFO] - Step 3055, Loss: {'train/loss': 0.2740124464035034, 'train/grad_norm_pre_clip': 11.812830924987793, 'timing/data': 0.013261557556688786, 'timing/model': 1.0879201460629702, 'learning_rate/qwen_vl_interface': 6.311492076997629e-06, 'learning_rate/action_model': 6.311492076997629e-05, 'epoch': 1.18}) +[2026-06-17 23:04:43,607][starVLA.training.train_starvla][INFO] - Step 3056, Loss: {'train/loss': 0.5226309895515442, 'train/grad_norm_pre_clip': 11.048277854919434, 'timing/data': 0.011456195265054703, 'timing/model': 1.4649253180250525, 'learning_rate/qwen_vl_interface': 6.309383970148922e-06, 'learning_rate/action_model': 6.309383970148922e-05, 'epoch': 1.18}) +[2026-06-17 23:04:57,102][starVLA.training.train_starvla][INFO] - Step 3057, Loss: {'train/loss': 0.8178024888038635, 'train/grad_norm_pre_clip': 16.146024703979492, 'timing/data': 0.01192866824567318, 'timing/model': 1.4347237711772323, 'learning_rate/qwen_vl_interface': 6.3072756436887805e-06, 'learning_rate/action_model': 6.30727564368878e-05, 'epoch': 1.18}) +[2026-06-17 23:05:04,744][starVLA.training.train_starvla][INFO] - Step 3058, Loss: {'train/loss': 0.4647004008293152, 'train/grad_norm_pre_clip': 11.194989204406738, 'timing/data': 0.02337403129786253, 'timing/model': 4.155174351297319, 'learning_rate/qwen_vl_interface': 6.30516709805426e-06, 'learning_rate/action_model': 6.30516709805426e-05, 'epoch': 1.18}) +[2026-06-17 23:05:09,770][starVLA.training.train_starvla][INFO] - Step 3059, Loss: {'train/loss': 0.4964827597141266, 'train/grad_norm_pre_clip': 9.297880172729492, 'timing/data': 0.01628733891993761, 'timing/model': 1.2158418372273445, 'learning_rate/qwen_vl_interface': 6.303058333682464e-06, 'learning_rate/action_model': 6.303058333682463e-05, 'epoch': 1.18}) +[2026-06-17 23:05:15,103][starVLA.training.train_starvla][INFO] - Step 3060, Loss: {'train/loss': 0.42412033677101135, 'train/grad_norm_pre_clip': 8.852624893188477, 'timing/data': 0.011760539375245571, 'timing/model': 1.0769121423363686, 'learning_rate/qwen_vl_interface': 6.300949351010542e-06, 'learning_rate/action_model': 6.300949351010543e-05, 'epoch': 1.18}) +[2026-06-17 23:05:20,187][starVLA.training.train_starvla][INFO] - Step 3061, Loss: {'train/loss': 0.5123714208602905, 'train/grad_norm_pre_clip': 9.169960975646973, 'timing/data': 0.012066301889717579, 'timing/model': 1.1551948972046375, 'learning_rate/qwen_vl_interface': 6.298840150475689e-06, 'learning_rate/action_model': 6.298840150475689e-05, 'epoch': 1.18}) +[2026-06-17 23:05:25,215][starVLA.training.train_starvla][INFO] - Step 3062, Loss: {'train/loss': 0.21455171704292297, 'train/grad_norm_pre_clip': 5.877018928527832, 'timing/data': 0.013101045042276382, 'timing/model': 1.6751710567623377, 'learning_rate/qwen_vl_interface': 6.296730732515145e-06, 'learning_rate/action_model': 6.296730732515145e-05, 'epoch': 1.18}) +[2026-06-17 23:05:34,364][starVLA.training.train_starvla][INFO] - Step 3063, Loss: {'train/loss': 0.5988596081733704, 'train/grad_norm_pre_clip': 13.093822479248047, 'timing/data': 0.011669453233480453, 'timing/model': 4.006211725994945, 'learning_rate/qwen_vl_interface': 6.294621097566193e-06, 'learning_rate/action_model': 6.294621097566193e-05, 'epoch': 1.18}) +[2026-06-17 23:05:41,324][starVLA.training.train_starvla][INFO] - Step 3064, Loss: {'train/loss': 0.3796645998954773, 'train/grad_norm_pre_clip': 9.553851127624512, 'timing/data': 0.01300605945289135, 'timing/model': 1.3453422747552395, 'learning_rate/qwen_vl_interface': 6.292511246066163e-06, 'learning_rate/action_model': 6.292511246066163e-05, 'epoch': 1.18}) +[2026-06-17 23:05:46,062][starVLA.training.train_starvla][INFO] - Step 3065, Loss: {'train/loss': 0.3502003252506256, 'train/grad_norm_pre_clip': 6.952072620391846, 'timing/data': 0.011712931096553802, 'timing/model': 1.1045286618173122, 'learning_rate/qwen_vl_interface': 6.290401178452431e-06, 'learning_rate/action_model': 6.29040117845243e-05, 'epoch': 1.18}) +[2026-06-17 23:05:50,796][starVLA.training.train_starvla][INFO] - Step 3066, Loss: {'train/loss': 0.5413879156112671, 'train/grad_norm_pre_clip': 10.3848237991333, 'timing/data': 0.011700421571731567, 'timing/model': 1.4078360497951508, 'learning_rate/qwen_vl_interface': 6.288290895162414e-06, 'learning_rate/action_model': 6.288290895162413e-05, 'epoch': 1.18}) +[2026-06-17 23:05:59,065][starVLA.training.train_starvla][INFO] - Step 3067, Loss: {'train/loss': 0.1862114667892456, 'train/grad_norm_pre_clip': 5.971040725708008, 'timing/data': 0.011401497758924961, 'timing/model': 3.097775649279356, 'learning_rate/qwen_vl_interface': 6.286180396633579e-06, 'learning_rate/action_model': 6.286180396633579e-05, 'epoch': 1.18}) +[2026-06-17 23:06:06,992][starVLA.training.train_starvla][INFO] - Step 3068, Loss: {'train/loss': 0.5610740780830383, 'train/grad_norm_pre_clip': 14.495487213134766, 'timing/data': 0.012559373863041401, 'timing/model': 1.2086473982781172, 'learning_rate/qwen_vl_interface': 6.28406968330343e-06, 'learning_rate/action_model': 6.28406968330343e-05, 'epoch': 1.18}) +[2026-06-17 23:06:12,411][starVLA.training.train_starvla][INFO] - Step 3069, Loss: {'train/loss': 0.778931736946106, 'train/grad_norm_pre_clip': 15.74487018585205, 'timing/data': 0.01565409731119871, 'timing/model': 1.0938671007752419, 'learning_rate/qwen_vl_interface': 6.281958755609526e-06, 'learning_rate/action_model': 6.281958755609526e-05, 'epoch': 1.18}) +[2026-06-17 23:06:17,282][starVLA.training.train_starvla][INFO] - Step 3070, Loss: {'train/loss': 0.6515246629714966, 'train/grad_norm_pre_clip': 12.696681022644043, 'timing/data': 0.011588472872972488, 'timing/model': 1.0847952766343951, 'learning_rate/qwen_vl_interface': 6.279847613989459e-06, 'learning_rate/action_model': 6.279847613989459e-05, 'epoch': 1.18}) +[2026-06-17 23:06:26,892][starVLA.training.train_starvla][INFO] - Step 3071, Loss: {'train/loss': 0.24438339471817017, 'train/grad_norm_pre_clip': 7.640949249267578, 'timing/data': 0.014140551909804344, 'timing/model': 1.3491911245509982, 'learning_rate/qwen_vl_interface': 6.2777362588808745e-06, 'learning_rate/action_model': 6.277736258880874e-05, 'epoch': 1.18}) +[2026-06-17 23:06:35,296][starVLA.training.train_starvla][INFO] - Step 3072, Loss: {'train/loss': 0.4012269973754883, 'train/grad_norm_pre_clip': 22.16424560546875, 'timing/data': 0.012273096479475498, 'timing/model': 1.484209437854588, 'learning_rate/qwen_vl_interface': 6.275624690721458e-06, 'learning_rate/action_model': 6.275624690721458e-05, 'epoch': 1.18}) +[2026-06-17 23:06:46,101][starVLA.training.train_starvla][INFO] - Step 3073, Loss: {'train/loss': 0.3233150839805603, 'train/grad_norm_pre_clip': 8.583954811096191, 'timing/data': 0.013361715711653233, 'timing/model': 1.2529770405963063, 'learning_rate/qwen_vl_interface': 6.27351290994894e-06, 'learning_rate/action_model': 6.27351290994894e-05, 'epoch': 1.18}) +[2026-06-17 23:06:51,422][starVLA.training.train_starvla][INFO] - Step 3074, Loss: {'train/loss': 0.35262051224708557, 'train/grad_norm_pre_clip': 9.268059730529785, 'timing/data': 0.01232738234102726, 'timing/model': 1.4304095227271318, 'learning_rate/qwen_vl_interface': 6.271400917001092e-06, 'learning_rate/action_model': 6.271400917001092e-05, 'epoch': 1.18}) +[2026-06-17 23:07:01,910][starVLA.training.train_starvla][INFO] - Step 3075, Loss: {'train/loss': 0.24508725106716156, 'train/grad_norm_pre_clip': 6.501260757446289, 'timing/data': 0.012263582088053226, 'timing/model': 5.0272684674710035, 'learning_rate/qwen_vl_interface': 6.269288712315737e-06, 'learning_rate/action_model': 6.269288712315737e-05, 'epoch': 1.18}) +[2026-06-17 23:07:11,836][starVLA.training.train_starvla][INFO] - Step 3076, Loss: {'train/loss': 0.5633099675178528, 'train/grad_norm_pre_clip': 12.162020683288574, 'timing/data': 0.014373479411005974, 'timing/model': 1.4973273212090135, 'learning_rate/qwen_vl_interface': 6.267176296330734e-06, 'learning_rate/action_model': 6.267176296330733e-05, 'epoch': 1.18}) +[2026-06-17 23:07:16,866][starVLA.training.train_starvla][INFO] - Step 3077, Loss: {'train/loss': 0.19659258425235748, 'train/grad_norm_pre_clip': 5.962105751037598, 'timing/data': 0.010674148797988892, 'timing/model': 1.4494771994650364, 'learning_rate/qwen_vl_interface': 6.265063669483992e-06, 'learning_rate/action_model': 6.265063669483991e-05, 'epoch': 1.18}) +[2026-06-17 23:07:21,333][starVLA.training.train_starvla][INFO] - Step 3078, Loss: {'train/loss': 0.1868048757314682, 'train/grad_norm_pre_clip': 5.381843090057373, 'timing/data': 0.012050538323819637, 'timing/model': 1.1529851723462343, 'learning_rate/qwen_vl_interface': 6.262950832213457e-06, 'learning_rate/action_model': 6.262950832213456e-05, 'epoch': 1.18}) +[2026-06-17 23:07:26,401][starVLA.training.train_starvla][INFO] - Step 3079, Loss: {'train/loss': 0.4004977345466614, 'train/grad_norm_pre_clip': 9.591107368469238, 'timing/data': 0.011927910149097443, 'timing/model': 1.1902292855083942, 'learning_rate/qwen_vl_interface': 6.260837784957125e-06, 'learning_rate/action_model': 6.260837784957126e-05, 'epoch': 1.18}) +[2026-06-17 23:07:31,398][starVLA.training.train_starvla][INFO] - Step 3080, Loss: {'train/loss': 0.596474826335907, 'train/grad_norm_pre_clip': 12.936128616333008, 'timing/data': 0.01202250923961401, 'timing/model': 1.244701392017305, 'learning_rate/qwen_vl_interface': 6.258724528153034e-06, 'learning_rate/action_model': 6.258724528153033e-05, 'epoch': 1.18}) +[2026-06-17 23:07:37,115][starVLA.training.train_starvla][INFO] - Step 3081, Loss: {'train/loss': 0.3109491765499115, 'train/grad_norm_pre_clip': 8.196063995361328, 'timing/data': 0.0151416826993227, 'timing/model': 1.1636048564687371, 'learning_rate/qwen_vl_interface': 6.256611062239261e-06, 'learning_rate/action_model': 6.25661106223926e-05, 'epoch': 1.19}) +[2026-06-17 23:07:46,132][starVLA.training.train_starvla][INFO] - Step 3082, Loss: {'train/loss': 0.5092626810073853, 'train/grad_norm_pre_clip': 11.473295211791992, 'timing/data': 0.011492040008306503, 'timing/model': 1.4127366058528423, 'learning_rate/qwen_vl_interface': 6.254497387653933e-06, 'learning_rate/action_model': 6.254497387653932e-05, 'epoch': 1.19}) +[2026-06-17 23:07:50,448][starVLA.training.train_starvla][INFO] - Step 3083, Loss: {'train/loss': 0.5361222624778748, 'train/grad_norm_pre_clip': 12.937785148620605, 'timing/data': 0.01166727114468813, 'timing/model': 1.1808204492554069, 'learning_rate/qwen_vl_interface': 6.252383504835213e-06, 'learning_rate/action_model': 6.252383504835213e-05, 'epoch': 1.19}) +[2026-06-17 23:07:55,432][starVLA.training.train_starvla][INFO] - Step 3084, Loss: {'train/loss': 0.3881007134914398, 'train/grad_norm_pre_clip': 12.28125286102295, 'timing/data': 0.018625774420797825, 'timing/model': 1.099381553940475, 'learning_rate/qwen_vl_interface': 6.250269414221317e-06, 'learning_rate/action_model': 6.250269414221317e-05, 'epoch': 1.19}) +[2026-06-17 23:08:00,715][starVLA.training.train_starvla][INFO] - Step 3085, Loss: {'train/loss': 0.23406368494033813, 'train/grad_norm_pre_clip': 6.793148994445801, 'timing/data': 0.012449054978787899, 'timing/model': 1.2819405402988195, 'learning_rate/qwen_vl_interface': 6.248155116250494e-06, 'learning_rate/action_model': 6.248155116250493e-05, 'epoch': 1.19}) +[2026-06-17 23:08:05,635][starVLA.training.train_starvla][INFO] - Step 3086, Loss: {'train/loss': 0.37710103392601013, 'train/grad_norm_pre_clip': 10.804591178894043, 'timing/data': 0.011148619465529919, 'timing/model': 1.646282110363245, 'learning_rate/qwen_vl_interface': 6.246040611361042e-06, 'learning_rate/action_model': 6.246040611361041e-05, 'epoch': 1.19}) +[2026-06-17 23:08:10,789][starVLA.training.train_starvla][INFO] - Step 3087, Loss: {'train/loss': 0.5247872471809387, 'train/grad_norm_pre_clip': 10.900802612304688, 'timing/data': 0.013505415990948677, 'timing/model': 1.3704712688922882, 'learning_rate/qwen_vl_interface': 6.2439258999912976e-06, 'learning_rate/action_model': 6.243925899991298e-05, 'epoch': 1.19}) +[2026-06-17 23:08:19,139][starVLA.training.train_starvla][INFO] - Step 3088, Loss: {'train/loss': 0.24332045018672943, 'train/grad_norm_pre_clip': 14.429365158081055, 'timing/data': 0.012082771398127079, 'timing/model': 1.3475194657221437, 'learning_rate/qwen_vl_interface': 6.241810982579646e-06, 'learning_rate/action_model': 6.241810982579646e-05, 'epoch': 1.19}) +[2026-06-17 23:08:31,857][starVLA.training.train_starvla][INFO] - Step 3089, Loss: {'train/loss': 0.2667715847492218, 'train/grad_norm_pre_clip': 10.069389343261719, 'timing/data': 0.01387215219438076, 'timing/model': 1.2714986372739077, 'learning_rate/qwen_vl_interface': 6.239695859564511e-06, 'learning_rate/action_model': 6.23969585956451e-05, 'epoch': 1.19}) +[2026-06-17 23:08:39,952][starVLA.training.train_starvla][INFO] - Step 3090, Loss: {'train/loss': 0.6273803114891052, 'train/grad_norm_pre_clip': 10.643326759338379, 'timing/data': 0.0166368642821908, 'timing/model': 2.6142069147899747, 'learning_rate/qwen_vl_interface': 6.2375805313843564e-06, 'learning_rate/action_model': 6.237580531384356e-05, 'epoch': 1.19}) +[2026-06-17 23:08:46,677][starVLA.training.train_starvla][INFO] - Step 3091, Loss: {'train/loss': 0.692232072353363, 'train/grad_norm_pre_clip': 14.658817291259766, 'timing/data': 0.01975557766854763, 'timing/model': 2.587874705903232, 'learning_rate/qwen_vl_interface': 6.2354649984777e-06, 'learning_rate/action_model': 6.235464998477699e-05, 'epoch': 1.19}) +[2026-06-17 23:08:54,608][starVLA.training.train_starvla][INFO] - Step 3092, Loss: {'train/loss': 0.8304515480995178, 'train/grad_norm_pre_clip': 14.649195671081543, 'timing/data': 0.012162715196609497, 'timing/model': 1.4715472720563412, 'learning_rate/qwen_vl_interface': 6.2333492612830835e-06, 'learning_rate/action_model': 6.233349261283083e-05, 'epoch': 1.19}) +[2026-06-17 23:08:59,140][starVLA.training.train_starvla][INFO] - Step 3093, Loss: {'train/loss': 0.44933968782424927, 'train/grad_norm_pre_clip': 9.602420806884766, 'timing/data': 0.015411578118801117, 'timing/model': 1.2025824831798673, 'learning_rate/qwen_vl_interface': 6.231233320239112e-06, 'learning_rate/action_model': 6.231233320239112e-05, 'epoch': 1.19}) +[2026-06-17 23:09:04,372][starVLA.training.train_starvla][INFO] - Step 3094, Loss: {'train/loss': 0.25450319051742554, 'train/grad_norm_pre_clip': 7.692743301391602, 'timing/data': 0.013220454566180706, 'timing/model': 1.3140456518158317, 'learning_rate/qwen_vl_interface': 6.229117175784414e-06, 'learning_rate/action_model': 6.229117175784413e-05, 'epoch': 1.19}) +[2026-06-17 23:09:09,174][starVLA.training.train_starvla][INFO] - Step 3095, Loss: {'train/loss': 0.27242541313171387, 'train/grad_norm_pre_clip': 7.515919208526611, 'timing/data': 0.011815682984888554, 'timing/model': 1.2081446656957269, 'learning_rate/qwen_vl_interface': 6.227000828357672e-06, 'learning_rate/action_model': 6.227000828357672e-05, 'epoch': 1.19}) +[2026-06-17 23:09:14,319][starVLA.training.train_starvla][INFO] - Step 3096, Loss: {'train/loss': 0.7803935408592224, 'train/grad_norm_pre_clip': 15.838459968566895, 'timing/data': 0.013077243231236935, 'timing/model': 1.5590329803526402, 'learning_rate/qwen_vl_interface': 6.2248842783976085e-06, 'learning_rate/action_model': 6.224884278397608e-05, 'epoch': 1.19}) +[2026-06-17 23:09:18,967][starVLA.training.train_starvla][INFO] - Step 3097, Loss: {'train/loss': 0.4469752311706543, 'train/grad_norm_pre_clip': 9.98810863494873, 'timing/data': 0.01211196556687355, 'timing/model': 1.362495587207377, 'learning_rate/qwen_vl_interface': 6.222767526342983e-06, 'learning_rate/action_model': 6.222767526342983e-05, 'epoch': 1.19}) +[2026-06-17 23:09:23,233][starVLA.training.train_starvla][INFO] - Step 3098, Loss: {'train/loss': 0.3747745156288147, 'train/grad_norm_pre_clip': 8.461959838867188, 'timing/data': 0.00969571340829134, 'timing/model': 1.0346058523282409, 'learning_rate/qwen_vl_interface': 6.220650572632605e-06, 'learning_rate/action_model': 6.220650572632604e-05, 'epoch': 1.19}) +[2026-06-17 23:09:27,463][starVLA.training.train_starvla][INFO] - Step 3099, Loss: {'train/loss': 0.49337148666381836, 'train/grad_norm_pre_clip': 12.493311882019043, 'timing/data': 0.013395941816270351, 'timing/model': 1.0538941286504269, 'learning_rate/qwen_vl_interface': 6.218533417705318e-06, 'learning_rate/action_model': 6.218533417705318e-05, 'epoch': 1.19}) +[2026-06-17 23:12:40,767][starVLA.training.train_starvla][INFO] - Step 3100, Loss: {'train/loss': 0.3507359027862549, 'train/grad_norm_pre_clip': 10.430567741394043, 'eval/loss': 0.4844139814376831, 'eval/latency_4/loss': 0.4844139814376831, 'eval/demon_fire_precision': 0.9652976758993951, 'eval/demon_fire_recall': 0.9847353036700227, 'eval/demon_fire_f1': 0.97491961414791, 'eval/demon_fire_rate_model': 0.9224669603524229, 'eval/demon_fire_rate_teacher': 0.9042584434654919, 'eval/demon_left_precision': 0.7373493975903614, 'eval/demon_left_recall': 0.6986301369863014, 'eval/demon_left_f1': 0.7174677608440797, 'eval/demon_left_rate_model': 0.12187958883994127, 'eval/demon_left_rate_teacher': 0.12863436123348018, 'eval/demon_right_precision': 0.816247582205029, 'eval/demon_right_recall': 0.7617328519855595, 'eval/demon_right_f1': 0.7880485527544352, 'eval/demon_right_rate_model': 0.15183553597650515, 'eval/demon_right_rate_teacher': 0.16270190895741557, 'eval/demon_move_f1': 0.7527581567992574, 'eval/demon_cc_f1': 0.8956221455181174, 'eval/latency_4/demon_fire_precision': 0.9652976758993951, 'eval/latency_4/demon_fire_recall': 0.9847353036700227, 'eval/latency_4/demon_fire_f1': 0.97491961414791, 'eval/latency_4/demon_fire_rate_model': 0.9224669603524229, 'eval/latency_4/demon_fire_rate_teacher': 0.9042584434654919, 'eval/latency_4/demon_left_precision': 0.7373493975903614, 'eval/latency_4/demon_left_recall': 0.6986301369863014, 'eval/latency_4/demon_left_f1': 0.7174677608440797, 'eval/latency_4/demon_left_rate_model': 0.12187958883994127, 'eval/latency_4/demon_left_rate_teacher': 0.12863436123348018, 'eval/latency_4/demon_right_precision': 0.816247582205029, 'eval/latency_4/demon_right_recall': 0.7617328519855595, 'eval/latency_4/demon_right_f1': 0.7880485527544352, 'eval/latency_4/demon_right_rate_model': 0.15183553597650515, 'eval/latency_4/demon_right_rate_teacher': 0.16270190895741557, 'eval/latency_4/demon_move_f1': 0.7527581567992574, 'eval/latency_4/demon_cc_f1': 0.8956221455181174, 'timing/data': 0.009791680611670017, 'timing/model': 1.034044485539198, 'learning_rate/qwen_vl_interface': 6.216416062000012e-06, 'learning_rate/action_model': 6.216416062000013e-05, 'epoch': 1.19}) +[2026-06-17 23:12:40,778][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_3100_state +[2026-06-17 23:12:40,787][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 23:12:50,896][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_3100_state/model.safetensors +[2026-06-17 23:13:13,611][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_3100_state/optimizer.bin +[2026-06-17 23:13:13,613][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_3100_state/sampler.bin +[2026-06-17 23:13:13,614][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_3100_state/sampler_1.bin +[2026-06-17 23:13:13,617][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_3100_state/random_states_0.pkl +[2026-06-17 23:13:13,621][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3100 +[2026-06-17 23:13:13,622][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 23:13:13,639][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 23:13:20,641][starVLA.training.train_starvla][INFO] - Step 3101, Loss: {'train/loss': 0.5031700134277344, 'train/grad_norm_pre_clip': 12.26396656036377, 'timing/data': 0.016376477666199207, 'timing/model': 2.030523171648383, 'learning_rate/qwen_vl_interface': 6.214298505955615e-06, 'learning_rate/action_model': 6.214298505955614e-05, 'epoch': 1.19}) +[2026-06-17 23:13:25,227][starVLA.training.train_starvla][INFO] - Step 3102, Loss: {'train/loss': 0.6312124133110046, 'train/grad_norm_pre_clip': 11.957846641540527, 'timing/data': 0.012129560112953186, 'timing/model': 1.087720451876521, 'learning_rate/qwen_vl_interface': 6.2121807500111e-06, 'learning_rate/action_model': 6.212180750011099e-05, 'epoch': 1.19}) +[2026-06-17 23:13:30,086][starVLA.training.train_starvla][INFO] - Step 3103, Loss: {'train/loss': 0.20993909239768982, 'train/grad_norm_pre_clip': 8.014742851257324, 'timing/data': 0.015185311436653137, 'timing/model': 1.201431549154222, 'learning_rate/qwen_vl_interface': 6.210062794605482e-06, 'learning_rate/action_model': 6.21006279460548e-05, 'epoch': 1.19}) +[2026-06-17 23:13:38,177][starVLA.training.train_starvla][INFO] - Step 3104, Loss: {'train/loss': 0.3597969114780426, 'train/grad_norm_pre_clip': 14.150003433227539, 'timing/data': 0.021316529251635075, 'timing/model': 4.104654849506915, 'learning_rate/qwen_vl_interface': 6.207944640177812e-06, 'learning_rate/action_model': 6.207944640177812e-05, 'epoch': 1.19}) +[2026-06-17 23:13:43,435][starVLA.training.train_starvla][INFO] - Step 3105, Loss: {'train/loss': 0.23762719333171844, 'train/grad_norm_pre_clip': 6.631361484527588, 'timing/data': 0.013238421641290188, 'timing/model': 1.2257055724039674, 'learning_rate/qwen_vl_interface': 6.205826287167188e-06, 'learning_rate/action_model': 6.205826287167188e-05, 'epoch': 1.19}) +[2026-06-17 23:13:48,585][starVLA.training.train_starvla][INFO] - Step 3106, Loss: {'train/loss': 0.6825696229934692, 'train/grad_norm_pre_clip': 12.835801124572754, 'timing/data': 0.022265460342168808, 'timing/model': 1.1360247367992997, 'learning_rate/qwen_vl_interface': 6.203707736012745e-06, 'learning_rate/action_model': 6.203707736012745e-05, 'epoch': 1.19}) +[2026-06-17 23:13:54,087][starVLA.training.train_starvla][INFO] - Step 3107, Loss: {'train/loss': 0.6880725026130676, 'train/grad_norm_pre_clip': 13.187458992004395, 'timing/data': 0.014908322133123875, 'timing/model': 1.7429368766024709, 'learning_rate/qwen_vl_interface': 6.201588987153665e-06, 'learning_rate/action_model': 6.201588987153664e-05, 'epoch': 1.2}) +[2026-06-17 23:13:59,020][starVLA.training.train_starvla][INFO] - Step 3108, Loss: {'train/loss': 0.2700806260108948, 'train/grad_norm_pre_clip': 7.700798034667969, 'timing/data': 0.011903511360287666, 'timing/model': 1.2379298694431782, 'learning_rate/qwen_vl_interface': 6.19947004102916e-06, 'learning_rate/action_model': 6.199470041029159e-05, 'epoch': 1.2}) +[2026-06-17 23:14:08,488][starVLA.training.train_starvla][INFO] - Step 3109, Loss: {'train/loss': 0.46900394558906555, 'train/grad_norm_pre_clip': 12.817828178405762, 'timing/data': 0.019772430881857872, 'timing/model': 1.3118449794128537, 'learning_rate/qwen_vl_interface': 6.197350898078496e-06, 'learning_rate/action_model': 6.197350898078495e-05, 'epoch': 1.2}) +[2026-06-17 23:14:13,557][starVLA.training.train_starvla][INFO] - Step 3110, Loss: {'train/loss': 0.4790741801261902, 'train/grad_norm_pre_clip': 10.974076271057129, 'timing/data': 0.012253624387085438, 'timing/model': 1.4042555065825582, 'learning_rate/qwen_vl_interface': 6.195231558740972e-06, 'learning_rate/action_model': 6.195231558740971e-05, 'epoch': 1.2}) +[2026-06-17 23:14:18,702][starVLA.training.train_starvla][INFO] - Step 3111, Loss: {'train/loss': 0.186673104763031, 'train/grad_norm_pre_clip': 8.052042961120605, 'timing/data': 0.013968403451144695, 'timing/model': 1.1531428061425686, 'learning_rate/qwen_vl_interface': 6.1931120234559275e-06, 'learning_rate/action_model': 6.193112023455927e-05, 'epoch': 1.2}) +[2026-06-17 23:14:26,034][starVLA.training.train_starvla][INFO] - Step 3112, Loss: {'train/loss': 0.29746997356414795, 'train/grad_norm_pre_clip': 9.031885147094727, 'timing/data': 0.011868849396705627, 'timing/model': 2.0621708184480667, 'learning_rate/qwen_vl_interface': 6.190992292662747e-06, 'learning_rate/action_model': 6.190992292662746e-05, 'epoch': 1.2}) +[2026-06-17 23:14:31,037][starVLA.training.train_starvla][INFO] - Step 3113, Loss: {'train/loss': 0.4311933219432831, 'train/grad_norm_pre_clip': 12.726202011108398, 'timing/data': 0.013344381004571915, 'timing/model': 1.1257861694321036, 'learning_rate/qwen_vl_interface': 6.18887236680085e-06, 'learning_rate/action_model': 6.18887236680085e-05, 'epoch': 1.2}) +[2026-06-17 23:14:35,938][starVLA.training.train_starvla][INFO] - Step 3114, Loss: {'train/loss': 0.4522954523563385, 'train/grad_norm_pre_clip': 13.997415542602539, 'timing/data': 0.013496625237166882, 'timing/model': 1.110143855214119, 'learning_rate/qwen_vl_interface': 6.186752246309702e-06, 'learning_rate/action_model': 6.186752246309702e-05, 'epoch': 1.2}) +[2026-06-17 23:14:45,297][starVLA.training.train_starvla][INFO] - Step 3115, Loss: {'train/loss': 0.4894283413887024, 'train/grad_norm_pre_clip': 25.470478057861328, 'timing/data': 0.012864414602518082, 'timing/model': 4.405151154845953, 'learning_rate/qwen_vl_interface': 6.184631931628806e-06, 'learning_rate/action_model': 6.184631931628806e-05, 'epoch': 1.2}) +[2026-06-17 23:14:52,880][starVLA.training.train_starvla][INFO] - Step 3116, Loss: {'train/loss': 0.3931726813316345, 'train/grad_norm_pre_clip': 8.455023765563965, 'timing/data': 0.06666693650186062, 'timing/model': 3.669624103233218, 'learning_rate/qwen_vl_interface': 6.1825114231977036e-06, 'learning_rate/action_model': 6.182511423197704e-05, 'epoch': 1.2}) +[2026-06-17 23:14:57,939][starVLA.training.train_starvla][INFO] - Step 3117, Loss: {'train/loss': 0.49042201042175293, 'train/grad_norm_pre_clip': 9.643087387084961, 'timing/data': 0.011952528730034828, 'timing/model': 1.3821653928607702, 'learning_rate/qwen_vl_interface': 6.18039072145598e-06, 'learning_rate/action_model': 6.18039072145598e-05, 'epoch': 1.2}) +[2026-06-17 23:15:03,208][starVLA.training.train_starvla][INFO] - Step 3118, Loss: {'train/loss': 0.6125579476356506, 'train/grad_norm_pre_clip': 12.037437438964844, 'timing/data': 0.013157883659005165, 'timing/model': 1.33337377384305, 'learning_rate/qwen_vl_interface': 6.17826982684326e-06, 'learning_rate/action_model': 6.178269826843259e-05, 'epoch': 1.2}) +[2026-06-17 23:15:14,412][starVLA.training.train_starvla][INFO] - Step 3119, Loss: {'train/loss': 0.48288846015930176, 'train/grad_norm_pre_clip': 11.28861141204834, 'timing/data': 0.01470392756164074, 'timing/model': 1.4937025848776102, 'learning_rate/qwen_vl_interface': 6.176148739799205e-06, 'learning_rate/action_model': 6.176148739799205e-05, 'epoch': 1.2}) +[2026-06-17 23:15:19,163][starVLA.training.train_starvla][INFO] - Step 3120, Loss: {'train/loss': 0.4280686676502228, 'train/grad_norm_pre_clip': 10.860026359558105, 'timing/data': 0.017600846476852894, 'timing/model': 1.1717161173000932, 'learning_rate/qwen_vl_interface': 6.1740274607635195e-06, 'learning_rate/action_model': 6.174027460763519e-05, 'epoch': 1.2}) +[2026-06-17 23:15:24,313][starVLA.training.train_starvla][INFO] - Step 3121, Loss: {'train/loss': 0.7864885330200195, 'train/grad_norm_pre_clip': 15.581250190734863, 'timing/data': 0.012516342103481293, 'timing/model': 1.6335380598902702, 'learning_rate/qwen_vl_interface': 6.1719059901759474e-06, 'learning_rate/action_model': 6.171905990175947e-05, 'epoch': 1.2}) +[2026-06-17 23:15:31,396][starVLA.training.train_starvla][INFO] - Step 3122, Loss: {'train/loss': 0.29423069953918457, 'train/grad_norm_pre_clip': 14.262746810913086, 'timing/data': 0.014496706426143646, 'timing/model': 3.8347688522189856, 'learning_rate/qwen_vl_interface': 6.169784328476272e-06, 'learning_rate/action_model': 6.169784328476271e-05, 'epoch': 1.2}) +[2026-06-17 23:15:36,429][starVLA.training.train_starvla][INFO] - Step 3123, Loss: {'train/loss': 0.44404149055480957, 'train/grad_norm_pre_clip': 10.413934707641602, 'timing/data': 0.02072118129581213, 'timing/model': 1.6008719000965357, 'learning_rate/qwen_vl_interface': 6.167662476104314e-06, 'learning_rate/action_model': 6.167662476104314e-05, 'epoch': 1.2}) +[2026-06-17 23:15:41,427][starVLA.training.train_starvla][INFO] - Step 3124, Loss: {'train/loss': 0.31405699253082275, 'train/grad_norm_pre_clip': 7.463569164276123, 'timing/data': 0.014798153191804886, 'timing/model': 1.0841750735417008, 'learning_rate/qwen_vl_interface': 6.165540433499939e-06, 'learning_rate/action_model': 6.165540433499939e-05, 'epoch': 1.2}) +[2026-06-17 23:15:50,243][starVLA.training.train_starvla][INFO] - Step 3125, Loss: {'train/loss': 0.4885754883289337, 'train/grad_norm_pre_clip': 9.332883834838867, 'timing/data': 0.01205168105661869, 'timing/model': 1.256868313997984, 'learning_rate/qwen_vl_interface': 6.163418201103045e-06, 'learning_rate/action_model': 6.163418201103045e-05, 'epoch': 1.2}) +[2026-06-17 23:15:54,961][starVLA.training.train_starvla][INFO] - Step 3126, Loss: {'train/loss': 0.25662708282470703, 'train/grad_norm_pre_clip': 8.176941871643066, 'timing/data': 0.011833596974611282, 'timing/model': 1.1376059781759977, 'learning_rate/qwen_vl_interface': 6.161295779353578e-06, 'learning_rate/action_model': 6.161295779353577e-05, 'epoch': 1.2}) +[2026-06-17 23:16:00,585][starVLA.training.train_starvla][INFO] - Step 3127, Loss: {'train/loss': 0.5931090116500854, 'train/grad_norm_pre_clip': 15.092679023742676, 'timing/data': 0.01565366890281439, 'timing/model': 1.7175254225730896, 'learning_rate/qwen_vl_interface': 6.159173168691512e-06, 'learning_rate/action_model': 6.159173168691512e-05, 'epoch': 1.2}) +[2026-06-17 23:16:05,471][starVLA.training.train_starvla][INFO] - Step 3128, Loss: {'train/loss': 0.43427127599716187, 'train/grad_norm_pre_clip': 9.191615104675293, 'timing/data': 0.013032692484557629, 'timing/model': 1.2597150383517146, 'learning_rate/qwen_vl_interface': 6.157050369556871e-06, 'learning_rate/action_model': 6.157050369556872e-05, 'epoch': 1.2}) +[2026-06-17 23:16:14,068][starVLA.training.train_starvla][INFO] - Step 3129, Loss: {'train/loss': 0.4035806953907013, 'train/grad_norm_pre_clip': 11.285374641418457, 'timing/data': 0.012888024561107159, 'timing/model': 4.509168430231512, 'learning_rate/qwen_vl_interface': 6.154927382389714e-06, 'learning_rate/action_model': 6.154927382389714e-05, 'epoch': 1.2}) +[2026-06-17 23:16:18,883][starVLA.training.train_starvla][INFO] - Step 3130, Loss: {'train/loss': 0.5297180414199829, 'train/grad_norm_pre_clip': 12.990304946899414, 'timing/data': 0.011972596868872643, 'timing/model': 1.2043019672855735, 'learning_rate/qwen_vl_interface': 6.152804207630135e-06, 'learning_rate/action_model': 6.152804207630136e-05, 'epoch': 1.2}) +[2026-06-17 23:16:23,475][starVLA.training.train_starvla][INFO] - Step 3131, Loss: {'train/loss': 0.5975396633148193, 'train/grad_norm_pre_clip': 10.913138389587402, 'timing/data': 0.01189532969146967, 'timing/model': 1.1036996291950345, 'learning_rate/qwen_vl_interface': 6.150680845718274e-06, 'learning_rate/action_model': 6.150680845718274e-05, 'epoch': 1.2}) +[2026-06-17 23:16:28,481][starVLA.training.train_starvla][INFO] - Step 3132, Loss: {'train/loss': 0.27082622051239014, 'train/grad_norm_pre_clip': 8.974693298339844, 'timing/data': 0.01399770937860012, 'timing/model': 1.607555934228003, 'learning_rate/qwen_vl_interface': 6.148557297094303e-06, 'learning_rate/action_model': 6.148557297094303e-05, 'epoch': 1.2}) +[2026-06-17 23:16:34,073][starVLA.training.train_starvla][INFO] - Step 3133, Loss: {'train/loss': 0.24928805232048035, 'train/grad_norm_pre_clip': 7.405873775482178, 'timing/data': 0.014348582364618778, 'timing/model': 1.153460992500186, 'learning_rate/qwen_vl_interface': 6.1464335621984375e-06, 'learning_rate/action_model': 6.146433562198437e-05, 'epoch': 1.21}) +[2026-06-17 23:16:39,026][starVLA.training.train_starvla][INFO] - Step 3134, Loss: {'train/loss': 0.2315814197063446, 'train/grad_norm_pre_clip': 6.766599655151367, 'timing/data': 0.01252224761992693, 'timing/model': 1.1060186363756657, 'learning_rate/qwen_vl_interface': 6.14430964147093e-06, 'learning_rate/action_model': 6.14430964147093e-05, 'epoch': 1.21}) +[2026-06-17 23:16:48,471][starVLA.training.train_starvla][INFO] - Step 3135, Loss: {'train/loss': 0.3164660632610321, 'train/grad_norm_pre_clip': 7.995407581329346, 'timing/data': 0.01520233042538166, 'timing/model': 2.442658506333828, 'learning_rate/qwen_vl_interface': 6.1421855353520735e-06, 'learning_rate/action_model': 6.142185535352073e-05, 'epoch': 1.21}) +[2026-06-17 23:16:53,540][starVLA.training.train_starvla][INFO] - Step 3136, Loss: {'train/loss': 0.5258702635765076, 'train/grad_norm_pre_clip': 9.451727867126465, 'timing/data': 0.011975266970694065, 'timing/model': 1.4805086199194193, 'learning_rate/qwen_vl_interface': 6.140061244282194e-06, 'learning_rate/action_model': 6.140061244282194e-05, 'epoch': 1.21}) +[2026-06-17 23:16:58,673][starVLA.training.train_starvla][INFO] - Step 3137, Loss: {'train/loss': 0.2894267737865448, 'train/grad_norm_pre_clip': 7.656859397888184, 'timing/data': 0.01272010337561369, 'timing/model': 1.6022201171144843, 'learning_rate/qwen_vl_interface': 6.137936768701659e-06, 'learning_rate/action_model': 6.137936768701659e-05, 'epoch': 1.21}) +[2026-06-17 23:17:03,242][starVLA.training.train_starvla][INFO] - Step 3138, Loss: {'train/loss': 0.22269664704799652, 'train/grad_norm_pre_clip': 7.604046821594238, 'timing/data': 0.01255539245903492, 'timing/model': 1.1175494119524956, 'learning_rate/qwen_vl_interface': 6.135812109050879e-06, 'learning_rate/action_model': 6.135812109050879e-05, 'epoch': 1.21}) +[2026-06-17 23:17:08,327][starVLA.training.train_starvla][INFO] - Step 3139, Loss: {'train/loss': 0.408520370721817, 'train/grad_norm_pre_clip': 9.58132266998291, 'timing/data': 0.012109505012631416, 'timing/model': 1.0861785300076008, 'learning_rate/qwen_vl_interface': 6.1336872657702936e-06, 'learning_rate/action_model': 6.133687265770293e-05, 'epoch': 1.21}) +[2026-06-17 23:17:13,084][starVLA.training.train_starvla][INFO] - Step 3140, Loss: {'train/loss': 0.2553165555000305, 'train/grad_norm_pre_clip': 6.557133674621582, 'timing/data': 0.011650674045085907, 'timing/model': 1.0708753317594528, 'learning_rate/qwen_vl_interface': 6.131562239300389e-06, 'learning_rate/action_model': 6.13156223930039e-05, 'epoch': 1.21}) +[2026-06-17 23:17:17,952][starVLA.training.train_starvla][INFO] - Step 3141, Loss: {'train/loss': 0.591149628162384, 'train/grad_norm_pre_clip': 13.451587677001953, 'timing/data': 0.014110077172517776, 'timing/model': 1.304933656938374, 'learning_rate/qwen_vl_interface': 6.129437030081683e-06, 'learning_rate/action_model': 6.129437030081682e-05, 'epoch': 1.21}) +[2026-06-17 23:17:22,840][starVLA.training.train_starvla][INFO] - Step 3142, Loss: {'train/loss': 0.4732767641544342, 'train/grad_norm_pre_clip': 15.090963363647461, 'timing/data': 0.011718101799488068, 'timing/model': 1.4279298083856702, 'learning_rate/qwen_vl_interface': 6.1273116385547336e-06, 'learning_rate/action_model': 6.127311638554734e-05, 'epoch': 1.21}) +[2026-06-17 23:17:27,215][starVLA.training.train_starvla][INFO] - Step 3143, Loss: {'train/loss': 0.5310035347938538, 'train/grad_norm_pre_clip': 9.443553924560547, 'timing/data': 0.012269035913050175, 'timing/model': 1.1104165334254503, 'learning_rate/qwen_vl_interface': 6.125186065160138e-06, 'learning_rate/action_model': 6.125186065160138e-05, 'epoch': 1.21}) +[2026-06-17 23:17:39,230][starVLA.training.train_starvla][INFO] - Step 3144, Loss: {'train/loss': 0.5915302038192749, 'train/grad_norm_pre_clip': 9.147655487060547, 'timing/data': 0.015822350047528744, 'timing/model': 5.4296888541430235, 'learning_rate/qwen_vl_interface': 6.123060310338529e-06, 'learning_rate/action_model': 6.123060310338529e-05, 'epoch': 1.21}) +[2026-06-17 23:17:48,406][starVLA.training.train_starvla][INFO] - Step 3145, Loss: {'train/loss': 0.3606954514980316, 'train/grad_norm_pre_clip': 9.163009643554688, 'timing/data': 0.01247019786387682, 'timing/model': 1.4358541471883655, 'learning_rate/qwen_vl_interface': 6.120934374530579e-06, 'learning_rate/action_model': 6.12093437453058e-05, 'epoch': 1.21}) +[2026-06-17 23:17:56,138][starVLA.training.train_starvla][INFO] - Step 3146, Loss: {'train/loss': 0.3216955363750458, 'train/grad_norm_pre_clip': 8.064364433288574, 'timing/data': 0.012158219702541828, 'timing/model': 4.457851688377559, 'learning_rate/qwen_vl_interface': 6.118808258176996e-06, 'learning_rate/action_model': 6.118808258176995e-05, 'epoch': 1.21}) +[2026-06-17 23:18:01,733][starVLA.training.train_starvla][INFO] - Step 3147, Loss: {'train/loss': 0.49962756037712097, 'train/grad_norm_pre_clip': 10.00688648223877, 'timing/data': 0.014098296873271465, 'timing/model': 1.5580847738310695, 'learning_rate/qwen_vl_interface': 6.116681961718525e-06, 'learning_rate/action_model': 6.116681961718524e-05, 'epoch': 1.21}) +[2026-06-17 23:18:12,494][starVLA.training.train_starvla][INFO] - Step 3148, Loss: {'train/loss': 0.3766375184059143, 'train/grad_norm_pre_clip': 8.245259284973145, 'timing/data': 0.012354577891528606, 'timing/model': 2.335285434499383, 'learning_rate/qwen_vl_interface': 6.114555485595952e-06, 'learning_rate/action_model': 6.114555485595952e-05, 'epoch': 1.21}) +[2026-06-17 23:18:20,502][starVLA.training.train_starvla][INFO] - Step 3149, Loss: {'train/loss': 0.3242948055267334, 'train/grad_norm_pre_clip': 9.178009033203125, 'timing/data': 0.012484579347074032, 'timing/model': 3.9267186326906085, 'learning_rate/qwen_vl_interface': 6.112428830250096e-06, 'learning_rate/action_model': 6.112428830250095e-05, 'epoch': 1.21}) +[2026-06-17 23:18:28,571][starVLA.training.train_starvla][INFO] - Step 3150, Loss: {'train/loss': 0.5409824848175049, 'train/grad_norm_pre_clip': 13.00513744354248, 'timing/data': 0.012190687470138073, 'timing/model': 1.1216947343200445, 'learning_rate/qwen_vl_interface': 6.110301996121816e-06, 'learning_rate/action_model': 6.110301996121816e-05, 'epoch': 1.21}) +[2026-06-17 23:18:33,622][starVLA.training.train_starvla][INFO] - Step 3151, Loss: {'train/loss': 0.553756833076477, 'train/grad_norm_pre_clip': 10.534499168395996, 'timing/data': 0.027685479260981083, 'timing/model': 1.0861211996525526, 'learning_rate/qwen_vl_interface': 6.108174983652007e-06, 'learning_rate/action_model': 6.108174983652006e-05, 'epoch': 1.21}) +[2026-06-17 23:18:38,326][starVLA.training.train_starvla][INFO] - Step 3152, Loss: {'train/loss': 0.6542396545410156, 'train/grad_norm_pre_clip': 12.471092224121094, 'timing/data': 0.01351242046803236, 'timing/model': 1.5034804064780474, 'learning_rate/qwen_vl_interface': 6.106047793281598e-06, 'learning_rate/action_model': 6.106047793281598e-05, 'epoch': 1.21}) +[2026-06-17 23:18:47,569][starVLA.training.train_starvla][INFO] - Step 3153, Loss: {'train/loss': 0.45278647541999817, 'train/grad_norm_pre_clip': 10.125062942504883, 'timing/data': 0.013152736239135265, 'timing/model': 3.39570199791342, 'learning_rate/qwen_vl_interface': 6.1039204254515626e-06, 'learning_rate/action_model': 6.103920425451563e-05, 'epoch': 1.21}) +[2026-06-17 23:18:56,294][starVLA.training.train_starvla][INFO] - Step 3154, Loss: {'train/loss': 0.4630013108253479, 'train/grad_norm_pre_clip': 9.348587989807129, 'timing/data': 0.011349705047905445, 'timing/model': 1.0917605506256223, 'learning_rate/qwen_vl_interface': 6.101792880602901e-06, 'learning_rate/action_model': 6.101792880602901e-05, 'epoch': 1.21}) +[2026-06-17 23:19:00,948][starVLA.training.train_starvla][INFO] - Step 3155, Loss: {'train/loss': 0.22908665239810944, 'train/grad_norm_pre_clip': 8.073206901550293, 'timing/data': 0.014691447839140892, 'timing/model': 1.1099832728505135, 'learning_rate/qwen_vl_interface': 6.099665159176663e-06, 'learning_rate/action_model': 6.099665159176663e-05, 'epoch': 1.21}) +[2026-06-17 23:19:05,677][starVLA.training.train_starvla][INFO] - Step 3156, Loss: {'train/loss': 0.28183797001838684, 'train/grad_norm_pre_clip': 7.649582386016846, 'timing/data': 0.01773842517286539, 'timing/model': 1.5650152079761028, 'learning_rate/qwen_vl_interface': 6.09753726161392e-06, 'learning_rate/action_model': 6.09753726161392e-05, 'epoch': 1.21}) +[2026-06-17 23:19:11,674][starVLA.training.train_starvla][INFO] - Step 3157, Loss: {'train/loss': 0.45565950870513916, 'train/grad_norm_pre_clip': 9.940874099731445, 'timing/data': 0.012807128019630909, 'timing/model': 1.4740193579345942, 'learning_rate/qwen_vl_interface': 6.095409188355793e-06, 'learning_rate/action_model': 6.095409188355793e-05, 'epoch': 1.21}) +[2026-06-17 23:19:17,535][starVLA.training.train_starvla][INFO] - Step 3158, Loss: {'train/loss': 0.33479785919189453, 'train/grad_norm_pre_clip': 10.6419038772583, 'timing/data': 0.017221493646502495, 'timing/model': 1.0776733700186014, 'learning_rate/qwen_vl_interface': 6.0932809398434315e-06, 'learning_rate/action_model': 6.093280939843431e-05, 'epoch': 1.21}) +[2026-06-17 23:19:22,121][starVLA.training.train_starvla][INFO] - Step 3159, Loss: {'train/loss': 0.36801284551620483, 'train/grad_norm_pre_clip': 7.3707499504089355, 'timing/data': 0.01384899765253067, 'timing/model': 1.0693430621176958, 'learning_rate/qwen_vl_interface': 6.091152516518022e-06, 'learning_rate/action_model': 6.0911525165180216e-05, 'epoch': 1.22}) +[2026-06-17 23:19:27,094][starVLA.training.train_starvla][INFO] - Step 3160, Loss: {'train/loss': 0.4648398458957672, 'train/grad_norm_pre_clip': 9.991678237915039, 'timing/data': 0.012431489303708076, 'timing/model': 1.2575653661042452, 'learning_rate/qwen_vl_interface': 6.089023918820792e-06, 'learning_rate/action_model': 6.089023918820792e-05, 'epoch': 1.22}) +[2026-06-17 23:19:31,931][starVLA.training.train_starvla][INFO] - Step 3161, Loss: {'train/loss': 0.2418801188468933, 'train/grad_norm_pre_clip': 5.942282199859619, 'timing/data': 0.012502516619861126, 'timing/model': 1.0725117353722453, 'learning_rate/qwen_vl_interface': 6.0868951471929986e-06, 'learning_rate/action_model': 6.086895147192998e-05, 'epoch': 1.22}) +[2026-06-17 23:19:36,829][starVLA.training.train_starvla][INFO] - Step 3162, Loss: {'train/loss': 0.35305845737457275, 'train/grad_norm_pre_clip': 8.230962753295898, 'timing/data': 0.012466171756386757, 'timing/model': 1.5563661800697446, 'learning_rate/qwen_vl_interface': 6.084766202075944e-06, 'learning_rate/action_model': 6.0847662020759435e-05, 'epoch': 1.22}) +[2026-06-17 23:19:41,322][starVLA.training.train_starvla][INFO] - Step 3163, Loss: {'train/loss': 0.38326624035835266, 'train/grad_norm_pre_clip': 12.846360206604004, 'timing/data': 0.012586962431669235, 'timing/model': 1.117962472140789, 'learning_rate/qwen_vl_interface': 6.082637083910955e-06, 'learning_rate/action_model': 6.0826370839109546e-05, 'epoch': 1.22}) +[2026-06-17 23:19:46,309][starVLA.training.train_starvla][INFO] - Step 3164, Loss: {'train/loss': 0.32184869050979614, 'train/grad_norm_pre_clip': 9.18993091583252, 'timing/data': 0.013069577515125275, 'timing/model': 1.2875418895855546, 'learning_rate/qwen_vl_interface': 6.080507793139403e-06, 'learning_rate/action_model': 6.080507793139403e-05, 'epoch': 1.22}) +[2026-06-17 23:19:53,198][starVLA.training.train_starvla][INFO] - Step 3165, Loss: {'train/loss': 0.3705914616584778, 'train/grad_norm_pre_clip': 12.019932746887207, 'timing/data': 0.012590146623551846, 'timing/model': 3.2783930469304323, 'learning_rate/qwen_vl_interface': 6.0783783302026925e-06, 'learning_rate/action_model': 6.078378330202692e-05, 'epoch': 1.22}) +[2026-06-17 23:19:58,117][starVLA.training.train_starvla][INFO] - Step 3166, Loss: {'train/loss': 0.34771728515625, 'train/grad_norm_pre_clip': 9.692072868347168, 'timing/data': 0.01688751019537449, 'timing/model': 1.5272575365379453, 'learning_rate/qwen_vl_interface': 6.076248695542262e-06, 'learning_rate/action_model': 6.076248695542261e-05, 'epoch': 1.22}) +[2026-06-17 23:20:02,983][starVLA.training.train_starvla][INFO] - Step 3167, Loss: {'train/loss': 0.24863074719905853, 'train/grad_norm_pre_clip': 13.589739799499512, 'timing/data': 0.018811947666108608, 'timing/model': 1.5523470686748624, 'learning_rate/qwen_vl_interface': 6.074118889599589e-06, 'learning_rate/action_model': 6.0741188895995884e-05, 'epoch': 1.22}) +[2026-06-17 23:20:07,344][starVLA.training.train_starvla][INFO] - Step 3168, Loss: {'train/loss': 0.5145342946052551, 'train/grad_norm_pre_clip': 15.242383003234863, 'timing/data': 0.011866521090269089, 'timing/model': 1.1447896333411336, 'learning_rate/qwen_vl_interface': 6.07198891281618e-06, 'learning_rate/action_model': 6.07198891281618e-05, 'epoch': 1.22}) +[2026-06-17 23:20:12,042][starVLA.training.train_starvla][INFO] - Step 3169, Loss: {'train/loss': 0.24544599652290344, 'train/grad_norm_pre_clip': 7.794412612915039, 'timing/data': 0.01582823507487774, 'timing/model': 1.1068917056545615, 'learning_rate/qwen_vl_interface': 6.0698587656335895e-06, 'learning_rate/action_model': 6.069858765633589e-05, 'epoch': 1.22}) +[2026-06-17 23:20:16,741][starVLA.training.train_starvla][INFO] - Step 3170, Loss: {'train/loss': 0.24623499810695648, 'train/grad_norm_pre_clip': 8.450034141540527, 'timing/data': 0.012347116135060787, 'timing/model': 1.1733349906280637, 'learning_rate/qwen_vl_interface': 6.067728448493392e-06, 'learning_rate/action_model': 6.067728448493392e-05, 'epoch': 1.22}) +[2026-06-17 23:20:21,572][starVLA.training.train_starvla][INFO] - Step 3171, Loss: {'train/loss': 0.5287231206893921, 'train/grad_norm_pre_clip': 11.479564666748047, 'timing/data': 0.013657305389642715, 'timing/model': 1.1756966738030314, 'learning_rate/qwen_vl_interface': 6.065597961837209e-06, 'learning_rate/action_model': 6.0655979618372086e-05, 'epoch': 1.22}) +[2026-06-17 23:20:26,431][starVLA.training.train_starvla][INFO] - Step 3172, Loss: {'train/loss': 0.5750221014022827, 'train/grad_norm_pre_clip': 16.04656410217285, 'timing/data': 0.013756624422967434, 'timing/model': 1.5058721555396914, 'learning_rate/qwen_vl_interface': 6.063467306106688e-06, 'learning_rate/action_model': 6.0634673061066885e-05, 'epoch': 1.22}) +[2026-06-17 23:20:30,825][starVLA.training.train_starvla][INFO] - Step 3173, Loss: {'train/loss': 0.8070984482765198, 'train/grad_norm_pre_clip': 15.700551986694336, 'timing/data': 0.013258962891995907, 'timing/model': 1.1263555670157075, 'learning_rate/qwen_vl_interface': 6.061336481743522e-06, 'learning_rate/action_model': 6.061336481743522e-05, 'epoch': 1.22}) +[2026-06-17 23:20:35,780][starVLA.training.train_starvla][INFO] - Step 3174, Loss: {'train/loss': 0.24495387077331543, 'train/grad_norm_pre_clip': 7.0059709548950195, 'timing/data': 0.013803314417600632, 'timing/model': 1.1538077853620052, 'learning_rate/qwen_vl_interface': 6.059205489189429e-06, 'learning_rate/action_model': 6.059205489189429e-05, 'epoch': 1.22}) +[2026-06-17 23:20:41,756][starVLA.training.train_starvla][INFO] - Step 3175, Loss: {'train/loss': 0.7258479595184326, 'train/grad_norm_pre_clip': 12.320296287536621, 'timing/data': 0.01229456253349781, 'timing/model': 1.1150561226531863, 'learning_rate/qwen_vl_interface': 6.057074328886168e-06, 'learning_rate/action_model': 6.057074328886168e-05, 'epoch': 1.22}) +[2026-06-17 23:20:46,768][starVLA.training.train_starvla][INFO] - Step 3176, Loss: {'train/loss': 0.6228628158569336, 'train/grad_norm_pre_clip': 13.331400871276855, 'timing/data': 0.012645924463868141, 'timing/model': 1.6642421623691916, 'learning_rate/qwen_vl_interface': 6.0549430012755295e-06, 'learning_rate/action_model': 6.0549430012755295e-05, 'epoch': 1.22}) +[2026-06-17 23:20:51,626][starVLA.training.train_starvla][INFO] - Step 3177, Loss: {'train/loss': 0.40453267097473145, 'train/grad_norm_pre_clip': 12.078499794006348, 'timing/data': 0.012473375536501408, 'timing/model': 1.4640520755201578, 'learning_rate/qwen_vl_interface': 6.052811506799341e-06, 'learning_rate/action_model': 6.052811506799341e-05, 'epoch': 1.22}) +[2026-06-17 23:20:55,903][starVLA.training.train_starvla][INFO] - Step 3178, Loss: {'train/loss': 0.28637781739234924, 'train/grad_norm_pre_clip': 7.455251693725586, 'timing/data': 0.013191510923206806, 'timing/model': 1.1056703869253397, 'learning_rate/qwen_vl_interface': 6.050679845899462e-06, 'learning_rate/action_model': 6.050679845899462e-05, 'epoch': 1.22}) +[2026-06-17 23:21:00,683][starVLA.training.train_starvla][INFO] - Step 3179, Loss: {'train/loss': 0.5461472868919373, 'train/grad_norm_pre_clip': 14.589895248413086, 'timing/data': 0.016600916162133217, 'timing/model': 1.145838189870119, 'learning_rate/qwen_vl_interface': 6.048548019017791e-06, 'learning_rate/action_model': 6.04854801901779e-05, 'epoch': 1.22}) +[2026-06-17 23:21:07,308][starVLA.training.train_starvla][INFO] - Step 3180, Loss: {'train/loss': 0.5009188055992126, 'train/grad_norm_pre_clip': 13.663838386535645, 'timing/data': 0.028059697709977627, 'timing/model': 1.1355717768892646, 'learning_rate/qwen_vl_interface': 6.046416026596256e-06, 'learning_rate/action_model': 6.046416026596255e-05, 'epoch': 1.22}) +[2026-06-17 23:21:12,150][starVLA.training.train_starvla][INFO] - Step 3181, Loss: {'train/loss': 0.3147622346878052, 'train/grad_norm_pre_clip': 7.294777870178223, 'timing/data': 0.014625062234699726, 'timing/model': 1.1568343741819263, 'learning_rate/qwen_vl_interface': 6.044283869076822e-06, 'learning_rate/action_model': 6.0442838690768224e-05, 'epoch': 1.22}) +[2026-06-17 23:21:19,420][starVLA.training.train_starvla][INFO] - Step 3182, Loss: {'train/loss': 0.5654351711273193, 'train/grad_norm_pre_clip': 10.416595458984375, 'timing/data': 0.014433566480875015, 'timing/model': 1.506812633946538, 'learning_rate/qwen_vl_interface': 6.042151546901487e-06, 'learning_rate/action_model': 6.042151546901487e-05, 'epoch': 1.22}) +[2026-06-17 23:21:28,208][starVLA.training.train_starvla][INFO] - Step 3183, Loss: {'train/loss': 0.472299188375473, 'train/grad_norm_pre_clip': 7.730026721954346, 'timing/data': 0.013633495196700096, 'timing/model': 5.4418614925816655, 'learning_rate/qwen_vl_interface': 6.0400190605122846e-06, 'learning_rate/action_model': 6.040019060512284e-05, 'epoch': 1.22}) +[2026-06-17 23:21:34,225][starVLA.training.train_starvla][INFO] - Step 3184, Loss: {'train/loss': 0.23743195831775665, 'train/grad_norm_pre_clip': 9.292376518249512, 'timing/data': 0.013744532130658627, 'timing/model': 1.1178276734426618, 'learning_rate/qwen_vl_interface': 6.0378864103512805e-06, 'learning_rate/action_model': 6.03788641035128e-05, 'epoch': 1.22}) +[2026-06-17 23:21:39,377][starVLA.training.train_starvla][INFO] - Step 3185, Loss: {'train/loss': 0.38421401381492615, 'train/grad_norm_pre_clip': 8.78680419921875, 'timing/data': 0.012666980735957623, 'timing/model': 1.1509716343134642, 'learning_rate/qwen_vl_interface': 6.035753596860573e-06, 'learning_rate/action_model': 6.0357535968605725e-05, 'epoch': 1.23}) +[2026-06-17 23:21:45,831][starVLA.training.train_starvla][INFO] - Step 3186, Loss: {'train/loss': 0.35851043462753296, 'train/grad_norm_pre_clip': 12.283026695251465, 'timing/data': 0.016479254700243473, 'timing/model': 3.106026627123356, 'learning_rate/qwen_vl_interface': 6.033620620482303e-06, 'learning_rate/action_model': 6.033620620482303e-05, 'epoch': 1.23}) +[2026-06-17 23:21:52,935][starVLA.training.train_starvla][INFO] - Step 3187, Loss: {'train/loss': 0.5220969319343567, 'train/grad_norm_pre_clip': 16.093252182006836, 'timing/data': 0.018717648461461067, 'timing/model': 1.5181731013581157, 'learning_rate/qwen_vl_interface': 6.031487481658633e-06, 'learning_rate/action_model': 6.031487481658633e-05, 'epoch': 1.23}) +[2026-06-17 23:21:57,367][starVLA.training.train_starvla][INFO] - Step 3188, Loss: {'train/loss': 0.4763343036174774, 'train/grad_norm_pre_clip': 13.9949312210083, 'timing/data': 0.012476184405386448, 'timing/model': 1.094710111618042, 'learning_rate/qwen_vl_interface': 6.029354180831769e-06, 'learning_rate/action_model': 6.029354180831769e-05, 'epoch': 1.23}) +[2026-06-17 23:22:01,993][starVLA.training.train_starvla][INFO] - Step 3189, Loss: {'train/loss': 0.36514580249786377, 'train/grad_norm_pre_clip': 10.475423812866211, 'timing/data': 0.012867718935012817, 'timing/model': 1.1086904164403677, 'learning_rate/qwen_vl_interface': 6.027220718443943e-06, 'learning_rate/action_model': 6.027220718443943e-05, 'epoch': 1.23}) +[2026-06-17 23:22:06,667][starVLA.training.train_starvla][INFO] - Step 3190, Loss: {'train/loss': 0.23571458458900452, 'train/grad_norm_pre_clip': 8.041616439819336, 'timing/data': 0.0127505948767066, 'timing/model': 1.1530000790953636, 'learning_rate/qwen_vl_interface': 6.025087094937426e-06, 'learning_rate/action_model': 6.025087094937426e-05, 'epoch': 1.23}) +[2026-06-17 23:22:11,753][starVLA.training.train_starvla][INFO] - Step 3191, Loss: {'train/loss': 0.7047592401504517, 'train/grad_norm_pre_clip': 13.574250221252441, 'timing/data': 0.01410343311727047, 'timing/model': 1.1566464891657233, 'learning_rate/qwen_vl_interface': 6.022953310754521e-06, 'learning_rate/action_model': 6.022953310754521e-05, 'epoch': 1.23}) +[2026-06-17 23:22:16,535][starVLA.training.train_starvla][INFO] - Step 3192, Loss: {'train/loss': 0.3233698308467865, 'train/grad_norm_pre_clip': 8.156373023986816, 'timing/data': 0.01159622147679329, 'timing/model': 1.6500711282715201, 'learning_rate/qwen_vl_interface': 6.020819366337562e-06, 'learning_rate/action_model': 6.020819366337562e-05, 'epoch': 1.23}) +[2026-06-17 23:22:20,880][starVLA.training.train_starvla][INFO] - Step 3193, Loss: {'train/loss': 0.37889280915260315, 'train/grad_norm_pre_clip': 14.79444694519043, 'timing/data': 0.01391101349145174, 'timing/model': 1.0782862221822143, 'learning_rate/qwen_vl_interface': 6.018685262128921e-06, 'learning_rate/action_model': 6.01868526212892e-05, 'epoch': 1.23}) +[2026-06-17 23:22:25,841][starVLA.training.train_starvla][INFO] - Step 3194, Loss: {'train/loss': 0.2987205982208252, 'train/grad_norm_pre_clip': 14.943129539489746, 'timing/data': 0.011434345506131649, 'timing/model': 1.1916232500225306, 'learning_rate/qwen_vl_interface': 6.016550998570995e-06, 'learning_rate/action_model': 6.0165509985709944e-05, 'epoch': 1.23}) +[2026-06-17 23:22:30,666][starVLA.training.train_starvla][INFO] - Step 3195, Loss: {'train/loss': 0.36953404545783997, 'train/grad_norm_pre_clip': 9.1654691696167, 'timing/data': 0.017527957446873188, 'timing/model': 1.159938788972795, 'learning_rate/qwen_vl_interface': 6.014416576106224e-06, 'learning_rate/action_model': 6.0144165761062244e-05, 'epoch': 1.23}) +[2026-06-17 23:22:35,371][starVLA.training.train_starvla][INFO] - Step 3196, Loss: {'train/loss': 0.3523905277252197, 'train/grad_norm_pre_clip': 8.37566089630127, 'timing/data': 0.020565450191497803, 'timing/model': 1.4751122305169702, 'learning_rate/qwen_vl_interface': 6.012281995177074e-06, 'learning_rate/action_model': 6.012281995177074e-05, 'epoch': 1.23}) +[2026-06-17 23:22:40,183][starVLA.training.train_starvla][INFO] - Step 3197, Loss: {'train/loss': 0.5455766320228577, 'train/grad_norm_pre_clip': 11.000066757202148, 'timing/data': 0.015234124846756458, 'timing/model': 1.5627046069130301, 'learning_rate/qwen_vl_interface': 6.010147256226046e-06, 'learning_rate/action_model': 6.0101472562260454e-05, 'epoch': 1.23}) +[2026-06-17 23:22:44,643][starVLA.training.train_starvla][INFO] - Step 3198, Loss: {'train/loss': 0.3781033754348755, 'train/grad_norm_pre_clip': 12.113094329833984, 'timing/data': 0.01560969091951847, 'timing/model': 1.086974836885929, 'learning_rate/qwen_vl_interface': 6.008012359695673e-06, 'learning_rate/action_model': 6.0080123596956726e-05, 'epoch': 1.23}) +[2026-06-17 23:22:49,393][starVLA.training.train_starvla][INFO] - Step 3199, Loss: {'train/loss': 0.33438825607299805, 'train/grad_norm_pre_clip': 9.430414199829102, 'timing/data': 0.012894148007035255, 'timing/model': 1.0716101415455341, 'learning_rate/qwen_vl_interface': 6.005877306028522e-06, 'learning_rate/action_model': 6.005877306028522e-05, 'epoch': 1.23}) +[2026-06-17 23:26:52,244][starVLA.training.train_starvla][INFO] - Step 3200, Loss: {'train/loss': 0.3651178181171417, 'train/grad_norm_pre_clip': 9.934779167175293, 'eval/loss': 0.48475131392478943, 'eval/latency_4/loss': 0.48475131392478943, 'eval/demon_fire_precision': 0.9582417582417583, 'eval/demon_fire_recall': 0.9912309191295875, 'eval/demon_fire_f1': 0.9744572158365262, 'eval/demon_fire_rate_model': 0.9353891336270191, 'eval/demon_fire_rate_teacher': 0.9042584434654919, 'eval/demon_left_precision': 0.773841961852861, 'eval/demon_left_recall': 0.6484018264840182, 'eval/demon_left_f1': 0.7055900621118013, 'eval/demon_left_rate_model': 0.10778267254038179, 'eval/demon_left_rate_teacher': 0.12863436123348018, 'eval/demon_right_precision': 0.6891495601173021, 'eval/demon_right_recall': 0.8483754512635379, 'eval/demon_right_f1': 0.7605177993527509, 'eval/demon_right_rate_model': 0.20029368575624082, 'eval/demon_right_rate_teacher': 0.16270190895741557, 'eval/demon_move_f1': 0.7330539307322761, 'eval/demon_cc_f1': 0.8868428773834928, 'eval/latency_4/demon_fire_precision': 0.9582417582417583, 'eval/latency_4/demon_fire_recall': 0.9912309191295875, 'eval/latency_4/demon_fire_f1': 0.9744572158365262, 'eval/latency_4/demon_fire_rate_model': 0.9353891336270191, 'eval/latency_4/demon_fire_rate_teacher': 0.9042584434654919, 'eval/latency_4/demon_left_precision': 0.773841961852861, 'eval/latency_4/demon_left_recall': 0.6484018264840182, 'eval/latency_4/demon_left_f1': 0.7055900621118013, 'eval/latency_4/demon_left_rate_model': 0.10778267254038179, 'eval/latency_4/demon_left_rate_teacher': 0.12863436123348018, 'eval/latency_4/demon_right_precision': 0.6891495601173021, 'eval/latency_4/demon_right_recall': 0.8483754512635379, 'eval/latency_4/demon_right_f1': 0.7605177993527509, 'eval/latency_4/demon_right_rate_model': 0.20029368575624082, 'eval/latency_4/demon_right_rate_teacher': 0.16270190895741557, 'eval/latency_4/demon_move_f1': 0.7330539307322761, 'eval/latency_4/demon_cc_f1': 0.8868428773834928, 'timing/data': 0.01274342555552721, 'timing/model': 1.0805874597281218, 'learning_rate/qwen_vl_interface': 6.003742095667194e-06, 'learning_rate/action_model': 6.003742095667194e-05, 'epoch': 1.23}) +[2026-06-17 23:26:52,299][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_3200_state +[2026-06-17 23:26:52,314][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 23:27:01,532][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_3200_state/model.safetensors +[2026-06-17 23:27:24,999][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_3200_state/optimizer.bin +[2026-06-17 23:27:25,001][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_3200_state/sampler.bin +[2026-06-17 23:27:25,001][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_3200_state/sampler_1.bin +[2026-06-17 23:27:25,004][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_3200_state/random_states_0.pkl +[2026-06-17 23:27:25,010][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3200 +[2026-06-17 23:27:25,011][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 23:27:25,032][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 23:27:29,878][starVLA.training.train_starvla][INFO] - Step 3201, Loss: {'train/loss': 0.35372689366340637, 'train/grad_norm_pre_clip': 13.338720321655273, 'timing/data': 0.014129681512713432, 'timing/model': 1.6788290869444609, 'learning_rate/qwen_vl_interface': 6.001606729054318e-06, 'learning_rate/action_model': 6.0016067290543176e-05, 'epoch': 1.23}) +[2026-06-17 23:27:34,275][starVLA.training.train_starvla][INFO] - Step 3202, Loss: {'train/loss': 0.43067100644111633, 'train/grad_norm_pre_clip': 13.542783737182617, 'timing/data': 0.01160093117505312, 'timing/model': 1.1405158499255776, 'learning_rate/qwen_vl_interface': 5.999471206632559e-06, 'learning_rate/action_model': 5.999471206632559e-05, 'epoch': 1.23}) +[2026-06-17 23:27:39,349][starVLA.training.train_starvla][INFO] - Step 3203, Loss: {'train/loss': 0.4214901328086853, 'train/grad_norm_pre_clip': 13.600018501281738, 'timing/data': 0.013790019787847996, 'timing/model': 1.1234689112752676, 'learning_rate/qwen_vl_interface': 5.99733552884461e-06, 'learning_rate/action_model': 5.99733552884461e-05, 'epoch': 1.23}) +[2026-06-17 23:27:44,136][starVLA.training.train_starvla][INFO] - Step 3204, Loss: {'train/loss': 0.687523603439331, 'train/grad_norm_pre_clip': 13.821675300598145, 'timing/data': 0.011425242759287357, 'timing/model': 1.0693071195855737, 'learning_rate/qwen_vl_interface': 5.995199696133204e-06, 'learning_rate/action_model': 5.9951996961332034e-05, 'epoch': 1.23}) +[2026-06-17 23:27:48,952][starVLA.training.train_starvla][INFO] - Step 3205, Loss: {'train/loss': 0.3894091844558716, 'train/grad_norm_pre_clip': 9.767005920410156, 'timing/data': 0.0117035498842597, 'timing/model': 1.3554625241085887, 'learning_rate/qwen_vl_interface': 5.993063708941098e-06, 'learning_rate/action_model': 5.993063708941098e-05, 'epoch': 1.23}) +[2026-06-17 23:27:55,152][starVLA.training.train_starvla][INFO] - Step 3206, Loss: {'train/loss': 0.5825156569480896, 'train/grad_norm_pre_clip': 13.820528984069824, 'timing/data': 0.011477452702820301, 'timing/model': 2.6117157572880387, 'learning_rate/qwen_vl_interface': 5.990927567711085e-06, 'learning_rate/action_model': 5.990927567711085e-05, 'epoch': 1.23}) +[2026-06-17 23:28:00,297][starVLA.training.train_starvla][INFO] - Step 3207, Loss: {'train/loss': 0.6104990839958191, 'train/grad_norm_pre_clip': 14.385214805603027, 'timing/data': 0.015950042754411697, 'timing/model': 1.627834914252162, 'learning_rate/qwen_vl_interface': 5.988791272885991e-06, 'learning_rate/action_model': 5.988791272885991e-05, 'epoch': 1.23}) +[2026-06-17 23:28:05,015][starVLA.training.train_starvla][INFO] - Step 3208, Loss: {'train/loss': 0.27352428436279297, 'train/grad_norm_pre_clip': 8.911981582641602, 'timing/data': 0.012537427246570587, 'timing/model': 1.1264332309365273, 'learning_rate/qwen_vl_interface': 5.986654824908671e-06, 'learning_rate/action_model': 5.9866548249086706e-05, 'epoch': 1.23}) +[2026-06-17 23:28:09,833][starVLA.training.train_starvla][INFO] - Step 3209, Loss: {'train/loss': 0.3757394552230835, 'train/grad_norm_pre_clip': 13.209821701049805, 'timing/data': 0.013498139567673206, 'timing/model': 1.1071848906576633, 'learning_rate/qwen_vl_interface': 5.9845182242220125e-06, 'learning_rate/action_model': 5.984518224222012e-05, 'epoch': 1.23}) +[2026-06-17 23:28:15,522][starVLA.training.train_starvla][INFO] - Step 3210, Loss: {'train/loss': 0.20043055713176727, 'train/grad_norm_pre_clip': 7.2978668212890625, 'timing/data': 0.012215381488204002, 'timing/model': 1.152872416190803, 'learning_rate/qwen_vl_interface': 5.982381471268935e-06, 'learning_rate/action_model': 5.9823814712689354e-05, 'epoch': 1.23}) +[2026-06-17 23:28:20,309][starVLA.training.train_starvla][INFO] - Step 3211, Loss: {'train/loss': 0.476290762424469, 'train/grad_norm_pre_clip': 11.639481544494629, 'timing/data': 0.013452389277517796, 'timing/model': 1.1752573940902948, 'learning_rate/qwen_vl_interface': 5.980244566492392e-06, 'learning_rate/action_model': 5.9802445664923925e-05, 'epoch': 1.24}) +[2026-06-17 23:28:25,146][starVLA.training.train_starvla][INFO] - Step 3212, Loss: {'train/loss': 0.27282601594924927, 'train/grad_norm_pre_clip': 10.312349319458008, 'timing/data': 0.02906856592744589, 'timing/model': 1.526767990551889, 'learning_rate/qwen_vl_interface': 5.978107510335365e-06, 'learning_rate/action_model': 5.9781075103353647e-05, 'epoch': 1.24}) +[2026-06-17 23:28:29,862][starVLA.training.train_starvla][INFO] - Step 3213, Loss: {'train/loss': 0.42571407556533813, 'train/grad_norm_pre_clip': 11.896661758422852, 'timing/data': 0.01867289189249277, 'timing/model': 1.177611413411796, 'learning_rate/qwen_vl_interface': 5.975970303240867e-06, 'learning_rate/action_model': 5.9759703032408664e-05, 'epoch': 1.24}) +[2026-06-17 23:28:34,589][starVLA.training.train_starvla][INFO] - Step 3214, Loss: {'train/loss': 0.35506197810173035, 'train/grad_norm_pre_clip': 10.913833618164062, 'timing/data': 0.01126535888761282, 'timing/model': 1.110235370695591, 'learning_rate/qwen_vl_interface': 5.973832945651944e-06, 'learning_rate/action_model': 5.9738329456519435e-05, 'epoch': 1.24}) +[2026-06-17 23:28:39,401][starVLA.training.train_starvla][INFO] - Step 3215, Loss: {'train/loss': 0.5743587613105774, 'train/grad_norm_pre_clip': 10.620040893554688, 'timing/data': 0.014089779928326607, 'timing/model': 1.0804165415465832, 'learning_rate/qwen_vl_interface': 5.971695438011672e-06, 'learning_rate/action_model': 5.971695438011672e-05, 'epoch': 1.24}) +[2026-06-17 23:28:44,012][starVLA.training.train_starvla][INFO] - Step 3216, Loss: {'train/loss': 0.39319872856140137, 'train/grad_norm_pre_clip': 8.199149131774902, 'timing/data': 0.011946615763008595, 'timing/model': 1.096066327765584, 'learning_rate/qwen_vl_interface': 5.969557780763159e-06, 'learning_rate/action_model': 5.9695577807631585e-05, 'epoch': 1.24}) +[2026-06-17 23:28:48,487][starVLA.training.train_starvla][INFO] - Step 3217, Loss: {'train/loss': 0.31178197264671326, 'train/grad_norm_pre_clip': 8.670652389526367, 'timing/data': 0.01207085233181715, 'timing/model': 1.2192139122635126, 'learning_rate/qwen_vl_interface': 5.9674199743495456e-06, 'learning_rate/action_model': 5.967419974349545e-05, 'epoch': 1.24}) +[2026-06-17 23:28:53,795][starVLA.training.train_starvla][INFO] - Step 3218, Loss: {'train/loss': 0.4577866196632385, 'train/grad_norm_pre_clip': 9.35932731628418, 'timing/data': 0.011564419604837894, 'timing/model': 1.106195910833776, 'learning_rate/qwen_vl_interface': 5.965282019213996e-06, 'learning_rate/action_model': 5.965282019213996e-05, 'epoch': 1.24}) +[2026-06-17 23:28:58,482][starVLA.training.train_starvla][INFO] - Step 3219, Loss: {'train/loss': 0.6605678796768188, 'train/grad_norm_pre_clip': 14.024420738220215, 'timing/data': 0.014398436062037945, 'timing/model': 1.0781704066321254, 'learning_rate/qwen_vl_interface': 5.963143915799718e-06, 'learning_rate/action_model': 5.9631439157997184e-05, 'epoch': 1.24}) +[2026-06-17 23:29:03,134][starVLA.training.train_starvla][INFO] - Step 3220, Loss: {'train/loss': 0.47637951374053955, 'train/grad_norm_pre_clip': 10.430501937866211, 'timing/data': 0.012682303786277771, 'timing/model': 1.174433839507401, 'learning_rate/qwen_vl_interface': 5.961005664549936e-06, 'learning_rate/action_model': 5.9610056645499354e-05, 'epoch': 1.24}) +[2026-06-17 23:29:14,293][starVLA.training.train_starvla][INFO] - Step 3221, Loss: {'train/loss': 0.37339508533477783, 'train/grad_norm_pre_clip': 11.167546272277832, 'timing/data': 0.011618704535067081, 'timing/model': 1.4902268555015326, 'learning_rate/qwen_vl_interface': 5.958867265907916e-06, 'learning_rate/action_model': 5.958867265907916e-05, 'epoch': 1.24}) +[2026-06-17 23:29:24,933][starVLA.training.train_starvla][INFO] - Step 3222, Loss: {'train/loss': 0.4882710874080658, 'train/grad_norm_pre_clip': 11.562294960021973, 'timing/data': 0.013128825463354588, 'timing/model': 1.6100541148334742, 'learning_rate/qwen_vl_interface': 5.956728720316949e-06, 'learning_rate/action_model': 5.9567287203169485e-05, 'epoch': 1.24}) +[2026-06-17 23:29:29,807][starVLA.training.train_starvla][INFO] - Step 3223, Loss: {'train/loss': 0.3584604859352112, 'train/grad_norm_pre_clip': 11.616640090942383, 'timing/data': 0.011991245672106743, 'timing/model': 1.255455806851387, 'learning_rate/qwen_vl_interface': 5.954590028220357e-06, 'learning_rate/action_model': 5.954590028220357e-05, 'epoch': 1.24}) +[2026-06-17 23:29:36,977][starVLA.training.train_starvla][INFO] - Step 3224, Loss: {'train/loss': 0.5310695171356201, 'train/grad_norm_pre_clip': 13.228251457214355, 'timing/data': 0.012340211309492588, 'timing/model': 2.2975517492741346, 'learning_rate/qwen_vl_interface': 5.952451190061496e-06, 'learning_rate/action_model': 5.9524511900614956e-05, 'epoch': 1.24}) +[2026-06-17 23:29:42,119][starVLA.training.train_starvla][INFO] - Step 3225, Loss: {'train/loss': 0.4006332457065582, 'train/grad_norm_pre_clip': 9.477788925170898, 'timing/data': 0.012038499116897583, 'timing/model': 1.1763938069343567, 'learning_rate/qwen_vl_interface': 5.950312206283743e-06, 'learning_rate/action_model': 5.9503122062837434e-05, 'epoch': 1.24}) +[2026-06-17 23:29:47,116][starVLA.training.train_starvla][INFO] - Step 3226, Loss: {'train/loss': 0.3645262122154236, 'train/grad_norm_pre_clip': 8.930803298950195, 'timing/data': 0.014295161701738834, 'timing/model': 1.1995445862412453, 'learning_rate/qwen_vl_interface': 5.948173077330522e-06, 'learning_rate/action_model': 5.948173077330522e-05, 'epoch': 1.24}) +[2026-06-17 23:29:51,853][starVLA.training.train_starvla][INFO] - Step 3227, Loss: {'train/loss': 0.4002116620540619, 'train/grad_norm_pre_clip': 8.283559799194336, 'timing/data': 0.016300706192851067, 'timing/model': 1.4718084251508117, 'learning_rate/qwen_vl_interface': 5.9460338036452665e-06, 'learning_rate/action_model': 5.946033803645267e-05, 'epoch': 1.24}) +[2026-06-17 23:29:56,418][starVLA.training.train_starvla][INFO] - Step 3228, Loss: {'train/loss': 0.6286793947219849, 'train/grad_norm_pre_clip': 12.553903579711914, 'timing/data': 0.013368269428610802, 'timing/model': 1.329584707506001, 'learning_rate/qwen_vl_interface': 5.943894385671456e-06, 'learning_rate/action_model': 5.9438943856714556e-05, 'epoch': 1.24}) +[2026-06-17 23:30:02,066][starVLA.training.train_starvla][INFO] - Step 3229, Loss: {'train/loss': 0.38663867115974426, 'train/grad_norm_pre_clip': 8.51893424987793, 'timing/data': 0.013135784305632114, 'timing/model': 1.9817723808810115, 'learning_rate/qwen_vl_interface': 5.941754823852593e-06, 'learning_rate/action_model': 5.9417548238525924e-05, 'epoch': 1.24}) +[2026-06-17 23:30:06,947][starVLA.training.train_starvla][INFO] - Step 3230, Loss: {'train/loss': 0.2754761278629303, 'train/grad_norm_pre_clip': 6.157220840454102, 'timing/data': 0.012019230984151363, 'timing/model': 1.1777801057323813, 'learning_rate/qwen_vl_interface': 5.939615118632208e-06, 'learning_rate/action_model': 5.939615118632208e-05, 'epoch': 1.24}) +[2026-06-17 23:30:12,476][starVLA.training.train_starvla][INFO] - Step 3231, Loss: {'train/loss': 0.3610847592353821, 'train/grad_norm_pre_clip': 8.508841514587402, 'timing/data': 0.015426340512931347, 'timing/model': 1.1296403631567955, 'learning_rate/qwen_vl_interface': 5.937475270453872e-06, 'learning_rate/action_model': 5.937475270453872e-05, 'epoch': 1.24}) +[2026-06-17 23:30:17,211][starVLA.training.train_starvla][INFO] - Step 3232, Loss: {'train/loss': 0.265687495470047, 'train/grad_norm_pre_clip': 8.316121101379395, 'timing/data': 0.015245061367750168, 'timing/model': 1.4827602952718735, 'learning_rate/qwen_vl_interface': 5.935335279761169e-06, 'learning_rate/action_model': 5.935335279761168e-05, 'epoch': 1.24}) +[2026-06-17 23:30:21,880][starVLA.training.train_starvla][INFO] - Step 3233, Loss: {'train/loss': 0.29649871587753296, 'train/grad_norm_pre_clip': 8.056612968444824, 'timing/data': 0.016028763726353645, 'timing/model': 1.1375169530510902, 'learning_rate/qwen_vl_interface': 5.933195146997725e-06, 'learning_rate/action_model': 5.9331951469977254e-05, 'epoch': 1.24}) +[2026-06-17 23:30:27,011][starVLA.training.train_starvla][INFO] - Step 3234, Loss: {'train/loss': 0.41697636246681213, 'train/grad_norm_pre_clip': 10.568638801574707, 'timing/data': 0.02207592222839594, 'timing/model': 1.1851643584668636, 'learning_rate/qwen_vl_interface': 5.931054872607193e-06, 'learning_rate/action_model': 5.9310548726071925e-05, 'epoch': 1.24}) +[2026-06-17 23:30:31,929][starVLA.training.train_starvla][INFO] - Step 3235, Loss: {'train/loss': 0.568710446357727, 'train/grad_norm_pre_clip': 10.561751365661621, 'timing/data': 0.011892339214682579, 'timing/model': 1.1396114761009812, 'learning_rate/qwen_vl_interface': 5.928914457033252e-06, 'learning_rate/action_model': 5.928914457033252e-05, 'epoch': 1.24}) +[2026-06-17 23:30:36,834][starVLA.training.train_starvla][INFO] - Step 3236, Loss: {'train/loss': 0.4451991319656372, 'train/grad_norm_pre_clip': 10.172403335571289, 'timing/data': 0.012614653445780277, 'timing/model': 1.4544154154136777, 'learning_rate/qwen_vl_interface': 5.926773900719616e-06, 'learning_rate/action_model': 5.9267739007196157e-05, 'epoch': 1.24}) +[2026-06-17 23:30:45,554][starVLA.training.train_starvla][INFO] - Step 3237, Loss: {'train/loss': 0.33971118927001953, 'train/grad_norm_pre_clip': 9.672945022583008, 'timing/data': 0.016029929742217064, 'timing/model': 1.5371614582836628, 'learning_rate/qwen_vl_interface': 5.92463320411002e-06, 'learning_rate/action_model': 5.92463320411002e-05, 'epoch': 1.25}) +[2026-06-17 23:30:50,281][starVLA.training.train_starvla][INFO] - Step 3238, Loss: {'train/loss': 0.5364262461662292, 'train/grad_norm_pre_clip': 12.404996871948242, 'timing/data': 0.03327414393424988, 'timing/model': 1.3135597417131066, 'learning_rate/qwen_vl_interface': 5.922492367648236e-06, 'learning_rate/action_model': 5.922492367648236e-05, 'epoch': 1.25}) +[2026-06-17 23:30:55,028][starVLA.training.train_starvla][INFO] - Step 3239, Loss: {'train/loss': 0.40410009026527405, 'train/grad_norm_pre_clip': 14.008938789367676, 'timing/data': 0.01662987284362316, 'timing/model': 1.1002865629270673, 'learning_rate/qwen_vl_interface': 5.920351391778059e-06, 'learning_rate/action_model': 5.920351391778059e-05, 'epoch': 1.25}) +[2026-06-17 23:31:03,064][starVLA.training.train_starvla][INFO] - Step 3240, Loss: {'train/loss': 0.3440271019935608, 'train/grad_norm_pre_clip': 8.796860694885254, 'timing/data': 0.012968016788363457, 'timing/model': 1.1445328574627638, 'learning_rate/qwen_vl_interface': 5.918210276943318e-06, 'learning_rate/action_model': 5.918210276943318e-05, 'epoch': 1.25}) +[2026-06-17 23:31:07,998][starVLA.training.train_starvla][INFO] - Step 3241, Loss: {'train/loss': 0.2296563684940338, 'train/grad_norm_pre_clip': 6.508836269378662, 'timing/data': 0.014610456302762032, 'timing/model': 1.3084127819165587, 'learning_rate/qwen_vl_interface': 5.916069023587867e-06, 'learning_rate/action_model': 5.916069023587867e-05, 'epoch': 1.25}) +[2026-06-17 23:31:12,864][starVLA.training.train_starvla][INFO] - Step 3242, Loss: {'train/loss': 0.24509422481060028, 'train/grad_norm_pre_clip': 7.424832820892334, 'timing/data': 0.012602810747921467, 'timing/model': 1.586031205020845, 'learning_rate/qwen_vl_interface': 5.913927632155592e-06, 'learning_rate/action_model': 5.913927632155591e-05, 'epoch': 1.25}) +[2026-06-17 23:31:17,365][starVLA.training.train_starvla][INFO] - Step 3243, Loss: {'train/loss': 0.3506360352039337, 'train/grad_norm_pre_clip': 10.104557991027832, 'timing/data': 0.013669432140886784, 'timing/model': 1.1250370051711798, 'learning_rate/qwen_vl_interface': 5.911786103090402e-06, 'learning_rate/action_model': 5.911786103090402e-05, 'epoch': 1.25}) +[2026-06-17 23:31:22,688][starVLA.training.train_starvla][INFO] - Step 3244, Loss: {'train/loss': 0.3961566984653473, 'train/grad_norm_pre_clip': 10.790164947509766, 'timing/data': 0.015258805826306343, 'timing/model': 1.2337595662102103, 'learning_rate/qwen_vl_interface': 5.909644436836243e-06, 'learning_rate/action_model': 5.909644436836244e-05, 'epoch': 1.25}) +[2026-06-17 23:31:34,709][starVLA.training.train_starvla][INFO] - Step 3245, Loss: {'train/loss': 0.46206846833229065, 'train/grad_norm_pre_clip': 11.116707801818848, 'timing/data': 0.0146098081022501, 'timing/model': 1.4496635785326362, 'learning_rate/qwen_vl_interface': 5.907502633837082e-06, 'learning_rate/action_model': 5.907502633837082e-05, 'epoch': 1.25}) +[2026-06-17 23:31:39,923][starVLA.training.train_starvla][INFO] - Step 3246, Loss: {'train/loss': 0.3971545696258545, 'train/grad_norm_pre_clip': 15.425652503967285, 'timing/data': 0.01822436135262251, 'timing/model': 1.6784230433404446, 'learning_rate/qwen_vl_interface': 5.905360694536918e-06, 'learning_rate/action_model': 5.905360694536918e-05, 'epoch': 1.25}) +[2026-06-17 23:31:46,955][starVLA.training.train_starvla][INFO] - Step 3247, Loss: {'train/loss': 0.48808473348617554, 'train/grad_norm_pre_clip': 10.60306167602539, 'timing/data': 0.012770420871675014, 'timing/model': 1.6838979749009013, 'learning_rate/qwen_vl_interface': 5.903218619379776e-06, 'learning_rate/action_model': 5.903218619379776e-05, 'epoch': 1.25}) +[2026-06-17 23:31:52,312][starVLA.training.train_starvla][INFO] - Step 3248, Loss: {'train/loss': 0.36068204045295715, 'train/grad_norm_pre_clip': 10.337295532226562, 'timing/data': 0.01235886663198471, 'timing/model': 1.3360090553760529, 'learning_rate/qwen_vl_interface': 5.901076408809712e-06, 'learning_rate/action_model': 5.9010764088097114e-05, 'epoch': 1.25}) +[2026-06-17 23:32:00,993][starVLA.training.train_starvla][INFO] - Step 3249, Loss: {'train/loss': 0.6523631811141968, 'train/grad_norm_pre_clip': 20.104867935180664, 'timing/data': 0.013013503514230251, 'timing/model': 1.2864522375166416, 'learning_rate/qwen_vl_interface': 5.898934063270806e-06, 'learning_rate/action_model': 5.8989340632708056e-05, 'epoch': 1.25}) +[2026-06-17 23:32:13,878][starVLA.training.train_starvla][INFO] - Step 3250, Loss: {'train/loss': 0.3515773415565491, 'train/grad_norm_pre_clip': 10.842761993408203, 'timing/data': 0.014944080263376236, 'timing/model': 4.384179985150695, 'learning_rate/qwen_vl_interface': 5.896791583207173e-06, 'learning_rate/action_model': 5.896791583207173e-05, 'epoch': 1.25}) +[2026-06-17 23:32:19,698][starVLA.training.train_starvla][INFO] - Step 3251, Loss: {'train/loss': 0.5558332204818726, 'train/grad_norm_pre_clip': 12.892589569091797, 'timing/data': 0.012651789002120495, 'timing/model': 1.358206955716014, 'learning_rate/qwen_vl_interface': 5.894648969062947e-06, 'learning_rate/action_model': 5.894648969062947e-05, 'epoch': 1.25}) +[2026-06-17 23:32:39,639][starVLA.training.train_starvla][INFO] - Step 3252, Loss: {'train/loss': 0.1861129105091095, 'train/grad_norm_pre_clip': 6.659151554107666, 'timing/data': 0.013879300095140934, 'timing/model': 5.515896033495665, 'learning_rate/qwen_vl_interface': 5.892506221282297e-06, 'learning_rate/action_model': 5.892506221282297e-05, 'epoch': 1.25}) +[2026-06-17 23:32:44,649][starVLA.training.train_starvla][INFO] - Step 3253, Loss: {'train/loss': 0.3152252733707428, 'train/grad_norm_pre_clip': 9.73697566986084, 'timing/data': 0.014423768036067486, 'timing/model': 1.1397176729515195, 'learning_rate/qwen_vl_interface': 5.890363340309415e-06, 'learning_rate/action_model': 5.890363340309415e-05, 'epoch': 1.25}) +[2026-06-17 23:32:53,357][starVLA.training.train_starvla][INFO] - Step 3254, Loss: {'train/loss': 0.5299037098884583, 'train/grad_norm_pre_clip': 9.778817176818848, 'timing/data': 0.013402491807937622, 'timing/model': 1.2885340554639697, 'learning_rate/qwen_vl_interface': 5.888220326588523e-06, 'learning_rate/action_model': 5.8882203265885226e-05, 'epoch': 1.25}) +[2026-06-17 23:33:03,838][starVLA.training.train_starvla][INFO] - Step 3255, Loss: {'train/loss': 0.6139901876449585, 'train/grad_norm_pre_clip': 24.88359260559082, 'timing/data': 0.016607502475380898, 'timing/model': 2.8614961951971054, 'learning_rate/qwen_vl_interface': 5.8860771805638725e-06, 'learning_rate/action_model': 5.886077180563872e-05, 'epoch': 1.25}) +[2026-06-17 23:33:08,817][starVLA.training.train_starvla][INFO] - Step 3256, Loss: {'train/loss': 0.4287150204181671, 'train/grad_norm_pre_clip': 12.054522514343262, 'timing/data': 0.011795719154179096, 'timing/model': 1.589687168598175, 'learning_rate/qwen_vl_interface': 5.883933902679736e-06, 'learning_rate/action_model': 5.8839339026797355e-05, 'epoch': 1.25}) +[2026-06-17 23:33:14,424][starVLA.training.train_starvla][INFO] - Step 3257, Loss: {'train/loss': 0.21604929864406586, 'train/grad_norm_pre_clip': 8.838027954101562, 'timing/data': 0.01154383271932602, 'timing/model': 1.8990470673888922, 'learning_rate/qwen_vl_interface': 5.8817904933804214e-06, 'learning_rate/action_model': 5.881790493380421e-05, 'epoch': 1.25}) +[2026-06-17 23:33:20,212][starVLA.training.train_starvla][INFO] - Step 3258, Loss: {'train/loss': 0.5688743591308594, 'train/grad_norm_pre_clip': 13.192671775817871, 'timing/data': 0.012454281561076641, 'timing/model': 2.506095165386796, 'learning_rate/qwen_vl_interface': 5.8796469531102565e-06, 'learning_rate/action_model': 5.879646953110256e-05, 'epoch': 1.25}) +[2026-06-17 23:33:25,042][starVLA.training.train_starvla][INFO] - Step 3259, Loss: {'train/loss': 0.2564002275466919, 'train/grad_norm_pre_clip': 11.161293983459473, 'timing/data': 0.013510618358850479, 'timing/model': 1.087325455620885, 'learning_rate/qwen_vl_interface': 5.877503282313601e-06, 'learning_rate/action_model': 5.877503282313601e-05, 'epoch': 1.25}) +[2026-06-17 23:33:31,163][starVLA.training.train_starvla][INFO] - Step 3260, Loss: {'train/loss': 0.5771687626838684, 'train/grad_norm_pre_clip': 11.17064094543457, 'timing/data': 0.013552909716963768, 'timing/model': 1.3079378390684724, 'learning_rate/qwen_vl_interface': 5.875359481434839e-06, 'learning_rate/action_model': 5.875359481434839e-05, 'epoch': 1.25}) +[2026-06-17 23:33:38,415][starVLA.training.train_starvla][INFO] - Step 3261, Loss: {'train/loss': 0.36341720819473267, 'train/grad_norm_pre_clip': 7.747138500213623, 'timing/data': 0.01583992224186659, 'timing/model': 3.294856294989586, 'learning_rate/qwen_vl_interface': 5.873215550918384e-06, 'learning_rate/action_model': 5.873215550918384e-05, 'epoch': 1.25}) +[2026-06-17 23:33:47,510][starVLA.training.train_starvla][INFO] - Step 3262, Loss: {'train/loss': 0.4544307589530945, 'train/grad_norm_pre_clip': 12.8709077835083, 'timing/data': 0.012424184940755367, 'timing/model': 5.760296464897692, 'learning_rate/qwen_vl_interface': 5.871071491208676e-06, 'learning_rate/action_model': 5.871071491208676e-05, 'epoch': 1.25}) +[2026-06-17 23:33:52,047][starVLA.training.train_starvla][INFO] - Step 3263, Loss: {'train/loss': 0.4134794771671295, 'train/grad_norm_pre_clip': 10.603795051574707, 'timing/data': 0.01435801014304161, 'timing/model': 1.1896642781794071, 'learning_rate/qwen_vl_interface': 5.868927302750178e-06, 'learning_rate/action_model': 5.868927302750178e-05, 'epoch': 1.26}) +[2026-06-17 23:33:57,080][starVLA.training.train_starvla][INFO] - Step 3264, Loss: {'train/loss': 0.1457437127828598, 'train/grad_norm_pre_clip': 5.943021297454834, 'timing/data': 0.013217143714427948, 'timing/model': 1.194444976747036, 'learning_rate/qwen_vl_interface': 5.866782985987386e-06, 'learning_rate/action_model': 5.866782985987386e-05, 'epoch': 1.26}) +[2026-06-17 23:34:03,601][starVLA.training.train_starvla][INFO] - Step 3265, Loss: {'train/loss': 0.7265762686729431, 'train/grad_norm_pre_clip': 17.478382110595703, 'timing/data': 0.017393244430422783, 'timing/model': 2.7603648509830236, 'learning_rate/qwen_vl_interface': 5.864638541364815e-06, 'learning_rate/action_model': 5.8646385413648154e-05, 'epoch': 1.26}) +[2026-06-17 23:34:08,923][starVLA.training.train_starvla][INFO] - Step 3266, Loss: {'train/loss': 0.35007244348526, 'train/grad_norm_pre_clip': 11.836885452270508, 'timing/data': 0.011770005337893963, 'timing/model': 1.5550573132932186, 'learning_rate/qwen_vl_interface': 5.862493969327015e-06, 'learning_rate/action_model': 5.862493969327014e-05, 'epoch': 1.26}) +[2026-06-17 23:34:14,611][starVLA.training.train_starvla][INFO] - Step 3267, Loss: {'train/loss': 0.4442782402038574, 'train/grad_norm_pre_clip': 9.957671165466309, 'timing/data': 0.016890074126422405, 'timing/model': 1.5156873911619186, 'learning_rate/qwen_vl_interface': 5.860349270318555e-06, 'learning_rate/action_model': 5.860349270318555e-05, 'epoch': 1.26}) +[2026-06-17 23:34:19,515][starVLA.training.train_starvla][INFO] - Step 3268, Loss: {'train/loss': 0.4024854600429535, 'train/grad_norm_pre_clip': 10.306137084960938, 'timing/data': 0.01574739348143339, 'timing/model': 1.0997297456488013, 'learning_rate/qwen_vl_interface': 5.8582044447840346e-06, 'learning_rate/action_model': 5.858204444784034e-05, 'epoch': 1.26}) +[2026-06-17 23:34:26,857][starVLA.training.train_starvla][INFO] - Step 3269, Loss: {'train/loss': 0.46486952900886536, 'train/grad_norm_pre_clip': 11.292303085327148, 'timing/data': 0.01586809568107128, 'timing/model': 2.0619282498955727, 'learning_rate/qwen_vl_interface': 5.8560594931680785e-06, 'learning_rate/action_model': 5.856059493168078e-05, 'epoch': 1.26}) +[2026-06-17 23:34:38,517][starVLA.training.train_starvla][INFO] - Step 3270, Loss: {'train/loss': 0.36889076232910156, 'train/grad_norm_pre_clip': 8.307990074157715, 'timing/data': 0.06089212652295828, 'timing/model': 1.6191707160323858, 'learning_rate/qwen_vl_interface': 5.853914415915337e-06, 'learning_rate/action_model': 5.853914415915337e-05, 'epoch': 1.26}) +[2026-06-17 23:34:43,456][starVLA.training.train_starvla][INFO] - Step 3271, Loss: {'train/loss': 0.6544166803359985, 'train/grad_norm_pre_clip': 14.879520416259766, 'timing/data': 0.061019862070679665, 'timing/model': 1.1168287200853229, 'learning_rate/qwen_vl_interface': 5.8517692134704875e-06, 'learning_rate/action_model': 5.851769213470487e-05, 'epoch': 1.26}) +[2026-06-17 23:34:51,820][starVLA.training.train_starvla][INFO] - Step 3272, Loss: {'train/loss': 0.18980267643928528, 'train/grad_norm_pre_clip': 6.436436653137207, 'timing/data': 0.05319666210561991, 'timing/model': 1.4899960365146399, 'learning_rate/qwen_vl_interface': 5.849623886278232e-06, 'learning_rate/action_model': 5.849623886278231e-05, 'epoch': 1.26}) +[2026-06-17 23:34:57,177][starVLA.training.train_starvla][INFO] - Step 3273, Loss: {'train/loss': 0.6898558735847473, 'train/grad_norm_pre_clip': 15.87846851348877, 'timing/data': 0.03883779235184193, 'timing/model': 1.3662937367334962, 'learning_rate/qwen_vl_interface': 5.8474784347833e-06, 'learning_rate/action_model': 5.8474784347833e-05, 'epoch': 1.26}) +[2026-06-17 23:35:10,030][starVLA.training.train_starvla][INFO] - Step 3274, Loss: {'train/loss': 0.11319704353809357, 'train/grad_norm_pre_clip': 4.457523822784424, 'timing/data': 0.045257529243826866, 'timing/model': 4.697907799854875, 'learning_rate/qwen_vl_interface': 5.845332859430446e-06, 'learning_rate/action_model': 5.845332859430446e-05, 'epoch': 1.26}) +[2026-06-17 23:35:21,835][starVLA.training.train_starvla][INFO] - Step 3275, Loss: {'train/loss': 0.5763360857963562, 'train/grad_norm_pre_clip': 12.184165000915527, 'timing/data': 0.05508721340447664, 'timing/model': 3.45481935236603, 'learning_rate/qwen_vl_interface': 5.843187160664449e-06, 'learning_rate/action_model': 5.8431871606644495e-05, 'epoch': 1.26}) +[2026-06-17 23:35:38,420][starVLA.training.train_starvla][INFO] - Step 3276, Loss: {'train/loss': 0.3859029710292816, 'train/grad_norm_pre_clip': 10.513437271118164, 'timing/data': 0.102072199806571, 'timing/model': 3.6970636369660497, 'learning_rate/qwen_vl_interface': 5.841041338930115e-06, 'learning_rate/action_model': 5.8410413389301155e-05, 'epoch': 1.26}) +[2026-06-17 23:35:49,855][starVLA.training.train_starvla][INFO] - Step 3277, Loss: {'train/loss': 0.8740847706794739, 'train/grad_norm_pre_clip': 15.064738273620605, 'timing/data': 0.05307038128376007, 'timing/model': 1.1597827291116118, 'learning_rate/qwen_vl_interface': 5.8388953946722775e-06, 'learning_rate/action_model': 5.8388953946722766e-05, 'epoch': 1.26}) +[2026-06-17 23:36:03,847][starVLA.training.train_starvla][INFO] - Step 3278, Loss: {'train/loss': 0.5204906463623047, 'train/grad_norm_pre_clip': 12.970722198486328, 'timing/data': 0.03825704474002123, 'timing/model': 5.360339595936239, 'learning_rate/qwen_vl_interface': 5.836749328335789e-06, 'learning_rate/action_model': 5.8367493283357885e-05, 'epoch': 1.26}) +[2026-06-17 23:36:17,941][starVLA.training.train_starvla][INFO] - Step 3279, Loss: {'train/loss': 0.26082363724708557, 'train/grad_norm_pre_clip': 9.1801118850708, 'timing/data': 0.05033613182604313, 'timing/model': 4.563354032114148, 'learning_rate/qwen_vl_interface': 5.8346031403655345e-06, 'learning_rate/action_model': 5.8346031403655345e-05, 'epoch': 1.26}) +[2026-06-17 23:36:29,967][starVLA.training.train_starvla][INFO] - Step 3280, Loss: {'train/loss': 0.4338797628879547, 'train/grad_norm_pre_clip': 11.499807357788086, 'timing/data': 0.05035558436065912, 'timing/model': 1.2279050247743726, 'learning_rate/qwen_vl_interface': 5.832456831206418e-06, 'learning_rate/action_model': 5.8324568312064176e-05, 'epoch': 1.26}) +[2026-06-17 23:36:35,523][starVLA.training.train_starvla][INFO] - Step 3281, Loss: {'train/loss': 0.3085441291332245, 'train/grad_norm_pre_clip': 8.249151229858398, 'timing/data': 0.011987297795712948, 'timing/model': 1.1648276317864656, 'learning_rate/qwen_vl_interface': 5.830310401303377e-06, 'learning_rate/action_model': 5.830310401303377e-05, 'epoch': 1.26}) +[2026-06-17 23:36:40,579][starVLA.training.train_starvla][INFO] - Step 3282, Loss: {'train/loss': 0.3300861716270447, 'train/grad_norm_pre_clip': 13.432476043701172, 'timing/data': 0.036743685603141785, 'timing/model': 1.30761375464499, 'learning_rate/qwen_vl_interface': 5.8281638511013625e-06, 'learning_rate/action_model': 5.828163851101363e-05, 'epoch': 1.26}) +[2026-06-17 23:36:46,390][starVLA.training.train_starvla][INFO] - Step 3283, Loss: {'train/loss': 0.3070206642150879, 'train/grad_norm_pre_clip': 6.257045745849609, 'timing/data': 0.03716643713414669, 'timing/model': 1.2350992057472467, 'learning_rate/qwen_vl_interface': 5.8260171810453605e-06, 'learning_rate/action_model': 5.82601718104536e-05, 'epoch': 1.26}) +[2026-06-17 23:36:51,370][starVLA.training.train_starvla][INFO] - Step 3284, Loss: {'train/loss': 0.2461036890745163, 'train/grad_norm_pre_clip': 9.1013822555542, 'timing/data': 0.03955809120088816, 'timing/model': 1.1449528150260448, 'learning_rate/qwen_vl_interface': 5.823870391580376e-06, 'learning_rate/action_model': 5.823870391580376e-05, 'epoch': 1.26}) +[2026-06-17 23:36:56,861][starVLA.training.train_starvla][INFO] - Step 3285, Loss: {'train/loss': 0.5021032094955444, 'train/grad_norm_pre_clip': 11.077560424804688, 'timing/data': 0.0395838962867856, 'timing/model': 1.1447311155498028, 'learning_rate/qwen_vl_interface': 5.821723483151444e-06, 'learning_rate/action_model': 5.821723483151443e-05, 'epoch': 1.26}) +[2026-06-17 23:37:02,087][starVLA.training.train_starvla][INFO] - Step 3286, Loss: {'train/loss': 0.46594539284706116, 'train/grad_norm_pre_clip': 8.79519271850586, 'timing/data': 0.034003826789557934, 'timing/model': 1.1593327214941382, 'learning_rate/qwen_vl_interface': 5.819576456203617e-06, 'learning_rate/action_model': 5.8195764562036174e-05, 'epoch': 1.26}) +[2026-06-17 23:37:07,361][starVLA.training.train_starvla][INFO] - Step 3287, Loss: {'train/loss': 0.49914366006851196, 'train/grad_norm_pre_clip': 21.034299850463867, 'timing/data': 0.05323045514523983, 'timing/model': 1.9330489411950111, 'learning_rate/qwen_vl_interface': 5.817429311181977e-06, 'learning_rate/action_model': 5.8174293111819764e-05, 'epoch': 1.26}) +[2026-06-17 23:37:12,156][starVLA.training.train_starvla][INFO] - Step 3288, Loss: {'train/loss': 0.42319536209106445, 'train/grad_norm_pre_clip': 10.587198257446289, 'timing/data': 0.034804423339664936, 'timing/model': 1.111761773005128, 'learning_rate/qwen_vl_interface': 5.815282048531633e-06, 'learning_rate/action_model': 5.815282048531633e-05, 'epoch': 1.26}) +[2026-06-17 23:37:17,649][starVLA.training.train_starvla][INFO] - Step 3289, Loss: {'train/loss': 0.3719731569290161, 'train/grad_norm_pre_clip': 10.2125825881958, 'timing/data': 0.039900398813188076, 'timing/model': 1.190864165313542, 'learning_rate/qwen_vl_interface': 5.81313466869771e-06, 'learning_rate/action_model': 5.8131346686977096e-05, 'epoch': 1.27}) +[2026-06-17 23:37:23,104][starVLA.training.train_starvla][INFO] - Step 3290, Loss: {'train/loss': 0.19431540369987488, 'train/grad_norm_pre_clip': 6.587296962738037, 'timing/data': 0.04302142094820738, 'timing/model': 1.223141647875309, 'learning_rate/qwen_vl_interface': 5.810987172125364e-06, 'learning_rate/action_model': 5.810987172125364e-05, 'epoch': 1.27}) +[2026-06-17 23:37:28,742][starVLA.training.train_starvla][INFO] - Step 3291, Loss: {'train/loss': 0.5209845900535583, 'train/grad_norm_pre_clip': 13.60976505279541, 'timing/data': 0.03870637062937021, 'timing/model': 1.5744769191369414, 'learning_rate/qwen_vl_interface': 5.808839559259774e-06, 'learning_rate/action_model': 5.8088395592597746e-05, 'epoch': 1.27}) +[2026-06-17 23:37:34,097][starVLA.training.train_starvla][INFO] - Step 3292, Loss: {'train/loss': 0.24263396859169006, 'train/grad_norm_pre_clip': 8.219696044921875, 'timing/data': 0.03468640521168709, 'timing/model': 1.582405660301447, 'learning_rate/qwen_vl_interface': 5.806691830546143e-06, 'learning_rate/action_model': 5.8066918305461427e-05, 'epoch': 1.27}) +[2026-06-17 23:37:38,757][starVLA.training.train_starvla][INFO] - Step 3293, Loss: {'train/loss': 0.5951913595199585, 'train/grad_norm_pre_clip': 13.563140869140625, 'timing/data': 0.04815212544053793, 'timing/model': 1.1509563084691763, 'learning_rate/qwen_vl_interface': 5.804543986429696e-06, 'learning_rate/action_model': 5.804543986429696e-05, 'epoch': 1.27}) +[2026-06-17 23:37:43,896][starVLA.training.train_starvla][INFO] - Step 3294, Loss: {'train/loss': 0.27524131536483765, 'train/grad_norm_pre_clip': 19.896604537963867, 'timing/data': 0.03568036016076803, 'timing/model': 1.1738976147025824, 'learning_rate/qwen_vl_interface': 5.802396027355683e-06, 'learning_rate/action_model': 5.8023960273556835e-05, 'epoch': 1.27}) +[2026-06-17 23:37:54,711][starVLA.training.train_starvla][INFO] - Step 3295, Loss: {'train/loss': 0.7407960891723633, 'train/grad_norm_pre_clip': 15.07030963897705, 'timing/data': 0.08067616447806358, 'timing/model': 2.7793832197785378, 'learning_rate/qwen_vl_interface': 5.800247953769381e-06, 'learning_rate/action_model': 5.800247953769381e-05, 'epoch': 1.27}) +[2026-06-17 23:38:05,129][starVLA.training.train_starvla][INFO] - Step 3296, Loss: {'train/loss': 0.27124154567718506, 'train/grad_norm_pre_clip': 9.65027141571045, 'timing/data': 0.036711281165480614, 'timing/model': 1.2935320734977722, 'learning_rate/qwen_vl_interface': 5.798099766116084e-06, 'learning_rate/action_model': 5.798099766116084e-05, 'epoch': 1.27}) +[2026-06-17 23:38:11,933][starVLA.training.train_starvla][INFO] - Step 3297, Loss: {'train/loss': 0.28458371758461, 'train/grad_norm_pre_clip': 8.381744384765625, 'timing/data': 0.04327013995498419, 'timing/model': 2.1037342697381973, 'learning_rate/qwen_vl_interface': 5.7959514648411175e-06, 'learning_rate/action_model': 5.795951464841117e-05, 'epoch': 1.27}) +[2026-06-17 23:38:22,604][starVLA.training.train_starvla][INFO] - Step 3298, Loss: {'train/loss': 0.4237987995147705, 'train/grad_norm_pre_clip': 11.519564628601074, 'timing/data': 0.04591123852878809, 'timing/model': 5.051627926528454, 'learning_rate/qwen_vl_interface': 5.793803050389824e-06, 'learning_rate/action_model': 5.793803050389824e-05, 'epoch': 1.27}) +[2026-06-17 23:38:39,967][starVLA.training.train_starvla][INFO] - Step 3299, Loss: {'train/loss': 0.2880101203918457, 'train/grad_norm_pre_clip': 10.444318771362305, 'timing/data': 0.034781455993652344, 'timing/model': 4.902436891570687, 'learning_rate/qwen_vl_interface': 5.7916545232075725e-06, 'learning_rate/action_model': 5.791654523207573e-05, 'epoch': 1.27}) +[2026-06-17 23:43:36,350][starVLA.training.train_starvla][INFO] - Step 3300, Loss: {'train/loss': 0.5531821846961975, 'train/grad_norm_pre_clip': 15.079007148742676, 'eval/loss': 0.48158150911331177, 'eval/latency_4/loss': 0.48158150911331177, 'eval/demon_fire_precision': 0.9569453174104336, 'eval/demon_fire_recall': 0.9889574537187399, 'eval/demon_fire_f1': 0.9726880689985624, 'eval/demon_fire_rate_model': 0.9345080763582966, 'eval/demon_fire_rate_teacher': 0.9042584434654919, 'eval/demon_left_precision': 0.7640449438202247, 'eval/demon_left_recall': 0.6210045662100456, 'eval/demon_left_f1': 0.6851385390428211, 'eval/demon_left_rate_model': 0.10455212922173275, 'eval/demon_left_rate_teacher': 0.12863436123348018, 'eval/demon_right_precision': 0.8396624472573839, 'eval/demon_right_recall': 0.7184115523465704, 'eval/demon_right_f1': 0.7743190661478598, 'eval/demon_right_rate_model': 0.13920704845814977, 'eval/demon_right_rate_teacher': 0.16270190895741557, 'eval/demon_move_f1': 0.7297288025953405, 'eval/demon_cc_f1': 0.88435563798568, 'eval/latency_4/demon_fire_precision': 0.9569453174104336, 'eval/latency_4/demon_fire_recall': 0.9889574537187399, 'eval/latency_4/demon_fire_f1': 0.9726880689985624, 'eval/latency_4/demon_fire_rate_model': 0.9345080763582966, 'eval/latency_4/demon_fire_rate_teacher': 0.9042584434654919, 'eval/latency_4/demon_left_precision': 0.7640449438202247, 'eval/latency_4/demon_left_recall': 0.6210045662100456, 'eval/latency_4/demon_left_f1': 0.6851385390428211, 'eval/latency_4/demon_left_rate_model': 0.10455212922173275, 'eval/latency_4/demon_left_rate_teacher': 0.12863436123348018, 'eval/latency_4/demon_right_precision': 0.8396624472573839, 'eval/latency_4/demon_right_recall': 0.7184115523465704, 'eval/latency_4/demon_right_f1': 0.7743190661478598, 'eval/latency_4/demon_right_rate_model': 0.13920704845814977, 'eval/latency_4/demon_right_rate_teacher': 0.16270190895741557, 'eval/latency_4/demon_move_f1': 0.7297288025953405, 'eval/latency_4/demon_cc_f1': 0.88435563798568, 'timing/data': 0.040991234593093395, 'timing/model': 2.9015521006658673, 'learning_rate/qwen_vl_interface': 5.789505883739756e-06, 'learning_rate/action_model': 5.7895058837397554e-05, 'epoch': 1.27}) +[2026-06-17 23:43:36,368][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_3300_state +[2026-06-17 23:43:36,376][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 23:43:46,081][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_3300_state/model.safetensors +[2026-06-17 23:44:09,171][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_3300_state/optimizer.bin +[2026-06-17 23:44:09,173][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_3300_state/sampler.bin +[2026-06-17 23:44:09,174][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_3300_state/sampler_1.bin +[2026-06-17 23:44:09,177][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_3300_state/random_states_0.pkl +[2026-06-17 23:44:09,184][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3300 +[2026-06-17 23:44:09,184][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 23:44:09,208][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 23:44:14,629][starVLA.training.train_starvla][INFO] - Step 3301, Loss: {'train/loss': 0.4792852997779846, 'train/grad_norm_pre_clip': 10.373817443847656, 'timing/data': 0.038698991760611534, 'timing/model': 1.1652146000415087, 'learning_rate/qwen_vl_interface': 5.78735713243179e-06, 'learning_rate/action_model': 5.78735713243179e-05, 'epoch': 1.27}) +[2026-06-17 23:44:20,291][starVLA.training.train_starvla][INFO] - Step 3302, Loss: {'train/loss': 0.36550748348236084, 'train/grad_norm_pre_clip': 9.71015739440918, 'timing/data': 0.03641734644770622, 'timing/model': 1.754662954248488, 'learning_rate/qwen_vl_interface': 5.785208269729111e-06, 'learning_rate/action_model': 5.785208269729111e-05, 'epoch': 1.27}) +[2026-06-17 23:44:25,169][starVLA.training.train_starvla][INFO] - Step 3303, Loss: {'train/loss': 0.5429516434669495, 'train/grad_norm_pre_clip': 15.01491641998291, 'timing/data': 0.03545439336448908, 'timing/model': 1.1232778672128916, 'learning_rate/qwen_vl_interface': 5.783059296077183e-06, 'learning_rate/action_model': 5.7830592960771826e-05, 'epoch': 1.27}) +[2026-06-17 23:44:31,806][starVLA.training.train_starvla][INFO] - Step 3304, Loss: {'train/loss': 0.2506372332572937, 'train/grad_norm_pre_clip': 7.647939205169678, 'timing/data': 0.04592312127351761, 'timing/model': 1.1562491497024894, 'learning_rate/qwen_vl_interface': 5.780910211921488e-06, 'learning_rate/action_model': 5.7809102119214874e-05, 'epoch': 1.27}) +[2026-06-17 23:44:36,848][starVLA.training.train_starvla][INFO] - Step 3305, Loss: {'train/loss': 0.32805535197257996, 'train/grad_norm_pre_clip': 8.260287284851074, 'timing/data': 0.03564617782831192, 'timing/model': 1.1034215651452541, 'learning_rate/qwen_vl_interface': 5.778761017707536e-06, 'learning_rate/action_model': 5.778761017707536e-05, 'epoch': 1.27}) +[2026-06-17 23:44:42,580][starVLA.training.train_starvla][INFO] - Step 3306, Loss: {'train/loss': 0.3674675524234772, 'train/grad_norm_pre_clip': 10.174397468566895, 'timing/data': 0.030501460656523705, 'timing/model': 1.1471880581229925, 'learning_rate/qwen_vl_interface': 5.7766117138808545e-06, 'learning_rate/action_model': 5.776611713880854e-05, 'epoch': 1.27}) +[2026-06-17 23:44:47,815][starVLA.training.train_starvla][INFO] - Step 3307, Loss: {'train/loss': 0.5474932193756104, 'train/grad_norm_pre_clip': 15.153416633605957, 'timing/data': 0.03346994426101446, 'timing/model': 1.6106355413794518, 'learning_rate/qwen_vl_interface': 5.774462300886999e-06, 'learning_rate/action_model': 5.774462300886999e-05, 'epoch': 1.27}) +[2026-06-17 23:44:55,750][starVLA.training.train_starvla][INFO] - Step 3308, Loss: {'train/loss': 0.4027321934700012, 'train/grad_norm_pre_clip': 10.2326078414917, 'timing/data': 0.03125707432627678, 'timing/model': 1.4598052743822336, 'learning_rate/qwen_vl_interface': 5.772312779171545e-06, 'learning_rate/action_model': 5.7723127791715445e-05, 'epoch': 1.27}) +[2026-06-17 23:45:03,190][starVLA.training.train_starvla][INFO] - Step 3309, Loss: {'train/loss': 0.26373207569122314, 'train/grad_norm_pre_clip': 8.835847854614258, 'timing/data': 0.04336169641464949, 'timing/model': 1.7290306827053428, 'learning_rate/qwen_vl_interface': 5.770163149180089e-06, 'learning_rate/action_model': 5.770163149180089e-05, 'epoch': 1.27}) +[2026-06-17 23:45:12,551][starVLA.training.train_starvla][INFO] - Step 3310, Loss: {'train/loss': 0.331981897354126, 'train/grad_norm_pre_clip': 9.534078598022461, 'timing/data': 0.03066921327263117, 'timing/model': 1.2562204403802752, 'learning_rate/qwen_vl_interface': 5.768013411358252e-06, 'learning_rate/action_model': 5.768013411358252e-05, 'epoch': 1.27}) +[2026-06-17 23:45:20,985][starVLA.training.train_starvla][INFO] - Step 3311, Loss: {'train/loss': 0.33897411823272705, 'train/grad_norm_pre_clip': 10.236123085021973, 'timing/data': 0.03389561548829079, 'timing/model': 1.1869718935340643, 'learning_rate/qwen_vl_interface': 5.765863566151677e-06, 'learning_rate/action_model': 5.7658635661516767e-05, 'epoch': 1.27}) +[2026-06-17 23:45:29,505][starVLA.training.train_starvla][INFO] - Step 3312, Loss: {'train/loss': 0.3419964015483856, 'train/grad_norm_pre_clip': 8.491300582885742, 'timing/data': 0.04383051302284002, 'timing/model': 4.645355327054858, 'learning_rate/qwen_vl_interface': 5.763713614006033e-06, 'learning_rate/action_model': 5.763713614006032e-05, 'epoch': 1.27}) +[2026-06-17 23:45:34,670][starVLA.training.train_starvla][INFO] - Step 3313, Loss: {'train/loss': 0.4716261625289917, 'train/grad_norm_pre_clip': 18.28863525390625, 'timing/data': 0.028381009586155415, 'timing/model': 1.151844517327845, 'learning_rate/qwen_vl_interface': 5.761563555367e-06, 'learning_rate/action_model': 5.761563555367e-05, 'epoch': 1.27}) +[2026-06-17 23:45:40,415][starVLA.training.train_starvla][INFO] - Step 3314, Loss: {'train/loss': 0.45467132329940796, 'train/grad_norm_pre_clip': 12.812129974365234, 'timing/data': 0.0333934323862195, 'timing/model': 1.351593716070056, 'learning_rate/qwen_vl_interface': 5.759413390680293e-06, 'learning_rate/action_model': 5.759413390680293e-05, 'epoch': 1.27}) +[2026-06-17 23:45:45,530][starVLA.training.train_starvla][INFO] - Step 3315, Loss: {'train/loss': 0.17834486067295074, 'train/grad_norm_pre_clip': 6.155673503875732, 'timing/data': 0.031767542473971844, 'timing/model': 1.1683715684339404, 'learning_rate/qwen_vl_interface': 5.757263120391643e-06, 'learning_rate/action_model': 5.757263120391643e-05, 'epoch': 1.28}) +[2026-06-17 23:45:59,221][starVLA.training.train_starvla][INFO] - Step 3316, Loss: {'train/loss': 0.3793756365776062, 'train/grad_norm_pre_clip': 9.703431129455566, 'timing/data': 0.0693390779197216, 'timing/model': 1.095154550857842, 'learning_rate/qwen_vl_interface': 5.7551127449468005e-06, 'learning_rate/action_model': 5.7551127449468004e-05, 'epoch': 1.28}) +[2026-06-17 23:46:04,678][starVLA.training.train_starvla][INFO] - Step 3317, Loss: {'train/loss': 0.575064480304718, 'train/grad_norm_pre_clip': 14.244690895080566, 'timing/data': 0.001143142580986023, 'timing/model': 1.7756891511380672, 'learning_rate/qwen_vl_interface': 5.752962264791543e-06, 'learning_rate/action_model': 5.752962264791543e-05, 'epoch': 1.28}) +[2026-06-17 23:46:09,446][starVLA.training.train_starvla][INFO] - Step 3318, Loss: {'train/loss': 0.3608289957046509, 'train/grad_norm_pre_clip': 9.17162799835205, 'timing/data': 0.0008904524147510529, 'timing/model': 1.1502155493944883, 'learning_rate/qwen_vl_interface': 5.750811680371666e-06, 'learning_rate/action_model': 5.750811680371666e-05, 'epoch': 1.28}) +[2026-06-17 23:46:21,918][starVLA.training.train_starvla][INFO] - Step 3319, Loss: {'train/loss': 0.35830244421958923, 'train/grad_norm_pre_clip': 12.951624870300293, 'timing/data': 0.0012681130319833755, 'timing/model': 4.347351705655456, 'learning_rate/qwen_vl_interface': 5.7486609921329914e-06, 'learning_rate/action_model': 5.748660992132992e-05, 'epoch': 1.28}) +[2026-06-17 23:46:27,126][starVLA.training.train_starvla][INFO] - Step 3320, Loss: {'train/loss': 0.14044679701328278, 'train/grad_norm_pre_clip': 9.52442455291748, 'timing/data': 0.0023429393768310547, 'timing/model': 1.1633180119097233, 'learning_rate/qwen_vl_interface': 5.746510200521354e-06, 'learning_rate/action_model': 5.746510200521355e-05, 'epoch': 1.28}) +[2026-06-17 23:46:33,324][starVLA.training.train_starvla][INFO] - Step 3321, Loss: {'train/loss': 0.3551383316516876, 'train/grad_norm_pre_clip': 12.500853538513184, 'timing/data': 0.0018281172960996628, 'timing/model': 1.8242215132340789, 'learning_rate/qwen_vl_interface': 5.7443593059826206e-06, 'learning_rate/action_model': 5.744359305982621e-05, 'epoch': 1.28}) +[2026-06-17 23:46:39,249][starVLA.training.train_starvla][INFO] - Step 3322, Loss: {'train/loss': 0.5929965972900391, 'train/grad_norm_pre_clip': 10.18514633178711, 'timing/data': 0.0009708963334560394, 'timing/model': 1.9511602520942688, 'learning_rate/qwen_vl_interface': 5.742208308962669e-06, 'learning_rate/action_model': 5.742208308962669e-05, 'epoch': 1.28}) +[2026-06-17 23:46:47,572][starVLA.training.train_starvla][INFO] - Step 3323, Loss: {'train/loss': 0.24938863515853882, 'train/grad_norm_pre_clip': 38.10588073730469, 'timing/data': 0.0011483170092105865, 'timing/model': 1.1709870165213943, 'learning_rate/qwen_vl_interface': 5.740057209907405e-06, 'learning_rate/action_model': 5.7400572099074046e-05, 'epoch': 1.28}) +[2026-06-17 23:46:57,033][starVLA.training.train_starvla][INFO] - Step 3324, Loss: {'train/loss': 0.3843805193901062, 'train/grad_norm_pre_clip': 10.426548957824707, 'timing/data': 0.0013614557683467865, 'timing/model': 1.2306169448420405, 'learning_rate/qwen_vl_interface': 5.7379060092627545e-06, 'learning_rate/action_model': 5.7379060092627544e-05, 'epoch': 1.28}) +[2026-06-17 23:47:10,485][starVLA.training.train_starvla][INFO] - Step 3325, Loss: {'train/loss': 0.6653746366500854, 'train/grad_norm_pre_clip': 15.523073196411133, 'timing/data': 0.001060985028743744, 'timing/model': 5.8625499829649925, 'learning_rate/qwen_vl_interface': 5.735754707474663e-06, 'learning_rate/action_model': 5.735754707474663e-05, 'epoch': 1.28}) +[2026-06-17 23:47:15,511][starVLA.training.train_starvla][INFO] - Step 3326, Loss: {'train/loss': 0.44074609875679016, 'train/grad_norm_pre_clip': 11.388222694396973, 'timing/data': 0.000947943888604641, 'timing/model': 1.5740816676989198, 'learning_rate/qwen_vl_interface': 5.7336033049891e-06, 'learning_rate/action_model': 5.733603304989099e-05, 'epoch': 1.28}) +[2026-06-17 23:47:20,536][starVLA.training.train_starvla][INFO] - Step 3327, Loss: {'train/loss': 0.3382333219051361, 'train/grad_norm_pre_clip': 9.577665328979492, 'timing/data': 0.0012366659939289093, 'timing/model': 1.5834726626053452, 'learning_rate/qwen_vl_interface': 5.731451802252049e-06, 'learning_rate/action_model': 5.731451802252048e-05, 'epoch': 1.28}) +[2026-06-17 23:47:26,962][starVLA.training.train_starvla][INFO] - Step 3328, Loss: {'train/loss': 0.3721136450767517, 'train/grad_norm_pre_clip': 11.17126178741455, 'timing/data': 0.0010961433872580528, 'timing/model': 2.761936734430492, 'learning_rate/qwen_vl_interface': 5.729300199709522e-06, 'learning_rate/action_model': 5.729300199709522e-05, 'epoch': 1.28}) +[2026-06-17 23:47:32,369][starVLA.training.train_starvla][INFO] - Step 3329, Loss: {'train/loss': 0.5174391269683838, 'train/grad_norm_pre_clip': 14.125750541687012, 'timing/data': 0.0010082395747303963, 'timing/model': 1.4079524222761393, 'learning_rate/qwen_vl_interface': 5.7271484978075475e-06, 'learning_rate/action_model': 5.7271484978075475e-05, 'epoch': 1.28}) +[2026-06-17 23:47:37,110][starVLA.training.train_starvla][INFO] - Step 3330, Loss: {'train/loss': 0.33235281705856323, 'train/grad_norm_pre_clip': 9.626459121704102, 'timing/data': 0.0017494503408670425, 'timing/model': 1.2190773282200098, 'learning_rate/qwen_vl_interface': 5.724996696992176e-06, 'learning_rate/action_model': 5.7249966969921764e-05, 'epoch': 1.28}) +[2026-06-17 23:47:41,896][starVLA.training.train_starvla][INFO] - Step 3331, Loss: {'train/loss': 0.5417407751083374, 'train/grad_norm_pre_clip': 12.291823387145996, 'timing/data': 0.0009011067450046539, 'timing/model': 1.1311851926147938, 'learning_rate/qwen_vl_interface': 5.72284479770948e-06, 'learning_rate/action_model': 5.72284479770948e-05, 'epoch': 1.28}) +[2026-06-17 23:47:46,591][starVLA.training.train_starvla][INFO] - Step 3332, Loss: {'train/loss': 0.26655101776123047, 'train/grad_norm_pre_clip': 10.897921562194824, 'timing/data': 0.0013179229572415352, 'timing/model': 1.5083633782342076, 'learning_rate/qwen_vl_interface': 5.720692800405547e-06, 'learning_rate/action_model': 5.720692800405547e-05, 'epoch': 1.28}) +[2026-06-17 23:47:50,948][starVLA.training.train_starvla][INFO] - Step 3333, Loss: {'train/loss': 0.6080852150917053, 'train/grad_norm_pre_clip': 12.0936279296875, 'timing/data': 0.0011756978929042816, 'timing/model': 1.125962134450674, 'learning_rate/qwen_vl_interface': 5.71854070552649e-06, 'learning_rate/action_model': 5.7185407055264904e-05, 'epoch': 1.28}) +[2026-06-17 23:47:58,538][starVLA.training.train_starvla][INFO] - Step 3334, Loss: {'train/loss': 0.5199873447418213, 'train/grad_norm_pre_clip': 14.300253868103027, 'timing/data': 0.000975240021944046, 'timing/model': 1.2830367917194963, 'learning_rate/qwen_vl_interface': 5.71638851351844e-06, 'learning_rate/action_model': 5.71638851351844e-05, 'epoch': 1.28}) +[2026-06-17 23:48:07,286][starVLA.training.train_starvla][INFO] - Step 3335, Loss: {'train/loss': 0.506070077419281, 'train/grad_norm_pre_clip': 12.477553367614746, 'timing/data': 0.0019082287326455116, 'timing/model': 1.1269669029861689, 'learning_rate/qwen_vl_interface': 5.714236224827551e-06, 'learning_rate/action_model': 5.714236224827551e-05, 'epoch': 1.28}) +[2026-06-17 23:48:12,637][starVLA.training.train_starvla][INFO] - Step 3336, Loss: {'train/loss': 0.6079438924789429, 'train/grad_norm_pre_clip': 12.648760795593262, 'timing/data': 0.0012328941375017166, 'timing/model': 1.086801040917635, 'learning_rate/qwen_vl_interface': 5.712083839899991e-06, 'learning_rate/action_model': 5.712083839899991e-05, 'epoch': 1.28}) +[2026-06-17 23:48:17,493][starVLA.training.train_starvla][INFO] - Step 3337, Loss: {'train/loss': 0.3353491723537445, 'train/grad_norm_pre_clip': 12.230978012084961, 'timing/data': 0.0010928288102149963, 'timing/model': 1.3523009857162833, 'learning_rate/qwen_vl_interface': 5.709931359181956e-06, 'learning_rate/action_model': 5.709931359181956e-05, 'epoch': 1.28}) +[2026-06-17 23:48:22,272][starVLA.training.train_starvla][INFO] - Step 3338, Loss: {'train/loss': 0.45289990305900574, 'train/grad_norm_pre_clip': 9.623153686523438, 'timing/data': 0.0032377177849411964, 'timing/model': 1.1334871090948582, 'learning_rate/qwen_vl_interface': 5.707778783119654e-06, 'learning_rate/action_model': 5.707778783119654e-05, 'epoch': 1.28}) +[2026-06-17 23:48:27,756][starVLA.training.train_starvla][INFO] - Step 3339, Loss: {'train/loss': 0.28761935234069824, 'train/grad_norm_pre_clip': 8.444416999816895, 'timing/data': 0.0011765724048018456, 'timing/model': 1.1222435012459755, 'learning_rate/qwen_vl_interface': 5.705626112159318e-06, 'learning_rate/action_model': 5.7056261121593177e-05, 'epoch': 1.28}) +[2026-06-17 23:48:33,596][starVLA.training.train_starvla][INFO] - Step 3340, Loss: {'train/loss': 0.4696170389652252, 'train/grad_norm_pre_clip': 8.52803897857666, 'timing/data': 0.001153554767370224, 'timing/model': 1.1640029978007078, 'learning_rate/qwen_vl_interface': 5.703473346747199e-06, 'learning_rate/action_model': 5.703473346747199e-05, 'epoch': 1.28}) +[2026-06-17 23:48:40,319][starVLA.training.train_starvla][INFO] - Step 3341, Loss: {'train/loss': 0.42345181107521057, 'train/grad_norm_pre_clip': 11.659798622131348, 'timing/data': 0.0013487450778484344, 'timing/model': 1.4510081810876727, 'learning_rate/qwen_vl_interface': 5.701320487329567e-06, 'learning_rate/action_model': 5.701320487329567e-05, 'epoch': 1.29}) +[2026-06-17 23:48:46,713][starVLA.training.train_starvla][INFO] - Step 3342, Loss: {'train/loss': 0.3324994146823883, 'train/grad_norm_pre_clip': 8.265999794006348, 'timing/data': 0.0011855168268084526, 'timing/model': 2.889208768494427, 'learning_rate/qwen_vl_interface': 5.699167534352712e-06, 'learning_rate/action_model': 5.699167534352712e-05, 'epoch': 1.29}) +[2026-06-17 23:48:51,435][starVLA.training.train_starvla][INFO] - Step 3343, Loss: {'train/loss': 0.3217502236366272, 'train/grad_norm_pre_clip': 11.330716133117676, 'timing/data': 0.001366766169667244, 'timing/model': 1.1125067360699177, 'learning_rate/qwen_vl_interface': 5.697014488262945e-06, 'learning_rate/action_model': 5.697014488262945e-05, 'epoch': 1.29}) +[2026-06-17 23:48:57,008][starVLA.training.train_starvla][INFO] - Step 3344, Loss: {'train/loss': 0.6776723861694336, 'train/grad_norm_pre_clip': 13.551325798034668, 'timing/data': 0.0013381382450461388, 'timing/model': 1.2558683464303613, 'learning_rate/qwen_vl_interface': 5.694861349506591e-06, 'learning_rate/action_model': 5.6948613495065903e-05, 'epoch': 1.29}) +[2026-06-17 23:49:01,839][starVLA.training.train_starvla][INFO] - Step 3345, Loss: {'train/loss': 0.42175933718681335, 'train/grad_norm_pre_clip': 7.621565341949463, 'timing/data': 0.0022830795496702194, 'timing/model': 1.187681706622243, 'learning_rate/qwen_vl_interface': 5.692708118530003e-06, 'learning_rate/action_model': 5.6927081185300025e-05, 'epoch': 1.29}) +[2026-06-17 23:49:06,689][starVLA.training.train_starvla][INFO] - Step 3346, Loss: {'train/loss': 0.34642261266708374, 'train/grad_norm_pre_clip': 9.854263305664062, 'timing/data': 0.0014208117499947548, 'timing/model': 1.490762148052454, 'learning_rate/qwen_vl_interface': 5.6905547957795446e-06, 'learning_rate/action_model': 5.690554795779545e-05, 'epoch': 1.29}) +[2026-06-17 23:49:11,540][starVLA.training.train_starvla][INFO] - Step 3347, Loss: {'train/loss': 0.36714401841163635, 'train/grad_norm_pre_clip': 11.498291969299316, 'timing/data': 0.0011442918330430984, 'timing/model': 1.5446863854303956, 'learning_rate/qwen_vl_interface': 5.688401381701604e-06, 'learning_rate/action_model': 5.688401381701603e-05, 'epoch': 1.29}) +[2026-06-17 23:49:16,405][starVLA.training.train_starvla][INFO] - Step 3348, Loss: {'train/loss': 0.5361590385437012, 'train/grad_norm_pre_clip': 13.477380752563477, 'timing/data': 0.0012429235503077507, 'timing/model': 1.0939243249595165, 'learning_rate/qwen_vl_interface': 5.686247876742583e-06, 'learning_rate/action_model': 5.686247876742583e-05, 'epoch': 1.29}) +[2026-06-17 23:49:25,514][starVLA.training.train_starvla][INFO] - Step 3349, Loss: {'train/loss': 0.3829376995563507, 'train/grad_norm_pre_clip': 11.145965576171875, 'timing/data': 0.0010518785566091537, 'timing/model': 1.2043980471789837, 'learning_rate/qwen_vl_interface': 5.684094281348908e-06, 'learning_rate/action_model': 5.684094281348908e-05, 'epoch': 1.29}) +[2026-06-17 23:49:33,298][starVLA.training.train_starvla][INFO] - Step 3350, Loss: {'train/loss': 0.19742918014526367, 'train/grad_norm_pre_clip': 5.368828296661377, 'timing/data': 0.001250244677066803, 'timing/model': 1.1681032441556454, 'learning_rate/qwen_vl_interface': 5.681940595967023e-06, 'learning_rate/action_model': 5.681940595967024e-05, 'epoch': 1.29}) +[2026-06-17 23:49:45,375][starVLA.training.train_starvla][INFO] - Step 3351, Loss: {'train/loss': 0.4163651168346405, 'train/grad_norm_pre_clip': 10.783859252929688, 'timing/data': 0.0010181469842791557, 'timing/model': 1.3801741153001785, 'learning_rate/qwen_vl_interface': 5.679786821043388e-06, 'learning_rate/action_model': 5.679786821043388e-05, 'epoch': 1.29}) +[2026-06-17 23:49:50,546][starVLA.training.train_starvla][INFO] - Step 3352, Loss: {'train/loss': 0.23622842133045197, 'train/grad_norm_pre_clip': 6.764491558074951, 'timing/data': 0.004997706040740013, 'timing/model': 1.735982588492334, 'learning_rate/qwen_vl_interface': 5.677632957024482e-06, 'learning_rate/action_model': 5.6776329570244826e-05, 'epoch': 1.29}) +[2026-06-17 23:50:08,540][starVLA.training.train_starvla][INFO] - Step 3353, Loss: {'train/loss': 0.3739028871059418, 'train/grad_norm_pre_clip': 10.835107803344727, 'timing/data': 0.0011204630136489868, 'timing/model': 5.6537651140242815, 'learning_rate/qwen_vl_interface': 5.675479004356803e-06, 'learning_rate/action_model': 5.675479004356803e-05, 'epoch': 1.29}) +[2026-06-17 23:50:13,854][starVLA.training.train_starvla][INFO] - Step 3354, Loss: {'train/loss': 0.6009388566017151, 'train/grad_norm_pre_clip': 12.30478286743164, 'timing/data': 0.001467033289372921, 'timing/model': 1.104955704882741, 'learning_rate/qwen_vl_interface': 5.673324963486871e-06, 'learning_rate/action_model': 5.67332496348687e-05, 'epoch': 1.29}) +[2026-06-17 23:50:27,778][starVLA.training.train_starvla][INFO] - Step 3355, Loss: {'train/loss': 0.31632283329963684, 'train/grad_norm_pre_clip': 10.112761497497559, 'timing/data': 0.003938734531402588, 'timing/model': 1.639253655448556, 'learning_rate/qwen_vl_interface': 5.671170834861215e-06, 'learning_rate/action_model': 5.671170834861215e-05, 'epoch': 1.29}) +[2026-06-17 23:50:37,493][starVLA.training.train_starvla][INFO] - Step 3356, Loss: {'train/loss': 0.40200817584991455, 'train/grad_norm_pre_clip': 10.579254150390625, 'timing/data': 0.0014708992093801498, 'timing/model': 1.3305010944604874, 'learning_rate/qwen_vl_interface': 5.6690166189263925e-06, 'learning_rate/action_model': 5.669016618926393e-05, 'epoch': 1.29}) +[2026-06-17 23:50:42,613][starVLA.training.train_starvla][INFO] - Step 3357, Loss: {'train/loss': 0.46993979811668396, 'train/grad_norm_pre_clip': 9.799692153930664, 'timing/data': 0.001119963824748993, 'timing/model': 1.4856012146919966, 'learning_rate/qwen_vl_interface': 5.666862316128977e-06, 'learning_rate/action_model': 5.666862316128977e-05, 'epoch': 1.29}) +[2026-06-17 23:50:50,473][starVLA.training.train_starvla][INFO] - Step 3358, Loss: {'train/loss': 0.5334625244140625, 'train/grad_norm_pre_clip': 13.46942138671875, 'timing/data': 0.001445690169930458, 'timing/model': 1.156147437170148, 'learning_rate/qwen_vl_interface': 5.664707926915551e-06, 'learning_rate/action_model': 5.66470792691555e-05, 'epoch': 1.29}) +[2026-06-17 23:50:55,369][starVLA.training.train_starvla][INFO] - Step 3359, Loss: {'train/loss': 0.2984096109867096, 'train/grad_norm_pre_clip': 11.197102546691895, 'timing/data': 0.001363210380077362, 'timing/model': 1.115178070962429, 'learning_rate/qwen_vl_interface': 5.662553451732729e-06, 'learning_rate/action_model': 5.6625534517327285e-05, 'epoch': 1.29}) +[2026-06-17 23:51:00,968][starVLA.training.train_starvla][INFO] - Step 3360, Loss: {'train/loss': 0.527887761592865, 'train/grad_norm_pre_clip': 13.901323318481445, 'timing/data': 0.001224820502102375, 'timing/model': 1.3476274404674768, 'learning_rate/qwen_vl_interface': 5.660398891027129e-06, 'learning_rate/action_model': 5.660398891027129e-05, 'epoch': 1.29}) +[2026-06-17 23:51:05,970][starVLA.training.train_starvla][INFO] - Step 3361, Loss: {'train/loss': 0.33854150772094727, 'train/grad_norm_pre_clip': 11.807089805603027, 'timing/data': 0.0012723933905363083, 'timing/model': 1.2165459832176566, 'learning_rate/qwen_vl_interface': 5.658244245245399e-06, 'learning_rate/action_model': 5.658244245245399e-05, 'epoch': 1.29}) +[2026-06-17 23:51:11,459][starVLA.training.train_starvla][INFO] - Step 3362, Loss: {'train/loss': 0.4509021043777466, 'train/grad_norm_pre_clip': 13.90937328338623, 'timing/data': 0.0009203879162669182, 'timing/model': 1.9630804127082229, 'learning_rate/qwen_vl_interface': 5.6560895148341985e-06, 'learning_rate/action_model': 5.656089514834199e-05, 'epoch': 1.29}) +[2026-06-17 23:51:15,972][starVLA.training.train_starvla][INFO] - Step 3363, Loss: {'train/loss': 0.5226814150810242, 'train/grad_norm_pre_clip': 10.38568115234375, 'timing/data': 0.0011175638064742088, 'timing/model': 1.204947566613555, 'learning_rate/qwen_vl_interface': 5.653934700240204e-06, 'learning_rate/action_model': 5.6539347002402034e-05, 'epoch': 1.29}) +[2026-06-17 23:51:21,051][starVLA.training.train_starvla][INFO] - Step 3364, Loss: {'train/loss': 0.2845679521560669, 'train/grad_norm_pre_clip': 12.134194374084473, 'timing/data': 0.000781775452196598, 'timing/model': 1.1838663527742028, 'learning_rate/qwen_vl_interface': 5.651779801910111e-06, 'learning_rate/action_model': 5.651779801910111e-05, 'epoch': 1.29}) +[2026-06-17 23:51:33,279][starVLA.training.train_starvla][INFO] - Step 3365, Loss: {'train/loss': 0.3257595896720886, 'train/grad_norm_pre_clip': 9.340557098388672, 'timing/data': 0.001285250298678875, 'timing/model': 1.1193033447489142, 'learning_rate/qwen_vl_interface': 5.649624820290631e-06, 'learning_rate/action_model': 5.649624820290631e-05, 'epoch': 1.29}) +[2026-06-17 23:51:38,404][starVLA.training.train_starvla][INFO] - Step 3366, Loss: {'train/loss': 0.4204588234424591, 'train/grad_norm_pre_clip': 9.73975944519043, 'timing/data': 0.0018931617960333824, 'timing/model': 1.6472267741337419, 'learning_rate/qwen_vl_interface': 5.647469755828501e-06, 'learning_rate/action_model': 5.647469755828501e-05, 'epoch': 1.29}) +[2026-06-17 23:51:43,116][starVLA.training.train_starvla][INFO] - Step 3367, Loss: {'train/loss': 0.5925788879394531, 'train/grad_norm_pre_clip': 12.879590034484863, 'timing/data': 0.001129666343331337, 'timing/model': 1.3765749651938677, 'learning_rate/qwen_vl_interface': 5.645314608970457e-06, 'learning_rate/action_model': 5.6453146089704566e-05, 'epoch': 1.3}) +[2026-06-17 23:51:50,093][starVLA.training.train_starvla][INFO] - Step 3368, Loss: {'train/loss': 0.422027051448822, 'train/grad_norm_pre_clip': 16.142990112304688, 'timing/data': 0.0013178559020161629, 'timing/model': 3.75596356485039, 'learning_rate/qwen_vl_interface': 5.643159380163273e-06, 'learning_rate/action_model': 5.643159380163273e-05, 'epoch': 1.3}) +[2026-06-17 23:51:54,798][starVLA.training.train_starvla][INFO] - Step 3369, Loss: {'train/loss': 0.4379732310771942, 'train/grad_norm_pre_clip': 13.25813102722168, 'timing/data': 0.0011612828820943832, 'timing/model': 1.162381887435913, 'learning_rate/qwen_vl_interface': 5.641004069853725e-06, 'learning_rate/action_model': 5.641004069853725e-05, 'epoch': 1.3}) +[2026-06-17 23:52:00,560][starVLA.training.train_starvla][INFO] - Step 3370, Loss: {'train/loss': 0.3143778443336487, 'train/grad_norm_pre_clip': 10.18006706237793, 'timing/data': 0.0009559234604239464, 'timing/model': 1.254335694015026, 'learning_rate/qwen_vl_interface': 5.638848678488613e-06, 'learning_rate/action_model': 5.638848678488613e-05, 'epoch': 1.3}) +[2026-06-17 23:52:05,423][starVLA.training.train_starvla][INFO] - Step 3371, Loss: {'train/loss': 0.33477362990379333, 'train/grad_norm_pre_clip': 11.461088180541992, 'timing/data': 0.001102609559893608, 'timing/model': 1.0902675790712237, 'learning_rate/qwen_vl_interface': 5.636693206514752e-06, 'learning_rate/action_model': 5.636693206514752e-05, 'epoch': 1.3}) +[2026-06-17 23:52:10,382][starVLA.training.train_starvla][INFO] - Step 3372, Loss: {'train/loss': 0.39405059814453125, 'train/grad_norm_pre_clip': 13.673309326171875, 'timing/data': 0.001232369802892208, 'timing/model': 1.607418855652213, 'learning_rate/qwen_vl_interface': 5.634537654378973e-06, 'learning_rate/action_model': 5.634537654378973e-05, 'epoch': 1.3}) +[2026-06-17 23:52:15,404][starVLA.training.train_starvla][INFO] - Step 3373, Loss: {'train/loss': 0.34553471207618713, 'train/grad_norm_pre_clip': 12.998126983642578, 'timing/data': 0.0013281134888529778, 'timing/model': 1.500610656104982, 'learning_rate/qwen_vl_interface': 5.632382022528123e-06, 'learning_rate/action_model': 5.632382022528123e-05, 'epoch': 1.3}) +[2026-06-17 23:52:20,292][starVLA.training.train_starvla][INFO] - Step 3374, Loss: {'train/loss': 0.2318098396062851, 'train/grad_norm_pre_clip': 8.688774108886719, 'timing/data': 0.0011571617797017097, 'timing/model': 1.1425181087106466, 'learning_rate/qwen_vl_interface': 5.6302263114090705e-06, 'learning_rate/action_model': 5.63022631140907e-05, 'epoch': 1.3}) +[2026-06-17 23:52:25,896][starVLA.training.train_starvla][INFO] - Step 3375, Loss: {'train/loss': 0.3531646132469177, 'train/grad_norm_pre_clip': 11.97843074798584, 'timing/data': 0.001137477345764637, 'timing/model': 1.2871349165216088, 'learning_rate/qwen_vl_interface': 5.628070521468693e-06, 'learning_rate/action_model': 5.628070521468692e-05, 'epoch': 1.3}) +[2026-06-17 23:52:30,958][starVLA.training.train_starvla][INFO] - Step 3376, Loss: {'train/loss': 0.42034468054771423, 'train/grad_norm_pre_clip': 16.27263832092285, 'timing/data': 0.00116741843521595, 'timing/model': 1.1982500711455941, 'learning_rate/qwen_vl_interface': 5.625914653153889e-06, 'learning_rate/action_model': 5.625914653153889e-05, 'epoch': 1.3}) +[2026-06-17 23:52:36,136][starVLA.training.train_starvla][INFO] - Step 3377, Loss: {'train/loss': 0.35950967669487, 'train/grad_norm_pre_clip': 13.765239715576172, 'timing/data': 0.0011284137144684792, 'timing/model': 1.7498378809541464, 'learning_rate/qwen_vl_interface': 5.623758706911574e-06, 'learning_rate/action_model': 5.6237587069115737e-05, 'epoch': 1.3}) +[2026-06-17 23:52:40,624][starVLA.training.train_starvla][INFO] - Step 3378, Loss: {'train/loss': 0.33739882707595825, 'train/grad_norm_pre_clip': 12.761645317077637, 'timing/data': 0.000954492948949337, 'timing/model': 1.1073871236294508, 'learning_rate/qwen_vl_interface': 5.621602683188677e-06, 'learning_rate/action_model': 5.621602683188677e-05, 'epoch': 1.3}) +[2026-06-17 23:52:45,882][starVLA.training.train_starvla][INFO] - Step 3379, Loss: {'train/loss': 0.3524973392486572, 'train/grad_norm_pre_clip': 10.951355934143066, 'timing/data': 0.0010416386649012566, 'timing/model': 1.2892290260642767, 'learning_rate/qwen_vl_interface': 5.619446582432143e-06, 'learning_rate/action_model': 5.619446582432143e-05, 'epoch': 1.3}) +[2026-06-17 23:52:50,679][starVLA.training.train_starvla][INFO] - Step 3380, Loss: {'train/loss': 0.4525245726108551, 'train/grad_norm_pre_clip': 16.270601272583008, 'timing/data': 0.0012051016092300415, 'timing/model': 1.2431531194597483, 'learning_rate/qwen_vl_interface': 5.617290405088936e-06, 'learning_rate/action_model': 5.617290405088936e-05, 'epoch': 1.3}) +[2026-06-17 23:52:55,575][starVLA.training.train_starvla][INFO] - Step 3381, Loss: {'train/loss': 0.44817787408828735, 'train/grad_norm_pre_clip': 16.087757110595703, 'timing/data': 0.0011319313198328018, 'timing/model': 1.2036856524646282, 'learning_rate/qwen_vl_interface': 5.615134151606033e-06, 'learning_rate/action_model': 5.615134151606033e-05, 'epoch': 1.3}) +[2026-06-17 23:53:00,512][starVLA.training.train_starvla][INFO] - Step 3382, Loss: {'train/loss': 0.3442223072052002, 'train/grad_norm_pre_clip': 11.262813568115234, 'timing/data': 0.0010822322219610214, 'timing/model': 1.4440773958340287, 'learning_rate/qwen_vl_interface': 5.612977822430425e-06, 'learning_rate/action_model': 5.612977822430425e-05, 'epoch': 1.3}) +[2026-06-17 23:53:05,149][starVLA.training.train_starvla][INFO] - Step 3383, Loss: {'train/loss': 0.42481517791748047, 'train/grad_norm_pre_clip': 10.723649024963379, 'timing/data': 0.00221241544932127, 'timing/model': 1.1126533029600978, 'learning_rate/qwen_vl_interface': 5.610821418009127e-06, 'learning_rate/action_model': 5.610821418009127e-05, 'epoch': 1.3}) +[2026-06-17 23:53:10,454][starVLA.training.train_starvla][INFO] - Step 3384, Loss: {'train/loss': 0.1833430677652359, 'train/grad_norm_pre_clip': 8.526987075805664, 'timing/data': 0.0017573237419128418, 'timing/model': 1.2037404021248221, 'learning_rate/qwen_vl_interface': 5.608664938789159e-06, 'learning_rate/action_model': 5.608664938789159e-05, 'epoch': 1.3}) +[2026-06-17 23:53:17,973][starVLA.training.train_starvla][INFO] - Step 3385, Loss: {'train/loss': 0.6102935075759888, 'train/grad_norm_pre_clip': 12.974614143371582, 'timing/data': 0.001367012970149517, 'timing/model': 2.509241985157132, 'learning_rate/qwen_vl_interface': 5.606508385217564e-06, 'learning_rate/action_model': 5.6065083852175636e-05, 'epoch': 1.3}) +[2026-06-17 23:53:28,433][starVLA.training.train_starvla][INFO] - Step 3386, Loss: {'train/loss': 0.4219391942024231, 'train/grad_norm_pre_clip': 17.135669708251953, 'timing/data': 0.0024273665621876717, 'timing/model': 1.604982104152441, 'learning_rate/qwen_vl_interface': 5.604351757741396e-06, 'learning_rate/action_model': 5.6043517577413964e-05, 'epoch': 1.3}) +[2026-06-17 23:53:36,052][starVLA.training.train_starvla][INFO] - Step 3387, Loss: {'train/loss': 0.3927272856235504, 'train/grad_norm_pre_clip': 14.2253999710083, 'timing/data': 0.001405579037964344, 'timing/model': 2.1553644873201847, 'learning_rate/qwen_vl_interface': 5.602195056807727e-06, 'learning_rate/action_model': 5.602195056807727e-05, 'epoch': 1.3}) +[2026-06-17 23:53:47,051][starVLA.training.train_starvla][INFO] - Step 3388, Loss: {'train/loss': 0.43243199586868286, 'train/grad_norm_pre_clip': 13.326239585876465, 'timing/data': 0.0012805694714188576, 'timing/model': 3.1847129678353667, 'learning_rate/qwen_vl_interface': 5.600038282863647e-06, 'learning_rate/action_model': 5.6000382828636466e-05, 'epoch': 1.3}) +[2026-06-17 23:54:03,842][starVLA.training.train_starvla][INFO] - Step 3389, Loss: {'train/loss': 0.2681569755077362, 'train/grad_norm_pre_clip': 11.304149627685547, 'timing/data': 0.0008377283811569214, 'timing/model': 4.733764436095953, 'learning_rate/qwen_vl_interface': 5.597881436356252e-06, 'learning_rate/action_model': 5.597881436356252e-05, 'epoch': 1.3}) +[2026-06-17 23:54:19,414][starVLA.training.train_starvla][INFO] - Step 3390, Loss: {'train/loss': 0.581335723400116, 'train/grad_norm_pre_clip': 18.206960678100586, 'timing/data': 0.0013687973842024803, 'timing/model': 5.9204164408147335, 'learning_rate/qwen_vl_interface': 5.595724517732661e-06, 'learning_rate/action_model': 5.595724517732661e-05, 'epoch': 1.3}) +[2026-06-17 23:54:36,034][starVLA.training.train_starvla][INFO] - Step 3391, Loss: {'train/loss': 0.5216994881629944, 'train/grad_norm_pre_clip': 11.38957691192627, 'timing/data': 0.0012133391574025154, 'timing/model': 2.6896190270781517, 'learning_rate/qwen_vl_interface': 5.5935675274400046e-06, 'learning_rate/action_model': 5.593567527440005e-05, 'epoch': 1.3}) +[2026-06-17 23:54:46,463][starVLA.training.train_starvla][INFO] - Step 3392, Loss: {'train/loss': 0.5604453086853027, 'train/grad_norm_pre_clip': 14.8538818359375, 'timing/data': 0.007081855088472366, 'timing/model': 1.7511076126247644, 'learning_rate/qwen_vl_interface': 5.591410465925431e-06, 'learning_rate/action_model': 5.5914104659254306e-05, 'epoch': 1.3}) +[2026-06-17 23:54:59,272][starVLA.training.train_starvla][INFO] - Step 3393, Loss: {'train/loss': 0.5043542981147766, 'train/grad_norm_pre_clip': 12.131221771240234, 'timing/data': 0.0011499067768454552, 'timing/model': 1.3090441413223743, 'learning_rate/qwen_vl_interface': 5.589253333636101e-06, 'learning_rate/action_model': 5.589253333636101e-05, 'epoch': 1.31}) +[2026-06-17 23:55:06,669][starVLA.training.train_starvla][INFO] - Step 3394, Loss: {'train/loss': 0.2998354136943817, 'train/grad_norm_pre_clip': 10.93752384185791, 'timing/data': 0.0014737574383616447, 'timing/model': 1.0984093602746725, 'learning_rate/qwen_vl_interface': 5.587096131019189e-06, 'learning_rate/action_model': 5.5870961310191896e-05, 'epoch': 1.31}) +[2026-06-17 23:55:11,321][starVLA.training.train_starvla][INFO] - Step 3395, Loss: {'train/loss': 0.3040491044521332, 'train/grad_norm_pre_clip': 8.445408821105957, 'timing/data': 0.0009339805692434311, 'timing/model': 1.0724180974066257, 'learning_rate/qwen_vl_interface': 5.5849388585218865e-06, 'learning_rate/action_model': 5.5849388585218866e-05, 'epoch': 1.31}) +[2026-06-17 23:55:15,976][starVLA.training.train_starvla][INFO] - Step 3396, Loss: {'train/loss': 0.3541526794433594, 'train/grad_norm_pre_clip': 8.369377136230469, 'timing/data': 0.001347462646663189, 'timing/model': 1.1765990080311894, 'learning_rate/qwen_vl_interface': 5.582781516591398e-06, 'learning_rate/action_model': 5.5827815165913977e-05, 'epoch': 1.31}) +[2026-06-17 23:55:20,857][starVLA.training.train_starvla][INFO] - Step 3397, Loss: {'train/loss': 0.13746947050094604, 'train/grad_norm_pre_clip': 7.3747992515563965, 'timing/data': 0.001249990426003933, 'timing/model': 1.575893571600318, 'learning_rate/qwen_vl_interface': 5.580624105674945e-06, 'learning_rate/action_model': 5.580624105674944e-05, 'epoch': 1.31}) +[2026-06-17 23:55:25,593][starVLA.training.train_starvla][INFO] - Step 3398, Loss: {'train/loss': 0.4062701165676117, 'train/grad_norm_pre_clip': 9.272128105163574, 'timing/data': 0.001115565188229084, 'timing/model': 1.3132485328242183, 'learning_rate/qwen_vl_interface': 5.578466626219756e-06, 'learning_rate/action_model': 5.578466626219756e-05, 'epoch': 1.31}) +[2026-06-17 23:55:30,711][starVLA.training.train_starvla][INFO] - Step 3399, Loss: {'train/loss': 0.4740046262741089, 'train/grad_norm_pre_clip': 9.716229438781738, 'timing/data': 0.0011645033955574036, 'timing/model': 1.2293214434757829, 'learning_rate/qwen_vl_interface': 5.576309078673083e-06, 'learning_rate/action_model': 5.576309078673083e-05, 'epoch': 1.31}) +[2026-06-17 23:59:49,969][starVLA.training.train_starvla][INFO] - Step 3400, Loss: {'train/loss': 0.40361934900283813, 'train/grad_norm_pre_clip': 13.807571411132812, 'eval/loss': 0.4926919639110565, 'eval/latency_4/loss': 0.4926919639110565, 'eval/demon_fire_precision': 0.9705406280349628, 'eval/demon_fire_recall': 0.9736927573887626, 'eval/demon_fire_f1': 0.9721141374837873, 'eval/demon_fire_rate_model': 0.9071953010279001, 'eval/demon_fire_rate_teacher': 0.9042584434654919, 'eval/demon_left_precision': 0.7731343283582089, 'eval/demon_left_recall': 0.591324200913242, 'eval/demon_left_f1': 0.6701164294954721, 'eval/demon_left_rate_model': 0.09838472834067548, 'eval/demon_left_rate_teacher': 0.12863436123348018, 'eval/demon_right_precision': 0.812133072407045, 'eval/demon_right_recall': 0.7490974729241877, 'eval/demon_right_f1': 0.7793427230046949, 'eval/demon_right_rate_model': 0.15007342143906022, 'eval/demon_right_rate_teacher': 0.16270190895741557, 'eval/demon_move_f1': 0.7247295762500835, 'eval/demon_cc_f1': 0.8818127853105796, 'eval/latency_4/demon_fire_precision': 0.9705406280349628, 'eval/latency_4/demon_fire_recall': 0.9736927573887626, 'eval/latency_4/demon_fire_f1': 0.9721141374837873, 'eval/latency_4/demon_fire_rate_model': 0.9071953010279001, 'eval/latency_4/demon_fire_rate_teacher': 0.9042584434654919, 'eval/latency_4/demon_left_precision': 0.7731343283582089, 'eval/latency_4/demon_left_recall': 0.591324200913242, 'eval/latency_4/demon_left_f1': 0.6701164294954721, 'eval/latency_4/demon_left_rate_model': 0.09838472834067548, 'eval/latency_4/demon_left_rate_teacher': 0.12863436123348018, 'eval/latency_4/demon_right_precision': 0.812133072407045, 'eval/latency_4/demon_right_recall': 0.7490974729241877, 'eval/latency_4/demon_right_f1': 0.7793427230046949, 'eval/latency_4/demon_right_rate_model': 0.15007342143906022, 'eval/latency_4/demon_right_rate_teacher': 0.16270190895741557, 'eval/latency_4/demon_move_f1': 0.7247295762500835, 'eval/latency_4/demon_cc_f1': 0.8818127853105796, 'timing/data': 0.0013145888224244118, 'timing/model': 1.5505159050226212, 'learning_rate/qwen_vl_interface': 5.5741514634821885e-06, 'learning_rate/action_model': 5.574151463482188e-05, 'epoch': 1.31}) +[2026-06-17 23:59:49,989][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_3400_state +[2026-06-17 23:59:49,998][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 23:59:59,109][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_3400_state/model.safetensors +[2026-06-18 00:00:22,843][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_3400_state/optimizer.bin +[2026-06-18 00:00:22,845][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_3400_state/sampler.bin +[2026-06-18 00:00:22,846][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_3400_state/sampler_1.bin +[2026-06-18 00:00:22,848][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_3400_state/random_states_0.pkl +[2026-06-18 00:00:22,853][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3400 +[2026-06-18 00:00:22,853][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 00:00:22,871][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 00:00:27,993][starVLA.training.train_starvla][INFO] - Step 3401, Loss: {'train/loss': 0.6478400230407715, 'train/grad_norm_pre_clip': 16.656774520874023, 'timing/data': 0.001036306843161583, 'timing/model': 1.6815975392237306, 'learning_rate/qwen_vl_interface': 5.571993781094345e-06, 'learning_rate/action_model': 5.5719937810943444e-05, 'epoch': 1.31}) +[2026-06-18 00:00:33,378][starVLA.training.train_starvla][INFO] - Step 3402, Loss: {'train/loss': 0.2613425552845001, 'train/grad_norm_pre_clip': 8.343023300170898, 'timing/data': 0.0011173393577337265, 'timing/model': 1.5095838773995638, 'learning_rate/qwen_vl_interface': 5.569836031956845e-06, 'learning_rate/action_model': 5.5698360319568454e-05, 'epoch': 1.31}) +[2026-06-18 00:00:50,970][starVLA.training.train_starvla][INFO] - Step 3403, Loss: {'train/loss': 0.4856358766555786, 'train/grad_norm_pre_clip': 13.26181697845459, 'timing/data': 0.0009643789380788803, 'timing/model': 5.661502318456769, 'learning_rate/qwen_vl_interface': 5.5676782165169915e-06, 'learning_rate/action_model': 5.567678216516991e-05, 'epoch': 1.31}) +[2026-06-18 00:00:56,558][starVLA.training.train_starvla][INFO] - Step 3404, Loss: {'train/loss': 0.45352545380592346, 'train/grad_norm_pre_clip': 11.24292278289795, 'timing/data': 0.001484476961195469, 'timing/model': 1.1416786769405007, 'learning_rate/qwen_vl_interface': 5.565520335222101e-06, 'learning_rate/action_model': 5.565520335222101e-05, 'epoch': 1.31}) +[2026-06-18 00:01:03,110][starVLA.training.train_starvla][INFO] - Step 3405, Loss: {'train/loss': 0.2677794396877289, 'train/grad_norm_pre_clip': 8.297101020812988, 'timing/data': 0.0011689923703670502, 'timing/model': 1.2357968157157302, 'learning_rate/qwen_vl_interface': 5.563362388519507e-06, 'learning_rate/action_model': 5.5633623885195066e-05, 'epoch': 1.31}) +[2026-06-18 00:01:08,176][starVLA.training.train_starvla][INFO] - Step 3406, Loss: {'train/loss': 0.7444006204605103, 'train/grad_norm_pre_clip': 15.22296142578125, 'timing/data': 0.0013895649462938309, 'timing/model': 1.300386456772685, 'learning_rate/qwen_vl_interface': 5.56120437685655e-06, 'learning_rate/action_model': 5.56120437685655e-05, 'epoch': 1.31}) +[2026-06-18 00:01:13,656][starVLA.training.train_starvla][INFO] - Step 3407, Loss: {'train/loss': 0.41807040572166443, 'train/grad_norm_pre_clip': 7.27999210357666, 'timing/data': 0.0016157831996679306, 'timing/model': 1.700938549824059, 'learning_rate/qwen_vl_interface': 5.559046300680593e-06, 'learning_rate/action_model': 5.5590463006805925e-05, 'epoch': 1.31}) +[2026-06-18 00:01:25,898][starVLA.training.train_starvla][INFO] - Step 3408, Loss: {'train/loss': 0.3903948962688446, 'train/grad_norm_pre_clip': 8.366264343261719, 'timing/data': 0.001125359907746315, 'timing/model': 2.705673759803176, 'learning_rate/qwen_vl_interface': 5.556888160439006e-06, 'learning_rate/action_model': 5.556888160439006e-05, 'epoch': 1.31}) +[2026-06-18 00:01:30,694][starVLA.training.train_starvla][INFO] - Step 3409, Loss: {'train/loss': 0.5988783836364746, 'train/grad_norm_pre_clip': 12.161486625671387, 'timing/data': 0.002243584953248501, 'timing/model': 1.098252622410655, 'learning_rate/qwen_vl_interface': 5.5547299565791725e-06, 'learning_rate/action_model': 5.5547299565791723e-05, 'epoch': 1.31}) +[2026-06-18 00:01:35,563][starVLA.training.train_starvla][INFO] - Step 3410, Loss: {'train/loss': 0.66783607006073, 'train/grad_norm_pre_clip': 12.148937225341797, 'timing/data': 0.0010248897597193718, 'timing/model': 1.1062802048400044, 'learning_rate/qwen_vl_interface': 5.55257168954849e-06, 'learning_rate/action_model': 5.5525716895484905e-05, 'epoch': 1.31}) +[2026-06-18 00:01:43,797][starVLA.training.train_starvla][INFO] - Step 3411, Loss: {'train/loss': 0.3445521891117096, 'train/grad_norm_pre_clip': 7.417237281799316, 'timing/data': 0.0010997578501701355, 'timing/model': 4.4224936394020915, 'learning_rate/qwen_vl_interface': 5.550413359794371e-06, 'learning_rate/action_model': 5.550413359794371e-05, 'epoch': 1.31}) +[2026-06-18 00:01:49,607][starVLA.training.train_starvla][INFO] - Step 3412, Loss: {'train/loss': 0.4212292432785034, 'train/grad_norm_pre_clip': 13.8446044921875, 'timing/data': 0.0016905982047319412, 'timing/model': 1.9019550988450646, 'learning_rate/qwen_vl_interface': 5.548254967764239e-06, 'learning_rate/action_model': 5.5482549677642394e-05, 'epoch': 1.31}) +[2026-06-18 00:01:55,635][starVLA.training.train_starvla][INFO] - Step 3413, Loss: {'train/loss': 0.1994355320930481, 'train/grad_norm_pre_clip': 7.058966636657715, 'timing/data': 0.0013421261683106422, 'timing/model': 1.087576786056161, 'learning_rate/qwen_vl_interface': 5.546096513905532e-06, 'learning_rate/action_model': 5.546096513905532e-05, 'epoch': 1.31}) +[2026-06-18 00:02:03,981][starVLA.training.train_starvla][INFO] - Step 3414, Loss: {'train/loss': 0.4730844497680664, 'train/grad_norm_pre_clip': 12.830429077148438, 'timing/data': 0.0018653897568583488, 'timing/model': 1.4898789636790752, 'learning_rate/qwen_vl_interface': 5.5439379986657e-06, 'learning_rate/action_model': 5.543937998665699e-05, 'epoch': 1.31}) +[2026-06-18 00:02:09,162][starVLA.training.train_starvla][INFO] - Step 3415, Loss: {'train/loss': 0.22773265838623047, 'train/grad_norm_pre_clip': 6.074087619781494, 'timing/data': 0.0008081430569291115, 'timing/model': 1.4717481723055243, 'learning_rate/qwen_vl_interface': 5.5417794224922025e-06, 'learning_rate/action_model': 5.5417794224922023e-05, 'epoch': 1.31}) +[2026-06-18 00:02:16,182][starVLA.training.train_starvla][INFO] - Step 3416, Loss: {'train/loss': 0.41302457451820374, 'train/grad_norm_pre_clip': 11.564607620239258, 'timing/data': 0.0007985644042491913, 'timing/model': 1.5924050714820623, 'learning_rate/qwen_vl_interface': 5.539620785832517e-06, 'learning_rate/action_model': 5.5396207858325166e-05, 'epoch': 1.31}) +[2026-06-18 00:02:23,011][starVLA.training.train_starvla][INFO] - Step 3417, Loss: {'train/loss': 0.3281934857368469, 'train/grad_norm_pre_clip': 11.8372163772583, 'timing/data': 0.0017641251906752586, 'timing/model': 1.5977308554574847, 'learning_rate/qwen_vl_interface': 5.537462089134133e-06, 'learning_rate/action_model': 5.537462089134132e-05, 'epoch': 1.31}) +[2026-06-18 00:02:32,019][starVLA.training.train_starvla][INFO] - Step 3418, Loss: {'train/loss': 0.6102103590965271, 'train/grad_norm_pre_clip': 11.404868125915527, 'timing/data': 0.001923849806189537, 'timing/model': 1.09435741789639, 'learning_rate/qwen_vl_interface': 5.535303332844546e-06, 'learning_rate/action_model': 5.5353033328445455e-05, 'epoch': 1.31}) +[2026-06-18 00:02:36,752][starVLA.training.train_starvla][INFO] - Step 3419, Loss: {'train/loss': 0.5299341082572937, 'train/grad_norm_pre_clip': 11.081085205078125, 'timing/data': 0.001102725975215435, 'timing/model': 1.1220025420188904, 'learning_rate/qwen_vl_interface': 5.533144517411273e-06, 'learning_rate/action_model': 5.533144517411273e-05, 'epoch': 1.32}) +[2026-06-18 00:02:41,550][starVLA.training.train_starvla][INFO] - Step 3420, Loss: {'train/loss': 0.5322189331054688, 'train/grad_norm_pre_clip': 12.416176795959473, 'timing/data': 0.000943467952311039, 'timing/model': 1.1110866041854024, 'learning_rate/qwen_vl_interface': 5.530985643281834e-06, 'learning_rate/action_model': 5.530985643281834e-05, 'epoch': 1.32}) +[2026-06-18 00:02:46,488][starVLA.training.train_starvla][INFO] - Step 3421, Loss: {'train/loss': 0.40587177872657776, 'train/grad_norm_pre_clip': 9.741606712341309, 'timing/data': 0.001082034781575203, 'timing/model': 1.0762682380154729, 'learning_rate/qwen_vl_interface': 5.528826710903774e-06, 'learning_rate/action_model': 5.528826710903774e-05, 'epoch': 1.32}) +[2026-06-18 00:02:57,332][starVLA.training.train_starvla][INFO] - Step 3422, Loss: {'train/loss': 0.4264482855796814, 'train/grad_norm_pre_clip': 8.743606567382812, 'timing/data': 0.0011796299368143082, 'timing/model': 3.2250663358718157, 'learning_rate/qwen_vl_interface': 5.526667720724632e-06, 'learning_rate/action_model': 5.526667720724632e-05, 'epoch': 1.32}) +[2026-06-18 00:03:02,957][starVLA.training.train_starvla][INFO] - Step 3423, Loss: {'train/loss': 0.21015767753124237, 'train/grad_norm_pre_clip': 6.057072639465332, 'timing/data': 0.001227731816470623, 'timing/model': 1.1246903762221336, 'learning_rate/qwen_vl_interface': 5.5245086731919764e-06, 'learning_rate/action_model': 5.524508673191976e-05, 'epoch': 1.32}) +[2026-06-18 00:03:09,185][starVLA.training.train_starvla][INFO] - Step 3424, Loss: {'train/loss': 0.4490474760532379, 'train/grad_norm_pre_clip': 8.068747520446777, 'timing/data': 0.001163526438176632, 'timing/model': 1.1334293922409415, 'learning_rate/qwen_vl_interface': 5.522349568753376e-06, 'learning_rate/action_model': 5.522349568753377e-05, 'epoch': 1.32}) +[2026-06-18 00:03:14,156][starVLA.training.train_starvla][INFO] - Step 3425, Loss: {'train/loss': 0.7274320125579834, 'train/grad_norm_pre_clip': 13.4390869140625, 'timing/data': 0.0013981536030769348, 'timing/model': 1.1048178737983108, 'learning_rate/qwen_vl_interface': 5.520190407856419e-06, 'learning_rate/action_model': 5.5201904078564185e-05, 'epoch': 1.32}) +[2026-06-18 00:03:19,124][starVLA.training.train_starvla][INFO] - Step 3426, Loss: {'train/loss': 0.43225106596946716, 'train/grad_norm_pre_clip': 8.92087459564209, 'timing/data': 0.0017595253884792328, 'timing/model': 1.5707557247951627, 'learning_rate/qwen_vl_interface': 5.5180311909486974e-06, 'learning_rate/action_model': 5.518031190948697e-05, 'epoch': 1.32}) +[2026-06-18 00:03:25,404][starVLA.training.train_starvla][INFO] - Step 3427, Loss: {'train/loss': 0.440578430891037, 'train/grad_norm_pre_clip': 9.734232902526855, 'timing/data': 0.001132492907345295, 'timing/model': 1.4532019905745983, 'learning_rate/qwen_vl_interface': 5.515871918477821e-06, 'learning_rate/action_model': 5.515871918477821e-05, 'epoch': 1.32}) +[2026-06-18 00:03:29,757][starVLA.training.train_starvla][INFO] - Step 3428, Loss: {'train/loss': 0.362510085105896, 'train/grad_norm_pre_clip': 8.616741180419922, 'timing/data': 0.0013283202424645424, 'timing/model': 1.1239825328812003, 'learning_rate/qwen_vl_interface': 5.513712590891414e-06, 'learning_rate/action_model': 5.513712590891413e-05, 'epoch': 1.32}) +[2026-06-18 00:03:36,279][starVLA.training.train_starvla][INFO] - Step 3429, Loss: {'train/loss': 0.5359625816345215, 'train/grad_norm_pre_clip': 13.591405868530273, 'timing/data': 0.0011629695072770119, 'timing/model': 1.1625101836398244, 'learning_rate/qwen_vl_interface': 5.511553208637098e-06, 'learning_rate/action_model': 5.511553208637098e-05, 'epoch': 1.32}) +[2026-06-18 00:03:40,950][starVLA.training.train_starvla][INFO] - Step 3430, Loss: {'train/loss': 0.29364103078842163, 'train/grad_norm_pre_clip': 10.835580825805664, 'timing/data': 0.0010977517813444138, 'timing/model': 1.116085254587233, 'learning_rate/qwen_vl_interface': 5.509393772162524e-06, 'learning_rate/action_model': 5.509393772162523e-05, 'epoch': 1.32}) +[2026-06-18 00:03:47,147][starVLA.training.train_starvla][INFO] - Step 3431, Loss: {'train/loss': 0.35631734132766724, 'train/grad_norm_pre_clip': 11.968067169189453, 'timing/data': 0.002350534312427044, 'timing/model': 2.2834531106054783, 'learning_rate/qwen_vl_interface': 5.50723428191534e-06, 'learning_rate/action_model': 5.50723428191534e-05, 'epoch': 1.32}) +[2026-06-18 00:03:55,137][starVLA.training.train_starvla][INFO] - Step 3432, Loss: {'train/loss': 0.48950403928756714, 'train/grad_norm_pre_clip': 9.907913208007812, 'timing/data': 0.0011615436524152756, 'timing/model': 1.6460532993078232, 'learning_rate/qwen_vl_interface': 5.505074738343212e-06, 'learning_rate/action_model': 5.505074738343212e-05, 'epoch': 1.32}) +[2026-06-18 00:03:59,960][starVLA.training.train_starvla][INFO] - Step 3433, Loss: {'train/loss': 0.341238796710968, 'train/grad_norm_pre_clip': 9.08470630645752, 'timing/data': 0.0009114015847444534, 'timing/model': 1.1048108283430338, 'learning_rate/qwen_vl_interface': 5.502915141893819e-06, 'learning_rate/action_model': 5.5029151418938184e-05, 'epoch': 1.32}) +[2026-06-18 00:04:05,227][starVLA.training.train_starvla][INFO] - Step 3434, Loss: {'train/loss': 0.2852744162082672, 'train/grad_norm_pre_clip': 7.627241611480713, 'timing/data': 0.000924769788980484, 'timing/model': 1.0953418780118227, 'learning_rate/qwen_vl_interface': 5.5007554930148436e-06, 'learning_rate/action_model': 5.5007554930148434e-05, 'epoch': 1.32}) +[2026-06-18 00:04:13,553][starVLA.training.train_starvla][INFO] - Step 3435, Loss: {'train/loss': 0.49438872933387756, 'train/grad_norm_pre_clip': 12.369144439697266, 'timing/data': 0.0010018059983849525, 'timing/model': 1.2563586020842195, 'learning_rate/qwen_vl_interface': 5.498595792153985e-06, 'learning_rate/action_model': 5.498595792153984e-05, 'epoch': 1.32}) +[2026-06-18 00:04:19,106][starVLA.training.train_starvla][INFO] - Step 3436, Loss: {'train/loss': 0.4409116208553314, 'train/grad_norm_pre_clip': 10.262140274047852, 'timing/data': 0.0009637987241148949, 'timing/model': 1.5950683886185288, 'learning_rate/qwen_vl_interface': 5.49643603975895e-06, 'learning_rate/action_model': 5.4964360397589495e-05, 'epoch': 1.32}) +[2026-06-18 00:04:23,996][starVLA.training.train_starvla][INFO] - Step 3437, Loss: {'train/loss': 0.34714770317077637, 'train/grad_norm_pre_clip': 10.922056198120117, 'timing/data': 0.0011466993018984795, 'timing/model': 1.6181231029331684, 'learning_rate/qwen_vl_interface': 5.494276236277458e-06, 'learning_rate/action_model': 5.4942762362774584e-05, 'epoch': 1.32}) +[2026-06-18 00:04:34,849][starVLA.training.train_starvla][INFO] - Step 3438, Loss: {'train/loss': 0.24818314611911774, 'train/grad_norm_pre_clip': 7.960408687591553, 'timing/data': 0.0008242232725024223, 'timing/model': 1.4876044932752848, 'learning_rate/qwen_vl_interface': 5.492116382157242e-06, 'learning_rate/action_model': 5.492116382157242e-05, 'epoch': 1.32}) +[2026-06-18 00:04:40,041][starVLA.training.train_starvla][INFO] - Step 3439, Loss: {'train/loss': 0.5603602528572083, 'train/grad_norm_pre_clip': 13.719531059265137, 'timing/data': 0.0013724956661462784, 'timing/model': 1.3231574958190322, 'learning_rate/qwen_vl_interface': 5.489956477846038e-06, 'learning_rate/action_model': 5.489956477846038e-05, 'epoch': 1.32}) +[2026-06-18 00:04:52,431][starVLA.training.train_starvla][INFO] - Step 3440, Loss: {'train/loss': 0.25810936093330383, 'train/grad_norm_pre_clip': 32.503318786621094, 'timing/data': 0.0010774480178952217, 'timing/model': 1.1432623770087957, 'learning_rate/qwen_vl_interface': 5.4877965237915984e-06, 'learning_rate/action_model': 5.4877965237915986e-05, 'epoch': 1.32}) +[2026-06-18 00:04:57,536][starVLA.training.train_starvla][INFO] - Step 3441, Loss: {'train/loss': 0.45099779963493347, 'train/grad_norm_pre_clip': 12.357604026794434, 'timing/data': 0.0015224460512399673, 'timing/model': 1.226989520713687, 'learning_rate/qwen_vl_interface': 5.485636520441684e-06, 'learning_rate/action_model': 5.4856365204416836e-05, 'epoch': 1.32}) +[2026-06-18 00:05:02,713][starVLA.training.train_starvla][INFO] - Step 3442, Loss: {'train/loss': 0.2709392309188843, 'train/grad_norm_pre_clip': 7.8608245849609375, 'timing/data': 0.0012255050241947174, 'timing/model': 1.9079674649983644, 'learning_rate/qwen_vl_interface': 5.483476468244063e-06, 'learning_rate/action_model': 5.483476468244063e-05, 'epoch': 1.32}) +[2026-06-18 00:05:07,531][starVLA.training.train_starvla][INFO] - Step 3443, Loss: {'train/loss': 0.33290567994117737, 'train/grad_norm_pre_clip': 7.7886528968811035, 'timing/data': 0.0009967982769012451, 'timing/model': 1.2657878510653973, 'learning_rate/qwen_vl_interface': 5.481316367646522e-06, 'learning_rate/action_model': 5.4813163676465216e-05, 'epoch': 1.32}) +[2026-06-18 00:05:12,399][starVLA.training.train_starvla][INFO] - Step 3444, Loss: {'train/loss': 0.32293978333473206, 'train/grad_norm_pre_clip': 9.080364227294922, 'timing/data': 0.001394941471517086, 'timing/model': 1.2126123970374465, 'learning_rate/qwen_vl_interface': 5.479156219096847e-06, 'learning_rate/action_model': 5.479156219096847e-05, 'epoch': 1.32}) +[2026-06-18 00:05:17,179][starVLA.training.train_starvla][INFO] - Step 3445, Loss: {'train/loss': 0.6375037431716919, 'train/grad_norm_pre_clip': 10.775773048400879, 'timing/data': 0.0010787677019834518, 'timing/model': 1.1162202507257462, 'learning_rate/qwen_vl_interface': 5.476996023042843e-06, 'learning_rate/action_model': 5.476996023042843e-05, 'epoch': 1.33}) +[2026-06-18 00:05:21,876][starVLA.training.train_starvla][INFO] - Step 3446, Loss: {'train/loss': 0.6271088123321533, 'train/grad_norm_pre_clip': 17.737356185913086, 'timing/data': 0.0011636940762400627, 'timing/model': 1.4633001619949937, 'learning_rate/qwen_vl_interface': 5.474835779932317e-06, 'learning_rate/action_model': 5.474835779932317e-05, 'epoch': 1.33}) +[2026-06-18 00:05:27,609][starVLA.training.train_starvla][INFO] - Step 3447, Loss: {'train/loss': 0.28876835107803345, 'train/grad_norm_pre_clip': 7.7846269607543945, 'timing/data': 0.0008371416479349136, 'timing/model': 2.440507118590176, 'learning_rate/qwen_vl_interface': 5.472675490213093e-06, 'learning_rate/action_model': 5.472675490213093e-05, 'epoch': 1.33}) +[2026-06-18 00:05:33,534][starVLA.training.train_starvla][INFO] - Step 3448, Loss: {'train/loss': 0.2900887429714203, 'train/grad_norm_pre_clip': 7.636016368865967, 'timing/data': 0.001134691759943962, 'timing/model': 2.7522005885839462, 'learning_rate/qwen_vl_interface': 5.470515154333e-06, 'learning_rate/action_model': 5.470515154333e-05, 'epoch': 1.33}) +[2026-06-18 00:05:39,358][starVLA.training.train_starvla][INFO] - Step 3449, Loss: {'train/loss': 0.5124931335449219, 'train/grad_norm_pre_clip': 17.226472854614258, 'timing/data': 0.0011231312528252602, 'timing/model': 1.2571836085990071, 'learning_rate/qwen_vl_interface': 5.468354772739879e-06, 'learning_rate/action_model': 5.468354772739879e-05, 'epoch': 1.33}) +[2026-06-18 00:05:49,156][starVLA.training.train_starvla][INFO] - Step 3450, Loss: {'train/loss': 0.298160195350647, 'train/grad_norm_pre_clip': 9.334590911865234, 'timing/data': 0.00228879414498806, 'timing/model': 5.857808758504689, 'learning_rate/qwen_vl_interface': 5.466194345881578e-06, 'learning_rate/action_model': 5.4661943458815776e-05, 'epoch': 1.33}) +[2026-06-18 00:05:57,742][starVLA.training.train_starvla][INFO] - Step 3451, Loss: {'train/loss': 0.4171339273452759, 'train/grad_norm_pre_clip': 9.72175121307373, 'timing/data': 0.0016320720314979553, 'timing/model': 1.149388324469328, 'learning_rate/qwen_vl_interface': 5.4640338742059564e-06, 'learning_rate/action_model': 5.464033874205956e-05, 'epoch': 1.33}) +[2026-06-18 00:06:07,605][starVLA.training.train_starvla][INFO] - Step 3452, Loss: {'train/loss': 0.774504542350769, 'train/grad_norm_pre_clip': 19.779266357421875, 'timing/data': 0.0008854232728481293, 'timing/model': 1.6990938344970345, 'learning_rate/qwen_vl_interface': 5.461873358160883e-06, 'learning_rate/action_model': 5.461873358160883e-05, 'epoch': 1.33}) +[2026-06-18 00:06:17,927][starVLA.training.train_starvla][INFO] - Step 3453, Loss: {'train/loss': 0.375383198261261, 'train/grad_norm_pre_clip': 7.7699408531188965, 'timing/data': 0.001627323217689991, 'timing/model': 1.1990784984081984, 'learning_rate/qwen_vl_interface': 5.459712798194234e-06, 'learning_rate/action_model': 5.459712798194233e-05, 'epoch': 1.33}) +[2026-06-18 00:06:25,617][starVLA.training.train_starvla][INFO] - Step 3454, Loss: {'train/loss': 0.46592286229133606, 'train/grad_norm_pre_clip': 13.179244041442871, 'timing/data': 0.0009902315214276314, 'timing/model': 1.2993510896340013, 'learning_rate/qwen_vl_interface': 5.457552194753896e-06, 'learning_rate/action_model': 5.457552194753895e-05, 'epoch': 1.33}) +[2026-06-18 00:06:34,547][starVLA.training.train_starvla][INFO] - Step 3455, Loss: {'train/loss': 0.34019309282302856, 'train/grad_norm_pre_clip': 8.871047973632812, 'timing/data': 0.001272459514439106, 'timing/model': 2.8875061264261603, 'learning_rate/qwen_vl_interface': 5.455391548287764e-06, 'learning_rate/action_model': 5.455391548287764e-05, 'epoch': 1.33}) +[2026-06-18 00:06:47,256][starVLA.training.train_starvla][INFO] - Step 3456, Loss: {'train/loss': 0.4933096766471863, 'train/grad_norm_pre_clip': 9.612194061279297, 'timing/data': 0.0008218018338084221, 'timing/model': 1.57710893265903, 'learning_rate/qwen_vl_interface': 5.453230859243745e-06, 'learning_rate/action_model': 5.4532308592437444e-05, 'epoch': 1.33}) +[2026-06-18 00:07:02,938][starVLA.training.train_starvla][INFO] - Step 3457, Loss: {'train/loss': 0.3102186322212219, 'train/grad_norm_pre_clip': 11.534987449645996, 'timing/data': 0.0012877881526947021, 'timing/model': 5.410014446824789, 'learning_rate/qwen_vl_interface': 5.451070128069748e-06, 'learning_rate/action_model': 5.4510701280697485e-05, 'epoch': 1.33}) +[2026-06-18 00:07:13,755][starVLA.training.train_starvla][INFO] - Step 3458, Loss: {'train/loss': 0.4861672818660736, 'train/grad_norm_pre_clip': 9.828080177307129, 'timing/data': 0.001329602673649788, 'timing/model': 1.1397464154288173, 'learning_rate/qwen_vl_interface': 5.4489093552136986e-06, 'learning_rate/action_model': 5.4489093552136984e-05, 'epoch': 1.33}) +[2026-06-18 00:07:24,260][starVLA.training.train_starvla][INFO] - Step 3459, Loss: {'train/loss': 0.29715999960899353, 'train/grad_norm_pre_clip': 9.591377258300781, 'timing/data': 0.002030489034950733, 'timing/model': 5.530926430597901, 'learning_rate/qwen_vl_interface': 5.446748541123526e-06, 'learning_rate/action_model': 5.4467485411235254e-05, 'epoch': 1.33}) +[2026-06-18 00:07:29,298][starVLA.training.train_starvla][INFO] - Step 3460, Loss: {'train/loss': 0.37478935718536377, 'train/grad_norm_pre_clip': 8.268729209899902, 'timing/data': 0.001184307038784027, 'timing/model': 1.124196058139205, 'learning_rate/qwen_vl_interface': 5.444587686247168e-06, 'learning_rate/action_model': 5.4445876862471675e-05, 'epoch': 1.33}) +[2026-06-18 00:07:34,074][starVLA.training.train_starvla][INFO] - Step 3461, Loss: {'train/loss': 0.5012182593345642, 'train/grad_norm_pre_clip': 9.494949340820312, 'timing/data': 0.0011975690722465515, 'timing/model': 1.119735373184085, 'learning_rate/qwen_vl_interface': 5.442426791032574e-06, 'learning_rate/action_model': 5.4424267910325746e-05, 'epoch': 1.33}) +[2026-06-18 00:07:38,708][starVLA.training.train_starvla][INFO] - Step 3462, Loss: {'train/loss': 0.6658077239990234, 'train/grad_norm_pre_clip': 13.630295753479004, 'timing/data': 0.0012409202754497528, 'timing/model': 1.4160073418170214, 'learning_rate/qwen_vl_interface': 5.4402658559277e-06, 'learning_rate/action_model': 5.4402658559276995e-05, 'epoch': 1.33}) +[2026-06-18 00:07:43,508][starVLA.training.train_starvla][INFO] - Step 3463, Loss: {'train/loss': 0.4007384181022644, 'train/grad_norm_pre_clip': 9.276780128479004, 'timing/data': 0.0011421870440244675, 'timing/model': 1.2040474554523826, 'learning_rate/qwen_vl_interface': 5.438104881380507e-06, 'learning_rate/action_model': 5.438104881380507e-05, 'epoch': 1.33}) +[2026-06-18 00:07:48,983][starVLA.training.train_starvla][INFO] - Step 3464, Loss: {'train/loss': 0.5064815878868103, 'train/grad_norm_pre_clip': 9.029419898986816, 'timing/data': 0.0017203204333782196, 'timing/model': 1.5159210953861475, 'learning_rate/qwen_vl_interface': 5.435943867838971e-06, 'learning_rate/action_model': 5.435943867838971e-05, 'epoch': 1.33}) +[2026-06-18 00:07:54,613][starVLA.training.train_starvla][INFO] - Step 3465, Loss: {'train/loss': 0.6098150610923767, 'train/grad_norm_pre_clip': 14.697808265686035, 'timing/data': 0.0015961527824401855, 'timing/model': 1.170048700645566, 'learning_rate/qwen_vl_interface': 5.433782815751072e-06, 'learning_rate/action_model': 5.433782815751072e-05, 'epoch': 1.33}) +[2026-06-18 00:08:00,157][starVLA.training.train_starvla][INFO] - Step 3466, Loss: {'train/loss': 0.21382126212120056, 'train/grad_norm_pre_clip': 13.793656349182129, 'timing/data': 0.0023573674261569977, 'timing/model': 1.6315261516720057, 'learning_rate/qwen_vl_interface': 5.431621725564797e-06, 'learning_rate/action_model': 5.4316217255647974e-05, 'epoch': 1.33}) +[2026-06-18 00:08:05,787][starVLA.training.train_starvla][INFO] - Step 3467, Loss: {'train/loss': 0.2874275743961334, 'train/grad_norm_pre_clip': 9.90102767944336, 'timing/data': 0.001177850179374218, 'timing/model': 2.2276697400957346, 'learning_rate/qwen_vl_interface': 5.429460597728142e-06, 'learning_rate/action_model': 5.429460597728142e-05, 'epoch': 1.33}) +[2026-06-18 00:08:11,746][starVLA.training.train_starvla][INFO] - Step 3468, Loss: {'train/loss': 0.29831835627555847, 'train/grad_norm_pre_clip': 10.523542404174805, 'timing/data': 0.0013766270130872726, 'timing/model': 1.2172940941527486, 'learning_rate/qwen_vl_interface': 5.4272994326891114e-06, 'learning_rate/action_model': 5.427299432689111e-05, 'epoch': 1.33}) +[2026-06-18 00:08:16,792][starVLA.training.train_starvla][INFO] - Step 3469, Loss: {'train/loss': 0.4545706808567047, 'train/grad_norm_pre_clip': 8.695893287658691, 'timing/data': 0.0011835414916276932, 'timing/model': 1.378591033630073, 'learning_rate/qwen_vl_interface': 5.425138230895719e-06, 'learning_rate/action_model': 5.425138230895719e-05, 'epoch': 1.33}) +[2026-06-18 00:08:21,958][starVLA.training.train_starvla][INFO] - Step 3470, Loss: {'train/loss': 0.3078419268131256, 'train/grad_norm_pre_clip': 8.010082244873047, 'timing/data': 0.0011028870940208435, 'timing/model': 1.1756670884788036, 'learning_rate/qwen_vl_interface': 5.422976992795984e-06, 'learning_rate/action_model': 5.422976992795984e-05, 'epoch': 1.33}) +[2026-06-18 00:08:27,425][starVLA.training.train_starvla][INFO] - Step 3471, Loss: {'train/loss': 0.47160136699676514, 'train/grad_norm_pre_clip': 15.68602466583252, 'timing/data': 0.0013437718152999878, 'timing/model': 1.1869599027559161, 'learning_rate/qwen_vl_interface': 5.420815718837929e-06, 'learning_rate/action_model': 5.42081571883793e-05, 'epoch': 1.34}) +[2026-06-18 00:08:32,787][starVLA.training.train_starvla][INFO] - Step 3472, Loss: {'train/loss': 0.37254413962364197, 'train/grad_norm_pre_clip': 10.164837837219238, 'timing/data': 0.0011026561260223389, 'timing/model': 1.7917076777666807, 'learning_rate/qwen_vl_interface': 5.418654409469595e-06, 'learning_rate/action_model': 5.4186544094695946e-05, 'epoch': 1.34}) +[2026-06-18 00:08:38,168][starVLA.training.train_starvla][INFO] - Step 3473, Loss: {'train/loss': 0.3200976848602295, 'train/grad_norm_pre_clip': 7.305387020111084, 'timing/data': 0.013094333931803703, 'timing/model': 1.4056278644129634, 'learning_rate/qwen_vl_interface': 5.416493065139019e-06, 'learning_rate/action_model': 5.4164930651390186e-05, 'epoch': 1.34}) +[2026-06-18 00:08:48,533][starVLA.training.train_starvla][INFO] - Step 3474, Loss: {'train/loss': 0.2956957519054413, 'train/grad_norm_pre_clip': 7.739630222320557, 'timing/data': 0.013390084728598595, 'timing/model': 1.4360711025074124, 'learning_rate/qwen_vl_interface': 5.4143316862942516e-06, 'learning_rate/action_model': 5.4143316862942514e-05, 'epoch': 1.34}) +[2026-06-18 00:09:01,955][starVLA.training.train_starvla][INFO] - Step 3475, Loss: {'train/loss': 0.38850080966949463, 'train/grad_norm_pre_clip': 11.729820251464844, 'timing/data': 0.016084841452538967, 'timing/model': 4.205554216168821, 'learning_rate/qwen_vl_interface': 5.412170273383348e-06, 'learning_rate/action_model': 5.412170273383348e-05, 'epoch': 1.34}) +[2026-06-18 00:09:11,464][starVLA.training.train_starvla][INFO] - Step 3476, Loss: {'train/loss': 0.34979766607284546, 'train/grad_norm_pre_clip': 10.814112663269043, 'timing/data': 0.01365919690579176, 'timing/model': 2.469324979931116, 'learning_rate/qwen_vl_interface': 5.410008826854374e-06, 'learning_rate/action_model': 5.410008826854373e-05, 'epoch': 1.34}) +[2026-06-18 00:09:24,460][starVLA.training.train_starvla][INFO] - Step 3477, Loss: {'train/loss': 0.6721547245979309, 'train/grad_norm_pre_clip': 13.648517608642578, 'timing/data': 0.014269854873418808, 'timing/model': 1.738260017707944, 'learning_rate/qwen_vl_interface': 5.407847347155395e-06, 'learning_rate/action_model': 5.407847347155395e-05, 'epoch': 1.34}) +[2026-06-18 00:09:45,391][starVLA.training.train_starvla][INFO] - Step 3478, Loss: {'train/loss': 0.30563026666641235, 'train/grad_norm_pre_clip': 7.799647808074951, 'timing/data': 0.013699622824788094, 'timing/model': 4.89768413733691, 'learning_rate/qwen_vl_interface': 5.4056858347344934e-06, 'learning_rate/action_model': 5.405685834734493e-05, 'epoch': 1.34}) +[2026-06-18 00:10:02,534][starVLA.training.train_starvla][INFO] - Step 3479, Loss: {'train/loss': 0.3249759078025818, 'train/grad_norm_pre_clip': 11.70279598236084, 'timing/data': 0.026372998021543026, 'timing/model': 5.748170369304717, 'learning_rate/qwen_vl_interface': 5.403524290039751e-06, 'learning_rate/action_model': 5.403524290039751e-05, 'epoch': 1.34}) +[2026-06-18 00:10:17,949][starVLA.training.train_starvla][INFO] - Step 3480, Loss: {'train/loss': 0.34638121724128723, 'train/grad_norm_pre_clip': 16.219099044799805, 'timing/data': 0.01716470904648304, 'timing/model': 3.009187097661197, 'learning_rate/qwen_vl_interface': 5.401362713519257e-06, 'learning_rate/action_model': 5.4013627135192565e-05, 'epoch': 1.34}) +[2026-06-18 00:10:29,107][starVLA.training.train_starvla][INFO] - Step 3481, Loss: {'train/loss': 0.5605323910713196, 'train/grad_norm_pre_clip': 13.014901161193848, 'timing/data': 0.014350327663123608, 'timing/model': 3.9546401454135776, 'learning_rate/qwen_vl_interface': 5.399201105621111e-06, 'learning_rate/action_model': 5.39920110562111e-05, 'epoch': 1.34}) +[2026-06-18 00:10:38,292][starVLA.training.train_starvla][INFO] - Step 3482, Loss: {'train/loss': 0.527377724647522, 'train/grad_norm_pre_clip': 10.143022537231445, 'timing/data': 0.014398780651390553, 'timing/model': 3.2554254047572613, 'learning_rate/qwen_vl_interface': 5.397039466793412e-06, 'learning_rate/action_model': 5.3970394667934123e-05, 'epoch': 1.34}) +[2026-06-18 00:10:55,108][starVLA.training.train_starvla][INFO] - Step 3483, Loss: {'train/loss': 0.3455527126789093, 'train/grad_norm_pre_clip': 12.428816795349121, 'timing/data': 0.01750895194709301, 'timing/model': 1.2700876127928495, 'learning_rate/qwen_vl_interface': 5.394877797484278e-06, 'learning_rate/action_model': 5.3948777974842777e-05, 'epoch': 1.34}) +[2026-06-18 00:11:03,750][starVLA.training.train_starvla][INFO] - Step 3484, Loss: {'train/loss': 0.2745673656463623, 'train/grad_norm_pre_clip': 9.390069007873535, 'timing/data': 0.015626981854438782, 'timing/model': 1.7659135749563575, 'learning_rate/qwen_vl_interface': 5.392716098141817e-06, 'learning_rate/action_model': 5.3927160981418165e-05, 'epoch': 1.34}) +[2026-06-18 00:11:08,737][starVLA.training.train_starvla][INFO] - Step 3485, Loss: {'train/loss': 0.5297228693962097, 'train/grad_norm_pre_clip': 13.89964485168457, 'timing/data': 0.01377489697188139, 'timing/model': 1.1069006463512778, 'learning_rate/qwen_vl_interface': 5.390554369214157e-06, 'learning_rate/action_model': 5.390554369214157e-05, 'epoch': 1.34}) +[2026-06-18 00:11:13,796][starVLA.training.train_starvla][INFO] - Step 3486, Loss: {'train/loss': 0.3634521961212158, 'train/grad_norm_pre_clip': 14.221569061279297, 'timing/data': 0.011718551628291607, 'timing/model': 1.3171855369582772, 'learning_rate/qwen_vl_interface': 5.388392611149425e-06, 'learning_rate/action_model': 5.388392611149424e-05, 'epoch': 1.34}) +[2026-06-18 00:11:19,391][starVLA.training.train_starvla][INFO] - Step 3487, Loss: {'train/loss': 0.43899011611938477, 'train/grad_norm_pre_clip': 11.958285331726074, 'timing/data': 0.012851923704147339, 'timing/model': 1.596125633455813, 'learning_rate/qwen_vl_interface': 5.386230824395754e-06, 'learning_rate/action_model': 5.386230824395754e-05, 'epoch': 1.34}) +[2026-06-18 00:11:24,444][starVLA.training.train_starvla][INFO] - Step 3488, Loss: {'train/loss': 0.3231506943702698, 'train/grad_norm_pre_clip': 9.275151252746582, 'timing/data': 0.014507063664495945, 'timing/model': 1.3271715743467212, 'learning_rate/qwen_vl_interface': 5.384069009401288e-06, 'learning_rate/action_model': 5.384069009401288e-05, 'epoch': 1.34}) +[2026-06-18 00:11:31,158][starVLA.training.train_starvla][INFO] - Step 3489, Loss: {'train/loss': 0.4879734516143799, 'train/grad_norm_pre_clip': 13.99755859375, 'timing/data': 0.022738704457879066, 'timing/model': 1.4400635743513703, 'learning_rate/qwen_vl_interface': 5.381907166614171e-06, 'learning_rate/action_model': 5.3819071666141704e-05, 'epoch': 1.34}) +[2026-06-18 00:11:36,924][starVLA.training.train_starvla][INFO] - Step 3490, Loss: {'train/loss': 0.4557051360607147, 'train/grad_norm_pre_clip': 27.9519100189209, 'timing/data': 0.0013708947226405144, 'timing/model': 1.5469891941174865, 'learning_rate/qwen_vl_interface': 5.379745296482559e-06, 'learning_rate/action_model': 5.379745296482559e-05, 'epoch': 1.34}) +[2026-06-18 00:11:41,935][starVLA.training.train_starvla][INFO] - Step 3491, Loss: {'train/loss': 0.41636067628860474, 'train/grad_norm_pre_clip': 14.104422569274902, 'timing/data': 0.0013325391337275505, 'timing/model': 1.5961000714451075, 'learning_rate/qwen_vl_interface': 5.377583399454605e-06, 'learning_rate/action_model': 5.3775833994546044e-05, 'epoch': 1.34}) +[2026-06-18 00:11:47,311][starVLA.training.train_starvla][INFO] - Step 3492, Loss: {'train/loss': 0.5673775672912598, 'train/grad_norm_pre_clip': 15.114656448364258, 'timing/data': 0.0011528199538588524, 'timing/model': 1.537224286235869, 'learning_rate/qwen_vl_interface': 5.375421475978476e-06, 'learning_rate/action_model': 5.375421475978476e-05, 'epoch': 1.34}) +[2026-06-18 00:11:58,318][starVLA.training.train_starvla][INFO] - Step 3493, Loss: {'train/loss': 0.4073880612850189, 'train/grad_norm_pre_clip': 11.209429740905762, 'timing/data': 0.00148699339479208, 'timing/model': 3.047114674001932, 'learning_rate/qwen_vl_interface': 5.37325952650234e-06, 'learning_rate/action_model': 5.37325952650234e-05, 'epoch': 1.34}) +[2026-06-18 00:12:03,938][starVLA.training.train_starvla][INFO] - Step 3494, Loss: {'train/loss': 0.5314658284187317, 'train/grad_norm_pre_clip': 12.534234046936035, 'timing/data': 0.0011452129110693932, 'timing/model': 1.6595429023727775, 'learning_rate/qwen_vl_interface': 5.371097551474371e-06, 'learning_rate/action_model': 5.371097551474371e-05, 'epoch': 1.34}) +[2026-06-18 00:12:08,960][starVLA.training.train_starvla][INFO] - Step 3495, Loss: {'train/loss': 0.4238486588001251, 'train/grad_norm_pre_clip': 13.605023384094238, 'timing/data': 0.0011163903400301933, 'timing/model': 1.1002020556479692, 'learning_rate/qwen_vl_interface': 5.36893555134275e-06, 'learning_rate/action_model': 5.3689355513427494e-05, 'epoch': 1.34}) +[2026-06-18 00:12:14,138][starVLA.training.train_starvla][INFO] - Step 3496, Loss: {'train/loss': 0.2681506276130676, 'train/grad_norm_pre_clip': 8.10448932647705, 'timing/data': 0.0012405738234519958, 'timing/model': 1.2371474746614695, 'learning_rate/qwen_vl_interface': 5.366773526555661e-06, 'learning_rate/action_model': 5.366773526555661e-05, 'epoch': 1.34}) +[2026-06-18 00:12:18,757][starVLA.training.train_starvla][INFO] - Step 3497, Loss: {'train/loss': 0.24005113542079926, 'train/grad_norm_pre_clip': 7.396295070648193, 'timing/data': 0.0012307502329349518, 'timing/model': 1.4419940514490008, 'learning_rate/qwen_vl_interface': 5.364611477561294e-06, 'learning_rate/action_model': 5.3646114775612945e-05, 'epoch': 1.35}) +[2026-06-18 00:12:23,983][starVLA.training.train_starvla][INFO] - Step 3498, Loss: {'train/loss': 0.3023848235607147, 'train/grad_norm_pre_clip': 8.656667709350586, 'timing/data': 0.0009454395622014999, 'timing/model': 1.0609714640304446, 'learning_rate/qwen_vl_interface': 5.362449404807845e-06, 'learning_rate/action_model': 5.362449404807844e-05, 'epoch': 1.35}) +[2026-06-18 00:12:31,783][starVLA.training.train_starvla][INFO] - Step 3499, Loss: {'train/loss': 0.5510536432266235, 'train/grad_norm_pre_clip': 11.178576469421387, 'timing/data': 0.001161225140094757, 'timing/model': 1.047672095708549, 'learning_rate/qwen_vl_interface': 5.360287308743512e-06, 'learning_rate/action_model': 5.3602873087435126e-05, 'epoch': 1.35}) +[2026-06-18 00:16:36,011][starVLA.training.train_starvla][INFO] - Step 3500, Loss: {'train/loss': 0.33845433592796326, 'train/grad_norm_pre_clip': 8.67175579071045, 'eval/loss': 0.4777756631374359, 'eval/latency_4/loss': 0.4777756631374359, 'eval/demon_fire_precision': 0.955583359399437, 'eval/demon_fire_recall': 0.9922052614485223, 'eval/demon_fire_f1': 0.9735500318674314, 'eval/demon_fire_rate_model': 0.938913362701909, 'eval/demon_fire_rate_teacher': 0.9042584434654919, 'eval/demon_left_precision': 0.8371335504885994, 'eval/demon_left_recall': 0.58675799086758, 'eval/demon_left_f1': 0.6899328859060403, 'eval/demon_left_rate_model': 0.09016152716593245, 'eval/demon_left_rate_teacher': 0.12863436123348018, 'eval/demon_right_precision': 0.8584269662921349, 'eval/demon_right_recall': 0.6895306859205776, 'eval/demon_right_f1': 0.7647647647647648, 'eval/demon_right_rate_model': 0.13069016152716592, 'eval/demon_right_rate_teacher': 0.16270190895741557, 'eval/demon_move_f1': 0.7273488253354026, 'eval/demon_cc_f1': 0.8838022370595813, 'eval/latency_4/demon_fire_precision': 0.955583359399437, 'eval/latency_4/demon_fire_recall': 0.9922052614485223, 'eval/latency_4/demon_fire_f1': 0.9735500318674314, 'eval/latency_4/demon_fire_rate_model': 0.938913362701909, 'eval/latency_4/demon_fire_rate_teacher': 0.9042584434654919, 'eval/latency_4/demon_left_precision': 0.8371335504885994, 'eval/latency_4/demon_left_recall': 0.58675799086758, 'eval/latency_4/demon_left_f1': 0.6899328859060403, 'eval/latency_4/demon_left_rate_model': 0.09016152716593245, 'eval/latency_4/demon_left_rate_teacher': 0.12863436123348018, 'eval/latency_4/demon_right_precision': 0.8584269662921349, 'eval/latency_4/demon_right_recall': 0.6895306859205776, 'eval/latency_4/demon_right_f1': 0.7647647647647648, 'eval/latency_4/demon_right_rate_model': 0.13069016152716592, 'eval/latency_4/demon_right_rate_teacher': 0.16270190895741557, 'eval/latency_4/demon_move_f1': 0.7273488253354026, 'eval/latency_4/demon_cc_f1': 0.8838022370595813, 'timing/data': 0.0009485110640525818, 'timing/model': 1.4572466546669602, 'learning_rate/qwen_vl_interface': 5.358125189816502e-06, 'learning_rate/action_model': 5.358125189816502e-05, 'epoch': 1.35}) +[2026-06-18 00:16:36,027][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_3500_state +[2026-06-18 00:16:36,035][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 00:16:45,231][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_3500_state/model.safetensors +[2026-06-18 00:17:09,183][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_3500_state/optimizer.bin +[2026-06-18 00:17:09,184][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_3500_state/sampler.bin +[2026-06-18 00:17:09,185][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_3500_state/sampler_1.bin +[2026-06-18 00:17:09,187][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_3500_state/random_states_0.pkl +[2026-06-18 00:17:09,197][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3500 +[2026-06-18 00:17:09,197][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 00:17:09,221][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 00:17:18,498][starVLA.training.train_starvla][INFO] - Step 3501, Loss: {'train/loss': 0.3972094655036926, 'train/grad_norm_pre_clip': 14.452832221984863, 'timing/data': 0.0011958079412579536, 'timing/model': 4.890572628937662, 'learning_rate/qwen_vl_interface': 5.355963048475024e-06, 'learning_rate/action_model': 5.3559630484750244e-05, 'epoch': 1.35}) +[2026-06-18 00:17:24,155][starVLA.training.train_starvla][INFO] - Step 3502, Loss: {'train/loss': 0.31099459528923035, 'train/grad_norm_pre_clip': 11.697904586791992, 'timing/data': 0.001375456340610981, 'timing/model': 1.9841570751741529, 'learning_rate/qwen_vl_interface': 5.3538008851672916e-06, 'learning_rate/action_model': 5.3538008851672916e-05, 'epoch': 1.35}) +[2026-06-18 00:17:31,891][starVLA.training.train_starvla][INFO] - Step 3503, Loss: {'train/loss': 0.42395663261413574, 'train/grad_norm_pre_clip': 16.43128204345703, 'timing/data': 0.001979040913283825, 'timing/model': 2.7606425415724516, 'learning_rate/qwen_vl_interface': 5.351638700341525e-06, 'learning_rate/action_model': 5.351638700341525e-05, 'epoch': 1.35}) +[2026-06-18 00:17:37,636][starVLA.training.train_starvla][INFO] - Step 3504, Loss: {'train/loss': 0.4522976577281952, 'train/grad_norm_pre_clip': 10.405512809753418, 'timing/data': 0.0012706434354186058, 'timing/model': 1.1073058126494288, 'learning_rate/qwen_vl_interface': 5.349476494445947e-06, 'learning_rate/action_model': 5.3494764944459466e-05, 'epoch': 1.35}) +[2026-06-18 00:17:43,521][starVLA.training.train_starvla][INFO] - Step 3505, Loss: {'train/loss': 0.43023163080215454, 'train/grad_norm_pre_clip': 11.796954154968262, 'timing/data': 0.0012154420837759972, 'timing/model': 1.430098233744502, 'learning_rate/qwen_vl_interface': 5.3473142679287824e-06, 'learning_rate/action_model': 5.3473142679287826e-05, 'epoch': 1.35}) +[2026-06-18 00:17:48,603][starVLA.training.train_starvla][INFO] - Step 3506, Loss: {'train/loss': 0.409487783908844, 'train/grad_norm_pre_clip': 12.240331649780273, 'timing/data': 0.0013783285394310951, 'timing/model': 1.0948819471523166, 'learning_rate/qwen_vl_interface': 5.345152021238268e-06, 'learning_rate/action_model': 5.345152021238268e-05, 'epoch': 1.35}) +[2026-06-18 00:17:57,100][starVLA.training.train_starvla][INFO] - Step 3507, Loss: {'train/loss': 0.3175450563430786, 'train/grad_norm_pre_clip': 8.607412338256836, 'timing/data': 0.0014853458851575851, 'timing/model': 1.6372374556958675, 'learning_rate/qwen_vl_interface': 5.342989754822635e-06, 'learning_rate/action_model': 5.342989754822635e-05, 'epoch': 1.35}) +[2026-06-18 00:18:02,763][starVLA.training.train_starvla][INFO] - Step 3508, Loss: {'train/loss': 0.36213257908821106, 'train/grad_norm_pre_clip': 8.852400779724121, 'timing/data': 0.0014497060328722, 'timing/model': 2.1098361369222403, 'learning_rate/qwen_vl_interface': 5.340827469130126e-06, 'learning_rate/action_model': 5.3408274691301255e-05, 'epoch': 1.35}) +[2026-06-18 00:18:08,118][starVLA.training.train_starvla][INFO] - Step 3509, Loss: {'train/loss': 0.44114500284194946, 'train/grad_norm_pre_clip': 12.834786415100098, 'timing/data': 0.0009531117975711823, 'timing/model': 1.6444376045837998, 'learning_rate/qwen_vl_interface': 5.338665164608985e-06, 'learning_rate/action_model': 5.3386651646089845e-05, 'epoch': 1.35}) +[2026-06-18 00:18:18,254][starVLA.training.train_starvla][INFO] - Step 3510, Loss: {'train/loss': 0.43152928352355957, 'train/grad_norm_pre_clip': 10.88737964630127, 'timing/data': 0.0013185692951083183, 'timing/model': 1.1884431866928935, 'learning_rate/qwen_vl_interface': 5.336502841707461e-06, 'learning_rate/action_model': 5.33650284170746e-05, 'epoch': 1.35}) +[2026-06-18 00:18:27,827][starVLA.training.train_starvla][INFO] - Step 3511, Loss: {'train/loss': 0.25211068987846375, 'train/grad_norm_pre_clip': 9.318694114685059, 'timing/data': 0.002088000066578388, 'timing/model': 1.2508584577590227, 'learning_rate/qwen_vl_interface': 5.334340500873803e-06, 'learning_rate/action_model': 5.3343405008738034e-05, 'epoch': 1.35}) +[2026-06-18 00:18:36,938][starVLA.training.train_starvla][INFO] - Step 3512, Loss: {'train/loss': 0.34162941575050354, 'train/grad_norm_pre_clip': 10.686474800109863, 'timing/data': 0.0011675301939249039, 'timing/model': 1.8146779658272862, 'learning_rate/qwen_vl_interface': 5.33217814255627e-06, 'learning_rate/action_model': 5.33217814255627e-05, 'epoch': 1.35}) +[2026-06-18 00:18:42,105][starVLA.training.train_starvla][INFO] - Step 3513, Loss: {'train/loss': 0.3451559543609619, 'train/grad_norm_pre_clip': 10.995489120483398, 'timing/data': 0.0011538779363036156, 'timing/model': 1.525847696699202, 'learning_rate/qwen_vl_interface': 5.330015767203118e-06, 'learning_rate/action_model': 5.330015767203118e-05, 'epoch': 1.35}) +[2026-06-18 00:18:47,555][starVLA.training.train_starvla][INFO] - Step 3514, Loss: {'train/loss': 0.22285941243171692, 'train/grad_norm_pre_clip': 7.926778316497803, 'timing/data': 0.0011074747890233994, 'timing/model': 1.1904732109978795, 'learning_rate/qwen_vl_interface': 5.3278533752626135e-06, 'learning_rate/action_model': 5.3278533752626136e-05, 'epoch': 1.35}) +[2026-06-18 00:18:55,156][starVLA.training.train_starvla][INFO] - Step 3515, Loss: {'train/loss': 0.22658412158489227, 'train/grad_norm_pre_clip': 6.998801231384277, 'timing/data': 0.0009647514671087265, 'timing/model': 1.3644979940727353, 'learning_rate/qwen_vl_interface': 5.325690967183019e-06, 'learning_rate/action_model': 5.325690967183019e-05, 'epoch': 1.35}) +[2026-06-18 00:19:02,778][starVLA.training.train_starvla][INFO] - Step 3516, Loss: {'train/loss': 0.421943336725235, 'train/grad_norm_pre_clip': 17.51729965209961, 'timing/data': 0.001161755993962288, 'timing/model': 1.1506604943424463, 'learning_rate/qwen_vl_interface': 5.323528543412607e-06, 'learning_rate/action_model': 5.323528543412607e-05, 'epoch': 1.35}) +[2026-06-18 00:19:08,331][starVLA.training.train_starvla][INFO] - Step 3517, Loss: {'train/loss': 0.5763967037200928, 'train/grad_norm_pre_clip': 12.32156753540039, 'timing/data': 0.0011421386152505875, 'timing/model': 1.7668446013703942, 'learning_rate/qwen_vl_interface': 5.321366104399648e-06, 'learning_rate/action_model': 5.321366104399648e-05, 'epoch': 1.35}) +[2026-06-18 00:19:13,381][starVLA.training.train_starvla][INFO] - Step 3518, Loss: {'train/loss': 0.5997360348701477, 'train/grad_norm_pre_clip': 13.436744689941406, 'timing/data': 0.0009130192920565605, 'timing/model': 1.6833665845915675, 'learning_rate/qwen_vl_interface': 5.319203650592419e-06, 'learning_rate/action_model': 5.3192036505924194e-05, 'epoch': 1.35}) +[2026-06-18 00:19:19,481][starVLA.training.train_starvla][INFO] - Step 3519, Loss: {'train/loss': 0.3423173725605011, 'train/grad_norm_pre_clip': 9.213136672973633, 'timing/data': 0.0011939574033021927, 'timing/model': 1.166334648616612, 'learning_rate/qwen_vl_interface': 5.317041182439199e-06, 'learning_rate/action_model': 5.3170411824391985e-05, 'epoch': 1.35}) +[2026-06-18 00:19:28,122][starVLA.training.train_starvla][INFO] - Step 3520, Loss: {'train/loss': 0.38070908188819885, 'train/grad_norm_pre_clip': 9.276007652282715, 'timing/data': 0.0011188294738531113, 'timing/model': 2.84477236866951, 'learning_rate/qwen_vl_interface': 5.314878700388267e-06, 'learning_rate/action_model': 5.314878700388267e-05, 'epoch': 1.35}) +[2026-06-18 00:19:33,019][starVLA.training.train_starvla][INFO] - Step 3521, Loss: {'train/loss': 0.24967126548290253, 'train/grad_norm_pre_clip': 8.069490432739258, 'timing/data': 0.0009242743253707886, 'timing/model': 1.0889200419187546, 'learning_rate/qwen_vl_interface': 5.312716204887914e-06, 'learning_rate/action_model': 5.3127162048879145e-05, 'epoch': 1.35}) +[2026-06-18 00:19:38,075][starVLA.training.train_starvla][INFO] - Step 3522, Loss: {'train/loss': 0.4203442633152008, 'train/grad_norm_pre_clip': 14.382874488830566, 'timing/data': 0.0012097647413611412, 'timing/model': 1.5783695606514812, 'learning_rate/qwen_vl_interface': 5.3105536963864226e-06, 'learning_rate/action_model': 5.310553696386422e-05, 'epoch': 1.35}) +[2026-06-18 00:19:42,476][starVLA.training.train_starvla][INFO] - Step 3523, Loss: {'train/loss': 0.18158848583698273, 'train/grad_norm_pre_clip': 8.339397430419922, 'timing/data': 0.001243731938302517, 'timing/model': 1.2128131976351142, 'learning_rate/qwen_vl_interface': 5.3083911753320875e-06, 'learning_rate/action_model': 5.308391175332087e-05, 'epoch': 1.36}) +[2026-06-18 00:19:49,157][starVLA.training.train_starvla][INFO] - Step 3524, Loss: {'train/loss': 0.4671664834022522, 'train/grad_norm_pre_clip': 13.106651306152344, 'timing/data': 0.001319010742008686, 'timing/model': 2.809644523076713, 'learning_rate/qwen_vl_interface': 5.306228642173194e-06, 'learning_rate/action_model': 5.306228642173194e-05, 'epoch': 1.36}) +[2026-06-18 00:19:54,545][starVLA.training.train_starvla][INFO] - Step 3525, Loss: {'train/loss': 0.44774121046066284, 'train/grad_norm_pre_clip': 12.078047752380371, 'timing/data': 0.0033280272036790848, 'timing/model': 1.3005609968677163, 'learning_rate/qwen_vl_interface': 5.304066097358046e-06, 'learning_rate/action_model': 5.304066097358046e-05, 'epoch': 1.36}) +[2026-06-18 00:20:00,159][starVLA.training.train_starvla][INFO] - Step 3526, Loss: {'train/loss': 0.23977254331111908, 'train/grad_norm_pre_clip': 9.050506591796875, 'timing/data': 0.001016806811094284, 'timing/model': 1.5520868198946118, 'learning_rate/qwen_vl_interface': 5.301903541334937e-06, 'learning_rate/action_model': 5.301903541334937e-05, 'epoch': 1.36}) +[2026-06-18 00:20:05,369][starVLA.training.train_starvla][INFO] - Step 3527, Loss: {'train/loss': 0.1903480738401413, 'train/grad_norm_pre_clip': 6.630851745605469, 'timing/data': 0.0009854426607489586, 'timing/model': 1.4214746272191405, 'learning_rate/qwen_vl_interface': 5.299740974552168e-06, 'learning_rate/action_model': 5.299740974552167e-05, 'epoch': 1.36}) +[2026-06-18 00:20:17,865][starVLA.training.train_starvla][INFO] - Step 3528, Loss: {'train/loss': 0.3263295590877533, 'train/grad_norm_pre_clip': 12.602599143981934, 'timing/data': 0.001397845335304737, 'timing/model': 1.7749736299738288, 'learning_rate/qwen_vl_interface': 5.297578397458042e-06, 'learning_rate/action_model': 5.297578397458042e-05, 'epoch': 1.36}) +[2026-06-18 00:20:26,146][starVLA.training.train_starvla][INFO] - Step 3529, Loss: {'train/loss': 0.35149654746055603, 'train/grad_norm_pre_clip': 7.3483357429504395, 'timing/data': 0.0066567594185471535, 'timing/model': 1.4327878737822175, 'learning_rate/qwen_vl_interface': 5.295415810500862e-06, 'learning_rate/action_model': 5.295415810500862e-05, 'epoch': 1.36}) +[2026-06-18 00:20:31,386][starVLA.training.train_starvla][INFO] - Step 3530, Loss: {'train/loss': 0.38198530673980713, 'train/grad_norm_pre_clip': 17.86077308654785, 'timing/data': 0.0014880858361721039, 'timing/model': 1.1348379654809833, 'learning_rate/qwen_vl_interface': 5.293253214128938e-06, 'learning_rate/action_model': 5.293253214128937e-05, 'epoch': 1.36}) +[2026-06-18 00:20:38,650][starVLA.training.train_starvla][INFO] - Step 3531, Loss: {'train/loss': 0.4196631908416748, 'train/grad_norm_pre_clip': 16.825618743896484, 'timing/data': 0.0013470770791172981, 'timing/model': 2.6008174559101462, 'learning_rate/qwen_vl_interface': 5.291090608790572e-06, 'learning_rate/action_model': 5.291090608790572e-05, 'epoch': 1.36}) +[2026-06-18 00:20:47,143][starVLA.training.train_starvla][INFO] - Step 3532, Loss: {'train/loss': 0.19671717286109924, 'train/grad_norm_pre_clip': 7.155463695526123, 'timing/data': 0.0011670151725411415, 'timing/model': 1.6444626320153475, 'learning_rate/qwen_vl_interface': 5.28892799493408e-06, 'learning_rate/action_model': 5.2889279949340796e-05, 'epoch': 1.36}) +[2026-06-18 00:20:52,476][starVLA.training.train_starvla][INFO] - Step 3533, Loss: {'train/loss': 0.22689728438854218, 'train/grad_norm_pre_clip': 7.86143159866333, 'timing/data': 0.0011330824345350266, 'timing/model': 1.3179752985015512, 'learning_rate/qwen_vl_interface': 5.286765373007774e-06, 'learning_rate/action_model': 5.2867653730077735e-05, 'epoch': 1.36}) +[2026-06-18 00:20:58,220][starVLA.training.train_starvla][INFO] - Step 3534, Loss: {'train/loss': 0.17483435571193695, 'train/grad_norm_pre_clip': 7.226797103881836, 'timing/data': 0.0012576915323734283, 'timing/model': 1.4198986142873764, 'learning_rate/qwen_vl_interface': 5.284602743459968e-06, 'learning_rate/action_model': 5.284602743459968e-05, 'epoch': 1.36}) +[2026-06-18 00:21:03,244][starVLA.training.train_starvla][INFO] - Step 3535, Loss: {'train/loss': 0.344333678483963, 'train/grad_norm_pre_clip': 15.197671890258789, 'timing/data': 0.00089995376765728, 'timing/model': 1.0973460478708148, 'learning_rate/qwen_vl_interface': 5.282440106738973e-06, 'learning_rate/action_model': 5.282440106738973e-05, 'epoch': 1.36}) +[2026-06-18 00:21:08,949][starVLA.training.train_starvla][INFO] - Step 3536, Loss: {'train/loss': 0.4838145673274994, 'train/grad_norm_pre_clip': 18.996484756469727, 'timing/data': 0.0010730326175689697, 'timing/model': 1.5377639820799232, 'learning_rate/qwen_vl_interface': 5.280277463293111e-06, 'learning_rate/action_model': 5.280277463293111e-05, 'epoch': 1.36}) +[2026-06-18 00:21:17,608][starVLA.training.train_starvla][INFO] - Step 3537, Loss: {'train/loss': 0.8681637048721313, 'train/grad_norm_pre_clip': 30.568832397460938, 'timing/data': 0.0009346753358840942, 'timing/model': 2.153729693032801, 'learning_rate/qwen_vl_interface': 5.2781148135706975e-06, 'learning_rate/action_model': 5.278114813570697e-05, 'epoch': 1.36}) +[2026-06-18 00:21:27,486][starVLA.training.train_starvla][INFO] - Step 3538, Loss: {'train/loss': 0.31850001215934753, 'train/grad_norm_pre_clip': 12.558582305908203, 'timing/data': 0.0013619009405374527, 'timing/model': 2.906797639094293, 'learning_rate/qwen_vl_interface': 5.275952158020054e-06, 'learning_rate/action_model': 5.275952158020053e-05, 'epoch': 1.36}) +[2026-06-18 00:21:38,636][starVLA.training.train_starvla][INFO] - Step 3539, Loss: {'train/loss': 0.42241233587265015, 'train/grad_norm_pre_clip': 9.808445930480957, 'timing/data': 0.0012992341071367264, 'timing/model': 1.1150877643376589, 'learning_rate/qwen_vl_interface': 5.273789497089497e-06, 'learning_rate/action_model': 5.2737894970894974e-05, 'epoch': 1.36}) +[2026-06-18 00:21:43,441][starVLA.training.train_starvla][INFO] - Step 3540, Loss: {'train/loss': 0.31016960740089417, 'train/grad_norm_pre_clip': 16.11186408996582, 'timing/data': 0.001484844833612442, 'timing/model': 1.094836501404643, 'learning_rate/qwen_vl_interface': 5.271626831227355e-06, 'learning_rate/action_model': 5.2716268312273544e-05, 'epoch': 1.36}) +[2026-06-18 00:21:48,695][starVLA.training.train_starvla][INFO] - Step 3541, Loss: {'train/loss': 0.3152853846549988, 'train/grad_norm_pre_clip': 13.553677558898926, 'timing/data': 0.00675545446574688, 'timing/model': 1.285282902419567, 'learning_rate/qwen_vl_interface': 5.2694641608819465e-06, 'learning_rate/action_model': 5.269464160881946e-05, 'epoch': 1.36}) +[2026-06-18 00:21:54,281][starVLA.training.train_starvla][INFO] - Step 3542, Loss: {'train/loss': 0.39172443747520447, 'train/grad_norm_pre_clip': 11.841604232788086, 'timing/data': 0.001260233111679554, 'timing/model': 1.5428873635828495, 'learning_rate/qwen_vl_interface': 5.267301486501595e-06, 'learning_rate/action_model': 5.2673014865015945e-05, 'epoch': 1.36}) +[2026-06-18 00:22:07,730][starVLA.training.train_starvla][INFO] - Step 3543, Loss: {'train/loss': 0.28249984979629517, 'train/grad_norm_pre_clip': 11.61031436920166, 'timing/data': 0.0011477097868919373, 'timing/model': 1.2557264622300863, 'learning_rate/qwen_vl_interface': 5.265138808534628e-06, 'learning_rate/action_model': 5.265138808534628e-05, 'epoch': 1.36}) +[2026-06-18 00:22:24,403][starVLA.training.train_starvla][INFO] - Step 3544, Loss: {'train/loss': 0.3159758448600769, 'train/grad_norm_pre_clip': 10.96410846710205, 'timing/data': 0.0011639995500445366, 'timing/model': 1.3655856391415, 'learning_rate/qwen_vl_interface': 5.26297612742937e-06, 'learning_rate/action_model': 5.26297612742937e-05, 'epoch': 1.36}) +[2026-06-18 00:22:32,494][starVLA.training.train_starvla][INFO] - Step 3545, Loss: {'train/loss': 0.1998339146375656, 'train/grad_norm_pre_clip': 9.309385299682617, 'timing/data': 0.00110530573874712, 'timing/model': 1.9215893158689141, 'learning_rate/qwen_vl_interface': 5.2608134436341476e-06, 'learning_rate/action_model': 5.260813443634147e-05, 'epoch': 1.36}) +[2026-06-18 00:22:43,081][starVLA.training.train_starvla][INFO] - Step 3546, Loss: {'train/loss': 0.3295191526412964, 'train/grad_norm_pre_clip': 17.526264190673828, 'timing/data': 0.0008391356095671654, 'timing/model': 1.654744254425168, 'learning_rate/qwen_vl_interface': 5.258650757597285e-06, 'learning_rate/action_model': 5.258650757597286e-05, 'epoch': 1.36}) +[2026-06-18 00:22:52,234][starVLA.training.train_starvla][INFO] - Step 3547, Loss: {'train/loss': 0.7298305630683899, 'train/grad_norm_pre_clip': 15.478229522705078, 'timing/data': 0.0017985496670007706, 'timing/model': 1.412908905185759, 'learning_rate/qwen_vl_interface': 5.256488069767113e-06, 'learning_rate/action_model': 5.256488069767113e-05, 'epoch': 1.36}) +[2026-06-18 00:22:56,687][starVLA.training.train_starvla][INFO] - Step 3548, Loss: {'train/loss': 0.5812422037124634, 'train/grad_norm_pre_clip': 14.705159187316895, 'timing/data': 0.001552932895720005, 'timing/model': 1.1323308888822794, 'learning_rate/qwen_vl_interface': 5.254325380591955e-06, 'learning_rate/action_model': 5.254325380591955e-05, 'epoch': 1.36}) +[2026-06-18 00:23:03,019][starVLA.training.train_starvla][INFO] - Step 3549, Loss: {'train/loss': 0.3834221065044403, 'train/grad_norm_pre_clip': 11.224327087402344, 'timing/data': 0.0012830104678869247, 'timing/model': 1.7694456642493606, 'learning_rate/qwen_vl_interface': 5.252162690520142e-06, 'learning_rate/action_model': 5.252162690520142e-05, 'epoch': 1.37}) +[2026-06-18 00:23:08,031][starVLA.training.train_starvla][INFO] - Step 3550, Loss: {'train/loss': 0.5521341562271118, 'train/grad_norm_pre_clip': 12.660382270812988, 'timing/data': 0.0016619358211755753, 'timing/model': 1.1330405808985233, 'learning_rate/qwen_vl_interface': 5.2500000000000006e-06, 'learning_rate/action_model': 5.25e-05, 'epoch': 1.37}) +[2026-06-18 00:23:13,966][starVLA.training.train_starvla][INFO] - Step 3551, Loss: {'train/loss': 0.29660022258758545, 'train/grad_norm_pre_clip': 9.453575134277344, 'timing/data': 0.0078305434435606, 'timing/model': 1.2385961236432195, 'learning_rate/qwen_vl_interface': 5.24783730947986e-06, 'learning_rate/action_model': 5.2478373094798594e-05, 'epoch': 1.37}) +[2026-06-18 00:23:18,941][starVLA.training.train_starvla][INFO] - Step 3552, Loss: {'train/loss': 0.4730605185031891, 'train/grad_norm_pre_clip': 12.234986305236816, 'timing/data': 0.0012786965817213058, 'timing/model': 1.7709860522300005, 'learning_rate/qwen_vl_interface': 5.245674619408045e-06, 'learning_rate/action_model': 5.245674619408045e-05, 'epoch': 1.37}) +[2026-06-18 00:23:24,066][starVLA.training.train_starvla][INFO] - Step 3553, Loss: {'train/loss': 0.3573848009109497, 'train/grad_norm_pre_clip': 10.897517204284668, 'timing/data': 0.0008972333744168282, 'timing/model': 1.293391041457653, 'learning_rate/qwen_vl_interface': 5.243511930232889e-06, 'learning_rate/action_model': 5.243511930232888e-05, 'epoch': 1.37}) +[2026-06-18 00:23:29,153][starVLA.training.train_starvla][INFO] - Step 3554, Loss: {'train/loss': 0.5555564761161804, 'train/grad_norm_pre_clip': 14.940106391906738, 'timing/data': 0.0028457418084144592, 'timing/model': 1.1023889416828752, 'learning_rate/qwen_vl_interface': 5.2413492424027165e-06, 'learning_rate/action_model': 5.241349242402717e-05, 'epoch': 1.37}) +[2026-06-18 00:23:34,036][starVLA.training.train_starvla][INFO] - Step 3555, Loss: {'train/loss': 0.25119778513908386, 'train/grad_norm_pre_clip': 10.058173179626465, 'timing/data': 0.0015485426411032677, 'timing/model': 1.1377824898809195, 'learning_rate/qwen_vl_interface': 5.239186556365855e-06, 'learning_rate/action_model': 5.239186556365855e-05, 'epoch': 1.37}) +[2026-06-18 00:23:39,002][starVLA.training.train_starvla][INFO] - Step 3556, Loss: {'train/loss': 0.14159908890724182, 'train/grad_norm_pre_clip': 6.509703636169434, 'timing/data': 0.0007932065054774284, 'timing/model': 1.4762691287323833, 'learning_rate/qwen_vl_interface': 5.23702387257063e-06, 'learning_rate/action_model': 5.23702387257063e-05, 'epoch': 1.37}) +[2026-06-18 00:23:44,340][starVLA.training.train_starvla][INFO] - Step 3557, Loss: {'train/loss': 0.2416238635778427, 'train/grad_norm_pre_clip': 8.932234764099121, 'timing/data': 0.0009196074679493904, 'timing/model': 1.7599960220977664, 'learning_rate/qwen_vl_interface': 5.234861191465373e-06, 'learning_rate/action_model': 5.234861191465372e-05, 'epoch': 1.37}) +[2026-06-18 00:23:49,072][starVLA.training.train_starvla][INFO] - Step 3558, Loss: {'train/loss': 0.40066269040107727, 'train/grad_norm_pre_clip': 11.863509178161621, 'timing/data': 0.0010840939357876778, 'timing/model': 1.2385676316916943, 'learning_rate/qwen_vl_interface': 5.232698513498407e-06, 'learning_rate/action_model': 5.2326985134984066e-05, 'epoch': 1.37}) +[2026-06-18 00:23:54,173][starVLA.training.train_starvla][INFO] - Step 3559, Loss: {'train/loss': 0.5461797118186951, 'train/grad_norm_pre_clip': 15.989543914794922, 'timing/data': 0.0011631809175014496, 'timing/model': 1.1675678482279181, 'learning_rate/qwen_vl_interface': 5.230535839118057e-06, 'learning_rate/action_model': 5.230535839118057e-05, 'epoch': 1.37}) +[2026-06-18 00:23:59,694][starVLA.training.train_starvla][INFO] - Step 3560, Loss: {'train/loss': 0.547305703163147, 'train/grad_norm_pre_clip': 15.42881965637207, 'timing/data': 0.0008043358102440834, 'timing/model': 1.1605601590126753, 'learning_rate/qwen_vl_interface': 5.2283731687726455e-06, 'learning_rate/action_model': 5.2283731687726454e-05, 'epoch': 1.37}) +[2026-06-18 00:24:04,675][starVLA.training.train_starvla][INFO] - Step 3561, Loss: {'train/loss': 0.5690831542015076, 'train/grad_norm_pre_clip': 16.935001373291016, 'timing/data': 0.0012930454686284065, 'timing/model': 1.3958194348961115, 'learning_rate/qwen_vl_interface': 5.226210502910504e-06, 'learning_rate/action_model': 5.226210502910504e-05, 'epoch': 1.37}) +[2026-06-18 00:24:10,224][starVLA.training.train_starvla][INFO] - Step 3562, Loss: {'train/loss': 0.3885001242160797, 'train/grad_norm_pre_clip': 10.450236320495605, 'timing/data': 0.0009718900546431541, 'timing/model': 1.7474775798618793, 'learning_rate/qwen_vl_interface': 5.224047841979949e-06, 'learning_rate/action_model': 5.2240478419799485e-05, 'epoch': 1.37}) +[2026-06-18 00:24:15,296][starVLA.training.train_starvla][INFO] - Step 3563, Loss: {'train/loss': 0.21399229764938354, 'train/grad_norm_pre_clip': 7.022731304168701, 'timing/data': 0.0013531502336263657, 'timing/model': 1.2742469320073724, 'learning_rate/qwen_vl_interface': 5.2218851864293044e-06, 'learning_rate/action_model': 5.221885186429304e-05, 'epoch': 1.37}) +[2026-06-18 00:24:20,272][starVLA.training.train_starvla][INFO] - Step 3564, Loss: {'train/loss': 0.20534493029117584, 'train/grad_norm_pre_clip': 9.045989036560059, 'timing/data': 0.00160939060151577, 'timing/model': 1.109887646511197, 'learning_rate/qwen_vl_interface': 5.21972253670689e-06, 'learning_rate/action_model': 5.21972253670689e-05, 'epoch': 1.37}) +[2026-06-18 00:24:31,191][starVLA.training.train_starvla][INFO] - Step 3565, Loss: {'train/loss': 0.36747369170188904, 'train/grad_norm_pre_clip': 14.923476219177246, 'timing/data': 0.0011481447145342827, 'timing/model': 5.399450811557472, 'learning_rate/qwen_vl_interface': 5.217559893261028e-06, 'learning_rate/action_model': 5.217559893261028e-05, 'epoch': 1.37}) +[2026-06-18 00:24:41,587][starVLA.training.train_starvla][INFO] - Step 3566, Loss: {'train/loss': 0.33080071210861206, 'train/grad_norm_pre_clip': 9.581735610961914, 'timing/data': 0.0011902786791324615, 'timing/model': 1.652932540513575, 'learning_rate/qwen_vl_interface': 5.2153972565400355e-06, 'learning_rate/action_model': 5.215397256540035e-05, 'epoch': 1.37}) +[2026-06-18 00:24:53,617][starVLA.training.train_starvla][INFO] - Step 3567, Loss: {'train/loss': 0.4309554994106293, 'train/grad_norm_pre_clip': 9.251801490783691, 'timing/data': 0.0013142600655555725, 'timing/model': 2.628220404498279, 'learning_rate/qwen_vl_interface': 5.213234626992227e-06, 'learning_rate/action_model': 5.213234626992226e-05, 'epoch': 1.37}) +[2026-06-18 00:25:04,142][starVLA.training.train_starvla][INFO] - Step 3568, Loss: {'train/loss': 0.5540441274642944, 'train/grad_norm_pre_clip': 16.74233055114746, 'timing/data': 0.0025523388758301735, 'timing/model': 4.990863123908639, 'learning_rate/qwen_vl_interface': 5.21107200506592e-06, 'learning_rate/action_model': 5.21107200506592e-05, 'epoch': 1.37}) +[2026-06-18 00:25:16,275][starVLA.training.train_starvla][INFO] - Step 3569, Loss: {'train/loss': 0.4104154407978058, 'train/grad_norm_pre_clip': 13.845453262329102, 'timing/data': 0.0011442266404628754, 'timing/model': 1.9164253175258636, 'learning_rate/qwen_vl_interface': 5.208909391209429e-06, 'learning_rate/action_model': 5.208909391209429e-05, 'epoch': 1.37}) +[2026-06-18 00:25:32,589][starVLA.training.train_starvla][INFO] - Step 3570, Loss: {'train/loss': 0.40856581926345825, 'train/grad_norm_pre_clip': 12.161200523376465, 'timing/data': 0.0009401477873325348, 'timing/model': 4.411298613995314, 'learning_rate/qwen_vl_interface': 5.206746785871066e-06, 'learning_rate/action_model': 5.2067467858710664e-05, 'epoch': 1.37}) +[2026-06-18 00:25:39,312][starVLA.training.train_starvla][INFO] - Step 3571, Loss: {'train/loss': 0.4698653221130371, 'train/grad_norm_pre_clip': 15.062055587768555, 'timing/data': 0.002176104113459587, 'timing/model': 1.991915013641119, 'learning_rate/qwen_vl_interface': 5.204584189499139e-06, 'learning_rate/action_model': 5.2045841894991385e-05, 'epoch': 1.37}) +[2026-06-18 00:25:54,541][starVLA.training.train_starvla][INFO] - Step 3572, Loss: {'train/loss': 0.1715582013130188, 'train/grad_norm_pre_clip': 7.456918239593506, 'timing/data': 0.0016525117680430412, 'timing/model': 3.460819869302213, 'learning_rate/qwen_vl_interface': 5.202421602541959e-06, 'learning_rate/action_model': 5.2024216025419594e-05, 'epoch': 1.37}) +[2026-06-18 00:26:05,286][starVLA.training.train_starvla][INFO] - Step 3573, Loss: {'train/loss': 0.5141850709915161, 'train/grad_norm_pre_clip': 13.113052368164062, 'timing/data': 0.0009842664003372192, 'timing/model': 2.6443888610228896, 'learning_rate/qwen_vl_interface': 5.200259025447835e-06, 'learning_rate/action_model': 5.2002590254478345e-05, 'epoch': 1.37}) +[2026-06-18 00:26:19,329][starVLA.training.train_starvla][INFO] - Step 3574, Loss: {'train/loss': 0.44452965259552, 'train/grad_norm_pre_clip': 10.078909873962402, 'timing/data': 0.0012960582971572876, 'timing/model': 1.1740803634747863, 'learning_rate/qwen_vl_interface': 5.198096458665064e-06, 'learning_rate/action_model': 5.198096458665064e-05, 'epoch': 1.37}) +[2026-06-18 00:26:25,191][starVLA.training.train_starvla][INFO] - Step 3575, Loss: {'train/loss': 0.47073522210121155, 'train/grad_norm_pre_clip': 13.82470989227295, 'timing/data': 0.0009540263563394547, 'timing/model': 1.1696707932278514, 'learning_rate/qwen_vl_interface': 5.195933902641955e-06, 'learning_rate/action_model': 5.195933902641955e-05, 'epoch': 1.38}) +[2026-06-18 00:26:30,626][starVLA.training.train_starvla][INFO] - Step 3576, Loss: {'train/loss': 0.32740360498428345, 'train/grad_norm_pre_clip': 9.010062217712402, 'timing/data': 0.0011404640972614288, 'timing/model': 1.6551858680322766, 'learning_rate/qwen_vl_interface': 5.193771357826807e-06, 'learning_rate/action_model': 5.193771357826806e-05, 'epoch': 1.38}) +[2026-06-18 00:26:36,049][starVLA.training.train_starvla][INFO] - Step 3577, Loss: {'train/loss': 0.34909874200820923, 'train/grad_norm_pre_clip': 9.963964462280273, 'timing/data': 0.001337636262178421, 'timing/model': 1.634006006643176, 'learning_rate/qwen_vl_interface': 5.191608824667916e-06, 'learning_rate/action_model': 5.1916088246679165e-05, 'epoch': 1.38}) +[2026-06-18 00:26:41,201][starVLA.training.train_starvla][INFO] - Step 3578, Loss: {'train/loss': 0.1248389333486557, 'train/grad_norm_pre_clip': 5.508608818054199, 'timing/data': 0.0009483546018600464, 'timing/model': 1.6042341385036707, 'learning_rate/qwen_vl_interface': 5.189446303613578e-06, 'learning_rate/action_model': 5.1894463036135775e-05, 'epoch': 1.38}) +[2026-06-18 00:26:46,717][starVLA.training.train_starvla][INFO] - Step 3579, Loss: {'train/loss': 0.2739746868610382, 'train/grad_norm_pre_clip': 10.520330429077148, 'timing/data': 0.0013736020773649216, 'timing/model': 1.1646190294995904, 'learning_rate/qwen_vl_interface': 5.187283795112087e-06, 'learning_rate/action_model': 5.1872837951120866e-05, 'epoch': 1.38}) +[2026-06-18 00:26:54,970][starVLA.training.train_starvla][INFO] - Step 3580, Loss: {'train/loss': 0.5997312068939209, 'train/grad_norm_pre_clip': 11.56773567199707, 'timing/data': 0.0014102160930633545, 'timing/model': 1.142300270497799, 'learning_rate/qwen_vl_interface': 5.185121299611734e-06, 'learning_rate/action_model': 5.1851212996117335e-05, 'epoch': 1.38}) +[2026-06-18 00:27:06,749][starVLA.training.train_starvla][INFO] - Step 3581, Loss: {'train/loss': 0.3139950633049011, 'train/grad_norm_pre_clip': 10.898589134216309, 'timing/data': 0.0015076026320457458, 'timing/model': 3.298964975401759, 'learning_rate/qwen_vl_interface': 5.182958817560803e-06, 'learning_rate/action_model': 5.1829588175608026e-05, 'epoch': 1.38}) +[2026-06-18 00:27:11,912][starVLA.training.train_starvla][INFO] - Step 3582, Loss: {'train/loss': 0.5641102194786072, 'train/grad_norm_pre_clip': 12.06154727935791, 'timing/data': 0.0011479426175355911, 'timing/model': 1.7221890715882182, 'learning_rate/qwen_vl_interface': 5.180796349407583e-06, 'learning_rate/action_model': 5.180796349407583e-05, 'epoch': 1.38}) +[2026-06-18 00:27:20,681][starVLA.training.train_starvla][INFO] - Step 3583, Loss: {'train/loss': 0.45233067870140076, 'train/grad_norm_pre_clip': 15.619499206542969, 'timing/data': 0.0012345537543296814, 'timing/model': 1.1336309127509594, 'learning_rate/qwen_vl_interface': 5.178633895600354e-06, 'learning_rate/action_model': 5.178633895600353e-05, 'epoch': 1.38}) +[2026-06-18 00:27:27,776][starVLA.training.train_starvla][INFO] - Step 3584, Loss: {'train/loss': 0.13048578798770905, 'train/grad_norm_pre_clip': 6.460495948791504, 'timing/data': 0.001208186149597168, 'timing/model': 1.206222196109593, 'learning_rate/qwen_vl_interface': 5.176471456587394e-06, 'learning_rate/action_model': 5.176471456587394e-05, 'epoch': 1.38}) +[2026-06-18 00:27:32,821][starVLA.training.train_starvla][INFO] - Step 3585, Loss: {'train/loss': 0.5240050554275513, 'train/grad_norm_pre_clip': 12.289896965026855, 'timing/data': 0.0016122981905937195, 'timing/model': 1.3082601325586438, 'learning_rate/qwen_vl_interface': 5.1743090328169805e-06, 'learning_rate/action_model': 5.1743090328169805e-05, 'epoch': 1.38}) +[2026-06-18 00:27:37,923][starVLA.training.train_starvla][INFO] - Step 3586, Loss: {'train/loss': 0.3139140009880066, 'train/grad_norm_pre_clip': 10.880472183227539, 'timing/data': 0.0013498533517122269, 'timing/model': 1.2686117244884372, 'learning_rate/qwen_vl_interface': 5.172146624737388e-06, 'learning_rate/action_model': 5.1721466247373875e-05, 'epoch': 1.38}) +[2026-06-18 00:27:47,268][starVLA.training.train_starvla][INFO] - Step 3587, Loss: {'train/loss': 0.40619927644729614, 'train/grad_norm_pre_clip': 12.577686309814453, 'timing/data': 0.0009144479408860207, 'timing/model': 1.9773051301017404, 'learning_rate/qwen_vl_interface': 5.169984232796883e-06, 'learning_rate/action_model': 5.169984232796883e-05, 'epoch': 1.38}) +[2026-06-18 00:27:56,657][starVLA.training.train_starvla][INFO] - Step 3588, Loss: {'train/loss': 0.33895790576934814, 'train/grad_norm_pre_clip': 12.605062484741211, 'timing/data': 0.0012390650808811188, 'timing/model': 1.3727070651948452, 'learning_rate/qwen_vl_interface': 5.1678218574437325e-06, 'learning_rate/action_model': 5.167821857443732e-05, 'epoch': 1.38}) +[2026-06-18 00:28:02,144][starVLA.training.train_starvla][INFO] - Step 3589, Loss: {'train/loss': 0.44760942459106445, 'train/grad_norm_pre_clip': 12.167072296142578, 'timing/data': 0.001371988095343113, 'timing/model': 1.1276126932352781, 'learning_rate/qwen_vl_interface': 5.165659499126198e-06, 'learning_rate/action_model': 5.165659499126198e-05, 'epoch': 1.38}) +[2026-06-18 00:28:07,315][starVLA.training.train_starvla][INFO] - Step 3590, Loss: {'train/loss': 0.6050585508346558, 'train/grad_norm_pre_clip': 13.554618835449219, 'timing/data': 0.0013621533289551735, 'timing/model': 1.1191626144573092, 'learning_rate/qwen_vl_interface': 5.1634971582925405e-06, 'learning_rate/action_model': 5.163497158292541e-05, 'epoch': 1.38}) +[2026-06-18 00:28:21,170][starVLA.training.train_starvla][INFO] - Step 3591, Loss: {'train/loss': 0.45349058508872986, 'train/grad_norm_pre_clip': 14.708285331726074, 'timing/data': 0.0014375243335962296, 'timing/model': 1.9700540835037827, 'learning_rate/qwen_vl_interface': 5.161334835391016e-06, 'learning_rate/action_model': 5.161334835391016e-05, 'epoch': 1.38}) +[2026-06-18 00:28:27,499][starVLA.training.train_starvla][INFO] - Step 3592, Loss: {'train/loss': 0.3385644853115082, 'train/grad_norm_pre_clip': 37.25659942626953, 'timing/data': 0.0032327892258763313, 'timing/model': 1.5964363189414144, 'learning_rate/qwen_vl_interface': 5.159172530869875e-06, 'learning_rate/action_model': 5.159172530869874e-05, 'epoch': 1.38}) +[2026-06-18 00:28:32,389][starVLA.training.train_starvla][INFO] - Step 3593, Loss: {'train/loss': 0.41183727979660034, 'train/grad_norm_pre_clip': 12.266236305236816, 'timing/data': 0.0011862730607390404, 'timing/model': 1.1403478644788265, 'learning_rate/qwen_vl_interface': 5.157010245177366e-06, 'learning_rate/action_model': 5.1570102451773664e-05, 'epoch': 1.38}) +[2026-06-18 00:28:37,243][starVLA.training.train_starvla][INFO] - Step 3594, Loss: {'train/loss': 0.3934706747531891, 'train/grad_norm_pre_clip': 19.47270965576172, 'timing/data': 0.015977206639945507, 'timing/model': 1.1262986958026886, 'learning_rate/qwen_vl_interface': 5.1548479787617346e-06, 'learning_rate/action_model': 5.1548479787617346e-05, 'epoch': 1.38}) +[2026-06-18 00:28:42,215][starVLA.training.train_starvla][INFO] - Step 3595, Loss: {'train/loss': 0.5947571992874146, 'train/grad_norm_pre_clip': 20.21204376220703, 'timing/data': 0.03500241134315729, 'timing/model': 1.246121489442885, 'learning_rate/qwen_vl_interface': 5.152685732071219e-06, 'learning_rate/action_model': 5.1526857320712185e-05, 'epoch': 1.38}) +[2026-06-18 00:28:52,926][starVLA.training.train_starvla][INFO] - Step 3596, Loss: {'train/loss': 0.21955272555351257, 'train/grad_norm_pre_clip': 8.566587448120117, 'timing/data': 0.0238558491691947, 'timing/model': 1.9405158935114741, 'learning_rate/qwen_vl_interface': 5.1505235055540545e-06, 'learning_rate/action_model': 5.1505235055540545e-05, 'epoch': 1.38}) +[2026-06-18 00:29:01,883][starVLA.training.train_starvla][INFO] - Step 3597, Loss: {'train/loss': 0.5380858182907104, 'train/grad_norm_pre_clip': 11.363410949707031, 'timing/data': 0.029485720209777355, 'timing/model': 1.6911552539095283, 'learning_rate/qwen_vl_interface': 5.148361299658477e-06, 'learning_rate/action_model': 5.148361299658476e-05, 'epoch': 1.38}) +[2026-06-18 00:29:06,520][starVLA.training.train_starvla][INFO] - Step 3598, Loss: {'train/loss': 0.41564902663230896, 'train/grad_norm_pre_clip': 31.85439682006836, 'timing/data': 0.039853379130363464, 'timing/model': 1.24016850348562, 'learning_rate/qwen_vl_interface': 5.1461991148327095e-06, 'learning_rate/action_model': 5.1461991148327095e-05, 'epoch': 1.38}) +[2026-06-18 00:29:12,118][starVLA.training.train_starvla][INFO] - Step 3599, Loss: {'train/loss': 0.2925366163253784, 'train/grad_norm_pre_clip': 7.406904220581055, 'timing/data': 0.026304892264306545, 'timing/model': 1.5189051572233438, 'learning_rate/qwen_vl_interface': 5.144036951524977e-06, 'learning_rate/action_model': 5.144036951524977e-05, 'epoch': 1.38}) +[2026-06-18 00:34:32,500][starVLA.training.train_starvla][INFO] - Step 3600, Loss: {'train/loss': 0.5251145362854004, 'train/grad_norm_pre_clip': 13.863733291625977, 'eval/loss': 0.4812329411506653, 'eval/latency_4/loss': 0.4812329411506653, 'eval/demon_fire_precision': 0.9586881110059918, 'eval/demon_fire_recall': 0.9873335498538487, 'eval/demon_fire_f1': 0.9728, 'eval/demon_fire_rate_model': 0.9312775330396476, 'eval/demon_fire_rate_teacher': 0.9042584434654919, 'eval/demon_left_precision': 0.753731343283582, 'eval/demon_left_recall': 0.6917808219178082, 'eval/demon_left_f1': 0.7214285714285713, 'eval/demon_left_rate_model': 0.11806167400881057, 'eval/demon_left_rate_teacher': 0.12863436123348018, 'eval/demon_right_precision': 0.6953846153846154, 'eval/demon_right_recall': 0.8158844765342961, 'eval/demon_right_f1': 0.7508305647840532, 'eval/demon_right_rate_model': 0.19089574155653452, 'eval/demon_right_rate_teacher': 0.16270190895741557, 'eval/demon_move_f1': 0.7361295681063122, 'eval/demon_cc_f1': 0.8872253848318934, 'eval/latency_4/demon_fire_precision': 0.9586881110059918, 'eval/latency_4/demon_fire_recall': 0.9873335498538487, 'eval/latency_4/demon_fire_f1': 0.9728, 'eval/latency_4/demon_fire_rate_model': 0.9312775330396476, 'eval/latency_4/demon_fire_rate_teacher': 0.9042584434654919, 'eval/latency_4/demon_left_precision': 0.753731343283582, 'eval/latency_4/demon_left_recall': 0.6917808219178082, 'eval/latency_4/demon_left_f1': 0.7214285714285713, 'eval/latency_4/demon_left_rate_model': 0.11806167400881057, 'eval/latency_4/demon_left_rate_teacher': 0.12863436123348018, 'eval/latency_4/demon_right_precision': 0.6953846153846154, 'eval/latency_4/demon_right_recall': 0.8158844765342961, 'eval/latency_4/demon_right_f1': 0.7508305647840532, 'eval/latency_4/demon_right_rate_model': 0.19089574155653452, 'eval/latency_4/demon_right_rate_teacher': 0.16270190895741557, 'eval/latency_4/demon_move_f1': 0.7361295681063122, 'eval/latency_4/demon_cc_f1': 0.8872253848318934, 'timing/data': 0.02356763184070587, 'timing/model': 1.3035215036943555, 'learning_rate/qwen_vl_interface': 5.141874810183498e-06, 'learning_rate/action_model': 5.141874810183498e-05, 'epoch': 1.38}) +[2026-06-18 00:34:32,503][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_3600_state +[2026-06-18 00:34:32,541][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 00:34:42,084][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_3600_state/model.safetensors +[2026-06-18 00:35:06,261][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_3600_state/optimizer.bin +[2026-06-18 00:35:06,262][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_3600_state/sampler.bin +[2026-06-18 00:35:06,263][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_3600_state/sampler_1.bin +[2026-06-18 00:35:06,265][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_3600_state/random_states_0.pkl +[2026-06-18 00:35:06,270][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3600 +[2026-06-18 00:35:06,270][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 00:35:06,289][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 00:35:20,461][starVLA.training.train_starvla][INFO] - Step 3601, Loss: {'train/loss': 0.49622198939323425, 'train/grad_norm_pre_clip': 13.759927749633789, 'timing/data': 0.02377642784267664, 'timing/model': 1.755368203856051, 'learning_rate/qwen_vl_interface': 5.139712691256489e-06, 'learning_rate/action_model': 5.1397126912564885e-05, 'epoch': 1.39}) +[2026-06-18 00:35:27,211][starVLA.training.train_starvla][INFO] - Step 3602, Loss: {'train/loss': 0.3459746539592743, 'train/grad_norm_pre_clip': 12.623783111572266, 'timing/data': 0.023695186711847782, 'timing/model': 3.0154901556670666, 'learning_rate/qwen_vl_interface': 5.137550595192156e-06, 'learning_rate/action_model': 5.137550595192156e-05, 'epoch': 1.39}) +[2026-06-18 00:35:33,656][starVLA.training.train_starvla][INFO] - Step 3603, Loss: {'train/loss': 0.5642431378364563, 'train/grad_norm_pre_clip': 13.197393417358398, 'timing/data': 0.03001989144831896, 'timing/model': 1.2613101424649358, 'learning_rate/qwen_vl_interface': 5.135388522438707e-06, 'learning_rate/action_model': 5.1353885224387066e-05, 'epoch': 1.39}) +[2026-06-18 00:35:40,654][starVLA.training.train_starvla][INFO] - Step 3604, Loss: {'train/loss': 0.36610883474349976, 'train/grad_norm_pre_clip': 14.043734550476074, 'timing/data': 0.024977565743029118, 'timing/model': 1.1772380024194717, 'learning_rate/qwen_vl_interface': 5.133226473444341e-06, 'learning_rate/action_model': 5.13322647344434e-05, 'epoch': 1.39}) +[2026-06-18 00:35:48,342][starVLA.training.train_starvla][INFO] - Step 3605, Loss: {'train/loss': 0.5561138987541199, 'train/grad_norm_pre_clip': 14.906140327453613, 'timing/data': 0.025245247408747673, 'timing/model': 1.3128746878355742, 'learning_rate/qwen_vl_interface': 5.1310644486572515e-06, 'learning_rate/action_model': 5.131064448657251e-05, 'epoch': 1.39}) +[2026-06-18 00:35:55,053][starVLA.training.train_starvla][INFO] - Step 3606, Loss: {'train/loss': 0.19438590109348297, 'train/grad_norm_pre_clip': 5.892891883850098, 'timing/data': 0.026964678429067135, 'timing/model': 2.609926540404558, 'learning_rate/qwen_vl_interface': 5.12890244852563e-06, 'learning_rate/action_model': 5.12890244852563e-05, 'epoch': 1.39}) +[2026-06-18 00:36:01,372][starVLA.training.train_starvla][INFO] - Step 3607, Loss: {'train/loss': 0.42179861664772034, 'train/grad_norm_pre_clip': 10.573755264282227, 'timing/data': 0.022555082105100155, 'timing/model': 2.8864882178604603, 'learning_rate/qwen_vl_interface': 5.126740473497661e-06, 'learning_rate/action_model': 5.126740473497661e-05, 'epoch': 1.39}) +[2026-06-18 00:36:06,810][starVLA.training.train_starvla][INFO] - Step 3608, Loss: {'train/loss': 0.289033442735672, 'train/grad_norm_pre_clip': 11.649904251098633, 'timing/data': 0.03537044394761324, 'timing/model': 1.3327016858384013, 'learning_rate/qwen_vl_interface': 5.124578524021525e-06, 'learning_rate/action_model': 5.124578524021525e-05, 'epoch': 1.39}) +[2026-06-18 00:36:11,990][starVLA.training.train_starvla][INFO] - Step 3609, Loss: {'train/loss': 0.4288252294063568, 'train/grad_norm_pre_clip': 12.13691234588623, 'timing/data': 0.03392611537128687, 'timing/model': 1.2385590448975563, 'learning_rate/qwen_vl_interface': 5.122416600545396e-06, 'learning_rate/action_model': 5.122416600545397e-05, 'epoch': 1.39}) +[2026-06-18 00:36:17,406][starVLA.training.train_starvla][INFO] - Step 3610, Loss: {'train/loss': 0.46902140974998474, 'train/grad_norm_pre_clip': 11.17747688293457, 'timing/data': 0.024152031168341637, 'timing/model': 1.0846520503982902, 'learning_rate/qwen_vl_interface': 5.1202547035174436e-06, 'learning_rate/action_model': 5.120254703517443e-05, 'epoch': 1.39}) +[2026-06-18 00:36:22,678][starVLA.training.train_starvla][INFO] - Step 3611, Loss: {'train/loss': 0.45042964816093445, 'train/grad_norm_pre_clip': 11.865118980407715, 'timing/data': 0.02110495325177908, 'timing/model': 1.2815096620470285, 'learning_rate/qwen_vl_interface': 5.118092833385829e-06, 'learning_rate/action_model': 5.1180928333858293e-05, 'epoch': 1.39}) +[2026-06-18 00:36:27,951][starVLA.training.train_starvla][INFO] - Step 3612, Loss: {'train/loss': 0.26712894439697266, 'train/grad_norm_pre_clip': 8.347665786743164, 'timing/data': 0.022661134600639343, 'timing/model': 1.6850653095170856, 'learning_rate/qwen_vl_interface': 5.115930990598713e-06, 'learning_rate/action_model': 5.115930990598713e-05, 'epoch': 1.39}) +[2026-06-18 00:36:38,622][starVLA.training.train_starvla][INFO] - Step 3613, Loss: {'train/loss': 0.3692255914211273, 'train/grad_norm_pre_clip': 9.812623977661133, 'timing/data': 0.024767516180872917, 'timing/model': 4.810795716941357, 'learning_rate/qwen_vl_interface': 5.113769175604248e-06, 'learning_rate/action_model': 5.113769175604247e-05, 'epoch': 1.39}) +[2026-06-18 00:36:43,658][starVLA.training.train_starvla][INFO] - Step 3614, Loss: {'train/loss': 0.47754785418510437, 'train/grad_norm_pre_clip': 19.105186462402344, 'timing/data': 0.023718449287116528, 'timing/model': 1.18727898132056, 'learning_rate/qwen_vl_interface': 5.111607388850576e-06, 'learning_rate/action_model': 5.111607388850576e-05, 'epoch': 1.39}) +[2026-06-18 00:36:48,580][starVLA.training.train_starvla][INFO] - Step 3615, Loss: {'train/loss': 0.2937747538089752, 'train/grad_norm_pre_clip': 12.079573631286621, 'timing/data': 0.032450142316520214, 'timing/model': 1.1661841608583927, 'learning_rate/qwen_vl_interface': 5.1094456307858444e-06, 'learning_rate/action_model': 5.109445630785844e-05, 'epoch': 1.39}) +[2026-06-18 00:36:58,953][starVLA.training.train_starvla][INFO] - Step 3616, Loss: {'train/loss': 0.6765977740287781, 'train/grad_norm_pre_clip': 16.6650333404541, 'timing/data': 0.02043972536921501, 'timing/model': 1.0878105415031314, 'learning_rate/qwen_vl_interface': 5.107283901858184e-06, 'learning_rate/action_model': 5.107283901858184e-05, 'epoch': 1.39}) +[2026-06-18 00:37:04,098][starVLA.training.train_starvla][INFO] - Step 3617, Loss: {'train/loss': 0.36918288469314575, 'train/grad_norm_pre_clip': 8.715065956115723, 'timing/data': 0.02374958246946335, 'timing/model': 1.4049879303202033, 'learning_rate/qwen_vl_interface': 5.105122202515726e-06, 'learning_rate/action_model': 5.1051222025157255e-05, 'epoch': 1.39}) +[2026-06-18 00:37:08,845][starVLA.training.train_starvla][INFO] - Step 3618, Loss: {'train/loss': 0.29938796162605286, 'train/grad_norm_pre_clip': 9.384344100952148, 'timing/data': 0.027829796075820923, 'timing/model': 1.095548409037292, 'learning_rate/qwen_vl_interface': 5.102960533206588e-06, 'learning_rate/action_model': 5.1029605332065874e-05, 'epoch': 1.39}) +[2026-06-18 00:37:14,023][starVLA.training.train_starvla][INFO] - Step 3619, Loss: {'train/loss': 0.3589768409729004, 'train/grad_norm_pre_clip': 10.892329216003418, 'timing/data': 0.02067372016608715, 'timing/model': 1.0867565674707294, 'learning_rate/qwen_vl_interface': 5.100798894378891e-06, 'learning_rate/action_model': 5.100798894378891e-05, 'epoch': 1.39}) +[2026-06-18 00:37:18,841][starVLA.training.train_starvla][INFO] - Step 3620, Loss: {'train/loss': 0.5487226843833923, 'train/grad_norm_pre_clip': 20.677309036254883, 'timing/data': 0.022689087316393852, 'timing/model': 1.0898343874141574, 'learning_rate/qwen_vl_interface': 5.098637286480744e-06, 'learning_rate/action_model': 5.098637286480744e-05, 'epoch': 1.39}) +[2026-06-18 00:37:25,517][starVLA.training.train_starvla][INFO] - Step 3621, Loss: {'train/loss': 0.2991863489151001, 'train/grad_norm_pre_clip': 10.581307411193848, 'timing/data': 0.0239607784897089, 'timing/model': 1.7539414800703526, 'learning_rate/qwen_vl_interface': 5.096475709960252e-06, 'learning_rate/action_model': 5.096475709960251e-05, 'epoch': 1.39}) +[2026-06-18 00:37:31,031][starVLA.training.train_starvla][INFO] - Step 3622, Loss: {'train/loss': 0.5032600164413452, 'train/grad_norm_pre_clip': 9.339274406433105, 'timing/data': 0.028656047768890858, 'timing/model': 1.485770171508193, 'learning_rate/qwen_vl_interface': 5.094314165265506e-06, 'learning_rate/action_model': 5.0943141652655056e-05, 'epoch': 1.39}) +[2026-06-18 00:37:35,693][starVLA.training.train_starvla][INFO] - Step 3623, Loss: {'train/loss': 0.178735613822937, 'train/grad_norm_pre_clip': 7.901559352874756, 'timing/data': 0.02162223495543003, 'timing/model': 1.1537069138139486, 'learning_rate/qwen_vl_interface': 5.092152652844606e-06, 'learning_rate/action_model': 5.092152652844606e-05, 'epoch': 1.39}) +[2026-06-18 00:37:42,592][starVLA.training.train_starvla][INFO] - Step 3624, Loss: {'train/loss': 0.26965463161468506, 'train/grad_norm_pre_clip': 7.032253265380859, 'timing/data': 0.04672395996749401, 'timing/model': 1.1383450329303741, 'learning_rate/qwen_vl_interface': 5.089991173145628e-06, 'learning_rate/action_model': 5.0899911731456285e-05, 'epoch': 1.39}) +[2026-06-18 00:37:49,113][starVLA.training.train_starvla][INFO] - Step 3625, Loss: {'train/loss': 0.3875177502632141, 'train/grad_norm_pre_clip': 12.400089263916016, 'timing/data': 0.023574592545628548, 'timing/model': 2.630223554559052, 'learning_rate/qwen_vl_interface': 5.087829726616653e-06, 'learning_rate/action_model': 5.087829726616653e-05, 'epoch': 1.39}) +[2026-06-18 00:37:53,968][starVLA.training.train_starvla][INFO] - Step 3626, Loss: {'train/loss': 0.47783827781677246, 'train/grad_norm_pre_clip': 11.708734512329102, 'timing/data': 0.022784700617194176, 'timing/model': 1.1615070728585124, 'learning_rate/qwen_vl_interface': 5.08566831370575e-06, 'learning_rate/action_model': 5.0856683137057504e-05, 'epoch': 1.39}) +[2026-06-18 00:37:58,994][starVLA.training.train_starvla][INFO] - Step 3627, Loss: {'train/loss': 0.3585266172885895, 'train/grad_norm_pre_clip': 11.229393005371094, 'timing/data': 0.022659001871943474, 'timing/model': 1.6346209719777107, 'learning_rate/qwen_vl_interface': 5.083506934860982e-06, 'learning_rate/action_model': 5.083506934860982e-05, 'epoch': 1.4}) +[2026-06-18 00:38:05,309][starVLA.training.train_starvla][INFO] - Step 3628, Loss: {'train/loss': 0.5830872058868408, 'train/grad_norm_pre_clip': 12.301748275756836, 'timing/data': 0.030276351608335972, 'timing/model': 1.171406309120357, 'learning_rate/qwen_vl_interface': 5.081345590530407e-06, 'learning_rate/action_model': 5.081345590530407e-05, 'epoch': 1.4}) +[2026-06-18 00:38:10,599][starVLA.training.train_starvla][INFO] - Step 3629, Loss: {'train/loss': 0.6554091572761536, 'train/grad_norm_pre_clip': 16.478532791137695, 'timing/data': 0.01996577437967062, 'timing/model': 1.5871275970712304, 'learning_rate/qwen_vl_interface': 5.079184281162069e-06, 'learning_rate/action_model': 5.0791842811620694e-05, 'epoch': 1.4}) +[2026-06-18 00:38:18,132][starVLA.training.train_starvla][INFO] - Step 3630, Loss: {'train/loss': 0.16717909276485443, 'train/grad_norm_pre_clip': 5.715126037597656, 'timing/data': 0.03983337618410587, 'timing/model': 1.1918323496356606, 'learning_rate/qwen_vl_interface': 5.077023007204018e-06, 'learning_rate/action_model': 5.077023007204018e-05, 'epoch': 1.4}) +[2026-06-18 00:38:22,798][starVLA.training.train_starvla][INFO] - Step 3631, Loss: {'train/loss': 0.4345155358314514, 'train/grad_norm_pre_clip': 15.413146018981934, 'timing/data': 0.03458932787179947, 'timing/model': 1.1097563561052084, 'learning_rate/qwen_vl_interface': 5.074861769104281e-06, 'learning_rate/action_model': 5.074861769104281e-05, 'epoch': 1.4}) +[2026-06-18 00:38:29,604][starVLA.training.train_starvla][INFO] - Step 3632, Loss: {'train/loss': 0.5125821828842163, 'train/grad_norm_pre_clip': 14.537580490112305, 'timing/data': 0.025408752262592316, 'timing/model': 1.6709843706339598, 'learning_rate/qwen_vl_interface': 5.07270056731089e-06, 'learning_rate/action_model': 5.07270056731089e-05, 'epoch': 1.4}) +[2026-06-18 00:38:37,957][starVLA.training.train_starvla][INFO] - Step 3633, Loss: {'train/loss': 0.38950711488723755, 'train/grad_norm_pre_clip': 9.300280570983887, 'timing/data': 0.020350908860564232, 'timing/model': 4.68076763022691, 'learning_rate/qwen_vl_interface': 5.070539402271859e-06, 'learning_rate/action_model': 5.070539402271859e-05, 'epoch': 1.4}) +[2026-06-18 00:38:50,780][starVLA.training.train_starvla][INFO] - Step 3634, Loss: {'train/loss': 0.192440927028656, 'train/grad_norm_pre_clip': 6.976321697235107, 'timing/data': 0.02394089847803116, 'timing/model': 1.2699482049793005, 'learning_rate/qwen_vl_interface': 5.068378274435205e-06, 'learning_rate/action_model': 5.068378274435205e-05, 'epoch': 1.4}) +[2026-06-18 00:39:02,794][starVLA.training.train_starvla][INFO] - Step 3635, Loss: {'train/loss': 0.33037465810775757, 'train/grad_norm_pre_clip': 16.689350128173828, 'timing/data': 0.024832445196807384, 'timing/model': 1.0716561088338494, 'learning_rate/qwen_vl_interface': 5.066217184248931e-06, 'learning_rate/action_model': 5.0662171842489305e-05, 'epoch': 1.4}) +[2026-06-18 00:39:14,864][starVLA.training.train_starvla][INFO] - Step 3636, Loss: {'train/loss': 0.5124053359031677, 'train/grad_norm_pre_clip': 13.852029800415039, 'timing/data': 0.0349025409668684, 'timing/model': 1.530903184786439, 'learning_rate/qwen_vl_interface': 5.0640561321610295e-06, 'learning_rate/action_model': 5.064056132161029e-05, 'epoch': 1.4}) +[2026-06-18 00:39:19,966][starVLA.training.train_starvla][INFO] - Step 3637, Loss: {'train/loss': 0.4196528494358063, 'train/grad_norm_pre_clip': 11.292145729064941, 'timing/data': 0.03294725343585014, 'timing/model': 1.4538829466328025, 'learning_rate/qwen_vl_interface': 5.061895118619493e-06, 'learning_rate/action_model': 5.061895118619493e-05, 'epoch': 1.4}) +[2026-06-18 00:39:24,969][starVLA.training.train_starvla][INFO] - Step 3638, Loss: {'train/loss': 0.4809492230415344, 'train/grad_norm_pre_clip': 10.707240104675293, 'timing/data': 0.0329012842848897, 'timing/model': 1.370129399932921, 'learning_rate/qwen_vl_interface': 5.059734144072302e-06, 'learning_rate/action_model': 5.059734144072302e-05, 'epoch': 1.4}) +[2026-06-18 00:39:29,849][starVLA.training.train_starvla][INFO] - Step 3639, Loss: {'train/loss': 0.4622013568878174, 'train/grad_norm_pre_clip': 14.899187088012695, 'timing/data': 0.02270493283867836, 'timing/model': 1.1496912706643343, 'learning_rate/qwen_vl_interface': 5.0575732089674275e-06, 'learning_rate/action_model': 5.057573208967428e-05, 'epoch': 1.4}) +[2026-06-18 00:39:35,253][starVLA.training.train_starvla][INFO] - Step 3640, Loss: {'train/loss': 0.6170399188995361, 'train/grad_norm_pre_clip': 10.759807586669922, 'timing/data': 0.022781986743211746, 'timing/model': 1.2566090002655983, 'learning_rate/qwen_vl_interface': 5.055412313752832e-06, 'learning_rate/action_model': 5.055412313752832e-05, 'epoch': 1.4}) +[2026-06-18 00:39:40,786][starVLA.training.train_starvla][INFO] - Step 3641, Loss: {'train/loss': 0.259414404630661, 'train/grad_norm_pre_clip': 9.771482467651367, 'timing/data': 0.025910289958119392, 'timing/model': 1.2217842619866133, 'learning_rate/qwen_vl_interface': 5.053251458876474e-06, 'learning_rate/action_model': 5.0532514588764744e-05, 'epoch': 1.4}) +[2026-06-18 00:39:45,616][starVLA.training.train_starvla][INFO] - Step 3642, Loss: {'train/loss': 0.3241030275821686, 'train/grad_norm_pre_clip': 8.002262115478516, 'timing/data': 0.022046033293008804, 'timing/model': 1.5413760943338275, 'learning_rate/qwen_vl_interface': 5.0510906447863025e-06, 'learning_rate/action_model': 5.051090644786303e-05, 'epoch': 1.4}) +[2026-06-18 00:39:50,117][starVLA.training.train_starvla][INFO] - Step 3643, Loss: {'train/loss': 0.512342631816864, 'train/grad_norm_pre_clip': 13.669285774230957, 'timing/data': 0.02246223483234644, 'timing/model': 1.3355083223432302, 'learning_rate/qwen_vl_interface': 5.048929871930252e-06, 'learning_rate/action_model': 5.048929871930251e-05, 'epoch': 1.4}) +[2026-06-18 00:39:55,149][starVLA.training.train_starvla][INFO] - Step 3644, Loss: {'train/loss': 0.3494572043418884, 'train/grad_norm_pre_clip': 13.130217552185059, 'timing/data': 0.02210313268005848, 'timing/model': 1.156259992159903, 'learning_rate/qwen_vl_interface': 5.046769140756257e-06, 'learning_rate/action_model': 5.046769140756257e-05, 'epoch': 1.4}) +[2026-06-18 00:40:00,492][starVLA.training.train_starvla][INFO] - Step 3645, Loss: {'train/loss': 0.5990970730781555, 'train/grad_norm_pre_clip': 12.728103637695312, 'timing/data': 0.025145653635263443, 'timing/model': 1.160334694199264, 'learning_rate/qwen_vl_interface': 5.044608451712237e-06, 'learning_rate/action_model': 5.044608451712237e-05, 'epoch': 1.4}) +[2026-06-18 00:40:05,177][starVLA.training.train_starvla][INFO] - Step 3646, Loss: {'train/loss': 0.3080756962299347, 'train/grad_norm_pre_clip': 8.671110153198242, 'timing/data': 0.02626824378967285, 'timing/model': 1.1561892656609416, 'learning_rate/qwen_vl_interface': 5.042447805246105e-06, 'learning_rate/action_model': 5.042447805246106e-05, 'epoch': 1.4}) +[2026-06-18 00:40:10,393][starVLA.training.train_starvla][INFO] - Step 3647, Loss: {'train/loss': 0.4393009543418884, 'train/grad_norm_pre_clip': 14.397095680236816, 'timing/data': 0.027229919098317623, 'timing/model': 1.6997106205672026, 'learning_rate/qwen_vl_interface': 5.040287201805767e-06, 'learning_rate/action_model': 5.040287201805767e-05, 'epoch': 1.4}) +[2026-06-18 00:40:14,992][starVLA.training.train_starvla][INFO] - Step 3648, Loss: {'train/loss': 0.27794235944747925, 'train/grad_norm_pre_clip': 9.88562297821045, 'timing/data': 0.020429156720638275, 'timing/model': 1.2128545036539435, 'learning_rate/qwen_vl_interface': 5.038126641839119e-06, 'learning_rate/action_model': 5.0381266418391184e-05, 'epoch': 1.4}) +[2026-06-18 00:40:20,293][starVLA.training.train_starvla][INFO] - Step 3649, Loss: {'train/loss': 0.28898486495018005, 'train/grad_norm_pre_clip': 10.059686660766602, 'timing/data': 0.03598161321133375, 'timing/model': 1.166734328493476, 'learning_rate/qwen_vl_interface': 5.0359661257940455e-06, 'learning_rate/action_model': 5.035966125794045e-05, 'epoch': 1.4}) +[2026-06-18 00:40:24,822][starVLA.training.train_starvla][INFO] - Step 3650, Loss: {'train/loss': 0.4600444734096527, 'train/grad_norm_pre_clip': 12.153020858764648, 'timing/data': 0.022889665327966213, 'timing/model': 1.0918273869901896, 'learning_rate/qwen_vl_interface': 5.033805654118424e-06, 'learning_rate/action_model': 5.033805654118424e-05, 'epoch': 1.4}) +[2026-06-18 00:40:29,433][starVLA.training.train_starvla][INFO] - Step 3651, Loss: {'train/loss': 0.21404576301574707, 'train/grad_norm_pre_clip': 7.601629257202148, 'timing/data': 0.023586035706102848, 'timing/model': 1.082080558873713, 'learning_rate/qwen_vl_interface': 5.031645227260122e-06, 'learning_rate/action_model': 5.0316452272601224e-05, 'epoch': 1.4}) +[2026-06-18 00:40:34,396][starVLA.training.train_starvla][INFO] - Step 3652, Loss: {'train/loss': 0.37057024240493774, 'train/grad_norm_pre_clip': 12.98060417175293, 'timing/data': 0.020740492269396782, 'timing/model': 1.6235009115189314, 'learning_rate/qwen_vl_interface': 5.029484845667001e-06, 'learning_rate/action_model': 5.029484845667001e-05, 'epoch': 1.4}) +[2026-06-18 00:40:49,494][starVLA.training.train_starvla][INFO] - Step 3653, Loss: {'train/loss': 0.17716634273529053, 'train/grad_norm_pre_clip': 5.946237564086914, 'timing/data': 0.026881265453994274, 'timing/model': 5.5816784938797355, 'learning_rate/qwen_vl_interface': 5.027324509786908e-06, 'learning_rate/action_model': 5.027324509786908e-05, 'epoch': 1.41}) +[2026-06-18 00:40:56,979][starVLA.training.train_starvla][INFO] - Step 3654, Loss: {'train/loss': 0.25921058654785156, 'train/grad_norm_pre_clip': 8.339254379272461, 'timing/data': 0.02105687838047743, 'timing/model': 3.592462331056595, 'learning_rate/qwen_vl_interface': 5.025164220067683e-06, 'learning_rate/action_model': 5.025164220067683e-05, 'epoch': 1.41}) +[2026-06-18 00:41:03,337][starVLA.training.train_starvla][INFO] - Step 3655, Loss: {'train/loss': 0.4806520938873291, 'train/grad_norm_pre_clip': 13.501494407653809, 'timing/data': 0.023401089943945408, 'timing/model': 1.3778095347806811, 'learning_rate/qwen_vl_interface': 5.023003976957159e-06, 'learning_rate/action_model': 5.023003976957159e-05, 'epoch': 1.41}) +[2026-06-18 00:41:19,015][starVLA.training.train_starvla][INFO] - Step 3656, Loss: {'train/loss': 0.3449048697948456, 'train/grad_norm_pre_clip': 12.295695304870605, 'timing/data': 0.029551398009061813, 'timing/model': 1.5834797201678157, 'learning_rate/qwen_vl_interface': 5.020843780903155e-06, 'learning_rate/action_model': 5.020843780903155e-05, 'epoch': 1.41}) +[2026-06-18 00:41:30,617][starVLA.training.train_starvla][INFO] - Step 3657, Loss: {'train/loss': 0.6080557703971863, 'train/grad_norm_pre_clip': 13.164923667907715, 'timing/data': 0.033455222845077515, 'timing/model': 1.7971656769514084, 'learning_rate/qwen_vl_interface': 5.01868363235348e-06, 'learning_rate/action_model': 5.01868363235348e-05, 'epoch': 1.41}) +[2026-06-18 00:41:38,179][starVLA.training.train_starvla][INFO] - Step 3658, Loss: {'train/loss': 0.8022702932357788, 'train/grad_norm_pre_clip': 16.768159866333008, 'timing/data': 0.027400191873311996, 'timing/model': 1.3235297836363316, 'learning_rate/qwen_vl_interface': 5.016523531755937e-06, 'learning_rate/action_model': 5.016523531755937e-05, 'epoch': 1.41}) +[2026-06-18 00:41:45,024][starVLA.training.train_starvla][INFO] - Step 3659, Loss: {'train/loss': 0.3992916941642761, 'train/grad_norm_pre_clip': 11.043646812438965, 'timing/data': 0.024325236678123474, 'timing/model': 1.2786859581246972, 'learning_rate/qwen_vl_interface': 5.014363479558318e-06, 'learning_rate/action_model': 5.014363479558318e-05, 'epoch': 1.41}) +[2026-06-18 00:41:57,377][starVLA.training.train_starvla][INFO] - Step 3660, Loss: {'train/loss': 0.478146493434906, 'train/grad_norm_pre_clip': 10.981801986694336, 'timing/data': 0.022249377332627773, 'timing/model': 1.40975162666291, 'learning_rate/qwen_vl_interface': 5.012203476208403e-06, 'learning_rate/action_model': 5.0122034762084025e-05, 'epoch': 1.41}) +[2026-06-18 00:42:02,755][starVLA.training.train_starvla][INFO] - Step 3661, Loss: {'train/loss': 0.5059309601783752, 'train/grad_norm_pre_clip': 13.997342109680176, 'timing/data': 0.023808584548532963, 'timing/model': 1.0931090228259563, 'learning_rate/qwen_vl_interface': 5.010043522153965e-06, 'learning_rate/action_model': 5.0100435221539646e-05, 'epoch': 1.41}) +[2026-06-18 00:42:07,565][starVLA.training.train_starvla][INFO] - Step 3662, Loss: {'train/loss': 0.6700608134269714, 'train/grad_norm_pre_clip': 12.481877326965332, 'timing/data': 0.023419243283569813, 'timing/model': 1.1561218844726682, 'learning_rate/qwen_vl_interface': 5.007883617842759e-06, 'learning_rate/action_model': 5.007883617842759e-05, 'epoch': 1.41}) +[2026-06-18 00:42:12,830][starVLA.training.train_starvla][INFO] - Step 3663, Loss: {'train/loss': 0.368918240070343, 'train/grad_norm_pre_clip': 13.123544692993164, 'timing/data': 0.023977561853826046, 'timing/model': 1.089949962683022, 'learning_rate/qwen_vl_interface': 5.005723763722543e-06, 'learning_rate/action_model': 5.005723763722543e-05, 'epoch': 1.41}) +[2026-06-18 00:42:17,642][starVLA.training.train_starvla][INFO] - Step 3664, Loss: {'train/loss': 0.4934234619140625, 'train/grad_norm_pre_clip': 10.986016273498535, 'timing/data': 0.02323973085731268, 'timing/model': 1.1503836046904325, 'learning_rate/qwen_vl_interface': 5.003563960241052e-06, 'learning_rate/action_model': 5.003563960241052e-05, 'epoch': 1.41}) +[2026-06-18 00:42:25,759][starVLA.training.train_starvla][INFO] - Step 3665, Loss: {'train/loss': 0.31859639286994934, 'train/grad_norm_pre_clip': 10.912275314331055, 'timing/data': 0.0288704801350832, 'timing/model': 4.166488444432616, 'learning_rate/qwen_vl_interface': 5.001404207846017e-06, 'learning_rate/action_model': 5.001404207846017e-05, 'epoch': 1.41}) +[2026-06-18 00:42:30,948][starVLA.training.train_starvla][INFO] - Step 3666, Loss: {'train/loss': 0.33470529317855835, 'train/grad_norm_pre_clip': 9.603468894958496, 'timing/data': 0.025844847783446312, 'timing/model': 1.6088939877226949, 'learning_rate/qwen_vl_interface': 4.999244506985158e-06, 'learning_rate/action_model': 4.999244506985158e-05, 'epoch': 1.41}) +[2026-06-18 00:42:35,557][starVLA.training.train_starvla][INFO] - Step 3667, Loss: {'train/loss': 0.13499681651592255, 'train/grad_norm_pre_clip': 5.466353416442871, 'timing/data': 0.021961047314107418, 'timing/model': 1.4299490954726934, 'learning_rate/qwen_vl_interface': 4.997084858106182e-06, 'learning_rate/action_model': 4.997084858106183e-05, 'epoch': 1.41}) +[2026-06-18 00:42:39,936][starVLA.training.train_starvla][INFO] - Step 3668, Loss: {'train/loss': 0.409204363822937, 'train/grad_norm_pre_clip': 14.177724838256836, 'timing/data': 0.04836582578718662, 'timing/model': 1.1051428206264973, 'learning_rate/qwen_vl_interface': 4.994925261656789e-06, 'learning_rate/action_model': 4.994925261656789e-05, 'epoch': 1.41}) +[2026-06-18 00:42:44,611][starVLA.training.train_starvla][INFO] - Step 3669, Loss: {'train/loss': 0.20862577855587006, 'train/grad_norm_pre_clip': 11.166934967041016, 'timing/data': 0.03480075113475323, 'timing/model': 1.1669466951861978, 'learning_rate/qwen_vl_interface': 4.99276571808466e-06, 'learning_rate/action_model': 4.9927657180846605e-05, 'epoch': 1.41}) +[2026-06-18 00:42:49,255][starVLA.training.train_starvla][INFO] - Step 3670, Loss: {'train/loss': 0.4727862477302551, 'train/grad_norm_pre_clip': 9.86652660369873, 'timing/data': 0.019692150875926018, 'timing/model': 1.2170070400461555, 'learning_rate/qwen_vl_interface': 4.9906062278374775e-06, 'learning_rate/action_model': 4.990606227837478e-05, 'epoch': 1.41}) +[2026-06-18 00:42:56,461][starVLA.training.train_starvla][INFO] - Step 3671, Loss: {'train/loss': 0.3262627422809601, 'train/grad_norm_pre_clip': 13.290033340454102, 'timing/data': 0.0289045637473464, 'timing/model': 1.0892782993614674, 'learning_rate/qwen_vl_interface': 4.988446791362903e-06, 'learning_rate/action_model': 4.988446791362903e-05, 'epoch': 1.41}) +[2026-06-18 00:43:03,994][starVLA.training.train_starvla][INFO] - Step 3672, Loss: {'train/loss': 0.49020445346832275, 'train/grad_norm_pre_clip': 14.591947555541992, 'timing/data': 0.023972846567630768, 'timing/model': 2.8312484407797456, 'learning_rate/qwen_vl_interface': 4.986287409108589e-06, 'learning_rate/action_model': 4.9862874091085893e-05, 'epoch': 1.41}) +[2026-06-18 00:43:08,919][starVLA.training.train_starvla][INFO] - Step 3673, Loss: {'train/loss': 0.6181233525276184, 'train/grad_norm_pre_clip': 12.650522232055664, 'timing/data': 0.023226206190884113, 'timing/model': 1.3300232077017426, 'learning_rate/qwen_vl_interface': 4.984128081522179e-06, 'learning_rate/action_model': 4.984128081522178e-05, 'epoch': 1.41}) +[2026-06-18 00:43:13,908][starVLA.training.train_starvla][INFO] - Step 3674, Loss: {'train/loss': 0.4235835075378418, 'train/grad_norm_pre_clip': 16.123321533203125, 'timing/data': 0.025580802001059055, 'timing/model': 1.1670045126229525, 'learning_rate/qwen_vl_interface': 4.981968809051304e-06, 'learning_rate/action_model': 4.981968809051304e-05, 'epoch': 1.41}) +[2026-06-18 00:43:18,851][starVLA.training.train_starvla][INFO] - Step 3675, Loss: {'train/loss': 0.44865307211875916, 'train/grad_norm_pre_clip': 11.002172470092773, 'timing/data': 0.021388008259236813, 'timing/model': 1.1469934340566397, 'learning_rate/qwen_vl_interface': 4.979809592143583e-06, 'learning_rate/action_model': 4.979809592143583e-05, 'epoch': 1.41}) +[2026-06-18 00:43:24,135][starVLA.training.train_starvla][INFO] - Step 3676, Loss: {'train/loss': 0.592376708984375, 'train/grad_norm_pre_clip': 12.869476318359375, 'timing/data': 0.022172358818352222, 'timing/model': 1.2343924595043063, 'learning_rate/qwen_vl_interface': 4.977650431246624e-06, 'learning_rate/action_model': 4.977650431246624e-05, 'epoch': 1.41}) +[2026-06-18 00:43:29,283][starVLA.training.train_starvla][INFO] - Step 3677, Loss: {'train/loss': 0.3027574419975281, 'train/grad_norm_pre_clip': 12.063031196594238, 'timing/data': 0.03464125283062458, 'timing/model': 1.70239251665771, 'learning_rate/qwen_vl_interface': 4.9754913268080255e-06, 'learning_rate/action_model': 4.975491326808025e-05, 'epoch': 1.41}) +[2026-06-18 00:43:34,442][starVLA.training.train_starvla][INFO] - Step 3678, Loss: {'train/loss': 0.43235841393470764, 'train/grad_norm_pre_clip': 13.407444953918457, 'timing/data': 0.024623152799904346, 'timing/model': 1.4428440872579813, 'learning_rate/qwen_vl_interface': 4.973332279275369e-06, 'learning_rate/action_model': 4.973332279275369e-05, 'epoch': 1.41}) +[2026-06-18 00:43:41,048][starVLA.training.train_starvla][INFO] - Step 3679, Loss: {'train/loss': 0.3689332604408264, 'train/grad_norm_pre_clip': 9.156062126159668, 'timing/data': 0.041388047859072685, 'timing/model': 1.1167567381635308, 'learning_rate/qwen_vl_interface': 4.971173289096229e-06, 'learning_rate/action_model': 4.9711732890962286e-05, 'epoch': 1.42}) +[2026-06-18 00:43:46,488][starVLA.training.train_starvla][INFO] - Step 3680, Loss: {'train/loss': 0.4739770293235779, 'train/grad_norm_pre_clip': 10.546342849731445, 'timing/data': 0.023510168306529522, 'timing/model': 1.2907895725220442, 'learning_rate/qwen_vl_interface': 4.969014356718165e-06, 'learning_rate/action_model': 4.969014356718165e-05, 'epoch': 1.42}) +[2026-06-18 00:44:02,243][starVLA.training.train_starvla][INFO] - Step 3681, Loss: {'train/loss': 0.4783839285373688, 'train/grad_norm_pre_clip': 9.300311088562012, 'timing/data': 0.02214016206562519, 'timing/model': 1.7740754494443536, 'learning_rate/qwen_vl_interface': 4.9668554825887285e-06, 'learning_rate/action_model': 4.966855482588728e-05, 'epoch': 1.42}) +[2026-06-18 00:44:08,065][starVLA.training.train_starvla][INFO] - Step 3682, Loss: {'train/loss': 0.316261887550354, 'train/grad_norm_pre_clip': 12.17275333404541, 'timing/data': 0.041556316427886486, 'timing/model': 1.5307238958775997, 'learning_rate/qwen_vl_interface': 4.964696667155455e-06, 'learning_rate/action_model': 4.964696667155454e-05, 'epoch': 1.42}) +[2026-06-18 00:44:13,212][starVLA.training.train_starvla][INFO] - Step 3683, Loss: {'train/loss': 0.7456681728363037, 'train/grad_norm_pre_clip': 16.872817993164062, 'timing/data': 0.026655863039195538, 'timing/model': 1.2395190466195345, 'learning_rate/qwen_vl_interface': 4.96253791086587e-06, 'learning_rate/action_model': 4.96253791086587e-05, 'epoch': 1.42}) +[2026-06-18 00:44:19,675][starVLA.training.train_starvla][INFO] - Step 3684, Loss: {'train/loss': 0.349295049905777, 'train/grad_norm_pre_clip': 10.046507835388184, 'timing/data': 0.023674857802689075, 'timing/model': 1.1218506190925837, 'learning_rate/qwen_vl_interface': 4.960379214167483e-06, 'learning_rate/action_model': 4.960379214167483e-05, 'epoch': 1.42}) +[2026-06-18 00:44:28,211][starVLA.training.train_starvla][INFO] - Step 3685, Loss: {'train/loss': 0.355410099029541, 'train/grad_norm_pre_clip': 9.502914428710938, 'timing/data': 0.02794519905000925, 'timing/model': 1.3206534665077925, 'learning_rate/qwen_vl_interface': 4.9582205775077994e-06, 'learning_rate/action_model': 4.958220577507799e-05, 'epoch': 1.42}) +[2026-06-18 00:44:33,366][starVLA.training.train_starvla][INFO] - Step 3686, Loss: {'train/loss': 0.2815203368663788, 'train/grad_norm_pre_clip': 11.032691955566406, 'timing/data': 0.027576320804655552, 'timing/model': 1.465240209363401, 'learning_rate/qwen_vl_interface': 4.956062001334302e-06, 'learning_rate/action_model': 4.956062001334302e-05, 'epoch': 1.42}) +[2026-06-18 00:44:39,294][starVLA.training.train_starvla][INFO] - Step 3687, Loss: {'train/loss': 0.16990047693252563, 'train/grad_norm_pre_clip': 5.485957622528076, 'timing/data': 0.02969358768314123, 'timing/model': 1.6120839798823, 'learning_rate/qwen_vl_interface': 4.953903486094469e-06, 'learning_rate/action_model': 4.953903486094469e-05, 'epoch': 1.42}) +[2026-06-18 00:44:44,609][starVLA.training.train_starvla][INFO] - Step 3688, Loss: {'train/loss': 0.34418338537216187, 'train/grad_norm_pre_clip': 8.1696138381958, 'timing/data': 0.031161190010607243, 'timing/model': 1.1399222640320659, 'learning_rate/qwen_vl_interface': 4.951745032235762e-06, 'learning_rate/action_model': 4.951745032235762e-05, 'epoch': 1.42}) +[2026-06-18 00:44:52,057][starVLA.training.train_starvla][INFO] - Step 3689, Loss: {'train/loss': 0.47313347458839417, 'train/grad_norm_pre_clip': 10.023859024047852, 'timing/data': 0.033395701088011265, 'timing/model': 1.242077276110649, 'learning_rate/qwen_vl_interface': 4.949586640205631e-06, 'learning_rate/action_model': 4.94958664020563e-05, 'epoch': 1.42}) +[2026-06-18 00:44:57,155][starVLA.training.train_starvla][INFO] - Step 3690, Loss: {'train/loss': 0.3375394642353058, 'train/grad_norm_pre_clip': 11.420547485351562, 'timing/data': 0.03033742494881153, 'timing/model': 1.2210134388878942, 'learning_rate/qwen_vl_interface': 4.9474283104515124e-06, 'learning_rate/action_model': 4.9474283104515126e-05, 'epoch': 1.42}) +[2026-06-18 00:45:02,242][starVLA.training.train_starvla][INFO] - Step 3691, Loss: {'train/loss': 0.45846036076545715, 'train/grad_norm_pre_clip': 13.979775428771973, 'timing/data': 0.021264481358230114, 'timing/model': 1.24938303232193, 'learning_rate/qwen_vl_interface': 4.945270043420829e-06, 'learning_rate/action_model': 4.945270043420828e-05, 'epoch': 1.42}) +[2026-06-18 00:45:08,819][starVLA.training.train_starvla][INFO] - Step 3692, Loss: {'train/loss': 0.4119194447994232, 'train/grad_norm_pre_clip': 11.322985649108887, 'timing/data': 0.03174196556210518, 'timing/model': 1.6695029027760029, 'learning_rate/qwen_vl_interface': 4.9431118395609954e-06, 'learning_rate/action_model': 4.9431118395609954e-05, 'epoch': 1.42}) +[2026-06-18 00:45:14,354][starVLA.training.train_starvla][INFO] - Step 3693, Loss: {'train/loss': 0.40649116039276123, 'train/grad_norm_pre_clip': 12.607028007507324, 'timing/data': 0.02333206497132778, 'timing/model': 1.880426318384707, 'learning_rate/qwen_vl_interface': 4.940953699319407e-06, 'learning_rate/action_model': 4.940953699319407e-05, 'epoch': 1.42}) +[2026-06-18 00:45:19,134][starVLA.training.train_starvla][INFO] - Step 3694, Loss: {'train/loss': 0.6167232394218445, 'train/grad_norm_pre_clip': 13.463093757629395, 'timing/data': 0.02884556818753481, 'timing/model': 1.114804282784462, 'learning_rate/qwen_vl_interface': 4.93879562314345e-06, 'learning_rate/action_model': 4.93879562314345e-05, 'epoch': 1.42}) +[2026-06-18 00:45:24,496][starVLA.training.train_starvla][INFO] - Step 3695, Loss: {'train/loss': 0.6187198162078857, 'train/grad_norm_pre_clip': 23.398799896240234, 'timing/data': 0.031414592638611794, 'timing/model': 1.340773151256144, 'learning_rate/qwen_vl_interface': 4.936637611480495e-06, 'learning_rate/action_model': 4.9366376114804945e-05, 'epoch': 1.42}) +[2026-06-18 00:45:34,321][starVLA.training.train_starvla][INFO] - Step 3696, Loss: {'train/loss': 0.4880448281764984, 'train/grad_norm_pre_clip': 10.44579029083252, 'timing/data': 0.028740203008055687, 'timing/model': 1.3453064216300845, 'learning_rate/qwen_vl_interface': 4.9344796647779e-06, 'learning_rate/action_model': 4.9344796647778995e-05, 'epoch': 1.42}) +[2026-06-18 00:45:39,475][starVLA.training.train_starvla][INFO] - Step 3697, Loss: {'train/loss': 0.3452160358428955, 'train/grad_norm_pre_clip': 12.276403427124023, 'timing/data': 0.02302346844226122, 'timing/model': 1.6010221196338534, 'learning_rate/qwen_vl_interface': 4.9323217834830105e-06, 'learning_rate/action_model': 4.9323217834830105e-05, 'epoch': 1.42}) +[2026-06-18 00:45:44,036][starVLA.training.train_starvla][INFO] - Step 3698, Loss: {'train/loss': 0.3035401999950409, 'train/grad_norm_pre_clip': 9.084790229797363, 'timing/data': 0.031156282871961594, 'timing/model': 1.1208193469792604, 'learning_rate/qwen_vl_interface': 4.930163968043155e-06, 'learning_rate/action_model': 4.930163968043155e-05, 'epoch': 1.42}) +[2026-06-18 00:45:48,971][starVLA.training.train_starvla][INFO] - Step 3699, Loss: {'train/loss': 0.2600899636745453, 'train/grad_norm_pre_clip': 12.218753814697266, 'timing/data': 0.02535634394735098, 'timing/model': 1.1463759234175086, 'learning_rate/qwen_vl_interface': 4.928006218905656e-06, 'learning_rate/action_model': 4.928006218905656e-05, 'epoch': 1.42}) +[2026-06-18 00:49:57,949][starVLA.training.train_starvla][INFO] - Step 3700, Loss: {'train/loss': 0.2714749574661255, 'train/grad_norm_pre_clip': 10.803528785705566, 'eval/loss': 0.4627450108528137, 'eval/latency_4/loss': 0.4627450108528137, 'eval/demon_fire_precision': 0.9627388535031847, 'eval/demon_fire_recall': 0.9818122767132186, 'eval/demon_fire_f1': 0.972182022833253, 'eval/demon_fire_rate_model': 0.922173274596182, 'eval/demon_fire_rate_teacher': 0.9042584434654919, 'eval/demon_left_precision': 0.7791044776119403, 'eval/demon_left_recall': 0.5958904109589042, 'eval/demon_left_f1': 0.6752910737386805, 'eval/demon_left_rate_model': 0.09838472834067548, 'eval/demon_left_rate_teacher': 0.12863436123348018, 'eval/demon_right_precision': 0.8243801652892562, 'eval/demon_right_recall': 0.720216606498195, 'eval/demon_right_f1': 0.7687861271676302, 'eval/demon_right_rate_model': 0.14214390602055801, 'eval/demon_right_rate_teacher': 0.16270190895741557, 'eval/demon_move_f1': 0.7220386004531554, 'eval/demon_cc_f1': 0.8806537790200598, 'eval/latency_4/demon_fire_precision': 0.9627388535031847, 'eval/latency_4/demon_fire_recall': 0.9818122767132186, 'eval/latency_4/demon_fire_f1': 0.972182022833253, 'eval/latency_4/demon_fire_rate_model': 0.922173274596182, 'eval/latency_4/demon_fire_rate_teacher': 0.9042584434654919, 'eval/latency_4/demon_left_precision': 0.7791044776119403, 'eval/latency_4/demon_left_recall': 0.5958904109589042, 'eval/latency_4/demon_left_f1': 0.6752910737386805, 'eval/latency_4/demon_left_rate_model': 0.09838472834067548, 'eval/latency_4/demon_left_rate_teacher': 0.12863436123348018, 'eval/latency_4/demon_right_precision': 0.8243801652892562, 'eval/latency_4/demon_right_recall': 0.720216606498195, 'eval/latency_4/demon_right_f1': 0.7687861271676302, 'eval/latency_4/demon_right_rate_model': 0.14214390602055801, 'eval/latency_4/demon_right_rate_teacher': 0.16270190895741557, 'eval/latency_4/demon_move_f1': 0.7220386004531554, 'eval/latency_4/demon_cc_f1': 0.8806537790200598, 'timing/data': 0.03169080149382353, 'timing/model': 1.8362301271408796, 'learning_rate/qwen_vl_interface': 4.9258485365178135e-06, 'learning_rate/action_model': 4.925848536517814e-05, 'epoch': 1.42}) +[2026-06-18 00:49:57,951][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_3700_state +[2026-06-18 00:49:57,959][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 00:50:07,300][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_3700_state/model.safetensors +[2026-06-18 00:50:30,696][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_3700_state/optimizer.bin +[2026-06-18 00:50:30,699][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_3700_state/sampler.bin +[2026-06-18 00:50:30,699][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_3700_state/sampler_1.bin +[2026-06-18 00:50:30,702][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_3700_state/random_states_0.pkl +[2026-06-18 00:50:30,708][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3700 +[2026-06-18 00:50:30,708][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 00:50:30,726][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 00:50:38,578][starVLA.training.train_starvla][INFO] - Step 3701, Loss: {'train/loss': 0.4909254014492035, 'train/grad_norm_pre_clip': 12.27737045288086, 'timing/data': 0.0228505190461874, 'timing/model': 1.1389103950932622, 'learning_rate/qwen_vl_interface': 4.923690921326918e-06, 'learning_rate/action_model': 4.9236909213269175e-05, 'epoch': 1.42}) +[2026-06-18 00:50:52,685][starVLA.training.train_starvla][INFO] - Step 3702, Loss: {'train/loss': 0.36165761947631836, 'train/grad_norm_pre_clip': 8.144462585449219, 'timing/data': 0.02744651585817337, 'timing/model': 2.6386424880474806, 'learning_rate/qwen_vl_interface': 4.921533373780244e-06, 'learning_rate/action_model': 4.921533373780244e-05, 'epoch': 1.42}) +[2026-06-18 00:51:03,393][starVLA.training.train_starvla][INFO] - Step 3703, Loss: {'train/loss': 0.4259820878505707, 'train/grad_norm_pre_clip': 11.450549125671387, 'timing/data': 0.02705443650484085, 'timing/model': 5.897417404688895, 'learning_rate/qwen_vl_interface': 4.919375894325057e-06, 'learning_rate/action_model': 4.9193758943250566e-05, 'epoch': 1.42}) +[2026-06-18 00:51:08,361][starVLA.training.train_starvla][INFO] - Step 3704, Loss: {'train/loss': 0.11232829838991165, 'train/grad_norm_pre_clip': 4.606339454650879, 'timing/data': 0.02358418982475996, 'timing/model': 1.1302131731063128, 'learning_rate/qwen_vl_interface': 4.917218483408603e-06, 'learning_rate/action_model': 4.9172184834086035e-05, 'epoch': 1.42}) +[2026-06-18 00:51:13,592][starVLA.training.train_starvla][INFO] - Step 3705, Loss: {'train/loss': 0.4368627667427063, 'train/grad_norm_pre_clip': 11.964910507202148, 'timing/data': 0.03195774182677269, 'timing/model': 1.1633941316977143, 'learning_rate/qwen_vl_interface': 4.915061141478114e-06, 'learning_rate/action_model': 4.915061141478113e-05, 'epoch': 1.43}) +[2026-06-18 00:51:18,647][starVLA.training.train_starvla][INFO] - Step 3706, Loss: {'train/loss': 0.43285807967185974, 'train/grad_norm_pre_clip': 16.691221237182617, 'timing/data': 0.022291201166808605, 'timing/model': 1.1379763558506966, 'learning_rate/qwen_vl_interface': 4.912903868980812e-06, 'learning_rate/action_model': 4.9129038689808115e-05, 'epoch': 1.43}) +[2026-06-18 00:51:23,742][starVLA.training.train_starvla][INFO] - Step 3707, Loss: {'train/loss': 0.4918454885482788, 'train/grad_norm_pre_clip': 11.135516166687012, 'timing/data': 0.029942290857434273, 'timing/model': 1.423565236851573, 'learning_rate/qwen_vl_interface': 4.9107466663639005e-06, 'learning_rate/action_model': 4.910746666363901e-05, 'epoch': 1.43}) +[2026-06-18 00:51:34,342][starVLA.training.train_starvla][INFO] - Step 3708, Loss: {'train/loss': 0.3544550836086273, 'train/grad_norm_pre_clip': 13.573474884033203, 'timing/data': 0.021527634002268314, 'timing/model': 4.737280189990997, 'learning_rate/qwen_vl_interface': 4.9085895340745705e-06, 'learning_rate/action_model': 4.9085895340745705e-05, 'epoch': 1.43}) +[2026-06-18 00:51:42,954][starVLA.training.train_starvla][INFO] - Step 3709, Loss: {'train/loss': 0.43964269757270813, 'train/grad_norm_pre_clip': 12.043679237365723, 'timing/data': 0.02503122016787529, 'timing/model': 1.139920936897397, 'learning_rate/qwen_vl_interface': 4.906432472559996e-06, 'learning_rate/action_model': 4.9064324725599955e-05, 'epoch': 1.43}) +[2026-06-18 00:51:48,537][starVLA.training.train_starvla][INFO] - Step 3710, Loss: {'train/loss': 0.5076841115951538, 'train/grad_norm_pre_clip': 10.63292407989502, 'timing/data': 0.0424833707511425, 'timing/model': 1.1717635681852698, 'learning_rate/qwen_vl_interface': 4.90427548226734e-06, 'learning_rate/action_model': 4.9042754822673396e-05, 'epoch': 1.43}) +[2026-06-18 00:51:53,543][starVLA.training.train_starvla][INFO] - Step 3711, Loss: {'train/loss': 0.27462247014045715, 'train/grad_norm_pre_clip': 12.133332252502441, 'timing/data': 0.022291135974228382, 'timing/model': 1.1055224565789104, 'learning_rate/qwen_vl_interface': 4.902118563643751e-06, 'learning_rate/action_model': 4.902118563643751e-05, 'epoch': 1.43}) +[2026-06-18 00:51:57,957][starVLA.training.train_starvla][INFO] - Step 3712, Loss: {'train/loss': 0.44677433371543884, 'train/grad_norm_pre_clip': 11.530579566955566, 'timing/data': 0.020280285738408566, 'timing/model': 1.1594778513535857, 'learning_rate/qwen_vl_interface': 4.899961717136356e-06, 'learning_rate/action_model': 4.899961717136355e-05, 'epoch': 1.43}) +[2026-06-18 00:52:02,713][starVLA.training.train_starvla][INFO] - Step 3713, Loss: {'train/loss': 0.5330489873886108, 'train/grad_norm_pre_clip': 11.43535327911377, 'timing/data': 0.029742413200438023, 'timing/model': 1.098162341862917, 'learning_rate/qwen_vl_interface': 4.897804943192273e-06, 'learning_rate/action_model': 4.897804943192273e-05, 'epoch': 1.43}) +[2026-06-18 00:52:07,913][starVLA.training.train_starvla][INFO] - Step 3714, Loss: {'train/loss': 0.40058639645576477, 'train/grad_norm_pre_clip': 15.305892944335938, 'timing/data': 0.02744625974446535, 'timing/model': 1.2099784901365638, 'learning_rate/qwen_vl_interface': 4.895648242258605e-06, 'learning_rate/action_model': 4.895648242258605e-05, 'epoch': 1.43}) +[2026-06-18 00:52:12,484][starVLA.training.train_starvla][INFO] - Step 3715, Loss: {'train/loss': 0.5728971362113953, 'train/grad_norm_pre_clip': 9.83504867553711, 'timing/data': 0.022929806262254715, 'timing/model': 1.10213253647089, 'learning_rate/qwen_vl_interface': 4.893491614782438e-06, 'learning_rate/action_model': 4.8934916147824375e-05, 'epoch': 1.43}) +[2026-06-18 00:52:17,260][starVLA.training.train_starvla][INFO] - Step 3716, Loss: {'train/loss': 0.24356982111930847, 'train/grad_norm_pre_clip': 10.8341703414917, 'timing/data': 0.034634267911314964, 'timing/model': 1.5430338885635138, 'learning_rate/qwen_vl_interface': 4.891335061210841e-06, 'learning_rate/action_model': 4.891335061210841e-05, 'epoch': 1.43}) +[2026-06-18 00:52:22,327][starVLA.training.train_starvla][INFO] - Step 3717, Loss: {'train/loss': 0.41167116165161133, 'train/grad_norm_pre_clip': 8.987515449523926, 'timing/data': 0.02403593808412552, 'timing/model': 1.6397487372159958, 'learning_rate/qwen_vl_interface': 4.889178581990874e-06, 'learning_rate/action_model': 4.8891785819908734e-05, 'epoch': 1.43}) +[2026-06-18 00:52:27,340][starVLA.training.train_starvla][INFO] - Step 3718, Loss: {'train/loss': 0.49056562781333923, 'train/grad_norm_pre_clip': 13.227618217468262, 'timing/data': 0.02380231861025095, 'timing/model': 1.371422422118485, 'learning_rate/qwen_vl_interface': 4.887022177569576e-06, 'learning_rate/action_model': 4.8870221775695754e-05, 'epoch': 1.43}) +[2026-06-18 00:52:35,259][starVLA.training.train_starvla][INFO] - Step 3719, Loss: {'train/loss': 0.2650986313819885, 'train/grad_norm_pre_clip': 14.06478500366211, 'timing/data': 0.023217150941491127, 'timing/model': 2.067380703985691, 'learning_rate/qwen_vl_interface': 4.88486584839397e-06, 'learning_rate/action_model': 4.8848658483939695e-05, 'epoch': 1.43}) +[2026-06-18 00:52:47,350][starVLA.training.train_starvla][INFO] - Step 3720, Loss: {'train/loss': 0.4175098240375519, 'train/grad_norm_pre_clip': 12.16574478149414, 'timing/data': 0.040919226594269276, 'timing/model': 1.29231021925807, 'learning_rate/qwen_vl_interface': 4.882709594911065e-06, 'learning_rate/action_model': 4.8827095949110655e-05, 'epoch': 1.43}) +[2026-06-18 00:53:03,828][starVLA.training.train_starvla][INFO] - Step 3721, Loss: {'train/loss': 0.47624456882476807, 'train/grad_norm_pre_clip': 11.744575500488281, 'timing/data': 0.035560532473027706, 'timing/model': 2.504278323613107, 'learning_rate/qwen_vl_interface': 4.880553417567858e-06, 'learning_rate/action_model': 4.880553417567858e-05, 'epoch': 1.43}) +[2026-06-18 00:53:18,565][starVLA.training.train_starvla][INFO] - Step 3722, Loss: {'train/loss': 0.17998121678829193, 'train/grad_norm_pre_clip': 6.475403785705566, 'timing/data': 0.0013468945398926735, 'timing/model': 1.9507542103528976, 'learning_rate/qwen_vl_interface': 4.878397316811324e-06, 'learning_rate/action_model': 4.878397316811324e-05, 'epoch': 1.43}) +[2026-06-18 00:53:35,144][starVLA.training.train_starvla][INFO] - Step 3723, Loss: {'train/loss': 0.42814159393310547, 'train/grad_norm_pre_clip': 11.265950202941895, 'timing/data': 0.002218896523118019, 'timing/model': 2.9977597668766975, 'learning_rate/qwen_vl_interface': 4.876241293088428e-06, 'learning_rate/action_model': 4.876241293088428e-05, 'epoch': 1.43}) +[2026-06-18 00:53:49,744][starVLA.training.train_starvla][INFO] - Step 3724, Loss: {'train/loss': 0.33735015988349915, 'train/grad_norm_pre_clip': 12.729552268981934, 'timing/data': 0.0010413918644189835, 'timing/model': 5.171455362811685, 'learning_rate/qwen_vl_interface': 4.874085346846111e-06, 'learning_rate/action_model': 4.87408534684611e-05, 'epoch': 1.43}) +[2026-06-18 00:54:00,643][starVLA.training.train_starvla][INFO] - Step 3725, Loss: {'train/loss': 0.3441292643547058, 'train/grad_norm_pre_clip': 10.138407707214355, 'timing/data': 0.0009718099609017372, 'timing/model': 1.1213082447648048, 'learning_rate/qwen_vl_interface': 4.871929478531309e-06, 'learning_rate/action_model': 4.871929478531309e-05, 'epoch': 1.43}) +[2026-06-18 00:54:13,185][starVLA.training.train_starvla][INFO] - Step 3726, Loss: {'train/loss': 0.5405703186988831, 'train/grad_norm_pre_clip': 14.586843490600586, 'timing/data': 0.0011689383536577225, 'timing/model': 1.4514812622219324, 'learning_rate/qwen_vl_interface': 4.869773688590932e-06, 'learning_rate/action_model': 4.8697736885909325e-05, 'epoch': 1.43}) +[2026-06-18 00:54:19,202][starVLA.training.train_starvla][INFO] - Step 3727, Loss: {'train/loss': 0.5496755838394165, 'train/grad_norm_pre_clip': 10.881808280944824, 'timing/data': 0.001414930447936058, 'timing/model': 2.4077377393841743, 'learning_rate/qwen_vl_interface': 4.867617977471878e-06, 'learning_rate/action_model': 4.867617977471878e-05, 'epoch': 1.43}) +[2026-06-18 00:54:23,593][starVLA.training.train_starvla][INFO] - Step 3728, Loss: {'train/loss': 0.3531741201877594, 'train/grad_norm_pre_clip': 15.096495628356934, 'timing/data': 0.001309288665652275, 'timing/model': 1.1434517921879888, 'learning_rate/qwen_vl_interface': 4.865462345621029e-06, 'learning_rate/action_model': 4.865462345621029e-05, 'epoch': 1.43}) +[2026-06-18 00:54:28,319][starVLA.training.train_starvla][INFO] - Step 3729, Loss: {'train/loss': 0.2528778612613678, 'train/grad_norm_pre_clip': 7.066916465759277, 'timing/data': 0.001137235201895237, 'timing/model': 1.0928379893302917, 'learning_rate/qwen_vl_interface': 4.86330679348525e-06, 'learning_rate/action_model': 4.863306793485249e-05, 'epoch': 1.43}) +[2026-06-18 00:54:33,741][starVLA.training.train_starvla][INFO] - Step 3730, Loss: {'train/loss': 0.19719944894313812, 'train/grad_norm_pre_clip': 7.470797061920166, 'timing/data': 0.0009836722165346146, 'timing/model': 1.54620442353189, 'learning_rate/qwen_vl_interface': 4.8611513215113894e-06, 'learning_rate/action_model': 4.861151321511389e-05, 'epoch': 1.43}) +[2026-06-18 00:54:38,824][starVLA.training.train_starvla][INFO] - Step 3731, Loss: {'train/loss': 0.6520857810974121, 'train/grad_norm_pre_clip': 15.104966163635254, 'timing/data': 0.0014272704720497131, 'timing/model': 1.0841616224497557, 'learning_rate/qwen_vl_interface': 4.858995930146276e-06, 'learning_rate/action_model': 4.858995930146275e-05, 'epoch': 1.44}) +[2026-06-18 00:54:43,628][starVLA.training.train_starvla][INFO] - Step 3732, Loss: {'train/loss': 0.42979851365089417, 'train/grad_norm_pre_clip': 14.69117259979248, 'timing/data': 0.001179930754005909, 'timing/model': 1.2635754449293017, 'learning_rate/qwen_vl_interface': 4.856840619836729e-06, 'learning_rate/action_model': 4.856840619836728e-05, 'epoch': 1.44}) +[2026-06-18 00:54:49,097][starVLA.training.train_starvla][INFO] - Step 3733, Loss: {'train/loss': 0.28333404660224915, 'train/grad_norm_pre_clip': 9.622014045715332, 'timing/data': 0.0014010248705744743, 'timing/model': 1.60128378495574, 'learning_rate/qwen_vl_interface': 4.854685391029543e-06, 'learning_rate/action_model': 4.854685391029543e-05, 'epoch': 1.44}) +[2026-06-18 00:54:54,211][starVLA.training.train_starvla][INFO] - Step 3734, Loss: {'train/loss': 0.3325459063053131, 'train/grad_norm_pre_clip': 14.596126556396484, 'timing/data': 0.0018965201452374458, 'timing/model': 1.157357333227992, 'learning_rate/qwen_vl_interface': 4.852530244171502e-06, 'learning_rate/action_model': 4.852530244171502e-05, 'epoch': 1.44}) +[2026-06-18 00:54:59,096][starVLA.training.train_starvla][INFO] - Step 3735, Loss: {'train/loss': 0.6340053081512451, 'train/grad_norm_pre_clip': 17.542516708374023, 'timing/data': 0.0012741060927510262, 'timing/model': 1.2448201915249228, 'learning_rate/qwen_vl_interface': 4.850375179709368e-06, 'learning_rate/action_model': 4.8503751797093674e-05, 'epoch': 1.44}) +[2026-06-18 00:55:04,369][starVLA.training.train_starvla][INFO] - Step 3736, Loss: {'train/loss': 0.20923493802547455, 'train/grad_norm_pre_clip': 6.142431735992432, 'timing/data': 0.0011087674647569656, 'timing/model': 1.1931645451113582, 'learning_rate/qwen_vl_interface': 4.84822019808989e-06, 'learning_rate/action_model': 4.84822019808989e-05, 'epoch': 1.44}) +[2026-06-18 00:55:09,857][starVLA.training.train_starvla][INFO] - Step 3737, Loss: {'train/loss': 0.2885989546775818, 'train/grad_norm_pre_clip': 9.240617752075195, 'timing/data': 0.0014457162469625473, 'timing/model': 1.4950376460328698, 'learning_rate/qwen_vl_interface': 4.846065299759799e-06, 'learning_rate/action_model': 4.846065299759799e-05, 'epoch': 1.44}) +[2026-06-18 00:55:14,593][starVLA.training.train_starvla][INFO] - Step 3738, Loss: {'train/loss': 0.39330101013183594, 'train/grad_norm_pre_clip': 13.425813674926758, 'timing/data': 0.0012869834899902344, 'timing/model': 1.338801957666874, 'learning_rate/qwen_vl_interface': 4.843910485165802e-06, 'learning_rate/action_model': 4.843910485165802e-05, 'epoch': 1.44}) +[2026-06-18 00:55:19,753][starVLA.training.train_starvla][INFO] - Step 3739, Loss: {'train/loss': 0.5467080473899841, 'train/grad_norm_pre_clip': 11.651500701904297, 'timing/data': 0.0010086912661790848, 'timing/model': 1.1456844247877598, 'learning_rate/qwen_vl_interface': 4.8417557547546015e-06, 'learning_rate/action_model': 4.8417557547546015e-05, 'epoch': 1.44}) +[2026-06-18 00:55:24,972][starVLA.training.train_starvla][INFO] - Step 3740, Loss: {'train/loss': 0.22412948310375214, 'train/grad_norm_pre_clip': 10.579354286193848, 'timing/data': 0.0014506299048662186, 'timing/model': 1.1464616935700178, 'learning_rate/qwen_vl_interface': 4.839601108972872e-06, 'learning_rate/action_model': 4.839601108972871e-05, 'epoch': 1.44}) +[2026-06-18 00:55:30,293][starVLA.training.train_starvla][INFO] - Step 3741, Loss: {'train/loss': 0.40577736496925354, 'train/grad_norm_pre_clip': 9.69843864440918, 'timing/data': 0.0014273039996623993, 'timing/model': 1.2879550037905574, 'learning_rate/qwen_vl_interface': 4.837446548267275e-06, 'learning_rate/action_model': 4.837446548267274e-05, 'epoch': 1.44}) +[2026-06-18 00:55:41,011][starVLA.training.train_starvla][INFO] - Step 3742, Loss: {'train/loss': 0.2221086621284485, 'train/grad_norm_pre_clip': 8.057353973388672, 'timing/data': 0.0012960107997059822, 'timing/model': 5.5699786972254515, 'learning_rate/qwen_vl_interface': 4.835292073084449e-06, 'learning_rate/action_model': 4.8352920730844486e-05, 'epoch': 1.44}) +[2026-06-18 00:55:53,146][starVLA.training.train_starvla][INFO] - Step 3743, Loss: {'train/loss': 0.373592734336853, 'train/grad_norm_pre_clip': 10.284650802612305, 'timing/data': 0.001304948702454567, 'timing/model': 5.920798999257386, 'learning_rate/qwen_vl_interface': 4.833137683871025e-06, 'learning_rate/action_model': 4.833137683871025e-05, 'epoch': 1.44}) +[2026-06-18 00:56:08,724][starVLA.training.train_starvla][INFO] - Step 3744, Loss: {'train/loss': 0.3599824607372284, 'train/grad_norm_pre_clip': 15.724812507629395, 'timing/data': 0.000984487123787403, 'timing/model': 5.527828642167151, 'learning_rate/qwen_vl_interface': 4.830983381073609e-06, 'learning_rate/action_model': 4.830983381073608e-05, 'epoch': 1.44}) +[2026-06-18 00:56:20,764][starVLA.training.train_starvla][INFO] - Step 3745, Loss: {'train/loss': 0.12289939820766449, 'train/grad_norm_pre_clip': 7.331798553466797, 'timing/data': 0.0009488454088568687, 'timing/model': 1.8548301439732313, 'learning_rate/qwen_vl_interface': 4.828829165138787e-06, 'learning_rate/action_model': 4.828829165138787e-05, 'epoch': 1.44}) +[2026-06-18 00:56:31,042][starVLA.training.train_starvla][INFO] - Step 3746, Loss: {'train/loss': 0.5275008678436279, 'train/grad_norm_pre_clip': 22.710981369018555, 'timing/data': 0.0015515973791480064, 'timing/model': 1.758996231481433, 'learning_rate/qwen_vl_interface': 4.826675036513132e-06, 'learning_rate/action_model': 4.826675036513132e-05, 'epoch': 1.44}) +[2026-06-18 00:56:39,204][starVLA.training.train_starvla][INFO] - Step 3747, Loss: {'train/loss': 0.2466086745262146, 'train/grad_norm_pre_clip': 12.400440216064453, 'timing/data': 0.0008472604677081108, 'timing/model': 4.335760273039341, 'learning_rate/qwen_vl_interface': 4.824520995643198e-06, 'learning_rate/action_model': 4.8245209956431984e-05, 'epoch': 1.44}) +[2026-06-18 00:56:49,039][starVLA.training.train_starvla][INFO] - Step 3748, Loss: {'train/loss': 0.7601011395454407, 'train/grad_norm_pre_clip': 21.28592872619629, 'timing/data': 0.001495138742029667, 'timing/model': 1.3725698236376047, 'learning_rate/qwen_vl_interface': 4.82236704297552e-06, 'learning_rate/action_model': 4.82236704297552e-05, 'epoch': 1.44}) +[2026-06-18 00:56:59,508][starVLA.training.train_starvla][INFO] - Step 3749, Loss: {'train/loss': 0.30649611353874207, 'train/grad_norm_pre_clip': 10.41930103302002, 'timing/data': 0.003977441228926182, 'timing/model': 6.200824360363185, 'learning_rate/qwen_vl_interface': 4.820213178956613e-06, 'learning_rate/action_model': 4.820213178956612e-05, 'epoch': 1.44}) +[2026-06-18 00:57:11,041][starVLA.training.train_starvla][INFO] - Step 3750, Loss: {'train/loss': 0.42423537373542786, 'train/grad_norm_pre_clip': 15.98344612121582, 'timing/data': 0.0012353993952274323, 'timing/model': 1.8399657541885972, 'learning_rate/qwen_vl_interface': 4.818059404032977e-06, 'learning_rate/action_model': 4.818059404032977e-05, 'epoch': 1.44}) +[2026-06-18 00:57:16,692][starVLA.training.train_starvla][INFO] - Step 3751, Loss: {'train/loss': 0.6101599335670471, 'train/grad_norm_pre_clip': 16.167675018310547, 'timing/data': 0.0016010226681828499, 'timing/model': 1.5647121910005808, 'learning_rate/qwen_vl_interface': 4.815905718651092e-06, 'learning_rate/action_model': 4.815905718651092e-05, 'epoch': 1.44}) +[2026-06-18 00:57:21,690][starVLA.training.train_starvla][INFO] - Step 3752, Loss: {'train/loss': 0.5456461310386658, 'train/grad_norm_pre_clip': 18.144527435302734, 'timing/data': 0.0013231923803687096, 'timing/model': 1.1701777959242463, 'learning_rate/qwen_vl_interface': 4.813752123257419e-06, 'learning_rate/action_model': 4.813752123257418e-05, 'epoch': 1.44}) +[2026-06-18 00:57:26,896][starVLA.training.train_starvla][INFO] - Step 3753, Loss: {'train/loss': 0.5095491409301758, 'train/grad_norm_pre_clip': 17.076297760009766, 'timing/data': 0.0012675663456320763, 'timing/model': 1.148098479025066, 'learning_rate/qwen_vl_interface': 4.811598618298398e-06, 'learning_rate/action_model': 4.811598618298398e-05, 'epoch': 1.44}) +[2026-06-18 00:57:31,870][starVLA.training.train_starvla][INFO] - Step 3754, Loss: {'train/loss': 0.46778902411460876, 'train/grad_norm_pre_clip': 10.344441413879395, 'timing/data': 0.0011807652190327644, 'timing/model': 1.281563488766551, 'learning_rate/qwen_vl_interface': 4.809445204220457e-06, 'learning_rate/action_model': 4.809445204220457e-05, 'epoch': 1.44}) +[2026-06-18 00:57:37,323][starVLA.training.train_starvla][INFO] - Step 3755, Loss: {'train/loss': 0.6130976676940918, 'train/grad_norm_pre_clip': 11.938494682312012, 'timing/data': 0.0012691756710410118, 'timing/model': 1.4379081949591637, 'learning_rate/qwen_vl_interface': 4.8072918814699985e-06, 'learning_rate/action_model': 4.8072918814699986e-05, 'epoch': 1.44}) +[2026-06-18 00:57:42,779][starVLA.training.train_starvla][INFO] - Step 3756, Loss: {'train/loss': 0.7346624732017517, 'train/grad_norm_pre_clip': 23.016616821289062, 'timing/data': 0.0014985259622335434, 'timing/model': 1.1989425467327237, 'learning_rate/qwen_vl_interface': 4.80513865049341e-06, 'learning_rate/action_model': 4.80513865049341e-05, 'epoch': 1.44}) +[2026-06-18 00:57:48,946][starVLA.training.train_starvla][INFO] - Step 3757, Loss: {'train/loss': 0.38657093048095703, 'train/grad_norm_pre_clip': 14.323470115661621, 'timing/data': 0.0030744262039661407, 'timing/model': 2.630424709059298, 'learning_rate/qwen_vl_interface': 4.802985511737057e-06, 'learning_rate/action_model': 4.802985511737057e-05, 'epoch': 1.45}) +[2026-06-18 00:57:53,616][starVLA.training.train_starvla][INFO] - Step 3758, Loss: {'train/loss': 0.4471064805984497, 'train/grad_norm_pre_clip': 10.11589241027832, 'timing/data': 0.0011511826887726784, 'timing/model': 1.082401872612536, 'learning_rate/qwen_vl_interface': 4.80083246564729e-06, 'learning_rate/action_model': 4.8008324656472894e-05, 'epoch': 1.45}) +[2026-06-18 00:57:59,616][starVLA.training.train_starvla][INFO] - Step 3759, Loss: {'train/loss': 0.26002049446105957, 'train/grad_norm_pre_clip': 11.135990142822266, 'timing/data': 0.0010286401957273483, 'timing/model': 1.4883189015090466, 'learning_rate/qwen_vl_interface': 4.798679512670436e-06, 'learning_rate/action_model': 4.798679512670435e-05, 'epoch': 1.45}) +[2026-06-18 00:58:04,521][starVLA.training.train_starvla][INFO] - Step 3760, Loss: {'train/loss': 0.5533946752548218, 'train/grad_norm_pre_clip': 13.34900188446045, 'timing/data': 0.0010180193930864334, 'timing/model': 1.4330682568252087, 'learning_rate/qwen_vl_interface': 4.796526653252802e-06, 'learning_rate/action_model': 4.796526653252802e-05, 'epoch': 1.45}) +[2026-06-18 00:58:11,627][starVLA.training.train_starvla][INFO] - Step 3761, Loss: {'train/loss': 0.13338039815425873, 'train/grad_norm_pre_clip': 5.276978492736816, 'timing/data': 0.0015863822773098946, 'timing/model': 1.6093080034479499, 'learning_rate/qwen_vl_interface': 4.794373887840683e-06, 'learning_rate/action_model': 4.794373887840682e-05, 'epoch': 1.45}) +[2026-06-18 00:58:24,091][starVLA.training.train_starvla][INFO] - Step 3762, Loss: {'train/loss': 0.3778092563152313, 'train/grad_norm_pre_clip': 14.191873550415039, 'timing/data': 0.0013246312737464905, 'timing/model': 1.6992232035845518, 'learning_rate/qwen_vl_interface': 4.792221216880347e-06, 'learning_rate/action_model': 4.792221216880347e-05, 'epoch': 1.45}) +[2026-06-18 00:58:28,632][starVLA.training.train_starvla][INFO] - Step 3763, Loss: {'train/loss': 0.3962629735469818, 'train/grad_norm_pre_clip': 14.80264663696289, 'timing/data': 0.002584279514849186, 'timing/model': 1.3322082264348865, 'learning_rate/qwen_vl_interface': 4.790068640818046e-06, 'learning_rate/action_model': 4.790068640818045e-05, 'epoch': 1.45}) +[2026-06-18 00:58:33,783][starVLA.training.train_starvla][INFO] - Step 3764, Loss: {'train/loss': 0.44767236709594727, 'train/grad_norm_pre_clip': 15.97758674621582, 'timing/data': 0.0011492688208818436, 'timing/model': 1.1987005453556776, 'learning_rate/qwen_vl_interface': 4.787916160100008e-06, 'learning_rate/action_model': 4.787916160100008e-05, 'epoch': 1.45}) +[2026-06-18 00:58:41,170][starVLA.training.train_starvla][INFO] - Step 3765, Loss: {'train/loss': 0.6515076756477356, 'train/grad_norm_pre_clip': 15.966626167297363, 'timing/data': 0.0012382958084344864, 'timing/model': 2.125599763356149, 'learning_rate/qwen_vl_interface': 4.78576377517245e-06, 'learning_rate/action_model': 4.7857637751724504e-05, 'epoch': 1.45}) +[2026-06-18 00:58:46,817][starVLA.training.train_starvla][INFO] - Step 3766, Loss: {'train/loss': 0.39049941301345825, 'train/grad_norm_pre_clip': 12.00001335144043, 'timing/data': 0.0012162020429968834, 'timing/model': 1.5445417752489448, 'learning_rate/qwen_vl_interface': 4.783611486481561e-06, 'learning_rate/action_model': 4.7836114864815604e-05, 'epoch': 1.45}) +[2026-06-18 00:58:52,263][starVLA.training.train_starvla][INFO] - Step 3767, Loss: {'train/loss': 0.38340866565704346, 'train/grad_norm_pre_clip': 12.76468563079834, 'timing/data': 0.0015375055372714996, 'timing/model': 1.5094641735777259, 'learning_rate/qwen_vl_interface': 4.781459294473511e-06, 'learning_rate/action_model': 4.781459294473511e-05, 'epoch': 1.45}) +[2026-06-18 00:58:57,152][starVLA.training.train_starvla][INFO] - Step 3768, Loss: {'train/loss': 0.6058884859085083, 'train/grad_norm_pre_clip': 14.357519149780273, 'timing/data': 0.0013943351805210114, 'timing/model': 1.247117057442665, 'learning_rate/qwen_vl_interface': 4.779307199594454e-06, 'learning_rate/action_model': 4.779307199594454e-05, 'epoch': 1.45}) +[2026-06-18 00:59:04,508][starVLA.training.train_starvla][INFO] - Step 3769, Loss: {'train/loss': 0.44845107197761536, 'train/grad_norm_pre_clip': 13.846678733825684, 'timing/data': 0.0009626848623156548, 'timing/model': 2.875514735467732, 'learning_rate/qwen_vl_interface': 4.777155202290522e-06, 'learning_rate/action_model': 4.777155202290521e-05, 'epoch': 1.45}) +[2026-06-18 00:59:13,414][starVLA.training.train_starvla][INFO] - Step 3770, Loss: {'train/loss': 0.16122759878635406, 'train/grad_norm_pre_clip': 8.392356872558594, 'timing/data': 0.002515028230845928, 'timing/model': 2.2577440524473786, 'learning_rate/qwen_vl_interface': 4.775003303007824e-06, 'learning_rate/action_model': 4.775003303007824e-05, 'epoch': 1.45}) +[2026-06-18 00:59:18,267][starVLA.training.train_starvla][INFO] - Step 3771, Loss: {'train/loss': 0.12530598044395447, 'train/grad_norm_pre_clip': 5.987120628356934, 'timing/data': 0.0012010252103209496, 'timing/model': 1.1175159234553576, 'learning_rate/qwen_vl_interface': 4.772851502192453e-06, 'learning_rate/action_model': 4.772851502192452e-05, 'epoch': 1.45}) +[2026-06-18 00:59:23,891][starVLA.training.train_starvla][INFO] - Step 3772, Loss: {'train/loss': 0.32319462299346924, 'train/grad_norm_pre_clip': 18.227521896362305, 'timing/data': 0.0011007795110344887, 'timing/model': 1.7230254858732224, 'learning_rate/qwen_vl_interface': 4.770699800290478e-06, 'learning_rate/action_model': 4.770699800290478e-05, 'epoch': 1.45}) +[2026-06-18 00:59:38,646][starVLA.training.train_starvla][INFO] - Step 3773, Loss: {'train/loss': 0.6121371388435364, 'train/grad_norm_pre_clip': 23.424734115600586, 'timing/data': 0.0012767389416694641, 'timing/model': 4.697740477509797, 'learning_rate/qwen_vl_interface': 4.768548197747953e-06, 'learning_rate/action_model': 4.768548197747953e-05, 'epoch': 1.45}) +[2026-06-18 00:59:44,096][starVLA.training.train_starvla][INFO] - Step 3774, Loss: {'train/loss': 0.3293767273426056, 'train/grad_norm_pre_clip': 13.863308906555176, 'timing/data': 0.001467932015657425, 'timing/model': 1.1468789298087358, 'learning_rate/qwen_vl_interface': 4.766396695010903e-06, 'learning_rate/action_model': 4.766396695010903e-05, 'epoch': 1.45}) +[2026-06-18 00:59:49,369][starVLA.training.train_starvla][INFO] - Step 3775, Loss: {'train/loss': 0.4948703646659851, 'train/grad_norm_pre_clip': 18.843059539794922, 'timing/data': 0.0009933877736330032, 'timing/model': 1.196156032383442, 'learning_rate/qwen_vl_interface': 4.764245292525337e-06, 'learning_rate/action_model': 4.7642452925253376e-05, 'epoch': 1.45}) +[2026-06-18 00:59:55,990][starVLA.training.train_starvla][INFO] - Step 3776, Loss: {'train/loss': 0.589758574962616, 'train/grad_norm_pre_clip': 13.79732608795166, 'timing/data': 0.0012595904991030693, 'timing/model': 1.4835150577127934, 'learning_rate/qwen_vl_interface': 4.762093990737247e-06, 'learning_rate/action_model': 4.762093990737246e-05, 'epoch': 1.45}) +[2026-06-18 01:00:00,634][starVLA.training.train_starvla][INFO] - Step 3777, Loss: {'train/loss': 0.19262900948524475, 'train/grad_norm_pre_clip': 7.696968078613281, 'timing/data': 0.0011617382988333702, 'timing/model': 1.0810219487175345, 'learning_rate/qwen_vl_interface': 4.759942790092596e-06, 'learning_rate/action_model': 4.759942790092596e-05, 'epoch': 1.45}) +[2026-06-18 01:00:07,078][starVLA.training.train_starvla][INFO] - Step 3778, Loss: {'train/loss': 0.9019182324409485, 'train/grad_norm_pre_clip': 19.268653869628906, 'timing/data': 0.0012677861377596855, 'timing/model': 2.6779794842004776, 'learning_rate/qwen_vl_interface': 4.757791691037332e-06, 'learning_rate/action_model': 4.757791691037332e-05, 'epoch': 1.45}) +[2026-06-18 01:00:13,033][starVLA.training.train_starvla][INFO] - Step 3779, Loss: {'train/loss': 0.29828083515167236, 'train/grad_norm_pre_clip': 10.050386428833008, 'timing/data': 0.001310368999838829, 'timing/model': 1.4484464768320322, 'learning_rate/qwen_vl_interface': 4.755640694017381e-06, 'learning_rate/action_model': 4.755640694017381e-05, 'epoch': 1.45}) +[2026-06-18 01:00:19,872][starVLA.training.train_starvla][INFO] - Step 3780, Loss: {'train/loss': 0.26198717951774597, 'train/grad_norm_pre_clip': 10.013866424560547, 'timing/data': 0.0016985414549708366, 'timing/model': 1.1219188896939158, 'learning_rate/qwen_vl_interface': 4.753489799478647e-06, 'learning_rate/action_model': 4.7534897994786464e-05, 'epoch': 1.45}) +[2026-06-18 01:00:24,517][starVLA.training.train_starvla][INFO] - Step 3781, Loss: {'train/loss': 0.4403092563152313, 'train/grad_norm_pre_clip': 9.094143867492676, 'timing/data': 0.0011011157184839249, 'timing/model': 1.1003973558545113, 'learning_rate/qwen_vl_interface': 4.7513390078670105e-06, 'learning_rate/action_model': 4.75133900786701e-05, 'epoch': 1.45}) +[2026-06-18 01:00:29,362][starVLA.training.train_starvla][INFO] - Step 3782, Loss: {'train/loss': 0.20907282829284668, 'train/grad_norm_pre_clip': 8.699806213378906, 'timing/data': 0.0011982666328549385, 'timing/model': 1.4384786421433091, 'learning_rate/qwen_vl_interface': 4.749188319628333e-06, 'learning_rate/action_model': 4.749188319628333e-05, 'epoch': 1.45}) +[2026-06-18 01:00:33,710][starVLA.training.train_starvla][INFO] - Step 3783, Loss: {'train/loss': 0.6772451400756836, 'train/grad_norm_pre_clip': 13.96957015991211, 'timing/data': 0.001451350748538971, 'timing/model': 1.115452527999878, 'learning_rate/qwen_vl_interface': 4.747037735208457e-06, 'learning_rate/action_model': 4.7470377352084575e-05, 'epoch': 1.46}) +[2026-06-18 01:00:38,353][starVLA.training.train_starvla][INFO] - Step 3784, Loss: {'train/loss': 0.438164085149765, 'train/grad_norm_pre_clip': 38.399723052978516, 'timing/data': 0.0011889347806572914, 'timing/model': 1.08962990809232, 'learning_rate/qwen_vl_interface': 4.7448872550532006e-06, 'learning_rate/action_model': 4.7448872550532e-05, 'epoch': 1.46}) +[2026-06-18 01:00:42,944][starVLA.training.train_starvla][INFO] - Step 3785, Loss: {'train/loss': 0.3367445468902588, 'train/grad_norm_pre_clip': 13.211670875549316, 'timing/data': 0.0014634616672992706, 'timing/model': 1.0750490510836244, 'learning_rate/qwen_vl_interface': 4.74273687960836e-06, 'learning_rate/action_model': 4.74273687960836e-05, 'epoch': 1.46}) +[2026-06-18 01:00:47,779][starVLA.training.train_starvla][INFO] - Step 3786, Loss: {'train/loss': 0.4943805932998657, 'train/grad_norm_pre_clip': 18.521167755126953, 'timing/data': 0.001034233719110489, 'timing/model': 1.1077673425897956, 'learning_rate/qwen_vl_interface': 4.7405866093197075e-06, 'learning_rate/action_model': 4.740586609319707e-05, 'epoch': 1.46}) +[2026-06-18 01:00:52,210][starVLA.training.train_starvla][INFO] - Step 3787, Loss: {'train/loss': 0.1944591999053955, 'train/grad_norm_pre_clip': 6.712932109832764, 'timing/data': 0.0012520058080554008, 'timing/model': 1.1237989393994212, 'learning_rate/qwen_vl_interface': 4.738436444633002e-06, 'learning_rate/action_model': 4.738436444633001e-05, 'epoch': 1.46}) +[2026-06-18 01:00:57,365][starVLA.training.train_starvla][INFO] - Step 3788, Loss: {'train/loss': 0.52634197473526, 'train/grad_norm_pre_clip': 13.796845436096191, 'timing/data': 0.0012856628745794296, 'timing/model': 1.1714779362082481, 'learning_rate/qwen_vl_interface': 4.73628638599397e-06, 'learning_rate/action_model': 4.73628638599397e-05, 'epoch': 1.46}) +[2026-06-18 01:01:06,160][starVLA.training.train_starvla][INFO] - Step 3789, Loss: {'train/loss': 0.5452173352241516, 'train/grad_norm_pre_clip': 16.15484619140625, 'timing/data': 0.0009229928255081177, 'timing/model': 1.122360316105187, 'learning_rate/qwen_vl_interface': 4.734136433848323e-06, 'learning_rate/action_model': 4.7341364338483224e-05, 'epoch': 1.46}) +[2026-06-18 01:01:11,589][starVLA.training.train_starvla][INFO] - Step 3790, Loss: {'train/loss': 0.23974502086639404, 'train/grad_norm_pre_clip': 8.122645378112793, 'timing/data': 0.0011324528604745865, 'timing/model': 1.3101251348853111, 'learning_rate/qwen_vl_interface': 4.731986588641748e-06, 'learning_rate/action_model': 4.7319865886417484e-05, 'epoch': 1.46}) +[2026-06-18 01:01:16,636][starVLA.training.train_starvla][INFO] - Step 3791, Loss: {'train/loss': 0.4831795394420624, 'train/grad_norm_pre_clip': 13.669839859008789, 'timing/data': 0.0011562341824173927, 'timing/model': 1.1136376662179828, 'learning_rate/qwen_vl_interface': 4.729836850819912e-06, 'learning_rate/action_model': 4.729836850819912e-05, 'epoch': 1.46}) +[2026-06-18 01:01:21,622][starVLA.training.train_starvla][INFO] - Step 3792, Loss: {'train/loss': 0.4805854558944702, 'train/grad_norm_pre_clip': 18.58424186706543, 'timing/data': 0.0011921552941203117, 'timing/model': 1.540187862701714, 'learning_rate/qwen_vl_interface': 4.727687220828458e-06, 'learning_rate/action_model': 4.727687220828458e-05, 'epoch': 1.46}) +[2026-06-18 01:01:26,486][starVLA.training.train_starvla][INFO] - Step 3793, Loss: {'train/loss': 0.18325698375701904, 'train/grad_norm_pre_clip': 7.494747638702393, 'timing/data': 0.0009890943765640259, 'timing/model': 1.4564191801473498, 'learning_rate/qwen_vl_interface': 4.7255376991130005e-06, 'learning_rate/action_model': 4.7255376991130005e-05, 'epoch': 1.46}) +[2026-06-18 01:01:34,965][starVLA.training.train_starvla][INFO] - Step 3794, Loss: {'train/loss': 0.4398108720779419, 'train/grad_norm_pre_clip': 16.549835205078125, 'timing/data': 0.0012045074254274368, 'timing/model': 1.2158704567700624, 'learning_rate/qwen_vl_interface': 4.723388286119147e-06, 'learning_rate/action_model': 4.723388286119146e-05, 'epoch': 1.46}) +[2026-06-18 01:01:40,051][starVLA.training.train_starvla][INFO] - Step 3795, Loss: {'train/loss': 0.7364603877067566, 'train/grad_norm_pre_clip': 18.025728225708008, 'timing/data': 0.0012839045375585556, 'timing/model': 1.2704008081927896, 'learning_rate/qwen_vl_interface': 4.721238982292467e-06, 'learning_rate/action_model': 4.721238982292466e-05, 'epoch': 1.46}) +[2026-06-18 01:01:45,356][starVLA.training.train_starvla][INFO] - Step 3796, Loss: {'train/loss': 0.23901791870594025, 'train/grad_norm_pre_clip': 16.920734405517578, 'timing/data': 0.0009236922487616539, 'timing/model': 1.2935892073437572, 'learning_rate/qwen_vl_interface': 4.719089788078514e-06, 'learning_rate/action_model': 4.719089788078514e-05, 'epoch': 1.46}) +[2026-06-18 01:01:50,071][starVLA.training.train_starvla][INFO] - Step 3797, Loss: {'train/loss': 0.5118992924690247, 'train/grad_norm_pre_clip': 17.722782135009766, 'timing/data': 0.0010334644466638565, 'timing/model': 1.1231213230639696, 'learning_rate/qwen_vl_interface': 4.716940703922818e-06, 'learning_rate/action_model': 4.716940703922818e-05, 'epoch': 1.46}) +[2026-06-18 01:01:55,403][starVLA.training.train_starvla][INFO] - Step 3798, Loss: {'train/loss': 0.4596005082130432, 'train/grad_norm_pre_clip': 14.382465362548828, 'timing/data': 0.0012268563732504845, 'timing/model': 1.3335179947316647, 'learning_rate/qwen_vl_interface': 4.714791730270889e-06, 'learning_rate/action_model': 4.71479173027089e-05, 'epoch': 1.46}) +[2026-06-18 01:02:00,181][starVLA.training.train_starvla][INFO] - Step 3799, Loss: {'train/loss': 0.5226789712905884, 'train/grad_norm_pre_clip': 18.53285026550293, 'timing/data': 0.002676107920706272, 'timing/model': 1.2712607299908996, 'learning_rate/qwen_vl_interface': 4.712642867568213e-06, 'learning_rate/action_model': 4.712642867568212e-05, 'epoch': 1.46}) +[2026-06-18 01:06:18,450][starVLA.training.train_starvla][INFO] - Step 3800, Loss: {'train/loss': 0.3558937609195709, 'train/grad_norm_pre_clip': 10.837579727172852, 'eval/loss': 0.46747925877571106, 'eval/latency_4/loss': 0.46747925877571106, 'eval/demon_fire_precision': 0.9656721206288098, 'eval/demon_fire_recall': 0.9775901266645015, 'eval/demon_fire_f1': 0.9715945771465462, 'eval/demon_fire_rate_model': 0.9154185022026432, 'eval/demon_fire_rate_teacher': 0.9042584434654919, 'eval/demon_left_precision': 0.7566137566137566, 'eval/demon_left_recall': 0.6529680365296804, 'eval/demon_left_f1': 0.7009803921568628, 'eval/demon_left_rate_model': 0.11101321585903083, 'eval/demon_left_rate_teacher': 0.12863436123348018, 'eval/demon_right_precision': 0.7867647058823529, 'eval/demon_right_recall': 0.7725631768953068, 'eval/demon_right_f1': 0.7795992714025501, 'eval/demon_right_rate_model': 0.15976505139500735, 'eval/demon_right_rate_teacher': 0.16270190895741557, 'eval/demon_move_f1': 0.7402898317797064, 'eval/demon_cc_f1': 0.8883268339065833, 'eval/latency_4/demon_fire_precision': 0.9656721206288098, 'eval/latency_4/demon_fire_recall': 0.9775901266645015, 'eval/latency_4/demon_fire_f1': 0.9715945771465462, 'eval/latency_4/demon_fire_rate_model': 0.9154185022026432, 'eval/latency_4/demon_fire_rate_teacher': 0.9042584434654919, 'eval/latency_4/demon_left_precision': 0.7566137566137566, 'eval/latency_4/demon_left_recall': 0.6529680365296804, 'eval/latency_4/demon_left_f1': 0.7009803921568628, 'eval/latency_4/demon_left_rate_model': 0.11101321585903083, 'eval/latency_4/demon_left_rate_teacher': 0.12863436123348018, 'eval/latency_4/demon_right_precision': 0.7867647058823529, 'eval/latency_4/demon_right_recall': 0.7725631768953068, 'eval/latency_4/demon_right_f1': 0.7795992714025501, 'eval/latency_4/demon_right_rate_model': 0.15976505139500735, 'eval/latency_4/demon_right_rate_teacher': 0.16270190895741557, 'eval/latency_4/demon_move_f1': 0.7402898317797064, 'eval/latency_4/demon_cc_f1': 0.8883268339065833, 'timing/data': 0.0013056360185146332, 'timing/model': 1.1247122399508953, 'learning_rate/qwen_vl_interface': 4.7104941162602436e-06, 'learning_rate/action_model': 4.710494116260244e-05, 'epoch': 1.46}) +[2026-06-18 01:06:18,454][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_3800_state +[2026-06-18 01:06:18,463][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 01:06:27,766][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_3800_state/model.safetensors +[2026-06-18 01:06:51,684][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_3800_state/optimizer.bin +[2026-06-18 01:06:51,685][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_3800_state/sampler.bin +[2026-06-18 01:06:51,686][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_3800_state/sampler_1.bin +[2026-06-18 01:06:51,690][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_3800_state/random_states_0.pkl +[2026-06-18 01:06:51,696][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3800 +[2026-06-18 01:06:51,696][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 01:06:51,720][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 01:06:57,222][starVLA.training.train_starvla][INFO] - Step 3801, Loss: {'train/loss': 0.4745979607105255, 'train/grad_norm_pre_clip': 13.070226669311523, 'timing/data': 0.0012798234820365906, 'timing/model': 1.586831210181117, 'learning_rate/qwen_vl_interface': 4.708345476792428e-06, 'learning_rate/action_model': 4.7083454767924276e-05, 'epoch': 1.46}) +[2026-06-18 01:07:06,523][starVLA.training.train_starvla][INFO] - Step 3802, Loss: {'train/loss': 0.7614967823028564, 'train/grad_norm_pre_clip': 16.362197875976562, 'timing/data': 0.0012675151228904724, 'timing/model': 5.288411269895732, 'learning_rate/qwen_vl_interface': 4.706196949610178e-06, 'learning_rate/action_model': 4.7061969496101776e-05, 'epoch': 1.46}) +[2026-06-18 01:07:13,893][starVLA.training.train_starvla][INFO] - Step 3803, Loss: {'train/loss': 0.39064353704452515, 'train/grad_norm_pre_clip': 12.353796005249023, 'timing/data': 0.0008116979151964188, 'timing/model': 1.4276195233687758, 'learning_rate/qwen_vl_interface': 4.7040485351588845e-06, 'learning_rate/action_model': 4.7040485351588845e-05, 'epoch': 1.46}) +[2026-06-18 01:07:19,248][starVLA.training.train_starvla][INFO] - Step 3804, Loss: {'train/loss': 0.3612818419933319, 'train/grad_norm_pre_clip': 11.919137001037598, 'timing/data': 0.009758147411048412, 'timing/model': 1.1338513614609838, 'learning_rate/qwen_vl_interface': 4.701900233883916e-06, 'learning_rate/action_model': 4.7019002338839156e-05, 'epoch': 1.46}) +[2026-06-18 01:07:24,344][starVLA.training.train_starvla][INFO] - Step 3805, Loss: {'train/loss': 0.7966325879096985, 'train/grad_norm_pre_clip': 15.694416999816895, 'timing/data': 0.0012169890105724335, 'timing/model': 1.225941974669695, 'learning_rate/qwen_vl_interface': 4.69975204623062e-06, 'learning_rate/action_model': 4.69975204623062e-05, 'epoch': 1.46}) +[2026-06-18 01:07:32,403][starVLA.training.train_starvla][INFO] - Step 3806, Loss: {'train/loss': 0.29935431480407715, 'train/grad_norm_pre_clip': 11.53656005859375, 'timing/data': 0.0012495415285229683, 'timing/model': 1.6021756893023849, 'learning_rate/qwen_vl_interface': 4.697603972644319e-06, 'learning_rate/action_model': 4.697603972644318e-05, 'epoch': 1.46}) +[2026-06-18 01:07:37,279][starVLA.training.train_starvla][INFO] - Step 3807, Loss: {'train/loss': 0.22679643332958221, 'train/grad_norm_pre_clip': 12.037805557250977, 'timing/data': 0.001168699935078621, 'timing/model': 1.5542174689471722, 'learning_rate/qwen_vl_interface': 4.695456013570306e-06, 'learning_rate/action_model': 4.695456013570306e-05, 'epoch': 1.46}) +[2026-06-18 01:07:44,591][starVLA.training.train_starvla][INFO] - Step 3808, Loss: {'train/loss': 0.35437464714050293, 'train/grad_norm_pre_clip': 7.6164350509643555, 'timing/data': 0.0013589151203632355, 'timing/model': 2.0019861357286572, 'learning_rate/qwen_vl_interface': 4.693308169453858e-06, 'learning_rate/action_model': 4.693308169453858e-05, 'epoch': 1.46}) +[2026-06-18 01:07:56,200][starVLA.training.train_starvla][INFO] - Step 3809, Loss: {'train/loss': 0.32482239603996277, 'train/grad_norm_pre_clip': 11.864500999450684, 'timing/data': 0.001185963861644268, 'timing/model': 2.990563591942191, 'learning_rate/qwen_vl_interface': 4.691160440740226e-06, 'learning_rate/action_model': 4.691160440740226e-05, 'epoch': 1.47}) +[2026-06-18 01:08:06,974][starVLA.training.train_starvla][INFO] - Step 3810, Loss: {'train/loss': 0.39934301376342773, 'train/grad_norm_pre_clip': 14.197391510009766, 'timing/data': 0.0014794906601309776, 'timing/model': 3.0462626479566097, 'learning_rate/qwen_vl_interface': 4.689012827874637e-06, 'learning_rate/action_model': 4.689012827874637e-05, 'epoch': 1.47}) +[2026-06-18 01:08:17,954][starVLA.training.train_starvla][INFO] - Step 3811, Loss: {'train/loss': 0.35685935616493225, 'train/grad_norm_pre_clip': 15.542736053466797, 'timing/data': 0.0013958495110273361, 'timing/model': 3.245364219881594, 'learning_rate/qwen_vl_interface': 4.6868653313022905e-06, 'learning_rate/action_model': 4.68686533130229e-05, 'epoch': 1.47}) +[2026-06-18 01:08:32,107][starVLA.training.train_starvla][INFO] - Step 3812, Loss: {'train/loss': 0.5018619894981384, 'train/grad_norm_pre_clip': 13.640028953552246, 'timing/data': 0.0014609917998313904, 'timing/model': 4.33080346789211, 'learning_rate/qwen_vl_interface': 4.684717951468368e-06, 'learning_rate/action_model': 4.6847179514683684e-05, 'epoch': 1.47}) +[2026-06-18 01:08:43,557][starVLA.training.train_starvla][INFO] - Step 3813, Loss: {'train/loss': 0.3164122700691223, 'train/grad_norm_pre_clip': 20.38170623779297, 'timing/data': 0.0009645754471421242, 'timing/model': 5.406046031974256, 'learning_rate/qwen_vl_interface': 4.682570688818024e-06, 'learning_rate/action_model': 4.682570688818024e-05, 'epoch': 1.47}) +[2026-06-18 01:08:48,820][starVLA.training.train_starvla][INFO] - Step 3814, Loss: {'train/loss': 0.3673166036605835, 'train/grad_norm_pre_clip': 10.517881393432617, 'timing/data': 0.00171711016446352, 'timing/model': 1.244118144735694, 'learning_rate/qwen_vl_interface': 4.680423543796385e-06, 'learning_rate/action_model': 4.680423543796385e-05, 'epoch': 1.47}) +[2026-06-18 01:08:56,375][starVLA.training.train_starvla][INFO] - Step 3815, Loss: {'train/loss': 0.24889758229255676, 'train/grad_norm_pre_clip': 9.536438941955566, 'timing/data': 0.001254752278327942, 'timing/model': 1.7710893554612994, 'learning_rate/qwen_vl_interface': 4.678276516848558e-06, 'learning_rate/action_model': 4.678276516848557e-05, 'epoch': 1.47}) +[2026-06-18 01:09:03,933][starVLA.training.train_starvla][INFO] - Step 3816, Loss: {'train/loss': 0.6016652584075928, 'train/grad_norm_pre_clip': 16.50948143005371, 'timing/data': 0.0010016094893217087, 'timing/model': 1.1617315588518977, 'learning_rate/qwen_vl_interface': 4.676129608419625e-06, 'learning_rate/action_model': 4.676129608419625e-05, 'epoch': 1.47}) +[2026-06-18 01:09:08,766][starVLA.training.train_starvla][INFO] - Step 3817, Loss: {'train/loss': 0.20488733053207397, 'train/grad_norm_pre_clip': 6.458158493041992, 'timing/data': 0.005452780053019524, 'timing/model': 1.5379936983808875, 'learning_rate/qwen_vl_interface': 4.6739828189546415e-06, 'learning_rate/action_model': 4.6739828189546417e-05, 'epoch': 1.47}) +[2026-06-18 01:09:13,534][starVLA.training.train_starvla][INFO] - Step 3818, Loss: {'train/loss': 0.520834743976593, 'train/grad_norm_pre_clip': 13.719574928283691, 'timing/data': 0.0011579487472772598, 'timing/model': 1.0973502835258842, 'learning_rate/qwen_vl_interface': 4.671836148898641e-06, 'learning_rate/action_model': 4.6718361488986404e-05, 'epoch': 1.47}) +[2026-06-18 01:09:18,665][starVLA.training.train_starvla][INFO] - Step 3819, Loss: {'train/loss': 0.31820520758628845, 'train/grad_norm_pre_clip': 8.065678596496582, 'timing/data': 0.0011083977296948433, 'timing/model': 1.1943023717030883, 'learning_rate/qwen_vl_interface': 4.6696895986966254e-06, 'learning_rate/action_model': 4.669689598696625e-05, 'epoch': 1.47}) +[2026-06-18 01:09:23,649][starVLA.training.train_starvla][INFO] - Step 3820, Loss: {'train/loss': 0.3708391487598419, 'train/grad_norm_pre_clip': 12.924357414245605, 'timing/data': 0.0011863354593515396, 'timing/model': 1.2440463546663523, 'learning_rate/qwen_vl_interface': 4.667543168793583e-06, 'learning_rate/action_model': 4.667543168793583e-05, 'epoch': 1.47}) +[2026-06-18 01:09:28,759][starVLA.training.train_starvla][INFO] - Step 3821, Loss: {'train/loss': 0.767760694026947, 'train/grad_norm_pre_clip': 19.56528091430664, 'timing/data': 0.0010471688583493233, 'timing/model': 1.161929765716195, 'learning_rate/qwen_vl_interface': 4.665396859634468e-06, 'learning_rate/action_model': 4.665396859634468e-05, 'epoch': 1.47}) +[2026-06-18 01:09:34,034][starVLA.training.train_starvla][INFO] - Step 3822, Loss: {'train/loss': 0.3277171552181244, 'train/grad_norm_pre_clip': 10.1578950881958, 'timing/data': 0.0009635183960199356, 'timing/model': 1.6124864825978875, 'learning_rate/qwen_vl_interface': 4.663250671664212e-06, 'learning_rate/action_model': 4.663250671664211e-05, 'epoch': 1.47}) +[2026-06-18 01:09:38,573][starVLA.training.train_starvla][INFO] - Step 3823, Loss: {'train/loss': 0.2876238226890564, 'train/grad_norm_pre_clip': 9.167182922363281, 'timing/data': 0.0012473175302147865, 'timing/model': 1.3000142443925142, 'learning_rate/qwen_vl_interface': 4.6611046053277245e-06, 'learning_rate/action_model': 4.661104605327724e-05, 'epoch': 1.47}) +[2026-06-18 01:09:43,807][starVLA.training.train_starvla][INFO] - Step 3824, Loss: {'train/loss': 0.3777664303779602, 'train/grad_norm_pre_clip': 9.234977722167969, 'timing/data': 0.001797742210328579, 'timing/model': 1.1770317740738392, 'learning_rate/qwen_vl_interface': 4.658958661069886e-06, 'learning_rate/action_model': 4.6589586610698856e-05, 'epoch': 1.47}) +[2026-06-18 01:09:49,071][starVLA.training.train_starvla][INFO] - Step 3825, Loss: {'train/loss': 0.5644692182540894, 'train/grad_norm_pre_clip': 13.604839324951172, 'timing/data': 0.0011578965932130814, 'timing/model': 1.4803198873996735, 'learning_rate/qwen_vl_interface': 4.6568128393355535e-06, 'learning_rate/action_model': 4.656812839335553e-05, 'epoch': 1.47}) +[2026-06-18 01:09:54,424][starVLA.training.train_starvla][INFO] - Step 3826, Loss: {'train/loss': 0.4023527503013611, 'train/grad_norm_pre_clip': 10.522778511047363, 'timing/data': 0.0011974787339568138, 'timing/model': 1.1908971769735217, 'learning_rate/qwen_vl_interface': 4.654667140569555e-06, 'learning_rate/action_model': 4.6546671405695544e-05, 'epoch': 1.47}) +[2026-06-18 01:09:59,212][starVLA.training.train_starvla][INFO] - Step 3827, Loss: {'train/loss': 0.4742342531681061, 'train/grad_norm_pre_clip': 17.480478286743164, 'timing/data': 0.0011705989018082619, 'timing/model': 1.550077316351235, 'learning_rate/qwen_vl_interface': 4.652521565216701e-06, 'learning_rate/action_model': 4.652521565216701e-05, 'epoch': 1.47}) +[2026-06-18 01:10:04,616][starVLA.training.train_starvla][INFO] - Step 3828, Loss: {'train/loss': 0.545945405960083, 'train/grad_norm_pre_clip': 14.482022285461426, 'timing/data': 0.0016993572935461998, 'timing/model': 1.6235719788819551, 'learning_rate/qwen_vl_interface': 4.65037611372177e-06, 'learning_rate/action_model': 4.650376113721769e-05, 'epoch': 1.47}) +[2026-06-18 01:10:10,488][starVLA.training.train_starvla][INFO] - Step 3829, Loss: {'train/loss': 0.296834796667099, 'train/grad_norm_pre_clip': 7.1396870613098145, 'timing/data': 0.0011812876909971237, 'timing/model': 1.1588373091071844, 'learning_rate/qwen_vl_interface': 4.648230786529514e-06, 'learning_rate/action_model': 4.648230786529513e-05, 'epoch': 1.47}) +[2026-06-18 01:10:15,583][starVLA.training.train_starvla][INFO] - Step 3830, Loss: {'train/loss': 0.13968637585639954, 'train/grad_norm_pre_clip': 8.218411445617676, 'timing/data': 0.001837143674492836, 'timing/model': 1.354850735515356, 'learning_rate/qwen_vl_interface': 4.646085584084664e-06, 'learning_rate/action_model': 4.6460855840846637e-05, 'epoch': 1.47}) +[2026-06-18 01:10:25,446][starVLA.training.train_starvla][INFO] - Step 3831, Loss: {'train/loss': 0.35952118039131165, 'train/grad_norm_pre_clip': 13.70102596282959, 'timing/data': 0.001130259595811367, 'timing/model': 1.3872729316353798, 'learning_rate/qwen_vl_interface': 4.643940506831923e-06, 'learning_rate/action_model': 4.6439405068319226e-05, 'epoch': 1.47}) +[2026-06-18 01:10:36,958][starVLA.training.train_starvla][INFO] - Step 3832, Loss: {'train/loss': 0.4383750259876251, 'train/grad_norm_pre_clip': 12.008755683898926, 'timing/data': 0.0014467621222138405, 'timing/model': 1.9090717220678926, 'learning_rate/qwen_vl_interface': 4.641795555215967e-06, 'learning_rate/action_model': 4.6417955552159676e-05, 'epoch': 1.47}) +[2026-06-18 01:10:43,097][starVLA.training.train_starvla][INFO] - Step 3833, Loss: {'train/loss': 0.2939387857913971, 'train/grad_norm_pre_clip': 9.7685546875, 'timing/data': 0.0011835843324661255, 'timing/model': 1.2176135005429387, 'learning_rate/qwen_vl_interface': 4.639650729681445e-06, 'learning_rate/action_model': 4.6396507296814454e-05, 'epoch': 1.47}) +[2026-06-18 01:10:58,837][starVLA.training.train_starvla][INFO] - Step 3834, Loss: {'train/loss': 0.37695613503456116, 'train/grad_norm_pre_clip': 11.467342376708984, 'timing/data': 0.0009636469185352325, 'timing/model': 3.3934729117900133, 'learning_rate/qwen_vl_interface': 4.637506030672987e-06, 'learning_rate/action_model': 4.637506030672987e-05, 'epoch': 1.47}) +[2026-06-18 01:11:12,254][starVLA.training.train_starvla][INFO] - Step 3835, Loss: {'train/loss': 0.3350788354873657, 'train/grad_norm_pre_clip': 12.058525085449219, 'timing/data': 0.0011865543201565742, 'timing/model': 1.2124923700466752, 'learning_rate/qwen_vl_interface': 4.635361458635186e-06, 'learning_rate/action_model': 4.6353614586351864e-05, 'epoch': 1.48}) +[2026-06-18 01:11:25,513][starVLA.training.train_starvla][INFO] - Step 3836, Loss: {'train/loss': 0.8241033554077148, 'train/grad_norm_pre_clip': 16.862598419189453, 'timing/data': 0.0014195498079061508, 'timing/model': 1.1599887320771813, 'learning_rate/qwen_vl_interface': 4.633217014012616e-06, 'learning_rate/action_model': 4.633217014012616e-05, 'epoch': 1.48}) +[2026-06-18 01:11:36,344][starVLA.training.train_starvla][INFO] - Step 3837, Loss: {'train/loss': 0.23131868243217468, 'train/grad_norm_pre_clip': 11.2691650390625, 'timing/data': 0.00108298659324646, 'timing/model': 1.8785156533122063, 'learning_rate/qwen_vl_interface': 4.631072697249822e-06, 'learning_rate/action_model': 4.631072697249822e-05, 'epoch': 1.48}) +[2026-06-18 01:11:49,381][starVLA.training.train_starvla][INFO] - Step 3838, Loss: {'train/loss': 0.47613656520843506, 'train/grad_norm_pre_clip': 10.535566329956055, 'timing/data': 0.0010709166526794434, 'timing/model': 4.427764942869544, 'learning_rate/qwen_vl_interface': 4.628928508791325e-06, 'learning_rate/action_model': 4.628928508791324e-05, 'epoch': 1.48}) +[2026-06-18 01:12:01,007][starVLA.training.train_starvla][INFO] - Step 3839, Loss: {'train/loss': 0.30513402819633484, 'train/grad_norm_pre_clip': 9.25069522857666, 'timing/data': 0.0012832274660468102, 'timing/model': 4.765247234143317, 'learning_rate/qwen_vl_interface': 4.626784449081617e-06, 'learning_rate/action_model': 4.626784449081617e-05, 'epoch': 1.48}) +[2026-06-18 01:12:10,499][starVLA.training.train_starvla][INFO] - Step 3840, Loss: {'train/loss': 0.2606929540634155, 'train/grad_norm_pre_clip': 10.527265548706055, 'timing/data': 0.0016924692317843437, 'timing/model': 1.1166478926315904, 'learning_rate/qwen_vl_interface': 4.62464051856516e-06, 'learning_rate/action_model': 4.6246405185651596e-05, 'epoch': 1.48}) +[2026-06-18 01:12:15,204][starVLA.training.train_starvla][INFO] - Step 3841, Loss: {'train/loss': 0.37616604566574097, 'train/grad_norm_pre_clip': 10.065650939941406, 'timing/data': 0.0012746108695864677, 'timing/model': 1.2128255423158407, 'learning_rate/qwen_vl_interface': 4.622496717686401e-06, 'learning_rate/action_model': 4.6224967176864005e-05, 'epoch': 1.48}) +[2026-06-18 01:12:20,029][starVLA.training.train_starvla][INFO] - Step 3842, Loss: {'train/loss': 0.13163243234157562, 'train/grad_norm_pre_clip': 7.676114559173584, 'timing/data': 0.0010023945942521095, 'timing/model': 1.5326860956847668, 'learning_rate/qwen_vl_interface': 4.620353046889745e-06, 'learning_rate/action_model': 4.620353046889745e-05, 'epoch': 1.48}) +[2026-06-18 01:12:24,872][starVLA.training.train_starvla][INFO] - Step 3843, Loss: {'train/loss': 0.33747604489326477, 'train/grad_norm_pre_clip': 15.01438045501709, 'timing/data': 0.001136678270995617, 'timing/model': 1.0922982217743993, 'learning_rate/qwen_vl_interface': 4.6182095066195805e-06, 'learning_rate/action_model': 4.61820950661958e-05, 'epoch': 1.48}) +[2026-06-18 01:12:29,779][starVLA.training.train_starvla][INFO] - Step 3844, Loss: {'train/loss': 0.5047286748886108, 'train/grad_norm_pre_clip': 14.907549858093262, 'timing/data': 0.0010059140622615814, 'timing/model': 1.2297292947769165, 'learning_rate/qwen_vl_interface': 4.616066097320264e-06, 'learning_rate/action_model': 4.6160660973202636e-05, 'epoch': 1.48}) +[2026-06-18 01:12:34,673][starVLA.training.train_starvla][INFO] - Step 3845, Loss: {'train/loss': 0.3884187638759613, 'train/grad_norm_pre_clip': 18.634967803955078, 'timing/data': 0.001582304947078228, 'timing/model': 1.1670722113922238, 'learning_rate/qwen_vl_interface': 4.613922819436129e-06, 'learning_rate/action_model': 4.613922819436128e-05, 'epoch': 1.48}) +[2026-06-18 01:12:39,786][starVLA.training.train_starvla][INFO] - Step 3846, Loss: {'train/loss': 0.516636848449707, 'train/grad_norm_pre_clip': 11.692873001098633, 'timing/data': 0.001241760328412056, 'timing/model': 1.3023566408082843, 'learning_rate/qwen_vl_interface': 4.611779673411478e-06, 'learning_rate/action_model': 4.6117796734114785e-05, 'epoch': 1.48}) +[2026-06-18 01:12:44,898][starVLA.training.train_starvla][INFO] - Step 3847, Loss: {'train/loss': 0.34225839376449585, 'train/grad_norm_pre_clip': 13.500819206237793, 'timing/data': 0.001245102845132351, 'timing/model': 1.6092038368806243, 'learning_rate/qwen_vl_interface': 4.609636659690588e-06, 'learning_rate/action_model': 4.609636659690587e-05, 'epoch': 1.48}) +[2026-06-18 01:12:59,470][starVLA.training.train_starvla][INFO] - Step 3848, Loss: {'train/loss': 0.3621705174446106, 'train/grad_norm_pre_clip': 9.225696563720703, 'timing/data': 0.0012162616476416588, 'timing/model': 1.2269427012652159, 'learning_rate/qwen_vl_interface': 4.607493778717705e-06, 'learning_rate/action_model': 4.6074937787177055e-05, 'epoch': 1.48}) +[2026-06-18 01:13:04,992][starVLA.training.train_starvla][INFO] - Step 3849, Loss: {'train/loss': 0.39861273765563965, 'train/grad_norm_pre_clip': 10.852133750915527, 'timing/data': 0.0014123274013400078, 'timing/model': 1.1672109700739384, 'learning_rate/qwen_vl_interface': 4.605351030937054e-06, 'learning_rate/action_model': 4.605351030937055e-05, 'epoch': 1.48}) +[2026-06-18 01:13:18,246][starVLA.training.train_starvla][INFO] - Step 3850, Loss: {'train/loss': 0.5888387560844421, 'train/grad_norm_pre_clip': 14.379989624023438, 'timing/data': 0.0009863870218396187, 'timing/model': 5.8269841354340315, 'learning_rate/qwen_vl_interface': 4.60320841679283e-06, 'learning_rate/action_model': 4.60320841679283e-05, 'epoch': 1.48}) +[2026-06-18 01:13:24,261][starVLA.training.train_starvla][INFO] - Step 3851, Loss: {'train/loss': 0.3757559657096863, 'train/grad_norm_pre_clip': 13.988725662231445, 'timing/data': 0.0014069825410842896, 'timing/model': 1.1856673453003168, 'learning_rate/qwen_vl_interface': 4.6010659367291945e-06, 'learning_rate/action_model': 4.601065936729194e-05, 'epoch': 1.48}) +[2026-06-18 01:13:29,454][starVLA.training.train_starvla][INFO] - Step 3852, Loss: {'train/loss': 0.43144407868385315, 'train/grad_norm_pre_clip': 13.072578430175781, 'timing/data': 0.0012007951736450195, 'timing/model': 1.5693162754178047, 'learning_rate/qwen_vl_interface': 4.598923591190289e-06, 'learning_rate/action_model': 4.598923591190289e-05, 'epoch': 1.48}) +[2026-06-18 01:13:34,342][starVLA.training.train_starvla][INFO] - Step 3853, Loss: {'train/loss': 0.6004168391227722, 'train/grad_norm_pre_clip': 12.559172630310059, 'timing/data': 0.0009143752977252007, 'timing/model': 1.1239272160455585, 'learning_rate/qwen_vl_interface': 4.596781380620225e-06, 'learning_rate/action_model': 4.5967813806202245e-05, 'epoch': 1.48}) +[2026-06-18 01:13:44,321][starVLA.training.train_starvla][INFO] - Step 3854, Loss: {'train/loss': 0.3923869729042053, 'train/grad_norm_pre_clip': 18.221773147583008, 'timing/data': 0.002186821773648262, 'timing/model': 1.564806953072548, 'learning_rate/qwen_vl_interface': 4.5946393054630845e-06, 'learning_rate/action_model': 4.5946393054630845e-05, 'epoch': 1.48}) +[2026-06-18 01:13:49,267][starVLA.training.train_starvla][INFO] - Step 3855, Loss: {'train/loss': 0.3340435028076172, 'train/grad_norm_pre_clip': 13.136740684509277, 'timing/data': 0.0012949667870998383, 'timing/model': 1.1345274103805423, 'learning_rate/qwen_vl_interface': 4.592497366162919e-06, 'learning_rate/action_model': 4.592497366162918e-05, 'epoch': 1.48}) +[2026-06-18 01:13:54,539][starVLA.training.train_starvla][INFO] - Step 3856, Loss: {'train/loss': 0.33720192313194275, 'train/grad_norm_pre_clip': 8.32614517211914, 'timing/data': 0.0012357169762253761, 'timing/model': 1.2208922123536468, 'learning_rate/qwen_vl_interface': 4.590355563163758e-06, 'learning_rate/action_model': 4.5903555631637574e-05, 'epoch': 1.48}) +[2026-06-18 01:14:03,317][starVLA.training.train_starvla][INFO] - Step 3857, Loss: {'train/loss': 0.38366082310676575, 'train/grad_norm_pre_clip': 12.475927352905273, 'timing/data': 0.0011383919045329094, 'timing/model': 1.7415958838537335, 'learning_rate/qwen_vl_interface': 4.588213896909598e-06, 'learning_rate/action_model': 4.588213896909598e-05, 'epoch': 1.48}) +[2026-06-18 01:14:09,842][starVLA.training.train_starvla][INFO] - Step 3858, Loss: {'train/loss': 0.7089166045188904, 'train/grad_norm_pre_clip': 11.233586311340332, 'timing/data': 0.0020199958235025406, 'timing/model': 1.6562983086332679, 'learning_rate/qwen_vl_interface': 4.58607236784441e-06, 'learning_rate/action_model': 4.58607236784441e-05, 'epoch': 1.48}) +[2026-06-18 01:14:15,441][starVLA.training.train_starvla][INFO] - Step 3859, Loss: {'train/loss': 0.3351449966430664, 'train/grad_norm_pre_clip': 9.88525104522705, 'timing/data': 0.000830397941172123, 'timing/model': 1.0920988637953997, 'learning_rate/qwen_vl_interface': 4.583930976412133e-06, 'learning_rate/action_model': 4.583930976412133e-05, 'epoch': 1.48}) +[2026-06-18 01:14:21,169][starVLA.training.train_starvla][INFO] - Step 3860, Loss: {'train/loss': 0.46074098348617554, 'train/grad_norm_pre_clip': 14.547985076904297, 'timing/data': 0.0011788075789809227, 'timing/model': 1.3773023402318358, 'learning_rate/qwen_vl_interface': 4.5817897230566834e-06, 'learning_rate/action_model': 4.581789723056683e-05, 'epoch': 1.48}) +[2026-06-18 01:14:26,649][starVLA.training.train_starvla][INFO] - Step 3861, Loss: {'train/loss': 0.5551218390464783, 'train/grad_norm_pre_clip': 10.859389305114746, 'timing/data': 0.0009536715224385262, 'timing/model': 1.8115760339424014, 'learning_rate/qwen_vl_interface': 4.579648608221943e-06, 'learning_rate/action_model': 4.579648608221943e-05, 'epoch': 1.49}) +[2026-06-18 01:14:35,588][starVLA.training.train_starvla][INFO] - Step 3862, Loss: {'train/loss': 0.4307776987552643, 'train/grad_norm_pre_clip': 12.410330772399902, 'timing/data': 0.0014909971505403519, 'timing/model': 1.5362040204927325, 'learning_rate/qwen_vl_interface': 4.577507632351766e-06, 'learning_rate/action_model': 4.5775076323517664e-05, 'epoch': 1.49}) +[2026-06-18 01:14:40,393][starVLA.training.train_starvla][INFO] - Step 3863, Loss: {'train/loss': 0.35641294717788696, 'train/grad_norm_pre_clip': 12.115827560424805, 'timing/data': 0.005924016237258911, 'timing/model': 1.2815657770261168, 'learning_rate/qwen_vl_interface': 4.575366795889981e-06, 'learning_rate/action_model': 4.575366795889981e-05, 'epoch': 1.49}) +[2026-06-18 01:14:45,643][starVLA.training.train_starvla][INFO] - Step 3864, Loss: {'train/loss': 0.23011991381645203, 'train/grad_norm_pre_clip': 6.376652717590332, 'timing/data': 0.0010923072695732117, 'timing/model': 1.244143039919436, 'learning_rate/qwen_vl_interface': 4.573226099280386e-06, 'learning_rate/action_model': 4.573226099280386e-05, 'epoch': 1.49}) +[2026-06-18 01:14:50,855][starVLA.training.train_starvla][INFO] - Step 3865, Loss: {'train/loss': 0.4912755489349365, 'train/grad_norm_pre_clip': 11.988905906677246, 'timing/data': 0.0011603040620684624, 'timing/model': 1.0998249473050237, 'learning_rate/qwen_vl_interface': 4.571085542966748e-06, 'learning_rate/action_model': 4.5710855429667485e-05, 'epoch': 1.49}) +[2026-06-18 01:14:56,389][starVLA.training.train_starvla][INFO] - Step 3866, Loss: {'train/loss': 0.6333687901496887, 'train/grad_norm_pre_clip': 15.457867622375488, 'timing/data': 0.0013343719765543938, 'timing/model': 1.3781526908278465, 'learning_rate/qwen_vl_interface': 4.568945127392807e-06, 'learning_rate/action_model': 4.5689451273928066e-05, 'epoch': 1.49}) +[2026-06-18 01:15:03,172][starVLA.training.train_starvla][INFO] - Step 3867, Loss: {'train/loss': 0.35443103313446045, 'train/grad_norm_pre_clip': 11.39119815826416, 'timing/data': 0.0011117244139313698, 'timing/model': 3.1846424490213394, 'learning_rate/qwen_vl_interface': 4.566804853002275e-06, 'learning_rate/action_model': 4.566804853002275e-05, 'epoch': 1.49}) +[2026-06-18 01:15:07,980][starVLA.training.train_starvla][INFO] - Step 3868, Loss: {'train/loss': 0.3540756106376648, 'train/grad_norm_pre_clip': 11.879884719848633, 'timing/data': 0.001203681342303753, 'timing/model': 1.3149129208177328, 'learning_rate/qwen_vl_interface': 4.564664720238833e-06, 'learning_rate/action_model': 4.564664720238833e-05, 'epoch': 1.49}) +[2026-06-18 01:15:12,995][starVLA.training.train_starvla][INFO] - Step 3869, Loss: {'train/loss': 0.4710375964641571, 'train/grad_norm_pre_clip': 13.944787979125977, 'timing/data': 0.0010141758248209953, 'timing/model': 1.1449567945674062, 'learning_rate/qwen_vl_interface': 4.5625247295461314e-06, 'learning_rate/action_model': 4.562524729546131e-05, 'epoch': 1.49}) +[2026-06-18 01:15:21,134][starVLA.training.train_starvla][INFO] - Step 3870, Loss: {'train/loss': 0.2794463634490967, 'train/grad_norm_pre_clip': 8.39168930053711, 'timing/data': 0.0018388880416750908, 'timing/model': 4.357311952859163, 'learning_rate/qwen_vl_interface': 4.560384881367791e-06, 'learning_rate/action_model': 4.5603848813677904e-05, 'epoch': 1.49}) +[2026-06-18 01:15:29,598][starVLA.training.train_starvla][INFO] - Step 3871, Loss: {'train/loss': 0.33018583059310913, 'train/grad_norm_pre_clip': 10.722493171691895, 'timing/data': 0.0013074381276965141, 'timing/model': 1.0735088912770152, 'learning_rate/qwen_vl_interface': 4.5582451761474085e-06, 'learning_rate/action_model': 4.558245176147408e-05, 'epoch': 1.49}) +[2026-06-18 01:15:34,155][starVLA.training.train_starvla][INFO] - Step 3872, Loss: {'train/loss': 0.23332862555980682, 'train/grad_norm_pre_clip': 10.905771255493164, 'timing/data': 0.0019248481839895248, 'timing/model': 1.4335515620186925, 'learning_rate/qwen_vl_interface': 4.556105614328545e-06, 'learning_rate/action_model': 4.5561056143285455e-05, 'epoch': 1.49}) +[2026-06-18 01:15:38,439][starVLA.training.train_starvla][INFO] - Step 3873, Loss: {'train/loss': 0.5550154447555542, 'train/grad_norm_pre_clip': 14.131096839904785, 'timing/data': 0.0012314626947045326, 'timing/model': 1.1056975293904543, 'learning_rate/qwen_vl_interface': 4.553966196354734e-06, 'learning_rate/action_model': 4.553966196354734e-05, 'epoch': 1.49}) +[2026-06-18 01:15:42,795][starVLA.training.train_starvla][INFO] - Step 3874, Loss: {'train/loss': 0.41370177268981934, 'train/grad_norm_pre_clip': 12.414050102233887, 'timing/data': 0.0010385075584053993, 'timing/model': 1.045237977989018, 'learning_rate/qwen_vl_interface': 4.55182692266948e-06, 'learning_rate/action_model': 4.551826922669479e-05, 'epoch': 1.49}) +[2026-06-18 01:15:47,283][starVLA.training.train_starvla][INFO] - Step 3875, Loss: {'train/loss': 0.49425607919692993, 'train/grad_norm_pre_clip': 11.631458282470703, 'timing/data': 0.0011471500620245934, 'timing/model': 1.068579701706767, 'learning_rate/qwen_vl_interface': 4.549687793716257e-06, 'learning_rate/action_model': 4.549687793716257e-05, 'epoch': 1.49}) +[2026-06-18 01:15:53,087][starVLA.training.train_starvla][INFO] - Step 3876, Loss: {'train/loss': 0.3135441839694977, 'train/grad_norm_pre_clip': 15.167033195495605, 'timing/data': 0.0010864222422242165, 'timing/model': 1.0806201491504908, 'learning_rate/qwen_vl_interface': 4.547548809938507e-06, 'learning_rate/action_model': 4.5475488099385075e-05, 'epoch': 1.49}) +[2026-06-18 01:15:57,451][starVLA.training.train_starvla][INFO] - Step 3877, Loss: {'train/loss': 0.22238892316818237, 'train/grad_norm_pre_clip': 5.786613464355469, 'timing/data': 0.0010291319340467453, 'timing/model': 1.3127717673778534, 'learning_rate/qwen_vl_interface': 4.545409971779644e-06, 'learning_rate/action_model': 4.545409971779644e-05, 'epoch': 1.49}) +[2026-06-18 01:16:01,696][starVLA.training.train_starvla][INFO] - Step 3878, Loss: {'train/loss': 0.5387550592422485, 'train/grad_norm_pre_clip': 14.189461708068848, 'timing/data': 0.0011828141286969185, 'timing/model': 1.059030237607658, 'learning_rate/qwen_vl_interface': 4.543271279683052e-06, 'learning_rate/action_model': 4.543271279683053e-05, 'epoch': 1.49}) +[2026-06-18 01:16:06,079][starVLA.training.train_starvla][INFO] - Step 3879, Loss: {'train/loss': 0.3864275813102722, 'train/grad_norm_pre_clip': 12.345856666564941, 'timing/data': 0.001431693322956562, 'timing/model': 1.1294510811567307, 'learning_rate/qwen_vl_interface': 4.541132734092086e-06, 'learning_rate/action_model': 4.541132734092086e-05, 'epoch': 1.49}) +[2026-06-18 01:16:10,609][starVLA.training.train_starvla][INFO] - Step 3880, Loss: {'train/loss': 0.17062252759933472, 'train/grad_norm_pre_clip': 7.2318267822265625, 'timing/data': 0.0010723872110247612, 'timing/model': 1.0667983023449779, 'learning_rate/qwen_vl_interface': 4.538994335450066e-06, 'learning_rate/action_model': 4.5389943354500664e-05, 'epoch': 1.49}) +[2026-06-18 01:16:15,055][starVLA.training.train_starvla][INFO] - Step 3881, Loss: {'train/loss': 0.1816035658121109, 'train/grad_norm_pre_clip': 5.672389030456543, 'timing/data': 0.0010104905813932419, 'timing/model': 1.097348626703024, 'learning_rate/qwen_vl_interface': 4.5368560842002836e-06, 'learning_rate/action_model': 4.5368560842002834e-05, 'epoch': 1.49}) +[2026-06-18 01:16:20,956][starVLA.training.train_starvla][INFO] - Step 3882, Loss: {'train/loss': 0.5110427141189575, 'train/grad_norm_pre_clip': 10.364395141601562, 'timing/data': 0.0007686642929911613, 'timing/model': 1.337093461304903, 'learning_rate/qwen_vl_interface': 4.534717980786005e-06, 'learning_rate/action_model': 4.5347179807860046e-05, 'epoch': 1.49}) +[2026-06-18 01:16:25,207][starVLA.training.train_starvla][INFO] - Step 3883, Loss: {'train/loss': 0.32382333278656006, 'train/grad_norm_pre_clip': 9.868377685546875, 'timing/data': 0.0010673534125089645, 'timing/model': 1.0655078357085586, 'learning_rate/qwen_vl_interface': 4.532580025650457e-06, 'learning_rate/action_model': 4.5325800256504574e-05, 'epoch': 1.49}) +[2026-06-18 01:16:29,858][starVLA.training.train_starvla][INFO] - Step 3884, Loss: {'train/loss': 0.48201265931129456, 'train/grad_norm_pre_clip': 16.337749481201172, 'timing/data': 0.0012085642665624619, 'timing/model': 1.1067386697977781, 'learning_rate/qwen_vl_interface': 4.530442219236842e-06, 'learning_rate/action_model': 4.530442219236842e-05, 'epoch': 1.49}) +[2026-06-18 01:16:40,679][starVLA.training.train_starvla][INFO] - Step 3885, Loss: {'train/loss': 0.23602792620658875, 'train/grad_norm_pre_clip': 7.46506929397583, 'timing/data': 0.0008368520066142082, 'timing/model': 4.12139676976949, 'learning_rate/qwen_vl_interface': 4.5283045619883285e-06, 'learning_rate/action_model': 4.5283045619883285e-05, 'epoch': 1.49}) +[2026-06-18 01:16:45,215][starVLA.training.train_starvla][INFO] - Step 3886, Loss: {'train/loss': 0.2534891366958618, 'train/grad_norm_pre_clip': 9.565668106079102, 'timing/data': 0.0010278578847646713, 'timing/model': 1.0672392882406712, 'learning_rate/qwen_vl_interface': 4.526167054348057e-06, 'learning_rate/action_model': 4.526167054348057e-05, 'epoch': 1.49}) +[2026-06-18 01:16:49,860][starVLA.training.train_starvla][INFO] - Step 3887, Loss: {'train/loss': 0.32614633440971375, 'train/grad_norm_pre_clip': 11.81032943725586, 'timing/data': 0.0010839765891432762, 'timing/model': 1.3913007183000445, 'learning_rate/qwen_vl_interface': 4.524029696759135e-06, 'learning_rate/action_model': 4.5240296967591354e-05, 'epoch': 1.5}) +[2026-06-18 01:16:54,080][starVLA.training.train_starvla][INFO] - Step 3888, Loss: {'train/loss': 0.3935157358646393, 'train/grad_norm_pre_clip': 11.350242614746094, 'timing/data': 0.0012312913313508034, 'timing/model': 1.062092331238091, 'learning_rate/qwen_vl_interface': 4.5218924896646354e-06, 'learning_rate/action_model': 4.521892489664635e-05, 'epoch': 1.5}) +[2026-06-18 01:16:58,391][starVLA.training.train_starvla][INFO] - Step 3889, Loss: {'train/loss': 0.8388079404830933, 'train/grad_norm_pre_clip': 18.486228942871094, 'timing/data': 0.0010813567787408829, 'timing/model': 1.0591056002303958, 'learning_rate/qwen_vl_interface': 4.519755433507609e-06, 'learning_rate/action_model': 4.5197554335076086e-05, 'epoch': 1.5}) +[2026-06-18 01:17:02,785][starVLA.training.train_starvla][INFO] - Step 3890, Loss: {'train/loss': 0.293841153383255, 'train/grad_norm_pre_clip': 9.701698303222656, 'timing/data': 0.001089438796043396, 'timing/model': 1.0885405177250504, 'learning_rate/qwen_vl_interface': 4.517618528731066e-06, 'learning_rate/action_model': 4.517618528731065e-05, 'epoch': 1.5}) +[2026-06-18 01:17:07,482][starVLA.training.train_starvla][INFO] - Step 3891, Loss: {'train/loss': 0.29628700017929077, 'train/grad_norm_pre_clip': 8.62161636352539, 'timing/data': 0.0015697553753852844, 'timing/model': 1.0814442755654454, 'learning_rate/qwen_vl_interface': 4.515481775777989e-06, 'learning_rate/action_model': 4.515481775777988e-05, 'epoch': 1.5}) +[2026-06-18 01:17:12,898][starVLA.training.train_starvla][INFO] - Step 3892, Loss: {'train/loss': 0.3284218907356262, 'train/grad_norm_pre_clip': 12.389908790588379, 'timing/data': 0.0010127155110239983, 'timing/model': 1.3310861820355058, 'learning_rate/qwen_vl_interface': 4.51334517509133e-06, 'learning_rate/action_model': 4.5133451750913305e-05, 'epoch': 1.5}) +[2026-06-18 01:17:16,985][starVLA.training.train_starvla][INFO] - Step 3893, Loss: {'train/loss': 0.5047500729560852, 'train/grad_norm_pre_clip': 13.348471641540527, 'timing/data': 0.0010139290243387222, 'timing/model': 1.0635518059134483, 'learning_rate/qwen_vl_interface': 4.511208727114011e-06, 'learning_rate/action_model': 4.51120872711401e-05, 'epoch': 1.5}) +[2026-06-18 01:17:21,652][starVLA.training.train_starvla][INFO] - Step 3894, Loss: {'train/loss': 0.21342600882053375, 'train/grad_norm_pre_clip': 7.945007801055908, 'timing/data': 0.0010143909603357315, 'timing/model': 1.1819903729483485, 'learning_rate/qwen_vl_interface': 4.5090724322889175e-06, 'learning_rate/action_model': 4.5090724322889174e-05, 'epoch': 1.5}) +[2026-06-18 01:17:26,077][starVLA.training.train_starvla][INFO] - Step 3895, Loss: {'train/loss': 0.22101224958896637, 'train/grad_norm_pre_clip': 8.594484329223633, 'timing/data': 0.0011777868494391441, 'timing/model': 1.060001665726304, 'learning_rate/qwen_vl_interface': 4.506936291058902e-06, 'learning_rate/action_model': 4.506936291058902e-05, 'epoch': 1.5}) +[2026-06-18 01:17:30,414][starVLA.training.train_starvla][INFO] - Step 3896, Loss: {'train/loss': 0.5520309209823608, 'train/grad_norm_pre_clip': 13.651054382324219, 'timing/data': 0.0009759813547134399, 'timing/model': 1.059568484313786, 'learning_rate/qwen_vl_interface': 4.504800303866798e-06, 'learning_rate/action_model': 4.504800303866798e-05, 'epoch': 1.5}) +[2026-06-18 01:17:41,754][starVLA.training.train_starvla][INFO] - Step 3897, Loss: {'train/loss': 0.6682601571083069, 'train/grad_norm_pre_clip': 26.818819046020508, 'timing/data': 0.0009666988626122475, 'timing/model': 1.672782407142222, 'learning_rate/qwen_vl_interface': 4.50266447115539e-06, 'learning_rate/action_model': 4.5026644711553904e-05, 'epoch': 1.5}) +[2026-06-18 01:17:49,293][starVLA.training.train_starvla][INFO] - Step 3898, Loss: {'train/loss': 0.416265070438385, 'train/grad_norm_pre_clip': 14.216934204101562, 'timing/data': 0.0009107636287808418, 'timing/model': 2.5188589952886105, 'learning_rate/qwen_vl_interface': 4.500528793367444e-06, 'learning_rate/action_model': 4.5005287933674435e-05, 'epoch': 1.5}) +[2026-06-18 01:17:55,151][starVLA.training.train_starvla][INFO] - Step 3899, Loss: {'train/loss': 0.6475815176963806, 'train/grad_norm_pre_clip': 12.871935844421387, 'timing/data': 0.0013339081779122353, 'timing/model': 1.121145878918469, 'learning_rate/qwen_vl_interface': 4.498393270945683e-06, 'learning_rate/action_model': 4.498393270945683e-05, 'epoch': 1.5}) +[2026-06-18 01:23:01,620][starVLA.training.train_starvla][INFO] - Step 3900, Loss: {'train/loss': 0.13361164927482605, 'train/grad_norm_pre_clip': 6.953939437866211, 'eval/loss': 0.47725462913513184, 'eval/latency_4/loss': 0.47725462913513184, 'eval/demon_fire_precision': 0.9550842170929508, 'eval/demon_fire_recall': 0.9944787268593699, 'eval/demon_fire_f1': 0.9743834526650756, 'eval/demon_fire_rate_model': 0.9415565345080763, 'eval/demon_fire_rate_teacher': 0.9042584434654919, 'eval/demon_left_precision': 0.8715953307392996, 'eval/demon_left_recall': 0.5114155251141552, 'eval/demon_left_f1': 0.6446043165467625, 'eval/demon_left_rate_model': 0.07547723935389133, 'eval/demon_left_rate_teacher': 0.12863436123348018, 'eval/demon_right_precision': 0.8292682926829268, 'eval/demon_right_recall': 0.7364620938628159, 'eval/demon_right_f1': 0.780114722753346, 'eval/demon_right_rate_model': 0.1444933920704846, 'eval/demon_right_rate_teacher': 0.16270190895741557, 'eval/demon_move_f1': 0.7123595196500543, 'eval/demon_cc_f1': 0.8775481049788179, 'eval/latency_4/demon_fire_precision': 0.9550842170929508, 'eval/latency_4/demon_fire_recall': 0.9944787268593699, 'eval/latency_4/demon_fire_f1': 0.9743834526650756, 'eval/latency_4/demon_fire_rate_model': 0.9415565345080763, 'eval/latency_4/demon_fire_rate_teacher': 0.9042584434654919, 'eval/latency_4/demon_left_precision': 0.8715953307392996, 'eval/latency_4/demon_left_recall': 0.5114155251141552, 'eval/latency_4/demon_left_f1': 0.6446043165467625, 'eval/latency_4/demon_left_rate_model': 0.07547723935389133, 'eval/latency_4/demon_left_rate_teacher': 0.12863436123348018, 'eval/latency_4/demon_right_precision': 0.8292682926829268, 'eval/latency_4/demon_right_recall': 0.7364620938628159, 'eval/latency_4/demon_right_f1': 0.780114722753346, 'eval/latency_4/demon_right_rate_model': 0.1444933920704846, 'eval/latency_4/demon_right_rate_teacher': 0.16270190895741557, 'eval/latency_4/demon_move_f1': 0.7123595196500543, 'eval/latency_4/demon_cc_f1': 0.8775481049788179, 'timing/data': 0.0012013111263513565, 'timing/model': 1.2657572133466601, 'learning_rate/qwen_vl_interface': 4.496257904332806e-06, 'learning_rate/action_model': 4.4962579043328064e-05, 'epoch': 1.5}) +[2026-06-18 01:23:01,622][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_3900_state +[2026-06-18 01:23:01,630][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 01:23:11,319][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_3900_state/model.safetensors +[2026-06-18 01:23:36,001][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_3900_state/optimizer.bin +[2026-06-18 01:23:36,004][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_3900_state/sampler.bin +[2026-06-18 01:23:36,005][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_3900_state/sampler_1.bin +[2026-06-18 01:23:36,030][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_3900_state/random_states_0.pkl +[2026-06-18 01:23:36,036][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3900 +[2026-06-18 01:23:36,037][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 01:23:36,065][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 01:23:49,998][starVLA.training.train_starvla][INFO] - Step 3901, Loss: {'train/loss': 0.13297197222709656, 'train/grad_norm_pre_clip': 9.069005966186523, 'timing/data': 0.001410938799381256, 'timing/model': 1.991197269409895, 'learning_rate/qwen_vl_interface': 4.494122693971478e-06, 'learning_rate/action_model': 4.4941226939714785e-05, 'epoch': 1.5}) +[2026-06-18 01:24:01,134][starVLA.training.train_starvla][INFO] - Step 3902, Loss: {'train/loss': 0.4749407172203064, 'train/grad_norm_pre_clip': 14.743481636047363, 'timing/data': 0.0015767691656947136, 'timing/model': 4.439826842397451, 'learning_rate/qwen_vl_interface': 4.491987640304327e-06, 'learning_rate/action_model': 4.4919876403043265e-05, 'epoch': 1.5}) +[2026-06-18 01:24:11,466][starVLA.training.train_starvla][INFO] - Step 3903, Loss: {'train/loss': 0.40610015392303467, 'train/grad_norm_pre_clip': 14.232991218566895, 'timing/data': 0.0010422607883810997, 'timing/model': 4.771642062813044, 'learning_rate/qwen_vl_interface': 4.489852743773955e-06, 'learning_rate/action_model': 4.489852743773955e-05, 'epoch': 1.5}) +[2026-06-18 01:24:19,927][starVLA.training.train_starvla][INFO] - Step 3904, Loss: {'train/loss': 0.4333858788013458, 'train/grad_norm_pre_clip': 15.661114692687988, 'timing/data': 0.0009241057559847832, 'timing/model': 1.1713007325306535, 'learning_rate/qwen_vl_interface': 4.487718004822928e-06, 'learning_rate/action_model': 4.487718004822928e-05, 'epoch': 1.5}) +[2026-06-18 01:24:25,320][starVLA.training.train_starvla][INFO] - Step 3905, Loss: {'train/loss': 0.3158806562423706, 'train/grad_norm_pre_clip': 12.598447799682617, 'timing/data': 0.0011961385607719421, 'timing/model': 1.2802700884640217, 'learning_rate/qwen_vl_interface': 4.485583423893778e-06, 'learning_rate/action_model': 4.4855834238937774e-05, 'epoch': 1.5}) +[2026-06-18 01:24:30,650][starVLA.training.train_starvla][INFO] - Step 3906, Loss: {'train/loss': 0.4017561674118042, 'train/grad_norm_pre_clip': 11.237751007080078, 'timing/data': 0.0011292574927210808, 'timing/model': 1.5337610095739365, 'learning_rate/qwen_vl_interface': 4.483449001429005e-06, 'learning_rate/action_model': 4.483449001429005e-05, 'epoch': 1.5}) +[2026-06-18 01:24:35,850][starVLA.training.train_starvla][INFO] - Step 3907, Loss: {'train/loss': 0.6342476010322571, 'train/grad_norm_pre_clip': 20.3480281829834, 'timing/data': 0.0012665707617998123, 'timing/model': 1.1026307260617614, 'learning_rate/qwen_vl_interface': 4.481314737871081e-06, 'learning_rate/action_model': 4.4813147378710804e-05, 'epoch': 1.5}) +[2026-06-18 01:24:42,114][starVLA.training.train_starvla][INFO] - Step 3908, Loss: {'train/loss': 0.4406364858150482, 'train/grad_norm_pre_clip': 15.510210037231445, 'timing/data': 0.0009106658399105072, 'timing/model': 1.6243762401863933, 'learning_rate/qwen_vl_interface': 4.47918063366244e-06, 'learning_rate/action_model': 4.4791806336624395e-05, 'epoch': 1.5}) +[2026-06-18 01:24:47,263][starVLA.training.train_starvla][INFO] - Step 3909, Loss: {'train/loss': 0.5183373093605042, 'train/grad_norm_pre_clip': 13.105857849121094, 'timing/data': 0.0014488361775875092, 'timing/model': 1.1347818514332175, 'learning_rate/qwen_vl_interface': 4.477046689245481e-06, 'learning_rate/action_model': 4.477046689245481e-05, 'epoch': 1.5}) +[2026-06-18 01:24:53,358][starVLA.training.train_starvla][INFO] - Step 3910, Loss: {'train/loss': 0.3093932569026947, 'train/grad_norm_pre_clip': 10.818406105041504, 'timing/data': 0.001285189762711525, 'timing/model': 1.4625476449728012, 'learning_rate/qwen_vl_interface': 4.474912905062575e-06, 'learning_rate/action_model': 4.474912905062575e-05, 'epoch': 1.5}) +[2026-06-18 01:24:58,693][starVLA.training.train_starvla][INFO] - Step 3911, Loss: {'train/loss': 0.3561048209667206, 'train/grad_norm_pre_clip': 10.380424499511719, 'timing/data': 0.0013910699635744095, 'timing/model': 1.5976651953533292, 'learning_rate/qwen_vl_interface': 4.472779281556059e-06, 'learning_rate/action_model': 4.4727792815560585e-05, 'epoch': 1.5}) +[2026-06-18 01:25:03,808][starVLA.training.train_starvla][INFO] - Step 3912, Loss: {'train/loss': 0.32993456721305847, 'train/grad_norm_pre_clip': 9.023103713989258, 'timing/data': 0.0013517681509256363, 'timing/model': 1.3427155949175358, 'learning_rate/qwen_vl_interface': 4.470645819168233e-06, 'learning_rate/action_model': 4.4706458191682324e-05, 'epoch': 1.5}) +[2026-06-18 01:25:09,993][starVLA.training.train_starvla][INFO] - Step 3913, Loss: {'train/loss': 0.35059142112731934, 'train/grad_norm_pre_clip': 9.890206336975098, 'timing/data': 0.0012675123289227486, 'timing/model': 1.6968074413016438, 'learning_rate/qwen_vl_interface': 4.468512518341367e-06, 'learning_rate/action_model': 4.4685125183413674e-05, 'epoch': 1.51}) +[2026-06-18 01:25:14,936][starVLA.training.train_starvla][INFO] - Step 3914, Loss: {'train/loss': 0.321612685918808, 'train/grad_norm_pre_clip': 11.23249340057373, 'timing/data': 0.009159771725535393, 'timing/model': 1.2253373758867383, 'learning_rate/qwen_vl_interface': 4.466379379517699e-06, 'learning_rate/action_model': 4.4663793795176985e-05, 'epoch': 1.51}) +[2026-06-18 01:25:20,108][starVLA.training.train_starvla][INFO] - Step 3915, Loss: {'train/loss': 0.29009345173835754, 'train/grad_norm_pre_clip': 11.03592300415039, 'timing/data': 0.0016591008752584457, 'timing/model': 1.3269506953656673, 'learning_rate/qwen_vl_interface': 4.464246403139429e-06, 'learning_rate/action_model': 4.464246403139428e-05, 'epoch': 1.51}) +[2026-06-18 01:25:25,867][starVLA.training.train_starvla][INFO] - Step 3916, Loss: {'train/loss': 0.3689986765384674, 'train/grad_norm_pre_clip': 10.329204559326172, 'timing/data': 0.0013044057413935661, 'timing/model': 1.6709658969193697, 'learning_rate/qwen_vl_interface': 4.462113589648723e-06, 'learning_rate/action_model': 4.462113589648723e-05, 'epoch': 1.51}) +[2026-06-18 01:25:31,299][starVLA.training.train_starvla][INFO] - Step 3917, Loss: {'train/loss': 0.371194988489151, 'train/grad_norm_pre_clip': 12.243491172790527, 'timing/data': 0.0012710914015769958, 'timing/model': 1.4489523563534021, 'learning_rate/qwen_vl_interface': 4.459980939487718e-06, 'learning_rate/action_model': 4.4599809394877176e-05, 'epoch': 1.51}) +[2026-06-18 01:25:37,311][starVLA.training.train_starvla][INFO] - Step 3918, Loss: {'train/loss': 0.33749037981033325, 'train/grad_norm_pre_clip': 10.17336368560791, 'timing/data': 0.001304621808230877, 'timing/model': 2.441071745939553, 'learning_rate/qwen_vl_interface': 4.457848453098514e-06, 'learning_rate/action_model': 4.4578484530985144e-05, 'epoch': 1.51}) +[2026-06-18 01:25:45,735][starVLA.training.train_starvla][INFO] - Step 3919, Loss: {'train/loss': 0.4278135299682617, 'train/grad_norm_pre_clip': 10.512561798095703, 'timing/data': 0.0013658693060278893, 'timing/model': 1.7376408213749528, 'learning_rate/qwen_vl_interface': 4.45571613092318e-06, 'learning_rate/action_model': 4.4557161309231794e-05, 'epoch': 1.51}) +[2026-06-18 01:25:53,788][starVLA.training.train_starvla][INFO] - Step 3920, Loss: {'train/loss': 0.4686257839202881, 'train/grad_norm_pre_clip': 14.186762809753418, 'timing/data': 0.0012763291597366333, 'timing/model': 1.1864743623882532, 'learning_rate/qwen_vl_interface': 4.453583973403746e-06, 'learning_rate/action_model': 4.453583973403746e-05, 'epoch': 1.51}) +[2026-06-18 01:26:03,583][starVLA.training.train_starvla][INFO] - Step 3921, Loss: {'train/loss': 0.6141325831413269, 'train/grad_norm_pre_clip': 14.44070053100586, 'timing/data': 0.001072118990123272, 'timing/model': 4.530726693570614, 'learning_rate/qwen_vl_interface': 4.451451980982209e-06, 'learning_rate/action_model': 4.4514519809822086e-05, 'epoch': 1.51}) +[2026-06-18 01:26:11,810][starVLA.training.train_starvla][INFO] - Step 3922, Loss: {'train/loss': 0.44947248697280884, 'train/grad_norm_pre_clip': 11.850318908691406, 'timing/data': 0.0012674424797296524, 'timing/model': 1.446023584343493, 'learning_rate/qwen_vl_interface': 4.4493201541005385e-06, 'learning_rate/action_model': 4.4493201541005384e-05, 'epoch': 1.51}) +[2026-06-18 01:26:27,417][starVLA.training.train_starvla][INFO] - Step 3923, Loss: {'train/loss': 0.6013160347938538, 'train/grad_norm_pre_clip': 14.537227630615234, 'timing/data': 0.000918346457183361, 'timing/model': 4.83868336956948, 'learning_rate/qwen_vl_interface': 4.447188493200661e-06, 'learning_rate/action_model': 4.447188493200661e-05, 'epoch': 1.51}) +[2026-06-18 01:26:35,589][starVLA.training.train_starvla][INFO] - Step 3924, Loss: {'train/loss': 0.23194603621959686, 'train/grad_norm_pre_clip': 7.6255621910095215, 'timing/data': 0.0014612115919589996, 'timing/model': 1.2743303468450904, 'learning_rate/qwen_vl_interface': 4.445056998724472e-06, 'learning_rate/action_model': 4.4450569987244716e-05, 'epoch': 1.51}) +[2026-06-18 01:26:45,508][starVLA.training.train_starvla][INFO] - Step 3925, Loss: {'train/loss': 0.328302800655365, 'train/grad_norm_pre_clip': 11.550869941711426, 'timing/data': 0.0012540677562355995, 'timing/model': 4.584798187017441, 'learning_rate/qwen_vl_interface': 4.4429256711138335e-06, 'learning_rate/action_model': 4.442925671113833e-05, 'epoch': 1.51}) +[2026-06-18 01:26:56,051][starVLA.training.train_starvla][INFO] - Step 3926, Loss: {'train/loss': 0.42786064743995667, 'train/grad_norm_pre_clip': 13.719931602478027, 'timing/data': 0.0016195308417081833, 'timing/model': 4.50361714232713, 'learning_rate/qwen_vl_interface': 4.440794510810572e-06, 'learning_rate/action_model': 4.440794510810572e-05, 'epoch': 1.51}) +[2026-06-18 01:27:04,239][starVLA.training.train_starvla][INFO] - Step 3927, Loss: {'train/loss': 0.5962916016578674, 'train/grad_norm_pre_clip': 11.638874053955078, 'timing/data': 0.0010712482035160065, 'timing/model': 1.626248779706657, 'learning_rate/qwen_vl_interface': 4.438663518256481e-06, 'learning_rate/action_model': 4.438663518256481e-05, 'epoch': 1.51}) +[2026-06-18 01:27:10,773][starVLA.training.train_starvla][INFO] - Step 3928, Loss: {'train/loss': 0.40192151069641113, 'train/grad_norm_pre_clip': 11.652006149291992, 'timing/data': 0.0013133324682712555, 'timing/model': 1.5989154586568475, 'learning_rate/qwen_vl_interface': 4.436532693893311e-06, 'learning_rate/action_model': 4.436532693893311e-05, 'epoch': 1.51}) +[2026-06-18 01:27:15,628][starVLA.training.train_starvla][INFO] - Step 3929, Loss: {'train/loss': 0.4292047619819641, 'train/grad_norm_pre_clip': 12.917076110839844, 'timing/data': 0.0015932563692331314, 'timing/model': 1.1894648019224405, 'learning_rate/qwen_vl_interface': 4.434402038162793e-06, 'learning_rate/action_model': 4.434402038162793e-05, 'epoch': 1.51}) +[2026-06-18 01:27:21,314][starVLA.training.train_starvla][INFO] - Step 3930, Loss: {'train/loss': 0.4981779456138611, 'train/grad_norm_pre_clip': 11.33830451965332, 'timing/data': 0.001538669690489769, 'timing/model': 1.3087176913395524, 'learning_rate/qwen_vl_interface': 4.432271551506609e-06, 'learning_rate/action_model': 4.4322715515066094e-05, 'epoch': 1.51}) +[2026-06-18 01:27:26,286][starVLA.training.train_starvla][INFO] - Step 3931, Loss: {'train/loss': 0.6685067415237427, 'train/grad_norm_pre_clip': 13.825148582458496, 'timing/data': 0.0014915112406015396, 'timing/model': 1.2276768097653985, 'learning_rate/qwen_vl_interface': 4.4301412343664125e-06, 'learning_rate/action_model': 4.4301412343664125e-05, 'epoch': 1.51}) +[2026-06-18 01:27:31,200][starVLA.training.train_starvla][INFO] - Step 3932, Loss: {'train/loss': 0.1436193734407425, 'train/grad_norm_pre_clip': 7.386646270751953, 'timing/data': 0.0032474612817168236, 'timing/model': 1.1549036670476198, 'learning_rate/qwen_vl_interface': 4.428011087183819e-06, 'learning_rate/action_model': 4.428011087183819e-05, 'epoch': 1.51}) +[2026-06-18 01:27:36,334][starVLA.training.train_starvla][INFO] - Step 3933, Loss: {'train/loss': 0.43823686242103577, 'train/grad_norm_pre_clip': 23.01123809814453, 'timing/data': 0.0015734555199742317, 'timing/model': 1.2593196853995323, 'learning_rate/qwen_vl_interface': 4.4258811104004125e-06, 'learning_rate/action_model': 4.425881110400413e-05, 'epoch': 1.51}) +[2026-06-18 01:27:41,215][starVLA.training.train_starvla][INFO] - Step 3934, Loss: {'train/loss': 0.22940725088119507, 'train/grad_norm_pre_clip': 7.362023830413818, 'timing/data': 0.0012997137382626534, 'timing/model': 1.0897041857242584, 'learning_rate/qwen_vl_interface': 4.4237513044577394e-06, 'learning_rate/action_model': 4.4237513044577386e-05, 'epoch': 1.51}) +[2026-06-18 01:27:49,735][starVLA.training.train_starvla][INFO] - Step 3935, Loss: {'train/loss': 0.4079994559288025, 'train/grad_norm_pre_clip': 16.85028839111328, 'timing/data': 0.0009374553337693214, 'timing/model': 4.805906631983817, 'learning_rate/qwen_vl_interface': 4.421621669797309e-06, 'learning_rate/action_model': 4.421621669797308e-05, 'epoch': 1.51}) +[2026-06-18 01:27:54,440][starVLA.training.train_starvla][INFO] - Step 3936, Loss: {'train/loss': 0.33290261030197144, 'train/grad_norm_pre_clip': 11.828120231628418, 'timing/data': 0.0013210969045758247, 'timing/model': 1.1221692031249404, 'learning_rate/qwen_vl_interface': 4.419492206860599e-06, 'learning_rate/action_model': 4.419492206860598e-05, 'epoch': 1.51}) +[2026-06-18 01:27:59,221][starVLA.training.train_starvla][INFO] - Step 3937, Loss: {'train/loss': 0.3489241898059845, 'train/grad_norm_pre_clip': 11.267098426818848, 'timing/data': 0.001042993739247322, 'timing/model': 1.4902022927999496, 'learning_rate/qwen_vl_interface': 4.417362916089048e-06, 'learning_rate/action_model': 4.417362916089047e-05, 'epoch': 1.51}) +[2026-06-18 01:28:03,725][starVLA.training.train_starvla][INFO] - Step 3938, Loss: {'train/loss': 0.3520483672618866, 'train/grad_norm_pre_clip': 12.580437660217285, 'timing/data': 0.002423258498311043, 'timing/model': 1.1390580190345645, 'learning_rate/qwen_vl_interface': 4.415233797924059e-06, 'learning_rate/action_model': 4.415233797924059e-05, 'epoch': 1.51}) +[2026-06-18 01:28:09,387][starVLA.training.train_starvla][INFO] - Step 3939, Loss: {'train/loss': 0.37544333934783936, 'train/grad_norm_pre_clip': 13.888954162597656, 'timing/data': 0.001294250600039959, 'timing/model': 1.507153069600463, 'learning_rate/qwen_vl_interface': 4.413104852807002e-06, 'learning_rate/action_model': 4.413104852807001e-05, 'epoch': 1.52}) +[2026-06-18 01:28:14,650][starVLA.training.train_starvla][INFO] - Step 3940, Loss: {'train/loss': 0.29982274770736694, 'train/grad_norm_pre_clip': 11.109631538391113, 'timing/data': 0.0012826640158891678, 'timing/model': 1.3385982122272253, 'learning_rate/qwen_vl_interface': 4.41097608117921e-06, 'learning_rate/action_model': 4.41097608117921e-05, 'epoch': 1.52}) +[2026-06-18 01:28:20,034][starVLA.training.train_starvla][INFO] - Step 3941, Loss: {'train/loss': 0.20741356909275055, 'train/grad_norm_pre_clip': 8.673873901367188, 'timing/data': 0.001184551976621151, 'timing/model': 1.2568734986707568, 'learning_rate/qwen_vl_interface': 4.40884748348198e-06, 'learning_rate/action_model': 4.40884748348198e-05, 'epoch': 1.52}) +[2026-06-18 01:28:24,789][starVLA.training.train_starvla][INFO] - Step 3942, Loss: {'train/loss': 0.4401511251926422, 'train/grad_norm_pre_clip': 11.333588600158691, 'timing/data': 0.001156090758740902, 'timing/model': 1.3230183189734817, 'learning_rate/qwen_vl_interface': 4.406719060156572e-06, 'learning_rate/action_model': 4.4067190601565716e-05, 'epoch': 1.52}) +[2026-06-18 01:28:29,856][starVLA.training.train_starvla][INFO] - Step 3943, Loss: {'train/loss': 0.18169750273227692, 'train/grad_norm_pre_clip': 7.425521373748779, 'timing/data': 0.0011838451027870178, 'timing/model': 1.2059654965996742, 'learning_rate/qwen_vl_interface': 4.404590811644209e-06, 'learning_rate/action_model': 4.4045908116442087e-05, 'epoch': 1.52}) +[2026-06-18 01:28:34,954][starVLA.training.train_starvla][INFO] - Step 3944, Loss: {'train/loss': 0.7338652610778809, 'train/grad_norm_pre_clip': 13.121965408325195, 'timing/data': 0.0010254886001348495, 'timing/model': 1.118088899180293, 'learning_rate/qwen_vl_interface': 4.40246273838608e-06, 'learning_rate/action_model': 4.40246273838608e-05, 'epoch': 1.52}) +[2026-06-18 01:28:44,125][starVLA.training.train_starvla][INFO] - Step 3945, Loss: {'train/loss': 0.5402069091796875, 'train/grad_norm_pre_clip': 14.05628490447998, 'timing/data': 0.0010348744690418243, 'timing/model': 2.3648913986980915, 'learning_rate/qwen_vl_interface': 4.400334840823339e-06, 'learning_rate/action_model': 4.400334840823339e-05, 'epoch': 1.52}) +[2026-06-18 01:28:49,758][starVLA.training.train_starvla][INFO] - Step 3946, Loss: {'train/loss': 0.5387285351753235, 'train/grad_norm_pre_clip': 15.551977157592773, 'timing/data': 0.0014018099755048752, 'timing/model': 1.2346942070871592, 'learning_rate/qwen_vl_interface': 4.398207119397098e-06, 'learning_rate/action_model': 4.398207119397098e-05, 'epoch': 1.52}) +[2026-06-18 01:28:54,980][starVLA.training.train_starvla][INFO] - Step 3947, Loss: {'train/loss': 0.24908383190631866, 'train/grad_norm_pre_clip': 9.075606346130371, 'timing/data': 0.0011745132505893707, 'timing/model': 1.455776683986187, 'learning_rate/qwen_vl_interface': 4.396079574548439e-06, 'learning_rate/action_model': 4.396079574548439e-05, 'epoch': 1.52}) +[2026-06-18 01:28:59,704][starVLA.training.train_starvla][INFO] - Step 3948, Loss: {'train/loss': 0.7434790730476379, 'train/grad_norm_pre_clip': 19.433940887451172, 'timing/data': 0.0009805932641029358, 'timing/model': 1.3036254681646824, 'learning_rate/qwen_vl_interface': 4.393952206718403e-06, 'learning_rate/action_model': 4.3939522067184026e-05, 'epoch': 1.52}) +[2026-06-18 01:29:04,402][starVLA.training.train_starvla][INFO] - Step 3949, Loss: {'train/loss': 0.4536934196949005, 'train/grad_norm_pre_clip': 14.50020980834961, 'timing/data': 0.0013545025140047073, 'timing/model': 1.1228001909330487, 'learning_rate/qwen_vl_interface': 4.391825016347996e-06, 'learning_rate/action_model': 4.391825016347996e-05, 'epoch': 1.52}) +[2026-06-18 01:29:09,480][starVLA.training.train_starvla][INFO] - Step 3950, Loss: {'train/loss': 0.4122788608074188, 'train/grad_norm_pre_clip': 17.234766006469727, 'timing/data': 0.001174665056169033, 'timing/model': 1.2753537828102708, 'learning_rate/qwen_vl_interface': 4.3896980038781844e-06, 'learning_rate/action_model': 4.3896980038781846e-05, 'epoch': 1.52}) +[2026-06-18 01:29:14,156][starVLA.training.train_starvla][INFO] - Step 3951, Loss: {'train/loss': 0.28083017468452454, 'train/grad_norm_pre_clip': 14.368317604064941, 'timing/data': 0.0011394862085580826, 'timing/model': 1.1016506301239133, 'learning_rate/qwen_vl_interface': 4.3875711697499054e-06, 'learning_rate/action_model': 4.387571169749905e-05, 'epoch': 1.52}) +[2026-06-18 01:29:18,847][starVLA.training.train_starvla][INFO] - Step 3952, Loss: {'train/loss': 0.4583781361579895, 'train/grad_norm_pre_clip': 14.287004470825195, 'timing/data': 0.0012842454016208649, 'timing/model': 1.3412246564403176, 'learning_rate/qwen_vl_interface': 4.38544451440405e-06, 'learning_rate/action_model': 4.3854445144040495e-05, 'epoch': 1.52}) +[2026-06-18 01:29:24,434][starVLA.training.train_starvla][INFO] - Step 3953, Loss: {'train/loss': 0.362621545791626, 'train/grad_norm_pre_clip': 10.266468048095703, 'timing/data': 0.0011411402374505997, 'timing/model': 1.5398083860054612, 'learning_rate/qwen_vl_interface': 4.383318038281475e-06, 'learning_rate/action_model': 4.383318038281475e-05, 'epoch': 1.52}) +[2026-06-18 01:29:31,235][starVLA.training.train_starvla][INFO] - Step 3954, Loss: {'train/loss': 0.48769301176071167, 'train/grad_norm_pre_clip': 18.225128173828125, 'timing/data': 0.0013960106298327446, 'timing/model': 1.1172722252085805, 'learning_rate/qwen_vl_interface': 4.381191741823005e-06, 'learning_rate/action_model': 4.3811917418230046e-05, 'epoch': 1.52}) +[2026-06-18 01:29:39,600][starVLA.training.train_starvla][INFO] - Step 3955, Loss: {'train/loss': 0.3696691393852234, 'train/grad_norm_pre_clip': 11.464303016662598, 'timing/data': 0.0011980105191469193, 'timing/model': 3.9237459180876613, 'learning_rate/qwen_vl_interface': 4.379065625469422e-06, 'learning_rate/action_model': 4.3790656254694215e-05, 'epoch': 1.52}) +[2026-06-18 01:29:53,060][starVLA.training.train_starvla][INFO] - Step 3956, Loss: {'train/loss': 0.5544100403785706, 'train/grad_norm_pre_clip': 13.889519691467285, 'timing/data': 0.001271476037800312, 'timing/model': 1.493165479041636, 'learning_rate/qwen_vl_interface': 4.3769396896614724e-06, 'learning_rate/action_model': 4.376939689661472e-05, 'epoch': 1.52}) +[2026-06-18 01:29:59,229][starVLA.training.train_starvla][INFO] - Step 3957, Loss: {'train/loss': 0.6335456371307373, 'train/grad_norm_pre_clip': 15.40396499633789, 'timing/data': 0.0014265431091189384, 'timing/model': 1.970342818647623, 'learning_rate/qwen_vl_interface': 4.374813934839862e-06, 'learning_rate/action_model': 4.374813934839862e-05, 'epoch': 1.52}) +[2026-06-18 01:30:03,865][starVLA.training.train_starvla][INFO] - Step 3958, Loss: {'train/loss': 0.5513251423835754, 'train/grad_norm_pre_clip': 28.4033260345459, 'timing/data': 0.0014959666877985, 'timing/model': 1.1433542054146528, 'learning_rate/qwen_vl_interface': 4.372688361445267e-06, 'learning_rate/action_model': 4.3726883614452665e-05, 'epoch': 1.52}) +[2026-06-18 01:30:09,280][starVLA.training.train_starvla][INFO] - Step 3959, Loss: {'train/loss': 0.3940197229385376, 'train/grad_norm_pre_clip': 11.911802291870117, 'timing/data': 0.0009775087237358093, 'timing/model': 1.148114925250411, 'learning_rate/qwen_vl_interface': 4.370562969918319e-06, 'learning_rate/action_model': 4.370562969918319e-05, 'epoch': 1.52}) +[2026-06-18 01:30:14,536][starVLA.training.train_starvla][INFO] - Step 3960, Loss: {'train/loss': 0.4678868353366852, 'train/grad_norm_pre_clip': 14.035788536071777, 'timing/data': 0.0011291569098830223, 'timing/model': 1.1122972667217255, 'learning_rate/qwen_vl_interface': 4.368437760699613e-06, 'learning_rate/action_model': 4.3684377606996124e-05, 'epoch': 1.52}) +[2026-06-18 01:30:19,857][starVLA.training.train_starvla][INFO] - Step 3961, Loss: {'train/loss': 0.4936925172805786, 'train/grad_norm_pre_clip': 17.357620239257812, 'timing/data': 0.0016798358410596848, 'timing/model': 1.3537904294207692, 'learning_rate/qwen_vl_interface': 4.366312734229706e-06, 'learning_rate/action_model': 4.366312734229706e-05, 'epoch': 1.52}) +[2026-06-18 01:30:24,912][starVLA.training.train_starvla][INFO] - Step 3962, Loss: {'train/loss': 0.5881948471069336, 'train/grad_norm_pre_clip': 10.694613456726074, 'timing/data': 0.0014713145792484283, 'timing/model': 1.309517651796341, 'learning_rate/qwen_vl_interface': 4.364187890949122e-06, 'learning_rate/action_model': 4.364187890949122e-05, 'epoch': 1.52}) +[2026-06-18 01:30:32,053][starVLA.training.train_starvla][INFO] - Step 3963, Loss: {'train/loss': 0.5418506264686584, 'train/grad_norm_pre_clip': 11.982039451599121, 'timing/data': 0.004147416912019253, 'timing/model': 2.832552340812981, 'learning_rate/qwen_vl_interface': 4.362063231298342e-06, 'learning_rate/action_model': 4.3620632312983425e-05, 'epoch': 1.52}) +[2026-06-18 01:30:38,878][starVLA.training.train_starvla][INFO] - Step 3964, Loss: {'train/loss': 0.31288743019104004, 'train/grad_norm_pre_clip': 10.081408500671387, 'timing/data': 0.0013666842132806778, 'timing/model': 1.514461835846305, 'learning_rate/qwen_vl_interface': 4.359938755717807e-06, 'learning_rate/action_model': 4.359938755717807e-05, 'epoch': 1.52}) +[2026-06-18 01:30:44,198][starVLA.training.train_starvla][INFO] - Step 3965, Loss: {'train/loss': 0.4260387718677521, 'train/grad_norm_pre_clip': 10.350584983825684, 'timing/data': 0.0013221856206655502, 'timing/model': 1.371828654780984, 'learning_rate/qwen_vl_interface': 4.357814464647929e-06, 'learning_rate/action_model': 4.357814464647929e-05, 'epoch': 1.53}) +[2026-06-18 01:30:49,510][starVLA.training.train_starvla][INFO] - Step 3966, Loss: {'train/loss': 0.12838754057884216, 'train/grad_norm_pre_clip': 6.149980545043945, 'timing/data': 0.0013887370005249977, 'timing/model': 1.5093824034556746, 'learning_rate/qwen_vl_interface': 4.355690358529071e-06, 'learning_rate/action_model': 4.3556903585290705e-05, 'epoch': 1.53}) +[2026-06-18 01:30:57,592][starVLA.training.train_starvla][INFO] - Step 3967, Loss: {'train/loss': 0.657481849193573, 'train/grad_norm_pre_clip': 18.012731552124023, 'timing/data': 0.001072426326572895, 'timing/model': 4.508629192598164, 'learning_rate/qwen_vl_interface': 4.353566437801564e-06, 'learning_rate/action_model': 4.353566437801563e-05, 'epoch': 1.53}) +[2026-06-18 01:31:02,256][starVLA.training.train_starvla][INFO] - Step 3968, Loss: {'train/loss': 0.4738399088382721, 'train/grad_norm_pre_clip': 13.322460174560547, 'timing/data': 0.0009644245728850365, 'timing/model': 1.2464007949456573, 'learning_rate/qwen_vl_interface': 4.351442702905698e-06, 'learning_rate/action_model': 4.3514427029056976e-05, 'epoch': 1.53}) +[2026-06-18 01:31:10,928][starVLA.training.train_starvla][INFO] - Step 3969, Loss: {'train/loss': 0.34205543994903564, 'train/grad_norm_pre_clip': 11.188922882080078, 'timing/data': 0.0013972790911793709, 'timing/model': 1.476132427342236, 'learning_rate/qwen_vl_interface': 4.349319154281727e-06, 'learning_rate/action_model': 4.349319154281728e-05, 'epoch': 1.53}) +[2026-06-18 01:31:15,649][starVLA.training.train_starvla][INFO] - Step 3970, Loss: {'train/loss': 0.5213853120803833, 'train/grad_norm_pre_clip': 12.716716766357422, 'timing/data': 0.0009803464636206627, 'timing/model': 1.0659507093951106, 'learning_rate/qwen_vl_interface': 4.347195792369867e-06, 'learning_rate/action_model': 4.347195792369867e-05, 'epoch': 1.53}) +[2026-06-18 01:31:20,606][starVLA.training.train_starvla][INFO] - Step 3971, Loss: {'train/loss': 1.0148197412490845, 'train/grad_norm_pre_clip': 16.529218673706055, 'timing/data': 0.0014303391799330711, 'timing/model': 1.1108247255906463, 'learning_rate/qwen_vl_interface': 4.345072617610288e-06, 'learning_rate/action_model': 4.345072617610288e-05, 'epoch': 1.53}) +[2026-06-18 01:31:25,099][starVLA.training.train_starvla][INFO] - Step 3972, Loss: {'train/loss': 0.5910882353782654, 'train/grad_norm_pre_clip': 13.872478485107422, 'timing/data': 0.0014465916901826859, 'timing/model': 1.1164586264640093, 'learning_rate/qwen_vl_interface': 4.342949630443129e-06, 'learning_rate/action_model': 4.342949630443129e-05, 'epoch': 1.53}) +[2026-06-18 01:31:29,969][starVLA.training.train_starvla][INFO] - Step 3973, Loss: {'train/loss': 0.42878806591033936, 'train/grad_norm_pre_clip': 10.66427230834961, 'timing/data': 0.0013387203216552734, 'timing/model': 1.1874657208099961, 'learning_rate/qwen_vl_interface': 4.340826831308488e-06, 'learning_rate/action_model': 4.340826831308488e-05, 'epoch': 1.53}) +[2026-06-18 01:31:34,729][starVLA.training.train_starvla][INFO] - Step 3974, Loss: {'train/loss': 0.5562777519226074, 'train/grad_norm_pre_clip': 14.725547790527344, 'timing/data': 0.0014495877549052238, 'timing/model': 1.082731100730598, 'learning_rate/qwen_vl_interface': 4.338704220646424e-06, 'learning_rate/action_model': 4.338704220646424e-05, 'epoch': 1.53}) +[2026-06-18 01:31:42,233][starVLA.training.train_starvla][INFO] - Step 3975, Loss: {'train/loss': 0.4484513998031616, 'train/grad_norm_pre_clip': 9.948471069335938, 'timing/data': 0.0021586325019598007, 'timing/model': 1.3609951566904783, 'learning_rate/qwen_vl_interface': 4.336581798896955e-06, 'learning_rate/action_model': 4.336581798896955e-05, 'epoch': 1.53}) +[2026-06-18 01:31:47,280][starVLA.training.train_starvla][INFO] - Step 3976, Loss: {'train/loss': 0.38193508982658386, 'train/grad_norm_pre_clip': 12.961624145507812, 'timing/data': 0.0008662566542625427, 'timing/model': 1.1183460596948862, 'learning_rate/qwen_vl_interface': 4.334459566500062e-06, 'learning_rate/action_model': 4.334459566500062e-05, 'epoch': 1.53}) +[2026-06-18 01:32:00,941][starVLA.training.train_starvla][INFO] - Step 3977, Loss: {'train/loss': 0.2896384298801422, 'train/grad_norm_pre_clip': 9.257765769958496, 'timing/data': 0.0016891416162252426, 'timing/model': 5.862689628265798, 'learning_rate/qwen_vl_interface': 4.332337523895688e-06, 'learning_rate/action_model': 4.3323375238956876e-05, 'epoch': 1.53}) +[2026-06-18 01:32:05,949][starVLA.training.train_starvla][INFO] - Step 3978, Loss: {'train/loss': 0.25041013956069946, 'train/grad_norm_pre_clip': 8.185826301574707, 'timing/data': 0.0012393714860081673, 'timing/model': 1.2485692268237472, 'learning_rate/qwen_vl_interface': 4.330215671523731e-06, 'learning_rate/action_model': 4.330215671523731e-05, 'epoch': 1.53}) +[2026-06-18 01:32:11,565][starVLA.training.train_starvla][INFO] - Step 3979, Loss: {'train/loss': 0.4545343816280365, 'train/grad_norm_pre_clip': 19.575668334960938, 'timing/data': 0.0010791625827550888, 'timing/model': 1.1632340177893639, 'learning_rate/qwen_vl_interface': 4.328094009824054e-06, 'learning_rate/action_model': 4.328094009824054e-05, 'epoch': 1.53}) +[2026-06-18 01:32:16,504][starVLA.training.train_starvla][INFO] - Step 3980, Loss: {'train/loss': 0.32897254824638367, 'train/grad_norm_pre_clip': 10.410505294799805, 'timing/data': 0.0010118000209331512, 'timing/model': 1.1717356480658054, 'learning_rate/qwen_vl_interface': 4.3259725392364825e-06, 'learning_rate/action_model': 4.325972539236482e-05, 'epoch': 1.53}) +[2026-06-18 01:32:23,490][starVLA.training.train_starvla][INFO] - Step 3981, Loss: {'train/loss': 0.2991068363189697, 'train/grad_norm_pre_clip': 10.16893482208252, 'timing/data': 0.0011311592534184456, 'timing/model': 3.0166519712656736, 'learning_rate/qwen_vl_interface': 4.323851260200797e-06, 'learning_rate/action_model': 4.323851260200796e-05, 'epoch': 1.53}) +[2026-06-18 01:32:34,327][starVLA.training.train_starvla][INFO] - Step 3982, Loss: {'train/loss': 0.30062881112098694, 'train/grad_norm_pre_clip': 11.09257698059082, 'timing/data': 0.0011271443217992783, 'timing/model': 1.1780013712123036, 'learning_rate/qwen_vl_interface': 4.321730173156743e-06, 'learning_rate/action_model': 4.321730173156743e-05, 'epoch': 1.53}) +[2026-06-18 01:32:39,484][starVLA.training.train_starvla][INFO] - Step 3983, Loss: {'train/loss': 0.2666269838809967, 'train/grad_norm_pre_clip': 8.669775009155273, 'timing/data': 0.0011702636256814003, 'timing/model': 1.3790400326251984, 'learning_rate/qwen_vl_interface': 4.31960927854402e-06, 'learning_rate/action_model': 4.31960927854402e-05, 'epoch': 1.53}) +[2026-06-18 01:32:47,561][starVLA.training.train_starvla][INFO] - Step 3984, Loss: {'train/loss': 0.32123079895973206, 'train/grad_norm_pre_clip': 8.286567687988281, 'timing/data': 0.0011532418429851532, 'timing/model': 4.072493481449783, 'learning_rate/qwen_vl_interface': 4.317488576802298e-06, 'learning_rate/action_model': 4.3174885768022986e-05, 'epoch': 1.53}) +[2026-06-18 01:32:52,717][starVLA.training.train_starvla][INFO] - Step 3985, Loss: {'train/loss': 0.1783360242843628, 'train/grad_norm_pre_clip': 7.475112438201904, 'timing/data': 0.0009874021634459496, 'timing/model': 1.392784626223147, 'learning_rate/qwen_vl_interface': 4.315368068371197e-06, 'learning_rate/action_model': 4.315368068371197e-05, 'epoch': 1.53}) +[2026-06-18 01:33:03,389][starVLA.training.train_starvla][INFO] - Step 3986, Loss: {'train/loss': 0.516092836856842, 'train/grad_norm_pre_clip': 16.04631996154785, 'timing/data': 0.0015583150088787079, 'timing/model': 1.1756316488608718, 'learning_rate/qwen_vl_interface': 4.313247753690298e-06, 'learning_rate/action_model': 4.3132477536902985e-05, 'epoch': 1.53}) +[2026-06-18 01:33:08,476][starVLA.training.train_starvla][INFO] - Step 3987, Loss: {'train/loss': 0.23604966700077057, 'train/grad_norm_pre_clip': 7.363866329193115, 'timing/data': 0.0010274266824126244, 'timing/model': 1.7833850411698222, 'learning_rate/qwen_vl_interface': 4.311127633199151e-06, 'learning_rate/action_model': 4.3111276331991504e-05, 'epoch': 1.53}) +[2026-06-18 01:33:13,707][starVLA.training.train_starvla][INFO] - Step 3988, Loss: {'train/loss': 0.40326350927352905, 'train/grad_norm_pre_clip': 11.422137260437012, 'timing/data': 0.0014789989218115807, 'timing/model': 1.505162263289094, 'learning_rate/qwen_vl_interface': 4.309007707337255e-06, 'learning_rate/action_model': 4.309007707337255e-05, 'epoch': 1.53}) +[2026-06-18 01:33:23,828][starVLA.training.train_starvla][INFO] - Step 3989, Loss: {'train/loss': 0.3885257840156555, 'train/grad_norm_pre_clip': 12.305834770202637, 'timing/data': 0.0012854617089033127, 'timing/model': 1.0946771912276745, 'learning_rate/qwen_vl_interface': 4.3068879765440745e-06, 'learning_rate/action_model': 4.306887976544074e-05, 'epoch': 1.53}) +[2026-06-18 01:33:31,627][starVLA.training.train_starvla][INFO] - Step 3990, Loss: {'train/loss': 0.5479641556739807, 'train/grad_norm_pre_clip': 10.738490104675293, 'timing/data': 0.0028028078377246857, 'timing/model': 1.1018116660416126, 'learning_rate/qwen_vl_interface': 4.304768441259029e-06, 'learning_rate/action_model': 4.3047684412590285e-05, 'epoch': 1.53}) +[2026-06-18 01:33:41,377][starVLA.training.train_starvla][INFO] - Step 3991, Loss: {'train/loss': 0.4265081584453583, 'train/grad_norm_pre_clip': 12.932541847229004, 'timing/data': 0.0012319674715399742, 'timing/model': 1.1386333415284753, 'learning_rate/qwen_vl_interface': 4.302649101921505e-06, 'learning_rate/action_model': 4.302649101921505e-05, 'epoch': 1.54}) +[2026-06-18 01:33:47,055][starVLA.training.train_starvla][INFO] - Step 3992, Loss: {'train/loss': 0.21427369117736816, 'train/grad_norm_pre_clip': 8.648447036743164, 'timing/data': 0.0032618846744298935, 'timing/model': 1.9361376874148846, 'learning_rate/qwen_vl_interface': 4.300529958970841e-06, 'learning_rate/action_model': 4.300529958970841e-05, 'epoch': 1.54}) +[2026-06-18 01:33:52,636][starVLA.training.train_starvla][INFO] - Step 3993, Loss: {'train/loss': 0.38305187225341797, 'train/grad_norm_pre_clip': 11.75080394744873, 'timing/data': 0.0012576356530189514, 'timing/model': 1.1204473841935396, 'learning_rate/qwen_vl_interface': 4.298411012846338e-06, 'learning_rate/action_model': 4.2984110128463376e-05, 'epoch': 1.54}) +[2026-06-18 01:33:57,542][starVLA.training.train_starvla][INFO] - Step 3994, Loss: {'train/loss': 0.2546199560165405, 'train/grad_norm_pre_clip': 9.905611991882324, 'timing/data': 0.0013507707044482231, 'timing/model': 1.1029984168708324, 'learning_rate/qwen_vl_interface': 4.296292263987255e-06, 'learning_rate/action_model': 4.2962922639872556e-05, 'epoch': 1.54}) +[2026-06-18 01:34:07,179][starVLA.training.train_starvla][INFO] - Step 3995, Loss: {'train/loss': 0.23077669739723206, 'train/grad_norm_pre_clip': 9.550285339355469, 'timing/data': 0.0009842319414019585, 'timing/model': 5.58303039893508, 'learning_rate/qwen_vl_interface': 4.294173712832813e-06, 'learning_rate/action_model': 4.294173712832813e-05, 'epoch': 1.54}) +[2026-06-18 01:34:13,858][starVLA.training.train_starvla][INFO] - Step 3996, Loss: {'train/loss': 0.20642384886741638, 'train/grad_norm_pre_clip': 7.302937984466553, 'timing/data': 0.001210436224937439, 'timing/model': 1.115209591574967, 'learning_rate/qwen_vl_interface': 4.292055359822191e-06, 'learning_rate/action_model': 4.29205535982219e-05, 'epoch': 1.54}) +[2026-06-18 01:34:19,506][starVLA.training.train_starvla][INFO] - Step 3997, Loss: {'train/loss': 0.4183274805545807, 'train/grad_norm_pre_clip': 11.501676559448242, 'timing/data': 0.0011569801717996597, 'timing/model': 1.8334152223542333, 'learning_rate/qwen_vl_interface': 4.289937205394519e-06, 'learning_rate/action_model': 4.2899372053945185e-05, 'epoch': 1.54}) +[2026-06-18 01:34:24,354][starVLA.training.train_starvla][INFO] - Step 3998, Loss: {'train/loss': 0.3155488073825836, 'train/grad_norm_pre_clip': 11.504864692687988, 'timing/data': 0.0018325727432966232, 'timing/model': 1.1080941651016474, 'learning_rate/qwen_vl_interface': 4.287819249988901e-06, 'learning_rate/action_model': 4.2878192499889007e-05, 'epoch': 1.54}) +[2026-06-18 01:34:30,353][starVLA.training.train_starvla][INFO] - Step 3999, Loss: {'train/loss': 0.607048749923706, 'train/grad_norm_pre_clip': 16.83727264404297, 'timing/data': 0.00127364881336689, 'timing/model': 1.1195073137059808, 'learning_rate/qwen_vl_interface': 4.285701494044387e-06, 'learning_rate/action_model': 4.285701494044387e-05, 'epoch': 1.54}) +[2026-06-18 01:41:03,934][starVLA.training.train_starvla][INFO] - Step 4000, Loss: {'train/loss': 0.5349904894828796, 'train/grad_norm_pre_clip': 20.224477767944336, 'eval/loss': 0.4686513841152191, 'eval/latency_4/loss': 0.4686513841152191, 'eval/demon_fire_precision': 0.9538365564566438, 'eval/demon_fire_recall': 0.993179603767457, 'eval/demon_fire_f1': 0.9731105807478122, 'eval/demon_fire_rate_model': 0.9415565345080763, 'eval/demon_fire_rate_teacher': 0.9042584434654919, 'eval/demon_left_precision': 0.909952606635071, 'eval/demon_left_recall': 0.4383561643835616, 'eval/demon_left_f1': 0.5916795069337442, 'eval/demon_left_rate_model': 0.06196769456681351, 'eval/demon_left_rate_teacher': 0.12863436123348018, 'eval/demon_right_precision': 0.804642166344294, 'eval/demon_right_recall': 0.7509025270758123, 'eval/demon_right_f1': 0.7768440709617179, 'eval/demon_right_rate_model': 0.15183553597650515, 'eval/demon_right_rate_teacher': 0.16270190895741557, 'eval/demon_move_f1': 0.684261788947731, 'eval/demon_cc_f1': 0.863728367885594, 'eval/latency_4/demon_fire_precision': 0.9538365564566438, 'eval/latency_4/demon_fire_recall': 0.993179603767457, 'eval/latency_4/demon_fire_f1': 0.9731105807478122, 'eval/latency_4/demon_fire_rate_model': 0.9415565345080763, 'eval/latency_4/demon_fire_rate_teacher': 0.9042584434654919, 'eval/latency_4/demon_left_precision': 0.909952606635071, 'eval/latency_4/demon_left_recall': 0.4383561643835616, 'eval/latency_4/demon_left_f1': 0.5916795069337442, 'eval/latency_4/demon_left_rate_model': 0.06196769456681351, 'eval/latency_4/demon_left_rate_teacher': 0.12863436123348018, 'eval/latency_4/demon_right_precision': 0.804642166344294, 'eval/latency_4/demon_right_recall': 0.7509025270758123, 'eval/latency_4/demon_right_f1': 0.7768440709617179, 'eval/latency_4/demon_right_rate_model': 0.15183553597650515, 'eval/latency_4/demon_right_rate_teacher': 0.16270190895741557, 'eval/latency_4/demon_move_f1': 0.684261788947731, 'eval/latency_4/demon_cc_f1': 0.863728367885594, 'timing/data': 0.00352554302662611, 'timing/model': 1.407225700095296, 'learning_rate/qwen_vl_interface': 4.283583937999991e-06, 'learning_rate/action_model': 4.2835839379999906e-05, 'epoch': 1.54}) +[2026-06-18 01:41:03,941][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_4000_state +[2026-06-18 01:41:03,951][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 01:41:13,313][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_4000_state/model.safetensors +[2026-06-18 01:41:36,280][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_4000_state/optimizer.bin +[2026-06-18 01:41:36,281][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_4000_state/sampler.bin +[2026-06-18 01:41:36,282][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_4000_state/sampler_1.bin +[2026-06-18 01:41:36,332][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_4000_state/random_states_0.pkl +[2026-06-18 01:41:36,368][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4000 +[2026-06-18 01:41:36,368][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 01:41:36,386][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 01:41:41,289][starVLA.training.train_starvla][INFO] - Step 4001, Loss: {'train/loss': 0.3772258162498474, 'train/grad_norm_pre_clip': 11.015788078308105, 'timing/data': 0.0012461664155125618, 'timing/model': 1.6024808743968606, 'learning_rate/qwen_vl_interface': 4.2814665822946825e-06, 'learning_rate/action_model': 4.281466582294682e-05, 'epoch': 1.54}) +[2026-06-18 01:41:46,106][starVLA.training.train_starvla][INFO] - Step 4002, Loss: {'train/loss': 0.585862398147583, 'train/grad_norm_pre_clip': 12.729597091674805, 'timing/data': 0.011514474637806416, 'timing/model': 1.4637945229187608, 'learning_rate/qwen_vl_interface': 4.279349427367395e-06, 'learning_rate/action_model': 4.279349427367395e-05, 'epoch': 1.54}) +[2026-06-18 01:41:50,680][starVLA.training.train_starvla][INFO] - Step 4003, Loss: {'train/loss': 0.46461033821105957, 'train/grad_norm_pre_clip': 11.936615943908691, 'timing/data': 0.026491847820580006, 'timing/model': 1.2064550705254078, 'learning_rate/qwen_vl_interface': 4.277232473657018e-06, 'learning_rate/action_model': 4.277232473657018e-05, 'epoch': 1.54}) +[2026-06-18 01:41:55,911][starVLA.training.train_starvla][INFO] - Step 4004, Loss: {'train/loss': 0.4609556496143341, 'train/grad_norm_pre_clip': 11.171581268310547, 'timing/data': 0.034016930498182774, 'timing/model': 1.17859151866287, 'learning_rate/qwen_vl_interface': 4.2751157216023934e-06, 'learning_rate/action_model': 4.2751157216023936e-05, 'epoch': 1.54}) +[2026-06-18 01:42:01,263][starVLA.training.train_starvla][INFO] - Step 4005, Loss: {'train/loss': 0.1978188306093216, 'train/grad_norm_pre_clip': 7.658858299255371, 'timing/data': 0.024616367183625698, 'timing/model': 1.2938172360882163, 'learning_rate/qwen_vl_interface': 4.272999171642329e-06, 'learning_rate/action_model': 4.272999171642329e-05, 'epoch': 1.54}) +[2026-06-18 01:42:06,738][starVLA.training.train_starvla][INFO] - Step 4006, Loss: {'train/loss': 0.5056073665618896, 'train/grad_norm_pre_clip': 13.029752731323242, 'timing/data': 0.028195220977067947, 'timing/model': 1.4631154760718346, 'learning_rate/qwen_vl_interface': 4.270882824215588e-06, 'learning_rate/action_model': 4.270882824215588e-05, 'epoch': 1.54}) +[2026-06-18 01:42:12,401][starVLA.training.train_starvla][INFO] - Step 4007, Loss: {'train/loss': 0.4294234812259674, 'train/grad_norm_pre_clip': 13.589524269104004, 'timing/data': 0.02144474908709526, 'timing/model': 1.726467634551227, 'learning_rate/qwen_vl_interface': 4.268766679760891e-06, 'learning_rate/action_model': 4.268766679760891e-05, 'epoch': 1.54}) +[2026-06-18 01:42:21,799][starVLA.training.train_starvla][INFO] - Step 4008, Loss: {'train/loss': 0.45732659101486206, 'train/grad_norm_pre_clip': 13.386072158813477, 'timing/data': 0.03849451802670956, 'timing/model': 6.033275785855949, 'learning_rate/qwen_vl_interface': 4.266650738716916e-06, 'learning_rate/action_model': 4.266650738716916e-05, 'epoch': 1.54}) +[2026-06-18 01:42:33,088][starVLA.training.train_starvla][INFO] - Step 4009, Loss: {'train/loss': 0.08641037344932556, 'train/grad_norm_pre_clip': 6.673099517822266, 'timing/data': 0.0011973343789577484, 'timing/model': 4.627175562083721, 'learning_rate/qwen_vl_interface': 4.264535001522302e-06, 'learning_rate/action_model': 4.2645350015223025e-05, 'epoch': 1.54}) +[2026-06-18 01:42:41,793][starVLA.training.train_starvla][INFO] - Step 4010, Loss: {'train/loss': 0.41121265292167664, 'train/grad_norm_pre_clip': 18.778560638427734, 'timing/data': 0.0020050499588251114, 'timing/model': 1.459498992189765, 'learning_rate/qwen_vl_interface': 4.262419468615644e-06, 'learning_rate/action_model': 4.2624194686156435e-05, 'epoch': 1.54}) +[2026-06-18 01:42:53,364][starVLA.training.train_starvla][INFO] - Step 4011, Loss: {'train/loss': 0.5784974694252014, 'train/grad_norm_pre_clip': 17.181793212890625, 'timing/data': 0.0013167131692171097, 'timing/model': 1.7822889937087893, 'learning_rate/qwen_vl_interface': 4.2603041404354925e-06, 'learning_rate/action_model': 4.260304140435492e-05, 'epoch': 1.54}) +[2026-06-18 01:43:07,327][starVLA.training.train_starvla][INFO] - Step 4012, Loss: {'train/loss': 0.3089790940284729, 'train/grad_norm_pre_clip': 11.561997413635254, 'timing/data': 0.0011970922350883484, 'timing/model': 3.4528050357475877, 'learning_rate/qwen_vl_interface': 4.258189017420356e-06, 'learning_rate/action_model': 4.258189017420355e-05, 'epoch': 1.54}) +[2026-06-18 01:43:25,691][starVLA.training.train_starvla][INFO] - Step 4013, Loss: {'train/loss': 0.41264960169792175, 'train/grad_norm_pre_clip': 13.738062858581543, 'timing/data': 0.0032863104715943336, 'timing/model': 2.6504661291837692, 'learning_rate/qwen_vl_interface': 4.2560741000087035e-06, 'learning_rate/action_model': 4.2560741000087034e-05, 'epoch': 1.54}) +[2026-06-18 01:43:35,848][starVLA.training.train_starvla][INFO] - Step 4014, Loss: {'train/loss': 0.367182195186615, 'train/grad_norm_pre_clip': 12.05126667022705, 'timing/data': 0.001779765821993351, 'timing/model': 4.101556498557329, 'learning_rate/qwen_vl_interface': 4.253959388638961e-06, 'learning_rate/action_model': 4.2539593886389604e-05, 'epoch': 1.54}) +[2026-06-18 01:43:52,661][starVLA.training.train_starvla][INFO] - Step 4015, Loss: {'train/loss': 0.3259393274784088, 'train/grad_norm_pre_clip': 9.453804969787598, 'timing/data': 0.00131208635866642, 'timing/model': 4.777009752579033, 'learning_rate/qwen_vl_interface': 4.251844883749507e-06, 'learning_rate/action_model': 4.2518448837495066e-05, 'epoch': 1.54}) +[2026-06-18 01:44:00,309][starVLA.training.train_starvla][INFO] - Step 4016, Loss: {'train/loss': 0.39018717408180237, 'train/grad_norm_pre_clip': 12.871997833251953, 'timing/data': 0.0010708728805184364, 'timing/model': 1.2480781208723783, 'learning_rate/qwen_vl_interface': 4.249730585778684e-06, 'learning_rate/action_model': 4.2497305857786834e-05, 'epoch': 1.54}) +[2026-06-18 01:44:06,021][starVLA.training.train_starvla][INFO] - Step 4017, Loss: {'train/loss': 0.6363438367843628, 'train/grad_norm_pre_clip': 21.320512771606445, 'timing/data': 0.0015635369345545769, 'timing/model': 1.8228515917435288, 'learning_rate/qwen_vl_interface': 4.247616495164788e-06, 'learning_rate/action_model': 4.247616495164787e-05, 'epoch': 1.55}) +[2026-06-18 01:44:13,588][starVLA.training.train_starvla][INFO] - Step 4018, Loss: {'train/loss': 0.4885837435722351, 'train/grad_norm_pre_clip': 9.613320350646973, 'timing/data': 0.0012113545089960098, 'timing/model': 1.140422873198986, 'learning_rate/qwen_vl_interface': 4.24550261234607e-06, 'learning_rate/action_model': 4.2455026123460694e-05, 'epoch': 1.55}) +[2026-06-18 01:44:18,745][starVLA.training.train_starvla][INFO] - Step 4019, Loss: {'train/loss': 0.22454240918159485, 'train/grad_norm_pre_clip': 8.927349090576172, 'timing/data': 0.0013153143227100372, 'timing/model': 1.0918767396360636, 'learning_rate/qwen_vl_interface': 4.2433889377607405e-06, 'learning_rate/action_model': 4.24338893776074e-05, 'epoch': 1.55}) +[2026-06-18 01:44:24,043][starVLA.training.train_starvla][INFO] - Step 4020, Loss: {'train/loss': 0.15771552920341492, 'train/grad_norm_pre_clip': 6.690724849700928, 'timing/data': 0.0014212233945727348, 'timing/model': 1.138281055726111, 'learning_rate/qwen_vl_interface': 4.241275471846968e-06, 'learning_rate/action_model': 4.241275471846968e-05, 'epoch': 1.55}) +[2026-06-18 01:44:29,779][starVLA.training.train_starvla][INFO] - Step 4021, Loss: {'train/loss': 0.3376498818397522, 'train/grad_norm_pre_clip': 12.081002235412598, 'timing/data': 0.0009169233962893486, 'timing/model': 1.1439486192539334, 'learning_rate/qwen_vl_interface': 4.239162215042876e-06, 'learning_rate/action_model': 4.239162215042876e-05, 'epoch': 1.55}) +[2026-06-18 01:44:35,637][starVLA.training.train_starvla][INFO] - Step 4022, Loss: {'train/loss': 0.5220009088516235, 'train/grad_norm_pre_clip': 10.889718055725098, 'timing/data': 0.0013691354542970657, 'timing/model': 1.5923155108466744, 'learning_rate/qwen_vl_interface': 4.237049167786544e-06, 'learning_rate/action_model': 4.2370491677865444e-05, 'epoch': 1.55}) +[2026-06-18 01:44:40,843][starVLA.training.train_starvla][INFO] - Step 4023, Loss: {'train/loss': 0.5210042595863342, 'train/grad_norm_pre_clip': 12.427021026611328, 'timing/data': 0.0018891729414463043, 'timing/model': 1.267391150817275, 'learning_rate/qwen_vl_interface': 4.234936330516009e-06, 'learning_rate/action_model': 4.234936330516009e-05, 'epoch': 1.55}) +[2026-06-18 01:44:54,243][starVLA.training.train_starvla][INFO] - Step 4024, Loss: {'train/loss': 0.5474048256874084, 'train/grad_norm_pre_clip': 20.861186981201172, 'timing/data': 0.002899477258324623, 'timing/model': 4.876435765996575, 'learning_rate/qwen_vl_interface': 4.2328237036692675e-06, 'learning_rate/action_model': 4.232823703669267e-05, 'epoch': 1.55}) +[2026-06-18 01:45:04,271][starVLA.training.train_starvla][INFO] - Step 4025, Loss: {'train/loss': 0.3668034076690674, 'train/grad_norm_pre_clip': 11.086677551269531, 'timing/data': 0.001161498948931694, 'timing/model': 1.3024525176733732, 'learning_rate/qwen_vl_interface': 4.2307112876842645e-06, 'learning_rate/action_model': 4.2307112876842645e-05, 'epoch': 1.55}) +[2026-06-18 01:45:09,376][starVLA.training.train_starvla][INFO] - Step 4026, Loss: {'train/loss': 0.6144306659698486, 'train/grad_norm_pre_clip': 15.67776870727539, 'timing/data': 0.0012056892737746239, 'timing/model': 1.1407855674624443, 'learning_rate/qwen_vl_interface': 4.228599082998908e-06, 'learning_rate/action_model': 4.228599082998908e-05, 'epoch': 1.55}) +[2026-06-18 01:45:15,043][starVLA.training.train_starvla][INFO] - Step 4027, Loss: {'train/loss': 0.30400362610816956, 'train/grad_norm_pre_clip': 12.53766918182373, 'timing/data': 0.002498367801308632, 'timing/model': 1.6511916182935238, 'learning_rate/qwen_vl_interface': 4.226487090051062e-06, 'learning_rate/action_model': 4.2264870900510624e-05, 'epoch': 1.55}) +[2026-06-18 01:45:19,793][starVLA.training.train_starvla][INFO] - Step 4028, Loss: {'train/loss': 0.3237398862838745, 'train/grad_norm_pre_clip': 11.932300567626953, 'timing/data': 0.0013178978115320206, 'timing/model': 1.398555988445878, 'learning_rate/qwen_vl_interface': 4.224375309278543e-06, 'learning_rate/action_model': 4.2243753092785434e-05, 'epoch': 1.55}) +[2026-06-18 01:45:24,516][starVLA.training.train_starvla][INFO] - Step 4029, Loss: {'train/loss': 0.5403754115104675, 'train/grad_norm_pre_clip': 10.471858978271484, 'timing/data': 0.0011764969676733017, 'timing/model': 1.08448270522058, 'learning_rate/qwen_vl_interface': 4.222263741119127e-06, 'learning_rate/action_model': 4.222263741119126e-05, 'epoch': 1.55}) +[2026-06-18 01:45:29,676][starVLA.training.train_starvla][INFO] - Step 4030, Loss: {'train/loss': 0.5252159833908081, 'train/grad_norm_pre_clip': 13.34492301940918, 'timing/data': 0.0009323609992861748, 'timing/model': 1.2012964338064194, 'learning_rate/qwen_vl_interface': 4.220152386010541e-06, 'learning_rate/action_model': 4.2201523860105405e-05, 'epoch': 1.55}) +[2026-06-18 01:45:34,654][starVLA.training.train_starvla][INFO] - Step 4031, Loss: {'train/loss': 0.5723831057548523, 'train/grad_norm_pre_clip': 15.45764446258545, 'timing/data': 0.002027585171163082, 'timing/model': 1.7149037653580308, 'learning_rate/qwen_vl_interface': 4.218041244390476e-06, 'learning_rate/action_model': 4.218041244390475e-05, 'epoch': 1.55}) +[2026-06-18 01:45:39,702][starVLA.training.train_starvla][INFO] - Step 4032, Loss: {'train/loss': 0.39496755599975586, 'train/grad_norm_pre_clip': 10.771622657775879, 'timing/data': 0.0045714592561125755, 'timing/model': 1.492998342961073, 'learning_rate/qwen_vl_interface': 4.215930316696571e-06, 'learning_rate/action_model': 4.215930316696571e-05, 'epoch': 1.55}) +[2026-06-18 01:45:45,592][starVLA.training.train_starvla][INFO] - Step 4033, Loss: {'train/loss': 0.7709760665893555, 'train/grad_norm_pre_clip': 12.265933990478516, 'timing/data': 0.0009032916277647018, 'timing/model': 1.1265033138915896, 'learning_rate/qwen_vl_interface': 4.213819603366424e-06, 'learning_rate/action_model': 4.213819603366424e-05, 'epoch': 1.55}) +[2026-06-18 01:45:50,720][starVLA.training.train_starvla][INFO] - Step 4034, Loss: {'train/loss': 0.20870184898376465, 'train/grad_norm_pre_clip': 8.853171348571777, 'timing/data': 0.005216691642999649, 'timing/model': 1.1054300954565406, 'learning_rate/qwen_vl_interface': 4.211709104837587e-06, 'learning_rate/action_model': 4.211709104837587e-05, 'epoch': 1.55}) +[2026-06-18 01:46:01,064][starVLA.training.train_starvla][INFO] - Step 4035, Loss: {'train/loss': 0.10294655710458755, 'train/grad_norm_pre_clip': 4.493890285491943, 'timing/data': 0.002655675634741783, 'timing/model': 2.9167437106370926, 'learning_rate/qwen_vl_interface': 4.2095988215475705e-06, 'learning_rate/action_model': 4.20959882154757e-05, 'epoch': 1.55}) +[2026-06-18 01:46:07,057][starVLA.training.train_starvla][INFO] - Step 4036, Loss: {'train/loss': 0.17456457018852234, 'train/grad_norm_pre_clip': 6.095841407775879, 'timing/data': 0.0015556542202830315, 'timing/model': 1.8672554166987538, 'learning_rate/qwen_vl_interface': 4.207488753933838e-06, 'learning_rate/action_model': 4.207488753933838e-05, 'epoch': 1.55}) +[2026-06-18 01:46:12,532][starVLA.training.train_starvla][INFO] - Step 4037, Loss: {'train/loss': 0.5335674285888672, 'train/grad_norm_pre_clip': 19.582353591918945, 'timing/data': 0.0018555931746959686, 'timing/model': 1.7045137602835894, 'learning_rate/qwen_vl_interface': 4.205378902433807e-06, 'learning_rate/action_model': 4.205378902433807e-05, 'epoch': 1.55}) +[2026-06-18 01:46:22,092][starVLA.training.train_starvla][INFO] - Step 4038, Loss: {'train/loss': 0.440114289522171, 'train/grad_norm_pre_clip': 19.577253341674805, 'timing/data': 0.010192178189754486, 'timing/model': 2.9779690857976675, 'learning_rate/qwen_vl_interface': 4.203269267484856e-06, 'learning_rate/action_model': 4.203269267484856e-05, 'epoch': 1.55}) +[2026-06-18 01:46:28,213][starVLA.training.train_starvla][INFO] - Step 4039, Loss: {'train/loss': 0.2653878629207611, 'train/grad_norm_pre_clip': 9.761367797851562, 'timing/data': 0.0013959389179944992, 'timing/model': 1.3306622579693794, 'learning_rate/qwen_vl_interface': 4.201159849524313e-06, 'learning_rate/action_model': 4.201159849524312e-05, 'epoch': 1.55}) +[2026-06-18 01:46:37,257][starVLA.training.train_starvla][INFO] - Step 4040, Loss: {'train/loss': 0.3611847162246704, 'train/grad_norm_pre_clip': 13.20310115814209, 'timing/data': 0.001826990395784378, 'timing/model': 1.10169424302876, 'learning_rate/qwen_vl_interface': 4.19905064898946e-06, 'learning_rate/action_model': 4.199050648989459e-05, 'epoch': 1.55}) +[2026-06-18 01:46:43,057][starVLA.training.train_starvla][INFO] - Step 4041, Loss: {'train/loss': 0.48220938444137573, 'train/grad_norm_pre_clip': 14.915962219238281, 'timing/data': 0.0019450215622782707, 'timing/model': 2.082292467355728, 'learning_rate/qwen_vl_interface': 4.196941666317537e-06, 'learning_rate/action_model': 4.196941666317537e-05, 'epoch': 1.55}) +[2026-06-18 01:46:52,441][starVLA.training.train_starvla][INFO] - Step 4042, Loss: {'train/loss': 0.38828375935554504, 'train/grad_norm_pre_clip': 13.269391059875488, 'timing/data': 0.0011920640245079994, 'timing/model': 1.4585517905652523, 'learning_rate/qwen_vl_interface': 4.194832901945741e-06, 'learning_rate/action_model': 4.194832901945741e-05, 'epoch': 1.55}) +[2026-06-18 01:46:58,870][starVLA.training.train_starvla][INFO] - Step 4043, Loss: {'train/loss': 0.4822303354740143, 'train/grad_norm_pre_clip': 12.544878959655762, 'timing/data': 0.0011084247380495071, 'timing/model': 1.4739790707826614, 'learning_rate/qwen_vl_interface': 4.192724356311221e-06, 'learning_rate/action_model': 4.19272435631122e-05, 'epoch': 1.56}) +[2026-06-18 01:47:03,996][starVLA.training.train_starvla][INFO] - Step 4044, Loss: {'train/loss': 0.3750840425491333, 'train/grad_norm_pre_clip': 11.291643142700195, 'timing/data': 0.0011303508654236794, 'timing/model': 1.1227199770510197, 'learning_rate/qwen_vl_interface': 4.190616029851079e-06, 'learning_rate/action_model': 4.190616029851079e-05, 'epoch': 1.56}) +[2026-06-18 01:47:11,075][starVLA.training.train_starvla][INFO] - Step 4045, Loss: {'train/loss': 0.4644656479358673, 'train/grad_norm_pre_clip': 16.412403106689453, 'timing/data': 0.001179765909910202, 'timing/model': 2.898366054520011, 'learning_rate/qwen_vl_interface': 4.1885079230023715e-06, 'learning_rate/action_model': 4.1885079230023715e-05, 'epoch': 1.56}) +[2026-06-18 01:47:16,526][starVLA.training.train_starvla][INFO] - Step 4046, Loss: {'train/loss': 0.6888629794120789, 'train/grad_norm_pre_clip': 16.28803253173828, 'timing/data': 0.0012517189607024193, 'timing/model': 1.3621023884043097, 'learning_rate/qwen_vl_interface': 4.186400036202117e-06, 'learning_rate/action_model': 4.1864000362021175e-05, 'epoch': 1.56}) +[2026-06-18 01:47:21,724][starVLA.training.train_starvla][INFO] - Step 4047, Loss: {'train/loss': 0.5355249047279358, 'train/grad_norm_pre_clip': 12.121066093444824, 'timing/data': 0.0010582050308585167, 'timing/model': 1.4684006534516811, 'learning_rate/qwen_vl_interface': 4.184292369887279e-06, 'learning_rate/action_model': 4.1842923698872785e-05, 'epoch': 1.56}) +[2026-06-18 01:47:27,645][starVLA.training.train_starvla][INFO] - Step 4048, Loss: {'train/loss': 0.5325266122817993, 'train/grad_norm_pre_clip': 15.503327369689941, 'timing/data': 0.0012318706139922142, 'timing/model': 1.3422168428078294, 'learning_rate/qwen_vl_interface': 4.182184924494778e-06, 'learning_rate/action_model': 4.1821849244947773e-05, 'epoch': 1.56}) +[2026-06-18 01:47:32,469][starVLA.training.train_starvla][INFO] - Step 4049, Loss: {'train/loss': 0.2879665791988373, 'train/grad_norm_pre_clip': 9.789118766784668, 'timing/data': 0.001334378495812416, 'timing/model': 1.1534377513453364, 'learning_rate/qwen_vl_interface': 4.1800777004614915e-06, 'learning_rate/action_model': 4.180077700461492e-05, 'epoch': 1.56}) +[2026-06-18 01:47:37,501][starVLA.training.train_starvla][INFO] - Step 4050, Loss: {'train/loss': 0.2570487856864929, 'train/grad_norm_pre_clip': 10.929842948913574, 'timing/data': 0.0007531018927693367, 'timing/model': 1.2897898498922586, 'learning_rate/qwen_vl_interface': 4.177970698224249e-06, 'learning_rate/action_model': 4.17797069822425e-05, 'epoch': 1.56}) +[2026-06-18 01:47:42,873][starVLA.training.train_starvla][INFO] - Step 4051, Loss: {'train/loss': 0.5386207699775696, 'train/grad_norm_pre_clip': 9.796415328979492, 'timing/data': 0.001342535950243473, 'timing/model': 1.9324198784306645, 'learning_rate/qwen_vl_interface': 4.175863918219836e-06, 'learning_rate/action_model': 4.175863918219836e-05, 'epoch': 1.56}) +[2026-06-18 01:47:47,894][starVLA.training.train_starvla][INFO] - Step 4052, Loss: {'train/loss': 0.45121878385543823, 'train/grad_norm_pre_clip': 12.10092830657959, 'timing/data': 0.002071438357234001, 'timing/model': 1.4752163589000702, 'learning_rate/qwen_vl_interface': 4.173757360884986e-06, 'learning_rate/action_model': 4.173757360884986e-05, 'epoch': 1.56}) +[2026-06-18 01:47:54,243][starVLA.training.train_starvla][INFO] - Step 4053, Loss: {'train/loss': 0.29037797451019287, 'train/grad_norm_pre_clip': 10.33590316772461, 'timing/data': 0.0015194350853562355, 'timing/model': 2.7770536141470075, 'learning_rate/qwen_vl_interface': 4.171651026656396e-06, 'learning_rate/action_model': 4.171651026656395e-05, 'epoch': 1.56}) +[2026-06-18 01:48:01,413][starVLA.training.train_starvla][INFO] - Step 4054, Loss: {'train/loss': 0.2264997959136963, 'train/grad_norm_pre_clip': 9.520477294921875, 'timing/data': 0.0011709313839673996, 'timing/model': 1.1773041607812047, 'learning_rate/qwen_vl_interface': 4.1695449159707066e-06, 'learning_rate/action_model': 4.169544915970707e-05, 'epoch': 1.56}) +[2026-06-18 01:48:09,446][starVLA.training.train_starvla][INFO] - Step 4055, Loss: {'train/loss': 0.4666033983230591, 'train/grad_norm_pre_clip': 14.94990062713623, 'timing/data': 0.0014312304556369781, 'timing/model': 1.1301786452531815, 'learning_rate/qwen_vl_interface': 4.167439029264519e-06, 'learning_rate/action_model': 4.167439029264519e-05, 'epoch': 1.56}) +[2026-06-18 01:48:14,810][starVLA.training.train_starvla][INFO] - Step 4056, Loss: {'train/loss': 0.4873093366622925, 'train/grad_norm_pre_clip': 12.040212631225586, 'timing/data': 0.0012202206999063492, 'timing/model': 1.4524275409057736, 'learning_rate/qwen_vl_interface': 4.165333366974385e-06, 'learning_rate/action_model': 4.165333366974385e-05, 'epoch': 1.56}) +[2026-06-18 01:48:19,961][starVLA.training.train_starvla][INFO] - Step 4057, Loss: {'train/loss': 0.33623960614204407, 'train/grad_norm_pre_clip': 9.629828453063965, 'timing/data': 0.001203545369207859, 'timing/model': 1.4949524132534862, 'learning_rate/qwen_vl_interface': 4.163227929536812e-06, 'learning_rate/action_model': 4.163227929536812e-05, 'epoch': 1.56}) +[2026-06-18 01:48:25,647][starVLA.training.train_starvla][INFO] - Step 4058, Loss: {'train/loss': 0.4811152517795563, 'train/grad_norm_pre_clip': 17.531843185424805, 'timing/data': 0.001167643815279007, 'timing/model': 2.4589086761698127, 'learning_rate/qwen_vl_interface': 4.1611227173882596e-06, 'learning_rate/action_model': 4.1611227173882596e-05, 'epoch': 1.56}) +[2026-06-18 01:48:30,781][starVLA.training.train_starvla][INFO] - Step 4059, Loss: {'train/loss': 0.2984572649002075, 'train/grad_norm_pre_clip': 8.508517265319824, 'timing/data': 0.001297404058277607, 'timing/model': 1.2769043985754251, 'learning_rate/qwen_vl_interface': 4.1590177309651365e-06, 'learning_rate/action_model': 4.159017730965136e-05, 'epoch': 1.56}) +[2026-06-18 01:48:35,814][starVLA.training.train_starvla][INFO] - Step 4060, Loss: {'train/loss': 0.19611966609954834, 'train/grad_norm_pre_clip': 8.265969276428223, 'timing/data': 0.00174715556204319, 'timing/model': 1.2886131778359413, 'learning_rate/qwen_vl_interface': 4.156912970703815e-06, 'learning_rate/action_model': 4.1569129707038155e-05, 'epoch': 1.56}) +[2026-06-18 01:48:40,634][starVLA.training.train_starvla][INFO] - Step 4061, Loss: {'train/loss': 0.27751147747039795, 'train/grad_norm_pre_clip': 10.617250442504883, 'timing/data': 0.000873875804245472, 'timing/model': 1.0909082191064954, 'learning_rate/qwen_vl_interface': 4.15480843704061e-06, 'learning_rate/action_model': 4.154808437040609e-05, 'epoch': 1.56}) +[2026-06-18 01:48:46,210][starVLA.training.train_starvla][INFO] - Step 4062, Loss: {'train/loss': 0.35353729128837585, 'train/grad_norm_pre_clip': 9.335330963134766, 'timing/data': 0.0011700782924890518, 'timing/model': 2.1287868870422244, 'learning_rate/qwen_vl_interface': 4.152704130411794e-06, 'learning_rate/action_model': 4.152704130411793e-05, 'epoch': 1.56}) +[2026-06-18 01:48:51,153][starVLA.training.train_starvla][INFO] - Step 4063, Loss: {'train/loss': 0.34961986541748047, 'train/grad_norm_pre_clip': 10.281843185424805, 'timing/data': 0.0016342950984835625, 'timing/model': 1.1337386490777135, 'learning_rate/qwen_vl_interface': 4.150600051253591e-06, 'learning_rate/action_model': 4.150600051253591e-05, 'epoch': 1.56}) +[2026-06-18 01:49:01,841][starVLA.training.train_starvla][INFO] - Step 4064, Loss: {'train/loss': 0.4850948750972748, 'train/grad_norm_pre_clip': 13.236268043518066, 'timing/data': 0.0012198584154248238, 'timing/model': 3.8473570700734854, 'learning_rate/qwen_vl_interface': 4.148496200002183e-06, 'learning_rate/action_model': 4.148496200002183e-05, 'epoch': 1.56}) +[2026-06-18 01:49:16,066][starVLA.training.train_starvla][INFO] - Step 4065, Loss: {'train/loss': 0.4943600594997406, 'train/grad_norm_pre_clip': 14.061890602111816, 'timing/data': 0.0015507349744439125, 'timing/model': 1.4579582810401917, 'learning_rate/qwen_vl_interface': 4.146392577093697e-06, 'learning_rate/action_model': 4.1463925770936974e-05, 'epoch': 1.56}) +[2026-06-18 01:49:21,352][starVLA.training.train_starvla][INFO] - Step 4066, Loss: {'train/loss': 0.43305426836013794, 'train/grad_norm_pre_clip': 16.468461990356445, 'timing/data': 0.0012205475941300392, 'timing/model': 1.1151238000020385, 'learning_rate/qwen_vl_interface': 4.144289182964218e-06, 'learning_rate/action_model': 4.144289182964218e-05, 'epoch': 1.56}) +[2026-06-18 01:49:26,042][starVLA.training.train_starvla][INFO] - Step 4067, Loss: {'train/loss': 0.26922908425331116, 'train/grad_norm_pre_clip': 10.447564125061035, 'timing/data': 0.0014110654592514038, 'timing/model': 1.4270474184304476, 'learning_rate/qwen_vl_interface': 4.14218601804978e-06, 'learning_rate/action_model': 4.1421860180497796e-05, 'epoch': 1.56}) +[2026-06-18 01:49:31,216][starVLA.training.train_starvla][INFO] - Step 4068, Loss: {'train/loss': 0.741223156452179, 'train/grad_norm_pre_clip': 15.75287914276123, 'timing/data': 0.001134222373366356, 'timing/model': 1.123874368146062, 'learning_rate/qwen_vl_interface': 4.1400830827863725e-06, 'learning_rate/action_model': 4.1400830827863725e-05, 'epoch': 1.56}) +[2026-06-18 01:49:39,226][starVLA.training.train_starvla][INFO] - Step 4069, Loss: {'train/loss': 0.3925777077674866, 'train/grad_norm_pre_clip': 13.147540092468262, 'timing/data': 0.001187717542052269, 'timing/model': 2.6482232864946127, 'learning_rate/qwen_vl_interface': 4.137980377609937e-06, 'learning_rate/action_model': 4.137980377609937e-05, 'epoch': 1.57}) +[2026-06-18 01:49:44,436][starVLA.training.train_starvla][INFO] - Step 4070, Loss: {'train/loss': 0.4187586307525635, 'train/grad_norm_pre_clip': 11.618717193603516, 'timing/data': 0.0010377289727330208, 'timing/model': 1.0936333611607552, 'learning_rate/qwen_vl_interface': 4.135877902956364e-06, 'learning_rate/action_model': 4.1358779029563636e-05, 'epoch': 1.57}) +[2026-06-18 01:49:50,168][starVLA.training.train_starvla][INFO] - Step 4071, Loss: {'train/loss': 0.4085630178451538, 'train/grad_norm_pre_clip': 11.805604934692383, 'timing/data': 0.0013185814023017883, 'timing/model': 1.5667292708531022, 'learning_rate/qwen_vl_interface': 4.133775659261503e-06, 'learning_rate/action_model': 4.133775659261502e-05, 'epoch': 1.57}) +[2026-06-18 01:49:55,503][starVLA.training.train_starvla][INFO] - Step 4072, Loss: {'train/loss': 0.3847014307975769, 'train/grad_norm_pre_clip': 12.196141242980957, 'timing/data': 0.001606806181371212, 'timing/model': 1.5921424198895693, 'learning_rate/qwen_vl_interface': 4.131673646961149e-06, 'learning_rate/action_model': 4.131673646961149e-05, 'epoch': 1.57}) +[2026-06-18 01:50:05,503][starVLA.training.train_starvla][INFO] - Step 4073, Loss: {'train/loss': 0.32151105999946594, 'train/grad_norm_pre_clip': 10.523161888122559, 'timing/data': 0.0012778723612427711, 'timing/model': 3.48691033013165, 'learning_rate/qwen_vl_interface': 4.129571866491051e-06, 'learning_rate/action_model': 4.1295718664910504e-05, 'epoch': 1.57}) +[2026-06-18 01:50:10,927][starVLA.training.train_starvla][INFO] - Step 4074, Loss: {'train/loss': 0.21672768890857697, 'train/grad_norm_pre_clip': 13.993820190429688, 'timing/data': 0.0012155193835496902, 'timing/model': 1.1074353540316224, 'learning_rate/qwen_vl_interface': 4.127470318286908e-06, 'learning_rate/action_model': 4.127470318286908e-05, 'epoch': 1.57}) +[2026-06-18 01:50:16,246][starVLA.training.train_starvla][INFO] - Step 4075, Loss: {'train/loss': 0.5643755197525024, 'train/grad_norm_pre_clip': 18.91624641418457, 'timing/data': 0.0009508421644568443, 'timing/model': 1.1360488906502724, 'learning_rate/qwen_vl_interface': 4.125369002784378e-06, 'learning_rate/action_model': 4.1253690027843776e-05, 'epoch': 1.57}) +[2026-06-18 01:50:21,862][starVLA.training.train_starvla][INFO] - Step 4076, Loss: {'train/loss': 0.5739078521728516, 'train/grad_norm_pre_clip': 17.70075035095215, 'timing/data': 0.0011016624048352242, 'timing/model': 1.340824669227004, 'learning_rate/qwen_vl_interface': 4.123267920419064e-06, 'learning_rate/action_model': 4.123267920419064e-05, 'epoch': 1.57}) +[2026-06-18 01:50:31,498][starVLA.training.train_starvla][INFO] - Step 4077, Loss: {'train/loss': 0.40270140767097473, 'train/grad_norm_pre_clip': 15.442809104919434, 'timing/data': 0.001308969222009182, 'timing/model': 6.006285418756306, 'learning_rate/qwen_vl_interface': 4.121167071626522e-06, 'learning_rate/action_model': 4.121167071626522e-05, 'epoch': 1.57}) +[2026-06-18 01:50:38,692][starVLA.training.train_starvla][INFO] - Step 4078, Loss: {'train/loss': 0.29643216729164124, 'train/grad_norm_pre_clip': 15.53602409362793, 'timing/data': 0.0009505711495876312, 'timing/model': 2.114700640551746, 'learning_rate/qwen_vl_interface': 4.119066456842261e-06, 'learning_rate/action_model': 4.119066456842261e-05, 'epoch': 1.57}) +[2026-06-18 01:50:49,092][starVLA.training.train_starvla][INFO] - Step 4079, Loss: {'train/loss': 0.5131289958953857, 'train/grad_norm_pre_clip': 14.651068687438965, 'timing/data': 0.0011598775163292885, 'timing/model': 1.2019906537607312, 'learning_rate/qwen_vl_interface': 4.1169660765017416e-06, 'learning_rate/action_model': 4.1169660765017414e-05, 'epoch': 1.57}) +[2026-06-18 01:50:54,108][starVLA.training.train_starvla][INFO] - Step 4080, Loss: {'train/loss': 0.49692344665527344, 'train/grad_norm_pre_clip': 17.081174850463867, 'timing/data': 0.0013004960492253304, 'timing/model': 1.1884655309841037, 'learning_rate/qwen_vl_interface': 4.1148659310403725e-06, 'learning_rate/action_model': 4.1148659310403723e-05, 'epoch': 1.57}) +[2026-06-18 01:50:58,888][starVLA.training.train_starvla][INFO] - Step 4081, Loss: {'train/loss': 0.40526601672172546, 'train/grad_norm_pre_clip': 10.5426025390625, 'timing/data': 0.0012378403916954994, 'timing/model': 1.1487973863258958, 'learning_rate/qwen_vl_interface': 4.112766020893516e-06, 'learning_rate/action_model': 4.1127660208935164e-05, 'epoch': 1.57}) +[2026-06-18 01:51:06,500][starVLA.training.train_starvla][INFO] - Step 4082, Loss: {'train/loss': 0.44346755743026733, 'train/grad_norm_pre_clip': 13.36801528930664, 'timing/data': 0.0011061560362577438, 'timing/model': 1.5164557304233313, 'learning_rate/qwen_vl_interface': 4.11066634649649e-06, 'learning_rate/action_model': 4.1106663464964895e-05, 'epoch': 1.57}) +[2026-06-18 01:51:11,498][starVLA.training.train_starvla][INFO] - Step 4083, Loss: {'train/loss': 0.7804422974586487, 'train/grad_norm_pre_clip': 13.035551071166992, 'timing/data': 0.001309509389102459, 'timing/model': 1.224823766387999, 'learning_rate/qwen_vl_interface': 4.1085669082845545e-06, 'learning_rate/action_model': 4.108566908284555e-05, 'epoch': 1.57}) +[2026-06-18 01:51:17,758][starVLA.training.train_starvla][INFO] - Step 4084, Loss: {'train/loss': 0.40094149112701416, 'train/grad_norm_pre_clip': 12.990769386291504, 'timing/data': 0.001270507462322712, 'timing/model': 1.099720124155283, 'learning_rate/qwen_vl_interface': 4.10646770669293e-06, 'learning_rate/action_model': 4.106467706692929e-05, 'epoch': 1.57}) +[2026-06-18 01:51:23,211][starVLA.training.train_starvla][INFO] - Step 4085, Loss: {'train/loss': 0.2889965772628784, 'train/grad_norm_pre_clip': 10.67057991027832, 'timing/data': 0.0014578383415937424, 'timing/model': 1.1557169491425157, 'learning_rate/qwen_vl_interface': 4.104368742156774e-06, 'learning_rate/action_model': 4.104368742156774e-05, 'epoch': 1.57}) +[2026-06-18 01:51:28,940][starVLA.training.train_starvla][INFO] - Step 4086, Loss: {'train/loss': 0.440746933221817, 'train/grad_norm_pre_clip': 14.482569694519043, 'timing/data': 0.001218223012983799, 'timing/model': 1.6011796249076724, 'learning_rate/qwen_vl_interface': 4.102270015111217e-06, 'learning_rate/action_model': 4.102270015111217e-05, 'epoch': 1.57}) +[2026-06-18 01:51:35,001][starVLA.training.train_starvla][INFO] - Step 4087, Loss: {'train/loss': 0.4705987274646759, 'train/grad_norm_pre_clip': 9.63469123840332, 'timing/data': 0.0014035701751708984, 'timing/model': 1.7277363985776901, 'learning_rate/qwen_vl_interface': 4.100171525991318e-06, 'learning_rate/action_model': 4.100171525991318e-05, 'epoch': 1.57}) +[2026-06-18 01:51:39,724][starVLA.training.train_starvla][INFO] - Step 4088, Loss: {'train/loss': 0.3763315975666046, 'train/grad_norm_pre_clip': 17.819324493408203, 'timing/data': 0.0014310181140899658, 'timing/model': 1.2157021667808294, 'learning_rate/qwen_vl_interface': 4.098073275232098e-06, 'learning_rate/action_model': 4.098073275232098e-05, 'epoch': 1.57}) +[2026-06-18 01:51:44,698][starVLA.training.train_starvla][INFO] - Step 4089, Loss: {'train/loss': 0.5409314036369324, 'train/grad_norm_pre_clip': 13.936944961547852, 'timing/data': 0.001246483065187931, 'timing/model': 1.3082351675257087, 'learning_rate/qwen_vl_interface': 4.09597526326853e-06, 'learning_rate/action_model': 4.095975263268529e-05, 'epoch': 1.57}) +[2026-06-18 01:51:49,563][starVLA.training.train_starvla][INFO] - Step 4090, Loss: {'train/loss': 0.24311837553977966, 'train/grad_norm_pre_clip': 13.560547828674316, 'timing/data': 0.0013233041390776634, 'timing/model': 1.1532434755936265, 'learning_rate/qwen_vl_interface': 4.09387749053553e-06, 'learning_rate/action_model': 4.0938774905355296e-05, 'epoch': 1.57}) +[2026-06-18 01:51:54,357][starVLA.training.train_starvla][INFO] - Step 4091, Loss: {'train/loss': 0.5976619124412537, 'train/grad_norm_pre_clip': 28.274017333984375, 'timing/data': 0.000799303874373436, 'timing/model': 1.4200093066319823, 'learning_rate/qwen_vl_interface': 4.0917799574679725e-06, 'learning_rate/action_model': 4.091779957467973e-05, 'epoch': 1.57}) +[2026-06-18 01:51:59,501][starVLA.training.train_starvla][INFO] - Step 4092, Loss: {'train/loss': 0.6064717769622803, 'train/grad_norm_pre_clip': 18.918550491333008, 'timing/data': 0.0013558343052864075, 'timing/model': 1.6019897926598787, 'learning_rate/qwen_vl_interface': 4.0896826645006725e-06, 'learning_rate/action_model': 4.089682664500673e-05, 'epoch': 1.57}) +[2026-06-18 01:52:05,493][starVLA.training.train_starvla][INFO] - Step 4093, Loss: {'train/loss': 0.15468829870224, 'train/grad_norm_pre_clip': 6.7597832679748535, 'timing/data': 0.0012430911883711815, 'timing/model': 2.699960515834391, 'learning_rate/qwen_vl_interface': 4.087585612068408e-06, 'learning_rate/action_model': 4.0875856120684074e-05, 'epoch': 1.57}) +[2026-06-18 01:52:13,788][starVLA.training.train_starvla][INFO] - Step 4094, Loss: {'train/loss': 0.33769282698631287, 'train/grad_norm_pre_clip': 12.528056144714355, 'timing/data': 0.0012570414692163467, 'timing/model': 1.2362187812104821, 'learning_rate/qwen_vl_interface': 4.0854888006058955e-06, 'learning_rate/action_model': 4.085488800605896e-05, 'epoch': 1.57}) +[2026-06-18 01:52:18,692][starVLA.training.train_starvla][INFO] - Step 4095, Loss: {'train/loss': 0.7399813532829285, 'train/grad_norm_pre_clip': 17.760194778442383, 'timing/data': 0.0012467941269278526, 'timing/model': 1.1353397723287344, 'learning_rate/qwen_vl_interface': 4.083392230547808e-06, 'learning_rate/action_model': 4.083392230547808e-05, 'epoch': 1.58}) +[2026-06-18 01:52:23,333][starVLA.training.train_starvla][INFO] - Step 4096, Loss: {'train/loss': 0.10608477890491486, 'train/grad_norm_pre_clip': 4.7949676513671875, 'timing/data': 0.001138792373239994, 'timing/model': 1.131966263987124, 'learning_rate/qwen_vl_interface': 4.081295902328765e-06, 'learning_rate/action_model': 4.081295902328765e-05, 'epoch': 1.58}) +[2026-06-18 01:52:28,379][starVLA.training.train_starvla][INFO] - Step 4097, Loss: {'train/loss': 0.447141170501709, 'train/grad_norm_pre_clip': 17.940380096435547, 'timing/data': 0.0009278934448957443, 'timing/model': 1.5232247794046998, 'learning_rate/qwen_vl_interface': 4.079199816383341e-06, 'learning_rate/action_model': 4.0791998163833404e-05, 'epoch': 1.58}) +[2026-06-18 01:52:32,928][starVLA.training.train_starvla][INFO] - Step 4098, Loss: {'train/loss': 0.2649816870689392, 'train/grad_norm_pre_clip': 16.54133415222168, 'timing/data': 0.0011179940775036812, 'timing/model': 1.1943467948585749, 'learning_rate/qwen_vl_interface': 4.077103973146054e-06, 'learning_rate/action_model': 4.077103973146054e-05, 'epoch': 1.58}) +[2026-06-18 01:52:43,143][starVLA.training.train_starvla][INFO] - Step 4099, Loss: {'train/loss': 0.37692633271217346, 'train/grad_norm_pre_clip': 13.002466201782227, 'timing/data': 0.0011265510693192482, 'timing/model': 5.4179281974211335, 'learning_rate/qwen_vl_interface': 4.075008373051375e-06, 'learning_rate/action_model': 4.075008373051374e-05, 'epoch': 1.58}) +[2026-06-18 01:55:46,139][starVLA.training.train_starvla][INFO] - Step 4100, Loss: {'train/loss': 0.34210819005966187, 'train/grad_norm_pre_clip': 14.797801971435547, 'eval/loss': 0.45452454686164856, 'eval/latency_4/loss': 0.45452454686164856, 'eval/demon_fire_precision': 0.9584382871536524, 'eval/demon_fire_recall': 0.9886326729457616, 'eval/demon_fire_f1': 0.9733013589128697, 'eval/demon_fire_rate_model': 0.9327459618208517, 'eval/demon_fire_rate_teacher': 0.9042584434654919, 'eval/demon_left_precision': 0.7692307692307693, 'eval/demon_left_recall': 0.639269406392694, 'eval/demon_left_f1': 0.6982543640897755, 'eval/demon_left_rate_model': 0.10690161527165933, 'eval/demon_left_rate_teacher': 0.12863436123348018, 'eval/demon_right_precision': 0.8038095238095239, 'eval/demon_right_recall': 0.7617328519855595, 'eval/demon_right_f1': 0.7822057460611678, 'eval/demon_right_rate_model': 0.15418502202643172, 'eval/demon_right_rate_teacher': 0.16270190895741557, 'eval/demon_move_f1': 0.7402300550754717, 'eval/demon_cc_f1': 0.8892990586744061, 'eval/latency_4/demon_fire_precision': 0.9584382871536524, 'eval/latency_4/demon_fire_recall': 0.9886326729457616, 'eval/latency_4/demon_fire_f1': 0.9733013589128697, 'eval/latency_4/demon_fire_rate_model': 0.9327459618208517, 'eval/latency_4/demon_fire_rate_teacher': 0.9042584434654919, 'eval/latency_4/demon_left_precision': 0.7692307692307693, 'eval/latency_4/demon_left_recall': 0.639269406392694, 'eval/latency_4/demon_left_f1': 0.6982543640897755, 'eval/latency_4/demon_left_rate_model': 0.10690161527165933, 'eval/latency_4/demon_left_rate_teacher': 0.12863436123348018, 'eval/latency_4/demon_right_precision': 0.8038095238095239, 'eval/latency_4/demon_right_recall': 0.7617328519855595, 'eval/latency_4/demon_right_f1': 0.7822057460611678, 'eval/latency_4/demon_right_rate_model': 0.15418502202643172, 'eval/latency_4/demon_right_rate_teacher': 0.16270190895741557, 'eval/latency_4/demon_move_f1': 0.7402300550754717, 'eval/latency_4/demon_cc_f1': 0.8892990586744061, 'timing/data': 0.002011604607105255, 'timing/model': 1.2175015453249216, 'learning_rate/qwen_vl_interface': 4.072913016533726e-06, 'learning_rate/action_model': 4.072913016533726e-05, 'epoch': 1.58}) +[2026-06-18 01:55:46,170][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_4100_state +[2026-06-18 01:55:46,178][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 01:55:55,039][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_4100_state/model.safetensors +[2026-06-18 01:56:18,660][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_4100_state/optimizer.bin +[2026-06-18 01:56:18,662][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_4100_state/sampler.bin +[2026-06-18 01:56:18,663][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_4100_state/sampler_1.bin +[2026-06-18 01:56:18,666][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_4100_state/random_states_0.pkl +[2026-06-18 01:56:18,672][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4100 +[2026-06-18 01:56:18,673][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 01:56:18,717][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 01:56:30,945][starVLA.training.train_starvla][INFO] - Step 4101, Loss: {'train/loss': 0.349645733833313, 'train/grad_norm_pre_clip': 10.018503189086914, 'timing/data': 0.0010721143335103989, 'timing/model': 1.5408146055415273, 'learning_rate/qwen_vl_interface': 4.070817904027474e-06, 'learning_rate/action_model': 4.070817904027474e-05, 'epoch': 1.58}) +[2026-06-18 01:56:40,170][starVLA.training.train_starvla][INFO] - Step 4102, Loss: {'train/loss': 0.4328167140483856, 'train/grad_norm_pre_clip': 15.393671989440918, 'timing/data': 0.0012621767818927765, 'timing/model': 5.3485212894156575, 'learning_rate/qwen_vl_interface': 4.06872303596694e-06, 'learning_rate/action_model': 4.06872303596694e-05, 'epoch': 1.58}) +[2026-06-18 01:56:46,459][starVLA.training.train_starvla][INFO] - Step 4103, Loss: {'train/loss': 0.6727404594421387, 'train/grad_norm_pre_clip': 15.876632690429688, 'timing/data': 0.021816289983689785, 'timing/model': 1.259941678494215, 'learning_rate/qwen_vl_interface': 4.066628412786391e-06, 'learning_rate/action_model': 4.066628412786391e-05, 'epoch': 1.58}) +[2026-06-18 01:56:58,797][starVLA.training.train_starvla][INFO] - Step 4104, Loss: {'train/loss': 0.47164830565452576, 'train/grad_norm_pre_clip': 13.495927810668945, 'timing/data': 0.031708777882158756, 'timing/model': 1.3120076041668653, 'learning_rate/qwen_vl_interface': 4.0645340349200466e-06, 'learning_rate/action_model': 4.0645340349200464e-05, 'epoch': 1.58}) +[2026-06-18 01:57:08,599][starVLA.training.train_starvla][INFO] - Step 4105, Loss: {'train/loss': 0.5176835060119629, 'train/grad_norm_pre_clip': 14.693570137023926, 'timing/data': 0.059968139976263046, 'timing/model': 4.440476334653795, 'learning_rate/qwen_vl_interface': 4.062439902802073e-06, 'learning_rate/action_model': 4.0624399028020725e-05, 'epoch': 1.58}) +[2026-06-18 01:57:13,820][starVLA.training.train_starvla][INFO] - Step 4106, Loss: {'train/loss': 0.3435910940170288, 'train/grad_norm_pre_clip': 12.671239852905273, 'timing/data': 0.03290037252008915, 'timing/model': 1.1608012095093727, 'learning_rate/qwen_vl_interface': 4.060346016866584e-06, 'learning_rate/action_model': 4.060346016866584e-05, 'epoch': 1.58}) +[2026-06-18 01:57:28,352][starVLA.training.train_starvla][INFO] - Step 4107, Loss: {'train/loss': 0.5599318742752075, 'train/grad_norm_pre_clip': 14.267866134643555, 'timing/data': 0.033116783015429974, 'timing/model': 3.421861114911735, 'learning_rate/qwen_vl_interface': 4.058252377547645e-06, 'learning_rate/action_model': 4.058252377547645e-05, 'epoch': 1.58}) +[2026-06-18 01:57:33,072][starVLA.training.train_starvla][INFO] - Step 4108, Loss: {'train/loss': 0.5838741660118103, 'train/grad_norm_pre_clip': 11.683571815490723, 'timing/data': 0.03503586072474718, 'timing/model': 1.0827577384188771, 'learning_rate/qwen_vl_interface': 4.056158985279272e-06, 'learning_rate/action_model': 4.0561589852792715e-05, 'epoch': 1.58}) +[2026-06-18 01:57:37,758][starVLA.training.train_starvla][INFO] - Step 4109, Loss: {'train/loss': 0.22130480408668518, 'train/grad_norm_pre_clip': 8.473164558410645, 'timing/data': 0.03178812377154827, 'timing/model': 1.118147935718298, 'learning_rate/qwen_vl_interface': 4.054065840495425e-06, 'learning_rate/action_model': 4.054065840495425e-05, 'epoch': 1.58}) +[2026-06-18 01:57:42,740][starVLA.training.train_starvla][INFO] - Step 4110, Loss: {'train/loss': 0.26024943590164185, 'train/grad_norm_pre_clip': 8.684839248657227, 'timing/data': 0.03897680342197418, 'timing/model': 1.150789761915803, 'learning_rate/qwen_vl_interface': 4.0519729436300165e-06, 'learning_rate/action_model': 4.051972943630016e-05, 'epoch': 1.58}) +[2026-06-18 01:57:47,811][starVLA.training.train_starvla][INFO] - Step 4111, Loss: {'train/loss': 0.3528052866458893, 'train/grad_norm_pre_clip': 12.415855407714844, 'timing/data': 0.04224187042564154, 'timing/model': 1.213190314359963, 'learning_rate/qwen_vl_interface': 4.049880295116906e-06, 'learning_rate/action_model': 4.049880295116906e-05, 'epoch': 1.58}) +[2026-06-18 01:57:52,710][starVLA.training.train_starvla][INFO] - Step 4112, Loss: {'train/loss': 0.3489343822002411, 'train/grad_norm_pre_clip': 12.840553283691406, 'timing/data': 0.03352691791951656, 'timing/model': 1.5458846390247345, 'learning_rate/qwen_vl_interface': 4.047787895389902e-06, 'learning_rate/action_model': 4.0477878953899016e-05, 'epoch': 1.58}) +[2026-06-18 01:57:57,634][starVLA.training.train_starvla][INFO] - Step 4113, Loss: {'train/loss': 0.41470402479171753, 'train/grad_norm_pre_clip': 11.16596508026123, 'timing/data': 0.033625937066972256, 'timing/model': 1.2017537048086524, 'learning_rate/qwen_vl_interface': 4.0456957448827615e-06, 'learning_rate/action_model': 4.045695744882761e-05, 'epoch': 1.58}) +[2026-06-18 01:58:12,346][starVLA.training.train_starvla][INFO] - Step 4114, Loss: {'train/loss': 0.5441370010375977, 'train/grad_norm_pre_clip': 18.128196716308594, 'timing/data': 0.0346421729773283, 'timing/model': 1.5922608729451895, 'learning_rate/qwen_vl_interface': 4.043603844029188e-06, 'learning_rate/action_model': 4.0436038440291874e-05, 'epoch': 1.58}) +[2026-06-18 01:58:18,975][starVLA.training.train_starvla][INFO] - Step 4115, Loss: {'train/loss': 0.48436814546585083, 'train/grad_norm_pre_clip': 14.34402847290039, 'timing/data': 0.04104882292449474, 'timing/model': 1.2193918703123927, 'learning_rate/qwen_vl_interface': 4.041512193262836e-06, 'learning_rate/action_model': 4.041512193262836e-05, 'epoch': 1.58}) +[2026-06-18 01:58:24,087][starVLA.training.train_starvla][INFO] - Step 4116, Loss: {'train/loss': 0.4437183737754822, 'train/grad_norm_pre_clip': 12.881989479064941, 'timing/data': 0.04283858835697174, 'timing/model': 1.230177086777985, 'learning_rate/qwen_vl_interface': 4.039420793017306e-06, 'learning_rate/action_model': 4.039420793017306e-05, 'epoch': 1.58}) +[2026-06-18 01:58:38,628][starVLA.training.train_starvla][INFO] - Step 4117, Loss: {'train/loss': 0.4051583409309387, 'train/grad_norm_pre_clip': 11.957533836364746, 'timing/data': 0.07067536935210228, 'timing/model': 2.9061154713854194, 'learning_rate/qwen_vl_interface': 4.03732964372615e-06, 'learning_rate/action_model': 4.037329643726149e-05, 'epoch': 1.58}) +[2026-06-18 01:58:44,368][starVLA.training.train_starvla][INFO] - Step 4118, Loss: {'train/loss': 0.49026402831077576, 'train/grad_norm_pre_clip': 14.601030349731445, 'timing/data': 0.042450047098100185, 'timing/model': 1.216600175946951, 'learning_rate/qwen_vl_interface': 4.035238745822861e-06, 'learning_rate/action_model': 4.035238745822861e-05, 'epoch': 1.58}) +[2026-06-18 01:58:49,282][starVLA.training.train_starvla][INFO] - Step 4119, Loss: {'train/loss': 0.3590551018714905, 'train/grad_norm_pre_clip': 16.34966468811035, 'timing/data': 0.04603543132543564, 'timing/model': 1.1306397235020995, 'learning_rate/qwen_vl_interface': 4.033148099740889e-06, 'learning_rate/action_model': 4.0331480997408885e-05, 'epoch': 1.58}) +[2026-06-18 01:58:59,836][starVLA.training.train_starvla][INFO] - Step 4120, Loss: {'train/loss': 0.5110985636711121, 'train/grad_norm_pre_clip': 11.406229019165039, 'timing/data': 0.03247751761227846, 'timing/model': 1.3814367586746812, 'learning_rate/qwen_vl_interface': 4.031057705913626e-06, 'learning_rate/action_model': 4.031057705913626e-05, 'epoch': 1.58}) +[2026-06-18 01:59:07,476][starVLA.training.train_starvla][INFO] - Step 4121, Loss: {'train/loss': 0.504300594329834, 'train/grad_norm_pre_clip': 14.000548362731934, 'timing/data': 0.035321260802447796, 'timing/model': 1.4667253885418177, 'learning_rate/qwen_vl_interface': 4.028967564774411e-06, 'learning_rate/action_model': 4.0289675647744106e-05, 'epoch': 1.59}) +[2026-06-18 01:59:13,381][starVLA.training.train_starvla][INFO] - Step 4122, Loss: {'train/loss': 0.22125259041786194, 'train/grad_norm_pre_clip': 8.549540519714355, 'timing/data': 0.03333858586847782, 'timing/model': 1.6896779090166092, 'learning_rate/qwen_vl_interface': 4.026877676756534e-06, 'learning_rate/action_model': 4.026877676756534e-05, 'epoch': 1.59}) +[2026-06-18 01:59:18,182][starVLA.training.train_starvla][INFO] - Step 4123, Loss: {'train/loss': 0.4218835234642029, 'train/grad_norm_pre_clip': 16.8398380279541, 'timing/data': 0.03265842515975237, 'timing/model': 1.1855537239462137, 'learning_rate/qwen_vl_interface': 4.0247880422932305e-06, 'learning_rate/action_model': 4.02478804229323e-05, 'epoch': 1.59}) +[2026-06-18 01:59:23,747][starVLA.training.train_starvla][INFO] - Step 4124, Loss: {'train/loss': 0.4449523687362671, 'train/grad_norm_pre_clip': 13.569865226745605, 'timing/data': 0.03627731837332249, 'timing/model': 1.4725093422457576, 'learning_rate/qwen_vl_interface': 4.0226986618176855e-06, 'learning_rate/action_model': 4.022698661817686e-05, 'epoch': 1.59}) +[2026-06-18 01:59:28,948][starVLA.training.train_starvla][INFO] - Step 4125, Loss: {'train/loss': 0.49074384570121765, 'train/grad_norm_pre_clip': 12.953912734985352, 'timing/data': 0.03832250740379095, 'timing/model': 1.0906839538365602, 'learning_rate/qwen_vl_interface': 4.0206095357630256e-06, 'learning_rate/action_model': 4.020609535763026e-05, 'epoch': 1.59}) +[2026-06-18 01:59:35,294][starVLA.training.train_starvla][INFO] - Step 4126, Loss: {'train/loss': 0.18082359433174133, 'train/grad_norm_pre_clip': 8.662242889404297, 'timing/data': 0.0395363150164485, 'timing/model': 1.411195769906044, 'learning_rate/qwen_vl_interface': 4.018520664562334e-06, 'learning_rate/action_model': 4.018520664562334e-05, 'epoch': 1.59}) +[2026-06-18 01:59:40,816][starVLA.training.train_starvla][INFO] - Step 4127, Loss: {'train/loss': 0.3716711401939392, 'train/grad_norm_pre_clip': 11.818535804748535, 'timing/data': 0.03919195756316185, 'timing/model': 1.6236402373760939, 'learning_rate/qwen_vl_interface': 4.016432048648635e-06, 'learning_rate/action_model': 4.016432048648635e-05, 'epoch': 1.59}) +[2026-06-18 01:59:48,506][starVLA.training.train_starvla][INFO] - Step 4128, Loss: {'train/loss': 0.5791787505149841, 'train/grad_norm_pre_clip': 15.011369705200195, 'timing/data': 0.037853071466088295, 'timing/model': 2.1576290344819427, 'learning_rate/qwen_vl_interface': 4.014343688454898e-06, 'learning_rate/action_model': 4.0143436884548976e-05, 'epoch': 1.59}) +[2026-06-18 01:59:54,024][starVLA.training.train_starvla][INFO] - Step 4129, Loss: {'train/loss': 0.7078818082809448, 'train/grad_norm_pre_clip': 14.750313758850098, 'timing/data': 0.04559345357120037, 'timing/model': 1.4565418604761362, 'learning_rate/qwen_vl_interface': 4.012255584414044e-06, 'learning_rate/action_model': 4.0122555844140436e-05, 'epoch': 1.59}) +[2026-06-18 01:59:59,339][starVLA.training.train_starvla][INFO] - Step 4130, Loss: {'train/loss': 0.46250835061073303, 'train/grad_norm_pre_clip': 14.227081298828125, 'timing/data': 0.041357120499014854, 'timing/model': 1.3905810471624136, 'learning_rate/qwen_vl_interface': 4.010167736958939e-06, 'learning_rate/action_model': 4.010167736958938e-05, 'epoch': 1.59}) +[2026-06-18 02:00:05,157][starVLA.training.train_starvla][INFO] - Step 4131, Loss: {'train/loss': 0.3119374215602875, 'train/grad_norm_pre_clip': 11.868792533874512, 'timing/data': 0.03469514288008213, 'timing/model': 1.9832301549613476, 'learning_rate/qwen_vl_interface': 4.008080146522397e-06, 'learning_rate/action_model': 4.008080146522396e-05, 'epoch': 1.59}) +[2026-06-18 02:00:14,786][starVLA.training.train_starvla][INFO] - Step 4132, Loss: {'train/loss': 0.5999171137809753, 'train/grad_norm_pre_clip': 17.38227653503418, 'timing/data': 0.051277254708111286, 'timing/model': 1.5953385718166828, 'learning_rate/qwen_vl_interface': 4.005992813537176e-06, 'learning_rate/action_model': 4.005992813537176e-05, 'epoch': 1.59}) +[2026-06-18 02:00:20,212][starVLA.training.train_starvla][INFO] - Step 4133, Loss: {'train/loss': 0.21766918897628784, 'train/grad_norm_pre_clip': 9.083539962768555, 'timing/data': 0.051475657150149345, 'timing/model': 1.8552356101572514, 'learning_rate/qwen_vl_interface': 4.003905738435984e-06, 'learning_rate/action_model': 4.003905738435984e-05, 'epoch': 1.59}) +[2026-06-18 02:00:30,618][starVLA.training.train_starvla][INFO] - Step 4134, Loss: {'train/loss': 0.14599116146564484, 'train/grad_norm_pre_clip': 6.216502666473389, 'timing/data': 0.054072228260338306, 'timing/model': 1.1588075226172805, 'learning_rate/qwen_vl_interface': 4.001818921651475e-06, 'learning_rate/action_model': 4.0018189216514746e-05, 'epoch': 1.59}) +[2026-06-18 02:00:35,723][starVLA.training.train_starvla][INFO] - Step 4135, Loss: {'train/loss': 0.5642743110656738, 'train/grad_norm_pre_clip': 14.985373497009277, 'timing/data': 0.030902058817446232, 'timing/model': 1.2367589883506298, 'learning_rate/qwen_vl_interface': 3.999732363616244e-06, 'learning_rate/action_model': 3.999732363616244e-05, 'epoch': 1.59}) +[2026-06-18 02:00:42,384][starVLA.training.train_starvla][INFO] - Step 4136, Loss: {'train/loss': 0.28761643171310425, 'train/grad_norm_pre_clip': 9.270035743713379, 'timing/data': 0.0360622787848115, 'timing/model': 1.6825710106641054, 'learning_rate/qwen_vl_interface': 3.9976460647628394e-06, 'learning_rate/action_model': 3.997646064762839e-05, 'epoch': 1.59}) +[2026-06-18 02:00:47,538][starVLA.training.train_starvla][INFO] - Step 4137, Loss: {'train/loss': 0.3905837833881378, 'train/grad_norm_pre_clip': 12.117515563964844, 'timing/data': 0.03024430200457573, 'timing/model': 1.6386188603937626, 'learning_rate/qwen_vl_interface': 3.995560025523755e-06, 'learning_rate/action_model': 3.995560025523754e-05, 'epoch': 1.59}) +[2026-06-18 02:00:52,249][starVLA.training.train_starvla][INFO] - Step 4138, Loss: {'train/loss': 0.4142841398715973, 'train/grad_norm_pre_clip': 12.954632759094238, 'timing/data': 0.04504486173391342, 'timing/model': 1.1134177641943097, 'learning_rate/qwen_vl_interface': 3.993474246331426e-06, 'learning_rate/action_model': 3.993474246331426e-05, 'epoch': 1.59}) +[2026-06-18 02:00:57,381][starVLA.training.train_starvla][INFO] - Step 4139, Loss: {'train/loss': 0.3416725695133209, 'train/grad_norm_pre_clip': 13.498438835144043, 'timing/data': 0.03856248874217272, 'timing/model': 1.222702268511057, 'learning_rate/qwen_vl_interface': 3.991388727618237e-06, 'learning_rate/action_model': 3.991388727618237e-05, 'epoch': 1.59}) +[2026-06-18 02:01:04,155][starVLA.training.train_starvla][INFO] - Step 4140, Loss: {'train/loss': 0.36483442783355713, 'train/grad_norm_pre_clip': 8.016619682312012, 'timing/data': 0.05429185088723898, 'timing/model': 1.1284821201115847, 'learning_rate/qwen_vl_interface': 3.989303469816521e-06, 'learning_rate/action_model': 3.9893034698165205e-05, 'epoch': 1.59}) +[2026-06-18 02:01:11,996][starVLA.training.train_starvla][INFO] - Step 4141, Loss: {'train/loss': 0.40192461013793945, 'train/grad_norm_pre_clip': 21.26007080078125, 'timing/data': 0.034517847932875156, 'timing/model': 3.1924154544249177, 'learning_rate/qwen_vl_interface': 3.9872184733585526e-06, 'learning_rate/action_model': 3.987218473358552e-05, 'epoch': 1.59}) +[2026-06-18 02:01:24,887][starVLA.training.train_starvla][INFO] - Step 4142, Loss: {'train/loss': 0.369777113199234, 'train/grad_norm_pre_clip': 9.722942352294922, 'timing/data': 0.03794137388467789, 'timing/model': 4.952175552025437, 'learning_rate/qwen_vl_interface': 3.985133738676552e-06, 'learning_rate/action_model': 3.985133738676552e-05, 'epoch': 1.59}) +[2026-06-18 02:01:31,592][starVLA.training.train_starvla][INFO] - Step 4143, Loss: {'train/loss': 0.4796462655067444, 'train/grad_norm_pre_clip': 13.431205749511719, 'timing/data': 0.012351957149803638, 'timing/model': 1.7205735854804516, 'learning_rate/qwen_vl_interface': 3.983049266202688e-06, 'learning_rate/action_model': 3.983049266202687e-05, 'epoch': 1.59}) +[2026-06-18 02:01:41,211][starVLA.training.train_starvla][INFO] - Step 4144, Loss: {'train/loss': 0.2785848379135132, 'train/grad_norm_pre_clip': 10.901925086975098, 'timing/data': 0.012407630681991577, 'timing/model': 1.2410772740840912, 'learning_rate/qwen_vl_interface': 3.980965056369075e-06, 'learning_rate/action_model': 3.980965056369075e-05, 'epoch': 1.59}) +[2026-06-18 02:01:46,255][starVLA.training.train_starvla][INFO] - Step 4145, Loss: {'train/loss': 0.28502577543258667, 'train/grad_norm_pre_clip': 9.14479923248291, 'timing/data': 0.01285108644515276, 'timing/model': 1.108909010887146, 'learning_rate/qwen_vl_interface': 3.978881109607772e-06, 'learning_rate/action_model': 3.978881109607772e-05, 'epoch': 1.59}) +[2026-06-18 02:01:52,244][starVLA.training.train_starvla][INFO] - Step 4146, Loss: {'train/loss': 0.18587784469127655, 'train/grad_norm_pre_clip': 9.165931701660156, 'timing/data': 0.012774546630680561, 'timing/model': 1.6495063863694668, 'learning_rate/qwen_vl_interface': 3.976797426350783e-06, 'learning_rate/action_model': 3.976797426350783e-05, 'epoch': 1.59}) +[2026-06-18 02:01:57,073][starVLA.training.train_starvla][INFO] - Step 4147, Loss: {'train/loss': 0.32233890891075134, 'train/grad_norm_pre_clip': 9.434245109558105, 'timing/data': 0.012262900359928608, 'timing/model': 1.4369612159207463, 'learning_rate/qwen_vl_interface': 3.974714007030055e-06, 'learning_rate/action_model': 3.974714007030056e-05, 'epoch': 1.6}) +[2026-06-18 02:02:01,838][starVLA.training.train_starvla][INFO] - Step 4148, Loss: {'train/loss': 0.29902929067611694, 'train/grad_norm_pre_clip': 10.805991172790527, 'timing/data': 0.012507067993283272, 'timing/model': 1.34846016112715, 'learning_rate/qwen_vl_interface': 3.972630852077488e-06, 'learning_rate/action_model': 3.972630852077488e-05, 'epoch': 1.6}) +[2026-06-18 02:02:06,532][starVLA.training.train_starvla][INFO] - Step 4149, Loss: {'train/loss': 0.33433473110198975, 'train/grad_norm_pre_clip': 11.80488109588623, 'timing/data': 0.014794883318245411, 'timing/model': 1.118568267673254, 'learning_rate/qwen_vl_interface': 3.9705479619249195e-06, 'learning_rate/action_model': 3.9705479619249197e-05, 'epoch': 1.6}) +[2026-06-18 02:02:12,626][starVLA.training.train_starvla][INFO] - Step 4150, Loss: {'train/loss': 0.30475229024887085, 'train/grad_norm_pre_clip': 11.771666526794434, 'timing/data': 0.013718752190470695, 'timing/model': 2.2503489032387733, 'learning_rate/qwen_vl_interface': 3.9684653370041345e-06, 'learning_rate/action_model': 3.968465337004134e-05, 'epoch': 1.6}) +[2026-06-18 02:02:17,556][starVLA.training.train_starvla][INFO] - Step 4151, Loss: {'train/loss': 0.39761480689048767, 'train/grad_norm_pre_clip': 11.73907470703125, 'timing/data': 0.011589556001126766, 'timing/model': 1.1208996018394828, 'learning_rate/qwen_vl_interface': 3.966382977746866e-06, 'learning_rate/action_model': 3.9663829777468656e-05, 'epoch': 1.6}) +[2026-06-18 02:02:22,310][starVLA.training.train_starvla][INFO] - Step 4152, Loss: {'train/loss': 0.5277528166770935, 'train/grad_norm_pre_clip': 13.13870620727539, 'timing/data': 0.018770206719636917, 'timing/model': 1.4589708056300879, 'learning_rate/qwen_vl_interface': 3.9643008845847845e-06, 'learning_rate/action_model': 3.964300884584785e-05, 'epoch': 1.6}) +[2026-06-18 02:02:26,829][starVLA.training.train_starvla][INFO] - Step 4153, Loss: {'train/loss': 0.19868294894695282, 'train/grad_norm_pre_clip': 6.9415202140808105, 'timing/data': 0.011640796437859535, 'timing/model': 1.1297735404223204, 'learning_rate/qwen_vl_interface': 3.962219057949517e-06, 'learning_rate/action_model': 3.9622190579495166e-05, 'epoch': 1.6}) +[2026-06-18 02:02:34,931][starVLA.training.train_starvla][INFO] - Step 4154, Loss: {'train/loss': 0.2964014708995819, 'train/grad_norm_pre_clip': 13.473370552062988, 'timing/data': 0.012264625169336796, 'timing/model': 4.442903166636825, 'learning_rate/qwen_vl_interface': 3.960137498272619e-06, 'learning_rate/action_model': 3.9601374982726194e-05, 'epoch': 1.6}) +[2026-06-18 02:02:39,882][starVLA.training.train_starvla][INFO] - Step 4155, Loss: {'train/loss': 0.21354766190052032, 'train/grad_norm_pre_clip': 7.233602523803711, 'timing/data': 0.012579443864524364, 'timing/model': 1.277825577184558, 'learning_rate/qwen_vl_interface': 3.958056205985611e-06, 'learning_rate/action_model': 3.95805620598561e-05, 'epoch': 1.6}) +[2026-06-18 02:02:46,722][starVLA.training.train_starvla][INFO] - Step 4156, Loss: {'train/loss': 0.4948597550392151, 'train/grad_norm_pre_clip': 16.11513900756836, 'timing/data': 0.0170700429007411, 'timing/model': 1.6964949974790215, 'learning_rate/qwen_vl_interface': 3.95597518151994e-06, 'learning_rate/action_model': 3.955975181519939e-05, 'epoch': 1.6}) +[2026-06-18 02:02:52,113][starVLA.training.train_starvla][INFO] - Step 4157, Loss: {'train/loss': 0.4178106188774109, 'train/grad_norm_pre_clip': 13.619244575500488, 'timing/data': 0.012663288973271847, 'timing/model': 1.7133752880617976, 'learning_rate/qwen_vl_interface': 3.953894425307005e-06, 'learning_rate/action_model': 3.9538944253070055e-05, 'epoch': 1.6}) +[2026-06-18 02:03:00,418][starVLA.training.train_starvla][INFO] - Step 4158, Loss: {'train/loss': 0.3133851885795593, 'train/grad_norm_pre_clip': 13.459668159484863, 'timing/data': 0.013382923789322376, 'timing/model': 1.0922597432509065, 'learning_rate/qwen_vl_interface': 3.951813937778151e-06, 'learning_rate/action_model': 3.95181393777815e-05, 'epoch': 1.6}) +[2026-06-18 02:03:05,527][starVLA.training.train_starvla][INFO] - Step 4159, Loss: {'train/loss': 0.6712499856948853, 'train/grad_norm_pre_clip': 15.22216510772705, 'timing/data': 0.011355230584740639, 'timing/model': 1.4247497553005815, 'learning_rate/qwen_vl_interface': 3.949733719364664e-06, 'learning_rate/action_model': 3.949733719364664e-05, 'epoch': 1.6}) +[2026-06-18 02:03:13,985][starVLA.training.train_starvla][INFO] - Step 4160, Loss: {'train/loss': 0.28165751695632935, 'train/grad_norm_pre_clip': 11.125244140625, 'timing/data': 0.01583449076861143, 'timing/model': 1.1311433492228389, 'learning_rate/qwen_vl_interface': 3.947653770497778e-06, 'learning_rate/action_model': 3.947653770497778e-05, 'epoch': 1.6}) +[2026-06-18 02:03:20,874][starVLA.training.train_starvla][INFO] - Step 4161, Loss: {'train/loss': 0.26566341519355774, 'train/grad_norm_pre_clip': 9.134760856628418, 'timing/data': 0.012949046678841114, 'timing/model': 2.7020866479724646, 'learning_rate/qwen_vl_interface': 3.945574091608664e-06, 'learning_rate/action_model': 3.9455740916086636e-05, 'epoch': 1.6}) +[2026-06-18 02:03:35,848][starVLA.training.train_starvla][INFO] - Step 4162, Loss: {'train/loss': 0.37727785110473633, 'train/grad_norm_pre_clip': 10.820963859558105, 'timing/data': 0.016630157828330994, 'timing/model': 4.912280838005245, 'learning_rate/qwen_vl_interface': 3.9434946831284465e-06, 'learning_rate/action_model': 3.943494683128447e-05, 'epoch': 1.6}) +[2026-06-18 02:03:40,721][starVLA.training.train_starvla][INFO] - Step 4163, Loss: {'train/loss': 0.23568785190582275, 'train/grad_norm_pre_clip': 10.486353874206543, 'timing/data': 0.016043666750192642, 'timing/model': 1.0902498411014676, 'learning_rate/qwen_vl_interface': 3.941415545488187e-06, 'learning_rate/action_model': 3.941415545488187e-05, 'epoch': 1.6}) +[2026-06-18 02:03:47,965][starVLA.training.train_starvla][INFO] - Step 4164, Loss: {'train/loss': 0.3933190107345581, 'train/grad_norm_pre_clip': 11.818270683288574, 'timing/data': 0.014791341498494148, 'timing/model': 2.397830073721707, 'learning_rate/qwen_vl_interface': 3.939336679118892e-06, 'learning_rate/action_model': 3.939336679118892e-05, 'epoch': 1.6}) +[2026-06-18 02:04:02,046][starVLA.training.train_starvla][INFO] - Step 4165, Loss: {'train/loss': 0.5286411046981812, 'train/grad_norm_pre_clip': 10.863563537597656, 'timing/data': 0.012869207188487053, 'timing/model': 5.6953249564394355, 'learning_rate/qwen_vl_interface': 3.937258084451514e-06, 'learning_rate/action_model': 3.937258084451514e-05, 'epoch': 1.6}) +[2026-06-18 02:04:15,612][starVLA.training.train_starvla][INFO] - Step 4166, Loss: {'train/loss': 0.16342417895793915, 'train/grad_norm_pre_clip': 7.22027063369751, 'timing/data': 0.012715584598481655, 'timing/model': 4.711550832726061, 'learning_rate/qwen_vl_interface': 3.9351797619169485e-06, 'learning_rate/action_model': 3.935179761916949e-05, 'epoch': 1.6}) +[2026-06-18 02:04:21,350][starVLA.training.train_starvla][INFO] - Step 4167, Loss: {'train/loss': 0.3937617838382721, 'train/grad_norm_pre_clip': 11.66304874420166, 'timing/data': 0.01598251797258854, 'timing/model': 1.785751361399889, 'learning_rate/qwen_vl_interface': 3.933101711946034e-06, 'learning_rate/action_model': 3.9331017119460336e-05, 'epoch': 1.6}) +[2026-06-18 02:04:29,710][starVLA.training.train_starvla][INFO] - Step 4168, Loss: {'train/loss': 0.27579745650291443, 'train/grad_norm_pre_clip': 11.974019050598145, 'timing/data': 0.024239608086645603, 'timing/model': 1.3739062380045652, 'learning_rate/qwen_vl_interface': 3.93102393496955e-06, 'learning_rate/action_model': 3.93102393496955e-05, 'epoch': 1.6}) +[2026-06-18 02:04:36,140][starVLA.training.train_starvla][INFO] - Step 4169, Loss: {'train/loss': 0.25641319155693054, 'train/grad_norm_pre_clip': 8.162254333496094, 'timing/data': 0.011463213711977005, 'timing/model': 1.1377559499815106, 'learning_rate/qwen_vl_interface': 3.928946431418223e-06, 'learning_rate/action_model': 3.9289464314182225e-05, 'epoch': 1.6}) +[2026-06-18 02:04:41,444][starVLA.training.train_starvla][INFO] - Step 4170, Loss: {'train/loss': 0.31989243626594543, 'train/grad_norm_pre_clip': 10.977397918701172, 'timing/data': 0.012945395894348621, 'timing/model': 1.243126129731536, 'learning_rate/qwen_vl_interface': 3.926869201722723e-06, 'learning_rate/action_model': 3.926869201722723e-05, 'epoch': 1.6}) +[2026-06-18 02:04:46,800][starVLA.training.train_starvla][INFO] - Step 4171, Loss: {'train/loss': 0.2439953088760376, 'train/grad_norm_pre_clip': 8.023408889770508, 'timing/data': 0.03923700004816055, 'timing/model': 1.2962631108239293, 'learning_rate/qwen_vl_interface': 3.9247922463136615e-06, 'learning_rate/action_model': 3.924792246313662e-05, 'epoch': 1.6}) +[2026-06-18 02:04:53,247][starVLA.training.train_starvla][INFO] - Step 4172, Loss: {'train/loss': 0.36196932196617126, 'train/grad_norm_pre_clip': 11.667885780334473, 'timing/data': 0.015234023332595825, 'timing/model': 1.5456727845594287, 'learning_rate/qwen_vl_interface': 3.922715565621593e-06, 'learning_rate/action_model': 3.9227155656215926e-05, 'epoch': 1.6}) +[2026-06-18 02:04:58,557][starVLA.training.train_starvla][INFO] - Step 4173, Loss: {'train/loss': 0.487948477268219, 'train/grad_norm_pre_clip': 16.343364715576172, 'timing/data': 0.014134894125163555, 'timing/model': 1.1053354628384113, 'learning_rate/qwen_vl_interface': 3.9206391600770164e-06, 'learning_rate/action_model': 3.920639160077016e-05, 'epoch': 1.61}) +[2026-06-18 02:05:03,635][starVLA.training.train_starvla][INFO] - Step 4174, Loss: {'train/loss': 0.20055334270000458, 'train/grad_norm_pre_clip': 8.353546142578125, 'timing/data': 0.012561027891933918, 'timing/model': 1.0801557628437877, 'learning_rate/qwen_vl_interface': 3.918563030110373e-06, 'learning_rate/action_model': 3.918563030110373e-05, 'epoch': 1.61}) +[2026-06-18 02:05:08,432][starVLA.training.train_starvla][INFO] - Step 4175, Loss: {'train/loss': 0.3462543785572052, 'train/grad_norm_pre_clip': 9.99127197265625, 'timing/data': 0.012826748192310333, 'timing/model': 1.0896738963201642, 'learning_rate/qwen_vl_interface': 3.916487176152045e-06, 'learning_rate/action_model': 3.916487176152045e-05, 'epoch': 1.61}) +[2026-06-18 02:05:14,225][starVLA.training.train_starvla][INFO] - Step 4176, Loss: {'train/loss': 0.5500535368919373, 'train/grad_norm_pre_clip': 20.33749008178711, 'timing/data': 0.016190068796277046, 'timing/model': 1.688208774663508, 'learning_rate/qwen_vl_interface': 3.914411598632359e-06, 'learning_rate/action_model': 3.914411598632359e-05, 'epoch': 1.61}) +[2026-06-18 02:05:20,045][starVLA.training.train_starvla][INFO] - Step 4177, Loss: {'train/loss': 0.40808844566345215, 'train/grad_norm_pre_clip': 11.911908149719238, 'timing/data': 0.01449562981724739, 'timing/model': 1.7657360387966037, 'learning_rate/qwen_vl_interface': 3.912336297981588e-06, 'learning_rate/action_model': 3.912336297981588e-05, 'epoch': 1.61}) +[2026-06-18 02:05:27,233][starVLA.training.train_starvla][INFO] - Step 4178, Loss: {'train/loss': 0.3740452826023102, 'train/grad_norm_pre_clip': 11.339798927307129, 'timing/data': 0.013271980918943882, 'timing/model': 1.330783473327756, 'learning_rate/qwen_vl_interface': 3.910261274629941e-06, 'learning_rate/action_model': 3.910261274629941e-05, 'epoch': 1.61}) +[2026-06-18 02:05:37,772][starVLA.training.train_starvla][INFO] - Step 4179, Loss: {'train/loss': 0.23192068934440613, 'train/grad_norm_pre_clip': 8.841280937194824, 'timing/data': 0.06299729458987713, 'timing/model': 2.8550910726189613, 'learning_rate/qwen_vl_interface': 3.908186529007576e-06, 'learning_rate/action_model': 3.908186529007575e-05, 'epoch': 1.61}) +[2026-06-18 02:05:43,358][starVLA.training.train_starvla][INFO] - Step 4180, Loss: {'train/loss': 0.427472323179245, 'train/grad_norm_pre_clip': 14.337518692016602, 'timing/data': 0.0013129469007253647, 'timing/model': 1.1450326861813664, 'learning_rate/qwen_vl_interface': 3.906112061544582e-06, 'learning_rate/action_model': 3.906112061544582e-05, 'epoch': 1.61}) +[2026-06-18 02:05:49,323][starVLA.training.train_starvla][INFO] - Step 4181, Loss: {'train/loss': 0.17855852842330933, 'train/grad_norm_pre_clip': 8.41801643371582, 'timing/data': 0.0015720156952738762, 'timing/model': 1.5242351610213518, 'learning_rate/qwen_vl_interface': 3.904037872671008e-06, 'learning_rate/action_model': 3.904037872671008e-05, 'epoch': 1.61}) +[2026-06-18 02:05:55,390][starVLA.training.train_starvla][INFO] - Step 4182, Loss: {'train/loss': 0.31076762080192566, 'train/grad_norm_pre_clip': 11.721658706665039, 'timing/data': 0.001256314106285572, 'timing/model': 2.5490812016651034, 'learning_rate/qwen_vl_interface': 3.90196396281683e-06, 'learning_rate/action_model': 3.9019639628168295e-05, 'epoch': 1.61}) +[2026-06-18 02:06:00,264][starVLA.training.train_starvla][INFO] - Step 4183, Loss: {'train/loss': 0.29736778140068054, 'train/grad_norm_pre_clip': 10.849825859069824, 'timing/data': 0.0010290071368217468, 'timing/model': 1.1558727426454425, 'learning_rate/qwen_vl_interface': 3.899890332411972e-06, 'learning_rate/action_model': 3.8998903324119724e-05, 'epoch': 1.61}) +[2026-06-18 02:06:10,256][starVLA.training.train_starvla][INFO] - Step 4184, Loss: {'train/loss': 0.34000322222709656, 'train/grad_norm_pre_clip': 13.539466857910156, 'timing/data': 0.0012416699901223183, 'timing/model': 1.3628655364736915, 'learning_rate/qwen_vl_interface': 3.897816981886302e-06, 'learning_rate/action_model': 3.897816981886302e-05, 'epoch': 1.61}) +[2026-06-18 02:06:17,913][starVLA.training.train_starvla][INFO] - Step 4185, Loss: {'train/loss': 0.45289966464042664, 'train/grad_norm_pre_clip': 13.44395637512207, 'timing/data': 0.0009460058063268661, 'timing/model': 3.3216477846726775, 'learning_rate/qwen_vl_interface': 3.895743911669627e-06, 'learning_rate/action_model': 3.895743911669627e-05, 'epoch': 1.61}) +[2026-06-18 02:06:22,550][starVLA.training.train_starvla][INFO] - Step 4186, Loss: {'train/loss': 0.4224492609500885, 'train/grad_norm_pre_clip': 12.886229515075684, 'timing/data': 0.0009959330782294273, 'timing/model': 1.4477093741297722, 'learning_rate/qwen_vl_interface': 3.893671122191698e-06, 'learning_rate/action_model': 3.8936711221916974e-05, 'epoch': 1.61}) +[2026-06-18 02:06:27,172][starVLA.training.train_starvla][INFO] - Step 4187, Loss: {'train/loss': 0.5681501626968384, 'train/grad_norm_pre_clip': 18.7372989654541, 'timing/data': 0.0016331048682332039, 'timing/model': 1.3996081352233887, 'learning_rate/qwen_vl_interface': 3.8915986138822e-06, 'learning_rate/action_model': 3.8915986138822e-05, 'epoch': 1.61}) +[2026-06-18 02:06:31,513][starVLA.training.train_starvla][INFO] - Step 4188, Loss: {'train/loss': 0.5404878854751587, 'train/grad_norm_pre_clip': 18.581695556640625, 'timing/data': 0.0014389269053936005, 'timing/model': 1.0710012391209602, 'learning_rate/qwen_vl_interface': 3.889526387170775e-06, 'learning_rate/action_model': 3.889526387170775e-05, 'epoch': 1.61}) +[2026-06-18 02:06:36,171][starVLA.training.train_starvla][INFO] - Step 4189, Loss: {'train/loss': 0.45717886090278625, 'train/grad_norm_pre_clip': 17.867826461791992, 'timing/data': 0.0008396981284022331, 'timing/model': 1.0908162780106068, 'learning_rate/qwen_vl_interface': 3.8874544424869924e-06, 'learning_rate/action_model': 3.8874544424869926e-05, 'epoch': 1.61}) +[2026-06-18 02:06:40,747][starVLA.training.train_starvla][INFO] - Step 4190, Loss: {'train/loss': 0.38503092527389526, 'train/grad_norm_pre_clip': 13.331628799438477, 'timing/data': 0.0013795839622616768, 'timing/model': 1.0805673487484455, 'learning_rate/qwen_vl_interface': 3.885382780260371e-06, 'learning_rate/action_model': 3.885382780260371e-05, 'epoch': 1.61}) +[2026-06-18 02:06:45,373][starVLA.training.train_starvla][INFO] - Step 4191, Loss: {'train/loss': 0.35900652408599854, 'train/grad_norm_pre_clip': 15.646513938903809, 'timing/data': 0.004302114248275757, 'timing/model': 1.1627527512609959, 'learning_rate/qwen_vl_interface': 3.883311400920366e-06, 'learning_rate/action_model': 3.883311400920366e-05, 'epoch': 1.61}) +[2026-06-18 02:06:50,442][starVLA.training.train_starvla][INFO] - Step 4192, Loss: {'train/loss': 0.3913274109363556, 'train/grad_norm_pre_clip': 13.068624496459961, 'timing/data': 0.0008815396577119827, 'timing/model': 1.670037703588605, 'learning_rate/qwen_vl_interface': 3.88124030489638e-06, 'learning_rate/action_model': 3.88124030489638e-05, 'epoch': 1.61}) +[2026-06-18 02:06:55,215][starVLA.training.train_starvla][INFO] - Step 4193, Loss: {'train/loss': 0.2595864236354828, 'train/grad_norm_pre_clip': 13.14527702331543, 'timing/data': 0.0017734412103891373, 'timing/model': 1.4137604450806975, 'learning_rate/qwen_vl_interface': 3.8791694926177495e-06, 'learning_rate/action_model': 3.87916949261775e-05, 'epoch': 1.61}) +[2026-06-18 02:07:05,751][starVLA.training.train_starvla][INFO] - Step 4194, Loss: {'train/loss': 0.3385397791862488, 'train/grad_norm_pre_clip': 8.412569046020508, 'timing/data': 0.001235119067132473, 'timing/model': 1.3954426608979702, 'learning_rate/qwen_vl_interface': 3.877098964513759e-06, 'learning_rate/action_model': 3.877098964513759e-05, 'epoch': 1.61}) +[2026-06-18 02:07:15,942][starVLA.training.train_starvla][INFO] - Step 4195, Loss: {'train/loss': 0.2637288272380829, 'train/grad_norm_pre_clip': 10.349889755249023, 'timing/data': 0.001198682002723217, 'timing/model': 1.248113633133471, 'learning_rate/qwen_vl_interface': 3.875028721013633e-06, 'learning_rate/action_model': 3.875028721013632e-05, 'epoch': 1.61}) +[2026-06-18 02:07:30,046][starVLA.training.train_starvla][INFO] - Step 4196, Loss: {'train/loss': 0.5616244673728943, 'train/grad_norm_pre_clip': 15.281977653503418, 'timing/data': 0.0008696531876921654, 'timing/model': 1.8866279320791364, 'learning_rate/qwen_vl_interface': 3.87295876254653e-06, 'learning_rate/action_model': 3.8729587625465304e-05, 'epoch': 1.61}) +[2026-06-18 02:07:35,159][starVLA.training.train_starvla][INFO] - Step 4197, Loss: {'train/loss': 0.8615779876708984, 'train/grad_norm_pre_clip': 15.485291481018066, 'timing/data': 0.0010449830442667007, 'timing/model': 1.3997865179553628, 'learning_rate/qwen_vl_interface': 3.870889089541558e-06, 'learning_rate/action_model': 3.870889089541558e-05, 'epoch': 1.61}) +[2026-06-18 02:07:42,242][starVLA.training.train_starvla][INFO] - Step 4198, Loss: {'train/loss': 0.3884176015853882, 'train/grad_norm_pre_clip': 12.744123458862305, 'timing/data': 0.0009544380009174347, 'timing/model': 1.2659199628978968, 'learning_rate/qwen_vl_interface': 3.868819702427762e-06, 'learning_rate/action_model': 3.8688197024277616e-05, 'epoch': 1.61}) +[2026-06-18 02:07:47,503][starVLA.training.train_starvla][INFO] - Step 4199, Loss: {'train/loss': 0.476358562707901, 'train/grad_norm_pre_clip': 20.225595474243164, 'timing/data': 0.0015269909054040909, 'timing/model': 1.5124040013179183, 'learning_rate/qwen_vl_interface': 3.866750601634128e-06, 'learning_rate/action_model': 3.8667506016341277e-05, 'epoch': 1.62}) +[2026-06-18 02:13:01,950][starVLA.training.train_starvla][INFO] - Step 4200, Loss: {'train/loss': 0.30947938561439514, 'train/grad_norm_pre_clip': 10.387177467346191, 'eval/loss': 0.46741876006126404, 'eval/latency_4/loss': 0.46741876006126404, 'eval/demon_fire_precision': 0.9614046187915217, 'eval/demon_fire_recall': 0.9870087690808704, 'eval/demon_fire_f1': 0.9740384615384615, 'eval/demon_fire_rate_model': 0.9283406754772393, 'eval/demon_fire_rate_teacher': 0.9042584434654919, 'eval/demon_left_precision': 0.82, 'eval/demon_left_recall': 0.5616438356164384, 'eval/demon_left_f1': 0.6666666666666667, 'eval/demon_left_rate_model': 0.0881057268722467, 'eval/demon_left_rate_teacher': 0.12863436123348018, 'eval/demon_right_precision': 0.8109640831758034, 'eval/demon_right_recall': 0.7743682310469314, 'eval/demon_right_f1': 0.7922437673130194, 'eval/demon_right_rate_model': 0.155359765051395, 'eval/demon_right_rate_teacher': 0.16270190895741557, 'eval/demon_move_f1': 0.7294552169898431, 'eval/demon_cc_f1': 0.8850158425212142, 'eval/latency_4/demon_fire_precision': 0.9614046187915217, 'eval/latency_4/demon_fire_recall': 0.9870087690808704, 'eval/latency_4/demon_fire_f1': 0.9740384615384615, 'eval/latency_4/demon_fire_rate_model': 0.9283406754772393, 'eval/latency_4/demon_fire_rate_teacher': 0.9042584434654919, 'eval/latency_4/demon_left_precision': 0.82, 'eval/latency_4/demon_left_recall': 0.5616438356164384, 'eval/latency_4/demon_left_f1': 0.6666666666666667, 'eval/latency_4/demon_left_rate_model': 0.0881057268722467, 'eval/latency_4/demon_left_rate_teacher': 0.12863436123348018, 'eval/latency_4/demon_right_precision': 0.8109640831758034, 'eval/latency_4/demon_right_recall': 0.7743682310469314, 'eval/latency_4/demon_right_f1': 0.7922437673130194, 'eval/latency_4/demon_right_rate_model': 0.155359765051395, 'eval/latency_4/demon_right_rate_teacher': 0.16270190895741557, 'eval/latency_4/demon_move_f1': 0.7294552169898431, 'eval/latency_4/demon_cc_f1': 0.8850158425212142, 'timing/data': 0.0011519603431224823, 'timing/model': 1.1812965832650661, 'learning_rate/qwen_vl_interface': 3.864681787589581e-06, 'learning_rate/action_model': 3.8646817875895816e-05, 'epoch': 1.62}) +[2026-06-18 02:13:01,981][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_4200_state +[2026-06-18 02:13:01,989][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 02:13:12,783][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_4200_state/model.safetensors +[2026-06-18 02:13:38,143][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_4200_state/optimizer.bin +[2026-06-18 02:13:38,144][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_4200_state/sampler.bin +[2026-06-18 02:13:38,145][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_4200_state/sampler_1.bin +[2026-06-18 02:13:38,149][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_4200_state/random_states_0.pkl +[2026-06-18 02:13:38,155][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4200 +[2026-06-18 02:13:38,155][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 02:13:38,172][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 02:13:43,651][starVLA.training.train_starvla][INFO] - Step 4201, Loss: {'train/loss': 0.2577696740627289, 'train/grad_norm_pre_clip': 8.472498893737793, 'timing/data': 0.001781628467142582, 'timing/model': 1.1911508068442345, 'learning_rate/qwen_vl_interface': 3.862613260722991e-06, 'learning_rate/action_model': 3.862613260722992e-05, 'epoch': 1.62}) +[2026-06-18 02:13:49,260][starVLA.training.train_starvla][INFO] - Step 4202, Loss: {'train/loss': 0.16537965834140778, 'train/grad_norm_pre_clip': 6.407248020172119, 'timing/data': 0.0011412017047405243, 'timing/model': 1.76441718544811, 'learning_rate/qwen_vl_interface': 3.860545021463166e-06, 'learning_rate/action_model': 3.860545021463167e-05, 'epoch': 1.62}) +[2026-06-18 02:13:54,092][starVLA.training.train_starvla][INFO] - Step 4203, Loss: {'train/loss': 0.20365609228610992, 'train/grad_norm_pre_clip': 10.520440101623535, 'timing/data': 0.0012182947248220444, 'timing/model': 1.1526386514306068, 'learning_rate/qwen_vl_interface': 3.858477070238854e-06, 'learning_rate/action_model': 3.8584770702388534e-05, 'epoch': 1.62}) +[2026-06-18 02:14:02,562][starVLA.training.train_starvla][INFO] - Step 4204, Loss: {'train/loss': 0.2240050882101059, 'train/grad_norm_pre_clip': 12.069401741027832, 'timing/data': 0.0013354010879993439, 'timing/model': 1.5673627676442266, 'learning_rate/qwen_vl_interface': 3.856409407478741e-06, 'learning_rate/action_model': 3.856409407478741e-05, 'epoch': 1.62}) +[2026-06-18 02:14:13,445][starVLA.training.train_starvla][INFO] - Step 4205, Loss: {'train/loss': 0.7538683414459229, 'train/grad_norm_pre_clip': 16.879995346069336, 'timing/data': 0.0010329727083444595, 'timing/model': 2.5634518591687083, 'learning_rate/qwen_vl_interface': 3.854342033611456e-06, 'learning_rate/action_model': 3.854342033611456e-05, 'epoch': 1.62}) +[2026-06-18 02:14:19,137][starVLA.training.train_starvla][INFO] - Step 4206, Loss: {'train/loss': 0.1534154862165451, 'train/grad_norm_pre_clip': 8.802367210388184, 'timing/data': 0.001241256482899189, 'timing/model': 1.1690107677131891, 'learning_rate/qwen_vl_interface': 3.852274949065571e-06, 'learning_rate/action_model': 3.852274949065571e-05, 'epoch': 1.62}) +[2026-06-18 02:14:24,051][starVLA.training.train_starvla][INFO] - Step 4207, Loss: {'train/loss': 0.3796752691268921, 'train/grad_norm_pre_clip': 14.178227424621582, 'timing/data': 0.0011219624429941177, 'timing/model': 1.194936653599143, 'learning_rate/qwen_vl_interface': 3.850208154269593e-06, 'learning_rate/action_model': 3.8502081542695934e-05, 'epoch': 1.62}) +[2026-06-18 02:14:31,317][starVLA.training.train_starvla][INFO] - Step 4208, Loss: {'train/loss': 0.5048781037330627, 'train/grad_norm_pre_clip': 17.164031982421875, 'timing/data': 0.0017436472699046135, 'timing/model': 1.4438772900030017, 'learning_rate/qwen_vl_interface': 3.8481416496519725e-06, 'learning_rate/action_model': 3.848141649651972e-05, 'epoch': 1.62}) +[2026-06-18 02:14:35,972][starVLA.training.train_starvla][INFO] - Step 4209, Loss: {'train/loss': 0.5964820384979248, 'train/grad_norm_pre_clip': 17.650768280029297, 'timing/data': 0.001723404973745346, 'timing/model': 1.105798102915287, 'learning_rate/qwen_vl_interface': 3.846075435641093e-06, 'learning_rate/action_model': 3.846075435641093e-05, 'epoch': 1.62}) +[2026-06-18 02:14:40,895][starVLA.training.train_starvla][INFO] - Step 4210, Loss: {'train/loss': 0.3649536967277527, 'train/grad_norm_pre_clip': 13.973910331726074, 'timing/data': 0.001050848513841629, 'timing/model': 1.223152226768434, 'learning_rate/qwen_vl_interface': 3.844009512665289e-06, 'learning_rate/action_model': 3.844009512665289e-05, 'epoch': 1.62}) +[2026-06-18 02:14:46,089][starVLA.training.train_starvla][INFO] - Step 4211, Loss: {'train/loss': 0.5527291297912598, 'train/grad_norm_pre_clip': 16.279338836669922, 'timing/data': 0.0012105423957109451, 'timing/model': 1.7107067005708814, 'learning_rate/qwen_vl_interface': 3.841943881152826e-06, 'learning_rate/action_model': 3.841943881152826e-05, 'epoch': 1.62}) +[2026-06-18 02:14:51,025][starVLA.training.train_starvla][INFO] - Step 4212, Loss: {'train/loss': 0.4369596540927887, 'train/grad_norm_pre_clip': 15.96009349822998, 'timing/data': 0.0011557424440979958, 'timing/model': 1.487216416746378, 'learning_rate/qwen_vl_interface': 3.83987854153191e-06, 'learning_rate/action_model': 3.83987854153191e-05, 'epoch': 1.62}) +[2026-06-18 02:14:55,686][starVLA.training.train_starvla][INFO] - Step 4213, Loss: {'train/loss': 0.5305461287498474, 'train/grad_norm_pre_clip': 13.871452331542969, 'timing/data': 0.0011478690430521965, 'timing/model': 1.1435426427051425, 'learning_rate/qwen_vl_interface': 3.83781349423069e-06, 'learning_rate/action_model': 3.83781349423069e-05, 'epoch': 1.62}) +[2026-06-18 02:15:04,018][starVLA.training.train_starvla][INFO] - Step 4214, Loss: {'train/loss': 0.28564444184303284, 'train/grad_norm_pre_clip': 14.394551277160645, 'timing/data': 0.0009548161178827286, 'timing/model': 4.370106622576714, 'learning_rate/qwen_vl_interface': 3.835748739677253e-06, 'learning_rate/action_model': 3.835748739677253e-05, 'epoch': 1.62}) +[2026-06-18 02:15:11,087][starVLA.training.train_starvla][INFO] - Step 4215, Loss: {'train/loss': 0.2530868649482727, 'train/grad_norm_pre_clip': 16.4333438873291, 'timing/data': 0.001186826266348362, 'timing/model': 1.0782995158806443, 'learning_rate/qwen_vl_interface': 3.8336842782996245e-06, 'learning_rate/action_model': 3.8336842782996244e-05, 'epoch': 1.62}) +[2026-06-18 02:15:16,012][starVLA.training.train_starvla][INFO] - Step 4216, Loss: {'train/loss': 0.4014663100242615, 'train/grad_norm_pre_clip': 10.096778869628906, 'timing/data': 0.000972433015704155, 'timing/model': 1.2395711056888103, 'learning_rate/qwen_vl_interface': 3.831620110525766e-06, 'learning_rate/action_model': 3.831620110525767e-05, 'epoch': 1.62}) +[2026-06-18 02:15:21,143][starVLA.training.train_starvla][INFO] - Step 4217, Loss: {'train/loss': 0.6811855435371399, 'train/grad_norm_pre_clip': 15.926618576049805, 'timing/data': 0.0014232266694307327, 'timing/model': 1.7455338407307863, 'learning_rate/qwen_vl_interface': 3.829556236783588e-06, 'learning_rate/action_model': 3.8295562367835884e-05, 'epoch': 1.62}) +[2026-06-18 02:15:26,239][starVLA.training.train_starvla][INFO] - Step 4218, Loss: {'train/loss': 0.36290058493614197, 'train/grad_norm_pre_clip': 12.77922534942627, 'timing/data': 0.0009155767038464546, 'timing/model': 1.1896282834932208, 'learning_rate/qwen_vl_interface': 3.827492657500931e-06, 'learning_rate/action_model': 3.8274926575009305e-05, 'epoch': 1.62}) +[2026-06-18 02:15:31,543][starVLA.training.train_starvla][INFO] - Step 4219, Loss: {'train/loss': 0.3760605752468109, 'train/grad_norm_pre_clip': 13.507473945617676, 'timing/data': 0.0015031937509775162, 'timing/model': 1.1643523992970586, 'learning_rate/qwen_vl_interface': 3.8254293731055754e-06, 'learning_rate/action_model': 3.825429373105576e-05, 'epoch': 1.62}) +[2026-06-18 02:15:38,072][starVLA.training.train_starvla][INFO] - Step 4220, Loss: {'train/loss': 0.477182000875473, 'train/grad_norm_pre_clip': 14.923970222473145, 'timing/data': 0.0013475893065333366, 'timing/model': 1.1443742644041777, 'learning_rate/qwen_vl_interface': 3.823366384025243e-06, 'learning_rate/action_model': 3.8233663840252426e-05, 'epoch': 1.62}) +[2026-06-18 02:15:51,441][starVLA.training.train_starvla][INFO] - Step 4221, Loss: {'train/loss': 0.1934952437877655, 'train/grad_norm_pre_clip': 7.236519813537598, 'timing/data': 0.0014367345720529556, 'timing/model': 1.4657791946083307, 'learning_rate/qwen_vl_interface': 3.821303690687595e-06, 'learning_rate/action_model': 3.821303690687595e-05, 'epoch': 1.62}) +[2026-06-18 02:15:56,564][starVLA.training.train_starvla][INFO] - Step 4222, Loss: {'train/loss': 0.4580090641975403, 'train/grad_norm_pre_clip': 15.421521186828613, 'timing/data': 0.0008902288973331451, 'timing/model': 1.5736421961337328, 'learning_rate/qwen_vl_interface': 3.81924129352023e-06, 'learning_rate/action_model': 3.8192412935202304e-05, 'epoch': 1.62}) +[2026-06-18 02:16:01,988][starVLA.training.train_starvla][INFO] - Step 4223, Loss: {'train/loss': 0.43253594636917114, 'train/grad_norm_pre_clip': 14.475961685180664, 'timing/data': 0.001190725713968277, 'timing/model': 1.4054438276216388, 'learning_rate/qwen_vl_interface': 3.817179192950682e-06, 'learning_rate/action_model': 3.8171791929506826e-05, 'epoch': 1.62}) +[2026-06-18 02:16:13,071][starVLA.training.train_starvla][INFO] - Step 4224, Loss: {'train/loss': 0.5000496506690979, 'train/grad_norm_pre_clip': 12.592643737792969, 'timing/data': 0.001054517924785614, 'timing/model': 1.1497897822409868, 'learning_rate/qwen_vl_interface': 3.815117389406432e-06, 'learning_rate/action_model': 3.8151173894064324e-05, 'epoch': 1.62}) +[2026-06-18 02:16:18,478][starVLA.training.train_starvla][INFO] - Step 4225, Loss: {'train/loss': 0.17290325462818146, 'train/grad_norm_pre_clip': 8.018619537353516, 'timing/data': 0.001149228774011135, 'timing/model': 1.391988355666399, 'learning_rate/qwen_vl_interface': 3.8130558833148906e-06, 'learning_rate/action_model': 3.813055883314891e-05, 'epoch': 1.63}) +[2026-06-18 02:16:23,969][starVLA.training.train_starvla][INFO] - Step 4226, Loss: {'train/loss': 0.5821685194969177, 'train/grad_norm_pre_clip': 23.512826919555664, 'timing/data': 0.0015816614031791687, 'timing/model': 1.4535859217867255, 'learning_rate/qwen_vl_interface': 3.81099467510341e-06, 'learning_rate/action_model': 3.81099467510341e-05, 'epoch': 1.63}) +[2026-06-18 02:16:29,322][starVLA.training.train_starvla][INFO] - Step 4227, Loss: {'train/loss': 0.4080609381198883, 'train/grad_norm_pre_clip': 15.93269157409668, 'timing/data': 0.0007943371310830116, 'timing/model': 1.5105594778433442, 'learning_rate/qwen_vl_interface': 3.8089337651992815e-06, 'learning_rate/action_model': 3.808933765199281e-05, 'epoch': 1.63}) +[2026-06-18 02:16:34,663][starVLA.training.train_starvla][INFO] - Step 4228, Loss: {'train/loss': 0.3801896274089813, 'train/grad_norm_pre_clip': 16.105798721313477, 'timing/data': 0.0011578211560845375, 'timing/model': 1.2987621994689107, 'learning_rate/qwen_vl_interface': 3.8068731540297344e-06, 'learning_rate/action_model': 3.806873154029734e-05, 'epoch': 1.63}) +[2026-06-18 02:16:41,977][starVLA.training.train_starvla][INFO] - Step 4229, Loss: {'train/loss': 0.21385975182056427, 'train/grad_norm_pre_clip': 9.868037223815918, 'timing/data': 0.001133602112531662, 'timing/model': 2.8193593993782997, 'learning_rate/qwen_vl_interface': 3.8048128420219366e-06, 'learning_rate/action_model': 3.804812842021936e-05, 'epoch': 1.63}) +[2026-06-18 02:16:47,544][starVLA.training.train_starvla][INFO] - Step 4230, Loss: {'train/loss': 0.20613501965999603, 'train/grad_norm_pre_clip': 11.0132474899292, 'timing/data': 0.0009935609996318817, 'timing/model': 1.1305464385077357, 'learning_rate/qwen_vl_interface': 3.80275282960299e-06, 'learning_rate/action_model': 3.80275282960299e-05, 'epoch': 1.63}) +[2026-06-18 02:16:55,291][starVLA.training.train_starvla][INFO] - Step 4231, Loss: {'train/loss': 0.4412177503108978, 'train/grad_norm_pre_clip': 12.116230964660645, 'timing/data': 0.001459241844713688, 'timing/model': 3.847233890555799, 'learning_rate/qwen_vl_interface': 3.8006931171999394e-06, 'learning_rate/action_model': 3.800693117199939e-05, 'epoch': 1.63}) +[2026-06-18 02:17:01,592][starVLA.training.train_starvla][INFO] - Step 4232, Loss: {'train/loss': 0.3671206533908844, 'train/grad_norm_pre_clip': 15.126731872558594, 'timing/data': 0.001072792336344719, 'timing/model': 2.2074491353705525, 'learning_rate/qwen_vl_interface': 3.798633705239766e-06, 'learning_rate/action_model': 3.7986337052397655e-05, 'epoch': 1.63}) +[2026-06-18 02:17:06,396][starVLA.training.train_starvla][INFO] - Step 4233, Loss: {'train/loss': 0.42383676767349243, 'train/grad_norm_pre_clip': 12.77392292022705, 'timing/data': 0.0012041404843330383, 'timing/model': 1.1384241553023458, 'learning_rate/qwen_vl_interface': 3.796574594149387e-06, 'learning_rate/action_model': 3.796574594149387e-05, 'epoch': 1.63}) +[2026-06-18 02:17:11,560][starVLA.training.train_starvla][INFO] - Step 4234, Loss: {'train/loss': 0.36980557441711426, 'train/grad_norm_pre_clip': 12.864282608032227, 'timing/data': 0.001400916837155819, 'timing/model': 1.1462879301980138, 'learning_rate/qwen_vl_interface': 3.794515784355657e-06, 'learning_rate/action_model': 3.7945157843556575e-05, 'epoch': 1.63}) +[2026-06-18 02:17:16,261][starVLA.training.train_starvla][INFO] - Step 4235, Loss: {'train/loss': 0.6883050203323364, 'train/grad_norm_pre_clip': 23.07672119140625, 'timing/data': 0.0008537601679563522, 'timing/model': 1.0864668553695083, 'learning_rate/qwen_vl_interface': 3.7924572762853727e-06, 'learning_rate/action_model': 3.7924572762853725e-05, 'epoch': 1.63}) +[2026-06-18 02:17:21,235][starVLA.training.train_starvla][INFO] - Step 4236, Loss: {'train/loss': 0.527097225189209, 'train/grad_norm_pre_clip': 11.676506996154785, 'timing/data': 0.0011511696502566338, 'timing/model': 1.7641257988288999, 'learning_rate/qwen_vl_interface': 3.7903990703652643e-06, 'learning_rate/action_model': 3.790399070365264e-05, 'epoch': 1.63}) +[2026-06-18 02:17:26,488][starVLA.training.train_starvla][INFO] - Step 4237, Loss: {'train/loss': 0.4452749490737915, 'train/grad_norm_pre_clip': 17.63886070251465, 'timing/data': 0.00119782704859972, 'timing/model': 1.5328248785808682, 'learning_rate/qwen_vl_interface': 3.7883411670219976e-06, 'learning_rate/action_model': 3.788341167021998e-05, 'epoch': 1.63}) +[2026-06-18 02:17:30,911][starVLA.training.train_starvla][INFO] - Step 4238, Loss: {'train/loss': 0.5390707850456238, 'train/grad_norm_pre_clip': 14.142050743103027, 'timing/data': 0.0011601196601986885, 'timing/model': 1.1717322580516338, 'learning_rate/qwen_vl_interface': 3.7862835666821795e-06, 'learning_rate/action_model': 3.78628356668218e-05, 'epoch': 1.63}) +[2026-06-18 02:17:36,116][starVLA.training.train_starvla][INFO] - Step 4239, Loss: {'train/loss': 0.40310221910476685, 'train/grad_norm_pre_clip': 10.663147926330566, 'timing/data': 0.0012228917330503464, 'timing/model': 1.0943827470764518, 'learning_rate/qwen_vl_interface': 3.7842262697723543e-06, 'learning_rate/action_model': 3.784226269772354e-05, 'epoch': 1.63}) +[2026-06-18 02:17:41,082][starVLA.training.train_starvla][INFO] - Step 4240, Loss: {'train/loss': 0.4027632474899292, 'train/grad_norm_pre_clip': 12.097208023071289, 'timing/data': 0.0010957438498735428, 'timing/model': 1.1199046047404408, 'learning_rate/qwen_vl_interface': 3.782169276719e-06, 'learning_rate/action_model': 3.782169276719e-05, 'epoch': 1.63}) +[2026-06-18 02:17:47,415][starVLA.training.train_starvla][INFO] - Step 4241, Loss: {'train/loss': 0.4675871729850769, 'train/grad_norm_pre_clip': 14.503993034362793, 'timing/data': 0.002261814661324024, 'timing/model': 2.701257781125605, 'learning_rate/qwen_vl_interface': 3.7801125879485355e-06, 'learning_rate/action_model': 3.7801125879485356e-05, 'epoch': 1.63}) +[2026-06-18 02:17:56,235][starVLA.training.train_starvla][INFO] - Step 4242, Loss: {'train/loss': 0.42737123370170593, 'train/grad_norm_pre_clip': 34.5068473815918, 'timing/data': 0.001108093187212944, 'timing/model': 1.6541054537519813, 'learning_rate/qwen_vl_interface': 3.77805620388731e-06, 'learning_rate/action_model': 3.7780562038873105e-05, 'epoch': 1.63}) +[2026-06-18 02:18:01,213][starVLA.training.train_starvla][INFO] - Step 4243, Loss: {'train/loss': 0.20656085014343262, 'train/grad_norm_pre_clip': 10.127704620361328, 'timing/data': 0.0011572428047657013, 'timing/model': 1.1917609414085746, 'learning_rate/qwen_vl_interface': 3.7760001249616204e-06, 'learning_rate/action_model': 3.77600012496162e-05, 'epoch': 1.63}) +[2026-06-18 02:18:05,952][starVLA.training.train_starvla][INFO] - Step 4244, Loss: {'train/loss': 0.3162987530231476, 'train/grad_norm_pre_clip': 11.697290420532227, 'timing/data': 0.0010422132909297943, 'timing/model': 1.101428333669901, 'learning_rate/qwen_vl_interface': 3.7739443515976898e-06, 'learning_rate/action_model': 3.7739443515976894e-05, 'epoch': 1.63}) +[2026-06-18 02:18:10,903][starVLA.training.train_starvla][INFO] - Step 4245, Loss: {'train/loss': 0.45713353157043457, 'train/grad_norm_pre_clip': 13.017882347106934, 'timing/data': 0.001155809499323368, 'timing/model': 1.1485152393579483, 'learning_rate/qwen_vl_interface': 3.7718888842216822e-06, 'learning_rate/action_model': 3.771888884221682e-05, 'epoch': 1.63}) +[2026-06-18 02:18:15,915][starVLA.training.train_starvla][INFO] - Step 4246, Loss: {'train/loss': 0.37670037150382996, 'train/grad_norm_pre_clip': 15.046147346496582, 'timing/data': 0.0013954425230622292, 'timing/model': 1.7009250959381461, 'learning_rate/qwen_vl_interface': 3.7698337232597002e-06, 'learning_rate/action_model': 3.7698337232597e-05, 'epoch': 1.63}) +[2026-06-18 02:18:31,101][starVLA.training.train_starvla][INFO] - Step 4247, Loss: {'train/loss': 0.6566891074180603, 'train/grad_norm_pre_clip': 18.30379295349121, 'timing/data': 0.001150989904999733, 'timing/model': 1.9720749855041504, 'learning_rate/qwen_vl_interface': 3.7677788691377814e-06, 'learning_rate/action_model': 3.7677788691377817e-05, 'epoch': 1.63}) +[2026-06-18 02:18:36,048][starVLA.training.train_starvla][INFO] - Step 4248, Loss: {'train/loss': 0.3587806522846222, 'train/grad_norm_pre_clip': 17.007274627685547, 'timing/data': 0.00247234757989645, 'timing/model': 1.3775654630735517, 'learning_rate/qwen_vl_interface': 3.7657243222818978e-06, 'learning_rate/action_model': 3.7657243222818976e-05, 'epoch': 1.63}) +[2026-06-18 02:18:41,047][starVLA.training.train_starvla][INFO] - Step 4249, Loss: {'train/loss': 0.11199852079153061, 'train/grad_norm_pre_clip': 6.091448783874512, 'timing/data': 0.0011203568428754807, 'timing/model': 1.2738539511337876, 'learning_rate/qwen_vl_interface': 3.7636700831179575e-06, 'learning_rate/action_model': 3.763670083117957e-05, 'epoch': 1.63}) +[2026-06-18 02:18:46,273][starVLA.training.train_starvla][INFO] - Step 4250, Loss: {'train/loss': 0.20124132931232452, 'train/grad_norm_pre_clip': 7.750075340270996, 'timing/data': 0.0013512205332517624, 'timing/model': 1.138009729795158, 'learning_rate/qwen_vl_interface': 3.7616161520718096e-06, 'learning_rate/action_model': 3.76161615207181e-05, 'epoch': 1.63}) +[2026-06-18 02:18:58,375][starVLA.training.train_starvla][INFO] - Step 4251, Loss: {'train/loss': 0.3223189413547516, 'train/grad_norm_pre_clip': 10.131072998046875, 'timing/data': 0.0022110715508461, 'timing/model': 1.382065523415804, 'learning_rate/qwen_vl_interface': 3.7595625295692352e-06, 'learning_rate/action_model': 3.759562529569235e-05, 'epoch': 1.64}) +[2026-06-18 02:19:03,278][starVLA.training.train_starvla][INFO] - Step 4252, Loss: {'train/loss': 0.27281898260116577, 'train/grad_norm_pre_clip': 11.403985977172852, 'timing/data': 0.0011529820039868355, 'timing/model': 1.5039592813700438, 'learning_rate/qwen_vl_interface': 3.7575092160359523e-06, 'learning_rate/action_model': 3.757509216035952e-05, 'epoch': 1.64}) +[2026-06-18 02:19:07,721][starVLA.training.train_starvla][INFO] - Step 4253, Loss: {'train/loss': 0.39071932435035706, 'train/grad_norm_pre_clip': 15.792092323303223, 'timing/data': 0.0010557621717453003, 'timing/model': 1.2207801369950175, 'learning_rate/qwen_vl_interface': 3.7554562118976135e-06, 'learning_rate/action_model': 3.7554562118976136e-05, 'epoch': 1.64}) +[2026-06-18 02:19:12,124][starVLA.training.train_starvla][INFO] - Step 4254, Loss: {'train/loss': 0.5230706930160522, 'train/grad_norm_pre_clip': 13.01964282989502, 'timing/data': 0.001027471385896206, 'timing/model': 1.0516065750271082, 'learning_rate/qwen_vl_interface': 3.7534035175798113e-06, 'learning_rate/action_model': 3.753403517579811e-05, 'epoch': 1.64}) +[2026-06-18 02:19:16,496][starVLA.training.train_starvla][INFO] - Step 4255, Loss: {'train/loss': 0.3914162814617157, 'train/grad_norm_pre_clip': 12.437397956848145, 'timing/data': 0.0010539563372731209, 'timing/model': 1.0528998961672187, 'learning_rate/qwen_vl_interface': 3.7513511335080706e-06, 'learning_rate/action_model': 3.75135113350807e-05, 'epoch': 1.64}) +[2026-06-18 02:19:20,875][starVLA.training.train_starvla][INFO] - Step 4256, Loss: {'train/loss': 0.288271427154541, 'train/grad_norm_pre_clip': 11.516988754272461, 'timing/data': 0.0010507432743906975, 'timing/model': 1.3621207065880299, 'learning_rate/qwen_vl_interface': 3.74929906010785e-06, 'learning_rate/action_model': 3.74929906010785e-05, 'epoch': 1.64}) +[2026-06-18 02:19:25,308][starVLA.training.train_starvla][INFO] - Step 4257, Loss: {'train/loss': 0.2906380593776703, 'train/grad_norm_pre_clip': 12.148202896118164, 'timing/data': 0.0010740263387560844, 'timing/model': 1.3374513406306505, 'learning_rate/qwen_vl_interface': 3.747247297804553e-06, 'learning_rate/action_model': 3.7472472978045526e-05, 'epoch': 1.64}) +[2026-06-18 02:19:29,407][starVLA.training.train_starvla][INFO] - Step 4258, Loss: {'train/loss': 0.561973512172699, 'train/grad_norm_pre_clip': 15.52640151977539, 'timing/data': 0.0010592611506581306, 'timing/model': 1.0458912812173367, 'learning_rate/qwen_vl_interface': 3.745195847023506e-06, 'learning_rate/action_model': 3.745195847023506e-05, 'epoch': 1.64}) +[2026-06-18 02:19:34,063][starVLA.training.train_starvla][INFO] - Step 4259, Loss: {'train/loss': 0.9631406664848328, 'train/grad_norm_pre_clip': 19.946468353271484, 'timing/data': 0.0011147074401378632, 'timing/model': 1.0464013163000345, 'learning_rate/qwen_vl_interface': 3.7431447081899795e-06, 'learning_rate/action_model': 3.7431447081899796e-05, 'epoch': 1.64}) +[2026-06-18 02:19:38,812][starVLA.training.train_starvla][INFO] - Step 4260, Loss: {'train/loss': 0.3086218237876892, 'train/grad_norm_pre_clip': 11.76169204711914, 'timing/data': 0.0010203691199421883, 'timing/model': 1.139252512715757, 'learning_rate/qwen_vl_interface': 3.7410938817291753e-06, 'learning_rate/action_model': 3.741093881729175e-05, 'epoch': 1.64}) +[2026-06-18 02:19:43,051][starVLA.training.train_starvla][INFO] - Step 4261, Loss: {'train/loss': 0.3707076609134674, 'train/grad_norm_pre_clip': 10.556963920593262, 'timing/data': 0.001011648215353489, 'timing/model': 1.0523781226947904, 'learning_rate/qwen_vl_interface': 3.739043368066234e-06, 'learning_rate/action_model': 3.739043368066234e-05, 'epoch': 1.64}) +[2026-06-18 02:19:47,453][starVLA.training.train_starvla][INFO] - Step 4262, Loss: {'train/loss': 0.4984655976295471, 'train/grad_norm_pre_clip': 12.987836837768555, 'timing/data': 0.0011332007125020027, 'timing/model': 1.3891756953671575, 'learning_rate/qwen_vl_interface': 3.7369931676262304e-06, 'learning_rate/action_model': 3.7369931676262304e-05, 'epoch': 1.64}) +[2026-06-18 02:19:51,483][starVLA.training.train_starvla][INFO] - Step 4263, Loss: {'train/loss': 0.2827589213848114, 'train/grad_norm_pre_clip': 10.564468383789062, 'timing/data': 0.0010551754385232925, 'timing/model': 1.0333544574677944, 'learning_rate/qwen_vl_interface': 3.7349432808341673e-06, 'learning_rate/action_model': 3.7349432808341674e-05, 'epoch': 1.64}) +[2026-06-18 02:19:55,743][starVLA.training.train_starvla][INFO] - Step 4264, Loss: {'train/loss': 0.35211992263793945, 'train/grad_norm_pre_clip': 13.245205879211426, 'timing/data': 0.0009579621255397797, 'timing/model': 1.0511664655059576, 'learning_rate/qwen_vl_interface': 3.7328937081149952e-06, 'learning_rate/action_model': 3.732893708114995e-05, 'epoch': 1.64}) +[2026-06-18 02:20:00,128][starVLA.training.train_starvla][INFO] - Step 4265, Loss: {'train/loss': 0.34815874695777893, 'train/grad_norm_pre_clip': 14.956954956054688, 'timing/data': 0.0010877978056669235, 'timing/model': 1.0705522475764155, 'learning_rate/qwen_vl_interface': 3.7308444498935894e-06, 'learning_rate/action_model': 3.730844449893589e-05, 'epoch': 1.64}) +[2026-06-18 02:20:07,358][starVLA.training.train_starvla][INFO] - Step 4266, Loss: {'train/loss': 0.4231252670288086, 'train/grad_norm_pre_clip': 14.319805145263672, 'timing/data': 0.0010515889152884483, 'timing/model': 1.4037518436089158, 'learning_rate/qwen_vl_interface': 3.7287955065947634e-06, 'learning_rate/action_model': 3.728795506594763e-05, 'epoch': 1.64}) +[2026-06-18 02:20:12,157][starVLA.training.train_starvla][INFO] - Step 4267, Loss: {'train/loss': 0.23999856412410736, 'train/grad_norm_pre_clip': 9.426302909851074, 'timing/data': 0.0011177780106663704, 'timing/model': 1.4233105089515448, 'learning_rate/qwen_vl_interface': 3.7267468786432655e-06, 'learning_rate/action_model': 3.7267468786432654e-05, 'epoch': 1.64}) +[2026-06-18 02:20:16,914][starVLA.training.train_starvla][INFO] - Step 4268, Loss: {'train/loss': 0.23219336569309235, 'train/grad_norm_pre_clip': 10.814322471618652, 'timing/data': 0.001271766610443592, 'timing/model': 1.1030808994546533, 'learning_rate/qwen_vl_interface': 3.7246985664637787e-06, 'learning_rate/action_model': 3.724698566463778e-05, 'epoch': 1.64}) +[2026-06-18 02:20:25,521][starVLA.training.train_starvla][INFO] - Step 4269, Loss: {'train/loss': 0.6416759490966797, 'train/grad_norm_pre_clip': 19.002212524414062, 'timing/data': 0.0012914203107357025, 'timing/model': 1.0925330510362983, 'learning_rate/qwen_vl_interface': 3.722650570480922e-06, 'learning_rate/action_model': 3.722650570480922e-05, 'epoch': 1.64}) +[2026-06-18 02:20:34,707][starVLA.training.train_starvla][INFO] - Step 4270, Loss: {'train/loss': 0.3715125024318695, 'train/grad_norm_pre_clip': 13.06070327758789, 'timing/data': 0.0011596493422985077, 'timing/model': 1.114421909675002, 'learning_rate/qwen_vl_interface': 3.7206028911192445e-06, 'learning_rate/action_model': 3.720602891119244e-05, 'epoch': 1.64}) +[2026-06-18 02:20:39,394][starVLA.training.train_starvla][INFO] - Step 4271, Loss: {'train/loss': 0.46764081716537476, 'train/grad_norm_pre_clip': 12.409052848815918, 'timing/data': 0.0011043921113014221, 'timing/model': 1.0978789804503322, 'learning_rate/qwen_vl_interface': 3.7185555288032315e-06, 'learning_rate/action_model': 3.7185555288032316e-05, 'epoch': 1.64}) +[2026-06-18 02:20:44,161][starVLA.training.train_starvla][INFO] - Step 4272, Loss: {'train/loss': 0.29421350359916687, 'train/grad_norm_pre_clip': 13.606014251708984, 'timing/data': 0.0010916497558355331, 'timing/model': 1.468958729878068, 'learning_rate/qwen_vl_interface': 3.716508483957308e-06, 'learning_rate/action_model': 3.716508483957308e-05, 'epoch': 1.64}) +[2026-06-18 02:20:50,012][starVLA.training.train_starvla][INFO] - Step 4273, Loss: {'train/loss': 0.4950568675994873, 'train/grad_norm_pre_clip': 20.867719650268555, 'timing/data': 0.0013402421027421951, 'timing/model': 1.1843428509309888, 'learning_rate/qwen_vl_interface': 3.7144617570058245e-06, 'learning_rate/action_model': 3.7144617570058245e-05, 'epoch': 1.64}) +[2026-06-18 02:20:59,509][starVLA.training.train_starvla][INFO] - Step 4274, Loss: {'train/loss': 0.6053082942962646, 'train/grad_norm_pre_clip': 15.780834197998047, 'timing/data': 0.000979800708591938, 'timing/model': 1.45637250225991, 'learning_rate/qwen_vl_interface': 3.7124153483730706e-06, 'learning_rate/action_model': 3.712415348373071e-05, 'epoch': 1.64}) +[2026-06-18 02:21:04,604][starVLA.training.train_starvla][INFO] - Step 4275, Loss: {'train/loss': 0.35332751274108887, 'train/grad_norm_pre_clip': 9.386788368225098, 'timing/data': 0.0014011021703481674, 'timing/model': 1.1870600720867515, 'learning_rate/qwen_vl_interface': 3.71036925848327e-06, 'learning_rate/action_model': 3.71036925848327e-05, 'epoch': 1.64}) +[2026-06-18 02:21:10,214][starVLA.training.train_starvla][INFO] - Step 4276, Loss: {'train/loss': 0.3896033465862274, 'train/grad_norm_pre_clip': 14.354554176330566, 'timing/data': 0.0019942782819271088, 'timing/model': 1.7719661109149456, 'learning_rate/qwen_vl_interface': 3.7083234877605804e-06, 'learning_rate/action_model': 3.7083234877605806e-05, 'epoch': 1.64}) +[2026-06-18 02:21:21,600][starVLA.training.train_starvla][INFO] - Step 4277, Loss: {'train/loss': 0.31572648882865906, 'train/grad_norm_pre_clip': 13.831525802612305, 'timing/data': 0.0012182723730802536, 'timing/model': 2.3709662472829223, 'learning_rate/qwen_vl_interface': 3.70627803662909e-06, 'learning_rate/action_model': 3.70627803662909e-05, 'epoch': 1.65}) +[2026-06-18 02:21:26,366][starVLA.training.train_starvla][INFO] - Step 4278, Loss: {'train/loss': 0.39166536927223206, 'train/grad_norm_pre_clip': 14.411967277526855, 'timing/data': 0.0011044982820749283, 'timing/model': 1.3051037900149822, 'learning_rate/qwen_vl_interface': 3.7042329055128216e-06, 'learning_rate/action_model': 3.704232905512821e-05, 'epoch': 1.65}) +[2026-06-18 02:21:33,629][starVLA.training.train_starvla][INFO] - Step 4279, Loss: {'train/loss': 0.21173222362995148, 'train/grad_norm_pre_clip': 9.069822311401367, 'timing/data': 0.0010880902409553528, 'timing/model': 1.2154972990974784, 'learning_rate/qwen_vl_interface': 3.7021880948357375e-06, 'learning_rate/action_model': 3.702188094835737e-05, 'epoch': 1.65}) +[2026-06-18 02:21:42,692][starVLA.training.train_starvla][INFO] - Step 4280, Loss: {'train/loss': 0.20721286535263062, 'train/grad_norm_pre_clip': 9.26807689666748, 'timing/data': 0.001165107823908329, 'timing/model': 1.4126858357340097, 'learning_rate/qwen_vl_interface': 3.700143605021726e-06, 'learning_rate/action_model': 3.7001436050217256e-05, 'epoch': 1.65}) +[2026-06-18 02:21:51,167][starVLA.training.train_starvla][INFO] - Step 4281, Loss: {'train/loss': 0.591689944267273, 'train/grad_norm_pre_clip': 14.719279289245605, 'timing/data': 0.0011266190558671951, 'timing/model': 1.1265898821875453, 'learning_rate/qwen_vl_interface': 3.6980994364946133e-06, 'learning_rate/action_model': 3.698099436494613e-05, 'epoch': 1.65}) +[2026-06-18 02:21:56,738][starVLA.training.train_starvla][INFO] - Step 4282, Loss: {'train/loss': 0.33709651231765747, 'train/grad_norm_pre_clip': 13.277856826782227, 'timing/data': 0.0012596920132637024, 'timing/model': 1.7570139672607183, 'learning_rate/qwen_vl_interface': 3.696055589678156e-06, 'learning_rate/action_model': 3.6960555896781554e-05, 'epoch': 1.65}) +[2026-06-18 02:22:02,892][starVLA.training.train_starvla][INFO] - Step 4283, Loss: {'train/loss': 0.44607096910476685, 'train/grad_norm_pre_clip': 14.736665725708008, 'timing/data': 0.0013327859342098236, 'timing/model': 2.674340601079166, 'learning_rate/qwen_vl_interface': 3.694012064996049e-06, 'learning_rate/action_model': 3.694012064996049e-05, 'epoch': 1.65}) +[2026-06-18 02:22:08,251][starVLA.training.train_starvla][INFO] - Step 4284, Loss: {'train/loss': 0.3169512450695038, 'train/grad_norm_pre_clip': 13.095633506774902, 'timing/data': 0.000948069617152214, 'timing/model': 1.119928983040154, 'learning_rate/qwen_vl_interface': 3.6919688628719153e-06, 'learning_rate/action_model': 3.691968862871915e-05, 'epoch': 1.65}) +[2026-06-18 02:22:13,371][starVLA.training.train_starvla][INFO] - Step 4285, Loss: {'train/loss': 0.6071622967720032, 'train/grad_norm_pre_clip': 14.731101036071777, 'timing/data': 0.001506788656115532, 'timing/model': 1.1984332893043756, 'learning_rate/qwen_vl_interface': 3.689925983729311e-06, 'learning_rate/action_model': 3.689925983729311e-05, 'epoch': 1.65}) +[2026-06-18 02:22:18,665][starVLA.training.train_starvla][INFO] - Step 4286, Loss: {'train/loss': 0.5311869382858276, 'train/grad_norm_pre_clip': 14.433664321899414, 'timing/data': 0.0014406712725758553, 'timing/model': 1.6100234761834145, 'learning_rate/qwen_vl_interface': 3.68788342799173e-06, 'learning_rate/action_model': 3.6878834279917294e-05, 'epoch': 1.65}) +[2026-06-18 02:22:23,344][starVLA.training.train_starvla][INFO] - Step 4287, Loss: {'train/loss': 0.17629586160182953, 'train/grad_norm_pre_clip': 6.544070720672607, 'timing/data': 0.0011299196630716324, 'timing/model': 1.3571551134809852, 'learning_rate/qwen_vl_interface': 3.6858411960825953e-06, 'learning_rate/action_model': 3.685841196082595e-05, 'epoch': 1.65}) +[2026-06-18 02:22:27,721][starVLA.training.train_starvla][INFO] - Step 4288, Loss: {'train/loss': 0.30777478218078613, 'train/grad_norm_pre_clip': 12.456591606140137, 'timing/data': 0.0011381767690181732, 'timing/model': 1.1059465184807777, 'learning_rate/qwen_vl_interface': 3.6837992884252636e-06, 'learning_rate/action_model': 3.683799288425264e-05, 'epoch': 1.65}) +[2026-06-18 02:22:32,726][starVLA.training.train_starvla][INFO] - Step 4289, Loss: {'train/loss': 0.4932151734828949, 'train/grad_norm_pre_clip': 14.563531875610352, 'timing/data': 0.0012188171967864037, 'timing/model': 1.2716128369793296, 'learning_rate/qwen_vl_interface': 3.6817577054430243e-06, 'learning_rate/action_model': 3.6817577054430244e-05, 'epoch': 1.65}) +[2026-06-18 02:22:37,523][starVLA.training.train_starvla][INFO] - Step 4290, Loss: {'train/loss': 0.20681555569171906, 'train/grad_norm_pre_clip': 6.460768222808838, 'timing/data': 0.00099867582321167, 'timing/model': 1.1651427587494254, 'learning_rate/qwen_vl_interface': 3.6797164475591003e-06, 'learning_rate/action_model': 3.6797164475591e-05, 'epoch': 1.65}) +[2026-06-18 02:22:43,909][starVLA.training.train_starvla][INFO] - Step 4291, Loss: {'train/loss': 0.443562775850296, 'train/grad_norm_pre_clip': 13.63593864440918, 'timing/data': 0.0015990650281310081, 'timing/model': 1.2525536641478539, 'learning_rate/qwen_vl_interface': 3.677675515196648e-06, 'learning_rate/action_model': 3.677675515196648e-05, 'epoch': 1.65}) +[2026-06-18 02:22:48,996][starVLA.training.train_starvla][INFO] - Step 4292, Loss: {'train/loss': 0.25014862418174744, 'train/grad_norm_pre_clip': 8.63284969329834, 'timing/data': 0.0011504264548420906, 'timing/model': 1.7107961978763342, 'learning_rate/qwen_vl_interface': 3.6756349087787524e-06, 'learning_rate/action_model': 3.6756349087787524e-05, 'epoch': 1.65}) +[2026-06-18 02:22:53,477][starVLA.training.train_starvla][INFO] - Step 4293, Loss: {'train/loss': 0.37729012966156006, 'train/grad_norm_pre_clip': 14.525083541870117, 'timing/data': 0.0012278761714696884, 'timing/model': 1.3156088199466467, 'learning_rate/qwen_vl_interface': 3.6735946287284325e-06, 'learning_rate/action_model': 3.6735946287284324e-05, 'epoch': 1.65}) +[2026-06-18 02:23:04,009][starVLA.training.train_starvla][INFO] - Step 4294, Loss: {'train/loss': 0.21084226667881012, 'train/grad_norm_pre_clip': 9.219011306762695, 'timing/data': 0.0009146621450781822, 'timing/model': 2.5434928545728326, 'learning_rate/qwen_vl_interface': 3.6715546754686444e-06, 'learning_rate/action_model': 3.6715546754686444e-05, 'epoch': 1.65}) +[2026-06-18 02:23:14,431][starVLA.training.train_starvla][INFO] - Step 4295, Loss: {'train/loss': 0.588532567024231, 'train/grad_norm_pre_clip': 13.437767028808594, 'timing/data': 0.0009815394878387451, 'timing/model': 1.1122675333172083, 'learning_rate/qwen_vl_interface': 3.6695150494222697e-06, 'learning_rate/action_model': 3.6695150494222695e-05, 'epoch': 1.65}) +[2026-06-18 02:23:21,911][starVLA.training.train_starvla][INFO] - Step 4296, Loss: {'train/loss': 0.4502316117286682, 'train/grad_norm_pre_clip': 19.15077781677246, 'timing/data': 0.0008058072999119759, 'timing/model': 4.085503880865872, 'learning_rate/qwen_vl_interface': 3.6674757510121255e-06, 'learning_rate/action_model': 3.6674757510121255e-05, 'epoch': 1.65}) +[2026-06-18 02:23:29,433][starVLA.training.train_starvla][INFO] - Step 4297, Loss: {'train/loss': 0.7685718536376953, 'train/grad_norm_pre_clip': 17.581466674804688, 'timing/data': 0.001109589822590351, 'timing/model': 1.4312798287719488, 'learning_rate/qwen_vl_interface': 3.6654367806609615e-06, 'learning_rate/action_model': 3.665436780660961e-05, 'epoch': 1.65}) +[2026-06-18 02:23:35,973][starVLA.training.train_starvla][INFO] - Step 4298, Loss: {'train/loss': 0.6079589128494263, 'train/grad_norm_pre_clip': 23.597999572753906, 'timing/data': 0.0012070219963788986, 'timing/model': 1.159800237044692, 'learning_rate/qwen_vl_interface': 3.6633981387914597e-06, 'learning_rate/action_model': 3.66339813879146e-05, 'epoch': 1.65}) +[2026-06-18 02:23:46,812][starVLA.training.train_starvla][INFO] - Step 4299, Loss: {'train/loss': 0.4596087336540222, 'train/grad_norm_pre_clip': 11.933107376098633, 'timing/data': 0.0014194613322615623, 'timing/model': 5.110841718502343, 'learning_rate/qwen_vl_interface': 3.6613598258262295e-06, 'learning_rate/action_model': 3.6613598258262296e-05, 'epoch': 1.65}) +[2026-06-18 02:29:14,241][starVLA.training.train_starvla][INFO] - Step 4300, Loss: {'train/loss': 0.39071014523506165, 'train/grad_norm_pre_clip': 10.849027633666992, 'eval/loss': 0.4615606665611267, 'eval/latency_4/loss': 0.4615606665611267, 'eval/demon_fire_precision': 0.9744010502133246, 'eval/demon_fire_recall': 0.9642741149723937, 'eval/demon_fire_f1': 0.9693111328762651, 'eval/demon_fire_rate_model': 0.8948604992657856, 'eval/demon_fire_rate_teacher': 0.9042584434654919, 'eval/demon_left_precision': 0.6802721088435374, 'eval/demon_left_recall': 0.684931506849315, 'eval/demon_left_f1': 0.68259385665529, 'eval/demon_left_rate_model': 0.12951541850220263, 'eval/demon_left_rate_teacher': 0.12863436123348018, 'eval/demon_right_precision': 0.7495826377295493, 'eval/demon_right_recall': 0.8104693140794224, 'eval/demon_right_f1': 0.7788378143972245, 'eval/demon_right_rate_model': 0.17591776798825257, 'eval/demon_right_rate_teacher': 0.16270190895741557, 'eval/demon_move_f1': 0.7307158355262573, 'eval/demon_cc_f1': 0.8828317541678489, 'eval/latency_4/demon_fire_precision': 0.9744010502133246, 'eval/latency_4/demon_fire_recall': 0.9642741149723937, 'eval/latency_4/demon_fire_f1': 0.9693111328762651, 'eval/latency_4/demon_fire_rate_model': 0.8948604992657856, 'eval/latency_4/demon_fire_rate_teacher': 0.9042584434654919, 'eval/latency_4/demon_left_precision': 0.6802721088435374, 'eval/latency_4/demon_left_recall': 0.684931506849315, 'eval/latency_4/demon_left_f1': 0.68259385665529, 'eval/latency_4/demon_left_rate_model': 0.12951541850220263, 'eval/latency_4/demon_left_rate_teacher': 0.12863436123348018, 'eval/latency_4/demon_right_precision': 0.7495826377295493, 'eval/latency_4/demon_right_recall': 0.8104693140794224, 'eval/latency_4/demon_right_f1': 0.7788378143972245, 'eval/latency_4/demon_right_rate_model': 0.17591776798825257, 'eval/latency_4/demon_right_rate_teacher': 0.16270190895741557, 'eval/latency_4/demon_move_f1': 0.7307158355262573, 'eval/latency_4/demon_cc_f1': 0.8828317541678489, 'timing/data': 0.0011728499084711075, 'timing/model': 5.331703091971576, 'learning_rate/qwen_vl_interface': 3.6593218421878157e-06, 'learning_rate/action_model': 3.659321842187816e-05, 'epoch': 1.65}) +[2026-06-18 02:29:14,254][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_4300_state +[2026-06-18 02:29:14,262][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 02:29:23,768][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_4300_state/model.safetensors +[2026-06-18 02:29:48,115][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_4300_state/optimizer.bin +[2026-06-18 02:29:48,116][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_4300_state/sampler.bin +[2026-06-18 02:29:48,117][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_4300_state/sampler_1.bin +[2026-06-18 02:29:48,120][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_4300_state/random_states_0.pkl +[2026-06-18 02:29:48,130][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4300 +[2026-06-18 02:29:48,131][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 02:29:48,150][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 02:29:59,303][starVLA.training.train_starvla][INFO] - Step 4301, Loss: {'train/loss': 0.2722662687301636, 'train/grad_norm_pre_clip': 9.689753532409668, 'timing/data': 0.0013377778232097626, 'timing/model': 1.9849587231874466, 'learning_rate/qwen_vl_interface': 3.657284188298697e-06, 'learning_rate/action_model': 3.657284188298697e-05, 'epoch': 1.65}) +[2026-06-18 02:30:04,354][starVLA.training.train_starvla][INFO] - Step 4302, Loss: {'train/loss': 0.4432658851146698, 'train/grad_norm_pre_clip': 20.479747772216797, 'timing/data': 0.0030923159793019295, 'timing/model': 1.3438445152714849, 'learning_rate/qwen_vl_interface': 3.65524686458128e-06, 'learning_rate/action_model': 3.65524686458128e-05, 'epoch': 1.65}) +[2026-06-18 02:30:17,384][starVLA.training.train_starvla][INFO] - Step 4303, Loss: {'train/loss': 0.39623212814331055, 'train/grad_norm_pre_clip': 13.603689193725586, 'timing/data': 0.0011567221954464912, 'timing/model': 5.824535200372338, 'learning_rate/qwen_vl_interface': 3.653209871457904e-06, 'learning_rate/action_model': 3.653209871457904e-05, 'epoch': 1.66}) +[2026-06-18 02:30:24,420][starVLA.training.train_starvla][INFO] - Step 4304, Loss: {'train/loss': 0.2584210932254791, 'train/grad_norm_pre_clip': 10.978949546813965, 'timing/data': 0.0015044845640659332, 'timing/model': 1.1737321177497506, 'learning_rate/qwen_vl_interface': 3.6511732093508376e-06, 'learning_rate/action_model': 3.651173209350838e-05, 'epoch': 1.66}) +[2026-06-18 02:30:34,877][starVLA.training.train_starvla][INFO] - Step 4305, Loss: {'train/loss': 0.5153489708900452, 'train/grad_norm_pre_clip': 18.890029907226562, 'timing/data': 0.0018515633419156075, 'timing/model': 2.7582650519907475, 'learning_rate/qwen_vl_interface': 3.6491368786822885e-06, 'learning_rate/action_model': 3.6491368786822884e-05, 'epoch': 1.66}) +[2026-06-18 02:30:43,473][starVLA.training.train_starvla][INFO] - Step 4306, Loss: {'train/loss': 0.4485923945903778, 'train/grad_norm_pre_clip': 11.986434936523438, 'timing/data': 0.0010396847501397133, 'timing/model': 3.2897176072001457, 'learning_rate/qwen_vl_interface': 3.647100879874386e-06, 'learning_rate/action_model': 3.6471008798743855e-05, 'epoch': 1.66}) +[2026-06-18 02:30:50,160][starVLA.training.train_starvla][INFO] - Step 4307, Loss: {'train/loss': 0.5382083654403687, 'train/grad_norm_pre_clip': 12.678434371948242, 'timing/data': 0.0020205965265631676, 'timing/model': 1.5370023595169187, 'learning_rate/qwen_vl_interface': 3.645065213349195e-06, 'learning_rate/action_model': 3.645065213349195e-05, 'epoch': 1.66}) +[2026-06-18 02:30:54,835][starVLA.training.train_starvla][INFO] - Step 4308, Loss: {'train/loss': 0.48120513558387756, 'train/grad_norm_pre_clip': 20.052425384521484, 'timing/data': 0.001195632852613926, 'timing/model': 1.1352947130799294, 'learning_rate/qwen_vl_interface': 3.6430298795287135e-06, 'learning_rate/action_model': 3.643029879528713e-05, 'epoch': 1.66}) +[2026-06-18 02:31:00,173][starVLA.training.train_starvla][INFO] - Step 4309, Loss: {'train/loss': 0.5133554339408875, 'train/grad_norm_pre_clip': 18.75403594970703, 'timing/data': 0.0032099056988954544, 'timing/model': 1.1577182356268167, 'learning_rate/qwen_vl_interface': 3.640994878834866e-06, 'learning_rate/action_model': 3.640994878834866e-05, 'epoch': 1.66}) +[2026-06-18 02:31:05,537][starVLA.training.train_starvla][INFO] - Step 4310, Loss: {'train/loss': 0.4363957643508911, 'train/grad_norm_pre_clip': 12.178464889526367, 'timing/data': 0.0015525994822382927, 'timing/model': 1.135567364282906, 'learning_rate/qwen_vl_interface': 3.6389602116895136e-06, 'learning_rate/action_model': 3.6389602116895134e-05, 'epoch': 1.66}) +[2026-06-18 02:31:17,503][starVLA.training.train_starvla][INFO] - Step 4311, Loss: {'train/loss': 0.3773953318595886, 'train/grad_norm_pre_clip': 8.172679901123047, 'timing/data': 0.0014756880700588226, 'timing/model': 1.278105047531426, 'learning_rate/qwen_vl_interface': 3.6369258785144395e-06, 'learning_rate/action_model': 3.636925878514439e-05, 'epoch': 1.66}) +[2026-06-18 02:31:21,907][starVLA.training.train_starvla][INFO] - Step 4312, Loss: {'train/loss': 0.3430415093898773, 'train/grad_norm_pre_clip': 13.17232608795166, 'timing/data': 0.0010784044861793518, 'timing/model': 1.1595232160761952, 'learning_rate/qwen_vl_interface': 3.6348918797313693e-06, 'learning_rate/action_model': 3.634891879731369e-05, 'epoch': 1.66}) +[2026-06-18 02:31:26,911][starVLA.training.train_starvla][INFO] - Step 4313, Loss: {'train/loss': 0.41202279925346375, 'train/grad_norm_pre_clip': 12.44581413269043, 'timing/data': 0.003097308799624443, 'timing/model': 1.1237087855115533, 'learning_rate/qwen_vl_interface': 3.6328582157619495e-06, 'learning_rate/action_model': 3.6328582157619496e-05, 'epoch': 1.66}) +[2026-06-18 02:31:31,632][starVLA.training.train_starvla][INFO] - Step 4314, Loss: {'train/loss': 0.5000443458557129, 'train/grad_norm_pre_clip': 14.877254486083984, 'timing/data': 0.0033680126070976257, 'timing/model': 1.207750516012311, 'learning_rate/qwen_vl_interface': 3.6308248870277613e-06, 'learning_rate/action_model': 3.630824887027761e-05, 'epoch': 1.66}) +[2026-06-18 02:31:36,350][starVLA.training.train_starvla][INFO] - Step 4315, Loss: {'train/loss': 0.29971179366111755, 'train/grad_norm_pre_clip': 12.71615982055664, 'timing/data': 0.0013676928356289864, 'timing/model': 1.1174074355512857, 'learning_rate/qwen_vl_interface': 3.628791893950316e-06, 'learning_rate/action_model': 3.628791893950316e-05, 'epoch': 1.66}) +[2026-06-18 02:31:44,755][starVLA.training.train_starvla][INFO] - Step 4316, Loss: {'train/loss': 0.32214629650115967, 'train/grad_norm_pre_clip': 12.049674987792969, 'timing/data': 0.0013520652428269386, 'timing/model': 2.7142695980146527, 'learning_rate/qwen_vl_interface': 3.6267592369510563e-06, 'learning_rate/action_model': 3.6267592369510564e-05, 'epoch': 1.66}) +[2026-06-18 02:31:52,607][starVLA.training.train_starvla][INFO] - Step 4317, Loss: {'train/loss': 0.2755224406719208, 'train/grad_norm_pre_clip': 11.97790241241455, 'timing/data': 0.0012947190552949905, 'timing/model': 1.829969366081059, 'learning_rate/qwen_vl_interface': 3.624726916451355e-06, 'learning_rate/action_model': 3.624726916451355e-05, 'epoch': 1.66}) +[2026-06-18 02:32:01,373][starVLA.training.train_starvla][INFO] - Step 4318, Loss: {'train/loss': 0.262286901473999, 'train/grad_norm_pre_clip': 12.731583595275879, 'timing/data': 0.0011816471815109253, 'timing/model': 1.3857720335945487, 'learning_rate/qwen_vl_interface': 3.622694932872509e-06, 'learning_rate/action_model': 3.622694932872509e-05, 'epoch': 1.66}) +[2026-06-18 02:32:06,648][starVLA.training.train_starvla][INFO] - Step 4319, Loss: {'train/loss': 0.6220624446868896, 'train/grad_norm_pre_clip': 15.994858741760254, 'timing/data': 0.001347293145954609, 'timing/model': 1.2003288008272648, 'learning_rate/qwen_vl_interface': 3.620663286635758e-06, 'learning_rate/action_model': 3.620663286635758e-05, 'epoch': 1.66}) +[2026-06-18 02:32:12,697][starVLA.training.train_starvla][INFO] - Step 4320, Loss: {'train/loss': 0.35212889313697815, 'train/grad_norm_pre_clip': 10.178168296813965, 'timing/data': 0.001179330050945282, 'timing/model': 1.5012141903862357, 'learning_rate/qwen_vl_interface': 3.618631978162259e-06, 'learning_rate/action_model': 3.618631978162259e-05, 'epoch': 1.66}) +[2026-06-18 02:32:23,284][starVLA.training.train_starvla][INFO] - Step 4321, Loss: {'train/loss': 0.3018280267715454, 'train/grad_norm_pre_clip': 10.628867149353027, 'timing/data': 0.0013376129791140556, 'timing/model': 4.775562341324985, 'learning_rate/qwen_vl_interface': 3.6166010078731073e-06, 'learning_rate/action_model': 3.6166010078731075e-05, 'epoch': 1.66}) +[2026-06-18 02:32:28,884][starVLA.training.train_starvla][INFO] - Step 4322, Loss: {'train/loss': 0.5233888626098633, 'train/grad_norm_pre_clip': 12.036015510559082, 'timing/data': 0.0010217102244496346, 'timing/model': 1.6460530860349536, 'learning_rate/qwen_vl_interface': 3.614570376189323e-06, 'learning_rate/action_model': 3.614570376189323e-05, 'epoch': 1.66}) +[2026-06-18 02:32:40,571][starVLA.training.train_starvla][INFO] - Step 4323, Loss: {'train/loss': 0.5058506727218628, 'train/grad_norm_pre_clip': 13.09294319152832, 'timing/data': 0.001038363203406334, 'timing/model': 4.360873942263424, 'learning_rate/qwen_vl_interface': 3.6125400835318604e-06, 'learning_rate/action_model': 3.61254008353186e-05, 'epoch': 1.66}) +[2026-06-18 02:32:46,212][starVLA.training.train_starvla][INFO] - Step 4324, Loss: {'train/loss': 0.20475807785987854, 'train/grad_norm_pre_clip': 12.394145965576172, 'timing/data': 0.0011373870074748993, 'timing/model': 1.2098040264099836, 'learning_rate/qwen_vl_interface': 3.6105101303216016e-06, 'learning_rate/action_model': 3.610510130321601e-05, 'epoch': 1.66}) +[2026-06-18 02:32:51,446][starVLA.training.train_starvla][INFO] - Step 4325, Loss: {'train/loss': 0.2261209785938263, 'train/grad_norm_pre_clip': 9.637200355529785, 'timing/data': 0.004590715281665325, 'timing/model': 1.1354197524487972, 'learning_rate/qwen_vl_interface': 3.608480516979354e-06, 'learning_rate/action_model': 3.608480516979354e-05, 'epoch': 1.66}) +[2026-06-18 02:33:02,535][starVLA.training.train_starvla][INFO] - Step 4326, Loss: {'train/loss': 0.36678045988082886, 'train/grad_norm_pre_clip': 11.485164642333984, 'timing/data': 0.0009509017691016197, 'timing/model': 2.7401008522138, 'learning_rate/qwen_vl_interface': 3.6064512439258646e-06, 'learning_rate/action_model': 3.606451243925865e-05, 'epoch': 1.66}) +[2026-06-18 02:33:08,004][starVLA.training.train_starvla][INFO] - Step 4327, Loss: {'train/loss': 0.1660763919353485, 'train/grad_norm_pre_clip': 7.7457051277160645, 'timing/data': 0.0010735541582107544, 'timing/model': 1.719742899760604, 'learning_rate/qwen_vl_interface': 3.6044223115818e-06, 'learning_rate/action_model': 3.6044223115818e-05, 'epoch': 1.66}) +[2026-06-18 02:33:12,828][starVLA.training.train_starvla][INFO] - Step 4328, Loss: {'train/loss': 0.3888022303581238, 'train/grad_norm_pre_clip': 15.56782054901123, 'timing/data': 0.0011246530339121819, 'timing/model': 1.2509191455319524, 'learning_rate/qwen_vl_interface': 3.6023937203677606e-06, 'learning_rate/action_model': 3.6023937203677604e-05, 'epoch': 1.66}) +[2026-06-18 02:33:22,585][starVLA.training.train_starvla][INFO] - Step 4329, Loss: {'train/loss': 0.2942727208137512, 'train/grad_norm_pre_clip': 11.035572052001953, 'timing/data': 0.0013042856007814407, 'timing/model': 1.1783919353038073, 'learning_rate/qwen_vl_interface': 3.6003654707042746e-06, 'learning_rate/action_model': 3.6003654707042745e-05, 'epoch': 1.67}) +[2026-06-18 02:33:31,662][starVLA.training.train_starvla][INFO] - Step 4330, Loss: {'train/loss': 0.3944166302680969, 'train/grad_norm_pre_clip': 11.663395881652832, 'timing/data': 0.0015598991885781288, 'timing/model': 1.3871568432077765, 'learning_rate/qwen_vl_interface': 3.5983375630118033e-06, 'learning_rate/action_model': 3.598337563011803e-05, 'epoch': 1.67}) +[2026-06-18 02:33:39,027][starVLA.training.train_starvla][INFO] - Step 4331, Loss: {'train/loss': 0.48254814743995667, 'train/grad_norm_pre_clip': 15.38447380065918, 'timing/data': 0.0014957552775740623, 'timing/model': 1.1023980658501387, 'learning_rate/qwen_vl_interface': 3.596309997710733e-06, 'learning_rate/action_model': 3.596309997710733e-05, 'epoch': 1.67}) +[2026-06-18 02:33:43,980][starVLA.training.train_starvla][INFO] - Step 4332, Loss: {'train/loss': 0.5350948572158813, 'train/grad_norm_pre_clip': 15.197212219238281, 'timing/data': 0.0020268820226192474, 'timing/model': 1.3444928722456098, 'learning_rate/qwen_vl_interface': 3.5942827752213788e-06, 'learning_rate/action_model': 3.594282775221379e-05, 'epoch': 1.67}) +[2026-06-18 02:33:50,245][starVLA.training.train_starvla][INFO] - Step 4333, Loss: {'train/loss': 0.13673639297485352, 'train/grad_norm_pre_clip': 7.432319641113281, 'timing/data': 0.0011954894289374352, 'timing/model': 1.8703933218494058, 'learning_rate/qwen_vl_interface': 3.592255895963986e-06, 'learning_rate/action_model': 3.592255895963986e-05, 'epoch': 1.67}) +[2026-06-18 02:33:55,354][starVLA.training.train_starvla][INFO] - Step 4334, Loss: {'train/loss': 0.47107288241386414, 'train/grad_norm_pre_clip': 15.818741798400879, 'timing/data': 0.001127786934375763, 'timing/model': 1.194528159685433, 'learning_rate/qwen_vl_interface': 3.590229360358731e-06, 'learning_rate/action_model': 3.590229360358731e-05, 'epoch': 1.67}) +[2026-06-18 02:34:01,514][starVLA.training.train_starvla][INFO] - Step 4335, Loss: {'train/loss': 0.20545457303524017, 'train/grad_norm_pre_clip': 8.116705894470215, 'timing/data': 0.0009594159200787544, 'timing/model': 1.3204505434259772, 'learning_rate/qwen_vl_interface': 3.5882031688257147e-06, 'learning_rate/action_model': 3.588203168825715e-05, 'epoch': 1.67}) +[2026-06-18 02:34:09,771][starVLA.training.train_starvla][INFO] - Step 4336, Loss: {'train/loss': 0.5286148190498352, 'train/grad_norm_pre_clip': 15.117705345153809, 'timing/data': 0.0008275024592876434, 'timing/model': 4.509159496054053, 'learning_rate/qwen_vl_interface': 3.5861773217849694e-06, 'learning_rate/action_model': 3.5861773217849693e-05, 'epoch': 1.67}) +[2026-06-18 02:34:15,692][starVLA.training.train_starvla][INFO] - Step 4337, Loss: {'train/loss': 0.38250085711479187, 'train/grad_norm_pre_clip': 9.578598976135254, 'timing/data': 0.0011393297463655472, 'timing/model': 2.304339487105608, 'learning_rate/qwen_vl_interface': 3.5841518196564566e-06, 'learning_rate/action_model': 3.584151819656457e-05, 'epoch': 1.67}) +[2026-06-18 02:34:21,132][starVLA.training.train_starvla][INFO] - Step 4338, Loss: {'train/loss': 0.41292715072631836, 'train/grad_norm_pre_clip': 14.374817848205566, 'timing/data': 0.001257321797311306, 'timing/model': 1.2779581127688289, 'learning_rate/qwen_vl_interface': 3.582126662860065e-06, 'learning_rate/action_model': 3.582126662860065e-05, 'epoch': 1.67}) +[2026-06-18 02:34:29,510][starVLA.training.train_starvla][INFO] - Step 4339, Loss: {'train/loss': 0.5212762355804443, 'train/grad_norm_pre_clip': 16.13065528869629, 'timing/data': 0.004926089197397232, 'timing/model': 4.400315367616713, 'learning_rate/qwen_vl_interface': 3.5801018518156096e-06, 'learning_rate/action_model': 3.5801018518156096e-05, 'epoch': 1.67}) +[2026-06-18 02:34:40,927][starVLA.training.train_starvla][INFO] - Step 4340, Loss: {'train/loss': 0.24104057252407074, 'train/grad_norm_pre_clip': 8.316930770874023, 'timing/data': 0.0010771295055747032, 'timing/model': 1.8698441591113806, 'learning_rate/qwen_vl_interface': 3.5780773869428366e-06, 'learning_rate/action_model': 3.5780773869428364e-05, 'epoch': 1.67}) +[2026-06-18 02:34:46,728][starVLA.training.train_starvla][INFO] - Step 4341, Loss: {'train/loss': 0.3897881805896759, 'train/grad_norm_pre_clip': 9.339194297790527, 'timing/data': 0.0012410124763846397, 'timing/model': 1.35744836833328, 'learning_rate/qwen_vl_interface': 3.576053268661422e-06, 'learning_rate/action_model': 3.5760532686614216e-05, 'epoch': 1.67}) +[2026-06-18 02:35:00,152][starVLA.training.train_starvla][INFO] - Step 4342, Loss: {'train/loss': 0.2126019299030304, 'train/grad_norm_pre_clip': 12.71610164642334, 'timing/data': 0.003093445673584938, 'timing/model': 5.330570914782584, 'learning_rate/qwen_vl_interface': 3.5740294973909643e-06, 'learning_rate/action_model': 3.5740294973909644e-05, 'epoch': 1.67}) +[2026-06-18 02:35:05,362][starVLA.training.train_starvla][INFO] - Step 4343, Loss: {'train/loss': 0.3167572021484375, 'train/grad_norm_pre_clip': 11.176456451416016, 'timing/data': 0.00126651581376791, 'timing/model': 1.10230559296906, 'learning_rate/qwen_vl_interface': 3.5720060735509976e-06, 'learning_rate/action_model': 3.5720060735509975e-05, 'epoch': 1.67}) +[2026-06-18 02:35:11,248][starVLA.training.train_starvla][INFO] - Step 4344, Loss: {'train/loss': 0.3861948549747467, 'train/grad_norm_pre_clip': 21.288175582885742, 'timing/data': 0.0011872649192810059, 'timing/model': 1.2506481129676104, 'learning_rate/qwen_vl_interface': 3.569982997560974e-06, 'learning_rate/action_model': 3.5699829975609736e-05, 'epoch': 1.67}) +[2026-06-18 02:35:16,597][starVLA.training.train_starvla][INFO] - Step 4345, Loss: {'train/loss': 0.19342491030693054, 'train/grad_norm_pre_clip': 7.508548736572266, 'timing/data': 0.0016162879765033722, 'timing/model': 1.40235382784158, 'learning_rate/qwen_vl_interface': 3.567960269840286e-06, 'learning_rate/action_model': 3.567960269840286e-05, 'epoch': 1.67}) +[2026-06-18 02:35:23,830][starVLA.training.train_starvla][INFO] - Step 4346, Loss: {'train/loss': 0.2092200517654419, 'train/grad_norm_pre_clip': 9.797325134277344, 'timing/data': 0.0011874940246343613, 'timing/model': 1.5834373831748962, 'learning_rate/qwen_vl_interface': 3.5659378908082427e-06, 'learning_rate/action_model': 3.5659378908082426e-05, 'epoch': 1.67}) +[2026-06-18 02:35:28,897][starVLA.training.train_starvla][INFO] - Step 4347, Loss: {'train/loss': 0.4195341169834137, 'train/grad_norm_pre_clip': 16.57756996154785, 'timing/data': 0.0016656648367643356, 'timing/model': 1.2353760665282607, 'learning_rate/qwen_vl_interface': 3.5639158608840862e-06, 'learning_rate/action_model': 3.563915860884086e-05, 'epoch': 1.67}) +[2026-06-18 02:35:34,319][starVLA.training.train_starvla][INFO] - Step 4348, Loss: {'train/loss': 0.34604015946388245, 'train/grad_norm_pre_clip': 12.33775520324707, 'timing/data': 0.0013461168855428696, 'timing/model': 1.2450890420004725, 'learning_rate/qwen_vl_interface': 3.5618941804869866e-06, 'learning_rate/action_model': 3.561894180486987e-05, 'epoch': 1.67}) +[2026-06-18 02:35:40,478][starVLA.training.train_starvla][INFO] - Step 4349, Loss: {'train/loss': 0.6133532524108887, 'train/grad_norm_pre_clip': 19.208566665649414, 'timing/data': 0.0013165371492505074, 'timing/model': 1.5676246993243694, 'learning_rate/qwen_vl_interface': 3.5598728500360395e-06, 'learning_rate/action_model': 3.55987285003604e-05, 'epoch': 1.67}) +[2026-06-18 02:35:50,405][starVLA.training.train_starvla][INFO] - Step 4350, Loss: {'train/loss': 0.811540424823761, 'train/grad_norm_pre_clip': 27.426013946533203, 'timing/data': 0.0010955380275845528, 'timing/model': 1.1779363080859184, 'learning_rate/qwen_vl_interface': 3.557851869950272e-06, 'learning_rate/action_model': 3.557851869950272e-05, 'epoch': 1.67}) +[2026-06-18 02:35:55,444][starVLA.training.train_starvla][INFO] - Step 4351, Loss: {'train/loss': 0.4588720500469208, 'train/grad_norm_pre_clip': 16.43098258972168, 'timing/data': 0.0012033013626933098, 'timing/model': 1.5993812093511224, 'learning_rate/qwen_vl_interface': 3.555831240648628e-06, 'learning_rate/action_model': 3.5558312406486274e-05, 'epoch': 1.67}) +[2026-06-18 02:36:00,317][starVLA.training.train_starvla][INFO] - Step 4352, Loss: {'train/loss': 0.3261403739452362, 'train/grad_norm_pre_clip': 12.046426773071289, 'timing/data': 0.0012577446177601814, 'timing/model': 1.56927125249058, 'learning_rate/qwen_vl_interface': 3.553810962549994e-06, 'learning_rate/action_model': 3.553810962549994e-05, 'epoch': 1.67}) +[2026-06-18 02:36:04,873][starVLA.training.train_starvla][INFO] - Step 4353, Loss: {'train/loss': 0.4735524654388428, 'train/grad_norm_pre_clip': 18.638242721557617, 'timing/data': 0.0012097395956516266, 'timing/model': 1.1839060075581074, 'learning_rate/qwen_vl_interface': 3.5517910360731723e-06, 'learning_rate/action_model': 3.551791036073172e-05, 'epoch': 1.67}) +[2026-06-18 02:36:10,283][starVLA.training.train_starvla][INFO] - Step 4354, Loss: {'train/loss': 0.301288902759552, 'train/grad_norm_pre_clip': 10.85912036895752, 'timing/data': 0.0015333546325564384, 'timing/model': 1.211812469176948, 'learning_rate/qwen_vl_interface': 3.549771461636896e-06, 'learning_rate/action_model': 3.549771461636896e-05, 'epoch': 1.67}) +[2026-06-18 02:36:15,176][starVLA.training.train_starvla][INFO] - Step 4355, Loss: {'train/loss': 0.392238974571228, 'train/grad_norm_pre_clip': 13.588418006896973, 'timing/data': 0.0009128283709287643, 'timing/model': 1.2397624468430877, 'learning_rate/qwen_vl_interface': 3.5477522396598236e-06, 'learning_rate/action_model': 3.5477522396598235e-05, 'epoch': 1.68}) +[2026-06-18 02:36:20,020][starVLA.training.train_starvla][INFO] - Step 4356, Loss: {'train/loss': 0.2817677855491638, 'train/grad_norm_pre_clip': 10.62331485748291, 'timing/data': 0.0013889875262975693, 'timing/model': 1.1902946699410677, 'learning_rate/qwen_vl_interface': 3.5457333705605456e-06, 'learning_rate/action_model': 3.5457333705605456e-05, 'epoch': 1.68}) +[2026-06-18 02:36:24,550][starVLA.training.train_starvla][INFO] - Step 4357, Loss: {'train/loss': 0.35645633935928345, 'train/grad_norm_pre_clip': 12.414484977722168, 'timing/data': 0.001389501616358757, 'timing/model': 1.1049196496605873, 'learning_rate/qwen_vl_interface': 3.543714854757574e-06, 'learning_rate/action_model': 3.5437148547575736e-05, 'epoch': 1.68}) +[2026-06-18 02:36:29,593][starVLA.training.train_starvla][INFO] - Step 4358, Loss: {'train/loss': 0.35293546319007874, 'train/grad_norm_pre_clip': 15.478143692016602, 'timing/data': 0.001212839037179947, 'timing/model': 1.2461359249427915, 'learning_rate/qwen_vl_interface': 3.5416966926693464e-06, 'learning_rate/action_model': 3.5416966926693465e-05, 'epoch': 1.68}) +[2026-06-18 02:36:34,447][starVLA.training.train_starvla][INFO] - Step 4359, Loss: {'train/loss': 0.22946318984031677, 'train/grad_norm_pre_clip': 11.057268142700195, 'timing/data': 0.0011739004403352737, 'timing/model': 1.1294925678521395, 'learning_rate/qwen_vl_interface': 3.539678884714235e-06, 'learning_rate/action_model': 3.539678884714235e-05, 'epoch': 1.68}) +[2026-06-18 02:36:39,217][starVLA.training.train_starvla][INFO] - Step 4360, Loss: {'train/loss': 0.1378334015607834, 'train/grad_norm_pre_clip': 8.36894416809082, 'timing/data': 0.0011492250487208366, 'timing/model': 1.1148144332692027, 'learning_rate/qwen_vl_interface': 3.5376614313105307e-06, 'learning_rate/action_model': 3.53766143131053e-05, 'epoch': 1.68}) +[2026-06-18 02:36:43,957][starVLA.training.train_starvla][INFO] - Step 4361, Loss: {'train/loss': 0.36168113350868225, 'train/grad_norm_pre_clip': 12.937527656555176, 'timing/data': 0.000790654681622982, 'timing/model': 1.2666768375784159, 'learning_rate/qwen_vl_interface': 3.5356443328764546e-06, 'learning_rate/action_model': 3.5356443328764546e-05, 'epoch': 1.68}) +[2026-06-18 02:36:48,953][starVLA.training.train_starvla][INFO] - Step 4362, Loss: {'train/loss': 0.549548864364624, 'train/grad_norm_pre_clip': 14.575180053710938, 'timing/data': 0.0013645654544234276, 'timing/model': 1.6216350384056568, 'learning_rate/qwen_vl_interface': 3.5336275898301515e-06, 'learning_rate/action_model': 3.5336275898301514e-05, 'epoch': 1.68}) +[2026-06-18 02:36:53,784][starVLA.training.train_starvla][INFO] - Step 4363, Loss: {'train/loss': 0.16857077181339264, 'train/grad_norm_pre_clip': 10.941913604736328, 'timing/data': 0.0013657892122864723, 'timing/model': 1.288465952500701, 'learning_rate/qwen_vl_interface': 3.531611202589697e-06, 'learning_rate/action_model': 3.531611202589697e-05, 'epoch': 1.68}) +[2026-06-18 02:36:59,391][starVLA.training.train_starvla][INFO] - Step 4364, Loss: {'train/loss': 0.5457174181938171, 'train/grad_norm_pre_clip': 14.955687522888184, 'timing/data': 0.0011101104319095612, 'timing/model': 1.2349217040464282, 'learning_rate/qwen_vl_interface': 3.5295951715730902e-06, 'learning_rate/action_model': 3.52959517157309e-05, 'epoch': 1.68}) +[2026-06-18 02:37:09,086][starVLA.training.train_starvla][INFO] - Step 4365, Loss: {'train/loss': 0.3380877673625946, 'train/grad_norm_pre_clip': 22.05315399169922, 'timing/data': 0.0013113273307681084, 'timing/model': 1.8491127481684089, 'learning_rate/qwen_vl_interface': 3.527579497198255e-06, 'learning_rate/action_model': 3.527579497198255e-05, 'epoch': 1.68}) +[2026-06-18 02:37:15,778][starVLA.training.train_starvla][INFO] - Step 4366, Loss: {'train/loss': 0.2708756923675537, 'train/grad_norm_pre_clip': 10.196516990661621, 'timing/data': 0.0015658177435398102, 'timing/model': 2.530566639266908, 'learning_rate/qwen_vl_interface': 3.525564179883042e-06, 'learning_rate/action_model': 3.525564179883042e-05, 'epoch': 1.68}) +[2026-06-18 02:37:25,622][starVLA.training.train_starvla][INFO] - Step 4367, Loss: {'train/loss': 0.2052120864391327, 'train/grad_norm_pre_clip': 8.824485778808594, 'timing/data': 0.001429927535355091, 'timing/model': 4.571310272440314, 'learning_rate/qwen_vl_interface': 3.5235492200452308e-06, 'learning_rate/action_model': 3.523549220045231e-05, 'epoch': 1.68}) +[2026-06-18 02:37:34,871][starVLA.training.train_starvla][INFO] - Step 4368, Loss: {'train/loss': 0.6601219773292542, 'train/grad_norm_pre_clip': 15.236108779907227, 'timing/data': 0.0011665541678667068, 'timing/model': 1.1287212278693914, 'learning_rate/qwen_vl_interface': 3.521534618102525e-06, 'learning_rate/action_model': 3.5215346181025246e-05, 'epoch': 1.68}) +[2026-06-18 02:37:41,530][starVLA.training.train_starvla][INFO] - Step 4369, Loss: {'train/loss': 0.35005247592926025, 'train/grad_norm_pre_clip': 15.038508415222168, 'timing/data': 0.0011681606993079185, 'timing/model': 1.0865320796146989, 'learning_rate/qwen_vl_interface': 3.5195203744725516e-06, 'learning_rate/action_model': 3.5195203744725516e-05, 'epoch': 1.68}) +[2026-06-18 02:37:49,773][starVLA.training.train_starvla][INFO] - Step 4370, Loss: {'train/loss': 0.4709303677082062, 'train/grad_norm_pre_clip': 17.150279998779297, 'timing/data': 0.0012533152475953102, 'timing/model': 1.298811400309205, 'learning_rate/qwen_vl_interface': 3.5175064895728665e-06, 'learning_rate/action_model': 3.5175064895728664e-05, 'epoch': 1.68}) +[2026-06-18 02:37:54,983][starVLA.training.train_starvla][INFO] - Step 4371, Loss: {'train/loss': 0.18474739789962769, 'train/grad_norm_pre_clip': 7.660402774810791, 'timing/data': 0.0011671865358948708, 'timing/model': 1.5037188734859228, 'learning_rate/qwen_vl_interface': 3.515492963820951e-06, 'learning_rate/action_model': 3.515492963820951e-05, 'epoch': 1.68}) +[2026-06-18 02:37:59,774][starVLA.training.train_starvla][INFO] - Step 4372, Loss: {'train/loss': 0.3124791383743286, 'train/grad_norm_pre_clip': 15.847110748291016, 'timing/data': 0.0010484838858246803, 'timing/model': 1.5772070540115237, 'learning_rate/qwen_vl_interface': 3.5134797976342123e-06, 'learning_rate/action_model': 3.513479797634212e-05, 'epoch': 1.68}) +[2026-06-18 02:38:11,850][starVLA.training.train_starvla][INFO] - Step 4373, Loss: {'train/loss': 0.2692522406578064, 'train/grad_norm_pre_clip': 11.927566528320312, 'timing/data': 0.0013262582942843437, 'timing/model': 1.0983680272474885, 'learning_rate/qwen_vl_interface': 3.5114669914299776e-06, 'learning_rate/action_model': 3.511466991429978e-05, 'epoch': 1.68}) +[2026-06-18 02:38:17,810][starVLA.training.train_starvla][INFO] - Step 4374, Loss: {'train/loss': 0.43176862597465515, 'train/grad_norm_pre_clip': 15.780027389526367, 'timing/data': 0.0014635780826210976, 'timing/model': 1.6190396137535572, 'learning_rate/qwen_vl_interface': 3.5094545456255065e-06, 'learning_rate/action_model': 3.509454545625506e-05, 'epoch': 1.68}) +[2026-06-18 02:38:22,829][starVLA.training.train_starvla][INFO] - Step 4375, Loss: {'train/loss': 0.4259721338748932, 'train/grad_norm_pre_clip': 12.331268310546875, 'timing/data': 0.0010128486901521683, 'timing/model': 1.1135272337123752, 'learning_rate/qwen_vl_interface': 3.5074424606379826e-06, 'learning_rate/action_model': 3.5074424606379824e-05, 'epoch': 1.68}) +[2026-06-18 02:38:28,196][starVLA.training.train_starvla][INFO] - Step 4376, Loss: {'train/loss': 0.34115949273109436, 'train/grad_norm_pre_clip': 10.844396591186523, 'timing/data': 0.001277979463338852, 'timing/model': 1.146863829344511, 'learning_rate/qwen_vl_interface': 3.505430736884511e-06, 'learning_rate/action_model': 3.5054307368845105e-05, 'epoch': 1.68}) +[2026-06-18 02:38:34,787][starVLA.training.train_starvla][INFO] - Step 4377, Loss: {'train/loss': 0.4038134515285492, 'train/grad_norm_pre_clip': 14.253081321716309, 'timing/data': 0.0008939048275351524, 'timing/model': 2.582163443788886, 'learning_rate/qwen_vl_interface': 3.503419374782123e-06, 'learning_rate/action_model': 3.503419374782123e-05, 'epoch': 1.68}) +[2026-06-18 02:38:40,397][starVLA.training.train_starvla][INFO] - Step 4378, Loss: {'train/loss': 0.4084371030330658, 'train/grad_norm_pre_clip': 16.064437866210938, 'timing/data': 0.005258448421955109, 'timing/model': 1.385767082683742, 'learning_rate/qwen_vl_interface': 3.50140837474778e-06, 'learning_rate/action_model': 3.50140837474778e-05, 'epoch': 1.68}) +[2026-06-18 02:38:46,267][starVLA.training.train_starvla][INFO] - Step 4379, Loss: {'train/loss': 0.40386512875556946, 'train/grad_norm_pre_clip': 20.24813461303711, 'timing/data': 0.0023564817383885384, 'timing/model': 1.2634577257558703, 'learning_rate/qwen_vl_interface': 3.4993977371983625e-06, 'learning_rate/action_model': 3.499397737198362e-05, 'epoch': 1.68}) +[2026-06-18 02:38:52,171][starVLA.training.train_starvla][INFO] - Step 4380, Loss: {'train/loss': 0.2791409194469452, 'train/grad_norm_pre_clip': 11.928715705871582, 'timing/data': 0.0017652614042162895, 'timing/model': 1.1963387429714203, 'learning_rate/qwen_vl_interface': 3.497387462550676e-06, 'learning_rate/action_model': 3.4973874625506757e-05, 'epoch': 1.68}) +[2026-06-18 02:39:01,634][starVLA.training.train_starvla][INFO] - Step 4381, Loss: {'train/loss': 0.08895678073167801, 'train/grad_norm_pre_clip': 5.70665979385376, 'timing/data': 0.0011221719905734062, 'timing/model': 1.116791721433401, 'learning_rate/qwen_vl_interface': 3.4953775512214568e-06, 'learning_rate/action_model': 3.495377551221457e-05, 'epoch': 1.69}) +[2026-06-18 02:39:06,553][starVLA.training.train_starvla][INFO] - Step 4382, Loss: {'train/loss': 0.25264957547187805, 'train/grad_norm_pre_clip': 7.991087436676025, 'timing/data': 0.0009081140160560608, 'timing/model': 1.2889548689126968, 'learning_rate/qwen_vl_interface': 3.4933680036273583e-06, 'learning_rate/action_model': 3.493368003627358e-05, 'epoch': 1.69}) +[2026-06-18 02:39:17,709][starVLA.training.train_starvla][INFO] - Step 4383, Loss: {'train/loss': 0.7979686260223389, 'train/grad_norm_pre_clip': 19.41298484802246, 'timing/data': 0.0016229161992669106, 'timing/model': 2.5785325653851032, 'learning_rate/qwen_vl_interface': 3.491358820184963e-06, 'learning_rate/action_model': 3.491358820184963e-05, 'epoch': 1.69}) +[2026-06-18 02:39:35,376][starVLA.training.train_starvla][INFO] - Step 4384, Loss: {'train/loss': 0.407856822013855, 'train/grad_norm_pre_clip': 14.936200141906738, 'timing/data': 0.0011928174644708633, 'timing/model': 4.624018458649516, 'learning_rate/qwen_vl_interface': 3.4893500013107758e-06, 'learning_rate/action_model': 3.4893500013107754e-05, 'epoch': 1.69}) +[2026-06-18 02:39:44,051][starVLA.training.train_starvla][INFO] - Step 4385, Loss: {'train/loss': 0.28364771604537964, 'train/grad_norm_pre_clip': 10.685199737548828, 'timing/data': 0.0017572389915585518, 'timing/model': 1.1391839738935232, 'learning_rate/qwen_vl_interface': 3.4873415474212287e-06, 'learning_rate/action_model': 3.487341547421228e-05, 'epoch': 1.69}) +[2026-06-18 02:39:49,304][starVLA.training.train_starvla][INFO] - Step 4386, Loss: {'train/loss': 0.660520076751709, 'train/grad_norm_pre_clip': 57.20357894897461, 'timing/data': 0.0023299632593989372, 'timing/model': 1.6038406221196055, 'learning_rate/qwen_vl_interface': 3.4853334589326764e-06, 'learning_rate/action_model': 3.4853334589326765e-05, 'epoch': 1.69}) +[2026-06-18 02:39:54,157][starVLA.training.train_starvla][INFO] - Step 4387, Loss: {'train/loss': 0.650691032409668, 'train/grad_norm_pre_clip': 19.970006942749023, 'timing/data': 0.0007923254743218422, 'timing/model': 1.5851170644164085, 'learning_rate/qwen_vl_interface': 3.483325736261394e-06, 'learning_rate/action_model': 3.483325736261394e-05, 'epoch': 1.69}) +[2026-06-18 02:39:59,832][starVLA.training.train_starvla][INFO] - Step 4388, Loss: {'train/loss': 0.22245268523693085, 'train/grad_norm_pre_clip': 8.640497207641602, 'timing/data': 0.0011999411508440971, 'timing/model': 1.5314972205087543, 'learning_rate/qwen_vl_interface': 3.48131837982359e-06, 'learning_rate/action_model': 3.48131837982359e-05, 'epoch': 1.69}) +[2026-06-18 02:40:12,007][starVLA.training.train_starvla][INFO] - Step 4389, Loss: {'train/loss': 0.29855042695999146, 'train/grad_norm_pre_clip': 10.288980484008789, 'timing/data': 0.0024782484397292137, 'timing/model': 4.9826626144349575, 'learning_rate/qwen_vl_interface': 3.4793113900353885e-06, 'learning_rate/action_model': 3.479311390035389e-05, 'epoch': 1.69}) +[2026-06-18 02:40:22,135][starVLA.training.train_starvla][INFO] - Step 4390, Loss: {'train/loss': 0.38870397210121155, 'train/grad_norm_pre_clip': 16.087005615234375, 'timing/data': 0.0012103309854865074, 'timing/model': 1.8694020062685013, 'learning_rate/qwen_vl_interface': 3.4773047673128412e-06, 'learning_rate/action_model': 3.477304767312841e-05, 'epoch': 1.69}) +[2026-06-18 02:40:29,920][starVLA.training.train_starvla][INFO] - Step 4391, Loss: {'train/loss': 0.33291059732437134, 'train/grad_norm_pre_clip': 12.591761589050293, 'timing/data': 0.0017604641616344452, 'timing/model': 1.1039385674521327, 'learning_rate/qwen_vl_interface': 3.4752985120719225e-06, 'learning_rate/action_model': 3.475298512071922e-05, 'epoch': 1.69}) +[2026-06-18 02:40:39,202][starVLA.training.train_starvla][INFO] - Step 4392, Loss: {'train/loss': 0.4214337170124054, 'train/grad_norm_pre_clip': 18.502704620361328, 'timing/data': 0.0009578149765729904, 'timing/model': 4.829889526590705, 'learning_rate/qwen_vl_interface': 3.4732926247285324e-06, 'learning_rate/action_model': 3.4732926247285325e-05, 'epoch': 1.69}) +[2026-06-18 02:40:44,566][starVLA.training.train_starvla][INFO] - Step 4393, Loss: {'train/loss': 0.2414613962173462, 'train/grad_norm_pre_clip': 12.231369972229004, 'timing/data': 0.0014246627688407898, 'timing/model': 1.2083935076370835, 'learning_rate/qwen_vl_interface': 3.471287105698494e-06, 'learning_rate/action_model': 3.471287105698494e-05, 'epoch': 1.69}) +[2026-06-18 02:40:50,753][starVLA.training.train_starvla][INFO] - Step 4394, Loss: {'train/loss': 0.5337198376655579, 'train/grad_norm_pre_clip': 14.020495414733887, 'timing/data': 0.0028788503259420395, 'timing/model': 1.1336859436705709, 'learning_rate/qwen_vl_interface': 3.4692819553975514e-06, 'learning_rate/action_model': 3.469281955397551e-05, 'epoch': 1.69}) +[2026-06-18 02:40:55,971][starVLA.training.train_starvla][INFO] - Step 4395, Loss: {'train/loss': 0.24731172621250153, 'train/grad_norm_pre_clip': 12.694286346435547, 'timing/data': 0.0012000128626823425, 'timing/model': 1.1184929851442575, 'learning_rate/qwen_vl_interface': 3.4672771742413754e-06, 'learning_rate/action_model': 3.467277174241375e-05, 'epoch': 1.69}) +[2026-06-18 02:41:01,331][starVLA.training.train_starvla][INFO] - Step 4396, Loss: {'train/loss': 0.3652525842189789, 'train/grad_norm_pre_clip': 19.675399780273438, 'timing/data': 0.0008442280814051628, 'timing/model': 1.3545560520142317, 'learning_rate/qwen_vl_interface': 3.46527276264556e-06, 'learning_rate/action_model': 3.46527276264556e-05, 'epoch': 1.69}) +[2026-06-18 02:41:06,733][starVLA.training.train_starvla][INFO] - Step 4397, Loss: {'train/loss': 0.4309745728969574, 'train/grad_norm_pre_clip': 13.732975959777832, 'timing/data': 0.0013819606974720955, 'timing/model': 1.8388950256630778, 'learning_rate/qwen_vl_interface': 3.4632687210256218e-06, 'learning_rate/action_model': 3.463268721025621e-05, 'epoch': 1.69}) +[2026-06-18 02:41:11,895][starVLA.training.train_starvla][INFO] - Step 4398, Loss: {'train/loss': 0.23246659338474274, 'train/grad_norm_pre_clip': 11.175745964050293, 'timing/data': 0.0010958695784211159, 'timing/model': 1.2876722998917103, 'learning_rate/qwen_vl_interface': 3.461265049796999e-06, 'learning_rate/action_model': 3.461265049796999e-05, 'epoch': 1.69}) +[2026-06-18 02:41:17,527][starVLA.training.train_starvla][INFO] - Step 4399, Loss: {'train/loss': 0.24043500423431396, 'train/grad_norm_pre_clip': 13.217073440551758, 'timing/data': 0.0011897953227162361, 'timing/model': 1.3282628143206239, 'learning_rate/qwen_vl_interface': 3.459261749375057e-06, 'learning_rate/action_model': 3.4592617493750567e-05, 'epoch': 1.69}) +[2026-06-18 02:46:25,145][starVLA.training.train_starvla][INFO] - Step 4400, Loss: {'train/loss': 0.24132755398750305, 'train/grad_norm_pre_clip': 13.034090042114258, 'eval/loss': 0.44470566511154175, 'eval/latency_4/loss': 0.44470566511154175, 'eval/demon_fire_precision': 0.9664536741214057, 'eval/demon_fire_recall': 0.9824618382591751, 'eval/demon_fire_f1': 0.9743920115960703, 'eval/demon_fire_rate_model': 0.9192364170337739, 'eval/demon_fire_rate_teacher': 0.9042584434654919, 'eval/demon_left_precision': 0.8, 'eval/demon_left_recall': 0.639269406392694, 'eval/demon_left_f1': 0.7106598984771573, 'eval/demon_left_rate_model': 0.1027900146842878, 'eval/demon_left_rate_teacher': 0.12863436123348018, 'eval/demon_right_precision': 0.7471451876019576, 'eval/demon_right_recall': 0.8267148014440433, 'eval/demon_right_f1': 0.7849185946872324, 'eval/demon_right_rate_model': 0.18002936857562407, 'eval/demon_right_rate_teacher': 0.16270190895741557, 'eval/demon_move_f1': 0.7477892465821948, 'eval/demon_cc_f1': 0.8931926894435398, 'eval/latency_4/demon_fire_precision': 0.9664536741214057, 'eval/latency_4/demon_fire_recall': 0.9824618382591751, 'eval/latency_4/demon_fire_f1': 0.9743920115960703, 'eval/latency_4/demon_fire_rate_model': 0.9192364170337739, 'eval/latency_4/demon_fire_rate_teacher': 0.9042584434654919, 'eval/latency_4/demon_left_precision': 0.8, 'eval/latency_4/demon_left_recall': 0.639269406392694, 'eval/latency_4/demon_left_f1': 0.7106598984771573, 'eval/latency_4/demon_left_rate_model': 0.1027900146842878, 'eval/latency_4/demon_left_rate_teacher': 0.12863436123348018, 'eval/latency_4/demon_right_precision': 0.7471451876019576, 'eval/latency_4/demon_right_recall': 0.8267148014440433, 'eval/latency_4/demon_right_f1': 0.7849185946872324, 'eval/latency_4/demon_right_rate_model': 0.18002936857562407, 'eval/latency_4/demon_right_rate_teacher': 0.16270190895741557, 'eval/latency_4/demon_move_f1': 0.7477892465821948, 'eval/latency_4/demon_cc_f1': 0.8931926894435398, 'timing/data': 0.0013789944350719452, 'timing/model': 1.1354740057140589, 'learning_rate/qwen_vl_interface': 3.457258820175082e-06, 'learning_rate/action_model': 3.457258820175082e-05, 'epoch': 1.69}) +[2026-06-18 02:46:25,159][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_4400_state +[2026-06-18 02:46:25,167][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 02:46:34,181][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_4400_state/model.safetensors +[2026-06-18 02:46:57,164][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_4400_state/optimizer.bin +[2026-06-18 02:46:57,166][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_4400_state/sampler.bin +[2026-06-18 02:46:57,167][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_4400_state/sampler_1.bin +[2026-06-18 02:46:57,170][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_4400_state/random_states_0.pkl +[2026-06-18 02:46:57,176][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4400 +[2026-06-18 02:46:57,176][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 02:46:57,194][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 02:47:03,124][starVLA.training.train_starvla][INFO] - Step 4401, Loss: {'train/loss': 0.40336528420448303, 'train/grad_norm_pre_clip': 12.200950622558594, 'timing/data': 0.0010645976290106773, 'timing/model': 1.0816730903461576, 'learning_rate/qwen_vl_interface': 3.4552562626122817e-06, 'learning_rate/action_model': 3.455256262612281e-05, 'epoch': 1.69}) +[2026-06-18 02:47:16,081][starVLA.training.train_starvla][INFO] - Step 4402, Loss: {'train/loss': 0.2434065341949463, 'train/grad_norm_pre_clip': 11.963356971740723, 'timing/data': 0.001080942340195179, 'timing/model': 3.2211263524368405, 'learning_rate/qwen_vl_interface': 3.453254077101789e-06, 'learning_rate/action_model': 3.453254077101789e-05, 'epoch': 1.69}) +[2026-06-18 02:47:26,808][starVLA.training.train_starvla][INFO] - Step 4403, Loss: {'train/loss': 0.6044511198997498, 'train/grad_norm_pre_clip': 16.34465217590332, 'timing/data': 0.0012309318408370018, 'timing/model': 5.998841653577983, 'learning_rate/qwen_vl_interface': 3.45125226405866e-06, 'learning_rate/action_model': 3.4512522640586595e-05, 'epoch': 1.69}) +[2026-06-18 02:47:32,356][starVLA.training.train_starvla][INFO] - Step 4404, Loss: {'train/loss': 0.5949658751487732, 'train/grad_norm_pre_clip': 18.06060791015625, 'timing/data': 0.0008598044514656067, 'timing/model': 1.4636049401015043, 'learning_rate/qwen_vl_interface': 3.449250823897872e-06, 'learning_rate/action_model': 3.449250823897872e-05, 'epoch': 1.69}) +[2026-06-18 02:47:37,831][starVLA.training.train_starvla][INFO] - Step 4405, Loss: {'train/loss': 0.284656286239624, 'train/grad_norm_pre_clip': 10.939292907714844, 'timing/data': 0.0008500255644321442, 'timing/model': 1.094149119220674, 'learning_rate/qwen_vl_interface': 3.447249757034328e-06, 'learning_rate/action_model': 3.4472497570343276e-05, 'epoch': 1.69}) +[2026-06-18 02:47:43,223][starVLA.training.train_starvla][INFO] - Step 4406, Loss: {'train/loss': 0.3634786307811737, 'train/grad_norm_pre_clip': 12.572288513183594, 'timing/data': 0.0011945059522986412, 'timing/model': 1.189166920259595, 'learning_rate/qwen_vl_interface': 3.445249063882845e-06, 'learning_rate/action_model': 3.445249063882845e-05, 'epoch': 1.69}) +[2026-06-18 02:47:49,725][starVLA.training.train_starvla][INFO] - Step 4407, Loss: {'train/loss': 0.49104592204093933, 'train/grad_norm_pre_clip': 14.895707130432129, 'timing/data': 0.0009283721446990967, 'timing/model': 1.3436877792701125, 'learning_rate/qwen_vl_interface': 3.4432487448581762e-06, 'learning_rate/action_model': 3.443248744858176e-05, 'epoch': 1.7}) +[2026-06-18 02:47:55,170][starVLA.training.train_starvla][INFO] - Step 4408, Loss: {'train/loss': 0.30045032501220703, 'train/grad_norm_pre_clip': 8.727768898010254, 'timing/data': 0.0008173901587724686, 'timing/model': 1.2092082863673568, 'learning_rate/qwen_vl_interface': 3.4412488003749867e-06, 'learning_rate/action_model': 3.4412488003749866e-05, 'epoch': 1.7}) +[2026-06-18 02:48:00,835][starVLA.training.train_starvla][INFO] - Step 4409, Loss: {'train/loss': 0.5084892511367798, 'train/grad_norm_pre_clip': 11.228900909423828, 'timing/data': 0.001147695817053318, 'timing/model': 1.1221088888123631, 'learning_rate/qwen_vl_interface': 3.4392492308478657e-06, 'learning_rate/action_model': 3.439249230847866e-05, 'epoch': 1.7}) +[2026-06-18 02:48:06,176][starVLA.training.train_starvla][INFO] - Step 4410, Loss: {'train/loss': 0.20154738426208496, 'train/grad_norm_pre_clip': 9.253124237060547, 'timing/data': 0.0010791867971420288, 'timing/model': 1.4714095685631037, 'learning_rate/qwen_vl_interface': 3.437250036691328e-06, 'learning_rate/action_model': 3.437250036691328e-05, 'epoch': 1.7}) +[2026-06-18 02:48:11,526][starVLA.training.train_starvla][INFO] - Step 4411, Loss: {'train/loss': 0.2994101643562317, 'train/grad_norm_pre_clip': 12.502266883850098, 'timing/data': 0.0010224292054772377, 'timing/model': 1.5955053055658937, 'learning_rate/qwen_vl_interface': 3.4352512183198094e-06, 'learning_rate/action_model': 3.4352512183198095e-05, 'epoch': 1.7}) +[2026-06-18 02:48:17,170][starVLA.training.train_starvla][INFO] - Step 4412, Loss: {'train/loss': 0.1693526953458786, 'train/grad_norm_pre_clip': 10.352032661437988, 'timing/data': 0.001227111555635929, 'timing/model': 1.8704554047435522, 'learning_rate/qwen_vl_interface': 3.4332527761476662e-06, 'learning_rate/action_model': 3.433252776147666e-05, 'epoch': 1.7}) +[2026-06-18 02:48:22,675][starVLA.training.train_starvla][INFO] - Step 4413, Loss: {'train/loss': 0.3277086317539215, 'train/grad_norm_pre_clip': 16.59702491760254, 'timing/data': 0.0012800106778740883, 'timing/model': 1.3886923464015126, 'learning_rate/qwen_vl_interface': 3.431254710589174e-06, 'learning_rate/action_model': 3.431254710589174e-05, 'epoch': 1.7}) +[2026-06-18 02:48:27,810][starVLA.training.train_starvla][INFO] - Step 4414, Loss: {'train/loss': 0.46734917163848877, 'train/grad_norm_pre_clip': 14.118152618408203, 'timing/data': 0.0011388827115297318, 'timing/model': 1.3240845939144492, 'learning_rate/qwen_vl_interface': 3.4292570220585407e-06, 'learning_rate/action_model': 3.4292570220585407e-05, 'epoch': 1.7}) +[2026-06-18 02:48:33,831][starVLA.training.train_starvla][INFO] - Step 4415, Loss: {'train/loss': 0.2274216115474701, 'train/grad_norm_pre_clip': 10.023781776428223, 'timing/data': 0.0010857665911316872, 'timing/model': 1.1871871100738645, 'learning_rate/qwen_vl_interface': 3.4272597109698857e-06, 'learning_rate/action_model': 3.427259710969886e-05, 'epoch': 1.7}) +[2026-06-18 02:48:40,373][starVLA.training.train_starvla][INFO] - Step 4416, Loss: {'train/loss': 0.41665130853652954, 'train/grad_norm_pre_clip': 12.476749420166016, 'timing/data': 0.0013349289074540138, 'timing/model': 1.1464006612077355, 'learning_rate/qwen_vl_interface': 3.4252627777372533e-06, 'learning_rate/action_model': 3.425262777737253e-05, 'epoch': 1.7}) +[2026-06-18 02:48:44,869][starVLA.training.train_starvla][INFO] - Step 4417, Loss: {'train/loss': 0.5455852150917053, 'train/grad_norm_pre_clip': 17.780559539794922, 'timing/data': 0.001355455256998539, 'timing/model': 1.1737867752090096, 'learning_rate/qwen_vl_interface': 3.4232662227746095e-06, 'learning_rate/action_model': 3.423266222774609e-05, 'epoch': 1.7}) +[2026-06-18 02:48:49,681][starVLA.training.train_starvla][INFO] - Step 4418, Loss: {'train/loss': 0.4443851411342621, 'train/grad_norm_pre_clip': 12.360584259033203, 'timing/data': 0.0012227632105350494, 'timing/model': 1.1693992977961898, 'learning_rate/qwen_vl_interface': 3.4212700464958448e-06, 'learning_rate/action_model': 3.421270046495845e-05, 'epoch': 1.7}) +[2026-06-18 02:48:57,978][starVLA.training.train_starvla][INFO] - Step 4419, Loss: {'train/loss': 0.26814815402030945, 'train/grad_norm_pre_clip': 13.653192520141602, 'timing/data': 0.0008943406865000725, 'timing/model': 1.519983446225524, 'learning_rate/qwen_vl_interface': 3.4192742493147686e-06, 'learning_rate/action_model': 3.419274249314768e-05, 'epoch': 1.7}) +[2026-06-18 02:49:02,985][starVLA.training.train_starvla][INFO] - Step 4420, Loss: {'train/loss': 0.09235984832048416, 'train/grad_norm_pre_clip': 5.614410877227783, 'timing/data': 0.0013701943680644035, 'timing/model': 1.2218911638483405, 'learning_rate/qwen_vl_interface': 3.4172788316451065e-06, 'learning_rate/action_model': 3.417278831645107e-05, 'epoch': 1.7}) +[2026-06-18 02:49:08,190][starVLA.training.train_starvla][INFO] - Step 4421, Loss: {'train/loss': 0.22922377288341522, 'train/grad_norm_pre_clip': 9.085907936096191, 'timing/data': 0.001306404359638691, 'timing/model': 1.237196421250701, 'learning_rate/qwen_vl_interface': 3.4152837939005177e-06, 'learning_rate/action_model': 3.415283793900518e-05, 'epoch': 1.7}) +[2026-06-18 02:49:17,023][starVLA.training.train_starvla][INFO] - Step 4422, Loss: {'train/loss': 0.5319266319274902, 'train/grad_norm_pre_clip': 20.71155548095703, 'timing/data': 0.003664071671664715, 'timing/model': 2.338469891808927, 'learning_rate/qwen_vl_interface': 3.413289136494572e-06, 'learning_rate/action_model': 3.413289136494572e-05, 'epoch': 1.7}) +[2026-06-18 02:49:22,824][starVLA.training.train_starvla][INFO] - Step 4423, Loss: {'train/loss': 0.43144702911376953, 'train/grad_norm_pre_clip': 14.97838020324707, 'timing/data': 0.0015572477132081985, 'timing/model': 1.4284424111247063, 'learning_rate/qwen_vl_interface': 3.411294859840765e-06, 'learning_rate/action_model': 3.411294859840764e-05, 'epoch': 1.7}) +[2026-06-18 02:49:29,768][starVLA.training.train_starvla][INFO] - Step 4424, Loss: {'train/loss': 0.2300533503293991, 'train/grad_norm_pre_clip': 9.766251564025879, 'timing/data': 0.0011873561888933182, 'timing/model': 2.9771236618980765, 'learning_rate/qwen_vl_interface': 3.409300964352511e-06, 'learning_rate/action_model': 3.409300964352511e-05, 'epoch': 1.7}) +[2026-06-18 02:49:39,085][starVLA.training.train_starvla][INFO] - Step 4425, Loss: {'train/loss': 0.34023359417915344, 'train/grad_norm_pre_clip': 13.946488380432129, 'timing/data': 0.0008114967495203018, 'timing/model': 5.362235278822482, 'learning_rate/qwen_vl_interface': 3.407307450443148e-06, 'learning_rate/action_model': 3.407307450443148e-05, 'epoch': 1.7}) +[2026-06-18 02:49:45,478][starVLA.training.train_starvla][INFO] - Step 4426, Loss: {'train/loss': 0.45521605014801025, 'train/grad_norm_pre_clip': 11.980951309204102, 'timing/data': 0.0012838048860430717, 'timing/model': 1.1184131456539035, 'learning_rate/qwen_vl_interface': 3.4053143185259353e-06, 'learning_rate/action_model': 3.405314318525935e-05, 'epoch': 1.7}) +[2026-06-18 02:49:50,258][starVLA.training.train_starvla][INFO] - Step 4427, Loss: {'train/loss': 0.4909052848815918, 'train/grad_norm_pre_clip': 25.05533218383789, 'timing/data': 0.0011023599654436111, 'timing/model': 1.3861124832183123, 'learning_rate/qwen_vl_interface': 3.4033215690140475e-06, 'learning_rate/action_model': 3.4033215690140476e-05, 'epoch': 1.7}) +[2026-06-18 02:49:54,719][starVLA.training.train_starvla][INFO] - Step 4428, Loss: {'train/loss': 0.3663731515407562, 'train/grad_norm_pre_clip': 21.635908126831055, 'timing/data': 0.0011592712253332138, 'timing/model': 1.0776650570333004, 'learning_rate/qwen_vl_interface': 3.4013292023205852e-06, 'learning_rate/action_model': 3.401329202320585e-05, 'epoch': 1.7}) +[2026-06-18 02:49:59,367][starVLA.training.train_starvla][INFO] - Step 4429, Loss: {'train/loss': 0.40361037850379944, 'train/grad_norm_pre_clip': 12.298100471496582, 'timing/data': 0.0015518460422754288, 'timing/model': 1.164499037899077, 'learning_rate/qwen_vl_interface': 3.3993372188585694e-06, 'learning_rate/action_model': 3.3993372188585694e-05, 'epoch': 1.7}) +[2026-06-18 02:50:08,356][starVLA.training.train_starvla][INFO] - Step 4430, Loss: {'train/loss': 0.29709509015083313, 'train/grad_norm_pre_clip': 11.262918472290039, 'timing/data': 0.001402595080435276, 'timing/model': 1.7663426427170634, 'learning_rate/qwen_vl_interface': 3.3973456190409397e-06, 'learning_rate/action_model': 3.39734561904094e-05, 'epoch': 1.7}) +[2026-06-18 02:50:14,689][starVLA.training.train_starvla][INFO] - Step 4431, Loss: {'train/loss': 0.41967523097991943, 'train/grad_norm_pre_clip': 15.643445014953613, 'timing/data': 0.0008176816627383232, 'timing/model': 1.184719966724515, 'learning_rate/qwen_vl_interface': 3.395354403280555e-06, 'learning_rate/action_model': 3.395354403280555e-05, 'epoch': 1.7}) +[2026-06-18 02:50:27,465][starVLA.training.train_starvla][INFO] - Step 4432, Loss: {'train/loss': 0.8086650967597961, 'train/grad_norm_pre_clip': 24.221139907836914, 'timing/data': 0.0012042755261063576, 'timing/model': 3.0713931238278747, 'learning_rate/qwen_vl_interface': 3.3933635719902e-06, 'learning_rate/action_model': 3.3933635719901995e-05, 'epoch': 1.7}) +[2026-06-18 02:50:34,993][starVLA.training.train_starvla][INFO] - Step 4433, Loss: {'train/loss': 0.17787866294384003, 'train/grad_norm_pre_clip': 9.196126937866211, 'timing/data': 0.0012790700420737267, 'timing/model': 1.2742537092417479, 'learning_rate/qwen_vl_interface': 3.3913731255825754e-06, 'learning_rate/action_model': 3.3913731255825754e-05, 'epoch': 1.71}) +[2026-06-18 02:50:40,714][starVLA.training.train_starvla][INFO] - Step 4434, Loss: {'train/loss': 0.5429486632347107, 'train/grad_norm_pre_clip': 15.985933303833008, 'timing/data': 0.0031182216480374336, 'timing/model': 1.613006859086454, 'learning_rate/qwen_vl_interface': 3.3893830644703003e-06, 'learning_rate/action_model': 3.3893830644703e-05, 'epoch': 1.71}) +[2026-06-18 02:50:45,668][starVLA.training.train_starvla][INFO] - Step 4435, Loss: {'train/loss': 0.2598176598548889, 'train/grad_norm_pre_clip': 10.425577163696289, 'timing/data': 0.001447359099984169, 'timing/model': 1.1575801884755492, 'learning_rate/qwen_vl_interface': 3.3873933890659174e-06, 'learning_rate/action_model': 3.3873933890659174e-05, 'epoch': 1.71}) +[2026-06-18 02:50:51,127][starVLA.training.train_starvla][INFO] - Step 4436, Loss: {'train/loss': 0.4316251277923584, 'train/grad_norm_pre_clip': 15.012147903442383, 'timing/data': 0.0012313676998019218, 'timing/model': 1.1614887621253729, 'learning_rate/qwen_vl_interface': 3.3854040997818897e-06, 'learning_rate/action_model': 3.38540409978189e-05, 'epoch': 1.71}) +[2026-06-18 02:50:56,744][starVLA.training.train_starvla][INFO] - Step 4437, Loss: {'train/loss': 0.3182973861694336, 'train/grad_norm_pre_clip': 12.461620330810547, 'timing/data': 0.0010529933497309685, 'timing/model': 1.8081283504143357, 'learning_rate/qwen_vl_interface': 3.383415197030599e-06, 'learning_rate/action_model': 3.383415197030599e-05, 'epoch': 1.71}) +[2026-06-18 02:51:07,890][starVLA.training.train_starvla][INFO] - Step 4438, Loss: {'train/loss': 0.3429277241230011, 'train/grad_norm_pre_clip': 12.564250946044922, 'timing/data': 0.001207517459988594, 'timing/model': 1.4288423508405685, 'learning_rate/qwen_vl_interface': 3.3814266812243473e-06, 'learning_rate/action_model': 3.3814266812243474e-05, 'epoch': 1.71}) +[2026-06-18 02:51:13,627][starVLA.training.train_starvla][INFO] - Step 4439, Loss: {'train/loss': 0.28575438261032104, 'train/grad_norm_pre_clip': 9.817266464233398, 'timing/data': 0.001104411669075489, 'timing/model': 1.7179019236937165, 'learning_rate/qwen_vl_interface': 3.3794385527753522e-06, 'learning_rate/action_model': 3.379438552775352e-05, 'epoch': 1.71}) +[2026-06-18 02:51:20,469][starVLA.training.train_starvla][INFO] - Step 4440, Loss: {'train/loss': 0.3356950879096985, 'train/grad_norm_pre_clip': 12.579094886779785, 'timing/data': 0.0011270511895418167, 'timing/model': 1.1143923243507743, 'learning_rate/qwen_vl_interface': 3.3774508120957606e-06, 'learning_rate/action_model': 3.37745081209576e-05, 'epoch': 1.71}) +[2026-06-18 02:51:25,687][starVLA.training.train_starvla][INFO] - Step 4441, Loss: {'train/loss': 0.34822362661361694, 'train/grad_norm_pre_clip': 15.258259773254395, 'timing/data': 0.0009235749021172523, 'timing/model': 1.1791523657739162, 'learning_rate/qwen_vl_interface': 3.37546345959763e-06, 'learning_rate/action_model': 3.3754634595976295e-05, 'epoch': 1.71}) +[2026-06-18 02:51:34,668][starVLA.training.train_starvla][INFO] - Step 4442, Loss: {'train/loss': 0.33794528245925903, 'train/grad_norm_pre_clip': 14.421723365783691, 'timing/data': 0.001308554783463478, 'timing/model': 1.2917591398581862, 'learning_rate/qwen_vl_interface': 3.3734764956929407e-06, 'learning_rate/action_model': 3.37347649569294e-05, 'epoch': 1.71}) +[2026-06-18 02:51:41,984][starVLA.training.train_starvla][INFO] - Step 4443, Loss: {'train/loss': 0.2560502588748932, 'train/grad_norm_pre_clip': 9.103255271911621, 'timing/data': 0.001002802513539791, 'timing/model': 1.2704729717224836, 'learning_rate/qwen_vl_interface': 3.3714899207935937e-06, 'learning_rate/action_model': 3.371489920793594e-05, 'epoch': 1.71}) +[2026-06-18 02:51:48,438][starVLA.training.train_starvla][INFO] - Step 4444, Loss: {'train/loss': 0.3290596008300781, 'train/grad_norm_pre_clip': 14.962329864501953, 'timing/data': 0.0011712107807397842, 'timing/model': 2.7423808500170708, 'learning_rate/qwen_vl_interface': 3.369503735311407e-06, 'learning_rate/action_model': 3.369503735311407e-05, 'epoch': 1.71}) +[2026-06-18 02:51:57,418][starVLA.training.train_starvla][INFO] - Step 4445, Loss: {'train/loss': 0.46889567375183105, 'train/grad_norm_pre_clip': 17.10359764099121, 'timing/data': 0.0014081783592700958, 'timing/model': 1.201859918422997, 'learning_rate/qwen_vl_interface': 3.3675179396581208e-06, 'learning_rate/action_model': 3.367517939658121e-05, 'epoch': 1.71}) +[2026-06-18 02:52:03,292][starVLA.training.train_starvla][INFO] - Step 4446, Loss: {'train/loss': 0.3022918999195099, 'train/grad_norm_pre_clip': 12.228389739990234, 'timing/data': 0.0036252550780773163, 'timing/model': 1.839528327807784, 'learning_rate/qwen_vl_interface': 3.365532534245389e-06, 'learning_rate/action_model': 3.3655325342453894e-05, 'epoch': 1.71}) +[2026-06-18 02:52:08,412][starVLA.training.train_starvla][INFO] - Step 4447, Loss: {'train/loss': 0.5770846009254456, 'train/grad_norm_pre_clip': 13.692261695861816, 'timing/data': 0.001223348081111908, 'timing/model': 1.6577644860371947, 'learning_rate/qwen_vl_interface': 3.3635475194847927e-06, 'learning_rate/action_model': 3.363547519484793e-05, 'epoch': 1.71}) +[2026-06-18 02:52:13,761][starVLA.training.train_starvla][INFO] - Step 4448, Loss: {'train/loss': 0.31225576996803284, 'train/grad_norm_pre_clip': 13.840991973876953, 'timing/data': 0.001157861202955246, 'timing/model': 1.3694008495658636, 'learning_rate/qwen_vl_interface': 3.361562895787825e-06, 'learning_rate/action_model': 3.361562895787825e-05, 'epoch': 1.71}) +[2026-06-18 02:52:18,586][starVLA.training.train_starvla][INFO] - Step 4449, Loss: {'train/loss': 0.4232332706451416, 'train/grad_norm_pre_clip': 15.593330383300781, 'timing/data': 0.0010971473529934883, 'timing/model': 1.102979519404471, 'learning_rate/qwen_vl_interface': 3.3595786635659e-06, 'learning_rate/action_model': 3.3595786635659e-05, 'epoch': 1.71}) +[2026-06-18 02:52:24,960][starVLA.training.train_starvla][INFO] - Step 4450, Loss: {'train/loss': 0.5526393055915833, 'train/grad_norm_pre_clip': 17.613540649414062, 'timing/data': 0.00096155796200037, 'timing/model': 1.1132743442431092, 'learning_rate/qwen_vl_interface': 3.357594823230353e-06, 'learning_rate/action_model': 3.3575948232303525e-05, 'epoch': 1.71}) +[2026-06-18 02:52:29,575][starVLA.training.train_starvla][INFO] - Step 4451, Loss: {'train/loss': 0.2325519472360611, 'train/grad_norm_pre_clip': 11.565940856933594, 'timing/data': 0.0011197412386536598, 'timing/model': 1.0794305875897408, 'learning_rate/qwen_vl_interface': 3.3556113751924357e-06, 'learning_rate/action_model': 3.355611375192436e-05, 'epoch': 1.71}) +[2026-06-18 02:52:34,078][starVLA.training.train_starvla][INFO] - Step 4452, Loss: {'train/loss': 0.20232774317264557, 'train/grad_norm_pre_clip': 13.34979248046875, 'timing/data': 0.0016501015052199364, 'timing/model': 1.149352259002626, 'learning_rate/qwen_vl_interface': 3.3536283198633183e-06, 'learning_rate/action_model': 3.353628319863318e-05, 'epoch': 1.71}) +[2026-06-18 02:52:38,845][starVLA.training.train_starvla][INFO] - Step 4453, Loss: {'train/loss': 0.23158042132854462, 'train/grad_norm_pre_clip': 9.743699073791504, 'timing/data': 0.0007706265896558762, 'timing/model': 1.131646305322647, 'learning_rate/qwen_vl_interface': 3.3516456576540913e-06, 'learning_rate/action_model': 3.3516456576540914e-05, 'epoch': 1.71}) +[2026-06-18 02:52:48,615][starVLA.training.train_starvla][INFO] - Step 4454, Loss: {'train/loss': 0.5098763704299927, 'train/grad_norm_pre_clip': 17.136077880859375, 'timing/data': 0.0013526389375329018, 'timing/model': 4.803258729167283, 'learning_rate/qwen_vl_interface': 3.349663388975763e-06, 'learning_rate/action_model': 3.3496633889757625e-05, 'epoch': 1.71}) +[2026-06-18 02:52:54,151][starVLA.training.train_starvla][INFO] - Step 4455, Loss: {'train/loss': 0.8701887726783752, 'train/grad_norm_pre_clip': 24.633047103881836, 'timing/data': 0.0011716876178979874, 'timing/model': 1.1303610280156136, 'learning_rate/qwen_vl_interface': 3.347681514239258e-06, 'learning_rate/action_model': 3.347681514239257e-05, 'epoch': 1.71}) +[2026-06-18 02:53:03,611][starVLA.training.train_starvla][INFO] - Step 4456, Loss: {'train/loss': 0.5206593871116638, 'train/grad_norm_pre_clip': 39.309234619140625, 'timing/data': 0.001299758441746235, 'timing/model': 2.9670404819771647, 'learning_rate/qwen_vl_interface': 3.345700033855421e-06, 'learning_rate/action_model': 3.345700033855421e-05, 'epoch': 1.71}) +[2026-06-18 02:53:10,809][starVLA.training.train_starvla][INFO] - Step 4457, Loss: {'train/loss': 0.4680262804031372, 'train/grad_norm_pre_clip': 14.463361740112305, 'timing/data': 0.0013477467000484467, 'timing/model': 3.11423200648278, 'learning_rate/qwen_vl_interface': 3.3437189482350173e-06, 'learning_rate/action_model': 3.343718948235017e-05, 'epoch': 1.71}) +[2026-06-18 02:53:19,283][starVLA.training.train_starvla][INFO] - Step 4458, Loss: {'train/loss': 0.3238580822944641, 'train/grad_norm_pre_clip': 13.517951965332031, 'timing/data': 0.0009442213922739029, 'timing/model': 1.1677712677046657, 'learning_rate/qwen_vl_interface': 3.341738257788727e-06, 'learning_rate/action_model': 3.3417382577887266e-05, 'epoch': 1.71}) +[2026-06-18 02:53:24,605][starVLA.training.train_starvla][INFO] - Step 4459, Loss: {'train/loss': 0.5187499523162842, 'train/grad_norm_pre_clip': 19.42909812927246, 'timing/data': 0.0011619022116065025, 'timing/model': 1.168776317499578, 'learning_rate/qwen_vl_interface': 3.3397579629271487e-06, 'learning_rate/action_model': 3.3397579629271486e-05, 'epoch': 1.72}) +[2026-06-18 02:53:29,888][starVLA.training.train_starvla][INFO] - Step 4460, Loss: {'train/loss': 0.4066229462623596, 'train/grad_norm_pre_clip': 14.166627883911133, 'timing/data': 0.0008360082283616066, 'timing/model': 1.1455061296001077, 'learning_rate/qwen_vl_interface': 3.3377780640607997e-06, 'learning_rate/action_model': 3.3377780640608e-05, 'epoch': 1.72}) +[2026-06-18 02:53:34,842][starVLA.training.train_starvla][INFO] - Step 4461, Loss: {'train/loss': 0.19836102426052094, 'train/grad_norm_pre_clip': 10.896470069885254, 'timing/data': 0.0009077135473489761, 'timing/model': 1.2023703921586275, 'learning_rate/qwen_vl_interface': 3.3357985616001166e-06, 'learning_rate/action_model': 3.335798561600116e-05, 'epoch': 1.72}) +[2026-06-18 02:53:39,367][starVLA.training.train_starvla][INFO] - Step 4462, Loss: {'train/loss': 0.5196685791015625, 'train/grad_norm_pre_clip': 13.344869613647461, 'timing/data': 0.0013801082968711853, 'timing/model': 1.1806205213069916, 'learning_rate/qwen_vl_interface': 3.3338194559554524e-06, 'learning_rate/action_model': 3.333819455955453e-05, 'epoch': 1.72}) +[2026-06-18 02:53:44,772][starVLA.training.train_starvla][INFO] - Step 4463, Loss: {'train/loss': 0.3479953408241272, 'train/grad_norm_pre_clip': 11.083213806152344, 'timing/data': 0.0017953859642148018, 'timing/model': 1.2231975682079792, 'learning_rate/qwen_vl_interface': 3.3318407475370758e-06, 'learning_rate/action_model': 3.3318407475370755e-05, 'epoch': 1.72}) +[2026-06-18 02:53:50,110][starVLA.training.train_starvla][INFO] - Step 4464, Loss: {'train/loss': 0.4342997968196869, 'train/grad_norm_pre_clip': 16.124547958374023, 'timing/data': 0.0008372198790311813, 'timing/model': 1.2312488732859492, 'learning_rate/qwen_vl_interface': 3.329862436755176e-06, 'learning_rate/action_model': 3.329862436755176e-05, 'epoch': 1.72}) +[2026-06-18 02:53:55,387][starVLA.training.train_starvla][INFO] - Step 4465, Loss: {'train/loss': 0.2868632376194, 'train/grad_norm_pre_clip': 16.715627670288086, 'timing/data': 0.0055750226601958275, 'timing/model': 1.17204665299505, 'learning_rate/qwen_vl_interface': 3.3278845240198592e-06, 'learning_rate/action_model': 3.3278845240198595e-05, 'epoch': 1.72}) +[2026-06-18 02:54:10,373][starVLA.training.train_starvla][INFO] - Step 4466, Loss: {'train/loss': 0.43989914655685425, 'train/grad_norm_pre_clip': 16.321231842041016, 'timing/data': 0.0011817533522844315, 'timing/model': 2.124406296759844, 'learning_rate/qwen_vl_interface': 3.325907009741149e-06, 'learning_rate/action_model': 3.3259070097411485e-05, 'epoch': 1.72}) +[2026-06-18 02:54:15,453][starVLA.training.train_starvla][INFO] - Step 4467, Loss: {'train/loss': 0.40759965777397156, 'train/grad_norm_pre_clip': 19.672500610351562, 'timing/data': 0.0012048697099089622, 'timing/model': 1.5686317086219788, 'learning_rate/qwen_vl_interface': 3.323929894328987e-06, 'learning_rate/action_model': 3.323929894328987e-05, 'epoch': 1.72}) +[2026-06-18 02:54:20,643][starVLA.training.train_starvla][INFO] - Step 4468, Loss: {'train/loss': 0.39178287982940674, 'train/grad_norm_pre_clip': 17.387907028198242, 'timing/data': 0.001252015121281147, 'timing/model': 1.0960128856822848, 'learning_rate/qwen_vl_interface': 3.321953178193228e-06, 'learning_rate/action_model': 3.321953178193228e-05, 'epoch': 1.72}) +[2026-06-18 02:54:25,347][starVLA.training.train_starvla][INFO] - Step 4469, Loss: {'train/loss': 0.5615190863609314, 'train/grad_norm_pre_clip': 30.05548095703125, 'timing/data': 0.0012713493779301643, 'timing/model': 1.0890029333531857, 'learning_rate/qwen_vl_interface': 3.319976861743653e-06, 'learning_rate/action_model': 3.319976861743653e-05, 'epoch': 1.72}) +[2026-06-18 02:54:33,892][starVLA.training.train_starvla][INFO] - Step 4470, Loss: {'train/loss': 0.2794414162635803, 'train/grad_norm_pre_clip': 11.878107070922852, 'timing/data': 0.0012379493564367294, 'timing/model': 1.3976318640634418, 'learning_rate/qwen_vl_interface': 3.3180009453899496e-06, 'learning_rate/action_model': 3.318000945389949e-05, 'epoch': 1.72}) +[2026-06-18 02:54:39,651][starVLA.training.train_starvla][INFO] - Step 4471, Loss: {'train/loss': 0.4969955086708069, 'train/grad_norm_pre_clip': 134.1100311279297, 'timing/data': 0.004889714531600475, 'timing/model': 1.1280740788206458, 'learning_rate/qwen_vl_interface': 3.3160254295417287e-06, 'learning_rate/action_model': 3.3160254295417285e-05, 'epoch': 1.72}) +[2026-06-18 02:54:44,477][starVLA.training.train_starvla][INFO] - Step 4472, Loss: {'train/loss': 0.2506141662597656, 'train/grad_norm_pre_clip': 12.676633834838867, 'timing/data': 0.0010528564453125, 'timing/model': 1.4749752189964056, 'learning_rate/qwen_vl_interface': 3.314050314608518e-06, 'learning_rate/action_model': 3.314050314608518e-05, 'epoch': 1.72}) +[2026-06-18 02:54:49,130][starVLA.training.train_starvla][INFO] - Step 4473, Loss: {'train/loss': 0.27317363023757935, 'train/grad_norm_pre_clip': 12.93619155883789, 'timing/data': 0.0012584459036588669, 'timing/model': 1.2106033591553569, 'learning_rate/qwen_vl_interface': 3.3120756009997606e-06, 'learning_rate/action_model': 3.312075600999761e-05, 'epoch': 1.72}) +[2026-06-18 02:54:55,837][starVLA.training.train_starvla][INFO] - Step 4474, Loss: {'train/loss': 0.46347445249557495, 'train/grad_norm_pre_clip': 35.05021667480469, 'timing/data': 0.001304936595261097, 'timing/model': 2.822966154664755, 'learning_rate/qwen_vl_interface': 3.310101289124817e-06, 'learning_rate/action_model': 3.310101289124817e-05, 'epoch': 1.72}) +[2026-06-18 02:55:01,966][starVLA.training.train_starvla][INFO] - Step 4475, Loss: {'train/loss': 0.22402521967887878, 'train/grad_norm_pre_clip': 12.102666854858398, 'timing/data': 0.0015713358297944069, 'timing/model': 2.216414278373122, 'learning_rate/qwen_vl_interface': 3.3081273793929615e-06, 'learning_rate/action_model': 3.308127379392961e-05, 'epoch': 1.72}) +[2026-06-18 02:55:10,375][starVLA.training.train_starvla][INFO] - Step 4476, Loss: {'train/loss': 0.30492886900901794, 'train/grad_norm_pre_clip': 13.691746711730957, 'timing/data': 0.0009504007175564766, 'timing/model': 2.815084840171039, 'learning_rate/qwen_vl_interface': 3.306153872213392e-06, 'learning_rate/action_model': 3.306153872213392e-05, 'epoch': 1.72}) +[2026-06-18 02:55:21,812][starVLA.training.train_starvla][INFO] - Step 4477, Loss: {'train/loss': 0.6551340818405151, 'train/grad_norm_pre_clip': 24.873394012451172, 'timing/data': 0.0012293225154280663, 'timing/model': 1.541633604094386, 'learning_rate/qwen_vl_interface': 3.304180767995215e-06, 'learning_rate/action_model': 3.3041807679952146e-05, 'epoch': 1.72}) +[2026-06-18 02:55:29,372][starVLA.training.train_starvla][INFO] - Step 4478, Loss: {'train/loss': 0.2791805565357208, 'train/grad_norm_pre_clip': 11.370869636535645, 'timing/data': 0.0011818921193480492, 'timing/model': 1.7995002921670675, 'learning_rate/qwen_vl_interface': 3.3022080671474588e-06, 'learning_rate/action_model': 3.302208067147459e-05, 'epoch': 1.72}) +[2026-06-18 02:55:38,826][starVLA.training.train_starvla][INFO] - Step 4479, Loss: {'train/loss': 0.4367160201072693, 'train/grad_norm_pre_clip': 16.516525268554688, 'timing/data': 0.0013906806707382202, 'timing/model': 1.9875554852187634, 'learning_rate/qwen_vl_interface': 3.300235770079066e-06, 'learning_rate/action_model': 3.300235770079066e-05, 'epoch': 1.72}) +[2026-06-18 02:55:52,821][starVLA.training.train_starvla][INFO] - Step 4480, Loss: {'train/loss': 0.1731596142053604, 'train/grad_norm_pre_clip': 10.707345962524414, 'timing/data': 0.0011710822582244873, 'timing/model': 3.1800034400075674, 'learning_rate/qwen_vl_interface': 3.298263877198896e-06, 'learning_rate/action_model': 3.298263877198896e-05, 'epoch': 1.72}) +[2026-06-18 02:56:05,385][starVLA.training.train_starvla][INFO] - Step 4481, Loss: {'train/loss': 0.23767079412937164, 'train/grad_norm_pre_clip': 14.71658706665039, 'timing/data': 0.0012811897322535515, 'timing/model': 1.8863710099831223, 'learning_rate/qwen_vl_interface': 3.296292388915726e-06, 'learning_rate/action_model': 3.296292388915726e-05, 'epoch': 1.72}) +[2026-06-18 02:56:12,025][starVLA.training.train_starvla][INFO] - Step 4482, Loss: {'train/loss': 0.245780348777771, 'train/grad_norm_pre_clip': 11.779619216918945, 'timing/data': 0.0009749326854944229, 'timing/model': 1.496410382911563, 'learning_rate/qwen_vl_interface': 3.294321305638243e-06, 'learning_rate/action_model': 3.294321305638243e-05, 'epoch': 1.72}) +[2026-06-18 02:56:16,892][starVLA.training.train_starvla][INFO] - Step 4483, Loss: {'train/loss': 0.22206078469753265, 'train/grad_norm_pre_clip': 8.989021301269531, 'timing/data': 0.0014440808445215225, 'timing/model': 1.510273864492774, 'learning_rate/qwen_vl_interface': 3.292350627775061e-06, 'learning_rate/action_model': 3.292350627775061e-05, 'epoch': 1.72}) +[2026-06-18 02:56:21,868][starVLA.training.train_starvla][INFO] - Step 4484, Loss: {'train/loss': 0.2794683575630188, 'train/grad_norm_pre_clip': 12.347005844116211, 'timing/data': 0.0009664855897426605, 'timing/model': 1.1281624604016542, 'learning_rate/qwen_vl_interface': 3.2903803557346997e-06, 'learning_rate/action_model': 3.2903803557347e-05, 'epoch': 1.72}) +[2026-06-18 02:56:27,662][starVLA.training.train_starvla][INFO] - Step 4485, Loss: {'train/loss': 0.6032154560089111, 'train/grad_norm_pre_clip': 18.777700424194336, 'timing/data': 0.0011343695223331451, 'timing/model': 1.3774025766178966, 'learning_rate/qwen_vl_interface': 3.2884104899255986e-06, 'learning_rate/action_model': 3.288410489925599e-05, 'epoch': 1.73}) +[2026-06-18 02:56:33,060][starVLA.training.train_starvla][INFO] - Step 4486, Loss: {'train/loss': 0.41856664419174194, 'train/grad_norm_pre_clip': 16.2469425201416, 'timing/data': 0.01130232959985733, 'timing/model': 1.1464153155684471, 'learning_rate/qwen_vl_interface': 3.2864410307561124e-06, 'learning_rate/action_model': 3.286441030756112e-05, 'epoch': 1.73}) +[2026-06-18 02:56:38,185][starVLA.training.train_starvla][INFO] - Step 4487, Loss: {'train/loss': 0.41398337483406067, 'train/grad_norm_pre_clip': 18.026119232177734, 'timing/data': 0.0011017275974154472, 'timing/model': 1.2117808973416686, 'learning_rate/qwen_vl_interface': 3.284471978634515e-06, 'learning_rate/action_model': 3.284471978634515e-05, 'epoch': 1.73}) +[2026-06-18 02:56:44,227][starVLA.training.train_starvla][INFO] - Step 4488, Loss: {'train/loss': 0.3938671052455902, 'train/grad_norm_pre_clip': 17.439355850219727, 'timing/data': 0.005444744601845741, 'timing/model': 1.417069697752595, 'learning_rate/qwen_vl_interface': 3.2825033339689926e-06, 'learning_rate/action_model': 3.282503333968993e-05, 'epoch': 1.73}) +[2026-06-18 02:56:48,975][starVLA.training.train_starvla][INFO] - Step 4489, Loss: {'train/loss': 0.3042871952056885, 'train/grad_norm_pre_clip': 10.634862899780273, 'timing/data': 0.00114425178617239, 'timing/model': 1.0845181392505765, 'learning_rate/qwen_vl_interface': 3.2805350971676447e-06, 'learning_rate/action_model': 3.280535097167645e-05, 'epoch': 1.73}) +[2026-06-18 02:56:53,615][starVLA.training.train_starvla][INFO] - Step 4490, Loss: {'train/loss': 0.5163663625717163, 'train/grad_norm_pre_clip': 14.335891723632812, 'timing/data': 0.0024154214188456535, 'timing/model': 1.1489644143730402, 'learning_rate/qwen_vl_interface': 3.2785672686384893e-06, 'learning_rate/action_model': 3.278567268638489e-05, 'epoch': 1.73}) +[2026-06-18 02:56:58,387][starVLA.training.train_starvla][INFO] - Step 4491, Loss: {'train/loss': 0.6087082624435425, 'train/grad_norm_pre_clip': 13.902457237243652, 'timing/data': 0.0018686838448047638, 'timing/model': 1.1403114283457398, 'learning_rate/qwen_vl_interface': 3.2765998487894622e-06, 'learning_rate/action_model': 3.276599848789462e-05, 'epoch': 1.73}) +[2026-06-18 02:57:03,585][starVLA.training.train_starvla][INFO] - Step 4492, Loss: {'train/loss': 0.36283645033836365, 'train/grad_norm_pre_clip': 26.190914154052734, 'timing/data': 0.001420716755092144, 'timing/model': 1.537034248933196, 'learning_rate/qwen_vl_interface': 3.2746328380284098e-06, 'learning_rate/action_model': 3.27463283802841e-05, 'epoch': 1.73}) +[2026-06-18 02:57:08,300][starVLA.training.train_starvla][INFO] - Step 4493, Loss: {'train/loss': 0.5140168070793152, 'train/grad_norm_pre_clip': 20.249082565307617, 'timing/data': 0.0009469464421272278, 'timing/model': 1.1674230266362429, 'learning_rate/qwen_vl_interface': 3.2726662367630953e-06, 'learning_rate/action_model': 3.272666236763095e-05, 'epoch': 1.73}) +[2026-06-18 02:57:13,867][starVLA.training.train_starvla][INFO] - Step 4494, Loss: {'train/loss': 0.3995967507362366, 'train/grad_norm_pre_clip': 12.981379508972168, 'timing/data': 0.0010483646765351295, 'timing/model': 1.3778577456250787, 'learning_rate/qwen_vl_interface': 3.270700045401197e-06, 'learning_rate/action_model': 3.2707000454011974e-05, 'epoch': 1.73}) +[2026-06-18 02:57:19,368][starVLA.training.train_starvla][INFO] - Step 4495, Loss: {'train/loss': 0.6047677993774414, 'train/grad_norm_pre_clip': 16.74197769165039, 'timing/data': 0.0010811872780323029, 'timing/model': 1.4014470065012574, 'learning_rate/qwen_vl_interface': 3.26873426435031e-06, 'learning_rate/action_model': 3.2687342643503096e-05, 'epoch': 1.73}) +[2026-06-18 02:57:30,022][starVLA.training.train_starvla][INFO] - Step 4496, Loss: {'train/loss': 0.4352445900440216, 'train/grad_norm_pre_clip': 16.745607376098633, 'timing/data': 0.0009776921942830086, 'timing/model': 4.646939723752439, 'learning_rate/qwen_vl_interface': 3.266768894017942e-06, 'learning_rate/action_model': 3.266768894017942e-05, 'epoch': 1.73}) +[2026-06-18 02:57:44,207][starVLA.training.train_starvla][INFO] - Step 4497, Loss: {'train/loss': 0.5147919654846191, 'train/grad_norm_pre_clip': 54.46244812011719, 'timing/data': 0.012525933794677258, 'timing/model': 1.849498494528234, 'learning_rate/qwen_vl_interface': 3.264803934811514e-06, 'learning_rate/action_model': 3.264803934811514e-05, 'epoch': 1.73}) +[2026-06-18 02:57:56,492][starVLA.training.train_starvla][INFO] - Step 4498, Loss: {'train/loss': 0.23284147679805756, 'train/grad_norm_pre_clip': 11.37957763671875, 'timing/data': 0.044957433827221394, 'timing/model': 1.2682369221001863, 'learning_rate/qwen_vl_interface': 3.2628393871383706e-06, 'learning_rate/action_model': 3.26283938713837e-05, 'epoch': 1.73}) +[2026-06-18 02:58:12,060][starVLA.training.train_starvla][INFO] - Step 4499, Loss: {'train/loss': 0.42840489745140076, 'train/grad_norm_pre_clip': 15.329874992370605, 'timing/data': 0.013515438884496689, 'timing/model': 4.75918105058372, 'learning_rate/qwen_vl_interface': 3.2608752514057585e-06, 'learning_rate/action_model': 3.260875251405758e-05, 'epoch': 1.73}) +[2026-06-18 03:03:16,992][starVLA.training.train_starvla][INFO] - Step 4500, Loss: {'train/loss': 0.29308998584747314, 'train/grad_norm_pre_clip': 12.882800102233887, 'eval/loss': 0.4447743892669678, 'eval/latency_4/loss': 0.4447743892669678, 'eval/demon_fire_precision': 0.9726473461413221, 'eval/demon_fire_recall': 0.9701201688860019, 'eval/demon_fire_f1': 0.9713821138211383, 'eval/demon_fire_rate_model': 0.9019089574155653, 'eval/demon_fire_rate_teacher': 0.9042584434654919, 'eval/demon_left_precision': 0.7244655581947743, 'eval/demon_left_recall': 0.6963470319634704, 'eval/demon_left_f1': 0.710128055878929, 'eval/demon_left_rate_model': 0.1236417033773862, 'eval/demon_left_rate_teacher': 0.12863436123348018, 'eval/demon_right_precision': 0.8063241106719368, 'eval/demon_right_recall': 0.7364620938628159, 'eval/demon_right_f1': 0.769811320754717, 'eval/demon_right_rate_model': 0.1486049926578561, 'eval/demon_right_rate_teacher': 0.16270190895741557, 'eval/demon_move_f1': 0.739969688316823, 'eval/demon_cc_f1': 0.8880642044030144, 'eval/latency_4/demon_fire_precision': 0.9726473461413221, 'eval/latency_4/demon_fire_recall': 0.9701201688860019, 'eval/latency_4/demon_fire_f1': 0.9713821138211383, 'eval/latency_4/demon_fire_rate_model': 0.9019089574155653, 'eval/latency_4/demon_fire_rate_teacher': 0.9042584434654919, 'eval/latency_4/demon_left_precision': 0.7244655581947743, 'eval/latency_4/demon_left_recall': 0.6963470319634704, 'eval/latency_4/demon_left_f1': 0.710128055878929, 'eval/latency_4/demon_left_rate_model': 0.1236417033773862, 'eval/latency_4/demon_left_rate_teacher': 0.12863436123348018, 'eval/latency_4/demon_right_precision': 0.8063241106719368, 'eval/latency_4/demon_right_recall': 0.7364620938628159, 'eval/latency_4/demon_right_f1': 0.769811320754717, 'eval/latency_4/demon_right_rate_model': 0.1486049926578561, 'eval/latency_4/demon_right_rate_teacher': 0.16270190895741557, 'eval/latency_4/demon_move_f1': 0.739969688316823, 'eval/latency_4/demon_cc_f1': 0.8880642044030144, 'timing/data': 0.020498458296060562, 'timing/model': 1.1854847613722086, 'learning_rate/qwen_vl_interface': 3.2589115280208473e-06, 'learning_rate/action_model': 3.258911528020847e-05, 'epoch': 1.73}) +[2026-06-18 03:03:17,005][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_4500_state +[2026-06-18 03:03:17,018][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 03:03:26,401][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_4500_state/model.safetensors +[2026-06-18 03:03:50,436][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_4500_state/optimizer.bin +[2026-06-18 03:03:50,437][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_4500_state/sampler.bin +[2026-06-18 03:03:50,437][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_4500_state/sampler_1.bin +[2026-06-18 03:03:50,440][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_4500_state/random_states_0.pkl +[2026-06-18 03:03:50,445][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4500 +[2026-06-18 03:03:50,446][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 03:03:50,463][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 03:03:55,533][starVLA.training.train_starvla][INFO] - Step 4501, Loss: {'train/loss': 0.37298378348350525, 'train/grad_norm_pre_clip': 13.679350852966309, 'timing/data': 0.012699874117970467, 'timing/model': 1.5672816187143326, 'learning_rate/qwen_vl_interface': 3.2569482173907164e-06, 'learning_rate/action_model': 3.2569482173907165e-05, 'epoch': 1.73}) +[2026-06-18 03:04:03,458][starVLA.training.train_starvla][INFO] - Step 4502, Loss: {'train/loss': 0.49531349539756775, 'train/grad_norm_pre_clip': 18.214008331298828, 'timing/data': 0.011727258563041687, 'timing/model': 1.10205851867795, 'learning_rate/qwen_vl_interface': 3.2549853199223668e-06, 'learning_rate/action_model': 3.254985319922367e-05, 'epoch': 1.73}) +[2026-06-18 03:04:17,392][starVLA.training.train_starvla][INFO] - Step 4503, Loss: {'train/loss': 0.3783119022846222, 'train/grad_norm_pre_clip': 16.380754470825195, 'timing/data': 0.01880633272230625, 'timing/model': 4.571199770085514, 'learning_rate/qwen_vl_interface': 3.253022836022708e-06, 'learning_rate/action_model': 3.253022836022708e-05, 'epoch': 1.73}) +[2026-06-18 03:04:22,972][starVLA.training.train_starvla][INFO] - Step 4504, Loss: {'train/loss': 0.4872857630252838, 'train/grad_norm_pre_clip': 14.746140480041504, 'timing/data': 0.015251961536705494, 'timing/model': 1.1283489894121885, 'learning_rate/qwen_vl_interface': 3.2510607660985627e-06, 'learning_rate/action_model': 3.251060766098563e-05, 'epoch': 1.73}) +[2026-06-18 03:04:29,893][starVLA.training.train_starvla][INFO] - Step 4505, Loss: {'train/loss': 0.3780719041824341, 'train/grad_norm_pre_clip': 14.512731552124023, 'timing/data': 0.014952396042644978, 'timing/model': 1.5580955669283867, 'learning_rate/qwen_vl_interface': 3.249099110556672e-06, 'learning_rate/action_model': 3.249099110556672e-05, 'epoch': 1.73}) +[2026-06-18 03:04:36,820][starVLA.training.train_starvla][INFO] - Step 4506, Loss: {'train/loss': 0.39433056116104126, 'train/grad_norm_pre_clip': 11.654585838317871, 'timing/data': 0.01556910015642643, 'timing/model': 1.102749546058476, 'learning_rate/qwen_vl_interface': 3.247137869803689e-06, 'learning_rate/action_model': 3.2471378698036886e-05, 'epoch': 1.73}) +[2026-06-18 03:04:41,573][starVLA.training.train_starvla][INFO] - Step 4507, Loss: {'train/loss': 0.29476070404052734, 'train/grad_norm_pre_clip': 13.08675765991211, 'timing/data': 0.012934493832290173, 'timing/model': 1.4370591510087252, 'learning_rate/qwen_vl_interface': 3.2451770442461813e-06, 'learning_rate/action_model': 3.245177044246181e-05, 'epoch': 1.73}) +[2026-06-18 03:04:47,021][starVLA.training.train_starvla][INFO] - Step 4508, Loss: {'train/loss': 0.5134699940681458, 'train/grad_norm_pre_clip': 13.513201713562012, 'timing/data': 0.013013486750423908, 'timing/model': 1.3457439569756389, 'learning_rate/qwen_vl_interface': 3.2432166342906286e-06, 'learning_rate/action_model': 3.2432166342906284e-05, 'epoch': 1.73}) +[2026-06-18 03:04:52,487][starVLA.training.train_starvla][INFO] - Step 4509, Loss: {'train/loss': 0.3017563223838806, 'train/grad_norm_pre_clip': 15.356661796569824, 'timing/data': 0.0299685625359416, 'timing/model': 1.5115430308505893, 'learning_rate/qwen_vl_interface': 3.241256640343427e-06, 'learning_rate/action_model': 3.241256640343427e-05, 'epoch': 1.73}) +[2026-06-18 03:04:58,223][starVLA.training.train_starvla][INFO] - Step 4510, Loss: {'train/loss': 0.205255925655365, 'train/grad_norm_pre_clip': 10.921296119689941, 'timing/data': 0.015915341675281525, 'timing/model': 1.230539602227509, 'learning_rate/qwen_vl_interface': 3.239297062810886e-06, 'learning_rate/action_model': 3.239297062810886e-05, 'epoch': 1.73}) +[2026-06-18 03:05:03,050][starVLA.training.train_starvla][INFO] - Step 4511, Loss: {'train/loss': 0.24332638084888458, 'train/grad_norm_pre_clip': 9.686084747314453, 'timing/data': 0.012816541828215122, 'timing/model': 1.1206656862050295, 'learning_rate/qwen_vl_interface': 3.2373379020992252e-06, 'learning_rate/action_model': 3.237337902099225e-05, 'epoch': 1.74}) +[2026-06-18 03:05:08,107][starVLA.training.train_starvla][INFO] - Step 4512, Loss: {'train/loss': 0.6005585789680481, 'train/grad_norm_pre_clip': 16.342485427856445, 'timing/data': 0.013393648900091648, 'timing/model': 1.5245728651061654, 'learning_rate/qwen_vl_interface': 3.2353791586145855e-06, 'learning_rate/action_model': 3.2353791586145855e-05, 'epoch': 1.74}) +[2026-06-18 03:05:12,467][starVLA.training.train_starvla][INFO] - Step 4513, Loss: {'train/loss': 0.31606823205947876, 'train/grad_norm_pre_clip': 12.268173217773438, 'timing/data': 0.012025835923850536, 'timing/model': 1.08239352889359, 'learning_rate/qwen_vl_interface': 3.2334208327630158e-06, 'learning_rate/action_model': 3.233420832763016e-05, 'epoch': 1.74}) +[2026-06-18 03:05:19,209][starVLA.training.train_starvla][INFO] - Step 4514, Loss: {'train/loss': 0.1522836834192276, 'train/grad_norm_pre_clip': 7.945253372192383, 'timing/data': 0.012369493022561073, 'timing/model': 1.1225753715261817, 'learning_rate/qwen_vl_interface': 3.231462924950475e-06, 'learning_rate/action_model': 3.2314629249504746e-05, 'epoch': 1.74}) +[2026-06-18 03:05:24,246][starVLA.training.train_starvla][INFO] - Step 4515, Loss: {'train/loss': 0.25735217332839966, 'train/grad_norm_pre_clip': 9.63077163696289, 'timing/data': 0.01774110086262226, 'timing/model': 1.2288655461743474, 'learning_rate/qwen_vl_interface': 3.2295054355828396e-06, 'learning_rate/action_model': 3.22950543558284e-05, 'epoch': 1.74}) +[2026-06-18 03:05:29,618][starVLA.training.train_starvla][INFO] - Step 4516, Loss: {'train/loss': 0.28387412428855896, 'train/grad_norm_pre_clip': 9.66385555267334, 'timing/data': 0.012865321710705757, 'timing/model': 1.1056890916079283, 'learning_rate/qwen_vl_interface': 3.227548365065905e-06, 'learning_rate/action_model': 3.227548365065905e-05, 'epoch': 1.74}) +[2026-06-18 03:05:34,578][starVLA.training.train_starvla][INFO] - Step 4517, Loss: {'train/loss': 0.15915654599666595, 'train/grad_norm_pre_clip': 7.569766044616699, 'timing/data': 0.012252981774508953, 'timing/model': 1.4386271191760898, 'learning_rate/qwen_vl_interface': 3.2255917138053695e-06, 'learning_rate/action_model': 3.2255917138053696e-05, 'epoch': 1.74}) +[2026-06-18 03:05:39,236][starVLA.training.train_starvla][INFO] - Step 4518, Loss: {'train/loss': 0.41234660148620605, 'train/grad_norm_pre_clip': 11.683815002441406, 'timing/data': 0.027729256078600883, 'timing/model': 1.129296414554119, 'learning_rate/qwen_vl_interface': 3.2236354822068515e-06, 'learning_rate/action_model': 3.223635482206852e-05, 'epoch': 1.74}) +[2026-06-18 03:05:49,099][starVLA.training.train_starvla][INFO] - Step 4519, Loss: {'train/loss': 0.4541010558605194, 'train/grad_norm_pre_clip': 15.964842796325684, 'timing/data': 0.013214020989835262, 'timing/model': 2.694515441544354, 'learning_rate/qwen_vl_interface': 3.2216796706758757e-06, 'learning_rate/action_model': 3.221679670675876e-05, 'epoch': 1.74}) +[2026-06-18 03:05:55,564][starVLA.training.train_starvla][INFO] - Step 4520, Loss: {'train/loss': 0.30773699283599854, 'train/grad_norm_pre_clip': 8.84874153137207, 'timing/data': 0.014364323578774929, 'timing/model': 1.113138527609408, 'learning_rate/qwen_vl_interface': 3.219724279617888e-06, 'learning_rate/action_model': 3.219724279617888e-05, 'epoch': 1.74}) +[2026-06-18 03:06:01,079][starVLA.training.train_starvla][INFO] - Step 4521, Loss: {'train/loss': 0.21230189502239227, 'train/grad_norm_pre_clip': 9.982270240783691, 'timing/data': 0.015256266109645367, 'timing/model': 1.374375798739493, 'learning_rate/qwen_vl_interface': 3.2177693094382424e-06, 'learning_rate/action_model': 3.2177693094382424e-05, 'epoch': 1.74}) +[2026-06-18 03:06:05,685][starVLA.training.train_starvla][INFO] - Step 4522, Loss: {'train/loss': 0.33390986919403076, 'train/grad_norm_pre_clip': 14.563074111938477, 'timing/data': 0.017334754578769207, 'timing/model': 1.2697723926976323, 'learning_rate/qwen_vl_interface': 3.2158147605422052e-06, 'learning_rate/action_model': 3.215814760542205e-05, 'epoch': 1.74}) +[2026-06-18 03:06:12,613][starVLA.training.train_starvla][INFO] - Step 4523, Loss: {'train/loss': 0.27002784609794617, 'train/grad_norm_pre_clip': 14.062122344970703, 'timing/data': 0.013807563111186028, 'timing/model': 1.319792790338397, 'learning_rate/qwen_vl_interface': 3.2138606333349576e-06, 'learning_rate/action_model': 3.213860633334957e-05, 'epoch': 1.74}) +[2026-06-18 03:06:17,586][starVLA.training.train_starvla][INFO] - Step 4524, Loss: {'train/loss': 0.22641979157924652, 'train/grad_norm_pre_clip': 35.491172790527344, 'timing/data': 0.012111406773328781, 'timing/model': 1.099211779423058, 'learning_rate/qwen_vl_interface': 3.2119069282215913e-06, 'learning_rate/action_model': 3.211906928221591e-05, 'epoch': 1.74}) +[2026-06-18 03:06:25,294][starVLA.training.train_starvla][INFO] - Step 4525, Loss: {'train/loss': 0.2505025565624237, 'train/grad_norm_pre_clip': 9.306183815002441, 'timing/data': 0.014562459662556648, 'timing/model': 1.732423922047019, 'learning_rate/qwen_vl_interface': 3.2099536456071118e-06, 'learning_rate/action_model': 3.209953645607112e-05, 'epoch': 1.74}) +[2026-06-18 03:06:42,396][starVLA.training.train_starvla][INFO] - Step 4526, Loss: {'train/loss': 0.40441596508026123, 'train/grad_norm_pre_clip': 16.7475643157959, 'timing/data': 0.023623532615602016, 'timing/model': 1.6708866283297539, 'learning_rate/qwen_vl_interface': 3.2080007858964347e-06, 'learning_rate/action_model': 3.2080007858964345e-05, 'epoch': 1.74}) +[2026-06-18 03:06:47,562][starVLA.training.train_starvla][INFO] - Step 4527, Loss: {'train/loss': 0.2993411421775818, 'train/grad_norm_pre_clip': 14.675687789916992, 'timing/data': 0.014225633814930916, 'timing/model': 1.5118545228615403, 'learning_rate/qwen_vl_interface': 3.206048349494396e-06, 'learning_rate/action_model': 3.206048349494396e-05, 'epoch': 1.74}) +[2026-06-18 03:06:52,596][starVLA.training.train_starvla][INFO] - Step 4528, Loss: {'train/loss': 0.49445411562919617, 'train/grad_norm_pre_clip': 15.948196411132812, 'timing/data': 0.012798555195331573, 'timing/model': 1.4021154511719942, 'learning_rate/qwen_vl_interface': 3.2040963368057323e-06, 'learning_rate/action_model': 3.204096336805732e-05, 'epoch': 1.74}) +[2026-06-18 03:06:59,977][starVLA.training.train_starvla][INFO] - Step 4529, Loss: {'train/loss': 0.5067497491836548, 'train/grad_norm_pre_clip': 14.832200050354004, 'timing/data': 0.01703036203980446, 'timing/model': 1.1552219679579139, 'learning_rate/qwen_vl_interface': 3.2021447482350994e-06, 'learning_rate/action_model': 3.202144748235099e-05, 'epoch': 1.74}) +[2026-06-18 03:07:05,816][starVLA.training.train_starvla][INFO] - Step 4530, Loss: {'train/loss': 0.21674075722694397, 'train/grad_norm_pre_clip': 9.858277320861816, 'timing/data': 0.016898807138204575, 'timing/model': 1.2485450757667422, 'learning_rate/qwen_vl_interface': 3.200193584187061e-06, 'learning_rate/action_model': 3.200193584187061e-05, 'epoch': 1.74}) +[2026-06-18 03:07:15,054][starVLA.training.train_starvla][INFO] - Step 4531, Loss: {'train/loss': 0.2616537809371948, 'train/grad_norm_pre_clip': 14.351043701171875, 'timing/data': 0.015733081847429276, 'timing/model': 1.465642036870122, 'learning_rate/qwen_vl_interface': 3.198242845066101e-06, 'learning_rate/action_model': 3.198242845066101e-05, 'epoch': 1.74}) +[2026-06-18 03:07:20,291][starVLA.training.train_starvla][INFO] - Step 4532, Loss: {'train/loss': 0.35177451372146606, 'train/grad_norm_pre_clip': 10.144686698913574, 'timing/data': 0.020733465440571308, 'timing/model': 1.752462049946189, 'learning_rate/qwen_vl_interface': 3.1962925312766087e-06, 'learning_rate/action_model': 3.196292531276608e-05, 'epoch': 1.74}) +[2026-06-18 03:07:25,015][starVLA.training.train_starvla][INFO] - Step 4533, Loss: {'train/loss': 0.5625255107879639, 'train/grad_norm_pre_clip': 16.268274307250977, 'timing/data': 0.012342888861894608, 'timing/model': 1.2134761875495315, 'learning_rate/qwen_vl_interface': 3.19434264322288e-06, 'learning_rate/action_model': 3.19434264322288e-05, 'epoch': 1.74}) +[2026-06-18 03:07:30,761][starVLA.training.train_starvla][INFO] - Step 4534, Loss: {'train/loss': 0.29098111391067505, 'train/grad_norm_pre_clip': 12.862119674682617, 'timing/data': 0.014089034870266914, 'timing/model': 1.3850477589294314, 'learning_rate/qwen_vl_interface': 3.1923931813091364e-06, 'learning_rate/action_model': 3.192393181309136e-05, 'epoch': 1.74}) +[2026-06-18 03:07:36,020][starVLA.training.train_starvla][INFO] - Step 4535, Loss: {'train/loss': 0.43659040331840515, 'train/grad_norm_pre_clip': 15.135214805603027, 'timing/data': 0.012793324887752533, 'timing/model': 1.2483387840911746, 'learning_rate/qwen_vl_interface': 3.1904441459395e-06, 'learning_rate/action_model': 3.1904441459395e-05, 'epoch': 1.74}) +[2026-06-18 03:07:44,840][starVLA.training.train_starvla][INFO] - Step 4536, Loss: {'train/loss': 0.5022329092025757, 'train/grad_norm_pre_clip': 19.828693389892578, 'timing/data': 0.01950477436184883, 'timing/model': 1.186714831739664, 'learning_rate/qwen_vl_interface': 3.1884955375180088e-06, 'learning_rate/action_model': 3.1884955375180085e-05, 'epoch': 1.74}) +[2026-06-18 03:07:52,276][starVLA.training.train_starvla][INFO] - Step 4537, Loss: {'train/loss': 0.24614079296588898, 'train/grad_norm_pre_clip': 11.847188949584961, 'timing/data': 0.015137613750994205, 'timing/model': 1.993773122318089, 'learning_rate/qwen_vl_interface': 3.18654735644861e-06, 'learning_rate/action_model': 3.1865473564486095e-05, 'epoch': 1.75}) +[2026-06-18 03:07:57,689][starVLA.training.train_starvla][INFO] - Step 4538, Loss: {'train/loss': 0.3216710090637207, 'train/grad_norm_pre_clip': 13.04842472076416, 'timing/data': 0.019267743453383446, 'timing/model': 1.1557527147233486, 'learning_rate/qwen_vl_interface': 3.184599603135165e-06, 'learning_rate/action_model': 3.1845996031351646e-05, 'epoch': 1.75}) +[2026-06-18 03:08:07,992][starVLA.training.train_starvla][INFO] - Step 4539, Loss: {'train/loss': 0.3503764569759369, 'train/grad_norm_pre_clip': 12.120405197143555, 'timing/data': 0.06980162300169468, 'timing/model': 1.546693455427885, 'learning_rate/qwen_vl_interface': 3.1826522779814453e-06, 'learning_rate/action_model': 3.182652277981445e-05, 'epoch': 1.75}) +[2026-06-18 03:08:13,537][starVLA.training.train_starvla][INFO] - Step 4540, Loss: {'train/loss': 0.474550724029541, 'train/grad_norm_pre_clip': 18.301380157470703, 'timing/data': 0.011995600536465645, 'timing/model': 1.296648120507598, 'learning_rate/qwen_vl_interface': 3.1807053813911317e-06, 'learning_rate/action_model': 3.180705381391132e-05, 'epoch': 1.75}) +[2026-06-18 03:08:19,008][starVLA.training.train_starvla][INFO] - Step 4541, Loss: {'train/loss': 0.28299716114997864, 'train/grad_norm_pre_clip': 12.678793907165527, 'timing/data': 0.06255657225847244, 'timing/model': 1.1495612720027566, 'learning_rate/qwen_vl_interface': 3.17875891376782e-06, 'learning_rate/action_model': 3.1787589137678196e-05, 'epoch': 1.75}) +[2026-06-18 03:08:24,576][starVLA.training.train_starvla][INFO] - Step 4542, Loss: {'train/loss': 0.36996594071388245, 'train/grad_norm_pre_clip': 14.34249496459961, 'timing/data': 0.06244906596839428, 'timing/model': 1.545708023943007, 'learning_rate/qwen_vl_interface': 3.176812875515014e-06, 'learning_rate/action_model': 3.176812875515014e-05, 'epoch': 1.75}) +[2026-06-18 03:08:36,245][starVLA.training.train_starvla][INFO] - Step 4543, Loss: {'train/loss': 0.3401455581188202, 'train/grad_norm_pre_clip': 14.18836498260498, 'timing/data': 0.04796671308577061, 'timing/model': 5.012262420728803, 'learning_rate/qwen_vl_interface': 3.1748672670361304e-06, 'learning_rate/action_model': 3.17486726703613e-05, 'epoch': 1.75}) +[2026-06-18 03:08:46,278][starVLA.training.train_starvla][INFO] - Step 4544, Loss: {'train/loss': 0.3400926887989044, 'train/grad_norm_pre_clip': 11.110426902770996, 'timing/data': 0.04478182829916477, 'timing/model': 1.4914139350876212, 'learning_rate/qwen_vl_interface': 3.172922088734493e-06, 'learning_rate/action_model': 3.172922088734493e-05, 'epoch': 1.75}) +[2026-06-18 03:08:51,576][starVLA.training.train_starvla][INFO] - Step 4545, Loss: {'train/loss': 0.38563448190689087, 'train/grad_norm_pre_clip': 16.8042049407959, 'timing/data': 0.0563991516828537, 'timing/model': 1.134105384349823, 'learning_rate/qwen_vl_interface': 3.1709773410133434e-06, 'learning_rate/action_model': 3.170977341013343e-05, 'epoch': 1.75}) +[2026-06-18 03:09:01,425][starVLA.training.train_starvla][INFO] - Step 4546, Loss: {'train/loss': 0.19892549514770508, 'train/grad_norm_pre_clip': 9.563159942626953, 'timing/data': 0.04950067400932312, 'timing/model': 5.54733706638217, 'learning_rate/qwen_vl_interface': 3.16903302427583e-06, 'learning_rate/action_model': 3.16903302427583e-05, 'epoch': 1.75}) +[2026-06-18 03:09:07,221][starVLA.training.train_starvla][INFO] - Step 4547, Loss: {'train/loss': 0.2177288830280304, 'train/grad_norm_pre_clip': 11.88082218170166, 'timing/data': 0.06489833444356918, 'timing/model': 1.9322162326425314, 'learning_rate/qwen_vl_interface': 3.1670891389250073e-06, 'learning_rate/action_model': 3.167089138925007e-05, 'epoch': 1.75}) +[2026-06-18 03:09:12,810][starVLA.training.train_starvla][INFO] - Step 4548, Loss: {'train/loss': 0.4414444863796234, 'train/grad_norm_pre_clip': 14.5931396484375, 'timing/data': 0.056200526654720306, 'timing/model': 1.6174191171303391, 'learning_rate/qwen_vl_interface': 3.1651456853638464e-06, 'learning_rate/action_model': 3.165145685363846e-05, 'epoch': 1.75}) +[2026-06-18 03:09:25,951][starVLA.training.train_starvla][INFO] - Step 4549, Loss: {'train/loss': 0.29042479395866394, 'train/grad_norm_pre_clip': 13.655694007873535, 'timing/data': 0.05119964387267828, 'timing/model': 6.039434289559722, 'learning_rate/qwen_vl_interface': 3.1632026639952303e-06, 'learning_rate/action_model': 3.16320266399523e-05, 'epoch': 1.75}) +[2026-06-18 03:09:31,318][starVLA.training.train_starvla][INFO] - Step 4550, Loss: {'train/loss': 0.5144172310829163, 'train/grad_norm_pre_clip': 19.774484634399414, 'timing/data': 0.014916260726749897, 'timing/model': 1.167323819361627, 'learning_rate/qwen_vl_interface': 3.161260075221947e-06, 'learning_rate/action_model': 3.161260075221947e-05, 'epoch': 1.75}) +[2026-06-18 03:09:40,729][starVLA.training.train_starvla][INFO] - Step 4551, Loss: {'train/loss': 0.2883758246898651, 'train/grad_norm_pre_clip': 11.766863822937012, 'timing/data': 0.035115561448037624, 'timing/model': 1.407790551893413, 'learning_rate/qwen_vl_interface': 3.1593179194467e-06, 'learning_rate/action_model': 3.1593179194467e-05, 'epoch': 1.75}) +[2026-06-18 03:09:45,686][starVLA.training.train_starvla][INFO] - Step 4552, Loss: {'train/loss': 0.2978596091270447, 'train/grad_norm_pre_clip': 13.35098648071289, 'timing/data': 0.015244649723172188, 'timing/model': 1.5279782405123115, 'learning_rate/qwen_vl_interface': 3.157376197072094e-06, 'learning_rate/action_model': 3.157376197072094e-05, 'epoch': 1.75}) +[2026-06-18 03:09:51,397][starVLA.training.train_starvla][INFO] - Step 4553, Loss: {'train/loss': 0.48282870650291443, 'train/grad_norm_pre_clip': 16.377931594848633, 'timing/data': 0.015341581776738167, 'timing/model': 1.8223942099139094, 'learning_rate/qwen_vl_interface': 3.1554349085006557e-06, 'learning_rate/action_model': 3.155434908500656e-05, 'epoch': 1.75}) +[2026-06-18 03:09:56,291][starVLA.training.train_starvla][INFO] - Step 4554, Loss: {'train/loss': 0.17388752102851868, 'train/grad_norm_pre_clip': 9.018260955810547, 'timing/data': 0.017178148962557316, 'timing/model': 1.1632305569946766, 'learning_rate/qwen_vl_interface': 3.1534940541348152e-06, 'learning_rate/action_model': 3.153494054134815e-05, 'epoch': 1.75}) +[2026-06-18 03:10:04,831][starVLA.training.train_starvla][INFO] - Step 4555, Loss: {'train/loss': 0.2935924828052521, 'train/grad_norm_pre_clip': 15.217967987060547, 'timing/data': 0.013487162999808788, 'timing/model': 1.183119686320424, 'learning_rate/qwen_vl_interface': 3.1515536343769133e-06, 'learning_rate/action_model': 3.151553634376913e-05, 'epoch': 1.75}) +[2026-06-18 03:10:10,881][starVLA.training.train_starvla][INFO] - Step 4556, Loss: {'train/loss': 0.30024635791778564, 'train/grad_norm_pre_clip': 10.285737991333008, 'timing/data': 0.017479658126831055, 'timing/model': 2.109300064854324, 'learning_rate/qwen_vl_interface': 3.1496136496292e-06, 'learning_rate/action_model': 3.1496136496292e-05, 'epoch': 1.75}) +[2026-06-18 03:10:19,034][starVLA.training.train_starvla][INFO] - Step 4557, Loss: {'train/loss': 0.4006444215774536, 'train/grad_norm_pre_clip': 14.253540992736816, 'timing/data': 0.019150017760694027, 'timing/model': 1.2169047314673662, 'learning_rate/qwen_vl_interface': 3.1476741002938383e-06, 'learning_rate/action_model': 3.147674100293838e-05, 'epoch': 1.75}) +[2026-06-18 03:10:26,145][starVLA.training.train_starvla][INFO] - Step 4558, Loss: {'train/loss': 0.4345943033695221, 'train/grad_norm_pre_clip': 18.40237045288086, 'timing/data': 0.015483720228075981, 'timing/model': 2.733754238113761, 'learning_rate/qwen_vl_interface': 3.145734986772897e-06, 'learning_rate/action_model': 3.1457349867728966e-05, 'epoch': 1.75}) +[2026-06-18 03:10:35,726][starVLA.training.train_starvla][INFO] - Step 4559, Loss: {'train/loss': 0.17244476079940796, 'train/grad_norm_pre_clip': 10.51569938659668, 'timing/data': 0.02526361122727394, 'timing/model': 1.5284116677939892, 'learning_rate/qwen_vl_interface': 3.1437963094683566e-06, 'learning_rate/action_model': 3.1437963094683564e-05, 'epoch': 1.75}) +[2026-06-18 03:10:42,518][starVLA.training.train_starvla][INFO] - Step 4560, Loss: {'train/loss': 0.19383275508880615, 'train/grad_norm_pre_clip': 11.687987327575684, 'timing/data': 0.020388170145452023, 'timing/model': 1.0947896307334304, 'learning_rate/qwen_vl_interface': 3.1418580687821103e-06, 'learning_rate/action_model': 3.14185806878211e-05, 'epoch': 1.75}) +[2026-06-18 03:10:47,419][starVLA.training.train_starvla][INFO] - Step 4561, Loss: {'train/loss': 0.4114781618118286, 'train/grad_norm_pre_clip': 17.80494499206543, 'timing/data': 0.01451485138386488, 'timing/model': 1.1928567998111248, 'learning_rate/qwen_vl_interface': 3.1399202651159532e-06, 'learning_rate/action_model': 3.1399202651159534e-05, 'epoch': 1.75}) +[2026-06-18 03:10:52,639][starVLA.training.train_starvla][INFO] - Step 4562, Loss: {'train/loss': 0.12229105085134506, 'train/grad_norm_pre_clip': 6.934704780578613, 'timing/data': 0.014284037053585052, 'timing/model': 1.5597387114539742, 'learning_rate/qwen_vl_interface': 3.1379828988715955e-06, 'learning_rate/action_model': 3.1379828988715956e-05, 'epoch': 1.75}) +[2026-06-18 03:11:01,357][starVLA.training.train_starvla][INFO] - Step 4563, Loss: {'train/loss': 0.5529268980026245, 'train/grad_norm_pre_clip': 15.697136878967285, 'timing/data': 0.015395070426166058, 'timing/model': 1.3082885760813951, 'learning_rate/qwen_vl_interface': 3.136045970450654e-06, 'learning_rate/action_model': 3.136045970450654e-05, 'epoch': 1.76}) +[2026-06-18 03:11:16,133][starVLA.training.train_starvla][INFO] - Step 4564, Loss: {'train/loss': 0.31557637453079224, 'train/grad_norm_pre_clip': 14.428503036499023, 'timing/data': 0.01533344853669405, 'timing/model': 4.391509003937244, 'learning_rate/qwen_vl_interface': 3.1341094802546595e-06, 'learning_rate/action_model': 3.1341094802546595e-05, 'epoch': 1.76}) +[2026-06-18 03:11:23,017][starVLA.training.train_starvla][INFO] - Step 4565, Loss: {'train/loss': 0.4810022711753845, 'train/grad_norm_pre_clip': 14.793525695800781, 'timing/data': 0.014870434068143368, 'timing/model': 1.2402238994836807, 'learning_rate/qwen_vl_interface': 3.132173428685048e-06, 'learning_rate/action_model': 3.132173428685048e-05, 'epoch': 1.76}) +[2026-06-18 03:11:28,407][starVLA.training.train_starvla][INFO] - Step 4566, Loss: {'train/loss': 0.5793155431747437, 'train/grad_norm_pre_clip': 23.80640983581543, 'timing/data': 0.030864913016557693, 'timing/model': 1.2599450889974833, 'learning_rate/qwen_vl_interface': 3.13023781614316e-06, 'learning_rate/action_model': 3.13023781614316e-05, 'epoch': 1.76}) +[2026-06-18 03:11:43,726][starVLA.training.train_starvla][INFO] - Step 4567, Loss: {'train/loss': 0.3039320111274719, 'train/grad_norm_pre_clip': 12.249900817871094, 'timing/data': 0.016168838366866112, 'timing/model': 5.8310949597507715, 'learning_rate/qwen_vl_interface': 3.1283026430302553e-06, 'learning_rate/action_model': 3.128302643030255e-05, 'epoch': 1.76}) +[2026-06-18 03:11:55,721][starVLA.training.train_starvla][INFO] - Step 4568, Loss: {'train/loss': 0.16105210781097412, 'train/grad_norm_pre_clip': 8.895045280456543, 'timing/data': 0.01504542212933302, 'timing/model': 2.8163089603185654, 'learning_rate/qwen_vl_interface': 3.1263679097474958e-06, 'learning_rate/action_model': 3.126367909747496e-05, 'epoch': 1.76}) +[2026-06-18 03:12:11,796][starVLA.training.train_starvla][INFO] - Step 4569, Loss: {'train/loss': 0.251773864030838, 'train/grad_norm_pre_clip': 11.454108238220215, 'timing/data': 0.014410790055990219, 'timing/model': 5.380741671659052, 'learning_rate/qwen_vl_interface': 3.124433616695955e-06, 'learning_rate/action_model': 3.1244336166959545e-05, 'epoch': 1.76}) +[2026-06-18 03:12:16,718][starVLA.training.train_starvla][INFO] - Step 4570, Loss: {'train/loss': 0.4435569643974304, 'train/grad_norm_pre_clip': 18.418380737304688, 'timing/data': 0.014829388819634914, 'timing/model': 1.250944989733398, 'learning_rate/qwen_vl_interface': 3.1224997642766115e-06, 'learning_rate/action_model': 3.122499764276611e-05, 'epoch': 1.76}) +[2026-06-18 03:12:22,539][starVLA.training.train_starvla][INFO] - Step 4571, Loss: {'train/loss': 0.15744510293006897, 'train/grad_norm_pre_clip': 8.351247787475586, 'timing/data': 0.022244703955948353, 'timing/model': 1.2294386960566044, 'learning_rate/qwen_vl_interface': 3.1205663528903563e-06, 'learning_rate/action_model': 3.1205663528903565e-05, 'epoch': 1.76}) +[2026-06-18 03:12:27,951][starVLA.training.train_starvla][INFO] - Step 4572, Loss: {'train/loss': 0.24164631962776184, 'train/grad_norm_pre_clip': 9.914763450622559, 'timing/data': 0.02030953485518694, 'timing/model': 1.5238744653761387, 'learning_rate/qwen_vl_interface': 3.118633382937987e-06, 'learning_rate/action_model': 3.118633382937987e-05, 'epoch': 1.76}) +[2026-06-18 03:12:33,154][starVLA.training.train_starvla][INFO] - Step 4573, Loss: {'train/loss': 0.3734739124774933, 'train/grad_norm_pre_clip': 14.717423439025879, 'timing/data': 0.014722581021487713, 'timing/model': 1.3575141755864024, 'learning_rate/qwen_vl_interface': 3.116700854820209e-06, 'learning_rate/action_model': 3.116700854820209e-05, 'epoch': 1.76}) +[2026-06-18 03:12:38,093][starVLA.training.train_starvla][INFO] - Step 4574, Loss: {'train/loss': 0.38390684127807617, 'train/grad_norm_pre_clip': 14.754655838012695, 'timing/data': 0.014588313177227974, 'timing/model': 1.1708961743861437, 'learning_rate/qwen_vl_interface': 3.1147687689376443e-06, 'learning_rate/action_model': 3.1147687689376444e-05, 'epoch': 1.76}) +[2026-06-18 03:12:42,874][starVLA.training.train_starvla][INFO] - Step 4575, Loss: {'train/loss': 0.5665187239646912, 'train/grad_norm_pre_clip': 16.374832153320312, 'timing/data': 0.015813047997653484, 'timing/model': 1.1209572525694966, 'learning_rate/qwen_vl_interface': 3.1128371256908087e-06, 'learning_rate/action_model': 3.112837125690809e-05, 'epoch': 1.76}) +[2026-06-18 03:12:47,553][starVLA.training.train_starvla][INFO] - Step 4576, Loss: {'train/loss': 0.30396437644958496, 'train/grad_norm_pre_clip': 13.968130111694336, 'timing/data': 0.015842373482882977, 'timing/model': 1.1017340626567602, 'learning_rate/qwen_vl_interface': 3.1109059254801368e-06, 'learning_rate/action_model': 3.1109059254801366e-05, 'epoch': 1.76}) +[2026-06-18 03:12:52,571][starVLA.training.train_starvla][INFO] - Step 4577, Loss: {'train/loss': 0.5044988393783569, 'train/grad_norm_pre_clip': 20.478364944458008, 'timing/data': 0.03627793863415718, 'timing/model': 1.5648329118266702, 'learning_rate/qwen_vl_interface': 3.108975168705966e-06, 'learning_rate/action_model': 3.1089751687059656e-05, 'epoch': 1.76}) +[2026-06-18 03:12:56,998][starVLA.training.train_starvla][INFO] - Step 4578, Loss: {'train/loss': 0.628418505191803, 'train/grad_norm_pre_clip': 17.264514923095703, 'timing/data': 0.012975642457604408, 'timing/model': 1.1007917625829577, 'learning_rate/qwen_vl_interface': 3.107044855768549e-06, 'learning_rate/action_model': 3.107044855768549e-05, 'epoch': 1.76}) +[2026-06-18 03:13:02,363][starVLA.training.train_starvla][INFO] - Step 4579, Loss: {'train/loss': 0.42052754759788513, 'train/grad_norm_pre_clip': 14.983926773071289, 'timing/data': 0.015700343064963818, 'timing/model': 1.2195902755483985, 'learning_rate/qwen_vl_interface': 3.1051149870680413e-06, 'learning_rate/action_model': 3.105114987068041e-05, 'epoch': 1.76}) +[2026-06-18 03:13:07,385][starVLA.training.train_starvla][INFO] - Step 4580, Loss: {'train/loss': 0.3991059958934784, 'train/grad_norm_pre_clip': 10.818346977233887, 'timing/data': 0.014599660411477089, 'timing/model': 1.2276652017608285, 'learning_rate/qwen_vl_interface': 3.103185563004503e-06, 'learning_rate/action_model': 3.103185563004503e-05, 'epoch': 1.76}) +[2026-06-18 03:13:12,400][starVLA.training.train_starvla][INFO] - Step 4581, Loss: {'train/loss': 0.08554491400718689, 'train/grad_norm_pre_clip': 7.138936996459961, 'timing/data': 0.015021161176264286, 'timing/model': 1.142233245074749, 'learning_rate/qwen_vl_interface': 3.101256583977905e-06, 'learning_rate/action_model': 3.101256583977905e-05, 'epoch': 1.76}) +[2026-06-18 03:13:17,470][starVLA.training.train_starvla][INFO] - Step 4582, Loss: {'train/loss': 0.3683319091796875, 'train/grad_norm_pre_clip': 12.666627883911133, 'timing/data': 0.013063660822808743, 'timing/model': 1.4864379707723856, 'learning_rate/qwen_vl_interface': 3.099328050388131e-06, 'learning_rate/action_model': 3.0993280503881305e-05, 'epoch': 1.76}) +[2026-06-18 03:13:22,621][starVLA.training.train_starvla][INFO] - Step 4583, Loss: {'train/loss': 0.23452171683311462, 'train/grad_norm_pre_clip': 12.244180679321289, 'timing/data': 0.01604864001274109, 'timing/model': 1.3582550249993801, 'learning_rate/qwen_vl_interface': 3.0973999626349667e-06, 'learning_rate/action_model': 3.097399962634966e-05, 'epoch': 1.76}) +[2026-06-18 03:13:28,530][starVLA.training.train_starvla][INFO] - Step 4584, Loss: {'train/loss': 0.12646634876728058, 'train/grad_norm_pre_clip': 9.744176864624023, 'timing/data': 0.025821072980761528, 'timing/model': 1.4078568778932095, 'learning_rate/qwen_vl_interface': 3.095472321118107e-06, 'learning_rate/action_model': 3.095472321118107e-05, 'epoch': 1.76}) +[2026-06-18 03:13:35,633][starVLA.training.train_starvla][INFO] - Step 4585, Loss: {'train/loss': 0.3019486367702484, 'train/grad_norm_pre_clip': 12.8757963180542, 'timing/data': 0.05979701038450003, 'timing/model': 1.3179152207449079, 'learning_rate/qwen_vl_interface': 3.093545126237153e-06, 'learning_rate/action_model': 3.0935451262371527e-05, 'epoch': 1.76}) +[2026-06-18 03:13:48,706][starVLA.training.train_starvla][INFO] - Step 4586, Loss: {'train/loss': 0.20072928071022034, 'train/grad_norm_pre_clip': 12.83289623260498, 'timing/data': 0.021138631738722324, 'timing/model': 1.4739651530981064, 'learning_rate/qwen_vl_interface': 3.091618378391615e-06, 'learning_rate/action_model': 3.0916183783916146e-05, 'epoch': 1.76}) +[2026-06-18 03:14:03,178][starVLA.training.train_starvla][INFO] - Step 4587, Loss: {'train/loss': 0.2616267204284668, 'train/grad_norm_pre_clip': 15.936762809753418, 'timing/data': 0.019521445967257023, 'timing/model': 5.947736667469144, 'learning_rate/qwen_vl_interface': 3.0896920779809097e-06, 'learning_rate/action_model': 3.08969207798091e-05, 'epoch': 1.76}) +[2026-06-18 03:14:09,933][starVLA.training.train_starvla][INFO] - Step 4588, Loss: {'train/loss': 0.31934303045272827, 'train/grad_norm_pre_clip': 11.803680419921875, 'timing/data': 0.017184020951390266, 'timing/model': 1.2027724785730243, 'learning_rate/qwen_vl_interface': 3.0877662254043605e-06, 'learning_rate/action_model': 3.0877662254043606e-05, 'epoch': 1.76}) +[2026-06-18 03:14:21,214][starVLA.training.train_starvla][INFO] - Step 4589, Loss: {'train/loss': 0.2771688997745514, 'train/grad_norm_pre_clip': 11.72329330444336, 'timing/data': 0.012398801743984222, 'timing/model': 3.1376578621566296, 'learning_rate/qwen_vl_interface': 3.0858408210612e-06, 'learning_rate/action_model': 3.0858408210612e-05, 'epoch': 1.77}) +[2026-06-18 03:14:28,514][starVLA.training.train_starvla][INFO] - Step 4590, Loss: {'train/loss': 0.17875336110591888, 'train/grad_norm_pre_clip': 10.157069206237793, 'timing/data': 0.01335048209875822, 'timing/model': 1.3136108350008726, 'learning_rate/qwen_vl_interface': 3.0839158653505663e-06, 'learning_rate/action_model': 3.0839158653505664e-05, 'epoch': 1.77}) +[2026-06-18 03:14:36,928][starVLA.training.train_starvla][INFO] - Step 4591, Loss: {'train/loss': 0.3246973156929016, 'train/grad_norm_pre_clip': 12.004741668701172, 'timing/data': 0.030316419899463654, 'timing/model': 1.0897961296141148, 'learning_rate/qwen_vl_interface': 3.081991358671504e-06, 'learning_rate/action_model': 3.081991358671504e-05, 'epoch': 1.77}) +[2026-06-18 03:14:50,847][starVLA.training.train_starvla][INFO] - Step 4592, Loss: {'train/loss': 0.39374709129333496, 'train/grad_norm_pre_clip': 16.477014541625977, 'timing/data': 0.014533302746713161, 'timing/model': 5.010781968943775, 'learning_rate/qwen_vl_interface': 3.080067301422965e-06, 'learning_rate/action_model': 3.080067301422965e-05, 'epoch': 1.77}) +[2026-06-18 03:15:08,141][starVLA.training.train_starvla][INFO] - Step 4593, Loss: {'train/loss': 0.2660202383995056, 'train/grad_norm_pre_clip': 14.977193832397461, 'timing/data': 0.04254481289535761, 'timing/model': 5.999149535782635, 'learning_rate/qwen_vl_interface': 3.07814369400381e-06, 'learning_rate/action_model': 3.07814369400381e-05, 'epoch': 1.77}) +[2026-06-18 03:15:18,220][starVLA.training.train_starvla][INFO] - Step 4594, Loss: {'train/loss': 0.6842465996742249, 'train/grad_norm_pre_clip': 15.38501262664795, 'timing/data': 0.04045581165701151, 'timing/model': 1.5303690787404776, 'learning_rate/qwen_vl_interface': 3.076220536812807e-06, 'learning_rate/action_model': 3.076220536812807e-05, 'epoch': 1.77}) +[2026-06-18 03:15:23,562][starVLA.training.train_starvla][INFO] - Step 4595, Loss: {'train/loss': 0.3295086920261383, 'train/grad_norm_pre_clip': 17.62774658203125, 'timing/data': 0.05338846147060394, 'timing/model': 1.286600866355002, 'learning_rate/qwen_vl_interface': 3.0742978302486214e-06, 'learning_rate/action_model': 3.074297830248621e-05, 'epoch': 1.77}) +[2026-06-18 03:15:29,987][starVLA.training.train_starvla][INFO] - Step 4596, Loss: {'train/loss': 0.18995487689971924, 'train/grad_norm_pre_clip': 8.953289985656738, 'timing/data': 0.04344580415636301, 'timing/model': 1.493642126210034, 'learning_rate/qwen_vl_interface': 3.0723755747098387e-06, 'learning_rate/action_model': 3.072375574709839e-05, 'epoch': 1.77}) +[2026-06-18 03:15:35,347][starVLA.training.train_starvla][INFO] - Step 4597, Loss: {'train/loss': 0.46698129177093506, 'train/grad_norm_pre_clip': 22.05845832824707, 'timing/data': 0.03441437613219023, 'timing/model': 1.7735095173120499, 'learning_rate/qwen_vl_interface': 3.070453770594943e-06, 'learning_rate/action_model': 3.070453770594943e-05, 'epoch': 1.77}) +[2026-06-18 03:15:40,779][starVLA.training.train_starvla][INFO] - Step 4598, Loss: {'train/loss': 0.504391610622406, 'train/grad_norm_pre_clip': 17.320127487182617, 'timing/data': 0.04272646363824606, 'timing/model': 1.4832412162795663, 'learning_rate/qwen_vl_interface': 3.0685324183023264e-06, 'learning_rate/action_model': 3.068532418302326e-05, 'epoch': 1.77}) +[2026-06-18 03:15:46,358][starVLA.training.train_starvla][INFO] - Step 4599, Loss: {'train/loss': 0.5061981081962585, 'train/grad_norm_pre_clip': 15.941737174987793, 'timing/data': 0.05295657832175493, 'timing/model': 1.6016648029908538, 'learning_rate/qwen_vl_interface': 3.0666115182302863e-06, 'learning_rate/action_model': 3.0666115182302865e-05, 'epoch': 1.77}) +[2026-06-18 03:19:43,389][starVLA.training.train_starvla][INFO] - Step 4600, Loss: {'train/loss': 0.446659654378891, 'train/grad_norm_pre_clip': 17.10130500793457, 'eval/loss': 0.4364909827709198, 'eval/latency_4/loss': 0.4364909827709198, 'eval/demon_fire_precision': 0.9624012638230648, 'eval/demon_fire_recall': 0.989282234491718, 'eval/demon_fire_f1': 0.9756566303651506, 'eval/demon_fire_rate_model': 0.9295154185022027, 'eval/demon_fire_rate_teacher': 0.9042584434654919, 'eval/demon_left_precision': 0.7780678851174935, 'eval/demon_left_recall': 0.680365296803653, 'eval/demon_left_f1': 0.7259439707673568, 'eval/demon_left_rate_model': 0.11248164464023495, 'eval/demon_left_rate_teacher': 0.12863436123348018, 'eval/demon_right_precision': 0.7690972222222222, 'eval/demon_right_recall': 0.7996389891696751, 'eval/demon_right_f1': 0.7840707964601771, 'eval/demon_right_rate_model': 0.16916299559471365, 'eval/demon_right_rate_teacher': 0.16270190895741557, 'eval/demon_move_f1': 0.755007383613767, 'eval/demon_cc_f1': 0.8970117198852691, 'eval/latency_4/demon_fire_precision': 0.9624012638230648, 'eval/latency_4/demon_fire_recall': 0.989282234491718, 'eval/latency_4/demon_fire_f1': 0.9756566303651506, 'eval/latency_4/demon_fire_rate_model': 0.9295154185022027, 'eval/latency_4/demon_fire_rate_teacher': 0.9042584434654919, 'eval/latency_4/demon_left_precision': 0.7780678851174935, 'eval/latency_4/demon_left_recall': 0.680365296803653, 'eval/latency_4/demon_left_f1': 0.7259439707673568, 'eval/latency_4/demon_left_rate_model': 0.11248164464023495, 'eval/latency_4/demon_left_rate_teacher': 0.12863436123348018, 'eval/latency_4/demon_right_precision': 0.7690972222222222, 'eval/latency_4/demon_right_recall': 0.7996389891696751, 'eval/latency_4/demon_right_f1': 0.7840707964601771, 'eval/latency_4/demon_right_rate_model': 0.16916299559471365, 'eval/latency_4/demon_right_rate_teacher': 0.16270190895741557, 'eval/latency_4/demon_move_f1': 0.755007383613767, 'eval/latency_4/demon_cc_f1': 0.8970117198852691, 'timing/data': 0.041823746636509895, 'timing/model': 1.3521073246374726, 'learning_rate/qwen_vl_interface': 3.064691070777027e-06, 'learning_rate/action_model': 3.064691070777027e-05, 'epoch': 1.77}) +[2026-06-18 03:19:43,393][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_4600_state +[2026-06-18 03:19:43,401][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 03:19:52,409][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_4600_state/model.safetensors +[2026-06-18 03:20:15,442][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_4600_state/optimizer.bin +[2026-06-18 03:20:15,443][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_4600_state/sampler.bin +[2026-06-18 03:20:15,444][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_4600_state/sampler_1.bin +[2026-06-18 03:20:15,446][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_4600_state/random_states_0.pkl +[2026-06-18 03:20:15,511][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4600 +[2026-06-18 03:20:15,511][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 03:20:15,530][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 03:20:23,895][starVLA.training.train_starvla][INFO] - Step 4601, Loss: {'train/loss': 0.3001236021518707, 'train/grad_norm_pre_clip': 11.220480918884277, 'timing/data': 0.04666437488049269, 'timing/model': 4.964540363289416, 'learning_rate/qwen_vl_interface': 3.062771076340662e-06, 'learning_rate/action_model': 3.0627710763406617e-05, 'epoch': 1.77}) +[2026-06-18 03:20:28,804][starVLA.training.train_starvla][INFO] - Step 4602, Loss: {'train/loss': 0.3329205811023712, 'train/grad_norm_pre_clip': 13.660101890563965, 'timing/data': 0.042767494916915894, 'timing/model': 1.290697455406189, 'learning_rate/qwen_vl_interface': 3.0608515353192046e-06, 'learning_rate/action_model': 3.060851535319205e-05, 'epoch': 1.77}) +[2026-06-18 03:20:45,737][starVLA.training.train_starvla][INFO] - Step 4603, Loss: {'train/loss': 0.47182172536849976, 'train/grad_norm_pre_clip': 22.62427520751953, 'timing/data': 0.04404348134994507, 'timing/model': 4.099478513002396, 'learning_rate/qwen_vl_interface': 3.0589324481105814e-06, 'learning_rate/action_model': 3.058932448110581e-05, 'epoch': 1.77}) +[2026-06-18 03:20:54,071][starVLA.training.train_starvla][INFO] - Step 4604, Loss: {'train/loss': 0.5017149448394775, 'train/grad_norm_pre_clip': 12.874224662780762, 'timing/data': 0.041201597079634666, 'timing/model': 1.1105119567364454, 'learning_rate/qwen_vl_interface': 3.0570138151126185e-06, 'learning_rate/action_model': 3.057013815112618e-05, 'epoch': 1.77}) +[2026-06-18 03:20:59,600][starVLA.training.train_starvla][INFO] - Step 4605, Loss: {'train/loss': 0.23404300212860107, 'train/grad_norm_pre_clip': 11.389982223510742, 'timing/data': 0.03579712100327015, 'timing/model': 1.4858428416773677, 'learning_rate/qwen_vl_interface': 3.0550956367230506e-06, 'learning_rate/action_model': 3.05509563672305e-05, 'epoch': 1.77}) +[2026-06-18 03:21:04,988][starVLA.training.train_starvla][INFO] - Step 4606, Loss: {'train/loss': 0.24656540155410767, 'train/grad_norm_pre_clip': 11.893289566040039, 'timing/data': 0.055394415743649006, 'timing/model': 1.398218016140163, 'learning_rate/qwen_vl_interface': 3.0531779133395184e-06, 'learning_rate/action_model': 3.053177913339518e-05, 'epoch': 1.77}) +[2026-06-18 03:21:09,916][starVLA.training.train_starvla][INFO] - Step 4607, Loss: {'train/loss': 0.3257308006286621, 'train/grad_norm_pre_clip': 11.184459686279297, 'timing/data': 0.04598035104572773, 'timing/model': 1.2808235129341483, 'learning_rate/qwen_vl_interface': 3.0512606453595708e-06, 'learning_rate/action_model': 3.0512606453595704e-05, 'epoch': 1.77}) +[2026-06-18 03:21:19,479][starVLA.training.train_starvla][INFO] - Step 4608, Loss: {'train/loss': 0.16760407388210297, 'train/grad_norm_pre_clip': 9.777663230895996, 'timing/data': 0.03301546163856983, 'timing/model': 5.299184361472726, 'learning_rate/qwen_vl_interface': 3.049343833180658e-06, 'learning_rate/action_model': 3.049343833180658e-05, 'epoch': 1.77}) +[2026-06-18 03:21:28,137][starVLA.training.train_starvla][INFO] - Step 4609, Loss: {'train/loss': 0.2724594175815582, 'train/grad_norm_pre_clip': 11.533828735351562, 'timing/data': 0.03553264867514372, 'timing/model': 1.1806771084666252, 'learning_rate/qwen_vl_interface': 3.0474274772001355e-06, 'learning_rate/action_model': 3.0474274772001355e-05, 'epoch': 1.77}) +[2026-06-18 03:21:33,235][starVLA.training.train_starvla][INFO] - Step 4610, Loss: {'train/loss': 0.5263102054595947, 'train/grad_norm_pre_clip': 18.8294734954834, 'timing/data': 0.031542666256427765, 'timing/model': 1.187061539851129, 'learning_rate/qwen_vl_interface': 3.045511577815265e-06, 'learning_rate/action_model': 3.045511577815265e-05, 'epoch': 1.77}) +[2026-06-18 03:21:38,130][starVLA.training.train_starvla][INFO] - Step 4611, Loss: {'train/loss': 0.47607290744781494, 'train/grad_norm_pre_clip': 16.194658279418945, 'timing/data': 0.0317414877936244, 'timing/model': 1.105406199581921, 'learning_rate/qwen_vl_interface': 3.043596135423218e-06, 'learning_rate/action_model': 3.043596135423218e-05, 'epoch': 1.77}) +[2026-06-18 03:21:46,290][starVLA.training.train_starvla][INFO] - Step 4612, Loss: {'train/loss': 0.5694985389709473, 'train/grad_norm_pre_clip': 15.97718334197998, 'timing/data': 0.034315258264541626, 'timing/model': 1.5227191569283605, 'learning_rate/qwen_vl_interface': 3.041681150421067e-06, 'learning_rate/action_model': 3.041681150421067e-05, 'epoch': 1.77}) +[2026-06-18 03:21:50,642][starVLA.training.train_starvla][INFO] - Step 4613, Loss: {'train/loss': 0.3140515387058258, 'train/grad_norm_pre_clip': 13.157855033874512, 'timing/data': 0.03180072735995054, 'timing/model': 1.0776762384921312, 'learning_rate/qwen_vl_interface': 3.039766623205791e-06, 'learning_rate/action_model': 3.039766623205791e-05, 'epoch': 1.77}) +[2026-06-18 03:21:55,431][starVLA.training.train_starvla][INFO] - Step 4614, Loss: {'train/loss': 0.2096620798110962, 'train/grad_norm_pre_clip': 15.29501724243164, 'timing/data': 0.04237130004912615, 'timing/model': 1.1114190053194761, 'learning_rate/qwen_vl_interface': 3.037852554174269e-06, 'learning_rate/action_model': 3.0378525541742692e-05, 'epoch': 1.77}) +[2026-06-18 03:22:00,364][starVLA.training.train_starvla][INFO] - Step 4615, Loss: {'train/loss': 0.3907301127910614, 'train/grad_norm_pre_clip': 12.127046585083008, 'timing/data': 0.03430724050849676, 'timing/model': 1.148566178046167, 'learning_rate/qwen_vl_interface': 3.035938943723295e-06, 'learning_rate/action_model': 3.035938943723295e-05, 'epoch': 1.78}) +[2026-06-18 03:22:05,153][starVLA.training.train_starvla][INFO] - Step 4616, Loss: {'train/loss': 0.6229580640792847, 'train/grad_norm_pre_clip': 22.235807418823242, 'timing/data': 0.030850748531520367, 'timing/model': 1.0977048175409436, 'learning_rate/qwen_vl_interface': 3.03402579224956e-06, 'learning_rate/action_model': 3.03402579224956e-05, 'epoch': 1.78}) +[2026-06-18 03:22:09,958][starVLA.training.train_starvla][INFO] - Step 4617, Loss: {'train/loss': 0.3092530369758606, 'train/grad_norm_pre_clip': 11.567083358764648, 'timing/data': 0.034901020117104053, 'timing/model': 1.2047233562916517, 'learning_rate/qwen_vl_interface': 3.032113100149663e-06, 'learning_rate/action_model': 3.0321131001496628e-05, 'epoch': 1.78}) +[2026-06-18 03:22:19,272][starVLA.training.train_starvla][INFO] - Step 4618, Loss: {'train/loss': 0.16656532883644104, 'train/grad_norm_pre_clip': 8.962645530700684, 'timing/data': 0.034423571079969406, 'timing/model': 1.2767024608328938, 'learning_rate/qwen_vl_interface': 3.0302008678201067e-06, 'learning_rate/action_model': 3.0302008678201067e-05, 'epoch': 1.78}) +[2026-06-18 03:22:24,045][starVLA.training.train_starvla][INFO] - Step 4619, Loss: {'train/loss': 0.5358072519302368, 'train/grad_norm_pre_clip': 17.071762084960938, 'timing/data': 0.03328364063054323, 'timing/model': 1.1775606283918023, 'learning_rate/qwen_vl_interface': 3.028289095657299e-06, 'learning_rate/action_model': 3.028289095657299e-05, 'epoch': 1.78}) +[2026-06-18 03:22:31,900][starVLA.training.train_starvla][INFO] - Step 4620, Loss: {'train/loss': 0.7311956882476807, 'train/grad_norm_pre_clip': 18.36697006225586, 'timing/data': 0.03719678055495024, 'timing/model': 1.0829122737050056, 'learning_rate/qwen_vl_interface': 3.026377784057552e-06, 'learning_rate/action_model': 3.026377784057552e-05, 'epoch': 1.78}) +[2026-06-18 03:22:36,822][starVLA.training.train_starvla][INFO] - Step 4621, Loss: {'train/loss': 0.569987952709198, 'train/grad_norm_pre_clip': 21.43219566345215, 'timing/data': 0.034632802940905094, 'timing/model': 1.5349331749603152, 'learning_rate/qwen_vl_interface': 3.0244669334170822e-06, 'learning_rate/action_model': 3.0244669334170822e-05, 'epoch': 1.78}) +[2026-06-18 03:22:41,530][starVLA.training.train_starvla][INFO] - Step 4622, Loss: {'train/loss': 0.3548298180103302, 'train/grad_norm_pre_clip': 17.8056583404541, 'timing/data': 0.04581499006599188, 'timing/model': 1.4097426859661937, 'learning_rate/qwen_vl_interface': 3.0225565441320156e-06, 'learning_rate/action_model': 3.0225565441320152e-05, 'epoch': 1.78}) +[2026-06-18 03:22:46,309][starVLA.training.train_starvla][INFO] - Step 4623, Loss: {'train/loss': 0.4351319670677185, 'train/grad_norm_pre_clip': 66.17498016357422, 'timing/data': 0.028697156347334385, 'timing/model': 1.139958694577217, 'learning_rate/qwen_vl_interface': 3.0206466165983716e-06, 'learning_rate/action_model': 3.0206466165983716e-05, 'epoch': 1.78}) +[2026-06-18 03:22:50,659][starVLA.training.train_starvla][INFO] - Step 4624, Loss: {'train/loss': 0.2365899235010147, 'train/grad_norm_pre_clip': 13.009057998657227, 'timing/data': 0.026345222257077694, 'timing/model': 1.0502382395789027, 'learning_rate/qwen_vl_interface': 3.0187371512120833e-06, 'learning_rate/action_model': 3.0187371512120832e-05, 'epoch': 1.78}) +[2026-06-18 03:22:55,195][starVLA.training.train_starvla][INFO] - Step 4625, Loss: {'train/loss': 0.3479669690132141, 'train/grad_norm_pre_clip': 15.102385520935059, 'timing/data': 0.028335374779999256, 'timing/model': 1.0502444757148623, 'learning_rate/qwen_vl_interface': 3.0168281483689825e-06, 'learning_rate/action_model': 3.0168281483689825e-05, 'epoch': 1.78}) +[2026-06-18 03:22:59,640][starVLA.training.train_starvla][INFO] - Step 4626, Loss: {'train/loss': 0.2557430565357208, 'train/grad_norm_pre_clip': 17.264867782592773, 'timing/data': 0.02691931463778019, 'timing/model': 1.065074222162366, 'learning_rate/qwen_vl_interface': 3.014919608464812e-06, 'learning_rate/action_model': 3.0149196084648117e-05, 'epoch': 1.78}) +[2026-06-18 03:23:03,905][starVLA.training.train_starvla][INFO] - Step 4627, Loss: {'train/loss': 0.548105776309967, 'train/grad_norm_pre_clip': 18.677106857299805, 'timing/data': 0.027874140068888664, 'timing/model': 1.270043408498168, 'learning_rate/qwen_vl_interface': 3.0130115318952124e-06, 'learning_rate/action_model': 3.0130115318952122e-05, 'epoch': 1.78}) +[2026-06-18 03:23:08,013][starVLA.training.train_starvla][INFO] - Step 4628, Loss: {'train/loss': 0.43537113070487976, 'train/grad_norm_pre_clip': 16.9288272857666, 'timing/data': 0.02876679040491581, 'timing/model': 1.0796554302796721, 'learning_rate/qwen_vl_interface': 3.011103919055726e-06, 'learning_rate/action_model': 3.011103919055726e-05, 'epoch': 1.78}) +[2026-06-18 03:23:15,845][starVLA.training.train_starvla][INFO] - Step 4629, Loss: {'train/loss': 0.4807201623916626, 'train/grad_norm_pre_clip': 17.215484619140625, 'timing/data': 0.026849187910556793, 'timing/model': 4.362454987131059, 'learning_rate/qwen_vl_interface': 3.009196770341809e-06, 'learning_rate/action_model': 3.009196770341809e-05, 'epoch': 1.78}) +[2026-06-18 03:23:20,386][starVLA.training.train_starvla][INFO] - Step 4630, Loss: {'train/loss': 0.3425489366054535, 'train/grad_norm_pre_clip': 14.558098793029785, 'timing/data': 0.03291577938944101, 'timing/model': 1.1163250608369708, 'learning_rate/qwen_vl_interface': 3.0072900861488125e-06, 'learning_rate/action_model': 3.0072900861488125e-05, 'epoch': 1.78}) +[2026-06-18 03:23:24,776][starVLA.training.train_starvla][INFO] - Step 4631, Loss: {'train/loss': 0.6912105083465576, 'train/grad_norm_pre_clip': 16.296964645385742, 'timing/data': 0.02699370589107275, 'timing/model': 1.0897634476423264, 'learning_rate/qwen_vl_interface': 3.0053838668719944e-06, 'learning_rate/action_model': 3.0053838668719943e-05, 'epoch': 1.78}) +[2026-06-18 03:23:29,374][starVLA.training.train_starvla][INFO] - Step 4632, Loss: {'train/loss': 0.2932855784893036, 'train/grad_norm_pre_clip': 10.86587142944336, 'timing/data': 0.02607234288007021, 'timing/model': 1.3729944778606296, 'learning_rate/qwen_vl_interface': 3.003478112906515e-06, 'learning_rate/action_model': 3.0034781129065148e-05, 'epoch': 1.78}) +[2026-06-18 03:23:33,579][starVLA.training.train_starvla][INFO] - Step 4633, Loss: {'train/loss': 0.29966333508491516, 'train/grad_norm_pre_clip': 15.22488021850586, 'timing/data': 0.027869099751114845, 'timing/model': 1.0811258312314749, 'learning_rate/qwen_vl_interface': 3.0015728246474405e-06, 'learning_rate/action_model': 3.0015728246474405e-05, 'epoch': 1.78}) +[2026-06-18 03:23:38,035][starVLA.training.train_starvla][INFO] - Step 4634, Loss: {'train/loss': 0.4415251314640045, 'train/grad_norm_pre_clip': 20.40212059020996, 'timing/data': 0.028892339207232, 'timing/model': 1.146829946897924, 'learning_rate/qwen_vl_interface': 2.99966800248974e-06, 'learning_rate/action_model': 2.99966800248974e-05, 'epoch': 1.78}) +[2026-06-18 03:23:42,982][starVLA.training.train_starvla][INFO] - Step 4635, Loss: {'train/loss': 0.23868438601493835, 'train/grad_norm_pre_clip': 8.510563850402832, 'timing/data': 0.03248391021043062, 'timing/model': 1.1375890662893653, 'learning_rate/qwen_vl_interface': 2.9977636468282817e-06, 'learning_rate/action_model': 2.9977636468282816e-05, 'epoch': 1.78}) +[2026-06-18 03:23:48,372][starVLA.training.train_starvla][INFO] - Step 4636, Loss: {'train/loss': 0.29580023884773254, 'train/grad_norm_pre_clip': 15.167398452758789, 'timing/data': 0.0476520424708724, 'timing/model': 1.5209785625338554, 'learning_rate/qwen_vl_interface': 2.995859758057846e-06, 'learning_rate/action_model': 2.9958597580578456e-05, 'epoch': 1.78}) +[2026-06-18 03:23:53,677][starVLA.training.train_starvla][INFO] - Step 4637, Loss: {'train/loss': 0.2779995799064636, 'train/grad_norm_pre_clip': 18.121957778930664, 'timing/data': 0.03867018781602383, 'timing/model': 1.4592173034325242, 'learning_rate/qwen_vl_interface': 2.993956336573106e-06, 'learning_rate/action_model': 2.9939563365731062e-05, 'epoch': 1.78}) +[2026-06-18 03:24:02,772][starVLA.training.train_starvla][INFO] - Step 4638, Loss: {'train/loss': 0.43391668796539307, 'train/grad_norm_pre_clip': 18.439414978027344, 'timing/data': 0.036456755362451077, 'timing/model': 4.214886322617531, 'learning_rate/qwen_vl_interface': 2.9920533827686453e-06, 'learning_rate/action_model': 2.9920533827686453e-05, 'epoch': 1.78}) +[2026-06-18 03:24:08,858][starVLA.training.train_starvla][INFO] - Step 4639, Loss: {'train/loss': 0.3455209732055664, 'train/grad_norm_pre_clip': 12.928858757019043, 'timing/data': 0.04311198182404041, 'timing/model': 1.6398415379226208, 'learning_rate/qwen_vl_interface': 2.9901508970389477e-06, 'learning_rate/action_model': 2.9901508970389476e-05, 'epoch': 1.78}) +[2026-06-18 03:24:16,315][starVLA.training.train_starvla][INFO] - Step 4640, Loss: {'train/loss': 0.20823794603347778, 'train/grad_norm_pre_clip': 9.443464279174805, 'timing/data': 0.03854662925004959, 'timing/model': 1.3313683308660984, 'learning_rate/qwen_vl_interface': 2.988248879778401e-06, 'learning_rate/action_model': 2.988248879778401e-05, 'epoch': 1.78}) +[2026-06-18 03:24:22,405][starVLA.training.train_starvla][INFO] - Step 4641, Loss: {'train/loss': 0.32672956585884094, 'train/grad_norm_pre_clip': 17.411706924438477, 'timing/data': 0.03684532642364502, 'timing/model': 2.287077872082591, 'learning_rate/qwen_vl_interface': 2.986347331381297e-06, 'learning_rate/action_model': 2.9863473313812967e-05, 'epoch': 1.79}) +[2026-06-18 03:24:28,520][starVLA.training.train_starvla][INFO] - Step 4642, Loss: {'train/loss': 0.4002660810947418, 'train/grad_norm_pre_clip': 18.709585189819336, 'timing/data': 0.039488730020821095, 'timing/model': 1.7534555569291115, 'learning_rate/qwen_vl_interface': 2.984446252241826e-06, 'learning_rate/action_model': 2.984446252241826e-05, 'epoch': 1.79}) +[2026-06-18 03:24:33,241][starVLA.training.train_starvla][INFO] - Step 4643, Loss: {'train/loss': 0.2214439958333969, 'train/grad_norm_pre_clip': 15.496119499206543, 'timing/data': 0.033216643147170544, 'timing/model': 1.2406557444483042, 'learning_rate/qwen_vl_interface': 2.982545642754083e-06, 'learning_rate/action_model': 2.9825456427540826e-05, 'epoch': 1.79}) +[2026-06-18 03:24:45,179][starVLA.training.train_starvla][INFO] - Step 4644, Loss: {'train/loss': 0.23012150824069977, 'train/grad_norm_pre_clip': 11.7332763671875, 'timing/data': 0.053095871582627296, 'timing/model': 4.23087733052671, 'learning_rate/qwen_vl_interface': 2.9806455033120697e-06, 'learning_rate/action_model': 2.9806455033120696e-05, 'epoch': 1.79}) +[2026-06-18 03:24:59,437][starVLA.training.train_starvla][INFO] - Step 4645, Loss: {'train/loss': 0.7369897365570068, 'train/grad_norm_pre_clip': 22.962064743041992, 'timing/data': 0.033873189240694046, 'timing/model': 1.5447676638141274, 'learning_rate/qwen_vl_interface': 2.978745834309685e-06, 'learning_rate/action_model': 2.9787458343096846e-05, 'epoch': 1.79}) +[2026-06-18 03:25:08,790][starVLA.training.train_starvla][INFO] - Step 4646, Loss: {'train/loss': 0.1474267542362213, 'train/grad_norm_pre_clip': 7.715342998504639, 'timing/data': 0.04587907809764147, 'timing/model': 1.5759131843224168, 'learning_rate/qwen_vl_interface': 2.9768466361407335e-06, 'learning_rate/action_model': 2.9768466361407334e-05, 'epoch': 1.79}) +[2026-06-18 03:25:20,732][starVLA.training.train_starvla][INFO] - Step 4647, Loss: {'train/loss': 0.30004966259002686, 'train/grad_norm_pre_clip': 21.36406707763672, 'timing/data': 0.042508047074079514, 'timing/model': 1.4808813473209739, 'learning_rate/qwen_vl_interface': 2.97494790919892e-06, 'learning_rate/action_model': 2.9749479091989197e-05, 'epoch': 1.79}) +[2026-06-18 03:25:26,443][starVLA.training.train_starvla][INFO] - Step 4648, Loss: {'train/loss': 0.5192787647247314, 'train/grad_norm_pre_clip': 17.76554298400879, 'timing/data': 0.0392792709171772, 'timing/model': 2.2092406880110502, 'learning_rate/qwen_vl_interface': 2.9730496538778524e-06, 'learning_rate/action_model': 2.9730496538778523e-05, 'epoch': 1.79}) +[2026-06-18 03:25:32,040][starVLA.training.train_starvla][INFO] - Step 4649, Loss: {'train/loss': 0.25252267718315125, 'train/grad_norm_pre_clip': 12.203557968139648, 'timing/data': 0.0321179423481226, 'timing/model': 1.0907208696007729, 'learning_rate/qwen_vl_interface': 2.9711518705710422e-06, 'learning_rate/action_model': 2.9711518705710424e-05, 'epoch': 1.79}) +[2026-06-18 03:25:36,854][starVLA.training.train_starvla][INFO] - Step 4650, Loss: {'train/loss': 0.3927825391292572, 'train/grad_norm_pre_clip': 12.84197998046875, 'timing/data': 0.03700848575681448, 'timing/model': 1.0941870613023639, 'learning_rate/qwen_vl_interface': 2.9692545596719014e-06, 'learning_rate/action_model': 2.969254559671901e-05, 'epoch': 1.79}) +[2026-06-18 03:25:42,854][starVLA.training.train_starvla][INFO] - Step 4651, Loss: {'train/loss': 0.45042139291763306, 'train/grad_norm_pre_clip': 11.621400833129883, 'timing/data': 0.037352997809648514, 'timing/model': 1.1999745285138488, 'learning_rate/qwen_vl_interface': 2.9673577215737454e-06, 'learning_rate/action_model': 2.9673577215737452e-05, 'epoch': 1.79}) +[2026-06-18 03:25:48,204][starVLA.training.train_starvla][INFO] - Step 4652, Loss: {'train/loss': 0.6404231190681458, 'train/grad_norm_pre_clip': 17.550931930541992, 'timing/data': 0.05344469752162695, 'timing/model': 1.6327222008258104, 'learning_rate/qwen_vl_interface': 2.9654613566697886e-06, 'learning_rate/action_model': 2.9654613566697887e-05, 'epoch': 1.79}) +[2026-06-18 03:25:57,704][starVLA.training.train_starvla][INFO] - Step 4653, Loss: {'train/loss': 0.5490934252738953, 'train/grad_norm_pre_clip': 19.53790283203125, 'timing/data': 0.055562809109687805, 'timing/model': 4.147364746779203, 'learning_rate/qwen_vl_interface': 2.9635654653531516e-06, 'learning_rate/action_model': 2.9635654653531515e-05, 'epoch': 1.79}) +[2026-06-18 03:26:07,813][starVLA.training.train_starvla][INFO] - Step 4654, Loss: {'train/loss': 0.28981199860572815, 'train/grad_norm_pre_clip': 11.351397514343262, 'timing/data': 0.03132070507854223, 'timing/model': 1.1616598470136523, 'learning_rate/qwen_vl_interface': 2.961670048016852e-06, 'learning_rate/action_model': 2.9616700480168518e-05, 'epoch': 1.79}) +[2026-06-18 03:26:14,820][starVLA.training.train_starvla][INFO] - Step 4655, Loss: {'train/loss': 0.3466411232948303, 'train/grad_norm_pre_clip': 13.425172805786133, 'timing/data': 0.040573193691670895, 'timing/model': 1.5624844413250685, 'learning_rate/qwen_vl_interface': 2.9597751050538177e-06, 'learning_rate/action_model': 2.9597751050538176e-05, 'epoch': 1.79}) +[2026-06-18 03:26:25,617][starVLA.training.train_starvla][INFO] - Step 4656, Loss: {'train/loss': 0.4020731449127197, 'train/grad_norm_pre_clip': 16.29515838623047, 'timing/data': 0.03658835031092167, 'timing/model': 4.629803997464478, 'learning_rate/qwen_vl_interface': 2.957880636856866e-06, 'learning_rate/action_model': 2.9578806368568655e-05, 'epoch': 1.79}) +[2026-06-18 03:26:37,851][starVLA.training.train_starvla][INFO] - Step 4657, Loss: {'train/loss': 0.41302722692489624, 'train/grad_norm_pre_clip': 19.024076461791992, 'timing/data': 0.09395082946866751, 'timing/model': 3.39214741345495, 'learning_rate/qwen_vl_interface': 2.955986643818723e-06, 'learning_rate/action_model': 2.9559866438187228e-05, 'epoch': 1.79}) +[2026-06-18 03:26:52,658][starVLA.training.train_starvla][INFO] - Step 4658, Loss: {'train/loss': 0.35572734475135803, 'train/grad_norm_pre_clip': 11.675346374511719, 'timing/data': 0.0371121009811759, 'timing/model': 3.110052957199514, 'learning_rate/qwen_vl_interface': 2.954093126332019e-06, 'learning_rate/action_model': 2.9540931263320194e-05, 'epoch': 1.79}) +[2026-06-18 03:27:00,982][starVLA.training.train_starvla][INFO] - Step 4659, Loss: {'train/loss': 0.3064368963241577, 'train/grad_norm_pre_clip': 11.526427268981934, 'timing/data': 0.03440672066062689, 'timing/model': 2.2277447767555714, 'learning_rate/qwen_vl_interface': 2.9522000847892807e-06, 'learning_rate/action_model': 2.9522000847892807e-05, 'epoch': 1.79}) +[2026-06-18 03:27:10,877][starVLA.training.train_starvla][INFO] - Step 4660, Loss: {'train/loss': 0.3607408106327057, 'train/grad_norm_pre_clip': 14.22298812866211, 'timing/data': 0.042298272252082825, 'timing/model': 1.1583391446620226, 'learning_rate/qwen_vl_interface': 2.9503075195829393e-06, 'learning_rate/action_model': 2.9503075195829393e-05, 'epoch': 1.79}) +[2026-06-18 03:27:16,209][starVLA.training.train_starvla][INFO] - Step 4661, Loss: {'train/loss': 0.19052289426326752, 'train/grad_norm_pre_clip': 8.43035888671875, 'timing/data': 0.03897538781166077, 'timing/model': 1.1440639551728964, 'learning_rate/qwen_vl_interface': 2.9484154311053203e-06, 'learning_rate/action_model': 2.94841543110532e-05, 'epoch': 1.79}) +[2026-06-18 03:27:21,751][starVLA.training.train_starvla][INFO] - Step 4662, Loss: {'train/loss': 0.1923222839832306, 'train/grad_norm_pre_clip': 10.504261016845703, 'timing/data': 0.03450552187860012, 'timing/model': 1.6600259728729725, 'learning_rate/qwen_vl_interface': 2.94652381974866e-06, 'learning_rate/action_model': 2.94652381974866e-05, 'epoch': 1.79}) +[2026-06-18 03:27:26,716][starVLA.training.train_starvla][INFO] - Step 4663, Loss: {'train/loss': 0.5453824400901794, 'train/grad_norm_pre_clip': 15.59213924407959, 'timing/data': 0.03842603974044323, 'timing/model': 1.082906779833138, 'learning_rate/qwen_vl_interface': 2.9446326859050916e-06, 'learning_rate/action_model': 2.9446326859050916e-05, 'epoch': 1.79}) +[2026-06-18 03:27:31,721][starVLA.training.train_starvla][INFO] - Step 4664, Loss: {'train/loss': 0.25039246678352356, 'train/grad_norm_pre_clip': 13.299568176269531, 'timing/data': 0.07582023832947016, 'timing/model': 1.1997926961630583, 'learning_rate/qwen_vl_interface': 2.9427420299666472e-06, 'learning_rate/action_model': 2.942742029966647e-05, 'epoch': 1.79}) +[2026-06-18 03:27:36,848][starVLA.training.train_starvla][INFO] - Step 4665, Loss: {'train/loss': 0.6003127694129944, 'train/grad_norm_pre_clip': 21.00663948059082, 'timing/data': 0.034811485558748245, 'timing/model': 1.2622979581356049, 'learning_rate/qwen_vl_interface': 2.940851852325263e-06, 'learning_rate/action_model': 2.9408518523252632e-05, 'epoch': 1.79}) +[2026-06-18 03:27:42,719][starVLA.training.train_starvla][INFO] - Step 4666, Loss: {'train/loss': 0.40003475546836853, 'train/grad_norm_pre_clip': 15.619393348693848, 'timing/data': 0.03433556202799082, 'timing/model': 1.8429678035899997, 'learning_rate/qwen_vl_interface': 2.9389621533727745e-06, 'learning_rate/action_model': 2.9389621533727747e-05, 'epoch': 1.79}) +[2026-06-18 03:27:47,986][starVLA.training.train_starvla][INFO] - Step 4667, Loss: {'train/loss': 0.6370075345039368, 'train/grad_norm_pre_clip': 15.3881254196167, 'timing/data': 0.033361163921654224, 'timing/model': 1.807379661127925, 'learning_rate/qwen_vl_interface': 2.9370729335009185e-06, 'learning_rate/action_model': 2.9370729335009184e-05, 'epoch': 1.8}) +[2026-06-18 03:27:53,006][starVLA.training.train_starvla][INFO] - Step 4668, Loss: {'train/loss': 0.2429894357919693, 'train/grad_norm_pre_clip': 11.186563491821289, 'timing/data': 0.034398576244711876, 'timing/model': 1.1465129302814603, 'learning_rate/qwen_vl_interface': 2.93518419310133e-06, 'learning_rate/action_model': 2.93518419310133e-05, 'epoch': 1.8}) +[2026-06-18 03:27:58,114][starVLA.training.train_starvla][INFO] - Step 4669, Loss: {'train/loss': 0.6071107983589172, 'train/grad_norm_pre_clip': 23.423114776611328, 'timing/data': 0.03223817143589258, 'timing/model': 1.1074057593941689, 'learning_rate/qwen_vl_interface': 2.9332959325655537e-06, 'learning_rate/action_model': 2.9332959325655535e-05, 'epoch': 1.8}) +[2026-06-18 03:28:03,252][starVLA.training.train_starvla][INFO] - Step 4670, Loss: {'train/loss': 0.2528552711009979, 'train/grad_norm_pre_clip': 10.698612213134766, 'timing/data': 0.03347182646393776, 'timing/model': 1.199474279768765, 'learning_rate/qwen_vl_interface': 2.9314081522850206e-06, 'learning_rate/action_model': 2.9314081522850205e-05, 'epoch': 1.8}) +[2026-06-18 03:28:08,283][starVLA.training.train_starvla][INFO] - Step 4671, Loss: {'train/loss': 0.32399141788482666, 'train/grad_norm_pre_clip': 15.460977554321289, 'timing/data': 0.03830742742866278, 'timing/model': 1.510840074159205, 'learning_rate/qwen_vl_interface': 2.9295208526510726e-06, 'learning_rate/action_model': 2.9295208526510725e-05, 'epoch': 1.8}) +[2026-06-18 03:28:13,963][starVLA.training.train_starvla][INFO] - Step 4672, Loss: {'train/loss': 0.3250516355037689, 'train/grad_norm_pre_clip': 10.19950008392334, 'timing/data': 0.03708635177463293, 'timing/model': 1.67760696914047, 'learning_rate/qwen_vl_interface': 2.9276340340549476e-06, 'learning_rate/action_model': 2.9276340340549473e-05, 'epoch': 1.8}) +[2026-06-18 03:28:18,263][starVLA.training.train_starvla][INFO] - Step 4673, Loss: {'train/loss': 0.4898907542228699, 'train/grad_norm_pre_clip': 16.466650009155273, 'timing/data': 0.03404711838811636, 'timing/model': 1.0753543488681316, 'learning_rate/qwen_vl_interface': 2.9257476968877875e-06, 'learning_rate/action_model': 2.9257476968877874e-05, 'epoch': 1.8}) +[2026-06-18 03:28:23,135][starVLA.training.train_starvla][INFO] - Step 4674, Loss: {'train/loss': 0.3161318898200989, 'train/grad_norm_pre_clip': 12.546221733093262, 'timing/data': 0.040795234963297844, 'timing/model': 1.1205656314268708, 'learning_rate/qwen_vl_interface': 2.923861841540631e-06, 'learning_rate/action_model': 2.923861841540631e-05, 'epoch': 1.8}) +[2026-06-18 03:28:28,614][starVLA.training.train_starvla][INFO] - Step 4675, Loss: {'train/loss': 0.348524808883667, 'train/grad_norm_pre_clip': 10.20692253112793, 'timing/data': 0.03128734044730663, 'timing/model': 1.7209221702069044, 'learning_rate/qwen_vl_interface': 2.9219764684044196e-06, 'learning_rate/action_model': 2.9219764684044193e-05, 'epoch': 1.8}) +[2026-06-18 03:28:38,190][starVLA.training.train_starvla][INFO] - Step 4676, Loss: {'train/loss': 0.3864839971065521, 'train/grad_norm_pre_clip': 13.754053115844727, 'timing/data': 0.03363672085106373, 'timing/model': 2.9940235083922744, 'learning_rate/qwen_vl_interface': 2.920091577869988e-06, 'learning_rate/action_model': 2.9200915778699877e-05, 'epoch': 1.8}) +[2026-06-18 03:28:51,580][starVLA.training.train_starvla][INFO] - Step 4677, Loss: {'train/loss': 0.46648108959198, 'train/grad_norm_pre_clip': 11.570591926574707, 'timing/data': 0.03979859221726656, 'timing/model': 5.0294205620884895, 'learning_rate/qwen_vl_interface': 2.918207170328081e-06, 'learning_rate/action_model': 2.918207170328081e-05, 'epoch': 1.8}) +[2026-06-18 03:29:03,711][starVLA.training.train_starvla][INFO] - Step 4678, Loss: {'train/loss': 0.5745552778244019, 'train/grad_norm_pre_clip': 16.381851196289062, 'timing/data': 0.035425673238933086, 'timing/model': 1.2516716215759516, 'learning_rate/qwen_vl_interface': 2.9163232461693373e-06, 'learning_rate/action_model': 2.9163232461693373e-05, 'epoch': 1.8}) +[2026-06-18 03:29:11,590][starVLA.training.train_starvla][INFO] - Step 4679, Loss: {'train/loss': 0.6135610938072205, 'train/grad_norm_pre_clip': 15.976677894592285, 'timing/data': 0.041318478994071484, 'timing/model': 1.3383284285664558, 'learning_rate/qwen_vl_interface': 2.9144398057842948e-06, 'learning_rate/action_model': 2.9144398057842948e-05, 'epoch': 1.8}) +[2026-06-18 03:29:16,508][starVLA.training.train_starvla][INFO] - Step 4680, Loss: {'train/loss': 0.40348026156425476, 'train/grad_norm_pre_clip': 15.25667953491211, 'timing/data': 0.03180443309247494, 'timing/model': 1.2089573023840785, 'learning_rate/qwen_vl_interface': 2.912556849563394e-06, 'learning_rate/action_model': 2.9125568495633943e-05, 'epoch': 1.8}) +[2026-06-18 03:29:28,501][starVLA.training.train_starvla][INFO] - Step 4681, Loss: {'train/loss': 0.11849412322044373, 'train/grad_norm_pre_clip': 7.50986909866333, 'timing/data': 0.05627071298658848, 'timing/model': 1.6914681252092123, 'learning_rate/qwen_vl_interface': 2.910674377896974e-06, 'learning_rate/action_model': 2.910674377896974e-05, 'epoch': 1.8}) +[2026-06-18 03:29:33,284][starVLA.training.train_starvla][INFO] - Step 4682, Loss: {'train/loss': 0.6054848432540894, 'train/grad_norm_pre_clip': 15.684183120727539, 'timing/data': 0.03484839852899313, 'timing/model': 1.4736403003335, 'learning_rate/qwen_vl_interface': 2.9087923911752723e-06, 'learning_rate/action_model': 2.9087923911752725e-05, 'epoch': 1.8}) +[2026-06-18 03:29:42,019][starVLA.training.train_starvla][INFO] - Step 4683, Loss: {'train/loss': 0.5443955063819885, 'train/grad_norm_pre_clip': 18.33937644958496, 'timing/data': 0.041015210561454296, 'timing/model': 1.1042455965653062, 'learning_rate/qwen_vl_interface': 2.906910889788425e-06, 'learning_rate/action_model': 2.906910889788425e-05, 'epoch': 1.8}) +[2026-06-18 03:29:46,836][starVLA.training.train_starvla][INFO] - Step 4684, Loss: {'train/loss': 0.3674984872341156, 'train/grad_norm_pre_clip': 12.268678665161133, 'timing/data': 0.04609955381602049, 'timing/model': 1.1018348783254623, 'learning_rate/qwen_vl_interface': 2.905029874126476e-06, 'learning_rate/action_model': 2.905029874126476e-05, 'epoch': 1.8}) +[2026-06-18 03:29:51,520][starVLA.training.train_starvla][INFO] - Step 4685, Loss: {'train/loss': 0.18724484741687775, 'train/grad_norm_pre_clip': 11.897991180419922, 'timing/data': 0.052456555888056755, 'timing/model': 1.102802955545485, 'learning_rate/qwen_vl_interface': 2.9031493445793558e-06, 'learning_rate/action_model': 2.9031493445793558e-05, 'epoch': 1.8}) +[2026-06-18 03:29:56,735][starVLA.training.train_starvla][INFO] - Step 4686, Loss: {'train/loss': 0.5310998558998108, 'train/grad_norm_pre_clip': 15.140231132507324, 'timing/data': 0.042986126616597176, 'timing/model': 1.4412444615736604, 'learning_rate/qwen_vl_interface': 2.9012693015369007e-06, 'learning_rate/action_model': 2.9012693015369008e-05, 'epoch': 1.8}) +[2026-06-18 03:30:02,149][starVLA.training.train_starvla][INFO] - Step 4687, Loss: {'train/loss': 0.27221447229385376, 'train/grad_norm_pre_clip': 10.59037971496582, 'timing/data': 0.04634595848619938, 'timing/model': 1.6751498160883784, 'learning_rate/qwen_vl_interface': 2.8993897453888454e-06, 'learning_rate/action_model': 2.899389745388845e-05, 'epoch': 1.8}) +[2026-06-18 03:30:06,942][starVLA.training.train_starvla][INFO] - Step 4688, Loss: {'train/loss': 0.27022066712379456, 'train/grad_norm_pre_clip': 11.271191596984863, 'timing/data': 0.0333594586700201, 'timing/model': 1.1994847618043423, 'learning_rate/qwen_vl_interface': 2.8975106765248267e-06, 'learning_rate/action_model': 2.8975106765248267e-05, 'epoch': 1.8}) +[2026-06-18 03:30:12,100][starVLA.training.train_starvla][INFO] - Step 4689, Loss: {'train/loss': 0.6181754469871521, 'train/grad_norm_pre_clip': 33.12582778930664, 'timing/data': 0.03145624045282602, 'timing/model': 1.078944580629468, 'learning_rate/qwen_vl_interface': 2.895632095334378e-06, 'learning_rate/action_model': 2.895632095334378e-05, 'epoch': 1.8}) +[2026-06-18 03:30:16,849][starVLA.training.train_starvla][INFO] - Step 4690, Loss: {'train/loss': 0.35059216618537903, 'train/grad_norm_pre_clip': 10.419126510620117, 'timing/data': 0.031896903179585934, 'timing/model': 1.100940503180027, 'learning_rate/qwen_vl_interface': 2.8937540022069267e-06, 'learning_rate/action_model': 2.8937540022069265e-05, 'epoch': 1.8}) +[2026-06-18 03:30:22,238][starVLA.training.train_starvla][INFO] - Step 4691, Loss: {'train/loss': 0.24976055324077606, 'train/grad_norm_pre_clip': 11.380528450012207, 'timing/data': 0.047461237758398056, 'timing/model': 1.6039418540894985, 'learning_rate/qwen_vl_interface': 2.8918763975318076e-06, 'learning_rate/action_model': 2.8918763975318076e-05, 'epoch': 1.8}) +[2026-06-18 03:30:27,511][starVLA.training.train_starvla][INFO] - Step 4692, Loss: {'train/loss': 0.3416956067085266, 'train/grad_norm_pre_clip': 14.074243545532227, 'timing/data': 0.03637915197759867, 'timing/model': 1.6588255474343896, 'learning_rate/qwen_vl_interface': 2.889999281698248e-06, 'learning_rate/action_model': 2.889999281698248e-05, 'epoch': 1.8}) +[2026-06-18 03:30:33,732][starVLA.training.train_starvla][INFO] - Step 4693, Loss: {'train/loss': 0.5411935448646545, 'train/grad_norm_pre_clip': 15.587911605834961, 'timing/data': 0.03709776513278484, 'timing/model': 1.2786401640623808, 'learning_rate/qwen_vl_interface': 2.8881226550953772e-06, 'learning_rate/action_model': 2.888122655095377e-05, 'epoch': 1.81}) +[2026-06-18 03:30:38,610][starVLA.training.train_starvla][INFO] - Step 4694, Loss: {'train/loss': 0.43415018916130066, 'train/grad_norm_pre_clip': 18.683406829833984, 'timing/data': 0.045919934287667274, 'timing/model': 1.1814170610159636, 'learning_rate/qwen_vl_interface': 2.8862465181122212e-06, 'learning_rate/action_model': 2.886246518112221e-05, 'epoch': 1.81}) +[2026-06-18 03:30:43,527][starVLA.training.train_starvla][INFO] - Step 4695, Loss: {'train/loss': 0.2447233647108078, 'train/grad_norm_pre_clip': 22.175676345825195, 'timing/data': 0.03415077645331621, 'timing/model': 1.0779460920020938, 'learning_rate/qwen_vl_interface': 2.884370871137706e-06, 'learning_rate/action_model': 2.884370871137706e-05, 'epoch': 1.81}) +[2026-06-18 03:30:49,171][starVLA.training.train_starvla][INFO] - Step 4696, Loss: {'train/loss': 0.642889142036438, 'train/grad_norm_pre_clip': 22.768413543701172, 'timing/data': 0.03445730730891228, 'timing/model': 1.1455710362643003, 'learning_rate/qwen_vl_interface': 2.8824957145606548e-06, 'learning_rate/action_model': 2.8824957145606548e-05, 'epoch': 1.81}) +[2026-06-18 03:30:59,485][starVLA.training.train_starvla][INFO] - Step 4697, Loss: {'train/loss': 0.4968639314174652, 'train/grad_norm_pre_clip': 15.457118034362793, 'timing/data': 0.04734954796731472, 'timing/model': 2.0874804379418492, 'learning_rate/qwen_vl_interface': 2.8806210487697866e-06, 'learning_rate/action_model': 2.8806210487697867e-05, 'epoch': 1.81}) +[2026-06-18 03:31:04,439][starVLA.training.train_starvla][INFO] - Step 4698, Loss: {'train/loss': 0.8354367613792419, 'train/grad_norm_pre_clip': 22.584760665893555, 'timing/data': 0.05840701796114445, 'timing/model': 1.2704648254439235, 'learning_rate/qwen_vl_interface': 2.87874687415373e-06, 'learning_rate/action_model': 2.87874687415373e-05, 'epoch': 1.81}) +[2026-06-18 03:31:10,770][starVLA.training.train_starvla][INFO] - Step 4699, Loss: {'train/loss': 0.23909999430179596, 'train/grad_norm_pre_clip': 8.42447566986084, 'timing/data': 0.03733794018626213, 'timing/model': 1.4734340822324157, 'learning_rate/qwen_vl_interface': 2.876873191100995e-06, 'learning_rate/action_model': 2.8768731911009948e-05, 'epoch': 1.81}) +[2026-06-18 03:34:58,413][starVLA.training.train_starvla][INFO] - Step 4700, Loss: {'train/loss': 0.5879411697387695, 'train/grad_norm_pre_clip': 19.61932373046875, 'eval/loss': 0.43081989884376526, 'eval/latency_4/loss': 0.43081989884376526, 'eval/demon_fire_precision': 0.9606547056971986, 'eval/demon_fire_recall': 0.9912309191295875, 'eval/demon_fire_f1': 0.975703324808184, 'eval/demon_fire_rate_model': 0.9330396475770925, 'eval/demon_fire_rate_teacher': 0.9042584434654919, 'eval/demon_left_precision': 0.7976878612716763, 'eval/demon_left_recall': 0.6301369863013698, 'eval/demon_left_f1': 0.7040816326530612, 'eval/demon_left_rate_model': 0.10161527165932452, 'eval/demon_left_rate_teacher': 0.12863436123348018, 'eval/demon_right_precision': 0.8227091633466136, 'eval/demon_right_recall': 0.7454873646209387, 'eval/demon_right_f1': 0.7821969696969697, 'eval/demon_right_rate_model': 0.1474302496328928, 'eval/demon_right_rate_teacher': 0.16270190895741557, 'eval/demon_move_f1': 0.7431393011750155, 'eval/demon_cc_f1': 0.8919619853649674, 'eval/latency_4/demon_fire_precision': 0.9606547056971986, 'eval/latency_4/demon_fire_recall': 0.9912309191295875, 'eval/latency_4/demon_fire_f1': 0.975703324808184, 'eval/latency_4/demon_fire_rate_model': 0.9330396475770925, 'eval/latency_4/demon_fire_rate_teacher': 0.9042584434654919, 'eval/latency_4/demon_left_precision': 0.7976878612716763, 'eval/latency_4/demon_left_recall': 0.6301369863013698, 'eval/latency_4/demon_left_f1': 0.7040816326530612, 'eval/latency_4/demon_left_rate_model': 0.10161527165932452, 'eval/latency_4/demon_left_rate_teacher': 0.12863436123348018, 'eval/latency_4/demon_right_precision': 0.8227091633466136, 'eval/latency_4/demon_right_recall': 0.7454873646209387, 'eval/latency_4/demon_right_f1': 0.7821969696969697, 'eval/latency_4/demon_right_rate_model': 0.1474302496328928, 'eval/latency_4/demon_right_rate_teacher': 0.16270190895741557, 'eval/latency_4/demon_move_f1': 0.7431393011750155, 'eval/latency_4/demon_cc_f1': 0.8919619853649674, 'timing/data': 0.0519937239587307, 'timing/model': 1.207452530041337, 'learning_rate/qwen_vl_interface': 2.8750000000000013e-06, 'learning_rate/action_model': 2.875000000000001e-05, 'epoch': 1.81}) +[2026-06-18 03:34:58,415][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_4700_state +[2026-06-18 03:34:58,421][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 03:35:06,274][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_4700_state/model.safetensors +[2026-06-18 03:35:26,596][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_4700_state/optimizer.bin +[2026-06-18 03:35:26,772][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_4700_state/sampler.bin +[2026-06-18 03:35:26,772][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_4700_state/sampler_1.bin +[2026-06-18 03:35:26,776][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_4700_state/random_states_0.pkl +[2026-06-18 03:35:26,780][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4700 +[2026-06-18 03:35:26,781][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 03:35:26,798][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 03:35:32,493][starVLA.training.train_starvla][INFO] - Step 4701, Loss: {'train/loss': 0.21196751296520233, 'train/grad_norm_pre_clip': 9.697665214538574, 'timing/data': 0.0397317111492157, 'timing/model': 1.1700182119384408, 'learning_rate/qwen_vl_interface': 2.8731273012390614e-06, 'learning_rate/action_model': 2.8731273012390615e-05, 'epoch': 1.81}) +[2026-06-18 03:35:45,658][starVLA.training.train_starvla][INFO] - Step 4702, Loss: {'train/loss': 0.23224063217639923, 'train/grad_norm_pre_clip': 10.926715850830078, 'timing/data': 0.03791645169258118, 'timing/model': 2.073452992364764, 'learning_rate/qwen_vl_interface': 2.8712550952063912e-06, 'learning_rate/action_model': 2.871255095206391e-05, 'epoch': 1.81}) +[2026-06-18 03:35:52,965][starVLA.training.train_starvla][INFO] - Step 4703, Loss: {'train/loss': 0.2034711390733719, 'train/grad_norm_pre_clip': 11.06064510345459, 'timing/data': 0.0413186252117157, 'timing/model': 1.1486983532086015, 'learning_rate/qwen_vl_interface': 2.8693833822901012e-06, 'learning_rate/action_model': 2.8693833822901012e-05, 'epoch': 1.81}) +[2026-06-18 03:36:00,484][starVLA.training.train_starvla][INFO] - Step 4704, Loss: {'train/loss': 0.21746988594532013, 'train/grad_norm_pre_clip': 12.46594524383545, 'timing/data': 0.05553815979510546, 'timing/model': 3.3981961319223046, 'learning_rate/qwen_vl_interface': 2.867512162878196e-06, 'learning_rate/action_model': 2.867512162878196e-05, 'epoch': 1.81}) +[2026-06-18 03:36:05,716][starVLA.training.train_starvla][INFO] - Step 4705, Loss: {'train/loss': 0.19660358130931854, 'train/grad_norm_pre_clip': 10.616369247436523, 'timing/data': 0.05947302281856537, 'timing/model': 1.1747199269011617, 'learning_rate/qwen_vl_interface': 2.8656414373585812e-06, 'learning_rate/action_model': 2.865641437358581e-05, 'epoch': 1.81}) +[2026-06-18 03:36:11,862][starVLA.training.train_starvla][INFO] - Step 4706, Loss: {'train/loss': 0.4022798538208008, 'train/grad_norm_pre_clip': 12.662566184997559, 'timing/data': 0.04690743889659643, 'timing/model': 1.3401793567463756, 'learning_rate/qwen_vl_interface': 2.8637712061190644e-06, 'learning_rate/action_model': 2.863771206119064e-05, 'epoch': 1.81}) +[2026-06-18 03:36:17,330][starVLA.training.train_starvla][INFO] - Step 4707, Loss: {'train/loss': 0.22069495916366577, 'train/grad_norm_pre_clip': 8.548484802246094, 'timing/data': 0.04352065175771713, 'timing/model': 1.513575498946011, 'learning_rate/qwen_vl_interface': 2.861901469547344e-06, 'learning_rate/action_model': 2.861901469547344e-05, 'epoch': 1.81}) +[2026-06-18 03:36:23,532][starVLA.training.train_starvla][INFO] - Step 4708, Loss: {'train/loss': 0.2776854336261749, 'train/grad_norm_pre_clip': 10.694549560546875, 'timing/data': 0.04300453141331673, 'timing/model': 1.1905970564112067, 'learning_rate/qwen_vl_interface': 2.8600322280310185e-06, 'learning_rate/action_model': 2.8600322280310183e-05, 'epoch': 1.81}) +[2026-06-18 03:36:29,723][starVLA.training.train_starvla][INFO] - Step 4709, Loss: {'train/loss': 0.45884692668914795, 'train/grad_norm_pre_clip': 12.556305885314941, 'timing/data': 0.03990272432565689, 'timing/model': 1.3658641362562776, 'learning_rate/qwen_vl_interface': 2.8581634819575843e-06, 'learning_rate/action_model': 2.8581634819575843e-05, 'epoch': 1.81}) +[2026-06-18 03:36:35,703][starVLA.training.train_starvla][INFO] - Step 4710, Loss: {'train/loss': 0.47965842485427856, 'train/grad_norm_pre_clip': 12.276835441589355, 'timing/data': 0.048616708256304264, 'timing/model': 1.3386045601218939, 'learning_rate/qwen_vl_interface': 2.856295231714434e-06, 'learning_rate/action_model': 2.8562952317144337e-05, 'epoch': 1.81}) +[2026-06-18 03:36:42,122][starVLA.training.train_starvla][INFO] - Step 4711, Loss: {'train/loss': 0.2556956112384796, 'train/grad_norm_pre_clip': 13.871614456176758, 'timing/data': 0.034295245073735714, 'timing/model': 1.7143117049708962, 'learning_rate/qwen_vl_interface': 2.854427477688857e-06, 'learning_rate/action_model': 2.854427477688857e-05, 'epoch': 1.81}) +[2026-06-18 03:36:47,428][starVLA.training.train_starvla][INFO] - Step 4712, Loss: {'train/loss': 0.39477819204330444, 'train/grad_norm_pre_clip': 13.155022621154785, 'timing/data': 0.03977557364851236, 'timing/model': 1.6083845915272832, 'learning_rate/qwen_vl_interface': 2.8525602202680424e-06, 'learning_rate/action_model': 2.8525602202680424e-05, 'epoch': 1.81}) +[2026-06-18 03:36:56,039][starVLA.training.train_starvla][INFO] - Step 4713, Loss: {'train/loss': 0.12285047769546509, 'train/grad_norm_pre_clip': 7.303382873535156, 'timing/data': 0.037062316201627254, 'timing/model': 1.6444470332935452, 'learning_rate/qwen_vl_interface': 2.8506934598390736e-06, 'learning_rate/action_model': 2.8506934598390735e-05, 'epoch': 1.81}) +[2026-06-18 03:37:04,681][starVLA.training.train_starvla][INFO] - Step 4714, Loss: {'train/loss': 0.40850305557250977, 'train/grad_norm_pre_clip': 12.782870292663574, 'timing/data': 0.046634295023977757, 'timing/model': 1.579683007672429, 'learning_rate/qwen_vl_interface': 2.848827196788932e-06, 'learning_rate/action_model': 2.848827196788932e-05, 'epoch': 1.81}) +[2026-06-18 03:37:10,989][starVLA.training.train_starvla][INFO] - Step 4715, Loss: {'train/loss': 0.17140153050422668, 'train/grad_norm_pre_clip': 19.87126350402832, 'timing/data': 0.04021295811980963, 'timing/model': 1.6127114482223988, 'learning_rate/qwen_vl_interface': 2.8469614315044975e-06, 'learning_rate/action_model': 2.8469614315044975e-05, 'epoch': 1.81}) +[2026-06-18 03:37:16,411][starVLA.training.train_starvla][INFO] - Step 4716, Loss: {'train/loss': 0.6738426685333252, 'train/grad_norm_pre_clip': 22.99442481994629, 'timing/data': 0.041165570728480816, 'timing/model': 1.5727559169754386, 'learning_rate/qwen_vl_interface': 2.8450961643725405e-06, 'learning_rate/action_model': 2.8450961643725405e-05, 'epoch': 1.81}) +[2026-06-18 03:37:23,559][starVLA.training.train_starvla][INFO] - Step 4717, Loss: {'train/loss': 0.4915694296360016, 'train/grad_norm_pre_clip': 12.900496482849121, 'timing/data': 0.042424049228429794, 'timing/model': 2.0003974558785558, 'learning_rate/qwen_vl_interface': 2.8432313957797415e-06, 'learning_rate/action_model': 2.8432313957797414e-05, 'epoch': 1.81}) +[2026-06-18 03:37:29,016][starVLA.training.train_starvla][INFO] - Step 4718, Loss: {'train/loss': 0.28032079339027405, 'train/grad_norm_pre_clip': 11.713583946228027, 'timing/data': 0.044468360021710396, 'timing/model': 1.184777069836855, 'learning_rate/qwen_vl_interface': 2.8413671261126613e-06, 'learning_rate/action_model': 2.8413671261126613e-05, 'epoch': 1.81}) +[2026-06-18 03:37:35,600][starVLA.training.train_starvla][INFO] - Step 4719, Loss: {'train/loss': 0.23744407296180725, 'train/grad_norm_pre_clip': 13.836553573608398, 'timing/data': 0.035314541310071945, 'timing/model': 2.3565087039023638, 'learning_rate/qwen_vl_interface': 2.8395033557577668e-06, 'learning_rate/action_model': 2.8395033557577668e-05, 'epoch': 1.82}) +[2026-06-18 03:37:48,711][starVLA.training.train_starvla][INFO] - Step 4720, Loss: {'train/loss': 0.45338544249534607, 'train/grad_norm_pre_clip': 17.76116371154785, 'timing/data': 0.09698150958865881, 'timing/model': 1.2637816350907087, 'learning_rate/qwen_vl_interface': 2.8376400851014214e-06, 'learning_rate/action_model': 2.8376400851014214e-05, 'epoch': 1.82}) +[2026-06-18 03:37:55,482][starVLA.training.train_starvla][INFO] - Step 4721, Loss: {'train/loss': 0.721604585647583, 'train/grad_norm_pre_clip': 15.80267333984375, 'timing/data': 0.043673260137438774, 'timing/model': 2.39915073197335, 'learning_rate/qwen_vl_interface': 2.8357773145298823e-06, 'learning_rate/action_model': 2.8357773145298823e-05, 'epoch': 1.82}) +[2026-06-18 03:38:03,143][starVLA.training.train_starvla][INFO] - Step 4722, Loss: {'train/loss': 0.3566321134567261, 'train/grad_norm_pre_clip': 16.255416870117188, 'timing/data': 0.04021692834794521, 'timing/model': 1.773347707465291, 'learning_rate/qwen_vl_interface': 2.8339150444293055e-06, 'learning_rate/action_model': 2.8339150444293055e-05, 'epoch': 1.82}) +[2026-06-18 03:38:08,568][starVLA.training.train_starvla][INFO] - Step 4723, Loss: {'train/loss': 0.3819510340690613, 'train/grad_norm_pre_clip': 12.671046257019043, 'timing/data': 0.05187139753252268, 'timing/model': 1.3993331799283624, 'learning_rate/qwen_vl_interface': 2.8320532751857365e-06, 'learning_rate/action_model': 2.8320532751857366e-05, 'epoch': 1.82}) +[2026-06-18 03:38:15,337][starVLA.training.train_starvla][INFO] - Step 4724, Loss: {'train/loss': 0.22505807876586914, 'train/grad_norm_pre_clip': 9.155232429504395, 'timing/data': 0.036935752257704735, 'timing/model': 1.223216649144888, 'learning_rate/qwen_vl_interface': 2.830192007185127e-06, 'learning_rate/action_model': 2.830192007185127e-05, 'epoch': 1.82}) +[2026-06-18 03:38:29,221][starVLA.training.train_starvla][INFO] - Step 4725, Loss: {'train/loss': 0.3968695402145386, 'train/grad_norm_pre_clip': 13.715422630310059, 'timing/data': 0.043982818722724915, 'timing/model': 4.660577963106334, 'learning_rate/qwen_vl_interface': 2.828331240813318e-06, 'learning_rate/action_model': 2.8283312408133176e-05, 'epoch': 1.82}) +[2026-06-18 03:38:37,012][starVLA.training.train_starvla][INFO] - Step 4726, Loss: {'train/loss': 0.40570634603500366, 'train/grad_norm_pre_clip': 13.55313777923584, 'timing/data': 0.04056366626173258, 'timing/model': 2.041765371337533, 'learning_rate/qwen_vl_interface': 2.826470976456049e-06, 'learning_rate/action_model': 2.826470976456049e-05, 'epoch': 1.82}) +[2026-06-18 03:38:42,558][starVLA.training.train_starvla][INFO] - Step 4727, Loss: {'train/loss': 0.41879305243492126, 'train/grad_norm_pre_clip': 15.749811172485352, 'timing/data': 0.04683588910847902, 'timing/model': 1.5150648839771748, 'learning_rate/qwen_vl_interface': 2.8246112144989548e-06, 'learning_rate/action_model': 2.8246112144989543e-05, 'epoch': 1.82}) +[2026-06-18 03:38:49,546][starVLA.training.train_starvla][INFO] - Step 4728, Loss: {'train/loss': 0.7861392498016357, 'train/grad_norm_pre_clip': 16.238115310668945, 'timing/data': 0.04035687446594238, 'timing/model': 3.078668004833162, 'learning_rate/qwen_vl_interface': 2.8227519553275644e-06, 'learning_rate/action_model': 2.8227519553275645e-05, 'epoch': 1.82}) +[2026-06-18 03:39:06,524][starVLA.training.train_starvla][INFO] - Step 4729, Loss: {'train/loss': 0.41369035840034485, 'train/grad_norm_pre_clip': 14.901001930236816, 'timing/data': 0.0477239741012454, 'timing/model': 5.4689546106383204, 'learning_rate/qwen_vl_interface': 2.8208931993273066e-06, 'learning_rate/action_model': 2.8208931993273062e-05, 'epoch': 1.82}) +[2026-06-18 03:39:15,758][starVLA.training.train_starvla][INFO] - Step 4730, Loss: {'train/loss': 0.43383049964904785, 'train/grad_norm_pre_clip': 17.419492721557617, 'timing/data': 0.034846194088459015, 'timing/model': 1.1267865588888526, 'learning_rate/qwen_vl_interface': 2.8190349468835006e-06, 'learning_rate/action_model': 2.8190349468835004e-05, 'epoch': 1.82}) +[2026-06-18 03:39:22,433][starVLA.training.train_starvla][INFO] - Step 4731, Loss: {'train/loss': 0.5419262647628784, 'train/grad_norm_pre_clip': 14.418659210205078, 'timing/data': 0.04544217977672815, 'timing/model': 1.8762079514563084, 'learning_rate/qwen_vl_interface': 2.817177198381369e-06, 'learning_rate/action_model': 2.817177198381369e-05, 'epoch': 1.82}) +[2026-06-18 03:39:36,624][starVLA.training.train_starvla][INFO] - Step 4732, Loss: {'train/loss': 0.292186975479126, 'train/grad_norm_pre_clip': 11.981988906860352, 'timing/data': 0.04382005147635937, 'timing/model': 2.332909361459315, 'learning_rate/qwen_vl_interface': 2.8153199542060206e-06, 'learning_rate/action_model': 2.8153199542060206e-05, 'epoch': 1.82}) +[2026-06-18 03:39:41,821][starVLA.training.train_starvla][INFO] - Step 4733, Loss: {'train/loss': 0.4513717591762543, 'train/grad_norm_pre_clip': 16.80897331237793, 'timing/data': 0.0448553841561079, 'timing/model': 1.1704273661598563, 'learning_rate/qwen_vl_interface': 2.813463214742465e-06, 'learning_rate/action_model': 2.813463214742465e-05, 'epoch': 1.82}) +[2026-06-18 03:39:47,508][starVLA.training.train_starvla][INFO] - Step 4734, Loss: {'train/loss': 0.4117574989795685, 'train/grad_norm_pre_clip': 16.08416175842285, 'timing/data': 0.041748808696866035, 'timing/model': 1.239264608360827, 'learning_rate/qwen_vl_interface': 2.8116069803756047e-06, 'learning_rate/action_model': 2.8116069803756047e-05, 'epoch': 1.82}) +[2026-06-18 03:39:53,531][starVLA.training.train_starvla][INFO] - Step 4735, Loss: {'train/loss': 0.14137844741344452, 'train/grad_norm_pre_clip': 6.812833786010742, 'timing/data': 0.03944140486419201, 'timing/model': 1.441162663511932, 'learning_rate/qwen_vl_interface': 2.809751251490244e-06, 'learning_rate/action_model': 2.809751251490244e-05, 'epoch': 1.82}) +[2026-06-18 03:39:59,090][starVLA.training.train_starvla][INFO] - Step 4736, Loss: {'train/loss': 0.14500021934509277, 'train/grad_norm_pre_clip': 8.75721263885498, 'timing/data': 0.04367372766137123, 'timing/model': 1.825377106666565, 'learning_rate/qwen_vl_interface': 2.807896028471076e-06, 'learning_rate/action_model': 2.807896028471076e-05, 'epoch': 1.82}) +[2026-06-18 03:40:15,592][starVLA.training.train_starvla][INFO] - Step 4737, Loss: {'train/loss': 0.20874591171741486, 'train/grad_norm_pre_clip': 9.036920547485352, 'timing/data': 0.05112086609005928, 'timing/model': 6.694536630995572, 'learning_rate/qwen_vl_interface': 2.806041311702688e-06, 'learning_rate/action_model': 2.806041311702688e-05, 'epoch': 1.82}) +[2026-06-18 03:40:21,054][starVLA.training.train_starvla][INFO] - Step 4738, Loss: {'train/loss': 0.4204820990562439, 'train/grad_norm_pre_clip': 24.42595100402832, 'timing/data': 0.03469815477728844, 'timing/model': 1.182550516910851, 'learning_rate/qwen_vl_interface': 2.8041871015695626e-06, 'learning_rate/action_model': 2.8041871015695626e-05, 'epoch': 1.82}) +[2026-06-18 03:40:28,141][starVLA.training.train_starvla][INFO] - Step 4739, Loss: {'train/loss': 0.28611254692077637, 'train/grad_norm_pre_clip': 13.08600902557373, 'timing/data': 0.08743368741124868, 'timing/model': 2.7911636056378484, 'learning_rate/qwen_vl_interface': 2.802333398456086e-06, 'learning_rate/action_model': 2.8023333984560856e-05, 'epoch': 1.82}) +[2026-06-18 03:40:33,323][starVLA.training.train_starvla][INFO] - Step 4740, Loss: {'train/loss': 0.3147711455821991, 'train/grad_norm_pre_clip': 14.540474891662598, 'timing/data': 0.05568150710314512, 'timing/model': 1.283781087026, 'learning_rate/qwen_vl_interface': 2.8004802027465296e-06, 'learning_rate/action_model': 2.8004802027465293e-05, 'epoch': 1.82}) +[2026-06-18 03:40:38,232][starVLA.training.train_starvla][INFO] - Step 4741, Loss: {'train/loss': 0.21432265639305115, 'train/grad_norm_pre_clip': 9.96680736541748, 'timing/data': 0.04798371437937021, 'timing/model': 1.1854545269161463, 'learning_rate/qwen_vl_interface': 2.7986275148250624e-06, 'learning_rate/action_model': 2.7986275148250622e-05, 'epoch': 1.82}) +[2026-06-18 03:40:46,930][starVLA.training.train_starvla][INFO] - Step 4742, Loss: {'train/loss': 0.38874655961990356, 'train/grad_norm_pre_clip': 22.29047966003418, 'timing/data': 0.04059810657054186, 'timing/model': 1.2928543062880635, 'learning_rate/qwen_vl_interface': 2.7967753350757498e-06, 'learning_rate/action_model': 2.7967753350757497e-05, 'epoch': 1.82}) +[2026-06-18 03:40:55,534][starVLA.training.train_starvla][INFO] - Step 4743, Loss: {'train/loss': 0.38331809639930725, 'train/grad_norm_pre_clip': 11.897320747375488, 'timing/data': 0.0539958355948329, 'timing/model': 2.7820406276732683, 'learning_rate/qwen_vl_interface': 2.7949236638825496e-06, 'learning_rate/action_model': 2.7949236638825495e-05, 'epoch': 1.82}) +[2026-06-18 03:41:13,123][starVLA.training.train_starvla][INFO] - Step 4744, Loss: {'train/loss': 0.21624065935611725, 'train/grad_norm_pre_clip': 11.38731861114502, 'timing/data': 0.0455497857183218, 'timing/model': 1.2932102531194687, 'learning_rate/qwen_vl_interface': 2.7930725016293155e-06, 'learning_rate/action_model': 2.793072501629315e-05, 'epoch': 1.82}) +[2026-06-18 03:41:26,461][starVLA.training.train_starvla][INFO] - Step 4745, Loss: {'train/loss': 0.25602367520332336, 'train/grad_norm_pre_clip': 13.15137004852295, 'timing/data': 0.039984884671866894, 'timing/model': 3.159545104019344, 'learning_rate/qwen_vl_interface': 2.7912218486997954e-06, 'learning_rate/action_model': 2.7912218486997954e-05, 'epoch': 1.83}) +[2026-06-18 03:41:43,770][starVLA.training.train_starvla][INFO] - Step 4746, Loss: {'train/loss': 0.41706064343452454, 'train/grad_norm_pre_clip': 15.80424976348877, 'timing/data': 0.03712044842541218, 'timing/model': 5.976951349526644, 'learning_rate/qwen_vl_interface': 2.7893717054776327e-06, 'learning_rate/action_model': 2.7893717054776327e-05, 'epoch': 1.83}) +[2026-06-18 03:41:58,116][starVLA.training.train_starvla][INFO] - Step 4747, Loss: {'train/loss': 0.17966483533382416, 'train/grad_norm_pre_clip': 7.2803778648376465, 'timing/data': 0.058076479472219944, 'timing/model': 2.8790310183539987, 'learning_rate/qwen_vl_interface': 2.7875220723463624e-06, 'learning_rate/action_model': 2.7875220723463625e-05, 'epoch': 1.83}) +[2026-06-18 03:42:11,863][starVLA.training.train_starvla][INFO] - Step 4748, Loss: {'train/loss': 0.3519170582294464, 'train/grad_norm_pre_clip': 13.230565071105957, 'timing/data': 0.04576609097421169, 'timing/model': 4.570052728056908, 'learning_rate/qwen_vl_interface': 2.7856729496894165e-06, 'learning_rate/action_model': 2.7856729496894162e-05, 'epoch': 1.83}) +[2026-06-18 03:42:29,693][starVLA.training.train_starvla][INFO] - Step 4749, Loss: {'train/loss': 0.2986006736755371, 'train/grad_norm_pre_clip': 16.544282913208008, 'timing/data': 0.037438168190419674, 'timing/model': 3.5042696464806795, 'learning_rate/qwen_vl_interface': 2.783824337890118e-06, 'learning_rate/action_model': 2.783824337890118e-05, 'epoch': 1.83}) +[2026-06-18 03:42:36,184][starVLA.training.train_starvla][INFO] - Step 4750, Loss: {'train/loss': 0.5494253635406494, 'train/grad_norm_pre_clip': 20.705045700073242, 'timing/data': 0.04276756849139929, 'timing/model': 1.1469249865040183, 'learning_rate/qwen_vl_interface': 2.7819762373316916e-06, 'learning_rate/action_model': 2.7819762373316917e-05, 'epoch': 1.83}) +[2026-06-18 03:42:41,372][starVLA.training.train_starvla][INFO] - Step 4751, Loss: {'train/loss': 0.6444770097732544, 'train/grad_norm_pre_clip': 17.454010009765625, 'timing/data': 0.035739716142416, 'timing/model': 1.3979954030364752, 'learning_rate/qwen_vl_interface': 2.7801286483972458e-06, 'learning_rate/action_model': 2.7801286483972457e-05, 'epoch': 1.83}) +[2026-06-18 03:42:46,832][starVLA.training.train_starvla][INFO] - Step 4752, Loss: {'train/loss': 0.3641641139984131, 'train/grad_norm_pre_clip': 14.072381019592285, 'timing/data': 0.05285041593015194, 'timing/model': 1.4852446699514985, 'learning_rate/qwen_vl_interface': 2.778281571469786e-06, 'learning_rate/action_model': 2.7782815714697858e-05, 'epoch': 1.83}) +[2026-06-18 03:42:52,603][starVLA.training.train_starvla][INFO] - Step 4753, Loss: {'train/loss': 0.2316950410604477, 'train/grad_norm_pre_clip': 9.755327224731445, 'timing/data': 0.04673280008137226, 'timing/model': 1.6088628405705094, 'learning_rate/qwen_vl_interface': 2.776435006932219e-06, 'learning_rate/action_model': 2.7764350069322187e-05, 'epoch': 1.83}) +[2026-06-18 03:42:57,425][starVLA.training.train_starvla][INFO] - Step 4754, Loss: {'train/loss': 0.5854776501655579, 'train/grad_norm_pre_clip': 18.941089630126953, 'timing/data': 0.04892733786255121, 'timing/model': 1.1065609892830253, 'learning_rate/qwen_vl_interface': 2.7745889551673355e-06, 'learning_rate/action_model': 2.7745889551673353e-05, 'epoch': 1.83}) +[2026-06-18 03:43:02,363][starVLA.training.train_starvla][INFO] - Step 4755, Loss: {'train/loss': 0.22955133020877838, 'train/grad_norm_pre_clip': 12.557615280151367, 'timing/data': 0.047408025711774826, 'timing/model': 1.1164632868021727, 'learning_rate/qwen_vl_interface': 2.772743416557826e-06, 'learning_rate/action_model': 2.772743416557826e-05, 'epoch': 1.83}) +[2026-06-18 03:43:08,228][starVLA.training.train_starvla][INFO] - Step 4756, Loss: {'train/loss': 0.22697272896766663, 'train/grad_norm_pre_clip': 11.508707046508789, 'timing/data': 0.03740380983799696, 'timing/model': 1.5971058048307896, 'learning_rate/qwen_vl_interface': 2.770898391486272e-06, 'learning_rate/action_model': 2.7708983914862718e-05, 'epoch': 1.83}) +[2026-06-18 03:43:13,646][starVLA.training.train_starvla][INFO] - Step 4757, Loss: {'train/loss': 0.19872982800006866, 'train/grad_norm_pre_clip': 11.940299034118652, 'timing/data': 0.0358055280521512, 'timing/model': 1.940882608294487, 'learning_rate/qwen_vl_interface': 2.7690538803351484e-06, 'learning_rate/action_model': 2.7690538803351484e-05, 'epoch': 1.83}) +[2026-06-18 03:43:19,565][starVLA.training.train_starvla][INFO] - Step 4758, Loss: {'train/loss': 0.35536906123161316, 'train/grad_norm_pre_clip': 14.561469078063965, 'timing/data': 0.03712077997624874, 'timing/model': 1.2999110640957952, 'learning_rate/qwen_vl_interface': 2.7672098834868253e-06, 'learning_rate/action_model': 2.7672098834868253e-05, 'epoch': 1.83}) +[2026-06-18 03:43:25,459][starVLA.training.train_starvla][INFO] - Step 4759, Loss: {'train/loss': 0.3163057565689087, 'train/grad_norm_pre_clip': 13.196920394897461, 'timing/data': 0.038602971471846104, 'timing/model': 1.1269170939922333, 'learning_rate/qwen_vl_interface': 2.7653664013235655e-06, 'learning_rate/action_model': 2.7653664013235653e-05, 'epoch': 1.83}) +[2026-06-18 03:43:30,765][starVLA.training.train_starvla][INFO] - Step 4760, Loss: {'train/loss': 0.26633650064468384, 'train/grad_norm_pre_clip': 10.684020042419434, 'timing/data': 0.04436310473829508, 'timing/model': 1.114278307184577, 'learning_rate/qwen_vl_interface': 2.763523434227524e-06, 'learning_rate/action_model': 2.763523434227524e-05, 'epoch': 1.83}) +[2026-06-18 03:43:35,675][starVLA.training.train_starvla][INFO] - Step 4761, Loss: {'train/loss': 0.22757795453071594, 'train/grad_norm_pre_clip': 10.754380226135254, 'timing/data': 0.04031661245971918, 'timing/model': 1.1580821834504604, 'learning_rate/qwen_vl_interface': 2.76168098258075e-06, 'learning_rate/action_model': 2.76168098258075e-05, 'epoch': 1.83}) +[2026-06-18 03:43:40,581][starVLA.training.train_starvla][INFO] - Step 4762, Loss: {'train/loss': 0.2509380877017975, 'train/grad_norm_pre_clip': 11.82587718963623, 'timing/data': 0.0420879228040576, 'timing/model': 1.4410009793937206, 'learning_rate/qwen_vl_interface': 2.7598390467651865e-06, 'learning_rate/action_model': 2.7598390467651863e-05, 'epoch': 1.83}) +[2026-06-18 03:43:46,232][starVLA.training.train_starvla][INFO] - Step 4763, Loss: {'train/loss': 0.319023996591568, 'train/grad_norm_pre_clip': 16.107349395751953, 'timing/data': 0.052264043129980564, 'timing/model': 1.3472820147871971, 'learning_rate/qwen_vl_interface': 2.757997627162665e-06, 'learning_rate/action_model': 2.7579976271626652e-05, 'epoch': 1.83}) +[2026-06-18 03:43:54,606][starVLA.training.train_starvla][INFO] - Step 4764, Loss: {'train/loss': 0.17102713882923126, 'train/grad_norm_pre_clip': 10.2325439453125, 'timing/data': 0.08632977399975061, 'timing/model': 2.323901171796024, 'learning_rate/qwen_vl_interface': 2.7561567241549203e-06, 'learning_rate/action_model': 2.7561567241549202e-05, 'epoch': 1.83}) +[2026-06-18 03:44:09,428][starVLA.training.train_starvla][INFO] - Step 4765, Loss: {'train/loss': 0.6984625458717346, 'train/grad_norm_pre_clip': 16.91419792175293, 'timing/data': 0.04461224749684334, 'timing/model': 2.3495389530435205, 'learning_rate/qwen_vl_interface': 2.7543163381235705e-06, 'learning_rate/action_model': 2.7543163381235702e-05, 'epoch': 1.83}) +[2026-06-18 03:44:25,058][starVLA.training.train_starvla][INFO] - Step 4766, Loss: {'train/loss': 0.25469526648521423, 'train/grad_norm_pre_clip': 10.434333801269531, 'timing/data': 0.04847678355872631, 'timing/model': 1.3228747230023146, 'learning_rate/qwen_vl_interface': 2.752476469450128e-06, 'learning_rate/action_model': 2.7524764694501283e-05, 'epoch': 1.83}) +[2026-06-18 03:44:39,381][starVLA.training.train_starvla][INFO] - Step 4767, Loss: {'train/loss': 0.29210177063941956, 'train/grad_norm_pre_clip': 18.10924530029297, 'timing/data': 0.0424716267734766, 'timing/model': 1.677900718525052, 'learning_rate/qwen_vl_interface': 2.750637118515999e-06, 'learning_rate/action_model': 2.750637118515999e-05, 'epoch': 1.83}) +[2026-06-18 03:44:55,814][starVLA.training.train_starvla][INFO] - Step 4768, Loss: {'train/loss': 0.4238702952861786, 'train/grad_norm_pre_clip': 15.458070755004883, 'timing/data': 0.03935245890170336, 'timing/model': 4.047003603540361, 'learning_rate/qwen_vl_interface': 2.7487982857024875e-06, 'learning_rate/action_model': 2.7487982857024876e-05, 'epoch': 1.83}) +[2026-06-18 03:45:09,546][starVLA.training.train_starvla][INFO] - Step 4769, Loss: {'train/loss': 0.466592937707901, 'train/grad_norm_pre_clip': 15.773828506469727, 'timing/data': 0.06277025118470192, 'timing/model': 1.7075115088373423, 'learning_rate/qwen_vl_interface': 2.746959971390782e-06, 'learning_rate/action_model': 2.7469599713907817e-05, 'epoch': 1.83}) +[2026-06-18 03:45:24,014][starVLA.training.train_starvla][INFO] - Step 4770, Loss: {'train/loss': 0.32625651359558105, 'train/grad_norm_pre_clip': 12.281648635864258, 'timing/data': 0.04503151308745146, 'timing/model': 1.2631572568789124, 'learning_rate/qwen_vl_interface': 2.745122175961967e-06, 'learning_rate/action_model': 2.7451221759619672e-05, 'epoch': 1.83}) +[2026-06-18 03:45:36,199][starVLA.training.train_starvla][INFO] - Step 4771, Loss: {'train/loss': 0.3417988717556, 'train/grad_norm_pre_clip': 11.100523948669434, 'timing/data': 0.041863931342959404, 'timing/model': 2.310475043952465, 'learning_rate/qwen_vl_interface': 2.743284899797022e-06, 'learning_rate/action_model': 2.7432848997970218e-05, 'epoch': 1.84}) +[2026-06-18 03:45:50,808][starVLA.training.train_starvla][INFO] - Step 4772, Loss: {'train/loss': 0.36104071140289307, 'train/grad_norm_pre_clip': 17.657581329345703, 'timing/data': 0.03840372245758772, 'timing/model': 5.817888864316046, 'learning_rate/qwen_vl_interface': 2.741448143276813e-06, 'learning_rate/action_model': 2.741448143276813e-05, 'epoch': 1.84}) +[2026-06-18 03:45:55,558][starVLA.training.train_starvla][INFO] - Step 4773, Loss: {'train/loss': 0.34657230973243713, 'train/grad_norm_pre_clip': 11.575095176696777, 'timing/data': 0.05152325052767992, 'timing/model': 1.2728482717648149, 'learning_rate/qwen_vl_interface': 2.7396119067821043e-06, 'learning_rate/action_model': 2.7396119067821043e-05, 'epoch': 1.84}) +[2026-06-18 03:46:01,164][starVLA.training.train_starvla][INFO] - Step 4774, Loss: {'train/loss': 0.13792598247528076, 'train/grad_norm_pre_clip': 6.862358570098877, 'timing/data': 0.04436543583869934, 'timing/model': 1.3989990931004286, 'learning_rate/qwen_vl_interface': 2.737776190693548e-06, 'learning_rate/action_model': 2.7377761906935477e-05, 'epoch': 1.84}) +[2026-06-18 03:46:07,054][starVLA.training.train_starvla][INFO] - Step 4775, Loss: {'train/loss': 0.2838311493396759, 'train/grad_norm_pre_clip': 14.30738639831543, 'timing/data': 0.03479237388819456, 'timing/model': 1.1407244028523564, 'learning_rate/qwen_vl_interface': 2.7359409953916914e-06, 'learning_rate/action_model': 2.7359409953916913e-05, 'epoch': 1.84}) +[2026-06-18 03:46:12,586][starVLA.training.train_starvla][INFO] - Step 4776, Loss: {'train/loss': 0.18658104538917542, 'train/grad_norm_pre_clip': 8.568211555480957, 'timing/data': 0.04272605385631323, 'timing/model': 1.15875783842057, 'learning_rate/qwen_vl_interface': 2.734106321256971e-06, 'learning_rate/action_model': 2.7341063212569705e-05, 'epoch': 1.84}) +[2026-06-18 03:46:18,030][starVLA.training.train_starvla][INFO] - Step 4777, Loss: {'train/loss': 0.2923572361469269, 'train/grad_norm_pre_clip': 9.752243041992188, 'timing/data': 0.04237408563494682, 'timing/model': 1.4861535094678402, 'learning_rate/qwen_vl_interface': 2.732272168669718e-06, 'learning_rate/action_model': 2.732272168669718e-05, 'epoch': 1.84}) +[2026-06-18 03:46:22,858][starVLA.training.train_starvla][INFO] - Step 4778, Loss: {'train/loss': 0.4746048152446747, 'train/grad_norm_pre_clip': 24.302059173583984, 'timing/data': 0.049835292622447014, 'timing/model': 1.1365595320239663, 'learning_rate/qwen_vl_interface': 2.7304385380101512e-06, 'learning_rate/action_model': 2.730438538010151e-05, 'epoch': 1.84}) +[2026-06-18 03:46:28,629][starVLA.training.train_starvla][INFO] - Step 4779, Loss: {'train/loss': 0.4438942074775696, 'train/grad_norm_pre_clip': 18.69945526123047, 'timing/data': 0.04556148871779442, 'timing/model': 1.5381110170856118, 'learning_rate/qwen_vl_interface': 2.7286054296583904e-06, 'learning_rate/action_model': 2.72860542965839e-05, 'epoch': 1.84}) +[2026-06-18 03:46:34,068][starVLA.training.train_starvla][INFO] - Step 4780, Loss: {'train/loss': 0.28855639696121216, 'train/grad_norm_pre_clip': 13.527557373046875, 'timing/data': 0.04276738874614239, 'timing/model': 1.4181021712720394, 'learning_rate/qwen_vl_interface': 2.726772843994435e-06, 'learning_rate/action_model': 2.7267728439944352e-05, 'epoch': 1.84}) +[2026-06-18 03:46:41,549][starVLA.training.train_starvla][INFO] - Step 4781, Loss: {'train/loss': 0.2951009273529053, 'train/grad_norm_pre_clip': 11.588250160217285, 'timing/data': 0.04491173569113016, 'timing/model': 3.1580346561968327, 'learning_rate/qwen_vl_interface': 2.7249407813981834e-06, 'learning_rate/action_model': 2.724940781398183e-05, 'epoch': 1.84}) +[2026-06-18 03:46:47,132][starVLA.training.train_starvla][INFO] - Step 4782, Loss: {'train/loss': 0.2172648012638092, 'train/grad_norm_pre_clip': 10.939509391784668, 'timing/data': 0.03585445508360863, 'timing/model': 1.652847989462316, 'learning_rate/qwen_vl_interface': 2.7231092422494255e-06, 'learning_rate/action_model': 2.7231092422494253e-05, 'epoch': 1.84}) +[2026-06-18 03:46:56,178][starVLA.training.train_starvla][INFO] - Step 4783, Loss: {'train/loss': 0.5539056062698364, 'train/grad_norm_pre_clip': 19.778398513793945, 'timing/data': 0.07309239357709885, 'timing/model': 2.0233834981918335, 'learning_rate/qwen_vl_interface': 2.7212782269278417e-06, 'learning_rate/action_model': 2.7212782269278413e-05, 'epoch': 1.84}) +[2026-06-18 03:47:03,493][starVLA.training.train_starvla][INFO] - Step 4784, Loss: {'train/loss': 0.26339203119277954, 'train/grad_norm_pre_clip': 12.8556489944458, 'timing/data': 0.026841524988412857, 'timing/model': 1.3157062800601125, 'learning_rate/qwen_vl_interface': 2.719447735813004e-06, 'learning_rate/action_model': 2.719447735813004e-05, 'epoch': 1.84}) +[2026-06-18 03:47:10,091][starVLA.training.train_starvla][INFO] - Step 4785, Loss: {'train/loss': 0.6028514504432678, 'train/grad_norm_pre_clip': 17.303321838378906, 'timing/data': 0.0411841981112957, 'timing/model': 1.1381598813459277, 'learning_rate/qwen_vl_interface': 2.7176177692843697e-06, 'learning_rate/action_model': 2.7176177692843697e-05, 'epoch': 1.84}) +[2026-06-18 03:47:18,073][starVLA.training.train_starvla][INFO] - Step 4786, Loss: {'train/loss': 0.2479255348443985, 'train/grad_norm_pre_clip': 10.996646881103516, 'timing/data': 0.06273798458278179, 'timing/model': 4.357968934811652, 'learning_rate/qwen_vl_interface': 2.7157883277212985e-06, 'learning_rate/action_model': 2.7157883277212985e-05, 'epoch': 1.84}) +[2026-06-18 03:47:22,971][starVLA.training.train_starvla][INFO] - Step 4787, Loss: {'train/loss': 0.1104196161031723, 'train/grad_norm_pre_clip': 7.371620178222656, 'timing/data': 0.04728567413985729, 'timing/model': 1.5543344328179955, 'learning_rate/qwen_vl_interface': 2.7139594115030347e-06, 'learning_rate/action_model': 2.7139594115030348e-05, 'epoch': 1.84}) +[2026-06-18 03:47:27,658][starVLA.training.train_starvla][INFO] - Step 4788, Loss: {'train/loss': 0.31363001465797424, 'train/grad_norm_pre_clip': 25.57380485534668, 'timing/data': 0.04877362120896578, 'timing/model': 1.2897819504141808, 'learning_rate/qwen_vl_interface': 2.712131021008714e-06, 'learning_rate/action_model': 2.7121310210087137e-05, 'epoch': 1.84}) +[2026-06-18 03:47:37,700][starVLA.training.train_starvla][INFO] - Step 4789, Loss: {'train/loss': 0.40524739027023315, 'train/grad_norm_pre_clip': 15.048006057739258, 'timing/data': 0.05417355988174677, 'timing/model': 4.222301400266588, 'learning_rate/qwen_vl_interface': 2.7103031566173622e-06, 'learning_rate/action_model': 2.710303156617362e-05, 'epoch': 1.84}) +[2026-06-18 03:47:43,249][starVLA.training.train_starvla][INFO] - Step 4790, Loss: {'train/loss': 0.7719272375106812, 'train/grad_norm_pre_clip': 19.127450942993164, 'timing/data': 0.046598127111792564, 'timing/model': 1.132360426709056, 'learning_rate/qwen_vl_interface': 2.7084758187078995e-06, 'learning_rate/action_model': 2.7084758187078997e-05, 'epoch': 1.84}) +[2026-06-18 03:47:48,128][starVLA.training.train_starvla][INFO] - Step 4791, Loss: {'train/loss': 0.1266401708126068, 'train/grad_norm_pre_clip': 7.884068965911865, 'timing/data': 0.07935076672583818, 'timing/model': 1.2822671113535762, 'learning_rate/qwen_vl_interface': 2.7066490076591344e-06, 'learning_rate/action_model': 2.7066490076591343e-05, 'epoch': 1.84}) +[2026-06-18 03:47:52,990][starVLA.training.train_starvla][INFO] - Step 4792, Loss: {'train/loss': 0.3053305447101593, 'train/grad_norm_pre_clip': 12.123306274414062, 'timing/data': 0.05822267569601536, 'timing/model': 1.106322981417179, 'learning_rate/qwen_vl_interface': 2.704822723849765e-06, 'learning_rate/action_model': 2.7048227238497648e-05, 'epoch': 1.84}) +[2026-06-18 03:48:01,210][starVLA.training.train_starvla][INFO] - Step 4793, Loss: {'train/loss': 0.27554216980934143, 'train/grad_norm_pre_clip': 15.289798736572266, 'timing/data': 0.05225616507232189, 'timing/model': 4.1327346647158265, 'learning_rate/qwen_vl_interface': 2.702996967658388e-06, 'learning_rate/action_model': 2.7029969676583876e-05, 'epoch': 1.84}) +[2026-06-18 03:48:06,223][starVLA.training.train_starvla][INFO] - Step 4794, Loss: {'train/loss': 0.2080863118171692, 'train/grad_norm_pre_clip': 11.133750915527344, 'timing/data': 0.05528371687978506, 'timing/model': 1.2845525424927473, 'learning_rate/qwen_vl_interface': 2.7011717394634767e-06, 'learning_rate/action_model': 2.701171739463477e-05, 'epoch': 1.84}) +[2026-06-18 03:48:12,670][starVLA.training.train_starvla][INFO] - Step 4795, Loss: {'train/loss': 0.26067572832107544, 'train/grad_norm_pre_clip': 12.143171310424805, 'timing/data': 0.06558391451835632, 'timing/model': 2.1178765250369906, 'learning_rate/qwen_vl_interface': 2.699347039643408e-06, 'learning_rate/action_model': 2.699347039643408e-05, 'epoch': 1.84}) +[2026-06-18 03:48:17,678][starVLA.training.train_starvla][INFO] - Step 4796, Loss: {'train/loss': 0.49082621932029724, 'train/grad_norm_pre_clip': 16.036962509155273, 'timing/data': 0.06395831424742937, 'timing/model': 1.320675214752555, 'learning_rate/qwen_vl_interface': 2.6975228685764398e-06, 'learning_rate/action_model': 2.69752286857644e-05, 'epoch': 1.84}) +[2026-06-18 03:48:22,412][starVLA.training.train_starvla][INFO] - Step 4797, Loss: {'train/loss': 0.19381293654441833, 'train/grad_norm_pre_clip': 8.902015686035156, 'timing/data': 0.043376083485782146, 'timing/model': 1.5221732994541526, 'learning_rate/qwen_vl_interface': 2.6956992266407307e-06, 'learning_rate/action_model': 2.6956992266407306e-05, 'epoch': 1.85}) +[2026-06-18 03:48:27,243][starVLA.training.train_starvla][INFO] - Step 4798, Loss: {'train/loss': 0.5131665468215942, 'train/grad_norm_pre_clip': 19.535655975341797, 'timing/data': 0.05730424355715513, 'timing/model': 1.392485679127276, 'learning_rate/qwen_vl_interface': 2.693876114214321e-06, 'learning_rate/action_model': 2.6938761142143208e-05, 'epoch': 1.85}) +[2026-06-18 03:48:32,215][starVLA.training.train_starvla][INFO] - Step 4799, Loss: {'train/loss': 0.20787401497364044, 'train/grad_norm_pre_clip': 9.921202659606934, 'timing/data': 0.0468897707760334, 'timing/model': 1.078244118951261, 'learning_rate/qwen_vl_interface': 2.6920535316751417e-06, 'learning_rate/action_model': 2.6920535316751416e-05, 'epoch': 1.85}) +[2026-06-18 03:52:12,957][starVLA.training.train_starvla][INFO] - Step 4800, Loss: {'train/loss': 0.334778368473053, 'train/grad_norm_pre_clip': 14.327373504638672, 'eval/loss': 0.4358205795288086, 'eval/latency_4/loss': 0.4358205795288086, 'eval/demon_fire_precision': 0.9617814276689829, 'eval/demon_fire_recall': 0.9889574537187399, 'eval/demon_fire_f1': 0.9751801441152923, 'eval/demon_fire_rate_model': 0.9298091042584434, 'eval/demon_fire_rate_teacher': 0.9042584434654919, 'eval/demon_left_precision': 0.8063583815028902, 'eval/demon_left_recall': 0.636986301369863, 'eval/demon_left_f1': 0.711734693877551, 'eval/demon_left_rate_model': 0.10161527165932452, 'eval/demon_left_rate_teacher': 0.12863436123348018, 'eval/demon_right_precision': 0.7902350813743219, 'eval/demon_right_recall': 0.7888086642599278, 'eval/demon_right_f1': 0.7895212285456188, 'eval/demon_right_rate_model': 0.16240822320117473, 'eval/demon_right_rate_teacher': 0.16270190895741557, 'eval/demon_move_f1': 0.7506279612115849, 'eval/demon_cc_f1': 0.8948694329630055, 'eval/latency_4/demon_fire_precision': 0.9617814276689829, 'eval/latency_4/demon_fire_recall': 0.9889574537187399, 'eval/latency_4/demon_fire_f1': 0.9751801441152923, 'eval/latency_4/demon_fire_rate_model': 0.9298091042584434, 'eval/latency_4/demon_fire_rate_teacher': 0.9042584434654919, 'eval/latency_4/demon_left_precision': 0.8063583815028902, 'eval/latency_4/demon_left_recall': 0.636986301369863, 'eval/latency_4/demon_left_f1': 0.711734693877551, 'eval/latency_4/demon_left_rate_model': 0.10161527165932452, 'eval/latency_4/demon_left_rate_teacher': 0.12863436123348018, 'eval/latency_4/demon_right_precision': 0.7902350813743219, 'eval/latency_4/demon_right_recall': 0.7888086642599278, 'eval/latency_4/demon_right_f1': 0.7895212285456188, 'eval/latency_4/demon_right_rate_model': 0.16240822320117473, 'eval/latency_4/demon_right_rate_teacher': 0.16270190895741557, 'eval/latency_4/demon_move_f1': 0.7506279612115849, 'eval/latency_4/demon_cc_f1': 0.8948694329630055, 'timing/data': 0.10339816194027662, 'timing/model': 1.2063880702480674, 'learning_rate/qwen_vl_interface': 2.6902314794010164e-06, 'learning_rate/action_model': 2.6902314794010165e-05, 'epoch': 1.85}) +[2026-06-18 03:52:12,959][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_4800_state +[2026-06-18 03:52:12,968][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 03:52:22,191][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_4800_state/model.safetensors +[2026-06-18 03:52:45,872][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_4800_state/optimizer.bin +[2026-06-18 03:52:45,874][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_4800_state/sampler.bin +[2026-06-18 03:52:45,875][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_4800_state/sampler_1.bin +[2026-06-18 03:52:45,888][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_4800_state/random_states_0.pkl +[2026-06-18 03:52:45,907][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4800 +[2026-06-18 03:52:45,907][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 03:52:45,925][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 03:52:51,240][starVLA.training.train_starvla][INFO] - Step 4801, Loss: {'train/loss': 0.41469940543174744, 'train/grad_norm_pre_clip': 17.133146286010742, 'timing/data': 0.04985013231635094, 'timing/model': 1.5766539834439754, 'learning_rate/qwen_vl_interface': 2.6884099577696615e-06, 'learning_rate/action_model': 2.6884099577696615e-05, 'epoch': 1.85}) +[2026-06-18 03:52:58,453][starVLA.training.train_starvla][INFO] - Step 4802, Loss: {'train/loss': 0.42576536536216736, 'train/grad_norm_pre_clip': 17.644329071044922, 'timing/data': 0.06265998631715775, 'timing/model': 1.6229166500270367, 'learning_rate/qwen_vl_interface': 2.686588967158679e-06, 'learning_rate/action_model': 2.6865889671586792e-05, 'epoch': 1.85}) +[2026-06-18 03:53:02,779][starVLA.training.train_starvla][INFO] - Step 4803, Loss: {'train/loss': 0.14886313676834106, 'train/grad_norm_pre_clip': 10.036847114562988, 'timing/data': 0.05245174653828144, 'timing/model': 1.1061818795278668, 'learning_rate/qwen_vl_interface': 2.6847685079455614e-06, 'learning_rate/action_model': 2.6847685079455613e-05, 'epoch': 1.85}) +[2026-06-18 03:53:12,642][starVLA.training.train_starvla][INFO] - Step 4804, Loss: {'train/loss': 0.2071814090013504, 'train/grad_norm_pre_clip': 11.470688819885254, 'timing/data': 0.04639723151922226, 'timing/model': 1.549237448722124, 'learning_rate/qwen_vl_interface': 2.6829485805076927e-06, 'learning_rate/action_model': 2.6829485805076925e-05, 'epoch': 1.85}) +[2026-06-18 03:53:18,355][starVLA.training.train_starvla][INFO] - Step 4805, Loss: {'train/loss': 0.5079429149627686, 'train/grad_norm_pre_clip': 20.482635498046875, 'timing/data': 0.04122789856046438, 'timing/model': 1.2575064096599817, 'learning_rate/qwen_vl_interface': 2.6811291852223452e-06, 'learning_rate/action_model': 2.6811291852223454e-05, 'epoch': 1.85}) +[2026-06-18 03:53:23,504][starVLA.training.train_starvla][INFO] - Step 4806, Loss: {'train/loss': 0.1292855441570282, 'train/grad_norm_pre_clip': 8.258724212646484, 'timing/data': 0.0457683177664876, 'timing/model': 1.4147792989388108, 'learning_rate/qwen_vl_interface': 2.679310322466683e-06, 'learning_rate/action_model': 2.6793103224666827e-05, 'epoch': 1.85}) +[2026-06-18 03:53:30,547][starVLA.training.train_starvla][INFO] - Step 4807, Loss: {'train/loss': 0.18427333235740662, 'train/grad_norm_pre_clip': 9.843042373657227, 'timing/data': 0.05472493264824152, 'timing/model': 1.708534867502749, 'learning_rate/qwen_vl_interface': 2.6774919926177563e-06, 'learning_rate/action_model': 2.677491992617756e-05, 'epoch': 1.85}) +[2026-06-18 03:53:36,285][starVLA.training.train_starvla][INFO] - Step 4808, Loss: {'train/loss': 0.4828440248966217, 'train/grad_norm_pre_clip': 21.756439208984375, 'timing/data': 0.044434268958866596, 'timing/model': 2.466498357243836, 'learning_rate/qwen_vl_interface': 2.675674196052507e-06, 'learning_rate/action_model': 2.6756741960525072e-05, 'epoch': 1.85}) +[2026-06-18 03:53:42,024][starVLA.training.train_starvla][INFO] - Step 4809, Loss: {'train/loss': 0.13071735203266144, 'train/grad_norm_pre_clip': 8.488662719726562, 'timing/data': 0.060171754099428654, 'timing/model': 1.2006908189505339, 'learning_rate/qwen_vl_interface': 2.6738569331477677e-06, 'learning_rate/action_model': 2.6738569331477674e-05, 'epoch': 1.85}) +[2026-06-18 03:53:50,152][starVLA.training.train_starvla][INFO] - Step 4810, Loss: {'train/loss': 0.2979486584663391, 'train/grad_norm_pre_clip': 14.206463813781738, 'timing/data': 0.04249646980315447, 'timing/model': 1.3805358968675137, 'learning_rate/qwen_vl_interface': 2.672040204280258e-06, 'learning_rate/action_model': 2.672040204280258e-05, 'epoch': 1.85}) +[2026-06-18 03:53:55,265][starVLA.training.train_starvla][INFO] - Step 4811, Loss: {'train/loss': 0.3105317950248718, 'train/grad_norm_pre_clip': 13.369344711303711, 'timing/data': 0.0714236218482256, 'timing/model': 1.2393484991043806, 'learning_rate/qwen_vl_interface': 2.6702240098265864e-06, 'learning_rate/action_model': 2.6702240098265864e-05, 'epoch': 1.85}) +[2026-06-18 03:54:00,539][starVLA.training.train_starvla][INFO] - Step 4812, Loss: {'train/loss': 0.2701805830001831, 'train/grad_norm_pre_clip': 15.941723823547363, 'timing/data': 0.046451011672616005, 'timing/model': 1.5667753517627716, 'learning_rate/qwen_vl_interface': 2.6684083501632573e-06, 'learning_rate/action_model': 2.6684083501632574e-05, 'epoch': 1.85}) +[2026-06-18 03:54:05,326][starVLA.training.train_starvla][INFO] - Step 4813, Loss: {'train/loss': 0.2931564450263977, 'train/grad_norm_pre_clip': 15.857197761535645, 'timing/data': 0.05926954746246338, 'timing/model': 1.139642215333879, 'learning_rate/qwen_vl_interface': 2.6665932256666525e-06, 'learning_rate/action_model': 2.6665932256666527e-05, 'epoch': 1.85}) +[2026-06-18 03:54:12,242][starVLA.training.train_starvla][INFO] - Step 4814, Loss: {'train/loss': 0.28346481919288635, 'train/grad_norm_pre_clip': 14.273333549499512, 'timing/data': 0.040530833415687084, 'timing/model': 1.1298038102686405, 'learning_rate/qwen_vl_interface': 2.6647786367130513e-06, 'learning_rate/action_model': 2.664778636713051e-05, 'epoch': 1.85}) +[2026-06-18 03:54:17,080][starVLA.training.train_starvla][INFO] - Step 4815, Loss: {'train/loss': 0.2963583469390869, 'train/grad_norm_pre_clip': 12.546849250793457, 'timing/data': 0.062334892340004444, 'timing/model': 1.1549108168110251, 'learning_rate/qwen_vl_interface': 2.6629645836786217e-06, 'learning_rate/action_model': 2.6629645836786217e-05, 'epoch': 1.85}) +[2026-06-18 03:54:21,789][starVLA.training.train_starvla][INFO] - Step 4816, Loss: {'train/loss': 0.1887315958738327, 'train/grad_norm_pre_clip': 10.189820289611816, 'timing/data': 0.043587760999798775, 'timing/model': 1.103616309352219, 'learning_rate/qwen_vl_interface': 2.661151066939417e-06, 'learning_rate/action_model': 2.6611510669394168e-05, 'epoch': 1.85}) +[2026-06-18 03:54:26,484][starVLA.training.train_starvla][INFO] - Step 4817, Loss: {'train/loss': 0.6720006465911865, 'train/grad_norm_pre_clip': 18.183700561523438, 'timing/data': 0.04633576236665249, 'timing/model': 1.4284830754622817, 'learning_rate/qwen_vl_interface': 2.6593380868713843e-06, 'learning_rate/action_model': 2.659338086871384e-05, 'epoch': 1.85}) +[2026-06-18 03:54:30,832][starVLA.training.train_starvla][INFO] - Step 4818, Loss: {'train/loss': 0.31685903668403625, 'train/grad_norm_pre_clip': 17.648115158081055, 'timing/data': 0.044090707786381245, 'timing/model': 1.148187156766653, 'learning_rate/qwen_vl_interface': 2.6575256438503507e-06, 'learning_rate/action_model': 2.6575256438503504e-05, 'epoch': 1.85}) +[2026-06-18 03:54:38,061][starVLA.training.train_starvla][INFO] - Step 4819, Loss: {'train/loss': 0.6463472247123718, 'train/grad_norm_pre_clip': 21.560922622680664, 'timing/data': 0.047415933571755886, 'timing/model': 1.2438028454780579, 'learning_rate/qwen_vl_interface': 2.6557137382520413e-06, 'learning_rate/action_model': 2.6557137382520412e-05, 'epoch': 1.85}) +[2026-06-18 03:54:42,764][starVLA.training.train_starvla][INFO] - Step 4820, Loss: {'train/loss': 0.4382573068141937, 'train/grad_norm_pre_clip': 18.044288635253906, 'timing/data': 0.0406273789703846, 'timing/model': 1.094565650448203, 'learning_rate/qwen_vl_interface': 2.6539023704520666e-06, 'learning_rate/action_model': 2.6539023704520664e-05, 'epoch': 1.85}) +[2026-06-18 03:54:49,658][starVLA.training.train_starvla][INFO] - Step 4821, Loss: {'train/loss': 0.08006352186203003, 'train/grad_norm_pre_clip': 6.282601833343506, 'timing/data': 0.04408374521881342, 'timing/model': 1.716720744036138, 'learning_rate/qwen_vl_interface': 2.6520915408259234e-06, 'learning_rate/action_model': 2.6520915408259235e-05, 'epoch': 1.85}) +[2026-06-18 03:54:54,860][starVLA.training.train_starvla][INFO] - Step 4822, Loss: {'train/loss': 0.17376399040222168, 'train/grad_norm_pre_clip': 12.735408782958984, 'timing/data': 0.058331907726824284, 'timing/model': 1.5116002019494772, 'learning_rate/qwen_vl_interface': 2.650281249749e-06, 'learning_rate/action_model': 2.6502812497489998e-05, 'epoch': 1.85}) +[2026-06-18 03:55:00,471][starVLA.training.train_starvla][INFO] - Step 4823, Loss: {'train/loss': 0.24534127116203308, 'train/grad_norm_pre_clip': 12.84205436706543, 'timing/data': 0.05978888738900423, 'timing/model': 1.3259992338716984, 'learning_rate/qwen_vl_interface': 2.6484714975965704e-06, 'learning_rate/action_model': 2.6484714975965703e-05, 'epoch': 1.86}) +[2026-06-18 03:55:05,790][starVLA.training.train_starvla][INFO] - Step 4824, Loss: {'train/loss': 0.2919013500213623, 'train/grad_norm_pre_clip': 14.814339637756348, 'timing/data': 0.04403669014573097, 'timing/model': 1.144929121248424, 'learning_rate/qwen_vl_interface': 2.646662284743799e-06, 'learning_rate/action_model': 2.646662284743799e-05, 'epoch': 1.86}) +[2026-06-18 03:55:13,075][starVLA.training.train_starvla][INFO] - Step 4825, Loss: {'train/loss': 0.2859680950641632, 'train/grad_norm_pre_clip': 13.135045051574707, 'timing/data': 0.04028189927339554, 'timing/model': 1.3045904282480478, 'learning_rate/qwen_vl_interface': 2.6448536115657353e-06, 'learning_rate/action_model': 2.6448536115657353e-05, 'epoch': 1.86}) +[2026-06-18 03:55:18,477][starVLA.training.train_starvla][INFO] - Step 4826, Loss: {'train/loss': 0.5186755657196045, 'train/grad_norm_pre_clip': 18.022672653198242, 'timing/data': 0.05183808412402868, 'timing/model': 1.2315886858850718, 'learning_rate/qwen_vl_interface': 2.6430454784373248e-06, 'learning_rate/action_model': 2.643045478437325e-05, 'epoch': 1.86}) +[2026-06-18 03:55:23,703][starVLA.training.train_starvla][INFO] - Step 4827, Loss: {'train/loss': 0.30509036779403687, 'train/grad_norm_pre_clip': 16.65531349182129, 'timing/data': 0.05162934120744467, 'timing/model': 1.4498105673119426, 'learning_rate/qwen_vl_interface': 2.6412378857333903e-06, 'learning_rate/action_model': 2.6412378857333903e-05, 'epoch': 1.86}) +[2026-06-18 03:55:29,128][starVLA.training.train_starvla][INFO] - Step 4828, Loss: {'train/loss': 0.32928958535194397, 'train/grad_norm_pre_clip': 13.008126258850098, 'timing/data': 0.039471606723964214, 'timing/model': 1.3641743557527661, 'learning_rate/qwen_vl_interface': 2.639430833828649e-06, 'learning_rate/action_model': 2.6394308338286488e-05, 'epoch': 1.86}) +[2026-06-18 03:55:36,695][starVLA.training.train_starvla][INFO] - Step 4829, Loss: {'train/loss': 0.4322660267353058, 'train/grad_norm_pre_clip': 17.63531494140625, 'timing/data': 0.0771645400673151, 'timing/model': 1.1706114569678903, 'learning_rate/qwen_vl_interface': 2.637624323097703e-06, 'learning_rate/action_model': 2.6376243230977026e-05, 'epoch': 1.86}) +[2026-06-18 03:55:47,336][starVLA.training.train_starvla][INFO] - Step 4830, Loss: {'train/loss': 0.39891481399536133, 'train/grad_norm_pre_clip': 12.738164901733398, 'timing/data': 0.04823557659983635, 'timing/model': 1.401294861920178, 'learning_rate/qwen_vl_interface': 2.635818353915048e-06, 'learning_rate/action_model': 2.635818353915048e-05, 'epoch': 1.86}) +[2026-06-18 03:55:54,084][starVLA.training.train_starvla][INFO] - Step 4831, Loss: {'train/loss': 0.2705618143081665, 'train/grad_norm_pre_clip': 13.07270622253418, 'timing/data': 0.04330088570713997, 'timing/model': 2.417204543016851, 'learning_rate/qwen_vl_interface': 2.634012926655063e-06, 'learning_rate/action_model': 2.6340129266550628e-05, 'epoch': 1.86}) +[2026-06-18 03:55:59,538][starVLA.training.train_starvla][INFO] - Step 4832, Loss: {'train/loss': 0.574080228805542, 'train/grad_norm_pre_clip': 14.705096244812012, 'timing/data': 0.04619611706584692, 'timing/model': 1.585296013392508, 'learning_rate/qwen_vl_interface': 2.632208041692009e-06, 'learning_rate/action_model': 2.6322080416920088e-05, 'epoch': 1.86}) +[2026-06-18 03:56:05,365][starVLA.training.train_starvla][INFO] - Step 4833, Loss: {'train/loss': 0.4613828957080841, 'train/grad_norm_pre_clip': 20.393600463867188, 'timing/data': 0.0598005848005414, 'timing/model': 2.4554974157363176, 'learning_rate/qwen_vl_interface': 2.630403699400047e-06, 'learning_rate/action_model': 2.6304036994000468e-05, 'epoch': 1.86}) +[2026-06-18 03:56:16,615][starVLA.training.train_starvla][INFO] - Step 4834, Loss: {'train/loss': 0.21225062012672424, 'train/grad_norm_pre_clip': 13.053912162780762, 'timing/data': 0.045718714594841, 'timing/model': 3.582597801461816, 'learning_rate/qwen_vl_interface': 2.628599900153217e-06, 'learning_rate/action_model': 2.628599900153217e-05, 'epoch': 1.86}) +[2026-06-18 03:56:25,919][starVLA.training.train_starvla][INFO] - Step 4835, Loss: {'train/loss': 0.18246577680110931, 'train/grad_norm_pre_clip': 7.804225444793701, 'timing/data': 0.044034598395228386, 'timing/model': 1.1138041531667113, 'learning_rate/qwen_vl_interface': 2.6267966443254484e-06, 'learning_rate/action_model': 2.6267966443254483e-05, 'epoch': 1.86}) +[2026-06-18 03:56:30,730][starVLA.training.train_starvla][INFO] - Step 4836, Loss: {'train/loss': 0.2834758758544922, 'train/grad_norm_pre_clip': 11.46695613861084, 'timing/data': 0.043355485424399376, 'timing/model': 1.0730817886069417, 'learning_rate/qwen_vl_interface': 2.6249939322905576e-06, 'learning_rate/action_model': 2.6249939322905576e-05, 'epoch': 1.86}) +[2026-06-18 03:56:38,212][starVLA.training.train_starvla][INFO] - Step 4837, Loss: {'train/loss': 0.3879559636116028, 'train/grad_norm_pre_clip': 17.9091854095459, 'timing/data': 0.04662454966455698, 'timing/model': 1.4482945119962096, 'learning_rate/qwen_vl_interface': 2.6231917644222505e-06, 'learning_rate/action_model': 2.6231917644222504e-05, 'epoch': 1.86}) +[2026-06-18 03:56:47,981][starVLA.training.train_starvla][INFO] - Step 4838, Loss: {'train/loss': 0.22980442643165588, 'train/grad_norm_pre_clip': 13.33418083190918, 'timing/data': 0.04217826947569847, 'timing/model': 1.3184771863743663, 'learning_rate/qwen_vl_interface': 2.621390141094116e-06, 'learning_rate/action_model': 2.621390141094116e-05, 'epoch': 1.86}) +[2026-06-18 03:57:00,668][starVLA.training.train_starvla][INFO] - Step 4839, Loss: {'train/loss': 0.388616144657135, 'train/grad_norm_pre_clip': 18.212915420532227, 'timing/data': 0.045532289892435074, 'timing/model': 1.674888987094164, 'learning_rate/qwen_vl_interface': 2.6195890626796345e-06, 'learning_rate/action_model': 2.6195890626796345e-05, 'epoch': 1.86}) +[2026-06-18 03:57:06,175][starVLA.training.train_starvla][INFO] - Step 4840, Loss: {'train/loss': 0.747684121131897, 'train/grad_norm_pre_clip': 21.41783332824707, 'timing/data': 0.04179306700825691, 'timing/model': 1.3936926731839776, 'learning_rate/qwen_vl_interface': 2.6177885295521706e-06, 'learning_rate/action_model': 2.6177885295521702e-05, 'epoch': 1.86}) +[2026-06-18 03:57:11,319][starVLA.training.train_starvla][INFO] - Step 4841, Loss: {'train/loss': 0.5243272185325623, 'train/grad_norm_pre_clip': 20.12251091003418, 'timing/data': 0.052581348456442356, 'timing/model': 1.1013010377064347, 'learning_rate/qwen_vl_interface': 2.6159885420849766e-06, 'learning_rate/action_model': 2.6159885420849766e-05, 'epoch': 1.86}) +[2026-06-18 03:57:16,259][starVLA.training.train_starvla][INFO] - Step 4842, Loss: {'train/loss': 0.13466718792915344, 'train/grad_norm_pre_clip': 9.535992622375488, 'timing/data': 0.04122584406286478, 'timing/model': 1.5611706236377358, 'learning_rate/qwen_vl_interface': 2.6141891006511927e-06, 'learning_rate/action_model': 2.6141891006511925e-05, 'epoch': 1.86}) +[2026-06-18 03:57:20,807][starVLA.training.train_starvla][INFO] - Step 4843, Loss: {'train/loss': 0.553714394569397, 'train/grad_norm_pre_clip': 14.624183654785156, 'timing/data': 0.04293401259928942, 'timing/model': 1.2331600431352854, 'learning_rate/qwen_vl_interface': 2.612390205623842e-06, 'learning_rate/action_model': 2.612390205623842e-05, 'epoch': 1.86}) +[2026-06-18 03:57:26,056][starVLA.training.train_starvla][INFO] - Step 4844, Loss: {'train/loss': 0.38568755984306335, 'train/grad_norm_pre_clip': 12.237610816955566, 'timing/data': 0.05382566526532173, 'timing/model': 1.2514791823923588, 'learning_rate/qwen_vl_interface': 2.6105918573758425e-06, 'learning_rate/action_model': 2.6105918573758427e-05, 'epoch': 1.86}) +[2026-06-18 03:57:30,930][starVLA.training.train_starvla][INFO] - Step 4845, Loss: {'train/loss': 0.33049294352531433, 'train/grad_norm_pre_clip': 15.497383117675781, 'timing/data': 0.04571689013391733, 'timing/model': 1.1340914089232683, 'learning_rate/qwen_vl_interface': 2.608794056279991e-06, 'learning_rate/action_model': 2.6087940562799906e-05, 'epoch': 1.86}) +[2026-06-18 03:57:36,040][starVLA.training.train_starvla][INFO] - Step 4846, Loss: {'train/loss': 0.267660915851593, 'train/grad_norm_pre_clip': 19.27191734313965, 'timing/data': 0.06271212082356215, 'timing/model': 1.0945430472493172, 'learning_rate/qwen_vl_interface': 2.6069968027089744e-06, 'learning_rate/action_model': 2.6069968027089742e-05, 'epoch': 1.86}) +[2026-06-18 03:57:41,224][starVLA.training.train_starvla][INFO] - Step 4847, Loss: {'train/loss': 0.335975706577301, 'train/grad_norm_pre_clip': 15.110321044921875, 'timing/data': 0.05027883779257536, 'timing/model': 1.6442290283739567, 'learning_rate/qwen_vl_interface': 2.605200097035362e-06, 'learning_rate/action_model': 2.605200097035362e-05, 'epoch': 1.86}) +[2026-06-18 03:57:45,831][starVLA.training.train_starvla][INFO] - Step 4848, Loss: {'train/loss': 0.3631730079650879, 'train/grad_norm_pre_clip': 13.065469741821289, 'timing/data': 0.04627731628715992, 'timing/model': 1.1911414358764887, 'learning_rate/qwen_vl_interface': 2.6034039396316167e-06, 'learning_rate/action_model': 2.6034039396316167e-05, 'epoch': 1.86}) +[2026-06-18 03:57:50,498][starVLA.training.train_starvla][INFO] - Step 4849, Loss: {'train/loss': 0.14612366259098053, 'train/grad_norm_pre_clip': 12.589640617370605, 'timing/data': 0.044970519840717316, 'timing/model': 1.1357308682054281, 'learning_rate/qwen_vl_interface': 2.601608330870083e-06, 'learning_rate/action_model': 2.601608330870083e-05, 'epoch': 1.87}) +[2026-06-18 03:57:55,680][starVLA.training.train_starvla][INFO] - Step 4850, Loss: {'train/loss': 0.3530164062976837, 'train/grad_norm_pre_clip': 12.097424507141113, 'timing/data': 0.04316453542560339, 'timing/model': 1.180267820134759, 'learning_rate/qwen_vl_interface': 2.5998132711229918e-06, 'learning_rate/action_model': 2.5998132711229915e-05, 'epoch': 1.87}) +[2026-06-18 03:58:00,727][starVLA.training.train_starvla][INFO] - Step 4851, Loss: {'train/loss': 0.26854225993156433, 'train/grad_norm_pre_clip': 11.750102996826172, 'timing/data': 0.053462112322449684, 'timing/model': 1.2065935358405113, 'learning_rate/qwen_vl_interface': 2.598018760762461e-06, 'learning_rate/action_model': 2.598018760762461e-05, 'epoch': 1.87}) +[2026-06-18 03:58:05,517][starVLA.training.train_starvla][INFO] - Step 4852, Loss: {'train/loss': 0.583974301815033, 'train/grad_norm_pre_clip': 25.289840698242188, 'timing/data': 0.05559597071260214, 'timing/model': 1.5502714291214943, 'learning_rate/qwen_vl_interface': 2.5962248001604958e-06, 'learning_rate/action_model': 2.5962248001604956e-05, 'epoch': 1.87}) +[2026-06-18 03:58:10,365][starVLA.training.train_starvla][INFO] - Step 4853, Loss: {'train/loss': 0.5250160098075867, 'train/grad_norm_pre_clip': 16.7640380859375, 'timing/data': 0.05271728429943323, 'timing/model': 1.295397812500596, 'learning_rate/qwen_vl_interface': 2.5944313896889837e-06, 'learning_rate/action_model': 2.594431389688984e-05, 'epoch': 1.87}) +[2026-06-18 03:58:15,140][starVLA.training.train_starvla][INFO] - Step 4854, Loss: {'train/loss': 0.2606659233570099, 'train/grad_norm_pre_clip': 11.710350036621094, 'timing/data': 0.03982452396303415, 'timing/model': 1.0817355290055275, 'learning_rate/qwen_vl_interface': 2.592638529719703e-06, 'learning_rate/action_model': 2.5926385297197033e-05, 'epoch': 1.87}) +[2026-06-18 03:58:23,315][starVLA.training.train_starvla][INFO] - Step 4855, Loss: {'train/loss': 0.19217343628406525, 'train/grad_norm_pre_clip': 10.888032913208008, 'timing/data': 0.15103548392653465, 'timing/model': 3.312897370196879, 'learning_rate/qwen_vl_interface': 2.590846220624314e-06, 'learning_rate/action_model': 2.5908462206243138e-05, 'epoch': 1.87}) +[2026-06-18 03:58:30,986][starVLA.training.train_starvla][INFO] - Step 4856, Loss: {'train/loss': 0.5291430354118347, 'train/grad_norm_pre_clip': 13.206964492797852, 'timing/data': 0.05764971952885389, 'timing/model': 2.018875396810472, 'learning_rate/qwen_vl_interface': 2.5890544627743652e-06, 'learning_rate/action_model': 2.589054462774365e-05, 'epoch': 1.87}) +[2026-06-18 03:58:48,346][starVLA.training.train_starvla][INFO] - Step 4857, Loss: {'train/loss': 0.1963365375995636, 'train/grad_norm_pre_clip': 11.699569702148438, 'timing/data': 0.04959412105381489, 'timing/model': 4.182377376593649, 'learning_rate/qwen_vl_interface': 2.58726325654129e-06, 'learning_rate/action_model': 2.5872632565412898e-05, 'epoch': 1.87}) +[2026-06-18 03:58:56,508][starVLA.training.train_starvla][INFO] - Step 4858, Loss: {'train/loss': 0.44486600160598755, 'train/grad_norm_pre_clip': 20.473241806030273, 'timing/data': 0.042723250575363636, 'timing/model': 1.2891216445714235, 'learning_rate/qwen_vl_interface': 2.5854726022964043e-06, 'learning_rate/action_model': 2.5854726022964042e-05, 'epoch': 1.87}) +[2026-06-18 03:59:03,586][starVLA.training.train_starvla][INFO] - Step 4859, Loss: {'train/loss': 0.3762606978416443, 'train/grad_norm_pre_clip': 14.495044708251953, 'timing/data': 0.06799427326768637, 'timing/model': 1.3760552611202002, 'learning_rate/qwen_vl_interface': 2.5836825004109173e-06, 'learning_rate/action_model': 2.5836825004109172e-05, 'epoch': 1.87}) +[2026-06-18 03:59:24,497][starVLA.training.train_starvla][INFO] - Step 4860, Loss: {'train/loss': 0.6122835278511047, 'train/grad_norm_pre_clip': 16.57156753540039, 'timing/data': 0.050495320931077003, 'timing/model': 5.123250612989068, 'learning_rate/qwen_vl_interface': 2.5818929512559193e-06, 'learning_rate/action_model': 2.5818929512559193e-05, 'epoch': 1.87}) +[2026-06-18 03:59:35,010][starVLA.training.train_starvla][INFO] - Step 4861, Loss: {'train/loss': 0.2001892328262329, 'train/grad_norm_pre_clip': 10.83487319946289, 'timing/data': 0.049211395904421806, 'timing/model': 4.782302189618349, 'learning_rate/qwen_vl_interface': 2.580103955202381e-06, 'learning_rate/action_model': 2.580103955202381e-05, 'epoch': 1.87}) +[2026-06-18 03:59:51,708][starVLA.training.train_starvla][INFO] - Step 4862, Loss: {'train/loss': 0.27925288677215576, 'train/grad_norm_pre_clip': 12.40611743927002, 'timing/data': 0.053818185813724995, 'timing/model': 4.893394438549876, 'learning_rate/qwen_vl_interface': 2.5783155126211647e-06, 'learning_rate/action_model': 2.5783155126211644e-05, 'epoch': 1.87}) +[2026-06-18 04:00:03,300][starVLA.training.train_starvla][INFO] - Step 4863, Loss: {'train/loss': 0.31744185090065, 'train/grad_norm_pre_clip': 17.37549591064453, 'timing/data': 0.04108247347176075, 'timing/model': 5.707747916691005, 'learning_rate/qwen_vl_interface': 2.576527623883019e-06, 'learning_rate/action_model': 2.576527623883019e-05, 'epoch': 1.87}) +[2026-06-18 04:00:10,445][starVLA.training.train_starvla][INFO] - Step 4864, Loss: {'train/loss': 0.59738689661026, 'train/grad_norm_pre_clip': 22.886070251464844, 'timing/data': 0.05395260825753212, 'timing/model': 1.4453102890402079, 'learning_rate/qwen_vl_interface': 2.5747402893585732e-06, 'learning_rate/action_model': 2.574740289358573e-05, 'epoch': 1.87}) +[2026-06-18 04:00:15,732][starVLA.training.train_starvla][INFO] - Step 4865, Loss: {'train/loss': 0.2013094574213028, 'train/grad_norm_pre_clip': 13.459461212158203, 'timing/data': 0.04911514651030302, 'timing/model': 1.224456300958991, 'learning_rate/qwen_vl_interface': 2.5729535094183445e-06, 'learning_rate/action_model': 2.5729535094183444e-05, 'epoch': 1.87}) +[2026-06-18 04:00:21,052][starVLA.training.train_starvla][INFO] - Step 4866, Loss: {'train/loss': 0.15854108333587646, 'train/grad_norm_pre_clip': 8.1245756149292, 'timing/data': 0.05826573446393013, 'timing/model': 1.3912859726697206, 'learning_rate/qwen_vl_interface': 2.571167284432734e-06, 'learning_rate/action_model': 2.5711672844327338e-05, 'epoch': 1.87}) +[2026-06-18 04:00:25,916][starVLA.training.train_starvla][INFO] - Step 4867, Loss: {'train/loss': 0.20911425352096558, 'train/grad_norm_pre_clip': 9.1303129196167, 'timing/data': 0.04406741727143526, 'timing/model': 1.4314412036910653, 'learning_rate/qwen_vl_interface': 2.569381614772028e-06, 'learning_rate/action_model': 2.5693816147720278e-05, 'epoch': 1.87}) +[2026-06-18 04:00:30,813][starVLA.training.train_starvla][INFO] - Step 4868, Loss: {'train/loss': 0.515356183052063, 'train/grad_norm_pre_clip': 21.70218276977539, 'timing/data': 0.05975511111319065, 'timing/model': 1.309635067358613, 'learning_rate/qwen_vl_interface': 2.5675965008063976e-06, 'learning_rate/action_model': 2.5675965008063974e-05, 'epoch': 1.87}) +[2026-06-18 04:00:39,216][starVLA.training.train_starvla][INFO] - Step 4869, Loss: {'train/loss': 0.17839814722537994, 'train/grad_norm_pre_clip': 11.277693748474121, 'timing/data': 0.05937240459024906, 'timing/model': 4.772862870246172, 'learning_rate/qwen_vl_interface': 2.565811942905899e-06, 'learning_rate/action_model': 2.5658119429058993e-05, 'epoch': 1.87}) +[2026-06-18 04:00:45,586][starVLA.training.train_starvla][INFO] - Step 4870, Loss: {'train/loss': 0.5029112100601196, 'train/grad_norm_pre_clip': 21.1020565032959, 'timing/data': 0.042870599776506424, 'timing/model': 1.28599426522851, 'learning_rate/qwen_vl_interface': 2.5640279414404723e-06, 'learning_rate/action_model': 2.5640279414404722e-05, 'epoch': 1.87}) +[2026-06-18 04:00:52,662][starVLA.training.train_starvla][INFO] - Step 4871, Loss: {'train/loss': 0.20029900968074799, 'train/grad_norm_pre_clip': 8.766590118408203, 'timing/data': 0.04634576290845871, 'timing/model': 1.1439465507864952, 'learning_rate/qwen_vl_interface': 2.5622444967799444e-06, 'learning_rate/action_model': 2.5622444967799443e-05, 'epoch': 1.87}) +[2026-06-18 04:00:57,807][starVLA.training.train_starvla][INFO] - Step 4872, Loss: {'train/loss': 0.252459317445755, 'train/grad_norm_pre_clip': 14.708867073059082, 'timing/data': 0.050481878221035004, 'timing/model': 1.6825108146294951, 'learning_rate/qwen_vl_interface': 2.5604616092940245e-06, 'learning_rate/action_model': 2.5604616092940242e-05, 'epoch': 1.87}) +[2026-06-18 04:01:02,314][starVLA.training.train_starvla][INFO] - Step 4873, Loss: {'train/loss': 0.6098343729972839, 'train/grad_norm_pre_clip': 21.412704467773438, 'timing/data': 0.05339022073894739, 'timing/model': 1.1576712066307664, 'learning_rate/qwen_vl_interface': 2.5586792793523052e-06, 'learning_rate/action_model': 2.5586792793523052e-05, 'epoch': 1.87}) +[2026-06-18 04:01:07,087][starVLA.training.train_starvla][INFO] - Step 4874, Loss: {'train/loss': 0.33321425318717957, 'train/grad_norm_pre_clip': 17.688526153564453, 'timing/data': 0.046818736009299755, 'timing/model': 1.1992847053334117, 'learning_rate/qwen_vl_interface': 2.556897507324271e-06, 'learning_rate/action_model': 2.5568975073242705e-05, 'epoch': 1.87}) +[2026-06-18 04:01:12,753][starVLA.training.train_starvla][INFO] - Step 4875, Loss: {'train/loss': 0.2053092122077942, 'train/grad_norm_pre_clip': 9.529617309570312, 'timing/data': 0.0396365849301219, 'timing/model': 1.1090571749955416, 'learning_rate/qwen_vl_interface': 2.5551162935792782e-06, 'learning_rate/action_model': 2.5551162935792784e-05, 'epoch': 1.88}) +[2026-06-18 04:01:17,454][starVLA.training.train_starvla][INFO] - Step 4876, Loss: {'train/loss': 0.4306573271751404, 'train/grad_norm_pre_clip': 15.594184875488281, 'timing/data': 0.04161663167178631, 'timing/model': 1.127850598655641, 'learning_rate/qwen_vl_interface': 2.5533356384865777e-06, 'learning_rate/action_model': 2.5533356384865776e-05, 'epoch': 1.88}) +[2026-06-18 04:01:22,627][starVLA.training.train_starvla][INFO] - Step 4877, Loss: {'train/loss': 0.5072804689407349, 'train/grad_norm_pre_clip': 17.74726676940918, 'timing/data': 0.04199211671948433, 'timing/model': 1.8402364291250706, 'learning_rate/qwen_vl_interface': 2.5515555424153017e-06, 'learning_rate/action_model': 2.5515555424153016e-05, 'epoch': 1.88}) +[2026-06-18 04:01:32,171][starVLA.training.train_starvla][INFO] - Step 4878, Loss: {'train/loss': 0.4263145327568054, 'train/grad_norm_pre_clip': 14.754714012145996, 'timing/data': 0.035381950438022614, 'timing/model': 1.2059987718239427, 'learning_rate/qwen_vl_interface': 2.5497760057344663e-06, 'learning_rate/action_model': 2.549776005734466e-05, 'epoch': 1.88}) +[2026-06-18 04:01:37,096][starVLA.training.train_starvla][INFO] - Step 4879, Loss: {'train/loss': 0.34693166613578796, 'train/grad_norm_pre_clip': 17.84278106689453, 'timing/data': 0.037773472256958485, 'timing/model': 1.0837981179356575, 'learning_rate/qwen_vl_interface': 2.5479970288129714e-06, 'learning_rate/action_model': 2.5479970288129714e-05, 'epoch': 1.88}) +[2026-06-18 04:01:46,010][starVLA.training.train_starvla][INFO] - Step 4880, Loss: {'train/loss': 0.3886008858680725, 'train/grad_norm_pre_clip': 16.898387908935547, 'timing/data': 0.05170154105871916, 'timing/model': 1.8684204323217273, 'learning_rate/qwen_vl_interface': 2.546218612019597e-06, 'learning_rate/action_model': 2.5462186120195967e-05, 'epoch': 1.88}) +[2026-06-18 04:01:51,514][starVLA.training.train_starvla][INFO] - Step 4881, Loss: {'train/loss': 0.37074440717697144, 'train/grad_norm_pre_clip': 16.699886322021484, 'timing/data': 0.03711484279483557, 'timing/model': 1.8183085080236197, 'learning_rate/qwen_vl_interface': 2.544440755723015e-06, 'learning_rate/action_model': 2.5444407557230147e-05, 'epoch': 1.88}) +[2026-06-18 04:01:57,688][starVLA.training.train_starvla][INFO] - Step 4882, Loss: {'train/loss': 0.44986334443092346, 'train/grad_norm_pre_clip': 13.628799438476562, 'timing/data': 0.023241154849529266, 'timing/model': 1.486726988106966, 'learning_rate/qwen_vl_interface': 2.5426634602917756e-06, 'learning_rate/action_model': 2.5426634602917755e-05, 'epoch': 1.88}) +[2026-06-18 04:02:07,325][starVLA.training.train_starvla][INFO] - Step 4883, Loss: {'train/loss': 0.5257609486579895, 'train/grad_norm_pre_clip': 19.67928123474121, 'timing/data': 0.02930478285998106, 'timing/model': 2.843063616193831, 'learning_rate/qwen_vl_interface': 2.540886726094315e-06, 'learning_rate/action_model': 2.540886726094315e-05, 'epoch': 1.88}) +[2026-06-18 04:02:17,739][starVLA.training.train_starvla][INFO] - Step 4884, Loss: {'train/loss': 0.19858983159065247, 'train/grad_norm_pre_clip': 10.021784782409668, 'timing/data': 0.024892769753932953, 'timing/model': 4.822727809660137, 'learning_rate/qwen_vl_interface': 2.5391105534989503e-06, 'learning_rate/action_model': 2.53911055349895e-05, 'epoch': 1.88}) +[2026-06-18 04:02:22,863][starVLA.training.train_starvla][INFO] - Step 4885, Loss: {'train/loss': 0.2561439871788025, 'train/grad_norm_pre_clip': 10.721709251403809, 'timing/data': 0.02411947399377823, 'timing/model': 1.1097097750753164, 'learning_rate/qwen_vl_interface': 2.537334942873885e-06, 'learning_rate/action_model': 2.5373349428738846e-05, 'epoch': 1.88}) +[2026-06-18 04:02:27,771][starVLA.training.train_starvla][INFO] - Step 4886, Loss: {'train/loss': 0.3897615075111389, 'train/grad_norm_pre_clip': 19.01871681213379, 'timing/data': 0.027404943481087685, 'timing/model': 1.1303886231034994, 'learning_rate/qwen_vl_interface': 2.535559894587205e-06, 'learning_rate/action_model': 2.535559894587205e-05, 'epoch': 1.88}) +[2026-06-18 04:02:33,738][starVLA.training.train_starvla][INFO] - Step 4887, Loss: {'train/loss': 0.336434006690979, 'train/grad_norm_pre_clip': 14.755736351013184, 'timing/data': 0.025362168438732624, 'timing/model': 1.7154871644452214, 'learning_rate/qwen_vl_interface': 2.533785409006878e-06, 'learning_rate/action_model': 2.533785409006878e-05, 'epoch': 1.88}) +[2026-06-18 04:02:38,163][starVLA.training.train_starvla][INFO] - Step 4888, Loss: {'train/loss': 0.19767460227012634, 'train/grad_norm_pre_clip': 12.71562671661377, 'timing/data': 0.039817326702177525, 'timing/model': 1.1175279123708606, 'learning_rate/qwen_vl_interface': 2.532011486500762e-06, 'learning_rate/action_model': 2.532011486500762e-05, 'epoch': 1.88}) +[2026-06-18 04:02:42,928][starVLA.training.train_starvla][INFO] - Step 4889, Loss: {'train/loss': 0.20692525804042816, 'train/grad_norm_pre_clip': 10.491410255432129, 'timing/data': 0.026330605149269104, 'timing/model': 1.106728640384972, 'learning_rate/qwen_vl_interface': 2.530238127436588e-06, 'learning_rate/action_model': 2.5302381274365882e-05, 'epoch': 1.88}) +[2026-06-18 04:02:47,759][starVLA.training.train_starvla][INFO] - Step 4890, Loss: {'train/loss': 0.3945106267929077, 'train/grad_norm_pre_clip': 18.298818588256836, 'timing/data': 0.030220544897019863, 'timing/model': 1.0838735485449433, 'learning_rate/qwen_vl_interface': 2.528465332181977e-06, 'learning_rate/action_model': 2.5284653321819767e-05, 'epoch': 1.88}) +[2026-06-18 04:02:53,334][starVLA.training.train_starvla][INFO] - Step 4891, Loss: {'train/loss': 0.2906804084777832, 'train/grad_norm_pre_clip': 13.223006248474121, 'timing/data': 0.02183704823255539, 'timing/model': 1.1299472823739052, 'learning_rate/qwen_vl_interface': 2.526693101104429e-06, 'learning_rate/action_model': 2.526693101104429e-05, 'epoch': 1.88}) +[2026-06-18 04:02:58,252][starVLA.training.train_starvla][INFO] - Step 4892, Loss: {'train/loss': 0.32598766684532166, 'train/grad_norm_pre_clip': 35.34787368774414, 'timing/data': 0.026526352390646935, 'timing/model': 1.6585220275446773, 'learning_rate/qwen_vl_interface': 2.524921434571333e-06, 'learning_rate/action_model': 2.524921434571333e-05, 'epoch': 1.88}) +[2026-06-18 04:03:02,853][starVLA.training.train_starvla][INFO] - Step 4893, Loss: {'train/loss': 0.30839282274246216, 'train/grad_norm_pre_clip': 13.799566268920898, 'timing/data': 0.020574812777340412, 'timing/model': 1.1074478421360254, 'learning_rate/qwen_vl_interface': 2.5231503329499585e-06, 'learning_rate/action_model': 2.5231503329499583e-05, 'epoch': 1.88}) +[2026-06-18 04:03:07,870][starVLA.training.train_starvla][INFO] - Step 4894, Loss: {'train/loss': 0.23801235854625702, 'train/grad_norm_pre_clip': 14.077313423156738, 'timing/data': 0.024957983754575253, 'timing/model': 1.2859269743785262, 'learning_rate/qwen_vl_interface': 2.5213797966074506e-06, 'learning_rate/action_model': 2.5213797966074505e-05, 'epoch': 1.88}) +[2026-06-18 04:03:12,775][starVLA.training.train_starvla][INFO] - Step 4895, Loss: {'train/loss': 0.504443347454071, 'train/grad_norm_pre_clip': 13.254526138305664, 'timing/data': 0.020392400212585926, 'timing/model': 1.1874167453497648, 'learning_rate/qwen_vl_interface': 2.519609825910848e-06, 'learning_rate/action_model': 2.5196098259108482e-05, 'epoch': 1.88}) +[2026-06-18 04:03:19,460][starVLA.training.train_starvla][INFO] - Step 4896, Loss: {'train/loss': 0.49166634678840637, 'train/grad_norm_pre_clip': 16.836681365966797, 'timing/data': 0.03278813511133194, 'timing/model': 1.2046730583533645, 'learning_rate/qwen_vl_interface': 2.517840421227067e-06, 'learning_rate/action_model': 2.517840421227067e-05, 'epoch': 1.88}) +[2026-06-18 04:03:24,236][starVLA.training.train_starvla][INFO] - Step 4897, Loss: {'train/loss': 0.33902591466903687, 'train/grad_norm_pre_clip': 15.095093727111816, 'timing/data': 0.0223108297213912, 'timing/model': 1.1374657964333892, 'learning_rate/qwen_vl_interface': 2.5160715829229067e-06, 'learning_rate/action_model': 2.5160715829229066e-05, 'epoch': 1.88}) +[2026-06-18 04:03:29,051][starVLA.training.train_starvla][INFO] - Step 4898, Loss: {'train/loss': 0.32644331455230713, 'train/grad_norm_pre_clip': 11.069685935974121, 'timing/data': 0.02660861611366272, 'timing/model': 1.1501866541802883, 'learning_rate/qwen_vl_interface': 2.5143033113650484e-06, 'learning_rate/action_model': 2.5143033113650483e-05, 'epoch': 1.88}) +[2026-06-18 04:03:34,073][starVLA.training.train_starvla][INFO] - Step 4899, Loss: {'train/loss': 0.5118903517723083, 'train/grad_norm_pre_clip': 32.71146011352539, 'timing/data': 0.02197800576686859, 'timing/model': 1.1211821986362338, 'learning_rate/qwen_vl_interface': 2.5125356069200572e-06, 'learning_rate/action_model': 2.5125356069200574e-05, 'epoch': 1.88}) +[2026-06-18 04:07:15,068][starVLA.training.train_starvla][INFO] - Step 4900, Loss: {'train/loss': 0.350433886051178, 'train/grad_norm_pre_clip': 17.053674697875977, 'eval/loss': 0.43588516116142273, 'eval/latency_4/loss': 0.43588516116142273, 'eval/demon_fire_precision': 0.9650015908367802, 'eval/demon_fire_recall': 0.985060084443001, 'eval/demon_fire_f1': 0.9749276759884281, 'eval/demon_fire_rate_model': 0.9230543318649046, 'eval/demon_fire_rate_teacher': 0.9042584434654919, 'eval/demon_left_precision': 0.7786885245901639, 'eval/demon_left_recall': 0.6506849315068494, 'eval/demon_left_f1': 0.7089552238805971, 'eval/demon_left_rate_model': 0.10748898678414097, 'eval/demon_left_rate_teacher': 0.12863436123348018, 'eval/demon_right_precision': 0.8471337579617835, 'eval/demon_right_recall': 0.720216606498195, 'eval/demon_right_f1': 0.7785365853658537, 'eval/demon_right_rate_model': 0.13832599118942732, 'eval/demon_right_rate_teacher': 0.16270190895741557, 'eval/demon_move_f1': 0.7437459046232253, 'eval/demon_cc_f1': 0.8917698653048005, 'eval/latency_4/demon_fire_precision': 0.9650015908367802, 'eval/latency_4/demon_fire_recall': 0.985060084443001, 'eval/latency_4/demon_fire_f1': 0.9749276759884281, 'eval/latency_4/demon_fire_rate_model': 0.9230543318649046, 'eval/latency_4/demon_fire_rate_teacher': 0.9042584434654919, 'eval/latency_4/demon_left_precision': 0.7786885245901639, 'eval/latency_4/demon_left_recall': 0.6506849315068494, 'eval/latency_4/demon_left_f1': 0.7089552238805971, 'eval/latency_4/demon_left_rate_model': 0.10748898678414097, 'eval/latency_4/demon_left_rate_teacher': 0.12863436123348018, 'eval/latency_4/demon_right_precision': 0.8471337579617835, 'eval/latency_4/demon_right_recall': 0.720216606498195, 'eval/latency_4/demon_right_f1': 0.7785365853658537, 'eval/latency_4/demon_right_rate_model': 0.13832599118942732, 'eval/latency_4/demon_right_rate_teacher': 0.16270190895741557, 'eval/latency_4/demon_move_f1': 0.7437459046232253, 'eval/latency_4/demon_cc_f1': 0.8917698653048005, 'timing/data': 0.022565850988030434, 'timing/model': 1.1011605458334088, 'learning_rate/qwen_vl_interface': 2.5107684699543817e-06, 'learning_rate/action_model': 2.5107684699543816e-05, 'epoch': 1.88}) +[2026-06-18 04:07:15,086][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_4900_state +[2026-06-18 04:07:15,096][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 04:07:24,248][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_4900_state/model.safetensors +[2026-06-18 04:07:47,458][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_4900_state/optimizer.bin +[2026-06-18 04:07:47,459][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_4900_state/sampler.bin +[2026-06-18 04:07:47,460][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_4900_state/sampler_1.bin +[2026-06-18 04:07:47,463][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_4900_state/random_states_0.pkl +[2026-06-18 04:07:47,473][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4900 +[2026-06-18 04:07:47,473][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 04:07:47,490][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 04:07:52,775][starVLA.training.train_starvla][INFO] - Step 4901, Loss: {'train/loss': 0.4755854606628418, 'train/grad_norm_pre_clip': 15.614632606506348, 'timing/data': 0.03496265318244696, 'timing/model': 1.757151335477829, 'learning_rate/qwen_vl_interface': 2.509001900834349e-06, 'learning_rate/action_model': 2.5090019008343492e-05, 'epoch': 1.89}) +[2026-06-18 04:07:57,221][starVLA.training.train_starvla][INFO] - Step 4902, Loss: {'train/loss': 0.31825631856918335, 'train/grad_norm_pre_clip': 23.071117401123047, 'timing/data': 0.029712560586631298, 'timing/model': 1.103780667297542, 'learning_rate/qwen_vl_interface': 2.507235899926172e-06, 'learning_rate/action_model': 2.5072358999261718e-05, 'epoch': 1.89}) +[2026-06-18 04:08:02,229][starVLA.training.train_starvla][INFO] - Step 4903, Loss: {'train/loss': 0.4920245409011841, 'train/grad_norm_pre_clip': 20.420080184936523, 'timing/data': 0.03448182623833418, 'timing/model': 1.112365740351379, 'learning_rate/qwen_vl_interface': 2.5054704675959445e-06, 'learning_rate/action_model': 2.5054704675959445e-05, 'epoch': 1.89}) +[2026-06-18 04:08:07,268][starVLA.training.train_starvla][INFO] - Step 4904, Loss: {'train/loss': 0.40470054745674133, 'train/grad_norm_pre_clip': 11.726896286010742, 'timing/data': 0.0258294939994812, 'timing/model': 1.2869657929986715, 'learning_rate/qwen_vl_interface': 2.5037056042096418e-06, 'learning_rate/action_model': 2.5037056042096417e-05, 'epoch': 1.89}) +[2026-06-18 04:08:18,140][starVLA.training.train_starvla][INFO] - Step 4905, Loss: {'train/loss': 0.18281394243240356, 'train/grad_norm_pre_clip': 10.609879493713379, 'timing/data': 0.025350715965032578, 'timing/model': 5.827364420518279, 'learning_rate/qwen_vl_interface': 2.5019413101331212e-06, 'learning_rate/action_model': 2.5019413101331214e-05, 'epoch': 1.89}) +[2026-06-18 04:08:23,668][starVLA.training.train_starvla][INFO] - Step 4906, Loss: {'train/loss': 0.295081228017807, 'train/grad_norm_pre_clip': 12.770549774169922, 'timing/data': 0.04226426687091589, 'timing/model': 1.7458183392882347, 'learning_rate/qwen_vl_interface': 2.5001775857321254e-06, 'learning_rate/action_model': 2.5001775857321253e-05, 'epoch': 1.89}) +[2026-06-18 04:08:28,936][starVLA.training.train_starvla][INFO] - Step 4907, Loss: {'train/loss': 0.2852526903152466, 'train/grad_norm_pre_clip': 14.159701347351074, 'timing/data': 0.023008246906101704, 'timing/model': 1.5027832798659801, 'learning_rate/qwen_vl_interface': 2.4984144313722765e-06, 'learning_rate/action_model': 2.4984144313722766e-05, 'epoch': 1.89}) +[2026-06-18 04:08:37,075][starVLA.training.train_starvla][INFO] - Step 4908, Loss: {'train/loss': 0.4415198266506195, 'train/grad_norm_pre_clip': 16.1434326171875, 'timing/data': 0.022852175869047642, 'timing/model': 1.1321191033348441, 'learning_rate/qwen_vl_interface': 2.496651847419076e-06, 'learning_rate/action_model': 2.496651847419076e-05, 'epoch': 1.89}) +[2026-06-18 04:08:45,420][starVLA.training.train_starvla][INFO] - Step 4909, Loss: {'train/loss': 0.6002432703971863, 'train/grad_norm_pre_clip': 16.421398162841797, 'timing/data': 0.028529677540063858, 'timing/model': 1.2070261053740978, 'learning_rate/qwen_vl_interface': 2.4948898342379076e-06, 'learning_rate/action_model': 2.4948898342379075e-05, 'epoch': 1.89}) +[2026-06-18 04:08:51,311][starVLA.training.train_starvla][INFO] - Step 4910, Loss: {'train/loss': 0.6132147312164307, 'train/grad_norm_pre_clip': 20.218164443969727, 'timing/data': 0.042578792199492455, 'timing/model': 1.2646258538588881, 'learning_rate/qwen_vl_interface': 2.4931283921940434e-06, 'learning_rate/action_model': 2.4931283921940434e-05, 'epoch': 1.89}) +[2026-06-18 04:08:56,259][starVLA.training.train_starvla][INFO] - Step 4911, Loss: {'train/loss': 0.3056808114051819, 'train/grad_norm_pre_clip': 12.609472274780273, 'timing/data': 0.022949058562517166, 'timing/model': 1.1412717085331678, 'learning_rate/qwen_vl_interface': 2.4913675216526302e-06, 'learning_rate/action_model': 2.4913675216526304e-05, 'epoch': 1.89}) +[2026-06-18 04:09:04,151][starVLA.training.train_starvla][INFO] - Step 4912, Loss: {'train/loss': 0.3436796963214874, 'train/grad_norm_pre_clip': 16.033855438232422, 'timing/data': 0.021314511075615883, 'timing/model': 4.528649354353547, 'learning_rate/qwen_vl_interface': 2.4896072229786996e-06, 'learning_rate/action_model': 2.4896072229786995e-05, 'epoch': 1.89}) +[2026-06-18 04:09:09,146][starVLA.training.train_starvla][INFO] - Step 4913, Loss: {'train/loss': 0.23098017275333405, 'train/grad_norm_pre_clip': 11.812832832336426, 'timing/data': 0.033837188966572285, 'timing/model': 1.2375960191711783, 'learning_rate/qwen_vl_interface': 2.4878474965371587e-06, 'learning_rate/action_model': 2.4878474965371586e-05, 'epoch': 1.89}) +[2026-06-18 04:09:15,010][starVLA.training.train_starvla][INFO] - Step 4914, Loss: {'train/loss': 0.2030562460422516, 'train/grad_norm_pre_clip': 6.8075270652771, 'timing/data': 0.03184737078845501, 'timing/model': 1.1758672464638948, 'learning_rate/qwen_vl_interface': 2.4860883426928053e-06, 'learning_rate/action_model': 2.4860883426928053e-05, 'epoch': 1.89}) +[2026-06-18 04:09:20,109][starVLA.training.train_starvla][INFO] - Step 4915, Loss: {'train/loss': 0.4395428001880646, 'train/grad_norm_pre_clip': 14.797696113586426, 'timing/data': 0.02238734718412161, 'timing/model': 1.2376565858721733, 'learning_rate/qwen_vl_interface': 2.4843297618103127e-06, 'learning_rate/action_model': 2.4843297618103128e-05, 'epoch': 1.89}) +[2026-06-18 04:09:24,834][starVLA.training.train_starvla][INFO] - Step 4916, Loss: {'train/loss': 0.1906118243932724, 'train/grad_norm_pre_clip': 13.788942337036133, 'timing/data': 0.023394224233925343, 'timing/model': 1.0985654890537262, 'learning_rate/qwen_vl_interface': 2.482571754254237e-06, 'learning_rate/action_model': 2.4825717542542372e-05, 'epoch': 1.89}) +[2026-06-18 04:09:30,632][starVLA.training.train_starvla][INFO] - Step 4917, Loss: {'train/loss': 0.3172896206378937, 'train/grad_norm_pre_clip': 15.271515846252441, 'timing/data': 0.024149740114808083, 'timing/model': 1.9436198621988297, 'learning_rate/qwen_vl_interface': 2.4808143203890137e-06, 'learning_rate/action_model': 2.480814320389014e-05, 'epoch': 1.89}) +[2026-06-18 04:09:36,091][starVLA.training.train_starvla][INFO] - Step 4918, Loss: {'train/loss': 0.5197256207466125, 'train/grad_norm_pre_clip': 17.48341178894043, 'timing/data': 0.025190898217260838, 'timing/model': 1.3765157759189606, 'learning_rate/qwen_vl_interface': 2.4790574605789627e-06, 'learning_rate/action_model': 2.4790574605789624e-05, 'epoch': 1.89}) +[2026-06-18 04:09:42,303][starVLA.training.train_starvla][INFO] - Step 4919, Loss: {'train/loss': 0.41316983103752136, 'train/grad_norm_pre_clip': 10.629470825195312, 'timing/data': 0.028844112530350685, 'timing/model': 1.3378151003271341, 'learning_rate/qwen_vl_interface': 2.4773011751882807e-06, 'learning_rate/action_model': 2.4773011751882804e-05, 'epoch': 1.89}) +[2026-06-18 04:09:55,533][starVLA.training.train_starvla][INFO] - Step 4920, Loss: {'train/loss': 0.22300618886947632, 'train/grad_norm_pre_clip': 9.920555114746094, 'timing/data': 0.041642322205007076, 'timing/model': 1.55281712859869, 'learning_rate/qwen_vl_interface': 2.4755454645810473e-06, 'learning_rate/action_model': 2.4755454645810473e-05, 'epoch': 1.89}) +[2026-06-18 04:10:00,874][starVLA.training.train_starvla][INFO] - Step 4921, Loss: {'train/loss': 0.631253719329834, 'train/grad_norm_pre_clip': 16.967477798461914, 'timing/data': 0.030472032725811005, 'timing/model': 1.5307287340983748, 'learning_rate/qwen_vl_interface': 2.473790329121228e-06, 'learning_rate/action_model': 2.473790329121228e-05, 'epoch': 1.89}) +[2026-06-18 04:10:06,165][starVLA.training.train_starvla][INFO] - Step 4922, Loss: {'train/loss': 0.47800880670547485, 'train/grad_norm_pre_clip': 22.491680145263672, 'timing/data': 0.023885851725935936, 'timing/model': 1.5614324295893312, 'learning_rate/qwen_vl_interface': 2.4720357691726595e-06, 'learning_rate/action_model': 2.4720357691726594e-05, 'epoch': 1.89}) +[2026-06-18 04:10:11,226][starVLA.training.train_starvla][INFO] - Step 4923, Loss: {'train/loss': 0.18691569566726685, 'train/grad_norm_pre_clip': 13.079710960388184, 'timing/data': 0.022974749095737934, 'timing/model': 1.2353553948923945, 'learning_rate/qwen_vl_interface': 2.4702817850990645e-06, 'learning_rate/action_model': 2.470281785099064e-05, 'epoch': 1.89}) +[2026-06-18 04:10:17,936][starVLA.training.train_starvla][INFO] - Step 4924, Loss: {'train/loss': 0.41300925612449646, 'train/grad_norm_pre_clip': 13.533001899719238, 'timing/data': 0.027438906021416187, 'timing/model': 2.7796793691813946, 'learning_rate/qwen_vl_interface': 2.4685283772640452e-06, 'learning_rate/action_model': 2.468528377264045e-05, 'epoch': 1.89}) +[2026-06-18 04:10:26,105][starVLA.training.train_starvla][INFO] - Step 4925, Loss: {'train/loss': 0.3515841066837311, 'train/grad_norm_pre_clip': 13.453935623168945, 'timing/data': 0.07545975502580404, 'timing/model': 1.2763994615525007, 'learning_rate/qwen_vl_interface': 2.466775546031088e-06, 'learning_rate/action_model': 2.4667755460310878e-05, 'epoch': 1.89}) +[2026-06-18 04:10:37,396][starVLA.training.train_starvla][INFO] - Step 4926, Loss: {'train/loss': 0.3775266408920288, 'train/grad_norm_pre_clip': 20.051631927490234, 'timing/data': 0.042729614302515984, 'timing/model': 1.5426817229017615, 'learning_rate/qwen_vl_interface': 2.465023291763555e-06, 'learning_rate/action_model': 2.465023291763555e-05, 'epoch': 1.89}) +[2026-06-18 04:10:44,562][starVLA.training.train_starvla][INFO] - Step 4927, Loss: {'train/loss': 0.436390221118927, 'train/grad_norm_pre_clip': 13.706124305725098, 'timing/data': 0.02347488235682249, 'timing/model': 3.487034058198333, 'learning_rate/qwen_vl_interface': 2.4632716148246903e-06, 'learning_rate/action_model': 2.4632716148246903e-05, 'epoch': 1.9}) +[2026-06-18 04:10:55,561][starVLA.training.train_starvla][INFO] - Step 4928, Loss: {'train/loss': 0.41565191745758057, 'train/grad_norm_pre_clip': 14.845184326171875, 'timing/data': 0.023402818478643894, 'timing/model': 1.2448676051571965, 'learning_rate/qwen_vl_interface': 2.4615205155776185e-06, 'learning_rate/action_model': 2.4615205155776188e-05, 'epoch': 1.9}) +[2026-06-18 04:11:01,411][starVLA.training.train_starvla][INFO] - Step 4929, Loss: {'train/loss': 0.8531330823898315, 'train/grad_norm_pre_clip': 19.852890014648438, 'timing/data': 0.022314720787107944, 'timing/model': 1.194839513860643, 'learning_rate/qwen_vl_interface': 2.4597699943853444e-06, 'learning_rate/action_model': 2.4597699943853442e-05, 'epoch': 1.9}) +[2026-06-18 04:11:06,243][starVLA.training.train_starvla][INFO] - Step 4930, Loss: {'train/loss': 0.07666395604610443, 'train/grad_norm_pre_clip': 4.450018405914307, 'timing/data': 0.025598476640880108, 'timing/model': 1.1079577775672078, 'learning_rate/qwen_vl_interface': 2.458020051610753e-06, 'learning_rate/action_model': 2.4580200516107528e-05, 'epoch': 1.9}) +[2026-06-18 04:11:11,189][starVLA.training.train_starvla][INFO] - Step 4931, Loss: {'train/loss': 0.27416542172431946, 'train/grad_norm_pre_clip': 11.082051277160645, 'timing/data': 0.026465541683137417, 'timing/model': 1.4717810051515698, 'learning_rate/qwen_vl_interface': 2.456270687616609e-06, 'learning_rate/action_model': 2.4562706876166087e-05, 'epoch': 1.9}) +[2026-06-18 04:11:16,116][starVLA.training.train_starvla][INFO] - Step 4932, Loss: {'train/loss': 0.45268070697784424, 'train/grad_norm_pre_clip': 16.794370651245117, 'timing/data': 0.029668615199625492, 'timing/model': 1.5434919595718384, 'learning_rate/qwen_vl_interface': 2.4545219027655577e-06, 'learning_rate/action_model': 2.4545219027655574e-05, 'epoch': 1.9}) +[2026-06-18 04:11:20,663][starVLA.training.train_starvla][INFO] - Step 4933, Loss: {'train/loss': 0.505120038986206, 'train/grad_norm_pre_clip': 21.131553649902344, 'timing/data': 0.035386729054152966, 'timing/model': 1.2220342308282852, 'learning_rate/qwen_vl_interface': 2.452773697420124e-06, 'learning_rate/action_model': 2.452773697420124e-05, 'epoch': 1.9}) +[2026-06-18 04:11:25,485][starVLA.training.train_starvla][INFO] - Step 4934, Loss: {'train/loss': 0.5728542804718018, 'train/grad_norm_pre_clip': 18.90546417236328, 'timing/data': 0.03630629554390907, 'timing/model': 1.1039409590885043, 'learning_rate/qwen_vl_interface': 2.4510260719427127e-06, 'learning_rate/action_model': 2.4510260719427127e-05, 'epoch': 1.9}) +[2026-06-18 04:11:30,231][starVLA.training.train_starvla][INFO] - Step 4935, Loss: {'train/loss': 0.20283232629299164, 'train/grad_norm_pre_clip': 7.862056732177734, 'timing/data': 0.025759435258805752, 'timing/model': 1.1487813144922256, 'learning_rate/qwen_vl_interface': 2.4492790266956092e-06, 'learning_rate/action_model': 2.449279026695609e-05, 'epoch': 1.9}) +[2026-06-18 04:11:35,386][starVLA.training.train_starvla][INFO] - Step 4936, Loss: {'train/loss': 0.5174722671508789, 'train/grad_norm_pre_clip': 14.892075538635254, 'timing/data': 0.024674183689057827, 'timing/model': 1.2003523269668221, 'learning_rate/qwen_vl_interface': 2.4475325620409765e-06, 'learning_rate/action_model': 2.4475325620409764e-05, 'epoch': 1.9}) +[2026-06-18 04:11:41,032][starVLA.training.train_starvla][INFO] - Step 4937, Loss: {'train/loss': 0.4496456980705261, 'train/grad_norm_pre_clip': 16.048931121826172, 'timing/data': 0.02345594670623541, 'timing/model': 1.765495203435421, 'learning_rate/qwen_vl_interface': 2.445786678340859e-06, 'learning_rate/action_model': 2.4457866783408588e-05, 'epoch': 1.9}) +[2026-06-18 04:11:45,826][starVLA.training.train_starvla][INFO] - Step 4938, Loss: {'train/loss': 0.793140709400177, 'train/grad_norm_pre_clip': 21.951120376586914, 'timing/data': 0.025428108870983124, 'timing/model': 1.1202438343316317, 'learning_rate/qwen_vl_interface': 2.444041375957179e-06, 'learning_rate/action_model': 2.444041375957179e-05, 'epoch': 1.9}) +[2026-06-18 04:11:50,651][starVLA.training.train_starvla][INFO] - Step 4939, Loss: {'train/loss': 0.312213659286499, 'train/grad_norm_pre_clip': 15.31972885131836, 'timing/data': 0.023719551041722298, 'timing/model': 1.149070618674159, 'learning_rate/qwen_vl_interface': 2.4422966552517434e-06, 'learning_rate/action_model': 2.4422966552517433e-05, 'epoch': 1.9}) +[2026-06-18 04:11:55,272][starVLA.training.train_starvla][INFO] - Step 4940, Loss: {'train/loss': 0.24258503317832947, 'train/grad_norm_pre_clip': 10.884180068969727, 'timing/data': 0.03580101113766432, 'timing/model': 1.1227249978110194, 'learning_rate/qwen_vl_interface': 2.440552516586232e-06, 'learning_rate/action_model': 2.440552516586232e-05, 'epoch': 1.9}) +[2026-06-18 04:12:00,321][starVLA.training.train_starvla][INFO] - Step 4941, Loss: {'train/loss': 0.2758232355117798, 'train/grad_norm_pre_clip': 13.418182373046875, 'timing/data': 0.024195905774831772, 'timing/model': 1.5945932380855083, 'learning_rate/qwen_vl_interface': 2.4388089603222085e-06, 'learning_rate/action_model': 2.4388089603222087e-05, 'epoch': 1.9}) +[2026-06-18 04:12:05,003][starVLA.training.train_starvla][INFO] - Step 4942, Loss: {'train/loss': 0.2994106411933899, 'train/grad_norm_pre_clip': 17.036338806152344, 'timing/data': 0.023634460754692554, 'timing/model': 1.4599385783076286, 'learning_rate/qwen_vl_interface': 2.4370659868211095e-06, 'learning_rate/action_model': 2.4370659868211097e-05, 'epoch': 1.9}) +[2026-06-18 04:12:09,485][starVLA.training.train_starvla][INFO] - Step 4943, Loss: {'train/loss': 0.519771158695221, 'train/grad_norm_pre_clip': 14.917059898376465, 'timing/data': 0.02549831662327051, 'timing/model': 1.1344034019857645, 'learning_rate/qwen_vl_interface': 2.4353235964442597e-06, 'learning_rate/action_model': 2.4353235964442598e-05, 'epoch': 1.9}) +[2026-06-18 04:12:14,170][starVLA.training.train_starvla][INFO] - Step 4944, Loss: {'train/loss': 0.43031859397888184, 'train/grad_norm_pre_clip': 14.778963088989258, 'timing/data': 0.026814267970621586, 'timing/model': 1.2232751809060574, 'learning_rate/qwen_vl_interface': 2.433581789552857e-06, 'learning_rate/action_model': 2.433581789552857e-05, 'epoch': 1.9}) +[2026-06-18 04:12:22,684][starVLA.training.train_starvla][INFO] - Step 4945, Loss: {'train/loss': 0.3410551846027374, 'train/grad_norm_pre_clip': 13.205836296081543, 'timing/data': 0.02323381695896387, 'timing/model': 3.098118619993329, 'learning_rate/qwen_vl_interface': 2.4318405665079806e-06, 'learning_rate/action_model': 2.4318405665079805e-05, 'epoch': 1.9}) +[2026-06-18 04:12:32,282][starVLA.training.train_starvla][INFO] - Step 4946, Loss: {'train/loss': 0.24846963584423065, 'train/grad_norm_pre_clip': 17.51300811767578, 'timing/data': 0.023473476991057396, 'timing/model': 1.9151245141401887, 'learning_rate/qwen_vl_interface': 2.4300999276705868e-06, 'learning_rate/action_model': 2.4300999276705868e-05, 'epoch': 1.9}) +[2026-06-18 04:12:38,975][starVLA.training.train_starvla][INFO] - Step 4947, Loss: {'train/loss': 0.34733858704566956, 'train/grad_norm_pre_clip': 17.59815216064453, 'timing/data': 0.02507948223501444, 'timing/model': 1.5167178688570857, 'learning_rate/qwen_vl_interface': 2.428359873401512e-06, 'learning_rate/action_model': 2.4283598734015124e-05, 'epoch': 1.9}) +[2026-06-18 04:12:51,236][starVLA.training.train_starvla][INFO] - Step 4948, Loss: {'train/loss': 0.24563689529895782, 'train/grad_norm_pre_clip': 13.54305648803711, 'timing/data': 0.024439796805381775, 'timing/model': 1.1918437220156193, 'learning_rate/qwen_vl_interface': 2.4266204040614722e-06, 'learning_rate/action_model': 2.4266204040614724e-05, 'epoch': 1.9}) +[2026-06-18 04:13:01,972][starVLA.training.train_starvla][INFO] - Step 4949, Loss: {'train/loss': 0.41458943486213684, 'train/grad_norm_pre_clip': 14.868435859680176, 'timing/data': 0.029576021246612072, 'timing/model': 1.400942767970264, 'learning_rate/qwen_vl_interface': 2.424881520011059e-06, 'learning_rate/action_model': 2.424881520011059e-05, 'epoch': 1.9}) +[2026-06-18 04:13:12,006][starVLA.training.train_starvla][INFO] - Step 4950, Loss: {'train/loss': 0.39130350947380066, 'train/grad_norm_pre_clip': 13.269499778747559, 'timing/data': 0.026724320836365223, 'timing/model': 1.1451665768399835, 'learning_rate/qwen_vl_interface': 2.4231432216107494e-06, 'learning_rate/action_model': 2.4231432216107494e-05, 'epoch': 1.9}) +[2026-06-18 04:13:20,454][starVLA.training.train_starvla][INFO] - Step 4951, Loss: {'train/loss': 0.23410427570343018, 'train/grad_norm_pre_clip': 10.656230926513672, 'timing/data': 0.024242663756012917, 'timing/model': 1.1060722628608346, 'learning_rate/qwen_vl_interface': 2.4214055092208907e-06, 'learning_rate/action_model': 2.4214055092208907e-05, 'epoch': 1.9}) +[2026-06-18 04:13:26,951][starVLA.training.train_starvla][INFO] - Step 4952, Loss: {'train/loss': 0.35003501176834106, 'train/grad_norm_pre_clip': 15.14682388305664, 'timing/data': 0.024356430396437645, 'timing/model': 1.5340470466762781, 'learning_rate/qwen_vl_interface': 2.4196683832017134e-06, 'learning_rate/action_model': 2.4196683832017135e-05, 'epoch': 1.9}) +[2026-06-18 04:13:31,386][starVLA.training.train_starvla][INFO] - Step 4953, Loss: {'train/loss': 0.3344222605228424, 'train/grad_norm_pre_clip': 13.1566743850708, 'timing/data': 0.030668284744024277, 'timing/model': 1.147855474613607, 'learning_rate/qwen_vl_interface': 2.4179318439133245e-06, 'learning_rate/action_model': 2.417931843913324e-05, 'epoch': 1.91}) +[2026-06-18 04:13:36,411][starVLA.training.train_starvla][INFO] - Step 4954, Loss: {'train/loss': 0.3422248661518097, 'train/grad_norm_pre_clip': 12.560558319091797, 'timing/data': 0.02968636527657509, 'timing/model': 1.2248581377789378, 'learning_rate/qwen_vl_interface': 2.4161958917157126e-06, 'learning_rate/action_model': 2.4161958917157126e-05, 'epoch': 1.91}) +[2026-06-18 04:13:41,158][starVLA.training.train_starvla][INFO] - Step 4955, Loss: {'train/loss': 0.4589127004146576, 'train/grad_norm_pre_clip': 15.6913423538208, 'timing/data': 0.02587416861206293, 'timing/model': 1.0785097386687994, 'learning_rate/qwen_vl_interface': 2.4144605269687445e-06, 'learning_rate/action_model': 2.4144605269687446e-05, 'epoch': 1.91}) +[2026-06-18 04:13:45,978][starVLA.training.train_starvla][INFO] - Step 4956, Loss: {'train/loss': 0.37709322571754456, 'train/grad_norm_pre_clip': 17.825828552246094, 'timing/data': 0.028101686388254166, 'timing/model': 1.079979546368122, 'learning_rate/qwen_vl_interface': 2.412725750032156e-06, 'learning_rate/action_model': 2.4127257500321563e-05, 'epoch': 1.91}) +[2026-06-18 04:13:50,887][starVLA.training.train_starvla][INFO] - Step 4957, Loss: {'train/loss': 0.6144919395446777, 'train/grad_norm_pre_clip': 18.528377532958984, 'timing/data': 0.03407454025000334, 'timing/model': 1.545895985327661, 'learning_rate/qwen_vl_interface': 2.4109915612655745e-06, 'learning_rate/action_model': 2.4109915612655745e-05, 'epoch': 1.91}) +[2026-06-18 04:13:55,995][starVLA.training.train_starvla][INFO] - Step 4958, Loss: {'train/loss': 0.23073416948318481, 'train/grad_norm_pre_clip': 12.645488739013672, 'timing/data': 0.021616012789309025, 'timing/model': 1.0968106463551521, 'learning_rate/qwen_vl_interface': 2.409257961028498e-06, 'learning_rate/action_model': 2.4092579610284978e-05, 'epoch': 1.91}) +[2026-06-18 04:14:00,945][starVLA.training.train_starvla][INFO] - Step 4959, Loss: {'train/loss': 0.24576954543590546, 'train/grad_norm_pre_clip': 15.243047714233398, 'timing/data': 0.021845036186277866, 'timing/model': 1.1712760431692004, 'learning_rate/qwen_vl_interface': 2.4075249496803026e-06, 'learning_rate/action_model': 2.4075249496803025e-05, 'epoch': 1.91}) +[2026-06-18 04:14:05,747][starVLA.training.train_starvla][INFO] - Step 4960, Loss: {'train/loss': 0.378041535615921, 'train/grad_norm_pre_clip': 13.383391380310059, 'timing/data': 0.03790792543441057, 'timing/model': 1.148361835628748, 'learning_rate/qwen_vl_interface': 2.405792527580243e-06, 'learning_rate/action_model': 2.405792527580243e-05, 'epoch': 1.91}) +[2026-06-18 04:14:10,523][starVLA.training.train_starvla][INFO] - Step 4961, Loss: {'train/loss': 0.42544615268707275, 'train/grad_norm_pre_clip': 18.922374725341797, 'timing/data': 0.0223380159586668, 'timing/model': 1.0685816695913672, 'learning_rate/qwen_vl_interface': 2.404060695087454e-06, 'learning_rate/action_model': 2.4040606950874536e-05, 'epoch': 1.91}) +[2026-06-18 04:14:15,107][starVLA.training.train_starvla][INFO] - Step 4962, Loss: {'train/loss': 0.5281146764755249, 'train/grad_norm_pre_clip': 21.404401779174805, 'timing/data': 0.021344080567359924, 'timing/model': 1.369522087275982, 'learning_rate/qwen_vl_interface': 2.4023294525609437e-06, 'learning_rate/action_model': 2.4023294525609437e-05, 'epoch': 1.91}) +[2026-06-18 04:14:19,422][starVLA.training.train_starvla][INFO] - Step 4963, Loss: {'train/loss': 0.262211412191391, 'train/grad_norm_pre_clip': 14.25436019897461, 'timing/data': 0.023178531788289547, 'timing/model': 1.0751592684537172, 'learning_rate/qwen_vl_interface': 2.400598800359604e-06, 'learning_rate/action_model': 2.4005988003596035e-05, 'epoch': 1.91}) +[2026-06-18 04:14:33,670][starVLA.training.train_starvla][INFO] - Step 4964, Loss: {'train/loss': 0.2750571370124817, 'train/grad_norm_pre_clip': 15.236597061157227, 'timing/data': 0.03012702614068985, 'timing/model': 3.269362053833902, 'learning_rate/qwen_vl_interface': 2.398868738842198e-06, 'learning_rate/action_model': 2.3988687388421978e-05, 'epoch': 1.91}) +[2026-06-18 04:14:38,381][starVLA.training.train_starvla][INFO] - Step 4965, Loss: {'train/loss': 0.24482539296150208, 'train/grad_norm_pre_clip': 10.592981338500977, 'timing/data': 0.023139122873544693, 'timing/model': 1.1330487476661801, 'learning_rate/qwen_vl_interface': 2.3971392683673705e-06, 'learning_rate/action_model': 2.3971392683673706e-05, 'epoch': 1.91}) +[2026-06-18 04:14:43,168][starVLA.training.train_starvla][INFO] - Step 4966, Loss: {'train/loss': 0.24822771549224854, 'train/grad_norm_pre_clip': 31.451974868774414, 'timing/data': 0.02251848205924034, 'timing/model': 1.503229796886444, 'learning_rate/qwen_vl_interface': 2.3954103892936425e-06, 'learning_rate/action_model': 2.3954103892936424e-05, 'epoch': 1.91}) +[2026-06-18 04:14:47,528][starVLA.training.train_starvla][INFO] - Step 4967, Loss: {'train/loss': 0.2047509402036667, 'train/grad_norm_pre_clip': 9.953145980834961, 'timing/data': 0.022771338932216167, 'timing/model': 1.1583635630086064, 'learning_rate/qwen_vl_interface': 2.393682101979411e-06, 'learning_rate/action_model': 2.3936821019794105e-05, 'epoch': 1.91}) +[2026-06-18 04:14:52,229][starVLA.training.train_starvla][INFO] - Step 4968, Loss: {'train/loss': 0.18110990524291992, 'train/grad_norm_pre_clip': 10.341938972473145, 'timing/data': 0.04083364550024271, 'timing/model': 1.074636533856392, 'learning_rate/qwen_vl_interface': 2.3919544067829546e-06, 'learning_rate/action_model': 2.3919544067829545e-05, 'epoch': 1.91}) +[2026-06-18 04:14:57,039][starVLA.training.train_starvla][INFO] - Step 4969, Loss: {'train/loss': 0.46684396266937256, 'train/grad_norm_pre_clip': 14.501993179321289, 'timing/data': 0.022893275134265423, 'timing/model': 1.2401499105617404, 'learning_rate/qwen_vl_interface': 2.3902273040624267e-06, 'learning_rate/action_model': 2.390227304062427e-05, 'epoch': 1.91}) +[2026-06-18 04:15:01,921][starVLA.training.train_starvla][INFO] - Step 4970, Loss: {'train/loss': 0.2013787031173706, 'train/grad_norm_pre_clip': 8.537223815917969, 'timing/data': 0.023523434065282345, 'timing/model': 1.0768700009211898, 'learning_rate/qwen_vl_interface': 2.388500794175855e-06, 'learning_rate/action_model': 2.3885007941758548e-05, 'epoch': 1.91}) +[2026-06-18 04:15:07,542][starVLA.training.train_starvla][INFO] - Step 4971, Loss: {'train/loss': 0.3528805673122406, 'train/grad_norm_pre_clip': 10.823866844177246, 'timing/data': 0.03085031546652317, 'timing/model': 2.307233070023358, 'learning_rate/qwen_vl_interface': 2.386774877481145e-06, 'learning_rate/action_model': 2.3867748774811452e-05, 'epoch': 1.91}) +[2026-06-18 04:15:12,905][starVLA.training.train_starvla][INFO] - Step 4972, Loss: {'train/loss': 0.4357815384864807, 'train/grad_norm_pre_clip': 17.850135803222656, 'timing/data': 0.022928684949874878, 'timing/model': 2.075193121097982, 'learning_rate/qwen_vl_interface': 2.385049554336087e-06, 'learning_rate/action_model': 2.385049554336087e-05, 'epoch': 1.91}) +[2026-06-18 04:15:17,349][starVLA.training.train_starvla][INFO] - Step 4973, Loss: {'train/loss': 0.26040416955947876, 'train/grad_norm_pre_clip': 14.183426856994629, 'timing/data': 0.028278972953557968, 'timing/model': 1.1654755538329482, 'learning_rate/qwen_vl_interface': 2.3833248250983383e-06, 'learning_rate/action_model': 2.3833248250983383e-05, 'epoch': 1.91}) +[2026-06-18 04:15:22,030][starVLA.training.train_starvla][INFO] - Step 4974, Loss: {'train/loss': 0.15180033445358276, 'train/grad_norm_pre_clip': 8.274942398071289, 'timing/data': 0.022812646813690662, 'timing/model': 1.0846766252070665, 'learning_rate/qwen_vl_interface': 2.3816006901254396e-06, 'learning_rate/action_model': 2.3816006901254396e-05, 'epoch': 1.91}) +[2026-06-18 04:15:26,856][starVLA.training.train_starvla][INFO] - Step 4975, Loss: {'train/loss': 0.3667348027229309, 'train/grad_norm_pre_clip': 16.685949325561523, 'timing/data': 0.025504288263618946, 'timing/model': 1.1837913701310754, 'learning_rate/qwen_vl_interface': 2.3798771497748012e-06, 'learning_rate/action_model': 2.379877149774801e-05, 'epoch': 1.91}) +[2026-06-18 04:15:31,587][starVLA.training.train_starvla][INFO] - Step 4976, Loss: {'train/loss': 0.22552798688411713, 'train/grad_norm_pre_clip': 12.85389518737793, 'timing/data': 0.02219512965530157, 'timing/model': 1.1105145961046219, 'learning_rate/qwen_vl_interface': 2.3781542044037195e-06, 'learning_rate/action_model': 2.3781542044037193e-05, 'epoch': 1.91}) +[2026-06-18 04:15:36,228][starVLA.training.train_starvla][INFO] - Step 4977, Loss: {'train/loss': 0.8003940582275391, 'train/grad_norm_pre_clip': 22.66962432861328, 'timing/data': 0.022518150508403778, 'timing/model': 1.3981253253296018, 'learning_rate/qwen_vl_interface': 2.376431854369362e-06, 'learning_rate/action_model': 2.3764318543693618e-05, 'epoch': 1.91}) +[2026-06-18 04:15:40,544][starVLA.training.train_starvla][INFO] - Step 4978, Loss: {'train/loss': 0.2976077198982239, 'train/grad_norm_pre_clip': 16.9880313873291, 'timing/data': 0.02450243104249239, 'timing/model': 1.0958952642977238, 'learning_rate/qwen_vl_interface': 2.374710100028773e-06, 'learning_rate/action_model': 2.3747101000287728e-05, 'epoch': 1.92}) +[2026-06-18 04:15:45,200][starVLA.training.train_starvla][INFO] - Step 4979, Loss: {'train/loss': 0.2560245990753174, 'train/grad_norm_pre_clip': 11.298761367797852, 'timing/data': 0.024615397676825523, 'timing/model': 1.0857311310246587, 'learning_rate/qwen_vl_interface': 2.3729889417388736e-06, 'learning_rate/action_model': 2.3729889417388737e-05, 'epoch': 1.92}) +[2026-06-18 04:15:50,112][starVLA.training.train_starvla][INFO] - Step 4980, Loss: {'train/loss': 0.5549687743186951, 'train/grad_norm_pre_clip': 22.64551544189453, 'timing/data': 0.02590125985443592, 'timing/model': 1.1064799912273884, 'learning_rate/qwen_vl_interface': 2.3712683798564628e-06, 'learning_rate/action_model': 2.3712683798564625e-05, 'epoch': 1.92}) +[2026-06-18 04:15:54,838][starVLA.training.train_starvla][INFO] - Step 4981, Loss: {'train/loss': 0.4075239598751068, 'train/grad_norm_pre_clip': 14.333815574645996, 'timing/data': 0.02346495632082224, 'timing/model': 1.1853738529607654, 'learning_rate/qwen_vl_interface': 2.3695484147382146e-06, 'learning_rate/action_model': 2.3695484147382147e-05, 'epoch': 1.92}) +[2026-06-18 04:16:00,655][starVLA.training.train_starvla][INFO] - Step 4982, Loss: {'train/loss': 0.16397063434123993, 'train/grad_norm_pre_clip': 6.2221455574035645, 'timing/data': 0.02288656122982502, 'timing/model': 1.5461646085605025, 'learning_rate/qwen_vl_interface': 2.367829046740678e-06, 'learning_rate/action_model': 2.3678290467406776e-05, 'epoch': 1.92}) +[2026-06-18 04:16:04,953][starVLA.training.train_starvla][INFO] - Step 4983, Loss: {'train/loss': 0.713340163230896, 'train/grad_norm_pre_clip': 21.088376998901367, 'timing/data': 0.030467439442873, 'timing/model': 1.1563360700383782, 'learning_rate/qwen_vl_interface': 2.3661102762202846e-06, 'learning_rate/action_model': 2.3661102762202843e-05, 'epoch': 1.92}) +[2026-06-18 04:16:09,766][starVLA.training.train_starvla][INFO] - Step 4984, Loss: {'train/loss': 0.34779998660087585, 'train/grad_norm_pre_clip': 11.149412155151367, 'timing/data': 0.02434573881328106, 'timing/model': 1.1707075769081712, 'learning_rate/qwen_vl_interface': 2.364392103533332e-06, 'learning_rate/action_model': 2.364392103533332e-05, 'epoch': 1.92}) +[2026-06-18 04:16:14,719][starVLA.training.train_starvla][INFO] - Step 4985, Loss: {'train/loss': 0.2859387695789337, 'train/grad_norm_pre_clip': 12.574514389038086, 'timing/data': 0.03000317234545946, 'timing/model': 1.090861958451569, 'learning_rate/qwen_vl_interface': 2.362674529036002e-06, 'learning_rate/action_model': 2.362674529036002e-05, 'epoch': 1.92}) +[2026-06-18 04:16:19,357][starVLA.training.train_starvla][INFO] - Step 4986, Loss: {'train/loss': 0.3335638642311096, 'train/grad_norm_pre_clip': 12.24413776397705, 'timing/data': 0.023623931221663952, 'timing/model': 1.1358780767768621, 'learning_rate/qwen_vl_interface': 2.360957553084347e-06, 'learning_rate/action_model': 2.360957553084347e-05, 'epoch': 1.92}) +[2026-06-18 04:16:24,060][starVLA.training.train_starvla][INFO] - Step 4987, Loss: {'train/loss': 0.33796441555023193, 'train/grad_norm_pre_clip': 12.032432556152344, 'timing/data': 0.02111243735998869, 'timing/model': 1.4287740644067526, 'learning_rate/qwen_vl_interface': 2.359241176034302e-06, 'learning_rate/action_model': 2.359241176034302e-05, 'epoch': 1.92}) +[2026-06-18 04:16:28,394][starVLA.training.train_starvla][INFO] - Step 4988, Loss: {'train/loss': 0.32076916098594666, 'train/grad_norm_pre_clip': 15.630389213562012, 'timing/data': 0.022914817556738853, 'timing/model': 1.0835621729493141, 'learning_rate/qwen_vl_interface': 2.3575253982416732e-06, 'learning_rate/action_model': 2.357525398241673e-05, 'epoch': 1.92}) +[2026-06-18 04:16:33,324][starVLA.training.train_starvla][INFO] - Step 4989, Loss: {'train/loss': 0.43116307258605957, 'train/grad_norm_pre_clip': 17.35580062866211, 'timing/data': 0.029578709974884987, 'timing/model': 1.2689814530313015, 'learning_rate/qwen_vl_interface': 2.3558102200621385e-06, 'learning_rate/action_model': 2.3558102200621386e-05, 'epoch': 1.92}) +[2026-06-18 04:16:38,049][starVLA.training.train_starvla][INFO] - Step 4990, Loss: {'train/loss': 0.21435712277889252, 'train/grad_norm_pre_clip': 11.375648498535156, 'timing/data': 0.0420344490557909, 'timing/model': 1.182913820259273, 'learning_rate/qwen_vl_interface': 2.3540956418512607e-06, 'learning_rate/action_model': 2.354095641851261e-05, 'epoch': 1.92}) +[2026-06-18 04:16:42,929][starVLA.training.train_starvla][INFO] - Step 4991, Loss: {'train/loss': 0.4828922748565674, 'train/grad_norm_pre_clip': 15.27099609375, 'timing/data': 0.0242654737085104, 'timing/model': 1.1252355547621846, 'learning_rate/qwen_vl_interface': 2.352381663964473e-06, 'learning_rate/action_model': 2.352381663964473e-05, 'epoch': 1.92}) +[2026-06-18 04:16:47,591][starVLA.training.train_starvla][INFO] - Step 4992, Loss: {'train/loss': 0.5759355425834656, 'train/grad_norm_pre_clip': 18.00914764404297, 'timing/data': 0.02823716402053833, 'timing/model': 1.4599587172269821, 'learning_rate/qwen_vl_interface': 2.3506682867570846e-06, 'learning_rate/action_model': 2.3506682867570844e-05, 'epoch': 1.92}) +[2026-06-18 04:16:51,893][starVLA.training.train_starvla][INFO] - Step 4993, Loss: {'train/loss': 0.4206405282020569, 'train/grad_norm_pre_clip': 14.548079490661621, 'timing/data': 0.02036184724420309, 'timing/model': 1.0885328743606806, 'learning_rate/qwen_vl_interface': 2.3489555105842796e-06, 'learning_rate/action_model': 2.3489555105842794e-05, 'epoch': 1.92}) +[2026-06-18 04:16:56,734][starVLA.training.train_starvla][INFO] - Step 4994, Loss: {'train/loss': 0.23215918242931366, 'train/grad_norm_pre_clip': 11.879910469055176, 'timing/data': 0.02639145217835903, 'timing/model': 1.1307006496936083, 'learning_rate/qwen_vl_interface': 2.347243335801119e-06, 'learning_rate/action_model': 2.347243335801119e-05, 'epoch': 1.92}) +[2026-06-18 04:17:01,513][starVLA.training.train_starvla][INFO] - Step 4995, Loss: {'train/loss': 0.3727963864803314, 'train/grad_norm_pre_clip': 11.803274154663086, 'timing/data': 0.021167734637856483, 'timing/model': 1.0995852425694466, 'learning_rate/qwen_vl_interface': 2.3455317627625385e-06, 'learning_rate/action_model': 2.3455317627625385e-05, 'epoch': 1.92}) +[2026-06-18 04:17:06,156][starVLA.training.train_starvla][INFO] - Step 4996, Loss: {'train/loss': 0.3181796073913574, 'train/grad_norm_pre_clip': 10.460540771484375, 'timing/data': 0.023638121783733368, 'timing/model': 1.128805079497397, 'learning_rate/qwen_vl_interface': 2.3438207918233495e-06, 'learning_rate/action_model': 2.3438207918233494e-05, 'epoch': 1.92}) +[2026-06-18 04:17:11,007][starVLA.training.train_starvla][INFO] - Step 4997, Loss: {'train/loss': 0.3095867335796356, 'train/grad_norm_pre_clip': 11.050179481506348, 'timing/data': 0.024209951981902122, 'timing/model': 1.4994596419855952, 'learning_rate/qwen_vl_interface': 2.3421104233382366e-06, 'learning_rate/action_model': 2.3421104233382364e-05, 'epoch': 1.92}) +[2026-06-18 04:17:15,345][starVLA.training.train_starvla][INFO] - Step 4998, Loss: {'train/loss': 0.17017364501953125, 'train/grad_norm_pre_clip': 9.853048324584961, 'timing/data': 0.023666823282837868, 'timing/model': 1.2006815494969487, 'learning_rate/qwen_vl_interface': 2.340400657661762e-06, 'learning_rate/action_model': 2.340400657661762e-05, 'epoch': 1.92}) +[2026-06-18 04:17:20,012][starVLA.training.train_starvla][INFO] - Step 4999, Loss: {'train/loss': 0.18483716249465942, 'train/grad_norm_pre_clip': 9.498763084411621, 'timing/data': 0.024546519853174686, 'timing/model': 1.1155278524383903, 'learning_rate/qwen_vl_interface': 2.338691495148363e-06, 'learning_rate/action_model': 2.3386914951483626e-05, 'epoch': 1.92}) +[2026-06-18 04:20:41,663][starVLA.training.train_starvla][INFO] - Step 5000, Loss: {'train/loss': 0.27366673946380615, 'train/grad_norm_pre_clip': 10.834854125976562, 'eval/loss': 0.4354496896266937, 'eval/latency_4/loss': 0.4354496896266937, 'eval/demon_fire_precision': 0.9631629088599556, 'eval/demon_fire_recall': 0.985060084443001, 'eval/demon_fire_f1': 0.9739884393063584, 'eval/demon_fire_rate_model': 0.9248164464023495, 'eval/demon_fire_rate_teacher': 0.9042584434654919, 'eval/demon_left_precision': 0.8359621451104101, 'eval/demon_left_recall': 0.6050228310502284, 'eval/demon_left_f1': 0.7019867549668874, 'eval/demon_left_rate_model': 0.09309838472834067, 'eval/demon_left_rate_teacher': 0.12863436123348018, 'eval/demon_right_precision': 0.7606837606837606, 'eval/demon_right_recall': 0.8032490974729242, 'eval/demon_right_f1': 0.781387181738367, 'eval/demon_right_rate_model': 0.17180616740088106, 'eval/demon_right_rate_teacher': 0.16270190895741557, 'eval/demon_move_f1': 0.7416869683526273, 'eval/demon_cc_f1': 0.8903310890566636, 'eval/latency_4/demon_fire_precision': 0.9631629088599556, 'eval/latency_4/demon_fire_recall': 0.985060084443001, 'eval/latency_4/demon_fire_f1': 0.9739884393063584, 'eval/latency_4/demon_fire_rate_model': 0.9248164464023495, 'eval/latency_4/demon_fire_rate_teacher': 0.9042584434654919, 'eval/latency_4/demon_left_precision': 0.8359621451104101, 'eval/latency_4/demon_left_recall': 0.6050228310502284, 'eval/latency_4/demon_left_f1': 0.7019867549668874, 'eval/latency_4/demon_left_rate_model': 0.09309838472834067, 'eval/latency_4/demon_left_rate_teacher': 0.12863436123348018, 'eval/latency_4/demon_right_precision': 0.7606837606837606, 'eval/latency_4/demon_right_recall': 0.8032490974729242, 'eval/latency_4/demon_right_f1': 0.781387181738367, 'eval/latency_4/demon_right_rate_model': 0.17180616740088106, 'eval/latency_4/demon_right_rate_teacher': 0.16270190895741557, 'eval/latency_4/demon_move_f1': 0.7416869683526273, 'eval/latency_4/demon_cc_f1': 0.8903310890566636, 'timing/data': 0.026643897406756878, 'timing/model': 1.1677451068535447, 'learning_rate/qwen_vl_interface': 2.336982936152347e-06, 'learning_rate/action_model': 2.3369829361523468e-05, 'epoch': 1.92}) +[2026-06-18 04:20:41,682][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_5000_state +[2026-06-18 04:20:41,690][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 04:20:50,720][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_5000_state/model.safetensors +[2026-06-18 04:21:14,031][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_5000_state/optimizer.bin +[2026-06-18 04:21:14,033][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_5000_state/sampler.bin +[2026-06-18 04:21:14,034][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_5000_state/sampler_1.bin +[2026-06-18 04:21:14,041][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_5000_state/random_states_0.pkl +[2026-06-18 04:21:14,056][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 5000 +[2026-06-18 04:21:14,056][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 04:21:14,074][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 04:21:19,132][starVLA.training.train_starvla][INFO] - Step 5001, Loss: {'train/loss': 0.32971110939979553, 'train/grad_norm_pre_clip': 12.994424819946289, 'timing/data': 0.024767639115452766, 'timing/model': 1.1448845257982612, 'learning_rate/qwen_vl_interface': 2.3352749810279047e-06, 'learning_rate/action_model': 2.3352749810279047e-05, 'epoch': 1.92}) +[2026-06-18 04:21:23,926][starVLA.training.train_starvla][INFO] - Step 5002, Loss: {'train/loss': 0.41571053862571716, 'train/grad_norm_pre_clip': 16.196401596069336, 'timing/data': 0.02758403867483139, 'timing/model': 1.6000131256878376, 'learning_rate/qwen_vl_interface': 2.3335676301290972e-06, 'learning_rate/action_model': 2.3335676301290973e-05, 'epoch': 1.92}) +[2026-06-18 04:21:28,335][starVLA.training.train_starvla][INFO] - Step 5003, Loss: {'train/loss': 0.5141813158988953, 'train/grad_norm_pre_clip': 15.22713851928711, 'timing/data': 0.02760419249534607, 'timing/model': 1.1081373998895288, 'learning_rate/qwen_vl_interface': 2.331860883809857e-06, 'learning_rate/action_model': 2.3318608838098565e-05, 'epoch': 1.92}) +[2026-06-18 04:21:33,266][starVLA.training.train_starvla][INFO] - Step 5004, Loss: {'train/loss': 0.6349990963935852, 'train/grad_norm_pre_clip': 16.889240264892578, 'timing/data': 0.03238246403634548, 'timing/model': 1.1348335836082697, 'learning_rate/qwen_vl_interface': 2.330154742423993e-06, 'learning_rate/action_model': 2.330154742423993e-05, 'epoch': 1.93}) +[2026-06-18 04:21:41,593][starVLA.training.train_starvla][INFO] - Step 5005, Loss: {'train/loss': 0.830753743648529, 'train/grad_norm_pre_clip': 20.36025047302246, 'timing/data': 0.027338731102645397, 'timing/model': 1.116507957689464, 'learning_rate/qwen_vl_interface': 2.328449206325195e-06, 'learning_rate/action_model': 2.3284492063251947e-05, 'epoch': 1.93}) +[2026-06-18 04:21:46,501][starVLA.training.train_starvla][INFO] - Step 5006, Loss: {'train/loss': 0.31257081031799316, 'train/grad_norm_pre_clip': 12.509181022644043, 'timing/data': 0.02298660669475794, 'timing/model': 1.1348544778302312, 'learning_rate/qwen_vl_interface': 2.32674427586702e-06, 'learning_rate/action_model': 2.3267442758670198e-05, 'epoch': 1.93}) +[2026-06-18 04:21:56,461][starVLA.training.train_starvla][INFO] - Step 5007, Loss: {'train/loss': 0.2052408754825592, 'train/grad_norm_pre_clip': 11.187129974365234, 'timing/data': 0.028169171884655952, 'timing/model': 1.5166474031284451, 'learning_rate/qwen_vl_interface': 2.3250399514029018e-06, 'learning_rate/action_model': 2.3250399514029016e-05, 'epoch': 1.93}) +[2026-06-18 04:22:00,810][starVLA.training.train_starvla][INFO] - Step 5008, Loss: {'train/loss': 0.3985552191734314, 'train/grad_norm_pre_clip': 13.932414054870605, 'timing/data': 0.02191898413002491, 'timing/model': 1.1398861967027187, 'learning_rate/qwen_vl_interface': 2.32333623328615e-06, 'learning_rate/action_model': 2.3233362332861498e-05, 'epoch': 1.93}) +[2026-06-18 04:22:05,612][starVLA.training.train_starvla][INFO] - Step 5009, Loss: {'train/loss': 0.15723741054534912, 'train/grad_norm_pre_clip': 8.420744895935059, 'timing/data': 0.028545914217829704, 'timing/model': 1.1821869919076562, 'learning_rate/qwen_vl_interface': 2.3216331218699457e-06, 'learning_rate/action_model': 2.3216331218699456e-05, 'epoch': 1.93}) +[2026-06-18 04:22:10,749][starVLA.training.train_starvla][INFO] - Step 5010, Loss: {'train/loss': 0.5116702914237976, 'train/grad_norm_pre_clip': 16.041481018066406, 'timing/data': 0.04043109901249409, 'timing/model': 1.2132073109969497, 'learning_rate/qwen_vl_interface': 2.3199306175073465e-06, 'learning_rate/action_model': 2.3199306175073467e-05, 'epoch': 1.93}) +[2026-06-18 04:22:16,425][starVLA.training.train_starvla][INFO] - Step 5011, Loss: {'train/loss': 0.5584168434143066, 'train/grad_norm_pre_clip': 16.003820419311523, 'timing/data': 0.029642554931342602, 'timing/model': 1.0970802558586001, 'learning_rate/qwen_vl_interface': 2.318228720551283e-06, 'learning_rate/action_model': 2.318228720551283e-05, 'epoch': 1.93}) +[2026-06-18 04:22:23,703][starVLA.training.train_starvla][INFO] - Step 5012, Loss: {'train/loss': 0.2096557915210724, 'train/grad_norm_pre_clip': 13.04498291015625, 'timing/data': 0.02563006803393364, 'timing/model': 1.6082616355270147, 'learning_rate/qwen_vl_interface': 2.3165274313545606e-06, 'learning_rate/action_model': 2.3165274313545606e-05, 'epoch': 1.93}) +[2026-06-18 04:22:28,277][starVLA.training.train_starvla][INFO] - Step 5013, Loss: {'train/loss': 0.38755324482917786, 'train/grad_norm_pre_clip': 21.128787994384766, 'timing/data': 0.02366775367408991, 'timing/model': 1.216349883005023, 'learning_rate/qwen_vl_interface': 2.3148267502698584e-06, 'learning_rate/action_model': 2.3148267502698584e-05, 'epoch': 1.93}) +[2026-06-18 04:22:33,509][starVLA.training.train_starvla][INFO] - Step 5014, Loss: {'train/loss': 0.7165668606758118, 'train/grad_norm_pre_clip': 16.244747161865234, 'timing/data': 0.022983630187809467, 'timing/model': 1.233054312877357, 'learning_rate/qwen_vl_interface': 2.313126677649729e-06, 'learning_rate/action_model': 2.313126677649729e-05, 'epoch': 1.93}) +[2026-06-18 04:22:42,397][starVLA.training.train_starvla][INFO] - Step 5015, Loss: {'train/loss': 0.4126417338848114, 'train/grad_norm_pre_clip': 12.826558113098145, 'timing/data': 0.022093189880251884, 'timing/model': 1.2265045465901494, 'learning_rate/qwen_vl_interface': 2.3114272138465975e-06, 'learning_rate/action_model': 2.3114272138465977e-05, 'epoch': 1.93}) +[2026-06-18 04:22:48,155][starVLA.training.train_starvla][INFO] - Step 5016, Loss: {'train/loss': 0.1649138480424881, 'train/grad_norm_pre_clip': 9.419719696044922, 'timing/data': 0.056384787894785404, 'timing/model': 1.7371279234066606, 'learning_rate/qwen_vl_interface': 2.3097283592127706e-06, 'learning_rate/action_model': 2.3097283592127706e-05, 'epoch': 1.93}) +[2026-06-18 04:23:01,268][starVLA.training.train_starvla][INFO] - Step 5017, Loss: {'train/loss': 0.3394365608692169, 'train/grad_norm_pre_clip': 15.326420783996582, 'timing/data': 0.030131932348012924, 'timing/model': 4.213307872414589, 'learning_rate/qwen_vl_interface': 2.3080301141004165e-06, 'learning_rate/action_model': 2.3080301141004165e-05, 'epoch': 1.93}) +[2026-06-18 04:23:12,280][starVLA.training.train_starvla][INFO] - Step 5018, Loss: {'train/loss': 0.6047849059104919, 'train/grad_norm_pre_clip': 18.089937210083008, 'timing/data': 0.043715289793908596, 'timing/model': 1.1715515768155456, 'learning_rate/qwen_vl_interface': 2.3063324788615843e-06, 'learning_rate/action_model': 2.306332478861584e-05, 'epoch': 1.93}) +[2026-06-18 04:23:18,632][starVLA.training.train_starvla][INFO] - Step 5019, Loss: {'train/loss': 0.5174261927604675, 'train/grad_norm_pre_clip': 18.123708724975586, 'timing/data': 0.02666524238884449, 'timing/model': 2.366213492117822, 'learning_rate/qwen_vl_interface': 2.304635453848198e-06, 'learning_rate/action_model': 2.3046354538481983e-05, 'epoch': 1.93}) +[2026-06-18 04:23:25,041][starVLA.training.train_starvla][INFO] - Step 5020, Loss: {'train/loss': 0.3396275043487549, 'train/grad_norm_pre_clip': 16.891902923583984, 'timing/data': 0.024728299118578434, 'timing/model': 2.7497473508119583, 'learning_rate/qwen_vl_interface': 2.3029390394120517e-06, 'learning_rate/action_model': 2.3029390394120513e-05, 'epoch': 1.93}) +[2026-06-18 04:23:34,314][starVLA.training.train_starvla][INFO] - Step 5021, Loss: {'train/loss': 0.2091471403837204, 'train/grad_norm_pre_clip': 11.531023979187012, 'timing/data': 0.027201424352824688, 'timing/model': 1.1873219422996044, 'learning_rate/qwen_vl_interface': 2.301243235904813e-06, 'learning_rate/action_model': 2.3012432359048133e-05, 'epoch': 1.93}) +[2026-06-18 04:23:43,144][starVLA.training.train_starvla][INFO] - Step 5022, Loss: {'train/loss': 0.1783035695552826, 'train/grad_norm_pre_clip': 37.8189811706543, 'timing/data': 0.028075555339455605, 'timing/model': 1.6548613822087646, 'learning_rate/qwen_vl_interface': 2.299548043678026e-06, 'learning_rate/action_model': 2.299548043678026e-05, 'epoch': 1.93}) +[2026-06-18 04:23:47,877][starVLA.training.train_starvla][INFO] - Step 5023, Loss: {'train/loss': 0.26873260736465454, 'train/grad_norm_pre_clip': 10.361576080322266, 'timing/data': 0.01315892394632101, 'timing/model': 1.1165700051933527, 'learning_rate/qwen_vl_interface': 2.297853463083103e-06, 'learning_rate/action_model': 2.297853463083103e-05, 'epoch': 1.93}) +[2026-06-18 04:23:52,681][starVLA.training.train_starvla][INFO] - Step 5024, Loss: {'train/loss': 0.25847887992858887, 'train/grad_norm_pre_clip': 13.283483505249023, 'timing/data': 0.01945133227854967, 'timing/model': 1.1840998297557235, 'learning_rate/qwen_vl_interface': 2.296159494471335e-06, 'learning_rate/action_model': 2.2961594944713347e-05, 'epoch': 1.93}) +[2026-06-18 04:23:57,547][starVLA.training.train_starvla][INFO] - Step 5025, Loss: {'train/loss': 0.32894235849380493, 'train/grad_norm_pre_clip': 18.123186111450195, 'timing/data': 0.013422718271613121, 'timing/model': 1.1095005879178643, 'learning_rate/qwen_vl_interface': 2.294466138193881e-06, 'learning_rate/action_model': 2.2944661381938812e-05, 'epoch': 1.93}) +[2026-06-18 04:24:02,574][starVLA.training.train_starvla][INFO] - Step 5026, Loss: {'train/loss': 0.2059928923845291, 'train/grad_norm_pre_clip': 12.60843276977539, 'timing/data': 0.015190932899713516, 'timing/model': 1.5975544285029173, 'learning_rate/qwen_vl_interface': 2.2927733946017777e-06, 'learning_rate/action_model': 2.2927733946017778e-05, 'epoch': 1.93}) +[2026-06-18 04:24:07,203][starVLA.training.train_starvla][INFO] - Step 5027, Loss: {'train/loss': 0.2011748105287552, 'train/grad_norm_pre_clip': 9.428611755371094, 'timing/data': 0.014356186613440514, 'timing/model': 1.1619213903322816, 'learning_rate/qwen_vl_interface': 2.2910812640459314e-06, 'learning_rate/action_model': 2.2910812640459315e-05, 'epoch': 1.93}) +[2026-06-18 04:24:12,144][starVLA.training.train_starvla][INFO] - Step 5028, Loss: {'train/loss': 0.30562594532966614, 'train/grad_norm_pre_clip': 16.225065231323242, 'timing/data': 0.013141747564077377, 'timing/model': 1.0937041491270065, 'learning_rate/qwen_vl_interface': 2.289389746877123e-06, 'learning_rate/action_model': 2.2893897468771228e-05, 'epoch': 1.93}) +[2026-06-18 04:24:16,885][starVLA.training.train_starvla][INFO] - Step 5029, Loss: {'train/loss': 0.3659653067588806, 'train/grad_norm_pre_clip': 14.593076705932617, 'timing/data': 0.02050554845482111, 'timing/model': 1.1664931429550052, 'learning_rate/qwen_vl_interface': 2.2876988434460046e-06, 'learning_rate/action_model': 2.2876988434460046e-05, 'epoch': 1.93}) +[2026-06-18 04:24:21,973][starVLA.training.train_starvla][INFO] - Step 5030, Loss: {'train/loss': 0.6486199498176575, 'train/grad_norm_pre_clip': 21.981184005737305, 'timing/data': 0.01316444855183363, 'timing/model': 1.1372426263988018, 'learning_rate/qwen_vl_interface': 2.2860085541031048e-06, 'learning_rate/action_model': 2.2860085541031047e-05, 'epoch': 1.94}) +[2026-06-18 04:24:26,783][starVLA.training.train_starvla][INFO] - Step 5031, Loss: {'train/loss': 0.41784754395484924, 'train/grad_norm_pre_clip': 13.107101440429688, 'timing/data': 0.014110727235674858, 'timing/model': 1.1020548464730382, 'learning_rate/qwen_vl_interface': 2.284318879198823e-06, 'learning_rate/action_model': 2.284318879198823e-05, 'epoch': 1.94}) +[2026-06-18 04:24:31,416][starVLA.training.train_starvla][INFO] - Step 5032, Loss: {'train/loss': 0.23047107458114624, 'train/grad_norm_pre_clip': 10.486722946166992, 'timing/data': 0.012793220579624176, 'timing/model': 1.4750783191993833, 'learning_rate/qwen_vl_interface': 2.2826298190834266e-06, 'learning_rate/action_model': 2.2826298190834266e-05, 'epoch': 1.94}) +[2026-06-18 04:24:35,816][starVLA.training.train_starvla][INFO] - Step 5033, Loss: {'train/loss': 0.2162994146347046, 'train/grad_norm_pre_clip': 13.632655143737793, 'timing/data': 0.012897354550659657, 'timing/model': 1.117720422334969, 'learning_rate/qwen_vl_interface': 2.2809413741070603e-06, 'learning_rate/action_model': 2.28094137410706e-05, 'epoch': 1.94}) +[2026-06-18 04:24:40,872][starVLA.training.train_starvla][INFO] - Step 5034, Loss: {'train/loss': 0.3701792061328888, 'train/grad_norm_pre_clip': 16.99383544921875, 'timing/data': 0.01484986674040556, 'timing/model': 1.2752428548410535, 'learning_rate/qwen_vl_interface': 2.2792535446197435e-06, 'learning_rate/action_model': 2.2792535446197436e-05, 'epoch': 1.94}) +[2026-06-18 04:24:45,729][starVLA.training.train_starvla][INFO] - Step 5035, Loss: {'train/loss': 0.3959991931915283, 'train/grad_norm_pre_clip': 13.126462936401367, 'timing/data': 0.01188337430357933, 'timing/model': 1.0901819178834558, 'learning_rate/qwen_vl_interface': 2.2775663309713643e-06, 'learning_rate/action_model': 2.2775663309713645e-05, 'epoch': 1.94}) +[2026-06-18 04:24:50,808][starVLA.training.train_starvla][INFO] - Step 5036, Loss: {'train/loss': 0.2671682834625244, 'train/grad_norm_pre_clip': 12.954639434814453, 'timing/data': 0.014894147403538227, 'timing/model': 1.1145909167826176, 'learning_rate/qwen_vl_interface': 2.2758797335116843e-06, 'learning_rate/action_model': 2.2758797335116844e-05, 'epoch': 1.94}) +[2026-06-18 04:24:57,049][starVLA.training.train_starvla][INFO] - Step 5037, Loss: {'train/loss': 0.32096168398857117, 'train/grad_norm_pre_clip': 16.608850479125977, 'timing/data': 0.012710369192063808, 'timing/model': 2.959295312874019, 'learning_rate/qwen_vl_interface': 2.2741937525903324e-06, 'learning_rate/action_model': 2.2741937525903324e-05, 'epoch': 1.94}) +[2026-06-18 04:25:07,445][starVLA.training.train_starvla][INFO] - Step 5038, Loss: {'train/loss': 0.44821256399154663, 'train/grad_norm_pre_clip': 16.38059425354004, 'timing/data': 0.012540780007839203, 'timing/model': 1.2547778384760022, 'learning_rate/qwen_vl_interface': 2.2725083885568198e-06, 'learning_rate/action_model': 2.2725083885568197e-05, 'epoch': 1.94}) +[2026-06-18 04:25:18,588][starVLA.training.train_starvla][INFO] - Step 5039, Loss: {'train/loss': 0.42570993304252625, 'train/grad_norm_pre_clip': 15.340533256530762, 'timing/data': 0.02049832046031952, 'timing/model': 1.181227931752801, 'learning_rate/qwen_vl_interface': 2.2708236417605217e-06, 'learning_rate/action_model': 2.2708236417605216e-05, 'epoch': 1.94}) +[2026-06-18 04:25:26,437][starVLA.training.train_starvla][INFO] - Step 5040, Loss: {'train/loss': 0.2806204557418823, 'train/grad_norm_pre_clip': 15.566723823547363, 'timing/data': 0.012698550708591938, 'timing/model': 2.1542452918365598, 'learning_rate/qwen_vl_interface': 2.2691395125506885e-06, 'learning_rate/action_model': 2.2691395125506885e-05, 'epoch': 1.94}) +[2026-06-18 04:25:34,721][starVLA.training.train_starvla][INFO] - Step 5041, Loss: {'train/loss': 0.2923453152179718, 'train/grad_norm_pre_clip': 18.673519134521484, 'timing/data': 0.0113870520144701, 'timing/model': 4.581513269804418, 'learning_rate/qwen_vl_interface': 2.2674560012764426e-06, 'learning_rate/action_model': 2.2674560012764423e-05, 'epoch': 1.94}) +[2026-06-18 04:25:43,190][starVLA.training.train_starvla][INFO] - Step 5042, Loss: {'train/loss': 0.2560192346572876, 'train/grad_norm_pre_clip': 15.73511791229248, 'timing/data': 0.016621113754808903, 'timing/model': 1.6349055608734488, 'learning_rate/qwen_vl_interface': 2.2657731082867764e-06, 'learning_rate/action_model': 2.265773108286776e-05, 'epoch': 1.94}) +[2026-06-18 04:25:49,885][starVLA.training.train_starvla][INFO] - Step 5043, Loss: {'train/loss': 0.39205434918403625, 'train/grad_norm_pre_clip': 24.1357364654541, 'timing/data': 0.01917209941893816, 'timing/model': 1.275158504024148, 'learning_rate/qwen_vl_interface': 2.264090833930555e-06, 'learning_rate/action_model': 2.2640908339305553e-05, 'epoch': 1.94}) +[2026-06-18 04:25:57,245][starVLA.training.train_starvla][INFO] - Step 5044, Loss: {'train/loss': 0.1343144327402115, 'train/grad_norm_pre_clip': 12.425032615661621, 'timing/data': 0.017740086652338505, 'timing/model': 2.813190500251949, 'learning_rate/qwen_vl_interface': 2.2624091785565165e-06, 'learning_rate/action_model': 2.2624091785565167e-05, 'epoch': 1.94}) +[2026-06-18 04:26:02,604][starVLA.training.train_starvla][INFO] - Step 5045, Loss: {'train/loss': 0.5236161947250366, 'train/grad_norm_pre_clip': 22.393146514892578, 'timing/data': 0.012389023788273335, 'timing/model': 1.3243087166920304, 'learning_rate/qwen_vl_interface': 2.2607281425132728e-06, 'learning_rate/action_model': 2.260728142513273e-05, 'epoch': 1.94}) +[2026-06-18 04:26:07,125][starVLA.training.train_starvla][INFO] - Step 5046, Loss: {'train/loss': 0.5205867886543274, 'train/grad_norm_pre_clip': 27.0842342376709, 'timing/data': 0.0117118451744318, 'timing/model': 1.0805713711306453, 'learning_rate/qwen_vl_interface': 2.2590477261493e-06, 'learning_rate/action_model': 2.2590477261493e-05, 'epoch': 1.94}) +[2026-06-18 04:26:11,512][starVLA.training.train_starvla][INFO] - Step 5047, Loss: {'train/loss': 0.24755699932575226, 'train/grad_norm_pre_clip': 12.093231201171875, 'timing/data': 0.010702370665967464, 'timing/model': 1.2902740091085434, 'learning_rate/qwen_vl_interface': 2.257367929812953e-06, 'learning_rate/action_model': 2.2573679298129525e-05, 'epoch': 1.94}) +[2026-06-18 04:26:15,595][starVLA.training.train_starvla][INFO] - Step 5048, Loss: {'train/loss': 0.3846642076969147, 'train/grad_norm_pre_clip': 13.465920448303223, 'timing/data': 0.00993119366466999, 'timing/model': 1.0418173940852284, 'learning_rate/qwen_vl_interface': 2.2556887538524514e-06, 'learning_rate/action_model': 2.2556887538524512e-05, 'epoch': 1.94}) +[2026-06-18 04:26:19,975][starVLA.training.train_starvla][INFO] - Step 5049, Loss: {'train/loss': 0.22166182100772858, 'train/grad_norm_pre_clip': 31.958946228027344, 'timing/data': 0.017139236442744732, 'timing/model': 1.0597571041435003, 'learning_rate/qwen_vl_interface': 2.2540101986158953e-06, 'learning_rate/action_model': 2.2540101986158953e-05, 'epoch': 1.94}) +[2026-06-18 04:26:24,389][starVLA.training.train_starvla][INFO] - Step 5050, Loss: {'train/loss': 0.381107360124588, 'train/grad_norm_pre_clip': 13.328734397888184, 'timing/data': 0.016987106762826443, 'timing/model': 1.0541643472388387, 'learning_rate/qwen_vl_interface': 2.2523322644512506e-06, 'learning_rate/action_model': 2.2523322644512504e-05, 'epoch': 1.94}) +[2026-06-18 04:26:28,770][starVLA.training.train_starvla][INFO] - Step 5051, Loss: {'train/loss': 0.17790073156356812, 'train/grad_norm_pre_clip': 8.914763450622559, 'timing/data': 0.01905869133770466, 'timing/model': 1.0470010414719582, 'learning_rate/qwen_vl_interface': 2.2506549517063497e-06, 'learning_rate/action_model': 2.2506549517063496e-05, 'epoch': 1.94}) +[2026-06-18 04:26:33,117][starVLA.training.train_starvla][INFO] - Step 5052, Loss: {'train/loss': 0.33538466691970825, 'train/grad_norm_pre_clip': 16.306018829345703, 'timing/data': 0.010805673897266388, 'timing/model': 1.3044774886220694, 'learning_rate/qwen_vl_interface': 2.248978260728907e-06, 'learning_rate/action_model': 2.2489782607289067e-05, 'epoch': 1.94}) +[2026-06-18 04:26:37,274][starVLA.training.train_starvla][INFO] - Step 5053, Loss: {'train/loss': 0.14282828569412231, 'train/grad_norm_pre_clip': 7.667361736297607, 'timing/data': 0.01502637006342411, 'timing/model': 1.061270578764379, 'learning_rate/qwen_vl_interface': 2.2473021918664987e-06, 'learning_rate/action_model': 2.247302191866499e-05, 'epoch': 1.94}) +[2026-06-18 04:26:41,679][starVLA.training.train_starvla][INFO] - Step 5054, Loss: {'train/loss': 0.48188450932502747, 'train/grad_norm_pre_clip': 15.402400970458984, 'timing/data': 0.009853817522525787, 'timing/model': 1.0551730617880821, 'learning_rate/qwen_vl_interface': 2.245626745466578e-06, 'learning_rate/action_model': 2.2456267454665782e-05, 'epoch': 1.94}) +[2026-06-18 04:26:45,938][starVLA.training.train_starvla][INFO] - Step 5055, Loss: {'train/loss': 0.36524292826652527, 'train/grad_norm_pre_clip': 10.192618370056152, 'timing/data': 0.010085535235702991, 'timing/model': 1.0382483331486583, 'learning_rate/qwen_vl_interface': 2.2439519218764657e-06, 'learning_rate/action_model': 2.2439519218764656e-05, 'epoch': 1.94}) +[2026-06-18 04:26:50,252][starVLA.training.train_starvla][INFO] - Step 5056, Loss: {'train/loss': 0.2715936303138733, 'train/grad_norm_pre_clip': 14.464216232299805, 'timing/data': 0.010156165808439255, 'timing/model': 1.0650716880336404, 'learning_rate/qwen_vl_interface': 2.242277721443355e-06, 'learning_rate/action_model': 2.2422777214433545e-05, 'epoch': 1.95}) +[2026-06-18 04:26:55,681][starVLA.training.train_starvla][INFO] - Step 5057, Loss: {'train/loss': 0.29309242963790894, 'train/grad_norm_pre_clip': 13.215987205505371, 'timing/data': 0.010153732262551785, 'timing/model': 1.2875228188931942, 'learning_rate/qwen_vl_interface': 2.2406041445143086e-06, 'learning_rate/action_model': 2.2406041445143086e-05, 'epoch': 1.95}) +[2026-06-18 04:26:59,870][starVLA.training.train_starvla][INFO] - Step 5058, Loss: {'train/loss': 0.18289826810359955, 'train/grad_norm_pre_clip': 11.914356231689453, 'timing/data': 0.011164040304720402, 'timing/model': 1.100130371749401, 'learning_rate/qwen_vl_interface': 2.2389311914362614e-06, 'learning_rate/action_model': 2.2389311914362616e-05, 'epoch': 1.95}) +[2026-06-18 04:27:04,379][starVLA.training.train_starvla][INFO] - Step 5059, Loss: {'train/loss': 0.4472554326057434, 'train/grad_norm_pre_clip': 15.799919128417969, 'timing/data': 0.01302199624478817, 'timing/model': 1.1035035606473684, 'learning_rate/qwen_vl_interface': 2.2372588625560185e-06, 'learning_rate/action_model': 2.2372588625560185e-05, 'epoch': 1.95}) +[2026-06-18 04:27:09,179][starVLA.training.train_starvla][INFO] - Step 5060, Loss: {'train/loss': 0.23851315677165985, 'train/grad_norm_pre_clip': 14.40883731842041, 'timing/data': 0.011961708776652813, 'timing/model': 1.156834869645536, 'learning_rate/qwen_vl_interface': 2.2355871582202544e-06, 'learning_rate/action_model': 2.2355871582202544e-05, 'epoch': 1.95}) +[2026-06-18 04:27:15,647][starVLA.training.train_starvla][INFO] - Step 5061, Loss: {'train/loss': 0.7084704041481018, 'train/grad_norm_pre_clip': 20.01987075805664, 'timing/data': 0.020245317369699478, 'timing/model': 1.1837785746902227, 'learning_rate/qwen_vl_interface': 2.2339160787755175e-06, 'learning_rate/action_model': 2.2339160787755174e-05, 'epoch': 1.95}) +[2026-06-18 04:27:20,694][starVLA.training.train_starvla][INFO] - Step 5062, Loss: {'train/loss': 0.3864983022212982, 'train/grad_norm_pre_clip': 16.727304458618164, 'timing/data': 0.013188754208385944, 'timing/model': 1.593481384217739, 'learning_rate/qwen_vl_interface': 2.23224562456822e-06, 'learning_rate/action_model': 2.23224562456822e-05, 'epoch': 1.95}) +[2026-06-18 04:27:25,167][starVLA.training.train_starvla][INFO] - Step 5063, Loss: {'train/loss': 0.5140072107315063, 'train/grad_norm_pre_clip': 17.06319236755371, 'timing/data': 0.01167299784719944, 'timing/model': 1.1870858445763588, 'learning_rate/qwen_vl_interface': 2.2305757959446538e-06, 'learning_rate/action_model': 2.2305757959446537e-05, 'epoch': 1.95}) +[2026-06-18 04:27:30,019][starVLA.training.train_starvla][INFO] - Step 5064, Loss: {'train/loss': 0.17394375801086426, 'train/grad_norm_pre_clip': 10.476480484008789, 'timing/data': 0.021231348626315594, 'timing/model': 1.0941683538258076, 'learning_rate/qwen_vl_interface': 2.228906593250976e-06, 'learning_rate/action_model': 2.228906593250976e-05, 'epoch': 1.95}) +[2026-06-18 04:27:41,021][starVLA.training.train_starvla][INFO] - Step 5065, Loss: {'train/loss': 0.29260125756263733, 'train/grad_norm_pre_clip': 14.559625625610352, 'timing/data': 0.012211647816002369, 'timing/model': 4.1920241909101605, 'learning_rate/qwen_vl_interface': 2.2272380168332105e-06, 'learning_rate/action_model': 2.2272380168332103e-05, 'epoch': 1.95}) +[2026-06-18 04:27:47,397][starVLA.training.train_starvla][INFO] - Step 5066, Loss: {'train/loss': 0.35264062881469727, 'train/grad_norm_pre_clip': 13.71678638458252, 'timing/data': 0.022773195058107376, 'timing/model': 1.1171552063897252, 'learning_rate/qwen_vl_interface': 2.225570067037255e-06, 'learning_rate/action_model': 2.2255700670372548e-05, 'epoch': 1.95}) +[2026-06-18 04:27:52,429][starVLA.training.train_starvla][INFO] - Step 5067, Loss: {'train/loss': 0.3716377913951874, 'train/grad_norm_pre_clip': 16.328872680664062, 'timing/data': 0.012589240446686745, 'timing/model': 1.632743127644062, 'learning_rate/qwen_vl_interface': 2.223902744208881e-06, 'learning_rate/action_model': 2.2239027442088808e-05, 'epoch': 1.95}) +[2026-06-18 04:27:57,211][starVLA.training.train_starvla][INFO] - Step 5068, Loss: {'train/loss': 0.14227117598056793, 'train/grad_norm_pre_clip': 8.759904861450195, 'timing/data': 0.012367494404315948, 'timing/model': 1.083827206864953, 'learning_rate/qwen_vl_interface': 2.2222360486937238e-06, 'learning_rate/action_model': 2.2222360486937235e-05, 'epoch': 1.95}) +[2026-06-18 04:28:02,149][starVLA.training.train_starvla][INFO] - Step 5069, Loss: {'train/loss': 0.1499553769826889, 'train/grad_norm_pre_clip': 6.740219593048096, 'timing/data': 0.014060065150260925, 'timing/model': 1.1749598728492856, 'learning_rate/qwen_vl_interface': 2.220569980837292e-06, 'learning_rate/action_model': 2.220569980837292e-05, 'epoch': 1.95}) +[2026-06-18 04:28:10,380][starVLA.training.train_starvla][INFO] - Step 5070, Loss: {'train/loss': 0.4377938508987427, 'train/grad_norm_pre_clip': 15.01053524017334, 'timing/data': 0.01379508338868618, 'timing/model': 1.4248027754947543, 'learning_rate/qwen_vl_interface': 2.2189045409849625e-06, 'learning_rate/action_model': 2.2189045409849628e-05, 'epoch': 1.95}) +[2026-06-18 04:28:21,464][starVLA.training.train_starvla][INFO] - Step 5071, Loss: {'train/loss': 0.532379150390625, 'train/grad_norm_pre_clip': 13.853353500366211, 'timing/data': 0.014752649702131748, 'timing/model': 5.41508199647069, 'learning_rate/qwen_vl_interface': 2.2172397294819843e-06, 'learning_rate/action_model': 2.2172397294819843e-05, 'epoch': 1.95}) +[2026-06-18 04:28:26,307][starVLA.training.train_starvla][INFO] - Step 5072, Loss: {'train/loss': 0.41635721921920776, 'train/grad_norm_pre_clip': 14.234105110168457, 'timing/data': 0.030291778966784477, 'timing/model': 1.5010718889534473, 'learning_rate/qwen_vl_interface': 2.215575546673473e-06, 'learning_rate/action_model': 2.215575546673473e-05, 'epoch': 1.95}) +[2026-06-18 04:28:30,782][starVLA.training.train_starvla][INFO] - Step 5073, Loss: {'train/loss': 0.19235552847385406, 'train/grad_norm_pre_clip': 12.421648025512695, 'timing/data': 0.03685665503144264, 'timing/model': 1.1457295482978225, 'learning_rate/qwen_vl_interface': 2.213911992904416e-06, 'learning_rate/action_model': 2.213911992904416e-05, 'epoch': 1.95}) +[2026-06-18 04:28:35,579][starVLA.training.train_starvla][INFO] - Step 5074, Loss: {'train/loss': 0.25544077157974243, 'train/grad_norm_pre_clip': 13.054107666015625, 'timing/data': 0.03741551376879215, 'timing/model': 1.1819514958187938, 'learning_rate/qwen_vl_interface': 2.2122490685196703e-06, 'learning_rate/action_model': 2.2122490685196704e-05, 'epoch': 1.95}) +[2026-06-18 04:28:40,854][starVLA.training.train_starvla][INFO] - Step 5075, Loss: {'train/loss': 0.34546756744384766, 'train/grad_norm_pre_clip': 17.953664779663086, 'timing/data': 0.032916377298533916, 'timing/model': 1.1842407630756497, 'learning_rate/qwen_vl_interface': 2.210586773863961e-06, 'learning_rate/action_model': 2.210586773863961e-05, 'epoch': 1.95}) +[2026-06-18 04:28:49,151][starVLA.training.train_starvla][INFO] - Step 5076, Loss: {'train/loss': 0.3760412633419037, 'train/grad_norm_pre_clip': 16.082294464111328, 'timing/data': 0.041429816745221615, 'timing/model': 1.6617923341691494, 'learning_rate/qwen_vl_interface': 2.208925109281884e-06, 'learning_rate/action_model': 2.208925109281884e-05, 'epoch': 1.95}) +[2026-06-18 04:28:54,257][starVLA.training.train_starvla][INFO] - Step 5077, Loss: {'train/loss': 0.5010309219360352, 'train/grad_norm_pre_clip': 16.28363609313965, 'timing/data': 0.04278183542191982, 'timing/model': 1.4651760226115584, 'learning_rate/qwen_vl_interface': 2.2072640751179026e-06, 'learning_rate/action_model': 2.2072640751179025e-05, 'epoch': 1.95}) +[2026-06-18 04:28:58,831][starVLA.training.train_starvla][INFO] - Step 5078, Loss: {'train/loss': 0.5590088963508606, 'train/grad_norm_pre_clip': 16.775283813476562, 'timing/data': 0.034115503542125225, 'timing/model': 1.225810238160193, 'learning_rate/qwen_vl_interface': 2.205603671716356e-06, 'learning_rate/action_model': 2.2056036717163557e-05, 'epoch': 1.95}) +[2026-06-18 04:29:03,727][starVLA.training.train_starvla][INFO] - Step 5079, Loss: {'train/loss': 0.32588648796081543, 'train/grad_norm_pre_clip': 18.58863639831543, 'timing/data': 0.030988403595983982, 'timing/model': 1.354668221436441, 'learning_rate/qwen_vl_interface': 2.2039438994214424e-06, 'learning_rate/action_model': 2.2039438994214424e-05, 'epoch': 1.95}) +[2026-06-18 04:29:08,605][starVLA.training.train_starvla][INFO] - Step 5080, Loss: {'train/loss': 0.1603027880191803, 'train/grad_norm_pre_clip': 9.933572769165039, 'timing/data': 0.035001976415514946, 'timing/model': 1.1281007397919893, 'learning_rate/qwen_vl_interface': 2.202284758577235e-06, 'learning_rate/action_model': 2.2022847585772347e-05, 'epoch': 1.95}) +[2026-06-18 04:29:13,377][starVLA.training.train_starvla][INFO] - Step 5081, Loss: {'train/loss': 0.2067449390888214, 'train/grad_norm_pre_clip': 11.22665786743164, 'timing/data': 0.03640840947628021, 'timing/model': 1.1326562659814954, 'learning_rate/qwen_vl_interface': 2.2006262495276774e-06, 'learning_rate/action_model': 2.2006262495276775e-05, 'epoch': 1.95}) +[2026-06-18 04:29:18,367][starVLA.training.train_starvla][INFO] - Step 5082, Loss: {'train/loss': 0.11461856961250305, 'train/grad_norm_pre_clip': 8.179736137390137, 'timing/data': 0.04292997904121876, 'timing/model': 1.436744506470859, 'learning_rate/qwen_vl_interface': 2.1989683726165795e-06, 'learning_rate/action_model': 2.1989683726165795e-05, 'epoch': 1.96}) +[2026-06-18 04:29:23,040][starVLA.training.train_starvla][INFO] - Step 5083, Loss: {'train/loss': 0.3702489137649536, 'train/grad_norm_pre_clip': 17.269874572753906, 'timing/data': 0.04490641783922911, 'timing/model': 1.0954826707020402, 'learning_rate/qwen_vl_interface': 2.197311128187623e-06, 'learning_rate/action_model': 2.1973111281876232e-05, 'epoch': 1.96}) +[2026-06-18 04:29:27,844][starVLA.training.train_starvla][INFO] - Step 5084, Loss: {'train/loss': 0.5252525210380554, 'train/grad_norm_pre_clip': 19.23354148864746, 'timing/data': 0.05008684005588293, 'timing/model': 1.1091363411396742, 'learning_rate/qwen_vl_interface': 2.1956545165843517e-06, 'learning_rate/action_model': 2.1956545165843517e-05, 'epoch': 1.96}) +[2026-06-18 04:29:32,700][starVLA.training.train_starvla][INFO] - Step 5085, Loss: {'train/loss': 0.5941529870033264, 'train/grad_norm_pre_clip': 19.12237548828125, 'timing/data': 0.0347151979804039, 'timing/model': 1.0863357549533248, 'learning_rate/qwen_vl_interface': 2.1939985381501866e-06, 'learning_rate/action_model': 2.1939985381501866e-05, 'epoch': 1.96}) +[2026-06-18 04:29:37,529][starVLA.training.train_starvla][INFO] - Step 5086, Loss: {'train/loss': 0.22405235469341278, 'train/grad_norm_pre_clip': 13.262285232543945, 'timing/data': 0.03548051696270704, 'timing/model': 1.100351714529097, 'learning_rate/qwen_vl_interface': 2.1923431932284135e-06, 'learning_rate/action_model': 2.1923431932284136e-05, 'epoch': 1.96}) +[2026-06-18 04:29:42,472][starVLA.training.train_starvla][INFO] - Step 5087, Loss: {'train/loss': 0.5712484121322632, 'train/grad_norm_pre_clip': 20.111337661743164, 'timing/data': 0.035979511216282845, 'timing/model': 1.639591838233173, 'learning_rate/qwen_vl_interface': 2.190688482162187e-06, 'learning_rate/action_model': 2.190688482162187e-05, 'epoch': 1.96}) +[2026-06-18 04:29:47,012][starVLA.training.train_starvla][INFO] - Step 5088, Loss: {'train/loss': 0.22353726625442505, 'train/grad_norm_pre_clip': 10.928452491760254, 'timing/data': 0.03912691492587328, 'timing/model': 1.0869167922064662, 'learning_rate/qwen_vl_interface': 2.189034405294529e-06, 'learning_rate/action_model': 2.189034405294529e-05, 'epoch': 1.96}) +[2026-06-18 04:29:52,815][starVLA.training.train_starvla][INFO] - Step 5089, Loss: {'train/loss': 0.12093943357467651, 'train/grad_norm_pre_clip': 7.934835910797119, 'timing/data': 0.045350885950028896, 'timing/model': 2.2016682364046574, 'learning_rate/qwen_vl_interface': 2.1873809629683333e-06, 'learning_rate/action_model': 2.1873809629683333e-05, 'epoch': 1.96}) +[2026-06-18 04:29:57,784][starVLA.training.train_starvla][INFO] - Step 5090, Loss: {'train/loss': 0.39129048585891724, 'train/grad_norm_pre_clip': 12.981904029846191, 'timing/data': 0.03884001914411783, 'timing/model': 1.1625207951292396, 'learning_rate/qwen_vl_interface': 2.185728155526359e-06, 'learning_rate/action_model': 2.1857281555263588e-05, 'epoch': 1.96}) +[2026-06-18 04:30:02,646][starVLA.training.train_starvla][INFO] - Step 5091, Loss: {'train/loss': 0.22768402099609375, 'train/grad_norm_pre_clip': 12.068442344665527, 'timing/data': 0.03212364111095667, 'timing/model': 1.2110371431335807, 'learning_rate/qwen_vl_interface': 2.184075983311234e-06, 'learning_rate/action_model': 2.184075983311234e-05, 'epoch': 1.96}) +[2026-06-18 04:30:07,410][starVLA.training.train_starvla][INFO] - Step 5092, Loss: {'train/loss': 0.31750941276550293, 'train/grad_norm_pre_clip': 14.203191757202148, 'timing/data': 0.0316533213481307, 'timing/model': 1.494819387793541, 'learning_rate/qwen_vl_interface': 2.18242444666546e-06, 'learning_rate/action_model': 2.18242444666546e-05, 'epoch': 1.96}) +[2026-06-18 04:30:15,391][starVLA.training.train_starvla][INFO] - Step 5093, Loss: {'train/loss': 0.6686203479766846, 'train/grad_norm_pre_clip': 16.0635929107666, 'timing/data': 0.0306082209572196, 'timing/model': 4.4995932802557945, 'learning_rate/qwen_vl_interface': 2.1807735459313974e-06, 'learning_rate/action_model': 2.180773545931397e-05, 'epoch': 1.96}) +[2026-06-18 04:30:20,270][starVLA.training.train_starvla][INFO] - Step 5094, Loss: {'train/loss': 0.11682569235563278, 'train/grad_norm_pre_clip': 8.544339179992676, 'timing/data': 0.04177578538656235, 'timing/model': 1.1155664855614305, 'learning_rate/qwen_vl_interface': 2.1791232814512815e-06, 'learning_rate/action_model': 2.1791232814512815e-05, 'epoch': 1.96}) +[2026-06-18 04:30:25,256][starVLA.training.train_starvla][INFO] - Step 5095, Loss: {'train/loss': 0.641814649105072, 'train/grad_norm_pre_clip': 18.074552536010742, 'timing/data': 0.047063883394002914, 'timing/model': 1.2896834257990122, 'learning_rate/qwen_vl_interface': 2.1774736535672117e-06, 'learning_rate/action_model': 2.1774736535672115e-05, 'epoch': 1.96}) +[2026-06-18 04:30:31,726][starVLA.training.train_starvla][INFO] - Step 5096, Loss: {'train/loss': 0.4838210642337799, 'train/grad_norm_pre_clip': 13.24183464050293, 'timing/data': 0.03420105669647455, 'timing/model': 2.7262468757107854, 'learning_rate/qwen_vl_interface': 2.1758246626211617e-06, 'learning_rate/action_model': 2.1758246626211617e-05, 'epoch': 1.96}) +[2026-06-18 04:30:36,595][starVLA.training.train_starvla][INFO] - Step 5097, Loss: {'train/loss': 0.4570409655570984, 'train/grad_norm_pre_clip': 17.40715789794922, 'timing/data': 0.03266957029700279, 'timing/model': 1.46848963201046, 'learning_rate/qwen_vl_interface': 2.174176308954969e-06, 'learning_rate/action_model': 2.174176308954969e-05, 'epoch': 1.96}) +[2026-06-18 04:30:41,208][starVLA.training.train_starvla][INFO] - Step 5098, Loss: {'train/loss': 0.2342536300420761, 'train/grad_norm_pre_clip': 15.739969253540039, 'timing/data': 0.03486595302820206, 'timing/model': 1.2197882989421487, 'learning_rate/qwen_vl_interface': 2.172528592910335e-06, 'learning_rate/action_model': 2.1725285929103348e-05, 'epoch': 1.96}) +[2026-06-18 04:30:46,036][starVLA.training.train_starvla][INFO] - Step 5099, Loss: {'train/loss': 0.5768597722053528, 'train/grad_norm_pre_clip': 17.8890438079834, 'timing/data': 0.0435700723901391, 'timing/model': 1.1109593380242586, 'learning_rate/qwen_vl_interface': 2.170881514828834e-06, 'learning_rate/action_model': 2.170881514828834e-05, 'epoch': 1.96}) +[2026-06-18 04:34:05,843][starVLA.training.train_starvla][INFO] - Step 5100, Loss: {'train/loss': 0.36581364274024963, 'train/grad_norm_pre_clip': 17.236934661865234, 'eval/loss': 0.43210047483444214, 'eval/latency_4/loss': 0.43210047483444214, 'eval/demon_fire_precision': 0.9659018483110261, 'eval/demon_fire_recall': 0.9844105228970444, 'eval/demon_fire_f1': 0.9750683609457937, 'eval/demon_fire_rate_model': 0.9215859030837005, 'eval/demon_fire_rate_teacher': 0.9042584434654919, 'eval/demon_left_precision': 0.7988980716253443, 'eval/demon_left_recall': 0.6621004566210046, 'eval/demon_left_f1': 0.7240948813982522, 'eval/demon_left_rate_model': 0.1066079295154185, 'eval/demon_left_rate_teacher': 0.12863436123348018, 'eval/demon_right_precision': 0.8653421633554084, 'eval/demon_right_recall': 0.7075812274368231, 'eval/demon_right_f1': 0.7785501489572989, 'eval/demon_right_rate_model': 0.13303964757709252, 'eval/demon_right_rate_teacher': 0.16270190895741557, 'eval/demon_move_f1': 0.7513225151777756, 'eval/demon_cc_f1': 0.8950994600944928, 'eval/latency_4/demon_fire_precision': 0.9659018483110261, 'eval/latency_4/demon_fire_recall': 0.9844105228970444, 'eval/latency_4/demon_fire_f1': 0.9750683609457937, 'eval/latency_4/demon_fire_rate_model': 0.9215859030837005, 'eval/latency_4/demon_fire_rate_teacher': 0.9042584434654919, 'eval/latency_4/demon_left_precision': 0.7988980716253443, 'eval/latency_4/demon_left_recall': 0.6621004566210046, 'eval/latency_4/demon_left_f1': 0.7240948813982522, 'eval/latency_4/demon_left_rate_model': 0.1066079295154185, 'eval/latency_4/demon_left_rate_teacher': 0.12863436123348018, 'eval/latency_4/demon_right_precision': 0.8653421633554084, 'eval/latency_4/demon_right_recall': 0.7075812274368231, 'eval/latency_4/demon_right_f1': 0.7785501489572989, 'eval/latency_4/demon_right_rate_model': 0.13303964757709252, 'eval/latency_4/demon_right_rate_teacher': 0.16270190895741557, 'eval/latency_4/demon_move_f1': 0.7513225151777756, 'eval/latency_4/demon_cc_f1': 0.8950994600944928, 'timing/data': 0.031035318970680237, 'timing/model': 1.238742216490209, 'learning_rate/qwen_vl_interface': 2.1692350750519093e-06, 'learning_rate/action_model': 2.1692350750519092e-05, 'epoch': 1.96}) +[2026-06-18 04:34:05,868][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_5100_state +[2026-06-18 04:34:05,876][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 04:34:15,621][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_5100_state/model.safetensors +[2026-06-18 04:34:39,218][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_5100_state/optimizer.bin +[2026-06-18 04:34:39,220][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_5100_state/sampler.bin +[2026-06-18 04:34:39,221][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_5100_state/sampler_1.bin +[2026-06-18 04:34:39,225][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_5100_state/random_states_0.pkl +[2026-06-18 04:34:39,232][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 5100 +[2026-06-18 04:34:39,232][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 04:34:39,257][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 04:34:44,176][starVLA.training.train_starvla][INFO] - Step 5101, Loss: {'train/loss': 0.5030393600463867, 'train/grad_norm_pre_clip': 15.0813570022583, 'timing/data': 0.03866610210388899, 'timing/model': 1.104714690707624, 'learning_rate/qwen_vl_interface': 2.167589273920869e-06, 'learning_rate/action_model': 2.167589273920869e-05, 'epoch': 1.96}) +[2026-06-18 04:34:49,097][starVLA.training.train_starvla][INFO] - Step 5102, Loss: {'train/loss': 0.7476766705513, 'train/grad_norm_pre_clip': 21.585851669311523, 'timing/data': 0.05196973402053118, 'timing/model': 1.6325871832668781, 'learning_rate/qwen_vl_interface': 2.1659441117768888e-06, 'learning_rate/action_model': 2.1659441117768886e-05, 'epoch': 1.96}) +[2026-06-18 04:34:53,738][starVLA.training.train_starvla][INFO] - Step 5103, Loss: {'train/loss': 0.4632309377193451, 'train/grad_norm_pre_clip': 16.50477409362793, 'timing/data': 0.03771649394184351, 'timing/model': 1.1122993044555187, 'learning_rate/qwen_vl_interface': 2.164299588961013e-06, 'learning_rate/action_model': 2.1642995889610128e-05, 'epoch': 1.96}) +[2026-06-18 04:34:59,138][starVLA.training.train_starvla][INFO] - Step 5104, Loss: {'train/loss': 0.36182132363319397, 'train/grad_norm_pre_clip': 14.993484497070312, 'timing/data': 0.03391561657190323, 'timing/model': 1.2360430769622326, 'learning_rate/qwen_vl_interface': 2.162655705814153e-06, 'learning_rate/action_model': 2.1626557058141527e-05, 'epoch': 1.96}) +[2026-06-18 04:35:03,912][starVLA.training.train_starvla][INFO] - Step 5105, Loss: {'train/loss': 0.5688759088516235, 'train/grad_norm_pre_clip': 16.68701171875, 'timing/data': 0.03759437706321478, 'timing/model': 1.0958819519728422, 'learning_rate/qwen_vl_interface': 2.161012462677087e-06, 'learning_rate/action_model': 2.161012462677087e-05, 'epoch': 1.96}) +[2026-06-18 04:35:08,979][starVLA.training.train_starvla][INFO] - Step 5106, Loss: {'train/loss': 0.46073779463768005, 'train/grad_norm_pre_clip': 15.039053916931152, 'timing/data': 0.05013416800647974, 'timing/model': 1.5955529101192951, 'learning_rate/qwen_vl_interface': 2.1593698598904612e-06, 'learning_rate/action_model': 2.1593698598904614e-05, 'epoch': 1.96}) +[2026-06-18 04:35:14,747][starVLA.training.train_starvla][INFO] - Step 5107, Loss: {'train/loss': 0.17751403152942657, 'train/grad_norm_pre_clip': 8.2886381149292, 'timing/data': 0.031335885636508465, 'timing/model': 2.248185153119266, 'learning_rate/qwen_vl_interface': 2.1577278977947896e-06, 'learning_rate/action_model': 2.1577278977947897e-05, 'epoch': 1.96}) +[2026-06-18 04:35:19,365][starVLA.training.train_starvla][INFO] - Step 5108, Loss: {'train/loss': 0.3122197389602661, 'train/grad_norm_pre_clip': 13.446345329284668, 'timing/data': 0.03296434972435236, 'timing/model': 1.113898110575974, 'learning_rate/qwen_vl_interface': 2.1560865767304524e-06, 'learning_rate/action_model': 2.156086576730452e-05, 'epoch': 1.97}) +[2026-06-18 04:35:24,484][starVLA.training.train_starvla][INFO] - Step 5109, Loss: {'train/loss': 0.3527539074420929, 'train/grad_norm_pre_clip': 21.635866165161133, 'timing/data': 0.04379571136087179, 'timing/model': 1.2466332390904427, 'learning_rate/qwen_vl_interface': 2.154445897037696e-06, 'learning_rate/action_model': 2.154445897037696e-05, 'epoch': 1.97}) +[2026-06-18 04:35:30,402][starVLA.training.train_starvla][INFO] - Step 5110, Loss: {'train/loss': 0.18745306134223938, 'train/grad_norm_pre_clip': 9.437291145324707, 'timing/data': 0.04165850020945072, 'timing/model': 1.352832956239581, 'learning_rate/qwen_vl_interface': 2.1528058590566353e-06, 'learning_rate/action_model': 2.1528058590566353e-05, 'epoch': 1.97}) +[2026-06-18 04:35:37,643][starVLA.training.train_starvla][INFO] - Step 5111, Loss: {'train/loss': 0.3971116840839386, 'train/grad_norm_pre_clip': 20.07176399230957, 'timing/data': 0.03923597373068333, 'timing/model': 2.7106496868655086, 'learning_rate/qwen_vl_interface': 2.1511664631272543e-06, 'learning_rate/action_model': 2.1511664631272545e-05, 'epoch': 1.97}) +[2026-06-18 04:35:42,633][starVLA.training.train_starvla][INFO] - Step 5112, Loss: {'train/loss': 0.540244460105896, 'train/grad_norm_pre_clip': 22.599853515625, 'timing/data': 0.04614158160984516, 'timing/model': 1.5782594149932265, 'learning_rate/qwen_vl_interface': 2.1495277095894015e-06, 'learning_rate/action_model': 2.1495277095894014e-05, 'epoch': 1.97}) +[2026-06-18 04:35:51,299][starVLA.training.train_starvla][INFO] - Step 5113, Loss: {'train/loss': 0.313353568315506, 'train/grad_norm_pre_clip': 13.735239028930664, 'timing/data': 0.037976911291480064, 'timing/model': 1.244966003112495, 'learning_rate/qwen_vl_interface': 2.147889598782787e-06, 'learning_rate/action_model': 2.147889598782787e-05, 'epoch': 1.97}) +[2026-06-18 04:36:01,350][starVLA.training.train_starvla][INFO] - Step 5114, Loss: {'train/loss': 0.4273357391357422, 'train/grad_norm_pre_clip': 19.004741668701172, 'timing/data': 0.03221006412059069, 'timing/model': 1.545878367498517, 'learning_rate/qwen_vl_interface': 2.1462521310469987e-06, 'learning_rate/action_model': 2.1462521310469987e-05, 'epoch': 1.97}) +[2026-06-18 04:36:07,874][starVLA.training.train_starvla][INFO] - Step 5115, Loss: {'train/loss': 0.2065361887216568, 'train/grad_norm_pre_clip': 9.199711799621582, 'timing/data': 0.03759701270610094, 'timing/model': 1.1435261331498623, 'learning_rate/qwen_vl_interface': 2.144615306721484e-06, 'learning_rate/action_model': 2.144615306721484e-05, 'epoch': 1.97}) +[2026-06-18 04:36:12,892][starVLA.training.train_starvla][INFO] - Step 5116, Loss: {'train/loss': 0.19855482876300812, 'train/grad_norm_pre_clip': 12.19725513458252, 'timing/data': 0.03515577409416437, 'timing/model': 1.3294795602560043, 'learning_rate/qwen_vl_interface': 2.1429791261455575e-06, 'learning_rate/action_model': 2.1429791261455572e-05, 'epoch': 1.97}) +[2026-06-18 04:36:17,646][starVLA.training.train_starvla][INFO] - Step 5117, Loss: {'train/loss': 0.37491562962532043, 'train/grad_norm_pre_clip': 17.667877197265625, 'timing/data': 0.030664293095469475, 'timing/model': 1.543304182589054, 'learning_rate/qwen_vl_interface': 2.1413435896584006e-06, 'learning_rate/action_model': 2.1413435896584005e-05, 'epoch': 1.97}) +[2026-06-18 04:36:22,010][starVLA.training.train_starvla][INFO] - Step 5118, Loss: {'train/loss': 0.2564547657966614, 'train/grad_norm_pre_clip': 14.745291709899902, 'timing/data': 0.02295180968940258, 'timing/model': 1.127209116704762, 'learning_rate/qwen_vl_interface': 2.1397086975990633e-06, 'learning_rate/action_model': 2.139708697599063e-05, 'epoch': 1.97}) +[2026-06-18 04:36:26,786][starVLA.training.train_starvla][INFO] - Step 5119, Loss: {'train/loss': 0.4774032235145569, 'train/grad_norm_pre_clip': 18.602108001708984, 'timing/data': 0.02639668434858322, 'timing/model': 1.1260800007730722, 'learning_rate/qwen_vl_interface': 2.1380744503064597e-06, 'learning_rate/action_model': 2.1380744503064598e-05, 'epoch': 1.97}) +[2026-06-18 04:36:31,842][starVLA.training.train_starvla][INFO] - Step 5120, Loss: {'train/loss': 0.46649107336997986, 'train/grad_norm_pre_clip': 17.91246795654297, 'timing/data': 0.022474066354334354, 'timing/model': 1.098320223391056, 'learning_rate/qwen_vl_interface': 2.1364408481193714e-06, 'learning_rate/action_model': 2.1364408481193712e-05, 'epoch': 1.97}) +[2026-06-18 04:36:36,864][starVLA.training.train_starvla][INFO] - Step 5121, Loss: {'train/loss': 0.37259238958358765, 'train/grad_norm_pre_clip': 27.621084213256836, 'timing/data': 0.025951649993658066, 'timing/model': 1.076840117573738, 'learning_rate/qwen_vl_interface': 2.1348078913764454e-06, 'learning_rate/action_model': 2.1348078913764452e-05, 'epoch': 1.97}) +[2026-06-18 04:36:41,895][starVLA.training.train_starvla][INFO] - Step 5122, Loss: {'train/loss': 0.22565403580665588, 'train/grad_norm_pre_clip': 11.884357452392578, 'timing/data': 0.038216568529605865, 'timing/model': 1.5211889147758484, 'learning_rate/qwen_vl_interface': 2.1331755804161957e-06, 'learning_rate/action_model': 2.133175580416196e-05, 'epoch': 1.97}) +[2026-06-18 04:36:46,377][starVLA.training.train_starvla][INFO] - Step 5123, Loss: {'train/loss': 0.2122655212879181, 'train/grad_norm_pre_clip': 9.26953411102295, 'timing/data': 0.02325347252190113, 'timing/model': 1.136115837842226, 'learning_rate/qwen_vl_interface': 2.131543915577002e-06, 'learning_rate/action_model': 2.131543915577002e-05, 'epoch': 1.97}) +[2026-06-18 04:36:51,429][starVLA.training.train_starvla][INFO] - Step 5124, Loss: {'train/loss': 0.4213898479938507, 'train/grad_norm_pre_clip': 17.14767074584961, 'timing/data': 0.027282094582915306, 'timing/model': 1.2327904533594847, 'learning_rate/qwen_vl_interface': 2.129912897197108e-06, 'learning_rate/action_model': 2.1299128971971084e-05, 'epoch': 1.97}) +[2026-06-18 04:36:56,180][starVLA.training.train_starvla][INFO] - Step 5125, Loss: {'train/loss': 0.2230314016342163, 'train/grad_norm_pre_clip': 10.97950267791748, 'timing/data': 0.024491596035659313, 'timing/model': 1.1044674636796117, 'learning_rate/qwen_vl_interface': 2.1282825256146305e-06, 'learning_rate/action_model': 2.12828252561463e-05, 'epoch': 1.97}) +[2026-06-18 04:37:01,104][starVLA.training.train_starvla][INFO] - Step 5126, Loss: {'train/loss': 0.4062333106994629, 'train/grad_norm_pre_clip': 17.077381134033203, 'timing/data': 0.021029328927397728, 'timing/model': 1.2094570957124233, 'learning_rate/qwen_vl_interface': 2.1266528011675453e-06, 'learning_rate/action_model': 2.126652801167545e-05, 'epoch': 1.97}) +[2026-06-18 04:37:06,026][starVLA.training.train_starvla][INFO] - Step 5127, Loss: {'train/loss': 0.39125123620033264, 'train/grad_norm_pre_clip': 15.641606330871582, 'timing/data': 0.03623440209776163, 'timing/model': 1.4090651152655482, 'learning_rate/qwen_vl_interface': 2.125023724193694e-06, 'learning_rate/action_model': 2.1250237241936936e-05, 'epoch': 1.97}) +[2026-06-18 04:37:11,107][starVLA.training.train_starvla][INFO] - Step 5128, Loss: {'train/loss': 0.4317368268966675, 'train/grad_norm_pre_clip': 16.625516891479492, 'timing/data': 0.02990804146975279, 'timing/model': 1.2191579127684236, 'learning_rate/qwen_vl_interface': 2.1233952950307857e-06, 'learning_rate/action_model': 2.1233952950307855e-05, 'epoch': 1.97}) +[2026-06-18 04:37:16,281][starVLA.training.train_starvla][INFO] - Step 5129, Loss: {'train/loss': 0.24960272014141083, 'train/grad_norm_pre_clip': 13.039729118347168, 'timing/data': 0.024237877689301968, 'timing/model': 1.2822074797004461, 'learning_rate/qwen_vl_interface': 2.1217675140163984e-06, 'learning_rate/action_model': 2.1217675140163983e-05, 'epoch': 1.97}) +[2026-06-18 04:37:23,432][starVLA.training.train_starvla][INFO] - Step 5130, Loss: {'train/loss': 0.34242498874664307, 'train/grad_norm_pre_clip': 12.591325759887695, 'timing/data': 0.034300196915864944, 'timing/model': 3.1127064414322376, 'learning_rate/qwen_vl_interface': 2.1201403814879715e-06, 'learning_rate/action_model': 2.1201403814879717e-05, 'epoch': 1.97}) +[2026-06-18 04:37:32,536][starVLA.training.train_starvla][INFO] - Step 5131, Loss: {'train/loss': 0.21806593239307404, 'train/grad_norm_pre_clip': 10.088122367858887, 'timing/data': 0.03660585265606642, 'timing/model': 4.248436647467315, 'learning_rate/qwen_vl_interface': 2.1185138977828122e-06, 'learning_rate/action_model': 2.118513897782812e-05, 'epoch': 1.97}) +[2026-06-18 04:37:53,204][starVLA.training.train_starvla][INFO] - Step 5132, Loss: {'train/loss': 0.0924227312207222, 'train/grad_norm_pre_clip': 8.152653694152832, 'timing/data': 0.02763829566538334, 'timing/model': 5.950207019224763, 'learning_rate/qwen_vl_interface': 2.116888063238091e-06, 'learning_rate/action_model': 2.116888063238091e-05, 'epoch': 1.97}) +[2026-06-18 04:37:58,928][starVLA.training.train_starvla][INFO] - Step 5133, Loss: {'train/loss': 0.3787306249141693, 'train/grad_norm_pre_clip': 15.708984375, 'timing/data': 0.07179892715066671, 'timing/model': 1.2876254636794329, 'learning_rate/qwen_vl_interface': 2.1152628781908455e-06, 'learning_rate/action_model': 2.1152628781908455e-05, 'epoch': 1.97}) +[2026-06-18 04:38:05,482][starVLA.training.train_starvla][INFO] - Step 5134, Loss: {'train/loss': 0.27028340101242065, 'train/grad_norm_pre_clip': 12.385880470275879, 'timing/data': 0.023125979118049145, 'timing/model': 1.1355077838525176, 'learning_rate/qwen_vl_interface': 2.1136383429779787e-06, 'learning_rate/action_model': 2.1136383429779788e-05, 'epoch': 1.98}) +[2026-06-18 04:38:11,979][starVLA.training.train_starvla][INFO] - Step 5135, Loss: {'train/loss': 0.2680773437023163, 'train/grad_norm_pre_clip': 12.823062896728516, 'timing/data': 0.03267577476799488, 'timing/model': 1.237393724732101, 'learning_rate/qwen_vl_interface': 2.1120144579362582e-06, 'learning_rate/action_model': 2.1120144579362582e-05, 'epoch': 1.98}) +[2026-06-18 04:38:20,743][starVLA.training.train_starvla][INFO] - Step 5136, Loss: {'train/loss': 0.37134671211242676, 'train/grad_norm_pre_clip': 16.683866500854492, 'timing/data': 0.028731568716466427, 'timing/model': 2.760781737975776, 'learning_rate/qwen_vl_interface': 2.1103912234023175e-06, 'learning_rate/action_model': 2.1103912234023175e-05, 'epoch': 1.98}) +[2026-06-18 04:38:29,883][starVLA.training.train_starvla][INFO] - Step 5137, Loss: {'train/loss': 0.4355604350566864, 'train/grad_norm_pre_clip': 18.716609954833984, 'timing/data': 0.0307095879688859, 'timing/model': 1.5043321456760168, 'learning_rate/qwen_vl_interface': 2.1087686397126544e-06, 'learning_rate/action_model': 2.1087686397126545e-05, 'epoch': 1.98}) +[2026-06-18 04:38:35,793][starVLA.training.train_starvla][INFO] - Step 5138, Loss: {'train/loss': 0.2587197721004486, 'train/grad_norm_pre_clip': 11.421174049377441, 'timing/data': 0.0574731444939971, 'timing/model': 2.264139970764518, 'learning_rate/qwen_vl_interface': 2.107146707203632e-06, 'learning_rate/action_model': 2.1071467072036323e-05, 'epoch': 1.98}) +[2026-06-18 04:38:43,194][starVLA.training.train_starvla][INFO] - Step 5139, Loss: {'train/loss': 0.22117677330970764, 'train/grad_norm_pre_clip': 13.0904541015625, 'timing/data': 0.025952151976525784, 'timing/model': 1.1166525315493345, 'learning_rate/qwen_vl_interface': 2.105525426211478e-06, 'learning_rate/action_model': 2.1055254262114778e-05, 'epoch': 1.98}) +[2026-06-18 04:38:47,934][starVLA.training.train_starvla][INFO] - Step 5140, Loss: {'train/loss': 0.4582020342350006, 'train/grad_norm_pre_clip': 17.25932502746582, 'timing/data': 0.0233461931347847, 'timing/model': 1.1291850302368402, 'learning_rate/qwen_vl_interface': 2.1039047970722894e-06, 'learning_rate/action_model': 2.1039047970722893e-05, 'epoch': 1.98}) +[2026-06-18 04:38:52,672][starVLA.training.train_starvla][INFO] - Step 5141, Loss: {'train/loss': 0.48413291573524475, 'train/grad_norm_pre_clip': 19.149044036865234, 'timing/data': 0.021336224861443043, 'timing/model': 1.4324489757418633, 'learning_rate/qwen_vl_interface': 2.1022848201220204e-06, 'learning_rate/action_model': 2.1022848201220204e-05, 'epoch': 1.98}) +[2026-06-18 04:38:57,894][starVLA.training.train_starvla][INFO] - Step 5142, Loss: {'train/loss': 0.22854653000831604, 'train/grad_norm_pre_clip': 10.0474853515625, 'timing/data': 0.03307495545595884, 'timing/model': 1.574004402384162, 'learning_rate/qwen_vl_interface': 2.1006654956964935e-06, 'learning_rate/action_model': 2.1006654956964936e-05, 'epoch': 1.98}) +[2026-06-18 04:39:02,449][starVLA.training.train_starvla][INFO] - Step 5143, Loss: {'train/loss': 0.377587229013443, 'train/grad_norm_pre_clip': 20.136064529418945, 'timing/data': 0.025682538747787476, 'timing/model': 1.1308927247300744, 'learning_rate/qwen_vl_interface': 2.0990468241314004e-06, 'learning_rate/action_model': 2.0990468241314002e-05, 'epoch': 1.98}) +[2026-06-18 04:39:07,145][starVLA.training.train_starvla][INFO] - Step 5144, Loss: {'train/loss': 0.39640018343925476, 'train/grad_norm_pre_clip': 15.525050163269043, 'timing/data': 0.022078600712120533, 'timing/model': 1.11066332180053, 'learning_rate/qwen_vl_interface': 2.0974288057622905e-06, 'learning_rate/action_model': 2.0974288057622903e-05, 'epoch': 1.98}) +[2026-06-18 04:39:12,672][starVLA.training.train_starvla][INFO] - Step 5145, Loss: {'train/loss': 0.15694549679756165, 'train/grad_norm_pre_clip': 9.685320854187012, 'timing/data': 0.03563474863767624, 'timing/model': 1.1936764977872372, 'learning_rate/qwen_vl_interface': 2.095811440924584e-06, 'learning_rate/action_model': 2.0958114409245835e-05, 'epoch': 1.98}) +[2026-06-18 04:39:17,683][starVLA.training.train_starvla][INFO] - Step 5146, Loss: {'train/loss': 0.332459032535553, 'train/grad_norm_pre_clip': 8.461423873901367, 'timing/data': 0.03354696091264486, 'timing/model': 1.1840128218755126, 'learning_rate/qwen_vl_interface': 2.0941947299535554e-06, 'learning_rate/action_model': 2.0941947299535552e-05, 'epoch': 1.98}) +[2026-06-18 04:39:22,369][starVLA.training.train_starvla][INFO] - Step 5147, Loss: {'train/loss': 0.4580114185810089, 'train/grad_norm_pre_clip': 18.150333404541016, 'timing/data': 0.03341858368366957, 'timing/model': 1.4423113595694304, 'learning_rate/qwen_vl_interface': 2.0925786731843573e-06, 'learning_rate/action_model': 2.0925786731843572e-05, 'epoch': 1.98}) +[2026-06-18 04:39:27,052][starVLA.training.train_starvla][INFO] - Step 5148, Loss: {'train/loss': 0.3782714903354645, 'train/grad_norm_pre_clip': 15.804945945739746, 'timing/data': 0.020960218273103237, 'timing/model': 1.1809093784540892, 'learning_rate/qwen_vl_interface': 2.0909632709519976e-06, 'learning_rate/action_model': 2.090963270951998e-05, 'epoch': 1.98}) +[2026-06-18 04:39:32,050][starVLA.training.train_starvla][INFO] - Step 5149, Loss: {'train/loss': 0.5718328952789307, 'train/grad_norm_pre_clip': 19.559471130371094, 'timing/data': 0.022811057046055794, 'timing/model': 1.2720776423811913, 'learning_rate/qwen_vl_interface': 2.0893485235913514e-06, 'learning_rate/action_model': 2.0893485235913514e-05, 'epoch': 1.98}) +[2026-06-18 04:39:40,889][starVLA.training.train_starvla][INFO] - Step 5150, Loss: {'train/loss': 0.48136645555496216, 'train/grad_norm_pre_clip': 16.666751861572266, 'timing/data': 0.02156196627765894, 'timing/model': 1.08960843924433, 'learning_rate/qwen_vl_interface': 2.0877344314371575e-06, 'learning_rate/action_model': 2.0877344314371573e-05, 'epoch': 1.98}) +[2026-06-18 04:39:51,643][starVLA.training.train_starvla][INFO] - Step 5151, Loss: {'train/loss': 0.3825460970401764, 'train/grad_norm_pre_clip': 16.389835357666016, 'timing/data': 0.027419223450124264, 'timing/model': 5.948079103603959, 'learning_rate/qwen_vl_interface': 2.086120994824018e-06, 'learning_rate/action_model': 2.0861209948240178e-05, 'epoch': 1.98}) +[2026-06-18 04:39:56,593][starVLA.training.train_starvla][INFO] - Step 5152, Loss: {'train/loss': 0.7202937006950378, 'train/grad_norm_pre_clip': 19.236160278320312, 'timing/data': 0.023568199947476387, 'timing/model': 1.633704242296517, 'learning_rate/qwen_vl_interface': 2.0845082140864012e-06, 'learning_rate/action_model': 2.084508214086401e-05, 'epoch': 1.98}) +[2026-06-18 04:40:01,225][starVLA.training.train_starvla][INFO] - Step 5153, Loss: {'train/loss': 0.647846519947052, 'train/grad_norm_pre_clip': 26.228675842285156, 'timing/data': 0.0303977457806468, 'timing/model': 1.0886639477685094, 'learning_rate/qwen_vl_interface': 2.0828960895586365e-06, 'learning_rate/action_model': 2.0828960895586364e-05, 'epoch': 1.98}) +[2026-06-18 04:40:06,698][starVLA.training.train_starvla][INFO] - Step 5154, Loss: {'train/loss': 0.15994170308113098, 'train/grad_norm_pre_clip': 7.988142967224121, 'timing/data': 0.025273581966757774, 'timing/model': 1.4140604007989168, 'learning_rate/qwen_vl_interface': 2.0812846215749227e-06, 'learning_rate/action_model': 2.0812846215749227e-05, 'epoch': 1.98}) +[2026-06-18 04:40:11,574][starVLA.training.train_starvla][INFO] - Step 5155, Loss: {'train/loss': 0.32767054438591003, 'train/grad_norm_pre_clip': 16.326236724853516, 'timing/data': 0.03311433922499418, 'timing/model': 1.1516168117523193, 'learning_rate/qwen_vl_interface': 2.079673810469316e-06, 'learning_rate/action_model': 2.0796738104693158e-05, 'epoch': 1.98}) +[2026-06-18 04:40:16,447][starVLA.training.train_starvla][INFO] - Step 5156, Loss: {'train/loss': 0.30908170342445374, 'train/grad_norm_pre_clip': 13.623980522155762, 'timing/data': 0.022731192409992218, 'timing/model': 1.0894233174622059, 'learning_rate/qwen_vl_interface': 2.0780636565757392e-06, 'learning_rate/action_model': 2.0780636565757392e-05, 'epoch': 1.98}) +[2026-06-18 04:40:21,265][starVLA.training.train_starvla][INFO] - Step 5157, Loss: {'train/loss': 0.26332882046699524, 'train/grad_norm_pre_clip': 12.777026176452637, 'timing/data': 0.02345528081059456, 'timing/model': 1.442212282679975, 'learning_rate/qwen_vl_interface': 2.0764541602279777e-06, 'learning_rate/action_model': 2.0764541602279778e-05, 'epoch': 1.98}) +[2026-06-18 04:40:25,735][starVLA.training.train_starvla][INFO] - Step 5158, Loss: {'train/loss': 0.46403342485427856, 'train/grad_norm_pre_clip': 15.840372085571289, 'timing/data': 0.0249327402561903, 'timing/model': 1.0895514506846666, 'learning_rate/qwen_vl_interface': 2.0748453217596856e-06, 'learning_rate/action_model': 2.0748453217596856e-05, 'epoch': 1.98}) +[2026-06-18 04:40:30,730][starVLA.training.train_starvla][INFO] - Step 5159, Loss: {'train/loss': 0.612115204334259, 'train/grad_norm_pre_clip': 23.43738555908203, 'timing/data': 0.02269130852073431, 'timing/model': 1.1860689101740718, 'learning_rate/qwen_vl_interface': 2.0732371415043766e-06, 'learning_rate/action_model': 2.0732371415043765e-05, 'epoch': 1.98}) +[2026-06-18 04:40:39,528][starVLA.training.train_starvla][INFO] - Step 5160, Loss: {'train/loss': 0.35534510016441345, 'train/grad_norm_pre_clip': 16.374780654907227, 'timing/data': 0.040240622125566006, 'timing/model': 1.1954254005104303, 'learning_rate/qwen_vl_interface': 2.071629619795425e-06, 'learning_rate/action_model': 2.071629619795425e-05, 'epoch': 1.99}) +[2026-06-18 04:40:44,481][starVLA.training.train_starvla][INFO] - Step 5161, Loss: {'train/loss': 0.44694095849990845, 'train/grad_norm_pre_clip': 17.62763023376465, 'timing/data': 0.02192895393818617, 'timing/model': 1.514447532594204, 'learning_rate/qwen_vl_interface': 2.070022756966071e-06, 'learning_rate/action_model': 2.0700227569660712e-05, 'epoch': 1.99}) +[2026-06-18 04:40:49,491][starVLA.training.train_starvla][INFO] - Step 5162, Loss: {'train/loss': 0.37392741441726685, 'train/grad_norm_pre_clip': 15.613855361938477, 'timing/data': 0.02695351094007492, 'timing/model': 1.5729922140017152, 'learning_rate/qwen_vl_interface': 2.068416553349424e-06, 'learning_rate/action_model': 2.068416553349424e-05, 'epoch': 1.99}) +[2026-06-18 04:40:55,726][starVLA.training.train_starvla][INFO] - Step 5163, Loss: {'train/loss': 0.2940112352371216, 'train/grad_norm_pre_clip': 16.920032501220703, 'timing/data': 0.029455121606588364, 'timing/model': 1.194107142277062, 'learning_rate/qwen_vl_interface': 2.0668110092784484e-06, 'learning_rate/action_model': 2.0668110092784484e-05, 'epoch': 1.99}) +[2026-06-18 04:41:06,743][starVLA.training.train_starvla][INFO] - Step 5164, Loss: {'train/loss': 0.4138081967830658, 'train/grad_norm_pre_clip': 15.200163841247559, 'timing/data': 0.023536162450909615, 'timing/model': 1.3147716168314219, 'learning_rate/qwen_vl_interface': 2.0652061250859758e-06, 'learning_rate/action_model': 2.0652061250859756e-05, 'epoch': 1.99}) +[2026-06-18 04:41:12,201][starVLA.training.train_starvla][INFO] - Step 5165, Loss: {'train/loss': 0.33028632402420044, 'train/grad_norm_pre_clip': 26.83363914489746, 'timing/data': 0.02792961709201336, 'timing/model': 1.545511488802731, 'learning_rate/qwen_vl_interface': 2.0636019011047005e-06, 'learning_rate/action_model': 2.0636019011047003e-05, 'epoch': 1.99}) +[2026-06-18 04:41:16,915][starVLA.training.train_starvla][INFO] - Step 5166, Loss: {'train/loss': 0.529996395111084, 'train/grad_norm_pre_clip': 13.859357833862305, 'timing/data': 0.026496052742004395, 'timing/model': 1.1126623516902328, 'learning_rate/qwen_vl_interface': 2.0619983376671796e-06, 'learning_rate/action_model': 2.0619983376671793e-05, 'epoch': 1.99}) +[2026-06-18 04:41:21,719][starVLA.training.train_starvla][INFO] - Step 5167, Loss: {'train/loss': 0.38177287578582764, 'train/grad_norm_pre_clip': 15.705533981323242, 'timing/data': 0.022481233812868595, 'timing/model': 1.6305090300738811, 'learning_rate/qwen_vl_interface': 2.060395435105834e-06, 'learning_rate/action_model': 2.060395435105834e-05, 'epoch': 1.99}) +[2026-06-18 04:41:27,681][starVLA.training.train_starvla][INFO] - Step 5168, Loss: {'train/loss': 0.2161714732646942, 'train/grad_norm_pre_clip': 10.074925422668457, 'timing/data': 0.02651596162468195, 'timing/model': 1.0794828962534666, 'learning_rate/qwen_vl_interface': 2.0587931937529465e-06, 'learning_rate/action_model': 2.0587931937529464e-05, 'epoch': 1.99}) +[2026-06-18 04:41:32,708][starVLA.training.train_starvla][INFO] - Step 5169, Loss: {'train/loss': 0.2849152684211731, 'train/grad_norm_pre_clip': 11.46902847290039, 'timing/data': 0.023642947897315025, 'timing/model': 1.2215306172147393, 'learning_rate/qwen_vl_interface': 2.0571916139406648e-06, 'learning_rate/action_model': 2.0571916139406647e-05, 'epoch': 1.99}) +[2026-06-18 04:41:37,656][starVLA.training.train_starvla][INFO] - Step 5170, Loss: {'train/loss': 0.459297776222229, 'train/grad_norm_pre_clip': 18.677549362182617, 'timing/data': 0.023288649506866932, 'timing/model': 1.1163853127509356, 'learning_rate/qwen_vl_interface': 2.055590696000996e-06, 'learning_rate/action_model': 2.0555906960009962e-05, 'epoch': 1.99}) +[2026-06-18 04:41:42,978][starVLA.training.train_starvla][INFO] - Step 5171, Loss: {'train/loss': 0.4257350564002991, 'train/grad_norm_pre_clip': 19.907777786254883, 'timing/data': 0.030391253530979156, 'timing/model': 1.1684501050040126, 'learning_rate/qwen_vl_interface': 2.0539904402658144e-06, 'learning_rate/action_model': 2.0539904402658142e-05, 'epoch': 1.99}) +[2026-06-18 04:41:48,040][starVLA.training.train_starvla][INFO] - Step 5172, Loss: {'train/loss': 0.32589268684387207, 'train/grad_norm_pre_clip': 13.605871200561523, 'timing/data': 0.02103897649794817, 'timing/model': 1.49810357298702, 'learning_rate/qwen_vl_interface': 2.052390847066851e-06, 'learning_rate/action_model': 2.052390847066851e-05, 'epoch': 1.99}) +[2026-06-18 04:42:00,795][starVLA.training.train_starvla][INFO] - Step 5173, Loss: {'train/loss': 0.17598117887973785, 'train/grad_norm_pre_clip': 10.013254165649414, 'timing/data': 0.026278290897607803, 'timing/model': 5.871781822293997, 'learning_rate/qwen_vl_interface': 2.0507919167357093e-06, 'learning_rate/action_model': 2.050791916735709e-05, 'epoch': 1.99}) +[2026-06-18 04:42:05,613][starVLA.training.train_starvla][INFO] - Step 5174, Loss: {'train/loss': 0.2667275071144104, 'train/grad_norm_pre_clip': 17.02046775817871, 'timing/data': 0.03429087810218334, 'timing/model': 1.1899678893387318, 'learning_rate/qwen_vl_interface': 2.049193649603844e-06, 'learning_rate/action_model': 2.049193649603844e-05, 'epoch': 1.99}) +[2026-06-18 04:42:10,622][starVLA.training.train_starvla][INFO] - Step 5175, Loss: {'train/loss': 0.35470178723335266, 'train/grad_norm_pre_clip': 15.264630317687988, 'timing/data': 0.021450252272188663, 'timing/model': 1.0951281925663352, 'learning_rate/qwen_vl_interface': 2.047596046002578e-06, 'learning_rate/action_model': 2.0475960460025778e-05, 'epoch': 1.99}) +[2026-06-18 04:42:17,241][starVLA.training.train_starvla][INFO] - Step 5176, Loss: {'train/loss': 0.30010825395584106, 'train/grad_norm_pre_clip': 15.196372985839844, 'timing/data': 0.03971379064023495, 'timing/model': 1.1299816546961665, 'learning_rate/qwen_vl_interface': 2.0459991062630994e-06, 'learning_rate/action_model': 2.045999106263099e-05, 'epoch': 1.99}) +[2026-06-18 04:42:22,249][starVLA.training.train_starvla][INFO] - Step 5177, Loss: {'train/loss': 0.36644357442855835, 'train/grad_norm_pre_clip': 11.451458930969238, 'timing/data': 0.03023610543459654, 'timing/model': 1.5186617784202099, 'learning_rate/qwen_vl_interface': 2.0444028307164532e-06, 'learning_rate/action_model': 2.0444028307164534e-05, 'epoch': 1.99}) +[2026-06-18 04:42:26,767][starVLA.training.train_starvla][INFO] - Step 5178, Loss: {'train/loss': 0.2122974991798401, 'train/grad_norm_pre_clip': 14.301674842834473, 'timing/data': 0.022533987648785114, 'timing/model': 1.1013937061652541, 'learning_rate/qwen_vl_interface': 2.0428072196935516e-06, 'learning_rate/action_model': 2.0428072196935514e-05, 'epoch': 1.99}) +[2026-06-18 04:42:31,997][starVLA.training.train_starvla][INFO] - Step 5179, Loss: {'train/loss': 0.48013070225715637, 'train/grad_norm_pre_clip': 16.282169342041016, 'timing/data': 0.024949591606855392, 'timing/model': 1.208836098201573, 'learning_rate/qwen_vl_interface': 2.041212273525161e-06, 'learning_rate/action_model': 2.041212273525161e-05, 'epoch': 1.99}) +[2026-06-18 04:42:38,114][starVLA.training.train_starvla][INFO] - Step 5180, Loss: {'train/loss': 0.31050893664360046, 'train/grad_norm_pre_clip': 19.721921920776367, 'timing/data': 0.026127946563065052, 'timing/model': 1.198378392495215, 'learning_rate/qwen_vl_interface': 2.03961799254192e-06, 'learning_rate/action_model': 2.03961799254192e-05, 'epoch': 1.99}) +[2026-06-18 04:42:46,200][starVLA.training.train_starvla][INFO] - Step 5181, Loss: {'train/loss': 0.43108925223350525, 'train/grad_norm_pre_clip': 21.982038497924805, 'timing/data': 0.02195310778915882, 'timing/model': 4.363351379521191, 'learning_rate/qwen_vl_interface': 2.0380243770743246e-06, 'learning_rate/action_model': 2.0380243770743247e-05, 'epoch': 1.99}) +[2026-06-18 04:42:53,053][starVLA.training.train_starvla][INFO] - Step 5182, Loss: {'train/loss': 0.2395879328250885, 'train/grad_norm_pre_clip': 11.027040481567383, 'timing/data': 0.02448410540819168, 'timing/model': 1.5091179935261607, 'learning_rate/qwen_vl_interface': 2.0364314274527318e-06, 'learning_rate/action_model': 2.0364314274527314e-05, 'epoch': 1.99}) +[2026-06-18 04:42:57,471][starVLA.training.train_starvla][INFO] - Step 5183, Loss: {'train/loss': 0.46747589111328125, 'train/grad_norm_pre_clip': 12.969063758850098, 'timing/data': 0.024589989334344864, 'timing/model': 1.1632893225178123, 'learning_rate/qwen_vl_interface': 2.034839144007362e-06, 'learning_rate/action_model': 2.034839144007362e-05, 'epoch': 1.99}) +[2026-06-18 04:43:03,314][starVLA.training.train_starvla][INFO] - Step 5184, Loss: {'train/loss': 0.2530573606491089, 'train/grad_norm_pre_clip': 10.123918533325195, 'timing/data': 0.022364623844623566, 'timing/model': 1.0801252853125334, 'learning_rate/qwen_vl_interface': 2.033247527068297e-06, 'learning_rate/action_model': 2.0332475270682972e-05, 'epoch': 1.99}) +[2026-06-18 04:43:08,093][starVLA.training.train_starvla][INFO] - Step 5185, Loss: {'train/loss': 0.17442984879016876, 'train/grad_norm_pre_clip': 13.159627914428711, 'timing/data': 0.030154841020703316, 'timing/model': 1.1361724650487304, 'learning_rate/qwen_vl_interface': 2.0316565769654816e-06, 'learning_rate/action_model': 2.0316565769654816e-05, 'epoch': 1.99}) +[2026-06-18 04:43:12,990][starVLA.training.train_starvla][INFO] - Step 5186, Loss: {'train/loss': 0.464781254529953, 'train/grad_norm_pre_clip': 16.138647079467773, 'timing/data': 0.03330247849225998, 'timing/model': 1.198677490465343, 'learning_rate/qwen_vl_interface': 2.0300662940287193e-06, 'learning_rate/action_model': 2.0300662940287192e-05, 'epoch': 2.0}) +[2026-06-18 04:43:17,811][starVLA.training.train_starvla][INFO] - Step 5187, Loss: {'train/loss': 0.2938803732395172, 'train/grad_norm_pre_clip': 16.22164535522461, 'timing/data': 0.022967789322137833, 'timing/model': 1.5657772924751043, 'learning_rate/qwen_vl_interface': 2.0284766785876812e-06, 'learning_rate/action_model': 2.0284766785876812e-05, 'epoch': 2.0}) +[2026-06-18 04:43:22,290][starVLA.training.train_starvla][INFO] - Step 5188, Loss: {'train/loss': 0.24752329289913177, 'train/grad_norm_pre_clip': 15.12823486328125, 'timing/data': 0.030671125277876854, 'timing/model': 1.1744059575721622, 'learning_rate/qwen_vl_interface': 2.026887730971894e-06, 'learning_rate/action_model': 2.0268877309718938e-05, 'epoch': 2.0}) +[2026-06-18 04:43:27,315][starVLA.training.train_starvla][INFO] - Step 5189, Loss: {'train/loss': 0.20639590919017792, 'train/grad_norm_pre_clip': 10.245370864868164, 'timing/data': 0.02457600738853216, 'timing/model': 1.0862631713971496, 'learning_rate/qwen_vl_interface': 2.0252994515107475e-06, 'learning_rate/action_model': 2.0252994515107472e-05, 'epoch': 2.0}) +[2026-06-18 04:43:34,519][starVLA.training.train_starvla][INFO] - Step 5190, Loss: {'train/loss': 0.1016320064663887, 'train/grad_norm_pre_clip': 10.537010192871094, 'timing/data': 0.02117555122822523, 'timing/model': 1.1262211464345455, 'learning_rate/qwen_vl_interface': 2.023711840533493e-06, 'learning_rate/action_model': 2.023711840533493e-05, 'epoch': 2.0}) +[2026-06-18 04:43:39,240][starVLA.training.train_starvla][INFO] - Step 5191, Loss: {'train/loss': 0.2306223213672638, 'train/grad_norm_pre_clip': 10.38128662109375, 'timing/data': 0.026302948594093323, 'timing/model': 1.0871315635740757, 'learning_rate/qwen_vl_interface': 2.0221248983692475e-06, 'learning_rate/action_model': 2.0221248983692474e-05, 'epoch': 2.0}) +[2026-06-18 04:43:44,051][starVLA.training.train_starvla][INFO] - Step 5192, Loss: {'train/loss': 0.2401835024356842, 'train/grad_norm_pre_clip': 10.83670711517334, 'timing/data': 0.029798460192978382, 'timing/model': 1.5278905713930726, 'learning_rate/qwen_vl_interface': 2.020538625346984e-06, 'learning_rate/action_model': 2.020538625346984e-05, 'epoch': 2.0}) +[2026-06-18 04:43:48,452][starVLA.training.train_starvla][INFO] - Step 5193, Loss: {'train/loss': 0.10366899520158768, 'train/grad_norm_pre_clip': 6.3451948165893555, 'timing/data': 0.03382145520299673, 'timing/model': 1.1745663238689303, 'learning_rate/qwen_vl_interface': 2.0189530217955382e-06, 'learning_rate/action_model': 2.018953021795538e-05, 'epoch': 2.0}) +[2026-06-18 04:43:53,390][starVLA.training.train_starvla][INFO] - Step 5194, Loss: {'train/loss': 0.16967731714248657, 'train/grad_norm_pre_clip': 11.985671043395996, 'timing/data': 0.024571520276367664, 'timing/model': 1.1477159820497036, 'learning_rate/qwen_vl_interface': 2.0173680880436076e-06, 'learning_rate/action_model': 2.0173680880436076e-05, 'epoch': 2.0}) +[2026-06-18 04:43:58,189][starVLA.training.train_starvla][INFO] - Step 5195, Loss: {'train/loss': 0.1845519244670868, 'train/grad_norm_pre_clip': 10.546745300292969, 'timing/data': 0.023042423650622368, 'timing/model': 1.1262561660259962, 'learning_rate/qwen_vl_interface': 2.015783824419751e-06, 'learning_rate/action_model': 2.0157838244197508e-05, 'epoch': 2.0}) +[2026-06-18 04:44:03,070][starVLA.training.train_starvla][INFO] - Step 5196, Loss: {'train/loss': 0.5516745448112488, 'train/grad_norm_pre_clip': 24.569541931152344, 'timing/data': 0.024057584814727306, 'timing/model': 1.5779641810804605, 'learning_rate/qwen_vl_interface': 2.0142002312523877e-06, 'learning_rate/action_model': 2.014200231252388e-05, 'epoch': 2.0}) +[2026-06-18 04:44:07,791][starVLA.training.train_starvla][INFO] - Step 5197, Loss: {'train/loss': 0.27170735597610474, 'train/grad_norm_pre_clip': 35.31332778930664, 'timing/data': 0.02323444839566946, 'timing/model': 1.4286047499626875, 'learning_rate/qwen_vl_interface': 2.0126173088697988e-06, 'learning_rate/action_model': 2.0126173088697987e-05, 'epoch': 2.0}) +[2026-06-18 04:44:12,175][starVLA.training.train_starvla][INFO] - Step 5198, Loss: {'train/loss': 0.2966934144496918, 'train/grad_norm_pre_clip': 12.498856544494629, 'timing/data': 0.02099297009408474, 'timing/model': 1.0956539250910282, 'learning_rate/qwen_vl_interface': 2.0110350576001244e-06, 'learning_rate/action_model': 2.0110350576001245e-05, 'epoch': 2.0}) +[2026-06-18 04:44:18,004][starVLA.training.train_starvla][INFO] - Step 5199, Loss: {'train/loss': 0.26495006680488586, 'train/grad_norm_pre_clip': 13.517329216003418, 'timing/data': 0.0010163448750972748, 'timing/model': 2.328945428133011, 'learning_rate/qwen_vl_interface': 2.0094534777713684e-06, 'learning_rate/action_model': 2.0094534777713684e-05, 'epoch': 2.0}) +[2026-06-18 04:47:12,344][starVLA.training.train_starvla][INFO] - Step 5200, Loss: {'train/loss': 0.31969791650772095, 'train/grad_norm_pre_clip': 12.799982070922852, 'eval/loss': 0.42516160011291504, 'eval/latency_4/loss': 0.42516160011291504, 'eval/demon_fire_precision': 0.9655941382605926, 'eval/demon_fire_recall': 0.9844105228970444, 'eval/demon_fire_f1': 0.9749115471212609, 'eval/demon_fire_rate_model': 0.9218795888399413, 'eval/demon_fire_rate_teacher': 0.9042584434654919, 'eval/demon_left_precision': 0.8074712643678161, 'eval/demon_left_recall': 0.6415525114155252, 'eval/demon_left_f1': 0.7150127226463104, 'eval/demon_left_rate_model': 0.10220264317180616, 'eval/demon_left_rate_teacher': 0.12863436123348018, 'eval/demon_right_precision': 0.8022181146025879, 'eval/demon_right_recall': 0.7833935018050542, 'eval/demon_right_f1': 0.7926940639269406, 'eval/demon_right_rate_model': 0.15888399412628487, 'eval/demon_right_rate_teacher': 0.16270190895741557, 'eval/demon_move_f1': 0.7538533932866256, 'eval/demon_cc_f1': 0.8960820631664449, 'eval/latency_4/demon_fire_precision': 0.9655941382605926, 'eval/latency_4/demon_fire_recall': 0.9844105228970444, 'eval/latency_4/demon_fire_f1': 0.9749115471212609, 'eval/latency_4/demon_fire_rate_model': 0.9218795888399413, 'eval/latency_4/demon_fire_rate_teacher': 0.9042584434654919, 'eval/latency_4/demon_left_precision': 0.8074712643678161, 'eval/latency_4/demon_left_recall': 0.6415525114155252, 'eval/latency_4/demon_left_f1': 0.7150127226463104, 'eval/latency_4/demon_left_rate_model': 0.10220264317180616, 'eval/latency_4/demon_left_rate_teacher': 0.12863436123348018, 'eval/latency_4/demon_right_precision': 0.8022181146025879, 'eval/latency_4/demon_right_recall': 0.7833935018050542, 'eval/latency_4/demon_right_f1': 0.7926940639269406, 'eval/latency_4/demon_right_rate_model': 0.15888399412628487, 'eval/latency_4/demon_right_rate_teacher': 0.16270190895741557, 'eval/latency_4/demon_move_f1': 0.7538533932866256, 'eval/latency_4/demon_cc_f1': 0.8960820631664449, 'timing/data': 0.03544298280030489, 'timing/model': 1.1618828112259507, 'learning_rate/qwen_vl_interface': 2.007872569711394e-06, 'learning_rate/action_model': 2.007872569711394e-05, 'epoch': 2.0}) +[2026-06-18 04:47:12,395][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_5200_state +[2026-06-18 04:47:12,403][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 04:47:21,458][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_5200_state/model.safetensors +[2026-06-18 04:47:44,613][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_5200_state/optimizer.bin +[2026-06-18 04:47:44,617][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_5200_state/sampler.bin +[2026-06-18 04:47:44,618][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_5200_state/sampler_1.bin +[2026-06-18 04:47:44,622][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_5200_state/random_states_0.pkl +[2026-06-18 04:47:44,631][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 5200 +[2026-06-18 04:47:44,631][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 04:47:44,666][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 04:47:49,574][starVLA.training.train_starvla][INFO] - Step 5201, Loss: {'train/loss': 0.44941431283950806, 'train/grad_norm_pre_clip': 21.573911666870117, 'timing/data': 0.044345561414957047, 'timing/model': 1.5366070345044136, 'learning_rate/qwen_vl_interface': 2.0062923337479223e-06, 'learning_rate/action_model': 2.006292333747922e-05, 'epoch': 2.0}) +[2026-06-18 04:47:55,982][starVLA.training.train_starvla][INFO] - Step 5202, Loss: {'train/loss': 0.2822105288505554, 'train/grad_norm_pre_clip': 13.282402038574219, 'timing/data': 0.05724501423537731, 'timing/model': 2.4411484645679593, 'learning_rate/qwen_vl_interface': 2.0047127702085437e-06, 'learning_rate/action_model': 2.0047127702085437e-05, 'epoch': 2.0}) +[2026-06-18 04:48:04,964][starVLA.training.train_starvla][INFO] - Step 5203, Loss: {'train/loss': 0.2679497003555298, 'train/grad_norm_pre_clip': 13.859126091003418, 'timing/data': 0.03367441240698099, 'timing/model': 1.1752055790275335, 'learning_rate/qwen_vl_interface': 2.003133879420698e-06, 'learning_rate/action_model': 2.003133879420698e-05, 'epoch': 2.0}) +[2026-06-18 04:48:11,835][starVLA.training.train_starvla][INFO] - Step 5204, Loss: {'train/loss': 0.3347817361354828, 'train/grad_norm_pre_clip': 15.45262336730957, 'timing/data': 0.03236516751348972, 'timing/model': 1.125129159539938, 'learning_rate/qwen_vl_interface': 2.001555661711691e-06, 'learning_rate/action_model': 2.001555661711691e-05, 'epoch': 2.0}) +[2026-06-18 04:48:16,653][starVLA.training.train_starvla][INFO] - Step 5205, Loss: {'train/loss': 0.24604003131389618, 'train/grad_norm_pre_clip': 11.73122787475586, 'timing/data': 0.04448826424777508, 'timing/model': 1.1043957835063338, 'learning_rate/qwen_vl_interface': 1.999978117408693e-06, 'learning_rate/action_model': 1.999978117408693e-05, 'epoch': 2.0}) +[2026-06-18 04:48:21,580][starVLA.training.train_starvla][INFO] - Step 5206, Loss: {'train/loss': 0.5656353831291199, 'train/grad_norm_pre_clip': 24.481399536132812, 'timing/data': 0.04075335431843996, 'timing/model': 1.0902487644925714, 'learning_rate/qwen_vl_interface': 1.9984012468387285e-06, 'learning_rate/action_model': 1.9984012468387285e-05, 'epoch': 2.0}) +[2026-06-18 04:48:28,063][starVLA.training.train_starvla][INFO] - Step 5207, Loss: {'train/loss': 0.16399407386779785, 'train/grad_norm_pre_clip': 8.91574478149414, 'timing/data': 0.031490092165768147, 'timing/model': 1.712757515721023, 'learning_rate/qwen_vl_interface': 1.9968250503286857e-06, 'learning_rate/action_model': 1.996825050328686e-05, 'epoch': 2.0}) +[2026-06-18 04:48:35,742][starVLA.training.train_starvla][INFO] - Step 5208, Loss: {'train/loss': 0.29689040780067444, 'train/grad_norm_pre_clip': 15.256443977355957, 'timing/data': 0.03306636493653059, 'timing/model': 1.1245603039860725, 'learning_rate/qwen_vl_interface': 1.9952495282053073e-06, 'learning_rate/action_model': 1.9952495282053073e-05, 'epoch': 2.0}) +[2026-06-18 04:48:42,957][starVLA.training.train_starvla][INFO] - Step 5209, Loss: {'train/loss': 0.4954644739627838, 'train/grad_norm_pre_clip': 18.739219665527344, 'timing/data': 0.031091036275029182, 'timing/model': 1.2135069472715259, 'learning_rate/qwen_vl_interface': 1.993674680795206e-06, 'learning_rate/action_model': 1.9936746807952058e-05, 'epoch': 2.0}) +[2026-06-18 04:48:47,757][starVLA.training.train_starvla][INFO] - Step 5210, Loss: {'train/loss': 0.23152773082256317, 'train/grad_norm_pre_clip': 13.022319793701172, 'timing/data': 0.03194639552384615, 'timing/model': 1.2188192559406161, 'learning_rate/qwen_vl_interface': 1.992100508424847e-06, 'learning_rate/action_model': 1.9921005084248467e-05, 'epoch': 2.0}) +[2026-06-18 04:48:52,759][starVLA.training.train_starvla][INFO] - Step 5211, Loss: {'train/loss': 0.26054948568344116, 'train/grad_norm_pre_clip': 9.485950469970703, 'timing/data': 0.04213285353034735, 'timing/model': 1.143382896669209, 'learning_rate/qwen_vl_interface': 1.990527011420558e-06, 'learning_rate/action_model': 1.990527011420558e-05, 'epoch': 2.0}) +[2026-06-18 04:48:57,442][starVLA.training.train_starvla][INFO] - Step 5212, Loss: {'train/loss': 0.30627062916755676, 'train/grad_norm_pre_clip': 13.075201988220215, 'timing/data': 0.03616233356297016, 'timing/model': 1.379550895653665, 'learning_rate/qwen_vl_interface': 1.9889541901085274e-06, 'learning_rate/action_model': 1.9889541901085273e-05, 'epoch': 2.01}) +[2026-06-18 04:49:02,046][starVLA.training.train_starvla][INFO] - Step 5213, Loss: {'train/loss': 0.32587599754333496, 'train/grad_norm_pre_clip': 11.510504722595215, 'timing/data': 0.039610263891518116, 'timing/model': 1.1115030637010932, 'learning_rate/qwen_vl_interface': 1.987382044814802e-06, 'learning_rate/action_model': 1.987382044814802e-05, 'epoch': 2.01}) +[2026-06-18 04:49:06,872][starVLA.training.train_starvla][INFO] - Step 5214, Loss: {'train/loss': 0.4430414140224457, 'train/grad_norm_pre_clip': 18.676191329956055, 'timing/data': 0.029284951277077198, 'timing/model': 1.282750453799963, 'learning_rate/qwen_vl_interface': 1.9858105758652895e-06, 'learning_rate/action_model': 1.9858105758652892e-05, 'epoch': 2.01}) +[2026-06-18 04:49:11,647][starVLA.training.train_starvla][INFO] - Step 5215, Loss: {'train/loss': 0.26769909262657166, 'train/grad_norm_pre_clip': 11.443344116210938, 'timing/data': 0.030599194578826427, 'timing/model': 1.2682309737429023, 'learning_rate/qwen_vl_interface': 1.9842397835857555e-06, 'learning_rate/action_model': 1.9842397835857554e-05, 'epoch': 2.01}) +[2026-06-18 04:49:16,528][starVLA.training.train_starvla][INFO] - Step 5216, Loss: {'train/loss': 0.442528635263443, 'train/grad_norm_pre_clip': 14.330656051635742, 'timing/data': 0.03319305367767811, 'timing/model': 1.1685495609417558, 'learning_rate/qwen_vl_interface': 1.982669668301831e-06, 'learning_rate/action_model': 1.982669668301831e-05, 'epoch': 2.01}) +[2026-06-18 04:49:21,229][starVLA.training.train_starvla][INFO] - Step 5217, Loss: {'train/loss': 0.25645023584365845, 'train/grad_norm_pre_clip': 11.769953727722168, 'timing/data': 0.033776238560676575, 'timing/model': 1.4528805455192924, 'learning_rate/qwen_vl_interface': 1.9811002303389993e-06, 'learning_rate/action_model': 1.9811002303389993e-05, 'epoch': 2.01}) +[2026-06-18 04:49:25,533][starVLA.training.train_starvla][INFO] - Step 5218, Loss: {'train/loss': 0.2505638599395752, 'train/grad_norm_pre_clip': 14.168845176696777, 'timing/data': 0.04298604838550091, 'timing/model': 1.1100611379370093, 'learning_rate/qwen_vl_interface': 1.9795314700226068e-06, 'learning_rate/action_model': 1.9795314700226066e-05, 'epoch': 2.01}) +[2026-06-18 04:49:30,579][starVLA.training.train_starvla][INFO] - Step 5219, Loss: {'train/loss': 0.3554905652999878, 'train/grad_norm_pre_clip': 15.338462829589844, 'timing/data': 0.03104604035615921, 'timing/model': 1.1725767478346825, 'learning_rate/qwen_vl_interface': 1.9779633876778587e-06, 'learning_rate/action_model': 1.9779633876778586e-05, 'epoch': 2.01}) +[2026-06-18 04:49:35,539][starVLA.training.train_starvla][INFO] - Step 5220, Loss: {'train/loss': 0.36773601174354553, 'train/grad_norm_pre_clip': 24.228740692138672, 'timing/data': 0.03444046247750521, 'timing/model': 1.1045730197802186, 'learning_rate/qwen_vl_interface': 1.9763959836298225e-06, 'learning_rate/action_model': 1.9763959836298225e-05, 'epoch': 2.01}) +[2026-06-18 04:49:40,325][starVLA.training.train_starvla][INFO] - Step 5221, Loss: {'train/loss': 0.30179163813591003, 'train/grad_norm_pre_clip': 14.109682083129883, 'timing/data': 0.03340105339884758, 'timing/model': 1.462834682315588, 'learning_rate/qwen_vl_interface': 1.9748292582034228e-06, 'learning_rate/action_model': 1.9748292582034228e-05, 'epoch': 2.01}) +[2026-06-18 04:49:45,099][starVLA.training.train_starvla][INFO] - Step 5222, Loss: {'train/loss': 0.3114475607872009, 'train/grad_norm_pre_clip': 12.064554214477539, 'timing/data': 0.03355093952268362, 'timing/model': 1.4580871211364865, 'learning_rate/qwen_vl_interface': 1.9732632117234405e-06, 'learning_rate/action_model': 1.9732632117234404e-05, 'epoch': 2.01}) +[2026-06-18 04:49:49,405][starVLA.training.train_starvla][INFO] - Step 5223, Loss: {'train/loss': 0.2977868914604187, 'train/grad_norm_pre_clip': 13.045843124389648, 'timing/data': 0.054582735523581505, 'timing/model': 1.1015131995081902, 'learning_rate/qwen_vl_interface': 1.9716978445145195e-06, 'learning_rate/action_model': 1.9716978445145193e-05, 'epoch': 2.01}) +[2026-06-18 04:49:54,499][starVLA.training.train_starvla][INFO] - Step 5224, Loss: {'train/loss': 0.26171478629112244, 'train/grad_norm_pre_clip': 13.077034950256348, 'timing/data': 0.03751803934574127, 'timing/model': 1.2707182224839926, 'learning_rate/qwen_vl_interface': 1.9701331569011634e-06, 'learning_rate/action_model': 1.9701331569011633e-05, 'epoch': 2.01}) +[2026-06-18 04:50:00,991][starVLA.training.train_starvla][INFO] - Step 5225, Loss: {'train/loss': 0.2867741882801056, 'train/grad_norm_pre_clip': 19.947723388671875, 'timing/data': 0.03279220499098301, 'timing/model': 1.1511495122686028, 'learning_rate/qwen_vl_interface': 1.968569149207734e-06, 'learning_rate/action_model': 1.968569149207734e-05, 'epoch': 2.01}) +[2026-06-18 04:50:13,831][starVLA.training.train_starvla][INFO] - Step 5226, Loss: {'train/loss': 0.2586882710456848, 'train/grad_norm_pre_clip': 16.339527130126953, 'timing/data': 0.0341872563585639, 'timing/model': 4.1206483552232385, 'learning_rate/qwen_vl_interface': 1.9670058217584505e-06, 'learning_rate/action_model': 1.9670058217584504e-05, 'epoch': 2.01}) +[2026-06-18 04:50:18,947][starVLA.training.train_starvla][INFO] - Step 5227, Loss: {'train/loss': 0.5081804990768433, 'train/grad_norm_pre_clip': 21.08591079711914, 'timing/data': 0.047644312493503094, 'timing/model': 1.5597636960446835, 'learning_rate/qwen_vl_interface': 1.965443174877393e-06, 'learning_rate/action_model': 1.965443174877393e-05, 'epoch': 2.01}) +[2026-06-18 04:50:28,940][starVLA.training.train_starvla][INFO] - Step 5228, Loss: {'train/loss': 0.30399590730667114, 'train/grad_norm_pre_clip': 18.376249313354492, 'timing/data': 0.041711119934916496, 'timing/model': 1.1107378397136927, 'learning_rate/qwen_vl_interface': 1.9638812088885003e-06, 'learning_rate/action_model': 1.9638812088885e-05, 'epoch': 2.01}) +[2026-06-18 04:50:34,208][starVLA.training.train_starvla][INFO] - Step 5229, Loss: {'train/loss': 0.24593324959278107, 'train/grad_norm_pre_clip': 17.659303665161133, 'timing/data': 0.0401134742423892, 'timing/model': 1.3611823534592986, 'learning_rate/qwen_vl_interface': 1.9623199241155684e-06, 'learning_rate/action_model': 1.9623199241155683e-05, 'epoch': 2.01}) +[2026-06-18 04:50:39,293][starVLA.training.train_starvla][INFO] - Step 5230, Loss: {'train/loss': 0.3355051577091217, 'train/grad_norm_pre_clip': 11.63420581817627, 'timing/data': 0.04491866659373045, 'timing/model': 1.1009350921958685, 'learning_rate/qwen_vl_interface': 1.9607593208822536e-06, 'learning_rate/action_model': 1.9607593208822535e-05, 'epoch': 2.01}) +[2026-06-18 04:50:45,997][starVLA.training.train_starvla][INFO] - Step 5231, Loss: {'train/loss': 0.41703733801841736, 'train/grad_norm_pre_clip': 17.835607528686523, 'timing/data': 0.03396053146570921, 'timing/model': 1.635038428939879, 'learning_rate/qwen_vl_interface': 1.9591993995120715e-06, 'learning_rate/action_model': 1.9591993995120712e-05, 'epoch': 2.01}) +[2026-06-18 04:50:50,623][starVLA.training.train_starvla][INFO] - Step 5232, Loss: {'train/loss': 0.28755274415016174, 'train/grad_norm_pre_clip': 12.815539360046387, 'timing/data': 0.030834523029625416, 'timing/model': 1.402696918696165, 'learning_rate/qwen_vl_interface': 1.9576401603283944e-06, 'learning_rate/action_model': 1.9576401603283944e-05, 'epoch': 2.01}) +[2026-06-18 04:50:55,239][starVLA.training.train_starvla][INFO] - Step 5233, Loss: {'train/loss': 0.3703840374946594, 'train/grad_norm_pre_clip': 14.536056518554688, 'timing/data': 0.031813446432352066, 'timing/model': 1.2341529922559857, 'learning_rate/qwen_vl_interface': 1.956081603654455e-06, 'learning_rate/action_model': 1.956081603654455e-05, 'epoch': 2.01}) +[2026-06-18 04:51:00,175][starVLA.training.train_starvla][INFO] - Step 5234, Loss: {'train/loss': 0.17852871119976044, 'train/grad_norm_pre_clip': 9.43298053741455, 'timing/data': 0.03963320795446634, 'timing/model': 1.1588844768702984, 'learning_rate/qwen_vl_interface': 1.9545237298133423e-06, 'learning_rate/action_model': 1.9545237298133424e-05, 'epoch': 2.01}) +[2026-06-18 04:51:04,995][starVLA.training.train_starvla][INFO] - Step 5235, Loss: {'train/loss': 0.3897860646247864, 'train/grad_norm_pre_clip': 13.353869438171387, 'timing/data': 0.02901138924062252, 'timing/model': 1.1118506053462625, 'learning_rate/qwen_vl_interface': 1.9529665391280096e-06, 'learning_rate/action_model': 1.95296653912801e-05, 'epoch': 2.01}) +[2026-06-18 04:51:09,978][starVLA.training.train_starvla][INFO] - Step 5236, Loss: {'train/loss': 0.4086272418498993, 'train/grad_norm_pre_clip': 15.924707412719727, 'timing/data': 0.03394368849694729, 'timing/model': 1.1248967899009585, 'learning_rate/qwen_vl_interface': 1.9514100319212594e-06, 'learning_rate/action_model': 1.9514100319212593e-05, 'epoch': 2.01}) +[2026-06-18 04:51:14,971][starVLA.training.train_starvla][INFO] - Step 5237, Loss: {'train/loss': 0.5494676828384399, 'train/grad_norm_pre_clip': 20.918601989746094, 'timing/data': 0.04409180488437414, 'timing/model': 1.5813117315992713, 'learning_rate/qwen_vl_interface': 1.949854208515757e-06, 'learning_rate/action_model': 1.9498542085157573e-05, 'epoch': 2.01}) +[2026-06-18 04:51:19,335][starVLA.training.train_starvla][INFO] - Step 5238, Loss: {'train/loss': 0.2723865807056427, 'train/grad_norm_pre_clip': 11.023208618164062, 'timing/data': 0.031845989637076855, 'timing/model': 1.1108570955693722, 'learning_rate/qwen_vl_interface': 1.9482990692340305e-06, 'learning_rate/action_model': 1.9482990692340305e-05, 'epoch': 2.02}) +[2026-06-18 04:51:24,099][starVLA.training.train_starvla][INFO] - Step 5239, Loss: {'train/loss': 0.5381481647491455, 'train/grad_norm_pre_clip': 16.163421630859375, 'timing/data': 0.0537574477493763, 'timing/model': 1.149227655492723, 'learning_rate/qwen_vl_interface': 1.94674461439846e-06, 'learning_rate/action_model': 1.9467446143984598e-05, 'epoch': 2.02}) +[2026-06-18 04:51:29,791][starVLA.training.train_starvla][INFO] - Step 5240, Loss: {'train/loss': 0.7012034058570862, 'train/grad_norm_pre_clip': 22.294322967529297, 'timing/data': 0.042333449237048626, 'timing/model': 1.1093345377594233, 'learning_rate/qwen_vl_interface': 1.9451908443312863e-06, 'learning_rate/action_model': 1.9451908443312865e-05, 'epoch': 2.02}) +[2026-06-18 04:51:34,495][starVLA.training.train_starvla][INFO] - Step 5241, Loss: {'train/loss': 0.3441646695137024, 'train/grad_norm_pre_clip': 15.283631324768066, 'timing/data': 0.03156725037842989, 'timing/model': 1.094205878674984, 'learning_rate/qwen_vl_interface': 1.943637759354604e-06, 'learning_rate/action_model': 1.9436377593546042e-05, 'epoch': 2.02}) +[2026-06-18 04:51:39,458][starVLA.training.train_starvla][INFO] - Step 5242, Loss: {'train/loss': 0.37348470091819763, 'train/grad_norm_pre_clip': 17.087825775146484, 'timing/data': 0.047553190030157566, 'timing/model': 1.5543816881254315, 'learning_rate/qwen_vl_interface': 1.942085359790373e-06, 'learning_rate/action_model': 1.942085359790373e-05, 'epoch': 2.02}) +[2026-06-18 04:51:43,879][starVLA.training.train_starvla][INFO] - Step 5243, Loss: {'train/loss': 0.1986042559146881, 'train/grad_norm_pre_clip': 25.60056495666504, 'timing/data': 0.03030944336205721, 'timing/model': 1.0880375560373068, 'learning_rate/qwen_vl_interface': 1.940533645960407e-06, 'learning_rate/action_model': 1.940533645960407e-05, 'epoch': 2.02}) +[2026-06-18 04:51:48,465][starVLA.training.train_starvla][INFO] - Step 5244, Loss: {'train/loss': 0.2931046783924103, 'train/grad_norm_pre_clip': 17.307918548583984, 'timing/data': 0.03332531452178955, 'timing/model': 1.0913262739777565, 'learning_rate/qwen_vl_interface': 1.9389826181863767e-06, 'learning_rate/action_model': 1.9389826181863767e-05, 'epoch': 2.02}) +[2026-06-18 04:51:54,286][starVLA.training.train_starvla][INFO] - Step 5245, Loss: {'train/loss': 0.11763821542263031, 'train/grad_norm_pre_clip': 7.174435615539551, 'timing/data': 0.03631345368921757, 'timing/model': 1.0731349159032106, 'learning_rate/qwen_vl_interface': 1.937432276789813e-06, 'learning_rate/action_model': 1.9374322767898127e-05, 'epoch': 2.02}) +[2026-06-18 04:51:58,834][starVLA.training.train_starvla][INFO] - Step 5246, Loss: {'train/loss': 0.11491013318300247, 'train/grad_norm_pre_clip': 6.7791876792907715, 'timing/data': 0.032290536910295486, 'timing/model': 1.0755968429148197, 'learning_rate/qwen_vl_interface': 1.935882622092103e-06, 'learning_rate/action_model': 1.935882622092103e-05, 'epoch': 2.02}) +[2026-06-18 04:52:03,536][starVLA.training.train_starvla][INFO] - Step 5247, Loss: {'train/loss': 0.29720908403396606, 'train/grad_norm_pre_clip': 13.642231941223145, 'timing/data': 0.037218693643808365, 'timing/model': 1.53803919441998, 'learning_rate/qwen_vl_interface': 1.9343336544144907e-06, 'learning_rate/action_model': 1.9343336544144906e-05, 'epoch': 2.02}) +[2026-06-18 04:52:11,161][starVLA.training.train_starvla][INFO] - Step 5248, Loss: {'train/loss': 0.1822209358215332, 'train/grad_norm_pre_clip': 8.731863975524902, 'timing/data': 0.04243909288197756, 'timing/model': 4.329523547552526, 'learning_rate/qwen_vl_interface': 1.9327853740780787e-06, 'learning_rate/action_model': 1.932785374078079e-05, 'epoch': 2.02}) +[2026-06-18 04:52:17,057][starVLA.training.train_starvla][INFO] - Step 5249, Loss: {'train/loss': 0.2952285706996918, 'train/grad_norm_pre_clip': 13.282933235168457, 'timing/data': 0.03291410859674215, 'timing/model': 1.1546336309984326, 'learning_rate/qwen_vl_interface': 1.9312377814038314e-06, 'learning_rate/action_model': 1.9312377814038312e-05, 'epoch': 2.02}) +[2026-06-18 04:52:22,062][starVLA.training.train_starvla][INFO] - Step 5250, Loss: {'train/loss': 0.21247032284736633, 'train/grad_norm_pre_clip': 10.214956283569336, 'timing/data': 0.03650937508791685, 'timing/model': 1.099722906947136, 'learning_rate/qwen_vl_interface': 1.9296908767125612e-06, 'learning_rate/action_model': 1.929690876712561e-05, 'epoch': 2.02}) +[2026-06-18 04:52:26,798][starVLA.training.train_starvla][INFO] - Step 5251, Loss: {'train/loss': 0.26171714067459106, 'train/grad_norm_pre_clip': 13.7853422164917, 'timing/data': 0.03039122000336647, 'timing/model': 1.116523484699428, 'learning_rate/qwen_vl_interface': 1.928144660324945e-06, 'learning_rate/action_model': 1.928144660324945e-05, 'epoch': 2.02}) +[2026-06-18 04:52:31,553][starVLA.training.train_starvla][INFO] - Step 5252, Loss: {'train/loss': 0.31623947620391846, 'train/grad_norm_pre_clip': 19.747739791870117, 'timing/data': 0.03970658779144287, 'timing/model': 1.5127831092104316, 'learning_rate/qwen_vl_interface': 1.9265991325615135e-06, 'learning_rate/action_model': 1.9265991325615135e-05, 'epoch': 2.02}) +[2026-06-18 04:52:35,991][starVLA.training.train_starvla][INFO] - Step 5253, Loss: {'train/loss': 0.3506200909614563, 'train/grad_norm_pre_clip': 14.4500150680542, 'timing/data': 0.030891322530806065, 'timing/model': 1.1796629121527076, 'learning_rate/qwen_vl_interface': 1.9250542937426604e-06, 'learning_rate/action_model': 1.9250542937426602e-05, 'epoch': 2.02}) +[2026-06-18 04:52:40,932][starVLA.training.train_starvla][INFO] - Step 5254, Loss: {'train/loss': 0.5663440823554993, 'train/grad_norm_pre_clip': 15.43641471862793, 'timing/data': 0.0358829265460372, 'timing/model': 1.1539381090551615, 'learning_rate/qwen_vl_interface': 1.9235101441886307e-06, 'learning_rate/action_model': 1.923510144188631e-05, 'epoch': 2.02}) +[2026-06-18 04:52:45,728][starVLA.training.train_starvla][INFO] - Step 5255, Loss: {'train/loss': 0.6911960244178772, 'train/grad_norm_pre_clip': 19.39358139038086, 'timing/data': 0.0383547218516469, 'timing/model': 1.1531572295352817, 'learning_rate/qwen_vl_interface': 1.921966684219525e-06, 'learning_rate/action_model': 1.921966684219525e-05, 'epoch': 2.02}) +[2026-06-18 04:52:51,373][starVLA.training.train_starvla][INFO] - Step 5256, Loss: {'train/loss': 0.2238437384366989, 'train/grad_norm_pre_clip': 10.557021141052246, 'timing/data': 0.04342201817780733, 'timing/model': 1.8826762177050114, 'learning_rate/qwen_vl_interface': 1.920423914155308e-06, 'learning_rate/action_model': 1.920423914155308e-05, 'epoch': 2.02}) +[2026-06-18 04:52:56,257][starVLA.training.train_starvla][INFO] - Step 5257, Loss: {'train/loss': 0.31183722615242004, 'train/grad_norm_pre_clip': 11.304553985595703, 'timing/data': 0.04273529537022114, 'timing/model': 1.5343189183622599, 'learning_rate/qwen_vl_interface': 1.9188818343157962e-06, 'learning_rate/action_model': 1.9188818343157964e-05, 'epoch': 2.02}) +[2026-06-18 04:53:01,779][starVLA.training.train_starvla][INFO] - Step 5258, Loss: {'train/loss': 0.6982889175415039, 'train/grad_norm_pre_clip': 21.16756820678711, 'timing/data': 0.036856538616120815, 'timing/model': 2.2829894898459315, 'learning_rate/qwen_vl_interface': 1.9173404450206654e-06, 'learning_rate/action_model': 1.9173404450206654e-05, 'epoch': 2.02}) +[2026-06-18 04:53:06,833][starVLA.training.train_starvla][INFO] - Step 5259, Loss: {'train/loss': 0.2539041340351105, 'train/grad_norm_pre_clip': 12.444901466369629, 'timing/data': 0.03062096145004034, 'timing/model': 1.1429291674867272, 'learning_rate/qwen_vl_interface': 1.9157997465894452e-06, 'learning_rate/action_model': 1.9157997465894454e-05, 'epoch': 2.02}) +[2026-06-18 04:53:11,624][starVLA.training.train_starvla][INFO] - Step 5260, Loss: {'train/loss': 0.4585016071796417, 'train/grad_norm_pre_clip': 17.294578552246094, 'timing/data': 0.03257037699222565, 'timing/model': 1.1089475667104125, 'learning_rate/qwen_vl_interface': 1.9142597393415264e-06, 'learning_rate/action_model': 1.914259739341526e-05, 'epoch': 2.02}) +[2026-06-18 04:53:16,462][starVLA.training.train_starvla][INFO] - Step 5261, Loss: {'train/loss': 0.4112147390842438, 'train/grad_norm_pre_clip': 12.67430305480957, 'timing/data': 0.03400855418294668, 'timing/model': 1.1668461039662361, 'learning_rate/qwen_vl_interface': 1.9127204235961524e-06, 'learning_rate/action_model': 1.9127204235961525e-05, 'epoch': 2.02}) +[2026-06-18 04:53:22,123][starVLA.training.train_starvla][INFO] - Step 5262, Loss: {'train/loss': 0.1714775264263153, 'train/grad_norm_pre_clip': 8.91318416595459, 'timing/data': 0.02764712553471327, 'timing/model': 1.5870173536241055, 'learning_rate/qwen_vl_interface': 1.911181799672426e-06, 'learning_rate/action_model': 1.9111817996724264e-05, 'epoch': 2.02}) +[2026-06-18 04:53:26,550][starVLA.training.train_starvla][INFO] - Step 5263, Loss: {'train/loss': 0.3807550370693207, 'train/grad_norm_pre_clip': 15.72531795501709, 'timing/data': 0.03235683962702751, 'timing/model': 1.0946545312181115, 'learning_rate/qwen_vl_interface': 1.909643867889305e-06, 'learning_rate/action_model': 1.909643867889305e-05, 'epoch': 2.02}) +[2026-06-18 04:53:31,238][starVLA.training.train_starvla][INFO] - Step 5264, Loss: {'train/loss': 0.47226905822753906, 'train/grad_norm_pre_clip': 91.16890716552734, 'timing/data': 0.045457491651177406, 'timing/model': 1.10734804533422, 'learning_rate/qwen_vl_interface': 1.908106628565605e-06, 'learning_rate/action_model': 1.9081066285656047e-05, 'epoch': 2.03}) +[2026-06-18 04:53:36,350][starVLA.training.train_starvla][INFO] - Step 5265, Loss: {'train/loss': 0.19524195790290833, 'train/grad_norm_pre_clip': 15.580145835876465, 'timing/data': 0.04890918079763651, 'timing/model': 1.487632249481976, 'learning_rate/qwen_vl_interface': 1.9065700820199965e-06, 'learning_rate/action_model': 1.9065700820199963e-05, 'epoch': 2.03}) +[2026-06-18 04:53:41,264][starVLA.training.train_starvla][INFO] - Step 5266, Loss: {'train/loss': 0.4091748297214508, 'train/grad_norm_pre_clip': 18.140853881835938, 'timing/data': 0.047382181510329247, 'timing/model': 1.150279887020588, 'learning_rate/qwen_vl_interface': 1.9050342285710063e-06, 'learning_rate/action_model': 1.9050342285710062e-05, 'epoch': 2.03}) +[2026-06-18 04:53:46,181][starVLA.training.train_starvla][INFO] - Step 5267, Loss: {'train/loss': 0.5584538578987122, 'train/grad_norm_pre_clip': 23.8912410736084, 'timing/data': 0.035390752367675304, 'timing/model': 1.5229923762381077, 'learning_rate/qwen_vl_interface': 1.9034990685370217e-06, 'learning_rate/action_model': 1.9034990685370217e-05, 'epoch': 2.03}) +[2026-06-18 04:53:50,829][starVLA.training.train_starvla][INFO] - Step 5268, Loss: {'train/loss': 0.2077038586139679, 'train/grad_norm_pre_clip': 14.96778678894043, 'timing/data': 0.03672411572188139, 'timing/model': 1.1044452907517552, 'learning_rate/qwen_vl_interface': 1.901964602236282e-06, 'learning_rate/action_model': 1.901964602236282e-05, 'epoch': 2.03}) +[2026-06-18 04:53:55,557][starVLA.training.train_starvla][INFO] - Step 5269, Loss: {'train/loss': 0.25681301951408386, 'train/grad_norm_pre_clip': 15.366631507873535, 'timing/data': 0.0311813922598958, 'timing/model': 1.1493304260075092, 'learning_rate/qwen_vl_interface': 1.900430829986882e-06, 'learning_rate/action_model': 1.900430829986882e-05, 'epoch': 2.03}) +[2026-06-18 04:54:00,349][starVLA.training.train_starvla][INFO] - Step 5270, Loss: {'train/loss': 0.4945690631866455, 'train/grad_norm_pre_clip': 15.299006462097168, 'timing/data': 0.04553695674985647, 'timing/model': 1.2177171586081386, 'learning_rate/qwen_vl_interface': 1.898897752106773e-06, 'learning_rate/action_model': 1.898897752106773e-05, 'epoch': 2.03}) +[2026-06-18 04:54:05,274][starVLA.training.train_starvla][INFO] - Step 5271, Loss: {'train/loss': 0.5872050523757935, 'train/grad_norm_pre_clip': 13.915637969970703, 'timing/data': 0.04442305024713278, 'timing/model': 1.2048011552542448, 'learning_rate/qwen_vl_interface': 1.8973653689137683e-06, 'learning_rate/action_model': 1.8973653689137684e-05, 'epoch': 2.03}) +[2026-06-18 04:54:10,129][starVLA.training.train_starvla][INFO] - Step 5272, Loss: {'train/loss': 0.5082209706306458, 'train/grad_norm_pre_clip': 16.126754760742188, 'timing/data': 0.039513117633759975, 'timing/model': 1.5022073769941926, 'learning_rate/qwen_vl_interface': 1.8958336807255298e-06, 'learning_rate/action_model': 1.8958336807255297e-05, 'epoch': 2.03}) +[2026-06-18 04:54:15,010][starVLA.training.train_starvla][INFO] - Step 5273, Loss: {'train/loss': 0.26845329999923706, 'train/grad_norm_pre_clip': 12.687816619873047, 'timing/data': 0.03844392206519842, 'timing/model': 1.2579791685566306, 'learning_rate/qwen_vl_interface': 1.8943026878595785e-06, 'learning_rate/action_model': 1.8943026878595786e-05, 'epoch': 2.03}) +[2026-06-18 04:54:19,988][starVLA.training.train_starvla][INFO] - Step 5274, Loss: {'train/loss': 0.28586116433143616, 'train/grad_norm_pre_clip': 12.380112648010254, 'timing/data': 0.0356834027916193, 'timing/model': 1.1857122657820582, 'learning_rate/qwen_vl_interface': 1.8927723906332913e-06, 'learning_rate/action_model': 1.8927723906332912e-05, 'epoch': 2.03}) +[2026-06-18 04:54:24,743][starVLA.training.train_starvla][INFO] - Step 5275, Loss: {'train/loss': 0.5137280821800232, 'train/grad_norm_pre_clip': 15.58688735961914, 'timing/data': 0.02962284255772829, 'timing/model': 1.1028558099642396, 'learning_rate/qwen_vl_interface': 1.8912427893638996e-06, 'learning_rate/action_model': 1.8912427893638996e-05, 'epoch': 2.03}) +[2026-06-18 04:54:30,092][starVLA.training.train_starvla][INFO] - Step 5276, Loss: {'train/loss': 0.37280070781707764, 'train/grad_norm_pre_clip': 14.96032428741455, 'timing/data': 0.05639089364558458, 'timing/model': 1.7177804987877607, 'learning_rate/qwen_vl_interface': 1.8897138843684918e-06, 'learning_rate/action_model': 1.8897138843684917e-05, 'epoch': 2.03}) +[2026-06-18 04:54:34,948][starVLA.training.train_starvla][INFO] - Step 5277, Loss: {'train/loss': 0.27039527893066406, 'train/grad_norm_pre_clip': 15.942851066589355, 'timing/data': 0.050666878931224346, 'timing/model': 1.4214660599827766, 'learning_rate/qwen_vl_interface': 1.8881856759640122e-06, 'learning_rate/action_model': 1.8881856759640122e-05, 'epoch': 2.03}) +[2026-06-18 04:54:39,379][starVLA.training.train_starvla][INFO] - Step 5278, Loss: {'train/loss': 0.28499409556388855, 'train/grad_norm_pre_clip': 23.626096725463867, 'timing/data': 0.03767411131411791, 'timing/model': 1.0948616042733192, 'learning_rate/qwen_vl_interface': 1.8866581644672589e-06, 'learning_rate/action_model': 1.8866581644672586e-05, 'epoch': 2.03}) +[2026-06-18 04:54:44,105][starVLA.training.train_starvla][INFO] - Step 5279, Loss: {'train/loss': 0.39296388626098633, 'train/grad_norm_pre_clip': 15.175219535827637, 'timing/data': 0.03744897339493036, 'timing/model': 1.0887102503329515, 'learning_rate/qwen_vl_interface': 1.8851313501948872e-06, 'learning_rate/action_model': 1.885131350194887e-05, 'epoch': 2.03}) +[2026-06-18 04:54:48,873][starVLA.training.train_starvla][INFO] - Step 5280, Loss: {'train/loss': 0.2386265993118286, 'train/grad_norm_pre_clip': 14.691180229187012, 'timing/data': 0.03208134975284338, 'timing/model': 1.1455824887380004, 'learning_rate/qwen_vl_interface': 1.8836052334634072e-06, 'learning_rate/action_model': 1.883605233463407e-05, 'epoch': 2.03}) +[2026-06-18 04:54:53,801][starVLA.training.train_starvla][INFO] - Step 5281, Loss: {'train/loss': 0.1579671949148178, 'train/grad_norm_pre_clip': 9.437267303466797, 'timing/data': 0.03518278896808624, 'timing/model': 1.2262307284399867, 'learning_rate/qwen_vl_interface': 1.8820798145891834e-06, 'learning_rate/action_model': 1.8820798145891832e-05, 'epoch': 2.03}) +[2026-06-18 04:54:59,468][starVLA.training.train_starvla][INFO] - Step 5282, Loss: {'train/loss': 0.2084817886352539, 'train/grad_norm_pre_clip': 8.61648941040039, 'timing/data': 0.034263454377651215, 'timing/model': 2.433713265694678, 'learning_rate/qwen_vl_interface': 1.88055509388844e-06, 'learning_rate/action_model': 1.88055509388844e-05, 'epoch': 2.03}) +[2026-06-18 04:55:04,246][starVLA.training.train_starvla][INFO] - Step 5283, Loss: {'train/loss': 0.3303399682044983, 'train/grad_norm_pre_clip': 11.205314636230469, 'timing/data': 0.03912076074630022, 'timing/model': 1.2615991299971938, 'learning_rate/qwen_vl_interface': 1.8790310716772525e-06, 'learning_rate/action_model': 1.8790310716772524e-05, 'epoch': 2.03}) +[2026-06-18 04:55:13,195][starVLA.training.train_starvla][INFO] - Step 5284, Loss: {'train/loss': 0.4378910958766937, 'train/grad_norm_pre_clip': 14.762872695922852, 'timing/data': 0.039693321101367474, 'timing/model': 1.1040812395513058, 'learning_rate/qwen_vl_interface': 1.8775077482715496e-06, 'learning_rate/action_model': 1.8775077482715494e-05, 'epoch': 2.03}) +[2026-06-18 04:55:17,976][starVLA.training.train_starvla][INFO] - Step 5285, Loss: {'train/loss': 0.593940794467926, 'train/grad_norm_pre_clip': 18.249753952026367, 'timing/data': 0.036620727740228176, 'timing/model': 1.1242210427299142, 'learning_rate/qwen_vl_interface': 1.8759851239871176e-06, 'learning_rate/action_model': 1.8759851239871175e-05, 'epoch': 2.03}) +[2026-06-18 04:55:22,707][starVLA.training.train_starvla][INFO] - Step 5286, Loss: {'train/loss': 0.6308366060256958, 'train/grad_norm_pre_clip': 23.413249969482422, 'timing/data': 0.04534166865050793, 'timing/model': 1.1085783364251256, 'learning_rate/qwen_vl_interface': 1.8744631991396023e-06, 'learning_rate/action_model': 1.8744631991396023e-05, 'epoch': 2.03}) +[2026-06-18 04:55:27,611][starVLA.training.train_starvla][INFO] - Step 5287, Loss: {'train/loss': 0.2016056478023529, 'train/grad_norm_pre_clip': 13.208589553833008, 'timing/data': 0.03429008275270462, 'timing/model': 1.56226974632591, 'learning_rate/qwen_vl_interface': 1.8729419740444977e-06, 'learning_rate/action_model': 1.8729419740444977e-05, 'epoch': 2.03}) +[2026-06-18 04:55:32,012][starVLA.training.train_starvla][INFO] - Step 5288, Loss: {'train/loss': 0.43881645798683167, 'train/grad_norm_pre_clip': 25.047143936157227, 'timing/data': 0.031052800826728344, 'timing/model': 1.1989165199920535, 'learning_rate/qwen_vl_interface': 1.8714214490171554e-06, 'learning_rate/action_model': 1.8714214490171552e-05, 'epoch': 2.03}) +[2026-06-18 04:55:36,835][starVLA.training.train_starvla][INFO] - Step 5289, Loss: {'train/loss': 0.15694621205329895, 'train/grad_norm_pre_clip': 13.770230293273926, 'timing/data': 0.04358487296849489, 'timing/model': 1.1694218339398503, 'learning_rate/qwen_vl_interface': 1.8699016243727826e-06, 'learning_rate/action_model': 1.8699016243727827e-05, 'epoch': 2.03}) +[2026-06-18 04:55:41,545][starVLA.training.train_starvla][INFO] - Step 5290, Loss: {'train/loss': 0.3011770248413086, 'train/grad_norm_pre_clip': 17.005640029907227, 'timing/data': 0.06454989779740572, 'timing/model': 1.1400714134797454, 'learning_rate/qwen_vl_interface': 1.8683825004264402e-06, 'learning_rate/action_model': 1.86838250042644e-05, 'epoch': 2.04}) +[2026-06-18 04:55:46,390][starVLA.training.train_starvla][INFO] - Step 5291, Loss: {'train/loss': 0.22689197957515717, 'train/grad_norm_pre_clip': 11.345934867858887, 'timing/data': 0.045298539102077484, 'timing/model': 1.0945849372074008, 'learning_rate/qwen_vl_interface': 1.8668640774930444e-06, 'learning_rate/action_model': 1.8668640774930442e-05, 'epoch': 2.04}) +[2026-06-18 04:55:51,098][starVLA.training.train_starvla][INFO] - Step 5292, Loss: {'train/loss': 0.24258944392204285, 'train/grad_norm_pre_clip': 9.550076484680176, 'timing/data': 0.049183244816958904, 'timing/model': 1.438571090810001, 'learning_rate/qwen_vl_interface': 1.8653463558873651e-06, 'learning_rate/action_model': 1.8653463558873652e-05, 'epoch': 2.04}) +[2026-06-18 04:55:58,277][starVLA.training.train_starvla][INFO] - Step 5293, Loss: {'train/loss': 0.2969537675380707, 'train/grad_norm_pre_clip': 12.96241569519043, 'timing/data': 0.03869644086807966, 'timing/model': 2.627409423701465, 'learning_rate/qwen_vl_interface': 1.863829335924029e-06, 'learning_rate/action_model': 1.8638293359240288e-05, 'epoch': 2.04}) +[2026-06-18 04:56:03,215][starVLA.training.train_starvla][INFO] - Step 5294, Loss: {'train/loss': 0.3578908443450928, 'train/grad_norm_pre_clip': 13.653571128845215, 'timing/data': 0.03977756109088659, 'timing/model': 1.1092941956594586, 'learning_rate/qwen_vl_interface': 1.8623130179175148e-06, 'learning_rate/action_model': 1.8623130179175146e-05, 'epoch': 2.04}) +[2026-06-18 04:56:08,128][starVLA.training.train_starvla][INFO] - Step 5295, Loss: {'train/loss': 0.39095357060432434, 'train/grad_norm_pre_clip': 14.439592361450195, 'timing/data': 0.05679764412343502, 'timing/model': 1.1089528519660234, 'learning_rate/qwen_vl_interface': 1.8607974021821573e-06, 'learning_rate/action_model': 1.8607974021821572e-05, 'epoch': 2.04}) +[2026-06-18 04:56:14,242][starVLA.training.train_starvla][INFO] - Step 5296, Loss: {'train/loss': 0.48617687821388245, 'train/grad_norm_pre_clip': 17.50439453125, 'timing/data': 0.036861520260572433, 'timing/model': 2.561659438535571, 'learning_rate/qwen_vl_interface': 1.8592824890321443e-06, 'learning_rate/action_model': 1.8592824890321442e-05, 'epoch': 2.04}) +[2026-06-18 04:56:18,997][starVLA.training.train_starvla][INFO] - Step 5297, Loss: {'train/loss': 0.19035494327545166, 'train/grad_norm_pre_clip': 8.598661422729492, 'timing/data': 0.03709505312144756, 'timing/model': 1.4726038742810488, 'learning_rate/qwen_vl_interface': 1.8577682787815223e-06, 'learning_rate/action_model': 1.8577682787815225e-05, 'epoch': 2.04}) +[2026-06-18 04:56:23,527][starVLA.training.train_starvla][INFO] - Step 5298, Loss: {'train/loss': 0.47965922951698303, 'train/grad_norm_pre_clip': 23.78173065185547, 'timing/data': 0.03853042237460613, 'timing/model': 1.0972114251926541, 'learning_rate/qwen_vl_interface': 1.856254771744185e-06, 'learning_rate/action_model': 1.856254771744185e-05, 'epoch': 2.04}) +[2026-06-18 04:56:28,171][starVLA.training.train_starvla][INFO] - Step 5299, Loss: {'train/loss': 0.43141090869903564, 'train/grad_norm_pre_clip': 21.690093994140625, 'timing/data': 0.03950766660273075, 'timing/model': 1.143071472644806, 'learning_rate/qwen_vl_interface': 1.8547419682338838e-06, 'learning_rate/action_model': 1.8547419682338836e-05, 'epoch': 2.04}) +[2026-06-18 04:59:54,925][starVLA.training.train_starvla][INFO] - Step 5300, Loss: {'train/loss': 0.21229754388332367, 'train/grad_norm_pre_clip': 12.802334785461426, 'eval/loss': 0.4246503710746765, 'eval/latency_4/loss': 0.4246503710746765, 'eval/demon_fire_precision': 0.960530470476792, 'eval/demon_fire_recall': 0.9879831113998051, 'eval/demon_fire_f1': 0.9740634005763689, 'eval/demon_fire_rate_model': 0.9301027900146843, 'eval/demon_fire_rate_teacher': 0.9042584434654919, 'eval/demon_left_precision': 0.8162650602409639, 'eval/demon_left_recall': 0.6187214611872146, 'eval/demon_left_f1': 0.703896103896104, 'eval/demon_left_rate_model': 0.09750367107195301, 'eval/demon_left_rate_teacher': 0.12863436123348018, 'eval/demon_right_precision': 0.811787072243346, 'eval/demon_right_recall': 0.7707581227436823, 'eval/demon_right_f1': 0.7907407407407407, 'eval/demon_right_rate_model': 0.15447870778267253, 'eval/demon_right_rate_teacher': 0.16270190895741557, 'eval/demon_move_f1': 0.7473184223184224, 'eval/demon_cc_f1': 0.8927978675625369, 'eval/latency_4/demon_fire_precision': 0.960530470476792, 'eval/latency_4/demon_fire_recall': 0.9879831113998051, 'eval/latency_4/demon_fire_f1': 0.9740634005763689, 'eval/latency_4/demon_fire_rate_model': 0.9301027900146843, 'eval/latency_4/demon_fire_rate_teacher': 0.9042584434654919, 'eval/latency_4/demon_left_precision': 0.8162650602409639, 'eval/latency_4/demon_left_recall': 0.6187214611872146, 'eval/latency_4/demon_left_f1': 0.703896103896104, 'eval/latency_4/demon_left_rate_model': 0.09750367107195301, 'eval/latency_4/demon_left_rate_teacher': 0.12863436123348018, 'eval/latency_4/demon_right_precision': 0.811787072243346, 'eval/latency_4/demon_right_recall': 0.7707581227436823, 'eval/latency_4/demon_right_f1': 0.7907407407407407, 'eval/latency_4/demon_right_rate_model': 0.15447870778267253, 'eval/latency_4/demon_right_rate_teacher': 0.16270190895741557, 'eval/latency_4/demon_move_f1': 0.7473184223184224, 'eval/latency_4/demon_cc_f1': 0.8927978675625369, 'timing/data': 0.04386690445244312, 'timing/model': 1.1351036028936505, 'learning_rate/qwen_vl_interface': 1.8532298685642275e-06, 'learning_rate/action_model': 1.8532298685642274e-05, 'epoch': 2.04}) +[2026-06-18 04:59:54,990][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_5300_state +[2026-06-18 04:59:54,999][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 05:00:05,070][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_5300_state/model.safetensors +[2026-06-18 05:00:30,747][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_5300_state/optimizer.bin +[2026-06-18 05:00:30,748][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_5300_state/sampler.bin +[2026-06-18 05:00:30,749][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_5300_state/sampler_1.bin +[2026-06-18 05:00:30,751][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_5300_state/random_states_0.pkl +[2026-06-18 05:00:30,757][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 5300 +[2026-06-18 05:00:30,758][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 05:00:30,817][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 05:00:37,641][starVLA.training.train_starvla][INFO] - Step 5301, Loss: {'train/loss': 0.5386031270027161, 'train/grad_norm_pre_clip': 18.781885147094727, 'timing/data': 0.04137334320694208, 'timing/model': 1.314817475154996, 'learning_rate/qwen_vl_interface': 1.8517184730486745e-06, 'learning_rate/action_model': 1.8517184730486746e-05, 'epoch': 2.04}) +[2026-06-18 05:00:42,666][starVLA.training.train_starvla][INFO] - Step 5302, Loss: {'train/loss': 0.19260689616203308, 'train/grad_norm_pre_clip': 9.775978088378906, 'timing/data': 0.038241940550506115, 'timing/model': 1.5105505734682083, 'learning_rate/qwen_vl_interface': 1.8502077820005408e-06, 'learning_rate/action_model': 1.8502077820005407e-05, 'epoch': 2.04}) +[2026-06-18 05:00:47,665][starVLA.training.train_starvla][INFO] - Step 5303, Loss: {'train/loss': 0.3561811149120331, 'train/grad_norm_pre_clip': 11.963643074035645, 'timing/data': 0.03058705013245344, 'timing/model': 1.2422095574438572, 'learning_rate/qwen_vl_interface': 1.8486977957329888e-06, 'learning_rate/action_model': 1.8486977957329886e-05, 'epoch': 2.04}) +[2026-06-18 05:00:53,284][starVLA.training.train_starvla][INFO] - Step 5304, Loss: {'train/loss': 0.29186293482780457, 'train/grad_norm_pre_clip': 12.042162895202637, 'timing/data': 0.04118799604475498, 'timing/model': 1.4274753788486123, 'learning_rate/qwen_vl_interface': 1.847188514559045e-06, 'learning_rate/action_model': 1.847188514559045e-05, 'epoch': 2.04}) +[2026-06-18 05:00:58,753][starVLA.training.train_starvla][INFO] - Step 5305, Loss: {'train/loss': 0.7067208886146545, 'train/grad_norm_pre_clip': 15.087370872497559, 'timing/data': 0.037886057049036026, 'timing/model': 1.191512712277472, 'learning_rate/qwen_vl_interface': 1.8456799387915839e-06, 'learning_rate/action_model': 1.8456799387915838e-05, 'epoch': 2.04}) +[2026-06-18 05:01:04,003][starVLA.training.train_starvla][INFO] - Step 5306, Loss: {'train/loss': 0.18296952545642853, 'train/grad_norm_pre_clip': 8.918745040893555, 'timing/data': 0.0673871748149395, 'timing/model': 1.2502540275454521, 'learning_rate/qwen_vl_interface': 1.8441720687433345e-06, 'learning_rate/action_model': 1.8441720687433344e-05, 'epoch': 2.04}) +[2026-06-18 05:01:09,407][starVLA.training.train_starvla][INFO] - Step 5307, Loss: {'train/loss': 0.26437801122665405, 'train/grad_norm_pre_clip': 13.163771629333496, 'timing/data': 0.032023114152252674, 'timing/model': 1.9143708134070039, 'learning_rate/qwen_vl_interface': 1.8426649047268808e-06, 'learning_rate/action_model': 1.8426649047268807e-05, 'epoch': 2.04}) +[2026-06-18 05:01:14,301][starVLA.training.train_starvla][INFO] - Step 5308, Loss: {'train/loss': 0.2252466380596161, 'train/grad_norm_pre_clip': 15.810633659362793, 'timing/data': 0.04855533316731453, 'timing/model': 1.3332423316314816, 'learning_rate/qwen_vl_interface': 1.8411584470546582e-06, 'learning_rate/action_model': 1.841158447054658e-05, 'epoch': 2.04}) +[2026-06-18 05:01:19,566][starVLA.training.train_starvla][INFO] - Step 5309, Loss: {'train/loss': 0.2643335461616516, 'train/grad_norm_pre_clip': 11.691274642944336, 'timing/data': 0.047333366237580776, 'timing/model': 1.2013469217345119, 'learning_rate/qwen_vl_interface': 1.8396526960389575e-06, 'learning_rate/action_model': 1.8396526960389575e-05, 'epoch': 2.04}) +[2026-06-18 05:01:24,564][starVLA.training.train_starvla][INFO] - Step 5310, Loss: {'train/loss': 0.3697201609611511, 'train/grad_norm_pre_clip': 13.668807983398438, 'timing/data': 0.0660110292956233, 'timing/model': 1.1484671048820019, 'learning_rate/qwen_vl_interface': 1.8381476519919215e-06, 'learning_rate/action_model': 1.8381476519919212e-05, 'epoch': 2.04}) +[2026-06-18 05:01:29,422][starVLA.training.train_starvla][INFO] - Step 5311, Loss: {'train/loss': 0.3597445487976074, 'train/grad_norm_pre_clip': 18.133403778076172, 'timing/data': 0.04965128470212221, 'timing/model': 1.1052324967458844, 'learning_rate/qwen_vl_interface': 1.8366433152255511e-06, 'learning_rate/action_model': 1.8366433152255512e-05, 'epoch': 2.04}) +[2026-06-18 05:01:34,471][starVLA.training.train_starvla][INFO] - Step 5312, Loss: {'train/loss': 0.3892160952091217, 'train/grad_norm_pre_clip': 13.618953704833984, 'timing/data': 0.052486727014184, 'timing/model': 1.4435598151758313, 'learning_rate/qwen_vl_interface': 1.8351396860516931e-06, 'learning_rate/action_model': 1.835139686051693e-05, 'epoch': 2.04}) +[2026-06-18 05:01:39,571][starVLA.training.train_starvla][INFO] - Step 5313, Loss: {'train/loss': 0.19424812495708466, 'train/grad_norm_pre_clip': 12.538683891296387, 'timing/data': 0.05714342184364796, 'timing/model': 1.345573098398745, 'learning_rate/qwen_vl_interface': 1.8336367647820527e-06, 'learning_rate/action_model': 1.8336367647820528e-05, 'epoch': 2.04}) +[2026-06-18 05:01:44,692][starVLA.training.train_starvla][INFO] - Step 5314, Loss: {'train/loss': 0.3746224641799927, 'train/grad_norm_pre_clip': 11.929548263549805, 'timing/data': 0.05191622767597437, 'timing/model': 1.149564211256802, 'learning_rate/qwen_vl_interface': 1.832134551728186e-06, 'learning_rate/action_model': 1.832134551728186e-05, 'epoch': 2.04}) +[2026-06-18 05:01:54,249][starVLA.training.train_starvla][INFO] - Step 5315, Loss: {'train/loss': 0.24900253117084503, 'train/grad_norm_pre_clip': 10.352622985839844, 'timing/data': 0.06337802577763796, 'timing/model': 2.427775197662413, 'learning_rate/qwen_vl_interface': 1.8306330472015051e-06, 'learning_rate/action_model': 1.8306330472015053e-05, 'epoch': 2.04}) +[2026-06-18 05:02:08,762][starVLA.training.train_starvla][INFO] - Step 5316, Loss: {'train/loss': 0.37321001291275024, 'train/grad_norm_pre_clip': 14.521051406860352, 'timing/data': 0.05070091784000397, 'timing/model': 5.568502499721944, 'learning_rate/qwen_vl_interface': 1.8291322515132757e-06, 'learning_rate/action_model': 1.8291322515132756e-05, 'epoch': 2.05}) +[2026-06-18 05:02:19,092][starVLA.training.train_starvla][INFO] - Step 5317, Loss: {'train/loss': 0.5899065732955933, 'train/grad_norm_pre_clip': 15.648991584777832, 'timing/data': 0.05462593026459217, 'timing/model': 3.4349207235500216, 'learning_rate/qwen_vl_interface': 1.8276321649746076e-06, 'learning_rate/action_model': 1.8276321649746076e-05, 'epoch': 2.05}) +[2026-06-18 05:02:33,661][starVLA.training.train_starvla][INFO] - Step 5318, Loss: {'train/loss': 0.3416305482387543, 'train/grad_norm_pre_clip': 21.655994415283203, 'timing/data': 0.11288137920200825, 'timing/model': 3.2030800320208073, 'learning_rate/qwen_vl_interface': 1.826132787896476e-06, 'learning_rate/action_model': 1.826132787896476e-05, 'epoch': 2.05}) +[2026-06-18 05:02:47,912][starVLA.training.train_starvla][INFO] - Step 5319, Loss: {'train/loss': 0.27853167057037354, 'train/grad_norm_pre_clip': 12.109156608581543, 'timing/data': 0.04076162539422512, 'timing/model': 4.914132772013545, 'learning_rate/qwen_vl_interface': 1.8246341205897018e-06, 'learning_rate/action_model': 1.8246341205897017e-05, 'epoch': 2.05}) +[2026-06-18 05:03:01,415][starVLA.training.train_starvla][INFO] - Step 5320, Loss: {'train/loss': 0.18640941381454468, 'train/grad_norm_pre_clip': 12.506534576416016, 'timing/data': 0.05760789941996336, 'timing/model': 2.9959637075662613, 'learning_rate/qwen_vl_interface': 1.8231361633649606e-06, 'learning_rate/action_model': 1.8231361633649606e-05, 'epoch': 2.05}) +[2026-06-18 05:03:18,841][starVLA.training.train_starvla][INFO] - Step 5321, Loss: {'train/loss': 0.21306392550468445, 'train/grad_norm_pre_clip': 18.847803115844727, 'timing/data': 0.0541897201910615, 'timing/model': 5.318960515782237, 'learning_rate/qwen_vl_interface': 1.8216389165327797e-06, 'learning_rate/action_model': 1.8216389165327798e-05, 'epoch': 2.05}) +[2026-06-18 05:03:28,115][starVLA.training.train_starvla][INFO] - Step 5322, Loss: {'train/loss': 0.5413442850112915, 'train/grad_norm_pre_clip': 20.2842960357666, 'timing/data': 0.1097544189542532, 'timing/model': 1.7623315528035164, 'learning_rate/qwen_vl_interface': 1.8201423804035406e-06, 'learning_rate/action_model': 1.8201423804035405e-05, 'epoch': 2.05}) +[2026-06-18 05:03:42,783][starVLA.training.train_starvla][INFO] - Step 5323, Loss: {'train/loss': 0.15038131177425385, 'train/grad_norm_pre_clip': 9.733006477355957, 'timing/data': 0.014428366906940937, 'timing/model': 4.244210271164775, 'learning_rate/qwen_vl_interface': 1.8186465552874769e-06, 'learning_rate/action_model': 1.8186465552874766e-05, 'epoch': 2.05}) +[2026-06-18 05:03:55,150][starVLA.training.train_starvla][INFO] - Step 5324, Loss: {'train/loss': 0.33615246415138245, 'train/grad_norm_pre_clip': 14.552931785583496, 'timing/data': 0.017004196532070637, 'timing/model': 1.3640771992504597, 'learning_rate/qwen_vl_interface': 1.8171514414946745e-06, 'learning_rate/action_model': 1.8171514414946745e-05, 'epoch': 2.05}) +[2026-06-18 05:04:01,945][starVLA.training.train_starvla][INFO] - Step 5325, Loss: {'train/loss': 0.3332242965698242, 'train/grad_norm_pre_clip': 14.914566993713379, 'timing/data': 0.01818770170211792, 'timing/model': 1.107467446476221, 'learning_rate/qwen_vl_interface': 1.815657039335072e-06, 'learning_rate/action_model': 1.815657039335072e-05, 'epoch': 2.05}) +[2026-06-18 05:04:07,489][starVLA.training.train_starvla][INFO] - Step 5326, Loss: {'train/loss': 0.3423025906085968, 'train/grad_norm_pre_clip': 13.491382598876953, 'timing/data': 0.011459793895483017, 'timing/model': 1.509966043755412, 'learning_rate/qwen_vl_interface': 1.8141633491184612e-06, 'learning_rate/action_model': 1.814163349118461e-05, 'epoch': 2.05}) +[2026-06-18 05:04:13,320][starVLA.training.train_starvla][INFO] - Step 5327, Loss: {'train/loss': 0.5081924200057983, 'train/grad_norm_pre_clip': 45.19728088378906, 'timing/data': 0.014503330923616886, 'timing/model': 1.7184443939477205, 'learning_rate/qwen_vl_interface': 1.8126703711544848e-06, 'learning_rate/action_model': 1.812670371154485e-05, 'epoch': 2.05}) +[2026-06-18 05:04:18,192][starVLA.training.train_starvla][INFO] - Step 5328, Loss: {'train/loss': 0.5358070135116577, 'train/grad_norm_pre_clip': 24.180368423461914, 'timing/data': 0.011745708994567394, 'timing/model': 1.5043719420209527, 'learning_rate/qwen_vl_interface': 1.8111781057526381e-06, 'learning_rate/action_model': 1.811178105752638e-05, 'epoch': 2.05}) +[2026-06-18 05:04:23,248][starVLA.training.train_starvla][INFO] - Step 5329, Loss: {'train/loss': 0.21277959644794464, 'train/grad_norm_pre_clip': 10.304158210754395, 'timing/data': 0.012059306725859642, 'timing/model': 1.096215351484716, 'learning_rate/qwen_vl_interface': 1.8096865532222714e-06, 'learning_rate/action_model': 1.8096865532222714e-05, 'epoch': 2.05}) +[2026-06-18 05:04:28,663][starVLA.training.train_starvla][INFO] - Step 5330, Loss: {'train/loss': 0.2686915993690491, 'train/grad_norm_pre_clip': 18.929182052612305, 'timing/data': 0.012914375402033329, 'timing/model': 1.262675704434514, 'learning_rate/qwen_vl_interface': 1.8081957138725861e-06, 'learning_rate/action_model': 1.808195713872586e-05, 'epoch': 2.05}) +[2026-06-18 05:04:37,133][starVLA.training.train_starvla][INFO] - Step 5331, Loss: {'train/loss': 0.35280174016952515, 'train/grad_norm_pre_clip': 13.76014232635498, 'timing/data': 0.021744628436863422, 'timing/model': 1.618362850509584, 'learning_rate/qwen_vl_interface': 1.8067055880126311e-06, 'learning_rate/action_model': 1.806705588012631e-05, 'epoch': 2.05}) +[2026-06-18 05:04:43,232][starVLA.training.train_starvla][INFO] - Step 5332, Loss: {'train/loss': 0.5813694596290588, 'train/grad_norm_pre_clip': 15.641448974609375, 'timing/data': 0.022001037374138832, 'timing/model': 2.116051304154098, 'learning_rate/qwen_vl_interface': 1.8052161759513114e-06, 'learning_rate/action_model': 1.8052161759513115e-05, 'epoch': 2.05}) +[2026-06-18 05:04:59,614][starVLA.training.train_starvla][INFO] - Step 5333, Loss: {'train/loss': 0.29784706234931946, 'train/grad_norm_pre_clip': 8.020853996276855, 'timing/data': 0.013513739220798016, 'timing/model': 5.655213092453778, 'learning_rate/qwen_vl_interface': 1.8037274779973868e-06, 'learning_rate/action_model': 1.8037274779973867e-05, 'epoch': 2.05}) +[2026-06-18 05:05:09,515][starVLA.training.train_starvla][INFO] - Step 5334, Loss: {'train/loss': 0.23366974294185638, 'train/grad_norm_pre_clip': 11.84514045715332, 'timing/data': 0.014954094775021076, 'timing/model': 1.4462029272690415, 'learning_rate/qwen_vl_interface': 1.8022394944594643e-06, 'learning_rate/action_model': 1.802239494459464e-05, 'epoch': 2.05}) +[2026-06-18 05:05:14,659][starVLA.training.train_starvla][INFO] - Step 5335, Loss: {'train/loss': 0.3438086211681366, 'train/grad_norm_pre_clip': 15.290786743164062, 'timing/data': 0.020084835588932037, 'timing/model': 1.1398688601329923, 'learning_rate/qwen_vl_interface': 1.8007522256460062e-06, 'learning_rate/action_model': 1.8007522256460062e-05, 'epoch': 2.05}) +[2026-06-18 05:05:20,159][starVLA.training.train_starvla][INFO] - Step 5336, Loss: {'train/loss': 0.41007980704307556, 'train/grad_norm_pre_clip': 15.49216365814209, 'timing/data': 0.012066550552845001, 'timing/model': 1.40242026373744, 'learning_rate/qwen_vl_interface': 1.7992656718653201e-06, 'learning_rate/action_model': 1.79926567186532e-05, 'epoch': 2.05}) +[2026-06-18 05:05:26,029][starVLA.training.train_starvla][INFO] - Step 5337, Loss: {'train/loss': 0.4632062017917633, 'train/grad_norm_pre_clip': 16.859073638916016, 'timing/data': 0.012913409620523453, 'timing/model': 1.757331581786275, 'learning_rate/qwen_vl_interface': 1.797779833425575e-06, 'learning_rate/action_model': 1.797779833425575e-05, 'epoch': 2.05}) +[2026-06-18 05:05:31,215][starVLA.training.train_starvla][INFO] - Step 5338, Loss: {'train/loss': 0.6012564897537231, 'train/grad_norm_pre_clip': 19.762571334838867, 'timing/data': 0.017712825909256935, 'timing/model': 1.0914210369810462, 'learning_rate/qwen_vl_interface': 1.7962947106347846e-06, 'learning_rate/action_model': 1.7962947106347845e-05, 'epoch': 2.05}) +[2026-06-18 05:05:37,921][starVLA.training.train_starvla][INFO] - Step 5339, Loss: {'train/loss': 0.3556319773197174, 'train/grad_norm_pre_clip': 15.669845581054688, 'timing/data': 0.01211217138916254, 'timing/model': 2.670953775756061, 'learning_rate/qwen_vl_interface': 1.794810303800818e-06, 'learning_rate/action_model': 1.794810303800818e-05, 'epoch': 2.05}) +[2026-06-18 05:05:46,632][starVLA.training.train_starvla][INFO] - Step 5340, Loss: {'train/loss': 0.2756139636039734, 'train/grad_norm_pre_clip': 15.581215858459473, 'timing/data': 0.021145650185644627, 'timing/model': 1.1996381916105747, 'learning_rate/qwen_vl_interface': 1.7933266132313924e-06, 'learning_rate/action_model': 1.7933266132313923e-05, 'epoch': 2.05}) +[2026-06-18 05:05:55,283][starVLA.training.train_starvla][INFO] - Step 5341, Loss: {'train/loss': 0.22145375609397888, 'train/grad_norm_pre_clip': 15.202641487121582, 'timing/data': 0.012464359402656555, 'timing/model': 1.7487286254763603, 'learning_rate/qwen_vl_interface': 1.7918436392340803e-06, 'learning_rate/action_model': 1.7918436392340802e-05, 'epoch': 2.05}) +[2026-06-18 05:06:00,786][starVLA.training.train_starvla][INFO] - Step 5342, Loss: {'train/loss': 0.1405985802412033, 'train/grad_norm_pre_clip': 7.513005256652832, 'timing/data': 0.013990623876452446, 'timing/model': 1.7099807495251298, 'learning_rate/qwen_vl_interface': 1.7903613821163022e-06, 'learning_rate/action_model': 1.790361382116302e-05, 'epoch': 2.06}) +[2026-06-18 05:06:06,136][starVLA.training.train_starvla][INFO] - Step 5343, Loss: {'train/loss': 0.3482184112071991, 'train/grad_norm_pre_clip': 15.313342094421387, 'timing/data': 0.011392748914659023, 'timing/model': 1.262333881109953, 'learning_rate/qwen_vl_interface': 1.7888798421853313e-06, 'learning_rate/action_model': 1.7888798421853313e-05, 'epoch': 2.06}) +[2026-06-18 05:06:12,993][starVLA.training.train_starvla][INFO] - Step 5344, Loss: {'train/loss': 0.3816249370574951, 'train/grad_norm_pre_clip': 14.6760835647583, 'timing/data': 0.024678039364516735, 'timing/model': 1.3335771234706044, 'learning_rate/qwen_vl_interface': 1.7873990197482962e-06, 'learning_rate/action_model': 1.787399019748296e-05, 'epoch': 2.06}) +[2026-06-18 05:06:21,374][starVLA.training.train_starvla][INFO] - Step 5345, Loss: {'train/loss': 0.43603140115737915, 'train/grad_norm_pre_clip': 15.064520835876465, 'timing/data': 0.011633140034973621, 'timing/model': 1.3818695722147822, 'learning_rate/qwen_vl_interface': 1.785918915112169e-06, 'learning_rate/action_model': 1.785918915112169e-05, 'epoch': 2.06}) +[2026-06-18 05:06:30,665][starVLA.training.train_starvla][INFO] - Step 5346, Loss: {'train/loss': 0.7780349850654602, 'train/grad_norm_pre_clip': 24.24942398071289, 'timing/data': 0.014266560785472393, 'timing/model': 3.0025664502754807, 'learning_rate/qwen_vl_interface': 1.7844395285837773e-06, 'learning_rate/action_model': 1.7844395285837774e-05, 'epoch': 2.06}) +[2026-06-18 05:06:44,633][starVLA.training.train_starvla][INFO] - Step 5347, Loss: {'train/loss': 0.18064787983894348, 'train/grad_norm_pre_clip': 9.221529960632324, 'timing/data': 0.011805010959506035, 'timing/model': 1.938805889338255, 'learning_rate/qwen_vl_interface': 1.7829608604698003e-06, 'learning_rate/action_model': 1.7829608604698e-05, 'epoch': 2.06}) +[2026-06-18 05:06:49,587][starVLA.training.train_starvla][INFO] - Step 5348, Loss: {'train/loss': 0.549881637096405, 'train/grad_norm_pre_clip': 15.507613182067871, 'timing/data': 0.01364707387983799, 'timing/model': 1.2231129063293338, 'learning_rate/qwen_vl_interface': 1.7814829110767685e-06, 'learning_rate/action_model': 1.7814829110767685e-05, 'epoch': 2.06}) +[2026-06-18 05:06:57,319][starVLA.training.train_starvla][INFO] - Step 5349, Loss: {'train/loss': 0.3607243597507477, 'train/grad_norm_pre_clip': 14.557765007019043, 'timing/data': 0.029834925197064877, 'timing/model': 2.797462562099099, 'learning_rate/qwen_vl_interface': 1.7800056807110627e-06, 'learning_rate/action_model': 1.7800056807110626e-05, 'epoch': 2.06}) +[2026-06-18 05:07:02,806][starVLA.training.train_starvla][INFO] - Step 5350, Loss: {'train/loss': 0.2567535638809204, 'train/grad_norm_pre_clip': 12.43366813659668, 'timing/data': 0.013859315775334835, 'timing/model': 1.3459708113223314, 'learning_rate/qwen_vl_interface': 1.7785291696789103e-06, 'learning_rate/action_model': 1.7785291696789102e-05, 'epoch': 2.06}) +[2026-06-18 05:07:14,941][starVLA.training.train_starvla][INFO] - Step 5351, Loss: {'train/loss': 0.47132816910743713, 'train/grad_norm_pre_clip': 14.789562225341797, 'timing/data': 0.013283242471516132, 'timing/model': 4.87794401217252, 'learning_rate/qwen_vl_interface': 1.7770533782863983e-06, 'learning_rate/action_model': 1.7770533782863983e-05, 'epoch': 2.06}) +[2026-06-18 05:07:20,237][starVLA.training.train_starvla][INFO] - Step 5352, Loss: {'train/loss': 0.4641939699649811, 'train/grad_norm_pre_clip': 24.74234962463379, 'timing/data': 0.015339150093495846, 'timing/model': 1.6103381467983127, 'learning_rate/qwen_vl_interface': 1.7755783068394578e-06, 'learning_rate/action_model': 1.7755783068394577e-05, 'epoch': 2.06}) +[2026-06-18 05:07:25,895][starVLA.training.train_starvla][INFO] - Step 5353, Loss: {'train/loss': 0.5315724015235901, 'train/grad_norm_pre_clip': 15.77995777130127, 'timing/data': 0.01929464191198349, 'timing/model': 1.513101807795465, 'learning_rate/qwen_vl_interface': 1.7741039556438733e-06, 'learning_rate/action_model': 1.7741039556438733e-05, 'epoch': 2.06}) +[2026-06-18 05:07:31,045][starVLA.training.train_starvla][INFO] - Step 5354, Loss: {'train/loss': 0.5893508195877075, 'train/grad_norm_pre_clip': 17.9459285736084, 'timing/data': 0.016245005652308464, 'timing/model': 1.2306376853957772, 'learning_rate/qwen_vl_interface': 1.7726303250052785e-06, 'learning_rate/action_model': 1.7726303250052784e-05, 'epoch': 2.06}) +[2026-06-18 05:07:38,309][starVLA.training.train_starvla][INFO] - Step 5355, Loss: {'train/loss': 0.30733954906463623, 'train/grad_norm_pre_clip': 11.830107688903809, 'timing/data': 0.01414676383137703, 'timing/model': 1.0927322460338473, 'learning_rate/qwen_vl_interface': 1.7711574152291591e-06, 'learning_rate/action_model': 1.771157415229159e-05, 'epoch': 2.06}) +[2026-06-18 05:07:45,250][starVLA.training.train_starvla][INFO] - Step 5356, Loss: {'train/loss': 0.10646748542785645, 'train/grad_norm_pre_clip': 9.838342666625977, 'timing/data': 0.014155425131320953, 'timing/model': 2.568880637176335, 'learning_rate/qwen_vl_interface': 1.7696852266208512e-06, 'learning_rate/action_model': 1.7696852266208512e-05, 'epoch': 2.06}) +[2026-06-18 05:07:58,401][starVLA.training.train_starvla][INFO] - Step 5357, Loss: {'train/loss': 0.17510411143302917, 'train/grad_norm_pre_clip': 10.490665435791016, 'timing/data': 0.015344166196882725, 'timing/model': 1.9055437091737986, 'learning_rate/qwen_vl_interface': 1.7682137594855408e-06, 'learning_rate/action_model': 1.768213759485541e-05, 'epoch': 2.06}) +[2026-06-18 05:08:07,180][starVLA.training.train_starvla][INFO] - Step 5358, Loss: {'train/loss': 0.3039987087249756, 'train/grad_norm_pre_clip': 17.125654220581055, 'timing/data': 0.015065331943333149, 'timing/model': 1.3051419844850898, 'learning_rate/qwen_vl_interface': 1.7667430141282649e-06, 'learning_rate/action_model': 1.7667430141282648e-05, 'epoch': 2.06}) +[2026-06-18 05:08:12,595][starVLA.training.train_starvla][INFO] - Step 5359, Loss: {'train/loss': 0.46533870697021484, 'train/grad_norm_pre_clip': 23.206817626953125, 'timing/data': 0.013947388157248497, 'timing/model': 1.225065934471786, 'learning_rate/qwen_vl_interface': 1.7652729908539101e-06, 'learning_rate/action_model': 1.76527299085391e-05, 'epoch': 2.06}) +[2026-06-18 05:08:18,447][starVLA.training.train_starvla][INFO] - Step 5360, Loss: {'train/loss': 0.3496398329734802, 'train/grad_norm_pre_clip': 17.71502113342285, 'timing/data': 0.013001589104533195, 'timing/model': 1.5657374504953623, 'learning_rate/qwen_vl_interface': 1.7638036899672137e-06, 'learning_rate/action_model': 1.7638036899672136e-05, 'epoch': 2.06}) +[2026-06-18 05:08:28,374][starVLA.training.train_starvla][INFO] - Step 5361, Loss: {'train/loss': 0.5769941210746765, 'train/grad_norm_pre_clip': 20.28716278076172, 'timing/data': 0.05669638514518738, 'timing/model': 1.6617277348414063, 'learning_rate/qwen_vl_interface': 1.7623351117727628e-06, 'learning_rate/action_model': 1.7623351117727626e-05, 'epoch': 2.06}) +[2026-06-18 05:08:33,646][starVLA.training.train_starvla][INFO] - Step 5362, Loss: {'train/loss': 0.25456929206848145, 'train/grad_norm_pre_clip': 16.57927131652832, 'timing/data': 0.014629228040575981, 'timing/model': 1.416025536134839, 'learning_rate/qwen_vl_interface': 1.7608672565749973e-06, 'learning_rate/action_model': 1.760867256574997e-05, 'epoch': 2.06}) +[2026-06-18 05:08:41,649][starVLA.training.train_starvla][INFO] - Step 5363, Loss: {'train/loss': 0.2443130612373352, 'train/grad_norm_pre_clip': 9.802569389343262, 'timing/data': 0.016378755681216717, 'timing/model': 1.2894331058487296, 'learning_rate/qwen_vl_interface': 1.7594001246782039e-06, 'learning_rate/action_model': 1.7594001246782037e-05, 'epoch': 2.06}) +[2026-06-18 05:08:47,612][starVLA.training.train_starvla][INFO] - Step 5364, Loss: {'train/loss': 0.3555912971496582, 'train/grad_norm_pre_clip': 13.60560131072998, 'timing/data': 0.020701371133327484, 'timing/model': 1.1762511478736997, 'learning_rate/qwen_vl_interface': 1.7579337163865221e-06, 'learning_rate/action_model': 1.757933716386522e-05, 'epoch': 2.06}) +[2026-06-18 05:08:53,039][starVLA.training.train_starvla][INFO] - Step 5365, Loss: {'train/loss': 0.36453142762184143, 'train/grad_norm_pre_clip': 13.091499328613281, 'timing/data': 0.01752692274749279, 'timing/model': 1.245822299271822, 'learning_rate/qwen_vl_interface': 1.7564680320039356e-06, 'learning_rate/action_model': 1.7564680320039358e-05, 'epoch': 2.06}) +[2026-06-18 05:08:59,259][starVLA.training.train_starvla][INFO] - Step 5366, Loss: {'train/loss': 0.644633412361145, 'train/grad_norm_pre_clip': 22.44048309326172, 'timing/data': 0.01375120971351862, 'timing/model': 2.1830305978655815, 'learning_rate/qwen_vl_interface': 1.7550030718342867e-06, 'learning_rate/action_model': 1.7550030718342867e-05, 'epoch': 2.06}) +[2026-06-18 05:09:04,727][starVLA.training.train_starvla][INFO] - Step 5367, Loss: {'train/loss': 0.617577314376831, 'train/grad_norm_pre_clip': 21.60834312438965, 'timing/data': 0.011801629327237606, 'timing/model': 1.6091283597052097, 'learning_rate/qwen_vl_interface': 1.753538836181262e-06, 'learning_rate/action_model': 1.753538836181262e-05, 'epoch': 2.06}) +[2026-06-18 05:09:10,010][starVLA.training.train_starvla][INFO] - Step 5368, Loss: {'train/loss': 0.2968933880329132, 'train/grad_norm_pre_clip': 22.132314682006836, 'timing/data': 0.011790215969085693, 'timing/model': 1.5074452171102166, 'learning_rate/qwen_vl_interface': 1.7520753253483987e-06, 'learning_rate/action_model': 1.7520753253483986e-05, 'epoch': 2.07}) +[2026-06-18 05:09:16,847][starVLA.training.train_starvla][INFO] - Step 5369, Loss: {'train/loss': 0.14452746510505676, 'train/grad_norm_pre_clip': 8.02384090423584, 'timing/data': 0.011900044046342373, 'timing/model': 2.6469953283667564, 'learning_rate/qwen_vl_interface': 1.7506125396390847e-06, 'learning_rate/action_model': 1.7506125396390847e-05, 'epoch': 2.07}) +[2026-06-18 05:09:29,361][starVLA.training.train_starvla][INFO] - Step 5370, Loss: {'train/loss': 0.12966765463352203, 'train/grad_norm_pre_clip': 8.554137229919434, 'timing/data': 0.014740526676177979, 'timing/model': 1.5860369950532913, 'learning_rate/qwen_vl_interface': 1.7491504793565564e-06, 'learning_rate/action_model': 1.7491504793565562e-05, 'epoch': 2.07}) +[2026-06-18 05:09:35,944][starVLA.training.train_starvla][INFO] - Step 5371, Loss: {'train/loss': 0.21916905045509338, 'train/grad_norm_pre_clip': 11.110283851623535, 'timing/data': 0.017249833792448044, 'timing/model': 1.7565472107380629, 'learning_rate/qwen_vl_interface': 1.7476891448039012e-06, 'learning_rate/action_model': 1.7476891448039013e-05, 'epoch': 2.07}) +[2026-06-18 05:09:41,428][starVLA.training.train_starvla][INFO] - Step 5372, Loss: {'train/loss': 0.34887468814849854, 'train/grad_norm_pre_clip': 17.09207534790039, 'timing/data': 0.017798871733248234, 'timing/model': 1.5223540076985955, 'learning_rate/qwen_vl_interface': 1.7462285362840534e-06, 'learning_rate/action_model': 1.746228536284053e-05, 'epoch': 2.07}) +[2026-06-18 05:09:46,383][starVLA.training.train_starvla][INFO] - Step 5373, Loss: {'train/loss': 0.23314721882343292, 'train/grad_norm_pre_clip': 14.30738639831543, 'timing/data': 0.01629447191953659, 'timing/model': 1.473571702837944, 'learning_rate/qwen_vl_interface': 1.7447686540998027e-06, 'learning_rate/action_model': 1.7447686540998026e-05, 'epoch': 2.07}) +[2026-06-18 05:09:52,089][starVLA.training.train_starvla][INFO] - Step 5374, Loss: {'train/loss': 0.20476587116718292, 'train/grad_norm_pre_clip': 7.0450029373168945, 'timing/data': 0.012235164642333984, 'timing/model': 1.3242895975708961, 'learning_rate/qwen_vl_interface': 1.743309498553781e-06, 'learning_rate/action_model': 1.743309498553781e-05, 'epoch': 2.07}) +[2026-06-18 05:10:02,718][starVLA.training.train_starvla][INFO] - Step 5375, Loss: {'train/loss': 0.3029947876930237, 'train/grad_norm_pre_clip': 10.54503059387207, 'timing/data': 0.015047968365252018, 'timing/model': 1.9397203903645277, 'learning_rate/qwen_vl_interface': 1.7418510699484728e-06, 'learning_rate/action_model': 1.7418510699484727e-05, 'epoch': 2.07}) +[2026-06-18 05:10:08,896][starVLA.training.train_starvla][INFO] - Step 5376, Loss: {'train/loss': 0.24591730535030365, 'train/grad_norm_pre_clip': 12.079062461853027, 'timing/data': 0.012453930452466011, 'timing/model': 1.8989969259127975, 'learning_rate/qwen_vl_interface': 1.7403933685862118e-06, 'learning_rate/action_model': 1.7403933685862118e-05, 'epoch': 2.07}) +[2026-06-18 05:10:14,472][starVLA.training.train_starvla][INFO] - Step 5377, Loss: {'train/loss': 0.07980158179998398, 'train/grad_norm_pre_clip': 6.273655891418457, 'timing/data': 0.013451348058879375, 'timing/model': 1.5244318572804332, 'learning_rate/qwen_vl_interface': 1.7389363947691837e-06, 'learning_rate/action_model': 1.7389363947691838e-05, 'epoch': 2.07}) +[2026-06-18 05:10:23,379][starVLA.training.train_starvla][INFO] - Step 5378, Loss: {'train/loss': 0.1855577677488327, 'train/grad_norm_pre_clip': 10.830015182495117, 'timing/data': 0.026874166913330555, 'timing/model': 2.1691261883825064, 'learning_rate/qwen_vl_interface': 1.7374801487994198e-06, 'learning_rate/action_model': 1.7374801487994198e-05, 'epoch': 2.07}) +[2026-06-18 05:10:29,055][starVLA.training.train_starvla][INFO] - Step 5379, Loss: {'train/loss': 0.24445688724517822, 'train/grad_norm_pre_clip': 12.483376502990723, 'timing/data': 0.011952649801969528, 'timing/model': 1.1264520166441798, 'learning_rate/qwen_vl_interface': 1.7360246309787982e-06, 'learning_rate/action_model': 1.736024630978798e-05, 'epoch': 2.07}) +[2026-06-18 05:10:40,391][starVLA.training.train_starvla][INFO] - Step 5380, Loss: {'train/loss': 0.46040835976600647, 'train/grad_norm_pre_clip': 18.775297164916992, 'timing/data': 0.013331259600818157, 'timing/model': 1.135803853161633, 'learning_rate/qwen_vl_interface': 1.7345698416090534e-06, 'learning_rate/action_model': 1.7345698416090533e-05, 'epoch': 2.07}) +[2026-06-18 05:10:45,448][starVLA.training.train_starvla][INFO] - Step 5381, Loss: {'train/loss': 0.3060075640678406, 'train/grad_norm_pre_clip': 12.712728500366211, 'timing/data': 0.014007623307406902, 'timing/model': 1.1262918561697006, 'learning_rate/qwen_vl_interface': 1.7331157809917632e-06, 'learning_rate/action_model': 1.733115780991763e-05, 'epoch': 2.07}) +[2026-06-18 05:10:51,455][starVLA.training.train_starvla][INFO] - Step 5382, Loss: {'train/loss': 0.3848143517971039, 'train/grad_norm_pre_clip': 16.478412628173828, 'timing/data': 0.013694005087018013, 'timing/model': 1.6391142262145877, 'learning_rate/qwen_vl_interface': 1.7316624494283557e-06, 'learning_rate/action_model': 1.7316624494283557e-05, 'epoch': 2.07}) +[2026-06-18 05:10:55,952][starVLA.training.train_starvla][INFO] - Step 5383, Loss: {'train/loss': 0.27768009901046753, 'train/grad_norm_pre_clip': 12.182683944702148, 'timing/data': 0.018017137423157692, 'timing/model': 1.1149777378886938, 'learning_rate/qwen_vl_interface': 1.7302098472201085e-06, 'learning_rate/action_model': 1.7302098472201084e-05, 'epoch': 2.07}) +[2026-06-18 05:11:04,610][starVLA.training.train_starvla][INFO] - Step 5384, Loss: {'train/loss': 0.19215506315231323, 'train/grad_norm_pre_clip': 9.608088493347168, 'timing/data': 0.012477309443056583, 'timing/model': 1.1505344482138753, 'learning_rate/qwen_vl_interface': 1.728757974668147e-06, 'learning_rate/action_model': 1.728757974668147e-05, 'epoch': 2.07}) +[2026-06-18 05:11:10,255][starVLA.training.train_starvla][INFO] - Step 5385, Loss: {'train/loss': 0.36745786666870117, 'train/grad_norm_pre_clip': 13.503914833068848, 'timing/data': 0.012173832394182682, 'timing/model': 1.2453300580382347, 'learning_rate/qwen_vl_interface': 1.727306832073447e-06, 'learning_rate/action_model': 1.7273068320734472e-05, 'epoch': 2.07}) +[2026-06-18 05:11:16,479][starVLA.training.train_starvla][INFO] - Step 5386, Loss: {'train/loss': 0.18394789099693298, 'train/grad_norm_pre_clip': 9.980426788330078, 'timing/data': 0.015076170675456524, 'timing/model': 1.4905185280367732, 'learning_rate/qwen_vl_interface': 1.7258564197368308e-06, 'learning_rate/action_model': 1.725856419736831e-05, 'epoch': 2.07}) +[2026-06-18 05:11:29,561][starVLA.training.train_starvla][INFO] - Step 5387, Loss: {'train/loss': 0.6298116445541382, 'train/grad_norm_pre_clip': 15.566709518432617, 'timing/data': 0.01592942513525486, 'timing/model': 6.022445220500231, 'learning_rate/qwen_vl_interface': 1.7244067379589708e-06, 'learning_rate/action_model': 1.7244067379589707e-05, 'epoch': 2.07}) +[2026-06-18 05:11:35,715][starVLA.training.train_starvla][INFO] - Step 5388, Loss: {'train/loss': 0.32868364453315735, 'train/grad_norm_pre_clip': 11.6531982421875, 'timing/data': 0.014798775315284729, 'timing/model': 1.3535377029329538, 'learning_rate/qwen_vl_interface': 1.7229577870403886e-06, 'learning_rate/action_model': 1.7229577870403886e-05, 'epoch': 2.07}) +[2026-06-18 05:11:40,687][starVLA.training.train_starvla][INFO] - Step 5389, Loss: {'train/loss': 0.2258090078830719, 'train/grad_norm_pre_clip': 15.960237503051758, 'timing/data': 0.012171057052910328, 'timing/model': 1.197837763465941, 'learning_rate/qwen_vl_interface': 1.7215095672814517e-06, 'learning_rate/action_model': 1.7215095672814516e-05, 'epoch': 2.07}) +[2026-06-18 05:11:46,212][starVLA.training.train_starvla][INFO] - Step 5390, Loss: {'train/loss': 0.37254300713539124, 'train/grad_norm_pre_clip': 14.530426979064941, 'timing/data': 0.01814191695302725, 'timing/model': 1.3582546468824148, 'learning_rate/qwen_vl_interface': 1.7200620789823769e-06, 'learning_rate/action_model': 1.7200620789823767e-05, 'epoch': 2.07}) +[2026-06-18 05:11:51,509][starVLA.training.train_starvla][INFO] - Step 5391, Loss: {'train/loss': 0.21350771188735962, 'train/grad_norm_pre_clip': 10.846304893493652, 'timing/data': 0.011715000495314598, 'timing/model': 1.1578190168365836, 'learning_rate/qwen_vl_interface': 1.7186153224432331e-06, 'learning_rate/action_model': 1.718615322443233e-05, 'epoch': 2.07}) +[2026-06-18 05:11:57,845][starVLA.training.train_starvla][INFO] - Step 5392, Loss: {'train/loss': 0.6578439474105835, 'train/grad_norm_pre_clip': 20.071439743041992, 'timing/data': 0.014724933542311192, 'timing/model': 2.8837788430973887, 'learning_rate/qwen_vl_interface': 1.7171692979639342e-06, 'learning_rate/action_model': 1.717169297963934e-05, 'epoch': 2.07}) +[2026-06-18 05:12:02,559][starVLA.training.train_starvla][INFO] - Step 5393, Loss: {'train/loss': 0.2077326476573944, 'train/grad_norm_pre_clip': 17.683006286621094, 'timing/data': 0.017144489102065563, 'timing/model': 1.1018741261214018, 'learning_rate/qwen_vl_interface': 1.7157240058442397e-06, 'learning_rate/action_model': 1.71572400584424e-05, 'epoch': 2.07}) +[2026-06-18 05:12:08,479][starVLA.training.train_starvla][INFO] - Step 5394, Loss: {'train/loss': 0.49850013852119446, 'train/grad_norm_pre_clip': 16.049436569213867, 'timing/data': 0.01174938678741455, 'timing/model': 1.3199091013520956, 'learning_rate/qwen_vl_interface': 1.7142794463837608e-06, 'learning_rate/action_model': 1.714279446383761e-05, 'epoch': 2.08}) +[2026-06-18 05:12:13,543][starVLA.training.train_starvla][INFO] - Step 5395, Loss: {'train/loss': 0.4383852779865265, 'train/grad_norm_pre_clip': 17.32605743408203, 'timing/data': 0.0141883734613657, 'timing/model': 1.1127055659890175, 'learning_rate/qwen_vl_interface': 1.712835619881959e-06, 'learning_rate/action_model': 1.7128356198819588e-05, 'epoch': 2.08}) +[2026-06-18 05:12:19,105][starVLA.training.train_starvla][INFO] - Step 5396, Loss: {'train/loss': 0.38968372344970703, 'train/grad_norm_pre_clip': 15.543127059936523, 'timing/data': 0.015492061153054237, 'timing/model': 1.2915545105934143, 'learning_rate/qwen_vl_interface': 1.7113925266381383e-06, 'learning_rate/action_model': 1.7113925266381383e-05, 'epoch': 2.08}) +[2026-06-18 05:12:23,813][starVLA.training.train_starvla][INFO] - Step 5397, Loss: {'train/loss': 0.19680586457252502, 'train/grad_norm_pre_clip': 11.72708511352539, 'timing/data': 0.014709913171827793, 'timing/model': 1.2176900496706367, 'learning_rate/qwen_vl_interface': 1.7099501669514569e-06, 'learning_rate/action_model': 1.7099501669514568e-05, 'epoch': 2.08}) +[2026-06-18 05:12:30,206][starVLA.training.train_starvla][INFO] - Step 5398, Loss: {'train/loss': 0.2730327844619751, 'train/grad_norm_pre_clip': 13.643244743347168, 'timing/data': 0.024187302216887474, 'timing/model': 1.1732969945296645, 'learning_rate/qwen_vl_interface': 1.7085085411209118e-06, 'learning_rate/action_model': 1.7085085411209118e-05, 'epoch': 2.08}) +[2026-06-18 05:12:35,376][starVLA.training.train_starvla][INFO] - Step 5399, Loss: {'train/loss': 0.33280524611473083, 'train/grad_norm_pre_clip': 13.26150894165039, 'timing/data': 0.013542014174163342, 'timing/model': 1.3811088986694813, 'learning_rate/qwen_vl_interface': 1.7070676494453587e-06, 'learning_rate/action_model': 1.7070676494453586e-05, 'epoch': 2.08}) +[2026-06-18 05:17:51,839][starVLA.training.train_starvla][INFO] - Step 5400, Loss: {'train/loss': 0.2855318784713745, 'train/grad_norm_pre_clip': 14.611127853393555, 'eval/loss': 0.42696845531463623, 'eval/latency_4/loss': 0.42696845531463623, 'eval/demon_fire_precision': 0.9650571791613723, 'eval/demon_fire_recall': 0.9866839883078922, 'eval/demon_fire_f1': 0.9757507628071302, 'eval/demon_fire_rate_model': 0.9245227606461086, 'eval/demon_fire_rate_teacher': 0.9042584434654919, 'eval/demon_left_precision': 0.8042328042328042, 'eval/demon_left_recall': 0.6940639269406392, 'eval/demon_left_f1': 0.7450980392156863, 'eval/demon_left_rate_model': 0.11101321585903083, 'eval/demon_left_rate_teacher': 0.12863436123348018, 'eval/demon_right_precision': 0.8469601677148847, 'eval/demon_right_recall': 0.7292418772563177, 'eval/demon_right_f1': 0.7837051406401552, 'eval/demon_right_rate_model': 0.14008810572687225, 'eval/demon_right_rate_teacher': 0.16270190895741557, 'eval/demon_move_f1': 0.7644015899279207, 'eval/demon_cc_f1': 0.9010144001178091, 'eval/latency_4/demon_fire_precision': 0.9650571791613723, 'eval/latency_4/demon_fire_recall': 0.9866839883078922, 'eval/latency_4/demon_fire_f1': 0.9757507628071302, 'eval/latency_4/demon_fire_rate_model': 0.9245227606461086, 'eval/latency_4/demon_fire_rate_teacher': 0.9042584434654919, 'eval/latency_4/demon_left_precision': 0.8042328042328042, 'eval/latency_4/demon_left_recall': 0.6940639269406392, 'eval/latency_4/demon_left_f1': 0.7450980392156863, 'eval/latency_4/demon_left_rate_model': 0.11101321585903083, 'eval/latency_4/demon_left_rate_teacher': 0.12863436123348018, 'eval/latency_4/demon_right_precision': 0.8469601677148847, 'eval/latency_4/demon_right_recall': 0.7292418772563177, 'eval/latency_4/demon_right_f1': 0.7837051406401552, 'eval/latency_4/demon_right_rate_model': 0.14008810572687225, 'eval/latency_4/demon_right_rate_teacher': 0.16270190895741557, 'eval/latency_4/demon_move_f1': 0.7644015899279207, 'eval/latency_4/demon_cc_f1': 0.9010144001178091, 'timing/data': 0.012113725766539574, 'timing/model': 1.2073475793004036, 'learning_rate/qwen_vl_interface': 1.705627492223494e-06, 'learning_rate/action_model': 1.705627492223494e-05, 'epoch': 2.08}) +[2026-06-18 05:17:51,908][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_5400_state +[2026-06-18 05:17:51,917][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 05:18:00,800][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_5400_state/model.safetensors +[2026-06-18 05:18:23,772][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_5400_state/optimizer.bin +[2026-06-18 05:18:23,773][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_5400_state/sampler.bin +[2026-06-18 05:18:23,774][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_5400_state/sampler_1.bin +[2026-06-18 05:18:23,776][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_5400_state/random_states_0.pkl +[2026-06-18 05:18:23,781][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 5400 +[2026-06-18 05:18:23,781][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 05:18:23,798][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 05:18:29,295][starVLA.training.train_starvla][INFO] - Step 5401, Loss: {'train/loss': 0.3061757981777191, 'train/grad_norm_pre_clip': 15.20877456665039, 'timing/data': 0.021138707175850868, 'timing/model': 1.6770993554964662, 'learning_rate/qwen_vl_interface': 1.7041880697538644e-06, 'learning_rate/action_model': 1.704188069753864e-05, 'epoch': 2.08}) +[2026-06-18 05:18:34,225][starVLA.training.train_starvla][INFO] - Step 5402, Loss: {'train/loss': 0.5561486482620239, 'train/grad_norm_pre_clip': 17.642684936523438, 'timing/data': 0.013817801140248775, 'timing/model': 1.1641953233629465, 'learning_rate/qwen_vl_interface': 1.702749382334863e-06, 'learning_rate/action_model': 1.7027493823348632e-05, 'epoch': 2.08}) +[2026-06-18 05:18:39,754][starVLA.training.train_starvla][INFO] - Step 5403, Loss: {'train/loss': 0.3895319998264313, 'train/grad_norm_pre_clip': 15.386208534240723, 'timing/data': 0.014007993042469025, 'timing/model': 1.4048617966473103, 'learning_rate/qwen_vl_interface': 1.7013114302647318e-06, 'learning_rate/action_model': 1.7013114302647318e-05, 'epoch': 2.08}) +[2026-06-18 05:18:51,319][starVLA.training.train_starvla][INFO] - Step 5404, Loss: {'train/loss': 0.4835106432437897, 'train/grad_norm_pre_clip': 16.368940353393555, 'timing/data': 0.015414617024362087, 'timing/model': 2.526264443062246, 'learning_rate/qwen_vl_interface': 1.69987421384156e-06, 'learning_rate/action_model': 1.69987421384156e-05, 'epoch': 2.08}) +[2026-06-18 05:19:08,498][starVLA.training.train_starvla][INFO] - Step 5405, Loss: {'train/loss': 0.3896295130252838, 'train/grad_norm_pre_clip': 14.257914543151855, 'timing/data': 0.012036262080073357, 'timing/model': 4.320455865934491, 'learning_rate/qwen_vl_interface': 1.6984377333632815e-06, 'learning_rate/action_model': 1.6984377333632813e-05, 'epoch': 2.08}) +[2026-06-18 05:19:20,763][starVLA.training.train_starvla][INFO] - Step 5406, Loss: {'train/loss': 0.4735685884952545, 'train/grad_norm_pre_clip': 15.493353843688965, 'timing/data': 0.012499880976974964, 'timing/model': 5.722779701463878, 'learning_rate/qwen_vl_interface': 1.697001989127686e-06, 'learning_rate/action_model': 1.697001989127686e-05, 'epoch': 2.08}) +[2026-06-18 05:19:27,651][starVLA.training.train_starvla][INFO] - Step 5407, Loss: {'train/loss': 0.2956549823284149, 'train/grad_norm_pre_clip': 181.7922821044922, 'timing/data': 0.011902092024683952, 'timing/model': 1.6231991592794657, 'learning_rate/qwen_vl_interface': 1.6955669814324003e-06, 'learning_rate/action_model': 1.6955669814324e-05, 'epoch': 2.08}) +[2026-06-18 05:19:35,650][starVLA.training.train_starvla][INFO] - Step 5408, Loss: {'train/loss': 0.14727352559566498, 'train/grad_norm_pre_clip': 22.82200050354004, 'timing/data': 0.01266245823353529, 'timing/model': 1.2719286922365427, 'learning_rate/qwen_vl_interface': 1.6941327105749034e-06, 'learning_rate/action_model': 1.694132710574903e-05, 'epoch': 2.08}) +[2026-06-18 05:19:45,063][starVLA.training.train_starvla][INFO] - Step 5409, Loss: {'train/loss': 0.349119633436203, 'train/grad_norm_pre_clip': 15.232966423034668, 'timing/data': 0.020010417327284813, 'timing/model': 4.139656906016171, 'learning_rate/qwen_vl_interface': 1.6926991768525205e-06, 'learning_rate/action_model': 1.6926991768525205e-05, 'epoch': 2.08}) +[2026-06-18 05:19:53,377][starVLA.training.train_starvla][INFO] - Step 5410, Loss: {'train/loss': 0.24223922193050385, 'train/grad_norm_pre_clip': 10.950570106506348, 'timing/data': 0.010811761021614075, 'timing/model': 1.1505413670092821, 'learning_rate/qwen_vl_interface': 1.6912663805624278e-06, 'learning_rate/action_model': 1.6912663805624275e-05, 'epoch': 2.08}) +[2026-06-18 05:19:59,556][starVLA.training.train_starvla][INFO] - Step 5411, Loss: {'train/loss': 0.41410714387893677, 'train/grad_norm_pre_clip': 12.391913414001465, 'timing/data': 0.013506694696843624, 'timing/model': 1.2217761697247624, 'learning_rate/qwen_vl_interface': 1.6898343220016446e-06, 'learning_rate/action_model': 1.6898343220016447e-05, 'epoch': 2.08}) +[2026-06-18 05:20:05,203][starVLA.training.train_starvla][INFO] - Step 5412, Loss: {'train/loss': 0.6052286624908447, 'train/grad_norm_pre_clip': 20.86407470703125, 'timing/data': 0.011449028737843037, 'timing/model': 1.4908962799236178, 'learning_rate/qwen_vl_interface': 1.6884030014670353e-06, 'learning_rate/action_model': 1.688403001467035e-05, 'epoch': 2.08}) +[2026-06-18 05:20:10,087][starVLA.training.train_starvla][INFO] - Step 5413, Loss: {'train/loss': 0.2556874454021454, 'train/grad_norm_pre_clip': 16.45314598083496, 'timing/data': 0.014013489708304405, 'timing/model': 1.5135208312422037, 'learning_rate/qwen_vl_interface': 1.6869724192553168e-06, 'learning_rate/action_model': 1.686972419255317e-05, 'epoch': 2.08}) +[2026-06-18 05:20:15,030][starVLA.training.train_starvla][INFO] - Step 5414, Loss: {'train/loss': 0.4015495181083679, 'train/grad_norm_pre_clip': 18.08151626586914, 'timing/data': 0.012900063768029213, 'timing/model': 1.130382132716477, 'learning_rate/qwen_vl_interface': 1.6855425756630506e-06, 'learning_rate/action_model': 1.6855425756630506e-05, 'epoch': 2.08}) +[2026-06-18 05:20:20,353][starVLA.training.train_starvla][INFO] - Step 5415, Loss: {'train/loss': 0.566531240940094, 'train/grad_norm_pre_clip': 19.19989585876465, 'timing/data': 0.011115455999970436, 'timing/model': 1.3436728175729513, 'learning_rate/qwen_vl_interface': 1.684113470986644e-06, 'learning_rate/action_model': 1.6841134709866438e-05, 'epoch': 2.08}) +[2026-06-18 05:20:25,766][starVLA.training.train_starvla][INFO] - Step 5416, Loss: {'train/loss': 0.41425251960754395, 'train/grad_norm_pre_clip': 16.031639099121094, 'timing/data': 0.012328147888183594, 'timing/model': 1.353982382453978, 'learning_rate/qwen_vl_interface': 1.682685105522351e-06, 'learning_rate/action_model': 1.682685105522351e-05, 'epoch': 2.08}) +[2026-06-18 05:20:34,855][starVLA.training.train_starvla][INFO] - Step 5417, Loss: {'train/loss': 0.14275851845741272, 'train/grad_norm_pre_clip': 14.604236602783203, 'timing/data': 0.011555260978639126, 'timing/model': 1.1482314867898822, 'learning_rate/qwen_vl_interface': 1.6812574795662751e-06, 'learning_rate/action_model': 1.681257479566275e-05, 'epoch': 2.08}) +[2026-06-18 05:20:41,054][starVLA.training.train_starvla][INFO] - Step 5418, Loss: {'train/loss': 0.40330642461776733, 'train/grad_norm_pre_clip': 15.663115501403809, 'timing/data': 0.012070994824171066, 'timing/model': 1.3163843844085932, 'learning_rate/qwen_vl_interface': 1.6798305934143635e-06, 'learning_rate/action_model': 1.6798305934143633e-05, 'epoch': 2.08}) +[2026-06-18 05:20:49,970][starVLA.training.train_starvla][INFO] - Step 5419, Loss: {'train/loss': 0.27897369861602783, 'train/grad_norm_pre_clip': 11.81253719329834, 'timing/data': 0.01623152196407318, 'timing/model': 4.307267134077847, 'learning_rate/qwen_vl_interface': 1.6784044473624118e-06, 'learning_rate/action_model': 1.6784044473624118e-05, 'epoch': 2.08}) +[2026-06-18 05:21:02,858][starVLA.training.train_starvla][INFO] - Step 5420, Loss: {'train/loss': 0.2075464129447937, 'train/grad_norm_pre_clip': 11.848332405090332, 'timing/data': 0.012222550809383392, 'timing/model': 1.1020965036004782, 'learning_rate/qwen_vl_interface': 1.6769790417060611e-06, 'learning_rate/action_model': 1.676979041706061e-05, 'epoch': 2.09}) +[2026-06-18 05:21:08,547][starVLA.training.train_starvla][INFO] - Step 5421, Loss: {'train/loss': 0.11660542339086533, 'train/grad_norm_pre_clip': 7.651913166046143, 'timing/data': 0.01707810629159212, 'timing/model': 1.5362317524850368, 'learning_rate/qwen_vl_interface': 1.6755543767407998e-06, 'learning_rate/action_model': 1.6755543767407998e-05, 'epoch': 2.09}) +[2026-06-18 05:21:14,186][starVLA.training.train_starvla][INFO] - Step 5422, Loss: {'train/loss': 0.3566907048225403, 'train/grad_norm_pre_clip': 15.4722900390625, 'timing/data': 0.01122659258544445, 'timing/model': 1.7797680627554655, 'learning_rate/qwen_vl_interface': 1.6741304527619624e-06, 'learning_rate/action_model': 1.674130452761962e-05, 'epoch': 2.09}) +[2026-06-18 05:21:21,547][starVLA.training.train_starvla][INFO] - Step 5423, Loss: {'train/loss': 0.26649007201194763, 'train/grad_norm_pre_clip': 24.991594314575195, 'timing/data': 0.01294123288244009, 'timing/model': 2.4268682692199945, 'learning_rate/qwen_vl_interface': 1.672707270064729e-06, 'learning_rate/action_model': 1.6727072700647288e-05, 'epoch': 2.09}) +[2026-06-18 05:21:30,520][starVLA.training.train_starvla][INFO] - Step 5424, Loss: {'train/loss': 0.14507442712783813, 'train/grad_norm_pre_clip': 8.927497863769531, 'timing/data': 0.011730327270925045, 'timing/model': 1.2545609502121806, 'learning_rate/qwen_vl_interface': 1.6712848289441288e-06, 'learning_rate/action_model': 1.6712848289441287e-05, 'epoch': 2.09}) +[2026-06-18 05:21:35,444][starVLA.training.train_starvla][INFO] - Step 5425, Loss: {'train/loss': 0.4364175796508789, 'train/grad_norm_pre_clip': 12.507400512695312, 'timing/data': 0.013503245078027248, 'timing/model': 1.1674810759723186, 'learning_rate/qwen_vl_interface': 1.669863129695036e-06, 'learning_rate/action_model': 1.669863129695036e-05, 'epoch': 2.09}) +[2026-06-18 05:21:40,573][starVLA.training.train_starvla][INFO] - Step 5426, Loss: {'train/loss': 0.35183826088905334, 'train/grad_norm_pre_clip': 13.679052352905273, 'timing/data': 0.012657816521823406, 'timing/model': 1.5436889976263046, 'learning_rate/qwen_vl_interface': 1.6684421726121675e-06, 'learning_rate/action_model': 1.6684421726121674e-05, 'epoch': 2.09}) +[2026-06-18 05:21:45,495][starVLA.training.train_starvla][INFO] - Step 5427, Loss: {'train/loss': 0.42496609687805176, 'train/grad_norm_pre_clip': 15.171953201293945, 'timing/data': 0.014588570222258568, 'timing/model': 1.6074963817372918, 'learning_rate/qwen_vl_interface': 1.6670219579900888e-06, 'learning_rate/action_model': 1.6670219579900887e-05, 'epoch': 2.09}) +[2026-06-18 05:21:50,676][starVLA.training.train_starvla][INFO] - Step 5428, Loss: {'train/loss': 0.35942307114601135, 'train/grad_norm_pre_clip': 15.361478805541992, 'timing/data': 0.01160283200442791, 'timing/model': 1.1967825563624501, 'learning_rate/qwen_vl_interface': 1.665602486123215e-06, 'learning_rate/action_model': 1.665602486123215e-05, 'epoch': 2.09}) +[2026-06-18 05:22:00,197][starVLA.training.train_starvla][INFO] - Step 5429, Loss: {'train/loss': 0.4047923982143402, 'train/grad_norm_pre_clip': 19.99454689025879, 'timing/data': 0.01590721122920513, 'timing/model': 1.2085136715322733, 'learning_rate/qwen_vl_interface': 1.6641837573058034e-06, 'learning_rate/action_model': 1.6641837573058035e-05, 'epoch': 2.09}) +[2026-06-18 05:22:05,733][starVLA.training.train_starvla][INFO] - Step 5430, Loss: {'train/loss': 0.5019978284835815, 'train/grad_norm_pre_clip': 18.00652503967285, 'timing/data': 0.02012817934155464, 'timing/model': 1.0947804860770702, 'learning_rate/qwen_vl_interface': 1.6627657718319568e-06, 'learning_rate/action_model': 1.6627657718319568e-05, 'epoch': 2.09}) +[2026-06-18 05:22:10,771][starVLA.training.train_starvla][INFO] - Step 5431, Loss: {'train/loss': 0.18590892851352692, 'train/grad_norm_pre_clip': 8.404694557189941, 'timing/data': 0.012088106013834476, 'timing/model': 1.1418035989627242, 'learning_rate/qwen_vl_interface': 1.6613485299956265e-06, 'learning_rate/action_model': 1.6613485299956262e-05, 'epoch': 2.09}) +[2026-06-18 05:22:15,629][starVLA.training.train_starvla][INFO] - Step 5432, Loss: {'train/loss': 0.22423452138900757, 'train/grad_norm_pre_clip': 19.005525588989258, 'timing/data': 0.021209645085036755, 'timing/model': 1.4809990227222443, 'learning_rate/qwen_vl_interface': 1.6599320320906064e-06, 'learning_rate/action_model': 1.6599320320906066e-05, 'epoch': 2.09}) +[2026-06-18 05:22:23,249][starVLA.training.train_starvla][INFO] - Step 5433, Loss: {'train/loss': 0.23748062551021576, 'train/grad_norm_pre_clip': 11.057290077209473, 'timing/data': 0.014735984615981579, 'timing/model': 1.1124828355386853, 'learning_rate/qwen_vl_interface': 1.65851627841054e-06, 'learning_rate/action_model': 1.65851627841054e-05, 'epoch': 2.09}) +[2026-06-18 05:22:28,454][starVLA.training.train_starvla][INFO] - Step 5434, Loss: {'train/loss': 0.13965831696987152, 'train/grad_norm_pre_clip': 7.360960006713867, 'timing/data': 0.013832518830895424, 'timing/model': 1.3196804728358984, 'learning_rate/qwen_vl_interface': 1.657101269248913e-06, 'learning_rate/action_model': 1.657101269248913e-05, 'epoch': 2.09}) +[2026-06-18 05:22:35,241][starVLA.training.train_starvla][INFO] - Step 5435, Loss: {'train/loss': 0.2808367609977722, 'train/grad_norm_pre_clip': 18.396644592285156, 'timing/data': 0.012499209493398666, 'timing/model': 1.5525824101641774, 'learning_rate/qwen_vl_interface': 1.65568700489906e-06, 'learning_rate/action_model': 1.65568700489906e-05, 'epoch': 2.09}) +[2026-06-18 05:22:40,612][starVLA.training.train_starvla][INFO] - Step 5436, Loss: {'train/loss': 0.09403014928102493, 'train/grad_norm_pre_clip': 4.863845348358154, 'timing/data': 0.02965129166841507, 'timing/model': 1.421659261919558, 'learning_rate/qwen_vl_interface': 1.6542734856541577e-06, 'learning_rate/action_model': 1.6542734856541576e-05, 'epoch': 2.09}) +[2026-06-18 05:22:51,957][starVLA.training.train_starvla][INFO] - Step 5437, Loss: {'train/loss': 0.36285218596458435, 'train/grad_norm_pre_clip': 12.765572547912598, 'timing/data': 0.013180376030504704, 'timing/model': 1.8791706915944815, 'learning_rate/qwen_vl_interface': 1.652860711807231e-06, 'learning_rate/action_model': 1.652860711807231e-05, 'epoch': 2.09}) +[2026-06-18 05:22:57,514][starVLA.training.train_starvla][INFO] - Step 5438, Loss: {'train/loss': 0.08279057592153549, 'train/grad_norm_pre_clip': 7.351780891418457, 'timing/data': 0.019159593619406223, 'timing/model': 1.8028229651972651, 'learning_rate/qwen_vl_interface': 1.6514486836511484e-06, 'learning_rate/action_model': 1.6514486836511484e-05, 'epoch': 2.09}) +[2026-06-18 05:23:03,806][starVLA.training.train_starvla][INFO] - Step 5439, Loss: {'train/loss': 0.24938935041427612, 'train/grad_norm_pre_clip': 15.516490936279297, 'timing/data': 0.010765104554593563, 'timing/model': 2.7177560236305, 'learning_rate/qwen_vl_interface': 1.6500374014786288e-06, 'learning_rate/action_model': 1.650037401478629e-05, 'epoch': 2.09}) +[2026-06-18 05:23:09,627][starVLA.training.train_starvla][INFO] - Step 5440, Loss: {'train/loss': 0.2167256623506546, 'train/grad_norm_pre_clip': 12.748540878295898, 'timing/data': 0.016597184352576733, 'timing/model': 1.5354115664958954, 'learning_rate/qwen_vl_interface': 1.6486268655822277e-06, 'learning_rate/action_model': 1.6486268655822277e-05, 'epoch': 2.09}) +[2026-06-18 05:23:15,253][starVLA.training.train_starvla][INFO] - Step 5441, Loss: {'train/loss': 0.3684525489807129, 'train/grad_norm_pre_clip': 26.0361270904541, 'timing/data': 0.014188443310558796, 'timing/model': 1.902490809559822, 'learning_rate/qwen_vl_interface': 1.6472170762543515e-06, 'learning_rate/action_model': 1.6472170762543514e-05, 'epoch': 2.09}) +[2026-06-18 05:23:20,386][starVLA.training.train_starvla][INFO] - Step 5442, Loss: {'train/loss': 0.2758602201938629, 'train/grad_norm_pre_clip': 15.765162467956543, 'timing/data': 0.02478845790028572, 'timing/model': 1.5702900970354676, 'learning_rate/qwen_vl_interface': 1.6458080337872534e-06, 'learning_rate/action_model': 1.6458080337872535e-05, 'epoch': 2.09}) +[2026-06-18 05:23:31,364][starVLA.training.train_starvla][INFO] - Step 5443, Loss: {'train/loss': 0.463499516248703, 'train/grad_norm_pre_clip': 21.431644439697266, 'timing/data': 0.0192001610994339, 'timing/model': 4.380055382847786, 'learning_rate/qwen_vl_interface': 1.644399738473029e-06, 'learning_rate/action_model': 1.644399738473029e-05, 'epoch': 2.09}) +[2026-06-18 05:23:36,813][starVLA.training.train_starvla][INFO] - Step 5444, Loss: {'train/loss': 0.3882197141647339, 'train/grad_norm_pre_clip': 14.314708709716797, 'timing/data': 0.014786867424845695, 'timing/model': 1.106464283540845, 'learning_rate/qwen_vl_interface': 1.6429921906036175e-06, 'learning_rate/action_model': 1.6429921906036176e-05, 'epoch': 2.09}) +[2026-06-18 05:23:45,142][starVLA.training.train_starvla][INFO] - Step 5445, Loss: {'train/loss': 0.33986279368400574, 'train/grad_norm_pre_clip': 13.843019485473633, 'timing/data': 0.011065958999097347, 'timing/model': 1.1592255029827356, 'learning_rate/qwen_vl_interface': 1.6415853904708068e-06, 'learning_rate/action_model': 1.6415853904708066e-05, 'epoch': 2.09}) +[2026-06-18 05:23:50,468][starVLA.training.train_starvla][INFO] - Step 5446, Loss: {'train/loss': 0.2538644075393677, 'train/grad_norm_pre_clip': 11.14896011352539, 'timing/data': 0.01997776050120592, 'timing/model': 1.3674030229449272, 'learning_rate/qwen_vl_interface': 1.6401793383662274e-06, 'learning_rate/action_model': 1.6401793383662273e-05, 'epoch': 2.1}) +[2026-06-18 05:23:55,788][starVLA.training.train_starvla][INFO] - Step 5447, Loss: {'train/loss': 0.405361533164978, 'train/grad_norm_pre_clip': 14.191490173339844, 'timing/data': 0.01208841148763895, 'timing/model': 1.5340132070705295, 'learning_rate/qwen_vl_interface': 1.6387740345813553e-06, 'learning_rate/action_model': 1.6387740345813552e-05, 'epoch': 2.1}) +[2026-06-18 05:24:00,377][starVLA.training.train_starvla][INFO] - Step 5448, Loss: {'train/loss': 0.27536046504974365, 'train/grad_norm_pre_clip': 15.569689750671387, 'timing/data': 0.011516233906149864, 'timing/model': 1.3045194959267974, 'learning_rate/qwen_vl_interface': 1.6373694794075115e-06, 'learning_rate/action_model': 1.6373694794075115e-05, 'epoch': 2.1}) +[2026-06-18 05:24:05,892][starVLA.training.train_starvla][INFO] - Step 5449, Loss: {'train/loss': 0.2116205245256424, 'train/grad_norm_pre_clip': 18.061925888061523, 'timing/data': 0.016243314370512962, 'timing/model': 1.2251280825585127, 'learning_rate/qwen_vl_interface': 1.6359656731358626e-06, 'learning_rate/action_model': 1.6359656731358626e-05, 'epoch': 2.1}) +[2026-06-18 05:24:10,799][starVLA.training.train_starvla][INFO] - Step 5450, Loss: {'train/loss': 0.6105428338050842, 'train/grad_norm_pre_clip': 16.44631004333496, 'timing/data': 0.01175383199006319, 'timing/model': 1.0691999308764935, 'learning_rate/qwen_vl_interface': 1.634562616057418e-06, 'learning_rate/action_model': 1.634562616057418e-05, 'epoch': 2.1}) +[2026-06-18 05:24:15,737][starVLA.training.train_starvla][INFO] - Step 5451, Loss: {'train/loss': 0.30364492535591125, 'train/grad_norm_pre_clip': 13.868022918701172, 'timing/data': 0.01161669846624136, 'timing/model': 1.0884801652282476, 'learning_rate/qwen_vl_interface': 1.6331603084630338e-06, 'learning_rate/action_model': 1.6331603084630338e-05, 'epoch': 2.1}) +[2026-06-18 05:24:25,816][starVLA.training.train_starvla][INFO] - Step 5452, Loss: {'train/loss': 0.17657946050167084, 'train/grad_norm_pre_clip': 13.015762329101562, 'timing/data': 0.018984174355864525, 'timing/model': 4.860883022658527, 'learning_rate/qwen_vl_interface': 1.6317587506434076e-06, 'learning_rate/action_model': 1.6317587506434076e-05, 'epoch': 2.1}) +[2026-06-18 05:24:31,417][starVLA.training.train_starvla][INFO] - Step 5453, Loss: {'train/loss': 0.2186160385608673, 'train/grad_norm_pre_clip': 10.965224266052246, 'timing/data': 0.015694337897002697, 'timing/model': 1.4432616718113422, 'learning_rate/qwen_vl_interface': 1.6303579428890881e-06, 'learning_rate/action_model': 1.630357942889088e-05, 'epoch': 2.1}) +[2026-06-18 05:24:36,386][starVLA.training.train_starvla][INFO] - Step 5454, Loss: {'train/loss': 0.07933070510625839, 'train/grad_norm_pre_clip': 7.368721008300781, 'timing/data': 0.0147364167496562, 'timing/model': 1.1422689482569695, 'learning_rate/qwen_vl_interface': 1.6289578854904626e-06, 'learning_rate/action_model': 1.6289578854904625e-05, 'epoch': 2.1}) +[2026-06-18 05:24:43,763][starVLA.training.train_starvla][INFO] - Step 5455, Loss: {'train/loss': 0.33482712507247925, 'train/grad_norm_pre_clip': 18.170087814331055, 'timing/data': 0.019129651598632336, 'timing/model': 1.5722350999712944, 'learning_rate/qwen_vl_interface': 1.6275585787377627e-06, 'learning_rate/action_model': 1.6275585787377626e-05, 'epoch': 2.1}) +[2026-06-18 05:24:52,807][starVLA.training.train_starvla][INFO] - Step 5456, Loss: {'train/loss': 0.18523012101650238, 'train/grad_norm_pre_clip': 13.142790794372559, 'timing/data': 0.028502040542662144, 'timing/model': 4.900841385126114, 'learning_rate/qwen_vl_interface': 1.6261600229210662e-06, 'learning_rate/action_model': 1.626160022921066e-05, 'epoch': 2.1}) +[2026-06-18 05:24:58,647][starVLA.training.train_starvla][INFO] - Step 5457, Loss: {'train/loss': 0.5113692283630371, 'train/grad_norm_pre_clip': 27.518354415893555, 'timing/data': 0.017997628077864647, 'timing/model': 1.857236891053617, 'learning_rate/qwen_vl_interface': 1.6247622183302972e-06, 'learning_rate/action_model': 1.624762218330297e-05, 'epoch': 2.1}) +[2026-06-18 05:25:07,003][starVLA.training.train_starvla][INFO] - Step 5458, Loss: {'train/loss': 0.17492426931858063, 'train/grad_norm_pre_clip': 13.976536750793457, 'timing/data': 0.015935726463794708, 'timing/model': 4.634479343891144, 'learning_rate/qwen_vl_interface': 1.6233651652552218e-06, 'learning_rate/action_model': 1.6233651652552217e-05, 'epoch': 2.1}) +[2026-06-18 05:25:14,573][starVLA.training.train_starvla][INFO] - Step 5459, Loss: {'train/loss': 0.2646399736404419, 'train/grad_norm_pre_clip': 12.654261589050293, 'timing/data': 0.014562292955815792, 'timing/model': 1.2132271295413375, 'learning_rate/qwen_vl_interface': 1.6219688639854507e-06, 'learning_rate/action_model': 1.6219688639854506e-05, 'epoch': 2.1}) +[2026-06-18 05:25:20,984][starVLA.training.train_starvla][INFO] - Step 5460, Loss: {'train/loss': 0.24825766682624817, 'train/grad_norm_pre_clip': 12.40233039855957, 'timing/data': 0.012320809997618198, 'timing/model': 1.163960418663919, 'learning_rate/qwen_vl_interface': 1.620573314810436e-06, 'learning_rate/action_model': 1.620573314810436e-05, 'epoch': 2.1}) +[2026-06-18 05:25:26,291][starVLA.training.train_starvla][INFO] - Step 5461, Loss: {'train/loss': 0.2399008721113205, 'train/grad_norm_pre_clip': 72.09937286376953, 'timing/data': 0.03728079795837402, 'timing/model': 1.4664445258677006, 'learning_rate/qwen_vl_interface': 1.6191785180194797e-06, 'learning_rate/action_model': 1.6191785180194796e-05, 'epoch': 2.1}) +[2026-06-18 05:25:32,245][starVLA.training.train_starvla][INFO] - Step 5462, Loss: {'train/loss': 0.6952954530715942, 'train/grad_norm_pre_clip': 21.392484664916992, 'timing/data': 0.013197449035942554, 'timing/model': 1.5047178883105516, 'learning_rate/qwen_vl_interface': 1.6177844739017236e-06, 'learning_rate/action_model': 1.6177844739017236e-05, 'epoch': 2.1}) +[2026-06-18 05:25:36,755][starVLA.training.train_starvla][INFO] - Step 5463, Loss: {'train/loss': 0.2812694013118744, 'train/grad_norm_pre_clip': 16.41085433959961, 'timing/data': 0.014366785064339638, 'timing/model': 1.2080876911059022, 'learning_rate/qwen_vl_interface': 1.6163911827461546e-06, 'learning_rate/action_model': 1.6163911827461546e-05, 'epoch': 2.1}) +[2026-06-18 05:25:41,999][starVLA.training.train_starvla][INFO] - Step 5464, Loss: {'train/loss': 0.2593264877796173, 'train/grad_norm_pre_clip': 12.206331253051758, 'timing/data': 0.01673855632543564, 'timing/model': 1.1768178083002567, 'learning_rate/qwen_vl_interface': 1.6149986448416032e-06, 'learning_rate/action_model': 1.614998644841603e-05, 'epoch': 2.1}) +[2026-06-18 05:25:47,108][starVLA.training.train_starvla][INFO] - Step 5465, Loss: {'train/loss': 0.36690187454223633, 'train/grad_norm_pre_clip': 16.729372024536133, 'timing/data': 0.015820475295186043, 'timing/model': 1.2109194276854396, 'learning_rate/qwen_vl_interface': 1.6136068604767444e-06, 'learning_rate/action_model': 1.6136068604767442e-05, 'epoch': 2.1}) +[2026-06-18 05:25:52,645][starVLA.training.train_starvla][INFO] - Step 5466, Loss: {'train/loss': 0.504460871219635, 'train/grad_norm_pre_clip': 18.642263412475586, 'timing/data': 0.01223536767065525, 'timing/model': 1.4499566406011581, 'learning_rate/qwen_vl_interface': 1.6122158299400962e-06, 'learning_rate/action_model': 1.612215829940096e-05, 'epoch': 2.1}) +[2026-06-18 05:26:04,772][starVLA.training.train_starvla][INFO] - Step 5467, Loss: {'train/loss': 0.19648267328739166, 'train/grad_norm_pre_clip': 13.441730499267578, 'timing/data': 0.021342670544981956, 'timing/model': 2.0946289598941803, 'learning_rate/qwen_vl_interface': 1.6108255535200193e-06, 'learning_rate/action_model': 1.6108255535200193e-05, 'epoch': 2.1}) +[2026-06-18 05:26:13,549][starVLA.training.train_starvla][INFO] - Step 5468, Loss: {'train/loss': 0.30736786127090454, 'train/grad_norm_pre_clip': 12.186399459838867, 'timing/data': 0.014859383925795555, 'timing/model': 1.2147354604676366, 'learning_rate/qwen_vl_interface': 1.6094360315047236e-06, 'learning_rate/action_model': 1.6094360315047237e-05, 'epoch': 2.1}) +[2026-06-18 05:26:19,238][starVLA.training.train_starvla][INFO] - Step 5469, Loss: {'train/loss': 0.23153136670589447, 'train/grad_norm_pre_clip': 14.43418025970459, 'timing/data': 0.01671546883881092, 'timing/model': 1.2836567340418696, 'learning_rate/qwen_vl_interface': 1.608047264182255e-06, 'learning_rate/action_model': 1.608047264182255e-05, 'epoch': 2.1}) +[2026-06-18 05:26:31,019][starVLA.training.train_starvla][INFO] - Step 5470, Loss: {'train/loss': 0.38407662510871887, 'train/grad_norm_pre_clip': 15.939790725708008, 'timing/data': 0.019454099237918854, 'timing/model': 1.1884354064241052, 'learning_rate/qwen_vl_interface': 1.6066592518405077e-06, 'learning_rate/action_model': 1.6066592518405076e-05, 'epoch': 2.1}) +[2026-06-18 05:26:45,707][starVLA.training.train_starvla][INFO] - Step 5471, Loss: {'train/loss': 0.2209467589855194, 'train/grad_norm_pre_clip': 19.854246139526367, 'timing/data': 0.01301935501396656, 'timing/model': 6.159815696999431, 'learning_rate/qwen_vl_interface': 1.605271994767216e-06, 'learning_rate/action_model': 1.605271994767216e-05, 'epoch': 2.1}) +[2026-06-18 05:26:53,250][starVLA.training.train_starvla][INFO] - Step 5472, Loss: {'train/loss': 0.3868176341056824, 'train/grad_norm_pre_clip': 20.467161178588867, 'timing/data': 0.0169226061552763, 'timing/model': 1.5045773163437843, 'learning_rate/qwen_vl_interface': 1.6038854932499635e-06, 'learning_rate/action_model': 1.6038854932499633e-05, 'epoch': 2.11}) +[2026-06-18 05:27:01,531][starVLA.training.train_starvla][INFO] - Step 5473, Loss: {'train/loss': 0.2548675537109375, 'train/grad_norm_pre_clip': 16.612123489379883, 'timing/data': 0.01348637044429779, 'timing/model': 1.0888806581497192, 'learning_rate/qwen_vl_interface': 1.6024997475761727e-06, 'learning_rate/action_model': 1.602499747576173e-05, 'epoch': 2.11}) +[2026-06-18 05:27:07,684][starVLA.training.train_starvla][INFO] - Step 5474, Loss: {'train/loss': 0.5100847482681274, 'train/grad_norm_pre_clip': 26.56429100036621, 'timing/data': 0.013045010156929493, 'timing/model': 1.7998039545491338, 'learning_rate/qwen_vl_interface': 1.6011147580331065e-06, 'learning_rate/action_model': 1.6011147580331066e-05, 'epoch': 2.11}) +[2026-06-18 05:27:18,565][starVLA.training.train_starvla][INFO] - Step 5475, Loss: {'train/loss': 0.20268814265727997, 'train/grad_norm_pre_clip': 14.539762496948242, 'timing/data': 0.01525188609957695, 'timing/model': 1.25934639852494, 'learning_rate/qwen_vl_interface': 1.5997305249078788e-06, 'learning_rate/action_model': 1.5997305249078787e-05, 'epoch': 2.11}) +[2026-06-18 05:27:24,317][starVLA.training.train_starvla][INFO] - Step 5476, Loss: {'train/loss': 0.22224849462509155, 'train/grad_norm_pre_clip': 41.66342544555664, 'timing/data': 0.011994658969342709, 'timing/model': 1.0891397399827838, 'learning_rate/qwen_vl_interface': 1.5983470484874412e-06, 'learning_rate/action_model': 1.598347048487441e-05, 'epoch': 2.11}) +[2026-06-18 05:27:29,587][starVLA.training.train_starvla][INFO] - Step 5477, Loss: {'train/loss': 0.6120306849479675, 'train/grad_norm_pre_clip': 23.89681625366211, 'timing/data': 0.012454497627913952, 'timing/model': 1.7084955181926489, 'learning_rate/qwen_vl_interface': 1.5969643290585893e-06, 'learning_rate/action_model': 1.5969643290585892e-05, 'epoch': 2.11}) +[2026-06-18 05:27:34,393][starVLA.training.train_starvla][INFO] - Step 5478, Loss: {'train/loss': 0.33556511998176575, 'train/grad_norm_pre_clip': 19.664901733398438, 'timing/data': 0.03960359003394842, 'timing/model': 1.3252675794064999, 'learning_rate/qwen_vl_interface': 1.5955823669079632e-06, 'learning_rate/action_model': 1.595582366907963e-05, 'epoch': 2.11}) +[2026-06-18 05:27:40,582][starVLA.training.train_starvla][INFO] - Step 5479, Loss: {'train/loss': 0.2439587116241455, 'train/grad_norm_pre_clip': 14.267121315002441, 'timing/data': 0.014965418726205826, 'timing/model': 1.315842225216329, 'learning_rate/qwen_vl_interface': 1.594201162322045e-06, 'learning_rate/action_model': 1.594201162322045e-05, 'epoch': 2.11}) +[2026-06-18 05:27:46,197][starVLA.training.train_starvla][INFO] - Step 5480, Loss: {'train/loss': 0.2717058062553406, 'train/grad_norm_pre_clip': 16.731643676757812, 'timing/data': 0.013309319503605366, 'timing/model': 1.5078137200325727, 'learning_rate/qwen_vl_interface': 1.592820715587159e-06, 'learning_rate/action_model': 1.5928207155871593e-05, 'epoch': 2.11}) +[2026-06-18 05:27:56,672][starVLA.training.train_starvla][INFO] - Step 5481, Loss: {'train/loss': 0.23023319244384766, 'train/grad_norm_pre_clip': 12.116135597229004, 'timing/data': 0.011619171127676964, 'timing/model': 1.5984696708619595, 'learning_rate/qwen_vl_interface': 1.5914410269894743e-06, 'learning_rate/action_model': 1.591441026989474e-05, 'epoch': 2.11}) +[2026-06-18 05:28:04,078][starVLA.training.train_starvla][INFO] - Step 5482, Loss: {'train/loss': 0.4620925784111023, 'train/grad_norm_pre_clip': 18.524290084838867, 'timing/data': 0.01133815199136734, 'timing/model': 1.7840581787750125, 'learning_rate/qwen_vl_interface': 1.5900620968150015e-06, 'learning_rate/action_model': 1.5900620968150015e-05, 'epoch': 2.11}) +[2026-06-18 05:28:09,659][starVLA.training.train_starvla][INFO] - Step 5483, Loss: {'train/loss': 0.37699955701828003, 'train/grad_norm_pre_clip': 24.127084732055664, 'timing/data': 0.012335475534200668, 'timing/model': 1.4018372669816017, 'learning_rate/qwen_vl_interface': 1.5886839253495944e-06, 'learning_rate/action_model': 1.5886839253495943e-05, 'epoch': 2.11}) +[2026-06-18 05:28:15,017][starVLA.training.train_starvla][INFO] - Step 5484, Loss: {'train/loss': 0.6016682982444763, 'train/grad_norm_pre_clip': 28.695600509643555, 'timing/data': 0.016542348079383373, 'timing/model': 1.1002998882904649, 'learning_rate/qwen_vl_interface': 1.5873065128789495e-06, 'learning_rate/action_model': 1.5873065128789493e-05, 'epoch': 2.11}) +[2026-06-18 05:28:20,541][starVLA.training.train_starvla][INFO] - Step 5485, Loss: {'train/loss': 0.13504065573215485, 'train/grad_norm_pre_clip': 7.597744941711426, 'timing/data': 0.0010657329112291336, 'timing/model': 1.485054725781083, 'learning_rate/qwen_vl_interface': 1.5859298596886038e-06, 'learning_rate/action_model': 1.585929859688604e-05, 'epoch': 2.11}) +[2026-06-18 05:28:25,794][starVLA.training.train_starvla][INFO] - Step 5486, Loss: {'train/loss': 0.26682865619659424, 'train/grad_norm_pre_clip': 14.443381309509277, 'timing/data': 0.001295541413128376, 'timing/model': 1.1506925225257874, 'learning_rate/qwen_vl_interface': 1.5845539660639429e-06, 'learning_rate/action_model': 1.5845539660639426e-05, 'epoch': 2.11}) +[2026-06-18 05:28:31,178][starVLA.training.train_starvla][INFO] - Step 5487, Loss: {'train/loss': 0.320753812789917, 'train/grad_norm_pre_clip': 13.575413703918457, 'timing/data': 0.0011293161660432816, 'timing/model': 1.4146751565858722, 'learning_rate/qwen_vl_interface': 1.5831788322901901e-06, 'learning_rate/action_model': 1.58317883229019e-05, 'epoch': 2.11}) +[2026-06-18 05:28:36,678][starVLA.training.train_starvla][INFO] - Step 5488, Loss: {'train/loss': 0.27498066425323486, 'train/grad_norm_pre_clip': 16.06807518005371, 'timing/data': 0.001057530753314495, 'timing/model': 1.6660777498036623, 'learning_rate/qwen_vl_interface': 1.5818044586524092e-06, 'learning_rate/action_model': 1.5818044586524092e-05, 'epoch': 2.11}) +[2026-06-18 05:28:43,605][starVLA.training.train_starvla][INFO] - Step 5489, Loss: {'train/loss': 0.2841093838214874, 'train/grad_norm_pre_clip': 17.03076171875, 'timing/data': 0.0014900146052241325, 'timing/model': 1.490015764720738, 'learning_rate/qwen_vl_interface': 1.5804308454355102e-06, 'learning_rate/action_model': 1.5804308454355102e-05, 'epoch': 2.11}) +[2026-06-18 05:28:48,908][starVLA.training.train_starvla][INFO] - Step 5490, Loss: {'train/loss': 0.48861557245254517, 'train/grad_norm_pre_clip': 15.991490364074707, 'timing/data': 0.0014041280373930931, 'timing/model': 1.5205601956695318, 'learning_rate/qwen_vl_interface': 1.5790579929242475e-06, 'learning_rate/action_model': 1.5790579929242475e-05, 'epoch': 2.11}) +[2026-06-18 05:29:02,479][starVLA.training.train_starvla][INFO] - Step 5491, Loss: {'train/loss': 0.40770211815834045, 'train/grad_norm_pre_clip': 19.232555389404297, 'timing/data': 0.0013714069500565529, 'timing/model': 1.159457820467651, 'learning_rate/qwen_vl_interface': 1.577685901403212e-06, 'learning_rate/action_model': 1.577685901403212e-05, 'epoch': 2.11}) +[2026-06-18 05:29:07,447][starVLA.training.train_starvla][INFO] - Step 5492, Loss: {'train/loss': 0.21704183518886566, 'train/grad_norm_pre_clip': 16.42713165283203, 'timing/data': 0.0012191534042358398, 'timing/model': 1.4962271265685558, 'learning_rate/qwen_vl_interface': 1.5763145711568414e-06, 'learning_rate/action_model': 1.5763145711568413e-05, 'epoch': 2.11}) +[2026-06-18 05:29:11,818][starVLA.training.train_starvla][INFO] - Step 5493, Loss: {'train/loss': 0.3169856071472168, 'train/grad_norm_pre_clip': 15.54574203491211, 'timing/data': 0.0010433131828904152, 'timing/model': 1.048610613681376, 'learning_rate/qwen_vl_interface': 1.5749440024694129e-06, 'learning_rate/action_model': 1.574944002469413e-05, 'epoch': 2.11}) +[2026-06-18 05:29:16,117][starVLA.training.train_starvla][INFO] - Step 5494, Loss: {'train/loss': 0.6044127345085144, 'train/grad_norm_pre_clip': 25.54375457763672, 'timing/data': 0.0010636607185006142, 'timing/model': 1.0635229041799903, 'learning_rate/qwen_vl_interface': 1.5735741956250476e-06, 'learning_rate/action_model': 1.5735741956250475e-05, 'epoch': 2.11}) +[2026-06-18 05:29:22,598][starVLA.training.train_starvla][INFO] - Step 5495, Loss: {'train/loss': 0.32524049282073975, 'train/grad_norm_pre_clip': 14.815374374389648, 'timing/data': 0.0011413739994168282, 'timing/model': 2.453936933539808, 'learning_rate/qwen_vl_interface': 1.5722051509077068e-06, 'learning_rate/action_model': 1.572205150907707e-05, 'epoch': 2.11}) +[2026-06-18 05:29:27,145][starVLA.training.train_starvla][INFO] - Step 5496, Loss: {'train/loss': 0.39864522218704224, 'train/grad_norm_pre_clip': 19.40958595275879, 'timing/data': 0.0009586866945028305, 'timing/model': 1.0787576995790005, 'learning_rate/qwen_vl_interface': 1.570836868601196e-06, 'learning_rate/action_model': 1.570836868601196e-05, 'epoch': 2.11}) +[2026-06-18 05:29:31,658][starVLA.training.train_starvla][INFO] - Step 5497, Loss: {'train/loss': 0.33340829610824585, 'train/grad_norm_pre_clip': 14.613351821899414, 'timing/data': 0.0010367371141910553, 'timing/model': 1.326587675139308, 'learning_rate/qwen_vl_interface': 1.5694693489891616e-06, 'learning_rate/action_model': 1.5694693489891617e-05, 'epoch': 2.11}) +[2026-06-18 05:29:35,793][starVLA.training.train_starvla][INFO] - Step 5498, Loss: {'train/loss': 0.4744512438774109, 'train/grad_norm_pre_clip': 19.15613555908203, 'timing/data': 0.0011295070871710777, 'timing/model': 1.0410243440419436, 'learning_rate/qwen_vl_interface': 1.5681025923550914e-06, 'learning_rate/action_model': 1.568102592355091e-05, 'epoch': 2.12}) +[2026-06-18 05:29:44,504][starVLA.training.train_starvla][INFO] - Step 5499, Loss: {'train/loss': 0.4650452136993408, 'train/grad_norm_pre_clip': 21.65369987487793, 'timing/data': 0.001034880056977272, 'timing/model': 1.1179534448310733, 'learning_rate/qwen_vl_interface': 1.5667365989823161e-06, 'learning_rate/action_model': 1.566736598982316e-05, 'epoch': 2.12}) +[2026-06-18 05:34:29,040][starVLA.training.train_starvla][INFO] - Step 5500, Loss: {'train/loss': 0.2524118721485138, 'train/grad_norm_pre_clip': 13.694833755493164, 'eval/loss': 0.4378877580165863, 'eval/latency_4/loss': 0.4378877580165863, 'eval/demon_fire_precision': 0.9751471550032701, 'eval/demon_fire_recall': 0.9684962650211107, 'eval/demon_fire_f1': 0.9718103307805117, 'eval/demon_fire_rate_model': 0.8980910425844346, 'eval/demon_fire_rate_teacher': 0.9042584434654919, 'eval/demon_left_precision': 0.7302631578947368, 'eval/demon_left_recall': 0.7602739726027398, 'eval/demon_left_f1': 0.7449664429530201, 'eval/demon_left_rate_model': 0.13392070484581498, 'eval/demon_left_rate_teacher': 0.12863436123348018, 'eval/demon_right_precision': 0.7934782608695652, 'eval/demon_right_recall': 0.7906137184115524, 'eval/demon_right_f1': 0.7920433996383365, 'eval/demon_right_rate_model': 0.16211453744493393, 'eval/demon_right_rate_teacher': 0.16270190895741557, 'eval/demon_move_f1': 0.7685049212956783, 'eval/demon_cc_f1': 0.900354591041839, 'eval/latency_4/demon_fire_precision': 0.9751471550032701, 'eval/latency_4/demon_fire_recall': 0.9684962650211107, 'eval/latency_4/demon_fire_f1': 0.9718103307805117, 'eval/latency_4/demon_fire_rate_model': 0.8980910425844346, 'eval/latency_4/demon_fire_rate_teacher': 0.9042584434654919, 'eval/latency_4/demon_left_precision': 0.7302631578947368, 'eval/latency_4/demon_left_recall': 0.7602739726027398, 'eval/latency_4/demon_left_f1': 0.7449664429530201, 'eval/latency_4/demon_left_rate_model': 0.13392070484581498, 'eval/latency_4/demon_left_rate_teacher': 0.12863436123348018, 'eval/latency_4/demon_right_precision': 0.7934782608695652, 'eval/latency_4/demon_right_recall': 0.7906137184115524, 'eval/latency_4/demon_right_f1': 0.7920433996383365, 'eval/latency_4/demon_right_rate_model': 0.16211453744493393, 'eval/latency_4/demon_right_rate_teacher': 0.16270190895741557, 'eval/latency_4/demon_move_f1': 0.7685049212956783, 'eval/latency_4/demon_cc_f1': 0.900354591041839, 'timing/data': 0.0011665243655443192, 'timing/model': 1.0903061730787158, 'learning_rate/qwen_vl_interface': 1.5653713691540053e-06, 'learning_rate/action_model': 1.5653713691540054e-05, 'epoch': 2.12}) +[2026-06-18 05:34:29,045][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_5500_state +[2026-06-18 05:34:29,057][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 05:34:40,189][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_5500_state/model.safetensors +[2026-06-18 05:35:06,837][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_5500_state/optimizer.bin +[2026-06-18 05:35:06,838][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_5500_state/sampler.bin +[2026-06-18 05:35:06,839][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_5500_state/sampler_1.bin +[2026-06-18 05:35:06,841][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_5500_state/random_states_0.pkl +[2026-06-18 05:35:06,851][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 5500 +[2026-06-18 05:35:06,852][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 05:35:06,869][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 05:35:18,608][starVLA.training.train_starvla][INFO] - Step 5501, Loss: {'train/loss': 0.49779370427131653, 'train/grad_norm_pre_clip': 15.29586124420166, 'timing/data': 0.0009828051552176476, 'timing/model': 4.495254465378821, 'learning_rate/qwen_vl_interface': 1.5640069031531774e-06, 'learning_rate/action_model': 1.5640069031531772e-05, 'epoch': 2.12}) +[2026-06-18 05:35:23,729][starVLA.training.train_starvla][INFO] - Step 5502, Loss: {'train/loss': 0.38010814785957336, 'train/grad_norm_pre_clip': 19.18829345703125, 'timing/data': 0.0016074366867542267, 'timing/model': 1.3620241191238165, 'learning_rate/qwen_vl_interface': 1.5626432012626835e-06, 'learning_rate/action_model': 1.5626432012626836e-05, 'epoch': 2.12}) +[2026-06-18 05:35:29,203][starVLA.training.train_starvla][INFO] - Step 5503, Loss: {'train/loss': 0.30388668179512024, 'train/grad_norm_pre_clip': 12.039337158203125, 'timing/data': 0.0013241702690720558, 'timing/model': 1.177385744638741, 'learning_rate/qwen_vl_interface': 1.5612802637652202e-06, 'learning_rate/action_model': 1.5612802637652203e-05, 'epoch': 2.12}) +[2026-06-18 05:35:34,528][starVLA.training.train_starvla][INFO] - Step 5504, Loss: {'train/loss': 0.39207401871681213, 'train/grad_norm_pre_clip': 14.247685432434082, 'timing/data': 0.0018668314442038536, 'timing/model': 1.2057255422696471, 'learning_rate/qwen_vl_interface': 1.5599180909433292e-06, 'learning_rate/action_model': 1.5599180909433292e-05, 'epoch': 2.12}) +[2026-06-18 05:35:39,942][starVLA.training.train_starvla][INFO] - Step 5505, Loss: {'train/loss': 0.28774920105934143, 'train/grad_norm_pre_clip': 18.101343154907227, 'timing/data': 0.0012957686558365822, 'timing/model': 1.1551905116066337, 'learning_rate/qwen_vl_interface': 1.5585566830793883e-06, 'learning_rate/action_model': 1.558556683079388e-05, 'epoch': 2.12}) +[2026-06-18 05:35:45,331][starVLA.training.train_starvla][INFO] - Step 5506, Loss: {'train/loss': 0.40975069999694824, 'train/grad_norm_pre_clip': 19.64360237121582, 'timing/data': 0.0012449482455849648, 'timing/model': 1.1783622093498707, 'learning_rate/qwen_vl_interface': 1.5571960404556204e-06, 'learning_rate/action_model': 1.5571960404556203e-05, 'epoch': 2.12}) +[2026-06-18 05:35:50,685][starVLA.training.train_starvla][INFO] - Step 5507, Loss: {'train/loss': 0.33285048604011536, 'train/grad_norm_pre_clip': 18.65546989440918, 'timing/data': 0.0015168990939855576, 'timing/model': 1.6402304330840707, 'learning_rate/qwen_vl_interface': 1.5558361633540848e-06, 'learning_rate/action_model': 1.5558361633540848e-05, 'epoch': 2.12}) +[2026-06-18 05:36:00,330][starVLA.training.train_starvla][INFO] - Step 5508, Loss: {'train/loss': 0.24412406980991364, 'train/grad_norm_pre_clip': 10.405591011047363, 'timing/data': 0.0013573477044701576, 'timing/model': 1.1878921054303646, 'learning_rate/qwen_vl_interface': 1.5544770520566884e-06, 'learning_rate/action_model': 1.5544770520566885e-05, 'epoch': 2.12}) +[2026-06-18 05:36:05,728][starVLA.training.train_starvla][INFO] - Step 5509, Loss: {'train/loss': 0.2967721223831177, 'train/grad_norm_pre_clip': 14.206331253051758, 'timing/data': 0.0013696160167455673, 'timing/model': 1.4507368989288807, 'learning_rate/qwen_vl_interface': 1.553118706845176e-06, 'learning_rate/action_model': 1.5531187068451757e-05, 'epoch': 2.12}) +[2026-06-18 05:36:12,800][starVLA.training.train_starvla][INFO] - Step 5510, Loss: {'train/loss': 0.5146512389183044, 'train/grad_norm_pre_clip': 19.309797286987305, 'timing/data': 0.0016454700380563736, 'timing/model': 1.288869496434927, 'learning_rate/qwen_vl_interface': 1.5517611280011333e-06, 'learning_rate/action_model': 1.551761128001133e-05, 'epoch': 2.12}) +[2026-06-18 05:36:17,865][starVLA.training.train_starvla][INFO] - Step 5511, Loss: {'train/loss': 0.36527538299560547, 'train/grad_norm_pre_clip': 15.093770980834961, 'timing/data': 0.0011134082451462746, 'timing/model': 1.1095655597746372, 'learning_rate/qwen_vl_interface': 1.5504043158059886e-06, 'learning_rate/action_model': 1.5504043158059885e-05, 'epoch': 2.12}) +[2026-06-18 05:36:23,449][starVLA.training.train_starvla][INFO] - Step 5512, Loss: {'train/loss': 0.45671290159225464, 'train/grad_norm_pre_clip': 14.581419944763184, 'timing/data': 0.0009971391409635544, 'timing/model': 1.500167103484273, 'learning_rate/qwen_vl_interface': 1.54904827054101e-06, 'learning_rate/action_model': 1.54904827054101e-05, 'epoch': 2.12}) +[2026-06-18 05:36:29,256][starVLA.training.train_starvla][INFO] - Step 5513, Loss: {'train/loss': 0.22090430557727814, 'train/grad_norm_pre_clip': 14.898868560791016, 'timing/data': 0.001331252045929432, 'timing/model': 2.45884757861495, 'learning_rate/qwen_vl_interface': 1.5476929924873074e-06, 'learning_rate/action_model': 1.5476929924873073e-05, 'epoch': 2.12}) +[2026-06-18 05:36:39,185][starVLA.training.train_starvla][INFO] - Step 5514, Loss: {'train/loss': 0.20861594378948212, 'train/grad_norm_pre_clip': 12.374664306640625, 'timing/data': 0.0011754650622606277, 'timing/model': 1.4197919527068734, 'learning_rate/qwen_vl_interface': 1.5463384819258302e-06, 'learning_rate/action_model': 1.54633848192583e-05, 'epoch': 2.12}) +[2026-06-18 05:36:44,440][starVLA.training.train_starvla][INFO] - Step 5515, Loss: {'train/loss': 0.24507151544094086, 'train/grad_norm_pre_clip': 10.327266693115234, 'timing/data': 0.0035937028005719185, 'timing/model': 1.3941372483968735, 'learning_rate/qwen_vl_interface': 1.544984739137373e-06, 'learning_rate/action_model': 1.544984739137373e-05, 'epoch': 2.12}) +[2026-06-18 05:36:49,923][starVLA.training.train_starvla][INFO] - Step 5516, Loss: {'train/loss': 0.269834965467453, 'train/grad_norm_pre_clip': 13.687339782714844, 'timing/data': 0.001327446661889553, 'timing/model': 1.201806136406958, 'learning_rate/qwen_vl_interface': 1.543631764402566e-06, 'learning_rate/action_model': 1.5436317644025658e-05, 'epoch': 2.12}) +[2026-06-18 05:36:56,021][starVLA.training.train_starvla][INFO] - Step 5517, Loss: {'train/loss': 0.42903944849967957, 'train/grad_norm_pre_clip': 15.251343727111816, 'timing/data': 0.0015923380851745605, 'timing/model': 2.1405661087483168, 'learning_rate/qwen_vl_interface': 1.5422795580018813e-06, 'learning_rate/action_model': 1.542279558001881e-05, 'epoch': 2.12}) +[2026-06-18 05:37:01,117][starVLA.training.train_starvla][INFO] - Step 5518, Loss: {'train/loss': 0.48537564277648926, 'train/grad_norm_pre_clip': 18.81583023071289, 'timing/data': 0.0014301026239991188, 'timing/model': 1.3148671733215451, 'learning_rate/qwen_vl_interface': 1.5409281202156329e-06, 'learning_rate/action_model': 1.540928120215633e-05, 'epoch': 2.12}) +[2026-06-18 05:37:06,617][starVLA.training.train_starvla][INFO] - Step 5519, Loss: {'train/loss': 0.15715183317661285, 'train/grad_norm_pre_clip': 10.261670112609863, 'timing/data': 0.0013436274603009224, 'timing/model': 1.5090738702565432, 'learning_rate/qwen_vl_interface': 1.5395774513239777e-06, 'learning_rate/action_model': 1.539577451323978e-05, 'epoch': 2.12}) +[2026-06-18 05:37:12,051][starVLA.training.train_starvla][INFO] - Step 5520, Loss: {'train/loss': 0.16190508008003235, 'train/grad_norm_pre_clip': 10.363385200500488, 'timing/data': 0.0011085672304034233, 'timing/model': 1.1776301795616746, 'learning_rate/qwen_vl_interface': 1.5382275516069112e-06, 'learning_rate/action_model': 1.5382275516069113e-05, 'epoch': 2.12}) +[2026-06-18 05:37:20,886][starVLA.training.train_starvla][INFO] - Step 5521, Loss: {'train/loss': 0.3176163136959076, 'train/grad_norm_pre_clip': 19.884153366088867, 'timing/data': 0.0009900564327836037, 'timing/model': 1.3198970761150122, 'learning_rate/qwen_vl_interface': 1.5368784213442653e-06, 'learning_rate/action_model': 1.5368784213442654e-05, 'epoch': 2.12}) +[2026-06-18 05:37:26,285][starVLA.training.train_starvla][INFO] - Step 5522, Loss: {'train/loss': 0.42965272068977356, 'train/grad_norm_pre_clip': 22.655990600585938, 'timing/data': 0.001366306096315384, 'timing/model': 1.6722236312925816, 'learning_rate/qwen_vl_interface': 1.5355300608157172e-06, 'learning_rate/action_model': 1.5355300608157173e-05, 'epoch': 2.12}) +[2026-06-18 05:37:34,830][starVLA.training.train_starvla][INFO] - Step 5523, Loss: {'train/loss': 0.34470176696777344, 'train/grad_norm_pre_clip': 14.731870651245117, 'timing/data': 0.0012430651113390923, 'timing/model': 1.1102316780015826, 'learning_rate/qwen_vl_interface': 1.5341824703007854e-06, 'learning_rate/action_model': 1.5341824703007854e-05, 'epoch': 2.12}) +[2026-06-18 05:37:40,518][starVLA.training.train_starvla][INFO] - Step 5524, Loss: {'train/loss': 0.21930676698684692, 'train/grad_norm_pre_clip': 12.256692886352539, 'timing/data': 0.0007829274982213974, 'timing/model': 1.3408418856561184, 'learning_rate/qwen_vl_interface': 1.5328356500788263e-06, 'learning_rate/action_model': 1.5328356500788264e-05, 'epoch': 2.13}) +[2026-06-18 05:37:46,086][starVLA.training.train_starvla][INFO] - Step 5525, Loss: {'train/loss': 0.4834645688533783, 'train/grad_norm_pre_clip': 16.606943130493164, 'timing/data': 0.0013285912573337555, 'timing/model': 1.288443080149591, 'learning_rate/qwen_vl_interface': 1.5314896004290366e-06, 'learning_rate/action_model': 1.5314896004290365e-05, 'epoch': 2.13}) +[2026-06-18 05:37:51,803][starVLA.training.train_starvla][INFO] - Step 5526, Loss: {'train/loss': 0.3629320561885834, 'train/grad_norm_pre_clip': 19.136804580688477, 'timing/data': 0.0008253641426563263, 'timing/model': 1.7107289880514145, 'learning_rate/qwen_vl_interface': 1.530144321630454e-06, 'learning_rate/action_model': 1.530144321630454e-05, 'epoch': 2.13}) +[2026-06-18 05:38:00,920][starVLA.training.train_starvla][INFO] - Step 5527, Loss: {'train/loss': 0.3531363010406494, 'train/grad_norm_pre_clip': 10.720170974731445, 'timing/data': 0.0012353938072919846, 'timing/model': 1.5654624123126268, 'learning_rate/qwen_vl_interface': 1.5287998139619564e-06, 'learning_rate/action_model': 1.5287998139619562e-05, 'epoch': 2.13}) +[2026-06-18 05:38:09,058][starVLA.training.train_starvla][INFO] - Step 5528, Loss: {'train/loss': 0.2485022246837616, 'train/grad_norm_pre_clip': 12.257071495056152, 'timing/data': 0.004859122447669506, 'timing/model': 4.344414194114506, 'learning_rate/qwen_vl_interface': 1.5274560777022613e-06, 'learning_rate/action_model': 1.5274560777022613e-05, 'epoch': 2.13}) +[2026-06-18 05:38:14,943][starVLA.training.train_starvla][INFO] - Step 5529, Loss: {'train/loss': 0.1318400353193283, 'train/grad_norm_pre_clip': 8.278436660766602, 'timing/data': 0.0018182750791311264, 'timing/model': 1.5348166832700372, 'learning_rate/qwen_vl_interface': 1.5261131131299274e-06, 'learning_rate/action_model': 1.5261131131299272e-05, 'epoch': 2.13}) +[2026-06-18 05:38:26,463][starVLA.training.train_starvla][INFO] - Step 5530, Loss: {'train/loss': 0.3315361738204956, 'train/grad_norm_pre_clip': 15.488838195800781, 'timing/data': 0.0017278986051678658, 'timing/model': 1.3303403090685606, 'learning_rate/qwen_vl_interface': 1.5247709205233518e-06, 'learning_rate/action_model': 1.5247709205233518e-05, 'epoch': 2.13}) +[2026-06-18 05:38:35,969][starVLA.training.train_starvla][INFO] - Step 5531, Loss: {'train/loss': 0.47401827573776245, 'train/grad_norm_pre_clip': 20.630292892456055, 'timing/data': 0.0014447402209043503, 'timing/model': 1.61805082205683, 'learning_rate/qwen_vl_interface': 1.5234295001607732e-06, 'learning_rate/action_model': 1.5234295001607732e-05, 'epoch': 2.13}) +[2026-06-18 05:38:41,736][starVLA.training.train_starvla][INFO] - Step 5532, Loss: {'train/loss': 0.3546083867549896, 'train/grad_norm_pre_clip': 14.812841415405273, 'timing/data': 0.0011529875919222832, 'timing/model': 1.5874487021937966, 'learning_rate/qwen_vl_interface': 1.5220888523202687e-06, 'learning_rate/action_model': 1.5220888523202686e-05, 'epoch': 2.13}) +[2026-06-18 05:38:54,010][starVLA.training.train_starvla][INFO] - Step 5533, Loss: {'train/loss': 0.42578163743019104, 'train/grad_norm_pre_clip': 14.56651496887207, 'timing/data': 0.0013807164505124092, 'timing/model': 1.4478210816159844, 'learning_rate/qwen_vl_interface': 1.5207489772797557e-06, 'learning_rate/action_model': 1.5207489772797556e-05, 'epoch': 2.13}) +[2026-06-18 05:38:59,459][starVLA.training.train_starvla][INFO] - Step 5534, Loss: {'train/loss': 0.341401070356369, 'train/grad_norm_pre_clip': 18.792449951171875, 'timing/data': 0.0009361375123262405, 'timing/model': 1.2607226623222232, 'learning_rate/qwen_vl_interface': 1.5194098753169933e-06, 'learning_rate/action_model': 1.5194098753169933e-05, 'epoch': 2.13}) +[2026-06-18 05:39:04,642][starVLA.training.train_starvla][INFO] - Step 5535, Loss: {'train/loss': 0.2437395602464676, 'train/grad_norm_pre_clip': 14.35871410369873, 'timing/data': 0.001791323535144329, 'timing/model': 1.463227866217494, 'learning_rate/qwen_vl_interface': 1.5180715467095799e-06, 'learning_rate/action_model': 1.5180715467095798e-05, 'epoch': 2.13}) +[2026-06-18 05:39:10,496][starVLA.training.train_starvla][INFO] - Step 5536, Loss: {'train/loss': 0.3463384807109833, 'train/grad_norm_pre_clip': 16.463523864746094, 'timing/data': 0.0011394983157515526, 'timing/model': 1.2111085457727313, 'learning_rate/qwen_vl_interface': 1.5167339917349482e-06, 'learning_rate/action_model': 1.5167339917349482e-05, 'epoch': 2.13}) +[2026-06-18 05:39:16,188][starVLA.training.train_starvla][INFO] - Step 5537, Loss: {'train/loss': 0.3394465744495392, 'train/grad_norm_pre_clip': 14.288871765136719, 'timing/data': 0.0011358512565493584, 'timing/model': 1.4943572636693716, 'learning_rate/qwen_vl_interface': 1.5153972106703778e-06, 'learning_rate/action_model': 1.5153972106703779e-05, 'epoch': 2.13}) +[2026-06-18 05:39:21,637][starVLA.training.train_starvla][INFO] - Step 5538, Loss: {'train/loss': 0.30771464109420776, 'train/grad_norm_pre_clip': 16.14304542541504, 'timing/data': 0.0014594485983252525, 'timing/model': 1.2679359875619411, 'learning_rate/qwen_vl_interface': 1.5140612037929844e-06, 'learning_rate/action_model': 1.5140612037929844e-05, 'epoch': 2.13}) +[2026-06-18 05:39:27,316][starVLA.training.train_starvla][INFO] - Step 5539, Loss: {'train/loss': 0.5938040018081665, 'train/grad_norm_pre_clip': 14.712488174438477, 'timing/data': 0.0022676875814795494, 'timing/model': 1.5272831600159407, 'learning_rate/qwen_vl_interface': 1.512725971379724e-06, 'learning_rate/action_model': 1.5127259713797239e-05, 'epoch': 2.13}) +[2026-06-18 05:39:32,805][starVLA.training.train_starvla][INFO] - Step 5540, Loss: {'train/loss': 0.3820855915546417, 'train/grad_norm_pre_clip': 16.823314666748047, 'timing/data': 0.0011531868949532509, 'timing/model': 1.0777183128520846, 'learning_rate/qwen_vl_interface': 1.5113915137073905e-06, 'learning_rate/action_model': 1.5113915137073906e-05, 'epoch': 2.13}) +[2026-06-18 05:39:43,315][starVLA.training.train_starvla][INFO] - Step 5541, Loss: {'train/loss': 0.433767169713974, 'train/grad_norm_pre_clip': 26.817705154418945, 'timing/data': 0.0012527666985988617, 'timing/model': 1.2390169678255916, 'learning_rate/qwen_vl_interface': 1.5100578310526198e-06, 'learning_rate/action_model': 1.5100578310526197e-05, 'epoch': 2.13}) +[2026-06-18 05:39:48,159][starVLA.training.train_starvla][INFO] - Step 5542, Loss: {'train/loss': 0.23578499257564545, 'train/grad_norm_pre_clip': 15.77489948272705, 'timing/data': 0.0008264174684882164, 'timing/model': 1.4218181790784001, 'learning_rate/qwen_vl_interface': 1.508724923691885e-06, 'learning_rate/action_model': 1.5087249236918848e-05, 'epoch': 2.13}) +[2026-06-18 05:39:55,841][starVLA.training.train_starvla][INFO] - Step 5543, Loss: {'train/loss': 0.42055395245552063, 'train/grad_norm_pre_clip': 18.48468589782715, 'timing/data': 0.0009758966043591499, 'timing/model': 1.7393568642437458, 'learning_rate/qwen_vl_interface': 1.5073927919014986e-06, 'learning_rate/action_model': 1.5073927919014985e-05, 'epoch': 2.13}) +[2026-06-18 05:40:01,513][starVLA.training.train_starvla][INFO] - Step 5544, Loss: {'train/loss': 0.2537789046764374, 'train/grad_norm_pre_clip': 11.961350440979004, 'timing/data': 0.0011446615681052208, 'timing/model': 1.3235909705981612, 'learning_rate/qwen_vl_interface': 1.506061435957614e-06, 'learning_rate/action_model': 1.506061435957614e-05, 'epoch': 2.13}) +[2026-06-18 05:40:10,465][starVLA.training.train_starvla][INFO] - Step 5545, Loss: {'train/loss': 0.3939935266971588, 'train/grad_norm_pre_clip': 18.572175979614258, 'timing/data': 0.0014121755957603455, 'timing/model': 4.2128111543133855, 'learning_rate/qwen_vl_interface': 1.5047308561362225e-06, 'learning_rate/action_model': 1.5047308561362225e-05, 'epoch': 2.13}) +[2026-06-18 05:40:16,480][starVLA.training.train_starvla][INFO] - Step 5546, Loss: {'train/loss': 0.18001402914524078, 'train/grad_norm_pre_clip': 11.832881927490234, 'timing/data': 0.0012408271431922913, 'timing/model': 1.5988298738375306, 'learning_rate/qwen_vl_interface': 1.5034010527131548e-06, 'learning_rate/action_model': 1.5034010527131549e-05, 'epoch': 2.13}) +[2026-06-18 05:40:20,918][starVLA.training.train_starvla][INFO] - Step 5547, Loss: {'train/loss': 0.6178722381591797, 'train/grad_norm_pre_clip': 15.821260452270508, 'timing/data': 0.0011374354362487793, 'timing/model': 1.22186085768044, 'learning_rate/qwen_vl_interface': 1.5020720259640792e-06, 'learning_rate/action_model': 1.5020720259640792e-05, 'epoch': 2.13}) +[2026-06-18 05:40:27,472][starVLA.training.train_starvla][INFO] - Step 5548, Loss: {'train/loss': 0.4222206175327301, 'train/grad_norm_pre_clip': 21.410097122192383, 'timing/data': 0.0011189961805939674, 'timing/model': 1.3184909345582128, 'learning_rate/qwen_vl_interface': 1.5007437761645064e-06, 'learning_rate/action_model': 1.5007437761645063e-05, 'epoch': 2.13}) +[2026-06-18 05:40:35,341][starVLA.training.train_starvla][INFO] - Step 5549, Loss: {'train/loss': 0.3123925030231476, 'train/grad_norm_pre_clip': 16.493316650390625, 'timing/data': 0.0009393533691763878, 'timing/model': 1.2627589832991362, 'learning_rate/qwen_vl_interface': 1.4994163035897845e-06, 'learning_rate/action_model': 1.4994163035897846e-05, 'epoch': 2.13}) +[2026-06-18 05:40:40,697][starVLA.training.train_starvla][INFO] - Step 5550, Loss: {'train/loss': 0.36181801557540894, 'train/grad_norm_pre_clip': 16.204326629638672, 'timing/data': 0.003141193650662899, 'timing/model': 1.1609195666387677, 'learning_rate/qwen_vl_interface': 1.4980896085150976e-06, 'learning_rate/action_model': 1.4980896085150975e-05, 'epoch': 2.14}) +[2026-06-18 05:40:45,480][starVLA.training.train_starvla][INFO] - Step 5551, Loss: {'train/loss': 0.28959590196609497, 'train/grad_norm_pre_clip': 12.113546371459961, 'timing/data': 0.001224924810230732, 'timing/model': 1.1258858842775226, 'learning_rate/qwen_vl_interface': 1.4967636912154707e-06, 'learning_rate/action_model': 1.4967636912154706e-05, 'epoch': 2.14}) +[2026-06-18 05:40:54,086][starVLA.training.train_starvla][INFO] - Step 5552, Loss: {'train/loss': 0.3549124300479889, 'train/grad_norm_pre_clip': 24.21140480041504, 'timing/data': 0.0009268997237086296, 'timing/model': 1.7699782131239772, 'learning_rate/qwen_vl_interface': 1.4954385519657701e-06, 'learning_rate/action_model': 1.49543855196577e-05, 'epoch': 2.14}) +[2026-06-18 05:41:00,979][starVLA.training.train_starvla][INFO] - Step 5553, Loss: {'train/loss': 0.32919058203697205, 'train/grad_norm_pre_clip': 16.54436683654785, 'timing/data': 0.002379646524786949, 'timing/model': 1.5201096013188362, 'learning_rate/qwen_vl_interface': 1.4941141910406984e-06, 'learning_rate/action_model': 1.4941141910406984e-05, 'epoch': 2.14}) +[2026-06-18 05:41:05,915][starVLA.training.train_starvla][INFO] - Step 5554, Loss: {'train/loss': 0.6993991732597351, 'train/grad_norm_pre_clip': 20.244747161865234, 'timing/data': 0.001383211463689804, 'timing/model': 1.0946944914758205, 'learning_rate/qwen_vl_interface': 1.4927906087147966e-06, 'learning_rate/action_model': 1.4927906087147967e-05, 'epoch': 2.14}) +[2026-06-18 05:41:10,889][starVLA.training.train_starvla][INFO] - Step 5555, Loss: {'train/loss': 0.21341538429260254, 'train/grad_norm_pre_clip': 12.755242347717285, 'timing/data': 0.0012300368398427963, 'timing/model': 1.1492829257622361, 'learning_rate/qwen_vl_interface': 1.491467805262444e-06, 'learning_rate/action_model': 1.4914678052624439e-05, 'epoch': 2.14}) +[2026-06-18 05:41:17,239][starVLA.training.train_starvla][INFO] - Step 5556, Loss: {'train/loss': 0.33824509382247925, 'train/grad_norm_pre_clip': 16.299789428710938, 'timing/data': 0.001315775327384472, 'timing/model': 1.3688746755942702, 'learning_rate/qwen_vl_interface': 1.4901457809578598e-06, 'learning_rate/action_model': 1.4901457809578597e-05, 'epoch': 2.14}) +[2026-06-18 05:41:22,230][starVLA.training.train_starvla][INFO] - Step 5557, Loss: {'train/loss': 0.17755119502544403, 'train/grad_norm_pre_clip': 11.499677658081055, 'timing/data': 0.0011411495506763458, 'timing/model': 1.5376104982569814, 'learning_rate/qwen_vl_interface': 1.4888245360751014e-06, 'learning_rate/action_model': 1.4888245360751013e-05, 'epoch': 2.14}) +[2026-06-18 05:41:30,180][starVLA.training.train_starvla][INFO] - Step 5558, Loss: {'train/loss': 0.3677826523780823, 'train/grad_norm_pre_clip': 15.394902229309082, 'timing/data': 0.001191064715385437, 'timing/model': 1.122437672689557, 'learning_rate/qwen_vl_interface': 1.4875040708880631e-06, 'learning_rate/action_model': 1.487504070888063e-05, 'epoch': 2.14}) +[2026-06-18 05:41:36,155][starVLA.training.train_starvla][INFO] - Step 5559, Loss: {'train/loss': 0.19245541095733643, 'train/grad_norm_pre_clip': 10.86355972290039, 'timing/data': 0.0011442909017205238, 'timing/model': 1.5606593936681747, 'learning_rate/qwen_vl_interface': 1.4861843856704802e-06, 'learning_rate/action_model': 1.4861843856704801e-05, 'epoch': 2.14}) +[2026-06-18 05:41:41,943][starVLA.training.train_starvla][INFO] - Step 5560, Loss: {'train/loss': 0.29453402757644653, 'train/grad_norm_pre_clip': 12.225279808044434, 'timing/data': 0.0011473633348941803, 'timing/model': 1.473758933134377, 'learning_rate/qwen_vl_interface': 1.4848654806959243e-06, 'learning_rate/action_model': 1.4848654806959242e-05, 'epoch': 2.14}) +[2026-06-18 05:41:48,694][starVLA.training.train_starvla][INFO] - Step 5561, Loss: {'train/loss': 0.5018463730812073, 'train/grad_norm_pre_clip': 22.495840072631836, 'timing/data': 0.0012259269133210182, 'timing/model': 1.2816969510167837, 'learning_rate/qwen_vl_interface': 1.483547356237805e-06, 'learning_rate/action_model': 1.4835473562378049e-05, 'epoch': 2.14}) +[2026-06-18 05:41:53,972][starVLA.training.train_starvla][INFO] - Step 5562, Loss: {'train/loss': 0.33694109320640564, 'train/grad_norm_pre_clip': 15.70970344543457, 'timing/data': 0.0011493014171719551, 'timing/model': 1.703365832567215, 'learning_rate/qwen_vl_interface': 1.4822300125693709e-06, 'learning_rate/action_model': 1.4822300125693708e-05, 'epoch': 2.14}) +[2026-06-18 05:41:58,550][starVLA.training.train_starvla][INFO] - Step 5563, Loss: {'train/loss': 0.5918697714805603, 'train/grad_norm_pre_clip': 19.224016189575195, 'timing/data': 0.0011525824666023254, 'timing/model': 1.2716013938188553, 'learning_rate/qwen_vl_interface': 1.480913449963712e-06, 'learning_rate/action_model': 1.480913449963712e-05, 'epoch': 2.14}) +[2026-06-18 05:42:03,718][starVLA.training.train_starvla][INFO] - Step 5564, Loss: {'train/loss': 0.3649880886077881, 'train/grad_norm_pre_clip': 12.235505104064941, 'timing/data': 0.0014632847160100937, 'timing/model': 1.2041898472234607, 'learning_rate/qwen_vl_interface': 1.4795976686937488e-06, 'learning_rate/action_model': 1.4795976686937487e-05, 'epoch': 2.14}) +[2026-06-18 05:42:08,816][starVLA.training.train_starvla][INFO] - Step 5565, Loss: {'train/loss': 0.4132630228996277, 'train/grad_norm_pre_clip': 14.958791732788086, 'timing/data': 0.0011030612513422966, 'timing/model': 1.3117415504530072, 'learning_rate/qwen_vl_interface': 1.4782826690322453e-06, 'learning_rate/action_model': 1.4782826690322451e-05, 'epoch': 2.14}) +[2026-06-18 05:42:13,848][starVLA.training.train_starvla][INFO] - Step 5566, Loss: {'train/loss': 0.31772297620773315, 'train/grad_norm_pre_clip': 13.812654495239258, 'timing/data': 0.0010849181562662125, 'timing/model': 1.3394771479070187, 'learning_rate/qwen_vl_interface': 1.4769684512518045e-06, 'learning_rate/action_model': 1.4769684512518045e-05, 'epoch': 2.14}) +[2026-06-18 05:42:19,133][starVLA.training.train_starvla][INFO] - Step 5567, Loss: {'train/loss': 0.3587934970855713, 'train/grad_norm_pre_clip': 16.558029174804688, 'timing/data': 0.0014287317171692848, 'timing/model': 1.5798417227342725, 'learning_rate/qwen_vl_interface': 1.4756550156248633e-06, 'learning_rate/action_model': 1.4756550156248632e-05, 'epoch': 2.14}) +[2026-06-18 05:42:24,889][starVLA.training.train_starvla][INFO] - Step 5568, Loss: {'train/loss': 0.3471209704875946, 'train/grad_norm_pre_clip': 58.045623779296875, 'timing/data': 0.0013464801013469696, 'timing/model': 1.3037524996325374, 'learning_rate/qwen_vl_interface': 1.4743423624237e-06, 'learning_rate/action_model': 1.4743423624236998e-05, 'epoch': 2.14}) +[2026-06-18 05:42:30,112][starVLA.training.train_starvla][INFO] - Step 5569, Loss: {'train/loss': 0.2431473582983017, 'train/grad_norm_pre_clip': 15.378016471862793, 'timing/data': 0.001184246502816677, 'timing/model': 1.1635718122124672, 'learning_rate/qwen_vl_interface': 1.4730304919204253e-06, 'learning_rate/action_model': 1.4730304919204252e-05, 'epoch': 2.14}) +[2026-06-18 05:42:36,857][starVLA.training.train_starvla][INFO] - Step 5570, Loss: {'train/loss': 0.30917051434516907, 'train/grad_norm_pre_clip': 26.78409194946289, 'timing/data': 0.0011295275762677193, 'timing/model': 2.5702680060639977, 'learning_rate/qwen_vl_interface': 1.4717194043869953e-06, 'learning_rate/action_model': 1.4717194043869953e-05, 'epoch': 2.14}) +[2026-06-18 05:42:43,920][starVLA.training.train_starvla][INFO] - Step 5571, Loss: {'train/loss': 0.29171937704086304, 'train/grad_norm_pre_clip': 11.020660400390625, 'timing/data': 0.0011449754238128662, 'timing/model': 1.5593327013775706, 'learning_rate/qwen_vl_interface': 1.4704091000951982e-06, 'learning_rate/action_model': 1.4704091000951981e-05, 'epoch': 2.14}) +[2026-06-18 05:42:49,915][starVLA.training.train_starvla][INFO] - Step 5572, Loss: {'train/loss': 0.16431112587451935, 'train/grad_norm_pre_clip': 9.20005989074707, 'timing/data': 0.0013867970556020737, 'timing/model': 1.909296853467822, 'learning_rate/qwen_vl_interface': 1.4690995793166615e-06, 'learning_rate/action_model': 1.4690995793166616e-05, 'epoch': 2.14}) +[2026-06-18 05:43:03,186][starVLA.training.train_starvla][INFO] - Step 5573, Loss: {'train/loss': 0.20667770504951477, 'train/grad_norm_pre_clip': 9.854879379272461, 'timing/data': 0.0010790890082716942, 'timing/model': 4.300314809195697, 'learning_rate/qwen_vl_interface': 1.4677908423228506e-06, 'learning_rate/action_model': 1.4677908423228507e-05, 'epoch': 2.14}) +[2026-06-18 05:43:10,730][starVLA.training.train_starvla][INFO] - Step 5574, Loss: {'train/loss': 0.1276797652244568, 'train/grad_norm_pre_clip': 7.002164840698242, 'timing/data': 0.002881750464439392, 'timing/model': 1.1050356207415462, 'learning_rate/qwen_vl_interface': 1.4664828893850677e-06, 'learning_rate/action_model': 1.4664828893850676e-05, 'epoch': 2.14}) +[2026-06-18 05:43:16,280][starVLA.training.train_starvla][INFO] - Step 5575, Loss: {'train/loss': 0.5805763006210327, 'train/grad_norm_pre_clip': 25.67949104309082, 'timing/data': 0.0014373613521456718, 'timing/model': 1.1303101014345884, 'learning_rate/qwen_vl_interface': 1.4651757207744539e-06, 'learning_rate/action_model': 1.4651757207744537e-05, 'epoch': 2.14}) +[2026-06-18 05:43:21,236][starVLA.training.train_starvla][INFO] - Step 5576, Loss: {'train/loss': 0.37004995346069336, 'train/grad_norm_pre_clip': 16.944656372070312, 'timing/data': 0.001124848611652851, 'timing/model': 1.0779221672564745, 'learning_rate/qwen_vl_interface': 1.4638693367619837e-06, 'learning_rate/action_model': 1.4638693367619834e-05, 'epoch': 2.15}) +[2026-06-18 05:43:27,284][starVLA.training.train_starvla][INFO] - Step 5577, Loss: {'train/loss': 0.3020230531692505, 'train/grad_norm_pre_clip': 11.776461601257324, 'timing/data': 0.0019876649603247643, 'timing/model': 1.9087169645354152, 'learning_rate/qwen_vl_interface': 1.4625637376184768e-06, 'learning_rate/action_model': 1.4625637376184768e-05, 'epoch': 2.15}) +[2026-06-18 05:43:31,693][starVLA.training.train_starvla][INFO] - Step 5578, Loss: {'train/loss': 0.3397294878959656, 'train/grad_norm_pre_clip': 17.46099281311035, 'timing/data': 0.001138533465564251, 'timing/model': 1.1992761278524995, 'learning_rate/qwen_vl_interface': 1.4612589236145811e-06, 'learning_rate/action_model': 1.461258923614581e-05, 'epoch': 2.15}) +[2026-06-18 05:43:36,983][starVLA.training.train_starvla][INFO] - Step 5579, Loss: {'train/loss': 0.4701750874519348, 'train/grad_norm_pre_clip': 15.106586456298828, 'timing/data': 0.0009632119908928871, 'timing/model': 1.4788909023627639, 'learning_rate/qwen_vl_interface': 1.4599548950207878e-06, 'learning_rate/action_model': 1.4599548950207878e-05, 'epoch': 2.15}) +[2026-06-18 05:43:42,269][starVLA.training.train_starvla][INFO] - Step 5580, Loss: {'train/loss': 0.4551822245121002, 'train/grad_norm_pre_clip': 16.178701400756836, 'timing/data': 0.0013597672805190086, 'timing/model': 1.1116554969921708, 'learning_rate/qwen_vl_interface': 1.4586516521074214e-06, 'learning_rate/action_model': 1.4586516521074214e-05, 'epoch': 2.15}) +[2026-06-18 05:43:48,664][starVLA.training.train_starvla][INFO] - Step 5581, Loss: {'train/loss': 0.3614138960838318, 'train/grad_norm_pre_clip': 13.208377838134766, 'timing/data': 0.0010293414816260338, 'timing/model': 1.1363229900598526, 'learning_rate/qwen_vl_interface': 1.457349195144649e-06, 'learning_rate/action_model': 1.457349195144649e-05, 'epoch': 2.15}) +[2026-06-18 05:44:05,824][starVLA.training.train_starvla][INFO] - Step 5582, Loss: {'train/loss': 0.4233798682689667, 'train/grad_norm_pre_clip': 18.88269805908203, 'timing/data': 0.0011093467473983765, 'timing/model': 3.636705712415278, 'learning_rate/qwen_vl_interface': 1.4560475244024702e-06, 'learning_rate/action_model': 1.4560475244024702e-05, 'epoch': 2.15}) +[2026-06-18 05:44:11,447][starVLA.training.train_starvla][INFO] - Step 5583, Loss: {'train/loss': 0.4446154832839966, 'train/grad_norm_pre_clip': 14.63316822052002, 'timing/data': 0.0009401338174939156, 'timing/model': 1.3736431552097201, 'learning_rate/qwen_vl_interface': 1.4547466401507213e-06, 'learning_rate/action_model': 1.4547466401507211e-05, 'epoch': 2.15}) +[2026-06-18 05:44:16,883][starVLA.training.train_starvla][INFO] - Step 5584, Loss: {'train/loss': 0.3024926781654358, 'train/grad_norm_pre_clip': 14.804988861083984, 'timing/data': 0.0013943128287792206, 'timing/model': 1.1795385209843516, 'learning_rate/qwen_vl_interface': 1.4534465426590763e-06, 'learning_rate/action_model': 1.4534465426590763e-05, 'epoch': 2.15}) +[2026-06-18 05:44:21,706][starVLA.training.train_starvla][INFO] - Step 5585, Loss: {'train/loss': 0.5676693320274353, 'train/grad_norm_pre_clip': 17.285316467285156, 'timing/data': 0.0011835936456918716, 'timing/model': 1.1453399015590549, 'learning_rate/qwen_vl_interface': 1.4521472321970496e-06, 'learning_rate/action_model': 1.4521472321970494e-05, 'epoch': 2.15}) +[2026-06-18 05:44:28,938][starVLA.training.train_starvla][INFO] - Step 5586, Loss: {'train/loss': 0.3178793489933014, 'train/grad_norm_pre_clip': 14.260862350463867, 'timing/data': 0.0012377258390188217, 'timing/model': 1.6815836345776916, 'learning_rate/qwen_vl_interface': 1.450848709033988e-06, 'learning_rate/action_model': 1.4508487090339878e-05, 'epoch': 2.15}) +[2026-06-18 05:44:34,952][starVLA.training.train_starvla][INFO] - Step 5587, Loss: {'train/loss': 0.4699687957763672, 'train/grad_norm_pre_clip': 16.28446388244629, 'timing/data': 0.0011413991451263428, 'timing/model': 2.3577190721407533, 'learning_rate/qwen_vl_interface': 1.4495509734390769e-06, 'learning_rate/action_model': 1.4495509734390769e-05, 'epoch': 2.15}) +[2026-06-18 05:44:40,769][starVLA.training.train_starvla][INFO] - Step 5588, Loss: {'train/loss': 0.361845463514328, 'train/grad_norm_pre_clip': 15.430822372436523, 'timing/data': 0.001291540451347828, 'timing/model': 1.297935807146132, 'learning_rate/qwen_vl_interface': 1.448254025681339e-06, 'learning_rate/action_model': 1.448254025681339e-05, 'epoch': 2.15}) +[2026-06-18 05:44:46,227][starVLA.training.train_starvla][INFO] - Step 5589, Loss: {'train/loss': 0.5619688630104065, 'train/grad_norm_pre_clip': 24.136308670043945, 'timing/data': 0.001197746954858303, 'timing/model': 1.1010601902380586, 'learning_rate/qwen_vl_interface': 1.4469578660296327e-06, 'learning_rate/action_model': 1.4469578660296326e-05, 'epoch': 2.15}) +[2026-06-18 05:44:55,718][starVLA.training.train_starvla][INFO] - Step 5590, Loss: {'train/loss': 0.47665148973464966, 'train/grad_norm_pre_clip': 22.3773193359375, 'timing/data': 0.0015003923326730728, 'timing/model': 1.6170204803347588, 'learning_rate/qwen_vl_interface': 1.4456624947526524e-06, 'learning_rate/action_model': 1.4456624947526524e-05, 'epoch': 2.15}) +[2026-06-18 05:45:01,613][starVLA.training.train_starvla][INFO] - Step 5591, Loss: {'train/loss': 0.20469115674495697, 'train/grad_norm_pre_clip': 10.997054100036621, 'timing/data': 0.0011628838256001472, 'timing/model': 2.1155961705371737, 'learning_rate/qwen_vl_interface': 1.4443679121189306e-06, 'learning_rate/action_model': 1.4443679121189305e-05, 'epoch': 2.15}) +[2026-06-18 05:45:06,582][starVLA.training.train_starvla][INFO] - Step 5592, Loss: {'train/loss': 0.3204385042190552, 'train/grad_norm_pre_clip': 16.41551399230957, 'timing/data': 0.0013634897768497467, 'timing/model': 1.4640123136341572, 'learning_rate/qwen_vl_interface': 1.4430741183968361e-06, 'learning_rate/action_model': 1.443074118396836e-05, 'epoch': 2.15}) +[2026-06-18 05:45:12,117][starVLA.training.train_starvla][INFO] - Step 5593, Loss: {'train/loss': 0.32450419664382935, 'train/grad_norm_pre_clip': 14.768726348876953, 'timing/data': 0.0013036048039793968, 'timing/model': 1.2967656813561916, 'learning_rate/qwen_vl_interface': 1.441781113854573e-06, 'learning_rate/action_model': 1.4417811138545728e-05, 'epoch': 2.15}) +[2026-06-18 05:45:17,455][starVLA.training.train_starvla][INFO] - Step 5594, Loss: {'train/loss': 0.2883664667606354, 'train/grad_norm_pre_clip': 11.800884246826172, 'timing/data': 0.001670120283961296, 'timing/model': 1.4571020137518644, 'learning_rate/qwen_vl_interface': 1.4404888987601828e-06, 'learning_rate/action_model': 1.4404888987601828e-05, 'epoch': 2.15}) +[2026-06-18 05:45:27,668][starVLA.training.train_starvla][INFO] - Step 5595, Loss: {'train/loss': 0.47401493787765503, 'train/grad_norm_pre_clip': 22.500478744506836, 'timing/data': 0.0010748058557510376, 'timing/model': 2.5446902215480804, 'learning_rate/qwen_vl_interface': 1.4391974733815417e-06, 'learning_rate/action_model': 1.4391974733815416e-05, 'epoch': 2.15}) +[2026-06-18 05:45:33,115][starVLA.training.train_starvla][INFO] - Step 5596, Loss: {'train/loss': 0.44312602281570435, 'train/grad_norm_pre_clip': 20.08098793029785, 'timing/data': 0.0011701444163918495, 'timing/model': 1.1914449026808143, 'learning_rate/qwen_vl_interface': 1.437906837986368e-06, 'learning_rate/action_model': 1.4379068379863677e-05, 'epoch': 2.15}) +[2026-06-18 05:45:46,599][starVLA.training.train_starvla][INFO] - Step 5597, Loss: {'train/loss': 0.4911579191684723, 'train/grad_norm_pre_clip': 20.312454223632812, 'timing/data': 0.001128855161368847, 'timing/model': 1.6478875381872058, 'learning_rate/qwen_vl_interface': 1.436616992842207e-06, 'learning_rate/action_model': 1.4366169928422068e-05, 'epoch': 2.15}) +[2026-06-18 05:45:54,630][starVLA.training.train_starvla][INFO] - Step 5598, Loss: {'train/loss': 0.5071080923080444, 'train/grad_norm_pre_clip': 19.464183807373047, 'timing/data': 0.001254894770681858, 'timing/model': 1.2241893243044615, 'learning_rate/qwen_vl_interface': 1.4353279382164459e-06, 'learning_rate/action_model': 1.4353279382164458e-05, 'epoch': 2.15}) +[2026-06-18 05:46:08,526][starVLA.training.train_starvla][INFO] - Step 5599, Loss: {'train/loss': 0.39204734563827515, 'train/grad_norm_pre_clip': 14.004639625549316, 'timing/data': 0.0009573325514793396, 'timing/model': 1.525594742037356, 'learning_rate/qwen_vl_interface': 1.4340396743763084e-06, 'learning_rate/action_model': 1.4340396743763085e-05, 'epoch': 2.15}) +[2026-06-18 05:51:40,056][starVLA.training.train_starvla][INFO] - Step 5600, Loss: {'train/loss': 0.27468374371528625, 'train/grad_norm_pre_clip': 14.214699745178223, 'eval/loss': 0.41881507635116577, 'eval/latency_4/loss': 0.41881507635116577, 'eval/demon_fire_precision': 0.9644331533820261, 'eval/demon_fire_recall': 0.9863592075349139, 'eval/demon_fire_f1': 0.9752729608220937, 'eval/demon_fire_rate_model': 0.9248164464023495, 'eval/demon_fire_rate_teacher': 0.9042584434654919, 'eval/demon_left_precision': 0.7921052631578948, 'eval/demon_left_recall': 0.6872146118721462, 'eval/demon_left_f1': 0.7359413202933985, 'eval/demon_left_rate_model': 0.11160058737151249, 'eval/demon_left_rate_teacher': 0.12863436123348018, 'eval/demon_right_precision': 0.8201581027667985, 'eval/demon_right_recall': 0.7490974729241877, 'eval/demon_right_f1': 0.7830188679245285, 'eval/demon_right_rate_model': 0.1486049926578561, 'eval/demon_right_rate_teacher': 0.16270190895741557, 'eval/demon_move_f1': 0.7594800941089634, 'eval/demon_cc_f1': 0.8986705474710422, 'eval/latency_4/demon_fire_precision': 0.9644331533820261, 'eval/latency_4/demon_fire_recall': 0.9863592075349139, 'eval/latency_4/demon_fire_f1': 0.9752729608220937, 'eval/latency_4/demon_fire_rate_model': 0.9248164464023495, 'eval/latency_4/demon_fire_rate_teacher': 0.9042584434654919, 'eval/latency_4/demon_left_precision': 0.7921052631578948, 'eval/latency_4/demon_left_recall': 0.6872146118721462, 'eval/latency_4/demon_left_f1': 0.7359413202933985, 'eval/latency_4/demon_left_rate_model': 0.11160058737151249, 'eval/latency_4/demon_left_rate_teacher': 0.12863436123348018, 'eval/latency_4/demon_right_precision': 0.8201581027667985, 'eval/latency_4/demon_right_recall': 0.7490974729241877, 'eval/latency_4/demon_right_f1': 0.7830188679245285, 'eval/latency_4/demon_right_rate_model': 0.1486049926578561, 'eval/latency_4/demon_right_rate_teacher': 0.16270190895741557, 'eval/latency_4/demon_move_f1': 0.7594800941089634, 'eval/latency_4/demon_cc_f1': 0.8986705474710422, 'timing/data': 0.0012066764757037163, 'timing/model': 1.1232218500226736, 'learning_rate/qwen_vl_interface': 1.4327522015888535e-06, 'learning_rate/action_model': 1.4327522015888534e-05, 'epoch': 2.15}) +[2026-06-18 05:51:40,058][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_5600_state +[2026-06-18 05:51:40,064][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 05:51:47,746][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_5600_state/model.safetensors +[2026-06-18 05:52:06,447][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_5600_state/optimizer.bin +[2026-06-18 05:52:06,449][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_5600_state/sampler.bin +[2026-06-18 05:52:06,450][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_5600_state/sampler_1.bin +[2026-06-18 05:52:06,452][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_5600_state/random_states_0.pkl +[2026-06-18 05:52:06,457][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 5600 +[2026-06-18 05:52:06,457][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 05:52:06,477][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 05:52:10,887][starVLA.training.train_starvla][INFO] - Step 5601, Loss: {'train/loss': 0.44554248452186584, 'train/grad_norm_pre_clip': 16.516469955444336, 'timing/data': 0.0008702855557203293, 'timing/model': 1.3628081884235144, 'learning_rate/qwen_vl_interface': 1.4314655201209756e-06, 'learning_rate/action_model': 1.4314655201209757e-05, 'epoch': 2.15}) +[2026-06-18 05:52:14,952][starVLA.training.train_starvla][INFO] - Step 5602, Loss: {'train/loss': 0.4912157952785492, 'train/grad_norm_pre_clip': 17.973058700561523, 'timing/data': 0.0010889898985624313, 'timing/model': 1.032697157934308, 'learning_rate/qwen_vl_interface': 1.4301796302394019e-06, 'learning_rate/action_model': 1.4301796302394018e-05, 'epoch': 2.16}) +[2026-06-18 05:52:20,859][starVLA.training.train_starvla][INFO] - Step 5603, Loss: {'train/loss': 0.407090961933136, 'train/grad_norm_pre_clip': 13.098564147949219, 'timing/data': 0.0015040654689073563, 'timing/model': 1.101143260486424, 'learning_rate/qwen_vl_interface': 1.4288945322107025e-06, 'learning_rate/action_model': 1.4288945322107026e-05, 'epoch': 2.16}) +[2026-06-18 05:52:32,436][starVLA.training.train_starvla][INFO] - Step 5604, Loss: {'train/loss': 0.3646114766597748, 'train/grad_norm_pre_clip': 18.770854949951172, 'timing/data': 0.001613035798072815, 'timing/model': 1.5575655000284314, 'learning_rate/qwen_vl_interface': 1.4276102263012781e-06, 'learning_rate/action_model': 1.427610226301278e-05, 'epoch': 2.16}) +[2026-06-18 05:52:38,736][starVLA.training.train_starvla][INFO] - Step 5605, Loss: {'train/loss': 0.29303595423698425, 'train/grad_norm_pre_clip': 8.59133529663086, 'timing/data': 0.001710161566734314, 'timing/model': 1.5710200378671288, 'learning_rate/qwen_vl_interface': 1.426326712777367e-06, 'learning_rate/action_model': 1.4263267127773669e-05, 'epoch': 2.16}) +[2026-06-18 05:52:49,114][starVLA.training.train_starvla][INFO] - Step 5606, Loss: {'train/loss': 0.41184523701667786, 'train/grad_norm_pre_clip': 22.664316177368164, 'timing/data': 0.001263277605175972, 'timing/model': 5.6135472962632775, 'learning_rate/qwen_vl_interface': 1.425043991905043e-06, 'learning_rate/action_model': 1.425043991905043e-05, 'epoch': 2.16}) +[2026-06-18 05:52:55,484][starVLA.training.train_starvla][INFO] - Step 5607, Loss: {'train/loss': 0.5836884379386902, 'train/grad_norm_pre_clip': 22.203813552856445, 'timing/data': 0.0009562727063894272, 'timing/model': 1.791169023141265, 'learning_rate/qwen_vl_interface': 1.4237620639502149e-06, 'learning_rate/action_model': 1.4237620639502147e-05, 'epoch': 2.16}) +[2026-06-18 05:53:13,296][starVLA.training.train_starvla][INFO] - Step 5608, Loss: {'train/loss': 0.4626353681087494, 'train/grad_norm_pre_clip': 15.982380867004395, 'timing/data': 0.0012780539691448212, 'timing/model': 1.675747686997056, 'learning_rate/qwen_vl_interface': 1.4224809291786281e-06, 'learning_rate/action_model': 1.4224809291786282e-05, 'epoch': 2.16}) +[2026-06-18 05:53:20,834][starVLA.training.train_starvla][INFO] - Step 5609, Loss: {'train/loss': 0.11986835300922394, 'train/grad_norm_pre_clip': 7.012231349945068, 'timing/data': 0.0038126762956380844, 'timing/model': 1.4248962346464396, 'learning_rate/qwen_vl_interface': 1.4212005878558622e-06, 'learning_rate/action_model': 1.4212005878558622e-05, 'epoch': 2.16}) +[2026-06-18 05:53:26,953][starVLA.training.train_starvla][INFO] - Step 5610, Loss: {'train/loss': 0.2939293682575226, 'train/grad_norm_pre_clip': 12.125457763671875, 'timing/data': 0.0010393168777227402, 'timing/model': 1.5050055142492056, 'learning_rate/qwen_vl_interface': 1.4199210402473368e-06, 'learning_rate/action_model': 1.4199210402473367e-05, 'epoch': 2.16}) +[2026-06-18 05:53:32,735][starVLA.training.train_starvla][INFO] - Step 5611, Loss: {'train/loss': 0.3346729874610901, 'train/grad_norm_pre_clip': 14.05160140991211, 'timing/data': 0.0013918448239564896, 'timing/model': 1.4923150585964322, 'learning_rate/qwen_vl_interface': 1.4186422866183e-06, 'learning_rate/action_model': 1.4186422866182999e-05, 'epoch': 2.16}) +[2026-06-18 05:53:44,861][starVLA.training.train_starvla][INFO] - Step 5612, Loss: {'train/loss': 0.40754228830337524, 'train/grad_norm_pre_clip': 14.49225902557373, 'timing/data': 0.0013791285455226898, 'timing/model': 1.5900751985609531, 'learning_rate/qwen_vl_interface': 1.41736432723384e-06, 'learning_rate/action_model': 1.41736432723384e-05, 'epoch': 2.16}) +[2026-06-18 05:53:54,261][starVLA.training.train_starvla][INFO] - Step 5613, Loss: {'train/loss': 0.18112699687480927, 'train/grad_norm_pre_clip': 9.288272857666016, 'timing/data': 0.0010386109352111816, 'timing/model': 1.2865171236917377, 'learning_rate/qwen_vl_interface': 1.4160871623588774e-06, 'learning_rate/action_model': 1.4160871623588772e-05, 'epoch': 2.16}) +[2026-06-18 05:54:06,470][starVLA.training.train_starvla][INFO] - Step 5614, Loss: {'train/loss': 0.15211330354213715, 'train/grad_norm_pre_clip': 10.497554779052734, 'timing/data': 0.002346809022128582, 'timing/model': 5.934796416200697, 'learning_rate/qwen_vl_interface': 1.414810792258173e-06, 'learning_rate/action_model': 1.414810792258173e-05, 'epoch': 2.16}) +[2026-06-18 05:54:14,958][starVLA.training.train_starvla][INFO] - Step 5615, Loss: {'train/loss': 0.2911924421787262, 'train/grad_norm_pre_clip': 11.407279968261719, 'timing/data': 0.001522325910627842, 'timing/model': 1.1198727497830987, 'learning_rate/qwen_vl_interface': 1.4135352171963186e-06, 'learning_rate/action_model': 1.4135352171963185e-05, 'epoch': 2.16}) +[2026-06-18 05:54:24,486][starVLA.training.train_starvla][INFO] - Step 5616, Loss: {'train/loss': 0.09404616802930832, 'train/grad_norm_pre_clip': 5.654796123504639, 'timing/data': 0.0016668159514665604, 'timing/model': 5.139122681692243, 'learning_rate/qwen_vl_interface': 1.412260437437741e-06, 'learning_rate/action_model': 1.412260437437741e-05, 'epoch': 2.16}) +[2026-06-18 05:54:30,778][starVLA.training.train_starvla][INFO] - Step 5617, Loss: {'train/loss': 0.42131179571151733, 'train/grad_norm_pre_clip': 18.61785125732422, 'timing/data': 0.0011091670021414757, 'timing/model': 2.386297671124339, 'learning_rate/qwen_vl_interface': 1.4109864532467051e-06, 'learning_rate/action_model': 1.4109864532467052e-05, 'epoch': 2.16}) +[2026-06-18 05:54:36,477][starVLA.training.train_starvla][INFO] - Step 5618, Loss: {'train/loss': 0.12182629853487015, 'train/grad_norm_pre_clip': 7.969577312469482, 'timing/data': 0.0015773214399814606, 'timing/model': 1.4951856452971697, 'learning_rate/qwen_vl_interface': 1.4097132648873082e-06, 'learning_rate/action_model': 1.4097132648873083e-05, 'epoch': 2.16}) +[2026-06-18 05:54:43,562][starVLA.training.train_starvla][INFO] - Step 5619, Loss: {'train/loss': 0.4552322030067444, 'train/grad_norm_pre_clip': 34.63185501098633, 'timing/data': 0.0012715039774775505, 'timing/model': 1.581142794340849, 'learning_rate/qwen_vl_interface': 1.4084408726234838e-06, 'learning_rate/action_model': 1.4084408726234838e-05, 'epoch': 2.16}) +[2026-06-18 05:54:55,786][starVLA.training.train_starvla][INFO] - Step 5620, Loss: {'train/loss': 0.18625840544700623, 'train/grad_norm_pre_clip': 9.527488708496094, 'timing/data': 0.0012335488572716713, 'timing/model': 1.182967715896666, 'learning_rate/qwen_vl_interface': 1.407169276719e-06, 'learning_rate/action_model': 1.4071692767190001e-05, 'epoch': 2.16}) +[2026-06-18 05:55:02,664][starVLA.training.train_starvla][INFO] - Step 5621, Loss: {'train/loss': 0.3043522536754608, 'train/grad_norm_pre_clip': 15.757412910461426, 'timing/data': 0.0008510332554578781, 'timing/model': 3.151333637535572, 'learning_rate/qwen_vl_interface': 1.4058984774374606e-06, 'learning_rate/action_model': 1.4058984774374607e-05, 'epoch': 2.16}) +[2026-06-18 05:55:07,715][starVLA.training.train_starvla][INFO] - Step 5622, Loss: {'train/loss': 0.14493420720100403, 'train/grad_norm_pre_clip': 8.320704460144043, 'timing/data': 0.0013236654922366142, 'timing/model': 1.5835113609209657, 'learning_rate/qwen_vl_interface': 1.404628475042303e-06, 'learning_rate/action_model': 1.404628475042303e-05, 'epoch': 2.16}) +[2026-06-18 05:55:13,187][starVLA.training.train_starvla][INFO] - Step 5623, Loss: {'train/loss': 0.22532138228416443, 'train/grad_norm_pre_clip': 15.966546058654785, 'timing/data': 0.004574242047965527, 'timing/model': 1.3769929464906454, 'learning_rate/qwen_vl_interface': 1.4033592697967998e-06, 'learning_rate/action_model': 1.4033592697967998e-05, 'epoch': 2.16}) +[2026-06-18 05:55:22,214][starVLA.training.train_starvla][INFO] - Step 5624, Loss: {'train/loss': 0.31355732679367065, 'train/grad_norm_pre_clip': 12.256592750549316, 'timing/data': 0.0012701433151960373, 'timing/model': 1.1652607843279839, 'learning_rate/qwen_vl_interface': 1.4020908619640602e-06, 'learning_rate/action_model': 1.4020908619640602e-05, 'epoch': 2.16}) +[2026-06-18 05:55:32,084][starVLA.training.train_starvla][INFO] - Step 5625, Loss: {'train/loss': 0.5268183946609497, 'train/grad_norm_pre_clip': 29.208425521850586, 'timing/data': 0.0015204092487692833, 'timing/model': 1.487366703338921, 'learning_rate/qwen_vl_interface': 1.4008232518070248e-06, 'learning_rate/action_model': 1.4008232518070247e-05, 'epoch': 2.16}) +[2026-06-18 05:55:37,225][starVLA.training.train_starvla][INFO] - Step 5626, Loss: {'train/loss': 0.5102179646492004, 'train/grad_norm_pre_clip': 17.62555503845215, 'timing/data': 0.004739650525152683, 'timing/model': 1.3187202839180827, 'learning_rate/qwen_vl_interface': 1.3995564395884712e-06, 'learning_rate/action_model': 1.399556439588471e-05, 'epoch': 2.16}) +[2026-06-18 05:55:45,190][starVLA.training.train_starvla][INFO] - Step 5627, Loss: {'train/loss': 0.34635743498802185, 'train/grad_norm_pre_clip': 16.515472412109375, 'timing/data': 0.0011024121195077896, 'timing/model': 2.824613275937736, 'learning_rate/qwen_vl_interface': 1.3982904255710094e-06, 'learning_rate/action_model': 1.3982904255710094e-05, 'epoch': 2.16}) +[2026-06-18 05:56:01,760][starVLA.training.train_starvla][INFO] - Step 5628, Loss: {'train/loss': 0.2553088366985321, 'train/grad_norm_pre_clip': 16.140167236328125, 'timing/data': 0.0033194152638316154, 'timing/model': 2.0447677560150623, 'learning_rate/qwen_vl_interface': 1.3970252100170892e-06, 'learning_rate/action_model': 1.3970252100170892e-05, 'epoch': 2.17}) +[2026-06-18 05:56:13,667][starVLA.training.train_starvla][INFO] - Step 5629, Loss: {'train/loss': 0.3256406784057617, 'train/grad_norm_pre_clip': 13.55745792388916, 'timing/data': 0.0010798033326864243, 'timing/model': 5.506331321783364, 'learning_rate/qwen_vl_interface': 1.3957607931889885e-06, 'learning_rate/action_model': 1.3957607931889884e-05, 'epoch': 2.17}) +[2026-06-18 05:56:22,948][starVLA.training.train_starvla][INFO] - Step 5630, Loss: {'train/loss': 0.348184198141098, 'train/grad_norm_pre_clip': 14.953571319580078, 'timing/data': 0.0010723071172833443, 'timing/model': 1.3321498706936836, 'learning_rate/qwen_vl_interface': 1.3944971753488238e-06, 'learning_rate/action_model': 1.3944971753488236e-05, 'epoch': 2.17}) +[2026-06-18 05:56:29,688][starVLA.training.train_starvla][INFO] - Step 5631, Loss: {'train/loss': 0.16344721615314484, 'train/grad_norm_pre_clip': 32.333030700683594, 'timing/data': 0.0011283792555332184, 'timing/model': 1.9056494887918234, 'learning_rate/qwen_vl_interface': 1.3932343567585408e-06, 'learning_rate/action_model': 1.3932343567585407e-05, 'epoch': 2.17}) +[2026-06-18 05:56:35,169][starVLA.training.train_starvla][INFO] - Step 5632, Loss: {'train/loss': 0.5734438896179199, 'train/grad_norm_pre_clip': 17.549053192138672, 'timing/data': 0.00120525062084198, 'timing/model': 1.6731895077973604, 'learning_rate/qwen_vl_interface': 1.3919723376799268e-06, 'learning_rate/action_model': 1.3919723376799269e-05, 'epoch': 2.17}) +[2026-06-18 05:56:53,551][starVLA.training.train_starvla][INFO] - Step 5633, Loss: {'train/loss': 0.3014398217201233, 'train/grad_norm_pre_clip': 15.686944961547852, 'timing/data': 0.0009213294833898544, 'timing/model': 3.9080871511250734, 'learning_rate/qwen_vl_interface': 1.3907111183745992e-06, 'learning_rate/action_model': 1.3907111183745991e-05, 'epoch': 2.17}) +[2026-06-18 05:57:02,785][starVLA.training.train_starvla][INFO] - Step 5634, Loss: {'train/loss': 0.474857896566391, 'train/grad_norm_pre_clip': 17.73412322998047, 'timing/data': 0.0015426278114318848, 'timing/model': 1.3011737056076527, 'learning_rate/qwen_vl_interface': 1.3894506991040089e-06, 'learning_rate/action_model': 1.3894506991040087e-05, 'epoch': 2.17}) +[2026-06-18 05:57:11,110][starVLA.training.train_starvla][INFO] - Step 5635, Loss: {'train/loss': 0.25549429655075073, 'train/grad_norm_pre_clip': 10.90589714050293, 'timing/data': 0.0016700448468327522, 'timing/model': 4.184048066847026, 'learning_rate/qwen_vl_interface': 1.3881910801294433e-06, 'learning_rate/action_model': 1.3881910801294431e-05, 'epoch': 2.17}) +[2026-06-18 05:57:18,438][starVLA.training.train_starvla][INFO] - Step 5636, Loss: {'train/loss': 0.38662487268447876, 'train/grad_norm_pre_clip': 13.425642967224121, 'timing/data': 0.0008948817849159241, 'timing/model': 2.975682624615729, 'learning_rate/qwen_vl_interface': 1.3869322617120223e-06, 'learning_rate/action_model': 1.3869322617120222e-05, 'epoch': 2.17}) +[2026-06-18 05:57:24,118][starVLA.training.train_starvla][INFO] - Step 5637, Loss: {'train/loss': 0.2367779016494751, 'train/grad_norm_pre_clip': 13.297194480895996, 'timing/data': 0.0009626951068639755, 'timing/model': 1.8300619572401047, 'learning_rate/qwen_vl_interface': 1.3856742441127e-06, 'learning_rate/action_model': 1.3856742441126999e-05, 'epoch': 2.17}) +[2026-06-18 05:57:35,129][starVLA.training.train_starvla][INFO] - Step 5638, Loss: {'train/loss': 0.42287683486938477, 'train/grad_norm_pre_clip': 23.235368728637695, 'timing/data': 0.0009718546643853188, 'timing/model': 5.611417935229838, 'learning_rate/qwen_vl_interface': 1.384417027592264e-06, 'learning_rate/action_model': 1.3844170275922638e-05, 'epoch': 2.17}) +[2026-06-18 05:57:42,758][starVLA.training.train_starvla][INFO] - Step 5639, Loss: {'train/loss': 0.19788528978824615, 'train/grad_norm_pre_clip': 11.969196319580078, 'timing/data': 0.001529669389128685, 'timing/model': 1.535699501633644, 'learning_rate/qwen_vl_interface': 1.3831606124113397e-06, 'learning_rate/action_model': 1.3831606124113398e-05, 'epoch': 2.17}) +[2026-06-18 05:57:50,121][starVLA.training.train_starvla][INFO] - Step 5640, Loss: {'train/loss': 0.20780403912067413, 'train/grad_norm_pre_clip': 11.225836753845215, 'timing/data': 0.0015817508101463318, 'timing/model': 1.1664082203060389, 'learning_rate/qwen_vl_interface': 1.3819049988303809e-06, 'learning_rate/action_model': 1.381904998830381e-05, 'epoch': 2.17}) +[2026-06-18 05:57:59,331][starVLA.training.train_starvla][INFO] - Step 5641, Loss: {'train/loss': 0.17788901925086975, 'train/grad_norm_pre_clip': 8.260632514953613, 'timing/data': 0.0015300260856747627, 'timing/model': 4.938151832669973, 'learning_rate/qwen_vl_interface': 1.3806501871096774e-06, 'learning_rate/action_model': 1.3806501871096774e-05, 'epoch': 2.17}) +[2026-06-18 05:58:12,618][starVLA.training.train_starvla][INFO] - Step 5642, Loss: {'train/loss': 0.34213149547576904, 'train/grad_norm_pre_clip': 12.692095756530762, 'timing/data': 0.0015070345252752304, 'timing/model': 2.9606370124965906, 'learning_rate/qwen_vl_interface': 1.3793961775093528e-06, 'learning_rate/action_model': 1.3793961775093528e-05, 'epoch': 2.17}) +[2026-06-18 05:58:18,063][starVLA.training.train_starvla][INFO] - Step 5643, Loss: {'train/loss': 0.23558801412582397, 'train/grad_norm_pre_clip': 15.457737922668457, 'timing/data': 0.0012677060440182686, 'timing/model': 1.227101099677384, 'learning_rate/qwen_vl_interface': 1.378142970289367e-06, 'learning_rate/action_model': 1.378142970289367e-05, 'epoch': 2.17}) +[2026-06-18 05:58:24,092][starVLA.training.train_starvla][INFO] - Step 5644, Loss: {'train/loss': 0.21197374165058136, 'train/grad_norm_pre_clip': 15.130899429321289, 'timing/data': 0.0011763274669647217, 'timing/model': 1.169694984331727, 'learning_rate/qwen_vl_interface': 1.3768905657095109e-06, 'learning_rate/action_model': 1.3768905657095107e-05, 'epoch': 2.17}) +[2026-06-18 05:58:29,125][starVLA.training.train_starvla][INFO] - Step 5645, Loss: {'train/loss': 0.31795936822891235, 'train/grad_norm_pre_clip': 13.732194900512695, 'timing/data': 0.0008280063048005104, 'timing/model': 1.1678100945428014, 'learning_rate/qwen_vl_interface': 1.3756389640294066e-06, 'learning_rate/action_model': 1.3756389640294064e-05, 'epoch': 2.17}) +[2026-06-18 05:58:37,283][starVLA.training.train_starvla][INFO] - Step 5646, Loss: {'train/loss': 0.287210077047348, 'train/grad_norm_pre_clip': 15.729073524475098, 'timing/data': 0.0011183209717273712, 'timing/model': 1.6115169329568744, 'learning_rate/qwen_vl_interface': 1.3743881655085131e-06, 'learning_rate/action_model': 1.3743881655085131e-05, 'epoch': 2.17}) +[2026-06-18 05:58:42,703][starVLA.training.train_starvla][INFO] - Step 5647, Loss: {'train/loss': 0.13974091410636902, 'train/grad_norm_pre_clip': 14.51839828491211, 'timing/data': 0.0013514608144760132, 'timing/model': 1.5460751745849848, 'learning_rate/qwen_vl_interface': 1.3731381704061248e-06, 'learning_rate/action_model': 1.3731381704061246e-05, 'epoch': 2.17}) +[2026-06-18 05:58:49,875][starVLA.training.train_starvla][INFO] - Step 5648, Loss: {'train/loss': 0.18714331090450287, 'train/grad_norm_pre_clip': 11.489253044128418, 'timing/data': 0.0013359738513827324, 'timing/model': 1.4272925769910216, 'learning_rate/qwen_vl_interface': 1.3718889789813647e-06, 'learning_rate/action_model': 1.3718889789813646e-05, 'epoch': 2.17}) +[2026-06-18 05:58:55,239][starVLA.training.train_starvla][INFO] - Step 5649, Loss: {'train/loss': 0.1790192574262619, 'train/grad_norm_pre_clip': 13.185050964355469, 'timing/data': 0.001224147155880928, 'timing/model': 1.09315044619143, 'learning_rate/qwen_vl_interface': 1.370640591493193e-06, 'learning_rate/action_model': 1.370640591493193e-05, 'epoch': 2.17}) +[2026-06-18 05:59:01,839][starVLA.training.train_starvla][INFO] - Step 5650, Loss: {'train/loss': 0.2347821593284607, 'train/grad_norm_pre_clip': 18.763856887817383, 'timing/data': 0.002373090013861656, 'timing/model': 1.328218849375844, 'learning_rate/qwen_vl_interface': 1.3693930082004006e-06, 'learning_rate/action_model': 1.3693930082004006e-05, 'epoch': 2.17}) +[2026-06-18 05:59:08,170][starVLA.training.train_starvla][INFO] - Step 5651, Loss: {'train/loss': 0.1927453875541687, 'train/grad_norm_pre_clip': 13.443951606750488, 'timing/data': 0.0011549117043614388, 'timing/model': 1.3128981050103903, 'learning_rate/qwen_vl_interface': 1.3681462293616131e-06, 'learning_rate/action_model': 1.3681462293616131e-05, 'epoch': 2.17}) +[2026-06-18 05:59:13,756][starVLA.training.train_starvla][INFO] - Step 5652, Loss: {'train/loss': 0.39410823583602905, 'train/grad_norm_pre_clip': 18.385730743408203, 'timing/data': 0.0010544303804636002, 'timing/model': 1.768669635988772, 'learning_rate/qwen_vl_interface': 1.3669002552352898e-06, 'learning_rate/action_model': 1.3669002552352897e-05, 'epoch': 2.17}) +[2026-06-18 05:59:19,253][starVLA.training.train_starvla][INFO] - Step 5653, Loss: {'train/loss': 0.1606118530035019, 'train/grad_norm_pre_clip': 13.404696464538574, 'timing/data': 0.0011723199859261513, 'timing/model': 1.668685763143003, 'learning_rate/qwen_vl_interface': 1.3656550860797223e-06, 'learning_rate/action_model': 1.3656550860797223e-05, 'epoch': 2.17}) +[2026-06-18 05:59:30,826][starVLA.training.train_starvla][INFO] - Step 5654, Loss: {'train/loss': 0.1857801079750061, 'train/grad_norm_pre_clip': 7.482469081878662, 'timing/data': 0.0013037500903010368, 'timing/model': 5.783284965902567, 'learning_rate/qwen_vl_interface': 1.3644107221530344e-06, 'learning_rate/action_model': 1.3644107221530345e-05, 'epoch': 2.18}) +[2026-06-18 05:59:45,157][starVLA.training.train_starvla][INFO] - Step 5655, Loss: {'train/loss': 0.23022016882896423, 'train/grad_norm_pre_clip': 13.920116424560547, 'timing/data': 0.0009857024997472763, 'timing/model': 1.1184488348662853, 'learning_rate/qwen_vl_interface': 1.3631671637131856e-06, 'learning_rate/action_model': 1.3631671637131855e-05, 'epoch': 2.18}) +[2026-06-18 05:59:50,410][starVLA.training.train_starvla][INFO] - Step 5656, Loss: {'train/loss': 0.6836831569671631, 'train/grad_norm_pre_clip': 19.99378204345703, 'timing/data': 0.002332218922674656, 'timing/model': 1.1798685537651181, 'learning_rate/qwen_vl_interface': 1.3619244110179655e-06, 'learning_rate/action_model': 1.3619244110179654e-05, 'epoch': 2.18}) +[2026-06-18 05:59:55,711][starVLA.training.train_starvla][INFO] - Step 5657, Loss: {'train/loss': 0.16077400743961334, 'train/grad_norm_pre_clip': 9.187246322631836, 'timing/data': 0.0015031155198812485, 'timing/model': 1.5884603699669242, 'learning_rate/qwen_vl_interface': 1.360682464324998e-06, 'learning_rate/action_model': 1.360682464324998e-05, 'epoch': 2.18}) +[2026-06-18 06:00:00,725][starVLA.training.train_starvla][INFO] - Step 5658, Loss: {'train/loss': 0.287141352891922, 'train/grad_norm_pre_clip': 11.904955863952637, 'timing/data': 0.0012706564739346504, 'timing/model': 1.3406515568494797, 'learning_rate/qwen_vl_interface': 1.3594413238917425e-06, 'learning_rate/action_model': 1.3594413238917425e-05, 'epoch': 2.18}) +[2026-06-18 06:00:12,944][starVLA.training.train_starvla][INFO] - Step 5659, Loss: {'train/loss': 0.449348509311676, 'train/grad_norm_pre_clip': 17.53890037536621, 'timing/data': 0.0019999751821160316, 'timing/model': 1.4963584328070283, 'learning_rate/qwen_vl_interface': 1.3582009899754864e-06, 'learning_rate/action_model': 1.3582009899754864e-05, 'epoch': 2.18}) +[2026-06-18 06:00:23,948][starVLA.training.train_starvla][INFO] - Step 5660, Loss: {'train/loss': 0.2944714426994324, 'train/grad_norm_pre_clip': 12.184226989746094, 'timing/data': 0.0009520528838038445, 'timing/model': 4.957302917726338, 'learning_rate/qwen_vl_interface': 1.3569614628333515e-06, 'learning_rate/action_model': 1.3569614628333516e-05, 'epoch': 2.18}) +[2026-06-18 06:00:38,352][starVLA.training.train_starvla][INFO] - Step 5661, Loss: {'train/loss': 0.5247385501861572, 'train/grad_norm_pre_clip': 21.012990951538086, 'timing/data': 0.0010914262384176254, 'timing/model': 1.8812830531969666, 'learning_rate/qwen_vl_interface': 1.3557227427222954e-06, 'learning_rate/action_model': 1.3557227427222954e-05, 'epoch': 2.18}) +[2026-06-18 06:00:43,478][starVLA.training.train_starvla][INFO] - Step 5662, Loss: {'train/loss': 0.47267088294029236, 'train/grad_norm_pre_clip': 16.600521087646484, 'timing/data': 0.0012648357078433037, 'timing/model': 1.7975254701450467, 'learning_rate/qwen_vl_interface': 1.3544848298991056e-06, 'learning_rate/action_model': 1.3544848298991056e-05, 'epoch': 2.18}) +[2026-06-18 06:00:48,574][starVLA.training.train_starvla][INFO] - Step 5663, Loss: {'train/loss': 0.241799995303154, 'train/grad_norm_pre_clip': 11.608973503112793, 'timing/data': 0.0010746018961071968, 'timing/model': 1.1377971628680825, 'learning_rate/qwen_vl_interface': 1.3532477246204037e-06, 'learning_rate/action_model': 1.3532477246204036e-05, 'epoch': 2.18}) +[2026-06-18 06:00:55,344][starVLA.training.train_starvla][INFO] - Step 5664, Loss: {'train/loss': 0.3954094350337982, 'train/grad_norm_pre_clip': 14.727062225341797, 'timing/data': 0.001997102051973343, 'timing/model': 1.2142167389392853, 'learning_rate/qwen_vl_interface': 1.3520114271426382e-06, 'learning_rate/action_model': 1.3520114271426381e-05, 'epoch': 2.18}) +[2026-06-18 06:01:01,853][starVLA.training.train_starvla][INFO] - Step 5665, Loss: {'train/loss': 0.3597242534160614, 'train/grad_norm_pre_clip': 15.680623054504395, 'timing/data': 0.0013595428317785263, 'timing/model': 1.7535886885598302, 'learning_rate/qwen_vl_interface': 1.3507759377221001e-06, 'learning_rate/action_model': 1.3507759377221003e-05, 'epoch': 2.18}) +[2026-06-18 06:01:11,752][starVLA.training.train_starvla][INFO] - Step 5666, Loss: {'train/loss': 0.18193191289901733, 'train/grad_norm_pre_clip': 13.99154281616211, 'timing/data': 0.0014576688408851624, 'timing/model': 1.5879073506221175, 'learning_rate/qwen_vl_interface': 1.3495412566149058e-06, 'learning_rate/action_model': 1.3495412566149057e-05, 'epoch': 2.18}) +[2026-06-18 06:01:17,307][starVLA.training.train_starvla][INFO] - Step 5667, Loss: {'train/loss': 0.44712698459625244, 'train/grad_norm_pre_clip': 15.737802505493164, 'timing/data': 0.0026750462129712105, 'timing/model': 1.4818931063637137, 'learning_rate/qwen_vl_interface': 1.3483073840770067e-06, 'learning_rate/action_model': 1.3483073840770066e-05, 'epoch': 2.18}) +[2026-06-18 06:01:26,828][starVLA.training.train_starvla][INFO] - Step 5668, Loss: {'train/loss': 0.5587372183799744, 'train/grad_norm_pre_clip': 23.641145706176758, 'timing/data': 0.001547020860016346, 'timing/model': 4.936343512497842, 'learning_rate/qwen_vl_interface': 1.3470743203641846e-06, 'learning_rate/action_model': 1.3470743203641844e-05, 'epoch': 2.18}) +[2026-06-18 06:01:32,183][starVLA.training.train_starvla][INFO] - Step 5669, Loss: {'train/loss': 0.21691426634788513, 'train/grad_norm_pre_clip': 13.728379249572754, 'timing/data': 0.0011663641780614853, 'timing/model': 1.1568396175280213, 'learning_rate/qwen_vl_interface': 1.3458420657320555e-06, 'learning_rate/action_model': 1.3458420657320555e-05, 'epoch': 2.18}) +[2026-06-18 06:01:41,170][starVLA.training.train_starvla][INFO] - Step 5670, Loss: {'train/loss': 0.24117444455623627, 'train/grad_norm_pre_clip': 16.462507247924805, 'timing/data': 0.0010699154809117317, 'timing/model': 4.806441561318934, 'learning_rate/qwen_vl_interface': 1.344610620436067e-06, 'learning_rate/action_model': 1.344610620436067e-05, 'epoch': 2.18}) +[2026-06-18 06:01:46,592][starVLA.training.train_starvla][INFO] - Step 5671, Loss: {'train/loss': 0.5164578557014465, 'train/grad_norm_pre_clip': 24.83917808532715, 'timing/data': 0.0010174596682190895, 'timing/model': 1.2346056327223778, 'learning_rate/qwen_vl_interface': 1.3433799847314989e-06, 'learning_rate/action_model': 1.3433799847314988e-05, 'epoch': 2.18}) +[2026-06-18 06:01:53,585][starVLA.training.train_starvla][INFO] - Step 5672, Loss: {'train/loss': 0.22102251648902893, 'train/grad_norm_pre_clip': 13.106867790222168, 'timing/data': 0.0014125416055321693, 'timing/model': 1.4653780590742826, 'learning_rate/qwen_vl_interface': 1.3421501588734655e-06, 'learning_rate/action_model': 1.3421501588734656e-05, 'epoch': 2.18}) +[2026-06-18 06:01:58,229][starVLA.training.train_starvla][INFO] - Step 5673, Loss: {'train/loss': 0.5629549026489258, 'train/grad_norm_pre_clip': 25.1863956451416, 'timing/data': 0.0012793373316526413, 'timing/model': 1.328767778351903, 'learning_rate/qwen_vl_interface': 1.3409211431169084e-06, 'learning_rate/action_model': 1.3409211431169083e-05, 'epoch': 2.18}) +[2026-06-18 06:02:03,068][starVLA.training.train_starvla][INFO] - Step 5674, Loss: {'train/loss': 0.3086133599281311, 'train/grad_norm_pre_clip': 18.042659759521484, 'timing/data': 0.003342130221426487, 'timing/model': 1.1143211890012026, 'learning_rate/qwen_vl_interface': 1.3396929377166042e-06, 'learning_rate/action_model': 1.3396929377166043e-05, 'epoch': 2.18}) +[2026-06-18 06:02:08,855][starVLA.training.train_starvla][INFO] - Step 5675, Loss: {'train/loss': 0.4924982190132141, 'train/grad_norm_pre_clip': 16.994651794433594, 'timing/data': 0.001056223176419735, 'timing/model': 1.452144023962319, 'learning_rate/qwen_vl_interface': 1.338465542927161e-06, 'learning_rate/action_model': 1.3384655429271608e-05, 'epoch': 2.18}) +[2026-06-18 06:02:16,360][starVLA.training.train_starvla][INFO] - Step 5676, Loss: {'train/loss': 0.47742098569869995, 'train/grad_norm_pre_clip': 15.702621459960938, 'timing/data': 0.0014662882313132286, 'timing/model': 1.2994537875056267, 'learning_rate/qwen_vl_interface': 1.3372389590030204e-06, 'learning_rate/action_model': 1.3372389590030204e-05, 'epoch': 2.18}) +[2026-06-18 06:02:22,080][starVLA.training.train_starvla][INFO] - Step 5677, Loss: {'train/loss': 0.23412536084651947, 'train/grad_norm_pre_clip': 13.532114028930664, 'timing/data': 0.0019758762791752815, 'timing/model': 1.533520195633173, 'learning_rate/qwen_vl_interface': 1.3360131861984551e-06, 'learning_rate/action_model': 1.3360131861984552e-05, 'epoch': 2.18}) +[2026-06-18 06:02:27,280][starVLA.training.train_starvla][INFO] - Step 5678, Loss: {'train/loss': 0.3378911018371582, 'train/grad_norm_pre_clip': 13.722702026367188, 'timing/data': 0.0011705774813890457, 'timing/model': 1.3918526964262128, 'learning_rate/qwen_vl_interface': 1.3347882247675662e-06, 'learning_rate/action_model': 1.3347882247675661e-05, 'epoch': 2.18}) +[2026-06-18 06:02:32,930][starVLA.training.train_starvla][INFO] - Step 5679, Loss: {'train/loss': 0.2545793056488037, 'train/grad_norm_pre_clip': 13.773704528808594, 'timing/data': 0.0010475283488631248, 'timing/model': 1.2468996495008469, 'learning_rate/qwen_vl_interface': 1.3335640749642923e-06, 'learning_rate/action_model': 1.3335640749642923e-05, 'epoch': 2.18}) +[2026-06-18 06:02:38,120][starVLA.training.train_starvla][INFO] - Step 5680, Loss: {'train/loss': 0.22012849152088165, 'train/grad_norm_pre_clip': 13.190916061401367, 'timing/data': 0.0013327701017260551, 'timing/model': 1.185503002256155, 'learning_rate/qwen_vl_interface': 1.3323407370423993e-06, 'learning_rate/action_model': 1.3323407370423994e-05, 'epoch': 2.19}) +[2026-06-18 06:02:52,354][starVLA.training.train_starvla][INFO] - Step 5681, Loss: {'train/loss': 0.5336245894432068, 'train/grad_norm_pre_clip': 15.812719345092773, 'timing/data': 0.0019699977710843086, 'timing/model': 1.6561633134260774, 'learning_rate/qwen_vl_interface': 1.3311182112554877e-06, 'learning_rate/action_model': 1.3311182112554876e-05, 'epoch': 2.19}) +[2026-06-18 06:02:58,479][starVLA.training.train_starvla][INFO] - Step 5682, Loss: {'train/loss': 0.474344938993454, 'train/grad_norm_pre_clip': 17.519615173339844, 'timing/data': 0.0022340770810842514, 'timing/model': 2.1958803357556462, 'learning_rate/qwen_vl_interface': 1.3298964978569873e-06, 'learning_rate/action_model': 1.3298964978569872e-05, 'epoch': 2.19}) +[2026-06-18 06:03:06,182][starVLA.training.train_starvla][INFO] - Step 5683, Loss: {'train/loss': 0.26623445749282837, 'train/grad_norm_pre_clip': 12.94940185546875, 'timing/data': 0.008840600959956646, 'timing/model': 2.7711335914209485, 'learning_rate/qwen_vl_interface': 1.3286755971001618e-06, 'learning_rate/action_model': 1.3286755971001618e-05, 'epoch': 2.19}) +[2026-06-18 06:03:13,387][starVLA.training.train_starvla][INFO] - Step 5684, Loss: {'train/loss': 0.4825372099876404, 'train/grad_norm_pre_clip': 18.999855041503906, 'timing/data': 0.0013130353763699532, 'timing/model': 1.1496637212112546, 'learning_rate/qwen_vl_interface': 1.327455509238104e-06, 'learning_rate/action_model': 1.3274555092381039e-05, 'epoch': 2.19}) +[2026-06-18 06:03:18,668][starVLA.training.train_starvla][INFO] - Step 5685, Loss: {'train/loss': 0.3485981523990631, 'train/grad_norm_pre_clip': 14.645171165466309, 'timing/data': 0.001126512885093689, 'timing/model': 1.3455237690359354, 'learning_rate/qwen_vl_interface': 1.32623623452374e-06, 'learning_rate/action_model': 1.32623623452374e-05, 'epoch': 2.19}) +[2026-06-18 06:03:24,183][starVLA.training.train_starvla][INFO] - Step 5686, Loss: {'train/loss': 0.5278986692428589, 'train/grad_norm_pre_clip': 21.022693634033203, 'timing/data': 0.0020348308607935905, 'timing/model': 1.5968707716092467, 'learning_rate/qwen_vl_interface': 1.325017773209827e-06, 'learning_rate/action_model': 1.325017773209827e-05, 'epoch': 2.19}) +[2026-06-18 06:03:29,419][starVLA.training.train_starvla][INFO] - Step 5687, Loss: {'train/loss': 0.22861497104167938, 'train/grad_norm_pre_clip': 10.923733711242676, 'timing/data': 0.0013236450031399727, 'timing/model': 1.4841272309422493, 'learning_rate/qwen_vl_interface': 1.323800125548953e-06, 'learning_rate/action_model': 1.323800125548953e-05, 'epoch': 2.19}) +[2026-06-18 06:03:33,926][starVLA.training.train_starvla][INFO] - Step 5688, Loss: {'train/loss': 0.3365914225578308, 'train/grad_norm_pre_clip': 11.790580749511719, 'timing/data': 0.00144136231392622, 'timing/model': 1.1008450631052256, 'learning_rate/qwen_vl_interface': 1.3225832917935376e-06, 'learning_rate/action_model': 1.3225832917935377e-05, 'epoch': 2.19}) +[2026-06-18 06:03:45,810][starVLA.training.train_starvla][INFO] - Step 5689, Loss: {'train/loss': 0.4508330225944519, 'train/grad_norm_pre_clip': 25.296396255493164, 'timing/data': 0.0008340403437614441, 'timing/model': 3.117432439699769, 'learning_rate/qwen_vl_interface': 1.3213672721958309e-06, 'learning_rate/action_model': 1.321367272195831e-05, 'epoch': 2.19}) +[2026-06-18 06:03:58,762][starVLA.training.train_starvla][INFO] - Step 5690, Loss: {'train/loss': 0.4231424033641815, 'train/grad_norm_pre_clip': 15.337248802185059, 'timing/data': 0.0011801989749073982, 'timing/model': 2.3496952326968312, 'learning_rate/qwen_vl_interface': 1.3201520670079172e-06, 'learning_rate/action_model': 1.3201520670079171e-05, 'epoch': 2.19}) +[2026-06-18 06:04:11,635][starVLA.training.train_starvla][INFO] - Step 5691, Loss: {'train/loss': 0.5488898158073425, 'train/grad_norm_pre_clip': 25.000215530395508, 'timing/data': 0.0008933376520872116, 'timing/model': 2.9859695630148053, 'learning_rate/qwen_vl_interface': 1.31893767648171e-06, 'learning_rate/action_model': 1.3189376764817099e-05, 'epoch': 2.19}) +[2026-06-18 06:04:26,047][starVLA.training.train_starvla][INFO] - Step 5692, Loss: {'train/loss': 0.33314642310142517, 'train/grad_norm_pre_clip': 13.246719360351562, 'timing/data': 0.0011939750984311104, 'timing/model': 5.880427300930023, 'learning_rate/qwen_vl_interface': 1.3177241008689513e-06, 'learning_rate/action_model': 1.3177241008689511e-05, 'epoch': 2.19}) +[2026-06-18 06:04:44,015][starVLA.training.train_starvla][INFO] - Step 5693, Loss: {'train/loss': 0.2576933205127716, 'train/grad_norm_pre_clip': 11.741345405578613, 'timing/data': 0.0010724514722824097, 'timing/model': 4.912960174493492, 'learning_rate/qwen_vl_interface': 1.316511340421216e-06, 'learning_rate/action_model': 1.316511340421216e-05, 'epoch': 2.19}) +[2026-06-18 06:04:55,993][starVLA.training.train_starvla][INFO] - Step 5694, Loss: {'train/loss': 0.36537086963653564, 'train/grad_norm_pre_clip': 17.7359561920166, 'timing/data': 0.0013398155570030212, 'timing/model': 1.5267160329967737, 'learning_rate/qwen_vl_interface': 1.315299395389914e-06, 'learning_rate/action_model': 1.315299395389914e-05, 'epoch': 2.19}) +[2026-06-18 06:05:10,544][starVLA.training.train_starvla][INFO] - Step 5695, Loss: {'train/loss': 0.34947845339775085, 'train/grad_norm_pre_clip': 14.462650299072266, 'timing/data': 0.0011260369792580605, 'timing/model': 5.702776811085641, 'learning_rate/qwen_vl_interface': 1.3140882660262818e-06, 'learning_rate/action_model': 1.3140882660262818e-05, 'epoch': 2.19}) +[2026-06-18 06:05:28,017][starVLA.training.train_starvla][INFO] - Step 5696, Loss: {'train/loss': 0.2703307569026947, 'train/grad_norm_pre_clip': 12.207406997680664, 'timing/data': 0.0009016823023557663, 'timing/model': 6.054817876778543, 'learning_rate/qwen_vl_interface': 1.3128779525813865e-06, 'learning_rate/action_model': 1.3128779525813864e-05, 'epoch': 2.19}) +[2026-06-18 06:05:33,474][starVLA.training.train_starvla][INFO] - Step 5697, Loss: {'train/loss': 0.5493917465209961, 'train/grad_norm_pre_clip': 17.051891326904297, 'timing/data': 0.0012077316641807556, 'timing/model': 1.5133142601698637, 'learning_rate/qwen_vl_interface': 1.3116684553061288e-06, 'learning_rate/action_model': 1.3116684553061287e-05, 'epoch': 2.19}) +[2026-06-18 06:05:38,622][starVLA.training.train_starvla][INFO] - Step 5698, Loss: {'train/loss': 0.4744592308998108, 'train/grad_norm_pre_clip': 20.346288681030273, 'timing/data': 0.002182677388191223, 'timing/model': 1.1040757289156318, 'learning_rate/qwen_vl_interface': 1.310459774451238e-06, 'learning_rate/action_model': 1.3104597744512379e-05, 'epoch': 2.19}) +[2026-06-18 06:05:44,271][starVLA.training.train_starvla][INFO] - Step 5699, Loss: {'train/loss': 0.3930002748966217, 'train/grad_norm_pre_clip': 18.40867042541504, 'timing/data': 0.0011635031551122665, 'timing/model': 1.406107329763472, 'learning_rate/qwen_vl_interface': 1.3092519102672744e-06, 'learning_rate/action_model': 1.3092519102672742e-05, 'epoch': 2.19}) +[2026-06-18 06:09:54,991][starVLA.training.train_starvla][INFO] - Step 5700, Loss: {'train/loss': 0.30909329652786255, 'train/grad_norm_pre_clip': 22.053712844848633, 'eval/loss': 0.42473965883255005, 'eval/latency_4/loss': 0.42473965883255005, 'eval/demon_fire_precision': 0.9722132471728594, 'eval/demon_fire_recall': 0.9772653458915233, 'eval/demon_fire_f1': 0.9747327502429544, 'eval/demon_fire_rate_model': 0.908957415565345, 'eval/demon_fire_rate_teacher': 0.9042584434654919, 'eval/demon_left_precision': 0.7829268292682927, 'eval/demon_left_recall': 0.7328767123287672, 'eval/demon_left_f1': 0.7570754716981133, 'eval/demon_left_rate_model': 0.12041116005873716, 'eval/demon_left_rate_teacher': 0.12863436123348018, 'eval/demon_right_precision': 0.796028880866426, 'eval/demon_right_recall': 0.796028880866426, 'eval/demon_right_f1': 0.796028880866426, 'eval/demon_right_rate_model': 0.16270190895741557, 'eval/demon_right_rate_teacher': 0.16270190895741557, 'eval/demon_move_f1': 0.7765521762822696, 'eval/demon_cc_f1': 0.905412787689495, 'eval/latency_4/demon_fire_precision': 0.9722132471728594, 'eval/latency_4/demon_fire_recall': 0.9772653458915233, 'eval/latency_4/demon_fire_f1': 0.9747327502429544, 'eval/latency_4/demon_fire_rate_model': 0.908957415565345, 'eval/latency_4/demon_fire_rate_teacher': 0.9042584434654919, 'eval/latency_4/demon_left_precision': 0.7829268292682927, 'eval/latency_4/demon_left_recall': 0.7328767123287672, 'eval/latency_4/demon_left_f1': 0.7570754716981133, 'eval/latency_4/demon_left_rate_model': 0.12041116005873716, 'eval/latency_4/demon_left_rate_teacher': 0.12863436123348018, 'eval/latency_4/demon_right_precision': 0.796028880866426, 'eval/latency_4/demon_right_recall': 0.796028880866426, 'eval/latency_4/demon_right_f1': 0.796028880866426, 'eval/latency_4/demon_right_rate_model': 0.16270190895741557, 'eval/latency_4/demon_right_rate_teacher': 0.16270190895741557, 'eval/latency_4/demon_move_f1': 0.7765521762822696, 'eval/latency_4/demon_cc_f1': 0.905412787689495, 'timing/data': 0.0011640554293990135, 'timing/model': 1.1024662358686328, 'learning_rate/qwen_vl_interface': 1.3080448630046312e-06, 'learning_rate/action_model': 1.308044863004631e-05, 'epoch': 2.19}) +[2026-06-18 06:09:54,993][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_5700_state +[2026-06-18 06:09:55,000][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 06:10:05,179][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_5700_state/model.safetensors +[2026-06-18 06:10:29,562][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_5700_state/optimizer.bin +[2026-06-18 06:10:29,631][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_5700_state/sampler.bin +[2026-06-18 06:10:29,632][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_5700_state/sampler_1.bin +[2026-06-18 06:10:29,637][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_5700_state/random_states_0.pkl +[2026-06-18 06:10:29,643][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 5700 +[2026-06-18 06:10:29,643][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 06:10:29,660][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 06:10:34,956][starVLA.training.train_starvla][INFO] - Step 5701, Loss: {'train/loss': 0.3954984247684479, 'train/grad_norm_pre_clip': 17.074111938476562, 'timing/data': 0.0009568445384502411, 'timing/model': 1.6086640488356352, 'learning_rate/qwen_vl_interface': 1.3068386329135287e-06, 'learning_rate/action_model': 1.3068386329135287e-05, 'epoch': 2.19}) +[2026-06-18 06:10:43,240][starVLA.training.train_starvla][INFO] - Step 5702, Loss: {'train/loss': 0.1865919828414917, 'train/grad_norm_pre_clip': 14.572275161743164, 'timing/data': 0.0015241624787449837, 'timing/model': 4.693870080634952, 'learning_rate/qwen_vl_interface': 1.3056332202440211e-06, 'learning_rate/action_model': 1.305633220244021e-05, 'epoch': 2.19}) +[2026-06-18 06:10:48,846][starVLA.training.train_starvla][INFO] - Step 5703, Loss: {'train/loss': 0.39715805649757385, 'train/grad_norm_pre_clip': 17.913562774658203, 'timing/data': 0.003958563320338726, 'timing/model': 1.431460315361619, 'learning_rate/qwen_vl_interface': 1.3044286252459914e-06, 'learning_rate/action_model': 1.3044286252459915e-05, 'epoch': 2.19}) +[2026-06-18 06:10:54,288][starVLA.training.train_starvla][INFO] - Step 5704, Loss: {'train/loss': 0.18870188295841217, 'train/grad_norm_pre_clip': 15.803375244140625, 'timing/data': 0.001549379900097847, 'timing/model': 1.2941136788576841, 'learning_rate/qwen_vl_interface': 1.3032248481691523e-06, 'learning_rate/action_model': 1.3032248481691523e-05, 'epoch': 2.19}) +[2026-06-18 06:10:59,450][starVLA.training.train_starvla][INFO] - Step 5705, Loss: {'train/loss': 0.4623136818408966, 'train/grad_norm_pre_clip': 16.32711410522461, 'timing/data': 0.0014570290222764015, 'timing/model': 1.1091309459879994, 'learning_rate/qwen_vl_interface': 1.302021889263051e-06, 'learning_rate/action_model': 1.3020218892630508e-05, 'epoch': 2.19}) +[2026-06-18 06:11:04,642][starVLA.training.train_starvla][INFO] - Step 5706, Loss: {'train/loss': 0.3292814791202545, 'train/grad_norm_pre_clip': 14.025795936584473, 'timing/data': 0.001175643876194954, 'timing/model': 1.3189988918602467, 'learning_rate/qwen_vl_interface': 1.3008197487770593e-06, 'learning_rate/action_model': 1.3008197487770593e-05, 'epoch': 2.2}) +[2026-06-18 06:11:09,866][starVLA.training.train_starvla][INFO] - Step 5707, Loss: {'train/loss': 0.2975318431854248, 'train/grad_norm_pre_clip': 12.897218704223633, 'timing/data': 0.0012000985443592072, 'timing/model': 1.4166733920574188, 'learning_rate/qwen_vl_interface': 1.2996184269603835e-06, 'learning_rate/action_model': 1.2996184269603833e-05, 'epoch': 2.2}) +[2026-06-18 06:11:14,794][starVLA.training.train_starvla][INFO] - Step 5708, Loss: {'train/loss': 0.2004658430814743, 'train/grad_norm_pre_clip': 9.153779029846191, 'timing/data': 0.001125594601035118, 'timing/model': 1.2171746473759413, 'learning_rate/qwen_vl_interface': 1.2984179240620567e-06, 'learning_rate/action_model': 1.2984179240620567e-05, 'epoch': 2.2}) +[2026-06-18 06:11:19,842][starVLA.training.train_starvla][INFO] - Step 5709, Loss: {'train/loss': 0.22365131974220276, 'train/grad_norm_pre_clip': 12.31626033782959, 'timing/data': 0.0013547036796808243, 'timing/model': 1.3232661252841353, 'learning_rate/qwen_vl_interface': 1.2972182403309474e-06, 'learning_rate/action_model': 1.2972182403309474e-05, 'epoch': 2.2}) +[2026-06-18 06:11:33,529][starVLA.training.train_starvla][INFO] - Step 5710, Loss: {'train/loss': 0.403251975774765, 'train/grad_norm_pre_clip': 15.909523963928223, 'timing/data': 0.001231856644153595, 'timing/model': 1.098405135795474, 'learning_rate/qwen_vl_interface': 1.2960193760157507e-06, 'learning_rate/action_model': 1.2960193760157507e-05, 'epoch': 2.2}) +[2026-06-18 06:11:43,059][starVLA.training.train_starvla][INFO] - Step 5711, Loss: {'train/loss': 0.3979712724685669, 'train/grad_norm_pre_clip': 19.81158447265625, 'timing/data': 0.0010688630864024162, 'timing/model': 1.5931418584659696, 'learning_rate/qwen_vl_interface': 1.294821331364992e-06, 'learning_rate/action_model': 1.294821331364992e-05, 'epoch': 2.2}) +[2026-06-18 06:11:48,170][starVLA.training.train_starvla][INFO] - Step 5712, Loss: {'train/loss': 0.20417091250419617, 'train/grad_norm_pre_clip': 24.313749313354492, 'timing/data': 0.0012084618210792542, 'timing/model': 1.4331987975165248, 'learning_rate/qwen_vl_interface': 1.2936241066270278e-06, 'learning_rate/action_model': 1.2936241066270278e-05, 'epoch': 2.2}) +[2026-06-18 06:11:52,734][starVLA.training.train_starvla][INFO] - Step 5713, Loss: {'train/loss': 0.3958219885826111, 'train/grad_norm_pre_clip': 14.077363014221191, 'timing/data': 0.0011790255084633827, 'timing/model': 1.1197184259071946, 'learning_rate/qwen_vl_interface': 1.2924277020500427e-06, 'learning_rate/action_model': 1.2924277020500428e-05, 'epoch': 2.2}) +[2026-06-18 06:12:07,586][starVLA.training.train_starvla][INFO] - Step 5714, Loss: {'train/loss': 0.30463066697120667, 'train/grad_norm_pre_clip': 16.959508895874023, 'timing/data': 0.00107574462890625, 'timing/model': 1.1340730711817741, 'learning_rate/qwen_vl_interface': 1.291232117882054e-06, 'learning_rate/action_model': 1.2912321178820538e-05, 'epoch': 2.2}) +[2026-06-18 06:12:13,381][starVLA.training.train_starvla][INFO] - Step 5715, Loss: {'train/loss': 0.268489807844162, 'train/grad_norm_pre_clip': 16.92168617248535, 'timing/data': 0.0014139963313937187, 'timing/model': 1.402808865532279, 'learning_rate/qwen_vl_interface': 1.2900373543709065e-06, 'learning_rate/action_model': 1.2900373543709065e-05, 'epoch': 2.2}) +[2026-06-18 06:12:24,181][starVLA.training.train_starvla][INFO] - Step 5716, Loss: {'train/loss': 0.40845051407814026, 'train/grad_norm_pre_clip': 51.67881393432617, 'timing/data': 0.0010939911007881165, 'timing/model': 1.135586996562779, 'learning_rate/qwen_vl_interface': 1.288843411764277e-06, 'learning_rate/action_model': 1.288843411764277e-05, 'epoch': 2.2}) +[2026-06-18 06:12:30,057][starVLA.training.train_starvla][INFO] - Step 5717, Loss: {'train/loss': 0.4270581007003784, 'train/grad_norm_pre_clip': 17.87513542175293, 'timing/data': 0.001448802649974823, 'timing/model': 1.6428193422034383, 'learning_rate/qwen_vl_interface': 1.2876502903096709e-06, 'learning_rate/action_model': 1.2876502903096708e-05, 'epoch': 2.2}) +[2026-06-18 06:12:39,574][starVLA.training.train_starvla][INFO] - Step 5718, Loss: {'train/loss': 0.7670816779136658, 'train/grad_norm_pre_clip': 30.03426742553711, 'timing/data': 0.0012653777375817299, 'timing/model': 1.2844530045986176, 'learning_rate/qwen_vl_interface': 1.286457990254423e-06, 'learning_rate/action_model': 1.286457990254423e-05, 'epoch': 2.2}) +[2026-06-18 06:12:51,739][starVLA.training.train_starvla][INFO] - Step 5719, Loss: {'train/loss': 0.25727578997612, 'train/grad_norm_pre_clip': 16.362241744995117, 'timing/data': 0.0013374751433730125, 'timing/model': 5.8334154430776834, 'learning_rate/qwen_vl_interface': 1.2852665118456971e-06, 'learning_rate/action_model': 1.2852665118456972e-05, 'epoch': 2.2}) +[2026-06-18 06:13:03,578][starVLA.training.train_starvla][INFO] - Step 5720, Loss: {'train/loss': 0.4398432672023773, 'train/grad_norm_pre_clip': 13.942800521850586, 'timing/data': 0.0008987411856651306, 'timing/model': 4.570392505265772, 'learning_rate/qwen_vl_interface': 1.2840758553304927e-06, 'learning_rate/action_model': 1.2840758553304927e-05, 'epoch': 2.2}) +[2026-06-18 06:13:11,118][starVLA.training.train_starvla][INFO] - Step 5721, Loss: {'train/loss': 0.3840840756893158, 'train/grad_norm_pre_clip': 18.048803329467773, 'timing/data': 0.029596594162285328, 'timing/model': 2.8600850412622094, 'learning_rate/qwen_vl_interface': 1.2828860209556293e-06, 'learning_rate/action_model': 1.2828860209556292e-05, 'epoch': 2.2}) +[2026-06-18 06:13:15,745][starVLA.training.train_starvla][INFO] - Step 5722, Loss: {'train/loss': 0.43501460552215576, 'train/grad_norm_pre_clip': 21.454587936401367, 'timing/data': 0.03822311386466026, 'timing/model': 1.1168593661859632, 'learning_rate/qwen_vl_interface': 1.2816970089677617e-06, 'learning_rate/action_model': 1.2816970089677615e-05, 'epoch': 2.2}) +[2026-06-18 06:13:22,458][starVLA.training.train_starvla][INFO] - Step 5723, Loss: {'train/loss': 0.38394060730934143, 'train/grad_norm_pre_clip': 16.81543731689453, 'timing/data': 0.04550028406083584, 'timing/model': 1.120661785826087, 'learning_rate/qwen_vl_interface': 1.280508819613375e-06, 'learning_rate/action_model': 1.280508819613375e-05, 'epoch': 2.2}) +[2026-06-18 06:13:27,660][starVLA.training.train_starvla][INFO] - Step 5724, Loss: {'train/loss': 0.3779575824737549, 'train/grad_norm_pre_clip': 16.95925521850586, 'timing/data': 0.051725996658205986, 'timing/model': 1.118611036799848, 'learning_rate/qwen_vl_interface': 1.2793214531387812e-06, 'learning_rate/action_model': 1.2793214531387812e-05, 'epoch': 2.2}) +[2026-06-18 06:13:32,456][starVLA.training.train_starvla][INFO] - Step 5725, Loss: {'train/loss': 0.3828659653663635, 'train/grad_norm_pre_clip': 15.032564163208008, 'timing/data': 0.03241140116006136, 'timing/model': 1.116562757641077, 'learning_rate/qwen_vl_interface': 1.2781349097901246e-06, 'learning_rate/action_model': 1.2781349097901247e-05, 'epoch': 2.2}) +[2026-06-18 06:13:40,322][starVLA.training.train_starvla][INFO] - Step 5726, Loss: {'train/loss': 0.11572743207216263, 'train/grad_norm_pre_clip': 13.111722946166992, 'timing/data': 0.034951894544065, 'timing/model': 1.5240944400429726, 'learning_rate/qwen_vl_interface': 1.2769491898133723e-06, 'learning_rate/action_model': 1.2769491898133722e-05, 'epoch': 2.2}) +[2026-06-18 06:13:46,281][starVLA.training.train_starvla][INFO] - Step 5727, Loss: {'train/loss': 0.24292533099651337, 'train/grad_norm_pre_clip': 14.147784233093262, 'timing/data': 0.04514662083238363, 'timing/model': 1.445709116756916, 'learning_rate/qwen_vl_interface': 1.2757642934543293e-06, 'learning_rate/action_model': 1.2757642934543293e-05, 'epoch': 2.2}) +[2026-06-18 06:13:52,169][starVLA.training.train_starvla][INFO] - Step 5728, Loss: {'train/loss': 0.2686459720134735, 'train/grad_norm_pre_clip': 13.494760513305664, 'timing/data': 0.03561178781092167, 'timing/model': 1.6165957376360893, 'learning_rate/qwen_vl_interface': 1.274580220958625e-06, 'learning_rate/action_model': 1.274580220958625e-05, 'epoch': 2.2}) +[2026-06-18 06:13:58,093][starVLA.training.train_starvla][INFO] - Step 5729, Loss: {'train/loss': 0.2591158151626587, 'train/grad_norm_pre_clip': 13.158312797546387, 'timing/data': 0.0411642175167799, 'timing/model': 1.3479121457785368, 'learning_rate/qwen_vl_interface': 1.2733969725717184e-06, 'learning_rate/action_model': 1.2733969725717184e-05, 'epoch': 2.2}) +[2026-06-18 06:14:03,118][starVLA.training.train_starvla][INFO] - Step 5730, Loss: {'train/loss': 0.5072311758995056, 'train/grad_norm_pre_clip': 21.299421310424805, 'timing/data': 0.03448924142867327, 'timing/model': 1.4177858149632812, 'learning_rate/qwen_vl_interface': 1.272214548538899e-06, 'learning_rate/action_model': 1.2722145485388989e-05, 'epoch': 2.2}) +[2026-06-18 06:14:09,587][starVLA.training.train_starvla][INFO] - Step 5731, Loss: {'train/loss': 0.12535318732261658, 'train/grad_norm_pre_clip': 19.012069702148438, 'timing/data': 0.040355321019887924, 'timing/model': 1.3907190663740039, 'learning_rate/qwen_vl_interface': 1.2710329491052834e-06, 'learning_rate/action_model': 1.2710329491052833e-05, 'epoch': 2.2}) +[2026-06-18 06:14:15,272][starVLA.training.train_starvla][INFO] - Step 5732, Loss: {'train/loss': 0.3630285859107971, 'train/grad_norm_pre_clip': 17.903522491455078, 'timing/data': 0.0388366524130106, 'timing/model': 1.9723454099148512, 'learning_rate/qwen_vl_interface': 1.2698521745158188e-06, 'learning_rate/action_model': 1.2698521745158187e-05, 'epoch': 2.21}) +[2026-06-18 06:14:20,449][starVLA.training.train_starvla][INFO] - Step 5733, Loss: {'train/loss': 0.2890113592147827, 'train/grad_norm_pre_clip': 12.717658042907715, 'timing/data': 0.03380894102156162, 'timing/model': 1.137713722884655, 'learning_rate/qwen_vl_interface': 1.2686722250152804e-06, 'learning_rate/action_model': 1.2686722250152806e-05, 'epoch': 2.21}) +[2026-06-18 06:14:26,265][starVLA.training.train_starvla][INFO] - Step 5734, Loss: {'train/loss': 0.40788376331329346, 'train/grad_norm_pre_clip': 18.52228355407715, 'timing/data': 0.04604025837033987, 'timing/model': 1.1143240500241518, 'learning_rate/qwen_vl_interface': 1.2674931008482763e-06, 'learning_rate/action_model': 1.2674931008482763e-05, 'epoch': 2.21}) +[2026-06-18 06:14:32,282][starVLA.training.train_starvla][INFO] - Step 5735, Loss: {'train/loss': 0.11104053258895874, 'train/grad_norm_pre_clip': 7.213113307952881, 'timing/data': 0.034692585468292236, 'timing/model': 1.5872335117310286, 'learning_rate/qwen_vl_interface': 1.2663148022592365e-06, 'learning_rate/action_model': 1.2663148022592363e-05, 'epoch': 2.21}) +[2026-06-18 06:14:38,080][starVLA.training.train_starvla][INFO] - Step 5736, Loss: {'train/loss': 0.35667693614959717, 'train/grad_norm_pre_clip': 10.88479232788086, 'timing/data': 0.044246867299079895, 'timing/model': 1.1873147320002317, 'learning_rate/qwen_vl_interface': 1.2651373294924249e-06, 'learning_rate/action_model': 1.2651373294924248e-05, 'epoch': 2.21}) +[2026-06-18 06:14:43,235][starVLA.training.train_starvla][INFO] - Step 5737, Loss: {'train/loss': 0.2434759885072708, 'train/grad_norm_pre_clip': 11.019721984863281, 'timing/data': 0.037943340837955475, 'timing/model': 1.5070890160277486, 'learning_rate/qwen_vl_interface': 1.2639606827919313e-06, 'learning_rate/action_model': 1.2639606827919311e-05, 'epoch': 2.21}) +[2026-06-18 06:14:48,267][starVLA.training.train_starvla][INFO] - Step 5738, Loss: {'train/loss': 0.14761163294315338, 'train/grad_norm_pre_clip': 12.040051460266113, 'timing/data': 0.03519980516284704, 'timing/model': 1.1134215518832207, 'learning_rate/qwen_vl_interface': 1.2627848624016788e-06, 'learning_rate/action_model': 1.2627848624016788e-05, 'epoch': 2.21}) +[2026-06-18 06:15:01,009][starVLA.training.train_starvla][INFO] - Step 5739, Loss: {'train/loss': 0.3709118068218231, 'train/grad_norm_pre_clip': 23.452394485473633, 'timing/data': 0.03870285674929619, 'timing/model': 1.3598006553947926, 'learning_rate/qwen_vl_interface': 1.261609868565416e-06, 'learning_rate/action_model': 1.2616098685654159e-05, 'epoch': 2.21}) +[2026-06-18 06:15:06,539][starVLA.training.train_starvla][INFO] - Step 5740, Loss: {'train/loss': 0.17180156707763672, 'train/grad_norm_pre_clip': 11.469032287597656, 'timing/data': 0.0471865339204669, 'timing/model': 1.3282195841893554, 'learning_rate/qwen_vl_interface': 1.2604357015267165e-06, 'learning_rate/action_model': 1.2604357015267166e-05, 'epoch': 2.21}) +[2026-06-18 06:15:13,002][starVLA.training.train_starvla][INFO] - Step 5741, Loss: {'train/loss': 0.4605557918548584, 'train/grad_norm_pre_clip': 19.482206344604492, 'timing/data': 0.035246556624770164, 'timing/model': 1.0709442319348454, 'learning_rate/qwen_vl_interface': 1.2592623615289906e-06, 'learning_rate/action_model': 1.2592623615289906e-05, 'epoch': 2.21}) +[2026-06-18 06:15:21,269][starVLA.training.train_starvla][INFO] - Step 5742, Loss: {'train/loss': 0.3762100040912628, 'train/grad_norm_pre_clip': 17.61349105834961, 'timing/data': 0.04485438950359821, 'timing/model': 4.938303231261671, 'learning_rate/qwen_vl_interface': 1.258089848815471e-06, 'learning_rate/action_model': 1.2580898488154708e-05, 'epoch': 2.21}) +[2026-06-18 06:15:26,173][starVLA.training.train_starvla][INFO] - Step 5743, Loss: {'train/loss': 0.21937713027000427, 'train/grad_norm_pre_clip': 13.35183334350586, 'timing/data': 0.03533657547086477, 'timing/model': 1.2287315223366022, 'learning_rate/qwen_vl_interface': 1.256918163629221e-06, 'learning_rate/action_model': 1.256918163629221e-05, 'epoch': 2.21}) +[2026-06-18 06:15:32,283][starVLA.training.train_starvla][INFO] - Step 5744, Loss: {'train/loss': 0.37363401055336, 'train/grad_norm_pre_clip': 15.616020202636719, 'timing/data': 0.03704558685421944, 'timing/model': 1.2646566089242697, 'learning_rate/qwen_vl_interface': 1.2557473062131332e-06, 'learning_rate/action_model': 1.2557473062131331e-05, 'epoch': 2.21}) +[2026-06-18 06:15:37,185][starVLA.training.train_starvla][INFO] - Step 5745, Loss: {'train/loss': 0.2998271882534027, 'train/grad_norm_pre_clip': 16.34551239013672, 'timing/data': 0.03464135713875294, 'timing/model': 1.3293371591717005, 'learning_rate/qwen_vl_interface': 1.2545772768099263e-06, 'learning_rate/action_model': 1.2545772768099263e-05, 'epoch': 2.21}) +[2026-06-18 06:15:42,335][starVLA.training.train_starvla][INFO] - Step 5746, Loss: {'train/loss': 0.29852524399757385, 'train/grad_norm_pre_clip': 12.007189750671387, 'timing/data': 0.042613510973751545, 'timing/model': 1.110708674415946, 'learning_rate/qwen_vl_interface': 1.2534080756621488e-06, 'learning_rate/action_model': 1.2534080756621488e-05, 'epoch': 2.21}) +[2026-06-18 06:15:53,467][starVLA.training.train_starvla][INFO] - Step 5747, Loss: {'train/loss': 0.3337978422641754, 'train/grad_norm_pre_clip': 14.07115650177002, 'timing/data': 0.09087029378861189, 'timing/model': 1.5419159606099129, 'learning_rate/qwen_vl_interface': 1.2522397030121782e-06, 'learning_rate/action_model': 1.2522397030121783e-05, 'epoch': 2.21}) +[2026-06-18 06:16:00,621][starVLA.training.train_starvla][INFO] - Step 5748, Loss: {'train/loss': 0.2963539958000183, 'train/grad_norm_pre_clip': 13.620705604553223, 'timing/data': 0.05382242891937494, 'timing/model': 1.5593188591301441, 'learning_rate/qwen_vl_interface': 1.2510721591022194e-06, 'learning_rate/action_model': 1.2510721591022192e-05, 'epoch': 2.21}) +[2026-06-18 06:16:06,971][starVLA.training.train_starvla][INFO] - Step 5749, Loss: {'train/loss': 0.20274579524993896, 'train/grad_norm_pre_clip': 7.564445972442627, 'timing/data': 0.03323472198098898, 'timing/model': 1.311106945388019, 'learning_rate/qwen_vl_interface': 1.2499054441743047e-06, 'learning_rate/action_model': 1.2499054441743047e-05, 'epoch': 2.21}) +[2026-06-18 06:16:18,426][starVLA.training.train_starvla][INFO] - Step 5750, Loss: {'train/loss': 0.34295597672462463, 'train/grad_norm_pre_clip': 15.143491744995117, 'timing/data': 0.03781204577535391, 'timing/model': 1.1878500636667013, 'learning_rate/qwen_vl_interface': 1.2487395584702955e-06, 'learning_rate/action_model': 1.2487395584702954e-05, 'epoch': 2.21}) +[2026-06-18 06:16:23,458][starVLA.training.train_starvla][INFO] - Step 5751, Loss: {'train/loss': 0.3264891803264618, 'train/grad_norm_pre_clip': 17.937108993530273, 'timing/data': 0.05119610484689474, 'timing/model': 1.219975263811648, 'learning_rate/qwen_vl_interface': 1.2475745022318803e-06, 'learning_rate/action_model': 1.2475745022318803e-05, 'epoch': 2.21}) +[2026-06-18 06:16:29,131][starVLA.training.train_starvla][INFO] - Step 5752, Loss: {'train/loss': 0.421210378408432, 'train/grad_norm_pre_clip': 14.964734077453613, 'timing/data': 0.05320549476891756, 'timing/model': 1.5865673972293735, 'learning_rate/qwen_vl_interface': 1.2464102757005788e-06, 'learning_rate/action_model': 1.2464102757005788e-05, 'epoch': 2.21}) +[2026-06-18 06:16:33,614][starVLA.training.train_starvla][INFO] - Step 5753, Loss: {'train/loss': 0.40015995502471924, 'train/grad_norm_pre_clip': 20.826574325561523, 'timing/data': 0.030641324818134308, 'timing/model': 1.1036542002111673, 'learning_rate/qwen_vl_interface': 1.2452468791177357e-06, 'learning_rate/action_model': 1.2452468791177357e-05, 'epoch': 2.21}) +[2026-06-18 06:16:38,705][starVLA.training.train_starvla][INFO] - Step 5754, Loss: {'train/loss': 0.3639907240867615, 'train/grad_norm_pre_clip': 17.330949783325195, 'timing/data': 0.034557404927909374, 'timing/model': 1.1488175736740232, 'learning_rate/qwen_vl_interface': 1.2440843127245235e-06, 'learning_rate/action_model': 1.2440843127245235e-05, 'epoch': 2.21}) +[2026-06-18 06:16:44,044][starVLA.training.train_starvla][INFO] - Step 5755, Loss: {'train/loss': 0.19935913383960724, 'train/grad_norm_pre_clip': 11.209842681884766, 'timing/data': 0.03806010726839304, 'timing/model': 1.1802276549860835, 'learning_rate/qwen_vl_interface': 1.2429225767619427e-06, 'learning_rate/action_model': 1.2429225767619427e-05, 'epoch': 2.21}) +[2026-06-18 06:16:54,237][starVLA.training.train_starvla][INFO] - Step 5756, Loss: {'train/loss': 0.2804148197174072, 'train/grad_norm_pre_clip': 16.96116065979004, 'timing/data': 0.034854563884437084, 'timing/model': 5.505333276465535, 'learning_rate/qwen_vl_interface': 1.2417616714708246e-06, 'learning_rate/action_model': 1.2417616714708247e-05, 'epoch': 2.21}) +[2026-06-18 06:16:59,941][starVLA.training.train_starvla][INFO] - Step 5757, Loss: {'train/loss': 0.9100192785263062, 'train/grad_norm_pre_clip': 28.89958381652832, 'timing/data': 0.034791333600878716, 'timing/model': 1.9471761789172888, 'learning_rate/qwen_vl_interface': 1.2406015970918247e-06, 'learning_rate/action_model': 1.2406015970918247e-05, 'epoch': 2.21}) +[2026-06-18 06:17:06,181][starVLA.training.train_starvla][INFO] - Step 5758, Loss: {'train/loss': 0.19243961572647095, 'train/grad_norm_pre_clip': 12.29780101776123, 'timing/data': 0.04667485784739256, 'timing/model': 1.1507867947220802, 'learning_rate/qwen_vl_interface': 1.2394423538654296e-06, 'learning_rate/action_model': 1.2394423538654296e-05, 'epoch': 2.22}) +[2026-06-18 06:17:11,911][starVLA.training.train_starvla][INFO] - Step 5759, Loss: {'train/loss': 0.3801727294921875, 'train/grad_norm_pre_clip': 15.509208679199219, 'timing/data': 0.0376154575496912, 'timing/model': 1.1721013840287924, 'learning_rate/qwen_vl_interface': 1.2382839420319478e-06, 'learning_rate/action_model': 1.2382839420319478e-05, 'epoch': 2.22}) +[2026-06-18 06:17:17,143][starVLA.training.train_starvla][INFO] - Step 5760, Loss: {'train/loss': 0.6425461769104004, 'train/grad_norm_pre_clip': 20.497528076171875, 'timing/data': 0.04336042143404484, 'timing/model': 1.4160214578732848, 'learning_rate/qwen_vl_interface': 1.2371263618315225e-06, 'learning_rate/action_model': 1.2371263618315224e-05, 'epoch': 2.22}) +[2026-06-18 06:17:22,632][starVLA.training.train_starvla][INFO] - Step 5761, Loss: {'train/loss': 0.16744036972522736, 'train/grad_norm_pre_clip': 9.832830429077148, 'timing/data': 0.041387674398720264, 'timing/model': 1.0969433998689055, 'learning_rate/qwen_vl_interface': 1.2359696135041202e-06, 'learning_rate/action_model': 1.2359696135041201e-05, 'epoch': 2.22}) +[2026-06-18 06:17:28,173][starVLA.training.train_starvla][INFO] - Step 5762, Loss: {'train/loss': 0.47455522418022156, 'train/grad_norm_pre_clip': 15.052774429321289, 'timing/data': 0.039179323241114616, 'timing/model': 1.5342744523659348, 'learning_rate/qwen_vl_interface': 1.2348136972895369e-06, 'learning_rate/action_model': 1.2348136972895368e-05, 'epoch': 2.22}) +[2026-06-18 06:17:33,391][starVLA.training.train_starvla][INFO] - Step 5763, Loss: {'train/loss': 0.2790569067001343, 'train/grad_norm_pre_clip': 14.80577278137207, 'timing/data': 0.04050187021493912, 'timing/model': 1.3262441456317902, 'learning_rate/qwen_vl_interface': 1.2336586134273945e-06, 'learning_rate/action_model': 1.2336586134273945e-05, 'epoch': 2.22}) +[2026-06-18 06:17:38,514][starVLA.training.train_starvla][INFO] - Step 5764, Loss: {'train/loss': 0.2724047601222992, 'train/grad_norm_pre_clip': 15.415658950805664, 'timing/data': 0.10359212569892406, 'timing/model': 1.2805304853245616, 'learning_rate/qwen_vl_interface': 1.2325043621571428e-06, 'learning_rate/action_model': 1.2325043621571427e-05, 'epoch': 2.22}) +[2026-06-18 06:17:45,226][starVLA.training.train_starvla][INFO] - Step 5765, Loss: {'train/loss': 0.3683732748031616, 'train/grad_norm_pre_clip': 15.580080032348633, 'timing/data': 0.040414102375507355, 'timing/model': 1.098247149027884, 'learning_rate/qwen_vl_interface': 1.2313509437180603e-06, 'learning_rate/action_model': 1.2313509437180604e-05, 'epoch': 2.22}) +[2026-06-18 06:17:52,520][starVLA.training.train_starvla][INFO] - Step 5766, Loss: {'train/loss': 0.27444201707839966, 'train/grad_norm_pre_clip': 12.144892692565918, 'timing/data': 0.03979762643575668, 'timing/model': 3.2516227727755904, 'learning_rate/qwen_vl_interface': 1.2301983583492502e-06, 'learning_rate/action_model': 1.2301983583492502e-05, 'epoch': 2.22}) +[2026-06-18 06:17:58,020][starVLA.training.train_starvla][INFO] - Step 5767, Loss: {'train/loss': 0.44171589612960815, 'train/grad_norm_pre_clip': 20.202369689941406, 'timing/data': 0.03679492976516485, 'timing/model': 1.9192354548722506, 'learning_rate/qwen_vl_interface': 1.229046606289648e-06, 'learning_rate/action_model': 1.229046606289648e-05, 'epoch': 2.22}) +[2026-06-18 06:18:13,337][starVLA.training.train_starvla][INFO] - Step 5768, Loss: {'train/loss': 0.4596911370754242, 'train/grad_norm_pre_clip': 12.103011131286621, 'timing/data': 0.029497298412024975, 'timing/model': 4.744735070504248, 'learning_rate/qwen_vl_interface': 1.2278956877780102e-06, 'learning_rate/action_model': 1.2278956877780101e-05, 'epoch': 2.22}) +[2026-06-18 06:18:27,762][starVLA.training.train_starvla][INFO] - Step 5769, Loss: {'train/loss': 0.30389854311943054, 'train/grad_norm_pre_clip': 15.478421211242676, 'timing/data': 0.03311200626194477, 'timing/model': 5.75360035803169, 'learning_rate/qwen_vl_interface': 1.2267456030529245e-06, 'learning_rate/action_model': 1.2267456030529245e-05, 'epoch': 2.22}) +[2026-06-18 06:18:33,674][starVLA.training.train_starvla][INFO] - Step 5770, Loss: {'train/loss': 0.5161613821983337, 'train/grad_norm_pre_clip': 17.960041046142578, 'timing/data': 0.044490398839116096, 'timing/model': 1.4593095863237977, 'learning_rate/qwen_vl_interface': 1.225596352352803e-06, 'learning_rate/action_model': 1.225596352352803e-05, 'epoch': 2.22}) +[2026-06-18 06:18:40,680][starVLA.training.train_starvla][INFO] - Step 5771, Loss: {'train/loss': 0.19929218292236328, 'train/grad_norm_pre_clip': 11.037290573120117, 'timing/data': 0.0434252405539155, 'timing/model': 2.700650821439922, 'learning_rate/qwen_vl_interface': 1.2244479359158895e-06, 'learning_rate/action_model': 1.2244479359158895e-05, 'epoch': 2.22}) +[2026-06-18 06:18:46,237][starVLA.training.train_starvla][INFO] - Step 5772, Loss: {'train/loss': 0.2913420796394348, 'train/grad_norm_pre_clip': 13.490211486816406, 'timing/data': 0.0349501334130764, 'timing/model': 1.6048311367630959, 'learning_rate/qwen_vl_interface': 1.2233003539802512e-06, 'learning_rate/action_model': 1.2233003539802513e-05, 'epoch': 2.22}) +[2026-06-18 06:18:50,989][starVLA.training.train_starvla][INFO] - Step 5773, Loss: {'train/loss': 0.31901782751083374, 'train/grad_norm_pre_clip': 15.128076553344727, 'timing/data': 0.03713962435722351, 'timing/model': 1.1642985474318266, 'learning_rate/qwen_vl_interface': 1.2221536067837807e-06, 'learning_rate/action_model': 1.2221536067837808e-05, 'epoch': 2.22}) +[2026-06-18 06:18:56,647][starVLA.training.train_starvla][INFO] - Step 5774, Loss: {'train/loss': 0.3562760353088379, 'train/grad_norm_pre_clip': 21.327974319458008, 'timing/data': 0.03665970638394356, 'timing/model': 1.5157868089154363, 'learning_rate/qwen_vl_interface': 1.2210076945642025e-06, 'learning_rate/action_model': 1.2210076945642024e-05, 'epoch': 2.22}) +[2026-06-18 06:19:08,503][starVLA.training.train_starvla][INFO] - Step 5775, Loss: {'train/loss': 0.4066702127456665, 'train/grad_norm_pre_clip': 17.203189849853516, 'timing/data': 0.05656429938971996, 'timing/model': 6.1534579480066895, 'learning_rate/qwen_vl_interface': 1.219862617559065e-06, 'learning_rate/action_model': 1.219862617559065e-05, 'epoch': 2.22}) +[2026-06-18 06:19:18,509][starVLA.training.train_starvla][INFO] - Step 5776, Loss: {'train/loss': 0.5699392557144165, 'train/grad_norm_pre_clip': 16.401592254638672, 'timing/data': 0.02989640925079584, 'timing/model': 1.5723509807139635, 'learning_rate/qwen_vl_interface': 1.2187183760057431e-06, 'learning_rate/action_model': 1.218718376005743e-05, 'epoch': 2.22}) +[2026-06-18 06:19:23,648][starVLA.training.train_starvla][INFO] - Step 5777, Loss: {'train/loss': 0.24448436498641968, 'train/grad_norm_pre_clip': 13.298561096191406, 'timing/data': 0.04191318526864052, 'timing/model': 1.524570726789534, 'learning_rate/qwen_vl_interface': 1.2175749701414393e-06, 'learning_rate/action_model': 1.2175749701414394e-05, 'epoch': 2.22}) +[2026-06-18 06:19:32,425][starVLA.training.train_starvla][INFO] - Step 5778, Loss: {'train/loss': 0.26188626885414124, 'train/grad_norm_pre_clip': 13.947759628295898, 'timing/data': 0.10727433860301971, 'timing/model': 2.7678019469603896, 'learning_rate/qwen_vl_interface': 1.216432400203184e-06, 'learning_rate/action_model': 1.2164324002031841e-05, 'epoch': 2.22}) +[2026-06-18 06:19:39,721][starVLA.training.train_starvla][INFO] - Step 5779, Loss: {'train/loss': 0.5978478193283081, 'train/grad_norm_pre_clip': 15.73973274230957, 'timing/data': 0.046331992372870445, 'timing/model': 2.959742158651352, 'learning_rate/qwen_vl_interface': 1.215290666427832e-06, 'learning_rate/action_model': 1.2152906664278318e-05, 'epoch': 2.22}) +[2026-06-18 06:19:58,495][starVLA.training.train_starvla][INFO] - Step 5780, Loss: {'train/loss': 0.33947181701660156, 'train/grad_norm_pre_clip': 13.33650016784668, 'timing/data': 0.06290776282548904, 'timing/model': 6.0504484958946705, 'learning_rate/qwen_vl_interface': 1.214149769052067e-06, 'learning_rate/action_model': 1.214149769052067e-05, 'epoch': 2.22}) +[2026-06-18 06:20:12,197][starVLA.training.train_starvla][INFO] - Step 5781, Loss: {'train/loss': 0.7417111992835999, 'train/grad_norm_pre_clip': 20.726463317871094, 'timing/data': 0.09779107291251421, 'timing/model': 5.246819416992366, 'learning_rate/qwen_vl_interface': 1.2130097083123971e-06, 'learning_rate/action_model': 1.213009708312397e-05, 'epoch': 2.22}) +[2026-06-18 06:20:24,234][starVLA.training.train_starvla][INFO] - Step 5782, Loss: {'train/loss': 0.36635905504226685, 'train/grad_norm_pre_clip': 16.062524795532227, 'timing/data': 0.017228659242391586, 'timing/model': 1.5014587752521038, 'learning_rate/qwen_vl_interface': 1.2118704844451587e-06, 'learning_rate/action_model': 1.2118704844451586e-05, 'epoch': 2.22}) +[2026-06-18 06:20:28,748][starVLA.training.train_starvla][INFO] - Step 5783, Loss: {'train/loss': 0.407980352640152, 'train/grad_norm_pre_clip': 18.666418075561523, 'timing/data': 0.03776467405259609, 'timing/model': 1.1900219013914466, 'learning_rate/qwen_vl_interface': 1.2107320976865138e-06, 'learning_rate/action_model': 1.2107320976865139e-05, 'epoch': 2.22}) +[2026-06-18 06:20:42,406][starVLA.training.train_starvla][INFO] - Step 5784, Loss: {'train/loss': 0.5393714904785156, 'train/grad_norm_pre_clip': 24.555641174316406, 'timing/data': 0.034813414327800274, 'timing/model': 2.6038348181173205, 'learning_rate/qwen_vl_interface': 1.209594548272451e-06, 'learning_rate/action_model': 1.209594548272451e-05, 'epoch': 2.23}) +[2026-06-18 06:20:53,129][starVLA.training.train_starvla][INFO] - Step 5785, Loss: {'train/loss': 0.2898332476615906, 'train/grad_norm_pre_clip': 13.771315574645996, 'timing/data': 0.05679526273161173, 'timing/model': 1.1099953204393387, 'learning_rate/qwen_vl_interface': 1.2084578364387869e-06, 'learning_rate/action_model': 1.208457836438787e-05, 'epoch': 2.23}) +[2026-06-18 06:20:57,913][starVLA.training.train_starvla][INFO] - Step 5786, Loss: {'train/loss': 0.23301759362220764, 'train/grad_norm_pre_clip': 10.204655647277832, 'timing/data': 0.032288878224790096, 'timing/model': 1.4532212726771832, 'learning_rate/qwen_vl_interface': 1.2073219624211635e-06, 'learning_rate/action_model': 1.2073219624211635e-05, 'epoch': 2.23}) +[2026-06-18 06:21:02,802][starVLA.training.train_starvla][INFO] - Step 5787, Loss: {'train/loss': 0.32970502972602844, 'train/grad_norm_pre_clip': 12.787422180175781, 'timing/data': 0.031235660426318645, 'timing/model': 1.6000269195064902, 'learning_rate/qwen_vl_interface': 1.2061869264550458e-06, 'learning_rate/action_model': 1.2061869264550457e-05, 'epoch': 2.23}) +[2026-06-18 06:21:07,956][starVLA.training.train_starvla][INFO] - Step 5788, Loss: {'train/loss': 0.21107521653175354, 'train/grad_norm_pre_clip': 10.803955078125, 'timing/data': 0.033117073588073254, 'timing/model': 1.2785383816808462, 'learning_rate/qwen_vl_interface': 1.2050527287757288e-06, 'learning_rate/action_model': 1.2050527287757289e-05, 'epoch': 2.23}) +[2026-06-18 06:21:13,245][starVLA.training.train_starvla][INFO] - Step 5789, Loss: {'train/loss': 0.1379745602607727, 'train/grad_norm_pre_clip': 9.932024002075195, 'timing/data': 0.034888578578829765, 'timing/model': 1.1074464339762926, 'learning_rate/qwen_vl_interface': 1.203919369618335e-06, 'learning_rate/action_model': 1.2039193696183348e-05, 'epoch': 2.23}) +[2026-06-18 06:21:18,639][starVLA.training.train_starvla][INFO] - Step 5790, Loss: {'train/loss': 0.16438958048820496, 'train/grad_norm_pre_clip': 10.283939361572266, 'timing/data': 0.03520233277231455, 'timing/model': 1.1218473687767982, 'learning_rate/qwen_vl_interface': 1.202786849217809e-06, 'learning_rate/action_model': 1.202786849217809e-05, 'epoch': 2.23}) +[2026-06-18 06:21:23,835][starVLA.training.train_starvla][INFO] - Step 5791, Loss: {'train/loss': 0.23508478701114655, 'train/grad_norm_pre_clip': 10.565491676330566, 'timing/data': 0.034457312896847725, 'timing/model': 1.1790776476264, 'learning_rate/qwen_vl_interface': 1.2016551678089247e-06, 'learning_rate/action_model': 1.2016551678089246e-05, 'epoch': 2.23}) +[2026-06-18 06:21:28,690][starVLA.training.train_starvla][INFO] - Step 5792, Loss: {'train/loss': 0.20604966580867767, 'train/grad_norm_pre_clip': 12.140730857849121, 'timing/data': 0.036096153780817986, 'timing/model': 1.553603082895279, 'learning_rate/qwen_vl_interface': 1.2005243256262803e-06, 'learning_rate/action_model': 1.2005243256262802e-05, 'epoch': 2.23}) +[2026-06-18 06:21:33,414][starVLA.training.train_starvla][INFO] - Step 5793, Loss: {'train/loss': 0.12158747017383575, 'train/grad_norm_pre_clip': 6.7195048332214355, 'timing/data': 0.02988758310675621, 'timing/model': 1.2038069050759077, 'learning_rate/qwen_vl_interface': 1.1993943229043003e-06, 'learning_rate/action_model': 1.1993943229043003e-05, 'epoch': 2.23}) +[2026-06-18 06:21:38,539][starVLA.training.train_starvla][INFO] - Step 5794, Loss: {'train/loss': 0.20289713144302368, 'train/grad_norm_pre_clip': 10.4781494140625, 'timing/data': 0.033547538332641125, 'timing/model': 1.1274848952889442, 'learning_rate/qwen_vl_interface': 1.1982651598772358e-06, 'learning_rate/action_model': 1.1982651598772356e-05, 'epoch': 2.23}) +[2026-06-18 06:21:43,383][starVLA.training.train_starvla][INFO] - Step 5795, Loss: {'train/loss': 0.34146782755851746, 'train/grad_norm_pre_clip': 15.35328483581543, 'timing/data': 0.04473056644201279, 'timing/model': 1.186960380524397, 'learning_rate/qwen_vl_interface': 1.197136836779164e-06, 'learning_rate/action_model': 1.1971368367791639e-05, 'epoch': 2.23}) +[2026-06-18 06:21:48,844][starVLA.training.train_starvla][INFO] - Step 5796, Loss: {'train/loss': 0.2672557532787323, 'train/grad_norm_pre_clip': 14.233363151550293, 'timing/data': 0.03040059097111225, 'timing/model': 1.090215123258531, 'learning_rate/qwen_vl_interface': 1.1960093538439866e-06, 'learning_rate/action_model': 1.1960093538439864e-05, 'epoch': 2.23}) +[2026-06-18 06:21:54,086][starVLA.training.train_starvla][INFO] - Step 5797, Loss: {'train/loss': 0.43265578150749207, 'train/grad_norm_pre_clip': 18.576356887817383, 'timing/data': 0.03122144192457199, 'timing/model': 1.6473944699391723, 'learning_rate/qwen_vl_interface': 1.1948827113054323e-06, 'learning_rate/action_model': 1.1948827113054323e-05, 'epoch': 2.23}) +[2026-06-18 06:21:59,199][starVLA.training.train_starvla][INFO] - Step 5798, Loss: {'train/loss': 0.32972416281700134, 'train/grad_norm_pre_clip': 13.29909610748291, 'timing/data': 0.04006355348974466, 'timing/model': 1.399269881658256, 'learning_rate/qwen_vl_interface': 1.1937569093970558e-06, 'learning_rate/action_model': 1.1937569093970558e-05, 'epoch': 2.23}) +[2026-06-18 06:22:04,417][starVLA.training.train_starvla][INFO] - Step 5799, Loss: {'train/loss': 0.20412664115428925, 'train/grad_norm_pre_clip': 13.150955200195312, 'timing/data': 0.0349019942805171, 'timing/model': 1.3077125931158662, 'learning_rate/qwen_vl_interface': 1.1926319483522353e-06, 'learning_rate/action_model': 1.1926319483522353e-05, 'epoch': 2.23}) +[2026-06-18 06:28:47,462][starVLA.training.train_starvla][INFO] - Step 5800, Loss: {'train/loss': 0.14109593629837036, 'train/grad_norm_pre_clip': 12.031106948852539, 'eval/loss': 0.4210118353366852, 'eval/latency_4/loss': 0.4210118353366852, 'eval/demon_fire_precision': 0.9685191133954385, 'eval/demon_fire_recall': 0.9792140305293927, 'eval/demon_fire_f1': 0.9738372093023256, 'eval/demon_fire_rate_model': 0.9142437591776799, 'eval/demon_fire_rate_teacher': 0.9042584434654919, 'eval/demon_left_precision': 0.7788944723618091, 'eval/demon_left_recall': 0.7077625570776256, 'eval/demon_left_f1': 0.7416267942583732, 'eval/demon_left_rate_model': 0.11688693098384728, 'eval/demon_left_rate_teacher': 0.12863436123348018, 'eval/demon_right_precision': 0.7939068100358423, 'eval/demon_right_recall': 0.7996389891696751, 'eval/demon_right_f1': 0.7967625899280575, 'eval/demon_right_rate_model': 0.16387665198237886, 'eval/demon_right_rate_teacher': 0.16270190895741557, 'eval/demon_move_f1': 0.7691946920932153, 'eval/demon_cc_f1': 0.9018561679844289, 'eval/latency_4/demon_fire_precision': 0.9685191133954385, 'eval/latency_4/demon_fire_recall': 0.9792140305293927, 'eval/latency_4/demon_fire_f1': 0.9738372093023256, 'eval/latency_4/demon_fire_rate_model': 0.9142437591776799, 'eval/latency_4/demon_fire_rate_teacher': 0.9042584434654919, 'eval/latency_4/demon_left_precision': 0.7788944723618091, 'eval/latency_4/demon_left_recall': 0.7077625570776256, 'eval/latency_4/demon_left_f1': 0.7416267942583732, 'eval/latency_4/demon_left_rate_model': 0.11688693098384728, 'eval/latency_4/demon_left_rate_teacher': 0.12863436123348018, 'eval/latency_4/demon_right_precision': 0.7939068100358423, 'eval/latency_4/demon_right_recall': 0.7996389891696751, 'eval/latency_4/demon_right_f1': 0.7967625899280575, 'eval/latency_4/demon_right_rate_model': 0.16387665198237886, 'eval/latency_4/demon_right_rate_teacher': 0.16270190895741557, 'eval/latency_4/demon_move_f1': 0.7691946920932153, 'eval/latency_4/demon_cc_f1': 0.9018561679844289, 'timing/data': 0.031347284093499184, 'timing/model': 1.2901709210127592, 'learning_rate/qwen_vl_interface': 1.1915078284041793e-06, 'learning_rate/action_model': 1.1915078284041793e-05, 'epoch': 2.23}) +[2026-06-18 06:28:47,466][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_5800_state +[2026-06-18 06:28:47,474][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 06:28:56,571][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_5800_state/model.safetensors +[2026-06-18 06:29:20,651][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_5800_state/optimizer.bin +[2026-06-18 06:29:20,653][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_5800_state/sampler.bin +[2026-06-18 06:29:20,654][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_5800_state/sampler_1.bin +[2026-06-18 06:29:20,673][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_5800_state/random_states_0.pkl +[2026-06-18 06:29:20,680][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 5800 +[2026-06-18 06:29:20,680][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 06:29:20,698][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 06:29:27,704][starVLA.training.train_starvla][INFO] - Step 5801, Loss: {'train/loss': 0.2356654852628708, 'train/grad_norm_pre_clip': 13.420353889465332, 'timing/data': 0.03515280783176422, 'timing/model': 2.7939359759911895, 'learning_rate/qwen_vl_interface': 1.1903845497859185e-06, 'learning_rate/action_model': 1.1903845497859184e-05, 'epoch': 2.23}) +[2026-06-18 06:29:37,605][starVLA.training.train_starvla][INFO] - Step 5802, Loss: {'train/loss': 0.2282375544309616, 'train/grad_norm_pre_clip': 12.435749053955078, 'timing/data': 0.04179415013641119, 'timing/model': 1.8284799698740244, 'learning_rate/qwen_vl_interface': 1.1892621127303067e-06, 'learning_rate/action_model': 1.1892621127303066e-05, 'epoch': 2.23}) +[2026-06-18 06:29:42,742][starVLA.training.train_starvla][INFO] - Step 5803, Loss: {'train/loss': 0.36781349778175354, 'train/grad_norm_pre_clip': 15.73867130279541, 'timing/data': 0.03186914883553982, 'timing/model': 1.4433578262105584, 'learning_rate/qwen_vl_interface': 1.1881405174700292e-06, 'learning_rate/action_model': 1.1881405174700291e-05, 'epoch': 2.23}) +[2026-06-18 06:29:48,329][starVLA.training.train_starvla][INFO] - Step 5804, Loss: {'train/loss': 0.4937073588371277, 'train/grad_norm_pre_clip': 22.764188766479492, 'timing/data': 0.0376173984259367, 'timing/model': 1.3669932028278708, 'learning_rate/qwen_vl_interface': 1.1870197642375935e-06, 'learning_rate/action_model': 1.1870197642375935e-05, 'epoch': 2.23}) +[2026-06-18 06:29:53,197][starVLA.training.train_starvla][INFO] - Step 5805, Loss: {'train/loss': 0.2737905979156494, 'train/grad_norm_pre_clip': 14.452742576599121, 'timing/data': 0.03258852940052748, 'timing/model': 1.0866993134841323, 'learning_rate/qwen_vl_interface': 1.1858998532653326e-06, 'learning_rate/action_model': 1.1858998532653325e-05, 'epoch': 2.23}) +[2026-06-18 06:29:58,453][starVLA.training.train_starvla][INFO] - Step 5806, Loss: {'train/loss': 0.2008231282234192, 'train/grad_norm_pre_clip': 12.36359691619873, 'timing/data': 0.04097239766269922, 'timing/model': 1.6242899354547262, 'learning_rate/qwen_vl_interface': 1.1847807847854054e-06, 'learning_rate/action_model': 1.1847807847854054e-05, 'epoch': 2.23}) +[2026-06-18 06:30:09,796][starVLA.training.train_starvla][INFO] - Step 5807, Loss: {'train/loss': 0.19640815258026123, 'train/grad_norm_pre_clip': 12.58466625213623, 'timing/data': 0.0256818700581789, 'timing/model': 4.995364642702043, 'learning_rate/qwen_vl_interface': 1.1836625590297959e-06, 'learning_rate/action_model': 1.1836625590297958e-05, 'epoch': 2.23}) +[2026-06-18 06:30:14,794][starVLA.training.train_starvla][INFO] - Step 5808, Loss: {'train/loss': 0.3878457844257355, 'train/grad_norm_pre_clip': 16.137954711914062, 'timing/data': 0.0412302790209651, 'timing/model': 1.607543721795082, 'learning_rate/qwen_vl_interface': 1.1825451762303134e-06, 'learning_rate/action_model': 1.1825451762303133e-05, 'epoch': 2.23}) +[2026-06-18 06:30:20,337][starVLA.training.train_starvla][INFO] - Step 5809, Loss: {'train/loss': 0.21017174422740936, 'train/grad_norm_pre_clip': 10.693915367126465, 'timing/data': 0.022790023125708103, 'timing/model': 1.237896946258843, 'learning_rate/qwen_vl_interface': 1.1814286366185923e-06, 'learning_rate/action_model': 1.1814286366185922e-05, 'epoch': 2.23}) +[2026-06-18 06:30:33,256][starVLA.training.train_starvla][INFO] - Step 5810, Loss: {'train/loss': 0.5954985618591309, 'train/grad_norm_pre_clip': 22.286277770996094, 'timing/data': 0.023000095039606094, 'timing/model': 1.103434938006103, 'learning_rate/qwen_vl_interface': 1.1803129404260932e-06, 'learning_rate/action_model': 1.1803129404260931e-05, 'epoch': 2.24}) +[2026-06-18 06:30:38,605][starVLA.training.train_starvla][INFO] - Step 5811, Loss: {'train/loss': 0.3791041970252991, 'train/grad_norm_pre_clip': 19.534515380859375, 'timing/data': 0.023416384123265743, 'timing/model': 1.1880680099129677, 'learning_rate/qwen_vl_interface': 1.1791980878841008e-06, 'learning_rate/action_model': 1.1791980878841008e-05, 'epoch': 2.24}) +[2026-06-18 06:30:43,900][starVLA.training.train_starvla][INFO] - Step 5812, Loss: {'train/loss': 0.41556215286254883, 'train/grad_norm_pre_clip': 15.026930809020996, 'timing/data': 0.02265507820993662, 'timing/model': 1.7137204864993691, 'learning_rate/qwen_vl_interface': 1.1780840792237246e-06, 'learning_rate/action_model': 1.1780840792237246e-05, 'epoch': 2.24}) +[2026-06-18 06:30:49,016][starVLA.training.train_starvla][INFO] - Step 5813, Loss: {'train/loss': 0.23716425895690918, 'train/grad_norm_pre_clip': 15.296526908874512, 'timing/data': 0.021841599605977535, 'timing/model': 1.4149788971990347, 'learning_rate/qwen_vl_interface': 1.1769709146759002e-06, 'learning_rate/action_model': 1.1769709146759e-05, 'epoch': 2.24}) +[2026-06-18 06:30:54,301][starVLA.training.train_starvla][INFO] - Step 5814, Loss: {'train/loss': 0.1372540295124054, 'train/grad_norm_pre_clip': 11.127941131591797, 'timing/data': 0.019688984379172325, 'timing/model': 1.3032853873446584, 'learning_rate/qwen_vl_interface': 1.175858594471388e-06, 'learning_rate/action_model': 1.175858594471388e-05, 'epoch': 2.24}) +[2026-06-18 06:30:58,968][starVLA.training.train_starvla][INFO] - Step 5815, Loss: {'train/loss': 0.27904632687568665, 'train/grad_norm_pre_clip': 12.336801528930664, 'timing/data': 0.028124450705945492, 'timing/model': 1.1381566617637873, 'learning_rate/qwen_vl_interface': 1.174747118840775e-06, 'learning_rate/action_model': 1.1747471188407749e-05, 'epoch': 2.24}) +[2026-06-18 06:31:03,964][starVLA.training.train_starvla][INFO] - Step 5816, Loss: {'train/loss': 0.3014012277126312, 'train/grad_norm_pre_clip': 17.487886428833008, 'timing/data': 0.027638789266347885, 'timing/model': 1.451301365159452, 'learning_rate/qwen_vl_interface': 1.173636488014468e-06, 'learning_rate/action_model': 1.173636488014468e-05, 'epoch': 2.24}) +[2026-06-18 06:31:09,039][starVLA.training.train_starvla][INFO] - Step 5817, Loss: {'train/loss': 0.4350489377975464, 'train/grad_norm_pre_clip': 18.94933319091797, 'timing/data': 0.025453818030655384, 'timing/model': 1.6379218120127916, 'learning_rate/qwen_vl_interface': 1.1725267022227024e-06, 'learning_rate/action_model': 1.1725267022227025e-05, 'epoch': 2.24}) +[2026-06-18 06:31:15,381][starVLA.training.train_starvla][INFO] - Step 5818, Loss: {'train/loss': 0.21350730955600739, 'train/grad_norm_pre_clip': 11.091057777404785, 'timing/data': 0.029016046784818172, 'timing/model': 1.1637851390987635, 'learning_rate/qwen_vl_interface': 1.17141776169554e-06, 'learning_rate/action_model': 1.17141776169554e-05, 'epoch': 2.24}) +[2026-06-18 06:31:20,866][starVLA.training.train_starvla][INFO] - Step 5819, Loss: {'train/loss': 0.5352204442024231, 'train/grad_norm_pre_clip': 24.170469284057617, 'timing/data': 0.024450493045151234, 'timing/model': 1.2340374663472176, 'learning_rate/qwen_vl_interface': 1.170309666662864e-06, 'learning_rate/action_model': 1.170309666662864e-05, 'epoch': 2.24}) +[2026-06-18 06:31:25,918][starVLA.training.train_starvla][INFO] - Step 5820, Loss: {'train/loss': 0.4028972387313843, 'train/grad_norm_pre_clip': 16.326313018798828, 'timing/data': 0.024579303339123726, 'timing/model': 1.3070218237116933, 'learning_rate/qwen_vl_interface': 1.1692024173543858e-06, 'learning_rate/action_model': 1.1692024173543859e-05, 'epoch': 2.24}) +[2026-06-18 06:31:30,875][starVLA.training.train_starvla][INFO] - Step 5821, Loss: {'train/loss': 0.4342955946922302, 'train/grad_norm_pre_clip': 17.257577896118164, 'timing/data': 0.01897740177810192, 'timing/model': 1.1154086459428072, 'learning_rate/qwen_vl_interface': 1.168096013999635e-06, 'learning_rate/action_model': 1.168096013999635e-05, 'epoch': 2.24}) +[2026-06-18 06:31:40,850][starVLA.training.train_starvla][INFO] - Step 5822, Loss: {'train/loss': 0.16300441324710846, 'train/grad_norm_pre_clip': 10.63512897491455, 'timing/data': 0.03459735494107008, 'timing/model': 1.6577718202024698, 'learning_rate/qwen_vl_interface': 1.1669904568279742e-06, 'learning_rate/action_model': 1.1669904568279741e-05, 'epoch': 2.24}) +[2026-06-18 06:31:45,659][starVLA.training.train_starvla][INFO] - Step 5823, Loss: {'train/loss': 0.45116931200027466, 'train/grad_norm_pre_clip': 20.747407913208008, 'timing/data': 0.021745169535279274, 'timing/model': 1.2013259986415505, 'learning_rate/qwen_vl_interface': 1.165885746068585e-06, 'learning_rate/action_model': 1.1658857460685848e-05, 'epoch': 2.24}) +[2026-06-18 06:31:50,947][starVLA.training.train_starvla][INFO] - Step 5824, Loss: {'train/loss': 0.29850298166275024, 'train/grad_norm_pre_clip': 13.259322166442871, 'timing/data': 0.037818100303411484, 'timing/model': 1.3177721658721566, 'learning_rate/qwen_vl_interface': 1.1647818819504754e-06, 'learning_rate/action_model': 1.1647818819504756e-05, 'epoch': 2.24}) +[2026-06-18 06:32:00,889][starVLA.training.train_starvla][INFO] - Step 5825, Loss: {'train/loss': 0.593845784664154, 'train/grad_norm_pre_clip': 21.803054809570312, 'timing/data': 0.02413463592529297, 'timing/model': 2.8546143332496285, 'learning_rate/qwen_vl_interface': 1.1636788647024776e-06, 'learning_rate/action_model': 1.1636788647024776e-05, 'epoch': 2.24}) +[2026-06-18 06:32:05,722][starVLA.training.train_starvla][INFO] - Step 5826, Loss: {'train/loss': 0.3510444462299347, 'train/grad_norm_pre_clip': 13.152390480041504, 'timing/data': 0.022702893242239952, 'timing/model': 1.0946316476911306, 'learning_rate/qwen_vl_interface': 1.1625766945532485e-06, 'learning_rate/action_model': 1.1625766945532484e-05, 'epoch': 2.24}) +[2026-06-18 06:32:10,550][starVLA.training.train_starvla][INFO] - Step 5827, Loss: {'train/loss': 0.3949079215526581, 'train/grad_norm_pre_clip': 14.564413070678711, 'timing/data': 0.02211504429578781, 'timing/model': 1.6223432188853621, 'learning_rate/qwen_vl_interface': 1.1614753717312683e-06, 'learning_rate/action_model': 1.1614753717312681e-05, 'epoch': 2.24}) +[2026-06-18 06:32:15,320][starVLA.training.train_starvla][INFO] - Step 5828, Loss: {'train/loss': 0.31628626585006714, 'train/grad_norm_pre_clip': 13.137990951538086, 'timing/data': 0.022057419642806053, 'timing/model': 1.1607924439013004, 'learning_rate/qwen_vl_interface': 1.1603748964648423e-06, 'learning_rate/action_model': 1.1603748964648422e-05, 'epoch': 2.24}) +[2026-06-18 06:32:34,628][starVLA.training.train_starvla][INFO] - Step 5829, Loss: {'train/loss': 0.3372995853424072, 'train/grad_norm_pre_clip': 18.107213973999023, 'timing/data': 0.022985373623669147, 'timing/model': 5.428113337606192, 'learning_rate/qwen_vl_interface': 1.1592752689821024e-06, 'learning_rate/action_model': 1.1592752689821024e-05, 'epoch': 2.24}) +[2026-06-18 06:32:39,953][starVLA.training.train_starvla][INFO] - Step 5830, Loss: {'train/loss': 0.40445464849472046, 'train/grad_norm_pre_clip': 17.575897216796875, 'timing/data': 0.025846420787274837, 'timing/model': 1.1274373456835747, 'learning_rate/qwen_vl_interface': 1.1581764895110007e-06, 'learning_rate/action_model': 1.1581764895110007e-05, 'epoch': 2.24}) +[2026-06-18 06:32:44,974][starVLA.training.train_starvla][INFO] - Step 5831, Loss: {'train/loss': 0.4436885714530945, 'train/grad_norm_pre_clip': 16.008243560791016, 'timing/data': 0.01996187772601843, 'timing/model': 1.0593716818839312, 'learning_rate/qwen_vl_interface': 1.1570785582793153e-06, 'learning_rate/action_model': 1.1570785582793154e-05, 'epoch': 2.24}) +[2026-06-18 06:32:49,514][starVLA.training.train_starvla][INFO] - Step 5832, Loss: {'train/loss': 0.3161158263683319, 'train/grad_norm_pre_clip': 13.030364990234375, 'timing/data': 0.019657472148537636, 'timing/model': 1.122312474064529, 'learning_rate/qwen_vl_interface': 1.1559814755146484e-06, 'learning_rate/action_model': 1.1559814755146483e-05, 'epoch': 2.24}) +[2026-06-18 06:32:53,877][starVLA.training.train_starvla][INFO] - Step 5833, Loss: {'train/loss': 0.15849153697490692, 'train/grad_norm_pre_clip': 8.076838493347168, 'timing/data': 0.019513112492859364, 'timing/model': 1.0461451606824994, 'learning_rate/qwen_vl_interface': 1.154885241444428e-06, 'learning_rate/action_model': 1.1548852414444277e-05, 'epoch': 2.24}) +[2026-06-18 06:32:59,930][starVLA.training.train_starvla][INFO] - Step 5834, Loss: {'train/loss': 0.11748798191547394, 'train/grad_norm_pre_clip': 5.722809791564941, 'timing/data': 0.01963010523468256, 'timing/model': 1.0633930675685406, 'learning_rate/qwen_vl_interface': 1.1537898562959039e-06, 'learning_rate/action_model': 1.1537898562959037e-05, 'epoch': 2.24}) +[2026-06-18 06:33:04,262][starVLA.training.train_starvla][INFO] - Step 5835, Loss: {'train/loss': 0.4585447311401367, 'train/grad_norm_pre_clip': 17.610626220703125, 'timing/data': 0.020770764909684658, 'timing/model': 1.0733432322740555, 'learning_rate/qwen_vl_interface': 1.152695320296149e-06, 'learning_rate/action_model': 1.1526953202961488e-05, 'epoch': 2.24}) +[2026-06-18 06:33:08,660][starVLA.training.train_starvla][INFO] - Step 5836, Loss: {'train/loss': 0.5429747104644775, 'train/grad_norm_pre_clip': 17.729665756225586, 'timing/data': 0.02841096092015505, 'timing/model': 1.0671109845861793, 'learning_rate/qwen_vl_interface': 1.1516016336720635e-06, 'learning_rate/action_model': 1.1516016336720633e-05, 'epoch': 2.25}) +[2026-06-18 06:33:13,005][starVLA.training.train_starvla][INFO] - Step 5837, Loss: {'train/loss': 0.30979666113853455, 'train/grad_norm_pre_clip': 15.600866317749023, 'timing/data': 0.011142352595925331, 'timing/model': 1.2707734080031514, 'learning_rate/qwen_vl_interface': 1.1505087966503697e-06, 'learning_rate/action_model': 1.1505087966503697e-05, 'epoch': 2.25}) +[2026-06-18 06:33:17,202][starVLA.training.train_starvla][INFO] - Step 5838, Loss: {'train/loss': 0.2884158790111542, 'train/grad_norm_pre_clip': 16.284557342529297, 'timing/data': 0.010982217267155647, 'timing/model': 1.0836911601945758, 'learning_rate/qwen_vl_interface': 1.1494168094576138e-06, 'learning_rate/action_model': 1.1494168094576138e-05, 'epoch': 2.25}) +[2026-06-18 06:33:21,531][starVLA.training.train_starvla][INFO] - Step 5839, Loss: {'train/loss': 0.3165033161640167, 'train/grad_norm_pre_clip': 11.380718231201172, 'timing/data': 0.010949919931590557, 'timing/model': 1.0803925525397062, 'learning_rate/qwen_vl_interface': 1.1483256723201656e-06, 'learning_rate/action_model': 1.1483256723201655e-05, 'epoch': 2.25}) +[2026-06-18 06:33:27,023][starVLA.training.train_starvla][INFO] - Step 5840, Loss: {'train/loss': 0.3617101013660431, 'train/grad_norm_pre_clip': 20.008195877075195, 'timing/data': 0.0240935068577528, 'timing/model': 1.080689176917076, 'learning_rate/qwen_vl_interface': 1.1472353854642187e-06, 'learning_rate/action_model': 1.1472353854642187e-05, 'epoch': 2.25}) +[2026-06-18 06:33:31,424][starVLA.training.train_starvla][INFO] - Step 5841, Loss: {'train/loss': 0.20065948367118835, 'train/grad_norm_pre_clip': 12.944005966186523, 'timing/data': 0.013813414610922337, 'timing/model': 1.0885300934314728, 'learning_rate/qwen_vl_interface': 1.1461459491157914e-06, 'learning_rate/action_model': 1.1461459491157914e-05, 'epoch': 2.25}) +[2026-06-18 06:33:35,668][starVLA.training.train_starvla][INFO] - Step 5842, Loss: {'train/loss': 0.40905141830444336, 'train/grad_norm_pre_clip': 19.563779830932617, 'timing/data': 0.012903519906103611, 'timing/model': 1.1731772357597947, 'learning_rate/qwen_vl_interface': 1.145057363500725e-06, 'learning_rate/action_model': 1.1450573635007251e-05, 'epoch': 2.25}) +[2026-06-18 06:33:44,319][starVLA.training.train_starvla][INFO] - Step 5843, Loss: {'train/loss': 0.4391934871673584, 'train/grad_norm_pre_clip': 20.16970443725586, 'timing/data': 0.01830297987908125, 'timing/model': 4.817658861167729, 'learning_rate/qwen_vl_interface': 1.1439696288446845e-06, 'learning_rate/action_model': 1.1439696288446844e-05, 'epoch': 2.25}) +[2026-06-18 06:33:49,543][starVLA.training.train_starvla][INFO] - Step 5844, Loss: {'train/loss': 0.4670470058917999, 'train/grad_norm_pre_clip': 21.06658172607422, 'timing/data': 0.016676459461450577, 'timing/model': 1.2592467330396175, 'learning_rate/qwen_vl_interface': 1.1428827453731575e-06, 'learning_rate/action_model': 1.1428827453731574e-05, 'epoch': 2.25}) +[2026-06-18 06:33:54,635][starVLA.training.train_starvla][INFO] - Step 5845, Loss: {'train/loss': 0.21986334025859833, 'train/grad_norm_pre_clip': 12.171174049377441, 'timing/data': 0.013525959104299545, 'timing/model': 1.116772647947073, 'learning_rate/qwen_vl_interface': 1.1417967133114575e-06, 'learning_rate/action_model': 1.1417967133114573e-05, 'epoch': 2.25}) +[2026-06-18 06:34:03,831][starVLA.training.train_starvla][INFO] - Step 5846, Loss: {'train/loss': 0.14952565729618073, 'train/grad_norm_pre_clip': 8.906774520874023, 'timing/data': 0.012344279326498508, 'timing/model': 4.28133440297097, 'learning_rate/qwen_vl_interface': 1.140711532884718e-06, 'learning_rate/action_model': 1.140711532884718e-05, 'epoch': 2.25}) +[2026-06-18 06:34:09,286][starVLA.training.train_starvla][INFO] - Step 5847, Loss: {'train/loss': 0.36473190784454346, 'train/grad_norm_pre_clip': 20.52041244506836, 'timing/data': 0.016031739301979542, 'timing/model': 1.7983625214546919, 'learning_rate/qwen_vl_interface': 1.1396272043179004e-06, 'learning_rate/action_model': 1.1396272043179005e-05, 'epoch': 2.25}) +[2026-06-18 06:34:17,703][starVLA.training.train_starvla][INFO] - Step 5848, Loss: {'train/loss': 0.2606768310070038, 'train/grad_norm_pre_clip': 16.484609603881836, 'timing/data': 0.019327052868902683, 'timing/model': 1.9943346455693245, 'learning_rate/qwen_vl_interface': 1.138543727835788e-06, 'learning_rate/action_model': 1.138543727835788e-05, 'epoch': 2.25}) +[2026-06-18 06:34:23,262][starVLA.training.train_starvla][INFO] - Step 5849, Loss: {'train/loss': 0.17414624989032745, 'train/grad_norm_pre_clip': 9.403691291809082, 'timing/data': 0.012465506792068481, 'timing/model': 1.2404790138825774, 'learning_rate/qwen_vl_interface': 1.1374611036629832e-06, 'learning_rate/action_model': 1.1374611036629831e-05, 'epoch': 2.25}) +[2026-06-18 06:34:28,406][starVLA.training.train_starvla][INFO] - Step 5850, Loss: {'train/loss': 0.2732033133506775, 'train/grad_norm_pre_clip': 19.14124870300293, 'timing/data': 0.012794431298971176, 'timing/model': 1.1880682688206434, 'learning_rate/qwen_vl_interface': 1.136379332023916e-06, 'learning_rate/action_model': 1.1363793320239161e-05, 'epoch': 2.25}) +[2026-06-18 06:34:40,284][starVLA.training.train_starvla][INFO] - Step 5851, Loss: {'train/loss': 0.25221529603004456, 'train/grad_norm_pre_clip': 12.699899673461914, 'timing/data': 0.016380449756979942, 'timing/model': 1.3993897987529635, 'learning_rate/qwen_vl_interface': 1.1352984131428413e-06, 'learning_rate/action_model': 1.1352984131428413e-05, 'epoch': 2.25}) +[2026-06-18 06:34:45,486][starVLA.training.train_starvla][INFO] - Step 5852, Loss: {'train/loss': 0.36075201630592346, 'train/grad_norm_pre_clip': 12.398791313171387, 'timing/data': 0.012393043376505375, 'timing/model': 1.7854887200519443, 'learning_rate/qwen_vl_interface': 1.1342183472438329e-06, 'learning_rate/action_model': 1.1342183472438328e-05, 'epoch': 2.25}) +[2026-06-18 06:34:50,200][starVLA.training.train_starvla][INFO] - Step 5853, Loss: {'train/loss': 0.37284329533576965, 'train/grad_norm_pre_clip': 16.27052116394043, 'timing/data': 0.01294033508747816, 'timing/model': 1.1779830567538738, 'learning_rate/qwen_vl_interface': 1.1331391345507905e-06, 'learning_rate/action_model': 1.1331391345507905e-05, 'epoch': 2.25}) +[2026-06-18 06:34:55,200][starVLA.training.train_starvla][INFO] - Step 5854, Loss: {'train/loss': 0.15864017605781555, 'train/grad_norm_pre_clip': 8.311741828918457, 'timing/data': 0.011552010662853718, 'timing/model': 1.1821646885946393, 'learning_rate/qwen_vl_interface': 1.1320607752874341e-06, 'learning_rate/action_model': 1.1320607752874341e-05, 'epoch': 2.25}) +[2026-06-18 06:35:00,543][starVLA.training.train_starvla][INFO] - Step 5855, Loss: {'train/loss': 0.4120633006095886, 'train/grad_norm_pre_clip': 17.346424102783203, 'timing/data': 0.01316991075873375, 'timing/model': 1.1519385436549783, 'learning_rate/qwen_vl_interface': 1.1309832696773104e-06, 'learning_rate/action_model': 1.1309832696773104e-05, 'epoch': 2.25}) +[2026-06-18 06:35:05,579][starVLA.training.train_starvla][INFO] - Step 5856, Loss: {'train/loss': 0.4735318720340729, 'train/grad_norm_pre_clip': 15.120166778564453, 'timing/data': 0.01281354296952486, 'timing/model': 1.199230633676052, 'learning_rate/qwen_vl_interface': 1.1299066179437876e-06, 'learning_rate/action_model': 1.1299066179437876e-05, 'epoch': 2.25}) +[2026-06-18 06:35:10,464][starVLA.training.train_starvla][INFO] - Step 5857, Loss: {'train/loss': 0.1396992951631546, 'train/grad_norm_pre_clip': 12.082605361938477, 'timing/data': 0.01435953937470913, 'timing/model': 1.4273041291162372, 'learning_rate/qwen_vl_interface': 1.1288308203100565e-06, 'learning_rate/action_model': 1.1288308203100564e-05, 'epoch': 2.25}) +[2026-06-18 06:35:15,519][starVLA.training.train_starvla][INFO] - Step 5858, Loss: {'train/loss': 0.2675602436065674, 'train/grad_norm_pre_clip': 15.755597114562988, 'timing/data': 0.01202897448092699, 'timing/model': 1.2187122097238898, 'learning_rate/qwen_vl_interface': 1.12775587699913e-06, 'learning_rate/action_model': 1.12775587699913e-05, 'epoch': 2.25}) +[2026-06-18 06:35:20,856][starVLA.training.train_starvla][INFO] - Step 5859, Loss: {'train/loss': 0.5155996084213257, 'train/grad_norm_pre_clip': 18.064495086669922, 'timing/data': 0.015135417692363262, 'timing/model': 1.3308210326358676, 'learning_rate/qwen_vl_interface': 1.126681788233847e-06, 'learning_rate/action_model': 1.1266817882338468e-05, 'epoch': 2.25}) +[2026-06-18 06:35:33,508][starVLA.training.train_starvla][INFO] - Step 5860, Loss: {'train/loss': 0.16332048177719116, 'train/grad_norm_pre_clip': 10.160608291625977, 'timing/data': 0.013900249265134335, 'timing/model': 5.2223919332027435, 'learning_rate/qwen_vl_interface': 1.1256085542368655e-06, 'learning_rate/action_model': 1.1256085542368655e-05, 'epoch': 2.25}) +[2026-06-18 06:35:38,969][starVLA.training.train_starvla][INFO] - Step 5861, Loss: {'train/loss': 0.33651602268218994, 'train/grad_norm_pre_clip': 30.310791015625, 'timing/data': 0.020657164976000786, 'timing/model': 1.3954649334773421, 'learning_rate/qwen_vl_interface': 1.1245361752306676e-06, 'learning_rate/action_model': 1.1245361752306677e-05, 'epoch': 2.25}) +[2026-06-18 06:35:44,202][starVLA.training.train_starvla][INFO] - Step 5862, Loss: {'train/loss': 0.12322445213794708, 'train/grad_norm_pre_clip': 7.096709728240967, 'timing/data': 0.012989038601517677, 'timing/model': 1.6310701658949256, 'learning_rate/qwen_vl_interface': 1.1234646514375625e-06, 'learning_rate/action_model': 1.1234646514375624e-05, 'epoch': 2.26}) +[2026-06-18 06:35:53,539][starVLA.training.train_starvla][INFO] - Step 5863, Loss: {'train/loss': 0.519174337387085, 'train/grad_norm_pre_clip': 18.364654541015625, 'timing/data': 0.012741665355861187, 'timing/model': 1.2572118993848562, 'learning_rate/qwen_vl_interface': 1.122393983079674e-06, 'learning_rate/action_model': 1.122393983079674e-05, 'epoch': 2.26}) +[2026-06-18 06:36:04,090][starVLA.training.train_starvla][INFO] - Step 5864, Loss: {'train/loss': 0.2788672149181366, 'train/grad_norm_pre_clip': 18.386747360229492, 'timing/data': 0.014444445259869099, 'timing/model': 2.0325126126408577, 'learning_rate/qwen_vl_interface': 1.1213241703789534e-06, 'learning_rate/action_model': 1.1213241703789534e-05, 'epoch': 2.26}) +[2026-06-18 06:36:09,590][starVLA.training.train_starvla][INFO] - Step 5865, Loss: {'train/loss': 0.12351801991462708, 'train/grad_norm_pre_clip': 6.787344455718994, 'timing/data': 0.015617596916854382, 'timing/model': 1.1538761164993048, 'learning_rate/qwen_vl_interface': 1.1202552135571759e-06, 'learning_rate/action_model': 1.1202552135571759e-05, 'epoch': 2.26}) +[2026-06-18 06:36:15,641][starVLA.training.train_starvla][INFO] - Step 5866, Loss: {'train/loss': 0.45388686656951904, 'train/grad_norm_pre_clip': 12.0598783493042, 'timing/data': 0.014199919998645782, 'timing/model': 1.5382645186036825, 'learning_rate/qwen_vl_interface': 1.1191871128359367e-06, 'learning_rate/action_model': 1.1191871128359365e-05, 'epoch': 2.26}) +[2026-06-18 06:36:21,089][starVLA.training.train_starvla][INFO] - Step 5867, Loss: {'train/loss': 0.18801403045654297, 'train/grad_norm_pre_clip': 9.962114334106445, 'timing/data': 0.012209434993565083, 'timing/model': 1.6585672218352556, 'learning_rate/qwen_vl_interface': 1.1181198684366548e-06, 'learning_rate/action_model': 1.1181198684366548e-05, 'epoch': 2.26}) +[2026-06-18 06:36:32,546][starVLA.training.train_starvla][INFO] - Step 5868, Loss: {'train/loss': 0.19073747098445892, 'train/grad_norm_pre_clip': 15.944681167602539, 'timing/data': 0.0012339111417531967, 'timing/model': 1.1043436098843813, 'learning_rate/qwen_vl_interface': 1.117053480580568e-06, 'learning_rate/action_model': 1.1170534805805679e-05, 'epoch': 2.26}) +[2026-06-18 06:36:44,224][starVLA.training.train_starvla][INFO] - Step 5869, Loss: {'train/loss': 0.4841582477092743, 'train/grad_norm_pre_clip': 16.1373348236084, 'timing/data': 0.0034956522285938263, 'timing/model': 3.1202698228880763, 'learning_rate/qwen_vl_interface': 1.1159879494887429e-06, 'learning_rate/action_model': 1.115987949488743e-05, 'epoch': 2.26}) +[2026-06-18 06:36:58,893][starVLA.training.train_starvla][INFO] - Step 5870, Loss: {'train/loss': 0.10668494552373886, 'train/grad_norm_pre_clip': 10.800153732299805, 'timing/data': 0.0012837490066885948, 'timing/model': 4.8976361667737365, 'learning_rate/qwen_vl_interface': 1.1149232753820645e-06, 'learning_rate/action_model': 1.1149232753820647e-05, 'epoch': 2.26}) +[2026-06-18 06:37:11,538][starVLA.training.train_starvla][INFO] - Step 5871, Loss: {'train/loss': 0.2686755061149597, 'train/grad_norm_pre_clip': 13.250348091125488, 'timing/data': 0.0015099449083209038, 'timing/model': 3.0161595586687326, 'learning_rate/qwen_vl_interface': 1.1138594584812406e-06, 'learning_rate/action_model': 1.1138594584812406e-05, 'epoch': 2.26}) +[2026-06-18 06:37:21,825][starVLA.training.train_starvla][INFO] - Step 5872, Loss: {'train/loss': 0.3096562623977661, 'train/grad_norm_pre_clip': 11.052542686462402, 'timing/data': 0.0012681055814027786, 'timing/model': 1.6333872126415372, 'learning_rate/qwen_vl_interface': 1.1127964990068017e-06, 'learning_rate/action_model': 1.1127964990068017e-05, 'epoch': 2.26}) +[2026-06-18 06:37:40,107][starVLA.training.train_starvla][INFO] - Step 5873, Loss: {'train/loss': 0.10616166144609451, 'train/grad_norm_pre_clip': 8.492600440979004, 'timing/data': 0.001454254612326622, 'timing/model': 3.6057245954871178, 'learning_rate/qwen_vl_interface': 1.1117343971791005e-06, 'learning_rate/action_model': 1.1117343971791006e-05, 'epoch': 2.26}) +[2026-06-18 06:37:55,732][starVLA.training.train_starvla][INFO] - Step 5874, Loss: {'train/loss': 0.2657504081726074, 'train/grad_norm_pre_clip': 13.037053108215332, 'timing/data': 0.001347181387245655, 'timing/model': 3.1825905283913016, 'learning_rate/qwen_vl_interface': 1.1106731532183114e-06, 'learning_rate/action_model': 1.1106731532183114e-05, 'epoch': 2.26}) +[2026-06-18 06:38:04,532][starVLA.training.train_starvla][INFO] - Step 5875, Loss: {'train/loss': 0.07868090271949768, 'train/grad_norm_pre_clip': 6.934245586395264, 'timing/data': 0.0014154240489006042, 'timing/model': 1.4337630029767752, 'learning_rate/qwen_vl_interface': 1.1096127673444309e-06, 'learning_rate/action_model': 1.109612767344431e-05, 'epoch': 2.26}) +[2026-06-18 06:38:09,506][starVLA.training.train_starvla][INFO] - Step 5876, Loss: {'train/loss': 0.10319741070270538, 'train/grad_norm_pre_clip': 7.8872880935668945, 'timing/data': 0.0015523293986916542, 'timing/model': 1.1599142225459218, 'learning_rate/qwen_vl_interface': 1.108553239777281e-06, 'learning_rate/action_model': 1.108553239777281e-05, 'epoch': 2.26}) +[2026-06-18 06:38:15,137][starVLA.training.train_starvla][INFO] - Step 5877, Loss: {'train/loss': 0.20816826820373535, 'train/grad_norm_pre_clip': 10.579824447631836, 'timing/data': 0.0011813649907708168, 'timing/model': 1.6501441523432732, 'learning_rate/qwen_vl_interface': 1.1074945707365004e-06, 'learning_rate/action_model': 1.1074945707365003e-05, 'epoch': 2.26}) +[2026-06-18 06:38:19,896][starVLA.training.train_starvla][INFO] - Step 5878, Loss: {'train/loss': 0.3050677180290222, 'train/grad_norm_pre_clip': 17.62333106994629, 'timing/data': 0.0011842399835586548, 'timing/model': 1.152906747534871, 'learning_rate/qwen_vl_interface': 1.1064367604415522e-06, 'learning_rate/action_model': 1.1064367604415521e-05, 'epoch': 2.26}) +[2026-06-18 06:38:25,437][starVLA.training.train_starvla][INFO] - Step 5879, Loss: {'train/loss': 0.22578170895576477, 'train/grad_norm_pre_clip': 10.196794509887695, 'timing/data': 0.0011273333802819252, 'timing/model': 1.3612046269699931, 'learning_rate/qwen_vl_interface': 1.105379809111721e-06, 'learning_rate/action_model': 1.105379809111721e-05, 'epoch': 2.26}) +[2026-06-18 06:38:30,958][starVLA.training.train_starvla][INFO] - Step 5880, Loss: {'train/loss': 0.3107674717903137, 'train/grad_norm_pre_clip': 18.38245964050293, 'timing/data': 0.0014418950304389, 'timing/model': 1.3591493489220738, 'learning_rate/qwen_vl_interface': 1.1043237169661167e-06, 'learning_rate/action_model': 1.1043237169661166e-05, 'epoch': 2.26}) +[2026-06-18 06:38:36,319][starVLA.training.train_starvla][INFO] - Step 5881, Loss: {'train/loss': 0.23443405330181122, 'train/grad_norm_pre_clip': 16.2571964263916, 'timing/data': 0.0013507278636097908, 'timing/model': 1.198873215354979, 'learning_rate/qwen_vl_interface': 1.103268484223667e-06, 'learning_rate/action_model': 1.1032684842236669e-05, 'epoch': 2.26}) +[2026-06-18 06:38:41,308][starVLA.training.train_starvla][INFO] - Step 5882, Loss: {'train/loss': 0.31947585940361023, 'train/grad_norm_pre_clip': 14.826560020446777, 'timing/data': 0.0008827382698655128, 'timing/model': 1.6268866825848818, 'learning_rate/qwen_vl_interface': 1.102214111103123e-06, 'learning_rate/action_model': 1.102214111103123e-05, 'epoch': 2.26}) +[2026-06-18 06:38:46,108][starVLA.training.train_starvla][INFO] - Step 5883, Loss: {'train/loss': 0.1578788012266159, 'train/grad_norm_pre_clip': 9.534812927246094, 'timing/data': 0.001184975728392601, 'timing/model': 1.1696444628760219, 'learning_rate/qwen_vl_interface': 1.1011605978230552e-06, 'learning_rate/action_model': 1.1011605978230551e-05, 'epoch': 2.26}) +[2026-06-18 06:38:51,462][starVLA.training.train_starvla][INFO] - Step 5884, Loss: {'train/loss': 0.3978580832481384, 'train/grad_norm_pre_clip': 17.18022918701172, 'timing/data': 0.000929882749915123, 'timing/model': 1.2784479903057218, 'learning_rate/qwen_vl_interface': 1.100107944601861e-06, 'learning_rate/action_model': 1.100107944601861e-05, 'epoch': 2.26}) +[2026-06-18 06:38:56,739][starVLA.training.train_starvla][INFO] - Step 5885, Loss: {'train/loss': 0.5258165597915649, 'train/grad_norm_pre_clip': 20.24013328552246, 'timing/data': 0.001753881573677063, 'timing/model': 1.3405076377093792, 'learning_rate/qwen_vl_interface': 1.0990561516577552e-06, 'learning_rate/action_model': 1.0990561516577552e-05, 'epoch': 2.26}) +[2026-06-18 06:39:02,183][starVLA.training.train_starvla][INFO] - Step 5886, Loss: {'train/loss': 0.5119353532791138, 'train/grad_norm_pre_clip': 17.271944046020508, 'timing/data': 0.0013443101197481155, 'timing/model': 1.1288137482479215, 'learning_rate/qwen_vl_interface': 1.0980052192087753e-06, 'learning_rate/action_model': 1.0980052192087754e-05, 'epoch': 2.26}) +[2026-06-18 06:39:06,985][starVLA.training.train_starvla][INFO] - Step 5887, Loss: {'train/loss': 0.36220523715019226, 'train/grad_norm_pre_clip': 13.1233549118042, 'timing/data': 0.0012115584686398506, 'timing/model': 1.4511690819635987, 'learning_rate/qwen_vl_interface': 1.096955147472782e-06, 'learning_rate/action_model': 1.0969551474727821e-05, 'epoch': 2.26}) +[2026-06-18 06:39:11,789][starVLA.training.train_starvla][INFO] - Step 5888, Loss: {'train/loss': 0.36795464158058167, 'train/grad_norm_pre_clip': 13.519664764404297, 'timing/data': 0.0012485114857554436, 'timing/model': 1.1420741928741336, 'learning_rate/qwen_vl_interface': 1.095905936667455e-06, 'learning_rate/action_model': 1.0959059366674552e-05, 'epoch': 2.27}) +[2026-06-18 06:39:16,737][starVLA.training.train_starvla][INFO] - Step 5889, Loss: {'train/loss': 0.23275113105773926, 'train/grad_norm_pre_clip': 11.703598022460938, 'timing/data': 0.0010821549221873283, 'timing/model': 1.1640480300411582, 'learning_rate/qwen_vl_interface': 1.0948575870102977e-06, 'learning_rate/action_model': 1.0948575870102978e-05, 'epoch': 2.27}) +[2026-06-18 06:39:25,536][starVLA.training.train_starvla][INFO] - Step 5890, Loss: {'train/loss': 0.31008195877075195, 'train/grad_norm_pre_clip': 13.778897285461426, 'timing/data': 0.0012040706351399422, 'timing/model': 1.313618116080761, 'learning_rate/qwen_vl_interface': 1.0938100987186325e-06, 'learning_rate/action_model': 1.0938100987186323e-05, 'epoch': 2.27}) +[2026-06-18 06:39:33,158][starVLA.training.train_starvla][INFO] - Step 5891, Loss: {'train/loss': 0.19701902568340302, 'train/grad_norm_pre_clip': 15.038629531860352, 'timing/data': 0.0011854767799377441, 'timing/model': 1.6074133394286036, 'learning_rate/qwen_vl_interface': 1.092763472009608e-06, 'learning_rate/action_model': 1.092763472009608e-05, 'epoch': 2.27}) +[2026-06-18 06:39:51,744][starVLA.training.train_starvla][INFO] - Step 5892, Loss: {'train/loss': 0.18943382799625397, 'train/grad_norm_pre_clip': 12.762632369995117, 'timing/data': 0.0008803969249129295, 'timing/model': 3.8359884871169925, 'learning_rate/qwen_vl_interface': 1.0917177071001884e-06, 'learning_rate/action_model': 1.0917177071001884e-05, 'epoch': 2.27}) +[2026-06-18 06:40:00,627][starVLA.training.train_starvla][INFO] - Step 5893, Loss: {'train/loss': 0.26845160126686096, 'train/grad_norm_pre_clip': 19.016191482543945, 'timing/data': 0.001254253089427948, 'timing/model': 1.4340108940377831, 'learning_rate/qwen_vl_interface': 1.0906728042071627e-06, 'learning_rate/action_model': 1.0906728042071627e-05, 'epoch': 2.27}) +[2026-06-18 06:40:06,615][starVLA.training.train_starvla][INFO] - Step 5894, Loss: {'train/loss': 0.26418253779411316, 'train/grad_norm_pre_clip': 16.605144500732422, 'timing/data': 0.0013018753379583359, 'timing/model': 1.1925510270521045, 'learning_rate/qwen_vl_interface': 1.089628763547139e-06, 'learning_rate/action_model': 1.0896287635471389e-05, 'epoch': 2.27}) +[2026-06-18 06:40:22,321][starVLA.training.train_starvla][INFO] - Step 5895, Loss: {'train/loss': 0.08234353363513947, 'train/grad_norm_pre_clip': 6.556406497955322, 'timing/data': 0.0013408716768026352, 'timing/model': 1.4801865564659238, 'learning_rate/qwen_vl_interface': 1.0885855853365505e-06, 'learning_rate/action_model': 1.0885855853365503e-05, 'epoch': 2.27}) +[2026-06-18 06:40:31,156][starVLA.training.train_starvla][INFO] - Step 5896, Loss: {'train/loss': 0.11976274102926254, 'train/grad_norm_pre_clip': 8.266828536987305, 'timing/data': 0.0011732438579201698, 'timing/model': 2.216533435508609, 'learning_rate/qwen_vl_interface': 1.0875432697916492e-06, 'learning_rate/action_model': 1.087543269791649e-05, 'epoch': 2.27}) +[2026-06-18 06:40:41,819][starVLA.training.train_starvla][INFO] - Step 5897, Loss: {'train/loss': 0.3976741135120392, 'train/grad_norm_pre_clip': 17.13861083984375, 'timing/data': 0.004229690879583359, 'timing/model': 1.7809287197887897, 'learning_rate/qwen_vl_interface': 1.0865018171285047e-06, 'learning_rate/action_model': 1.0865018171285047e-05, 'epoch': 2.27}) +[2026-06-18 06:40:50,724][starVLA.training.train_starvla][INFO] - Step 5898, Loss: {'train/loss': 0.24815842509269714, 'train/grad_norm_pre_clip': 14.83467960357666, 'timing/data': 0.0011976473033428192, 'timing/model': 1.1974656721577048, 'learning_rate/qwen_vl_interface': 1.0854612275630146e-06, 'learning_rate/action_model': 1.0854612275630146e-05, 'epoch': 2.27}) +[2026-06-18 06:41:02,294][starVLA.training.train_starvla][INFO] - Step 5899, Loss: {'train/loss': 0.5394116044044495, 'train/grad_norm_pre_clip': 23.895299911499023, 'timing/data': 0.0011627217754721642, 'timing/model': 1.1455554068088531, 'learning_rate/qwen_vl_interface': 1.0844215013108936e-06, 'learning_rate/action_model': 1.0844215013108936e-05, 'epoch': 2.27}) +[2026-06-18 06:45:59,310][starVLA.training.train_starvla][INFO] - Step 5900, Loss: {'train/loss': 0.409302681684494, 'train/grad_norm_pre_clip': 16.07605743408203, 'eval/loss': 0.42309409379959106, 'eval/latency_4/loss': 0.42309409379959106, 'eval/demon_fire_precision': 0.9689500640204866, 'eval/demon_fire_recall': 0.9831113998051315, 'eval/demon_fire_f1': 0.9759793648234726, 'eval/demon_fire_rate_model': 0.9174743024963289, 'eval/demon_fire_rate_teacher': 0.9042584434654919, 'eval/demon_left_precision': 0.8113207547169812, 'eval/demon_left_recall': 0.6872146118721462, 'eval/demon_left_f1': 0.7441285537700865, 'eval/demon_left_rate_model': 0.10895741556534508, 'eval/demon_left_rate_teacher': 0.12863436123348018, 'eval/demon_right_precision': 0.7924187725631769, 'eval/demon_right_recall': 0.7924187725631769, 'eval/demon_right_f1': 0.7924187725631769, 'eval/demon_right_rate_model': 0.16270190895741557, 'eval/demon_right_rate_teacher': 0.16270190895741557, 'eval/demon_move_f1': 0.7682736631666317, 'eval/demon_cc_f1': 0.9027599877909533, 'eval/latency_4/demon_fire_precision': 0.9689500640204866, 'eval/latency_4/demon_fire_recall': 0.9831113998051315, 'eval/latency_4/demon_fire_f1': 0.9759793648234726, 'eval/latency_4/demon_fire_rate_model': 0.9174743024963289, 'eval/latency_4/demon_fire_rate_teacher': 0.9042584434654919, 'eval/latency_4/demon_left_precision': 0.8113207547169812, 'eval/latency_4/demon_left_recall': 0.6872146118721462, 'eval/latency_4/demon_left_f1': 0.7441285537700865, 'eval/latency_4/demon_left_rate_model': 0.10895741556534508, 'eval/latency_4/demon_left_rate_teacher': 0.12863436123348018, 'eval/latency_4/demon_right_precision': 0.7924187725631769, 'eval/latency_4/demon_right_recall': 0.7924187725631769, 'eval/latency_4/demon_right_f1': 0.7924187725631769, 'eval/latency_4/demon_right_rate_model': 0.16270190895741557, 'eval/latency_4/demon_right_rate_teacher': 0.16270190895741557, 'eval/latency_4/demon_move_f1': 0.7682736631666317, 'eval/latency_4/demon_cc_f1': 0.9027599877909533, 'timing/data': 0.0011513996869325638, 'timing/model': 1.6850237334147096, 'learning_rate/qwen_vl_interface': 1.0833826385876772e-06, 'learning_rate/action_model': 1.0833826385876771e-05, 'epoch': 2.27}) +[2026-06-18 06:45:59,312][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_5900_state +[2026-06-18 06:45:59,320][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 06:46:08,474][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_5900_state/model.safetensors +[2026-06-18 06:46:31,979][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_5900_state/optimizer.bin +[2026-06-18 06:46:31,984][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_5900_state/sampler.bin +[2026-06-18 06:46:31,985][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_5900_state/sampler_1.bin +[2026-06-18 06:46:31,988][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_5900_state/random_states_0.pkl +[2026-06-18 06:46:32,005][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 5900 +[2026-06-18 06:46:32,006][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 06:46:32,027][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 06:46:37,812][starVLA.training.train_starvla][INFO] - Step 5901, Loss: {'train/loss': 0.309349000453949, 'train/grad_norm_pre_clip': 14.102412223815918, 'timing/data': 0.0016350923106074333, 'timing/model': 1.8277627853676677, 'learning_rate/qwen_vl_interface': 1.0823446396087239e-06, 'learning_rate/action_model': 1.0823446396087239e-05, 'epoch': 2.27}) +[2026-06-18 06:46:43,285][starVLA.training.train_starvla][INFO] - Step 5902, Loss: {'train/loss': 0.1426476687192917, 'train/grad_norm_pre_clip': 8.186524391174316, 'timing/data': 0.0012274831533432007, 'timing/model': 1.4561888994649053, 'learning_rate/qwen_vl_interface': 1.0813075045892116e-06, 'learning_rate/action_model': 1.0813075045892116e-05, 'epoch': 2.27}) +[2026-06-18 06:46:47,966][starVLA.training.train_starvla][INFO] - Step 5903, Loss: {'train/loss': 0.16586226224899292, 'train/grad_norm_pre_clip': 12.0326509475708, 'timing/data': 0.0012934748083353043, 'timing/model': 1.1407107999548316, 'learning_rate/qwen_vl_interface': 1.0802712337441383e-06, 'learning_rate/action_model': 1.0802712337441384e-05, 'epoch': 2.27}) +[2026-06-18 06:46:58,599][starVLA.training.train_starvla][INFO] - Step 5904, Loss: {'train/loss': 0.4004340171813965, 'train/grad_norm_pre_clip': 16.400163650512695, 'timing/data': 0.0012050475925207138, 'timing/model': 1.6171254441142082, 'learning_rate/qwen_vl_interface': 1.0792358272883262e-06, 'learning_rate/action_model': 1.0792358272883262e-05, 'epoch': 2.27}) +[2026-06-18 06:47:12,309][starVLA.training.train_starvla][INFO] - Step 5905, Loss: {'train/loss': 0.17820577323436737, 'train/grad_norm_pre_clip': 11.563591003417969, 'timing/data': 0.001638774760067463, 'timing/model': 1.4740171758458018, 'learning_rate/qwen_vl_interface': 1.0782012854364145e-06, 'learning_rate/action_model': 1.0782012854364145e-05, 'epoch': 2.27}) +[2026-06-18 06:47:21,667][starVLA.training.train_starvla][INFO] - Step 5906, Loss: {'train/loss': 0.2452586591243744, 'train/grad_norm_pre_clip': 16.661270141601562, 'timing/data': 0.00305327121168375, 'timing/model': 5.207082177512348, 'learning_rate/qwen_vl_interface': 1.0771676084028652e-06, 'learning_rate/action_model': 1.0771676084028652e-05, 'epoch': 2.27}) +[2026-06-18 06:47:28,301][starVLA.training.train_starvla][INFO] - Step 5907, Loss: {'train/loss': 0.21454794704914093, 'train/grad_norm_pre_clip': 13.619793891906738, 'timing/data': 0.0011977199465036392, 'timing/model': 1.9815514096990228, 'learning_rate/qwen_vl_interface': 1.0761347964019611e-06, 'learning_rate/action_model': 1.076134796401961e-05, 'epoch': 2.27}) +[2026-06-18 06:47:34,375][starVLA.training.train_starvla][INFO] - Step 5908, Loss: {'train/loss': 0.34024038910865784, 'train/grad_norm_pre_clip': 19.830944061279297, 'timing/data': 0.0023079561069607735, 'timing/model': 1.4769472740590572, 'learning_rate/qwen_vl_interface': 1.075102849647804e-06, 'learning_rate/action_model': 1.0751028496478038e-05, 'epoch': 2.27}) +[2026-06-18 06:47:39,896][starVLA.training.train_starvla][INFO] - Step 5909, Loss: {'train/loss': 0.1999971866607666, 'train/grad_norm_pre_clip': 15.258864402770996, 'timing/data': 0.0011713309213519096, 'timing/model': 1.2953281607478857, 'learning_rate/qwen_vl_interface': 1.0740717683543193e-06, 'learning_rate/action_model': 1.0740717683543191e-05, 'epoch': 2.27}) +[2026-06-18 06:47:45,346][starVLA.training.train_starvla][INFO] - Step 5910, Loss: {'train/loss': 0.19477927684783936, 'train/grad_norm_pre_clip': 10.708736419677734, 'timing/data': 0.0013676965609192848, 'timing/model': 1.4731401707977057, 'learning_rate/qwen_vl_interface': 1.0730415527352518e-06, 'learning_rate/action_model': 1.0730415527352517e-05, 'epoch': 2.27}) +[2026-06-18 06:47:50,730][starVLA.training.train_starvla][INFO] - Step 5911, Loss: {'train/loss': 0.34457945823669434, 'train/grad_norm_pre_clip': 19.527523040771484, 'timing/data': 0.001080760732293129, 'timing/model': 1.0837517324835062, 'learning_rate/qwen_vl_interface': 1.0720122030041638e-06, 'learning_rate/action_model': 1.0720122030041637e-05, 'epoch': 2.27}) +[2026-06-18 06:48:00,828][starVLA.training.train_starvla][INFO] - Step 5912, Loss: {'train/loss': 0.406374454498291, 'train/grad_norm_pre_clip': 12.346307754516602, 'timing/data': 0.0011682314798235893, 'timing/model': 1.779456534422934, 'learning_rate/qwen_vl_interface': 1.0709837193744412e-06, 'learning_rate/action_model': 1.070983719374441e-05, 'epoch': 2.27}) +[2026-06-18 06:48:06,407][starVLA.training.train_starvla][INFO] - Step 5913, Loss: {'train/loss': 0.6939845085144043, 'train/grad_norm_pre_clip': 22.17668342590332, 'timing/data': 0.0012375777587294579, 'timing/model': 2.137280808761716, 'learning_rate/qwen_vl_interface': 1.069956102059291e-06, 'learning_rate/action_model': 1.069956102059291e-05, 'epoch': 2.27}) +[2026-06-18 06:48:11,771][starVLA.training.train_starvla][INFO] - Step 5914, Loss: {'train/loss': 0.7097526788711548, 'train/grad_norm_pre_clip': 19.67527198791504, 'timing/data': 0.0009676180779933929, 'timing/model': 1.1164981042966247, 'learning_rate/qwen_vl_interface': 1.0689293512717387e-06, 'learning_rate/action_model': 1.0689293512717385e-05, 'epoch': 2.28}) +[2026-06-18 06:48:20,023][starVLA.training.train_starvla][INFO] - Step 5915, Loss: {'train/loss': 0.2312309443950653, 'train/grad_norm_pre_clip': 11.797690391540527, 'timing/data': 0.0011248188093304634, 'timing/model': 4.488049538806081, 'learning_rate/qwen_vl_interface': 1.0679034672246322e-06, 'learning_rate/action_model': 1.0679034672246322e-05, 'epoch': 2.28}) +[2026-06-18 06:48:31,091][starVLA.training.train_starvla][INFO] - Step 5916, Loss: {'train/loss': 0.3247635066509247, 'train/grad_norm_pre_clip': 16.998306274414062, 'timing/data': 0.0016171894967556, 'timing/model': 1.6117242043837905, 'learning_rate/qwen_vl_interface': 1.0668784501306348e-06, 'learning_rate/action_model': 1.0668784501306348e-05, 'epoch': 2.28}) +[2026-06-18 06:48:35,944][starVLA.training.train_starvla][INFO] - Step 5917, Loss: {'train/loss': 0.2296232432126999, 'train/grad_norm_pre_clip': 13.609427452087402, 'timing/data': 0.022122960537672043, 'timing/model': 1.525339706800878, 'learning_rate/qwen_vl_interface': 1.0658543002022377e-06, 'learning_rate/action_model': 1.0658543002022375e-05, 'epoch': 2.28}) +[2026-06-18 06:48:41,141][starVLA.training.train_starvla][INFO] - Step 5918, Loss: {'train/loss': 0.20640763640403748, 'train/grad_norm_pre_clip': 14.871909141540527, 'timing/data': 0.02110488060861826, 'timing/model': 1.3396781785413623, 'learning_rate/qwen_vl_interface': 1.0648310176517454e-06, 'learning_rate/action_model': 1.0648310176517453e-05, 'epoch': 2.28}) +[2026-06-18 06:48:46,699][starVLA.training.train_starvla][INFO] - Step 5919, Loss: {'train/loss': 0.48931005597114563, 'train/grad_norm_pre_clip': 20.87446403503418, 'timing/data': 0.032606239430606365, 'timing/model': 1.1612637909129262, 'learning_rate/qwen_vl_interface': 1.0638086026912873e-06, 'learning_rate/action_model': 1.0638086026912872e-05, 'epoch': 2.28}) +[2026-06-18 06:48:54,382][starVLA.training.train_starvla][INFO] - Step 5920, Loss: {'train/loss': 0.31406885385513306, 'train/grad_norm_pre_clip': 13.52514934539795, 'timing/data': 0.027757400646805763, 'timing/model': 2.0458121905103326, 'learning_rate/qwen_vl_interface': 1.0627870555328102e-06, 'learning_rate/action_model': 1.06278705553281e-05, 'epoch': 2.28}) +[2026-06-18 06:48:59,753][starVLA.training.train_starvla][INFO] - Step 5921, Loss: {'train/loss': 0.1463215947151184, 'train/grad_norm_pre_clip': 9.61145305633545, 'timing/data': 0.024106628261506557, 'timing/model': 1.201219329610467, 'learning_rate/qwen_vl_interface': 1.061766376388082e-06, 'learning_rate/action_model': 1.0617663763880821e-05, 'epoch': 2.28}) +[2026-06-18 06:49:04,999][starVLA.training.train_starvla][INFO] - Step 5922, Loss: {'train/loss': 0.2582590878009796, 'train/grad_norm_pre_clip': 13.130897521972656, 'timing/data': 0.02223511878401041, 'timing/model': 1.454835201613605, 'learning_rate/qwen_vl_interface': 1.0607465654686908e-06, 'learning_rate/action_model': 1.0607465654686907e-05, 'epoch': 2.28}) +[2026-06-18 06:49:10,257][starVLA.training.train_starvla][INFO] - Step 5923, Loss: {'train/loss': 0.49573200941085815, 'train/grad_norm_pre_clip': 17.32916259765625, 'timing/data': 0.021800782531499863, 'timing/model': 1.6612415965646505, 'learning_rate/qwen_vl_interface': 1.0597276229860434e-06, 'learning_rate/action_model': 1.0597276229860434e-05, 'epoch': 2.28}) +[2026-06-18 06:49:15,532][starVLA.training.train_starvla][INFO] - Step 5924, Loss: {'train/loss': 0.39203035831451416, 'train/grad_norm_pre_clip': 15.35400104522705, 'timing/data': 0.030082800425589085, 'timing/model': 1.1142401024699211, 'learning_rate/qwen_vl_interface': 1.0587095491513709e-06, 'learning_rate/action_model': 1.0587095491513709e-05, 'epoch': 2.28}) +[2026-06-18 06:49:27,077][starVLA.training.train_starvla][INFO] - Step 5925, Loss: {'train/loss': 0.2865590751171112, 'train/grad_norm_pre_clip': 12.211031913757324, 'timing/data': 0.02466694451868534, 'timing/model': 1.1610606806352735, 'learning_rate/qwen_vl_interface': 1.0576923441757174e-06, 'learning_rate/action_model': 1.0576923441757176e-05, 'epoch': 2.28}) +[2026-06-18 06:49:33,621][starVLA.training.train_starvla][INFO] - Step 5926, Loss: {'train/loss': 0.27451372146606445, 'train/grad_norm_pre_clip': 16.855688095092773, 'timing/data': 0.034600380808115005, 'timing/model': 2.154921667650342, 'learning_rate/qwen_vl_interface': 1.0566760082699515e-06, 'learning_rate/action_model': 1.0566760082699514e-05, 'epoch': 2.28}) +[2026-06-18 06:49:41,831][starVLA.training.train_starvla][INFO] - Step 5927, Loss: {'train/loss': 0.21900247037410736, 'train/grad_norm_pre_clip': 13.070938110351562, 'timing/data': 0.03118007630109787, 'timing/model': 4.764634232968092, 'learning_rate/qwen_vl_interface': 1.0556605416447626e-06, 'learning_rate/action_model': 1.0556605416447625e-05, 'epoch': 2.28}) +[2026-06-18 06:49:46,640][starVLA.training.train_starvla][INFO] - Step 5928, Loss: {'train/loss': 0.2297758311033249, 'train/grad_norm_pre_clip': 14.738567352294922, 'timing/data': 0.036989880725741386, 'timing/model': 1.2070120396092534, 'learning_rate/qwen_vl_interface': 1.054645944510656e-06, 'learning_rate/action_model': 1.054645944510656e-05, 'epoch': 2.28}) +[2026-06-18 06:49:54,379][starVLA.training.train_starvla][INFO] - Step 5929, Loss: {'train/loss': 0.20098857581615448, 'train/grad_norm_pre_clip': 14.63367748260498, 'timing/data': 0.021906550973653793, 'timing/model': 1.953798970207572, 'learning_rate/qwen_vl_interface': 1.0536322170779605e-06, 'learning_rate/action_model': 1.0536322170779606e-05, 'epoch': 2.28}) +[2026-06-18 06:49:59,891][starVLA.training.train_starvla][INFO] - Step 5930, Loss: {'train/loss': 0.32830578088760376, 'train/grad_norm_pre_clip': 14.270681381225586, 'timing/data': 0.023834045976400375, 'timing/model': 1.1456752149388194, 'learning_rate/qwen_vl_interface': 1.0526193595568203e-06, 'learning_rate/action_model': 1.0526193595568203e-05, 'epoch': 2.28}) +[2026-06-18 06:50:09,015][starVLA.training.train_starvla][INFO] - Step 5931, Loss: {'train/loss': 0.1998162567615509, 'train/grad_norm_pre_clip': 9.866964340209961, 'timing/data': 0.02612450160086155, 'timing/model': 1.7615999542176723, 'learning_rate/qwen_vl_interface': 1.0516073721572043e-06, 'learning_rate/action_model': 1.0516073721572044e-05, 'epoch': 2.28}) +[2026-06-18 06:50:17,147][starVLA.training.train_starvla][INFO] - Step 5932, Loss: {'train/loss': 0.1364126354455948, 'train/grad_norm_pre_clip': 8.366755485534668, 'timing/data': 0.021031538024544716, 'timing/model': 4.522047580219805, 'learning_rate/qwen_vl_interface': 1.0505962550888982e-06, 'learning_rate/action_model': 1.0505962550888982e-05, 'epoch': 2.28}) +[2026-06-18 06:50:22,522][starVLA.training.train_starvla][INFO] - Step 5933, Loss: {'train/loss': 0.3643067479133606, 'train/grad_norm_pre_clip': 16.29324722290039, 'timing/data': 0.027061409316956997, 'timing/model': 1.1910813003778458, 'learning_rate/qwen_vl_interface': 1.0495860085615065e-06, 'learning_rate/action_model': 1.0495860085615064e-05, 'epoch': 2.28}) +[2026-06-18 06:50:28,453][starVLA.training.train_starvla][INFO] - Step 5934, Loss: {'train/loss': 0.31555160880088806, 'train/grad_norm_pre_clip': 20.449344635009766, 'timing/data': 0.019998584873974323, 'timing/model': 1.652101369574666, 'learning_rate/qwen_vl_interface': 1.0485766327844558e-06, 'learning_rate/action_model': 1.0485766327844557e-05, 'epoch': 2.28}) +[2026-06-18 06:50:35,680][starVLA.training.train_starvla][INFO] - Step 5935, Loss: {'train/loss': 0.299714058637619, 'train/grad_norm_pre_clip': 12.034228324890137, 'timing/data': 0.03518662974238396, 'timing/model': 1.4902044450864196, 'learning_rate/qwen_vl_interface': 1.0475681279669892e-06, 'learning_rate/action_model': 1.0475681279669891e-05, 'epoch': 2.28}) +[2026-06-18 06:50:42,970][starVLA.training.train_starvla][INFO] - Step 5936, Loss: {'train/loss': 0.21973595023155212, 'train/grad_norm_pre_clip': 14.675017356872559, 'timing/data': 0.026124021038413048, 'timing/model': 1.7066618567332625, 'learning_rate/qwen_vl_interface': 1.0465604943181717e-06, 'learning_rate/action_model': 1.0465604943181717e-05, 'epoch': 2.28}) +[2026-06-18 06:50:54,811][starVLA.training.train_starvla][INFO] - Step 5937, Loss: {'train/loss': 0.4846841096878052, 'train/grad_norm_pre_clip': 17.21095085144043, 'timing/data': 0.0391522366553545, 'timing/model': 1.6285936227068305, 'learning_rate/qwen_vl_interface': 1.0455537320468861e-06, 'learning_rate/action_model': 1.045553732046886e-05, 'epoch': 2.28}) +[2026-06-18 06:50:59,773][starVLA.training.train_starvla][INFO] - Step 5938, Loss: {'train/loss': 0.23648223280906677, 'train/grad_norm_pre_clip': 16.60157012939453, 'timing/data': 0.022634951397776604, 'timing/model': 1.5800787406042218, 'learning_rate/qwen_vl_interface': 1.0445478413618382e-06, 'learning_rate/action_model': 1.0445478413618382e-05, 'epoch': 2.28}) +[2026-06-18 06:51:05,254][starVLA.training.train_starvla][INFO] - Step 5939, Loss: {'train/loss': 0.1598028689622879, 'train/grad_norm_pre_clip': 9.399552345275879, 'timing/data': 0.030275209806859493, 'timing/model': 1.6710184179246426, 'learning_rate/qwen_vl_interface': 1.0435428224715472e-06, 'learning_rate/action_model': 1.0435428224715471e-05, 'epoch': 2.28}) +[2026-06-18 06:51:12,021][starVLA.training.train_starvla][INFO] - Step 5940, Loss: {'train/loss': 0.4232681095600128, 'train/grad_norm_pre_clip': 15.22149658203125, 'timing/data': 0.018920247443020344, 'timing/model': 1.1094050640240312, 'learning_rate/qwen_vl_interface': 1.0425386755843547e-06, 'learning_rate/action_model': 1.0425386755843547e-05, 'epoch': 2.29}) +[2026-06-18 06:51:17,290][starVLA.training.train_starvla][INFO] - Step 5941, Loss: {'train/loss': 0.5768115520477295, 'train/grad_norm_pre_clip': 24.07850456237793, 'timing/data': 0.028761236928403378, 'timing/model': 1.1720227589830756, 'learning_rate/qwen_vl_interface': 1.0415354009084219e-06, 'learning_rate/action_model': 1.041535400908422e-05, 'epoch': 2.29}) +[2026-06-18 06:51:23,126][starVLA.training.train_starvla][INFO] - Step 5942, Loss: {'train/loss': 0.4037332534790039, 'train/grad_norm_pre_clip': 18.47598648071289, 'timing/data': 0.03025718219578266, 'timing/model': 1.5845369659364223, 'learning_rate/qwen_vl_interface': 1.0405329986517301e-06, 'learning_rate/action_model': 1.0405329986517301e-05, 'epoch': 2.29}) +[2026-06-18 06:51:27,789][starVLA.training.train_starvla][INFO] - Step 5943, Loss: {'train/loss': 0.4948592483997345, 'train/grad_norm_pre_clip': 22.610618591308594, 'timing/data': 0.024088469333946705, 'timing/model': 1.1446613939478993, 'learning_rate/qwen_vl_interface': 1.0395314690220776e-06, 'learning_rate/action_model': 1.0395314690220776e-05, 'epoch': 2.29}) +[2026-06-18 06:51:36,036][starVLA.training.train_starvla][INFO] - Step 5944, Loss: {'train/loss': 0.2750512361526489, 'train/grad_norm_pre_clip': 13.294526100158691, 'timing/data': 0.0206765066832304, 'timing/model': 1.1301372544839978, 'learning_rate/qwen_vl_interface': 1.0385308122270815e-06, 'learning_rate/action_model': 1.0385308122270814e-05, 'epoch': 2.29}) +[2026-06-18 06:51:41,074][starVLA.training.train_starvla][INFO] - Step 5945, Loss: {'train/loss': 0.1439831256866455, 'train/grad_norm_pre_clip': 10.775322914123535, 'timing/data': 0.02341260015964508, 'timing/model': 1.12337932176888, 'learning_rate/qwen_vl_interface': 1.0375310284741787e-06, 'learning_rate/action_model': 1.0375310284741788e-05, 'epoch': 2.29}) +[2026-06-18 06:51:46,548][starVLA.training.train_starvla][INFO] - Step 5946, Loss: {'train/loss': 0.3413373529911041, 'train/grad_norm_pre_clip': 14.72067642211914, 'timing/data': 0.023698698729276657, 'timing/model': 1.8604403110221028, 'learning_rate/qwen_vl_interface': 1.0365321179706276e-06, 'learning_rate/action_model': 1.0365321179706276e-05, 'epoch': 2.29}) +[2026-06-18 06:51:55,609][starVLA.training.train_starvla][INFO] - Step 5947, Loss: {'train/loss': 0.1720445156097412, 'train/grad_norm_pre_clip': 12.502028465270996, 'timing/data': 0.03133912570774555, 'timing/model': 2.0247514182701707, 'learning_rate/qwen_vl_interface': 1.035534080923503e-06, 'learning_rate/action_model': 1.0355340809235029e-05, 'epoch': 2.29}) +[2026-06-18 06:52:02,696][starVLA.training.train_starvla][INFO] - Step 5948, Loss: {'train/loss': 0.1825745701789856, 'train/grad_norm_pre_clip': 12.459590911865234, 'timing/data': 0.02525464352220297, 'timing/model': 2.9865242885425687, 'learning_rate/qwen_vl_interface': 1.0345369175396972e-06, 'learning_rate/action_model': 1.0345369175396972e-05, 'epoch': 2.29}) +[2026-06-18 06:52:07,892][starVLA.training.train_starvla][INFO] - Step 5949, Loss: {'train/loss': 0.5179320573806763, 'train/grad_norm_pre_clip': 20.550010681152344, 'timing/data': 0.026693206280469894, 'timing/model': 1.2844142718240619, 'learning_rate/qwen_vl_interface': 1.0335406280259255e-06, 'learning_rate/action_model': 1.0335406280259254e-05, 'epoch': 2.29}) +[2026-06-18 06:52:13,612][starVLA.training.train_starvla][INFO] - Step 5950, Loss: {'train/loss': 0.1476043164730072, 'train/grad_norm_pre_clip': 12.539085388183594, 'timing/data': 0.02261463087052107, 'timing/model': 1.4068989539518952, 'learning_rate/qwen_vl_interface': 1.0325452125887177e-06, 'learning_rate/action_model': 1.0325452125887178e-05, 'epoch': 2.29}) +[2026-06-18 06:52:25,719][starVLA.training.train_starvla][INFO] - Step 5951, Loss: {'train/loss': 0.4151821732521057, 'train/grad_norm_pre_clip': 20.68074607849121, 'timing/data': 0.020636217668652534, 'timing/model': 1.2168503822758794, 'learning_rate/qwen_vl_interface': 1.031550671434426e-06, 'learning_rate/action_model': 1.031550671434426e-05, 'epoch': 2.29}) +[2026-06-18 06:52:35,961][starVLA.training.train_starvla][INFO] - Step 5952, Loss: {'train/loss': 0.36827710270881653, 'train/grad_norm_pre_clip': 19.025415420532227, 'timing/data': 0.019992276094853878, 'timing/model': 1.5763931311666965, 'learning_rate/qwen_vl_interface': 1.0305570047692197e-06, 'learning_rate/action_model': 1.0305570047692195e-05, 'epoch': 2.29}) +[2026-06-18 06:52:40,907][starVLA.training.train_starvla][INFO] - Step 5953, Loss: {'train/loss': 0.4578374922275543, 'train/grad_norm_pre_clip': 17.215253829956055, 'timing/data': 0.02136258129030466, 'timing/model': 1.4391969162970781, 'learning_rate/qwen_vl_interface': 1.0295642127990863e-06, 'learning_rate/action_model': 1.0295642127990863e-05, 'epoch': 2.29}) +[2026-06-18 06:52:46,884][starVLA.training.train_starvla][INFO] - Step 5954, Loss: {'train/loss': 0.4913521409034729, 'train/grad_norm_pre_clip': 21.365074157714844, 'timing/data': 0.040524243377149105, 'timing/model': 2.3499743044376373, 'learning_rate/qwen_vl_interface': 1.0285722957298326e-06, 'learning_rate/action_model': 1.0285722957298326e-05, 'epoch': 2.29}) +[2026-06-18 06:52:54,526][starVLA.training.train_starvla][INFO] - Step 5955, Loss: {'train/loss': 0.7328974008560181, 'train/grad_norm_pre_clip': 23.494430541992188, 'timing/data': 0.02368517965078354, 'timing/model': 2.9299856405705214, 'learning_rate/qwen_vl_interface': 1.027581253767085e-06, 'learning_rate/action_model': 1.027581253767085e-05, 'epoch': 2.29}) +[2026-06-18 06:53:04,805][starVLA.training.train_starvla][INFO] - Step 5956, Loss: {'train/loss': 0.1555091291666031, 'train/grad_norm_pre_clip': 11.783068656921387, 'timing/data': 0.025420963764190674, 'timing/model': 3.1558771524578333, 'learning_rate/qwen_vl_interface': 1.0265910871162858e-06, 'learning_rate/action_model': 1.0265910871162858e-05, 'epoch': 2.29}) +[2026-06-18 06:53:21,456][starVLA.training.train_starvla][INFO] - Step 5957, Loss: {'train/loss': 0.37992677092552185, 'train/grad_norm_pre_clip': 21.86053466796875, 'timing/data': 0.021929203532636166, 'timing/model': 6.27171182166785, 'learning_rate/qwen_vl_interface': 1.0256017959827008e-06, 'learning_rate/action_model': 1.0256017959827008e-05, 'epoch': 2.29}) +[2026-06-18 06:53:35,771][starVLA.training.train_starvla][INFO] - Step 5958, Loss: {'train/loss': 0.17522038519382477, 'train/grad_norm_pre_clip': 13.160937309265137, 'timing/data': 0.024906405247747898, 'timing/model': 2.8607820458710194, 'learning_rate/qwen_vl_interface': 1.0246133805714084e-06, 'learning_rate/action_model': 1.0246133805714084e-05, 'epoch': 2.29}) +[2026-06-18 06:53:44,914][starVLA.training.train_starvla][INFO] - Step 5959, Loss: {'train/loss': 0.21011468768119812, 'train/grad_norm_pre_clip': 13.742203712463379, 'timing/data': 0.02508462965488434, 'timing/model': 4.7123167319223285, 'learning_rate/qwen_vl_interface': 1.0236258410873076e-06, 'learning_rate/action_model': 1.0236258410873076e-05, 'epoch': 2.29}) +[2026-06-18 06:53:59,170][starVLA.training.train_starvla][INFO] - Step 5960, Loss: {'train/loss': 0.17391915619373322, 'train/grad_norm_pre_clip': 8.57949161529541, 'timing/data': 0.021952851675450802, 'timing/model': 2.0860306080430746, 'learning_rate/qwen_vl_interface': 1.0226391777351188e-06, 'learning_rate/action_model': 1.0226391777351189e-05, 'epoch': 2.29}) +[2026-06-18 06:54:10,764][starVLA.training.train_starvla][INFO] - Step 5961, Loss: {'train/loss': 0.21066176891326904, 'train/grad_norm_pre_clip': 11.330979347229004, 'timing/data': 0.021897297352552414, 'timing/model': 4.605486079119146, 'learning_rate/qwen_vl_interface': 1.0216533907193777e-06, 'learning_rate/action_model': 1.0216533907193776e-05, 'epoch': 2.29}) +[2026-06-18 06:54:27,675][starVLA.training.train_starvla][INFO] - Step 5962, Loss: {'train/loss': 0.15817314386367798, 'train/grad_norm_pre_clip': 11.811640739440918, 'timing/data': 0.03196010831743479, 'timing/model': 4.657167511060834, 'learning_rate/qwen_vl_interface': 1.020668480244438e-06, 'learning_rate/action_model': 1.020668480244438e-05, 'epoch': 2.29}) +[2026-06-18 06:54:32,845][starVLA.training.train_starvla][INFO] - Step 5963, Loss: {'train/loss': 0.2138836681842804, 'train/grad_norm_pre_clip': 12.194110870361328, 'timing/data': 0.029824371449649334, 'timing/model': 1.1099350592121482, 'learning_rate/qwen_vl_interface': 1.0196844465144735e-06, 'learning_rate/action_model': 1.0196844465144734e-05, 'epoch': 2.29}) +[2026-06-18 06:54:38,408][starVLA.training.train_starvla][INFO] - Step 5964, Loss: {'train/loss': 0.3349250853061676, 'train/grad_norm_pre_clip': 16.792165756225586, 'timing/data': 0.024521146900951862, 'timing/model': 1.5424679396674037, 'learning_rate/qwen_vl_interface': 1.0187012897334755e-06, 'learning_rate/action_model': 1.0187012897334755e-05, 'epoch': 2.29}) +[2026-06-18 06:54:43,348][starVLA.training.train_starvla][INFO] - Step 5965, Loss: {'train/loss': 0.4042294919490814, 'train/grad_norm_pre_clip': 21.526031494140625, 'timing/data': 0.02632990013808012, 'timing/model': 1.1240205885842443, 'learning_rate/qwen_vl_interface': 1.017719010105253e-06, 'learning_rate/action_model': 1.017719010105253e-05, 'epoch': 2.29}) +[2026-06-18 06:54:48,540][starVLA.training.train_starvla][INFO] - Step 5966, Loss: {'train/loss': 0.17232440412044525, 'train/grad_norm_pre_clip': 8.46333122253418, 'timing/data': 0.03318061027675867, 'timing/model': 1.5114312330260873, 'learning_rate/qwen_vl_interface': 1.0167376078334337e-06, 'learning_rate/action_model': 1.0167376078334337e-05, 'epoch': 2.3}) +[2026-06-18 06:54:53,908][starVLA.training.train_starvla][INFO] - Step 5967, Loss: {'train/loss': 0.1175321713089943, 'train/grad_norm_pre_clip': 10.413867950439453, 'timing/data': 0.02243446744978428, 'timing/model': 1.814626956358552, 'learning_rate/qwen_vl_interface': 1.0157570831214634e-06, 'learning_rate/action_model': 1.0157570831214634e-05, 'epoch': 2.3}) +[2026-06-18 06:54:59,274][starVLA.training.train_starvla][INFO] - Step 5968, Loss: {'train/loss': 0.24551880359649658, 'train/grad_norm_pre_clip': 12.029311180114746, 'timing/data': 0.02599463425576687, 'timing/model': 1.1542473640292883, 'learning_rate/qwen_vl_interface': 1.0147774361726066e-06, 'learning_rate/action_model': 1.0147774361726064e-05, 'epoch': 2.3}) +[2026-06-18 06:55:04,802][starVLA.training.train_starvla][INFO] - Step 5969, Loss: {'train/loss': 0.17384806275367737, 'train/grad_norm_pre_clip': 10.38991928100586, 'timing/data': 0.023795832879841328, 'timing/model': 1.401405368000269, 'learning_rate/qwen_vl_interface': 1.013798667189944e-06, 'learning_rate/action_model': 1.013798667189944e-05, 'epoch': 2.3}) +[2026-06-18 06:55:10,517][starVLA.training.train_starvla][INFO] - Step 5970, Loss: {'train/loss': 0.08114038407802582, 'train/grad_norm_pre_clip': 4.697902202606201, 'timing/data': 0.03155034687370062, 'timing/model': 1.1666717417538166, 'learning_rate/qwen_vl_interface': 1.0128207763763755e-06, 'learning_rate/action_model': 1.0128207763763755e-05, 'epoch': 2.3}) +[2026-06-18 06:55:15,511][starVLA.training.train_starvla][INFO] - Step 5971, Loss: {'train/loss': 0.7554753422737122, 'train/grad_norm_pre_clip': 21.577465057373047, 'timing/data': 0.021502568386495113, 'timing/model': 1.1443578405305743, 'learning_rate/qwen_vl_interface': 1.0118437639346204e-06, 'learning_rate/action_model': 1.0118437639346204e-05, 'epoch': 2.3}) +[2026-06-18 06:55:20,640][starVLA.training.train_starvla][INFO] - Step 5972, Loss: {'train/loss': 0.26361483335494995, 'train/grad_norm_pre_clip': 12.017980575561523, 'timing/data': 0.0316344341263175, 'timing/model': 1.5863531678915024, 'learning_rate/qwen_vl_interface': 1.0108676300672144e-06, 'learning_rate/action_model': 1.0108676300672144e-05, 'epoch': 2.3}) +[2026-06-18 06:55:25,525][starVLA.training.train_starvla][INFO] - Step 5973, Loss: {'train/loss': 0.2719462215900421, 'train/grad_norm_pre_clip': 16.530710220336914, 'timing/data': 0.047302721068263054, 'timing/model': 1.1081265937536955, 'learning_rate/qwen_vl_interface': 1.0098923749765086e-06, 'learning_rate/action_model': 1.0098923749765086e-05, 'epoch': 2.3}) +[2026-06-18 06:55:30,351][starVLA.training.train_starvla][INFO] - Step 5974, Loss: {'train/loss': 0.38011929392814636, 'train/grad_norm_pre_clip': 23.00301170349121, 'timing/data': 0.020813637413084507, 'timing/model': 1.1047363467514515, 'learning_rate/qwen_vl_interface': 1.008917998864676e-06, 'learning_rate/action_model': 1.008917998864676e-05, 'epoch': 2.3}) +[2026-06-18 06:55:35,808][starVLA.training.train_starvla][INFO] - Step 5975, Loss: {'train/loss': 0.43313872814178467, 'train/grad_norm_pre_clip': 25.903242111206055, 'timing/data': 0.020839018747210503, 'timing/model': 1.3357028188183904, 'learning_rate/qwen_vl_interface': 1.0079445019337063e-06, 'learning_rate/action_model': 1.0079445019337063e-05, 'epoch': 2.3}) +[2026-06-18 06:55:41,296][starVLA.training.train_starvla][INFO] - Step 5976, Loss: {'train/loss': 0.368629515171051, 'train/grad_norm_pre_clip': 22.652206420898438, 'timing/data': 0.0215328773483634, 'timing/model': 1.73391152638942, 'learning_rate/qwen_vl_interface': 1.0069718843854058e-06, 'learning_rate/action_model': 1.0069718843854058e-05, 'epoch': 2.3}) +[2026-06-18 06:55:47,137][starVLA.training.train_starvla][INFO] - Step 5977, Loss: {'train/loss': 0.5580975413322449, 'train/grad_norm_pre_clip': 18.280601501464844, 'timing/data': 0.030941154807806015, 'timing/model': 2.09392561390996, 'learning_rate/qwen_vl_interface': 1.0060001464214007e-06, 'learning_rate/action_model': 1.0060001464214006e-05, 'epoch': 2.3}) +[2026-06-18 06:55:58,959][starVLA.training.train_starvla][INFO] - Step 5978, Loss: {'train/loss': 0.1983027458190918, 'train/grad_norm_pre_clip': 10.00651741027832, 'timing/data': 0.07365962211042643, 'timing/model': 1.335779364220798, 'learning_rate/qwen_vl_interface': 1.00502928824313e-06, 'learning_rate/action_model': 1.00502928824313e-05, 'epoch': 2.3}) +[2026-06-18 06:56:12,948][starVLA.training.train_starvla][INFO] - Step 5979, Loss: {'train/loss': 0.5904730558395386, 'train/grad_norm_pre_clip': 25.10736846923828, 'timing/data': 0.03061378188431263, 'timing/model': 1.3243004875257611, 'learning_rate/qwen_vl_interface': 1.0040593100518567e-06, 'learning_rate/action_model': 1.0040593100518566e-05, 'epoch': 2.3}) +[2026-06-18 06:56:19,756][starVLA.training.train_starvla][INFO] - Step 5980, Loss: {'train/loss': 0.30949658155441284, 'train/grad_norm_pre_clip': 19.42926025390625, 'timing/data': 0.023915673606097698, 'timing/model': 1.1137214647606015, 'learning_rate/qwen_vl_interface': 1.003090212048657e-06, 'learning_rate/action_model': 1.003090212048657e-05, 'epoch': 2.3}) +[2026-06-18 06:56:29,151][starVLA.training.train_starvla][INFO] - Step 5981, Loss: {'train/loss': 0.33411821722984314, 'train/grad_norm_pre_clip': 19.704999923706055, 'timing/data': 0.02459865529090166, 'timing/model': 1.2657388243824244, 'learning_rate/qwen_vl_interface': 1.0021219944344263e-06, 'learning_rate/action_model': 1.0021219944344262e-05, 'epoch': 2.3}) +[2026-06-18 06:56:38,447][starVLA.training.train_starvla][INFO] - Step 5982, Loss: {'train/loss': 0.34447380900382996, 'train/grad_norm_pre_clip': 15.198698043823242, 'timing/data': 0.02637386415153742, 'timing/model': 4.096496642567217, 'learning_rate/qwen_vl_interface': 1.0011546574098773e-06, 'learning_rate/action_model': 1.0011546574098772e-05, 'epoch': 2.3}) +[2026-06-18 06:56:48,081][starVLA.training.train_starvla][INFO] - Step 5983, Loss: {'train/loss': 0.19971273839473724, 'train/grad_norm_pre_clip': 11.314508438110352, 'timing/data': 0.025941187515854836, 'timing/model': 1.52252955455333, 'learning_rate/qwen_vl_interface': 1.0001882011755396e-06, 'learning_rate/action_model': 1.0001882011755395e-05, 'epoch': 2.3}) +[2026-06-18 06:57:00,209][starVLA.training.train_starvla][INFO] - Step 5984, Loss: {'train/loss': 0.4149326980113983, 'train/grad_norm_pre_clip': 17.01693344116211, 'timing/data': 0.07218918576836586, 'timing/model': 3.187236883677542, 'learning_rate/qwen_vl_interface': 9.992226259317605e-07, 'learning_rate/action_model': 9.992226259317605e-06, 'epoch': 2.3}) +[2026-06-18 06:57:06,816][starVLA.training.train_starvla][INFO] - Step 5985, Loss: {'train/loss': 0.46892717480659485, 'train/grad_norm_pre_clip': 23.831424713134766, 'timing/data': 0.02537278551608324, 'timing/model': 2.578796546906233, 'learning_rate/qwen_vl_interface': 9.982579318787044e-07, 'learning_rate/action_model': 9.982579318787044e-06, 'epoch': 2.3}) +[2026-06-18 06:57:23,466][starVLA.training.train_starvla][INFO] - Step 5986, Loss: {'train/loss': 0.26019570231437683, 'train/grad_norm_pre_clip': 16.57481575012207, 'timing/data': 0.057385693304240704, 'timing/model': 3.7513176826760173, 'learning_rate/qwen_vl_interface': 9.97294119216356e-07, 'learning_rate/action_model': 9.97294119216356e-06, 'epoch': 2.3}) +[2026-06-18 06:57:30,814][starVLA.training.train_starvla][INFO] - Step 5987, Loss: {'train/loss': 0.38637521862983704, 'train/grad_norm_pre_clip': 26.15757179260254, 'timing/data': 0.023084482178092003, 'timing/model': 1.7373308958485723, 'learning_rate/qwen_vl_interface': 9.96331188144512e-07, 'learning_rate/action_model': 9.96331188144512e-06, 'epoch': 2.3}) +[2026-06-18 06:57:35,755][starVLA.training.train_starvla][INFO] - Step 5988, Loss: {'train/loss': 0.35790765285491943, 'train/grad_norm_pre_clip': 14.015645027160645, 'timing/data': 0.024115655571222305, 'timing/model': 1.2841683942824602, 'learning_rate/qwen_vl_interface': 9.953691388627887e-07, 'learning_rate/action_model': 9.953691388627886e-06, 'epoch': 2.3}) +[2026-06-18 06:57:40,863][starVLA.training.train_starvla][INFO] - Step 5989, Loss: {'train/loss': 0.29968422651290894, 'train/grad_norm_pre_clip': 11.982867240905762, 'timing/data': 0.02068305481225252, 'timing/model': 1.1015764214098454, 'learning_rate/qwen_vl_interface': 9.94407971570622e-07, 'learning_rate/action_model': 9.94407971570622e-06, 'epoch': 2.3}) +[2026-06-18 06:57:45,788][starVLA.training.train_starvla][INFO] - Step 5990, Loss: {'train/loss': 0.3404615521430969, 'train/grad_norm_pre_clip': 14.690573692321777, 'timing/data': 0.03270462341606617, 'timing/model': 1.2980276523157954, 'learning_rate/qwen_vl_interface': 9.934476864672618e-07, 'learning_rate/action_model': 9.934476864672617e-06, 'epoch': 2.3}) +[2026-06-18 06:57:50,932][starVLA.training.train_starvla][INFO] - Step 5991, Loss: {'train/loss': 0.38287338614463806, 'train/grad_norm_pre_clip': 16.13067626953125, 'timing/data': 0.021650055423378944, 'timing/model': 1.3367035379633307, 'learning_rate/qwen_vl_interface': 9.924882837517781e-07, 'learning_rate/action_model': 9.924882837517782e-06, 'epoch': 2.3}) +[2026-06-18 06:57:55,806][starVLA.training.train_starvla][INFO] - Step 5992, Loss: {'train/loss': 0.09118686616420746, 'train/grad_norm_pre_clip': 6.396468639373779, 'timing/data': 0.025841706432402134, 'timing/model': 1.6349483104422688, 'learning_rate/qwen_vl_interface': 9.91529763623052e-07, 'learning_rate/action_model': 9.91529763623052e-06, 'epoch': 2.31}) +[2026-06-18 06:58:00,639][starVLA.training.train_starvla][INFO] - Step 5993, Loss: {'train/loss': 0.2854044735431671, 'train/grad_norm_pre_clip': 14.16805362701416, 'timing/data': 0.019951866008341312, 'timing/model': 1.0842783283442259, 'learning_rate/qwen_vl_interface': 9.905721262797894e-07, 'learning_rate/action_model': 9.905721262797892e-06, 'epoch': 2.31}) +[2026-06-18 06:58:12,336][starVLA.training.train_starvla][INFO] - Step 5994, Loss: {'train/loss': 0.3951927721500397, 'train/grad_norm_pre_clip': 17.923185348510742, 'timing/data': 0.02200052421540022, 'timing/model': 1.8469544760882854, 'learning_rate/qwen_vl_interface': 9.896153719205082e-07, 'learning_rate/action_model': 9.89615371920508e-06, 'epoch': 2.31}) +[2026-06-18 06:58:17,402][starVLA.training.train_starvla][INFO] - Step 5995, Loss: {'train/loss': 0.3146325647830963, 'train/grad_norm_pre_clip': 13.454387664794922, 'timing/data': 0.020093095488846302, 'timing/model': 1.1696928823366761, 'learning_rate/qwen_vl_interface': 9.886595007435448e-07, 'learning_rate/action_model': 9.886595007435447e-06, 'epoch': 2.31}) +[2026-06-18 06:58:22,842][starVLA.training.train_starvla][INFO] - Step 5996, Loss: {'train/loss': 0.10998664796352386, 'train/grad_norm_pre_clip': 16.516008377075195, 'timing/data': 0.0330641558393836, 'timing/model': 1.1034700786694884, 'learning_rate/qwen_vl_interface': 9.877045129470524e-07, 'learning_rate/action_model': 9.877045129470524e-06, 'epoch': 2.31}) +[2026-06-18 06:58:28,642][starVLA.training.train_starvla][INFO] - Step 5997, Loss: {'train/loss': 0.1517968326807022, 'train/grad_norm_pre_clip': 8.761019706726074, 'timing/data': 0.024203340522944927, 'timing/model': 1.8678063713014126, 'learning_rate/qwen_vl_interface': 9.86750408729001e-07, 'learning_rate/action_model': 9.867504087290009e-06, 'epoch': 2.31}) +[2026-06-18 06:58:33,116][starVLA.training.train_starvla][INFO] - Step 5998, Loss: {'train/loss': 0.21850398182868958, 'train/grad_norm_pre_clip': 25.303546905517578, 'timing/data': 0.02068364806473255, 'timing/model': 1.2386995945125818, 'learning_rate/qwen_vl_interface': 9.857971882871773e-07, 'learning_rate/action_model': 9.857971882871772e-06, 'epoch': 2.31}) +[2026-06-18 06:58:42,169][starVLA.training.train_starvla][INFO] - Step 5999, Loss: {'train/loss': 0.31050169467926025, 'train/grad_norm_pre_clip': 20.13768196105957, 'timing/data': 0.02959385607391596, 'timing/model': 1.2799019198864698, 'learning_rate/qwen_vl_interface': 9.848448518191838e-07, 'learning_rate/action_model': 9.848448518191836e-06, 'epoch': 2.31}) +[2026-06-18 07:02:52,396][starVLA.training.train_starvla][INFO] - Step 6000, Loss: {'train/loss': 0.12056228518486023, 'train/grad_norm_pre_clip': 8.342884063720703, 'eval/loss': 0.4206308424472809, 'eval/latency_4/loss': 0.4206308424472809, 'eval/demon_fire_precision': 0.9641383687718185, 'eval/demon_fire_recall': 0.9866839883078922, 'eval/demon_fire_f1': 0.9752808988764045, 'eval/demon_fire_rate_model': 0.9254038179148312, 'eval/demon_fire_rate_teacher': 0.9042584434654919, 'eval/demon_left_precision': 0.8106666666666666, 'eval/demon_left_recall': 0.6940639269406392, 'eval/demon_left_f1': 0.7478474784747847, 'eval/demon_left_rate_model': 0.11013215859030837, 'eval/demon_left_rate_teacher': 0.12863436123348018, 'eval/demon_right_precision': 0.801094890510949, 'eval/demon_right_recall': 0.7924187725631769, 'eval/demon_right_f1': 0.7967332123411978, 'eval/demon_right_rate_model': 0.16093979441997064, 'eval/demon_right_rate_teacher': 0.16270190895741557, 'eval/demon_move_f1': 0.7722903454079912, 'eval/demon_cc_f1': 0.9039982139838119, 'eval/latency_4/demon_fire_precision': 0.9641383687718185, 'eval/latency_4/demon_fire_recall': 0.9866839883078922, 'eval/latency_4/demon_fire_f1': 0.9752808988764045, 'eval/latency_4/demon_fire_rate_model': 0.9254038179148312, 'eval/latency_4/demon_fire_rate_teacher': 0.9042584434654919, 'eval/latency_4/demon_left_precision': 0.8106666666666666, 'eval/latency_4/demon_left_recall': 0.6940639269406392, 'eval/latency_4/demon_left_f1': 0.7478474784747847, 'eval/latency_4/demon_left_rate_model': 0.11013215859030837, 'eval/latency_4/demon_left_rate_teacher': 0.12863436123348018, 'eval/latency_4/demon_right_precision': 0.801094890510949, 'eval/latency_4/demon_right_recall': 0.7924187725631769, 'eval/latency_4/demon_right_f1': 0.7967332123411978, 'eval/latency_4/demon_right_rate_model': 0.16093979441997064, 'eval/latency_4/demon_right_rate_teacher': 0.16270190895741557, 'eval/latency_4/demon_move_f1': 0.7722903454079912, 'eval/latency_4/demon_cc_f1': 0.9039982139838119, 'timing/data': 0.021344168111681938, 'timing/model': 1.2606349615380168, 'learning_rate/qwen_vl_interface': 9.838933995224439e-07, 'learning_rate/action_model': 9.83893399522444e-06, 'epoch': 2.31}) +[2026-06-18 07:02:52,418][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_6000_state +[2026-06-18 07:02:52,427][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 07:03:01,339][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_6000_state/model.safetensors +[2026-06-18 07:03:24,532][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_6000_state/optimizer.bin +[2026-06-18 07:03:24,533][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_6000_state/sampler.bin +[2026-06-18 07:03:24,533][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_6000_state/sampler_1.bin +[2026-06-18 07:03:24,538][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_6000_state/random_states_0.pkl +[2026-06-18 07:03:24,543][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 6000 +[2026-06-18 07:03:24,543][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 07:03:24,560][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 07:03:30,357][starVLA.training.train_starvla][INFO] - Step 6001, Loss: {'train/loss': 0.3999222218990326, 'train/grad_norm_pre_clip': 22.45745849609375, 'timing/data': 0.02612851094454527, 'timing/model': 1.6397046772763133, 'learning_rate/qwen_vl_interface': 9.829428315941921e-07, 'learning_rate/action_model': 9.829428315941922e-06, 'epoch': 2.31}) +[2026-06-18 07:03:35,642][starVLA.training.train_starvla][INFO] - Step 6002, Loss: {'train/loss': 0.4953082501888275, 'train/grad_norm_pre_clip': 52.58074188232422, 'timing/data': 0.019945775158703327, 'timing/model': 1.4741174820810556, 'learning_rate/qwen_vl_interface': 9.819931482314823e-07, 'learning_rate/action_model': 9.819931482314823e-06, 'epoch': 2.31}) +[2026-06-18 07:03:43,197][starVLA.training.train_starvla][INFO] - Step 6003, Loss: {'train/loss': 0.24474182724952698, 'train/grad_norm_pre_clip': 10.362163543701172, 'timing/data': 0.02034528460353613, 'timing/model': 1.1618705671280622, 'learning_rate/qwen_vl_interface': 9.81044349631185e-07, 'learning_rate/action_model': 9.81044349631185e-06, 'epoch': 2.31}) +[2026-06-18 07:03:50,514][starVLA.training.train_starvla][INFO] - Step 6004, Loss: {'train/loss': 0.46079742908477783, 'train/grad_norm_pre_clip': 17.435537338256836, 'timing/data': 0.028983334079384804, 'timing/model': 1.7255708761513233, 'learning_rate/qwen_vl_interface': 9.800964359899886e-07, 'learning_rate/action_model': 9.800964359899886e-06, 'epoch': 2.31}) +[2026-06-18 07:03:58,638][starVLA.training.train_starvla][INFO] - Step 6005, Loss: {'train/loss': 0.36539143323898315, 'train/grad_norm_pre_clip': 18.94460105895996, 'timing/data': 0.023113712668418884, 'timing/model': 4.219369436614215, 'learning_rate/qwen_vl_interface': 9.791494075043963e-07, 'learning_rate/action_model': 9.791494075043963e-06, 'epoch': 2.31}) +[2026-06-18 07:04:07,296][starVLA.training.train_starvla][INFO] - Step 6006, Loss: {'train/loss': 0.5727148652076721, 'train/grad_norm_pre_clip': 18.92223358154297, 'timing/data': 0.02099546417593956, 'timing/model': 1.4056508736684918, 'learning_rate/qwen_vl_interface': 9.78203264370727e-07, 'learning_rate/action_model': 9.78203264370727e-06, 'epoch': 2.31}) +[2026-06-18 07:04:12,339][starVLA.training.train_starvla][INFO] - Step 6007, Loss: {'train/loss': 0.2380310744047165, 'train/grad_norm_pre_clip': 14.053979873657227, 'timing/data': 0.028483832255005836, 'timing/model': 1.6180559834465384, 'learning_rate/qwen_vl_interface': 9.772580067851162e-07, 'learning_rate/action_model': 9.772580067851161e-06, 'epoch': 2.31}) +[2026-06-18 07:04:18,642][starVLA.training.train_starvla][INFO] - Step 6008, Loss: {'train/loss': 0.4181365966796875, 'train/grad_norm_pre_clip': 15.437143325805664, 'timing/data': 0.023486635647714138, 'timing/model': 2.373056437820196, 'learning_rate/qwen_vl_interface': 9.763136349435195e-07, 'learning_rate/action_model': 9.763136349435194e-06, 'epoch': 2.31}) +[2026-06-18 07:04:28,852][starVLA.training.train_starvla][INFO] - Step 6009, Loss: {'train/loss': 0.34639284014701843, 'train/grad_norm_pre_clip': 17.62639617919922, 'timing/data': 0.023470056243240833, 'timing/model': 4.754680928774178, 'learning_rate/qwen_vl_interface': 9.753701490417048e-07, 'learning_rate/action_model': 9.753701490417047e-06, 'epoch': 2.31}) +[2026-06-18 07:04:39,959][starVLA.training.train_starvla][INFO] - Step 6010, Loss: {'train/loss': 0.1324232965707779, 'train/grad_norm_pre_clip': 10.682044982910156, 'timing/data': 0.022470824420452118, 'timing/model': 1.2343041021376848, 'learning_rate/qwen_vl_interface': 9.744275492752584e-07, 'learning_rate/action_model': 9.744275492752583e-06, 'epoch': 2.31}) +[2026-06-18 07:04:45,225][starVLA.training.train_starvla][INFO] - Step 6011, Loss: {'train/loss': 0.36701226234436035, 'train/grad_norm_pre_clip': 18.39542007446289, 'timing/data': 0.025263182818889618, 'timing/model': 1.0830807695165277, 'learning_rate/qwen_vl_interface': 9.734858358395808e-07, 'learning_rate/action_model': 9.734858358395807e-06, 'epoch': 2.31}) +[2026-06-18 07:04:53,029][starVLA.training.train_starvla][INFO] - Step 6012, Loss: {'train/loss': 0.35761648416519165, 'train/grad_norm_pre_clip': 17.731462478637695, 'timing/data': 0.024157912470400333, 'timing/model': 1.6638288777321577, 'learning_rate/qwen_vl_interface': 9.72545008929892e-07, 'learning_rate/action_model': 9.725450089298921e-06, 'epoch': 2.31}) +[2026-06-18 07:04:58,037][starVLA.training.train_starvla][INFO] - Step 6013, Loss: {'train/loss': 0.3157760500907898, 'train/grad_norm_pre_clip': 16.660438537597656, 'timing/data': 0.023517418652772903, 'timing/model': 1.3387156473472714, 'learning_rate/qwen_vl_interface': 9.71605068741225e-07, 'learning_rate/action_model': 9.716050687412251e-06, 'epoch': 2.31}) +[2026-06-18 07:05:14,102][starVLA.training.train_starvla][INFO] - Step 6014, Loss: {'train/loss': 0.33075568079948425, 'train/grad_norm_pre_clip': 15.177998542785645, 'timing/data': 0.022512812167406082, 'timing/model': 4.653090790845454, 'learning_rate/qwen_vl_interface': 9.706660154684312e-07, 'learning_rate/action_model': 9.706660154684312e-06, 'epoch': 2.31}) +[2026-06-18 07:05:23,903][starVLA.training.train_starvla][INFO] - Step 6015, Loss: {'train/loss': 0.47296640276908875, 'train/grad_norm_pre_clip': 16.374858856201172, 'timing/data': 0.02062523551285267, 'timing/model': 1.5303213624283671, 'learning_rate/qwen_vl_interface': 9.697278493061773e-07, 'learning_rate/action_model': 9.697278493061773e-06, 'epoch': 2.31}) +[2026-06-18 07:05:28,939][starVLA.training.train_starvla][INFO] - Step 6016, Loss: {'train/loss': 0.519722044467926, 'train/grad_norm_pre_clip': 15.1906156539917, 'timing/data': 0.028134018182754517, 'timing/model': 1.6389314075931907, 'learning_rate/qwen_vl_interface': 9.687905704489461e-07, 'learning_rate/action_model': 9.687905704489462e-06, 'epoch': 2.31}) +[2026-06-18 07:05:35,668][starVLA.training.train_starvla][INFO] - Step 6017, Loss: {'train/loss': 0.3258991837501526, 'train/grad_norm_pre_clip': 19.418176651000977, 'timing/data': 0.027362233959138393, 'timing/model': 1.674407446756959, 'learning_rate/qwen_vl_interface': 9.678541790910361e-07, 'learning_rate/action_model': 9.678541790910361e-06, 'epoch': 2.31}) +[2026-06-18 07:05:47,007][starVLA.training.train_starvla][INFO] - Step 6018, Loss: {'train/loss': 0.09607457369565964, 'train/grad_norm_pre_clip': 15.816736221313477, 'timing/data': 0.026774567551910877, 'timing/model': 1.4132175827398896, 'learning_rate/qwen_vl_interface': 9.669186754265619e-07, 'learning_rate/action_model': 9.669186754265618e-06, 'epoch': 2.32}) +[2026-06-18 07:05:53,503][starVLA.training.train_starvla][INFO] - Step 6019, Loss: {'train/loss': 0.45944342017173767, 'train/grad_norm_pre_clip': 16.33924674987793, 'timing/data': 0.02117238100618124, 'timing/model': 2.6990334624424577, 'learning_rate/qwen_vl_interface': 9.65984059649457e-07, 'learning_rate/action_model': 9.65984059649457e-06, 'epoch': 2.32}) +[2026-06-18 07:05:58,783][starVLA.training.train_starvla][INFO] - Step 6020, Loss: {'train/loss': 0.35272955894470215, 'train/grad_norm_pre_clip': 20.29877471923828, 'timing/data': 0.02077456284314394, 'timing/model': 1.141957783140242, 'learning_rate/qwen_vl_interface': 9.650503319534652e-07, 'learning_rate/action_model': 9.650503319534651e-06, 'epoch': 2.32}) +[2026-06-18 07:06:03,706][starVLA.training.train_starvla][INFO] - Step 6021, Loss: {'train/loss': 0.5815987586975098, 'train/grad_norm_pre_clip': 19.101896286010742, 'timing/data': 0.02027007658034563, 'timing/model': 1.0848155850544572, 'learning_rate/qwen_vl_interface': 9.641174925321494e-07, 'learning_rate/action_model': 9.641174925321494e-06, 'epoch': 2.32}) +[2026-06-18 07:06:13,418][starVLA.training.train_starvla][INFO] - Step 6022, Loss: {'train/loss': 0.4568553566932678, 'train/grad_norm_pre_clip': 23.280153274536133, 'timing/data': 0.022706789895892143, 'timing/model': 6.124203070066869, 'learning_rate/qwen_vl_interface': 9.631855415788904e-07, 'learning_rate/action_model': 9.631855415788904e-06, 'epoch': 2.32}) +[2026-06-18 07:06:18,669][starVLA.training.train_starvla][INFO] - Step 6023, Loss: {'train/loss': 0.4561667740345001, 'train/grad_norm_pre_clip': 18.46797752380371, 'timing/data': 0.021749253384768963, 'timing/model': 1.1335530076175928, 'learning_rate/qwen_vl_interface': 9.622544792868812e-07, 'learning_rate/action_model': 9.622544792868812e-06, 'epoch': 2.32}) +[2026-06-18 07:06:31,549][starVLA.training.train_starvla][INFO] - Step 6024, Loss: {'train/loss': 0.31497085094451904, 'train/grad_norm_pre_clip': 14.491631507873535, 'timing/data': 0.027643457986414433, 'timing/model': 5.5521077597513795, 'learning_rate/qwen_vl_interface': 9.613243058491334e-07, 'learning_rate/action_model': 9.613243058491334e-06, 'epoch': 2.32}) +[2026-06-18 07:06:37,251][starVLA.training.train_starvla][INFO] - Step 6025, Loss: {'train/loss': 0.3948000371456146, 'train/grad_norm_pre_clip': 16.3728084564209, 'timing/data': 0.022513494826853275, 'timing/model': 1.553816583007574, 'learning_rate/qwen_vl_interface': 9.603950214584695e-07, 'learning_rate/action_model': 9.603950214584696e-06, 'epoch': 2.32}) +[2026-06-18 07:06:44,259][starVLA.training.train_starvla][INFO] - Step 6026, Loss: {'train/loss': 0.1435554325580597, 'train/grad_norm_pre_clip': 9.785815238952637, 'timing/data': 0.022713985294103622, 'timing/model': 1.2115608276799321, 'learning_rate/qwen_vl_interface': 9.59466626307535e-07, 'learning_rate/action_model': 9.594666263075349e-06, 'epoch': 2.32}) +[2026-06-18 07:06:50,079][starVLA.training.train_starvla][INFO] - Step 6027, Loss: {'train/loss': 0.25013110041618347, 'train/grad_norm_pre_clip': 11.797941207885742, 'timing/data': 0.023977220058441162, 'timing/model': 1.78505555447191, 'learning_rate/qwen_vl_interface': 9.585391205887851e-07, 'learning_rate/action_model': 9.585391205887851e-06, 'epoch': 2.32}) +[2026-06-18 07:06:55,455][starVLA.training.train_starvla][INFO] - Step 6028, Loss: {'train/loss': 0.24329480528831482, 'train/grad_norm_pre_clip': 11.408064842224121, 'timing/data': 0.020617123693227768, 'timing/model': 1.2779925521463156, 'learning_rate/qwen_vl_interface': 9.57612504494494e-07, 'learning_rate/action_model': 9.57612504494494e-06, 'epoch': 2.32}) +[2026-06-18 07:07:08,228][starVLA.training.train_starvla][INFO] - Step 6029, Loss: {'train/loss': 0.1790667027235031, 'train/grad_norm_pre_clip': 10.883051872253418, 'timing/data': 0.02448262646794319, 'timing/model': 1.5305022904649377, 'learning_rate/qwen_vl_interface': 9.566867782167494e-07, 'learning_rate/action_model': 9.566867782167494e-06, 'epoch': 2.32}) +[2026-06-18 07:07:13,406][starVLA.training.train_starvla][INFO] - Step 6030, Loss: {'train/loss': 0.26307618618011475, 'train/grad_norm_pre_clip': 15.756360054016113, 'timing/data': 0.021594926714897156, 'timing/model': 1.3284072643145919, 'learning_rate/qwen_vl_interface': 9.55761941947455e-07, 'learning_rate/action_model': 9.557619419474551e-06, 'epoch': 2.32}) +[2026-06-18 07:07:18,478][starVLA.training.train_starvla][INFO] - Step 6031, Loss: {'train/loss': 0.32865849137306213, 'train/grad_norm_pre_clip': 15.55547046661377, 'timing/data': 0.023545279167592525, 'timing/model': 1.2056781761348248, 'learning_rate/qwen_vl_interface': 9.548379958783308e-07, 'learning_rate/action_model': 9.548379958783308e-06, 'epoch': 2.32}) +[2026-06-18 07:07:23,847][starVLA.training.train_starvla][INFO] - Step 6032, Loss: {'train/loss': 0.20598389208316803, 'train/grad_norm_pre_clip': 12.483372688293457, 'timing/data': 0.027343968860805035, 'timing/model': 1.501030127517879, 'learning_rate/qwen_vl_interface': 9.539149402009114e-07, 'learning_rate/action_model': 9.539149402009115e-06, 'epoch': 2.32}) +[2026-06-18 07:07:29,113][starVLA.training.train_starvla][INFO] - Step 6033, Loss: {'train/loss': 0.26771754026412964, 'train/grad_norm_pre_clip': 14.22330379486084, 'timing/data': 0.010914861224591732, 'timing/model': 1.3152994392439723, 'learning_rate/qwen_vl_interface': 9.529927751065493e-07, 'learning_rate/action_model': 9.529927751065492e-06, 'epoch': 2.32}) +[2026-06-18 07:07:36,034][starVLA.training.train_starvla][INFO] - Step 6034, Loss: {'train/loss': 0.3398129642009735, 'train/grad_norm_pre_clip': 14.336451530456543, 'timing/data': 0.014668803662061691, 'timing/model': 1.2787864580750465, 'learning_rate/qwen_vl_interface': 9.520715007864082e-07, 'learning_rate/action_model': 9.52071500786408e-06, 'epoch': 2.32}) +[2026-06-18 07:07:45,927][starVLA.training.train_starvla][INFO] - Step 6035, Loss: {'train/loss': 0.20203988254070282, 'train/grad_norm_pre_clip': 10.518815040588379, 'timing/data': 0.07035892084240913, 'timing/model': 4.096319663338363, 'learning_rate/qwen_vl_interface': 9.511511174314689e-07, 'learning_rate/action_model': 9.511511174314688e-06, 'epoch': 2.32}) +[2026-06-18 07:07:51,748][starVLA.training.train_starvla][INFO] - Step 6036, Loss: {'train/loss': 0.10467738658189774, 'train/grad_norm_pre_clip': 6.098169803619385, 'timing/data': 0.015825157053768635, 'timing/model': 1.6563922530040145, 'learning_rate/qwen_vl_interface': 9.502316252325283e-07, 'learning_rate/action_model': 9.502316252325283e-06, 'epoch': 2.32}) +[2026-06-18 07:07:59,972][starVLA.training.train_starvla][INFO] - Step 6037, Loss: {'train/loss': 0.3543686270713806, 'train/grad_norm_pre_clip': 12.22747802734375, 'timing/data': 0.014166978187859058, 'timing/model': 1.9779056757688522, 'learning_rate/qwen_vl_interface': 9.493130243801993e-07, 'learning_rate/action_model': 9.493130243801992e-06, 'epoch': 2.32}) +[2026-06-18 07:08:13,171][starVLA.training.train_starvla][INFO] - Step 6038, Loss: {'train/loss': 0.2861807644367218, 'train/grad_norm_pre_clip': 15.921740531921387, 'timing/data': 0.023024188354611397, 'timing/model': 4.739903152920306, 'learning_rate/qwen_vl_interface': 9.483953150649091e-07, 'learning_rate/action_model': 9.483953150649091e-06, 'epoch': 2.32}) +[2026-06-18 07:08:19,659][starVLA.training.train_starvla][INFO] - Step 6039, Loss: {'train/loss': 0.25195446610450745, 'train/grad_norm_pre_clip': 9.723298072814941, 'timing/data': 0.011819564737379551, 'timing/model': 1.2215771386399865, 'learning_rate/qwen_vl_interface': 9.474784974768982e-07, 'learning_rate/action_model': 9.474784974768982e-06, 'epoch': 2.32}) +[2026-06-18 07:08:24,876][starVLA.training.train_starvla][INFO] - Step 6040, Loss: {'train/loss': 0.3749869167804718, 'train/grad_norm_pre_clip': 16.208621978759766, 'timing/data': 0.011691314168274403, 'timing/model': 1.2953449096530676, 'learning_rate/qwen_vl_interface': 9.465625718062236e-07, 'learning_rate/action_model': 9.465625718062235e-06, 'epoch': 2.32}) +[2026-06-18 07:08:29,871][starVLA.training.train_starvla][INFO] - Step 6041, Loss: {'train/loss': 0.17034198343753815, 'train/grad_norm_pre_clip': 13.761359214782715, 'timing/data': 0.011410406790673733, 'timing/model': 1.1195103032514453, 'learning_rate/qwen_vl_interface': 9.456475382427591e-07, 'learning_rate/action_model': 9.45647538242759e-06, 'epoch': 2.32}) +[2026-06-18 07:08:39,085][starVLA.training.train_starvla][INFO] - Step 6042, Loss: {'train/loss': 0.38018426299095154, 'train/grad_norm_pre_clip': 14.689655303955078, 'timing/data': 0.017354131676256657, 'timing/model': 1.4985095001757145, 'learning_rate/qwen_vl_interface': 9.447333969761915e-07, 'learning_rate/action_model': 9.447333969761915e-06, 'epoch': 2.32}) +[2026-06-18 07:08:44,016][starVLA.training.train_starvla][INFO] - Step 6043, Loss: {'train/loss': 0.46308502554893494, 'train/grad_norm_pre_clip': 14.618212699890137, 'timing/data': 0.01565225888043642, 'timing/model': 1.084999236278236, 'learning_rate/qwen_vl_interface': 9.438201481960235e-07, 'learning_rate/action_model': 9.438201481960234e-06, 'epoch': 2.32}) +[2026-06-18 07:08:49,251][starVLA.training.train_starvla][INFO] - Step 6044, Loss: {'train/loss': 0.431045264005661, 'train/grad_norm_pre_clip': 15.120556831359863, 'timing/data': 0.012757012620568275, 'timing/model': 1.250701416283846, 'learning_rate/qwen_vl_interface': 9.429077920915724e-07, 'learning_rate/action_model': 9.429077920915724e-06, 'epoch': 2.33}) +[2026-06-18 07:08:58,032][starVLA.training.train_starvla][INFO] - Step 6045, Loss: {'train/loss': 0.30286258459091187, 'train/grad_norm_pre_clip': 19.957454681396484, 'timing/data': 0.014883222058415413, 'timing/model': 1.2171607231721282, 'learning_rate/qwen_vl_interface': 9.419963288519711e-07, 'learning_rate/action_model': 9.41996328851971e-06, 'epoch': 2.33}) +[2026-06-18 07:09:10,950][starVLA.training.train_starvla][INFO] - Step 6046, Loss: {'train/loss': 0.05971263349056244, 'train/grad_norm_pre_clip': 5.892050266265869, 'timing/data': 0.010502400808036327, 'timing/model': 4.7589449705556035, 'learning_rate/qwen_vl_interface': 9.410857586661659e-07, 'learning_rate/action_model': 9.410857586661659e-06, 'epoch': 2.33}) +[2026-06-18 07:09:25,930][starVLA.training.train_starvla][INFO] - Step 6047, Loss: {'train/loss': 0.35775354504585266, 'train/grad_norm_pre_clip': 18.214452743530273, 'timing/data': 0.01702672243118286, 'timing/model': 4.454251437447965, 'learning_rate/qwen_vl_interface': 9.401760817229197e-07, 'learning_rate/action_model': 9.401760817229196e-06, 'epoch': 2.33}) +[2026-06-18 07:09:33,595][starVLA.training.train_starvla][INFO] - Step 6048, Loss: {'train/loss': 0.20874173939228058, 'train/grad_norm_pre_clip': 16.070100784301758, 'timing/data': 0.013297476805746555, 'timing/model': 1.158423826098442, 'learning_rate/qwen_vl_interface': 9.392672982108091e-07, 'learning_rate/action_model': 9.392672982108091e-06, 'epoch': 2.33}) +[2026-06-18 07:09:46,149][starVLA.training.train_starvla][INFO] - Step 6049, Loss: {'train/loss': 0.6288736462593079, 'train/grad_norm_pre_clip': 20.75157928466797, 'timing/data': 0.014842603355646133, 'timing/model': 3.315712545067072, 'learning_rate/qwen_vl_interface': 9.383594083182265e-07, 'learning_rate/action_model': 9.383594083182265e-06, 'epoch': 2.33}) +[2026-06-18 07:09:55,093][starVLA.training.train_starvla][INFO] - Step 6050, Loss: {'train/loss': 0.4062153995037079, 'train/grad_norm_pre_clip': 16.37533950805664, 'timing/data': 0.017955129966139793, 'timing/model': 1.1460104444995522, 'learning_rate/qwen_vl_interface': 9.374524122333769e-07, 'learning_rate/action_model': 9.374524122333769e-06, 'epoch': 2.33}) +[2026-06-18 07:10:07,356][starVLA.training.train_starvla][INFO] - Step 6051, Loss: {'train/loss': 0.310986191034317, 'train/grad_norm_pre_clip': 13.576643943786621, 'timing/data': 0.011764227412641048, 'timing/model': 1.103994881734252, 'learning_rate/qwen_vl_interface': 9.365463101442842e-07, 'learning_rate/action_model': 9.365463101442842e-06, 'epoch': 2.33}) +[2026-06-18 07:10:12,912][starVLA.training.train_starvla][INFO] - Step 6052, Loss: {'train/loss': 0.2447819858789444, 'train/grad_norm_pre_clip': 10.474215507507324, 'timing/data': 0.02472520899027586, 'timing/model': 1.5842849323526025, 'learning_rate/qwen_vl_interface': 9.356411022387823e-07, 'learning_rate/action_model': 9.356411022387824e-06, 'epoch': 2.33}) +[2026-06-18 07:10:17,424][starVLA.training.train_starvla][INFO] - Step 6053, Loss: {'train/loss': 0.26878058910369873, 'train/grad_norm_pre_clip': 29.4233455657959, 'timing/data': 0.022954190149903297, 'timing/model': 1.325851246714592, 'learning_rate/qwen_vl_interface': 9.347367887045236e-07, 'learning_rate/action_model': 9.347367887045236e-06, 'epoch': 2.33}) +[2026-06-18 07:10:22,793][starVLA.training.train_starvla][INFO] - Step 6054, Loss: {'train/loss': 0.32769277691841125, 'train/grad_norm_pre_clip': 23.434328079223633, 'timing/data': 0.017859062179923058, 'timing/model': 1.2638068366795778, 'learning_rate/qwen_vl_interface': 9.338333697289697e-07, 'learning_rate/action_model': 9.338333697289696e-06, 'epoch': 2.33}) +[2026-06-18 07:10:28,848][starVLA.training.train_starvla][INFO] - Step 6055, Loss: {'train/loss': 0.1926751583814621, 'train/grad_norm_pre_clip': 9.488409996032715, 'timing/data': 0.015162882395088673, 'timing/model': 1.5826489767059684, 'learning_rate/qwen_vl_interface': 9.329308454994044e-07, 'learning_rate/action_model': 9.329308454994043e-06, 'epoch': 2.33}) +[2026-06-18 07:10:34,014][starVLA.training.train_starvla][INFO] - Step 6056, Loss: {'train/loss': 0.5853129029273987, 'train/grad_norm_pre_clip': 17.397064208984375, 'timing/data': 0.010740293189883232, 'timing/model': 1.2342466646805406, 'learning_rate/qwen_vl_interface': 9.320292162029192e-07, 'learning_rate/action_model': 9.320292162029191e-06, 'epoch': 2.33}) +[2026-06-18 07:10:39,666][starVLA.training.train_starvla][INFO] - Step 6057, Loss: {'train/loss': 0.34095391631126404, 'train/grad_norm_pre_clip': 23.598495483398438, 'timing/data': 0.01296587660908699, 'timing/model': 1.6657365718856454, 'learning_rate/qwen_vl_interface': 9.311284820264245e-07, 'learning_rate/action_model': 9.311284820264244e-06, 'epoch': 2.33}) +[2026-06-18 07:10:44,454][starVLA.training.train_starvla][INFO] - Step 6058, Loss: {'train/loss': 0.1539698839187622, 'train/grad_norm_pre_clip': 10.766396522521973, 'timing/data': 0.011894749477505684, 'timing/model': 1.459964346140623, 'learning_rate/qwen_vl_interface': 9.302286431566428e-07, 'learning_rate/action_model': 9.302286431566428e-06, 'epoch': 2.33}) +[2026-06-18 07:10:49,708][starVLA.training.train_starvla][INFO] - Step 6059, Loss: {'train/loss': 0.12057054787874222, 'train/grad_norm_pre_clip': 10.01447868347168, 'timing/data': 0.011802037246525288, 'timing/model': 1.3041357528418303, 'learning_rate/qwen_vl_interface': 9.293296997801113e-07, 'learning_rate/action_model': 9.293296997801112e-06, 'epoch': 2.33}) +[2026-06-18 07:10:54,634][starVLA.training.train_starvla][INFO] - Step 6060, Loss: {'train/loss': 0.3496769368648529, 'train/grad_norm_pre_clip': 17.795055389404297, 'timing/data': 0.019538518972694874, 'timing/model': 1.1119896788150072, 'learning_rate/qwen_vl_interface': 9.284316520831825e-07, 'learning_rate/action_model': 9.284316520831825e-06, 'epoch': 2.33}) +[2026-06-18 07:10:59,544][starVLA.training.train_starvla][INFO] - Step 6061, Loss: {'train/loss': 0.2531260848045349, 'train/grad_norm_pre_clip': 13.950630187988281, 'timing/data': 0.013178606517612934, 'timing/model': 1.140390177257359, 'learning_rate/qwen_vl_interface': 9.275345002520225e-07, 'learning_rate/action_model': 9.275345002520225e-06, 'epoch': 2.33}) +[2026-06-18 07:11:04,476][starVLA.training.train_starvla][INFO] - Step 6062, Loss: {'train/loss': 0.7004337310791016, 'train/grad_norm_pre_clip': 25.64741325378418, 'timing/data': 0.011720897629857063, 'timing/model': 1.4286146396771073, 'learning_rate/qwen_vl_interface': 9.266382444726117e-07, 'learning_rate/action_model': 9.266382444726117e-06, 'epoch': 2.33}) +[2026-06-18 07:11:09,245][starVLA.training.train_starvla][INFO] - Step 6063, Loss: {'train/loss': 0.18521445989608765, 'train/grad_norm_pre_clip': 9.51417064666748, 'timing/data': 0.016601470299065113, 'timing/model': 1.1364947017282248, 'learning_rate/qwen_vl_interface': 9.257428849307446e-07, 'learning_rate/action_model': 9.257428849307445e-06, 'epoch': 2.33}) +[2026-06-18 07:11:14,049][starVLA.training.train_starvla][INFO] - Step 6064, Loss: {'train/loss': 0.607430636882782, 'train/grad_norm_pre_clip': 22.936479568481445, 'timing/data': 0.010824753902852535, 'timing/model': 1.0916126063093543, 'learning_rate/qwen_vl_interface': 9.248484218120307e-07, 'learning_rate/action_model': 9.248484218120307e-06, 'epoch': 2.33}) +[2026-06-18 07:11:19,452][starVLA.training.train_starvla][INFO] - Step 6065, Loss: {'train/loss': 0.6516145467758179, 'train/grad_norm_pre_clip': 22.67043685913086, 'timing/data': 0.013490275479853153, 'timing/model': 1.1910830112174153, 'learning_rate/qwen_vl_interface': 9.239548553018925e-07, 'learning_rate/action_model': 9.239548553018924e-06, 'epoch': 2.33}) +[2026-06-18 07:11:24,485][starVLA.training.train_starvla][INFO] - Step 6066, Loss: {'train/loss': 0.3004685044288635, 'train/grad_norm_pre_clip': 13.256169319152832, 'timing/data': 0.0127432681620121, 'timing/model': 1.155763553455472, 'learning_rate/qwen_vl_interface': 9.230621855855687e-07, 'learning_rate/action_model': 9.230621855855687e-06, 'epoch': 2.33}) +[2026-06-18 07:11:36,713][starVLA.training.train_starvla][INFO] - Step 6067, Loss: {'train/loss': 0.2104298323392868, 'train/grad_norm_pre_clip': 12.132569313049316, 'timing/data': 0.02540880162268877, 'timing/model': 2.1961272666230798, 'learning_rate/qwen_vl_interface': 9.221704128481107e-07, 'learning_rate/action_model': 9.221704128481106e-06, 'epoch': 2.33}) +[2026-06-18 07:11:47,663][starVLA.training.train_starvla][INFO] - Step 6068, Loss: {'train/loss': 0.24019956588745117, 'train/grad_norm_pre_clip': 10.140735626220703, 'timing/data': 0.021416330710053444, 'timing/model': 5.540988525375724, 'learning_rate/qwen_vl_interface': 9.212795372743815e-07, 'learning_rate/action_model': 9.212795372743813e-06, 'epoch': 2.33}) +[2026-06-18 07:12:03,789][starVLA.training.train_starvla][INFO] - Step 6069, Loss: {'train/loss': 0.3467908799648285, 'train/grad_norm_pre_clip': 19.868087768554688, 'timing/data': 0.01508930791169405, 'timing/model': 1.8211015816777945, 'learning_rate/qwen_vl_interface': 9.203895590490617e-07, 'learning_rate/action_model': 9.203895590490616e-06, 'epoch': 2.33}) +[2026-06-18 07:12:14,667][starVLA.training.train_starvla][INFO] - Step 6070, Loss: {'train/loss': 0.3705008327960968, 'train/grad_norm_pre_clip': 15.62877368927002, 'timing/data': 0.013017535209655762, 'timing/model': 2.5278019569814205, 'learning_rate/qwen_vl_interface': 9.195004783566464e-07, 'learning_rate/action_model': 9.195004783566463e-06, 'epoch': 2.34}) +[2026-06-18 07:12:30,052][starVLA.training.train_starvla][INFO] - Step 6071, Loss: {'train/loss': 0.24817205965518951, 'train/grad_norm_pre_clip': 12.469488143920898, 'timing/data': 0.021096394397318363, 'timing/model': 2.826637231744826, 'learning_rate/qwen_vl_interface': 9.186122953814414e-07, 'learning_rate/action_model': 9.186122953814413e-06, 'epoch': 2.34}) +[2026-06-18 07:12:50,714][starVLA.training.train_starvla][INFO] - Step 6072, Loss: {'train/loss': 0.2814503610134125, 'train/grad_norm_pre_clip': 21.957380294799805, 'timing/data': 0.0008881194517016411, 'timing/model': 5.42737205978483, 'learning_rate/qwen_vl_interface': 9.177250103075686e-07, 'learning_rate/action_model': 9.177250103075686e-06, 'epoch': 2.34}) +[2026-06-18 07:13:04,287][starVLA.training.train_starvla][INFO] - Step 6073, Loss: {'train/loss': 0.19002766907215118, 'train/grad_norm_pre_clip': 10.894221305847168, 'timing/data': 0.0014010714367032051, 'timing/model': 1.1667864732444286, 'learning_rate/qwen_vl_interface': 9.168386233189627e-07, 'learning_rate/action_model': 9.168386233189627e-06, 'epoch': 2.34}) +[2026-06-18 07:13:10,216][starVLA.training.train_starvla][INFO] - Step 6074, Loss: {'train/loss': 0.25620946288108826, 'train/grad_norm_pre_clip': 15.028918266296387, 'timing/data': 0.000999322161078453, 'timing/model': 1.5690121408551931, 'learning_rate/qwen_vl_interface': 9.159531345993726e-07, 'learning_rate/action_model': 9.159531345993725e-06, 'epoch': 2.34}) +[2026-06-18 07:13:22,329][starVLA.training.train_starvla][INFO] - Step 6075, Loss: {'train/loss': 0.26786255836486816, 'train/grad_norm_pre_clip': 23.767391204833984, 'timing/data': 0.0016223099082708359, 'timing/model': 1.2074891095981002, 'learning_rate/qwen_vl_interface': 9.150685443323619e-07, 'learning_rate/action_model': 9.150685443323618e-06, 'epoch': 2.34}) +[2026-06-18 07:13:27,805][starVLA.training.train_starvla][INFO] - Step 6076, Loss: {'train/loss': 0.26765626668930054, 'train/grad_norm_pre_clip': 13.106352806091309, 'timing/data': 0.0011839605867862701, 'timing/model': 1.2301075365394354, 'learning_rate/qwen_vl_interface': 9.141848527013059e-07, 'learning_rate/action_model': 9.141848527013059e-06, 'epoch': 2.34}) +[2026-06-18 07:13:32,934][starVLA.training.train_starvla][INFO] - Step 6077, Loss: {'train/loss': 0.34225907921791077, 'train/grad_norm_pre_clip': 15.425434112548828, 'timing/data': 0.0013173297047615051, 'timing/model': 1.649554549716413, 'learning_rate/qwen_vl_interface': 9.13302059889396e-07, 'learning_rate/action_model': 9.13302059889396e-06, 'epoch': 2.34}) +[2026-06-18 07:13:37,859][starVLA.training.train_starvla][INFO] - Step 6078, Loss: {'train/loss': 0.3892422914505005, 'train/grad_norm_pre_clip': 19.042499542236328, 'timing/data': 0.0013617277145385742, 'timing/model': 1.28553656488657, 'learning_rate/qwen_vl_interface': 9.12420166079635e-07, 'learning_rate/action_model': 9.12420166079635e-06, 'epoch': 2.34}) +[2026-06-18 07:13:43,380][starVLA.training.train_starvla][INFO] - Step 6079, Loss: {'train/loss': 0.30112865567207336, 'train/grad_norm_pre_clip': 18.783113479614258, 'timing/data': 0.0010369820520281792, 'timing/model': 1.2155469739809632, 'learning_rate/qwen_vl_interface': 9.115391714548418e-07, 'learning_rate/action_model': 9.115391714548418e-06, 'epoch': 2.34}) +[2026-06-18 07:13:48,412][starVLA.training.train_starvla][INFO] - Step 6080, Loss: {'train/loss': 0.26207417249679565, 'train/grad_norm_pre_clip': 19.11343002319336, 'timing/data': 0.001014532521367073, 'timing/model': 1.2343908362090588, 'learning_rate/qwen_vl_interface': 9.106590761976454e-07, 'learning_rate/action_model': 9.106590761976455e-06, 'epoch': 2.34}) +[2026-06-18 07:13:53,815][starVLA.training.train_starvla][INFO] - Step 6081, Loss: {'train/loss': 0.33710575103759766, 'train/grad_norm_pre_clip': 13.735228538513184, 'timing/data': 0.001601451076567173, 'timing/model': 1.4282505875453353, 'learning_rate/qwen_vl_interface': 9.097798804904934e-07, 'learning_rate/action_model': 9.097798804904933e-06, 'epoch': 2.34}) +[2026-06-18 07:14:08,823][starVLA.training.train_starvla][INFO] - Step 6082, Loss: {'train/loss': 0.20225505530834198, 'train/grad_norm_pre_clip': 12.223307609558105, 'timing/data': 0.0013298820704221725, 'timing/model': 4.7095151003450155, 'learning_rate/qwen_vl_interface': 9.089015845156422e-07, 'learning_rate/action_model': 9.089015845156422e-06, 'epoch': 2.34}) +[2026-06-18 07:14:14,061][starVLA.training.train_starvla][INFO] - Step 6083, Loss: {'train/loss': 0.24300403892993927, 'train/grad_norm_pre_clip': 15.01058292388916, 'timing/data': 0.0031303279101848602, 'timing/model': 1.1858651088550687, 'learning_rate/qwen_vl_interface': 9.080241884551627e-07, 'learning_rate/action_model': 9.080241884551627e-06, 'epoch': 2.34}) +[2026-06-18 07:14:27,647][starVLA.training.train_starvla][INFO] - Step 6084, Loss: {'train/loss': 0.29943937063217163, 'train/grad_norm_pre_clip': 13.337701797485352, 'timing/data': 0.0009613418951630592, 'timing/model': 2.8096310580149293, 'learning_rate/qwen_vl_interface': 9.071476924909417e-07, 'learning_rate/action_model': 9.071476924909416e-06, 'epoch': 2.34}) +[2026-06-18 07:14:33,418][starVLA.training.train_starvla][INFO] - Step 6085, Loss: {'train/loss': 0.23237210512161255, 'train/grad_norm_pre_clip': 13.094893455505371, 'timing/data': 0.0011710477992892265, 'timing/model': 1.2984243109822273, 'learning_rate/qwen_vl_interface': 9.062720968046773e-07, 'learning_rate/action_model': 9.062720968046773e-06, 'epoch': 2.34}) +[2026-06-18 07:14:39,137][starVLA.training.train_starvla][INFO] - Step 6086, Loss: {'train/loss': 0.2851300537586212, 'train/grad_norm_pre_clip': 11.441021919250488, 'timing/data': 0.001263774000108242, 'timing/model': 1.6246708119288087, 'learning_rate/qwen_vl_interface': 9.053974015778817e-07, 'learning_rate/action_model': 9.053974015778817e-06, 'epoch': 2.34}) +[2026-06-18 07:14:44,383][starVLA.training.train_starvla][INFO] - Step 6087, Loss: {'train/loss': 0.2580328583717346, 'train/grad_norm_pre_clip': 14.889796257019043, 'timing/data': 0.0009557269513607025, 'timing/model': 1.4362009763717651, 'learning_rate/qwen_vl_interface': 9.045236069918788e-07, 'learning_rate/action_model': 9.045236069918788e-06, 'epoch': 2.34}) +[2026-06-18 07:14:54,034][starVLA.training.train_starvla][INFO] - Step 6088, Loss: {'train/loss': 0.219156876206398, 'train/grad_norm_pre_clip': 12.664994239807129, 'timing/data': 0.0012291781604290009, 'timing/model': 5.8732644421979785, 'learning_rate/qwen_vl_interface': 9.036507132278085e-07, 'learning_rate/action_model': 9.036507132278086e-06, 'epoch': 2.34}) +[2026-06-18 07:15:01,045][starVLA.training.train_starvla][INFO] - Step 6089, Loss: {'train/loss': 0.37270596623420715, 'train/grad_norm_pre_clip': 17.424287796020508, 'timing/data': 0.0012298505753278732, 'timing/model': 1.0993707412853837, 'learning_rate/qwen_vl_interface': 9.02778720466622e-07, 'learning_rate/action_model': 9.02778720466622e-06, 'epoch': 2.34}) +[2026-06-18 07:15:13,168][starVLA.training.train_starvla][INFO] - Step 6090, Loss: {'train/loss': 0.29313740134239197, 'train/grad_norm_pre_clip': 18.871191024780273, 'timing/data': 0.001486477442085743, 'timing/model': 4.773314733989537, 'learning_rate/qwen_vl_interface': 9.019076288890852e-07, 'learning_rate/action_model': 9.019076288890852e-06, 'epoch': 2.34}) +[2026-06-18 07:15:20,219][starVLA.training.train_starvla][INFO] - Step 6091, Loss: {'train/loss': 0.20042085647583008, 'train/grad_norm_pre_clip': 29.31792449951172, 'timing/data': 0.001232583075761795, 'timing/model': 1.4522483255714178, 'learning_rate/qwen_vl_interface': 9.010374386757755e-07, 'learning_rate/action_model': 9.010374386757754e-06, 'epoch': 2.34}) +[2026-06-18 07:15:25,356][starVLA.training.train_starvla][INFO] - Step 6092, Loss: {'train/loss': 0.12342808395624161, 'train/grad_norm_pre_clip': 8.157525062561035, 'timing/data': 0.001302650198340416, 'timing/model': 1.6460207486525178, 'learning_rate/qwen_vl_interface': 9.001681500070842e-07, 'learning_rate/action_model': 9.001681500070842e-06, 'epoch': 2.34}) +[2026-06-18 07:15:30,344][starVLA.training.train_starvla][INFO] - Step 6093, Loss: {'train/loss': 0.25342050194740295, 'train/grad_norm_pre_clip': 13.975971221923828, 'timing/data': 0.0009141964837908745, 'timing/model': 1.3615679126232862, 'learning_rate/qwen_vl_interface': 8.992997630632161e-07, 'learning_rate/action_model': 8.99299763063216e-06, 'epoch': 2.34}) +[2026-06-18 07:15:35,920][starVLA.training.train_starvla][INFO] - Step 6094, Loss: {'train/loss': 0.23115630447864532, 'train/grad_norm_pre_clip': 11.084300994873047, 'timing/data': 0.0012144315987825394, 'timing/model': 1.1841677576303482, 'learning_rate/qwen_vl_interface': 8.984322780241879e-07, 'learning_rate/action_model': 8.984322780241878e-06, 'epoch': 2.34}) +[2026-06-18 07:15:54,698][starVLA.training.train_starvla][INFO] - Step 6095, Loss: {'train/loss': 0.45316454768180847, 'train/grad_norm_pre_clip': 20.13429832458496, 'timing/data': 0.0012034978717565536, 'timing/model': 3.834976659156382, 'learning_rate/qwen_vl_interface': 8.975656950698325e-07, 'learning_rate/action_model': 8.975656950698324e-06, 'epoch': 2.34}) +[2026-06-18 07:16:00,410][starVLA.training.train_starvla][INFO] - Step 6096, Loss: {'train/loss': 0.31383615732192993, 'train/grad_norm_pre_clip': 15.52216625213623, 'timing/data': 0.0011646673083305359, 'timing/model': 1.3389395279809833, 'learning_rate/qwen_vl_interface': 8.967000143797913e-07, 'learning_rate/action_model': 8.967000143797913e-06, 'epoch': 2.35}) +[2026-06-18 07:16:05,830][starVLA.training.train_starvla][INFO] - Step 6097, Loss: {'train/loss': 0.33825719356536865, 'train/grad_norm_pre_clip': 14.91189956665039, 'timing/data': 0.0013555334880948067, 'timing/model': 1.3767936266958714, 'learning_rate/qwen_vl_interface': 8.958352361335214e-07, 'learning_rate/action_model': 8.958352361335213e-06, 'epoch': 2.35}) +[2026-06-18 07:16:14,386][starVLA.training.train_starvla][INFO] - Step 6098, Loss: {'train/loss': 0.2288360595703125, 'train/grad_norm_pre_clip': 13.40328598022461, 'timing/data': 0.0012068450450897217, 'timing/model': 4.205607743002474, 'learning_rate/qwen_vl_interface': 8.949713605102917e-07, 'learning_rate/action_model': 8.949713605102916e-06, 'epoch': 2.35}) +[2026-06-18 07:16:21,315][starVLA.training.train_starvla][INFO] - Step 6099, Loss: {'train/loss': 0.28243669867515564, 'train/grad_norm_pre_clip': 17.500629425048828, 'timing/data': 0.009430047124624252, 'timing/model': 2.3960331501439214, 'learning_rate/qwen_vl_interface': 8.941083876891857e-07, 'learning_rate/action_model': 8.941083876891856e-06, 'epoch': 2.35}) +[2026-06-18 07:21:05,687][starVLA.training.train_starvla][INFO] - Step 6100, Loss: {'train/loss': 0.15246473252773285, 'train/grad_norm_pre_clip': 8.424225807189941, 'eval/loss': 0.4202488362789154, 'eval/latency_4/loss': 0.4202488362789154, 'eval/demon_fire_precision': 0.967443345036706, 'eval/demon_fire_recall': 0.9844105228970444, 'eval/demon_fire_f1': 0.9758531873792659, 'eval/demon_fire_rate_model': 0.9201174743024964, 'eval/demon_fire_rate_teacher': 0.9042584434654919, 'eval/demon_left_precision': 0.8037135278514589, 'eval/demon_left_recall': 0.6917808219178082, 'eval/demon_left_f1': 0.7435582822085889, 'eval/demon_left_rate_model': 0.11071953010279001, 'eval/demon_left_rate_teacher': 0.12863436123348018, 'eval/demon_right_precision': 0.8210116731517509, 'eval/demon_right_recall': 0.7617328519855595, 'eval/demon_right_f1': 0.7902621722846441, 'eval/demon_right_rate_model': 0.15095447870778267, 'eval/demon_right_rate_teacher': 0.16270190895741557, 'eval/demon_move_f1': 0.7669102272466165, 'eval/demon_cc_f1': 0.9021190941706294, 'eval/latency_4/demon_fire_precision': 0.967443345036706, 'eval/latency_4/demon_fire_recall': 0.9844105228970444, 'eval/latency_4/demon_fire_f1': 0.9758531873792659, 'eval/latency_4/demon_fire_rate_model': 0.9201174743024964, 'eval/latency_4/demon_fire_rate_teacher': 0.9042584434654919, 'eval/latency_4/demon_left_precision': 0.8037135278514589, 'eval/latency_4/demon_left_recall': 0.6917808219178082, 'eval/latency_4/demon_left_f1': 0.7435582822085889, 'eval/latency_4/demon_left_rate_model': 0.11071953010279001, 'eval/latency_4/demon_left_rate_teacher': 0.12863436123348018, 'eval/latency_4/demon_right_precision': 0.8210116731517509, 'eval/latency_4/demon_right_recall': 0.7617328519855595, 'eval/latency_4/demon_right_f1': 0.7902621722846441, 'eval/latency_4/demon_right_rate_model': 0.15095447870778267, 'eval/latency_4/demon_right_rate_teacher': 0.16270190895741557, 'eval/latency_4/demon_move_f1': 0.7669102272466165, 'eval/latency_4/demon_cc_f1': 0.9021190941706294, 'timing/data': 0.0010461080819368362, 'timing/model': 1.3903415063396096, 'learning_rate/qwen_vl_interface': 8.932463178490991e-07, 'learning_rate/action_model': 8.93246317849099e-06, 'epoch': 2.35}) +[2026-06-18 07:21:05,719][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_6100_state +[2026-06-18 07:21:05,729][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 07:21:15,316][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_6100_state/model.safetensors +[2026-06-18 07:21:38,317][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_6100_state/optimizer.bin +[2026-06-18 07:21:38,320][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_6100_state/sampler.bin +[2026-06-18 07:21:38,320][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_6100_state/sampler_1.bin +[2026-06-18 07:21:38,323][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_6100_state/random_states_0.pkl +[2026-06-18 07:21:38,329][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 6100 +[2026-06-18 07:21:38,329][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 07:21:38,347][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 07:21:43,307][starVLA.training.train_starvla][INFO] - Step 6101, Loss: {'train/loss': 0.28854069113731384, 'train/grad_norm_pre_clip': 13.364822387695312, 'timing/data': 0.00125031266361475, 'timing/model': 1.2936134142801166, 'learning_rate/qwen_vl_interface': 8.923851511687379e-07, 'learning_rate/action_model': 8.923851511687378e-06, 'epoch': 2.35}) +[2026-06-18 07:21:48,695][starVLA.training.train_starvla][INFO] - Step 6102, Loss: {'train/loss': 0.5035542249679565, 'train/grad_norm_pre_clip': 33.989295959472656, 'timing/data': 0.0008329069241881371, 'timing/model': 1.7784923929721117, 'learning_rate/qwen_vl_interface': 8.915248878266228e-07, 'learning_rate/action_model': 8.915248878266227e-06, 'epoch': 2.35}) +[2026-06-18 07:21:53,471][starVLA.training.train_starvla][INFO] - Step 6103, Loss: {'train/loss': 0.24389733374118805, 'train/grad_norm_pre_clip': 15.155081748962402, 'timing/data': 0.0012120883911848068, 'timing/model': 1.157703467644751, 'learning_rate/qwen_vl_interface': 8.906655280010895e-07, 'learning_rate/action_model': 8.906655280010896e-06, 'epoch': 2.35}) +[2026-06-18 07:21:58,544][starVLA.training.train_starvla][INFO] - Step 6104, Loss: {'train/loss': 0.2610848546028137, 'train/grad_norm_pre_clip': 17.59809112548828, 'timing/data': 0.0011076843366026878, 'timing/model': 1.3061886904761195, 'learning_rate/qwen_vl_interface': 8.898070718702831e-07, 'learning_rate/action_model': 8.898070718702831e-06, 'epoch': 2.35}) +[2026-06-18 07:22:04,210][starVLA.training.train_starvla][INFO] - Step 6105, Loss: {'train/loss': 0.2843187153339386, 'train/grad_norm_pre_clip': 14.66219425201416, 'timing/data': 0.000867847353219986, 'timing/model': 1.3373244497925043, 'learning_rate/qwen_vl_interface': 8.889495196121628e-07, 'learning_rate/action_model': 8.889495196121627e-06, 'epoch': 2.35}) +[2026-06-18 07:22:09,245][starVLA.training.train_starvla][INFO] - Step 6106, Loss: {'train/loss': 0.22431506216526031, 'train/grad_norm_pre_clip': 12.992461204528809, 'timing/data': 0.0014125797897577286, 'timing/model': 1.3556720931082964, 'learning_rate/qwen_vl_interface': 8.880928714044994e-07, 'learning_rate/action_model': 8.880928714044993e-06, 'epoch': 2.35}) +[2026-06-18 07:22:20,621][starVLA.training.train_starvla][INFO] - Step 6107, Loss: {'train/loss': 0.4027189016342163, 'train/grad_norm_pre_clip': 16.928020477294922, 'timing/data': 0.0012684138491749763, 'timing/model': 1.5009122723713517, 'learning_rate/qwen_vl_interface': 8.872371274248773e-07, 'learning_rate/action_model': 8.872371274248772e-06, 'epoch': 2.35}) +[2026-06-18 07:22:27,445][starVLA.training.train_starvla][INFO] - Step 6108, Loss: {'train/loss': 0.28130051493644714, 'train/grad_norm_pre_clip': 15.846210479736328, 'timing/data': 0.0010339654982089996, 'timing/model': 1.1466384204104543, 'learning_rate/qwen_vl_interface': 8.863822878506931e-07, 'learning_rate/action_model': 8.86382287850693e-06, 'epoch': 2.35}) +[2026-06-18 07:22:32,892][starVLA.training.train_starvla][INFO] - Step 6109, Loss: {'train/loss': 0.3590414524078369, 'train/grad_norm_pre_clip': 21.068565368652344, 'timing/data': 0.0011445404961705208, 'timing/model': 1.4821799341589212, 'learning_rate/qwen_vl_interface': 8.855283528591566e-07, 'learning_rate/action_model': 8.855283528591565e-06, 'epoch': 2.35}) +[2026-06-18 07:22:38,844][starVLA.training.train_starvla][INFO] - Step 6110, Loss: {'train/loss': 0.27620694041252136, 'train/grad_norm_pre_clip': 14.164774894714355, 'timing/data': 0.001203860156238079, 'timing/model': 2.050980537198484, 'learning_rate/qwen_vl_interface': 8.846753226272884e-07, 'learning_rate/action_model': 8.846753226272883e-06, 'epoch': 2.35}) +[2026-06-18 07:22:53,090][starVLA.training.train_starvla][INFO] - Step 6111, Loss: {'train/loss': 0.5385094881057739, 'train/grad_norm_pre_clip': 17.44910430908203, 'timing/data': 0.0011603087186813354, 'timing/model': 4.904883268289268, 'learning_rate/qwen_vl_interface': 8.838231973319235e-07, 'learning_rate/action_model': 8.838231973319235e-06, 'epoch': 2.35}) +[2026-06-18 07:22:58,076][starVLA.training.train_starvla][INFO] - Step 6112, Loss: {'train/loss': 0.41412585973739624, 'train/grad_norm_pre_clip': 52.00419616699219, 'timing/data': 0.0016465643420815468, 'timing/model': 1.5484848394989967, 'learning_rate/qwen_vl_interface': 8.829719771497072e-07, 'learning_rate/action_model': 8.829719771497071e-06, 'epoch': 2.35}) +[2026-06-18 07:23:02,853][starVLA.training.train_starvla][INFO] - Step 6113, Loss: {'train/loss': 0.3384449779987335, 'train/grad_norm_pre_clip': 9.968353271484375, 'timing/data': 0.0012499522417783737, 'timing/model': 1.4980236981064081, 'learning_rate/qwen_vl_interface': 8.821216622570999e-07, 'learning_rate/action_model': 8.821216622570998e-06, 'epoch': 2.35}) +[2026-06-18 07:23:07,951][starVLA.training.train_starvla][INFO] - Step 6114, Loss: {'train/loss': 0.19513621926307678, 'train/grad_norm_pre_clip': 12.936128616333008, 'timing/data': 0.0017261626198887825, 'timing/model': 1.2324025826528668, 'learning_rate/qwen_vl_interface': 8.812722528303727e-07, 'learning_rate/action_model': 8.812722528303727e-06, 'epoch': 2.35}) +[2026-06-18 07:23:13,609][starVLA.training.train_starvla][INFO] - Step 6115, Loss: {'train/loss': 0.4040769636631012, 'train/grad_norm_pre_clip': 17.871307373046875, 'timing/data': 0.0033334679901599884, 'timing/model': 1.1156891146674752, 'learning_rate/qwen_vl_interface': 8.804237490456076e-07, 'learning_rate/action_model': 8.804237490456075e-06, 'epoch': 2.35}) +[2026-06-18 07:23:20,688][starVLA.training.train_starvla][INFO] - Step 6116, Loss: {'train/loss': 0.113739512860775, 'train/grad_norm_pre_clip': 18.6463680267334, 'timing/data': 0.001028609462082386, 'timing/model': 1.4899622537195683, 'learning_rate/qwen_vl_interface': 8.795761510787007e-07, 'learning_rate/action_model': 8.795761510787006e-06, 'epoch': 2.35}) +[2026-06-18 07:23:31,984][starVLA.training.train_starvla][INFO] - Step 6117, Loss: {'train/loss': 0.42591920495033264, 'train/grad_norm_pre_clip': 17.734413146972656, 'timing/data': 0.0010032802820205688, 'timing/model': 1.5429788688197732, 'learning_rate/qwen_vl_interface': 8.78729459105361e-07, 'learning_rate/action_model': 8.78729459105361e-06, 'epoch': 2.35}) +[2026-06-18 07:23:36,734][starVLA.training.train_starvla][INFO] - Step 6118, Loss: {'train/loss': 0.20179666578769684, 'train/grad_norm_pre_clip': 13.22561264038086, 'timing/data': 0.001182730309665203, 'timing/model': 1.0884612947702408, 'learning_rate/qwen_vl_interface': 8.778836733011079e-07, 'learning_rate/action_model': 8.778836733011079e-06, 'epoch': 2.35}) +[2026-06-18 07:23:42,202][starVLA.training.train_starvla][INFO] - Step 6119, Loss: {'train/loss': 0.14888744056224823, 'train/grad_norm_pre_clip': 9.557987213134766, 'timing/data': 0.0009985407814383507, 'timing/model': 1.3996921656653285, 'learning_rate/qwen_vl_interface': 8.770387938412753e-07, 'learning_rate/action_model': 8.770387938412753e-06, 'epoch': 2.35}) +[2026-06-18 07:23:47,353][starVLA.training.train_starvla][INFO] - Step 6120, Loss: {'train/loss': 0.38357993960380554, 'train/grad_norm_pre_clip': 17.820770263671875, 'timing/data': 0.0007825382053852081, 'timing/model': 1.122883977368474, 'learning_rate/qwen_vl_interface': 8.76194820901004e-07, 'learning_rate/action_model': 8.76194820901004e-06, 'epoch': 2.35}) +[2026-06-18 07:23:52,437][starVLA.training.train_starvla][INFO] - Step 6121, Loss: {'train/loss': 0.2989972233772278, 'train/grad_norm_pre_clip': 19.33037567138672, 'timing/data': 0.0014468804001808167, 'timing/model': 1.1899877553805709, 'learning_rate/qwen_vl_interface': 8.753517546552545e-07, 'learning_rate/action_model': 8.753517546552544e-06, 'epoch': 2.35}) +[2026-06-18 07:23:58,560][starVLA.training.train_starvla][INFO] - Step 6122, Loss: {'train/loss': 0.19051630795001984, 'train/grad_norm_pre_clip': 15.316149711608887, 'timing/data': 0.0011084973812103271, 'timing/model': 1.4449403369799256, 'learning_rate/qwen_vl_interface': 8.74509595278793e-07, 'learning_rate/action_model': 8.74509595278793e-06, 'epoch': 2.36}) +[2026-06-18 07:24:06,516][starVLA.training.train_starvla][INFO] - Step 6123, Loss: {'train/loss': 0.044719163328409195, 'train/grad_norm_pre_clip': 2.5622475147247314, 'timing/data': 0.0017227716743946075, 'timing/model': 4.208590437658131, 'learning_rate/qwen_vl_interface': 8.736683429462003e-07, 'learning_rate/action_model': 8.736683429462004e-06, 'epoch': 2.36}) +[2026-06-18 07:24:15,544][starVLA.training.train_starvla][INFO] - Step 6124, Loss: {'train/loss': 0.30773553252220154, 'train/grad_norm_pre_clip': 13.658397674560547, 'timing/data': 0.0009254040196537971, 'timing/model': 1.3533258391544223, 'learning_rate/qwen_vl_interface': 8.728279978318696e-07, 'learning_rate/action_model': 8.728279978318695e-06, 'epoch': 2.36}) +[2026-06-18 07:24:20,647][starVLA.training.train_starvla][INFO] - Step 6125, Loss: {'train/loss': 0.3078015446662903, 'train/grad_norm_pre_clip': 14.943316459655762, 'timing/data': 0.0011078035458922386, 'timing/model': 1.0791510632261634, 'learning_rate/qwen_vl_interface': 8.719885601100052e-07, 'learning_rate/action_model': 8.719885601100051e-06, 'epoch': 2.36}) +[2026-06-18 07:24:25,586][starVLA.training.train_starvla][INFO] - Step 6126, Loss: {'train/loss': 0.29705581068992615, 'train/grad_norm_pre_clip': 15.48204517364502, 'timing/data': 0.0015659835189580917, 'timing/model': 1.1860688040032983, 'learning_rate/qwen_vl_interface': 8.71150029954623e-07, 'learning_rate/action_model': 8.71150029954623e-06, 'epoch': 2.36}) +[2026-06-18 07:24:30,601][starVLA.training.train_starvla][INFO] - Step 6127, Loss: {'train/loss': 0.3207767903804779, 'train/grad_norm_pre_clip': 14.449118614196777, 'timing/data': 0.0014959527179598808, 'timing/model': 1.562788269482553, 'learning_rate/qwen_vl_interface': 8.703124075395504e-07, 'learning_rate/action_model': 8.703124075395504e-06, 'epoch': 2.36}) +[2026-06-18 07:24:39,335][starVLA.training.train_starvla][INFO] - Step 6128, Loss: {'train/loss': 0.42360520362854004, 'train/grad_norm_pre_clip': 18.617000579833984, 'timing/data': 0.0025801220908761024, 'timing/model': 1.5867331996560097, 'learning_rate/qwen_vl_interface': 8.694756930384305e-07, 'learning_rate/action_model': 8.694756930384306e-06, 'epoch': 2.36}) +[2026-06-18 07:24:49,016][starVLA.training.train_starvla][INFO] - Step 6129, Loss: {'train/loss': 0.16912156343460083, 'train/grad_norm_pre_clip': 8.364357948303223, 'timing/data': 0.0012678010389208794, 'timing/model': 4.0163740096613765, 'learning_rate/qwen_vl_interface': 8.686398866247121e-07, 'learning_rate/action_model': 8.68639886624712e-06, 'epoch': 2.36}) +[2026-06-18 07:24:54,467][starVLA.training.train_starvla][INFO] - Step 6130, Loss: {'train/loss': 0.37311500310897827, 'train/grad_norm_pre_clip': 15.05063247680664, 'timing/data': 0.0012732269242405891, 'timing/model': 1.1956737991422415, 'learning_rate/qwen_vl_interface': 8.6780498847166e-07, 'learning_rate/action_model': 8.678049884716599e-06, 'epoch': 2.36}) +[2026-06-18 07:24:59,839][starVLA.training.train_starvla][INFO] - Step 6131, Loss: {'train/loss': 0.2673593759536743, 'train/grad_norm_pre_clip': 14.003087997436523, 'timing/data': 0.0011789845302700996, 'timing/model': 1.8433423219248652, 'learning_rate/qwen_vl_interface': 8.669709987523479e-07, 'learning_rate/action_model': 8.669709987523477e-06, 'epoch': 2.36}) +[2026-06-18 07:25:04,798][starVLA.training.train_starvla][INFO] - Step 6132, Loss: {'train/loss': 0.3928752541542053, 'train/grad_norm_pre_clip': 19.066129684448242, 'timing/data': 0.0012724250555038452, 'timing/model': 1.4294927073642612, 'learning_rate/qwen_vl_interface': 8.661379176396651e-07, 'learning_rate/action_model': 8.66137917639665e-06, 'epoch': 2.36}) +[2026-06-18 07:25:13,919][starVLA.training.train_starvla][INFO] - Step 6133, Loss: {'train/loss': 0.289550244808197, 'train/grad_norm_pre_clip': 17.840065002441406, 'timing/data': 0.0015732115134596825, 'timing/model': 1.2034119823947549, 'learning_rate/qwen_vl_interface': 8.653057453063092e-07, 'learning_rate/action_model': 8.653057453063091e-06, 'epoch': 2.36}) +[2026-06-18 07:25:22,205][starVLA.training.train_starvla][INFO] - Step 6134, Loss: {'train/loss': 0.15384811162948608, 'train/grad_norm_pre_clip': 11.966085433959961, 'timing/data': 0.0008447002619504929, 'timing/model': 2.574068211019039, 'learning_rate/qwen_vl_interface': 8.644744819247908e-07, 'learning_rate/action_model': 8.644744819247908e-06, 'epoch': 2.36}) +[2026-06-18 07:25:27,712][starVLA.training.train_starvla][INFO] - Step 6135, Loss: {'train/loss': 0.1647057980298996, 'train/grad_norm_pre_clip': 13.829560279846191, 'timing/data': 0.0014771101996302605, 'timing/model': 1.2532889945432544, 'learning_rate/qwen_vl_interface': 8.636441276674308e-07, 'learning_rate/action_model': 8.636441276674308e-06, 'epoch': 2.36}) +[2026-06-18 07:25:41,422][starVLA.training.train_starvla][INFO] - Step 6136, Loss: {'train/loss': 0.20513014495372772, 'train/grad_norm_pre_clip': 18.13396644592285, 'timing/data': 0.0012123668566346169, 'timing/model': 5.388558853417635, 'learning_rate/qwen_vl_interface': 8.628146827063632e-07, 'learning_rate/action_model': 8.628146827063633e-06, 'epoch': 2.36}) +[2026-06-18 07:25:52,154][starVLA.training.train_starvla][INFO] - Step 6137, Loss: {'train/loss': 0.21865330636501312, 'train/grad_norm_pre_clip': 12.062156677246094, 'timing/data': 0.0013702195137739182, 'timing/model': 3.002209429629147, 'learning_rate/qwen_vl_interface': 8.619861472135332e-07, 'learning_rate/action_model': 8.619861472135332e-06, 'epoch': 2.36}) +[2026-06-18 07:26:07,312][starVLA.training.train_starvla][INFO] - Step 6138, Loss: {'train/loss': 0.24748383462429047, 'train/grad_norm_pre_clip': 13.790106773376465, 'timing/data': 0.001235567033290863, 'timing/model': 1.1658603055402637, 'learning_rate/qwen_vl_interface': 8.611585213606964e-07, 'learning_rate/action_model': 8.611585213606964e-06, 'epoch': 2.36}) +[2026-06-18 07:26:20,587][starVLA.training.train_starvla][INFO] - Step 6139, Loss: {'train/loss': 0.45939403772354126, 'train/grad_norm_pre_clip': 17.907245635986328, 'timing/data': 0.0013954732567071915, 'timing/model': 2.0319388257339597, 'learning_rate/qwen_vl_interface': 8.603318053194206e-07, 'learning_rate/action_model': 8.603318053194205e-06, 'epoch': 2.36}) +[2026-06-18 07:26:27,986][starVLA.training.train_starvla][INFO] - Step 6140, Loss: {'train/loss': 0.15514573454856873, 'train/grad_norm_pre_clip': 9.442333221435547, 'timing/data': 0.001112702302634716, 'timing/model': 1.7066268417984247, 'learning_rate/qwen_vl_interface': 8.595059992610846e-07, 'learning_rate/action_model': 8.595059992610846e-06, 'epoch': 2.36}) +[2026-06-18 07:26:32,962][starVLA.training.train_starvla][INFO] - Step 6141, Loss: {'train/loss': 0.4407200813293457, 'train/grad_norm_pre_clip': 19.18044090270996, 'timing/data': 0.0008488409221172333, 'timing/model': 1.2792094377800822, 'learning_rate/qwen_vl_interface': 8.586811033568796e-07, 'learning_rate/action_model': 8.586811033568797e-06, 'epoch': 2.36}) +[2026-06-18 07:26:38,028][starVLA.training.train_starvla][INFO] - Step 6142, Loss: {'train/loss': 0.3524981737136841, 'train/grad_norm_pre_clip': 19.221904754638672, 'timing/data': 0.0012076469138264656, 'timing/model': 1.4774856558069587, 'learning_rate/qwen_vl_interface': 8.578571177778076e-07, 'learning_rate/action_model': 8.578571177778075e-06, 'epoch': 2.36}) +[2026-06-18 07:26:42,870][starVLA.training.train_starvla][INFO] - Step 6143, Loss: {'train/loss': 0.15293708443641663, 'train/grad_norm_pre_clip': 11.00374698638916, 'timing/data': 0.0012063952162861824, 'timing/model': 1.1710990816354752, 'learning_rate/qwen_vl_interface': 8.570340426946805e-07, 'learning_rate/action_model': 8.570340426946804e-06, 'epoch': 2.36}) +[2026-06-18 07:26:47,650][starVLA.training.train_starvla][INFO] - Step 6144, Loss: {'train/loss': 0.3817704916000366, 'train/grad_norm_pre_clip': 13.471230506896973, 'timing/data': 0.0013464922085404396, 'timing/model': 1.1890634195879102, 'learning_rate/qwen_vl_interface': 8.562118782781241e-07, 'learning_rate/action_model': 8.56211878278124e-06, 'epoch': 2.36}) +[2026-06-18 07:26:53,111][starVLA.training.train_starvla][INFO] - Step 6145, Loss: {'train/loss': 0.2819070816040039, 'train/grad_norm_pre_clip': 16.6785945892334, 'timing/data': 0.001727985218167305, 'timing/model': 1.2101210448890924, 'learning_rate/qwen_vl_interface': 8.553906246985722e-07, 'learning_rate/action_model': 8.553906246985721e-06, 'epoch': 2.36}) +[2026-06-18 07:26:58,373][starVLA.training.train_starvla][INFO] - Step 6146, Loss: {'train/loss': 0.31352347135543823, 'train/grad_norm_pre_clip': 21.69593048095703, 'timing/data': 0.0011635376140475273, 'timing/model': 1.4019671231508255, 'learning_rate/qwen_vl_interface': 8.545702821262733e-07, 'learning_rate/action_model': 8.545702821262733e-06, 'epoch': 2.36}) +[2026-06-18 07:27:03,335][starVLA.training.train_starvla][INFO] - Step 6147, Loss: {'train/loss': 0.2648872137069702, 'train/grad_norm_pre_clip': 14.587666511535645, 'timing/data': 0.000838499516248703, 'timing/model': 1.5349106257781386, 'learning_rate/qwen_vl_interface': 8.537508507312844e-07, 'learning_rate/action_model': 8.537508507312844e-06, 'epoch': 2.36}) +[2026-06-18 07:27:07,979][starVLA.training.train_starvla][INFO] - Step 6148, Loss: {'train/loss': 0.5063579082489014, 'train/grad_norm_pre_clip': 17.5354061126709, 'timing/data': 0.0009871739894151688, 'timing/model': 1.1519523216411471, 'learning_rate/qwen_vl_interface': 8.52932330683476e-07, 'learning_rate/action_model': 8.529323306834759e-06, 'epoch': 2.37}) +[2026-06-18 07:27:13,365][starVLA.training.train_starvla][INFO] - Step 6149, Loss: {'train/loss': 0.33419129252433777, 'train/grad_norm_pre_clip': 13.717683792114258, 'timing/data': 0.0013245847076177597, 'timing/model': 1.1809780756011605, 'learning_rate/qwen_vl_interface': 8.52114722152525e-07, 'learning_rate/action_model': 8.521147221525249e-06, 'epoch': 2.37}) +[2026-06-18 07:27:18,575][starVLA.training.train_starvla][INFO] - Step 6150, Loss: {'train/loss': 0.3105045258998871, 'train/grad_norm_pre_clip': 18.02107048034668, 'timing/data': 0.0011119972914457321, 'timing/model': 1.1330754412338138, 'learning_rate/qwen_vl_interface': 8.51298025307925e-07, 'learning_rate/action_model': 8.51298025307925e-06, 'epoch': 2.37}) +[2026-06-18 07:27:23,581][starVLA.training.train_starvla][INFO] - Step 6151, Loss: {'train/loss': 0.18667851388454437, 'train/grad_norm_pre_clip': 12.236348152160645, 'timing/data': 0.0011426527053117752, 'timing/model': 1.5694994842633605, 'learning_rate/qwen_vl_interface': 8.504822403189783e-07, 'learning_rate/action_model': 8.504822403189782e-06, 'epoch': 2.37}) +[2026-06-18 07:27:28,781][starVLA.training.train_starvla][INFO] - Step 6152, Loss: {'train/loss': 0.24827036261558533, 'train/grad_norm_pre_clip': 13.155951499938965, 'timing/data': 0.0013471860438585281, 'timing/model': 1.7963532879948616, 'learning_rate/qwen_vl_interface': 8.496673673547967e-07, 'learning_rate/action_model': 8.496673673547968e-06, 'epoch': 2.37}) +[2026-06-18 07:27:33,438][starVLA.training.train_starvla][INFO] - Step 6153, Loss: {'train/loss': 0.2346169501543045, 'train/grad_norm_pre_clip': 13.923518180847168, 'timing/data': 0.001300819218158722, 'timing/model': 1.1103519899770617, 'learning_rate/qwen_vl_interface': 8.488534065843046e-07, 'learning_rate/action_model': 8.488534065843045e-06, 'epoch': 2.37}) +[2026-06-18 07:27:42,632][starVLA.training.train_starvla][INFO] - Step 6154, Loss: {'train/loss': 0.11721960455179214, 'train/grad_norm_pre_clip': 15.699740409851074, 'timing/data': 0.00140355434268713, 'timing/model': 5.4415108831599355, 'learning_rate/qwen_vl_interface': 8.480403581762378e-07, 'learning_rate/action_model': 8.480403581762378e-06, 'epoch': 2.37}) +[2026-06-18 07:27:51,820][starVLA.training.train_starvla][INFO] - Step 6155, Loss: {'train/loss': 0.4089927077293396, 'train/grad_norm_pre_clip': 22.002443313598633, 'timing/data': 0.0013285186141729355, 'timing/model': 1.912539173848927, 'learning_rate/qwen_vl_interface': 8.472282222991411e-07, 'learning_rate/action_model': 8.47228222299141e-06, 'epoch': 2.37}) +[2026-06-18 07:28:00,396][starVLA.training.train_starvla][INFO] - Step 6156, Loss: {'train/loss': 0.4498998522758484, 'train/grad_norm_pre_clip': 23.663475036621094, 'timing/data': 0.006574471481144428, 'timing/model': 4.387206686660647, 'learning_rate/qwen_vl_interface': 8.464169991213705e-07, 'learning_rate/action_model': 8.464169991213706e-06, 'epoch': 2.37}) +[2026-06-18 07:28:15,734][starVLA.training.train_starvla][INFO] - Step 6157, Loss: {'train/loss': 0.20362558960914612, 'train/grad_norm_pre_clip': 14.513456344604492, 'timing/data': 0.008491561748087406, 'timing/model': 5.023930201306939, 'learning_rate/qwen_vl_interface': 8.456066888110964e-07, 'learning_rate/action_model': 8.456066888110964e-06, 'epoch': 2.37}) +[2026-06-18 07:28:25,043][starVLA.training.train_starvla][INFO] - Step 6158, Loss: {'train/loss': 0.26623883843421936, 'train/grad_norm_pre_clip': 16.352386474609375, 'timing/data': 0.0013002809137105942, 'timing/model': 1.3045683410018682, 'learning_rate/qwen_vl_interface': 8.447972915362941e-07, 'learning_rate/action_model': 8.44797291536294e-06, 'epoch': 2.37}) +[2026-06-18 07:28:32,328][starVLA.training.train_starvla][INFO] - Step 6159, Loss: {'train/loss': 0.24256780743598938, 'train/grad_norm_pre_clip': 19.63078498840332, 'timing/data': 0.0012600421905517578, 'timing/model': 1.4649502532556653, 'learning_rate/qwen_vl_interface': 8.439888074647536e-07, 'learning_rate/action_model': 8.439888074647536e-06, 'epoch': 2.37}) +[2026-06-18 07:28:45,891][starVLA.training.train_starvla][INFO] - Step 6160, Loss: {'train/loss': 0.44733840227127075, 'train/grad_norm_pre_clip': 17.887548446655273, 'timing/data': 0.001863742247223854, 'timing/model': 1.380411647260189, 'learning_rate/qwen_vl_interface': 8.431812367640732e-07, 'learning_rate/action_model': 8.431812367640731e-06, 'epoch': 2.37}) +[2026-06-18 07:29:00,275][starVLA.training.train_starvla][INFO] - Step 6161, Loss: {'train/loss': 0.39375039935112, 'train/grad_norm_pre_clip': 19.270252227783203, 'timing/data': 0.0018122131004929543, 'timing/model': 3.460448794066906, 'learning_rate/qwen_vl_interface': 8.423745796016659e-07, 'learning_rate/action_model': 8.42374579601666e-06, 'epoch': 2.37}) +[2026-06-18 07:29:12,050][starVLA.training.train_starvla][INFO] - Step 6162, Loss: {'train/loss': 0.49593231081962585, 'train/grad_norm_pre_clip': 18.981361389160156, 'timing/data': 0.0009165974333882332, 'timing/model': 1.8977509904652834, 'learning_rate/qwen_vl_interface': 8.41568836144751e-07, 'learning_rate/action_model': 8.41568836144751e-06, 'epoch': 2.37}) +[2026-06-18 07:29:26,208][starVLA.training.train_starvla][INFO] - Step 6163, Loss: {'train/loss': 0.4277634620666504, 'train/grad_norm_pre_clip': 17.110881805419922, 'timing/data': 0.005306919105350971, 'timing/model': 3.1934076994657516, 'learning_rate/qwen_vl_interface': 8.407640065603601e-07, 'learning_rate/action_model': 8.4076400656036e-06, 'epoch': 2.37}) +[2026-06-18 07:29:36,016][starVLA.training.train_starvla][INFO] - Step 6164, Loss: {'train/loss': 0.416629433631897, 'train/grad_norm_pre_clip': 14.133161544799805, 'timing/data': 0.0009545320644974709, 'timing/model': 1.228422592394054, 'learning_rate/qwen_vl_interface': 8.399600910153336e-07, 'learning_rate/action_model': 8.399600910153336e-06, 'epoch': 2.37}) +[2026-06-18 07:29:41,778][starVLA.training.train_starvla][INFO] - Step 6165, Loss: {'train/loss': 0.11871323734521866, 'train/grad_norm_pre_clip': 9.642995834350586, 'timing/data': 0.0010023564100265503, 'timing/model': 1.385832641273737, 'learning_rate/qwen_vl_interface': 8.391570896763266e-07, 'learning_rate/action_model': 8.391570896763267e-06, 'epoch': 2.37}) +[2026-06-18 07:29:47,324][starVLA.training.train_starvla][INFO] - Step 6166, Loss: {'train/loss': 0.3646891415119171, 'train/grad_norm_pre_clip': 14.972029685974121, 'timing/data': 0.0012823017314076424, 'timing/model': 1.4234275752678514, 'learning_rate/qwen_vl_interface': 8.38355002709802e-07, 'learning_rate/action_model': 8.38355002709802e-06, 'epoch': 2.37}) +[2026-06-18 07:29:53,338][starVLA.training.train_starvla][INFO] - Step 6167, Loss: {'train/loss': 0.3503628373146057, 'train/grad_norm_pre_clip': 12.261900901794434, 'timing/data': 0.0009270105510950089, 'timing/model': 1.7557273395359516, 'learning_rate/qwen_vl_interface': 8.375538302820314e-07, 'learning_rate/action_model': 8.375538302820315e-06, 'epoch': 2.37}) +[2026-06-18 07:29:58,353][starVLA.training.train_starvla][INFO] - Step 6168, Loss: {'train/loss': 0.29919520020484924, 'train/grad_norm_pre_clip': 16.316268920898438, 'timing/data': 0.0012666601687669754, 'timing/model': 1.2852916838601232, 'learning_rate/qwen_vl_interface': 8.367535725591003e-07, 'learning_rate/action_model': 8.367535725591003e-06, 'epoch': 2.37}) +[2026-06-18 07:30:04,064][starVLA.training.train_starvla][INFO] - Step 6169, Loss: {'train/loss': 0.22505821287631989, 'train/grad_norm_pre_clip': 11.967743873596191, 'timing/data': 0.0014385273680090904, 'timing/model': 1.1084904205054045, 'learning_rate/qwen_vl_interface': 8.359542297069024e-07, 'learning_rate/action_model': 8.359542297069024e-06, 'epoch': 2.37}) +[2026-06-18 07:30:10,036][starVLA.training.train_starvla][INFO] - Step 6170, Loss: {'train/loss': 0.43283969163894653, 'train/grad_norm_pre_clip': 14.745049476623535, 'timing/data': 0.0009322548285126686, 'timing/model': 1.5270252646878362, 'learning_rate/qwen_vl_interface': 8.351558018911423e-07, 'learning_rate/action_model': 8.351558018911423e-06, 'epoch': 2.37}) +[2026-06-18 07:30:15,977][starVLA.training.train_starvla][INFO] - Step 6171, Loss: {'train/loss': 0.3669334352016449, 'train/grad_norm_pre_clip': 18.090856552124023, 'timing/data': 0.001226421445608139, 'timing/model': 1.7493422348052263, 'learning_rate/qwen_vl_interface': 8.343582892773352e-07, 'learning_rate/action_model': 8.343582892773351e-06, 'epoch': 2.37}) +[2026-06-18 07:30:21,121][starVLA.training.train_starvla][INFO] - Step 6172, Loss: {'train/loss': 0.4037834703922272, 'train/grad_norm_pre_clip': 20.423736572265625, 'timing/data': 0.0011471342295408249, 'timing/model': 1.7550596799701452, 'learning_rate/qwen_vl_interface': 8.335616920308061e-07, 'learning_rate/action_model': 8.33561692030806e-06, 'epoch': 2.37}) +[2026-06-18 07:30:33,947][starVLA.training.train_starvla][INFO] - Step 6173, Loss: {'train/loss': 0.6219685673713684, 'train/grad_norm_pre_clip': 12.683401107788086, 'timing/data': 0.0009993473067879677, 'timing/model': 1.1115737576037645, 'learning_rate/qwen_vl_interface': 8.327660103166907e-07, 'learning_rate/action_model': 8.327660103166906e-06, 'epoch': 2.37}) +[2026-06-18 07:30:42,447][starVLA.training.train_starvla][INFO] - Step 6174, Loss: {'train/loss': 0.4302828013896942, 'train/grad_norm_pre_clip': 20.727943420410156, 'timing/data': 0.0011646486818790436, 'timing/model': 1.2810249840840697, 'learning_rate/qwen_vl_interface': 8.319712442999335e-07, 'learning_rate/action_model': 8.319712442999335e-06, 'epoch': 2.38}) +[2026-06-18 07:30:47,620][starVLA.training.train_starvla][INFO] - Step 6175, Loss: {'train/loss': 0.38991987705230713, 'train/grad_norm_pre_clip': 13.554444313049316, 'timing/data': 0.014333377592265606, 'timing/model': 1.098549255169928, 'learning_rate/qwen_vl_interface': 8.311773941452924e-07, 'learning_rate/action_model': 8.311773941452925e-06, 'epoch': 2.38}) +[2026-06-18 07:30:52,778][starVLA.training.train_starvla][INFO] - Step 6176, Loss: {'train/loss': 0.2645924985408783, 'train/grad_norm_pre_clip': 13.000133514404297, 'timing/data': 0.0009135864675045013, 'timing/model': 1.1373397102579474, 'learning_rate/qwen_vl_interface': 8.303844600173326e-07, 'learning_rate/action_model': 8.303844600173325e-06, 'epoch': 2.38}) +[2026-06-18 07:31:00,714][starVLA.training.train_starvla][INFO] - Step 6177, Loss: {'train/loss': 0.2456829845905304, 'train/grad_norm_pre_clip': 12.532624244689941, 'timing/data': 0.007125326432287693, 'timing/model': 1.696332030929625, 'learning_rate/qwen_vl_interface': 8.295924420804297e-07, 'learning_rate/action_model': 8.295924420804296e-06, 'epoch': 2.38}) +[2026-06-18 07:31:09,534][starVLA.training.train_starvla][INFO] - Step 6178, Loss: {'train/loss': 0.2283318191766739, 'train/grad_norm_pre_clip': 13.08460807800293, 'timing/data': 0.0027786651626229286, 'timing/model': 1.3018212746828794, 'learning_rate/qwen_vl_interface': 8.288013404987691e-07, 'learning_rate/action_model': 8.28801340498769e-06, 'epoch': 2.38}) +[2026-06-18 07:31:15,104][starVLA.training.train_starvla][INFO] - Step 6179, Loss: {'train/loss': 0.6161190867424011, 'train/grad_norm_pre_clip': 14.924202919006348, 'timing/data': 0.0020447522401809692, 'timing/model': 1.0867197653278708, 'learning_rate/qwen_vl_interface': 8.28011155436349e-07, 'learning_rate/action_model': 8.28011155436349e-06, 'epoch': 2.38}) +[2026-06-18 07:31:20,027][starVLA.training.train_starvla][INFO] - Step 6180, Loss: {'train/loss': 0.24370376765727997, 'train/grad_norm_pre_clip': 13.249988555908203, 'timing/data': 0.000961795449256897, 'timing/model': 1.1810778006911278, 'learning_rate/qwen_vl_interface': 8.272218870569748e-07, 'learning_rate/action_model': 8.272218870569748e-06, 'epoch': 2.38}) +[2026-06-18 07:31:25,471][starVLA.training.train_starvla][INFO] - Step 6181, Loss: {'train/loss': 0.2552127242088318, 'train/grad_norm_pre_clip': 15.801520347595215, 'timing/data': 0.0011137435212731361, 'timing/model': 1.19013237580657, 'learning_rate/qwen_vl_interface': 8.264335355242625e-07, 'learning_rate/action_model': 8.264335355242625e-06, 'epoch': 2.38}) +[2026-06-18 07:31:30,178][starVLA.training.train_starvla][INFO] - Step 6182, Loss: {'train/loss': 0.5313648581504822, 'train/grad_norm_pre_clip': 22.88935661315918, 'timing/data': 0.0013575898483395576, 'timing/model': 1.495236693881452, 'learning_rate/qwen_vl_interface': 8.256461010016376e-07, 'learning_rate/action_model': 8.256461010016376e-06, 'epoch': 2.38}) +[2026-06-18 07:31:35,982][starVLA.training.train_starvla][INFO] - Step 6183, Loss: {'train/loss': 0.5650642514228821, 'train/grad_norm_pre_clip': 22.302278518676758, 'timing/data': 0.0012254351750016212, 'timing/model': 1.2077393010258675, 'learning_rate/qwen_vl_interface': 8.248595836523368e-07, 'learning_rate/action_model': 8.248595836523368e-06, 'epoch': 2.38}) +[2026-06-18 07:31:42,270][starVLA.training.train_starvla][INFO] - Step 6184, Loss: {'train/loss': 0.6861560344696045, 'train/grad_norm_pre_clip': 22.179738998413086, 'timing/data': 0.0010523851960897446, 'timing/model': 1.132291086949408, 'learning_rate/qwen_vl_interface': 8.240739836394068e-07, 'learning_rate/action_model': 8.240739836394068e-06, 'epoch': 2.38}) +[2026-06-18 07:31:47,567][starVLA.training.train_starvla][INFO] - Step 6185, Loss: {'train/loss': 0.20038291811943054, 'train/grad_norm_pre_clip': 13.5084810256958, 'timing/data': 0.0012357356026768684, 'timing/model': 1.1728873969987035, 'learning_rate/qwen_vl_interface': 8.232893011257019e-07, 'learning_rate/action_model': 8.232893011257018e-06, 'epoch': 2.38}) +[2026-06-18 07:31:56,466][starVLA.training.train_starvla][INFO] - Step 6186, Loss: {'train/loss': 0.3908330500125885, 'train/grad_norm_pre_clip': 13.804102897644043, 'timing/data': 0.0009190291166305542, 'timing/model': 1.1425711438059807, 'learning_rate/qwen_vl_interface': 8.225055362738886e-07, 'learning_rate/action_model': 8.225055362738885e-06, 'epoch': 2.38}) +[2026-06-18 07:32:01,788][starVLA.training.train_starvla][INFO] - Step 6187, Loss: {'train/loss': 0.20084582269191742, 'train/grad_norm_pre_clip': 9.84225082397461, 'timing/data': 0.0018554935231804848, 'timing/model': 1.7486487505957484, 'learning_rate/qwen_vl_interface': 8.217226892464411e-07, 'learning_rate/action_model': 8.21722689246441e-06, 'epoch': 2.38}) +[2026-06-18 07:32:06,537][starVLA.training.train_starvla][INFO] - Step 6188, Loss: {'train/loss': 0.2607674300670624, 'train/grad_norm_pre_clip': 18.41242027282715, 'timing/data': 0.0009309062734246254, 'timing/model': 1.267454031854868, 'learning_rate/qwen_vl_interface': 8.209407602056457e-07, 'learning_rate/action_model': 8.209407602056457e-06, 'epoch': 2.38}) +[2026-06-18 07:32:15,949][starVLA.training.train_starvla][INFO] - Step 6189, Loss: {'train/loss': 0.3024621307849884, 'train/grad_norm_pre_clip': 11.08507251739502, 'timing/data': 0.0013599786907434464, 'timing/model': 1.997064022347331, 'learning_rate/qwen_vl_interface': 8.201597493135962e-07, 'learning_rate/action_model': 8.201597493135961e-06, 'epoch': 2.38}) +[2026-06-18 07:32:21,514][starVLA.training.train_starvla][INFO] - Step 6190, Loss: {'train/loss': 0.262071430683136, 'train/grad_norm_pre_clip': 14.185672760009766, 'timing/data': 0.0008239550516009331, 'timing/model': 1.2276960285380483, 'learning_rate/qwen_vl_interface': 8.193796567321982e-07, 'learning_rate/action_model': 8.193796567321982e-06, 'epoch': 2.38}) +[2026-06-18 07:32:26,641][starVLA.training.train_starvla][INFO] - Step 6191, Loss: {'train/loss': 0.18576563894748688, 'train/grad_norm_pre_clip': 10.715356826782227, 'timing/data': 0.012807468883693218, 'timing/model': 1.3728316519409418, 'learning_rate/qwen_vl_interface': 8.186004826231644e-07, 'learning_rate/action_model': 8.186004826231644e-06, 'epoch': 2.38}) +[2026-06-18 07:32:31,797][starVLA.training.train_starvla][INFO] - Step 6192, Loss: {'train/loss': 0.4048575758934021, 'train/grad_norm_pre_clip': 18.4051570892334, 'timing/data': 0.015098733827471733, 'timing/model': 1.726720786653459, 'learning_rate/qwen_vl_interface': 8.178222271480192e-07, 'learning_rate/action_model': 8.178222271480191e-06, 'epoch': 2.38}) +[2026-06-18 07:32:43,265][starVLA.training.train_starvla][INFO] - Step 6193, Loss: {'train/loss': 0.30968570709228516, 'train/grad_norm_pre_clip': 20.97774314880371, 'timing/data': 0.01896011922508478, 'timing/model': 5.109805233776569, 'learning_rate/qwen_vl_interface': 8.170448904680943e-07, 'learning_rate/action_model': 8.170448904680943e-06, 'epoch': 2.38}) +[2026-06-18 07:32:49,089][starVLA.training.train_starvla][INFO] - Step 6194, Loss: {'train/loss': 0.44307994842529297, 'train/grad_norm_pre_clip': 13.943424224853516, 'timing/data': 0.01133708842098713, 'timing/model': 1.3228246718645096, 'learning_rate/qwen_vl_interface': 8.162684727445347e-07, 'learning_rate/action_model': 8.162684727445346e-06, 'epoch': 2.38}) +[2026-06-18 07:33:06,053][starVLA.training.train_starvla][INFO] - Step 6195, Loss: {'train/loss': 0.1278865784406662, 'train/grad_norm_pre_clip': 8.633513450622559, 'timing/data': 0.022672351449728012, 'timing/model': 4.7280474081635475, 'learning_rate/qwen_vl_interface': 8.154929741382922e-07, 'learning_rate/action_model': 8.154929741382922e-06, 'epoch': 2.38}) +[2026-06-18 07:33:12,573][starVLA.training.train_starvla][INFO] - Step 6196, Loss: {'train/loss': 0.3517618179321289, 'train/grad_norm_pre_clip': 17.02672004699707, 'timing/data': 0.014970912598073483, 'timing/model': 1.7839495949447155, 'learning_rate/qwen_vl_interface': 8.147183948101266e-07, 'learning_rate/action_model': 8.147183948101265e-06, 'epoch': 2.38}) +[2026-06-18 07:33:17,643][starVLA.training.train_starvla][INFO] - Step 6197, Loss: {'train/loss': 0.2595195174217224, 'train/grad_norm_pre_clip': 12.083009719848633, 'timing/data': 0.013609478250145912, 'timing/model': 1.1146160429343581, 'learning_rate/qwen_vl_interface': 8.139447349206111e-07, 'learning_rate/action_model': 8.139447349206111e-06, 'epoch': 2.38}) +[2026-06-18 07:33:27,646][starVLA.training.train_starvla][INFO] - Step 6198, Loss: {'train/loss': 0.2881814241409302, 'train/grad_norm_pre_clip': 11.831787109375, 'timing/data': 0.014553806744515896, 'timing/model': 4.376789910718799, 'learning_rate/qwen_vl_interface': 8.131719946301255e-07, 'learning_rate/action_model': 8.131719946301255e-06, 'epoch': 2.38}) +[2026-06-18 07:33:33,599][starVLA.training.train_starvla][INFO] - Step 6199, Loss: {'train/loss': 0.3986097276210785, 'train/grad_norm_pre_clip': 15.356822967529297, 'timing/data': 0.01700405403971672, 'timing/model': 1.1792188696563244, 'learning_rate/qwen_vl_interface': 8.124001740988595e-07, 'learning_rate/action_model': 8.124001740988595e-06, 'epoch': 2.38}) +[2026-06-18 07:37:40,666][starVLA.training.train_starvla][INFO] - Step 6200, Loss: {'train/loss': 0.3285382390022278, 'train/grad_norm_pre_clip': 21.75212287902832, 'eval/loss': 0.42319339513778687, 'eval/latency_4/loss': 0.42319339513778687, 'eval/demon_fire_precision': 0.9685998077539251, 'eval/demon_fire_recall': 0.9818122767132186, 'eval/demon_fire_f1': 0.9751612903225807, 'eval/demon_fire_rate_model': 0.9165932452276064, 'eval/demon_fire_rate_teacher': 0.9042584434654919, 'eval/demon_left_precision': 0.7505773672055427, 'eval/demon_left_recall': 0.7420091324200914, 'eval/demon_left_f1': 0.7462686567164181, 'eval/demon_left_rate_model': 0.12716593245227606, 'eval/demon_left_rate_teacher': 0.12863436123348018, 'eval/demon_right_precision': 0.7684391080617495, 'eval/demon_right_recall': 0.8086642599277978, 'eval/demon_right_f1': 0.7880386983289357, 'eval/demon_right_rate_model': 0.17121879588839942, 'eval/demon_right_rate_teacher': 0.16270190895741557, 'eval/demon_move_f1': 0.7671536775226768, 'eval/demon_cc_f1': 0.9018060322535473, 'eval/latency_4/demon_fire_precision': 0.9685998077539251, 'eval/latency_4/demon_fire_recall': 0.9818122767132186, 'eval/latency_4/demon_fire_f1': 0.9751612903225807, 'eval/latency_4/demon_fire_rate_model': 0.9165932452276064, 'eval/latency_4/demon_fire_rate_teacher': 0.9042584434654919, 'eval/latency_4/demon_left_precision': 0.7505773672055427, 'eval/latency_4/demon_left_recall': 0.7420091324200914, 'eval/latency_4/demon_left_f1': 0.7462686567164181, 'eval/latency_4/demon_left_rate_model': 0.12716593245227606, 'eval/latency_4/demon_left_rate_teacher': 0.12863436123348018, 'eval/latency_4/demon_right_precision': 0.7684391080617495, 'eval/latency_4/demon_right_recall': 0.8086642599277978, 'eval/latency_4/demon_right_f1': 0.7880386983289357, 'eval/latency_4/demon_right_rate_model': 0.17121879588839942, 'eval/latency_4/demon_right_rate_teacher': 0.16270190895741557, 'eval/latency_4/demon_move_f1': 0.7671536775226768, 'eval/latency_4/demon_cc_f1': 0.9018060322535473, 'timing/data': 0.013696854934096336, 'timing/model': 1.3519174112007022, 'learning_rate/qwen_vl_interface': 8.116292734868126e-07, 'learning_rate/action_model': 8.116292734868125e-06, 'epoch': 2.39}) +[2026-06-18 07:37:40,702][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_6200_state +[2026-06-18 07:37:40,714][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 07:37:49,661][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_6200_state/model.safetensors +[2026-06-18 07:38:13,732][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_6200_state/optimizer.bin +[2026-06-18 07:38:13,733][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_6200_state/sampler.bin +[2026-06-18 07:38:13,734][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_6200_state/sampler_1.bin +[2026-06-18 07:38:13,737][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_6200_state/random_states_0.pkl +[2026-06-18 07:38:13,743][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 6200 +[2026-06-18 07:38:13,744][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 07:38:13,761][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 07:38:25,116][starVLA.training.train_starvla][INFO] - Step 6201, Loss: {'train/loss': 0.3990200161933899, 'train/grad_norm_pre_clip': 19.50375747680664, 'timing/data': 0.011665496975183487, 'timing/model': 1.824894237332046, 'learning_rate/qwen_vl_interface': 8.108592929537928e-07, 'learning_rate/action_model': 8.108592929537928e-06, 'epoch': 2.39}) +[2026-06-18 07:38:30,224][starVLA.training.train_starvla][INFO] - Step 6202, Loss: {'train/loss': 0.11381252855062485, 'train/grad_norm_pre_clip': 8.885995864868164, 'timing/data': 0.01467453595250845, 'timing/model': 1.2680648183450103, 'learning_rate/qwen_vl_interface': 8.100902326594186e-07, 'learning_rate/action_model': 8.100902326594186e-06, 'epoch': 2.39}) +[2026-06-18 07:38:35,359][starVLA.training.train_starvla][INFO] - Step 6203, Loss: {'train/loss': 0.1974383443593979, 'train/grad_norm_pre_clip': 15.242055892944336, 'timing/data': 0.014837938360869884, 'timing/model': 1.1330886129289865, 'learning_rate/qwen_vl_interface': 8.093220927631171e-07, 'learning_rate/action_model': 8.09322092763117e-06, 'epoch': 2.39}) +[2026-06-18 07:38:40,234][starVLA.training.train_starvla][INFO] - Step 6204, Loss: {'train/loss': 0.20904646813869476, 'train/grad_norm_pre_clip': 12.131973266601562, 'timing/data': 0.024502701126039028, 'timing/model': 1.1060121478512883, 'learning_rate/qwen_vl_interface': 8.085548734241233e-07, 'learning_rate/action_model': 8.085548734241232e-06, 'epoch': 2.39}) +[2026-06-18 07:38:45,403][starVLA.training.train_starvla][INFO] - Step 6205, Loss: {'train/loss': 0.2984682619571686, 'train/grad_norm_pre_clip': 17.845003128051758, 'timing/data': 0.011927803047001362, 'timing/model': 1.238301133736968, 'learning_rate/qwen_vl_interface': 8.077885748014843e-07, 'learning_rate/action_model': 8.077885748014841e-06, 'epoch': 2.39}) +[2026-06-18 07:38:52,149][starVLA.training.train_starvla][INFO] - Step 6206, Loss: {'train/loss': 0.17610174417495728, 'train/grad_norm_pre_clip': 11.600373268127441, 'timing/data': 0.01380942016839981, 'timing/model': 1.4674623552709818, 'learning_rate/qwen_vl_interface': 8.070231970540532e-07, 'learning_rate/action_model': 8.070231970540532e-06, 'epoch': 2.39}) +[2026-06-18 07:38:56,955][starVLA.training.train_starvla][INFO] - Step 6207, Loss: {'train/loss': 0.44471752643585205, 'train/grad_norm_pre_clip': 16.058992385864258, 'timing/data': 0.016586272977292538, 'timing/model': 1.1277820905670524, 'learning_rate/qwen_vl_interface': 8.062587403404934e-07, 'learning_rate/action_model': 8.062587403404934e-06, 'epoch': 2.39}) +[2026-06-18 07:39:02,385][starVLA.training.train_starvla][INFO] - Step 6208, Loss: {'train/loss': 0.23113463819026947, 'train/grad_norm_pre_clip': 20.386503219604492, 'timing/data': 0.014123788103461266, 'timing/model': 1.1900444636121392, 'learning_rate/qwen_vl_interface': 8.054952048192793e-07, 'learning_rate/action_model': 8.054952048192794e-06, 'epoch': 2.39}) +[2026-06-18 07:39:08,293][starVLA.training.train_starvla][INFO] - Step 6209, Loss: {'train/loss': 0.20840200781822205, 'train/grad_norm_pre_clip': 16.499679565429688, 'timing/data': 0.018065414391458035, 'timing/model': 1.4284070814028382, 'learning_rate/qwen_vl_interface': 8.047325906486926e-07, 'learning_rate/action_model': 8.047325906486925e-06, 'epoch': 2.39}) +[2026-06-18 07:39:13,524][starVLA.training.train_starvla][INFO] - Step 6210, Loss: {'train/loss': 0.3913254141807556, 'train/grad_norm_pre_clip': 13.431717872619629, 'timing/data': 0.01634489744901657, 'timing/model': 1.0970431054010987, 'learning_rate/qwen_vl_interface': 8.039708979868229e-07, 'learning_rate/action_model': 8.039708979868227e-06, 'epoch': 2.39}) +[2026-06-18 07:39:25,225][starVLA.training.train_starvla][INFO] - Step 6211, Loss: {'train/loss': 0.2686332166194916, 'train/grad_norm_pre_clip': 11.842935562133789, 'timing/data': 0.011605402454733849, 'timing/model': 1.3613962391391397, 'learning_rate/qwen_vl_interface': 8.032101269915687e-07, 'learning_rate/action_model': 8.032101269915687e-06, 'epoch': 2.39}) +[2026-06-18 07:39:30,687][starVLA.training.train_starvla][INFO] - Step 6212, Loss: {'train/loss': 0.3689488172531128, 'train/grad_norm_pre_clip': 14.311119079589844, 'timing/data': 0.022854631766676903, 'timing/model': 1.557615466415882, 'learning_rate/qwen_vl_interface': 8.024502778206411e-07, 'learning_rate/action_model': 8.02450277820641e-06, 'epoch': 2.39}) +[2026-06-18 07:39:35,788][starVLA.training.train_starvla][INFO] - Step 6213, Loss: {'train/loss': 0.315425306558609, 'train/grad_norm_pre_clip': 19.794191360473633, 'timing/data': 0.01183208916336298, 'timing/model': 1.503629107028246, 'learning_rate/qwen_vl_interface': 8.016913506315566e-07, 'learning_rate/action_model': 8.016913506315565e-06, 'epoch': 2.39}) +[2026-06-18 07:39:42,980][starVLA.training.train_starvla][INFO] - Step 6214, Loss: {'train/loss': 0.3822721838951111, 'train/grad_norm_pre_clip': 18.502187728881836, 'timing/data': 0.01490634772926569, 'timing/model': 2.2025618217885494, 'learning_rate/qwen_vl_interface': 8.009333455816416e-07, 'learning_rate/action_model': 8.009333455816417e-06, 'epoch': 2.39}) +[2026-06-18 07:39:48,269][starVLA.training.train_starvla][INFO] - Step 6215, Loss: {'train/loss': 0.13971973955631256, 'train/grad_norm_pre_clip': 10.627010345458984, 'timing/data': 0.017360761761665344, 'timing/model': 1.1514243334531784, 'learning_rate/qwen_vl_interface': 8.001762628280318e-07, 'learning_rate/action_model': 8.001762628280319e-06, 'epoch': 2.39}) +[2026-06-18 07:39:53,654][starVLA.training.train_starvla][INFO] - Step 6216, Loss: {'train/loss': 0.20301687717437744, 'train/grad_norm_pre_clip': 14.259281158447266, 'timing/data': 0.021490395069122314, 'timing/model': 1.3811361901462078, 'learning_rate/qwen_vl_interface': 7.994201025276708e-07, 'learning_rate/action_model': 7.994201025276708e-06, 'epoch': 2.39}) +[2026-06-18 07:39:58,278][starVLA.training.train_starvla][INFO] - Step 6217, Loss: {'train/loss': 0.30436357855796814, 'train/grad_norm_pre_clip': 24.66877555847168, 'timing/data': 0.012380083091557026, 'timing/model': 1.121363283134997, 'learning_rate/qwen_vl_interface': 7.986648648373115e-07, 'learning_rate/action_model': 7.986648648373115e-06, 'epoch': 2.39}) +[2026-06-18 07:40:03,417][starVLA.training.train_starvla][INFO] - Step 6218, Loss: {'train/loss': 0.435577392578125, 'train/grad_norm_pre_clip': 12.739123344421387, 'timing/data': 0.01199703011661768, 'timing/model': 1.2105549136176705, 'learning_rate/qwen_vl_interface': 7.979105499135157e-07, 'learning_rate/action_model': 7.979105499135155e-06, 'epoch': 2.39}) +[2026-06-18 07:40:08,618][starVLA.training.train_starvla][INFO] - Step 6219, Loss: {'train/loss': 0.37300872802734375, 'train/grad_norm_pre_clip': 16.62495994567871, 'timing/data': 0.02922012098133564, 'timing/model': 1.3891060333698988, 'learning_rate/qwen_vl_interface': 7.971571579126533e-07, 'learning_rate/action_model': 7.971571579126534e-06, 'epoch': 2.39}) +[2026-06-18 07:40:14,831][starVLA.training.train_starvla][INFO] - Step 6220, Loss: {'train/loss': 0.3273738920688629, 'train/grad_norm_pre_clip': 18.824676513671875, 'timing/data': 0.012378043495118618, 'timing/model': 1.09490477014333, 'learning_rate/qwen_vl_interface': 7.964046889909038e-07, 'learning_rate/action_model': 7.964046889909038e-06, 'epoch': 2.39}) +[2026-06-18 07:40:20,133][starVLA.training.train_starvla][INFO] - Step 6221, Loss: {'train/loss': 0.4271523654460907, 'train/grad_norm_pre_clip': 18.406307220458984, 'timing/data': 0.01537272334098816, 'timing/model': 1.5908563947305083, 'learning_rate/qwen_vl_interface': 7.956531433042543e-07, 'learning_rate/action_model': 7.956531433042542e-06, 'epoch': 2.39}) +[2026-06-18 07:40:26,478][starVLA.training.train_starvla][INFO] - Step 6222, Loss: {'train/loss': 0.2331882268190384, 'train/grad_norm_pre_clip': 15.535704612731934, 'timing/data': 0.010974419303238392, 'timing/model': 1.4251975966617465, 'learning_rate/qwen_vl_interface': 7.94902521008501e-07, 'learning_rate/action_model': 7.94902521008501e-06, 'epoch': 2.39}) +[2026-06-18 07:40:42,022][starVLA.training.train_starvla][INFO] - Step 6223, Loss: {'train/loss': 0.2561044692993164, 'train/grad_norm_pre_clip': 13.495783805847168, 'timing/data': 0.012133021838963032, 'timing/model': 4.373522212728858, 'learning_rate/qwen_vl_interface': 7.941528222592501e-07, 'learning_rate/action_model': 7.941528222592501e-06, 'epoch': 2.39}) +[2026-06-18 07:40:54,596][starVLA.training.train_starvla][INFO] - Step 6224, Loss: {'train/loss': 0.2964596152305603, 'train/grad_norm_pre_clip': 14.221199035644531, 'timing/data': 0.014065899886190891, 'timing/model': 4.270254367031157, 'learning_rate/qwen_vl_interface': 7.934040472119132e-07, 'learning_rate/action_model': 7.934040472119133e-06, 'epoch': 2.39}) +[2026-06-18 07:41:03,570][starVLA.training.train_starvla][INFO] - Step 6225, Loss: {'train/loss': 0.31010133028030396, 'train/grad_norm_pre_clip': 14.421151161193848, 'timing/data': 0.011247455142438412, 'timing/model': 1.2983785569667816, 'learning_rate/qwen_vl_interface': 7.926561960217127e-07, 'learning_rate/action_model': 7.926561960217126e-06, 'epoch': 2.39}) +[2026-06-18 07:41:20,222][starVLA.training.train_starvla][INFO] - Step 6226, Loss: {'train/loss': 0.23034042119979858, 'train/grad_norm_pre_clip': 11.076273918151855, 'timing/data': 0.018924487754702568, 'timing/model': 4.3287418046966195, 'learning_rate/qwen_vl_interface': 7.919092688436783e-07, 'learning_rate/action_model': 7.919092688436783e-06, 'epoch': 2.4}) +[2026-06-18 07:41:30,844][starVLA.training.train_starvla][INFO] - Step 6227, Loss: {'train/loss': 0.08879531174898148, 'train/grad_norm_pre_clip': 5.995895862579346, 'timing/data': 0.022020812146365643, 'timing/model': 2.1106285164132714, 'learning_rate/qwen_vl_interface': 7.911632658326505e-07, 'learning_rate/action_model': 7.911632658326505e-06, 'epoch': 2.4}) +[2026-06-18 07:41:46,188][starVLA.training.train_starvla][INFO] - Step 6228, Loss: {'train/loss': 0.6600892543792725, 'train/grad_norm_pre_clip': 21.927396774291992, 'timing/data': 0.0010966798290610313, 'timing/model': 3.3822875805199146, 'learning_rate/qwen_vl_interface': 7.904181871432757e-07, 'learning_rate/action_model': 7.904181871432757e-06, 'epoch': 2.4}) +[2026-06-18 07:41:54,021][starVLA.training.train_starvla][INFO] - Step 6229, Loss: {'train/loss': 0.19406075775623322, 'train/grad_norm_pre_clip': 12.462536811828613, 'timing/data': 0.00433886144310236, 'timing/model': 1.0970347132533789, 'learning_rate/qwen_vl_interface': 7.896740329300091e-07, 'learning_rate/action_model': 7.89674032930009e-06, 'epoch': 2.4}) +[2026-06-18 07:41:59,041][starVLA.training.train_starvla][INFO] - Step 6230, Loss: {'train/loss': 0.3915429413318634, 'train/grad_norm_pre_clip': 16.086034774780273, 'timing/data': 0.0011614281684160233, 'timing/model': 1.3578021628782153, 'learning_rate/qwen_vl_interface': 7.889308033471156e-07, 'learning_rate/action_model': 7.889308033471155e-06, 'epoch': 2.4}) +[2026-06-18 07:42:04,672][starVLA.training.train_starvla][INFO] - Step 6231, Loss: {'train/loss': 0.3920896053314209, 'train/grad_norm_pre_clip': 15.261207580566406, 'timing/data': 0.007285239174962044, 'timing/model': 1.557970923371613, 'learning_rate/qwen_vl_interface': 7.881884985486662e-07, 'learning_rate/action_model': 7.881884985486662e-06, 'epoch': 2.4}) +[2026-06-18 07:42:10,427][starVLA.training.train_starvla][INFO] - Step 6232, Loss: {'train/loss': 0.23200605809688568, 'train/grad_norm_pre_clip': 18.840435028076172, 'timing/data': 0.0015133516862988472, 'timing/model': 1.7682189410552382, 'learning_rate/qwen_vl_interface': 7.874471186885427e-07, 'learning_rate/action_model': 7.874471186885427e-06, 'epoch': 2.4}) +[2026-06-18 07:42:15,682][starVLA.training.train_starvla][INFO] - Step 6233, Loss: {'train/loss': 0.6099504232406616, 'train/grad_norm_pre_clip': 20.66106414794922, 'timing/data': 0.0020157145336270332, 'timing/model': 1.6081324461847544, 'learning_rate/qwen_vl_interface': 7.867066639204326e-07, 'learning_rate/action_model': 7.867066639204327e-06, 'epoch': 2.4}) +[2026-06-18 07:42:20,848][starVLA.training.train_starvla][INFO] - Step 6234, Loss: {'train/loss': 0.2142254263162613, 'train/grad_norm_pre_clip': 11.788421630859375, 'timing/data': 0.001056554727256298, 'timing/model': 1.149418881163001, 'learning_rate/qwen_vl_interface': 7.859671343978348e-07, 'learning_rate/action_model': 7.859671343978348e-06, 'epoch': 2.4}) +[2026-06-18 07:42:26,361][starVLA.training.train_starvla][INFO] - Step 6235, Loss: {'train/loss': 0.26303979754447937, 'train/grad_norm_pre_clip': 15.11116886138916, 'timing/data': 0.0011627310886979103, 'timing/model': 1.397339604794979, 'learning_rate/qwen_vl_interface': 7.852285302740528e-07, 'learning_rate/action_model': 7.852285302740527e-06, 'epoch': 2.4}) +[2026-06-18 07:42:31,344][starVLA.training.train_starvla][INFO] - Step 6236, Loss: {'train/loss': 0.5168030261993408, 'train/grad_norm_pre_clip': 15.983551025390625, 'timing/data': 0.0009901812300086021, 'timing/model': 1.0967911463230848, 'learning_rate/qwen_vl_interface': 7.844908517022005e-07, 'learning_rate/action_model': 7.844908517022005e-06, 'epoch': 2.4}) +[2026-06-18 07:42:36,293][starVLA.training.train_starvla][INFO] - Step 6237, Loss: {'train/loss': 0.25866127014160156, 'train/grad_norm_pre_clip': 18.00612449645996, 'timing/data': 0.0017272382974624634, 'timing/model': 1.1648303931578994, 'learning_rate/qwen_vl_interface': 7.837540988352001e-07, 'learning_rate/action_model': 7.837540988352e-06, 'epoch': 2.4}) +[2026-06-18 07:42:42,159][starVLA.training.train_starvla][INFO] - Step 6238, Loss: {'train/loss': 0.543323278427124, 'train/grad_norm_pre_clip': 19.3570613861084, 'timing/data': 0.0012816134840250015, 'timing/model': 1.1285618962720037, 'learning_rate/qwen_vl_interface': 7.83018271825782e-07, 'learning_rate/action_model': 7.83018271825782e-06, 'epoch': 2.4}) +[2026-06-18 07:42:46,954][starVLA.training.train_starvla][INFO] - Step 6239, Loss: {'train/loss': 0.24023880064487457, 'train/grad_norm_pre_clip': 11.319217681884766, 'timing/data': 0.0015139570459723473, 'timing/model': 1.1317282328382134, 'learning_rate/qwen_vl_interface': 7.822833708264821e-07, 'learning_rate/action_model': 7.822833708264821e-06, 'epoch': 2.4}) +[2026-06-18 07:42:52,162][starVLA.training.train_starvla][INFO] - Step 6240, Loss: {'train/loss': 0.27887633442878723, 'train/grad_norm_pre_clip': 11.402215957641602, 'timing/data': 0.001827627420425415, 'timing/model': 1.1297917570918798, 'learning_rate/qwen_vl_interface': 7.81549395989646e-07, 'learning_rate/action_model': 7.81549395989646e-06, 'epoch': 2.4}) +[2026-06-18 07:42:57,810][starVLA.training.train_starvla][INFO] - Step 6241, Loss: {'train/loss': 0.27203992009162903, 'train/grad_norm_pre_clip': 13.93044376373291, 'timing/data': 0.006029428914189339, 'timing/model': 1.6852620271965861, 'learning_rate/qwen_vl_interface': 7.808163474674291e-07, 'learning_rate/action_model': 7.80816347467429e-06, 'epoch': 2.4}) +[2026-06-18 07:43:03,065][starVLA.training.train_starvla][INFO] - Step 6242, Loss: {'train/loss': 0.4743027091026306, 'train/grad_norm_pre_clip': 18.474252700805664, 'timing/data': 0.0014870660379529, 'timing/model': 1.5234972322359681, 'learning_rate/qwen_vl_interface': 7.800842254117922e-07, 'learning_rate/action_model': 7.80084225411792e-06, 'epoch': 2.4}) +[2026-06-18 07:43:07,921][starVLA.training.train_starvla][INFO] - Step 6243, Loss: {'train/loss': 0.2627932131290436, 'train/grad_norm_pre_clip': 14.707239151000977, 'timing/data': 0.0012968527153134346, 'timing/model': 1.355837962590158, 'learning_rate/qwen_vl_interface': 7.793530299745061e-07, 'learning_rate/action_model': 7.793530299745061e-06, 'epoch': 2.4}) +[2026-06-18 07:43:16,928][starVLA.training.train_starvla][INFO] - Step 6244, Loss: {'train/loss': 0.15662261843681335, 'train/grad_norm_pre_clip': 19.218595504760742, 'timing/data': 0.0016842391341924667, 'timing/model': 4.690025528892875, 'learning_rate/qwen_vl_interface': 7.786227613071464e-07, 'learning_rate/action_model': 7.786227613071464e-06, 'epoch': 2.4}) +[2026-06-18 07:43:34,784][starVLA.training.train_starvla][INFO] - Step 6245, Loss: {'train/loss': 0.29382720589637756, 'train/grad_norm_pre_clip': 18.594627380371094, 'timing/data': 0.0014271605759859085, 'timing/model': 3.2102946816012263, 'learning_rate/qwen_vl_interface': 7.778934195610998e-07, 'learning_rate/action_model': 7.778934195610998e-06, 'epoch': 2.4}) +[2026-06-18 07:43:45,232][starVLA.training.train_starvla][INFO] - Step 6246, Loss: {'train/loss': 0.5526684522628784, 'train/grad_norm_pre_clip': 24.00461196899414, 'timing/data': 0.001370125450193882, 'timing/model': 1.2793833017349243, 'learning_rate/qwen_vl_interface': 7.771650048875598e-07, 'learning_rate/action_model': 7.771650048875597e-06, 'epoch': 2.4}) +[2026-06-18 07:43:56,192][starVLA.training.train_starvla][INFO] - Step 6247, Loss: {'train/loss': 0.5114840865135193, 'train/grad_norm_pre_clip': 17.801971435546875, 'timing/data': 0.004350455477833748, 'timing/model': 3.072748708538711, 'learning_rate/qwen_vl_interface': 7.764375174375276e-07, 'learning_rate/action_model': 7.764375174375276e-06, 'epoch': 2.4}) +[2026-06-18 07:44:07,832][starVLA.training.train_starvla][INFO] - Step 6248, Loss: {'train/loss': 0.13340604305267334, 'train/grad_norm_pre_clip': 7.199343204498291, 'timing/data': 0.002440759912133217, 'timing/model': 1.9625131553038955, 'learning_rate/qwen_vl_interface': 7.757109573618116e-07, 'learning_rate/action_model': 7.757109573618115e-06, 'epoch': 2.4}) +[2026-06-18 07:44:22,315][starVLA.training.train_starvla][INFO] - Step 6249, Loss: {'train/loss': 0.6517133712768555, 'train/grad_norm_pre_clip': 24.872161865234375, 'timing/data': 0.0013917041942477226, 'timing/model': 6.08876749034971, 'learning_rate/qwen_vl_interface': 7.749853248110287e-07, 'learning_rate/action_model': 7.749853248110287e-06, 'epoch': 2.4}) +[2026-06-18 07:44:32,005][starVLA.training.train_starvla][INFO] - Step 6250, Loss: {'train/loss': 0.234745055437088, 'train/grad_norm_pre_clip': 16.626920700073242, 'timing/data': 0.0012232596054673195, 'timing/model': 4.318462022580206, 'learning_rate/qwen_vl_interface': 7.742606199356032e-07, 'learning_rate/action_model': 7.74260619935603e-06, 'epoch': 2.4}) +[2026-06-18 07:44:42,721][starVLA.training.train_starvla][INFO] - Step 6251, Loss: {'train/loss': 0.2632864713668823, 'train/grad_norm_pre_clip': 13.932350158691406, 'timing/data': 0.0028887903317809105, 'timing/model': 3.4374686693772674, 'learning_rate/qwen_vl_interface': 7.735368428857665e-07, 'learning_rate/action_model': 7.735368428857665e-06, 'epoch': 2.4}) +[2026-06-18 07:44:49,548][starVLA.training.train_starvla][INFO] - Step 6252, Loss: {'train/loss': 0.3600192368030548, 'train/grad_norm_pre_clip': 15.608861923217773, 'timing/data': 0.0013831621035933495, 'timing/model': 2.9010645328089595, 'learning_rate/qwen_vl_interface': 7.728139938115605e-07, 'learning_rate/action_model': 7.728139938115604e-06, 'epoch': 2.41}) +[2026-06-18 07:44:58,697][starVLA.training.train_starvla][INFO] - Step 6253, Loss: {'train/loss': 0.24340657889842987, 'train/grad_norm_pre_clip': 14.873313903808594, 'timing/data': 0.0011809412389993668, 'timing/model': 3.232766388915479, 'learning_rate/qwen_vl_interface': 7.720920728628306e-07, 'learning_rate/action_model': 7.720920728628306e-06, 'epoch': 2.41}) +[2026-06-18 07:45:06,053][starVLA.training.train_starvla][INFO] - Step 6254, Loss: {'train/loss': 0.4319891035556793, 'train/grad_norm_pre_clip': 14.723577499389648, 'timing/data': 0.0013210466131567955, 'timing/model': 1.6967067290097475, 'learning_rate/qwen_vl_interface': 7.713710801892327e-07, 'learning_rate/action_model': 7.713710801892326e-06, 'epoch': 2.41}) +[2026-06-18 07:45:13,066][starVLA.training.train_starvla][INFO] - Step 6255, Loss: {'train/loss': 0.30160394310951233, 'train/grad_norm_pre_clip': 12.607152938842773, 'timing/data': 0.0008296500891447067, 'timing/model': 1.1205630488693714, 'learning_rate/qwen_vl_interface': 7.70651015940228e-07, 'learning_rate/action_model': 7.70651015940228e-06, 'epoch': 2.41}) +[2026-06-18 07:45:18,457][starVLA.training.train_starvla][INFO] - Step 6256, Loss: {'train/loss': 0.23412159085273743, 'train/grad_norm_pre_clip': 15.61596393585205, 'timing/data': 0.000980490818619728, 'timing/model': 1.1943426793441176, 'learning_rate/qwen_vl_interface': 7.699318802650891e-07, 'learning_rate/action_model': 7.69931880265089e-06, 'epoch': 2.41}) +[2026-06-18 07:45:23,667][starVLA.training.train_starvla][INFO] - Step 6257, Loss: {'train/loss': 0.4106144607067108, 'train/grad_norm_pre_clip': 16.948171615600586, 'timing/data': 0.012808375991880894, 'timing/model': 1.7005128730088472, 'learning_rate/qwen_vl_interface': 7.692136733128918e-07, 'learning_rate/action_model': 7.692136733128917e-06, 'epoch': 2.41}) +[2026-06-18 07:45:28,884][starVLA.training.train_starvla][INFO] - Step 6258, Loss: {'train/loss': 0.15243032574653625, 'train/grad_norm_pre_clip': 10.006729125976562, 'timing/data': 0.013332825154066086, 'timing/model': 1.3187919007614255, 'learning_rate/qwen_vl_interface': 7.684963952325207e-07, 'learning_rate/action_model': 7.684963952325206e-06, 'epoch': 2.41}) +[2026-06-18 07:45:34,372][starVLA.training.train_starvla][INFO] - Step 6259, Loss: {'train/loss': 0.22080761194229126, 'train/grad_norm_pre_clip': 12.055329322814941, 'timing/data': 0.035273477435112, 'timing/model': 1.3489002352580428, 'learning_rate/qwen_vl_interface': 7.6778004617267e-07, 'learning_rate/action_model': 7.6778004617267e-06, 'epoch': 2.41}) +[2026-06-18 07:45:40,200][starVLA.training.train_starvla][INFO] - Step 6260, Loss: {'train/loss': 0.26095783710479736, 'train/grad_norm_pre_clip': 17.206632614135742, 'timing/data': 0.015288449823856354, 'timing/model': 1.4208735227584839, 'learning_rate/qwen_vl_interface': 7.670646262818387e-07, 'learning_rate/action_model': 7.670646262818387e-06, 'epoch': 2.41}) +[2026-06-18 07:45:45,494][starVLA.training.train_starvla][INFO] - Step 6261, Loss: {'train/loss': 0.4240010976791382, 'train/grad_norm_pre_clip': 18.329212188720703, 'timing/data': 0.011955679394304752, 'timing/model': 1.2768739517778158, 'learning_rate/qwen_vl_interface': 7.663501357083342e-07, 'learning_rate/action_model': 7.66350135708334e-06, 'epoch': 2.41}) +[2026-06-18 07:45:51,579][starVLA.training.train_starvla][INFO] - Step 6262, Loss: {'train/loss': 0.20134352147579193, 'train/grad_norm_pre_clip': 16.121414184570312, 'timing/data': 0.014619355089962482, 'timing/model': 1.9955490110442042, 'learning_rate/qwen_vl_interface': 7.656365746002715e-07, 'learning_rate/action_model': 7.656365746002716e-06, 'epoch': 2.41}) +[2026-06-18 07:45:56,907][starVLA.training.train_starvla][INFO] - Step 6263, Loss: {'train/loss': 0.2274128645658493, 'train/grad_norm_pre_clip': 14.19478988647461, 'timing/data': 0.023320632986724377, 'timing/model': 1.5465373322367668, 'learning_rate/qwen_vl_interface': 7.649239431055718e-07, 'learning_rate/action_model': 7.649239431055717e-06, 'epoch': 2.41}) +[2026-06-18 07:46:03,497][starVLA.training.train_starvla][INFO] - Step 6264, Loss: {'train/loss': 0.4369621276855469, 'train/grad_norm_pre_clip': 22.957990646362305, 'timing/data': 0.019664423540234566, 'timing/model': 1.3339580576866865, 'learning_rate/qwen_vl_interface': 7.642122413719652e-07, 'learning_rate/action_model': 7.642122413719651e-06, 'epoch': 2.41}) +[2026-06-18 07:46:21,547][starVLA.training.train_starvla][INFO] - Step 6265, Loss: {'train/loss': 0.14559634029865265, 'train/grad_norm_pre_clip': 10.251315116882324, 'timing/data': 0.020996364764869213, 'timing/model': 1.114077660255134, 'learning_rate/qwen_vl_interface': 7.635014695469877e-07, 'learning_rate/action_model': 7.635014695469878e-06, 'epoch': 2.41}) +[2026-06-18 07:46:27,083][starVLA.training.train_starvla][INFO] - Step 6266, Loss: {'train/loss': 0.3401279151439667, 'train/grad_norm_pre_clip': 13.177661895751953, 'timing/data': 0.014235926792025566, 'timing/model': 1.2339150197803974, 'learning_rate/qwen_vl_interface': 7.627916277779832e-07, 'learning_rate/action_model': 7.627916277779831e-06, 'epoch': 2.41}) +[2026-06-18 07:46:37,336][starVLA.training.train_starvla][INFO] - Step 6267, Loss: {'train/loss': 0.1699637472629547, 'train/grad_norm_pre_clip': 11.590023040771484, 'timing/data': 0.011410962790250778, 'timing/model': 6.2012793850153685, 'learning_rate/qwen_vl_interface': 7.62082716212103e-07, 'learning_rate/action_model': 7.62082716212103e-06, 'epoch': 2.41}) +[2026-06-18 07:46:54,698][starVLA.training.train_starvla][INFO] - Step 6268, Loss: {'train/loss': 0.1799900382757187, 'train/grad_norm_pre_clip': 9.66901969909668, 'timing/data': 0.017750738188624382, 'timing/model': 4.6781239761039615, 'learning_rate/qwen_vl_interface': 7.613747349963048e-07, 'learning_rate/action_model': 7.613747349963048e-06, 'epoch': 2.41}) +[2026-06-18 07:47:03,562][starVLA.training.train_starvla][INFO] - Step 6269, Loss: {'train/loss': 0.12234315276145935, 'train/grad_norm_pre_clip': 10.214320182800293, 'timing/data': 0.01471868995577097, 'timing/model': 1.4682694990187883, 'learning_rate/qwen_vl_interface': 7.606676842773531e-07, 'learning_rate/action_model': 7.60667684277353e-06, 'epoch': 2.41}) +[2026-06-18 07:47:08,780][starVLA.training.train_starvla][INFO] - Step 6270, Loss: {'train/loss': 0.3103558421134949, 'train/grad_norm_pre_clip': 16.583425521850586, 'timing/data': 0.012081668712198734, 'timing/model': 1.1312405690550804, 'learning_rate/qwen_vl_interface': 7.599615642018218e-07, 'learning_rate/action_model': 7.599615642018218e-06, 'epoch': 2.41}) +[2026-06-18 07:47:16,451][starVLA.training.train_starvla][INFO] - Step 6271, Loss: {'train/loss': 0.1396394968032837, 'train/grad_norm_pre_clip': 10.863410949707031, 'timing/data': 0.014386152848601341, 'timing/model': 1.977495463564992, 'learning_rate/qwen_vl_interface': 7.592563749160908e-07, 'learning_rate/action_model': 7.592563749160908e-06, 'epoch': 2.41}) +[2026-06-18 07:47:21,782][starVLA.training.train_starvla][INFO] - Step 6272, Loss: {'train/loss': 0.16287170350551605, 'train/grad_norm_pre_clip': 12.01275634765625, 'timing/data': 0.012154413387179375, 'timing/model': 1.455886723473668, 'learning_rate/qwen_vl_interface': 7.585521165663445e-07, 'learning_rate/action_model': 7.585521165663445e-06, 'epoch': 2.41}) +[2026-06-18 07:47:28,190][starVLA.training.train_starvla][INFO] - Step 6273, Loss: {'train/loss': 0.3538108468055725, 'train/grad_norm_pre_clip': 23.57053565979004, 'timing/data': 0.028995062224566936, 'timing/model': 1.2597318878397346, 'learning_rate/qwen_vl_interface': 7.578487892985769e-07, 'learning_rate/action_model': 7.578487892985769e-06, 'epoch': 2.41}) +[2026-06-18 07:47:33,653][starVLA.training.train_starvla][INFO] - Step 6274, Loss: {'train/loss': 0.2903711795806885, 'train/grad_norm_pre_clip': 15.352334976196289, 'timing/data': 0.012537973932921886, 'timing/model': 1.1103019388392568, 'learning_rate/qwen_vl_interface': 7.571463932585901e-07, 'learning_rate/action_model': 7.571463932585901e-06, 'epoch': 2.41}) +[2026-06-18 07:47:45,424][starVLA.training.train_starvla][INFO] - Step 6275, Loss: {'train/loss': 0.09203458577394485, 'train/grad_norm_pre_clip': 8.809931755065918, 'timing/data': 0.014338241890072823, 'timing/model': 4.585089121945202, 'learning_rate/qwen_vl_interface': 7.564449285919897e-07, 'learning_rate/action_model': 7.564449285919897e-06, 'epoch': 2.41}) +[2026-06-18 07:47:55,215][starVLA.training.train_starvla][INFO] - Step 6276, Loss: {'train/loss': 0.22976545989513397, 'train/grad_norm_pre_clip': 13.72326374053955, 'timing/data': 0.014734646305441856, 'timing/model': 1.574394497089088, 'learning_rate/qwen_vl_interface': 7.557443954441924e-07, 'learning_rate/action_model': 7.557443954441923e-06, 'epoch': 2.41}) +[2026-06-18 07:48:00,713][starVLA.training.train_starvla][INFO] - Step 6277, Loss: {'train/loss': 0.15104657411575317, 'train/grad_norm_pre_clip': 9.942523956298828, 'timing/data': 0.041820162907242775, 'timing/model': 1.5700948908925056, 'learning_rate/qwen_vl_interface': 7.550447939604161e-07, 'learning_rate/action_model': 7.550447939604161e-06, 'epoch': 2.41}) +[2026-06-18 07:48:06,039][starVLA.training.train_starvla][INFO] - Step 6278, Loss: {'train/loss': 0.261860728263855, 'train/grad_norm_pre_clip': 12.59402847290039, 'timing/data': 0.02385102678090334, 'timing/model': 1.5440690452232957, 'learning_rate/qwen_vl_interface': 7.543461242856916e-07, 'learning_rate/action_model': 7.543461242856916e-06, 'epoch': 2.42}) +[2026-06-18 07:48:11,980][starVLA.training.train_starvla][INFO] - Step 6279, Loss: {'train/loss': 0.30246877670288086, 'train/grad_norm_pre_clip': 12.294483184814453, 'timing/data': 0.04774269089102745, 'timing/model': 1.2415851978585124, 'learning_rate/qwen_vl_interface': 7.536483865648531e-07, 'learning_rate/action_model': 7.536483865648531e-06, 'epoch': 2.42}) +[2026-06-18 07:48:16,768][starVLA.training.train_starvla][INFO] - Step 6280, Loss: {'train/loss': 0.5268490314483643, 'train/grad_norm_pre_clip': 22.94794273376465, 'timing/data': 0.04292637389153242, 'timing/model': 1.2011316400021315, 'learning_rate/qwen_vl_interface': 7.52951580942542e-07, 'learning_rate/action_model': 7.5295158094254195e-06, 'epoch': 2.42}) +[2026-06-18 07:48:25,601][starVLA.training.train_starvla][INFO] - Step 6281, Loss: {'train/loss': 0.34132274985313416, 'train/grad_norm_pre_clip': 11.382478713989258, 'timing/data': 0.04653498250991106, 'timing/model': 4.899480698630214, 'learning_rate/qwen_vl_interface': 7.522557075632078e-07, 'learning_rate/action_model': 7.522557075632078e-06, 'epoch': 2.42}) +[2026-06-18 07:48:30,619][starVLA.training.train_starvla][INFO] - Step 6282, Loss: {'train/loss': 0.3361808955669403, 'train/grad_norm_pre_clip': 19.373050689697266, 'timing/data': 0.04576793126761913, 'timing/model': 1.0817008195444942, 'learning_rate/qwen_vl_interface': 7.515607665711051e-07, 'learning_rate/action_model': 7.515607665711051e-06, 'epoch': 2.42}) +[2026-06-18 07:48:35,426][starVLA.training.train_starvla][INFO] - Step 6283, Loss: {'train/loss': 0.2563926577568054, 'train/grad_norm_pre_clip': 20.136619567871094, 'timing/data': 0.0444259699434042, 'timing/model': 1.0916454875841737, 'learning_rate/qwen_vl_interface': 7.508667581102959e-07, 'learning_rate/action_model': 7.508667581102958e-06, 'epoch': 2.42}) +[2026-06-18 07:48:41,061][starVLA.training.train_starvla][INFO] - Step 6284, Loss: {'train/loss': 0.414072722196579, 'train/grad_norm_pre_clip': 17.483842849731445, 'timing/data': 0.04836764466017485, 'timing/model': 1.2183530572801828, 'learning_rate/qwen_vl_interface': 7.50173682324648e-07, 'learning_rate/action_model': 7.501736823246481e-06, 'epoch': 2.42}) +[2026-06-18 07:48:46,523][starVLA.training.train_starvla][INFO] - Step 6285, Loss: {'train/loss': 0.46186861395835876, 'train/grad_norm_pre_clip': 15.791651725769043, 'timing/data': 0.058183339424431324, 'timing/model': 1.5759157305583358, 'learning_rate/qwen_vl_interface': 7.494815393578394e-07, 'learning_rate/action_model': 7.494815393578393e-06, 'epoch': 2.42}) +[2026-06-18 07:48:55,769][starVLA.training.train_starvla][INFO] - Step 6286, Loss: {'train/loss': 0.5069578289985657, 'train/grad_norm_pre_clip': 19.0330753326416, 'timing/data': 0.0423651784658432, 'timing/model': 2.291857392527163, 'learning_rate/qwen_vl_interface': 7.487903293533501e-07, 'learning_rate/action_model': 7.4879032935335e-06, 'epoch': 2.42}) +[2026-06-18 07:49:08,780][starVLA.training.train_starvla][INFO] - Step 6287, Loss: {'train/loss': 0.1014719158411026, 'train/grad_norm_pre_clip': 7.917959213256836, 'timing/data': 0.09868963155895472, 'timing/model': 1.4395608473569155, 'learning_rate/qwen_vl_interface': 7.481000524544686e-07, 'learning_rate/action_model': 7.481000524544686e-06, 'epoch': 2.42}) +[2026-06-18 07:49:24,461][starVLA.training.train_starvla][INFO] - Step 6288, Loss: {'train/loss': 0.1781264990568161, 'train/grad_norm_pre_clip': 10.0875883102417, 'timing/data': 0.04535327386111021, 'timing/model': 4.6973576080054045, 'learning_rate/qwen_vl_interface': 7.474107088042898e-07, 'learning_rate/action_model': 7.474107088042898e-06, 'epoch': 2.42}) +[2026-06-18 07:49:35,992][starVLA.training.train_starvla][INFO] - Step 6289, Loss: {'train/loss': 0.5753790736198425, 'train/grad_norm_pre_clip': 20.457374572753906, 'timing/data': 0.057114980183541775, 'timing/model': 1.5745956599712372, 'learning_rate/qwen_vl_interface': 7.467222985457167e-07, 'learning_rate/action_model': 7.4672229854571675e-06, 'epoch': 2.42}) +[2026-06-18 07:49:46,237][starVLA.training.train_starvla][INFO] - Step 6290, Loss: {'train/loss': 0.3525349199771881, 'train/grad_norm_pre_clip': 17.13514518737793, 'timing/data': 0.06428784225136042, 'timing/model': 4.514874752610922, 'learning_rate/qwen_vl_interface': 7.46034821821458e-07, 'learning_rate/action_model': 7.46034821821458e-06, 'epoch': 2.42}) +[2026-06-18 07:49:52,424][starVLA.training.train_starvla][INFO] - Step 6291, Loss: {'train/loss': 0.27594929933547974, 'train/grad_norm_pre_clip': 11.538799285888672, 'timing/data': 0.043340081349015236, 'timing/model': 2.0289394855499268, 'learning_rate/qwen_vl_interface': 7.453482787740255e-07, 'learning_rate/action_model': 7.4534827877402545e-06, 'epoch': 2.42}) +[2026-06-18 07:49:57,286][starVLA.training.train_starvla][INFO] - Step 6292, Loss: {'train/loss': 0.3911627531051636, 'train/grad_norm_pre_clip': 19.741064071655273, 'timing/data': 0.04736574273556471, 'timing/model': 1.4622670663520694, 'learning_rate/qwen_vl_interface': 7.446626695457427e-07, 'learning_rate/action_model': 7.446626695457427e-06, 'epoch': 2.42}) +[2026-06-18 07:50:11,968][starVLA.training.train_starvla][INFO] - Step 6293, Loss: {'train/loss': 0.5132755041122437, 'train/grad_norm_pre_clip': 14.986690521240234, 'timing/data': 0.05505168903619051, 'timing/model': 1.540950171649456, 'learning_rate/qwen_vl_interface': 7.439779942787367e-07, 'learning_rate/action_model': 7.439779942787367e-06, 'epoch': 2.42}) +[2026-06-18 07:50:21,430][starVLA.training.train_starvla][INFO] - Step 6294, Loss: {'train/loss': 0.30251604318618774, 'train/grad_norm_pre_clip': 12.355511665344238, 'timing/data': 0.06794646754860878, 'timing/model': 1.2559750620275736, 'learning_rate/qwen_vl_interface': 7.432942531149409e-07, 'learning_rate/action_model': 7.432942531149409e-06, 'epoch': 2.42}) +[2026-06-18 07:50:26,247][starVLA.training.train_starvla][INFO] - Step 6295, Loss: {'train/loss': 0.22488024830818176, 'train/grad_norm_pre_clip': 15.450677871704102, 'timing/data': 0.05189328733831644, 'timing/model': 1.107625455595553, 'learning_rate/qwen_vl_interface': 7.42611446196096e-07, 'learning_rate/action_model': 7.426114461960959e-06, 'epoch': 2.42}) +[2026-06-18 07:50:31,420][starVLA.training.train_starvla][INFO] - Step 6296, Loss: {'train/loss': 0.3483796715736389, 'train/grad_norm_pre_clip': 16.04503059387207, 'timing/data': 0.05058062542229891, 'timing/model': 1.450741003267467, 'learning_rate/qwen_vl_interface': 7.419295736637483e-07, 'learning_rate/action_model': 7.419295736637483e-06, 'epoch': 2.42}) +[2026-06-18 07:50:36,129][starVLA.training.train_starvla][INFO] - Step 6297, Loss: {'train/loss': 0.2308456003665924, 'train/grad_norm_pre_clip': 14.176679611206055, 'timing/data': 0.052178836427628994, 'timing/model': 1.10827028285712, 'learning_rate/qwen_vl_interface': 7.412486356592512e-07, 'learning_rate/action_model': 7.412486356592511e-06, 'epoch': 2.42}) +[2026-06-18 07:50:41,104][starVLA.training.train_starvla][INFO] - Step 6298, Loss: {'train/loss': 0.48150306940078735, 'train/grad_norm_pre_clip': 22.247455596923828, 'timing/data': 0.042993294075131416, 'timing/model': 1.1330494973808527, 'learning_rate/qwen_vl_interface': 7.405686323237623e-07, 'learning_rate/action_model': 7.405686323237623e-06, 'epoch': 2.42}) +[2026-06-18 07:50:49,835][starVLA.training.train_starvla][INFO] - Step 6299, Loss: {'train/loss': 0.22322021424770355, 'train/grad_norm_pre_clip': 13.341660499572754, 'timing/data': 0.050504411570727825, 'timing/model': 4.378414677456021, 'learning_rate/qwen_vl_interface': 7.398895637982504e-07, 'learning_rate/action_model': 7.398895637982503e-06, 'epoch': 2.42}) +[2026-06-18 07:55:05,339][starVLA.training.train_starvla][INFO] - Step 6300, Loss: {'train/loss': 0.09396309405565262, 'train/grad_norm_pre_clip': 6.596848964691162, 'eval/loss': 0.42140519618988037, 'eval/latency_4/loss': 0.42140519618988037, 'eval/demon_fire_precision': 0.9680306905370843, 'eval/demon_fire_recall': 0.9834361805781098, 'eval/demon_fire_f1': 0.9756726276784276, 'eval/demon_fire_rate_model': 0.9186490455212922, 'eval/demon_fire_rate_teacher': 0.9042584434654919, 'eval/demon_left_precision': 0.7873417721518987, 'eval/demon_left_recall': 0.7100456621004566, 'eval/demon_left_f1': 0.7466986794717887, 'eval/demon_left_rate_model': 0.11600587371512482, 'eval/demon_left_rate_teacher': 0.12863436123348018, 'eval/demon_right_precision': 0.7858407079646018, 'eval/demon_right_recall': 0.8014440433212996, 'eval/demon_right_f1': 0.7935656836461126, 'eval/demon_right_rate_model': 0.16593245227606462, 'eval/demon_right_rate_teacher': 0.16270190895741557, 'eval/demon_move_f1': 0.7701321815589506, 'eval/demon_cc_f1': 0.9033445833116885, 'eval/latency_4/demon_fire_precision': 0.9680306905370843, 'eval/latency_4/demon_fire_recall': 0.9834361805781098, 'eval/latency_4/demon_fire_f1': 0.9756726276784276, 'eval/latency_4/demon_fire_rate_model': 0.9186490455212922, 'eval/latency_4/demon_fire_rate_teacher': 0.9042584434654919, 'eval/latency_4/demon_left_precision': 0.7873417721518987, 'eval/latency_4/demon_left_recall': 0.7100456621004566, 'eval/latency_4/demon_left_f1': 0.7466986794717887, 'eval/latency_4/demon_left_rate_model': 0.11600587371512482, 'eval/latency_4/demon_left_rate_teacher': 0.12863436123348018, 'eval/latency_4/demon_right_precision': 0.7858407079646018, 'eval/latency_4/demon_right_recall': 0.8014440433212996, 'eval/latency_4/demon_right_f1': 0.7935656836461126, 'eval/latency_4/demon_right_rate_model': 0.16593245227606462, 'eval/latency_4/demon_right_rate_teacher': 0.16270190895741557, 'eval/latency_4/demon_move_f1': 0.7701321815589506, 'eval/latency_4/demon_cc_f1': 0.9033445833116885, 'timing/data': 0.07704207580536604, 'timing/model': 1.1205198541283607, 'learning_rate/qwen_vl_interface': 7.392114302234838e-07, 'learning_rate/action_model': 7.3921143022348385e-06, 'epoch': 2.42}) +[2026-06-18 07:55:05,425][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_6300_state +[2026-06-18 07:55:05,433][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 07:55:14,488][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_6300_state/model.safetensors +[2026-06-18 07:55:38,030][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_6300_state/optimizer.bin +[2026-06-18 07:55:38,031][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_6300_state/sampler.bin +[2026-06-18 07:55:38,031][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_6300_state/sampler_1.bin +[2026-06-18 07:55:38,034][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_6300_state/random_states_0.pkl +[2026-06-18 07:55:38,039][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 6300 +[2026-06-18 07:55:38,040][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 07:55:38,059][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 07:55:46,797][starVLA.training.train_starvla][INFO] - Step 6301, Loss: {'train/loss': 0.2860410809516907, 'train/grad_norm_pre_clip': 13.63415813446045, 'timing/data': 0.04819290433079004, 'timing/model': 4.112261815927923, 'learning_rate/qwen_vl_interface': 7.385342317400418e-07, 'learning_rate/action_model': 7.385342317400417e-06, 'epoch': 2.42}) +[2026-06-18 07:55:57,609][starVLA.training.train_starvla][INFO] - Step 6302, Loss: {'train/loss': 0.3969270586967468, 'train/grad_norm_pre_clip': 15.172477722167969, 'timing/data': 0.042707053013145924, 'timing/model': 1.4940078305080533, 'learning_rate/qwen_vl_interface': 7.378579684883069e-07, 'learning_rate/action_model': 7.378579684883069e-06, 'epoch': 2.42}) +[2026-06-18 07:56:02,137][starVLA.training.train_starvla][INFO] - Step 6303, Loss: {'train/loss': 0.22510036826133728, 'train/grad_norm_pre_clip': 11.13811206817627, 'timing/data': 0.03843645192682743, 'timing/model': 1.368149533867836, 'learning_rate/qwen_vl_interface': 7.371826406084714e-07, 'learning_rate/action_model': 7.371826406084714e-06, 'epoch': 2.42}) +[2026-06-18 07:56:07,464][starVLA.training.train_starvla][INFO] - Step 6304, Loss: {'train/loss': 0.35065826773643494, 'train/grad_norm_pre_clip': 17.625547409057617, 'timing/data': 0.043073153123259544, 'timing/model': 1.1021497724577785, 'learning_rate/qwen_vl_interface': 7.365082482405315e-07, 'learning_rate/action_model': 7.365082482405315e-06, 'epoch': 2.43}) +[2026-06-18 07:56:14,136][starVLA.training.train_starvla][INFO] - Step 6305, Loss: {'train/loss': 0.11801683902740479, 'train/grad_norm_pre_clip': 6.9175615310668945, 'timing/data': 0.04417731333523989, 'timing/model': 1.2551924353465438, 'learning_rate/qwen_vl_interface': 7.358347915242877e-07, 'learning_rate/action_model': 7.358347915242876e-06, 'epoch': 2.43}) +[2026-06-18 07:56:19,386][starVLA.training.train_starvla][INFO] - Step 6306, Loss: {'train/loss': 0.15650619566440582, 'train/grad_norm_pre_clip': 14.581256866455078, 'timing/data': 0.041457719169557095, 'timing/model': 1.2027531983330846, 'learning_rate/qwen_vl_interface': 7.351622705993481e-07, 'learning_rate/action_model': 7.351622705993481e-06, 'epoch': 2.43}) +[2026-06-18 07:56:26,052][starVLA.training.train_starvla][INFO] - Step 6307, Loss: {'train/loss': 0.32966628670692444, 'train/grad_norm_pre_clip': 15.93600082397461, 'timing/data': 0.054533411748707294, 'timing/model': 3.0210902374237776, 'learning_rate/qwen_vl_interface': 7.344906856051288e-07, 'learning_rate/action_model': 7.344906856051288e-06, 'epoch': 2.43}) +[2026-06-18 07:56:31,136][starVLA.training.train_starvla][INFO] - Step 6308, Loss: {'train/loss': 0.364449679851532, 'train/grad_norm_pre_clip': 20.269325256347656, 'timing/data': 0.05166273284703493, 'timing/model': 1.1233101962134242, 'learning_rate/qwen_vl_interface': 7.338200366808493e-07, 'learning_rate/action_model': 7.338200366808492e-06, 'epoch': 2.43}) +[2026-06-18 07:56:36,389][starVLA.training.train_starvla][INFO] - Step 6309, Loss: {'train/loss': 0.1257324516773224, 'train/grad_norm_pre_clip': 9.256169319152832, 'timing/data': 0.04621188808232546, 'timing/model': 1.3948564920574427, 'learning_rate/qwen_vl_interface': 7.331503239655361e-07, 'learning_rate/action_model': 7.33150323965536e-06, 'epoch': 2.43}) +[2026-06-18 07:56:49,128][starVLA.training.train_starvla][INFO] - Step 6310, Loss: {'train/loss': 0.34108686447143555, 'train/grad_norm_pre_clip': 17.413440704345703, 'timing/data': 0.04337014351040125, 'timing/model': 1.1862560231238604, 'learning_rate/qwen_vl_interface': 7.324815475980207e-07, 'learning_rate/action_model': 7.324815475980207e-06, 'epoch': 2.43}) +[2026-06-18 07:56:55,553][starVLA.training.train_starvla][INFO] - Step 6311, Loss: {'train/loss': 0.23178508877754211, 'train/grad_norm_pre_clip': 14.64441204071045, 'timing/data': 0.05053337570279837, 'timing/model': 2.593328509479761, 'learning_rate/qwen_vl_interface': 7.318137077169416e-07, 'learning_rate/action_model': 7.318137077169415e-06, 'epoch': 2.43}) +[2026-06-18 07:57:03,942][starVLA.training.train_starvla][INFO] - Step 6312, Loss: {'train/loss': 0.43669819831848145, 'train/grad_norm_pre_clip': 13.177239418029785, 'timing/data': 0.045952619053423405, 'timing/model': 4.644655806943774, 'learning_rate/qwen_vl_interface': 7.311468044607425e-07, 'learning_rate/action_model': 7.311468044607425e-06, 'epoch': 2.43}) +[2026-06-18 07:57:14,101][starVLA.training.train_starvla][INFO] - Step 6313, Loss: {'train/loss': 0.363288551568985, 'train/grad_norm_pre_clip': 25.582908630371094, 'timing/data': 0.03935293294489384, 'timing/model': 5.347943044267595, 'learning_rate/qwen_vl_interface': 7.304808379676739e-07, 'learning_rate/action_model': 7.3048083796767394e-06, 'epoch': 2.43}) +[2026-06-18 07:57:24,148][starVLA.training.train_starvla][INFO] - Step 6314, Loss: {'train/loss': 0.462242990732193, 'train/grad_norm_pre_clip': 16.089075088500977, 'timing/data': 0.0512565802782774, 'timing/model': 1.1069260155782104, 'learning_rate/qwen_vl_interface': 7.298158083757903e-07, 'learning_rate/action_model': 7.298158083757903e-06, 'epoch': 2.43}) +[2026-06-18 07:57:37,026][starVLA.training.train_starvla][INFO] - Step 6315, Loss: {'train/loss': 0.23728211224079132, 'train/grad_norm_pre_clip': 11.920247077941895, 'timing/data': 0.05202706344425678, 'timing/model': 2.14130590390414, 'learning_rate/qwen_vl_interface': 7.291517158229536e-07, 'learning_rate/action_model': 7.291517158229535e-06, 'epoch': 2.43}) +[2026-06-18 07:57:46,046][starVLA.training.train_starvla][INFO] - Step 6316, Loss: {'train/loss': 0.2736548185348511, 'train/grad_norm_pre_clip': 12.245261192321777, 'timing/data': 0.03804188501089811, 'timing/model': 2.083271927200258, 'learning_rate/qwen_vl_interface': 7.284885604468312e-07, 'learning_rate/action_model': 7.284885604468311e-06, 'epoch': 2.43}) +[2026-06-18 07:57:54,833][starVLA.training.train_starvla][INFO] - Step 6317, Loss: {'train/loss': 0.2518755793571472, 'train/grad_norm_pre_clip': 14.871337890625, 'timing/data': 0.041984799318015575, 'timing/model': 1.7236146302893758, 'learning_rate/qwen_vl_interface': 7.278263423848948e-07, 'learning_rate/action_model': 7.278263423848948e-06, 'epoch': 2.43}) +[2026-06-18 07:58:03,301][starVLA.training.train_starvla][INFO] - Step 6318, Loss: {'train/loss': 0.28326645493507385, 'train/grad_norm_pre_clip': 13.741536140441895, 'timing/data': 0.04721699561923742, 'timing/model': 4.69044719543308, 'learning_rate/qwen_vl_interface': 7.271650617744239e-07, 'learning_rate/action_model': 7.271650617744238e-06, 'epoch': 2.43}) +[2026-06-18 07:58:18,963][starVLA.training.train_starvla][INFO] - Step 6319, Loss: {'train/loss': 0.3603375256061554, 'train/grad_norm_pre_clip': 16.76746940612793, 'timing/data': 0.04111101105809212, 'timing/model': 4.101310968399048, 'learning_rate/qwen_vl_interface': 7.265047187525033e-07, 'learning_rate/action_model': 7.265047187525033e-06, 'epoch': 2.43}) +[2026-06-18 07:58:27,876][starVLA.training.train_starvla][INFO] - Step 6320, Loss: {'train/loss': 0.3068784475326538, 'train/grad_norm_pre_clip': 12.873814582824707, 'timing/data': 0.06364159379154444, 'timing/model': 1.1770044639706612, 'learning_rate/qwen_vl_interface': 7.258453134560212e-07, 'learning_rate/action_model': 7.258453134560211e-06, 'epoch': 2.43}) +[2026-06-18 07:58:33,455][starVLA.training.train_starvla][INFO] - Step 6321, Loss: {'train/loss': 0.31796273589134216, 'train/grad_norm_pre_clip': 12.908824920654297, 'timing/data': 0.04710444342344999, 'timing/model': 1.2732801958918571, 'learning_rate/qwen_vl_interface': 7.251868460216743e-07, 'learning_rate/action_model': 7.251868460216743e-06, 'epoch': 2.43}) +[2026-06-18 07:58:38,363][starVLA.training.train_starvla][INFO] - Step 6322, Loss: {'train/loss': 0.24652890861034393, 'train/grad_norm_pre_clip': 13.500758171081543, 'timing/data': 0.054928784258663654, 'timing/model': 1.3894717134535313, 'learning_rate/qwen_vl_interface': 7.245293165859629e-07, 'learning_rate/action_model': 7.245293165859629e-06, 'epoch': 2.43}) +[2026-06-18 07:58:43,565][starVLA.training.train_starvla][INFO] - Step 6323, Loss: {'train/loss': 0.07269074022769928, 'train/grad_norm_pre_clip': 4.54461669921875, 'timing/data': 0.051374671049416065, 'timing/model': 1.3668924011290073, 'learning_rate/qwen_vl_interface': 7.238727252851935e-07, 'learning_rate/action_model': 7.238727252851935e-06, 'epoch': 2.43}) +[2026-06-18 07:58:48,872][starVLA.training.train_starvla][INFO] - Step 6324, Loss: {'train/loss': 0.301893949508667, 'train/grad_norm_pre_clip': 18.985239028930664, 'timing/data': 0.052216784097254276, 'timing/model': 1.2045158613473177, 'learning_rate/qwen_vl_interface': 7.23217072255479e-07, 'learning_rate/action_model': 7.23217072255479e-06, 'epoch': 2.43}) +[2026-06-18 07:58:53,938][starVLA.training.train_starvla][INFO] - Step 6325, Loss: {'train/loss': 0.14564430713653564, 'train/grad_norm_pre_clip': 10.327369689941406, 'timing/data': 0.046380666084587574, 'timing/model': 1.1086927326396108, 'learning_rate/qwen_vl_interface': 7.225623576327361e-07, 'learning_rate/action_model': 7.22562357632736e-06, 'epoch': 2.43}) +[2026-06-18 07:58:58,856][starVLA.training.train_starvla][INFO] - Step 6326, Loss: {'train/loss': 0.30172955989837646, 'train/grad_norm_pre_clip': 13.862970352172852, 'timing/data': 0.03956423792988062, 'timing/model': 1.1135160541161895, 'learning_rate/qwen_vl_interface': 7.219085815526887e-07, 'learning_rate/action_model': 7.219085815526887e-06, 'epoch': 2.43}) +[2026-06-18 07:59:03,765][starVLA.training.train_starvla][INFO] - Step 6327, Loss: {'train/loss': 0.33252525329589844, 'train/grad_norm_pre_clip': 12.885948181152344, 'timing/data': 0.05399164371192455, 'timing/model': 1.4398871017619967, 'learning_rate/qwen_vl_interface': 7.212557441508641e-07, 'learning_rate/action_model': 7.212557441508641e-06, 'epoch': 2.43}) +[2026-06-18 07:59:08,693][starVLA.training.train_starvla][INFO] - Step 6328, Loss: {'train/loss': 0.4714881479740143, 'train/grad_norm_pre_clip': 17.631685256958008, 'timing/data': 0.0408182218670845, 'timing/model': 1.2361994674429297, 'learning_rate/qwen_vl_interface': 7.206038455625972e-07, 'learning_rate/action_model': 7.206038455625971e-06, 'epoch': 2.43}) +[2026-06-18 07:59:13,948][starVLA.training.train_starvla][INFO] - Step 6329, Loss: {'train/loss': 0.2703690528869629, 'train/grad_norm_pre_clip': 33.074607849121094, 'timing/data': 0.0418206425383687, 'timing/model': 1.2169570680707693, 'learning_rate/qwen_vl_interface': 7.199528859230271e-07, 'learning_rate/action_model': 7.199528859230272e-06, 'epoch': 2.43}) +[2026-06-18 07:59:18,728][starVLA.training.train_starvla][INFO] - Step 6330, Loss: {'train/loss': 0.37995967268943787, 'train/grad_norm_pre_clip': 15.542428016662598, 'timing/data': 0.053876008838415146, 'timing/model': 1.0980121558532119, 'learning_rate/qwen_vl_interface': 7.19302865367098e-07, 'learning_rate/action_model': 7.19302865367098e-06, 'epoch': 2.44}) +[2026-06-18 07:59:23,674][starVLA.training.train_starvla][INFO] - Step 6331, Loss: {'train/loss': 0.0998043641448021, 'train/grad_norm_pre_clip': 9.22272777557373, 'timing/data': 0.053920043632388115, 'timing/model': 1.138637823984027, 'learning_rate/qwen_vl_interface': 7.186537840295599e-07, 'learning_rate/action_model': 7.186537840295598e-06, 'epoch': 2.44}) +[2026-06-18 07:59:29,065][starVLA.training.train_starvla][INFO] - Step 6332, Loss: {'train/loss': 0.2550067901611328, 'train/grad_norm_pre_clip': 12.387127876281738, 'timing/data': 0.05031559709459543, 'timing/model': 1.57138039637357, 'learning_rate/qwen_vl_interface': 7.18005642044969e-07, 'learning_rate/action_model': 7.180056420449689e-06, 'epoch': 2.44}) +[2026-06-18 07:59:33,822][starVLA.training.train_starvla][INFO] - Step 6333, Loss: {'train/loss': 0.16537396609783173, 'train/grad_norm_pre_clip': 10.98425579071045, 'timing/data': 0.04705125093460083, 'timing/model': 1.0814756015315652, 'learning_rate/qwen_vl_interface': 7.173584395476852e-07, 'learning_rate/action_model': 7.173584395476851e-06, 'epoch': 2.44}) +[2026-06-18 07:59:38,843][starVLA.training.train_starvla][INFO] - Step 6334, Loss: {'train/loss': 0.39834117889404297, 'train/grad_norm_pre_clip': 17.64529037475586, 'timing/data': 0.03936920687556267, 'timing/model': 1.0951254134997725, 'learning_rate/qwen_vl_interface': 7.16712176671874e-07, 'learning_rate/action_model': 7.16712176671874e-06, 'epoch': 2.44}) +[2026-06-18 07:59:48,077][starVLA.training.train_starvla][INFO] - Step 6335, Loss: {'train/loss': 0.31733906269073486, 'train/grad_norm_pre_clip': 13.69960880279541, 'timing/data': 0.0444433419033885, 'timing/model': 1.3646525191143155, 'learning_rate/qwen_vl_interface': 7.160668535515059e-07, 'learning_rate/action_model': 7.1606685355150585e-06, 'epoch': 2.44}) +[2026-06-18 08:00:03,637][starVLA.training.train_starvla][INFO] - Step 6336, Loss: {'train/loss': 0.30956438183784485, 'train/grad_norm_pre_clip': 15.420680046081543, 'timing/data': 0.03955759201198816, 'timing/model': 4.606408290565014, 'learning_rate/qwen_vl_interface': 7.154224703203584e-07, 'learning_rate/action_model': 7.154224703203584e-06, 'epoch': 2.44}) +[2026-06-18 08:00:21,133][starVLA.training.train_starvla][INFO] - Step 6337, Loss: {'train/loss': 0.4645119607448578, 'train/grad_norm_pre_clip': 21.3761043548584, 'timing/data': 0.10780456569045782, 'timing/model': 2.18591193202883, 'learning_rate/qwen_vl_interface': 7.147790271120122e-07, 'learning_rate/action_model': 7.147790271120122e-06, 'epoch': 2.44}) +[2026-06-18 08:00:27,780][starVLA.training.train_starvla][INFO] - Step 6338, Loss: {'train/loss': 0.44659408926963806, 'train/grad_norm_pre_clip': 18.12186050415039, 'timing/data': 0.048337701708078384, 'timing/model': 2.3051335215568542, 'learning_rate/qwen_vl_interface': 7.141365240598545e-07, 'learning_rate/action_model': 7.141365240598545e-06, 'epoch': 2.44}) +[2026-06-18 08:00:37,239][starVLA.training.train_starvla][INFO] - Step 6339, Loss: {'train/loss': 0.18187852203845978, 'train/grad_norm_pre_clip': 10.294050216674805, 'timing/data': 0.04321620799601078, 'timing/model': 2.8604893777519464, 'learning_rate/qwen_vl_interface': 7.134949612970749e-07, 'learning_rate/action_model': 7.134949612970748e-06, 'epoch': 2.44}) +[2026-06-18 08:00:52,660][starVLA.training.train_starvla][INFO] - Step 6340, Loss: {'train/loss': 0.1320820301771164, 'train/grad_norm_pre_clip': 15.762393951416016, 'timing/data': 0.15172749757766724, 'timing/model': 3.414150496944785, 'learning_rate/qwen_vl_interface': 7.128543389566718e-07, 'learning_rate/action_model': 7.128543389566718e-06, 'epoch': 2.44}) +[2026-06-18 08:00:59,648][starVLA.training.train_starvla][INFO] - Step 6341, Loss: {'train/loss': 0.17466822266578674, 'train/grad_norm_pre_clip': 12.897187232971191, 'timing/data': 0.013057167641818523, 'timing/model': 2.2686526076868176, 'learning_rate/qwen_vl_interface': 7.122146571714464e-07, 'learning_rate/action_model': 7.122146571714463e-06, 'epoch': 2.44}) +[2026-06-18 08:01:08,901][starVLA.training.train_starvla][INFO] - Step 6342, Loss: {'train/loss': 0.2346012443304062, 'train/grad_norm_pre_clip': 14.866930961608887, 'timing/data': 0.10332669876515865, 'timing/model': 1.8346677999943495, 'learning_rate/qwen_vl_interface': 7.115759160740058e-07, 'learning_rate/action_model': 7.115759160740057e-06, 'epoch': 2.44}) +[2026-06-18 08:01:19,363][starVLA.training.train_starvla][INFO] - Step 6343, Loss: {'train/loss': 0.5999943614006042, 'train/grad_norm_pre_clip': 20.013765335083008, 'timing/data': 0.04644524957984686, 'timing/model': 5.746117347851396, 'learning_rate/qwen_vl_interface': 7.109381157967607e-07, 'learning_rate/action_model': 7.109381157967607e-06, 'epoch': 2.44}) +[2026-06-18 08:01:24,442][starVLA.training.train_starvla][INFO] - Step 6344, Loss: {'train/loss': 0.23852330446243286, 'train/grad_norm_pre_clip': 14.199257850646973, 'timing/data': 0.05177305359393358, 'timing/model': 1.1277610817924142, 'learning_rate/qwen_vl_interface': 7.103012564719289e-07, 'learning_rate/action_model': 7.103012564719288e-06, 'epoch': 2.44}) +[2026-06-18 08:01:29,341][starVLA.training.train_starvla][INFO] - Step 6345, Loss: {'train/loss': 0.2769213020801544, 'train/grad_norm_pre_clip': 19.671194076538086, 'timing/data': 0.04690505098551512, 'timing/model': 1.1250794986262918, 'learning_rate/qwen_vl_interface': 7.096653382315315e-07, 'learning_rate/action_model': 7.096653382315315e-06, 'epoch': 2.44}) +[2026-06-18 08:01:34,078][starVLA.training.train_starvla][INFO] - Step 6346, Loss: {'train/loss': 0.14201398193836212, 'train/grad_norm_pre_clip': 8.527297973632812, 'timing/data': 0.04377937130630016, 'timing/model': 1.1483801621943712, 'learning_rate/qwen_vl_interface': 7.090303612073946e-07, 'learning_rate/action_model': 7.090303612073946e-06, 'epoch': 2.44}) +[2026-06-18 08:01:38,862][starVLA.training.train_starvla][INFO] - Step 6347, Loss: {'train/loss': 0.30039411783218384, 'train/grad_norm_pre_clip': 9.714200019836426, 'timing/data': 0.049915445037186146, 'timing/model': 1.4304717238992453, 'learning_rate/qwen_vl_interface': 7.083963255311508e-07, 'learning_rate/action_model': 7.083963255311509e-06, 'epoch': 2.44}) +[2026-06-18 08:01:43,233][starVLA.training.train_starvla][INFO] - Step 6348, Loss: {'train/loss': 0.37516847252845764, 'train/grad_norm_pre_clip': 16.457927703857422, 'timing/data': 0.052339665591716766, 'timing/model': 1.1289495695382357, 'learning_rate/qwen_vl_interface': 7.07763231334236e-07, 'learning_rate/action_model': 7.077632313342359e-06, 'epoch': 2.44}) +[2026-06-18 08:01:50,849][starVLA.training.train_starvla][INFO] - Step 6349, Loss: {'train/loss': 0.4458157420158386, 'train/grad_norm_pre_clip': 19.634714126586914, 'timing/data': 0.041676671244204044, 'timing/model': 4.063048772513866, 'learning_rate/qwen_vl_interface': 7.071310787478902e-07, 'learning_rate/action_model': 7.071310787478902e-06, 'epoch': 2.44}) +[2026-06-18 08:01:55,846][starVLA.training.train_starvla][INFO] - Step 6350, Loss: {'train/loss': 0.21366840600967407, 'train/grad_norm_pre_clip': 12.005789756774902, 'timing/data': 0.04359644651412964, 'timing/model': 1.3577578412368894, 'learning_rate/qwen_vl_interface': 7.0649986790316e-07, 'learning_rate/action_model': 7.0649986790315995e-06, 'epoch': 2.44}) +[2026-06-18 08:02:00,728][starVLA.training.train_starvla][INFO] - Step 6351, Loss: {'train/loss': 0.1452508270740509, 'train/grad_norm_pre_clip': 9.735589981079102, 'timing/data': 0.04474534187465906, 'timing/model': 1.168226396664977, 'learning_rate/qwen_vl_interface': 7.058695989308965e-07, 'learning_rate/action_model': 7.058695989308965e-06, 'epoch': 2.44}) +[2026-06-18 08:02:05,068][starVLA.training.train_starvla][INFO] - Step 6352, Loss: {'train/loss': 0.10780415683984756, 'train/grad_norm_pre_clip': 7.403589248657227, 'timing/data': 0.0394568108022213, 'timing/model': 1.183596696704626, 'learning_rate/qwen_vl_interface': 7.052402719617554e-07, 'learning_rate/action_model': 7.052402719617554e-06, 'epoch': 2.44}) +[2026-06-18 08:02:12,075][starVLA.training.train_starvla][INFO] - Step 6353, Loss: {'train/loss': 0.30177080631256104, 'train/grad_norm_pre_clip': 13.103431701660156, 'timing/data': 0.044954718090593815, 'timing/model': 1.8082847306504846, 'learning_rate/qwen_vl_interface': 7.046118871261952e-07, 'learning_rate/action_model': 7.0461188712619514e-06, 'epoch': 2.44}) +[2026-06-18 08:02:19,946][starVLA.training.train_starvla][INFO] - Step 6354, Loss: {'train/loss': 0.20254625380039215, 'train/grad_norm_pre_clip': 10.706930160522461, 'timing/data': 0.041895593516528606, 'timing/model': 1.1226402558386326, 'learning_rate/qwen_vl_interface': 7.039844445544828e-07, 'learning_rate/action_model': 7.039844445544827e-06, 'epoch': 2.44}) +[2026-06-18 08:02:24,758][starVLA.training.train_starvla][INFO] - Step 6355, Loss: {'train/loss': 0.2242773473262787, 'train/grad_norm_pre_clip': 12.816934585571289, 'timing/data': 0.05879886448383331, 'timing/model': 1.1082827197387815, 'learning_rate/qwen_vl_interface': 7.033579443766864e-07, 'learning_rate/action_model': 7.033579443766864e-06, 'epoch': 2.44}) +[2026-06-18 08:02:29,845][starVLA.training.train_starvla][INFO] - Step 6356, Loss: {'train/loss': 0.6765265464782715, 'train/grad_norm_pre_clip': 19.79887580871582, 'timing/data': 0.05145331285893917, 'timing/model': 1.1433496829122305, 'learning_rate/qwen_vl_interface': 7.027323867226807e-07, 'learning_rate/action_model': 7.0273238672268065e-06, 'epoch': 2.45}) +[2026-06-18 08:02:34,512][starVLA.training.train_starvla][INFO] - Step 6357, Loss: {'train/loss': 0.18121561408042908, 'train/grad_norm_pre_clip': 9.750571250915527, 'timing/data': 0.04400401096791029, 'timing/model': 1.420773595571518, 'learning_rate/qwen_vl_interface': 7.021077717221441e-07, 'learning_rate/action_model': 7.0210777172214415e-06, 'epoch': 2.45}) +[2026-06-18 08:02:39,043][starVLA.training.train_starvla][INFO] - Step 6358, Loss: {'train/loss': 0.23355425894260406, 'train/grad_norm_pre_clip': 11.599385261535645, 'timing/data': 0.04679825343191624, 'timing/model': 1.0980101628229022, 'learning_rate/qwen_vl_interface': 7.014840995045602e-07, 'learning_rate/action_model': 7.014840995045602e-06, 'epoch': 2.45}) +[2026-06-18 08:02:43,748][starVLA.training.train_starvla][INFO] - Step 6359, Loss: {'train/loss': 0.3548602759838104, 'train/grad_norm_pre_clip': 15.089045524597168, 'timing/data': 0.040677216835319996, 'timing/model': 1.0869768485426903, 'learning_rate/qwen_vl_interface': 7.008613701992172e-07, 'learning_rate/action_model': 7.008613701992171e-06, 'epoch': 2.45}) +[2026-06-18 08:02:49,213][starVLA.training.train_starvla][INFO] - Step 6360, Loss: {'train/loss': 0.09425196051597595, 'train/grad_norm_pre_clip': 6.757370948791504, 'timing/data': 0.04885287117213011, 'timing/model': 1.6178270243108273, 'learning_rate/qwen_vl_interface': 7.002395839352063e-07, 'learning_rate/action_model': 7.002395839352063e-06, 'epoch': 2.45}) +[2026-06-18 08:02:54,125][starVLA.training.train_starvla][INFO] - Step 6361, Loss: {'train/loss': 0.18632686138153076, 'train/grad_norm_pre_clip': 16.21266746520996, 'timing/data': 0.06162824109196663, 'timing/model': 1.103229658678174, 'learning_rate/qwen_vl_interface': 6.996187408414268e-07, 'learning_rate/action_model': 6.996187408414268e-06, 'epoch': 2.45}) +[2026-06-18 08:02:59,592][starVLA.training.train_starvla][INFO] - Step 6362, Loss: {'train/loss': 0.308160662651062, 'train/grad_norm_pre_clip': 14.066915512084961, 'timing/data': 0.040422530844807625, 'timing/model': 1.6257486585527658, 'learning_rate/qwen_vl_interface': 6.98998841046578e-07, 'learning_rate/action_model': 6.989988410465779e-06, 'epoch': 2.45}) +[2026-06-18 08:03:03,933][starVLA.training.train_starvla][INFO] - Step 6363, Loss: {'train/loss': 0.3962782919406891, 'train/grad_norm_pre_clip': 35.251930236816406, 'timing/data': 0.0444810651242733, 'timing/model': 1.1330210780724883, 'learning_rate/qwen_vl_interface': 6.983798846791667e-07, 'learning_rate/action_model': 6.983798846791667e-06, 'epoch': 2.45}) +[2026-06-18 08:03:08,839][starVLA.training.train_starvla][INFO] - Step 6364, Loss: {'train/loss': 0.1538599282503128, 'train/grad_norm_pre_clip': 11.182884216308594, 'timing/data': 0.040953490883111954, 'timing/model': 1.3505192873999476, 'learning_rate/qwen_vl_interface': 6.977618718675023e-07, 'learning_rate/action_model': 6.977618718675024e-06, 'epoch': 2.45}) +[2026-06-18 08:03:18,395][starVLA.training.train_starvla][INFO] - Step 6365, Loss: {'train/loss': 0.14671364426612854, 'train/grad_norm_pre_clip': 9.59083366394043, 'timing/data': 0.055134717375040054, 'timing/model': 4.54648554418236, 'learning_rate/qwen_vl_interface': 6.971448027397007e-07, 'learning_rate/action_model': 6.971448027397007e-06, 'epoch': 2.45}) +[2026-06-18 08:03:23,103][starVLA.training.train_starvla][INFO] - Step 6366, Loss: {'train/loss': 0.22229671478271484, 'train/grad_norm_pre_clip': 15.430410385131836, 'timing/data': 0.044944094493985176, 'timing/model': 1.0960027677938342, 'learning_rate/qwen_vl_interface': 6.965286774236814e-07, 'learning_rate/action_model': 6.965286774236813e-06, 'epoch': 2.45}) +[2026-06-18 08:03:27,812][starVLA.training.train_starvla][INFO] - Step 6367, Loss: {'train/loss': 0.391463041305542, 'train/grad_norm_pre_clip': 14.78746509552002, 'timing/data': 0.04711091425269842, 'timing/model': 1.4979041339829564, 'learning_rate/qwen_vl_interface': 6.959134960471664e-07, 'learning_rate/action_model': 6.959134960471664e-06, 'epoch': 2.45}) +[2026-06-18 08:03:32,344][starVLA.training.train_starvla][INFO] - Step 6368, Loss: {'train/loss': 0.19183628261089325, 'train/grad_norm_pre_clip': 10.721648216247559, 'timing/data': 0.046557086519896984, 'timing/model': 1.1149139907211065, 'learning_rate/qwen_vl_interface': 6.952992587376836e-07, 'learning_rate/action_model': 6.952992587376836e-06, 'epoch': 2.45}) +[2026-06-18 08:03:37,467][starVLA.training.train_starvla][INFO] - Step 6369, Loss: {'train/loss': 0.12278454005718231, 'train/grad_norm_pre_clip': 8.940376281738281, 'timing/data': 0.05912628397345543, 'timing/model': 1.20170936640352, 'learning_rate/qwen_vl_interface': 6.946859656225662e-07, 'learning_rate/action_model': 6.946859656225661e-06, 'epoch': 2.45}) +[2026-06-18 08:03:42,246][starVLA.training.train_starvla][INFO] - Step 6370, Loss: {'train/loss': 0.16710612177848816, 'train/grad_norm_pre_clip': 9.890926361083984, 'timing/data': 0.053219336085021496, 'timing/model': 1.1663101175799966, 'learning_rate/qwen_vl_interface': 6.940736168289497e-07, 'learning_rate/action_model': 6.940736168289497e-06, 'epoch': 2.45}) +[2026-06-18 08:03:47,207][starVLA.training.train_starvla][INFO] - Step 6371, Loss: {'train/loss': 0.3406580090522766, 'train/grad_norm_pre_clip': 15.631545066833496, 'timing/data': 0.04142905957996845, 'timing/model': 1.1933247176930308, 'learning_rate/qwen_vl_interface': 6.934622124837751e-07, 'learning_rate/action_model': 6.93462212483775e-06, 'epoch': 2.45}) +[2026-06-18 08:03:52,439][starVLA.training.train_starvla][INFO] - Step 6372, Loss: {'train/loss': 0.19836388528347015, 'train/grad_norm_pre_clip': 13.612882614135742, 'timing/data': 0.06335046980530024, 'timing/model': 1.4603252578526735, 'learning_rate/qwen_vl_interface': 6.928517527137864e-07, 'learning_rate/action_model': 6.9285175271378635e-06, 'epoch': 2.45}) +[2026-06-18 08:03:58,178][starVLA.training.train_starvla][INFO] - Step 6373, Loss: {'train/loss': 0.1679757684469223, 'train/grad_norm_pre_clip': 13.644073486328125, 'timing/data': 0.04554200265556574, 'timing/model': 2.4692953219637275, 'learning_rate/qwen_vl_interface': 6.922422376455333e-07, 'learning_rate/action_model': 6.9224223764553326e-06, 'epoch': 2.45}) +[2026-06-18 08:04:03,125][starVLA.training.train_starvla][INFO] - Step 6374, Loss: {'train/loss': 0.261947900056839, 'train/grad_norm_pre_clip': 14.909317970275879, 'timing/data': 0.04374799691140652, 'timing/model': 1.2249934924766421, 'learning_rate/qwen_vl_interface': 6.916336674053688e-07, 'learning_rate/action_model': 6.9163366740536875e-06, 'epoch': 2.45}) +[2026-06-18 08:04:08,060][starVLA.training.train_starvla][INFO] - Step 6375, Loss: {'train/loss': 0.47055479884147644, 'train/grad_norm_pre_clip': 19.32863426208496, 'timing/data': 0.0548610370606184, 'timing/model': 1.1305329985916615, 'learning_rate/qwen_vl_interface': 6.9102604211945e-07, 'learning_rate/action_model': 6.910260421194501e-06, 'epoch': 2.45}) +[2026-06-18 08:04:12,982][starVLA.training.train_starvla][INFO] - Step 6376, Loss: {'train/loss': 0.29516392946243286, 'train/grad_norm_pre_clip': 17.820690155029297, 'timing/data': 0.046940251253545284, 'timing/model': 1.0881278244778514, 'learning_rate/qwen_vl_interface': 6.904193619137381e-07, 'learning_rate/action_model': 6.904193619137381e-06, 'epoch': 2.45}) +[2026-06-18 08:04:17,879][starVLA.training.train_starvla][INFO] - Step 6377, Loss: {'train/loss': 0.20409533381462097, 'train/grad_norm_pre_clip': 18.789398193359375, 'timing/data': 0.04976764880120754, 'timing/model': 1.446449245326221, 'learning_rate/qwen_vl_interface': 6.898136269139989e-07, 'learning_rate/action_model': 6.8981362691399885e-06, 'epoch': 2.45}) +[2026-06-18 08:04:27,254][starVLA.training.train_starvla][INFO] - Step 6378, Loss: {'train/loss': 0.3609437942504883, 'train/grad_norm_pre_clip': 15.097904205322266, 'timing/data': 0.0686911977827549, 'timing/model': 1.3937454195693135, 'learning_rate/qwen_vl_interface': 6.89208837245802e-07, 'learning_rate/action_model': 6.892088372458019e-06, 'epoch': 2.45}) +[2026-06-18 08:04:33,612][starVLA.training.train_starvla][INFO] - Step 6379, Loss: {'train/loss': 0.34476929903030396, 'train/grad_norm_pre_clip': 20.150569915771484, 'timing/data': 0.044870403595268726, 'timing/model': 1.1538359578698874, 'learning_rate/qwen_vl_interface': 6.886049930345204e-07, 'learning_rate/action_model': 6.8860499303452035e-06, 'epoch': 2.45}) +[2026-06-18 08:04:43,156][starVLA.training.train_starvla][INFO] - Step 6380, Loss: {'train/loss': 0.25796639919281006, 'train/grad_norm_pre_clip': 15.229515075683594, 'timing/data': 0.04596993699669838, 'timing/model': 1.1101836049929261, 'learning_rate/qwen_vl_interface': 6.880020944053326e-07, 'learning_rate/action_model': 6.880020944053325e-06, 'epoch': 2.45}) +[2026-06-18 08:04:48,439][starVLA.training.train_starvla][INFO] - Step 6381, Loss: {'train/loss': 0.232070192694664, 'train/grad_norm_pre_clip': 11.97120475769043, 'timing/data': 0.04657372925430536, 'timing/model': 1.3400602461770177, 'learning_rate/qwen_vl_interface': 6.87400141483219e-07, 'learning_rate/action_model': 6.87400141483219e-06, 'epoch': 2.45}) +[2026-06-18 08:04:53,160][starVLA.training.train_starvla][INFO] - Step 6382, Loss: {'train/loss': 0.32443878054618835, 'train/grad_norm_pre_clip': 14.698103904724121, 'timing/data': 0.0469477828592062, 'timing/model': 1.472007429227233, 'learning_rate/qwen_vl_interface': 6.867991343929655e-07, 'learning_rate/action_model': 6.867991343929656e-06, 'epoch': 2.46}) +[2026-06-18 08:04:57,892][starVLA.training.train_starvla][INFO] - Step 6383, Loss: {'train/loss': 0.5479305386543274, 'train/grad_norm_pre_clip': 15.635992050170898, 'timing/data': 0.04866441898047924, 'timing/model': 1.1444784495979548, 'learning_rate/qwen_vl_interface': 6.861990732591622e-07, 'learning_rate/action_model': 6.861990732591622e-06, 'epoch': 2.46}) +[2026-06-18 08:05:02,861][starVLA.training.train_starvla][INFO] - Step 6384, Loss: {'train/loss': 0.33141398429870605, 'train/grad_norm_pre_clip': 16.596742630004883, 'timing/data': 0.044605644419789314, 'timing/model': 1.1268099127337337, 'learning_rate/qwen_vl_interface': 6.855999582062017e-07, 'learning_rate/action_model': 6.855999582062017e-06, 'epoch': 2.46}) +[2026-06-18 08:05:08,876][starVLA.training.train_starvla][INFO] - Step 6385, Loss: {'train/loss': 0.1850701868534088, 'train/grad_norm_pre_clip': 9.109976768493652, 'timing/data': 0.050244709476828575, 'timing/model': 1.2504907939583063, 'learning_rate/qwen_vl_interface': 6.850017893582818e-07, 'learning_rate/action_model': 6.850017893582818e-06, 'epoch': 2.46}) +[2026-06-18 08:05:13,647][starVLA.training.train_starvla][INFO] - Step 6386, Loss: {'train/loss': 0.20384962856769562, 'train/grad_norm_pre_clip': 9.644009590148926, 'timing/data': 0.04619574639946222, 'timing/model': 1.1644897926598787, 'learning_rate/qwen_vl_interface': 6.844045668394024e-07, 'learning_rate/action_model': 6.844045668394024e-06, 'epoch': 2.46}) +[2026-06-18 08:05:18,790][starVLA.training.train_starvla][INFO] - Step 6387, Loss: {'train/loss': 0.25347819924354553, 'train/grad_norm_pre_clip': 19.57994270324707, 'timing/data': 0.04376822616904974, 'timing/model': 1.8005657270550728, 'learning_rate/qwen_vl_interface': 6.838082907733694e-07, 'learning_rate/action_model': 6.838082907733694e-06, 'epoch': 2.46}) +[2026-06-18 08:05:31,230][starVLA.training.train_starvla][INFO] - Step 6388, Loss: {'train/loss': 0.428646057844162, 'train/grad_norm_pre_clip': 15.282527923583984, 'timing/data': 0.04456325713545084, 'timing/model': 1.3471000725403428, 'learning_rate/qwen_vl_interface': 6.832129612837912e-07, 'learning_rate/action_model': 6.832129612837912e-06, 'epoch': 2.46}) +[2026-06-18 08:05:36,070][starVLA.training.train_starvla][INFO] - Step 6389, Loss: {'train/loss': 0.3119449317455292, 'train/grad_norm_pre_clip': 16.20598602294922, 'timing/data': 0.047999536618590355, 'timing/model': 1.1527653532102704, 'learning_rate/qwen_vl_interface': 6.826185784940801e-07, 'learning_rate/action_model': 6.826185784940801e-06, 'epoch': 2.46}) +[2026-06-18 08:05:41,822][starVLA.training.train_starvla][INFO] - Step 6390, Loss: {'train/loss': 0.07989423722028732, 'train/grad_norm_pre_clip': 5.498863697052002, 'timing/data': 0.0503033297136426, 'timing/model': 1.5252778325229883, 'learning_rate/qwen_vl_interface': 6.820251425274526e-07, 'learning_rate/action_model': 6.820251425274527e-06, 'epoch': 2.46}) +[2026-06-18 08:05:46,477][starVLA.training.train_starvla][INFO] - Step 6391, Loss: {'train/loss': 0.3620348572731018, 'train/grad_norm_pre_clip': 17.23613166809082, 'timing/data': 0.04333908390253782, 'timing/model': 1.0943263452500105, 'learning_rate/qwen_vl_interface': 6.81432653506928e-07, 'learning_rate/action_model': 6.814326535069279e-06, 'epoch': 2.46}) +[2026-06-18 08:05:51,891][starVLA.training.train_starvla][INFO] - Step 6392, Loss: {'train/loss': 0.47195473313331604, 'train/grad_norm_pre_clip': 15.877599716186523, 'timing/data': 0.044975029304623604, 'timing/model': 1.6154931979253888, 'learning_rate/qwen_vl_interface': 6.808411115553305e-07, 'learning_rate/action_model': 6.808411115553305e-06, 'epoch': 2.46}) +[2026-06-18 08:05:57,048][starVLA.training.train_starvla][INFO] - Step 6393, Loss: {'train/loss': 0.4468769431114197, 'train/grad_norm_pre_clip': 13.470023155212402, 'timing/data': 0.051140681840479374, 'timing/model': 1.678360684774816, 'learning_rate/qwen_vl_interface': 6.802505167952871e-07, 'learning_rate/action_model': 6.80250516795287e-06, 'epoch': 2.46}) +[2026-06-18 08:06:05,831][starVLA.training.train_starvla][INFO] - Step 6394, Loss: {'train/loss': 0.43563324213027954, 'train/grad_norm_pre_clip': 20.827905654907227, 'timing/data': 0.05583416298031807, 'timing/model': 1.4062024680897593, 'learning_rate/qwen_vl_interface': 6.796608693492293e-07, 'learning_rate/action_model': 6.7966086934922925e-06, 'epoch': 2.46}) +[2026-06-18 08:06:13,563][starVLA.training.train_starvla][INFO] - Step 6395, Loss: {'train/loss': 0.27103158831596375, 'train/grad_norm_pre_clip': 15.662755966186523, 'timing/data': 0.039140306413173676, 'timing/model': 1.1638233177363873, 'learning_rate/qwen_vl_interface': 6.790721693393912e-07, 'learning_rate/action_model': 6.790721693393912e-06, 'epoch': 2.46}) +[2026-06-18 08:06:18,742][starVLA.training.train_starvla][INFO] - Step 6396, Loss: {'train/loss': 0.23452499508857727, 'train/grad_norm_pre_clip': 13.04237174987793, 'timing/data': 0.046235148794949055, 'timing/model': 1.266509309411049, 'learning_rate/qwen_vl_interface': 6.784844168878108e-07, 'learning_rate/action_model': 6.784844168878107e-06, 'epoch': 2.46}) +[2026-06-18 08:06:23,787][starVLA.training.train_starvla][INFO] - Step 6397, Loss: {'train/loss': 0.2927422523498535, 'train/grad_norm_pre_clip': 15.960816383361816, 'timing/data': 0.04742822423577309, 'timing/model': 1.6148585844784975, 'learning_rate/qwen_vl_interface': 6.778976121163289e-07, 'learning_rate/action_model': 6.778976121163288e-06, 'epoch': 2.46}) +[2026-06-18 08:06:28,686][starVLA.training.train_starvla][INFO] - Step 6398, Loss: {'train/loss': 0.10165686905384064, 'train/grad_norm_pre_clip': 8.066431045532227, 'timing/data': 0.0489548584446311, 'timing/model': 1.4059546738862991, 'learning_rate/qwen_vl_interface': 6.773117551465927e-07, 'learning_rate/action_model': 6.773117551465927e-06, 'epoch': 2.46}) +[2026-06-18 08:06:39,084][starVLA.training.train_starvla][INFO] - Step 6399, Loss: {'train/loss': 0.569614589214325, 'train/grad_norm_pre_clip': 23.800296783447266, 'timing/data': 0.050331950187683105, 'timing/model': 5.949856637977064, 'learning_rate/qwen_vl_interface': 6.767268461000498e-07, 'learning_rate/action_model': 6.7672684610004975e-06, 'epoch': 2.46}) +[2026-06-18 08:09:54,405][starVLA.training.train_starvla][INFO] - Step 6400, Loss: {'train/loss': 0.20523613691329956, 'train/grad_norm_pre_clip': 13.894756317138672, 'eval/loss': 0.422702819108963, 'eval/latency_4/loss': 0.422702819108963, 'eval/demon_fire_precision': 0.9674121405750798, 'eval/demon_fire_recall': 0.9834361805781098, 'eval/demon_fire_f1': 0.9753583507811242, 'eval/demon_fire_rate_model': 0.9192364170337739, 'eval/demon_fire_rate_teacher': 0.9042584434654919, 'eval/demon_left_precision': 0.8164383561643835, 'eval/demon_left_recall': 0.680365296803653, 'eval/demon_left_f1': 0.742216687422167, 'eval/demon_left_rate_model': 0.10719530102790015, 'eval/demon_left_rate_teacher': 0.12863436123348018, 'eval/demon_right_precision': 0.8048327137546468, 'eval/demon_right_recall': 0.7815884476534296, 'eval/demon_right_f1': 0.7930402930402931, 'eval/demon_right_rate_model': 0.15800293685756242, 'eval/demon_right_rate_teacher': 0.16270190895741557, 'eval/demon_move_f1': 0.76762849023123, 'eval/demon_cc_f1': 0.9021208025777533, 'eval/latency_4/demon_fire_precision': 0.9674121405750798, 'eval/latency_4/demon_fire_recall': 0.9834361805781098, 'eval/latency_4/demon_fire_f1': 0.9753583507811242, 'eval/latency_4/demon_fire_rate_model': 0.9192364170337739, 'eval/latency_4/demon_fire_rate_teacher': 0.9042584434654919, 'eval/latency_4/demon_left_precision': 0.8164383561643835, 'eval/latency_4/demon_left_recall': 0.680365296803653, 'eval/latency_4/demon_left_f1': 0.742216687422167, 'eval/latency_4/demon_left_rate_model': 0.10719530102790015, 'eval/latency_4/demon_left_rate_teacher': 0.12863436123348018, 'eval/latency_4/demon_right_precision': 0.8048327137546468, 'eval/latency_4/demon_right_recall': 0.7815884476534296, 'eval/latency_4/demon_right_f1': 0.7930402930402931, 'eval/latency_4/demon_right_rate_model': 0.15800293685756242, 'eval/latency_4/demon_right_rate_teacher': 0.16270190895741557, 'eval/latency_4/demon_move_f1': 0.76762849023123, 'eval/latency_4/demon_cc_f1': 0.9021208025777533, 'timing/data': 0.04454027954488993, 'timing/model': 4.587246057577431, 'learning_rate/qwen_vl_interface': 6.761428850979542e-07, 'learning_rate/action_model': 6.7614288509795415e-06, 'epoch': 2.46}) +[2026-06-18 08:09:54,442][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_6400_state +[2026-06-18 08:09:54,448][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 08:10:02,426][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_6400_state/model.safetensors +[2026-06-18 08:10:26,100][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_6400_state/optimizer.bin +[2026-06-18 08:10:26,101][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_6400_state/sampler.bin +[2026-06-18 08:10:26,101][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_6400_state/sampler_1.bin +[2026-06-18 08:10:26,104][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_6400_state/random_states_0.pkl +[2026-06-18 08:10:26,115][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 6400 +[2026-06-18 08:10:26,115][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 08:10:26,133][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 08:10:31,308][starVLA.training.train_starvla][INFO] - Step 6401, Loss: {'train/loss': 0.1778889000415802, 'train/grad_norm_pre_clip': 13.533695220947266, 'timing/data': 0.04585928190499544, 'timing/model': 1.182906330563128, 'learning_rate/qwen_vl_interface': 6.755598722613579e-07, 'learning_rate/action_model': 6.755598722613579e-06, 'epoch': 2.46}) +[2026-06-18 08:10:36,049][starVLA.training.train_starvla][INFO] - Step 6402, Loss: {'train/loss': 0.4277713894844055, 'train/grad_norm_pre_clip': 17.559846878051758, 'timing/data': 0.06022180989384651, 'timing/model': 1.510116477496922, 'learning_rate/qwen_vl_interface': 6.749778077111241e-07, 'learning_rate/action_model': 6.749778077111241e-06, 'epoch': 2.46}) +[2026-06-18 08:10:40,639][starVLA.training.train_starvla][INFO] - Step 6403, Loss: {'train/loss': 0.11455640196800232, 'train/grad_norm_pre_clip': 8.549243927001953, 'timing/data': 0.04695842228829861, 'timing/model': 1.1515415338799357, 'learning_rate/qwen_vl_interface': 6.743966915679128e-07, 'learning_rate/action_model': 6.743966915679129e-06, 'epoch': 2.46}) +[2026-06-18 08:10:46,414][starVLA.training.train_starvla][INFO] - Step 6404, Loss: {'train/loss': 0.1844855397939682, 'train/grad_norm_pre_clip': 13.840262413024902, 'timing/data': 0.045854005962610245, 'timing/model': 2.005911121144891, 'learning_rate/qwen_vl_interface': 6.738165239521909e-07, 'learning_rate/action_model': 6.7381652395219094e-06, 'epoch': 2.46}) +[2026-06-18 08:10:51,228][starVLA.training.train_starvla][INFO] - Step 6405, Loss: {'train/loss': 0.40747785568237305, 'train/grad_norm_pre_clip': 17.854429244995117, 'timing/data': 0.04242758546024561, 'timing/model': 1.113828214816749, 'learning_rate/qwen_vl_interface': 6.732373049842284e-07, 'learning_rate/action_model': 6.7323730498422846e-06, 'epoch': 2.46}) +[2026-06-18 08:10:58,574][starVLA.training.train_starvla][INFO] - Step 6406, Loss: {'train/loss': 0.18534520268440247, 'train/grad_norm_pre_clip': 14.9570894241333, 'timing/data': 0.04511528927832842, 'timing/model': 1.1073539312928915, 'learning_rate/qwen_vl_interface': 6.726590347840972e-07, 'learning_rate/action_model': 6.726590347840972e-06, 'epoch': 2.46}) +[2026-06-18 08:11:03,439][starVLA.training.train_starvla][INFO] - Step 6407, Loss: {'train/loss': 0.20718427002429962, 'train/grad_norm_pre_clip': 17.99654769897461, 'timing/data': 0.043295145966112614, 'timing/model': 1.4613512977957726, 'learning_rate/qwen_vl_interface': 6.720817134716736e-07, 'learning_rate/action_model': 6.720817134716736e-06, 'epoch': 2.46}) +[2026-06-18 08:11:07,868][starVLA.training.train_starvla][INFO] - Step 6408, Loss: {'train/loss': 0.292892724275589, 'train/grad_norm_pre_clip': 23.86624526977539, 'timing/data': 0.04052051994949579, 'timing/model': 1.09128921572119, 'learning_rate/qwen_vl_interface': 6.71505341166636e-07, 'learning_rate/action_model': 6.71505341166636e-06, 'epoch': 2.47}) +[2026-06-18 08:11:14,207][starVLA.training.train_starvla][INFO] - Step 6409, Loss: {'train/loss': 0.20845085382461548, 'train/grad_norm_pre_clip': 10.692995071411133, 'timing/data': 0.05283671524375677, 'timing/model': 1.274201818741858, 'learning_rate/qwen_vl_interface': 6.709299179884689e-07, 'learning_rate/action_model': 6.709299179884689e-06, 'epoch': 2.47}) +[2026-06-18 08:11:21,135][starVLA.training.train_starvla][INFO] - Step 6410, Loss: {'train/loss': 0.7711473107337952, 'train/grad_norm_pre_clip': 21.756990432739258, 'timing/data': 0.040960706770420074, 'timing/model': 1.1200322210788727, 'learning_rate/qwen_vl_interface': 6.703554440564569e-07, 'learning_rate/action_model': 6.703554440564569e-06, 'epoch': 2.47}) +[2026-06-18 08:11:25,882][starVLA.training.train_starvla][INFO] - Step 6411, Loss: {'train/loss': 0.4270975887775421, 'train/grad_norm_pre_clip': 15.309328079223633, 'timing/data': 0.04965443257242441, 'timing/model': 1.1284105749800801, 'learning_rate/qwen_vl_interface': 6.697819194896889e-07, 'learning_rate/action_model': 6.697819194896889e-06, 'epoch': 2.47}) +[2026-06-18 08:11:33,983][starVLA.training.train_starvla][INFO] - Step 6412, Loss: {'train/loss': 0.18990182876586914, 'train/grad_norm_pre_clip': 19.72620964050293, 'timing/data': 0.04753364156931639, 'timing/model': 1.5409722719341516, 'learning_rate/qwen_vl_interface': 6.692093444070571e-07, 'learning_rate/action_model': 6.692093444070572e-06, 'epoch': 2.47}) +[2026-06-18 08:11:38,458][starVLA.training.train_starvla][INFO] - Step 6413, Loss: {'train/loss': 0.09045306593179703, 'train/grad_norm_pre_clip': 7.143528938293457, 'timing/data': 0.08074757643043995, 'timing/model': 1.1555232387036085, 'learning_rate/qwen_vl_interface': 6.686377189272584e-07, 'learning_rate/action_model': 6.686377189272583e-06, 'epoch': 2.47}) +[2026-06-18 08:11:43,754][starVLA.training.train_starvla][INFO] - Step 6414, Loss: {'train/loss': 0.24668344855308533, 'train/grad_norm_pre_clip': 12.030130386352539, 'timing/data': 0.061578935012221336, 'timing/model': 1.2817591857165098, 'learning_rate/qwen_vl_interface': 6.680670431687898e-07, 'learning_rate/action_model': 6.680670431687898e-06, 'epoch': 2.47}) +[2026-06-18 08:11:53,166][starVLA.training.train_starvla][INFO] - Step 6415, Loss: {'train/loss': 0.3628767430782318, 'train/grad_norm_pre_clip': 19.55777359008789, 'timing/data': 0.06360907293856144, 'timing/model': 1.2322087222710252, 'learning_rate/qwen_vl_interface': 6.674973172499536e-07, 'learning_rate/action_model': 6.674973172499535e-06, 'epoch': 2.47}) +[2026-06-18 08:11:58,033][starVLA.training.train_starvla][INFO] - Step 6416, Loss: {'train/loss': 0.6097263693809509, 'train/grad_norm_pre_clip': 18.96255111694336, 'timing/data': 0.06012027524411678, 'timing/model': 1.145543402992189, 'learning_rate/qwen_vl_interface': 6.669285412888544e-07, 'learning_rate/action_model': 6.669285412888544e-06, 'epoch': 2.47}) +[2026-06-18 08:12:02,871][starVLA.training.train_starvla][INFO] - Step 6417, Loss: {'train/loss': 0.22434745728969574, 'train/grad_norm_pre_clip': 12.013026237487793, 'timing/data': 0.04274826217442751, 'timing/model': 1.5148358643054962, 'learning_rate/qwen_vl_interface': 6.663607154034003e-07, 'learning_rate/action_model': 6.663607154034003e-06, 'epoch': 2.47}) +[2026-06-18 08:12:07,278][starVLA.training.train_starvla][INFO] - Step 6418, Loss: {'train/loss': 0.23224222660064697, 'train/grad_norm_pre_clip': 11.47377872467041, 'timing/data': 0.05136504862457514, 'timing/model': 1.1817635921761394, 'learning_rate/qwen_vl_interface': 6.657938397113019e-07, 'learning_rate/action_model': 6.657938397113018e-06, 'epoch': 2.47}) +[2026-06-18 08:12:12,218][starVLA.training.train_starvla][INFO] - Step 6419, Loss: {'train/loss': 0.25255724787712097, 'train/grad_norm_pre_clip': 11.055158615112305, 'timing/data': 0.04771112930029631, 'timing/model': 1.182972639799118, 'learning_rate/qwen_vl_interface': 6.652279143300734e-07, 'learning_rate/action_model': 6.652279143300734e-06, 'epoch': 2.47}) +[2026-06-18 08:12:17,385][starVLA.training.train_starvla][INFO] - Step 6420, Loss: {'train/loss': 0.3760842978954315, 'train/grad_norm_pre_clip': 19.711368560791016, 'timing/data': 0.04567009210586548, 'timing/model': 1.1966788535937667, 'learning_rate/qwen_vl_interface': 6.64662939377032e-07, 'learning_rate/action_model': 6.646629393770321e-06, 'epoch': 2.47}) +[2026-06-18 08:12:22,243][starVLA.training.train_starvla][INFO] - Step 6421, Loss: {'train/loss': 0.21408690512180328, 'train/grad_norm_pre_clip': 12.802087783813477, 'timing/data': 0.04289344139397144, 'timing/model': 1.0895103020593524, 'learning_rate/qwen_vl_interface': 6.64098914969297e-07, 'learning_rate/action_model': 6.6409891496929695e-06, 'epoch': 2.47}) +[2026-06-18 08:12:27,263][starVLA.training.train_starvla][INFO] - Step 6422, Loss: {'train/loss': 0.45205751061439514, 'train/grad_norm_pre_clip': 19.453784942626953, 'timing/data': 0.04440726898610592, 'timing/model': 1.5998042225837708, 'learning_rate/qwen_vl_interface': 6.635358412237912e-07, 'learning_rate/action_model': 6.635358412237912e-06, 'epoch': 2.47}) +[2026-06-18 08:12:31,789][starVLA.training.train_starvla][INFO] - Step 6423, Loss: {'train/loss': 0.30525434017181396, 'train/grad_norm_pre_clip': 15.112968444824219, 'timing/data': 0.05441777966916561, 'timing/model': 1.1146788755431771, 'learning_rate/qwen_vl_interface': 6.629737182572418e-07, 'learning_rate/action_model': 6.629737182572418e-06, 'epoch': 2.47}) +[2026-06-18 08:12:36,736][starVLA.training.train_starvla][INFO] - Step 6424, Loss: {'train/loss': 0.1856127828359604, 'train/grad_norm_pre_clip': 10.05051326751709, 'timing/data': 0.051608811132609844, 'timing/model': 1.157589172013104, 'learning_rate/qwen_vl_interface': 6.62412546186176e-07, 'learning_rate/action_model': 6.62412546186176e-06, 'epoch': 2.47}) +[2026-06-18 08:12:41,531][starVLA.training.train_starvla][INFO] - Step 6425, Loss: {'train/loss': 0.31203407049179077, 'train/grad_norm_pre_clip': 17.018911361694336, 'timing/data': 0.046085567213594913, 'timing/model': 1.09829072188586, 'learning_rate/qwen_vl_interface': 6.61852325126926e-07, 'learning_rate/action_model': 6.61852325126926e-06, 'epoch': 2.47}) +[2026-06-18 08:12:46,373][starVLA.training.train_starvla][INFO] - Step 6426, Loss: {'train/loss': 0.3850446045398712, 'train/grad_norm_pre_clip': 17.232969284057617, 'timing/data': 0.04418411385267973, 'timing/model': 1.1067758845165372, 'learning_rate/qwen_vl_interface': 6.61293055195625e-07, 'learning_rate/action_model': 6.6129305519562504e-06, 'epoch': 2.47}) +[2026-06-18 08:12:51,366][starVLA.training.train_starvla][INFO] - Step 6427, Loss: {'train/loss': 0.36751750111579895, 'train/grad_norm_pre_clip': 28.24339485168457, 'timing/data': 0.05003323592245579, 'timing/model': 1.6303926492109895, 'learning_rate/qwen_vl_interface': 6.607347365082127e-07, 'learning_rate/action_model': 6.607347365082127e-06, 'epoch': 2.47}) +[2026-06-18 08:12:55,655][starVLA.training.train_starvla][INFO] - Step 6428, Loss: {'train/loss': 0.29062941670417786, 'train/grad_norm_pre_clip': 17.437442779541016, 'timing/data': 0.03910450544208288, 'timing/model': 1.0912861516699195, 'learning_rate/qwen_vl_interface': 6.601773691804273e-07, 'learning_rate/action_model': 6.6017736918042735e-06, 'epoch': 2.47}) +[2026-06-18 08:13:00,516][starVLA.training.train_starvla][INFO] - Step 6429, Loss: {'train/loss': 0.1845582127571106, 'train/grad_norm_pre_clip': 14.740490913391113, 'timing/data': 0.051522535271942616, 'timing/model': 1.1073685893788934, 'learning_rate/qwen_vl_interface': 6.596209533278122e-07, 'learning_rate/action_model': 6.596209533278122e-06, 'epoch': 2.47}) +[2026-06-18 08:13:05,428][starVLA.training.train_starvla][INFO] - Step 6430, Loss: {'train/loss': 0.16093523800373077, 'train/grad_norm_pre_clip': 10.494610786437988, 'timing/data': 0.043808821588754654, 'timing/model': 1.0849071890115738, 'learning_rate/qwen_vl_interface': 6.590654890657125e-07, 'learning_rate/action_model': 6.590654890657125e-06, 'epoch': 2.47}) +[2026-06-18 08:13:10,333][starVLA.training.train_starvla][INFO] - Step 6431, Loss: {'train/loss': 0.25517547130584717, 'train/grad_norm_pre_clip': 12.773873329162598, 'timing/data': 0.044118392281234264, 'timing/model': 1.1151419440284371, 'learning_rate/qwen_vl_interface': 6.585109765092772e-07, 'learning_rate/action_model': 6.585109765092772e-06, 'epoch': 2.47}) +[2026-06-18 08:13:15,265][starVLA.training.train_starvla][INFO] - Step 6432, Loss: {'train/loss': 0.2908588945865631, 'train/grad_norm_pre_clip': 14.526026725769043, 'timing/data': 0.07910490781068802, 'timing/model': 1.4749910412356257, 'learning_rate/qwen_vl_interface': 6.579574157734575e-07, 'learning_rate/action_model': 6.579574157734575e-06, 'epoch': 2.47}) +[2026-06-18 08:13:21,223][starVLA.training.train_starvla][INFO] - Step 6433, Loss: {'train/loss': 0.39708295464515686, 'train/grad_norm_pre_clip': 16.715330123901367, 'timing/data': 0.0418263953179121, 'timing/model': 1.0903658550232649, 'learning_rate/qwen_vl_interface': 6.574048069730062e-07, 'learning_rate/action_model': 6.574048069730063e-06, 'epoch': 2.47}) +[2026-06-18 08:13:25,842][starVLA.training.train_starvla][INFO] - Step 6434, Loss: {'train/loss': 0.2884283661842346, 'train/grad_norm_pre_clip': 22.159297943115234, 'timing/data': 0.044415674172341824, 'timing/model': 1.0949843749403954, 'learning_rate/qwen_vl_interface': 6.568531502224808e-07, 'learning_rate/action_model': 6.568531502224808e-06, 'epoch': 2.48}) +[2026-06-18 08:13:35,143][starVLA.training.train_starvla][INFO] - Step 6435, Loss: {'train/loss': 0.16113418340682983, 'train/grad_norm_pre_clip': 11.244779586791992, 'timing/data': 0.054758199490606785, 'timing/model': 2.424344900995493, 'learning_rate/qwen_vl_interface': 6.563024456362395e-07, 'learning_rate/action_model': 6.563024456362395e-06, 'epoch': 2.48}) +[2026-06-18 08:13:39,984][starVLA.training.train_starvla][INFO] - Step 6436, Loss: {'train/loss': 0.3600834012031555, 'train/grad_norm_pre_clip': 17.589134216308594, 'timing/data': 0.07120653055608273, 'timing/model': 1.105054758489132, 'learning_rate/qwen_vl_interface': 6.557526933284443e-07, 'learning_rate/action_model': 6.557526933284443e-06, 'epoch': 2.48}) +[2026-06-18 08:13:45,780][starVLA.training.train_starvla][INFO] - Step 6437, Loss: {'train/loss': 0.1736394464969635, 'train/grad_norm_pre_clip': 8.688450813293457, 'timing/data': 0.04221384134143591, 'timing/model': 2.417478277347982, 'learning_rate/qwen_vl_interface': 6.552038934130595e-07, 'learning_rate/action_model': 6.552038934130594e-06, 'epoch': 2.48}) +[2026-06-18 08:13:50,237][starVLA.training.train_starvla][INFO] - Step 6438, Loss: {'train/loss': 0.28365036845207214, 'train/grad_norm_pre_clip': 15.937370300292969, 'timing/data': 0.04475458525121212, 'timing/model': 1.0770383030176163, 'learning_rate/qwen_vl_interface': 6.546560460038514e-07, 'learning_rate/action_model': 6.546560460038514e-06, 'epoch': 2.48}) +[2026-06-18 08:13:54,870][starVLA.training.train_starvla][INFO] - Step 6439, Loss: {'train/loss': 0.24290183186531067, 'train/grad_norm_pre_clip': 13.553258895874023, 'timing/data': 0.05054119136184454, 'timing/model': 1.127235067076981, 'learning_rate/qwen_vl_interface': 6.5410915121439e-07, 'learning_rate/action_model': 6.5410915121439e-06, 'epoch': 2.48}) +[2026-06-18 08:13:59,994][starVLA.training.train_starvla][INFO] - Step 6440, Loss: {'train/loss': 0.19469383358955383, 'train/grad_norm_pre_clip': 9.486900329589844, 'timing/data': 0.06092526111751795, 'timing/model': 1.1070025339722633, 'learning_rate/qwen_vl_interface': 6.535632091580474e-07, 'learning_rate/action_model': 6.535632091580474e-06, 'epoch': 2.48}) +[2026-06-18 08:14:04,734][starVLA.training.train_starvla][INFO] - Step 6441, Loss: {'train/loss': 0.20145542919635773, 'train/grad_norm_pre_clip': 13.931318283081055, 'timing/data': 0.04026769660413265, 'timing/model': 1.1467578709125519, 'learning_rate/qwen_vl_interface': 6.530182199479965e-07, 'learning_rate/action_model': 6.5301821994799645e-06, 'epoch': 2.48}) +[2026-06-18 08:14:09,355][starVLA.training.train_starvla][INFO] - Step 6442, Loss: {'train/loss': 0.16147226095199585, 'train/grad_norm_pre_clip': 11.909964561462402, 'timing/data': 0.04367354419082403, 'timing/model': 1.4211398903280497, 'learning_rate/qwen_vl_interface': 6.524741836972162e-07, 'learning_rate/action_model': 6.524741836972162e-06, 'epoch': 2.48}) +[2026-06-18 08:14:14,043][starVLA.training.train_starvla][INFO] - Step 6443, Loss: {'train/loss': 0.28849363327026367, 'train/grad_norm_pre_clip': 15.788421630859375, 'timing/data': 0.04131922125816345, 'timing/model': 1.120665586553514, 'learning_rate/qwen_vl_interface': 6.519311005184848e-07, 'learning_rate/action_model': 6.5193110051848476e-06, 'epoch': 2.48}) +[2026-06-18 08:14:18,652][starVLA.training.train_starvla][INFO] - Step 6444, Loss: {'train/loss': 0.4826464056968689, 'train/grad_norm_pre_clip': 19.85431480407715, 'timing/data': 0.050022855401039124, 'timing/model': 1.0825277511030436, 'learning_rate/qwen_vl_interface': 6.51388970524383e-07, 'learning_rate/action_model': 6.513889705243829e-06, 'epoch': 2.48}) +[2026-06-18 08:14:23,251][starVLA.training.train_starvla][INFO] - Step 6445, Loss: {'train/loss': 0.35027942061424255, 'train/grad_norm_pre_clip': 19.38401985168457, 'timing/data': 0.03963243681937456, 'timing/model': 1.1412646044045687, 'learning_rate/qwen_vl_interface': 6.508477938272963e-07, 'learning_rate/action_model': 6.508477938272962e-06, 'epoch': 2.48}) +[2026-06-18 08:14:28,035][starVLA.training.train_starvla][INFO] - Step 6446, Loss: {'train/loss': 0.4319446086883545, 'train/grad_norm_pre_clip': 15.262679100036621, 'timing/data': 0.0490862587466836, 'timing/model': 1.151111057959497, 'learning_rate/qwen_vl_interface': 6.503075705394113e-07, 'learning_rate/action_model': 6.5030757053941125e-06, 'epoch': 2.48}) +[2026-06-18 08:14:32,729][starVLA.training.train_starvla][INFO] - Step 6447, Loss: {'train/loss': 0.14003731310367584, 'train/grad_norm_pre_clip': 9.156249046325684, 'timing/data': 0.04511339031159878, 'timing/model': 1.35286068264395, 'learning_rate/qwen_vl_interface': 6.497683007727166e-07, 'learning_rate/action_model': 6.497683007727166e-06, 'epoch': 2.48}) +[2026-06-18 08:14:37,017][starVLA.training.train_starvla][INFO] - Step 6448, Loss: {'train/loss': 0.2390909194946289, 'train/grad_norm_pre_clip': 13.177385330200195, 'timing/data': 0.04220359958708286, 'timing/model': 1.1260585011914372, 'learning_rate/qwen_vl_interface': 6.492299846390024e-07, 'learning_rate/action_model': 6.492299846390023e-06, 'epoch': 2.48}) +[2026-06-18 08:14:41,643][starVLA.training.train_starvla][INFO] - Step 6449, Loss: {'train/loss': 0.30219992995262146, 'train/grad_norm_pre_clip': 15.600756645202637, 'timing/data': 0.053830891847610474, 'timing/model': 1.0926140202209353, 'learning_rate/qwen_vl_interface': 6.486926222498633e-07, 'learning_rate/action_model': 6.486926222498633e-06, 'epoch': 2.48}) +[2026-06-18 08:14:50,067][starVLA.training.train_starvla][INFO] - Step 6450, Loss: {'train/loss': 0.1886613965034485, 'train/grad_norm_pre_clip': 12.991792678833008, 'timing/data': 0.04392263200134039, 'timing/model': 4.350687755271792, 'learning_rate/qwen_vl_interface': 6.481562137166952e-07, 'learning_rate/action_model': 6.481562137166952e-06, 'epoch': 2.48}) +[2026-06-18 08:14:56,762][starVLA.training.train_starvla][INFO] - Step 6451, Loss: {'train/loss': 0.33979034423828125, 'train/grad_norm_pre_clip': 17.725980758666992, 'timing/data': 0.04103485215455294, 'timing/model': 1.361233314499259, 'learning_rate/qwen_vl_interface': 6.476207591506954e-07, 'learning_rate/action_model': 6.476207591506954e-06, 'epoch': 2.48}) +[2026-06-18 08:15:01,542][starVLA.training.train_starvla][INFO] - Step 6452, Loss: {'train/loss': 0.26000669598579407, 'train/grad_norm_pre_clip': 14.924605369567871, 'timing/data': 0.04591077473014593, 'timing/model': 1.4791586166247725, 'learning_rate/qwen_vl_interface': 6.470862586628645e-07, 'learning_rate/action_model': 6.470862586628646e-06, 'epoch': 2.48}) +[2026-06-18 08:15:06,040][starVLA.training.train_starvla][INFO] - Step 6453, Loss: {'train/loss': 0.45829105377197266, 'train/grad_norm_pre_clip': 17.652164459228516, 'timing/data': 0.0388722438365221, 'timing/model': 1.2026181062683463, 'learning_rate/qwen_vl_interface': 6.465527123640056e-07, 'learning_rate/action_model': 6.465527123640056e-06, 'epoch': 2.48}) +[2026-06-18 08:15:10,795][starVLA.training.train_starvla][INFO] - Step 6454, Loss: {'train/loss': 0.5078655481338501, 'train/grad_norm_pre_clip': 20.734230041503906, 'timing/data': 0.04158073477447033, 'timing/model': 1.1245597042143345, 'learning_rate/qwen_vl_interface': 6.460201203647229e-07, 'learning_rate/action_model': 6.460201203647229e-06, 'epoch': 2.48}) +[2026-06-18 08:15:15,702][starVLA.training.train_starvla][INFO] - Step 6455, Loss: {'train/loss': 0.22756506502628326, 'train/grad_norm_pre_clip': 18.378948211669922, 'timing/data': 0.05196359660476446, 'timing/model': 1.1435389900580049, 'learning_rate/qwen_vl_interface': 6.454884827754224e-07, 'learning_rate/action_model': 6.4548848277542235e-06, 'epoch': 2.48}) +[2026-06-18 08:15:20,347][starVLA.training.train_starvla][INFO] - Step 6456, Loss: {'train/loss': 0.2844305634498596, 'train/grad_norm_pre_clip': 13.56618881225586, 'timing/data': 0.04842090234160423, 'timing/model': 1.0929212514311075, 'learning_rate/qwen_vl_interface': 6.449577997063156e-07, 'learning_rate/action_model': 6.449577997063155e-06, 'epoch': 2.48}) +[2026-06-18 08:15:24,959][starVLA.training.train_starvla][INFO] - Step 6457, Loss: {'train/loss': 0.14779792726039886, 'train/grad_norm_pre_clip': 9.983402252197266, 'timing/data': 0.04261447675526142, 'timing/model': 1.3943310491740704, 'learning_rate/qwen_vl_interface': 6.444280712674112e-07, 'learning_rate/action_model': 6.444280712674111e-06, 'epoch': 2.48}) +[2026-06-18 08:15:29,225][starVLA.training.train_starvla][INFO] - Step 6458, Loss: {'train/loss': 0.2775880694389343, 'train/grad_norm_pre_clip': 13.641144752502441, 'timing/data': 0.04282015189528465, 'timing/model': 1.1092509450390935, 'learning_rate/qwen_vl_interface': 6.438992975685237e-07, 'learning_rate/action_model': 6.438992975685237e-06, 'epoch': 2.48}) +[2026-06-18 08:15:33,884][starVLA.training.train_starvla][INFO] - Step 6459, Loss: {'train/loss': 0.34881386160850525, 'train/grad_norm_pre_clip': 18.33827781677246, 'timing/data': 0.04407844226807356, 'timing/model': 1.1572035672143102, 'learning_rate/qwen_vl_interface': 6.433714787192677e-07, 'learning_rate/action_model': 6.4337147871926766e-06, 'epoch': 2.48}) +[2026-06-18 08:15:39,141][starVLA.training.train_starvla][INFO] - Step 6460, Loss: {'train/loss': 0.10960264503955841, 'train/grad_norm_pre_clip': 11.599316596984863, 'timing/data': 0.0614403085783124, 'timing/model': 1.12047388125211, 'learning_rate/qwen_vl_interface': 6.428446148290615e-07, 'learning_rate/action_model': 6.428446148290615e-06, 'epoch': 2.49}) +[2026-06-18 08:15:43,857][starVLA.training.train_starvla][INFO] - Step 6461, Loss: {'train/loss': 0.38106703758239746, 'train/grad_norm_pre_clip': 18.329051971435547, 'timing/data': 0.042488062754273415, 'timing/model': 1.092144382186234, 'learning_rate/qwen_vl_interface': 6.423187060071248e-07, 'learning_rate/action_model': 6.423187060071248e-06, 'epoch': 2.49}) +[2026-06-18 08:15:48,514][starVLA.training.train_starvla][INFO] - Step 6462, Loss: {'train/loss': 0.15565159916877747, 'train/grad_norm_pre_clip': 9.827179908752441, 'timing/data': 0.03917039930820465, 'timing/model': 1.423343289643526, 'learning_rate/qwen_vl_interface': 6.417937523624778e-07, 'learning_rate/action_model': 6.417937523624778e-06, 'epoch': 2.49}) +[2026-06-18 08:15:52,748][starVLA.training.train_starvla][INFO] - Step 6463, Loss: {'train/loss': 0.34438639879226685, 'train/grad_norm_pre_clip': 15.722282409667969, 'timing/data': 0.037456230260431767, 'timing/model': 1.0870483322069049, 'learning_rate/qwen_vl_interface': 6.412697540039442e-07, 'learning_rate/action_model': 6.412697540039442e-06, 'epoch': 2.49}) +[2026-06-18 08:15:58,205][starVLA.training.train_starvla][INFO] - Step 6464, Loss: {'train/loss': 0.21426904201507568, 'train/grad_norm_pre_clip': 15.659027099609375, 'timing/data': 0.03943505324423313, 'timing/model': 1.1151913609355688, 'learning_rate/qwen_vl_interface': 6.407467110401505e-07, 'learning_rate/action_model': 6.407467110401505e-06, 'epoch': 2.49}) +[2026-06-18 08:16:02,929][starVLA.training.train_starvla][INFO] - Step 6465, Loss: {'train/loss': 0.48624542355537415, 'train/grad_norm_pre_clip': 18.951629638671875, 'timing/data': 0.041321951895952225, 'timing/model': 1.1436703819781542, 'learning_rate/qwen_vl_interface': 6.402246235795228e-07, 'learning_rate/action_model': 6.402246235795227e-06, 'epoch': 2.49}) +[2026-06-18 08:16:07,727][starVLA.training.train_starvla][INFO] - Step 6466, Loss: {'train/loss': 0.30026543140411377, 'train/grad_norm_pre_clip': 17.518903732299805, 'timing/data': 0.0423461627215147, 'timing/model': 1.1471214797347784, 'learning_rate/qwen_vl_interface': 6.397034917302913e-07, 'learning_rate/action_model': 6.397034917302913e-06, 'epoch': 2.49}) +[2026-06-18 08:16:12,641][starVLA.training.train_starvla][INFO] - Step 6467, Loss: {'train/loss': 0.20210619270801544, 'train/grad_norm_pre_clip': 12.686803817749023, 'timing/data': 0.054279061034321785, 'timing/model': 1.458369287662208, 'learning_rate/qwen_vl_interface': 6.391833156004871e-07, 'learning_rate/action_model': 6.39183315600487e-06, 'epoch': 2.49}) +[2026-06-18 08:16:16,914][starVLA.training.train_starvla][INFO] - Step 6468, Loss: {'train/loss': 0.2920144200325012, 'train/grad_norm_pre_clip': 13.826689720153809, 'timing/data': 0.04154479783028364, 'timing/model': 1.0746691282838583, 'learning_rate/qwen_vl_interface': 6.386640952979427e-07, 'learning_rate/action_model': 6.386640952979426e-06, 'epoch': 2.49}) +[2026-06-18 08:16:21,604][starVLA.training.train_starvla][INFO] - Step 6469, Loss: {'train/loss': 0.48414942622184753, 'train/grad_norm_pre_clip': 24.654556274414062, 'timing/data': 0.03964743297547102, 'timing/model': 1.2836474003270268, 'learning_rate/qwen_vl_interface': 6.381458309302937e-07, 'learning_rate/action_model': 6.381458309302937e-06, 'epoch': 2.49}) +[2026-06-18 08:16:26,389][starVLA.training.train_starvla][INFO] - Step 6470, Loss: {'train/loss': 0.2560007572174072, 'train/grad_norm_pre_clip': 15.465771675109863, 'timing/data': 0.042633069679141045, 'timing/model': 1.1107885679230094, 'learning_rate/qwen_vl_interface': 6.376285226049762e-07, 'learning_rate/action_model': 6.376285226049762e-06, 'epoch': 2.49}) +[2026-06-18 08:16:33,209][starVLA.training.train_starvla][INFO] - Step 6471, Loss: {'train/loss': 0.4922630786895752, 'train/grad_norm_pre_clip': 19.076906204223633, 'timing/data': 0.04078679904341698, 'timing/model': 1.2283037677407265, 'learning_rate/qwen_vl_interface': 6.371121704292292e-07, 'learning_rate/action_model': 6.371121704292292e-06, 'epoch': 2.49}) +[2026-06-18 08:16:38,007][starVLA.training.train_starvla][INFO] - Step 6472, Loss: {'train/loss': 0.4276541471481323, 'train/grad_norm_pre_clip': 18.41360092163086, 'timing/data': 0.050402951426804066, 'timing/model': 1.5221651019528508, 'learning_rate/qwen_vl_interface': 6.365967745100926e-07, 'learning_rate/action_model': 6.3659677451009256e-06, 'epoch': 2.49}) +[2026-06-18 08:16:42,546][starVLA.training.train_starvla][INFO] - Step 6473, Loss: {'train/loss': 0.16809293627738953, 'train/grad_norm_pre_clip': 13.23865795135498, 'timing/data': 0.03768893051892519, 'timing/model': 1.2705553686246276, 'learning_rate/qwen_vl_interface': 6.360823349544092e-07, 'learning_rate/action_model': 6.360823349544092e-06, 'epoch': 2.49}) +[2026-06-18 08:16:47,295][starVLA.training.train_starvla][INFO] - Step 6474, Loss: {'train/loss': 0.19412389397621155, 'train/grad_norm_pre_clip': 10.650433540344238, 'timing/data': 0.06392782274633646, 'timing/model': 1.2015826143324375, 'learning_rate/qwen_vl_interface': 6.355688518688222e-07, 'learning_rate/action_model': 6.355688518688221e-06, 'epoch': 2.49}) +[2026-06-18 08:16:51,902][starVLA.training.train_starvla][INFO] - Step 6475, Loss: {'train/loss': 0.3627032935619354, 'train/grad_norm_pre_clip': 14.86405086517334, 'timing/data': 0.03897832240909338, 'timing/model': 1.0893529746681452, 'learning_rate/qwen_vl_interface': 6.350563253597781e-07, 'learning_rate/action_model': 6.350563253597781e-06, 'epoch': 2.49}) +[2026-06-18 08:16:56,541][starVLA.training.train_starvla][INFO] - Step 6476, Loss: {'train/loss': 0.17459101974964142, 'train/grad_norm_pre_clip': 11.262310028076172, 'timing/data': 0.042993881739676, 'timing/model': 1.0939056100323796, 'learning_rate/qwen_vl_interface': 6.345447555335236e-07, 'learning_rate/action_model': 6.3454475553352345e-06, 'epoch': 2.49}) +[2026-06-18 08:17:01,342][starVLA.training.train_starvla][INFO] - Step 6477, Loss: {'train/loss': 0.41369983553886414, 'train/grad_norm_pre_clip': 15.970064163208008, 'timing/data': 0.04697071760892868, 'timing/model': 1.551527675241232, 'learning_rate/qwen_vl_interface': 6.340341424961072e-07, 'learning_rate/action_model': 6.340341424961071e-06, 'epoch': 2.49}) +[2026-06-18 08:17:06,683][starVLA.training.train_starvla][INFO] - Step 6478, Loss: {'train/loss': 0.4926496148109436, 'train/grad_norm_pre_clip': 24.161319732666016, 'timing/data': 0.04293058253824711, 'timing/model': 1.2360656224191189, 'learning_rate/qwen_vl_interface': 6.335244863533802e-07, 'learning_rate/action_model': 6.3352448635338016e-06, 'epoch': 2.49}) +[2026-06-18 08:17:11,585][starVLA.training.train_starvla][INFO] - Step 6479, Loss: {'train/loss': 0.2766657769680023, 'train/grad_norm_pre_clip': 15.29345703125, 'timing/data': 0.04520292952656746, 'timing/model': 1.1010438846424222, 'learning_rate/qwen_vl_interface': 6.330157872109953e-07, 'learning_rate/action_model': 6.330157872109953e-06, 'epoch': 2.49}) +[2026-06-18 08:17:16,293][starVLA.training.train_starvla][INFO] - Step 6480, Loss: {'train/loss': 0.40984389185905457, 'train/grad_norm_pre_clip': 16.54353904724121, 'timing/data': 0.04317774623632431, 'timing/model': 1.099491854198277, 'learning_rate/qwen_vl_interface': 6.325080451744053e-07, 'learning_rate/action_model': 6.325080451744053e-06, 'epoch': 2.49}) +[2026-06-18 08:17:21,043][starVLA.training.train_starvla][INFO] - Step 6481, Loss: {'train/loss': 0.3117610514163971, 'train/grad_norm_pre_clip': 13.566341400146484, 'timing/data': 0.03727415669709444, 'timing/model': 1.0884016957134008, 'learning_rate/qwen_vl_interface': 6.320012603488663e-07, 'learning_rate/action_model': 6.320012603488662e-06, 'epoch': 2.49}) +[2026-06-18 08:17:25,767][starVLA.training.train_starvla][INFO] - Step 6482, Loss: {'train/loss': 0.16012226045131683, 'train/grad_norm_pre_clip': 7.123279094696045, 'timing/data': 0.04586453549563885, 'timing/model': 1.5599023718386889, 'learning_rate/qwen_vl_interface': 6.314954328394351e-07, 'learning_rate/action_model': 6.3149543283943515e-06, 'epoch': 2.49}) +[2026-06-18 08:17:30,184][starVLA.training.train_starvla][INFO] - Step 6483, Loss: {'train/loss': 0.32668766379356384, 'train/grad_norm_pre_clip': 21.6500301361084, 'timing/data': 0.040511745028197765, 'timing/model': 1.0942479576915503, 'learning_rate/qwen_vl_interface': 6.309905627509707e-07, 'learning_rate/action_model': 6.309905627509706e-06, 'epoch': 2.49}) +[2026-06-18 08:17:34,967][starVLA.training.train_starvla][INFO] - Step 6484, Loss: {'train/loss': 0.32142671942710876, 'train/grad_norm_pre_clip': 23.502779006958008, 'timing/data': 0.041672226041555405, 'timing/model': 1.1580497696995735, 'learning_rate/qwen_vl_interface': 6.304866501881327e-07, 'learning_rate/action_model': 6.304866501881327e-06, 'epoch': 2.49}) +[2026-06-18 08:17:41,812][starVLA.training.train_starvla][INFO] - Step 6485, Loss: {'train/loss': 0.13921698927879333, 'train/grad_norm_pre_clip': 9.865704536437988, 'timing/data': 0.07485113572329283, 'timing/model': 2.7810482420027256, 'learning_rate/qwen_vl_interface': 6.299836952553831e-07, 'learning_rate/action_model': 6.299836952553831e-06, 'epoch': 2.49}) +[2026-06-18 08:17:47,293][starVLA.training.train_starvla][INFO] - Step 6486, Loss: {'train/loss': 0.5985038876533508, 'train/grad_norm_pre_clip': 18.618165969848633, 'timing/data': 0.03991616144776344, 'timing/model': 1.1081166444346309, 'learning_rate/qwen_vl_interface': 6.294816980569845e-07, 'learning_rate/action_model': 6.294816980569845e-06, 'epoch': 2.5}) +[2026-06-18 08:17:51,854][starVLA.training.train_starvla][INFO] - Step 6487, Loss: {'train/loss': 0.24217547476291656, 'train/grad_norm_pre_clip': 15.098498344421387, 'timing/data': 0.04333858843892813, 'timing/model': 1.3996480908244848, 'learning_rate/qwen_vl_interface': 6.289806586970021e-07, 'learning_rate/action_model': 6.289806586970021e-06, 'epoch': 2.5}) +[2026-06-18 08:17:56,298][starVLA.training.train_starvla][INFO] - Step 6488, Loss: {'train/loss': 0.2820203900337219, 'train/grad_norm_pre_clip': 13.747713088989258, 'timing/data': 0.04071499779820442, 'timing/model': 1.0911324117332697, 'learning_rate/qwen_vl_interface': 6.284805772793011e-07, 'learning_rate/action_model': 6.284805772793011e-06, 'epoch': 2.5}) +[2026-06-18 08:18:04,300][starVLA.training.train_starvla][INFO] - Step 6489, Loss: {'train/loss': 0.14694921672344208, 'train/grad_norm_pre_clip': 13.126089096069336, 'timing/data': 0.04283038154244423, 'timing/model': 1.0891051217913628, 'learning_rate/qwen_vl_interface': 6.279814539075499e-07, 'learning_rate/action_model': 6.2798145390754985e-06, 'epoch': 2.5}) +[2026-06-18 08:18:10,436][starVLA.training.train_starvla][INFO] - Step 6490, Loss: {'train/loss': 0.35743996500968933, 'train/grad_norm_pre_clip': 17.486324310302734, 'timing/data': 0.04036503005772829, 'timing/model': 1.086159223690629, 'learning_rate/qwen_vl_interface': 6.274832886852172e-07, 'learning_rate/action_model': 6.2748328868521724e-06, 'epoch': 2.5}) +[2026-06-18 08:18:15,273][starVLA.training.train_starvla][INFO] - Step 6491, Loss: {'train/loss': 0.4159693717956543, 'train/grad_norm_pre_clip': 19.402326583862305, 'timing/data': 0.044810534454882145, 'timing/model': 1.132683478295803, 'learning_rate/qwen_vl_interface': 6.269860817155721e-07, 'learning_rate/action_model': 6.269860817155721e-06, 'epoch': 2.5}) +[2026-06-18 08:18:20,643][starVLA.training.train_starvla][INFO] - Step 6492, Loss: {'train/loss': 0.2988761365413666, 'train/grad_norm_pre_clip': 14.7485990524292, 'timing/data': 0.040526581928133965, 'timing/model': 1.4281685911118984, 'learning_rate/qwen_vl_interface': 6.264898331016865e-07, 'learning_rate/action_model': 6.264898331016865e-06, 'epoch': 2.5}) +[2026-06-18 08:18:24,924][starVLA.training.train_starvla][INFO] - Step 6493, Loss: {'train/loss': 0.2971908450126648, 'train/grad_norm_pre_clip': 21.11676025390625, 'timing/data': 0.04486209712922573, 'timing/model': 1.112956965342164, 'learning_rate/qwen_vl_interface': 6.259945429464343e-07, 'learning_rate/action_model': 6.259945429464342e-06, 'epoch': 2.5}) +[2026-06-18 08:18:29,675][starVLA.training.train_starvla][INFO] - Step 6494, Loss: {'train/loss': 0.21438419818878174, 'train/grad_norm_pre_clip': 14.190352439880371, 'timing/data': 0.035982996225357056, 'timing/model': 1.0797230070456862, 'learning_rate/qwen_vl_interface': 6.255002113524885e-07, 'learning_rate/action_model': 6.255002113524886e-06, 'epoch': 2.5}) +[2026-06-18 08:18:34,447][starVLA.training.train_starvla][INFO] - Step 6495, Loss: {'train/loss': 0.14765354990959167, 'train/grad_norm_pre_clip': 9.747221946716309, 'timing/data': 0.038328791968524456, 'timing/model': 1.1293076844885945, 'learning_rate/qwen_vl_interface': 6.250068384223255e-07, 'learning_rate/action_model': 6.250068384223255e-06, 'epoch': 2.5}) +[2026-06-18 08:18:39,034][starVLA.training.train_starvla][INFO] - Step 6496, Loss: {'train/loss': 0.2760634422302246, 'train/grad_norm_pre_clip': 15.502083778381348, 'timing/data': 0.029023081064224243, 'timing/model': 1.1199518097564578, 'learning_rate/qwen_vl_interface': 6.245144242582214e-07, 'learning_rate/action_model': 6.245144242582214e-06, 'epoch': 2.5}) +[2026-06-18 08:18:43,560][starVLA.training.train_starvla][INFO] - Step 6497, Loss: {'train/loss': 0.35207319259643555, 'train/grad_norm_pre_clip': 15.204658508300781, 'timing/data': 0.028755699284374714, 'timing/model': 1.370862246491015, 'learning_rate/qwen_vl_interface': 6.240229689622546e-07, 'learning_rate/action_model': 6.240229689622545e-06, 'epoch': 2.5}) +[2026-06-18 08:18:47,821][starVLA.training.train_starvla][INFO] - Step 6498, Loss: {'train/loss': 0.3467143177986145, 'train/grad_norm_pre_clip': 18.932205200195312, 'timing/data': 0.03085155412554741, 'timing/model': 1.094213075004518, 'learning_rate/qwen_vl_interface': 6.235324726363042e-07, 'learning_rate/action_model': 6.235324726363041e-06, 'epoch': 2.5}) +[2026-06-18 08:18:52,578][starVLA.training.train_starvla][INFO] - Step 6499, Loss: {'train/loss': 0.4170228838920593, 'train/grad_norm_pre_clip': 14.699458122253418, 'timing/data': 0.03385934326797724, 'timing/model': 1.1069441083818674, 'learning_rate/qwen_vl_interface': 6.230429353820503e-07, 'learning_rate/action_model': 6.230429353820503e-06, 'epoch': 2.5}) +[2026-06-18 08:21:14,090][starVLA.training.train_starvla][INFO] - Step 6500, Loss: {'train/loss': 0.20249053835868835, 'train/grad_norm_pre_clip': 13.357163429260254, 'eval/loss': 0.4244554936885834, 'eval/latency_4/loss': 0.4244554936885834, 'eval/demon_fire_precision': 0.9695317511225144, 'eval/demon_fire_recall': 0.9818122767132186, 'eval/demon_fire_f1': 0.9756333709859609, 'eval/demon_fire_rate_model': 0.915712187958884, 'eval/demon_fire_rate_teacher': 0.9042584434654919, 'eval/demon_left_precision': 0.7814070351758794, 'eval/demon_left_recall': 0.7100456621004566, 'eval/demon_left_f1': 0.7440191387559809, 'eval/demon_left_rate_model': 0.11688693098384728, 'eval/demon_left_rate_teacher': 0.12863436123348018, 'eval/demon_right_precision': 0.7870036101083032, 'eval/demon_right_recall': 0.7870036101083032, 'eval/demon_right_f1': 0.7870036101083032, 'eval/demon_right_rate_model': 0.16270190895741557, 'eval/demon_right_rate_teacher': 0.16270190895741557, 'eval/demon_move_f1': 0.765511374432142, 'eval/demon_cc_f1': 0.9014063961825755, 'eval/latency_4/demon_fire_precision': 0.9695317511225144, 'eval/latency_4/demon_fire_recall': 0.9818122767132186, 'eval/latency_4/demon_fire_f1': 0.9756333709859609, 'eval/latency_4/demon_fire_rate_model': 0.915712187958884, 'eval/latency_4/demon_fire_rate_teacher': 0.9042584434654919, 'eval/latency_4/demon_left_precision': 0.7814070351758794, 'eval/latency_4/demon_left_recall': 0.7100456621004566, 'eval/latency_4/demon_left_f1': 0.7440191387559809, 'eval/latency_4/demon_left_rate_model': 0.11688693098384728, 'eval/latency_4/demon_left_rate_teacher': 0.12863436123348018, 'eval/latency_4/demon_right_precision': 0.7870036101083032, 'eval/latency_4/demon_right_recall': 0.7870036101083032, 'eval/latency_4/demon_right_f1': 0.7870036101083032, 'eval/latency_4/demon_right_rate_model': 0.16270190895741557, 'eval/latency_4/demon_right_rate_teacher': 0.16270190895741557, 'eval/latency_4/demon_move_f1': 0.765511374432142, 'eval/latency_4/demon_cc_f1': 0.9014063961825755, 'timing/data': 0.029945937916636467, 'timing/model': 1.0756824892014265, 'learning_rate/qwen_vl_interface': 6.225543573009747e-07, 'learning_rate/action_model': 6.225543573009748e-06, 'epoch': 2.5}) +[2026-06-18 08:21:14,092][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_6500_state +[2026-06-18 08:21:14,099][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 08:21:22,937][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_6500_state/model.safetensors +[2026-06-18 08:21:44,952][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_6500_state/optimizer.bin +[2026-06-18 08:21:44,954][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_6500_state/sampler.bin +[2026-06-18 08:21:44,954][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_6500_state/sampler_1.bin +[2026-06-18 08:21:44,957][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_6500_state/random_states_0.pkl +[2026-06-18 08:21:44,962][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 6500 +[2026-06-18 08:21:44,963][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 08:21:44,981][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 08:21:49,964][starVLA.training.train_starvla][INFO] - Step 6501, Loss: {'train/loss': 0.23559081554412842, 'train/grad_norm_pre_clip': 17.302940368652344, 'timing/data': 0.032989393919706345, 'timing/model': 1.0976205728948116, 'learning_rate/qwen_vl_interface': 6.2206673849436e-07, 'learning_rate/action_model': 6.2206673849436e-06, 'epoch': 2.5}) +[2026-06-18 08:21:54,676][starVLA.training.train_starvla][INFO] - Step 6502, Loss: {'train/loss': 0.14551913738250732, 'train/grad_norm_pre_clip': 10.64387321472168, 'timing/data': 0.035710218362510204, 'timing/model': 1.348501873202622, 'learning_rate/qwen_vl_interface': 6.215800790632905e-07, 'learning_rate/action_model': 6.215800790632904e-06, 'epoch': 2.5}) +[2026-06-18 08:21:59,147][starVLA.training.train_starvla][INFO] - Step 6503, Loss: {'train/loss': 0.20874260365962982, 'train/grad_norm_pre_clip': 12.352895736694336, 'timing/data': 0.03373074345290661, 'timing/model': 1.1234836485236883, 'learning_rate/qwen_vl_interface': 6.210943791086506e-07, 'learning_rate/action_model': 6.210943791086506e-06, 'epoch': 2.5}) +[2026-06-18 08:22:05,722][starVLA.training.train_starvla][INFO] - Step 6504, Loss: {'train/loss': 0.16110087931156158, 'train/grad_norm_pre_clip': 8.679047584533691, 'timing/data': 0.03086903505027294, 'timing/model': 1.232548982836306, 'learning_rate/qwen_vl_interface': 6.206096387311278e-07, 'learning_rate/action_model': 6.206096387311277e-06, 'epoch': 2.5}) +[2026-06-18 08:22:11,733][starVLA.training.train_starvla][INFO] - Step 6505, Loss: {'train/loss': 0.4068731963634491, 'train/grad_norm_pre_clip': 18.94571876525879, 'timing/data': 0.03160715475678444, 'timing/model': 1.107525055296719, 'learning_rate/qwen_vl_interface': 6.201258580312078e-07, 'learning_rate/action_model': 6.201258580312077e-06, 'epoch': 2.5}) +[2026-06-18 08:22:18,780][starVLA.training.train_starvla][INFO] - Step 6506, Loss: {'train/loss': 0.12246187031269073, 'train/grad_norm_pre_clip': 10.674450874328613, 'timing/data': 0.03122993092983961, 'timing/model': 2.108225670643151, 'learning_rate/qwen_vl_interface': 6.196430371091792e-07, 'learning_rate/action_model': 6.196430371091792e-06, 'epoch': 2.5}) +[2026-06-18 08:22:24,883][starVLA.training.train_starvla][INFO] - Step 6507, Loss: {'train/loss': 0.0915215015411377, 'train/grad_norm_pre_clip': 5.91496467590332, 'timing/data': 0.04321076720952988, 'timing/model': 2.791080314666033, 'learning_rate/qwen_vl_interface': 6.191611760651312e-07, 'learning_rate/action_model': 6.191611760651312e-06, 'epoch': 2.5}) +[2026-06-18 08:22:30,022][starVLA.training.train_starvla][INFO] - Step 6508, Loss: {'train/loss': 0.17498788237571716, 'train/grad_norm_pre_clip': 11.85273551940918, 'timing/data': 0.03095600474625826, 'timing/model': 1.1752515276893973, 'learning_rate/qwen_vl_interface': 6.18680274998955e-07, 'learning_rate/action_model': 6.186802749989549e-06, 'epoch': 2.5}) +[2026-06-18 08:22:34,827][starVLA.training.train_starvla][INFO] - Step 6509, Loss: {'train/loss': 0.10038747638463974, 'train/grad_norm_pre_clip': 7.7385029792785645, 'timing/data': 0.03138938080519438, 'timing/model': 1.132678009569645, 'learning_rate/qwen_vl_interface': 6.182003340103413e-07, 'learning_rate/action_model': 6.1820033401034135e-06, 'epoch': 2.5}) +[2026-06-18 08:22:39,463][starVLA.training.train_starvla][INFO] - Step 6510, Loss: {'train/loss': 0.4167702794075012, 'train/grad_norm_pre_clip': 14.510222434997559, 'timing/data': 0.030056153424084187, 'timing/model': 1.0915403198450804, 'learning_rate/qwen_vl_interface': 6.177213531987819e-07, 'learning_rate/action_model': 6.177213531987819e-06, 'epoch': 2.5}) +[2026-06-18 08:22:44,285][starVLA.training.train_starvla][INFO] - Step 6511, Loss: {'train/loss': 0.25512224435806274, 'train/grad_norm_pre_clip': 16.619001388549805, 'timing/data': 0.033741530030965805, 'timing/model': 1.6180314784869552, 'learning_rate/qwen_vl_interface': 6.172433326635714e-07, 'learning_rate/action_model': 6.172433326635713e-06, 'epoch': 2.5}) +[2026-06-18 08:22:49,132][starVLA.training.train_starvla][INFO] - Step 6512, Loss: {'train/loss': 0.33792826533317566, 'train/grad_norm_pre_clip': 20.375980377197266, 'timing/data': 0.032010975293815136, 'timing/model': 1.5340201435610652, 'learning_rate/qwen_vl_interface': 6.167662725038025e-07, 'learning_rate/action_model': 6.1676627250380255e-06, 'epoch': 2.51}) +[2026-06-18 08:22:53,428][starVLA.training.train_starvla][INFO] - Step 6513, Loss: {'train/loss': 0.4171849191188812, 'train/grad_norm_pre_clip': 18.16707992553711, 'timing/data': 0.032247269526124, 'timing/model': 1.096901928074658, 'learning_rate/qwen_vl_interface': 6.16290172818371e-07, 'learning_rate/action_model': 6.16290172818371e-06, 'epoch': 2.51}) +[2026-06-18 08:22:58,084][starVLA.training.train_starvla][INFO] - Step 6514, Loss: {'train/loss': 0.13002197444438934, 'train/grad_norm_pre_clip': 8.083659172058105, 'timing/data': 0.0312010170891881, 'timing/model': 1.109848054125905, 'learning_rate/qwen_vl_interface': 6.158150337059725e-07, 'learning_rate/action_model': 6.158150337059725e-06, 'epoch': 2.51}) +[2026-06-18 08:23:02,851][starVLA.training.train_starvla][INFO] - Step 6515, Loss: {'train/loss': 0.42109397053718567, 'train/grad_norm_pre_clip': 21.151334762573242, 'timing/data': 0.03337698243558407, 'timing/model': 1.1179649969562888, 'learning_rate/qwen_vl_interface': 6.153408552651045e-07, 'learning_rate/action_model': 6.153408552651045e-06, 'epoch': 2.51}) +[2026-06-18 08:23:07,672][starVLA.training.train_starvla][INFO] - Step 6516, Loss: {'train/loss': 0.09870076924562454, 'train/grad_norm_pre_clip': 11.445162773132324, 'timing/data': 0.03824332170188427, 'timing/model': 1.5625261142849922, 'learning_rate/qwen_vl_interface': 6.14867637594064e-07, 'learning_rate/action_model': 6.14867637594064e-06, 'epoch': 2.51}) +[2026-06-18 08:23:12,546][starVLA.training.train_starvla][INFO] - Step 6517, Loss: {'train/loss': 0.06245856359601021, 'train/grad_norm_pre_clip': 29.370054244995117, 'timing/data': 0.03276418335735798, 'timing/model': 1.606951167806983, 'learning_rate/qwen_vl_interface': 6.143953807909497e-07, 'learning_rate/action_model': 6.143953807909496e-06, 'epoch': 2.51}) +[2026-06-18 08:23:16,867][starVLA.training.train_starvla][INFO] - Step 6518, Loss: {'train/loss': 0.15954133868217468, 'train/grad_norm_pre_clip': 11.876221656799316, 'timing/data': 0.03318885527551174, 'timing/model': 1.1083990838378668, 'learning_rate/qwen_vl_interface': 6.139240849536616e-07, 'learning_rate/action_model': 6.1392408495366156e-06, 'epoch': 2.51}) +[2026-06-18 08:23:21,596][starVLA.training.train_starvla][INFO] - Step 6519, Loss: {'train/loss': 0.2792229950428009, 'train/grad_norm_pre_clip': 16.888744354248047, 'timing/data': 0.03224517032504082, 'timing/model': 1.1511855265125632, 'learning_rate/qwen_vl_interface': 6.134537501798993e-07, 'learning_rate/action_model': 6.1345375017989925e-06, 'epoch': 2.51}) +[2026-06-18 08:23:26,615][starVLA.training.train_starvla][INFO] - Step 6520, Loss: {'train/loss': 0.306519478559494, 'train/grad_norm_pre_clip': 19.56293296813965, 'timing/data': 0.05817484110593796, 'timing/model': 1.1175877526402473, 'learning_rate/qwen_vl_interface': 6.129843765671634e-07, 'learning_rate/action_model': 6.129843765671633e-06, 'epoch': 2.51}) +[2026-06-18 08:23:31,320][starVLA.training.train_starvla][INFO] - Step 6521, Loss: {'train/loss': 0.1856740266084671, 'train/grad_norm_pre_clip': 22.79127311706543, 'timing/data': 0.02981510292738676, 'timing/model': 1.1001029554754496, 'learning_rate/qwen_vl_interface': 6.125159642127559e-07, 'learning_rate/action_model': 6.125159642127558e-06, 'epoch': 2.51}) +[2026-06-18 08:23:36,003][starVLA.training.train_starvla][INFO] - Step 6522, Loss: {'train/loss': 0.38464370369911194, 'train/grad_norm_pre_clip': 14.7376070022583, 'timing/data': 0.03042863868176937, 'timing/model': 1.5030718399211764, 'learning_rate/qwen_vl_interface': 6.120485132137793e-07, 'learning_rate/action_model': 6.120485132137794e-06, 'epoch': 2.51}) +[2026-06-18 08:23:42,999][starVLA.training.train_starvla][INFO] - Step 6523, Loss: {'train/loss': 0.19687402248382568, 'train/grad_norm_pre_clip': 12.665738105773926, 'timing/data': 0.03183938190340996, 'timing/model': 2.196203908883035, 'learning_rate/qwen_vl_interface': 6.115820236671369e-07, 'learning_rate/action_model': 6.115820236671369e-06, 'epoch': 2.51}) +[2026-06-18 08:23:49,739][starVLA.training.train_starvla][INFO] - Step 6524, Loss: {'train/loss': 0.23540565371513367, 'train/grad_norm_pre_clip': 15.7821683883667, 'timing/data': 0.0361566636711359, 'timing/model': 1.5868518752977252, 'learning_rate/qwen_vl_interface': 6.111164956695316e-07, 'learning_rate/action_model': 6.111164956695316e-06, 'epoch': 2.51}) +[2026-06-18 08:23:55,997][starVLA.training.train_starvla][INFO] - Step 6525, Loss: {'train/loss': 0.5151214599609375, 'train/grad_norm_pre_clip': 23.443105697631836, 'timing/data': 0.0313192130997777, 'timing/model': 1.1390940127894282, 'learning_rate/qwen_vl_interface': 6.106519293174684e-07, 'learning_rate/action_model': 6.106519293174684e-06, 'epoch': 2.51}) +[2026-06-18 08:24:00,764][starVLA.training.train_starvla][INFO] - Step 6526, Loss: {'train/loss': 0.45751821994781494, 'train/grad_norm_pre_clip': 17.427242279052734, 'timing/data': 0.03157393168658018, 'timing/model': 1.0916616199538112, 'learning_rate/qwen_vl_interface': 6.101883247072525e-07, 'learning_rate/action_model': 6.1018832470725245e-06, 'epoch': 2.51}) +[2026-06-18 08:24:06,262][starVLA.training.train_starvla][INFO] - Step 6527, Loss: {'train/loss': 0.2228560894727707, 'train/grad_norm_pre_clip': 16.532846450805664, 'timing/data': 0.06308077741414309, 'timing/model': 1.5143360821530223, 'learning_rate/qwen_vl_interface': 6.097256819349899e-07, 'learning_rate/action_model': 6.0972568193498985e-06, 'epoch': 2.51}) +[2026-06-18 08:24:13,184][starVLA.training.train_starvla][INFO] - Step 6528, Loss: {'train/loss': 0.3251543641090393, 'train/grad_norm_pre_clip': 15.458203315734863, 'timing/data': 0.04130949545651674, 'timing/model': 1.1469971546903253, 'learning_rate/qwen_vl_interface': 6.092640010965864e-07, 'learning_rate/action_model': 6.092640010965864e-06, 'epoch': 2.51}) +[2026-06-18 08:24:18,600][starVLA.training.train_starvla][INFO] - Step 6529, Loss: {'train/loss': 0.21782514452934265, 'train/grad_norm_pre_clip': 12.41985034942627, 'timing/data': 0.03387684002518654, 'timing/model': 1.1572470972314477, 'learning_rate/qwen_vl_interface': 6.088032822877488e-07, 'learning_rate/action_model': 6.088032822877488e-06, 'epoch': 2.51}) +[2026-06-18 08:24:24,256][starVLA.training.train_starvla][INFO] - Step 6530, Loss: {'train/loss': 0.6387373805046082, 'train/grad_norm_pre_clip': 15.959551811218262, 'timing/data': 0.04198740888386965, 'timing/model': 1.7853991519659758, 'learning_rate/qwen_vl_interface': 6.083435256039854e-07, 'learning_rate/action_model': 6.083435256039854e-06, 'epoch': 2.51}) +[2026-06-18 08:24:29,325][starVLA.training.train_starvla][INFO] - Step 6531, Loss: {'train/loss': 0.28855040669441223, 'train/grad_norm_pre_clip': 12.950766563415527, 'timing/data': 0.034155127592384815, 'timing/model': 1.096193851903081, 'learning_rate/qwen_vl_interface': 6.078847311406033e-07, 'learning_rate/action_model': 6.078847311406032e-06, 'epoch': 2.51}) +[2026-06-18 08:24:33,786][starVLA.training.train_starvla][INFO] - Step 6532, Loss: {'train/loss': 0.18925835192203522, 'train/grad_norm_pre_clip': 16.89556121826172, 'timing/data': 0.048284866847097874, 'timing/model': 1.1917200637981296, 'learning_rate/qwen_vl_interface': 6.07426898992712e-07, 'learning_rate/action_model': 6.074268989927119e-06, 'epoch': 2.51}) +[2026-06-18 08:24:38,567][starVLA.training.train_starvla][INFO] - Step 6533, Loss: {'train/loss': 0.5173966884613037, 'train/grad_norm_pre_clip': 21.913578033447266, 'timing/data': 0.03435794450342655, 'timing/model': 1.130603350698948, 'learning_rate/qwen_vl_interface': 6.069700292552202e-07, 'learning_rate/action_model': 6.069700292552202e-06, 'epoch': 2.51}) +[2026-06-18 08:24:43,394][starVLA.training.train_starvla][INFO] - Step 6534, Loss: {'train/loss': 0.14701144397258759, 'train/grad_norm_pre_clip': 15.279012680053711, 'timing/data': 0.03149195574223995, 'timing/model': 1.1240205550566316, 'learning_rate/qwen_vl_interface': 6.065141220228372e-07, 'learning_rate/action_model': 6.065141220228372e-06, 'epoch': 2.51}) +[2026-06-18 08:24:48,043][starVLA.training.train_starvla][INFO] - Step 6535, Loss: {'train/loss': 0.35433971881866455, 'train/grad_norm_pre_clip': 17.560522079467773, 'timing/data': 0.03206373751163483, 'timing/model': 1.1045214273035526, 'learning_rate/qwen_vl_interface': 6.060591773900732e-07, 'learning_rate/action_model': 6.060591773900733e-06, 'epoch': 2.51}) +[2026-06-18 08:24:52,925][starVLA.training.train_starvla][INFO] - Step 6536, Loss: {'train/loss': 0.13498100638389587, 'train/grad_norm_pre_clip': 16.621675491333008, 'timing/data': 0.03391411621123552, 'timing/model': 1.15079584158957, 'learning_rate/qwen_vl_interface': 6.056051954512388e-07, 'learning_rate/action_model': 6.0560519545123875e-06, 'epoch': 2.51}) +[2026-06-18 08:24:57,666][starVLA.training.train_starvla][INFO] - Step 6537, Loss: {'train/loss': 0.19922275841236115, 'train/grad_norm_pre_clip': 14.622350692749023, 'timing/data': 0.041704319417476654, 'timing/model': 1.5236995806917548, 'learning_rate/qwen_vl_interface': 6.051521763004455e-07, 'learning_rate/action_model': 6.051521763004456e-06, 'epoch': 2.51}) +[2026-06-18 08:25:02,134][starVLA.training.train_starvla][INFO] - Step 6538, Loss: {'train/loss': 0.3112388849258423, 'train/grad_norm_pre_clip': 16.78972816467285, 'timing/data': 0.049716025590896606, 'timing/model': 1.1469651805236936, 'learning_rate/qwen_vl_interface': 6.047001200316038e-07, 'learning_rate/action_model': 6.047001200316037e-06, 'epoch': 2.52}) +[2026-06-18 08:25:06,910][starVLA.training.train_starvla][INFO] - Step 6539, Loss: {'train/loss': 0.32709333300590515, 'train/grad_norm_pre_clip': 13.837331771850586, 'timing/data': 0.03653107117861509, 'timing/model': 1.134309479035437, 'learning_rate/qwen_vl_interface': 6.042490267384255e-07, 'learning_rate/action_model': 6.0424902673842545e-06, 'epoch': 2.52}) +[2026-06-18 08:25:11,739][starVLA.training.train_starvla][INFO] - Step 6540, Loss: {'train/loss': 0.19192537665367126, 'train/grad_norm_pre_clip': 15.525592803955078, 'timing/data': 0.03410243336111307, 'timing/model': 1.1637966493144631, 'learning_rate/qwen_vl_interface': 6.03798896514423e-07, 'learning_rate/action_model': 6.0379889651442295e-06, 'epoch': 2.52}) +[2026-06-18 08:25:16,717][starVLA.training.train_starvla][INFO] - Step 6541, Loss: {'train/loss': 0.2094150334596634, 'train/grad_norm_pre_clip': 17.45526123046875, 'timing/data': 0.03215489070862532, 'timing/model': 1.165771022439003, 'learning_rate/qwen_vl_interface': 6.033497294529093e-07, 'learning_rate/action_model': 6.033497294529093e-06, 'epoch': 2.52}) +[2026-06-18 08:25:21,050][starVLA.training.train_starvla][INFO] - Step 6542, Loss: {'train/loss': 0.20809021592140198, 'train/grad_norm_pre_clip': 15.172077178955078, 'timing/data': 0.03702751733362675, 'timing/model': 1.1144210565835238, 'learning_rate/qwen_vl_interface': 6.029015256469967e-07, 'learning_rate/action_model': 6.0290152564699666e-06, 'epoch': 2.52}) +[2026-06-18 08:25:26,053][starVLA.training.train_starvla][INFO] - Step 6543, Loss: {'train/loss': 0.14915241301059723, 'train/grad_norm_pre_clip': 37.793739318847656, 'timing/data': 0.037601382471621037, 'timing/model': 1.1244032755494118, 'learning_rate/qwen_vl_interface': 6.024542851895976e-07, 'learning_rate/action_model': 6.0245428518959756e-06, 'epoch': 2.52}) +[2026-06-18 08:25:30,958][starVLA.training.train_starvla][INFO] - Step 6544, Loss: {'train/loss': 0.224297896027565, 'train/grad_norm_pre_clip': 11.889805793762207, 'timing/data': 0.04686241876333952, 'timing/model': 1.216120583936572, 'learning_rate/qwen_vl_interface': 6.020080081734267e-07, 'learning_rate/action_model': 6.020080081734266e-06, 'epoch': 2.52}) +[2026-06-18 08:25:35,891][starVLA.training.train_starvla][INFO] - Step 6545, Loss: {'train/loss': 0.3634102940559387, 'train/grad_norm_pre_clip': 19.113121032714844, 'timing/data': 0.04536943882703781, 'timing/model': 1.1367448307573795, 'learning_rate/qwen_vl_interface': 6.015626946909974e-07, 'learning_rate/action_model': 6.015626946909973e-06, 'epoch': 2.52}) +[2026-06-18 08:25:40,652][starVLA.training.train_starvla][INFO] - Step 6546, Loss: {'train/loss': 0.26794007420539856, 'train/grad_norm_pre_clip': 18.46324348449707, 'timing/data': 0.031676266342401505, 'timing/model': 1.5123878130689263, 'learning_rate/qwen_vl_interface': 6.011183448346233e-07, 'learning_rate/action_model': 6.011183448346233e-06, 'epoch': 2.52}) +[2026-06-18 08:25:45,281][starVLA.training.train_starvla][INFO] - Step 6547, Loss: {'train/loss': 0.20285996794700623, 'train/grad_norm_pre_clip': 11.459318161010742, 'timing/data': 0.03181970492005348, 'timing/model': 1.409428388811648, 'learning_rate/qwen_vl_interface': 6.006749586964183e-07, 'learning_rate/action_model': 6.006749586964183e-06, 'epoch': 2.52}) +[2026-06-18 08:25:49,962][starVLA.training.train_starvla][INFO] - Step 6548, Loss: {'train/loss': 0.22072294354438782, 'train/grad_norm_pre_clip': 14.253070831298828, 'timing/data': 0.03200354985892773, 'timing/model': 1.211692868731916, 'learning_rate/qwen_vl_interface': 6.002325363682974e-07, 'learning_rate/action_model': 6.0023253636829736e-06, 'epoch': 2.52}) +[2026-06-18 08:25:54,807][starVLA.training.train_starvla][INFO] - Step 6549, Loss: {'train/loss': 0.5234028697013855, 'train/grad_norm_pre_clip': 19.029834747314453, 'timing/data': 0.03442135080695152, 'timing/model': 1.1381581341847777, 'learning_rate/qwen_vl_interface': 5.997910779419756e-07, 'learning_rate/action_model': 5.9979107794197565e-06, 'epoch': 2.52}) +[2026-06-18 08:25:59,758][starVLA.training.train_starvla][INFO] - Step 6550, Loss: {'train/loss': 0.12483144551515579, 'train/grad_norm_pre_clip': 12.618420600891113, 'timing/data': 0.03154024574905634, 'timing/model': 1.2618140541017056, 'learning_rate/qwen_vl_interface': 5.993505835089661e-07, 'learning_rate/action_model': 5.993505835089661e-06, 'epoch': 2.52}) +[2026-06-18 08:26:04,616][starVLA.training.train_starvla][INFO] - Step 6551, Loss: {'train/loss': 0.32082265615463257, 'train/grad_norm_pre_clip': 19.49578285217285, 'timing/data': 0.028711074963212013, 'timing/model': 1.5031248275190592, 'learning_rate/qwen_vl_interface': 5.98911053160586e-07, 'learning_rate/action_model': 5.989110531605859e-06, 'epoch': 2.52}) +[2026-06-18 08:26:09,712][starVLA.training.train_starvla][INFO] - Step 6552, Loss: {'train/loss': 0.44471749663352966, 'train/grad_norm_pre_clip': 21.141496658325195, 'timing/data': 0.04246125090867281, 'timing/model': 1.655575473792851, 'learning_rate/qwen_vl_interface': 5.984724869879494e-07, 'learning_rate/action_model': 5.984724869879493e-06, 'epoch': 2.52}) +[2026-06-18 08:26:14,297][starVLA.training.train_starvla][INFO] - Step 6553, Loss: {'train/loss': 0.10941970348358154, 'train/grad_norm_pre_clip': 6.601065635681152, 'timing/data': 0.030044199898838997, 'timing/model': 1.2037517447024584, 'learning_rate/qwen_vl_interface': 5.980348850819711e-07, 'learning_rate/action_model': 5.98034885081971e-06, 'epoch': 2.52}) +[2026-06-18 08:26:18,962][starVLA.training.train_starvla][INFO] - Step 6554, Loss: {'train/loss': 0.3315240144729614, 'train/grad_norm_pre_clip': 18.18002700805664, 'timing/data': 0.030317913740873337, 'timing/model': 1.0910732625052333, 'learning_rate/qwen_vl_interface': 5.975982475333664e-07, 'learning_rate/action_model': 5.975982475333663e-06, 'epoch': 2.52}) +[2026-06-18 08:26:23,776][starVLA.training.train_starvla][INFO] - Step 6555, Loss: {'train/loss': 0.2885381579399109, 'train/grad_norm_pre_clip': 18.371700286865234, 'timing/data': 0.036334370262920856, 'timing/model': 1.1636363435536623, 'learning_rate/qwen_vl_interface': 5.971625744326516e-07, 'learning_rate/action_model': 5.971625744326515e-06, 'epoch': 2.52}) +[2026-06-18 08:26:28,540][starVLA.training.train_starvla][INFO] - Step 6556, Loss: {'train/loss': 0.12651191651821136, 'train/grad_norm_pre_clip': 13.515044212341309, 'timing/data': 0.03398390207439661, 'timing/model': 1.0933128269389272, 'learning_rate/qwen_vl_interface': 5.967278658701421e-07, 'learning_rate/action_model': 5.967278658701421e-06, 'epoch': 2.52}) +[2026-06-18 08:26:33,439][starVLA.training.train_starvla][INFO] - Step 6557, Loss: {'train/loss': 0.10021540522575378, 'train/grad_norm_pre_clip': 8.490144729614258, 'timing/data': 0.04232924152165651, 'timing/model': 1.597519119270146, 'learning_rate/qwen_vl_interface': 5.962941219359527e-07, 'learning_rate/action_model': 5.962941219359527e-06, 'epoch': 2.52}) +[2026-06-18 08:26:37,950][starVLA.training.train_starvla][INFO] - Step 6558, Loss: {'train/loss': 0.432237833738327, 'train/grad_norm_pre_clip': 26.793180465698242, 'timing/data': 0.04027815442532301, 'timing/model': 1.140583386644721, 'learning_rate/qwen_vl_interface': 5.958613427199998e-07, 'learning_rate/action_model': 5.958613427199997e-06, 'epoch': 2.52}) +[2026-06-18 08:26:42,939][starVLA.training.train_starvla][INFO] - Step 6559, Loss: {'train/loss': 0.38410302996635437, 'train/grad_norm_pre_clip': 23.295928955078125, 'timing/data': 0.03541139978915453, 'timing/model': 1.2592418184503913, 'learning_rate/qwen_vl_interface': 5.954295283119985e-07, 'learning_rate/action_model': 5.954295283119985e-06, 'epoch': 2.52}) +[2026-06-18 08:26:47,683][starVLA.training.train_starvla][INFO] - Step 6560, Loss: {'train/loss': 0.10479512810707092, 'train/grad_norm_pre_clip': 8.64945125579834, 'timing/data': 0.04360033571720123, 'timing/model': 1.1382225109264255, 'learning_rate/qwen_vl_interface': 5.949986788014643e-07, 'learning_rate/action_model': 5.949986788014643e-06, 'epoch': 2.52}) +[2026-06-18 08:26:52,357][starVLA.training.train_starvla][INFO] - Step 6561, Loss: {'train/loss': 0.3727620542049408, 'train/grad_norm_pre_clip': 61.17095947265625, 'timing/data': 0.03525155317038298, 'timing/model': 1.4451047824695706, 'learning_rate/qwen_vl_interface': 5.94568794277713e-07, 'learning_rate/action_model': 5.94568794277713e-06, 'epoch': 2.52}) +[2026-06-18 08:26:57,054][starVLA.training.train_starvla][INFO] - Step 6562, Loss: {'train/loss': 0.38840413093566895, 'train/grad_norm_pre_clip': 11.471651077270508, 'timing/data': 0.035878950729966164, 'timing/model': 1.4945646226406097, 'learning_rate/qwen_vl_interface': 5.941398748298603e-07, 'learning_rate/action_model': 5.941398748298603e-06, 'epoch': 2.52}) +[2026-06-18 08:27:01,437][starVLA.training.train_starvla][INFO] - Step 6563, Loss: {'train/loss': 0.39098021388053894, 'train/grad_norm_pre_clip': 21.4152774810791, 'timing/data': 0.033510600216686726, 'timing/model': 1.1166034862399101, 'learning_rate/qwen_vl_interface': 5.937119205468217e-07, 'learning_rate/action_model': 5.9371192054682165e-06, 'epoch': 2.52}) +[2026-06-18 08:27:06,222][starVLA.training.train_starvla][INFO] - Step 6564, Loss: {'train/loss': 0.16166119277477264, 'train/grad_norm_pre_clip': 12.449328422546387, 'timing/data': 0.032735067419707775, 'timing/model': 1.111965373158455, 'learning_rate/qwen_vl_interface': 5.93284931517312e-07, 'learning_rate/action_model': 5.93284931517312e-06, 'epoch': 2.53}) +[2026-06-18 08:27:11,291][starVLA.training.train_starvla][INFO] - Step 6565, Loss: {'train/loss': 0.25337550044059753, 'train/grad_norm_pre_clip': 18.62510108947754, 'timing/data': 0.05427048821002245, 'timing/model': 1.133563227020204, 'learning_rate/qwen_vl_interface': 5.928589078298468e-07, 'learning_rate/action_model': 5.928589078298468e-06, 'epoch': 2.53}) +[2026-06-18 08:27:16,034][starVLA.training.train_starvla][INFO] - Step 6566, Loss: {'train/loss': 0.2867930233478546, 'train/grad_norm_pre_clip': 19.99114418029785, 'timing/data': 0.039948402903974056, 'timing/model': 1.119486128911376, 'learning_rate/qwen_vl_interface': 5.924338495727416e-07, 'learning_rate/action_model': 5.9243384957274155e-06, 'epoch': 2.53}) +[2026-06-18 08:27:20,891][starVLA.training.train_starvla][INFO] - Step 6567, Loss: {'train/loss': 0.7139372825622559, 'train/grad_norm_pre_clip': 19.967512130737305, 'timing/data': 0.036945879459381104, 'timing/model': 1.527759350836277, 'learning_rate/qwen_vl_interface': 5.92009756834111e-07, 'learning_rate/action_model': 5.9200975683411094e-06, 'epoch': 2.53}) +[2026-06-18 08:27:25,307][starVLA.training.train_starvla][INFO] - Step 6568, Loss: {'train/loss': 0.2888907194137573, 'train/grad_norm_pre_clip': 484.42041015625, 'timing/data': 0.03220323100686073, 'timing/model': 1.1652666674926877, 'learning_rate/qwen_vl_interface': 5.915866297018697e-07, 'learning_rate/action_model': 5.915866297018696e-06, 'epoch': 2.53}) +[2026-06-18 08:27:30,184][starVLA.training.train_starvla][INFO] - Step 6569, Loss: {'train/loss': 0.4101406931877136, 'train/grad_norm_pre_clip': 15.506068229675293, 'timing/data': 0.032719068229198456, 'timing/model': 1.1361877378076315, 'learning_rate/qwen_vl_interface': 5.911644682637329e-07, 'learning_rate/action_model': 5.911644682637329e-06, 'epoch': 2.53}) +[2026-06-18 08:27:35,063][starVLA.training.train_starvla][INFO] - Step 6570, Loss: {'train/loss': 0.22308075428009033, 'train/grad_norm_pre_clip': 14.314898490905762, 'timing/data': 0.03254710976034403, 'timing/model': 1.1197629692032933, 'learning_rate/qwen_vl_interface': 5.907432726072149e-07, 'learning_rate/action_model': 5.9074327260721485e-06, 'epoch': 2.53}) +[2026-06-18 08:27:40,152][starVLA.training.train_starvla][INFO] - Step 6571, Loss: {'train/loss': 0.313364714384079, 'train/grad_norm_pre_clip': 14.382030487060547, 'timing/data': 0.04117053281515837, 'timing/model': 1.2083833441138268, 'learning_rate/qwen_vl_interface': 5.903230428196305e-07, 'learning_rate/action_model': 5.9032304281963045e-06, 'epoch': 2.53}) +[2026-06-18 08:27:45,141][starVLA.training.train_starvla][INFO] - Step 6572, Loss: {'train/loss': 0.3133499324321747, 'train/grad_norm_pre_clip': 15.505658149719238, 'timing/data': 0.034714896231889725, 'timing/model': 1.5402969401329756, 'learning_rate/qwen_vl_interface': 5.899037789880929e-07, 'learning_rate/action_model': 5.899037789880929e-06, 'epoch': 2.53}) +[2026-06-18 08:27:49,698][starVLA.training.train_starvla][INFO] - Step 6573, Loss: {'train/loss': 0.0638311430811882, 'train/grad_norm_pre_clip': 4.660569667816162, 'timing/data': 0.03115310799330473, 'timing/model': 1.1156368544325233, 'learning_rate/qwen_vl_interface': 5.894854811995166e-07, 'learning_rate/action_model': 5.894854811995165e-06, 'epoch': 2.53}) +[2026-06-18 08:27:54,577][starVLA.training.train_starvla][INFO] - Step 6574, Loss: {'train/loss': 0.1752503514289856, 'train/grad_norm_pre_clip': 13.186264038085938, 'timing/data': 0.028661546297371387, 'timing/model': 1.1605715630576015, 'learning_rate/qwen_vl_interface': 5.890681495406152e-07, 'learning_rate/action_model': 5.890681495406151e-06, 'epoch': 2.53}) +[2026-06-18 08:27:59,569][starVLA.training.train_starvla][INFO] - Step 6575, Loss: {'train/loss': 0.2579911947250366, 'train/grad_norm_pre_clip': 18.095653533935547, 'timing/data': 0.03580387681722641, 'timing/model': 1.2141266744583845, 'learning_rate/qwen_vl_interface': 5.886517840979015e-07, 'learning_rate/action_model': 5.886517840979016e-06, 'epoch': 2.53}) +[2026-06-18 08:28:04,350][starVLA.training.train_starvla][INFO] - Step 6576, Loss: {'train/loss': 0.47560545802116394, 'train/grad_norm_pre_clip': 23.799041748046875, 'timing/data': 0.032554345205426216, 'timing/model': 1.125557036139071, 'learning_rate/qwen_vl_interface': 5.882363849576889e-07, 'learning_rate/action_model': 5.882363849576889e-06, 'epoch': 2.53}) +[2026-06-18 08:28:09,326][starVLA.training.train_starvla][INFO] - Step 6577, Loss: {'train/loss': 0.20039589703083038, 'train/grad_norm_pre_clip': 10.345185279846191, 'timing/data': 0.04412715137004852, 'timing/model': 1.5678885048255324, 'learning_rate/qwen_vl_interface': 5.8782195220609e-07, 'learning_rate/action_model': 5.878219522060899e-06, 'epoch': 2.53}) +[2026-06-18 08:28:13,729][starVLA.training.train_starvla][INFO] - Step 6578, Loss: {'train/loss': 0.2619416117668152, 'train/grad_norm_pre_clip': 17.894515991210938, 'timing/data': 0.036857916973531246, 'timing/model': 1.1607515336945653, 'learning_rate/qwen_vl_interface': 5.874084859290172e-07, 'learning_rate/action_model': 5.874084859290171e-06, 'epoch': 2.53}) +[2026-06-18 08:28:18,562][starVLA.training.train_starvla][INFO] - Step 6579, Loss: {'train/loss': 0.3099558353424072, 'train/grad_norm_pre_clip': 15.86949634552002, 'timing/data': 0.03018641471862793, 'timing/model': 1.095470391213894, 'learning_rate/qwen_vl_interface': 5.869959862121822e-07, 'learning_rate/action_model': 5.869959862121822e-06, 'epoch': 2.53}) +[2026-06-18 08:28:23,370][starVLA.training.train_starvla][INFO] - Step 6580, Loss: {'train/loss': 0.2869950234889984, 'train/grad_norm_pre_clip': 18.48587989807129, 'timing/data': 0.034639074467122555, 'timing/model': 1.143297865986824, 'learning_rate/qwen_vl_interface': 5.86584453141097e-07, 'learning_rate/action_model': 5.86584453141097e-06, 'epoch': 2.53}) +[2026-06-18 08:28:28,178][starVLA.training.train_starvla][INFO] - Step 6581, Loss: {'train/loss': 0.1647685021162033, 'train/grad_norm_pre_clip': 8.371554374694824, 'timing/data': 0.05320214480161667, 'timing/model': 1.4565495010465384, 'learning_rate/qwen_vl_interface': 5.861738868010723e-07, 'learning_rate/action_model': 5.861738868010723e-06, 'epoch': 2.53}) +[2026-06-18 08:28:33,044][starVLA.training.train_starvla][INFO] - Step 6582, Loss: {'train/loss': 0.2186051309108734, 'train/grad_norm_pre_clip': 15.995895385742188, 'timing/data': 0.03152035176753998, 'timing/model': 1.5119196800515056, 'learning_rate/qwen_vl_interface': 5.857642872772198e-07, 'learning_rate/action_model': 5.8576428727721986e-06, 'epoch': 2.53}) +[2026-06-18 08:28:37,535][starVLA.training.train_starvla][INFO] - Step 6583, Loss: {'train/loss': 0.38428249955177307, 'train/grad_norm_pre_clip': 18.245040893554688, 'timing/data': 0.032291026785969734, 'timing/model': 1.2273198897019029, 'learning_rate/qwen_vl_interface': 5.853556546544491e-07, 'learning_rate/action_model': 5.8535565465444904e-06, 'epoch': 2.53}) +[2026-06-18 08:28:42,487][starVLA.training.train_starvla][INFO] - Step 6584, Loss: {'train/loss': 0.2924811840057373, 'train/grad_norm_pre_clip': 16.164318084716797, 'timing/data': 0.034328688867390156, 'timing/model': 1.1328396024182439, 'learning_rate/qwen_vl_interface': 5.849479890174708e-07, 'learning_rate/action_model': 5.8494798901747075e-06, 'epoch': 2.53}) +[2026-06-18 08:28:47,362][starVLA.training.train_starvla][INFO] - Step 6585, Loss: {'train/loss': 0.14646106958389282, 'train/grad_norm_pre_clip': 9.237427711486816, 'timing/data': 0.036868334747850895, 'timing/model': 1.208218039944768, 'learning_rate/qwen_vl_interface': 5.845412904507948e-07, 'learning_rate/action_model': 5.845412904507948e-06, 'epoch': 2.53}) +[2026-06-18 08:28:52,108][starVLA.training.train_starvla][INFO] - Step 6586, Loss: {'train/loss': 0.19994179904460907, 'train/grad_norm_pre_clip': 11.343381881713867, 'timing/data': 0.03971083369106054, 'timing/model': 1.123582391999662, 'learning_rate/qwen_vl_interface': 5.841355590387291e-07, 'learning_rate/action_model': 5.841355590387291e-06, 'epoch': 2.53}) +[2026-06-18 08:28:56,847][starVLA.training.train_starvla][INFO] - Step 6587, Loss: {'train/loss': 0.1603747308254242, 'train/grad_norm_pre_clip': 15.527033805847168, 'timing/data': 0.03110253531485796, 'timing/model': 1.5030774045735598, 'learning_rate/qwen_vl_interface': 5.837307948653821e-07, 'learning_rate/action_model': 5.837307948653821e-06, 'epoch': 2.53}) +[2026-06-18 08:29:01,331][starVLA.training.train_starvla][INFO] - Step 6588, Loss: {'train/loss': 0.27648183703422546, 'train/grad_norm_pre_clip': 14.816549301147461, 'timing/data': 0.032076128758490086, 'timing/model': 1.1065707299858332, 'learning_rate/qwen_vl_interface': 5.833269980146631e-07, 'learning_rate/action_model': 5.83326998014663e-06, 'epoch': 2.53}) +[2026-06-18 08:29:06,157][starVLA.training.train_starvla][INFO] - Step 6589, Loss: {'train/loss': 0.3005308508872986, 'train/grad_norm_pre_clip': 8.888012886047363, 'timing/data': 0.03196304477751255, 'timing/model': 1.1948999520391226, 'learning_rate/qwen_vl_interface': 5.829241685702791e-07, 'learning_rate/action_model': 5.82924168570279e-06, 'epoch': 2.53}) +[2026-06-18 08:29:11,040][starVLA.training.train_starvla][INFO] - Step 6590, Loss: {'train/loss': 0.7765424251556396, 'train/grad_norm_pre_clip': 22.5893497467041, 'timing/data': 0.03248995263129473, 'timing/model': 1.0945240808650851, 'learning_rate/qwen_vl_interface': 5.825223066157365e-07, 'learning_rate/action_model': 5.8252230661573645e-06, 'epoch': 2.54}) +[2026-06-18 08:29:15,889][starVLA.training.train_starvla][INFO] - Step 6591, Loss: {'train/loss': 0.10865893959999084, 'train/grad_norm_pre_clip': 6.799234867095947, 'timing/data': 0.03425065707415342, 'timing/model': 1.2349734921008348, 'learning_rate/qwen_vl_interface': 5.821214122343423e-07, 'learning_rate/action_model': 5.821214122343422e-06, 'epoch': 2.54}) +[2026-06-18 08:29:20,758][starVLA.training.train_starvla][INFO] - Step 6592, Loss: {'train/loss': 0.48037099838256836, 'train/grad_norm_pre_clip': 22.682390213012695, 'timing/data': 0.036152370274066925, 'timing/model': 1.567194014787674, 'learning_rate/qwen_vl_interface': 5.817214855092024e-07, 'learning_rate/action_model': 5.817214855092024e-06, 'epoch': 2.54}) +[2026-06-18 08:29:25,296][starVLA.training.train_starvla][INFO] - Step 6593, Loss: {'train/loss': 0.11944667994976044, 'train/grad_norm_pre_clip': 9.274140357971191, 'timing/data': 0.042139504104852676, 'timing/model': 1.127661264501512, 'learning_rate/qwen_vl_interface': 5.813225265232216e-07, 'learning_rate/action_model': 5.813225265232215e-06, 'epoch': 2.54}) +[2026-06-18 08:29:30,380][starVLA.training.train_starvla][INFO] - Step 6594, Loss: {'train/loss': 0.2932451665401459, 'train/grad_norm_pre_clip': 20.8807373046875, 'timing/data': 0.041608063504099846, 'timing/model': 1.2243165271356702, 'learning_rate/qwen_vl_interface': 5.809245353591048e-07, 'learning_rate/action_model': 5.8092453535910485e-06, 'epoch': 2.54}) +[2026-06-18 08:29:35,190][starVLA.training.train_starvla][INFO] - Step 6595, Loss: {'train/loss': 0.07913315296173096, 'train/grad_norm_pre_clip': 8.451048851013184, 'timing/data': 0.030769026838243008, 'timing/model': 1.2247195029631257, 'learning_rate/qwen_vl_interface': 5.805275120993561e-07, 'learning_rate/action_model': 5.805275120993561e-06, 'epoch': 2.54}) +[2026-06-18 08:29:40,599][starVLA.training.train_starvla][INFO] - Step 6596, Loss: {'train/loss': 0.24960957467556, 'train/grad_norm_pre_clip': 12.817524909973145, 'timing/data': 0.03865078557282686, 'timing/model': 1.206723302602768, 'learning_rate/qwen_vl_interface': 5.801314568262787e-07, 'learning_rate/action_model': 5.801314568262787e-06, 'epoch': 2.54}) +[2026-06-18 08:29:45,413][starVLA.training.train_starvla][INFO] - Step 6597, Loss: {'train/loss': 0.26136812567710876, 'train/grad_norm_pre_clip': 17.12626838684082, 'timing/data': 0.02941071055829525, 'timing/model': 1.4710963908582926, 'learning_rate/qwen_vl_interface': 5.797363696219751e-07, 'learning_rate/action_model': 5.797363696219751e-06, 'epoch': 2.54}) +[2026-06-18 08:29:50,002][starVLA.training.train_starvla][INFO] - Step 6598, Loss: {'train/loss': 0.2285938411951065, 'train/grad_norm_pre_clip': 15.869827270507812, 'timing/data': 0.029189560562372208, 'timing/model': 1.2289978666231036, 'learning_rate/qwen_vl_interface': 5.793422505683474e-07, 'learning_rate/action_model': 5.793422505683474e-06, 'epoch': 2.54}) +[2026-06-18 08:29:54,896][starVLA.training.train_starvla][INFO] - Step 6599, Loss: {'train/loss': 0.1619102507829666, 'train/grad_norm_pre_clip': 13.625901222229004, 'timing/data': 0.032992079854011536, 'timing/model': 1.1621291190385818, 'learning_rate/qwen_vl_interface': 5.789490997470978e-07, 'learning_rate/action_model': 5.789490997470978e-06, 'epoch': 2.54}) +[2026-06-18 08:32:36,412][starVLA.training.train_starvla][INFO] - Step 6600, Loss: {'train/loss': 0.16856122016906738, 'train/grad_norm_pre_clip': 10.366447448730469, 'eval/loss': 0.42546597123146057, 'eval/latency_4/loss': 0.42546597123146057, 'eval/demon_fire_precision': 0.9640699523052464, 'eval/demon_fire_recall': 0.9847353036700227, 'eval/demon_fire_f1': 0.974293059125964, 'eval/demon_fire_rate_model': 0.9236417033773862, 'eval/demon_fire_rate_teacher': 0.9042584434654919, 'eval/demon_left_precision': 0.8202247191011236, 'eval/demon_left_recall': 0.6666666666666666, 'eval/demon_left_f1': 0.7355163727959698, 'eval/demon_left_rate_model': 0.10455212922173275, 'eval/demon_left_rate_teacher': 0.12863436123348018, 'eval/demon_right_precision': 0.8128544423440454, 'eval/demon_right_recall': 0.776173285198556, 'eval/demon_right_f1': 0.7940904893813481, 'eval/demon_right_rate_model': 0.155359765051395, 'eval/demon_right_rate_teacher': 0.16270190895741557, 'eval/demon_move_f1': 0.7648034310886589, 'eval/demon_cc_f1': 0.9003109851461099, 'eval/latency_4/demon_fire_precision': 0.9640699523052464, 'eval/latency_4/demon_fire_recall': 0.9847353036700227, 'eval/latency_4/demon_fire_f1': 0.974293059125964, 'eval/latency_4/demon_fire_rate_model': 0.9236417033773862, 'eval/latency_4/demon_fire_rate_teacher': 0.9042584434654919, 'eval/latency_4/demon_left_precision': 0.8202247191011236, 'eval/latency_4/demon_left_recall': 0.6666666666666666, 'eval/latency_4/demon_left_f1': 0.7355163727959698, 'eval/latency_4/demon_left_rate_model': 0.10455212922173275, 'eval/latency_4/demon_left_rate_teacher': 0.12863436123348018, 'eval/latency_4/demon_right_precision': 0.8128544423440454, 'eval/latency_4/demon_right_recall': 0.776173285198556, 'eval/latency_4/demon_right_f1': 0.7940904893813481, 'eval/latency_4/demon_right_rate_model': 0.155359765051395, 'eval/latency_4/demon_right_rate_teacher': 0.16270190895741557, 'eval/latency_4/demon_move_f1': 0.7648034310886589, 'eval/latency_4/demon_cc_f1': 0.9003109851461099, 'timing/data': 0.030655421316623688, 'timing/model': 1.086154043674469, 'learning_rate/qwen_vl_interface': 5.785569172397254e-07, 'learning_rate/action_model': 5.785569172397254e-06, 'epoch': 2.54}) +[2026-06-18 08:32:36,414][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_6600_state +[2026-06-18 08:32:36,431][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 08:32:46,100][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_6600_state/model.safetensors +[2026-06-18 08:33:09,989][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_6600_state/optimizer.bin +[2026-06-18 08:33:09,991][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_6600_state/sampler.bin +[2026-06-18 08:33:10,042][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_6600_state/sampler_1.bin +[2026-06-18 08:33:10,045][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_6600_state/random_states_0.pkl +[2026-06-18 08:33:10,051][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 6600 +[2026-06-18 08:33:10,051][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 08:33:10,069][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 08:33:15,060][starVLA.training.train_starvla][INFO] - Step 6601, Loss: {'train/loss': 0.3441198170185089, 'train/grad_norm_pre_clip': 14.322220802307129, 'timing/data': 0.034314521588385105, 'timing/model': 1.6665633618831635, 'learning_rate/qwen_vl_interface': 5.781657031275306e-07, 'learning_rate/action_model': 5.781657031275306e-06, 'epoch': 2.54}) +[2026-06-18 08:33:19,719][starVLA.training.train_starvla][INFO] - Step 6602, Loss: {'train/loss': 0.21369025111198425, 'train/grad_norm_pre_clip': 12.783058166503906, 'timing/data': 0.03305340092629194, 'timing/model': 1.1265854816883802, 'learning_rate/qwen_vl_interface': 5.777754574916129e-07, 'learning_rate/action_model': 5.7777545749161285e-06, 'epoch': 2.54}) +[2026-06-18 08:33:24,781][starVLA.training.train_starvla][INFO] - Step 6603, Loss: {'train/loss': 0.2695574164390564, 'train/grad_norm_pre_clip': 17.498836517333984, 'timing/data': 0.0455827945843339, 'timing/model': 1.1493012243881822, 'learning_rate/qwen_vl_interface': 5.773861804128701e-07, 'learning_rate/action_model': 5.773861804128701e-06, 'epoch': 2.54}) +[2026-06-18 08:33:29,751][starVLA.training.train_starvla][INFO] - Step 6604, Loss: {'train/loss': 0.42389100790023804, 'train/grad_norm_pre_clip': 19.299551010131836, 'timing/data': 0.05813313089311123, 'timing/model': 1.1020019836723804, 'learning_rate/qwen_vl_interface': 5.769978719720011e-07, 'learning_rate/action_model': 5.76997871972001e-06, 'epoch': 2.54}) +[2026-06-18 08:33:34,767][starVLA.training.train_starvla][INFO] - Step 6605, Loss: {'train/loss': 0.3802540600299835, 'train/grad_norm_pre_clip': 21.99090576171875, 'timing/data': 0.03595553431659937, 'timing/model': 1.1672026626765728, 'learning_rate/qwen_vl_interface': 5.766105322495002e-07, 'learning_rate/action_model': 5.766105322495002e-06, 'epoch': 2.54}) +[2026-06-18 08:33:39,813][starVLA.training.train_starvla][INFO] - Step 6606, Loss: {'train/loss': 0.2841153144836426, 'train/grad_norm_pre_clip': 15.474695205688477, 'timing/data': 0.045966253615915775, 'timing/model': 1.209196725860238, 'learning_rate/qwen_vl_interface': 5.762241613256654e-07, 'learning_rate/action_model': 5.762241613256654e-06, 'epoch': 2.54}) +[2026-06-18 08:33:45,070][starVLA.training.train_starvla][INFO] - Step 6607, Loss: {'train/loss': 0.09851737320423126, 'train/grad_norm_pre_clip': 7.438230037689209, 'timing/data': 0.03823321592062712, 'timing/model': 1.506686300970614, 'learning_rate/qwen_vl_interface': 5.758387592805916e-07, 'learning_rate/action_model': 5.758387592805915e-06, 'epoch': 2.54}) +[2026-06-18 08:33:49,742][starVLA.training.train_starvla][INFO] - Step 6608, Loss: {'train/loss': 0.5341363549232483, 'train/grad_norm_pre_clip': 21.867097854614258, 'timing/data': 0.037775314413011074, 'timing/model': 1.249268228188157, 'learning_rate/qwen_vl_interface': 5.754543261941719e-07, 'learning_rate/action_model': 5.754543261941718e-06, 'epoch': 2.54}) +[2026-06-18 08:33:55,079][starVLA.training.train_starvla][INFO] - Step 6609, Loss: {'train/loss': 0.18625852465629578, 'train/grad_norm_pre_clip': 18.125425338745117, 'timing/data': 0.0380468750372529, 'timing/model': 1.2624141816049814, 'learning_rate/qwen_vl_interface': 5.750708621461011e-07, 'learning_rate/action_model': 5.75070862146101e-06, 'epoch': 2.54}) +[2026-06-18 08:34:00,317][starVLA.training.train_starvla][INFO] - Step 6610, Loss: {'train/loss': 0.17058366537094116, 'train/grad_norm_pre_clip': 15.61202335357666, 'timing/data': 0.037650576792657375, 'timing/model': 1.2206985093653202, 'learning_rate/qwen_vl_interface': 5.746883672158705e-07, 'learning_rate/action_model': 5.746883672158705e-06, 'epoch': 2.54}) +[2026-06-18 08:34:05,213][starVLA.training.train_starvla][INFO] - Step 6611, Loss: {'train/loss': 0.098858542740345, 'train/grad_norm_pre_clip': 7.5370402336120605, 'timing/data': 0.03025989793241024, 'timing/model': 1.1556624844670296, 'learning_rate/qwen_vl_interface': 5.743068414827726e-07, 'learning_rate/action_model': 5.743068414827726e-06, 'epoch': 2.54}) +[2026-06-18 08:34:10,356][starVLA.training.train_starvla][INFO] - Step 6612, Loss: {'train/loss': 0.36446481943130493, 'train/grad_norm_pre_clip': 23.934356689453125, 'timing/data': 0.04971396829932928, 'timing/model': 1.6785534098744392, 'learning_rate/qwen_vl_interface': 5.739262850258968e-07, 'learning_rate/action_model': 5.7392628502589684e-06, 'epoch': 2.54}) +[2026-06-18 08:34:14,908][starVLA.training.train_starvla][INFO] - Step 6613, Loss: {'train/loss': 0.43453070521354675, 'train/grad_norm_pre_clip': 12.763273239135742, 'timing/data': 0.029690970666706562, 'timing/model': 1.1548328343778849, 'learning_rate/qwen_vl_interface': 5.735466979241352e-07, 'learning_rate/action_model': 5.735466979241352e-06, 'epoch': 2.54}) +[2026-06-18 08:34:20,057][starVLA.training.train_starvla][INFO] - Step 6614, Loss: {'train/loss': 0.2093994915485382, 'train/grad_norm_pre_clip': 17.31865882873535, 'timing/data': 0.046963038854300976, 'timing/model': 1.1369261220097542, 'learning_rate/qwen_vl_interface': 5.731680802561744e-07, 'learning_rate/action_model': 5.731680802561745e-06, 'epoch': 2.54}) +[2026-06-18 08:34:25,003][starVLA.training.train_starvla][INFO] - Step 6615, Loss: {'train/loss': 0.21148599684238434, 'train/grad_norm_pre_clip': 13.51939868927002, 'timing/data': 0.04306750185787678, 'timing/model': 1.1611306387931108, 'learning_rate/qwen_vl_interface': 5.727904321005031e-07, 'learning_rate/action_model': 5.727904321005031e-06, 'epoch': 2.54}) +[2026-06-18 08:34:29,979][starVLA.training.train_starvla][INFO] - Step 6616, Loss: {'train/loss': 0.346510112285614, 'train/grad_norm_pre_clip': 18.4340763092041, 'timing/data': 0.03781290538609028, 'timing/model': 1.3401783118024468, 'learning_rate/qwen_vl_interface': 5.724137535354077e-07, 'learning_rate/action_model': 5.724137535354077e-06, 'epoch': 2.55}) +[2026-06-18 08:34:34,827][starVLA.training.train_starvla][INFO] - Step 6617, Loss: {'train/loss': 0.25289586186408997, 'train/grad_norm_pre_clip': 13.607319831848145, 'timing/data': 0.04571413714438677, 'timing/model': 1.2610528413206339, 'learning_rate/qwen_vl_interface': 5.720380446389748e-07, 'learning_rate/action_model': 5.720380446389748e-06, 'epoch': 2.55}) +[2026-06-18 08:34:40,075][starVLA.training.train_starvla][INFO] - Step 6618, Loss: {'train/loss': 0.4662849009037018, 'train/grad_norm_pre_clip': 32.33318328857422, 'timing/data': 0.02995824348181486, 'timing/model': 1.2882937779650092, 'learning_rate/qwen_vl_interface': 5.716633054890897e-07, 'learning_rate/action_model': 5.716633054890896e-06, 'epoch': 2.55}) +[2026-06-18 08:34:44,965][starVLA.training.train_starvla][INFO] - Step 6619, Loss: {'train/loss': 0.23780211806297302, 'train/grad_norm_pre_clip': 42.59870910644531, 'timing/data': 0.03653727937489748, 'timing/model': 1.188556332141161, 'learning_rate/qwen_vl_interface': 5.712895361634345e-07, 'learning_rate/action_model': 5.712895361634345e-06, 'epoch': 2.55}) +[2026-06-18 08:34:49,739][starVLA.training.train_starvla][INFO] - Step 6620, Loss: {'train/loss': 0.32165488600730896, 'train/grad_norm_pre_clip': 13.333028793334961, 'timing/data': 0.05020565539598465, 'timing/model': 1.1766122309491038, 'learning_rate/qwen_vl_interface': 5.709167367394936e-07, 'learning_rate/action_model': 5.709167367394936e-06, 'epoch': 2.55}) +[2026-06-18 08:34:54,815][starVLA.training.train_starvla][INFO] - Step 6621, Loss: {'train/loss': 0.2715204656124115, 'train/grad_norm_pre_clip': 12.042693138122559, 'timing/data': 0.03508228622376919, 'timing/model': 1.3332885559648275, 'learning_rate/qwen_vl_interface': 5.705449072945479e-07, 'learning_rate/action_model': 5.705449072945479e-06, 'epoch': 2.55}) +[2026-06-18 08:35:00,014][starVLA.training.train_starvla][INFO] - Step 6622, Loss: {'train/loss': 0.1676858514547348, 'train/grad_norm_pre_clip': 12.04612922668457, 'timing/data': 0.039810044690966606, 'timing/model': 1.575572825036943, 'learning_rate/qwen_vl_interface': 5.70174047905678e-07, 'learning_rate/action_model': 5.701740479056779e-06, 'epoch': 2.55}) +[2026-06-18 08:35:04,482][starVLA.training.train_starvla][INFO] - Step 6623, Loss: {'train/loss': 0.18664099276065826, 'train/grad_norm_pre_clip': 11.524774551391602, 'timing/data': 0.03270499221980572, 'timing/model': 1.1975288605317473, 'learning_rate/qwen_vl_interface': 5.69804158649764e-07, 'learning_rate/action_model': 5.6980415864976405e-06, 'epoch': 2.55}) +[2026-06-18 08:35:09,422][starVLA.training.train_starvla][INFO] - Step 6624, Loss: {'train/loss': 0.19875477254390717, 'train/grad_norm_pre_clip': 13.754849433898926, 'timing/data': 0.037612161599099636, 'timing/model': 1.09561587870121, 'learning_rate/qwen_vl_interface': 5.69435239603484e-07, 'learning_rate/action_model': 5.69435239603484e-06, 'epoch': 2.55}) +[2026-06-18 08:35:14,445][starVLA.training.train_starvla][INFO] - Step 6625, Loss: {'train/loss': 0.29065608978271484, 'train/grad_norm_pre_clip': 16.292827606201172, 'timing/data': 0.04002377297729254, 'timing/model': 1.1219522645696998, 'learning_rate/qwen_vl_interface': 5.690672908433149e-07, 'learning_rate/action_model': 5.690672908433149e-06, 'epoch': 2.55}) +[2026-06-18 08:35:19,154][starVLA.training.train_starvla][INFO] - Step 6626, Loss: {'train/loss': 0.35195016860961914, 'train/grad_norm_pre_clip': 19.90277099609375, 'timing/data': 0.04067708645015955, 'timing/model': 1.1405555456876755, 'learning_rate/qwen_vl_interface': 5.68700312445534e-07, 'learning_rate/action_model': 5.687003124455339e-06, 'epoch': 2.55}) +[2026-06-18 08:35:23,978][starVLA.training.train_starvla][INFO] - Step 6627, Loss: {'train/loss': 0.38199561834335327, 'train/grad_norm_pre_clip': 22.66289710998535, 'timing/data': 0.03539843577891588, 'timing/model': 1.4018283281475306, 'learning_rate/qwen_vl_interface': 5.683343044862151e-07, 'learning_rate/action_model': 5.68334304486215e-06, 'epoch': 2.55}) +[2026-06-18 08:35:29,127][starVLA.training.train_starvla][INFO] - Step 6628, Loss: {'train/loss': 0.35608750581741333, 'train/grad_norm_pre_clip': 16.476707458496094, 'timing/data': 0.029312129132449627, 'timing/model': 1.3425613986328244, 'learning_rate/qwen_vl_interface': 5.679692670412324e-07, 'learning_rate/action_model': 5.679692670412324e-06, 'epoch': 2.55}) +[2026-06-18 08:35:34,037][starVLA.training.train_starvla][INFO] - Step 6629, Loss: {'train/loss': 0.16486923396587372, 'train/grad_norm_pre_clip': 11.320717811584473, 'timing/data': 0.0439170291647315, 'timing/model': 1.1174969216808677, 'learning_rate/qwen_vl_interface': 5.67605200186259e-07, 'learning_rate/action_model': 5.67605200186259e-06, 'epoch': 2.55}) +[2026-06-18 08:35:39,074][starVLA.training.train_starvla][INFO] - Step 6630, Loss: {'train/loss': 0.1560305804014206, 'train/grad_norm_pre_clip': 11.932004928588867, 'timing/data': 0.03230235818773508, 'timing/model': 1.2724825209006667, 'learning_rate/qwen_vl_interface': 5.67242103996766e-07, 'learning_rate/action_model': 5.67242103996766e-06, 'epoch': 2.55}) +[2026-06-18 08:35:44,231][starVLA.training.train_starvla][INFO] - Step 6631, Loss: {'train/loss': 0.2391394078731537, 'train/grad_norm_pre_clip': 24.402795791625977, 'timing/data': 0.03976402524858713, 'timing/model': 1.6599793555215, 'learning_rate/qwen_vl_interface': 5.668799785480242e-07, 'learning_rate/action_model': 5.668799785480242e-06, 'epoch': 2.55}) +[2026-06-18 08:35:49,296][starVLA.training.train_starvla][INFO] - Step 6632, Loss: {'train/loss': 0.16160519421100616, 'train/grad_norm_pre_clip': 9.498442649841309, 'timing/data': 0.031187868677079678, 'timing/model': 1.6185781257227063, 'learning_rate/qwen_vl_interface': 5.665188239151019e-07, 'learning_rate/action_model': 5.665188239151018e-06, 'epoch': 2.55}) +[2026-06-18 08:35:53,832][starVLA.training.train_starvla][INFO] - Step 6633, Loss: {'train/loss': 0.19383610785007477, 'train/grad_norm_pre_clip': 12.641794204711914, 'timing/data': 0.0420237397775054, 'timing/model': 1.1458070389926434, 'learning_rate/qwen_vl_interface': 5.661586401728671e-07, 'learning_rate/action_model': 5.6615864017286705e-06, 'epoch': 2.55}) +[2026-06-18 08:35:58,778][starVLA.training.train_starvla][INFO] - Step 6634, Loss: {'train/loss': 0.39584609866142273, 'train/grad_norm_pre_clip': 11.674077987670898, 'timing/data': 0.031023365445435047, 'timing/model': 1.2190118078142405, 'learning_rate/qwen_vl_interface': 5.657994273959857e-07, 'learning_rate/action_model': 5.657994273959856e-06, 'epoch': 2.55}) +[2026-06-18 08:36:03,869][starVLA.training.train_starvla][INFO] - Step 6635, Loss: {'train/loss': 0.2013380229473114, 'train/grad_norm_pre_clip': 11.87641429901123, 'timing/data': 0.0379085773602128, 'timing/model': 1.1255372241139412, 'learning_rate/qwen_vl_interface': 5.654411856589233e-07, 'learning_rate/action_model': 5.654411856589233e-06, 'epoch': 2.55}) +[2026-06-18 08:36:09,516][starVLA.training.train_starvla][INFO] - Step 6636, Loss: {'train/loss': 0.3333539366722107, 'train/grad_norm_pre_clip': 20.607494354248047, 'timing/data': 0.031603192910552025, 'timing/model': 1.8113585114479065, 'learning_rate/qwen_vl_interface': 5.650839150359444e-07, 'learning_rate/action_model': 5.650839150359444e-06, 'epoch': 2.55}) +[2026-06-18 08:36:14,520][starVLA.training.train_starvla][INFO] - Step 6637, Loss: {'train/loss': 0.24705654382705688, 'train/grad_norm_pre_clip': 16.0615234375, 'timing/data': 0.035321129485964775, 'timing/model': 1.6348687028512359, 'learning_rate/qwen_vl_interface': 5.647276156011114e-07, 'learning_rate/action_model': 5.6472761560111135e-06, 'epoch': 2.55}) +[2026-06-18 08:36:19,033][starVLA.training.train_starvla][INFO] - Step 6638, Loss: {'train/loss': 0.1020660325884819, 'train/grad_norm_pre_clip': 10.56569766998291, 'timing/data': 0.035827876068651676, 'timing/model': 1.1157561894506216, 'learning_rate/qwen_vl_interface': 5.643722874282846e-07, 'learning_rate/action_model': 5.643722874282846e-06, 'epoch': 2.55}) +[2026-06-18 08:36:23,887][starVLA.training.train_starvla][INFO] - Step 6639, Loss: {'train/loss': 0.2962268590927124, 'train/grad_norm_pre_clip': 12.567924499511719, 'timing/data': 0.032022914849221706, 'timing/model': 1.1399704832583666, 'learning_rate/qwen_vl_interface': 5.640179305911253e-07, 'learning_rate/action_model': 5.640179305911253e-06, 'epoch': 2.55}) +[2026-06-18 08:36:29,185][starVLA.training.train_starvla][INFO] - Step 6640, Loss: {'train/loss': 0.2665000557899475, 'train/grad_norm_pre_clip': 16.04762077331543, 'timing/data': 0.029577740468084812, 'timing/model': 1.6807041503489017, 'learning_rate/qwen_vl_interface': 5.636645451630908e-07, 'learning_rate/action_model': 5.6366454516309084e-06, 'epoch': 2.55}) +[2026-06-18 08:36:34,381][starVLA.training.train_starvla][INFO] - Step 6641, Loss: {'train/loss': 0.12061050534248352, 'train/grad_norm_pre_clip': 11.575479507446289, 'timing/data': 0.039980288594961166, 'timing/model': 1.1960106929764152, 'learning_rate/qwen_vl_interface': 5.63312131217439e-07, 'learning_rate/action_model': 5.633121312174389e-06, 'epoch': 2.55}) +[2026-06-18 08:36:39,828][starVLA.training.train_starvla][INFO] - Step 6642, Loss: {'train/loss': 0.22054564952850342, 'train/grad_norm_pre_clip': 13.773982048034668, 'timing/data': 0.037861546501517296, 'timing/model': 1.7638007961213589, 'learning_rate/qwen_vl_interface': 5.629606888272253e-07, 'learning_rate/action_model': 5.629606888272252e-06, 'epoch': 2.56}) +[2026-06-18 08:36:44,321][starVLA.training.train_starvla][INFO] - Step 6643, Loss: {'train/loss': 0.3724786639213562, 'train/grad_norm_pre_clip': 20.581567764282227, 'timing/data': 0.029622756876051426, 'timing/model': 1.122500255703926, 'learning_rate/qwen_vl_interface': 5.626102180653046e-07, 'learning_rate/action_model': 5.626102180653045e-06, 'epoch': 2.56}) +[2026-06-18 08:36:49,544][starVLA.training.train_starvla][INFO] - Step 6644, Loss: {'train/loss': 0.4271824359893799, 'train/grad_norm_pre_clip': 21.45639419555664, 'timing/data': 0.03602653928101063, 'timing/model': 1.2803734270855784, 'learning_rate/qwen_vl_interface': 5.62260719004329e-07, 'learning_rate/action_model': 5.6226071900432895e-06, 'epoch': 2.56}) +[2026-06-18 08:36:54,425][starVLA.training.train_starvla][INFO] - Step 6645, Loss: {'train/loss': 0.39714837074279785, 'train/grad_norm_pre_clip': 17.516523361206055, 'timing/data': 0.035415783524513245, 'timing/model': 1.1100756293162704, 'learning_rate/qwen_vl_interface': 5.619121917167504e-07, 'learning_rate/action_model': 5.619121917167505e-06, 'epoch': 2.56}) +[2026-06-18 08:36:59,613][starVLA.training.train_starvla][INFO] - Step 6646, Loss: {'train/loss': 0.36310333013534546, 'train/grad_norm_pre_clip': 15.05724811553955, 'timing/data': 0.033789834938943386, 'timing/model': 1.7295235451310873, 'learning_rate/qwen_vl_interface': 5.615646362748199e-07, 'learning_rate/action_model': 5.6156463627482e-06, 'epoch': 2.56}) +[2026-06-18 08:37:04,664][starVLA.training.train_starvla][INFO] - Step 6647, Loss: {'train/loss': 0.3474632799625397, 'train/grad_norm_pre_clip': 17.153135299682617, 'timing/data': 0.03895420767366886, 'timing/model': 1.506606857292354, 'learning_rate/qwen_vl_interface': 5.612180527505854e-07, 'learning_rate/action_model': 5.612180527505854e-06, 'epoch': 2.56}) +[2026-06-18 08:37:09,497][starVLA.training.train_starvla][INFO] - Step 6648, Loss: {'train/loss': 0.10589325428009033, 'train/grad_norm_pre_clip': 7.090353488922119, 'timing/data': 0.04046729579567909, 'timing/model': 1.404395624063909, 'learning_rate/qwen_vl_interface': 5.608724412158931e-07, 'learning_rate/action_model': 5.608724412158931e-06, 'epoch': 2.56}) +[2026-06-18 08:37:14,339][starVLA.training.train_starvla][INFO] - Step 6649, Loss: {'train/loss': 0.2966426908969879, 'train/grad_norm_pre_clip': 16.61335563659668, 'timing/data': 0.03127008303999901, 'timing/model': 1.110004147514701, 'learning_rate/qwen_vl_interface': 5.605278017423897e-07, 'learning_rate/action_model': 5.605278017423897e-06, 'epoch': 2.56}) +[2026-06-18 08:37:19,105][starVLA.training.train_starvla][INFO] - Step 6650, Loss: {'train/loss': 0.2331046164035797, 'train/grad_norm_pre_clip': 13.912668228149414, 'timing/data': 0.03706999495625496, 'timing/model': 1.185834113508463, 'learning_rate/qwen_vl_interface': 5.601841344015193e-07, 'learning_rate/action_model': 5.601841344015193e-06, 'epoch': 2.56}) +[2026-06-18 08:37:24,066][starVLA.training.train_starvla][INFO] - Step 6651, Loss: {'train/loss': 0.3253985345363617, 'train/grad_norm_pre_clip': 15.697535514831543, 'timing/data': 0.03173573035746813, 'timing/model': 1.2226365422829986, 'learning_rate/qwen_vl_interface': 5.598414392645239e-07, 'learning_rate/action_model': 5.598414392645239e-06, 'epoch': 2.56}) +[2026-06-18 08:37:28,918][starVLA.training.train_starvla][INFO] - Step 6652, Loss: {'train/loss': 0.49595052003860474, 'train/grad_norm_pre_clip': 15.446686744689941, 'timing/data': 0.046773575246334076, 'timing/model': 1.4861411936581135, 'learning_rate/qwen_vl_interface': 5.594997164024454e-07, 'learning_rate/action_model': 5.594997164024454e-06, 'epoch': 2.56}) +[2026-06-18 08:37:33,361][starVLA.training.train_starvla][INFO] - Step 6653, Loss: {'train/loss': 0.2614762485027313, 'train/grad_norm_pre_clip': 14.154535293579102, 'timing/data': 0.03298873081803322, 'timing/model': 1.084420752711594, 'learning_rate/qwen_vl_interface': 5.591589658861228e-07, 'learning_rate/action_model': 5.591589658861228e-06, 'epoch': 2.56}) +[2026-06-18 08:37:38,995][starVLA.training.train_starvla][INFO] - Step 6654, Loss: {'train/loss': 0.37761053442955017, 'train/grad_norm_pre_clip': 42.55133056640625, 'timing/data': 0.05362649168819189, 'timing/model': 1.50692471396178, 'learning_rate/qwen_vl_interface': 5.588191877861945e-07, 'learning_rate/action_model': 5.588191877861944e-06, 'epoch': 2.56}) +[2026-06-18 08:37:43,809][starVLA.training.train_starvla][INFO] - Step 6655, Loss: {'train/loss': 0.2512769103050232, 'train/grad_norm_pre_clip': 12.332627296447754, 'timing/data': 0.030883303843438625, 'timing/model': 1.095008261501789, 'learning_rate/qwen_vl_interface': 5.584803821730961e-07, 'learning_rate/action_model': 5.58480382173096e-06, 'epoch': 2.56}) +[2026-06-18 08:37:49,156][starVLA.training.train_starvla][INFO] - Step 6656, Loss: {'train/loss': 0.1783621907234192, 'train/grad_norm_pre_clip': 13.439728736877441, 'timing/data': 0.03596468176692724, 'timing/model': 1.5788664603605866, 'learning_rate/qwen_vl_interface': 5.581425491170627e-07, 'learning_rate/action_model': 5.581425491170626e-06, 'epoch': 2.56}) +[2026-06-18 08:37:54,494][starVLA.training.train_starvla][INFO] - Step 6657, Loss: {'train/loss': 0.19739091396331787, 'train/grad_norm_pre_clip': 18.18336296081543, 'timing/data': 0.031224854290485382, 'timing/model': 1.4248112700879574, 'learning_rate/qwen_vl_interface': 5.578056886881274e-07, 'learning_rate/action_model': 5.578056886881274e-06, 'epoch': 2.56}) +[2026-06-18 08:37:59,443][starVLA.training.train_starvla][INFO] - Step 6658, Loss: {'train/loss': 0.37084800004959106, 'train/grad_norm_pre_clip': 18.823087692260742, 'timing/data': 0.045736213214695454, 'timing/model': 1.0941938990727067, 'learning_rate/qwen_vl_interface': 5.574698009561223e-07, 'learning_rate/action_model': 5.574698009561223e-06, 'epoch': 2.56}) +[2026-06-18 08:38:04,442][starVLA.training.train_starvla][INFO] - Step 6659, Loss: {'train/loss': 0.39369192719459534, 'train/grad_norm_pre_clip': 14.792750358581543, 'timing/data': 0.03542435634881258, 'timing/model': 1.1792626176029444, 'learning_rate/qwen_vl_interface': 5.571348859906763e-07, 'learning_rate/action_model': 5.571348859906764e-06, 'epoch': 2.56}) +[2026-06-18 08:38:09,834][starVLA.training.train_starvla][INFO] - Step 6660, Loss: {'train/loss': 0.2167985439300537, 'train/grad_norm_pre_clip': 10.692450523376465, 'timing/data': 0.03378216549754143, 'timing/model': 1.3661908637732267, 'learning_rate/qwen_vl_interface': 5.568009438612186e-07, 'learning_rate/action_model': 5.568009438612186e-06, 'epoch': 2.56}) +[2026-06-18 08:38:14,575][starVLA.training.train_starvla][INFO] - Step 6661, Loss: {'train/loss': 0.2268834412097931, 'train/grad_norm_pre_clip': 12.020698547363281, 'timing/data': 0.03331805020570755, 'timing/model': 1.0907764043658972, 'learning_rate/qwen_vl_interface': 5.564679746369755e-07, 'learning_rate/action_model': 5.5646797463697556e-06, 'epoch': 2.56}) +[2026-06-18 08:38:19,340][starVLA.training.train_starvla][INFO] - Step 6662, Loss: {'train/loss': 0.1691448986530304, 'train/grad_norm_pre_clip': 11.850875854492188, 'timing/data': 0.04632977116852999, 'timing/model': 1.5082600181922317, 'learning_rate/qwen_vl_interface': 5.561359783869712e-07, 'learning_rate/action_model': 5.5613597838697116e-06, 'epoch': 2.56}) +[2026-06-18 08:38:23,999][starVLA.training.train_starvla][INFO] - Step 6663, Loss: {'train/loss': 0.2463746964931488, 'train/grad_norm_pre_clip': 20.54195213317871, 'timing/data': 0.049313273280858994, 'timing/model': 1.2186440089717507, 'learning_rate/qwen_vl_interface': 5.558049551800291e-07, 'learning_rate/action_model': 5.558049551800292e-06, 'epoch': 2.56}) +[2026-06-18 08:38:28,889][starVLA.training.train_starvla][INFO] - Step 6664, Loss: {'train/loss': 0.37677937746047974, 'train/grad_norm_pre_clip': 23.77278709411621, 'timing/data': 0.04257429298013449, 'timing/model': 1.1270022056996822, 'learning_rate/qwen_vl_interface': 5.554749050847718e-07, 'learning_rate/action_model': 5.5547490508477176e-06, 'epoch': 2.56}) +[2026-06-18 08:38:34,143][starVLA.training.train_starvla][INFO] - Step 6665, Loss: {'train/loss': 0.26591378450393677, 'train/grad_norm_pre_clip': 15.548792839050293, 'timing/data': 0.03647431451827288, 'timing/model': 1.1332466546446085, 'learning_rate/qwen_vl_interface': 5.55145828169618e-07, 'learning_rate/action_model': 5.5514582816961805e-06, 'epoch': 2.56}) +[2026-06-18 08:38:39,122][starVLA.training.train_starvla][INFO] - Step 6666, Loss: {'train/loss': 0.19634807109832764, 'train/grad_norm_pre_clip': 9.234899520874023, 'timing/data': 0.03740218095481396, 'timing/model': 1.1044845273718238, 'learning_rate/qwen_vl_interface': 5.548177245027859e-07, 'learning_rate/action_model': 5.548177245027859e-06, 'epoch': 2.56}) +[2026-06-18 08:38:44,384][starVLA.training.train_starvla][INFO] - Step 6667, Loss: {'train/loss': 0.223475843667984, 'train/grad_norm_pre_clip': 14.771920204162598, 'timing/data': 0.029783868230879307, 'timing/model': 1.890871367417276, 'learning_rate/qwen_vl_interface': 5.544905941522913e-07, 'learning_rate/action_model': 5.544905941522913e-06, 'epoch': 2.56}) +[2026-06-18 08:38:48,949][starVLA.training.train_starvla][INFO] - Step 6668, Loss: {'train/loss': 0.2624876797199249, 'train/grad_norm_pre_clip': 13.06551742553711, 'timing/data': 0.04428804013878107, 'timing/model': 1.1242189928889275, 'learning_rate/qwen_vl_interface': 5.541644371859498e-07, 'learning_rate/action_model': 5.541644371859498e-06, 'epoch': 2.57}) +[2026-06-18 08:38:53,951][starVLA.training.train_starvla][INFO] - Step 6669, Loss: {'train/loss': 0.27815642952919006, 'train/grad_norm_pre_clip': 17.411048889160156, 'timing/data': 0.03604811429977417, 'timing/model': 1.111854737624526, 'learning_rate/qwen_vl_interface': 5.53839253671373e-07, 'learning_rate/action_model': 5.5383925367137305e-06, 'epoch': 2.57}) +[2026-06-18 08:38:59,032][starVLA.training.train_starvla][INFO] - Step 6670, Loss: {'train/loss': 0.09720304608345032, 'train/grad_norm_pre_clip': 7.29660701751709, 'timing/data': 0.03450593538582325, 'timing/model': 1.1761514591053128, 'learning_rate/qwen_vl_interface': 5.535150436759725e-07, 'learning_rate/action_model': 5.535150436759725e-06, 'epoch': 2.57}) +[2026-06-18 08:39:03,695][starVLA.training.train_starvla][INFO] - Step 6671, Loss: {'train/loss': 0.47560742497444153, 'train/grad_norm_pre_clip': 19.907028198242188, 'timing/data': 0.052045359276235104, 'timing/model': 1.1498355939984322, 'learning_rate/qwen_vl_interface': 5.531918072669566e-07, 'learning_rate/action_model': 5.531918072669565e-06, 'epoch': 2.57}) +[2026-06-18 08:39:08,689][starVLA.training.train_starvla][INFO] - Step 6672, Loss: {'train/loss': 0.3614867925643921, 'train/grad_norm_pre_clip': 17.365062713623047, 'timing/data': 0.03934333100914955, 'timing/model': 1.534235194325447, 'learning_rate/qwen_vl_interface': 5.528695445113336e-07, 'learning_rate/action_model': 5.528695445113335e-06, 'epoch': 2.57}) +[2026-06-18 08:39:13,439][starVLA.training.train_starvla][INFO] - Step 6673, Loss: {'train/loss': 0.46575692296028137, 'train/grad_norm_pre_clip': 18.750707626342773, 'timing/data': 0.031462074257433414, 'timing/model': 1.0774338385090232, 'learning_rate/qwen_vl_interface': 5.525482554759079e-07, 'learning_rate/action_model': 5.5254825547590795e-06, 'epoch': 2.57}) +[2026-06-18 08:39:18,050][starVLA.training.train_starvla][INFO] - Step 6674, Loss: {'train/loss': 0.21640653908252716, 'train/grad_norm_pre_clip': 14.074544906616211, 'timing/data': 0.033723688684403896, 'timing/model': 1.1361582465469837, 'learning_rate/qwen_vl_interface': 5.522279402272834e-07, 'learning_rate/action_model': 5.522279402272834e-06, 'epoch': 2.57}) +[2026-06-18 08:39:22,588][starVLA.training.train_starvla][INFO] - Step 6675, Loss: {'train/loss': 0.2597249746322632, 'train/grad_norm_pre_clip': 16.42218589782715, 'timing/data': 0.027265416458249092, 'timing/model': 1.098812802694738, 'learning_rate/qwen_vl_interface': 5.519085988318627e-07, 'learning_rate/action_model': 5.5190859883186265e-06, 'epoch': 2.57}) +[2026-06-18 08:39:27,245][starVLA.training.train_starvla][INFO] - Step 6676, Loss: {'train/loss': 0.39079299569129944, 'train/grad_norm_pre_clip': 18.505348205566406, 'timing/data': 0.033988917246460915, 'timing/model': 1.4364756420254707, 'learning_rate/qwen_vl_interface': 5.515902313558443e-07, 'learning_rate/action_model': 5.515902313558444e-06, 'epoch': 2.57}) +[2026-06-18 08:39:31,777][starVLA.training.train_starvla][INFO] - Step 6677, Loss: {'train/loss': 0.36197251081466675, 'train/grad_norm_pre_clip': 15.911251068115234, 'timing/data': 0.034949785098433495, 'timing/model': 1.353174690157175, 'learning_rate/qwen_vl_interface': 5.512728378652266e-07, 'learning_rate/action_model': 5.512728378652266e-06, 'epoch': 2.57}) +[2026-06-18 08:39:36,311][starVLA.training.train_starvla][INFO] - Step 6678, Loss: {'train/loss': 0.25373199582099915, 'train/grad_norm_pre_clip': 12.624086380004883, 'timing/data': 0.029101415537297726, 'timing/model': 1.0738778989762068, 'learning_rate/qwen_vl_interface': 5.509564184258055e-07, 'learning_rate/action_model': 5.509564184258055e-06, 'epoch': 2.57}) +[2026-06-18 08:39:40,918][starVLA.training.train_starvla][INFO] - Step 6679, Loss: {'train/loss': 0.27975699305534363, 'train/grad_norm_pre_clip': 14.276092529296875, 'timing/data': 0.03784969262778759, 'timing/model': 1.0914584323763847, 'learning_rate/qwen_vl_interface': 5.506409731031754e-07, 'learning_rate/action_model': 5.506409731031755e-06, 'epoch': 2.57}) +[2026-06-18 08:39:45,399][starVLA.training.train_starvla][INFO] - Step 6680, Loss: {'train/loss': 0.2966809570789337, 'train/grad_norm_pre_clip': 27.841901779174805, 'timing/data': 0.029666364192962646, 'timing/model': 1.0658552097156644, 'learning_rate/qwen_vl_interface': 5.503265019627289e-07, 'learning_rate/action_model': 5.503265019627289e-06, 'epoch': 2.57}) +[2026-06-18 08:39:49,791][starVLA.training.train_starvla][INFO] - Step 6681, Loss: {'train/loss': 0.20974071323871613, 'train/grad_norm_pre_clip': 15.853923797607422, 'timing/data': 0.0328110633417964, 'timing/model': 1.0935542061924934, 'learning_rate/qwen_vl_interface': 5.500130050696548e-07, 'learning_rate/action_model': 5.500130050696548e-06, 'epoch': 2.57}) +[2026-06-18 08:39:54,367][starVLA.training.train_starvla][INFO] - Step 6682, Loss: {'train/loss': 0.27301985025405884, 'train/grad_norm_pre_clip': 14.523914337158203, 'timing/data': 0.028807351365685463, 'timing/model': 1.3217610511928797, 'learning_rate/qwen_vl_interface': 5.497004824889424e-07, 'learning_rate/action_model': 5.497004824889423e-06, 'epoch': 2.57}) +[2026-06-18 08:39:58,666][starVLA.training.train_starvla][INFO] - Step 6683, Loss: {'train/loss': 0.36800578236579895, 'train/grad_norm_pre_clip': 18.481382369995117, 'timing/data': 0.02857463527470827, 'timing/model': 1.1792359221726656, 'learning_rate/qwen_vl_interface': 5.493889342853779e-07, 'learning_rate/action_model': 5.493889342853778e-06, 'epoch': 2.57}) +[2026-06-18 08:40:03,137][starVLA.training.train_starvla][INFO] - Step 6684, Loss: {'train/loss': 0.4068857729434967, 'train/grad_norm_pre_clip': 16.930431365966797, 'timing/data': 0.026517395861446857, 'timing/model': 1.0831444608047605, 'learning_rate/qwen_vl_interface': 5.490783605235451e-07, 'learning_rate/action_model': 5.49078360523545e-06, 'epoch': 2.57}) +[2026-06-18 08:40:07,724][starVLA.training.train_starvla][INFO] - Step 6685, Loss: {'train/loss': 0.22311978042125702, 'train/grad_norm_pre_clip': 15.67536735534668, 'timing/data': 0.0344186881557107, 'timing/model': 1.1130338162183762, 'learning_rate/qwen_vl_interface': 5.487687612678265e-07, 'learning_rate/action_model': 5.4876876126782645e-06, 'epoch': 2.57}) +[2026-06-18 08:40:12,492][starVLA.training.train_starvla][INFO] - Step 6686, Loss: {'train/loss': 0.17090873420238495, 'train/grad_norm_pre_clip': 23.231098175048828, 'timing/data': 0.0390434255823493, 'timing/model': 1.5387163423001766, 'learning_rate/qwen_vl_interface': 5.484601365824032e-07, 'learning_rate/action_model': 5.4846013658240315e-06, 'epoch': 2.57}) +[2026-06-18 08:40:17,419][starVLA.training.train_starvla][INFO] - Step 6687, Loss: {'train/loss': 0.2729446589946747, 'train/grad_norm_pre_clip': 19.04349708557129, 'timing/data': 0.04128595255315304, 'timing/model': 1.4925001049414277, 'learning_rate/qwen_vl_interface': 5.481524865312518e-07, 'learning_rate/action_model': 5.481524865312518e-06, 'epoch': 2.57}) +[2026-06-18 08:40:21,836][starVLA.training.train_starvla][INFO] - Step 6688, Loss: {'train/loss': 0.3188737630844116, 'train/grad_norm_pre_clip': 10.290486335754395, 'timing/data': 0.035229407250881195, 'timing/model': 1.204524657689035, 'learning_rate/qwen_vl_interface': 5.478458111781496e-07, 'learning_rate/action_model': 5.478458111781496e-06, 'epoch': 2.57}) +[2026-06-18 08:40:26,969][starVLA.training.train_starvla][INFO] - Step 6689, Loss: {'train/loss': 0.2308780550956726, 'train/grad_norm_pre_clip': 19.24030876159668, 'timing/data': 0.03520475793629885, 'timing/model': 1.1271134791895747, 'learning_rate/qwen_vl_interface': 5.475401105866712e-07, 'learning_rate/action_model': 5.4754011058667114e-06, 'epoch': 2.57}) +[2026-06-18 08:40:31,825][starVLA.training.train_starvla][INFO] - Step 6690, Loss: {'train/loss': 0.3609601557254791, 'train/grad_norm_pre_clip': 14.34626293182373, 'timing/data': 0.03518132492899895, 'timing/model': 1.1153286341577768, 'learning_rate/qwen_vl_interface': 5.47235384820187e-07, 'learning_rate/action_model': 5.47235384820187e-06, 'epoch': 2.57}) +[2026-06-18 08:40:36,657][starVLA.training.train_starvla][INFO] - Step 6691, Loss: {'train/loss': 0.2964269518852234, 'train/grad_norm_pre_clip': 17.439350128173828, 'timing/data': 0.04918251093477011, 'timing/model': 1.1322913533076644, 'learning_rate/qwen_vl_interface': 5.469316339418681e-07, 'learning_rate/action_model': 5.469316339418681e-06, 'epoch': 2.57}) +[2026-06-18 08:40:41,405][starVLA.training.train_starvla][INFO] - Step 6692, Loss: {'train/loss': 0.3018046021461487, 'train/grad_norm_pre_clip': 15.536656379699707, 'timing/data': 0.03484780341386795, 'timing/model': 1.4623310593888164, 'learning_rate/qwen_vl_interface': 5.466288580146825e-07, 'learning_rate/action_model': 5.466288580146825e-06, 'epoch': 2.57}) +[2026-06-18 08:40:46,026][starVLA.training.train_starvla][INFO] - Step 6693, Loss: {'train/loss': 0.13256558775901794, 'train/grad_norm_pre_clip': 12.20449447631836, 'timing/data': 0.03272056486457586, 'timing/model': 1.2495667887851596, 'learning_rate/qwen_vl_interface': 5.463270571013954e-07, 'learning_rate/action_model': 5.463270571013954e-06, 'epoch': 2.57}) +[2026-06-18 08:40:51,032][starVLA.training.train_starvla][INFO] - Step 6694, Loss: {'train/loss': 0.29063984751701355, 'train/grad_norm_pre_clip': 16.265357971191406, 'timing/data': 0.038890198804438114, 'timing/model': 1.144409322179854, 'learning_rate/qwen_vl_interface': 5.460262312645709e-07, 'learning_rate/action_model': 5.460262312645709e-06, 'epoch': 2.58}) +[2026-06-18 08:40:55,713][starVLA.training.train_starvla][INFO] - Step 6695, Loss: {'train/loss': 0.2525142729282379, 'train/grad_norm_pre_clip': 21.488689422607422, 'timing/data': 0.028420492075383663, 'timing/model': 1.0978293968364596, 'learning_rate/qwen_vl_interface': 5.457263805665707e-07, 'learning_rate/action_model': 5.457263805665707e-06, 'epoch': 2.58}) +[2026-06-18 08:41:00,836][starVLA.training.train_starvla][INFO] - Step 6696, Loss: {'train/loss': 0.1379857212305069, 'train/grad_norm_pre_clip': 11.447270393371582, 'timing/data': 0.03375409636646509, 'timing/model': 1.4477123115211725, 'learning_rate/qwen_vl_interface': 5.454275050695529e-07, 'learning_rate/action_model': 5.454275050695529e-06, 'epoch': 2.58}) +[2026-06-18 08:41:05,641][starVLA.training.train_starvla][INFO] - Step 6697, Loss: {'train/loss': 0.28790053725242615, 'train/grad_norm_pre_clip': 17.725570678710938, 'timing/data': 0.04713507927954197, 'timing/model': 1.4546050084754825, 'learning_rate/qwen_vl_interface': 5.451296048354764e-07, 'learning_rate/action_model': 5.4512960483547635e-06, 'epoch': 2.58}) +[2026-06-18 08:41:10,371][starVLA.training.train_starvla][INFO] - Step 6698, Loss: {'train/loss': 0.26569682359695435, 'train/grad_norm_pre_clip': 14.481097221374512, 'timing/data': 0.05025643203407526, 'timing/model': 1.2686799969524145, 'learning_rate/qwen_vl_interface': 5.448326799260951e-07, 'learning_rate/action_model': 5.448326799260952e-06, 'epoch': 2.58}) +[2026-06-18 08:41:15,517][starVLA.training.train_starvla][INFO] - Step 6699, Loss: {'train/loss': 0.42872053384780884, 'train/grad_norm_pre_clip': 19.849517822265625, 'timing/data': 0.038395777344703674, 'timing/model': 1.3553973156958818, 'learning_rate/qwen_vl_interface': 5.445367304029627e-07, 'learning_rate/action_model': 5.445367304029626e-06, 'epoch': 2.58}) +[2026-06-18 08:43:55,123][starVLA.training.train_starvla][INFO] - Step 6700, Loss: {'train/loss': 0.2739364504814148, 'train/grad_norm_pre_clip': 13.3808012008667, 'eval/loss': 0.42900779843330383, 'eval/latency_4/loss': 0.42900779843330383, 'eval/demon_fire_precision': 0.9671136653895275, 'eval/demon_fire_recall': 0.983760961351088, 'eval/demon_fire_f1': 0.975366285622283, 'eval/demon_fire_rate_model': 0.9198237885462555, 'eval/demon_fire_rate_teacher': 0.9042584434654919, 'eval/demon_left_precision': 0.8063660477453581, 'eval/demon_left_recall': 0.6940639269406392, 'eval/demon_left_f1': 0.7460122699386503, 'eval/demon_left_rate_model': 0.11071953010279001, 'eval/demon_left_rate_teacher': 0.12863436123348018, 'eval/demon_right_precision': 0.7924187725631769, 'eval/demon_right_recall': 0.7924187725631769, 'eval/demon_right_f1': 0.7924187725631769, 'eval/demon_right_rate_model': 0.16270190895741557, 'eval/demon_right_rate_teacher': 0.16270190895741557, 'eval/demon_move_f1': 0.7692155212509136, 'eval/demon_cc_f1': 0.902782239803595, 'eval/latency_4/demon_fire_precision': 0.9671136653895275, 'eval/latency_4/demon_fire_recall': 0.983760961351088, 'eval/latency_4/demon_fire_f1': 0.975366285622283, 'eval/latency_4/demon_fire_rate_model': 0.9198237885462555, 'eval/latency_4/demon_fire_rate_teacher': 0.9042584434654919, 'eval/latency_4/demon_left_precision': 0.8063660477453581, 'eval/latency_4/demon_left_recall': 0.6940639269406392, 'eval/latency_4/demon_left_f1': 0.7460122699386503, 'eval/latency_4/demon_left_rate_model': 0.11071953010279001, 'eval/latency_4/demon_left_rate_teacher': 0.12863436123348018, 'eval/latency_4/demon_right_precision': 0.7924187725631769, 'eval/latency_4/demon_right_recall': 0.7924187725631769, 'eval/latency_4/demon_right_f1': 0.7924187725631769, 'eval/latency_4/demon_right_rate_model': 0.16270190895741557, 'eval/latency_4/demon_right_rate_teacher': 0.16270190895741557, 'eval/latency_4/demon_move_f1': 0.7692155212509136, 'eval/latency_4/demon_cc_f1': 0.902782239803595, 'timing/data': 0.04108897875994444, 'timing/model': 1.2423442648723722, 'learning_rate/qwen_vl_interface': 5.442417563274289e-07, 'learning_rate/action_model': 5.442417563274289e-06, 'epoch': 2.58}) +[2026-06-18 08:43:55,124][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_6700_state +[2026-06-18 08:43:55,131][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 08:44:04,678][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_6700_state/model.safetensors +[2026-06-18 08:44:27,803][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_6700_state/optimizer.bin +[2026-06-18 08:44:27,805][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_6700_state/sampler.bin +[2026-06-18 08:44:27,806][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_6700_state/sampler_1.bin +[2026-06-18 08:44:27,809][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_6700_state/random_states_0.pkl +[2026-06-18 08:44:27,831][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 6700 +[2026-06-18 08:44:27,831][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 08:44:27,849][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 08:44:32,698][starVLA.training.train_starvla][INFO] - Step 6701, Loss: {'train/loss': 0.2173326164484024, 'train/grad_norm_pre_clip': 11.517054557800293, 'timing/data': 0.039598449133336544, 'timing/model': 1.5808341884985566, 'learning_rate/qwen_vl_interface': 5.439477577606429e-07, 'learning_rate/action_model': 5.4394775776064295e-06, 'epoch': 2.58}) +[2026-06-18 08:44:37,456][starVLA.training.train_starvla][INFO] - Step 6702, Loss: {'train/loss': 0.3211553394794464, 'train/grad_norm_pre_clip': 14.8537015914917, 'timing/data': 0.03292183857411146, 'timing/model': 1.4903903529047966, 'learning_rate/qwen_vl_interface': 5.436547347635506e-07, 'learning_rate/action_model': 5.436547347635506e-06, 'epoch': 2.58}) +[2026-06-18 08:44:41,772][starVLA.training.train_starvla][INFO] - Step 6703, Loss: {'train/loss': 0.22983230650424957, 'train/grad_norm_pre_clip': 13.775845527648926, 'timing/data': 0.0329756923019886, 'timing/model': 1.1055761706084013, 'learning_rate/qwen_vl_interface': 5.433626873968962e-07, 'learning_rate/action_model': 5.433626873968962e-06, 'epoch': 2.58}) +[2026-06-18 08:44:46,609][starVLA.training.train_starvla][INFO] - Step 6704, Loss: {'train/loss': 0.14095693826675415, 'train/grad_norm_pre_clip': 10.342299461364746, 'timing/data': 0.03910861164331436, 'timing/model': 1.0914341760799289, 'learning_rate/qwen_vl_interface': 5.43071615721221e-07, 'learning_rate/action_model': 5.43071615721221e-06, 'epoch': 2.58}) +[2026-06-18 08:44:51,402][starVLA.training.train_starvla][INFO] - Step 6705, Loss: {'train/loss': 0.38398322463035583, 'train/grad_norm_pre_clip': 19.874849319458008, 'timing/data': 0.02927678357809782, 'timing/model': 1.1403518412262201, 'learning_rate/qwen_vl_interface': 5.427815197968647e-07, 'learning_rate/action_model': 5.4278151979686465e-06, 'epoch': 2.58}) +[2026-06-18 08:44:56,172][starVLA.training.train_starvla][INFO] - Step 6706, Loss: {'train/loss': 0.202137291431427, 'train/grad_norm_pre_clip': 13.607650756835938, 'timing/data': 0.04261422622948885, 'timing/model': 1.1475852904841304, 'learning_rate/qwen_vl_interface': 5.424923996839647e-07, 'learning_rate/action_model': 5.4249239968396466e-06, 'epoch': 2.58}) +[2026-06-18 08:45:00,997][starVLA.training.train_starvla][INFO] - Step 6707, Loss: {'train/loss': 0.35474374890327454, 'train/grad_norm_pre_clip': 20.518178939819336, 'timing/data': 0.03681542631238699, 'timing/model': 1.4432014031335711, 'learning_rate/qwen_vl_interface': 5.422042554424557e-07, 'learning_rate/action_model': 5.4220425544245565e-06, 'epoch': 2.58}) +[2026-06-18 08:45:05,679][starVLA.training.train_starvla][INFO] - Step 6708, Loss: {'train/loss': 0.42688336968421936, 'train/grad_norm_pre_clip': 20.00811767578125, 'timing/data': 0.04103308543562889, 'timing/model': 1.2125189220532775, 'learning_rate/qwen_vl_interface': 5.419170871320705e-07, 'learning_rate/action_model': 5.419170871320706e-06, 'epoch': 2.58}) +[2026-06-18 08:45:10,520][starVLA.training.train_starvla][INFO] - Step 6709, Loss: {'train/loss': 0.40788546204566956, 'train/grad_norm_pre_clip': 22.947168350219727, 'timing/data': 0.03296888247132301, 'timing/model': 1.0788011969998479, 'learning_rate/qwen_vl_interface': 5.41630894812339e-07, 'learning_rate/action_model': 5.41630894812339e-06, 'epoch': 2.58}) +[2026-06-18 08:45:15,460][starVLA.training.train_starvla][INFO] - Step 6710, Loss: {'train/loss': 0.14042764902114868, 'train/grad_norm_pre_clip': 10.657144546508789, 'timing/data': 0.03994155861437321, 'timing/model': 1.1991037856787443, 'learning_rate/qwen_vl_interface': 5.413456785425896e-07, 'learning_rate/action_model': 5.413456785425896e-06, 'epoch': 2.58}) +[2026-06-18 08:45:20,466][starVLA.training.train_starvla][INFO] - Step 6711, Loss: {'train/loss': 0.2449515461921692, 'train/grad_norm_pre_clip': 14.463042259216309, 'timing/data': 0.0345027856528759, 'timing/model': 1.2712144441902637, 'learning_rate/qwen_vl_interface': 5.410614383819474e-07, 'learning_rate/action_model': 5.4106143838194736e-06, 'epoch': 2.58}) +[2026-06-18 08:45:25,269][starVLA.training.train_starvla][INFO] - Step 6712, Loss: {'train/loss': 0.1677590161561966, 'train/grad_norm_pre_clip': 12.654776573181152, 'timing/data': 0.034724499098956585, 'timing/model': 1.535676509141922, 'learning_rate/qwen_vl_interface': 5.407781743893364e-07, 'learning_rate/action_model': 5.407781743893364e-06, 'epoch': 2.58}) +[2026-06-18 08:45:29,738][starVLA.training.train_starvla][INFO] - Step 6713, Loss: {'train/loss': 0.11662033200263977, 'train/grad_norm_pre_clip': 14.123540878295898, 'timing/data': 0.046707297675311565, 'timing/model': 1.1056369505822659, 'learning_rate/qwen_vl_interface': 5.404958866234772e-07, 'learning_rate/action_model': 5.404958866234772e-06, 'epoch': 2.58}) +[2026-06-18 08:45:34,631][starVLA.training.train_starvla][INFO] - Step 6714, Loss: {'train/loss': 0.3658922016620636, 'train/grad_norm_pre_clip': 22.824975967407227, 'timing/data': 0.038834053091704845, 'timing/model': 1.1957687307149172, 'learning_rate/qwen_vl_interface': 5.402145751428884e-07, 'learning_rate/action_model': 5.402145751428884e-06, 'epoch': 2.58}) +[2026-06-18 08:45:39,502][starVLA.training.train_starvla][INFO] - Step 6715, Loss: {'train/loss': 0.3067591190338135, 'train/grad_norm_pre_clip': 16.472158432006836, 'timing/data': 0.033624510280787945, 'timing/model': 1.1446380419656634, 'learning_rate/qwen_vl_interface': 5.399342400058856e-07, 'learning_rate/action_model': 5.399342400058856e-06, 'epoch': 2.58}) +[2026-06-18 08:45:44,248][starVLA.training.train_starvla][INFO] - Step 6716, Loss: {'train/loss': 0.4020431339740753, 'train/grad_norm_pre_clip': 17.133100509643555, 'timing/data': 0.0431442940607667, 'timing/model': 1.122418432496488, 'learning_rate/qwen_vl_interface': 5.396548812705838e-07, 'learning_rate/action_model': 5.396548812705839e-06, 'epoch': 2.58}) +[2026-06-18 08:45:48,626][starVLA.training.train_starvla][INFO] - Step 6717, Loss: {'train/loss': 0.26550257205963135, 'train/grad_norm_pre_clip': 18.0664119720459, 'timing/data': 0.0350856464356184, 'timing/model': 1.10605579148978, 'learning_rate/qwen_vl_interface': 5.393764989948933e-07, 'learning_rate/action_model': 5.393764989948933e-06, 'epoch': 2.58}) +[2026-06-18 08:45:53,783][starVLA.training.train_starvla][INFO] - Step 6718, Loss: {'train/loss': 0.2930189371109009, 'train/grad_norm_pre_clip': 21.870807647705078, 'timing/data': 0.03361089341342449, 'timing/model': 1.2039108239114285, 'learning_rate/qwen_vl_interface': 5.390990932365234e-07, 'learning_rate/action_model': 5.390990932365233e-06, 'epoch': 2.58}) +[2026-06-18 08:45:59,019][starVLA.training.train_starvla][INFO] - Step 6719, Loss: {'train/loss': 0.11867062747478485, 'train/grad_norm_pre_clip': 9.460601806640625, 'timing/data': 0.04173856042325497, 'timing/model': 1.2564069014042616, 'learning_rate/qwen_vl_interface': 5.388226640529808e-07, 'learning_rate/action_model': 5.388226640529808e-06, 'epoch': 2.58}) +[2026-06-18 08:46:03,949][starVLA.training.train_starvla][INFO] - Step 6720, Loss: {'train/loss': 0.07434205710887909, 'train/grad_norm_pre_clip': 6.245284557342529, 'timing/data': 0.04788431245833635, 'timing/model': 1.145395090803504, 'learning_rate/qwen_vl_interface': 5.385472115015697e-07, 'learning_rate/action_model': 5.385472115015696e-06, 'epoch': 2.59}) +[2026-06-18 08:46:09,319][starVLA.training.train_starvla][INFO] - Step 6721, Loss: {'train/loss': 0.21663987636566162, 'train/grad_norm_pre_clip': 12.405311584472656, 'timing/data': 0.03212498687207699, 'timing/model': 1.6437951838597655, 'learning_rate/qwen_vl_interface': 5.382727356393909e-07, 'learning_rate/action_model': 5.382727356393909e-06, 'epoch': 2.59}) +[2026-06-18 08:46:14,131][starVLA.training.train_starvla][INFO] - Step 6722, Loss: {'train/loss': 0.35847726464271545, 'train/grad_norm_pre_clip': 17.921663284301758, 'timing/data': 0.030491854064166546, 'timing/model': 1.4690216220915318, 'learning_rate/qwen_vl_interface': 5.379992365233442e-07, 'learning_rate/action_model': 5.379992365233442e-06, 'epoch': 2.59}) +[2026-06-18 08:46:18,663][starVLA.training.train_starvla][INFO] - Step 6723, Loss: {'train/loss': 0.3223065137863159, 'train/grad_norm_pre_clip': 22.054542541503906, 'timing/data': 0.03795082867145538, 'timing/model': 1.2731714202091098, 'learning_rate/qwen_vl_interface': 5.377267142101261e-07, 'learning_rate/action_model': 5.37726714210126e-06, 'epoch': 2.59}) +[2026-06-18 08:46:23,587][starVLA.training.train_starvla][INFO] - Step 6724, Loss: {'train/loss': 0.4984782338142395, 'train/grad_norm_pre_clip': 15.856413841247559, 'timing/data': 0.031541433185338974, 'timing/model': 1.0965995583683252, 'learning_rate/qwen_vl_interface': 5.374551687562306e-07, 'learning_rate/action_model': 5.374551687562306e-06, 'epoch': 2.59}) +[2026-06-18 08:46:28,512][starVLA.training.train_starvla][INFO] - Step 6725, Loss: {'train/loss': 0.16700759530067444, 'train/grad_norm_pre_clip': 17.09052276611328, 'timing/data': 0.030828834511339664, 'timing/model': 1.1046382710337639, 'learning_rate/qwen_vl_interface': 5.371846002179494e-07, 'learning_rate/action_model': 5.371846002179493e-06, 'epoch': 2.59}) +[2026-06-18 08:46:33,321][starVLA.training.train_starvla][INFO] - Step 6726, Loss: {'train/loss': 0.19887971878051758, 'train/grad_norm_pre_clip': 16.51381492614746, 'timing/data': 0.030644801445305347, 'timing/model': 1.521845955401659, 'learning_rate/qwen_vl_interface': 5.36915008651372e-07, 'learning_rate/action_model': 5.369150086513719e-06, 'epoch': 2.59}) +[2026-06-18 08:46:38,350][starVLA.training.train_starvla][INFO] - Step 6727, Loss: {'train/loss': 0.10635321587324142, 'train/grad_norm_pre_clip': 10.333436965942383, 'timing/data': 0.03330073133111, 'timing/model': 1.7190897762775421, 'learning_rate/qwen_vl_interface': 5.366463941123853e-07, 'learning_rate/action_model': 5.366463941123853e-06, 'epoch': 2.59}) +[2026-06-18 08:46:42,858][starVLA.training.train_starvla][INFO] - Step 6728, Loss: {'train/loss': 0.16780732572078705, 'train/grad_norm_pre_clip': 11.71638011932373, 'timing/data': 0.04232122749090195, 'timing/model': 1.1838683895766735, 'learning_rate/qwen_vl_interface': 5.36378756656672e-07, 'learning_rate/action_model': 5.36378756656672e-06, 'epoch': 2.59}) +[2026-06-18 08:46:47,648][starVLA.training.train_starvla][INFO] - Step 6729, Loss: {'train/loss': 0.107399120926857, 'train/grad_norm_pre_clip': 8.546223640441895, 'timing/data': 0.03695374447852373, 'timing/model': 1.1855701049789786, 'learning_rate/qwen_vl_interface': 5.361120963397151e-07, 'learning_rate/action_model': 5.361120963397152e-06, 'epoch': 2.59}) +[2026-06-18 08:46:53,103][starVLA.training.train_starvla][INFO] - Step 6730, Loss: {'train/loss': 0.05304534360766411, 'train/grad_norm_pre_clip': 5.693912506103516, 'timing/data': 0.05331265460699797, 'timing/model': 1.515874894335866, 'learning_rate/qwen_vl_interface': 5.358464132167927e-07, 'learning_rate/action_model': 5.358464132167927e-06, 'epoch': 2.59}) +[2026-06-18 08:46:58,105][starVLA.training.train_starvla][INFO] - Step 6731, Loss: {'train/loss': 0.20148968696594238, 'train/grad_norm_pre_clip': 19.80886459350586, 'timing/data': 0.03847418911755085, 'timing/model': 1.1480143209919333, 'learning_rate/qwen_vl_interface': 5.355817073429818e-07, 'learning_rate/action_model': 5.355817073429818e-06, 'epoch': 2.59}) +[2026-06-18 08:47:03,005][starVLA.training.train_starvla][INFO] - Step 6732, Loss: {'train/loss': 0.18344493210315704, 'train/grad_norm_pre_clip': 12.930129051208496, 'timing/data': 0.034802865236997604, 'timing/model': 1.5978886177763343, 'learning_rate/qwen_vl_interface': 5.353179787731557e-07, 'learning_rate/action_model': 5.353179787731557e-06, 'epoch': 2.59}) +[2026-06-18 08:47:07,715][starVLA.training.train_starvla][INFO] - Step 6733, Loss: {'train/loss': 0.347976952791214, 'train/grad_norm_pre_clip': 21.1937313079834, 'timing/data': 0.0314703369513154, 'timing/model': 1.1384304873645306, 'learning_rate/qwen_vl_interface': 5.350552275619857e-07, 'learning_rate/action_model': 5.350552275619857e-06, 'epoch': 2.59}) +[2026-06-18 08:47:12,518][starVLA.training.train_starvla][INFO] - Step 6734, Loss: {'train/loss': 0.3409949839115143, 'train/grad_norm_pre_clip': 20.007034301757812, 'timing/data': 0.033370484597980976, 'timing/model': 1.1156014939770103, 'learning_rate/qwen_vl_interface': 5.347934537639405e-07, 'learning_rate/action_model': 5.347934537639405e-06, 'epoch': 2.59}) +[2026-06-18 08:47:17,279][starVLA.training.train_starvla][INFO] - Step 6735, Loss: {'train/loss': 0.2902342677116394, 'train/grad_norm_pre_clip': 15.100875854492188, 'timing/data': 0.036428920924663544, 'timing/model': 1.1276726955547929, 'learning_rate/qwen_vl_interface': 5.345326574332865e-07, 'learning_rate/action_model': 5.345326574332865e-06, 'epoch': 2.59}) +[2026-06-18 08:47:22,228][starVLA.training.train_starvla][INFO] - Step 6736, Loss: {'train/loss': 0.28286412358283997, 'train/grad_norm_pre_clip': 14.874731063842773, 'timing/data': 0.03346074838191271, 'timing/model': 1.1152904881164432, 'learning_rate/qwen_vl_interface': 5.342728386240867e-07, 'learning_rate/action_model': 5.3427283862408665e-06, 'epoch': 2.59}) +[2026-06-18 08:47:26,885][starVLA.training.train_starvla][INFO] - Step 6737, Loss: {'train/loss': 0.4393863379955292, 'train/grad_norm_pre_clip': 20.818174362182617, 'timing/data': 0.047798584215343, 'timing/model': 1.4685971587896347, 'learning_rate/qwen_vl_interface': 5.340139973902013e-07, 'learning_rate/action_model': 5.340139973902013e-06, 'epoch': 2.59}) +[2026-06-18 08:47:31,630][starVLA.training.train_starvla][INFO] - Step 6738, Loss: {'train/loss': 0.1984257847070694, 'train/grad_norm_pre_clip': 13.1719331741333, 'timing/data': 0.05400987435132265, 'timing/model': 1.201419087126851, 'learning_rate/qwen_vl_interface': 5.337561337852892e-07, 'learning_rate/action_model': 5.337561337852892e-06, 'epoch': 2.59}) +[2026-06-18 08:47:36,440][starVLA.training.train_starvla][INFO] - Step 6739, Loss: {'train/loss': 0.28524091839790344, 'train/grad_norm_pre_clip': 12.752758979797363, 'timing/data': 0.031237243674695492, 'timing/model': 1.1255650790408254, 'learning_rate/qwen_vl_interface': 5.334992478628054e-07, 'learning_rate/action_model': 5.334992478628054e-06, 'epoch': 2.59}) +[2026-06-18 08:47:41,453][starVLA.training.train_starvla][INFO] - Step 6740, Loss: {'train/loss': 0.5344706177711487, 'train/grad_norm_pre_clip': 22.235239028930664, 'timing/data': 0.03015175834298134, 'timing/model': 1.1062122946605086, 'learning_rate/qwen_vl_interface': 5.332433396760027e-07, 'learning_rate/action_model': 5.332433396760028e-06, 'epoch': 2.59}) +[2026-06-18 08:47:46,307][starVLA.training.train_starvla][INFO] - Step 6741, Loss: {'train/loss': 0.41677695512771606, 'train/grad_norm_pre_clip': 18.110811233520508, 'timing/data': 0.03448351100087166, 'timing/model': 1.123893879354, 'learning_rate/qwen_vl_interface': 5.329884092779312e-07, 'learning_rate/action_model': 5.329884092779312e-06, 'epoch': 2.59}) +[2026-06-18 08:47:50,732][starVLA.training.train_starvla][INFO] - Step 6742, Loss: {'train/loss': 0.33505216240882874, 'train/grad_norm_pre_clip': 15.41747760772705, 'timing/data': 0.0400601988658309, 'timing/model': 1.0976586202159524, 'learning_rate/qwen_vl_interface': 5.327344567214386e-07, 'learning_rate/action_model': 5.327344567214385e-06, 'epoch': 2.59}) +[2026-06-18 08:47:55,387][starVLA.training.train_starvla][INFO] - Step 6743, Loss: {'train/loss': 0.24201220273971558, 'train/grad_norm_pre_clip': 17.10666275024414, 'timing/data': 0.03251634631305933, 'timing/model': 1.1039327858015895, 'learning_rate/qwen_vl_interface': 5.324814820591683e-07, 'learning_rate/action_model': 5.324814820591682e-06, 'epoch': 2.59}) +[2026-06-18 08:48:00,125][starVLA.training.train_starvla][INFO] - Step 6744, Loss: {'train/loss': 0.3725476861000061, 'train/grad_norm_pre_clip': 21.217660903930664, 'timing/data': 0.03097113035619259, 'timing/model': 1.0694211702793837, 'learning_rate/qwen_vl_interface': 5.322294853435631e-07, 'learning_rate/action_model': 5.322294853435631e-06, 'epoch': 2.59}) +[2026-06-18 08:48:05,121][starVLA.training.train_starvla][INFO] - Step 6745, Loss: {'train/loss': 0.442551851272583, 'train/grad_norm_pre_clip': 19.535472869873047, 'timing/data': 0.03178182151168585, 'timing/model': 1.2206932893022895, 'learning_rate/qwen_vl_interface': 5.319784666268626e-07, 'learning_rate/action_model': 5.319784666268626e-06, 'epoch': 2.59}) +[2026-06-18 08:48:09,805][starVLA.training.train_starvla][INFO] - Step 6746, Loss: {'train/loss': 0.4705653488636017, 'train/grad_norm_pre_clip': 24.27707862854004, 'timing/data': 0.03814005944877863, 'timing/model': 1.1124398577958345, 'learning_rate/qwen_vl_interface': 5.317284259611024e-07, 'learning_rate/action_model': 5.317284259611024e-06, 'epoch': 2.6}) +[2026-06-18 08:48:14,501][starVLA.training.train_starvla][INFO] - Step 6747, Loss: {'train/loss': 0.4183560013771057, 'train/grad_norm_pre_clip': 15.453929901123047, 'timing/data': 0.03594476543366909, 'timing/model': 1.4172234693542123, 'learning_rate/qwen_vl_interface': 5.314793633981169e-07, 'learning_rate/action_model': 5.314793633981169e-06, 'epoch': 2.6}) +[2026-06-18 08:48:18,927][starVLA.training.train_starvla][INFO] - Step 6748, Loss: {'train/loss': 0.2924290895462036, 'train/grad_norm_pre_clip': 15.427719116210938, 'timing/data': 0.03326855227351189, 'timing/model': 1.1253240257501602, 'learning_rate/qwen_vl_interface': 5.312312789895369e-07, 'learning_rate/action_model': 5.3123127898953685e-06, 'epoch': 2.6}) +[2026-06-18 08:48:23,680][starVLA.training.train_starvla][INFO] - Step 6749, Loss: {'train/loss': 0.42676272988319397, 'train/grad_norm_pre_clip': 22.784542083740234, 'timing/data': 0.04050033539533615, 'timing/model': 1.1979547664523125, 'learning_rate/qwen_vl_interface': 5.3098417278679e-07, 'learning_rate/action_model': 5.3098417278679e-06, 'epoch': 2.6}) +[2026-06-18 08:48:28,501][starVLA.training.train_starvla][INFO] - Step 6750, Loss: {'train/loss': 0.32665178179740906, 'train/grad_norm_pre_clip': 25.73921012878418, 'timing/data': 0.03492306359112263, 'timing/model': 1.147832253947854, 'learning_rate/qwen_vl_interface': 5.307380448411022e-07, 'learning_rate/action_model': 5.307380448411022e-06, 'epoch': 2.6}) +[2026-06-18 08:48:33,200][starVLA.training.train_starvla][INFO] - Step 6751, Loss: {'train/loss': 0.3184470236301422, 'train/grad_norm_pre_clip': 16.620685577392578, 'timing/data': 0.03473576158285141, 'timing/model': 1.1371722612529993, 'learning_rate/qwen_vl_interface': 5.304928952034955e-07, 'learning_rate/action_model': 5.304928952034956e-06, 'epoch': 2.6}) +[2026-06-18 08:48:37,789][starVLA.training.train_starvla][INFO] - Step 6752, Loss: {'train/loss': 0.27756258845329285, 'train/grad_norm_pre_clip': 82.12630462646484, 'timing/data': 0.032266635447740555, 'timing/model': 1.1890480853617191, 'learning_rate/qwen_vl_interface': 5.302487239247905e-07, 'learning_rate/action_model': 5.302487239247904e-06, 'epoch': 2.6}) +[2026-06-18 08:48:42,534][starVLA.training.train_starvla][INFO] - Step 6753, Loss: {'train/loss': 0.3599393963813782, 'train/grad_norm_pre_clip': 20.497528076171875, 'timing/data': 0.03245903830975294, 'timing/model': 1.085104745812714, 'learning_rate/qwen_vl_interface': 5.300055310556034e-07, 'learning_rate/action_model': 5.300055310556034e-06, 'epoch': 2.6}) +[2026-06-18 08:48:47,475][starVLA.training.train_starvla][INFO] - Step 6754, Loss: {'train/loss': 0.13890039920806885, 'train/grad_norm_pre_clip': 17.51034927368164, 'timing/data': 0.0343023557215929, 'timing/model': 1.1197148067876697, 'learning_rate/qwen_vl_interface': 5.297633166463482e-07, 'learning_rate/action_model': 5.297633166463481e-06, 'epoch': 2.6}) +[2026-06-18 08:48:52,387][starVLA.training.train_starvla][INFO] - Step 6755, Loss: {'train/loss': 0.5109567046165466, 'train/grad_norm_pre_clip': 25.719449996948242, 'timing/data': 0.05425981618463993, 'timing/model': 1.1451665163040161, 'learning_rate/qwen_vl_interface': 5.295220807472373e-07, 'learning_rate/action_model': 5.295220807472374e-06, 'epoch': 2.6}) +[2026-06-18 08:48:57,478][starVLA.training.train_starvla][INFO] - Step 6756, Loss: {'train/loss': 0.5774737596511841, 'train/grad_norm_pre_clip': 23.123817443847656, 'timing/data': 0.04436888359487057, 'timing/model': 1.5374354347586632, 'learning_rate/qwen_vl_interface': 5.292818234082786e-07, 'learning_rate/action_model': 5.292818234082787e-06, 'epoch': 2.6}) +[2026-06-18 08:49:02,267][starVLA.training.train_starvla][INFO] - Step 6757, Loss: {'train/loss': 0.07536539435386658, 'train/grad_norm_pre_clip': 7.023900032043457, 'timing/data': 0.032691788859665394, 'timing/model': 1.483506322838366, 'learning_rate/qwen_vl_interface': 5.290425446792772e-07, 'learning_rate/action_model': 5.2904254467927716e-06, 'epoch': 2.6}) +[2026-06-18 08:49:07,026][starVLA.training.train_starvla][INFO] - Step 6758, Loss: {'train/loss': 0.167123943567276, 'train/grad_norm_pre_clip': 9.955707550048828, 'timing/data': 0.042970698326826096, 'timing/model': 1.2754903556779027, 'learning_rate/qwen_vl_interface': 5.288042446098364e-07, 'learning_rate/action_model': 5.2880424460983634e-06, 'epoch': 2.6}) +[2026-06-18 08:49:11,814][starVLA.training.train_starvla][INFO] - Step 6759, Loss: {'train/loss': 0.315586119890213, 'train/grad_norm_pre_clip': 13.05029296875, 'timing/data': 0.03881641756743193, 'timing/model': 1.1071904813870788, 'learning_rate/qwen_vl_interface': 5.28566923249356e-07, 'learning_rate/action_model': 5.28566923249356e-06, 'epoch': 2.6}) +[2026-06-18 08:49:16,682][starVLA.training.train_starvla][INFO] - Step 6760, Loss: {'train/loss': 0.2263079434633255, 'train/grad_norm_pre_clip': 12.68132495880127, 'timing/data': 0.03391159325838089, 'timing/model': 1.1604578187689185, 'learning_rate/qwen_vl_interface': 5.283305806470329e-07, 'learning_rate/action_model': 5.2833058064703286e-06, 'epoch': 2.6}) +[2026-06-18 08:49:21,615][starVLA.training.train_starvla][INFO] - Step 6761, Loss: {'train/loss': 0.2110845446586609, 'train/grad_norm_pre_clip': 13.372435569763184, 'timing/data': 0.04101836308836937, 'timing/model': 1.6779373763129115, 'learning_rate/qwen_vl_interface': 5.280952168518612e-07, 'learning_rate/action_model': 5.280952168518612e-06, 'epoch': 2.6}) +[2026-06-18 08:49:26,576][starVLA.training.train_starvla][INFO] - Step 6762, Loss: {'train/loss': 0.32142552733421326, 'train/grad_norm_pre_clip': 15.648844718933105, 'timing/data': 0.034498696215450764, 'timing/model': 1.7389643024653196, 'learning_rate/qwen_vl_interface': 5.278608319126318e-07, 'learning_rate/action_model': 5.278608319126318e-06, 'epoch': 2.6}) +[2026-06-18 08:49:31,091][starVLA.training.train_starvla][INFO] - Step 6763, Loss: {'train/loss': 0.42378830909729004, 'train/grad_norm_pre_clip': 19.342687606811523, 'timing/data': 0.04559220466762781, 'timing/model': 1.1657417295500636, 'learning_rate/qwen_vl_interface': 5.276274258779332e-07, 'learning_rate/action_model': 5.2762742587793315e-06, 'epoch': 2.6}) +[2026-06-18 08:49:35,917][starVLA.training.train_starvla][INFO] - Step 6764, Loss: {'train/loss': 0.24538783729076385, 'train/grad_norm_pre_clip': 13.729199409484863, 'timing/data': 0.033957662992179394, 'timing/model': 1.1596448542550206, 'learning_rate/qwen_vl_interface': 5.273949987961508e-07, 'learning_rate/action_model': 5.273949987961508e-06, 'epoch': 2.6}) +[2026-06-18 08:49:40,670][starVLA.training.train_starvla][INFO] - Step 6765, Loss: {'train/loss': 0.39744260907173157, 'train/grad_norm_pre_clip': 17.428447723388672, 'timing/data': 0.046633546240627766, 'timing/model': 1.0778319453820586, 'learning_rate/qwen_vl_interface': 5.271635507154671e-07, 'learning_rate/action_model': 5.2716355071546706e-06, 'epoch': 2.6}) +[2026-06-18 08:49:45,189][starVLA.training.train_starvla][INFO] - Step 6766, Loss: {'train/loss': 0.1484888195991516, 'train/grad_norm_pre_clip': 13.782475471496582, 'timing/data': 0.03669339790940285, 'timing/model': 1.134997713379562, 'learning_rate/qwen_vl_interface': 5.269330816838614e-07, 'learning_rate/action_model': 5.269330816838613e-06, 'epoch': 2.6}) +[2026-06-18 08:49:49,803][starVLA.training.train_starvla][INFO] - Step 6767, Loss: {'train/loss': 0.12689734995365143, 'train/grad_norm_pre_clip': 14.98163890838623, 'timing/data': 0.03006318211555481, 'timing/model': 1.430870084092021, 'learning_rate/qwen_vl_interface': 5.267035917491095e-07, 'learning_rate/action_model': 5.267035917491094e-06, 'epoch': 2.6}) +[2026-06-18 08:49:54,168][starVLA.training.train_starvla][INFO] - Step 6768, Loss: {'train/loss': 0.21059265732765198, 'train/grad_norm_pre_clip': 15.569150924682617, 'timing/data': 0.030134650878608227, 'timing/model': 1.0760222030803561, 'learning_rate/qwen_vl_interface': 5.264750809587862e-07, 'learning_rate/action_model': 5.264750809587863e-06, 'epoch': 2.6}) +[2026-06-18 08:49:58,811][starVLA.training.train_starvla][INFO] - Step 6769, Loss: {'train/loss': 0.22587856650352478, 'train/grad_norm_pre_clip': 11.89535140991211, 'timing/data': 0.030445444397628307, 'timing/model': 1.072200944647193, 'learning_rate/qwen_vl_interface': 5.262475493602609e-07, 'learning_rate/action_model': 5.26247549360261e-06, 'epoch': 2.6}) +[2026-06-18 08:50:03,633][starVLA.training.train_starvla][INFO] - Step 6770, Loss: {'train/loss': 0.14209449291229248, 'train/grad_norm_pre_clip': 11.178210258483887, 'timing/data': 0.036047366447746754, 'timing/model': 1.145325968042016, 'learning_rate/qwen_vl_interface': 5.260209970007018e-07, 'learning_rate/action_model': 5.260209970007018e-06, 'epoch': 2.6}) +[2026-06-18 08:50:08,243][starVLA.training.train_starvla][INFO] - Step 6771, Loss: {'train/loss': 0.4559116065502167, 'train/grad_norm_pre_clip': 30.974580764770508, 'timing/data': 0.031726179644465446, 'timing/model': 1.4943094868212938, 'learning_rate/qwen_vl_interface': 5.257954239270732e-07, 'learning_rate/action_model': 5.257954239270732e-06, 'epoch': 2.6}) +[2026-06-18 08:50:12,924][starVLA.training.train_starvla][INFO] - Step 6772, Loss: {'train/loss': 0.11633351445198059, 'train/grad_norm_pre_clip': 10.689310073852539, 'timing/data': 0.03158401884138584, 'timing/model': 1.4117601960897446, 'learning_rate/qwen_vl_interface': 5.255708301861367e-07, 'learning_rate/action_model': 5.255708301861367e-06, 'epoch': 2.61}) +[2026-06-18 08:50:17,309][starVLA.training.train_starvla][INFO] - Step 6773, Loss: {'train/loss': 0.17221547663211823, 'train/grad_norm_pre_clip': 9.836334228515625, 'timing/data': 0.052782042883336544, 'timing/model': 1.1697745565325022, 'learning_rate/qwen_vl_interface': 5.253472158244506e-07, 'learning_rate/action_model': 5.253472158244505e-06, 'epoch': 2.61}) +[2026-06-18 08:50:21,897][starVLA.training.train_starvla][INFO] - Step 6774, Loss: {'train/loss': 0.2349918782711029, 'train/grad_norm_pre_clip': 23.261371612548828, 'timing/data': 0.03179693873971701, 'timing/model': 1.0810441635549068, 'learning_rate/qwen_vl_interface': 5.251245808883707e-07, 'learning_rate/action_model': 5.251245808883707e-06, 'epoch': 2.61}) +[2026-06-18 08:50:26,624][starVLA.training.train_starvla][INFO] - Step 6775, Loss: {'train/loss': 0.19168458878993988, 'train/grad_norm_pre_clip': 14.480792045593262, 'timing/data': 0.03683078847825527, 'timing/model': 1.09138947352767, 'learning_rate/qwen_vl_interface': 5.249029254240495e-07, 'learning_rate/action_model': 5.249029254240495e-06, 'epoch': 2.61}) +[2026-06-18 08:50:31,156][starVLA.training.train_starvla][INFO] - Step 6776, Loss: {'train/loss': 0.37762486934661865, 'train/grad_norm_pre_clip': 17.795347213745117, 'timing/data': 0.03384892176836729, 'timing/model': 1.0818808684125543, 'learning_rate/qwen_vl_interface': 5.246822494774364e-07, 'learning_rate/action_model': 5.246822494774363e-06, 'epoch': 2.61}) +[2026-06-18 08:50:35,833][starVLA.training.train_starvla][INFO] - Step 6777, Loss: {'train/loss': 0.12823405861854553, 'train/grad_norm_pre_clip': 12.89990234375, 'timing/data': 0.03219319228082895, 'timing/model': 1.4826716212555766, 'learning_rate/qwen_vl_interface': 5.244625530942774e-07, 'learning_rate/action_model': 5.244625530942774e-06, 'epoch': 2.61}) +[2026-06-18 08:50:40,099][starVLA.training.train_starvla][INFO] - Step 6778, Loss: {'train/loss': 0.3489430546760559, 'train/grad_norm_pre_clip': 14.0702486038208, 'timing/data': 0.03016556054353714, 'timing/model': 1.0837097764015198, 'learning_rate/qwen_vl_interface': 5.242438363201161e-07, 'learning_rate/action_model': 5.2424383632011605e-06, 'epoch': 2.61}) +[2026-06-18 08:50:44,904][starVLA.training.train_starvla][INFO] - Step 6779, Loss: {'train/loss': 0.44393450021743774, 'train/grad_norm_pre_clip': 16.01737403869629, 'timing/data': 0.0473073972389102, 'timing/model': 1.1195101737976074, 'learning_rate/qwen_vl_interface': 5.240260992002928e-07, 'learning_rate/action_model': 5.240260992002928e-06, 'epoch': 2.61}) +[2026-06-18 08:50:49,421][starVLA.training.train_starvla][INFO] - Step 6780, Loss: {'train/loss': 0.12983198463916779, 'train/grad_norm_pre_clip': 10.703092575073242, 'timing/data': 0.029626320116221905, 'timing/model': 1.0652662329375744, 'learning_rate/qwen_vl_interface': 5.238093417799446e-07, 'learning_rate/action_model': 5.238093417799446e-06, 'epoch': 2.61}) +[2026-06-18 08:50:53,919][starVLA.training.train_starvla][INFO] - Step 6781, Loss: {'train/loss': 0.11026404798030853, 'train/grad_norm_pre_clip': 8.61257266998291, 'timing/data': 0.030453004874289036, 'timing/model': 1.3960533495992422, 'learning_rate/qwen_vl_interface': 5.235935641040048e-07, 'learning_rate/action_model': 5.235935641040048e-06, 'epoch': 2.61}) +[2026-06-18 08:50:58,638][starVLA.training.train_starvla][INFO] - Step 6782, Loss: {'train/loss': 0.17030680179595947, 'train/grad_norm_pre_clip': 14.580436706542969, 'timing/data': 0.03420138172805309, 'timing/model': 1.5359589168801904, 'learning_rate/qwen_vl_interface': 5.233787662172059e-07, 'learning_rate/action_model': 5.23378766217206e-06, 'epoch': 2.61}) +[2026-06-18 08:51:02,934][starVLA.training.train_starvla][INFO] - Step 6783, Loss: {'train/loss': 0.33733823895454407, 'train/grad_norm_pre_clip': 16.02467918395996, 'timing/data': 0.030740242451429367, 'timing/model': 1.0702725751325488, 'learning_rate/qwen_vl_interface': 5.231649481640741e-07, 'learning_rate/action_model': 5.231649481640741e-06, 'epoch': 2.61}) +[2026-06-18 08:51:07,543][starVLA.training.train_starvla][INFO] - Step 6784, Loss: {'train/loss': 0.17661412060260773, 'train/grad_norm_pre_clip': 13.02385425567627, 'timing/data': 0.029659089632332325, 'timing/model': 1.1279181893914938, 'learning_rate/qwen_vl_interface': 5.229521099889351e-07, 'learning_rate/action_model': 5.229521099889352e-06, 'epoch': 2.61}) +[2026-06-18 08:51:12,142][starVLA.training.train_starvla][INFO] - Step 6785, Loss: {'train/loss': 0.31862348318099976, 'train/grad_norm_pre_clip': 14.820161819458008, 'timing/data': 0.037163713946938515, 'timing/model': 1.0893224086612463, 'learning_rate/qwen_vl_interface': 5.227402517359105e-07, 'learning_rate/action_model': 5.2274025173591045e-06, 'epoch': 2.61}) +[2026-06-18 08:51:16,869][starVLA.training.train_starvla][INFO] - Step 6786, Loss: {'train/loss': 0.25666093826293945, 'train/grad_norm_pre_clip': 19.103574752807617, 'timing/data': 0.031667293049395084, 'timing/model': 1.1173204686492682, 'learning_rate/qwen_vl_interface': 5.225293734489182e-07, 'learning_rate/action_model': 5.225293734489181e-06, 'epoch': 2.61}) +[2026-06-18 08:51:21,540][starVLA.training.train_starvla][INFO] - Step 6787, Loss: {'train/loss': 0.28414392471313477, 'train/grad_norm_pre_clip': 16.604095458984375, 'timing/data': 0.03288611024618149, 'timing/model': 1.4599224273115396, 'learning_rate/qwen_vl_interface': 5.223194751716733e-07, 'learning_rate/action_model': 5.223194751716733e-06, 'epoch': 2.61}) +[2026-06-18 08:51:25,861][starVLA.training.train_starvla][INFO] - Step 6788, Loss: {'train/loss': 0.2887631058692932, 'train/grad_norm_pre_clip': 20.77376937866211, 'timing/data': 0.03206505533307791, 'timing/model': 1.099241964519024, 'learning_rate/qwen_vl_interface': 5.221105569476893e-07, 'learning_rate/action_model': 5.221105569476892e-06, 'epoch': 2.61}) +[2026-06-18 08:51:30,715][starVLA.training.train_starvla][INFO] - Step 6789, Loss: {'train/loss': 0.24498192965984344, 'train/grad_norm_pre_clip': 14.728273391723633, 'timing/data': 0.030666731297969818, 'timing/model': 1.0888710021972656, 'learning_rate/qwen_vl_interface': 5.21902618820274e-07, 'learning_rate/action_model': 5.21902618820274e-06, 'epoch': 2.61}) +[2026-06-18 08:51:35,463][starVLA.training.train_starvla][INFO] - Step 6790, Loss: {'train/loss': 0.12001999467611313, 'train/grad_norm_pre_clip': 8.644877433776855, 'timing/data': 0.033475833013653755, 'timing/model': 1.1060249842703342, 'learning_rate/qwen_vl_interface': 5.216956608325331e-07, 'learning_rate/action_model': 5.2169566083253304e-06, 'epoch': 2.61}) +[2026-06-18 08:51:40,243][starVLA.training.train_starvla][INFO] - Step 6791, Loss: {'train/loss': 0.2983207702636719, 'train/grad_norm_pre_clip': 15.899669647216797, 'timing/data': 0.033850930631160736, 'timing/model': 1.6153643932193518, 'learning_rate/qwen_vl_interface': 5.2148968302737e-07, 'learning_rate/action_model': 5.2148968302737005e-06, 'epoch': 2.61}) +[2026-06-18 08:51:44,943][starVLA.training.train_starvla][INFO] - Step 6792, Loss: {'train/loss': 0.277486115694046, 'train/grad_norm_pre_clip': 24.02628517150879, 'timing/data': 0.03686212282627821, 'timing/model': 1.4940626909956336, 'learning_rate/qwen_vl_interface': 5.212846854474834e-07, 'learning_rate/action_model': 5.212846854474833e-06, 'epoch': 2.61}) +[2026-06-18 08:51:49,164][starVLA.training.train_starvla][INFO] - Step 6793, Loss: {'train/loss': 0.23850493133068085, 'train/grad_norm_pre_clip': 11.676859855651855, 'timing/data': 0.03305622097104788, 'timing/model': 1.0638494417071342, 'learning_rate/qwen_vl_interface': 5.2108066813537e-07, 'learning_rate/action_model': 5.2108066813537e-06, 'epoch': 2.61}) +[2026-06-18 08:51:53,730][starVLA.training.train_starvla][INFO] - Step 6794, Loss: {'train/loss': 0.5315850377082825, 'train/grad_norm_pre_clip': 24.302906036376953, 'timing/data': 0.029351751320064068, 'timing/model': 1.0966710094362497, 'learning_rate/qwen_vl_interface': 5.208776311333223e-07, 'learning_rate/action_model': 5.208776311333223e-06, 'epoch': 2.61}) +[2026-06-18 08:51:58,387][starVLA.training.train_starvla][INFO] - Step 6795, Loss: {'train/loss': 0.228797048330307, 'train/grad_norm_pre_clip': 18.289295196533203, 'timing/data': 0.03413514234125614, 'timing/model': 1.120756447315216, 'learning_rate/qwen_vl_interface': 5.206755744834312e-07, 'learning_rate/action_model': 5.2067557448343125e-06, 'epoch': 2.61}) +[2026-06-18 08:52:03,152][starVLA.training.train_starvla][INFO] - Step 6796, Loss: {'train/loss': 0.3818851709365845, 'train/grad_norm_pre_clip': 22.190832138061523, 'timing/data': 0.03141282685101032, 'timing/model': 1.1099450513720512, 'learning_rate/qwen_vl_interface': 5.20474498227582e-07, 'learning_rate/action_model': 5.2047449822758196e-06, 'epoch': 2.61}) +[2026-06-18 08:52:07,888][starVLA.training.train_starvla][INFO] - Step 6797, Loss: {'train/loss': 0.19127309322357178, 'train/grad_norm_pre_clip': 12.079452514648438, 'timing/data': 0.03165341168642044, 'timing/model': 1.6321757500991225, 'learning_rate/qwen_vl_interface': 5.202744024074583e-07, 'learning_rate/action_model': 5.202744024074584e-06, 'epoch': 2.61}) +[2026-06-18 08:52:12,152][starVLA.training.train_starvla][INFO] - Step 6798, Loss: {'train/loss': 0.29402413964271545, 'train/grad_norm_pre_clip': 18.157550811767578, 'timing/data': 0.03400911018252373, 'timing/model': 1.1001023957505822, 'learning_rate/qwen_vl_interface': 5.200752870645407e-07, 'learning_rate/action_model': 5.200752870645406e-06, 'epoch': 2.62}) +[2026-06-18 08:52:16,961][starVLA.training.train_starvla][INFO] - Step 6799, Loss: {'train/loss': 0.35689011216163635, 'train/grad_norm_pre_clip': 16.009422302246094, 'timing/data': 0.04237609263509512, 'timing/model': 1.0932330703362823, 'learning_rate/qwen_vl_interface': 5.198771522401054e-07, 'learning_rate/action_model': 5.198771522401054e-06, 'epoch': 2.62}) +[2026-06-18 08:54:37,681][starVLA.training.train_starvla][INFO] - Step 6800, Loss: {'train/loss': 0.2522401511669159, 'train/grad_norm_pre_clip': 14.369672775268555, 'eval/loss': 0.4286957085132599, 'eval/latency_4/loss': 0.4286957085132599, 'eval/demon_fire_precision': 0.9670610809082187, 'eval/demon_fire_recall': 0.9821370574861968, 'eval/demon_fire_f1': 0.9745407669996778, 'eval/demon_fire_rate_model': 0.9183553597650514, 'eval/demon_fire_rate_teacher': 0.9042584434654919, 'eval/demon_left_precision': 0.8100263852242744, 'eval/demon_left_recall': 0.7009132420091324, 'eval/demon_left_f1': 0.7515299877600979, 'eval/demon_left_rate_model': 0.11130690161527165, 'eval/demon_left_rate_teacher': 0.12863436123348018, 'eval/demon_right_precision': 0.8018518518518518, 'eval/demon_right_recall': 0.7815884476534296, 'eval/demon_right_f1': 0.7915904936014624, 'eval/demon_right_rate_model': 0.15859030837004406, 'eval/demon_right_rate_teacher': 0.16270190895741557, 'eval/demon_move_f1': 0.7715602406807802, 'eval/demon_cc_f1': 0.9032529710039731, 'eval/latency_4/demon_fire_precision': 0.9670610809082187, 'eval/latency_4/demon_fire_recall': 0.9821370574861968, 'eval/latency_4/demon_fire_f1': 0.9745407669996778, 'eval/latency_4/demon_fire_rate_model': 0.9183553597650514, 'eval/latency_4/demon_fire_rate_teacher': 0.9042584434654919, 'eval/latency_4/demon_left_precision': 0.8100263852242744, 'eval/latency_4/demon_left_recall': 0.7009132420091324, 'eval/latency_4/demon_left_f1': 0.7515299877600979, 'eval/latency_4/demon_left_rate_model': 0.11130690161527165, 'eval/latency_4/demon_left_rate_teacher': 0.12863436123348018, 'eval/latency_4/demon_right_precision': 0.8018518518518518, 'eval/latency_4/demon_right_recall': 0.7815884476534296, 'eval/latency_4/demon_right_f1': 0.7915904936014624, 'eval/latency_4/demon_right_rate_model': 0.15859030837004406, 'eval/latency_4/demon_right_rate_teacher': 0.16270190895741557, 'eval/latency_4/demon_move_f1': 0.7715602406807802, 'eval/latency_4/demon_cc_f1': 0.9032529710039731, 'timing/data': 0.033898682333528996, 'timing/model': 1.106471411883831, 'learning_rate/qwen_vl_interface': 5.196799979752265e-07, 'learning_rate/action_model': 5.196799979752264e-06, 'epoch': 2.62}) +[2026-06-18 08:54:37,728][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_6800_state +[2026-06-18 08:54:37,738][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 08:54:46,596][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_6800_state/model.safetensors +[2026-06-18 08:55:08,719][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_6800_state/optimizer.bin +[2026-06-18 08:55:08,720][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_6800_state/sampler.bin +[2026-06-18 08:55:08,721][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_6800_state/sampler_1.bin +[2026-06-18 08:55:08,724][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_6800_state/random_states_0.pkl +[2026-06-18 08:55:08,732][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 6800 +[2026-06-18 08:55:08,733][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 08:55:08,753][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 08:55:13,543][starVLA.training.train_starvla][INFO] - Step 6801, Loss: {'train/loss': 0.376236230134964, 'train/grad_norm_pre_clip': 17.72385025024414, 'timing/data': 0.0333477733656764, 'timing/model': 1.140835927799344, 'learning_rate/qwen_vl_interface': 5.194838243107742e-07, 'learning_rate/action_model': 5.194838243107742e-06, 'epoch': 2.62}) +[2026-06-18 08:55:17,878][starVLA.training.train_starvla][INFO] - Step 6802, Loss: {'train/loss': 0.19176097214221954, 'train/grad_norm_pre_clip': 16.985822677612305, 'timing/data': 0.032811809331178665, 'timing/model': 1.129027085378766, 'learning_rate/qwen_vl_interface': 5.192886312874148e-07, 'learning_rate/action_model': 5.1928863128741475e-06, 'epoch': 2.62}) +[2026-06-18 08:55:22,624][starVLA.training.train_starvla][INFO] - Step 6803, Loss: {'train/loss': 0.31274983286857605, 'train/grad_norm_pre_clip': 15.466212272644043, 'timing/data': 0.030459188856184483, 'timing/model': 1.0732828481122851, 'learning_rate/qwen_vl_interface': 5.190944189456133e-07, 'learning_rate/action_model': 5.190944189456132e-06, 'epoch': 2.62}) +[2026-06-18 08:55:27,188][starVLA.training.train_starvla][INFO] - Step 6804, Loss: {'train/loss': 0.20178484916687012, 'train/grad_norm_pre_clip': 14.552515983581543, 'timing/data': 0.050149792805314064, 'timing/model': 1.078606623224914, 'learning_rate/qwen_vl_interface': 5.189011873256295e-07, 'learning_rate/action_model': 5.189011873256295e-06, 'epoch': 2.62}) +[2026-06-18 08:55:31,980][starVLA.training.train_starvla][INFO] - Step 6805, Loss: {'train/loss': 0.33003613352775574, 'train/grad_norm_pre_clip': 20.4781494140625, 'timing/data': 0.03538601007312536, 'timing/model': 1.171217399649322, 'learning_rate/qwen_vl_interface': 5.187089364675195e-07, 'learning_rate/action_model': 5.187089364675194e-06, 'epoch': 2.62}) +[2026-06-18 08:55:36,737][starVLA.training.train_starvla][INFO] - Step 6806, Loss: {'train/loss': 0.326673299074173, 'train/grad_norm_pre_clip': 22.52841567993164, 'timing/data': 0.032565788365900517, 'timing/model': 1.100170262157917, 'learning_rate/qwen_vl_interface': 5.18517666411139e-07, 'learning_rate/action_model': 5.18517666411139e-06, 'epoch': 2.62}) +[2026-06-18 08:55:41,281][starVLA.training.train_starvla][INFO] - Step 6807, Loss: {'train/loss': 0.19070561230182648, 'train/grad_norm_pre_clip': 9.06789493560791, 'timing/data': 0.03469339944422245, 'timing/model': 1.375837161205709, 'learning_rate/qwen_vl_interface': 5.18327377196137e-07, 'learning_rate/action_model': 5.18327377196137e-06, 'epoch': 2.62}) +[2026-06-18 08:55:45,719][starVLA.training.train_starvla][INFO] - Step 6808, Loss: {'train/loss': 0.34628376364707947, 'train/grad_norm_pre_clip': 12.974523544311523, 'timing/data': 0.02850315533578396, 'timing/model': 1.0958098666742444, 'learning_rate/qwen_vl_interface': 5.181380688619617e-07, 'learning_rate/action_model': 5.181380688619617e-06, 'epoch': 2.62}) +[2026-06-18 08:55:50,301][starVLA.training.train_starvla][INFO] - Step 6809, Loss: {'train/loss': 0.2968645691871643, 'train/grad_norm_pre_clip': 16.00605010986328, 'timing/data': 0.030357984825968742, 'timing/model': 1.0884013902395964, 'learning_rate/qwen_vl_interface': 5.179497414478561e-07, 'learning_rate/action_model': 5.179497414478562e-06, 'epoch': 2.62}) +[2026-06-18 08:55:54,832][starVLA.training.train_starvla][INFO] - Step 6810, Loss: {'train/loss': 0.5243409276008606, 'train/grad_norm_pre_clip': 22.992002487182617, 'timing/data': 0.029436144977808, 'timing/model': 1.078695842064917, 'learning_rate/qwen_vl_interface': 5.177623949928608e-07, 'learning_rate/action_model': 5.177623949928608e-06, 'epoch': 2.62}) +[2026-06-18 08:55:59,344][starVLA.training.train_starvla][INFO] - Step 6811, Loss: {'train/loss': 0.36499783396720886, 'train/grad_norm_pre_clip': 14.71449089050293, 'timing/data': 0.03262847848236561, 'timing/model': 1.0980701437219977, 'learning_rate/qwen_vl_interface': 5.175760295358137e-07, 'learning_rate/action_model': 5.175760295358136e-06, 'epoch': 2.62}) +[2026-06-18 08:56:03,618][starVLA.training.train_starvla][INFO] - Step 6812, Loss: {'train/loss': 0.3863184452056885, 'train/grad_norm_pre_clip': 19.800634384155273, 'timing/data': 0.031044160947203636, 'timing/model': 1.0716864308342338, 'learning_rate/qwen_vl_interface': 5.173906451153477e-07, 'learning_rate/action_model': 5.173906451153477e-06, 'epoch': 2.62}) +[2026-06-18 08:56:08,357][starVLA.training.train_starvla][INFO] - Step 6813, Loss: {'train/loss': 0.19566695392131805, 'train/grad_norm_pre_clip': 14.237504005432129, 'timing/data': 0.03083636984229088, 'timing/model': 1.075061940588057, 'learning_rate/qwen_vl_interface': 5.172062417698931e-07, 'learning_rate/action_model': 5.1720624176989304e-06, 'epoch': 2.62}) +[2026-06-18 08:56:13,084][starVLA.training.train_starvla][INFO] - Step 6814, Loss: {'train/loss': 0.30132800340652466, 'train/grad_norm_pre_clip': 16.99558448791504, 'timing/data': 0.03131980635225773, 'timing/model': 1.1525170719251037, 'learning_rate/qwen_vl_interface': 5.170228195376778e-07, 'learning_rate/action_model': 5.170228195376778e-06, 'epoch': 2.62}) +[2026-06-18 08:56:17,796][starVLA.training.train_starvla][INFO] - Step 6815, Loss: {'train/loss': 0.3167513608932495, 'train/grad_norm_pre_clip': 16.643064498901367, 'timing/data': 0.0370982913300395, 'timing/model': 1.123377215117216, 'learning_rate/qwen_vl_interface': 5.168403784567247e-07, 'learning_rate/action_model': 5.1684037845672475e-06, 'epoch': 2.62}) +[2026-06-18 08:56:22,826][starVLA.training.train_starvla][INFO] - Step 6816, Loss: {'train/loss': 0.23540568351745605, 'train/grad_norm_pre_clip': 14.993810653686523, 'timing/data': 0.042320006527006626, 'timing/model': 1.6092044962570071, 'learning_rate/qwen_vl_interface': 5.166589185648543e-07, 'learning_rate/action_model': 5.1665891856485425e-06, 'epoch': 2.62}) +[2026-06-18 08:56:27,503][starVLA.training.train_starvla][INFO] - Step 6817, Loss: {'train/loss': 0.26023462414741516, 'train/grad_norm_pre_clip': 17.65188217163086, 'timing/data': 0.03348791226744652, 'timing/model': 1.3853771230205894, 'learning_rate/qwen_vl_interface': 5.164784398996836e-07, 'learning_rate/action_model': 5.164784398996836e-06, 'epoch': 2.62}) +[2026-06-18 08:56:31,967][starVLA.training.train_starvla][INFO] - Step 6818, Loss: {'train/loss': 0.21586905419826508, 'train/grad_norm_pre_clip': 14.334315299987793, 'timing/data': 0.033279492519795895, 'timing/model': 1.1573626631870866, 'learning_rate/qwen_vl_interface': 5.162989424986253e-07, 'learning_rate/action_model': 5.162989424986253e-06, 'epoch': 2.62}) +[2026-06-18 08:56:36,748][starVLA.training.train_starvla][INFO] - Step 6819, Loss: {'train/loss': 0.2424074113368988, 'train/grad_norm_pre_clip': 16.765880584716797, 'timing/data': 0.03262229450047016, 'timing/model': 1.0767518542706966, 'learning_rate/qwen_vl_interface': 5.161204263988904e-07, 'learning_rate/action_model': 5.161204263988904e-06, 'epoch': 2.62}) +[2026-06-18 08:56:41,411][starVLA.training.train_starvla][INFO] - Step 6820, Loss: {'train/loss': 0.33229732513427734, 'train/grad_norm_pre_clip': 12.891498565673828, 'timing/data': 0.03332164790481329, 'timing/model': 1.091351905837655, 'learning_rate/qwen_vl_interface': 5.159428916374845e-07, 'learning_rate/action_model': 5.159428916374845e-06, 'epoch': 2.62}) +[2026-06-18 08:56:46,016][starVLA.training.train_starvla][INFO] - Step 6821, Loss: {'train/loss': 0.32319802045822144, 'train/grad_norm_pre_clip': 14.118783950805664, 'timing/data': 0.04838506877422333, 'timing/model': 1.487496447749436, 'learning_rate/qwen_vl_interface': 5.157663382512118e-07, 'learning_rate/action_model': 5.157663382512119e-06, 'epoch': 2.62}) +[2026-06-18 08:56:50,585][starVLA.training.train_starvla][INFO] - Step 6822, Loss: {'train/loss': 0.30251988768577576, 'train/grad_norm_pre_clip': 18.996618270874023, 'timing/data': 0.04249554127454758, 'timing/model': 1.4130918364971876, 'learning_rate/qwen_vl_interface': 5.155907662766711e-07, 'learning_rate/action_model': 5.155907662766711e-06, 'epoch': 2.62}) +[2026-06-18 08:56:54,848][starVLA.training.train_starvla][INFO] - Step 6823, Loss: {'train/loss': 0.6025663614273071, 'train/grad_norm_pre_clip': 16.764089584350586, 'timing/data': 0.02304015588015318, 'timing/model': 1.1135855177417397, 'learning_rate/qwen_vl_interface': 5.15416175750259e-07, 'learning_rate/action_model': 5.154161757502589e-06, 'epoch': 2.62}) +[2026-06-18 08:56:59,507][starVLA.training.train_starvla][INFO] - Step 6824, Loss: {'train/loss': 0.5056982636451721, 'train/grad_norm_pre_clip': 20.494524002075195, 'timing/data': 0.02109846193343401, 'timing/model': 1.1052275551483035, 'learning_rate/qwen_vl_interface': 5.15242566708168e-07, 'learning_rate/action_model': 5.15242566708168e-06, 'epoch': 2.63}) +[2026-06-18 08:57:04,081][starVLA.training.train_starvla][INFO] - Step 6825, Loss: {'train/loss': 0.17646263539791107, 'train/grad_norm_pre_clip': 14.740198135375977, 'timing/data': 0.024415893480181694, 'timing/model': 1.0870606992393732, 'learning_rate/qwen_vl_interface': 5.150699391863884e-07, 'learning_rate/action_model': 5.150699391863884e-06, 'epoch': 2.63}) +[2026-06-18 08:57:08,825][starVLA.training.train_starvla][INFO] - Step 6826, Loss: {'train/loss': 0.11010590940713882, 'train/grad_norm_pre_clip': 8.966830253601074, 'timing/data': 0.02502063475549221, 'timing/model': 1.243637909181416, 'learning_rate/qwen_vl_interface': 5.148982932207052e-07, 'learning_rate/action_model': 5.148982932207052e-06, 'epoch': 2.63}) +[2026-06-18 08:57:13,461][starVLA.training.train_starvla][INFO] - Step 6827, Loss: {'train/loss': 0.29116642475128174, 'train/grad_norm_pre_clip': 19.37830352783203, 'timing/data': 0.023724842816591263, 'timing/model': 1.5202703727409244, 'learning_rate/qwen_vl_interface': 5.147276288467008e-07, 'learning_rate/action_model': 5.147276288467008e-06, 'epoch': 2.63}) +[2026-06-18 08:57:17,820][starVLA.training.train_starvla][INFO] - Step 6828, Loss: {'train/loss': 0.3333292007446289, 'train/grad_norm_pre_clip': 18.219587326049805, 'timing/data': 0.026572274044156075, 'timing/model': 1.185306248255074, 'learning_rate/qwen_vl_interface': 5.145579460997544e-07, 'learning_rate/action_model': 5.145579460997544e-06, 'epoch': 2.63}) +[2026-06-18 08:57:22,508][starVLA.training.train_starvla][INFO] - Step 6829, Loss: {'train/loss': 0.35076674818992615, 'train/grad_norm_pre_clip': 19.72819709777832, 'timing/data': 0.022267738357186317, 'timing/model': 1.0933844204992056, 'learning_rate/qwen_vl_interface': 5.143892450150411e-07, 'learning_rate/action_model': 5.143892450150411e-06, 'epoch': 2.63}) +[2026-06-18 08:57:27,242][starVLA.training.train_starvla][INFO] - Step 6830, Loss: {'train/loss': 0.342899352312088, 'train/grad_norm_pre_clip': 16.79452133178711, 'timing/data': 0.023087210953235626, 'timing/model': 1.1327854935079813, 'learning_rate/qwen_vl_interface': 5.142215256275338e-07, 'learning_rate/action_model': 5.142215256275338e-06, 'epoch': 2.63}) +[2026-06-18 08:57:31,885][starVLA.training.train_starvla][INFO] - Step 6831, Loss: {'train/loss': 0.10689728707075119, 'train/grad_norm_pre_clip': 16.423511505126953, 'timing/data': 0.020982108078897, 'timing/model': 1.4382724966853857, 'learning_rate/qwen_vl_interface': 5.140547879719993e-07, 'learning_rate/action_model': 5.140547879719993e-06, 'epoch': 2.63}) +[2026-06-18 08:57:36,583][starVLA.training.train_starvla][INFO] - Step 6832, Loss: {'train/loss': 0.16438639163970947, 'train/grad_norm_pre_clip': 10.04941177368164, 'timing/data': 0.031134371645748615, 'timing/model': 1.5126387821510434, 'learning_rate/qwen_vl_interface': 5.138890320830038e-07, 'learning_rate/action_model': 5.138890320830038e-06, 'epoch': 2.63}) +[2026-06-18 08:57:40,802][starVLA.training.train_starvla][INFO] - Step 6833, Loss: {'train/loss': 0.14277632534503937, 'train/grad_norm_pre_clip': 10.767332077026367, 'timing/data': 0.021806716918945312, 'timing/model': 1.0821367986500263, 'learning_rate/qwen_vl_interface': 5.137242579949081e-07, 'learning_rate/action_model': 5.137242579949081e-06, 'epoch': 2.63}) +[2026-06-18 08:57:45,406][starVLA.training.train_starvla][INFO] - Step 6834, Loss: {'train/loss': 0.718181312084198, 'train/grad_norm_pre_clip': 22.431108474731445, 'timing/data': 0.025651736184954643, 'timing/model': 1.088942864909768, 'learning_rate/qwen_vl_interface': 5.135604657418702e-07, 'learning_rate/action_model': 5.135604657418702e-06, 'epoch': 2.63}) +[2026-06-18 08:57:50,144][starVLA.training.train_starvla][INFO] - Step 6835, Loss: {'train/loss': 0.3640706539154053, 'train/grad_norm_pre_clip': 17.556312561035156, 'timing/data': 0.031086564995348454, 'timing/model': 1.1103053744882345, 'learning_rate/qwen_vl_interface': 5.133976553578443e-07, 'learning_rate/action_model': 5.1339765535784425e-06, 'epoch': 2.63}) +[2026-06-18 08:57:54,901][starVLA.training.train_starvla][INFO] - Step 6836, Loss: {'train/loss': 0.30864188075065613, 'train/grad_norm_pre_clip': 17.091007232666016, 'timing/data': 0.023648476228117943, 'timing/model': 1.1055202605202794, 'learning_rate/qwen_vl_interface': 5.132358268765813e-07, 'learning_rate/action_model': 5.132358268765813e-06, 'epoch': 2.63}) +[2026-06-18 08:57:59,536][starVLA.training.train_starvla][INFO] - Step 6837, Loss: {'train/loss': 0.18484798073768616, 'train/grad_norm_pre_clip': 12.746700286865234, 'timing/data': 0.022620340809226036, 'timing/model': 1.4090471789240837, 'learning_rate/qwen_vl_interface': 5.130749803316284e-07, 'learning_rate/action_model': 5.130749803316284e-06, 'epoch': 2.63}) +[2026-06-18 08:58:03,855][starVLA.training.train_starvla][INFO] - Step 6838, Loss: {'train/loss': 0.19901178777217865, 'train/grad_norm_pre_clip': 14.967034339904785, 'timing/data': 0.02436669636517763, 'timing/model': 1.1795243043452501, 'learning_rate/qwen_vl_interface': 5.129151157563294e-07, 'learning_rate/action_model': 5.129151157563293e-06, 'epoch': 2.63}) +[2026-06-18 08:58:08,566][starVLA.training.train_starvla][INFO] - Step 6839, Loss: {'train/loss': 0.21416912972927094, 'train/grad_norm_pre_clip': 20.056772232055664, 'timing/data': 0.02144006546586752, 'timing/model': 1.149383433163166, 'learning_rate/qwen_vl_interface': 5.127562331838241e-07, 'learning_rate/action_model': 5.12756233183824e-06, 'epoch': 2.63}) +[2026-06-18 08:58:13,159][starVLA.training.train_starvla][INFO] - Step 6840, Loss: {'train/loss': 0.19540362060070038, 'train/grad_norm_pre_clip': 13.140162467956543, 'timing/data': 0.021760222502052784, 'timing/model': 1.1039591385051608, 'learning_rate/qwen_vl_interface': 5.125983326470492e-07, 'learning_rate/action_model': 5.125983326470492e-06, 'epoch': 2.63}) +[2026-06-18 08:58:18,063][starVLA.training.train_starvla][INFO] - Step 6841, Loss: {'train/loss': 0.44001635909080505, 'train/grad_norm_pre_clip': 21.988161087036133, 'timing/data': 0.02740396000444889, 'timing/model': 1.1786801805719733, 'learning_rate/qwen_vl_interface': 5.124414141787372e-07, 'learning_rate/action_model': 5.124414141787372e-06, 'epoch': 2.63}) +[2026-06-18 08:58:22,288][starVLA.training.train_starvla][INFO] - Step 6842, Loss: {'train/loss': 0.5040481686592102, 'train/grad_norm_pre_clip': 17.877235412597656, 'timing/data': 0.02414083294570446, 'timing/model': 1.0981755927205086, 'learning_rate/qwen_vl_interface': 5.122854778114185e-07, 'learning_rate/action_model': 5.122854778114185e-06, 'epoch': 2.63}) +[2026-06-18 08:58:26,844][starVLA.training.train_starvla][INFO] - Step 6843, Loss: {'train/loss': 0.4627074599266052, 'train/grad_norm_pre_clip': 17.93006134033203, 'timing/data': 0.019248615950345993, 'timing/model': 1.0594639731571078, 'learning_rate/qwen_vl_interface': 5.12130523577418e-07, 'learning_rate/action_model': 5.12130523577418e-06, 'epoch': 2.63}) +[2026-06-18 08:58:31,404][starVLA.training.train_starvla][INFO] - Step 6844, Loss: {'train/loss': 0.10410752892494202, 'train/grad_norm_pre_clip': 8.093585968017578, 'timing/data': 0.031970204785466194, 'timing/model': 1.0705389231443405, 'learning_rate/qwen_vl_interface': 5.119765515088579e-07, 'learning_rate/action_model': 5.1197655150885785e-06, 'epoch': 2.63}) +[2026-06-18 08:58:36,025][starVLA.training.train_starvla][INFO] - Step 6845, Loss: {'train/loss': 0.11353248357772827, 'train/grad_norm_pre_clip': 8.966374397277832, 'timing/data': 0.022664178162813187, 'timing/model': 1.0677322456613183, 'learning_rate/qwen_vl_interface': 5.11823561637657e-07, 'learning_rate/action_model': 5.1182356163765705e-06, 'epoch': 2.63}) +[2026-06-18 08:58:40,854][starVLA.training.train_starvla][INFO] - Step 6846, Loss: {'train/loss': 0.49023520946502686, 'train/grad_norm_pre_clip': 25.137500762939453, 'timing/data': 0.02316581830382347, 'timing/model': 1.2646417329087853, 'learning_rate/qwen_vl_interface': 5.116715539955306e-07, 'learning_rate/action_model': 5.116715539955305e-06, 'epoch': 2.63}) +[2026-06-18 08:58:45,516][starVLA.training.train_starvla][INFO] - Step 6847, Loss: {'train/loss': 0.21167035400867462, 'train/grad_norm_pre_clip': 19.559825897216797, 'timing/data': 0.021012572571635246, 'timing/model': 1.3911443622782826, 'learning_rate/qwen_vl_interface': 5.115205286139895e-07, 'learning_rate/action_model': 5.115205286139895e-06, 'epoch': 2.63}) +[2026-06-18 08:58:49,830][starVLA.training.train_starvla][INFO] - Step 6848, Loss: {'train/loss': 0.1814318746328354, 'train/grad_norm_pre_clip': 14.702069282531738, 'timing/data': 0.022149897180497646, 'timing/model': 1.1267304848879576, 'learning_rate/qwen_vl_interface': 5.11370485524342e-07, 'learning_rate/action_model': 5.1137048552434205e-06, 'epoch': 2.63}) +[2026-06-18 08:58:54,543][starVLA.training.train_starvla][INFO] - Step 6849, Loss: {'train/loss': 0.22989434003829956, 'train/grad_norm_pre_clip': 16.491859436035156, 'timing/data': 0.023223254829645157, 'timing/model': 1.1270915502682328, 'learning_rate/qwen_vl_interface': 5.112214247576915e-07, 'learning_rate/action_model': 5.1122142475769155e-06, 'epoch': 2.63}) +[2026-06-18 08:58:59,203][starVLA.training.train_starvla][INFO] - Step 6850, Loss: {'train/loss': 0.14899230003356934, 'train/grad_norm_pre_clip': 22.69563865661621, 'timing/data': 0.020938646979629993, 'timing/model': 1.1253311773762107, 'learning_rate/qwen_vl_interface': 5.110733463449384e-07, 'learning_rate/action_model': 5.110733463449385e-06, 'epoch': 2.64}) +[2026-06-18 08:59:04,065][starVLA.training.train_starvla][INFO] - Step 6851, Loss: {'train/loss': 0.2153618335723877, 'train/grad_norm_pre_clip': 15.073908805847168, 'timing/data': 0.022153842262923717, 'timing/model': 1.608233661390841, 'learning_rate/qwen_vl_interface': 5.109262503167808e-07, 'learning_rate/action_model': 5.109262503167808e-06, 'epoch': 2.64}) +[2026-06-18 08:59:08,722][starVLA.training.train_starvla][INFO] - Step 6852, Loss: {'train/loss': 0.11026649177074432, 'train/grad_norm_pre_clip': 9.44603443145752, 'timing/data': 0.024556217715144157, 'timing/model': 1.511191206984222, 'learning_rate/qwen_vl_interface': 5.107801367037106e-07, 'learning_rate/action_model': 5.107801367037107e-06, 'epoch': 2.64}) +[2026-06-18 08:59:12,933][starVLA.training.train_starvla][INFO] - Step 6853, Loss: {'train/loss': 0.10285531729459763, 'train/grad_norm_pre_clip': 10.524480819702148, 'timing/data': 0.02107710763812065, 'timing/model': 1.080031644552946, 'learning_rate/qwen_vl_interface': 5.106350055360173e-07, 'learning_rate/action_model': 5.106350055360173e-06, 'epoch': 2.64}) +[2026-06-18 08:59:17,591][starVLA.training.train_starvla][INFO] - Step 6854, Loss: {'train/loss': 0.5667940974235535, 'train/grad_norm_pre_clip': 22.096372604370117, 'timing/data': 0.03209713473916054, 'timing/model': 1.0982987694442272, 'learning_rate/qwen_vl_interface': 5.104908568437875e-07, 'learning_rate/action_model': 5.104908568437875e-06, 'epoch': 2.64}) +[2026-06-18 08:59:22,329][starVLA.training.train_starvla][INFO] - Step 6855, Loss: {'train/loss': 0.24227021634578705, 'train/grad_norm_pre_clip': 18.20513153076172, 'timing/data': 0.024194443598389626, 'timing/model': 1.1122262766584754, 'learning_rate/qwen_vl_interface': 5.103476906569034e-07, 'learning_rate/action_model': 5.103476906569033e-06, 'epoch': 2.64}) +[2026-06-18 08:59:26,797][starVLA.training.train_starvla][INFO] - Step 6856, Loss: {'train/loss': 0.28608912229537964, 'train/grad_norm_pre_clip': 15.374186515808105, 'timing/data': 0.022002930752933025, 'timing/model': 1.0772476494312286, 'learning_rate/qwen_vl_interface': 5.102055070050428e-07, 'learning_rate/action_model': 5.102055070050427e-06, 'epoch': 2.64}) +[2026-06-18 08:59:31,524][starVLA.training.train_starvla][INFO] - Step 6857, Loss: {'train/loss': 0.33571135997772217, 'train/grad_norm_pre_clip': 19.72078514099121, 'timing/data': 0.02270568162202835, 'timing/model': 1.4283383814617991, 'learning_rate/qwen_vl_interface': 5.100643059176803e-07, 'learning_rate/action_model': 5.100643059176802e-06, 'epoch': 2.64}) +[2026-06-18 08:59:35,766][starVLA.training.train_starvla][INFO] - Step 6858, Loss: {'train/loss': 0.4276665151119232, 'train/grad_norm_pre_clip': 17.857852935791016, 'timing/data': 0.028637945652008057, 'timing/model': 1.0706354090943933, 'learning_rate/qwen_vl_interface': 5.099240874240883e-07, 'learning_rate/action_model': 5.099240874240883e-06, 'epoch': 2.64}) +[2026-06-18 08:59:40,404][starVLA.training.train_starvla][INFO] - Step 6859, Loss: {'train/loss': 0.19335351884365082, 'train/grad_norm_pre_clip': 40.64745330810547, 'timing/data': 0.025995799340307713, 'timing/model': 1.0948608024045825, 'learning_rate/qwen_vl_interface': 5.097848515533331e-07, 'learning_rate/action_model': 5.0978485155333305e-06, 'epoch': 2.64}) +[2026-06-18 08:59:44,945][starVLA.training.train_starvla][INFO] - Step 6860, Loss: {'train/loss': 0.2171863615512848, 'train/grad_norm_pre_clip': 15.78123950958252, 'timing/data': 0.024515897035598755, 'timing/model': 1.094867318868637, 'learning_rate/qwen_vl_interface': 5.096465983342791e-07, 'learning_rate/action_model': 5.096465983342791e-06, 'epoch': 2.64}) +[2026-06-18 08:59:49,549][starVLA.training.train_starvla][INFO] - Step 6861, Loss: {'train/loss': 0.2741609215736389, 'train/grad_norm_pre_clip': 18.797624588012695, 'timing/data': 0.020380469970405102, 'timing/model': 1.4002640480175614, 'learning_rate/qwen_vl_interface': 5.095093277955859e-07, 'learning_rate/action_model': 5.095093277955859e-06, 'epoch': 2.64}) +[2026-06-18 08:59:54,190][starVLA.training.train_starvla][INFO] - Step 6862, Loss: {'train/loss': 0.19649748504161835, 'train/grad_norm_pre_clip': 15.923287391662598, 'timing/data': 0.023626250214874744, 'timing/model': 1.459280977025628, 'learning_rate/qwen_vl_interface': 5.093730399657101e-07, 'learning_rate/action_model': 5.0937303996571e-06, 'epoch': 2.64}) +[2026-06-18 08:59:58,575][starVLA.training.train_starvla][INFO] - Step 6863, Loss: {'train/loss': 0.302315890789032, 'train/grad_norm_pre_clip': 15.01353645324707, 'timing/data': 0.04066237714141607, 'timing/model': 1.184151149354875, 'learning_rate/qwen_vl_interface': 5.092377348729043e-07, 'learning_rate/action_model': 5.092377348729043e-06, 'epoch': 2.64}) +[2026-06-18 09:00:03,456][starVLA.training.train_starvla][INFO] - Step 6864, Loss: {'train/loss': 0.07513561844825745, 'train/grad_norm_pre_clip': 6.991805553436279, 'timing/data': 0.02378845028579235, 'timing/model': 1.1242040442302823, 'learning_rate/qwen_vl_interface': 5.091034125452171e-07, 'learning_rate/action_model': 5.091034125452171e-06, 'epoch': 2.64}) +[2026-06-18 09:00:08,242][starVLA.training.train_starvla][INFO] - Step 6865, Loss: {'train/loss': 0.29212486743927, 'train/grad_norm_pre_clip': 17.466245651245117, 'timing/data': 0.028069731779396534, 'timing/model': 1.2143473690375686, 'learning_rate/qwen_vl_interface': 5.089700730104944e-07, 'learning_rate/action_model': 5.089700730104944e-06, 'epoch': 2.64}) +[2026-06-18 09:00:12,955][starVLA.training.train_starvla][INFO] - Step 6866, Loss: {'train/loss': 0.29870080947875977, 'train/grad_norm_pre_clip': 10.280374526977539, 'timing/data': 0.024727394804358482, 'timing/model': 1.1163262268528342, 'learning_rate/qwen_vl_interface': 5.08837716296377e-07, 'learning_rate/action_model': 5.0883771629637696e-06, 'epoch': 2.64}) +[2026-06-18 09:00:17,527][starVLA.training.train_starvla][INFO] - Step 6867, Loss: {'train/loss': 0.1478937864303589, 'train/grad_norm_pre_clip': 11.7516508102417, 'timing/data': 0.02936753910034895, 'timing/model': 1.4112279079854488, 'learning_rate/qwen_vl_interface': 5.087063424303026e-07, 'learning_rate/action_model': 5.087063424303026e-06, 'epoch': 2.64}) +[2026-06-18 09:00:21,738][starVLA.training.train_starvla][INFO] - Step 6868, Loss: {'train/loss': 0.15084224939346313, 'train/grad_norm_pre_clip': 15.129033088684082, 'timing/data': 0.020551173947751522, 'timing/model': 1.064699862152338, 'learning_rate/qwen_vl_interface': 5.085759514395052e-07, 'learning_rate/action_model': 5.085759514395051e-06, 'epoch': 2.64}) +[2026-06-18 09:00:26,193][starVLA.training.train_starvla][INFO] - Step 6869, Loss: {'train/loss': 0.051646094769239426, 'train/grad_norm_pre_clip': 5.071476936340332, 'timing/data': 0.02340673189610243, 'timing/model': 1.0652493741363287, 'learning_rate/qwen_vl_interface': 5.084465433510151e-07, 'learning_rate/action_model': 5.0844654335101515e-06, 'epoch': 2.64}) +[2026-06-18 09:00:30,927][starVLA.training.train_starvla][INFO] - Step 6870, Loss: {'train/loss': 0.20430760085582733, 'train/grad_norm_pre_clip': 12.53541088104248, 'timing/data': 0.02298764605075121, 'timing/model': 1.190728010609746, 'learning_rate/qwen_vl_interface': 5.083181181916588e-07, 'learning_rate/action_model': 5.083181181916588e-06, 'epoch': 2.64}) +[2026-06-18 09:00:35,475][starVLA.training.train_starvla][INFO] - Step 6871, Loss: {'train/loss': 0.1854475736618042, 'train/grad_norm_pre_clip': 10.282721519470215, 'timing/data': 0.022201807238161564, 'timing/model': 1.0719354189932346, 'learning_rate/qwen_vl_interface': 5.081906759880585e-07, 'learning_rate/action_model': 5.081906759880585e-06, 'epoch': 2.64}) +[2026-06-18 09:00:39,938][starVLA.training.train_starvla][INFO] - Step 6872, Loss: {'train/loss': 0.31216537952423096, 'train/grad_norm_pre_clip': 15.188684463500977, 'timing/data': 0.02244466356933117, 'timing/model': 1.1513017024844885, 'learning_rate/qwen_vl_interface': 5.080642167666338e-07, 'learning_rate/action_model': 5.080642167666337e-06, 'epoch': 2.64}) +[2026-06-18 09:00:44,569][starVLA.training.train_starvla][INFO] - Step 6873, Loss: {'train/loss': 0.28905370831489563, 'train/grad_norm_pre_clip': 16.436872482299805, 'timing/data': 0.02347459178417921, 'timing/model': 1.1047466062009335, 'learning_rate/qwen_vl_interface': 5.079387405535995e-07, 'learning_rate/action_model': 5.079387405535995e-06, 'epoch': 2.64}) +[2026-06-18 09:00:49,097][starVLA.training.train_starvla][INFO] - Step 6874, Loss: {'train/loss': 0.37120679020881653, 'train/grad_norm_pre_clip': 16.526384353637695, 'timing/data': 0.02191959787160158, 'timing/model': 1.0714537594467402, 'learning_rate/qwen_vl_interface': 5.07814247374967e-07, 'learning_rate/action_model': 5.078142473749671e-06, 'epoch': 2.64}) +[2026-06-18 09:00:53,609][starVLA.training.train_starvla][INFO] - Step 6875, Loss: {'train/loss': 0.18999935686588287, 'train/grad_norm_pre_clip': 13.555469512939453, 'timing/data': 0.020464928820729256, 'timing/model': 1.065325053408742, 'learning_rate/qwen_vl_interface': 5.07690737256544e-07, 'learning_rate/action_model': 5.07690737256544e-06, 'epoch': 2.64}) +[2026-06-18 09:00:58,063][starVLA.training.train_starvla][INFO] - Step 6876, Loss: {'train/loss': 0.21098193526268005, 'train/grad_norm_pre_clip': 11.064117431640625, 'timing/data': 0.021185920573771, 'timing/model': 1.0734527334570885, 'learning_rate/qwen_vl_interface': 5.075682102239336e-07, 'learning_rate/action_model': 5.075682102239335e-06, 'epoch': 2.65}) +[2026-06-18 09:01:02,733][starVLA.training.train_starvla][INFO] - Step 6877, Loss: {'train/loss': 0.23373521864414215, 'train/grad_norm_pre_clip': 17.460983276367188, 'timing/data': 0.021846991032361984, 'timing/model': 1.3968156455084682, 'learning_rate/qwen_vl_interface': 5.074466663025369e-07, 'learning_rate/action_model': 5.074466663025369e-06, 'epoch': 2.65}) +[2026-06-18 09:01:07,030][starVLA.training.train_starvla][INFO] - Step 6878, Loss: {'train/loss': 0.19426631927490234, 'train/grad_norm_pre_clip': 15.523968696594238, 'timing/data': 0.02105826884508133, 'timing/model': 1.0689548030495644, 'learning_rate/qwen_vl_interface': 5.07326105517549e-07, 'learning_rate/action_model': 5.07326105517549e-06, 'epoch': 2.65}) +[2026-06-18 09:01:11,618][starVLA.training.train_starvla][INFO] - Step 6879, Loss: {'train/loss': 0.6018519997596741, 'train/grad_norm_pre_clip': 20.234193801879883, 'timing/data': 0.02334574516862631, 'timing/model': 1.069770042784512, 'learning_rate/qwen_vl_interface': 5.072065278939628e-07, 'learning_rate/action_model': 5.072065278939627e-06, 'epoch': 2.65}) +[2026-06-18 09:01:16,220][starVLA.training.train_starvla][INFO] - Step 6880, Loss: {'train/loss': 0.13771973550319672, 'train/grad_norm_pre_clip': 12.790848731994629, 'timing/data': 0.022958182729780674, 'timing/model': 1.1274820370599627, 'learning_rate/qwen_vl_interface': 5.070879334565671e-07, 'learning_rate/action_model': 5.070879334565671e-06, 'epoch': 2.65}) +[2026-06-18 09:01:20,916][starVLA.training.train_starvla][INFO] - Step 6881, Loss: {'train/loss': 0.09358998388051987, 'train/grad_norm_pre_clip': 9.026893615722656, 'timing/data': 0.03489320818334818, 'timing/model': 1.441676875576377, 'learning_rate/qwen_vl_interface': 5.069703222299464e-07, 'learning_rate/action_model': 5.069703222299463e-06, 'epoch': 2.65}) +[2026-06-18 09:01:25,437][starVLA.training.train_starvla][INFO] - Step 6882, Loss: {'train/loss': 0.10111138969659805, 'train/grad_norm_pre_clip': 9.514629364013672, 'timing/data': 0.021817386150360107, 'timing/model': 1.412196135148406, 'learning_rate/qwen_vl_interface': 5.068536942384813e-07, 'learning_rate/action_model': 5.068536942384812e-06, 'epoch': 2.65}) +[2026-06-18 09:01:29,702][starVLA.training.train_starvla][INFO] - Step 6883, Loss: {'train/loss': 0.4700479805469513, 'train/grad_norm_pre_clip': 30.221689224243164, 'timing/data': 0.023036446422338486, 'timing/model': 1.1340516451746225, 'learning_rate/qwen_vl_interface': 5.067380495063494e-07, 'learning_rate/action_model': 5.067380495063494e-06, 'epoch': 2.65}) +[2026-06-18 09:01:34,238][starVLA.training.train_starvla][INFO] - Step 6884, Loss: {'train/loss': 0.23328769207000732, 'train/grad_norm_pre_clip': 18.286197662353516, 'timing/data': 0.02113087847828865, 'timing/model': 1.0809236280620098, 'learning_rate/qwen_vl_interface': 5.066233880575239e-07, 'learning_rate/action_model': 5.0662338805752395e-06, 'epoch': 2.65}) +[2026-06-18 09:01:38,798][starVLA.training.train_starvla][INFO] - Step 6885, Loss: {'train/loss': 0.24620011448860168, 'train/grad_norm_pre_clip': 12.810101509094238, 'timing/data': 0.021550278179347515, 'timing/model': 1.1280368389561772, 'learning_rate/qwen_vl_interface': 5.065097099157743e-07, 'learning_rate/action_model': 5.065097099157743e-06, 'epoch': 2.65}) +[2026-06-18 09:01:43,364][starVLA.training.train_starvla][INFO] - Step 6886, Loss: {'train/loss': 0.33284851908683777, 'train/grad_norm_pre_clip': 13.472822189331055, 'timing/data': 0.022139104083180428, 'timing/model': 1.0907040853053331, 'learning_rate/qwen_vl_interface': 5.063970151046654e-07, 'learning_rate/action_model': 5.0639701510466545e-06, 'epoch': 2.65}) +[2026-06-18 09:01:47,987][starVLA.training.train_starvla][INFO] - Step 6887, Loss: {'train/loss': 0.2617555558681488, 'train/grad_norm_pre_clip': 16.369474411010742, 'timing/data': 0.02689501829445362, 'timing/model': 1.4084399277344346, 'learning_rate/qwen_vl_interface': 5.062853036475604e-07, 'learning_rate/action_model': 5.062853036475603e-06, 'epoch': 2.65}) +[2026-06-18 09:01:52,223][starVLA.training.train_starvla][INFO] - Step 6888, Loss: {'train/loss': 0.36227405071258545, 'train/grad_norm_pre_clip': 16.700464248657227, 'timing/data': 0.024561970494687557, 'timing/model': 1.0813285568729043, 'learning_rate/qwen_vl_interface': 5.061745755676161e-07, 'learning_rate/action_model': 5.061745755676161e-06, 'epoch': 2.65}) +[2026-06-18 09:01:56,763][starVLA.training.train_starvla][INFO] - Step 6889, Loss: {'train/loss': 0.24416129291057587, 'train/grad_norm_pre_clip': 10.95121955871582, 'timing/data': 0.020553654991090298, 'timing/model': 1.0749055743217468, 'learning_rate/qwen_vl_interface': 5.060648308877866e-07, 'learning_rate/action_model': 5.0606483088778656e-06, 'epoch': 2.65}) +[2026-06-18 09:02:01,280][starVLA.training.train_starvla][INFO] - Step 6890, Loss: {'train/loss': 0.29378047585487366, 'train/grad_norm_pre_clip': 15.128369331359863, 'timing/data': 0.022611528635025024, 'timing/model': 1.082769681699574, 'learning_rate/qwen_vl_interface': 5.059560696308229e-07, 'learning_rate/action_model': 5.059560696308228e-06, 'epoch': 2.65}) +[2026-06-18 09:02:05,773][starVLA.training.train_starvla][INFO] - Step 6891, Loss: {'train/loss': 0.21526135504245758, 'train/grad_norm_pre_clip': 13.074483871459961, 'timing/data': 0.021864701062440872, 'timing/model': 1.3883912162855268, 'learning_rate/qwen_vl_interface': 5.0584829181927e-07, 'learning_rate/action_model': 5.058482918192699e-06, 'epoch': 2.65}) +[2026-06-18 09:02:10,569][starVLA.training.train_starvla][INFO] - Step 6892, Loss: {'train/loss': 0.4111497402191162, 'train/grad_norm_pre_clip': 16.047622680664062, 'timing/data': 0.021950097754597664, 'timing/model': 1.5468612601980567, 'learning_rate/qwen_vl_interface': 5.05741497475472e-07, 'learning_rate/action_model': 5.05741497475472e-06, 'epoch': 2.65}) +[2026-06-18 09:02:14,888][starVLA.training.train_starvla][INFO] - Step 6893, Loss: {'train/loss': 0.3711533844470978, 'train/grad_norm_pre_clip': 16.99597930908203, 'timing/data': 0.024109011515975, 'timing/model': 1.0748790428042412, 'learning_rate/qwen_vl_interface': 5.056356866215663e-07, 'learning_rate/action_model': 5.056356866215662e-06, 'epoch': 2.65}) +[2026-06-18 09:02:19,489][starVLA.training.train_starvla][INFO] - Step 6894, Loss: {'train/loss': 0.22582371532917023, 'train/grad_norm_pre_clip': 17.169740676879883, 'timing/data': 0.0233409833163023, 'timing/model': 1.0832099672406912, 'learning_rate/qwen_vl_interface': 5.055308592794876e-07, 'learning_rate/action_model': 5.0553085927948756e-06, 'epoch': 2.65}) +[2026-06-18 09:02:24,139][starVLA.training.train_starvla][INFO] - Step 6895, Loss: {'train/loss': 0.1822604238986969, 'train/grad_norm_pre_clip': 17.038238525390625, 'timing/data': 0.02483859471976757, 'timing/model': 1.0849267039448023, 'learning_rate/qwen_vl_interface': 5.054270154709674e-07, 'learning_rate/action_model': 5.054270154709673e-06, 'epoch': 2.65}) +[2026-06-18 09:02:28,749][starVLA.training.train_starvla][INFO] - Step 6896, Loss: {'train/loss': 0.11742468923330307, 'train/grad_norm_pre_clip': 11.631773948669434, 'timing/data': 0.021281037479639053, 'timing/model': 1.0948193287476897, 'learning_rate/qwen_vl_interface': 5.053241552175322e-07, 'learning_rate/action_model': 5.053241552175321e-06, 'epoch': 2.65}) +[2026-06-18 09:02:33,319][starVLA.training.train_starvla][INFO] - Step 6897, Loss: {'train/loss': 0.16812904179096222, 'train/grad_norm_pre_clip': 11.286633491516113, 'timing/data': 0.022051045671105385, 'timing/model': 1.4157401006668806, 'learning_rate/qwen_vl_interface': 5.05222278540505e-07, 'learning_rate/action_model': 5.05222278540505e-06, 'epoch': 2.65}) +[2026-06-18 09:02:37,508][starVLA.training.train_starvla][INFO] - Step 6898, Loss: {'train/loss': 0.2185472697019577, 'train/grad_norm_pre_clip': 10.75057315826416, 'timing/data': 0.01986095029860735, 'timing/model': 1.0716714151203632, 'learning_rate/qwen_vl_interface': 5.051213854610051e-07, 'learning_rate/action_model': 5.0512138546100514e-06, 'epoch': 2.65}) +[2026-06-18 09:02:42,187][starVLA.training.train_starvla][INFO] - Step 6899, Loss: {'train/loss': 0.11717391014099121, 'train/grad_norm_pre_clip': 17.475265502929688, 'timing/data': 0.023722107522189617, 'timing/model': 1.1157483775168657, 'learning_rate/qwen_vl_interface': 5.050214759999478e-07, 'learning_rate/action_model': 5.050214759999478e-06, 'epoch': 2.65}) +[2026-06-18 09:05:19,419][starVLA.training.train_starvla][INFO] - Step 6900, Loss: {'train/loss': 0.10573501139879227, 'train/grad_norm_pre_clip': 11.150376319885254, 'eval/loss': 0.4297175705432892, 'eval/latency_4/loss': 0.4297175705432892, 'eval/demon_fire_precision': 0.9698234349919743, 'eval/demon_fire_recall': 0.9811627151672621, 'eval/demon_fire_f1': 0.9754601226993864, 'eval/demon_fire_rate_model': 0.9148311306901615, 'eval/demon_fire_rate_teacher': 0.9042584434654919, 'eval/demon_left_precision': 0.7906976744186046, 'eval/demon_left_recall': 0.6986301369863014, 'eval/demon_left_f1': 0.7418181818181818, 'eval/demon_left_rate_model': 0.11365638766519824, 'eval/demon_left_rate_teacher': 0.12863436123348018, 'eval/demon_right_precision': 0.8116760828625236, 'eval/demon_right_recall': 0.7779783393501805, 'eval/demon_right_f1': 0.7944700460829494, 'eval/demon_right_rate_model': 0.15594713656387665, 'eval/demon_right_rate_teacher': 0.16270190895741557, 'eval/demon_move_f1': 0.7681441139505656, 'eval/demon_cc_f1': 0.9023953191734563, 'eval/latency_4/demon_fire_precision': 0.9698234349919743, 'eval/latency_4/demon_fire_recall': 0.9811627151672621, 'eval/latency_4/demon_fire_f1': 0.9754601226993864, 'eval/latency_4/demon_fire_rate_model': 0.9148311306901615, 'eval/latency_4/demon_fire_rate_teacher': 0.9042584434654919, 'eval/latency_4/demon_left_precision': 0.7906976744186046, 'eval/latency_4/demon_left_recall': 0.6986301369863014, 'eval/latency_4/demon_left_f1': 0.7418181818181818, 'eval/latency_4/demon_left_rate_model': 0.11365638766519824, 'eval/latency_4/demon_left_rate_teacher': 0.12863436123348018, 'eval/latency_4/demon_right_precision': 0.8116760828625236, 'eval/latency_4/demon_right_recall': 0.7779783393501805, 'eval/latency_4/demon_right_f1': 0.7944700460829494, 'eval/latency_4/demon_right_rate_model': 0.15594713656387665, 'eval/latency_4/demon_right_rate_teacher': 0.16270190895741557, 'eval/latency_4/demon_move_f1': 0.7681441139505656, 'eval/latency_4/demon_cc_f1': 0.9023953191734563, 'timing/data': 0.024549581110477448, 'timing/model': 1.070142355747521, 'learning_rate/qwen_vl_interface': 5.049225501780441e-07, 'learning_rate/action_model': 5.049225501780441e-06, 'epoch': 2.65}) +[2026-06-18 09:05:19,439][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_6900_state +[2026-06-18 09:05:19,447][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 09:05:29,881][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_6900_state/model.safetensors +[2026-06-18 09:05:53,528][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_6900_state/optimizer.bin +[2026-06-18 09:05:53,530][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_6900_state/sampler.bin +[2026-06-18 09:05:53,531][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_6900_state/sampler_1.bin +[2026-06-18 09:05:53,534][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_6900_state/random_states_0.pkl +[2026-06-18 09:05:53,540][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 6900 +[2026-06-18 09:05:53,540][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 09:05:53,560][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 09:05:58,710][starVLA.training.train_starvla][INFO] - Step 6901, Loss: {'train/loss': 0.4113090932369232, 'train/grad_norm_pre_clip': 24.448122024536133, 'timing/data': 0.03365160804241896, 'timing/model': 1.2391863195225596, 'learning_rate/qwen_vl_interface': 5.048246080158016e-07, 'learning_rate/action_model': 5.048246080158016e-06, 'epoch': 2.65}) +[2026-06-18 09:06:03,341][starVLA.training.train_starvla][INFO] - Step 6902, Loss: {'train/loss': 0.2664245367050171, 'train/grad_norm_pre_clip': 13.252823829650879, 'timing/data': 0.024623610079288483, 'timing/model': 1.159818748012185, 'learning_rate/qwen_vl_interface': 5.047276495335243e-07, 'learning_rate/action_model': 5.047276495335242e-06, 'epoch': 2.66}) +[2026-06-18 09:06:08,490][starVLA.training.train_starvla][INFO] - Step 6903, Loss: {'train/loss': 0.1960563361644745, 'train/grad_norm_pre_clip': 18.146900177001953, 'timing/data': 0.023920290172100067, 'timing/model': 1.4059410691261292, 'learning_rate/qwen_vl_interface': 5.046316747513115e-07, 'learning_rate/action_model': 5.046316747513115e-06, 'epoch': 2.66}) +[2026-06-18 09:06:13,713][starVLA.training.train_starvla][INFO] - Step 6904, Loss: {'train/loss': 0.33687907457351685, 'train/grad_norm_pre_clip': 20.29535484313965, 'timing/data': 0.02545082662254572, 'timing/model': 1.121400136500597, 'learning_rate/qwen_vl_interface': 5.045366836890585e-07, 'learning_rate/action_model': 5.045366836890584e-06, 'epoch': 2.66}) +[2026-06-18 09:06:18,814][starVLA.training.train_starvla][INFO] - Step 6905, Loss: {'train/loss': 0.13543486595153809, 'train/grad_norm_pre_clip': 15.869884490966797, 'timing/data': 0.024784001521766186, 'timing/model': 1.1744988095015287, 'learning_rate/qwen_vl_interface': 5.044426763664572e-07, 'learning_rate/action_model': 5.044426763664571e-06, 'epoch': 2.66}) +[2026-06-18 09:06:23,562][starVLA.training.train_starvla][INFO] - Step 6906, Loss: {'train/loss': 0.10433144867420197, 'train/grad_norm_pre_clip': 10.189495086669922, 'timing/data': 0.03897547721862793, 'timing/model': 1.1243987027555704, 'learning_rate/qwen_vl_interface': 5.043496528029954e-07, 'learning_rate/action_model': 5.043496528029954e-06, 'epoch': 2.66}) +[2026-06-18 09:06:28,778][starVLA.training.train_starvla][INFO] - Step 6907, Loss: {'train/loss': 0.2954588830471039, 'train/grad_norm_pre_clip': 27.367475509643555, 'timing/data': 0.024142789654433727, 'timing/model': 1.4503078470006585, 'learning_rate/qwen_vl_interface': 5.042576130179573e-07, 'learning_rate/action_model': 5.042576130179572e-06, 'epoch': 2.66}) +[2026-06-18 09:06:33,568][starVLA.training.train_starvla][INFO] - Step 6908, Loss: {'train/loss': 0.2757115960121155, 'train/grad_norm_pre_clip': 18.255250930786133, 'timing/data': 0.021550949662923813, 'timing/model': 1.2318181712180376, 'learning_rate/qwen_vl_interface': 5.041665570304227e-07, 'learning_rate/action_model': 5.041665570304226e-06, 'epoch': 2.66}) +[2026-06-18 09:06:38,599][starVLA.training.train_starvla][INFO] - Step 6909, Loss: {'train/loss': 0.3349924683570862, 'train/grad_norm_pre_clip': 14.704529762268066, 'timing/data': 0.026190467178821564, 'timing/model': 1.309499679133296, 'learning_rate/qwen_vl_interface': 5.040764848592675e-07, 'learning_rate/action_model': 5.040764848592675e-06, 'epoch': 2.66}) +[2026-06-18 09:06:43,481][starVLA.training.train_starvla][INFO] - Step 6910, Loss: {'train/loss': 0.3685409426689148, 'train/grad_norm_pre_clip': 21.36382293701172, 'timing/data': 0.02444553840905428, 'timing/model': 1.1961818989366293, 'learning_rate/qwen_vl_interface': 5.039873965231644e-07, 'learning_rate/action_model': 5.039873965231643e-06, 'epoch': 2.66}) +[2026-06-18 09:06:48,312][starVLA.training.train_starvla][INFO] - Step 6911, Loss: {'train/loss': 0.27636030316352844, 'train/grad_norm_pre_clip': 11.25230598449707, 'timing/data': 0.028284930624067783, 'timing/model': 1.1150431130081415, 'learning_rate/qwen_vl_interface': 5.0389929204058e-07, 'learning_rate/action_model': 5.0389929204058e-06, 'epoch': 2.66}) +[2026-06-18 09:06:52,815][starVLA.training.train_starvla][INFO] - Step 6912, Loss: {'train/loss': 0.18155544996261597, 'train/grad_norm_pre_clip': 14.87589168548584, 'timing/data': 0.02721086610108614, 'timing/model': 1.0910022538155317, 'learning_rate/qwen_vl_interface': 5.038121714297802e-07, 'learning_rate/action_model': 5.038121714297802e-06, 'epoch': 2.66}) +[2026-06-18 09:06:57,824][starVLA.training.train_starvla][INFO] - Step 6913, Loss: {'train/loss': 0.33386895060539246, 'train/grad_norm_pre_clip': 20.664438247680664, 'timing/data': 0.023077775724232197, 'timing/model': 1.1081263301894069, 'learning_rate/qwen_vl_interface': 5.037260347088236e-07, 'learning_rate/action_model': 5.037260347088236e-06, 'epoch': 2.66}) +[2026-06-18 09:07:02,879][starVLA.training.train_starvla][INFO] - Step 6914, Loss: {'train/loss': 0.40250203013420105, 'train/grad_norm_pre_clip': 18.238370895385742, 'timing/data': 0.030232269316911697, 'timing/model': 1.3792936662212014, 'learning_rate/qwen_vl_interface': 5.03640881895568e-07, 'learning_rate/action_model': 5.03640881895568e-06, 'epoch': 2.66}) +[2026-06-18 09:07:07,822][starVLA.training.train_starvla][INFO] - Step 6915, Loss: {'train/loss': 0.5120146870613098, 'train/grad_norm_pre_clip': 23.945375442504883, 'timing/data': 0.02564518339931965, 'timing/model': 1.1925916308537126, 'learning_rate/qwen_vl_interface': 5.035567130076642e-07, 'learning_rate/action_model': 5.035567130076642e-06, 'epoch': 2.66}) +[2026-06-18 09:07:13,009][starVLA.training.train_starvla][INFO] - Step 6916, Loss: {'train/loss': 0.22628173232078552, 'train/grad_norm_pre_clip': 20.182750701904297, 'timing/data': 0.001105939969420433, 'timing/model': 1.6486920164898038, 'learning_rate/qwen_vl_interface': 5.034735280625617e-07, 'learning_rate/action_model': 5.034735280625616e-06, 'epoch': 2.66}) +[2026-06-18 09:07:18,137][starVLA.training.train_starvla][INFO] - Step 6917, Loss: {'train/loss': 0.12033647298812866, 'train/grad_norm_pre_clip': 8.069720268249512, 'timing/data': 0.0011203838512301445, 'timing/model': 1.4164056675508618, 'learning_rate/qwen_vl_interface': 5.033913270775039e-07, 'learning_rate/action_model': 5.033913270775039e-06, 'epoch': 2.66}) +[2026-06-18 09:07:22,555][starVLA.training.train_starvla][INFO] - Step 6918, Loss: {'train/loss': 0.19435206055641174, 'train/grad_norm_pre_clip': 79.15243530273438, 'timing/data': 0.001366766169667244, 'timing/model': 1.136599611490965, 'learning_rate/qwen_vl_interface': 5.03310110069532e-07, 'learning_rate/action_model': 5.03310110069532e-06, 'epoch': 2.66}) +[2026-06-18 09:07:27,390][starVLA.training.train_starvla][INFO] - Step 6919, Loss: {'train/loss': 0.3808359205722809, 'train/grad_norm_pre_clip': 21.795879364013672, 'timing/data': 0.0012312382459640503, 'timing/model': 1.1746802888810635, 'learning_rate/qwen_vl_interface': 5.032298770554818e-07, 'learning_rate/action_model': 5.032298770554818e-06, 'epoch': 2.66}) +[2026-06-18 09:07:32,341][starVLA.training.train_starvla][INFO] - Step 6920, Loss: {'train/loss': 0.11768632382154465, 'train/grad_norm_pre_clip': 12.591135025024414, 'timing/data': 0.001133945770561695, 'timing/model': 1.08310139272362, 'learning_rate/qwen_vl_interface': 5.031506280519855e-07, 'learning_rate/action_model': 5.031506280519855e-06, 'epoch': 2.66}) +[2026-06-18 09:07:37,274][starVLA.training.train_starvla][INFO] - Step 6921, Loss: {'train/loss': 0.1853543370962143, 'train/grad_norm_pre_clip': 14.873512268066406, 'timing/data': 0.0019850023090839386, 'timing/model': 1.1940862154588103, 'learning_rate/qwen_vl_interface': 5.030723630754723e-07, 'learning_rate/action_model': 5.030723630754723e-06, 'epoch': 2.66}) +[2026-06-18 09:07:41,988][starVLA.training.train_starvla][INFO] - Step 6922, Loss: {'train/loss': 0.07090188562870026, 'train/grad_norm_pre_clip': 7.50974178314209, 'timing/data': 0.0012564081698656082, 'timing/model': 1.0770637122914195, 'learning_rate/qwen_vl_interface': 5.02995082142166e-07, 'learning_rate/action_model': 5.029950821421659e-06, 'epoch': 2.66}) +[2026-06-18 09:07:46,648][starVLA.training.train_starvla][INFO] - Step 6923, Loss: {'train/loss': 0.16786539554595947, 'train/grad_norm_pre_clip': 14.014336585998535, 'timing/data': 0.00115138478577137, 'timing/model': 1.1817508824169636, 'learning_rate/qwen_vl_interface': 5.029187852680869e-07, 'learning_rate/action_model': 5.029187852680868e-06, 'epoch': 2.66}) +[2026-06-18 09:07:51,226][starVLA.training.train_starvla][INFO] - Step 6924, Loss: {'train/loss': 0.13528181612491608, 'train/grad_norm_pre_clip': 16.759180068969727, 'timing/data': 0.001535799354314804, 'timing/model': 1.1341628301888704, 'learning_rate/qwen_vl_interface': 5.028434724690521e-07, 'learning_rate/action_model': 5.028434724690521e-06, 'epoch': 2.66}) +[2026-06-18 09:07:56,157][starVLA.training.train_starvla][INFO] - Step 6925, Loss: {'train/loss': 0.5169638991355896, 'train/grad_norm_pre_clip': 22.149850845336914, 'timing/data': 0.0010412214323878288, 'timing/model': 1.2711967639625072, 'learning_rate/qwen_vl_interface': 5.027691437606734e-07, 'learning_rate/action_model': 5.027691437606733e-06, 'epoch': 2.66}) +[2026-06-18 09:08:01,052][starVLA.training.train_starvla][INFO] - Step 6926, Loss: {'train/loss': 0.5145572423934937, 'train/grad_norm_pre_clip': 18.286062240600586, 'timing/data': 0.0011433577165007591, 'timing/model': 1.4771394552662969, 'learning_rate/qwen_vl_interface': 5.026957991583597e-07, 'learning_rate/action_model': 5.0269579915835965e-06, 'epoch': 2.66}) +[2026-06-18 09:08:06,029][starVLA.training.train_starvla][INFO] - Step 6927, Loss: {'train/loss': 0.3994898200035095, 'train/grad_norm_pre_clip': 15.545845031738281, 'timing/data': 0.0014694146811962128, 'timing/model': 1.5562009140849113, 'learning_rate/qwen_vl_interface': 5.026234386773145e-07, 'learning_rate/action_model': 5.026234386773145e-06, 'epoch': 2.66}) +[2026-06-18 09:08:10,342][starVLA.training.train_starvla][INFO] - Step 6928, Loss: {'train/loss': 0.5156589150428772, 'train/grad_norm_pre_clip': 24.200563430786133, 'timing/data': 0.0014143334701657295, 'timing/model': 1.0973170287907124, 'learning_rate/qwen_vl_interface': 5.025520623325398e-07, 'learning_rate/action_model': 5.025520623325397e-06, 'epoch': 2.67}) +[2026-06-18 09:08:15,113][starVLA.training.train_starvla][INFO] - Step 6929, Loss: {'train/loss': 0.20162007212638855, 'train/grad_norm_pre_clip': 11.232100486755371, 'timing/data': 0.0019180551171302795, 'timing/model': 1.1001016274094582, 'learning_rate/qwen_vl_interface': 5.024816701388302e-07, 'learning_rate/action_model': 5.024816701388302e-06, 'epoch': 2.67}) +[2026-06-18 09:08:19,853][starVLA.training.train_starvla][INFO] - Step 6930, Loss: {'train/loss': 0.10480913519859314, 'train/grad_norm_pre_clip': 10.586697578430176, 'timing/data': 0.0067258477210998535, 'timing/model': 1.200976655818522, 'learning_rate/qwen_vl_interface': 5.024122621107795e-07, 'learning_rate/action_model': 5.024122621107795e-06, 'epoch': 2.67}) +[2026-06-18 09:08:25,020][starVLA.training.train_starvla][INFO] - Step 6931, Loss: {'train/loss': 0.3929588496685028, 'train/grad_norm_pre_clip': 12.651790618896484, 'timing/data': 0.0012299595400691032, 'timing/model': 1.137254238128662, 'learning_rate/qwen_vl_interface': 5.023438382627749e-07, 'learning_rate/action_model': 5.0234383826277486e-06, 'epoch': 2.67}) +[2026-06-18 09:08:30,056][starVLA.training.train_starvla][INFO] - Step 6932, Loss: {'train/loss': 0.09021027386188507, 'train/grad_norm_pre_clip': 8.77150821685791, 'timing/data': 0.0009528147056698799, 'timing/model': 1.752546307630837, 'learning_rate/qwen_vl_interface': 5.022763986090018e-07, 'learning_rate/action_model': 5.022763986090018e-06, 'epoch': 2.67}) +[2026-06-18 09:08:35,152][starVLA.training.train_starvla][INFO] - Step 6933, Loss: {'train/loss': 0.24298006296157837, 'train/grad_norm_pre_clip': 20.984291076660156, 'timing/data': 0.001296517439186573, 'timing/model': 1.3638481078669429, 'learning_rate/qwen_vl_interface': 5.022099431634398e-07, 'learning_rate/action_model': 5.022099431634398e-06, 'epoch': 2.67}) +[2026-06-18 09:08:40,071][starVLA.training.train_starvla][INFO] - Step 6934, Loss: {'train/loss': 0.3381960690021515, 'train/grad_norm_pre_clip': 15.162801742553711, 'timing/data': 0.001456553116440773, 'timing/model': 1.265843266621232, 'learning_rate/qwen_vl_interface': 5.021444719398656e-07, 'learning_rate/action_model': 5.021444719398656e-06, 'epoch': 2.67}) +[2026-06-18 09:08:45,030][starVLA.training.train_starvla][INFO] - Step 6935, Loss: {'train/loss': 0.14335587620735168, 'train/grad_norm_pre_clip': 12.672701835632324, 'timing/data': 0.0011488143354654312, 'timing/model': 1.1024882849305868, 'learning_rate/qwen_vl_interface': 5.020799849518513e-07, 'learning_rate/action_model': 5.020799849518513e-06, 'epoch': 2.67}) +[2026-06-18 09:08:50,152][starVLA.training.train_starvla][INFO] - Step 6936, Loss: {'train/loss': 0.3967253267765045, 'train/grad_norm_pre_clip': 21.85038185119629, 'timing/data': 0.0013204561546444893, 'timing/model': 1.3591315438970923, 'learning_rate/qwen_vl_interface': 5.020164822127652e-07, 'learning_rate/action_model': 5.020164822127652e-06, 'epoch': 2.67}) +[2026-06-18 09:08:54,780][starVLA.training.train_starvla][INFO] - Step 6937, Loss: {'train/loss': 0.2042146474123001, 'train/grad_norm_pre_clip': 15.518986701965332, 'timing/data': 0.0012534167617559433, 'timing/model': 1.1917059002444148, 'learning_rate/qwen_vl_interface': 5.019539637357715e-07, 'learning_rate/action_model': 5.019539637357714e-06, 'epoch': 2.67}) +[2026-06-18 09:08:59,846][starVLA.training.train_starvla][INFO] - Step 6938, Loss: {'train/loss': 0.24251818656921387, 'train/grad_norm_pre_clip': 18.329689025878906, 'timing/data': 0.0012326762080192566, 'timing/model': 1.2296848632395267, 'learning_rate/qwen_vl_interface': 5.0189242953383e-07, 'learning_rate/action_model': 5.018924295338299e-06, 'epoch': 2.67}) +[2026-06-18 09:09:04,730][starVLA.training.train_starvla][INFO] - Step 6939, Loss: {'train/loss': 0.35368433594703674, 'train/grad_norm_pre_clip': 25.58412742614746, 'timing/data': 0.001589217223227024, 'timing/model': 1.1151368375867605, 'learning_rate/qwen_vl_interface': 5.018318796196967e-07, 'learning_rate/action_model': 5.018318796196967e-06, 'epoch': 2.67}) +[2026-06-18 09:09:09,922][starVLA.training.train_starvla][INFO] - Step 6940, Loss: {'train/loss': 0.24798166751861572, 'train/grad_norm_pre_clip': 17.721574783325195, 'timing/data': 0.0010186964645981789, 'timing/model': 1.1059203073382378, 'learning_rate/qwen_vl_interface': 5.017723140059243e-07, 'learning_rate/action_model': 5.017723140059242e-06, 'epoch': 2.67}) +[2026-06-18 09:09:15,033][starVLA.training.train_starvla][INFO] - Step 6941, Loss: {'train/loss': 0.1667354553937912, 'train/grad_norm_pre_clip': 17.248424530029297, 'timing/data': 0.0019481265917420387, 'timing/model': 1.5525370668619871, 'learning_rate/qwen_vl_interface': 5.017137327048606e-07, 'learning_rate/action_model': 5.017137327048606e-06, 'epoch': 2.67}) +[2026-06-18 09:09:20,047][starVLA.training.train_starvla][INFO] - Step 6942, Loss: {'train/loss': 0.1761891096830368, 'train/grad_norm_pre_clip': 14.203255653381348, 'timing/data': 0.0011617811396718025, 'timing/model': 1.6590211456641555, 'learning_rate/qwen_vl_interface': 5.016561357286498e-07, 'learning_rate/action_model': 5.016561357286498e-06, 'epoch': 2.67}) +[2026-06-18 09:09:24,382][starVLA.training.train_starvla][INFO] - Step 6943, Loss: {'train/loss': 0.1730242669582367, 'train/grad_norm_pre_clip': 14.701159477233887, 'timing/data': 0.0011405162513256073, 'timing/model': 1.1261408869177103, 'learning_rate/qwen_vl_interface': 5.015995230892312e-07, 'learning_rate/action_model': 5.015995230892311e-06, 'epoch': 2.67}) +[2026-06-18 09:09:29,360][starVLA.training.train_starvla][INFO] - Step 6944, Loss: {'train/loss': 0.12298238277435303, 'train/grad_norm_pre_clip': 11.784560203552246, 'timing/data': 0.0014656232669949532, 'timing/model': 1.1133267162367702, 'learning_rate/qwen_vl_interface': 5.015438947983409e-07, 'learning_rate/action_model': 5.015438947983409e-06, 'epoch': 2.67}) +[2026-06-18 09:09:34,090][starVLA.training.train_starvla][INFO] - Step 6945, Loss: {'train/loss': 0.4153238832950592, 'train/grad_norm_pre_clip': 16.5580997467041, 'timing/data': 0.0010210899636149406, 'timing/model': 1.1323802433907986, 'learning_rate/qwen_vl_interface': 5.014892508675108e-07, 'learning_rate/action_model': 5.014892508675107e-06, 'epoch': 2.67}) +[2026-06-18 09:09:39,294][starVLA.training.train_starvla][INFO] - Step 6946, Loss: {'train/loss': 0.11102326214313507, 'train/grad_norm_pre_clip': 7.602924346923828, 'timing/data': 0.0012418357655405998, 'timing/model': 1.1133083533495665, 'learning_rate/qwen_vl_interface': 5.014355913080683e-07, 'learning_rate/action_model': 5.014355913080683e-06, 'epoch': 2.67}) +[2026-06-18 09:09:43,685][starVLA.training.train_starvla][INFO] - Step 6947, Loss: {'train/loss': 0.37665048241615295, 'train/grad_norm_pre_clip': 19.460729598999023, 'timing/data': 0.0016377093270421028, 'timing/model': 1.1379929864779115, 'learning_rate/qwen_vl_interface': 5.013829161311377e-07, 'learning_rate/action_model': 5.013829161311377e-06, 'epoch': 2.67}) +[2026-06-18 09:09:48,619][starVLA.training.train_starvla][INFO] - Step 6948, Loss: {'train/loss': 0.07270202040672302, 'train/grad_norm_pre_clip': 7.519294738769531, 'timing/data': 0.0010539889335632324, 'timing/model': 1.0982003258541226, 'learning_rate/qwen_vl_interface': 5.013312253476383e-07, 'learning_rate/action_model': 5.0133122534763826e-06, 'epoch': 2.67}) +[2026-06-18 09:09:53,651][starVLA.training.train_starvla][INFO] - Step 6949, Loss: {'train/loss': 0.22636373341083527, 'train/grad_norm_pre_clip': 12.175924301147461, 'timing/data': 0.0013136407360434532, 'timing/model': 1.0948917483910918, 'learning_rate/qwen_vl_interface': 5.012805189682861e-07, 'learning_rate/action_model': 5.01280518968286e-06, 'epoch': 2.67}) +[2026-06-18 09:09:58,517][starVLA.training.train_starvla][INFO] - Step 6950, Loss: {'train/loss': 0.7791870832443237, 'train/grad_norm_pre_clip': 31.420867919921875, 'timing/data': 0.001241704449057579, 'timing/model': 1.0879180328920484, 'learning_rate/qwen_vl_interface': 5.012307970035915e-07, 'learning_rate/action_model': 5.012307970035915e-06, 'epoch': 2.67}) +[2026-06-18 09:10:03,191][starVLA.training.train_starvla][INFO] - Step 6951, Loss: {'train/loss': 0.4273601472377777, 'train/grad_norm_pre_clip': 19.567020416259766, 'timing/data': 0.0011951504275202751, 'timing/model': 1.1275746831670403, 'learning_rate/qwen_vl_interface': 5.011820594638626e-07, 'learning_rate/action_model': 5.0118205946386256e-06, 'epoch': 2.67}) +[2026-06-18 09:10:08,016][starVLA.training.train_starvla][INFO] - Step 6952, Loss: {'train/loss': 0.2682296931743622, 'train/grad_norm_pre_clip': 19.17351531982422, 'timing/data': 0.0013253837823867798, 'timing/model': 1.4853711817413568, 'learning_rate/qwen_vl_interface': 5.011343063592032e-07, 'learning_rate/action_model': 5.011343063592032e-06, 'epoch': 2.67}) +[2026-06-18 09:10:12,662][starVLA.training.train_starvla][INFO] - Step 6953, Loss: {'train/loss': 0.31689509749412537, 'train/grad_norm_pre_clip': 15.448567390441895, 'timing/data': 0.0012480933219194412, 'timing/model': 1.1669866535812616, 'learning_rate/qwen_vl_interface': 5.010875376995121e-07, 'learning_rate/action_model': 5.01087537699512e-06, 'epoch': 2.67}) +[2026-06-18 09:10:17,396][starVLA.training.train_starvla][INFO] - Step 6954, Loss: {'train/loss': 0.48673662543296814, 'train/grad_norm_pre_clip': 20.28789710998535, 'timing/data': 0.0011290526017546654, 'timing/model': 1.0809360323473811, 'learning_rate/qwen_vl_interface': 5.010417534944839e-07, 'learning_rate/action_model': 5.010417534944838e-06, 'epoch': 2.68}) +[2026-06-18 09:10:22,203][starVLA.training.train_starvla][INFO] - Step 6955, Loss: {'train/loss': 0.11784851551055908, 'train/grad_norm_pre_clip': 10.046064376831055, 'timing/data': 0.0011787563562393188, 'timing/model': 1.081772598437965, 'learning_rate/qwen_vl_interface': 5.009969537536106e-07, 'learning_rate/action_model': 5.009969537536106e-06, 'epoch': 2.68}) +[2026-06-18 09:10:26,984][starVLA.training.train_starvla][INFO] - Step 6956, Loss: {'train/loss': 0.23393703997135162, 'train/grad_norm_pre_clip': 20.379703521728516, 'timing/data': 0.0011450322344899178, 'timing/model': 1.099845633842051, 'learning_rate/qwen_vl_interface': 5.00953138486179e-07, 'learning_rate/action_model': 5.00953138486179e-06, 'epoch': 2.68}) +[2026-06-18 09:10:31,813][starVLA.training.train_starvla][INFO] - Step 6957, Loss: {'train/loss': 0.3441527783870697, 'train/grad_norm_pre_clip': 16.97734260559082, 'timing/data': 0.0014236457645893097, 'timing/model': 1.4904687833040953, 'learning_rate/qwen_vl_interface': 5.009103077012723e-07, 'learning_rate/action_model': 5.009103077012723e-06, 'epoch': 2.68}) +[2026-06-18 09:10:36,346][starVLA.training.train_starvla][INFO] - Step 6958, Loss: {'train/loss': 0.21525262296199799, 'train/grad_norm_pre_clip': 14.806978225708008, 'timing/data': 0.0012159282341599464, 'timing/model': 1.2282337620854378, 'learning_rate/qwen_vl_interface': 5.008684614077684e-07, 'learning_rate/action_model': 5.008684614077684e-06, 'epoch': 2.68}) +[2026-06-18 09:10:41,272][starVLA.training.train_starvla][INFO] - Step 6959, Loss: {'train/loss': 0.2956320345401764, 'train/grad_norm_pre_clip': 15.614094734191895, 'timing/data': 0.0012082336470484734, 'timing/model': 1.1127971019595861, 'learning_rate/qwen_vl_interface': 5.008275996143429e-07, 'learning_rate/action_model': 5.0082759961434295e-06, 'epoch': 2.68}) +[2026-06-18 09:10:46,085][starVLA.training.train_starvla][INFO] - Step 6960, Loss: {'train/loss': 0.2526531219482422, 'train/grad_norm_pre_clip': 15.103998184204102, 'timing/data': 0.0013988027349114418, 'timing/model': 1.101472396403551, 'learning_rate/qwen_vl_interface': 5.007877223294668e-07, 'learning_rate/action_model': 5.0078772232946675e-06, 'epoch': 2.68}) +[2026-06-18 09:10:50,934][starVLA.training.train_starvla][INFO] - Step 6961, Loss: {'train/loss': 0.344027578830719, 'train/grad_norm_pre_clip': 12.943768501281738, 'timing/data': 0.0008898572996258736, 'timing/model': 1.1195404957979918, 'learning_rate/qwen_vl_interface': 5.007488295614058e-07, 'learning_rate/action_model': 5.007488295614058e-06, 'epoch': 2.68}) +[2026-06-18 09:10:55,630][starVLA.training.train_starvla][INFO] - Step 6962, Loss: {'train/loss': 0.46145322918891907, 'train/grad_norm_pre_clip': 17.038253784179688, 'timing/data': 0.001089523546397686, 'timing/model': 1.579499171115458, 'learning_rate/qwen_vl_interface': 5.007109213182231e-07, 'learning_rate/action_model': 5.007109213182231e-06, 'epoch': 2.68}) +[2026-06-18 09:10:59,990][starVLA.training.train_starvla][INFO] - Step 6963, Loss: {'train/loss': 0.2036966234445572, 'train/grad_norm_pre_clip': 13.305046081542969, 'timing/data': 0.0012032808735966682, 'timing/model': 1.112397094257176, 'learning_rate/qwen_vl_interface': 5.006739976077769e-07, 'learning_rate/action_model': 5.0067399760777695e-06, 'epoch': 2.68}) +[2026-06-18 09:11:05,429][starVLA.training.train_starvla][INFO] - Step 6964, Loss: {'train/loss': 0.2278253734111786, 'train/grad_norm_pre_clip': 15.125858306884766, 'timing/data': 0.0017672833055257797, 'timing/model': 1.2861351715400815, 'learning_rate/qwen_vl_interface': 5.006380584377213e-07, 'learning_rate/action_model': 5.006380584377213e-06, 'epoch': 2.68}) +[2026-06-18 09:11:10,511][starVLA.training.train_starvla][INFO] - Step 6965, Loss: {'train/loss': 0.19717328250408173, 'train/grad_norm_pre_clip': 10.807580947875977, 'timing/data': 0.0012006647884845734, 'timing/model': 1.1173364985734224, 'learning_rate/qwen_vl_interface': 5.006031038155068e-07, 'learning_rate/action_model': 5.006031038155067e-06, 'epoch': 2.68}) +[2026-06-18 09:11:15,161][starVLA.training.train_starvla][INFO] - Step 6966, Loss: {'train/loss': 0.38876426219940186, 'train/grad_norm_pre_clip': 26.02126693725586, 'timing/data': 0.0014103297144174576, 'timing/model': 1.1099729733541608, 'learning_rate/qwen_vl_interface': 5.00569133748379e-07, 'learning_rate/action_model': 5.00569133748379e-06, 'epoch': 2.68}) +[2026-06-18 09:11:19,942][starVLA.training.train_starvla][INFO] - Step 6967, Loss: {'train/loss': 0.33834415674209595, 'train/grad_norm_pre_clip': 22.251461029052734, 'timing/data': 0.008020007982850075, 'timing/model': 1.5303880982100964, 'learning_rate/qwen_vl_interface': 5.005361482433811e-07, 'learning_rate/action_model': 5.005361482433811e-06, 'epoch': 2.68}) +[2026-06-18 09:11:24,315][starVLA.training.train_starvla][INFO] - Step 6968, Loss: {'train/loss': 0.3998570144176483, 'train/grad_norm_pre_clip': 25.36393928527832, 'timing/data': 0.0011584237217903137, 'timing/model': 1.1353488452732563, 'learning_rate/qwen_vl_interface': 5.005041473073501e-07, 'learning_rate/action_model': 5.005041473073501e-06, 'epoch': 2.68}) +[2026-06-18 09:11:29,250][starVLA.training.train_starvla][INFO] - Step 6969, Loss: {'train/loss': 0.3312393128871918, 'train/grad_norm_pre_clip': 19.94259262084961, 'timing/data': 0.001617254689335823, 'timing/model': 1.1223374046385288, 'learning_rate/qwen_vl_interface': 5.004731309469201e-07, 'learning_rate/action_model': 5.004731309469201e-06, 'epoch': 2.68}) +[2026-06-18 09:11:33,885][starVLA.training.train_starvla][INFO] - Step 6970, Loss: {'train/loss': 0.49618491530418396, 'train/grad_norm_pre_clip': 25.464580535888672, 'timing/data': 0.0018561594188213348, 'timing/model': 1.1078894473612309, 'learning_rate/qwen_vl_interface': 5.004430991685201e-07, 'learning_rate/action_model': 5.0044309916852016e-06, 'epoch': 2.68}) +[2026-06-18 09:11:38,919][starVLA.training.train_starvla][INFO] - Step 6971, Loss: {'train/loss': 0.25810933113098145, 'train/grad_norm_pre_clip': 13.788557052612305, 'timing/data': 0.00135823804885149, 'timing/model': 1.27453035954386, 'learning_rate/qwen_vl_interface': 5.004140519783773e-07, 'learning_rate/action_model': 5.0041405197837725e-06, 'epoch': 2.68}) +[2026-06-18 09:11:43,529][starVLA.training.train_starvla][INFO] - Step 6972, Loss: {'train/loss': 0.08736130595207214, 'train/grad_norm_pre_clip': 7.541001796722412, 'timing/data': 0.0019060363993048668, 'timing/model': 1.1111280443146825, 'learning_rate/qwen_vl_interface': 5.003859893825117e-07, 'learning_rate/action_model': 5.003859893825117e-06, 'epoch': 2.68}) +[2026-06-18 09:11:48,357][starVLA.training.train_starvla][INFO] - Step 6973, Loss: {'train/loss': 0.11953595280647278, 'train/grad_norm_pre_clip': 11.562633514404297, 'timing/data': 0.0012469198554754257, 'timing/model': 1.1295704720541835, 'learning_rate/qwen_vl_interface': 5.003589113867419e-07, 'learning_rate/action_model': 5.003589113867419e-06, 'epoch': 2.68}) +[2026-06-18 09:11:53,180][starVLA.training.train_starvla][INFO] - Step 6974, Loss: {'train/loss': 0.20353803038597107, 'train/grad_norm_pre_clip': 10.608264923095703, 'timing/data': 0.0013037929311394691, 'timing/model': 1.168861079029739, 'learning_rate/qwen_vl_interface': 5.003328179966799e-07, 'learning_rate/action_model': 5.003328179966798e-06, 'epoch': 2.68}) +[2026-06-18 09:11:58,113][starVLA.training.train_starvla][INFO] - Step 6975, Loss: {'train/loss': 0.48618215322494507, 'train/grad_norm_pre_clip': 25.861024856567383, 'timing/data': 0.0018796203657984734, 'timing/model': 1.1305937319993973, 'learning_rate/qwen_vl_interface': 5.003077092177363e-07, 'learning_rate/action_model': 5.003077092177363e-06, 'epoch': 2.68}) +[2026-06-18 09:12:02,976][starVLA.training.train_starvla][INFO] - Step 6976, Loss: {'train/loss': 0.21801146864891052, 'train/grad_norm_pre_clip': 19.56922721862793, 'timing/data': 0.0012419866397976875, 'timing/model': 1.5602200543507934, 'learning_rate/qwen_vl_interface': 5.002835850551146e-07, 'learning_rate/action_model': 5.002835850551146e-06, 'epoch': 2.68}) +[2026-06-18 09:12:08,125][starVLA.training.train_starvla][INFO] - Step 6977, Loss: {'train/loss': 0.3287433087825775, 'train/grad_norm_pre_clip': 17.218687057495117, 'timing/data': 0.0016228631138801575, 'timing/model': 1.568075766786933, 'learning_rate/qwen_vl_interface': 5.002604455138174e-07, 'learning_rate/action_model': 5.002604455138173e-06, 'epoch': 2.68}) +[2026-06-18 09:12:12,568][starVLA.training.train_starvla][INFO] - Step 6978, Loss: {'train/loss': 0.3331686854362488, 'train/grad_norm_pre_clip': 18.148513793945312, 'timing/data': 0.0023329928517341614, 'timing/model': 1.1163289528340101, 'learning_rate/qwen_vl_interface': 5.002382905986407e-07, 'learning_rate/action_model': 5.002382905986407e-06, 'epoch': 2.68}) +[2026-06-18 09:12:17,249][starVLA.training.train_starvla][INFO] - Step 6979, Loss: {'train/loss': 0.1510581225156784, 'train/grad_norm_pre_clip': 10.038782119750977, 'timing/data': 0.0010868851095438004, 'timing/model': 1.100702190771699, 'learning_rate/qwen_vl_interface': 5.002171203141771e-07, 'learning_rate/action_model': 5.0021712031417704e-06, 'epoch': 2.68}) +[2026-06-18 09:12:21,995][starVLA.training.train_starvla][INFO] - Step 6980, Loss: {'train/loss': 0.1216232031583786, 'train/grad_norm_pre_clip': 9.954169273376465, 'timing/data': 0.0010453620925545692, 'timing/model': 1.1989639606326818, 'learning_rate/qwen_vl_interface': 5.001969346648154e-07, 'learning_rate/action_model': 5.001969346648154e-06, 'epoch': 2.69}) +[2026-06-18 09:12:26,996][starVLA.training.train_starvla][INFO] - Step 6981, Loss: {'train/loss': 0.34118136763572693, 'train/grad_norm_pre_clip': 16.163503646850586, 'timing/data': 0.0015574097633361816, 'timing/model': 1.1486829994246364, 'learning_rate/qwen_vl_interface': 5.0017773365474e-07, 'learning_rate/action_model': 5.001777336547399e-06, 'epoch': 2.69}) +[2026-06-18 09:12:32,065][starVLA.training.train_starvla][INFO] - Step 6982, Loss: {'train/loss': 0.3993474245071411, 'train/grad_norm_pre_clip': 21.994600296020508, 'timing/data': 0.0011858455836772919, 'timing/model': 1.673276654444635, 'learning_rate/qwen_vl_interface': 5.001595172879317e-07, 'learning_rate/action_model': 5.001595172879316e-06, 'epoch': 2.69}) +[2026-06-18 09:12:36,526][starVLA.training.train_starvla][INFO] - Step 6983, Loss: {'train/loss': 0.44379085302352905, 'train/grad_norm_pre_clip': 24.63362693786621, 'timing/data': 0.0014444990083575249, 'timing/model': 1.0849316362291574, 'learning_rate/qwen_vl_interface': 5.001422855681669e-07, 'learning_rate/action_model': 5.001422855681669e-06, 'epoch': 2.69}) +[2026-06-18 09:12:41,398][starVLA.training.train_starvla][INFO] - Step 6984, Loss: {'train/loss': 0.3793107867240906, 'train/grad_norm_pre_clip': 18.642478942871094, 'timing/data': 0.0012591741979122162, 'timing/model': 1.1135735092684627, 'learning_rate/qwen_vl_interface': 5.001260384990169e-07, 'learning_rate/action_model': 5.001260384990169e-06, 'epoch': 2.69}) +[2026-06-18 09:12:46,134][starVLA.training.train_starvla][INFO] - Step 6985, Loss: {'train/loss': 0.1314917802810669, 'train/grad_norm_pre_clip': 10.430039405822754, 'timing/data': 0.0014861617237329483, 'timing/model': 1.1166665069758892, 'learning_rate/qwen_vl_interface': 5.001107760838504e-07, 'learning_rate/action_model': 5.001107760838504e-06, 'epoch': 2.69}) +[2026-06-18 09:12:51,062][starVLA.training.train_starvla][INFO] - Step 6986, Loss: {'train/loss': 0.2202455997467041, 'train/grad_norm_pre_clip': 15.42418384552002, 'timing/data': 0.0020601367577910423, 'timing/model': 1.0983279971405864, 'learning_rate/qwen_vl_interface': 5.000964983258306e-07, 'learning_rate/action_model': 5.000964983258306e-06, 'epoch': 2.69}) +[2026-06-18 09:12:55,979][starVLA.training.train_starvla][INFO] - Step 6987, Loss: {'train/loss': 0.2738795876502991, 'train/grad_norm_pre_clip': 13.147989273071289, 'timing/data': 0.001235191710293293, 'timing/model': 1.4395540403202176, 'learning_rate/qwen_vl_interface': 5.000832052279185e-07, 'learning_rate/action_model': 5.000832052279185e-06, 'epoch': 2.69}) +[2026-06-18 09:13:00,246][starVLA.training.train_starvla][INFO] - Step 6988, Loss: {'train/loss': 0.17845913767814636, 'train/grad_norm_pre_clip': 11.658912658691406, 'timing/data': 0.0010122554376721382, 'timing/model': 1.0868798149749637, 'learning_rate/qwen_vl_interface': 5.000708967928686e-07, 'learning_rate/action_model': 5.000708967928685e-06, 'epoch': 2.69}) +[2026-06-18 09:13:05,031][starVLA.training.train_starvla][INFO] - Step 6989, Loss: {'train/loss': 0.4275328814983368, 'train/grad_norm_pre_clip': 20.77572250366211, 'timing/data': 0.0048158979043364525, 'timing/model': 1.1727319620549679, 'learning_rate/qwen_vl_interface': 5.00059573023233e-07, 'learning_rate/action_model': 5.0005957302323305e-06, 'epoch': 2.69}) +[2026-06-18 09:13:09,951][starVLA.training.train_starvla][INFO] - Step 6990, Loss: {'train/loss': 0.3685551881790161, 'train/grad_norm_pre_clip': 18.303478240966797, 'timing/data': 0.001541266217827797, 'timing/model': 1.1325091067701578, 'learning_rate/qwen_vl_interface': 5.000492339213594e-07, 'learning_rate/action_model': 5.000492339213594e-06, 'epoch': 2.69}) +[2026-06-18 09:13:15,252][starVLA.training.train_starvla][INFO] - Step 6991, Loss: {'train/loss': 0.4172786474227905, 'train/grad_norm_pre_clip': 17.613712310791016, 'timing/data': 0.0012003341689705849, 'timing/model': 1.364673776552081, 'learning_rate/qwen_vl_interface': 5.000398794893907e-07, 'learning_rate/action_model': 5.000398794893906e-06, 'epoch': 2.69}) +[2026-06-18 09:13:20,030][starVLA.training.train_starvla][INFO] - Step 6992, Loss: {'train/loss': 0.12597811222076416, 'train/grad_norm_pre_clip': 12.369794845581055, 'timing/data': 0.0013352921232581139, 'timing/model': 1.5710020354017615, 'learning_rate/qwen_vl_interface': 5.000315097292659e-07, 'learning_rate/action_model': 5.000315097292659e-06, 'epoch': 2.69}) +[2026-06-18 09:13:24,409][starVLA.training.train_starvla][INFO] - Step 6993, Loss: {'train/loss': 0.13776978850364685, 'train/grad_norm_pre_clip': 10.526579856872559, 'timing/data': 0.0009944383054971695, 'timing/model': 1.1318858582526445, 'learning_rate/qwen_vl_interface': 5.000241246427207e-07, 'learning_rate/action_model': 5.000241246427207e-06, 'epoch': 2.69}) +[2026-06-18 09:13:29,122][starVLA.training.train_starvla][INFO] - Step 6994, Loss: {'train/loss': 0.3732002079486847, 'train/grad_norm_pre_clip': 22.62209129333496, 'timing/data': 0.001155761070549488, 'timing/model': 1.1411540815606713, 'learning_rate/qwen_vl_interface': 5.000177242312855e-07, 'learning_rate/action_model': 5.000177242312855e-06, 'epoch': 2.69}) +[2026-06-18 09:13:34,204][starVLA.training.train_starvla][INFO] - Step 6995, Loss: {'train/loss': 0.14496172964572906, 'train/grad_norm_pre_clip': 45.89853286743164, 'timing/data': 0.0011973483487963676, 'timing/model': 1.3169431649148464, 'learning_rate/qwen_vl_interface': 5.00012308496287e-07, 'learning_rate/action_model': 5.000123084962871e-06, 'epoch': 2.69}) +[2026-06-18 09:13:39,276][starVLA.training.train_starvla][INFO] - Step 6996, Loss: {'train/loss': 0.16488969326019287, 'train/grad_norm_pre_clip': 11.949756622314453, 'timing/data': 0.001693987287580967, 'timing/model': 1.293151818215847, 'learning_rate/qwen_vl_interface': 5.000078774388487e-07, 'learning_rate/action_model': 5.000078774388487e-06, 'epoch': 2.69}) +[2026-06-18 09:13:43,731][starVLA.training.train_starvla][INFO] - Step 6997, Loss: {'train/loss': 0.16960826516151428, 'train/grad_norm_pre_clip': 11.771444320678711, 'timing/data': 0.001154816709458828, 'timing/model': 1.097850939258933, 'learning_rate/qwen_vl_interface': 5.000044310598879e-07, 'learning_rate/action_model': 5.00004431059888e-06, 'epoch': 2.69}) +[2026-06-18 09:13:48,663][starVLA.training.train_starvla][INFO] - Step 6998, Loss: {'train/loss': 0.2483251988887787, 'train/grad_norm_pre_clip': 16.20953369140625, 'timing/data': 0.0009847339242696762, 'timing/model': 1.2700571864843369, 'learning_rate/qwen_vl_interface': 5.000019693601204e-07, 'learning_rate/action_model': 5.000019693601204e-06, 'epoch': 2.69}) +[2026-06-18 09:13:53,643][starVLA.training.train_starvla][INFO] - Step 6999, Loss: {'train/loss': 0.5030854940414429, 'train/grad_norm_pre_clip': 19.154529571533203, 'timing/data': 0.001141163520514965, 'timing/model': 1.1576565559953451, 'learning_rate/qwen_vl_interface': 5.000004923400556e-07, 'learning_rate/action_model': 5.0000049234005565e-06, 'epoch': 2.69}) +[2026-06-18 09:16:44,511][starVLA.training.train_starvla][INFO] - Step 7000, Loss: {'train/loss': 0.14866994321346283, 'train/grad_norm_pre_clip': 10.291830062866211, 'eval/loss': 0.4332925081253052, 'eval/latency_4/loss': 0.4332925081253052, 'eval/demon_fire_precision': 0.969220904135941, 'eval/demon_fire_recall': 0.9818122767132186, 'eval/demon_fire_f1': 0.9754759599870927, 'eval/demon_fire_rate_model': 0.9160058737151248, 'eval/demon_fire_rate_teacher': 0.9042584434654919, 'eval/demon_left_precision': 0.8, 'eval/demon_left_recall': 0.7123287671232876, 'eval/demon_left_f1': 0.7536231884057971, 'eval/demon_left_rate_model': 0.1145374449339207, 'eval/demon_left_rate_teacher': 0.12863436123348018, 'eval/demon_right_precision': 0.8097928436911488, 'eval/demon_right_recall': 0.776173285198556, 'eval/demon_right_f1': 0.7926267281105991, 'eval/demon_right_rate_model': 0.15594713656387665, 'eval/demon_right_rate_teacher': 0.16270190895741557, 'eval/demon_move_f1': 0.773124958258198, 'eval/demon_cc_f1': 0.9044584349255569, 'eval/latency_4/demon_fire_precision': 0.969220904135941, 'eval/latency_4/demon_fire_recall': 0.9818122767132186, 'eval/latency_4/demon_fire_f1': 0.9754759599870927, 'eval/latency_4/demon_fire_rate_model': 0.9160058737151248, 'eval/latency_4/demon_fire_rate_teacher': 0.9042584434654919, 'eval/latency_4/demon_left_precision': 0.8, 'eval/latency_4/demon_left_recall': 0.7123287671232876, 'eval/latency_4/demon_left_f1': 0.7536231884057971, 'eval/latency_4/demon_left_rate_model': 0.1145374449339207, 'eval/latency_4/demon_left_rate_teacher': 0.12863436123348018, 'eval/latency_4/demon_right_precision': 0.8097928436911488, 'eval/latency_4/demon_right_recall': 0.776173285198556, 'eval/latency_4/demon_right_f1': 0.7926267281105991, 'eval/latency_4/demon_right_rate_model': 0.15594713656387665, 'eval/latency_4/demon_right_rate_teacher': 0.16270190895741557, 'eval/latency_4/demon_move_f1': 0.773124958258198, 'eval/latency_4/demon_cc_f1': 0.9044584349255569, 'timing/data': 0.0010414337739348412, 'timing/model': 1.1099280258640647, 'learning_rate/qwen_vl_interface': 5e-07, 'learning_rate/action_model': 4.9999999999999996e-06, 'epoch': 2.69}) +[2026-06-18 09:16:44,526][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_7000_state +[2026-06-18 09:16:44,534][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 09:16:53,877][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_7000_state/model.safetensors +[2026-06-18 09:17:17,748][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_7000_state/optimizer.bin +[2026-06-18 09:17:17,749][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_7000_state/sampler.bin +[2026-06-18 09:17:17,750][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_7000_state/sampler_1.bin +[2026-06-18 09:17:17,753][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints/steps_7000_state/random_states_0.pkl +[2026-06-18 09:17:17,768][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 7000 +[2026-06-18 09:17:17,768][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 09:17:17,785][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 09:17:17,786][starVLA.training.train_starvla][INFO] - Training complete. Final training state is saved under /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_4_200ep/checkpoints +[2026-06-18 09:17:17,843][starVLA.training.train_starvla][INFO] - ... and that's all, folks! diff --git a/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_4_200ep/summary.jsonl b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_4_200ep/summary.jsonl new file mode 100644 index 0000000000000000000000000000000000000000..67f4020ac3871f475553f86f0c342379cb905502 --- /dev/null +++ b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_4_200ep/summary.jsonl @@ -0,0 +1,70 @@ +{"steps": 100} +{"steps": 200} +{"steps": 300} +{"steps": 400} +{"steps": 500} +{"steps": 600} +{"steps": 700} +{"steps": 800} +{"steps": 900} +{"steps": 1000} +{"steps": 1100} +{"steps": 1200} +{"steps": 1300} +{"steps": 1400} +{"steps": 1500} +{"steps": 1600} +{"steps": 1700} +{"steps": 1800} +{"steps": 1900} +{"steps": 2000} +{"steps": 2100} +{"steps": 2200} +{"steps": 2300} +{"steps": 2400} +{"steps": 2500} +{"steps": 2600} +{"steps": 2700} +{"steps": 2800} +{"steps": 2900} +{"steps": 3000} +{"steps": 3100} +{"steps": 3200} +{"steps": 3300} +{"steps": 3400} +{"steps": 3500} +{"steps": 3600} +{"steps": 3700} +{"steps": 3800} +{"steps": 3900} +{"steps": 4000} +{"steps": 4100} +{"steps": 4200} +{"steps": 4300} +{"steps": 4400} +{"steps": 4500} +{"steps": 4600} +{"steps": 4700} +{"steps": 4800} +{"steps": 4900} +{"steps": 5000} +{"steps": 5100} +{"steps": 5200} +{"steps": 5300} +{"steps": 5400} +{"steps": 5500} +{"steps": 5600} +{"steps": 5700} +{"steps": 5800} +{"steps": 5900} +{"steps": 6000} +{"steps": 6100} +{"steps": 6200} +{"steps": 6300} +{"steps": 6400} +{"steps": 6500} +{"steps": 6600} +{"steps": 6700} +{"steps": 6800} +{"steps": 6900} +{"steps": 7000} diff --git a/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_6_200ep/checkpoints/steps_7000_model.safetensors b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_6_200ep/checkpoints/steps_7000_model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..964783d2894384ce0773b4dcbc779199a3b58043 --- /dev/null +++ b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_6_200ep/checkpoints/steps_7000_model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:375ff9f24479e219e8197bbc2eee3e71e61e1853e0c80eefe438ca947d5c5b1b +size 9138230516 diff --git a/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_6_200ep/config.full.yaml b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_6_200ep/config.full.yaml new file mode 100644 index 0000000000000000000000000000000000000000..0dac30603410755b25d194b0c993ff4ee8cf035e --- /dev/null +++ b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_6_200ep/config.full.yaml @@ -0,0 +1,209 @@ +framework: + name: QwenOFT + qwenvl: + base_vlm: /workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + attn_implementation: flash_attention_2 + enable_gradient_checkpointing: true + action_model: + action_model_type: MLP + action_dim: 7 + action_hidden_dim: 2560 + future_action_window_size: 0 + past_action_window_size: 0 + loss_type: discrete_ce + state_dim: 7 + action_horizon: 1 + action_env_dim: 6 +datasets: + vla_data: + dataset_py: lerobot_datasets + include_state: true + data_root_dir: /workspace/starVLA/data/demon_attack_fix_latency_6_200ep + data_mix: demon_attack_train__bridge + eval_data_mix: demon_attack_train__bridge__val + custom_mixtures_path: null + action_type: discrete + sequential_step_sampling: false + shuffle: true + action_balance: + enabled: false + strategy: balanced_epoch + action_key: action_id + target_flap_fraction: 0.3 + noop_id: 0 + flap_id: 1 + latency_curriculum: + enabled: false + strategy: exclusive + latencies: null + phase_steps: null + per_device_batch_size: 32 + load_all_data_for_training: true + obs_image_size: null + video_backend: torchvision_av +dataset: + source_hf: '' + config_name: null + source_subdir: null + converted_name: demon_attack_train + single_source_hf: '' + mixed_source_hf: '' + single_converted_name: demon_attack_train + mixed_converted_name: demon_attack_mixed_latency_train + single_latency_filter: null + mixed_latency_filter: null + force_download: false + setup_force: false + skip_verification: false + verify_rows: 200 + max_episodes: null + episodes_per_latency: null + latency_filter: null + debug_subset: + enabled: false + max_episodes: 5 + suffix: debug +base_model: + repo_id: Qwen/Qwen3-VL-4B-Instruct +initialization: + checkpoint_local_dir: playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_hf_repo_id: StarVLA/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_filename: checkpoints/steps_5000_pytorch_model.pt +trainer: + max_train_steps: 7000 + num_warmup_steps: 100 + save_interval: 100 + eval_interval: 100 + eval_num_batches: 100 + per_latency_eval_num_batches: null + eval_action_classification: true + cc_f1_tolerance: 1 + learning_rate: + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + action_model: 0.0001 + lr_scheduler_type: cosine_with_min_lr + scheduler_specific_kwargs: + min_lr: 1.0e-06 + freeze_modules: '' + loss_scale: + vla: 1.0 + vlm: 0.1 + max_grad_norm: 1.0 + weight_decay: 0.0 + logging_frequency: 1 + gradient_clipping: 1.0 + gradient_accumulation_steps: 4 + distributed_backend: none + is_resume: false + pretrained_checkpoint: null + resume_step: 0 + reload_modules: null + optimizer: + name: AdamW + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + weight_decay: 1.0e-08 + fused: false + save_format: pt +workspace_dir: WORKSPACE_DIR +run_root_dir: /workspace/starVLA/results/Checkpoints +seed: 42 +wandb_entity: dongqianyu99-zhejiang-university +wandb_project: starVLA_rl_games +auth: + env_file: null + hf_token_env: HF_TOKEN + wandb_api_key_env: WANDB_API_KEY +paths: + run_root_dir: results/Checkpoints + dataset_local_dir: playground/Datasets/rl_games + dataset_cache_dir: null + base_model_dir: playground/Pretrained_models/Qwen3-VL-4B-Instruct + accelerate_config: starVLA/config/deepseeds/deepspeed_zero2.yaml +launch: + use_accelerate: true + gpus: null + num_processes: 1 + dry_run: false +conda: + enabled: true + env_name: null +rl_games: + model_alias: openvla + env_eval: + image_size: 224 + frameskip: 4 + seed: 42 + fixed_episode_seeds: true + latency_seed_stride: 0 + task_seed_stride: 0 + task_description: 'You are playing Demon Attack from a single game image. Choose + exactly one action from: NOOP, FIRE, RIGHT, LEFT, RIGHTFIRE, LEFTFIRE.' + demon_attack: + noop_max: 30 + enabled: true + distributed_mode: none + vectorized: + enabled: false + batch_size: 1 + latency: + prompt_map_path: /workspace/starVLA/data/demon_attack_fix_latency_6_200ep/demon_attack_train__bridge/latency_prompt_map.json + mode: single + values: + - 0 + mid_train: + enabled: false + interval_steps: 100 + latencies: + - 0 + num_episodes: 5 + max_steps_per_episode: 3600 + post_train: + enabled: false + latencies: + - 0 + - 1 + - 2 + - 3 + - 4 + num_episodes: 5 + max_steps_per_episode: 3600 + task: demon_attack + initialization_mode: bridge + action_carrier: bridge +model: openvla +env: demon_attack +init: bridge +bridge_base_model: + repo_id: + openvla: Qwen/Qwen3-VL-4B-Instruct + pi0: StarVLA/Qwen2.5-VL-3B-Instruct-Action + pi05: Qwen/Qwen3-VL-4B-Instruct + gr00t: Qwen/Qwen3-VL-4B-Instruct + local_dir: + openvla: playground/Pretrained_models/Qwen3-VL-4B-Instruct + pi0: playground/Pretrained_models/Qwen2.5-VL-3B-Instruct-Action + pi05: playground/Pretrained_models/Qwen3-VL-4B-Instruct + gr00t: playground/Pretrained_models/Qwen3-VL-4B-Instruct +mode: single +checkpoint: + load: auto + hf_repo_id: null + save_best_model: false + save_pt_file: false + local: + keep_last_n: 1 + sync: + enabled: false + repo_id: null + keep_last_n: 0 + sync_every_n_checkpoints: 1 + resume_policy: local_latest +run_id: demon_attack_fix_latency_6_200ep +output_dir: /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_6_200ep +config_yaml: null +is_debug: false +version_id: '0.21' diff --git a/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_6_200ep/config.yaml b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_6_200ep/config.yaml new file mode 100644 index 0000000000000000000000000000000000000000..817903d559cbd8c405de111abb296b07a5dd4f8c --- /dev/null +++ b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_6_200ep/config.yaml @@ -0,0 +1,75 @@ +checkpoint: + local: + keep_last_n: 1 + save_best_model: false + save_pt_file: false + sync: + enabled: false + keep_last_n: 0 + repo_id: null +datasets: + vla_data: + data_mix: demon_attack_train__bridge + dataset_py: lerobot_datasets + eval_data_mix: demon_attack_train__bridge__val + latency_curriculum: + enabled: false + obs_image_size: null + per_device_batch_size: 32 +framework: + action_model: + action_dim: 7 + action_env_dim: 6 + action_hidden_dim: 2560 + action_horizon: 1 + action_model_type: MLP + loss_type: discrete_ce + name: QwenOFT + qwenvl: + attn_implementation: flash_attention_2 + base_vlm: /workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + enable_gradient_checkpointing: true +output_dir: /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_6_200ep +rl_games: + env_eval: + enabled: true + mid_train: + enabled: false + interval_steps: 100 + task: demon_attack +run_id: demon_attack_fix_latency_6_200ep +run_root_dir: /workspace/starVLA/results/Checkpoints +seed: 42 +trainer: + cc_f1_tolerance: 1 + distributed_backend: none + eval_action_classification: true + eval_interval: 100 + eval_num_batches: 100 + freeze_modules: '' + gradient_accumulation_steps: 4 + gradient_clipping: 1.0 + is_resume: false + learning_rate: + action_model: 0.0001 + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + logging_frequency: 1 + lr_scheduler_type: cosine_with_min_lr + max_train_steps: 7000 + num_warmup_steps: 100 + optimizer: + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + fused: false + weight_decay: 1.0e-08 + per_latency_eval_num_batches: null + pretrained_checkpoint: null + reload_modules: null + save_interval: 100 + scheduler_specific_kwargs: + min_lr: 1.0e-06 +wandb_entity: dongqianyu99-zhejiang-university +wandb_project: starVLA_rl_games diff --git a/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_6_200ep/dataset_statistics.json b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_6_200ep/dataset_statistics.json new file mode 100644 index 0000000000000000000000000000000000000000..e59b9bd8df4642f46eab492ea9cc4538976ee592 --- /dev/null +++ b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_6_200ep/dataset_statistics.json @@ -0,0 +1,127 @@ +{ + "new_embodiment": { + "action": { + "mean": [ + 0.06990445405244827, + 0.6139481067657471, + 0.008142709732055664, + 0.061076752841472626, + 0.1516113430261612, + 0.09531665593385696, + 0.0 + ], + "std": [ + 0.2553316354751587, + 0.48614779114723206, + 0.08989176899194717, + 0.23972871899604797, + 0.3591064512729645, + 0.29394957423210144, + 0.0 + ], + "max": [ + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 1.0, + 1.0, + 0.0, + 1.0, + 1.0, + 1.0, + 0.0 + ], + "mask": [ + true, + true, + true, + true, + true, + true, + true + ] + }, + "state": { + "mean": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ] + }, + "num_transitions": 310953, + "num_trajectories": 180 + } +} \ No newline at end of file diff --git a/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_6_200ep/dataset_statistics_eval.json b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_6_200ep/dataset_statistics_eval.json new file mode 100644 index 0000000000000000000000000000000000000000..b4f731b76c7940127f8e3f1ce9b15ac56a033778 --- /dev/null +++ b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_6_200ep/dataset_statistics_eval.json @@ -0,0 +1,127 @@ +{ + "new_embodiment": { + "action": { + "mean": [ + 0.07470002770423889, + 0.6113184094429016, + 0.008066286332905293, + 0.058384548872709274, + 0.15318627655506134, + 0.09434445202350616, + 0.0 + ], + "std": [ + 0.2629486620426178, + 0.48751503229141235, + 0.08943457156419754, + 0.23448707163333893, + 0.36014819145202637, + 0.2922261357307434, + 0.0 + ], + "max": [ + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 1.0, + 1.0, + 0.0, + 1.0, + 1.0, + 1.0, + 0.0 + ], + "mask": [ + true, + true, + true, + true, + true, + true, + true + ] + }, + "state": { + "mean": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ] + }, + "num_transitions": 54672, + "num_trajectories": 20 + } +} \ No newline at end of file diff --git a/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_6_200ep/hydra/.hydra/config.yaml b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_6_200ep/hydra/.hydra/config.yaml new file mode 100644 index 0000000000000000000000000000000000000000..87e0039de280199af7ccb827cc403e525e936433 --- /dev/null +++ b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_6_200ep/hydra/.hydra/config.yaml @@ -0,0 +1,207 @@ +framework: + qwenvl: + base_vlm: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/playground/Pretrained_models/Qwen3-VL-4B-Instruct + attn_implementation: flash_attention_2 + enable_gradient_checkpointing: true + action_model: + state_dim: 7 + loss_type: discrete_ce + action_horizon: 1 + future_action_window_size: 0 + past_action_window_size: 0 + action_dim: 7 + action_env_dim: 6 + name: QwenOFT +datasets: + vla_data: + dataset_py: lerobot_datasets + include_state: true + data_root_dir: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/data/demon_attack_fix_latency_6_200ep + data_mix: demon_attack_train__bridge + eval_data_mix: demon_attack_train__bridge__val + custom_mixtures_path: null + action_type: discrete + sequential_step_sampling: false + shuffle: true + action_balance: + enabled: false + strategy: balanced_epoch + action_key: action_id + target_flap_fraction: 0.3 + noop_id: 0 + flap_id: 1 + latency_curriculum: + enabled: false + strategy: exclusive + latencies: null + phase_steps: null + per_device_batch_size: 32 + load_all_data_for_training: true + obs_image_size: null + video_backend: torchvision_av +dataset: + source_hf: ${dataset.single_source_hf} + config_name: null + source_subdir: null + converted_name: ${dataset.single_converted_name} + single_source_hf: '' + mixed_source_hf: '' + single_converted_name: demon_attack_train + mixed_converted_name: demon_attack_mixed_latency_train + single_latency_filter: null + mixed_latency_filter: null + force_download: false + setup_force: false + skip_verification: false + verify_rows: 200 + max_episodes: null + episodes_per_latency: null + latency_filter: ${dataset.single_latency_filter} + debug_subset: + enabled: false + max_episodes: 5 + suffix: debug +base_model: + repo_id: ${bridge_base_model.repo_id.${model}} +initialization: + checkpoint_local_dir: playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_hf_repo_id: StarVLA/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_filename: checkpoints/steps_5000_pytorch_model.pt +trainer: + max_train_steps: 7000 + num_warmup_steps: 100 + save_interval: 100 + eval_interval: 100 + eval_num_batches: 100 + per_latency_eval_num_batches: null + eval_action_classification: true + cc_f1_tolerance: 1 + learning_rate: + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + action_model: 0.0001 + lr_scheduler_type: cosine_with_min_lr + scheduler_specific_kwargs: + min_lr: 1.0e-06 + freeze_modules: '' + loss_scale: + vla: 1.0 + vlm: 0.1 + max_grad_norm: 1.0 + weight_decay: 0.0 + logging_frequency: 1 + gradient_clipping: 1.0 + gradient_accumulation_steps: 4 + distributed_backend: none + is_resume: false + pretrained_checkpoint: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt + resume_step: 0 + reload_modules: null + optimizer: + name: AdamW + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + weight_decay: 1.0e-08 + fused: false + save_format: pt +workspace_dir: WORKSPACE_DIR +run_root_dir: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints +seed: 42 +wandb_entity: dongqianyu99-zhejiang-university +wandb_project: starVLA_rl_games +auth: + env_file: null + hf_token_env: HF_TOKEN + wandb_api_key_env: WANDB_API_KEY +paths: + run_root_dir: results/Checkpoints + dataset_local_dir: playground/Datasets/rl_games + dataset_cache_dir: null + base_model_dir: ${bridge_base_model.local_dir.${model}} + accelerate_config: starVLA/config/deepseeds/deepspeed_zero2.yaml +launch: + use_accelerate: true + gpus: null + num_processes: 1 + dry_run: false +conda: + enabled: true + env_name: null +rl_games: + model_alias: openvla + env_eval: + image_size: 224 + frameskip: 4 + seed: 42 + fixed_episode_seeds: true + latency_seed_stride: 0 + task_seed_stride: 0 + task_description: 'You are playing Demon Attack from a single game image. Choose + exactly one action from: NOOP, FIRE, RIGHT, LEFT, RIGHTFIRE, LEFTFIRE.' + demon_attack: + noop_max: 30 + enabled: true + distributed_mode: none + vectorized: + enabled: false + batch_size: 1 + latency: + prompt_map_path: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/data/demon_attack_fix_latency_6_200ep/demon_attack_train__bridge/latency_prompt_map.json + mode: single + values: + - 0 + mid_train: + enabled: false + interval_steps: 100 + latencies: + - 0 + num_episodes: 5 + max_steps_per_episode: 3600 + post_train: + enabled: false + latencies: + - 0 + - 1 + - 2 + - 3 + - 4 + num_episodes: 5 + max_steps_per_episode: 3600 + task: demon_attack + initialization_mode: bridge + action_carrier: bridge +model: openvla +env: demon_attack +init: bridge +bridge_base_model: + repo_id: + openvla: Qwen/Qwen3-VL-4B-Instruct + pi0: StarVLA/Qwen2.5-VL-3B-Instruct-Action + pi05: Qwen/Qwen3-VL-4B-Instruct + gr00t: Qwen/Qwen3-VL-4B-Instruct + local_dir: + openvla: playground/Pretrained_models/Qwen3-VL-4B-Instruct + pi0: playground/Pretrained_models/Qwen2.5-VL-3B-Instruct-Action + pi05: playground/Pretrained_models/Qwen3-VL-4B-Instruct + gr00t: playground/Pretrained_models/Qwen3-VL-4B-Instruct +mode: single +checkpoint: + load: auto + hf_repo_id: null + save_best_model: false + save_pt_file: false + local: + keep_last_n: 1 + sync: + enabled: false + repo_id: null + keep_last_n: 0 + sync_every_n_checkpoints: 1 + resume_policy: local_latest +run_id: demon_attack_fix_latency_6_200ep +output_dir: null +config_yaml: null +is_debug: false +version_id: '0.21' diff --git a/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_6_200ep/hydra/.hydra/hydra.yaml b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_6_200ep/hydra/.hydra/hydra.yaml new file mode 100644 index 0000000000000000000000000000000000000000..ab72805df6aeec31117cefac1759e4c3a32a94b8 --- /dev/null +++ b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_6_200ep/hydra/.hydra/hydra.yaml @@ -0,0 +1,259 @@ +hydra: + run: + dir: ${run_root_dir}/${run_id}/hydra + sweep: + dir: multirun/${now:%Y-%m-%d}/${now:%H-%M-%S} + subdir: ${hydra.job.num} + launcher: + _target_: hydra._internal.core_plugins.basic_launcher.BasicLauncher + sweeper: + _target_: hydra._internal.core_plugins.basic_sweeper.BasicSweeper + max_batch_size: null + params: null + help: + app_name: ${hydra.job.name} + header: '${hydra.help.app_name} is powered by Hydra. + + ' + footer: 'Powered by Hydra (https://hydra.cc) + + Use --hydra-help to view Hydra specific help + + ' + template: '${hydra.help.header} + + == Configuration groups == + + Compose your configuration from those groups (group=option) + + + $APP_CONFIG_GROUPS + + + == Config == + + Override anything in the config (foo.bar=value) + + + $CONFIG + + + ${hydra.help.footer} + + ' + hydra_help: + template: 'Hydra (${hydra.runtime.version}) + + See https://hydra.cc for more info. + + + == Flags == + + $FLAGS_HELP + + + == Configuration groups == + + Compose your configuration from those groups (For example, append hydra/job_logging=disabled + to command line) + + + $HYDRA_CONFIG_GROUPS + + + Use ''--cfg hydra'' to Show the Hydra config. + + ' + hydra_help: ??? + hydra_logging: + version: 1 + formatters: + simple: + format: '[%(asctime)s][HYDRA] %(message)s' + handlers: + console: + class: logging.StreamHandler + formatter: simple + stream: ext://sys.stdout + root: + level: INFO + handlers: + - console + loggers: + logging_example: + level: DEBUG + disable_existing_loggers: false + job_logging: + version: 1 + formatters: + simple: + format: '[%(asctime)s][%(name)s][%(levelname)s] - %(message)s' + handlers: + console: + class: logging.StreamHandler + formatter: simple + stream: ext://sys.stdout + file: + class: logging.FileHandler + formatter: simple + filename: ${hydra.runtime.output_dir}/${hydra.job.name}.log + root: + level: INFO + handlers: + - console + - file + disable_existing_loggers: false + env: {} + mode: RUN + searchpath: [] + callbacks: {} + output_subdir: .hydra + overrides: + hydra: + - hydra.mode=RUN + task: + - model=openvla + - env=demon_attack + - init=bridge + - mode=single + - run_id=demon_attack_fix_latency_6_200ep + - run_root_dir=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints + - seed=42 + - wandb_entity=dongqianyu99-zhejiang-university + - wandb_project=starVLA_rl_games + - rl_games.env_eval.enabled=true + - checkpoint.sync.enabled=false + - checkpoint.sync.keep_last_n=0 + - checkpoint.local.keep_last_n=1 + - checkpoint.save_best_model=false + - checkpoint.save_pt_file=false + - trainer.is_resume=false + - trainer.pretrained_checkpoint=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt + - trainer.resume_step=0 + - trainer.max_train_steps=7000 + - trainer.num_warmup_steps=100 + - trainer.save_interval=100 + - trainer.eval_interval=100 + - trainer.eval_num_batches=100 + - trainer.logging_frequency=1 + - trainer.gradient_accumulation_steps=4 + - trainer.distributed_backend=none + - trainer.learning_rate.base=2e-05 + - trainer.learning_rate.qwen_vl_interface=1e-05 + - trainer.learning_rate.action_model=0.0001 + - trainer.lr_scheduler_type=cosine_with_min_lr + - trainer.scheduler_specific_kwargs.min_lr=1e-06 + - trainer.loss_scale.vla=1.0 + - trainer.loss_scale.vlm=0.1 + - trainer.max_grad_norm=1.0 + - trainer.weight_decay=0.0 + - trainer.gradient_clipping=1.0 + - trainer.optimizer.name=AdamW + - trainer.optimizer.betas=[0.9,0.95] + - trainer.optimizer.eps=1e-08 + - trainer.optimizer.weight_decay=1e-08 + - trainer.optimizer.fused=false + - trainer.save_format=pt + - framework.name=QwenOFT + - framework.qwenvl.attn_implementation=flash_attention_2 + - framework.qwenvl.enable_gradient_checkpointing=true + - framework.action_model.action_dim=7 + - framework.action_model.action_env_dim=6 + - framework.action_model.state_dim=7 + - framework.action_model.loss_type=discrete_ce + - framework.action_model.action_horizon=1 + - framework.action_model.future_action_window_size=0 + - framework.action_model.past_action_window_size=0 + - datasets.vla_data.include_state=true + - datasets.vla_data.action_type=discrete + - datasets.vla_data.sequential_step_sampling=false + - datasets.vla_data.shuffle=true + - datasets.vla_data.action_balance.enabled=false + - datasets.vla_data.action_balance.strategy=balanced_epoch + - datasets.vla_data.action_balance.action_key=action_id + - datasets.vla_data.action_balance.target_flap_fraction=0.3 + - datasets.vla_data.action_balance.noop_id=0 + - datasets.vla_data.action_balance.flap_id=1 + - datasets.vla_data.latency_curriculum.enabled=false + - datasets.vla_data.latency_curriculum.strategy=exclusive + - datasets.vla_data.per_device_batch_size=32 + - datasets.vla_data.load_all_data_for_training=true + - datasets.vla_data.video_backend=torchvision_av + - datasets.vla_data.data_root_dir=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/data/demon_attack_fix_latency_6_200ep + - datasets.vla_data.data_mix=demon_attack_train__bridge + - datasets.vla_data.eval_data_mix=demon_attack_train__bridge__val + - framework.qwenvl.base_vlm=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/playground/Pretrained_models/Qwen3-VL-4B-Instruct + - rl_games.task=demon_attack + - rl_games.model_alias=openvla + - rl_games.initialization_mode=bridge + - rl_games.action_carrier=bridge + - rl_games.env_eval.distributed_mode=none + - rl_games.env_eval.latency.mode=single + - rl_games.env_eval.frameskip=4 + - rl_games.env_eval.image_size=224 + - rl_games.env_eval.seed=42 + - rl_games.env_eval.fixed_episode_seeds=true + - rl_games.env_eval.latency_seed_stride=0 + - rl_games.env_eval.task_seed_stride=0 + - 'rl_games.env_eval.task_description=''You are playing Demon Attack from a single + game image. Choose exactly one action from: NOOP, FIRE, RIGHT, LEFT, RIGHTFIRE, + LEFTFIRE.''' + - rl_games.env_eval.latency.values=[0] + - rl_games.env_eval.latency.prompt_map_path=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/data/demon_attack_fix_latency_6_200ep/demon_attack_train__bridge/latency_prompt_map.json + - rl_games.env_eval.mid_train.enabled=false + - rl_games.env_eval.mid_train.interval_steps=100 + - rl_games.env_eval.mid_train.num_episodes=5 + - rl_games.env_eval.mid_train.max_steps_per_episode=3600 + - rl_games.env_eval.mid_train.latencies=[0] + - rl_games.env_eval.post_train.enabled=false + - rl_games.env_eval.post_train.num_episodes=5 + - rl_games.env_eval.post_train.max_steps_per_episode=3600 + - rl_games.env_eval.post_train.latencies=[0,1,2,3,4] + job: + name: train_starvla_hydra + chdir: false + override_dirname: 'checkpoint.local.keep_last_n=1,checkpoint.save_best_model=false,checkpoint.save_pt_file=false,checkpoint.sync.enabled=false,checkpoint.sync.keep_last_n=0,datasets.vla_data.action_balance.action_key=action_id,datasets.vla_data.action_balance.enabled=false,datasets.vla_data.action_balance.flap_id=1,datasets.vla_data.action_balance.noop_id=0,datasets.vla_data.action_balance.strategy=balanced_epoch,datasets.vla_data.action_balance.target_flap_fraction=0.3,datasets.vla_data.action_type=discrete,datasets.vla_data.data_mix=demon_attack_train__bridge,datasets.vla_data.data_root_dir=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/data/demon_attack_fix_latency_6_200ep,datasets.vla_data.eval_data_mix=demon_attack_train__bridge__val,datasets.vla_data.include_state=true,datasets.vla_data.latency_curriculum.enabled=false,datasets.vla_data.latency_curriculum.strategy=exclusive,datasets.vla_data.load_all_data_for_training=true,datasets.vla_data.per_device_batch_size=32,datasets.vla_data.sequential_step_sampling=false,datasets.vla_data.shuffle=true,datasets.vla_data.video_backend=torchvision_av,env=demon_attack,framework.action_model.action_dim=7,framework.action_model.action_env_dim=6,framework.action_model.action_horizon=1,framework.action_model.future_action_window_size=0,framework.action_model.loss_type=discrete_ce,framework.action_model.past_action_window_size=0,framework.action_model.state_dim=7,framework.name=QwenOFT,framework.qwenvl.attn_implementation=flash_attention_2,framework.qwenvl.base_vlm=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/playground/Pretrained_models/Qwen3-VL-4B-Instruct,framework.qwenvl.enable_gradient_checkpointing=true,init=bridge,mode=single,model=openvla,rl_games.action_carrier=bridge,rl_games.env_eval.distributed_mode=none,rl_games.env_eval.enabled=true,rl_games.env_eval.fixed_episode_seeds=true,rl_games.env_eval.frameskip=4,rl_games.env_eval.image_size=224,rl_games.env_eval.latency.mode=single,rl_games.env_eval.latency.prompt_map_path=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/data/demon_attack_fix_latency_6_200ep/demon_attack_train__bridge/latency_prompt_map.json,rl_games.env_eval.latency.values=[0],rl_games.env_eval.latency_seed_stride=0,rl_games.env_eval.mid_train.enabled=false,rl_games.env_eval.mid_train.interval_steps=100,rl_games.env_eval.mid_train.latencies=[0],rl_games.env_eval.mid_train.max_steps_per_episode=3600,rl_games.env_eval.mid_train.num_episodes=5,rl_games.env_eval.post_train.enabled=false,rl_games.env_eval.post_train.latencies=[0,1,2,3,4],rl_games.env_eval.post_train.max_steps_per_episode=3600,rl_games.env_eval.post_train.num_episodes=5,rl_games.env_eval.seed=42,rl_games.env_eval.task_description=''You + are playing Demon Attack from a single game image. Choose exactly one action + from: NOOP, FIRE, RIGHT, LEFT, RIGHTFIRE, LEFTFIRE.'',rl_games.env_eval.task_seed_stride=0,rl_games.initialization_mode=bridge,rl_games.model_alias=openvla,rl_games.task=demon_attack,run_id=demon_attack_fix_latency_6_200ep,run_root_dir=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints,seed=42,trainer.distributed_backend=none,trainer.eval_interval=100,trainer.eval_num_batches=100,trainer.gradient_accumulation_steps=4,trainer.gradient_clipping=1.0,trainer.is_resume=false,trainer.learning_rate.action_model=0.0001,trainer.learning_rate.base=2e-05,trainer.learning_rate.qwen_vl_interface=1e-05,trainer.logging_frequency=1,trainer.loss_scale.vla=1.0,trainer.loss_scale.vlm=0.1,trainer.lr_scheduler_type=cosine_with_min_lr,trainer.max_grad_norm=1.0,trainer.max_train_steps=7000,trainer.num_warmup_steps=100,trainer.optimizer.betas=[0.9,0.95],trainer.optimizer.eps=1e-08,trainer.optimizer.fused=false,trainer.optimizer.name=AdamW,trainer.optimizer.weight_decay=1e-08,trainer.pretrained_checkpoint=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt,trainer.resume_step=0,trainer.save_format=pt,trainer.save_interval=100,trainer.scheduler_specific_kwargs.min_lr=1e-06,trainer.weight_decay=0.0,wandb_entity=dongqianyu99-zhejiang-university,wandb_project=starVLA_rl_games' + id: ??? + num: ??? + config_name: train + env_set: {} + env_copy: [] + config: + override_dirname: + kv_sep: '=' + item_sep: ',' + exclude_keys: [] + runtime: + version: 1.3.2 + version_base: '1.1' + cwd: /inspire/hdd/project/spatialintelligence/public/lzj/starVLA + config_sources: + - path: hydra.conf + schema: pkg + provider: hydra + - path: /inspire/hdd/project/spatialintelligence/public/lzj/starVLA/examples/rl_games/config + schema: file + provider: main + - path: '' + schema: structured + provider: schema + output_dir: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/hydra + choices: + cross_task_setup: null + checkpoint: default + mode: single + init: bridge + env: demon_attack + model: openvla + hydra/env: default + hydra/callbacks: null + hydra/job_logging: default + hydra/hydra_logging: default + hydra/hydra_help: default + hydra/help: default + hydra/sweeper: basic + hydra/launcher: basic + hydra/output: default + verbose: false diff --git a/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_6_200ep/hydra/.hydra/overrides.yaml b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_6_200ep/hydra/.hydra/overrides.yaml new file mode 100644 index 0000000000000000000000000000000000000000..134852bad6474ef00ffc03a574c6d962edc62880 --- /dev/null +++ b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_6_200ep/hydra/.hydra/overrides.yaml @@ -0,0 +1,97 @@ +- model=openvla +- env=demon_attack +- init=bridge +- mode=single +- run_id=demon_attack_fix_latency_6_200ep +- run_root_dir=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints +- seed=42 +- wandb_entity=dongqianyu99-zhejiang-university +- wandb_project=starVLA_rl_games +- rl_games.env_eval.enabled=true +- checkpoint.sync.enabled=false +- checkpoint.sync.keep_last_n=0 +- checkpoint.local.keep_last_n=1 +- checkpoint.save_best_model=false +- checkpoint.save_pt_file=false +- trainer.is_resume=false +- trainer.pretrained_checkpoint=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt +- trainer.resume_step=0 +- trainer.max_train_steps=7000 +- trainer.num_warmup_steps=100 +- trainer.save_interval=100 +- trainer.eval_interval=100 +- trainer.eval_num_batches=100 +- trainer.logging_frequency=1 +- trainer.gradient_accumulation_steps=4 +- trainer.distributed_backend=none +- trainer.learning_rate.base=2e-05 +- trainer.learning_rate.qwen_vl_interface=1e-05 +- trainer.learning_rate.action_model=0.0001 +- trainer.lr_scheduler_type=cosine_with_min_lr +- trainer.scheduler_specific_kwargs.min_lr=1e-06 +- trainer.loss_scale.vla=1.0 +- trainer.loss_scale.vlm=0.1 +- trainer.max_grad_norm=1.0 +- trainer.weight_decay=0.0 +- trainer.gradient_clipping=1.0 +- trainer.optimizer.name=AdamW +- trainer.optimizer.betas=[0.9,0.95] +- trainer.optimizer.eps=1e-08 +- trainer.optimizer.weight_decay=1e-08 +- trainer.optimizer.fused=false +- trainer.save_format=pt +- framework.name=QwenOFT +- framework.qwenvl.attn_implementation=flash_attention_2 +- framework.qwenvl.enable_gradient_checkpointing=true +- framework.action_model.action_dim=7 +- framework.action_model.action_env_dim=6 +- framework.action_model.state_dim=7 +- framework.action_model.loss_type=discrete_ce +- framework.action_model.action_horizon=1 +- framework.action_model.future_action_window_size=0 +- framework.action_model.past_action_window_size=0 +- datasets.vla_data.include_state=true +- datasets.vla_data.action_type=discrete +- datasets.vla_data.sequential_step_sampling=false +- datasets.vla_data.shuffle=true +- datasets.vla_data.action_balance.enabled=false +- datasets.vla_data.action_balance.strategy=balanced_epoch +- datasets.vla_data.action_balance.action_key=action_id +- datasets.vla_data.action_balance.target_flap_fraction=0.3 +- datasets.vla_data.action_balance.noop_id=0 +- datasets.vla_data.action_balance.flap_id=1 +- datasets.vla_data.latency_curriculum.enabled=false +- datasets.vla_data.latency_curriculum.strategy=exclusive +- datasets.vla_data.per_device_batch_size=32 +- datasets.vla_data.load_all_data_for_training=true +- datasets.vla_data.video_backend=torchvision_av +- datasets.vla_data.data_root_dir=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/data/demon_attack_fix_latency_6_200ep +- datasets.vla_data.data_mix=demon_attack_train__bridge +- datasets.vla_data.eval_data_mix=demon_attack_train__bridge__val +- framework.qwenvl.base_vlm=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/playground/Pretrained_models/Qwen3-VL-4B-Instruct +- rl_games.task=demon_attack +- rl_games.model_alias=openvla +- rl_games.initialization_mode=bridge +- rl_games.action_carrier=bridge +- rl_games.env_eval.distributed_mode=none +- rl_games.env_eval.latency.mode=single +- rl_games.env_eval.frameskip=4 +- rl_games.env_eval.image_size=224 +- rl_games.env_eval.seed=42 +- rl_games.env_eval.fixed_episode_seeds=true +- rl_games.env_eval.latency_seed_stride=0 +- rl_games.env_eval.task_seed_stride=0 +- 'rl_games.env_eval.task_description=''You are playing Demon Attack from a single + game image. Choose exactly one action from: NOOP, FIRE, RIGHT, LEFT, RIGHTFIRE, + LEFTFIRE.''' +- rl_games.env_eval.latency.values=[0] +- rl_games.env_eval.latency.prompt_map_path=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/data/demon_attack_fix_latency_6_200ep/demon_attack_train__bridge/latency_prompt_map.json +- rl_games.env_eval.mid_train.enabled=false +- rl_games.env_eval.mid_train.interval_steps=100 +- rl_games.env_eval.mid_train.num_episodes=5 +- rl_games.env_eval.mid_train.max_steps_per_episode=3600 +- rl_games.env_eval.mid_train.latencies=[0] +- rl_games.env_eval.post_train.enabled=false +- rl_games.env_eval.post_train.num_episodes=5 +- rl_games.env_eval.post_train.max_steps_per_episode=3600 +- rl_games.env_eval.post_train.latencies=[0,1,2,3,4] diff --git a/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_6_200ep/hydra/train_starvla_hydra.log b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_6_200ep/hydra/train_starvla_hydra.log new file mode 100644 index 0000000000000000000000000000000000000000..2c38bdd924b98971012f5760c59b4adaba199821 --- /dev/null +++ b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_6_200ep/hydra/train_starvla_hydra.log @@ -0,0 +1,7752 @@ +[2026-06-17 16:47:00,355][starVLA.training.train_starvla][INFO] - VLA Training :: Warming Up +[2026-06-17 16:47:00,366][starVLA.training.train_starvla][INFO] - ✅ Configuration wrapped for access tracking +[2026-06-17 16:49:02,846][starVLA.training.train_starvla][INFO] - Preloaded model checkpoint before Accelerator/DeepSpeed initialization: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt +[2026-06-17 16:49:03,000][starVLA.training.train_starvla][INFO] - Creating VLA Dataset with Mixture `demon_attack_train__bridge` +[2026-06-17 16:49:13,826][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Behavior (data_config): ['BEHAVIOR_challenge'] +[2026-06-17 16:49:13,826][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from DOMINO: ['robotwin'] +[2026-06-17 16:49:13,826][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from DOMINO (data_config): [] +[2026-06-17 16:49:13,826][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from DOMINO (data_config): ['domino', 'domino_clean', 'domino_random', 'domino_cotrain'] +[2026-06-17 16:49:13,826][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Franka: ['custom_robot_config', 'demo_sim_franka_delta_joints', 'SO101'] +[2026-06-17 16:49:13,827][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Franka (data_config): [] +[2026-06-17 16:49:13,827][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Franka (data_config): ['custom_dataset', 'custom_dataset_2', 'demo_sim_pick_place', 'SO101_pick'] +[2026-06-17 16:49:13,827][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from LIBERO: ['libero_franka'] +[2026-06-17 16:49:13,827][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from LIBERO (data_config): [] +[2026-06-17 16:49:13,827][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from LIBERO (data_config): ['libero_all', 'libero_goal', 'multi_robot'] +[2026-06-17 16:49:13,827][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from RoboChallenge_table30v2: ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-17 16:49:13,827][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from RoboChallenge_table30v2 (data_config): ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-17 16:49:13,827][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from RoboChallenge_table30v2 (data_config): ['robochallenge_table30v2_shred_paper', 'robochallenge_table30v2_ur5_all', 'robochallenge_table30v2_arx5_all', 'robochallenge_table30v2_dosw1_all'] +[2026-06-17 16:49:13,827][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_365: ['panda_omron_robocasa365'] +[2026-06-17 16:49:13,827][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_365 (data_config): [] +[2026-06-17 16:49:13,827][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_365 (data_config): ['robocasa365_open_drawer_target_human', 'robocasa365_atomic_target_human_all', 'robocasa365_composite_target_human_all', 'robocasa365_target_human_all'] +[2026-06-17 16:49:13,828][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_tabletop: ['fourier_gr1_arms_waist'] +[2026-06-17 16:49:13,828][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_tabletop (data_config): [] +[2026-06-17 16:49:13,828][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_tabletop (data_config): ['fourier_gr1_unified_1000'] +[2026-06-17 16:49:13,830][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robotwin: ['robotwin', 'robotwin50', 'arx_x5'] +[2026-06-17 16:49:13,830][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robotwin (data_config): [] +[2026-06-17 16:49:13,830][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robotwin (data_config): ['robotwin_all', 'robotwin_all_50', 'robotwin', 'robotwin_task1', 'robotwin_task2', 'arx_x5'] +[2026-06-17 16:49:13,831][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from SimplerEnv: ['oxe_droid', 'oxe_bridge', 'oxe_rt1'] +[2026-06-17 16:49:13,831][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from SimplerEnv (data_config): [] +[2026-06-17 16:49:13,831][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from SimplerEnv (data_config): ['bridge', 'bridge_rt_1'] +[2026-06-17 16:49:13,831][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from VLA-Arena: ['vla_arena_franka'] +[2026-06-17 16:49:13,831][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from VLA-Arena (data_config): [] +[2026-06-17 16:49:13,831][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from VLA-Arena (data_config): ['vla_arena_L0_S', 'vla_arena_L0_M', 'vla_arena_L0_L'] +[2026-06-17 16:49:13,831][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from rl_games: ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-17 16:49:13,831][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from rl_games (data_config): ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-17 16:49:13,831][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from rl_games (data_config): ['flappy_train', 'flappy_train__bridge', 'flappy_mixed_latency_train', 'flappy_mixed_latency_train__bridge', 'demon_attack_train', 'demon_attack_train__bridge', 'demon_attack_mixed_latency_train', 'demon_attack_mixed_latency_train__bridge', 'deadly_corridor_train', 'deadly_corridor_train__bridge', 'deadly_corridor_mixed_latency_train', 'deadly_corridor_mixed_latency_train__bridge'] +[2026-06-17 16:49:13,978][starVLA.training.train_starvla][INFO] - Creating VLA Eval Dataset with Mixture `demon_attack_train__bridge__val` +[2026-06-17 16:49:14,228][starVLA.training.train_starvla][INFO] - 📝 Full config saved at /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/config.full.yaml +[2026-06-17 16:49:14,236][starVLA.training.train_starvla][INFO] - 📊 Accessed config snapshot saved at /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/config.yaml +[2026-06-17 16:49:14,237][starVLA.training.train_starvla][INFO] - No pretrained checkpoint provided. Starting training from scratch. +[2026-06-17 16:49:26,270][starVLA.training.train_starvla][INFO] - ***** Training Configuration ***** +[2026-06-17 16:49:26,271][starVLA.training.train_starvla][INFO] - Total optimization steps = 7000 +[2026-06-17 16:49:26,271][starVLA.training.train_starvla][INFO] - Per device batch size = 32 +[2026-06-17 16:49:26,272][starVLA.training.train_starvla][INFO] - Gradient accumulation steps = 4 +[2026-06-17 16:49:26,272][starVLA.training.train_starvla][INFO] - Total batch size = 128 +[2026-06-17 16:50:33,599][starVLA.training.train_starvla][INFO] - Step 1, Loss: {'train/loss': 1.7677783966064453, 'train/grad_norm_pre_clip': 7.542159557342529, 'timing/data': 0.03133136406540871, 'timing/model': 0.8664270043373108, 'learning_rate/qwen_vl_interface': 1.0000000000000001e-07, 'learning_rate/action_model': 1.0000000000000002e-06, 'epoch': 0.0}) +[2026-06-17 16:50:36,779][starVLA.training.train_starvla][INFO] - Step 2, Loss: {'train/loss': 1.7718952894210815, 'train/grad_norm_pre_clip': 11.631218910217285, 'timing/data': 0.03903497941792011, 'timing/model': 0.9979685181751847, 'learning_rate/qwen_vl_interface': 2.0000000000000002e-07, 'learning_rate/action_model': 2.0000000000000003e-06, 'epoch': 0.0}) +[2026-06-17 16:50:39,636][starVLA.training.train_starvla][INFO] - Step 3, Loss: {'train/loss': 1.7642384767532349, 'train/grad_norm_pre_clip': 8.489670753479004, 'timing/data': 0.03526883013546467, 'timing/model': 0.7341658677905798, 'learning_rate/qwen_vl_interface': 3.0000000000000004e-07, 'learning_rate/action_model': 3e-06, 'epoch': 0.0}) +[2026-06-17 16:50:42,495][starVLA.training.train_starvla][INFO] - Step 4, Loss: {'train/loss': 1.7479965686798096, 'train/grad_norm_pre_clip': 9.855609893798828, 'timing/data': 0.03225105535238981, 'timing/model': 0.7343582110479474, 'learning_rate/qwen_vl_interface': 4.0000000000000003e-07, 'learning_rate/action_model': 4.000000000000001e-06, 'epoch': 0.0}) +[2026-06-17 16:50:45,407][starVLA.training.train_starvla][INFO] - Step 5, Loss: {'train/loss': 1.713245153427124, 'train/grad_norm_pre_clip': 10.316790580749512, 'timing/data': 0.03595211822539568, 'timing/model': 0.791488703340292, 'learning_rate/qwen_vl_interface': 5.000000000000001e-07, 'learning_rate/action_model': 5e-06, 'epoch': 0.0}) +[2026-06-17 16:50:48,436][starVLA.training.train_starvla][INFO] - Step 6, Loss: {'train/loss': 1.681128978729248, 'train/grad_norm_pre_clip': 8.572081565856934, 'timing/data': 0.031236346811056137, 'timing/model': 0.7758892262354493, 'learning_rate/qwen_vl_interface': 6.000000000000001e-07, 'learning_rate/action_model': 6e-06, 'epoch': 0.0}) +[2026-06-17 16:50:51,365][starVLA.training.train_starvla][INFO] - Step 7, Loss: {'train/loss': 1.6284877061843872, 'train/grad_norm_pre_clip': 8.449813842773438, 'timing/data': 0.029988901689648628, 'timing/model': 0.7450479464605451, 'learning_rate/qwen_vl_interface': 7.000000000000001e-07, 'learning_rate/action_model': 7.000000000000001e-06, 'epoch': 0.0}) +[2026-06-17 16:50:54,286][starVLA.training.train_starvla][INFO] - Step 8, Loss: {'train/loss': 1.624849796295166, 'train/grad_norm_pre_clip': 6.0658040046691895, 'timing/data': 0.02229093201458454, 'timing/model': 0.7400676002725959, 'learning_rate/qwen_vl_interface': 8.000000000000001e-07, 'learning_rate/action_model': 8.000000000000001e-06, 'epoch': 0.0}) +[2026-06-17 16:50:57,984][starVLA.training.train_starvla][INFO] - Step 9, Loss: {'train/loss': 1.5194815397262573, 'train/grad_norm_pre_clip': 7.461197853088379, 'timing/data': 0.03500268142670393, 'timing/model': 0.7377362074330449, 'learning_rate/qwen_vl_interface': 9.000000000000001e-07, 'learning_rate/action_model': 9e-06, 'epoch': 0.0}) +[2026-06-17 16:51:00,875][starVLA.training.train_starvla][INFO] - Step 10, Loss: {'train/loss': 1.3827991485595703, 'train/grad_norm_pre_clip': 9.029788970947266, 'timing/data': 0.03565686475485563, 'timing/model': 0.7437174133956432, 'learning_rate/qwen_vl_interface': 1.0000000000000002e-06, 'learning_rate/action_model': 1e-05, 'epoch': 0.0}) +[2026-06-17 16:51:03,743][starVLA.training.train_starvla][INFO] - Step 11, Loss: {'train/loss': 1.4223417043685913, 'train/grad_norm_pre_clip': 5.878377914428711, 'timing/data': 0.03138535562902689, 'timing/model': 0.7535898080095649, 'learning_rate/qwen_vl_interface': 1.1e-06, 'learning_rate/action_model': 1.1000000000000001e-05, 'epoch': 0.0}) +[2026-06-17 16:51:06,626][starVLA.training.train_starvla][INFO] - Step 12, Loss: {'train/loss': 1.5178956985473633, 'train/grad_norm_pre_clip': 3.334635019302368, 'timing/data': 0.028387383557856083, 'timing/model': 0.7247487613931298, 'learning_rate/qwen_vl_interface': 1.2000000000000002e-06, 'learning_rate/action_model': 1.2e-05, 'epoch': 0.0}) +[2026-06-17 16:51:09,630][starVLA.training.train_starvla][INFO] - Step 13, Loss: {'train/loss': 1.3000582456588745, 'train/grad_norm_pre_clip': 4.384140491485596, 'timing/data': 0.03517978359013796, 'timing/model': 0.8705192795023322, 'learning_rate/qwen_vl_interface': 1.3e-06, 'learning_rate/action_model': 1.3000000000000001e-05, 'epoch': 0.01}) +[2026-06-17 16:51:12,599][starVLA.training.train_starvla][INFO] - Step 14, Loss: {'train/loss': 1.0781853199005127, 'train/grad_norm_pre_clip': 5.287487030029297, 'timing/data': 0.02773313596844673, 'timing/model': 0.800530968233943, 'learning_rate/qwen_vl_interface': 1.4000000000000001e-06, 'learning_rate/action_model': 1.4000000000000001e-05, 'epoch': 0.01}) +[2026-06-17 16:51:15,527][starVLA.training.train_starvla][INFO] - Step 15, Loss: {'train/loss': 1.2272852659225464, 'train/grad_norm_pre_clip': 2.9499263763427734, 'timing/data': 0.029846124351024628, 'timing/model': 0.8208014136180282, 'learning_rate/qwen_vl_interface': 1.5e-06, 'learning_rate/action_model': 1.5e-05, 'epoch': 0.01}) +[2026-06-17 16:51:18,395][starVLA.training.train_starvla][INFO] - Step 16, Loss: {'train/loss': 1.3030308485031128, 'train/grad_norm_pre_clip': 2.187178134918213, 'timing/data': 0.044831094332039356, 'timing/model': 0.7303721681237221, 'learning_rate/qwen_vl_interface': 1.6000000000000001e-06, 'learning_rate/action_model': 1.6000000000000003e-05, 'epoch': 0.01}) +[2026-06-17 16:51:21,292][starVLA.training.train_starvla][INFO] - Step 17, Loss: {'train/loss': 1.5377713441848755, 'train/grad_norm_pre_clip': 4.3974738121032715, 'timing/data': 0.02643996011465788, 'timing/model': 0.7445986792445183, 'learning_rate/qwen_vl_interface': 1.7000000000000002e-06, 'learning_rate/action_model': 1.7000000000000003e-05, 'epoch': 0.01}) +[2026-06-17 16:51:24,199][starVLA.training.train_starvla][INFO] - Step 18, Loss: {'train/loss': 1.2827528715133667, 'train/grad_norm_pre_clip': 3.103233814239502, 'timing/data': 0.02622952125966549, 'timing/model': 0.7292012665420771, 'learning_rate/qwen_vl_interface': 1.8000000000000001e-06, 'learning_rate/action_model': 1.8e-05, 'epoch': 0.01}) +[2026-06-17 16:51:27,163][starVLA.training.train_starvla][INFO] - Step 19, Loss: {'train/loss': 1.0531405210494995, 'train/grad_norm_pre_clip': 1.6118340492248535, 'timing/data': 0.022884031757712364, 'timing/model': 0.7437222050502896, 'learning_rate/qwen_vl_interface': 1.9000000000000002e-06, 'learning_rate/action_model': 1.9e-05, 'epoch': 0.01}) +[2026-06-17 16:51:30,143][starVLA.training.train_starvla][INFO] - Step 20, Loss: {'train/loss': 1.2801979780197144, 'train/grad_norm_pre_clip': 3.9890925884246826, 'timing/data': 0.030511289834976196, 'timing/model': 0.7685529105365276, 'learning_rate/qwen_vl_interface': 2.0000000000000003e-06, 'learning_rate/action_model': 2e-05, 'epoch': 0.01}) +[2026-06-17 16:51:32,971][starVLA.training.train_starvla][INFO] - Step 21, Loss: {'train/loss': 1.608650803565979, 'train/grad_norm_pre_clip': 6.1863813400268555, 'timing/data': 0.029345613904297352, 'timing/model': 0.7368445135653019, 'learning_rate/qwen_vl_interface': 2.1000000000000002e-06, 'learning_rate/action_model': 2.1e-05, 'epoch': 0.01}) +[2026-06-17 16:51:36,179][starVLA.training.train_starvla][INFO] - Step 22, Loss: {'train/loss': 1.791437029838562, 'train/grad_norm_pre_clip': 7.025918006896973, 'timing/data': 0.03685290738940239, 'timing/model': 1.043835399672389, 'learning_rate/qwen_vl_interface': 2.2e-06, 'learning_rate/action_model': 2.2000000000000003e-05, 'epoch': 0.01}) +[2026-06-17 16:51:39,091][starVLA.training.train_starvla][INFO] - Step 23, Loss: {'train/loss': 1.128738284111023, 'train/grad_norm_pre_clip': 2.312258243560791, 'timing/data': 0.023773379623889923, 'timing/model': 0.7410866040736437, 'learning_rate/qwen_vl_interface': 2.3000000000000004e-06, 'learning_rate/action_model': 2.3000000000000003e-05, 'epoch': 0.01}) +[2026-06-17 16:51:41,953][starVLA.training.train_starvla][INFO] - Step 24, Loss: {'train/loss': 1.1243475675582886, 'train/grad_norm_pre_clip': 1.6597427129745483, 'timing/data': 0.031437444500625134, 'timing/model': 0.735942654311657, 'learning_rate/qwen_vl_interface': 2.4000000000000003e-06, 'learning_rate/action_model': 2.4e-05, 'epoch': 0.01}) +[2026-06-17 16:51:44,826][starVLA.training.train_starvla][INFO] - Step 25, Loss: {'train/loss': 1.3503332138061523, 'train/grad_norm_pre_clip': 3.243929862976074, 'timing/data': 0.02279081381857395, 'timing/model': 0.7533096075057983, 'learning_rate/qwen_vl_interface': 2.5e-06, 'learning_rate/action_model': 2.5e-05, 'epoch': 0.01}) +[2026-06-17 16:51:47,990][starVLA.training.train_starvla][INFO] - Step 26, Loss: {'train/loss': 1.75709867477417, 'train/grad_norm_pre_clip': 5.130255222320557, 'timing/data': 0.022222165018320084, 'timing/model': 0.728011965751648, 'learning_rate/qwen_vl_interface': 2.6e-06, 'learning_rate/action_model': 2.6000000000000002e-05, 'epoch': 0.01}) +[2026-06-17 16:51:51,291][starVLA.training.train_starvla][INFO] - Step 27, Loss: {'train/loss': 1.3893500566482544, 'train/grad_norm_pre_clip': 2.098816394805908, 'timing/data': 0.02250005677342415, 'timing/model': 1.1926098195835948, 'learning_rate/qwen_vl_interface': 2.7000000000000004e-06, 'learning_rate/action_model': 2.7000000000000002e-05, 'epoch': 0.01}) +[2026-06-17 16:51:55,288][starVLA.training.train_starvla][INFO] - Step 28, Loss: {'train/loss': 1.3268569707870483, 'train/grad_norm_pre_clip': 3.214137077331543, 'timing/data': 0.02693642769008875, 'timing/model': 0.8026251941919327, 'learning_rate/qwen_vl_interface': 2.8000000000000003e-06, 'learning_rate/action_model': 2.8000000000000003e-05, 'epoch': 0.01}) +[2026-06-17 16:51:58,218][starVLA.training.train_starvla][INFO] - Step 29, Loss: {'train/loss': 1.5054012537002563, 'train/grad_norm_pre_clip': 2.798760414123535, 'timing/data': 0.03941219858825207, 'timing/model': 0.7766925692558289, 'learning_rate/qwen_vl_interface': 2.9e-06, 'learning_rate/action_model': 2.9e-05, 'epoch': 0.01}) +[2026-06-17 16:52:01,056][starVLA.training.train_starvla][INFO] - Step 30, Loss: {'train/loss': 1.381118655204773, 'train/grad_norm_pre_clip': 3.2952070236206055, 'timing/data': 0.026600555516779423, 'timing/model': 0.737499033100903, 'learning_rate/qwen_vl_interface': 3e-06, 'learning_rate/action_model': 3e-05, 'epoch': 0.01}) +[2026-06-17 16:52:03,961][starVLA.training.train_starvla][INFO] - Step 31, Loss: {'train/loss': 1.4493359327316284, 'train/grad_norm_pre_clip': 6.56747579574585, 'timing/data': 0.024533742107450962, 'timing/model': 0.7310443259775639, 'learning_rate/qwen_vl_interface': 3.1000000000000004e-06, 'learning_rate/action_model': 3.1e-05, 'epoch': 0.01}) +[2026-06-17 16:52:06,807][starVLA.training.train_starvla][INFO] - Step 32, Loss: {'train/loss': 1.4353135824203491, 'train/grad_norm_pre_clip': 4.19177770614624, 'timing/data': 0.02654108591377735, 'timing/model': 0.7526568211615086, 'learning_rate/qwen_vl_interface': 3.2000000000000003e-06, 'learning_rate/action_model': 3.2000000000000005e-05, 'epoch': 0.01}) +[2026-06-17 16:52:09,651][starVLA.training.train_starvla][INFO] - Step 33, Loss: {'train/loss': 1.4844188690185547, 'train/grad_norm_pre_clip': 4.658655643463135, 'timing/data': 0.03898353409022093, 'timing/model': 0.7351239705458283, 'learning_rate/qwen_vl_interface': 3.3000000000000006e-06, 'learning_rate/action_model': 3.3e-05, 'epoch': 0.01}) +[2026-06-17 16:52:12,500][starVLA.training.train_starvla][INFO] - Step 34, Loss: {'train/loss': 1.3188748359680176, 'train/grad_norm_pre_clip': 3.3138957023620605, 'timing/data': 0.02154109813272953, 'timing/model': 0.7538152290508151, 'learning_rate/qwen_vl_interface': 3.4000000000000005e-06, 'learning_rate/action_model': 3.4000000000000007e-05, 'epoch': 0.01}) +[2026-06-17 16:52:15,576][starVLA.training.train_starvla][INFO] - Step 35, Loss: {'train/loss': 1.2506980895996094, 'train/grad_norm_pre_clip': 5.630130290985107, 'timing/data': 0.024054021574556828, 'timing/model': 0.7541371043771505, 'learning_rate/qwen_vl_interface': 3.5e-06, 'learning_rate/action_model': 3.5e-05, 'epoch': 0.01}) +[2026-06-17 16:52:18,428][starVLA.training.train_starvla][INFO] - Step 36, Loss: {'train/loss': 1.0970264673233032, 'train/grad_norm_pre_clip': 5.084575653076172, 'timing/data': 0.030118771828711033, 'timing/model': 0.7380721215158701, 'learning_rate/qwen_vl_interface': 3.6000000000000003e-06, 'learning_rate/action_model': 3.6e-05, 'epoch': 0.01}) +[2026-06-17 16:52:21,232][starVLA.training.train_starvla][INFO] - Step 37, Loss: {'train/loss': 1.1728572845458984, 'train/grad_norm_pre_clip': 3.334322929382324, 'timing/data': 0.02278466522693634, 'timing/model': 0.7350204288959503, 'learning_rate/qwen_vl_interface': 3.7e-06, 'learning_rate/action_model': 3.7e-05, 'epoch': 0.02}) +[2026-06-17 16:52:24,138][starVLA.training.train_starvla][INFO] - Step 38, Loss: {'train/loss': 1.2587661743164062, 'train/grad_norm_pre_clip': 1.107424020767212, 'timing/data': 0.04035290516912937, 'timing/model': 0.7300616512075067, 'learning_rate/qwen_vl_interface': 3.8000000000000005e-06, 'learning_rate/action_model': 3.8e-05, 'epoch': 0.02}) +[2026-06-17 16:52:26,949][starVLA.training.train_starvla][INFO] - Step 39, Loss: {'train/loss': 1.1639000177383423, 'train/grad_norm_pre_clip': 1.5405142307281494, 'timing/data': 0.03459041006863117, 'timing/model': 0.7252713656052947, 'learning_rate/qwen_vl_interface': 3.900000000000001e-06, 'learning_rate/action_model': 3.9000000000000006e-05, 'epoch': 0.02}) +[2026-06-17 16:52:29,745][starVLA.training.train_starvla][INFO] - Step 40, Loss: {'train/loss': 1.1226606369018555, 'train/grad_norm_pre_clip': 1.703951120376587, 'timing/data': 0.029259820468723774, 'timing/model': 0.7239900687709451, 'learning_rate/qwen_vl_interface': 4.000000000000001e-06, 'learning_rate/action_model': 4e-05, 'epoch': 0.02}) +[2026-06-17 16:52:32,579][starVLA.training.train_starvla][INFO] - Step 41, Loss: {'train/loss': 1.2045732736587524, 'train/grad_norm_pre_clip': 3.623197555541992, 'timing/data': 0.03303102031350136, 'timing/model': 0.7335802111774683, 'learning_rate/qwen_vl_interface': 4.1e-06, 'learning_rate/action_model': 4.1e-05, 'epoch': 0.02}) +[2026-06-17 16:52:35,363][starVLA.training.train_starvla][INFO] - Step 42, Loss: {'train/loss': 0.866359531879425, 'train/grad_norm_pre_clip': 1.5777981281280518, 'timing/data': 0.02449987269937992, 'timing/model': 0.7243705429136753, 'learning_rate/qwen_vl_interface': 4.2000000000000004e-06, 'learning_rate/action_model': 4.2e-05, 'epoch': 0.02}) +[2026-06-17 16:52:38,179][starVLA.training.train_starvla][INFO] - Step 43, Loss: {'train/loss': 1.185181975364685, 'train/grad_norm_pre_clip': 3.321855306625366, 'timing/data': 0.025182459503412247, 'timing/model': 0.7337141912430525, 'learning_rate/qwen_vl_interface': 4.3e-06, 'learning_rate/action_model': 4.3e-05, 'epoch': 0.02}) +[2026-06-17 16:52:41,193][starVLA.training.train_starvla][INFO] - Step 44, Loss: {'train/loss': 1.1057389974594116, 'train/grad_norm_pre_clip': 1.8340305089950562, 'timing/data': 0.029362629167735577, 'timing/model': 0.8131993766874075, 'learning_rate/qwen_vl_interface': 4.4e-06, 'learning_rate/action_model': 4.4000000000000006e-05, 'epoch': 0.02}) +[2026-06-17 16:52:44,040][starVLA.training.train_starvla][INFO] - Step 45, Loss: {'train/loss': 1.3681353330612183, 'train/grad_norm_pre_clip': 3.9247069358825684, 'timing/data': 0.03313850145787001, 'timing/model': 0.7311669867485762, 'learning_rate/qwen_vl_interface': 4.5e-06, 'learning_rate/action_model': 4.5e-05, 'epoch': 0.02}) +[2026-06-17 16:52:47,800][starVLA.training.train_starvla][INFO] - Step 46, Loss: {'train/loss': 1.5595983266830444, 'train/grad_norm_pre_clip': 4.839189052581787, 'timing/data': 0.023650795221328735, 'timing/model': 0.7427543215453625, 'learning_rate/qwen_vl_interface': 4.600000000000001e-06, 'learning_rate/action_model': 4.600000000000001e-05, 'epoch': 0.02}) +[2026-06-17 16:52:50,663][starVLA.training.train_starvla][INFO] - Step 47, Loss: {'train/loss': 1.6439322233200073, 'train/grad_norm_pre_clip': 7.125220775604248, 'timing/data': 0.022027133964002132, 'timing/model': 0.7341340547427535, 'learning_rate/qwen_vl_interface': 4.7e-06, 'learning_rate/action_model': 4.7e-05, 'epoch': 0.02}) +[2026-06-17 16:52:53,470][starVLA.training.train_starvla][INFO] - Step 48, Loss: {'train/loss': 1.1887353658676147, 'train/grad_norm_pre_clip': 2.204472780227661, 'timing/data': 0.03255991078913212, 'timing/model': 0.7376848636195064, 'learning_rate/qwen_vl_interface': 4.800000000000001e-06, 'learning_rate/action_model': 4.8e-05, 'epoch': 0.02}) +[2026-06-17 16:52:56,371][starVLA.training.train_starvla][INFO] - Step 49, Loss: {'train/loss': 1.2511540651321411, 'train/grad_norm_pre_clip': 1.841915488243103, 'timing/data': 0.03454593103379011, 'timing/model': 0.8073273543268442, 'learning_rate/qwen_vl_interface': 4.9000000000000005e-06, 'learning_rate/action_model': 4.9e-05, 'epoch': 0.02}) +[2026-06-17 16:52:59,211][starVLA.training.train_starvla][INFO] - Step 50, Loss: {'train/loss': 0.9799042344093323, 'train/grad_norm_pre_clip': 6.559732437133789, 'timing/data': 0.029698215425014496, 'timing/model': 0.7369725918397307, 'learning_rate/qwen_vl_interface': 5e-06, 'learning_rate/action_model': 5e-05, 'epoch': 0.02}) +[2026-06-17 16:53:02,031][starVLA.training.train_starvla][INFO] - Step 51, Loss: {'train/loss': 1.4007155895233154, 'train/grad_norm_pre_clip': 2.9626803398132324, 'timing/data': 0.034014783799648285, 'timing/model': 0.731739018112421, 'learning_rate/qwen_vl_interface': 5.1e-06, 'learning_rate/action_model': 5.1000000000000006e-05, 'epoch': 0.02}) +[2026-06-17 16:53:04,881][starVLA.training.train_starvla][INFO] - Step 52, Loss: {'train/loss': 1.182476282119751, 'train/grad_norm_pre_clip': 3.860836982727051, 'timing/data': 0.02230601105839014, 'timing/model': 0.7389605501666665, 'learning_rate/qwen_vl_interface': 5.2e-06, 'learning_rate/action_model': 5.2000000000000004e-05, 'epoch': 0.02}) +[2026-06-17 16:53:07,771][starVLA.training.train_starvla][INFO] - Step 53, Loss: {'train/loss': 1.2425079345703125, 'train/grad_norm_pre_clip': 5.811413764953613, 'timing/data': 0.027033108286559582, 'timing/model': 0.8015445638448, 'learning_rate/qwen_vl_interface': 5.300000000000001e-06, 'learning_rate/action_model': 5.300000000000001e-05, 'epoch': 0.02}) +[2026-06-17 16:53:10,679][starVLA.training.train_starvla][INFO] - Step 54, Loss: {'train/loss': 1.2155145406723022, 'train/grad_norm_pre_clip': 3.8866653442382812, 'timing/data': 0.02737412229180336, 'timing/model': 0.7869038181379437, 'learning_rate/qwen_vl_interface': 5.400000000000001e-06, 'learning_rate/action_model': 5.4000000000000005e-05, 'epoch': 0.02}) +[2026-06-17 16:53:13,564][starVLA.training.train_starvla][INFO] - Step 55, Loss: {'train/loss': 1.046223521232605, 'train/grad_norm_pre_clip': 3.687878370285034, 'timing/data': 0.030469322577118874, 'timing/model': 0.7385937990620732, 'learning_rate/qwen_vl_interface': 5.500000000000001e-06, 'learning_rate/action_model': 5.500000000000001e-05, 'epoch': 0.02}) +[2026-06-17 16:53:16,601][starVLA.training.train_starvla][INFO] - Step 56, Loss: {'train/loss': 0.9490922093391418, 'train/grad_norm_pre_clip': 2.8102498054504395, 'timing/data': 0.030246028676629066, 'timing/model': 0.7464559441432357, 'learning_rate/qwen_vl_interface': 5.600000000000001e-06, 'learning_rate/action_model': 5.6000000000000006e-05, 'epoch': 0.02}) +[2026-06-17 16:53:19,486][starVLA.training.train_starvla][INFO] - Step 57, Loss: {'train/loss': 1.5415455102920532, 'train/grad_norm_pre_clip': 4.3488545417785645, 'timing/data': 0.035694449208676815, 'timing/model': 0.8064426807686687, 'learning_rate/qwen_vl_interface': 5.7e-06, 'learning_rate/action_model': 5.6999999999999996e-05, 'epoch': 0.02}) +[2026-06-17 16:53:23,223][starVLA.training.train_starvla][INFO] - Step 58, Loss: {'train/loss': 0.9768402576446533, 'train/grad_norm_pre_clip': 2.081549882888794, 'timing/data': 0.02756981272250414, 'timing/model': 0.7326722592115402, 'learning_rate/qwen_vl_interface': 5.8e-06, 'learning_rate/action_model': 5.8e-05, 'epoch': 0.02}) +[2026-06-17 16:53:26,026][starVLA.training.train_starvla][INFO] - Step 59, Loss: {'train/loss': 1.2179125547409058, 'train/grad_norm_pre_clip': 3.422036647796631, 'timing/data': 0.028708080761134624, 'timing/model': 0.7278051404282451, 'learning_rate/qwen_vl_interface': 5.9e-06, 'learning_rate/action_model': 5.9e-05, 'epoch': 0.02}) +[2026-06-17 16:53:28,861][starVLA.training.train_starvla][INFO] - Step 60, Loss: {'train/loss': 1.1519458293914795, 'train/grad_norm_pre_clip': 2.615889310836792, 'timing/data': 0.03671980928629637, 'timing/model': 0.7401050021871924, 'learning_rate/qwen_vl_interface': 6e-06, 'learning_rate/action_model': 6e-05, 'epoch': 0.02}) +[2026-06-17 16:53:31,855][starVLA.training.train_starvla][INFO] - Step 61, Loss: {'train/loss': 1.6132564544677734, 'train/grad_norm_pre_clip': 5.595757961273193, 'timing/data': 0.03534538857638836, 'timing/model': 0.729612329043448, 'learning_rate/qwen_vl_interface': 6.1e-06, 'learning_rate/action_model': 6.1e-05, 'epoch': 0.03}) +[2026-06-17 16:53:34,728][starVLA.training.train_starvla][INFO] - Step 62, Loss: {'train/loss': 1.153441309928894, 'train/grad_norm_pre_clip': 1.632203459739685, 'timing/data': 0.025630400516092777, 'timing/model': 0.7380608264356852, 'learning_rate/qwen_vl_interface': 6.200000000000001e-06, 'learning_rate/action_model': 6.2e-05, 'epoch': 0.03}) +[2026-06-17 16:53:37,579][starVLA.training.train_starvla][INFO] - Step 63, Loss: {'train/loss': 1.0926724672317505, 'train/grad_norm_pre_clip': 3.3338162899017334, 'timing/data': 0.03066803514957428, 'timing/model': 0.7317371033132076, 'learning_rate/qwen_vl_interface': 6.300000000000001e-06, 'learning_rate/action_model': 6.3e-05, 'epoch': 0.03}) +[2026-06-17 16:53:40,410][starVLA.training.train_starvla][INFO] - Step 64, Loss: {'train/loss': 1.165225863456726, 'train/grad_norm_pre_clip': 1.8484482765197754, 'timing/data': 0.03354648780077696, 'timing/model': 0.7373220967128873, 'learning_rate/qwen_vl_interface': 6.4000000000000006e-06, 'learning_rate/action_model': 6.400000000000001e-05, 'epoch': 0.03}) +[2026-06-17 16:53:43,226][starVLA.training.train_starvla][INFO] - Step 65, Loss: {'train/loss': 1.153017282485962, 'train/grad_norm_pre_clip': 3.204357624053955, 'timing/data': 0.03466645907610655, 'timing/model': 0.7317366320639849, 'learning_rate/qwen_vl_interface': 6.5000000000000004e-06, 'learning_rate/action_model': 6.500000000000001e-05, 'epoch': 0.03}) +[2026-06-17 16:53:46,062][starVLA.training.train_starvla][INFO] - Step 66, Loss: {'train/loss': 1.1862541437149048, 'train/grad_norm_pre_clip': 2.311345100402832, 'timing/data': 0.03009805455803871, 'timing/model': 0.7369400421157479, 'learning_rate/qwen_vl_interface': 6.600000000000001e-06, 'learning_rate/action_model': 6.6e-05, 'epoch': 0.03}) +[2026-06-17 16:53:49,243][starVLA.training.train_starvla][INFO] - Step 67, Loss: {'train/loss': 1.0921509265899658, 'train/grad_norm_pre_clip': 3.3679211139678955, 'timing/data': 0.035633526742458344, 'timing/model': 0.8070738753303885, 'learning_rate/qwen_vl_interface': 6.700000000000001e-06, 'learning_rate/action_model': 6.7e-05, 'epoch': 0.03}) +[2026-06-17 16:53:52,031][starVLA.training.train_starvla][INFO] - Step 68, Loss: {'train/loss': 0.9484646320343018, 'train/grad_norm_pre_clip': 2.966639518737793, 'timing/data': 0.03465524408966303, 'timing/model': 0.7241710890084505, 'learning_rate/qwen_vl_interface': 6.800000000000001e-06, 'learning_rate/action_model': 6.800000000000001e-05, 'epoch': 0.03}) +[2026-06-17 16:53:54,808][starVLA.training.train_starvla][INFO] - Step 69, Loss: {'train/loss': 1.475743055343628, 'train/grad_norm_pre_clip': 3.321868419647217, 'timing/data': 0.03857172839343548, 'timing/model': 0.7254781257361174, 'learning_rate/qwen_vl_interface': 6.9e-06, 'learning_rate/action_model': 6.9e-05, 'epoch': 0.03}) +[2026-06-17 16:53:57,603][starVLA.training.train_starvla][INFO] - Step 70, Loss: {'train/loss': 1.307031512260437, 'train/grad_norm_pre_clip': 2.908792018890381, 'timing/data': 0.03687770199030638, 'timing/model': 0.7241410221904516, 'learning_rate/qwen_vl_interface': 7e-06, 'learning_rate/action_model': 7e-05, 'epoch': 0.03}) +[2026-06-17 16:54:00,398][starVLA.training.train_starvla][INFO] - Step 71, Loss: {'train/loss': 1.274011254310608, 'train/grad_norm_pre_clip': 3.1226320266723633, 'timing/data': 0.03541738074272871, 'timing/model': 0.7216150388121605, 'learning_rate/qwen_vl_interface': 7.100000000000001e-06, 'learning_rate/action_model': 7.1e-05, 'epoch': 0.03}) +[2026-06-17 16:54:03,186][starVLA.training.train_starvla][INFO] - Step 72, Loss: {'train/loss': 1.664035677909851, 'train/grad_norm_pre_clip': 4.075622081756592, 'timing/data': 0.022790925577282906, 'timing/model': 0.7272270899266005, 'learning_rate/qwen_vl_interface': 7.2000000000000005e-06, 'learning_rate/action_model': 7.2e-05, 'epoch': 0.03}) +[2026-06-17 16:54:05,972][starVLA.training.train_starvla][INFO] - Step 73, Loss: {'train/loss': 1.368842601776123, 'train/grad_norm_pre_clip': 2.4067020416259766, 'timing/data': 0.02238294016569853, 'timing/model': 0.7250576457008719, 'learning_rate/qwen_vl_interface': 7.3e-06, 'learning_rate/action_model': 7.3e-05, 'epoch': 0.03}) +[2026-06-17 16:54:08,739][starVLA.training.train_starvla][INFO] - Step 74, Loss: {'train/loss': 1.10593843460083, 'train/grad_norm_pre_clip': 2.9827651977539062, 'timing/data': 0.022995106875896454, 'timing/model': 0.7258972870185971, 'learning_rate/qwen_vl_interface': 7.4e-06, 'learning_rate/action_model': 7.4e-05, 'epoch': 0.03}) +[2026-06-17 16:54:11,511][starVLA.training.train_starvla][INFO] - Step 75, Loss: {'train/loss': 1.2290217876434326, 'train/grad_norm_pre_clip': 2.084385395050049, 'timing/data': 0.02372277621179819, 'timing/model': 0.7250581393018365, 'learning_rate/qwen_vl_interface': 7.500000000000001e-06, 'learning_rate/action_model': 7.500000000000001e-05, 'epoch': 0.03}) +[2026-06-17 16:54:14,293][starVLA.training.train_starvla][INFO] - Step 76, Loss: {'train/loss': 1.132291555404663, 'train/grad_norm_pre_clip': 3.0805373191833496, 'timing/data': 0.024501713924109936, 'timing/model': 0.7254724288359284, 'learning_rate/qwen_vl_interface': 7.600000000000001e-06, 'learning_rate/action_model': 7.6e-05, 'epoch': 0.03}) +[2026-06-17 16:54:17,086][starVLA.training.train_starvla][INFO] - Step 77, Loss: {'train/loss': 1.542353868484497, 'train/grad_norm_pre_clip': 2.534780740737915, 'timing/data': 0.03388581145554781, 'timing/model': 0.7258150232955813, 'learning_rate/qwen_vl_interface': 7.7e-06, 'learning_rate/action_model': 7.7e-05, 'epoch': 0.03}) +[2026-06-17 16:54:19,862][starVLA.training.train_starvla][INFO] - Step 78, Loss: {'train/loss': 1.3784271478652954, 'train/grad_norm_pre_clip': 1.9875683784484863, 'timing/data': 0.0303813349455595, 'timing/model': 0.7259455006569624, 'learning_rate/qwen_vl_interface': 7.800000000000002e-06, 'learning_rate/action_model': 7.800000000000001e-05, 'epoch': 0.03}) +[2026-06-17 16:54:22,643][starVLA.training.train_starvla][INFO] - Step 79, Loss: {'train/loss': 1.033862590789795, 'train/grad_norm_pre_clip': 3.3563361167907715, 'timing/data': 0.03655994310975075, 'timing/model': 0.7250093575567007, 'learning_rate/qwen_vl_interface': 7.9e-06, 'learning_rate/action_model': 7.900000000000001e-05, 'epoch': 0.03}) +[2026-06-17 16:54:25,435][starVLA.training.train_starvla][INFO] - Step 80, Loss: {'train/loss': 1.0007184743881226, 'train/grad_norm_pre_clip': 3.9422104358673096, 'timing/data': 0.03434806875884533, 'timing/model': 0.7276592804118991, 'learning_rate/qwen_vl_interface': 8.000000000000001e-06, 'learning_rate/action_model': 8e-05, 'epoch': 0.03}) +[2026-06-17 16:54:28,247][starVLA.training.train_starvla][INFO] - Step 81, Loss: {'train/loss': 1.344183087348938, 'train/grad_norm_pre_clip': 2.036050796508789, 'timing/data': 0.03636965807527304, 'timing/model': 0.7315747607499361, 'learning_rate/qwen_vl_interface': 8.1e-06, 'learning_rate/action_model': 8.1e-05, 'epoch': 0.03}) +[2026-06-17 16:54:31,040][starVLA.training.train_starvla][INFO] - Step 82, Loss: {'train/loss': 1.3099939823150635, 'train/grad_norm_pre_clip': 2.6316111087799072, 'timing/data': 0.030960766598582268, 'timing/model': 0.7274641152471304, 'learning_rate/qwen_vl_interface': 8.2e-06, 'learning_rate/action_model': 8.2e-05, 'epoch': 0.03}) +[2026-06-17 16:54:33,865][starVLA.training.train_starvla][INFO] - Step 83, Loss: {'train/loss': 1.6158578395843506, 'train/grad_norm_pre_clip': 6.712810039520264, 'timing/data': 0.04428055416792631, 'timing/model': 0.7406748440116644, 'learning_rate/qwen_vl_interface': 8.3e-06, 'learning_rate/action_model': 8.3e-05, 'epoch': 0.03}) +[2026-06-17 16:54:36,674][starVLA.training.train_starvla][INFO] - Step 84, Loss: {'train/loss': 1.2632864713668823, 'train/grad_norm_pre_clip': 1.6905760765075684, 'timing/data': 0.03251462522894144, 'timing/model': 0.7300836676731706, 'learning_rate/qwen_vl_interface': 8.400000000000001e-06, 'learning_rate/action_model': 8.4e-05, 'epoch': 0.03}) +[2026-06-17 16:54:39,552][starVLA.training.train_starvla][INFO] - Step 85, Loss: {'train/loss': 1.1139744520187378, 'train/grad_norm_pre_clip': 2.2776355743408203, 'timing/data': 0.035344392992556095, 'timing/model': 0.7381484434008598, 'learning_rate/qwen_vl_interface': 8.5e-06, 'learning_rate/action_model': 8.5e-05, 'epoch': 0.03}) +[2026-06-17 16:54:42,558][starVLA.training.train_starvla][INFO] - Step 86, Loss: {'train/loss': 1.263928771018982, 'train/grad_norm_pre_clip': 1.7473770380020142, 'timing/data': 0.03652821760624647, 'timing/model': 0.7246596580371261, 'learning_rate/qwen_vl_interface': 8.6e-06, 'learning_rate/action_model': 8.6e-05, 'epoch': 0.04}) +[2026-06-17 16:54:45,542][starVLA.training.train_starvla][INFO] - Step 87, Loss: {'train/loss': 1.3006387948989868, 'train/grad_norm_pre_clip': 3.408562183380127, 'timing/data': 0.028432106599211693, 'timing/model': 0.7336010225117207, 'learning_rate/qwen_vl_interface': 8.700000000000001e-06, 'learning_rate/action_model': 8.7e-05, 'epoch': 0.04}) +[2026-06-17 16:54:48,562][starVLA.training.train_starvla][INFO] - Step 88, Loss: {'train/loss': 1.2178783416748047, 'train/grad_norm_pre_clip': 3.419106960296631, 'timing/data': 0.031574442982673645, 'timing/model': 0.8147929860278964, 'learning_rate/qwen_vl_interface': 8.8e-06, 'learning_rate/action_model': 8.800000000000001e-05, 'epoch': 0.04}) +[2026-06-17 16:54:51,588][starVLA.training.train_starvla][INFO] - Step 89, Loss: {'train/loss': 0.9604470729827881, 'train/grad_norm_pre_clip': 3.4328486919403076, 'timing/data': 0.02655185479670763, 'timing/model': 0.7579113310202956, 'learning_rate/qwen_vl_interface': 8.900000000000001e-06, 'learning_rate/action_model': 8.900000000000001e-05, 'epoch': 0.04}) +[2026-06-17 16:54:54,763][starVLA.training.train_starvla][INFO] - Step 90, Loss: {'train/loss': 1.151175856590271, 'train/grad_norm_pre_clip': 3.458531141281128, 'timing/data': 0.03532099537551403, 'timing/model': 0.8139584120362997, 'learning_rate/qwen_vl_interface': 9e-06, 'learning_rate/action_model': 9e-05, 'epoch': 0.04}) +[2026-06-17 16:54:58,009][starVLA.training.train_starvla][INFO] - Step 91, Loss: {'train/loss': 1.5326250791549683, 'train/grad_norm_pre_clip': 5.788142681121826, 'timing/data': 0.026970146223902702, 'timing/model': 0.7626590169966221, 'learning_rate/qwen_vl_interface': 9.100000000000001e-06, 'learning_rate/action_model': 9.1e-05, 'epoch': 0.04}) +[2026-06-17 16:55:00,988][starVLA.training.train_starvla][INFO] - Step 92, Loss: {'train/loss': 1.359979271888733, 'train/grad_norm_pre_clip': 4.210568428039551, 'timing/data': 0.030280078761279583, 'timing/model': 0.8081577103585005, 'learning_rate/qwen_vl_interface': 9.200000000000002e-06, 'learning_rate/action_model': 9.200000000000001e-05, 'epoch': 0.04}) +[2026-06-17 16:55:03,793][starVLA.training.train_starvla][INFO] - Step 93, Loss: {'train/loss': 1.3038709163665771, 'train/grad_norm_pre_clip': 5.210051536560059, 'timing/data': 0.025790677405893803, 'timing/model': 0.7327974606305361, 'learning_rate/qwen_vl_interface': 9.3e-06, 'learning_rate/action_model': 9.300000000000001e-05, 'epoch': 0.04}) +[2026-06-17 16:55:06,631][starVLA.training.train_starvla][INFO] - Step 94, Loss: {'train/loss': 0.919599711894989, 'train/grad_norm_pre_clip': 4.878929615020752, 'timing/data': 0.02619838062673807, 'timing/model': 0.7290292298421264, 'learning_rate/qwen_vl_interface': 9.4e-06, 'learning_rate/action_model': 9.4e-05, 'epoch': 0.04}) +[2026-06-17 16:55:09,415][starVLA.training.train_starvla][INFO] - Step 95, Loss: {'train/loss': 1.1056841611862183, 'train/grad_norm_pre_clip': 5.388809680938721, 'timing/data': 0.02663818746805191, 'timing/model': 0.7290931427851319, 'learning_rate/qwen_vl_interface': 9.5e-06, 'learning_rate/action_model': 9.5e-05, 'epoch': 0.04}) +[2026-06-17 16:55:12,391][starVLA.training.train_starvla][INFO] - Step 96, Loss: {'train/loss': 1.283048391342163, 'train/grad_norm_pre_clip': 3.912376642227173, 'timing/data': 0.024738834239542484, 'timing/model': 0.7616807259619236, 'learning_rate/qwen_vl_interface': 9.600000000000001e-06, 'learning_rate/action_model': 9.6e-05, 'epoch': 0.04}) +[2026-06-17 16:55:15,384][starVLA.training.train_starvla][INFO] - Step 97, Loss: {'train/loss': 1.3843461275100708, 'train/grad_norm_pre_clip': 3.6364192962646484, 'timing/data': 0.027709681540727615, 'timing/model': 0.8574044331908226, 'learning_rate/qwen_vl_interface': 9.7e-06, 'learning_rate/action_model': 9.7e-05, 'epoch': 0.04}) +[2026-06-17 16:55:18,371][starVLA.training.train_starvla][INFO] - Step 98, Loss: {'train/loss': 1.259334683418274, 'train/grad_norm_pre_clip': 3.843705654144287, 'timing/data': 0.027993435971438885, 'timing/model': 0.8525521624833345, 'learning_rate/qwen_vl_interface': 9.800000000000001e-06, 'learning_rate/action_model': 9.8e-05, 'epoch': 0.04}) +[2026-06-17 16:55:21,157][starVLA.training.train_starvla][INFO] - Step 99, Loss: {'train/loss': 1.1774131059646606, 'train/grad_norm_pre_clip': 2.945322275161743, 'timing/data': 0.026810642331838608, 'timing/model': 0.7295409673824906, 'learning_rate/qwen_vl_interface': 9.9e-06, 'learning_rate/action_model': 9.900000000000001e-05, 'epoch': 0.04}) +[2026-06-17 16:56:53,843][root][INFO] - gcc -pthread -B /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/include -fPIC -O2 -isystem /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/include -fPIC -c /tmp/tmpt62l4b0v/test.c -o /tmp/tmpt62l4b0v/test.o +[2026-06-17 16:56:53,883][root][INFO] - gcc -pthread -B /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/compiler_compat /tmp/tmpt62l4b0v/test.o -laio -o /tmp/tmpt62l4b0v/a.out +[2026-06-17 16:56:54,526][root][INFO] - gcc -pthread -B /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/include -fPIC -O2 -isystem /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/include -fPIC -c /tmp/tmp10bet3zu/test.c -o /tmp/tmp10bet3zu/test.o +[2026-06-17 16:56:54,541][root][INFO] - gcc -pthread -B /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/compiler_compat /tmp/tmp10bet3zu/test.o -L/usr/local/cuda -L/usr/local/cuda/lib64 -lcufile -o /tmp/tmp10bet3zu/a.out +[2026-06-17 16:56:54,584][root][INFO] - gcc -pthread -B /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/include -fPIC -O2 -isystem /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/include -fPIC -c /tmp/tmpjd94xmbz/test.c -o /tmp/tmpjd94xmbz/test.o +[2026-06-17 16:56:54,597][root][INFO] - gcc -pthread -B /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/compiler_compat /tmp/tmpjd94xmbz/test.o -laio -o /tmp/tmpjd94xmbz/a.out +[2026-06-17 16:58:27,870][starVLA.training.train_starvla][INFO] - Step 100, Loss: {'train/loss': 1.3480448722839355, 'train/grad_norm_pre_clip': 3.5242934226989746, 'eval/loss': 1.1526362895965576, 'eval/latency_6/loss': 1.1526362895965576, 'eval/demon_fire_precision': 0.8819325328759291, 'eval/demon_fire_recall': 0.9970911441499677, 'eval/demon_fire_f1': 0.9359830097087378, 'eval/demon_fire_rate_model': 0.9732888146911519, 'eval/demon_fire_rate_teacher': 0.860879243183083, 'eval/demon_left_precision': 0.8541666666666666, 'eval/demon_left_recall': 0.14385964912280702, 'eval/demon_left_f1': 0.24624624624624625, 'eval/demon_left_rate_model': 0.02671118530884808, 'eval/demon_left_rate_teacher': 0.15859766277128548, 'eval/demon_right_precision': 0.0, 'eval/demon_right_recall': 0.0, 'eval/demon_right_f1': 0.0, 'eval/demon_right_rate_model': 0.0, 'eval/demon_right_rate_teacher': 0.18113522537562604, 'eval/demon_move_f1': 0.12312312312312312, 'eval/demon_cc_f1': 0.314024767947899, 'eval/latency_6/demon_fire_precision': 0.8819325328759291, 'eval/latency_6/demon_fire_recall': 0.9970911441499677, 'eval/latency_6/demon_fire_f1': 0.9359830097087378, 'eval/latency_6/demon_fire_rate_model': 0.9732888146911519, 'eval/latency_6/demon_fire_rate_teacher': 0.860879243183083, 'eval/latency_6/demon_left_precision': 0.8541666666666666, 'eval/latency_6/demon_left_recall': 0.14385964912280702, 'eval/latency_6/demon_left_f1': 0.24624624624624625, 'eval/latency_6/demon_left_rate_model': 0.02671118530884808, 'eval/latency_6/demon_left_rate_teacher': 0.15859766277128548, 'eval/latency_6/demon_right_precision': 0.0, 'eval/latency_6/demon_right_recall': 0.0, 'eval/latency_6/demon_right_f1': 0.0, 'eval/latency_6/demon_right_rate_model': 0.0, 'eval/latency_6/demon_right_rate_teacher': 0.18113522537562604, 'eval/latency_6/demon_move_f1': 0.12312312312312312, 'eval/latency_6/demon_cc_f1': 0.314024767947899, 'timing/data': 0.03165320586413145, 'timing/model': 0.7304291427135468, 'learning_rate/qwen_vl_interface': 1e-05, 'learning_rate/action_model': 0.0001, 'epoch': 0.04}) +[2026-06-17 16:58:27,875][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_100_state +[2026-06-17 16:58:27,895][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 16:58:35,257][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_100_state/model.safetensors +[2026-06-17 16:58:49,956][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_100_state/optimizer.bin +[2026-06-17 16:58:49,957][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_100_state/sampler.bin +[2026-06-17 16:58:49,957][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_100_state/sampler_1.bin +[2026-06-17 16:58:49,960][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_100_state/random_states_0.pkl +[2026-06-17 16:58:49,965][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 100 +[2026-06-17 16:58:49,965][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 16:58:49,977][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 16:58:53,850][starVLA.training.train_starvla][INFO] - Step 101, Loss: {'train/loss': 0.9588661193847656, 'train/grad_norm_pre_clip': 7.675518035888672, 'timing/data': 0.029392601922154427, 'timing/model': 0.9315469367429614, 'learning_rate/qwen_vl_interface': 9.999999507659945e-06, 'learning_rate/action_model': 9.999999507659945e-05, 'epoch': 0.04}) +[2026-06-17 16:58:57,670][starVLA.training.train_starvla][INFO] - Step 102, Loss: {'train/loss': 0.9509338140487671, 'train/grad_norm_pre_clip': 3.3490049839019775, 'timing/data': 0.047467658296227455, 'timing/model': 0.9243249539285898, 'learning_rate/qwen_vl_interface': 9.99999803063988e-06, 'learning_rate/action_model': 9.99999803063988e-05, 'epoch': 0.04}) +[2026-06-17 16:59:01,537][starVLA.training.train_starvla][INFO] - Step 103, Loss: {'train/loss': 1.285292148590088, 'train/grad_norm_pre_clip': 3.9635326862335205, 'timing/data': 0.030581247992813587, 'timing/model': 0.9301352193579078, 'learning_rate/qwen_vl_interface': 9.999995568940113e-06, 'learning_rate/action_model': 9.999995568940113e-05, 'epoch': 0.04}) +[2026-06-17 16:59:05,411][starVLA.training.train_starvla][INFO] - Step 104, Loss: {'train/loss': 1.030258297920227, 'train/grad_norm_pre_clip': 3.357456922531128, 'timing/data': 0.04398365505039692, 'timing/model': 1.1194816185161471, 'learning_rate/qwen_vl_interface': 9.999992122561152e-06, 'learning_rate/action_model': 9.999992122561152e-05, 'epoch': 0.04}) +[2026-06-17 16:59:09,175][starVLA.training.train_starvla][INFO] - Step 105, Loss: {'train/loss': 0.8225131630897522, 'train/grad_norm_pre_clip': 2.9087471961975098, 'timing/data': 0.03221812658011913, 'timing/model': 1.0556585462763906, 'learning_rate/qwen_vl_interface': 9.999987691503714e-06, 'learning_rate/action_model': 9.999987691503714e-05, 'epoch': 0.04}) +[2026-06-17 16:59:12,763][starVLA.training.train_starvla][INFO] - Step 106, Loss: {'train/loss': 0.9373887777328491, 'train/grad_norm_pre_clip': 2.5959525108337402, 'timing/data': 0.02304824721068144, 'timing/model': 0.9280778467655182, 'learning_rate/qwen_vl_interface': 9.999982275768716e-06, 'learning_rate/action_model': 9.999982275768715e-05, 'epoch': 0.04}) +[2026-06-17 16:59:16,572][starVLA.training.train_starvla][INFO] - Step 107, Loss: {'train/loss': 0.7749138474464417, 'train/grad_norm_pre_clip': 12.392507553100586, 'timing/data': 0.03167587332427502, 'timing/model': 0.9322601892054081, 'learning_rate/qwen_vl_interface': 9.99997587535728e-06, 'learning_rate/action_model': 9.99997587535728e-05, 'epoch': 0.04}) +[2026-06-17 16:59:20,406][starVLA.training.train_starvla][INFO] - Step 108, Loss: {'train/loss': 1.6919559240341187, 'train/grad_norm_pre_clip': 7.027068614959717, 'timing/data': 0.03412092383950949, 'timing/model': 0.9333336697891355, 'learning_rate/qwen_vl_interface': 9.999968490270735e-06, 'learning_rate/action_model': 9.999968490270735e-05, 'epoch': 0.04}) +[2026-06-17 16:59:24,252][starVLA.training.train_starvla][INFO] - Step 109, Loss: {'train/loss': 1.5748251676559448, 'train/grad_norm_pre_clip': 23.7506046295166, 'timing/data': 0.028398350812494755, 'timing/model': 0.9721512887626886, 'learning_rate/qwen_vl_interface': 9.99996012051061e-06, 'learning_rate/action_model': 9.99996012051061e-05, 'epoch': 0.04}) +[2026-06-17 16:59:28,054][starVLA.training.train_starvla][INFO] - Step 110, Loss: {'train/loss': 1.322676658630371, 'train/grad_norm_pre_clip': 8.81274127960205, 'timing/data': 0.03469647187739611, 'timing/model': 0.9172915406525135, 'learning_rate/qwen_vl_interface': 9.999950766078641e-06, 'learning_rate/action_model': 9.99995076607864e-05, 'epoch': 0.05}) +[2026-06-17 16:59:31,849][starVLA.training.train_starvla][INFO] - Step 111, Loss: {'train/loss': 0.8474093675613403, 'train/grad_norm_pre_clip': 4.769794940948486, 'timing/data': 0.02309525664895773, 'timing/model': 1.0904220379889011, 'learning_rate/qwen_vl_interface': 9.999940426976768e-06, 'learning_rate/action_model': 9.999940426976768e-05, 'epoch': 0.05}) +[2026-06-17 16:59:35,694][starVLA.training.train_starvla][INFO] - Step 112, Loss: {'train/loss': 1.0950353145599365, 'train/grad_norm_pre_clip': 4.450963497161865, 'timing/data': 0.033026328310370445, 'timing/model': 0.9282127665355802, 'learning_rate/qwen_vl_interface': 9.999929103207133e-06, 'learning_rate/action_model': 9.999929103207132e-05, 'epoch': 0.05}) +[2026-06-17 16:59:39,560][starVLA.training.train_starvla][INFO] - Step 113, Loss: {'train/loss': 1.0100781917572021, 'train/grad_norm_pre_clip': 2.31376314163208, 'timing/data': 0.03431946597993374, 'timing/model': 0.9254641672596335, 'learning_rate/qwen_vl_interface': 9.999916794772083e-06, 'learning_rate/action_model': 9.999916794772082e-05, 'epoch': 0.05}) +[2026-06-17 16:59:43,344][starVLA.training.train_starvla][INFO] - Step 114, Loss: {'train/loss': 1.2155752182006836, 'train/grad_norm_pre_clip': 5.1757283210754395, 'timing/data': 0.02526514884084463, 'timing/model': 0.9308338826522231, 'learning_rate/qwen_vl_interface': 9.99990350167417e-06, 'learning_rate/action_model': 9.999903501674169e-05, 'epoch': 0.05}) +[2026-06-17 16:59:47,184][starVLA.training.train_starvla][INFO] - Step 115, Loss: {'train/loss': 1.156023383140564, 'train/grad_norm_pre_clip': 3.7056992053985596, 'timing/data': 0.033241056837141514, 'timing/model': 0.9283605655655265, 'learning_rate/qwen_vl_interface': 9.99988922391615e-06, 'learning_rate/action_model': 9.99988922391615e-05, 'epoch': 0.05}) +[2026-06-17 16:59:50,978][starVLA.training.train_starvla][INFO] - Step 116, Loss: {'train/loss': 1.4875361919403076, 'train/grad_norm_pre_clip': 21.29111099243164, 'timing/data': 0.025281300768256187, 'timing/model': 0.9554001893848181, 'learning_rate/qwen_vl_interface': 9.999873961500985e-06, 'learning_rate/action_model': 9.999873961500984e-05, 'epoch': 0.05}) +[2026-06-17 16:59:54,787][starVLA.training.train_starvla][INFO] - Step 117, Loss: {'train/loss': 0.9514855146408081, 'train/grad_norm_pre_clip': 3.0184171199798584, 'timing/data': 0.030365724116563797, 'timing/model': 1.1319968337193131, 'learning_rate/qwen_vl_interface': 9.999857714431835e-06, 'learning_rate/action_model': 9.999857714431834e-05, 'epoch': 0.05}) +[2026-06-17 16:59:58,932][starVLA.training.train_starvla][INFO] - Step 118, Loss: {'train/loss': 0.9795684218406677, 'train/grad_norm_pre_clip': 2.8826091289520264, 'timing/data': 0.03442489355802536, 'timing/model': 1.1740000694990158, 'learning_rate/qwen_vl_interface': 9.999840482712069e-06, 'learning_rate/action_model': 9.999840482712068e-05, 'epoch': 0.05}) +[2026-06-17 17:00:02,707][starVLA.training.train_starvla][INFO] - Step 119, Loss: {'train/loss': 1.006898283958435, 'train/grad_norm_pre_clip': 2.5128681659698486, 'timing/data': 0.029418179765343666, 'timing/model': 0.9320771303027868, 'learning_rate/qwen_vl_interface': 9.99982226634526e-06, 'learning_rate/action_model': 9.99982226634526e-05, 'epoch': 0.05}) +[2026-06-17 17:00:06,797][starVLA.training.train_starvla][INFO] - Step 120, Loss: {'train/loss': 1.1503492593765259, 'train/grad_norm_pre_clip': 3.9576103687286377, 'timing/data': 0.02746611088514328, 'timing/model': 1.0174946915358305, 'learning_rate/qwen_vl_interface': 9.999803065335185e-06, 'learning_rate/action_model': 9.999803065335185e-05, 'epoch': 0.05}) +[2026-06-17 17:00:10,750][starVLA.training.train_starvla][INFO] - Step 121, Loss: {'train/loss': 1.204123616218567, 'train/grad_norm_pre_clip': 26.321571350097656, 'timing/data': 0.03637769725173712, 'timing/model': 0.933509198948741, 'learning_rate/qwen_vl_interface': 9.999782879685825e-06, 'learning_rate/action_model': 9.999782879685824e-05, 'epoch': 0.05}) +[2026-06-17 17:00:14,624][starVLA.training.train_starvla][INFO] - Step 122, Loss: {'train/loss': 1.1878794431686401, 'train/grad_norm_pre_clip': 16.720489501953125, 'timing/data': 0.02733741607517004, 'timing/model': 0.9288881989195943, 'learning_rate/qwen_vl_interface': 9.99976170940136e-06, 'learning_rate/action_model': 9.99976170940136e-05, 'epoch': 0.05}) +[2026-06-17 17:00:18,381][starVLA.training.train_starvla][INFO] - Step 123, Loss: {'train/loss': 0.997840940952301, 'train/grad_norm_pre_clip': 4.1902241706848145, 'timing/data': 0.03717606421560049, 'timing/model': 0.9176028119400144, 'learning_rate/qwen_vl_interface': 9.999739554486184e-06, 'learning_rate/action_model': 9.999739554486182e-05, 'epoch': 0.05}) +[2026-06-17 17:00:22,321][starVLA.training.train_starvla][INFO] - Step 124, Loss: {'train/loss': 1.0875290632247925, 'train/grad_norm_pre_clip': 7.044910430908203, 'timing/data': 0.02732889261096716, 'timing/model': 1.2749726073816419, 'learning_rate/qwen_vl_interface': 9.999716414944887e-06, 'learning_rate/action_model': 9.999716414944886e-05, 'epoch': 0.05}) +[2026-06-17 17:00:26,231][starVLA.training.train_starvla][INFO] - Step 125, Loss: {'train/loss': 1.0651607513427734, 'train/grad_norm_pre_clip': 3.818603515625, 'timing/data': 0.03311284724622965, 'timing/model': 1.1797785377129912, 'learning_rate/qwen_vl_interface': 9.999692290782264e-06, 'learning_rate/action_model': 9.999692290782263e-05, 'epoch': 0.05}) +[2026-06-17 17:00:29,837][starVLA.training.train_starvla][INFO] - Step 126, Loss: {'train/loss': 0.8929380774497986, 'train/grad_norm_pre_clip': 3.7832136154174805, 'timing/data': 0.02732514962553978, 'timing/model': 0.9339259276166558, 'learning_rate/qwen_vl_interface': 9.999667182003321e-06, 'learning_rate/action_model': 9.999667182003321e-05, 'epoch': 0.05}) +[2026-06-17 17:00:33,633][starVLA.training.train_starvla][INFO] - Step 127, Loss: {'train/loss': 1.129773736000061, 'train/grad_norm_pre_clip': 3.3831093311309814, 'timing/data': 0.029473052360117435, 'timing/model': 0.9359773164615035, 'learning_rate/qwen_vl_interface': 9.99964108861326e-06, 'learning_rate/action_model': 9.99964108861326e-05, 'epoch': 0.05}) +[2026-06-17 17:00:37,483][starVLA.training.train_starvla][INFO] - Step 128, Loss: {'train/loss': 1.2002336978912354, 'train/grad_norm_pre_clip': 4.675418853759766, 'timing/data': 0.024057692848145962, 'timing/model': 0.9395097279921174, 'learning_rate/qwen_vl_interface': 9.99961401061749e-06, 'learning_rate/action_model': 9.99961401061749e-05, 'epoch': 0.05}) +[2026-06-17 17:00:41,274][starVLA.training.train_starvla][INFO] - Step 129, Loss: {'train/loss': 0.9460384249687195, 'train/grad_norm_pre_clip': 2.6955697536468506, 'timing/data': 0.04137038718909025, 'timing/model': 0.9214792819693685, 'learning_rate/qwen_vl_interface': 9.999585948021623e-06, 'learning_rate/action_model': 9.999585948021624e-05, 'epoch': 0.05}) +[2026-06-17 17:00:45,031][starVLA.training.train_starvla][INFO] - Step 130, Loss: {'train/loss': 0.9410619139671326, 'train/grad_norm_pre_clip': 3.76922607421875, 'timing/data': 0.043185354210436344, 'timing/model': 0.9307640055194497, 'learning_rate/qwen_vl_interface': 9.99955690083148e-06, 'learning_rate/action_model': 9.99955690083148e-05, 'epoch': 0.05}) +[2026-06-17 17:00:48,939][starVLA.training.train_starvla][INFO] - Step 131, Loss: {'train/loss': 0.9665225148200989, 'train/grad_norm_pre_clip': 4.254268169403076, 'timing/data': 0.03522156551480293, 'timing/model': 1.1762445848435163, 'learning_rate/qwen_vl_interface': 9.99952686905308e-06, 'learning_rate/action_model': 9.999526869053081e-05, 'epoch': 0.05}) +[2026-06-17 17:00:52,659][starVLA.training.train_starvla][INFO] - Step 132, Loss: {'train/loss': 1.0618094205856323, 'train/grad_norm_pre_clip': 3.4135634899139404, 'timing/data': 0.028564725071191788, 'timing/model': 0.9604992438107729, 'learning_rate/qwen_vl_interface': 9.99949585269265e-06, 'learning_rate/action_model': 9.99949585269265e-05, 'epoch': 0.05}) +[2026-06-17 17:00:56,611][starVLA.training.train_starvla][INFO] - Step 133, Loss: {'train/loss': 1.358404517173767, 'train/grad_norm_pre_clip': 5.929444789886475, 'timing/data': 0.03466242924332619, 'timing/model': 0.9379665059968829, 'learning_rate/qwen_vl_interface': 9.99946385175662e-06, 'learning_rate/action_model': 9.999463851756619e-05, 'epoch': 0.05}) +[2026-06-17 17:01:00,365][starVLA.training.train_starvla][INFO] - Step 134, Loss: {'train/loss': 0.7487068772315979, 'train/grad_norm_pre_clip': 9.994955062866211, 'timing/data': 0.04245682153850794, 'timing/model': 0.9219269454479218, 'learning_rate/qwen_vl_interface': 9.999430866251623e-06, 'learning_rate/action_model': 9.999430866251622e-05, 'epoch': 0.06}) +[2026-06-17 17:01:04,171][starVLA.training.train_starvla][INFO] - Step 135, Loss: {'train/loss': 1.0896178483963013, 'train/grad_norm_pre_clip': 5.91492223739624, 'timing/data': 0.03175914287567139, 'timing/model': 0.9232125654816628, 'learning_rate/qwen_vl_interface': 9.999396896184494e-06, 'learning_rate/action_model': 9.999396896184494e-05, 'epoch': 0.06}) +[2026-06-17 17:01:08,169][starVLA.training.train_starvla][INFO] - Step 136, Loss: {'train/loss': 0.9649459719657898, 'train/grad_norm_pre_clip': 5.565429210662842, 'timing/data': 0.037252502515912056, 'timing/model': 0.9419841840863228, 'learning_rate/qwen_vl_interface': 9.99936194156228e-06, 'learning_rate/action_model': 9.99936194156228e-05, 'epoch': 0.06}) +[2026-06-17 17:01:12,367][starVLA.training.train_starvla][INFO] - Step 137, Loss: {'train/loss': 1.058288812637329, 'train/grad_norm_pre_clip': 4.831766605377197, 'timing/data': 0.0276834387332201, 'timing/model': 1.172334817238152, 'learning_rate/qwen_vl_interface': 9.999326002392225e-06, 'learning_rate/action_model': 9.999326002392224e-05, 'epoch': 0.06}) +[2026-06-17 17:01:16,138][starVLA.training.train_starvla][INFO] - Step 138, Loss: {'train/loss': 1.0352579355239868, 'train/grad_norm_pre_clip': 5.542057991027832, 'timing/data': 0.02769668959081173, 'timing/model': 1.0838880250230432, 'learning_rate/qwen_vl_interface': 9.999289078681779e-06, 'learning_rate/action_model': 9.999289078681778e-05, 'epoch': 0.06}) +[2026-06-17 17:01:19,821][starVLA.training.train_starvla][INFO] - Step 139, Loss: {'train/loss': 1.408981204032898, 'train/grad_norm_pre_clip': 13.296619415283203, 'timing/data': 0.034401776269078255, 'timing/model': 0.9493987979367375, 'learning_rate/qwen_vl_interface': 9.999251170438594e-06, 'learning_rate/action_model': 9.999251170438594e-05, 'epoch': 0.06}) +[2026-06-17 17:01:23,681][starVLA.training.train_starvla][INFO] - Step 140, Loss: {'train/loss': 0.959984540939331, 'train/grad_norm_pre_clip': 6.544445991516113, 'timing/data': 0.030213911086320877, 'timing/model': 0.9502377258613706, 'learning_rate/qwen_vl_interface': 9.999212277670533e-06, 'learning_rate/action_model': 9.999212277670533e-05, 'epoch': 0.06}) +[2026-06-17 17:01:27,515][starVLA.training.train_starvla][INFO] - Step 141, Loss: {'train/loss': 1.3343138694763184, 'train/grad_norm_pre_clip': 7.823055267333984, 'timing/data': 0.04029128234833479, 'timing/model': 0.9301185803487897, 'learning_rate/qwen_vl_interface': 9.999172400385658e-06, 'learning_rate/action_model': 9.999172400385658e-05, 'epoch': 0.06}) +[2026-06-17 17:01:31,321][starVLA.training.train_starvla][INFO] - Step 142, Loss: {'train/loss': 0.8532314300537109, 'train/grad_norm_pre_clip': 4.148621082305908, 'timing/data': 0.03356822766363621, 'timing/model': 0.9265033081173897, 'learning_rate/qwen_vl_interface': 9.999131538592232e-06, 'learning_rate/action_model': 9.999131538592232e-05, 'epoch': 0.06}) +[2026-06-17 17:01:35,070][starVLA.training.train_starvla][INFO] - Step 143, Loss: {'train/loss': 0.825861394405365, 'train/grad_norm_pre_clip': 5.250245571136475, 'timing/data': 0.028562544845044613, 'timing/model': 0.9368657059967518, 'learning_rate/qwen_vl_interface': 9.999089692298729e-06, 'learning_rate/action_model': 9.99908969229873e-05, 'epoch': 0.06}) +[2026-06-17 17:01:38,977][starVLA.training.train_starvla][INFO] - Step 144, Loss: {'train/loss': 0.9852825999259949, 'train/grad_norm_pre_clip': 6.250959873199463, 'timing/data': 0.03155360370874405, 'timing/model': 1.22412190400064, 'learning_rate/qwen_vl_interface': 9.999046861513822e-06, 'learning_rate/action_model': 9.999046861513821e-05, 'epoch': 0.06}) +[2026-06-17 17:01:42,749][starVLA.training.train_starvla][INFO] - Step 145, Loss: {'train/loss': 1.0275958776474, 'train/grad_norm_pre_clip': 5.160058975219727, 'timing/data': 0.027088706381618977, 'timing/model': 1.1098774774000049, 'learning_rate/qwen_vl_interface': 9.99900304624639e-06, 'learning_rate/action_model': 9.99900304624639e-05, 'epoch': 0.06}) +[2026-06-17 17:01:46,348][starVLA.training.train_starvla][INFO] - Step 146, Loss: {'train/loss': 1.0270488262176514, 'train/grad_norm_pre_clip': 6.231433391571045, 'timing/data': 0.03209032863378525, 'timing/model': 0.953248230740428, 'learning_rate/qwen_vl_interface': 9.998958246505518e-06, 'learning_rate/action_model': 9.998958246505516e-05, 'epoch': 0.06}) +[2026-06-17 17:01:50,270][starVLA.training.train_starvla][INFO] - Step 147, Loss: {'train/loss': 1.410858154296875, 'train/grad_norm_pre_clip': 9.589488983154297, 'timing/data': 0.03657853044569492, 'timing/model': 0.971637949347496, 'learning_rate/qwen_vl_interface': 9.998912462300489e-06, 'learning_rate/action_model': 9.998912462300488e-05, 'epoch': 0.06}) +[2026-06-17 17:01:54,309][starVLA.training.train_starvla][INFO] - Step 148, Loss: {'train/loss': 1.2234948873519897, 'train/grad_norm_pre_clip': 6.944392681121826, 'timing/data': 0.025187211111187935, 'timing/model': 1.1221448052674532, 'learning_rate/qwen_vl_interface': 9.998865693640797e-06, 'learning_rate/action_model': 9.998865693640796e-05, 'epoch': 0.06}) +[2026-06-17 17:01:58,796][starVLA.training.train_starvla][INFO] - Step 149, Loss: {'train/loss': 1.3963509798049927, 'train/grad_norm_pre_clip': 4.750202178955078, 'timing/data': 0.027415617369115353, 'timing/model': 1.6018994888290763, 'learning_rate/qwen_vl_interface': 9.998817940536138e-06, 'learning_rate/action_model': 9.998817940536137e-05, 'epoch': 0.06}) +[2026-06-17 17:02:02,626][starVLA.training.train_starvla][INFO] - Step 150, Loss: {'train/loss': 1.1611618995666504, 'train/grad_norm_pre_clip': 3.820514678955078, 'timing/data': 0.025406479835510254, 'timing/model': 0.9436605246737599, 'learning_rate/qwen_vl_interface': 9.99876920299641e-06, 'learning_rate/action_model': 9.998769202996408e-05, 'epoch': 0.06}) +[2026-06-17 17:02:06,427][starVLA.training.train_starvla][INFO] - Step 151, Loss: {'train/loss': 1.1981970071792603, 'train/grad_norm_pre_clip': 3.508578062057495, 'timing/data': 0.03485609404742718, 'timing/model': 1.1051854183897376, 'learning_rate/qwen_vl_interface': 9.998719481031715e-06, 'learning_rate/action_model': 9.998719481031715e-05, 'epoch': 0.06}) +[2026-06-17 17:02:10,223][starVLA.training.train_starvla][INFO] - Step 152, Loss: {'train/loss': 1.0401201248168945, 'train/grad_norm_pre_clip': 6.039717197418213, 'timing/data': 0.023975325748324394, 'timing/model': 1.1104478677734733, 'learning_rate/qwen_vl_interface': 9.998668774652363e-06, 'learning_rate/action_model': 9.998668774652362e-05, 'epoch': 0.06}) +[2026-06-17 17:02:13,806][starVLA.training.train_starvla][INFO] - Step 153, Loss: {'train/loss': 0.992409348487854, 'train/grad_norm_pre_clip': 5.38098669052124, 'timing/data': 0.027298325672745705, 'timing/model': 0.9243744676932693, 'learning_rate/qwen_vl_interface': 9.998617083868863e-06, 'learning_rate/action_model': 9.998617083868863e-05, 'epoch': 0.06}) +[2026-06-17 17:02:17,558][starVLA.training.train_starvla][INFO] - Step 154, Loss: {'train/loss': 1.0714558362960815, 'train/grad_norm_pre_clip': 4.019418716430664, 'timing/data': 0.026734184473752975, 'timing/model': 0.9239160707220435, 'learning_rate/qwen_vl_interface': 9.998564408691933e-06, 'learning_rate/action_model': 9.998564408691932e-05, 'epoch': 0.06}) +[2026-06-17 17:02:21,345][starVLA.training.train_starvla][INFO] - Step 155, Loss: {'train/loss': 0.8930748701095581, 'train/grad_norm_pre_clip': 11.6248140335083, 'timing/data': 0.033938026055693626, 'timing/model': 0.9383513713255525, 'learning_rate/qwen_vl_interface': 9.998510749132491e-06, 'learning_rate/action_model': 9.998510749132491e-05, 'epoch': 0.06}) +[2026-06-17 17:02:25,234][starVLA.training.train_starvla][INFO] - Step 156, Loss: {'train/loss': 0.985824465751648, 'train/grad_norm_pre_clip': 3.1118130683898926, 'timing/data': 0.025872177444398403, 'timing/model': 0.9433598583564162, 'learning_rate/qwen_vl_interface': 9.998456105201659e-06, 'learning_rate/action_model': 9.99845610520166e-05, 'epoch': 0.06}) +[2026-06-17 17:02:29,142][starVLA.training.train_starvla][INFO] - Step 157, Loss: {'train/loss': 1.3241157531738281, 'train/grad_norm_pre_clip': 5.583334445953369, 'timing/data': 0.026972477324306965, 'timing/model': 1.1654188493266702, 'learning_rate/qwen_vl_interface': 9.99840047691077e-06, 'learning_rate/action_model': 9.99840047691077e-05, 'epoch': 0.06}) +[2026-06-17 17:02:32,909][starVLA.training.train_starvla][INFO] - Step 158, Loss: {'train/loss': 0.7519146800041199, 'train/grad_norm_pre_clip': 4.484188556671143, 'timing/data': 0.02611042745411396, 'timing/model': 1.0460741203278303, 'learning_rate/qwen_vl_interface': 9.998343864271352e-06, 'learning_rate/action_model': 9.998343864271351e-05, 'epoch': 0.07}) +[2026-06-17 17:02:36,567][starVLA.training.train_starvla][INFO] - Step 159, Loss: {'train/loss': 0.8777592182159424, 'train/grad_norm_pre_clip': 7.258093357086182, 'timing/data': 0.034148698672652245, 'timing/model': 0.9386423425748944, 'learning_rate/qwen_vl_interface': 9.99828626729514e-06, 'learning_rate/action_model': 9.998286267295139e-05, 'epoch': 0.07}) +[2026-06-17 17:02:40,424][starVLA.training.train_starvla][INFO] - Step 160, Loss: {'train/loss': 1.1789977550506592, 'train/grad_norm_pre_clip': 5.431256294250488, 'timing/data': 0.024700009264051914, 'timing/model': 0.9308034274727106, 'learning_rate/qwen_vl_interface': 9.998227685994077e-06, 'learning_rate/action_model': 9.998227685994075e-05, 'epoch': 0.07}) +[2026-06-17 17:02:44,232][starVLA.training.train_starvla][INFO] - Step 161, Loss: {'train/loss': 0.795444905757904, 'train/grad_norm_pre_clip': 4.242764472961426, 'timing/data': 0.02590761985629797, 'timing/model': 0.9373730719089508, 'learning_rate/qwen_vl_interface': 9.998168120380303e-06, 'learning_rate/action_model': 9.998168120380303e-05, 'epoch': 0.07}) +[2026-06-17 17:02:48,372][starVLA.training.train_starvla][INFO] - Step 162, Loss: {'train/loss': 1.438750982284546, 'train/grad_norm_pre_clip': 6.1550679206848145, 'timing/data': 0.023935656994581223, 'timing/model': 0.9955261889845133, 'learning_rate/qwen_vl_interface': 9.99810757046617e-06, 'learning_rate/action_model': 9.99810757046617e-05, 'epoch': 0.07}) +[2026-06-17 17:02:52,537][starVLA.training.train_starvla][INFO] - Step 163, Loss: {'train/loss': 1.3288265466690063, 'train/grad_norm_pre_clip': 5.537344455718994, 'timing/data': 0.028156829066574574, 'timing/model': 1.0865609562024474, 'learning_rate/qwen_vl_interface': 9.998046036264229e-06, 'learning_rate/action_model': 9.99804603626423e-05, 'epoch': 0.07}) +[2026-06-17 17:02:57,220][starVLA.training.train_starvla][INFO] - Step 164, Loss: {'train/loss': 1.0957082509994507, 'train/grad_norm_pre_clip': 8.050115585327148, 'timing/data': 0.027714683674275875, 'timing/model': 1.2116469126194715, 'learning_rate/qwen_vl_interface': 9.997983517787236e-06, 'learning_rate/action_model': 9.997983517787236e-05, 'epoch': 0.07}) +[2026-06-17 17:03:00,886][starVLA.training.train_starvla][INFO] - Step 165, Loss: {'train/loss': 0.7671588659286499, 'train/grad_norm_pre_clip': 5.68677282333374, 'timing/data': 0.03185967355966568, 'timing/model': 0.9445857210084796, 'learning_rate/qwen_vl_interface': 9.99792001504815e-06, 'learning_rate/action_model': 9.997920015048149e-05, 'epoch': 0.07}) +[2026-06-17 17:03:05,014][starVLA.training.train_starvla][INFO] - Step 166, Loss: {'train/loss': 1.0236743688583374, 'train/grad_norm_pre_clip': 5.827538967132568, 'timing/data': 0.029927179217338562, 'timing/model': 1.035430840216577, 'learning_rate/qwen_vl_interface': 9.997855528060136e-06, 'learning_rate/action_model': 9.997855528060135e-05, 'epoch': 0.07}) +[2026-06-17 17:03:09,036][starVLA.training.train_starvla][INFO] - Step 167, Loss: {'train/loss': 0.816388726234436, 'train/grad_norm_pre_clip': 4.53543758392334, 'timing/data': 0.028232690878212452, 'timing/model': 0.9727829676121473, 'learning_rate/qwen_vl_interface': 9.997790056836562e-06, 'learning_rate/action_model': 9.997790056836562e-05, 'epoch': 0.07}) +[2026-06-17 17:03:12,877][starVLA.training.train_starvla][INFO] - Step 168, Loss: {'train/loss': 1.1245149374008179, 'train/grad_norm_pre_clip': 9.32034683227539, 'timing/data': 0.023656563833355904, 'timing/model': 0.9428051467984915, 'learning_rate/qwen_vl_interface': 9.997723601390998e-06, 'learning_rate/action_model': 9.997723601390998e-05, 'epoch': 0.07}) +[2026-06-17 17:03:16,860][starVLA.training.train_starvla][INFO] - Step 169, Loss: {'train/loss': 1.2898703813552856, 'train/grad_norm_pre_clip': 5.536014556884766, 'timing/data': 0.02311130240559578, 'timing/model': 0.9461980732157826, 'learning_rate/qwen_vl_interface': 9.997656161737227e-06, 'learning_rate/action_model': 9.997656161737226e-05, 'epoch': 0.07}) +[2026-06-17 17:03:20,753][starVLA.training.train_starvla][INFO] - Step 170, Loss: {'train/loss': 1.026314377784729, 'train/grad_norm_pre_clip': 3.73359751701355, 'timing/data': 0.030232666991651058, 'timing/model': 1.1244305251166224, 'learning_rate/qwen_vl_interface': 9.997587737889222e-06, 'learning_rate/action_model': 9.997587737889221e-05, 'epoch': 0.07}) +[2026-06-17 17:03:24,394][starVLA.training.train_starvla][INFO] - Step 171, Loss: {'train/loss': 0.821370542049408, 'train/grad_norm_pre_clip': 2.802161931991577, 'timing/data': 0.024673432111740112, 'timing/model': 0.9520930889993906, 'learning_rate/qwen_vl_interface': 9.997518329861172e-06, 'learning_rate/action_model': 9.997518329861171e-05, 'epoch': 0.07}) +[2026-06-17 17:03:28,536][starVLA.training.train_starvla][INFO] - Step 172, Loss: {'train/loss': 0.9725300073623657, 'train/grad_norm_pre_clip': 3.442382335662842, 'timing/data': 0.03938248660415411, 'timing/model': 0.9889860088005662, 'learning_rate/qwen_vl_interface': 9.997447937667462e-06, 'learning_rate/action_model': 9.997447937667462e-05, 'epoch': 0.07}) +[2026-06-17 17:03:32,329][starVLA.training.train_starvla][INFO] - Step 173, Loss: {'train/loss': 1.15476393699646, 'train/grad_norm_pre_clip': 2.942708969116211, 'timing/data': 0.03248700778931379, 'timing/model': 0.9409588752314448, 'learning_rate/qwen_vl_interface': 9.997376561322686e-06, 'learning_rate/action_model': 9.997376561322686e-05, 'epoch': 0.07}) +[2026-06-17 17:03:36,163][starVLA.training.train_starvla][INFO] - Step 174, Loss: {'train/loss': 1.0730605125427246, 'train/grad_norm_pre_clip': 6.493447303771973, 'timing/data': 0.02998459432274103, 'timing/model': 0.9404965005815029, 'learning_rate/qwen_vl_interface': 9.997304200841642e-06, 'learning_rate/action_model': 9.997304200841641e-05, 'epoch': 0.07}) +[2026-06-17 17:03:40,065][starVLA.training.train_starvla][INFO] - Step 175, Loss: {'train/loss': 1.1149427890777588, 'train/grad_norm_pre_clip': 5.827326774597168, 'timing/data': 0.023922041058540344, 'timing/model': 0.93258320633322, 'learning_rate/qwen_vl_interface': 9.997230856239327e-06, 'learning_rate/action_model': 9.997230856239327e-05, 'epoch': 0.07}) +[2026-06-17 17:03:43,940][starVLA.training.train_starvla][INFO] - Step 176, Loss: {'train/loss': 1.0035972595214844, 'train/grad_norm_pre_clip': 4.688045024871826, 'timing/data': 0.027134605683386326, 'timing/model': 0.9477875223383307, 'learning_rate/qwen_vl_interface': 9.997156527530948e-06, 'learning_rate/action_model': 9.997156527530948e-05, 'epoch': 0.07}) +[2026-06-17 17:03:47,776][starVLA.training.train_starvla][INFO] - Step 177, Loss: {'train/loss': 0.9152765274047852, 'train/grad_norm_pre_clip': 6.512076377868652, 'timing/data': 0.027628393843770027, 'timing/model': 1.1041180323809385, 'learning_rate/qwen_vl_interface': 9.997081214731915e-06, 'learning_rate/action_model': 9.997081214731914e-05, 'epoch': 0.07}) +[2026-06-17 17:03:51,436][starVLA.training.train_starvla][INFO] - Step 178, Loss: {'train/loss': 0.8769719004631042, 'train/grad_norm_pre_clip': 7.969386100769043, 'timing/data': 0.024203134700655937, 'timing/model': 0.9422246245667338, 'learning_rate/qwen_vl_interface': 9.997004917857834e-06, 'learning_rate/action_model': 9.997004917857834e-05, 'epoch': 0.07}) +[2026-06-17 17:03:55,345][starVLA.training.train_starvla][INFO] - Step 179, Loss: {'train/loss': 0.8915597796440125, 'train/grad_norm_pre_clip': 3.288235902786255, 'timing/data': 0.029705369845032692, 'timing/model': 0.9526590090245008, 'learning_rate/qwen_vl_interface': 9.996927636924529e-06, 'learning_rate/action_model': 9.996927636924528e-05, 'epoch': 0.07}) +[2026-06-17 17:03:59,164][starVLA.training.train_starvla][INFO] - Step 180, Loss: {'train/loss': 0.6484648585319519, 'train/grad_norm_pre_clip': 2.3151695728302, 'timing/data': 0.032404630444943905, 'timing/model': 0.9281425829976797, 'learning_rate/qwen_vl_interface': 9.996849371948016e-06, 'learning_rate/action_model': 9.996849371948016e-05, 'epoch': 0.07}) +[2026-06-17 17:04:03,410][starVLA.training.train_starvla][INFO] - Step 181, Loss: {'train/loss': 1.1629959344863892, 'train/grad_norm_pre_clip': 5.503312110900879, 'timing/data': 0.029106310568749905, 'timing/model': 1.3292785687372088, 'learning_rate/qwen_vl_interface': 9.996770122944518e-06, 'learning_rate/action_model': 9.996770122944518e-05, 'epoch': 0.07}) +[2026-06-17 17:04:07,379][starVLA.training.train_starvla][INFO] - Step 182, Loss: {'train/loss': 1.0290441513061523, 'train/grad_norm_pre_clip': 5.416391372680664, 'timing/data': 0.030479551292955875, 'timing/model': 1.017916640266776, 'learning_rate/qwen_vl_interface': 9.99668988993047e-06, 'learning_rate/action_model': 9.99668988993047e-05, 'epoch': 0.07}) +[2026-06-17 17:04:11,348][starVLA.training.train_starvla][INFO] - Step 183, Loss: {'train/loss': 0.6880783438682556, 'train/grad_norm_pre_clip': 3.034147262573242, 'timing/data': 0.03319822531193495, 'timing/model': 1.1752590602263808, 'learning_rate/qwen_vl_interface': 9.996608672922496e-06, 'learning_rate/action_model': 9.996608672922496e-05, 'epoch': 0.08}) +[2026-06-17 17:04:15,195][starVLA.training.train_starvla][INFO] - Step 184, Loss: {'train/loss': 1.1047557592391968, 'train/grad_norm_pre_clip': 4.253321170806885, 'timing/data': 0.04769880510866642, 'timing/model': 1.1393185332417488, 'learning_rate/qwen_vl_interface': 9.996526471937438e-06, 'learning_rate/action_model': 9.996526471937439e-05, 'epoch': 0.08}) +[2026-06-17 17:04:18,818][starVLA.training.train_starvla][INFO] - Step 185, Loss: {'train/loss': 1.0506821870803833, 'train/grad_norm_pre_clip': 4.374868392944336, 'timing/data': 0.033268761821091175, 'timing/model': 0.9318341678008437, 'learning_rate/qwen_vl_interface': 9.996443286992337e-06, 'learning_rate/action_model': 9.996443286992336e-05, 'epoch': 0.08}) +[2026-06-17 17:04:22,687][starVLA.training.train_starvla][INFO] - Step 186, Loss: {'train/loss': 1.1555919647216797, 'train/grad_norm_pre_clip': 9.654940605163574, 'timing/data': 0.027883240021765232, 'timing/model': 0.9286302588880062, 'learning_rate/qwen_vl_interface': 9.996359118104432e-06, 'learning_rate/action_model': 9.996359118104432e-05, 'epoch': 0.08}) +[2026-06-17 17:04:26,538][starVLA.training.train_starvla][INFO] - Step 187, Loss: {'train/loss': 1.106263279914856, 'train/grad_norm_pre_clip': 5.682295322418213, 'timing/data': 0.031046603806316853, 'timing/model': 0.9361033020541072, 'learning_rate/qwen_vl_interface': 9.996273965291177e-06, 'learning_rate/action_model': 9.996273965291176e-05, 'epoch': 0.08}) +[2026-06-17 17:04:30,333][starVLA.training.train_starvla][INFO] - Step 188, Loss: {'train/loss': 1.2738279104232788, 'train/grad_norm_pre_clip': 5.570979595184326, 'timing/data': 0.045827596448361874, 'timing/model': 0.9294722825288773, 'learning_rate/qwen_vl_interface': 9.996187828570221e-06, 'learning_rate/action_model': 9.996187828570221e-05, 'epoch': 0.08}) +[2026-06-17 17:04:34,252][starVLA.training.train_starvla][INFO] - Step 189, Loss: {'train/loss': 0.9462074637413025, 'train/grad_norm_pre_clip': 8.385214805603027, 'timing/data': 0.03287522681057453, 'timing/model': 0.9291642066091299, 'learning_rate/qwen_vl_interface': 9.996100707959422e-06, 'learning_rate/action_model': 9.996100707959421e-05, 'epoch': 0.08}) +[2026-06-17 17:04:38,162][starVLA.training.train_starvla][INFO] - Step 190, Loss: {'train/loss': 0.8811458945274353, 'train/grad_norm_pre_clip': 4.9179205894470215, 'timing/data': 0.026173428632318974, 'timing/model': 1.2122866911813617, 'learning_rate/qwen_vl_interface': 9.996012603476836e-06, 'learning_rate/action_model': 9.996012603476837e-05, 'epoch': 0.08}) +[2026-06-17 17:04:41,904][starVLA.training.train_starvla][INFO] - Step 191, Loss: {'train/loss': 0.8780217170715332, 'train/grad_norm_pre_clip': 3.481311798095703, 'timing/data': 0.03132582176476717, 'timing/model': 1.0787652656435966, 'learning_rate/qwen_vl_interface': 9.995923515140734e-06, 'learning_rate/action_model': 9.995923515140732e-05, 'epoch': 0.08}) +[2026-06-17 17:04:45,508][starVLA.training.train_starvla][INFO] - Step 192, Loss: {'train/loss': 1.199061632156372, 'train/grad_norm_pre_clip': 4.81370210647583, 'timing/data': 0.023042072542011738, 'timing/model': 0.9221781454980373, 'learning_rate/qwen_vl_interface': 9.995833442969578e-06, 'learning_rate/action_model': 9.995833442969578e-05, 'epoch': 0.08}) +[2026-06-17 17:04:49,325][starVLA.training.train_starvla][INFO] - Step 193, Loss: {'train/loss': 0.7067267894744873, 'train/grad_norm_pre_clip': 3.2204396724700928, 'timing/data': 0.03308019880205393, 'timing/model': 0.9306658953428268, 'learning_rate/qwen_vl_interface': 9.995742386982045e-06, 'learning_rate/action_model': 9.995742386982045e-05, 'epoch': 0.08}) +[2026-06-17 17:04:53,133][starVLA.training.train_starvla][INFO] - Step 194, Loss: {'train/loss': 0.9975900650024414, 'train/grad_norm_pre_clip': 4.601963043212891, 'timing/data': 0.025073451921343803, 'timing/model': 0.9309687064960599, 'learning_rate/qwen_vl_interface': 9.995650347197006e-06, 'learning_rate/action_model': 9.995650347197005e-05, 'epoch': 0.08}) +[2026-06-17 17:04:56,925][starVLA.training.train_starvla][INFO] - Step 195, Loss: {'train/loss': 0.7625583410263062, 'train/grad_norm_pre_clip': 5.228476047515869, 'timing/data': 0.03891703113913536, 'timing/model': 0.9331238083541393, 'learning_rate/qwen_vl_interface': 9.995557323633544e-06, 'learning_rate/action_model': 9.995557323633543e-05, 'epoch': 0.08}) +[2026-06-17 17:05:00,731][starVLA.training.train_starvla][INFO] - Step 196, Loss: {'train/loss': 0.7331489324569702, 'train/grad_norm_pre_clip': 3.5098938941955566, 'timing/data': 0.03331869747489691, 'timing/model': 0.9377709254622459, 'learning_rate/qwen_vl_interface': 9.995463316310942e-06, 'learning_rate/action_model': 9.995463316310943e-05, 'epoch': 0.08}) +[2026-06-17 17:05:04,643][starVLA.training.train_starvla][INFO] - Step 197, Loss: {'train/loss': 1.0954182147979736, 'train/grad_norm_pre_clip': 7.018911361694336, 'timing/data': 0.03150777705013752, 'timing/model': 1.1822997387498617, 'learning_rate/qwen_vl_interface': 9.995368325248688e-06, 'learning_rate/action_model': 9.995368325248688e-05, 'epoch': 0.08}) +[2026-06-17 17:05:08,299][starVLA.training.train_starvla][INFO] - Step 198, Loss: {'train/loss': 0.9589937925338745, 'train/grad_norm_pre_clip': 4.896287441253662, 'timing/data': 0.02984762191772461, 'timing/model': 0.9443986527621746, 'learning_rate/qwen_vl_interface': 9.995272350466476e-06, 'learning_rate/action_model': 9.995272350466476e-05, 'epoch': 0.08}) +[2026-06-17 17:05:12,202][starVLA.training.train_starvla][INFO] - Step 199, Loss: {'train/loss': 1.112766146659851, 'train/grad_norm_pre_clip': 6.77903413772583, 'timing/data': 0.021632248535752296, 'timing/model': 0.9410506170243025, 'learning_rate/qwen_vl_interface': 9.995175391984199e-06, 'learning_rate/action_model': 9.995175391984198e-05, 'epoch': 0.08}) +[2026-06-17 17:07:16,313][starVLA.training.train_starvla][INFO] - Step 200, Loss: {'train/loss': 1.6827484369277954, 'train/grad_norm_pre_clip': 10.940857887268066, 'eval/loss': 1.031978964805603, 'eval/latency_6/loss': 1.031978964805603, 'eval/demon_fire_precision': 0.9009960760639903, 'eval/demon_fire_recall': 0.964770523594053, 'eval/demon_fire_f1': 0.9317933510223193, 'eval/demon_fire_rate_model': 0.9218141346688926, 'eval/demon_fire_rate_teacher': 0.860879243183083, 'eval/demon_left_precision': 0.5435540069686411, 'eval/demon_left_recall': 0.2736842105263158, 'eval/demon_left_f1': 0.36406067677946324, 'eval/demon_left_rate_model': 0.07985531441291041, 'eval/demon_left_rate_teacher': 0.15859766277128548, 'eval/demon_right_precision': 0.698744769874477, 'eval/demon_right_recall': 0.2565284178187404, 'eval/demon_right_f1': 0.3752808988764045, 'eval/demon_right_rate_model': 0.06649972175848637, 'eval/demon_right_rate_teacher': 0.18113522537562604, 'eval/demon_move_f1': 0.36967078782793383, 'eval/demon_cc_f1': 0.6398884389097065, 'eval/latency_6/demon_fire_precision': 0.9009960760639903, 'eval/latency_6/demon_fire_recall': 0.964770523594053, 'eval/latency_6/demon_fire_f1': 0.9317933510223193, 'eval/latency_6/demon_fire_rate_model': 0.9218141346688926, 'eval/latency_6/demon_fire_rate_teacher': 0.860879243183083, 'eval/latency_6/demon_left_precision': 0.5435540069686411, 'eval/latency_6/demon_left_recall': 0.2736842105263158, 'eval/latency_6/demon_left_f1': 0.36406067677946324, 'eval/latency_6/demon_left_rate_model': 0.07985531441291041, 'eval/latency_6/demon_left_rate_teacher': 0.15859766277128548, 'eval/latency_6/demon_right_precision': 0.698744769874477, 'eval/latency_6/demon_right_recall': 0.2565284178187404, 'eval/latency_6/demon_right_f1': 0.3752808988764045, 'eval/latency_6/demon_right_rate_model': 0.06649972175848637, 'eval/latency_6/demon_right_rate_teacher': 0.18113522537562604, 'eval/latency_6/demon_move_f1': 0.36967078782793383, 'eval/latency_6/demon_cc_f1': 0.6398884389097065, 'timing/data': 0.026903929188847542, 'timing/model': 0.960696917027235, 'learning_rate/qwen_vl_interface': 9.995077449821957e-06, 'learning_rate/action_model': 9.995077449821956e-05, 'epoch': 0.08}) +[2026-06-17 17:07:16,325][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_200_state +[2026-06-17 17:07:16,329][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 17:07:23,060][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_200_state/model.safetensors +[2026-06-17 17:07:38,832][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_200_state/optimizer.bin +[2026-06-17 17:07:38,833][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_200_state/sampler.bin +[2026-06-17 17:07:38,833][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_200_state/sampler_1.bin +[2026-06-17 17:07:38,836][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_200_state/random_states_0.pkl +[2026-06-17 17:07:38,844][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 200 +[2026-06-17 17:07:38,844][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 17:07:38,856][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 17:07:42,755][starVLA.training.train_starvla][INFO] - Step 201, Loss: {'train/loss': 1.2037951946258545, 'train/grad_norm_pre_clip': 6.974680423736572, 'timing/data': 0.029101855121552944, 'timing/model': 0.9316193582490087, 'learning_rate/qwen_vl_interface': 9.994978524000053e-06, 'learning_rate/action_model': 9.994978524000052e-05, 'epoch': 0.08}) +[2026-06-17 17:07:46,717][starVLA.training.train_starvla][INFO] - Step 202, Loss: {'train/loss': 1.283625841140747, 'train/grad_norm_pre_clip': 6.1625823974609375, 'timing/data': 0.025731858797371387, 'timing/model': 0.9338127886876464, 'learning_rate/qwen_vl_interface': 9.994878614538996e-06, 'learning_rate/action_model': 9.994878614538996e-05, 'epoch': 0.08}) +[2026-06-17 17:07:50,575][starVLA.training.train_starvla][INFO] - Step 203, Loss: {'train/loss': 1.175360918045044, 'train/grad_norm_pre_clip': 5.354290962219238, 'timing/data': 0.025205468758940697, 'timing/model': 0.9329278348013759, 'learning_rate/qwen_vl_interface': 9.994777721459496e-06, 'learning_rate/action_model': 9.994777721459497e-05, 'epoch': 0.08}) +[2026-06-17 17:07:54,446][starVLA.training.train_starvla][INFO] - Step 204, Loss: {'train/loss': 0.9465813636779785, 'train/grad_norm_pre_clip': 7.627247333526611, 'timing/data': 0.028025085106492043, 'timing/model': 0.9272393686696887, 'learning_rate/qwen_vl_interface': 9.994675844782468e-06, 'learning_rate/action_model': 9.994675844782468e-05, 'epoch': 0.08}) +[2026-06-17 17:07:58,252][starVLA.training.train_starvla][INFO] - Step 205, Loss: {'train/loss': 0.9130039811134338, 'train/grad_norm_pre_clip': 5.112088680267334, 'timing/data': 0.025770134292542934, 'timing/model': 0.9529949612915516, 'learning_rate/qwen_vl_interface': 9.994572984529033e-06, 'learning_rate/action_model': 9.994572984529034e-05, 'epoch': 0.08}) +[2026-06-17 17:08:02,186][starVLA.training.train_starvla][INFO] - Step 206, Loss: {'train/loss': 1.0054094791412354, 'train/grad_norm_pre_clip': 3.365388870239258, 'timing/data': 0.0317652877420187, 'timing/model': 1.192192348651588, 'learning_rate/qwen_vl_interface': 9.994469140720513e-06, 'learning_rate/action_model': 9.994469140720512e-05, 'epoch': 0.08}) +[2026-06-17 17:08:06,013][starVLA.training.train_starvla][INFO] - Step 207, Loss: {'train/loss': 1.4865140914916992, 'train/grad_norm_pre_clip': 5.836053371429443, 'timing/data': 0.03197195753455162, 'timing/model': 0.9314821884036064, 'learning_rate/qwen_vl_interface': 9.994364313378435e-06, 'learning_rate/action_model': 9.994364313378433e-05, 'epoch': 0.09}) +[2026-06-17 17:08:09,950][starVLA.training.train_starvla][INFO] - Step 208, Loss: {'train/loss': 1.0239747762680054, 'train/grad_norm_pre_clip': 3.7393670082092285, 'timing/data': 0.028667454607784748, 'timing/model': 0.9667763048782945, 'learning_rate/qwen_vl_interface': 9.99425850252453e-06, 'learning_rate/action_model': 9.994258502524529e-05, 'epoch': 0.09}) +[2026-06-17 17:08:13,770][starVLA.training.train_starvla][INFO] - Step 209, Loss: {'train/loss': 1.3980116844177246, 'train/grad_norm_pre_clip': 5.102783679962158, 'timing/data': 0.023747370578348637, 'timing/model': 0.9314188715070486, 'learning_rate/qwen_vl_interface': 9.994151708180731e-06, 'learning_rate/action_model': 9.99415170818073e-05, 'epoch': 0.09}) +[2026-06-17 17:08:17,598][starVLA.training.train_starvla][INFO] - Step 210, Loss: {'train/loss': 1.563408613204956, 'train/grad_norm_pre_clip': 6.684407711029053, 'timing/data': 0.03984505496919155, 'timing/model': 0.9206528989598155, 'learning_rate/qwen_vl_interface': 9.994043930369179e-06, 'learning_rate/action_model': 9.994043930369178e-05, 'epoch': 0.09}) +[2026-06-17 17:08:21,696][starVLA.training.train_starvla][INFO] - Step 211, Loss: {'train/loss': 1.1898431777954102, 'train/grad_norm_pre_clip': 5.190353870391846, 'timing/data': 0.0425297487527132, 'timing/model': 0.9283462651073933, 'learning_rate/qwen_vl_interface': 9.993935169112213e-06, 'learning_rate/action_model': 9.993935169112213e-05, 'epoch': 0.09}) +[2026-06-17 17:08:25,667][starVLA.training.train_starvla][INFO] - Step 212, Loss: {'train/loss': 1.0972384214401245, 'train/grad_norm_pre_clip': 3.4612395763397217, 'timing/data': 0.028128876350820065, 'timing/model': 0.9348646914586425, 'learning_rate/qwen_vl_interface': 9.993825424432385e-06, 'learning_rate/action_model': 9.993825424432385e-05, 'epoch': 0.09}) +[2026-06-17 17:08:29,443][starVLA.training.train_starvla][INFO] - Step 213, Loss: {'train/loss': 1.0329020023345947, 'train/grad_norm_pre_clip': 6.98179817199707, 'timing/data': 0.02360984031111002, 'timing/model': 1.0935327988117933, 'learning_rate/qwen_vl_interface': 9.99371469635244e-06, 'learning_rate/action_model': 9.99371469635244e-05, 'epoch': 0.09}) +[2026-06-17 17:08:33,103][starVLA.training.train_starvla][INFO] - Step 214, Loss: {'train/loss': 0.8268360495567322, 'train/grad_norm_pre_clip': 4.026712894439697, 'timing/data': 0.03258915804326534, 'timing/model': 0.9338419651612639, 'learning_rate/qwen_vl_interface': 9.993602984895336e-06, 'learning_rate/action_model': 9.993602984895335e-05, 'epoch': 0.09}) +[2026-06-17 17:08:37,008][starVLA.training.train_starvla][INFO] - Step 215, Loss: {'train/loss': 0.9602454900741577, 'train/grad_norm_pre_clip': 3.121676445007324, 'timing/data': 0.030507463961839676, 'timing/model': 1.0167747866362333, 'learning_rate/qwen_vl_interface': 9.993490290084225e-06, 'learning_rate/action_model': 9.993490290084225e-05, 'epoch': 0.09}) +[2026-06-17 17:08:40,739][starVLA.training.train_starvla][INFO] - Step 216, Loss: {'train/loss': 0.9325868487358093, 'train/grad_norm_pre_clip': 3.6289684772491455, 'timing/data': 0.023645696230232716, 'timing/model': 0.9398933406919241, 'learning_rate/qwen_vl_interface': 9.993376611942477e-06, 'learning_rate/action_model': 9.993376611942477e-05, 'epoch': 0.09}) +[2026-06-17 17:08:45,923][starVLA.training.train_starvla][INFO] - Step 217, Loss: {'train/loss': 0.9976270794868469, 'train/grad_norm_pre_clip': 5.200857162475586, 'timing/data': 0.031219310127198696, 'timing/model': 0.9551633195951581, 'learning_rate/qwen_vl_interface': 9.99326195049365e-06, 'learning_rate/action_model': 9.99326195049365e-05, 'epoch': 0.09}) +[2026-06-17 17:08:50,098][starVLA.training.train_starvla][INFO] - Step 218, Loss: {'train/loss': 0.5756504535675049, 'train/grad_norm_pre_clip': 2.815321207046509, 'timing/data': 0.012697990983724594, 'timing/model': 1.0182519555091858, 'learning_rate/qwen_vl_interface': 9.99314630576152e-06, 'learning_rate/action_model': 9.993146305761519e-05, 'epoch': 0.09}) +[2026-06-17 17:08:53,902][starVLA.training.train_starvla][INFO] - Step 219, Loss: {'train/loss': 0.8533605337142944, 'train/grad_norm_pre_clip': 2.5319149494171143, 'timing/data': 0.022579815238714218, 'timing/model': 0.9315842883661389, 'learning_rate/qwen_vl_interface': 9.993029677770055e-06, 'learning_rate/action_model': 9.993029677770055e-05, 'epoch': 0.09}) +[2026-06-17 17:08:57,700][starVLA.training.train_starvla][INFO] - Step 220, Loss: {'train/loss': 1.64481520652771, 'train/grad_norm_pre_clip': 7.146904945373535, 'timing/data': 0.026484242640435696, 'timing/model': 1.0936817526817322, 'learning_rate/qwen_vl_interface': 9.992912066543433e-06, 'learning_rate/action_model': 9.992912066543433e-05, 'epoch': 0.09}) +[2026-06-17 17:09:01,360][starVLA.training.train_starvla][INFO] - Step 221, Loss: {'train/loss': 0.7852765321731567, 'train/grad_norm_pre_clip': 3.6264090538024902, 'timing/data': 0.035286055877804756, 'timing/model': 0.9421031940728426, 'learning_rate/qwen_vl_interface': 9.992793472106039e-06, 'learning_rate/action_model': 9.992793472106038e-05, 'epoch': 0.09}) +[2026-06-17 17:09:05,286][starVLA.training.train_starvla][INFO] - Step 222, Loss: {'train/loss': 0.7742725610733032, 'train/grad_norm_pre_clip': 3.59533429145813, 'timing/data': 0.024607237428426743, 'timing/model': 0.9864112511277199, 'learning_rate/qwen_vl_interface': 9.992673894482452e-06, 'learning_rate/action_model': 9.992673894482451e-05, 'epoch': 0.09}) +[2026-06-17 17:09:09,357][starVLA.training.train_starvla][INFO] - Step 223, Loss: {'train/loss': 1.492499828338623, 'train/grad_norm_pre_clip': 5.441126346588135, 'timing/data': 0.027484814636409283, 'timing/model': 0.9895581910386682, 'learning_rate/qwen_vl_interface': 9.992553333697464e-06, 'learning_rate/action_model': 9.992553333697463e-05, 'epoch': 0.09}) +[2026-06-17 17:09:14,023][starVLA.training.train_starvla][INFO] - Step 224, Loss: {'train/loss': 1.3006789684295654, 'train/grad_norm_pre_clip': 4.8199076652526855, 'timing/data': 0.023019894026219845, 'timing/model': 1.3876226395368576, 'learning_rate/qwen_vl_interface': 9.992431789776067e-06, 'learning_rate/action_model': 9.992431789776066e-05, 'epoch': 0.09}) +[2026-06-17 17:09:18,101][starVLA.training.train_starvla][INFO] - Step 225, Loss: {'train/loss': 1.3179428577423096, 'train/grad_norm_pre_clip': 6.1672773361206055, 'timing/data': 0.028565519489347935, 'timing/model': 0.9339598976075649, 'learning_rate/qwen_vl_interface': 9.992309262743458e-06, 'learning_rate/action_model': 9.992309262743457e-05, 'epoch': 0.09}) +[2026-06-17 17:09:22,094][starVLA.training.train_starvla][INFO] - Step 226, Loss: {'train/loss': 1.1384837627410889, 'train/grad_norm_pre_clip': 5.004791259765625, 'timing/data': 0.03951134718954563, 'timing/model': 1.2045173346996307, 'learning_rate/qwen_vl_interface': 9.992185752625035e-06, 'learning_rate/action_model': 9.992185752625033e-05, 'epoch': 0.09}) +[2026-06-17 17:09:30,101][starVLA.training.train_starvla][INFO] - Step 227, Loss: {'train/loss': 0.8762181401252747, 'train/grad_norm_pre_clip': 5.274242877960205, 'timing/data': 0.02830960601568222, 'timing/model': 2.529791667126119, 'learning_rate/qwen_vl_interface': 9.992061259446402e-06, 'learning_rate/action_model': 9.992061259446401e-05, 'epoch': 0.09}) +[2026-06-17 17:09:34,312][starVLA.training.train_starvla][INFO] - Step 228, Loss: {'train/loss': 1.0869786739349365, 'train/grad_norm_pre_clip': 4.173563003540039, 'timing/data': 0.03053385019302368, 'timing/model': 1.03826339635998, 'learning_rate/qwen_vl_interface': 9.991935783233368e-06, 'learning_rate/action_model': 9.991935783233367e-05, 'epoch': 0.09}) +[2026-06-17 17:09:38,331][starVLA.training.train_starvla][INFO] - Step 229, Loss: {'train/loss': 1.0988168716430664, 'train/grad_norm_pre_clip': 7.902601718902588, 'timing/data': 0.028437175787985325, 'timing/model': 0.9791965410113335, 'learning_rate/qwen_vl_interface': 9.991809324011943e-06, 'learning_rate/action_model': 9.991809324011943e-05, 'epoch': 0.09}) +[2026-06-17 17:09:42,223][starVLA.training.train_starvla][INFO] - Step 230, Loss: {'train/loss': 1.3523037433624268, 'train/grad_norm_pre_clip': 3.252086877822876, 'timing/data': 0.03392944764345884, 'timing/model': 0.947446939535439, 'learning_rate/qwen_vl_interface': 9.991681881808343e-06, 'learning_rate/action_model': 9.991681881808343e-05, 'epoch': 0.09}) +[2026-06-17 17:09:46,110][starVLA.training.train_starvla][INFO] - Step 231, Loss: {'train/loss': 0.941709041595459, 'train/grad_norm_pre_clip': 4.079599380493164, 'timing/data': 0.025268377736210823, 'timing/model': 1.034801822155714, 'learning_rate/qwen_vl_interface': 9.991553456648986e-06, 'learning_rate/action_model': 9.991553456648985e-05, 'epoch': 0.1}) +[2026-06-17 17:09:50,069][starVLA.training.train_starvla][INFO] - Step 232, Loss: {'train/loss': 1.1038010120391846, 'train/grad_norm_pre_clip': 3.767055034637451, 'timing/data': 0.0296232420951128, 'timing/model': 0.944904956035316, 'learning_rate/qwen_vl_interface': 9.991424048560497e-06, 'learning_rate/action_model': 9.991424048560496e-05, 'epoch': 0.1}) +[2026-06-17 17:09:54,074][starVLA.training.train_starvla][INFO] - Step 233, Loss: {'train/loss': 1.1439452171325684, 'train/grad_norm_pre_clip': 2.9793593883514404, 'timing/data': 0.038911919109523296, 'timing/model': 1.2046818314120173, 'learning_rate/qwen_vl_interface': 9.991293657569698e-06, 'learning_rate/action_model': 9.991293657569698e-05, 'epoch': 0.1}) +[2026-06-17 17:09:57,771][starVLA.training.train_starvla][INFO] - Step 234, Loss: {'train/loss': 1.005284309387207, 'train/grad_norm_pre_clip': 2.864314556121826, 'timing/data': 0.023154753260314465, 'timing/model': 0.9483400080353022, 'learning_rate/qwen_vl_interface': 9.991162283703624e-06, 'learning_rate/action_model': 9.991162283703624e-05, 'epoch': 0.1}) +[2026-06-17 17:10:01,743][starVLA.training.train_starvla][INFO] - Step 235, Loss: {'train/loss': 1.202212929725647, 'train/grad_norm_pre_clip': 4.669708251953125, 'timing/data': 0.02906810399144888, 'timing/model': 0.9409990403801203, 'learning_rate/qwen_vl_interface': 9.991029926989506e-06, 'learning_rate/action_model': 9.991029926989506e-05, 'epoch': 0.1}) +[2026-06-17 17:10:05,557][starVLA.training.train_starvla][INFO] - Step 236, Loss: {'train/loss': 1.250766396522522, 'train/grad_norm_pre_clip': 4.566940784454346, 'timing/data': 0.026173511520028114, 'timing/model': 0.9425790729001164, 'learning_rate/qwen_vl_interface': 9.990896587454783e-06, 'learning_rate/action_model': 9.990896587454783e-05, 'epoch': 0.1}) +[2026-06-17 17:10:09,631][starVLA.training.train_starvla][INFO] - Step 237, Loss: {'train/loss': 1.052626371383667, 'train/grad_norm_pre_clip': 3.995710849761963, 'timing/data': 0.03511663340032101, 'timing/model': 1.0561577640473843, 'learning_rate/qwen_vl_interface': 9.990762265127096e-06, 'learning_rate/action_model': 9.990762265127096e-05, 'epoch': 0.1}) +[2026-06-17 17:10:13,562][starVLA.training.train_starvla][INFO] - Step 238, Loss: {'train/loss': 0.8441969752311707, 'train/grad_norm_pre_clip': 2.0822863578796387, 'timing/data': 0.022823362611234188, 'timing/model': 0.9834181126207113, 'learning_rate/qwen_vl_interface': 9.990626960034291e-06, 'learning_rate/action_model': 9.990626960034291e-05, 'epoch': 0.1}) +[2026-06-17 17:10:17,573][starVLA.training.train_starvla][INFO] - Step 239, Loss: {'train/loss': 0.9262311458587646, 'train/grad_norm_pre_clip': 3.056568145751953, 'timing/data': 0.027741076424717903, 'timing/model': 1.165222325362265, 'learning_rate/qwen_vl_interface': 9.990490672204417e-06, 'learning_rate/action_model': 9.990490672204415e-05, 'epoch': 0.1}) +[2026-06-17 17:10:21,510][starVLA.training.train_starvla][INFO] - Step 240, Loss: {'train/loss': 0.9738989472389221, 'train/grad_norm_pre_clip': 3.4195854663848877, 'timing/data': 0.02467017900198698, 'timing/model': 1.094340237788856, 'learning_rate/qwen_vl_interface': 9.99035340166572e-06, 'learning_rate/action_model': 9.990353401665721e-05, 'epoch': 0.1}) +[2026-06-17 17:10:25,221][starVLA.training.train_starvla][INFO] - Step 241, Loss: {'train/loss': 1.1204184293746948, 'train/grad_norm_pre_clip': 4.386373519897461, 'timing/data': 0.026222549378871918, 'timing/model': 1.0017456449568272, 'learning_rate/qwen_vl_interface': 9.990215148446667e-06, 'learning_rate/action_model': 9.990215148446666e-05, 'epoch': 0.1}) +[2026-06-17 17:10:29,140][starVLA.training.train_starvla][INFO] - Step 242, Loss: {'train/loss': 1.0676888227462769, 'train/grad_norm_pre_clip': 2.970843553543091, 'timing/data': 0.03251369949430227, 'timing/model': 0.9561957018449903, 'learning_rate/qwen_vl_interface': 9.990075912575912e-06, 'learning_rate/action_model': 9.990075912575911e-05, 'epoch': 0.1}) +[2026-06-17 17:10:32,988][starVLA.training.train_starvla][INFO] - Step 243, Loss: {'train/loss': 0.9958487153053284, 'train/grad_norm_pre_clip': 8.418432235717773, 'timing/data': 0.0260080648586154, 'timing/model': 0.9453804465010762, 'learning_rate/qwen_vl_interface': 9.98993569408232e-06, 'learning_rate/action_model': 9.989935694082321e-05, 'epoch': 0.1}) +[2026-06-17 17:10:36,836][starVLA.training.train_starvla][INFO] - Step 244, Loss: {'train/loss': 1.0063807964324951, 'train/grad_norm_pre_clip': 2.887946128845215, 'timing/data': 0.02626472245901823, 'timing/model': 0.9306677663698792, 'learning_rate/qwen_vl_interface': 9.989794492994958e-06, 'learning_rate/action_model': 9.989794492994958e-05, 'epoch': 0.1}) +[2026-06-17 17:10:40,632][starVLA.training.train_starvla][INFO] - Step 245, Loss: {'train/loss': 1.2692806720733643, 'train/grad_norm_pre_clip': 3.767899751663208, 'timing/data': 0.034868789836764336, 'timing/model': 0.9249753877520561, 'learning_rate/qwen_vl_interface': 9.989652309343097e-06, 'learning_rate/action_model': 9.989652309343098e-05, 'epoch': 0.1}) +[2026-06-17 17:10:44,436][starVLA.training.train_starvla][INFO] - Step 246, Loss: {'train/loss': 1.0930652618408203, 'train/grad_norm_pre_clip': 4.8303542137146, 'timing/data': 0.0319575285539031, 'timing/model': 1.1049078460782766, 'learning_rate/qwen_vl_interface': 9.989509143156213e-06, 'learning_rate/action_model': 9.989509143156213e-05, 'epoch': 0.1}) +[2026-06-17 17:10:48,188][starVLA.training.train_starvla][INFO] - Step 247, Loss: {'train/loss': 1.1629916429519653, 'train/grad_norm_pre_clip': 6.565506935119629, 'timing/data': 0.02694634348154068, 'timing/model': 1.093670779839158, 'learning_rate/qwen_vl_interface': 9.989364994463983e-06, 'learning_rate/action_model': 9.989364994463983e-05, 'epoch': 0.1}) +[2026-06-17 17:10:51,829][starVLA.training.train_starvla][INFO] - Step 248, Loss: {'train/loss': 0.9179755449295044, 'train/grad_norm_pre_clip': 5.968934535980225, 'timing/data': 0.034740899689495564, 'timing/model': 0.9359792033210397, 'learning_rate/qwen_vl_interface': 9.989219863296291e-06, 'learning_rate/action_model': 9.989219863296291e-05, 'epoch': 0.1}) +[2026-06-17 17:10:55,792][starVLA.training.train_starvla][INFO] - Step 249, Loss: {'train/loss': 0.8672153353691101, 'train/grad_norm_pre_clip': 3.8184404373168945, 'timing/data': 0.028072461485862732, 'timing/model': 1.0290148090571165, 'learning_rate/qwen_vl_interface': 9.98907374968322e-06, 'learning_rate/action_model': 9.989073749683219e-05, 'epoch': 0.1}) +[2026-06-17 17:10:59,646][starVLA.training.train_starvla][INFO] - Step 250, Loss: {'train/loss': 1.0892819166183472, 'train/grad_norm_pre_clip': 4.955741882324219, 'timing/data': 0.03094958234578371, 'timing/model': 0.9390615522861481, 'learning_rate/qwen_vl_interface': 9.988926653655062e-06, 'learning_rate/action_model': 9.988926653655062e-05, 'epoch': 0.1}) +[2026-06-17 17:11:03,541][starVLA.training.train_starvla][INFO] - Step 251, Loss: {'train/loss': 0.9540948867797852, 'train/grad_norm_pre_clip': 3.8106460571289062, 'timing/data': 0.028336521238088608, 'timing/model': 0.9427300803363323, 'learning_rate/qwen_vl_interface': 9.988778575242309e-06, 'learning_rate/action_model': 9.988778575242309e-05, 'epoch': 0.1}) +[2026-06-17 17:11:07,535][starVLA.training.train_starvla][INFO] - Step 252, Loss: {'train/loss': 0.9652049541473389, 'train/grad_norm_pre_clip': 5.068089962005615, 'timing/data': 0.02789443451911211, 'timing/model': 1.0606211936101317, 'learning_rate/qwen_vl_interface': 9.98862951447566e-06, 'learning_rate/action_model': 9.988629514475659e-05, 'epoch': 0.1}) +[2026-06-17 17:11:11,542][starVLA.training.train_starvla][INFO] - Step 253, Loss: {'train/loss': 0.9392480850219727, 'train/grad_norm_pre_clip': 3.7912819385528564, 'timing/data': 0.031803044490516186, 'timing/model': 1.144265104085207, 'learning_rate/qwen_vl_interface': 9.98847947138601e-06, 'learning_rate/action_model': 9.98847947138601e-05, 'epoch': 0.1}) +[2026-06-17 17:11:15,407][starVLA.training.train_starvla][INFO] - Step 254, Loss: {'train/loss': 1.2414628267288208, 'train/grad_norm_pre_clip': 7.094559669494629, 'timing/data': 0.03314865194261074, 'timing/model': 1.125091795809567, 'learning_rate/qwen_vl_interface': 9.98832844600447e-06, 'learning_rate/action_model': 9.98832844600447e-05, 'epoch': 0.1}) +[2026-06-17 17:11:19,911][starVLA.training.train_starvla][INFO] - Step 255, Loss: {'train/loss': 0.9991640448570251, 'train/grad_norm_pre_clip': 7.093717575073242, 'timing/data': 0.034624937921762466, 'timing/model': 1.0329056261107326, 'learning_rate/qwen_vl_interface': 9.988176438362344e-06, 'learning_rate/action_model': 9.988176438362343e-05, 'epoch': 0.1}) +[2026-06-17 17:11:23,805][starVLA.training.train_starvla][INFO] - Step 256, Loss: {'train/loss': 1.127163052558899, 'train/grad_norm_pre_clip': 6.51407527923584, 'timing/data': 0.031094171106815338, 'timing/model': 0.940671437419951, 'learning_rate/qwen_vl_interface': 9.988023448491143e-06, 'learning_rate/action_model': 9.988023448491143e-05, 'epoch': 0.11}) +[2026-06-17 17:11:27,697][starVLA.training.train_starvla][INFO] - Step 257, Loss: {'train/loss': 1.3201740980148315, 'train/grad_norm_pre_clip': 7.5602312088012695, 'timing/data': 0.026104195043444633, 'timing/model': 0.9759895792230964, 'learning_rate/qwen_vl_interface': 9.987869476422582e-06, 'learning_rate/action_model': 9.987869476422583e-05, 'epoch': 0.11}) +[2026-06-17 17:11:31,686][starVLA.training.train_starvla][INFO] - Step 258, Loss: {'train/loss': 1.1122294664382935, 'train/grad_norm_pre_clip': 6.091132640838623, 'timing/data': 0.032417066395282745, 'timing/model': 0.9589077979326248, 'learning_rate/qwen_vl_interface': 9.987714522188583e-06, 'learning_rate/action_model': 9.987714522188582e-05, 'epoch': 0.11}) +[2026-06-17 17:11:35,587][starVLA.training.train_starvla][INFO] - Step 259, Loss: {'train/loss': 1.1235508918762207, 'train/grad_norm_pre_clip': 4.44393253326416, 'timing/data': 0.023765111342072487, 'timing/model': 1.124673799611628, 'learning_rate/qwen_vl_interface': 9.987558585821264e-06, 'learning_rate/action_model': 9.987558585821263e-05, 'epoch': 0.11}) +[2026-06-17 17:11:39,415][starVLA.training.train_starvla][INFO] - Step 260, Loss: {'train/loss': 1.1556521654129028, 'train/grad_norm_pre_clip': 5.798076629638672, 'timing/data': 0.02823967207223177, 'timing/model': 1.0787082687020302, 'learning_rate/qwen_vl_interface': 9.987401667352952e-06, 'learning_rate/action_model': 9.987401667352951e-05, 'epoch': 0.11}) +[2026-06-17 17:11:43,198][starVLA.training.train_starvla][INFO] - Step 261, Loss: {'train/loss': 1.1705392599105835, 'train/grad_norm_pre_clip': 5.12846565246582, 'timing/data': 0.02903873659670353, 'timing/model': 1.0455234367400408, 'learning_rate/qwen_vl_interface': 9.987243766816175e-06, 'learning_rate/action_model': 9.987243766816176e-05, 'epoch': 0.11}) +[2026-06-17 17:11:47,066][starVLA.training.train_starvla][INFO] - Step 262, Loss: {'train/loss': 1.0745271444320679, 'train/grad_norm_pre_clip': 5.697298526763916, 'timing/data': 0.02322258148342371, 'timing/model': 0.9489177707582712, 'learning_rate/qwen_vl_interface': 9.987084884243671e-06, 'learning_rate/action_model': 9.987084884243672e-05, 'epoch': 0.11}) +[2026-06-17 17:11:50,980][starVLA.training.train_starvla][INFO] - Step 263, Loss: {'train/loss': 1.388623833656311, 'train/grad_norm_pre_clip': 6.7144775390625, 'timing/data': 0.04106101766228676, 'timing/model': 0.92509432323277, 'learning_rate/qwen_vl_interface': 9.986925019668374e-06, 'learning_rate/action_model': 9.986925019668373e-05, 'epoch': 0.11}) +[2026-06-17 17:11:54,880][starVLA.training.train_starvla][INFO] - Step 264, Loss: {'train/loss': 0.9131360650062561, 'train/grad_norm_pre_clip': 5.721681118011475, 'timing/data': 0.03450523689389229, 'timing/model': 0.9211648274213076, 'learning_rate/qwen_vl_interface': 9.98676417312342e-06, 'learning_rate/action_model': 9.986764173123419e-05, 'epoch': 0.11}) +[2026-06-17 17:11:58,765][starVLA.training.train_starvla][INFO] - Step 265, Loss: {'train/loss': 0.9988989233970642, 'train/grad_norm_pre_clip': 5.1879353523254395, 'timing/data': 0.03693228214979172, 'timing/model': 0.9521037293598056, 'learning_rate/qwen_vl_interface': 9.986602344642157e-06, 'learning_rate/action_model': 9.986602344642156e-05, 'epoch': 0.11}) +[2026-06-17 17:12:02,578][starVLA.training.train_starvla][INFO] - Step 266, Loss: {'train/loss': 0.8472131490707397, 'train/grad_norm_pre_clip': 8.250686645507812, 'timing/data': 0.026431672275066376, 'timing/model': 1.1086972644552588, 'learning_rate/qwen_vl_interface': 9.986439534258132e-06, 'learning_rate/action_model': 9.986439534258131e-05, 'epoch': 0.11}) +[2026-06-17 17:12:06,240][starVLA.training.train_starvla][INFO] - Step 267, Loss: {'train/loss': 1.0943751335144043, 'train/grad_norm_pre_clip': 5.9041748046875, 'timing/data': 0.023719999007880688, 'timing/model': 0.9498610375449061, 'learning_rate/qwen_vl_interface': 9.986275742005093e-06, 'learning_rate/action_model': 9.986275742005093e-05, 'epoch': 0.11}) +[2026-06-17 17:12:10,036][starVLA.training.train_starvla][INFO] - Step 268, Loss: {'train/loss': 1.2149978876113892, 'train/grad_norm_pre_clip': 5.797401428222656, 'timing/data': 0.024827820248901844, 'timing/model': 0.9369156388565898, 'learning_rate/qwen_vl_interface': 9.986110967916997e-06, 'learning_rate/action_model': 9.986110967916996e-05, 'epoch': 0.11}) +[2026-06-17 17:12:13,907][starVLA.training.train_starvla][INFO] - Step 269, Loss: {'train/loss': 1.1072323322296143, 'train/grad_norm_pre_clip': 7.897913932800293, 'timing/data': 0.025705316103994846, 'timing/model': 0.9631225718185306, 'learning_rate/qwen_vl_interface': 9.985945212028002e-06, 'learning_rate/action_model': 9.985945212028001e-05, 'epoch': 0.11}) +[2026-06-17 17:12:17,710][starVLA.training.train_starvla][INFO] - Step 270, Loss: {'train/loss': 0.9354102611541748, 'train/grad_norm_pre_clip': 6.138509273529053, 'timing/data': 0.02379935421049595, 'timing/model': 0.9357499536126852, 'learning_rate/qwen_vl_interface': 9.985778474372467e-06, 'learning_rate/action_model': 9.985778474372466e-05, 'epoch': 0.11}) +[2026-06-17 17:12:21,753][starVLA.training.train_starvla][INFO] - Step 271, Loss: {'train/loss': 1.0092095136642456, 'train/grad_norm_pre_clip': 5.7616424560546875, 'timing/data': 0.03362342435866594, 'timing/model': 0.9510182598605752, 'learning_rate/qwen_vl_interface': 9.98561075498496e-06, 'learning_rate/action_model': 9.98561075498496e-05, 'epoch': 0.11}) +[2026-06-17 17:12:25,579][starVLA.training.train_starvla][INFO] - Step 272, Loss: {'train/loss': 1.1122978925704956, 'train/grad_norm_pre_clip': 3.5834927558898926, 'timing/data': 0.033585879020392895, 'timing/model': 1.1652590548619628, 'learning_rate/qwen_vl_interface': 9.985442053900246e-06, 'learning_rate/action_model': 9.985442053900246e-05, 'epoch': 0.11}) +[2026-06-17 17:12:29,247][starVLA.training.train_starvla][INFO] - Step 273, Loss: {'train/loss': 0.8297041654586792, 'train/grad_norm_pre_clip': 2.9244978427886963, 'timing/data': 0.029660871252417564, 'timing/model': 0.939438221976161, 'learning_rate/qwen_vl_interface': 9.9852723711533e-06, 'learning_rate/action_model': 9.9852723711533e-05, 'epoch': 0.11}) +[2026-06-17 17:12:33,346][starVLA.training.train_starvla][INFO] - Step 274, Loss: {'train/loss': 0.9893621206283569, 'train/grad_norm_pre_clip': 3.441561698913574, 'timing/data': 0.030844377353787422, 'timing/model': 0.9290696186944842, 'learning_rate/qwen_vl_interface': 9.985101706779297e-06, 'learning_rate/action_model': 9.985101706779296e-05, 'epoch': 0.11}) +[2026-06-17 17:12:37,473][starVLA.training.train_starvla][INFO] - Step 275, Loss: {'train/loss': 0.9155720472335815, 'train/grad_norm_pre_clip': 3.2377121448516846, 'timing/data': 0.03494873270392418, 'timing/model': 1.0034760981798172, 'learning_rate/qwen_vl_interface': 9.984930060813613e-06, 'learning_rate/action_model': 9.984930060813613e-05, 'epoch': 0.11}) +[2026-06-17 17:12:41,426][starVLA.training.train_starvla][INFO] - Step 276, Loss: {'train/loss': 0.8636743426322937, 'train/grad_norm_pre_clip': 9.793158531188965, 'timing/data': 0.02908089105039835, 'timing/model': 0.9440249539911747, 'learning_rate/qwen_vl_interface': 9.984757433291832e-06, 'learning_rate/action_model': 9.984757433291832e-05, 'epoch': 0.11}) +[2026-06-17 17:12:45,279][starVLA.training.train_starvla][INFO] - Step 277, Loss: {'train/loss': 0.7919426560401917, 'train/grad_norm_pre_clip': 5.104252338409424, 'timing/data': 0.037461753003299236, 'timing/model': 0.9556523505598307, 'learning_rate/qwen_vl_interface': 9.984583824249742e-06, 'learning_rate/action_model': 9.984583824249742e-05, 'epoch': 0.11}) +[2026-06-17 17:12:49,173][starVLA.training.train_starvla][INFO] - Step 278, Loss: {'train/loss': 0.8887015581130981, 'train/grad_norm_pre_clip': 3.140183210372925, 'timing/data': 0.031244839541614056, 'timing/model': 0.9312306344509125, 'learning_rate/qwen_vl_interface': 9.98440923372333e-06, 'learning_rate/action_model': 9.984409233723328e-05, 'epoch': 0.11}) +[2026-06-17 17:12:53,070][starVLA.training.train_starvla][INFO] - Step 279, Loss: {'train/loss': 0.7381633520126343, 'train/grad_norm_pre_clip': 3.0311145782470703, 'timing/data': 0.027330229058861732, 'timing/model': 1.2314608916640282, 'learning_rate/qwen_vl_interface': 9.984233661748788e-06, 'learning_rate/action_model': 9.984233661748788e-05, 'epoch': 0.11}) +[2026-06-17 17:12:56,693][starVLA.training.train_starvla][INFO] - Step 280, Loss: {'train/loss': 1.102643370628357, 'train/grad_norm_pre_clip': 6.222664833068848, 'timing/data': 0.03932130616158247, 'timing/model': 0.9285092549398541, 'learning_rate/qwen_vl_interface': 9.984057108362516e-06, 'learning_rate/action_model': 9.984057108362515e-05, 'epoch': 0.12}) +[2026-06-17 17:13:00,494][starVLA.training.train_starvla][INFO] - Step 281, Loss: {'train/loss': 1.124111294746399, 'train/grad_norm_pre_clip': 4.374555587768555, 'timing/data': 0.0334434574469924, 'timing/model': 0.931202688254416, 'learning_rate/qwen_vl_interface': 9.98387957360111e-06, 'learning_rate/action_model': 9.98387957360111e-05, 'epoch': 0.12}) +[2026-06-17 17:13:04,507][starVLA.training.train_starvla][INFO] - Step 282, Loss: {'train/loss': 0.8829134702682495, 'train/grad_norm_pre_clip': 4.936490058898926, 'timing/data': 0.03401971701532602, 'timing/model': 0.9311628751456738, 'learning_rate/qwen_vl_interface': 9.983701057501375e-06, 'learning_rate/action_model': 9.983701057501375e-05, 'epoch': 0.12}) +[2026-06-17 17:13:08,372][starVLA.training.train_starvla][INFO] - Step 283, Loss: {'train/loss': 0.9759172201156616, 'train/grad_norm_pre_clip': 4.799968242645264, 'timing/data': 0.02303577959537506, 'timing/model': 0.9462526459246874, 'learning_rate/qwen_vl_interface': 9.983521560100317e-06, 'learning_rate/action_model': 9.983521560100316e-05, 'epoch': 0.12}) +[2026-06-17 17:13:12,255][starVLA.training.train_starvla][INFO] - Step 284, Loss: {'train/loss': 1.181197166442871, 'train/grad_norm_pre_clip': 4.994071006774902, 'timing/data': 0.028009088709950447, 'timing/model': 0.9435787117108703, 'learning_rate/qwen_vl_interface': 9.983341081435145e-06, 'learning_rate/action_model': 9.983341081435145e-05, 'epoch': 0.12}) +[2026-06-17 17:13:16,041][starVLA.training.train_starvla][INFO] - Step 285, Loss: {'train/loss': 1.0733604431152344, 'train/grad_norm_pre_clip': 5.261009216308594, 'timing/data': 0.024745916947722435, 'timing/model': 1.109064464457333, 'learning_rate/qwen_vl_interface': 9.983159621543275e-06, 'learning_rate/action_model': 9.983159621543275e-05, 'epoch': 0.12}) +[2026-06-17 17:13:19,901][starVLA.training.train_starvla][INFO] - Step 286, Loss: {'train/loss': 1.2107856273651123, 'train/grad_norm_pre_clip': 5.5483174324035645, 'timing/data': 0.0285642109811306, 'timing/model': 1.1846872009336948, 'learning_rate/qwen_vl_interface': 9.982977180462324e-06, 'learning_rate/action_model': 9.982977180462323e-05, 'epoch': 0.12}) +[2026-06-17 17:13:23,578][starVLA.training.train_starvla][INFO] - Step 287, Loss: {'train/loss': 0.8773943185806274, 'train/grad_norm_pre_clip': 6.753994464874268, 'timing/data': 0.025728448294103146, 'timing/model': 0.9358330508694053, 'learning_rate/qwen_vl_interface': 9.982793758230107e-06, 'learning_rate/action_model': 9.982793758230108e-05, 'epoch': 0.12}) +[2026-06-17 17:13:27,426][starVLA.training.train_starvla][INFO] - Step 288, Loss: {'train/loss': 0.8861513733863831, 'train/grad_norm_pre_clip': 5.484282970428467, 'timing/data': 0.028852779418230057, 'timing/model': 0.9390428317710757, 'learning_rate/qwen_vl_interface': 9.982609354884653e-06, 'learning_rate/action_model': 9.982609354884653e-05, 'epoch': 0.12}) +[2026-06-17 17:13:31,255][starVLA.training.train_starvla][INFO] - Step 289, Loss: {'train/loss': 0.8589937686920166, 'train/grad_norm_pre_clip': 7.427407264709473, 'timing/data': 0.02433164231479168, 'timing/model': 0.9379357984289527, 'learning_rate/qwen_vl_interface': 9.982423970464187e-06, 'learning_rate/action_model': 9.982423970464186e-05, 'epoch': 0.12}) +[2026-06-17 17:13:35,249][starVLA.training.train_starvla][INFO] - Step 290, Loss: {'train/loss': 0.8576459288597107, 'train/grad_norm_pre_clip': 8.06863021850586, 'timing/data': 0.028370419517159462, 'timing/model': 0.9379910696297884, 'learning_rate/qwen_vl_interface': 9.98223760500714e-06, 'learning_rate/action_model': 9.982237605007139e-05, 'epoch': 0.12}) +[2026-06-17 17:13:39,483][starVLA.training.train_starvla][INFO] - Step 291, Loss: {'train/loss': 1.253879189491272, 'train/grad_norm_pre_clip': 14.11229133605957, 'timing/data': 0.02694757841527462, 'timing/model': 1.1122299786657095, 'learning_rate/qwen_vl_interface': 9.982050258552146e-06, 'learning_rate/action_model': 9.982050258552145e-05, 'epoch': 0.12}) +[2026-06-17 17:13:43,373][starVLA.training.train_starvla][INFO] - Step 292, Loss: {'train/loss': 0.9132180213928223, 'train/grad_norm_pre_clip': 6.697291374206543, 'timing/data': 0.024800905026495457, 'timing/model': 1.1183321727439761, 'learning_rate/qwen_vl_interface': 9.981861931138039e-06, 'learning_rate/action_model': 9.981861931138039e-05, 'epoch': 0.12}) +[2026-06-17 17:13:47,253][starVLA.training.train_starvla][INFO] - Step 293, Loss: {'train/loss': 0.9506781101226807, 'train/grad_norm_pre_clip': 5.59562873840332, 'timing/data': 0.028165440075099468, 'timing/model': 1.13036875333637, 'learning_rate/qwen_vl_interface': 9.981672622803863e-06, 'learning_rate/action_model': 9.981672622803862e-05, 'epoch': 0.12}) +[2026-06-17 17:13:51,048][starVLA.training.train_starvla][INFO] - Step 294, Loss: {'train/loss': 0.9492883086204529, 'train/grad_norm_pre_clip': 5.808786392211914, 'timing/data': 0.03350941836833954, 'timing/model': 0.9758818726986647, 'learning_rate/qwen_vl_interface': 9.981482333588863e-06, 'learning_rate/action_model': 9.981482333588862e-05, 'epoch': 0.12}) +[2026-06-17 17:13:55,017][starVLA.training.train_starvla][INFO] - Step 295, Loss: {'train/loss': 1.0010663270950317, 'train/grad_norm_pre_clip': 5.428448677062988, 'timing/data': 0.025440840981900692, 'timing/model': 0.9469463974237442, 'learning_rate/qwen_vl_interface': 9.981291063532481e-06, 'learning_rate/action_model': 9.981291063532481e-05, 'epoch': 0.12}) +[2026-06-17 17:13:58,952][starVLA.training.train_starvla][INFO] - Step 296, Loss: {'train/loss': 1.1948580741882324, 'train/grad_norm_pre_clip': 7.403757572174072, 'timing/data': 0.02754107490181923, 'timing/model': 0.9453547755256295, 'learning_rate/qwen_vl_interface': 9.981098812674372e-06, 'learning_rate/action_model': 9.981098812674371e-05, 'epoch': 0.12}) +[2026-06-17 17:14:02,994][starVLA.training.train_starvla][INFO] - Step 297, Loss: {'train/loss': 1.3275991678237915, 'train/grad_norm_pre_clip': 7.9507527351379395, 'timing/data': 0.025416980497539043, 'timing/model': 1.0237063942477107, 'learning_rate/qwen_vl_interface': 9.980905581054388e-06, 'learning_rate/action_model': 9.980905581054387e-05, 'epoch': 0.12}) +[2026-06-17 17:14:06,985][starVLA.training.train_starvla][INFO] - Step 298, Loss: {'train/loss': 1.1290985345840454, 'train/grad_norm_pre_clip': 5.248310089111328, 'timing/data': 0.033130522817373276, 'timing/model': 0.988737584091723, 'learning_rate/qwen_vl_interface': 9.980711368712586e-06, 'learning_rate/action_model': 9.980711368712585e-05, 'epoch': 0.12}) +[2026-06-17 17:14:12,363][starVLA.training.train_starvla][INFO] - Step 299, Loss: {'train/loss': 0.9455457925796509, 'train/grad_norm_pre_clip': 6.871269226074219, 'timing/data': 0.03239284176379442, 'timing/model': 2.6106174988672137, 'learning_rate/qwen_vl_interface': 9.980516175689228e-06, 'learning_rate/action_model': 9.980516175689227e-05, 'epoch': 0.12}) +[2026-06-17 17:16:28,638][starVLA.training.train_starvla][INFO] - Step 300, Loss: {'train/loss': 1.0011287927627563, 'train/grad_norm_pre_clip': 6.694736957550049, 'eval/loss': 0.9868867993354797, 'eval/latency_6/loss': 0.9868867993354797, 'eval/demon_fire_precision': 0.917991892734643, 'eval/demon_fire_recall': 0.9515190691661279, 'eval/demon_fire_f1': 0.934454848436756, 'eval/demon_fire_rate_model': 0.8923205342237062, 'eval/demon_fire_rate_teacher': 0.860879243183083, 'eval/demon_left_precision': 0.5677083333333334, 'eval/demon_left_recall': 0.3824561403508772, 'eval/demon_left_f1': 0.45702306079664573, 'eval/demon_left_rate_model': 0.10684474123539232, 'eval/demon_left_rate_teacher': 0.15859766277128548, 'eval/demon_right_precision': 0.7054794520547946, 'eval/demon_right_recall': 0.15821812596006143, 'eval/demon_right_f1': 0.25846925972396484, 'eval/demon_right_rate_model': 0.04062326099053979, 'eval/demon_right_rate_teacher': 0.18113522537562604, 'eval/demon_move_f1': 0.35774616026030526, 'eval/demon_cc_f1': 0.6298485393562798, 'eval/latency_6/demon_fire_precision': 0.917991892734643, 'eval/latency_6/demon_fire_recall': 0.9515190691661279, 'eval/latency_6/demon_fire_f1': 0.934454848436756, 'eval/latency_6/demon_fire_rate_model': 0.8923205342237062, 'eval/latency_6/demon_fire_rate_teacher': 0.860879243183083, 'eval/latency_6/demon_left_precision': 0.5677083333333334, 'eval/latency_6/demon_left_recall': 0.3824561403508772, 'eval/latency_6/demon_left_f1': 0.45702306079664573, 'eval/latency_6/demon_left_rate_model': 0.10684474123539232, 'eval/latency_6/demon_left_rate_teacher': 0.15859766277128548, 'eval/latency_6/demon_right_precision': 0.7054794520547946, 'eval/latency_6/demon_right_recall': 0.15821812596006143, 'eval/latency_6/demon_right_f1': 0.25846925972396484, 'eval/latency_6/demon_right_rate_model': 0.04062326099053979, 'eval/latency_6/demon_right_rate_teacher': 0.18113522537562604, 'eval/latency_6/demon_move_f1': 0.35774616026030526, 'eval/latency_6/demon_cc_f1': 0.6298485393562798, 'timing/data': 0.03160991799086332, 'timing/model': 0.9397676782682538, 'learning_rate/qwen_vl_interface': 9.980320002024774e-06, 'learning_rate/action_model': 9.980320002024774e-05, 'epoch': 0.12}) +[2026-06-17 17:16:28,639][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_300_state +[2026-06-17 17:16:28,644][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 17:16:36,205][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_300_state/model.safetensors +[2026-06-17 17:16:51,174][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_300_state/optimizer.bin +[2026-06-17 17:16:51,175][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_300_state/sampler.bin +[2026-06-17 17:16:51,176][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_300_state/sampler_1.bin +[2026-06-17 17:16:51,185][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_300_state/random_states_0.pkl +[2026-06-17 17:16:51,190][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 300 +[2026-06-17 17:16:51,190][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 17:16:51,202][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 17:16:55,074][starVLA.training.train_starvla][INFO] - Step 301, Loss: {'train/loss': 0.9465001821517944, 'train/grad_norm_pre_clip': 6.256583213806152, 'timing/data': 0.03034697100520134, 'timing/model': 1.1894733486697078, 'learning_rate/qwen_vl_interface': 9.980122847759896e-06, 'learning_rate/action_model': 9.980122847759896e-05, 'epoch': 0.12}) +[2026-06-17 17:16:58,974][starVLA.training.train_starvla][INFO] - Step 302, Loss: {'train/loss': 1.0326662063598633, 'train/grad_norm_pre_clip': 6.217584609985352, 'timing/data': 0.028989039361476898, 'timing/model': 1.110286365263164, 'learning_rate/qwen_vl_interface': 9.97992471293546e-06, 'learning_rate/action_model': 9.97992471293546e-05, 'epoch': 0.12}) +[2026-06-17 17:17:02,873][starVLA.training.train_starvla][INFO] - Step 303, Loss: {'train/loss': 0.8684980869293213, 'train/grad_norm_pre_clip': 3.5704894065856934, 'timing/data': 0.02811509557068348, 'timing/model': 0.9300805479288101, 'learning_rate/qwen_vl_interface': 9.979725597592543e-06, 'learning_rate/action_model': 9.979725597592541e-05, 'epoch': 0.12}) +[2026-06-17 17:17:07,702][starVLA.training.train_starvla][INFO] - Step 304, Loss: {'train/loss': 0.8866482973098755, 'train/grad_norm_pre_clip': 5.0827412605285645, 'timing/data': 0.02357630617916584, 'timing/model': 1.5187375340610743, 'learning_rate/qwen_vl_interface': 9.979525501772418e-06, 'learning_rate/action_model': 9.979525501772417e-05, 'epoch': 0.13}) +[2026-06-17 17:17:11,865][starVLA.training.train_starvla][INFO] - Step 305, Loss: {'train/loss': 0.9596419334411621, 'train/grad_norm_pre_clip': 4.263095855712891, 'timing/data': 0.025245861150324345, 'timing/model': 1.1633470635861158, 'learning_rate/qwen_vl_interface': 9.97932442551657e-06, 'learning_rate/action_model': 9.979324425516569e-05, 'epoch': 0.13}) +[2026-06-17 17:17:15,844][starVLA.training.train_starvla][INFO] - Step 306, Loss: {'train/loss': 0.5574119091033936, 'train/grad_norm_pre_clip': 3.6983933448791504, 'timing/data': 0.03665362112224102, 'timing/model': 0.934662907384336, 'learning_rate/qwen_vl_interface': 9.979122368866678e-06, 'learning_rate/action_model': 9.979122368866679e-05, 'epoch': 0.13}) +[2026-06-17 17:17:19,774][starVLA.training.train_starvla][INFO] - Step 307, Loss: {'train/loss': 1.0702478885650635, 'train/grad_norm_pre_clip': 4.617365837097168, 'timing/data': 0.026124448515474796, 'timing/model': 0.9430047431960702, 'learning_rate/qwen_vl_interface': 9.978919331864631e-06, 'learning_rate/action_model': 9.97891933186463e-05, 'epoch': 0.13}) +[2026-06-17 17:17:23,795][starVLA.training.train_starvla][INFO] - Step 308, Loss: {'train/loss': 0.8155800104141235, 'train/grad_norm_pre_clip': 2.9416847229003906, 'timing/data': 0.02324666455388069, 'timing/model': 1.1788000343367457, 'learning_rate/qwen_vl_interface': 9.978715314552518e-06, 'learning_rate/action_model': 9.978715314552518e-05, 'epoch': 0.13}) +[2026-06-17 17:17:27,633][starVLA.training.train_starvla][INFO] - Step 309, Loss: {'train/loss': 1.491929054260254, 'train/grad_norm_pre_clip': 8.198837280273438, 'timing/data': 0.025045267306268215, 'timing/model': 1.0504255909472704, 'learning_rate/qwen_vl_interface': 9.978510316972631e-06, 'learning_rate/action_model': 9.978510316972631e-05, 'epoch': 0.13}) +[2026-06-17 17:17:31,421][starVLA.training.train_starvla][INFO] - Step 310, Loss: {'train/loss': 0.7437817454338074, 'train/grad_norm_pre_clip': 3.2278659343719482, 'timing/data': 0.029679245315492153, 'timing/model': 0.932277369312942, 'learning_rate/qwen_vl_interface': 9.978304339167467e-06, 'learning_rate/action_model': 9.978304339167467e-05, 'epoch': 0.13}) +[2026-06-17 17:17:37,629][starVLA.training.train_starvla][INFO] - Step 311, Loss: {'train/loss': 0.8449097871780396, 'train/grad_norm_pre_clip': 5.111281871795654, 'timing/data': 0.030522008426487446, 'timing/model': 2.365185140632093, 'learning_rate/qwen_vl_interface': 9.978097381179726e-06, 'learning_rate/action_model': 9.978097381179726e-05, 'epoch': 0.13}) +[2026-06-17 17:17:41,681][starVLA.training.train_starvla][INFO] - Step 312, Loss: {'train/loss': 0.9308344125747681, 'train/grad_norm_pre_clip': 3.614887237548828, 'timing/data': 0.009724169969558716, 'timing/model': 0.9393768822774291, 'learning_rate/qwen_vl_interface': 9.977889443052312e-06, 'learning_rate/action_model': 9.977889443052311e-05, 'epoch': 0.13}) +[2026-06-17 17:17:48,327][starVLA.training.train_starvla][INFO] - Step 313, Loss: {'train/loss': 0.8326602578163147, 'train/grad_norm_pre_clip': 5.0064191818237305, 'timing/data': 0.008568799123167992, 'timing/model': 3.7383935702964664, 'learning_rate/qwen_vl_interface': 9.977680524828327e-06, 'learning_rate/action_model': 9.977680524828327e-05, 'epoch': 0.13}) +[2026-06-17 17:17:52,237][starVLA.training.train_starvla][INFO] - Step 314, Loss: {'train/loss': 0.7878844141960144, 'train/grad_norm_pre_clip': 7.069552421569824, 'timing/data': 0.01739733200520277, 'timing/model': 1.1682556495070457, 'learning_rate/qwen_vl_interface': 9.977470626551083e-06, 'learning_rate/action_model': 9.977470626551083e-05, 'epoch': 0.13}) +[2026-06-17 17:17:56,079][starVLA.training.train_starvla][INFO] - Step 315, Loss: {'train/loss': 0.6940783262252808, 'train/grad_norm_pre_clip': 6.117353439331055, 'timing/data': 0.009115306660532951, 'timing/model': 1.1089976709336042, 'learning_rate/qwen_vl_interface': 9.977259748264089e-06, 'learning_rate/action_model': 9.97725974826409e-05, 'epoch': 0.13}) +[2026-06-17 17:17:59,765][starVLA.training.train_starvla][INFO] - Step 316, Loss: {'train/loss': 1.2842857837677002, 'train/grad_norm_pre_clip': 5.742411136627197, 'timing/data': 0.011388705112040043, 'timing/model': 0.9199369801208377, 'learning_rate/qwen_vl_interface': 9.977047890011065e-06, 'learning_rate/action_model': 9.977047890011065e-05, 'epoch': 0.13}) +[2026-06-17 17:18:03,567][starVLA.training.train_starvla][INFO] - Step 317, Loss: {'train/loss': 1.225363850593567, 'train/grad_norm_pre_clip': 6.631336212158203, 'timing/data': 0.008869846351444721, 'timing/model': 0.9241574686020613, 'learning_rate/qwen_vl_interface': 9.976835051835928e-06, 'learning_rate/action_model': 9.976835051835926e-05, 'epoch': 0.13}) +[2026-06-17 17:18:07,349][starVLA.training.train_starvla][INFO] - Step 318, Loss: {'train/loss': 0.8372784852981567, 'train/grad_norm_pre_clip': 7.523259162902832, 'timing/data': 0.01155161764472723, 'timing/model': 0.9423201698809862, 'learning_rate/qwen_vl_interface': 9.976621233782795e-06, 'learning_rate/action_model': 9.976621233782795e-05, 'epoch': 0.13}) +[2026-06-17 17:18:11,106][starVLA.training.train_starvla][INFO] - Step 319, Loss: {'train/loss': 0.9370667338371277, 'train/grad_norm_pre_clip': 6.477059364318848, 'timing/data': 0.007884849794209003, 'timing/model': 0.9345687227323651, 'learning_rate/qwen_vl_interface': 9.976406435895995e-06, 'learning_rate/action_model': 9.976406435895995e-05, 'epoch': 0.13}) +[2026-06-17 17:18:15,349][starVLA.training.train_starvla][INFO] - Step 320, Loss: {'train/loss': 1.0924116373062134, 'train/grad_norm_pre_clip': 5.6173095703125, 'timing/data': 0.010200227610766888, 'timing/model': 0.9440268827602267, 'learning_rate/qwen_vl_interface': 9.976190658220056e-06, 'learning_rate/action_model': 9.976190658220056e-05, 'epoch': 0.13}) +[2026-06-17 17:18:19,346][starVLA.training.train_starvla][INFO] - Step 321, Loss: {'train/loss': 0.9969226121902466, 'train/grad_norm_pre_clip': 2.996591329574585, 'timing/data': 0.010826663114130497, 'timing/model': 1.1590262753888965, 'learning_rate/qwen_vl_interface': 9.975973900799709e-06, 'learning_rate/action_model': 9.975973900799708e-05, 'epoch': 0.13}) +[2026-06-17 17:18:23,416][starVLA.training.train_starvla][INFO] - Step 322, Loss: {'train/loss': 1.099170446395874, 'train/grad_norm_pre_clip': 7.347977638244629, 'timing/data': 0.00889892503619194, 'timing/model': 1.0114131234586239, 'learning_rate/qwen_vl_interface': 9.975756163679884e-06, 'learning_rate/action_model': 9.975756163679885e-05, 'epoch': 0.13}) +[2026-06-17 17:18:27,582][starVLA.training.train_starvla][INFO] - Step 323, Loss: {'train/loss': 0.9877744913101196, 'train/grad_norm_pre_clip': 4.898098945617676, 'timing/data': 0.008303833194077015, 'timing/model': 1.0441557802259922, 'learning_rate/qwen_vl_interface': 9.975537446905722e-06, 'learning_rate/action_model': 9.975537446905723e-05, 'epoch': 0.13}) +[2026-06-17 17:18:31,614][starVLA.training.train_starvla][INFO] - Step 324, Loss: {'train/loss': 1.069100260734558, 'train/grad_norm_pre_clip': 7.026169300079346, 'timing/data': 0.010681229643523693, 'timing/model': 0.9389944858849049, 'learning_rate/qwen_vl_interface': 9.975317750522565e-06, 'learning_rate/action_model': 9.975317750522564e-05, 'epoch': 0.13}) +[2026-06-17 17:18:35,410][starVLA.training.train_starvla][INFO] - Step 325, Loss: {'train/loss': 0.7590100765228271, 'train/grad_norm_pre_clip': 3.49336576461792, 'timing/data': 0.008787451311945915, 'timing/model': 0.9398194346576929, 'learning_rate/qwen_vl_interface': 9.97509707457595e-06, 'learning_rate/action_model': 9.97509707457595e-05, 'epoch': 0.13}) +[2026-06-17 17:18:39,233][starVLA.training.train_starvla][INFO] - Step 326, Loss: {'train/loss': 0.9173271656036377, 'train/grad_norm_pre_clip': 4.638777732849121, 'timing/data': 0.009745459072291851, 'timing/model': 0.9423095677047968, 'learning_rate/qwen_vl_interface': 9.974875419111631e-06, 'learning_rate/action_model': 9.97487541911163e-05, 'epoch': 0.13}) +[2026-06-17 17:18:43,027][starVLA.training.train_starvla][INFO] - Step 327, Loss: {'train/loss': 0.9305368065834045, 'train/grad_norm_pre_clip': 4.461155414581299, 'timing/data': 0.009066374972462654, 'timing/model': 0.9343819813802838, 'learning_rate/qwen_vl_interface': 9.974652784175552e-06, 'learning_rate/action_model': 9.974652784175551e-05, 'epoch': 0.13}) +[2026-06-17 17:18:47,232][starVLA.training.train_starvla][INFO] - Step 328, Loss: {'train/loss': 1.1839135885238647, 'train/grad_norm_pre_clip': 8.958039283752441, 'timing/data': 0.010646602138876915, 'timing/model': 1.1512137949466705, 'learning_rate/qwen_vl_interface': 9.974429169813865e-06, 'learning_rate/action_model': 9.974429169813864e-05, 'epoch': 0.14}) +[2026-06-17 17:18:50,856][starVLA.training.train_starvla][INFO] - Step 329, Loss: {'train/loss': 0.9035511612892151, 'train/grad_norm_pre_clip': 5.649453163146973, 'timing/data': 0.012937808409333229, 'timing/model': 0.9362905947491527, 'learning_rate/qwen_vl_interface': 9.974204576072928e-06, 'learning_rate/action_model': 9.974204576072928e-05, 'epoch': 0.14}) +[2026-06-17 17:18:54,626][starVLA.training.train_starvla][INFO] - Step 330, Loss: {'train/loss': 1.1813515424728394, 'train/grad_norm_pre_clip': 5.344922065734863, 'timing/data': 0.008309322409331799, 'timing/model': 0.9322071420028806, 'learning_rate/qwen_vl_interface': 9.973979002999299e-06, 'learning_rate/action_model': 9.973979002999299e-05, 'epoch': 0.14}) +[2026-06-17 17:18:58,468][starVLA.training.train_starvla][INFO] - Step 331, Loss: {'train/loss': 1.0513219833374023, 'train/grad_norm_pre_clip': 31.75385093688965, 'timing/data': 0.006643461994826794, 'timing/model': 0.9538769228383899, 'learning_rate/qwen_vl_interface': 9.97375245063974e-06, 'learning_rate/action_model': 9.97375245063974e-05, 'epoch': 0.14}) +[2026-06-17 17:19:02,500][starVLA.training.train_starvla][INFO] - Step 332, Loss: {'train/loss': 1.0255619287490845, 'train/grad_norm_pre_clip': 4.554740905761719, 'timing/data': 0.009440751746296883, 'timing/model': 1.0209778053686023, 'learning_rate/qwen_vl_interface': 9.973524919041215e-06, 'learning_rate/action_model': 9.973524919041215e-05, 'epoch': 0.14}) +[2026-06-17 17:19:06,527][starVLA.training.train_starvla][INFO] - Step 333, Loss: {'train/loss': 1.1423956155776978, 'train/grad_norm_pre_clip': 4.749995708465576, 'timing/data': 0.00961067620664835, 'timing/model': 1.0739697562530637, 'learning_rate/qwen_vl_interface': 9.97329640825089e-06, 'learning_rate/action_model': 9.973296408250891e-05, 'epoch': 0.14}) +[2026-06-17 17:19:10,622][starVLA.training.train_starvla][INFO] - Step 334, Loss: {'train/loss': 1.0227710008621216, 'train/grad_norm_pre_clip': 7.730557441711426, 'timing/data': 0.009996904991567135, 'timing/model': 1.2638095542788506, 'learning_rate/qwen_vl_interface': 9.97306691831614e-06, 'learning_rate/action_model': 9.97306691831614e-05, 'epoch': 0.14}) +[2026-06-17 17:19:14,493][starVLA.training.train_starvla][INFO] - Step 335, Loss: {'train/loss': 1.187670350074768, 'train/grad_norm_pre_clip': 14.264938354492188, 'timing/data': 0.008843259885907173, 'timing/model': 1.1445571770891547, 'learning_rate/qwen_vl_interface': 9.972836449284533e-06, 'learning_rate/action_model': 9.972836449284532e-05, 'epoch': 0.14}) +[2026-06-17 17:19:18,155][starVLA.training.train_starvla][INFO] - Step 336, Loss: {'train/loss': 0.994176983833313, 'train/grad_norm_pre_clip': 5.714041233062744, 'timing/data': 0.01021492201834917, 'timing/model': 0.9505459116771817, 'learning_rate/qwen_vl_interface': 9.97260500120385e-06, 'learning_rate/action_model': 9.97260500120385e-05, 'epoch': 0.14}) +[2026-06-17 17:19:22,038][starVLA.training.train_starvla][INFO] - Step 337, Loss: {'train/loss': 0.8513259291648865, 'train/grad_norm_pre_clip': 5.021971702575684, 'timing/data': 0.0069086551666259766, 'timing/model': 0.9467524448409677, 'learning_rate/qwen_vl_interface': 9.972372574122067e-06, 'learning_rate/action_model': 9.972372574122068e-05, 'epoch': 0.14}) +[2026-06-17 17:19:25,959][starVLA.training.train_starvla][INFO] - Step 338, Loss: {'train/loss': 1.3248411417007446, 'train/grad_norm_pre_clip': 5.830910682678223, 'timing/data': 0.01029652263969183, 'timing/model': 0.9375446615740657, 'learning_rate/qwen_vl_interface': 9.972139168087368e-06, 'learning_rate/action_model': 9.972139168087369e-05, 'epoch': 0.14}) +[2026-06-17 17:19:32,329][starVLA.training.train_starvla][INFO] - Step 339, Loss: {'train/loss': 0.8534904718399048, 'train/grad_norm_pre_clip': 5.498149394989014, 'timing/data': 0.008013226091861725, 'timing/model': 1.1878222972154617, 'learning_rate/qwen_vl_interface': 9.97190478314814e-06, 'learning_rate/action_model': 9.97190478314814e-05, 'epoch': 0.14}) +[2026-06-17 17:19:36,549][starVLA.training.train_starvla][INFO] - Step 340, Loss: {'train/loss': 0.7973954081535339, 'train/grad_norm_pre_clip': 4.183435440063477, 'timing/data': 0.012227263301610947, 'timing/model': 0.9503735806792974, 'learning_rate/qwen_vl_interface': 9.971669419352968e-06, 'learning_rate/action_model': 9.971669419352966e-05, 'epoch': 0.14}) +[2026-06-17 17:19:40,624][starVLA.training.train_starvla][INFO] - Step 341, Loss: {'train/loss': 1.1214144229888916, 'train/grad_norm_pre_clip': 5.280436992645264, 'timing/data': 0.01406660582870245, 'timing/model': 1.1300957798957825, 'learning_rate/qwen_vl_interface': 9.971433076750644e-06, 'learning_rate/action_model': 9.971433076750643e-05, 'epoch': 0.14}) +[2026-06-17 17:19:44,533][starVLA.training.train_starvla][INFO] - Step 342, Loss: {'train/loss': 1.3902689218521118, 'train/grad_norm_pre_clip': 8.265727996826172, 'timing/data': 0.009471551515161991, 'timing/model': 1.1598020857200027, 'learning_rate/qwen_vl_interface': 9.971195755390166e-06, 'learning_rate/action_model': 9.971195755390164e-05, 'epoch': 0.14}) +[2026-06-17 17:19:48,254][starVLA.training.train_starvla][INFO] - Step 343, Loss: {'train/loss': 0.9453909397125244, 'train/grad_norm_pre_clip': 4.3181376457214355, 'timing/data': 0.01888410560786724, 'timing/model': 0.9282553028315306, 'learning_rate/qwen_vl_interface': 9.970957455320723e-06, 'learning_rate/action_model': 9.970957455320723e-05, 'epoch': 0.14}) +[2026-06-17 17:19:52,040][starVLA.training.train_starvla][INFO] - Step 344, Loss: {'train/loss': 0.8963777422904968, 'train/grad_norm_pre_clip': 5.934500694274902, 'timing/data': 0.00974983349442482, 'timing/model': 0.9357235049828887, 'learning_rate/qwen_vl_interface': 9.970718176591723e-06, 'learning_rate/action_model': 9.970718176591722e-05, 'epoch': 0.14}) +[2026-06-17 17:19:55,894][starVLA.training.train_starvla][INFO] - Step 345, Loss: {'train/loss': 0.9110164046287537, 'train/grad_norm_pre_clip': 6.454567909240723, 'timing/data': 0.008419914171099663, 'timing/model': 0.9548664633184671, 'learning_rate/qwen_vl_interface': 9.970477919252763e-06, 'learning_rate/action_model': 9.970477919252763e-05, 'epoch': 0.14}) +[2026-06-17 17:19:59,963][starVLA.training.train_starvla][INFO] - Step 346, Loss: {'train/loss': 1.2170761823654175, 'train/grad_norm_pre_clip': 6.005058288574219, 'timing/data': 0.008883029222488403, 'timing/model': 1.2915358673781157, 'learning_rate/qwen_vl_interface': 9.970236683353652e-06, 'learning_rate/action_model': 9.970236683353652e-05, 'epoch': 0.14}) +[2026-06-17 17:20:03,775][starVLA.training.train_starvla][INFO] - Step 347, Loss: {'train/loss': 0.9901326298713684, 'train/grad_norm_pre_clip': 7.986317157745361, 'timing/data': 0.012016642838716507, 'timing/model': 0.9268188085407019, 'learning_rate/qwen_vl_interface': 9.969994468944399e-06, 'learning_rate/action_model': 9.969994468944398e-05, 'epoch': 0.14}) +[2026-06-17 17:20:07,649][starVLA.training.train_starvla][INFO] - Step 348, Loss: {'train/loss': 0.9219748377799988, 'train/grad_norm_pre_clip': 6.009446144104004, 'timing/data': 0.014959537424147129, 'timing/model': 1.1443185675889254, 'learning_rate/qwen_vl_interface': 9.969751276075212e-06, 'learning_rate/action_model': 9.969751276075212e-05, 'epoch': 0.14}) +[2026-06-17 17:20:11,380][starVLA.training.train_starvla][INFO] - Step 349, Loss: {'train/loss': 0.8708532452583313, 'train/grad_norm_pre_clip': 6.797059059143066, 'timing/data': 0.008454225957393646, 'timing/model': 0.9662031922489405, 'learning_rate/qwen_vl_interface': 9.969507104796506e-06, 'learning_rate/action_model': 9.969507104796505e-05, 'epoch': 0.14}) +[2026-06-17 17:20:15,314][starVLA.training.train_starvla][INFO] - Step 350, Loss: {'train/loss': 0.8903600573539734, 'train/grad_norm_pre_clip': 4.581815242767334, 'timing/data': 0.008151381276547909, 'timing/model': 0.9668234661221504, 'learning_rate/qwen_vl_interface': 9.969261955158898e-06, 'learning_rate/action_model': 9.969261955158898e-05, 'epoch': 0.14}) +[2026-06-17 17:20:19,219][starVLA.training.train_starvla][INFO] - Step 351, Loss: {'train/loss': 0.7495279908180237, 'train/grad_norm_pre_clip': 3.889214277267456, 'timing/data': 0.01106733176857233, 'timing/model': 0.949091337621212, 'learning_rate/qwen_vl_interface': 9.96901582721321e-06, 'learning_rate/action_model': 9.96901582721321e-05, 'epoch': 0.14}) +[2026-06-17 17:20:23,128][starVLA.training.train_starvla][INFO] - Step 352, Loss: {'train/loss': 1.1081387996673584, 'train/grad_norm_pre_clip': 6.28065824508667, 'timing/data': 0.01231347769498825, 'timing/model': 0.9761878103017807, 'learning_rate/qwen_vl_interface': 9.968768721010463e-06, 'learning_rate/action_model': 9.968768721010463e-05, 'epoch': 0.14}) +[2026-06-17 17:20:27,489][starVLA.training.train_starvla][INFO] - Step 353, Loss: {'train/loss': 1.0789645910263062, 'train/grad_norm_pre_clip': 6.0041632652282715, 'timing/data': 0.009757631458342075, 'timing/model': 0.9401834672316909, 'learning_rate/qwen_vl_interface': 9.968520636601883e-06, 'learning_rate/action_model': 9.968520636601883e-05, 'epoch': 0.15}) +[2026-06-17 17:20:31,531][starVLA.training.train_starvla][INFO] - Step 354, Loss: {'train/loss': 1.0845646858215332, 'train/grad_norm_pre_clip': 6.3234100341796875, 'timing/data': 0.011703201569616795, 'timing/model': 1.1279169153422117, 'learning_rate/qwen_vl_interface': 9.968271574038899e-06, 'learning_rate/action_model': 9.968271574038898e-05, 'epoch': 0.15}) +[2026-06-17 17:20:35,383][starVLA.training.train_starvla][INFO] - Step 355, Loss: {'train/loss': 0.9961779713630676, 'train/grad_norm_pre_clip': 6.5656819343566895, 'timing/data': 0.008806927129626274, 'timing/model': 1.1275914460420609, 'learning_rate/qwen_vl_interface': 9.968021533373138e-06, 'learning_rate/action_model': 9.968021533373138e-05, 'epoch': 0.15}) +[2026-06-17 17:20:39,037][starVLA.training.train_starvla][INFO] - Step 356, Loss: {'train/loss': 1.2393782138824463, 'train/grad_norm_pre_clip': 7.063343524932861, 'timing/data': 0.010518028400838375, 'timing/model': 0.9263249970972538, 'learning_rate/qwen_vl_interface': 9.967770514656439e-06, 'learning_rate/action_model': 9.967770514656438e-05, 'epoch': 0.15}) +[2026-06-17 17:20:42,825][starVLA.training.train_starvla][INFO] - Step 357, Loss: {'train/loss': 1.0259039402008057, 'train/grad_norm_pre_clip': 8.290533065795898, 'timing/data': 0.007402484305202961, 'timing/model': 0.962081397883594, 'learning_rate/qwen_vl_interface': 9.967518517940832e-06, 'learning_rate/action_model': 9.967518517940832e-05, 'epoch': 0.15}) +[2026-06-17 17:20:46,676][starVLA.training.train_starvla][INFO] - Step 358, Loss: {'train/loss': 1.2264988422393799, 'train/grad_norm_pre_clip': 7.3892388343811035, 'timing/data': 0.008334451355040073, 'timing/model': 1.0077348686754704, 'learning_rate/qwen_vl_interface': 9.967265543278563e-06, 'learning_rate/action_model': 9.967265543278562e-05, 'epoch': 0.15}) +[2026-06-17 17:20:50,443][starVLA.training.train_starvla][INFO] - Step 359, Loss: {'train/loss': 0.7241579294204712, 'train/grad_norm_pre_clip': 3.5674097537994385, 'timing/data': 0.009562023915350437, 'timing/model': 0.929819593206048, 'learning_rate/qwen_vl_interface': 9.96701159072207e-06, 'learning_rate/action_model': 9.96701159072207e-05, 'epoch': 0.15}) +[2026-06-17 17:20:54,322][starVLA.training.train_starvla][INFO] - Step 360, Loss: {'train/loss': 0.7660156488418579, 'train/grad_norm_pre_clip': 4.694231986999512, 'timing/data': 0.008478574454784393, 'timing/model': 0.9300472484901547, 'learning_rate/qwen_vl_interface': 9.966756660323998e-06, 'learning_rate/action_model': 9.966756660323999e-05, 'epoch': 0.15}) +[2026-06-17 17:20:58,224][starVLA.training.train_starvla][INFO] - Step 361, Loss: {'train/loss': 0.9123981595039368, 'train/grad_norm_pre_clip': 7.666846752166748, 'timing/data': 0.008214987814426422, 'timing/model': 1.1527450159192085, 'learning_rate/qwen_vl_interface': 9.966500752137195e-06, 'learning_rate/action_model': 9.966500752137196e-05, 'epoch': 0.15}) +[2026-06-17 17:21:02,035][starVLA.training.train_starvla][INFO] - Step 362, Loss: {'train/loss': 1.1167542934417725, 'train/grad_norm_pre_clip': 5.696255207061768, 'timing/data': 0.009863684885203838, 'timing/model': 1.1336785005405545, 'learning_rate/qwen_vl_interface': 9.966243866214712e-06, 'learning_rate/action_model': 9.966243866214712e-05, 'epoch': 0.15}) +[2026-06-17 17:21:05,977][starVLA.training.train_starvla][INFO] - Step 363, Loss: {'train/loss': 0.9393739700317383, 'train/grad_norm_pre_clip': 5.382136821746826, 'timing/data': 0.009109186939895153, 'timing/model': 0.9563414081931114, 'learning_rate/qwen_vl_interface': 9.965986002609799e-06, 'learning_rate/action_model': 9.965986002609799e-05, 'epoch': 0.15}) +[2026-06-17 17:21:10,353][starVLA.training.train_starvla][INFO] - Step 364, Loss: {'train/loss': 0.9987660646438599, 'train/grad_norm_pre_clip': 4.89346981048584, 'timing/data': 0.011149252764880657, 'timing/model': 0.9699246101081371, 'learning_rate/qwen_vl_interface': 9.965727161375915e-06, 'learning_rate/action_model': 9.965727161375914e-05, 'epoch': 0.15}) +[2026-06-17 17:21:14,386][starVLA.training.train_starvla][INFO] - Step 365, Loss: {'train/loss': 0.7413080334663391, 'train/grad_norm_pre_clip': 6.757819175720215, 'timing/data': 0.00930645689368248, 'timing/model': 0.9933438496664166, 'learning_rate/qwen_vl_interface': 9.965467342566715e-06, 'learning_rate/action_model': 9.965467342566715e-05, 'epoch': 0.15}) +[2026-06-17 17:21:18,225][starVLA.training.train_starvla][INFO] - Step 366, Loss: {'train/loss': 0.9595904350280762, 'train/grad_norm_pre_clip': 3.382558822631836, 'timing/data': 0.012527505867183208, 'timing/model': 0.9789933804422617, 'learning_rate/qwen_vl_interface': 9.965206546236061e-06, 'learning_rate/action_model': 9.96520654623606e-05, 'epoch': 0.15}) +[2026-06-17 17:21:22,226][starVLA.training.train_starvla][INFO] - Step 367, Loss: {'train/loss': 0.6957175731658936, 'train/grad_norm_pre_clip': 4.106905460357666, 'timing/data': 0.014932182617485523, 'timing/model': 1.0543616628274322, 'learning_rate/qwen_vl_interface': 9.964944772438014e-06, 'learning_rate/action_model': 9.964944772438013e-05, 'epoch': 0.15}) +[2026-06-17 17:21:26,148][starVLA.training.train_starvla][INFO] - Step 368, Loss: {'train/loss': 0.774116039276123, 'train/grad_norm_pre_clip': 4.556819915771484, 'timing/data': 0.0070427171885967255, 'timing/model': 1.1946750422939658, 'learning_rate/qwen_vl_interface': 9.964682021226844e-06, 'learning_rate/action_model': 9.964682021226844e-05, 'epoch': 0.15}) +[2026-06-17 17:21:30,043][starVLA.training.train_starvla][INFO] - Step 369, Loss: {'train/loss': 0.9090219736099243, 'train/grad_norm_pre_clip': 5.8006439208984375, 'timing/data': 0.013358752243220806, 'timing/model': 0.9369672313332558, 'learning_rate/qwen_vl_interface': 9.964418292657019e-06, 'learning_rate/action_model': 9.964418292657018e-05, 'epoch': 0.15}) +[2026-06-17 17:21:33,842][starVLA.training.train_starvla][INFO] - Step 370, Loss: {'train/loss': 0.7943227887153625, 'train/grad_norm_pre_clip': 4.028494358062744, 'timing/data': 0.012284607626497746, 'timing/model': 0.9397134128957987, 'learning_rate/qwen_vl_interface': 9.96415358678321e-06, 'learning_rate/action_model': 9.964153586783209e-05, 'epoch': 0.15}) +[2026-06-17 17:21:37,576][starVLA.training.train_starvla][INFO] - Step 371, Loss: {'train/loss': 0.7739865183830261, 'train/grad_norm_pre_clip': 4.6136651039123535, 'timing/data': 0.009593991562724113, 'timing/model': 0.9262457927688956, 'learning_rate/qwen_vl_interface': 9.963887903660285e-06, 'learning_rate/action_model': 9.963887903660285e-05, 'epoch': 0.15}) +[2026-06-17 17:21:41,448][starVLA.training.train_starvla][INFO] - Step 372, Loss: {'train/loss': 1.2959489822387695, 'train/grad_norm_pre_clip': 7.574708938598633, 'timing/data': 0.011867250315845013, 'timing/model': 0.9774410044774413, 'learning_rate/qwen_vl_interface': 9.96362124334333e-06, 'learning_rate/action_model': 9.963621243343329e-05, 'epoch': 0.15}) +[2026-06-17 17:21:45,248][starVLA.training.train_starvla][INFO] - Step 373, Loss: {'train/loss': 0.7504653930664062, 'train/grad_norm_pre_clip': 4.996454238891602, 'timing/data': 0.00774986669421196, 'timing/model': 0.9307096637785435, 'learning_rate/qwen_vl_interface': 9.963353605887617e-06, 'learning_rate/action_model': 9.963353605887616e-05, 'epoch': 0.15}) +[2026-06-17 17:21:49,054][starVLA.training.train_starvla][INFO] - Step 374, Loss: {'train/loss': 1.0183175802230835, 'train/grad_norm_pre_clip': 15.526904106140137, 'timing/data': 0.009517000056803226, 'timing/model': 0.9486847845837474, 'learning_rate/qwen_vl_interface': 9.963084991348628e-06, 'learning_rate/action_model': 9.963084991348628e-05, 'epoch': 0.15}) +[2026-06-17 17:21:52,929][starVLA.training.train_starvla][INFO] - Step 375, Loss: {'train/loss': 1.3564561605453491, 'train/grad_norm_pre_clip': 9.456191062927246, 'timing/data': 0.008657609112560749, 'timing/model': 1.1783115668222308, 'learning_rate/qwen_vl_interface': 9.962815399782051e-06, 'learning_rate/action_model': 9.96281539978205e-05, 'epoch': 0.15}) +[2026-06-17 17:21:56,705][starVLA.training.train_starvla][INFO] - Step 376, Loss: {'train/loss': 0.9029784202575684, 'train/grad_norm_pre_clip': 17.131427764892578, 'timing/data': 0.007217892445623875, 'timing/model': 0.9548935759812593, 'learning_rate/qwen_vl_interface': 9.96254483124377e-06, 'learning_rate/action_model': 9.96254483124377e-05, 'epoch': 0.15}) +[2026-06-17 17:22:00,755][starVLA.training.train_starvla][INFO] - Step 377, Loss: {'train/loss': 1.0117430686950684, 'train/grad_norm_pre_clip': 7.809257507324219, 'timing/data': 0.009712226688861847, 'timing/model': 0.977316434495151, 'learning_rate/qwen_vl_interface': 9.962273285789876e-06, 'learning_rate/action_model': 9.962273285789875e-05, 'epoch': 0.16}) +[2026-06-17 17:22:07,386][starVLA.training.train_starvla][INFO] - Step 378, Loss: {'train/loss': 0.8761756420135498, 'train/grad_norm_pre_clip': 4.465548038482666, 'timing/data': 0.008932013995945454, 'timing/model': 1.0074349259957671, 'learning_rate/qwen_vl_interface': 9.962000763476658e-06, 'learning_rate/action_model': 9.962000763476658e-05, 'epoch': 0.16}) +[2026-06-17 17:22:13,287][starVLA.training.train_starvla][INFO] - Step 379, Loss: {'train/loss': 1.0016558170318604, 'train/grad_norm_pre_clip': 4.540921211242676, 'timing/data': 0.018094396218657494, 'timing/model': 1.0589210465550423, 'learning_rate/qwen_vl_interface': 9.96172726436061e-06, 'learning_rate/action_model': 9.96172726436061e-05, 'epoch': 0.16}) +[2026-06-17 17:22:17,920][starVLA.training.train_starvla][INFO] - Step 380, Loss: {'train/loss': 0.765945553779602, 'train/grad_norm_pre_clip': 5.950633525848389, 'timing/data': 0.014940843917429447, 'timing/model': 1.1291770674288273, 'learning_rate/qwen_vl_interface': 9.961452788498432e-06, 'learning_rate/action_model': 9.961452788498433e-05, 'epoch': 0.16}) +[2026-06-17 17:22:21,944][starVLA.training.train_starvla][INFO] - Step 381, Loss: {'train/loss': 0.9440561532974243, 'train/grad_norm_pre_clip': 6.383771896362305, 'timing/data': 0.008855322375893593, 'timing/model': 1.24796640034765, 'learning_rate/qwen_vl_interface': 9.96117733594702e-06, 'learning_rate/action_model': 9.96117733594702e-05, 'epoch': 0.16}) +[2026-06-17 17:22:25,926][starVLA.training.train_starvla][INFO] - Step 382, Loss: {'train/loss': 1.0799659490585327, 'train/grad_norm_pre_clip': 16.49148178100586, 'timing/data': 0.006631101481616497, 'timing/model': 1.1832642471417785, 'learning_rate/qwen_vl_interface': 9.960900906763478e-06, 'learning_rate/action_model': 9.960900906763477e-05, 'epoch': 0.16}) +[2026-06-17 17:22:29,823][starVLA.training.train_starvla][INFO] - Step 383, Loss: {'train/loss': 0.9842439293861389, 'train/grad_norm_pre_clip': 4.61693811416626, 'timing/data': 0.013168319128453732, 'timing/model': 1.0199582818895578, 'learning_rate/qwen_vl_interface': 9.960623501005108e-06, 'learning_rate/action_model': 9.960623501005107e-05, 'epoch': 0.16}) +[2026-06-17 17:22:33,784][starVLA.training.train_starvla][INFO] - Step 384, Loss: {'train/loss': 0.9048459529876709, 'train/grad_norm_pre_clip': 5.911394119262695, 'timing/data': 0.00964722316712141, 'timing/model': 0.9353229766711593, 'learning_rate/qwen_vl_interface': 9.960345118729416e-06, 'learning_rate/action_model': 9.960345118729417e-05, 'epoch': 0.16}) +[2026-06-17 17:22:38,217][starVLA.training.train_starvla][INFO] - Step 385, Loss: {'train/loss': 0.8718271255493164, 'train/grad_norm_pre_clip': 4.115174770355225, 'timing/data': 0.011979915201663971, 'timing/model': 1.1967743085697293, 'learning_rate/qwen_vl_interface': 9.960065759994116e-06, 'learning_rate/action_model': 9.960065759994114e-05, 'epoch': 0.16}) +[2026-06-17 17:22:42,089][starVLA.training.train_starvla][INFO] - Step 386, Loss: {'train/loss': 0.7630161046981812, 'train/grad_norm_pre_clip': 5.306571006774902, 'timing/data': 0.01759959664195776, 'timing/model': 0.941718720830977, 'learning_rate/qwen_vl_interface': 9.959785424857113e-06, 'learning_rate/action_model': 9.959785424857112e-05, 'epoch': 0.16}) +[2026-06-17 17:22:46,194][starVLA.training.train_starvla][INFO] - Step 387, Loss: {'train/loss': 0.9473713040351868, 'train/grad_norm_pre_clip': 5.391946315765381, 'timing/data': 0.010571314021945, 'timing/model': 0.9824059326201677, 'learning_rate/qwen_vl_interface': 9.959504113376523e-06, 'learning_rate/action_model': 9.959504113376523e-05, 'epoch': 0.16}) +[2026-06-17 17:22:50,769][starVLA.training.train_starvla][INFO] - Step 388, Loss: {'train/loss': 0.9256943464279175, 'train/grad_norm_pre_clip': 5.179825782775879, 'timing/data': 0.012051014229655266, 'timing/model': 1.4914568969979882, 'learning_rate/qwen_vl_interface': 9.959221825610664e-06, 'learning_rate/action_model': 9.959221825610663e-05, 'epoch': 0.16}) +[2026-06-17 17:22:56,521][starVLA.training.train_starvla][INFO] - Step 389, Loss: {'train/loss': 1.2176626920700073, 'train/grad_norm_pre_clip': 5.78532075881958, 'timing/data': 0.008437040261924267, 'timing/model': 2.9457456432282925, 'learning_rate/qwen_vl_interface': 9.958938561618053e-06, 'learning_rate/action_model': 9.958938561618053e-05, 'epoch': 0.16}) +[2026-06-17 17:23:00,399][starVLA.training.train_starvla][INFO] - Step 390, Loss: {'train/loss': 1.1171343326568604, 'train/grad_norm_pre_clip': 6.228641510009766, 'timing/data': 0.013500199653208256, 'timing/model': 0.952122276648879, 'learning_rate/qwen_vl_interface': 9.958654321457412e-06, 'learning_rate/action_model': 9.958654321457412e-05, 'epoch': 0.16}) +[2026-06-17 17:23:04,315][starVLA.training.train_starvla][INFO] - Step 391, Loss: {'train/loss': 0.6870694756507874, 'train/grad_norm_pre_clip': 4.780137062072754, 'timing/data': 0.00729712750762701, 'timing/model': 0.9832796072587371, 'learning_rate/qwen_vl_interface': 9.958369105187661e-06, 'learning_rate/action_model': 9.958369105187661e-05, 'epoch': 0.16}) +[2026-06-17 17:23:09,673][starVLA.training.train_starvla][INFO] - Step 392, Loss: {'train/loss': 0.7567085027694702, 'train/grad_norm_pre_clip': 4.8521833419799805, 'timing/data': 0.00944120530039072, 'timing/model': 0.9525789692997932, 'learning_rate/qwen_vl_interface': 9.958082912867932e-06, 'learning_rate/action_model': 9.958082912867931e-05, 'epoch': 0.16}) +[2026-06-17 17:23:13,458][starVLA.training.train_starvla][INFO] - Step 393, Loss: {'train/loss': 0.914644181728363, 'train/grad_norm_pre_clip': 5.52445650100708, 'timing/data': 0.008598178625106812, 'timing/model': 0.9431265955790877, 'learning_rate/qwen_vl_interface': 9.957795744557545e-06, 'learning_rate/action_model': 9.957795744557546e-05, 'epoch': 0.16}) +[2026-06-17 17:23:17,607][starVLA.training.train_starvla][INFO] - Step 394, Loss: {'train/loss': 0.6636749505996704, 'train/grad_norm_pre_clip': 5.417159080505371, 'timing/data': 0.01041136309504509, 'timing/model': 0.9563011098653078, 'learning_rate/qwen_vl_interface': 9.957507600316036e-06, 'learning_rate/action_model': 9.957507600316036e-05, 'epoch': 0.16}) +[2026-06-17 17:23:21,490][starVLA.training.train_starvla][INFO] - Step 395, Loss: {'train/loss': 0.850659191608429, 'train/grad_norm_pre_clip': 3.9162542819976807, 'timing/data': 0.009584261104464531, 'timing/model': 1.123522107489407, 'learning_rate/qwen_vl_interface': 9.957218480203137e-06, 'learning_rate/action_model': 9.957218480203136e-05, 'epoch': 0.16}) +[2026-06-17 17:23:25,115][starVLA.training.train_starvla][INFO] - Step 396, Loss: {'train/loss': 0.9139876365661621, 'train/grad_norm_pre_clip': 7.4194183349609375, 'timing/data': 0.010679662227630615, 'timing/model': 0.9246000992134213, 'learning_rate/qwen_vl_interface': 9.95692838427878e-06, 'learning_rate/action_model': 9.956928384278778e-05, 'epoch': 0.16}) +[2026-06-17 17:23:29,028][starVLA.training.train_starvla][INFO] - Step 397, Loss: {'train/loss': 0.860065221786499, 'train/grad_norm_pre_clip': 5.976511478424072, 'timing/data': 0.009768090210855007, 'timing/model': 1.0352840013802052, 'learning_rate/qwen_vl_interface': 9.956637312603105e-06, 'learning_rate/action_model': 9.956637312603104e-05, 'epoch': 0.16}) +[2026-06-17 17:23:32,763][starVLA.training.train_starvla][INFO] - Step 398, Loss: {'train/loss': 0.7507843375205994, 'train/grad_norm_pre_clip': 4.681503772735596, 'timing/data': 0.01963159069418907, 'timing/model': 0.9406653698533773, 'learning_rate/qwen_vl_interface': 9.95634526523645e-06, 'learning_rate/action_model': 9.956345265236449e-05, 'epoch': 0.16}) +[2026-06-17 17:23:37,348][starVLA.training.train_starvla][INFO] - Step 399, Loss: {'train/loss': 0.7972618937492371, 'train/grad_norm_pre_clip': 6.0549211502075195, 'timing/data': 0.011590790003538132, 'timing/model': 1.66136987041682, 'learning_rate/qwen_vl_interface': 9.956052242239358e-06, 'learning_rate/action_model': 9.956052242239358e-05, 'epoch': 0.16}) +[2026-06-17 17:26:00,134][starVLA.training.train_starvla][INFO] - Step 400, Loss: {'train/loss': 0.7263655066490173, 'train/grad_norm_pre_clip': 5.220306873321533, 'eval/loss': 0.9545953273773193, 'eval/latency_6/loss': 0.9545953273773193, 'eval/demon_fire_precision': 0.9227129337539433, 'eval/demon_fire_recall': 0.9453781512605042, 'eval/demon_fire_f1': 0.9339080459770116, 'eval/demon_fire_rate_model': 0.8820255982192543, 'eval/demon_fire_rate_teacher': 0.860879243183083, 'eval/demon_left_precision': 0.9381443298969072, 'eval/demon_left_recall': 0.15964912280701754, 'eval/demon_left_f1': 0.272863568215892, 'eval/demon_left_rate_model': 0.026989426822481913, 'eval/demon_left_rate_teacher': 0.15859766277128548, 'eval/demon_right_precision': 0.7800829875518672, 'eval/demon_right_recall': 0.2887864823348694, 'eval/demon_right_f1': 0.42152466367713, 'eval/demon_right_rate_model': 0.06705620478575404, 'eval/demon_right_rate_teacher': 0.18113522537562604, 'eval/demon_move_f1': 0.347194115946511, 'eval/demon_cc_f1': 0.6197290477319366, 'eval/latency_6/demon_fire_precision': 0.9227129337539433, 'eval/latency_6/demon_fire_recall': 0.9453781512605042, 'eval/latency_6/demon_fire_f1': 0.9339080459770116, 'eval/latency_6/demon_fire_rate_model': 0.8820255982192543, 'eval/latency_6/demon_fire_rate_teacher': 0.860879243183083, 'eval/latency_6/demon_left_precision': 0.9381443298969072, 'eval/latency_6/demon_left_recall': 0.15964912280701754, 'eval/latency_6/demon_left_f1': 0.272863568215892, 'eval/latency_6/demon_left_rate_model': 0.026989426822481913, 'eval/latency_6/demon_left_rate_teacher': 0.15859766277128548, 'eval/latency_6/demon_right_precision': 0.7800829875518672, 'eval/latency_6/demon_right_recall': 0.2887864823348694, 'eval/latency_6/demon_right_f1': 0.42152466367713, 'eval/latency_6/demon_right_rate_model': 0.06705620478575404, 'eval/latency_6/demon_right_rate_teacher': 0.18113522537562604, 'eval/latency_6/demon_move_f1': 0.347194115946511, 'eval/latency_6/demon_cc_f1': 0.6197290477319366, 'timing/data': 0.010902609676122665, 'timing/model': 3.345526876859367, 'learning_rate/qwen_vl_interface': 9.955758243672572e-06, 'learning_rate/action_model': 9.955758243672572e-05, 'epoch': 0.16}) +[2026-06-17 17:26:00,136][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_400_state +[2026-06-17 17:26:00,146][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 17:26:07,312][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_400_state/model.safetensors +[2026-06-17 17:26:23,852][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_400_state/optimizer.bin +[2026-06-17 17:26:23,854][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_400_state/sampler.bin +[2026-06-17 17:26:23,856][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_400_state/sampler_1.bin +[2026-06-17 17:26:23,859][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_400_state/random_states_0.pkl +[2026-06-17 17:26:23,863][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 400 +[2026-06-17 17:26:23,863][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 17:26:23,874][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 17:26:27,804][starVLA.training.train_starvla][INFO] - Step 401, Loss: {'train/loss': 0.9371806979179382, 'train/grad_norm_pre_clip': 4.83161735534668, 'timing/data': 0.00791352242231369, 'timing/model': 1.003983622416854, 'learning_rate/qwen_vl_interface': 9.955463269597038e-06, 'learning_rate/action_model': 9.955463269597038e-05, 'epoch': 0.17}) +[2026-06-17 17:26:31,745][starVLA.training.train_starvla][INFO] - Step 402, Loss: {'train/loss': 1.1988580226898193, 'train/grad_norm_pre_clip': 7.784094333648682, 'timing/data': 0.009241796098649502, 'timing/model': 0.9457773007452488, 'learning_rate/qwen_vl_interface': 9.955167320073905e-06, 'learning_rate/action_model': 9.955167320073905e-05, 'epoch': 0.17}) +[2026-06-17 17:26:35,694][starVLA.training.train_starvla][INFO] - Step 403, Loss: {'train/loss': 1.331876516342163, 'train/grad_norm_pre_clip': 7.984849452972412, 'timing/data': 0.010504020377993584, 'timing/model': 0.9490309478715062, 'learning_rate/qwen_vl_interface': 9.954870395164524e-06, 'learning_rate/action_model': 9.954870395164524e-05, 'epoch': 0.17}) +[2026-06-17 17:26:39,630][starVLA.training.train_starvla][INFO] - Step 404, Loss: {'train/loss': 1.0513668060302734, 'train/grad_norm_pre_clip': 6.861976623535156, 'timing/data': 0.009286543354392052, 'timing/model': 1.1830505328252912, 'learning_rate/qwen_vl_interface': 9.954572494930448e-06, 'learning_rate/action_model': 9.954572494930448e-05, 'epoch': 0.17}) +[2026-06-17 17:26:43,417][starVLA.training.train_starvla][INFO] - Step 405, Loss: {'train/loss': 0.9171653985977173, 'train/grad_norm_pre_clip': 7.520148277282715, 'timing/data': 0.009565377607941628, 'timing/model': 1.0382921677082777, 'learning_rate/qwen_vl_interface': 9.95427361943343e-06, 'learning_rate/action_model': 9.954273619433431e-05, 'epoch': 0.17}) +[2026-06-17 17:26:47,561][starVLA.training.train_starvla][INFO] - Step 406, Loss: {'train/loss': 1.002213478088379, 'train/grad_norm_pre_clip': 5.64812707901001, 'timing/data': 0.00708810705691576, 'timing/model': 0.9458799790591002, 'learning_rate/qwen_vl_interface': 9.95397376873543e-06, 'learning_rate/action_model': 9.95397376873543e-05, 'epoch': 0.17}) +[2026-06-17 17:26:51,718][starVLA.training.train_starvla][INFO] - Step 407, Loss: {'train/loss': 0.7825546264648438, 'train/grad_norm_pre_clip': 5.447281360626221, 'timing/data': 0.007055444642901421, 'timing/model': 0.9257187098264694, 'learning_rate/qwen_vl_interface': 9.953672942898606e-06, 'learning_rate/action_model': 9.953672942898605e-05, 'epoch': 0.17}) +[2026-06-17 17:26:55,544][starVLA.training.train_starvla][INFO] - Step 408, Loss: {'train/loss': 1.2078536748886108, 'train/grad_norm_pre_clip': 9.433669090270996, 'timing/data': 0.010791189037263393, 'timing/model': 0.9575586849823594, 'learning_rate/qwen_vl_interface': 9.953371141985318e-06, 'learning_rate/action_model': 9.953371141985319e-05, 'epoch': 0.17}) +[2026-06-17 17:26:59,563][starVLA.training.train_starvla][INFO] - Step 409, Loss: {'train/loss': 0.7876221537590027, 'train/grad_norm_pre_clip': 5.87847375869751, 'timing/data': 0.010523738339543343, 'timing/model': 0.9904003692790866, 'learning_rate/qwen_vl_interface': 9.953068366058133e-06, 'learning_rate/action_model': 9.953068366058133e-05, 'epoch': 0.17}) +[2026-06-17 17:27:03,899][starVLA.training.train_starvla][INFO] - Step 410, Loss: {'train/loss': 0.9072302579879761, 'train/grad_norm_pre_clip': 3.466850757598877, 'timing/data': 0.0092705013230443, 'timing/model': 0.931362584233284, 'learning_rate/qwen_vl_interface': 9.952764615179813e-06, 'learning_rate/action_model': 9.952764615179812e-05, 'epoch': 0.17}) +[2026-06-17 17:27:07,749][starVLA.training.train_starvla][INFO] - Step 411, Loss: {'train/loss': 0.872046172618866, 'train/grad_norm_pre_clip': 11.27094841003418, 'timing/data': 0.00993333663791418, 'timing/model': 1.171874899417162, 'learning_rate/qwen_vl_interface': 9.95245988941333e-06, 'learning_rate/action_model': 9.95245988941333e-05, 'epoch': 0.17}) +[2026-06-17 17:27:11,411][starVLA.training.train_starvla][INFO] - Step 412, Loss: {'train/loss': 1.0375367403030396, 'train/grad_norm_pre_clip': 5.779135227203369, 'timing/data': 0.008628278970718384, 'timing/model': 0.9244106775149703, 'learning_rate/qwen_vl_interface': 9.952154188821852e-06, 'learning_rate/action_model': 9.952154188821851e-05, 'epoch': 0.17}) +[2026-06-17 17:27:15,249][starVLA.training.train_starvla][INFO] - Step 413, Loss: {'train/loss': 1.1408231258392334, 'train/grad_norm_pre_clip': 6.402139663696289, 'timing/data': 0.011108219623565674, 'timing/model': 0.9264021180570126, 'learning_rate/qwen_vl_interface': 9.95184751346875e-06, 'learning_rate/action_model': 9.95184751346875e-05, 'epoch': 0.17}) +[2026-06-17 17:27:19,266][starVLA.training.train_starvla][INFO] - Step 414, Loss: {'train/loss': 0.7943485975265503, 'train/grad_norm_pre_clip': 10.49790096282959, 'timing/data': 0.009172149933874607, 'timing/model': 0.9461144423112273, 'learning_rate/qwen_vl_interface': 9.951539863417598e-06, 'learning_rate/action_model': 9.951539863417599e-05, 'epoch': 0.17}) +[2026-06-17 17:27:23,350][starVLA.training.train_starvla][INFO] - Step 415, Loss: {'train/loss': 0.8993294835090637, 'train/grad_norm_pre_clip': 5.511012554168701, 'timing/data': 0.010039453394711018, 'timing/model': 1.1181746814399958, 'learning_rate/qwen_vl_interface': 9.951231238732174e-06, 'learning_rate/action_model': 9.951231238732173e-05, 'epoch': 0.17}) +[2026-06-17 17:27:27,238][starVLA.training.train_starvla][INFO] - Step 416, Loss: {'train/loss': 0.9826258420944214, 'train/grad_norm_pre_clip': 5.946445941925049, 'timing/data': 0.013516361825168133, 'timing/model': 0.9717648951336741, 'learning_rate/qwen_vl_interface': 9.950921639476455e-06, 'learning_rate/action_model': 9.950921639476455e-05, 'epoch': 0.17}) +[2026-06-17 17:27:31,144][starVLA.training.train_starvla][INFO] - Step 417, Loss: {'train/loss': 0.8480111360549927, 'train/grad_norm_pre_clip': 6.089911937713623, 'timing/data': 0.012019741348922253, 'timing/model': 1.1548018883913755, 'learning_rate/qwen_vl_interface': 9.950611065714623e-06, 'learning_rate/action_model': 9.950611065714623e-05, 'epoch': 0.17}) +[2026-06-17 17:27:35,169][starVLA.training.train_starvla][INFO] - Step 418, Loss: {'train/loss': 0.8808249831199646, 'train/grad_norm_pre_clip': 6.014618873596191, 'timing/data': 0.011298893019557, 'timing/model': 1.1759629277512431, 'learning_rate/qwen_vl_interface': 9.950299517511058e-06, 'learning_rate/action_model': 9.950299517511058e-05, 'epoch': 0.17}) +[2026-06-17 17:27:39,149][starVLA.training.train_starvla][INFO] - Step 419, Loss: {'train/loss': 0.9357298612594604, 'train/grad_norm_pre_clip': 3.621469259262085, 'timing/data': 0.01426113024353981, 'timing/model': 0.9781312542036176, 'learning_rate/qwen_vl_interface': 9.949986994930346e-06, 'learning_rate/action_model': 9.949986994930346e-05, 'epoch': 0.17}) +[2026-06-17 17:27:43,118][starVLA.training.train_starvla][INFO] - Step 420, Loss: {'train/loss': 0.9626823663711548, 'train/grad_norm_pre_clip': 5.02434778213501, 'timing/data': 0.009610394015908241, 'timing/model': 0.9633796140551567, 'learning_rate/qwen_vl_interface': 9.949673498037272e-06, 'learning_rate/action_model': 9.949673498037272e-05, 'epoch': 0.17}) +[2026-06-17 17:27:47,198][starVLA.training.train_starvla][INFO] - Step 421, Loss: {'train/loss': 0.6720459461212158, 'train/grad_norm_pre_clip': 4.267676830291748, 'timing/data': 0.011298569850623608, 'timing/model': 1.0027202563360333, 'learning_rate/qwen_vl_interface': 9.949359026896826e-06, 'learning_rate/action_model': 9.949359026896826e-05, 'epoch': 0.17}) +[2026-06-17 17:27:51,085][starVLA.training.train_starvla][INFO] - Step 422, Loss: {'train/loss': 0.7466602325439453, 'train/grad_norm_pre_clip': 4.692008972167969, 'timing/data': 0.00872411485761404, 'timing/model': 0.9519103998318315, 'learning_rate/qwen_vl_interface': 9.949043581574196e-06, 'learning_rate/action_model': 9.949043581574196e-05, 'epoch': 0.17}) +[2026-06-17 17:27:55,005][starVLA.training.train_starvla][INFO] - Step 423, Loss: {'train/loss': 0.7013243436813354, 'train/grad_norm_pre_clip': 12.249846458435059, 'timing/data': 0.008085216395556927, 'timing/model': 0.9761359849944711, 'learning_rate/qwen_vl_interface': 9.948727162134775e-06, 'learning_rate/action_model': 9.948727162134775e-05, 'epoch': 0.17}) +[2026-06-17 17:27:59,010][starVLA.training.train_starvla][INFO] - Step 424, Loss: {'train/loss': 0.9047554731369019, 'train/grad_norm_pre_clip': 4.340712547302246, 'timing/data': 0.007952926680445671, 'timing/model': 1.2365046357735991, 'learning_rate/qwen_vl_interface': 9.948409768644157e-06, 'learning_rate/action_model': 9.948409768644156e-05, 'epoch': 0.17}) +[2026-06-17 17:28:02,922][starVLA.training.train_starvla][INFO] - Step 425, Loss: {'train/loss': 1.6521847248077393, 'train/grad_norm_pre_clip': 12.078049659729004, 'timing/data': 0.009109551087021828, 'timing/model': 1.1222654795274138, 'learning_rate/qwen_vl_interface': 9.948091401168139e-06, 'learning_rate/action_model': 9.948091401168138e-05, 'epoch': 0.17}) +[2026-06-17 17:28:06,747][starVLA.training.train_starvla][INFO] - Step 426, Loss: {'train/loss': 0.8524062037467957, 'train/grad_norm_pre_clip': 4.082324028015137, 'timing/data': 0.010322410613298416, 'timing/model': 0.9524775519967079, 'learning_rate/qwen_vl_interface': 9.947772059772718e-06, 'learning_rate/action_model': 9.947772059772717e-05, 'epoch': 0.18}) +[2026-06-17 17:28:10,651][starVLA.training.train_starvla][INFO] - Step 427, Loss: {'train/loss': 0.6745243668556213, 'train/grad_norm_pre_clip': 6.773190021514893, 'timing/data': 0.015661150217056274, 'timing/model': 0.9311752635985613, 'learning_rate/qwen_vl_interface': 9.947451744524094e-06, 'learning_rate/action_model': 9.947451744524093e-05, 'epoch': 0.18}) +[2026-06-17 17:28:14,420][starVLA.training.train_starvla][INFO] - Step 428, Loss: {'train/loss': 0.7895666360855103, 'train/grad_norm_pre_clip': 8.620501518249512, 'timing/data': 0.010636454448103905, 'timing/model': 0.9414988243952394, 'learning_rate/qwen_vl_interface': 9.947130455488668e-06, 'learning_rate/action_model': 9.947130455488666e-05, 'epoch': 0.18}) +[2026-06-17 17:28:18,378][starVLA.training.train_starvla][INFO] - Step 429, Loss: {'train/loss': 1.0594788789749146, 'train/grad_norm_pre_clip': 6.131842613220215, 'timing/data': 0.017203226685523987, 'timing/model': 0.9614482801407576, 'learning_rate/qwen_vl_interface': 9.946808192733045e-06, 'learning_rate/action_model': 9.946808192733044e-05, 'epoch': 0.18}) +[2026-06-17 17:28:22,336][starVLA.training.train_starvla][INFO] - Step 430, Loss: {'train/loss': 1.0210944414138794, 'train/grad_norm_pre_clip': 5.20814323425293, 'timing/data': 0.014224935322999954, 'timing/model': 0.9456364326179028, 'learning_rate/qwen_vl_interface': 9.946484956324028e-06, 'learning_rate/action_model': 9.946484956324028e-05, 'epoch': 0.18}) +[2026-06-17 17:28:26,406][starVLA.training.train_starvla][INFO] - Step 431, Loss: {'train/loss': 1.0123581886291504, 'train/grad_norm_pre_clip': 7.234260082244873, 'timing/data': 0.008421600796282291, 'timing/model': 1.2771072247996926, 'learning_rate/qwen_vl_interface': 9.946160746328628e-06, 'learning_rate/action_model': 9.946160746328627e-05, 'epoch': 0.18}) +[2026-06-17 17:28:30,215][starVLA.training.train_starvla][INFO] - Step 432, Loss: {'train/loss': 0.8822365999221802, 'train/grad_norm_pre_clip': 5.281203269958496, 'timing/data': 0.012132576666772366, 'timing/model': 0.9463031636551023, 'learning_rate/qwen_vl_interface': 9.945835562814052e-06, 'learning_rate/action_model': 9.945835562814051e-05, 'epoch': 0.18}) +[2026-06-17 17:28:34,462][starVLA.training.train_starvla][INFO] - Step 433, Loss: {'train/loss': 0.8574592471122742, 'train/grad_norm_pre_clip': 6.452758312225342, 'timing/data': 0.019922276958823204, 'timing/model': 0.9544324763119221, 'learning_rate/qwen_vl_interface': 9.94550940584771e-06, 'learning_rate/action_model': 9.945509405847709e-05, 'epoch': 0.18}) +[2026-06-17 17:28:38,559][starVLA.training.train_starvla][INFO] - Step 434, Loss: {'train/loss': 1.1046509742736816, 'train/grad_norm_pre_clip': 9.854415893554688, 'timing/data': 0.011420127004384995, 'timing/model': 0.964698439463973, 'learning_rate/qwen_vl_interface': 9.945182275497215e-06, 'learning_rate/action_model': 9.945182275497215e-05, 'epoch': 0.18}) +[2026-06-17 17:28:42,509][starVLA.training.train_starvla][INFO] - Step 435, Loss: {'train/loss': 0.85541832447052, 'train/grad_norm_pre_clip': 8.457954406738281, 'timing/data': 0.011188353411853313, 'timing/model': 1.0019421074539423, 'learning_rate/qwen_vl_interface': 9.944854171830382e-06, 'learning_rate/action_model': 9.944854171830383e-05, 'epoch': 0.18}) +[2026-06-17 17:28:46,427][starVLA.training.train_starvla][INFO] - Step 436, Loss: {'train/loss': 0.9464181661605835, 'train/grad_norm_pre_clip': 5.55089807510376, 'timing/data': 0.01742707286030054, 'timing/model': 0.9687461163848639, 'learning_rate/qwen_vl_interface': 9.94452509491523e-06, 'learning_rate/action_model': 9.94452509491523e-05, 'epoch': 0.18}) +[2026-06-17 17:28:50,330][starVLA.training.train_starvla][INFO] - Step 437, Loss: {'train/loss': 1.001469373703003, 'train/grad_norm_pre_clip': 6.026028633117676, 'timing/data': 0.009430371224880219, 'timing/model': 1.172346342355013, 'learning_rate/qwen_vl_interface': 9.944195044819973e-06, 'learning_rate/action_model': 9.944195044819972e-05, 'epoch': 0.18}) +[2026-06-17 17:28:54,366][starVLA.training.train_starvla][INFO] - Step 438, Loss: {'train/loss': 1.133847951889038, 'train/grad_norm_pre_clip': 5.670760631561279, 'timing/data': 0.009462264366447926, 'timing/model': 1.2088293759152293, 'learning_rate/qwen_vl_interface': 9.94386402161303e-06, 'learning_rate/action_model': 9.943864021613029e-05, 'epoch': 0.18}) +[2026-06-17 17:28:58,044][starVLA.training.train_starvla][INFO] - Step 439, Loss: {'train/loss': 0.7736909985542297, 'train/grad_norm_pre_clip': 6.025125503540039, 'timing/data': 0.010387045331299305, 'timing/model': 0.9278882313519716, 'learning_rate/qwen_vl_interface': 9.943532025363026e-06, 'learning_rate/action_model': 9.943532025363025e-05, 'epoch': 0.18}) +[2026-06-17 17:29:02,040][starVLA.training.train_starvla][INFO] - Step 440, Loss: {'train/loss': 1.0095298290252686, 'train/grad_norm_pre_clip': 6.706571102142334, 'timing/data': 0.010365411639213562, 'timing/model': 0.9447457743808627, 'learning_rate/qwen_vl_interface': 9.943199056138781e-06, 'learning_rate/action_model': 9.943199056138782e-05, 'epoch': 0.18}) +[2026-06-17 17:29:06,363][starVLA.training.train_starvla][INFO] - Step 441, Loss: {'train/loss': 1.0730057954788208, 'train/grad_norm_pre_clip': 5.683958053588867, 'timing/data': 0.009593483991920948, 'timing/model': 0.9605083111673594, 'learning_rate/qwen_vl_interface': 9.942865114009325e-06, 'learning_rate/action_model': 9.942865114009324e-05, 'epoch': 0.18}) +[2026-06-17 17:29:10,188][starVLA.training.train_starvla][INFO] - Step 442, Loss: {'train/loss': 0.8656884431838989, 'train/grad_norm_pre_clip': 6.525867938995361, 'timing/data': 0.018009664490818977, 'timing/model': 0.9367024078965187, 'learning_rate/qwen_vl_interface': 9.94253019904388e-06, 'learning_rate/action_model': 9.942530199043879e-05, 'epoch': 0.18}) +[2026-06-17 17:29:14,046][starVLA.training.train_starvla][INFO] - Step 443, Loss: {'train/loss': 0.9543613195419312, 'train/grad_norm_pre_clip': 5.88787317276001, 'timing/data': 0.009457018226385117, 'timing/model': 0.9302562894299626, 'learning_rate/qwen_vl_interface': 9.942194311311874e-06, 'learning_rate/action_model': 9.942194311311874e-05, 'epoch': 0.18}) +[2026-06-17 17:29:18,063][starVLA.training.train_starvla][INFO] - Step 444, Loss: {'train/loss': 0.9111111164093018, 'train/grad_norm_pre_clip': 27.093250274658203, 'timing/data': 0.010039245709776878, 'timing/model': 1.3043989418074489, 'learning_rate/qwen_vl_interface': 9.941857450882939e-06, 'learning_rate/action_model': 9.941857450882937e-05, 'epoch': 0.18}) +[2026-06-17 17:29:22,359][starVLA.training.train_starvla][INFO] - Step 445, Loss: {'train/loss': 0.8672192096710205, 'train/grad_norm_pre_clip': 4.878993034362793, 'timing/data': 0.0084415627643466, 'timing/model': 1.3619220750406384, 'learning_rate/qwen_vl_interface': 9.941519617826906e-06, 'learning_rate/action_model': 9.941519617826906e-05, 'epoch': 0.18}) +[2026-06-17 17:29:26,040][starVLA.training.train_starvla][INFO] - Step 446, Loss: {'train/loss': 0.8289043307304382, 'train/grad_norm_pre_clip': 6.852829456329346, 'timing/data': 0.008337759412825108, 'timing/model': 0.9442864116281271, 'learning_rate/qwen_vl_interface': 9.941180812213807e-06, 'learning_rate/action_model': 9.941180812213806e-05, 'epoch': 0.18}) +[2026-06-17 17:29:29,832][starVLA.training.train_starvla][INFO] - Step 447, Loss: {'train/loss': 0.9876192808151245, 'train/grad_norm_pre_clip': 4.468613624572754, 'timing/data': 0.00965544767677784, 'timing/model': 0.9238839028403163, 'learning_rate/qwen_vl_interface': 9.940841034113878e-06, 'learning_rate/action_model': 9.940841034113877e-05, 'epoch': 0.18}) +[2026-06-17 17:29:33,713][starVLA.training.train_starvla][INFO] - Step 448, Loss: {'train/loss': 0.7720854878425598, 'train/grad_norm_pre_clip': 3.7909584045410156, 'timing/data': 0.00878132227808237, 'timing/model': 0.930591868236661, 'learning_rate/qwen_vl_interface': 9.940500283597557e-06, 'learning_rate/action_model': 9.940500283597556e-05, 'epoch': 0.18}) +[2026-06-17 17:29:37,588][starVLA.training.train_starvla][INFO] - Step 449, Loss: {'train/loss': 0.9220399856567383, 'train/grad_norm_pre_clip': 5.245026111602783, 'timing/data': 0.010411771945655346, 'timing/model': 0.9495295314118266, 'learning_rate/qwen_vl_interface': 9.940158560735478e-06, 'learning_rate/action_model': 9.940158560735477e-05, 'epoch': 0.18}) +[2026-06-17 17:29:41,706][starVLA.training.train_starvla][INFO] - Step 450, Loss: {'train/loss': 1.0454877614974976, 'train/grad_norm_pre_clip': 7.90311336517334, 'timing/data': 0.011828056536614895, 'timing/model': 1.0125666419044137, 'learning_rate/qwen_vl_interface': 9.939815865598482e-06, 'learning_rate/action_model': 9.93981586559848e-05, 'epoch': 0.19}) +[2026-06-17 17:29:45,627][starVLA.training.train_starvla][INFO] - Step 451, Loss: {'train/loss': 1.1297954320907593, 'train/grad_norm_pre_clip': 6.980548858642578, 'timing/data': 0.010057258419692516, 'timing/model': 1.1947941463440657, 'learning_rate/qwen_vl_interface': 9.939472198257611e-06, 'learning_rate/action_model': 9.93947219825761e-05, 'epoch': 0.19}) +[2026-06-17 17:29:49,293][starVLA.training.train_starvla][INFO] - Step 452, Loss: {'train/loss': 0.8785340189933777, 'train/grad_norm_pre_clip': 4.723301410675049, 'timing/data': 0.007746036164462566, 'timing/model': 0.9460750976577401, 'learning_rate/qwen_vl_interface': 9.939127558784108e-06, 'learning_rate/action_model': 9.939127558784107e-05, 'epoch': 0.19}) +[2026-06-17 17:29:53,197][starVLA.training.train_starvla][INFO] - Step 453, Loss: {'train/loss': 0.8110120892524719, 'train/grad_norm_pre_clip': 4.083715438842773, 'timing/data': 0.008638158440589905, 'timing/model': 0.9471495458856225, 'learning_rate/qwen_vl_interface': 9.938781947249417e-06, 'learning_rate/action_model': 9.938781947249416e-05, 'epoch': 0.19}) +[2026-06-17 17:29:57,196][starVLA.training.train_starvla][INFO] - Step 454, Loss: {'train/loss': 0.858218252658844, 'train/grad_norm_pre_clip': 4.754546642303467, 'timing/data': 0.01955690234899521, 'timing/model': 1.0199742270633578, 'learning_rate/qwen_vl_interface': 9.93843536372518e-06, 'learning_rate/action_model': 9.93843536372518e-05, 'epoch': 0.19}) +[2026-06-17 17:30:01,459][starVLA.training.train_starvla][INFO] - Step 455, Loss: {'train/loss': 0.8598190546035767, 'train/grad_norm_pre_clip': 4.073235511779785, 'timing/data': 0.009492892771959305, 'timing/model': 0.9296806482598186, 'learning_rate/qwen_vl_interface': 9.938087808283248e-06, 'learning_rate/action_model': 9.938087808283248e-05, 'epoch': 0.19}) +[2026-06-17 17:30:05,443][starVLA.training.train_starvla][INFO] - Step 456, Loss: {'train/loss': 0.9710469841957092, 'train/grad_norm_pre_clip': 4.71563196182251, 'timing/data': 0.010338333435356617, 'timing/model': 0.9487797552719712, 'learning_rate/qwen_vl_interface': 9.937739280995672e-06, 'learning_rate/action_model': 9.937739280995672e-05, 'epoch': 0.19}) +[2026-06-17 17:30:09,927][starVLA.training.train_starvla][INFO] - Step 457, Loss: {'train/loss': 1.0249862670898438, 'train/grad_norm_pre_clip': 5.951079368591309, 'timing/data': 0.010701196268200874, 'timing/model': 1.1707529602572322, 'learning_rate/qwen_vl_interface': 9.937389781934697e-06, 'learning_rate/action_model': 9.937389781934696e-05, 'epoch': 0.19}) +[2026-06-17 17:30:14,804][starVLA.training.train_starvla][INFO] - Step 458, Loss: {'train/loss': 0.9111879467964172, 'train/grad_norm_pre_clip': 5.20176887512207, 'timing/data': 0.009890297427773476, 'timing/model': 1.1329212738201022, 'learning_rate/qwen_vl_interface': 9.937039311172776e-06, 'learning_rate/action_model': 9.937039311172776e-05, 'epoch': 0.19}) +[2026-06-17 17:30:18,661][starVLA.training.train_starvla][INFO] - Step 459, Loss: {'train/loss': 0.7583003044128418, 'train/grad_norm_pre_clip': 4.568290710449219, 'timing/data': 0.01264218334108591, 'timing/model': 1.016549470834434, 'learning_rate/qwen_vl_interface': 9.936687868782562e-06, 'learning_rate/action_model': 9.936687868782561e-05, 'epoch': 0.19}) +[2026-06-17 17:30:22,571][starVLA.training.train_starvla][INFO] - Step 460, Loss: {'train/loss': 0.7975368499755859, 'train/grad_norm_pre_clip': 3.725216865539551, 'timing/data': 0.010640027932822704, 'timing/model': 0.9741085208952427, 'learning_rate/qwen_vl_interface': 9.93633545483691e-06, 'learning_rate/action_model': 9.93633545483691e-05, 'epoch': 0.19}) +[2026-06-17 17:30:26,878][starVLA.training.train_starvla][INFO] - Step 461, Loss: {'train/loss': 0.7599043846130371, 'train/grad_norm_pre_clip': 4.5387983322143555, 'timing/data': 0.014050302095711231, 'timing/model': 0.9329228643327951, 'learning_rate/qwen_vl_interface': 9.935982069408877e-06, 'learning_rate/action_model': 9.935982069408877e-05, 'epoch': 0.19}) +[2026-06-17 17:30:30,695][starVLA.training.train_starvla][INFO] - Step 462, Loss: {'train/loss': 0.5715254545211792, 'train/grad_norm_pre_clip': 5.342864036560059, 'timing/data': 0.009586692787706852, 'timing/model': 0.9280755426734686, 'learning_rate/qwen_vl_interface': 9.935627712571716e-06, 'learning_rate/action_model': 9.935627712571715e-05, 'epoch': 0.19}) +[2026-06-17 17:30:34,574][starVLA.training.train_starvla][INFO] - Step 463, Loss: {'train/loss': 0.6115420460700989, 'train/grad_norm_pre_clip': 11.556761741638184, 'timing/data': 0.008531283587217331, 'timing/model': 0.9265472032129765, 'learning_rate/qwen_vl_interface': 9.93527238439889e-06, 'learning_rate/action_model': 9.935272384398889e-05, 'epoch': 0.19}) +[2026-06-17 17:30:38,359][starVLA.training.train_starvla][INFO] - Step 464, Loss: {'train/loss': 1.1099437475204468, 'train/grad_norm_pre_clip': 6.93626070022583, 'timing/data': 0.010855483822524548, 'timing/model': 1.1028259424492717, 'learning_rate/qwen_vl_interface': 9.934916084964056e-06, 'learning_rate/action_model': 9.934916084964056e-05, 'epoch': 0.19}) +[2026-06-17 17:30:42,223][starVLA.training.train_starvla][INFO] - Step 465, Loss: {'train/loss': 0.8629621863365173, 'train/grad_norm_pre_clip': 6.026217937469482, 'timing/data': 0.01857650838792324, 'timing/model': 1.086792984046042, 'learning_rate/qwen_vl_interface': 9.934558814341077e-06, 'learning_rate/action_model': 9.934558814341077e-05, 'epoch': 0.19}) +[2026-06-17 17:30:45,909][starVLA.training.train_starvla][INFO] - Step 466, Loss: {'train/loss': 0.9751178026199341, 'train/grad_norm_pre_clip': 7.920859336853027, 'timing/data': 0.00929077249020338, 'timing/model': 1.0328744007274508, 'learning_rate/qwen_vl_interface': 9.934200572604015e-06, 'learning_rate/action_model': 9.934200572604015e-05, 'epoch': 0.19}) +[2026-06-17 17:30:49,839][starVLA.training.train_starvla][INFO] - Step 467, Loss: {'train/loss': 0.9226689338684082, 'train/grad_norm_pre_clip': 13.328703880310059, 'timing/data': 0.012572386302053928, 'timing/model': 0.9267845181748271, 'learning_rate/qwen_vl_interface': 9.933841359827133e-06, 'learning_rate/action_model': 9.933841359827133e-05, 'epoch': 0.19}) +[2026-06-17 17:30:53,682][starVLA.training.train_starvla][INFO] - Step 468, Loss: {'train/loss': 0.7071545720100403, 'train/grad_norm_pre_clip': 8.301617622375488, 'timing/data': 0.008802161552011967, 'timing/model': 0.9380510309711099, 'learning_rate/qwen_vl_interface': 9.933481176084899e-06, 'learning_rate/action_model': 9.933481176084899e-05, 'epoch': 0.19}) +[2026-06-17 17:30:57,456][starVLA.training.train_starvla][INFO] - Step 469, Loss: {'train/loss': 1.0702707767486572, 'train/grad_norm_pre_clip': 4.860035419464111, 'timing/data': 0.011019209399819374, 'timing/model': 0.9304087515920401, 'learning_rate/qwen_vl_interface': 9.933120021451976e-06, 'learning_rate/action_model': 9.933120021451976e-05, 'epoch': 0.19}) +[2026-06-17 17:31:01,276][starVLA.training.train_starvla][INFO] - Step 470, Loss: {'train/loss': 0.5646106004714966, 'train/grad_norm_pre_clip': 5.309661865234375, 'timing/data': 0.007123264484107494, 'timing/model': 0.9331006472930312, 'learning_rate/qwen_vl_interface': 9.932757896003235e-06, 'learning_rate/action_model': 9.932757896003235e-05, 'epoch': 0.19}) +[2026-06-17 17:31:05,212][starVLA.training.train_starvla][INFO] - Step 471, Loss: {'train/loss': 0.9608416557312012, 'train/grad_norm_pre_clip': 5.483511924743652, 'timing/data': 0.0088787991553545, 'timing/model': 1.1491342829540372, 'learning_rate/qwen_vl_interface': 9.932394799813742e-06, 'learning_rate/action_model': 9.932394799813741e-05, 'epoch': 0.19}) +[2026-06-17 17:31:09,277][starVLA.training.train_starvla][INFO] - Step 472, Loss: {'train/loss': 0.9825370907783508, 'train/grad_norm_pre_clip': 7.522464752197266, 'timing/data': 0.010101447813212872, 'timing/model': 1.190497606061399, 'learning_rate/qwen_vl_interface': 9.932030732958768e-06, 'learning_rate/action_model': 9.932030732958769e-05, 'epoch': 0.19}) +[2026-06-17 17:31:13,421][starVLA.training.train_starvla][INFO] - Step 473, Loss: {'train/loss': 0.7183011770248413, 'train/grad_norm_pre_clip': 8.843823432922363, 'timing/data': 0.012001103721559048, 'timing/model': 1.044087323360145, 'learning_rate/qwen_vl_interface': 9.931665695513786e-06, 'learning_rate/action_model': 9.931665695513786e-05, 'epoch': 0.19}) +[2026-06-17 17:31:17,350][starVLA.training.train_starvla][INFO] - Step 474, Loss: {'train/loss': 0.6983698606491089, 'train/grad_norm_pre_clip': 5.78865909576416, 'timing/data': 0.011268505826592445, 'timing/model': 0.9453314626589417, 'learning_rate/qwen_vl_interface': 9.931299687554466e-06, 'learning_rate/action_model': 9.931299687554466e-05, 'epoch': 0.2}) +[2026-06-17 17:31:21,435][starVLA.training.train_starvla][INFO] - Step 475, Loss: {'train/loss': 0.7255786061286926, 'train/grad_norm_pre_clip': 5.972423076629639, 'timing/data': 0.00741505715996027, 'timing/model': 1.1108915517106652, 'learning_rate/qwen_vl_interface': 9.930932709156686e-06, 'learning_rate/action_model': 9.930932709156685e-05, 'epoch': 0.2}) +[2026-06-17 17:31:25,441][starVLA.training.train_starvla][INFO] - Step 476, Loss: {'train/loss': 1.1102733612060547, 'train/grad_norm_pre_clip': 7.2779765129089355, 'timing/data': 0.008819038979709148, 'timing/model': 0.9805652722716331, 'learning_rate/qwen_vl_interface': 9.930564760396518e-06, 'learning_rate/action_model': 9.930564760396518e-05, 'epoch': 0.2}) +[2026-06-17 17:31:29,499][starVLA.training.train_starvla][INFO] - Step 477, Loss: {'train/loss': 1.045577883720398, 'train/grad_norm_pre_clip': 8.427919387817383, 'timing/data': 0.013319311663508415, 'timing/model': 1.0416896799579263, 'learning_rate/qwen_vl_interface': 9.930195841350237e-06, 'learning_rate/action_model': 9.930195841350236e-05, 'epoch': 0.2}) +[2026-06-17 17:31:33,395][starVLA.training.train_starvla][INFO] - Step 478, Loss: {'train/loss': 0.9823065996170044, 'train/grad_norm_pre_clip': 7.139065265655518, 'timing/data': 0.01655543316155672, 'timing/model': 1.1703686844557524, 'learning_rate/qwen_vl_interface': 9.929825952094323e-06, 'learning_rate/action_model': 9.929825952094323e-05, 'epoch': 0.2}) +[2026-06-17 17:31:37,073][starVLA.training.train_starvla][INFO] - Step 479, Loss: {'train/loss': 0.6777743101119995, 'train/grad_norm_pre_clip': 5.496505260467529, 'timing/data': 0.010685622692108154, 'timing/model': 0.9674307061359286, 'learning_rate/qwen_vl_interface': 9.929455092705453e-06, 'learning_rate/action_model': 9.929455092705453e-05, 'epoch': 0.2}) +[2026-06-17 17:31:41,305][starVLA.training.train_starvla][INFO] - Step 480, Loss: {'train/loss': 0.9624590277671814, 'train/grad_norm_pre_clip': 23.803794860839844, 'timing/data': 0.01241741981357336, 'timing/model': 1.034892168827355, 'learning_rate/qwen_vl_interface': 9.929083263260508e-06, 'learning_rate/action_model': 9.929083263260507e-05, 'epoch': 0.2}) +[2026-06-17 17:31:50,110][starVLA.training.train_starvla][INFO] - Step 481, Loss: {'train/loss': 1.0650750398635864, 'train/grad_norm_pre_clip': 7.717925071716309, 'timing/data': 0.007358189672231674, 'timing/model': 2.465527015738189, 'learning_rate/qwen_vl_interface': 9.928710463836566e-06, 'learning_rate/action_model': 9.928710463836565e-05, 'epoch': 0.2}) +[2026-06-17 17:31:54,239][starVLA.training.train_starvla][INFO] - Step 482, Loss: {'train/loss': 0.754955530166626, 'train/grad_norm_pre_clip': 6.145718097686768, 'timing/data': 0.010029049590229988, 'timing/model': 0.9458397245034575, 'learning_rate/qwen_vl_interface': 9.928336694510912e-06, 'learning_rate/action_model': 9.928336694510911e-05, 'epoch': 0.2}) +[2026-06-17 17:31:58,059][starVLA.training.train_starvla][INFO] - Step 483, Loss: {'train/loss': 0.7893609404563904, 'train/grad_norm_pre_clip': 5.032792091369629, 'timing/data': 0.019930253736674786, 'timing/model': 0.925812073983252, 'learning_rate/qwen_vl_interface': 9.927961955361025e-06, 'learning_rate/action_model': 9.927961955361025e-05, 'epoch': 0.2}) +[2026-06-17 17:32:01,848][starVLA.training.train_starvla][INFO] - Step 484, Loss: {'train/loss': 0.5578418970108032, 'train/grad_norm_pre_clip': 5.312914848327637, 'timing/data': 0.009925073944032192, 'timing/model': 1.0897398451343179, 'learning_rate/qwen_vl_interface': 9.927586246464593e-06, 'learning_rate/action_model': 9.927586246464593e-05, 'epoch': 0.2}) +[2026-06-17 17:32:05,814][starVLA.training.train_starvla][INFO] - Step 485, Loss: {'train/loss': 0.643745481967926, 'train/grad_norm_pre_clip': 5.416767597198486, 'timing/data': 0.011452628299593925, 'timing/model': 1.257998458109796, 'learning_rate/qwen_vl_interface': 9.927209567899499e-06, 'learning_rate/action_model': 9.927209567899498e-05, 'epoch': 0.2}) +[2026-06-17 17:32:09,538][starVLA.training.train_starvla][INFO] - Step 486, Loss: {'train/loss': 0.555840015411377, 'train/grad_norm_pre_clip': 3.9438724517822266, 'timing/data': 0.00961905624717474, 'timing/model': 0.967637893743813, 'learning_rate/qwen_vl_interface': 9.926831919743827e-06, 'learning_rate/action_model': 9.926831919743827e-05, 'epoch': 0.2}) +[2026-06-17 17:32:13,415][starVLA.training.train_starvla][INFO] - Step 487, Loss: {'train/loss': 1.014259934425354, 'train/grad_norm_pre_clip': 8.138873100280762, 'timing/data': 0.009576241485774517, 'timing/model': 0.9511662991717458, 'learning_rate/qwen_vl_interface': 9.926453302075866e-06, 'learning_rate/action_model': 9.926453302075867e-05, 'epoch': 0.2}) +[2026-06-17 17:32:17,318][starVLA.training.train_starvla][INFO] - Step 488, Loss: {'train/loss': 1.0146201848983765, 'train/grad_norm_pre_clip': 7.9289679527282715, 'timing/data': 0.010038604959845543, 'timing/model': 0.9638161212205887, 'learning_rate/qwen_vl_interface': 9.926073714974103e-06, 'learning_rate/action_model': 9.926073714974104e-05, 'epoch': 0.2}) +[2026-06-17 17:32:21,142][starVLA.training.train_starvla][INFO] - Step 489, Loss: {'train/loss': 0.9151197075843811, 'train/grad_norm_pre_clip': 6.7059102058410645, 'timing/data': 0.010948912240564823, 'timing/model': 0.9448978826403618, 'learning_rate/qwen_vl_interface': 9.925693158517228e-06, 'learning_rate/action_model': 9.925693158517228e-05, 'epoch': 0.2}) +[2026-06-17 17:32:25,284][starVLA.training.train_starvla][INFO] - Step 490, Loss: {'train/loss': 1.15742027759552, 'train/grad_norm_pre_clip': 10.335506439208984, 'timing/data': 0.011710011400282383, 'timing/model': 1.0031068045645952, 'learning_rate/qwen_vl_interface': 9.92531163278413e-06, 'learning_rate/action_model': 9.92531163278413e-05, 'epoch': 0.2}) +[2026-06-17 17:32:29,661][starVLA.training.train_starvla][INFO] - Step 491, Loss: {'train/loss': 1.0285460948944092, 'train/grad_norm_pre_clip': 7.016514301300049, 'timing/data': 0.018424704670906067, 'timing/model': 1.48246034886688, 'learning_rate/qwen_vl_interface': 9.924929137853899e-06, 'learning_rate/action_model': 9.924929137853898e-05, 'epoch': 0.2}) +[2026-06-17 17:32:35,265][starVLA.training.train_starvla][INFO] - Step 492, Loss: {'train/loss': 0.757539689540863, 'train/grad_norm_pre_clip': 6.022542476654053, 'timing/data': 0.01154261827468872, 'timing/model': 2.6055095652118325, 'learning_rate/qwen_vl_interface': 9.92454567380583e-06, 'learning_rate/action_model': 9.924545673805829e-05, 'epoch': 0.2}) +[2026-06-17 17:32:39,467][starVLA.training.train_starvla][INFO] - Step 493, Loss: {'train/loss': 0.8501863479614258, 'train/grad_norm_pre_clip': 6.112825870513916, 'timing/data': 0.008228285238146782, 'timing/model': 0.9325325582176447, 'learning_rate/qwen_vl_interface': 9.924161240719409e-06, 'learning_rate/action_model': 9.924161240719408e-05, 'epoch': 0.2}) +[2026-06-17 17:32:43,314][starVLA.training.train_starvla][INFO] - Step 494, Loss: {'train/loss': 0.8265863060951233, 'train/grad_norm_pre_clip': 7.927798271179199, 'timing/data': 0.00923432782292366, 'timing/model': 0.9719282146543264, 'learning_rate/qwen_vl_interface': 9.923775838674336e-06, 'learning_rate/action_model': 9.923775838674335e-05, 'epoch': 0.2}) +[2026-06-17 17:32:47,207][starVLA.training.train_starvla][INFO] - Step 495, Loss: {'train/loss': 0.4108070731163025, 'train/grad_norm_pre_clip': 5.836637020111084, 'timing/data': 0.008860061876475811, 'timing/model': 0.9584225863218307, 'learning_rate/qwen_vl_interface': 9.9233894677505e-06, 'learning_rate/action_model': 9.9233894677505e-05, 'epoch': 0.2}) +[2026-06-17 17:32:51,417][starVLA.training.train_starvla][INFO] - Step 496, Loss: {'train/loss': 1.026962161064148, 'train/grad_norm_pre_clip': 6.5063018798828125, 'timing/data': 0.011615341529250145, 'timing/model': 1.204985247924924, 'learning_rate/qwen_vl_interface': 9.923002128028e-06, 'learning_rate/action_model': 9.923002128028e-05, 'epoch': 0.2}) +[2026-06-17 17:32:56,700][starVLA.training.train_starvla][INFO] - Step 497, Loss: {'train/loss': 0.7823006510734558, 'train/grad_norm_pre_clip': 6.911802768707275, 'timing/data': 0.00992779154330492, 'timing/model': 0.966851613484323, 'learning_rate/qwen_vl_interface': 9.922613819587131e-06, 'learning_rate/action_model': 9.92261381958713e-05, 'epoch': 0.2}) +[2026-06-17 17:33:00,643][starVLA.training.train_starvla][INFO] - Step 498, Loss: {'train/loss': 0.8251343369483948, 'train/grad_norm_pre_clip': 12.700631141662598, 'timing/data': 0.011071541346609592, 'timing/model': 1.135338575579226, 'learning_rate/qwen_vl_interface': 9.922224542508389e-06, 'learning_rate/action_model': 9.922224542508388e-05, 'epoch': 0.2}) +[2026-06-17 17:33:04,810][starVLA.training.train_starvla][INFO] - Step 499, Loss: {'train/loss': 0.9532020092010498, 'train/grad_norm_pre_clip': 7.143864154815674, 'timing/data': 0.010989480651915073, 'timing/model': 1.252679475583136, 'learning_rate/qwen_vl_interface': 9.92183429687247e-06, 'learning_rate/action_model': 9.92183429687247e-05, 'epoch': 0.21}) +[2026-06-17 17:35:08,667][starVLA.training.train_starvla][INFO] - Step 500, Loss: {'train/loss': 1.0747655630111694, 'train/grad_norm_pre_clip': 8.752114295959473, 'eval/loss': 0.8695785403251648, 'eval/latency_6/loss': 0.8695785403251648, 'eval/demon_fire_precision': 0.9056434756643774, 'eval/demon_fire_recall': 0.980284421460892, 'eval/demon_fire_f1': 0.9414868849914636, 'eval/demon_fire_rate_model': 0.9318308291597106, 'eval/demon_fire_rate_teacher': 0.860879243183083, 'eval/demon_left_precision': 0.5333333333333333, 'eval/demon_left_recall': 0.3929824561403509, 'eval/demon_left_f1': 0.45252525252525255, 'eval/demon_left_rate_model': 0.11686143572621036, 'eval/demon_left_rate_teacher': 0.15859766277128548, 'eval/demon_right_precision': 0.8163265306122449, 'eval/demon_right_recall': 0.06144393241167435, 'eval/demon_right_f1': 0.11428571428571428, 'eval/demon_right_rate_model': 0.013633834168057874, 'eval/demon_right_rate_teacher': 0.18113522537562604, 'eval/demon_move_f1': 0.28340548340548344, 'eval/demon_cc_f1': 0.5549208641291373, 'eval/latency_6/demon_fire_precision': 0.9056434756643774, 'eval/latency_6/demon_fire_recall': 0.980284421460892, 'eval/latency_6/demon_fire_f1': 0.9414868849914636, 'eval/latency_6/demon_fire_rate_model': 0.9318308291597106, 'eval/latency_6/demon_fire_rate_teacher': 0.860879243183083, 'eval/latency_6/demon_left_precision': 0.5333333333333333, 'eval/latency_6/demon_left_recall': 0.3929824561403509, 'eval/latency_6/demon_left_f1': 0.45252525252525255, 'eval/latency_6/demon_left_rate_model': 0.11686143572621036, 'eval/latency_6/demon_left_rate_teacher': 0.15859766277128548, 'eval/latency_6/demon_right_precision': 0.8163265306122449, 'eval/latency_6/demon_right_recall': 0.06144393241167435, 'eval/latency_6/demon_right_f1': 0.11428571428571428, 'eval/latency_6/demon_right_rate_model': 0.013633834168057874, 'eval/latency_6/demon_right_rate_teacher': 0.18113522537562604, 'eval/latency_6/demon_move_f1': 0.28340548340548344, 'eval/latency_6/demon_cc_f1': 0.5549208641291373, 'timing/data': 0.006968220695853233, 'timing/model': 1.0045640198513865, 'learning_rate/qwen_vl_interface': 9.921443082760276e-06, 'learning_rate/action_model': 9.921443082760276e-05, 'epoch': 0.21}) +[2026-06-17 17:35:08,670][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_500_state +[2026-06-17 17:35:08,694][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 17:35:16,319][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_500_state/model.safetensors +[2026-06-17 17:35:31,865][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_500_state/optimizer.bin +[2026-06-17 17:35:31,866][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_500_state/sampler.bin +[2026-06-17 17:35:31,866][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_500_state/sampler_1.bin +[2026-06-17 17:35:31,879][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_500_state/random_states_0.pkl +[2026-06-17 17:35:31,884][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 500 +[2026-06-17 17:35:31,885][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 17:35:31,900][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 17:35:35,885][starVLA.training.train_starvla][INFO] - Step 501, Loss: {'train/loss': 0.7638300061225891, 'train/grad_norm_pre_clip': 5.193850517272949, 'timing/data': 0.010887518525123596, 'timing/model': 1.1934410212561488, 'learning_rate/qwen_vl_interface': 9.921050900252904e-06, 'learning_rate/action_model': 9.921050900252903e-05, 'epoch': 0.21}) +[2026-06-17 17:35:39,496][starVLA.training.train_starvla][INFO] - Step 502, Loss: {'train/loss': 0.4413439631462097, 'train/grad_norm_pre_clip': 5.152349948883057, 'timing/data': 0.00841242540627718, 'timing/model': 0.9206219930201769, 'learning_rate/qwen_vl_interface': 9.920657749431654e-06, 'learning_rate/action_model': 9.920657749431654e-05, 'epoch': 0.21}) +[2026-06-17 17:35:43,282][starVLA.training.train_starvla][INFO] - Step 503, Loss: {'train/loss': 0.8330145478248596, 'train/grad_norm_pre_clip': 31.1603946685791, 'timing/data': 0.013648300431668758, 'timing/model': 0.9362933952361345, 'learning_rate/qwen_vl_interface': 9.920263630378026e-06, 'learning_rate/action_model': 9.920263630378024e-05, 'epoch': 0.21}) +[2026-06-17 17:35:47,198][starVLA.training.train_starvla][INFO] - Step 504, Loss: {'train/loss': 0.8569438457489014, 'train/grad_norm_pre_clip': 12.419939041137695, 'timing/data': 0.01225246675312519, 'timing/model': 0.980122628621757, 'learning_rate/qwen_vl_interface': 9.919868543173722e-06, 'learning_rate/action_model': 9.919868543173721e-05, 'epoch': 0.21}) +[2026-06-17 17:35:51,217][starVLA.training.train_starvla][INFO] - Step 505, Loss: {'train/loss': 0.9708070755004883, 'train/grad_norm_pre_clip': 5.70446252822876, 'timing/data': 0.011476008221507072, 'timing/model': 0.9561024662107229, 'learning_rate/qwen_vl_interface': 9.919472487900645e-06, 'learning_rate/action_model': 9.919472487900644e-05, 'epoch': 0.21}) +[2026-06-17 17:35:56,693][starVLA.training.train_starvla][INFO] - Step 506, Loss: {'train/loss': 0.7416186332702637, 'train/grad_norm_pre_clip': 11.854762077331543, 'timing/data': 0.008081002160906792, 'timing/model': 2.096416817046702, 'learning_rate/qwen_vl_interface': 9.919075464640895e-06, 'learning_rate/action_model': 9.919075464640895e-05, 'epoch': 0.21}) +[2026-06-17 17:36:00,620][starVLA.training.train_starvla][INFO] - Step 507, Loss: {'train/loss': 1.166374921798706, 'train/grad_norm_pre_clip': 6.455967903137207, 'timing/data': 0.01259877160191536, 'timing/model': 1.1299261981621385, 'learning_rate/qwen_vl_interface': 9.91867747347678e-06, 'learning_rate/action_model': 9.918677473476779e-05, 'epoch': 0.21}) +[2026-06-17 17:36:04,560][starVLA.training.train_starvla][INFO] - Step 508, Loss: {'train/loss': 0.6368259191513062, 'train/grad_norm_pre_clip': 6.402385711669922, 'timing/data': 0.009555554017424583, 'timing/model': 1.1752419667318463, 'learning_rate/qwen_vl_interface': 9.918278514490798e-06, 'learning_rate/action_model': 9.918278514490799e-05, 'epoch': 0.21}) +[2026-06-17 17:36:08,321][starVLA.training.train_starvla][INFO] - Step 509, Loss: {'train/loss': 1.1880829334259033, 'train/grad_norm_pre_clip': 7.881405830383301, 'timing/data': 0.00904911570250988, 'timing/model': 0.9838996957987547, 'learning_rate/qwen_vl_interface': 9.917878587765658e-06, 'learning_rate/action_model': 9.917878587765658e-05, 'epoch': 0.21}) +[2026-06-17 17:36:12,250][starVLA.training.train_starvla][INFO] - Step 510, Loss: {'train/loss': 0.9436204433441162, 'train/grad_norm_pre_clip': 8.228885650634766, 'timing/data': 0.01034472044557333, 'timing/model': 0.9722441313788295, 'learning_rate/qwen_vl_interface': 9.917477693384264e-06, 'learning_rate/action_model': 9.917477693384264e-05, 'epoch': 0.21}) +[2026-06-17 17:36:16,147][starVLA.training.train_starvla][INFO] - Step 511, Loss: {'train/loss': 0.8074350953102112, 'train/grad_norm_pre_clip': 6.801384449005127, 'timing/data': 0.009097268804907799, 'timing/model': 0.958045044913888, 'learning_rate/qwen_vl_interface': 9.917075831429721e-06, 'learning_rate/action_model': 9.917075831429722e-05, 'epoch': 0.21}) +[2026-06-17 17:36:20,004][starVLA.training.train_starvla][INFO] - Step 512, Loss: {'train/loss': 1.0199499130249023, 'train/grad_norm_pre_clip': 8.97625732421875, 'timing/data': 0.008682984858751297, 'timing/model': 0.9407607894390821, 'learning_rate/qwen_vl_interface': 9.916673001985339e-06, 'learning_rate/action_model': 9.916673001985338e-05, 'epoch': 0.21}) +[2026-06-17 17:36:24,060][starVLA.training.train_starvla][INFO] - Step 513, Loss: {'train/loss': 0.8008959889411926, 'train/grad_norm_pre_clip': 6.920738697052002, 'timing/data': 0.01173077430576086, 'timing/model': 0.9409446604549885, 'learning_rate/qwen_vl_interface': 9.91626920513462e-06, 'learning_rate/action_model': 9.916269205134619e-05, 'epoch': 0.21}) +[2026-06-17 17:36:29,037][starVLA.training.train_starvla][INFO] - Step 514, Loss: {'train/loss': 0.8309561014175415, 'train/grad_norm_pre_clip': 5.544281482696533, 'timing/data': 0.010521424002945423, 'timing/model': 1.170154712162912, 'learning_rate/qwen_vl_interface': 9.915864440961272e-06, 'learning_rate/action_model': 9.915864440961272e-05, 'epoch': 0.21}) +[2026-06-17 17:36:32,896][starVLA.training.train_starvla][INFO] - Step 515, Loss: {'train/loss': 1.0228891372680664, 'train/grad_norm_pre_clip': 8.716962814331055, 'timing/data': 0.010858722031116486, 'timing/model': 0.9791336162015796, 'learning_rate/qwen_vl_interface': 9.915458709549205e-06, 'learning_rate/action_model': 9.915458709549206e-05, 'epoch': 0.21}) +[2026-06-17 17:36:39,212][starVLA.training.train_starvla][INFO] - Step 516, Loss: {'train/loss': 1.1214134693145752, 'train/grad_norm_pre_clip': 8.76842975616455, 'timing/data': 0.010144490748643875, 'timing/model': 1.762981672771275, 'learning_rate/qwen_vl_interface': 9.91505201098253e-06, 'learning_rate/action_model': 9.91505201098253e-05, 'epoch': 0.21}) +[2026-06-17 17:36:43,482][starVLA.training.train_starvla][INFO] - Step 517, Loss: {'train/loss': 0.6672785878181458, 'train/grad_norm_pre_clip': 5.47489070892334, 'timing/data': 0.01649887766689062, 'timing/model': 1.0303045446053147, 'learning_rate/qwen_vl_interface': 9.914644345345551e-06, 'learning_rate/action_model': 9.914644345345552e-05, 'epoch': 0.21}) +[2026-06-17 17:36:47,369][starVLA.training.train_starvla][INFO] - Step 518, Loss: {'train/loss': 1.0118640661239624, 'train/grad_norm_pre_clip': 7.115229606628418, 'timing/data': 0.012506721541285515, 'timing/model': 0.9457824435085058, 'learning_rate/qwen_vl_interface': 9.914235712722782e-06, 'learning_rate/action_model': 9.914235712722781e-05, 'epoch': 0.21}) +[2026-06-17 17:36:51,411][starVLA.training.train_starvla][INFO] - Step 519, Loss: {'train/loss': 0.8322829008102417, 'train/grad_norm_pre_clip': 7.232666969299316, 'timing/data': 0.014187001623213291, 'timing/model': 0.9414525646716356, 'learning_rate/qwen_vl_interface': 9.913826113198927e-06, 'learning_rate/action_model': 9.913826113198927e-05, 'epoch': 0.21}) +[2026-06-17 17:36:55,393][starVLA.training.train_starvla][INFO] - Step 520, Loss: {'train/loss': 1.089276671409607, 'train/grad_norm_pre_clip': 9.131351470947266, 'timing/data': 0.011039781384170055, 'timing/model': 1.1976730320602655, 'learning_rate/qwen_vl_interface': 9.913415546858904e-06, 'learning_rate/action_model': 9.913415546858904e-05, 'epoch': 0.21}) +[2026-06-17 17:36:59,396][starVLA.training.train_starvla][INFO] - Step 521, Loss: {'train/loss': 0.8128294348716736, 'train/grad_norm_pre_clip': 6.853999137878418, 'timing/data': 0.01922213565558195, 'timing/model': 1.124314310029149, 'learning_rate/qwen_vl_interface': 9.913004013787818e-06, 'learning_rate/action_model': 9.913004013787819e-05, 'epoch': 0.21}) +[2026-06-17 17:37:03,576][starVLA.training.train_starvla][INFO] - Step 522, Loss: {'train/loss': 0.8072890043258667, 'train/grad_norm_pre_clip': 7.886904716491699, 'timing/data': 0.012303082272410393, 'timing/model': 1.1152290580794215, 'learning_rate/qwen_vl_interface': 9.912591514070983e-06, 'learning_rate/action_model': 9.912591514070983e-05, 'epoch': 0.21}) +[2026-06-17 17:37:07,843][starVLA.training.train_starvla][INFO] - Step 523, Loss: {'train/loss': 0.7504894137382507, 'train/grad_norm_pre_clip': 9.313494682312012, 'timing/data': 0.009615136310458183, 'timing/model': 0.9601026950404048, 'learning_rate/qwen_vl_interface': 9.912178047793912e-06, 'learning_rate/action_model': 9.91217804779391e-05, 'epoch': 0.22}) +[2026-06-17 17:37:11,971][starVLA.training.train_starvla][INFO] - Step 524, Loss: {'train/loss': 0.7172946333885193, 'train/grad_norm_pre_clip': 5.861557960510254, 'timing/data': 0.010774902068078518, 'timing/model': 1.0957500590011477, 'learning_rate/qwen_vl_interface': 9.911763615042311e-06, 'learning_rate/action_model': 9.91176361504231e-05, 'epoch': 0.22}) +[2026-06-17 17:37:15,922][starVLA.training.train_starvla][INFO] - Step 525, Loss: {'train/loss': 0.7059494853019714, 'train/grad_norm_pre_clip': 5.783107757568359, 'timing/data': 0.009314991533756256, 'timing/model': 0.9254204798489809, 'learning_rate/qwen_vl_interface': 9.911348215902099e-06, 'learning_rate/action_model': 9.911348215902099e-05, 'epoch': 0.22}) +[2026-06-17 17:37:20,008][starVLA.training.train_starvla][INFO] - Step 526, Loss: {'train/loss': 0.9581197500228882, 'train/grad_norm_pre_clip': 5.192107677459717, 'timing/data': 0.00988046545535326, 'timing/model': 0.9905147533863783, 'learning_rate/qwen_vl_interface': 9.910931850459385e-06, 'learning_rate/action_model': 9.910931850459385e-05, 'epoch': 0.22}) +[2026-06-17 17:37:23,823][starVLA.training.train_starvla][INFO] - Step 527, Loss: {'train/loss': 0.8880410194396973, 'train/grad_norm_pre_clip': 6.709724426269531, 'timing/data': 0.010441011749207973, 'timing/model': 1.120927401818335, 'learning_rate/qwen_vl_interface': 9.910514518800485e-06, 'learning_rate/action_model': 9.910514518800484e-05, 'epoch': 0.22}) +[2026-06-17 17:37:27,752][starVLA.training.train_starvla][INFO] - Step 528, Loss: {'train/loss': 0.8829753398895264, 'train/grad_norm_pre_clip': 31.08085060119629, 'timing/data': 0.009977038949728012, 'timing/model': 1.1829951889812946, 'learning_rate/qwen_vl_interface': 9.910096221011908e-06, 'learning_rate/action_model': 9.910096221011907e-05, 'epoch': 0.22}) +[2026-06-17 17:37:31,886][starVLA.training.train_starvla][INFO] - Step 529, Loss: {'train/loss': 0.7694953680038452, 'train/grad_norm_pre_clip': 5.213215351104736, 'timing/data': 0.014680287800729275, 'timing/model': 1.021115661598742, 'learning_rate/qwen_vl_interface': 9.90967695718037e-06, 'learning_rate/action_model': 9.90967695718037e-05, 'epoch': 0.22}) +[2026-06-17 17:37:35,911][starVLA.training.train_starvla][INFO] - Step 530, Loss: {'train/loss': 1.076841950416565, 'train/grad_norm_pre_clip': 8.342802047729492, 'timing/data': 0.009062951430678368, 'timing/model': 1.0129313068464398, 'learning_rate/qwen_vl_interface': 9.909256727392786e-06, 'learning_rate/action_model': 9.909256727392784e-05, 'epoch': 0.22}) +[2026-06-17 17:37:39,928][starVLA.training.train_starvla][INFO] - Step 531, Loss: {'train/loss': 0.8055122494697571, 'train/grad_norm_pre_clip': 9.26270580291748, 'timing/data': 0.015893826261162758, 'timing/model': 1.1019245125353336, 'learning_rate/qwen_vl_interface': 9.908835531736269e-06, 'learning_rate/action_model': 9.908835531736267e-05, 'epoch': 0.22}) +[2026-06-17 17:37:43,817][starVLA.training.train_starvla][INFO] - Step 532, Loss: {'train/loss': 0.7256643772125244, 'train/grad_norm_pre_clip': 6.645920276641846, 'timing/data': 0.008093666285276413, 'timing/model': 0.9628772083669901, 'learning_rate/qwen_vl_interface': 9.908413370298131e-06, 'learning_rate/action_model': 9.90841337029813e-05, 'epoch': 0.22}) +[2026-06-17 17:37:47,959][starVLA.training.train_starvla][INFO] - Step 533, Loss: {'train/loss': 0.8895166516304016, 'train/grad_norm_pre_clip': 10.825919151306152, 'timing/data': 0.01334923692047596, 'timing/model': 0.9671677509322762, 'learning_rate/qwen_vl_interface': 9.90799024316589e-06, 'learning_rate/action_model': 9.90799024316589e-05, 'epoch': 0.22}) +[2026-06-17 17:37:51,842][starVLA.training.train_starvla][INFO] - Step 534, Loss: {'train/loss': 0.9623377919197083, 'train/grad_norm_pre_clip': 8.21665096282959, 'timing/data': 0.009564826264977455, 'timing/model': 1.1308789439499378, 'learning_rate/qwen_vl_interface': 9.90756615042726e-06, 'learning_rate/action_model': 9.907566150427259e-05, 'epoch': 0.22}) +[2026-06-17 17:37:55,767][starVLA.training.train_starvla][INFO] - Step 535, Loss: {'train/loss': 0.9226388931274414, 'train/grad_norm_pre_clip': 6.514520645141602, 'timing/data': 0.010537766851484776, 'timing/model': 1.0735514173284173, 'learning_rate/qwen_vl_interface': 9.907141092170154e-06, 'learning_rate/action_model': 9.907141092170153e-05, 'epoch': 0.22}) +[2026-06-17 17:38:01,492][starVLA.training.train_starvla][INFO] - Step 536, Loss: {'train/loss': 0.6255658864974976, 'train/grad_norm_pre_clip': 7.379212856292725, 'timing/data': 0.010278253816068172, 'timing/model': 2.208016444928944, 'learning_rate/qwen_vl_interface': 9.906715068482689e-06, 'learning_rate/action_model': 9.90671506848269e-05, 'epoch': 0.22}) +[2026-06-17 17:38:06,959][starVLA.training.train_starvla][INFO] - Step 537, Loss: {'train/loss': 0.8025428652763367, 'train/grad_norm_pre_clip': 9.099443435668945, 'timing/data': 0.009571936912834644, 'timing/model': 1.0044623473659158, 'learning_rate/qwen_vl_interface': 9.90628807945318e-06, 'learning_rate/action_model': 9.90628807945318e-05, 'epoch': 0.22}) +[2026-06-17 17:38:10,892][starVLA.training.train_starvla][INFO] - Step 538, Loss: {'train/loss': 1.086547613143921, 'train/grad_norm_pre_clip': 6.714651107788086, 'timing/data': 0.007336829788982868, 'timing/model': 0.9609313178807497, 'learning_rate/qwen_vl_interface': 9.90586012517014e-06, 'learning_rate/action_model': 9.90586012517014e-05, 'epoch': 0.22}) +[2026-06-17 17:38:14,834][starVLA.training.train_starvla][INFO] - Step 539, Loss: {'train/loss': 0.6063501238822937, 'train/grad_norm_pre_clip': 8.29489517211914, 'timing/data': 0.009822716936469078, 'timing/model': 0.9677493562921882, 'learning_rate/qwen_vl_interface': 9.905431205722288e-06, 'learning_rate/action_model': 9.905431205722288e-05, 'epoch': 0.22}) +[2026-06-17 17:38:18,691][starVLA.training.train_starvla][INFO] - Step 540, Loss: {'train/loss': 0.6144570112228394, 'train/grad_norm_pre_clip': 4.904439449310303, 'timing/data': 0.010647392831742764, 'timing/model': 1.1335641732439399, 'learning_rate/qwen_vl_interface': 9.905001321198536e-06, 'learning_rate/action_model': 9.905001321198536e-05, 'epoch': 0.22}) +[2026-06-17 17:38:22,643][starVLA.training.train_starvla][INFO] - Step 541, Loss: {'train/loss': 0.6359403133392334, 'train/grad_norm_pre_clip': 5.924860000610352, 'timing/data': 0.013837960548698902, 'timing/model': 1.0637640012428164, 'learning_rate/qwen_vl_interface': 9.904570471688003e-06, 'learning_rate/action_model': 9.904570471688002e-05, 'epoch': 0.22}) +[2026-06-17 17:38:26,571][starVLA.training.train_starvla][INFO] - Step 542, Loss: {'train/loss': 0.6818676590919495, 'train/grad_norm_pre_clip': 9.671289443969727, 'timing/data': 0.009295097552239895, 'timing/model': 1.023867342621088, 'learning_rate/qwen_vl_interface': 9.90413865728e-06, 'learning_rate/action_model': 9.90413865728e-05, 'epoch': 0.22}) +[2026-06-17 17:38:30,441][starVLA.training.train_starvla][INFO] - Step 543, Loss: {'train/loss': 0.5961329340934753, 'train/grad_norm_pre_clip': 4.87800407409668, 'timing/data': 0.009991844184696674, 'timing/model': 0.9668998708948493, 'learning_rate/qwen_vl_interface': 9.903705878064047e-06, 'learning_rate/action_model': 9.903705878064046e-05, 'epoch': 0.22}) +[2026-06-17 17:38:34,285][starVLA.training.train_starvla][INFO] - Step 544, Loss: {'train/loss': 0.9619706273078918, 'train/grad_norm_pre_clip': 7.5246052742004395, 'timing/data': 0.010112646967172623, 'timing/model': 0.94220763258636, 'learning_rate/qwen_vl_interface': 9.903272134129858e-06, 'learning_rate/action_model': 9.903272134129858e-05, 'epoch': 0.22}) +[2026-06-17 17:38:38,076][starVLA.training.train_starvla][INFO] - Step 545, Loss: {'train/loss': 1.0069316625595093, 'train/grad_norm_pre_clip': 9.125818252563477, 'timing/data': 0.010422483086585999, 'timing/model': 0.9427481023594737, 'learning_rate/qwen_vl_interface': 9.90283742556735e-06, 'learning_rate/action_model': 9.902837425567349e-05, 'epoch': 0.22}) +[2026-06-17 17:38:41,873][starVLA.training.train_starvla][INFO] - Step 546, Loss: {'train/loss': 0.7051733136177063, 'train/grad_norm_pre_clip': 7.249682426452637, 'timing/data': 0.009948505088686943, 'timing/model': 0.9376047551631927, 'learning_rate/qwen_vl_interface': 9.902401752466635e-06, 'learning_rate/action_model': 9.902401752466634e-05, 'epoch': 0.22}) +[2026-06-17 17:38:45,692][starVLA.training.train_starvla][INFO] - Step 547, Loss: {'train/loss': 0.9531843662261963, 'train/grad_norm_pre_clip': 6.306337833404541, 'timing/data': 0.008830143138766289, 'timing/model': 1.1289502112194896, 'learning_rate/qwen_vl_interface': 9.90196511491803e-06, 'learning_rate/action_model': 9.901965114918029e-05, 'epoch': 0.23}) +[2026-06-17 17:38:49,299][starVLA.training.train_starvla][INFO] - Step 548, Loss: {'train/loss': 0.68720942735672, 'train/grad_norm_pre_clip': 6.6576151847839355, 'timing/data': 0.01471694279462099, 'timing/model': 0.9291568184271455, 'learning_rate/qwen_vl_interface': 9.901527513012053e-06, 'learning_rate/action_model': 9.901527513012052e-05, 'epoch': 0.23}) +[2026-06-17 17:38:53,146][starVLA.training.train_starvla][INFO] - Step 549, Loss: {'train/loss': 0.756256103515625, 'train/grad_norm_pre_clip': 9.227982521057129, 'timing/data': 0.009208585135638714, 'timing/model': 0.9312770795077085, 'learning_rate/qwen_vl_interface': 9.901088946839415e-06, 'learning_rate/action_model': 9.901088946839414e-05, 'epoch': 0.23}) +[2026-06-17 17:38:56,959][starVLA.training.train_starvla][INFO] - Step 550, Loss: {'train/loss': 0.69655442237854, 'train/grad_norm_pre_clip': 16.159317016601562, 'timing/data': 0.008075249381363392, 'timing/model': 0.9403592990711331, 'learning_rate/qwen_vl_interface': 9.900649416491035e-06, 'learning_rate/action_model': 9.900649416491035e-05, 'epoch': 0.23}) +[2026-06-17 17:39:01,130][starVLA.training.train_starvla][INFO] - Step 551, Loss: {'train/loss': 0.5895818471908569, 'train/grad_norm_pre_clip': 6.552199840545654, 'timing/data': 0.013662009499967098, 'timing/model': 1.076333730481565, 'learning_rate/qwen_vl_interface': 9.900208922058026e-06, 'learning_rate/action_model': 9.900208922058025e-05, 'epoch': 0.23}) +[2026-06-17 17:39:05,245][starVLA.training.train_starvla][INFO] - Step 552, Loss: {'train/loss': 0.7794767022132874, 'train/grad_norm_pre_clip': 6.30858850479126, 'timing/data': 0.01042515505105257, 'timing/model': 0.9699421254917979, 'learning_rate/qwen_vl_interface': 9.899767463631704e-06, 'learning_rate/action_model': 9.899767463631704e-05, 'epoch': 0.23}) +[2026-06-17 17:39:09,197][starVLA.training.train_starvla][INFO] - Step 553, Loss: {'train/loss': 0.9258288145065308, 'train/grad_norm_pre_clip': 7.890223979949951, 'timing/data': 0.01059531606733799, 'timing/model': 1.1871572490781546, 'learning_rate/qwen_vl_interface': 9.899325041303582e-06, 'learning_rate/action_model': 9.899325041303583e-05, 'epoch': 0.23}) +[2026-06-17 17:39:12,870][starVLA.training.train_starvla][INFO] - Step 554, Loss: {'train/loss': 0.9893174171447754, 'train/grad_norm_pre_clip': 8.567317008972168, 'timing/data': 0.010581131093204021, 'timing/model': 0.9444949254393578, 'learning_rate/qwen_vl_interface': 9.89888165516538e-06, 'learning_rate/action_model': 9.898881655165379e-05, 'epoch': 0.23}) +[2026-06-17 17:39:16,829][starVLA.training.train_starvla][INFO] - Step 555, Loss: {'train/loss': 0.7042867541313171, 'train/grad_norm_pre_clip': 6.625666618347168, 'timing/data': 0.006794381886720657, 'timing/model': 0.9694659374654293, 'learning_rate/qwen_vl_interface': 9.898437305309004e-06, 'learning_rate/action_model': 9.898437305309003e-05, 'epoch': 0.23}) +[2026-06-17 17:39:20,659][starVLA.training.train_starvla][INFO] - Step 556, Loss: {'train/loss': 1.1567984819412231, 'train/grad_norm_pre_clip': 12.863792419433594, 'timing/data': 0.009347090497612953, 'timing/model': 0.9393814746290445, 'learning_rate/qwen_vl_interface': 9.897991991826575e-06, 'learning_rate/action_model': 9.897991991826574e-05, 'epoch': 0.23}) +[2026-06-17 17:39:24,504][starVLA.training.train_starvla][INFO] - Step 557, Loss: {'train/loss': 0.6504749059677124, 'train/grad_norm_pre_clip': 4.705261707305908, 'timing/data': 0.008473344147205353, 'timing/model': 0.9443033132702112, 'learning_rate/qwen_vl_interface': 9.897545714810404e-06, 'learning_rate/action_model': 9.897545714810403e-05, 'epoch': 0.23}) +[2026-06-17 17:39:28,501][starVLA.training.train_starvla][INFO] - Step 558, Loss: {'train/loss': 0.7079424262046814, 'train/grad_norm_pre_clip': 14.024725914001465, 'timing/data': 0.007480104453861713, 'timing/model': 1.102477390319109, 'learning_rate/qwen_vl_interface': 9.897098474353004e-06, 'learning_rate/action_model': 9.897098474353004e-05, 'epoch': 0.23}) +[2026-06-17 17:39:32,451][starVLA.training.train_starvla][INFO] - Step 559, Loss: {'train/loss': 0.5695770978927612, 'train/grad_norm_pre_clip': 6.645133018493652, 'timing/data': 0.009875115007162094, 'timing/model': 0.9923997223377228, 'learning_rate/qwen_vl_interface': 9.896650270547092e-06, 'learning_rate/action_model': 9.89665027054709e-05, 'epoch': 0.23}) +[2026-06-17 17:39:36,280][starVLA.training.train_starvla][INFO] - Step 560, Loss: {'train/loss': 0.828447699546814, 'train/grad_norm_pre_clip': 7.032575607299805, 'timing/data': 0.010404516011476517, 'timing/model': 1.123251048848033, 'learning_rate/qwen_vl_interface': 9.896201103485577e-06, 'learning_rate/action_model': 9.896201103485577e-05, 'epoch': 0.23}) +[2026-06-17 17:39:40,489][starVLA.training.train_starvla][INFO] - Step 561, Loss: {'train/loss': 0.9843849539756775, 'train/grad_norm_pre_clip': 7.826930999755859, 'timing/data': 0.009058571420609951, 'timing/model': 0.936048966832459, 'learning_rate/qwen_vl_interface': 9.895750973261576e-06, 'learning_rate/action_model': 9.895750973261574e-05, 'epoch': 0.23}) +[2026-06-17 17:39:44,373][starVLA.training.train_starvla][INFO] - Step 562, Loss: {'train/loss': 0.8333992958068848, 'train/grad_norm_pre_clip': 6.416029453277588, 'timing/data': 0.01239991094917059, 'timing/model': 0.9264359250664711, 'learning_rate/qwen_vl_interface': 9.895299879968398e-06, 'learning_rate/action_model': 9.895299879968398e-05, 'epoch': 0.23}) +[2026-06-17 17:39:48,287][starVLA.training.train_starvla][INFO] - Step 563, Loss: {'train/loss': 0.6733723878860474, 'train/grad_norm_pre_clip': 7.892154216766357, 'timing/data': 0.014826888218522072, 'timing/model': 0.9298880184069276, 'learning_rate/qwen_vl_interface': 9.894847823699555e-06, 'learning_rate/action_model': 9.894847823699555e-05, 'epoch': 0.23}) +[2026-06-17 17:39:52,268][starVLA.training.train_starvla][INFO] - Step 564, Loss: {'train/loss': 0.6949414610862732, 'train/grad_norm_pre_clip': 7.505761623382568, 'timing/data': 0.015785586088895798, 'timing/model': 1.1095268055796623, 'learning_rate/qwen_vl_interface': 9.894394804548763e-06, 'learning_rate/action_model': 9.894394804548762e-05, 'epoch': 0.23}) +[2026-06-17 17:39:56,039][starVLA.training.train_starvla][INFO] - Step 565, Loss: {'train/loss': 0.5729672312736511, 'train/grad_norm_pre_clip': 9.751432418823242, 'timing/data': 0.009382514283061028, 'timing/model': 0.9361668769270182, 'learning_rate/qwen_vl_interface': 9.893940822609928e-06, 'learning_rate/action_model': 9.893940822609927e-05, 'epoch': 0.23}) +[2026-06-17 17:39:59,904][starVLA.training.train_starvla][INFO] - Step 566, Loss: {'train/loss': 1.024003267288208, 'train/grad_norm_pre_clip': 12.38231372833252, 'timing/data': 0.008677925914525986, 'timing/model': 1.1886481093242764, 'learning_rate/qwen_vl_interface': 9.893485877977163e-06, 'learning_rate/action_model': 9.893485877977163e-05, 'epoch': 0.23}) +[2026-06-17 17:40:03,760][starVLA.training.train_starvla][INFO] - Step 567, Loss: {'train/loss': 0.5661014914512634, 'train/grad_norm_pre_clip': 5.773358345031738, 'timing/data': 0.014231047593057156, 'timing/model': 1.1481649773195386, 'learning_rate/qwen_vl_interface': 9.89302997074478e-06, 'learning_rate/action_model': 9.89302997074478e-05, 'epoch': 0.23}) +[2026-06-17 17:40:07,542][starVLA.training.train_starvla][INFO] - Step 568, Loss: {'train/loss': 0.8503215312957764, 'train/grad_norm_pre_clip': 9.416067123413086, 'timing/data': 0.014146106317639351, 'timing/model': 0.9388286992907524, 'learning_rate/qwen_vl_interface': 9.892573101007288e-06, 'learning_rate/action_model': 9.892573101007289e-05, 'epoch': 0.23}) +[2026-06-17 17:40:11,865][starVLA.training.train_starvla][INFO] - Step 569, Loss: {'train/loss': 1.05852210521698, 'train/grad_norm_pre_clip': 7.433675289154053, 'timing/data': 0.009867451153695583, 'timing/model': 0.9415753679350019, 'learning_rate/qwen_vl_interface': 9.892115268859398e-06, 'learning_rate/action_model': 9.892115268859397e-05, 'epoch': 0.23}) +[2026-06-17 17:40:15,669][starVLA.training.train_starvla][INFO] - Step 570, Loss: {'train/loss': 0.8152559995651245, 'train/grad_norm_pre_clip': 6.925981521606445, 'timing/data': 0.010797083377838135, 'timing/model': 0.9532549120485783, 'learning_rate/qwen_vl_interface': 9.891656474396015e-06, 'learning_rate/action_model': 9.891656474396015e-05, 'epoch': 0.23}) +[2026-06-17 17:40:19,466][starVLA.training.train_starvla][INFO] - Step 571, Loss: {'train/loss': 1.040486216545105, 'train/grad_norm_pre_clip': 7.94730806350708, 'timing/data': 0.009383536875247955, 'timing/model': 0.9231606759130955, 'learning_rate/qwen_vl_interface': 9.891196717712253e-06, 'learning_rate/action_model': 9.891196717712252e-05, 'epoch': 0.24}) +[2026-06-17 17:40:23,278][starVLA.training.train_starvla][INFO] - Step 572, Loss: {'train/loss': 0.752295732498169, 'train/grad_norm_pre_clip': 8.61825180053711, 'timing/data': 0.008919536136090755, 'timing/model': 0.9309552228078246, 'learning_rate/qwen_vl_interface': 9.890735998903414e-06, 'learning_rate/action_model': 9.890735998903414e-05, 'epoch': 0.24}) +[2026-06-17 17:40:27,206][starVLA.training.train_starvla][INFO] - Step 573, Loss: {'train/loss': 0.7611224055290222, 'train/grad_norm_pre_clip': 8.3848237991333, 'timing/data': 0.0184283209964633, 'timing/model': 1.2249765843153, 'learning_rate/qwen_vl_interface': 9.890274318065012e-06, 'learning_rate/action_model': 9.890274318065011e-05, 'epoch': 0.24}) +[2026-06-17 17:40:31,073][starVLA.training.train_starvla][INFO] - Step 574, Loss: {'train/loss': 0.7481508255004883, 'train/grad_norm_pre_clip': 7.463810443878174, 'timing/data': 0.009414486587047577, 'timing/model': 1.1406597141176462, 'learning_rate/qwen_vl_interface': 9.889811675292748e-06, 'learning_rate/action_model': 9.889811675292749e-05, 'epoch': 0.24}) +[2026-06-17 17:40:34,670][starVLA.training.train_starvla][INFO] - Step 575, Loss: {'train/loss': 0.4509608745574951, 'train/grad_norm_pre_clip': 4.6392717361450195, 'timing/data': 0.009146430529654026, 'timing/model': 0.9315654719248414, 'learning_rate/qwen_vl_interface': 9.889348070682533e-06, 'learning_rate/action_model': 9.889348070682533e-05, 'epoch': 0.24}) +[2026-06-17 17:40:38,614][starVLA.training.train_starvla][INFO] - Step 576, Loss: {'train/loss': 0.8383532762527466, 'train/grad_norm_pre_clip': 10.57717227935791, 'timing/data': 0.00854761153459549, 'timing/model': 0.9590919902548194, 'learning_rate/qwen_vl_interface': 9.888883504330469e-06, 'learning_rate/action_model': 9.888883504330468e-05, 'epoch': 0.24}) +[2026-06-17 17:40:42,443][starVLA.training.train_starvla][INFO] - Step 577, Loss: {'train/loss': 0.8920983672142029, 'train/grad_norm_pre_clip': 9.11438274383545, 'timing/data': 0.016315560787916183, 'timing/model': 0.9274850990623236, 'learning_rate/qwen_vl_interface': 9.888417976332865e-06, 'learning_rate/action_model': 9.888417976332865e-05, 'epoch': 0.24}) +[2026-06-17 17:40:46,251][starVLA.training.train_starvla][INFO] - Step 578, Loss: {'train/loss': 0.6657795906066895, 'train/grad_norm_pre_clip': 8.421354293823242, 'timing/data': 0.013577538542449474, 'timing/model': 0.9280374683439732, 'learning_rate/qwen_vl_interface': 9.887951486786222e-06, 'learning_rate/action_model': 9.887951486786221e-05, 'epoch': 0.24}) +[2026-06-17 17:40:50,141][starVLA.training.train_starvla][INFO] - Step 579, Loss: {'train/loss': 1.0093492269515991, 'train/grad_norm_pre_clip': 14.985892295837402, 'timing/data': 0.011982835829257965, 'timing/model': 0.939019126817584, 'learning_rate/qwen_vl_interface': 9.887484035787246e-06, 'learning_rate/action_model': 9.887484035787246e-05, 'epoch': 0.24}) +[2026-06-17 17:40:53,977][starVLA.training.train_starvla][INFO] - Step 580, Loss: {'train/loss': 1.0782051086425781, 'train/grad_norm_pre_clip': 13.312501907348633, 'timing/data': 0.00787044595927, 'timing/model': 1.1652282876893878, 'learning_rate/qwen_vl_interface': 9.887015623432837e-06, 'learning_rate/action_model': 9.887015623432837e-05, 'epoch': 0.24}) +[2026-06-17 17:40:57,877][starVLA.training.train_starvla][INFO] - Step 581, Loss: {'train/loss': 0.6676252484321594, 'train/grad_norm_pre_clip': 7.434388160705566, 'timing/data': 0.02640837337821722, 'timing/model': 1.1174280093982816, 'learning_rate/qwen_vl_interface': 9.886546249820101e-06, 'learning_rate/action_model': 9.886546249820102e-05, 'epoch': 0.24}) +[2026-06-17 17:41:01,990][starVLA.training.train_starvla][INFO] - Step 582, Loss: {'train/loss': 1.154865026473999, 'train/grad_norm_pre_clip': 12.334907531738281, 'timing/data': 0.019237946718931198, 'timing/model': 0.9394547995179892, 'learning_rate/qwen_vl_interface': 9.886075915046339e-06, 'learning_rate/action_model': 9.886075915046338e-05, 'epoch': 0.24}) +[2026-06-17 17:41:06,665][starVLA.training.train_starvla][INFO] - Step 583, Loss: {'train/loss': 0.8709728121757507, 'train/grad_norm_pre_clip': 8.150355339050293, 'timing/data': 0.015901888720691204, 'timing/model': 1.030373108573258, 'learning_rate/qwen_vl_interface': 9.885604619209052e-06, 'learning_rate/action_model': 9.885604619209051e-05, 'epoch': 0.24}) +[2026-06-17 17:41:10,652][starVLA.training.train_starvla][INFO] - Step 584, Loss: {'train/loss': 0.8366225361824036, 'train/grad_norm_pre_clip': 7.143681526184082, 'timing/data': 0.020434831269085407, 'timing/model': 0.9526112340390682, 'learning_rate/qwen_vl_interface': 9.885132362405935e-06, 'learning_rate/action_model': 9.885132362405936e-05, 'epoch': 0.24}) +[2026-06-17 17:41:14,647][starVLA.training.train_starvla][INFO] - Step 585, Loss: {'train/loss': 0.7990860939025879, 'train/grad_norm_pre_clip': 8.581887245178223, 'timing/data': 0.014603924006223679, 'timing/model': 0.9322692723944783, 'learning_rate/qwen_vl_interface': 9.884659144734896e-06, 'learning_rate/action_model': 9.884659144734896e-05, 'epoch': 0.24}) +[2026-06-17 17:41:18,725][starVLA.training.train_starvla][INFO] - Step 586, Loss: {'train/loss': 0.7830513119697571, 'train/grad_norm_pre_clip': 7.3886799812316895, 'timing/data': 0.01675491314381361, 'timing/model': 1.15438396576792, 'learning_rate/qwen_vl_interface': 9.884184966294028e-06, 'learning_rate/action_model': 9.884184966294028e-05, 'epoch': 0.24}) +[2026-06-17 17:41:22,546][starVLA.training.train_starvla][INFO] - Step 587, Loss: {'train/loss': 0.9305186867713928, 'train/grad_norm_pre_clip': 9.620954513549805, 'timing/data': 0.022074691019952297, 'timing/model': 1.114972143433988, 'learning_rate/qwen_vl_interface': 9.883709827181632e-06, 'learning_rate/action_model': 9.88370982718163e-05, 'epoch': 0.24}) +[2026-06-17 17:41:26,308][starVLA.training.train_starvla][INFO] - Step 588, Loss: {'train/loss': 1.045304298400879, 'train/grad_norm_pre_clip': 9.370702743530273, 'timing/data': 0.01294972375035286, 'timing/model': 0.9611797071993351, 'learning_rate/qwen_vl_interface': 9.8832337274962e-06, 'learning_rate/action_model': 9.883233727496198e-05, 'epoch': 0.24}) +[2026-06-17 17:41:30,283][starVLA.training.train_starvla][INFO] - Step 589, Loss: {'train/loss': 0.7775374054908752, 'train/grad_norm_pre_clip': 7.469885349273682, 'timing/data': 0.017447689548134804, 'timing/model': 0.9565061619505286, 'learning_rate/qwen_vl_interface': 9.88275666733643e-06, 'learning_rate/action_model': 9.882756667336429e-05, 'epoch': 0.24}) +[2026-06-17 17:41:34,623][starVLA.training.train_starvla][INFO] - Step 590, Loss: {'train/loss': 0.557211697101593, 'train/grad_norm_pre_clip': 6.199609279632568, 'timing/data': 0.01927023008465767, 'timing/model': 0.9679296035319567, 'learning_rate/qwen_vl_interface': 9.88227864680122e-06, 'learning_rate/action_model': 9.88227864680122e-05, 'epoch': 0.24}) +[2026-06-17 17:41:38,661][starVLA.training.train_starvla][INFO] - Step 591, Loss: {'train/loss': 0.6368412375450134, 'train/grad_norm_pre_clip': 6.649439811706543, 'timing/data': 0.016462021507322788, 'timing/model': 0.9414019444957376, 'learning_rate/qwen_vl_interface': 9.88179966598966e-06, 'learning_rate/action_model': 9.88179966598966e-05, 'epoch': 0.24}) +[2026-06-17 17:41:42,659][starVLA.training.train_starvla][INFO] - Step 592, Loss: {'train/loss': 1.2074289321899414, 'train/grad_norm_pre_clip': 14.735152244567871, 'timing/data': 0.01635877974331379, 'timing/model': 0.9418102568015456, 'learning_rate/qwen_vl_interface': 9.881319725001046e-06, 'learning_rate/action_model': 9.881319725001046e-05, 'epoch': 0.24}) +[2026-06-17 17:41:46,853][starVLA.training.train_starvla][INFO] - Step 593, Loss: {'train/loss': 0.9597420692443848, 'train/grad_norm_pre_clip': 8.768388748168945, 'timing/data': 0.01797270867973566, 'timing/model': 1.380937465466559, 'learning_rate/qwen_vl_interface': 9.88083882393487e-06, 'learning_rate/action_model': 9.880838823934869e-05, 'epoch': 0.24}) +[2026-06-17 17:41:50,805][starVLA.training.train_starvla][INFO] - Step 594, Loss: {'train/loss': 0.9195674657821655, 'train/grad_norm_pre_clip': 7.747944355010986, 'timing/data': 0.01714204717427492, 'timing/model': 1.238987105898559, 'learning_rate/qwen_vl_interface': 9.880356962890823e-06, 'learning_rate/action_model': 9.880356962890822e-05, 'epoch': 0.24}) +[2026-06-17 17:41:56,730][starVLA.training.train_starvla][INFO] - Step 595, Loss: {'train/loss': 0.5947238206863403, 'train/grad_norm_pre_clip': 7.152231693267822, 'timing/data': 0.05165389459580183, 'timing/model': 2.0326898572966456, 'learning_rate/qwen_vl_interface': 9.879874141968792e-06, 'learning_rate/action_model': 9.879874141968792e-05, 'epoch': 0.24}) +[2026-06-17 17:42:01,218][starVLA.training.train_starvla][INFO] - Step 596, Loss: {'train/loss': 0.45572471618652344, 'train/grad_norm_pre_clip': 7.040951728820801, 'timing/data': 0.02062141429632902, 'timing/model': 0.9631605353206396, 'learning_rate/qwen_vl_interface': 9.879390361268873e-06, 'learning_rate/action_model': 9.879390361268873e-05, 'epoch': 0.25}) +[2026-06-17 17:42:05,229][starVLA.training.train_starvla][INFO] - Step 597, Loss: {'train/loss': 0.6175529956817627, 'train/grad_norm_pre_clip': 5.212855339050293, 'timing/data': 0.018996169790625572, 'timing/model': 0.9576454274356365, 'learning_rate/qwen_vl_interface': 9.878905620891349e-06, 'learning_rate/action_model': 9.878905620891349e-05, 'epoch': 0.25}) +[2026-06-17 17:42:10,158][starVLA.training.train_starvla][INFO] - Step 598, Loss: {'train/loss': 0.6801563501358032, 'train/grad_norm_pre_clip': 6.794193744659424, 'timing/data': 0.01581541169434786, 'timing/model': 0.9469664394855499, 'learning_rate/qwen_vl_interface': 9.87841992093671e-06, 'learning_rate/action_model': 9.87841992093671e-05, 'epoch': 0.25}) +[2026-06-17 17:42:15,627][starVLA.training.train_starvla][INFO] - Step 599, Loss: {'train/loss': 0.6718193292617798, 'train/grad_norm_pre_clip': 6.084566116333008, 'timing/data': 0.01944339368492365, 'timing/model': 0.9574151234701276, 'learning_rate/qwen_vl_interface': 9.877933261505642e-06, 'learning_rate/action_model': 9.877933261505641e-05, 'epoch': 0.25}) +[2026-06-17 17:44:19,849][starVLA.training.train_starvla][INFO] - Step 600, Loss: {'train/loss': 0.9940696954727173, 'train/grad_norm_pre_clip': 8.988385200500488, 'eval/loss': 0.8974494338035583, 'eval/latency_6/loss': 0.8974494338035583, 'eval/demon_fire_precision': 0.9311897106109325, 'eval/demon_fire_recall': 0.936005171299289, 'eval/demon_fire_f1': 0.9335912314635719, 'eval/demon_fire_rate_model': 0.8653311074012242, 'eval/demon_fire_rate_teacher': 0.860879243183083, 'eval/demon_left_precision': 0.8558951965065502, 'eval/demon_left_recall': 0.34385964912280703, 'eval/demon_left_f1': 0.4906132665832291, 'eval/demon_left_rate_model': 0.06371730662214803, 'eval/demon_left_rate_teacher': 0.15859766277128548, 'eval/demon_right_precision': 0.5217391304347826, 'eval/demon_right_recall': 0.8110599078341014, 'eval/demon_right_f1': 0.634996993385448, 'eval/demon_right_rate_model': 0.2815804117974402, 'eval/demon_right_rate_teacher': 0.18113522537562604, 'eval/demon_move_f1': 0.5628051299843386, 'eval/demon_cc_f1': 0.7795222623301323, 'eval/latency_6/demon_fire_precision': 0.9311897106109325, 'eval/latency_6/demon_fire_recall': 0.936005171299289, 'eval/latency_6/demon_fire_f1': 0.9335912314635719, 'eval/latency_6/demon_fire_rate_model': 0.8653311074012242, 'eval/latency_6/demon_fire_rate_teacher': 0.860879243183083, 'eval/latency_6/demon_left_precision': 0.8558951965065502, 'eval/latency_6/demon_left_recall': 0.34385964912280703, 'eval/latency_6/demon_left_f1': 0.4906132665832291, 'eval/latency_6/demon_left_rate_model': 0.06371730662214803, 'eval/latency_6/demon_left_rate_teacher': 0.15859766277128548, 'eval/latency_6/demon_right_precision': 0.5217391304347826, 'eval/latency_6/demon_right_recall': 0.8110599078341014, 'eval/latency_6/demon_right_f1': 0.634996993385448, 'eval/latency_6/demon_right_rate_model': 0.2815804117974402, 'eval/latency_6/demon_right_rate_teacher': 0.18113522537562604, 'eval/latency_6/demon_move_f1': 0.5628051299843386, 'eval/latency_6/demon_cc_f1': 0.7795222623301323, 'timing/data': 0.018751345574855804, 'timing/model': 1.2391882101073861, 'learning_rate/qwen_vl_interface': 9.877445642699026e-06, 'learning_rate/action_model': 9.877445642699026e-05, 'epoch': 0.25}) +[2026-06-17 17:44:19,912][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_600_state +[2026-06-17 17:44:19,917][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 17:44:27,530][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_600_state/model.safetensors +[2026-06-17 17:44:43,923][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_600_state/optimizer.bin +[2026-06-17 17:44:43,924][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_600_state/sampler.bin +[2026-06-17 17:44:43,924][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_600_state/sampler_1.bin +[2026-06-17 17:44:43,926][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_600_state/random_states_0.pkl +[2026-06-17 17:44:43,931][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 600 +[2026-06-17 17:44:43,931][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 17:44:43,942][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 17:44:47,869][starVLA.training.train_starvla][INFO] - Step 601, Loss: {'train/loss': 0.6562047600746155, 'train/grad_norm_pre_clip': 26.96673011779785, 'timing/data': 0.026415260508656502, 'timing/model': 0.9624925386160612, 'learning_rate/qwen_vl_interface': 9.876957064617951e-06, 'learning_rate/action_model': 9.876957064617951e-05, 'epoch': 0.25}) +[2026-06-17 17:44:51,679][starVLA.training.train_starvla][INFO] - Step 602, Loss: {'train/loss': 0.8553215265274048, 'train/grad_norm_pre_clip': 8.10464096069336, 'timing/data': 0.019431845284998417, 'timing/model': 0.9392696991562843, 'learning_rate/qwen_vl_interface': 9.876467527363696e-06, 'learning_rate/action_model': 9.876467527363697e-05, 'epoch': 0.25}) +[2026-06-17 17:44:55,681][starVLA.training.train_starvla][INFO] - Step 603, Loss: {'train/loss': 0.7283724546432495, 'train/grad_norm_pre_clip': 17.234844207763672, 'timing/data': 0.018450372852385044, 'timing/model': 1.243019008077681, 'learning_rate/qwen_vl_interface': 9.875977031037746e-06, 'learning_rate/action_model': 9.875977031037745e-05, 'epoch': 0.25}) +[2026-06-17 17:44:59,603][starVLA.training.train_starvla][INFO] - Step 604, Loss: {'train/loss': 0.9453368782997131, 'train/grad_norm_pre_clip': 6.786897659301758, 'timing/data': 0.01955164596438408, 'timing/model': 1.0466720517724752, 'learning_rate/qwen_vl_interface': 9.875485575741779e-06, 'learning_rate/action_model': 9.875485575741779e-05, 'epoch': 0.25}) +[2026-06-17 17:45:03,660][starVLA.training.train_starvla][INFO] - Step 605, Loss: {'train/loss': 0.8825141787528992, 'train/grad_norm_pre_clip': 9.67930793762207, 'timing/data': 0.02061222866177559, 'timing/model': 0.9405116885900497, 'learning_rate/qwen_vl_interface': 9.874993161577674e-06, 'learning_rate/action_model': 9.874993161577675e-05, 'epoch': 0.25}) +[2026-06-17 17:45:07,747][starVLA.training.train_starvla][INFO] - Step 606, Loss: {'train/loss': 1.0339974164962769, 'train/grad_norm_pre_clip': 9.201629638671875, 'timing/data': 0.019446364603936672, 'timing/model': 0.9444515416398644, 'learning_rate/qwen_vl_interface': 9.874499788647512e-06, 'learning_rate/action_model': 9.874499788647513e-05, 'epoch': 0.25}) +[2026-06-17 17:45:11,597][starVLA.training.train_starvla][INFO] - Step 607, Loss: {'train/loss': 1.0909080505371094, 'train/grad_norm_pre_clip': 10.774759292602539, 'timing/data': 0.01759233418852091, 'timing/model': 0.9470278285443783, 'learning_rate/qwen_vl_interface': 9.874005457053565e-06, 'learning_rate/action_model': 9.874005457053565e-05, 'epoch': 0.25}) +[2026-06-17 17:45:15,788][starVLA.training.train_starvla][INFO] - Step 608, Loss: {'train/loss': 0.9068801999092102, 'train/grad_norm_pre_clip': 9.313222885131836, 'timing/data': 0.034321135841310024, 'timing/model': 1.03673829510808, 'learning_rate/qwen_vl_interface': 9.873510166898315e-06, 'learning_rate/action_model': 9.873510166898314e-05, 'epoch': 0.25}) +[2026-06-17 17:45:19,724][starVLA.training.train_starvla][INFO] - Step 609, Loss: {'train/loss': 0.825261652469635, 'train/grad_norm_pre_clip': 7.537093162536621, 'timing/data': 0.025247691199183464, 'timing/model': 1.187123385258019, 'learning_rate/qwen_vl_interface': 9.87301391828443e-06, 'learning_rate/action_model': 9.873013918284428e-05, 'epoch': 0.25}) +[2026-06-17 17:45:24,005][starVLA.training.train_starvla][INFO] - Step 610, Loss: {'train/loss': 0.8992776274681091, 'train/grad_norm_pre_clip': 13.986710548400879, 'timing/data': 0.02040658611804247, 'timing/model': 1.4287042925134301, 'learning_rate/qwen_vl_interface': 9.872516711314784e-06, 'learning_rate/action_model': 9.872516711314784e-05, 'epoch': 0.25}) +[2026-06-17 17:45:27,875][starVLA.training.train_starvla][INFO] - Step 611, Loss: {'train/loss': 0.7865341901779175, 'train/grad_norm_pre_clip': 7.952881336212158, 'timing/data': 0.020722446031868458, 'timing/model': 0.9711494371294975, 'learning_rate/qwen_vl_interface': 9.872018546092452e-06, 'learning_rate/action_model': 9.872018546092451e-05, 'epoch': 0.25}) +[2026-06-17 17:45:31,739][starVLA.training.train_starvla][INFO] - Step 612, Loss: {'train/loss': 0.9374820590019226, 'train/grad_norm_pre_clip': 8.783112525939941, 'timing/data': 0.018179134465754032, 'timing/model': 0.9333475036546588, 'learning_rate/qwen_vl_interface': 9.871519422720699e-06, 'learning_rate/action_model': 9.871519422720698e-05, 'epoch': 0.25}) +[2026-06-17 17:45:35,606][starVLA.training.train_starvla][INFO] - Step 613, Loss: {'train/loss': 0.7772431373596191, 'train/grad_norm_pre_clip': 7.8559465408325195, 'timing/data': 0.02519835066050291, 'timing/model': 0.9577823393046856, 'learning_rate/qwen_vl_interface': 9.871019341303e-06, 'learning_rate/action_model': 9.871019341303e-05, 'epoch': 0.25}) +[2026-06-17 17:45:39,459][starVLA.training.train_starvla][INFO] - Step 614, Loss: {'train/loss': 1.2290195226669312, 'train/grad_norm_pre_clip': 29.108966827392578, 'timing/data': 0.021095935255289078, 'timing/model': 0.934564171358943, 'learning_rate/qwen_vl_interface': 9.870518301943018e-06, 'learning_rate/action_model': 9.870518301943018e-05, 'epoch': 0.25}) +[2026-06-17 17:45:43,385][starVLA.training.train_starvla][INFO] - Step 615, Loss: {'train/loss': 1.172542691230774, 'train/grad_norm_pre_clip': 13.36317253112793, 'timing/data': 0.020309485495090485, 'timing/model': 0.9492747681215405, 'learning_rate/qwen_vl_interface': 9.870016304744617e-06, 'learning_rate/action_model': 9.870016304744616e-05, 'epoch': 0.25}) +[2026-06-17 17:45:48,331][starVLA.training.train_starvla][INFO] - Step 616, Loss: {'train/loss': 0.7817427515983582, 'train/grad_norm_pre_clip': 7.319106101989746, 'timing/data': 0.02198367938399315, 'timing/model': 1.237255297601223, 'learning_rate/qwen_vl_interface': 9.869513349811868e-06, 'learning_rate/action_model': 9.869513349811868e-05, 'epoch': 0.25}) +[2026-06-17 17:45:51,979][starVLA.training.train_starvla][INFO] - Step 617, Loss: {'train/loss': 0.5532926321029663, 'train/grad_norm_pre_clip': 5.803870677947998, 'timing/data': 0.02089494839310646, 'timing/model': 0.9604725884273648, 'learning_rate/qwen_vl_interface': 9.86900943724903e-06, 'learning_rate/action_model': 9.869009437249031e-05, 'epoch': 0.25}) +[2026-06-17 17:45:58,838][starVLA.training.train_starvla][INFO] - Step 618, Loss: {'train/loss': 0.6178798675537109, 'train/grad_norm_pre_clip': 5.659029006958008, 'timing/data': 0.03218111768364906, 'timing/model': 0.9433707678690553, 'learning_rate/qwen_vl_interface': 9.868504567160566e-06, 'learning_rate/action_model': 9.868504567160565e-05, 'epoch': 0.25}) +[2026-06-17 17:46:02,647][starVLA.training.train_starvla][INFO] - Step 619, Loss: {'train/loss': 0.5057773590087891, 'train/grad_norm_pre_clip': 8.685578346252441, 'timing/data': 0.024269108660519123, 'timing/model': 0.9361550947651267, 'learning_rate/qwen_vl_interface': 9.867998739651134e-06, 'learning_rate/action_model': 9.867998739651134e-05, 'epoch': 0.25}) +[2026-06-17 17:46:06,564][starVLA.training.train_starvla][INFO] - Step 620, Loss: {'train/loss': 0.8081028461456299, 'train/grad_norm_pre_clip': 7.930338382720947, 'timing/data': 0.01990113314241171, 'timing/model': 0.9371858481317759, 'learning_rate/qwen_vl_interface': 9.867491954825595e-06, 'learning_rate/action_model': 9.867491954825595e-05, 'epoch': 0.26}) +[2026-06-17 17:46:10,461][starVLA.training.train_starvla][INFO] - Step 621, Loss: {'train/loss': 0.6863448023796082, 'train/grad_norm_pre_clip': 6.061117649078369, 'timing/data': 0.021834262646734715, 'timing/model': 0.9553340701386333, 'learning_rate/qwen_vl_interface': 9.866984212789007e-06, 'learning_rate/action_model': 9.866984212789006e-05, 'epoch': 0.26}) +[2026-06-17 17:46:14,268][starVLA.training.train_starvla][INFO] - Step 622, Loss: {'train/loss': 0.9478551149368286, 'train/grad_norm_pre_clip': 7.485236644744873, 'timing/data': 0.022623431868851185, 'timing/model': 1.0945147657766938, 'learning_rate/qwen_vl_interface': 9.86647551364662e-06, 'learning_rate/action_model': 9.86647551364662e-05, 'epoch': 0.26}) +[2026-06-17 17:46:18,185][starVLA.training.train_starvla][INFO] - Step 623, Loss: {'train/loss': 0.900421142578125, 'train/grad_norm_pre_clip': 13.809069633483887, 'timing/data': 0.0202958220615983, 'timing/model': 1.2381667476147413, 'learning_rate/qwen_vl_interface': 9.865965857503893e-06, 'learning_rate/action_model': 9.865965857503894e-05, 'epoch': 0.26}) +[2026-06-17 17:46:21,870][starVLA.training.train_starvla][INFO] - Step 624, Loss: {'train/loss': 1.4238629341125488, 'train/grad_norm_pre_clip': 11.625184059143066, 'timing/data': 0.031863363459706306, 'timing/model': 0.9593939557671547, 'learning_rate/qwen_vl_interface': 9.865455244466478e-06, 'learning_rate/action_model': 9.865455244466477e-05, 'epoch': 0.26}) +[2026-06-17 17:46:25,849][starVLA.training.train_starvla][INFO] - Step 625, Loss: {'train/loss': 1.1834157705307007, 'train/grad_norm_pre_clip': 9.946203231811523, 'timing/data': 0.01967137400060892, 'timing/model': 0.949395140632987, 'learning_rate/qwen_vl_interface': 9.864943674640223e-06, 'learning_rate/action_model': 9.864943674640222e-05, 'epoch': 0.26}) +[2026-06-17 17:46:32,633][starVLA.training.train_starvla][INFO] - Step 626, Loss: {'train/loss': 0.8645181059837341, 'train/grad_norm_pre_clip': 14.932482719421387, 'timing/data': 0.021645276807248592, 'timing/model': 1.064266444183886, 'learning_rate/qwen_vl_interface': 9.864431148131178e-06, 'learning_rate/action_model': 9.864431148131178e-05, 'epoch': 0.26}) +[2026-06-17 17:46:36,713][starVLA.training.train_starvla][INFO] - Step 627, Loss: {'train/loss': 0.9599322080612183, 'train/grad_norm_pre_clip': 12.893366813659668, 'timing/data': 0.02482429798692465, 'timing/model': 0.950340305455029, 'learning_rate/qwen_vl_interface': 9.863917665045591e-06, 'learning_rate/action_model': 9.863917665045592e-05, 'epoch': 0.26}) +[2026-06-17 17:46:40,635][starVLA.training.train_starvla][INFO] - Step 628, Loss: {'train/loss': 1.0705313682556152, 'train/grad_norm_pre_clip': 7.855632305145264, 'timing/data': 0.01719077117741108, 'timing/model': 1.019766729325056, 'learning_rate/qwen_vl_interface': 9.863403225489909e-06, 'learning_rate/action_model': 9.863403225489909e-05, 'epoch': 0.26}) +[2026-06-17 17:46:44,588][starVLA.training.train_starvla][INFO] - Step 629, Loss: {'train/loss': 0.8415378332138062, 'train/grad_norm_pre_clip': 11.700496673583984, 'timing/data': 0.021259737201035023, 'timing/model': 1.189713821746409, 'learning_rate/qwen_vl_interface': 9.862887829570771e-06, 'learning_rate/action_model': 9.862887829570772e-05, 'epoch': 0.26}) +[2026-06-17 17:46:48,611][starVLA.training.train_starvla][INFO] - Step 630, Loss: {'train/loss': 0.8453064560890198, 'train/grad_norm_pre_clip': 14.727307319641113, 'timing/data': 0.021752660162746906, 'timing/model': 1.2509448938071728, 'learning_rate/qwen_vl_interface': 9.862371477395024e-06, 'learning_rate/action_model': 9.862371477395025e-05, 'epoch': 0.26}) +[2026-06-17 17:46:54,684][starVLA.training.train_starvla][INFO] - Step 631, Loss: {'train/loss': 0.8932382464408875, 'train/grad_norm_pre_clip': 6.180320739746094, 'timing/data': 0.019348652102053165, 'timing/model': 3.1634783912450075, 'learning_rate/qwen_vl_interface': 9.861854169069708e-06, 'learning_rate/action_model': 9.861854169069707e-05, 'epoch': 0.26}) +[2026-06-17 17:46:59,185][starVLA.training.train_starvla][INFO] - Step 632, Loss: {'train/loss': 0.8797361850738525, 'train/grad_norm_pre_clip': 7.998074531555176, 'timing/data': 0.020444896072149277, 'timing/model': 1.010081423446536, 'learning_rate/qwen_vl_interface': 9.861335904702058e-06, 'learning_rate/action_model': 9.861335904702058e-05, 'epoch': 0.26}) +[2026-06-17 17:47:04,221][starVLA.training.train_starvla][INFO] - Step 633, Loss: {'train/loss': 1.0221188068389893, 'train/grad_norm_pre_clip': 5.997286796569824, 'timing/data': 0.020926130935549736, 'timing/model': 1.0429576020687819, 'learning_rate/qwen_vl_interface': 9.860816684399514e-06, 'learning_rate/action_model': 9.860816684399514e-05, 'epoch': 0.26}) +[2026-06-17 17:47:08,356][starVLA.training.train_starvla][INFO] - Step 634, Loss: {'train/loss': 0.7652612328529358, 'train/grad_norm_pre_clip': 5.742353916168213, 'timing/data': 0.021500924602150917, 'timing/model': 1.0957321133464575, 'learning_rate/qwen_vl_interface': 9.86029650826971e-06, 'learning_rate/action_model': 9.86029650826971e-05, 'epoch': 0.26}) +[2026-06-17 17:47:12,597][starVLA.training.train_starvla][INFO] - Step 635, Loss: {'train/loss': 0.6117081046104431, 'train/grad_norm_pre_clip': 5.408457279205322, 'timing/data': 0.021042142994701862, 'timing/model': 1.0610131369903684, 'learning_rate/qwen_vl_interface': 9.859775376420478e-06, 'learning_rate/action_model': 9.859775376420478e-05, 'epoch': 0.26}) +[2026-06-17 17:47:16,851][starVLA.training.train_starvla][INFO] - Step 636, Loss: {'train/loss': 0.7744820713996887, 'train/grad_norm_pre_clip': 6.629265785217285, 'timing/data': 0.02334785182029009, 'timing/model': 1.190668011084199, 'learning_rate/qwen_vl_interface': 9.85925328895985e-06, 'learning_rate/action_model': 9.85925328895985e-05, 'epoch': 0.26}) +[2026-06-17 17:47:20,823][starVLA.training.train_starvla][INFO] - Step 637, Loss: {'train/loss': 0.965034008026123, 'train/grad_norm_pre_clip': 7.726936340332031, 'timing/data': 0.016622480936348438, 'timing/model': 1.114503319375217, 'learning_rate/qwen_vl_interface': 9.858730245996057e-06, 'learning_rate/action_model': 9.858730245996057e-05, 'epoch': 0.26}) +[2026-06-17 17:47:24,740][starVLA.training.train_starvla][INFO] - Step 638, Loss: {'train/loss': 0.8370904326438904, 'train/grad_norm_pre_clip': 5.667871952056885, 'timing/data': 0.025736593641340733, 'timing/model': 0.9517401456832886, 'learning_rate/qwen_vl_interface': 9.858206247637522e-06, 'learning_rate/action_model': 9.858206247637523e-05, 'epoch': 0.26}) +[2026-06-17 17:47:28,697][starVLA.training.train_starvla][INFO] - Step 639, Loss: {'train/loss': 0.8240709900856018, 'train/grad_norm_pre_clip': 6.750852584838867, 'timing/data': 0.029381017200648785, 'timing/model': 0.9733268022537231, 'learning_rate/qwen_vl_interface': 9.857681293992876e-06, 'learning_rate/action_model': 9.857681293992877e-05, 'epoch': 0.26}) +[2026-06-17 17:47:32,729][starVLA.training.train_starvla][INFO] - Step 640, Loss: {'train/loss': 0.8188256025314331, 'train/grad_norm_pre_clip': 6.75993013381958, 'timing/data': 0.017616464756429195, 'timing/model': 1.072650014422834, 'learning_rate/qwen_vl_interface': 9.85715538517094e-06, 'learning_rate/action_model': 9.857155385170938e-05, 'epoch': 0.26}) +[2026-06-17 17:47:36,943][starVLA.training.train_starvla][INFO] - Step 641, Loss: {'train/loss': 0.6974513530731201, 'train/grad_norm_pre_clip': 7.2185139656066895, 'timing/data': 0.01872736867517233, 'timing/model': 0.9568792982026935, 'learning_rate/qwen_vl_interface': 9.856628521280733e-06, 'learning_rate/action_model': 9.856628521280733e-05, 'epoch': 0.26}) +[2026-06-17 17:47:40,895][starVLA.training.train_starvla][INFO] - Step 642, Loss: {'train/loss': 0.7592262625694275, 'train/grad_norm_pre_clip': 6.565021514892578, 'timing/data': 0.019852508790791035, 'timing/model': 1.2095319097861648, 'learning_rate/qwen_vl_interface': 9.856100702431477e-06, 'learning_rate/action_model': 9.856100702431477e-05, 'epoch': 0.26}) +[2026-06-17 17:47:44,778][starVLA.training.train_starvla][INFO] - Step 643, Loss: {'train/loss': 0.7967854738235474, 'train/grad_norm_pre_clip': 7.152450084686279, 'timing/data': 0.02211802452802658, 'timing/model': 1.0928391721099615, 'learning_rate/qwen_vl_interface': 9.85557192873259e-06, 'learning_rate/action_model': 9.85557192873259e-05, 'epoch': 0.26}) +[2026-06-17 17:47:48,476][starVLA.training.train_starvla][INFO] - Step 644, Loss: {'train/loss': 0.6843027472496033, 'train/grad_norm_pre_clip': 12.162444114685059, 'timing/data': 0.021016934886574745, 'timing/model': 0.9592793332412839, 'learning_rate/qwen_vl_interface': 9.855042200293684e-06, 'learning_rate/action_model': 9.855042200293685e-05, 'epoch': 0.27}) +[2026-06-17 17:47:52,307][starVLA.training.train_starvla][INFO] - Step 645, Loss: {'train/loss': 0.8191759586334229, 'train/grad_norm_pre_clip': 9.302614212036133, 'timing/data': 0.028607388958334923, 'timing/model': 0.9391623958945274, 'learning_rate/qwen_vl_interface': 9.854511517224578e-06, 'learning_rate/action_model': 9.854511517224578e-05, 'epoch': 0.27}) +[2026-06-17 17:47:56,174][starVLA.training.train_starvla][INFO] - Step 646, Loss: {'train/loss': 1.0088982582092285, 'train/grad_norm_pre_clip': 10.318021774291992, 'timing/data': 0.018847987987101078, 'timing/model': 0.9226212734356523, 'learning_rate/qwen_vl_interface': 9.853979879635278e-06, 'learning_rate/action_model': 9.853979879635278e-05, 'epoch': 0.27}) +[2026-06-17 17:48:00,121][starVLA.training.train_starvla][INFO] - Step 647, Loss: {'train/loss': 0.7616872787475586, 'train/grad_norm_pre_clip': 8.838310241699219, 'timing/data': 0.02323331870138645, 'timing/model': 0.9218708351254463, 'learning_rate/qwen_vl_interface': 9.853447287635995e-06, 'learning_rate/action_model': 9.853447287635995e-05, 'epoch': 0.27}) +[2026-06-17 17:48:03,951][starVLA.training.train_starvla][INFO] - Step 648, Loss: {'train/loss': 0.8822622299194336, 'train/grad_norm_pre_clip': 7.987016677856445, 'timing/data': 0.026557212695479393, 'timing/model': 0.9363113576546311, 'learning_rate/qwen_vl_interface': 9.852913741337136e-06, 'learning_rate/action_model': 9.852913741337136e-05, 'epoch': 0.27}) +[2026-06-17 17:48:07,723][starVLA.training.train_starvla][INFO] - Step 649, Loss: {'train/loss': 0.7796125411987305, 'train/grad_norm_pre_clip': 7.066770076751709, 'timing/data': 0.0161186084151268, 'timing/model': 1.10692647844553, 'learning_rate/qwen_vl_interface': 9.852379240849306e-06, 'learning_rate/action_model': 9.852379240849305e-05, 'epoch': 0.27}) +[2026-06-17 17:48:11,483][starVLA.training.train_starvla][INFO] - Step 650, Loss: {'train/loss': 0.7407610416412354, 'train/grad_norm_pre_clip': 7.515966892242432, 'timing/data': 0.01800814364105463, 'timing/model': 1.0721386466175318, 'learning_rate/qwen_vl_interface': 9.851843786283306e-06, 'learning_rate/action_model': 9.851843786283306e-05, 'epoch': 0.27}) +[2026-06-17 17:48:15,789][starVLA.training.train_starvla][INFO] - Step 651, Loss: {'train/loss': 0.9529151320457458, 'train/grad_norm_pre_clip': 9.34341812133789, 'timing/data': 0.031146244145929813, 'timing/model': 1.16705886926502, 'learning_rate/qwen_vl_interface': 9.851307377750137e-06, 'learning_rate/action_model': 9.851307377750137e-05, 'epoch': 0.27}) +[2026-06-17 17:48:19,574][starVLA.training.train_starvla][INFO] - Step 652, Loss: {'train/loss': 0.6431813836097717, 'train/grad_norm_pre_clip': 7.2060089111328125, 'timing/data': 0.024014814756810665, 'timing/model': 0.9408885771408677, 'learning_rate/qwen_vl_interface': 9.850770015360999e-06, 'learning_rate/action_model': 9.850770015360999e-05, 'epoch': 0.27}) +[2026-06-17 17:48:23,660][starVLA.training.train_starvla][INFO] - Step 653, Loss: {'train/loss': 0.864936351776123, 'train/grad_norm_pre_clip': 7.529792785644531, 'timing/data': 0.01949838362634182, 'timing/model': 0.9403691189363599, 'learning_rate/qwen_vl_interface': 9.850231699227284e-06, 'learning_rate/action_model': 9.850231699227284e-05, 'epoch': 0.27}) +[2026-06-17 17:48:27,857][starVLA.training.train_starvla][INFO] - Step 654, Loss: {'train/loss': 0.7047008275985718, 'train/grad_norm_pre_clip': 7.463517665863037, 'timing/data': 0.020217486657202244, 'timing/model': 0.9490816295146942, 'learning_rate/qwen_vl_interface': 9.84969242946059e-06, 'learning_rate/action_model': 9.84969242946059e-05, 'epoch': 0.27}) +[2026-06-17 17:48:31,893][starVLA.training.train_starvla][INFO] - Step 655, Loss: {'train/loss': 0.9045925736427307, 'train/grad_norm_pre_clip': 7.9922871589660645, 'timing/data': 0.024975675158202648, 'timing/model': 1.2018338283523917, 'learning_rate/qwen_vl_interface': 9.849152206172704e-06, 'learning_rate/action_model': 9.849152206172704e-05, 'epoch': 0.27}) +[2026-06-17 17:48:35,632][starVLA.training.train_starvla][INFO] - Step 656, Loss: {'train/loss': 0.8741995692253113, 'train/grad_norm_pre_clip': 42.88241195678711, 'timing/data': 0.019737372174859047, 'timing/model': 0.9576892899349332, 'learning_rate/qwen_vl_interface': 9.84861102947562e-06, 'learning_rate/action_model': 9.848611029475619e-05, 'epoch': 0.27}) +[2026-06-17 17:48:39,472][starVLA.training.train_starvla][INFO] - Step 657, Loss: {'train/loss': 0.6211074590682983, 'train/grad_norm_pre_clip': 6.516239166259766, 'timing/data': 0.016986670903861523, 'timing/model': 0.9451651657000184, 'learning_rate/qwen_vl_interface': 9.848068899481516e-06, 'learning_rate/action_model': 9.848068899481516e-05, 'epoch': 0.27}) +[2026-06-17 17:48:43,478][starVLA.training.train_starvla][INFO] - Step 658, Loss: {'train/loss': 0.7035248875617981, 'train/grad_norm_pre_clip': 12.239957809448242, 'timing/data': 0.026366825215518475, 'timing/model': 0.9477577190846205, 'learning_rate/qwen_vl_interface': 9.847525816302784e-06, 'learning_rate/action_model': 9.847525816302784e-05, 'epoch': 0.27}) +[2026-06-17 17:48:47,412][starVLA.training.train_starvla][INFO] - Step 659, Loss: {'train/loss': 0.8694092631340027, 'train/grad_norm_pre_clip': 7.177662372589111, 'timing/data': 0.025091159157454967, 'timing/model': 0.987945961765945, 'learning_rate/qwen_vl_interface': 9.846981780052005e-06, 'learning_rate/action_model': 9.846981780052003e-05, 'epoch': 0.27}) +[2026-06-17 17:48:51,365][starVLA.training.train_starvla][INFO] - Step 660, Loss: {'train/loss': 1.0426191091537476, 'train/grad_norm_pre_clip': 8.820268630981445, 'timing/data': 0.025577981024980545, 'timing/model': 0.9734905501827598, 'learning_rate/qwen_vl_interface': 9.846436790841955e-06, 'learning_rate/action_model': 9.846436790841955e-05, 'epoch': 0.27}) +[2026-06-17 17:48:55,531][starVLA.training.train_starvla][INFO] - Step 661, Loss: {'train/loss': 0.8681411743164062, 'train/grad_norm_pre_clip': 8.505315780639648, 'timing/data': 0.016737003810703754, 'timing/model': 0.9628684464842081, 'learning_rate/qwen_vl_interface': 9.84589084878561e-06, 'learning_rate/action_model': 9.845890848785611e-05, 'epoch': 0.27}) +[2026-06-17 17:48:59,557][starVLA.training.train_starvla][INFO] - Step 662, Loss: {'train/loss': 0.728276789188385, 'train/grad_norm_pre_clip': 6.239592552185059, 'timing/data': 0.02200312539935112, 'timing/model': 1.182544196024537, 'learning_rate/qwen_vl_interface': 9.84534395399615e-06, 'learning_rate/action_model': 9.84534395399615e-05, 'epoch': 0.27}) +[2026-06-17 17:49:03,593][starVLA.training.train_starvla][INFO] - Step 663, Loss: {'train/loss': 0.8265572786331177, 'train/grad_norm_pre_clip': 6.669013977050781, 'timing/data': 0.01953009981662035, 'timing/model': 1.0015342365950346, 'learning_rate/qwen_vl_interface': 9.844796106586942e-06, 'learning_rate/action_model': 9.844796106586941e-05, 'epoch': 0.27}) +[2026-06-17 17:49:07,514][starVLA.training.train_starvla][INFO] - Step 664, Loss: {'train/loss': 1.0754181146621704, 'train/grad_norm_pre_clip': 10.799076080322266, 'timing/data': 0.025784428231418133, 'timing/model': 0.9705946808680892, 'learning_rate/qwen_vl_interface': 9.844247306671557e-06, 'learning_rate/action_model': 9.844247306671555e-05, 'epoch': 0.27}) +[2026-06-17 17:49:11,570][starVLA.training.train_starvla][INFO] - Step 665, Loss: {'train/loss': 0.8378474116325378, 'train/grad_norm_pre_clip': 19.1083984375, 'timing/data': 0.02424236573278904, 'timing/model': 0.9442652892321348, 'learning_rate/qwen_vl_interface': 9.843697554363762e-06, 'learning_rate/action_model': 9.843697554363762e-05, 'epoch': 0.27}) +[2026-06-17 17:49:15,559][starVLA.training.train_starvla][INFO] - Step 666, Loss: {'train/loss': 0.7103745341300964, 'train/grad_norm_pre_clip': 6.076900005340576, 'timing/data': 0.02909709792584181, 'timing/model': 0.957880743779242, 'learning_rate/qwen_vl_interface': 9.84314684977752e-06, 'learning_rate/action_model': 9.84314684977752e-05, 'epoch': 0.27}) +[2026-06-17 17:49:19,519][starVLA.training.train_starvla][INFO] - Step 667, Loss: {'train/loss': 0.815952479839325, 'train/grad_norm_pre_clip': 6.963968276977539, 'timing/data': 0.017064767889678478, 'timing/model': 1.0409389361739159, 'learning_rate/qwen_vl_interface': 9.842595193026993e-06, 'learning_rate/action_model': 9.842595193026993e-05, 'epoch': 0.27}) +[2026-06-17 17:49:23,439][starVLA.training.train_starvla][INFO] - Step 668, Loss: {'train/loss': 0.6970619559288025, 'train/grad_norm_pre_clip': 5.782436847686768, 'timing/data': 0.02353107463568449, 'timing/model': 1.1854084003716707, 'learning_rate/qwen_vl_interface': 9.842042584226545e-06, 'learning_rate/action_model': 9.842042584226544e-05, 'epoch': 0.27}) +[2026-06-17 17:49:27,579][starVLA.training.train_starvla][INFO] - Step 669, Loss: {'train/loss': 0.7693707346916199, 'train/grad_norm_pre_clip': 6.774971008300781, 'timing/data': 0.024824918247759342, 'timing/model': 1.1320147197693586, 'learning_rate/qwen_vl_interface': 9.841489023490723e-06, 'learning_rate/action_model': 9.841489023490723e-05, 'epoch': 0.28}) +[2026-06-17 17:49:33,932][starVLA.training.train_starvla][INFO] - Step 670, Loss: {'train/loss': 0.7606831789016724, 'train/grad_norm_pre_clip': 5.813389301300049, 'timing/data': 0.022806663997471333, 'timing/model': 1.071772656403482, 'learning_rate/qwen_vl_interface': 9.840934510934288e-06, 'learning_rate/action_model': 9.840934510934288e-05, 'epoch': 0.28}) +[2026-06-17 17:49:38,507][starVLA.training.train_starvla][INFO] - Step 671, Loss: {'train/loss': 0.5817278027534485, 'train/grad_norm_pre_clip': 5.038552761077881, 'timing/data': 0.01758153922855854, 'timing/model': 0.9572969991713762, 'learning_rate/qwen_vl_interface': 9.840379046672189e-06, 'learning_rate/action_model': 9.840379046672189e-05, 'epoch': 0.28}) +[2026-06-17 17:49:43,076][starVLA.training.train_starvla][INFO] - Step 672, Loss: {'train/loss': 0.6711752414703369, 'train/grad_norm_pre_clip': 5.084009647369385, 'timing/data': 0.023044681176543236, 'timing/model': 1.0520061729475856, 'learning_rate/qwen_vl_interface': 9.839822630819573e-06, 'learning_rate/action_model': 9.839822630819572e-05, 'epoch': 0.28}) +[2026-06-17 17:49:47,298][starVLA.training.train_starvla][INFO] - Step 673, Loss: {'train/loss': 0.569383442401886, 'train/grad_norm_pre_clip': 6.206186294555664, 'timing/data': 0.02148047275841236, 'timing/model': 1.1133531127125025, 'learning_rate/qwen_vl_interface': 9.839265263491789e-06, 'learning_rate/action_model': 9.839265263491789e-05, 'epoch': 0.28}) +[2026-06-17 17:49:51,120][starVLA.training.train_starvla][INFO] - Step 674, Loss: {'train/loss': 0.9650801420211792, 'train/grad_norm_pre_clip': 14.243332862854004, 'timing/data': 0.02849606890231371, 'timing/model': 0.9276123382151127, 'learning_rate/qwen_vl_interface': 9.838706944804376e-06, 'learning_rate/action_model': 9.838706944804374e-05, 'epoch': 0.28}) +[2026-06-17 17:49:55,137][starVLA.training.train_starvla][INFO] - Step 675, Loss: {'train/loss': 0.6702191233634949, 'train/grad_norm_pre_clip': 17.875829696655273, 'timing/data': 0.018769551999866962, 'timing/model': 1.157955921255052, 'learning_rate/qwen_vl_interface': 9.838147674873076e-06, 'learning_rate/action_model': 9.838147674873075e-05, 'epoch': 0.28}) +[2026-06-17 17:49:59,221][starVLA.training.train_starvla][INFO] - Step 676, Loss: {'train/loss': 0.9596227407455444, 'train/grad_norm_pre_clip': 8.749773979187012, 'timing/data': 0.02573327347636223, 'timing/model': 1.3589209476485848, 'learning_rate/qwen_vl_interface': 9.837587453813826e-06, 'learning_rate/action_model': 9.837587453813824e-05, 'epoch': 0.28}) +[2026-06-17 17:50:03,031][starVLA.training.train_starvla][INFO] - Step 677, Loss: {'train/loss': 0.7321817874908447, 'train/grad_norm_pre_clip': 7.434157371520996, 'timing/data': 0.022058402188122272, 'timing/model': 0.9609810430556536, 'learning_rate/qwen_vl_interface': 9.83702628174276e-06, 'learning_rate/action_model': 9.837026281742759e-05, 'epoch': 0.28}) +[2026-06-17 17:50:07,108][starVLA.training.train_starvla][INFO] - Step 678, Loss: {'train/loss': 1.0123753547668457, 'train/grad_norm_pre_clip': 8.00418472290039, 'timing/data': 0.021595114842057228, 'timing/model': 1.010993174277246, 'learning_rate/qwen_vl_interface': 9.83646415877621e-06, 'learning_rate/action_model': 9.83646415877621e-05, 'epoch': 0.28}) +[2026-06-17 17:50:11,225][starVLA.training.train_starvla][INFO] - Step 679, Loss: {'train/loss': 0.5870279669761658, 'train/grad_norm_pre_clip': 7.0030517578125, 'timing/data': 0.021374099887907505, 'timing/model': 1.0807335739955306, 'learning_rate/qwen_vl_interface': 9.835901085030703e-06, 'learning_rate/action_model': 9.835901085030703e-05, 'epoch': 0.28}) +[2026-06-17 17:50:15,816][starVLA.training.train_starvla][INFO] - Step 680, Loss: {'train/loss': 0.7301754355430603, 'train/grad_norm_pre_clip': 6.600776672363281, 'timing/data': 0.0207514651119709, 'timing/model': 1.426421475596726, 'learning_rate/qwen_vl_interface': 9.83533706062297e-06, 'learning_rate/action_model': 9.835337060622969e-05, 'epoch': 0.28}) +[2026-06-17 17:50:19,701][starVLA.training.train_starvla][INFO] - Step 681, Loss: {'train/loss': 0.7303656339645386, 'train/grad_norm_pre_clip': 8.96471881866455, 'timing/data': 0.02664982620626688, 'timing/model': 0.9270513458177447, 'learning_rate/qwen_vl_interface': 9.834772085669927e-06, 'learning_rate/action_model': 9.834772085669927e-05, 'epoch': 0.28}) +[2026-06-17 17:50:23,795][starVLA.training.train_starvla][INFO] - Step 682, Loss: {'train/loss': 0.5922955870628357, 'train/grad_norm_pre_clip': 8.601513862609863, 'timing/data': 0.018437687307596207, 'timing/model': 1.1417437111958861, 'learning_rate/qwen_vl_interface': 9.834206160288698e-06, 'learning_rate/action_model': 9.834206160288698e-05, 'epoch': 0.28}) +[2026-06-17 17:50:27,503][starVLA.training.train_starvla][INFO] - Step 683, Loss: {'train/loss': 0.7114226818084717, 'train/grad_norm_pre_clip': 6.042293071746826, 'timing/data': 0.026887964457273483, 'timing/model': 0.9442754350602627, 'learning_rate/qwen_vl_interface': 9.8336392845966e-06, 'learning_rate/action_model': 9.833639284596599e-05, 'epoch': 0.28}) +[2026-06-17 17:50:31,453][starVLA.training.train_starvla][INFO] - Step 684, Loss: {'train/loss': 0.7718777060508728, 'train/grad_norm_pre_clip': 10.22819709777832, 'timing/data': 0.02157184947282076, 'timing/model': 0.9432180756703019, 'learning_rate/qwen_vl_interface': 9.833071458711146e-06, 'learning_rate/action_model': 9.833071458711146e-05, 'epoch': 0.28}) +[2026-06-17 17:50:35,509][starVLA.training.train_starvla][INFO] - Step 685, Loss: {'train/loss': 0.8667235374450684, 'train/grad_norm_pre_clip': 9.050042152404785, 'timing/data': 0.022791066206991673, 'timing/model': 1.0979244569316506, 'learning_rate/qwen_vl_interface': 9.832502682750048e-06, 'learning_rate/action_model': 9.832502682750047e-05, 'epoch': 0.28}) +[2026-06-17 17:50:39,606][starVLA.training.train_starvla][INFO] - Step 686, Loss: {'train/loss': 0.6038234233856201, 'train/grad_norm_pre_clip': 6.397466659545898, 'timing/data': 0.022785911336541176, 'timing/model': 1.0277169523760676, 'learning_rate/qwen_vl_interface': 9.831932956831212e-06, 'learning_rate/action_model': 9.831932956831211e-05, 'epoch': 0.28}) +[2026-06-17 17:50:43,515][starVLA.training.train_starvla][INFO] - Step 687, Loss: {'train/loss': 0.30020883679389954, 'train/grad_norm_pre_clip': 4.274683475494385, 'timing/data': 0.016620347276329994, 'timing/model': 0.960429978556931, 'learning_rate/qwen_vl_interface': 9.831362281072743e-06, 'learning_rate/action_model': 9.831362281072742e-05, 'epoch': 0.28}) +[2026-06-17 17:50:47,360][starVLA.training.train_starvla][INFO] - Step 688, Loss: {'train/loss': 1.077029824256897, 'train/grad_norm_pre_clip': 11.547016143798828, 'timing/data': 0.02244698815047741, 'timing/model': 1.1466520112007856, 'learning_rate/qwen_vl_interface': 9.830790655592944e-06, 'learning_rate/action_model': 9.830790655592943e-05, 'epoch': 0.28}) +[2026-06-17 17:50:51,180][starVLA.training.train_starvla][INFO] - Step 689, Loss: {'train/loss': 0.6132045388221741, 'train/grad_norm_pre_clip': 7.48215389251709, 'timing/data': 0.016498793847858906, 'timing/model': 1.1137572834268212, 'learning_rate/qwen_vl_interface': 9.830218080510313e-06, 'learning_rate/action_model': 9.830218080510312e-05, 'epoch': 0.28}) +[2026-06-17 17:50:54,990][starVLA.training.train_starvla][INFO] - Step 690, Loss: {'train/loss': 0.6856497526168823, 'train/grad_norm_pre_clip': 8.195115089416504, 'timing/data': 0.022237582132220268, 'timing/model': 1.043091950006783, 'learning_rate/qwen_vl_interface': 9.829644555943545e-06, 'learning_rate/action_model': 9.829644555943545e-05, 'epoch': 0.28}) +[2026-06-17 17:50:59,219][starVLA.training.train_starvla][INFO] - Step 691, Loss: {'train/loss': 0.7681445479393005, 'train/grad_norm_pre_clip': 8.333395957946777, 'timing/data': 0.026742253452539444, 'timing/model': 1.0606579668819904, 'learning_rate/qwen_vl_interface': 9.82907008201153e-06, 'learning_rate/action_model': 9.829070082011531e-05, 'epoch': 0.28}) +[2026-06-17 17:51:03,156][starVLA.training.train_starvla][INFO] - Step 692, Loss: {'train/loss': 0.788176953792572, 'train/grad_norm_pre_clip': 6.410250663757324, 'timing/data': 0.019029706716537476, 'timing/model': 0.9451778503134847, 'learning_rate/qwen_vl_interface': 9.828494658833366e-06, 'learning_rate/action_model': 9.828494658833365e-05, 'epoch': 0.28}) +[2026-06-17 17:51:09,628][starVLA.training.train_starvla][INFO] - Step 693, Loss: {'train/loss': 0.4452824592590332, 'train/grad_norm_pre_clip': 13.028225898742676, 'timing/data': 0.028079467825591564, 'timing/model': 1.1012198738753796, 'learning_rate/qwen_vl_interface': 9.82791828652833e-06, 'learning_rate/action_model': 9.827918286528329e-05, 'epoch': 0.29}) +[2026-06-17 17:51:15,584][starVLA.training.train_starvla][INFO] - Step 694, Loss: {'train/loss': 0.903874933719635, 'train/grad_norm_pre_clip': 9.95539379119873, 'timing/data': 0.02371141780167818, 'timing/model': 1.0149772502481937, 'learning_rate/qwen_vl_interface': 9.827340965215903e-06, 'learning_rate/action_model': 9.827340965215904e-05, 'epoch': 0.29}) +[2026-06-17 17:51:19,840][starVLA.training.train_starvla][INFO] - Step 695, Loss: {'train/loss': 0.8370166420936584, 'train/grad_norm_pre_clip': 60.24484634399414, 'timing/data': 0.020042676478624344, 'timing/model': 1.3429699875414371, 'learning_rate/qwen_vl_interface': 9.826762695015773e-06, 'learning_rate/action_model': 9.826762695015772e-05, 'epoch': 0.29}) +[2026-06-17 17:51:25,689][starVLA.training.train_starvla][INFO] - Step 696, Loss: {'train/loss': 1.140477180480957, 'train/grad_norm_pre_clip': 10.096115112304688, 'timing/data': 0.017173120751976967, 'timing/model': 2.5825135177001357, 'learning_rate/qwen_vl_interface': 9.82618347604781e-06, 'learning_rate/action_model': 9.82618347604781e-05, 'epoch': 0.29}) +[2026-06-17 17:51:30,508][starVLA.training.train_starvla][INFO] - Step 697, Loss: {'train/loss': 0.9684523940086365, 'train/grad_norm_pre_clip': 8.814656257629395, 'timing/data': 0.018091331236064434, 'timing/model': 0.9665102884173393, 'learning_rate/qwen_vl_interface': 9.825603308432089e-06, 'learning_rate/action_model': 9.825603308432088e-05, 'epoch': 0.29}) +[2026-06-17 17:51:34,532][starVLA.training.train_starvla][INFO] - Step 698, Loss: {'train/loss': 0.8826870918273926, 'train/grad_norm_pre_clip': 11.302913665771484, 'timing/data': 0.021491246297955513, 'timing/model': 0.9705184595659375, 'learning_rate/qwen_vl_interface': 9.825022192288878e-06, 'learning_rate/action_model': 9.825022192288877e-05, 'epoch': 0.29}) +[2026-06-17 17:51:38,425][starVLA.training.train_starvla][INFO] - Step 699, Loss: {'train/loss': 1.1296615600585938, 'train/grad_norm_pre_clip': 12.604227066040039, 'timing/data': 0.017864982597529888, 'timing/model': 0.9627321986481547, 'learning_rate/qwen_vl_interface': 9.824440127738643e-06, 'learning_rate/action_model': 9.824440127738642e-05, 'epoch': 0.29}) +[2026-06-17 17:53:34,767][starVLA.training.train_starvla][INFO] - Step 700, Loss: {'train/loss': 0.7686453461647034, 'train/grad_norm_pre_clip': 7.8849873542785645, 'eval/loss': 0.8005172610282898, 'eval/latency_6/loss': 0.8005172610282898, 'eval/demon_fire_precision': 0.9270603094411115, 'eval/demon_fire_recall': 0.9489334195216548, 'eval/demon_fire_f1': 0.9378693499440983, 'eval/demon_fire_rate_model': 0.8811908736783528, 'eval/demon_fire_rate_teacher': 0.860879243183083, 'eval/demon_left_precision': 0.5976095617529881, 'eval/demon_left_recall': 0.5263157894736842, 'eval/demon_left_f1': 0.5597014925373134, 'eval/demon_left_rate_model': 0.13967723984418476, 'eval/demon_left_rate_teacher': 0.15859766277128548, 'eval/demon_right_precision': 0.7454128440366973, 'eval/demon_right_recall': 0.49923195084485406, 'eval/demon_right_f1': 0.5979760809567617, 'eval/demon_right_rate_model': 0.1213132999443517, 'eval/demon_right_rate_teacher': 0.18113522537562604, 'eval/demon_move_f1': 0.5788387867470375, 'eval/demon_cc_f1': 0.7907315900321849, 'eval/latency_6/demon_fire_precision': 0.9270603094411115, 'eval/latency_6/demon_fire_recall': 0.9489334195216548, 'eval/latency_6/demon_fire_f1': 0.9378693499440983, 'eval/latency_6/demon_fire_rate_model': 0.8811908736783528, 'eval/latency_6/demon_fire_rate_teacher': 0.860879243183083, 'eval/latency_6/demon_left_precision': 0.5976095617529881, 'eval/latency_6/demon_left_recall': 0.5263157894736842, 'eval/latency_6/demon_left_f1': 0.5597014925373134, 'eval/latency_6/demon_left_rate_model': 0.13967723984418476, 'eval/latency_6/demon_left_rate_teacher': 0.15859766277128548, 'eval/latency_6/demon_right_precision': 0.7454128440366973, 'eval/latency_6/demon_right_recall': 0.49923195084485406, 'eval/latency_6/demon_right_f1': 0.5979760809567617, 'eval/latency_6/demon_right_rate_model': 0.1213132999443517, 'eval/latency_6/demon_right_rate_teacher': 0.18113522537562604, 'eval/latency_6/demon_move_f1': 0.5788387867470375, 'eval/latency_6/demon_cc_f1': 0.7907315900321849, 'timing/data': 0.0274173216894269, 'timing/model': 0.9536214675754309, 'learning_rate/qwen_vl_interface': 9.823857114902047e-06, 'learning_rate/action_model': 9.823857114902047e-05, 'epoch': 0.29}) +[2026-06-17 17:53:34,770][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_700_state +[2026-06-17 17:53:34,776][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 17:53:41,594][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_700_state/model.safetensors +[2026-06-17 17:53:57,768][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_700_state/optimizer.bin +[2026-06-17 17:53:57,768][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_700_state/sampler.bin +[2026-06-17 17:53:57,769][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_700_state/sampler_1.bin +[2026-06-17 17:53:57,772][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_700_state/random_states_0.pkl +[2026-06-17 17:53:57,790][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 700 +[2026-06-17 17:53:57,791][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 17:53:57,802][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 17:54:01,943][starVLA.training.train_starvla][INFO] - Step 701, Loss: {'train/loss': 0.8359478712081909, 'train/grad_norm_pre_clip': 9.407339096069336, 'timing/data': 0.021204261109232903, 'timing/model': 1.1165848737582564, 'learning_rate/qwen_vl_interface': 9.823273153899951e-06, 'learning_rate/action_model': 9.82327315389995e-05, 'epoch': 0.29}) +[2026-06-17 17:54:05,860][starVLA.training.train_starvla][INFO] - Step 702, Loss: {'train/loss': 0.8485422730445862, 'train/grad_norm_pre_clip': 9.889839172363281, 'timing/data': 0.019435079768300056, 'timing/model': 0.9448810257017612, 'learning_rate/qwen_vl_interface': 9.822688244853407e-06, 'learning_rate/action_model': 9.822688244853408e-05, 'epoch': 0.29}) +[2026-06-17 17:54:10,024][starVLA.training.train_starvla][INFO] - Step 703, Loss: {'train/loss': 0.9787039160728455, 'train/grad_norm_pre_clip': 14.350738525390625, 'timing/data': 0.01765655167400837, 'timing/model': 1.0241569047793746, 'learning_rate/qwen_vl_interface': 9.822102387883672e-06, 'learning_rate/action_model': 9.822102387883671e-05, 'epoch': 0.29}) +[2026-06-17 17:54:13,889][starVLA.training.train_starvla][INFO] - Step 704, Loss: {'train/loss': 0.6795839667320251, 'train/grad_norm_pre_clip': 8.667576789855957, 'timing/data': 0.015299838967621326, 'timing/model': 1.1602199263870716, 'learning_rate/qwen_vl_interface': 9.82151558311219e-06, 'learning_rate/action_model': 9.821515583112191e-05, 'epoch': 0.29}) +[2026-06-17 17:54:17,780][starVLA.training.train_starvla][INFO] - Step 705, Loss: {'train/loss': 0.785545289516449, 'train/grad_norm_pre_clip': 12.050324440002441, 'timing/data': 0.019957300275564194, 'timing/model': 1.1588589623570442, 'learning_rate/qwen_vl_interface': 9.82092783066061e-06, 'learning_rate/action_model': 9.82092783066061e-05, 'epoch': 0.29}) +[2026-06-17 17:54:21,972][starVLA.training.train_starvla][INFO] - Step 706, Loss: {'train/loss': 1.0256038904190063, 'train/grad_norm_pre_clip': 12.483625411987305, 'timing/data': 0.01881353836506605, 'timing/model': 1.3185672890394926, 'learning_rate/qwen_vl_interface': 9.820339130650771e-06, 'learning_rate/action_model': 9.820339130650771e-05, 'epoch': 0.29}) +[2026-06-17 17:54:25,902][starVLA.training.train_starvla][INFO] - Step 707, Loss: {'train/loss': 0.9385353326797485, 'train/grad_norm_pre_clip': 10.267094612121582, 'timing/data': 0.018901439383625984, 'timing/model': 0.930744232609868, 'learning_rate/qwen_vl_interface': 9.819749483204715e-06, 'learning_rate/action_model': 9.819749483204715e-05, 'epoch': 0.29}) +[2026-06-17 17:54:29,788][starVLA.training.train_starvla][INFO] - Step 708, Loss: {'train/loss': 0.6004462242126465, 'train/grad_norm_pre_clip': 7.385281085968018, 'timing/data': 0.017350505106151104, 'timing/model': 0.9312659846618772, 'learning_rate/qwen_vl_interface': 9.81915888844467e-06, 'learning_rate/action_model': 9.81915888844467e-05, 'epoch': 0.29}) +[2026-06-17 17:54:33,607][starVLA.training.train_starvla][INFO] - Step 709, Loss: {'train/loss': 1.061552882194519, 'train/grad_norm_pre_clip': 8.423709869384766, 'timing/data': 0.02307797595858574, 'timing/model': 0.9906974770128727, 'learning_rate/qwen_vl_interface': 9.818567346493072e-06, 'learning_rate/action_model': 9.818567346493071e-05, 'epoch': 0.29}) +[2026-06-17 17:54:38,006][starVLA.training.train_starvla][INFO] - Step 710, Loss: {'train/loss': 0.6692125797271729, 'train/grad_norm_pre_clip': 8.69096851348877, 'timing/data': 0.020035856403410435, 'timing/model': 0.9318494768813252, 'learning_rate/qwen_vl_interface': 9.817974857472549e-06, 'learning_rate/action_model': 9.817974857472548e-05, 'epoch': 0.29}) +[2026-06-17 17:54:41,829][starVLA.training.train_starvla][INFO] - Step 711, Loss: {'train/loss': 0.9047051668167114, 'train/grad_norm_pre_clip': 17.637046813964844, 'timing/data': 0.01777276862412691, 'timing/model': 1.1406919164583087, 'learning_rate/qwen_vl_interface': 9.817381421505922e-06, 'learning_rate/action_model': 9.817381421505921e-05, 'epoch': 0.29}) +[2026-06-17 17:54:45,750][starVLA.training.train_starvla][INFO] - Step 712, Loss: {'train/loss': 0.783793568611145, 'train/grad_norm_pre_clip': 8.791497230529785, 'timing/data': 0.01923431921750307, 'timing/model': 1.20509485155344, 'learning_rate/qwen_vl_interface': 9.81678703871621e-06, 'learning_rate/action_model': 9.81678703871621e-05, 'epoch': 0.29}) +[2026-06-17 17:54:49,489][starVLA.training.train_starvla][INFO] - Step 713, Loss: {'train/loss': 0.8198676109313965, 'train/grad_norm_pre_clip': 9.23483943939209, 'timing/data': 0.0192193491384387, 'timing/model': 0.9579034997150302, 'learning_rate/qwen_vl_interface': 9.816191709226631e-06, 'learning_rate/action_model': 9.816191709226632e-05, 'epoch': 0.29}) +[2026-06-17 17:54:53,374][starVLA.training.train_starvla][INFO] - Step 714, Loss: {'train/loss': 0.847575843334198, 'train/grad_norm_pre_clip': 7.117984294891357, 'timing/data': 0.016591940075159073, 'timing/model': 0.9627146050333977, 'learning_rate/qwen_vl_interface': 9.815595433160599e-06, 'learning_rate/action_model': 9.815595433160598e-05, 'epoch': 0.29}) +[2026-06-17 17:54:57,381][starVLA.training.train_starvla][INFO] - Step 715, Loss: {'train/loss': 0.6972998380661011, 'train/grad_norm_pre_clip': 6.776851177215576, 'timing/data': 0.025817804969847202, 'timing/model': 0.9442881653085351, 'learning_rate/qwen_vl_interface': 9.81499821064172e-06, 'learning_rate/action_model': 9.81499821064172e-05, 'epoch': 0.29}) +[2026-06-17 17:55:01,220][starVLA.training.train_starvla][INFO] - Step 716, Loss: {'train/loss': 0.7358497977256775, 'train/grad_norm_pre_clip': 7.080821990966797, 'timing/data': 0.026107548736035824, 'timing/model': 0.9277843963354826, 'learning_rate/qwen_vl_interface': 9.814400041793799e-06, 'learning_rate/action_model': 9.814400041793798e-05, 'epoch': 0.29}) +[2026-06-17 17:55:05,399][starVLA.training.train_starvla][INFO] - Step 717, Loss: {'train/loss': 0.5915205478668213, 'train/grad_norm_pre_clip': 5.351985454559326, 'timing/data': 0.016247105784714222, 'timing/model': 1.1320327296853065, 'learning_rate/qwen_vl_interface': 9.813800926740839e-06, 'learning_rate/action_model': 9.813800926740838e-05, 'epoch': 0.3}) +[2026-06-17 17:55:09,378][starVLA.training.train_starvla][INFO] - Step 718, Loss: {'train/loss': 0.7858110666275024, 'train/grad_norm_pre_clip': 7.872401714324951, 'timing/data': 0.013628337532281876, 'timing/model': 1.2067965790629387, 'learning_rate/qwen_vl_interface': 9.813200865607034e-06, 'learning_rate/action_model': 9.813200865607034e-05, 'epoch': 0.3}) +[2026-06-17 17:55:13,067][starVLA.training.train_starvla][INFO] - Step 719, Loss: {'train/loss': 0.6577262878417969, 'train/grad_norm_pre_clip': 9.209912300109863, 'timing/data': 0.02004153188318014, 'timing/model': 0.9696348821744323, 'learning_rate/qwen_vl_interface': 9.812599858516782e-06, 'learning_rate/action_model': 9.812599858516782e-05, 'epoch': 0.3}) +[2026-06-17 17:55:16,974][starVLA.training.train_starvla][INFO] - Step 720, Loss: {'train/loss': 0.6272037625312805, 'train/grad_norm_pre_clip': 7.779623508453369, 'timing/data': 0.022882173769176006, 'timing/model': 0.9603934735059738, 'learning_rate/qwen_vl_interface': 9.81199790559467e-06, 'learning_rate/action_model': 9.811997905594668e-05, 'epoch': 0.3}) +[2026-06-17 17:55:20,803][starVLA.training.train_starvla][INFO] - Step 721, Loss: {'train/loss': 0.8499436974525452, 'train/grad_norm_pre_clip': 7.618723392486572, 'timing/data': 0.015274506993591785, 'timing/model': 0.9362543169409037, 'learning_rate/qwen_vl_interface': 9.811395006965481e-06, 'learning_rate/action_model': 9.81139500696548e-05, 'epoch': 0.3}) +[2026-06-17 17:55:24,756][starVLA.training.train_starvla][INFO] - Step 722, Loss: {'train/loss': 0.9177958965301514, 'train/grad_norm_pre_clip': 11.488932609558105, 'timing/data': 0.020681499503552914, 'timing/model': 0.9337323810905218, 'learning_rate/qwen_vl_interface': 9.810791162754199e-06, 'learning_rate/action_model': 9.810791162754198e-05, 'epoch': 0.3}) +[2026-06-17 17:55:28,811][starVLA.training.train_starvla][INFO] - Step 723, Loss: {'train/loss': 0.7979747653007507, 'train/grad_norm_pre_clip': 7.062660217285156, 'timing/data': 0.023967484943568707, 'timing/model': 0.9352356158196926, 'learning_rate/qwen_vl_interface': 9.810186373086002e-06, 'learning_rate/action_model': 9.810186373086002e-05, 'epoch': 0.3}) +[2026-06-17 17:55:32,741][starVLA.training.train_starvla][INFO] - Step 724, Loss: {'train/loss': 0.7701035737991333, 'train/grad_norm_pre_clip': 7.0082316398620605, 'timing/data': 0.021979606710374355, 'timing/model': 1.1781239286065102, 'learning_rate/qwen_vl_interface': 9.809580638086264e-06, 'learning_rate/action_model': 9.809580638086263e-05, 'epoch': 0.3}) +[2026-06-17 17:55:36,650][starVLA.training.train_starvla][INFO] - Step 725, Loss: {'train/loss': 0.7599844336509705, 'train/grad_norm_pre_clip': 7.757591247558594, 'timing/data': 0.02420433610677719, 'timing/model': 1.1329303160309792, 'learning_rate/qwen_vl_interface': 9.80897395788055e-06, 'learning_rate/action_model': 9.808973957880551e-05, 'epoch': 0.3}) +[2026-06-17 17:55:40,301][starVLA.training.train_starvla][INFO] - Step 726, Loss: {'train/loss': 0.9733573794364929, 'train/grad_norm_pre_clip': 9.559284210205078, 'timing/data': 0.023398594930768013, 'timing/model': 0.9640273619443178, 'learning_rate/qwen_vl_interface': 9.808366332594632e-06, 'learning_rate/action_model': 9.808366332594632e-05, 'epoch': 0.3}) +[2026-06-17 17:55:44,157][starVLA.training.train_starvla][INFO] - Step 727, Loss: {'train/loss': 0.9852672219276428, 'train/grad_norm_pre_clip': 7.494624614715576, 'timing/data': 0.019365468062460423, 'timing/model': 0.9295125128701329, 'learning_rate/qwen_vl_interface': 9.807757762354467e-06, 'learning_rate/action_model': 9.807757762354467e-05, 'epoch': 0.3}) +[2026-06-17 17:55:48,230][starVLA.training.train_starvla][INFO] - Step 728, Loss: {'train/loss': 0.8974481821060181, 'train/grad_norm_pre_clip': 10.512076377868652, 'timing/data': 0.021532424725592136, 'timing/model': 0.9287157915532589, 'learning_rate/qwen_vl_interface': 9.807148247286215e-06, 'learning_rate/action_model': 9.807148247286215e-05, 'epoch': 0.3}) +[2026-06-17 17:55:52,107][starVLA.training.train_starvla][INFO] - Step 729, Loss: {'train/loss': 0.7530903220176697, 'train/grad_norm_pre_clip': 8.455613136291504, 'timing/data': 0.016922272741794586, 'timing/model': 0.9447837295010686, 'learning_rate/qwen_vl_interface': 9.806537787516226e-06, 'learning_rate/action_model': 9.806537787516226e-05, 'epoch': 0.3}) +[2026-06-17 17:55:56,149][starVLA.training.train_starvla][INFO] - Step 730, Loss: {'train/loss': 1.2495471239089966, 'train/grad_norm_pre_clip': 10.179099082946777, 'timing/data': 0.03750938642770052, 'timing/model': 0.9436046089977026, 'learning_rate/qwen_vl_interface': 9.80592638317105e-06, 'learning_rate/action_model': 9.80592638317105e-05, 'epoch': 0.3}) +[2026-06-17 17:56:00,270][starVLA.training.train_starvla][INFO] - Step 731, Loss: {'train/loss': 0.7869071960449219, 'train/grad_norm_pre_clip': 8.401729583740234, 'timing/data': 0.018449874594807625, 'timing/model': 1.1941032968461514, 'learning_rate/qwen_vl_interface': 9.805314034377434e-06, 'learning_rate/action_model': 9.805314034377434e-05, 'epoch': 0.3}) +[2026-06-17 17:56:04,109][starVLA.training.train_starvla][INFO] - Step 732, Loss: {'train/loss': 0.7103153467178345, 'train/grad_norm_pre_clip': 6.42212438583374, 'timing/data': 0.015243546105921268, 'timing/model': 1.1621796311810613, 'learning_rate/qwen_vl_interface': 9.804700741262318e-06, 'learning_rate/action_model': 9.804700741262317e-05, 'epoch': 0.3}) +[2026-06-17 17:56:08,920][starVLA.training.train_starvla][INFO] - Step 733, Loss: {'train/loss': 0.8750436902046204, 'train/grad_norm_pre_clip': 7.908463001251221, 'timing/data': 0.026263357140123844, 'timing/model': 0.9506219550967216, 'learning_rate/qwen_vl_interface': 9.804086503952835e-06, 'learning_rate/action_model': 9.804086503952835e-05, 'epoch': 0.3}) +[2026-06-17 17:56:13,157][starVLA.training.train_starvla][INFO] - Step 734, Loss: {'train/loss': 0.6896485686302185, 'train/grad_norm_pre_clip': 7.977433681488037, 'timing/data': 0.017049038782715797, 'timing/model': 0.9421028876677155, 'learning_rate/qwen_vl_interface': 9.803471322576319e-06, 'learning_rate/action_model': 9.803471322576319e-05, 'epoch': 0.3}) +[2026-06-17 17:56:17,293][starVLA.training.train_starvla][INFO] - Step 735, Loss: {'train/loss': 0.6541957855224609, 'train/grad_norm_pre_clip': 6.5452399253845215, 'timing/data': 0.018233228474855423, 'timing/model': 0.9291010098531842, 'learning_rate/qwen_vl_interface': 9.8028551972603e-06, 'learning_rate/action_model': 9.802855197260299e-05, 'epoch': 0.3}) +[2026-06-17 17:56:21,115][starVLA.training.train_starvla][INFO] - Step 736, Loss: {'train/loss': 0.6158105134963989, 'train/grad_norm_pre_clip': 7.606716632843018, 'timing/data': 0.023712921887636185, 'timing/model': 0.9259667191654444, 'learning_rate/qwen_vl_interface': 9.802238128132498e-06, 'learning_rate/action_model': 9.802238128132498e-05, 'epoch': 0.3}) +[2026-06-17 17:56:24,970][starVLA.training.train_starvla][INFO] - Step 737, Loss: {'train/loss': 0.7570365071296692, 'train/grad_norm_pre_clip': 29.557600021362305, 'timing/data': 0.02307156752794981, 'timing/model': 0.9322131024673581, 'learning_rate/qwen_vl_interface': 9.801620115320836e-06, 'learning_rate/action_model': 9.801620115320835e-05, 'epoch': 0.3}) +[2026-06-17 17:56:28,827][starVLA.training.train_starvla][INFO] - Step 738, Loss: {'train/loss': 0.9471485614776611, 'train/grad_norm_pre_clip': 6.862945556640625, 'timing/data': 0.01923621352761984, 'timing/model': 1.1530671687796712, 'learning_rate/qwen_vl_interface': 9.801001158953425e-06, 'learning_rate/action_model': 9.801001158953423e-05, 'epoch': 0.3}) +[2026-06-17 17:56:32,889][starVLA.training.train_starvla][INFO] - Step 739, Loss: {'train/loss': 0.652957022190094, 'train/grad_norm_pre_clip': 7.212001323699951, 'timing/data': 0.0208056028932333, 'timing/model': 1.1998503832146525, 'learning_rate/qwen_vl_interface': 9.800381259158575e-06, 'learning_rate/action_model': 9.800381259158574e-05, 'epoch': 0.3}) +[2026-06-17 17:56:36,490][starVLA.training.train_starvla][INFO] - Step 740, Loss: {'train/loss': 0.982516348361969, 'train/grad_norm_pre_clip': 8.936702728271484, 'timing/data': 0.02349445130676031, 'timing/model': 0.9313807636499405, 'learning_rate/qwen_vl_interface': 9.799760416064794e-06, 'learning_rate/action_model': 9.799760416064793e-05, 'epoch': 0.3}) +[2026-06-17 17:56:40,625][starVLA.training.train_starvla][INFO] - Step 741, Loss: {'train/loss': 0.6298196315765381, 'train/grad_norm_pre_clip': 6.095052719116211, 'timing/data': 0.018738589249551296, 'timing/model': 0.9450233327224851, 'learning_rate/qwen_vl_interface': 9.799138629800785e-06, 'learning_rate/action_model': 9.799138629800785e-05, 'epoch': 0.31}) +[2026-06-17 17:56:44,480][starVLA.training.train_starvla][INFO] - Step 742, Loss: {'train/loss': 0.7178869843482971, 'train/grad_norm_pre_clip': 6.934845924377441, 'timing/data': 0.01942856702953577, 'timing/model': 0.9256621981039643, 'learning_rate/qwen_vl_interface': 9.798515900495441e-06, 'learning_rate/action_model': 9.79851590049544e-05, 'epoch': 0.31}) +[2026-06-17 17:56:48,376][starVLA.training.train_starvla][INFO] - Step 743, Loss: {'train/loss': 0.7823391556739807, 'train/grad_norm_pre_clip': 7.995687484741211, 'timing/data': 0.01779443770647049, 'timing/model': 0.943733905442059, 'learning_rate/qwen_vl_interface': 9.797892228277857e-06, 'learning_rate/action_model': 9.797892228277857e-05, 'epoch': 0.31}) +[2026-06-17 17:56:52,447][starVLA.training.train_starvla][INFO] - Step 744, Loss: {'train/loss': 0.816717267036438, 'train/grad_norm_pre_clip': 7.602428436279297, 'timing/data': 0.024595752358436584, 'timing/model': 1.1873239381238818, 'learning_rate/qwen_vl_interface': 9.797267613277319e-06, 'learning_rate/action_model': 9.797267613277319e-05, 'epoch': 0.31}) +[2026-06-17 17:56:56,227][starVLA.training.train_starvla][INFO] - Step 745, Loss: {'train/loss': 0.8780652284622192, 'train/grad_norm_pre_clip': 8.46386432647705, 'timing/data': 0.019821345806121826, 'timing/model': 1.0756338136270642, 'learning_rate/qwen_vl_interface': 9.796642055623314e-06, 'learning_rate/action_model': 9.796642055623314e-05, 'epoch': 0.31}) +[2026-06-17 17:56:59,918][starVLA.training.train_starvla][INFO] - Step 746, Loss: {'train/loss': 0.4804147183895111, 'train/grad_norm_pre_clip': 7.714803695678711, 'timing/data': 0.026030697859823704, 'timing/model': 0.9766791304573417, 'learning_rate/qwen_vl_interface': 9.796015555445518e-06, 'learning_rate/action_model': 9.796015555445517e-05, 'epoch': 0.31}) +[2026-06-17 17:57:03,811][starVLA.training.train_starvla][INFO] - Step 747, Loss: {'train/loss': 0.5354642271995544, 'train/grad_norm_pre_clip': 7.847177028656006, 'timing/data': 0.025682099163532257, 'timing/model': 0.9308931538835168, 'learning_rate/qwen_vl_interface': 9.795388112873806e-06, 'learning_rate/action_model': 9.795388112873806e-05, 'epoch': 0.31}) +[2026-06-17 17:57:07,732][starVLA.training.train_starvla][INFO] - Step 748, Loss: {'train/loss': 0.7315133213996887, 'train/grad_norm_pre_clip': 8.1111478805542, 'timing/data': 0.018211319111287594, 'timing/model': 0.9833870269358158, 'learning_rate/qwen_vl_interface': 9.794759728038246e-06, 'learning_rate/action_model': 9.794759728038246e-05, 'epoch': 0.31}) +[2026-06-17 17:57:12,305][starVLA.training.train_starvla][INFO] - Step 749, Loss: {'train/loss': 0.6091644167900085, 'train/grad_norm_pre_clip': 20.293663024902344, 'timing/data': 0.019490284845232964, 'timing/model': 0.9369393233209848, 'learning_rate/qwen_vl_interface': 9.794130401069105e-06, 'learning_rate/action_model': 9.794130401069104e-05, 'epoch': 0.31}) +[2026-06-17 17:57:16,114][starVLA.training.train_starvla][INFO] - Step 750, Loss: {'train/loss': 0.8149057030677795, 'train/grad_norm_pre_clip': 9.861868858337402, 'timing/data': 0.01826859451830387, 'timing/model': 0.9314732076600194, 'learning_rate/qwen_vl_interface': 9.79350013209684e-06, 'learning_rate/action_model': 9.79350013209684e-05, 'epoch': 0.31}) +[2026-06-17 17:57:20,087][starVLA.training.train_starvla][INFO] - Step 751, Loss: {'train/loss': 0.8888570070266724, 'train/grad_norm_pre_clip': 10.742637634277344, 'timing/data': 0.022989606484770775, 'timing/model': 1.2238206332549453, 'learning_rate/qwen_vl_interface': 9.792868921252112e-06, 'learning_rate/action_model': 9.792868921252111e-05, 'epoch': 0.31}) +[2026-06-17 17:57:23,713][starVLA.training.train_starvla][INFO] - Step 752, Loss: {'train/loss': 0.7305892705917358, 'train/grad_norm_pre_clip': 9.02098560333252, 'timing/data': 0.0170593848451972, 'timing/model': 0.9312734268605709, 'learning_rate/qwen_vl_interface': 9.792236768665765e-06, 'learning_rate/action_model': 9.792236768665764e-05, 'epoch': 0.31}) +[2026-06-17 17:57:27,653][starVLA.training.train_starvla][INFO] - Step 753, Loss: {'train/loss': 0.9050251245498657, 'train/grad_norm_pre_clip': 10.947958946228027, 'timing/data': 0.01771292556077242, 'timing/model': 1.031632793135941, 'learning_rate/qwen_vl_interface': 9.79160367446885e-06, 'learning_rate/action_model': 9.79160367446885e-05, 'epoch': 0.31}) +[2026-06-17 17:57:31,854][starVLA.training.train_starvla][INFO] - Step 754, Loss: {'train/loss': 0.7850938439369202, 'train/grad_norm_pre_clip': 7.582993984222412, 'timing/data': 0.024236783385276794, 'timing/model': 1.0603209203109145, 'learning_rate/qwen_vl_interface': 9.790969638792606e-06, 'learning_rate/action_model': 9.790969638792606e-05, 'epoch': 0.31}) +[2026-06-17 17:57:35,675][starVLA.training.train_starvla][INFO] - Step 755, Loss: {'train/loss': 0.8320557475090027, 'train/grad_norm_pre_clip': 7.444880485534668, 'timing/data': 0.018541913479566574, 'timing/model': 0.936743394471705, 'learning_rate/qwen_vl_interface': 9.790334661768471e-06, 'learning_rate/action_model': 9.79033466176847e-05, 'epoch': 0.31}) +[2026-06-17 17:57:39,560][starVLA.training.train_starvla][INFO] - Step 756, Loss: {'train/loss': 0.7801932096481323, 'train/grad_norm_pre_clip': 8.56041431427002, 'timing/data': 0.023720919154584408, 'timing/model': 0.9361003618687391, 'learning_rate/qwen_vl_interface': 9.78969874352807e-06, 'learning_rate/action_model': 9.78969874352807e-05, 'epoch': 0.31}) +[2026-06-17 17:57:43,411][starVLA.training.train_starvla][INFO] - Step 757, Loss: {'train/loss': 0.8385761380195618, 'train/grad_norm_pre_clip': 8.819290161132812, 'timing/data': 0.01580975390970707, 'timing/model': 1.1670092195272446, 'learning_rate/qwen_vl_interface': 9.78906188420324e-06, 'learning_rate/action_model': 9.78906188420324e-05, 'epoch': 0.31}) +[2026-06-17 17:57:47,011][starVLA.training.train_starvla][INFO] - Step 758, Loss: {'train/loss': 0.959964394569397, 'train/grad_norm_pre_clip': 14.59009075164795, 'timing/data': 0.023298725485801697, 'timing/model': 0.9332174817100167, 'learning_rate/qwen_vl_interface': 9.788424083925996e-06, 'learning_rate/action_model': 9.788424083925995e-05, 'epoch': 0.31}) +[2026-06-17 17:57:51,134][starVLA.training.train_starvla][INFO] - Step 759, Loss: {'train/loss': 0.8702788949012756, 'train/grad_norm_pre_clip': 8.330917358398438, 'timing/data': 0.01972161326557398, 'timing/model': 0.9759100042283535, 'learning_rate/qwen_vl_interface': 9.787785342828554e-06, 'learning_rate/action_model': 9.787785342828553e-05, 'epoch': 0.31}) +[2026-06-17 17:57:55,014][starVLA.training.train_starvla][INFO] - Step 760, Loss: {'train/loss': 0.6701930165290833, 'train/grad_norm_pre_clip': 8.927044868469238, 'timing/data': 0.019232481718063354, 'timing/model': 0.9273030459880829, 'learning_rate/qwen_vl_interface': 9.78714566104333e-06, 'learning_rate/action_model': 9.787145661043329e-05, 'epoch': 0.31}) +[2026-06-17 17:57:58,933][starVLA.training.train_starvla][INFO] - Step 761, Loss: {'train/loss': 1.0418055057525635, 'train/grad_norm_pre_clip': 8.134398460388184, 'timing/data': 0.024340609088540077, 'timing/model': 0.9258805215358734, 'learning_rate/qwen_vl_interface': 9.786505038702926e-06, 'learning_rate/action_model': 9.786505038702925e-05, 'epoch': 0.31}) +[2026-06-17 17:58:02,829][starVLA.training.train_starvla][INFO] - Step 762, Loss: {'train/loss': 0.9932255148887634, 'train/grad_norm_pre_clip': 9.58301830291748, 'timing/data': 0.02279143873602152, 'timing/model': 0.9674950018525124, 'learning_rate/qwen_vl_interface': 9.785863475940147e-06, 'learning_rate/action_model': 9.785863475940145e-05, 'epoch': 0.31}) +[2026-06-17 17:58:06,911][starVLA.training.train_starvla][INFO] - Step 763, Loss: {'train/loss': 0.6730255484580994, 'train/grad_norm_pre_clip': 7.624610900878906, 'timing/data': 0.05691188480705023, 'timing/model': 1.017984819598496, 'learning_rate/qwen_vl_interface': 9.785220972887988e-06, 'learning_rate/action_model': 9.785220972887989e-05, 'epoch': 0.31}) +[2026-06-17 17:58:10,854][starVLA.training.train_starvla][INFO] - Step 764, Loss: {'train/loss': 0.7463670372962952, 'train/grad_norm_pre_clip': 8.173852920532227, 'timing/data': 0.01992974244058132, 'timing/model': 1.2477291487157345, 'learning_rate/qwen_vl_interface': 9.784577529679643e-06, 'learning_rate/action_model': 9.784577529679643e-05, 'epoch': 0.31}) +[2026-06-17 17:58:14,537][starVLA.training.train_starvla][INFO] - Step 765, Loss: {'train/loss': 1.3032187223434448, 'train/grad_norm_pre_clip': 11.11182689666748, 'timing/data': 0.01758970320224762, 'timing/model': 0.9343556370586157, 'learning_rate/qwen_vl_interface': 9.783933146448495e-06, 'learning_rate/action_model': 9.783933146448495e-05, 'epoch': 0.31}) +[2026-06-17 17:58:18,438][starVLA.training.train_starvla][INFO] - Step 766, Loss: {'train/loss': 0.7568320035934448, 'train/grad_norm_pre_clip': 9.454200744628906, 'timing/data': 0.020270426757633686, 'timing/model': 0.9362637512385845, 'learning_rate/qwen_vl_interface': 9.783287823328127e-06, 'learning_rate/action_model': 9.783287823328127e-05, 'epoch': 0.32}) +[2026-06-17 17:58:22,505][starVLA.training.train_starvla][INFO] - Step 767, Loss: {'train/loss': 0.7294342517852783, 'train/grad_norm_pre_clip': 8.575688362121582, 'timing/data': 0.025247502140700817, 'timing/model': 0.9471023892983794, 'learning_rate/qwen_vl_interface': 9.782641560452317e-06, 'learning_rate/action_model': 9.782641560452316e-05, 'epoch': 0.32}) +[2026-06-17 17:58:26,388][starVLA.training.train_starvla][INFO] - Step 768, Loss: {'train/loss': 0.9119451642036438, 'train/grad_norm_pre_clip': 13.095746994018555, 'timing/data': 0.02245945855975151, 'timing/model': 0.9455788526684046, 'learning_rate/qwen_vl_interface': 9.781994357955032e-06, 'learning_rate/action_model': 9.781994357955032e-05, 'epoch': 0.32}) +[2026-06-17 17:58:30,258][starVLA.training.train_starvla][INFO] - Step 769, Loss: {'train/loss': 1.0433571338653564, 'train/grad_norm_pre_clip': 6.647625923156738, 'timing/data': 0.021252382546663284, 'timing/model': 0.936340045183897, 'learning_rate/qwen_vl_interface': 9.78134621597044e-06, 'learning_rate/action_model': 9.781346215970441e-05, 'epoch': 0.32}) +[2026-06-17 17:58:34,077][starVLA.training.train_starvla][INFO] - Step 770, Loss: {'train/loss': 0.6431719660758972, 'train/grad_norm_pre_clip': 8.722763061523438, 'timing/data': 0.017044142819941044, 'timing/model': 1.1613407926633954, 'learning_rate/qwen_vl_interface': 9.780697134632903e-06, 'learning_rate/action_model': 9.780697134632902e-05, 'epoch': 0.32}) +[2026-06-17 17:58:38,406][starVLA.training.train_starvla][INFO] - Step 771, Loss: {'train/loss': 0.8266761302947998, 'train/grad_norm_pre_clip': 7.298286437988281, 'timing/data': 0.023852181620895863, 'timing/model': 1.1781247491016984, 'learning_rate/qwen_vl_interface': 9.780047114076974e-06, 'learning_rate/action_model': 9.780047114076973e-05, 'epoch': 0.32}) +[2026-06-17 17:58:42,106][starVLA.training.train_starvla][INFO] - Step 772, Loss: {'train/loss': 0.8148244619369507, 'train/grad_norm_pre_clip': 7.539830207824707, 'timing/data': 0.018482406623661518, 'timing/model': 0.9454729603603482, 'learning_rate/qwen_vl_interface': 9.779396154437404e-06, 'learning_rate/action_model': 9.779396154437403e-05, 'epoch': 0.32}) +[2026-06-17 17:58:46,316][starVLA.training.train_starvla][INFO] - Step 773, Loss: {'train/loss': 0.9138509631156921, 'train/grad_norm_pre_clip': 11.344307899475098, 'timing/data': 0.019315628334879875, 'timing/model': 0.9362779762595892, 'learning_rate/qwen_vl_interface': 9.778744255849138e-06, 'learning_rate/action_model': 9.778744255849138e-05, 'epoch': 0.32}) +[2026-06-17 17:58:50,375][starVLA.training.train_starvla][INFO] - Step 774, Loss: {'train/loss': 0.707522451877594, 'train/grad_norm_pre_clip': 111.29299926757812, 'timing/data': 0.020976271480321884, 'timing/model': 0.9329226016998291, 'learning_rate/qwen_vl_interface': 9.778091418447313e-06, 'learning_rate/action_model': 9.778091418447312e-05, 'epoch': 0.32}) +[2026-06-17 17:58:54,209][starVLA.training.train_starvla][INFO] - Step 775, Loss: {'train/loss': 0.5794689059257507, 'train/grad_norm_pre_clip': 7.45355224609375, 'timing/data': 0.018491116352379322, 'timing/model': 0.9306829506531358, 'learning_rate/qwen_vl_interface': 9.777437642367265e-06, 'learning_rate/action_model': 9.777437642367264e-05, 'epoch': 0.32}) +[2026-06-17 17:58:58,097][starVLA.training.train_starvla][INFO] - Step 776, Loss: {'train/loss': 1.0438780784606934, 'train/grad_norm_pre_clip': 8.132948875427246, 'timing/data': 0.024261525832116604, 'timing/model': 0.9769179048016667, 'learning_rate/qwen_vl_interface': 9.776782927744523e-06, 'learning_rate/action_model': 9.776782927744522e-05, 'epoch': 0.32}) +[2026-06-17 17:59:01,999][starVLA.training.train_starvla][INFO] - Step 777, Loss: {'train/loss': 0.5186972618103027, 'train/grad_norm_pre_clip': 5.39792537689209, 'timing/data': 0.022952290251851082, 'timing/model': 1.1556983543559909, 'learning_rate/qwen_vl_interface': 9.776127274714808e-06, 'learning_rate/action_model': 9.776127274714807e-05, 'epoch': 0.32}) +[2026-06-17 17:59:05,837][starVLA.training.train_starvla][INFO] - Step 778, Loss: {'train/loss': 0.8896787762641907, 'train/grad_norm_pre_clip': 7.052087306976318, 'timing/data': 0.01953138690441847, 'timing/model': 1.131501717492938, 'learning_rate/qwen_vl_interface': 9.775470683414038e-06, 'learning_rate/action_model': 9.775470683414038e-05, 'epoch': 0.32}) +[2026-06-17 17:59:09,566][starVLA.training.train_starvla][INFO] - Step 779, Loss: {'train/loss': 0.7824618220329285, 'train/grad_norm_pre_clip': 8.325907707214355, 'timing/data': 0.019059212878346443, 'timing/model': 0.9303612383082509, 'learning_rate/qwen_vl_interface': 9.774813153978327e-06, 'learning_rate/action_model': 9.774813153978327e-05, 'epoch': 0.32}) +[2026-06-17 17:59:13,522][starVLA.training.train_starvla][INFO] - Step 780, Loss: {'train/loss': 0.7565576434135437, 'train/grad_norm_pre_clip': 6.707614898681641, 'timing/data': 0.016802079044282436, 'timing/model': 0.9377774484455585, 'learning_rate/qwen_vl_interface': 9.77415468654398e-06, 'learning_rate/action_model': 9.77415468654398e-05, 'epoch': 0.32}) +[2026-06-17 17:59:17,411][starVLA.training.train_starvla][INFO] - Step 781, Loss: {'train/loss': 0.8621962666511536, 'train/grad_norm_pre_clip': 15.093369483947754, 'timing/data': 0.015556510537862778, 'timing/model': 0.9561688620597124, 'learning_rate/qwen_vl_interface': 9.773495281247497e-06, 'learning_rate/action_model': 9.773495281247497e-05, 'epoch': 0.32}) +[2026-06-17 17:59:21,224][starVLA.training.train_starvla][INFO] - Step 782, Loss: {'train/loss': 0.5566017627716064, 'train/grad_norm_pre_clip': 6.477084159851074, 'timing/data': 0.01829254999756813, 'timing/model': 0.9364285878837109, 'learning_rate/qwen_vl_interface': 9.772834938225577e-06, 'learning_rate/action_model': 9.772834938225576e-05, 'epoch': 0.32}) +[2026-06-17 17:59:25,066][starVLA.training.train_starvla][INFO] - Step 783, Loss: {'train/loss': 0.7863778471946716, 'train/grad_norm_pre_clip': 8.876215934753418, 'timing/data': 0.01755954883992672, 'timing/model': 0.9270505998283625, 'learning_rate/qwen_vl_interface': 9.772173657615107e-06, 'learning_rate/action_model': 9.772173657615106e-05, 'epoch': 0.32}) +[2026-06-17 17:59:28,989][starVLA.training.train_starvla][INFO] - Step 784, Loss: {'train/loss': 0.6701474785804749, 'train/grad_norm_pre_clip': 8.018202781677246, 'timing/data': 0.018152807839214802, 'timing/model': 1.193912890739739, 'learning_rate/qwen_vl_interface': 9.77151143955317e-06, 'learning_rate/action_model': 9.77151143955317e-05, 'epoch': 0.32}) +[2026-06-17 17:59:32,898][starVLA.training.train_starvla][INFO] - Step 785, Loss: {'train/loss': 0.6693690419197083, 'train/grad_norm_pre_clip': 7.737159729003906, 'timing/data': 0.026080798357725143, 'timing/model': 1.216172469779849, 'learning_rate/qwen_vl_interface': 9.770848284177048e-06, 'learning_rate/action_model': 9.770848284177047e-05, 'epoch': 0.32}) +[2026-06-17 17:59:36,727][starVLA.training.train_starvla][INFO] - Step 786, Loss: {'train/loss': 0.7406420111656189, 'train/grad_norm_pre_clip': 7.545047283172607, 'timing/data': 0.020824783481657505, 'timing/model': 0.9465479636564851, 'learning_rate/qwen_vl_interface': 9.77018419162421e-06, 'learning_rate/action_model': 9.77018419162421e-05, 'epoch': 0.32}) +[2026-06-17 17:59:40,698][starVLA.training.train_starvla][INFO] - Step 787, Loss: {'train/loss': 0.8544395565986633, 'train/grad_norm_pre_clip': 126.58412170410156, 'timing/data': 0.016516624949872494, 'timing/model': 0.9390246793627739, 'learning_rate/qwen_vl_interface': 9.769519162032328e-06, 'learning_rate/action_model': 9.769519162032327e-05, 'epoch': 0.32}) +[2026-06-17 17:59:44,821][starVLA.training.train_starvla][INFO] - Step 788, Loss: {'train/loss': 0.5476062893867493, 'train/grad_norm_pre_clip': 7.0841264724731445, 'timing/data': 0.017438880167901516, 'timing/model': 0.9540886403992772, 'learning_rate/qwen_vl_interface': 9.768853195539259e-06, 'learning_rate/action_model': 9.768853195539258e-05, 'epoch': 0.32}) +[2026-06-17 17:59:49,345][starVLA.training.train_starvla][INFO] - Step 789, Loss: {'train/loss': 0.7565081119537354, 'train/grad_norm_pre_clip': 9.557600975036621, 'timing/data': 0.01894314866513014, 'timing/model': 0.9399272613227367, 'learning_rate/qwen_vl_interface': 9.76818629228306e-06, 'learning_rate/action_model': 9.768186292283059e-05, 'epoch': 0.32}) +[2026-06-17 17:59:53,311][starVLA.training.train_starvla][INFO] - Step 790, Loss: {'train/loss': 0.9141137003898621, 'train/grad_norm_pre_clip': 11.847234725952148, 'timing/data': 0.03545869421213865, 'timing/model': 1.2142644776031375, 'learning_rate/qwen_vl_interface': 9.767518452401981e-06, 'learning_rate/action_model': 9.76751845240198e-05, 'epoch': 0.33}) +[2026-06-17 17:59:57,253][starVLA.training.train_starvla][INFO] - Step 791, Loss: {'train/loss': 0.8169746398925781, 'train/grad_norm_pre_clip': 10.208011627197266, 'timing/data': 0.025669346563518047, 'timing/model': 1.2093798285350204, 'learning_rate/qwen_vl_interface': 9.766849676034465e-06, 'learning_rate/action_model': 9.766849676034464e-05, 'epoch': 0.33}) +[2026-06-17 18:00:00,894][starVLA.training.train_starvla][INFO] - Step 792, Loss: {'train/loss': 0.8475116491317749, 'train/grad_norm_pre_clip': 11.51978588104248, 'timing/data': 0.02024312410503626, 'timing/model': 0.9404134443029761, 'learning_rate/qwen_vl_interface': 9.766179963319152e-06, 'learning_rate/action_model': 9.76617996331915e-05, 'epoch': 0.33}) +[2026-06-17 18:00:05,115][starVLA.training.train_starvla][INFO] - Step 793, Loss: {'train/loss': 0.6906598210334778, 'train/grad_norm_pre_clip': 8.911534309387207, 'timing/data': 0.027313959784805775, 'timing/model': 0.9391892943531275, 'learning_rate/qwen_vl_interface': 9.765509314394872e-06, 'learning_rate/action_model': 9.765509314394872e-05, 'epoch': 0.33}) +[2026-06-17 18:00:09,695][starVLA.training.train_starvla][INFO] - Step 794, Loss: {'train/loss': 1.0851203203201294, 'train/grad_norm_pre_clip': 10.497940063476562, 'timing/data': 0.020945179276168346, 'timing/model': 0.9924955815076828, 'learning_rate/qwen_vl_interface': 9.764837729400652e-06, 'learning_rate/action_model': 9.764837729400653e-05, 'epoch': 0.33}) +[2026-06-17 18:00:14,165][starVLA.training.train_starvla][INFO] - Step 795, Loss: {'train/loss': 0.8253476023674011, 'train/grad_norm_pre_clip': 13.084026336669922, 'timing/data': 0.018025286495685577, 'timing/model': 0.9890047833323479, 'learning_rate/qwen_vl_interface': 9.764165208475714e-06, 'learning_rate/action_model': 9.764165208475714e-05, 'epoch': 0.33}) +[2026-06-17 18:00:18,393][starVLA.training.train_starvla][INFO] - Step 796, Loss: {'train/loss': 0.5871754884719849, 'train/grad_norm_pre_clip': 9.502707481384277, 'timing/data': 0.02404639031738043, 'timing/model': 0.9439182635396719, 'learning_rate/qwen_vl_interface': 9.76349175175947e-06, 'learning_rate/action_model': 9.763491751759469e-05, 'epoch': 0.33}) +[2026-06-17 18:00:22,636][starVLA.training.train_starvla][INFO] - Step 797, Loss: {'train/loss': 0.7991657853126526, 'train/grad_norm_pre_clip': 8.633838653564453, 'timing/data': 0.021373217925429344, 'timing/model': 1.3467278480529785, 'learning_rate/qwen_vl_interface': 9.76281735939153e-06, 'learning_rate/action_model': 9.762817359391529e-05, 'epoch': 0.33}) +[2026-06-17 18:00:26,871][starVLA.training.train_starvla][INFO] - Step 798, Loss: {'train/loss': 0.5678424835205078, 'train/grad_norm_pre_clip': 8.858271598815918, 'timing/data': 0.021200792863965034, 'timing/model': 1.5144801922142506, 'learning_rate/qwen_vl_interface': 9.762142031511695e-06, 'learning_rate/action_model': 9.762142031511695e-05, 'epoch': 0.33}) +[2026-06-17 18:00:30,818][starVLA.training.train_starvla][INFO] - Step 799, Loss: {'train/loss': 0.7378435134887695, 'train/grad_norm_pre_clip': 11.910106658935547, 'timing/data': 0.01720591075718403, 'timing/model': 0.9340083915740252, 'learning_rate/qwen_vl_interface': 9.76146576825996e-06, 'learning_rate/action_model': 9.761465768259959e-05, 'epoch': 0.33}) +[2026-06-17 18:02:25,604][starVLA.training.train_starvla][INFO] - Step 800, Loss: {'train/loss': 0.9084669947624207, 'train/grad_norm_pre_clip': 11.335236549377441, 'eval/loss': 0.7523888349533081, 'eval/latency_6/loss': 0.7523888349533081, 'eval/demon_fire_precision': 0.9263224181360201, 'eval/demon_fire_recall': 0.9508726567550096, 'eval/demon_fire_f1': 0.9384370015948963, 'eval/demon_fire_rate_model': 0.8836950473010573, 'eval/demon_fire_rate_teacher': 0.860879243183083, 'eval/demon_left_precision': 0.7533875338753387, 'eval/demon_left_recall': 0.48771929824561405, 'eval/demon_left_f1': 0.5921192758253462, 'eval/demon_left_rate_model': 0.10267111853088481, 'eval/demon_left_rate_teacher': 0.15859766277128548, 'eval/demon_right_precision': 0.7052117263843648, 'eval/demon_right_recall': 0.6651305683563749, 'eval/demon_right_f1': 0.6845849802371542, 'eval/demon_right_rate_model': 0.17084028937117418, 'eval/demon_right_rate_teacher': 0.18113522537562604, 'eval/demon_move_f1': 0.6383521280312502, 'eval/demon_cc_f1': 0.8224487753787701, 'eval/latency_6/demon_fire_precision': 0.9263224181360201, 'eval/latency_6/demon_fire_recall': 0.9508726567550096, 'eval/latency_6/demon_fire_f1': 0.9384370015948963, 'eval/latency_6/demon_fire_rate_model': 0.8836950473010573, 'eval/latency_6/demon_fire_rate_teacher': 0.860879243183083, 'eval/latency_6/demon_left_precision': 0.7533875338753387, 'eval/latency_6/demon_left_recall': 0.48771929824561405, 'eval/latency_6/demon_left_f1': 0.5921192758253462, 'eval/latency_6/demon_left_rate_model': 0.10267111853088481, 'eval/latency_6/demon_left_rate_teacher': 0.15859766277128548, 'eval/latency_6/demon_right_precision': 0.7052117263843648, 'eval/latency_6/demon_right_recall': 0.6651305683563749, 'eval/latency_6/demon_right_f1': 0.6845849802371542, 'eval/latency_6/demon_right_rate_model': 0.17084028937117418, 'eval/latency_6/demon_right_rate_teacher': 0.18113522537562604, 'eval/latency_6/demon_move_f1': 0.6383521280312502, 'eval/latency_6/demon_cc_f1': 0.8224487753787701, 'timing/data': 0.023230338469147682, 'timing/model': 0.9280024813488126, 'learning_rate/qwen_vl_interface': 9.760788569776518e-06, 'learning_rate/action_model': 9.760788569776517e-05, 'epoch': 0.33}) +[2026-06-17 18:02:25,605][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_800_state +[2026-06-17 18:02:25,612][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 18:02:32,390][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_800_state/model.safetensors +[2026-06-17 18:02:48,412][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_800_state/optimizer.bin +[2026-06-17 18:02:48,413][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_800_state/sampler.bin +[2026-06-17 18:02:48,413][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_800_state/sampler_1.bin +[2026-06-17 18:02:48,416][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_800_state/random_states_0.pkl +[2026-06-17 18:02:48,422][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 800 +[2026-06-17 18:02:48,422][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 18:02:48,433][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 18:02:52,120][starVLA.training.train_starvla][INFO] - Step 801, Loss: {'train/loss': 0.627765953540802, 'train/grad_norm_pre_clip': 10.5226411819458, 'timing/data': 0.027923036366701126, 'timing/model': 0.9424429293721914, 'learning_rate/qwen_vl_interface': 9.760110436201752e-06, 'learning_rate/action_model': 9.76011043620175e-05, 'epoch': 0.33}) +[2026-06-17 18:02:56,256][starVLA.training.train_starvla][INFO] - Step 802, Loss: {'train/loss': 0.5439384579658508, 'train/grad_norm_pre_clip': 6.327301979064941, 'timing/data': 0.021414317190647125, 'timing/model': 1.1580057637766004, 'learning_rate/qwen_vl_interface': 9.759431367676238e-06, 'learning_rate/action_model': 9.759431367676237e-05, 'epoch': 0.33}) +[2026-06-17 18:03:00,194][starVLA.training.train_starvla][INFO] - Step 803, Loss: {'train/loss': 1.3713226318359375, 'train/grad_norm_pre_clip': 12.407090187072754, 'timing/data': 0.032312932424247265, 'timing/model': 1.007899433374405, 'learning_rate/qwen_vl_interface': 9.758751364340751e-06, 'learning_rate/action_model': 9.75875136434075e-05, 'epoch': 0.33}) +[2026-06-17 18:03:04,046][starVLA.training.train_starvla][INFO] - Step 804, Loss: {'train/loss': 0.6261720061302185, 'train/grad_norm_pre_clip': 8.925271034240723, 'timing/data': 0.01938759908080101, 'timing/model': 0.9475804399698973, 'learning_rate/qwen_vl_interface': 9.758070426336252e-06, 'learning_rate/action_model': 9.758070426336252e-05, 'epoch': 0.33}) +[2026-06-17 18:03:07,894][starVLA.training.train_starvla][INFO] - Step 805, Loss: {'train/loss': 0.47168847918510437, 'train/grad_norm_pre_clip': 9.254009246826172, 'timing/data': 0.016803056001663208, 'timing/model': 0.9485860671848059, 'learning_rate/qwen_vl_interface': 9.757388553803905e-06, 'learning_rate/action_model': 9.757388553803905e-05, 'epoch': 0.33}) +[2026-06-17 18:03:11,734][starVLA.training.train_starvla][INFO] - Step 806, Loss: {'train/loss': 1.0377650260925293, 'train/grad_norm_pre_clip': 11.13586139678955, 'timing/data': 0.025671524927020073, 'timing/model': 1.1325692804530263, 'learning_rate/qwen_vl_interface': 9.75670574688506e-06, 'learning_rate/action_model': 9.75670574688506e-05, 'epoch': 0.33}) +[2026-06-17 18:03:15,361][starVLA.training.train_starvla][INFO] - Step 807, Loss: {'train/loss': 0.5316060781478882, 'train/grad_norm_pre_clip': 6.685639381408691, 'timing/data': 0.030866037122905254, 'timing/model': 0.935182286426425, 'learning_rate/qwen_vl_interface': 9.756022005721265e-06, 'learning_rate/action_model': 9.756022005721265e-05, 'epoch': 0.33}) +[2026-06-17 18:03:19,390][starVLA.training.train_starvla][INFO] - Step 808, Loss: {'train/loss': 0.8633579611778259, 'train/grad_norm_pre_clip': 8.728691101074219, 'timing/data': 0.015608653426170349, 'timing/model': 0.934989214874804, 'learning_rate/qwen_vl_interface': 9.755337330454257e-06, 'learning_rate/action_model': 9.755337330454257e-05, 'epoch': 0.33}) +[2026-06-17 18:03:23,991][starVLA.training.train_starvla][INFO] - Step 809, Loss: {'train/loss': 0.6821229457855225, 'train/grad_norm_pre_clip': 9.904091835021973, 'timing/data': 0.019574743695557117, 'timing/model': 1.6121389046311378, 'learning_rate/qwen_vl_interface': 9.754651721225975e-06, 'learning_rate/action_model': 9.754651721225975e-05, 'epoch': 0.33}) +[2026-06-17 18:03:27,926][starVLA.training.train_starvla][INFO] - Step 810, Loss: {'train/loss': 0.6380163431167603, 'train/grad_norm_pre_clip': 10.143654823303223, 'timing/data': 0.026627602986991405, 'timing/model': 0.931763781234622, 'learning_rate/qwen_vl_interface': 9.753965178178543e-06, 'learning_rate/action_model': 9.753965178178543e-05, 'epoch': 0.33}) +[2026-06-17 18:03:31,805][starVLA.training.train_starvla][INFO] - Step 811, Loss: {'train/loss': 0.8934614658355713, 'train/grad_norm_pre_clip': 12.2706880569458, 'timing/data': 0.02500747051090002, 'timing/model': 0.9266527583822608, 'learning_rate/qwen_vl_interface': 9.753277701454285e-06, 'learning_rate/action_model': 9.753277701454284e-05, 'epoch': 0.33}) +[2026-06-17 18:03:35,644][starVLA.training.train_starvla][INFO] - Step 812, Loss: {'train/loss': 0.9983747005462646, 'train/grad_norm_pre_clip': 10.051773071289062, 'timing/data': 0.01653364859521389, 'timing/model': 1.1650898642838001, 'learning_rate/qwen_vl_interface': 9.752589291195711e-06, 'learning_rate/action_model': 9.75258929119571e-05, 'epoch': 0.33}) +[2026-06-17 18:03:39,463][starVLA.training.train_starvla][INFO] - Step 813, Loss: {'train/loss': 0.6658852100372314, 'train/grad_norm_pre_clip': 8.84055233001709, 'timing/data': 0.0243093091994524, 'timing/model': 1.1402508346363902, 'learning_rate/qwen_vl_interface': 9.751899947545534e-06, 'learning_rate/action_model': 9.751899947545533e-05, 'epoch': 0.33}) +[2026-06-17 18:03:43,219][starVLA.training.train_starvla][INFO] - Step 814, Loss: {'train/loss': 0.8954350352287292, 'train/grad_norm_pre_clip': 9.808334350585938, 'timing/data': 0.01650973316282034, 'timing/model': 0.9443802200257778, 'learning_rate/qwen_vl_interface': 9.751209670646653e-06, 'learning_rate/action_model': 9.751209670646652e-05, 'epoch': 0.34}) +[2026-06-17 18:03:47,038][starVLA.training.train_starvla][INFO] - Step 815, Loss: {'train/loss': 0.5752018094062805, 'train/grad_norm_pre_clip': 9.136157989501953, 'timing/data': 0.023580160923302174, 'timing/model': 0.951104799285531, 'learning_rate/qwen_vl_interface': 9.750518460642162e-06, 'learning_rate/action_model': 9.75051846064216e-05, 'epoch': 0.34}) +[2026-06-17 18:03:50,893][starVLA.training.train_starvla][INFO] - Step 816, Loss: {'train/loss': 0.9985460638999939, 'train/grad_norm_pre_clip': 7.491085052490234, 'timing/data': 0.0161740155890584, 'timing/model': 0.9341751709580421, 'learning_rate/qwen_vl_interface': 9.749826317675352e-06, 'learning_rate/action_model': 9.749826317675352e-05, 'epoch': 0.34}) +[2026-06-17 18:03:55,069][starVLA.training.train_starvla][INFO] - Step 817, Loss: {'train/loss': 0.8740149736404419, 'train/grad_norm_pre_clip': 8.881814956665039, 'timing/data': 0.015612880699336529, 'timing/model': 1.0695606926456094, 'learning_rate/qwen_vl_interface': 9.749133241889706e-06, 'learning_rate/action_model': 9.749133241889705e-05, 'epoch': 0.34}) +[2026-06-17 18:03:59,241][starVLA.training.train_starvla][INFO] - Step 818, Loss: {'train/loss': 0.8984747529029846, 'train/grad_norm_pre_clip': 32.14179229736328, 'timing/data': 0.02159335371106863, 'timing/model': 0.9365043351426721, 'learning_rate/qwen_vl_interface': 9.748439233428897e-06, 'learning_rate/action_model': 9.748439233428896e-05, 'epoch': 0.34}) +[2026-06-17 18:04:03,030][starVLA.training.train_starvla][INFO] - Step 819, Loss: {'train/loss': 0.8404138684272766, 'train/grad_norm_pre_clip': 7.4835968017578125, 'timing/data': 0.02197778783738613, 'timing/model': 1.1239620810374618, 'learning_rate/qwen_vl_interface': 9.747744292436794e-06, 'learning_rate/action_model': 9.747744292436793e-05, 'epoch': 0.34}) +[2026-06-17 18:04:08,739][starVLA.training.train_starvla][INFO] - Step 820, Loss: {'train/loss': 0.7644608020782471, 'train/grad_norm_pre_clip': 6.737951278686523, 'timing/data': 0.0183449424803257, 'timing/model': 0.9386625476181507, 'learning_rate/qwen_vl_interface': 9.747048419057458e-06, 'learning_rate/action_model': 9.747048419057458e-05, 'epoch': 0.34}) +[2026-06-17 18:04:12,702][starVLA.training.train_starvla][INFO] - Step 821, Loss: {'train/loss': 0.46606501936912537, 'train/grad_norm_pre_clip': 16.454893112182617, 'timing/data': 0.018933359533548355, 'timing/model': 0.9472399111837149, 'learning_rate/qwen_vl_interface': 9.746351613435148e-06, 'learning_rate/action_model': 9.746351613435147e-05, 'epoch': 0.34}) +[2026-06-17 18:04:16,527][starVLA.training.train_starvla][INFO] - Step 822, Loss: {'train/loss': 0.9566643834114075, 'train/grad_norm_pre_clip': 10.535064697265625, 'timing/data': 0.020746195688843727, 'timing/model': 0.9365868559107184, 'learning_rate/qwen_vl_interface': 9.74565387571431e-06, 'learning_rate/action_model': 9.745653875714309e-05, 'epoch': 0.34}) +[2026-06-17 18:04:20,368][starVLA.training.train_starvla][INFO] - Step 823, Loss: {'train/loss': 0.8879728317260742, 'train/grad_norm_pre_clip': 7.1077165603637695, 'timing/data': 0.018867009319365025, 'timing/model': 0.9875180060043931, 'learning_rate/qwen_vl_interface': 9.744955206039585e-06, 'learning_rate/action_model': 9.744955206039585e-05, 'epoch': 0.34}) +[2026-06-17 18:04:24,601][starVLA.training.train_starvla][INFO] - Step 824, Loss: {'train/loss': 1.0264729261398315, 'train/grad_norm_pre_clip': 10.14441204071045, 'timing/data': 0.014129919931292534, 'timing/model': 0.9712280668318272, 'learning_rate/qwen_vl_interface': 9.74425560455581e-06, 'learning_rate/action_model': 9.74425560455581e-05, 'epoch': 0.34}) +[2026-06-17 18:04:28,428][starVLA.training.train_starvla][INFO] - Step 825, Loss: {'train/loss': 0.7432241439819336, 'train/grad_norm_pre_clip': 10.651080131530762, 'timing/data': 0.01770514901727438, 'timing/model': 0.9978536758571863, 'learning_rate/qwen_vl_interface': 9.743555071408011e-06, 'learning_rate/action_model': 9.743555071408011e-05, 'epoch': 0.34}) +[2026-06-17 18:04:32,807][starVLA.training.train_starvla][INFO] - Step 826, Loss: {'train/loss': 0.5362133979797363, 'train/grad_norm_pre_clip': 6.033130645751953, 'timing/data': 0.01851547881960869, 'timing/model': 1.2426322344690561, 'learning_rate/qwen_vl_interface': 9.742853606741412e-06, 'learning_rate/action_model': 9.742853606741411e-05, 'epoch': 0.34}) +[2026-06-17 18:04:36,527][starVLA.training.train_starvla][INFO] - Step 827, Loss: {'train/loss': 0.9208733439445496, 'train/grad_norm_pre_clip': 8.398090362548828, 'timing/data': 0.024293423630297184, 'timing/model': 1.0095936693251133, 'learning_rate/qwen_vl_interface': 9.742151210701423e-06, 'learning_rate/action_model': 9.742151210701424e-05, 'epoch': 0.34}) +[2026-06-17 18:04:41,484][starVLA.training.train_starvla][INFO] - Step 828, Loss: {'train/loss': 0.5999300479888916, 'train/grad_norm_pre_clip': 6.859809398651123, 'timing/data': 0.022057593800127506, 'timing/model': 0.929049008525908, 'learning_rate/qwen_vl_interface': 9.741447883433656e-06, 'learning_rate/action_model': 9.741447883433656e-05, 'epoch': 0.34}) +[2026-06-17 18:04:45,441][starVLA.training.train_starvla][INFO] - Step 829, Loss: {'train/loss': 0.8369256258010864, 'train/grad_norm_pre_clip': 7.347023963928223, 'timing/data': 0.018457523547112942, 'timing/model': 0.9468871913850307, 'learning_rate/qwen_vl_interface': 9.74074362508391e-06, 'learning_rate/action_model': 9.740743625083911e-05, 'epoch': 0.34}) +[2026-06-17 18:04:49,270][starVLA.training.train_starvla][INFO] - Step 830, Loss: {'train/loss': 0.8657187223434448, 'train/grad_norm_pre_clip': 8.56814956665039, 'timing/data': 0.01841303799301386, 'timing/model': 0.9507800955325365, 'learning_rate/qwen_vl_interface': 9.74003843579818e-06, 'learning_rate/action_model': 9.740038435798178e-05, 'epoch': 0.34}) +[2026-06-17 18:04:55,082][starVLA.training.train_starvla][INFO] - Step 831, Loss: {'train/loss': 0.8112252354621887, 'train/grad_norm_pre_clip': 8.674067497253418, 'timing/data': 0.025395726785063744, 'timing/model': 1.756458641961217, 'learning_rate/qwen_vl_interface': 9.739332315722647e-06, 'learning_rate/action_model': 9.739332315722646e-05, 'epoch': 0.34}) +[2026-06-17 18:04:59,130][starVLA.training.train_starvla][INFO] - Step 832, Loss: {'train/loss': 0.4994988739490509, 'train/grad_norm_pre_clip': 5.287795066833496, 'timing/data': 0.021055215038359165, 'timing/model': 1.175422029569745, 'learning_rate/qwen_vl_interface': 9.738625265003695e-06, 'learning_rate/action_model': 9.738625265003695e-05, 'epoch': 0.34}) +[2026-06-17 18:05:03,068][starVLA.training.train_starvla][INFO] - Step 833, Loss: {'train/loss': 0.5409356951713562, 'train/grad_norm_pre_clip': 5.760388374328613, 'timing/data': 0.01873953267931938, 'timing/model': 1.1548777716234326, 'learning_rate/qwen_vl_interface': 9.7379172837879e-06, 'learning_rate/action_model': 9.737917283787898e-05, 'epoch': 0.34}) +[2026-06-17 18:05:06,948][starVLA.training.train_starvla][INFO] - Step 834, Loss: {'train/loss': 0.563267707824707, 'train/grad_norm_pre_clip': 10.201687812805176, 'timing/data': 0.015832291916012764, 'timing/model': 0.9402191964909434, 'learning_rate/qwen_vl_interface': 9.737208372222018e-06, 'learning_rate/action_model': 9.737208372222017e-05, 'epoch': 0.34}) +[2026-06-17 18:05:10,868][starVLA.training.train_starvla][INFO] - Step 835, Loss: {'train/loss': 1.0510859489440918, 'train/grad_norm_pre_clip': 11.82974910736084, 'timing/data': 0.02463117614388466, 'timing/model': 0.942972794175148, 'learning_rate/qwen_vl_interface': 9.736498530453015e-06, 'learning_rate/action_model': 9.736498530453014e-05, 'epoch': 0.34}) +[2026-06-17 18:05:14,803][starVLA.training.train_starvla][INFO] - Step 836, Loss: {'train/loss': 0.730401873588562, 'train/grad_norm_pre_clip': 7.603038311004639, 'timing/data': 0.022941624745726585, 'timing/model': 0.9436473893001676, 'learning_rate/qwen_vl_interface': 9.735787758628036e-06, 'learning_rate/action_model': 9.735787758628035e-05, 'epoch': 0.34}) +[2026-06-17 18:05:20,704][starVLA.training.train_starvla][INFO] - Step 837, Loss: {'train/loss': 0.7294151186943054, 'train/grad_norm_pre_clip': 7.434839725494385, 'timing/data': 0.024547106586396694, 'timing/model': 0.9485493283718824, 'learning_rate/qwen_vl_interface': 9.73507605689443e-06, 'learning_rate/action_model': 9.735076056894429e-05, 'epoch': 0.34}) +[2026-06-17 18:05:24,548][starVLA.training.train_starvla][INFO] - Step 838, Loss: {'train/loss': 0.8054508566856384, 'train/grad_norm_pre_clip': 9.687552452087402, 'timing/data': 0.028354317881166935, 'timing/model': 0.9441337753087282, 'learning_rate/qwen_vl_interface': 9.73436342539973e-06, 'learning_rate/action_model': 9.734363425399728e-05, 'epoch': 0.34}) +[2026-06-17 18:05:28,539][starVLA.training.train_starvla][INFO] - Step 839, Loss: {'train/loss': 0.9408599734306335, 'train/grad_norm_pre_clip': 8.86129093170166, 'timing/data': 0.01613911148160696, 'timing/model': 1.2525047082453966, 'learning_rate/qwen_vl_interface': 9.733649864291667e-06, 'learning_rate/action_model': 9.733649864291666e-05, 'epoch': 0.35}) +[2026-06-17 18:05:32,306][starVLA.training.train_starvla][INFO] - Step 840, Loss: {'train/loss': 0.5904466509819031, 'train/grad_norm_pre_clip': 9.09722900390625, 'timing/data': 0.019863667897880077, 'timing/model': 1.0798442577943206, 'learning_rate/qwen_vl_interface': 9.732935373718162e-06, 'learning_rate/action_model': 9.732935373718161e-05, 'epoch': 0.35}) +[2026-06-17 18:05:35,945][starVLA.training.train_starvla][INFO] - Step 841, Loss: {'train/loss': 0.6537793874740601, 'train/grad_norm_pre_clip': 6.55725622177124, 'timing/data': 0.024241776205599308, 'timing/model': 0.9423935767263174, 'learning_rate/qwen_vl_interface': 9.73221995382733e-06, 'learning_rate/action_model': 9.73221995382733e-05, 'epoch': 0.35}) +[2026-06-17 18:05:41,792][starVLA.training.train_starvla][INFO] - Step 842, Loss: {'train/loss': 0.6368400454521179, 'train/grad_norm_pre_clip': 8.938065528869629, 'timing/data': 0.02514870185405016, 'timing/model': 1.6265760930255055, 'learning_rate/qwen_vl_interface': 9.73150360476748e-06, 'learning_rate/action_model': 9.73150360476748e-05, 'epoch': 0.35}) +[2026-06-17 18:05:45,688][starVLA.training.train_starvla][INFO] - Step 843, Loss: {'train/loss': 0.3518329858779907, 'train/grad_norm_pre_clip': 4.445579528808594, 'timing/data': 0.024437489919364452, 'timing/model': 1.014896098524332, 'learning_rate/qwen_vl_interface': 9.73078632668711e-06, 'learning_rate/action_model': 9.73078632668711e-05, 'epoch': 0.35}) +[2026-06-17 18:05:49,874][starVLA.training.train_starvla][INFO] - Step 844, Loss: {'train/loss': 0.6894619464874268, 'train/grad_norm_pre_clip': 8.065214157104492, 'timing/data': 0.015595181845128536, 'timing/model': 0.9963808171451092, 'learning_rate/qwen_vl_interface': 9.730068119734912e-06, 'learning_rate/action_model': 9.730068119734912e-05, 'epoch': 0.35}) +[2026-06-17 18:05:53,759][starVLA.training.train_starvla][INFO] - Step 845, Loss: {'train/loss': 0.889064371585846, 'train/grad_norm_pre_clip': 12.13424015045166, 'timing/data': 0.017756075598299503, 'timing/model': 0.9368862798437476, 'learning_rate/qwen_vl_interface': 9.729348984059773e-06, 'learning_rate/action_model': 9.729348984059774e-05, 'epoch': 0.35}) +[2026-06-17 18:05:57,863][starVLA.training.train_starvla][INFO] - Step 846, Loss: {'train/loss': 0.5975646376609802, 'train/grad_norm_pre_clip': 7.2766828536987305, 'timing/data': 0.023313491605222225, 'timing/model': 1.2884589675813913, 'learning_rate/qwen_vl_interface': 9.728628919810769e-06, 'learning_rate/action_model': 9.728628919810768e-05, 'epoch': 0.35}) +[2026-06-17 18:06:01,729][starVLA.training.train_starvla][INFO] - Step 847, Loss: {'train/loss': 0.6689332127571106, 'train/grad_norm_pre_clip': 7.385623931884766, 'timing/data': 0.01789784897118807, 'timing/model': 1.051848796196282, 'learning_rate/qwen_vl_interface': 9.727907927137171e-06, 'learning_rate/action_model': 9.727907927137171e-05, 'epoch': 0.35}) +[2026-06-17 18:06:05,535][starVLA.training.train_starvla][INFO] - Step 848, Loss: {'train/loss': 0.9725785851478577, 'train/grad_norm_pre_clip': 8.403714179992676, 'timing/data': 0.017232556827366352, 'timing/model': 0.9358288962393999, 'learning_rate/qwen_vl_interface': 9.72718600618844e-06, 'learning_rate/action_model': 9.727186006188441e-05, 'epoch': 0.35}) +[2026-06-17 18:06:09,418][starVLA.training.train_starvla][INFO] - Step 849, Loss: {'train/loss': 0.744726836681366, 'train/grad_norm_pre_clip': 10.269180297851562, 'timing/data': 0.01751964259892702, 'timing/model': 0.9864892195910215, 'learning_rate/qwen_vl_interface': 9.726463157114233e-06, 'learning_rate/action_model': 9.726463157114234e-05, 'epoch': 0.35}) +[2026-06-17 18:06:13,239][starVLA.training.train_starvla][INFO] - Step 850, Loss: {'train/loss': 0.9395287036895752, 'train/grad_norm_pre_clip': 8.74052906036377, 'timing/data': 0.02066913340240717, 'timing/model': 0.9459042241796851, 'learning_rate/qwen_vl_interface': 9.725739380064399e-06, 'learning_rate/action_model': 9.725739380064398e-05, 'epoch': 0.35}) +[2026-06-17 18:06:19,775][starVLA.training.train_starvla][INFO] - Step 851, Loss: {'train/loss': 0.626045286655426, 'train/grad_norm_pre_clip': 8.195284843444824, 'timing/data': 0.01864817924797535, 'timing/model': 0.9892007475718856, 'learning_rate/qwen_vl_interface': 9.725014675188972e-06, 'learning_rate/action_model': 9.725014675188973e-05, 'epoch': 0.35}) +[2026-06-17 18:06:23,704][starVLA.training.train_starvla][INFO] - Step 852, Loss: {'train/loss': 0.6552414298057556, 'train/grad_norm_pre_clip': 10.022858619689941, 'timing/data': 0.0171252628788352, 'timing/model': 1.1544006392359734, 'learning_rate/qwen_vl_interface': 9.72428904263819e-06, 'learning_rate/action_model': 9.72428904263819e-05, 'epoch': 0.35}) +[2026-06-17 18:06:27,570][starVLA.training.train_starvla][INFO] - Step 853, Loss: {'train/loss': 0.39562541246414185, 'train/grad_norm_pre_clip': 7.472501277923584, 'timing/data': 0.02226193156093359, 'timing/model': 1.0857417462393641, 'learning_rate/qwen_vl_interface': 9.723562482562474e-06, 'learning_rate/action_model': 9.723562482562473e-05, 'epoch': 0.35}) +[2026-06-17 18:06:31,229][starVLA.training.train_starvla][INFO] - Step 854, Loss: {'train/loss': 0.7664228081703186, 'train/grad_norm_pre_clip': 9.249773979187012, 'timing/data': 0.027802759781479836, 'timing/model': 0.9334046598523855, 'learning_rate/qwen_vl_interface': 9.72283499511244e-06, 'learning_rate/action_model': 9.722834995112439e-05, 'epoch': 0.35}) +[2026-06-17 18:06:35,143][starVLA.training.train_starvla][INFO] - Step 855, Loss: {'train/loss': 0.9320746064186096, 'train/grad_norm_pre_clip': 11.0977201461792, 'timing/data': 0.01866091974079609, 'timing/model': 0.9814306050539017, 'learning_rate/qwen_vl_interface': 9.722106580438901e-06, 'learning_rate/action_model': 9.7221065804389e-05, 'epoch': 0.35}) +[2026-06-17 18:06:38,999][starVLA.training.train_starvla][INFO] - Step 856, Loss: {'train/loss': 0.5608800053596497, 'train/grad_norm_pre_clip': 8.35383415222168, 'timing/data': 0.019657252356410027, 'timing/model': 0.9531837422400713, 'learning_rate/qwen_vl_interface': 9.721377238692855e-06, 'learning_rate/action_model': 9.721377238692854e-05, 'epoch': 0.35}) +[2026-06-17 18:06:42,831][starVLA.training.train_starvla][INFO] - Step 857, Loss: {'train/loss': 0.8267955780029297, 'train/grad_norm_pre_clip': 9.693252563476562, 'timing/data': 0.01611841283738613, 'timing/model': 0.9351498847827315, 'learning_rate/qwen_vl_interface': 9.720646970025494e-06, 'learning_rate/action_model': 9.720646970025494e-05, 'epoch': 0.35}) +[2026-06-17 18:06:47,010][starVLA.training.train_starvla][INFO] - Step 858, Loss: {'train/loss': 0.7336252331733704, 'train/grad_norm_pre_clip': 8.86837100982666, 'timing/data': 0.01919172052294016, 'timing/model': 0.9627313790842891, 'learning_rate/qwen_vl_interface': 9.719915774588209e-06, 'learning_rate/action_model': 9.719915774588209e-05, 'epoch': 0.35}) +[2026-06-17 18:06:50,909][starVLA.training.train_starvla][INFO] - Step 859, Loss: {'train/loss': 0.6830182671546936, 'train/grad_norm_pre_clip': 9.141934394836426, 'timing/data': 0.017196577042341232, 'timing/model': 1.15079325158149, 'learning_rate/qwen_vl_interface': 9.719183652532572e-06, 'learning_rate/action_model': 9.719183652532572e-05, 'epoch': 0.35}) +[2026-06-17 18:06:54,970][starVLA.training.train_starvla][INFO] - Step 860, Loss: {'train/loss': 0.6678404211997986, 'train/grad_norm_pre_clip': 13.1715669631958, 'timing/data': 0.01824683416634798, 'timing/model': 1.3664098270237446, 'learning_rate/qwen_vl_interface': 9.718450604010355e-06, 'learning_rate/action_model': 9.718450604010355e-05, 'epoch': 0.35}) +[2026-06-17 18:06:58,672][starVLA.training.train_starvla][INFO] - Step 861, Loss: {'train/loss': 0.5856393575668335, 'train/grad_norm_pre_clip': 7.302308559417725, 'timing/data': 0.02861351426690817, 'timing/model': 0.9373217429965734, 'learning_rate/qwen_vl_interface': 9.717716629173521e-06, 'learning_rate/action_model': 9.71771662917352e-05, 'epoch': 0.35}) +[2026-06-17 18:07:02,555][starVLA.training.train_starvla][INFO] - Step 862, Loss: {'train/loss': 0.6486563086509705, 'train/grad_norm_pre_clip': 10.009306907653809, 'timing/data': 0.017949890345335007, 'timing/model': 0.954330300912261, 'learning_rate/qwen_vl_interface': 9.71698172817422e-06, 'learning_rate/action_model': 9.716981728174219e-05, 'epoch': 0.35}) +[2026-06-17 18:07:06,441][starVLA.training.train_starvla][INFO] - Step 863, Loss: {'train/loss': 0.8928724527359009, 'train/grad_norm_pre_clip': 11.15186595916748, 'timing/data': 0.01379514578729868, 'timing/model': 0.9391006464138627, 'learning_rate/qwen_vl_interface': 9.7162459011648e-06, 'learning_rate/action_model': 9.7162459011648e-05, 'epoch': 0.36}) +[2026-06-17 18:07:10,296][starVLA.training.train_starvla][INFO] - Step 864, Loss: {'train/loss': 0.7428218126296997, 'train/grad_norm_pre_clip': 13.196060180664062, 'timing/data': 0.020944693125784397, 'timing/model': 0.9460390107706189, 'learning_rate/qwen_vl_interface': 9.7155091482978e-06, 'learning_rate/action_model': 9.7155091482978e-05, 'epoch': 0.36}) +[2026-06-17 18:07:14,211][starVLA.training.train_starvla][INFO] - Step 865, Loss: {'train/loss': 0.9660561680793762, 'train/grad_norm_pre_clip': 16.365459442138672, 'timing/data': 0.021601294167339802, 'timing/model': 0.9533008448779583, 'learning_rate/qwen_vl_interface': 9.714771469725947e-06, 'learning_rate/action_model': 9.714771469725947e-05, 'epoch': 0.36}) +[2026-06-17 18:07:18,066][starVLA.training.train_starvla][INFO] - Step 866, Loss: {'train/loss': 0.9831652045249939, 'train/grad_norm_pre_clip': 11.456158638000488, 'timing/data': 0.020045162178575993, 'timing/model': 1.139883746393025, 'learning_rate/qwen_vl_interface': 9.714032865602167e-06, 'learning_rate/action_model': 9.714032865602166e-05, 'epoch': 0.36}) +[2026-06-17 18:07:22,029][starVLA.training.train_starvla][INFO] - Step 867, Loss: {'train/loss': 0.9162019491195679, 'train/grad_norm_pre_clip': 11.661245346069336, 'timing/data': 0.022972576320171356, 'timing/model': 1.0875176973640919, 'learning_rate/qwen_vl_interface': 9.713293336079568e-06, 'learning_rate/action_model': 9.713293336079568e-05, 'epoch': 0.36}) +[2026-06-17 18:07:25,688][starVLA.training.train_starvla][INFO] - Step 868, Loss: {'train/loss': 0.6749180555343628, 'train/grad_norm_pre_clip': 11.314424514770508, 'timing/data': 0.019076447933912277, 'timing/model': 0.9434757037088275, 'learning_rate/qwen_vl_interface': 9.712552881311459e-06, 'learning_rate/action_model': 9.712552881311458e-05, 'epoch': 0.36}) +[2026-06-17 18:07:29,592][starVLA.training.train_starvla][INFO] - Step 869, Loss: {'train/loss': 0.7487765550613403, 'train/grad_norm_pre_clip': 6.828997611999512, 'timing/data': 0.018666849471628666, 'timing/model': 0.9371594786643982, 'learning_rate/qwen_vl_interface': 9.711811501451336e-06, 'learning_rate/action_model': 9.711811501451335e-05, 'epoch': 0.36}) +[2026-06-17 18:07:33,536][starVLA.training.train_starvla][INFO] - Step 870, Loss: {'train/loss': 0.6995618939399719, 'train/grad_norm_pre_clip': 9.631294250488281, 'timing/data': 0.02552318572998047, 'timing/model': 0.952176658436656, 'learning_rate/qwen_vl_interface': 9.711069196652886e-06, 'learning_rate/action_model': 9.711069196652886e-05, 'epoch': 0.36}) +[2026-06-17 18:07:37,343][starVLA.training.train_starvla][INFO] - Step 871, Loss: {'train/loss': 0.9690790176391602, 'train/grad_norm_pre_clip': 11.627488136291504, 'timing/data': 0.015955996699631214, 'timing/model': 0.9349755980074406, 'learning_rate/qwen_vl_interface': 9.710325967069992e-06, 'learning_rate/action_model': 9.710325967069991e-05, 'epoch': 0.36}) +[2026-06-17 18:07:41,213][starVLA.training.train_starvla][INFO] - Step 872, Loss: {'train/loss': 0.6212801933288574, 'train/grad_norm_pre_clip': 9.764427185058594, 'timing/data': 0.020387067459523678, 'timing/model': 1.1477321153506637, 'learning_rate/qwen_vl_interface': 9.709581812856725e-06, 'learning_rate/action_model': 9.709581812856726e-05, 'epoch': 0.36}) +[2026-06-17 18:07:44,981][starVLA.training.train_starvla][INFO] - Step 873, Loss: {'train/loss': 0.5835247039794922, 'train/grad_norm_pre_clip': 10.378315925598145, 'timing/data': 0.015463070943951607, 'timing/model': 1.065851727500558, 'learning_rate/qwen_vl_interface': 9.708836734167351e-06, 'learning_rate/action_model': 9.708836734167351e-05, 'epoch': 0.36}) +[2026-06-17 18:07:48,742][starVLA.training.train_starvla][INFO] - Step 874, Loss: {'train/loss': 0.6581798195838928, 'train/grad_norm_pre_clip': 12.495633125305176, 'timing/data': 0.01837047003209591, 'timing/model': 0.9721922045573592, 'learning_rate/qwen_vl_interface': 9.708090731156321e-06, 'learning_rate/action_model': 9.708090731156321e-05, 'epoch': 0.36}) +[2026-06-17 18:07:52,624][starVLA.training.train_starvla][INFO] - Step 875, Loss: {'train/loss': 0.6540235877037048, 'train/grad_norm_pre_clip': 12.82097339630127, 'timing/data': 0.01823695097118616, 'timing/model': 0.9587743496522307, 'learning_rate/qwen_vl_interface': 9.707343803978288e-06, 'learning_rate/action_model': 9.707343803978288e-05, 'epoch': 0.36}) +[2026-06-17 18:07:56,441][starVLA.training.train_starvla][INFO] - Step 876, Loss: {'train/loss': 0.5690151453018188, 'train/grad_norm_pre_clip': 9.44172191619873, 'timing/data': 0.017755253240466118, 'timing/model': 0.9401169819757342, 'learning_rate/qwen_vl_interface': 9.706595952788088e-06, 'learning_rate/action_model': 9.706595952788088e-05, 'epoch': 0.36}) +[2026-06-17 18:08:00,239][starVLA.training.train_starvla][INFO] - Step 877, Loss: {'train/loss': 0.8033486604690552, 'train/grad_norm_pre_clip': 10.568493843078613, 'timing/data': 0.01649659499526024, 'timing/model': 0.932438038289547, 'learning_rate/qwen_vl_interface': 9.705847177740751e-06, 'learning_rate/action_model': 9.705847177740751e-05, 'epoch': 0.36}) +[2026-06-17 18:08:04,108][starVLA.training.train_starvla][INFO] - Step 878, Loss: {'train/loss': 0.9389031529426575, 'train/grad_norm_pre_clip': 15.933777809143066, 'timing/data': 0.016119503416121006, 'timing/model': 1.0332232555374503, 'learning_rate/qwen_vl_interface': 9.7050974789915e-06, 'learning_rate/action_model': 9.7050974789915e-05, 'epoch': 0.36}) +[2026-06-17 18:08:07,968][starVLA.training.train_starvla][INFO] - Step 879, Loss: {'train/loss': 0.9830306172370911, 'train/grad_norm_pre_clip': 12.583572387695312, 'timing/data': 0.01521807536482811, 'timing/model': 1.1609502825886011, 'learning_rate/qwen_vl_interface': 9.704346856695747e-06, 'learning_rate/action_model': 9.704346856695746e-05, 'epoch': 0.36}) +[2026-06-17 18:08:11,647][starVLA.training.train_starvla][INFO] - Step 880, Loss: {'train/loss': 0.7734421491622925, 'train/grad_norm_pre_clip': 8.10110855102539, 'timing/data': 0.01626688428223133, 'timing/model': 0.9548741355538368, 'learning_rate/qwen_vl_interface': 9.703595311009098e-06, 'learning_rate/action_model': 9.703595311009097e-05, 'epoch': 0.36}) +[2026-06-17 18:08:15,473][starVLA.training.train_starvla][INFO] - Step 881, Loss: {'train/loss': 0.7761653661727905, 'train/grad_norm_pre_clip': 11.342422485351562, 'timing/data': 0.022012196481227875, 'timing/model': 0.9403026662766933, 'learning_rate/qwen_vl_interface': 9.702842842087348e-06, 'learning_rate/action_model': 9.702842842087348e-05, 'epoch': 0.36}) +[2026-06-17 18:08:21,006][starVLA.training.train_starvla][INFO] - Step 882, Loss: {'train/loss': 0.4278463125228882, 'train/grad_norm_pre_clip': 11.151042938232422, 'timing/data': 0.014951576478779316, 'timing/model': 2.1900121737271547, 'learning_rate/qwen_vl_interface': 9.702089450086486e-06, 'learning_rate/action_model': 9.702089450086486e-05, 'epoch': 0.36}) +[2026-06-17 18:08:24,873][starVLA.training.train_starvla][INFO] - Step 883, Loss: {'train/loss': 0.9186257719993591, 'train/grad_norm_pre_clip': 11.091999053955078, 'timing/data': 0.03706845920532942, 'timing/model': 0.944664902985096, 'learning_rate/qwen_vl_interface': 9.70133513516269e-06, 'learning_rate/action_model': 9.701335135162688e-05, 'epoch': 0.36}) +[2026-06-17 18:08:31,661][starVLA.training.train_starvla][INFO] - Step 884, Loss: {'train/loss': 0.7225660681724548, 'train/grad_norm_pre_clip': 11.632505416870117, 'timing/data': 0.02134075202047825, 'timing/model': 0.9345602644607425, 'learning_rate/qwen_vl_interface': 9.70057989747233e-06, 'learning_rate/action_model': 9.700579897472329e-05, 'epoch': 0.36}) +[2026-06-17 18:08:35,572][starVLA.training.train_starvla][INFO] - Step 885, Loss: {'train/loss': 0.5022500157356262, 'train/grad_norm_pre_clip': 7.281116962432861, 'timing/data': 0.01675775647163391, 'timing/model': 1.1168303498998284, 'learning_rate/qwen_vl_interface': 9.699823737171969e-06, 'learning_rate/action_model': 9.699823737171969e-05, 'epoch': 0.36}) +[2026-06-17 18:08:39,439][starVLA.training.train_starvla][INFO] - Step 886, Loss: {'train/loss': 0.7004436254501343, 'train/grad_norm_pre_clip': 19.01494026184082, 'timing/data': 0.017501025460660458, 'timing/model': 0.9390043392777443, 'learning_rate/qwen_vl_interface': 9.699066654418358e-06, 'learning_rate/action_model': 9.699066654418358e-05, 'epoch': 0.36}) +[2026-06-17 18:08:44,231][starVLA.training.train_starvla][INFO] - Step 887, Loss: {'train/loss': 0.5625098943710327, 'train/grad_norm_pre_clip': 9.929805755615234, 'timing/data': 0.020304279401898384, 'timing/model': 1.730050046928227, 'learning_rate/qwen_vl_interface': 9.698308649368444e-06, 'learning_rate/action_model': 9.698308649368444e-05, 'epoch': 0.37}) +[2026-06-17 18:08:48,185][starVLA.training.train_starvla][INFO] - Step 888, Loss: {'train/loss': 1.0128384828567505, 'train/grad_norm_pre_clip': 10.150542259216309, 'timing/data': 0.028754360042512417, 'timing/model': 0.9469203846529126, 'learning_rate/qwen_vl_interface': 9.69754972217936e-06, 'learning_rate/action_model': 9.69754972217936e-05, 'epoch': 0.37}) +[2026-06-17 18:08:52,040][starVLA.training.train_starvla][INFO] - Step 889, Loss: {'train/loss': 0.5437558889389038, 'train/grad_norm_pre_clip': 10.225237846374512, 'timing/data': 0.01319844089448452, 'timing/model': 0.9414539048448205, 'learning_rate/qwen_vl_interface': 9.696789873008432e-06, 'learning_rate/action_model': 9.696789873008433e-05, 'epoch': 0.37}) +[2026-06-17 18:08:55,942][starVLA.training.train_starvla][INFO] - Step 890, Loss: {'train/loss': 1.0205556154251099, 'train/grad_norm_pre_clip': 9.894064903259277, 'timing/data': 0.018854103982448578, 'timing/model': 0.9493936933577061, 'learning_rate/qwen_vl_interface': 9.69602910201318e-06, 'learning_rate/action_model': 9.696029102013179e-05, 'epoch': 0.37}) +[2026-06-17 18:09:00,944][starVLA.training.train_starvla][INFO] - Step 891, Loss: {'train/loss': 0.6486411094665527, 'train/grad_norm_pre_clip': 8.748960494995117, 'timing/data': 0.026438375934958458, 'timing/model': 1.0183314848691225, 'learning_rate/qwen_vl_interface': 9.695267409351308e-06, 'learning_rate/action_model': 9.695267409351307e-05, 'epoch': 0.37}) +[2026-06-17 18:09:05,153][starVLA.training.train_starvla][INFO] - Step 892, Loss: {'train/loss': 0.540367841720581, 'train/grad_norm_pre_clip': 8.701358795166016, 'timing/data': 0.0182036105543375, 'timing/model': 1.2217423655092716, 'learning_rate/qwen_vl_interface': 9.694504795180722e-06, 'learning_rate/action_model': 9.694504795180721e-05, 'epoch': 0.37}) +[2026-06-17 18:09:09,780][starVLA.training.train_starvla][INFO] - Step 893, Loss: {'train/loss': 0.6932492256164551, 'train/grad_norm_pre_clip': 10.864858627319336, 'timing/data': 0.023112338036298752, 'timing/model': 1.0069754952564836, 'learning_rate/qwen_vl_interface': 9.693741259659507e-06, 'learning_rate/action_model': 9.693741259659507e-05, 'epoch': 0.37}) +[2026-06-17 18:09:14,028][starVLA.training.train_starvla][INFO] - Step 894, Loss: {'train/loss': 0.7600285410881042, 'train/grad_norm_pre_clip': 8.978915214538574, 'timing/data': 0.01638460624963045, 'timing/model': 1.012051334604621, 'learning_rate/qwen_vl_interface': 9.692976802945947e-06, 'learning_rate/action_model': 9.692976802945948e-05, 'epoch': 0.37}) +[2026-06-17 18:09:18,044][starVLA.training.train_starvla][INFO] - Step 895, Loss: {'train/loss': 0.5370109677314758, 'train/grad_norm_pre_clip': 5.8483452796936035, 'timing/data': 0.021228873170912266, 'timing/model': 1.0173194846138358, 'learning_rate/qwen_vl_interface': 9.692211425198517e-06, 'learning_rate/action_model': 9.692211425198517e-05, 'epoch': 0.37}) +[2026-06-17 18:09:24,147][starVLA.training.train_starvla][INFO] - Step 896, Loss: {'train/loss': 0.8592706918716431, 'train/grad_norm_pre_clip': 14.701906204223633, 'timing/data': 0.015060251578688622, 'timing/model': 0.9383509987965226, 'learning_rate/qwen_vl_interface': 9.691445126575879e-06, 'learning_rate/action_model': 9.691445126575877e-05, 'epoch': 0.37}) +[2026-06-17 18:09:28,248][starVLA.training.train_starvla][INFO] - Step 897, Loss: {'train/loss': 0.8225535154342651, 'train/grad_norm_pre_clip': 8.846359252929688, 'timing/data': 0.019868810661137104, 'timing/model': 1.1895382925868034, 'learning_rate/qwen_vl_interface': 9.690677907236884e-06, 'learning_rate/action_model': 9.690677907236884e-05, 'epoch': 0.37}) +[2026-06-17 18:09:33,258][starVLA.training.train_starvla][INFO] - Step 898, Loss: {'train/loss': 0.793616533279419, 'train/grad_norm_pre_clip': 16.46549415588379, 'timing/data': 0.01522229053080082, 'timing/model': 1.2472057966515422, 'learning_rate/qwen_vl_interface': 9.689909767340582e-06, 'learning_rate/action_model': 9.689909767340583e-05, 'epoch': 0.37}) +[2026-06-17 18:09:37,122][starVLA.training.train_starvla][INFO] - Step 899, Loss: {'train/loss': 0.7610236406326294, 'train/grad_norm_pre_clip': 8.218768119812012, 'timing/data': 0.016425677575170994, 'timing/model': 1.1519397525116801, 'learning_rate/qwen_vl_interface': 9.689140707046209e-06, 'learning_rate/action_model': 9.689140707046209e-05, 'epoch': 0.37}) +[2026-06-17 18:11:34,561][starVLA.training.train_starvla][INFO] - Step 900, Loss: {'train/loss': 0.7351146936416626, 'train/grad_norm_pre_clip': 10.448846817016602, 'eval/loss': 0.7445945739746094, 'eval/latency_6/loss': 0.7445945739746094, 'eval/demon_fire_precision': 0.9317544406357121, 'eval/demon_fire_recall': 0.9663865546218487, 'eval/demon_fire_f1': 0.9487545613200064, 'eval/demon_fire_rate_model': 0.8928770172509738, 'eval/demon_fire_rate_teacher': 0.860879243183083, 'eval/demon_left_precision': 0.8391959798994975, 'eval/demon_left_recall': 0.2929824561403509, 'eval/demon_left_f1': 0.4343302990897269, 'eval/demon_left_rate_model': 0.055370061213133, 'eval/demon_left_rate_teacher': 0.15859766277128548, 'eval/demon_right_precision': 0.7087542087542088, 'eval/demon_right_recall': 0.6466973886328725, 'eval/demon_right_f1': 0.6763052208835342, 'eval/demon_right_rate_model': 0.1652754590984975, 'eval/demon_right_rate_teacher': 0.18113522537562604, 'eval/demon_move_f1': 0.5553177599866306, 'eval/demon_cc_f1': 0.7824477558537213, 'eval/latency_6/demon_fire_precision': 0.9317544406357121, 'eval/latency_6/demon_fire_recall': 0.9663865546218487, 'eval/latency_6/demon_fire_f1': 0.9487545613200064, 'eval/latency_6/demon_fire_rate_model': 0.8928770172509738, 'eval/latency_6/demon_fire_rate_teacher': 0.860879243183083, 'eval/latency_6/demon_left_precision': 0.8391959798994975, 'eval/latency_6/demon_left_recall': 0.2929824561403509, 'eval/latency_6/demon_left_f1': 0.4343302990897269, 'eval/latency_6/demon_left_rate_model': 0.055370061213133, 'eval/latency_6/demon_left_rate_teacher': 0.15859766277128548, 'eval/latency_6/demon_right_precision': 0.7087542087542088, 'eval/latency_6/demon_right_recall': 0.6466973886328725, 'eval/latency_6/demon_right_f1': 0.6763052208835342, 'eval/latency_6/demon_right_rate_model': 0.1652754590984975, 'eval/latency_6/demon_right_rate_teacher': 0.18113522537562604, 'eval/latency_6/demon_move_f1': 0.5553177599866306, 'eval/latency_6/demon_cc_f1': 0.7824477558537213, 'timing/data': 0.018176286481320858, 'timing/model': 0.9368613790720701, 'learning_rate/qwen_vl_interface': 9.68837072651319e-06, 'learning_rate/action_model': 9.688370726513188e-05, 'epoch': 0.37}) +[2026-06-17 18:11:34,562][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_900_state +[2026-06-17 18:11:34,583][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 18:11:42,553][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_900_state/model.safetensors +[2026-06-17 18:11:57,282][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_900_state/optimizer.bin +[2026-06-17 18:11:57,283][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_900_state/sampler.bin +[2026-06-17 18:11:57,283][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_900_state/sampler_1.bin +[2026-06-17 18:11:57,286][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_900_state/random_states_0.pkl +[2026-06-17 18:11:57,290][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 900 +[2026-06-17 18:11:57,290][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 18:11:57,306][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 18:12:01,277][starVLA.training.train_starvla][INFO] - Step 901, Loss: {'train/loss': 0.568428099155426, 'train/grad_norm_pre_clip': 8.004034996032715, 'timing/data': 0.01443327683955431, 'timing/model': 1.1737211858853698, 'learning_rate/qwen_vl_interface': 9.687599825901141e-06, 'learning_rate/action_model': 9.687599825901142e-05, 'epoch': 0.37}) +[2026-06-17 18:12:05,801][starVLA.training.train_starvla][INFO] - Step 902, Loss: {'train/loss': 0.986903965473175, 'train/grad_norm_pre_clip': 9.770370483398438, 'timing/data': 0.016959370113909245, 'timing/model': 0.9541524117812514, 'learning_rate/qwen_vl_interface': 9.686828005369876e-06, 'learning_rate/action_model': 9.686828005369875e-05, 'epoch': 0.37}) +[2026-06-17 18:12:09,711][starVLA.training.train_starvla][INFO] - Step 903, Loss: {'train/loss': 0.7023460268974304, 'train/grad_norm_pre_clip': 6.6711530685424805, 'timing/data': 0.018582328222692013, 'timing/model': 0.9492407273501158, 'learning_rate/qwen_vl_interface': 9.68605526507939e-06, 'learning_rate/action_model': 9.686055265079389e-05, 'epoch': 0.37}) +[2026-06-17 18:12:13,536][starVLA.training.train_starvla][INFO] - Step 904, Loss: {'train/loss': 0.5968084931373596, 'train/grad_norm_pre_clip': 11.96800708770752, 'timing/data': 0.021515564993023872, 'timing/model': 0.9371034363284707, 'learning_rate/qwen_vl_interface': 9.685281605189874e-06, 'learning_rate/action_model': 9.685281605189874e-05, 'epoch': 0.37}) +[2026-06-17 18:12:17,545][starVLA.training.train_starvla][INFO] - Step 905, Loss: {'train/loss': 0.4347188472747803, 'train/grad_norm_pre_clip': 15.527525901794434, 'timing/data': 0.01641830988228321, 'timing/model': 0.9362266967073083, 'learning_rate/qwen_vl_interface': 9.68450702586171e-06, 'learning_rate/action_model': 9.684507025861709e-05, 'epoch': 0.37}) +[2026-06-17 18:12:21,394][starVLA.training.train_starvla][INFO] - Step 906, Loss: {'train/loss': 0.9928750395774841, 'train/grad_norm_pre_clip': 14.014971733093262, 'timing/data': 0.01762774959206581, 'timing/model': 0.9493658961728215, 'learning_rate/qwen_vl_interface': 9.683731527255467e-06, 'learning_rate/action_model': 9.683731527255467e-05, 'epoch': 0.37}) +[2026-06-17 18:12:25,359][starVLA.training.train_starvla][INFO] - Step 907, Loss: {'train/loss': 0.6723691821098328, 'train/grad_norm_pre_clip': 10.613882064819336, 'timing/data': 0.02292601950466633, 'timing/model': 1.1672036154195666, 'learning_rate/qwen_vl_interface': 9.682955109531906e-06, 'learning_rate/action_model': 9.682955109531905e-05, 'epoch': 0.37}) +[2026-06-17 18:12:29,210][starVLA.training.train_starvla][INFO] - Step 908, Loss: {'train/loss': 0.6255540251731873, 'train/grad_norm_pre_clip': 9.173934936523438, 'timing/data': 0.01800430752336979, 'timing/model': 1.0676252599805593, 'learning_rate/qwen_vl_interface': 9.682177772851983e-06, 'learning_rate/action_model': 9.682177772851982e-05, 'epoch': 0.37}) +[2026-06-17 18:12:32,844][starVLA.training.train_starvla][INFO] - Step 909, Loss: {'train/loss': 1.1129968166351318, 'train/grad_norm_pre_clip': 13.030426979064941, 'timing/data': 0.026350381784141064, 'timing/model': 0.9452719744294882, 'learning_rate/qwen_vl_interface': 9.681399517376837e-06, 'learning_rate/action_model': 9.681399517376837e-05, 'epoch': 0.37}) +[2026-06-17 18:12:36,700][starVLA.training.train_starvla][INFO] - Step 910, Loss: {'train/loss': 0.8061583638191223, 'train/grad_norm_pre_clip': 12.513343811035156, 'timing/data': 0.023287583142518997, 'timing/model': 0.95550502743572, 'learning_rate/qwen_vl_interface': 9.680620343267803e-06, 'learning_rate/action_model': 9.680620343267802e-05, 'epoch': 0.37}) +[2026-06-17 18:12:40,573][starVLA.training.train_starvla][INFO] - Step 911, Loss: {'train/loss': 0.9676247835159302, 'train/grad_norm_pre_clip': 11.888395309448242, 'timing/data': 0.018196153454482555, 'timing/model': 0.9387593809515238, 'learning_rate/qwen_vl_interface': 9.679840250686403e-06, 'learning_rate/action_model': 9.679840250686404e-05, 'epoch': 0.38}) +[2026-06-17 18:12:44,396][starVLA.training.train_starvla][INFO] - Step 912, Loss: {'train/loss': 0.7680978178977966, 'train/grad_norm_pre_clip': 8.285584449768066, 'timing/data': 0.02292205300182104, 'timing/model': 0.941450516693294, 'learning_rate/qwen_vl_interface': 9.679059239794355e-06, 'learning_rate/action_model': 9.679059239794355e-05, 'epoch': 0.38}) +[2026-06-17 18:12:48,528][starVLA.training.train_starvla][INFO] - Step 913, Loss: {'train/loss': 0.8937594294548035, 'train/grad_norm_pre_clip': 94.19886016845703, 'timing/data': 0.018781875260174274, 'timing/model': 1.0618637483567, 'learning_rate/qwen_vl_interface': 9.67827731075356e-06, 'learning_rate/action_model': 9.67827731075356e-05, 'epoch': 0.38}) +[2026-06-17 18:12:52,330][starVLA.training.train_starvla][INFO] - Step 914, Loss: {'train/loss': 0.4305342733860016, 'train/grad_norm_pre_clip': 10.839999198913574, 'timing/data': 0.015311246737837791, 'timing/model': 1.1243444178253412, 'learning_rate/qwen_vl_interface': 9.677494463726112e-06, 'learning_rate/action_model': 9.677494463726112e-05, 'epoch': 0.38}) +[2026-06-17 18:12:56,135][starVLA.training.train_starvla][INFO] - Step 915, Loss: {'train/loss': 0.9538570642471313, 'train/grad_norm_pre_clip': 13.254878044128418, 'timing/data': 0.018799525685608387, 'timing/model': 1.0942498268559575, 'learning_rate/qwen_vl_interface': 9.676710698874299e-06, 'learning_rate/action_model': 9.676710698874298e-05, 'epoch': 0.38}) +[2026-06-17 18:12:59,741][starVLA.training.train_starvla][INFO] - Step 916, Loss: {'train/loss': 0.8283867239952087, 'train/grad_norm_pre_clip': 11.64975357055664, 'timing/data': 0.01542660966515541, 'timing/model': 0.9261804549023509, 'learning_rate/qwen_vl_interface': 9.675926016360595e-06, 'learning_rate/action_model': 9.675926016360595e-05, 'epoch': 0.38}) +[2026-06-17 18:13:03,958][starVLA.training.train_starvla][INFO] - Step 917, Loss: {'train/loss': 0.750156581401825, 'train/grad_norm_pre_clip': 11.489758491516113, 'timing/data': 0.02251648809760809, 'timing/model': 0.948543238453567, 'learning_rate/qwen_vl_interface': 9.675140416347664e-06, 'learning_rate/action_model': 9.675140416347663e-05, 'epoch': 0.38}) +[2026-06-17 18:13:07,941][starVLA.training.train_starvla][INFO] - Step 918, Loss: {'train/loss': 0.8976112008094788, 'train/grad_norm_pre_clip': 10.237375259399414, 'timing/data': 0.02542963344603777, 'timing/model': 0.9412012491375208, 'learning_rate/qwen_vl_interface': 9.674353898998362e-06, 'learning_rate/action_model': 9.674353898998363e-05, 'epoch': 0.38}) +[2026-06-17 18:13:11,753][starVLA.training.train_starvla][INFO] - Step 919, Loss: {'train/loss': 0.7210477590560913, 'train/grad_norm_pre_clip': 8.400870323181152, 'timing/data': 0.014327778480947018, 'timing/model': 0.9467678954824805, 'learning_rate/qwen_vl_interface': 9.673566464475738e-06, 'learning_rate/action_model': 9.673566464475739e-05, 'epoch': 0.38}) +[2026-06-17 18:13:15,661][starVLA.training.train_starvla][INFO] - Step 920, Loss: {'train/loss': 0.8543732762336731, 'train/grad_norm_pre_clip': 12.486263275146484, 'timing/data': 0.023657124489545822, 'timing/model': 0.991740845143795, 'learning_rate/qwen_vl_interface': 9.672778112943026e-06, 'learning_rate/action_model': 9.672778112943026e-05, 'epoch': 0.38}) +[2026-06-17 18:13:19,573][starVLA.training.train_starvla][INFO] - Step 921, Loss: {'train/loss': 0.719300389289856, 'train/grad_norm_pre_clip': 7.989875793457031, 'timing/data': 0.02392469346523285, 'timing/model': 1.1831365451216698, 'learning_rate/qwen_vl_interface': 9.671988844563652e-06, 'learning_rate/action_model': 9.671988844563652e-05, 'epoch': 0.38}) +[2026-06-17 18:13:23,450][starVLA.training.train_starvla][INFO] - Step 922, Loss: {'train/loss': 0.9136157035827637, 'train/grad_norm_pre_clip': 14.722905158996582, 'timing/data': 0.01656433381140232, 'timing/model': 1.0471199648454785, 'learning_rate/qwen_vl_interface': 9.671198659501233e-06, 'learning_rate/action_model': 9.671198659501231e-05, 'epoch': 0.38}) +[2026-06-17 18:13:27,410][starVLA.training.train_starvla][INFO] - Step 923, Loss: {'train/loss': 0.7724889516830444, 'train/grad_norm_pre_clip': 7.774807453155518, 'timing/data': 0.022860145196318626, 'timing/model': 0.9935906641185284, 'learning_rate/qwen_vl_interface': 9.670407557919573e-06, 'learning_rate/action_model': 9.670407557919572e-05, 'epoch': 0.38}) +[2026-06-17 18:13:31,286][starVLA.training.train_starvla][INFO] - Step 924, Loss: {'train/loss': 1.0523741245269775, 'train/grad_norm_pre_clip': 15.53705883026123, 'timing/data': 0.02186738047748804, 'timing/model': 1.0077434284612536, 'learning_rate/qwen_vl_interface': 9.669615539982668e-06, 'learning_rate/action_model': 9.669615539982668e-05, 'epoch': 0.38}) +[2026-06-17 18:13:35,354][starVLA.training.train_starvla][INFO] - Step 925, Loss: {'train/loss': 0.37579765915870667, 'train/grad_norm_pre_clip': 6.922720909118652, 'timing/data': 0.019623725675046444, 'timing/model': 0.9361021658405662, 'learning_rate/qwen_vl_interface': 9.668822605854708e-06, 'learning_rate/action_model': 9.668822605854706e-05, 'epoch': 0.38}) +[2026-06-17 18:13:39,329][starVLA.training.train_starvla][INFO] - Step 926, Loss: {'train/loss': 0.9871515035629272, 'train/grad_norm_pre_clip': 8.365154266357422, 'timing/data': 0.02182089537382126, 'timing/model': 0.9523599920794368, 'learning_rate/qwen_vl_interface': 9.668028755700067e-06, 'learning_rate/action_model': 9.668028755700067e-05, 'epoch': 0.38}) +[2026-06-17 18:13:43,204][starVLA.training.train_starvla][INFO] - Step 927, Loss: {'train/loss': 0.5997859835624695, 'train/grad_norm_pre_clip': 23.129711151123047, 'timing/data': 0.020030046813189983, 'timing/model': 1.1777294194325805, 'learning_rate/qwen_vl_interface': 9.66723398968331e-06, 'learning_rate/action_model': 9.66723398968331e-05, 'epoch': 0.38}) +[2026-06-17 18:13:46,942][starVLA.training.train_starvla][INFO] - Step 928, Loss: {'train/loss': 0.641811192035675, 'train/grad_norm_pre_clip': 9.198023796081543, 'timing/data': 0.02678620722144842, 'timing/model': 1.0365371480584145, 'learning_rate/qwen_vl_interface': 9.666438307969194e-06, 'learning_rate/action_model': 9.666438307969195e-05, 'epoch': 0.38}) +[2026-06-17 18:13:50,643][starVLA.training.train_starvla][INFO] - Step 929, Loss: {'train/loss': 0.7298482060432434, 'train/grad_norm_pre_clip': 10.56368350982666, 'timing/data': 0.018788671121001244, 'timing/model': 0.9696056097745895, 'learning_rate/qwen_vl_interface': 9.665641710722666e-06, 'learning_rate/action_model': 9.665641710722665e-05, 'epoch': 0.38}) +[2026-06-17 18:13:54,545][starVLA.training.train_starvla][INFO] - Step 930, Loss: {'train/loss': 0.8711655139923096, 'train/grad_norm_pre_clip': 10.005436897277832, 'timing/data': 0.022562561556696892, 'timing/model': 0.941136859357357, 'learning_rate/qwen_vl_interface': 9.664844198108858e-06, 'learning_rate/action_model': 9.664844198108858e-05, 'epoch': 0.38}) +[2026-06-17 18:13:58,458][starVLA.training.train_starvla][INFO] - Step 931, Loss: {'train/loss': 1.0131396055221558, 'train/grad_norm_pre_clip': 12.832566261291504, 'timing/data': 0.016121291555464268, 'timing/model': 0.9441998340189457, 'learning_rate/qwen_vl_interface': 9.6640457702931e-06, 'learning_rate/action_model': 9.664045770293099e-05, 'epoch': 0.38}) +[2026-06-17 18:14:02,404][starVLA.training.train_starvla][INFO] - Step 932, Loss: {'train/loss': 0.5950077176094055, 'train/grad_norm_pre_clip': 257.5274658203125, 'timing/data': 0.017912350594997406, 'timing/model': 1.02949932590127, 'learning_rate/qwen_vl_interface': 9.663246427440901e-06, 'learning_rate/action_model': 9.6632464274409e-05, 'epoch': 0.38}) +[2026-06-17 18:14:06,259][starVLA.training.train_starvla][INFO] - Step 933, Loss: {'train/loss': 0.7382846474647522, 'train/grad_norm_pre_clip': 8.79119873046875, 'timing/data': 0.016412653028964996, 'timing/model': 0.9643638534471393, 'learning_rate/qwen_vl_interface': 9.662446169717969e-06, 'learning_rate/action_model': 9.662446169717969e-05, 'epoch': 0.38}) +[2026-06-17 18:14:10,171][starVLA.training.train_starvla][INFO] - Step 934, Loss: {'train/loss': 0.5890690684318542, 'train/grad_norm_pre_clip': 7.81870698928833, 'timing/data': 0.0005164649337530136, 'timing/model': 1.150062009692192, 'learning_rate/qwen_vl_interface': 9.6616449972902e-06, 'learning_rate/action_model': 9.6616449972902e-05, 'epoch': 0.38}) +[2026-06-17 18:14:13,856][starVLA.training.train_starvla][INFO] - Step 935, Loss: {'train/loss': 0.651365339756012, 'train/grad_norm_pre_clip': 25.58650016784668, 'timing/data': 0.0008667269721627235, 'timing/model': 0.951935950666666, 'learning_rate/qwen_vl_interface': 9.660842910323675e-06, 'learning_rate/action_model': 9.660842910323674e-05, 'epoch': 0.38}) +[2026-06-17 18:14:18,031][starVLA.training.train_starvla][INFO] - Step 936, Loss: {'train/loss': 0.8966015577316284, 'train/grad_norm_pre_clip': 13.778403282165527, 'timing/data': 0.0012839259579777718, 'timing/model': 1.0108947847038507, 'learning_rate/qwen_vl_interface': 9.660039908984667e-06, 'learning_rate/action_model': 9.660039908984666e-05, 'epoch': 0.39}) +[2026-06-17 18:14:22,147][starVLA.training.train_starvla][INFO] - Step 937, Loss: {'train/loss': 0.6560455560684204, 'train/grad_norm_pre_clip': 11.906220436096191, 'timing/data': 0.0005061943084001541, 'timing/model': 1.076877074316144, 'learning_rate/qwen_vl_interface': 9.659235993439643e-06, 'learning_rate/action_model': 9.659235993439643e-05, 'epoch': 0.39}) +[2026-06-17 18:14:26,091][starVLA.training.train_starvla][INFO] - Step 938, Loss: {'train/loss': 0.8195801377296448, 'train/grad_norm_pre_clip': 11.877161979675293, 'timing/data': 0.0004993230104446411, 'timing/model': 0.9500771248713136, 'learning_rate/qwen_vl_interface': 9.658431163855251e-06, 'learning_rate/action_model': 9.658431163855251e-05, 'epoch': 0.39}) +[2026-06-17 18:14:30,018][starVLA.training.train_starvla][INFO] - Step 939, Loss: {'train/loss': 0.6202020645141602, 'train/grad_norm_pre_clip': 17.26495361328125, 'timing/data': 0.0004548998549580574, 'timing/model': 0.9582319604232907, 'learning_rate/qwen_vl_interface': 9.657625420398334e-06, 'learning_rate/action_model': 9.657625420398333e-05, 'epoch': 0.39}) +[2026-06-17 18:14:33,918][starVLA.training.train_starvla][INFO] - Step 940, Loss: {'train/loss': 1.0138808488845825, 'train/grad_norm_pre_clip': 11.820890426635742, 'timing/data': 0.0013037072494626045, 'timing/model': 1.1666359016671777, 'learning_rate/qwen_vl_interface': 9.656818763235927e-06, 'learning_rate/action_model': 9.656818763235926e-05, 'epoch': 0.39}) +[2026-06-17 18:14:37,651][starVLA.training.train_starvla][INFO] - Step 941, Loss: {'train/loss': 0.9714383482933044, 'train/grad_norm_pre_clip': 8.898823738098145, 'timing/data': 0.0006625363603234291, 'timing/model': 0.9433756144717336, 'learning_rate/qwen_vl_interface': 9.656011192535248e-06, 'learning_rate/action_model': 9.656011192535247e-05, 'epoch': 0.39}) +[2026-06-17 18:14:42,484][starVLA.training.train_starvla][INFO] - Step 942, Loss: {'train/loss': 0.9130679368972778, 'train/grad_norm_pre_clip': 8.429553985595703, 'timing/data': 0.000496160238981247, 'timing/model': 1.007197704166174, 'learning_rate/qwen_vl_interface': 9.655202708463706e-06, 'learning_rate/action_model': 9.655202708463707e-05, 'epoch': 0.39}) +[2026-06-17 18:14:46,364][starVLA.training.train_starvla][INFO] - Step 943, Loss: {'train/loss': 0.9871572852134705, 'train/grad_norm_pre_clip': 12.255219459533691, 'timing/data': 0.0006426321342587471, 'timing/model': 0.9433912634849548, 'learning_rate/qwen_vl_interface': 9.654393311188905e-06, 'learning_rate/action_model': 9.654393311188905e-05, 'epoch': 0.39}) +[2026-06-17 18:14:50,717][starVLA.training.train_starvla][INFO] - Step 944, Loss: {'train/loss': 0.853069007396698, 'train/grad_norm_pre_clip': 11.09141731262207, 'timing/data': 0.0004653939977288246, 'timing/model': 0.9423124697059393, 'learning_rate/qwen_vl_interface': 9.653583000878631e-06, 'learning_rate/action_model': 9.65358300087863e-05, 'epoch': 0.39}) +[2026-06-17 18:14:54,889][starVLA.training.train_starvla][INFO] - Step 945, Loss: {'train/loss': 0.7847539782524109, 'train/grad_norm_pre_clip': 16.150081634521484, 'timing/data': 0.0004490138962864876, 'timing/model': 0.9399542305618525, 'learning_rate/qwen_vl_interface': 9.65277177770086e-06, 'learning_rate/action_model': 9.65277177770086e-05, 'epoch': 0.39}) +[2026-06-17 18:14:58,705][starVLA.training.train_starvla][INFO] - Step 946, Loss: {'train/loss': 0.8654099702835083, 'train/grad_norm_pre_clip': 13.885480880737305, 'timing/data': 0.0004849722608923912, 'timing/model': 0.9426520122215152, 'learning_rate/qwen_vl_interface': 9.651959641823764e-06, 'learning_rate/action_model': 9.651959641823763e-05, 'epoch': 0.39}) +[2026-06-17 18:15:02,603][starVLA.training.train_starvla][INFO] - Step 947, Loss: {'train/loss': 0.7793090343475342, 'train/grad_norm_pre_clip': 11.016337394714355, 'timing/data': 0.0004874998703598976, 'timing/model': 1.2178027722984552, 'learning_rate/qwen_vl_interface': 9.651146593415696e-06, 'learning_rate/action_model': 9.651146593415695e-05, 'epoch': 0.39}) +[2026-06-17 18:15:06,407][starVLA.training.train_starvla][INFO] - Step 948, Loss: {'train/loss': 0.9127688407897949, 'train/grad_norm_pre_clip': 17.14642906188965, 'timing/data': 0.0006097638979554176, 'timing/model': 1.035002144984901, 'learning_rate/qwen_vl_interface': 9.650332632645206e-06, 'learning_rate/action_model': 9.650332632645205e-05, 'epoch': 0.39}) +[2026-06-17 18:15:13,737][starVLA.training.train_starvla][INFO] - Step 949, Loss: {'train/loss': 0.8272178769111633, 'train/grad_norm_pre_clip': 10.664529800415039, 'timing/data': 0.0012014005333185196, 'timing/model': 1.3013419285416603, 'learning_rate/qwen_vl_interface': 9.649517759681022e-06, 'learning_rate/action_model': 9.649517759681023e-05, 'epoch': 0.39}) +[2026-06-17 18:15:17,857][starVLA.training.train_starvla][INFO] - Step 950, Loss: {'train/loss': 0.8369829058647156, 'train/grad_norm_pre_clip': 8.160425186157227, 'timing/data': 0.000858500599861145, 'timing/model': 1.0030627213418484, 'learning_rate/qwen_vl_interface': 9.648701974692075e-06, 'learning_rate/action_model': 9.648701974692075e-05, 'epoch': 0.39}) +[2026-06-17 18:15:21,696][starVLA.training.train_starvla][INFO] - Step 951, Loss: {'train/loss': 0.843249499797821, 'train/grad_norm_pre_clip': 9.631879806518555, 'timing/data': 0.00048810988664627075, 'timing/model': 0.9418292893096805, 'learning_rate/qwen_vl_interface': 9.647885277847476e-06, 'learning_rate/action_model': 9.647885277847476e-05, 'epoch': 0.39}) +[2026-06-17 18:15:25,511][starVLA.training.train_starvla][INFO] - Step 952, Loss: {'train/loss': 0.6945482492446899, 'train/grad_norm_pre_clip': 8.902942657470703, 'timing/data': 0.0008511887863278389, 'timing/model': 0.9510119231417775, 'learning_rate/qwen_vl_interface': 9.647067669316525e-06, 'learning_rate/action_model': 9.647067669316525e-05, 'epoch': 0.39}) +[2026-06-17 18:15:29,382][starVLA.training.train_starvla][INFO] - Step 953, Loss: {'train/loss': 0.8991311192512512, 'train/grad_norm_pre_clip': 11.29459285736084, 'timing/data': 0.0007234355434775352, 'timing/model': 1.1933255214244127, 'learning_rate/qwen_vl_interface': 9.646249149268716e-06, 'learning_rate/action_model': 9.646249149268715e-05, 'epoch': 0.39}) +[2026-06-17 18:15:33,161][starVLA.training.train_starvla][INFO] - Step 954, Loss: {'train/loss': 0.504940390586853, 'train/grad_norm_pre_clip': 6.678740978240967, 'timing/data': 0.0004863971844315529, 'timing/model': 1.1111999172717333, 'learning_rate/qwen_vl_interface': 9.645429717873728e-06, 'learning_rate/action_model': 9.645429717873727e-05, 'epoch': 0.39}) +[2026-06-17 18:15:36,832][starVLA.training.train_starvla][INFO] - Step 955, Loss: {'train/loss': 0.8921528458595276, 'train/grad_norm_pre_clip': 9.292696952819824, 'timing/data': 0.0011485572904348373, 'timing/model': 0.9410356311127543, 'learning_rate/qwen_vl_interface': 9.644609375301429e-06, 'learning_rate/action_model': 9.64460937530143e-05, 'epoch': 0.39}) +[2026-06-17 18:15:40,702][starVLA.training.train_starvla][INFO] - Step 956, Loss: {'train/loss': 0.9792156219482422, 'train/grad_norm_pre_clip': 10.099376678466797, 'timing/data': 0.0005127331241965294, 'timing/model': 0.94999626185745, 'learning_rate/qwen_vl_interface': 9.643788121721877e-06, 'learning_rate/action_model': 9.643788121721876e-05, 'epoch': 0.39}) +[2026-06-17 18:15:44,557][starVLA.training.train_starvla][INFO] - Step 957, Loss: {'train/loss': 0.7717302441596985, 'train/grad_norm_pre_clip': 9.987099647521973, 'timing/data': 0.00046565383672714233, 'timing/model': 0.9342250442132354, 'learning_rate/qwen_vl_interface': 9.64296595730532e-06, 'learning_rate/action_model': 9.64296595730532e-05, 'epoch': 0.39}) +[2026-06-17 18:15:48,497][starVLA.training.train_starvla][INFO] - Step 958, Loss: {'train/loss': 0.6290750503540039, 'train/grad_norm_pre_clip': 9.256620407104492, 'timing/data': 0.0009432509541511536, 'timing/model': 0.9555375194177032, 'learning_rate/qwen_vl_interface': 9.642142882222193e-06, 'learning_rate/action_model': 9.642142882222193e-05, 'epoch': 0.39}) +[2026-06-17 18:15:52,382][starVLA.training.train_starvla][INFO] - Step 959, Loss: {'train/loss': 0.7899699807167053, 'train/grad_norm_pre_clip': 8.4430513381958, 'timing/data': 0.00046674441546201706, 'timing/model': 0.9613363286480308, 'learning_rate/qwen_vl_interface': 9.641318896643123e-06, 'learning_rate/action_model': 9.641318896643122e-05, 'epoch': 0.39}) +[2026-06-17 18:15:56,387][starVLA.training.train_starvla][INFO] - Step 960, Loss: {'train/loss': 0.8901194930076599, 'train/grad_norm_pre_clip': 19.34041976928711, 'timing/data': 0.0009664865210652351, 'timing/model': 1.1868073754012585, 'learning_rate/qwen_vl_interface': 9.640494000738917e-06, 'learning_rate/action_model': 9.640494000738917e-05, 'epoch': 0.4}) +[2026-06-17 18:16:00,726][starVLA.training.train_starvla][INFO] - Step 961, Loss: {'train/loss': 0.8514004349708557, 'train/grad_norm_pre_clip': 8.332075119018555, 'timing/data': 0.0007410142570734024, 'timing/model': 1.1494186474010348, 'learning_rate/qwen_vl_interface': 9.63966819468058e-06, 'learning_rate/action_model': 9.63966819468058e-05, 'epoch': 0.4}) +[2026-06-17 18:16:04,434][starVLA.training.train_starvla][INFO] - Step 962, Loss: {'train/loss': 0.7251708507537842, 'train/grad_norm_pre_clip': 8.30591106414795, 'timing/data': 0.0010095452889800072, 'timing/model': 0.9571775989606977, 'learning_rate/qwen_vl_interface': 9.638841478639306e-06, 'learning_rate/action_model': 9.638841478639306e-05, 'epoch': 0.4}) +[2026-06-17 18:16:08,354][starVLA.training.train_starvla][INFO] - Step 963, Loss: {'train/loss': 0.634480893611908, 'train/grad_norm_pre_clip': 8.229684829711914, 'timing/data': 0.0018868762999773026, 'timing/model': 0.9445839459076524, 'learning_rate/qwen_vl_interface': 9.638013852786468e-06, 'learning_rate/action_model': 9.638013852786467e-05, 'epoch': 0.4}) +[2026-06-17 18:16:12,423][starVLA.training.train_starvla][INFO] - Step 964, Loss: {'train/loss': 0.9003153443336487, 'train/grad_norm_pre_clip': 11.185465812683105, 'timing/data': 0.000521521084010601, 'timing/model': 0.9582988796755672, 'learning_rate/qwen_vl_interface': 9.637185317293638e-06, 'learning_rate/action_model': 9.637185317293638e-05, 'epoch': 0.4}) +[2026-06-17 18:16:16,308][starVLA.training.train_starvla][INFO] - Step 965, Loss: {'train/loss': 0.6834059357643127, 'train/grad_norm_pre_clip': 10.838552474975586, 'timing/data': 0.0005311379209160805, 'timing/model': 0.9426538022235036, 'learning_rate/qwen_vl_interface': 9.63635587233257e-06, 'learning_rate/action_model': 9.63635587233257e-05, 'epoch': 0.4}) +[2026-06-17 18:16:20,131][starVLA.training.train_starvla][INFO] - Step 966, Loss: {'train/loss': 0.8975211977958679, 'train/grad_norm_pre_clip': 19.57518196105957, 'timing/data': 0.0010227300226688385, 'timing/model': 0.9326315233483911, 'learning_rate/qwen_vl_interface': 9.63552551807521e-06, 'learning_rate/action_model': 9.63552551807521e-05, 'epoch': 0.4}) +[2026-06-17 18:16:24,190][starVLA.training.train_starvla][INFO] - Step 967, Loss: {'train/loss': 0.8025758862495422, 'train/grad_norm_pre_clip': 44.7547607421875, 'timing/data': 0.0008163824677467346, 'timing/model': 1.2601540824398398, 'learning_rate/qwen_vl_interface': 9.634694254693691e-06, 'learning_rate/action_model': 9.634694254693692e-05, 'epoch': 0.4}) +[2026-06-17 18:16:27,848][starVLA.training.train_starvla][INFO] - Step 968, Loss: {'train/loss': 0.8366476893424988, 'train/grad_norm_pre_clip': 8.105472564697266, 'timing/data': 0.0011558495461940765, 'timing/model': 0.9615189563483, 'learning_rate/qwen_vl_interface': 9.633862082360336e-06, 'learning_rate/action_model': 9.633862082360335e-05, 'epoch': 0.4}) +[2026-06-17 18:16:31,647][starVLA.training.train_starvla][INFO] - Step 969, Loss: {'train/loss': 1.099237322807312, 'train/grad_norm_pre_clip': 13.670680046081543, 'timing/data': 0.0008433889597654343, 'timing/model': 0.938756987452507, 'learning_rate/qwen_vl_interface': 9.633029001247654e-06, 'learning_rate/action_model': 9.633029001247653e-05, 'epoch': 0.4}) +[2026-06-17 18:16:35,492][starVLA.training.train_starvla][INFO] - Step 970, Loss: {'train/loss': 0.7932227253913879, 'train/grad_norm_pre_clip': 8.404784202575684, 'timing/data': 0.0011880164965987206, 'timing/model': 0.9772402094677091, 'learning_rate/qwen_vl_interface': 9.63219501152834e-06, 'learning_rate/action_model': 9.632195011528341e-05, 'epoch': 0.4}) +[2026-06-17 18:16:39,569][starVLA.training.train_starvla][INFO] - Step 971, Loss: {'train/loss': 0.7140571475028992, 'train/grad_norm_pre_clip': 9.00961685180664, 'timing/data': 0.0007161879912018776, 'timing/model': 0.943323484621942, 'learning_rate/qwen_vl_interface': 9.63136011337529e-06, 'learning_rate/action_model': 9.63136011337529e-05, 'epoch': 0.4}) +[2026-06-17 18:16:43,870][starVLA.training.train_starvla][INFO] - Step 972, Loss: {'train/loss': 0.84794682264328, 'train/grad_norm_pre_clip': 11.034668922424316, 'timing/data': 0.0007355064153671265, 'timing/model': 1.13457319047302, 'learning_rate/qwen_vl_interface': 9.63052430696157e-06, 'learning_rate/action_model': 9.63052430696157e-05, 'epoch': 0.4}) +[2026-06-17 18:16:48,128][starVLA.training.train_starvla][INFO] - Step 973, Loss: {'train/loss': 0.9953624606132507, 'train/grad_norm_pre_clip': 10.189057350158691, 'timing/data': 0.0004894081503152847, 'timing/model': 1.513359130360186, 'learning_rate/qwen_vl_interface': 9.629687592460449e-06, 'learning_rate/action_model': 9.62968759246045e-05, 'epoch': 0.4}) +[2026-06-17 18:16:52,113][starVLA.training.train_starvla][INFO] - Step 974, Loss: {'train/loss': 0.7885243892669678, 'train/grad_norm_pre_clip': 13.577821731567383, 'timing/data': 0.00125920120626688, 'timing/model': 1.158796121366322, 'learning_rate/qwen_vl_interface': 9.628849970045378e-06, 'learning_rate/action_model': 9.628849970045378e-05, 'epoch': 0.4}) +[2026-06-17 18:16:55,782][starVLA.training.train_starvla][INFO] - Step 975, Loss: {'train/loss': 0.7219098806381226, 'train/grad_norm_pre_clip': 16.87641143798828, 'timing/data': 0.0006938846781849861, 'timing/model': 0.9466585125774145, 'learning_rate/qwen_vl_interface': 9.628011439889997e-06, 'learning_rate/action_model': 9.628011439889996e-05, 'epoch': 0.4}) +[2026-06-17 18:16:59,888][starVLA.training.train_starvla][INFO] - Step 976, Loss: {'train/loss': 0.914286732673645, 'train/grad_norm_pre_clip': 11.013343811035156, 'timing/data': 0.000667765736579895, 'timing/model': 1.0219586985185742, 'learning_rate/qwen_vl_interface': 9.62717200216813e-06, 'learning_rate/action_model': 9.62717200216813e-05, 'epoch': 0.4}) +[2026-06-17 18:17:03,765][starVLA.training.train_starvla][INFO] - Step 977, Loss: {'train/loss': 0.48043450713157654, 'train/grad_norm_pre_clip': 8.268754959106445, 'timing/data': 0.00046622566878795624, 'timing/model': 0.9378270618617535, 'learning_rate/qwen_vl_interface': 9.6263316570538e-06, 'learning_rate/action_model': 9.6263316570538e-05, 'epoch': 0.4}) +[2026-06-17 18:17:10,042][starVLA.training.train_starvla][INFO] - Step 978, Loss: {'train/loss': 0.7111073136329651, 'train/grad_norm_pre_clip': 9.622152328491211, 'timing/data': 0.0006446549668908119, 'timing/model': 1.1714435275644064, 'learning_rate/qwen_vl_interface': 9.625490404721208e-06, 'learning_rate/action_model': 9.625490404721208e-05, 'epoch': 0.4}) +[2026-06-17 18:17:15,236][starVLA.training.train_starvla][INFO] - Step 979, Loss: {'train/loss': 0.618389368057251, 'train/grad_norm_pre_clip': 10.11671257019043, 'timing/data': 0.0005105109885334969, 'timing/model': 1.888400623574853, 'learning_rate/qwen_vl_interface': 9.624648245344747e-06, 'learning_rate/action_model': 9.624648245344746e-05, 'epoch': 0.4}) +[2026-06-17 18:17:19,265][starVLA.training.train_starvla][INFO] - Step 980, Loss: {'train/loss': 0.43270808458328247, 'train/grad_norm_pre_clip': 9.558170318603516, 'timing/data': 0.0007503703236579895, 'timing/model': 1.2741970857605338, 'learning_rate/qwen_vl_interface': 9.623805179098996e-06, 'learning_rate/action_model': 9.623805179098996e-05, 'epoch': 0.4}) +[2026-06-17 18:17:23,164][starVLA.training.train_starvla][INFO] - Step 981, Loss: {'train/loss': 0.7568904161453247, 'train/grad_norm_pre_clip': 21.667531967163086, 'timing/data': 0.0005505681037902832, 'timing/model': 1.1775146052241325, 'learning_rate/qwen_vl_interface': 9.622961206158726e-06, 'learning_rate/action_model': 9.622961206158726e-05, 'epoch': 0.4}) +[2026-06-17 18:17:26,859][starVLA.training.train_starvla][INFO] - Step 982, Loss: {'train/loss': 1.043715000152588, 'train/grad_norm_pre_clip': 25.25872039794922, 'timing/data': 0.0005452875047922134, 'timing/model': 0.9373681144788861, 'learning_rate/qwen_vl_interface': 9.622116326698894e-06, 'learning_rate/action_model': 9.622116326698893e-05, 'epoch': 0.4}) +[2026-06-17 18:17:30,698][starVLA.training.train_starvla][INFO] - Step 983, Loss: {'train/loss': 0.9696041345596313, 'train/grad_norm_pre_clip': 10.284581184387207, 'timing/data': 0.000721207819879055, 'timing/model': 0.9330798648297787, 'learning_rate/qwen_vl_interface': 9.621270540894639e-06, 'learning_rate/action_model': 9.621270540894639e-05, 'epoch': 0.4}) +[2026-06-17 18:17:34,837][starVLA.training.train_starvla][INFO] - Step 984, Loss: {'train/loss': 0.9263932704925537, 'train/grad_norm_pre_clip': 9.947482109069824, 'timing/data': 0.0011096121743321419, 'timing/model': 0.943543205037713, 'learning_rate/qwen_vl_interface': 9.6204238489213e-06, 'learning_rate/action_model': 9.6204238489213e-05, 'epoch': 0.41}) +[2026-06-17 18:17:39,338][starVLA.training.train_starvla][INFO] - Step 985, Loss: {'train/loss': 0.7092572450637817, 'train/grad_norm_pre_clip': 7.331665515899658, 'timing/data': 0.0009385719895362854, 'timing/model': 0.946300215087831, 'learning_rate/qwen_vl_interface': 9.619576250954393e-06, 'learning_rate/action_model': 9.619576250954393e-05, 'epoch': 0.41}) +[2026-06-17 18:17:43,235][starVLA.training.train_starvla][INFO] - Step 986, Loss: {'train/loss': 0.6141731142997742, 'train/grad_norm_pre_clip': 15.219178199768066, 'timing/data': 0.0008176006376743317, 'timing/model': 0.9430578052997589, 'learning_rate/qwen_vl_interface': 9.618727747169628e-06, 'learning_rate/action_model': 9.618727747169627e-05, 'epoch': 0.41}) +[2026-06-17 18:17:47,133][starVLA.training.train_starvla][INFO] - Step 987, Loss: {'train/loss': 0.9940195083618164, 'train/grad_norm_pre_clip': 13.288531303405762, 'timing/data': 0.0005474509671330452, 'timing/model': 1.1749633429571986, 'learning_rate/qwen_vl_interface': 9.617878337742902e-06, 'learning_rate/action_model': 9.617878337742902e-05, 'epoch': 0.41}) +[2026-06-17 18:17:50,975][starVLA.training.train_starvla][INFO] - Step 988, Loss: {'train/loss': 0.4303423762321472, 'train/grad_norm_pre_clip': 9.134958267211914, 'timing/data': 0.0008123181760311127, 'timing/model': 1.0709288334473968, 'learning_rate/qwen_vl_interface': 9.617028022850294e-06, 'learning_rate/action_model': 9.617028022850293e-05, 'epoch': 0.41}) +[2026-06-17 18:17:54,641][starVLA.training.train_starvla][INFO] - Step 989, Loss: {'train/loss': 1.0208503007888794, 'train/grad_norm_pre_clip': 11.16800308227539, 'timing/data': 0.0007277196273207664, 'timing/model': 0.933822076767683, 'learning_rate/qwen_vl_interface': 9.616176802668078e-06, 'learning_rate/action_model': 9.616176802668078e-05, 'epoch': 0.41}) +[2026-06-17 18:17:59,341][starVLA.training.train_starvla][INFO] - Step 990, Loss: {'train/loss': 0.517763078212738, 'train/grad_norm_pre_clip': 6.504798889160156, 'timing/data': 0.0008162232115864754, 'timing/model': 1.022070842795074, 'learning_rate/qwen_vl_interface': 9.615324677372712e-06, 'learning_rate/action_model': 9.615324677372713e-05, 'epoch': 0.41}) +[2026-06-17 18:18:04,542][starVLA.training.train_starvla][INFO] - Step 991, Loss: {'train/loss': 0.46438437700271606, 'train/grad_norm_pre_clip': 8.03693675994873, 'timing/data': 0.0008221184834837914, 'timing/model': 0.9426663359627128, 'learning_rate/qwen_vl_interface': 9.614471647140844e-06, 'learning_rate/action_model': 9.614471647140844e-05, 'epoch': 0.41}) +[2026-06-17 18:18:08,616][starVLA.training.train_starvla][INFO] - Step 992, Loss: {'train/loss': 0.6788498163223267, 'train/grad_norm_pre_clip': 7.5380754470825195, 'timing/data': 0.000589245930314064, 'timing/model': 0.969701093621552, 'learning_rate/qwen_vl_interface': 9.613617712149309e-06, 'learning_rate/action_model': 9.613617712149308e-05, 'epoch': 0.41}) +[2026-06-17 18:18:13,968][starVLA.training.train_starvla][INFO] - Step 993, Loss: {'train/loss': 0.6186342835426331, 'train/grad_norm_pre_clip': 40.82078552246094, 'timing/data': 0.0009216796606779099, 'timing/model': 1.1693696090951562, 'learning_rate/qwen_vl_interface': 9.612762872575124e-06, 'learning_rate/action_model': 9.612762872575124e-05, 'epoch': 0.41}) +[2026-06-17 18:18:19,966][starVLA.training.train_starvla][INFO] - Step 994, Loss: {'train/loss': 0.7576866149902344, 'train/grad_norm_pre_clip': 7.769775390625, 'timing/data': 0.0006216904148459435, 'timing/model': 1.1231812834739685, 'learning_rate/qwen_vl_interface': 9.611907128595503e-06, 'learning_rate/action_model': 9.611907128595502e-05, 'epoch': 0.41}) +[2026-06-17 18:18:26,115][starVLA.training.train_starvla][INFO] - Step 995, Loss: {'train/loss': 0.5434502959251404, 'train/grad_norm_pre_clip': 6.647316932678223, 'timing/data': 0.0008473647758364677, 'timing/model': 3.2242026003077626, 'learning_rate/qwen_vl_interface': 9.61105048038784e-06, 'learning_rate/action_model': 9.611050480387839e-05, 'epoch': 0.41}) +[2026-06-17 18:18:30,150][starVLA.training.train_starvla][INFO] - Step 996, Loss: {'train/loss': 0.326421856880188, 'train/grad_norm_pre_clip': 8.645630836486816, 'timing/data': 0.000715441070497036, 'timing/model': 0.9524541990831494, 'learning_rate/qwen_vl_interface': 9.610192928129716e-06, 'learning_rate/action_model': 9.610192928129717e-05, 'epoch': 0.41}) +[2026-06-17 18:18:34,137][starVLA.training.train_starvla][INFO] - Step 997, Loss: {'train/loss': 1.194290041923523, 'train/grad_norm_pre_clip': 12.695147514343262, 'timing/data': 0.0005809823051095009, 'timing/model': 0.9484249670058489, 'learning_rate/qwen_vl_interface': 9.609334471998911e-06, 'learning_rate/action_model': 9.609334471998912e-05, 'epoch': 0.41}) +[2026-06-17 18:18:37,976][starVLA.training.train_starvla][INFO] - Step 998, Loss: {'train/loss': 0.7200819253921509, 'train/grad_norm_pre_clip': 10.893388748168945, 'timing/data': 0.001536605879664421, 'timing/model': 0.9405283723026514, 'learning_rate/qwen_vl_interface': 9.608475112173379e-06, 'learning_rate/action_model': 9.608475112173379e-05, 'epoch': 0.41}) +[2026-06-17 18:18:41,801][starVLA.training.train_starvla][INFO] - Step 999, Loss: {'train/loss': 0.5895612835884094, 'train/grad_norm_pre_clip': 6.062784194946289, 'timing/data': 0.00048775970935821533, 'timing/model': 0.9381203781813383, 'learning_rate/qwen_vl_interface': 9.607614848831264e-06, 'learning_rate/action_model': 9.607614848831264e-05, 'epoch': 0.41}) +[2026-06-17 18:20:41,478][starVLA.training.train_starvla][INFO] - Step 1000, Loss: {'train/loss': 0.7368380427360535, 'train/grad_norm_pre_clip': 73.22738647460938, 'eval/loss': 0.7780559659004211, 'eval/latency_6/loss': 0.7780559659004211, 'eval/demon_fire_precision': 0.8983847283406755, 'eval/demon_fire_recall': 0.9886877828054299, 'eval/demon_fire_f1': 0.9413755962455764, 'eval/demon_fire_rate_model': 0.9474123539232053, 'eval/demon_fire_rate_teacher': 0.860879243183083, 'eval/demon_left_precision': 0.8333333333333334, 'eval/demon_left_recall': 0.37719298245614036, 'eval/demon_left_f1': 0.5193236714975845, 'eval/demon_left_rate_model': 0.07178631051752922, 'eval/demon_left_rate_teacher': 0.15859766277128548, 'eval/demon_right_precision': 0.7429193899782135, 'eval/demon_right_recall': 0.5238095238095238, 'eval/demon_right_f1': 0.6144144144144145, 'eval/demon_right_rate_model': 0.12771285475792987, 'eval/demon_right_rate_teacher': 0.18113522537562604, 'eval/demon_move_f1': 0.5668690429559995, 'eval/demon_cc_f1': 0.7856598607519603, 'eval/latency_6/demon_fire_precision': 0.8983847283406755, 'eval/latency_6/demon_fire_recall': 0.9886877828054299, 'eval/latency_6/demon_fire_f1': 0.9413755962455764, 'eval/latency_6/demon_fire_rate_model': 0.9474123539232053, 'eval/latency_6/demon_fire_rate_teacher': 0.860879243183083, 'eval/latency_6/demon_left_precision': 0.8333333333333334, 'eval/latency_6/demon_left_recall': 0.37719298245614036, 'eval/latency_6/demon_left_f1': 0.5193236714975845, 'eval/latency_6/demon_left_rate_model': 0.07178631051752922, 'eval/latency_6/demon_left_rate_teacher': 0.15859766277128548, 'eval/latency_6/demon_right_precision': 0.7429193899782135, 'eval/latency_6/demon_right_recall': 0.5238095238095238, 'eval/latency_6/demon_right_f1': 0.6144144144144145, 'eval/latency_6/demon_right_rate_model': 0.12771285475792987, 'eval/latency_6/demon_right_rate_teacher': 0.18113522537562604, 'eval/latency_6/demon_move_f1': 0.5668690429559995, 'eval/latency_6/demon_cc_f1': 0.7856598607519603, 'timing/data': 0.0007984191179275513, 'timing/model': 1.1293379198759794, 'learning_rate/qwen_vl_interface': 9.606753682150904e-06, 'learning_rate/action_model': 9.606753682150903e-05, 'epoch': 0.41}) +[2026-06-17 18:20:41,480][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_1000_state +[2026-06-17 18:20:41,485][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 18:20:48,558][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_1000_state/model.safetensors +[2026-06-17 18:21:03,413][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_1000_state/optimizer.bin +[2026-06-17 18:21:03,414][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_1000_state/sampler.bin +[2026-06-17 18:21:03,414][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_1000_state/sampler_1.bin +[2026-06-17 18:21:03,417][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_1000_state/random_states_0.pkl +[2026-06-17 18:21:03,423][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1000 +[2026-06-17 18:21:03,424][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 18:21:03,434][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 18:21:07,303][starVLA.training.train_starvla][INFO] - Step 1001, Loss: {'train/loss': 0.8217639923095703, 'train/grad_norm_pre_clip': 11.855392456054688, 'timing/data': 0.0005217287689447403, 'timing/model': 0.9432396497577429, 'learning_rate/qwen_vl_interface': 9.605891612310815e-06, 'learning_rate/action_model': 9.605891612310814e-05, 'epoch': 0.41}) +[2026-06-17 18:21:11,153][starVLA.training.train_starvla][INFO] - Step 1002, Loss: {'train/loss': 1.1873427629470825, 'train/grad_norm_pre_clip': 11.868879318237305, 'timing/data': 0.0004960056394338608, 'timing/model': 1.173048290424049, 'learning_rate/qwen_vl_interface': 9.60502863948971e-06, 'learning_rate/action_model': 9.60502863948971e-05, 'epoch': 0.41}) +[2026-06-17 18:21:14,892][starVLA.training.train_starvla][INFO] - Step 1003, Loss: {'train/loss': 0.6568732261657715, 'train/grad_norm_pre_clip': 9.874500274658203, 'timing/data': 0.000544094480574131, 'timing/model': 1.0184114705771208, 'learning_rate/qwen_vl_interface': 9.604164763866479e-06, 'learning_rate/action_model': 9.60416476386648e-05, 'epoch': 0.41}) +[2026-06-17 18:21:18,724][starVLA.training.train_starvla][INFO] - Step 1004, Loss: {'train/loss': 0.9420323371887207, 'train/grad_norm_pre_clip': 12.640281677246094, 'timing/data': 0.0014405492693185806, 'timing/model': 0.945272658020258, 'learning_rate/qwen_vl_interface': 9.60329998562021e-06, 'learning_rate/action_model': 9.60329998562021e-05, 'epoch': 0.41}) +[2026-06-17 18:21:22,542][starVLA.training.train_starvla][INFO] - Step 1005, Loss: {'train/loss': 0.7950208187103271, 'train/grad_norm_pre_clip': 8.568693161010742, 'timing/data': 0.0009353961795568466, 'timing/model': 0.9469945570454001, 'learning_rate/qwen_vl_interface': 9.602434304930168e-06, 'learning_rate/action_model': 9.602434304930168e-05, 'epoch': 0.41}) +[2026-06-17 18:21:26,388][starVLA.training.train_starvla][INFO] - Step 1006, Loss: {'train/loss': 0.48268961906433105, 'train/grad_norm_pre_clip': 9.244319915771484, 'timing/data': 0.0005478858947753906, 'timing/model': 0.9539909204468131, 'learning_rate/qwen_vl_interface': 9.601567721975811e-06, 'learning_rate/action_model': 9.601567721975812e-05, 'epoch': 0.41}) +[2026-06-17 18:21:30,237][starVLA.training.train_starvla][INFO] - Step 1007, Loss: {'train/loss': 0.8447514176368713, 'train/grad_norm_pre_clip': 9.259821891784668, 'timing/data': 0.0007443791255354881, 'timing/model': 0.9391526123508811, 'learning_rate/qwen_vl_interface': 9.600700236936785e-06, 'learning_rate/action_model': 9.600700236936785e-05, 'epoch': 0.41}) +[2026-06-17 18:21:34,025][starVLA.training.train_starvla][INFO] - Step 1008, Loss: {'train/loss': 0.6331679224967957, 'train/grad_norm_pre_clip': 7.479476451873779, 'timing/data': 0.0005157822743058205, 'timing/model': 0.9476831899955869, 'learning_rate/qwen_vl_interface': 9.599831849992916e-06, 'learning_rate/action_model': 9.599831849992916e-05, 'epoch': 0.41}) +[2026-06-17 18:21:37,908][starVLA.training.train_starvla][INFO] - Step 1009, Loss: {'train/loss': 0.6473451256752014, 'train/grad_norm_pre_clip': 9.847854614257812, 'timing/data': 0.0004697078838944435, 'timing/model': 1.0980189992114902, 'learning_rate/qwen_vl_interface': 9.598962561324225e-06, 'learning_rate/action_model': 9.598962561324225e-05, 'epoch': 0.42}) +[2026-06-17 18:21:41,653][starVLA.training.train_starvla][INFO] - Step 1010, Loss: {'train/loss': 0.9787244200706482, 'train/grad_norm_pre_clip': 8.6519193649292, 'timing/data': 0.00046208593994379044, 'timing/model': 1.075387205928564, 'learning_rate/qwen_vl_interface': 9.598092371110916e-06, 'learning_rate/action_model': 9.598092371110916e-05, 'epoch': 0.42}) +[2026-06-17 18:21:45,493][starVLA.training.train_starvla][INFO] - Step 1011, Loss: {'train/loss': 0.8959676027297974, 'train/grad_norm_pre_clip': 10.682281494140625, 'timing/data': 0.0005830144509673119, 'timing/model': 0.9489478850737214, 'learning_rate/qwen_vl_interface': 9.59722127953338e-06, 'learning_rate/action_model': 9.597221279533379e-05, 'epoch': 0.42}) +[2026-06-17 18:21:49,347][starVLA.training.train_starvla][INFO] - Step 1012, Loss: {'train/loss': 0.650958776473999, 'train/grad_norm_pre_clip': 7.3440423011779785, 'timing/data': 0.0004943320527672768, 'timing/model': 0.9334707651287317, 'learning_rate/qwen_vl_interface': 9.596349286772194e-06, 'learning_rate/action_model': 9.596349286772193e-05, 'epoch': 0.42}) +[2026-06-17 18:21:53,270][starVLA.training.train_starvla][INFO] - Step 1013, Loss: {'train/loss': 0.7308644652366638, 'train/grad_norm_pre_clip': 9.175968170166016, 'timing/data': 0.000619005411863327, 'timing/model': 0.9372295653447509, 'learning_rate/qwen_vl_interface': 9.595476393008123e-06, 'learning_rate/action_model': 9.595476393008123e-05, 'epoch': 0.42}) +[2026-06-17 18:21:57,165][starVLA.training.train_starvla][INFO] - Step 1014, Loss: {'train/loss': 0.6806505918502808, 'train/grad_norm_pre_clip': 8.0632963180542, 'timing/data': 0.000835757702589035, 'timing/model': 0.9323755642399192, 'learning_rate/qwen_vl_interface': 9.59460259842212e-06, 'learning_rate/action_model': 9.59460259842212e-05, 'epoch': 0.42}) +[2026-06-17 18:22:01,039][starVLA.training.train_starvla][INFO] - Step 1015, Loss: {'train/loss': 0.783003568649292, 'train/grad_norm_pre_clip': 7.899710655212402, 'timing/data': 0.0007229549810290337, 'timing/model': 1.1146389674395323, 'learning_rate/qwen_vl_interface': 9.593727903195323e-06, 'learning_rate/action_model': 9.593727903195324e-05, 'epoch': 0.42}) +[2026-06-17 18:22:04,891][starVLA.training.train_starvla][INFO] - Step 1016, Loss: {'train/loss': 0.911876380443573, 'train/grad_norm_pre_clip': 8.013483047485352, 'timing/data': 0.0011529270559549332, 'timing/model': 1.1025226535275578, 'learning_rate/qwen_vl_interface': 9.59285230750906e-06, 'learning_rate/action_model': 9.592852307509059e-05, 'epoch': 0.42}) +[2026-06-17 18:22:08,603][starVLA.training.train_starvla][INFO] - Step 1017, Loss: {'train/loss': 0.6334162950515747, 'train/grad_norm_pre_clip': 18.216514587402344, 'timing/data': 0.0011021001264452934, 'timing/model': 0.9528972906991839, 'learning_rate/qwen_vl_interface': 9.591975811544839e-06, 'learning_rate/action_model': 9.591975811544839e-05, 'epoch': 0.42}) +[2026-06-17 18:22:13,898][starVLA.training.train_starvla][INFO] - Step 1018, Loss: {'train/loss': 0.9236034750938416, 'train/grad_norm_pre_clip': 11.06916332244873, 'timing/data': 0.000504310242831707, 'timing/model': 0.9591397559270263, 'learning_rate/qwen_vl_interface': 9.59109841548436e-06, 'learning_rate/action_model': 9.591098415484359e-05, 'epoch': 0.42}) +[2026-06-17 18:22:17,741][starVLA.training.train_starvla][INFO] - Step 1019, Loss: {'train/loss': 0.46471038460731506, 'train/grad_norm_pre_clip': 18.880653381347656, 'timing/data': 0.0010483535006642342, 'timing/model': 0.9363210704177618, 'learning_rate/qwen_vl_interface': 9.590220119509508e-06, 'learning_rate/action_model': 9.590220119509508e-05, 'epoch': 0.42}) +[2026-06-17 18:22:22,262][starVLA.training.train_starvla][INFO] - Step 1020, Loss: {'train/loss': 0.7048102617263794, 'train/grad_norm_pre_clip': 9.092120170593262, 'timing/data': 0.0013210270553827286, 'timing/model': 0.9357787873595953, 'learning_rate/qwen_vl_interface': 9.589340923802355e-06, 'learning_rate/action_model': 9.589340923802354e-05, 'epoch': 0.42}) +[2026-06-17 18:22:26,366][starVLA.training.train_starvla][INFO] - Step 1021, Loss: {'train/loss': 0.8287215232849121, 'train/grad_norm_pre_clip': 9.323131561279297, 'timing/data': 0.00140313059091568, 'timing/model': 0.9299149280413985, 'learning_rate/qwen_vl_interface': 9.58846082854516e-06, 'learning_rate/action_model': 9.58846082854516e-05, 'epoch': 0.42}) +[2026-06-17 18:22:30,237][starVLA.training.train_starvla][INFO] - Step 1022, Loss: {'train/loss': 0.8177849054336548, 'train/grad_norm_pre_clip': 8.328696250915527, 'timing/data': 0.0011181402951478958, 'timing/model': 1.1496211998164654, 'learning_rate/qwen_vl_interface': 9.587579833920366e-06, 'learning_rate/action_model': 9.587579833920365e-05, 'epoch': 0.42}) +[2026-06-17 18:22:33,913][starVLA.training.train_starvla][INFO] - Step 1023, Loss: {'train/loss': 0.8302760124206543, 'train/grad_norm_pre_clip': 18.475086212158203, 'timing/data': 0.0005568200722336769, 'timing/model': 0.94107885658741, 'learning_rate/qwen_vl_interface': 9.586697940110605e-06, 'learning_rate/action_model': 9.586697940110605e-05, 'epoch': 0.42}) +[2026-06-17 18:22:37,760][starVLA.training.train_starvla][INFO] - Step 1024, Loss: {'train/loss': 0.8479679226875305, 'train/grad_norm_pre_clip': 11.167049407958984, 'timing/data': 0.00059525016695261, 'timing/model': 0.9427192956209183, 'learning_rate/qwen_vl_interface': 9.585815147298696e-06, 'learning_rate/action_model': 9.585815147298696e-05, 'epoch': 0.42}) +[2026-06-17 18:22:41,637][starVLA.training.train_starvla][INFO] - Step 1025, Loss: {'train/loss': 0.6595759391784668, 'train/grad_norm_pre_clip': 6.1770920753479, 'timing/data': 0.0005474463105201721, 'timing/model': 0.9459594255313277, 'learning_rate/qwen_vl_interface': 9.58493145566764e-06, 'learning_rate/action_model': 9.58493145566764e-05, 'epoch': 0.42}) +[2026-06-17 18:22:45,438][starVLA.training.train_starvla][INFO] - Step 1026, Loss: {'train/loss': 0.8192562460899353, 'train/grad_norm_pre_clip': 7.108494281768799, 'timing/data': 0.0010958854109048843, 'timing/model': 0.9350908594205976, 'learning_rate/qwen_vl_interface': 9.584046865400627e-06, 'learning_rate/action_model': 9.584046865400627e-05, 'epoch': 0.42}) +[2026-06-17 18:22:49,413][starVLA.training.train_starvla][INFO] - Step 1027, Loss: {'train/loss': 0.5665321946144104, 'train/grad_norm_pre_clip': 11.17258358001709, 'timing/data': 0.0005606589838862419, 'timing/model': 0.9266350427642465, 'learning_rate/qwen_vl_interface': 9.583161376681039e-06, 'learning_rate/action_model': 9.583161376681038e-05, 'epoch': 0.42}) +[2026-06-17 18:22:53,339][starVLA.training.train_starvla][INFO] - Step 1028, Loss: {'train/loss': 0.7321172952651978, 'train/grad_norm_pre_clip': 8.498114585876465, 'timing/data': 0.0005938978865742683, 'timing/model': 1.1580441258847713, 'learning_rate/qwen_vl_interface': 9.582274989692432e-06, 'learning_rate/action_model': 9.582274989692432e-05, 'epoch': 0.42}) +[2026-06-17 18:22:57,247][starVLA.training.train_starvla][INFO] - Step 1029, Loss: {'train/loss': 0.5183296799659729, 'train/grad_norm_pre_clip': 9.002959251403809, 'timing/data': 0.0005664899945259094, 'timing/model': 1.1324576465412974, 'learning_rate/qwen_vl_interface': 9.58138770461856e-06, 'learning_rate/action_model': 9.58138770461856e-05, 'epoch': 0.42}) +[2026-06-17 18:23:00,948][starVLA.training.train_starvla][INFO] - Step 1030, Loss: {'train/loss': 0.8093466758728027, 'train/grad_norm_pre_clip': 8.9468412399292, 'timing/data': 0.0014562439173460007, 'timing/model': 0.9562675952911377, 'learning_rate/qwen_vl_interface': 9.580499521643355e-06, 'learning_rate/action_model': 9.580499521643355e-05, 'epoch': 0.42}) +[2026-06-17 18:23:04,835][starVLA.training.train_starvla][INFO] - Step 1031, Loss: {'train/loss': 0.7379070520401001, 'train/grad_norm_pre_clip': 9.996756553649902, 'timing/data': 0.0005381377413868904, 'timing/model': 0.9534508138895035, 'learning_rate/qwen_vl_interface': 9.579610440950939e-06, 'learning_rate/action_model': 9.579610440950937e-05, 'epoch': 0.42}) +[2026-06-17 18:23:08,745][starVLA.training.train_starvla][INFO] - Step 1032, Loss: {'train/loss': 0.6688230037689209, 'train/grad_norm_pre_clip': 9.176279067993164, 'timing/data': 0.00046913884580135345, 'timing/model': 0.9631526153534651, 'learning_rate/qwen_vl_interface': 9.57872046272562e-06, 'learning_rate/action_model': 9.57872046272562e-05, 'epoch': 0.42}) +[2026-06-17 18:23:13,266][starVLA.training.train_starvla][INFO] - Step 1033, Loss: {'train/loss': 0.8048276305198669, 'train/grad_norm_pre_clip': 10.185540199279785, 'timing/data': 0.0005142334848642349, 'timing/model': 0.9473008438944817, 'learning_rate/qwen_vl_interface': 9.577829587151892e-06, 'learning_rate/action_model': 9.577829587151891e-05, 'epoch': 0.43}) +[2026-06-17 18:23:17,487][starVLA.training.train_starvla][INFO] - Step 1034, Loss: {'train/loss': 0.8821942806243896, 'train/grad_norm_pre_clip': 9.903322219848633, 'timing/data': 0.0005153501406311989, 'timing/model': 1.0803514504805207, 'learning_rate/qwen_vl_interface': 9.576937814414433e-06, 'learning_rate/action_model': 9.576937814414432e-05, 'epoch': 0.43}) +[2026-06-17 18:23:21,890][starVLA.training.train_starvla][INFO] - Step 1035, Loss: {'train/loss': 0.8125377297401428, 'train/grad_norm_pre_clip': 9.552062034606934, 'timing/data': 0.0005472693592309952, 'timing/model': 1.170704659074545, 'learning_rate/qwen_vl_interface': 9.576045144698109e-06, 'learning_rate/action_model': 9.576045144698108e-05, 'epoch': 0.43}) +[2026-06-17 18:23:25,530][starVLA.training.train_starvla][INFO] - Step 1036, Loss: {'train/loss': 0.5921106338500977, 'train/grad_norm_pre_clip': 8.038443565368652, 'timing/data': 0.000843157060444355, 'timing/model': 0.9366730870679021, 'learning_rate/qwen_vl_interface': 9.57515157818797e-06, 'learning_rate/action_model': 9.57515157818797e-05, 'epoch': 0.43}) +[2026-06-17 18:23:29,418][starVLA.training.train_starvla][INFO] - Step 1037, Loss: {'train/loss': 0.7011517882347107, 'train/grad_norm_pre_clip': 9.148394584655762, 'timing/data': 0.000485430471599102, 'timing/model': 1.0483935000374913, 'learning_rate/qwen_vl_interface': 9.574257115069256e-06, 'learning_rate/action_model': 9.574257115069256e-05, 'epoch': 0.43}) +[2026-06-17 18:23:33,232][starVLA.training.train_starvla][INFO] - Step 1038, Loss: {'train/loss': 0.5516272783279419, 'train/grad_norm_pre_clip': 9.04040813446045, 'timing/data': 0.00048635993152856827, 'timing/model': 0.9478751737624407, 'learning_rate/qwen_vl_interface': 9.573361755527389e-06, 'learning_rate/action_model': 9.573361755527389e-05, 'epoch': 0.43}) +[2026-06-17 18:23:37,182][starVLA.training.train_starvla][INFO] - Step 1039, Loss: {'train/loss': 0.7114361524581909, 'train/grad_norm_pre_clip': 9.107869148254395, 'timing/data': 0.0008587455376982689, 'timing/model': 0.9414289696142077, 'learning_rate/qwen_vl_interface': 9.572465499747978e-06, 'learning_rate/action_model': 9.572465499747978e-05, 'epoch': 0.43}) +[2026-06-17 18:23:40,984][starVLA.training.train_starvla][INFO] - Step 1040, Loss: {'train/loss': 0.9732491970062256, 'train/grad_norm_pre_clip': 8.672022819519043, 'timing/data': 0.0005964701995253563, 'timing/model': 0.941969039849937, 'learning_rate/qwen_vl_interface': 9.571568347916819e-06, 'learning_rate/action_model': 9.571568347916819e-05, 'epoch': 0.43}) +[2026-06-17 18:23:44,828][starVLA.training.train_starvla][INFO] - Step 1041, Loss: {'train/loss': 0.6982120275497437, 'train/grad_norm_pre_clip': 8.680441856384277, 'timing/data': 0.0004852982237935066, 'timing/model': 0.929353566840291, 'learning_rate/qwen_vl_interface': 9.570670300219889e-06, 'learning_rate/action_model': 9.570670300219889e-05, 'epoch': 0.43}) +[2026-06-17 18:23:48,685][starVLA.training.train_starvla][INFO] - Step 1042, Loss: {'train/loss': 0.6395763754844666, 'train/grad_norm_pre_clip': 9.519966125488281, 'timing/data': 0.0004891008138656616, 'timing/model': 1.1386548792943358, 'learning_rate/qwen_vl_interface': 9.56977135684336e-06, 'learning_rate/action_model': 9.569771356843359e-05, 'epoch': 0.43}) +[2026-06-17 18:23:52,399][starVLA.training.train_starvla][INFO] - Step 1043, Loss: {'train/loss': 0.45198652148246765, 'train/grad_norm_pre_clip': 7.759263515472412, 'timing/data': 0.0004909588024020195, 'timing/model': 1.015332748182118, 'learning_rate/qwen_vl_interface': 9.568871517973576e-06, 'learning_rate/action_model': 9.568871517973577e-05, 'epoch': 0.43}) +[2026-06-17 18:23:56,233][starVLA.training.train_starvla][INFO] - Step 1044, Loss: {'train/loss': 1.0880279541015625, 'train/grad_norm_pre_clip': 9.457906723022461, 'timing/data': 0.0008186288177967072, 'timing/model': 0.9538453137502074, 'learning_rate/qwen_vl_interface': 9.567970783797081e-06, 'learning_rate/action_model': 9.567970783797082e-05, 'epoch': 0.43}) +[2026-06-17 18:24:00,006][starVLA.training.train_starvla][INFO] - Step 1045, Loss: {'train/loss': 0.5379818677902222, 'train/grad_norm_pre_clip': 7.3065714836120605, 'timing/data': 0.000506378710269928, 'timing/model': 0.9314026916399598, 'learning_rate/qwen_vl_interface': 9.567069154500597e-06, 'learning_rate/action_model': 9.567069154500596e-05, 'epoch': 0.43}) +[2026-06-17 18:24:03,876][starVLA.training.train_starvla][INFO] - Step 1046, Loss: {'train/loss': 0.7910473346710205, 'train/grad_norm_pre_clip': 9.933435440063477, 'timing/data': 0.0004813270643353462, 'timing/model': 0.9367442531511188, 'learning_rate/qwen_vl_interface': 9.56616663027103e-06, 'learning_rate/action_model': 9.566166630271029e-05, 'epoch': 0.43}) +[2026-06-17 18:24:07,765][starVLA.training.train_starvla][INFO] - Step 1047, Loss: {'train/loss': 0.8850772380828857, 'train/grad_norm_pre_clip': 8.981366157531738, 'timing/data': 0.0006895242258906364, 'timing/model': 0.939681114628911, 'learning_rate/qwen_vl_interface': 9.56526321129548e-06, 'learning_rate/action_model': 9.565263211295479e-05, 'epoch': 0.43}) +[2026-06-17 18:24:11,609][starVLA.training.train_starvla][INFO] - Step 1048, Loss: {'train/loss': 0.9027581214904785, 'train/grad_norm_pre_clip': 7.447039604187012, 'timing/data': 0.0004708012565970421, 'timing/model': 0.9436597339808941, 'learning_rate/qwen_vl_interface': 9.56435889776122e-06, 'learning_rate/action_model': 9.564358897761218e-05, 'epoch': 0.43}) +[2026-06-17 18:24:15,490][starVLA.training.train_starvla][INFO] - Step 1049, Loss: {'train/loss': 0.8971244692802429, 'train/grad_norm_pre_clip': 8.60027027130127, 'timing/data': 0.0004628533497452736, 'timing/model': 1.138620807789266, 'learning_rate/qwen_vl_interface': 9.563453689855717e-06, 'learning_rate/action_model': 9.563453689855716e-05, 'epoch': 0.43}) +[2026-06-17 18:24:19,109][starVLA.training.train_starvla][INFO] - Step 1050, Loss: {'train/loss': 0.4614916443824768, 'train/grad_norm_pre_clip': 7.012834548950195, 'timing/data': 0.00046203378587961197, 'timing/model': 0.9505809312686324, 'learning_rate/qwen_vl_interface': 9.562547587766624e-06, 'learning_rate/action_model': 9.562547587766623e-05, 'epoch': 0.43}) +[2026-06-17 18:24:22,904][starVLA.training.train_starvla][INFO] - Step 1051, Loss: {'train/loss': 0.7496984004974365, 'train/grad_norm_pre_clip': 10.83260726928711, 'timing/data': 0.0004988797008991241, 'timing/model': 0.9483466325327754, 'learning_rate/qwen_vl_interface': 9.561640591681775e-06, 'learning_rate/action_model': 9.561640591681774e-05, 'epoch': 0.43}) +[2026-06-17 18:24:26,729][starVLA.training.train_starvla][INFO] - Step 1052, Loss: {'train/loss': 0.9645218253135681, 'train/grad_norm_pre_clip': 8.806595802307129, 'timing/data': 0.0004896577447652817, 'timing/model': 0.9503631442785263, 'learning_rate/qwen_vl_interface': 9.560732701789193e-06, 'learning_rate/action_model': 9.560732701789193e-05, 'epoch': 0.43}) +[2026-06-17 18:24:30,715][starVLA.training.train_starvla][INFO] - Step 1053, Loss: {'train/loss': 0.7925190925598145, 'train/grad_norm_pre_clip': 10.764525413513184, 'timing/data': 0.0005442090332508087, 'timing/model': 0.9566460903733969, 'learning_rate/qwen_vl_interface': 9.55982391827708e-06, 'learning_rate/action_model': 9.559823918277081e-05, 'epoch': 0.43}) +[2026-06-17 18:24:34,564][starVLA.training.train_starvla][INFO] - Step 1054, Loss: {'train/loss': 0.5658198595046997, 'train/grad_norm_pre_clip': 10.107617378234863, 'timing/data': 0.0005496107041835785, 'timing/model': 0.9433479653671384, 'learning_rate/qwen_vl_interface': 9.558914241333835e-06, 'learning_rate/action_model': 9.558914241333835e-05, 'epoch': 0.43}) +[2026-06-17 18:24:38,465][starVLA.training.train_starvla][INFO] - Step 1055, Loss: {'train/loss': 0.38881915807724, 'train/grad_norm_pre_clip': 7.183718681335449, 'timing/data': 0.0008165361359715462, 'timing/model': 1.1683743773028255, 'learning_rate/qwen_vl_interface': 9.55800367114803e-06, 'learning_rate/action_model': 9.55800367114803e-05, 'epoch': 0.43}) +[2026-06-17 18:24:42,562][starVLA.training.train_starvla][INFO] - Step 1056, Loss: {'train/loss': 0.5689182877540588, 'train/grad_norm_pre_clip': 6.675412654876709, 'timing/data': 0.0009658476337790489, 'timing/model': 1.3759305123239756, 'learning_rate/qwen_vl_interface': 9.557092207908428e-06, 'learning_rate/action_model': 9.557092207908428e-05, 'epoch': 0.43}) +[2026-06-17 18:24:46,329][starVLA.training.train_starvla][INFO] - Step 1057, Loss: {'train/loss': 0.5310717821121216, 'train/grad_norm_pre_clip': 11.988436698913574, 'timing/data': 0.0004864893853664398, 'timing/model': 0.9497181018814445, 'learning_rate/qwen_vl_interface': 9.556179851803978e-06, 'learning_rate/action_model': 9.556179851803978e-05, 'epoch': 0.44}) +[2026-06-17 18:24:50,129][starVLA.training.train_starvla][INFO] - Step 1058, Loss: {'train/loss': 0.8017977476119995, 'train/grad_norm_pre_clip': 9.87886905670166, 'timing/data': 0.0005007116124033928, 'timing/model': 0.947921484708786, 'learning_rate/qwen_vl_interface': 9.555266603023809e-06, 'learning_rate/action_model': 9.555266603023808e-05, 'epoch': 0.44}) +[2026-06-17 18:24:54,052][starVLA.training.train_starvla][INFO] - Step 1059, Loss: {'train/loss': 0.8006748557090759, 'train/grad_norm_pre_clip': 11.658726692199707, 'timing/data': 0.0007626973092556, 'timing/model': 0.9389634858816862, 'learning_rate/qwen_vl_interface': 9.554352461757243e-06, 'learning_rate/action_model': 9.554352461757242e-05, 'epoch': 0.44}) +[2026-06-17 18:24:57,940][starVLA.training.train_starvla][INFO] - Step 1060, Loss: {'train/loss': 0.6969214081764221, 'train/grad_norm_pre_clip': 12.753541946411133, 'timing/data': 0.0008167680352926254, 'timing/model': 0.9520386150106788, 'learning_rate/qwen_vl_interface': 9.553437428193778e-06, 'learning_rate/action_model': 9.553437428193778e-05, 'epoch': 0.44}) +[2026-06-17 18:25:01,954][starVLA.training.train_starvla][INFO] - Step 1061, Loss: {'train/loss': 0.8923922777175903, 'train/grad_norm_pre_clip': 9.432453155517578, 'timing/data': 0.0006502056494355202, 'timing/model': 0.9411758035421371, 'learning_rate/qwen_vl_interface': 9.552521502523104e-06, 'learning_rate/action_model': 9.552521502523102e-05, 'epoch': 0.44}) +[2026-06-17 18:25:07,496][starVLA.training.train_starvla][INFO] - Step 1062, Loss: {'train/loss': 0.7485520839691162, 'train/grad_norm_pre_clip': 8.549723625183105, 'timing/data': 0.0009481925517320633, 'timing/model': 1.2028141412883997, 'learning_rate/qwen_vl_interface': 9.551604684935093e-06, 'learning_rate/action_model': 9.551604684935092e-05, 'epoch': 0.44}) +[2026-06-17 18:25:11,329][starVLA.training.train_starvla][INFO] - Step 1063, Loss: {'train/loss': 0.8298632502555847, 'train/grad_norm_pre_clip': 21.04558563232422, 'timing/data': 0.0005751354619860649, 'timing/model': 1.1176919667050242, 'learning_rate/qwen_vl_interface': 9.550686975619803e-06, 'learning_rate/action_model': 9.550686975619802e-05, 'epoch': 0.44}) +[2026-06-17 18:25:15,001][starVLA.training.train_starvla][INFO] - Step 1064, Loss: {'train/loss': 0.6876731514930725, 'train/grad_norm_pre_clip': 6.871984004974365, 'timing/data': 0.00050331000238657, 'timing/model': 0.9429876375943422, 'learning_rate/qwen_vl_interface': 9.549768374767473e-06, 'learning_rate/action_model': 9.549768374767473e-05, 'epoch': 0.44}) +[2026-06-17 18:25:18,884][starVLA.training.train_starvla][INFO] - Step 1065, Loss: {'train/loss': 0.6649245023727417, 'train/grad_norm_pre_clip': 8.005505561828613, 'timing/data': 0.0004906849935650826, 'timing/model': 0.942063732072711, 'learning_rate/qwen_vl_interface': 9.548848882568533e-06, 'learning_rate/action_model': 9.548848882568532e-05, 'epoch': 0.44}) +[2026-06-17 18:25:22,928][starVLA.training.train_starvla][INFO] - Step 1066, Loss: {'train/loss': 0.7377211451530457, 'train/grad_norm_pre_clip': 8.82801628112793, 'timing/data': 0.0008630175143480301, 'timing/model': 1.0535560892894864, 'learning_rate/qwen_vl_interface': 9.547928499213593e-06, 'learning_rate/action_model': 9.547928499213593e-05, 'epoch': 0.44}) +[2026-06-17 18:25:26,783][starVLA.training.train_starvla][INFO] - Step 1067, Loss: {'train/loss': 0.9220854043960571, 'train/grad_norm_pre_clip': 11.413344383239746, 'timing/data': 0.0011217733845114708, 'timing/model': 0.9457565555348992, 'learning_rate/qwen_vl_interface': 9.547007224893452e-06, 'learning_rate/action_model': 9.547007224893452e-05, 'epoch': 0.44}) +[2026-06-17 18:25:30,627][starVLA.training.train_starvla][INFO] - Step 1068, Loss: {'train/loss': 0.702394962310791, 'train/grad_norm_pre_clip': 8.740939140319824, 'timing/data': 0.0007196897640824318, 'timing/model': 0.9512748951092362, 'learning_rate/qwen_vl_interface': 9.546085059799088e-06, 'learning_rate/action_model': 9.546085059799088e-05, 'epoch': 0.44}) +[2026-06-17 18:25:34,518][starVLA.training.train_starvla][INFO] - Step 1069, Loss: {'train/loss': 0.8997341394424438, 'train/grad_norm_pre_clip': 11.164026260375977, 'timing/data': 0.0005064606666564941, 'timing/model': 1.169866118580103, 'learning_rate/qwen_vl_interface': 9.545162004121672e-06, 'learning_rate/action_model': 9.54516200412167e-05, 'epoch': 0.44}) +[2026-06-17 18:25:38,239][starVLA.training.train_starvla][INFO] - Step 1070, Loss: {'train/loss': 0.5506115555763245, 'train/grad_norm_pre_clip': 6.8523335456848145, 'timing/data': 0.0005842652171850204, 'timing/model': 0.9375894162803888, 'learning_rate/qwen_vl_interface': 9.544238058052546e-06, 'learning_rate/action_model': 9.544238058052545e-05, 'epoch': 0.44}) +[2026-06-17 18:25:42,180][starVLA.training.train_starvla][INFO] - Step 1071, Loss: {'train/loss': 0.656434953212738, 'train/grad_norm_pre_clip': 7.282461166381836, 'timing/data': 0.0005325805395841599, 'timing/model': 0.9432590864598751, 'learning_rate/qwen_vl_interface': 9.54331322178325e-06, 'learning_rate/action_model': 9.543313221783251e-05, 'epoch': 0.44}) +[2026-06-17 18:25:47,119][starVLA.training.train_starvla][INFO] - Step 1072, Loss: {'train/loss': 0.8568690419197083, 'train/grad_norm_pre_clip': 9.432084083557129, 'timing/data': 0.0005255080759525299, 'timing/model': 0.9879547692835331, 'learning_rate/qwen_vl_interface': 9.542387495505505e-06, 'learning_rate/action_model': 9.542387495505506e-05, 'epoch': 0.44}) +[2026-06-17 18:25:52,119][starVLA.training.train_starvla][INFO] - Step 1073, Loss: {'train/loss': 0.8310189247131348, 'train/grad_norm_pre_clip': 8.848329544067383, 'timing/data': 0.0005318485200405121, 'timing/model': 0.957146193832159, 'learning_rate/qwen_vl_interface': 9.541460879411214e-06, 'learning_rate/action_model': 9.541460879411215e-05, 'epoch': 0.44}) +[2026-06-17 18:25:56,011][starVLA.training.train_starvla][INFO] - Step 1074, Loss: {'train/loss': 0.4895744323730469, 'train/grad_norm_pre_clip': 17.517873764038086, 'timing/data': 0.0006205365061759949, 'timing/model': 0.9584354944527149, 'learning_rate/qwen_vl_interface': 9.540533373692466e-06, 'learning_rate/action_model': 9.540533373692466e-05, 'epoch': 0.44}) +[2026-06-17 18:25:59,844][starVLA.training.train_starvla][INFO] - Step 1075, Loss: {'train/loss': 0.45937225222587585, 'train/grad_norm_pre_clip': 33.42354965209961, 'timing/data': 0.0007623499259352684, 'timing/model': 1.1230624904856086, 'learning_rate/qwen_vl_interface': 9.539604978541531e-06, 'learning_rate/action_model': 9.539604978541531e-05, 'epoch': 0.44}) +[2026-06-17 18:26:03,702][starVLA.training.train_starvla][INFO] - Step 1076, Loss: {'train/loss': 0.7929589152336121, 'train/grad_norm_pre_clip': 9.660781860351562, 'timing/data': 0.000712166540324688, 'timing/model': 1.1211283905431628, 'learning_rate/qwen_vl_interface': 9.538675694150868e-06, 'learning_rate/action_model': 9.538675694150868e-05, 'epoch': 0.44}) +[2026-06-17 18:26:07,370][starVLA.training.train_starvla][INFO] - Step 1077, Loss: {'train/loss': 0.5821544528007507, 'train/grad_norm_pre_clip': 10.642850875854492, 'timing/data': 0.0006309589371085167, 'timing/model': 0.9803817011415958, 'learning_rate/qwen_vl_interface': 9.53774552071312e-06, 'learning_rate/action_model': 9.537745520713119e-05, 'epoch': 0.44}) +[2026-06-17 18:26:11,388][starVLA.training.train_starvla][INFO] - Step 1078, Loss: {'train/loss': 0.5470746159553528, 'train/grad_norm_pre_clip': 11.360564231872559, 'timing/data': 0.0005229869857430458, 'timing/model': 1.0200830679386854, 'learning_rate/qwen_vl_interface': 9.53681445842111e-06, 'learning_rate/action_model': 9.53681445842111e-05, 'epoch': 0.44}) +[2026-06-17 18:26:15,286][starVLA.training.train_starvla][INFO] - Step 1079, Loss: {'train/loss': 0.6509811282157898, 'train/grad_norm_pre_clip': 9.622578620910645, 'timing/data': 0.0006180787459015846, 'timing/model': 0.9425747180357575, 'learning_rate/qwen_vl_interface': 9.53588250746785e-06, 'learning_rate/action_model': 9.53588250746785e-05, 'epoch': 0.44}) +[2026-06-17 18:26:19,166][starVLA.training.train_starvla][INFO] - Step 1080, Loss: {'train/loss': 0.693759024143219, 'train/grad_norm_pre_clip': 9.226065635681152, 'timing/data': 0.0005664601922035217, 'timing/model': 0.9359947079792619, 'learning_rate/qwen_vl_interface': 9.534949668046537e-06, 'learning_rate/action_model': 9.534949668046536e-05, 'epoch': 0.44}) +[2026-06-17 18:26:23,053][starVLA.training.train_starvla][INFO] - Step 1081, Loss: {'train/loss': 0.6605113744735718, 'train/grad_norm_pre_clip': 10.441561698913574, 'timing/data': 0.0009064944460988045, 'timing/model': 0.9394631516188383, 'learning_rate/qwen_vl_interface': 9.534015940350545e-06, 'learning_rate/action_model': 9.534015940350544e-05, 'epoch': 0.44}) +[2026-06-17 18:26:26,944][starVLA.training.train_starvla][INFO] - Step 1082, Loss: {'train/loss': 0.7333496809005737, 'train/grad_norm_pre_clip': 7.823351860046387, 'timing/data': 0.0005601244047284126, 'timing/model': 1.114739197306335, 'learning_rate/qwen_vl_interface': 9.533081324573437e-06, 'learning_rate/action_model': 9.533081324573437e-05, 'epoch': 0.45}) +[2026-06-17 18:26:30,605][starVLA.training.train_starvla][INFO] - Step 1083, Loss: {'train/loss': 0.6875132918357849, 'train/grad_norm_pre_clip': 8.64002513885498, 'timing/data': 0.0005285833030939102, 'timing/model': 0.9468145482242107, 'learning_rate/qwen_vl_interface': 9.532145820908965e-06, 'learning_rate/action_model': 9.532145820908964e-05, 'epoch': 0.45}) +[2026-06-17 18:26:34,462][starVLA.training.train_starvla][INFO] - Step 1084, Loss: {'train/loss': 0.6528534889221191, 'train/grad_norm_pre_clip': 7.432962894439697, 'timing/data': 0.0005387384444475174, 'timing/model': 0.9330552015453577, 'learning_rate/qwen_vl_interface': 9.531209429551056e-06, 'learning_rate/action_model': 9.531209429551054e-05, 'epoch': 0.45}) +[2026-06-17 18:26:38,489][starVLA.training.train_starvla][INFO] - Step 1085, Loss: {'train/loss': 0.8083947896957397, 'train/grad_norm_pre_clip': 10.093430519104004, 'timing/data': 0.00048680417239665985, 'timing/model': 0.9416621709242463, 'learning_rate/qwen_vl_interface': 9.530272150693823e-06, 'learning_rate/action_model': 9.530272150693824e-05, 'epoch': 0.45}) +[2026-06-17 18:26:42,558][starVLA.training.train_starvla][INFO] - Step 1086, Loss: {'train/loss': 0.5047013163566589, 'train/grad_norm_pre_clip': 9.117761611938477, 'timing/data': 0.0005304105579853058, 'timing/model': 0.9310224875807762, 'learning_rate/qwen_vl_interface': 9.52933398453157e-06, 'learning_rate/action_model': 9.52933398453157e-05, 'epoch': 0.45}) +[2026-06-17 18:26:46,355][starVLA.training.train_starvla][INFO] - Step 1087, Loss: {'train/loss': 0.5954076051712036, 'train/grad_norm_pre_clip': 10.774371147155762, 'timing/data': 0.0008816970512270927, 'timing/model': 0.9386192662641406, 'learning_rate/qwen_vl_interface': 9.528394931258777e-06, 'learning_rate/action_model': 9.528394931258776e-05, 'epoch': 0.45}) +[2026-06-17 18:26:50,270][starVLA.training.train_starvla][INFO] - Step 1088, Loss: {'train/loss': 0.5061103701591492, 'train/grad_norm_pre_clip': 9.892725944519043, 'timing/data': 0.0007062796503305435, 'timing/model': 0.9408074095845222, 'learning_rate/qwen_vl_interface': 9.527454991070108e-06, 'learning_rate/action_model': 9.527454991070108e-05, 'epoch': 0.45}) +[2026-06-17 18:26:54,159][starVLA.training.train_starvla][INFO] - Step 1089, Loss: {'train/loss': 0.5719276666641235, 'train/grad_norm_pre_clip': 7.388179779052734, 'timing/data': 0.0010062549263238907, 'timing/model': 1.1671361979097128, 'learning_rate/qwen_vl_interface': 9.52651416416042e-06, 'learning_rate/action_model': 9.526514164160421e-05, 'epoch': 0.45}) +[2026-06-17 18:26:57,889][starVLA.training.train_starvla][INFO] - Step 1090, Loss: {'train/loss': 1.1181946992874146, 'train/grad_norm_pre_clip': 14.64579963684082, 'timing/data': 0.0005345232784748077, 'timing/model': 0.955431648530066, 'learning_rate/qwen_vl_interface': 9.525572450724743e-06, 'learning_rate/action_model': 9.525572450724743e-05, 'epoch': 0.45}) +[2026-06-17 18:27:01,934][starVLA.training.train_starvla][INFO] - Step 1091, Loss: {'train/loss': 0.6212105751037598, 'train/grad_norm_pre_clip': 9.932028770446777, 'timing/data': 0.0008917981758713722, 'timing/model': 1.0083781676366925, 'learning_rate/qwen_vl_interface': 9.524629850958297e-06, 'learning_rate/action_model': 9.524629850958297e-05, 'epoch': 0.45}) +[2026-06-17 18:27:06,146][starVLA.training.train_starvla][INFO] - Step 1092, Loss: {'train/loss': 0.6909304261207581, 'train/grad_norm_pre_clip': 11.35546875, 'timing/data': 0.0018596230074763298, 'timing/model': 0.94763317797333, 'learning_rate/qwen_vl_interface': 9.523686365056482e-06, 'learning_rate/action_model': 9.523686365056481e-05, 'epoch': 0.45}) +[2026-06-17 18:27:12,533][starVLA.training.train_starvla][INFO] - Step 1093, Loss: {'train/loss': 0.8309475183486938, 'train/grad_norm_pre_clip': 11.890015602111816, 'timing/data': 0.0017268871888518333, 'timing/model': 3.2719125719740987, 'learning_rate/qwen_vl_interface': 9.522741993214886e-06, 'learning_rate/action_model': 9.522741993214886e-05, 'epoch': 0.45}) +[2026-06-17 18:27:16,705][starVLA.training.train_starvla][INFO] - Step 1094, Loss: {'train/loss': 0.8035938739776611, 'train/grad_norm_pre_clip': 10.61373233795166, 'timing/data': 0.0006228890269994736, 'timing/model': 1.0751324892044067, 'learning_rate/qwen_vl_interface': 9.521796735629275e-06, 'learning_rate/action_model': 9.521796735629275e-05, 'epoch': 0.45}) +[2026-06-17 18:27:23,583][starVLA.training.train_starvla][INFO] - Step 1095, Loss: {'train/loss': 0.9242160320281982, 'train/grad_norm_pre_clip': 11.601090431213379, 'timing/data': 0.0005047740414738655, 'timing/model': 2.442591682076454, 'learning_rate/qwen_vl_interface': 9.520850592495603e-06, 'learning_rate/action_model': 9.520850592495604e-05, 'epoch': 0.45}) +[2026-06-17 18:27:27,923][starVLA.training.train_starvla][INFO] - Step 1096, Loss: {'train/loss': 0.6625896096229553, 'train/grad_norm_pre_clip': 9.380048751831055, 'timing/data': 0.0008408715948462486, 'timing/model': 1.4361018585041165, 'learning_rate/qwen_vl_interface': 9.519903564010012e-06, 'learning_rate/action_model': 9.51990356401001e-05, 'epoch': 0.45}) +[2026-06-17 18:27:33,588][starVLA.training.train_starvla][INFO] - Step 1097, Loss: {'train/loss': 0.7601760029792786, 'train/grad_norm_pre_clip': 11.701776504516602, 'timing/data': 0.0005181543529033661, 'timing/model': 1.671380652114749, 'learning_rate/qwen_vl_interface': 9.518955650368816e-06, 'learning_rate/action_model': 9.518955650368814e-05, 'epoch': 0.45}) +[2026-06-17 18:27:37,586][starVLA.training.train_starvla][INFO] - Step 1098, Loss: {'train/loss': 0.6720028519630432, 'train/grad_norm_pre_clip': 10.176985740661621, 'timing/data': 0.0005422439426183701, 'timing/model': 0.9391608322039247, 'learning_rate/qwen_vl_interface': 9.51800685176852e-06, 'learning_rate/action_model': 9.518006851768518e-05, 'epoch': 0.45}) +[2026-06-17 18:27:41,666][starVLA.training.train_starvla][INFO] - Step 1099, Loss: {'train/loss': 0.42107850313186646, 'train/grad_norm_pre_clip': 10.668285369873047, 'timing/data': 0.0008241450414061546, 'timing/model': 0.9786597527563572, 'learning_rate/qwen_vl_interface': 9.51705716840581e-06, 'learning_rate/action_model': 9.517057168405808e-05, 'epoch': 0.45}) +[2026-06-17 18:29:43,597][starVLA.training.train_starvla][INFO] - Step 1100, Loss: {'train/loss': 1.007182002067566, 'train/grad_norm_pre_clip': 11.751142501831055, 'eval/loss': 0.7072373032569885, 'eval/latency_6/loss': 0.7072373032569885, 'eval/demon_fire_precision': 0.9274595267745953, 'eval/demon_fire_recall': 0.9628312863606981, 'eval/demon_fire_f1': 0.9448144624167459, 'eval/demon_fire_rate_model': 0.8937117417918754, 'eval/demon_fire_rate_teacher': 0.860879243183083, 'eval/demon_left_precision': 0.7219730941704036, 'eval/demon_left_recall': 0.5649122807017544, 'eval/demon_left_f1': 0.6338582677165354, 'eval/demon_left_rate_model': 0.12409571508069003, 'eval/demon_left_rate_teacher': 0.15859766277128548, 'eval/demon_right_precision': 0.7661290322580645, 'eval/demon_right_recall': 0.5837173579109063, 'eval/demon_right_f1': 0.6625980819529207, 'eval/demon_right_rate_model': 0.13800779076238176, 'eval/demon_right_rate_teacher': 0.18113522537562604, 'eval/demon_move_f1': 0.648228174834728, 'eval/demon_cc_f1': 0.830781308513478, 'eval/latency_6/demon_fire_precision': 0.9274595267745953, 'eval/latency_6/demon_fire_recall': 0.9628312863606981, 'eval/latency_6/demon_fire_f1': 0.9448144624167459, 'eval/latency_6/demon_fire_rate_model': 0.8937117417918754, 'eval/latency_6/demon_fire_rate_teacher': 0.860879243183083, 'eval/latency_6/demon_left_precision': 0.7219730941704036, 'eval/latency_6/demon_left_recall': 0.5649122807017544, 'eval/latency_6/demon_left_f1': 0.6338582677165354, 'eval/latency_6/demon_left_rate_model': 0.12409571508069003, 'eval/latency_6/demon_left_rate_teacher': 0.15859766277128548, 'eval/latency_6/demon_right_precision': 0.7661290322580645, 'eval/latency_6/demon_right_recall': 0.5837173579109063, 'eval/latency_6/demon_right_f1': 0.6625980819529207, 'eval/latency_6/demon_right_rate_model': 0.13800779076238176, 'eval/latency_6/demon_right_rate_teacher': 0.18113522537562604, 'eval/latency_6/demon_move_f1': 0.648228174834728, 'eval/latency_6/demon_cc_f1': 0.830781308513478, 'timing/data': 0.0006121834740042686, 'timing/model': 0.9378590453416109, 'learning_rate/qwen_vl_interface': 9.516106600477557e-06, 'learning_rate/action_model': 9.516106600477557e-05, 'epoch': 0.45}) +[2026-06-17 18:29:43,619][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_1100_state +[2026-06-17 18:29:43,624][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 18:29:50,472][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_1100_state/model.safetensors +[2026-06-17 18:30:06,456][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_1100_state/optimizer.bin +[2026-06-17 18:30:06,457][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_1100_state/sampler.bin +[2026-06-17 18:30:06,457][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_1100_state/sampler_1.bin +[2026-06-17 18:30:06,461][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_1100_state/random_states_0.pkl +[2026-06-17 18:30:06,465][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1100 +[2026-06-17 18:30:06,465][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 18:30:06,476][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 18:30:10,443][starVLA.training.train_starvla][INFO] - Step 1101, Loss: {'train/loss': 0.6859258413314819, 'train/grad_norm_pre_clip': 8.683375358581543, 'timing/data': 0.0005720704793930054, 'timing/model': 0.9363430719822645, 'learning_rate/qwen_vl_interface': 9.515155148180816e-06, 'learning_rate/action_model': 9.515155148180816e-05, 'epoch': 0.45}) +[2026-06-17 18:30:14,371][starVLA.training.train_starvla][INFO] - Step 1102, Loss: {'train/loss': 0.6117326021194458, 'train/grad_norm_pre_clip': 9.497825622558594, 'timing/data': 0.0006156833842396736, 'timing/model': 0.9475260470062494, 'learning_rate/qwen_vl_interface': 9.514202811712825e-06, 'learning_rate/action_model': 9.514202811712825e-05, 'epoch': 0.45}) +[2026-06-17 18:30:18,685][starVLA.training.train_starvla][INFO] - Step 1103, Loss: {'train/loss': 0.42879197001457214, 'train/grad_norm_pre_clip': 5.400619983673096, 'timing/data': 0.000539545901119709, 'timing/model': 1.1209908751770854, 'learning_rate/qwen_vl_interface': 9.513249591271e-06, 'learning_rate/action_model': 9.513249591271e-05, 'epoch': 0.45}) +[2026-06-17 18:30:22,814][starVLA.training.train_starvla][INFO] - Step 1104, Loss: {'train/loss': 1.0304445028305054, 'train/grad_norm_pre_clip': 8.959818840026855, 'timing/data': 0.001084793359041214, 'timing/model': 1.4343255376443267, 'learning_rate/qwen_vl_interface': 9.512295487052949e-06, 'learning_rate/action_model': 9.512295487052949e-05, 'epoch': 0.45}) +[2026-06-17 18:30:26,500][starVLA.training.train_starvla][INFO] - Step 1105, Loss: {'train/loss': 0.995478630065918, 'train/grad_norm_pre_clip': 12.534586906433105, 'timing/data': 0.0005326149985194206, 'timing/model': 0.9753281576558948, 'learning_rate/qwen_vl_interface': 9.511340499256455e-06, 'learning_rate/action_model': 9.511340499256455e-05, 'epoch': 0.45}) +[2026-06-17 18:30:30,357][starVLA.training.train_starvla][INFO] - Step 1106, Loss: {'train/loss': 0.9894108176231384, 'train/grad_norm_pre_clip': 12.546265602111816, 'timing/data': 0.0008661011233925819, 'timing/model': 0.9374025603756309, 'learning_rate/qwen_vl_interface': 9.510384628079493e-06, 'learning_rate/action_model': 9.510384628079493e-05, 'epoch': 0.46}) +[2026-06-17 18:30:34,249][starVLA.training.train_starvla][INFO] - Step 1107, Loss: {'train/loss': 0.6892677545547485, 'train/grad_norm_pre_clip': 10.83649730682373, 'timing/data': 0.0010101543739438057, 'timing/model': 0.947993665933609, 'learning_rate/qwen_vl_interface': 9.50942787372021e-06, 'learning_rate/action_model': 9.50942787372021e-05, 'epoch': 0.46}) +[2026-06-17 18:30:38,842][starVLA.training.train_starvla][INFO] - Step 1108, Loss: {'train/loss': 0.622089147567749, 'train/grad_norm_pre_clip': 7.889560222625732, 'timing/data': 0.0006869295611977577, 'timing/model': 0.9881585650146008, 'learning_rate/qwen_vl_interface': 9.508470236376948e-06, 'learning_rate/action_model': 9.508470236376948e-05, 'epoch': 0.46}) +[2026-06-17 18:30:42,794][starVLA.training.train_starvla][INFO] - Step 1109, Loss: {'train/loss': 0.8172174692153931, 'train/grad_norm_pre_clip': 11.89318561553955, 'timing/data': 0.0011681821197271347, 'timing/model': 0.9540646178647876, 'learning_rate/qwen_vl_interface': 9.507511716248224e-06, 'learning_rate/action_model': 9.507511716248224e-05, 'epoch': 0.46}) +[2026-06-17 18:30:46,850][starVLA.training.train_starvla][INFO] - Step 1110, Loss: {'train/loss': 1.040497064590454, 'train/grad_norm_pre_clip': 16.9237117767334, 'timing/data': 0.0005284780636429787, 'timing/model': 1.2015894297510386, 'learning_rate/qwen_vl_interface': 9.506552313532739e-06, 'learning_rate/action_model': 9.506552313532739e-05, 'epoch': 0.46}) +[2026-06-17 18:30:50,916][starVLA.training.train_starvla][INFO] - Step 1111, Loss: {'train/loss': 0.8247024416923523, 'train/grad_norm_pre_clip': 10.398611068725586, 'timing/data': 0.0005554733797907829, 'timing/model': 1.1660657692700624, 'learning_rate/qwen_vl_interface': 9.505592028429377e-06, 'learning_rate/action_model': 9.505592028429377e-05, 'epoch': 0.46}) +[2026-06-17 18:30:54,557][starVLA.training.train_starvla][INFO] - Step 1112, Loss: {'train/loss': 0.5666373372077942, 'train/grad_norm_pre_clip': 8.842610359191895, 'timing/data': 0.0007425602525472641, 'timing/model': 0.9400834608823061, 'learning_rate/qwen_vl_interface': 9.504630861137212e-06, 'learning_rate/action_model': 9.504630861137212e-05, 'epoch': 0.46}) +[2026-06-17 18:30:58,452][starVLA.training.train_starvla][INFO] - Step 1113, Loss: {'train/loss': 0.5156227946281433, 'train/grad_norm_pre_clip': 7.712825298309326, 'timing/data': 0.0006711035966873169, 'timing/model': 0.9462315561249852, 'learning_rate/qwen_vl_interface': 9.503668811855489e-06, 'learning_rate/action_model': 9.503668811855488e-05, 'epoch': 0.46}) +[2026-06-17 18:31:02,472][starVLA.training.train_starvla][INFO] - Step 1114, Loss: {'train/loss': 0.7287274599075317, 'train/grad_norm_pre_clip': 24.740612030029297, 'timing/data': 0.0005615251138806343, 'timing/model': 0.9872669335454702, 'learning_rate/qwen_vl_interface': 9.502705880783644e-06, 'learning_rate/action_model': 9.502705880783645e-05, 'epoch': 0.46}) +[2026-06-17 18:31:06,377][starVLA.training.train_starvla][INFO] - Step 1115, Loss: {'train/loss': 0.7969681024551392, 'train/grad_norm_pre_clip': 10.259121894836426, 'timing/data': 0.0005355160683393478, 'timing/model': 0.9472680995240808, 'learning_rate/qwen_vl_interface': 9.501742068121296e-06, 'learning_rate/action_model': 9.501742068121296e-05, 'epoch': 0.46}) +[2026-06-17 18:31:10,315][starVLA.training.train_starvla][INFO] - Step 1116, Loss: {'train/loss': 1.0160309076309204, 'train/grad_norm_pre_clip': 13.111640930175781, 'timing/data': 0.0005433596670627594, 'timing/model': 0.9389563929289579, 'learning_rate/qwen_vl_interface': 9.500777374068241e-06, 'learning_rate/action_model': 9.50077737406824e-05, 'epoch': 0.46}) +[2026-06-17 18:31:14,456][starVLA.training.train_starvla][INFO] - Step 1117, Loss: {'train/loss': 0.8782660365104675, 'train/grad_norm_pre_clip': 12.014894485473633, 'timing/data': 0.0011937925592064857, 'timing/model': 1.290097525343299, 'learning_rate/qwen_vl_interface': 9.49981179882446e-06, 'learning_rate/action_model': 9.499811798824461e-05, 'epoch': 0.46}) +[2026-06-17 18:31:18,249][starVLA.training.train_starvla][INFO] - Step 1118, Loss: {'train/loss': 0.6348261833190918, 'train/grad_norm_pre_clip': 18.685272216796875, 'timing/data': 0.0008709030225872993, 'timing/model': 0.9483655272051692, 'learning_rate/qwen_vl_interface': 9.498845342590124e-06, 'learning_rate/action_model': 9.498845342590124e-05, 'epoch': 0.46}) +[2026-06-17 18:31:22,331][starVLA.training.train_starvla][INFO] - Step 1119, Loss: {'train/loss': 0.6457728147506714, 'train/grad_norm_pre_clip': 7.890676975250244, 'timing/data': 0.0005326615646481514, 'timing/model': 1.0722981579601765, 'learning_rate/qwen_vl_interface': 9.497878005565575e-06, 'learning_rate/action_model': 9.497878005565574e-05, 'epoch': 0.46}) +[2026-06-17 18:31:28,196][starVLA.training.train_starvla][INFO] - Step 1120, Loss: {'train/loss': 0.7578719258308411, 'train/grad_norm_pre_clip': 9.66543960571289, 'timing/data': 0.001566820777952671, 'timing/model': 0.950648844242096, 'learning_rate/qwen_vl_interface': 9.496909787951344e-06, 'learning_rate/action_model': 9.496909787951343e-05, 'epoch': 0.46}) +[2026-06-17 18:31:32,186][starVLA.training.train_starvla][INFO] - Step 1121, Loss: {'train/loss': 0.5589559078216553, 'train/grad_norm_pre_clip': 9.98507308959961, 'timing/data': 0.0006613461300730705, 'timing/model': 0.9502537623047829, 'learning_rate/qwen_vl_interface': 9.495940689948144e-06, 'learning_rate/action_model': 9.495940689948144e-05, 'epoch': 0.46}) +[2026-06-17 18:31:36,027][starVLA.training.train_starvla][INFO] - Step 1122, Loss: {'train/loss': 0.522493302822113, 'train/grad_norm_pre_clip': 7.761404037475586, 'timing/data': 0.000509326346218586, 'timing/model': 0.9500634344294667, 'learning_rate/qwen_vl_interface': 9.49497071175687e-06, 'learning_rate/action_model': 9.49497071175687e-05, 'epoch': 0.46}) +[2026-06-17 18:31:40,039][starVLA.training.train_starvla][INFO] - Step 1123, Loss: {'train/loss': 0.6065750122070312, 'train/grad_norm_pre_clip': 8.3827486038208, 'timing/data': 0.0007436517626047134, 'timing/model': 0.9421281442046165, 'learning_rate/qwen_vl_interface': 9.493999853578601e-06, 'learning_rate/action_model': 9.493999853578602e-05, 'epoch': 0.46}) +[2026-06-17 18:31:43,999][starVLA.training.train_starvla][INFO] - Step 1124, Loss: {'train/loss': 0.7419034838676453, 'train/grad_norm_pre_clip': 8.93818187713623, 'timing/data': 0.0011574812233448029, 'timing/model': 1.2071602065116167, 'learning_rate/qwen_vl_interface': 9.493028115614595e-06, 'learning_rate/action_model': 9.493028115614595e-05, 'epoch': 0.46}) +[2026-06-17 18:31:47,660][starVLA.training.train_starvla][INFO] - Step 1125, Loss: {'train/loss': 0.7368078231811523, 'train/grad_norm_pre_clip': 9.286988258361816, 'timing/data': 0.0007867058739066124, 'timing/model': 0.9540675096213818, 'learning_rate/qwen_vl_interface': 9.492055498066295e-06, 'learning_rate/action_model': 9.492055498066294e-05, 'epoch': 0.46}) +[2026-06-17 18:31:51,545][starVLA.training.train_starvla][INFO] - Step 1126, Loss: {'train/loss': 0.6620091795921326, 'train/grad_norm_pre_clip': 7.169151306152344, 'timing/data': 0.000562014989554882, 'timing/model': 0.9595103356987238, 'learning_rate/qwen_vl_interface': 9.491082001135326e-06, 'learning_rate/action_model': 9.491082001135325e-05, 'epoch': 0.46}) +[2026-06-17 18:31:55,507][starVLA.training.train_starvla][INFO] - Step 1127, Loss: {'train/loss': 0.5010814666748047, 'train/grad_norm_pre_clip': 8.789831161499023, 'timing/data': 0.000572490505874157, 'timing/model': 0.9613995300605893, 'learning_rate/qwen_vl_interface': 9.490107625023492e-06, 'learning_rate/action_model': 9.490107625023492e-05, 'epoch': 0.46}) +[2026-06-17 18:31:59,758][starVLA.training.train_starvla][INFO] - Step 1128, Loss: {'train/loss': 0.6010617613792419, 'train/grad_norm_pre_clip': 7.6607136726379395, 'timing/data': 0.0005424804985523224, 'timing/model': 0.9755276162177324, 'learning_rate/qwen_vl_interface': 9.489132369932788e-06, 'learning_rate/action_model': 9.489132369932788e-05, 'epoch': 0.46}) +[2026-06-17 18:32:03,817][starVLA.training.train_starvla][INFO] - Step 1129, Loss: {'train/loss': 0.9076828956604004, 'train/grad_norm_pre_clip': 10.007323265075684, 'timing/data': 0.000873025506734848, 'timing/model': 0.9601446408778429, 'learning_rate/qwen_vl_interface': 9.488156236065378e-06, 'learning_rate/action_model': 9.488156236065379e-05, 'epoch': 0.46}) +[2026-06-17 18:32:07,764][starVLA.training.train_starvla][INFO] - Step 1130, Loss: {'train/loss': 0.7094109058380127, 'train/grad_norm_pre_clip': 10.598437309265137, 'timing/data': 0.0005631949752569199, 'timing/model': 1.1962922858074307, 'learning_rate/qwen_vl_interface': 9.487179223623626e-06, 'learning_rate/action_model': 9.487179223623624e-05, 'epoch': 0.47}) +[2026-06-17 18:32:11,670][starVLA.training.train_starvla][INFO] - Step 1131, Loss: {'train/loss': 0.9971142411231995, 'train/grad_norm_pre_clip': 12.14207649230957, 'timing/data': 0.0005253935232758522, 'timing/model': 1.1100098062306643, 'learning_rate/qwen_vl_interface': 9.486201332810058e-06, 'learning_rate/action_model': 9.486201332810057e-05, 'epoch': 0.47}) +[2026-06-17 18:32:15,356][starVLA.training.train_starvla][INFO] - Step 1132, Loss: {'train/loss': 0.6196843385696411, 'train/grad_norm_pre_clip': 7.490480422973633, 'timing/data': 0.0007792459800839424, 'timing/model': 0.9553502444177866, 'learning_rate/qwen_vl_interface': 9.485222563827395e-06, 'learning_rate/action_model': 9.485222563827395e-05, 'epoch': 0.47}) +[2026-06-17 18:32:20,029][starVLA.training.train_starvla][INFO] - Step 1133, Loss: {'train/loss': 0.6966941356658936, 'train/grad_norm_pre_clip': 8.55400562286377, 'timing/data': 0.000970049761235714, 'timing/model': 1.059031218290329, 'learning_rate/qwen_vl_interface': 9.484242916878537e-06, 'learning_rate/action_model': 9.484242916878536e-05, 'epoch': 0.47}) +[2026-06-17 18:32:24,055][starVLA.training.train_starvla][INFO] - Step 1134, Loss: {'train/loss': 0.46179381012916565, 'train/grad_norm_pre_clip': 5.5217671394348145, 'timing/data': 0.0012049432843923569, 'timing/model': 0.9469975531101227, 'learning_rate/qwen_vl_interface': 9.483262392166567e-06, 'learning_rate/action_model': 9.483262392166566e-05, 'epoch': 0.47}) +[2026-06-17 18:32:31,078][starVLA.training.train_starvla][INFO] - Step 1135, Loss: {'train/loss': 0.5420072674751282, 'train/grad_norm_pre_clip': 7.537679195404053, 'timing/data': 0.0005320608615875244, 'timing/model': 0.9443413354456425, 'learning_rate/qwen_vl_interface': 9.482280989894748e-06, 'learning_rate/action_model': 9.482280989894748e-05, 'epoch': 0.47}) +[2026-06-17 18:32:34,957][starVLA.training.train_starvla][INFO] - Step 1136, Loss: {'train/loss': 0.7402355074882507, 'train/grad_norm_pre_clip': 8.101805686950684, 'timing/data': 0.00048653315752744675, 'timing/model': 0.9968472141772509, 'learning_rate/qwen_vl_interface': 9.481298710266525e-06, 'learning_rate/action_model': 9.481298710266526e-05, 'epoch': 0.47}) +[2026-06-17 18:32:39,000][starVLA.training.train_starvla][INFO] - Step 1137, Loss: {'train/loss': 0.8850869536399841, 'train/grad_norm_pre_clip': 8.940083503723145, 'timing/data': 0.0006475904956459999, 'timing/model': 1.2718600304797292, 'learning_rate/qwen_vl_interface': 9.480315553485527e-06, 'learning_rate/action_model': 9.480315553485527e-05, 'epoch': 0.47}) +[2026-06-17 18:32:42,817][starVLA.training.train_starvla][INFO] - Step 1138, Loss: {'train/loss': 0.6677448749542236, 'train/grad_norm_pre_clip': 7.402877330780029, 'timing/data': 0.0011437293142080307, 'timing/model': 1.1212192131206393, 'learning_rate/qwen_vl_interface': 9.479331519755563e-06, 'learning_rate/action_model': 9.479331519755562e-05, 'epoch': 0.47}) +[2026-06-17 18:32:46,628][starVLA.training.train_starvla][INFO] - Step 1139, Loss: {'train/loss': 0.6469197273254395, 'train/grad_norm_pre_clip': 7.422668933868408, 'timing/data': 0.0005509620532393456, 'timing/model': 1.0536434091627598, 'learning_rate/qwen_vl_interface': 9.478346609280622e-06, 'learning_rate/action_model': 9.478346609280623e-05, 'epoch': 0.47}) +[2026-06-17 18:32:50,556][starVLA.training.train_starvla][INFO] - Step 1140, Loss: {'train/loss': 0.8220000267028809, 'train/grad_norm_pre_clip': 7.996811389923096, 'timing/data': 0.0011063320562243462, 'timing/model': 0.9511538799852133, 'learning_rate/qwen_vl_interface': 9.477360822264883e-06, 'learning_rate/action_model': 9.477360822264882e-05, 'epoch': 0.47}) +[2026-06-17 18:32:54,559][starVLA.training.train_starvla][INFO] - Step 1141, Loss: {'train/loss': 0.5368980765342712, 'train/grad_norm_pre_clip': 6.5630412101745605, 'timing/data': 0.0005378331989049911, 'timing/model': 0.9505418483167887, 'learning_rate/qwen_vl_interface': 9.476374158912692e-06, 'learning_rate/action_model': 9.476374158912691e-05, 'epoch': 0.47}) +[2026-06-17 18:32:58,460][starVLA.training.train_starvla][INFO] - Step 1142, Loss: {'train/loss': 0.9997379779815674, 'train/grad_norm_pre_clip': 10.700474739074707, 'timing/data': 0.0006953179836273193, 'timing/model': 1.011095568537712, 'learning_rate/qwen_vl_interface': 9.475386619428593e-06, 'learning_rate/action_model': 9.475386619428594e-05, 'epoch': 0.47}) +[2026-06-17 18:33:03,616][starVLA.training.train_starvla][INFO] - Step 1143, Loss: {'train/loss': 0.6542487144470215, 'train/grad_norm_pre_clip': 8.259273529052734, 'timing/data': 0.0005510319024324417, 'timing/model': 1.5420290566980839, 'learning_rate/qwen_vl_interface': 9.4743982040173e-06, 'learning_rate/action_model': 9.4743982040173e-05, 'epoch': 0.47}) +[2026-06-17 18:33:07,563][starVLA.training.train_starvla][INFO] - Step 1144, Loss: {'train/loss': 0.7226898670196533, 'train/grad_norm_pre_clip': 7.42283296585083, 'timing/data': 0.0005160002037882805, 'timing/model': 1.180241639725864, 'learning_rate/qwen_vl_interface': 9.473408912883714e-06, 'learning_rate/action_model': 9.473408912883714e-05, 'epoch': 0.47}) +[2026-06-17 18:33:11,514][starVLA.training.train_starvla][INFO] - Step 1145, Loss: {'train/loss': 0.8109447956085205, 'train/grad_norm_pre_clip': 9.251903533935547, 'timing/data': 0.0011228537186980247, 'timing/model': 1.1147067053243518, 'learning_rate/qwen_vl_interface': 9.472418746232916e-06, 'learning_rate/action_model': 9.472418746232916e-05, 'epoch': 0.47}) +[2026-06-17 18:33:15,268][starVLA.training.train_starvla][INFO] - Step 1146, Loss: {'train/loss': 0.6535742282867432, 'train/grad_norm_pre_clip': 7.142029762268066, 'timing/data': 0.0008696448057889938, 'timing/model': 0.962865823879838, 'learning_rate/qwen_vl_interface': 9.471427704270168e-06, 'learning_rate/action_model': 9.471427704270169e-05, 'epoch': 0.47}) +[2026-06-17 18:33:19,229][starVLA.training.train_starvla][INFO] - Step 1147, Loss: {'train/loss': 0.6320481300354004, 'train/grad_norm_pre_clip': 6.931243896484375, 'timing/data': 0.0005413778126239777, 'timing/model': 0.9664890794083476, 'learning_rate/qwen_vl_interface': 9.470435787200914e-06, 'learning_rate/action_model': 9.470435787200915e-05, 'epoch': 0.47}) +[2026-06-17 18:33:23,190][starVLA.training.train_starvla][INFO] - Step 1148, Loss: {'train/loss': 0.5680312514305115, 'train/grad_norm_pre_clip': 10.506023406982422, 'timing/data': 0.0004986235871911049, 'timing/model': 0.950475849211216, 'learning_rate/qwen_vl_interface': 9.469442995230781e-06, 'learning_rate/action_model': 9.469442995230782e-05, 'epoch': 0.47}) +[2026-06-17 18:33:27,042][starVLA.training.train_starvla][INFO] - Step 1149, Loss: {'train/loss': 0.4949762225151062, 'train/grad_norm_pre_clip': 7.5405592918396, 'timing/data': 0.0005908533930778503, 'timing/model': 0.9422820722684264, 'learning_rate/qwen_vl_interface': 9.468449328565574e-06, 'learning_rate/action_model': 9.468449328565573e-05, 'epoch': 0.47}) +[2026-06-17 18:33:32,844][starVLA.training.train_starvla][INFO] - Step 1150, Loss: {'train/loss': 0.6545027494430542, 'train/grad_norm_pre_clip': 20.333049774169922, 'timing/data': 0.0009580953046679497, 'timing/model': 2.002286915667355, 'learning_rate/qwen_vl_interface': 9.467454787411282e-06, 'learning_rate/action_model': 9.467454787411281e-05, 'epoch': 0.47}) +[2026-06-17 18:33:38,237][starVLA.training.train_starvla][INFO] - Step 1151, Loss: {'train/loss': 0.6350992918014526, 'train/grad_norm_pre_clip': 8.915488243103027, 'timing/data': 0.0005634212866425514, 'timing/model': 2.472687905654311, 'learning_rate/qwen_vl_interface': 9.466459371974076e-06, 'learning_rate/action_model': 9.466459371974076e-05, 'epoch': 0.47}) +[2026-06-17 18:33:42,005][starVLA.training.train_starvla][INFO] - Step 1152, Loss: {'train/loss': 0.6265347003936768, 'train/grad_norm_pre_clip': 9.686264038085938, 'timing/data': 0.0008227396756410599, 'timing/model': 0.9560252837836742, 'learning_rate/qwen_vl_interface': 9.465463082460304e-06, 'learning_rate/action_model': 9.465463082460304e-05, 'epoch': 0.47}) +[2026-06-17 18:33:46,500][starVLA.training.train_starvla][INFO] - Step 1153, Loss: {'train/loss': 0.6407198905944824, 'train/grad_norm_pre_clip': 9.061114311218262, 'timing/data': 0.0005696751177310944, 'timing/model': 1.0418493300676346, 'learning_rate/qwen_vl_interface': 9.4644659190765e-06, 'learning_rate/action_model': 9.464465919076499e-05, 'epoch': 0.47}) +[2026-06-17 18:33:50,605][starVLA.training.train_starvla][INFO] - Step 1154, Loss: {'train/loss': 1.0273410081863403, 'train/grad_norm_pre_clip': 15.440984725952148, 'timing/data': 0.000654284842312336, 'timing/model': 0.9483650270849466, 'learning_rate/qwen_vl_interface': 9.463467882029374e-06, 'learning_rate/action_model': 9.463467882029373e-05, 'epoch': 0.48}) +[2026-06-17 18:33:54,590][starVLA.training.train_starvla][INFO] - Step 1155, Loss: {'train/loss': 1.0945855379104614, 'train/grad_norm_pre_clip': 11.596880912780762, 'timing/data': 0.0006148330867290497, 'timing/model': 0.971992613747716, 'learning_rate/qwen_vl_interface': 9.462468971525822e-06, 'learning_rate/action_model': 9.462468971525821e-05, 'epoch': 0.48}) +[2026-06-17 18:33:58,460][starVLA.training.train_starvla][INFO] - Step 1156, Loss: {'train/loss': 0.6304644346237183, 'train/grad_norm_pre_clip': 8.072566032409668, 'timing/data': 0.0005174325779080391, 'timing/model': 0.9463313221931458, 'learning_rate/qwen_vl_interface': 9.461469187772921e-06, 'learning_rate/action_model': 9.46146918777292e-05, 'epoch': 0.48}) +[2026-06-17 18:34:02,417][starVLA.training.train_starvla][INFO] - Step 1157, Loss: {'train/loss': 0.8907821774482727, 'train/grad_norm_pre_clip': 9.423221588134766, 'timing/data': 0.0005557667464017868, 'timing/model': 1.194282486103475, 'learning_rate/qwen_vl_interface': 9.460468530977924e-06, 'learning_rate/action_model': 9.460468530977924e-05, 'epoch': 0.48}) +[2026-06-17 18:34:06,239][starVLA.training.train_starvla][INFO] - Step 1158, Loss: {'train/loss': 0.8823601603507996, 'train/grad_norm_pre_clip': 37.42436599731445, 'timing/data': 0.0005373461171984673, 'timing/model': 0.9804726997390389, 'learning_rate/qwen_vl_interface': 9.459467001348271e-06, 'learning_rate/action_model': 9.45946700134827e-05, 'epoch': 0.48}) +[2026-06-17 18:34:10,154][starVLA.training.train_starvla][INFO] - Step 1159, Loss: {'train/loss': 0.6311783194541931, 'train/grad_norm_pre_clip': 8.49333381652832, 'timing/data': 0.0005479538813233376, 'timing/model': 0.946049134247005, 'learning_rate/qwen_vl_interface': 9.458464599091578e-06, 'learning_rate/action_model': 9.458464599091578e-05, 'epoch': 0.48}) +[2026-06-17 18:34:14,176][starVLA.training.train_starvla][INFO] - Step 1160, Loss: {'train/loss': 0.6966087222099304, 'train/grad_norm_pre_clip': 9.039677619934082, 'timing/data': 0.0007145954295992851, 'timing/model': 0.9912697542458773, 'learning_rate/qwen_vl_interface': 9.457461324415647e-06, 'learning_rate/action_model': 9.457461324415646e-05, 'epoch': 0.48}) +[2026-06-17 18:34:18,095][starVLA.training.train_starvla][INFO] - Step 1161, Loss: {'train/loss': 0.7323076128959656, 'train/grad_norm_pre_clip': 46.381839752197266, 'timing/data': 0.0005229515954852104, 'timing/model': 0.9629112463444471, 'learning_rate/qwen_vl_interface': 9.456457177528454e-06, 'learning_rate/action_model': 9.456457177528453e-05, 'epoch': 0.48}) +[2026-06-17 18:34:22,036][starVLA.training.train_starvla][INFO] - Step 1162, Loss: {'train/loss': 0.6678223609924316, 'train/grad_norm_pre_clip': 9.274641990661621, 'timing/data': 0.0005541006103157997, 'timing/model': 0.9689599080011249, 'learning_rate/qwen_vl_interface': 9.455452158638163e-06, 'learning_rate/action_model': 9.455452158638163e-05, 'epoch': 0.48}) +[2026-06-17 18:34:25,926][starVLA.training.train_starvla][INFO] - Step 1163, Loss: {'train/loss': 0.7399789094924927, 'train/grad_norm_pre_clip': 10.219919204711914, 'timing/data': 0.0005998723208904266, 'timing/model': 1.1340154046192765, 'learning_rate/qwen_vl_interface': 9.454446267953114e-06, 'learning_rate/action_model': 9.454446267953114e-05, 'epoch': 0.48}) +[2026-06-17 18:34:29,738][starVLA.training.train_starvla][INFO] - Step 1164, Loss: {'train/loss': 0.8827012181282043, 'train/grad_norm_pre_clip': 9.032121658325195, 'timing/data': 0.0011198315769433975, 'timing/model': 1.0077934805303812, 'learning_rate/qwen_vl_interface': 9.453439505681829e-06, 'learning_rate/action_model': 9.453439505681829e-05, 'epoch': 0.48}) +[2026-06-17 18:34:34,274][starVLA.training.train_starvla][INFO] - Step 1165, Loss: {'train/loss': 0.6307756304740906, 'train/grad_norm_pre_clip': 10.116276741027832, 'timing/data': 0.0005654264241456985, 'timing/model': 0.954031596891582, 'learning_rate/qwen_vl_interface': 9.452431872033012e-06, 'learning_rate/action_model': 9.452431872033012e-05, 'epoch': 0.48}) +[2026-06-17 18:34:38,162][starVLA.training.train_starvla][INFO] - Step 1166, Loss: {'train/loss': 0.5353506803512573, 'train/grad_norm_pre_clip': 9.541873931884766, 'timing/data': 0.0006854888051748276, 'timing/model': 0.9397928090766072, 'learning_rate/qwen_vl_interface': 9.451423367215546e-06, 'learning_rate/action_model': 9.451423367215545e-05, 'epoch': 0.48}) +[2026-06-17 18:34:42,160][starVLA.training.train_starvla][INFO] - Step 1167, Loss: {'train/loss': 0.6766106486320496, 'train/grad_norm_pre_clip': 8.843215942382812, 'timing/data': 0.0010346509516239166, 'timing/model': 0.9610136169940233, 'learning_rate/qwen_vl_interface': 9.450413991438494e-06, 'learning_rate/action_model': 9.450413991438494e-05, 'epoch': 0.48}) +[2026-06-17 18:34:46,066][starVLA.training.train_starvla][INFO] - Step 1168, Loss: {'train/loss': 0.6958468556404114, 'train/grad_norm_pre_clip': 9.983870506286621, 'timing/data': 0.0008981972932815552, 'timing/model': 0.9670328833162785, 'learning_rate/qwen_vl_interface': 9.449403744911104e-06, 'learning_rate/action_model': 9.449403744911103e-05, 'epoch': 0.48}) +[2026-06-17 18:34:50,038][starVLA.training.train_starvla][INFO] - Step 1169, Loss: {'train/loss': 0.5294210314750671, 'train/grad_norm_pre_clip': 8.680164337158203, 'timing/data': 0.0012259790673851967, 'timing/model': 0.9473986178636551, 'learning_rate/qwen_vl_interface': 9.448392627842797e-06, 'learning_rate/action_model': 9.448392627842795e-05, 'epoch': 0.48}) +[2026-06-17 18:34:53,993][starVLA.training.train_starvla][INFO] - Step 1170, Loss: {'train/loss': 0.623367965221405, 'train/grad_norm_pre_clip': 7.799173355102539, 'timing/data': 0.00072470773011446, 'timing/model': 1.1586022451519966, 'learning_rate/qwen_vl_interface': 9.44738064044318e-06, 'learning_rate/action_model': 9.44738064044318e-05, 'epoch': 0.48}) +[2026-06-17 18:34:57,750][starVLA.training.train_starvla][INFO] - Step 1171, Loss: {'train/loss': 0.6330068707466125, 'train/grad_norm_pre_clip': 161.86087036132812, 'timing/data': 0.0016731377691030502, 'timing/model': 0.9427876537665725, 'learning_rate/qwen_vl_interface': 9.44636778292204e-06, 'learning_rate/action_model': 9.44636778292204e-05, 'epoch': 0.48}) +[2026-06-17 18:35:01,630][starVLA.training.train_starvla][INFO] - Step 1172, Loss: {'train/loss': 0.4677451252937317, 'train/grad_norm_pre_clip': 12.966954231262207, 'timing/data': 0.0007639564573764801, 'timing/model': 0.9557720925658941, 'learning_rate/qwen_vl_interface': 9.445354055489345e-06, 'learning_rate/action_model': 9.445354055489345e-05, 'epoch': 0.48}) +[2026-06-17 18:35:05,544][starVLA.training.train_starvla][INFO] - Step 1173, Loss: {'train/loss': 0.6561225056648254, 'train/grad_norm_pre_clip': 7.831578254699707, 'timing/data': 0.0012402897700667381, 'timing/model': 0.944925949908793, 'learning_rate/qwen_vl_interface': 9.444339458355239e-06, 'learning_rate/action_model': 9.444339458355239e-05, 'epoch': 0.48}) +[2026-06-17 18:35:09,712][starVLA.training.train_starvla][INFO] - Step 1174, Loss: {'train/loss': 0.7490177750587463, 'train/grad_norm_pre_clip': 12.305435180664062, 'timing/data': 0.0006834659725427628, 'timing/model': 0.9654692523181438, 'learning_rate/qwen_vl_interface': 9.443323991730049e-06, 'learning_rate/action_model': 9.443323991730049e-05, 'epoch': 0.48}) +[2026-06-17 18:35:13,688][starVLA.training.train_starvla][INFO] - Step 1175, Loss: {'train/loss': 0.5950086712837219, 'train/grad_norm_pre_clip': 25.678037643432617, 'timing/data': 0.0009784037247300148, 'timing/model': 0.9614397855475545, 'learning_rate/qwen_vl_interface': 9.442307655824284e-06, 'learning_rate/action_model': 9.442307655824285e-05, 'epoch': 0.48}) +[2026-06-17 18:35:17,547][starVLA.training.train_starvla][INFO] - Step 1176, Loss: {'train/loss': 0.6214767098426819, 'train/grad_norm_pre_clip': 7.624270439147949, 'timing/data': 0.0009016860276460648, 'timing/model': 1.1467440417036414, 'learning_rate/qwen_vl_interface': 9.441290450848631e-06, 'learning_rate/action_model': 9.44129045084863e-05, 'epoch': 0.48}) +[2026-06-17 18:35:21,617][starVLA.training.train_starvla][INFO] - Step 1177, Loss: {'train/loss': 0.9383158683776855, 'train/grad_norm_pre_clip': 10.165115356445312, 'timing/data': 0.0010876022279262543, 'timing/model': 1.3035874152556062, 'learning_rate/qwen_vl_interface': 9.440272377013957e-06, 'learning_rate/action_model': 9.440272377013956e-05, 'epoch': 0.48}) +[2026-06-17 18:35:25,454][starVLA.training.train_starvla][INFO] - Step 1178, Loss: {'train/loss': 1.0988801717758179, 'train/grad_norm_pre_clip': 10.917778968811035, 'timing/data': 0.0006948960945010185, 'timing/model': 0.9460460487753153, 'learning_rate/qwen_vl_interface': 9.43925343453131e-06, 'learning_rate/action_model': 9.43925343453131e-05, 'epoch': 0.48}) +[2026-06-17 18:35:29,490][starVLA.training.train_starvla][INFO] - Step 1179, Loss: {'train/loss': 0.6889311075210571, 'train/grad_norm_pre_clip': 7.9942827224731445, 'timing/data': 0.0006363345310091972, 'timing/model': 0.9558372832834721, 'learning_rate/qwen_vl_interface': 9.438233623611918e-06, 'learning_rate/action_model': 9.438233623611918e-05, 'epoch': 0.49}) +[2026-06-17 18:35:33,419][starVLA.training.train_starvla][INFO] - Step 1180, Loss: {'train/loss': 0.6037057042121887, 'train/grad_norm_pre_clip': 9.130049705505371, 'timing/data': 0.0005014669150114059, 'timing/model': 0.9518999643623829, 'learning_rate/qwen_vl_interface': 9.43721294446719e-06, 'learning_rate/action_model': 9.437212944467191e-05, 'epoch': 0.49}) +[2026-06-17 18:35:39,000][starVLA.training.train_starvla][INFO] - Step 1181, Loss: {'train/loss': 0.5943271517753601, 'train/grad_norm_pre_clip': 7.529227256774902, 'timing/data': 0.0005689943209290504, 'timing/model': 0.9727307232096791, 'learning_rate/qwen_vl_interface': 9.436191397308713e-06, 'learning_rate/action_model': 9.436191397308713e-05, 'epoch': 0.49}) +[2026-06-17 18:35:43,060][starVLA.training.train_starvla][INFO] - Step 1182, Loss: {'train/loss': 0.4744589626789093, 'train/grad_norm_pre_clip': 7.599636554718018, 'timing/data': 0.00047882460057735443, 'timing/model': 0.9974337238818407, 'learning_rate/qwen_vl_interface': 9.435168982348254e-06, 'learning_rate/action_model': 9.435168982348255e-05, 'epoch': 0.49}) +[2026-06-17 18:35:47,153][starVLA.training.train_starvla][INFO] - Step 1183, Loss: {'train/loss': 0.5724349617958069, 'train/grad_norm_pre_clip': 7.539844989776611, 'timing/data': 0.0005860347300767899, 'timing/model': 1.2340216394513845, 'learning_rate/qwen_vl_interface': 9.434145699797763e-06, 'learning_rate/action_model': 9.434145699797764e-05, 'epoch': 0.49}) +[2026-06-17 18:35:51,042][starVLA.training.train_starvla][INFO] - Step 1184, Loss: {'train/loss': 0.2751193344593048, 'train/grad_norm_pre_clip': 4.909114360809326, 'timing/data': 0.001270887441933155, 'timing/model': 1.1458945162594318, 'learning_rate/qwen_vl_interface': 9.433121549869365e-06, 'learning_rate/action_model': 9.433121549869365e-05, 'epoch': 0.49}) +[2026-06-17 18:35:54,906][starVLA.training.train_starvla][INFO] - Step 1185, Loss: {'train/loss': 0.6694881319999695, 'train/grad_norm_pre_clip': 9.248658180236816, 'timing/data': 0.0009876452386379242, 'timing/model': 0.9525384027510881, 'learning_rate/qwen_vl_interface': 9.43209653277537e-06, 'learning_rate/action_model': 9.432096532775369e-05, 'epoch': 0.49}) +[2026-06-17 18:35:58,886][starVLA.training.train_starvla][INFO] - Step 1186, Loss: {'train/loss': 0.5073751211166382, 'train/grad_norm_pre_clip': 7.96865701675415, 'timing/data': 0.0008407328277826309, 'timing/model': 0.9613158795982599, 'learning_rate/qwen_vl_interface': 9.431070648728262e-06, 'learning_rate/action_model': 9.431070648728261e-05, 'epoch': 0.49}) +[2026-06-17 18:36:02,782][starVLA.training.train_starvla][INFO] - Step 1187, Loss: {'train/loss': 0.6120427250862122, 'train/grad_norm_pre_clip': 11.327491760253906, 'timing/data': 0.0010898606851696968, 'timing/model': 0.9546039886772633, 'learning_rate/qwen_vl_interface': 9.430043897940709e-06, 'learning_rate/action_model': 9.43004389794071e-05, 'epoch': 0.49}) +[2026-06-17 18:36:06,595][starVLA.training.train_starvla][INFO] - Step 1188, Loss: {'train/loss': 0.8219190239906311, 'train/grad_norm_pre_clip': 10.300148963928223, 'timing/data': 0.0007921867072582245, 'timing/model': 0.9576418269425631, 'learning_rate/qwen_vl_interface': 9.429016280625559e-06, 'learning_rate/action_model': 9.429016280625559e-05, 'epoch': 0.49}) +[2026-06-17 18:36:10,621][starVLA.training.train_starvla][INFO] - Step 1189, Loss: {'train/loss': 0.3901335597038269, 'train/grad_norm_pre_clip': 6.446084976196289, 'timing/data': 0.0008024526759982109, 'timing/model': 1.0424293475225568, 'learning_rate/qwen_vl_interface': 9.427987796995838e-06, 'learning_rate/action_model': 9.427987796995837e-05, 'epoch': 0.49}) +[2026-06-17 18:36:14,733][starVLA.training.train_starvla][INFO] - Step 1190, Loss: {'train/loss': 0.4515053927898407, 'train/grad_norm_pre_clip': 6.1185126304626465, 'timing/data': 0.0005755526944994926, 'timing/model': 1.2828304376453161, 'learning_rate/qwen_vl_interface': 9.42695844726475e-06, 'learning_rate/action_model': 9.426958447264749e-05, 'epoch': 0.49}) +[2026-06-17 18:36:18,472][starVLA.training.train_starvla][INFO] - Step 1191, Loss: {'train/loss': 0.7408892512321472, 'train/grad_norm_pre_clip': 12.153436660766602, 'timing/data': 0.0011166995391249657, 'timing/model': 0.9734013993293047, 'learning_rate/qwen_vl_interface': 9.42592823164568e-06, 'learning_rate/action_model': 9.42592823164568e-05, 'epoch': 0.49}) +[2026-06-17 18:36:22,667][starVLA.training.train_starvla][INFO] - Step 1192, Loss: {'train/loss': 0.8944109082221985, 'train/grad_norm_pre_clip': 12.796204566955566, 'timing/data': 0.0005682501941919327, 'timing/model': 0.994413728825748, 'learning_rate/qwen_vl_interface': 9.424897150352197e-06, 'learning_rate/action_model': 9.424897150352197e-05, 'epoch': 0.49}) +[2026-06-17 18:36:27,345][starVLA.training.train_starvla][INFO] - Step 1193, Loss: {'train/loss': 0.7769636511802673, 'train/grad_norm_pre_clip': 8.398209571838379, 'timing/data': 0.0013831304386258125, 'timing/model': 0.951253242790699, 'learning_rate/qwen_vl_interface': 9.42386520359804e-06, 'learning_rate/action_model': 9.42386520359804e-05, 'epoch': 0.49}) +[2026-06-17 18:36:32,055][starVLA.training.train_starvla][INFO] - Step 1194, Loss: {'train/loss': 0.5874313712120056, 'train/grad_norm_pre_clip': 10.519020080566406, 'timing/data': 0.0006691934540867805, 'timing/model': 0.9607755048200488, 'learning_rate/qwen_vl_interface': 9.422832391597136e-06, 'learning_rate/action_model': 9.422832391597136e-05, 'epoch': 0.49}) +[2026-06-17 18:36:36,070][starVLA.training.train_starvla][INFO] - Step 1195, Loss: {'train/loss': 0.5513835549354553, 'train/grad_norm_pre_clip': 10.2759428024292, 'timing/data': 0.0009284727275371552, 'timing/model': 0.968739865347743, 'learning_rate/qwen_vl_interface': 9.421798714563586e-06, 'learning_rate/action_model': 9.421798714563586e-05, 'epoch': 0.49}) +[2026-06-17 18:36:40,582][starVLA.training.train_starvla][INFO] - Step 1196, Loss: {'train/loss': 1.0736230611801147, 'train/grad_norm_pre_clip': 13.359922409057617, 'timing/data': 0.0008474988862872124, 'timing/model': 1.2542479410767555, 'learning_rate/qwen_vl_interface': 9.420764172711676e-06, 'learning_rate/action_model': 9.420764172711676e-05, 'epoch': 0.49}) +[2026-06-17 18:36:44,461][starVLA.training.train_starvla][INFO] - Step 1197, Loss: {'train/loss': 0.9042174816131592, 'train/grad_norm_pre_clip': 10.173114776611328, 'timing/data': 0.0014664549380540848, 'timing/model': 1.1151683432981372, 'learning_rate/qwen_vl_interface': 9.419728766255863e-06, 'learning_rate/action_model': 9.419728766255862e-05, 'epoch': 0.49}) +[2026-06-17 18:36:48,355][starVLA.training.train_starvla][INFO] - Step 1198, Loss: {'train/loss': 0.5603525042533875, 'train/grad_norm_pre_clip': 8.328765869140625, 'timing/data': 0.0008926009759306908, 'timing/model': 1.0183415161445737, 'learning_rate/qwen_vl_interface': 9.418692495410789e-06, 'learning_rate/action_model': 9.418692495410789e-05, 'epoch': 0.49}) +[2026-06-17 18:36:52,366][starVLA.training.train_starvla][INFO] - Step 1199, Loss: {'train/loss': 0.838706374168396, 'train/grad_norm_pre_clip': 11.877137184143066, 'timing/data': 0.0011636102572083473, 'timing/model': 0.9598239986225963, 'learning_rate/qwen_vl_interface': 9.417655360391277e-06, 'learning_rate/action_model': 9.417655360391277e-05, 'epoch': 0.49}) +[2026-06-17 18:38:59,952][starVLA.training.train_starvla][INFO] - Step 1200, Loss: {'train/loss': 0.7407113909721375, 'train/grad_norm_pre_clip': 10.19324779510498, 'eval/loss': 0.6772504448890686, 'eval/latency_6/loss': 0.6772504448890686, 'eval/demon_fire_precision': 0.9423868312757202, 'eval/demon_fire_recall': 0.9621848739495799, 'eval/demon_fire_f1': 0.9521829521829522, 'eval/demon_fire_rate_model': 0.8789649415692822, 'eval/demon_fire_rate_teacher': 0.860879243183083, 'eval/demon_left_precision': 0.6915322580645161, 'eval/demon_left_recall': 0.6017543859649123, 'eval/demon_left_f1': 0.6435272045028142, 'eval/demon_left_rate_model': 0.13800779076238176, 'eval/demon_left_rate_teacher': 0.15859766277128548, 'eval/demon_right_precision': 0.751004016064257, 'eval/demon_right_recall': 0.5745007680491552, 'eval/demon_right_f1': 0.6510008703220191, 'eval/demon_right_rate_model': 0.13856427378964942, 'eval/demon_right_rate_teacher': 0.18113522537562604, 'eval/demon_move_f1': 0.6472640374124167, 'eval/demon_cc_f1': 0.8342773132463084, 'eval/latency_6/demon_fire_precision': 0.9423868312757202, 'eval/latency_6/demon_fire_recall': 0.9621848739495799, 'eval/latency_6/demon_fire_f1': 0.9521829521829522, 'eval/latency_6/demon_fire_rate_model': 0.8789649415692822, 'eval/latency_6/demon_fire_rate_teacher': 0.860879243183083, 'eval/latency_6/demon_left_precision': 0.6915322580645161, 'eval/latency_6/demon_left_recall': 0.6017543859649123, 'eval/latency_6/demon_left_f1': 0.6435272045028142, 'eval/latency_6/demon_left_rate_model': 0.13800779076238176, 'eval/latency_6/demon_left_rate_teacher': 0.15859766277128548, 'eval/latency_6/demon_right_precision': 0.751004016064257, 'eval/latency_6/demon_right_recall': 0.5745007680491552, 'eval/latency_6/demon_right_f1': 0.6510008703220191, 'eval/latency_6/demon_right_rate_model': 0.13856427378964942, 'eval/latency_6/demon_right_rate_teacher': 0.18113522537562604, 'eval/latency_6/demon_move_f1': 0.6472640374124167, 'eval/latency_6/demon_cc_f1': 0.8342773132463084, 'timing/data': 0.0005411775782704353, 'timing/model': 0.9967928975820541, 'learning_rate/qwen_vl_interface': 9.416617361412325e-06, 'learning_rate/action_model': 9.416617361412325e-05, 'epoch': 0.49}) +[2026-06-17 18:38:59,966][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_1200_state +[2026-06-17 18:38:59,971][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 18:39:06,686][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_1200_state/model.safetensors +[2026-06-17 18:39:21,627][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_1200_state/optimizer.bin +[2026-06-17 18:39:21,629][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_1200_state/sampler.bin +[2026-06-17 18:39:21,629][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_1200_state/sampler_1.bin +[2026-06-17 18:39:21,632][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_1200_state/random_states_0.pkl +[2026-06-17 18:39:21,638][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1200 +[2026-06-17 18:39:21,638][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 18:39:21,649][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 18:39:25,995][starVLA.training.train_starvla][INFO] - Step 1201, Loss: {'train/loss': 0.5673428177833557, 'train/grad_norm_pre_clip': 8.668540000915527, 'timing/data': 0.001016564667224884, 'timing/model': 0.961088833399117, 'learning_rate/qwen_vl_interface': 9.415578498689106e-06, 'learning_rate/action_model': 9.415578498689106e-05, 'epoch': 0.49}) +[2026-06-17 18:39:30,167][starVLA.training.train_starvla][INFO] - Step 1202, Loss: {'train/loss': 0.5212225317955017, 'train/grad_norm_pre_clip': 9.736268043518066, 'timing/data': 0.0012483205646276474, 'timing/model': 1.095719770528376, 'learning_rate/qwen_vl_interface': 9.414538772436987e-06, 'learning_rate/action_model': 9.414538772436986e-05, 'epoch': 0.49}) +[2026-06-17 18:39:36,100][starVLA.training.train_starvla][INFO] - Step 1203, Loss: {'train/loss': 0.7914652824401855, 'train/grad_norm_pre_clip': 11.622838020324707, 'timing/data': 0.0005532000213861465, 'timing/model': 2.6878144638612866, 'learning_rate/qwen_vl_interface': 9.413498182871496e-06, 'learning_rate/action_model': 9.413498182871496e-05, 'epoch': 0.5}) +[2026-06-17 18:39:41,719][starVLA.training.train_starvla][INFO] - Step 1204, Loss: {'train/loss': 0.5002698302268982, 'train/grad_norm_pre_clip': 11.132174491882324, 'timing/data': 0.0005399361252784729, 'timing/model': 0.9751001270487905, 'learning_rate/qwen_vl_interface': 9.412456730208352e-06, 'learning_rate/action_model': 9.412456730208352e-05, 'epoch': 0.5}) +[2026-06-17 18:39:48,839][starVLA.training.train_starvla][INFO] - Step 1205, Loss: {'train/loss': 0.5575304627418518, 'train/grad_norm_pre_clip': 8.052539825439453, 'timing/data': 0.0011879615485668182, 'timing/model': 1.0446063131093979, 'learning_rate/qwen_vl_interface': 9.41141441466345e-06, 'learning_rate/action_model': 9.41141441466345e-05, 'epoch': 0.5}) +[2026-06-17 18:39:53,198][starVLA.training.train_starvla][INFO] - Step 1206, Loss: {'train/loss': 0.7198522090911865, 'train/grad_norm_pre_clip': 9.830401420593262, 'timing/data': 0.0005659842863678932, 'timing/model': 1.168680283240974, 'learning_rate/qwen_vl_interface': 9.41037123645286e-06, 'learning_rate/action_model': 9.41037123645286e-05, 'epoch': 0.5}) +[2026-06-17 18:39:56,851][starVLA.training.train_starvla][INFO] - Step 1207, Loss: {'train/loss': 0.8234420418739319, 'train/grad_norm_pre_clip': 9.563154220581055, 'timing/data': 0.0005315598100423813, 'timing/model': 0.9353042338043451, 'learning_rate/qwen_vl_interface': 9.409327195792838e-06, 'learning_rate/action_model': 9.409327195792838e-05, 'epoch': 0.5}) +[2026-06-17 18:40:00,805][starVLA.training.train_starvla][INFO] - Step 1208, Loss: {'train/loss': 0.5780898332595825, 'train/grad_norm_pre_clip': 7.2580976486206055, 'timing/data': 0.0008816737681627274, 'timing/model': 1.0207434706389904, 'learning_rate/qwen_vl_interface': 9.408282292899812e-06, 'learning_rate/action_model': 9.408282292899812e-05, 'epoch': 0.5}) +[2026-06-17 18:40:04,800][starVLA.training.train_starvla][INFO] - Step 1209, Loss: {'train/loss': 0.7347649931907654, 'train/grad_norm_pre_clip': 11.241174697875977, 'timing/data': 0.001064506359398365, 'timing/model': 1.0641827518120408, 'learning_rate/qwen_vl_interface': 9.407236527990394e-06, 'learning_rate/action_model': 9.407236527990393e-05, 'epoch': 0.5}) +[2026-06-17 18:40:08,896][starVLA.training.train_starvla][INFO] - Step 1210, Loss: {'train/loss': 0.6204994916915894, 'train/grad_norm_pre_clip': 8.846151351928711, 'timing/data': 0.0007057851180434227, 'timing/model': 0.9520214814692736, 'learning_rate/qwen_vl_interface': 9.406189901281369e-06, 'learning_rate/action_model': 9.406189901281368e-05, 'epoch': 0.5}) +[2026-06-17 18:40:12,815][starVLA.training.train_starvla][INFO] - Step 1211, Loss: {'train/loss': 0.5567498207092285, 'train/grad_norm_pre_clip': 15.782051086425781, 'timing/data': 0.001004100777208805, 'timing/model': 0.9649311322718859, 'learning_rate/qwen_vl_interface': 9.405142412989705e-06, 'learning_rate/action_model': 9.405142412989704e-05, 'epoch': 0.5}) +[2026-06-17 18:40:16,801][starVLA.training.train_starvla][INFO] - Step 1212, Loss: {'train/loss': 1.0033423900604248, 'train/grad_norm_pre_clip': 10.73214340209961, 'timing/data': 0.0010395655408501625, 'timing/model': 1.0098825749009848, 'learning_rate/qwen_vl_interface': 9.404094063332546e-06, 'learning_rate/action_model': 9.404094063332546e-05, 'epoch': 0.5}) +[2026-06-17 18:40:20,762][starVLA.training.train_starvla][INFO] - Step 1213, Loss: {'train/loss': 0.5935816168785095, 'train/grad_norm_pre_clip': 7.475174903869629, 'timing/data': 0.0008767498657107353, 'timing/model': 1.176626207306981, 'learning_rate/qwen_vl_interface': 9.40304485252722e-06, 'learning_rate/action_model': 9.403044852527219e-05, 'epoch': 0.5}) +[2026-06-17 18:40:24,516][starVLA.training.train_starvla][INFO] - Step 1214, Loss: {'train/loss': 0.5585964918136597, 'train/grad_norm_pre_clip': 10.047184944152832, 'timing/data': 0.0006084144115447998, 'timing/model': 0.9592249784618616, 'learning_rate/qwen_vl_interface': 9.401994780791226e-06, 'learning_rate/action_model': 9.401994780791226e-05, 'epoch': 0.5}) +[2026-06-17 18:40:28,539][starVLA.training.train_starvla][INFO] - Step 1215, Loss: {'train/loss': 0.6954125165939331, 'train/grad_norm_pre_clip': 9.860665321350098, 'timing/data': 0.0011031199246644974, 'timing/model': 0.9684044355526567, 'learning_rate/qwen_vl_interface': 9.400943848342246e-06, 'learning_rate/action_model': 9.400943848342246e-05, 'epoch': 0.5}) +[2026-06-17 18:40:32,658][starVLA.training.train_starvla][INFO] - Step 1216, Loss: {'train/loss': 0.936927318572998, 'train/grad_norm_pre_clip': 8.781192779541016, 'timing/data': 0.0018974095582962036, 'timing/model': 0.9648432498797774, 'learning_rate/qwen_vl_interface': 9.39989205539814e-06, 'learning_rate/action_model': 9.39989205539814e-05, 'epoch': 0.5}) +[2026-06-17 18:40:36,746][starVLA.training.train_starvla][INFO] - Step 1217, Loss: {'train/loss': 0.42573782801628113, 'train/grad_norm_pre_clip': 8.419722557067871, 'timing/data': 0.0009208396077156067, 'timing/model': 0.9552997322753072, 'learning_rate/qwen_vl_interface': 9.398839402176946e-06, 'learning_rate/action_model': 9.398839402176946e-05, 'epoch': 0.5}) +[2026-06-17 18:40:40,833][starVLA.training.train_starvla][INFO] - Step 1218, Loss: {'train/loss': 0.5966525077819824, 'train/grad_norm_pre_clip': 16.032033920288086, 'timing/data': 0.0006104176864027977, 'timing/model': 1.0683921175077558, 'learning_rate/qwen_vl_interface': 9.39778588889688e-06, 'learning_rate/action_model': 9.39778588889688e-05, 'epoch': 0.5}) +[2026-06-17 18:40:44,865][starVLA.training.train_starvla][INFO] - Step 1219, Loss: {'train/loss': 0.7317638993263245, 'train/grad_norm_pre_clip': 8.668839454650879, 'timing/data': 0.0005471715703606606, 'timing/model': 1.2177457436919212, 'learning_rate/qwen_vl_interface': 9.396731515776335e-06, 'learning_rate/action_model': 9.396731515776333e-05, 'epoch': 0.5}) +[2026-06-17 18:40:48,914][starVLA.training.train_starvla][INFO] - Step 1220, Loss: {'train/loss': 0.8326981663703918, 'train/grad_norm_pre_clip': 10.521201133728027, 'timing/data': 0.0006358008831739426, 'timing/model': 1.1977038783952594, 'learning_rate/qwen_vl_interface': 9.395676283033884e-06, 'learning_rate/action_model': 9.395676283033885e-05, 'epoch': 0.5}) +[2026-06-17 18:40:52,560][starVLA.training.train_starvla][INFO] - Step 1221, Loss: {'train/loss': 0.8994572758674622, 'train/grad_norm_pre_clip': 10.352493286132812, 'timing/data': 0.0006137257441878319, 'timing/model': 0.9525307919830084, 'learning_rate/qwen_vl_interface': 9.39462019088828e-06, 'learning_rate/action_model': 9.394620190888279e-05, 'epoch': 0.5}) +[2026-06-17 18:40:56,865][starVLA.training.train_starvla][INFO] - Step 1222, Loss: {'train/loss': 0.6311103701591492, 'train/grad_norm_pre_clip': 9.409421920776367, 'timing/data': 0.001125073991715908, 'timing/model': 1.2649107621982694, 'learning_rate/qwen_vl_interface': 9.393563239558449e-06, 'learning_rate/action_model': 9.39356323955845e-05, 'epoch': 0.5}) +[2026-06-17 18:41:02,036][starVLA.training.train_starvla][INFO] - Step 1223, Loss: {'train/loss': 0.850861668586731, 'train/grad_norm_pre_clip': 10.343233108520508, 'timing/data': 0.0013876557350158691, 'timing/model': 1.1570958672091365, 'learning_rate/qwen_vl_interface': 9.392505429263501e-06, 'learning_rate/action_model': 9.392505429263501e-05, 'epoch': 0.5}) +[2026-06-17 18:41:06,322][starVLA.training.train_starvla][INFO] - Step 1224, Loss: {'train/loss': 0.617935061454773, 'train/grad_norm_pre_clip': 8.406964302062988, 'timing/data': 0.0008784392848610878, 'timing/model': 0.9683485329151154, 'learning_rate/qwen_vl_interface': 9.39144676022272e-06, 'learning_rate/action_model': 9.39144676022272e-05, 'epoch': 0.5}) +[2026-06-17 18:41:10,326][starVLA.training.train_starvla][INFO] - Step 1225, Loss: {'train/loss': 0.5431358814239502, 'train/grad_norm_pre_clip': 8.598345756530762, 'timing/data': 0.0005417913198471069, 'timing/model': 0.9861037768423557, 'learning_rate/qwen_vl_interface': 9.390387232655568e-06, 'learning_rate/action_model': 9.390387232655569e-05, 'epoch': 0.5}) +[2026-06-17 18:41:14,356][starVLA.training.train_starvla][INFO] - Step 1226, Loss: {'train/loss': 0.7420427799224854, 'train/grad_norm_pre_clip': 10.687439918518066, 'timing/data': 0.0007920721545815468, 'timing/model': 1.240726470015943, 'learning_rate/qwen_vl_interface': 9.38932684678169e-06, 'learning_rate/action_model': 9.38932684678169e-05, 'epoch': 0.5}) +[2026-06-17 18:41:18,154][starVLA.training.train_starvla][INFO] - Step 1227, Loss: {'train/loss': 0.6673049330711365, 'train/grad_norm_pre_clip': 8.562650680541992, 'timing/data': 0.0005419878289103508, 'timing/model': 1.0015005022287369, 'learning_rate/qwen_vl_interface': 9.388265602820901e-06, 'learning_rate/action_model': 9.3882656028209e-05, 'epoch': 0.51}) +[2026-06-17 18:41:22,186][starVLA.training.train_starvla][INFO] - Step 1228, Loss: {'train/loss': 0.43726128339767456, 'train/grad_norm_pre_clip': 8.093883514404297, 'timing/data': 0.0007948670536279678, 'timing/model': 0.9650220656767488, 'learning_rate/qwen_vl_interface': 9.387203500993199e-06, 'learning_rate/action_model': 9.387203500993199e-05, 'epoch': 0.51}) +[2026-06-17 18:41:26,091][starVLA.training.train_starvla][INFO] - Step 1229, Loss: {'train/loss': 0.6689238548278809, 'train/grad_norm_pre_clip': 9.883613586425781, 'timing/data': 0.0014807218685746193, 'timing/model': 0.9952549552544951, 'learning_rate/qwen_vl_interface': 9.38614054151876e-06, 'learning_rate/action_model': 9.38614054151876e-05, 'epoch': 0.51}) +[2026-06-17 18:41:30,744][starVLA.training.train_starvla][INFO] - Step 1230, Loss: {'train/loss': 1.0324169397354126, 'train/grad_norm_pre_clip': 67.16178894042969, 'timing/data': 0.0008419454097747803, 'timing/model': 1.6423057969659567, 'learning_rate/qwen_vl_interface': 9.385076724617937e-06, 'learning_rate/action_model': 9.385076724617935e-05, 'epoch': 0.51}) +[2026-06-17 18:41:34,757][starVLA.training.train_starvla][INFO] - Step 1231, Loss: {'train/loss': 0.8456703424453735, 'train/grad_norm_pre_clip': 11.695184707641602, 'timing/data': 0.0009922804310917854, 'timing/model': 0.960933524183929, 'learning_rate/qwen_vl_interface': 9.384012050511258e-06, 'learning_rate/action_model': 9.384012050511258e-05, 'epoch': 0.51}) +[2026-06-17 18:41:38,822][starVLA.training.train_starvla][INFO] - Step 1232, Loss: {'train/loss': 0.5475855469703674, 'train/grad_norm_pre_clip': 7.70259952545166, 'timing/data': 0.0008426597341895103, 'timing/model': 1.2190145133063197, 'learning_rate/qwen_vl_interface': 9.382946519419433e-06, 'learning_rate/action_model': 9.382946519419434e-05, 'epoch': 0.51}) +[2026-06-17 18:41:42,896][starVLA.training.train_starvla][INFO] - Step 1233, Loss: {'train/loss': 0.9046988487243652, 'train/grad_norm_pre_clip': 13.204141616821289, 'timing/data': 0.0009582741186022758, 'timing/model': 1.215219784528017, 'learning_rate/qwen_vl_interface': 9.381880131563347e-06, 'learning_rate/action_model': 9.381880131563346e-05, 'epoch': 0.51}) +[2026-06-17 18:41:46,621][starVLA.training.train_starvla][INFO] - Step 1234, Loss: {'train/loss': 0.40927401185035706, 'train/grad_norm_pre_clip': 7.2962493896484375, 'timing/data': 0.0006639938801527023, 'timing/model': 0.9456275589764118, 'learning_rate/qwen_vl_interface': 9.380812887164065e-06, 'learning_rate/action_model': 9.380812887164065e-05, 'epoch': 0.51}) +[2026-06-17 18:41:50,564][starVLA.training.train_starvla][INFO] - Step 1235, Loss: {'train/loss': 0.7423568964004517, 'train/grad_norm_pre_clip': 11.77513313293457, 'timing/data': 0.0009640911594033241, 'timing/model': 0.9856501203030348, 'learning_rate/qwen_vl_interface': 9.379744786442824e-06, 'learning_rate/action_model': 9.379744786442824e-05, 'epoch': 0.51}) +[2026-06-17 18:41:54,586][starVLA.training.train_starvla][INFO] - Step 1236, Loss: {'train/loss': 0.3783535063266754, 'train/grad_norm_pre_clip': 7.536160945892334, 'timing/data': 0.0008730972185730934, 'timing/model': 1.0060759568586946, 'learning_rate/qwen_vl_interface': 9.378675829621047e-06, 'learning_rate/action_model': 9.378675829621046e-05, 'epoch': 0.51}) +[2026-06-17 18:41:58,925][starVLA.training.train_starvla][INFO] - Step 1237, Loss: {'train/loss': 0.9424550533294678, 'train/grad_norm_pre_clip': 11.440800666809082, 'timing/data': 0.000962182879447937, 'timing/model': 0.9584911745041609, 'learning_rate/qwen_vl_interface': 9.377606016920328e-06, 'learning_rate/action_model': 9.377606016920327e-05, 'epoch': 0.51}) +[2026-06-17 18:42:02,875][starVLA.training.train_starvla][INFO] - Step 1238, Loss: {'train/loss': 0.6674060225486755, 'train/grad_norm_pre_clip': 9.430020332336426, 'timing/data': 0.0008066082373261452, 'timing/model': 0.9518425296992064, 'learning_rate/qwen_vl_interface': 9.376535348562438e-06, 'learning_rate/action_model': 9.376535348562439e-05, 'epoch': 0.51}) +[2026-06-17 18:42:06,845][starVLA.training.train_starvla][INFO] - Step 1239, Loss: {'train/loss': 0.8234670758247375, 'train/grad_norm_pre_clip': 26.17399787902832, 'timing/data': 0.0009136321023106575, 'timing/model': 1.2144243093207479, 'learning_rate/qwen_vl_interface': 9.375463824769332e-06, 'learning_rate/action_model': 9.375463824769333e-05, 'epoch': 0.51}) +[2026-06-17 18:42:10,785][starVLA.training.train_starvla][INFO] - Step 1240, Loss: {'train/loss': 0.35217201709747314, 'train/grad_norm_pre_clip': 5.910861492156982, 'timing/data': 0.0010118447244167328, 'timing/model': 1.1967890774831176, 'learning_rate/qwen_vl_interface': 9.374391445763135e-06, 'learning_rate/action_model': 9.374391445763135e-05, 'epoch': 0.51}) +[2026-06-17 18:42:14,566][starVLA.training.train_starvla][INFO] - Step 1241, Loss: {'train/loss': 0.7570763230323792, 'train/grad_norm_pre_clip': 10.912129402160645, 'timing/data': 0.0010594343766570091, 'timing/model': 0.9567703241482377, 'learning_rate/qwen_vl_interface': 9.373318211766155e-06, 'learning_rate/action_model': 9.373318211766155e-05, 'epoch': 0.51}) +[2026-06-17 18:42:19,413][starVLA.training.train_starvla][INFO] - Step 1242, Loss: {'train/loss': 0.7305552959442139, 'train/grad_norm_pre_clip': 9.738180160522461, 'timing/data': 0.0008208770304918289, 'timing/model': 0.9648142606019974, 'learning_rate/qwen_vl_interface': 9.37224412300087e-06, 'learning_rate/action_model': 9.37224412300087e-05, 'epoch': 0.51}) +[2026-06-17 18:42:23,527][starVLA.training.train_starvla][INFO] - Step 1243, Loss: {'train/loss': 0.9901371002197266, 'train/grad_norm_pre_clip': 15.300129890441895, 'timing/data': 0.0007556704804301262, 'timing/model': 0.9590435363352299, 'learning_rate/qwen_vl_interface': 9.371169179689945e-06, 'learning_rate/action_model': 9.371169179689945e-05, 'epoch': 0.51}) +[2026-06-17 18:42:27,545][starVLA.training.train_starvla][INFO] - Step 1244, Loss: {'train/loss': 0.8530500531196594, 'train/grad_norm_pre_clip': 8.943320274353027, 'timing/data': 0.0007961494848132133, 'timing/model': 0.9446883453056216, 'learning_rate/qwen_vl_interface': 9.370093382056213e-06, 'learning_rate/action_model': 9.370093382056214e-05, 'epoch': 0.51}) +[2026-06-17 18:42:31,481][starVLA.training.train_starvla][INFO] - Step 1245, Loss: {'train/loss': 0.9646260142326355, 'train/grad_norm_pre_clip': 9.66988468170166, 'timing/data': 0.0005160234868526459, 'timing/model': 0.9448959622532129, 'learning_rate/qwen_vl_interface': 9.369016730322692e-06, 'learning_rate/action_model': 9.36901673032269e-05, 'epoch': 0.51}) +[2026-06-17 18:42:35,383][starVLA.training.train_starvla][INFO] - Step 1246, Loss: {'train/loss': 0.9388888478279114, 'train/grad_norm_pre_clip': 10.614293098449707, 'timing/data': 0.0009371526539325714, 'timing/model': 1.1697892034426332, 'learning_rate/qwen_vl_interface': 9.367939224712566e-06, 'learning_rate/action_model': 9.367939224712566e-05, 'epoch': 0.51}) +[2026-06-17 18:42:39,616][starVLA.training.train_starvla][INFO] - Step 1247, Loss: {'train/loss': 0.702627420425415, 'train/grad_norm_pre_clip': 9.774972915649414, 'timing/data': 0.0016381219029426575, 'timing/model': 1.1094364486634731, 'learning_rate/qwen_vl_interface': 9.366860865449213e-06, 'learning_rate/action_model': 9.366860865449212e-05, 'epoch': 0.51}) +[2026-06-17 18:42:43,302][starVLA.training.train_starvla][INFO] - Step 1248, Loss: {'train/loss': 0.584315836429596, 'train/grad_norm_pre_clip': 10.142582893371582, 'timing/data': 0.0010009435936808586, 'timing/model': 0.9563571568578482, 'learning_rate/qwen_vl_interface': 9.365781652756168e-06, 'learning_rate/action_model': 9.365781652756168e-05, 'epoch': 0.51}) +[2026-06-17 18:42:47,254][starVLA.training.train_starvla][INFO] - Step 1249, Loss: {'train/loss': 1.0401116609573364, 'train/grad_norm_pre_clip': 10.022354125976562, 'timing/data': 0.0007886309176683426, 'timing/model': 0.9579304317012429, 'learning_rate/qwen_vl_interface': 9.36470158685716e-06, 'learning_rate/action_model': 9.36470158685716e-05, 'epoch': 0.51}) +[2026-06-17 18:42:51,254][starVLA.training.train_starvla][INFO] - Step 1250, Loss: {'train/loss': 0.8725125789642334, 'train/grad_norm_pre_clip': 15.174687385559082, 'timing/data': 0.0008767284452915192, 'timing/model': 0.9388727257028222, 'learning_rate/qwen_vl_interface': 9.363620667976085e-06, 'learning_rate/action_model': 9.363620667976085e-05, 'epoch': 0.51}) +[2026-06-17 18:42:55,296][starVLA.training.train_starvla][INFO] - Step 1251, Loss: {'train/loss': 0.4609428942203522, 'train/grad_norm_pre_clip': 7.779464244842529, 'timing/data': 0.000571858137845993, 'timing/model': 1.0107041411101818, 'learning_rate/qwen_vl_interface': 9.362538896337019e-06, 'learning_rate/action_model': 9.362538896337019e-05, 'epoch': 0.51}) +[2026-06-17 18:42:59,251][starVLA.training.train_starvla][INFO] - Step 1252, Loss: {'train/loss': 0.6769753694534302, 'train/grad_norm_pre_clip': 11.99644660949707, 'timing/data': 0.0005008522421121597, 'timing/model': 1.176877949386835, 'learning_rate/qwen_vl_interface': 9.361456272164215e-06, 'learning_rate/action_model': 9.361456272164214e-05, 'epoch': 0.52}) +[2026-06-17 18:43:03,427][starVLA.training.train_starvla][INFO] - Step 1253, Loss: {'train/loss': 0.2954445779323578, 'train/grad_norm_pre_clip': 6.076766014099121, 'timing/data': 0.0007669180631637573, 'timing/model': 1.2499049240723252, 'learning_rate/qwen_vl_interface': 9.3603727956821e-06, 'learning_rate/action_model': 9.3603727956821e-05, 'epoch': 0.52}) +[2026-06-17 18:43:07,222][starVLA.training.train_starvla][INFO] - Step 1254, Loss: {'train/loss': 0.8585386276245117, 'train/grad_norm_pre_clip': 10.673946380615234, 'timing/data': 0.000805993564426899, 'timing/model': 1.0164038613438606, 'learning_rate/qwen_vl_interface': 9.359288467115283e-06, 'learning_rate/action_model': 9.359288467115283e-05, 'epoch': 0.52}) +[2026-06-17 18:43:11,190][starVLA.training.train_starvla][INFO] - Step 1255, Loss: {'train/loss': 0.8204737901687622, 'train/grad_norm_pre_clip': 19.756561279296875, 'timing/data': 0.0007848059758543968, 'timing/model': 0.9447455117478967, 'learning_rate/qwen_vl_interface': 9.358203286688544e-06, 'learning_rate/action_model': 9.358203286688544e-05, 'epoch': 0.52}) +[2026-06-17 18:43:15,305][starVLA.training.train_starvla][INFO] - Step 1256, Loss: {'train/loss': 0.8152007460594177, 'train/grad_norm_pre_clip': 11.543265342712402, 'timing/data': 0.0007543032988905907, 'timing/model': 1.024341999553144, 'learning_rate/qwen_vl_interface': 9.357117254626844e-06, 'learning_rate/action_model': 9.357117254626844e-05, 'epoch': 0.52}) +[2026-06-17 18:43:22,553][starVLA.training.train_starvla][INFO] - Step 1257, Loss: {'train/loss': 0.7109907865524292, 'train/grad_norm_pre_clip': 9.30260944366455, 'timing/data': 0.0012134788557887077, 'timing/model': 0.9813490593805909, 'learning_rate/qwen_vl_interface': 9.356030371155316e-06, 'learning_rate/action_model': 9.356030371155316e-05, 'epoch': 0.52}) +[2026-06-17 18:43:26,786][starVLA.training.train_starvla][INFO] - Step 1258, Loss: {'train/loss': 0.6358022689819336, 'train/grad_norm_pre_clip': 9.958475112915039, 'timing/data': 0.000986647792160511, 'timing/model': 0.9603014718741179, 'learning_rate/qwen_vl_interface': 9.354942636499275e-06, 'learning_rate/action_model': 9.354942636499276e-05, 'epoch': 0.52}) +[2026-06-17 18:43:30,769][starVLA.training.train_starvla][INFO] - Step 1259, Loss: {'train/loss': 0.924364447593689, 'train/grad_norm_pre_clip': 9.90847396850586, 'timing/data': 0.0009584417566657066, 'timing/model': 1.1812324272468686, 'learning_rate/qwen_vl_interface': 9.35385405088421e-06, 'learning_rate/action_model': 9.353854050884208e-05, 'epoch': 0.52}) +[2026-06-17 18:43:34,514][starVLA.training.train_starvla][INFO] - Step 1260, Loss: {'train/loss': 0.7731323838233948, 'train/grad_norm_pre_clip': 15.91179084777832, 'timing/data': 0.001115943305194378, 'timing/model': 0.9612587364390492, 'learning_rate/qwen_vl_interface': 9.352764614535782e-06, 'learning_rate/action_model': 9.352764614535781e-05, 'epoch': 0.52}) +[2026-06-17 18:43:38,744][starVLA.training.train_starvla][INFO] - Step 1261, Loss: {'train/loss': 0.5810036063194275, 'train/grad_norm_pre_clip': 9.221885681152344, 'timing/data': 0.0013221511617302895, 'timing/model': 0.9402196770533919, 'learning_rate/qwen_vl_interface': 9.351674327679837e-06, 'learning_rate/action_model': 9.351674327679836e-05, 'epoch': 0.52}) +[2026-06-17 18:43:43,126][starVLA.training.train_starvla][INFO] - Step 1262, Loss: {'train/loss': 0.5913545489311218, 'train/grad_norm_pre_clip': 8.008790016174316, 'timing/data': 0.002105075865983963, 'timing/model': 1.021109377965331, 'learning_rate/qwen_vl_interface': 9.350583190542387e-06, 'learning_rate/action_model': 9.350583190542387e-05, 'epoch': 0.52}) +[2026-06-17 18:43:47,657][starVLA.training.train_starvla][INFO] - Step 1263, Loss: {'train/loss': 0.46476951241493225, 'train/grad_norm_pre_clip': 6.912255764007568, 'timing/data': 0.0005759177729487419, 'timing/model': 0.9906780915334821, 'learning_rate/qwen_vl_interface': 9.34949120334963e-06, 'learning_rate/action_model': 9.34949120334963e-05, 'epoch': 0.52}) +[2026-06-17 18:43:51,711][starVLA.training.train_starvla][INFO] - Step 1264, Loss: {'train/loss': 0.4729855954647064, 'train/grad_norm_pre_clip': 8.136056900024414, 'timing/data': 0.0012560179457068443, 'timing/model': 0.9752346687018871, 'learning_rate/qwen_vl_interface': 9.348398366327938e-06, 'learning_rate/action_model': 9.348398366327937e-05, 'epoch': 0.52}) +[2026-06-17 18:43:55,630][starVLA.training.train_starvla][INFO] - Step 1265, Loss: {'train/loss': 0.8377394080162048, 'train/grad_norm_pre_clip': 26.2048282623291, 'timing/data': 0.0006449883803725243, 'timing/model': 1.1931642480194569, 'learning_rate/qwen_vl_interface': 9.347304679703852e-06, 'learning_rate/action_model': 9.347304679703851e-05, 'epoch': 0.52}) +[2026-06-17 18:43:59,358][starVLA.training.train_starvla][INFO] - Step 1266, Loss: {'train/loss': 0.9895896911621094, 'train/grad_norm_pre_clip': 21.190046310424805, 'timing/data': 0.001002708449959755, 'timing/model': 0.952686870470643, 'learning_rate/qwen_vl_interface': 9.346210143704099e-06, 'learning_rate/action_model': 9.346210143704098e-05, 'epoch': 0.52}) +[2026-06-17 18:44:03,412][starVLA.training.train_starvla][INFO] - Step 1267, Loss: {'train/loss': 0.6471399068832397, 'train/grad_norm_pre_clip': 7.987306118011475, 'timing/data': 0.0009415047243237495, 'timing/model': 1.0495827766135335, 'learning_rate/qwen_vl_interface': 9.345114758555573e-06, 'learning_rate/action_model': 9.345114758555573e-05, 'epoch': 0.52}) +[2026-06-17 18:44:07,283][starVLA.training.train_starvla][INFO] - Step 1268, Loss: {'train/loss': 0.9510213732719421, 'train/grad_norm_pre_clip': 9.928983688354492, 'timing/data': 0.0007983315736055374, 'timing/model': 0.9453626750037074, 'learning_rate/qwen_vl_interface': 9.344018524485352e-06, 'learning_rate/action_model': 9.344018524485351e-05, 'epoch': 0.52}) +[2026-06-17 18:44:11,194][starVLA.training.train_starvla][INFO] - Step 1269, Loss: {'train/loss': 0.641163170337677, 'train/grad_norm_pre_clip': 10.742786407470703, 'timing/data': 0.0006373496726155281, 'timing/model': 0.9488162789493799, 'learning_rate/qwen_vl_interface': 9.342921441720686e-06, 'learning_rate/action_model': 9.342921441720685e-05, 'epoch': 0.52}) +[2026-06-17 18:44:15,123][starVLA.training.train_starvla][INFO] - Step 1270, Loss: {'train/loss': 0.5877872109413147, 'train/grad_norm_pre_clip': 9.657829284667969, 'timing/data': 0.0005119713023304939, 'timing/model': 0.9586849585175514, 'learning_rate/qwen_vl_interface': 9.341823510489e-06, 'learning_rate/action_model': 9.341823510489e-05, 'epoch': 0.52}) +[2026-06-17 18:44:18,964][starVLA.training.train_starvla][INFO] - Step 1271, Loss: {'train/loss': 0.4095594584941864, 'train/grad_norm_pre_clip': 6.623251438140869, 'timing/data': 0.0011020274832844734, 'timing/model': 1.1177758825942874, 'learning_rate/qwen_vl_interface': 9.340724731017898e-06, 'learning_rate/action_model': 9.340724731017898e-05, 'epoch': 0.52}) +[2026-06-17 18:44:22,830][starVLA.training.train_starvla][INFO] - Step 1272, Loss: {'train/loss': 0.6959367990493774, 'train/grad_norm_pre_clip': 10.077856063842773, 'timing/data': 0.0008675083518028259, 'timing/model': 1.1308998074382544, 'learning_rate/qwen_vl_interface': 9.339625103535158e-06, 'learning_rate/action_model': 9.339625103535156e-05, 'epoch': 0.52}) +[2026-06-17 18:44:26,832][starVLA.training.train_starvla][INFO] - Step 1273, Loss: {'train/loss': 0.6994051337242126, 'train/grad_norm_pre_clip': 9.959912300109863, 'timing/data': 0.0005224402993917465, 'timing/model': 0.9493623031303287, 'learning_rate/qwen_vl_interface': 9.338524628268734e-06, 'learning_rate/action_model': 9.338524628268733e-05, 'epoch': 0.52}) +[2026-06-17 18:44:30,791][starVLA.training.train_starvla][INFO] - Step 1274, Loss: {'train/loss': 0.6534664630889893, 'train/grad_norm_pre_clip': 57.1141471862793, 'timing/data': 0.0014114929363131523, 'timing/model': 0.9470175364986062, 'learning_rate/qwen_vl_interface': 9.337423305446753e-06, 'learning_rate/action_model': 9.337423305446753e-05, 'epoch': 0.52}) +[2026-06-17 18:44:34,641][starVLA.training.train_starvla][INFO] - Step 1275, Loss: {'train/loss': 1.0691699981689453, 'train/grad_norm_pre_clip': 14.556839942932129, 'timing/data': 0.0011577773839235306, 'timing/model': 0.9414974907413125, 'learning_rate/qwen_vl_interface': 9.336321135297525e-06, 'learning_rate/action_model': 9.336321135297525e-05, 'epoch': 0.52}) +[2026-06-17 18:44:38,523][starVLA.training.train_starvla][INFO] - Step 1276, Loss: {'train/loss': 0.5086542963981628, 'train/grad_norm_pre_clip': 9.017563819885254, 'timing/data': 0.00051124207675457, 'timing/model': 0.9718966530635953, 'learning_rate/qwen_vl_interface': 9.335218118049526e-06, 'learning_rate/action_model': 9.335218118049526e-05, 'epoch': 0.53}) +[2026-06-17 18:44:42,333][starVLA.training.train_starvla][INFO] - Step 1277, Loss: {'train/loss': 0.5209290385246277, 'train/grad_norm_pre_clip': 11.455161094665527, 'timing/data': 0.0006707692518830299, 'timing/model': 0.9379635192453861, 'learning_rate/qwen_vl_interface': 9.334114253931417e-06, 'learning_rate/action_model': 9.334114253931416e-05, 'epoch': 0.53}) +[2026-06-17 18:44:46,158][starVLA.training.train_starvla][INFO] - Step 1278, Loss: {'train/loss': 0.7381072044372559, 'train/grad_norm_pre_clip': 11.817010879516602, 'timing/data': 0.0005114730447530746, 'timing/model': 1.1208688225597143, 'learning_rate/qwen_vl_interface': 9.333009543172027e-06, 'learning_rate/action_model': 9.333009543172027e-05, 'epoch': 0.53}) +[2026-06-17 18:44:49,847][starVLA.training.train_starvla][INFO] - Step 1279, Loss: {'train/loss': 0.6396561861038208, 'train/grad_norm_pre_clip': 11.262835502624512, 'timing/data': 0.0008505769073963165, 'timing/model': 0.9563285289332271, 'learning_rate/qwen_vl_interface': 9.331903986000365e-06, 'learning_rate/action_model': 9.331903986000365e-05, 'epoch': 0.53}) +[2026-06-17 18:44:53,767][starVLA.training.train_starvla][INFO] - Step 1280, Loss: {'train/loss': 0.42766451835632324, 'train/grad_norm_pre_clip': 8.297887802124023, 'timing/data': 0.0006816862151026726, 'timing/model': 0.9391193557530642, 'learning_rate/qwen_vl_interface': 9.330797582645615e-06, 'learning_rate/action_model': 9.330797582645615e-05, 'epoch': 0.53}) +[2026-06-17 18:44:57,646][starVLA.training.train_starvla][INFO] - Step 1281, Loss: {'train/loss': 0.6978847980499268, 'train/grad_norm_pre_clip': 12.189810752868652, 'timing/data': 0.0007272176444530487, 'timing/model': 0.9499702202156186, 'learning_rate/qwen_vl_interface': 9.329690333337137e-06, 'learning_rate/action_model': 9.329690333337137e-05, 'epoch': 0.53}) +[2026-06-17 18:45:01,554][starVLA.training.train_starvla][INFO] - Step 1282, Loss: {'train/loss': 0.46082669496536255, 'train/grad_norm_pre_clip': 10.119863510131836, 'timing/data': 0.0005777683109045029, 'timing/model': 0.9365546116605401, 'learning_rate/qwen_vl_interface': 9.328582238304461e-06, 'learning_rate/action_model': 9.328582238304461e-05, 'epoch': 0.53}) +[2026-06-17 18:45:05,539][starVLA.training.train_starvla][INFO] - Step 1283, Loss: {'train/loss': 0.7132018804550171, 'train/grad_norm_pre_clip': 8.75837516784668, 'timing/data': 0.0005444921553134918, 'timing/model': 0.9791535129770637, 'learning_rate/qwen_vl_interface': 9.327473297777297e-06, 'learning_rate/action_model': 9.327473297777297e-05, 'epoch': 0.53}) +[2026-06-17 18:45:09,484][starVLA.training.train_starvla][INFO] - Step 1284, Loss: {'train/loss': 0.6975109577178955, 'train/grad_norm_pre_clip': 16.532875061035156, 'timing/data': 0.0005197497084736824, 'timing/model': 0.943276391364634, 'learning_rate/qwen_vl_interface': 9.326363511985533e-06, 'learning_rate/action_model': 9.326363511985533e-05, 'epoch': 0.53}) +[2026-06-17 18:45:13,416][starVLA.training.train_starvla][INFO] - Step 1285, Loss: {'train/loss': 0.7378334403038025, 'train/grad_norm_pre_clip': 9.313519477844238, 'timing/data': 0.0012022154405713081, 'timing/model': 1.167387655004859, 'learning_rate/qwen_vl_interface': 9.325252881159226e-06, 'learning_rate/action_model': 9.325252881159226e-05, 'epoch': 0.53}) +[2026-06-17 18:45:17,115][starVLA.training.train_starvla][INFO] - Step 1286, Loss: {'train/loss': 0.913828432559967, 'train/grad_norm_pre_clip': 12.77946662902832, 'timing/data': 0.0005985843017697334, 'timing/model': 0.9382987460121512, 'learning_rate/qwen_vl_interface': 9.324141405528612e-06, 'learning_rate/action_model': 9.324141405528612e-05, 'epoch': 0.53}) +[2026-06-17 18:45:20,987][starVLA.training.train_starvla][INFO] - Step 1287, Loss: {'train/loss': 0.6533483862876892, 'train/grad_norm_pre_clip': 10.246708869934082, 'timing/data': 0.0016791243106126785, 'timing/model': 0.9411349901929498, 'learning_rate/qwen_vl_interface': 9.3230290853241e-06, 'learning_rate/action_model': 9.3230290853241e-05, 'epoch': 0.53}) +[2026-06-17 18:45:24,956][starVLA.training.train_starvla][INFO] - Step 1288, Loss: {'train/loss': 0.8947355151176453, 'train/grad_norm_pre_clip': 12.827756881713867, 'timing/data': 0.0008037229999899864, 'timing/model': 0.9534699833020568, 'learning_rate/qwen_vl_interface': 9.321915920776278e-06, 'learning_rate/action_model': 9.321915920776278e-05, 'epoch': 0.53}) +[2026-06-17 18:45:28,942][starVLA.training.train_starvla][INFO] - Step 1289, Loss: {'train/loss': 0.36064472794532776, 'train/grad_norm_pre_clip': 7.343925476074219, 'timing/data': 0.0008856570348143578, 'timing/model': 0.9393947459757328, 'learning_rate/qwen_vl_interface': 9.320801912115901e-06, 'learning_rate/action_model': 9.320801912115901e-05, 'epoch': 0.53}) +[2026-06-17 18:45:32,976][starVLA.training.train_starvla][INFO] - Step 1290, Loss: {'train/loss': 0.8237794637680054, 'train/grad_norm_pre_clip': 12.547372817993164, 'timing/data': 0.000545538030564785, 'timing/model': 1.0040262788534164, 'learning_rate/qwen_vl_interface': 9.319687059573909e-06, 'learning_rate/action_model': 9.319687059573909e-05, 'epoch': 0.53}) +[2026-06-17 18:45:37,059][starVLA.training.train_starvla][INFO] - Step 1291, Loss: {'train/loss': 0.4237012267112732, 'train/grad_norm_pre_clip': 7.428271770477295, 'timing/data': 0.0007916651666164398, 'timing/model': 1.213210585527122, 'learning_rate/qwen_vl_interface': 9.318571363381408e-06, 'learning_rate/action_model': 9.318571363381408e-05, 'epoch': 0.53}) +[2026-06-17 18:45:40,739][starVLA.training.train_starvla][INFO] - Step 1292, Loss: {'train/loss': 0.5034367442131042, 'train/grad_norm_pre_clip': 10.80797004699707, 'timing/data': 0.0010921219363808632, 'timing/model': 0.9441602313891053, 'learning_rate/qwen_vl_interface': 9.317454823769686e-06, 'learning_rate/action_model': 9.317454823769687e-05, 'epoch': 0.53}) +[2026-06-17 18:45:44,820][starVLA.training.train_starvla][INFO] - Step 1293, Loss: {'train/loss': 0.584122896194458, 'train/grad_norm_pre_clip': 10.764253616333008, 'timing/data': 0.0006551714614033699, 'timing/model': 1.0431756190955639, 'learning_rate/qwen_vl_interface': 9.316337440970206e-06, 'learning_rate/action_model': 9.316337440970206e-05, 'epoch': 0.53}) +[2026-06-17 18:45:49,389][starVLA.training.train_starvla][INFO] - Step 1294, Loss: {'train/loss': 0.8758516907691956, 'train/grad_norm_pre_clip': 9.911972045898438, 'timing/data': 0.0005227178335189819, 'timing/model': 0.9541854942217469, 'learning_rate/qwen_vl_interface': 9.315219215214596e-06, 'learning_rate/action_model': 9.315219215214595e-05, 'epoch': 0.53}) +[2026-06-17 18:45:54,677][starVLA.training.train_starvla][INFO] - Step 1295, Loss: {'train/loss': 0.6538746953010559, 'train/grad_norm_pre_clip': 10.969521522521973, 'timing/data': 0.0006629005074501038, 'timing/model': 1.0209640730172396, 'learning_rate/qwen_vl_interface': 9.314100146734668e-06, 'learning_rate/action_model': 9.314100146734668e-05, 'epoch': 0.53}) +[2026-06-17 18:45:58,628][starVLA.training.train_starvla][INFO] - Step 1296, Loss: {'train/loss': 0.5083378553390503, 'train/grad_norm_pre_clip': 6.744411945343018, 'timing/data': 0.0008435044437646866, 'timing/model': 0.9604656109586358, 'learning_rate/qwen_vl_interface': 9.312980235762408e-06, 'learning_rate/action_model': 9.312980235762408e-05, 'epoch': 0.53}) +[2026-06-17 18:46:02,721][starVLA.training.train_starvla][INFO] - Step 1297, Loss: {'train/loss': 0.6115683913230896, 'train/grad_norm_pre_clip': 8.12295150756836, 'timing/data': 0.0005959458649158478, 'timing/model': 1.2457766300067306, 'learning_rate/qwen_vl_interface': 9.311859482529971e-06, 'learning_rate/action_model': 9.31185948252997e-05, 'epoch': 0.53}) +[2026-06-17 18:46:06,840][starVLA.training.train_starvla][INFO] - Step 1298, Loss: {'train/loss': 0.41318193078041077, 'train/grad_norm_pre_clip': 6.033342361450195, 'timing/data': 0.0007975604385137558, 'timing/model': 1.2128273313865066, 'learning_rate/qwen_vl_interface': 9.310737887269694e-06, 'learning_rate/action_model': 9.310737887269694e-05, 'epoch': 0.53}) +[2026-06-17 18:46:11,146][starVLA.training.train_starvla][INFO] - Step 1299, Loss: {'train/loss': 0.6692973971366882, 'train/grad_norm_pre_clip': 16.837324142456055, 'timing/data': 0.0005654152482748032, 'timing/model': 1.0110475299879909, 'learning_rate/qwen_vl_interface': 9.309615450214084e-06, 'learning_rate/action_model': 9.309615450214084e-05, 'epoch': 0.53}) +[2026-06-17 18:48:34,730][starVLA.training.train_starvla][INFO] - Step 1300, Loss: {'train/loss': 0.5743420124053955, 'train/grad_norm_pre_clip': 9.312026023864746, 'eval/loss': 0.7338804006576538, 'eval/latency_6/loss': 0.7338804006576538, 'eval/demon_fire_precision': 0.9237209302325582, 'eval/demon_fire_recall': 0.9628312863606981, 'eval/demon_fire_f1': 0.94287070739041, 'eval/demon_fire_rate_model': 0.8973288814691152, 'eval/demon_fire_rate_teacher': 0.860879243183083, 'eval/demon_left_precision': 0.774869109947644, 'eval/demon_left_recall': 0.519298245614035, 'eval/demon_left_f1': 0.6218487394957983, 'eval/demon_left_rate_model': 0.10628825820812465, 'eval/demon_left_rate_teacher': 0.15859766277128548, 'eval/demon_right_precision': 0.9458333333333333, 'eval/demon_right_recall': 0.3486943164362519, 'eval/demon_right_f1': 0.5095398428731762, 'eval/demon_right_rate_model': 0.0667779632721202, 'eval/demon_right_rate_teacher': 0.18113522537562604, 'eval/demon_move_f1': 0.5656942911844873, 'eval/demon_cc_f1': 0.7857093082038714, 'eval/latency_6/demon_fire_precision': 0.9237209302325582, 'eval/latency_6/demon_fire_recall': 0.9628312863606981, 'eval/latency_6/demon_fire_f1': 0.94287070739041, 'eval/latency_6/demon_fire_rate_model': 0.8973288814691152, 'eval/latency_6/demon_fire_rate_teacher': 0.860879243183083, 'eval/latency_6/demon_left_precision': 0.774869109947644, 'eval/latency_6/demon_left_recall': 0.519298245614035, 'eval/latency_6/demon_left_f1': 0.6218487394957983, 'eval/latency_6/demon_left_rate_model': 0.10628825820812465, 'eval/latency_6/demon_left_rate_teacher': 0.15859766277128548, 'eval/latency_6/demon_right_precision': 0.9458333333333333, 'eval/latency_6/demon_right_recall': 0.3486943164362519, 'eval/latency_6/demon_right_f1': 0.5095398428731762, 'eval/latency_6/demon_right_rate_model': 0.0667779632721202, 'eval/latency_6/demon_right_rate_teacher': 0.18113522537562604, 'eval/latency_6/demon_move_f1': 0.5656942911844873, 'eval/latency_6/demon_cc_f1': 0.7857093082038714, 'timing/data': 0.0006992043927311897, 'timing/model': 0.9507304765284061, 'learning_rate/qwen_vl_interface': 9.308492171595821e-06, 'learning_rate/action_model': 9.30849217159582e-05, 'epoch': 0.54}) +[2026-06-17 18:48:34,734][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_1300_state +[2026-06-17 18:48:34,755][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 18:48:41,514][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_1300_state/model.safetensors +[2026-06-17 18:48:57,902][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_1300_state/optimizer.bin +[2026-06-17 18:48:57,907][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_1300_state/sampler.bin +[2026-06-17 18:48:57,912][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_1300_state/sampler_1.bin +[2026-06-17 18:48:57,915][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_1300_state/random_states_0.pkl +[2026-06-17 18:48:57,921][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1300 +[2026-06-17 18:48:57,922][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 18:48:57,936][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 18:49:01,847][starVLA.training.train_starvla][INFO] - Step 1301, Loss: {'train/loss': 1.1100353002548218, 'train/grad_norm_pre_clip': 14.379768371582031, 'timing/data': 0.000562707893550396, 'timing/model': 0.9301850162446499, 'learning_rate/qwen_vl_interface': 9.307368051647765e-06, 'learning_rate/action_model': 9.307368051647765e-05, 'epoch': 0.54}) +[2026-06-17 18:49:11,588][starVLA.training.train_starvla][INFO] - Step 1302, Loss: {'train/loss': 0.2743315100669861, 'train/grad_norm_pre_clip': 5.271212577819824, 'timing/data': 0.0006503555923700333, 'timing/model': 3.553293545730412, 'learning_rate/qwen_vl_interface': 9.306243090602945e-06, 'learning_rate/action_model': 9.306243090602946e-05, 'epoch': 0.54}) +[2026-06-17 18:49:16,809][starVLA.training.train_starvla][INFO] - Step 1303, Loss: {'train/loss': 0.49568235874176025, 'train/grad_norm_pre_clip': 7.375885486602783, 'timing/data': 0.006383874453604221, 'timing/model': 2.1976705407723784, 'learning_rate/qwen_vl_interface': 9.30511728869457e-06, 'learning_rate/action_model': 9.305117288694569e-05, 'epoch': 0.54}) +[2026-06-17 18:49:20,695][starVLA.training.train_starvla][INFO] - Step 1304, Loss: {'train/loss': 0.7491968870162964, 'train/grad_norm_pre_clip': 14.313941955566406, 'timing/data': 0.0005806414410471916, 'timing/model': 0.9650707868859172, 'learning_rate/qwen_vl_interface': 9.303990646156015e-06, 'learning_rate/action_model': 9.303990646156014e-05, 'epoch': 0.54}) +[2026-06-17 18:49:24,565][starVLA.training.train_starvla][INFO] - Step 1305, Loss: {'train/loss': 0.7654174566268921, 'train/grad_norm_pre_clip': 14.209869384765625, 'timing/data': 0.0008680708706378937, 'timing/model': 0.9444650392979383, 'learning_rate/qwen_vl_interface': 9.302863163220837e-06, 'learning_rate/action_model': 9.302863163220837e-05, 'epoch': 0.54}) +[2026-06-17 18:49:28,448][starVLA.training.train_starvla][INFO] - Step 1306, Loss: {'train/loss': 0.6855078339576721, 'train/grad_norm_pre_clip': 8.659287452697754, 'timing/data': 0.0007666340097784996, 'timing/model': 0.947531065903604, 'learning_rate/qwen_vl_interface': 9.301734840122765e-06, 'learning_rate/action_model': 9.301734840122766e-05, 'epoch': 0.54}) +[2026-06-17 18:49:32,397][starVLA.training.train_starvla][INFO] - Step 1307, Loss: {'train/loss': 0.34891971945762634, 'train/grad_norm_pre_clip': 7.844758987426758, 'timing/data': 0.0010141637176275253, 'timing/model': 1.1748664574697614, 'learning_rate/qwen_vl_interface': 9.300605677095701e-06, 'learning_rate/action_model': 9.3006056770957e-05, 'epoch': 0.54}) +[2026-06-17 18:49:36,051][starVLA.training.train_starvla][INFO] - Step 1308, Loss: {'train/loss': 0.5422908663749695, 'train/grad_norm_pre_clip': 12.959667205810547, 'timing/data': 0.0005522584542632103, 'timing/model': 0.9599995240569115, 'learning_rate/qwen_vl_interface': 9.299475674373722e-06, 'learning_rate/action_model': 9.299475674373721e-05, 'epoch': 0.54}) +[2026-06-17 18:49:40,169][starVLA.training.train_starvla][INFO] - Step 1309, Loss: {'train/loss': 0.5335355401039124, 'train/grad_norm_pre_clip': 7.291814804077148, 'timing/data': 0.0009648576378822327, 'timing/model': 0.9358162032440305, 'learning_rate/qwen_vl_interface': 9.298344832191076e-06, 'learning_rate/action_model': 9.298344832191076e-05, 'epoch': 0.54}) +[2026-06-17 18:49:43,988][starVLA.training.train_starvla][INFO] - Step 1310, Loss: {'train/loss': 0.7224768400192261, 'train/grad_norm_pre_clip': 23.03415870666504, 'timing/data': 0.0005326829850673676, 'timing/model': 0.9321572137996554, 'learning_rate/qwen_vl_interface': 9.297213150782192e-06, 'learning_rate/action_model': 9.297213150782192e-05, 'epoch': 0.54}) +[2026-06-17 18:49:48,076][starVLA.training.train_starvla][INFO] - Step 1311, Loss: {'train/loss': 0.9582643508911133, 'train/grad_norm_pre_clip': 18.173372268676758, 'timing/data': 0.0005719112232327461, 'timing/model': 0.9377369992434978, 'learning_rate/qwen_vl_interface': 9.296080630381666e-06, 'learning_rate/action_model': 9.296080630381666e-05, 'epoch': 0.54}) +[2026-06-17 18:49:52,039][starVLA.training.train_starvla][INFO] - Step 1312, Loss: {'train/loss': 0.718508243560791, 'train/grad_norm_pre_clip': 9.931171417236328, 'timing/data': 0.000582314096391201, 'timing/model': 1.0591164696961641, 'learning_rate/qwen_vl_interface': 9.294947271224271e-06, 'learning_rate/action_model': 9.294947271224272e-05, 'epoch': 0.54}) +[2026-06-17 18:49:55,930][starVLA.training.train_starvla][INFO] - Step 1313, Loss: {'train/loss': 0.652643084526062, 'train/grad_norm_pre_clip': 15.330696105957031, 'timing/data': 0.0007091853767633438, 'timing/model': 1.1259611304849386, 'learning_rate/qwen_vl_interface': 9.293813073544957e-06, 'learning_rate/action_model': 9.293813073544956e-05, 'epoch': 0.54}) +[2026-06-17 18:50:00,005][starVLA.training.train_starvla][INFO] - Step 1314, Loss: {'train/loss': 0.723588764667511, 'train/grad_norm_pre_clip': 10.169720649719238, 'timing/data': 0.0009822715073823929, 'timing/model': 1.228824763558805, 'learning_rate/qwen_vl_interface': 9.292678037578838e-06, 'learning_rate/action_model': 9.292678037578837e-05, 'epoch': 0.54}) +[2026-06-17 18:50:03,792][starVLA.training.train_starvla][INFO] - Step 1315, Loss: {'train/loss': 0.7686702609062195, 'train/grad_norm_pre_clip': 15.535295486450195, 'timing/data': 0.0009284866973757744, 'timing/model': 0.9416850637644529, 'learning_rate/qwen_vl_interface': 9.291542163561214e-06, 'learning_rate/action_model': 9.291542163561213e-05, 'epoch': 0.54}) +[2026-06-17 18:50:10,290][starVLA.training.train_starvla][INFO] - Step 1316, Loss: {'train/loss': 0.8257102370262146, 'train/grad_norm_pre_clip': 14.45589828491211, 'timing/data': 0.0007741041481494904, 'timing/model': 1.941433323547244, 'learning_rate/qwen_vl_interface': 9.29040545172755e-06, 'learning_rate/action_model': 9.290405451727549e-05, 'epoch': 0.54}) +[2026-06-17 18:50:15,917][starVLA.training.train_starvla][INFO] - Step 1317, Loss: {'train/loss': 0.9621297717094421, 'train/grad_norm_pre_clip': 12.76760482788086, 'timing/data': 0.0012945393100380898, 'timing/model': 0.9579731468111277, 'learning_rate/qwen_vl_interface': 9.289267902313486e-06, 'learning_rate/action_model': 9.289267902313487e-05, 'epoch': 0.54}) +[2026-06-17 18:50:20,137][starVLA.training.train_starvla][INFO] - Step 1318, Loss: {'train/loss': 0.6527222990989685, 'train/grad_norm_pre_clip': 10.824020385742188, 'timing/data': 0.0008288584649562836, 'timing/model': 0.9506649132817984, 'learning_rate/qwen_vl_interface': 9.288129515554843e-06, 'learning_rate/action_model': 9.288129515554843e-05, 'epoch': 0.54}) +[2026-06-17 18:50:23,999][starVLA.training.train_starvla][INFO] - Step 1319, Loss: {'train/loss': 0.639253556728363, 'train/grad_norm_pre_clip': 13.177485466003418, 'timing/data': 0.0006614970043301582, 'timing/model': 0.9413009425625205, 'learning_rate/qwen_vl_interface': 9.286990291687605e-06, 'learning_rate/action_model': 9.286990291687604e-05, 'epoch': 0.54}) +[2026-06-17 18:50:27,911][starVLA.training.train_starvla][INFO] - Step 1320, Loss: {'train/loss': 0.7428253889083862, 'train/grad_norm_pre_clip': 9.663869857788086, 'timing/data': 0.0007854839786887169, 'timing/model': 1.206983813084662, 'learning_rate/qwen_vl_interface': 9.285850230947934e-06, 'learning_rate/action_model': 9.285850230947934e-05, 'epoch': 0.54}) +[2026-06-17 18:50:31,616][starVLA.training.train_starvla][INFO] - Step 1321, Loss: {'train/loss': 0.6164360642433167, 'train/grad_norm_pre_clip': 10.557385444641113, 'timing/data': 0.0007101045921444893, 'timing/model': 0.9387689586728811, 'learning_rate/qwen_vl_interface': 9.28470933357217e-06, 'learning_rate/action_model': 9.28470933357217e-05, 'epoch': 0.54}) +[2026-06-17 18:50:35,565][starVLA.training.train_starvla][INFO] - Step 1322, Loss: {'train/loss': 0.6826621890068054, 'train/grad_norm_pre_clip': 9.997342109680176, 'timing/data': 0.0005580801516771317, 'timing/model': 0.9470133977010846, 'learning_rate/qwen_vl_interface': 9.283567599796819e-06, 'learning_rate/action_model': 9.283567599796818e-05, 'epoch': 0.54}) +[2026-06-17 18:50:39,580][starVLA.training.train_starvla][INFO] - Step 1323, Loss: {'train/loss': 0.6077120900154114, 'train/grad_norm_pre_clip': 8.398734092712402, 'timing/data': 0.0005130767822265625, 'timing/model': 1.007950996980071, 'learning_rate/qwen_vl_interface': 9.282425029858561e-06, 'learning_rate/action_model': 9.282425029858561e-05, 'epoch': 0.54}) +[2026-06-17 18:50:43,536][starVLA.training.train_starvla][INFO] - Step 1324, Loss: {'train/loss': 0.7496606111526489, 'train/grad_norm_pre_clip': 7.818686485290527, 'timing/data': 0.000904901884496212, 'timing/model': 0.9654470672830939, 'learning_rate/qwen_vl_interface': 9.281281623994258e-06, 'learning_rate/action_model': 9.281281623994258e-05, 'epoch': 0.55}) +[2026-06-17 18:50:47,711][starVLA.training.train_starvla][INFO] - Step 1325, Loss: {'train/loss': 0.44274187088012695, 'train/grad_norm_pre_clip': 8.179807662963867, 'timing/data': 0.0006904909387230873, 'timing/model': 1.0564376208931208, 'learning_rate/qwen_vl_interface': 9.280137382440936e-06, 'learning_rate/action_model': 9.280137382440937e-05, 'epoch': 0.55}) +[2026-06-17 18:50:51,608][starVLA.training.train_starvla][INFO] - Step 1326, Loss: {'train/loss': 0.6408405900001526, 'train/grad_norm_pre_clip': 8.42703914642334, 'timing/data': 0.0004937015473842621, 'timing/model': 0.9572410713881254, 'learning_rate/qwen_vl_interface': 9.278992305435798e-06, 'learning_rate/action_model': 9.278992305435797e-05, 'epoch': 0.55}) +[2026-06-17 18:50:55,566][starVLA.training.train_starvla][INFO] - Step 1327, Loss: {'train/loss': 0.3823924958705902, 'train/grad_norm_pre_clip': 8.48656177520752, 'timing/data': 0.0005811518058180809, 'timing/model': 1.1864712499082088, 'learning_rate/qwen_vl_interface': 9.27784639321622e-06, 'learning_rate/action_model': 9.27784639321622e-05, 'epoch': 0.55}) +[2026-06-17 18:50:59,242][starVLA.training.train_starvla][INFO] - Step 1328, Loss: {'train/loss': 0.8193504214286804, 'train/grad_norm_pre_clip': 10.00504207611084, 'timing/data': 0.0008805692195892334, 'timing/model': 0.9493593359366059, 'learning_rate/qwen_vl_interface': 9.27669964601975e-06, 'learning_rate/action_model': 9.27669964601975e-05, 'epoch': 0.55}) +[2026-06-17 18:51:03,147][starVLA.training.train_starvla][INFO] - Step 1329, Loss: {'train/loss': 0.4806694984436035, 'train/grad_norm_pre_clip': 9.046491622924805, 'timing/data': 0.0010188845917582512, 'timing/model': 0.9545961311087012, 'learning_rate/qwen_vl_interface': 9.27555206408411e-06, 'learning_rate/action_model': 9.275552064084111e-05, 'epoch': 0.55}) +[2026-06-17 18:51:07,074][starVLA.training.train_starvla][INFO] - Step 1330, Loss: {'train/loss': 0.49548056721687317, 'train/grad_norm_pre_clip': 9.172154426574707, 'timing/data': 0.0005775699391961098, 'timing/model': 0.9515416752547026, 'learning_rate/qwen_vl_interface': 9.274403647647197e-06, 'learning_rate/action_model': 9.274403647647197e-05, 'epoch': 0.55}) +[2026-06-17 18:51:11,056][starVLA.training.train_starvla][INFO] - Step 1331, Loss: {'train/loss': 0.4339667856693268, 'train/grad_norm_pre_clip': 9.508811950683594, 'timing/data': 0.0005048606544733047, 'timing/model': 1.004609507508576, 'learning_rate/qwen_vl_interface': 9.273254396947079e-06, 'learning_rate/action_model': 9.273254396947078e-05, 'epoch': 0.55}) +[2026-06-17 18:51:15,014][starVLA.training.train_starvla][INFO] - Step 1332, Loss: {'train/loss': 0.47650331258773804, 'train/grad_norm_pre_clip': 6.147271633148193, 'timing/data': 0.0006401464343070984, 'timing/model': 1.034638972952962, 'learning_rate/qwen_vl_interface': 9.27210431222199e-06, 'learning_rate/action_model': 9.27210431222199e-05, 'epoch': 0.55}) +[2026-06-17 18:51:19,038][starVLA.training.train_starvla][INFO] - Step 1333, Loss: {'train/loss': 0.4927767813205719, 'train/grad_norm_pre_clip': 11.439373970031738, 'timing/data': 0.0005731591954827309, 'timing/model': 1.1860912069678307, 'learning_rate/qwen_vl_interface': 9.270953393710353e-06, 'learning_rate/action_model': 9.270953393710352e-05, 'epoch': 0.55}) +[2026-06-17 18:51:22,972][starVLA.training.train_starvla][INFO] - Step 1334, Loss: {'train/loss': 0.552751362323761, 'train/grad_norm_pre_clip': 9.649154663085938, 'timing/data': 0.0005355300381779671, 'timing/model': 1.116966838017106, 'learning_rate/qwen_vl_interface': 9.269801641650751e-06, 'learning_rate/action_model': 9.26980164165075e-05, 'epoch': 0.55}) +[2026-06-17 18:51:27,016][starVLA.training.train_starvla][INFO] - Step 1335, Loss: {'train/loss': 0.6356589198112488, 'train/grad_norm_pre_clip': 13.514402389526367, 'timing/data': 0.0007811766117811203, 'timing/model': 1.0366345830261707, 'learning_rate/qwen_vl_interface': 9.268649056281941e-06, 'learning_rate/action_model': 9.268649056281941e-05, 'epoch': 0.55}) +[2026-06-17 18:51:30,842][starVLA.training.train_starvla][INFO] - Step 1336, Loss: {'train/loss': 0.6334707736968994, 'train/grad_norm_pre_clip': 9.385433197021484, 'timing/data': 0.0006434256210923195, 'timing/model': 0.9459217889234424, 'learning_rate/qwen_vl_interface': 9.267495637842859e-06, 'learning_rate/action_model': 9.267495637842857e-05, 'epoch': 0.55}) +[2026-06-17 18:51:34,749][starVLA.training.train_starvla][INFO] - Step 1337, Loss: {'train/loss': 0.5207692980766296, 'train/grad_norm_pre_clip': 11.347729682922363, 'timing/data': 0.0009736772626638412, 'timing/model': 0.9511180212721229, 'learning_rate/qwen_vl_interface': 9.266341386572607e-06, 'learning_rate/action_model': 9.266341386572605e-05, 'epoch': 0.55}) +[2026-06-17 18:51:38,739][starVLA.training.train_starvla][INFO] - Step 1338, Loss: {'train/loss': 0.7642250657081604, 'train/grad_norm_pre_clip': 14.751313209533691, 'timing/data': 0.0007695872336626053, 'timing/model': 0.9713381044566631, 'learning_rate/qwen_vl_interface': 9.265186302710465e-06, 'learning_rate/action_model': 9.265186302710465e-05, 'epoch': 0.55}) +[2026-06-17 18:51:42,666][starVLA.training.train_starvla][INFO] - Step 1339, Loss: {'train/loss': 0.6576305627822876, 'train/grad_norm_pre_clip': 9.366174697875977, 'timing/data': 0.001216731034219265, 'timing/model': 0.9472295865416527, 'learning_rate/qwen_vl_interface': 9.26403038649588e-06, 'learning_rate/action_model': 9.26403038649588e-05, 'epoch': 0.55}) +[2026-06-17 18:51:46,576][starVLA.training.train_starvla][INFO] - Step 1340, Loss: {'train/loss': 0.7195205092430115, 'train/grad_norm_pre_clip': 12.138258934020996, 'timing/data': 0.0005613779649138451, 'timing/model': 1.136029141023755, 'learning_rate/qwen_vl_interface': 9.262873638168478e-06, 'learning_rate/action_model': 9.262873638168478e-05, 'epoch': 0.55}) +[2026-06-17 18:51:50,259][starVLA.training.train_starvla][INFO] - Step 1341, Loss: {'train/loss': 0.3950548768043518, 'train/grad_norm_pre_clip': 8.044818878173828, 'timing/data': 0.0007583582773804665, 'timing/model': 0.9725800827145576, 'learning_rate/qwen_vl_interface': 9.261716057968054e-06, 'learning_rate/action_model': 9.261716057968053e-05, 'epoch': 0.55}) +[2026-06-17 18:51:54,257][starVLA.training.train_starvla][INFO] - Step 1342, Loss: {'train/loss': 0.45668554306030273, 'train/grad_norm_pre_clip': 10.87801456451416, 'timing/data': 0.0009432705119252205, 'timing/model': 0.9540259223431349, 'learning_rate/qwen_vl_interface': 9.260557646134572e-06, 'learning_rate/action_model': 9.260557646134572e-05, 'epoch': 0.55}) +[2026-06-17 18:51:58,363][starVLA.training.train_starvla][INFO] - Step 1343, Loss: {'train/loss': 0.7010017037391663, 'train/grad_norm_pre_clip': 12.066882133483887, 'timing/data': 0.0006084395572543144, 'timing/model': 0.9616876784712076, 'learning_rate/qwen_vl_interface': 9.259398402908175e-06, 'learning_rate/action_model': 9.259398402908174e-05, 'epoch': 0.55}) +[2026-06-17 18:52:03,193][starVLA.training.train_starvla][INFO] - Step 1344, Loss: {'train/loss': 0.6948012113571167, 'train/grad_norm_pre_clip': 11.194855690002441, 'timing/data': 0.0007456885650753975, 'timing/model': 1.0081147942692041, 'learning_rate/qwen_vl_interface': 9.258238328529177e-06, 'learning_rate/action_model': 9.258238328529176e-05, 'epoch': 0.55}) +[2026-06-17 18:52:07,162][starVLA.training.train_starvla][INFO] - Step 1345, Loss: {'train/loss': 0.6421117186546326, 'train/grad_norm_pre_clip': 13.853283882141113, 'timing/data': 0.0009115627035498619, 'timing/model': 0.9502473371103406, 'learning_rate/qwen_vl_interface': 9.257077423238058e-06, 'learning_rate/action_model': 9.257077423238058e-05, 'epoch': 0.55}) +[2026-06-17 18:52:11,149][starVLA.training.train_starvla][INFO] - Step 1346, Loss: {'train/loss': 0.7796298861503601, 'train/grad_norm_pre_clip': 26.511484146118164, 'timing/data': 0.0008176695555448532, 'timing/model': 0.9698664471507072, 'learning_rate/qwen_vl_interface': 9.25591568727548e-06, 'learning_rate/action_model': 9.255915687275478e-05, 'epoch': 0.55}) +[2026-06-17 18:52:15,295][starVLA.training.train_starvla][INFO] - Step 1347, Loss: {'train/loss': 0.676750123500824, 'train/grad_norm_pre_clip': 12.656862258911133, 'timing/data': 0.0009535234421491623, 'timing/model': 1.3158109402284026, 'learning_rate/qwen_vl_interface': 9.254753120882265e-06, 'learning_rate/action_model': 9.254753120882266e-05, 'epoch': 0.55}) +[2026-06-17 18:52:19,005][starVLA.training.train_starvla][INFO] - Step 1348, Loss: {'train/loss': 0.955420732498169, 'train/grad_norm_pre_clip': 12.12655258178711, 'timing/data': 0.0011559799313545227, 'timing/model': 0.982398672029376, 'learning_rate/qwen_vl_interface': 9.25358972429942e-06, 'learning_rate/action_model': 9.253589724299421e-05, 'epoch': 0.55}) +[2026-06-17 18:52:23,039][starVLA.training.train_starvla][INFO] - Step 1349, Loss: {'train/loss': 0.494762122631073, 'train/grad_norm_pre_clip': 9.492655754089355, 'timing/data': 0.0008947532624006271, 'timing/model': 0.9919803095981479, 'learning_rate/qwen_vl_interface': 9.25242549776812e-06, 'learning_rate/action_model': 9.252425497768119e-05, 'epoch': 0.56}) +[2026-06-17 18:52:27,058][starVLA.training.train_starvla][INFO] - Step 1350, Loss: {'train/loss': 0.5840858817100525, 'train/grad_norm_pre_clip': 12.69466781616211, 'timing/data': 0.0010489430278539658, 'timing/model': 0.9606745494529605, 'learning_rate/qwen_vl_interface': 9.251260441529706e-06, 'learning_rate/action_model': 9.251260441529705e-05, 'epoch': 0.56}) +[2026-06-17 18:52:30,979][starVLA.training.train_starvla][INFO] - Step 1351, Loss: {'train/loss': 0.7704414129257202, 'train/grad_norm_pre_clip': 10.405828475952148, 'timing/data': 0.0010531209409236908, 'timing/model': 0.9489421546459198, 'learning_rate/qwen_vl_interface': 9.250094555825697e-06, 'learning_rate/action_model': 9.250094555825696e-05, 'epoch': 0.56}) +[2026-06-17 18:52:34,833][starVLA.training.train_starvla][INFO] - Step 1352, Loss: {'train/loss': 0.47991621494293213, 'train/grad_norm_pre_clip': 7.5651984214782715, 'timing/data': 0.0013882536441087723, 'timing/model': 0.9657830446958542, 'learning_rate/qwen_vl_interface': 9.248927840897782e-06, 'learning_rate/action_model': 9.248927840897781e-05, 'epoch': 0.56}) +[2026-06-17 18:52:38,794][starVLA.training.train_starvla][INFO] - Step 1353, Loss: {'train/loss': 0.8579170107841492, 'train/grad_norm_pre_clip': 11.096723556518555, 'timing/data': 0.0008459081873297691, 'timing/model': 1.2295365892350674, 'learning_rate/qwen_vl_interface': 9.247760296987822e-06, 'learning_rate/action_model': 9.247760296987821e-05, 'epoch': 0.56}) +[2026-06-17 18:52:42,699][starVLA.training.train_starvla][INFO] - Step 1354, Loss: {'train/loss': 0.4147586524486542, 'train/grad_norm_pre_clip': 18.02544593811035, 'timing/data': 0.000737232156097889, 'timing/model': 1.1978257661685348, 'learning_rate/qwen_vl_interface': 9.246591924337852e-06, 'learning_rate/action_model': 9.246591924337852e-05, 'epoch': 0.56}) +[2026-06-17 18:52:46,457][starVLA.training.train_starvla][INFO] - Step 1355, Loss: {'train/loss': 0.6163488626480103, 'train/grad_norm_pre_clip': 9.287117958068848, 'timing/data': 0.0008118636906147003, 'timing/model': 0.9564993483945727, 'learning_rate/qwen_vl_interface': 9.245422723190076e-06, 'learning_rate/action_model': 9.245422723190074e-05, 'epoch': 0.56}) +[2026-06-17 18:52:50,354][starVLA.training.train_starvla][INFO] - Step 1356, Loss: {'train/loss': 0.7110565900802612, 'train/grad_norm_pre_clip': 12.592554092407227, 'timing/data': 0.001689063385128975, 'timing/model': 0.9477425953373313, 'learning_rate/qwen_vl_interface': 9.244252693786868e-06, 'learning_rate/action_model': 9.244252693786869e-05, 'epoch': 0.56}) +[2026-06-17 18:52:54,396][starVLA.training.train_starvla][INFO] - Step 1357, Loss: {'train/loss': 0.5428317189216614, 'train/grad_norm_pre_clip': 11.278858184814453, 'timing/data': 0.0012839734554290771, 'timing/model': 1.00958002358675, 'learning_rate/qwen_vl_interface': 9.24308183637078e-06, 'learning_rate/action_model': 9.24308183637078e-05, 'epoch': 0.56}) +[2026-06-17 18:52:58,347][starVLA.training.train_starvla][INFO] - Step 1358, Loss: {'train/loss': 0.6100343465805054, 'train/grad_norm_pre_clip': 11.979134559631348, 'timing/data': 0.0009470386430621147, 'timing/model': 0.9655321994796395, 'learning_rate/qwen_vl_interface': 9.24191015118453e-06, 'learning_rate/action_model': 9.24191015118453e-05, 'epoch': 0.56}) +[2026-06-17 18:53:02,600][starVLA.training.train_starvla][INFO] - Step 1359, Loss: {'train/loss': 0.7588009238243103, 'train/grad_norm_pre_clip': 11.254940032958984, 'timing/data': 0.0006019845604896545, 'timing/model': 0.9515368593856692, 'learning_rate/qwen_vl_interface': 9.24073763847101e-06, 'learning_rate/action_model': 9.24073763847101e-05, 'epoch': 0.56}) +[2026-06-17 18:53:06,575][starVLA.training.train_starvla][INFO] - Step 1360, Loss: {'train/loss': 0.6348243355751038, 'train/grad_norm_pre_clip': 11.091196060180664, 'timing/data': 0.0009203851222991943, 'timing/model': 1.1203772500157356, 'learning_rate/qwen_vl_interface': 9.239564298473284e-06, 'learning_rate/action_model': 9.239564298473284e-05, 'epoch': 0.56}) +[2026-06-17 18:53:10,339][starVLA.training.train_starvla][INFO] - Step 1361, Loss: {'train/loss': 0.6846556663513184, 'train/grad_norm_pre_clip': 28.201013565063477, 'timing/data': 0.0009021209552884102, 'timing/model': 0.9663383141160011, 'learning_rate/qwen_vl_interface': 9.238390131434586e-06, 'learning_rate/action_model': 9.238390131434586e-05, 'epoch': 0.56}) +[2026-06-17 18:53:14,341][starVLA.training.train_starvla][INFO] - Step 1362, Loss: {'train/loss': 0.4323881268501282, 'train/grad_norm_pre_clip': 9.564325332641602, 'timing/data': 0.0012390092015266418, 'timing/model': 0.9466707641258836, 'learning_rate/qwen_vl_interface': 9.237215137598322e-06, 'learning_rate/action_model': 9.237215137598322e-05, 'epoch': 0.56}) +[2026-06-17 18:53:19,229][starVLA.training.train_starvla][INFO] - Step 1363, Loss: {'train/loss': 0.6812321543693542, 'train/grad_norm_pre_clip': 12.052144050598145, 'timing/data': 0.0011681979522109032, 'timing/model': 1.4724788954481483, 'learning_rate/qwen_vl_interface': 9.236039317208068e-06, 'learning_rate/action_model': 9.236039317208068e-05, 'epoch': 0.56}) +[2026-06-17 18:53:25,404][starVLA.training.train_starvla][INFO] - Step 1364, Loss: {'train/loss': 0.5712278485298157, 'train/grad_norm_pre_clip': 11.688102722167969, 'timing/data': 0.001152845099568367, 'timing/model': 1.0372992837801576, 'learning_rate/qwen_vl_interface': 9.234862670507577e-06, 'learning_rate/action_model': 9.234862670507576e-05, 'epoch': 0.56}) +[2026-06-17 18:53:29,899][starVLA.training.train_starvla][INFO] - Step 1365, Loss: {'train/loss': 0.9197057485580444, 'train/grad_norm_pre_clip': 13.184561729431152, 'timing/data': 0.0007418384775519371, 'timing/model': 0.999695441685617, 'learning_rate/qwen_vl_interface': 9.233685197740765e-06, 'learning_rate/action_model': 9.233685197740764e-05, 'epoch': 0.56}) +[2026-06-17 18:53:33,851][starVLA.training.train_starvla][INFO] - Step 1366, Loss: {'train/loss': 0.479707807302475, 'train/grad_norm_pre_clip': 9.194292068481445, 'timing/data': 0.0006876429542899132, 'timing/model': 0.9377077743411064, 'learning_rate/qwen_vl_interface': 9.232506899151725e-06, 'learning_rate/action_model': 9.232506899151725e-05, 'epoch': 0.56}) +[2026-06-17 18:53:37,783][starVLA.training.train_starvla][INFO] - Step 1367, Loss: {'train/loss': 0.6313525438308716, 'train/grad_norm_pre_clip': 12.43604564666748, 'timing/data': 0.0010491367429494858, 'timing/model': 1.1789537267759442, 'learning_rate/qwen_vl_interface': 9.23132777498472e-06, 'learning_rate/action_model': 9.231327774984719e-05, 'epoch': 0.56}) +[2026-06-17 18:53:41,441][starVLA.training.train_starvla][INFO] - Step 1368, Loss: {'train/loss': 0.6268091201782227, 'train/grad_norm_pre_clip': 12.283102035522461, 'timing/data': 0.0007957415655255318, 'timing/model': 0.9322226997464895, 'learning_rate/qwen_vl_interface': 9.230147825484182e-06, 'learning_rate/action_model': 9.230147825484183e-05, 'epoch': 0.56}) +[2026-06-17 18:53:45,391][starVLA.training.train_starvla][INFO] - Step 1369, Loss: {'train/loss': 0.40646857023239136, 'train/grad_norm_pre_clip': 14.837665557861328, 'timing/data': 0.0004950752481818199, 'timing/model': 0.9526378689333797, 'learning_rate/qwen_vl_interface': 9.228967050894717e-06, 'learning_rate/action_model': 9.228967050894716e-05, 'epoch': 0.56}) +[2026-06-17 18:53:49,918][starVLA.training.train_starvla][INFO] - Step 1370, Loss: {'train/loss': 0.7775076031684875, 'train/grad_norm_pre_clip': 12.80479621887207, 'timing/data': 0.0007542064413428307, 'timing/model': 1.5164676131680608, 'learning_rate/qwen_vl_interface': 9.227785451461103e-06, 'learning_rate/action_model': 9.227785451461102e-05, 'epoch': 0.56}) +[2026-06-17 18:53:54,006][starVLA.training.train_starvla][INFO] - Step 1371, Loss: {'train/loss': 0.801361620426178, 'train/grad_norm_pre_clip': 12.167914390563965, 'timing/data': 0.0005218349397182465, 'timing/model': 1.0459678713232279, 'learning_rate/qwen_vl_interface': 9.226603027428281e-06, 'learning_rate/action_model': 9.226603027428281e-05, 'epoch': 0.56}) +[2026-06-17 18:53:58,042][starVLA.training.train_starvla][INFO] - Step 1372, Loss: {'train/loss': 0.6136161088943481, 'train/grad_norm_pre_clip': 14.85537338256836, 'timing/data': 0.001138973981142044, 'timing/model': 0.9499015612527728, 'learning_rate/qwen_vl_interface': 9.225419779041376e-06, 'learning_rate/action_model': 9.225419779041376e-05, 'epoch': 0.56}) +[2026-06-17 18:54:02,062][starVLA.training.train_starvla][INFO] - Step 1373, Loss: {'train/loss': 0.5468882918357849, 'train/grad_norm_pre_clip': 8.534756660461426, 'timing/data': 0.0006255414336919785, 'timing/model': 1.126484096981585, 'learning_rate/qwen_vl_interface': 9.224235706545672e-06, 'learning_rate/action_model': 9.224235706545671e-05, 'epoch': 0.57}) +[2026-06-17 18:54:05,948][starVLA.training.train_starvla][INFO] - Step 1374, Loss: {'train/loss': 0.5783451199531555, 'train/grad_norm_pre_clip': 9.497783660888672, 'timing/data': 0.0006983857601881027, 'timing/model': 1.144744017161429, 'learning_rate/qwen_vl_interface': 9.22305081018663e-06, 'learning_rate/action_model': 9.223050810186629e-05, 'epoch': 0.57}) +[2026-06-17 18:54:09,810][starVLA.training.train_starvla][INFO] - Step 1375, Loss: {'train/loss': 0.47163915634155273, 'train/grad_norm_pre_clip': 8.093244552612305, 'timing/data': 0.0005379645153880119, 'timing/model': 1.0300512947142124, 'learning_rate/qwen_vl_interface': 9.221865090209878e-06, 'learning_rate/action_model': 9.221865090209878e-05, 'epoch': 0.57}) +[2026-06-17 18:54:13,797][starVLA.training.train_starvla][INFO] - Step 1376, Loss: {'train/loss': 0.6870338320732117, 'train/grad_norm_pre_clip': 11.200888633728027, 'timing/data': 0.0007341383025050163, 'timing/model': 0.9805984208360314, 'learning_rate/qwen_vl_interface': 9.22067854686122e-06, 'learning_rate/action_model': 9.22067854686122e-05, 'epoch': 0.57}) +[2026-06-17 18:54:17,817][starVLA.training.train_starvla][INFO] - Step 1377, Loss: {'train/loss': 0.7865096926689148, 'train/grad_norm_pre_clip': 10.448596000671387, 'timing/data': 0.0005550095811486244, 'timing/model': 0.966259534470737, 'learning_rate/qwen_vl_interface': 9.219491180386626e-06, 'learning_rate/action_model': 9.219491180386626e-05, 'epoch': 0.57}) +[2026-06-17 18:54:21,793][starVLA.training.train_starvla][INFO] - Step 1378, Loss: {'train/loss': 0.7204163670539856, 'train/grad_norm_pre_clip': 9.145698547363281, 'timing/data': 0.0005611041560769081, 'timing/model': 0.9506932403892279, 'learning_rate/qwen_vl_interface': 9.21830299103224e-06, 'learning_rate/action_model': 9.21830299103224e-05, 'epoch': 0.57}) +[2026-06-17 18:54:25,607][starVLA.training.train_starvla][INFO] - Step 1379, Loss: {'train/loss': 0.4485905170440674, 'train/grad_norm_pre_clip': 9.399752616882324, 'timing/data': 0.000924108549952507, 'timing/model': 0.9502792758867145, 'learning_rate/qwen_vl_interface': 9.217113979044373e-06, 'learning_rate/action_model': 9.217113979044373e-05, 'epoch': 0.57}) +[2026-06-17 18:54:29,457][starVLA.training.train_starvla][INFO] - Step 1380, Loss: {'train/loss': 0.5964468717575073, 'train/grad_norm_pre_clip': 76.22937774658203, 'timing/data': 0.000798448920249939, 'timing/model': 1.112200847826898, 'learning_rate/qwen_vl_interface': 9.215924144669509e-06, 'learning_rate/action_model': 9.215924144669509e-05, 'epoch': 0.57}) +[2026-06-17 18:54:34,717][starVLA.training.train_starvla][INFO] - Step 1381, Loss: {'train/loss': 0.684439480304718, 'train/grad_norm_pre_clip': 10.818500518798828, 'timing/data': 0.0005503026768565178, 'timing/model': 2.5332939038053155, 'learning_rate/qwen_vl_interface': 9.214733488154304e-06, 'learning_rate/action_model': 9.214733488154304e-05, 'epoch': 0.57}) +[2026-06-17 18:54:38,804][starVLA.training.train_starvla][INFO] - Step 1382, Loss: {'train/loss': 0.37324953079223633, 'train/grad_norm_pre_clip': 6.475648880004883, 'timing/data': 0.0005221469327807426, 'timing/model': 1.1494590183719993, 'learning_rate/qwen_vl_interface': 9.213542009745579e-06, 'learning_rate/action_model': 9.213542009745578e-05, 'epoch': 0.57}) +[2026-06-17 18:54:44,006][starVLA.training.train_starvla][INFO] - Step 1383, Loss: {'train/loss': 0.7433810830116272, 'train/grad_norm_pre_clip': 11.840712547302246, 'timing/data': 0.0012851804494857788, 'timing/model': 0.9563811989501119, 'learning_rate/qwen_vl_interface': 9.21234970969033e-06, 'learning_rate/action_model': 9.21234970969033e-05, 'epoch': 0.57}) +[2026-06-17 18:54:48,676][starVLA.training.train_starvla][INFO] - Step 1384, Loss: {'train/loss': 0.659206211566925, 'train/grad_norm_pre_clip': 9.636655807495117, 'timing/data': 0.0007723849266767502, 'timing/model': 1.0008606938645244, 'learning_rate/qwen_vl_interface': 9.211156588235724e-06, 'learning_rate/action_model': 9.211156588235725e-05, 'epoch': 0.57}) +[2026-06-17 18:54:52,673][starVLA.training.train_starvla][INFO] - Step 1385, Loss: {'train/loss': 0.3767112195491791, 'train/grad_norm_pre_clip': 8.11852741241455, 'timing/data': 0.000560482032597065, 'timing/model': 0.9623646503314376, 'learning_rate/qwen_vl_interface': 9.209962645629095e-06, 'learning_rate/action_model': 9.209962645629095e-05, 'epoch': 0.57}) +[2026-06-17 18:54:56,601][starVLA.training.train_starvla][INFO] - Step 1386, Loss: {'train/loss': 0.5246288776397705, 'train/grad_norm_pre_clip': 9.80377197265625, 'timing/data': 0.0005248589441180229, 'timing/model': 0.9484446840360761, 'learning_rate/qwen_vl_interface': 9.208767882117947e-06, 'learning_rate/action_model': 9.208767882117947e-05, 'epoch': 0.57}) +[2026-06-17 18:55:00,561][starVLA.training.train_starvla][INFO] - Step 1387, Loss: {'train/loss': 0.846699059009552, 'train/grad_norm_pre_clip': 14.901111602783203, 'timing/data': 0.0005229366943240166, 'timing/model': 1.2206430472433567, 'learning_rate/qwen_vl_interface': 9.207572297949958e-06, 'learning_rate/action_model': 9.207572297949958e-05, 'epoch': 0.57}) +[2026-06-17 18:55:04,361][starVLA.training.train_starvla][INFO] - Step 1388, Loss: {'train/loss': 0.6726235151290894, 'train/grad_norm_pre_clip': 11.7582426071167, 'timing/data': 0.0006681699305772781, 'timing/model': 0.9911859491840005, 'learning_rate/qwen_vl_interface': 9.206375893372973e-06, 'learning_rate/action_model': 9.206375893372973e-05, 'epoch': 0.57}) +[2026-06-17 18:55:08,502][starVLA.training.train_starvla][INFO] - Step 1389, Loss: {'train/loss': 0.9015906453132629, 'train/grad_norm_pre_clip': 11.917417526245117, 'timing/data': 0.0007960526272654533, 'timing/model': 1.0306071089580655, 'learning_rate/qwen_vl_interface': 9.20517866863501e-06, 'learning_rate/action_model': 9.205178668635009e-05, 'epoch': 0.57}) +[2026-06-17 18:55:12,459][starVLA.training.train_starvla][INFO] - Step 1390, Loss: {'train/loss': 0.6933709979057312, 'train/grad_norm_pre_clip': 14.609613418579102, 'timing/data': 0.0008206097409129143, 'timing/model': 0.9539862629026175, 'learning_rate/qwen_vl_interface': 9.20398062398425e-06, 'learning_rate/action_model': 9.20398062398425e-05, 'epoch': 0.57}) +[2026-06-17 18:55:16,526][starVLA.training.train_starvla][INFO] - Step 1391, Loss: {'train/loss': 0.7406238913536072, 'train/grad_norm_pre_clip': 17.102401733398438, 'timing/data': 0.00060313381254673, 'timing/model': 0.9864005241543055, 'learning_rate/qwen_vl_interface': 9.202781759669053e-06, 'learning_rate/action_model': 9.202781759669052e-05, 'epoch': 0.57}) +[2026-06-17 18:55:21,838][starVLA.training.train_starvla][INFO] - Step 1392, Loss: {'train/loss': 0.5078229308128357, 'train/grad_norm_pre_clip': 9.64645767211914, 'timing/data': 0.0012103533372282982, 'timing/model': 2.145854638889432, 'learning_rate/qwen_vl_interface': 9.201582075937943e-06, 'learning_rate/action_model': 9.201582075937944e-05, 'epoch': 0.57}) +[2026-06-17 18:55:25,943][starVLA.training.train_starvla][INFO] - Step 1393, Loss: {'train/loss': 0.9107826352119446, 'train/grad_norm_pre_clip': 12.172380447387695, 'timing/data': 0.0010080533102154732, 'timing/model': 1.2449252549558878, 'learning_rate/qwen_vl_interface': 9.200381573039619e-06, 'learning_rate/action_model': 9.200381573039619e-05, 'epoch': 0.57}) +[2026-06-17 18:55:29,947][starVLA.training.train_starvla][INFO] - Step 1394, Loss: {'train/loss': 0.5342667102813721, 'train/grad_norm_pre_clip': 9.773672103881836, 'timing/data': 0.0009514540433883667, 'timing/model': 1.0974043030291796, 'learning_rate/qwen_vl_interface': 9.199180251222942e-06, 'learning_rate/action_model': 9.199180251222942e-05, 'epoch': 0.57}) +[2026-06-17 18:55:33,805][starVLA.training.train_starvla][INFO] - Step 1395, Loss: {'train/loss': 0.5188108086585999, 'train/grad_norm_pre_clip': 8.804768562316895, 'timing/data': 0.0005779601633548737, 'timing/model': 0.9540656311437488, 'learning_rate/qwen_vl_interface': 9.197978110736951e-06, 'learning_rate/action_model': 9.19797811073695e-05, 'epoch': 0.57}) +[2026-06-17 18:55:38,551][starVLA.training.train_starvla][INFO] - Step 1396, Loss: {'train/loss': 0.8507390022277832, 'train/grad_norm_pre_clip': 12.031319618225098, 'timing/data': 0.000665898434817791, 'timing/model': 0.9970564041286707, 'learning_rate/qwen_vl_interface': 9.196775151830849e-06, 'learning_rate/action_model': 9.196775151830849e-05, 'epoch': 0.57}) +[2026-06-17 18:55:42,560][starVLA.training.train_starvla][INFO] - Step 1397, Loss: {'train/loss': 0.6694700717926025, 'train/grad_norm_pre_clip': 11.57894515991211, 'timing/data': 0.0010084239766001701, 'timing/model': 0.9549297345802188, 'learning_rate/qwen_vl_interface': 9.19557137475401e-06, 'learning_rate/action_model': 9.19557137475401e-05, 'epoch': 0.58}) +[2026-06-17 18:55:50,954][starVLA.training.train_starvla][INFO] - Step 1398, Loss: {'train/loss': 0.7337036728858948, 'train/grad_norm_pre_clip': 10.556821823120117, 'timing/data': 0.0005723424255847931, 'timing/model': 1.0870992243289948, 'learning_rate/qwen_vl_interface': 9.19436677975598e-06, 'learning_rate/action_model': 9.19436677975598e-05, 'epoch': 0.58}) +[2026-06-17 18:55:55,113][starVLA.training.train_starvla][INFO] - Step 1399, Loss: {'train/loss': 0.5758387446403503, 'train/grad_norm_pre_clip': 9.40530014038086, 'timing/data': 0.0006040139123797417, 'timing/model': 1.0358645236119628, 'learning_rate/qwen_vl_interface': 9.193161367086472e-06, 'learning_rate/action_model': 9.193161367086472e-05, 'epoch': 0.58}) +[2026-06-17 18:58:02,212][starVLA.training.train_starvla][INFO] - Step 1400, Loss: {'train/loss': 0.7416254281997681, 'train/grad_norm_pre_clip': 12.97986125946045, 'eval/loss': 0.676602303981781, 'eval/latency_6/loss': 0.676602303981781, 'eval/demon_fire_precision': 0.9457950065703022, 'eval/demon_fire_recall': 0.9305106658047835, 'eval/demon_fire_f1': 0.9380905832518736, 'eval/demon_fire_rate_model': 0.8469671675013912, 'eval/demon_fire_rate_teacher': 0.860879243183083, 'eval/demon_left_precision': 0.7446808510638298, 'eval/demon_left_recall': 0.49122807017543857, 'eval/demon_left_f1': 0.5919661733615222, 'eval/demon_left_rate_model': 0.10461880912632164, 'eval/demon_left_rate_teacher': 0.15859766277128548, 'eval/demon_right_precision': 0.8869565217391304, 'eval/demon_right_recall': 0.4700460829493088, 'eval/demon_right_f1': 0.6144578313253013, 'eval/demon_right_rate_model': 0.09599332220367279, 'eval/demon_right_rate_teacher': 0.18113522537562604, 'eval/demon_move_f1': 0.6032120023434118, 'eval/demon_cc_f1': 0.804159641043928, 'eval/latency_6/demon_fire_precision': 0.9457950065703022, 'eval/latency_6/demon_fire_recall': 0.9305106658047835, 'eval/latency_6/demon_fire_f1': 0.9380905832518736, 'eval/latency_6/demon_fire_rate_model': 0.8469671675013912, 'eval/latency_6/demon_fire_rate_teacher': 0.860879243183083, 'eval/latency_6/demon_left_precision': 0.7446808510638298, 'eval/latency_6/demon_left_recall': 0.49122807017543857, 'eval/latency_6/demon_left_f1': 0.5919661733615222, 'eval/latency_6/demon_left_rate_model': 0.10461880912632164, 'eval/latency_6/demon_left_rate_teacher': 0.15859766277128548, 'eval/latency_6/demon_right_precision': 0.8869565217391304, 'eval/latency_6/demon_right_recall': 0.4700460829493088, 'eval/latency_6/demon_right_f1': 0.6144578313253013, 'eval/latency_6/demon_right_rate_model': 0.09599332220367279, 'eval/latency_6/demon_right_rate_teacher': 0.18113522537562604, 'eval/latency_6/demon_move_f1': 0.6032120023434118, 'eval/latency_6/demon_cc_f1': 0.804159641043928, 'timing/data': 0.0007707951590418816, 'timing/model': 1.450707757845521, 'learning_rate/qwen_vl_interface': 9.191955136995371e-06, 'learning_rate/action_model': 9.191955136995371e-05, 'epoch': 0.58}) +[2026-06-17 18:58:02,216][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_1400_state +[2026-06-17 18:58:02,221][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 18:58:08,768][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_1400_state/model.safetensors +[2026-06-17 18:58:24,099][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_1400_state/optimizer.bin +[2026-06-17 18:58:24,101][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_1400_state/sampler.bin +[2026-06-17 18:58:24,101][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_1400_state/sampler_1.bin +[2026-06-17 18:58:24,103][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_1400_state/random_states_0.pkl +[2026-06-17 18:58:24,107][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1400 +[2026-06-17 18:58:24,107][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 18:58:24,118][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 18:58:28,313][starVLA.training.train_starvla][INFO] - Step 1401, Loss: {'train/loss': 0.6125980019569397, 'train/grad_norm_pre_clip': 8.541030883789062, 'timing/data': 0.0009624883532524109, 'timing/model': 0.9496846683323383, 'learning_rate/qwen_vl_interface': 9.190748089732727e-06, 'learning_rate/action_model': 9.190748089732728e-05, 'epoch': 0.58}) +[2026-06-17 18:58:32,225][starVLA.training.train_starvla][INFO] - Step 1402, Loss: {'train/loss': 0.3926691710948944, 'train/grad_norm_pre_clip': 6.3127875328063965, 'timing/data': 0.001173817552626133, 'timing/model': 1.1749093756079674, 'learning_rate/qwen_vl_interface': 9.189540225548763e-06, 'learning_rate/action_model': 9.189540225548764e-05, 'epoch': 0.58}) +[2026-06-17 18:58:36,608][starVLA.training.train_starvla][INFO] - Step 1403, Loss: {'train/loss': 0.6487212777137756, 'train/grad_norm_pre_clip': 11.67036247253418, 'timing/data': 0.0008147507905960083, 'timing/model': 0.964048245921731, 'learning_rate/qwen_vl_interface': 9.188331544693872e-06, 'learning_rate/action_model': 9.188331544693873e-05, 'epoch': 0.58}) +[2026-06-17 18:58:42,107][starVLA.training.train_starvla][INFO] - Step 1404, Loss: {'train/loss': 1.1781593561172485, 'train/grad_norm_pre_clip': 14.202770233154297, 'timing/data': 0.000498979352414608, 'timing/model': 0.9649205328896642, 'learning_rate/qwen_vl_interface': 9.187122047418614e-06, 'learning_rate/action_model': 9.187122047418614e-05, 'epoch': 0.58}) +[2026-06-17 18:58:46,279][starVLA.training.train_starvla][INFO] - Step 1405, Loss: {'train/loss': 0.734721839427948, 'train/grad_norm_pre_clip': 581.9654541015625, 'timing/data': 0.0010789278894662857, 'timing/model': 0.984905481338501, 'learning_rate/qwen_vl_interface': 9.18591173397372e-06, 'learning_rate/action_model': 9.185911733973719e-05, 'epoch': 0.58}) +[2026-06-17 18:58:50,452][starVLA.training.train_starvla][INFO] - Step 1406, Loss: {'train/loss': 1.116497278213501, 'train/grad_norm_pre_clip': 9.93350601196289, 'timing/data': 0.0007506841793656349, 'timing/model': 1.1477889874950051, 'learning_rate/qwen_vl_interface': 9.184700604610087e-06, 'learning_rate/action_model': 9.184700604610085e-05, 'epoch': 0.58}) +[2026-06-17 18:58:54,419][starVLA.training.train_starvla][INFO] - Step 1407, Loss: {'train/loss': 0.5997995734214783, 'train/grad_norm_pre_clip': 7.086209774017334, 'timing/data': 0.0012615900486707687, 'timing/model': 0.9919526102021337, 'learning_rate/qwen_vl_interface': 9.183488659578786e-06, 'learning_rate/action_model': 9.183488659578785e-05, 'epoch': 0.58}) +[2026-06-17 18:58:58,743][starVLA.training.train_starvla][INFO] - Step 1408, Loss: {'train/loss': 0.7473287582397461, 'train/grad_norm_pre_clip': 10.399345397949219, 'timing/data': 0.0010852348059415817, 'timing/model': 1.3180958982557058, 'learning_rate/qwen_vl_interface': 9.18227589913105e-06, 'learning_rate/action_model': 9.18227589913105e-05, 'epoch': 0.58}) +[2026-06-17 18:59:02,627][starVLA.training.train_starvla][INFO] - Step 1409, Loss: {'train/loss': 0.5222935676574707, 'train/grad_norm_pre_clip': 8.856278419494629, 'timing/data': 0.0012955442070960999, 'timing/model': 1.129722986370325, 'learning_rate/qwen_vl_interface': 9.181062323518293e-06, 'learning_rate/action_model': 9.181062323518292e-05, 'epoch': 0.58}) +[2026-06-17 18:59:06,365][starVLA.training.train_starvla][INFO] - Step 1410, Loss: {'train/loss': 0.7028672099113464, 'train/grad_norm_pre_clip': 9.759650230407715, 'timing/data': 0.0009071696549654007, 'timing/model': 0.9529446074739099, 'learning_rate/qwen_vl_interface': 9.179847932992084e-06, 'learning_rate/action_model': 9.179847932992083e-05, 'epoch': 0.58}) +[2026-06-17 18:59:10,432][starVLA.training.train_starvla][INFO] - Step 1411, Loss: {'train/loss': 0.6860755085945129, 'train/grad_norm_pre_clip': 13.471264839172363, 'timing/data': 0.0008899476379156113, 'timing/model': 1.0727560799568892, 'learning_rate/qwen_vl_interface': 9.17863272780417e-06, 'learning_rate/action_model': 9.178632727804168e-05, 'epoch': 0.58}) +[2026-06-17 18:59:15,188][starVLA.training.train_starvla][INFO] - Step 1412, Loss: {'train/loss': 0.5689226388931274, 'train/grad_norm_pre_clip': 7.427921772003174, 'timing/data': 0.0005261870101094246, 'timing/model': 1.1356917303055525, 'learning_rate/qwen_vl_interface': 9.177416708206463e-06, 'learning_rate/action_model': 9.177416708206463e-05, 'epoch': 0.58}) +[2026-06-17 18:59:19,825][starVLA.training.train_starvla][INFO] - Step 1413, Loss: {'train/loss': 0.7214831709861755, 'train/grad_norm_pre_clip': 15.747178077697754, 'timing/data': 0.0005377568304538727, 'timing/model': 1.1542029427364469, 'learning_rate/qwen_vl_interface': 9.176199874451049e-06, 'learning_rate/action_model': 9.176199874451048e-05, 'epoch': 0.58}) +[2026-06-17 18:59:24,332][starVLA.training.train_starvla][INFO] - Step 1414, Loss: {'train/loss': 0.736832857131958, 'train/grad_norm_pre_clip': 21.12389373779297, 'timing/data': 0.000555860809981823, 'timing/model': 1.166717610321939, 'learning_rate/qwen_vl_interface': 9.174982226790175e-06, 'learning_rate/action_model': 9.174982226790175e-05, 'epoch': 0.58}) +[2026-06-17 18:59:28,468][starVLA.training.train_starvla][INFO] - Step 1415, Loss: {'train/loss': 0.5180888772010803, 'train/grad_norm_pre_clip': 8.77558708190918, 'timing/data': 0.0009122053161263466, 'timing/model': 1.1520705996081233, 'learning_rate/qwen_vl_interface': 9.173763765476262e-06, 'learning_rate/action_model': 9.17376376547626e-05, 'epoch': 0.58}) +[2026-06-17 18:59:32,063][starVLA.training.train_starvla][INFO] - Step 1416, Loss: {'train/loss': 0.6769168972969055, 'train/grad_norm_pre_clip': 12.583738327026367, 'timing/data': 0.0005588876083493233, 'timing/model': 0.941092380322516, 'learning_rate/qwen_vl_interface': 9.172544490761897e-06, 'learning_rate/action_model': 9.172544490761897e-05, 'epoch': 0.58}) +[2026-06-17 18:59:35,967][starVLA.training.train_starvla][INFO] - Step 1417, Loss: {'train/loss': 0.5496063232421875, 'train/grad_norm_pre_clip': 12.571561813354492, 'timing/data': 0.000952485017478466, 'timing/model': 0.934855711646378, 'learning_rate/qwen_vl_interface': 9.171324402899841e-06, 'learning_rate/action_model': 9.17132440289984e-05, 'epoch': 0.58}) +[2026-06-17 18:59:39,931][starVLA.training.train_starvla][INFO] - Step 1418, Loss: {'train/loss': 0.6801601052284241, 'train/grad_norm_pre_clip': 9.02484130859375, 'timing/data': 0.0005214670673012733, 'timing/model': 0.9704758552834392, 'learning_rate/qwen_vl_interface': 9.170103502143014e-06, 'learning_rate/action_model': 9.170103502143013e-05, 'epoch': 0.58}) +[2026-06-17 18:59:44,196][starVLA.training.train_starvla][INFO] - Step 1419, Loss: {'train/loss': 0.813809335231781, 'train/grad_norm_pre_clip': 12.163823127746582, 'timing/data': 0.0009862249717116356, 'timing/model': 0.952306604012847, 'learning_rate/qwen_vl_interface': 9.168881788744513e-06, 'learning_rate/action_model': 9.168881788744513e-05, 'epoch': 0.58}) +[2026-06-17 18:59:48,071][starVLA.training.train_starvla][INFO] - Step 1420, Loss: {'train/loss': 0.7472567558288574, 'train/grad_norm_pre_clip': 11.391475677490234, 'timing/data': 0.001241518184542656, 'timing/model': 0.9292875872924924, 'learning_rate/qwen_vl_interface': 9.1676592629576e-06, 'learning_rate/action_model': 9.167659262957601e-05, 'epoch': 0.58}) +[2026-06-17 18:59:52,204][starVLA.training.train_starvla][INFO] - Step 1421, Loss: {'train/loss': 0.9378995299339294, 'train/grad_norm_pre_clip': 12.695632934570312, 'timing/data': 0.0006510177627205849, 'timing/model': 0.9943044483661652, 'learning_rate/qwen_vl_interface': 9.16643592503571e-06, 'learning_rate/action_model': 9.166435925035709e-05, 'epoch': 0.58}) +[2026-06-17 18:59:55,992][starVLA.training.train_starvla][INFO] - Step 1422, Loss: {'train/loss': 0.7461837530136108, 'train/grad_norm_pre_clip': 20.16651725769043, 'timing/data': 0.0005467673763632774, 'timing/model': 1.1116706049069762, 'learning_rate/qwen_vl_interface': 9.165211775232436e-06, 'learning_rate/action_model': 9.165211775232435e-05, 'epoch': 0.59}) +[2026-06-17 19:00:00,597][starVLA.training.train_starvla][INFO] - Step 1423, Loss: {'train/loss': 0.6759960055351257, 'train/grad_norm_pre_clip': 9.058942794799805, 'timing/data': 0.000533049926161766, 'timing/model': 1.105884545482695, 'learning_rate/qwen_vl_interface': 9.163986813801548e-06, 'learning_rate/action_model': 9.163986813801548e-05, 'epoch': 0.59}) +[2026-06-17 19:00:06,664][starVLA.training.train_starvla][INFO] - Step 1424, Loss: {'train/loss': 0.6065274477005005, 'train/grad_norm_pre_clip': 8.676748275756836, 'timing/data': 0.0005288533866405487, 'timing/model': 0.9409192521125078, 'learning_rate/qwen_vl_interface': 9.162761040996979e-06, 'learning_rate/action_model': 9.162761040996978e-05, 'epoch': 0.59}) +[2026-06-17 19:00:10,558][starVLA.training.train_starvla][INFO] - Step 1425, Loss: {'train/loss': 0.6442622542381287, 'train/grad_norm_pre_clip': 10.357095718383789, 'timing/data': 0.0007210597395896912, 'timing/model': 0.9495861819013953, 'learning_rate/qwen_vl_interface': 9.161534457072839e-06, 'learning_rate/action_model': 9.161534457072839e-05, 'epoch': 0.59}) +[2026-06-17 19:00:17,354][starVLA.training.train_starvla][INFO] - Step 1426, Loss: {'train/loss': 0.6456229090690613, 'train/grad_norm_pre_clip': 11.256063461303711, 'timing/data': 0.0010992130264639854, 'timing/model': 0.9506575707346201, 'learning_rate/qwen_vl_interface': 9.160307062283397e-06, 'learning_rate/action_model': 9.160307062283397e-05, 'epoch': 0.59}) +[2026-06-17 19:00:23,522][starVLA.training.train_starvla][INFO] - Step 1427, Loss: {'train/loss': 0.7296343445777893, 'train/grad_norm_pre_clip': 11.057602882385254, 'timing/data': 0.0012811152264475822, 'timing/model': 0.9499535774812102, 'learning_rate/qwen_vl_interface': 9.159078856883093e-06, 'learning_rate/action_model': 9.159078856883093e-05, 'epoch': 0.59}) +[2026-06-17 19:00:27,568][starVLA.training.train_starvla][INFO] - Step 1428, Loss: {'train/loss': 0.6892143487930298, 'train/grad_norm_pre_clip': 12.226052284240723, 'timing/data': 0.0005857199430465698, 'timing/model': 1.2257711552083492, 'learning_rate/qwen_vl_interface': 9.157849841126535e-06, 'learning_rate/action_model': 9.157849841126535e-05, 'epoch': 0.59}) +[2026-06-17 19:00:31,402][starVLA.training.train_starvla][INFO] - Step 1429, Loss: {'train/loss': 0.4504550099372864, 'train/grad_norm_pre_clip': 9.643386840820312, 'timing/data': 0.0007572853937745094, 'timing/model': 1.0615813545882702, 'learning_rate/qwen_vl_interface': 9.156620015268502e-06, 'learning_rate/action_model': 9.1566200152685e-05, 'epoch': 0.59}) +[2026-06-17 19:00:35,345][starVLA.training.train_starvla][INFO] - Step 1430, Loss: {'train/loss': 0.7572632431983948, 'train/grad_norm_pre_clip': 11.446924209594727, 'timing/data': 0.0006240773946046829, 'timing/model': 1.080075797624886, 'learning_rate/qwen_vl_interface': 9.155389379563934e-06, 'learning_rate/action_model': 9.155389379563933e-05, 'epoch': 0.59}) +[2026-06-17 19:00:39,722][starVLA.training.train_starvla][INFO] - Step 1431, Loss: {'train/loss': 0.8027743697166443, 'train/grad_norm_pre_clip': 61.35142135620117, 'timing/data': 0.0005092369392514229, 'timing/model': 0.9683623984456062, 'learning_rate/qwen_vl_interface': 9.154157934267946e-06, 'learning_rate/action_model': 9.154157934267944e-05, 'epoch': 0.59}) +[2026-06-17 19:00:43,970][starVLA.training.train_starvla][INFO] - Step 1432, Loss: {'train/loss': 0.44535455107688904, 'train/grad_norm_pre_clip': 6.529899597167969, 'timing/data': 0.0006870441138744354, 'timing/model': 1.0170239144936204, 'learning_rate/qwen_vl_interface': 9.152925679635817e-06, 'learning_rate/action_model': 9.152925679635817e-05, 'epoch': 0.59}) +[2026-06-17 19:00:48,169][starVLA.training.train_starvla][INFO] - Step 1433, Loss: {'train/loss': 0.4812391400337219, 'train/grad_norm_pre_clip': 8.003104209899902, 'timing/data': 0.0010680072009563446, 'timing/model': 0.9372347621247172, 'learning_rate/qwen_vl_interface': 9.151692615922995e-06, 'learning_rate/action_model': 9.151692615922994e-05, 'epoch': 0.59}) +[2026-06-17 19:00:52,612][starVLA.training.train_starvla][INFO] - Step 1434, Loss: {'train/loss': 0.7015939950942993, 'train/grad_norm_pre_clip': 12.069757461547852, 'timing/data': 0.0008132783696055412, 'timing/model': 1.5232089655473828, 'learning_rate/qwen_vl_interface': 9.150458743385095e-06, 'learning_rate/action_model': 9.150458743385094e-05, 'epoch': 0.59}) +[2026-06-17 19:00:56,621][starVLA.training.train_starvla][INFO] - Step 1435, Loss: {'train/loss': 0.6111434102058411, 'train/grad_norm_pre_clip': 10.708006858825684, 'timing/data': 0.0005323132500052452, 'timing/model': 1.1520724445581436, 'learning_rate/qwen_vl_interface': 9.1492240622779e-06, 'learning_rate/action_model': 9.1492240622779e-05, 'epoch': 0.59}) +[2026-06-17 19:01:00,441][starVLA.training.train_starvla][INFO] - Step 1436, Loss: {'train/loss': 0.36982062458992004, 'train/grad_norm_pre_clip': 19.288677215576172, 'timing/data': 0.000767391175031662, 'timing/model': 0.9844359178096056, 'learning_rate/qwen_vl_interface': 9.147988572857362e-06, 'learning_rate/action_model': 9.147988572857362e-05, 'epoch': 0.59}) +[2026-06-17 19:01:05,621][starVLA.training.train_starvla][INFO] - Step 1437, Loss: {'train/loss': 0.4656480848789215, 'train/grad_norm_pre_clip': 8.449420928955078, 'timing/data': 0.0039052842184901237, 'timing/model': 1.2195096164941788, 'learning_rate/qwen_vl_interface': 9.146752275379599e-06, 'learning_rate/action_model': 9.146752275379598e-05, 'epoch': 0.59}) +[2026-06-17 19:01:10,105][starVLA.training.train_starvla][INFO] - Step 1438, Loss: {'train/loss': 0.8128553628921509, 'train/grad_norm_pre_clip': 12.251008033752441, 'timing/data': 0.000729014165699482, 'timing/model': 1.0365578969940543, 'learning_rate/qwen_vl_interface': 9.145515170100894e-06, 'learning_rate/action_model': 9.145515170100894e-05, 'epoch': 0.59}) +[2026-06-17 19:01:15,708][starVLA.training.train_starvla][INFO] - Step 1439, Loss: {'train/loss': 0.8353434205055237, 'train/grad_norm_pre_clip': 12.225833892822266, 'timing/data': 0.0009888792410492897, 'timing/model': 1.01962682697922, 'learning_rate/qwen_vl_interface': 9.144277257277705e-06, 'learning_rate/action_model': 9.144277257277703e-05, 'epoch': 0.59}) +[2026-06-17 19:01:20,089][starVLA.training.train_starvla][INFO] - Step 1440, Loss: {'train/loss': 0.8629474639892578, 'train/grad_norm_pre_clip': 11.565863609313965, 'timing/data': 0.0008187098428606987, 'timing/model': 1.1414371337741613, 'learning_rate/qwen_vl_interface': 9.14303853716665e-06, 'learning_rate/action_model': 9.143038537166649e-05, 'epoch': 0.59}) +[2026-06-17 19:01:24,838][starVLA.training.train_starvla][INFO] - Step 1441, Loss: {'train/loss': 0.556914210319519, 'train/grad_norm_pre_clip': 9.662263870239258, 'timing/data': 0.0005213525146245956, 'timing/model': 0.9768469389528036, 'learning_rate/qwen_vl_interface': 9.141799010024516e-06, 'learning_rate/action_model': 9.141799010024515e-05, 'epoch': 0.59}) +[2026-06-17 19:01:28,919][starVLA.training.train_starvla][INFO] - Step 1442, Loss: {'train/loss': 0.612027645111084, 'train/grad_norm_pre_clip': 9.242104530334473, 'timing/data': 0.0005400087684392929, 'timing/model': 1.2638254538178444, 'learning_rate/qwen_vl_interface': 9.140558676108258e-06, 'learning_rate/action_model': 9.140558676108259e-05, 'epoch': 0.59}) +[2026-06-17 19:01:32,710][starVLA.training.train_starvla][INFO] - Step 1443, Loss: {'train/loss': 0.5197789669036865, 'train/grad_norm_pre_clip': 7.045263767242432, 'timing/data': 0.0006000623106956482, 'timing/model': 0.9665437005460262, 'learning_rate/qwen_vl_interface': 9.139317535675002e-06, 'learning_rate/action_model': 9.139317535675002e-05, 'epoch': 0.59}) +[2026-06-17 19:01:36,664][starVLA.training.train_starvla][INFO] - Step 1444, Loss: {'train/loss': 0.7367759943008423, 'train/grad_norm_pre_clip': 11.014961242675781, 'timing/data': 0.0005395989865064621, 'timing/model': 0.949894180521369, 'learning_rate/qwen_vl_interface': 9.138075588982035e-06, 'learning_rate/action_model': 9.138075588982036e-05, 'epoch': 0.59}) +[2026-06-17 19:01:40,629][starVLA.training.train_starvla][INFO] - Step 1445, Loss: {'train/loss': 0.7357335090637207, 'train/grad_norm_pre_clip': 12.99428653717041, 'timing/data': 0.0005210898816585541, 'timing/model': 1.057799375616014, 'learning_rate/qwen_vl_interface': 9.136832836286816e-06, 'learning_rate/action_model': 9.136832836286816e-05, 'epoch': 0.59}) +[2026-06-17 19:01:47,325][starVLA.training.train_starvla][INFO] - Step 1446, Loss: {'train/loss': 0.48601987957954407, 'train/grad_norm_pre_clip': 9.989228248596191, 'timing/data': 0.0006261197850108147, 'timing/model': 1.2392587373033166, 'learning_rate/qwen_vl_interface': 9.135589277846967e-06, 'learning_rate/action_model': 9.135589277846966e-05, 'epoch': 0.6}) +[2026-06-17 19:01:52,177][starVLA.training.train_starvla][INFO] - Step 1447, Loss: {'train/loss': 0.6681569814682007, 'train/grad_norm_pre_clip': 9.65760612487793, 'timing/data': 0.0016857869923114777, 'timing/model': 1.8033204497769475, 'learning_rate/qwen_vl_interface': 9.134344913920278e-06, 'learning_rate/action_model': 9.134344913920278e-05, 'epoch': 0.6}) +[2026-06-17 19:01:59,132][starVLA.training.train_starvla][INFO] - Step 1448, Loss: {'train/loss': 0.7818047404289246, 'train/grad_norm_pre_clip': 11.878791809082031, 'timing/data': 0.0005507683381438255, 'timing/model': 1.1638741036877036, 'learning_rate/qwen_vl_interface': 9.133099744764711e-06, 'learning_rate/action_model': 9.133099744764711e-05, 'epoch': 0.6}) +[2026-06-17 19:02:02,934][starVLA.training.train_starvla][INFO] - Step 1449, Loss: {'train/loss': 0.9192196726799011, 'train/grad_norm_pre_clip': 12.006038665771484, 'timing/data': 0.0005711913108825684, 'timing/model': 1.0739452932029963, 'learning_rate/qwen_vl_interface': 9.131853770638386e-06, 'learning_rate/action_model': 9.131853770638386e-05, 'epoch': 0.6}) +[2026-06-17 19:02:06,746][starVLA.training.train_starvla][INFO] - Step 1450, Loss: {'train/loss': 0.5540263652801514, 'train/grad_norm_pre_clip': 14.266498565673828, 'timing/data': 0.0009309081360697746, 'timing/model': 0.9492757422849536, 'learning_rate/qwen_vl_interface': 9.130606991799599e-06, 'learning_rate/action_model': 9.130606991799599e-05, 'epoch': 0.6}) +[2026-06-17 19:02:10,780][starVLA.training.train_starvla][INFO] - Step 1451, Loss: {'train/loss': 0.4808008074760437, 'train/grad_norm_pre_clip': 15.234847068786621, 'timing/data': 0.0005914270877838135, 'timing/model': 0.999795313924551, 'learning_rate/qwen_vl_interface': 9.129359408506808e-06, 'learning_rate/action_model': 9.129359408506808e-05, 'epoch': 0.6}) +[2026-06-17 19:02:14,719][starVLA.training.train_starvla][INFO] - Step 1452, Loss: {'train/loss': 0.5521412491798401, 'train/grad_norm_pre_clip': 10.960336685180664, 'timing/data': 0.0007393611595034599, 'timing/model': 0.9577013552188873, 'learning_rate/qwen_vl_interface': 9.128111021018636e-06, 'learning_rate/action_model': 9.128111021018636e-05, 'epoch': 0.6}) +[2026-06-17 19:02:18,571][starVLA.training.train_starvla][INFO] - Step 1453, Loss: {'train/loss': 0.6214666366577148, 'train/grad_norm_pre_clip': 8.650348663330078, 'timing/data': 0.0005305036902427673, 'timing/model': 0.9686221526935697, 'learning_rate/qwen_vl_interface': 9.126861829593876e-06, 'learning_rate/action_model': 9.126861829593876e-05, 'epoch': 0.6}) +[2026-06-17 19:02:22,556][starVLA.training.train_starvla][INFO] - Step 1454, Loss: {'train/loss': 0.7676475048065186, 'train/grad_norm_pre_clip': 11.571993827819824, 'timing/data': 0.0010087555274367332, 'timing/model': 0.9587655011564493, 'learning_rate/qwen_vl_interface': 9.125611834491487e-06, 'learning_rate/action_model': 9.125611834491486e-05, 'epoch': 0.6}) +[2026-06-17 19:02:26,466][starVLA.training.train_starvla][INFO] - Step 1455, Loss: {'train/loss': 0.6621435880661011, 'train/grad_norm_pre_clip': 13.50035572052002, 'timing/data': 0.0005310336127877235, 'timing/model': 1.1826176838949323, 'learning_rate/qwen_vl_interface': 9.124361035970594e-06, 'learning_rate/action_model': 9.124361035970594e-05, 'epoch': 0.6}) +[2026-06-17 19:02:30,201][starVLA.training.train_starvla][INFO] - Step 1456, Loss: {'train/loss': 0.8494009375572205, 'train/grad_norm_pre_clip': 11.242231369018555, 'timing/data': 0.0009303679689764977, 'timing/model': 1.0167961493134499, 'learning_rate/qwen_vl_interface': 9.123109434290491e-06, 'learning_rate/action_model': 9.123109434290492e-05, 'epoch': 0.6}) +[2026-06-17 19:02:34,143][starVLA.training.train_starvla][INFO] - Step 1457, Loss: {'train/loss': 1.0002521276474, 'train/grad_norm_pre_clip': 707.8775634765625, 'timing/data': 0.0008467147126793861, 'timing/model': 0.9546604789793491, 'learning_rate/qwen_vl_interface': 9.121857029710632e-06, 'learning_rate/action_model': 9.121857029710633e-05, 'epoch': 0.6}) +[2026-06-17 19:02:38,043][starVLA.training.train_starvla][INFO] - Step 1458, Loss: {'train/loss': 0.5707245469093323, 'train/grad_norm_pre_clip': 8.496294975280762, 'timing/data': 0.0011499058455228806, 'timing/model': 0.938716558739543, 'learning_rate/qwen_vl_interface': 9.120603822490647e-06, 'learning_rate/action_model': 9.120603822490647e-05, 'epoch': 0.6}) +[2026-06-17 19:02:42,047][starVLA.training.train_starvla][INFO] - Step 1459, Loss: {'train/loss': 0.657284677028656, 'train/grad_norm_pre_clip': 21.468042373657227, 'timing/data': 0.0007637506350874901, 'timing/model': 0.9665732346475124, 'learning_rate/qwen_vl_interface': 9.119349812890323e-06, 'learning_rate/action_model': 9.119349812890323e-05, 'epoch': 0.6}) +[2026-06-17 19:02:46,116][starVLA.training.train_starvla][INFO] - Step 1460, Loss: {'train/loss': 0.6291716694831848, 'train/grad_norm_pre_clip': 9.451104164123535, 'timing/data': 0.0005335211753845215, 'timing/model': 1.044183176010847, 'learning_rate/qwen_vl_interface': 9.11809500116962e-06, 'learning_rate/action_model': 9.11809500116962e-05, 'epoch': 0.6}) +[2026-06-17 19:02:50,020][starVLA.training.train_starvla][INFO] - Step 1461, Loss: {'train/loss': 0.49427908658981323, 'train/grad_norm_pre_clip': 9.579936981201172, 'timing/data': 0.0004789372906088829, 'timing/model': 0.9563561985269189, 'learning_rate/qwen_vl_interface': 9.116839387588662e-06, 'learning_rate/action_model': 9.116839387588661e-05, 'epoch': 0.6}) +[2026-06-17 19:02:53,992][starVLA.training.train_starvla][INFO] - Step 1462, Loss: {'train/loss': 0.6532509922981262, 'train/grad_norm_pre_clip': 12.055771827697754, 'timing/data': 0.0006947387009859085, 'timing/model': 1.1924598515033722, 'learning_rate/qwen_vl_interface': 9.115582972407735e-06, 'learning_rate/action_model': 9.115582972407736e-05, 'epoch': 0.6}) +[2026-06-17 19:02:57,905][starVLA.training.train_starvla][INFO] - Step 1463, Loss: {'train/loss': 0.40210631489753723, 'train/grad_norm_pre_clip': 8.866538047790527, 'timing/data': 0.001002245582640171, 'timing/model': 1.0473785242065787, 'learning_rate/qwen_vl_interface': 9.114325755887301e-06, 'learning_rate/action_model': 9.114325755887301e-05, 'epoch': 0.6}) +[2026-06-17 19:03:01,849][starVLA.training.train_starvla][INFO] - Step 1464, Loss: {'train/loss': 0.6704474687576294, 'train/grad_norm_pre_clip': 12.117246627807617, 'timing/data': 0.0006213383749127388, 'timing/model': 0.9650968024507165, 'learning_rate/qwen_vl_interface': 9.113067738287978e-06, 'learning_rate/action_model': 9.113067738287979e-05, 'epoch': 0.6}) +[2026-06-17 19:03:05,752][starVLA.training.train_starvla][INFO] - Step 1465, Loss: {'train/loss': 0.7873020768165588, 'train/grad_norm_pre_clip': 11.414595603942871, 'timing/data': 0.001054864376783371, 'timing/model': 0.9474898846819997, 'learning_rate/qwen_vl_interface': 9.111808919870557e-06, 'learning_rate/action_model': 9.111808919870557e-05, 'epoch': 0.6}) +[2026-06-17 19:03:09,661][starVLA.training.train_starvla][INFO] - Step 1466, Loss: {'train/loss': 0.7822771072387695, 'train/grad_norm_pre_clip': 14.708636283874512, 'timing/data': 0.001231054775416851, 'timing/model': 0.9546700716018677, 'learning_rate/qwen_vl_interface': 9.11054930089599e-06, 'learning_rate/action_model': 9.110549300895991e-05, 'epoch': 0.6}) +[2026-06-17 19:03:15,133][starVLA.training.train_starvla][INFO] - Step 1467, Loss: {'train/loss': 0.6557120084762573, 'train/grad_norm_pre_clip': 11.585151672363281, 'timing/data': 0.0007922099903225899, 'timing/model': 2.455484250560403, 'learning_rate/qwen_vl_interface': 9.109288881625402e-06, 'learning_rate/action_model': 9.109288881625402e-05, 'epoch': 0.6}) +[2026-06-17 19:03:19,413][starVLA.training.train_starvla][INFO] - Step 1468, Loss: {'train/loss': 0.5906097888946533, 'train/grad_norm_pre_clip': 15.019247055053711, 'timing/data': 0.0005992930382490158, 'timing/model': 1.2313603749498725, 'learning_rate/qwen_vl_interface': 9.108027662320073e-06, 'learning_rate/action_model': 9.108027662320073e-05, 'epoch': 0.6}) +[2026-06-17 19:03:23,337][starVLA.training.train_starvla][INFO] - Step 1469, Loss: {'train/loss': 0.35770073533058167, 'train/grad_norm_pre_clip': 7.693807601928711, 'timing/data': 0.001046968623995781, 'timing/model': 1.128932329826057, 'learning_rate/qwen_vl_interface': 9.10676564324146e-06, 'learning_rate/action_model': 9.10676564324146e-05, 'epoch': 0.6}) +[2026-06-17 19:03:27,171][starVLA.training.train_starvla][INFO] - Step 1470, Loss: {'train/loss': 0.6050494909286499, 'train/grad_norm_pre_clip': 11.22276496887207, 'timing/data': 0.0008295709267258644, 'timing/model': 1.0105398409068584, 'learning_rate/qwen_vl_interface': 9.105502824651179e-06, 'learning_rate/action_model': 9.105502824651178e-05, 'epoch': 0.61}) +[2026-06-17 19:03:31,122][starVLA.training.train_starvla][INFO] - Step 1471, Loss: {'train/loss': 0.6997743248939514, 'train/grad_norm_pre_clip': 17.050222396850586, 'timing/data': 0.0006180554628372192, 'timing/model': 0.9532822323963046, 'learning_rate/qwen_vl_interface': 9.104239206811012e-06, 'learning_rate/action_model': 9.104239206811011e-05, 'epoch': 0.61}) +[2026-06-17 19:03:35,056][starVLA.training.train_starvla][INFO] - Step 1472, Loss: {'train/loss': 0.6462275981903076, 'train/grad_norm_pre_clip': 10.747772216796875, 'timing/data': 0.0009515099227428436, 'timing/model': 0.9504189854487777, 'learning_rate/qwen_vl_interface': 9.102974789982911e-06, 'learning_rate/action_model': 9.102974789982912e-05, 'epoch': 0.61}) +[2026-06-17 19:03:39,091][starVLA.training.train_starvla][INFO] - Step 1473, Loss: {'train/loss': 0.6305491924285889, 'train/grad_norm_pre_clip': 21.510150909423828, 'timing/data': 0.0012413319200277328, 'timing/model': 0.9688581405207515, 'learning_rate/qwen_vl_interface': 9.10170957442899e-06, 'learning_rate/action_model': 9.101709574428991e-05, 'epoch': 0.61}) +[2026-06-17 19:03:43,012][starVLA.training.train_starvla][INFO] - Step 1474, Loss: {'train/loss': 0.41228076815605164, 'train/grad_norm_pre_clip': 9.096431732177734, 'timing/data': 0.0005579562857747078, 'timing/model': 1.040901992470026, 'learning_rate/qwen_vl_interface': 9.10044356041153e-06, 'learning_rate/action_model': 9.10044356041153e-05, 'epoch': 0.61}) +[2026-06-17 19:03:46,907][starVLA.training.train_starvla][INFO] - Step 1475, Loss: {'train/loss': 0.9271581172943115, 'train/grad_norm_pre_clip': 10.952980995178223, 'timing/data': 0.0009190617129206657, 'timing/model': 1.179790547117591, 'learning_rate/qwen_vl_interface': 9.099176748192976e-06, 'learning_rate/action_model': 9.099176748192976e-05, 'epoch': 0.61}) +[2026-06-17 19:03:50,604][starVLA.training.train_starvla][INFO] - Step 1476, Loss: {'train/loss': 0.7808799147605896, 'train/grad_norm_pre_clip': 11.949969291687012, 'timing/data': 0.0006098626181483269, 'timing/model': 0.9518751548603177, 'learning_rate/qwen_vl_interface': 9.097909138035943e-06, 'learning_rate/action_model': 9.097909138035941e-05, 'epoch': 0.61}) +[2026-06-17 19:03:54,568][starVLA.training.train_starvla][INFO] - Step 1477, Loss: {'train/loss': 0.6202033758163452, 'train/grad_norm_pre_clip': 12.28117561340332, 'timing/data': 0.0005154553800821304, 'timing/model': 0.9475307362154126, 'learning_rate/qwen_vl_interface': 9.096640730203202e-06, 'learning_rate/action_model': 9.096640730203202e-05, 'epoch': 0.61}) +[2026-06-17 19:04:02,175][starVLA.training.train_starvla][INFO] - Step 1478, Loss: {'train/loss': 0.6818605065345764, 'train/grad_norm_pre_clip': 9.639902114868164, 'timing/data': 0.0011706668883562088, 'timing/model': 0.9872602336108685, 'learning_rate/qwen_vl_interface': 9.095371524957698e-06, 'learning_rate/action_model': 9.095371524957698e-05, 'epoch': 0.61}) +[2026-06-17 19:04:06,165][starVLA.training.train_starvla][INFO] - Step 1479, Loss: {'train/loss': 0.5507432818412781, 'train/grad_norm_pre_clip': 9.425796508789062, 'timing/data': 0.0005412818863987923, 'timing/model': 0.9472053525969386, 'learning_rate/qwen_vl_interface': 9.094101522562541e-06, 'learning_rate/action_model': 9.09410152256254e-05, 'epoch': 0.61}) +[2026-06-17 19:04:10,192][starVLA.training.train_starvla][INFO] - Step 1480, Loss: {'train/loss': 0.4940974712371826, 'train/grad_norm_pre_clip': 8.775909423828125, 'timing/data': 0.0015428820624947548, 'timing/model': 0.9963088426738977, 'learning_rate/qwen_vl_interface': 9.092830723281e-06, 'learning_rate/action_model': 9.092830723281e-05, 'epoch': 0.61}) +[2026-06-17 19:04:14,191][starVLA.training.train_starvla][INFO] - Step 1481, Loss: {'train/loss': 0.9921098351478577, 'train/grad_norm_pre_clip': 16.634151458740234, 'timing/data': 0.000502399168908596, 'timing/model': 1.0271077593788505, 'learning_rate/qwen_vl_interface': 9.091559127376519e-06, 'learning_rate/action_model': 9.091559127376517e-05, 'epoch': 0.61}) +[2026-06-17 19:04:18,410][starVLA.training.train_starvla][INFO] - Step 1482, Loss: {'train/loss': 0.6877655386924744, 'train/grad_norm_pre_clip': 10.936753273010254, 'timing/data': 0.0010727066546678543, 'timing/model': 1.237444557249546, 'learning_rate/qwen_vl_interface': 9.090286735112693e-06, 'learning_rate/action_model': 9.090286735112693e-05, 'epoch': 0.61}) +[2026-06-17 19:04:22,212][starVLA.training.train_starvla][INFO] - Step 1483, Loss: {'train/loss': 0.5375478267669678, 'train/grad_norm_pre_clip': 9.60879135131836, 'timing/data': 0.0005473531782627106, 'timing/model': 0.9901019791141152, 'learning_rate/qwen_vl_interface': 9.089013546753296e-06, 'learning_rate/action_model': 9.089013546753295e-05, 'epoch': 0.61}) +[2026-06-17 19:04:26,341][starVLA.training.train_starvla][INFO] - Step 1484, Loss: {'train/loss': 0.599169135093689, 'train/grad_norm_pre_clip': 10.695151329040527, 'timing/data': 0.0005409885197877884, 'timing/model': 1.0976869789883494, 'learning_rate/qwen_vl_interface': 9.08773956256226e-06, 'learning_rate/action_model': 9.087739562562259e-05, 'epoch': 0.61}) +[2026-06-17 19:04:30,349][starVLA.training.train_starvla][INFO] - Step 1485, Loss: {'train/loss': 0.623745322227478, 'train/grad_norm_pre_clip': 19.976911544799805, 'timing/data': 0.0016299895942211151, 'timing/model': 0.9488399215042591, 'learning_rate/qwen_vl_interface': 9.086464782803682e-06, 'learning_rate/action_model': 9.086464782803682e-05, 'epoch': 0.61}) +[2026-06-17 19:04:36,390][starVLA.training.train_starvla][INFO] - Step 1486, Loss: {'train/loss': 0.5261691808700562, 'train/grad_norm_pre_clip': 8.933513641357422, 'timing/data': 0.000940161757171154, 'timing/model': 0.9470887612551451, 'learning_rate/qwen_vl_interface': 9.085189207741827e-06, 'learning_rate/action_model': 9.085189207741827e-05, 'epoch': 0.61}) +[2026-06-17 19:04:40,280][starVLA.training.train_starvla][INFO] - Step 1487, Loss: {'train/loss': 0.3589365780353546, 'train/grad_norm_pre_clip': 9.68763256072998, 'timing/data': 0.0013469671830534935, 'timing/model': 0.9499493567273021, 'learning_rate/qwen_vl_interface': 9.083912837641123e-06, 'learning_rate/action_model': 9.083912837641123e-05, 'epoch': 0.61}) +[2026-06-17 19:04:44,246][starVLA.training.train_starvla][INFO] - Step 1488, Loss: {'train/loss': 0.9675340056419373, 'train/grad_norm_pre_clip': 12.102307319641113, 'timing/data': 0.0005687568336725235, 'timing/model': 1.2136365370824933, 'learning_rate/qwen_vl_interface': 9.082635672766162e-06, 'learning_rate/action_model': 9.082635672766161e-05, 'epoch': 0.61}) +[2026-06-17 19:04:48,212][starVLA.training.train_starvla][INFO] - Step 1489, Loss: {'train/loss': 0.734836757183075, 'train/grad_norm_pre_clip': 11.346907615661621, 'timing/data': 0.0011581974104046822, 'timing/model': 1.1314976466819644, 'learning_rate/qwen_vl_interface': 9.081357713381702e-06, 'learning_rate/action_model': 9.081357713381701e-05, 'epoch': 0.61}) +[2026-06-17 19:04:51,854][starVLA.training.train_starvla][INFO] - Step 1490, Loss: {'train/loss': 0.5734305381774902, 'train/grad_norm_pre_clip': 15.009848594665527, 'timing/data': 0.0005724355578422546, 'timing/model': 0.9402798591181636, 'learning_rate/qwen_vl_interface': 9.080078959752664e-06, 'learning_rate/action_model': 9.080078959752664e-05, 'epoch': 0.61}) +[2026-06-17 19:04:55,737][starVLA.training.train_starvla][INFO] - Step 1491, Loss: {'train/loss': 0.4378308057785034, 'train/grad_norm_pre_clip': 10.080778121948242, 'timing/data': 0.0011668680235743523, 'timing/model': 0.9445932479575276, 'learning_rate/qwen_vl_interface': 9.078799412144137e-06, 'learning_rate/action_model': 9.078799412144137e-05, 'epoch': 0.61}) +[2026-06-17 19:04:59,718][starVLA.training.train_starvla][INFO] - Step 1492, Loss: {'train/loss': 0.41874533891677856, 'train/grad_norm_pre_clip': 7.640468597412109, 'timing/data': 0.0005437768995761871, 'timing/model': 0.9780191648751497, 'learning_rate/qwen_vl_interface': 9.077519070821373e-06, 'learning_rate/action_model': 9.077519070821373e-05, 'epoch': 0.61}) +[2026-06-17 19:05:03,806][starVLA.training.train_starvla][INFO] - Step 1493, Loss: {'train/loss': 0.7492466568946838, 'train/grad_norm_pre_clip': 12.695198059082031, 'timing/data': 0.00073190126568079, 'timing/model': 0.9546018447726965, 'learning_rate/qwen_vl_interface': 9.076237936049787e-06, 'learning_rate/action_model': 9.076237936049786e-05, 'epoch': 0.61}) +[2026-06-17 19:05:07,706][starVLA.training.train_starvla][INFO] - Step 1494, Loss: {'train/loss': 0.5231207609176636, 'train/grad_norm_pre_clip': 8.157227516174316, 'timing/data': 0.0011151321232318878, 'timing/model': 0.9499414404854178, 'learning_rate/qwen_vl_interface': 9.074956008094958e-06, 'learning_rate/action_model': 9.074956008094958e-05, 'epoch': 0.61}) +[2026-06-17 19:05:11,584][starVLA.training.train_starvla][INFO] - Step 1495, Loss: {'train/loss': 0.4936434030532837, 'train/grad_norm_pre_clip': 10.508577346801758, 'timing/data': 0.0008383793756365776, 'timing/model': 1.1431540604680777, 'learning_rate/qwen_vl_interface': 9.073673287222634e-06, 'learning_rate/action_model': 9.073673287222633e-05, 'epoch': 0.62}) +[2026-06-17 19:05:15,222][starVLA.training.train_starvla][INFO] - Step 1496, Loss: {'train/loss': 0.6147783398628235, 'train/grad_norm_pre_clip': 11.419088363647461, 'timing/data': 0.0009289886802434921, 'timing/model': 0.9288044143468142, 'learning_rate/qwen_vl_interface': 9.072389773698723e-06, 'learning_rate/action_model': 9.072389773698723e-05, 'epoch': 0.62}) +[2026-06-17 19:05:19,265][starVLA.training.train_starvla][INFO] - Step 1497, Loss: {'train/loss': 0.5967706441879272, 'train/grad_norm_pre_clip': 9.631956100463867, 'timing/data': 0.000660398043692112, 'timing/model': 0.9472501492127776, 'learning_rate/qwen_vl_interface': 9.071105467789299e-06, 'learning_rate/action_model': 9.071105467789298e-05, 'epoch': 0.62}) +[2026-06-17 19:05:23,395][starVLA.training.train_starvla][INFO] - Step 1498, Loss: {'train/loss': 0.679637610912323, 'train/grad_norm_pre_clip': 10.118935585021973, 'timing/data': 0.0016346219927072525, 'timing/model': 1.0074698589742184, 'learning_rate/qwen_vl_interface': 9.0698203697606e-06, 'learning_rate/action_model': 9.0698203697606e-05, 'epoch': 0.62}) +[2026-06-17 19:05:28,377][starVLA.training.train_starvla][INFO] - Step 1499, Loss: {'train/loss': 0.5000624656677246, 'train/grad_norm_pre_clip': 32.23580551147461, 'timing/data': 0.0007471544668078423, 'timing/model': 0.9777920767664909, 'learning_rate/qwen_vl_interface': 9.068534479879026e-06, 'learning_rate/action_model': 9.068534479879024e-05, 'epoch': 0.62}) +[2026-06-17 19:07:39,023][starVLA.training.train_starvla][INFO] - Step 1500, Loss: {'train/loss': 0.44427552819252014, 'train/grad_norm_pre_clip': 8.555424690246582, 'eval/loss': 0.6252224445343018, 'eval/latency_6/loss': 0.6252224445343018, 'eval/demon_fire_precision': 0.9336463223787167, 'eval/demon_fire_recall': 0.9641241111829347, 'eval/demon_fire_f1': 0.9486404833836858, 'eval/demon_fire_rate_model': 0.8889816360601002, 'eval/demon_fire_rate_teacher': 0.860879243183083, 'eval/demon_left_precision': 0.730844793713163, 'eval/demon_left_recall': 0.6526315789473685, 'eval/demon_left_f1': 0.6895273401297498, 'eval/demon_left_rate_model': 0.1416249304396216, 'eval/demon_left_rate_teacher': 0.15859766277128548, 'eval/demon_right_precision': 0.7857142857142857, 'eval/demon_right_recall': 0.6082949308755761, 'eval/demon_right_f1': 0.6857142857142858, 'eval/demon_right_rate_model': 0.14023372287145242, 'eval/demon_right_rate_teacher': 0.18113522537562604, 'eval/demon_move_f1': 0.6876208129220178, 'eval/demon_cc_f1': 0.8516545099946538, 'eval/latency_6/demon_fire_precision': 0.9336463223787167, 'eval/latency_6/demon_fire_recall': 0.9641241111829347, 'eval/latency_6/demon_fire_f1': 0.9486404833836858, 'eval/latency_6/demon_fire_rate_model': 0.8889816360601002, 'eval/latency_6/demon_fire_rate_teacher': 0.860879243183083, 'eval/latency_6/demon_left_precision': 0.730844793713163, 'eval/latency_6/demon_left_recall': 0.6526315789473685, 'eval/latency_6/demon_left_f1': 0.6895273401297498, 'eval/latency_6/demon_left_rate_model': 0.1416249304396216, 'eval/latency_6/demon_left_rate_teacher': 0.15859766277128548, 'eval/latency_6/demon_right_precision': 0.7857142857142857, 'eval/latency_6/demon_right_recall': 0.6082949308755761, 'eval/latency_6/demon_right_f1': 0.6857142857142858, 'eval/latency_6/demon_right_rate_model': 0.14023372287145242, 'eval/latency_6/demon_right_rate_teacher': 0.18113522537562604, 'eval/latency_6/demon_move_f1': 0.6876208129220178, 'eval/latency_6/demon_cc_f1': 0.8516545099946538, 'timing/data': 0.0005122534930706024, 'timing/model': 1.0789349721744657, 'learning_rate/qwen_vl_interface': 9.067247798411147e-06, 'learning_rate/action_model': 9.067247798411147e-05, 'epoch': 0.62}) +[2026-06-17 19:07:39,025][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_1500_state +[2026-06-17 19:07:39,034][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 19:07:45,488][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_1500_state/model.safetensors +[2026-06-17 19:08:01,089][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_1500_state/optimizer.bin +[2026-06-17 19:08:01,091][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_1500_state/sampler.bin +[2026-06-17 19:08:01,091][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_1500_state/sampler_1.bin +[2026-06-17 19:08:01,093][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_1500_state/random_states_0.pkl +[2026-06-17 19:08:01,097][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1500 +[2026-06-17 19:08:01,097][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 19:08:01,110][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 19:08:05,151][starVLA.training.train_starvla][INFO] - Step 1501, Loss: {'train/loss': 0.7598141431808472, 'train/grad_norm_pre_clip': 10.21596908569336, 'timing/data': 0.0007374323904514313, 'timing/model': 0.9424124388024211, 'learning_rate/qwen_vl_interface': 9.065960325623692e-06, 'learning_rate/action_model': 9.065960325623692e-05, 'epoch': 0.62}) +[2026-06-17 19:08:11,374][starVLA.training.train_starvla][INFO] - Step 1502, Loss: {'train/loss': 0.4946843385696411, 'train/grad_norm_pre_clip': 9.52900505065918, 'timing/data': 0.0005234302952885628, 'timing/model': 0.9389873705804348, 'learning_rate/qwen_vl_interface': 9.064672061783555e-06, 'learning_rate/action_model': 9.064672061783554e-05, 'epoch': 0.62}) +[2026-06-17 19:08:15,206][starVLA.training.train_starvla][INFO] - Step 1503, Loss: {'train/loss': 0.6964267492294312, 'train/grad_norm_pre_clip': 13.841705322265625, 'timing/data': 0.0008579278364777565, 'timing/model': 1.1183003950864077, 'learning_rate/qwen_vl_interface': 9.063383007157795e-06, 'learning_rate/action_model': 9.063383007157796e-05, 'epoch': 0.62}) +[2026-06-17 19:08:18,973][starVLA.training.train_starvla][INFO] - Step 1504, Loss: {'train/loss': 0.6475362181663513, 'train/grad_norm_pre_clip': 10.850359916687012, 'timing/data': 0.0006994539871811867, 'timing/model': 1.077942012809217, 'learning_rate/qwen_vl_interface': 9.062093162013635e-06, 'learning_rate/action_model': 9.062093162013634e-05, 'epoch': 0.62}) +[2026-06-17 19:08:22,708][starVLA.training.train_starvla][INFO] - Step 1505, Loss: {'train/loss': 1.0410319566726685, 'train/grad_norm_pre_clip': 14.045113563537598, 'timing/data': 0.0005653304979205132, 'timing/model': 0.9610003940761089, 'learning_rate/qwen_vl_interface': 9.06080252661846e-06, 'learning_rate/action_model': 9.060802526618458e-05, 'epoch': 0.62}) +[2026-06-17 19:08:26,659][starVLA.training.train_starvla][INFO] - Step 1506, Loss: {'train/loss': 0.8147943615913391, 'train/grad_norm_pre_clip': 8.932291984558105, 'timing/data': 0.0005870722234249115, 'timing/model': 0.9367700582370162, 'learning_rate/qwen_vl_interface': 9.059511101239818e-06, 'learning_rate/action_model': 9.059511101239818e-05, 'epoch': 0.62}) +[2026-06-17 19:08:30,556][starVLA.training.train_starvla][INFO] - Step 1507, Loss: {'train/loss': 0.4693808853626251, 'train/grad_norm_pre_clip': 8.127747535705566, 'timing/data': 0.0005237571895122528, 'timing/model': 0.9411973087117076, 'learning_rate/qwen_vl_interface': 9.058218886145427e-06, 'learning_rate/action_model': 9.058218886145427e-05, 'epoch': 0.62}) +[2026-06-17 19:08:35,148][starVLA.training.train_starvla][INFO] - Step 1508, Loss: {'train/loss': 0.7737843990325928, 'train/grad_norm_pre_clip': 11.508368492126465, 'timing/data': 0.0007849941030144691, 'timing/model': 0.942582375369966, 'learning_rate/qwen_vl_interface': 9.056925881603165e-06, 'learning_rate/action_model': 9.056925881603166e-05, 'epoch': 0.62}) +[2026-06-17 19:08:39,597][starVLA.training.train_starvla][INFO] - Step 1509, Loss: {'train/loss': 0.6416523456573486, 'train/grad_norm_pre_clip': 9.834107398986816, 'timing/data': 0.0005623055621981621, 'timing/model': 0.9552937680855393, 'learning_rate/qwen_vl_interface': 9.05563208788107e-06, 'learning_rate/action_model': 9.05563208788107e-05, 'epoch': 0.62}) +[2026-06-17 19:08:43,509][starVLA.training.train_starvla][INFO] - Step 1510, Loss: {'train/loss': 0.6527760028839111, 'train/grad_norm_pre_clip': 11.719489097595215, 'timing/data': 0.0005112728103995323, 'timing/model': 1.1790584083646536, 'learning_rate/qwen_vl_interface': 9.054337505247349e-06, 'learning_rate/action_model': 9.054337505247348e-05, 'epoch': 0.62}) +[2026-06-17 19:08:47,194][starVLA.training.train_starvla][INFO] - Step 1511, Loss: {'train/loss': 0.912011981010437, 'train/grad_norm_pre_clip': 10.483325958251953, 'timing/data': 0.0005429014563560486, 'timing/model': 0.956001402810216, 'learning_rate/qwen_vl_interface': 9.05304213397037e-06, 'learning_rate/action_model': 9.053042133970369e-05, 'epoch': 0.62}) +[2026-06-17 19:08:51,050][starVLA.training.train_starvla][INFO] - Step 1512, Loss: {'train/loss': 0.9448821544647217, 'train/grad_norm_pre_clip': 14.723104476928711, 'timing/data': 0.0007828641682863235, 'timing/model': 0.9771006209775805, 'learning_rate/qwen_vl_interface': 9.051745974318663e-06, 'learning_rate/action_model': 9.051745974318662e-05, 'epoch': 0.62}) +[2026-06-17 19:08:55,548][starVLA.training.train_starvla][INFO] - Step 1513, Loss: {'train/loss': 0.4673345685005188, 'train/grad_norm_pre_clip': 9.413990020751953, 'timing/data': 0.0006337855011224747, 'timing/model': 0.958929605782032, 'learning_rate/qwen_vl_interface': 9.050449026560923e-06, 'learning_rate/action_model': 9.050449026560924e-05, 'epoch': 0.62}) +[2026-06-17 19:08:59,714][starVLA.training.train_starvla][INFO] - Step 1514, Loss: {'train/loss': 0.5457044839859009, 'train/grad_norm_pre_clip': 9.569989204406738, 'timing/data': 0.0005353726446628571, 'timing/model': 0.9994164807721972, 'learning_rate/qwen_vl_interface': 9.049151290966013e-06, 'learning_rate/action_model': 9.049151290966012e-05, 'epoch': 0.62}) +[2026-06-17 19:09:05,398][starVLA.training.train_starvla][INFO] - Step 1515, Loss: {'train/loss': 0.5895625352859497, 'train/grad_norm_pre_clip': 10.618016242980957, 'timing/data': 0.000567813403904438, 'timing/model': 1.0208963844925165, 'learning_rate/qwen_vl_interface': 9.04785276780295e-06, 'learning_rate/action_model': 9.04785276780295e-05, 'epoch': 0.62}) +[2026-06-17 19:09:10,389][starVLA.training.train_starvla][INFO] - Step 1516, Loss: {'train/loss': 0.870297372341156, 'train/grad_norm_pre_clip': 11.766529083251953, 'timing/data': 0.00048562511801719666, 'timing/model': 2.0207476764917374, 'learning_rate/qwen_vl_interface': 9.046553457340924e-06, 'learning_rate/action_model': 9.046553457340924e-05, 'epoch': 0.62}) +[2026-06-17 19:09:16,067][starVLA.training.train_starvla][INFO] - Step 1517, Loss: {'train/loss': 0.5101059675216675, 'train/grad_norm_pre_clip': 8.88050365447998, 'timing/data': 0.0005615008994936943, 'timing/model': 1.204146663658321, 'learning_rate/qwen_vl_interface': 9.04525335984928e-06, 'learning_rate/action_model': 9.04525335984928e-05, 'epoch': 0.62}) +[2026-06-17 19:09:19,828][starVLA.training.train_starvla][INFO] - Step 1518, Loss: {'train/loss': 0.606624960899353, 'train/grad_norm_pre_clip': 9.699546813964844, 'timing/data': 0.0011845100671052933, 'timing/model': 0.9673244804143906, 'learning_rate/qwen_vl_interface': 9.043952475597532e-06, 'learning_rate/action_model': 9.043952475597532e-05, 'epoch': 0.62}) +[2026-06-17 19:09:23,867][starVLA.training.train_starvla][INFO] - Step 1519, Loss: {'train/loss': 0.8756406307220459, 'train/grad_norm_pre_clip': 10.56635856628418, 'timing/data': 0.0006852252408862114, 'timing/model': 0.9669278915971518, 'learning_rate/qwen_vl_interface': 9.04265080485535e-06, 'learning_rate/action_model': 9.042650804855351e-05, 'epoch': 0.63}) +[2026-06-17 19:09:27,761][starVLA.training.train_starvla][INFO] - Step 1520, Loss: {'train/loss': 0.3915223479270935, 'train/grad_norm_pre_clip': 7.288628578186035, 'timing/data': 0.0009984225034713745, 'timing/model': 0.9570405343547463, 'learning_rate/qwen_vl_interface': 9.041348347892578e-06, 'learning_rate/action_model': 9.041348347892579e-05, 'epoch': 0.63}) +[2026-06-17 19:09:31,741][starVLA.training.train_starvla][INFO] - Step 1521, Loss: {'train/loss': 0.6534750461578369, 'train/grad_norm_pre_clip': 10.467137336730957, 'timing/data': 0.0005522165447473526, 'timing/model': 0.974957768805325, 'learning_rate/qwen_vl_interface': 9.040045104979213e-06, 'learning_rate/action_model': 9.040045104979214e-05, 'epoch': 0.63}) +[2026-06-17 19:09:35,652][starVLA.training.train_starvla][INFO] - Step 1522, Loss: {'train/loss': 0.6643916964530945, 'train/grad_norm_pre_clip': 9.675395011901855, 'timing/data': 0.0009070718660950661, 'timing/model': 0.9649915434420109, 'learning_rate/qwen_vl_interface': 9.03874107638542e-06, 'learning_rate/action_model': 9.038741076385419e-05, 'epoch': 0.63}) +[2026-06-17 19:09:39,639][starVLA.training.train_starvla][INFO] - Step 1523, Loss: {'train/loss': 0.38871410489082336, 'train/grad_norm_pre_clip': 7.92979097366333, 'timing/data': 0.0007228143513202667, 'timing/model': 1.1983511075377464, 'learning_rate/qwen_vl_interface': 9.037436262381525e-06, 'learning_rate/action_model': 9.037436262381525e-05, 'epoch': 0.63}) +[2026-06-17 19:09:43,618][starVLA.training.train_starvla][INFO] - Step 1524, Loss: {'train/loss': 0.706372082233429, 'train/grad_norm_pre_clip': 13.526287078857422, 'timing/data': 0.0006484240293502808, 'timing/model': 1.1777545427903533, 'learning_rate/qwen_vl_interface': 9.036130663238018e-06, 'learning_rate/action_model': 9.036130663238017e-05, 'epoch': 0.63}) +[2026-06-17 19:09:47,434][starVLA.training.train_starvla][INFO] - Step 1525, Loss: {'train/loss': 0.6864957213401794, 'train/grad_norm_pre_clip': 9.470358848571777, 'timing/data': 0.0005423706024885178, 'timing/model': 1.055925901979208, 'learning_rate/qwen_vl_interface': 9.034824279225547e-06, 'learning_rate/action_model': 9.034824279225547e-05, 'epoch': 0.63}) +[2026-06-17 19:09:51,459][starVLA.training.train_starvla][INFO] - Step 1526, Loss: {'train/loss': 0.7758185863494873, 'train/grad_norm_pre_clip': 10.309616088867188, 'timing/data': 0.0007366733625531197, 'timing/model': 0.9688050588592887, 'learning_rate/qwen_vl_interface': 9.033517110614933e-06, 'learning_rate/action_model': 9.033517110614933e-05, 'epoch': 0.63}) +[2026-06-17 19:09:56,924][starVLA.training.train_starvla][INFO] - Step 1527, Loss: {'train/loss': 0.7757823467254639, 'train/grad_norm_pre_clip': 12.199999809265137, 'timing/data': 0.0005266321823000908, 'timing/model': 2.3696301644667983, 'learning_rate/qwen_vl_interface': 9.032209157677151e-06, 'learning_rate/action_model': 9.032209157677151e-05, 'epoch': 0.63}) +[2026-06-17 19:10:00,919][starVLA.training.train_starvla][INFO] - Step 1528, Loss: {'train/loss': 0.613724946975708, 'train/grad_norm_pre_clip': 10.310672760009766, 'timing/data': 0.0007592756301164627, 'timing/model': 0.9910945249721408, 'learning_rate/qwen_vl_interface': 9.03090042068334e-06, 'learning_rate/action_model': 9.030900420683339e-05, 'epoch': 0.63}) +[2026-06-17 19:10:04,959][starVLA.training.train_starvla][INFO] - Step 1529, Loss: {'train/loss': 0.9555121660232544, 'train/grad_norm_pre_clip': 12.992515563964844, 'timing/data': 0.0007854066789150238, 'timing/model': 1.0855801859870553, 'learning_rate/qwen_vl_interface': 9.029590899904803e-06, 'learning_rate/action_model': 9.029590899904803e-05, 'epoch': 0.63}) +[2026-06-17 19:10:08,842][starVLA.training.train_starvla][INFO] - Step 1530, Loss: {'train/loss': 0.5906011462211609, 'train/grad_norm_pre_clip': 8.893640518188477, 'timing/data': 0.0011406056582927704, 'timing/model': 1.1233000932261348, 'learning_rate/qwen_vl_interface': 9.028280595613006e-06, 'learning_rate/action_model': 9.028280595613005e-05, 'epoch': 0.63}) +[2026-06-17 19:10:12,548][starVLA.training.train_starvla][INFO] - Step 1531, Loss: {'train/loss': 0.5108704566955566, 'train/grad_norm_pre_clip': 9.067490577697754, 'timing/data': 0.0010672705247998238, 'timing/model': 0.9483616836369038, 'learning_rate/qwen_vl_interface': 9.026969508079575e-06, 'learning_rate/action_model': 9.026969508079575e-05, 'epoch': 0.63}) +[2026-06-17 19:10:16,478][starVLA.training.train_starvla][INFO] - Step 1532, Loss: {'train/loss': 0.42396488785743713, 'train/grad_norm_pre_clip': 7.416531085968018, 'timing/data': 0.0005716066807508469, 'timing/model': 0.952473065815866, 'learning_rate/qwen_vl_interface': 9.025657637576302e-06, 'learning_rate/action_model': 9.025657637576302e-05, 'epoch': 0.63}) +[2026-06-17 19:10:20,334][starVLA.training.train_starvla][INFO] - Step 1533, Loss: {'train/loss': 0.5970576405525208, 'train/grad_norm_pre_clip': 11.391963005065918, 'timing/data': 0.0007763002067804337, 'timing/model': 0.939427712932229, 'learning_rate/qwen_vl_interface': 9.024344984375138e-06, 'learning_rate/action_model': 9.024344984375137e-05, 'epoch': 0.63}) +[2026-06-17 19:10:24,430][starVLA.training.train_starvla][INFO] - Step 1534, Loss: {'train/loss': 0.5722433924674988, 'train/grad_norm_pre_clip': 12.505536079406738, 'timing/data': 0.0015739798545837402, 'timing/model': 0.9749147854745388, 'learning_rate/qwen_vl_interface': 9.023031548748197e-06, 'learning_rate/action_model': 9.023031548748196e-05, 'epoch': 0.63}) +[2026-06-17 19:10:28,454][starVLA.training.train_starvla][INFO] - Step 1535, Loss: {'train/loss': 0.9353699684143066, 'train/grad_norm_pre_clip': 14.725058555603027, 'timing/data': 0.0008079921826720238, 'timing/model': 0.9497198984026909, 'learning_rate/qwen_vl_interface': 9.021717330967755e-06, 'learning_rate/action_model': 9.021717330967754e-05, 'epoch': 0.63}) +[2026-06-17 19:10:32,334][starVLA.training.train_starvla][INFO] - Step 1536, Loss: {'train/loss': 0.708976686000824, 'train/grad_norm_pre_clip': 9.413806915283203, 'timing/data': 0.0007495544850826263, 'timing/model': 0.963415670208633, 'learning_rate/qwen_vl_interface': 9.020402331306254e-06, 'learning_rate/action_model': 9.020402331306253e-05, 'epoch': 0.63}) +[2026-06-17 19:10:36,342][starVLA.training.train_starvla][INFO] - Step 1537, Loss: {'train/loss': 0.585857093334198, 'train/grad_norm_pre_clip': 10.028615951538086, 'timing/data': 0.0005579330027103424, 'timing/model': 1.1999016171321273, 'learning_rate/qwen_vl_interface': 9.01908655003629e-06, 'learning_rate/action_model': 9.01908655003629e-05, 'epoch': 0.63}) +[2026-06-17 19:10:41,491][starVLA.training.train_starvla][INFO] - Step 1538, Loss: {'train/loss': 0.5139613747596741, 'train/grad_norm_pre_clip': 6.908317565917969, 'timing/data': 0.0008813189342617989, 'timing/model': 0.9882429512217641, 'learning_rate/qwen_vl_interface': 9.017769987430629e-06, 'learning_rate/action_model': 9.017769987430628e-05, 'epoch': 0.63}) +[2026-06-17 19:10:46,310][starVLA.training.train_starvla][INFO] - Step 1539, Loss: {'train/loss': 0.5658520460128784, 'train/grad_norm_pre_clip': 8.866332054138184, 'timing/data': 0.000582721084356308, 'timing/model': 0.9695223905146122, 'learning_rate/qwen_vl_interface': 9.016452643762197e-06, 'learning_rate/action_model': 9.016452643762197e-05, 'epoch': 0.63}) +[2026-06-17 19:10:53,169][starVLA.training.train_starvla][INFO] - Step 1540, Loss: {'train/loss': 1.054746389389038, 'train/grad_norm_pre_clip': 18.028213500976562, 'timing/data': 0.0013334192335605621, 'timing/model': 1.3162775412201881, 'learning_rate/qwen_vl_interface': 9.015134519304078e-06, 'learning_rate/action_model': 9.015134519304078e-05, 'epoch': 0.63}) +[2026-06-17 19:10:58,063][starVLA.training.train_starvla][INFO] - Step 1541, Loss: {'train/loss': 0.6821140646934509, 'train/grad_norm_pre_clip': 10.633026123046875, 'timing/data': 0.0004644058644771576, 'timing/model': 0.9404997993260622, 'learning_rate/qwen_vl_interface': 9.013815614329521e-06, 'learning_rate/action_model': 9.013815614329522e-05, 'epoch': 0.63}) +[2026-06-17 19:11:03,231][starVLA.training.train_starvla][INFO] - Step 1542, Loss: {'train/loss': 0.660551130771637, 'train/grad_norm_pre_clip': 11.601641654968262, 'timing/data': 0.0031270040199160576, 'timing/model': 1.0226784199476242, 'learning_rate/qwen_vl_interface': 9.01249592911194e-06, 'learning_rate/action_model': 9.012495929111938e-05, 'epoch': 0.63}) +[2026-06-17 19:11:07,180][starVLA.training.train_starvla][INFO] - Step 1543, Loss: {'train/loss': 0.5963566303253174, 'train/grad_norm_pre_clip': 11.19882869720459, 'timing/data': 0.0005478393286466599, 'timing/model': 1.2003732975572348, 'learning_rate/qwen_vl_interface': 9.0111754639249e-06, 'learning_rate/action_model': 9.0111754639249e-05, 'epoch': 0.64}) +[2026-06-17 19:11:11,164][starVLA.training.train_starvla][INFO] - Step 1544, Loss: {'train/loss': 0.7523895502090454, 'train/grad_norm_pre_clip': 13.549334526062012, 'timing/data': 0.0010514678433537483, 'timing/model': 1.26293764449656, 'learning_rate/qwen_vl_interface': 9.009854219042142e-06, 'learning_rate/action_model': 9.009854219042142e-05, 'epoch': 0.64}) +[2026-06-17 19:11:14,859][starVLA.training.train_starvla][INFO] - Step 1545, Loss: {'train/loss': 0.45865216851234436, 'train/grad_norm_pre_clip': 8.900569915771484, 'timing/data': 0.0008593425154685974, 'timing/model': 0.9453057758510113, 'learning_rate/qwen_vl_interface': 9.008532194737556e-06, 'learning_rate/action_model': 9.008532194737557e-05, 'epoch': 0.64}) +[2026-06-17 19:11:18,814][starVLA.training.train_starvla][INFO] - Step 1546, Loss: {'train/loss': 0.6989187002182007, 'train/grad_norm_pre_clip': 12.9845552444458, 'timing/data': 0.0006379084661602974, 'timing/model': 0.951386178843677, 'learning_rate/qwen_vl_interface': 9.007209391285205e-06, 'learning_rate/action_model': 9.007209391285205e-05, 'epoch': 0.64}) +[2026-06-17 19:11:23,002][starVLA.training.train_starvla][INFO] - Step 1547, Loss: {'train/loss': 0.44942352175712585, 'train/grad_norm_pre_clip': 9.316069602966309, 'timing/data': 0.0005747564136981964, 'timing/model': 1.0283024217933416, 'learning_rate/qwen_vl_interface': 9.005885808959301e-06, 'learning_rate/action_model': 9.005885808959302e-05, 'epoch': 0.64}) +[2026-06-17 19:11:26,900][starVLA.training.train_starvla][INFO] - Step 1548, Loss: {'train/loss': 0.4261666238307953, 'train/grad_norm_pre_clip': 7.634100437164307, 'timing/data': 0.0011165179312229156, 'timing/model': 0.9513945346698165, 'learning_rate/qwen_vl_interface': 9.00456144803423e-06, 'learning_rate/action_model': 9.00456144803423e-05, 'epoch': 0.64}) +[2026-06-17 19:11:30,807][starVLA.training.train_starvla][INFO] - Step 1549, Loss: {'train/loss': 0.7197768092155457, 'train/grad_norm_pre_clip': 10.930497169494629, 'timing/data': 0.0008359737694263458, 'timing/model': 0.9648125562816858, 'learning_rate/qwen_vl_interface': 9.00323630878453e-06, 'learning_rate/action_model': 9.00323630878453e-05, 'epoch': 0.64}) +[2026-06-17 19:11:34,696][starVLA.training.train_starvla][INFO] - Step 1550, Loss: {'train/loss': 0.545878529548645, 'train/grad_norm_pre_clip': 19.574146270751953, 'timing/data': 0.0006622755900025368, 'timing/model': 1.1945166075602174, 'learning_rate/qwen_vl_interface': 9.001910391484903e-06, 'learning_rate/action_model': 9.001910391484904e-05, 'epoch': 0.64}) +[2026-06-17 19:11:38,615][starVLA.training.train_starvla][INFO] - Step 1551, Loss: {'train/loss': 0.6290659308433533, 'train/grad_norm_pre_clip': 13.22944164276123, 'timing/data': 0.0005001416429877281, 'timing/model': 1.161314164288342, 'learning_rate/qwen_vl_interface': 9.000583696410216e-06, 'learning_rate/action_model': 9.000583696410217e-05, 'epoch': 0.64}) +[2026-06-17 19:11:42,390][starVLA.training.train_starvla][INFO] - Step 1552, Loss: {'train/loss': 0.6687746047973633, 'train/grad_norm_pre_clip': 10.298008918762207, 'timing/data': 0.0005465038120746613, 'timing/model': 1.0093294708058238, 'learning_rate/qwen_vl_interface': 8.999256223835493e-06, 'learning_rate/action_model': 8.999256223835494e-05, 'epoch': 0.64}) +[2026-06-17 19:11:46,321][starVLA.training.train_starvla][INFO] - Step 1553, Loss: {'train/loss': 0.541429340839386, 'train/grad_norm_pre_clip': 8.739230155944824, 'timing/data': 0.0005205795168876648, 'timing/model': 0.9625080460682511, 'learning_rate/qwen_vl_interface': 8.997927974035922e-06, 'learning_rate/action_model': 8.997927974035921e-05, 'epoch': 0.64}) +[2026-06-17 19:11:50,338][starVLA.training.train_starvla][INFO] - Step 1554, Loss: {'train/loss': 0.603867769241333, 'train/grad_norm_pre_clip': 14.466336250305176, 'timing/data': 0.000523839145898819, 'timing/model': 0.9714621528983116, 'learning_rate/qwen_vl_interface': 8.996598947286846e-06, 'learning_rate/action_model': 8.996598947286846e-05, 'epoch': 0.64}) +[2026-06-17 19:11:54,330][starVLA.training.train_starvla][INFO] - Step 1555, Loss: {'train/loss': 0.8189467191696167, 'train/grad_norm_pre_clip': 13.998644828796387, 'timing/data': 0.0009100371971726418, 'timing/model': 0.9669707017019391, 'learning_rate/qwen_vl_interface': 8.99526914386378e-06, 'learning_rate/action_model': 8.995269143863779e-05, 'epoch': 0.64}) +[2026-06-17 19:11:58,365][starVLA.training.train_starvla][INFO] - Step 1556, Loss: {'train/loss': 0.5374833941459656, 'train/grad_norm_pre_clip': 8.805309295654297, 'timing/data': 0.0007295887917280197, 'timing/model': 0.9359099827706814, 'learning_rate/qwen_vl_interface': 8.993938564042386e-06, 'learning_rate/action_model': 8.993938564042387e-05, 'epoch': 0.64}) +[2026-06-17 19:12:02,297][starVLA.training.train_starvla][INFO] - Step 1557, Loss: {'train/loss': 0.47909632325172424, 'train/grad_norm_pre_clip': 43.601680755615234, 'timing/data': 0.0011330517008900642, 'timing/model': 1.1610104590654373, 'learning_rate/qwen_vl_interface': 8.992607208098502e-06, 'learning_rate/action_model': 8.992607208098503e-05, 'epoch': 0.64}) +[2026-06-17 19:12:06,177][starVLA.training.train_starvla][INFO] - Step 1558, Loss: {'train/loss': 0.6909852623939514, 'train/grad_norm_pre_clip': 8.703974723815918, 'timing/data': 0.0009382534772157669, 'timing/model': 1.0838606301695108, 'learning_rate/qwen_vl_interface': 8.991275076308117e-06, 'learning_rate/action_model': 8.991275076308117e-05, 'epoch': 0.64}) +[2026-06-17 19:12:10,008][starVLA.training.train_starvla][INFO] - Step 1559, Loss: {'train/loss': 0.7302660346031189, 'train/grad_norm_pre_clip': 10.810724258422852, 'timing/data': 0.0007539317011833191, 'timing/model': 0.9641488650813699, 'learning_rate/qwen_vl_interface': 8.989942168947382e-06, 'learning_rate/action_model': 8.989942168947381e-05, 'epoch': 0.64}) +[2026-06-17 19:12:14,037][starVLA.training.train_starvla][INFO] - Step 1560, Loss: {'train/loss': 0.7190282940864563, 'train/grad_norm_pre_clip': 11.352404594421387, 'timing/data': 0.0006145983934402466, 'timing/model': 0.9960334282368422, 'learning_rate/qwen_vl_interface': 8.98860848629261e-06, 'learning_rate/action_model': 8.988608486292611e-05, 'epoch': 0.64}) +[2026-06-17 19:12:17,977][starVLA.training.train_starvla][INFO] - Step 1561, Loss: {'train/loss': 0.47896796464920044, 'train/grad_norm_pre_clip': 8.398728370666504, 'timing/data': 0.0005037179216742516, 'timing/model': 0.9491564929485321, 'learning_rate/qwen_vl_interface': 8.987274028620276e-06, 'learning_rate/action_model': 8.987274028620276e-05, 'epoch': 0.64}) +[2026-06-17 19:12:21,847][starVLA.training.train_starvla][INFO] - Step 1562, Loss: {'train/loss': 0.5536047220230103, 'train/grad_norm_pre_clip': 9.470458984375, 'timing/data': 0.0007291007786989212, 'timing/model': 0.9416808243840933, 'learning_rate/qwen_vl_interface': 8.985938796207017e-06, 'learning_rate/action_model': 8.985938796207017e-05, 'epoch': 0.64}) +[2026-06-17 19:12:25,832][starVLA.training.train_starvla][INFO] - Step 1563, Loss: {'train/loss': 0.7838853597640991, 'train/grad_norm_pre_clip': 12.150803565979004, 'timing/data': 0.0011557331308722496, 'timing/model': 1.1243590647354722, 'learning_rate/qwen_vl_interface': 8.984602789329623e-06, 'learning_rate/action_model': 8.984602789329623e-05, 'epoch': 0.64}) +[2026-06-17 19:12:30,008][starVLA.training.train_starvla][INFO] - Step 1564, Loss: {'train/loss': 0.621741771697998, 'train/grad_norm_pre_clip': 11.87023639678955, 'timing/data': 0.0007534055039286613, 'timing/model': 1.315953348763287, 'learning_rate/qwen_vl_interface': 8.983266008265052e-06, 'learning_rate/action_model': 8.983266008265052e-05, 'epoch': 0.64}) +[2026-06-17 19:12:33,843][starVLA.training.train_starvla][INFO] - Step 1565, Loss: {'train/loss': 0.45097148418426514, 'train/grad_norm_pre_clip': 9.510014533996582, 'timing/data': 0.0005693864077329636, 'timing/model': 0.9372410569339991, 'learning_rate/qwen_vl_interface': 8.981928453290421e-06, 'learning_rate/action_model': 8.981928453290421e-05, 'epoch': 0.64}) +[2026-06-17 19:12:37,738][starVLA.training.train_starvla][INFO] - Step 1566, Loss: {'train/loss': 0.3973439335823059, 'train/grad_norm_pre_clip': 8.122639656066895, 'timing/data': 0.0014046523720026016, 'timing/model': 0.9506311165168881, 'learning_rate/qwen_vl_interface': 8.980590124683007e-06, 'learning_rate/action_model': 8.980590124683007e-05, 'epoch': 0.64}) +[2026-06-17 19:12:41,646][starVLA.training.train_starvla][INFO] - Step 1567, Loss: {'train/loss': 0.530698835849762, 'train/grad_norm_pre_clip': 11.855259895324707, 'timing/data': 0.0006391294300556183, 'timing/model': 0.9799714023247361, 'learning_rate/qwen_vl_interface': 8.979251022720244e-06, 'learning_rate/action_model': 8.979251022720244e-05, 'epoch': 0.65}) +[2026-06-17 19:12:45,662][starVLA.training.train_starvla][INFO] - Step 1568, Loss: {'train/loss': 0.6962104439735413, 'train/grad_norm_pre_clip': 10.418434143066406, 'timing/data': 0.0008113328367471695, 'timing/model': 1.0023296307772398, 'learning_rate/qwen_vl_interface': 8.977911147679733e-06, 'learning_rate/action_model': 8.977911147679733e-05, 'epoch': 0.65}) +[2026-06-17 19:12:49,541][starVLA.training.train_starvla][INFO] - Step 1569, Loss: {'train/loss': 0.4412676990032196, 'train/grad_norm_pre_clip': 18.573673248291016, 'timing/data': 0.0009805066511034966, 'timing/model': 0.9428412578999996, 'learning_rate/qwen_vl_interface': 8.976570499839229e-06, 'learning_rate/action_model': 8.976570499839228e-05, 'epoch': 0.65}) +[2026-06-17 19:12:53,393][starVLA.training.train_starvla][INFO] - Step 1570, Loss: {'train/loss': 0.4989651143550873, 'train/grad_norm_pre_clip': 8.628622055053711, 'timing/data': 0.0008065802976489067, 'timing/model': 1.1257612314075232, 'learning_rate/qwen_vl_interface': 8.97522907947665e-06, 'learning_rate/action_model': 8.975229079476649e-05, 'epoch': 0.65}) +[2026-06-17 19:12:57,032][starVLA.training.train_starvla][INFO] - Step 1571, Loss: {'train/loss': 0.687054455280304, 'train/grad_norm_pre_clip': 8.630147933959961, 'timing/data': 0.0004842085763812065, 'timing/model': 0.9532150039449334, 'learning_rate/qwen_vl_interface': 8.973886886870075e-06, 'learning_rate/action_model': 8.973886886870074e-05, 'epoch': 0.65}) +[2026-06-17 19:13:01,059][starVLA.training.train_starvla][INFO] - Step 1572, Loss: {'train/loss': 0.5170512795448303, 'train/grad_norm_pre_clip': 11.486190795898438, 'timing/data': 0.0005532121285796165, 'timing/model': 0.9337447248399258, 'learning_rate/qwen_vl_interface': 8.972543922297739e-06, 'learning_rate/action_model': 8.972543922297739e-05, 'epoch': 0.65}) +[2026-06-17 19:13:05,990][starVLA.training.train_starvla][INFO] - Step 1573, Loss: {'train/loss': 0.786960244178772, 'train/grad_norm_pre_clip': 11.067895889282227, 'timing/data': 0.0013881837949156761, 'timing/model': 0.947541868314147, 'learning_rate/qwen_vl_interface': 8.971200186038046e-06, 'learning_rate/action_model': 8.971200186038045e-05, 'epoch': 0.65}) +[2026-06-17 19:13:09,794][starVLA.training.train_starvla][INFO] - Step 1574, Loss: {'train/loss': 0.6258918046951294, 'train/grad_norm_pre_clip': 11.922722816467285, 'timing/data': 0.0008352939039468765, 'timing/model': 0.9446219922974706, 'learning_rate/qwen_vl_interface': 8.969855678369547e-06, 'learning_rate/action_model': 8.969855678369546e-05, 'epoch': 0.65}) +[2026-06-17 19:13:13,726][starVLA.training.train_starvla][INFO] - Step 1575, Loss: {'train/loss': 0.5337695479393005, 'train/grad_norm_pre_clip': 8.939452171325684, 'timing/data': 0.0004889843985438347, 'timing/model': 0.9586722608655691, 'learning_rate/qwen_vl_interface': 8.968510399570964e-06, 'learning_rate/action_model': 8.968510399570965e-05, 'epoch': 0.65}) +[2026-06-17 19:13:18,131][starVLA.training.train_starvla][INFO] - Step 1576, Loss: {'train/loss': 0.4778076708316803, 'train/grad_norm_pre_clip': 7.630599021911621, 'timing/data': 0.0005434807389974594, 'timing/model': 1.1861981274560094, 'learning_rate/qwen_vl_interface': 8.967164349921173e-06, 'learning_rate/action_model': 8.967164349921174e-05, 'epoch': 0.65}) +[2026-06-17 19:13:21,892][starVLA.training.train_starvla][INFO] - Step 1577, Loss: {'train/loss': 0.8908430337905884, 'train/grad_norm_pre_clip': 22.833402633666992, 'timing/data': 0.0008770423009991646, 'timing/model': 0.9667554628103971, 'learning_rate/qwen_vl_interface': 8.965817529699216e-06, 'learning_rate/action_model': 8.965817529699215e-05, 'epoch': 0.65}) +[2026-06-17 19:13:25,910][starVLA.training.train_starvla][INFO] - Step 1578, Loss: {'train/loss': 0.5751798748970032, 'train/grad_norm_pre_clip': 10.390827178955078, 'timing/data': 0.0005930326879024506, 'timing/model': 1.0323142986744642, 'learning_rate/qwen_vl_interface': 8.964469939184283e-06, 'learning_rate/action_model': 8.964469939184282e-05, 'epoch': 0.65}) +[2026-06-17 19:13:29,937][starVLA.training.train_starvla][INFO] - Step 1579, Loss: {'train/loss': 0.5918391346931458, 'train/grad_norm_pre_clip': 11.6248197555542, 'timing/data': 0.000479002483189106, 'timing/model': 0.9535363782197237, 'learning_rate/qwen_vl_interface': 8.963121578655736e-06, 'learning_rate/action_model': 8.963121578655736e-05, 'epoch': 0.65}) +[2026-06-17 19:13:34,267][starVLA.training.train_starvla][INFO] - Step 1580, Loss: {'train/loss': 0.589713454246521, 'train/grad_norm_pre_clip': 39.67389678955078, 'timing/data': 0.0007118098437786102, 'timing/model': 0.9449506932869554, 'learning_rate/qwen_vl_interface': 8.96177244839309e-06, 'learning_rate/action_model': 8.961772448393092e-05, 'epoch': 0.65}) +[2026-06-17 19:13:38,207][starVLA.training.train_starvla][INFO] - Step 1581, Loss: {'train/loss': 0.49331390857696533, 'train/grad_norm_pre_clip': 10.429227828979492, 'timing/data': 0.0006415573880076408, 'timing/model': 0.963192630559206, 'learning_rate/qwen_vl_interface': 8.960422548676024e-06, 'learning_rate/action_model': 8.960422548676022e-05, 'epoch': 0.65}) +[2026-06-17 19:13:42,289][starVLA.training.train_starvla][INFO] - Step 1582, Loss: {'train/loss': 0.45786142349243164, 'train/grad_norm_pre_clip': 9.068881034851074, 'timing/data': 0.0004769768565893173, 'timing/model': 0.943442402407527, 'learning_rate/qwen_vl_interface': 8.959071879784368e-06, 'learning_rate/action_model': 8.959071879784368e-05, 'epoch': 0.65}) +[2026-06-17 19:13:46,137][starVLA.training.train_starvla][INFO] - Step 1583, Loss: {'train/loss': 0.23366022109985352, 'train/grad_norm_pre_clip': 12.25415325164795, 'timing/data': 0.000851532444357872, 'timing/model': 1.151634563691914, 'learning_rate/qwen_vl_interface': 8.95772044199812e-06, 'learning_rate/action_model': 8.95772044199812e-05, 'epoch': 0.65}) +[2026-06-17 19:13:49,832][starVLA.training.train_starvla][INFO] - Step 1584, Loss: {'train/loss': 0.6716287136077881, 'train/grad_norm_pre_clip': 11.735980987548828, 'timing/data': 0.0011981390416622162, 'timing/model': 0.9383313255384564, 'learning_rate/qwen_vl_interface': 8.956368235597437e-06, 'learning_rate/action_model': 8.956368235597436e-05, 'epoch': 0.65}) +[2026-06-17 19:13:55,103][starVLA.training.train_starvla][INFO] - Step 1585, Loss: {'train/loss': 0.4994489550590515, 'train/grad_norm_pre_clip': 7.734918594360352, 'timing/data': 0.0008488325402140617, 'timing/model': 0.9539311593398452, 'learning_rate/qwen_vl_interface': 8.955015260862627e-06, 'learning_rate/action_model': 8.955015260862627e-05, 'epoch': 0.65}) +[2026-06-17 19:13:59,343][starVLA.training.train_starvla][INFO] - Step 1586, Loss: {'train/loss': 0.586571216583252, 'train/grad_norm_pre_clip': 9.464544296264648, 'timing/data': 0.0004882225766777992, 'timing/model': 1.0911451010033488, 'learning_rate/qwen_vl_interface': 8.95366151807417e-06, 'learning_rate/action_model': 8.95366151807417e-05, 'epoch': 0.65}) +[2026-06-17 19:14:03,342][starVLA.training.train_starvla][INFO] - Step 1587, Loss: {'train/loss': 0.6660024523735046, 'train/grad_norm_pre_clip': 11.378561973571777, 'timing/data': 0.0008807415142655373, 'timing/model': 0.9834816930815578, 'learning_rate/qwen_vl_interface': 8.952307007512694e-06, 'learning_rate/action_model': 8.952307007512694e-05, 'epoch': 0.65}) +[2026-06-17 19:14:07,326][starVLA.training.train_starvla][INFO] - Step 1588, Loss: {'train/loss': 0.5402671098709106, 'train/grad_norm_pre_clip': 9.14306354522705, 'timing/data': 0.0009595528244972229, 'timing/model': 0.984781657345593, 'learning_rate/qwen_vl_interface': 8.950951729458991e-06, 'learning_rate/action_model': 8.950951729458991e-05, 'epoch': 0.65}) +[2026-06-17 19:14:11,534][starVLA.training.train_starvla][INFO] - Step 1589, Loss: {'train/loss': 0.3818879723548889, 'train/grad_norm_pre_clip': 7.884503364562988, 'timing/data': 0.000529065728187561, 'timing/model': 1.254469407722354, 'learning_rate/qwen_vl_interface': 8.949595684194012e-06, 'learning_rate/action_model': 8.949595684194012e-05, 'epoch': 0.65}) +[2026-06-17 19:14:15,667][starVLA.training.train_starvla][INFO] - Step 1590, Loss: {'train/loss': 0.6662383675575256, 'train/grad_norm_pre_clip': 12.53140926361084, 'timing/data': 0.0006679166108369827, 'timing/model': 1.2057300927117467, 'learning_rate/qwen_vl_interface': 8.948238871998866e-06, 'learning_rate/action_model': 8.948238871998866e-05, 'epoch': 0.65}) +[2026-06-17 19:14:19,602][starVLA.training.train_starvla][INFO] - Step 1591, Loss: {'train/loss': 0.9678633809089661, 'train/grad_norm_pre_clip': 12.756752014160156, 'timing/data': 0.0005108555778861046, 'timing/model': 0.96365718357265, 'learning_rate/qwen_vl_interface': 8.946881293154825e-06, 'learning_rate/action_model': 8.946881293154825e-05, 'epoch': 0.65}) +[2026-06-17 19:14:23,512][starVLA.training.train_starvla][INFO] - Step 1592, Loss: {'train/loss': 0.8332377076148987, 'train/grad_norm_pre_clip': 11.567639350891113, 'timing/data': 0.001256641000509262, 'timing/model': 0.9619442075490952, 'learning_rate/qwen_vl_interface': 8.945522947943312e-06, 'learning_rate/action_model': 8.945522947943313e-05, 'epoch': 0.66}) +[2026-06-17 19:14:27,453][starVLA.training.train_starvla][INFO] - Step 1593, Loss: {'train/loss': 0.6223781108856201, 'train/grad_norm_pre_clip': 18.240028381347656, 'timing/data': 0.0005375118926167488, 'timing/model': 0.9430262390524149, 'learning_rate/qwen_vl_interface': 8.944163836645916e-06, 'learning_rate/action_model': 8.944163836645915e-05, 'epoch': 0.66}) +[2026-06-17 19:14:31,482][starVLA.training.train_starvla][INFO] - Step 1594, Loss: {'train/loss': 0.6089549660682678, 'train/grad_norm_pre_clip': 13.133329391479492, 'timing/data': 0.0006479229778051376, 'timing/model': 1.0053310943767428, 'learning_rate/qwen_vl_interface': 8.942803959544381e-06, 'learning_rate/action_model': 8.942803959544381e-05, 'epoch': 0.66}) +[2026-06-17 19:14:35,342][starVLA.training.train_starvla][INFO] - Step 1595, Loss: {'train/loss': 0.6453714966773987, 'train/grad_norm_pre_clip': 10.503386497497559, 'timing/data': 0.0008464939892292023, 'timing/model': 0.942884580232203, 'learning_rate/qwen_vl_interface': 8.941443316920613e-06, 'learning_rate/action_model': 8.941443316920612e-05, 'epoch': 0.66}) +[2026-06-17 19:14:39,323][starVLA.training.train_starvla][INFO] - Step 1596, Loss: {'train/loss': 0.3768588602542877, 'train/grad_norm_pre_clip': 7.021100044250488, 'timing/data': 0.0008851420134305954, 'timing/model': 1.1864336133003235, 'learning_rate/qwen_vl_interface': 8.940081909056671e-06, 'learning_rate/action_model': 8.940081909056671e-05, 'epoch': 0.66}) +[2026-06-17 19:14:43,159][starVLA.training.train_starvla][INFO] - Step 1597, Loss: {'train/loss': 0.6182616949081421, 'train/grad_norm_pre_clip': 21.823640823364258, 'timing/data': 0.0008060773834586143, 'timing/model': 1.1033346559852362, 'learning_rate/qwen_vl_interface': 8.93871973623478e-06, 'learning_rate/action_model': 8.93871973623478e-05, 'epoch': 0.66}) +[2026-06-17 19:14:48,939][starVLA.training.train_starvla][INFO] - Step 1598, Loss: {'train/loss': 0.5760987997055054, 'train/grad_norm_pre_clip': 10.996634483337402, 'timing/data': 0.0032214736565947533, 'timing/model': 2.9347566412761807, 'learning_rate/qwen_vl_interface': 8.937356798737317e-06, 'learning_rate/action_model': 8.937356798737317e-05, 'epoch': 0.66}) +[2026-06-17 19:14:52,896][starVLA.training.train_starvla][INFO] - Step 1599, Loss: {'train/loss': 0.4565242528915405, 'train/grad_norm_pre_clip': 12.081815719604492, 'timing/data': 0.0012359898537397385, 'timing/model': 0.9631489720195532, 'learning_rate/qwen_vl_interface': 8.935993096846825e-06, 'learning_rate/action_model': 8.935993096846825e-05, 'epoch': 0.66}) +[2026-06-17 19:17:44,352][starVLA.training.train_starvla][INFO] - Step 1600, Loss: {'train/loss': 0.5683773159980774, 'train/grad_norm_pre_clip': 8.842771530151367, 'eval/loss': 0.6514930725097656, 'eval/latency_6/loss': 0.6514930725097656, 'eval/demon_fire_precision': 0.958922558922559, 'eval/demon_fire_recall': 0.9204912734324499, 'eval/demon_fire_f1': 0.9393139841688654, 'eval/demon_fire_rate_model': 0.8263772954924875, 'eval/demon_fire_rate_teacher': 0.860879243183083, 'eval/demon_left_precision': 0.7902869757174393, 'eval/demon_left_recall': 0.6280701754385964, 'eval/demon_left_f1': 0.699902248289345, 'eval/demon_left_rate_model': 0.12604340567612687, 'eval/demon_left_rate_teacher': 0.15859766277128548, 'eval/demon_right_precision': 0.6901408450704225, 'eval/demon_right_recall': 0.7526881720430108, 'eval/demon_right_f1': 0.7200587803085966, 'eval/demon_right_rate_model': 0.19755147468002227, 'eval/demon_right_rate_teacher': 0.18113522537562604, 'eval/demon_move_f1': 0.7099805142989708, 'eval/demon_cc_f1': 0.8563318910111105, 'eval/latency_6/demon_fire_precision': 0.958922558922559, 'eval/latency_6/demon_fire_recall': 0.9204912734324499, 'eval/latency_6/demon_fire_f1': 0.9393139841688654, 'eval/latency_6/demon_fire_rate_model': 0.8263772954924875, 'eval/latency_6/demon_fire_rate_teacher': 0.860879243183083, 'eval/latency_6/demon_left_precision': 0.7902869757174393, 'eval/latency_6/demon_left_recall': 0.6280701754385964, 'eval/latency_6/demon_left_f1': 0.699902248289345, 'eval/latency_6/demon_left_rate_model': 0.12604340567612687, 'eval/latency_6/demon_left_rate_teacher': 0.15859766277128548, 'eval/latency_6/demon_right_precision': 0.6901408450704225, 'eval/latency_6/demon_right_recall': 0.7526881720430108, 'eval/latency_6/demon_right_f1': 0.7200587803085966, 'eval/latency_6/demon_right_rate_model': 0.19755147468002227, 'eval/latency_6/demon_right_rate_teacher': 0.18113522537562604, 'eval/latency_6/demon_move_f1': 0.7099805142989708, 'eval/latency_6/demon_cc_f1': 0.8563318910111105, 'timing/data': 0.0007271263748407364, 'timing/model': 2.08794107940048, 'learning_rate/qwen_vl_interface': 8.934628630845994e-06, 'learning_rate/action_model': 8.934628630845995e-05, 'epoch': 0.66}) +[2026-06-17 19:17:44,354][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_1600_state +[2026-06-17 19:17:44,359][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 19:17:51,241][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_1600_state/model.safetensors +[2026-06-17 19:18:07,173][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_1600_state/optimizer.bin +[2026-06-17 19:18:07,175][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_1600_state/sampler.bin +[2026-06-17 19:18:07,175][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_1600_state/sampler_1.bin +[2026-06-17 19:18:07,178][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_1600_state/random_states_0.pkl +[2026-06-17 19:18:07,182][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1600 +[2026-06-17 19:18:07,182][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 19:18:07,195][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 19:18:12,714][starVLA.training.train_starvla][INFO] - Step 1601, Loss: {'train/loss': 0.6333987712860107, 'train/grad_norm_pre_clip': 9.148033142089844, 'timing/data': 0.0005491608753800392, 'timing/model': 0.9589145798236132, 'learning_rate/qwen_vl_interface': 8.933263401017685e-06, 'learning_rate/action_model': 8.933263401017684e-05, 'epoch': 0.66}) +[2026-06-17 19:18:16,705][starVLA.training.train_starvla][INFO] - Step 1602, Loss: {'train/loss': 0.559747576713562, 'train/grad_norm_pre_clip': 10.177517890930176, 'timing/data': 0.0009136730805039406, 'timing/model': 0.9774089837446809, 'learning_rate/qwen_vl_interface': 8.93189740764491e-06, 'learning_rate/action_model': 8.931897407644909e-05, 'epoch': 0.66}) +[2026-06-17 19:18:21,076][starVLA.training.train_starvla][INFO] - Step 1603, Loss: {'train/loss': 0.722423791885376, 'train/grad_norm_pre_clip': 11.113358497619629, 'timing/data': 0.0005426490679383278, 'timing/model': 1.1579767307266593, 'learning_rate/qwen_vl_interface': 8.93053065101084e-06, 'learning_rate/action_model': 8.93053065101084e-05, 'epoch': 0.66}) +[2026-06-17 19:18:25,025][starVLA.training.train_starvla][INFO] - Step 1604, Loss: {'train/loss': 0.6215742826461792, 'train/grad_norm_pre_clip': 12.910323143005371, 'timing/data': 0.0007556183263659477, 'timing/model': 0.9402678664773703, 'learning_rate/qwen_vl_interface': 8.929163131398804e-06, 'learning_rate/action_model': 8.929163131398805e-05, 'epoch': 0.66}) +[2026-06-17 19:18:28,895][starVLA.training.train_starvla][INFO] - Step 1605, Loss: {'train/loss': 0.48341190814971924, 'train/grad_norm_pre_clip': 11.851234436035156, 'timing/data': 0.0006778538227081299, 'timing/model': 1.1625680988654494, 'learning_rate/qwen_vl_interface': 8.927794849092294e-06, 'learning_rate/action_model': 8.927794849092293e-05, 'epoch': 0.66}) +[2026-06-17 19:18:32,671][starVLA.training.train_starvla][INFO] - Step 1606, Loss: {'train/loss': 0.48844024538993835, 'train/grad_norm_pre_clip': 8.269584655761719, 'timing/data': 0.0008729705587029457, 'timing/model': 1.0197521150112152, 'learning_rate/qwen_vl_interface': 8.926425804374953e-06, 'learning_rate/action_model': 8.926425804374954e-05, 'epoch': 0.66}) +[2026-06-17 19:18:36,702][starVLA.training.train_starvla][INFO] - Step 1607, Loss: {'train/loss': 0.7564467191696167, 'train/grad_norm_pre_clip': 13.255038261413574, 'timing/data': 0.0006637899205088615, 'timing/model': 0.9565009539946914, 'learning_rate/qwen_vl_interface': 8.925055997530587e-06, 'learning_rate/action_model': 8.925055997530587e-05, 'epoch': 0.66}) +[2026-06-17 19:18:41,078][starVLA.training.train_starvla][INFO] - Step 1608, Loss: {'train/loss': 0.5607331395149231, 'train/grad_norm_pre_clip': 10.440131187438965, 'timing/data': 0.0008360948413610458, 'timing/model': 0.9611720656976104, 'learning_rate/qwen_vl_interface': 8.92368542884316e-06, 'learning_rate/action_model': 8.923685428843159e-05, 'epoch': 0.66}) +[2026-06-17 19:18:44,984][starVLA.training.train_starvla][INFO] - Step 1609, Loss: {'train/loss': 0.5282989144325256, 'train/grad_norm_pre_clip': 7.943619728088379, 'timing/data': 0.0007749376818537712, 'timing/model': 0.9558104602620006, 'learning_rate/qwen_vl_interface': 8.922314098596789e-06, 'learning_rate/action_model': 8.922314098596789e-05, 'epoch': 0.66}) +[2026-06-17 19:18:49,124][starVLA.training.train_starvla][INFO] - Step 1610, Loss: {'train/loss': 0.5824049711227417, 'train/grad_norm_pre_clip': 8.66834545135498, 'timing/data': 0.000623771920800209, 'timing/model': 1.1390532357618213, 'learning_rate/qwen_vl_interface': 8.920942007075752e-06, 'learning_rate/action_model': 8.920942007075752e-05, 'epoch': 0.66}) +[2026-06-17 19:18:53,437][starVLA.training.train_starvla][INFO] - Step 1611, Loss: {'train/loss': 0.7494796514511108, 'train/grad_norm_pre_clip': 13.149109840393066, 'timing/data': 0.00148063525557518, 'timing/model': 1.4581809025257826, 'learning_rate/qwen_vl_interface': 8.91956915456449e-06, 'learning_rate/action_model': 8.91956915456449e-05, 'epoch': 0.66}) +[2026-06-17 19:18:57,393][starVLA.training.train_starvla][INFO] - Step 1612, Loss: {'train/loss': 0.3061797618865967, 'train/grad_norm_pre_clip': 8.182170867919922, 'timing/data': 0.0006283819675445557, 'timing/model': 1.1594978040084243, 'learning_rate/qwen_vl_interface': 8.918195541347593e-06, 'learning_rate/action_model': 8.918195541347593e-05, 'epoch': 0.66}) +[2026-06-17 19:19:01,518][starVLA.training.train_starvla][INFO] - Step 1613, Loss: {'train/loss': 0.7100552916526794, 'train/grad_norm_pre_clip': 12.909364700317383, 'timing/data': 0.0010698698461055756, 'timing/model': 1.1243694052100182, 'learning_rate/qwen_vl_interface': 8.916821167709811e-06, 'learning_rate/action_model': 8.916821167709812e-05, 'epoch': 0.66}) +[2026-06-17 19:19:07,204][starVLA.training.train_starvla][INFO] - Step 1614, Loss: {'train/loss': 0.9026638269424438, 'train/grad_norm_pre_clip': 15.94509220123291, 'timing/data': 0.0014507006853818893, 'timing/model': 1.118889338336885, 'learning_rate/qwen_vl_interface': 8.915446033936057e-06, 'learning_rate/action_model': 8.915446033936056e-05, 'epoch': 0.66}) +[2026-06-17 19:19:11,250][starVLA.training.train_starvla][INFO] - Step 1615, Loss: {'train/loss': 0.6203889846801758, 'train/grad_norm_pre_clip': 12.568758010864258, 'timing/data': 0.000657067634165287, 'timing/model': 0.9775487072765827, 'learning_rate/qwen_vl_interface': 8.914070140311395e-06, 'learning_rate/action_model': 8.914070140311395e-05, 'epoch': 0.66}) +[2026-06-17 19:19:15,575][starVLA.training.train_starvla][INFO] - Step 1616, Loss: {'train/loss': 0.7998542189598083, 'train/grad_norm_pre_clip': 13.066200256347656, 'timing/data': 0.0009924713522195816, 'timing/model': 0.934281587600708, 'learning_rate/qwen_vl_interface': 8.912693487121053e-06, 'learning_rate/action_model': 8.912693487121052e-05, 'epoch': 0.67}) +[2026-06-17 19:19:19,707][starVLA.training.train_starvla][INFO] - Step 1617, Loss: {'train/loss': 0.5734326839447021, 'train/grad_norm_pre_clip': 9.989167213439941, 'timing/data': 0.0006355522200465202, 'timing/model': 1.1455622306093574, 'learning_rate/qwen_vl_interface': 8.911316074650407e-06, 'learning_rate/action_model': 8.911316074650406e-05, 'epoch': 0.67}) +[2026-06-17 19:19:23,689][starVLA.training.train_starvla][INFO] - Step 1618, Loss: {'train/loss': 0.7857862114906311, 'train/grad_norm_pre_clip': 11.7695951461792, 'timing/data': 0.0012130122631788254, 'timing/model': 1.2319249920547009, 'learning_rate/qwen_vl_interface': 8.909937903185e-06, 'learning_rate/action_model': 8.909937903185e-05, 'epoch': 0.67}) +[2026-06-17 19:19:27,752][starVLA.training.train_starvla][INFO] - Step 1619, Loss: {'train/loss': 0.8121421933174133, 'train/grad_norm_pre_clip': 11.769637107849121, 'timing/data': 0.0006661657243967056, 'timing/model': 1.2111048270016909, 'learning_rate/qwen_vl_interface': 8.908558973010528e-06, 'learning_rate/action_model': 8.908558973010526e-05, 'epoch': 0.67}) +[2026-06-17 19:19:31,827][starVLA.training.train_starvla][INFO] - Step 1620, Loss: {'train/loss': 0.5984392762184143, 'train/grad_norm_pre_clip': 10.037528991699219, 'timing/data': 0.0005700085312128067, 'timing/model': 0.9739807294681668, 'learning_rate/qwen_vl_interface': 8.907179284412841e-06, 'learning_rate/action_model': 8.90717928441284e-05, 'epoch': 0.67}) +[2026-06-17 19:19:35,914][starVLA.training.train_starvla][INFO] - Step 1621, Loss: {'train/loss': 0.5811993479728699, 'train/grad_norm_pre_clip': 15.896466255187988, 'timing/data': 0.0005623111501336098, 'timing/model': 1.0837693437933922, 'learning_rate/qwen_vl_interface': 8.905798837677956e-06, 'learning_rate/action_model': 8.905798837677956e-05, 'epoch': 0.67}) +[2026-06-17 19:19:39,800][starVLA.training.train_starvla][INFO] - Step 1622, Loss: {'train/loss': 0.7215732336044312, 'train/grad_norm_pre_clip': 9.850990295410156, 'timing/data': 0.000578257255256176, 'timing/model': 0.9539868319407105, 'learning_rate/qwen_vl_interface': 8.904417633092037e-06, 'learning_rate/action_model': 8.904417633092037e-05, 'epoch': 0.67}) +[2026-06-17 19:19:43,699][starVLA.training.train_starvla][INFO] - Step 1623, Loss: {'train/loss': 0.6635705828666687, 'train/grad_norm_pre_clip': 12.986065864562988, 'timing/data': 0.0005519939586520195, 'timing/model': 0.9444959312677383, 'learning_rate/qwen_vl_interface': 8.903035670941411e-06, 'learning_rate/action_model': 8.903035670941412e-05, 'epoch': 0.67}) +[2026-06-17 19:19:47,573][starVLA.training.train_starvla][INFO] - Step 1624, Loss: {'train/loss': 0.5431494116783142, 'train/grad_norm_pre_clip': 10.577842712402344, 'timing/data': 0.0005267718806862831, 'timing/model': 0.938434244133532, 'learning_rate/qwen_vl_interface': 8.901652951512559e-06, 'learning_rate/action_model': 8.90165295151256e-05, 'epoch': 0.67}) +[2026-06-17 19:19:54,750][starVLA.training.train_starvla][INFO] - Step 1625, Loss: {'train/loss': 0.6354718804359436, 'train/grad_norm_pre_clip': 11.04629898071289, 'timing/data': 0.001039082184433937, 'timing/model': 2.292094728909433, 'learning_rate/qwen_vl_interface': 8.900269475092122e-06, 'learning_rate/action_model': 8.900269475092122e-05, 'epoch': 0.67}) +[2026-06-17 19:19:58,914][starVLA.training.train_starvla][INFO] - Step 1626, Loss: {'train/loss': 0.7465035319328308, 'train/grad_norm_pre_clip': 12.55903434753418, 'timing/data': 0.0005762996152043343, 'timing/model': 1.0480995802208781, 'learning_rate/qwen_vl_interface': 8.898885241966894e-06, 'learning_rate/action_model': 8.898885241966893e-05, 'epoch': 0.67}) +[2026-06-17 19:20:02,553][starVLA.training.train_starvla][INFO] - Step 1627, Loss: {'train/loss': 1.130671739578247, 'train/grad_norm_pre_clip': 13.419798851013184, 'timing/data': 0.0006353119388222694, 'timing/model': 0.9459028905257583, 'learning_rate/qwen_vl_interface': 8.89750025242383e-06, 'learning_rate/action_model': 8.89750025242383e-05, 'epoch': 0.67}) +[2026-06-17 19:20:06,451][starVLA.training.train_starvla][INFO] - Step 1628, Loss: {'train/loss': 0.6086174845695496, 'train/grad_norm_pre_clip': 10.136087417602539, 'timing/data': 0.0010222308337688446, 'timing/model': 1.0056914575397968, 'learning_rate/qwen_vl_interface': 8.896114506750037e-06, 'learning_rate/action_model': 8.896114506750036e-05, 'epoch': 0.67}) +[2026-06-17 19:20:10,332][starVLA.training.train_starvla][INFO] - Step 1629, Loss: {'train/loss': 0.4095289707183838, 'train/grad_norm_pre_clip': 8.008527755737305, 'timing/data': 0.000575006939470768, 'timing/model': 1.0025002276524901, 'learning_rate/qwen_vl_interface': 8.894728005232785e-06, 'learning_rate/action_model': 8.894728005232785e-05, 'epoch': 0.67}) +[2026-06-17 19:20:14,302][starVLA.training.train_starvla][INFO] - Step 1630, Loss: {'train/loss': 0.6234021782875061, 'train/grad_norm_pre_clip': 9.192798614501953, 'timing/data': 0.0005175657570362091, 'timing/model': 0.9509593257680535, 'learning_rate/qwen_vl_interface': 8.893340748159494e-06, 'learning_rate/action_model': 8.893340748159493e-05, 'epoch': 0.67}) +[2026-06-17 19:20:18,246][starVLA.training.train_starvla][INFO] - Step 1631, Loss: {'train/loss': 0.4660157859325409, 'train/grad_norm_pre_clip': 13.887874603271484, 'timing/data': 0.0005453042685985565, 'timing/model': 1.197288915514946, 'learning_rate/qwen_vl_interface': 8.891952735817746e-06, 'learning_rate/action_model': 8.891952735817745e-05, 'epoch': 0.67}) +[2026-06-17 19:20:22,111][starVLA.training.train_starvla][INFO] - Step 1632, Loss: {'train/loss': 0.731523871421814, 'train/grad_norm_pre_clip': 10.220077514648438, 'timing/data': 0.001123923808336258, 'timing/model': 1.1259501250460744, 'learning_rate/qwen_vl_interface': 8.890563968495278e-06, 'learning_rate/action_model': 8.890563968495278e-05, 'epoch': 0.67}) +[2026-06-17 19:20:25,824][starVLA.training.train_starvla][INFO] - Step 1633, Loss: {'train/loss': 0.516324520111084, 'train/grad_norm_pre_clip': 9.758252143859863, 'timing/data': 0.0005291858687996864, 'timing/model': 0.9513313099741936, 'learning_rate/qwen_vl_interface': 8.88917444647998e-06, 'learning_rate/action_model': 8.889174446479981e-05, 'epoch': 0.67}) +[2026-06-17 19:20:29,611][starVLA.training.train_starvla][INFO] - Step 1634, Loss: {'train/loss': 0.6558523774147034, 'train/grad_norm_pre_clip': 9.834399223327637, 'timing/data': 0.0004910649731755257, 'timing/model': 0.9397144932299852, 'learning_rate/qwen_vl_interface': 8.887784170059905e-06, 'learning_rate/action_model': 8.887784170059905e-05, 'epoch': 0.67}) +[2026-06-17 19:20:33,482][starVLA.training.train_starvla][INFO] - Step 1635, Loss: {'train/loss': 0.6900612115859985, 'train/grad_norm_pre_clip': 11.652608871459961, 'timing/data': 0.0010330434888601303, 'timing/model': 0.9533152393996716, 'learning_rate/qwen_vl_interface': 8.886393139523258e-06, 'learning_rate/action_model': 8.886393139523257e-05, 'epoch': 0.67}) +[2026-06-17 19:20:37,407][starVLA.training.train_starvla][INFO] - Step 1636, Loss: {'train/loss': 0.6188147664070129, 'train/grad_norm_pre_clip': 13.373762130737305, 'timing/data': 0.001610761508345604, 'timing/model': 0.9628640422597528, 'learning_rate/qwen_vl_interface': 8.885001355158397e-06, 'learning_rate/action_model': 8.885001355158397e-05, 'epoch': 0.67}) +[2026-06-17 19:20:41,237][starVLA.training.train_starvla][INFO] - Step 1637, Loss: {'train/loss': 0.8113424181938171, 'train/grad_norm_pre_clip': 11.57373332977295, 'timing/data': 0.0007481556385755539, 'timing/model': 0.9713808055967093, 'learning_rate/qwen_vl_interface': 8.883608817253847e-06, 'learning_rate/action_model': 8.883608817253847e-05, 'epoch': 0.67}) +[2026-06-17 19:20:45,224][starVLA.training.train_starvla][INFO] - Step 1638, Loss: {'train/loss': 0.7925341129302979, 'train/grad_norm_pre_clip': 11.893383026123047, 'timing/data': 0.0008946629241108894, 'timing/model': 1.2393976021558046, 'learning_rate/qwen_vl_interface': 8.882215526098276e-06, 'learning_rate/action_model': 8.882215526098277e-05, 'epoch': 0.67}) +[2026-06-17 19:20:48,981][starVLA.training.train_starvla][INFO] - Step 1639, Loss: {'train/loss': 0.5361899137496948, 'train/grad_norm_pre_clip': 11.758101463317871, 'timing/data': 0.0005707377567887306, 'timing/model': 0.9537008432671428, 'learning_rate/qwen_vl_interface': 8.880821481980521e-06, 'learning_rate/action_model': 8.880821481980521e-05, 'epoch': 0.67}) +[2026-06-17 19:20:53,148][starVLA.training.train_starvla][INFO] - Step 1640, Loss: {'train/loss': 0.706890344619751, 'train/grad_norm_pre_clip': 18.45624351501465, 'timing/data': 0.0005269292742013931, 'timing/model': 0.9639022639021277, 'learning_rate/qwen_vl_interface': 8.879426685189565e-06, 'learning_rate/action_model': 8.879426685189565e-05, 'epoch': 0.68}) +[2026-06-17 19:20:57,409][starVLA.training.train_starvla][INFO] - Step 1641, Loss: {'train/loss': 0.7685720324516296, 'train/grad_norm_pre_clip': 10.055863380432129, 'timing/data': 0.0010230336338281631, 'timing/model': 1.0404577543959022, 'learning_rate/qwen_vl_interface': 8.878031136014552e-06, 'learning_rate/action_model': 8.878031136014551e-05, 'epoch': 0.68}) +[2026-06-17 19:21:01,386][starVLA.training.train_starvla][INFO] - Step 1642, Loss: {'train/loss': 0.8167762756347656, 'train/grad_norm_pre_clip': 12.533158302307129, 'timing/data': 0.0005023637786507607, 'timing/model': 0.9629210522398353, 'learning_rate/qwen_vl_interface': 8.87663483474478e-06, 'learning_rate/action_model': 8.87663483474478e-05, 'epoch': 0.68}) +[2026-06-17 19:21:05,420][starVLA.training.train_starvla][INFO] - Step 1643, Loss: {'train/loss': 0.49802064895629883, 'train/grad_norm_pre_clip': 9.193289756774902, 'timing/data': 0.000496281310915947, 'timing/model': 0.9671843629330397, 'learning_rate/qwen_vl_interface': 8.875237781669704e-06, 'learning_rate/action_model': 8.875237781669703e-05, 'epoch': 0.68}) +[2026-06-17 19:21:09,748][starVLA.training.train_starvla][INFO] - Step 1644, Loss: {'train/loss': 0.8704679012298584, 'train/grad_norm_pre_clip': 13.191558837890625, 'timing/data': 0.0005097473040223122, 'timing/model': 1.1519840862601995, 'learning_rate/qwen_vl_interface': 8.873839977078933e-06, 'learning_rate/action_model': 8.873839977078934e-05, 'epoch': 0.68}) +[2026-06-17 19:21:13,595][starVLA.training.train_starvla][INFO] - Step 1645, Loss: {'train/loss': 0.6238470673561096, 'train/grad_norm_pre_clip': 9.317464828491211, 'timing/data': 0.0013527357950806618, 'timing/model': 0.9977224143221974, 'learning_rate/qwen_vl_interface': 8.87244142126224e-06, 'learning_rate/action_model': 8.87244142126224e-05, 'epoch': 0.68}) +[2026-06-17 19:21:17,638][starVLA.training.train_starvla][INFO] - Step 1646, Loss: {'train/loss': 0.6429418921470642, 'train/grad_norm_pre_clip': 8.747343063354492, 'timing/data': 0.000878477469086647, 'timing/model': 0.943349719978869, 'learning_rate/qwen_vl_interface': 8.871042114509539e-06, 'learning_rate/action_model': 8.87104211450954e-05, 'epoch': 0.68}) +[2026-06-17 19:21:21,605][starVLA.training.train_starvla][INFO] - Step 1647, Loss: {'train/loss': 0.6555504202842712, 'train/grad_norm_pre_clip': 10.040092468261719, 'timing/data': 0.0005406197160482407, 'timing/model': 0.9630985539406538, 'learning_rate/qwen_vl_interface': 8.869642057110912e-06, 'learning_rate/action_model': 8.869642057110911e-05, 'epoch': 0.68}) +[2026-06-17 19:21:25,726][starVLA.training.train_starvla][INFO] - Step 1648, Loss: {'train/loss': 0.556115448474884, 'train/grad_norm_pre_clip': 10.464061737060547, 'timing/data': 0.0004951180890202522, 'timing/model': 1.046823744662106, 'learning_rate/qwen_vl_interface': 8.868241249356592e-06, 'learning_rate/action_model': 8.868241249356592e-05, 'epoch': 0.68}) +[2026-06-17 19:21:29,596][starVLA.training.train_starvla][INFO] - Step 1649, Loss: {'train/loss': 0.774713397026062, 'train/grad_norm_pre_clip': 10.332837104797363, 'timing/data': 0.0005248039960861206, 'timing/model': 0.9423549985513091, 'learning_rate/qwen_vl_interface': 8.866839691536968e-06, 'learning_rate/action_model': 8.866839691536968e-05, 'epoch': 0.68}) +[2026-06-17 19:21:34,554][starVLA.training.train_starvla][INFO] - Step 1650, Loss: {'train/loss': 0.3938777446746826, 'train/grad_norm_pre_clip': 8.543992042541504, 'timing/data': 0.0005531469359993935, 'timing/model': 0.9894629875198007, 'learning_rate/qwen_vl_interface': 8.865437383942584e-06, 'learning_rate/action_model': 8.865437383942583e-05, 'epoch': 0.68}) +[2026-06-17 19:21:38,598][starVLA.training.train_starvla][INFO] - Step 1651, Loss: {'train/loss': 0.6447564363479614, 'train/grad_norm_pre_clip': 9.298537254333496, 'timing/data': 0.0007809381932020187, 'timing/model': 1.212660820223391, 'learning_rate/qwen_vl_interface': 8.86403432686414e-06, 'learning_rate/action_model': 8.86403432686414e-05, 'epoch': 0.68}) +[2026-06-17 19:21:42,309][starVLA.training.train_starvla][INFO] - Step 1652, Loss: {'train/loss': 0.39029884338378906, 'train/grad_norm_pre_clip': 7.896063327789307, 'timing/data': 0.0005371058359742165, 'timing/model': 1.024364646524191, 'learning_rate/qwen_vl_interface': 8.86263052059249e-06, 'learning_rate/action_model': 8.862630520592489e-05, 'epoch': 0.68}) +[2026-06-17 19:21:46,395][starVLA.training.train_starvla][INFO] - Step 1653, Loss: {'train/loss': 0.8969573378562927, 'train/grad_norm_pre_clip': 12.987720489501953, 'timing/data': 0.0010655475780367851, 'timing/model': 1.02864926494658, 'learning_rate/qwen_vl_interface': 8.861225965418647e-06, 'learning_rate/action_model': 8.861225965418646e-05, 'epoch': 0.68}) +[2026-06-17 19:21:50,381][starVLA.training.train_starvla][INFO] - Step 1654, Loss: {'train/loss': 0.7771687507629395, 'train/grad_norm_pre_clip': 10.372279167175293, 'timing/data': 0.0005394183099269867, 'timing/model': 1.0076239118352532, 'learning_rate/qwen_vl_interface': 8.859820661633775e-06, 'learning_rate/action_model': 8.859820661633775e-05, 'epoch': 0.68}) +[2026-06-17 19:21:54,307][starVLA.training.train_starvla][INFO] - Step 1655, Loss: {'train/loss': 0.5092846751213074, 'train/grad_norm_pre_clip': 10.576094627380371, 'timing/data': 0.0009579872712492943, 'timing/model': 0.968735946342349, 'learning_rate/qwen_vl_interface': 8.858414609529193e-06, 'learning_rate/action_model': 8.858414609529193e-05, 'epoch': 0.68}) +[2026-06-17 19:21:58,512][starVLA.training.train_starvla][INFO] - Step 1656, Loss: {'train/loss': 0.5961869359016418, 'train/grad_norm_pre_clip': 11.595610618591309, 'timing/data': 0.000605415552854538, 'timing/model': 1.0547104701399803, 'learning_rate/qwen_vl_interface': 8.857007809396384e-06, 'learning_rate/action_model': 8.857007809396383e-05, 'epoch': 0.68}) +[2026-06-17 19:22:02,599][starVLA.training.train_starvla][INFO] - Step 1657, Loss: {'train/loss': 0.5854626297950745, 'train/grad_norm_pre_clip': 10.000513076782227, 'timing/data': 0.0011451756581664085, 'timing/model': 1.298383011482656, 'learning_rate/qwen_vl_interface': 8.855600261526973e-06, 'learning_rate/action_model': 8.855600261526973e-05, 'epoch': 0.68}) +[2026-06-17 19:22:06,550][starVLA.training.train_starvla][INFO] - Step 1658, Loss: {'train/loss': 0.3516375720500946, 'train/grad_norm_pre_clip': 7.192662239074707, 'timing/data': 0.000742611475288868, 'timing/model': 1.2082038186490536, 'learning_rate/qwen_vl_interface': 8.854191966212747e-06, 'learning_rate/action_model': 8.854191966212748e-05, 'epoch': 0.68}) +[2026-06-17 19:22:10,306][starVLA.training.train_starvla][INFO] - Step 1659, Loss: {'train/loss': 0.47967344522476196, 'train/grad_norm_pre_clip': 12.398906707763672, 'timing/data': 0.0009066173806786537, 'timing/model': 0.9691797383129597, 'learning_rate/qwen_vl_interface': 8.852782923745648e-06, 'learning_rate/action_model': 8.852782923745648e-05, 'epoch': 0.68}) +[2026-06-17 19:22:14,284][starVLA.training.train_starvla][INFO] - Step 1660, Loss: {'train/loss': 0.12985540926456451, 'train/grad_norm_pre_clip': 5.276652812957764, 'timing/data': 0.0007181642577052116, 'timing/model': 0.9589345175772905, 'learning_rate/qwen_vl_interface': 8.851373134417773e-06, 'learning_rate/action_model': 8.851373134417773e-05, 'epoch': 0.68}) +[2026-06-17 19:22:18,401][starVLA.training.train_starvla][INFO] - Step 1661, Loss: {'train/loss': 0.852924644947052, 'train/grad_norm_pre_clip': 12.90205192565918, 'timing/data': 0.0005631493404507637, 'timing/model': 0.956744534894824, 'learning_rate/qwen_vl_interface': 8.849962598521373e-06, 'learning_rate/action_model': 8.849962598521373e-05, 'epoch': 0.68}) +[2026-06-17 19:22:22,462][starVLA.training.train_starvla][INFO] - Step 1662, Loss: {'train/loss': 0.4597530961036682, 'train/grad_norm_pre_clip': 18.13365936279297, 'timing/data': 0.0009506382048130035, 'timing/model': 0.9689207151532173, 'learning_rate/qwen_vl_interface': 8.848551316348852e-06, 'learning_rate/action_model': 8.848551316348853e-05, 'epoch': 0.68}) +[2026-06-17 19:22:26,519][starVLA.training.train_starvla][INFO] - Step 1663, Loss: {'train/loss': 0.37078389525413513, 'train/grad_norm_pre_clip': 12.139734268188477, 'timing/data': 0.0007858583703637123, 'timing/model': 1.058003113605082, 'learning_rate/qwen_vl_interface': 8.847139288192771e-06, 'learning_rate/action_model': 8.84713928819277e-05, 'epoch': 0.68}) +[2026-06-17 19:22:30,709][starVLA.training.train_starvla][INFO] - Step 1664, Loss: {'train/loss': 0.5783408880233765, 'train/grad_norm_pre_clip': 15.843703269958496, 'timing/data': 0.0009453194215893745, 'timing/model': 1.3943997453898191, 'learning_rate/qwen_vl_interface': 8.845726514345845e-06, 'learning_rate/action_model': 8.845726514345843e-05, 'epoch': 0.68}) +[2026-06-17 19:22:34,682][starVLA.training.train_starvla][INFO] - Step 1665, Loss: {'train/loss': 0.7794926762580872, 'train/grad_norm_pre_clip': 11.19849967956543, 'timing/data': 0.001266689971089363, 'timing/model': 1.1946834465488791, 'learning_rate/qwen_vl_interface': 8.844312995100943e-06, 'learning_rate/action_model': 8.844312995100942e-05, 'epoch': 0.69}) +[2026-06-17 19:22:38,471][starVLA.training.train_starvla][INFO] - Step 1666, Loss: {'train/loss': 0.5521708130836487, 'train/grad_norm_pre_clip': 13.134349822998047, 'timing/data': 0.0009624240919947624, 'timing/model': 1.0107186995446682, 'learning_rate/qwen_vl_interface': 8.842898730751088e-06, 'learning_rate/action_model': 8.842898730751087e-05, 'epoch': 0.69}) +[2026-06-17 19:22:43,546][starVLA.training.train_starvla][INFO] - Step 1667, Loss: {'train/loss': 0.8476728796958923, 'train/grad_norm_pre_clip': 15.083807945251465, 'timing/data': 0.0017656255513429642, 'timing/model': 0.9462818875908852, 'learning_rate/qwen_vl_interface': 8.841483721589462e-06, 'learning_rate/action_model': 8.84148372158946e-05, 'epoch': 0.69}) +[2026-06-17 19:22:47,635][starVLA.training.train_starvla][INFO] - Step 1668, Loss: {'train/loss': 0.6192731857299805, 'train/grad_norm_pre_clip': 16.338350296020508, 'timing/data': 0.0007853321731090546, 'timing/model': 0.9781436678022146, 'learning_rate/qwen_vl_interface': 8.840067967909396e-06, 'learning_rate/action_model': 8.840067967909394e-05, 'epoch': 0.69}) +[2026-06-17 19:22:52,273][starVLA.training.train_starvla][INFO] - Step 1669, Loss: {'train/loss': 0.7936965823173523, 'train/grad_norm_pre_clip': 17.4001407623291, 'timing/data': 0.0006608311086893082, 'timing/model': 1.0675310706719756, 'learning_rate/qwen_vl_interface': 8.838651470004376e-06, 'learning_rate/action_model': 8.838651470004375e-05, 'epoch': 0.69}) +[2026-06-17 19:22:57,706][starVLA.training.train_starvla][INFO] - Step 1670, Loss: {'train/loss': 0.5502663850784302, 'train/grad_norm_pre_clip': 12.280149459838867, 'timing/data': 0.0005451133474707603, 'timing/model': 2.482165362685919, 'learning_rate/qwen_vl_interface': 8.837234228168045e-06, 'learning_rate/action_model': 8.837234228168044e-05, 'epoch': 0.69}) +[2026-06-17 19:23:01,784][starVLA.training.train_starvla][INFO] - Step 1671, Loss: {'train/loss': 0.4109917879104614, 'train/grad_norm_pre_clip': 8.27920150756836, 'timing/data': 0.0009193122386932373, 'timing/model': 1.3072605542838573, 'learning_rate/qwen_vl_interface': 8.835816242694197e-06, 'learning_rate/action_model': 8.835816242694197e-05, 'epoch': 0.69}) +[2026-06-17 19:23:05,729][starVLA.training.train_starvla][INFO] - Step 1672, Loss: {'train/loss': 0.7163332104682922, 'train/grad_norm_pre_clip': 14.725603103637695, 'timing/data': 0.0007917061448097229, 'timing/model': 1.0382373314350843, 'learning_rate/qwen_vl_interface': 8.834397513876786e-06, 'learning_rate/action_model': 8.834397513876785e-05, 'epoch': 0.69}) +[2026-06-17 19:23:09,817][starVLA.training.train_starvla][INFO] - Step 1673, Loss: {'train/loss': 0.45723670721054077, 'train/grad_norm_pre_clip': 16.264751434326172, 'timing/data': 0.0005481112748384476, 'timing/model': 0.9614359205588698, 'learning_rate/qwen_vl_interface': 8.832978042009913e-06, 'learning_rate/action_model': 8.832978042009912e-05, 'epoch': 0.69}) +[2026-06-17 19:23:15,025][starVLA.training.train_starvla][INFO] - Step 1674, Loss: {'train/loss': 0.8604686260223389, 'train/grad_norm_pre_clip': 14.065269470214844, 'timing/data': 0.0005275579169392586, 'timing/model': 0.9512024382129312, 'learning_rate/qwen_vl_interface': 8.831557827387835e-06, 'learning_rate/action_model': 8.831557827387834e-05, 'epoch': 0.69}) +[2026-06-17 19:23:19,138][starVLA.training.train_starvla][INFO] - Step 1675, Loss: {'train/loss': 0.4896203875541687, 'train/grad_norm_pre_clip': 11.1544828414917, 'timing/data': 0.0012409016489982605, 'timing/model': 0.9657680951058865, 'learning_rate/qwen_vl_interface': 8.830136870304966e-06, 'learning_rate/action_model': 8.830136870304965e-05, 'epoch': 0.69}) +[2026-06-17 19:23:23,224][starVLA.training.train_starvla][INFO] - Step 1676, Loss: {'train/loss': 0.8152465224266052, 'train/grad_norm_pre_clip': 11.874577522277832, 'timing/data': 0.001033250242471695, 'timing/model': 1.0245460597798228, 'learning_rate/qwen_vl_interface': 8.828715171055871e-06, 'learning_rate/action_model': 8.828715171055871e-05, 'epoch': 0.69}) +[2026-06-17 19:23:27,121][starVLA.training.train_starvla][INFO] - Step 1677, Loss: {'train/loss': 0.5961477756500244, 'train/grad_norm_pre_clip': 11.684294700622559, 'timing/data': 0.0007122727110981941, 'timing/model': 1.124737374484539, 'learning_rate/qwen_vl_interface': 8.827292729935271e-06, 'learning_rate/action_model': 8.827292729935271e-05, 'epoch': 0.69}) +[2026-06-17 19:23:30,950][starVLA.training.train_starvla][INFO] - Step 1678, Loss: {'train/loss': 0.6357775330543518, 'train/grad_norm_pre_clip': 14.15324592590332, 'timing/data': 0.0005600433796644211, 'timing/model': 1.111757175065577, 'learning_rate/qwen_vl_interface': 8.825869547238038e-06, 'learning_rate/action_model': 8.825869547238038e-05, 'epoch': 0.69}) +[2026-06-17 19:23:34,639][starVLA.training.train_starvla][INFO] - Step 1679, Loss: {'train/loss': 0.6993447542190552, 'train/grad_norm_pre_clip': 12.000691413879395, 'timing/data': 0.0006269244477152824, 'timing/model': 0.9523172238841653, 'learning_rate/qwen_vl_interface': 8.824445623259202e-06, 'learning_rate/action_model': 8.8244456232592e-05, 'epoch': 0.69}) +[2026-06-17 19:23:38,512][starVLA.training.train_starvla][INFO] - Step 1680, Loss: {'train/loss': 0.36792322993278503, 'train/grad_norm_pre_clip': 9.378403663635254, 'timing/data': 0.0005246251821517944, 'timing/model': 0.9473116984590888, 'learning_rate/qwen_vl_interface': 8.82302095829394e-06, 'learning_rate/action_model': 8.82302095829394e-05, 'epoch': 0.69}) +[2026-06-17 19:23:42,349][starVLA.training.train_starvla][INFO] - Step 1681, Loss: {'train/loss': 0.5299404263496399, 'train/grad_norm_pre_clip': 9.238757133483887, 'timing/data': 0.0005769096314907074, 'timing/model': 0.9383738050237298, 'learning_rate/qwen_vl_interface': 8.82159555263759e-06, 'learning_rate/action_model': 8.821595552637589e-05, 'epoch': 0.69}) +[2026-06-17 19:23:46,412][starVLA.training.train_starvla][INFO] - Step 1682, Loss: {'train/loss': 0.5711009502410889, 'train/grad_norm_pre_clip': 10.835756301879883, 'timing/data': 0.0011663036420941353, 'timing/model': 0.9878060109913349, 'learning_rate/qwen_vl_interface': 8.820169406585638e-06, 'learning_rate/action_model': 8.820169406585637e-05, 'epoch': 0.69}) +[2026-06-17 19:23:50,353][starVLA.training.train_starvla][INFO] - Step 1683, Loss: {'train/loss': 0.4302619397640228, 'train/grad_norm_pre_clip': 12.491751670837402, 'timing/data': 0.0008455999195575714, 'timing/model': 0.9646193413063884, 'learning_rate/qwen_vl_interface': 8.818742520433727e-06, 'learning_rate/action_model': 8.818742520433726e-05, 'epoch': 0.69}) +[2026-06-17 19:23:54,228][starVLA.training.train_starvla][INFO] - Step 1684, Loss: {'train/loss': 0.7182151675224304, 'train/grad_norm_pre_clip': 12.148731231689453, 'timing/data': 0.0009355293586850166, 'timing/model': 1.1290593845769763, 'learning_rate/qwen_vl_interface': 8.81731489447765e-06, 'learning_rate/action_model': 8.81731489447765e-05, 'epoch': 0.69}) +[2026-06-17 19:23:58,222][starVLA.training.train_starvla][INFO] - Step 1685, Loss: {'train/loss': 0.4548836946487427, 'train/grad_norm_pre_clip': 8.183228492736816, 'timing/data': 0.0015366105362772942, 'timing/model': 1.181147588416934, 'learning_rate/qwen_vl_interface': 8.815886529013356e-06, 'learning_rate/action_model': 8.815886529013356e-05, 'epoch': 0.69}) +[2026-06-17 19:24:01,981][starVLA.training.train_starvla][INFO] - Step 1686, Loss: {'train/loss': 0.5110263824462891, 'train/grad_norm_pre_clip': 9.566771507263184, 'timing/data': 0.0008464409038424492, 'timing/model': 0.9741143938153982, 'learning_rate/qwen_vl_interface': 8.814457424336951e-06, 'learning_rate/action_model': 8.81445742433695e-05, 'epoch': 0.69}) +[2026-06-17 19:24:05,946][starVLA.training.train_starvla][INFO] - Step 1687, Loss: {'train/loss': 0.5929707288742065, 'train/grad_norm_pre_clip': 9.027861595153809, 'timing/data': 0.0014122547581791878, 'timing/model': 0.9725710237398744, 'learning_rate/qwen_vl_interface': 8.813027580744685e-06, 'learning_rate/action_model': 8.813027580744684e-05, 'epoch': 0.69}) +[2026-06-17 19:24:10,110][starVLA.training.train_starvla][INFO] - Step 1688, Loss: {'train/loss': 0.45314812660217285, 'train/grad_norm_pre_clip': 8.182645797729492, 'timing/data': 0.0008876165375113487, 'timing/model': 1.0314241480082273, 'learning_rate/qwen_vl_interface': 8.811596998532965e-06, 'learning_rate/action_model': 8.811596998532965e-05, 'epoch': 0.69}) +[2026-06-17 19:24:14,079][starVLA.training.train_starvla][INFO] - Step 1689, Loss: {'train/loss': 0.46546122431755066, 'train/grad_norm_pre_clip': 8.458457946777344, 'timing/data': 0.001064320094883442, 'timing/model': 0.9699959130957723, 'learning_rate/qwen_vl_interface': 8.810165677998356e-06, 'learning_rate/action_model': 8.810165677998356e-05, 'epoch': 0.7}) +[2026-06-17 19:24:18,658][starVLA.training.train_starvla][INFO] - Step 1690, Loss: {'train/loss': 0.7513409852981567, 'train/grad_norm_pre_clip': 10.497875213623047, 'timing/data': 0.0009196214377880096, 'timing/model': 0.9500641161575913, 'learning_rate/qwen_vl_interface': 8.808733619437572e-06, 'learning_rate/action_model': 8.808733619437572e-05, 'epoch': 0.7}) +[2026-06-17 19:24:25,845][starVLA.training.train_starvla][INFO] - Step 1691, Loss: {'train/loss': 0.7986249923706055, 'train/grad_norm_pre_clip': 13.169517517089844, 'timing/data': 0.000557248480618, 'timing/model': 2.864118093624711, 'learning_rate/qwen_vl_interface': 8.80730082314748e-06, 'learning_rate/action_model': 8.80730082314748e-05, 'epoch': 0.7}) +[2026-06-17 19:24:29,758][starVLA.training.train_starvla][INFO] - Step 1692, Loss: {'train/loss': 0.5895224809646606, 'train/grad_norm_pre_clip': 10.743836402893066, 'timing/data': 0.000734223984181881, 'timing/model': 1.1245362414047122, 'learning_rate/qwen_vl_interface': 8.805867289425098e-06, 'learning_rate/action_model': 8.805867289425098e-05, 'epoch': 0.7}) +[2026-06-17 19:24:33,463][starVLA.training.train_starvla][INFO] - Step 1693, Loss: {'train/loss': 0.39654287695884705, 'train/grad_norm_pre_clip': 11.207548141479492, 'timing/data': 0.0011853883042931557, 'timing/model': 0.9676156984642148, 'learning_rate/qwen_vl_interface': 8.804433018567601e-06, 'learning_rate/action_model': 8.8044330185676e-05, 'epoch': 0.7}) +[2026-06-17 19:24:37,708][starVLA.training.train_starvla][INFO] - Step 1694, Loss: {'train/loss': 0.8304067254066467, 'train/grad_norm_pre_clip': 16.05600929260254, 'timing/data': 0.0007656579837203026, 'timing/model': 1.0257027177140117, 'learning_rate/qwen_vl_interface': 8.802998010872316e-06, 'learning_rate/action_model': 8.802998010872316e-05, 'epoch': 0.7}) +[2026-06-17 19:24:41,851][starVLA.training.train_starvla][INFO] - Step 1695, Loss: {'train/loss': 0.36756613850593567, 'train/grad_norm_pre_clip': 11.065667152404785, 'timing/data': 0.0007942542433738708, 'timing/model': 0.9608798064291477, 'learning_rate/qwen_vl_interface': 8.801562266636719e-06, 'learning_rate/action_model': 8.801562266636719e-05, 'epoch': 0.7}) +[2026-06-17 19:24:46,171][starVLA.training.train_starvla][INFO] - Step 1696, Loss: {'train/loss': 0.4090323746204376, 'train/grad_norm_pre_clip': 13.752008438110352, 'timing/data': 0.0006742803379893303, 'timing/model': 1.0013947999104857, 'learning_rate/qwen_vl_interface': 8.800125786158442e-06, 'learning_rate/action_model': 8.800125786158442e-05, 'epoch': 0.7}) +[2026-06-17 19:24:50,129][starVLA.training.train_starvla][INFO] - Step 1697, Loss: {'train/loss': 0.3289681673049927, 'train/grad_norm_pre_clip': 8.041406631469727, 'timing/data': 0.0009080478921532631, 'timing/model': 1.2174592232331634, 'learning_rate/qwen_vl_interface': 8.798688569735268e-06, 'learning_rate/action_model': 8.798688569735268e-05, 'epoch': 0.7}) +[2026-06-17 19:24:54,233][starVLA.training.train_starvla][INFO] - Step 1698, Loss: {'train/loss': 0.7998164892196655, 'train/grad_norm_pre_clip': 10.761259078979492, 'timing/data': 0.0013837525621056557, 'timing/model': 1.241280305199325, 'learning_rate/qwen_vl_interface': 8.797250617665139e-06, 'learning_rate/action_model': 8.797250617665138e-05, 'epoch': 0.7}) +[2026-06-17 19:24:58,033][starVLA.training.train_starvla][INFO] - Step 1699, Loss: {'train/loss': 0.46329912543296814, 'train/grad_norm_pre_clip': 9.595909118652344, 'timing/data': 0.0013112640008330345, 'timing/model': 1.0205206014215946, 'learning_rate/qwen_vl_interface': 8.795811930246136e-06, 'learning_rate/action_model': 8.795811930246136e-05, 'epoch': 0.7}) +[2026-06-17 19:27:21,520][starVLA.training.train_starvla][INFO] - Step 1700, Loss: {'train/loss': 0.8730661869049072, 'train/grad_norm_pre_clip': 17.41976547241211, 'eval/loss': 0.6108598113059998, 'eval/latency_6/loss': 0.6108598113059998, 'eval/demon_fire_precision': 0.9366175086288046, 'eval/demon_fire_recall': 0.964770523594053, 'eval/demon_fire_f1': 0.950485591466327, 'eval/demon_fire_rate_model': 0.8867557039510295, 'eval/demon_fire_rate_teacher': 0.860879243183083, 'eval/demon_left_precision': 0.7242888402625821, 'eval/demon_left_recall': 0.5807017543859649, 'eval/demon_left_f1': 0.6445959104186954, 'eval/demon_left_rate_model': 0.12715637173066222, 'eval/demon_left_rate_teacher': 0.15859766277128548, 'eval/demon_right_precision': 0.727447216890595, 'eval/demon_right_recall': 0.5821812596006144, 'eval/demon_right_f1': 0.6467576791808873, 'eval/demon_right_rate_model': 0.1449638286032276, 'eval/demon_right_rate_teacher': 0.18113522537562604, 'eval/demon_move_f1': 0.6456767947997913, 'eval/demon_cc_f1': 0.8325740199763655, 'eval/latency_6/demon_fire_precision': 0.9366175086288046, 'eval/latency_6/demon_fire_recall': 0.964770523594053, 'eval/latency_6/demon_fire_f1': 0.950485591466327, 'eval/latency_6/demon_fire_rate_model': 0.8867557039510295, 'eval/latency_6/demon_fire_rate_teacher': 0.860879243183083, 'eval/latency_6/demon_left_precision': 0.7242888402625821, 'eval/latency_6/demon_left_recall': 0.5807017543859649, 'eval/latency_6/demon_left_f1': 0.6445959104186954, 'eval/latency_6/demon_left_rate_model': 0.12715637173066222, 'eval/latency_6/demon_left_rate_teacher': 0.15859766277128548, 'eval/latency_6/demon_right_precision': 0.727447216890595, 'eval/latency_6/demon_right_recall': 0.5821812596006144, 'eval/latency_6/demon_right_f1': 0.6467576791808873, 'eval/latency_6/demon_right_rate_model': 0.1449638286032276, 'eval/latency_6/demon_right_rate_teacher': 0.18113522537562604, 'eval/latency_6/demon_move_f1': 0.6456767947997913, 'eval/latency_6/demon_cc_f1': 0.8325740199763655, 'timing/data': 0.0009415559470653534, 'timing/model': 1.0033501340076327, 'learning_rate/qwen_vl_interface': 8.794372507776508e-06, 'learning_rate/action_model': 8.794372507776508e-05, 'epoch': 0.7}) +[2026-06-17 19:27:21,528][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_1700_state +[2026-06-17 19:27:21,533][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 19:27:27,926][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_1700_state/model.safetensors +[2026-06-17 19:27:43,903][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_1700_state/optimizer.bin +[2026-06-17 19:27:43,904][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_1700_state/sampler.bin +[2026-06-17 19:27:43,906][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_1700_state/sampler_1.bin +[2026-06-17 19:27:43,908][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_1700_state/random_states_0.pkl +[2026-06-17 19:27:43,913][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1700 +[2026-06-17 19:27:43,914][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 19:27:43,925][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 19:27:47,691][starVLA.training.train_starvla][INFO] - Step 1701, Loss: {'train/loss': 0.9029438495635986, 'train/grad_norm_pre_clip': 13.345854759216309, 'timing/data': 0.0005963509902358055, 'timing/model': 0.9409022983163595, 'learning_rate/qwen_vl_interface': 8.792932350554642e-06, 'learning_rate/action_model': 8.792932350554643e-05, 'epoch': 0.7}) +[2026-06-17 19:27:51,781][starVLA.training.train_starvla][INFO] - Step 1702, Loss: {'train/loss': 0.7283211350440979, 'train/grad_norm_pre_clip': 11.7250337600708, 'timing/data': 0.0007311068475246429, 'timing/model': 1.0160375144332647, 'learning_rate/qwen_vl_interface': 8.791491458879089e-06, 'learning_rate/action_model': 8.791491458879089e-05, 'epoch': 0.7}) +[2026-06-17 19:27:55,969][starVLA.training.train_starvla][INFO] - Step 1703, Loss: {'train/loss': 0.4947676658630371, 'train/grad_norm_pre_clip': 9.96105670928955, 'timing/data': 0.0006005251780152321, 'timing/model': 0.9370935885235667, 'learning_rate/qwen_vl_interface': 8.790049833048545e-06, 'learning_rate/action_model': 8.790049833048545e-05, 'epoch': 0.7}) +[2026-06-17 19:28:00,047][starVLA.training.train_starvla][INFO] - Step 1704, Loss: {'train/loss': 0.5536010265350342, 'train/grad_norm_pre_clip': 10.260895729064941, 'timing/data': 0.0005294354632496834, 'timing/model': 1.0232328027486801, 'learning_rate/qwen_vl_interface': 8.788607473361862e-06, 'learning_rate/action_model': 8.788607473361862e-05, 'epoch': 0.7}) +[2026-06-17 19:28:03,981][starVLA.training.train_starvla][INFO] - Step 1705, Loss: {'train/loss': 0.6638315916061401, 'train/grad_norm_pre_clip': 13.680201530456543, 'timing/data': 0.0006125094369053841, 'timing/model': 0.9822730300948024, 'learning_rate/qwen_vl_interface': 8.787164380118044e-06, 'learning_rate/action_model': 8.787164380118043e-05, 'epoch': 0.7}) +[2026-06-17 19:28:08,204][starVLA.training.train_starvla][INFO] - Step 1706, Loss: {'train/loss': 0.425922155380249, 'train/grad_norm_pre_clip': 8.636820793151855, 'timing/data': 0.0009402027353644371, 'timing/model': 1.3308423962444067, 'learning_rate/qwen_vl_interface': 8.785720553616241e-06, 'learning_rate/action_model': 8.78572055361624e-05, 'epoch': 0.7}) +[2026-06-17 19:28:11,953][starVLA.training.train_starvla][INFO] - Step 1707, Loss: {'train/loss': 0.5962126851081848, 'train/grad_norm_pre_clip': 16.65205192565918, 'timing/data': 0.0008699037134647369, 'timing/model': 0.952174499630928, 'learning_rate/qwen_vl_interface': 8.784275994155762e-06, 'learning_rate/action_model': 8.784275994155762e-05, 'epoch': 0.7}) +[2026-06-17 19:28:15,919][starVLA.training.train_starvla][INFO] - Step 1708, Loss: {'train/loss': 0.8240475654602051, 'train/grad_norm_pre_clip': 22.052841186523438, 'timing/data': 0.0005296971648931503, 'timing/model': 0.9598027663305402, 'learning_rate/qwen_vl_interface': 8.782830702036067e-06, 'learning_rate/action_model': 8.782830702036067e-05, 'epoch': 0.7}) +[2026-06-17 19:28:20,169][starVLA.training.train_starvla][INFO] - Step 1709, Loss: {'train/loss': 0.40108510851860046, 'train/grad_norm_pre_clip': 11.428006172180176, 'timing/data': 0.0011735493317246437, 'timing/model': 1.1786688072606921, 'learning_rate/qwen_vl_interface': 8.781384677556767e-06, 'learning_rate/action_model': 8.781384677556767e-05, 'epoch': 0.7}) +[2026-06-17 19:28:24,825][starVLA.training.train_starvla][INFO] - Step 1710, Loss: {'train/loss': 0.6528024077415466, 'train/grad_norm_pre_clip': 13.051286697387695, 'timing/data': 0.0005223052576184273, 'timing/model': 1.0238326005637646, 'learning_rate/qwen_vl_interface': 8.779937921017624e-06, 'learning_rate/action_model': 8.779937921017623e-05, 'epoch': 0.7}) +[2026-06-17 19:28:32,610][starVLA.training.train_starvla][INFO] - Step 1711, Loss: {'train/loss': 0.7929370403289795, 'train/grad_norm_pre_clip': 15.484533309936523, 'timing/data': 0.0009172307327389717, 'timing/model': 4.090209927409887, 'learning_rate/qwen_vl_interface': 8.77849043271855e-06, 'learning_rate/action_model': 8.778490432718551e-05, 'epoch': 0.7}) +[2026-06-17 19:28:40,260][starVLA.training.train_starvla][INFO] - Step 1712, Loss: {'train/loss': 0.6403928995132446, 'train/grad_norm_pre_clip': 11.339651107788086, 'timing/data': 0.0005714185535907745, 'timing/model': 1.4510333994403481, 'learning_rate/qwen_vl_interface': 8.777042212959615e-06, 'learning_rate/action_model': 8.777042212959614e-05, 'epoch': 0.7}) +[2026-06-17 19:28:45,714][starVLA.training.train_starvla][INFO] - Step 1713, Loss: {'train/loss': 0.9338428974151611, 'train/grad_norm_pre_clip': 25.08595085144043, 'timing/data': 0.0005346247926354408, 'timing/model': 1.0817572744563222, 'learning_rate/qwen_vl_interface': 8.775593262041032e-06, 'learning_rate/action_model': 8.77559326204103e-05, 'epoch': 0.71}) +[2026-06-17 19:28:51,607][starVLA.training.train_starvla][INFO] - Step 1714, Loss: {'train/loss': 0.6124792695045471, 'train/grad_norm_pre_clip': 34.97937774658203, 'timing/data': 0.001166192814707756, 'timing/model': 1.053899115882814, 'learning_rate/qwen_vl_interface': 8.774143580263172e-06, 'learning_rate/action_model': 8.77414358026317e-05, 'epoch': 0.71}) +[2026-06-17 19:28:55,927][starVLA.training.train_starvla][INFO] - Step 1715, Loss: {'train/loss': 0.4565909802913666, 'train/grad_norm_pre_clip': 28.013370513916016, 'timing/data': 0.0012062303721904755, 'timing/model': 0.9705734802410007, 'learning_rate/qwen_vl_interface': 8.772693167926554e-06, 'learning_rate/action_model': 8.772693167926555e-05, 'epoch': 0.71}) +[2026-06-17 19:28:59,842][starVLA.training.train_starvla][INFO] - Step 1716, Loss: {'train/loss': 0.5741861462593079, 'train/grad_norm_pre_clip': 18.666284561157227, 'timing/data': 0.0009810598567128181, 'timing/model': 0.9564615106210113, 'learning_rate/qwen_vl_interface': 8.771242025331855e-06, 'learning_rate/action_model': 8.771242025331854e-05, 'epoch': 0.71}) +[2026-06-17 19:29:03,872][starVLA.training.train_starvla][INFO] - Step 1717, Loss: {'train/loss': 0.595619797706604, 'train/grad_norm_pre_clip': 16.72882843017578, 'timing/data': 0.0012646643444895744, 'timing/model': 0.9572923090308905, 'learning_rate/qwen_vl_interface': 8.769790152779893e-06, 'learning_rate/action_model': 8.769790152779892e-05, 'epoch': 0.71}) +[2026-06-17 19:29:07,836][starVLA.training.train_starvla][INFO] - Step 1718, Loss: {'train/loss': 0.6359207034111023, 'train/grad_norm_pre_clip': 24.967605590820312, 'timing/data': 0.0005155308172106743, 'timing/model': 1.1778232250362635, 'learning_rate/qwen_vl_interface': 8.768337550571644e-06, 'learning_rate/action_model': 8.768337550571645e-05, 'epoch': 0.71}) +[2026-06-17 19:29:11,781][starVLA.training.train_starvla][INFO] - Step 1719, Loss: {'train/loss': 0.5195634961128235, 'train/grad_norm_pre_clip': 11.069464683532715, 'timing/data': 0.0007699187844991684, 'timing/model': 1.1566174272447824, 'learning_rate/qwen_vl_interface': 8.766884219008238e-06, 'learning_rate/action_model': 8.766884219008238e-05, 'epoch': 0.71}) +[2026-06-17 19:29:15,902][starVLA.training.train_starvla][INFO] - Step 1720, Loss: {'train/loss': 0.5615964531898499, 'train/grad_norm_pre_clip': 18.175994873046875, 'timing/data': 0.0005339225754141808, 'timing/model': 1.255420788191259, 'learning_rate/qwen_vl_interface': 8.765430158390948e-06, 'learning_rate/action_model': 8.765430158390947e-05, 'epoch': 0.71}) +[2026-06-17 19:29:20,017][starVLA.training.train_starvla][INFO] - Step 1721, Loss: {'train/loss': 0.6513829827308655, 'train/grad_norm_pre_clip': 10.293423652648926, 'timing/data': 0.0005282722413539886, 'timing/model': 1.0015440825372934, 'learning_rate/qwen_vl_interface': 8.763975369021202e-06, 'learning_rate/action_model': 8.763975369021203e-05, 'epoch': 0.71}) +[2026-06-17 19:29:23,987][starVLA.training.train_starvla][INFO] - Step 1722, Loss: {'train/loss': 0.6844611167907715, 'train/grad_norm_pre_clip': 15.50584602355957, 'timing/data': 0.0006233435124158859, 'timing/model': 0.9955321401357651, 'learning_rate/qwen_vl_interface': 8.762519851200581e-06, 'learning_rate/action_model': 8.762519851200581e-05, 'epoch': 0.71}) +[2026-06-17 19:29:27,866][starVLA.training.train_starvla][INFO] - Step 1723, Loss: {'train/loss': 0.6137325167655945, 'train/grad_norm_pre_clip': 12.08284854888916, 'timing/data': 0.0005909837782382965, 'timing/model': 0.939694688655436, 'learning_rate/qwen_vl_interface': 8.761063605230818e-06, 'learning_rate/action_model': 8.761063605230818e-05, 'epoch': 0.71}) +[2026-06-17 19:29:32,013][starVLA.training.train_starvla][INFO] - Step 1724, Loss: {'train/loss': 0.4612378478050232, 'train/grad_norm_pre_clip': 8.048754692077637, 'timing/data': 0.0008795075118541718, 'timing/model': 0.9487859671935439, 'learning_rate/qwen_vl_interface': 8.759606631413788e-06, 'learning_rate/action_model': 8.759606631413788e-05, 'epoch': 0.71}) +[2026-06-17 19:29:35,994][starVLA.training.train_starvla][INFO] - Step 1725, Loss: {'train/loss': 0.710044264793396, 'train/grad_norm_pre_clip': 12.589848518371582, 'timing/data': 0.0005376478657126427, 'timing/model': 1.1504145013168454, 'learning_rate/qwen_vl_interface': 8.75814893005153e-06, 'learning_rate/action_model': 8.75814893005153e-05, 'epoch': 0.71}) +[2026-06-17 19:29:39,762][starVLA.training.train_starvla][INFO] - Step 1726, Loss: {'train/loss': 0.6479647755622864, 'train/grad_norm_pre_clip': 13.331792831420898, 'timing/data': 0.0005384944379329681, 'timing/model': 0.9356201235204935, 'learning_rate/qwen_vl_interface': 8.756690501446222e-06, 'learning_rate/action_model': 8.756690501446221e-05, 'epoch': 0.71}) +[2026-06-17 19:29:43,988][starVLA.training.train_starvla][INFO] - Step 1727, Loss: {'train/loss': 0.8049362897872925, 'train/grad_norm_pre_clip': 10.696428298950195, 'timing/data': 0.0011184783652424812, 'timing/model': 0.9650683989748359, 'learning_rate/qwen_vl_interface': 8.755231345900199e-06, 'learning_rate/action_model': 8.755231345900198e-05, 'epoch': 0.71}) +[2026-06-17 19:29:48,562][starVLA.training.train_starvla][INFO] - Step 1728, Loss: {'train/loss': 0.5572523474693298, 'train/grad_norm_pre_clip': 8.377543449401855, 'timing/data': 0.0005487585440278053, 'timing/model': 0.9467098535969853, 'learning_rate/qwen_vl_interface': 8.753771463715947e-06, 'learning_rate/action_model': 8.753771463715947e-05, 'epoch': 0.71}) +[2026-06-17 19:29:52,727][starVLA.training.train_starvla][INFO] - Step 1729, Loss: {'train/loss': 0.590794563293457, 'train/grad_norm_pre_clip': 8.977280616760254, 'timing/data': 0.0011594099923968315, 'timing/model': 0.9779298519715667, 'learning_rate/qwen_vl_interface': 8.7523108551961e-06, 'learning_rate/action_model': 8.7523108551961e-05, 'epoch': 0.71}) +[2026-06-17 19:29:57,687][starVLA.training.train_starvla][INFO] - Step 1730, Loss: {'train/loss': 0.45136022567749023, 'train/grad_norm_pre_clip': 9.701743125915527, 'timing/data': 0.0006690435111522675, 'timing/model': 0.9503285801038146, 'learning_rate/qwen_vl_interface': 8.750849520643445e-06, 'learning_rate/action_model': 8.750849520643444e-05, 'epoch': 0.71}) +[2026-06-17 19:30:01,815][starVLA.training.train_starvla][INFO] - Step 1731, Loss: {'train/loss': 0.3424430191516876, 'train/grad_norm_pre_clip': 11.928056716918945, 'timing/data': 0.0011356296017765999, 'timing/model': 1.3069279929623008, 'learning_rate/qwen_vl_interface': 8.749387460360917e-06, 'learning_rate/action_model': 8.749387460360916e-05, 'epoch': 0.71}) +[2026-06-17 19:30:07,972][starVLA.training.train_starvla][INFO] - Step 1732, Loss: {'train/loss': 0.49236491322517395, 'train/grad_norm_pre_clip': 8.253046989440918, 'timing/data': 0.0012552477419376373, 'timing/model': 1.2639925703406334, 'learning_rate/qwen_vl_interface': 8.747924674651602e-06, 'learning_rate/action_model': 8.747924674651602e-05, 'epoch': 0.71}) +[2026-06-17 19:30:13,287][starVLA.training.train_starvla][INFO] - Step 1733, Loss: {'train/loss': 0.5707833766937256, 'train/grad_norm_pre_clip': 11.153202056884766, 'timing/data': 0.0006633149459958076, 'timing/model': 0.9634320363402367, 'learning_rate/qwen_vl_interface': 8.74646116381874e-06, 'learning_rate/action_model': 8.74646116381874e-05, 'epoch': 0.71}) +[2026-06-17 19:30:17,663][starVLA.training.train_starvla][INFO] - Step 1734, Loss: {'train/loss': 0.7122363448143005, 'train/grad_norm_pre_clip': 21.54926300048828, 'timing/data': 0.0015546148642897606, 'timing/model': 0.9679502015933394, 'learning_rate/qwen_vl_interface': 8.744996928165714e-06, 'learning_rate/action_model': 8.744996928165714e-05, 'epoch': 0.71}) +[2026-06-17 19:30:22,048][starVLA.training.train_starvla][INFO] - Step 1735, Loss: {'train/loss': 0.6208193898200989, 'train/grad_norm_pre_clip': 10.466754913330078, 'timing/data': 0.0005354937165975571, 'timing/model': 1.086191114038229, 'learning_rate/qwen_vl_interface': 8.743531967996066e-06, 'learning_rate/action_model': 8.743531967996065e-05, 'epoch': 0.71}) +[2026-06-17 19:30:26,504][starVLA.training.train_starvla][INFO] - Step 1736, Loss: {'train/loss': 0.5765489339828491, 'train/grad_norm_pre_clip': 12.477030754089355, 'timing/data': 0.000858403742313385, 'timing/model': 1.0161805376410484, 'learning_rate/qwen_vl_interface': 8.742066283613481e-06, 'learning_rate/action_model': 8.74206628361348e-05, 'epoch': 0.71}) +[2026-06-17 19:30:30,590][starVLA.training.train_starvla][INFO] - Step 1737, Loss: {'train/loss': 0.8412196040153503, 'train/grad_norm_pre_clip': 13.317705154418945, 'timing/data': 0.0007435409352183342, 'timing/model': 0.9489051187410951, 'learning_rate/qwen_vl_interface': 8.740599875321797e-06, 'learning_rate/action_model': 8.740599875321797e-05, 'epoch': 0.72}) +[2026-06-17 19:30:35,820][starVLA.training.train_starvla][INFO] - Step 1738, Loss: {'train/loss': 0.7309727072715759, 'train/grad_norm_pre_clip': 12.06684398651123, 'timing/data': 0.0006309691816568375, 'timing/model': 2.2584792841225863, 'learning_rate/qwen_vl_interface': 8.739132743425004e-06, 'learning_rate/action_model': 8.739132743425002e-05, 'epoch': 0.72}) +[2026-06-17 19:30:39,793][starVLA.training.train_starvla][INFO] - Step 1739, Loss: {'train/loss': 0.5764902830123901, 'train/grad_norm_pre_clip': 17.162046432495117, 'timing/data': 0.0008185086771845818, 'timing/model': 1.0076665924862027, 'learning_rate/qwen_vl_interface': 8.737664888227239e-06, 'learning_rate/action_model': 8.737664888227239e-05, 'epoch': 0.72}) +[2026-06-17 19:30:45,049][starVLA.training.train_starvla][INFO] - Step 1740, Loss: {'train/loss': 0.6969308257102966, 'train/grad_norm_pre_clip': 17.419328689575195, 'timing/data': 0.0009010899811983109, 'timing/model': 1.4630062486976385, 'learning_rate/qwen_vl_interface': 8.736196310032787e-06, 'learning_rate/action_model': 8.736196310032787e-05, 'epoch': 0.72}) +[2026-06-17 19:30:50,209][starVLA.training.train_starvla][INFO] - Step 1741, Loss: {'train/loss': 0.654036819934845, 'train/grad_norm_pre_clip': 9.9993896484375, 'timing/data': 0.001326393336057663, 'timing/model': 0.9470039419829845, 'learning_rate/qwen_vl_interface': 8.734727009146092e-06, 'learning_rate/action_model': 8.734727009146092e-05, 'epoch': 0.72}) +[2026-06-17 19:30:54,146][starVLA.training.train_starvla][INFO] - Step 1742, Loss: {'train/loss': 0.4808800518512726, 'train/grad_norm_pre_clip': 11.452034950256348, 'timing/data': 0.0012647397816181183, 'timing/model': 0.9378376873210073, 'learning_rate/qwen_vl_interface': 8.733256985871737e-06, 'learning_rate/action_model': 8.733256985871737e-05, 'epoch': 0.72}) +[2026-06-17 19:30:58,130][starVLA.training.train_starvla][INFO] - Step 1743, Loss: {'train/loss': 0.5738039016723633, 'train/grad_norm_pre_clip': 9.755385398864746, 'timing/data': 0.000933157280087471, 'timing/model': 0.9721715599298477, 'learning_rate/qwen_vl_interface': 8.73178624051446e-06, 'learning_rate/action_model': 8.73178624051446e-05, 'epoch': 0.72}) +[2026-06-17 19:31:02,117][starVLA.training.train_starvla][INFO] - Step 1744, Loss: {'train/loss': 0.5078287124633789, 'train/grad_norm_pre_clip': 10.48806381225586, 'timing/data': 0.000967671163380146, 'timing/model': 0.9811171712353826, 'learning_rate/qwen_vl_interface': 8.73031477337915e-06, 'learning_rate/action_model': 8.73031477337915e-05, 'epoch': 0.72}) +[2026-06-17 19:31:06,006][starVLA.training.train_starvla][INFO] - Step 1745, Loss: {'train/loss': 0.6624869704246521, 'train/grad_norm_pre_clip': 12.729942321777344, 'timing/data': 0.0007080798968672752, 'timing/model': 1.1707363463938236, 'learning_rate/qwen_vl_interface': 8.728842584770842e-06, 'learning_rate/action_model': 8.728842584770842e-05, 'epoch': 0.72}) +[2026-06-17 19:31:09,651][starVLA.training.train_starvla][INFO] - Step 1746, Loss: {'train/loss': 0.8510351181030273, 'train/grad_norm_pre_clip': 13.588911056518555, 'timing/data': 0.0011466527357697487, 'timing/model': 0.9408469181507826, 'learning_rate/qwen_vl_interface': 8.727369674994722e-06, 'learning_rate/action_model': 8.727369674994722e-05, 'epoch': 0.72}) +[2026-06-17 19:31:13,570][starVLA.training.train_starvla][INFO] - Step 1747, Loss: {'train/loss': 0.6710626482963562, 'train/grad_norm_pre_clip': 13.713716506958008, 'timing/data': 0.0005621826276183128, 'timing/model': 0.9459021799266338, 'learning_rate/qwen_vl_interface': 8.725896044356128e-06, 'learning_rate/action_model': 8.725896044356128e-05, 'epoch': 0.72}) +[2026-06-17 19:31:18,527][starVLA.training.train_starvla][INFO] - Step 1748, Loss: {'train/loss': 0.5806523561477661, 'train/grad_norm_pre_clip': 11.526145935058594, 'timing/data': 0.0009595351293683052, 'timing/model': 1.0468312436714768, 'learning_rate/qwen_vl_interface': 8.724421693160544e-06, 'learning_rate/action_model': 8.724421693160543e-05, 'epoch': 0.72}) +[2026-06-17 19:31:23,027][starVLA.training.train_starvla][INFO] - Step 1749, Loss: {'train/loss': 0.6537315845489502, 'train/grad_norm_pre_clip': 10.483485221862793, 'timing/data': 0.0005276575684547424, 'timing/model': 0.9353422708809376, 'learning_rate/qwen_vl_interface': 8.722946621713603e-06, 'learning_rate/action_model': 8.722946621713602e-05, 'epoch': 0.72}) +[2026-06-17 19:31:26,953][starVLA.training.train_starvla][INFO] - Step 1750, Loss: {'train/loss': 0.4182570278644562, 'train/grad_norm_pre_clip': 9.919622421264648, 'timing/data': 0.000593358650803566, 'timing/model': 0.9498060559853911, 'learning_rate/qwen_vl_interface': 8.721470830321091e-06, 'learning_rate/action_model': 8.72147083032109e-05, 'epoch': 0.72}) +[2026-06-17 19:31:31,102][starVLA.training.train_starvla][INFO] - Step 1751, Loss: {'train/loss': 0.6375613808631897, 'train/grad_norm_pre_clip': 11.917402267456055, 'timing/data': 0.0005276864394545555, 'timing/model': 1.175632911734283, 'learning_rate/qwen_vl_interface': 8.719994319288939e-06, 'learning_rate/action_model': 8.71999431928894e-05, 'epoch': 0.72}) +[2026-06-17 19:31:38,973][starVLA.training.train_starvla][INFO] - Step 1752, Loss: {'train/loss': 0.43170252442359924, 'train/grad_norm_pre_clip': 8.357719421386719, 'timing/data': 0.001053420826792717, 'timing/model': 1.4378423355519772, 'learning_rate/qwen_vl_interface': 8.718517088923233e-06, 'learning_rate/action_model': 8.718517088923233e-05, 'epoch': 0.72}) +[2026-06-17 19:31:43,080][starVLA.training.train_starvla][INFO] - Step 1753, Loss: {'train/loss': 0.8307385444641113, 'train/grad_norm_pre_clip': 15.744609832763672, 'timing/data': 0.0005350783467292786, 'timing/model': 1.025291495025158, 'learning_rate/qwen_vl_interface': 8.7170391395302e-06, 'learning_rate/action_model': 8.7170391395302e-05, 'epoch': 0.72}) +[2026-06-17 19:31:47,598][starVLA.training.train_starvla][INFO] - Step 1754, Loss: {'train/loss': 0.6154903769493103, 'train/grad_norm_pre_clip': 12.288186073303223, 'timing/data': 0.000826910138130188, 'timing/model': 1.0271446984261274, 'learning_rate/qwen_vl_interface': 8.715560471416223e-06, 'learning_rate/action_model': 8.715560471416224e-05, 'epoch': 0.72}) +[2026-06-17 19:31:51,708][starVLA.training.train_starvla][INFO] - Step 1755, Loss: {'train/loss': 1.0573139190673828, 'train/grad_norm_pre_clip': 16.366056442260742, 'timing/data': 0.0005514100193977356, 'timing/model': 0.9645548313856125, 'learning_rate/qwen_vl_interface': 8.714081084887833e-06, 'learning_rate/action_model': 8.714081084887833e-05, 'epoch': 0.72}) +[2026-06-17 19:31:55,609][starVLA.training.train_starvla][INFO] - Step 1756, Loss: {'train/loss': 0.6067619323730469, 'train/grad_norm_pre_clip': 11.555188179016113, 'timing/data': 0.0005241641774773598, 'timing/model': 0.9544015191495419, 'learning_rate/qwen_vl_interface': 8.712600980251706e-06, 'learning_rate/action_model': 8.712600980251705e-05, 'epoch': 0.72}) +[2026-06-17 19:31:59,520][starVLA.training.train_starvla][INFO] - Step 1757, Loss: {'train/loss': 1.0107793807983398, 'train/grad_norm_pre_clip': 13.43494701385498, 'timing/data': 0.0009439652785658836, 'timing/model': 0.9617877937853336, 'learning_rate/qwen_vl_interface': 8.71112015781467e-06, 'learning_rate/action_model': 8.71112015781467e-05, 'epoch': 0.72}) +[2026-06-17 19:32:03,689][starVLA.training.train_starvla][INFO] - Step 1758, Loss: {'train/loss': 0.6649699807167053, 'train/grad_norm_pre_clip': 11.94018268585205, 'timing/data': 0.001066369004547596, 'timing/model': 1.3194360081106424, 'learning_rate/qwen_vl_interface': 8.709638617883699e-06, 'learning_rate/action_model': 8.709638617883699e-05, 'epoch': 0.72}) +[2026-06-17 19:32:07,639][starVLA.training.train_starvla][INFO] - Step 1759, Loss: {'train/loss': 1.0173448324203491, 'train/grad_norm_pre_clip': 11.8495454788208, 'timing/data': 0.0008054915815591812, 'timing/model': 1.2166695101186633, 'learning_rate/qwen_vl_interface': 8.708156360765922e-06, 'learning_rate/action_model': 8.708156360765922e-05, 'epoch': 0.72}) +[2026-06-17 19:32:11,392][starVLA.training.train_starvla][INFO] - Step 1760, Loss: {'train/loss': 0.341100811958313, 'train/grad_norm_pre_clip': 7.257775783538818, 'timing/data': 0.0012352541089057922, 'timing/model': 0.9462392861023545, 'learning_rate/qwen_vl_interface': 8.706673386768609e-06, 'learning_rate/action_model': 8.706673386768609e-05, 'epoch': 0.72}) +[2026-06-17 19:32:15,598][starVLA.training.train_starvla][INFO] - Step 1761, Loss: {'train/loss': 0.4346611797809601, 'train/grad_norm_pre_clip': 8.256488800048828, 'timing/data': 0.0005449727177619934, 'timing/model': 1.1216231239959598, 'learning_rate/qwen_vl_interface': 8.705189696199184e-06, 'learning_rate/action_model': 8.705189696199183e-05, 'epoch': 0.72}) +[2026-06-17 19:32:19,610][starVLA.training.train_starvla][INFO] - Step 1762, Loss: {'train/loss': 0.6035901308059692, 'train/grad_norm_pre_clip': 11.370964050292969, 'timing/data': 0.00088474340736866, 'timing/model': 0.9509326303377748, 'learning_rate/qwen_vl_interface': 8.703705289365216e-06, 'learning_rate/action_model': 8.703705289365216e-05, 'epoch': 0.73}) +[2026-06-17 19:32:23,614][starVLA.training.train_starvla][INFO] - Step 1763, Loss: {'train/loss': 0.3883238136768341, 'train/grad_norm_pre_clip': 8.554373741149902, 'timing/data': 0.0005212659016251564, 'timing/model': 1.0479732053354383, 'learning_rate/qwen_vl_interface': 8.702220166574426e-06, 'learning_rate/action_model': 8.702220166574425e-05, 'epoch': 0.73}) +[2026-06-17 19:32:27,798][starVLA.training.train_starvla][INFO] - Step 1764, Loss: {'train/loss': 0.24063916504383087, 'train/grad_norm_pre_clip': 5.674194812774658, 'timing/data': 0.0005536507815122604, 'timing/model': 1.0227518947795033, 'learning_rate/qwen_vl_interface': 8.700734328134681e-06, 'learning_rate/action_model': 8.70073432813468e-05, 'epoch': 0.73}) +[2026-06-17 19:32:31,894][starVLA.training.train_starvla][INFO] - Step 1765, Loss: {'train/loss': 0.8659085035324097, 'train/grad_norm_pre_clip': 14.058771133422852, 'timing/data': 0.0009035486727952957, 'timing/model': 1.2011927487328649, 'learning_rate/qwen_vl_interface': 8.699247774353997e-06, 'learning_rate/action_model': 8.699247774353996e-05, 'epoch': 0.73}) +[2026-06-17 19:32:35,809][starVLA.training.train_starvla][INFO] - Step 1766, Loss: {'train/loss': 0.9373167753219604, 'train/grad_norm_pre_clip': 15.729677200317383, 'timing/data': 0.000834224745631218, 'timing/model': 1.120519726537168, 'learning_rate/qwen_vl_interface': 8.697760505540536e-06, 'learning_rate/action_model': 8.697760505540536e-05, 'epoch': 0.73}) +[2026-06-17 19:32:39,520][starVLA.training.train_starvla][INFO] - Step 1767, Loss: {'train/loss': 0.6548720002174377, 'train/grad_norm_pre_clip': 11.581978797912598, 'timing/data': 0.0005346033722162247, 'timing/model': 0.9493101369589567, 'learning_rate/qwen_vl_interface': 8.696272522002614e-06, 'learning_rate/action_model': 8.696272522002615e-05, 'epoch': 0.73}) +[2026-06-17 19:32:43,629][starVLA.training.train_starvla][INFO] - Step 1768, Loss: {'train/loss': 0.690413773059845, 'train/grad_norm_pre_clip': 9.186619758605957, 'timing/data': 0.0009142458438873291, 'timing/model': 0.9678685273975134, 'learning_rate/qwen_vl_interface': 8.69478382404869e-06, 'learning_rate/action_model': 8.69478382404869e-05, 'epoch': 0.73}) +[2026-06-17 19:32:47,746][starVLA.training.train_starvla][INFO] - Step 1769, Loss: {'train/loss': 0.27840518951416016, 'train/grad_norm_pre_clip': 6.371433258056641, 'timing/data': 0.0005325442180037498, 'timing/model': 0.9787326417863369, 'learning_rate/qwen_vl_interface': 8.693294411987371e-06, 'learning_rate/action_model': 8.693294411987371e-05, 'epoch': 0.73}) +[2026-06-17 19:32:51,794][starVLA.training.train_starvla][INFO] - Step 1770, Loss: {'train/loss': 0.6858196258544922, 'train/grad_norm_pre_clip': 11.189743041992188, 'timing/data': 0.0005203662440180779, 'timing/model': 1.0104954447597265, 'learning_rate/qwen_vl_interface': 8.691804286127416e-06, 'learning_rate/action_model': 8.691804286127415e-05, 'epoch': 0.73}) +[2026-06-17 19:32:55,790][starVLA.training.train_starvla][INFO] - Step 1771, Loss: {'train/loss': 0.5378892421722412, 'train/grad_norm_pre_clip': 9.286090850830078, 'timing/data': 0.001054692082107067, 'timing/model': 1.2321202075108886, 'learning_rate/qwen_vl_interface': 8.690313446777729e-06, 'learning_rate/action_model': 8.690313446777728e-05, 'epoch': 0.73}) +[2026-06-17 19:33:03,388][starVLA.training.train_starvla][INFO] - Step 1772, Loss: {'train/loss': 0.384568452835083, 'train/grad_norm_pre_clip': 10.168279647827148, 'timing/data': 0.0010753665119409561, 'timing/model': 2.1487550055608153, 'learning_rate/qwen_vl_interface': 8.688821894247363e-06, 'learning_rate/action_model': 8.688821894247362e-05, 'epoch': 0.73}) +[2026-06-17 19:33:12,518][starVLA.training.train_starvla][INFO] - Step 1773, Loss: {'train/loss': 0.6187180876731873, 'train/grad_norm_pre_clip': 18.700050354003906, 'timing/data': 0.0010261107236146927, 'timing/model': 3.0484387520700693, 'learning_rate/qwen_vl_interface': 8.687329628845517e-06, 'learning_rate/action_model': 8.687329628845516e-05, 'epoch': 0.73}) +[2026-06-17 19:33:16,593][starVLA.training.train_starvla][INFO] - Step 1774, Loss: {'train/loss': 1.202215313911438, 'train/grad_norm_pre_clip': 24.9735050201416, 'timing/data': 0.0005116080865263939, 'timing/model': 1.041152598336339, 'learning_rate/qwen_vl_interface': 8.68583665088154e-06, 'learning_rate/action_model': 8.68583665088154e-05, 'epoch': 0.73}) +[2026-06-17 19:33:20,518][starVLA.training.train_starvla][INFO] - Step 1775, Loss: {'train/loss': 0.5873821973800659, 'train/grad_norm_pre_clip': 20.27547836303711, 'timing/data': 0.0005533639341592789, 'timing/model': 0.9526210585609078, 'learning_rate/qwen_vl_interface': 8.684342960664928e-06, 'learning_rate/action_model': 8.684342960664928e-05, 'epoch': 0.73}) +[2026-06-17 19:33:25,234][starVLA.training.train_starvla][INFO] - Step 1776, Loss: {'train/loss': 0.41252821683883667, 'train/grad_norm_pre_clip': 9.751154899597168, 'timing/data': 0.000728449784219265, 'timing/model': 1.6461208630353212, 'learning_rate/qwen_vl_interface': 8.682848558505327e-06, 'learning_rate/action_model': 8.682848558505326e-05, 'epoch': 0.73}) +[2026-06-17 19:33:29,231][starVLA.training.train_starvla][INFO] - Step 1777, Loss: {'train/loss': 0.5573641061782837, 'train/grad_norm_pre_clip': 12.567194938659668, 'timing/data': 0.001186043955385685, 'timing/model': 0.9734833259135485, 'learning_rate/qwen_vl_interface': 8.681353444712525e-06, 'learning_rate/action_model': 8.681353444712524e-05, 'epoch': 0.73}) +[2026-06-17 19:33:33,262][starVLA.training.train_starvla][INFO] - Step 1778, Loss: {'train/loss': 0.7244036197662354, 'train/grad_norm_pre_clip': 12.762581825256348, 'timing/data': 0.0014403825625777245, 'timing/model': 1.1537993578240275, 'learning_rate/qwen_vl_interface': 8.67985761959646e-06, 'learning_rate/action_model': 8.67985761959646e-05, 'epoch': 0.73}) +[2026-06-17 19:33:36,950][starVLA.training.train_starvla][INFO] - Step 1779, Loss: {'train/loss': 0.6023474335670471, 'train/grad_norm_pre_clip': 13.000604629516602, 'timing/data': 0.0013904403895139694, 'timing/model': 0.9561901949346066, 'learning_rate/qwen_vl_interface': 8.678361083467221e-06, 'learning_rate/action_model': 8.67836108346722e-05, 'epoch': 0.73}) +[2026-06-17 19:33:41,013][starVLA.training.train_starvla][INFO] - Step 1780, Loss: {'train/loss': 0.8415546417236328, 'train/grad_norm_pre_clip': 28.173900604248047, 'timing/data': 0.0005615334957838058, 'timing/model': 0.9955354556441307, 'learning_rate/qwen_vl_interface': 8.67686383663504e-06, 'learning_rate/action_model': 8.67686383663504e-05, 'epoch': 0.73}) +[2026-06-17 19:33:44,925][starVLA.training.train_starvla][INFO] - Step 1781, Loss: {'train/loss': 0.48564955592155457, 'train/grad_norm_pre_clip': 9.953851699829102, 'timing/data': 0.0005339840427041054, 'timing/model': 0.9505778662860394, 'learning_rate/qwen_vl_interface': 8.6753658794103e-06, 'learning_rate/action_model': 8.675365879410298e-05, 'epoch': 0.73}) +[2026-06-17 19:33:48,891][starVLA.training.train_starvla][INFO] - Step 1782, Loss: {'train/loss': 0.7225555777549744, 'train/grad_norm_pre_clip': 14.446616172790527, 'timing/data': 0.0005944166332483292, 'timing/model': 0.9558860333636403, 'learning_rate/qwen_vl_interface': 8.673867212103525e-06, 'learning_rate/action_model': 8.673867212103524e-05, 'epoch': 0.73}) +[2026-06-17 19:33:52,900][starVLA.training.train_starvla][INFO] - Step 1783, Loss: {'train/loss': 0.45573002099990845, 'train/grad_norm_pre_clip': 8.215996742248535, 'timing/data': 0.0005485508590936661, 'timing/model': 0.9725719224661589, 'learning_rate/qwen_vl_interface': 8.672367835025395e-06, 'learning_rate/action_model': 8.672367835025393e-05, 'epoch': 0.73}) +[2026-06-17 19:33:56,825][starVLA.training.train_starvla][INFO] - Step 1784, Loss: {'train/loss': 0.7377732992172241, 'train/grad_norm_pre_clip': 11.900715827941895, 'timing/data': 0.0005233930423855782, 'timing/model': 1.174252918921411, 'learning_rate/qwen_vl_interface': 8.670867748486727e-06, 'learning_rate/action_model': 8.670867748486726e-05, 'epoch': 0.73}) +[2026-06-17 19:34:00,611][starVLA.training.train_starvla][INFO] - Step 1785, Loss: {'train/loss': 0.3855908215045929, 'train/grad_norm_pre_clip': 13.30240249633789, 'timing/data': 0.000599963590502739, 'timing/model': 0.9686780069023371, 'learning_rate/qwen_vl_interface': 8.669366952798494e-06, 'learning_rate/action_model': 8.669366952798494e-05, 'epoch': 0.73}) +[2026-06-17 19:34:04,601][starVLA.training.train_starvla][INFO] - Step 1786, Loss: {'train/loss': 0.7871617674827576, 'train/grad_norm_pre_clip': 17.874292373657227, 'timing/data': 0.0005492027848958969, 'timing/model': 1.0176526056602597, 'learning_rate/qwen_vl_interface': 8.667865448271814e-06, 'learning_rate/action_model': 8.667865448271814e-05, 'epoch': 0.74}) +[2026-06-17 19:34:08,592][starVLA.training.train_starvla][INFO] - Step 1787, Loss: {'train/loss': 0.5902933478355408, 'train/grad_norm_pre_clip': 10.775986671447754, 'timing/data': 0.0005154022946953773, 'timing/model': 0.9518312802538276, 'learning_rate/qwen_vl_interface': 8.666363235217949e-06, 'learning_rate/action_model': 8.66636323521795e-05, 'epoch': 0.74}) +[2026-06-17 19:34:12,666][starVLA.training.train_starvla][INFO] - Step 1788, Loss: {'train/loss': 0.506803035736084, 'train/grad_norm_pre_clip': 10.826942443847656, 'timing/data': 0.0005515636876225471, 'timing/model': 0.9987403117120266, 'learning_rate/qwen_vl_interface': 8.66486031394831e-06, 'learning_rate/action_model': 8.664860313948309e-05, 'epoch': 0.74}) +[2026-06-17 19:34:16,545][starVLA.training.train_starvla][INFO] - Step 1789, Loss: {'train/loss': 0.5756855010986328, 'train/grad_norm_pre_clip': 14.689969062805176, 'timing/data': 0.0008551785722374916, 'timing/model': 0.9582514567300677, 'learning_rate/qwen_vl_interface': 8.66335668477445e-06, 'learning_rate/action_model': 8.66335668477445e-05, 'epoch': 0.74}) +[2026-06-17 19:34:20,531][starVLA.training.train_starvla][INFO] - Step 1790, Loss: {'train/loss': 0.48663967847824097, 'train/grad_norm_pre_clip': 12.489070892333984, 'timing/data': 0.0007404694333672523, 'timing/model': 0.9624908240512013, 'learning_rate/qwen_vl_interface': 8.661852348008079e-06, 'learning_rate/action_model': 8.661852348008078e-05, 'epoch': 0.74}) +[2026-06-17 19:34:24,548][starVLA.training.train_starvla][INFO] - Step 1791, Loss: {'train/loss': 0.37989574670791626, 'train/grad_norm_pre_clip': 10.869532585144043, 'timing/data': 0.0005172966048121452, 'timing/model': 1.1384231084957719, 'learning_rate/qwen_vl_interface': 8.660347303961044e-06, 'learning_rate/action_model': 8.660347303961044e-05, 'epoch': 0.74}) +[2026-06-17 19:34:30,357][starVLA.training.train_starvla][INFO] - Step 1792, Loss: {'train/loss': 0.4679459035396576, 'train/grad_norm_pre_clip': 9.501802444458008, 'timing/data': 0.000575067475438118, 'timing/model': 0.9433357045054436, 'learning_rate/qwen_vl_interface': 8.658841552945342e-06, 'learning_rate/action_model': 8.658841552945342e-05, 'epoch': 0.74}) +[2026-06-17 19:34:34,289][starVLA.training.train_starvla][INFO] - Step 1793, Loss: {'train/loss': 0.7919272184371948, 'train/grad_norm_pre_clip': 11.067754745483398, 'timing/data': 0.0008806455880403519, 'timing/model': 0.950351252220571, 'learning_rate/qwen_vl_interface': 8.65733509527312e-06, 'learning_rate/action_model': 8.65733509527312e-05, 'epoch': 0.74}) +[2026-06-17 19:34:38,302][starVLA.training.train_starvla][INFO] - Step 1794, Loss: {'train/loss': 0.7278817892074585, 'train/grad_norm_pre_clip': 11.009638786315918, 'timing/data': 0.0007901350036263466, 'timing/model': 1.0247344989329576, 'learning_rate/qwen_vl_interface': 8.655827931256667e-06, 'learning_rate/action_model': 8.655827931256667e-05, 'epoch': 0.74}) +[2026-06-17 19:34:42,187][starVLA.training.train_starvla][INFO] - Step 1795, Loss: {'train/loss': 0.7551392316818237, 'train/grad_norm_pre_clip': 11.326560974121094, 'timing/data': 0.0004923278465867043, 'timing/model': 0.9434781363233924, 'learning_rate/qwen_vl_interface': 8.654320061208418e-06, 'learning_rate/action_model': 8.654320061208417e-05, 'epoch': 0.74}) +[2026-06-17 19:34:46,043][starVLA.training.train_starvla][INFO] - Step 1796, Loss: {'train/loss': 0.59056556224823, 'train/grad_norm_pre_clip': 9.650822639465332, 'timing/data': 0.0008172197267413139, 'timing/model': 0.9364087833091617, 'learning_rate/qwen_vl_interface': 8.652811485440956e-06, 'learning_rate/action_model': 8.652811485440956e-05, 'epoch': 0.74}) +[2026-06-17 19:34:49,998][starVLA.training.train_starvla][INFO] - Step 1797, Loss: {'train/loss': 0.397077351808548, 'train/grad_norm_pre_clip': 9.568855285644531, 'timing/data': 0.000580725260078907, 'timing/model': 1.1515766978263855, 'learning_rate/qwen_vl_interface': 8.651302204267013e-06, 'learning_rate/action_model': 8.651302204267013e-05, 'epoch': 0.74}) +[2026-06-17 19:34:54,201][starVLA.training.train_starvla][INFO] - Step 1798, Loss: {'train/loss': 0.5214667916297913, 'train/grad_norm_pre_clip': 8.420966148376465, 'timing/data': 0.0009066620841622353, 'timing/model': 1.2107059881091118, 'learning_rate/qwen_vl_interface': 8.649792217999461e-06, 'learning_rate/action_model': 8.649792217999462e-05, 'epoch': 0.74}) +[2026-06-17 19:34:57,977][starVLA.training.train_starvla][INFO] - Step 1799, Loss: {'train/loss': 0.46355384588241577, 'train/grad_norm_pre_clip': 10.40335464477539, 'timing/data': 0.0005157412961125374, 'timing/model': 0.9463529605418444, 'learning_rate/qwen_vl_interface': 8.648281526951327e-06, 'learning_rate/action_model': 8.648281526951327e-05, 'epoch': 0.74}) +[2026-06-17 19:37:40,303][starVLA.training.train_starvla][INFO] - Step 1800, Loss: {'train/loss': 0.5889568328857422, 'train/grad_norm_pre_clip': 10.845065116882324, 'eval/loss': 0.6317704319953918, 'eval/latency_6/loss': 0.6317704319953918, 'eval/demon_fire_precision': 0.9241778319123021, 'eval/demon_fire_recall': 0.9809308338720103, 'eval/demon_fire_f1': 0.9517089996864221, 'eval/demon_fire_rate_model': 0.9137451307735114, 'eval/demon_fire_rate_teacher': 0.860879243183083, 'eval/demon_left_precision': 0.875, 'eval/demon_left_recall': 0.5035087719298246, 'eval/demon_left_f1': 0.6391982182628062, 'eval/demon_left_rate_model': 0.09126321647189761, 'eval/demon_left_rate_teacher': 0.15859766277128548, 'eval/demon_right_precision': 0.868020304568528, 'eval/demon_right_recall': 0.5253456221198156, 'eval/demon_right_f1': 0.6545454545454545, 'eval/demon_right_rate_model': 0.10962715637173066, 'eval/demon_right_rate_teacher': 0.18113522537562604, 'eval/demon_move_f1': 0.6468718364041304, 'eval/demon_cc_f1': 0.8338271468350493, 'eval/latency_6/demon_fire_precision': 0.9241778319123021, 'eval/latency_6/demon_fire_recall': 0.9809308338720103, 'eval/latency_6/demon_fire_f1': 0.9517089996864221, 'eval/latency_6/demon_fire_rate_model': 0.9137451307735114, 'eval/latency_6/demon_fire_rate_teacher': 0.860879243183083, 'eval/latency_6/demon_left_precision': 0.875, 'eval/latency_6/demon_left_recall': 0.5035087719298246, 'eval/latency_6/demon_left_f1': 0.6391982182628062, 'eval/latency_6/demon_left_rate_model': 0.09126321647189761, 'eval/latency_6/demon_left_rate_teacher': 0.15859766277128548, 'eval/latency_6/demon_right_precision': 0.868020304568528, 'eval/latency_6/demon_right_recall': 0.5253456221198156, 'eval/latency_6/demon_right_f1': 0.6545454545454545, 'eval/latency_6/demon_right_rate_model': 0.10962715637173066, 'eval/latency_6/demon_right_rate_teacher': 0.18113522537562604, 'eval/latency_6/demon_move_f1': 0.6468718364041304, 'eval/latency_6/demon_cc_f1': 0.8338271468350493, 'timing/data': 0.0005107233300805092, 'timing/model': 0.9812123831361532, 'learning_rate/qwen_vl_interface': 8.646770131435774e-06, 'learning_rate/action_model': 8.646770131435773e-05, 'epoch': 0.74}) +[2026-06-17 19:37:40,307][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_1800_state +[2026-06-17 19:37:40,312][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 19:37:47,142][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_1800_state/model.safetensors +[2026-06-17 19:38:03,424][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_1800_state/optimizer.bin +[2026-06-17 19:38:03,424][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_1800_state/sampler.bin +[2026-06-17 19:38:03,424][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_1800_state/sampler_1.bin +[2026-06-17 19:38:03,426][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_1800_state/random_states_0.pkl +[2026-06-17 19:38:03,432][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1800 +[2026-06-17 19:38:03,432][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 19:38:03,442][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 19:38:07,413][starVLA.training.train_starvla][INFO] - Step 1801, Loss: {'train/loss': 0.340169757604599, 'train/grad_norm_pre_clip': 7.151381969451904, 'timing/data': 0.0006262930110096931, 'timing/model': 1.046922616660595, 'learning_rate/qwen_vl_interface': 8.645258031766116e-06, 'learning_rate/action_model': 8.645258031766116e-05, 'epoch': 0.74}) +[2026-06-17 19:38:11,488][starVLA.training.train_starvla][INFO] - Step 1802, Loss: {'train/loss': 0.6644881963729858, 'train/grad_norm_pre_clip': 11.604907035827637, 'timing/data': 0.001217159442603588, 'timing/model': 0.9463079702109098, 'learning_rate/qwen_vl_interface': 8.643745228255816e-06, 'learning_rate/action_model': 8.643745228255817e-05, 'epoch': 0.74}) +[2026-06-17 19:38:15,600][starVLA.training.train_starvla][INFO] - Step 1803, Loss: {'train/loss': 0.34989091753959656, 'train/grad_norm_pre_clip': 7.5215678215026855, 'timing/data': 0.0005384963005781174, 'timing/model': 0.9513101987540722, 'learning_rate/qwen_vl_interface': 8.64223172121848e-06, 'learning_rate/action_model': 8.642231721218479e-05, 'epoch': 0.74}) +[2026-06-17 19:38:19,635][starVLA.training.train_starvla][INFO] - Step 1804, Loss: {'train/loss': 0.8289217948913574, 'train/grad_norm_pre_clip': 10.83818244934082, 'timing/data': 0.0005567977204918861, 'timing/model': 0.9506100090220571, 'learning_rate/qwen_vl_interface': 8.640717510967857e-06, 'learning_rate/action_model': 8.640717510967856e-05, 'epoch': 0.74}) +[2026-06-17 19:38:23,697][starVLA.training.train_starvla][INFO] - Step 1805, Loss: {'train/loss': 0.9843608736991882, 'train/grad_norm_pre_clip': 14.700246810913086, 'timing/data': 0.0005687298253178596, 'timing/model': 0.9382920200005174, 'learning_rate/qwen_vl_interface': 8.639202597817844e-06, 'learning_rate/action_model': 8.639202597817844e-05, 'epoch': 0.74}) +[2026-06-17 19:38:27,713][starVLA.training.train_starvla][INFO] - Step 1806, Loss: {'train/loss': 0.5515238642692566, 'train/grad_norm_pre_clip': 11.337096214294434, 'timing/data': 0.0009769275784492493, 'timing/model': 1.1808453863486648, 'learning_rate/qwen_vl_interface': 8.637686982082487e-06, 'learning_rate/action_model': 8.637686982082487e-05, 'epoch': 0.74}) +[2026-06-17 19:38:31,893][starVLA.training.train_starvla][INFO] - Step 1807, Loss: {'train/loss': 0.8720192313194275, 'train/grad_norm_pre_clip': 12.381464004516602, 'timing/data': 0.0010059261694550514, 'timing/model': 1.1175524909049273, 'learning_rate/qwen_vl_interface': 8.636170664075972e-06, 'learning_rate/action_model': 8.636170664075972e-05, 'epoch': 0.74}) +[2026-06-17 19:38:35,883][starVLA.training.train_starvla][INFO] - Step 1808, Loss: {'train/loss': 0.7740227580070496, 'train/grad_norm_pre_clip': 14.356176376342773, 'timing/data': 0.0009921388700604439, 'timing/model': 0.9539317572489381, 'learning_rate/qwen_vl_interface': 8.634653644112636e-06, 'learning_rate/action_model': 8.634653644112635e-05, 'epoch': 0.74}) +[2026-06-17 19:38:40,673][starVLA.training.train_starvla][INFO] - Step 1809, Loss: {'train/loss': 0.4393341541290283, 'train/grad_norm_pre_clip': 11.546554565429688, 'timing/data': 0.0010371562093496323, 'timing/model': 1.0022755106911063, 'learning_rate/qwen_vl_interface': 8.633135922506957e-06, 'learning_rate/action_model': 8.633135922506957e-05, 'epoch': 0.74}) +[2026-06-17 19:38:46,055][starVLA.training.train_starvla][INFO] - Step 1810, Loss: {'train/loss': 0.6080886721611023, 'train/grad_norm_pre_clip': 14.472825050354004, 'timing/data': 0.0006899740546941757, 'timing/model': 0.954075513407588, 'learning_rate/qwen_vl_interface': 8.63161749957356e-06, 'learning_rate/action_model': 8.63161749957356e-05, 'epoch': 0.75}) +[2026-06-17 19:38:50,356][starVLA.training.train_starvla][INFO] - Step 1811, Loss: {'train/loss': 0.5389265418052673, 'train/grad_norm_pre_clip': 9.308564186096191, 'timing/data': 0.0006994195282459259, 'timing/model': 0.959057473577559, 'learning_rate/qwen_vl_interface': 8.630098375627216e-06, 'learning_rate/action_model': 8.630098375627217e-05, 'epoch': 0.75}) +[2026-06-17 19:38:55,009][starVLA.training.train_starvla][INFO] - Step 1812, Loss: {'train/loss': 0.801308810710907, 'train/grad_norm_pre_clip': 9.918447494506836, 'timing/data': 0.0007802741602063179, 'timing/model': 0.9606069093570113, 'learning_rate/qwen_vl_interface': 8.628578550982846e-06, 'learning_rate/action_model': 8.628578550982845e-05, 'epoch': 0.75}) +[2026-06-17 19:38:59,371][starVLA.training.train_starvla][INFO] - Step 1813, Loss: {'train/loss': 0.5403419733047485, 'train/grad_norm_pre_clip': 12.383915901184082, 'timing/data': 0.0015423614531755447, 'timing/model': 1.2873421590775251, 'learning_rate/qwen_vl_interface': 8.627058025955504e-06, 'learning_rate/action_model': 8.627058025955503e-05, 'epoch': 0.75}) +[2026-06-17 19:39:03,095][starVLA.training.train_starvla][INFO] - Step 1814, Loss: {'train/loss': 0.6109269261360168, 'train/grad_norm_pre_clip': 11.193787574768066, 'timing/data': 0.0010186908766627312, 'timing/model': 0.9534610826522112, 'learning_rate/qwen_vl_interface': 8.6255368008604e-06, 'learning_rate/action_model': 8.625536800860398e-05, 'epoch': 0.75}) +[2026-06-17 19:39:07,135][starVLA.training.train_starvla][INFO] - Step 1815, Loss: {'train/loss': 0.5226764678955078, 'train/grad_norm_pre_clip': 9.140393257141113, 'timing/data': 0.0006712423637509346, 'timing/model': 1.0130332559347153, 'learning_rate/qwen_vl_interface': 8.624014876012883e-06, 'learning_rate/action_model': 8.624014876012882e-05, 'epoch': 0.75}) +[2026-06-17 19:39:11,438][starVLA.training.train_starvla][INFO] - Step 1816, Loss: {'train/loss': 0.4508930742740631, 'train/grad_norm_pre_clip': 9.799336433410645, 'timing/data': 0.0005392106249928474, 'timing/model': 1.197056457400322, 'learning_rate/qwen_vl_interface': 8.622492251728453e-06, 'learning_rate/action_model': 8.622492251728452e-05, 'epoch': 0.75}) +[2026-06-17 19:39:15,821][starVLA.training.train_starvla][INFO] - Step 1817, Loss: {'train/loss': 0.6091952919960022, 'train/grad_norm_pre_clip': 11.531697273254395, 'timing/data': 0.0013272715732455254, 'timing/model': 0.9575714431703091, 'learning_rate/qwen_vl_interface': 8.62096892832275e-06, 'learning_rate/action_model': 8.620968928322749e-05, 'epoch': 0.75}) +[2026-06-17 19:39:20,283][starVLA.training.train_starvla][INFO] - Step 1818, Loss: {'train/loss': 1.0940172672271729, 'train/grad_norm_pre_clip': 13.872916221618652, 'timing/data': 0.0009868340566754341, 'timing/model': 0.9498458132147789, 'learning_rate/qwen_vl_interface': 8.619444906111562e-06, 'learning_rate/action_model': 8.619444906111561e-05, 'epoch': 0.75}) +[2026-06-17 19:39:24,275][starVLA.training.train_starvla][INFO] - Step 1819, Loss: {'train/loss': 0.5447882413864136, 'train/grad_norm_pre_clip': 13.171889305114746, 'timing/data': 0.001239168457686901, 'timing/model': 1.2237432394176722, 'learning_rate/qwen_vl_interface': 8.617920185410817e-06, 'learning_rate/action_model': 8.617920185410817e-05, 'epoch': 0.75}) +[2026-06-17 19:39:28,270][starVLA.training.train_starvla][INFO] - Step 1820, Loss: {'train/loss': 0.38741788268089294, 'train/grad_norm_pre_clip': 10.094901084899902, 'timing/data': 0.0009753573685884476, 'timing/model': 1.223570748232305, 'learning_rate/qwen_vl_interface': 8.616394766536594e-06, 'learning_rate/action_model': 8.616394766536593e-05, 'epoch': 0.75}) +[2026-06-17 19:39:31,951][starVLA.training.train_starvla][INFO] - Step 1821, Loss: {'train/loss': 0.5863426923751831, 'train/grad_norm_pre_clip': 10.685404777526855, 'timing/data': 0.0009891865774989128, 'timing/model': 0.9481985978782177, 'learning_rate/qwen_vl_interface': 8.614868649805115e-06, 'learning_rate/action_model': 8.614868649805114e-05, 'epoch': 0.75}) +[2026-06-17 19:39:36,320][starVLA.training.train_starvla][INFO] - Step 1822, Loss: {'train/loss': 0.7394651770591736, 'train/grad_norm_pre_clip': 10.617725372314453, 'timing/data': 0.000643949955701828, 'timing/model': 1.3428298784419894, 'learning_rate/qwen_vl_interface': 8.613341835532743e-06, 'learning_rate/action_model': 8.613341835532743e-05, 'epoch': 0.75}) +[2026-06-17 19:39:41,895][starVLA.training.train_starvla][INFO] - Step 1823, Loss: {'train/loss': 0.35274946689605713, 'train/grad_norm_pre_clip': 8.511880874633789, 'timing/data': 0.0013832515105605125, 'timing/model': 2.2902589701116085, 'learning_rate/qwen_vl_interface': 8.611814324035989e-06, 'learning_rate/action_model': 8.61181432403599e-05, 'epoch': 0.75}) +[2026-06-17 19:39:47,200][starVLA.training.train_starvla][INFO] - Step 1824, Loss: {'train/loss': 0.5831405520439148, 'train/grad_norm_pre_clip': 9.235325813293457, 'timing/data': 0.0006096037104725838, 'timing/model': 0.9774982845410705, 'learning_rate/qwen_vl_interface': 8.610286115631508e-06, 'learning_rate/action_model': 8.610286115631509e-05, 'epoch': 0.75}) +[2026-06-17 19:39:51,585][starVLA.training.train_starvla][INFO] - Step 1825, Loss: {'train/loss': 0.8193405866622925, 'train/grad_norm_pre_clip': 12.627816200256348, 'timing/data': 0.0012955600395798683, 'timing/model': 1.103756700642407, 'learning_rate/qwen_vl_interface': 8.608757210636101e-06, 'learning_rate/action_model': 8.608757210636101e-05, 'epoch': 0.75}) +[2026-06-17 19:39:55,798][starVLA.training.train_starvla][INFO] - Step 1826, Loss: {'train/loss': 0.38530054688453674, 'train/grad_norm_pre_clip': 10.045825958251953, 'timing/data': 0.0011149002239108086, 'timing/model': 1.2722367448732257, 'learning_rate/qwen_vl_interface': 8.60722760936671e-06, 'learning_rate/action_model': 8.60722760936671e-05, 'epoch': 0.75}) +[2026-06-17 19:39:59,658][starVLA.training.train_starvla][INFO] - Step 1827, Loss: {'train/loss': 0.4057074189186096, 'train/grad_norm_pre_clip': 11.541595458984375, 'timing/data': 0.0007078740745782852, 'timing/model': 0.970797254703939, 'learning_rate/qwen_vl_interface': 8.605697312140421e-06, 'learning_rate/action_model': 8.605697312140422e-05, 'epoch': 0.75}) +[2026-06-17 19:40:03,566][starVLA.training.train_starvla][INFO] - Step 1828, Loss: {'train/loss': 0.6023079752922058, 'train/grad_norm_pre_clip': 14.180977821350098, 'timing/data': 0.0009698318317532539, 'timing/model': 0.950156832113862, 'learning_rate/qwen_vl_interface': 8.604166319274472e-06, 'learning_rate/action_model': 8.604166319274471e-05, 'epoch': 0.75}) +[2026-06-17 19:40:07,696][starVLA.training.train_starvla][INFO] - Step 1829, Loss: {'train/loss': 0.6447035670280457, 'train/grad_norm_pre_clip': 8.936245918273926, 'timing/data': 0.0005314275622367859, 'timing/model': 1.0403301436454058, 'learning_rate/qwen_vl_interface': 8.602634631086233e-06, 'learning_rate/action_model': 8.602634631086233e-05, 'epoch': 0.75}) +[2026-06-17 19:40:11,869][starVLA.training.train_starvla][INFO] - Step 1830, Loss: {'train/loss': 0.586624026298523, 'train/grad_norm_pre_clip': 10.614653587341309, 'timing/data': 0.0005938364192843437, 'timing/model': 0.9550870731472969, 'learning_rate/qwen_vl_interface': 8.601102247893227e-06, 'learning_rate/action_model': 8.601102247893226e-05, 'epoch': 0.75}) +[2026-06-17 19:40:15,757][starVLA.training.train_starvla][INFO] - Step 1831, Loss: {'train/loss': 0.45980769395828247, 'train/grad_norm_pre_clip': 8.291875839233398, 'timing/data': 0.0005368636921048164, 'timing/model': 0.9429424442350864, 'learning_rate/qwen_vl_interface': 8.599569170013121e-06, 'learning_rate/action_model': 8.59956917001312e-05, 'epoch': 0.75}) +[2026-06-17 19:40:19,926][starVLA.training.train_starvla][INFO] - Step 1832, Loss: {'train/loss': 0.34141528606414795, 'train/grad_norm_pre_clip': 9.016270637512207, 'timing/data': 0.000655345618724823, 'timing/model': 1.0429729735478759, 'learning_rate/qwen_vl_interface': 8.59803539776372e-06, 'learning_rate/action_model': 8.59803539776372e-05, 'epoch': 0.75}) +[2026-06-17 19:40:24,046][starVLA.training.train_starvla][INFO] - Step 1833, Loss: {'train/loss': 0.3690353333950043, 'train/grad_norm_pre_clip': 12.725685119628906, 'timing/data': 0.0006068628281354904, 'timing/model': 1.2597123039886355, 'learning_rate/qwen_vl_interface': 8.596500931462979e-06, 'learning_rate/action_model': 8.596500931462979e-05, 'epoch': 0.75}) +[2026-06-17 19:40:27,808][starVLA.training.train_starvla][INFO] - Step 1834, Loss: {'train/loss': 0.42683786153793335, 'train/grad_norm_pre_clip': 9.896695137023926, 'timing/data': 0.0005469182506203651, 'timing/model': 0.9549140920862556, 'learning_rate/qwen_vl_interface': 8.594965771428993e-06, 'learning_rate/action_model': 8.594965771428993e-05, 'epoch': 0.75}) +[2026-06-17 19:40:31,766][starVLA.training.train_starvla][INFO] - Step 1835, Loss: {'train/loss': 0.6513925790786743, 'train/grad_norm_pre_clip': 13.266284942626953, 'timing/data': 0.0011249976232647896, 'timing/model': 0.966364249587059, 'learning_rate/qwen_vl_interface': 8.593429917980006e-06, 'learning_rate/action_model': 8.593429917980006e-05, 'epoch': 0.76}) +[2026-06-17 19:40:35,711][starVLA.training.train_starvla][INFO] - Step 1836, Loss: {'train/loss': 0.4620838761329651, 'train/grad_norm_pre_clip': 15.284220695495605, 'timing/data': 0.001252502202987671, 'timing/model': 0.9632843108847737, 'learning_rate/qwen_vl_interface': 8.591893371434397e-06, 'learning_rate/action_model': 8.591893371434396e-05, 'epoch': 0.76}) +[2026-06-17 19:40:39,958][starVLA.training.train_starvla][INFO] - Step 1837, Loss: {'train/loss': 0.9911330938339233, 'train/grad_norm_pre_clip': 14.812830924987793, 'timing/data': 0.0005232440307736397, 'timing/model': 0.9990770230069757, 'learning_rate/qwen_vl_interface': 8.590356132110695e-06, 'learning_rate/action_model': 8.590356132110695e-05, 'epoch': 0.76}) +[2026-06-17 19:40:43,894][starVLA.training.train_starvla][INFO] - Step 1838, Loss: {'train/loss': 0.8525612950325012, 'train/grad_norm_pre_clip': 17.82785987854004, 'timing/data': 0.0005405964329838753, 'timing/model': 0.9598686881363392, 'learning_rate/qwen_vl_interface': 8.588818200327575e-06, 'learning_rate/action_model': 8.588818200327575e-05, 'epoch': 0.76}) +[2026-06-17 19:40:47,861][starVLA.training.train_starvla][INFO] - Step 1839, Loss: {'train/loss': 0.5181607604026794, 'train/grad_norm_pre_clip': 9.172797203063965, 'timing/data': 0.0007700305432081223, 'timing/model': 0.9605535510927439, 'learning_rate/qwen_vl_interface': 8.587279576403848e-06, 'learning_rate/action_model': 8.587279576403847e-05, 'epoch': 0.76}) +[2026-06-17 19:40:51,830][starVLA.training.train_starvla][INFO] - Step 1840, Loss: {'train/loss': 0.5489553213119507, 'train/grad_norm_pre_clip': 8.75977897644043, 'timing/data': 0.0005838135257363319, 'timing/model': 1.192293118685484, 'learning_rate/qwen_vl_interface': 8.585740260658475e-06, 'learning_rate/action_model': 8.585740260658475e-05, 'epoch': 0.76}) +[2026-06-17 19:40:55,691][starVLA.training.train_starvla][INFO] - Step 1841, Loss: {'train/loss': 0.8244385123252869, 'train/grad_norm_pre_clip': 19.421955108642578, 'timing/data': 0.0005573555827140808, 'timing/model': 1.004234554246068, 'learning_rate/qwen_vl_interface': 8.584200253410557e-06, 'learning_rate/action_model': 8.584200253410556e-05, 'epoch': 0.76}) +[2026-06-17 19:41:01,626][starVLA.training.train_starvla][INFO] - Step 1842, Loss: {'train/loss': 0.4403000771999359, 'train/grad_norm_pre_clip': 11.345900535583496, 'timing/data': 0.0005415631458163261, 'timing/model': 0.9623893601819873, 'learning_rate/qwen_vl_interface': 8.582659554979336e-06, 'learning_rate/action_model': 8.582659554979336e-05, 'epoch': 0.76}) +[2026-06-17 19:41:05,714][starVLA.training.train_starvla][INFO] - Step 1843, Loss: {'train/loss': 0.503373920917511, 'train/grad_norm_pre_clip': 11.878034591674805, 'timing/data': 0.0005805017426609993, 'timing/model': 1.047957488335669, 'learning_rate/qwen_vl_interface': 8.581118165684204e-06, 'learning_rate/action_model': 8.581118165684204e-05, 'epoch': 0.76}) +[2026-06-17 19:41:09,860][starVLA.training.train_starvla][INFO] - Step 1844, Loss: {'train/loss': 0.5558934211730957, 'train/grad_norm_pre_clip': 8.214400291442871, 'timing/data': 0.0005592210218310356, 'timing/model': 0.991652543656528, 'learning_rate/qwen_vl_interface': 8.579576085844695e-06, 'learning_rate/action_model': 8.579576085844694e-05, 'epoch': 0.76}) +[2026-06-17 19:41:13,889][starVLA.training.train_starvla][INFO] - Step 1845, Loss: {'train/loss': 0.6184086799621582, 'train/grad_norm_pre_clip': 14.313958168029785, 'timing/data': 0.0006308676674962044, 'timing/model': 0.9791165292263031, 'learning_rate/qwen_vl_interface': 8.578033315780474e-06, 'learning_rate/action_model': 8.578033315780475e-05, 'epoch': 0.76}) +[2026-06-17 19:41:17,901][starVLA.training.train_starvla][INFO] - Step 1846, Loss: {'train/loss': 0.6053375005722046, 'train/grad_norm_pre_clip': 10.994551658630371, 'timing/data': 0.0005563916638493538, 'timing/model': 1.1806272575631738, 'learning_rate/qwen_vl_interface': 8.576489855811371e-06, 'learning_rate/action_model': 8.57648985581137e-05, 'epoch': 0.76}) +[2026-06-17 19:41:21,805][starVLA.training.train_starvla][INFO] - Step 1847, Loss: {'train/loss': 0.6714494824409485, 'train/grad_norm_pre_clip': 12.226458549499512, 'timing/data': 0.0006823446601629257, 'timing/model': 1.1814019568264484, 'learning_rate/qwen_vl_interface': 8.57494570625734e-06, 'learning_rate/action_model': 8.57494570625734e-05, 'epoch': 0.76}) +[2026-06-17 19:41:25,631][starVLA.training.train_starvla][INFO] - Step 1848, Loss: {'train/loss': 0.6965957283973694, 'train/grad_norm_pre_clip': 15.703136444091797, 'timing/data': 0.0008772220462560654, 'timing/model': 0.9685870306566358, 'learning_rate/qwen_vl_interface': 8.573400867438486e-06, 'learning_rate/action_model': 8.573400867438486e-05, 'epoch': 0.76}) +[2026-06-17 19:41:29,575][starVLA.training.train_starvla][INFO] - Step 1849, Loss: {'train/loss': 0.4694700241088867, 'train/grad_norm_pre_clip': 9.829593658447266, 'timing/data': 0.0005767811089754105, 'timing/model': 0.9414228154346347, 'learning_rate/qwen_vl_interface': 8.571855339675057e-06, 'learning_rate/action_model': 8.571855339675057e-05, 'epoch': 0.76}) +[2026-06-17 19:41:33,399][starVLA.training.train_starvla][INFO] - Step 1850, Loss: {'train/loss': 0.2751614451408386, 'train/grad_norm_pre_clip': 7.110803604125977, 'timing/data': 0.0004955539479851723, 'timing/model': 0.9503931263461709, 'learning_rate/qwen_vl_interface': 8.57030912328744e-06, 'learning_rate/action_model': 8.57030912328744e-05, 'epoch': 0.76}) +[2026-06-17 19:41:37,254][starVLA.training.train_starvla][INFO] - Step 1851, Loss: {'train/loss': 0.5983112454414368, 'train/grad_norm_pre_clip': 14.229808807373047, 'timing/data': 0.0008128946647047997, 'timing/model': 0.961973543278873, 'learning_rate/qwen_vl_interface': 8.568762218596172e-06, 'learning_rate/action_model': 8.568762218596171e-05, 'epoch': 0.76}) +[2026-06-17 19:41:41,141][starVLA.training.train_starvla][INFO] - Step 1852, Loss: {'train/loss': 0.40649497509002686, 'train/grad_norm_pre_clip': 9.245691299438477, 'timing/data': 0.0006336532533168793, 'timing/model': 0.9335242975503206, 'learning_rate/qwen_vl_interface': 8.567214625921922e-06, 'learning_rate/action_model': 8.567214625921922e-05, 'epoch': 0.76}) +[2026-06-17 19:41:45,017][starVLA.training.train_starvla][INFO] - Step 1853, Loss: {'train/loss': 0.14308252930641174, 'train/grad_norm_pre_clip': 4.570524215698242, 'timing/data': 0.0010652216151356697, 'timing/model': 1.1641219407320023, 'learning_rate/qwen_vl_interface': 8.565666345585511e-06, 'learning_rate/action_model': 8.56566634558551e-05, 'epoch': 0.76}) +[2026-06-17 19:41:48,878][starVLA.training.train_starvla][INFO] - Step 1854, Loss: {'train/loss': 0.680753767490387, 'train/grad_norm_pre_clip': 15.375835418701172, 'timing/data': 0.000711662694811821, 'timing/model': 1.1296138148754835, 'learning_rate/qwen_vl_interface': 8.564117377907898e-06, 'learning_rate/action_model': 8.564117377907898e-05, 'epoch': 0.76}) +[2026-06-17 19:41:52,584][starVLA.training.train_starvla][INFO] - Step 1855, Loss: {'train/loss': 0.4886434078216553, 'train/grad_norm_pre_clip': 14.077929496765137, 'timing/data': 0.0005623167380690575, 'timing/model': 1.010556573048234, 'learning_rate/qwen_vl_interface': 8.56256772321019e-06, 'learning_rate/action_model': 8.562567723210188e-05, 'epoch': 0.76}) +[2026-06-17 19:41:56,440][starVLA.training.train_starvla][INFO] - Step 1856, Loss: {'train/loss': 0.5040447115898132, 'train/grad_norm_pre_clip': 11.54848861694336, 'timing/data': 0.0008861077949404716, 'timing/model': 0.9459935296326876, 'learning_rate/qwen_vl_interface': 8.561017381813624e-06, 'learning_rate/action_model': 8.561017381813624e-05, 'epoch': 0.76}) +[2026-06-17 19:42:00,352][starVLA.training.train_starvla][INFO] - Step 1857, Loss: {'train/loss': 0.7509999871253967, 'train/grad_norm_pre_clip': 14.386418342590332, 'timing/data': 0.0007157847285270691, 'timing/model': 0.9503367738798261, 'learning_rate/qwen_vl_interface': 8.559466354039594e-06, 'learning_rate/action_model': 8.559466354039594e-05, 'epoch': 0.76}) +[2026-06-17 19:42:04,354][starVLA.training.train_starvla][INFO] - Step 1858, Loss: {'train/loss': 0.3708251714706421, 'train/grad_norm_pre_clip': 10.210619926452637, 'timing/data': 0.0012788735330104828, 'timing/model': 0.9470178643241525, 'learning_rate/qwen_vl_interface': 8.557914640209628e-06, 'learning_rate/action_model': 8.557914640209627e-05, 'epoch': 0.76}) +[2026-06-17 19:42:08,230][starVLA.training.train_starvla][INFO] - Step 1859, Loss: {'train/loss': 0.679653525352478, 'train/grad_norm_pre_clip': 12.988511085510254, 'timing/data': 0.0005053449422121048, 'timing/model': 0.9550311928614974, 'learning_rate/qwen_vl_interface': 8.556362240645397e-06, 'learning_rate/action_model': 8.556362240645397e-05, 'epoch': 0.77}) +[2026-06-17 19:42:12,252][starVLA.training.train_starvla][INFO] - Step 1860, Loss: {'train/loss': 0.6199822425842285, 'train/grad_norm_pre_clip': 11.675655364990234, 'timing/data': 0.0009774137288331985, 'timing/model': 1.2498168060556054, 'learning_rate/qwen_vl_interface': 8.554809155668716e-06, 'learning_rate/action_model': 8.554809155668716e-05, 'epoch': 0.77}) +[2026-06-17 19:42:15,952][starVLA.training.train_starvla][INFO] - Step 1861, Loss: {'train/loss': 0.4395984709262848, 'train/grad_norm_pre_clip': 14.694120407104492, 'timing/data': 0.000733385793864727, 'timing/model': 0.9551499476656318, 'learning_rate/qwen_vl_interface': 8.553255385601541e-06, 'learning_rate/action_model': 8.55325538560154e-05, 'epoch': 0.77}) +[2026-06-17 19:42:19,945][starVLA.training.train_starvla][INFO] - Step 1862, Loss: {'train/loss': 0.6642930507659912, 'train/grad_norm_pre_clip': 11.18802547454834, 'timing/data': 0.0010263444855809212, 'timing/model': 0.9491240754723549, 'learning_rate/qwen_vl_interface': 8.551700930765969e-06, 'learning_rate/action_model': 8.551700930765969e-05, 'epoch': 0.77}) +[2026-06-17 19:42:23,899][starVLA.training.train_starvla][INFO] - Step 1863, Loss: {'train/loss': 0.33604753017425537, 'train/grad_norm_pre_clip': 8.825737953186035, 'timing/data': 0.001549242064356804, 'timing/model': 0.9523099111393094, 'learning_rate/qwen_vl_interface': 8.550145791484243e-06, 'learning_rate/action_model': 8.550145791484242e-05, 'epoch': 0.77}) +[2026-06-17 19:42:27,945][starVLA.training.train_starvla][INFO] - Step 1864, Loss: {'train/loss': 0.8718470931053162, 'train/grad_norm_pre_clip': 13.205606460571289, 'timing/data': 0.0012755049392580986, 'timing/model': 0.942517526447773, 'learning_rate/qwen_vl_interface': 8.548589968078743e-06, 'learning_rate/action_model': 8.548589968078743e-05, 'epoch': 0.77}) +[2026-06-17 19:42:31,930][starVLA.training.train_starvla][INFO] - Step 1865, Loss: {'train/loss': 0.5355097055435181, 'train/grad_norm_pre_clip': 9.837653160095215, 'timing/data': 0.0010063257068395615, 'timing/model': 0.9721042765304446, 'learning_rate/qwen_vl_interface': 8.547033460871992e-06, 'learning_rate/action_model': 8.547033460871992e-05, 'epoch': 0.77}) +[2026-06-17 19:42:35,978][starVLA.training.train_starvla][INFO] - Step 1866, Loss: {'train/loss': 0.5636391639709473, 'train/grad_norm_pre_clip': 11.485532760620117, 'timing/data': 0.0010540718212723732, 'timing/model': 1.1774026276543736, 'learning_rate/qwen_vl_interface': 8.545476270186658e-06, 'learning_rate/action_model': 8.545476270186659e-05, 'epoch': 0.77}) +[2026-06-17 19:42:40,017][starVLA.training.train_starvla][INFO] - Step 1867, Loss: {'train/loss': 0.5289270281791687, 'train/grad_norm_pre_clip': 9.078407287597656, 'timing/data': 0.0006442740559577942, 'timing/model': 1.2443741969764233, 'learning_rate/qwen_vl_interface': 8.543918396345546e-06, 'learning_rate/action_model': 8.543918396345546e-05, 'epoch': 0.77}) +[2026-06-17 19:42:43,841][starVLA.training.train_starvla][INFO] - Step 1868, Loss: {'train/loss': 0.5547624826431274, 'train/grad_norm_pre_clip': 11.594199180603027, 'timing/data': 0.0005696304142475128, 'timing/model': 1.0649168835952878, 'learning_rate/qwen_vl_interface': 8.542359839671607e-06, 'learning_rate/action_model': 8.542359839671606e-05, 'epoch': 0.77}) +[2026-06-17 19:42:47,931][starVLA.training.train_starvla][INFO] - Step 1869, Loss: {'train/loss': 0.6479628682136536, 'train/grad_norm_pre_clip': 16.586273193359375, 'timing/data': 0.001009264960885048, 'timing/model': 1.061735318042338, 'learning_rate/qwen_vl_interface': 8.540800600487931e-06, 'learning_rate/action_model': 8.54080060048793e-05, 'epoch': 0.77}) +[2026-06-17 19:42:51,906][starVLA.training.train_starvla][INFO] - Step 1870, Loss: {'train/loss': 0.46610549092292786, 'train/grad_norm_pre_clip': 11.865673065185547, 'timing/data': 0.0013148030266165733, 'timing/model': 0.9742981102317572, 'learning_rate/qwen_vl_interface': 8.539240679117748e-06, 'learning_rate/action_model': 8.539240679117748e-05, 'epoch': 0.77}) +[2026-06-17 19:42:55,873][starVLA.training.train_starvla][INFO] - Step 1871, Loss: {'train/loss': 0.5903840065002441, 'train/grad_norm_pre_clip': 10.5901517868042, 'timing/data': 0.0011677136644721031, 'timing/model': 0.9704668736085296, 'learning_rate/qwen_vl_interface': 8.537680075884433e-06, 'learning_rate/action_model': 8.537680075884433e-05, 'epoch': 0.77}) +[2026-06-17 19:42:59,981][starVLA.training.train_starvla][INFO] - Step 1872, Loss: {'train/loss': 0.5708610415458679, 'train/grad_norm_pre_clip': 14.475107192993164, 'timing/data': 0.0012316806241869926, 'timing/model': 1.033519983291626, 'learning_rate/qwen_vl_interface': 8.5361187911115e-06, 'learning_rate/action_model': 8.5361187911115e-05, 'epoch': 0.77}) +[2026-06-17 19:43:04,082][starVLA.training.train_starvla][INFO] - Step 1873, Loss: {'train/loss': 0.47490552067756653, 'train/grad_norm_pre_clip': 10.149889945983887, 'timing/data': 0.0006018253043293953, 'timing/model': 1.2136705396696925, 'learning_rate/qwen_vl_interface': 8.534556825122608e-06, 'learning_rate/action_model': 8.534556825122607e-05, 'epoch': 0.77}) +[2026-06-17 19:43:08,066][starVLA.training.train_starvla][INFO] - Step 1874, Loss: {'train/loss': 0.8637768626213074, 'train/grad_norm_pre_clip': 12.151369094848633, 'timing/data': 0.0007256316021084785, 'timing/model': 1.1602106411010027, 'learning_rate/qwen_vl_interface': 8.532994178241551e-06, 'learning_rate/action_model': 8.53299417824155e-05, 'epoch': 0.77}) +[2026-06-17 19:43:11,751][starVLA.training.train_starvla][INFO] - Step 1875, Loss: {'train/loss': 0.5360058546066284, 'train/grad_norm_pre_clip': 11.449155807495117, 'timing/data': 0.0008690515533089638, 'timing/model': 0.9452750263735652, 'learning_rate/qwen_vl_interface': 8.531430850792267e-06, 'learning_rate/action_model': 8.531430850792266e-05, 'epoch': 0.77}) +[2026-06-17 19:43:15,892][starVLA.training.train_starvla][INFO] - Step 1876, Loss: {'train/loss': 0.6898523569107056, 'train/grad_norm_pre_clip': 16.493070602416992, 'timing/data': 0.0008193477988243103, 'timing/model': 1.0237392205744982, 'learning_rate/qwen_vl_interface': 8.529866843098838e-06, 'learning_rate/action_model': 8.529866843098837e-05, 'epoch': 0.77}) +[2026-06-17 19:43:19,859][starVLA.training.train_starvla][INFO] - Step 1877, Loss: {'train/loss': 0.3362904489040375, 'train/grad_norm_pre_clip': 6.941533088684082, 'timing/data': 0.0005571749061346054, 'timing/model': 0.9685177421197295, 'learning_rate/qwen_vl_interface': 8.528302155485482e-06, 'learning_rate/action_model': 8.528302155485481e-05, 'epoch': 0.77}) +[2026-06-17 19:43:23,922][starVLA.training.train_starvla][INFO] - Step 1878, Loss: {'train/loss': 0.5509181022644043, 'train/grad_norm_pre_clip': 12.736783027648926, 'timing/data': 0.0009082704782485962, 'timing/model': 0.9758001836016774, 'learning_rate/qwen_vl_interface': 8.526736788276563e-06, 'learning_rate/action_model': 8.526736788276563e-05, 'epoch': 0.77}) +[2026-06-17 19:43:27,847][starVLA.training.train_starvla][INFO] - Step 1879, Loss: {'train/loss': 0.7463093996047974, 'train/grad_norm_pre_clip': 11.89865493774414, 'timing/data': 0.0005342792719602585, 'timing/model': 0.9566706130281091, 'learning_rate/qwen_vl_interface': 8.525170741796578e-06, 'learning_rate/action_model': 8.525170741796579e-05, 'epoch': 0.77}) +[2026-06-17 19:43:31,957][starVLA.training.train_starvla][INFO] - Step 1880, Loss: {'train/loss': 0.6309425830841064, 'train/grad_norm_pre_clip': 10.533638954162598, 'timing/data': 0.0008722022175788879, 'timing/model': 1.242228495888412, 'learning_rate/qwen_vl_interface': 8.523604016370179e-06, 'learning_rate/action_model': 8.523604016370179e-05, 'epoch': 0.77}) +[2026-06-17 19:43:35,990][starVLA.training.train_starvla][INFO] - Step 1881, Loss: {'train/loss': 0.8777255415916443, 'train/grad_norm_pre_clip': 11.172359466552734, 'timing/data': 0.0007659317925572395, 'timing/model': 1.135415288619697, 'learning_rate/qwen_vl_interface': 8.522036612322143e-06, 'learning_rate/action_model': 8.522036612322142e-05, 'epoch': 0.77}) +[2026-06-17 19:43:39,814][starVLA.training.train_starvla][INFO] - Step 1882, Loss: {'train/loss': 0.34620240330696106, 'train/grad_norm_pre_clip': 7.461679458618164, 'timing/data': 0.0006103403866291046, 'timing/model': 0.9487323155626655, 'learning_rate/qwen_vl_interface': 8.520468529977395e-06, 'learning_rate/action_model': 8.520468529977395e-05, 'epoch': 0.77}) +[2026-06-17 19:43:43,807][starVLA.training.train_starvla][INFO] - Step 1883, Loss: {'train/loss': 0.7074633836746216, 'train/grad_norm_pre_clip': 12.532087326049805, 'timing/data': 0.0007569119334220886, 'timing/model': 0.9709047535434365, 'learning_rate/qwen_vl_interface': 8.518899769661002e-06, 'learning_rate/action_model': 8.518899769661002e-05, 'epoch': 0.78}) +[2026-06-17 19:43:48,061][starVLA.training.train_starvla][INFO] - Step 1884, Loss: {'train/loss': 0.9125326871871948, 'train/grad_norm_pre_clip': 12.271930694580078, 'timing/data': 0.0009287530556321144, 'timing/model': 0.970482025295496, 'learning_rate/qwen_vl_interface': 8.517330331698171e-06, 'learning_rate/action_model': 8.51733033169817e-05, 'epoch': 0.78}) +[2026-06-17 19:43:52,070][starVLA.training.train_starvla][INFO] - Step 1885, Loss: {'train/loss': 0.6182239055633545, 'train/grad_norm_pre_clip': 11.458898544311523, 'timing/data': 0.0016071340069174767, 'timing/model': 0.9501671148464084, 'learning_rate/qwen_vl_interface': 8.515760216414244e-06, 'learning_rate/action_model': 8.515760216414244e-05, 'epoch': 0.78}) +[2026-06-17 19:43:56,067][starVLA.training.train_starvla][INFO] - Step 1886, Loss: {'train/loss': 0.7472465634346008, 'train/grad_norm_pre_clip': 14.238616943359375, 'timing/data': 0.0010186899453401566, 'timing/model': 1.234829724766314, 'learning_rate/qwen_vl_interface': 8.514189424134712e-06, 'learning_rate/action_model': 8.514189424134713e-05, 'epoch': 0.78}) +[2026-06-17 19:43:59,925][starVLA.training.train_starvla][INFO] - Step 1887, Loss: {'train/loss': 0.8139404058456421, 'train/grad_norm_pre_clip': 20.351478576660156, 'timing/data': 0.0005432842299342155, 'timing/model': 1.1454679323360324, 'learning_rate/qwen_vl_interface': 8.512617955185199e-06, 'learning_rate/action_model': 8.5126179551852e-05, 'epoch': 0.78}) +[2026-06-17 19:44:03,716][starVLA.training.train_starvla][INFO] - Step 1888, Loss: {'train/loss': 0.47083595395088196, 'train/grad_norm_pre_clip': 8.902018547058105, 'timing/data': 0.0008983425796031952, 'timing/model': 0.9634194383397698, 'learning_rate/qwen_vl_interface': 8.511045809891472e-06, 'learning_rate/action_model': 8.511045809891472e-05, 'epoch': 0.78}) +[2026-06-17 19:44:07,671][starVLA.training.train_starvla][INFO] - Step 1889, Loss: {'train/loss': 0.8704613447189331, 'train/grad_norm_pre_clip': 12.545235633850098, 'timing/data': 0.0012687146663665771, 'timing/model': 0.9575066333636642, 'learning_rate/qwen_vl_interface': 8.509472988579443e-06, 'learning_rate/action_model': 8.509472988579443e-05, 'epoch': 0.78}) +[2026-06-17 19:44:11,692][starVLA.training.train_starvla][INFO] - Step 1890, Loss: {'train/loss': 0.7626463770866394, 'train/grad_norm_pre_clip': 12.47671890258789, 'timing/data': 0.0007044617086648941, 'timing/model': 0.9496158342808485, 'learning_rate/qwen_vl_interface': 8.507899491575153e-06, 'learning_rate/action_model': 8.507899491575154e-05, 'epoch': 0.78}) +[2026-06-17 19:44:15,650][starVLA.training.train_starvla][INFO] - Step 1891, Loss: {'train/loss': 0.6247223019599915, 'train/grad_norm_pre_clip': 14.658842086791992, 'timing/data': 0.0007964828982949257, 'timing/model': 0.956188322044909, 'learning_rate/qwen_vl_interface': 8.506325319204795e-06, 'learning_rate/action_model': 8.506325319204796e-05, 'epoch': 0.78}) +[2026-06-17 19:44:19,602][starVLA.training.train_starvla][INFO] - Step 1892, Loss: {'train/loss': 0.909637451171875, 'train/grad_norm_pre_clip': 12.238332748413086, 'timing/data': 0.0008040228858590126, 'timing/model': 0.954474026337266, 'learning_rate/qwen_vl_interface': 8.504750471794694e-06, 'learning_rate/action_model': 8.504750471794693e-05, 'epoch': 0.78}) +[2026-06-17 19:44:23,605][starVLA.training.train_starvla][INFO] - Step 1893, Loss: {'train/loss': 0.7626647353172302, 'train/grad_norm_pre_clip': 12.829081535339355, 'timing/data': 0.0009392453357577324, 'timing/model': 1.252378516830504, 'learning_rate/qwen_vl_interface': 8.503174949671315e-06, 'learning_rate/action_model': 8.503174949671315e-05, 'epoch': 0.78}) +[2026-06-17 19:44:27,368][starVLA.training.train_starvla][INFO] - Step 1894, Loss: {'train/loss': 0.7607042193412781, 'train/grad_norm_pre_clip': 11.956018447875977, 'timing/data': 0.0007654009386897087, 'timing/model': 0.9739084476605058, 'learning_rate/qwen_vl_interface': 8.501598753161273e-06, 'learning_rate/action_model': 8.501598753161272e-05, 'epoch': 0.78}) +[2026-06-17 19:44:31,346][starVLA.training.train_starvla][INFO] - Step 1895, Loss: {'train/loss': 0.626312792301178, 'train/grad_norm_pre_clip': 10.616289138793945, 'timing/data': 0.000733494758605957, 'timing/model': 0.9507908523082733, 'learning_rate/qwen_vl_interface': 8.500021882591308e-06, 'learning_rate/action_model': 8.500021882591308e-05, 'epoch': 0.78}) +[2026-06-17 19:44:35,351][starVLA.training.train_starvla][INFO] - Step 1896, Loss: {'train/loss': 0.30930355191230774, 'train/grad_norm_pre_clip': 6.6874680519104, 'timing/data': 0.0008731093257665634, 'timing/model': 0.9518023347482085, 'learning_rate/qwen_vl_interface': 8.498444338288308e-06, 'learning_rate/action_model': 8.498444338288309e-05, 'epoch': 0.78}) +[2026-06-17 19:44:39,314][starVLA.training.train_starvla][INFO] - Step 1897, Loss: {'train/loss': 0.42696964740753174, 'train/grad_norm_pre_clip': 8.03231430053711, 'timing/data': 0.0006716540083289146, 'timing/model': 0.9495854079723358, 'learning_rate/qwen_vl_interface': 8.496866120579304e-06, 'learning_rate/action_model': 8.496866120579303e-05, 'epoch': 0.78}) +[2026-06-17 19:44:43,355][starVLA.training.train_starvla][INFO] - Step 1898, Loss: {'train/loss': 0.381777822971344, 'train/grad_norm_pre_clip': 9.034717559814453, 'timing/data': 0.0005584321916103363, 'timing/model': 0.9941158620640635, 'learning_rate/qwen_vl_interface': 8.495287229791459e-06, 'learning_rate/action_model': 8.495287229791459e-05, 'epoch': 0.78}) +[2026-06-17 19:44:47,478][starVLA.training.train_starvla][INFO] - Step 1899, Loss: {'train/loss': 0.6821515560150146, 'train/grad_norm_pre_clip': 10.80012035369873, 'timing/data': 0.0011428054422140121, 'timing/model': 1.2243337836116552, 'learning_rate/qwen_vl_interface': 8.493707666252078e-06, 'learning_rate/action_model': 8.493707666252078e-05, 'epoch': 0.78}) +[2026-06-17 19:47:02,858][starVLA.training.train_starvla][INFO] - Step 1900, Loss: {'train/loss': 0.8933191895484924, 'train/grad_norm_pre_clip': 13.389811515808105, 'eval/loss': 0.6103129982948303, 'eval/latency_6/loss': 0.6103129982948303, 'eval/demon_fire_precision': 0.9208611279563371, 'eval/demon_fire_recall': 0.9815772462831286, 'eval/demon_fire_f1': 0.9502503128911138, 'eval/demon_fire_rate_model': 0.9176405119643851, 'eval/demon_fire_rate_teacher': 0.860879243183083, 'eval/demon_left_precision': 0.7684210526315789, 'eval/demon_left_recall': 0.6403508771929824, 'eval/demon_left_f1': 0.6985645933014354, 'eval/demon_left_rate_model': 0.13216471897607124, 'eval/demon_left_rate_teacher': 0.15859766277128548, 'eval/demon_right_precision': 0.7454545454545455, 'eval/demon_right_recall': 0.6298003072196621, 'eval/demon_right_f1': 0.6827643630308077, 'eval/demon_right_rate_model': 0.1530328324986088, 'eval/demon_right_rate_teacher': 0.18113522537562604, 'eval/demon_move_f1': 0.6906644781661215, 'eval/demon_cc_f1': 0.853961981521345, 'eval/latency_6/demon_fire_precision': 0.9208611279563371, 'eval/latency_6/demon_fire_recall': 0.9815772462831286, 'eval/latency_6/demon_fire_f1': 0.9502503128911138, 'eval/latency_6/demon_fire_rate_model': 0.9176405119643851, 'eval/latency_6/demon_fire_rate_teacher': 0.860879243183083, 'eval/latency_6/demon_left_precision': 0.7684210526315789, 'eval/latency_6/demon_left_recall': 0.6403508771929824, 'eval/latency_6/demon_left_f1': 0.6985645933014354, 'eval/latency_6/demon_left_rate_model': 0.13216471897607124, 'eval/latency_6/demon_left_rate_teacher': 0.15859766277128548, 'eval/latency_6/demon_right_precision': 0.7454545454545455, 'eval/latency_6/demon_right_recall': 0.6298003072196621, 'eval/latency_6/demon_right_f1': 0.6827643630308077, 'eval/latency_6/demon_right_rate_model': 0.1530328324986088, 'eval/latency_6/demon_right_rate_teacher': 0.18113522537562604, 'eval/latency_6/demon_move_f1': 0.6906644781661215, 'eval/latency_6/demon_cc_f1': 0.853961981521345, 'timing/data': 0.0009105131030082703, 'timing/model': 0.9783801017329097, 'learning_rate/qwen_vl_interface': 8.492127430288609e-06, 'learning_rate/action_model': 8.492127430288608e-05, 'epoch': 0.78}) +[2026-06-17 19:47:02,862][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_1900_state +[2026-06-17 19:47:02,867][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 19:47:09,525][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_1900_state/model.safetensors +[2026-06-17 19:47:25,525][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_1900_state/optimizer.bin +[2026-06-17 19:47:25,526][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_1900_state/sampler.bin +[2026-06-17 19:47:25,526][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_1900_state/sampler_1.bin +[2026-06-17 19:47:25,550][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_1900_state/random_states_0.pkl +[2026-06-17 19:47:25,554][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1900 +[2026-06-17 19:47:25,554][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 19:47:25,565][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 19:47:29,693][starVLA.training.train_starvla][INFO] - Step 1901, Loss: {'train/loss': 0.29194504022598267, 'train/grad_norm_pre_clip': 6.866480350494385, 'timing/data': 0.0005518170073628426, 'timing/model': 0.9668569257482886, 'learning_rate/qwen_vl_interface': 8.490546522228633e-06, 'learning_rate/action_model': 8.490546522228633e-05, 'epoch': 0.78}) +[2026-06-17 19:47:37,247][starVLA.training.train_starvla][INFO] - Step 1902, Loss: {'train/loss': 0.888551652431488, 'train/grad_norm_pre_clip': 17.652740478515625, 'timing/data': 0.0005540810525417328, 'timing/model': 1.2046006107702851, 'learning_rate/qwen_vl_interface': 8.488964942399876e-06, 'learning_rate/action_model': 8.488964942399877e-05, 'epoch': 0.78}) +[2026-06-17 19:47:41,702][starVLA.training.train_starvla][INFO] - Step 1903, Loss: {'train/loss': 0.8178572058677673, 'train/grad_norm_pre_clip': 9.929869651794434, 'timing/data': 0.0005523962900042534, 'timing/model': 0.9558616895228624, 'learning_rate/qwen_vl_interface': 8.487382691130203e-06, 'learning_rate/action_model': 8.487382691130202e-05, 'epoch': 0.78}) +[2026-06-17 19:47:45,663][starVLA.training.train_starvla][INFO] - Step 1904, Loss: {'train/loss': 0.8271669745445251, 'train/grad_norm_pre_clip': 14.935831069946289, 'timing/data': 0.0005596969276666641, 'timing/model': 0.9506928976625204, 'learning_rate/qwen_vl_interface': 8.485799768747614e-06, 'learning_rate/action_model': 8.485799768747614e-05, 'epoch': 0.78}) +[2026-06-17 19:47:49,801][starVLA.training.train_starvla][INFO] - Step 1905, Loss: {'train/loss': 0.7053523063659668, 'train/grad_norm_pre_clip': 11.541101455688477, 'timing/data': 0.0020966166630387306, 'timing/model': 1.109538353048265, 'learning_rate/qwen_vl_interface': 8.484216175580248e-06, 'learning_rate/action_model': 8.484216175580249e-05, 'epoch': 0.78}) +[2026-06-17 19:47:54,447][starVLA.training.train_starvla][INFO] - Step 1906, Loss: {'train/loss': 0.657154381275177, 'train/grad_norm_pre_clip': 23.718671798706055, 'timing/data': 0.0009756004437804222, 'timing/model': 0.9853512486442924, 'learning_rate/qwen_vl_interface': 8.482631911956395e-06, 'learning_rate/action_model': 8.482631911956395e-05, 'epoch': 0.78}) +[2026-06-17 19:47:59,430][starVLA.training.train_starvla][INFO] - Step 1907, Loss: {'train/loss': 0.5531750321388245, 'train/grad_norm_pre_clip': 237.19815063476562, 'timing/data': 0.0005617458373308182, 'timing/model': 0.9698134604841471, 'learning_rate/qwen_vl_interface': 8.481046978204462e-06, 'learning_rate/action_model': 8.481046978204463e-05, 'epoch': 0.78}) +[2026-06-17 19:48:03,386][starVLA.training.train_starvla][INFO] - Step 1908, Loss: {'train/loss': 0.7995888590812683, 'train/grad_norm_pre_clip': 14.151528358459473, 'timing/data': 0.000957355834543705, 'timing/model': 1.1690917210653424, 'learning_rate/qwen_vl_interface': 8.479461374653017e-06, 'learning_rate/action_model': 8.479461374653018e-05, 'epoch': 0.79}) +[2026-06-17 19:48:07,348][starVLA.training.train_starvla][INFO] - Step 1909, Loss: {'train/loss': 0.487979918718338, 'train/grad_norm_pre_clip': 8.64255142211914, 'timing/data': 0.0005881646648049355, 'timing/model': 1.1187516702339053, 'learning_rate/qwen_vl_interface': 8.477875101630752e-06, 'learning_rate/action_model': 8.477875101630753e-05, 'epoch': 0.79}) +[2026-06-17 19:48:12,117][starVLA.training.train_starvla][INFO] - Step 1910, Loss: {'train/loss': 0.6151144504547119, 'train/grad_norm_pre_clip': 9.003927230834961, 'timing/data': 0.000635114498436451, 'timing/model': 1.4602188980206847, 'learning_rate/qwen_vl_interface': 8.476288159466508e-06, 'learning_rate/action_model': 8.476288159466508e-05, 'epoch': 0.79}) +[2026-06-17 19:48:16,196][starVLA.training.train_starvla][INFO] - Step 1911, Loss: {'train/loss': 0.5922863483428955, 'train/grad_norm_pre_clip': 10.215699195861816, 'timing/data': 0.0005051745101809502, 'timing/model': 1.0220859171822667, 'learning_rate/qwen_vl_interface': 8.474700548489255e-06, 'learning_rate/action_model': 8.474700548489255e-05, 'epoch': 0.79}) +[2026-06-17 19:48:22,160][starVLA.training.train_starvla][INFO] - Step 1912, Loss: {'train/loss': 0.45149338245391846, 'train/grad_norm_pre_clip': 9.37409496307373, 'timing/data': 0.0004989858716726303, 'timing/model': 0.9368092594668269, 'learning_rate/qwen_vl_interface': 8.473112269028108e-06, 'learning_rate/action_model': 8.473112269028107e-05, 'epoch': 0.79}) +[2026-06-17 19:48:26,211][starVLA.training.train_starvla][INFO] - Step 1913, Loss: {'train/loss': 0.7761970162391663, 'train/grad_norm_pre_clip': 11.447074890136719, 'timing/data': 0.0014653867110610008, 'timing/model': 1.1758798575028777, 'learning_rate/qwen_vl_interface': 8.47152332141232e-06, 'learning_rate/action_model': 8.47152332141232e-05, 'epoch': 0.79}) +[2026-06-17 19:48:31,130][starVLA.training.train_starvla][INFO] - Step 1914, Loss: {'train/loss': 0.8284752368927002, 'train/grad_norm_pre_clip': 21.790931701660156, 'timing/data': 0.000598931685090065, 'timing/model': 1.1199647756293416, 'learning_rate/qwen_vl_interface': 8.469933705971282e-06, 'learning_rate/action_model': 8.469933705971282e-05, 'epoch': 0.79}) +[2026-06-17 19:48:35,574][starVLA.training.train_starvla][INFO] - Step 1915, Loss: {'train/loss': 0.6130421757698059, 'train/grad_norm_pre_clip': 10.050827026367188, 'timing/data': 0.0012887977063655853, 'timing/model': 1.16632298566401, 'learning_rate/qwen_vl_interface': 8.468343423034521e-06, 'learning_rate/action_model': 8.46834342303452e-05, 'epoch': 0.79}) +[2026-06-17 19:48:40,845][starVLA.training.train_starvla][INFO] - Step 1916, Loss: {'train/loss': 0.6039801239967346, 'train/grad_norm_pre_clip': 11.471802711486816, 'timing/data': 0.0005306005477905273, 'timing/model': 0.930568553507328, 'learning_rate/qwen_vl_interface': 8.466752472931704e-06, 'learning_rate/action_model': 8.466752472931704e-05, 'epoch': 0.79}) +[2026-06-17 19:48:44,630][starVLA.training.train_starvla][INFO] - Step 1917, Loss: {'train/loss': 0.5306596159934998, 'train/grad_norm_pre_clip': 9.457056999206543, 'timing/data': 0.00048654433339834213, 'timing/model': 0.9349382156506181, 'learning_rate/qwen_vl_interface': 8.465160855992639e-06, 'learning_rate/action_model': 8.465160855992638e-05, 'epoch': 0.79}) +[2026-06-17 19:48:48,446][starVLA.training.train_starvla][INFO] - Step 1918, Loss: {'train/loss': 0.41694387793540955, 'train/grad_norm_pre_clip': 8.796831130981445, 'timing/data': 0.0005241846665740013, 'timing/model': 0.9269344052299857, 'learning_rate/qwen_vl_interface': 8.46356857254727e-06, 'learning_rate/action_model': 8.46356857254727e-05, 'epoch': 0.79}) +[2026-06-17 19:48:52,316][starVLA.training.train_starvla][INFO] - Step 1919, Loss: {'train/loss': 0.6992383599281311, 'train/grad_norm_pre_clip': 10.867308616638184, 'timing/data': 0.0005284976214170456, 'timing/model': 0.94185377843678, 'learning_rate/qwen_vl_interface': 8.461975622925676e-06, 'learning_rate/action_model': 8.461975622925676e-05, 'epoch': 0.79}) +[2026-06-17 19:48:56,207][starVLA.training.train_starvla][INFO] - Step 1920, Loss: {'train/loss': 0.5513219833374023, 'train/grad_norm_pre_clip': 10.720000267028809, 'timing/data': 0.0009916657581925392, 'timing/model': 0.940964269451797, 'learning_rate/qwen_vl_interface': 8.460382007458081e-06, 'learning_rate/action_model': 8.460382007458081e-05, 'epoch': 0.79}) +[2026-06-17 19:49:00,042][starVLA.training.train_starvla][INFO] - Step 1921, Loss: {'train/loss': 0.5405291318893433, 'train/grad_norm_pre_clip': 12.071903228759766, 'timing/data': 0.0014135092496871948, 'timing/model': 1.1676610000431538, 'learning_rate/qwen_vl_interface': 8.45878772647484e-06, 'learning_rate/action_model': 8.458787726474839e-05, 'epoch': 0.79}) +[2026-06-17 19:49:03,929][starVLA.training.train_starvla][INFO] - Step 1922, Loss: {'train/loss': 0.5066704154014587, 'train/grad_norm_pre_clip': 10.436150550842285, 'timing/data': 0.0008817026391625404, 'timing/model': 1.1471954649314284, 'learning_rate/qwen_vl_interface': 8.457192780306452e-06, 'learning_rate/action_model': 8.457192780306451e-05, 'epoch': 0.79}) +[2026-06-17 19:49:07,770][starVLA.training.train_starvla][INFO] - Step 1923, Loss: {'train/loss': 0.8585216403007507, 'train/grad_norm_pre_clip': 11.021894454956055, 'timing/data': 0.0006214529275894165, 'timing/model': 0.9832134572789073, 'learning_rate/qwen_vl_interface': 8.455597169283549e-06, 'learning_rate/action_model': 8.455597169283548e-05, 'epoch': 0.79}) +[2026-06-17 19:49:11,977][starVLA.training.train_starvla][INFO] - Step 1924, Loss: {'train/loss': 0.3814319968223572, 'train/grad_norm_pre_clip': 7.905327796936035, 'timing/data': 0.001007487066090107, 'timing/model': 1.1094412440434098, 'learning_rate/qwen_vl_interface': 8.454000893736901e-06, 'learning_rate/action_model': 8.4540008937369e-05, 'epoch': 0.79}) +[2026-06-17 19:49:16,026][starVLA.training.train_starvla][INFO] - Step 1925, Loss: {'train/loss': 0.5540140271186829, 'train/grad_norm_pre_clip': 11.04958438873291, 'timing/data': 0.0006074355915188789, 'timing/model': 0.9747534338384867, 'learning_rate/qwen_vl_interface': 8.452403953997422e-06, 'learning_rate/action_model': 8.452403953997422e-05, 'epoch': 0.79}) +[2026-06-17 19:49:19,989][starVLA.training.train_starvla][INFO] - Step 1926, Loss: {'train/loss': 0.6382951736450195, 'train/grad_norm_pre_clip': 44.63839340209961, 'timing/data': 0.0005586091428995132, 'timing/model': 0.9578115651383996, 'learning_rate/qwen_vl_interface': 8.450806350396158e-06, 'learning_rate/action_model': 8.450806350396158e-05, 'epoch': 0.79}) +[2026-06-17 19:49:23,935][starVLA.training.train_starvla][INFO] - Step 1927, Loss: {'train/loss': 0.5059622526168823, 'train/grad_norm_pre_clip': 10.771624565124512, 'timing/data': 0.0005285711959004402, 'timing/model': 0.9563578655943274, 'learning_rate/qwen_vl_interface': 8.449208083264291e-06, 'learning_rate/action_model': 8.449208083264292e-05, 'epoch': 0.79}) +[2026-06-17 19:49:28,031][starVLA.training.train_starvla][INFO] - Step 1928, Loss: {'train/loss': 0.9320716261863708, 'train/grad_norm_pre_clip': 14.035865783691406, 'timing/data': 0.000866706483066082, 'timing/model': 1.2503111232072115, 'learning_rate/qwen_vl_interface': 8.447609152933149e-06, 'learning_rate/action_model': 8.447609152933149e-05, 'epoch': 0.79}) +[2026-06-17 19:49:35,947][starVLA.training.train_starvla][INFO] - Step 1929, Loss: {'train/loss': 0.7018998265266418, 'train/grad_norm_pre_clip': 11.444891929626465, 'timing/data': 0.0005688173696398735, 'timing/model': 1.183085479773581, 'learning_rate/qwen_vl_interface': 8.44600955973419e-06, 'learning_rate/action_model': 8.446009559734188e-05, 'epoch': 0.79}) +[2026-06-17 19:49:40,929][starVLA.training.train_starvla][INFO] - Step 1930, Loss: {'train/loss': 0.8346351385116577, 'train/grad_norm_pre_clip': 18.49686622619629, 'timing/data': 0.0010379999876022339, 'timing/model': 0.972272046841681, 'learning_rate/qwen_vl_interface': 8.444409303999005e-06, 'learning_rate/action_model': 8.444409303999005e-05, 'epoch': 0.79}) +[2026-06-17 19:49:44,914][starVLA.training.train_starvla][INFO] - Step 1931, Loss: {'train/loss': 0.574674129486084, 'train/grad_norm_pre_clip': 10.81490421295166, 'timing/data': 0.000731348991394043, 'timing/model': 0.9772054580971599, 'learning_rate/qwen_vl_interface': 8.442808386059337e-06, 'learning_rate/action_model': 8.442808386059337e-05, 'epoch': 0.79}) +[2026-06-17 19:49:48,943][starVLA.training.train_starvla][INFO] - Step 1932, Loss: {'train/loss': 0.5000476837158203, 'train/grad_norm_pre_clip': 8.594542503356934, 'timing/data': 0.002309606410562992, 'timing/model': 0.9564133631065488, 'learning_rate/qwen_vl_interface': 8.441206806247055e-06, 'learning_rate/action_model': 8.441206806247054e-05, 'epoch': 0.8}) +[2026-06-17 19:49:53,625][starVLA.training.train_starvla][INFO] - Step 1933, Loss: {'train/loss': 0.3240305781364441, 'train/grad_norm_pre_clip': 7.468793869018555, 'timing/data': 0.0005539963021874428, 'timing/model': 1.6696669328957796, 'learning_rate/qwen_vl_interface': 8.439604564894167e-06, 'learning_rate/action_model': 8.439604564894167e-05, 'epoch': 0.8}) +[2026-06-17 19:49:57,969][starVLA.training.train_starvla][INFO] - Step 1934, Loss: {'train/loss': 0.7808676958084106, 'train/grad_norm_pre_clip': 11.466888427734375, 'timing/data': 0.0005279025062918663, 'timing/model': 0.984187088906765, 'learning_rate/qwen_vl_interface': 8.438001662332821e-06, 'learning_rate/action_model': 8.43800166233282e-05, 'epoch': 0.8}) +[2026-06-17 19:50:04,117][starVLA.training.train_starvla][INFO] - Step 1935, Loss: {'train/loss': 0.35662364959716797, 'train/grad_norm_pre_clip': 8.648781776428223, 'timing/data': 0.0009020715951919556, 'timing/model': 1.2256738422438502, 'learning_rate/qwen_vl_interface': 8.4363980988953e-06, 'learning_rate/action_model': 8.436398098895301e-05, 'epoch': 0.8}) +[2026-06-17 19:50:08,239][starVLA.training.train_starvla][INFO] - Step 1936, Loss: {'train/loss': 0.7171412110328674, 'train/grad_norm_pre_clip': 16.654457092285156, 'timing/data': 0.0010009277611970901, 'timing/model': 1.1348978132009506, 'learning_rate/qwen_vl_interface': 8.434793874914026e-06, 'learning_rate/action_model': 8.434793874914026e-05, 'epoch': 0.8}) +[2026-06-17 19:50:12,303][starVLA.training.train_starvla][INFO] - Step 1937, Loss: {'train/loss': 0.4854694902896881, 'train/grad_norm_pre_clip': 9.303664207458496, 'timing/data': 0.0005215415731072426, 'timing/model': 0.963359241373837, 'learning_rate/qwen_vl_interface': 8.433188990721553e-06, 'learning_rate/action_model': 8.433188990721552e-05, 'epoch': 0.8}) +[2026-06-17 19:50:16,338][starVLA.training.train_starvla][INFO] - Step 1938, Loss: {'train/loss': 0.5148129463195801, 'train/grad_norm_pre_clip': 9.03367805480957, 'timing/data': 0.0006071794778108597, 'timing/model': 0.9709820207208395, 'learning_rate/qwen_vl_interface': 8.431583446650577e-06, 'learning_rate/action_model': 8.431583446650578e-05, 'epoch': 0.8}) +[2026-06-17 19:50:23,307][starVLA.training.train_starvla][INFO] - Step 1939, Loss: {'train/loss': 0.4409250020980835, 'train/grad_norm_pre_clip': 10.828124046325684, 'timing/data': 0.0011787023395299911, 'timing/model': 2.421039587818086, 'learning_rate/qwen_vl_interface': 8.429977243033928e-06, 'learning_rate/action_model': 8.429977243033928e-05, 'epoch': 0.8}) +[2026-06-17 19:50:27,359][starVLA.training.train_starvla][INFO] - Step 1940, Loss: {'train/loss': 0.38020995259284973, 'train/grad_norm_pre_clip': 10.451250076293945, 'timing/data': 0.0011005876585841179, 'timing/model': 0.9693454802036285, 'learning_rate/qwen_vl_interface': 8.428370380204577e-06, 'learning_rate/action_model': 8.428370380204577e-05, 'epoch': 0.8}) +[2026-06-17 19:50:31,402][starVLA.training.train_starvla][INFO] - Step 1941, Loss: {'train/loss': 0.5558710694313049, 'train/grad_norm_pre_clip': 10.418351173400879, 'timing/data': 0.000880868174135685, 'timing/model': 1.1973618790507317, 'learning_rate/qwen_vl_interface': 8.426762858495625e-06, 'learning_rate/action_model': 8.426762858495625e-05, 'epoch': 0.8}) +[2026-06-17 19:50:35,255][starVLA.training.train_starvla][INFO] - Step 1942, Loss: {'train/loss': 0.4194072186946869, 'train/grad_norm_pre_clip': 9.376984596252441, 'timing/data': 0.0007844613865017891, 'timing/model': 1.1064499914646149, 'learning_rate/qwen_vl_interface': 8.425154678240315e-06, 'learning_rate/action_model': 8.425154678240314e-05, 'epoch': 0.8}) +[2026-06-17 19:50:38,950][starVLA.training.train_starvla][INFO] - Step 1943, Loss: {'train/loss': 0.4432847499847412, 'train/grad_norm_pre_clip': 14.449503898620605, 'timing/data': 0.0008543273434042931, 'timing/model': 0.9492085380479693, 'learning_rate/qwen_vl_interface': 8.423545839772021e-06, 'learning_rate/action_model': 8.423545839772021e-05, 'epoch': 0.8}) +[2026-06-17 19:50:43,045][starVLA.training.train_starvla][INFO] - Step 1944, Loss: {'train/loss': 0.33455270528793335, 'train/grad_norm_pre_clip': 10.696341514587402, 'timing/data': 0.0008432390168309212, 'timing/model': 0.973909336142242, 'learning_rate/qwen_vl_interface': 8.421936343424264e-06, 'learning_rate/action_model': 8.421936343424262e-05, 'epoch': 0.8}) +[2026-06-17 19:50:48,865][starVLA.training.train_starvla][INFO] - Step 1945, Loss: {'train/loss': 0.8554477095603943, 'train/grad_norm_pre_clip': 16.6916561126709, 'timing/data': 0.0007426580414175987, 'timing/model': 1.081590467132628, 'learning_rate/qwen_vl_interface': 8.420326189530685e-06, 'learning_rate/action_model': 8.420326189530685e-05, 'epoch': 0.8}) +[2026-06-17 19:50:53,025][starVLA.training.train_starvla][INFO] - Step 1946, Loss: {'train/loss': 0.7236486673355103, 'train/grad_norm_pre_clip': 16.248104095458984, 'timing/data': 0.0006717368960380554, 'timing/model': 1.0508371172472835, 'learning_rate/qwen_vl_interface': 8.418715378425078e-06, 'learning_rate/action_model': 8.418715378425078e-05, 'epoch': 0.8}) +[2026-06-17 19:50:57,099][starVLA.training.train_starvla][INFO] - Step 1947, Loss: {'train/loss': 0.37381935119628906, 'train/grad_norm_pre_clip': 11.627123832702637, 'timing/data': 0.0005319388583302498, 'timing/model': 0.9662369396537542, 'learning_rate/qwen_vl_interface': 8.417103910441364e-06, 'learning_rate/action_model': 8.417103910441363e-05, 'epoch': 0.8}) +[2026-06-17 19:51:01,374][starVLA.training.train_starvla][INFO] - Step 1948, Loss: {'train/loss': 0.9502974152565002, 'train/grad_norm_pre_clip': 15.16323184967041, 'timing/data': 0.0009336713701486588, 'timing/model': 1.2364259967580438, 'learning_rate/qwen_vl_interface': 8.415491785913599e-06, 'learning_rate/action_model': 8.4154917859136e-05, 'epoch': 0.8}) +[2026-06-17 19:51:05,161][starVLA.training.train_starvla][INFO] - Step 1949, Loss: {'train/loss': 0.369129478931427, 'train/grad_norm_pre_clip': 15.295905113220215, 'timing/data': 0.0011231359094381332, 'timing/model': 0.9653208451345563, 'learning_rate/qwen_vl_interface': 8.413879005175983e-06, 'learning_rate/action_model': 8.413879005175983e-05, 'epoch': 0.8}) +[2026-06-17 19:51:09,257][starVLA.training.train_starvla][INFO] - Step 1950, Loss: {'train/loss': 0.6644058227539062, 'train/grad_norm_pre_clip': 110.46646881103516, 'timing/data': 0.001434648409485817, 'timing/model': 0.9602195359766483, 'learning_rate/qwen_vl_interface': 8.412265568562843e-06, 'learning_rate/action_model': 8.412265568562842e-05, 'epoch': 0.8}) +[2026-06-17 19:51:13,356][starVLA.training.train_starvla][INFO] - Step 1951, Loss: {'train/loss': 1.0194536447525024, 'train/grad_norm_pre_clip': 16.153587341308594, 'timing/data': 0.0012416401877999306, 'timing/model': 0.9569957945495844, 'learning_rate/qwen_vl_interface': 8.41065147640865e-06, 'learning_rate/action_model': 8.41065147640865e-05, 'epoch': 0.8}) +[2026-06-17 19:51:17,355][starVLA.training.train_starvla][INFO] - Step 1952, Loss: {'train/loss': 0.4565179944038391, 'train/grad_norm_pre_clip': 14.374425888061523, 'timing/data': 0.0007737781852483749, 'timing/model': 0.9421934429556131, 'learning_rate/qwen_vl_interface': 8.409036729048002e-06, 'learning_rate/action_model': 8.409036729048002e-05, 'epoch': 0.8}) +[2026-06-17 19:51:21,506][starVLA.training.train_starvla][INFO] - Step 1953, Loss: {'train/loss': 0.5477597117424011, 'train/grad_norm_pre_clip': 12.751978874206543, 'timing/data': 0.0005213739350438118, 'timing/model': 1.0232267472893, 'learning_rate/qwen_vl_interface': 8.407421326815644e-06, 'learning_rate/action_model': 8.407421326815644e-05, 'epoch': 0.8}) +[2026-06-17 19:51:27,936][starVLA.training.train_starvla][INFO] - Step 1954, Loss: {'train/loss': 0.6360659003257751, 'train/grad_norm_pre_clip': 11.536727905273438, 'timing/data': 0.0006713103502988815, 'timing/model': 1.187028151936829, 'learning_rate/qwen_vl_interface': 8.405805270046445e-06, 'learning_rate/action_model': 8.405805270046445e-05, 'epoch': 0.8}) +[2026-06-17 19:51:32,519][starVLA.training.train_starvla][INFO] - Step 1955, Loss: {'train/loss': 0.7574361562728882, 'train/grad_norm_pre_clip': 14.044839859008789, 'timing/data': 0.0011117393150925636, 'timing/model': 0.969571421854198, 'learning_rate/qwen_vl_interface': 8.404188559075419e-06, 'learning_rate/action_model': 8.404188559075419e-05, 'epoch': 0.8}) +[2026-06-17 19:51:36,493][starVLA.training.train_starvla][INFO] - Step 1956, Loss: {'train/loss': 0.6876894235610962, 'train/grad_norm_pre_clip': 11.68004035949707, 'timing/data': 0.0007548993453383446, 'timing/model': 0.9838563520461321, 'learning_rate/qwen_vl_interface': 8.40257119423771e-06, 'learning_rate/action_model': 8.402571194237709e-05, 'epoch': 0.81}) +[2026-06-17 19:51:40,450][starVLA.training.train_starvla][INFO] - Step 1957, Loss: {'train/loss': 0.7527514696121216, 'train/grad_norm_pre_clip': 14.456483840942383, 'timing/data': 0.0006481055170297623, 'timing/model': 0.9456185074523091, 'learning_rate/qwen_vl_interface': 8.4009531758686e-06, 'learning_rate/action_model': 8.4009531758686e-05, 'epoch': 0.81}) +[2026-06-17 19:51:44,521][starVLA.training.train_starvla][INFO] - Step 1958, Loss: {'train/loss': 0.6063268780708313, 'train/grad_norm_pre_clip': 9.207344055175781, 'timing/data': 0.0006331177428364754, 'timing/model': 0.9790040189400315, 'learning_rate/qwen_vl_interface': 8.399334504303507e-06, 'learning_rate/action_model': 8.399334504303507e-05, 'epoch': 0.81}) +[2026-06-17 19:51:48,666][starVLA.training.train_starvla][INFO] - Step 1959, Loss: {'train/loss': 0.5777422785758972, 'train/grad_norm_pre_clip': 16.246599197387695, 'timing/data': 0.0005595032125711441, 'timing/model': 1.0131346341222525, 'learning_rate/qwen_vl_interface': 8.397715179877982e-06, 'learning_rate/action_model': 8.397715179877982e-05, 'epoch': 0.81}) +[2026-06-17 19:51:52,810][starVLA.training.train_starvla][INFO] - Step 1960, Loss: {'train/loss': 0.8011375665664673, 'train/grad_norm_pre_clip': 138.74484252929688, 'timing/data': 0.0006042830646038055, 'timing/model': 1.032162619754672, 'learning_rate/qwen_vl_interface': 8.396095202927712e-06, 'learning_rate/action_model': 8.396095202927712e-05, 'epoch': 0.81}) +[2026-06-17 19:51:58,074][starVLA.training.train_starvla][INFO] - Step 1961, Loss: {'train/loss': 0.5579657554626465, 'train/grad_norm_pre_clip': 12.495222091674805, 'timing/data': 0.000796598382294178, 'timing/model': 1.3094296380877495, 'learning_rate/qwen_vl_interface': 8.394474573788522e-06, 'learning_rate/action_model': 8.394474573788522e-05, 'epoch': 0.81}) +[2026-06-17 19:52:01,952][starVLA.training.train_starvla][INFO] - Step 1962, Loss: {'train/loss': 0.6355804800987244, 'train/grad_norm_pre_clip': 14.295724868774414, 'timing/data': 0.0008288882672786713, 'timing/model': 1.0117221614345908, 'learning_rate/qwen_vl_interface': 8.39285329279637e-06, 'learning_rate/action_model': 8.39285329279637e-05, 'epoch': 0.81}) +[2026-06-17 19:52:05,971][starVLA.training.train_starvla][INFO] - Step 1963, Loss: {'train/loss': 0.5217135548591614, 'train/grad_norm_pre_clip': 11.047660827636719, 'timing/data': 0.0011854907497763634, 'timing/model': 0.9546658117324114, 'learning_rate/qwen_vl_interface': 8.391231360287348e-06, 'learning_rate/action_model': 8.391231360287347e-05, 'epoch': 0.81}) +[2026-06-17 19:52:10,080][starVLA.training.train_starvla][INFO] - Step 1964, Loss: {'train/loss': 0.6009876728057861, 'train/grad_norm_pre_clip': 10.577515602111816, 'timing/data': 0.0006837360560894012, 'timing/model': 0.9873835239559412, 'learning_rate/qwen_vl_interface': 8.389608776597684e-06, 'learning_rate/action_model': 8.389608776597684e-05, 'epoch': 0.81}) +[2026-06-17 19:52:14,188][starVLA.training.train_starvla][INFO] - Step 1965, Loss: {'train/loss': 0.41465720534324646, 'train/grad_norm_pre_clip': 26.472740173339844, 'timing/data': 0.0012991400435566902, 'timing/model': 0.9930408829823136, 'learning_rate/qwen_vl_interface': 8.387985542063743e-06, 'learning_rate/action_model': 8.387985542063743e-05, 'epoch': 0.81}) +[2026-06-17 19:52:18,834][starVLA.training.train_starvla][INFO] - Step 1966, Loss: {'train/loss': 0.416860967874527, 'train/grad_norm_pre_clip': 8.630207061767578, 'timing/data': 0.0007558669894933701, 'timing/model': 0.9651207271963358, 'learning_rate/qwen_vl_interface': 8.386361657022024e-06, 'learning_rate/action_model': 8.386361657022023e-05, 'epoch': 0.81}) +[2026-06-17 19:52:22,972][starVLA.training.train_starvla][INFO] - Step 1967, Loss: {'train/loss': 0.7458863854408264, 'train/grad_norm_pre_clip': 14.64622974395752, 'timing/data': 0.0005822321400046349, 'timing/model': 1.2079703779891133, 'learning_rate/qwen_vl_interface': 8.384737121809156e-06, 'learning_rate/action_model': 8.384737121809156e-05, 'epoch': 0.81}) +[2026-06-17 19:52:26,916][starVLA.training.train_starvla][INFO] - Step 1968, Loss: {'train/loss': 0.5933908820152283, 'train/grad_norm_pre_clip': 11.686354637145996, 'timing/data': 0.000535275787115097, 'timing/model': 1.177408087067306, 'learning_rate/qwen_vl_interface': 8.38311193676191e-06, 'learning_rate/action_model': 8.383111936761909e-05, 'epoch': 0.81}) +[2026-06-17 19:52:30,902][starVLA.training.train_starvla][INFO] - Step 1969, Loss: {'train/loss': 0.40645864605903625, 'train/grad_norm_pre_clip': 10.46855354309082, 'timing/data': 0.000969727523624897, 'timing/model': 1.1238953564316034, 'learning_rate/qwen_vl_interface': 8.381486102217188e-06, 'learning_rate/action_model': 8.381486102217188e-05, 'epoch': 0.81}) +[2026-06-17 19:52:34,888][starVLA.training.train_starvla][INFO] - Step 1970, Loss: {'train/loss': 0.508939266204834, 'train/grad_norm_pre_clip': 9.358842849731445, 'timing/data': 0.001645130105316639, 'timing/model': 0.9882395677268505, 'learning_rate/qwen_vl_interface': 8.379859618512029e-06, 'learning_rate/action_model': 8.379859618512029e-05, 'epoch': 0.81}) +[2026-06-17 19:52:38,908][starVLA.training.train_starvla][INFO] - Step 1971, Loss: {'train/loss': 0.507394552230835, 'train/grad_norm_pre_clip': 24.527273178100586, 'timing/data': 0.0005393177270889282, 'timing/model': 0.96393589861691, 'learning_rate/qwen_vl_interface': 8.378232485983603e-06, 'learning_rate/action_model': 8.378232485983603e-05, 'epoch': 0.81}) +[2026-06-17 19:52:42,914][starVLA.training.train_starvla][INFO] - Step 1972, Loss: {'train/loss': 0.49525701999664307, 'train/grad_norm_pre_clip': 11.689451217651367, 'timing/data': 0.0008966345340013504, 'timing/model': 0.959829761646688, 'learning_rate/qwen_vl_interface': 8.376604704969216e-06, 'learning_rate/action_model': 8.376604704969216e-05, 'epoch': 0.81}) +[2026-06-17 19:52:47,964][starVLA.training.train_starvla][INFO] - Step 1973, Loss: {'train/loss': 0.32819613814353943, 'train/grad_norm_pre_clip': 7.632140159606934, 'timing/data': 0.0009698541834950447, 'timing/model': 2.058156010694802, 'learning_rate/qwen_vl_interface': 8.374976275806307e-06, 'learning_rate/action_model': 8.374976275806308e-05, 'epoch': 0.81}) +[2026-06-17 19:52:52,139][starVLA.training.train_starvla][INFO] - Step 1974, Loss: {'train/loss': 0.7216432094573975, 'train/grad_norm_pre_clip': 18.230207443237305, 'timing/data': 0.0009210072457790375, 'timing/model': 1.2965492820367217, 'learning_rate/qwen_vl_interface': 8.373347198832457e-06, 'learning_rate/action_model': 8.373347198832457e-05, 'epoch': 0.81}) +[2026-06-17 19:52:56,299][starVLA.training.train_starvla][INFO] - Step 1975, Loss: {'train/loss': 0.5358411073684692, 'train/grad_norm_pre_clip': 10.744330406188965, 'timing/data': 0.0005908627063035965, 'timing/model': 1.1802451834082603, 'learning_rate/qwen_vl_interface': 8.37171747438537e-06, 'learning_rate/action_model': 8.37171747438537e-05, 'epoch': 0.81}) +[2026-06-17 19:53:00,061][starVLA.training.train_starvla][INFO] - Step 1976, Loss: {'train/loss': 0.8302936553955078, 'train/grad_norm_pre_clip': 16.951242446899414, 'timing/data': 0.0007855091243982315, 'timing/model': 0.9542390154674649, 'learning_rate/qwen_vl_interface': 8.370087102802892e-06, 'learning_rate/action_model': 8.370087102802892e-05, 'epoch': 0.81}) +[2026-06-17 19:53:04,332][starVLA.training.train_starvla][INFO] - Step 1977, Loss: {'train/loss': 0.9779230952262878, 'train/grad_norm_pre_clip': 21.09896469116211, 'timing/data': 0.0014189155772328377, 'timing/model': 1.0917079877108335, 'learning_rate/qwen_vl_interface': 8.368456084423e-06, 'learning_rate/action_model': 8.368456084422999e-05, 'epoch': 0.81}) +[2026-06-17 19:53:10,098][starVLA.training.train_starvla][INFO] - Step 1978, Loss: {'train/loss': 0.5712795853614807, 'train/grad_norm_pre_clip': 13.91469955444336, 'timing/data': 0.0005294959992170334, 'timing/model': 1.0393456118181348, 'learning_rate/qwen_vl_interface': 8.366824419583806e-06, 'learning_rate/action_model': 8.366824419583806e-05, 'epoch': 0.81}) +[2026-06-17 19:53:14,027][starVLA.training.train_starvla][INFO] - Step 1979, Loss: {'train/loss': 0.6286332607269287, 'train/grad_norm_pre_clip': 11.384785652160645, 'timing/data': 0.0007744012400507927, 'timing/model': 0.9926653672009706, 'learning_rate/qwen_vl_interface': 8.365192108623555e-06, 'learning_rate/action_model': 8.365192108623555e-05, 'epoch': 0.81}) +[2026-06-17 19:53:18,013][starVLA.training.train_starvla][INFO] - Step 1980, Loss: {'train/loss': 0.5449795126914978, 'train/grad_norm_pre_clip': 12.137641906738281, 'timing/data': 0.0005402658134698868, 'timing/model': 0.9576658923178911, 'learning_rate/qwen_vl_interface': 8.363559151880631e-06, 'learning_rate/action_model': 8.36355915188063e-05, 'epoch': 0.82}) +[2026-06-17 19:53:23,965][starVLA.training.train_starvla][INFO] - Step 1981, Loss: {'train/loss': 0.4373221695423126, 'train/grad_norm_pre_clip': 9.352240562438965, 'timing/data': 0.0005636774003505707, 'timing/model': 1.218968115746975, 'learning_rate/qwen_vl_interface': 8.361925549693541e-06, 'learning_rate/action_model': 8.36192554969354e-05, 'epoch': 0.82}) +[2026-06-17 19:53:28,182][starVLA.training.train_starvla][INFO] - Step 1982, Loss: {'train/loss': 0.5162602066993713, 'train/grad_norm_pre_clip': 9.991959571838379, 'timing/data': 0.0005227094516158104, 'timing/model': 1.014075574465096, 'learning_rate/qwen_vl_interface': 8.360291302400937e-06, 'learning_rate/action_model': 8.360291302400937e-05, 'epoch': 0.82}) +[2026-06-17 19:53:32,618][starVLA.training.train_starvla][INFO] - Step 1983, Loss: {'train/loss': 0.4191041588783264, 'train/grad_norm_pre_clip': 10.031268119812012, 'timing/data': 0.0005409745499491692, 'timing/model': 0.9716774420812726, 'learning_rate/qwen_vl_interface': 8.3586564103416e-06, 'learning_rate/action_model': 8.358656410341599e-05, 'epoch': 0.82}) +[2026-06-17 19:53:36,644][starVLA.training.train_starvla][INFO] - Step 1984, Loss: {'train/loss': 0.9124832153320312, 'train/grad_norm_pre_clip': 19.605186462402344, 'timing/data': 0.0013041654601693153, 'timing/model': 0.9484028918668628, 'learning_rate/qwen_vl_interface': 8.357020873854445e-06, 'learning_rate/action_model': 8.357020873854444e-05, 'epoch': 0.82}) +[2026-06-17 19:53:40,743][starVLA.training.train_starvla][INFO] - Step 1985, Loss: {'train/loss': 0.4783327281475067, 'train/grad_norm_pre_clip': 12.642525672912598, 'timing/data': 0.002547311596572399, 'timing/model': 1.0383290071040392, 'learning_rate/qwen_vl_interface': 8.355384693278516e-06, 'learning_rate/action_model': 8.355384693278517e-05, 'epoch': 0.82}) +[2026-06-17 19:53:44,739][starVLA.training.train_starvla][INFO] - Step 1986, Loss: {'train/loss': 0.6204337477684021, 'train/grad_norm_pre_clip': 10.324161529541016, 'timing/data': 0.0014004399999976158, 'timing/model': 0.9790922235697508, 'learning_rate/qwen_vl_interface': 8.353747868953002e-06, 'learning_rate/action_model': 8.353747868953003e-05, 'epoch': 0.82}) +[2026-06-17 19:53:48,741][starVLA.training.train_starvla][INFO] - Step 1987, Loss: {'train/loss': 0.6469821333885193, 'train/grad_norm_pre_clip': 12.339282035827637, 'timing/data': 0.000790049321949482, 'timing/model': 1.2408845964819193, 'learning_rate/qwen_vl_interface': 8.352110401217213e-06, 'learning_rate/action_model': 8.352110401217213e-05, 'epoch': 0.82}) +[2026-06-17 19:53:52,620][starVLA.training.train_starvla][INFO] - Step 1988, Loss: {'train/loss': 0.6833381056785583, 'train/grad_norm_pre_clip': 10.542951583862305, 'timing/data': 0.0011130338534712791, 'timing/model': 1.1220414796844125, 'learning_rate/qwen_vl_interface': 8.350472290410602e-06, 'learning_rate/action_model': 8.350472290410602e-05, 'epoch': 0.82}) +[2026-06-17 19:53:56,490][starVLA.training.train_starvla][INFO] - Step 1989, Loss: {'train/loss': 0.6390601992607117, 'train/grad_norm_pre_clip': 14.601129531860352, 'timing/data': 0.0005507310852408409, 'timing/model': 0.9769497066736221, 'learning_rate/qwen_vl_interface': 8.348833536872747e-06, 'learning_rate/action_model': 8.348833536872747e-05, 'epoch': 0.82}) +[2026-06-17 19:54:00,519][starVLA.training.train_starvla][INFO] - Step 1990, Loss: {'train/loss': 0.5357513427734375, 'train/grad_norm_pre_clip': 10.461403846740723, 'timing/data': 0.000527367927134037, 'timing/model': 1.0014247670769691, 'learning_rate/qwen_vl_interface': 8.347194140943366e-06, 'learning_rate/action_model': 8.347194140943366e-05, 'epoch': 0.82}) +[2026-06-17 19:54:04,537][starVLA.training.train_starvla][INFO] - Step 1991, Loss: {'train/loss': 0.22037121653556824, 'train/grad_norm_pre_clip': 5.641294956207275, 'timing/data': 0.0005791047587990761, 'timing/model': 0.9721512421965599, 'learning_rate/qwen_vl_interface': 8.345554102962306e-06, 'learning_rate/action_model': 8.345554102962305e-05, 'epoch': 0.82}) +[2026-06-17 19:54:08,737][starVLA.training.train_starvla][INFO] - Step 1992, Loss: {'train/loss': 0.4696694612503052, 'train/grad_norm_pre_clip': 11.593210220336914, 'timing/data': 0.0014501307159662247, 'timing/model': 1.0688011785969138, 'learning_rate/qwen_vl_interface': 8.34391342326955e-06, 'learning_rate/action_model': 8.343913423269549e-05, 'epoch': 0.82}) +[2026-06-17 19:54:12,807][starVLA.training.train_starvla][INFO] - Step 1993, Loss: {'train/loss': 0.6366252303123474, 'train/grad_norm_pre_clip': 14.603510856628418, 'timing/data': 0.0005324874073266983, 'timing/model': 1.039104113355279, 'learning_rate/qwen_vl_interface': 8.342272102205213e-06, 'learning_rate/action_model': 8.342272102205212e-05, 'epoch': 0.82}) +[2026-06-17 19:54:16,962][starVLA.training.train_starvla][INFO] - Step 1994, Loss: {'train/loss': 0.633194625377655, 'train/grad_norm_pre_clip': 14.561792373657227, 'timing/data': 0.0006734719499945641, 'timing/model': 1.2602941412478685, 'learning_rate/qwen_vl_interface': 8.340630140109539e-06, 'learning_rate/action_model': 8.340630140109539e-05, 'epoch': 0.82}) +[2026-06-17 19:54:20,895][starVLA.training.train_starvla][INFO] - Step 1995, Loss: {'train/loss': 0.5739619731903076, 'train/grad_norm_pre_clip': 13.308452606201172, 'timing/data': 0.0010275226086378098, 'timing/model': 1.1857405146583915, 'learning_rate/qwen_vl_interface': 8.338987537322916e-06, 'learning_rate/action_model': 8.338987537322916e-05, 'epoch': 0.82}) +[2026-06-17 19:54:25,946][starVLA.training.train_starvla][INFO] - Step 1996, Loss: {'train/loss': 0.7254252433776855, 'train/grad_norm_pre_clip': 11.964674949645996, 'timing/data': 0.0006824424490332603, 'timing/model': 0.9614226007834077, 'learning_rate/qwen_vl_interface': 8.337344294185848e-06, 'learning_rate/action_model': 8.337344294185849e-05, 'epoch': 0.82}) +[2026-06-17 19:54:29,931][starVLA.training.train_starvla][INFO] - Step 1997, Loss: {'train/loss': 0.4301907420158386, 'train/grad_norm_pre_clip': 9.879775047302246, 'timing/data': 0.0016502048820257187, 'timing/model': 0.9603729899972677, 'learning_rate/qwen_vl_interface': 8.335700411038988e-06, 'learning_rate/action_model': 8.335700411038988e-05, 'epoch': 0.82}) +[2026-06-17 19:54:34,022][starVLA.training.train_starvla][INFO] - Step 1998, Loss: {'train/loss': 0.4818148910999298, 'train/grad_norm_pre_clip': 11.336310386657715, 'timing/data': 0.0005402481183409691, 'timing/model': 1.0580072309821844, 'learning_rate/qwen_vl_interface': 8.334055888223111e-06, 'learning_rate/action_model': 8.334055888223111e-05, 'epoch': 0.82}) +[2026-06-17 19:54:39,139][starVLA.training.train_starvla][INFO] - Step 1999, Loss: {'train/loss': 0.47669872641563416, 'train/grad_norm_pre_clip': 8.258983612060547, 'timing/data': 0.001079578883945942, 'timing/model': 0.9647402372211218, 'learning_rate/qwen_vl_interface': 8.332410726079132e-06, 'learning_rate/action_model': 8.332410726079132e-05, 'epoch': 0.82}) +[2026-06-17 19:56:50,754][starVLA.training.train_starvla][INFO] - Step 2000, Loss: {'train/loss': 0.6023102402687073, 'train/grad_norm_pre_clip': 14.173213958740234, 'eval/loss': 0.5931057929992676, 'eval/latency_6/loss': 0.5931057929992676, 'eval/demon_fire_precision': 0.9359850839030454, 'eval/demon_fire_recall': 0.97349709114415, 'eval/demon_fire_f1': 0.9543726235741444, 'eval/demon_fire_rate_model': 0.8953811908736784, 'eval/demon_fire_rate_teacher': 0.860879243183083, 'eval/demon_left_precision': 0.7644710578842315, 'eval/demon_left_recall': 0.6719298245614035, 'eval/demon_left_f1': 0.7152194211017739, 'eval/demon_left_rate_model': 0.13939899833055092, 'eval/demon_left_rate_teacher': 0.15859766277128548, 'eval/demon_right_precision': 0.7064606741573034, 'eval/demon_right_recall': 0.7726574500768049, 'eval/demon_right_f1': 0.7380777696258253, 'eval/demon_right_rate_model': 0.19810795770728992, 'eval/demon_right_rate_teacher': 0.18113522537562604, 'eval/demon_move_f1': 0.7266485953637996, 'eval/demon_cc_f1': 0.872356424685682, 'eval/latency_6/demon_fire_precision': 0.9359850839030454, 'eval/latency_6/demon_fire_recall': 0.97349709114415, 'eval/latency_6/demon_fire_f1': 0.9543726235741444, 'eval/latency_6/demon_fire_rate_model': 0.8953811908736784, 'eval/latency_6/demon_fire_rate_teacher': 0.860879243183083, 'eval/latency_6/demon_left_precision': 0.7644710578842315, 'eval/latency_6/demon_left_recall': 0.6719298245614035, 'eval/latency_6/demon_left_f1': 0.7152194211017739, 'eval/latency_6/demon_left_rate_model': 0.13939899833055092, 'eval/latency_6/demon_left_rate_teacher': 0.15859766277128548, 'eval/latency_6/demon_right_precision': 0.7064606741573034, 'eval/latency_6/demon_right_recall': 0.7726574500768049, 'eval/latency_6/demon_right_f1': 0.7380777696258253, 'eval/latency_6/demon_right_rate_model': 0.19810795770728992, 'eval/latency_6/demon_right_rate_teacher': 0.18113522537562604, 'eval/latency_6/demon_move_f1': 0.7266485953637996, 'eval/latency_6/demon_cc_f1': 0.872356424685682, 'timing/data': 0.0008902633562684059, 'timing/model': 0.9724645558744669, 'learning_rate/qwen_vl_interface': 8.330764924948093e-06, 'learning_rate/action_model': 8.330764924948093e-05, 'epoch': 0.82}) +[2026-06-17 19:56:50,757][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_2000_state +[2026-06-17 19:56:50,762][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 19:56:57,443][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_2000_state/model.safetensors +[2026-06-17 19:57:12,803][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_2000_state/optimizer.bin +[2026-06-17 19:57:12,803][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_2000_state/sampler.bin +[2026-06-17 19:57:12,803][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_2000_state/sampler_1.bin +[2026-06-17 19:57:12,806][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_2000_state/random_states_0.pkl +[2026-06-17 19:57:12,811][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2000 +[2026-06-17 19:57:12,811][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 19:57:12,822][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 19:57:16,928][starVLA.training.train_starvla][INFO] - Step 2001, Loss: {'train/loss': 0.1743953675031662, 'train/grad_norm_pre_clip': 8.07904052734375, 'timing/data': 0.0006039189174771309, 'timing/model': 1.0624967636540532, 'learning_rate/qwen_vl_interface': 8.329118485171167e-06, 'learning_rate/action_model': 8.329118485171167e-05, 'epoch': 0.82}) +[2026-06-17 19:57:20,901][starVLA.training.train_starvla][INFO] - Step 2002, Loss: {'train/loss': 0.5098868608474731, 'train/grad_norm_pre_clip': 9.253128051757812, 'timing/data': 0.000521589070558548, 'timing/model': 0.9497327227145433, 'learning_rate/qwen_vl_interface': 8.327471407089667e-06, 'learning_rate/action_model': 8.327471407089668e-05, 'epoch': 0.82}) +[2026-06-17 19:57:24,879][starVLA.training.train_starvla][INFO] - Step 2003, Loss: {'train/loss': 0.6210407018661499, 'train/grad_norm_pre_clip': 13.809468269348145, 'timing/data': 0.0005146162584424019, 'timing/model': 1.1221723221242428, 'learning_rate/qwen_vl_interface': 8.325823691045032e-06, 'learning_rate/action_model': 8.325823691045033e-05, 'epoch': 0.82}) +[2026-06-17 19:57:28,715][starVLA.training.train_starvla][INFO] - Step 2004, Loss: {'train/loss': 0.5308905243873596, 'train/grad_norm_pre_clip': 11.572436332702637, 'timing/data': 0.0007469085976481438, 'timing/model': 1.1246844185516238, 'learning_rate/qwen_vl_interface': 8.32417533737884e-06, 'learning_rate/action_model': 8.324175337378839e-05, 'epoch': 0.82}) +[2026-06-17 19:57:32,431][starVLA.training.train_starvla][INFO] - Step 2005, Loss: {'train/loss': 0.5760976076126099, 'train/grad_norm_pre_clip': 12.557849884033203, 'timing/data': 0.0005644308403134346, 'timing/model': 0.9780199779197574, 'learning_rate/qwen_vl_interface': 8.322526346432789e-06, 'learning_rate/action_model': 8.322526346432789e-05, 'epoch': 0.83}) +[2026-06-17 19:57:36,515][starVLA.training.train_starvla][INFO] - Step 2006, Loss: {'train/loss': 0.6757209897041321, 'train/grad_norm_pre_clip': 12.565309524536133, 'timing/data': 0.0009405985474586487, 'timing/model': 1.196629162877798, 'learning_rate/qwen_vl_interface': 8.32087671854872e-06, 'learning_rate/action_model': 8.32087671854872e-05, 'epoch': 0.83}) +[2026-06-17 19:57:41,158][starVLA.training.train_starvla][INFO] - Step 2007, Loss: {'train/loss': 0.6763486266136169, 'train/grad_norm_pre_clip': 13.046083450317383, 'timing/data': 0.00048655644059181213, 'timing/model': 0.9960219664499164, 'learning_rate/qwen_vl_interface': 8.319226454068604e-06, 'learning_rate/action_model': 8.319226454068604e-05, 'epoch': 0.83}) +[2026-06-17 19:57:45,704][starVLA.training.train_starvla][INFO] - Step 2008, Loss: {'train/loss': 0.7681810259819031, 'train/grad_norm_pre_clip': 14.683405876159668, 'timing/data': 0.0004984447732567787, 'timing/model': 0.9608686892315745, 'learning_rate/qwen_vl_interface': 8.317575553334542e-06, 'learning_rate/action_model': 8.317575553334542e-05, 'epoch': 0.83}) +[2026-06-17 19:57:50,423][starVLA.training.train_starvla][INFO] - Step 2009, Loss: {'train/loss': 0.5859500765800476, 'train/grad_norm_pre_clip': 11.591270446777344, 'timing/data': 0.0005470858886837959, 'timing/model': 1.1531505640596151, 'learning_rate/qwen_vl_interface': 8.315924016688767e-06, 'learning_rate/action_model': 8.315924016688766e-05, 'epoch': 0.83}) +[2026-06-17 19:57:54,229][starVLA.training.train_starvla][INFO] - Step 2010, Loss: {'train/loss': 0.4479745030403137, 'train/grad_norm_pre_clip': 8.306949615478516, 'timing/data': 0.0005199657753109932, 'timing/model': 1.0841418225318193, 'learning_rate/qwen_vl_interface': 8.314271844473644e-06, 'learning_rate/action_model': 8.314271844473644e-05, 'epoch': 0.83}) +[2026-06-17 19:57:58,090][starVLA.training.train_starvla][INFO] - Step 2011, Loss: {'train/loss': 0.7246114015579224, 'train/grad_norm_pre_clip': 12.890024185180664, 'timing/data': 0.001107584685087204, 'timing/model': 1.0186801655218005, 'learning_rate/qwen_vl_interface': 8.312619037031668e-06, 'learning_rate/action_model': 8.312619037031668e-05, 'epoch': 0.83}) +[2026-06-17 19:58:01,989][starVLA.training.train_starvla][INFO] - Step 2012, Loss: {'train/loss': 0.46935874223709106, 'train/grad_norm_pre_clip': 9.744316101074219, 'timing/data': 0.0008109798654913902, 'timing/model': 0.942270671017468, 'learning_rate/qwen_vl_interface': 8.310965594705473e-06, 'learning_rate/action_model': 8.310965594705472e-05, 'epoch': 0.83}) +[2026-06-17 19:58:06,026][starVLA.training.train_starvla][INFO] - Step 2013, Loss: {'train/loss': 0.6732703447341919, 'train/grad_norm_pre_clip': 11.554276466369629, 'timing/data': 0.0008816169574856758, 'timing/model': 0.9686784017831087, 'learning_rate/qwen_vl_interface': 8.309311517837814e-06, 'learning_rate/action_model': 8.309311517837813e-05, 'epoch': 0.83}) +[2026-06-17 19:58:10,011][starVLA.training.train_starvla][INFO] - Step 2014, Loss: {'train/loss': 0.5774884223937988, 'train/grad_norm_pre_clip': 12.731241226196289, 'timing/data': 0.0007655750960111618, 'timing/model': 0.9582870993763208, 'learning_rate/qwen_vl_interface': 8.307656806771587e-06, 'learning_rate/action_model': 8.307656806771588e-05, 'epoch': 0.83}) +[2026-06-17 19:58:13,981][starVLA.training.train_starvla][INFO] - Step 2015, Loss: {'train/loss': 0.5315285325050354, 'train/grad_norm_pre_clip': 94.4114761352539, 'timing/data': 0.0007368922233581543, 'timing/model': 0.9479314181953669, 'learning_rate/qwen_vl_interface': 8.306001461849814e-06, 'learning_rate/action_model': 8.306001461849814e-05, 'epoch': 0.83}) +[2026-06-17 19:58:17,870][starVLA.training.train_starvla][INFO] - Step 2016, Loss: {'train/loss': 0.49007686972618103, 'train/grad_norm_pre_clip': 14.706902503967285, 'timing/data': 0.0009140167385339737, 'timing/model': 1.172882903367281, 'learning_rate/qwen_vl_interface': 8.30434548341565e-06, 'learning_rate/action_model': 8.304345483415649e-05, 'epoch': 0.83}) +[2026-06-17 19:58:21,544][starVLA.training.train_starvla][INFO] - Step 2017, Loss: {'train/loss': 0.6343547701835632, 'train/grad_norm_pre_clip': 13.688432693481445, 'timing/data': 0.0005150279030203819, 'timing/model': 0.9372665286064148, 'learning_rate/qwen_vl_interface': 8.30268887181238e-06, 'learning_rate/action_model': 8.302688871812378e-05, 'epoch': 0.83}) +[2026-06-17 19:58:25,498][starVLA.training.train_starvla][INFO] - Step 2018, Loss: {'train/loss': 0.9447695016860962, 'train/grad_norm_pre_clip': 21.284025192260742, 'timing/data': 0.000906781293451786, 'timing/model': 1.0026449589058757, 'learning_rate/qwen_vl_interface': 8.301031627383422e-06, 'learning_rate/action_model': 8.301031627383421e-05, 'epoch': 0.83}) +[2026-06-17 19:58:29,635][starVLA.training.train_starvla][INFO] - Step 2019, Loss: {'train/loss': 0.4423857629299164, 'train/grad_norm_pre_clip': 10.888279914855957, 'timing/data': 0.000546642579138279, 'timing/model': 0.968776123598218, 'learning_rate/qwen_vl_interface': 8.299373750472323e-06, 'learning_rate/action_model': 8.299373750472324e-05, 'epoch': 0.83}) +[2026-06-17 19:58:33,673][starVLA.training.train_starvla][INFO] - Step 2020, Loss: {'train/loss': 0.6192806959152222, 'train/grad_norm_pre_clip': 12.758624076843262, 'timing/data': 0.0006987759843468666, 'timing/model': 0.9426413485780358, 'learning_rate/qwen_vl_interface': 8.297715241422765e-06, 'learning_rate/action_model': 8.297715241422766e-05, 'epoch': 0.83}) +[2026-06-17 19:58:37,621][starVLA.training.train_starvla][INFO] - Step 2021, Loss: {'train/loss': 0.5108350515365601, 'train/grad_norm_pre_clip': 10.317509651184082, 'timing/data': 0.0005301171913743019, 'timing/model': 0.9652318879961967, 'learning_rate/qwen_vl_interface': 8.29605610057856e-06, 'learning_rate/action_model': 8.29605610057856e-05, 'epoch': 0.83}) +[2026-06-17 19:58:41,563][starVLA.training.train_starvla][INFO] - Step 2022, Loss: {'train/loss': 0.532922089099884, 'train/grad_norm_pre_clip': 12.834415435791016, 'timing/data': 0.00122099369764328, 'timing/model': 1.213528429158032, 'learning_rate/qwen_vl_interface': 8.294396328283647e-06, 'learning_rate/action_model': 8.294396328283645e-05, 'epoch': 0.83}) +[2026-06-17 19:58:45,406][starVLA.training.train_starvla][INFO] - Step 2023, Loss: {'train/loss': 0.6692761182785034, 'train/grad_norm_pre_clip': 12.423276901245117, 'timing/data': 0.0009042862802743912, 'timing/model': 0.9665115280076861, 'learning_rate/qwen_vl_interface': 8.292735924882097e-06, 'learning_rate/action_model': 8.292735924882097e-05, 'epoch': 0.83}) +[2026-06-17 19:58:49,291][starVLA.training.train_starvla][INFO] - Step 2024, Loss: {'train/loss': 0.6366166472434998, 'train/grad_norm_pre_clip': 11.868149757385254, 'timing/data': 0.0010786037892103195, 'timing/model': 0.9383259536698461, 'learning_rate/qwen_vl_interface': 8.291074890718117e-06, 'learning_rate/action_model': 8.291074890718117e-05, 'epoch': 0.83}) +[2026-06-17 19:58:53,183][starVLA.training.train_starvla][INFO] - Step 2025, Loss: {'train/loss': 0.5791046619415283, 'train/grad_norm_pre_clip': 12.285585403442383, 'timing/data': 0.0007328800857067108, 'timing/model': 0.9458567649126053, 'learning_rate/qwen_vl_interface': 8.28941322613604e-06, 'learning_rate/action_model': 8.289413226136039e-05, 'epoch': 0.83}) +[2026-06-17 19:58:57,084][starVLA.training.train_starvla][INFO] - Step 2026, Loss: {'train/loss': 0.5250357985496521, 'train/grad_norm_pre_clip': 11.25147819519043, 'timing/data': 0.0005538659170269966, 'timing/model': 0.9521985575556755, 'learning_rate/qwen_vl_interface': 8.287750931480332e-06, 'learning_rate/action_model': 8.287750931480332e-05, 'epoch': 0.83}) +[2026-06-17 19:59:00,996][starVLA.training.train_starvla][INFO] - Step 2027, Loss: {'train/loss': 0.5667345523834229, 'train/grad_norm_pre_clip': 10.175860404968262, 'timing/data': 0.0005219178274273872, 'timing/model': 1.0179234091192484, 'learning_rate/qwen_vl_interface': 8.286088007095584e-06, 'learning_rate/action_model': 8.286088007095584e-05, 'epoch': 0.83}) +[2026-06-17 19:59:04,852][starVLA.training.train_starvla][INFO] - Step 2028, Loss: {'train/loss': 0.6174013614654541, 'train/grad_norm_pre_clip': 10.268033027648926, 'timing/data': 0.0005896426737308502, 'timing/model': 0.9386457456275821, 'learning_rate/qwen_vl_interface': 8.284424453326528e-06, 'learning_rate/action_model': 8.284424453326528e-05, 'epoch': 0.83}) +[2026-06-17 19:59:08,767][starVLA.training.train_starvla][INFO] - Step 2029, Loss: {'train/loss': 0.548217236995697, 'train/grad_norm_pre_clip': 11.924217224121094, 'timing/data': 0.0005034543573856354, 'timing/model': 1.187689338810742, 'learning_rate/qwen_vl_interface': 8.282760270518017e-06, 'learning_rate/action_model': 8.282760270518017e-05, 'epoch': 0.84}) +[2026-06-17 19:59:12,458][starVLA.training.train_starvla][INFO] - Step 2030, Loss: {'train/loss': 0.32028454542160034, 'train/grad_norm_pre_clip': 7.68495512008667, 'timing/data': 0.000493946485221386, 'timing/model': 0.9468564698472619, 'learning_rate/qwen_vl_interface': 8.281095459015037e-06, 'learning_rate/action_model': 8.281095459015036e-05, 'epoch': 0.84}) +[2026-06-17 19:59:16,496][starVLA.training.train_starvla][INFO] - Step 2031, Loss: {'train/loss': 0.5420997142791748, 'train/grad_norm_pre_clip': 12.878975868225098, 'timing/data': 0.000506165437400341, 'timing/model': 0.953861610032618, 'learning_rate/qwen_vl_interface': 8.279430019162709e-06, 'learning_rate/action_model': 8.279430019162708e-05, 'epoch': 0.84}) +[2026-06-17 19:59:20,407][starVLA.training.train_starvla][INFO] - Step 2032, Loss: {'train/loss': 0.5398651957511902, 'train/grad_norm_pre_clip': 9.39737606048584, 'timing/data': 0.0007450319826602936, 'timing/model': 0.9587692776694894, 'learning_rate/qwen_vl_interface': 8.277763951306277e-06, 'learning_rate/action_model': 8.277763951306277e-05, 'epoch': 0.84}) +[2026-06-17 19:59:27,337][starVLA.training.train_starvla][INFO] - Step 2033, Loss: {'train/loss': 0.5082122683525085, 'train/grad_norm_pre_clip': 11.627209663391113, 'timing/data': 0.00046226009726524353, 'timing/model': 1.008947771973908, 'learning_rate/qwen_vl_interface': 8.27609725579112e-06, 'learning_rate/action_model': 8.27609725579112e-05, 'epoch': 0.84}) +[2026-06-17 19:59:31,321][starVLA.training.train_starvla][INFO] - Step 2034, Loss: {'train/loss': 0.48284271359443665, 'train/grad_norm_pre_clip': 12.263520240783691, 'timing/data': 0.00048788171261548996, 'timing/model': 0.9668043544515967, 'learning_rate/qwen_vl_interface': 8.274429932962745e-06, 'learning_rate/action_model': 8.274429932962746e-05, 'epoch': 0.84}) +[2026-06-17 19:59:35,249][starVLA.training.train_starvla][INFO] - Step 2035, Loss: {'train/loss': 0.4828154742717743, 'train/grad_norm_pre_clip': 11.348438262939453, 'timing/data': 0.0007651699706912041, 'timing/model': 1.1693491404876113, 'learning_rate/qwen_vl_interface': 8.272761983166793e-06, 'learning_rate/action_model': 8.272761983166792e-05, 'epoch': 0.84}) +[2026-06-17 19:59:39,195][starVLA.training.train_starvla][INFO] - Step 2036, Loss: {'train/loss': 0.4075597822666168, 'train/grad_norm_pre_clip': 10.672043800354004, 'timing/data': 0.00046620145440101624, 'timing/model': 1.2242217268794775, 'learning_rate/qwen_vl_interface': 8.271093406749025e-06, 'learning_rate/action_model': 8.271093406749025e-05, 'epoch': 0.84}) +[2026-06-17 19:59:42,842][starVLA.training.train_starvla][INFO] - Step 2037, Loss: {'train/loss': 0.6557506918907166, 'train/grad_norm_pre_clip': 11.501935005187988, 'timing/data': 0.00106715876609087, 'timing/model': 0.9475114718079567, 'learning_rate/qwen_vl_interface': 8.269424204055346e-06, 'learning_rate/action_model': 8.269424204055346e-05, 'epoch': 0.84}) +[2026-06-17 19:59:46,737][starVLA.training.train_starvla][INFO] - Step 2038, Loss: {'train/loss': 0.5394653081893921, 'train/grad_norm_pre_clip': 12.689183235168457, 'timing/data': 0.0008493149653077126, 'timing/model': 0.9503469886258245, 'learning_rate/qwen_vl_interface': 8.26775437543178e-06, 'learning_rate/action_model': 8.26775437543178e-05, 'epoch': 0.84}) +[2026-06-17 19:59:50,795][starVLA.training.train_starvla][INFO] - Step 2039, Loss: {'train/loss': 0.6776275038719177, 'train/grad_norm_pre_clip': 12.759167671203613, 'timing/data': 0.0005241502076387405, 'timing/model': 1.0307542840018868, 'learning_rate/qwen_vl_interface': 8.266083921224484e-06, 'learning_rate/action_model': 8.266083921224485e-05, 'epoch': 0.84}) +[2026-06-17 19:59:54,640][starVLA.training.train_starvla][INFO] - Step 2040, Loss: {'train/loss': 0.3759714663028717, 'train/grad_norm_pre_clip': 13.519885063171387, 'timing/data': 0.0006326623260974884, 'timing/model': 0.9498476143926382, 'learning_rate/qwen_vl_interface': 8.264412841779747e-06, 'learning_rate/action_model': 8.264412841779746e-05, 'epoch': 0.84}) +[2026-06-17 19:59:58,659][starVLA.training.train_starvla][INFO] - Step 2041, Loss: {'train/loss': 0.8265770077705383, 'train/grad_norm_pre_clip': 12.432299613952637, 'timing/data': 0.0008137114346027374, 'timing/model': 0.9494957709684968, 'learning_rate/qwen_vl_interface': 8.262741137443984e-06, 'learning_rate/action_model': 8.262741137443985e-05, 'epoch': 0.84}) +[2026-06-17 20:00:02,642][starVLA.training.train_starvla][INFO] - Step 2042, Loss: {'train/loss': 0.4289867877960205, 'train/grad_norm_pre_clip': 10.315103530883789, 'timing/data': 0.0007441667839884758, 'timing/model': 1.1482417210936546, 'learning_rate/qwen_vl_interface': 8.26106880856374e-06, 'learning_rate/action_model': 8.26106880856374e-05, 'epoch': 0.84}) +[2026-06-17 20:00:06,592][starVLA.training.train_starvla][INFO] - Step 2043, Loss: {'train/loss': 0.38053303956985474, 'train/grad_norm_pre_clip': 9.685319900512695, 'timing/data': 0.0005342643707990646, 'timing/model': 1.1691529648378491, 'learning_rate/qwen_vl_interface': 8.259395855485693e-06, 'learning_rate/action_model': 8.259395855485694e-05, 'epoch': 0.84}) +[2026-06-17 20:00:10,272][starVLA.training.train_starvla][INFO] - Step 2044, Loss: {'train/loss': 0.6091669201850891, 'train/grad_norm_pre_clip': 12.879493713378906, 'timing/data': 0.001150752417743206, 'timing/model': 0.9698107382282615, 'learning_rate/qwen_vl_interface': 8.257722278556647e-06, 'learning_rate/action_model': 8.257722278556646e-05, 'epoch': 0.84}) +[2026-06-17 20:00:14,155][starVLA.training.train_starvla][INFO] - Step 2045, Loss: {'train/loss': 0.44462043046951294, 'train/grad_norm_pre_clip': 9.0286283493042, 'timing/data': 0.001429954543709755, 'timing/model': 0.951427249237895, 'learning_rate/qwen_vl_interface': 8.256048078123536e-06, 'learning_rate/action_model': 8.256048078123535e-05, 'epoch': 0.84}) +[2026-06-17 20:00:18,806][starVLA.training.train_starvla][INFO] - Step 2046, Loss: {'train/loss': 0.7202077507972717, 'train/grad_norm_pre_clip': 13.385701179504395, 'timing/data': 0.000989183783531189, 'timing/model': 1.0185301508754492, 'learning_rate/qwen_vl_interface': 8.254373254533424e-06, 'learning_rate/action_model': 8.254373254533423e-05, 'epoch': 0.84}) +[2026-06-17 20:00:23,887][starVLA.training.train_starvla][INFO] - Step 2047, Loss: {'train/loss': 0.5754228830337524, 'train/grad_norm_pre_clip': 16.273479461669922, 'timing/data': 0.0007177367806434631, 'timing/model': 1.0045221149921417, 'learning_rate/qwen_vl_interface': 8.252697808133501e-06, 'learning_rate/action_model': 8.252697808133501e-05, 'epoch': 0.84}) +[2026-06-17 20:00:27,927][starVLA.training.train_starvla][INFO] - Step 2048, Loss: {'train/loss': 0.6326836347579956, 'train/grad_norm_pre_clip': 9.998739242553711, 'timing/data': 0.0007213931530714035, 'timing/model': 0.9837100803852081, 'learning_rate/qwen_vl_interface': 8.251021739271095e-06, 'learning_rate/action_model': 8.251021739271094e-05, 'epoch': 0.84}) +[2026-06-17 20:00:31,941][starVLA.training.train_starvla][INFO] - Step 2049, Loss: {'train/loss': 0.7908704876899719, 'train/grad_norm_pre_clip': 9.109877586364746, 'timing/data': 0.0007379297167062759, 'timing/model': 1.2383740842342377, 'learning_rate/qwen_vl_interface': 8.249345048293652e-06, 'learning_rate/action_model': 8.24934504829365e-05, 'epoch': 0.84}) +[2026-06-17 20:00:35,794][starVLA.training.train_starvla][INFO] - Step 2050, Loss: {'train/loss': 0.4930039346218109, 'train/grad_norm_pre_clip': 11.565863609313965, 'timing/data': 0.0009401459246873856, 'timing/model': 1.0162441115826368, 'learning_rate/qwen_vl_interface': 8.247667735548753e-06, 'learning_rate/action_model': 8.247667735548752e-05, 'epoch': 0.84}) +[2026-06-17 20:00:39,832][starVLA.training.train_starvla][INFO] - Step 2051, Loss: {'train/loss': 0.7817572355270386, 'train/grad_norm_pre_clip': 16.038171768188477, 'timing/data': 0.0015484634786844254, 'timing/model': 0.9794016582891345, 'learning_rate/qwen_vl_interface': 8.245989801384104e-06, 'learning_rate/action_model': 8.245989801384104e-05, 'epoch': 0.84}) +[2026-06-17 20:00:43,798][starVLA.training.train_starvla][INFO] - Step 2052, Loss: {'train/loss': 0.6663885116577148, 'train/grad_norm_pre_clip': 11.195211410522461, 'timing/data': 0.0005519231781363487, 'timing/model': 0.9389620032161474, 'learning_rate/qwen_vl_interface': 8.24431124614755e-06, 'learning_rate/action_model': 8.244311246147549e-05, 'epoch': 0.84}) +[2026-06-17 20:00:47,797][starVLA.training.train_starvla][INFO] - Step 2053, Loss: {'train/loss': 0.4110208749771118, 'train/grad_norm_pre_clip': 8.203266143798828, 'timing/data': 0.0008296864107251167, 'timing/model': 0.9784696577116847, 'learning_rate/qwen_vl_interface': 8.24263207018705e-06, 'learning_rate/action_model': 8.24263207018705e-05, 'epoch': 0.85}) +[2026-06-17 20:00:51,987][starVLA.training.train_starvla][INFO] - Step 2054, Loss: {'train/loss': 0.4958571791648865, 'train/grad_norm_pre_clip': 8.916868209838867, 'timing/data': 0.0008352166041731834, 'timing/model': 0.9750029798597097, 'learning_rate/qwen_vl_interface': 8.2409522738507e-06, 'learning_rate/action_model': 8.240952273850701e-05, 'epoch': 0.85}) +[2026-06-17 20:00:57,688][starVLA.training.train_starvla][INFO] - Step 2055, Loss: {'train/loss': 0.6407402753829956, 'train/grad_norm_pre_clip': 13.597087860107422, 'timing/data': 0.0007916446775197983, 'timing/model': 1.3108073631301522, 'learning_rate/qwen_vl_interface': 8.23927185748673e-06, 'learning_rate/action_model': 8.239271857486729e-05, 'epoch': 0.85}) +[2026-06-17 20:01:02,407][starVLA.training.train_starvla][INFO] - Step 2056, Loss: {'train/loss': 0.335589736700058, 'train/grad_norm_pre_clip': 7.6327996253967285, 'timing/data': 0.0005176449194550514, 'timing/model': 1.229096875526011, 'learning_rate/qwen_vl_interface': 8.237590821443484e-06, 'learning_rate/action_model': 8.237590821443483e-05, 'epoch': 0.85}) +[2026-06-17 20:01:06,300][starVLA.training.train_starvla][INFO] - Step 2057, Loss: {'train/loss': 0.4337431490421295, 'train/grad_norm_pre_clip': 9.160160064697266, 'timing/data': 0.0005376432090997696, 'timing/model': 1.0253044124692678, 'learning_rate/qwen_vl_interface': 8.235909166069447e-06, 'learning_rate/action_model': 8.235909166069447e-05, 'epoch': 0.85}) +[2026-06-17 20:01:10,403][starVLA.training.train_starvla][INFO] - Step 2058, Loss: {'train/loss': 0.884851336479187, 'train/grad_norm_pre_clip': 15.532179832458496, 'timing/data': 0.00066376943141222, 'timing/model': 1.0303495656698942, 'learning_rate/qwen_vl_interface': 8.234226891713225e-06, 'learning_rate/action_model': 8.234226891713225e-05, 'epoch': 0.85}) +[2026-06-17 20:01:14,411][starVLA.training.train_starvla][INFO] - Step 2059, Loss: {'train/loss': 0.46373847126960754, 'train/grad_norm_pre_clip': 9.380488395690918, 'timing/data': 0.0008199987933039665, 'timing/model': 0.9578698119148612, 'learning_rate/qwen_vl_interface': 8.232543998723559e-06, 'learning_rate/action_model': 8.232543998723559e-05, 'epoch': 0.85}) +[2026-06-17 20:01:18,400][starVLA.training.train_starvla][INFO] - Step 2060, Loss: {'train/loss': 0.39179718494415283, 'train/grad_norm_pre_clip': 12.1058349609375, 'timing/data': 0.0006916355341672897, 'timing/model': 0.9713582610711455, 'learning_rate/qwen_vl_interface': 8.230860487449314e-06, 'learning_rate/action_model': 8.230860487449313e-05, 'epoch': 0.85}) +[2026-06-17 20:01:23,374][starVLA.training.train_starvla][INFO] - Step 2061, Loss: {'train/loss': 0.6846739649772644, 'train/grad_norm_pre_clip': 13.28824520111084, 'timing/data': 0.0006687138229608536, 'timing/model': 1.0232252497226, 'learning_rate/qwen_vl_interface': 8.22917635823948e-06, 'learning_rate/action_model': 8.229176358239478e-05, 'epoch': 0.85}) +[2026-06-17 20:01:28,646][starVLA.training.train_starvla][INFO] - Step 2062, Loss: {'train/loss': 0.4271782338619232, 'train/grad_norm_pre_clip': 9.532209396362305, 'timing/data': 0.0006998982280492783, 'timing/model': 1.21094929613173, 'learning_rate/qwen_vl_interface': 8.227491611443182e-06, 'learning_rate/action_model': 8.227491611443182e-05, 'epoch': 0.85}) +[2026-06-17 20:01:32,852][starVLA.training.train_starvla][INFO] - Step 2063, Loss: {'train/loss': 0.7162690758705139, 'train/grad_norm_pre_clip': 9.535612106323242, 'timing/data': 0.0010707695037126541, 'timing/model': 1.4306123843416572, 'learning_rate/qwen_vl_interface': 8.225806247409667e-06, 'learning_rate/action_model': 8.225806247409668e-05, 'epoch': 0.85}) +[2026-06-17 20:01:36,580][starVLA.training.train_starvla][INFO] - Step 2064, Loss: {'train/loss': 0.6560193300247192, 'train/grad_norm_pre_clip': 12.078893661499023, 'timing/data': 0.0013997815549373627, 'timing/model': 0.9651508750393987, 'learning_rate/qwen_vl_interface': 8.224120266488318e-06, 'learning_rate/action_model': 8.224120266488318e-05, 'epoch': 0.85}) +[2026-06-17 20:01:40,760][starVLA.training.train_starvla][INFO] - Step 2065, Loss: {'train/loss': 0.5815463662147522, 'train/grad_norm_pre_clip': 12.544058799743652, 'timing/data': 0.0006330432370305061, 'timing/model': 1.0809256555512547, 'learning_rate/qwen_vl_interface': 8.222433669028636e-06, 'learning_rate/action_model': 8.222433669028635e-05, 'epoch': 0.85}) +[2026-06-17 20:01:44,775][starVLA.training.train_starvla][INFO] - Step 2066, Loss: {'train/loss': 0.6289199590682983, 'train/grad_norm_pre_clip': 10.257807731628418, 'timing/data': 0.0006425566971302032, 'timing/model': 0.9602638492360711, 'learning_rate/qwen_vl_interface': 8.220746455380257e-06, 'learning_rate/action_model': 8.220746455380257e-05, 'epoch': 0.85}) +[2026-06-17 20:01:48,931][starVLA.training.train_starvla][INFO] - Step 2067, Loss: {'train/loss': 0.7144796848297119, 'train/grad_norm_pre_clip': 12.372321128845215, 'timing/data': 0.0006501581519842148, 'timing/model': 0.9639004161581397, 'learning_rate/qwen_vl_interface': 8.219058625892939e-06, 'learning_rate/action_model': 8.21905862589294e-05, 'epoch': 0.85}) +[2026-06-17 20:01:52,936][starVLA.training.train_starvla][INFO] - Step 2068, Loss: {'train/loss': 0.5119565725326538, 'train/grad_norm_pre_clip': 11.652153968811035, 'timing/data': 0.0008056154474616051, 'timing/model': 0.9755615349858999, 'learning_rate/qwen_vl_interface': 8.217370180916574e-06, 'learning_rate/action_model': 8.217370180916575e-05, 'epoch': 0.85}) +[2026-06-17 20:01:56,910][starVLA.training.train_starvla][INFO] - Step 2069, Loss: {'train/loss': 0.5175071358680725, 'train/grad_norm_pre_clip': 10.357918739318848, 'timing/data': 0.0006383759900927544, 'timing/model': 1.1770681301131845, 'learning_rate/qwen_vl_interface': 8.215681120801179e-06, 'learning_rate/action_model': 8.215681120801179e-05, 'epoch': 0.85}) +[2026-06-17 20:02:01,042][starVLA.training.train_starvla][INFO] - Step 2070, Loss: {'train/loss': 0.43874210119247437, 'train/grad_norm_pre_clip': 9.19539737701416, 'timing/data': 0.0010751066729426384, 'timing/model': 1.2247257251292467, 'learning_rate/qwen_vl_interface': 8.213991445896895e-06, 'learning_rate/action_model': 8.213991445896895e-05, 'epoch': 0.85}) +[2026-06-17 20:02:04,921][starVLA.training.train_starvla][INFO] - Step 2071, Loss: {'train/loss': 0.7248079180717468, 'train/grad_norm_pre_clip': 12.318044662475586, 'timing/data': 0.0006964225322008133, 'timing/model': 1.047252144664526, 'learning_rate/qwen_vl_interface': 8.212301156553995e-06, 'learning_rate/action_model': 8.212301156553995e-05, 'epoch': 0.85}) +[2026-06-17 20:02:09,211][starVLA.training.train_starvla][INFO] - Step 2072, Loss: {'train/loss': 0.3907393217086792, 'train/grad_norm_pre_clip': 11.860260009765625, 'timing/data': 0.0007212674245238304, 'timing/model': 0.9769267002120614, 'learning_rate/qwen_vl_interface': 8.210610253122879e-06, 'learning_rate/action_model': 8.210610253122879e-05, 'epoch': 0.85}) +[2026-06-17 20:02:13,432][starVLA.training.train_starvla][INFO] - Step 2073, Loss: {'train/loss': 0.7845221161842346, 'train/grad_norm_pre_clip': 9.890227317810059, 'timing/data': 0.0008762441575527191, 'timing/model': 1.0645504547283053, 'learning_rate/qwen_vl_interface': 8.20891873595407e-06, 'learning_rate/action_model': 8.20891873595407e-05, 'epoch': 0.85}) +[2026-06-17 20:02:17,575][starVLA.training.train_starvla][INFO] - Step 2074, Loss: {'train/loss': 0.8561897277832031, 'train/grad_norm_pre_clip': 17.61013412475586, 'timing/data': 0.001025545410811901, 'timing/model': 0.9814028386026621, 'learning_rate/qwen_vl_interface': 8.207226605398225e-06, 'learning_rate/action_model': 8.207226605398224e-05, 'epoch': 0.85}) +[2026-06-17 20:02:21,764][starVLA.training.train_starvla][INFO] - Step 2075, Loss: {'train/loss': 0.6071831583976746, 'train/grad_norm_pre_clip': 10.678751945495605, 'timing/data': 0.0007373904809355736, 'timing/model': 1.3873741766437888, 'learning_rate/qwen_vl_interface': 8.20553386180612e-06, 'learning_rate/action_model': 8.205533861806119e-05, 'epoch': 0.85}) +[2026-06-17 20:02:25,849][starVLA.training.train_starvla][INFO] - Step 2076, Loss: {'train/loss': 0.48521867394447327, 'train/grad_norm_pre_clip': 11.031271934509277, 'timing/data': 0.0006956364959478378, 'timing/model': 1.2321749068796635, 'learning_rate/qwen_vl_interface': 8.203840505528665e-06, 'learning_rate/action_model': 8.203840505528665e-05, 'epoch': 0.85}) +[2026-06-17 20:02:29,989][starVLA.training.train_starvla][INFO] - Step 2077, Loss: {'train/loss': 0.792107880115509, 'train/grad_norm_pre_clip': 14.223551750183105, 'timing/data': 0.0009570708498358727, 'timing/model': 1.0141780972480774, 'learning_rate/qwen_vl_interface': 8.202146536916899e-06, 'learning_rate/action_model': 8.202146536916899e-05, 'epoch': 0.85}) +[2026-06-17 20:02:34,191][starVLA.training.train_starvla][INFO] - Step 2078, Loss: {'train/loss': 0.9630030989646912, 'train/grad_norm_pre_clip': 11.479954719543457, 'timing/data': 0.0011349590495228767, 'timing/model': 1.0477873533964157, 'learning_rate/qwen_vl_interface': 8.200451956321974e-06, 'learning_rate/action_model': 8.200451956321974e-05, 'epoch': 0.86}) +[2026-06-17 20:02:38,236][starVLA.training.train_starvla][INFO] - Step 2079, Loss: {'train/loss': 0.7044219970703125, 'train/grad_norm_pre_clip': 11.221997261047363, 'timing/data': 0.0009329048916697502, 'timing/model': 0.9653650838881731, 'learning_rate/qwen_vl_interface': 8.198756764095187e-06, 'learning_rate/action_model': 8.198756764095187e-05, 'epoch': 0.86}) +[2026-06-17 20:02:42,356][starVLA.training.train_starvla][INFO] - Step 2080, Loss: {'train/loss': 0.3837745785713196, 'train/grad_norm_pre_clip': 8.751458168029785, 'timing/data': 0.001275121234357357, 'timing/model': 1.0699267657473683, 'learning_rate/qwen_vl_interface': 8.19706096058795e-06, 'learning_rate/action_model': 8.197060960587949e-05, 'epoch': 0.86}) +[2026-06-17 20:02:46,208][starVLA.training.train_starvla][INFO] - Step 2081, Loss: {'train/loss': 0.6106175184249878, 'train/grad_norm_pre_clip': 10.060062408447266, 'timing/data': 0.0013963999226689339, 'timing/model': 0.9485724596306682, 'learning_rate/qwen_vl_interface': 8.195364546151802e-06, 'learning_rate/action_model': 8.195364546151801e-05, 'epoch': 0.86}) +[2026-06-17 20:02:50,084][starVLA.training.train_starvla][INFO] - Step 2082, Loss: {'train/loss': 0.7610113024711609, 'train/grad_norm_pre_clip': 17.12919044494629, 'timing/data': 0.0010245200246572495, 'timing/model': 1.1767558278515935, 'learning_rate/qwen_vl_interface': 8.193667521138416e-06, 'learning_rate/action_model': 8.193667521138416e-05, 'epoch': 0.86}) +[2026-06-17 20:02:53,735][starVLA.training.train_starvla][INFO] - Step 2083, Loss: {'train/loss': 0.6039395332336426, 'train/grad_norm_pre_clip': 12.74735164642334, 'timing/data': 0.0008151745423674583, 'timing/model': 0.9466812321916223, 'learning_rate/qwen_vl_interface': 8.191969885899585e-06, 'learning_rate/action_model': 8.191969885899585e-05, 'epoch': 0.86}) +[2026-06-17 20:02:57,602][starVLA.training.train_starvla][INFO] - Step 2084, Loss: {'train/loss': 0.6274811625480652, 'train/grad_norm_pre_clip': 12.633190155029297, 'timing/data': 0.0005265045911073685, 'timing/model': 0.9341530175879598, 'learning_rate/qwen_vl_interface': 8.190271640787232e-06, 'learning_rate/action_model': 8.190271640787232e-05, 'epoch': 0.86}) +[2026-06-17 20:03:01,456][starVLA.training.train_starvla][INFO] - Step 2085, Loss: {'train/loss': 0.8418604135513306, 'train/grad_norm_pre_clip': 11.409282684326172, 'timing/data': 0.0005747433751821518, 'timing/model': 0.9495492586866021, 'learning_rate/qwen_vl_interface': 8.188572786153403e-06, 'learning_rate/action_model': 8.188572786153402e-05, 'epoch': 0.86}) +[2026-06-17 20:03:05,326][starVLA.training.train_starvla][INFO] - Step 2086, Loss: {'train/loss': 0.657034158706665, 'train/grad_norm_pre_clip': 11.671337127685547, 'timing/data': 0.0005612345412373543, 'timing/model': 0.938598895445466, 'learning_rate/qwen_vl_interface': 8.186873322350274e-06, 'learning_rate/action_model': 8.186873322350273e-05, 'epoch': 0.86}) +[2026-06-17 20:03:09,209][starVLA.training.train_starvla][INFO] - Step 2087, Loss: {'train/loss': 0.5874961018562317, 'train/grad_norm_pre_clip': 10.074674606323242, 'timing/data': 0.0009413929656147957, 'timing/model': 0.9461868703365326, 'learning_rate/qwen_vl_interface': 8.185173249730143e-06, 'learning_rate/action_model': 8.185173249730142e-05, 'epoch': 0.86}) +[2026-06-17 20:03:13,032][starVLA.training.train_starvla][INFO] - Step 2088, Loss: {'train/loss': 0.5077150464057922, 'train/grad_norm_pre_clip': 10.295564651489258, 'timing/data': 0.0005251076072454453, 'timing/model': 1.114046891219914, 'learning_rate/qwen_vl_interface': 8.183472568645441e-06, 'learning_rate/action_model': 8.18347256864544e-05, 'epoch': 0.86}) +[2026-06-17 20:03:16,686][starVLA.training.train_starvla][INFO] - Step 2089, Loss: {'train/loss': 0.5704161524772644, 'train/grad_norm_pre_clip': 8.43294620513916, 'timing/data': 0.0010593654587864876, 'timing/model': 0.9363373322412372, 'learning_rate/qwen_vl_interface': 8.181771279448718e-06, 'learning_rate/action_model': 8.181771279448718e-05, 'epoch': 0.86}) +[2026-06-17 20:03:20,614][starVLA.training.train_starvla][INFO] - Step 2090, Loss: {'train/loss': 0.48892492055892944, 'train/grad_norm_pre_clip': 9.154654502868652, 'timing/data': 0.0008039912208914757, 'timing/model': 0.9353934163227677, 'learning_rate/qwen_vl_interface': 8.180069382492655e-06, 'learning_rate/action_model': 8.180069382492654e-05, 'epoch': 0.86}) +[2026-06-17 20:03:24,429][starVLA.training.train_starvla][INFO] - Step 2091, Loss: {'train/loss': 0.48733651638031006, 'train/grad_norm_pre_clip': 9.126224517822266, 'timing/data': 0.0005203494802117348, 'timing/model': 0.9439298771321774, 'learning_rate/qwen_vl_interface': 8.178366878130056e-06, 'learning_rate/action_model': 8.178366878130056e-05, 'epoch': 0.86}) +[2026-06-17 20:03:28,288][starVLA.training.train_starvla][INFO] - Step 2092, Loss: {'train/loss': 0.47694888710975647, 'train/grad_norm_pre_clip': 9.109634399414062, 'timing/data': 0.0009462377056479454, 'timing/model': 0.9538594055920839, 'learning_rate/qwen_vl_interface': 8.176663766713852e-06, 'learning_rate/action_model': 8.176663766713852e-05, 'epoch': 0.86}) +[2026-06-17 20:03:32,266][starVLA.training.train_starvla][INFO] - Step 2093, Loss: {'train/loss': 0.6040961146354675, 'train/grad_norm_pre_clip': 23.783891677856445, 'timing/data': 0.0008545592427253723, 'timing/model': 0.9488012855872512, 'learning_rate/qwen_vl_interface': 8.174960048597099e-06, 'learning_rate/action_model': 8.174960048597099e-05, 'epoch': 0.86}) +[2026-06-17 20:03:36,193][starVLA.training.train_starvla][INFO] - Step 2094, Loss: {'train/loss': 0.5601958632469177, 'train/grad_norm_pre_clip': 11.299112319946289, 'timing/data': 0.0005913656204938889, 'timing/model': 0.9625794105231762, 'learning_rate/qwen_vl_interface': 8.173255724132982e-06, 'learning_rate/action_model': 8.17325572413298e-05, 'epoch': 0.86}) +[2026-06-17 20:03:40,320][starVLA.training.train_starvla][INFO] - Step 2095, Loss: {'train/loss': 0.26986145973205566, 'train/grad_norm_pre_clip': 7.35410213470459, 'timing/data': 0.00133469607681036, 'timing/model': 1.1822141343727708, 'learning_rate/qwen_vl_interface': 8.171550793674806e-06, 'learning_rate/action_model': 8.171550793674805e-05, 'epoch': 0.86}) +[2026-06-17 20:03:44,175][starVLA.training.train_starvla][INFO] - Step 2096, Loss: {'train/loss': 0.5900068283081055, 'train/grad_norm_pre_clip': 10.12014102935791, 'timing/data': 0.0008096806704998016, 'timing/model': 1.0167779717594385, 'learning_rate/qwen_vl_interface': 8.169845257576008e-06, 'learning_rate/action_model': 8.169845257576007e-05, 'epoch': 0.86}) +[2026-06-17 20:03:48,250][starVLA.training.train_starvla][INFO] - Step 2097, Loss: {'train/loss': 0.7587155699729919, 'train/grad_norm_pre_clip': 10.576899528503418, 'timing/data': 0.0010610604658722878, 'timing/model': 0.9536152807995677, 'learning_rate/qwen_vl_interface': 8.168139116190146e-06, 'learning_rate/action_model': 8.168139116190146e-05, 'epoch': 0.86}) +[2026-06-17 20:03:52,498][starVLA.training.train_starvla][INFO] - Step 2098, Loss: {'train/loss': 0.6312779188156128, 'train/grad_norm_pre_clip': 11.14371109008789, 'timing/data': 0.0005283085629343987, 'timing/model': 1.0183953437954187, 'learning_rate/qwen_vl_interface': 8.166432369870905e-06, 'learning_rate/action_model': 8.166432369870904e-05, 'epoch': 0.86}) +[2026-06-17 20:03:56,437][starVLA.training.train_starvla][INFO] - Step 2099, Loss: {'train/loss': 0.5133888721466064, 'train/grad_norm_pre_clip': 11.836092948913574, 'timing/data': 0.0005286820232868195, 'timing/model': 0.9661451745778322, 'learning_rate/qwen_vl_interface': 8.164725018972096e-06, 'learning_rate/action_model': 8.164725018972096e-05, 'epoch': 0.86}) +[2026-06-17 20:06:22,502][starVLA.training.train_starvla][INFO] - Step 2100, Loss: {'train/loss': 0.38344284892082214, 'train/grad_norm_pre_clip': 9.656116485595703, 'eval/loss': 0.6006090044975281, 'eval/latency_6/loss': 0.6006090044975281, 'eval/demon_fire_precision': 0.9333126550868487, 'eval/demon_fire_recall': 0.9725274725274725, 'eval/demon_fire_f1': 0.9525166191832858, 'eval/demon_fire_rate_model': 0.8970506399554814, 'eval/demon_fire_rate_teacher': 0.860879243183083, 'eval/demon_left_precision': 0.8179775280898877, 'eval/demon_left_recall': 0.6385964912280702, 'eval/demon_left_f1': 0.7172413793103448, 'eval/demon_left_rate_model': 0.1238174735670562, 'eval/demon_left_rate_teacher': 0.15859766277128548, 'eval/demon_right_precision': 0.8412698412698413, 'eval/demon_right_recall': 0.5698924731182796, 'eval/demon_right_f1': 0.6794871794871795, 'eval/demon_right_rate_model': 0.12270450751252086, 'eval/demon_right_rate_teacher': 0.18113522537562604, 'eval/demon_move_f1': 0.6983642793987621, 'eval/demon_cc_f1': 0.8587593736212812, 'eval/latency_6/demon_fire_precision': 0.9333126550868487, 'eval/latency_6/demon_fire_recall': 0.9725274725274725, 'eval/latency_6/demon_fire_f1': 0.9525166191832858, 'eval/latency_6/demon_fire_rate_model': 0.8970506399554814, 'eval/latency_6/demon_fire_rate_teacher': 0.860879243183083, 'eval/latency_6/demon_left_precision': 0.8179775280898877, 'eval/latency_6/demon_left_recall': 0.6385964912280702, 'eval/latency_6/demon_left_f1': 0.7172413793103448, 'eval/latency_6/demon_left_rate_model': 0.1238174735670562, 'eval/latency_6/demon_left_rate_teacher': 0.15859766277128548, 'eval/latency_6/demon_right_precision': 0.8412698412698413, 'eval/latency_6/demon_right_recall': 0.5698924731182796, 'eval/latency_6/demon_right_f1': 0.6794871794871795, 'eval/latency_6/demon_right_rate_model': 0.12270450751252086, 'eval/latency_6/demon_right_rate_teacher': 0.18113522537562604, 'eval/latency_6/demon_move_f1': 0.6983642793987621, 'eval/latency_6/demon_cc_f1': 0.8587593736212812, 'timing/data': 0.0005091950297355652, 'timing/model': 1.1645081546157598, 'learning_rate/qwen_vl_interface': 8.163017063847654e-06, 'learning_rate/action_model': 8.163017063847654e-05, 'epoch': 0.86}) +[2026-06-17 20:06:22,506][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_2100_state +[2026-06-17 20:06:22,511][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 20:06:29,070][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_2100_state/model.safetensors +[2026-06-17 20:06:44,446][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_2100_state/optimizer.bin +[2026-06-17 20:06:44,447][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_2100_state/sampler.bin +[2026-06-17 20:06:44,447][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_2100_state/sampler_1.bin +[2026-06-17 20:06:44,449][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_2100_state/random_states_0.pkl +[2026-06-17 20:06:44,453][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2100 +[2026-06-17 20:06:44,453][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 20:06:44,467][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 20:06:50,797][starVLA.training.train_starvla][INFO] - Step 2101, Loss: {'train/loss': 0.48725321888923645, 'train/grad_norm_pre_clip': 10.179007530212402, 'timing/data': 0.0007469765841960907, 'timing/model': 0.9512315634638071, 'learning_rate/qwen_vl_interface': 8.16130850485164e-06, 'learning_rate/action_model': 8.161308504851639e-05, 'epoch': 0.86}) +[2026-06-17 20:06:54,821][starVLA.training.train_starvla][INFO] - Step 2102, Loss: {'train/loss': 0.7884722948074341, 'train/grad_norm_pre_clip': 17.579641342163086, 'timing/data': 0.0005202163010835648, 'timing/model': 0.9529943382367492, 'learning_rate/qwen_vl_interface': 8.15959934233824e-06, 'learning_rate/action_model': 8.15959934233824e-05, 'epoch': 0.87}) +[2026-06-17 20:06:58,628][starVLA.training.train_starvla][INFO] - Step 2103, Loss: {'train/loss': 0.41431355476379395, 'train/grad_norm_pre_clip': 9.916229248046875, 'timing/data': 0.0005199508741497993, 'timing/model': 1.1206027418375015, 'learning_rate/qwen_vl_interface': 8.157889576661765e-06, 'learning_rate/action_model': 8.157889576661764e-05, 'epoch': 0.87}) +[2026-06-17 20:07:02,530][starVLA.training.train_starvla][INFO] - Step 2104, Loss: {'train/loss': 0.7863548398017883, 'train/grad_norm_pre_clip': 13.576281547546387, 'timing/data': 0.0005217725411057472, 'timing/model': 1.2057971712201834, 'learning_rate/qwen_vl_interface': 8.156179208176652e-06, 'learning_rate/action_model': 8.156179208176652e-05, 'epoch': 0.87}) +[2026-06-17 20:07:06,454][starVLA.training.train_starvla][INFO] - Step 2105, Loss: {'train/loss': 0.6464212536811829, 'train/grad_norm_pre_clip': 11.97659683227539, 'timing/data': 0.0009551960974931717, 'timing/model': 1.205012284219265, 'learning_rate/qwen_vl_interface': 8.154468237237463e-06, 'learning_rate/action_model': 8.154468237237462e-05, 'epoch': 0.87}) +[2026-06-17 20:07:11,058][starVLA.training.train_starvla][INFO] - Step 2106, Loss: {'train/loss': 0.7319570779800415, 'train/grad_norm_pre_clip': 11.684463500976562, 'timing/data': 0.0011767121031880379, 'timing/model': 1.1837242376059294, 'learning_rate/qwen_vl_interface': 8.152756664198882e-06, 'learning_rate/action_model': 8.152756664198882e-05, 'epoch': 0.87}) +[2026-06-17 20:07:15,272][starVLA.training.train_starvla][INFO] - Step 2107, Loss: {'train/loss': 0.4407470226287842, 'train/grad_norm_pre_clip': 15.08420181274414, 'timing/data': 0.0009862901642918587, 'timing/model': 0.9480350269004703, 'learning_rate/qwen_vl_interface': 8.151044489415722e-06, 'learning_rate/action_model': 8.151044489415722e-05, 'epoch': 0.87}) +[2026-06-17 20:07:19,329][starVLA.training.train_starvla][INFO] - Step 2108, Loss: {'train/loss': 0.5298742055892944, 'train/grad_norm_pre_clip': 12.605754852294922, 'timing/data': 0.0008640093728899956, 'timing/model': 0.9911277899518609, 'learning_rate/qwen_vl_interface': 8.149331713242916e-06, 'learning_rate/action_model': 8.149331713242916e-05, 'epoch': 0.87}) +[2026-06-17 20:07:25,307][starVLA.training.train_starvla][INFO] - Step 2109, Loss: {'train/loss': 0.6746922731399536, 'train/grad_norm_pre_clip': 11.878070831298828, 'timing/data': 0.0007200445979833603, 'timing/model': 2.6757784942165017, 'learning_rate/qwen_vl_interface': 8.147618336035526e-06, 'learning_rate/action_model': 8.147618336035526e-05, 'epoch': 0.87}) +[2026-06-17 20:07:29,340][starVLA.training.train_starvla][INFO] - Step 2110, Loss: {'train/loss': 0.7611711621284485, 'train/grad_norm_pre_clip': 14.964618682861328, 'timing/data': 0.0005402546375989914, 'timing/model': 1.2097417702898383, 'learning_rate/qwen_vl_interface': 8.14590435814874e-06, 'learning_rate/action_model': 8.14590435814874e-05, 'epoch': 0.87}) +[2026-06-17 20:07:34,435][starVLA.training.train_starvla][INFO] - Step 2111, Loss: {'train/loss': 0.6767382621765137, 'train/grad_norm_pre_clip': 13.191746711730957, 'timing/data': 0.0005748216062784195, 'timing/model': 1.1872721826657653, 'learning_rate/qwen_vl_interface': 8.144189779937862e-06, 'learning_rate/action_model': 8.144189779937862e-05, 'epoch': 0.87}) +[2026-06-17 20:07:38,673][starVLA.training.train_starvla][INFO] - Step 2112, Loss: {'train/loss': 0.9031326770782471, 'train/grad_norm_pre_clip': 17.517099380493164, 'timing/data': 0.0009853411465883255, 'timing/model': 1.1643557343631983, 'learning_rate/qwen_vl_interface': 8.142474601758331e-06, 'learning_rate/action_model': 8.142474601758329e-05, 'epoch': 0.87}) +[2026-06-17 20:07:43,269][starVLA.training.train_starvla][INFO] - Step 2113, Loss: {'train/loss': 0.6473134160041809, 'train/grad_norm_pre_clip': 13.382546424865723, 'timing/data': 0.0005443952977657318, 'timing/model': 0.9546835413202643, 'learning_rate/qwen_vl_interface': 8.1407588239657e-06, 'learning_rate/action_model': 8.140758823965698e-05, 'epoch': 0.87}) +[2026-06-17 20:07:47,510][starVLA.training.train_starvla][INFO] - Step 2114, Loss: {'train/loss': 0.5036612749099731, 'train/grad_norm_pre_clip': 16.888460159301758, 'timing/data': 0.0011670542880892754, 'timing/model': 1.0401592878624797, 'learning_rate/qwen_vl_interface': 8.139042446915653e-06, 'learning_rate/action_model': 8.139042446915654e-05, 'epoch': 0.87}) +[2026-06-17 20:07:53,057][starVLA.training.train_starvla][INFO] - Step 2115, Loss: {'train/loss': 0.6253913640975952, 'train/grad_norm_pre_clip': 13.547287940979004, 'timing/data': 0.0012204376980662346, 'timing/model': 0.9388312585651875, 'learning_rate/qwen_vl_interface': 8.137325470964001e-06, 'learning_rate/action_model': 8.137325470964e-05, 'epoch': 0.87}) +[2026-06-17 20:07:57,067][starVLA.training.train_starvla][INFO] - Step 2116, Loss: {'train/loss': 0.5762268304824829, 'train/grad_norm_pre_clip': 12.486645698547363, 'timing/data': 0.0007006069645285606, 'timing/model': 0.9428815646097064, 'learning_rate/qwen_vl_interface': 8.135607896466669e-06, 'learning_rate/action_model': 8.13560789646667e-05, 'epoch': 0.87}) +[2026-06-17 20:08:01,087][starVLA.training.train_starvla][INFO] - Step 2117, Loss: {'train/loss': 0.6109685301780701, 'train/grad_norm_pre_clip': 12.737844467163086, 'timing/data': 0.0007690684869885445, 'timing/model': 1.2818730156868696, 'learning_rate/qwen_vl_interface': 8.133889723779718e-06, 'learning_rate/action_model': 8.133889723779717e-05, 'epoch': 0.87}) +[2026-06-17 20:08:05,016][starVLA.training.train_starvla][INFO] - Step 2118, Loss: {'train/loss': 0.6577268838882446, 'train/grad_norm_pre_clip': 14.097930908203125, 'timing/data': 0.0005494002252817154, 'timing/model': 1.1997480420395732, 'learning_rate/qwen_vl_interface': 8.132170953259321e-06, 'learning_rate/action_model': 8.132170953259322e-05, 'epoch': 0.87}) +[2026-06-17 20:08:08,947][starVLA.training.train_starvla][INFO] - Step 2119, Loss: {'train/loss': 0.7648077607154846, 'train/grad_norm_pre_clip': 14.115335464477539, 'timing/data': 0.0007188916206359863, 'timing/model': 0.9574862755835056, 'learning_rate/qwen_vl_interface': 8.130451585261786e-06, 'learning_rate/action_model': 8.130451585261786e-05, 'epoch': 0.87}) +[2026-06-17 20:08:15,200][starVLA.training.train_starvla][INFO] - Step 2120, Loss: {'train/loss': 0.43582502007484436, 'train/grad_norm_pre_clip': 11.36363410949707, 'timing/data': 0.0008449461311101913, 'timing/model': 0.965263245627284, 'learning_rate/qwen_vl_interface': 8.128731620143538e-06, 'learning_rate/action_model': 8.128731620143538e-05, 'epoch': 0.87}) +[2026-06-17 20:08:19,086][starVLA.training.train_starvla][INFO] - Step 2121, Loss: {'train/loss': 0.21801087260246277, 'train/grad_norm_pre_clip': 7.428999900817871, 'timing/data': 0.0006232587620615959, 'timing/model': 0.9494193242862821, 'learning_rate/qwen_vl_interface': 8.127011058261129e-06, 'learning_rate/action_model': 8.127011058261128e-05, 'epoch': 0.87}) +[2026-06-17 20:08:23,532][starVLA.training.train_starvla][INFO] - Step 2122, Loss: {'train/loss': 0.9136070609092712, 'train/grad_norm_pre_clip': 24.640457153320312, 'timing/data': 0.0006475001573562622, 'timing/model': 0.9394935928285122, 'learning_rate/qwen_vl_interface': 8.125289899971229e-06, 'learning_rate/action_model': 8.125289899971228e-05, 'epoch': 0.87}) +[2026-06-17 20:08:27,457][starVLA.training.train_starvla][INFO] - Step 2123, Loss: {'train/loss': 0.41362160444259644, 'train/grad_norm_pre_clip': 10.497673988342285, 'timing/data': 0.0005175229161977768, 'timing/model': 0.9349577436223626, 'learning_rate/qwen_vl_interface': 8.123568145630641e-06, 'learning_rate/action_model': 8.12356814563064e-05, 'epoch': 0.87}) +[2026-06-17 20:08:31,356][starVLA.training.train_starvla][INFO] - Step 2124, Loss: {'train/loss': 0.287884384393692, 'train/grad_norm_pre_clip': 8.698490142822266, 'timing/data': 0.000612935982644558, 'timing/model': 1.1464717090129852, 'learning_rate/qwen_vl_interface': 8.121845795596282e-06, 'learning_rate/action_model': 8.121845795596282e-05, 'epoch': 0.87}) +[2026-06-17 20:08:35,513][starVLA.training.train_starvla][INFO] - Step 2125, Loss: {'train/loss': 0.47759172320365906, 'train/grad_norm_pre_clip': 11.288655281066895, 'timing/data': 0.0014374079182744026, 'timing/model': 1.1264622490853071, 'learning_rate/qwen_vl_interface': 8.1201228502252e-06, 'learning_rate/action_model': 8.120122850225199e-05, 'epoch': 0.87}) +[2026-06-17 20:08:39,161][starVLA.training.train_starvla][INFO] - Step 2126, Loss: {'train/loss': 0.6058235168457031, 'train/grad_norm_pre_clip': 11.93376350402832, 'timing/data': 0.00048484280705451965, 'timing/model': 0.9866844983771443, 'learning_rate/qwen_vl_interface': 8.118399309874563e-06, 'learning_rate/action_model': 8.118399309874563e-05, 'epoch': 0.88}) +[2026-06-17 20:08:43,699][starVLA.training.train_starvla][INFO] - Step 2127, Loss: {'train/loss': 0.5161870121955872, 'train/grad_norm_pre_clip': 26.057907104492188, 'timing/data': 0.0008761892095208168, 'timing/model': 0.9377205092459917, 'learning_rate/qwen_vl_interface': 8.116675174901663e-06, 'learning_rate/action_model': 8.116675174901662e-05, 'epoch': 0.88}) +[2026-06-17 20:08:47,629][starVLA.training.train_starvla][INFO] - Step 2128, Loss: {'train/loss': 0.5948293209075928, 'train/grad_norm_pre_clip': 13.069735527038574, 'timing/data': 0.0005449140444397926, 'timing/model': 0.9426080975681543, 'learning_rate/qwen_vl_interface': 8.114950445663915e-06, 'learning_rate/action_model': 8.114950445663913e-05, 'epoch': 0.88}) +[2026-06-17 20:08:51,647][starVLA.training.train_starvla][INFO] - Step 2129, Loss: {'train/loss': 0.517768383026123, 'train/grad_norm_pre_clip': 13.608258247375488, 'timing/data': 0.0012758010998368263, 'timing/model': 0.9550933763384819, 'learning_rate/qwen_vl_interface': 8.113225122518854e-06, 'learning_rate/action_model': 8.113225122518854e-05, 'epoch': 0.88}) +[2026-06-17 20:08:55,655][starVLA.training.train_starvla][INFO] - Step 2130, Loss: {'train/loss': 0.6351035833358765, 'train/grad_norm_pre_clip': 13.524880409240723, 'timing/data': 0.000559704378247261, 'timing/model': 1.290094105526805, 'learning_rate/qwen_vl_interface': 8.111499205824148e-06, 'learning_rate/action_model': 8.111499205824147e-05, 'epoch': 0.88}) +[2026-06-17 20:08:59,546][starVLA.training.train_starvla][INFO] - Step 2131, Loss: {'train/loss': 0.5703592300415039, 'train/grad_norm_pre_clip': 10.066483497619629, 'timing/data': 0.0005248542875051498, 'timing/model': 1.110027845017612, 'learning_rate/qwen_vl_interface': 8.109772695937575e-06, 'learning_rate/action_model': 8.109772695937574e-05, 'epoch': 0.88}) +[2026-06-17 20:09:03,286][starVLA.training.train_starvla][INFO] - Step 2132, Loss: {'train/loss': 0.3764425218105316, 'train/grad_norm_pre_clip': 10.021952629089355, 'timing/data': 0.0005439966917037964, 'timing/model': 0.9557387242093682, 'learning_rate/qwen_vl_interface': 8.108045593217045e-06, 'learning_rate/action_model': 8.108045593217045e-05, 'epoch': 0.88}) +[2026-06-17 20:09:07,435][starVLA.training.train_starvla][INFO] - Step 2133, Loss: {'train/loss': 0.637437641620636, 'train/grad_norm_pre_clip': 10.997734069824219, 'timing/data': 0.0005296682938933372, 'timing/model': 1.0562434801831841, 'learning_rate/qwen_vl_interface': 8.106317898020589e-06, 'learning_rate/action_model': 8.106317898020589e-05, 'epoch': 0.88}) +[2026-06-17 20:09:11,425][starVLA.training.train_starvla][INFO] - Step 2134, Loss: {'train/loss': 0.9159635305404663, 'train/grad_norm_pre_clip': 14.586689949035645, 'timing/data': 0.0005707228556275368, 'timing/model': 0.9625387890264392, 'learning_rate/qwen_vl_interface': 8.104589610706359e-06, 'learning_rate/action_model': 8.104589610706359e-05, 'epoch': 0.88}) +[2026-06-17 20:09:15,826][starVLA.training.train_starvla][INFO] - Step 2135, Loss: {'train/loss': 0.3579868972301483, 'train/grad_norm_pre_clip': 10.43656063079834, 'timing/data': 0.0006186459213495255, 'timing/model': 1.3127439469099045, 'learning_rate/qwen_vl_interface': 8.102860731632632e-06, 'learning_rate/action_model': 8.102860731632631e-05, 'epoch': 0.88}) +[2026-06-17 20:09:19,804][starVLA.training.train_starvla][INFO] - Step 2136, Loss: {'train/loss': 0.7240557670593262, 'train/grad_norm_pre_clip': 15.096734046936035, 'timing/data': 0.0008766278624534607, 'timing/model': 0.9601364331319928, 'learning_rate/qwen_vl_interface': 8.101131261157802e-06, 'learning_rate/action_model': 8.101131261157802e-05, 'epoch': 0.88}) +[2026-06-17 20:09:23,760][starVLA.training.train_starvla][INFO] - Step 2137, Loss: {'train/loss': 0.729643702507019, 'train/grad_norm_pre_clip': 13.260430335998535, 'timing/data': 0.0006234161555767059, 'timing/model': 1.2216000370681286, 'learning_rate/qwen_vl_interface': 8.099401199640397e-06, 'learning_rate/action_model': 8.099401199640397e-05, 'epoch': 0.88}) +[2026-06-17 20:09:27,645][starVLA.training.train_starvla][INFO] - Step 2138, Loss: {'train/loss': 0.8310203552246094, 'train/grad_norm_pre_clip': 16.139427185058594, 'timing/data': 0.0009089410305023193, 'timing/model': 0.9474005214869976, 'learning_rate/qwen_vl_interface': 8.097670547439057e-06, 'learning_rate/action_model': 8.097670547439058e-05, 'epoch': 0.88}) +[2026-06-17 20:09:31,677][starVLA.training.train_starvla][INFO] - Step 2139, Loss: {'train/loss': 0.5371533036231995, 'train/grad_norm_pre_clip': 12.586653709411621, 'timing/data': 0.001127125695347786, 'timing/model': 1.024066916666925, 'learning_rate/qwen_vl_interface': 8.095939304912549e-06, 'learning_rate/action_model': 8.095939304912548e-05, 'epoch': 0.88}) +[2026-06-17 20:09:35,682][starVLA.training.train_starvla][INFO] - Step 2140, Loss: {'train/loss': 0.6434499025344849, 'train/grad_norm_pre_clip': 12.239970207214355, 'timing/data': 0.0005553262308239937, 'timing/model': 0.9530164673924446, 'learning_rate/qwen_vl_interface': 8.094207472419759e-06, 'learning_rate/action_model': 8.094207472419759e-05, 'epoch': 0.88}) +[2026-06-17 20:09:39,570][starVLA.training.train_starvla][INFO] - Step 2141, Loss: {'train/loss': 0.33808547258377075, 'train/grad_norm_pre_clip': 10.204693794250488, 'timing/data': 0.0011590495705604553, 'timing/model': 0.9759078677743673, 'learning_rate/qwen_vl_interface': 8.092475050319699e-06, 'learning_rate/action_model': 8.092475050319698e-05, 'epoch': 0.88}) +[2026-06-17 20:09:43,526][starVLA.training.train_starvla][INFO] - Step 2142, Loss: {'train/loss': 0.5446242690086365, 'train/grad_norm_pre_clip': 14.846697807312012, 'timing/data': 0.0005456497892737389, 'timing/model': 0.9791538445279002, 'learning_rate/qwen_vl_interface': 8.090742038971503e-06, 'learning_rate/action_model': 8.090742038971503e-05, 'epoch': 0.88}) +[2026-06-17 20:09:47,530][starVLA.training.train_starvla][INFO] - Step 2143, Loss: {'train/loss': 0.3892222046852112, 'train/grad_norm_pre_clip': 9.322113037109375, 'timing/data': 0.0006784610450267792, 'timing/model': 1.1817844957113266, 'learning_rate/qwen_vl_interface': 8.089008438734427e-06, 'learning_rate/action_model': 8.089008438734427e-05, 'epoch': 0.88}) +[2026-06-17 20:09:51,275][starVLA.training.train_starvla][INFO] - Step 2144, Loss: {'train/loss': 0.4865315556526184, 'train/grad_norm_pre_clip': 10.18131160736084, 'timing/data': 0.0008123219013214111, 'timing/model': 0.967003065161407, 'learning_rate/qwen_vl_interface': 8.087274249967844e-06, 'learning_rate/action_model': 8.087274249967845e-05, 'epoch': 0.88}) +[2026-06-17 20:09:55,257][starVLA.training.train_starvla][INFO] - Step 2145, Loss: {'train/loss': 0.42038407921791077, 'train/grad_norm_pre_clip': 8.6281099319458, 'timing/data': 0.0005490966141223907, 'timing/model': 0.9518889728933573, 'learning_rate/qwen_vl_interface': 8.08553947303126e-06, 'learning_rate/action_model': 8.085539473031259e-05, 'epoch': 0.88}) +[2026-06-17 20:09:59,214][starVLA.training.train_starvla][INFO] - Step 2146, Loss: {'train/loss': 0.6781948804855347, 'train/grad_norm_pre_clip': 15.07091236114502, 'timing/data': 0.00124267116189003, 'timing/model': 1.0158102754503489, 'learning_rate/qwen_vl_interface': 8.083804108284288e-06, 'learning_rate/action_model': 8.083804108284287e-05, 'epoch': 0.88}) +[2026-06-17 20:10:03,074][starVLA.training.train_starvla][INFO] - Step 2147, Loss: {'train/loss': 0.6268749237060547, 'train/grad_norm_pre_clip': 34.48862075805664, 'timing/data': 0.0005045812577009201, 'timing/model': 0.9429205842316151, 'learning_rate/qwen_vl_interface': 8.082068156086675e-06, 'learning_rate/action_model': 8.082068156086675e-05, 'epoch': 0.88}) +[2026-06-17 20:10:06,947][starVLA.training.train_starvla][INFO] - Step 2148, Loss: {'train/loss': 0.6023487448692322, 'train/grad_norm_pre_clip': 12.375986099243164, 'timing/data': 0.0004828600212931633, 'timing/model': 0.9787401184439659, 'learning_rate/qwen_vl_interface': 8.080331616798289e-06, 'learning_rate/action_model': 8.080331616798287e-05, 'epoch': 0.88}) +[2026-06-17 20:10:10,927][starVLA.training.train_starvla][INFO] - Step 2149, Loss: {'train/loss': 0.7417754530906677, 'train/grad_norm_pre_clip': 13.490866661071777, 'timing/data': 0.0006643123924732208, 'timing/model': 0.9595822552219033, 'learning_rate/qwen_vl_interface': 8.078594490779112e-06, 'learning_rate/action_model': 8.078594490779111e-05, 'epoch': 0.88}) +[2026-06-17 20:10:15,948][starVLA.training.train_starvla][INFO] - Step 2150, Loss: {'train/loss': 0.6399482488632202, 'train/grad_norm_pre_clip': 11.623403549194336, 'timing/data': 0.0004865769296884537, 'timing/model': 1.1308059226721525, 'learning_rate/qwen_vl_interface': 8.076856778389253e-06, 'learning_rate/action_model': 8.076856778389252e-05, 'epoch': 0.89}) +[2026-06-17 20:10:19,722][starVLA.training.train_starvla][INFO] - Step 2151, Loss: {'train/loss': 0.4548753499984741, 'train/grad_norm_pre_clip': 17.416244506835938, 'timing/data': 0.0007785046473145485, 'timing/model': 1.0283282939344645, 'learning_rate/qwen_vl_interface': 8.075118479988942e-06, 'learning_rate/action_model': 8.075118479988941e-05, 'epoch': 0.89}) +[2026-06-17 20:10:23,545][starVLA.training.train_starvla][INFO] - Step 2152, Loss: {'train/loss': 0.5047752261161804, 'train/grad_norm_pre_clip': 11.945335388183594, 'timing/data': 0.0007575331255793571, 'timing/model': 0.956032826565206, 'learning_rate/qwen_vl_interface': 8.07337959593853e-06, 'learning_rate/action_model': 8.073379595938529e-05, 'epoch': 0.89}) +[2026-06-17 20:10:27,436][starVLA.training.train_starvla][INFO] - Step 2153, Loss: {'train/loss': 0.6172167062759399, 'train/grad_norm_pre_clip': 10.213828086853027, 'timing/data': 0.0008724033832550049, 'timing/model': 0.9443211257457733, 'learning_rate/qwen_vl_interface': 8.071640126598488e-06, 'learning_rate/action_model': 8.071640126598488e-05, 'epoch': 0.89}) +[2026-06-17 20:10:31,337][starVLA.training.train_starvla][INFO] - Step 2154, Loss: {'train/loss': 0.4712009131908417, 'train/grad_norm_pre_clip': 13.001822471618652, 'timing/data': 0.0009538242593407631, 'timing/model': 0.9511730633676052, 'learning_rate/qwen_vl_interface': 8.069900072329414e-06, 'learning_rate/action_model': 8.069900072329415e-05, 'epoch': 0.89}) +[2026-06-17 20:10:35,255][starVLA.training.train_starvla][INFO] - Step 2155, Loss: {'train/loss': 0.5887786746025085, 'train/grad_norm_pre_clip': 12.791645050048828, 'timing/data': 0.0005268631502985954, 'timing/model': 0.9429327510297298, 'learning_rate/qwen_vl_interface': 8.068159433492021e-06, 'learning_rate/action_model': 8.068159433492021e-05, 'epoch': 0.89}) +[2026-06-17 20:10:39,296][starVLA.training.train_starvla][INFO] - Step 2156, Loss: {'train/loss': 0.419138103723526, 'train/grad_norm_pre_clip': 8.829463005065918, 'timing/data': 0.0009826067835092545, 'timing/model': 1.1852245489135385, 'learning_rate/qwen_vl_interface': 8.066418210447144e-06, 'learning_rate/action_model': 8.066418210447143e-05, 'epoch': 0.89}) +[2026-06-17 20:10:43,173][starVLA.training.train_starvla][INFO] - Step 2157, Loss: {'train/loss': 0.3981937766075134, 'train/grad_norm_pre_clip': 9.978862762451172, 'timing/data': 0.0010722121223807335, 'timing/model': 1.1268611187115312, 'learning_rate/qwen_vl_interface': 8.064676403555741e-06, 'learning_rate/action_model': 8.064676403555741e-05, 'epoch': 0.89}) +[2026-06-17 20:10:46,947][starVLA.training.train_starvla][INFO] - Step 2158, Loss: {'train/loss': 0.3000944256782532, 'train/grad_norm_pre_clip': 9.034035682678223, 'timing/data': 0.001195157878100872, 'timing/model': 0.9432504130527377, 'learning_rate/qwen_vl_interface': 8.06293401317889e-06, 'learning_rate/action_model': 8.062934013178891e-05, 'epoch': 0.89}) +[2026-06-17 20:10:50,888][starVLA.training.train_starvla][INFO] - Step 2159, Loss: {'train/loss': 0.9699928760528564, 'train/grad_norm_pre_clip': 22.498016357421875, 'timing/data': 0.0006695138290524483, 'timing/model': 0.938559977337718, 'learning_rate/qwen_vl_interface': 8.061191039677793e-06, 'learning_rate/action_model': 8.061191039677793e-05, 'epoch': 0.89}) +[2026-06-17 20:10:55,529][starVLA.training.train_starvla][INFO] - Step 2160, Loss: {'train/loss': 0.5696122646331787, 'train/grad_norm_pre_clip': 13.984657287597656, 'timing/data': 0.000754253938794136, 'timing/model': 1.3211232228204608, 'learning_rate/qwen_vl_interface': 8.059447483413768e-06, 'learning_rate/action_model': 8.059447483413767e-05, 'epoch': 0.89}) +[2026-06-17 20:10:59,560][starVLA.training.train_starvla][INFO] - Step 2161, Loss: {'train/loss': 0.7934843301773071, 'train/grad_norm_pre_clip': 14.730764389038086, 'timing/data': 0.0015015201643109322, 'timing/model': 1.0133516574278474, 'learning_rate/qwen_vl_interface': 8.057703344748258e-06, 'learning_rate/action_model': 8.057703344748257e-05, 'epoch': 0.89}) +[2026-06-17 20:11:03,652][starVLA.training.train_starvla][INFO] - Step 2162, Loss: {'train/loss': 0.7986438870429993, 'train/grad_norm_pre_clip': 12.002217292785645, 'timing/data': 0.0008132848888635635, 'timing/model': 0.9565489031374454, 'learning_rate/qwen_vl_interface': 8.05595862404282e-06, 'learning_rate/action_model': 8.05595862404282e-05, 'epoch': 0.89}) +[2026-06-17 20:11:07,563][starVLA.training.train_starvla][INFO] - Step 2163, Loss: {'train/loss': 0.8117341995239258, 'train/grad_norm_pre_clip': 22.17942237854004, 'timing/data': 0.0009303083643317223, 'timing/model': 1.1612774739041924, 'learning_rate/qwen_vl_interface': 8.054213321659143e-06, 'learning_rate/action_model': 8.054213321659142e-05, 'epoch': 0.89}) +[2026-06-17 20:11:11,473][starVLA.training.train_starvla][INFO] - Step 2164, Loss: {'train/loss': 0.42487621307373047, 'train/grad_norm_pre_clip': 8.884573936462402, 'timing/data': 0.0005209241062402725, 'timing/model': 1.1312042875215411, 'learning_rate/qwen_vl_interface': 8.052467437959025e-06, 'learning_rate/action_model': 8.052467437959025e-05, 'epoch': 0.89}) +[2026-06-17 20:11:15,257][starVLA.training.train_starvla][INFO] - Step 2165, Loss: {'train/loss': 0.6547988653182983, 'train/grad_norm_pre_clip': 11.105996131896973, 'timing/data': 0.0005257017910480499, 'timing/model': 1.0171809028834105, 'learning_rate/qwen_vl_interface': 8.050720973304393e-06, 'learning_rate/action_model': 8.050720973304391e-05, 'epoch': 0.89}) +[2026-06-17 20:11:19,226][starVLA.training.train_starvla][INFO] - Step 2166, Loss: {'train/loss': 0.5025679469108582, 'train/grad_norm_pre_clip': 16.075908660888672, 'timing/data': 0.0005984483286738396, 'timing/model': 1.036062485538423, 'learning_rate/qwen_vl_interface': 8.048973928057288e-06, 'learning_rate/action_model': 8.048973928057288e-05, 'epoch': 0.89}) +[2026-06-17 20:11:23,303][starVLA.training.train_starvla][INFO] - Step 2167, Loss: {'train/loss': 0.6570514440536499, 'train/grad_norm_pre_clip': 10.33656120300293, 'timing/data': 0.0006795357912778854, 'timing/model': 0.9582093097269535, 'learning_rate/qwen_vl_interface': 8.047226302579877e-06, 'learning_rate/action_model': 8.047226302579877e-05, 'epoch': 0.89}) +[2026-06-17 20:11:27,193][starVLA.training.train_starvla][INFO] - Step 2168, Loss: {'train/loss': 0.8628707528114319, 'train/grad_norm_pre_clip': 15.501396179199219, 'timing/data': 0.0005865786224603653, 'timing/model': 0.9667173568159342, 'learning_rate/qwen_vl_interface': 8.045478097234443e-06, 'learning_rate/action_model': 8.045478097234444e-05, 'epoch': 0.89}) +[2026-06-17 20:11:31,050][starVLA.training.train_starvla][INFO] - Step 2169, Loss: {'train/loss': 0.8287204504013062, 'train/grad_norm_pre_clip': 14.226958274841309, 'timing/data': 0.0008310545235872269, 'timing/model': 0.951571635901928, 'learning_rate/qwen_vl_interface': 8.043729312383393e-06, 'learning_rate/action_model': 8.043729312383392e-05, 'epoch': 0.89}) +[2026-06-17 20:11:35,047][starVLA.training.train_starvla][INFO] - Step 2170, Loss: {'train/loss': 0.5230119228363037, 'train/grad_norm_pre_clip': 26.84748077392578, 'timing/data': 0.0005657002329826355, 'timing/model': 1.165334852412343, 'learning_rate/qwen_vl_interface': 8.041979948389248e-06, 'learning_rate/action_model': 8.041979948389248e-05, 'epoch': 0.89}) +[2026-06-17 20:11:38,838][starVLA.training.train_starvla][INFO] - Step 2171, Loss: {'train/loss': 0.5756477117538452, 'train/grad_norm_pre_clip': 13.180816650390625, 'timing/data': 0.0005129221826791763, 'timing/model': 0.9593936502933502, 'learning_rate/qwen_vl_interface': 8.040230005614656e-06, 'learning_rate/action_model': 8.040230005614656e-05, 'epoch': 0.89}) +[2026-06-17 20:11:42,812][starVLA.training.train_starvla][INFO] - Step 2172, Loss: {'train/loss': 0.23262359201908112, 'train/grad_norm_pre_clip': 7.035514831542969, 'timing/data': 0.0008142134174704552, 'timing/model': 1.0104386750608683, 'learning_rate/qwen_vl_interface': 8.038479484422383e-06, 'learning_rate/action_model': 8.038479484422383e-05, 'epoch': 0.89}) +[2026-06-17 20:11:46,969][starVLA.training.train_starvla][INFO] - Step 2173, Loss: {'train/loss': 0.38765212893486023, 'train/grad_norm_pre_clip': 9.794384002685547, 'timing/data': 0.000752660445868969, 'timing/model': 1.274684239178896, 'learning_rate/qwen_vl_interface': 8.03672838517531e-06, 'learning_rate/action_model': 8.036728385175311e-05, 'epoch': 0.89}) +[2026-06-17 20:11:50,883][starVLA.training.train_starvla][INFO] - Step 2174, Loss: {'train/loss': 0.5967901945114136, 'train/grad_norm_pre_clip': 12.258208274841309, 'timing/data': 0.001286693848669529, 'timing/model': 0.9507453478872776, 'learning_rate/qwen_vl_interface': 8.034976708236446e-06, 'learning_rate/action_model': 8.034976708236445e-05, 'epoch': 0.89}) +[2026-06-17 20:11:54,737][starVLA.training.train_starvla][INFO] - Step 2175, Loss: {'train/loss': 0.7755168080329895, 'train/grad_norm_pre_clip': 13.583477973937988, 'timing/data': 0.000582503154873848, 'timing/model': 0.9656391339376569, 'learning_rate/qwen_vl_interface': 8.033224453968913e-06, 'learning_rate/action_model': 8.033224453968913e-05, 'epoch': 0.9}) +[2026-06-17 20:11:58,615][starVLA.training.train_starvla][INFO] - Step 2176, Loss: {'train/loss': 0.5809015035629272, 'train/grad_norm_pre_clip': 12.660627365112305, 'timing/data': 0.0009130574762821198, 'timing/model': 1.1254798052832484, 'learning_rate/qwen_vl_interface': 8.031471622735956e-06, 'learning_rate/action_model': 8.031471622735956e-05, 'epoch': 0.9}) +[2026-06-17 20:12:02,373][starVLA.training.train_starvla][INFO] - Step 2177, Loss: {'train/loss': 0.7033186554908752, 'train/grad_norm_pre_clip': 12.91633129119873, 'timing/data': 0.0008686650544404984, 'timing/model': 1.0681410757824779, 'learning_rate/qwen_vl_interface': 8.029718214900937e-06, 'learning_rate/action_model': 8.029718214900937e-05, 'epoch': 0.9}) +[2026-06-17 20:12:06,051][starVLA.training.train_starvla][INFO] - Step 2178, Loss: {'train/loss': 0.6083306074142456, 'train/grad_norm_pre_clip': 9.718377113342285, 'timing/data': 0.0009260820224881172, 'timing/model': 0.9607013463973999, 'learning_rate/qwen_vl_interface': 8.027964230827342e-06, 'learning_rate/action_model': 8.027964230827342e-05, 'epoch': 0.9}) +[2026-06-17 20:12:09,942][starVLA.training.train_starvla][INFO] - Step 2179, Loss: {'train/loss': 0.4886195957660675, 'train/grad_norm_pre_clip': 9.49555492401123, 'timing/data': 0.0004892898723483086, 'timing/model': 0.9602944767102599, 'learning_rate/qwen_vl_interface': 8.026209670878776e-06, 'learning_rate/action_model': 8.026209670878775e-05, 'epoch': 0.9}) +[2026-06-17 20:12:13,917][starVLA.training.train_starvla][INFO] - Step 2180, Loss: {'train/loss': 0.5035552978515625, 'train/grad_norm_pre_clip': 13.463847160339355, 'timing/data': 0.0007035480812191963, 'timing/model': 1.0393012864515185, 'learning_rate/qwen_vl_interface': 8.024454535418952e-06, 'learning_rate/action_model': 8.024454535418951e-05, 'epoch': 0.9}) +[2026-06-17 20:12:17,765][starVLA.training.train_starvla][INFO] - Step 2181, Loss: {'train/loss': 0.3576665222644806, 'train/grad_norm_pre_clip': 9.127856254577637, 'timing/data': 0.000972433015704155, 'timing/model': 0.947367686778307, 'learning_rate/qwen_vl_interface': 8.02269882481172e-06, 'learning_rate/action_model': 8.022698824811721e-05, 'epoch': 0.9}) +[2026-06-17 20:12:21,645][starVLA.training.train_starvla][INFO] - Step 2182, Loss: {'train/loss': 0.7258255481719971, 'train/grad_norm_pre_clip': 14.7601957321167, 'timing/data': 0.0012512169778347015, 'timing/model': 0.9462991999462247, 'learning_rate/qwen_vl_interface': 8.02094253942104e-06, 'learning_rate/action_model': 8.02094253942104e-05, 'epoch': 0.9}) +[2026-06-17 20:12:25,680][starVLA.training.train_starvla][INFO] - Step 2183, Loss: {'train/loss': 0.5616225600242615, 'train/grad_norm_pre_clip': 11.811809539794922, 'timing/data': 0.001197158358991146, 'timing/model': 1.1805181112140417, 'learning_rate/qwen_vl_interface': 8.019185679610988e-06, 'learning_rate/action_model': 8.019185679610988e-05, 'epoch': 0.9}) +[2026-06-17 20:12:29,534][starVLA.training.train_starvla][INFO] - Step 2184, Loss: {'train/loss': 0.5193107724189758, 'train/grad_norm_pre_clip': 11.38355541229248, 'timing/data': 0.0006866874173283577, 'timing/model': 1.1536637488752604, 'learning_rate/qwen_vl_interface': 8.017428245745765e-06, 'learning_rate/action_model': 8.017428245745763e-05, 'epoch': 0.9}) +[2026-06-17 20:12:33,309][starVLA.training.train_starvla][INFO] - Step 2185, Loss: {'train/loss': 0.6368452310562134, 'train/grad_norm_pre_clip': 13.009383201599121, 'timing/data': 0.0005409382283687592, 'timing/model': 0.9519191095605493, 'learning_rate/qwen_vl_interface': 8.015670238189688e-06, 'learning_rate/action_model': 8.015670238189688e-05, 'epoch': 0.9}) +[2026-06-17 20:12:37,198][starVLA.training.train_starvla][INFO] - Step 2186, Loss: {'train/loss': 0.31082427501678467, 'train/grad_norm_pre_clip': 7.6603102684021, 'timing/data': 0.0005140285938978195, 'timing/model': 0.9498005816712976, 'learning_rate/qwen_vl_interface': 8.013911657307196e-06, 'learning_rate/action_model': 8.013911657307196e-05, 'epoch': 0.9}) +[2026-06-17 20:12:41,142][starVLA.training.train_starvla][INFO] - Step 2187, Loss: {'train/loss': 0.35409751534461975, 'train/grad_norm_pre_clip': 7.982658386230469, 'timing/data': 0.001077788881957531, 'timing/model': 0.9543108623474836, 'learning_rate/qwen_vl_interface': 8.012152503462843e-06, 'learning_rate/action_model': 8.012152503462842e-05, 'epoch': 0.9}) +[2026-06-17 20:12:45,158][starVLA.training.train_starvla][INFO] - Step 2188, Loss: {'train/loss': 0.7905974984169006, 'train/grad_norm_pre_clip': 12.763652801513672, 'timing/data': 0.0010506818071007729, 'timing/model': 0.9691139748319983, 'learning_rate/qwen_vl_interface': 8.010392777021304e-06, 'learning_rate/action_model': 8.010392777021303e-05, 'epoch': 0.9}) +[2026-06-17 20:12:49,155][starVLA.training.train_starvla][INFO] - Step 2189, Loss: {'train/loss': 0.631453275680542, 'train/grad_norm_pre_clip': 10.54888916015625, 'timing/data': 0.0011314451694488525, 'timing/model': 0.9524784693494439, 'learning_rate/qwen_vl_interface': 8.008632478347372e-06, 'learning_rate/action_model': 8.008632478347371e-05, 'epoch': 0.9}) +[2026-06-17 20:12:53,257][starVLA.training.train_starvla][INFO] - Step 2190, Loss: {'train/loss': 0.6718952655792236, 'train/grad_norm_pre_clip': 16.67821502685547, 'timing/data': 0.0008299639448523521, 'timing/model': 1.3479878259822726, 'learning_rate/qwen_vl_interface': 8.006871607805957e-06, 'learning_rate/action_model': 8.006871607805957e-05, 'epoch': 0.9}) +[2026-06-17 20:13:01,407][starVLA.training.train_starvla][INFO] - Step 2191, Loss: {'train/loss': 0.5004340410232544, 'train/grad_norm_pre_clip': 10.2354736328125, 'timing/data': 0.0005465270951390266, 'timing/model': 1.1830811966210604, 'learning_rate/qwen_vl_interface': 8.005110165762093e-06, 'learning_rate/action_model': 8.005110165762093e-05, 'epoch': 0.9}) +[2026-06-17 20:13:05,168][starVLA.training.train_starvla][INFO] - Step 2192, Loss: {'train/loss': 0.40346595644950867, 'train/grad_norm_pre_clip': 32.53037643432617, 'timing/data': 0.0013363976031541824, 'timing/model': 0.9523863652721047, 'learning_rate/qwen_vl_interface': 8.003348152580927e-06, 'learning_rate/action_model': 8.003348152580927e-05, 'epoch': 0.9}) +[2026-06-17 20:13:09,359][starVLA.training.train_starvla][INFO] - Step 2193, Loss: {'train/loss': 0.7810356020927429, 'train/grad_norm_pre_clip': 14.811234474182129, 'timing/data': 0.0008815648034214973, 'timing/model': 0.9749195259064436, 'learning_rate/qwen_vl_interface': 8.001585568627724e-06, 'learning_rate/action_model': 8.001585568627725e-05, 'epoch': 0.9}) +[2026-06-17 20:13:14,235][starVLA.training.train_starvla][INFO] - Step 2194, Loss: {'train/loss': 0.41914719343185425, 'train/grad_norm_pre_clip': 10.178540229797363, 'timing/data': 0.0005393717437982559, 'timing/model': 1.7611654568463564, 'learning_rate/qwen_vl_interface': 7.999822414267874e-06, 'learning_rate/action_model': 7.999822414267874e-05, 'epoch': 0.9}) +[2026-06-17 20:13:21,008][starVLA.training.train_starvla][INFO] - Step 2195, Loss: {'train/loss': 0.5977632999420166, 'train/grad_norm_pre_clip': 11.93103313446045, 'timing/data': 0.0006546387448906898, 'timing/model': 3.6173685835674405, 'learning_rate/qwen_vl_interface': 7.998058689866879e-06, 'learning_rate/action_model': 7.998058689866879e-05, 'epoch': 0.9}) +[2026-06-17 20:13:27,666][starVLA.training.train_starvla][INFO] - Step 2196, Loss: {'train/loss': 0.621052086353302, 'train/grad_norm_pre_clip': 14.757894515991211, 'timing/data': 0.0008885813876986504, 'timing/model': 2.871021619066596, 'learning_rate/qwen_vl_interface': 7.99629439579036e-06, 'learning_rate/action_model': 7.99629439579036e-05, 'epoch': 0.9}) +[2026-06-17 20:13:31,938][starVLA.training.train_starvla][INFO] - Step 2197, Loss: {'train/loss': 0.6089439988136292, 'train/grad_norm_pre_clip': 15.659045219421387, 'timing/data': 0.0009072292596101761, 'timing/model': 1.0951606538146734, 'learning_rate/qwen_vl_interface': 7.994529532404059e-06, 'learning_rate/action_model': 7.994529532404058e-05, 'epoch': 0.9}) +[2026-06-17 20:13:38,013][starVLA.training.train_starvla][INFO] - Step 2198, Loss: {'train/loss': 0.360589861869812, 'train/grad_norm_pre_clip': 8.330151557922363, 'timing/data': 0.000529969111084938, 'timing/model': 1.0292644537985325, 'learning_rate/qwen_vl_interface': 7.99276410007383e-06, 'learning_rate/action_model': 7.99276410007383e-05, 'epoch': 0.9}) +[2026-06-17 20:13:41,992][starVLA.training.train_starvla][INFO] - Step 2199, Loss: {'train/loss': 0.7498315572738647, 'train/grad_norm_pre_clip': 23.137386322021484, 'timing/data': 0.0012861629948019981, 'timing/model': 0.95299984049052, 'learning_rate/qwen_vl_interface': 7.990998099165652e-06, 'learning_rate/action_model': 7.990998099165652e-05, 'epoch': 0.91}) +[2026-06-17 20:16:21,682][starVLA.training.train_starvla][INFO] - Step 2200, Loss: {'train/loss': 0.7127363681793213, 'train/grad_norm_pre_clip': 15.518760681152344, 'eval/loss': 0.5720392465591431, 'eval/latency_6/loss': 0.5720392465591431, 'eval/demon_fire_precision': 0.940578328566889, 'eval/demon_fire_recall': 0.9566903684550744, 'eval/demon_fire_f1': 0.9485659349463227, 'eval/demon_fire_rate_model': 0.8756260434056762, 'eval/demon_fire_rate_teacher': 0.860879243183083, 'eval/demon_left_precision': 0.76953125, 'eval/demon_left_recall': 0.6912280701754386, 'eval/demon_left_f1': 0.7282809611829945, 'eval/demon_left_rate_model': 0.1424596549805231, 'eval/demon_left_rate_teacher': 0.15859766277128548, 'eval/demon_right_precision': 0.7741379310344828, 'eval/demon_right_recall': 0.6897081413210445, 'eval/demon_right_f1': 0.7294882209585702, 'eval/demon_right_rate_model': 0.16138007790762382, 'eval/demon_right_rate_teacher': 0.18113522537562604, 'eval/demon_move_f1': 0.7288845910707824, 'eval/demon_cc_f1': 0.8699102511652912, 'eval/latency_6/demon_fire_precision': 0.940578328566889, 'eval/latency_6/demon_fire_recall': 0.9566903684550744, 'eval/latency_6/demon_fire_f1': 0.9485659349463227, 'eval/latency_6/demon_fire_rate_model': 0.8756260434056762, 'eval/latency_6/demon_fire_rate_teacher': 0.860879243183083, 'eval/latency_6/demon_left_precision': 0.76953125, 'eval/latency_6/demon_left_recall': 0.6912280701754386, 'eval/latency_6/demon_left_f1': 0.7282809611829945, 'eval/latency_6/demon_left_rate_model': 0.1424596549805231, 'eval/latency_6/demon_left_rate_teacher': 0.15859766277128548, 'eval/latency_6/demon_right_precision': 0.7741379310344828, 'eval/latency_6/demon_right_recall': 0.6897081413210445, 'eval/latency_6/demon_right_f1': 0.7294882209585702, 'eval/latency_6/demon_right_rate_model': 0.16138007790762382, 'eval/latency_6/demon_right_rate_teacher': 0.18113522537562604, 'eval/latency_6/demon_move_f1': 0.7288845910707824, 'eval/latency_6/demon_cc_f1': 0.8699102511652912, 'timing/data': 0.0007102470844984055, 'timing/model': 0.956541271880269, 'learning_rate/qwen_vl_interface': 7.98923153004562e-06, 'learning_rate/action_model': 7.98923153004562e-05, 'epoch': 0.91}) +[2026-06-17 20:16:21,686][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_2200_state +[2026-06-17 20:16:21,691][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 20:16:28,277][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_2200_state/model.safetensors +[2026-06-17 20:16:44,474][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_2200_state/optimizer.bin +[2026-06-17 20:16:44,475][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_2200_state/sampler.bin +[2026-06-17 20:16:44,476][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_2200_state/sampler_1.bin +[2026-06-17 20:16:44,478][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_2200_state/random_states_0.pkl +[2026-06-17 20:16:44,484][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2200 +[2026-06-17 20:16:44,484][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 20:16:44,496][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 20:16:48,594][starVLA.training.train_starvla][INFO] - Step 2201, Loss: {'train/loss': 0.6007330417633057, 'train/grad_norm_pre_clip': 14.286230087280273, 'timing/data': 0.000623529776930809, 'timing/model': 0.9567529400810599, 'learning_rate/qwen_vl_interface': 7.987464393079943e-06, 'learning_rate/action_model': 7.987464393079943e-05, 'epoch': 0.91}) +[2026-06-17 20:16:52,836][starVLA.training.train_starvla][INFO] - Step 2202, Loss: {'train/loss': 0.7240551114082336, 'train/grad_norm_pre_clip': 15.21876049041748, 'timing/data': 0.0006786435842514038, 'timing/model': 1.0728304656222463, 'learning_rate/qwen_vl_interface': 7.985696688634953e-06, 'learning_rate/action_model': 7.985696688634952e-05, 'epoch': 0.91}) +[2026-06-17 20:16:56,821][starVLA.training.train_starvla][INFO] - Step 2203, Loss: {'train/loss': 0.7093086242675781, 'train/grad_norm_pre_clip': 12.390002250671387, 'timing/data': 0.001021779142320156, 'timing/model': 0.9581018462777138, 'learning_rate/qwen_vl_interface': 7.983928417077095e-06, 'learning_rate/action_model': 7.983928417077095e-05, 'epoch': 0.91}) +[2026-06-17 20:17:00,922][starVLA.training.train_starvla][INFO] - Step 2204, Loss: {'train/loss': 0.572995126247406, 'train/grad_norm_pre_clip': 11.712000846862793, 'timing/data': 0.0005573760718107224, 'timing/model': 1.038275944069028, 'learning_rate/qwen_vl_interface': 7.982159578772935e-06, 'learning_rate/action_model': 7.982159578772934e-05, 'epoch': 0.91}) +[2026-06-17 20:17:04,980][starVLA.training.train_starvla][INFO] - Step 2205, Loss: {'train/loss': 0.6698823571205139, 'train/grad_norm_pre_clip': 12.634589195251465, 'timing/data': 0.0010744957253336906, 'timing/model': 1.1863161362707615, 'learning_rate/qwen_vl_interface': 7.980390174089154e-06, 'learning_rate/action_model': 7.980390174089153e-05, 'epoch': 0.91}) +[2026-06-17 20:17:08,743][starVLA.training.train_starvla][INFO] - Step 2206, Loss: {'train/loss': 0.5854729413986206, 'train/grad_norm_pre_clip': 10.052739143371582, 'timing/data': 0.0008201142773032188, 'timing/model': 0.9705451242625713, 'learning_rate/qwen_vl_interface': 7.978620203392551e-06, 'learning_rate/action_model': 7.97862020339255e-05, 'epoch': 0.91}) +[2026-06-17 20:17:12,784][starVLA.training.train_starvla][INFO] - Step 2207, Loss: {'train/loss': 0.5887125134468079, 'train/grad_norm_pre_clip': 13.782492637634277, 'timing/data': 0.0009886883199214935, 'timing/model': 0.9533345056697726, 'learning_rate/qwen_vl_interface': 7.976849667050044e-06, 'learning_rate/action_model': 7.976849667050044e-05, 'epoch': 0.91}) +[2026-06-17 20:17:16,937][starVLA.training.train_starvla][INFO] - Step 2208, Loss: {'train/loss': 0.586912989616394, 'train/grad_norm_pre_clip': 12.404440879821777, 'timing/data': 0.001125267706811428, 'timing/model': 0.9747123634442687, 'learning_rate/qwen_vl_interface': 7.975078565428667e-06, 'learning_rate/action_model': 7.975078565428668e-05, 'epoch': 0.91}) +[2026-06-17 20:17:20,951][starVLA.training.train_starvla][INFO] - Step 2209, Loss: {'train/loss': 0.4837871193885803, 'train/grad_norm_pre_clip': 13.332063674926758, 'timing/data': 0.0005290498957037926, 'timing/model': 0.9581691399216652, 'learning_rate/qwen_vl_interface': 7.97330689889557e-06, 'learning_rate/action_model': 7.97330689889557e-05, 'epoch': 0.91}) +[2026-06-17 20:17:25,024][starVLA.training.train_starvla][INFO] - Step 2210, Loss: {'train/loss': 0.577788770198822, 'train/grad_norm_pre_clip': 12.673194885253906, 'timing/data': 0.0009451285004615784, 'timing/model': 1.050997033715248, 'learning_rate/qwen_vl_interface': 7.971534667818026e-06, 'learning_rate/action_model': 7.971534667818025e-05, 'epoch': 0.91}) +[2026-06-17 20:17:29,076][starVLA.training.train_starvla][INFO] - Step 2211, Loss: {'train/loss': 0.551966667175293, 'train/grad_norm_pre_clip': 13.149168968200684, 'timing/data': 0.0007309475913643837, 'timing/model': 1.2592316418886185, 'learning_rate/qwen_vl_interface': 7.969761872563413e-06, 'learning_rate/action_model': 7.969761872563413e-05, 'epoch': 0.91}) +[2026-06-17 20:17:32,832][starVLA.training.train_starvla][INFO] - Step 2212, Loss: {'train/loss': 0.20923110842704773, 'train/grad_norm_pre_clip': 6.509467601776123, 'timing/data': 0.0006561707705259323, 'timing/model': 0.9659453798085451, 'learning_rate/qwen_vl_interface': 7.96798851349924e-06, 'learning_rate/action_model': 7.96798851349924e-05, 'epoch': 0.91}) +[2026-06-17 20:17:36,969][starVLA.training.train_starvla][INFO] - Step 2213, Loss: {'train/loss': 0.6042374968528748, 'train/grad_norm_pre_clip': 16.651681900024414, 'timing/data': 0.0012166239321231842, 'timing/model': 0.9465712560340762, 'learning_rate/qwen_vl_interface': 7.966214590993122e-06, 'learning_rate/action_model': 7.966214590993122e-05, 'epoch': 0.91}) +[2026-06-17 20:17:40,891][starVLA.training.train_starvla][INFO] - Step 2214, Loss: {'train/loss': 0.7421984672546387, 'train/grad_norm_pre_clip': 32.73408889770508, 'timing/data': 0.0007120408117771149, 'timing/model': 0.9615837354212999, 'learning_rate/qwen_vl_interface': 7.964440105412797e-06, 'learning_rate/action_model': 7.964440105412796e-05, 'epoch': 0.91}) +[2026-06-17 20:17:45,008][starVLA.training.train_starvla][INFO] - Step 2215, Loss: {'train/loss': 0.6306624412536621, 'train/grad_norm_pre_clip': 15.55103588104248, 'timing/data': 0.0008158963173627853, 'timing/model': 0.9648974109441042, 'learning_rate/qwen_vl_interface': 7.962665057126116e-06, 'learning_rate/action_model': 7.962665057126117e-05, 'epoch': 0.91}) +[2026-06-17 20:17:49,473][starVLA.training.train_starvla][INFO] - Step 2216, Loss: {'train/loss': 0.3986988067626953, 'train/grad_norm_pre_clip': 9.319660186767578, 'timing/data': 0.0008749421685934067, 'timing/model': 1.5236531095579267, 'learning_rate/qwen_vl_interface': 7.960889446501052e-06, 'learning_rate/action_model': 7.960889446501052e-05, 'epoch': 0.91}) +[2026-06-17 20:17:53,365][starVLA.training.train_starvla][INFO] - Step 2217, Loss: {'train/loss': 0.772587776184082, 'train/grad_norm_pre_clip': 15.891706466674805, 'timing/data': 0.0007712328806519508, 'timing/model': 1.1342553682625294, 'learning_rate/qwen_vl_interface': 7.959113273905688e-06, 'learning_rate/action_model': 7.959113273905688e-05, 'epoch': 0.91}) +[2026-06-17 20:17:58,372][starVLA.training.train_starvla][INFO] - Step 2218, Loss: {'train/loss': 0.5975213646888733, 'train/grad_norm_pre_clip': 9.364232063293457, 'timing/data': 0.001528640277683735, 'timing/model': 1.0846386710181832, 'learning_rate/qwen_vl_interface': 7.957336539708226e-06, 'learning_rate/action_model': 7.957336539708225e-05, 'epoch': 0.91}) +[2026-06-17 20:18:02,077][starVLA.training.train_starvla][INFO] - Step 2219, Loss: {'train/loss': 0.5701444745063782, 'train/grad_norm_pre_clip': 11.56022834777832, 'timing/data': 0.0005060825496912003, 'timing/model': 0.994494722224772, 'learning_rate/qwen_vl_interface': 7.955559244276987e-06, 'learning_rate/action_model': 7.955559244276987e-05, 'epoch': 0.91}) +[2026-06-17 20:18:06,030][starVLA.training.train_starvla][INFO] - Step 2220, Loss: {'train/loss': 0.5651236176490784, 'train/grad_norm_pre_clip': 17.331850051879883, 'timing/data': 0.0005231304094195366, 'timing/model': 1.0512266512960196, 'learning_rate/qwen_vl_interface': 7.953781387980404e-06, 'learning_rate/action_model': 7.953781387980404e-05, 'epoch': 0.91}) +[2026-06-17 20:18:10,037][starVLA.training.train_starvla][INFO] - Step 2221, Loss: {'train/loss': 0.4574592411518097, 'train/grad_norm_pre_clip': 10.226433753967285, 'timing/data': 0.0004966994747519493, 'timing/model': 0.9270759802311659, 'learning_rate/qwen_vl_interface': 7.952002971187032e-06, 'learning_rate/action_model': 7.952002971187032e-05, 'epoch': 0.91}) +[2026-06-17 20:18:14,019][starVLA.training.train_starvla][INFO] - Step 2222, Loss: {'train/loss': 0.8658500909805298, 'train/grad_norm_pre_clip': 20.841386795043945, 'timing/data': 0.0005948040634393692, 'timing/model': 0.9693098943680525, 'learning_rate/qwen_vl_interface': 7.950223994265534e-06, 'learning_rate/action_model': 7.950223994265533e-05, 'epoch': 0.91}) +[2026-06-17 20:18:18,614][starVLA.training.train_starvla][INFO] - Step 2223, Loss: {'train/loss': 0.48478591442108154, 'train/grad_norm_pre_clip': 9.950815200805664, 'timing/data': 0.000879405066370964, 'timing/model': 0.9853154709562659, 'learning_rate/qwen_vl_interface': 7.9484444575847e-06, 'learning_rate/action_model': 7.948444457584699e-05, 'epoch': 0.92}) +[2026-06-17 20:18:26,113][starVLA.training.train_starvla][INFO] - Step 2224, Loss: {'train/loss': 0.8202155232429504, 'train/grad_norm_pre_clip': 16.27478790283203, 'timing/data': 0.0005880314856767654, 'timing/model': 1.9512777058407664, 'learning_rate/qwen_vl_interface': 7.946664361513423e-06, 'learning_rate/action_model': 7.946664361513421e-05, 'epoch': 0.92}) +[2026-06-17 20:18:29,897][starVLA.training.train_starvla][INFO] - Step 2225, Loss: {'train/loss': 0.49164414405822754, 'train/grad_norm_pre_clip': 10.289587020874023, 'timing/data': 0.0007117362692952156, 'timing/model': 1.0210050093010068, 'learning_rate/qwen_vl_interface': 7.944883706420724e-06, 'learning_rate/action_model': 7.944883706420724e-05, 'epoch': 0.92}) +[2026-06-17 20:18:33,863][starVLA.training.train_starvla][INFO] - Step 2226, Loss: {'train/loss': 0.895767331123352, 'train/grad_norm_pre_clip': 13.54488468170166, 'timing/data': 0.0006727809086441994, 'timing/model': 0.9395431131124496, 'learning_rate/qwen_vl_interface': 7.943102492675732e-06, 'learning_rate/action_model': 7.943102492675732e-05, 'epoch': 0.92}) +[2026-06-17 20:18:37,797][starVLA.training.train_starvla][INFO] - Step 2227, Loss: {'train/loss': 1.0179117918014526, 'train/grad_norm_pre_clip': 17.565189361572266, 'timing/data': 0.0005459235981106758, 'timing/model': 0.9566057147458196, 'learning_rate/qwen_vl_interface': 7.941320720647694e-06, 'learning_rate/action_model': 7.941320720647695e-05, 'epoch': 0.92}) +[2026-06-17 20:18:41,831][starVLA.training.train_starvla][INFO] - Step 2228, Loss: {'train/loss': 0.37031039595603943, 'train/grad_norm_pre_clip': 10.004687309265137, 'timing/data': 0.0005846712738275528, 'timing/model': 0.9717452106997371, 'learning_rate/qwen_vl_interface': 7.939538390705976e-06, 'learning_rate/action_model': 7.939538390705976e-05, 'epoch': 0.92}) +[2026-06-17 20:18:45,793][starVLA.training.train_starvla][INFO] - Step 2229, Loss: {'train/loss': 0.45563533902168274, 'train/grad_norm_pre_clip': 11.709538459777832, 'timing/data': 0.0005197310820221901, 'timing/model': 0.9513077968731523, 'learning_rate/qwen_vl_interface': 7.937755503220057e-06, 'learning_rate/action_model': 7.937755503220056e-05, 'epoch': 0.92}) +[2026-06-17 20:18:49,688][starVLA.training.train_starvla][INFO] - Step 2230, Loss: {'train/loss': 0.41383928060531616, 'train/grad_norm_pre_clip': 10.588751792907715, 'timing/data': 0.001132912002503872, 'timing/model': 1.193219074048102, 'learning_rate/qwen_vl_interface': 7.93597205855953e-06, 'learning_rate/action_model': 7.935972058559529e-05, 'epoch': 0.92}) +[2026-06-17 20:18:53,574][starVLA.training.train_starvla][INFO] - Step 2231, Loss: {'train/loss': 0.42766937613487244, 'train/grad_norm_pre_clip': 11.563861846923828, 'timing/data': 0.0007177507504820824, 'timing/model': 1.1636004839092493, 'learning_rate/qwen_vl_interface': 7.934188057094103e-06, 'learning_rate/action_model': 7.934188057094103e-05, 'epoch': 0.92}) +[2026-06-17 20:18:57,258][starVLA.training.train_starvla][INFO] - Step 2232, Loss: {'train/loss': 0.45169931650161743, 'train/grad_norm_pre_clip': 11.668917655944824, 'timing/data': 0.000926189124584198, 'timing/model': 0.9660449763759971, 'learning_rate/qwen_vl_interface': 7.932403499193604e-06, 'learning_rate/action_model': 7.932403499193605e-05, 'epoch': 0.92}) +[2026-06-17 20:19:01,284][starVLA.training.train_starvla][INFO] - Step 2233, Loss: {'train/loss': 0.44389721751213074, 'train/grad_norm_pre_clip': 11.263949394226074, 'timing/data': 0.0010602446272969246, 'timing/model': 1.023652714677155, 'learning_rate/qwen_vl_interface': 7.930618385227975e-06, 'learning_rate/action_model': 7.930618385227973e-05, 'epoch': 0.92}) +[2026-06-17 20:19:05,206][starVLA.training.train_starvla][INFO] - Step 2234, Loss: {'train/loss': 0.5690252184867859, 'train/grad_norm_pre_clip': 11.782625198364258, 'timing/data': 0.0005616806447505951, 'timing/model': 0.9634419716894627, 'learning_rate/qwen_vl_interface': 7.928832715567268e-06, 'learning_rate/action_model': 7.928832715567267e-05, 'epoch': 0.92}) +[2026-06-17 20:19:09,174][starVLA.training.train_starvla][INFO] - Step 2235, Loss: {'train/loss': 0.8667307496070862, 'train/grad_norm_pre_clip': 17.616941452026367, 'timing/data': 0.0005343956872820854, 'timing/model': 0.9440948953852057, 'learning_rate/qwen_vl_interface': 7.927046490581655e-06, 'learning_rate/action_model': 7.927046490581655e-05, 'epoch': 0.92}) +[2026-06-17 20:19:13,102][starVLA.training.train_starvla][INFO] - Step 2236, Loss: {'train/loss': 0.36531543731689453, 'train/grad_norm_pre_clip': 8.830878257751465, 'timing/data': 0.0007064184173941612, 'timing/model': 0.95387820340693, 'learning_rate/qwen_vl_interface': 7.925259710641428e-06, 'learning_rate/action_model': 7.925259710641427e-05, 'epoch': 0.92}) +[2026-06-17 20:19:17,078][starVLA.training.train_starvla][INFO] - Step 2237, Loss: {'train/loss': 0.4878736436367035, 'train/grad_norm_pre_clip': 11.638448715209961, 'timing/data': 0.0007524788379669189, 'timing/model': 1.2222598688676953, 'learning_rate/qwen_vl_interface': 7.923472376116982e-06, 'learning_rate/action_model': 7.923472376116982e-05, 'epoch': 0.92}) +[2026-06-17 20:19:21,128][starVLA.training.train_starvla][INFO] - Step 2238, Loss: {'train/loss': 0.410462886095047, 'train/grad_norm_pre_clip': 9.582722663879395, 'timing/data': 0.0011938335373997688, 'timing/model': 1.1287857107818127, 'learning_rate/qwen_vl_interface': 7.921684487378836e-06, 'learning_rate/action_model': 7.921684487378837e-05, 'epoch': 0.92}) +[2026-06-17 20:19:24,860][starVLA.training.train_starvla][INFO] - Step 2239, Loss: {'train/loss': 0.4888489246368408, 'train/grad_norm_pre_clip': 13.393956184387207, 'timing/data': 0.0017737345770001411, 'timing/model': 0.968443563207984, 'learning_rate/qwen_vl_interface': 7.919896044797622e-06, 'learning_rate/action_model': 7.91989604479762e-05, 'epoch': 0.92}) +[2026-06-17 20:19:28,789][starVLA.training.train_starvla][INFO] - Step 2240, Loss: {'train/loss': 0.61344975233078, 'train/grad_norm_pre_clip': 22.320308685302734, 'timing/data': 0.0007814839482307434, 'timing/model': 0.9456839002668858, 'learning_rate/qwen_vl_interface': 7.918107048744083e-06, 'learning_rate/action_model': 7.918107048744083e-05, 'epoch': 0.92}) +[2026-06-17 20:19:32,643][starVLA.training.train_starvla][INFO] - Step 2241, Loss: {'train/loss': 0.34758374094963074, 'train/grad_norm_pre_clip': 21.378190994262695, 'timing/data': 0.0008923495188355446, 'timing/model': 0.9406072990968823, 'learning_rate/qwen_vl_interface': 7.916317499589083e-06, 'learning_rate/action_model': 7.916317499589083e-05, 'epoch': 0.92}) +[2026-06-17 20:19:36,576][starVLA.training.train_starvla][INFO] - Step 2242, Loss: {'train/loss': 0.5922238230705261, 'train/grad_norm_pre_clip': 11.722057342529297, 'timing/data': 0.001072918064892292, 'timing/model': 1.006505362689495, 'learning_rate/qwen_vl_interface': 7.914527397703596e-06, 'learning_rate/action_model': 7.914527397703595e-05, 'epoch': 0.92}) +[2026-06-17 20:19:40,377][starVLA.training.train_starvla][INFO] - Step 2243, Loss: {'train/loss': 0.6448332667350769, 'train/grad_norm_pre_clip': 14.85991096496582, 'timing/data': 0.0004983935505151749, 'timing/model': 0.9444890869781375, 'learning_rate/qwen_vl_interface': 7.912736743458712e-06, 'learning_rate/action_model': 7.912736743458711e-05, 'epoch': 0.92}) +[2026-06-17 20:19:44,290][starVLA.training.train_starvla][INFO] - Step 2244, Loss: {'train/loss': 0.6028947234153748, 'train/grad_norm_pre_clip': 12.301033020019531, 'timing/data': 0.0005555795505642891, 'timing/model': 1.1624214053153992, 'learning_rate/qwen_vl_interface': 7.910945537225638e-06, 'learning_rate/action_model': 7.910945537225637e-05, 'epoch': 0.92}) +[2026-06-17 20:19:47,950][starVLA.training.train_starvla][INFO] - Step 2245, Loss: {'train/loss': 0.7794967889785767, 'train/grad_norm_pre_clip': 15.348333358764648, 'timing/data': 0.0007005073130130768, 'timing/model': 0.9620054718106985, 'learning_rate/qwen_vl_interface': 7.909153779375688e-06, 'learning_rate/action_model': 7.909153779375688e-05, 'epoch': 0.92}) +[2026-06-17 20:19:51,842][starVLA.training.train_starvla][INFO] - Step 2246, Loss: {'train/loss': 0.8939030170440674, 'train/grad_norm_pre_clip': 50.513763427734375, 'timing/data': 0.0005092229694128036, 'timing/model': 0.962918465025723, 'learning_rate/qwen_vl_interface': 7.907361470280299e-06, 'learning_rate/action_model': 7.907361470280299e-05, 'epoch': 0.92}) +[2026-06-17 20:19:55,696][starVLA.training.train_starvla][INFO] - Step 2247, Loss: {'train/loss': 0.3536916971206665, 'train/grad_norm_pre_clip': 9.033560752868652, 'timing/data': 0.0006157495081424713, 'timing/model': 0.9533372158184648, 'learning_rate/qwen_vl_interface': 7.905568610311018e-06, 'learning_rate/action_model': 7.905568610311017e-05, 'epoch': 0.92}) +[2026-06-17 20:19:59,711][starVLA.training.train_starvla][INFO] - Step 2248, Loss: {'train/loss': 0.33006536960601807, 'train/grad_norm_pre_clip': 8.414047241210938, 'timing/data': 0.0007259966805577278, 'timing/model': 0.9583367882296443, 'learning_rate/qwen_vl_interface': 7.903775199839506e-06, 'learning_rate/action_model': 7.903775199839505e-05, 'epoch': 0.93}) +[2026-06-17 20:20:03,758][starVLA.training.train_starvla][INFO] - Step 2249, Loss: {'train/loss': 0.6151354908943176, 'train/grad_norm_pre_clip': 12.692323684692383, 'timing/data': 0.0013868454843759537, 'timing/model': 0.9507275242358446, 'learning_rate/qwen_vl_interface': 7.901981239237539e-06, 'learning_rate/action_model': 7.901981239237539e-05, 'epoch': 0.93}) +[2026-06-17 20:20:07,720][starVLA.training.train_starvla][INFO] - Step 2250, Loss: {'train/loss': 0.7049434781074524, 'train/grad_norm_pre_clip': 14.66716194152832, 'timing/data': 0.001324513927102089, 'timing/model': 1.2285703513771296, 'learning_rate/qwen_vl_interface': 7.900186728877009e-06, 'learning_rate/action_model': 7.90018672887701e-05, 'epoch': 0.93}) +[2026-06-17 20:20:11,663][starVLA.training.train_starvla][INFO] - Step 2251, Loss: {'train/loss': 0.8140401244163513, 'train/grad_norm_pre_clip': 14.805598258972168, 'timing/data': 0.0005122674629092216, 'timing/model': 1.1119495434686542, 'learning_rate/qwen_vl_interface': 7.898391669129919e-06, 'learning_rate/action_model': 7.898391669129918e-05, 'epoch': 0.93}) +[2026-06-17 20:20:15,363][starVLA.training.train_starvla][INFO] - Step 2252, Loss: {'train/loss': 0.6394728422164917, 'train/grad_norm_pre_clip': 13.541973114013672, 'timing/data': 0.0009138006716966629, 'timing/model': 0.9684279868379235, 'learning_rate/qwen_vl_interface': 7.896596060368384e-06, 'learning_rate/action_model': 7.896596060368384e-05, 'epoch': 0.93}) +[2026-06-17 20:20:19,370][starVLA.training.train_starvla][INFO] - Step 2253, Loss: {'train/loss': 0.45975297689437866, 'train/grad_norm_pre_clip': 13.68139362335205, 'timing/data': 0.000721566379070282, 'timing/model': 0.9579135375097394, 'learning_rate/qwen_vl_interface': 7.89479990296464e-06, 'learning_rate/action_model': 7.894799902964639e-05, 'epoch': 0.93}) +[2026-06-17 20:20:23,507][starVLA.training.train_starvla][INFO] - Step 2254, Loss: {'train/loss': 0.677740752696991, 'train/grad_norm_pre_clip': 13.050313949584961, 'timing/data': 0.0010454906150698662, 'timing/model': 1.0340242218226194, 'learning_rate/qwen_vl_interface': 7.893003197291028e-06, 'learning_rate/action_model': 7.893003197291027e-05, 'epoch': 0.93}) +[2026-06-17 20:20:27,476][starVLA.training.train_starvla][INFO] - Step 2255, Loss: {'train/loss': 0.5933636426925659, 'train/grad_norm_pre_clip': 18.229951858520508, 'timing/data': 0.0005829250440001488, 'timing/model': 0.9467863198369741, 'learning_rate/qwen_vl_interface': 7.89120594372001e-06, 'learning_rate/action_model': 7.89120594372001e-05, 'epoch': 0.93}) +[2026-06-17 20:20:31,311][starVLA.training.train_starvla][INFO] - Step 2256, Loss: {'train/loss': 0.28754377365112305, 'train/grad_norm_pre_clip': 8.255585670471191, 'timing/data': 0.0006305910646915436, 'timing/model': 0.9572108583524823, 'learning_rate/qwen_vl_interface': 7.88940814262416e-06, 'learning_rate/action_model': 7.889408142624158e-05, 'epoch': 0.93}) +[2026-06-17 20:20:35,286][starVLA.training.train_starvla][INFO] - Step 2257, Loss: {'train/loss': 0.5190932750701904, 'train/grad_norm_pre_clip': 11.536931991577148, 'timing/data': 0.0005113789811730385, 'timing/model': 1.2125281495973468, 'learning_rate/qwen_vl_interface': 7.887609794376158e-06, 'learning_rate/action_model': 7.887609794376158e-05, 'epoch': 0.93}) +[2026-06-17 20:20:39,128][starVLA.training.train_starvla][INFO] - Step 2258, Loss: {'train/loss': 0.7281944155693054, 'train/grad_norm_pre_clip': 10.973295211791992, 'timing/data': 0.0004909420385956764, 'timing/model': 1.1141501991078258, 'learning_rate/qwen_vl_interface': 7.88581089934881e-06, 'learning_rate/action_model': 7.88581089934881e-05, 'epoch': 0.93}) +[2026-06-17 20:20:42,865][starVLA.training.train_starvla][INFO] - Step 2259, Loss: {'train/loss': 0.6195377111434937, 'train/grad_norm_pre_clip': 11.571352005004883, 'timing/data': 0.0009823273867368698, 'timing/model': 0.9706631097942591, 'learning_rate/qwen_vl_interface': 7.884011457915025e-06, 'learning_rate/action_model': 7.884011457915024e-05, 'epoch': 0.93}) +[2026-06-17 20:20:46,803][starVLA.training.train_starvla][INFO] - Step 2260, Loss: {'train/loss': 0.3318556845188141, 'train/grad_norm_pre_clip': 10.424783706665039, 'timing/data': 0.0007971739396452904, 'timing/model': 0.973209273070097, 'learning_rate/qwen_vl_interface': 7.882211470447832e-06, 'learning_rate/action_model': 7.882211470447831e-05, 'epoch': 0.93}) +[2026-06-17 20:20:52,333][starVLA.training.train_starvla][INFO] - Step 2261, Loss: {'train/loss': 0.47587278485298157, 'train/grad_norm_pre_clip': 18.339256286621094, 'timing/data': 0.0007865829393267632, 'timing/model': 0.9765362190082669, 'learning_rate/qwen_vl_interface': 7.880410937320367e-06, 'learning_rate/action_model': 7.880410937320368e-05, 'epoch': 0.93}) +[2026-06-17 20:20:56,233][starVLA.training.train_starvla][INFO] - Step 2262, Loss: {'train/loss': 0.6103426218032837, 'train/grad_norm_pre_clip': 11.471369743347168, 'timing/data': 0.0008089151233434677, 'timing/model': 0.9446905199438334, 'learning_rate/qwen_vl_interface': 7.878609858905885e-06, 'learning_rate/action_model': 7.878609858905886e-05, 'epoch': 0.93}) +[2026-06-17 20:21:00,152][starVLA.training.train_starvla][INFO] - Step 2263, Loss: {'train/loss': 0.32208043336868286, 'train/grad_norm_pre_clip': 8.980904579162598, 'timing/data': 0.0007283687591552734, 'timing/model': 0.9499182524159551, 'learning_rate/qwen_vl_interface': 7.876808235577751e-06, 'learning_rate/action_model': 7.87680823557775e-05, 'epoch': 0.93}) +[2026-06-17 20:21:04,119][starVLA.training.train_starvla][INFO] - Step 2264, Loss: {'train/loss': 0.44686195254325867, 'train/grad_norm_pre_clip': 10.836248397827148, 'timing/data': 0.0008793119341135025, 'timing/model': 1.1941645918413997, 'learning_rate/qwen_vl_interface': 7.875006067709444e-06, 'learning_rate/action_model': 7.875006067709443e-05, 'epoch': 0.93}) +[2026-06-17 20:21:08,157][starVLA.training.train_starvla][INFO] - Step 2265, Loss: {'train/loss': 0.4734276533126831, 'train/grad_norm_pre_clip': 9.603577613830566, 'timing/data': 0.000522049143910408, 'timing/model': 1.212916312739253, 'learning_rate/qwen_vl_interface': 7.873203355674553e-06, 'learning_rate/action_model': 7.873203355674552e-05, 'epoch': 0.93}) +[2026-06-17 20:21:11,900][starVLA.training.train_starvla][INFO] - Step 2266, Loss: {'train/loss': 0.5722850561141968, 'train/grad_norm_pre_clip': 14.894299507141113, 'timing/data': 0.0005347942933440208, 'timing/model': 0.9613934466615319, 'learning_rate/qwen_vl_interface': 7.871400099846784e-06, 'learning_rate/action_model': 7.871400099846783e-05, 'epoch': 0.93}) +[2026-06-17 20:21:15,895][starVLA.training.train_starvla][INFO] - Step 2267, Loss: {'train/loss': 0.811001181602478, 'train/grad_norm_pre_clip': 14.564183235168457, 'timing/data': 0.0008480548858642578, 'timing/model': 0.9517513643950224, 'learning_rate/qwen_vl_interface': 7.869596300599954e-06, 'learning_rate/action_model': 7.869596300599953e-05, 'epoch': 0.93}) +[2026-06-17 20:21:19,908][starVLA.training.train_starvla][INFO] - Step 2268, Loss: {'train/loss': 0.44386643171310425, 'train/grad_norm_pre_clip': 14.109146118164062, 'timing/data': 0.0009144367650151253, 'timing/model': 0.9775441754609346, 'learning_rate/qwen_vl_interface': 7.867791958307992e-06, 'learning_rate/action_model': 7.867791958307992e-05, 'epoch': 0.93}) +[2026-06-17 20:21:24,460][starVLA.training.train_starvla][INFO] - Step 2269, Loss: {'train/loss': 0.482925683259964, 'train/grad_norm_pre_clip': 11.28718090057373, 'timing/data': 0.0012864898890256882, 'timing/model': 0.94970419164747, 'learning_rate/qwen_vl_interface': 7.865987073344942e-06, 'learning_rate/action_model': 7.865987073344942e-05, 'epoch': 0.93}) +[2026-06-17 20:21:28,434][starVLA.training.train_starvla][INFO] - Step 2270, Loss: {'train/loss': 0.4651198387145996, 'train/grad_norm_pre_clip': 11.748663902282715, 'timing/data': 0.0006186310201883316, 'timing/model': 1.2418060908094049, 'learning_rate/qwen_vl_interface': 7.864181646084953e-06, 'learning_rate/action_model': 7.864181646084953e-05, 'epoch': 0.93}) +[2026-06-17 20:21:32,328][starVLA.training.train_starvla][INFO] - Step 2271, Loss: {'train/loss': 0.6117199063301086, 'train/grad_norm_pre_clip': 14.194910049438477, 'timing/data': 0.0005453182384371758, 'timing/model': 1.121288982219994, 'learning_rate/qwen_vl_interface': 7.862375676902298e-06, 'learning_rate/action_model': 7.862375676902297e-05, 'epoch': 0.93}) +[2026-06-17 20:21:36,162][starVLA.training.train_starvla][INFO] - Step 2272, Loss: {'train/loss': 0.4012681543827057, 'train/grad_norm_pre_clip': 9.157389640808105, 'timing/data': 0.0008873920887708664, 'timing/model': 0.948991734534502, 'learning_rate/qwen_vl_interface': 7.860569166171353e-06, 'learning_rate/action_model': 7.860569166171354e-05, 'epoch': 0.94}) +[2026-06-17 20:21:40,235][starVLA.training.train_starvla][INFO] - Step 2273, Loss: {'train/loss': 0.41530099511146545, 'train/grad_norm_pre_clip': 9.79051399230957, 'timing/data': 0.000525166280567646, 'timing/model': 0.9477028595283628, 'learning_rate/qwen_vl_interface': 7.858762114266613e-06, 'learning_rate/action_model': 7.858762114266612e-05, 'epoch': 0.94}) +[2026-06-17 20:21:44,191][starVLA.training.train_starvla][INFO] - Step 2274, Loss: {'train/loss': 0.3916149139404297, 'train/grad_norm_pre_clip': 15.13187026977539, 'timing/data': 0.0005508307367563248, 'timing/model': 0.9547078050673008, 'learning_rate/qwen_vl_interface': 7.856954521562677e-06, 'learning_rate/action_model': 7.856954521562677e-05, 'epoch': 0.94}) +[2026-06-17 20:21:48,216][starVLA.training.train_starvla][INFO] - Step 2275, Loss: {'train/loss': 0.41351768374443054, 'train/grad_norm_pre_clip': 13.37876033782959, 'timing/data': 0.000738147646188736, 'timing/model': 1.0496556228026748, 'learning_rate/qwen_vl_interface': 7.855146388434264e-06, 'learning_rate/action_model': 7.855146388434264e-05, 'epoch': 0.94}) +[2026-06-17 20:21:52,154][starVLA.training.train_starvla][INFO] - Step 2276, Loss: {'train/loss': 0.5162652134895325, 'train/grad_norm_pre_clip': 13.516243934631348, 'timing/data': 0.0007219845429062843, 'timing/model': 0.9766492592170835, 'learning_rate/qwen_vl_interface': 7.853337715256202e-06, 'learning_rate/action_model': 7.853337715256203e-05, 'epoch': 0.94}) +[2026-06-17 20:21:56,038][starVLA.training.train_starvla][INFO] - Step 2277, Loss: {'train/loss': 0.5636057257652283, 'train/grad_norm_pre_clip': 12.304123878479004, 'timing/data': 0.0012617046013474464, 'timing/model': 1.1047050217166543, 'learning_rate/qwen_vl_interface': 7.851528502403431e-06, 'learning_rate/action_model': 7.851528502403431e-05, 'epoch': 0.94}) +[2026-06-17 20:21:59,630][starVLA.training.train_starvla][INFO] - Step 2278, Loss: {'train/loss': 0.8879127502441406, 'train/grad_norm_pre_clip': 18.074800491333008, 'timing/data': 0.0008389223366975784, 'timing/model': 0.9340588673949242, 'learning_rate/qwen_vl_interface': 7.849718750251001e-06, 'learning_rate/action_model': 7.849718750251002e-05, 'epoch': 0.94}) +[2026-06-17 20:22:03,438][starVLA.training.train_starvla][INFO] - Step 2279, Loss: {'train/loss': 1.0065168142318726, 'train/grad_norm_pre_clip': 18.341522216796875, 'timing/data': 0.0013332674279808998, 'timing/model': 0.937442984431982, 'learning_rate/qwen_vl_interface': 7.847908459174077e-06, 'learning_rate/action_model': 7.847908459174078e-05, 'epoch': 0.94}) +[2026-06-17 20:22:07,513][starVLA.training.train_starvla][INFO] - Step 2280, Loss: {'train/loss': 0.48168686032295227, 'train/grad_norm_pre_clip': 11.685100555419922, 'timing/data': 0.0006116637960076332, 'timing/model': 1.0121134370565414, 'learning_rate/qwen_vl_interface': 7.846097629547934e-06, 'learning_rate/action_model': 7.846097629547933e-05, 'epoch': 0.94}) +[2026-06-17 20:22:11,790][starVLA.training.train_starvla][INFO] - Step 2281, Loss: {'train/loss': 0.5362381935119629, 'train/grad_norm_pre_clip': 18.489126205444336, 'timing/data': 0.0009365472942590714, 'timing/model': 0.9524447107687593, 'learning_rate/qwen_vl_interface': 7.844286261747959e-06, 'learning_rate/action_model': 7.844286261747958e-05, 'epoch': 0.94}) +[2026-06-17 20:22:15,846][starVLA.training.train_starvla][INFO] - Step 2282, Loss: {'train/loss': 0.3527897000312805, 'train/grad_norm_pre_clip': 11.11667537689209, 'timing/data': 0.0005058906972408295, 'timing/model': 0.9751019282266498, 'learning_rate/qwen_vl_interface': 7.84247435614965e-06, 'learning_rate/action_model': 7.84247435614965e-05, 'epoch': 0.94}) +[2026-06-17 20:22:19,744][starVLA.training.train_starvla][INFO] - Step 2283, Loss: {'train/loss': 0.3283489942550659, 'train/grad_norm_pre_clip': 9.72746467590332, 'timing/data': 0.0005294373258948326, 'timing/model': 1.2052406407892704, 'learning_rate/qwen_vl_interface': 7.840661913128618e-06, 'learning_rate/action_model': 7.840661913128619e-05, 'epoch': 0.94}) +[2026-06-17 20:22:23,498][starVLA.training.train_starvla][INFO] - Step 2284, Loss: {'train/loss': 0.2896508276462555, 'train/grad_norm_pre_clip': 10.839310646057129, 'timing/data': 0.0005210041999816895, 'timing/model': 1.0111778201535344, 'learning_rate/qwen_vl_interface': 7.838848933060583e-06, 'learning_rate/action_model': 7.838848933060583e-05, 'epoch': 0.94}) +[2026-06-17 20:22:27,574][starVLA.training.train_starvla][INFO] - Step 2285, Loss: {'train/loss': 0.3339504897594452, 'train/grad_norm_pre_clip': 10.067610740661621, 'timing/data': 0.0007064482197165489, 'timing/model': 0.961278285831213, 'learning_rate/qwen_vl_interface': 7.83703541632138e-06, 'learning_rate/action_model': 7.83703541632138e-05, 'epoch': 0.94}) +[2026-06-17 20:22:31,538][starVLA.training.train_starvla][INFO] - Step 2286, Loss: {'train/loss': 0.6458309888839722, 'train/grad_norm_pre_clip': 16.704687118530273, 'timing/data': 0.0007545249536633492, 'timing/model': 0.9853898407891393, 'learning_rate/qwen_vl_interface': 7.835221363286949e-06, 'learning_rate/action_model': 7.835221363286949e-05, 'epoch': 0.94}) +[2026-06-17 20:22:35,510][starVLA.training.train_starvla][INFO] - Step 2287, Loss: {'train/loss': 0.558504045009613, 'train/grad_norm_pre_clip': 11.863404273986816, 'timing/data': 0.0005280738696455956, 'timing/model': 0.9471365576609969, 'learning_rate/qwen_vl_interface': 7.83340677433335e-06, 'learning_rate/action_model': 7.83340677433335e-05, 'epoch': 0.94}) +[2026-06-17 20:22:39,501][starVLA.training.train_starvla][INFO] - Step 2288, Loss: {'train/loss': 0.4167533814907074, 'train/grad_norm_pre_clip': 10.92822265625, 'timing/data': 0.0012745475396513939, 'timing/model': 1.0255555175244808, 'learning_rate/qwen_vl_interface': 7.831591649836745e-06, 'learning_rate/action_model': 7.831591649836744e-05, 'epoch': 0.94}) +[2026-06-17 20:22:44,116][starVLA.training.train_starvla][INFO] - Step 2289, Loss: {'train/loss': 0.659558892250061, 'train/grad_norm_pre_clip': 13.73538589477539, 'timing/data': 0.0009935563430190086, 'timing/model': 0.9454808318987489, 'learning_rate/qwen_vl_interface': 7.829775990173413e-06, 'learning_rate/action_model': 7.829775990173413e-05, 'epoch': 0.94}) +[2026-06-17 20:22:48,101][starVLA.training.train_starvla][INFO] - Step 2290, Loss: {'train/loss': 0.5136194229125977, 'train/grad_norm_pre_clip': 11.242485046386719, 'timing/data': 0.0007421644404530525, 'timing/model': 1.199045654386282, 'learning_rate/qwen_vl_interface': 7.827959795719744e-06, 'learning_rate/action_model': 7.827959795719744e-05, 'epoch': 0.94}) +[2026-06-17 20:22:51,818][starVLA.training.train_starvla][INFO] - Step 2291, Loss: {'train/loss': 0.6914559602737427, 'train/grad_norm_pre_clip': 17.96080207824707, 'timing/data': 0.0006796913221478462, 'timing/model': 0.9538883548229933, 'learning_rate/qwen_vl_interface': 7.826143066852233e-06, 'learning_rate/action_model': 7.826143066852233e-05, 'epoch': 0.94}) +[2026-06-17 20:22:56,042][starVLA.training.train_starvla][INFO] - Step 2292, Loss: {'train/loss': 0.4716152548789978, 'train/grad_norm_pre_clip': 12.052111625671387, 'timing/data': 0.0009167594835162163, 'timing/model': 0.96901666931808, 'learning_rate/qwen_vl_interface': 7.824325803947494e-06, 'learning_rate/action_model': 7.824325803947495e-05, 'epoch': 0.94}) +[2026-06-17 20:23:00,063][starVLA.training.train_starvla][INFO] - Step 2293, Loss: {'train/loss': 0.6927199363708496, 'train/grad_norm_pre_clip': 13.536981582641602, 'timing/data': 0.0009924154728651047, 'timing/model': 0.9683058215305209, 'learning_rate/qwen_vl_interface': 7.822508007382246e-06, 'learning_rate/action_model': 7.822508007382245e-05, 'epoch': 0.94}) +[2026-06-17 20:23:04,097][starVLA.training.train_starvla][INFO] - Step 2294, Loss: {'train/loss': 0.468553364276886, 'train/grad_norm_pre_clip': 14.500218391418457, 'timing/data': 0.0007306840270757675, 'timing/model': 0.9607789674773812, 'learning_rate/qwen_vl_interface': 7.820689677533319e-06, 'learning_rate/action_model': 7.82068967753332e-05, 'epoch': 0.94}) +[2026-06-17 20:23:08,069][starVLA.training.train_starvla][INFO] - Step 2295, Loss: {'train/loss': 0.3296159505844116, 'train/grad_norm_pre_clip': 10.29994010925293, 'timing/data': 0.0010618381202220917, 'timing/model': 0.9778615878894925, 'learning_rate/qwen_vl_interface': 7.818870814777655e-06, 'learning_rate/action_model': 7.818870814777655e-05, 'epoch': 0.94}) +[2026-06-17 20:23:12,359][starVLA.training.train_starvla][INFO] - Step 2296, Loss: {'train/loss': 0.5800767540931702, 'train/grad_norm_pre_clip': 19.688732147216797, 'timing/data': 0.0007856171578168869, 'timing/model': 1.2942270562052727, 'learning_rate/qwen_vl_interface': 7.81705141949231e-06, 'learning_rate/action_model': 7.817051419492308e-05, 'epoch': 0.95}) +[2026-06-17 20:23:17,306][starVLA.training.train_starvla][INFO] - Step 2297, Loss: {'train/loss': 0.5468135476112366, 'train/grad_norm_pre_clip': 17.16809844970703, 'timing/data': 0.0013485029339790344, 'timing/model': 1.1113629834726453, 'learning_rate/qwen_vl_interface': 7.81523149205444e-06, 'learning_rate/action_model': 7.81523149205444e-05, 'epoch': 0.95}) +[2026-06-17 20:23:21,089][starVLA.training.train_starvla][INFO] - Step 2298, Loss: {'train/loss': 0.6183455586433411, 'train/grad_norm_pre_clip': 12.696378707885742, 'timing/data': 0.0006232084706425667, 'timing/model': 1.0710069155320525, 'learning_rate/qwen_vl_interface': 7.813411032841323e-06, 'learning_rate/action_model': 7.813411032841323e-05, 'epoch': 0.95}) +[2026-06-17 20:23:25,103][starVLA.training.train_starvla][INFO] - Step 2299, Loss: {'train/loss': 0.5414940714836121, 'train/grad_norm_pre_clip': 14.565317153930664, 'timing/data': 0.0010994160547852516, 'timing/model': 0.9430050505325198, 'learning_rate/qwen_vl_interface': 7.81159004223034e-06, 'learning_rate/action_model': 7.81159004223034e-05, 'epoch': 0.95}) +[2026-06-17 20:25:40,386][starVLA.training.train_starvla][INFO] - Step 2300, Loss: {'train/loss': 0.38637131452560425, 'train/grad_norm_pre_clip': 9.71130084991455, 'eval/loss': 0.5745959281921387, 'eval/latency_6/loss': 0.5745959281921387, 'eval/demon_fire_precision': 0.9274168953949374, 'eval/demon_fire_recall': 0.9828700711053652, 'eval/demon_fire_f1': 0.9543386160364035, 'eval/demon_fire_rate_model': 0.9123539232053423, 'eval/demon_fire_rate_teacher': 0.860879243183083, 'eval/demon_left_precision': 0.8202764976958525, 'eval/demon_left_recall': 0.624561403508772, 'eval/demon_left_f1': 0.709163346613546, 'eval/demon_left_rate_model': 0.12075681691708402, 'eval/demon_left_rate_teacher': 0.15859766277128548, 'eval/demon_right_precision': 0.7441860465116279, 'eval/demon_right_recall': 0.7373271889400922, 'eval/demon_right_f1': 0.7407407407407408, 'eval/demon_right_rate_model': 0.17946577629382304, 'eval/demon_right_rate_teacher': 0.18113522537562604, 'eval/demon_move_f1': 0.7249520436771434, 'eval/demon_cc_f1': 0.8716019248018063, 'eval/latency_6/demon_fire_precision': 0.9274168953949374, 'eval/latency_6/demon_fire_recall': 0.9828700711053652, 'eval/latency_6/demon_fire_f1': 0.9543386160364035, 'eval/latency_6/demon_fire_rate_model': 0.9123539232053423, 'eval/latency_6/demon_fire_rate_teacher': 0.860879243183083, 'eval/latency_6/demon_left_precision': 0.8202764976958525, 'eval/latency_6/demon_left_recall': 0.624561403508772, 'eval/latency_6/demon_left_f1': 0.709163346613546, 'eval/latency_6/demon_left_rate_model': 0.12075681691708402, 'eval/latency_6/demon_left_rate_teacher': 0.15859766277128548, 'eval/latency_6/demon_right_precision': 0.7441860465116279, 'eval/latency_6/demon_right_recall': 0.7373271889400922, 'eval/latency_6/demon_right_f1': 0.7407407407407408, 'eval/latency_6/demon_right_rate_model': 0.17946577629382304, 'eval/latency_6/demon_right_rate_teacher': 0.18113522537562604, 'eval/latency_6/demon_move_f1': 0.7249520436771434, 'eval/latency_6/demon_cc_f1': 0.8716019248018063, 'timing/data': 0.0006101792678236961, 'timing/model': 1.0191979734227061, 'learning_rate/qwen_vl_interface': 7.809768520598986e-06, 'learning_rate/action_model': 7.809768520598984e-05, 'epoch': 0.95}) +[2026-06-17 20:25:40,387][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_2300_state +[2026-06-17 20:25:40,392][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 20:25:46,823][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_2300_state/model.safetensors +[2026-06-17 20:26:02,473][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_2300_state/optimizer.bin +[2026-06-17 20:26:02,473][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_2300_state/sampler.bin +[2026-06-17 20:26:02,474][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_2300_state/sampler_1.bin +[2026-06-17 20:26:02,476][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_2300_state/random_states_0.pkl +[2026-06-17 20:26:02,481][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2300 +[2026-06-17 20:26:02,481][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 20:26:02,492][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 20:26:06,524][starVLA.training.train_starvla][INFO] - Step 2301, Loss: {'train/loss': 0.3513820469379425, 'train/grad_norm_pre_clip': 10.959834098815918, 'timing/data': 0.0005125943571329117, 'timing/model': 1.0444849161431193, 'learning_rate/qwen_vl_interface': 7.80794646832486e-06, 'learning_rate/action_model': 7.80794646832486e-05, 'epoch': 0.95}) +[2026-06-17 20:26:10,348][starVLA.training.train_starvla][INFO] - Step 2302, Loss: {'train/loss': 0.5896632671356201, 'train/grad_norm_pre_clip': 14.478538513183594, 'timing/data': 0.0007928730919957161, 'timing/model': 0.9550555059686303, 'learning_rate/qwen_vl_interface': 7.80612388578568e-06, 'learning_rate/action_model': 7.80612388578568e-05, 'epoch': 0.95}) +[2026-06-17 20:26:14,179][starVLA.training.train_starvla][INFO] - Step 2303, Loss: {'train/loss': 0.7715420722961426, 'train/grad_norm_pre_clip': 17.324365615844727, 'timing/data': 0.001329728402197361, 'timing/model': 0.9998206719756126, 'learning_rate/qwen_vl_interface': 7.80430077335927e-06, 'learning_rate/action_model': 7.80430077335927e-05, 'epoch': 0.95}) +[2026-06-17 20:26:18,207][starVLA.training.train_starvla][INFO] - Step 2304, Loss: {'train/loss': 0.6477625370025635, 'train/grad_norm_pre_clip': 11.433438301086426, 'timing/data': 0.0005585197359323502, 'timing/model': 0.977636069059372, 'learning_rate/qwen_vl_interface': 7.802477131423561e-06, 'learning_rate/action_model': 7.80247713142356e-05, 'epoch': 0.95}) +[2026-06-17 20:26:22,014][starVLA.training.train_starvla][INFO] - Step 2305, Loss: {'train/loss': 0.9232544302940369, 'train/grad_norm_pre_clip': 18.133405685424805, 'timing/data': 0.0005073528736829758, 'timing/model': 1.1100475434213877, 'learning_rate/qwen_vl_interface': 7.800652960356594e-06, 'learning_rate/action_model': 7.800652960356594e-05, 'epoch': 0.95}) +[2026-06-17 20:26:25,968][starVLA.training.train_starvla][INFO] - Step 2306, Loss: {'train/loss': 0.5789932012557983, 'train/grad_norm_pre_clip': 12.228565216064453, 'timing/data': 0.0007831715047359467, 'timing/model': 0.9436116749420762, 'learning_rate/qwen_vl_interface': 7.798828260536524e-06, 'learning_rate/action_model': 7.798828260536524e-05, 'epoch': 0.95}) +[2026-06-17 20:26:31,070][starVLA.training.train_starvla][INFO] - Step 2307, Loss: {'train/loss': 0.6773943305015564, 'train/grad_norm_pre_clip': 13.039300918579102, 'timing/data': 0.0007377667352557182, 'timing/model': 1.0024898741394281, 'learning_rate/qwen_vl_interface': 7.797003032341615e-06, 'learning_rate/action_model': 7.797003032341614e-05, 'epoch': 0.95}) +[2026-06-17 20:26:35,142][starVLA.training.train_starvla][INFO] - Step 2308, Loss: {'train/loss': 0.4682377278804779, 'train/grad_norm_pre_clip': 9.6826753616333, 'timing/data': 0.0005276454612612724, 'timing/model': 0.9623847603797913, 'learning_rate/qwen_vl_interface': 7.795177276150234e-06, 'learning_rate/action_model': 7.795177276150234e-05, 'epoch': 0.95}) +[2026-06-17 20:26:39,034][starVLA.training.train_starvla][INFO] - Step 2309, Loss: {'train/loss': 0.5650990605354309, 'train/grad_norm_pre_clip': 14.561633110046387, 'timing/data': 0.0006744628772139549, 'timing/model': 0.9466774947941303, 'learning_rate/qwen_vl_interface': 7.793350992340867e-06, 'learning_rate/action_model': 7.793350992340867e-05, 'epoch': 0.95}) +[2026-06-17 20:26:42,899][starVLA.training.train_starvla][INFO] - Step 2310, Loss: {'train/loss': 0.7261388301849365, 'train/grad_norm_pre_clip': 12.158548355102539, 'timing/data': 0.0005331235006451607, 'timing/model': 0.954048402607441, 'learning_rate/qwen_vl_interface': 7.791524181292102e-06, 'learning_rate/action_model': 7.791524181292101e-05, 'epoch': 0.95}) +[2026-06-17 20:26:46,768][starVLA.training.train_starvla][INFO] - Step 2311, Loss: {'train/loss': 0.44566527009010315, 'train/grad_norm_pre_clip': 11.194753646850586, 'timing/data': 0.0005141431465744972, 'timing/model': 1.1216964209452271, 'learning_rate/qwen_vl_interface': 7.789696843382638e-06, 'learning_rate/action_model': 7.789696843382639e-05, 'epoch': 0.95}) +[2026-06-17 20:26:50,452][starVLA.training.train_starvla][INFO] - Step 2312, Loss: {'train/loss': 0.5089622139930725, 'train/grad_norm_pre_clip': 11.804340362548828, 'timing/data': 0.0004677698016166687, 'timing/model': 0.9342001052573323, 'learning_rate/qwen_vl_interface': 7.787868978991288e-06, 'learning_rate/action_model': 7.787868978991287e-05, 'epoch': 0.95}) +[2026-06-17 20:26:54,273][starVLA.training.train_starvla][INFO] - Step 2313, Loss: {'train/loss': 0.48322466015815735, 'train/grad_norm_pre_clip': 19.42829132080078, 'timing/data': 0.001807195134460926, 'timing/model': 0.934332437813282, 'learning_rate/qwen_vl_interface': 7.786040588496966e-06, 'learning_rate/action_model': 7.786040588496965e-05, 'epoch': 0.95}) +[2026-06-17 20:26:58,181][starVLA.training.train_starvla][INFO] - Step 2314, Loss: {'train/loss': 0.6011094450950623, 'train/grad_norm_pre_clip': 13.362275123596191, 'timing/data': 0.0006557591259479523, 'timing/model': 0.9434085479006171, 'learning_rate/qwen_vl_interface': 7.784211672278703e-06, 'learning_rate/action_model': 7.784211672278703e-05, 'epoch': 0.95}) +[2026-06-17 20:27:02,046][starVLA.training.train_starvla][INFO] - Step 2315, Loss: {'train/loss': 0.26074904203414917, 'train/grad_norm_pre_clip': 11.090184211730957, 'timing/data': 0.0004978282377123833, 'timing/model': 0.9332350045442581, 'learning_rate/qwen_vl_interface': 7.78238223071563e-06, 'learning_rate/action_model': 7.78238223071563e-05, 'epoch': 0.95}) +[2026-06-17 20:27:06,137][starVLA.training.train_starvla][INFO] - Step 2316, Loss: {'train/loss': 0.5317490696907043, 'train/grad_norm_pre_clip': 13.029213905334473, 'timing/data': 0.0008150599896907806, 'timing/model': 0.9626523982733488, 'learning_rate/qwen_vl_interface': 7.780552264187e-06, 'learning_rate/action_model': 7.780552264186999e-05, 'epoch': 0.95}) +[2026-06-17 20:27:10,035][starVLA.training.train_starvla][INFO] - Step 2317, Loss: {'train/loss': 0.3214818239212036, 'train/grad_norm_pre_clip': 10.870862007141113, 'timing/data': 0.0004821140319108963, 'timing/model': 0.9478869792073965, 'learning_rate/qwen_vl_interface': 7.77872177307216e-06, 'learning_rate/action_model': 7.778721773072159e-05, 'epoch': 0.95}) +[2026-06-17 20:27:14,025][starVLA.training.train_starvla][INFO] - Step 2318, Loss: {'train/loss': 0.48127803206443787, 'train/grad_norm_pre_clip': 11.434883117675781, 'timing/data': 0.001206282526254654, 'timing/model': 1.1262228768318892, 'learning_rate/qwen_vl_interface': 7.776890757750575e-06, 'learning_rate/action_model': 7.776890757750575e-05, 'epoch': 0.95}) +[2026-06-17 20:27:17,695][starVLA.training.train_starvla][INFO] - Step 2319, Loss: {'train/loss': 0.6257427334785461, 'train/grad_norm_pre_clip': 12.584508895874023, 'timing/data': 0.0013950886204838753, 'timing/model': 0.9516097577288747, 'learning_rate/qwen_vl_interface': 7.775059218601816e-06, 'learning_rate/action_model': 7.775059218601816e-05, 'epoch': 0.95}) +[2026-06-17 20:27:21,570][starVLA.training.train_starvla][INFO] - Step 2320, Loss: {'train/loss': 0.8581839799880981, 'train/grad_norm_pre_clip': 15.729673385620117, 'timing/data': 0.000506560318171978, 'timing/model': 0.9427043283358216, 'learning_rate/qwen_vl_interface': 7.773227156005567e-06, 'learning_rate/action_model': 7.773227156005567e-05, 'epoch': 0.95}) +[2026-06-17 20:27:25,974][starVLA.training.train_starvla][INFO] - Step 2321, Loss: {'train/loss': 0.7001829147338867, 'train/grad_norm_pre_clip': 13.383463859558105, 'timing/data': 0.0006690211594104767, 'timing/model': 0.9503178475424647, 'learning_rate/qwen_vl_interface': 7.771394570341612e-06, 'learning_rate/action_model': 7.771394570341612e-05, 'epoch': 0.96}) +[2026-06-17 20:27:29,880][starVLA.training.train_starvla][INFO] - Step 2322, Loss: {'train/loss': 0.7166197896003723, 'train/grad_norm_pre_clip': 17.254602432250977, 'timing/data': 0.0013932418078184128, 'timing/model': 0.9405338503420353, 'learning_rate/qwen_vl_interface': 7.769561461989848e-06, 'learning_rate/action_model': 7.769561461989848e-05, 'epoch': 0.96}) +[2026-06-17 20:27:33,797][starVLA.training.train_starvla][INFO] - Step 2323, Loss: {'train/loss': 0.3321038484573364, 'train/grad_norm_pre_clip': 11.83305549621582, 'timing/data': 0.0005105156451463699, 'timing/model': 0.9396966323256493, 'learning_rate/qwen_vl_interface': 7.767727831330285e-06, 'learning_rate/action_model': 7.767727831330285e-05, 'epoch': 0.96}) +[2026-06-17 20:27:37,647][starVLA.training.train_starvla][INFO] - Step 2324, Loss: {'train/loss': 0.45505571365356445, 'train/grad_norm_pre_clip': 10.005495071411133, 'timing/data': 0.0005231602117419243, 'timing/model': 1.1454193163663149, 'learning_rate/qwen_vl_interface': 7.765893678743029e-06, 'learning_rate/action_model': 7.76589367874303e-05, 'epoch': 0.96}) +[2026-06-17 20:27:41,529][starVLA.training.train_starvla][INFO] - Step 2325, Loss: {'train/loss': 0.38444435596466064, 'train/grad_norm_pre_clip': 10.96714973449707, 'timing/data': 0.0009306911379098892, 'timing/model': 1.1273230919614434, 'learning_rate/qwen_vl_interface': 7.76405900460831e-06, 'learning_rate/action_model': 7.76405900460831e-05, 'epoch': 0.96}) +[2026-06-17 20:27:46,344][starVLA.training.train_starvla][INFO] - Step 2326, Loss: {'train/loss': 0.42401057481765747, 'train/grad_norm_pre_clip': 13.932876586914062, 'timing/data': 0.0008293287828564644, 'timing/model': 0.9377035340294242, 'learning_rate/qwen_vl_interface': 7.762223809306453e-06, 'learning_rate/action_model': 7.762223809306452e-05, 'epoch': 0.96}) +[2026-06-17 20:27:50,198][starVLA.training.train_starvla][INFO] - Step 2327, Loss: {'train/loss': 0.3836754262447357, 'train/grad_norm_pre_clip': 10.36507511138916, 'timing/data': 0.0005266973748803139, 'timing/model': 0.9513555318117142, 'learning_rate/qwen_vl_interface': 7.760388093217897e-06, 'learning_rate/action_model': 7.760388093217898e-05, 'epoch': 0.96}) +[2026-06-17 20:27:54,057][starVLA.training.train_starvla][INFO] - Step 2328, Loss: {'train/loss': 0.6107243299484253, 'train/grad_norm_pre_clip': 13.176475524902344, 'timing/data': 0.00047835614532232285, 'timing/model': 0.9350387351587415, 'learning_rate/qwen_vl_interface': 7.758551856723187e-06, 'learning_rate/action_model': 7.758551856723188e-05, 'epoch': 0.96}) +[2026-06-17 20:27:57,913][starVLA.training.train_starvla][INFO] - Step 2329, Loss: {'train/loss': 0.40937095880508423, 'train/grad_norm_pre_clip': 11.014838218688965, 'timing/data': 0.0006864313036203384, 'timing/model': 0.9397113770246506, 'learning_rate/qwen_vl_interface': 7.756715100202979e-06, 'learning_rate/action_model': 7.75671510020298e-05, 'epoch': 0.96}) +[2026-06-17 20:28:01,839][starVLA.training.train_starvla][INFO] - Step 2330, Loss: {'train/loss': 0.5631526708602905, 'train/grad_norm_pre_clip': 12.875509262084961, 'timing/data': 0.0005143210291862488, 'timing/model': 0.9556617876514792, 'learning_rate/qwen_vl_interface': 7.754877824038035e-06, 'learning_rate/action_model': 7.754877824038035e-05, 'epoch': 0.96}) +[2026-06-17 20:28:05,793][starVLA.training.train_starvla][INFO] - Step 2331, Loss: {'train/loss': 0.7346673011779785, 'train/grad_norm_pre_clip': 12.540461540222168, 'timing/data': 0.0004777936264872551, 'timing/model': 1.1538296090438962, 'learning_rate/qwen_vl_interface': 7.75304002860922e-06, 'learning_rate/action_model': 7.75304002860922e-05, 'epoch': 0.96}) +[2026-06-17 20:28:09,441][starVLA.training.train_starvla][INFO] - Step 2332, Loss: {'train/loss': 0.4920992851257324, 'train/grad_norm_pre_clip': 12.625663757324219, 'timing/data': 0.0004597604274749756, 'timing/model': 0.9568760208785534, 'learning_rate/qwen_vl_interface': 7.751201714297513e-06, 'learning_rate/action_model': 7.751201714297512e-05, 'epoch': 0.96}) +[2026-06-17 20:28:13,316][starVLA.training.train_starvla][INFO] - Step 2333, Loss: {'train/loss': 0.7142353057861328, 'train/grad_norm_pre_clip': 16.749425888061523, 'timing/data': 0.0005876291543245316, 'timing/model': 0.949457187205553, 'learning_rate/qwen_vl_interface': 7.749362881484002e-06, 'learning_rate/action_model': 7.749362881484001e-05, 'epoch': 0.96}) +[2026-06-17 20:28:17,191][starVLA.training.train_starvla][INFO] - Step 2334, Loss: {'train/loss': 0.6770380139350891, 'train/grad_norm_pre_clip': 11.443222045898438, 'timing/data': 0.0004990296438336372, 'timing/model': 0.9501573862507939, 'learning_rate/qwen_vl_interface': 7.747523530549874e-06, 'learning_rate/action_model': 7.747523530549875e-05, 'epoch': 0.96}) +[2026-06-17 20:28:21,047][starVLA.training.train_starvla][INFO] - Step 2335, Loss: {'train/loss': 0.577256441116333, 'train/grad_norm_pre_clip': 15.246618270874023, 'timing/data': 0.0005822246894240379, 'timing/model': 0.9455295819789171, 'learning_rate/qwen_vl_interface': 7.745683661876432e-06, 'learning_rate/action_model': 7.745683661876432e-05, 'epoch': 0.96}) +[2026-06-17 20:28:25,061][starVLA.training.train_starvla][INFO] - Step 2336, Loss: {'train/loss': 0.42666375637054443, 'train/grad_norm_pre_clip': 8.875091552734375, 'timing/data': 0.001040617935359478, 'timing/model': 0.9506579395383596, 'learning_rate/qwen_vl_interface': 7.743843275845081e-06, 'learning_rate/action_model': 7.74384327584508e-05, 'epoch': 0.96}) +[2026-06-17 20:28:29,200][starVLA.training.train_starvla][INFO] - Step 2337, Loss: {'train/loss': 0.5543668866157532, 'train/grad_norm_pre_clip': 13.427574157714844, 'timing/data': 0.000506032258272171, 'timing/model': 0.9476866843178868, 'learning_rate/qwen_vl_interface': 7.742002372837336e-06, 'learning_rate/action_model': 7.742002372837335e-05, 'epoch': 0.96}) +[2026-06-17 20:28:33,244][starVLA.training.train_starvla][INFO] - Step 2338, Loss: {'train/loss': 0.7497310638427734, 'train/grad_norm_pre_clip': 17.455629348754883, 'timing/data': 0.0008081737905740738, 'timing/model': 1.1456056935712695, 'learning_rate/qwen_vl_interface': 7.740160953234815e-06, 'learning_rate/action_model': 7.740160953234815e-05, 'epoch': 0.96}) +[2026-06-17 20:28:36,958][starVLA.training.train_starvla][INFO] - Step 2339, Loss: {'train/loss': 0.41821324825286865, 'train/grad_norm_pre_clip': 10.181059837341309, 'timing/data': 0.0006007859483361244, 'timing/model': 0.9438849128782749, 'learning_rate/qwen_vl_interface': 7.738319017419252e-06, 'learning_rate/action_model': 7.738319017419252e-05, 'epoch': 0.96}) +[2026-06-17 20:28:41,265][starVLA.training.train_starvla][INFO] - Step 2340, Loss: {'train/loss': 0.31645336747169495, 'train/grad_norm_pre_clip': 11.294098854064941, 'timing/data': 0.0006494782865047455, 'timing/model': 0.9595635524019599, 'learning_rate/qwen_vl_interface': 7.736476565772478e-06, 'learning_rate/action_model': 7.736476565772478e-05, 'epoch': 0.96}) +[2026-06-17 20:28:46,986][starVLA.training.train_starvla][INFO] - Step 2341, Loss: {'train/loss': 0.515402615070343, 'train/grad_norm_pre_clip': 16.75140953063965, 'timing/data': 0.0005874466150999069, 'timing/model': 0.9466229267418385, 'learning_rate/qwen_vl_interface': 7.734633598676435e-06, 'learning_rate/action_model': 7.734633598676434e-05, 'epoch': 0.96}) +[2026-06-17 20:28:50,867][starVLA.training.train_starvla][INFO] - Step 2342, Loss: {'train/loss': 0.6390779614448547, 'train/grad_norm_pre_clip': 13.88298511505127, 'timing/data': 0.0006121154874563217, 'timing/model': 0.9358639726415277, 'learning_rate/qwen_vl_interface': 7.732790116513176e-06, 'learning_rate/action_model': 7.732790116513176e-05, 'epoch': 0.96}) +[2026-06-17 20:28:55,000][starVLA.training.train_starvla][INFO] - Step 2343, Loss: {'train/loss': 0.6349865198135376, 'train/grad_norm_pre_clip': 12.142531394958496, 'timing/data': 0.0005021318793296814, 'timing/model': 1.020285346545279, 'learning_rate/qwen_vl_interface': 7.730946119664852e-06, 'learning_rate/action_model': 7.730946119664853e-05, 'epoch': 0.96}) +[2026-06-17 20:28:59,078][starVLA.training.train_starvla][INFO] - Step 2344, Loss: {'train/loss': 0.8375356197357178, 'train/grad_norm_pre_clip': 15.157150268554688, 'timing/data': 0.00103826355189085, 'timing/model': 0.9889099784195423, 'learning_rate/qwen_vl_interface': 7.729101608513729e-06, 'learning_rate/action_model': 7.72910160851373e-05, 'epoch': 0.96}) +[2026-06-17 20:29:02,996][starVLA.training.train_starvla][INFO] - Step 2345, Loss: {'train/loss': 0.6406167149543762, 'train/grad_norm_pre_clip': 11.464616775512695, 'timing/data': 0.0004852376878261566, 'timing/model': 1.1530955312773585, 'learning_rate/qwen_vl_interface': 7.727256583442175e-06, 'learning_rate/action_model': 7.727256583442175e-05, 'epoch': 0.97}) +[2026-06-17 20:29:06,686][starVLA.training.train_starvla][INFO] - Step 2346, Loss: {'train/loss': 0.5802357792854309, 'train/grad_norm_pre_clip': 12.394347190856934, 'timing/data': 0.0008186418563127518, 'timing/model': 0.9846424302086234, 'learning_rate/qwen_vl_interface': 7.725411044832665e-06, 'learning_rate/action_model': 7.725411044832665e-05, 'epoch': 0.97}) +[2026-06-17 20:29:10,795][starVLA.training.train_starvla][INFO] - Step 2347, Loss: {'train/loss': 0.3979291617870331, 'train/grad_norm_pre_clip': 8.07508373260498, 'timing/data': 0.0007960079237818718, 'timing/model': 0.9611852569505572, 'learning_rate/qwen_vl_interface': 7.723564993067783e-06, 'learning_rate/action_model': 7.723564993067783e-05, 'epoch': 0.97}) +[2026-06-17 20:29:14,717][starVLA.training.train_starvla][INFO] - Step 2348, Loss: {'train/loss': 0.23111800849437714, 'train/grad_norm_pre_clip': 7.038736343383789, 'timing/data': 0.0005367416888475418, 'timing/model': 0.9480019249022007, 'learning_rate/qwen_vl_interface': 7.721718428530214e-06, 'learning_rate/action_model': 7.721718428530214e-05, 'epoch': 0.97}) +[2026-06-17 20:29:18,658][starVLA.training.train_starvla][INFO] - Step 2349, Loss: {'train/loss': 0.5103062391281128, 'train/grad_norm_pre_clip': 9.745856285095215, 'timing/data': 0.0008041160181164742, 'timing/model': 0.9677850445732474, 'learning_rate/qwen_vl_interface': 7.719871351602757e-06, 'learning_rate/action_model': 7.719871351602757e-05, 'epoch': 0.97}) +[2026-06-17 20:29:22,544][starVLA.training.train_starvla][INFO] - Step 2350, Loss: {'train/loss': 0.4191296398639679, 'train/grad_norm_pre_clip': 10.646147727966309, 'timing/data': 0.0017982721328735352, 'timing/model': 0.9476620890200138, 'learning_rate/qwen_vl_interface': 7.718023762668309e-06, 'learning_rate/action_model': 7.718023762668309e-05, 'epoch': 0.97}) +[2026-06-17 20:29:26,404][starVLA.training.train_starvla][INFO] - Step 2351, Loss: {'train/loss': 0.43497616052627563, 'train/grad_norm_pre_clip': 9.149614334106445, 'timing/data': 0.0005070716142654419, 'timing/model': 0.9523097276687622, 'learning_rate/qwen_vl_interface': 7.716175662109882e-06, 'learning_rate/action_model': 7.716175662109881e-05, 'epoch': 0.97}) +[2026-06-17 20:29:30,206][starVLA.training.train_starvla][INFO] - Step 2352, Loss: {'train/loss': 0.38867151737213135, 'train/grad_norm_pre_clip': 11.114158630371094, 'timing/data': 0.0005243038758635521, 'timing/model': 1.11924350541085, 'learning_rate/qwen_vl_interface': 7.714327050310586e-06, 'learning_rate/action_model': 7.714327050310585e-05, 'epoch': 0.97}) +[2026-06-17 20:29:33,846][starVLA.training.train_starvla][INFO] - Step 2353, Loss: {'train/loss': 0.5922443866729736, 'train/grad_norm_pre_clip': 23.990530014038086, 'timing/data': 0.0008554086089134216, 'timing/model': 0.943305972032249, 'learning_rate/qwen_vl_interface': 7.71247792765364e-06, 'learning_rate/action_model': 7.712477927653638e-05, 'epoch': 0.97}) +[2026-06-17 20:29:37,757][starVLA.training.train_starvla][INFO] - Step 2354, Loss: {'train/loss': 0.32360976934432983, 'train/grad_norm_pre_clip': 7.836001396179199, 'timing/data': 0.0004817526787519455, 'timing/model': 0.9421223774552345, 'learning_rate/qwen_vl_interface': 7.71062829452237e-06, 'learning_rate/action_model': 7.710628294522369e-05, 'epoch': 0.97}) +[2026-06-17 20:29:41,540][starVLA.training.train_starvla][INFO] - Step 2355, Loss: {'train/loss': 0.43832558393478394, 'train/grad_norm_pre_clip': 10.707317352294922, 'timing/data': 0.0010116761550307274, 'timing/model': 0.9460218073800206, 'learning_rate/qwen_vl_interface': 7.708778151300206e-06, 'learning_rate/action_model': 7.708778151300206e-05, 'epoch': 0.97}) +[2026-06-17 20:29:45,389][starVLA.training.train_starvla][INFO] - Step 2356, Loss: {'train/loss': 0.8599193692207336, 'train/grad_norm_pre_clip': 17.82622528076172, 'timing/data': 0.001036916859447956, 'timing/model': 0.9545403700321913, 'learning_rate/qwen_vl_interface': 7.706927498370687e-06, 'learning_rate/action_model': 7.706927498370686e-05, 'epoch': 0.97}) +[2026-06-17 20:29:49,259][starVLA.training.train_starvla][INFO] - Step 2357, Loss: {'train/loss': 0.36796000599861145, 'train/grad_norm_pre_clip': 12.078468322753906, 'timing/data': 0.0011321725323796272, 'timing/model': 0.9358540279790759, 'learning_rate/qwen_vl_interface': 7.705076336117452e-06, 'learning_rate/action_model': 7.705076336117451e-05, 'epoch': 0.97}) +[2026-06-17 20:29:53,158][starVLA.training.train_starvla][INFO] - Step 2358, Loss: {'train/loss': 0.6971685886383057, 'train/grad_norm_pre_clip': 13.945791244506836, 'timing/data': 0.0005135424435138702, 'timing/model': 1.1405376754701138, 'learning_rate/qwen_vl_interface': 7.703224664924252e-06, 'learning_rate/action_model': 7.703224664924251e-05, 'epoch': 0.97}) +[2026-06-17 20:29:56,977][starVLA.training.train_starvla][INFO] - Step 2359, Loss: {'train/loss': 0.4135655164718628, 'train/grad_norm_pre_clip': 13.66756820678711, 'timing/data': 0.0005030697211623192, 'timing/model': 1.1227656342089176, 'learning_rate/qwen_vl_interface': 7.70137248517494e-06, 'learning_rate/action_model': 7.701372485174939e-05, 'epoch': 0.97}) +[2026-06-17 20:30:00,622][starVLA.training.train_starvla][INFO] - Step 2360, Loss: {'train/loss': 0.4178333282470703, 'train/grad_norm_pre_clip': 13.877554893493652, 'timing/data': 0.0004909764975309372, 'timing/model': 0.9397017201408744, 'learning_rate/qwen_vl_interface': 7.699519797253471e-06, 'learning_rate/action_model': 7.699519797253472e-05, 'epoch': 0.97}) +[2026-06-17 20:30:04,470][starVLA.training.train_starvla][INFO] - Step 2361, Loss: {'train/loss': 0.3366067707538605, 'train/grad_norm_pre_clip': 12.428336143493652, 'timing/data': 0.0008590565994381905, 'timing/model': 0.9311548983678222, 'learning_rate/qwen_vl_interface': 7.697666601543915e-06, 'learning_rate/action_model': 7.697666601543915e-05, 'epoch': 0.97}) +[2026-06-17 20:30:08,423][starVLA.training.train_starvla][INFO] - Step 2362, Loss: {'train/loss': 0.6142516136169434, 'train/grad_norm_pre_clip': 14.23840045928955, 'timing/data': 0.0007328903302550316, 'timing/model': 0.9519373197108507, 'learning_rate/qwen_vl_interface': 7.695812898430438e-06, 'learning_rate/action_model': 7.695812898430437e-05, 'epoch': 0.97}) +[2026-06-17 20:30:12,268][starVLA.training.train_starvla][INFO] - Step 2363, Loss: {'train/loss': 0.4149480164051056, 'train/grad_norm_pre_clip': 10.741358757019043, 'timing/data': 0.001024516299366951, 'timing/model': 0.9423319455236197, 'learning_rate/qwen_vl_interface': 7.693958688297316e-06, 'learning_rate/action_model': 7.693958688297314e-05, 'epoch': 0.97}) +[2026-06-17 20:30:16,108][starVLA.training.train_starvla][INFO] - Step 2364, Loss: {'train/loss': 0.7235345244407654, 'train/grad_norm_pre_clip': 17.262460708618164, 'timing/data': 0.0010610604658722878, 'timing/model': 0.9474662691354752, 'learning_rate/qwen_vl_interface': 7.692103971528925e-06, 'learning_rate/action_model': 7.692103971528926e-05, 'epoch': 0.97}) +[2026-06-17 20:30:19,906][starVLA.training.train_starvla][INFO] - Step 2365, Loss: {'train/loss': 0.5945817232131958, 'train/grad_norm_pre_clip': 14.54985237121582, 'timing/data': 0.0010685306042432785, 'timing/model': 1.1180928749963641, 'learning_rate/qwen_vl_interface': 7.690248748509757e-06, 'learning_rate/action_model': 7.690248748509757e-05, 'epoch': 0.97}) +[2026-06-17 20:30:23,535][starVLA.training.train_starvla][INFO] - Step 2366, Loss: {'train/loss': 0.3201403021812439, 'train/grad_norm_pre_clip': 11.412970542907715, 'timing/data': 0.001348152756690979, 'timing/model': 0.9414442786946893, 'learning_rate/qwen_vl_interface': 7.688393019624396e-06, 'learning_rate/action_model': 7.688393019624395e-05, 'epoch': 0.97}) +[2026-06-17 20:30:27,397][starVLA.training.train_starvla][INFO] - Step 2367, Loss: {'train/loss': 0.44815793633461, 'train/grad_norm_pre_clip': 16.8200626373291, 'timing/data': 0.0005344860255718231, 'timing/model': 1.0172296781092882, 'learning_rate/qwen_vl_interface': 7.686536785257539e-06, 'learning_rate/action_model': 7.686536785257538e-05, 'epoch': 0.97}) +[2026-06-17 20:30:31,234][starVLA.training.train_starvla][INFO] - Step 2368, Loss: {'train/loss': 0.6484781503677368, 'train/grad_norm_pre_clip': 23.39986801147461, 'timing/data': 0.0005082413554191589, 'timing/model': 0.9462410435080528, 'learning_rate/qwen_vl_interface': 7.684680045793982e-06, 'learning_rate/action_model': 7.684680045793983e-05, 'epoch': 0.97}) +[2026-06-17 20:30:35,006][starVLA.training.train_starvla][INFO] - Step 2369, Loss: {'train/loss': 0.3083282709121704, 'train/grad_norm_pre_clip': 8.673727035522461, 'timing/data': 0.0007329490035772324, 'timing/model': 0.9366384223103523, 'learning_rate/qwen_vl_interface': 7.682822801618634e-06, 'learning_rate/action_model': 7.682822801618633e-05, 'epoch': 0.98}) +[2026-06-17 20:30:38,837][starVLA.training.train_starvla][INFO] - Step 2370, Loss: {'train/loss': 0.5081430673599243, 'train/grad_norm_pre_clip': 15.752202033996582, 'timing/data': 0.0007336875423789024, 'timing/model': 0.9331092452630401, 'learning_rate/qwen_vl_interface': 7.680965053116498e-06, 'learning_rate/action_model': 7.680965053116499e-05, 'epoch': 0.98}) +[2026-06-17 20:30:42,974][starVLA.training.train_starvla][INFO] - Step 2371, Loss: {'train/loss': 0.3529747724533081, 'train/grad_norm_pre_clip': 84.00341033935547, 'timing/data': 0.0004948601126670837, 'timing/model': 1.3067948212847114, 'learning_rate/qwen_vl_interface': 7.679106800672695e-06, 'learning_rate/action_model': 7.679106800672695e-05, 'epoch': 0.98}) +[2026-06-17 20:30:46,786][starVLA.training.train_starvla][INFO] - Step 2372, Loss: {'train/loss': 0.37677738070487976, 'train/grad_norm_pre_clip': 11.042999267578125, 'timing/data': 0.0005105650052428246, 'timing/model': 1.125905230641365, 'learning_rate/qwen_vl_interface': 7.677248044672438e-06, 'learning_rate/action_model': 7.677248044672437e-05, 'epoch': 0.98}) +[2026-06-17 20:30:50,821][starVLA.training.train_starvla][INFO] - Step 2373, Loss: {'train/loss': 0.3443988561630249, 'train/grad_norm_pre_clip': 11.701127052307129, 'timing/data': 0.0005578584969043732, 'timing/model': 0.9307562736794353, 'learning_rate/qwen_vl_interface': 7.675388785501048e-06, 'learning_rate/action_model': 7.675388785501047e-05, 'epoch': 0.98}) +[2026-06-17 20:30:54,619][starVLA.training.train_starvla][INFO] - Step 2374, Loss: {'train/loss': 0.7454245090484619, 'train/grad_norm_pre_clip': 15.533890724182129, 'timing/data': 0.00048618391156196594, 'timing/model': 0.9378553526476026, 'learning_rate/qwen_vl_interface': 7.673529023543951e-06, 'learning_rate/action_model': 7.673529023543951e-05, 'epoch': 0.98}) +[2026-06-17 20:30:58,361][starVLA.training.train_starvla][INFO] - Step 2375, Loss: {'train/loss': 0.8031330108642578, 'train/grad_norm_pre_clip': 15.784371376037598, 'timing/data': 0.0005134977400302887, 'timing/model': 0.9334293436259031, 'learning_rate/qwen_vl_interface': 7.671668759186683e-06, 'learning_rate/action_model': 7.671668759186683e-05, 'epoch': 0.98}) +[2026-06-17 20:31:02,169][starVLA.training.train_starvla][INFO] - Step 2376, Loss: {'train/loss': 0.3960384726524353, 'train/grad_norm_pre_clip': 12.143269538879395, 'timing/data': 0.0007975520566105843, 'timing/model': 0.9823686648160219, 'learning_rate/qwen_vl_interface': 7.669807992814875e-06, 'learning_rate/action_model': 7.669807992814875e-05, 'epoch': 0.98}) +[2026-06-17 20:31:06,086][starVLA.training.train_starvla][INFO] - Step 2377, Loss: {'train/loss': 0.5454398393630981, 'train/grad_norm_pre_clip': 11.254290580749512, 'timing/data': 0.0011002598330378532, 'timing/model': 1.0175508186221123, 'learning_rate/qwen_vl_interface': 7.667946724814264e-06, 'learning_rate/action_model': 7.667946724814265e-05, 'epoch': 0.98}) +[2026-06-17 20:31:09,940][starVLA.training.train_starvla][INFO] - Step 2378, Loss: {'train/loss': 0.418959379196167, 'train/grad_norm_pre_clip': 12.749165534973145, 'timing/data': 0.000799546018242836, 'timing/model': 1.1309942184016109, 'learning_rate/qwen_vl_interface': 7.666084955570698e-06, 'learning_rate/action_model': 7.666084955570698e-05, 'epoch': 0.98}) +[2026-06-17 20:31:13,728][starVLA.training.train_starvla][INFO] - Step 2379, Loss: {'train/loss': 0.5237528085708618, 'train/grad_norm_pre_clip': 10.506139755249023, 'timing/data': 0.0005656955763697624, 'timing/model': 0.9738040789961815, 'learning_rate/qwen_vl_interface': 7.664222685470119e-06, 'learning_rate/action_model': 7.664222685470118e-05, 'epoch': 0.98}) +[2026-06-17 20:31:17,538][starVLA.training.train_starvla][INFO] - Step 2380, Loss: {'train/loss': 0.22928716242313385, 'train/grad_norm_pre_clip': 7.1089067459106445, 'timing/data': 0.001258869655430317, 'timing/model': 0.9501990862190723, 'learning_rate/qwen_vl_interface': 7.66235991489858e-06, 'learning_rate/action_model': 7.66235991489858e-05, 'epoch': 0.98}) +[2026-06-17 20:31:21,700][starVLA.training.train_starvla][INFO] - Step 2381, Loss: {'train/loss': 0.5588774681091309, 'train/grad_norm_pre_clip': 13.571356773376465, 'timing/data': 0.0004956750199198723, 'timing/model': 0.9483727039769292, 'learning_rate/qwen_vl_interface': 7.660496644242233e-06, 'learning_rate/action_model': 7.660496644242233e-05, 'epoch': 0.98}) +[2026-06-17 20:31:25,659][starVLA.training.train_starvla][INFO] - Step 2382, Loss: {'train/loss': 0.5178155303001404, 'train/grad_norm_pre_clip': 14.309113502502441, 'timing/data': 0.0005332194268703461, 'timing/model': 0.9850525036454201, 'learning_rate/qwen_vl_interface': 7.658632873887342e-06, 'learning_rate/action_model': 7.658632873887342e-05, 'epoch': 0.98}) +[2026-06-17 20:31:29,510][starVLA.training.train_starvla][INFO] - Step 2383, Loss: {'train/loss': 0.36730635166168213, 'train/grad_norm_pre_clip': 9.554530143737793, 'timing/data': 0.0004975451156497002, 'timing/model': 0.9508373904973269, 'learning_rate/qwen_vl_interface': 7.656768604220262e-06, 'learning_rate/action_model': 7.656768604220261e-05, 'epoch': 0.98}) +[2026-06-17 20:31:33,457][starVLA.training.train_starvla][INFO] - Step 2384, Loss: {'train/loss': 0.2703112065792084, 'train/grad_norm_pre_clip': 10.524682998657227, 'timing/data': 0.000652506947517395, 'timing/model': 0.9453050754964352, 'learning_rate/qwen_vl_interface': 7.654903835627459e-06, 'learning_rate/action_model': 7.654903835627458e-05, 'epoch': 0.98}) +[2026-06-17 20:31:37,308][starVLA.training.train_starvla][INFO] - Step 2385, Loss: {'train/loss': 0.39452865719795227, 'train/grad_norm_pre_clip': 9.64969539642334, 'timing/data': 0.0005059372633695602, 'timing/model': 1.1324596023187041, 'learning_rate/qwen_vl_interface': 7.653038568495505e-06, 'learning_rate/action_model': 7.653038568495506e-05, 'epoch': 0.98}) +[2026-06-17 20:31:41,437][starVLA.training.train_starvla][INFO] - Step 2386, Loss: {'train/loss': 0.5565407276153564, 'train/grad_norm_pre_clip': 12.900465965270996, 'timing/data': 0.0008240034803748131, 'timing/model': 0.9373175529763103, 'learning_rate/qwen_vl_interface': 7.651172803211068e-06, 'learning_rate/action_model': 7.651172803211068e-05, 'epoch': 0.98}) +[2026-06-17 20:31:45,362][starVLA.training.train_starvla][INFO] - Step 2387, Loss: {'train/loss': 0.43838071823120117, 'train/grad_norm_pre_clip': 10.527023315429688, 'timing/data': 0.0013121552765369415, 'timing/model': 0.9723828863352537, 'learning_rate/qwen_vl_interface': 7.649306540160928e-06, 'learning_rate/action_model': 7.649306540160927e-05, 'epoch': 0.98}) +[2026-06-17 20:31:49,236][starVLA.training.train_starvla][INFO] - Step 2388, Loss: {'train/loss': 0.3885285258293152, 'train/grad_norm_pre_clip': 9.52818775177002, 'timing/data': 0.0005260221660137177, 'timing/model': 0.9602855686098337, 'learning_rate/qwen_vl_interface': 7.647439779731959e-06, 'learning_rate/action_model': 7.647439779731958e-05, 'epoch': 0.98}) +[2026-06-17 20:31:53,038][starVLA.training.train_starvla][INFO] - Step 2389, Loss: {'train/loss': 0.8653073310852051, 'train/grad_norm_pre_clip': 11.191373825073242, 'timing/data': 0.0006385650485754013, 'timing/model': 0.9312100503593683, 'learning_rate/qwen_vl_interface': 7.645572522311144e-06, 'learning_rate/action_model': 7.645572522311144e-05, 'epoch': 0.98}) +[2026-06-17 20:31:56,901][starVLA.training.train_starvla][INFO] - Step 2390, Loss: {'train/loss': 0.3597866892814636, 'train/grad_norm_pre_clip': 13.676025390625, 'timing/data': 0.0005240589380264282, 'timing/model': 0.9482554392889142, 'learning_rate/qwen_vl_interface': 7.643704768285568e-06, 'learning_rate/action_model': 7.643704768285568e-05, 'epoch': 0.98}) +[2026-06-17 20:32:00,714][starVLA.training.train_starvla][INFO] - Step 2391, Loss: {'train/loss': 0.5034844875335693, 'train/grad_norm_pre_clip': 14.349408149719238, 'timing/data': 0.0005036415532231331, 'timing/model': 1.1309326821938157, 'learning_rate/qwen_vl_interface': 7.641836518042417e-06, 'learning_rate/action_model': 7.641836518042416e-05, 'epoch': 0.98}) +[2026-06-17 20:32:04,522][starVLA.training.train_starvla][INFO] - Step 2392, Loss: {'train/loss': 0.46911144256591797, 'train/grad_norm_pre_clip': 12.486623764038086, 'timing/data': 0.0005878480151295662, 'timing/model': 1.112455246038735, 'learning_rate/qwen_vl_interface': 7.639967771968983e-06, 'learning_rate/action_model': 7.639967771968983e-05, 'epoch': 0.98}) +[2026-06-17 20:32:08,196][starVLA.training.train_starvla][INFO] - Step 2393, Loss: {'train/loss': 0.33275794982910156, 'train/grad_norm_pre_clip': 10.557239532470703, 'timing/data': 0.0005172742530703545, 'timing/model': 0.9501718580722809, 'learning_rate/qwen_vl_interface': 7.638098530452657e-06, 'learning_rate/action_model': 7.638098530452658e-05, 'epoch': 0.99}) +[2026-06-17 20:32:12,058][starVLA.training.train_starvla][INFO] - Step 2394, Loss: {'train/loss': 0.7485030889511108, 'train/grad_norm_pre_clip': 17.28492546081543, 'timing/data': 0.0013192994520068169, 'timing/model': 0.9653333183377981, 'learning_rate/qwen_vl_interface': 7.636228793880935e-06, 'learning_rate/action_model': 7.636228793880935e-05, 'epoch': 0.99}) +[2026-06-17 20:32:16,038][starVLA.training.train_starvla][INFO] - Step 2395, Loss: {'train/loss': 0.9235817193984985, 'train/grad_norm_pre_clip': 16.874942779541016, 'timing/data': 0.0007274020463228226, 'timing/model': 0.9402729840949178, 'learning_rate/qwen_vl_interface': 7.63435856264142e-06, 'learning_rate/action_model': 7.63435856264142e-05, 'epoch': 0.99}) +[2026-06-17 20:32:19,828][starVLA.training.train_starvla][INFO] - Step 2396, Loss: {'train/loss': 0.5082586407661438, 'train/grad_norm_pre_clip': 10.36424446105957, 'timing/data': 0.0016413712874054909, 'timing/model': 0.9330683629959822, 'learning_rate/qwen_vl_interface': 7.632487837121806e-06, 'learning_rate/action_model': 7.632487837121806e-05, 'epoch': 0.99}) +[2026-06-17 20:32:23,635][starVLA.training.train_starvla][INFO] - Step 2397, Loss: {'train/loss': 0.6641250848770142, 'train/grad_norm_pre_clip': 19.291547775268555, 'timing/data': 0.0011263135820627213, 'timing/model': 0.9391179159283638, 'learning_rate/qwen_vl_interface': 7.630616617709902e-06, 'learning_rate/action_model': 7.630616617709902e-05, 'epoch': 0.99}) +[2026-06-17 20:32:27,540][starVLA.training.train_starvla][INFO] - Step 2398, Loss: {'train/loss': 0.5847627520561218, 'train/grad_norm_pre_clip': 17.457834243774414, 'timing/data': 0.0010882820934057236, 'timing/model': 1.1898157354444265, 'learning_rate/qwen_vl_interface': 7.6287449047936095e-06, 'learning_rate/action_model': 7.628744904793609e-05, 'epoch': 0.99}) +[2026-06-17 20:32:31,308][starVLA.training.train_starvla][INFO] - Step 2399, Loss: {'train/loss': 0.3689310848712921, 'train/grad_norm_pre_clip': 10.390801429748535, 'timing/data': 0.0006034579128026962, 'timing/model': 1.007824994623661, 'learning_rate/qwen_vl_interface': 7.6268726987609384e-06, 'learning_rate/action_model': 7.626872698760939e-05, 'epoch': 0.99}) +[2026-06-17 20:34:37,816][starVLA.training.train_starvla][INFO] - Step 2400, Loss: {'train/loss': 0.6453414559364319, 'train/grad_norm_pre_clip': 13.869504928588867, 'eval/loss': 0.552358865737915, 'eval/latency_6/loss': 0.552358865737915, 'eval/demon_fire_precision': 0.9344413665743305, 'eval/demon_fire_recall': 0.9812540400775694, 'eval/demon_fire_f1': 0.9572757370329495, 'eval/demon_fire_rate_model': 0.9040066777963273, 'eval/demon_fire_rate_teacher': 0.860879243183083, 'eval/demon_left_precision': 0.8429951690821256, 'eval/demon_left_recall': 0.612280701754386, 'eval/demon_left_f1': 0.709349593495935, 'eval/demon_left_rate_model': 0.11519198664440734, 'eval/demon_left_rate_teacher': 0.15859766277128548, 'eval/demon_right_precision': 0.8369351669941061, 'eval/demon_right_recall': 0.6543778801843319, 'eval/demon_right_f1': 0.7344827586206897, 'eval/demon_right_rate_model': 0.1416249304396216, 'eval/demon_right_rate_teacher': 0.18113522537562604, 'eval/demon_move_f1': 0.7219161760583124, 'eval/demon_cc_f1': 0.8719897452176476, 'eval/latency_6/demon_fire_precision': 0.9344413665743305, 'eval/latency_6/demon_fire_recall': 0.9812540400775694, 'eval/latency_6/demon_fire_f1': 0.9572757370329495, 'eval/latency_6/demon_fire_rate_model': 0.9040066777963273, 'eval/latency_6/demon_fire_rate_teacher': 0.860879243183083, 'eval/latency_6/demon_left_precision': 0.8429951690821256, 'eval/latency_6/demon_left_recall': 0.612280701754386, 'eval/latency_6/demon_left_f1': 0.709349593495935, 'eval/latency_6/demon_left_rate_model': 0.11519198664440734, 'eval/latency_6/demon_left_rate_teacher': 0.15859766277128548, 'eval/latency_6/demon_right_precision': 0.8369351669941061, 'eval/latency_6/demon_right_recall': 0.6543778801843319, 'eval/latency_6/demon_right_f1': 0.7344827586206897, 'eval/latency_6/demon_right_rate_model': 0.1416249304396216, 'eval/latency_6/demon_right_rate_teacher': 0.18113522537562604, 'eval/latency_6/demon_move_f1': 0.7219161760583124, 'eval/latency_6/demon_cc_f1': 0.8719897452176476, 'timing/data': 0.0009577935561537743, 'timing/model': 0.9449946815147996, 'learning_rate/qwen_vl_interface': 7.625e-06, 'learning_rate/action_model': 7.625e-05, 'epoch': 0.99}) +[2026-06-17 20:34:37,818][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_2400_state +[2026-06-17 20:34:37,823][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 20:34:44,395][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_2400_state/model.safetensors +[2026-06-17 20:34:59,765][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_2400_state/optimizer.bin +[2026-06-17 20:34:59,767][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_2400_state/sampler.bin +[2026-06-17 20:34:59,767][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_2400_state/sampler_1.bin +[2026-06-17 20:34:59,770][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_2400_state/random_states_0.pkl +[2026-06-17 20:34:59,775][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2400 +[2026-06-17 20:34:59,775][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 20:34:59,786][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 20:35:03,671][starVLA.training.train_starvla][INFO] - Step 2401, Loss: {'train/loss': 0.42045024037361145, 'train/grad_norm_pre_clip': 12.277437210083008, 'timing/data': 0.0005290685221552849, 'timing/model': 1.1871752077713609, 'learning_rate/qwen_vl_interface': 7.6231268088990075e-06, 'learning_rate/action_model': 7.623126808899008e-05, 'epoch': 0.99}) +[2026-06-17 20:35:07,343][starVLA.training.train_starvla][INFO] - Step 2402, Loss: {'train/loss': 0.4783845543861389, 'train/grad_norm_pre_clip': 11.403200149536133, 'timing/data': 0.0005316575989127159, 'timing/model': 0.9594451365992427, 'learning_rate/qwen_vl_interface': 7.621253125846273e-06, 'learning_rate/action_model': 7.621253125846273e-05, 'epoch': 0.99}) +[2026-06-17 20:35:11,318][starVLA.training.train_starvla][INFO] - Step 2403, Loss: {'train/loss': 0.5545110702514648, 'train/grad_norm_pre_clip': 13.676319122314453, 'timing/data': 0.0005679568275809288, 'timing/model': 0.9532828917726874, 'learning_rate/qwen_vl_interface': 7.619378951230213e-06, 'learning_rate/action_model': 7.619378951230213e-05, 'epoch': 0.99}) +[2026-06-17 20:35:15,137][starVLA.training.train_starvla][INFO] - Step 2404, Loss: {'train/loss': 0.505955696105957, 'train/grad_norm_pre_clip': 14.01021957397461, 'timing/data': 0.0009570075199007988, 'timing/model': 0.9461565650999546, 'learning_rate/qwen_vl_interface': 7.617504285439346e-06, 'learning_rate/action_model': 7.617504285439346e-05, 'epoch': 0.99}) +[2026-06-17 20:35:19,033][starVLA.training.train_starvla][INFO] - Step 2405, Loss: {'train/loss': 0.46282029151916504, 'train/grad_norm_pre_clip': 12.267291069030762, 'timing/data': 0.0008067870512604713, 'timing/model': 0.9394875317811966, 'learning_rate/qwen_vl_interface': 7.615629128862295e-06, 'learning_rate/action_model': 7.615629128862295e-05, 'epoch': 0.99}) +[2026-06-17 20:35:22,900][starVLA.training.train_starvla][INFO] - Step 2406, Loss: {'train/loss': 0.40775448083877563, 'train/grad_norm_pre_clip': 9.73907470703125, 'timing/data': 0.0006871316581964493, 'timing/model': 0.9856816614046693, 'learning_rate/qwen_vl_interface': 7.61375348188778e-06, 'learning_rate/action_model': 7.61375348188778e-05, 'epoch': 0.99}) +[2026-06-17 20:35:26,719][starVLA.training.train_starvla][INFO] - Step 2407, Loss: {'train/loss': 0.6816818714141846, 'train/grad_norm_pre_clip': 14.423321723937988, 'timing/data': 0.0004697982221841812, 'timing/model': 0.9346525315195322, 'learning_rate/qwen_vl_interface': 7.611877344904624e-06, 'learning_rate/action_model': 7.611877344904624e-05, 'epoch': 0.99}) +[2026-06-17 20:35:30,996][starVLA.training.train_starvla][INFO] - Step 2408, Loss: {'train/loss': 0.38982218503952026, 'train/grad_norm_pre_clip': 16.222885131835938, 'timing/data': 0.0005335137248039246, 'timing/model': 1.32576088514179, 'learning_rate/qwen_vl_interface': 7.610000718301753e-06, 'learning_rate/action_model': 7.610000718301753e-05, 'epoch': 0.99}) +[2026-06-17 20:35:34,658][starVLA.training.train_starvla][INFO] - Step 2409, Loss: {'train/loss': 0.6299652457237244, 'train/grad_norm_pre_clip': 15.435879707336426, 'timing/data': 0.0004711644724011421, 'timing/model': 0.9409701563417912, 'learning_rate/qwen_vl_interface': 7.608123602468195e-06, 'learning_rate/action_model': 7.608123602468194e-05, 'epoch': 0.99}) +[2026-06-17 20:35:38,670][starVLA.training.train_starvla][INFO] - Step 2410, Loss: {'train/loss': 0.4757464826107025, 'train/grad_norm_pre_clip': 10.547123908996582, 'timing/data': 0.0005821734666824341, 'timing/model': 0.9478013049811125, 'learning_rate/qwen_vl_interface': 7.606245997793074e-06, 'learning_rate/action_model': 7.606245997793075e-05, 'epoch': 0.99}) +[2026-06-17 20:35:42,645][starVLA.training.train_starvla][INFO] - Step 2411, Loss: {'train/loss': 0.5753324627876282, 'train/grad_norm_pre_clip': 17.801931381225586, 'timing/data': 0.0005321484059095383, 'timing/model': 0.9499700292944908, 'learning_rate/qwen_vl_interface': 7.604367904665625e-06, 'learning_rate/action_model': 7.604367904665624e-05, 'epoch': 0.99}) +[2026-06-17 20:35:46,610][starVLA.training.train_starvla][INFO] - Step 2412, Loss: {'train/loss': 0.6302406787872314, 'train/grad_norm_pre_clip': 13.331232070922852, 'timing/data': 0.0008938666433095932, 'timing/model': 0.9632162051275373, 'learning_rate/qwen_vl_interface': 7.602489323475174e-06, 'learning_rate/action_model': 7.602489323475174e-05, 'epoch': 0.99}) +[2026-06-17 20:35:50,663][starVLA.training.train_starvla][INFO] - Step 2413, Loss: {'train/loss': 0.35794830322265625, 'train/grad_norm_pre_clip': 10.096719741821289, 'timing/data': 0.0014754077419638634, 'timing/model': 0.9949286105111241, 'learning_rate/qwen_vl_interface': 7.600610254611155e-06, 'learning_rate/action_model': 7.600610254611155e-05, 'epoch': 0.99}) +[2026-06-17 20:35:55,294][starVLA.training.train_starvla][INFO] - Step 2414, Loss: {'train/loss': 0.46063339710235596, 'train/grad_norm_pre_clip': 11.707666397094727, 'timing/data': 0.00045730918645858765, 'timing/model': 1.162808750756085, 'learning_rate/qwen_vl_interface': 7.5987306984631025e-06, 'learning_rate/action_model': 7.598730698463102e-05, 'epoch': 0.99}) +[2026-06-17 20:35:59,106][starVLA.training.train_starvla][INFO] - Step 2415, Loss: {'train/loss': 0.44073134660720825, 'train/grad_norm_pre_clip': 11.011218070983887, 'timing/data': 0.0008072126656770706, 'timing/model': 1.1217969795688987, 'learning_rate/qwen_vl_interface': 7.596850655420647e-06, 'learning_rate/action_model': 7.596850655420646e-05, 'epoch': 0.99}) +[2026-06-17 20:36:02,797][starVLA.training.train_starvla][INFO] - Step 2416, Loss: {'train/loss': 0.5217739343643188, 'train/grad_norm_pre_clip': 12.609023094177246, 'timing/data': 0.0009139245375990868, 'timing/model': 1.004493254236877, 'learning_rate/qwen_vl_interface': 7.5949701258735265e-06, 'learning_rate/action_model': 7.594970125873526e-05, 'epoch': 0.99}) +[2026-06-17 20:36:06,755][starVLA.training.train_starvla][INFO] - Step 2417, Loss: {'train/loss': 0.45060068368911743, 'train/grad_norm_pre_clip': 11.135433197021484, 'timing/data': 0.0008179973810911179, 'timing/model': 0.9526192704215646, 'learning_rate/qwen_vl_interface': 7.5930891102115735e-06, 'learning_rate/action_model': 7.593089110211573e-05, 'epoch': 0.99}) +[2026-06-17 20:36:10,741][starVLA.training.train_starvla][INFO] - Step 2418, Loss: {'train/loss': 0.43435314297676086, 'train/grad_norm_pre_clip': 11.647944450378418, 'timing/data': 0.0005336347967386246, 'timing/model': 0.956658280454576, 'learning_rate/qwen_vl_interface': 7.59120760882473e-06, 'learning_rate/action_model': 7.59120760882473e-05, 'epoch': 1.0}) +[2026-06-17 20:36:14,672][starVLA.training.train_starvla][INFO] - Step 2419, Loss: {'train/loss': 0.27758973836898804, 'train/grad_norm_pre_clip': 9.335912704467773, 'timing/data': 0.000528654083609581, 'timing/model': 0.9294842937961221, 'learning_rate/qwen_vl_interface': 7.5893256221030275e-06, 'learning_rate/action_model': 7.589325622103027e-05, 'epoch': 1.0}) +[2026-06-17 20:36:18,585][starVLA.training.train_starvla][INFO] - Step 2420, Loss: {'train/loss': 0.32829129695892334, 'train/grad_norm_pre_clip': 9.220372200012207, 'timing/data': 0.000500422902405262, 'timing/model': 1.012075143866241, 'learning_rate/qwen_vl_interface': 7.587443150436608e-06, 'learning_rate/action_model': 7.587443150436608e-05, 'epoch': 1.0}) +[2026-06-17 20:36:22,392][starVLA.training.train_starvla][INFO] - Step 2421, Loss: {'train/loss': 0.45949453115463257, 'train/grad_norm_pre_clip': 13.476531982421875, 'timing/data': 0.0004825415089726448, 'timing/model': 1.1202092273160815, 'learning_rate/qwen_vl_interface': 7.585560194215706e-06, 'learning_rate/action_model': 7.585560194215706e-05, 'epoch': 1.0}) +[2026-06-17 20:36:26,189][starVLA.training.train_starvla][INFO] - Step 2422, Loss: {'train/loss': 0.49631696939468384, 'train/grad_norm_pre_clip': 14.442399024963379, 'timing/data': 0.0005172984674572945, 'timing/model': 1.1142118507996202, 'learning_rate/qwen_vl_interface': 7.583676753830665e-06, 'learning_rate/action_model': 7.583676753830664e-05, 'epoch': 1.0}) +[2026-06-17 20:36:29,824][starVLA.training.train_starvla][INFO] - Step 2423, Loss: {'train/loss': 0.46572110056877136, 'train/grad_norm_pre_clip': 12.996397018432617, 'timing/data': 0.0007683625444769859, 'timing/model': 0.9429222997277975, 'learning_rate/qwen_vl_interface': 7.58179282967192e-06, 'learning_rate/action_model': 7.58179282967192e-05, 'epoch': 1.0}) +[2026-06-17 20:36:33,690][starVLA.training.train_starvla][INFO] - Step 2424, Loss: {'train/loss': 0.7884418964385986, 'train/grad_norm_pre_clip': 24.97675132751465, 'timing/data': 0.0008978052064776421, 'timing/model': 0.9447949072346091, 'learning_rate/qwen_vl_interface': 7.579908422130013e-06, 'learning_rate/action_model': 7.579908422130014e-05, 'epoch': 1.0}) +[2026-06-17 20:36:37,601][starVLA.training.train_starvla][INFO] - Step 2425, Loss: {'train/loss': 0.26696139574050903, 'train/grad_norm_pre_clip': 10.969165802001953, 'timing/data': 0.0005909772589802742, 'timing/model': 0.9510705284774303, 'learning_rate/qwen_vl_interface': 7.578023531595583e-06, 'learning_rate/action_model': 7.578023531595583e-05, 'epoch': 1.0}) +[2026-06-17 20:36:41,416][starVLA.training.train_starvla][INFO] - Step 2426, Loss: {'train/loss': 0.3650651276111603, 'train/grad_norm_pre_clip': 11.554516792297363, 'timing/data': 0.0009533204138278961, 'timing/model': 0.9398840442299843, 'learning_rate/qwen_vl_interface': 7.576138158459369e-06, 'learning_rate/action_model': 7.576138158459368e-05, 'epoch': 1.0}) +[2026-06-17 20:36:45,297][starVLA.training.train_starvla][INFO] - Step 2427, Loss: {'train/loss': 0.3754531443119049, 'train/grad_norm_pre_clip': 10.631322860717773, 'timing/data': 0.0004885643720626831, 'timing/model': 0.9581457749009132, 'learning_rate/qwen_vl_interface': 7.574252303112213e-06, 'learning_rate/action_model': 7.574252303112212e-05, 'epoch': 1.0}) +[2026-06-17 20:36:50,604][starVLA.training.train_starvla][INFO] - Step 2428, Loss: {'train/loss': 0.5749176740646362, 'train/grad_norm_pre_clip': 15.266728401184082, 'timing/data': 0.001442071981728077, 'timing/model': 1.6967777200043201, 'learning_rate/qwen_vl_interface': 7.572365965945053e-06, 'learning_rate/action_model': 7.572365965945052e-05, 'epoch': 1.0}) +[2026-06-17 20:36:54,347][starVLA.training.train_starvla][INFO] - Step 2429, Loss: {'train/loss': 0.38348978757858276, 'train/grad_norm_pre_clip': 13.06849479675293, 'timing/data': 0.0009931791573762894, 'timing/model': 0.9737059492617846, 'learning_rate/qwen_vl_interface': 7.570479147348929e-06, 'learning_rate/action_model': 7.570479147348929e-05, 'epoch': 1.0}) +[2026-06-17 20:36:58,422][starVLA.training.train_starvla][INFO] - Step 2430, Loss: {'train/loss': 0.4758540391921997, 'train/grad_norm_pre_clip': 15.065321922302246, 'timing/data': 0.011691253632307053, 'timing/model': 0.9515225924551487, 'learning_rate/qwen_vl_interface': 7.568591847714983e-06, 'learning_rate/action_model': 7.568591847714982e-05, 'epoch': 1.0}) +[2026-06-17 20:37:02,302][starVLA.training.train_starvla][INFO] - Step 2431, Loss: {'train/loss': 0.4703812897205353, 'train/grad_norm_pre_clip': 12.87095832824707, 'timing/data': 0.018310775980353355, 'timing/model': 0.9511569524183869, 'learning_rate/qwen_vl_interface': 7.5667040674344495e-06, 'learning_rate/action_model': 7.566704067434449e-05, 'epoch': 1.0}) +[2026-06-17 20:37:06,232][starVLA.training.train_starvla][INFO] - Step 2432, Loss: {'train/loss': 0.34259381890296936, 'train/grad_norm_pre_clip': 22.062379837036133, 'timing/data': 0.0180481169372797, 'timing/model': 1.0085316207259893, 'learning_rate/qwen_vl_interface': 7.564815806898669e-06, 'learning_rate/action_model': 7.564815806898668e-05, 'epoch': 1.0}) +[2026-06-17 20:37:10,137][starVLA.training.train_starvla][INFO] - Step 2433, Loss: {'train/loss': 0.38687464594841003, 'train/grad_norm_pre_clip': 11.932975769042969, 'timing/data': 0.01799831259995699, 'timing/model': 0.9353136634454131, 'learning_rate/qwen_vl_interface': 7.562927066499084e-06, 'learning_rate/action_model': 7.562927066499083e-05, 'epoch': 1.0}) +[2026-06-17 20:37:13,974][starVLA.training.train_starvla][INFO] - Step 2434, Loss: {'train/loss': 0.5071812272071838, 'train/grad_norm_pre_clip': 14.153158187866211, 'timing/data': 0.02222986426204443, 'timing/model': 1.1228474602103233, 'learning_rate/qwen_vl_interface': 7.5610378466272274e-06, 'learning_rate/action_model': 7.561037846627227e-05, 'epoch': 1.0}) +[2026-06-17 20:37:17,796][starVLA.training.train_starvla][INFO] - Step 2435, Loss: {'train/loss': 0.8692299723625183, 'train/grad_norm_pre_clip': 19.361045837402344, 'timing/data': 0.021092606708407402, 'timing/model': 1.089534517377615, 'learning_rate/qwen_vl_interface': 7.5591481476747384e-06, 'learning_rate/action_model': 7.559148147674739e-05, 'epoch': 1.0}) +[2026-06-17 20:37:21,448][starVLA.training.train_starvla][INFO] - Step 2436, Loss: {'train/loss': 0.7041240930557251, 'train/grad_norm_pre_clip': 16.824003219604492, 'timing/data': 0.016182322055101395, 'timing/model': 0.9430148033425212, 'learning_rate/qwen_vl_interface': 7.557257970033355e-06, 'learning_rate/action_model': 7.557257970033354e-05, 'epoch': 1.0}) +[2026-06-17 20:37:25,738][starVLA.training.train_starvla][INFO] - Step 2437, Loss: {'train/loss': 0.46656328439712524, 'train/grad_norm_pre_clip': 14.945863723754883, 'timing/data': 0.021747556515038013, 'timing/model': 0.9374008625745773, 'learning_rate/qwen_vl_interface': 7.55536731409491e-06, 'learning_rate/action_model': 7.55536731409491e-05, 'epoch': 1.0}) +[2026-06-17 20:37:29,648][starVLA.training.train_starvla][INFO] - Step 2438, Loss: {'train/loss': 0.5215704441070557, 'train/grad_norm_pre_clip': 13.879779815673828, 'timing/data': 0.02244576159864664, 'timing/model': 0.9653330110013485, 'learning_rate/qwen_vl_interface': 7.553476180251341e-06, 'learning_rate/action_model': 7.55347618025134e-05, 'epoch': 1.0}) +[2026-06-17 20:37:33,500][starVLA.training.train_starvla][INFO] - Step 2439, Loss: {'train/loss': 0.3853391706943512, 'train/grad_norm_pre_clip': 11.614608764648438, 'timing/data': 0.018118239007890224, 'timing/model': 0.9381429897621274, 'learning_rate/qwen_vl_interface': 7.551584568894681e-06, 'learning_rate/action_model': 7.551584568894681e-05, 'epoch': 1.0}) +[2026-06-17 20:37:37,337][starVLA.training.train_starvla][INFO] - Step 2440, Loss: {'train/loss': 0.5201866030693054, 'train/grad_norm_pre_clip': 12.638931274414062, 'timing/data': 0.027329490520060062, 'timing/model': 0.932363829575479, 'learning_rate/qwen_vl_interface': 7.5496924804170635e-06, 'learning_rate/action_model': 7.549692480417063e-05, 'epoch': 1.0}) +[2026-06-17 20:37:41,208][starVLA.training.train_starvla][INFO] - Step 2441, Loss: {'train/loss': 0.37514933943748474, 'train/grad_norm_pre_clip': 14.52087688446045, 'timing/data': 0.022702375426888466, 'timing/model': 1.1753481477499008, 'learning_rate/qwen_vl_interface': 7.5477999152107195e-06, 'learning_rate/action_model': 7.54779991521072e-05, 'epoch': 1.0}) +[2026-06-17 20:37:46,298][starVLA.training.train_starvla][INFO] - Step 2442, Loss: {'train/loss': 0.5830870866775513, 'train/grad_norm_pre_clip': 16.76158905029297, 'timing/data': 0.01800096035003662, 'timing/model': 0.9449091088026762, 'learning_rate/qwen_vl_interface': 7.545906873667982e-06, 'learning_rate/action_model': 7.545906873667981e-05, 'epoch': 1.01}) +[2026-06-17 20:37:52,148][starVLA.training.train_starvla][INFO] - Step 2443, Loss: {'train/loss': 0.47647127509117126, 'train/grad_norm_pre_clip': 13.162125587463379, 'timing/data': 0.024749877862632275, 'timing/model': 0.9466921044513583, 'learning_rate/qwen_vl_interface': 7.544013356181277e-06, 'learning_rate/action_model': 7.544013356181277e-05, 'epoch': 1.01}) +[2026-06-17 20:37:55,941][starVLA.training.train_starvla][INFO] - Step 2444, Loss: {'train/loss': 0.5739311575889587, 'train/grad_norm_pre_clip': 21.72478485107422, 'timing/data': 0.019041680730879307, 'timing/model': 0.9400636870414019, 'learning_rate/qwen_vl_interface': 7.542119363143135e-06, 'learning_rate/action_model': 7.542119363143136e-05, 'epoch': 1.01}) +[2026-06-17 20:38:00,446][starVLA.training.train_starvla][INFO] - Step 2445, Loss: {'train/loss': 0.425416499376297, 'train/grad_norm_pre_clip': 12.01113510131836, 'timing/data': 0.020206770859658718, 'timing/model': 1.0193516472354531, 'learning_rate/qwen_vl_interface': 7.540224894946184e-06, 'learning_rate/action_model': 7.540224894946184e-05, 'epoch': 1.01}) +[2026-06-17 20:38:04,356][starVLA.training.train_starvla][INFO] - Step 2446, Loss: {'train/loss': 0.6279128789901733, 'train/grad_norm_pre_clip': 16.672727584838867, 'timing/data': 0.029719890095293522, 'timing/model': 0.9949601497501135, 'learning_rate/qwen_vl_interface': 7.538329951983148e-06, 'learning_rate/action_model': 7.538329951983147e-05, 'epoch': 1.01}) +[2026-06-17 20:38:08,291][starVLA.training.train_starvla][INFO] - Step 2447, Loss: {'train/loss': 0.4908016622066498, 'train/grad_norm_pre_clip': 12.608091354370117, 'timing/data': 0.017507675103843212, 'timing/model': 1.1761742997914553, 'learning_rate/qwen_vl_interface': 7.536434534646851e-06, 'learning_rate/action_model': 7.53643453464685e-05, 'epoch': 1.01}) +[2026-06-17 20:38:12,218][starVLA.training.train_starvla][INFO] - Step 2448, Loss: {'train/loss': 0.3988822102546692, 'train/grad_norm_pre_clip': 14.98036003112793, 'timing/data': 0.021460507065057755, 'timing/model': 1.2004033140838146, 'learning_rate/qwen_vl_interface': 7.534538643330213e-06, 'learning_rate/action_model': 7.534538643330214e-05, 'epoch': 1.01}) +[2026-06-17 20:38:15,936][starVLA.training.train_starvla][INFO] - Step 2449, Loss: {'train/loss': 0.32081443071365356, 'train/grad_norm_pre_clip': 12.52440357208252, 'timing/data': 0.018028215505182743, 'timing/model': 0.96950245462358, 'learning_rate/qwen_vl_interface': 7.5326422784262566e-06, 'learning_rate/action_model': 7.532642278426256e-05, 'epoch': 1.01}) +[2026-06-17 20:38:19,798][starVLA.training.train_starvla][INFO] - Step 2450, Loss: {'train/loss': 0.5504708290100098, 'train/grad_norm_pre_clip': 13.926828384399414, 'timing/data': 0.019273496232926846, 'timing/model': 0.9562204293906689, 'learning_rate/qwen_vl_interface': 7.5307454403281e-06, 'learning_rate/action_model': 7.5307454403281e-05, 'epoch': 1.01}) +[2026-06-17 20:38:25,017][starVLA.training.train_starvla][INFO] - Step 2451, Loss: {'train/loss': 0.19778260588645935, 'train/grad_norm_pre_clip': 7.0920586585998535, 'timing/data': 0.022856125608086586, 'timing/model': 0.9382371073588729, 'learning_rate/qwen_vl_interface': 7.528848129428958e-06, 'learning_rate/action_model': 7.528848129428958e-05, 'epoch': 1.01}) +[2026-06-17 20:38:28,841][starVLA.training.train_starvla][INFO] - Step 2452, Loss: {'train/loss': 0.5529100894927979, 'train/grad_norm_pre_clip': 12.104337692260742, 'timing/data': 0.018525149673223495, 'timing/model': 0.9275646321475506, 'learning_rate/qwen_vl_interface': 7.526950346122149e-06, 'learning_rate/action_model': 7.526950346122148e-05, 'epoch': 1.01}) +[2026-06-17 20:38:32,771][starVLA.training.train_starvla][INFO] - Step 2453, Loss: {'train/loss': 0.1809854656457901, 'train/grad_norm_pre_clip': 6.993059158325195, 'timing/data': 0.012711400166153908, 'timing/model': 1.0242416514083743, 'learning_rate/qwen_vl_interface': 7.525052090801083e-06, 'learning_rate/action_model': 7.525052090801083e-05, 'epoch': 1.01}) +[2026-06-17 20:38:36,656][starVLA.training.train_starvla][INFO] - Step 2454, Loss: {'train/loss': 0.39697733521461487, 'train/grad_norm_pre_clip': 13.333340644836426, 'timing/data': 0.009950145147740841, 'timing/model': 1.1207926580682397, 'learning_rate/qwen_vl_interface': 7.523153363859269e-06, 'learning_rate/action_model': 7.523153363859269e-05, 'epoch': 1.01}) +[2026-06-17 20:38:40,688][starVLA.training.train_starvla][INFO] - Step 2455, Loss: {'train/loss': 0.3513695001602173, 'train/grad_norm_pre_clip': 11.768902778625488, 'timing/data': 0.0083913030102849, 'timing/model': 1.2137755835428834, 'learning_rate/qwen_vl_interface': 7.521254165690317e-06, 'learning_rate/action_model': 7.521254165690317e-05, 'epoch': 1.01}) +[2026-06-17 20:38:44,371][starVLA.training.train_starvla][INFO] - Step 2456, Loss: {'train/loss': 0.615500807762146, 'train/grad_norm_pre_clip': 14.448722839355469, 'timing/data': 0.009304317645728588, 'timing/model': 0.9362236643210053, 'learning_rate/qwen_vl_interface': 7.519354496687933e-06, 'learning_rate/action_model': 7.519354496687933e-05, 'epoch': 1.01}) +[2026-06-17 20:38:49,168][starVLA.training.train_starvla][INFO] - Step 2457, Loss: {'train/loss': 0.4777984917163849, 'train/grad_norm_pre_clip': 14.414661407470703, 'timing/data': 0.009202255867421627, 'timing/model': 0.9277549898251891, 'learning_rate/qwen_vl_interface': 7.517454357245918e-06, 'learning_rate/action_model': 7.517454357245917e-05, 'epoch': 1.01}) +[2026-06-17 20:38:52,977][starVLA.training.train_starvla][INFO] - Step 2458, Loss: {'train/loss': 0.5348309874534607, 'train/grad_norm_pre_clip': 16.289438247680664, 'timing/data': 0.00965940859168768, 'timing/model': 0.9336984781548381, 'learning_rate/qwen_vl_interface': 7.515553747758176e-06, 'learning_rate/action_model': 7.515553747758176e-05, 'epoch': 1.01}) +[2026-06-17 20:38:56,843][starVLA.training.train_starvla][INFO] - Step 2459, Loss: {'train/loss': 0.5461174845695496, 'train/grad_norm_pre_clip': 13.773919105529785, 'timing/data': 0.012520773336291313, 'timing/model': 0.9489734619855881, 'learning_rate/qwen_vl_interface': 7.513652668618705e-06, 'learning_rate/action_model': 7.513652668618705e-05, 'epoch': 1.01}) +[2026-06-17 20:39:00,640][starVLA.training.train_starvla][INFO] - Step 2460, Loss: {'train/loss': 0.565762460231781, 'train/grad_norm_pre_clip': 17.209943771362305, 'timing/data': 0.008711094036698341, 'timing/model': 0.9281617403030396, 'learning_rate/qwen_vl_interface': 7.5117511202215994e-06, 'learning_rate/action_model': 7.5117511202216e-05, 'epoch': 1.01}) +[2026-06-17 20:39:04,620][starVLA.training.train_starvla][INFO] - Step 2461, Loss: {'train/loss': 0.44704440236091614, 'train/grad_norm_pre_clip': 15.295181274414062, 'timing/data': 0.007772447541356087, 'timing/model': 1.115089986473322, 'learning_rate/qwen_vl_interface': 7.509849102961053e-06, 'learning_rate/action_model': 7.509849102961053e-05, 'epoch': 1.01}) +[2026-06-17 20:39:08,330][starVLA.training.train_starvla][INFO] - Step 2462, Loss: {'train/loss': 0.5537676811218262, 'train/grad_norm_pre_clip': 16.94601058959961, 'timing/data': 0.00885180477052927, 'timing/model': 0.953683951869607, 'learning_rate/qwen_vl_interface': 7.507946617231357e-06, 'learning_rate/action_model': 7.507946617231356e-05, 'epoch': 1.01}) +[2026-06-17 20:39:12,159][starVLA.training.train_starvla][INFO] - Step 2463, Loss: {'train/loss': 0.42032304406166077, 'train/grad_norm_pre_clip': 13.161824226379395, 'timing/data': 0.010126344859600067, 'timing/model': 0.9371527042239904, 'learning_rate/qwen_vl_interface': 7.5060436634268965e-06, 'learning_rate/action_model': 7.506043663426896e-05, 'epoch': 1.01}) +[2026-06-17 20:39:16,047][starVLA.training.train_starvla][INFO] - Step 2464, Loss: {'train/loss': 0.5718916058540344, 'train/grad_norm_pre_clip': 15.964425086975098, 'timing/data': 0.010415689088404179, 'timing/model': 0.9446314927190542, 'learning_rate/qwen_vl_interface': 7.504140241942157e-06, 'learning_rate/action_model': 7.504140241942157e-05, 'epoch': 1.01}) +[2026-06-17 20:39:19,893][starVLA.training.train_starvla][INFO] - Step 2465, Loss: {'train/loss': 0.5416136384010315, 'train/grad_norm_pre_clip': 15.110086441040039, 'timing/data': 0.01010961551219225, 'timing/model': 0.9479342810809612, 'learning_rate/qwen_vl_interface': 7.502236353171718e-06, 'learning_rate/action_model': 7.502236353171718e-05, 'epoch': 1.01}) +[2026-06-17 20:39:23,791][starVLA.training.train_starvla][INFO] - Step 2466, Loss: {'train/loss': 0.35902470350265503, 'train/grad_norm_pre_clip': 12.319122314453125, 'timing/data': 0.008452036418020725, 'timing/model': 1.00322456471622, 'learning_rate/qwen_vl_interface': 7.500331997510262e-06, 'learning_rate/action_model': 7.500331997510262e-05, 'epoch': 1.02}) +[2026-06-17 20:39:27,758][starVLA.training.train_starvla][INFO] - Step 2467, Loss: {'train/loss': 0.8268030881881714, 'train/grad_norm_pre_clip': 15.727032661437988, 'timing/data': 0.021381733007729053, 'timing/model': 0.9564291564747691, 'learning_rate/qwen_vl_interface': 7.498427175352561e-06, 'learning_rate/action_model': 7.49842717535256e-05, 'epoch': 1.02}) +[2026-06-17 20:39:33,005][starVLA.training.train_starvla][INFO] - Step 2468, Loss: {'train/loss': 0.4134899973869324, 'train/grad_norm_pre_clip': 24.261058807373047, 'timing/data': 0.01687481440603733, 'timing/model': 1.152411038056016, 'learning_rate/qwen_vl_interface': 7.496521887093487e-06, 'learning_rate/action_model': 7.496521887093486e-05, 'epoch': 1.02}) +[2026-06-17 20:39:37,818][starVLA.training.train_starvla][INFO] - Step 2469, Loss: {'train/loss': 0.3704516887664795, 'train/grad_norm_pre_clip': 10.028947830200195, 'timing/data': 0.012579902075231075, 'timing/model': 0.9490616200491786, 'learning_rate/qwen_vl_interface': 7.494616133128008e-06, 'learning_rate/action_model': 7.494616133128008e-05, 'epoch': 1.02}) +[2026-06-17 20:39:41,977][starVLA.training.train_starvla][INFO] - Step 2470, Loss: {'train/loss': 0.5829881429672241, 'train/grad_norm_pre_clip': 11.014812469482422, 'timing/data': 0.013205678202211857, 'timing/model': 0.9409457901492715, 'learning_rate/qwen_vl_interface': 7.4927099138511895e-06, 'learning_rate/action_model': 7.492709913851189e-05, 'epoch': 1.02}) +[2026-06-17 20:39:46,167][starVLA.training.train_starvla][INFO] - Step 2471, Loss: {'train/loss': 0.630339503288269, 'train/grad_norm_pre_clip': 14.228358268737793, 'timing/data': 0.009521132335066795, 'timing/model': 0.9411895181983709, 'learning_rate/qwen_vl_interface': 7.490803229658192e-06, 'learning_rate/action_model': 7.490803229658192e-05, 'epoch': 1.02}) +[2026-06-17 20:39:51,784][starVLA.training.train_starvla][INFO] - Step 2472, Loss: {'train/loss': 0.29365062713623047, 'train/grad_norm_pre_clip': 10.590254783630371, 'timing/data': 0.015154841355979443, 'timing/model': 0.9549561599269509, 'learning_rate/qwen_vl_interface': 7.488896080944274e-06, 'learning_rate/action_model': 7.488896080944273e-05, 'epoch': 1.02}) +[2026-06-17 20:39:55,627][starVLA.training.train_starvla][INFO] - Step 2473, Loss: {'train/loss': 0.5343669056892395, 'train/grad_norm_pre_clip': 12.851381301879883, 'timing/data': 0.026359226554632187, 'timing/model': 0.9365038825199008, 'learning_rate/qwen_vl_interface': 7.4869884681047896e-06, 'learning_rate/action_model': 7.486988468104789e-05, 'epoch': 1.02}) +[2026-06-17 20:39:59,790][starVLA.training.train_starvla][INFO] - Step 2474, Loss: {'train/loss': 0.11236913502216339, 'train/grad_norm_pre_clip': 5.145423412322998, 'timing/data': 0.01120890211313963, 'timing/model': 1.1339356880635023, 'learning_rate/qwen_vl_interface': 7.485080391535189e-06, 'learning_rate/action_model': 7.485080391535188e-05, 'epoch': 1.02}) +[2026-06-17 20:40:04,017][starVLA.training.train_starvla][INFO] - Step 2475, Loss: {'train/loss': 0.3662545680999756, 'train/grad_norm_pre_clip': 9.874027252197266, 'timing/data': 0.018148214556276798, 'timing/model': 1.1587966801598668, 'learning_rate/qwen_vl_interface': 7.483171851631018e-06, 'learning_rate/action_model': 7.483171851631017e-05, 'epoch': 1.02}) +[2026-06-17 20:40:07,853][starVLA.training.train_starvla][INFO] - Step 2476, Loss: {'train/loss': 0.42299914360046387, 'train/grad_norm_pre_clip': 13.671843528747559, 'timing/data': 0.017052489332854748, 'timing/model': 0.956922217272222, 'learning_rate/qwen_vl_interface': 7.481262848787919e-06, 'learning_rate/action_model': 7.481262848787919e-05, 'epoch': 1.02}) +[2026-06-17 20:40:11,741][starVLA.training.train_starvla][INFO] - Step 2477, Loss: {'train/loss': 0.586201548576355, 'train/grad_norm_pre_clip': 17.984262466430664, 'timing/data': 0.00821663811802864, 'timing/model': 0.9416912132874131, 'learning_rate/qwen_vl_interface': 7.479353383401631e-06, 'learning_rate/action_model': 7.47935338340163e-05, 'epoch': 1.02}) +[2026-06-17 20:40:16,219][starVLA.training.train_starvla][INFO] - Step 2478, Loss: {'train/loss': 0.4806314706802368, 'train/grad_norm_pre_clip': 14.980732917785645, 'timing/data': 0.010910469107329845, 'timing/model': 1.0424599330872297, 'learning_rate/qwen_vl_interface': 7.477443455867987e-06, 'learning_rate/action_model': 7.477443455867988e-05, 'epoch': 1.02}) +[2026-06-17 20:40:20,071][starVLA.training.train_starvla][INFO] - Step 2479, Loss: {'train/loss': 0.37631654739379883, 'train/grad_norm_pre_clip': 12.559113502502441, 'timing/data': 0.009942781180143356, 'timing/model': 0.941990896128118, 'learning_rate/qwen_vl_interface': 7.475533066582917e-06, 'learning_rate/action_model': 7.475533066582918e-05, 'epoch': 1.02}) +[2026-06-17 20:40:24,145][starVLA.training.train_starvla][INFO] - Step 2480, Loss: {'train/loss': 0.516782820224762, 'train/grad_norm_pre_clip': 11.699831008911133, 'timing/data': 0.0083985710516572, 'timing/model': 0.9896749909967184, 'learning_rate/qwen_vl_interface': 7.47362221594245e-06, 'learning_rate/action_model': 7.47362221594245e-05, 'epoch': 1.02}) +[2026-06-17 20:40:29,656][starVLA.training.train_starvla][INFO] - Step 2481, Loss: {'train/loss': 0.5646703839302063, 'train/grad_norm_pre_clip': 19.602354049682617, 'timing/data': 0.01696253102272749, 'timing/model': 1.188969835639, 'learning_rate/qwen_vl_interface': 7.4717109043427025e-06, 'learning_rate/action_model': 7.471710904342702e-05, 'epoch': 1.02}) +[2026-06-17 20:40:34,034][starVLA.training.train_starvla][INFO] - Step 2482, Loss: {'train/loss': 0.5946413278579712, 'train/grad_norm_pre_clip': 15.480978012084961, 'timing/data': 0.009604926221072674, 'timing/model': 1.6265230365097523, 'learning_rate/qwen_vl_interface': 7.469799132179894e-06, 'learning_rate/action_model': 7.469799132179895e-05, 'epoch': 1.02}) +[2026-06-17 20:40:38,758][starVLA.training.train_starvla][INFO] - Step 2483, Loss: {'train/loss': 0.6045425534248352, 'train/grad_norm_pre_clip': 21.866106033325195, 'timing/data': 0.010812086053192616, 'timing/model': 1.5013519553467631, 'learning_rate/qwen_vl_interface': 7.467886899850338e-06, 'learning_rate/action_model': 7.467886899850338e-05, 'epoch': 1.02}) +[2026-06-17 20:40:42,798][starVLA.training.train_starvla][INFO] - Step 2484, Loss: {'train/loss': 0.6181947588920593, 'train/grad_norm_pre_clip': 16.454355239868164, 'timing/data': 0.013932658359408379, 'timing/model': 1.0095771681517363, 'learning_rate/qwen_vl_interface': 7.465974207750442e-06, 'learning_rate/action_model': 7.465974207750442e-05, 'epoch': 1.02}) +[2026-06-17 20:40:47,942][starVLA.training.train_starvla][INFO] - Step 2485, Loss: {'train/loss': 0.572786808013916, 'train/grad_norm_pre_clip': 14.486217498779297, 'timing/data': 0.008090436458587646, 'timing/model': 1.050194394774735, 'learning_rate/qwen_vl_interface': 7.464061056276707e-06, 'learning_rate/action_model': 7.464061056276707e-05, 'epoch': 1.02}) +[2026-06-17 20:40:51,814][starVLA.training.train_starvla][INFO] - Step 2486, Loss: {'train/loss': 0.46557632088661194, 'train/grad_norm_pre_clip': 11.712185859680176, 'timing/data': 0.011620910838246346, 'timing/model': 0.9640993680804968, 'learning_rate/qwen_vl_interface': 7.462147445825733e-06, 'learning_rate/action_model': 7.462147445825733e-05, 'epoch': 1.02}) +[2026-06-17 20:40:55,656][starVLA.training.train_starvla][INFO] - Step 2487, Loss: {'train/loss': 0.3839716613292694, 'train/grad_norm_pre_clip': 14.156560897827148, 'timing/data': 0.018718679435551167, 'timing/model': 0.9295262563973665, 'learning_rate/qwen_vl_interface': 7.460233376794211e-06, 'learning_rate/action_model': 7.46023337679421e-05, 'epoch': 1.02}) +[2026-06-17 20:40:59,521][starVLA.training.train_starvla][INFO] - Step 2488, Loss: {'train/loss': 0.8472347259521484, 'train/grad_norm_pre_clip': 28.308677673339844, 'timing/data': 0.019197222776710987, 'timing/model': 1.1051270915195346, 'learning_rate/qwen_vl_interface': 7.458318849578934e-06, 'learning_rate/action_model': 7.458318849578933e-05, 'epoch': 1.02}) +[2026-06-17 20:41:03,148][starVLA.training.train_starvla][INFO] - Step 2489, Loss: {'train/loss': 0.4863339364528656, 'train/grad_norm_pre_clip': 18.050617218017578, 'timing/data': 0.01107085682451725, 'timing/model': 0.9479443142190576, 'learning_rate/qwen_vl_interface': 7.456403864576782e-06, 'learning_rate/action_model': 7.456403864576782e-05, 'epoch': 1.02}) +[2026-06-17 20:41:07,037][starVLA.training.train_starvla][INFO] - Step 2490, Loss: {'train/loss': 0.5994855165481567, 'train/grad_norm_pre_clip': 20.10628890991211, 'timing/data': 0.013313633389770985, 'timing/model': 0.9410341521725059, 'learning_rate/qwen_vl_interface': 7.4544884221847354e-06, 'learning_rate/action_model': 7.454488422184735e-05, 'epoch': 1.02}) +[2026-06-17 20:41:10,852][starVLA.training.train_starvla][INFO] - Step 2491, Loss: {'train/loss': 0.3640013039112091, 'train/grad_norm_pre_clip': 10.115294456481934, 'timing/data': 0.01334193255752325, 'timing/model': 0.9553296742960811, 'learning_rate/qwen_vl_interface': 7.452572522799867e-06, 'learning_rate/action_model': 7.452572522799866e-05, 'epoch': 1.03}) +[2026-06-17 20:41:14,754][starVLA.training.train_starvla][INFO] - Step 2492, Loss: {'train/loss': 0.28263628482818604, 'train/grad_norm_pre_clip': 9.96471881866455, 'timing/data': 0.010782650671899319, 'timing/model': 0.9574523409828544, 'learning_rate/qwen_vl_interface': 7.450656166819344e-06, 'learning_rate/action_model': 7.450656166819344e-05, 'epoch': 1.03}) +[2026-06-17 20:41:18,717][starVLA.training.train_starvla][INFO] - Step 2493, Loss: {'train/loss': 0.20709724724292755, 'train/grad_norm_pre_clip': 9.040621757507324, 'timing/data': 0.011136490851640701, 'timing/model': 0.9541427846997976, 'learning_rate/qwen_vl_interface': 7.4487393546404295e-06, 'learning_rate/action_model': 7.44873935464043e-05, 'epoch': 1.03}) +[2026-06-17 20:41:22,552][starVLA.training.train_starvla][INFO] - Step 2494, Loss: {'train/loss': 0.4298270642757416, 'train/grad_norm_pre_clip': 10.384063720703125, 'timing/data': 0.015468351542949677, 'timing/model': 1.1486413879320025, 'learning_rate/qwen_vl_interface': 7.446822086660481e-06, 'learning_rate/action_model': 7.446822086660481e-05, 'epoch': 1.03}) +[2026-06-17 20:41:26,402][starVLA.training.train_starvla][INFO] - Step 2495, Loss: {'train/loss': 0.6908589601516724, 'train/grad_norm_pre_clip': 21.822126388549805, 'timing/data': 0.013236590661108494, 'timing/model': 1.1277714036405087, 'learning_rate/qwen_vl_interface': 7.444904363276951e-06, 'learning_rate/action_model': 7.444904363276951e-05, 'epoch': 1.03}) +[2026-06-17 20:41:30,041][starVLA.training.train_starvla][INFO] - Step 2496, Loss: {'train/loss': 0.6771374940872192, 'train/grad_norm_pre_clip': 20.043834686279297, 'timing/data': 0.014308355748653412, 'timing/model': 0.9399155480787158, 'learning_rate/qwen_vl_interface': 7.442986184887383e-06, 'learning_rate/action_model': 7.442986184887383e-05, 'epoch': 1.03}) +[2026-06-17 20:41:33,887][starVLA.training.train_starvla][INFO] - Step 2497, Loss: {'train/loss': 0.28317421674728394, 'train/grad_norm_pre_clip': 12.06893539428711, 'timing/data': 0.010124586522579193, 'timing/model': 0.9494457449764013, 'learning_rate/qwen_vl_interface': 7.4410675518894205e-06, 'learning_rate/action_model': 7.44106755188942e-05, 'epoch': 1.03}) +[2026-06-17 20:41:37,800][starVLA.training.train_starvla][INFO] - Step 2498, Loss: {'train/loss': 0.7027490735054016, 'train/grad_norm_pre_clip': 19.66883659362793, 'timing/data': 0.01201015803962946, 'timing/model': 0.9491018615663052, 'learning_rate/qwen_vl_interface': 7.439148464680795e-06, 'learning_rate/action_model': 7.439148464680796e-05, 'epoch': 1.03}) +[2026-06-17 20:41:41,746][starVLA.training.train_starvla][INFO] - Step 2499, Loss: {'train/loss': 0.923427164554596, 'train/grad_norm_pre_clip': 21.621049880981445, 'timing/data': 0.010082803666591644, 'timing/model': 1.0054638842120767, 'learning_rate/qwen_vl_interface': 7.4372289236593395e-06, 'learning_rate/action_model': 7.43722892365934e-05, 'epoch': 1.03}) +[2026-06-17 20:43:58,930][starVLA.training.train_starvla][INFO] - Step 2500, Loss: {'train/loss': 0.553831160068512, 'train/grad_norm_pre_clip': 14.840588569641113, 'eval/loss': 0.5423181653022766, 'eval/latency_6/loss': 0.5423181653022766, 'eval/demon_fire_precision': 0.9390168014934661, 'eval/demon_fire_recall': 0.9754363283775048, 'eval/demon_fire_f1': 0.9568801521876983, 'eval/demon_fire_rate_model': 0.894268224819143, 'eval/demon_fire_rate_teacher': 0.860879243183083, 'eval/demon_left_precision': 0.8894736842105263, 'eval/demon_left_recall': 0.5929824561403508, 'eval/demon_left_f1': 0.711578947368421, 'eval/demon_left_rate_model': 0.10573177518085698, 'eval/demon_left_rate_teacher': 0.15859766277128548, 'eval/demon_right_precision': 0.8253358925143954, 'eval/demon_right_recall': 0.6605222734254992, 'eval/demon_right_f1': 0.7337883959044368, 'eval/demon_right_rate_model': 0.1449638286032276, 'eval/demon_right_rate_teacher': 0.18113522537562604, 'eval/demon_move_f1': 0.7226836716364289, 'eval/demon_cc_f1': 0.8720954680935847, 'eval/latency_6/demon_fire_precision': 0.9390168014934661, 'eval/latency_6/demon_fire_recall': 0.9754363283775048, 'eval/latency_6/demon_fire_f1': 0.9568801521876983, 'eval/latency_6/demon_fire_rate_model': 0.894268224819143, 'eval/latency_6/demon_fire_rate_teacher': 0.860879243183083, 'eval/latency_6/demon_left_precision': 0.8894736842105263, 'eval/latency_6/demon_left_recall': 0.5929824561403508, 'eval/latency_6/demon_left_f1': 0.711578947368421, 'eval/latency_6/demon_left_rate_model': 0.10573177518085698, 'eval/latency_6/demon_left_rate_teacher': 0.15859766277128548, 'eval/latency_6/demon_right_precision': 0.8253358925143954, 'eval/latency_6/demon_right_recall': 0.6605222734254992, 'eval/latency_6/demon_right_f1': 0.7337883959044368, 'eval/latency_6/demon_right_rate_model': 0.1449638286032276, 'eval/latency_6/demon_right_rate_teacher': 0.18113522537562604, 'eval/latency_6/demon_move_f1': 0.7226836716364289, 'eval/latency_6/demon_cc_f1': 0.8720954680935847, 'timing/data': 0.009792853146791458, 'timing/model': 0.9857711074873805, 'learning_rate/qwen_vl_interface': 7.435308929222973e-06, 'learning_rate/action_model': 7.435308929222972e-05, 'epoch': 1.03}) +[2026-06-17 20:43:58,931][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_2500_state +[2026-06-17 20:43:58,936][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 20:44:05,588][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_2500_state/model.safetensors +[2026-06-17 20:44:21,111][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_2500_state/optimizer.bin +[2026-06-17 20:44:21,111][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_2500_state/sampler.bin +[2026-06-17 20:44:21,112][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_2500_state/sampler_1.bin +[2026-06-17 20:44:21,113][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_2500_state/random_states_0.pkl +[2026-06-17 20:44:21,117][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2500 +[2026-06-17 20:44:21,117][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 20:44:21,128][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 20:44:24,964][starVLA.training.train_starvla][INFO] - Step 2501, Loss: {'train/loss': 0.5932740569114685, 'train/grad_norm_pre_clip': 15.218204498291016, 'timing/data': 0.008997006341814995, 'timing/model': 0.9296136051416397, 'learning_rate/qwen_vl_interface': 7.433388481769716e-06, 'learning_rate/action_model': 7.433388481769715e-05, 'epoch': 1.03}) +[2026-06-17 20:44:30,550][starVLA.training.train_starvla][INFO] - Step 2502, Loss: {'train/loss': 0.635003924369812, 'train/grad_norm_pre_clip': 18.238359451293945, 'timing/data': 0.021584002301096916, 'timing/model': 2.633261845447123, 'learning_rate/qwen_vl_interface': 7.431467581697675e-06, 'learning_rate/action_model': 7.431467581697674e-05, 'epoch': 1.03}) +[2026-06-17 20:44:34,495][starVLA.training.train_starvla][INFO] - Step 2503, Loss: {'train/loss': 0.7084900736808777, 'train/grad_norm_pre_clip': 17.1047420501709, 'timing/data': 0.01809835247695446, 'timing/model': 1.1456295093521476, 'learning_rate/qwen_vl_interface': 7.429546229405057e-06, 'learning_rate/action_model': 7.429546229405057e-05, 'epoch': 1.03}) +[2026-06-17 20:44:38,445][starVLA.training.train_starvla][INFO] - Step 2504, Loss: {'train/loss': 0.6343079805374146, 'train/grad_norm_pre_clip': 12.770241737365723, 'timing/data': 0.008445852436125278, 'timing/model': 1.117761041969061, 'learning_rate/qwen_vl_interface': 7.427624425290162e-06, 'learning_rate/action_model': 7.427624425290161e-05, 'epoch': 1.03}) +[2026-06-17 20:44:43,441][starVLA.training.train_starvla][INFO] - Step 2505, Loss: {'train/loss': 0.4977148473262787, 'train/grad_norm_pre_clip': 10.909541130065918, 'timing/data': 0.009089800529181957, 'timing/model': 0.9285418223589659, 'learning_rate/qwen_vl_interface': 7.425702169751379e-06, 'learning_rate/action_model': 7.425702169751378e-05, 'epoch': 1.03}) +[2026-06-17 20:44:47,249][starVLA.training.train_starvla][INFO] - Step 2506, Loss: {'train/loss': 0.7880975008010864, 'train/grad_norm_pre_clip': 54.125831604003906, 'timing/data': 0.009016787633299828, 'timing/model': 0.945461904630065, 'learning_rate/qwen_vl_interface': 7.4237794631871965e-06, 'learning_rate/action_model': 7.423779463187196e-05, 'epoch': 1.03}) +[2026-06-17 20:44:51,169][starVLA.training.train_starvla][INFO] - Step 2507, Loss: {'train/loss': 0.3312249481678009, 'train/grad_norm_pre_clip': 11.468962669372559, 'timing/data': 0.009049355052411556, 'timing/model': 0.939158221706748, 'learning_rate/qwen_vl_interface': 7.42185630599619e-06, 'learning_rate/action_model': 7.42185630599619e-05, 'epoch': 1.03}) +[2026-06-17 20:44:55,213][starVLA.training.train_starvla][INFO] - Step 2508, Loss: {'train/loss': 0.6281086206436157, 'train/grad_norm_pre_clip': 9.909133911132812, 'timing/data': 0.0157100111246109, 'timing/model': 0.9647191753610969, 'learning_rate/qwen_vl_interface': 7.419932698577036e-06, 'learning_rate/action_model': 7.419932698577035e-05, 'epoch': 1.03}) +[2026-06-17 20:44:59,060][starVLA.training.train_starvla][INFO] - Step 2509, Loss: {'train/loss': 0.4390602707862854, 'train/grad_norm_pre_clip': 11.809611320495605, 'timing/data': 0.007044924423098564, 'timing/model': 1.1623043036088347, 'learning_rate/qwen_vl_interface': 7.418008641328498e-06, 'learning_rate/action_model': 7.418008641328497e-05, 'epoch': 1.03}) +[2026-06-17 20:45:03,230][starVLA.training.train_starvla][INFO] - Step 2510, Loss: {'train/loss': 0.5983375906944275, 'train/grad_norm_pre_clip': 14.170106887817383, 'timing/data': 0.011526207439601421, 'timing/model': 1.25512186717242, 'learning_rate/qwen_vl_interface': 7.416084134649435e-06, 'learning_rate/action_model': 7.416084134649435e-05, 'epoch': 1.03}) +[2026-06-17 20:45:07,057][starVLA.training.train_starvla][INFO] - Step 2511, Loss: {'train/loss': 0.5090303421020508, 'train/grad_norm_pre_clip': 14.020772933959961, 'timing/data': 0.01051597110927105, 'timing/model': 0.9477521954104304, 'learning_rate/qwen_vl_interface': 7.414159178938801e-06, 'learning_rate/action_model': 7.414159178938801e-05, 'epoch': 1.03}) +[2026-06-17 20:45:12,607][starVLA.training.train_starvla][INFO] - Step 2512, Loss: {'train/loss': 0.47753697633743286, 'train/grad_norm_pre_clip': 12.315892219543457, 'timing/data': 0.012446231208741665, 'timing/model': 0.9528985088691115, 'learning_rate/qwen_vl_interface': 7.412233774595639e-06, 'learning_rate/action_model': 7.41223377459564e-05, 'epoch': 1.03}) +[2026-06-17 20:45:17,410][starVLA.training.train_starvla][INFO] - Step 2513, Loss: {'train/loss': 0.4070340692996979, 'train/grad_norm_pre_clip': 8.778766632080078, 'timing/data': 0.013861292041838169, 'timing/model': 0.9468441670760512, 'learning_rate/qwen_vl_interface': 7.410307922019091e-06, 'learning_rate/action_model': 7.41030792201909e-05, 'epoch': 1.03}) +[2026-06-17 20:45:21,417][starVLA.training.train_starvla][INFO] - Step 2514, Loss: {'train/loss': 0.3622388243675232, 'train/grad_norm_pre_clip': 9.636527061462402, 'timing/data': 0.008825422264635563, 'timing/model': 1.002193683758378, 'learning_rate/qwen_vl_interface': 7.408381621608387e-06, 'learning_rate/action_model': 7.408381621608387e-05, 'epoch': 1.03}) +[2026-06-17 20:45:25,245][starVLA.training.train_starvla][INFO] - Step 2515, Loss: {'train/loss': 0.5841068625450134, 'train/grad_norm_pre_clip': 20.321918487548828, 'timing/data': 0.008424314670264721, 'timing/model': 0.9420102573931217, 'learning_rate/qwen_vl_interface': 7.406454873762848e-06, 'learning_rate/action_model': 7.406454873762848e-05, 'epoch': 1.04}) +[2026-06-17 20:45:29,042][starVLA.training.train_starvla][INFO] - Step 2516, Loss: {'train/loss': 0.43302202224731445, 'train/grad_norm_pre_clip': 34.55661392211914, 'timing/data': 0.010082601569592953, 'timing/model': 0.9690829571336508, 'learning_rate/qwen_vl_interface': 7.404527678881894e-06, 'learning_rate/action_model': 7.404527678881894e-05, 'epoch': 1.04}) +[2026-06-17 20:45:32,960][starVLA.training.train_starvla][INFO] - Step 2517, Loss: {'train/loss': 0.32108068466186523, 'train/grad_norm_pre_clip': 12.13078784942627, 'timing/data': 0.008036074228584766, 'timing/model': 1.1217668745666742, 'learning_rate/qwen_vl_interface': 7.402600037365034e-06, 'learning_rate/action_model': 7.402600037365034e-05, 'epoch': 1.04}) +[2026-06-17 20:45:36,614][starVLA.training.train_starvla][INFO] - Step 2518, Loss: {'train/loss': 0.38864704966545105, 'train/grad_norm_pre_clip': 11.02453899383545, 'timing/data': 0.012435299344360828, 'timing/model': 0.9487253855913877, 'learning_rate/qwen_vl_interface': 7.400671949611869e-06, 'learning_rate/action_model': 7.40067194961187e-05, 'epoch': 1.04}) +[2026-06-17 20:45:40,437][starVLA.training.train_starvla][INFO] - Step 2519, Loss: {'train/loss': 0.3504473865032196, 'train/grad_norm_pre_clip': 9.76274299621582, 'timing/data': 0.007386018522083759, 'timing/model': 0.9414466600865126, 'learning_rate/qwen_vl_interface': 7.398743416022095e-06, 'learning_rate/action_model': 7.398743416022095e-05, 'epoch': 1.04}) +[2026-06-17 20:45:44,356][starVLA.training.train_starvla][INFO] - Step 2520, Loss: {'train/loss': 0.5765832662582397, 'train/grad_norm_pre_clip': 19.182138442993164, 'timing/data': 0.009823806583881378, 'timing/model': 1.0021171681582928, 'learning_rate/qwen_vl_interface': 7.3968144369954995e-06, 'learning_rate/action_model': 7.396814436995499e-05, 'epoch': 1.04}) +[2026-06-17 20:45:48,184][starVLA.training.train_starvla][INFO] - Step 2521, Loss: {'train/loss': 0.7070353031158447, 'train/grad_norm_pre_clip': 17.65516471862793, 'timing/data': 0.016969378106296062, 'timing/model': 0.9479146860539913, 'learning_rate/qwen_vl_interface': 7.39488501293196e-06, 'learning_rate/action_model': 7.39488501293196e-05, 'epoch': 1.04}) +[2026-06-17 20:45:52,021][starVLA.training.train_starvla][INFO] - Step 2522, Loss: {'train/loss': 0.22867026925086975, 'train/grad_norm_pre_clip': 7.942709922790527, 'timing/data': 0.00956751499325037, 'timing/model': 0.946779360063374, 'learning_rate/qwen_vl_interface': 7.392955144231452e-06, 'learning_rate/action_model': 7.392955144231452e-05, 'epoch': 1.04}) +[2026-06-17 20:45:55,855][starVLA.training.train_starvla][INFO] - Step 2523, Loss: {'train/loss': 0.3990218937397003, 'train/grad_norm_pre_clip': 16.947938919067383, 'timing/data': 0.014177321456372738, 'timing/model': 1.1139227375388145, 'learning_rate/qwen_vl_interface': 7.391024831294034e-06, 'learning_rate/action_model': 7.391024831294033e-05, 'epoch': 1.04}) +[2026-06-17 20:45:59,640][starVLA.training.train_starvla][INFO] - Step 2524, Loss: {'train/loss': 0.27262696623802185, 'train/grad_norm_pre_clip': 11.189854621887207, 'timing/data': 0.007275186479091644, 'timing/model': 1.0768038276582956, 'learning_rate/qwen_vl_interface': 7.389094074519866e-06, 'learning_rate/action_model': 7.389094074519866e-05, 'epoch': 1.04}) +[2026-06-17 20:46:03,300][starVLA.training.train_starvla][INFO] - Step 2525, Loss: {'train/loss': 0.42672014236450195, 'train/grad_norm_pre_clip': 11.314963340759277, 'timing/data': 0.009198473766446114, 'timing/model': 0.9456992968916893, 'learning_rate/qwen_vl_interface': 7.387162874309194e-06, 'learning_rate/action_model': 7.387162874309193e-05, 'epoch': 1.04}) +[2026-06-17 20:46:07,142][starVLA.training.train_starvla][INFO] - Step 2526, Loss: {'train/loss': 0.3500872850418091, 'train/grad_norm_pre_clip': 13.429582595825195, 'timing/data': 0.011395566165447235, 'timing/model': 0.9491358287632465, 'learning_rate/qwen_vl_interface': 7.385231231062359e-06, 'learning_rate/action_model': 7.385231231062358e-05, 'epoch': 1.04}) +[2026-06-17 20:46:11,037][starVLA.training.train_starvla][INFO] - Step 2527, Loss: {'train/loss': 0.4131104350090027, 'train/grad_norm_pre_clip': 13.948986053466797, 'timing/data': 0.011067490093410015, 'timing/model': 0.9554025596007705, 'learning_rate/qwen_vl_interface': 7.383299145179789e-06, 'learning_rate/action_model': 7.383299145179789e-05, 'epoch': 1.04}) +[2026-06-17 20:46:14,840][starVLA.training.train_starvla][INFO] - Step 2528, Loss: {'train/loss': 0.8488327264785767, 'train/grad_norm_pre_clip': 18.383337020874023, 'timing/data': 0.008949724026024342, 'timing/model': 0.9443031195551157, 'learning_rate/qwen_vl_interface': 7.381366617062014e-06, 'learning_rate/action_model': 7.381366617062014e-05, 'epoch': 1.04}) +[2026-06-17 20:46:18,674][starVLA.training.train_starvla][INFO] - Step 2529, Loss: {'train/loss': 0.7776392698287964, 'train/grad_norm_pre_clip': 16.193756103515625, 'timing/data': 0.008700940757989883, 'timing/model': 0.9528300501406193, 'learning_rate/qwen_vl_interface': 7.379433647109645e-06, 'learning_rate/action_model': 7.379433647109645e-05, 'epoch': 1.04}) +[2026-06-17 20:46:22,701][starVLA.training.train_starvla][INFO] - Step 2530, Loss: {'train/loss': 0.3964536786079407, 'train/grad_norm_pre_clip': 12.941964149475098, 'timing/data': 0.008701025508344173, 'timing/model': 1.1512738205492496, 'learning_rate/qwen_vl_interface': 7.37750023572339e-06, 'learning_rate/action_model': 7.37750023572339e-05, 'epoch': 1.04}) +[2026-06-17 20:46:26,362][starVLA.training.train_starvla][INFO] - Step 2531, Loss: {'train/loss': 0.32203128933906555, 'train/grad_norm_pre_clip': 18.018028259277344, 'timing/data': 0.011400114744901657, 'timing/model': 0.9476727787405252, 'learning_rate/qwen_vl_interface': 7.375566383304047e-06, 'learning_rate/action_model': 7.375566383304047e-05, 'epoch': 1.04}) +[2026-06-17 20:46:30,702][starVLA.training.train_starvla][INFO] - Step 2532, Loss: {'train/loss': 0.5871410965919495, 'train/grad_norm_pre_clip': 13.554015159606934, 'timing/data': 0.010691632516682148, 'timing/model': 0.9476017216220498, 'learning_rate/qwen_vl_interface': 7.373632090252505e-06, 'learning_rate/action_model': 7.373632090252505e-05, 'epoch': 1.04}) +[2026-06-17 20:46:34,497][starVLA.training.train_starvla][INFO] - Step 2533, Loss: {'train/loss': 0.30785006284713745, 'train/grad_norm_pre_clip': 9.399698257446289, 'timing/data': 0.008108554407954216, 'timing/model': 0.9469919558614492, 'learning_rate/qwen_vl_interface': 7.371697356969746e-06, 'learning_rate/action_model': 7.371697356969746e-05, 'epoch': 1.04}) +[2026-06-17 20:46:39,027][starVLA.training.train_starvla][INFO] - Step 2534, Loss: {'train/loss': 0.6008270978927612, 'train/grad_norm_pre_clip': 14.668776512145996, 'timing/data': 0.006474626250565052, 'timing/model': 0.9461191073060036, 'learning_rate/qwen_vl_interface': 7.36976218385684e-06, 'learning_rate/action_model': 7.36976218385684e-05, 'epoch': 1.04}) +[2026-06-17 20:46:42,926][starVLA.training.train_starvla][INFO] - Step 2535, Loss: {'train/loss': 0.20479416847229004, 'train/grad_norm_pre_clip': 7.39318323135376, 'timing/data': 0.009801529347896576, 'timing/model': 0.9525165865197778, 'learning_rate/qwen_vl_interface': 7.367826571314955e-06, 'learning_rate/action_model': 7.367826571314954e-05, 'epoch': 1.04}) +[2026-06-17 20:46:47,019][starVLA.training.train_starvla][INFO] - Step 2536, Loss: {'train/loss': 0.3621460795402527, 'train/grad_norm_pre_clip': 12.803564071655273, 'timing/data': 0.009537729434669018, 'timing/model': 1.3403867660090327, 'learning_rate/qwen_vl_interface': 7.36589051974534e-06, 'learning_rate/action_model': 7.36589051974534e-05, 'epoch': 1.04}) +[2026-06-17 20:46:50,903][starVLA.training.train_starvla][INFO] - Step 2537, Loss: {'train/loss': 0.5986480712890625, 'train/grad_norm_pre_clip': 12.72905158996582, 'timing/data': 0.013594034127891064, 'timing/model': 1.1839140774682164, 'learning_rate/qwen_vl_interface': 7.363954029549348e-06, 'learning_rate/action_model': 7.363954029549348e-05, 'epoch': 1.04}) +[2026-06-17 20:46:54,578][starVLA.training.train_starvla][INFO] - Step 2538, Loss: {'train/loss': 0.2721866965293884, 'train/grad_norm_pre_clip': 8.955100059509277, 'timing/data': 0.010455658659338951, 'timing/model': 0.9354111179709435, 'learning_rate/qwen_vl_interface': 7.362017101128406e-06, 'learning_rate/action_model': 7.362017101128406e-05, 'epoch': 1.04}) +[2026-06-17 20:46:58,432][starVLA.training.train_starvla][INFO] - Step 2539, Loss: {'train/loss': 0.46810290217399597, 'train/grad_norm_pre_clip': 14.64826488494873, 'timing/data': 0.007564106024801731, 'timing/model': 0.9364051800221205, 'learning_rate/qwen_vl_interface': 7.360079734884049e-06, 'learning_rate/action_model': 7.360079734884049e-05, 'epoch': 1.05}) +[2026-06-17 20:47:02,330][starVLA.training.train_starvla][INFO] - Step 2540, Loss: {'train/loss': 0.6819605827331543, 'train/grad_norm_pre_clip': 18.031166076660156, 'timing/data': 0.018666300922632217, 'timing/model': 0.9581584893167019, 'learning_rate/qwen_vl_interface': 7.358141931217893e-06, 'learning_rate/action_model': 7.358141931217892e-05, 'epoch': 1.05}) +[2026-06-17 20:47:06,294][starVLA.training.train_starvla][INFO] - Step 2541, Loss: {'train/loss': 0.5945618152618408, 'train/grad_norm_pre_clip': 22.44719696044922, 'timing/data': 0.009889516979455948, 'timing/model': 1.0194839788600802, 'learning_rate/qwen_vl_interface': 7.356203690531643e-06, 'learning_rate/action_model': 7.356203690531643e-05, 'epoch': 1.05}) +[2026-06-17 20:47:10,135][starVLA.training.train_starvla][INFO] - Step 2542, Loss: {'train/loss': 0.4103466272354126, 'train/grad_norm_pre_clip': 11.572434425354004, 'timing/data': 0.01019562128931284, 'timing/model': 0.9567777430638671, 'learning_rate/qwen_vl_interface': 7.354265013227104e-06, 'learning_rate/action_model': 7.354265013227104e-05, 'epoch': 1.05}) +[2026-06-17 20:47:14,087][starVLA.training.train_starvla][INFO] - Step 2543, Loss: {'train/loss': 0.41804203391075134, 'train/grad_norm_pre_clip': 13.92864990234375, 'timing/data': 0.010194080881774426, 'timing/model': 1.1478218408301473, 'learning_rate/qwen_vl_interface': 7.352325899706163e-06, 'learning_rate/action_model': 7.352325899706162e-05, 'epoch': 1.05}) +[2026-06-17 20:47:17,955][starVLA.training.train_starvla][INFO] - Step 2544, Loss: {'train/loss': 0.5126838088035583, 'train/grad_norm_pre_clip': 16.013092041015625, 'timing/data': 0.01045333780348301, 'timing/model': 1.1619454743340611, 'learning_rate/qwen_vl_interface': 7.350386350370802e-06, 'learning_rate/action_model': 7.350386350370801e-05, 'epoch': 1.05}) +[2026-06-17 20:47:21,723][starVLA.training.train_starvla][INFO] - Step 2545, Loss: {'train/loss': 0.5257980823516846, 'train/grad_norm_pre_clip': 14.003716468811035, 'timing/data': 0.012684712186455727, 'timing/model': 0.9685693522915244, 'learning_rate/qwen_vl_interface': 7.348446365623089e-06, 'learning_rate/action_model': 7.348446365623088e-05, 'epoch': 1.05}) +[2026-06-17 20:47:25,691][starVLA.training.train_starvla][INFO] - Step 2546, Loss: {'train/loss': 0.5753629207611084, 'train/grad_norm_pre_clip': 16.455148696899414, 'timing/data': 0.010103152133524418, 'timing/model': 0.970939370803535, 'learning_rate/qwen_vl_interface': 7.346505945865186e-06, 'learning_rate/action_model': 7.346505945865186e-05, 'epoch': 1.05}) +[2026-06-17 20:47:29,575][starVLA.training.train_starvla][INFO] - Step 2547, Loss: {'train/loss': 0.45281118154525757, 'train/grad_norm_pre_clip': 12.384078979492188, 'timing/data': 0.009277060627937317, 'timing/model': 0.9444568371400237, 'learning_rate/qwen_vl_interface': 7.344565091499345e-06, 'learning_rate/action_model': 7.344565091499344e-05, 'epoch': 1.05}) +[2026-06-17 20:47:33,422][starVLA.training.train_starvla][INFO] - Step 2548, Loss: {'train/loss': 0.3210464119911194, 'train/grad_norm_pre_clip': 11.249716758728027, 'timing/data': 0.010400745086371899, 'timing/model': 0.9596683187410235, 'learning_rate/qwen_vl_interface': 7.342623802927907e-06, 'learning_rate/action_model': 7.342623802927907e-05, 'epoch': 1.05}) +[2026-06-17 20:47:37,904][starVLA.training.train_starvla][INFO] - Step 2549, Loss: {'train/loss': 0.8212480545043945, 'train/grad_norm_pre_clip': 17.769628524780273, 'timing/data': 0.008118299767374992, 'timing/model': 1.3054392067715526, 'learning_rate/qwen_vl_interface': 7.340682080553303e-06, 'learning_rate/action_model': 7.340682080553303e-05, 'epoch': 1.05}) +[2026-06-17 20:47:41,934][starVLA.training.train_starvla][INFO] - Step 2550, Loss: {'train/loss': 0.530203104019165, 'train/grad_norm_pre_clip': 10.603949546813965, 'timing/data': 0.012302262708544731, 'timing/model': 1.266300174407661, 'learning_rate/qwen_vl_interface': 7.338739924778053e-06, 'learning_rate/action_model': 7.338739924778053e-05, 'epoch': 1.05}) +[2026-06-17 20:47:45,644][starVLA.training.train_starvla][INFO] - Step 2551, Loss: {'train/loss': 0.7309991717338562, 'train/grad_norm_pre_clip': 16.689977645874023, 'timing/data': 0.010060328990221024, 'timing/model': 0.973856745287776, 'learning_rate/qwen_vl_interface': 7.336797336004771e-06, 'learning_rate/action_model': 7.33679733600477e-05, 'epoch': 1.05}) +[2026-06-17 20:47:49,537][starVLA.training.train_starvla][INFO] - Step 2552, Loss: {'train/loss': 0.6603871583938599, 'train/grad_norm_pre_clip': 14.003189086914062, 'timing/data': 0.008663329295814037, 'timing/model': 0.9451169762760401, 'learning_rate/qwen_vl_interface': 7.334854314636153e-06, 'learning_rate/action_model': 7.334854314636152e-05, 'epoch': 1.05}) +[2026-06-17 20:47:53,392][starVLA.training.train_starvla][INFO] - Step 2553, Loss: {'train/loss': 0.5198932886123657, 'train/grad_norm_pre_clip': 13.28414535522461, 'timing/data': 0.008057739585638046, 'timing/model': 0.9497907068580389, 'learning_rate/qwen_vl_interface': 7.3329108610749955e-06, 'learning_rate/action_model': 7.332910861074995e-05, 'epoch': 1.05}) +[2026-06-17 20:47:57,302][starVLA.training.train_starvla][INFO] - Step 2554, Loss: {'train/loss': 0.7930928468704224, 'train/grad_norm_pre_clip': 17.176034927368164, 'timing/data': 0.008082428015768528, 'timing/model': 0.9421697296202183, 'learning_rate/qwen_vl_interface': 7.330966975724173e-06, 'learning_rate/action_model': 7.330966975724173e-05, 'epoch': 1.05}) +[2026-06-17 20:48:01,130][starVLA.training.train_starvla][INFO] - Step 2555, Loss: {'train/loss': 0.7341328859329224, 'train/grad_norm_pre_clip': 19.58401870727539, 'timing/data': 0.010521410033106804, 'timing/model': 0.9368234938010573, 'learning_rate/qwen_vl_interface': 7.329022658986657e-06, 'learning_rate/action_model': 7.329022658986658e-05, 'epoch': 1.05}) +[2026-06-17 20:48:04,940][starVLA.training.train_starvla][INFO] - Step 2556, Loss: {'train/loss': 0.3603970408439636, 'train/grad_norm_pre_clip': 11.601677894592285, 'timing/data': 0.007700907997786999, 'timing/model': 0.9438237119466066, 'learning_rate/qwen_vl_interface': 7.327077911265508e-06, 'learning_rate/action_model': 7.327077911265507e-05, 'epoch': 1.05}) +[2026-06-17 20:48:08,910][starVLA.training.train_starvla][INFO] - Step 2557, Loss: {'train/loss': 0.4506652057170868, 'train/grad_norm_pre_clip': 10.973113059997559, 'timing/data': 0.006667237728834152, 'timing/model': 1.2338882312178612, 'learning_rate/qwen_vl_interface': 7.325132732963872e-06, 'learning_rate/action_model': 7.325132732963872e-05, 'epoch': 1.05}) +[2026-06-17 20:48:12,552][starVLA.training.train_starvla][INFO] - Step 2558, Loss: {'train/loss': 0.654674768447876, 'train/grad_norm_pre_clip': 14.8909330368042, 'timing/data': 0.01026953849941492, 'timing/model': 0.9372580219060183, 'learning_rate/qwen_vl_interface': 7.323187124484988e-06, 'learning_rate/action_model': 7.323187124484988e-05, 'epoch': 1.05}) +[2026-06-17 20:48:16,484][starVLA.training.train_starvla][INFO] - Step 2559, Loss: {'train/loss': 0.4295414686203003, 'train/grad_norm_pre_clip': 12.157275199890137, 'timing/data': 0.010612639598548412, 'timing/model': 1.0183263197541237, 'learning_rate/qwen_vl_interface': 7.3212410862321815e-06, 'learning_rate/action_model': 7.321241086232182e-05, 'epoch': 1.05}) +[2026-06-17 20:48:20,381][starVLA.training.train_starvla][INFO] - Step 2560, Loss: {'train/loss': 0.4881281852722168, 'train/grad_norm_pre_clip': 13.106049537658691, 'timing/data': 0.011565818451344967, 'timing/model': 1.025563650764525, 'learning_rate/qwen_vl_interface': 7.319294618608869e-06, 'learning_rate/action_model': 7.319294618608868e-05, 'epoch': 1.05}) +[2026-06-17 20:48:26,108][starVLA.training.train_starvla][INFO] - Step 2561, Loss: {'train/loss': 0.29569461941719055, 'train/grad_norm_pre_clip': 6.769082069396973, 'timing/data': 0.007728710770606995, 'timing/model': 2.8075865749269724, 'learning_rate/qwen_vl_interface': 7.317347722018557e-06, 'learning_rate/action_model': 7.317347722018557e-05, 'epoch': 1.05}) +[2026-06-17 20:48:30,071][starVLA.training.train_starvla][INFO] - Step 2562, Loss: {'train/loss': 0.27664366364479065, 'train/grad_norm_pre_clip': 38.487361907958984, 'timing/data': 0.009275990538299084, 'timing/model': 0.9692128710448742, 'learning_rate/qwen_vl_interface': 7.3154003968648355e-06, 'learning_rate/action_model': 7.315400396864835e-05, 'epoch': 1.05}) +[2026-06-17 20:48:35,859][starVLA.training.train_starvla][INFO] - Step 2563, Loss: {'train/loss': 0.4868830442428589, 'train/grad_norm_pre_clip': 10.066972732543945, 'timing/data': 0.010562004521489143, 'timing/model': 1.1450359839946032, 'learning_rate/qwen_vl_interface': 7.313452643551392e-06, 'learning_rate/action_model': 7.313452643551392e-05, 'epoch': 1.06}) +[2026-06-17 20:48:39,708][starVLA.training.train_starvla][INFO] - Step 2564, Loss: {'train/loss': 0.329071968793869, 'train/grad_norm_pre_clip': 9.424761772155762, 'timing/data': 0.010694191791117191, 'timing/model': 1.110352797433734, 'learning_rate/qwen_vl_interface': 7.311504462481993e-06, 'learning_rate/action_model': 7.311504462481993e-05, 'epoch': 1.06}) +[2026-06-17 20:48:43,339][starVLA.training.train_starvla][INFO] - Step 2565, Loss: {'train/loss': 0.5917242765426636, 'train/grad_norm_pre_clip': 12.867417335510254, 'timing/data': 0.011742956936359406, 'timing/model': 0.9460218027234077, 'learning_rate/qwen_vl_interface': 7.3095558540605e-06, 'learning_rate/action_model': 7.309555854060499e-05, 'epoch': 1.06}) +[2026-06-17 20:48:47,156][starVLA.training.train_starvla][INFO] - Step 2566, Loss: {'train/loss': 0.48285573720932007, 'train/grad_norm_pre_clip': 12.950325965881348, 'timing/data': 0.011028754524886608, 'timing/model': 0.9381947368383408, 'learning_rate/qwen_vl_interface': 7.307606818690865e-06, 'learning_rate/action_model': 7.307606818690864e-05, 'epoch': 1.06}) +[2026-06-17 20:48:51,193][starVLA.training.train_starvla][INFO] - Step 2567, Loss: {'train/loss': 0.3400379419326782, 'train/grad_norm_pre_clip': 10.988938331604004, 'timing/data': 0.01152095291763544, 'timing/model': 1.0009086728096008, 'learning_rate/qwen_vl_interface': 7.30565735677712e-06, 'learning_rate/action_model': 7.305657356777119e-05, 'epoch': 1.06}) +[2026-06-17 20:48:55,124][starVLA.training.train_starvla][INFO] - Step 2568, Loss: {'train/loss': 0.2629651427268982, 'train/grad_norm_pre_clip': 10.823726654052734, 'timing/data': 0.013652071356773376, 'timing/model': 0.9628344792872667, 'learning_rate/qwen_vl_interface': 7.303707468723395e-06, 'learning_rate/action_model': 7.303707468723395e-05, 'epoch': 1.06}) +[2026-06-17 20:48:58,957][starVLA.training.train_starvla][INFO] - Step 2569, Loss: {'train/loss': 0.7129914164543152, 'train/grad_norm_pre_clip': 15.094877243041992, 'timing/data': 0.008891211822628975, 'timing/model': 0.963624975644052, 'learning_rate/qwen_vl_interface': 7.301757154933899e-06, 'learning_rate/action_model': 7.3017571549339e-05, 'epoch': 1.06}) +[2026-06-17 20:49:02,825][starVLA.training.train_starvla][INFO] - Step 2570, Loss: {'train/loss': 0.8094164133071899, 'train/grad_norm_pre_clip': 16.985130310058594, 'timing/data': 0.008835596032440662, 'timing/model': 1.1751063214614987, 'learning_rate/qwen_vl_interface': 7.29980641581294e-06, 'learning_rate/action_model': 7.299806415812939e-05, 'epoch': 1.06}) +[2026-06-17 20:49:06,638][starVLA.training.train_starvla][INFO] - Step 2571, Loss: {'train/loss': 0.3402150869369507, 'train/grad_norm_pre_clip': 9.433058738708496, 'timing/data': 0.0089999008923769, 'timing/model': 1.0888700690120459, 'learning_rate/qwen_vl_interface': 7.297855251764903e-06, 'learning_rate/action_model': 7.297855251764903e-05, 'epoch': 1.06}) +[2026-06-17 20:49:11,008][starVLA.training.train_starvla][INFO] - Step 2572, Loss: {'train/loss': 0.3565386235713959, 'train/grad_norm_pre_clip': 10.07320785522461, 'timing/data': 0.01338493824005127, 'timing/model': 1.3255768176168203, 'learning_rate/qwen_vl_interface': 7.295903663194269e-06, 'learning_rate/action_model': 7.29590366319427e-05, 'epoch': 1.06}) +[2026-06-17 20:49:15,385][starVLA.training.train_starvla][INFO] - Step 2573, Loss: {'train/loss': 0.476441890001297, 'train/grad_norm_pre_clip': 10.84335994720459, 'timing/data': 0.01317624282091856, 'timing/model': 0.9354383377358317, 'learning_rate/qwen_vl_interface': 7.293951650505606e-06, 'learning_rate/action_model': 7.293951650505607e-05, 'epoch': 1.06}) +[2026-06-17 20:49:19,598][starVLA.training.train_starvla][INFO] - Step 2574, Loss: {'train/loss': 0.49857810139656067, 'train/grad_norm_pre_clip': 12.41572380065918, 'timing/data': 0.010375553742051125, 'timing/model': 0.9997465964406729, 'learning_rate/qwen_vl_interface': 7.291999214103565e-06, 'learning_rate/action_model': 7.291999214103564e-05, 'epoch': 1.06}) +[2026-06-17 20:49:23,372][starVLA.training.train_starvla][INFO] - Step 2575, Loss: {'train/loss': 0.4980217218399048, 'train/grad_norm_pre_clip': 16.38689613342285, 'timing/data': 0.010043533518910408, 'timing/model': 0.9417705396190286, 'learning_rate/qwen_vl_interface': 7.29004635439289e-06, 'learning_rate/action_model': 7.29004635439289e-05, 'epoch': 1.06}) +[2026-06-17 20:49:27,244][starVLA.training.train_starvla][INFO] - Step 2576, Loss: {'train/loss': 0.4892144203186035, 'train/grad_norm_pre_clip': 11.868796348571777, 'timing/data': 0.00796264223754406, 'timing/model': 0.9399447832256556, 'learning_rate/qwen_vl_interface': 7.28809307177841e-06, 'learning_rate/action_model': 7.28809307177841e-05, 'epoch': 1.06}) +[2026-06-17 20:49:31,203][starVLA.training.train_starvla][INFO] - Step 2577, Loss: {'train/loss': 0.5217615365982056, 'train/grad_norm_pre_clip': 14.015762329101562, 'timing/data': 0.007925517857074738, 'timing/model': 1.1572904977947474, 'learning_rate/qwen_vl_interface': 7.286139366665044e-06, 'learning_rate/action_model': 7.286139366665043e-05, 'epoch': 1.06}) +[2026-06-17 20:49:34,832][starVLA.training.train_starvla][INFO] - Step 2578, Loss: {'train/loss': 0.6350795030593872, 'train/grad_norm_pre_clip': 17.10605239868164, 'timing/data': 0.010198931209743023, 'timing/model': 0.9363484084606171, 'learning_rate/qwen_vl_interface': 7.284185239457796e-06, 'learning_rate/action_model': 7.284185239457795e-05, 'epoch': 1.06}) +[2026-06-17 20:49:38,729][starVLA.training.train_starvla][INFO] - Step 2579, Loss: {'train/loss': 0.5853769779205322, 'train/grad_norm_pre_clip': 14.189553260803223, 'timing/data': 0.00791159551590681, 'timing/model': 0.9551829444244504, 'learning_rate/qwen_vl_interface': 7.282230690561759e-06, 'learning_rate/action_model': 7.28223069056176e-05, 'epoch': 1.06}) +[2026-06-17 20:49:42,586][starVLA.training.train_starvla][INFO] - Step 2580, Loss: {'train/loss': 0.5576921105384827, 'train/grad_norm_pre_clip': 14.778800010681152, 'timing/data': 0.0078718988224864, 'timing/model': 0.9463970083743334, 'learning_rate/qwen_vl_interface': 7.280275720382113e-06, 'learning_rate/action_model': 7.280275720382113e-05, 'epoch': 1.06}) +[2026-06-17 20:49:46,655][starVLA.training.train_starvla][INFO] - Step 2581, Loss: {'train/loss': 0.7016474008560181, 'train/grad_norm_pre_clip': 15.497127532958984, 'timing/data': 0.008131278678774834, 'timing/model': 0.9388241907581687, 'learning_rate/qwen_vl_interface': 7.2783203293241246e-06, 'learning_rate/action_model': 7.278320329324125e-05, 'epoch': 1.06}) +[2026-06-17 20:49:50,555][starVLA.training.train_starvla][INFO] - Step 2582, Loss: {'train/loss': 0.4614695906639099, 'train/grad_norm_pre_clip': 124.55048370361328, 'timing/data': 0.007676534354686737, 'timing/model': 0.9582276549190283, 'learning_rate/qwen_vl_interface': 7.276364517793152e-06, 'learning_rate/action_model': 7.276364517793151e-05, 'epoch': 1.06}) +[2026-06-17 20:49:54,454][starVLA.training.train_starvla][INFO] - Step 2583, Loss: {'train/loss': 0.4553200304508209, 'train/grad_norm_pre_clip': 11.163625717163086, 'timing/data': 0.013002275489270687, 'timing/model': 1.1674114046618342, 'learning_rate/qwen_vl_interface': 7.274408286194631e-06, 'learning_rate/action_model': 7.27440828619463e-05, 'epoch': 1.06}) +[2026-06-17 20:49:58,436][starVLA.training.train_starvla][INFO] - Step 2584, Loss: {'train/loss': 0.6025179028511047, 'train/grad_norm_pre_clip': 17.781112670898438, 'timing/data': 0.010000172071158886, 'timing/model': 1.1501788292080164, 'learning_rate/qwen_vl_interface': 7.272451634934096e-06, 'learning_rate/action_model': 7.272451634934096e-05, 'epoch': 1.06}) +[2026-06-17 20:50:02,221][starVLA.training.train_starvla][INFO] - Step 2585, Loss: {'train/loss': 0.3932863771915436, 'train/grad_norm_pre_clip': 13.301791191101074, 'timing/data': 0.012189836241304874, 'timing/model': 1.0383828533813357, 'learning_rate/qwen_vl_interface': 7.270494564417161e-06, 'learning_rate/action_model': 7.270494564417161e-05, 'epoch': 1.06}) +[2026-06-17 20:50:06,144][starVLA.training.train_starvla][INFO] - Step 2586, Loss: {'train/loss': 0.44189345836639404, 'train/grad_norm_pre_clip': 12.753775596618652, 'timing/data': 0.007739013992249966, 'timing/model': 0.9407958323135972, 'learning_rate/qwen_vl_interface': 7.268537075049527e-06, 'learning_rate/action_model': 7.268537075049526e-05, 'epoch': 1.06}) +[2026-06-17 20:50:10,139][starVLA.training.train_starvla][INFO] - Step 2587, Loss: {'train/loss': 0.23817026615142822, 'train/grad_norm_pre_clip': 9.356745719909668, 'timing/data': 0.008875733241438866, 'timing/model': 0.9351214058697224, 'learning_rate/qwen_vl_interface': 7.2665791672369875e-06, 'learning_rate/action_model': 7.266579167236988e-05, 'epoch': 1.06}) +[2026-06-17 20:50:13,963][starVLA.training.train_starvla][INFO] - Step 2588, Loss: {'train/loss': 0.6757768392562866, 'train/grad_norm_pre_clip': 17.73393440246582, 'timing/data': 0.014206894673407078, 'timing/model': 0.9419678784906864, 'learning_rate/qwen_vl_interface': 7.264620841385415e-06, 'learning_rate/action_model': 7.264620841385415e-05, 'epoch': 1.07}) +[2026-06-17 20:50:17,832][starVLA.training.train_starvla][INFO] - Step 2589, Loss: {'train/loss': 0.5774420499801636, 'train/grad_norm_pre_clip': 11.543501853942871, 'timing/data': 0.007445497438311577, 'timing/model': 0.9409095635637641, 'learning_rate/qwen_vl_interface': 7.262662097900775e-06, 'learning_rate/action_model': 7.262662097900774e-05, 'epoch': 1.07}) +[2026-06-17 20:50:21,799][starVLA.training.train_starvla][INFO] - Step 2590, Loss: {'train/loss': 0.6648957133293152, 'train/grad_norm_pre_clip': 15.17060375213623, 'timing/data': 0.007650159299373627, 'timing/model': 1.2606143644079566, 'learning_rate/qwen_vl_interface': 7.260702937189116e-06, 'learning_rate/action_model': 7.260702937189115e-05, 'epoch': 1.07}) +[2026-06-17 20:50:25,644][starVLA.training.train_starvla][INFO] - Step 2591, Loss: {'train/loss': 0.5389437079429626, 'train/grad_norm_pre_clip': 13.639830589294434, 'timing/data': 0.008811124600470066, 'timing/model': 1.1331113949418068, 'learning_rate/qwen_vl_interface': 7.258743359656574e-06, 'learning_rate/action_model': 7.258743359656573e-05, 'epoch': 1.07}) +[2026-06-17 20:50:29,360][starVLA.training.train_starvla][INFO] - Step 2592, Loss: {'train/loss': 0.3714671730995178, 'train/grad_norm_pre_clip': 94.476318359375, 'timing/data': 0.010461810044944286, 'timing/model': 0.9462912222370505, 'learning_rate/qwen_vl_interface': 7.2567833657093725e-06, 'learning_rate/action_model': 7.256783365709371e-05, 'epoch': 1.07}) +[2026-06-17 20:50:33,701][starVLA.training.train_starvla][INFO] - Step 2593, Loss: {'train/loss': 0.44122838973999023, 'train/grad_norm_pre_clip': 12.730290412902832, 'timing/data': 0.009823598898947239, 'timing/model': 0.9496191991493106, 'learning_rate/qwen_vl_interface': 7.254822955753821e-06, 'learning_rate/action_model': 7.25482295575382e-05, 'epoch': 1.07}) +[2026-06-17 20:50:37,536][starVLA.training.train_starvla][INFO] - Step 2594, Loss: {'train/loss': 0.4872390031814575, 'train/grad_norm_pre_clip': 13.34986400604248, 'timing/data': 0.008823882788419724, 'timing/model': 0.9354695407673717, 'learning_rate/qwen_vl_interface': 7.2528621301963135e-06, 'learning_rate/action_model': 7.252862130196313e-05, 'epoch': 1.07}) +[2026-06-17 20:50:41,337][starVLA.training.train_starvla][INFO] - Step 2595, Loss: {'train/loss': 0.48021844029426575, 'train/grad_norm_pre_clip': 14.267210960388184, 'timing/data': 0.01417316123843193, 'timing/model': 0.9479649998247623, 'learning_rate/qwen_vl_interface': 7.25090088944333e-06, 'learning_rate/action_model': 7.250900889443329e-05, 'epoch': 1.07}) +[2026-06-17 20:50:45,318][starVLA.training.train_starvla][INFO] - Step 2596, Loss: {'train/loss': 0.47286757826805115, 'train/grad_norm_pre_clip': 9.833730697631836, 'timing/data': 0.015124784782528877, 'timing/model': 0.9576052688062191, 'learning_rate/qwen_vl_interface': 7.248939233901438e-06, 'learning_rate/action_model': 7.248939233901439e-05, 'epoch': 1.07}) +[2026-06-17 20:50:49,175][starVLA.training.train_starvla][INFO] - Step 2597, Loss: {'train/loss': 0.32019561529159546, 'train/grad_norm_pre_clip': 9.945737838745117, 'timing/data': 0.008048499934375286, 'timing/model': 1.143990502692759, 'learning_rate/qwen_vl_interface': 7.2469771639772944e-06, 'learning_rate/action_model': 7.246977163977294e-05, 'epoch': 1.07}) +[2026-06-17 20:50:53,024][starVLA.training.train_starvla][INFO] - Step 2598, Loss: {'train/loss': 0.24127990007400513, 'train/grad_norm_pre_clip': 8.890518188476562, 'timing/data': 0.008653167635202408, 'timing/model': 0.9735899679362774, 'learning_rate/qwen_vl_interface': 7.245014680077633e-06, 'learning_rate/action_model': 7.245014680077633e-05, 'epoch': 1.07}) +[2026-06-17 20:50:56,919][starVLA.training.train_starvla][INFO] - Step 2599, Loss: {'train/loss': 0.35510021448135376, 'train/grad_norm_pre_clip': 16.50772476196289, 'timing/data': 0.00834622886031866, 'timing/model': 0.9441713308915496, 'learning_rate/qwen_vl_interface': 7.243051782609285e-06, 'learning_rate/action_model': 7.243051782609285e-05, 'epoch': 1.07}) +[2026-06-17 20:53:10,704][starVLA.training.train_starvla][INFO] - Step 2600, Loss: {'train/loss': 0.4430547058582306, 'train/grad_norm_pre_clip': 13.142711639404297, 'eval/loss': 0.5369272232055664, 'eval/latency_6/loss': 0.5369272232055664, 'eval/demon_fire_precision': 0.9424214130096483, 'eval/demon_fire_recall': 0.9786683904330963, 'eval/demon_fire_f1': 0.96020294910417, 'eval/demon_fire_rate_model': 0.8939899833055092, 'eval/demon_fire_rate_teacher': 0.860879243183083, 'eval/demon_left_precision': 0.8375870069605569, 'eval/demon_left_recall': 0.6333333333333333, 'eval/demon_left_f1': 0.7212787212787213, 'eval/demon_left_rate_model': 0.11992209237618252, 'eval/demon_left_rate_teacher': 0.15859766277128548, 'eval/demon_right_precision': 0.7639751552795031, 'eval/demon_right_recall': 0.7557603686635944, 'eval/demon_right_f1': 0.7598455598455598, 'eval/demon_right_rate_model': 0.1791875347801892, 'eval/demon_right_rate_teacher': 0.18113522537562604, 'eval/demon_move_f1': 0.7405621405621405, 'eval/demon_cc_f1': 0.881748510817712, 'eval/latency_6/demon_fire_precision': 0.9424214130096483, 'eval/latency_6/demon_fire_recall': 0.9786683904330963, 'eval/latency_6/demon_fire_f1': 0.96020294910417, 'eval/latency_6/demon_fire_rate_model': 0.8939899833055092, 'eval/latency_6/demon_fire_rate_teacher': 0.860879243183083, 'eval/latency_6/demon_left_precision': 0.8375870069605569, 'eval/latency_6/demon_left_recall': 0.6333333333333333, 'eval/latency_6/demon_left_f1': 0.7212787212787213, 'eval/latency_6/demon_left_rate_model': 0.11992209237618252, 'eval/latency_6/demon_left_rate_teacher': 0.15859766277128548, 'eval/latency_6/demon_right_precision': 0.7639751552795031, 'eval/latency_6/demon_right_recall': 0.7557603686635944, 'eval/latency_6/demon_right_f1': 0.7598455598455598, 'eval/latency_6/demon_right_rate_model': 0.1791875347801892, 'eval/latency_6/demon_right_rate_teacher': 0.18113522537562604, 'eval/latency_6/demon_move_f1': 0.7405621405621405, 'eval/latency_6/demon_cc_f1': 0.881748510817712, 'timing/data': 0.00993653479963541, 'timing/model': 0.9478369308635592, 'learning_rate/qwen_vl_interface': 7.241088471979155e-06, 'learning_rate/action_model': 7.241088471979155e-05, 'epoch': 1.07}) +[2026-06-17 20:53:10,712][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_2600_state +[2026-06-17 20:53:10,717][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 20:53:17,428][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_2600_state/model.safetensors +[2026-06-17 20:53:32,812][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_2600_state/optimizer.bin +[2026-06-17 20:53:32,813][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_2600_state/sampler.bin +[2026-06-17 20:53:32,813][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_2600_state/sampler_1.bin +[2026-06-17 20:53:32,816][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_2600_state/random_states_0.pkl +[2026-06-17 20:53:32,822][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2600 +[2026-06-17 20:53:32,823][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 20:53:32,834][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 20:53:36,545][starVLA.training.train_starvla][INFO] - Step 2601, Loss: {'train/loss': 0.6336659789085388, 'train/grad_norm_pre_clip': 13.034509658813477, 'timing/data': 0.011248955503106117, 'timing/model': 0.9857295695692301, 'learning_rate/qwen_vl_interface': 7.239124748594244e-06, 'learning_rate/action_model': 7.239124748594244e-05, 'epoch': 1.07}) +[2026-06-17 20:53:40,880][starVLA.training.train_starvla][INFO] - Step 2602, Loss: {'train/loss': 0.6209363341331482, 'train/grad_norm_pre_clip': 23.895145416259766, 'timing/data': 0.008373329415917397, 'timing/model': 0.9422415569424629, 'learning_rate/qwen_vl_interface': 7.237160612861632e-06, 'learning_rate/action_model': 7.237160612861632e-05, 'epoch': 1.07}) +[2026-06-17 20:53:44,830][starVLA.training.train_starvla][INFO] - Step 2603, Loss: {'train/loss': 0.2844514846801758, 'train/grad_norm_pre_clip': 8.978734016418457, 'timing/data': 0.006868455559015274, 'timing/model': 0.9537130957469344, 'learning_rate/qwen_vl_interface': 7.235196065188485e-06, 'learning_rate/action_model': 7.235196065188485e-05, 'epoch': 1.07}) +[2026-06-17 20:53:48,597][starVLA.training.train_starvla][INFO] - Step 2604, Loss: {'train/loss': 0.4604373574256897, 'train/grad_norm_pre_clip': 16.39971160888672, 'timing/data': 0.012614806182682514, 'timing/model': 0.9504560315981507, 'learning_rate/qwen_vl_interface': 7.23323110598206e-06, 'learning_rate/action_model': 7.23323110598206e-05, 'epoch': 1.07}) +[2026-06-17 20:53:52,477][starVLA.training.train_starvla][INFO] - Step 2605, Loss: {'train/loss': 0.5377238988876343, 'train/grad_norm_pre_clip': 14.594915390014648, 'timing/data': 0.01046501100063324, 'timing/model': 1.0088192373514175, 'learning_rate/qwen_vl_interface': 7.231265735649691e-06, 'learning_rate/action_model': 7.231265735649691e-05, 'epoch': 1.07}) +[2026-06-17 20:53:56,325][starVLA.training.train_starvla][INFO] - Step 2606, Loss: {'train/loss': 0.44494858384132385, 'train/grad_norm_pre_clip': 13.9882173538208, 'timing/data': 0.010868627578020096, 'timing/model': 1.114149914123118, 'learning_rate/qwen_vl_interface': 7.229299954598805e-06, 'learning_rate/action_model': 7.229299954598804e-05, 'epoch': 1.07}) +[2026-06-17 20:54:00,066][starVLA.training.train_starvla][INFO] - Step 2607, Loss: {'train/loss': 0.6225740313529968, 'train/grad_norm_pre_clip': 13.80257797241211, 'timing/data': 0.007786299102008343, 'timing/model': 1.0631743967533112, 'learning_rate/qwen_vl_interface': 7.227333763236907e-06, 'learning_rate/action_model': 7.227333763236906e-05, 'epoch': 1.07}) +[2026-06-17 20:54:03,738][starVLA.training.train_starvla][INFO] - Step 2608, Loss: {'train/loss': 0.347962349653244, 'train/grad_norm_pre_clip': 11.894083023071289, 'timing/data': 0.007273049093782902, 'timing/model': 0.9715188983827829, 'learning_rate/qwen_vl_interface': 7.225367161971592e-06, 'learning_rate/action_model': 7.225367161971591e-05, 'epoch': 1.07}) +[2026-06-17 20:54:07,641][starVLA.training.train_starvla][INFO] - Step 2609, Loss: {'train/loss': 0.29571205377578735, 'train/grad_norm_pre_clip': 9.705331802368164, 'timing/data': 0.012817270122468472, 'timing/model': 0.9396282229572535, 'learning_rate/qwen_vl_interface': 7.223400151210539e-06, 'learning_rate/action_model': 7.223400151210538e-05, 'epoch': 1.07}) +[2026-06-17 20:54:11,443][starVLA.training.train_starvla][INFO] - Step 2610, Loss: {'train/loss': 0.263347327709198, 'train/grad_norm_pre_clip': 9.921041488647461, 'timing/data': 0.008309916593134403, 'timing/model': 0.9388175792992115, 'learning_rate/qwen_vl_interface': 7.2214327313615105e-06, 'learning_rate/action_model': 7.22143273136151e-05, 'epoch': 1.07}) +[2026-06-17 20:54:15,262][starVLA.training.train_starvla][INFO] - Step 2611, Loss: {'train/loss': 0.3728175759315491, 'train/grad_norm_pre_clip': 9.353351593017578, 'timing/data': 0.01125432550907135, 'timing/model': 0.9405450923368335, 'learning_rate/qwen_vl_interface': 7.219464902832357e-06, 'learning_rate/action_model': 7.219464902832356e-05, 'epoch': 1.07}) +[2026-06-17 20:54:19,158][starVLA.training.train_starvla][INFO] - Step 2612, Loss: {'train/loss': 0.6014427542686462, 'train/grad_norm_pre_clip': 14.709569931030273, 'timing/data': 0.009584193117916584, 'timing/model': 0.9719794932752848, 'learning_rate/qwen_vl_interface': 7.2174966660310085e-06, 'learning_rate/action_model': 7.217496666031008e-05, 'epoch': 1.08}) +[2026-06-17 20:54:23,046][starVLA.training.train_starvla][INFO] - Step 2613, Loss: {'train/loss': 0.33781829476356506, 'train/grad_norm_pre_clip': 10.77971363067627, 'timing/data': 0.008237069472670555, 'timing/model': 1.147156022489071, 'learning_rate/qwen_vl_interface': 7.215528021365487e-06, 'learning_rate/action_model': 7.215528021365487e-05, 'epoch': 1.08}) +[2026-06-17 20:54:26,726][starVLA.training.train_starvla][INFO] - Step 2614, Loss: {'train/loss': 0.554987370967865, 'train/grad_norm_pre_clip': 13.33492660522461, 'timing/data': 0.009537400677800179, 'timing/model': 0.9479319201782346, 'learning_rate/qwen_vl_interface': 7.213558969243888e-06, 'learning_rate/action_model': 7.213558969243888e-05, 'epoch': 1.08}) +[2026-06-17 20:54:30,527][starVLA.training.train_starvla][INFO] - Step 2615, Loss: {'train/loss': 1.0407968759536743, 'train/grad_norm_pre_clip': 22.08748435974121, 'timing/data': 0.00979558564722538, 'timing/model': 0.9412023788318038, 'learning_rate/qwen_vl_interface': 7.211589510074404e-06, 'learning_rate/action_model': 7.211589510074404e-05, 'epoch': 1.08}) +[2026-06-17 20:54:34,292][starVLA.training.train_starvla][INFO] - Step 2616, Loss: {'train/loss': 0.4234301745891571, 'train/grad_norm_pre_clip': 11.22370719909668, 'timing/data': 0.009798618964850903, 'timing/model': 0.9398288987576962, 'learning_rate/qwen_vl_interface': 7.209619644265302e-06, 'learning_rate/action_model': 7.209619644265302e-05, 'epoch': 1.08}) +[2026-06-17 20:54:38,105][starVLA.training.train_starvla][INFO] - Step 2617, Loss: {'train/loss': 0.3591272532939911, 'train/grad_norm_pre_clip': 10.907666206359863, 'timing/data': 0.011927195824682713, 'timing/model': 0.938802681863308, 'learning_rate/qwen_vl_interface': 7.20764937222494e-06, 'learning_rate/action_model': 7.20764937222494e-05, 'epoch': 1.08}) +[2026-06-17 20:54:41,947][starVLA.training.train_starvla][INFO] - Step 2618, Loss: {'train/loss': 0.6908483505249023, 'train/grad_norm_pre_clip': 15.657644271850586, 'timing/data': 0.015650018118321896, 'timing/model': 0.9304261673241854, 'learning_rate/qwen_vl_interface': 7.205678694361757e-06, 'learning_rate/action_model': 7.205678694361757e-05, 'epoch': 1.08}) +[2026-06-17 20:54:45,839][starVLA.training.train_starvla][INFO] - Step 2619, Loss: {'train/loss': 0.559607744216919, 'train/grad_norm_pre_clip': 15.673270225524902, 'timing/data': 0.0073645636439323425, 'timing/model': 1.1829370111227036, 'learning_rate/qwen_vl_interface': 7.203707611084276e-06, 'learning_rate/action_model': 7.203707611084275e-05, 'epoch': 1.08}) +[2026-06-17 20:54:49,689][starVLA.training.train_starvla][INFO] - Step 2620, Loss: {'train/loss': 0.38696226477622986, 'train/grad_norm_pre_clip': 13.154308319091797, 'timing/data': 0.014124972745776176, 'timing/model': 1.1278360318392515, 'learning_rate/qwen_vl_interface': 7.2017361228011055e-06, 'learning_rate/action_model': 7.201736122801104e-05, 'epoch': 1.08}) +[2026-06-17 20:54:53,394][starVLA.training.train_starvla][INFO] - Step 2621, Loss: {'train/loss': 0.5144549012184143, 'train/grad_norm_pre_clip': 20.581321716308594, 'timing/data': 0.009681250900030136, 'timing/model': 0.997464525513351, 'learning_rate/qwen_vl_interface': 7.199764229920935e-06, 'learning_rate/action_model': 7.199764229920934e-05, 'epoch': 1.08}) +[2026-06-17 20:54:57,290][starVLA.training.train_starvla][INFO] - Step 2622, Loss: {'train/loss': 0.7173107862472534, 'train/grad_norm_pre_clip': 16.865015029907227, 'timing/data': 0.011414164677262306, 'timing/model': 0.9485886376351118, 'learning_rate/qwen_vl_interface': 7.197791932852543e-06, 'learning_rate/action_model': 7.197791932852542e-05, 'epoch': 1.08}) +[2026-06-17 20:55:01,123][starVLA.training.train_starvla][INFO] - Step 2623, Loss: {'train/loss': 0.36955443024635315, 'train/grad_norm_pre_clip': 10.21121597290039, 'timing/data': 0.009705898351967335, 'timing/model': 0.9493327690288424, 'learning_rate/qwen_vl_interface': 7.195819232004787e-06, 'learning_rate/action_model': 7.195819232004787e-05, 'epoch': 1.08}) +[2026-06-17 20:55:04,927][starVLA.training.train_starvla][INFO] - Step 2624, Loss: {'train/loss': 0.660727858543396, 'train/grad_norm_pre_clip': 19.270065307617188, 'timing/data': 0.010597586631774902, 'timing/model': 0.9519241014495492, 'learning_rate/qwen_vl_interface': 7.1938461277866105e-06, 'learning_rate/action_model': 7.19384612778661e-05, 'epoch': 1.08}) +[2026-06-17 20:55:08,859][starVLA.training.train_starvla][INFO] - Step 2625, Loss: {'train/loss': 0.6306671500205994, 'train/grad_norm_pre_clip': 15.888216972351074, 'timing/data': 0.00906774215400219, 'timing/model': 0.9638957884162664, 'learning_rate/qwen_vl_interface': 7.19187262060704e-06, 'learning_rate/action_model': 7.19187262060704e-05, 'epoch': 1.08}) +[2026-06-17 20:55:12,817][starVLA.training.train_starvla][INFO] - Step 2626, Loss: {'train/loss': 0.5832181572914124, 'train/grad_norm_pre_clip': 13.774096488952637, 'timing/data': 0.009335371665656567, 'timing/model': 1.2248376188799739, 'learning_rate/qwen_vl_interface': 7.189898710875185e-06, 'learning_rate/action_model': 7.189898710875184e-05, 'epoch': 1.08}) +[2026-06-17 20:55:16,722][starVLA.training.train_starvla][INFO] - Step 2627, Loss: {'train/loss': 0.4955262839794159, 'train/grad_norm_pre_clip': 13.034510612487793, 'timing/data': 0.009335394948720932, 'timing/model': 1.1345402104780078, 'learning_rate/qwen_vl_interface': 7.187924399000241e-06, 'learning_rate/action_model': 7.18792439900024e-05, 'epoch': 1.08}) +[2026-06-17 20:55:20,397][starVLA.training.train_starvla][INFO] - Step 2628, Loss: {'train/loss': 0.4374334514141083, 'train/grad_norm_pre_clip': 12.96109390258789, 'timing/data': 0.01020199153572321, 'timing/model': 0.9552028104662895, 'learning_rate/qwen_vl_interface': 7.185949685391483e-06, 'learning_rate/action_model': 7.185949685391483e-05, 'epoch': 1.08}) +[2026-06-17 20:55:24,238][starVLA.training.train_starvla][INFO] - Step 2629, Loss: {'train/loss': 0.5504887104034424, 'train/grad_norm_pre_clip': 18.651702880859375, 'timing/data': 0.009112679399549961, 'timing/model': 0.9402424432337284, 'learning_rate/qwen_vl_interface': 7.183974570458272e-06, 'learning_rate/action_model': 7.183974570458272e-05, 'epoch': 1.08}) +[2026-06-17 20:55:28,034][starVLA.training.train_starvla][INFO] - Step 2630, Loss: {'train/loss': 0.6399896144866943, 'train/grad_norm_pre_clip': 19.582731246948242, 'timing/data': 0.010613145306706429, 'timing/model': 0.9300686633214355, 'learning_rate/qwen_vl_interface': 7.181999054610053e-06, 'learning_rate/action_model': 7.181999054610052e-05, 'epoch': 1.08}) +[2026-06-17 20:55:31,802][starVLA.training.train_starvla][INFO] - Step 2631, Loss: {'train/loss': 0.40480485558509827, 'train/grad_norm_pre_clip': 12.95773983001709, 'timing/data': 0.010335840284824371, 'timing/model': 0.9550050189718604, 'learning_rate/qwen_vl_interface': 7.180023138256349e-06, 'learning_rate/action_model': 7.180023138256348e-05, 'epoch': 1.08}) +[2026-06-17 20:55:35,643][starVLA.training.train_starvla][INFO] - Step 2632, Loss: {'train/loss': 0.6523873805999756, 'train/grad_norm_pre_clip': 12.500478744506836, 'timing/data': 0.008504623547196388, 'timing/model': 0.9506663279607892, 'learning_rate/qwen_vl_interface': 7.178046821806771e-06, 'learning_rate/action_model': 7.17804682180677e-05, 'epoch': 1.08}) +[2026-06-17 20:55:39,651][starVLA.training.train_starvla][INFO] - Step 2633, Loss: {'train/loss': 0.4196341633796692, 'train/grad_norm_pre_clip': 13.991487503051758, 'timing/data': 0.01255229115486145, 'timing/model': 1.1922991089522839, 'learning_rate/qwen_vl_interface': 7.176070105671016e-06, 'learning_rate/action_model': 7.176070105671015e-05, 'epoch': 1.08}) +[2026-06-17 20:55:43,465][starVLA.training.train_starvla][INFO] - Step 2634, Loss: {'train/loss': 0.5895938873291016, 'train/grad_norm_pre_clip': 13.190135955810547, 'timing/data': 0.010351195000112057, 'timing/model': 0.9403154384344816, 'learning_rate/qwen_vl_interface': 7.174092990258852e-06, 'learning_rate/action_model': 7.174092990258851e-05, 'epoch': 1.08}) +[2026-06-17 20:55:47,624][starVLA.training.train_starvla][INFO] - Step 2635, Loss: {'train/loss': 0.46011075377464294, 'train/grad_norm_pre_clip': 14.597676277160645, 'timing/data': 0.011390909552574158, 'timing/model': 1.019527188502252, 'learning_rate/qwen_vl_interface': 7.172115475980142e-06, 'learning_rate/action_model': 7.172115475980142e-05, 'epoch': 1.08}) +[2026-06-17 20:55:51,543][starVLA.training.train_starvla][INFO] - Step 2636, Loss: {'train/loss': 0.3485881984233856, 'train/grad_norm_pre_clip': 9.292993545532227, 'timing/data': 0.007847950793802738, 'timing/model': 0.9516092417761683, 'learning_rate/qwen_vl_interface': 7.170137563244825e-06, 'learning_rate/action_model': 7.170137563244824e-05, 'epoch': 1.09}) +[2026-06-17 20:55:55,394][starVLA.training.train_starvla][INFO] - Step 2637, Loss: {'train/loss': 0.7395541071891785, 'train/grad_norm_pre_clip': 15.660416603088379, 'timing/data': 0.01238001137971878, 'timing/model': 0.9592079035937786, 'learning_rate/qwen_vl_interface': 7.168159252462927e-06, 'learning_rate/action_model': 7.168159252462926e-05, 'epoch': 1.09}) +[2026-06-17 20:55:59,416][starVLA.training.train_starvla][INFO] - Step 2638, Loss: {'train/loss': 0.27205008268356323, 'train/grad_norm_pre_clip': 10.29410171508789, 'timing/data': 0.008971672505140305, 'timing/model': 0.9483204493299127, 'learning_rate/qwen_vl_interface': 7.166180544044549e-06, 'learning_rate/action_model': 7.166180544044549e-05, 'epoch': 1.09}) +[2026-06-17 20:56:03,249][starVLA.training.train_starvla][INFO] - Step 2639, Loss: {'train/loss': 0.5291453003883362, 'train/grad_norm_pre_clip': 13.118041038513184, 'timing/data': 0.009171335957944393, 'timing/model': 0.9347169687971473, 'learning_rate/qwen_vl_interface': 7.164201438399884e-06, 'learning_rate/action_model': 7.164201438399884e-05, 'epoch': 1.09}) +[2026-06-17 20:56:07,101][starVLA.training.train_starvla][INFO] - Step 2640, Loss: {'train/loss': 0.374781996011734, 'train/grad_norm_pre_clip': 10.105079650878906, 'timing/data': 0.0073446230962872505, 'timing/model': 1.1415754333138466, 'learning_rate/qwen_vl_interface': 7.1622219359392e-06, 'learning_rate/action_model': 7.162221935939199e-05, 'epoch': 1.09}) +[2026-06-17 20:56:10,766][starVLA.training.train_starvla][INFO] - Step 2641, Loss: {'train/loss': 0.846694827079773, 'train/grad_norm_pre_clip': 14.240280151367188, 'timing/data': 0.014095078222453594, 'timing/model': 0.9409986538812518, 'learning_rate/qwen_vl_interface': 7.160242037072852e-06, 'learning_rate/action_model': 7.160242037072851e-05, 'epoch': 1.09}) +[2026-06-17 20:56:14,629][starVLA.training.train_starvla][INFO] - Step 2642, Loss: {'train/loss': 0.43164417147636414, 'train/grad_norm_pre_clip': 14.587059020996094, 'timing/data': 0.011058834381401539, 'timing/model': 0.9421586645767093, 'learning_rate/qwen_vl_interface': 7.158261742211275e-06, 'learning_rate/action_model': 7.158261742211274e-05, 'epoch': 1.09}) +[2026-06-17 20:56:18,443][starVLA.training.train_starvla][INFO] - Step 2643, Loss: {'train/loss': 0.3576952815055847, 'train/grad_norm_pre_clip': 16.486846923828125, 'timing/data': 0.011189931072294712, 'timing/model': 0.9334904318675399, 'learning_rate/qwen_vl_interface': 7.156281051764983e-06, 'learning_rate/action_model': 7.156281051764983e-05, 'epoch': 1.09}) +[2026-06-17 20:56:22,335][starVLA.training.train_starvla][INFO] - Step 2644, Loss: {'train/loss': 0.586175262928009, 'train/grad_norm_pre_clip': 12.639607429504395, 'timing/data': 0.008600921370089054, 'timing/model': 0.9404134228825569, 'learning_rate/qwen_vl_interface': 7.15429996614458e-06, 'learning_rate/action_model': 7.15429996614458e-05, 'epoch': 1.09}) +[2026-06-17 20:56:26,178][starVLA.training.train_starvla][INFO] - Step 2645, Loss: {'train/loss': 0.5033840537071228, 'train/grad_norm_pre_clip': 13.328166961669922, 'timing/data': 0.009249760769307613, 'timing/model': 0.9460059469565749, 'learning_rate/qwen_vl_interface': 7.152318485760743e-06, 'learning_rate/action_model': 7.152318485760742e-05, 'epoch': 1.09}) +[2026-06-17 20:56:30,029][starVLA.training.train_starvla][INFO] - Step 2646, Loss: {'train/loss': 0.31206104159355164, 'train/grad_norm_pre_clip': 9.651382446289062, 'timing/data': 0.007730142213404179, 'timing/model': 1.135799253359437, 'learning_rate/qwen_vl_interface': 7.150336611024239e-06, 'learning_rate/action_model': 7.150336611024239e-05, 'epoch': 1.09}) +[2026-06-17 20:56:33,863][starVLA.training.train_starvla][INFO] - Step 2647, Loss: {'train/loss': 0.6018503308296204, 'train/grad_norm_pre_clip': 14.0452241897583, 'timing/data': 0.00969090685248375, 'timing/model': 1.1421716781333089, 'learning_rate/qwen_vl_interface': 7.1483543423459086e-06, 'learning_rate/action_model': 7.148354342345908e-05, 'epoch': 1.09}) +[2026-06-17 20:56:37,553][starVLA.training.train_starvla][INFO] - Step 2648, Loss: {'train/loss': 0.2906684875488281, 'train/grad_norm_pre_clip': 8.556838989257812, 'timing/data': 0.012598361819982529, 'timing/model': 0.9517262140288949, 'learning_rate/qwen_vl_interface': 7.146371680136682e-06, 'learning_rate/action_model': 7.146371680136681e-05, 'epoch': 1.09}) +[2026-06-17 20:56:41,473][starVLA.training.train_starvla][INFO] - Step 2649, Loss: {'train/loss': 0.6357027292251587, 'train/grad_norm_pre_clip': 14.719372749328613, 'timing/data': 0.011923127807676792, 'timing/model': 0.9426308590918779, 'learning_rate/qwen_vl_interface': 7.144388624807565e-06, 'learning_rate/action_model': 7.144388624807566e-05, 'epoch': 1.09}) +[2026-06-17 20:56:45,333][starVLA.training.train_starvla][INFO] - Step 2650, Loss: {'train/loss': 0.4527006149291992, 'train/grad_norm_pre_clip': 12.14867115020752, 'timing/data': 0.011821307241916656, 'timing/model': 0.9392426982522011, 'learning_rate/qwen_vl_interface': 7.142405176769648e-06, 'learning_rate/action_model': 7.142405176769648e-05, 'epoch': 1.09}) +[2026-06-17 20:56:49,144][starVLA.training.train_starvla][INFO] - Step 2651, Loss: {'train/loss': 0.5420740842819214, 'train/grad_norm_pre_clip': 12.241003036499023, 'timing/data': 0.009338922798633575, 'timing/model': 0.9256580006331205, 'learning_rate/qwen_vl_interface': 7.140421336434102e-06, 'learning_rate/action_model': 7.140421336434102e-05, 'epoch': 1.09}) +[2026-06-17 20:56:53,002][starVLA.training.train_starvla][INFO] - Step 2652, Loss: {'train/loss': 0.3079986572265625, 'train/grad_norm_pre_clip': 9.471214294433594, 'timing/data': 0.014561064541339874, 'timing/model': 0.94699377566576, 'learning_rate/qwen_vl_interface': 7.138437104212176e-06, 'learning_rate/action_model': 7.138437104212176e-05, 'epoch': 1.09}) +[2026-06-17 20:56:56,845][starVLA.training.train_starvla][INFO] - Step 2653, Loss: {'train/loss': 0.3363097906112671, 'train/grad_norm_pre_clip': 10.367429733276367, 'timing/data': 0.010339422151446342, 'timing/model': 1.1425554109737277, 'learning_rate/qwen_vl_interface': 7.136452480515208e-06, 'learning_rate/action_model': 7.136452480515208e-05, 'epoch': 1.09}) +[2026-06-17 20:57:00,718][starVLA.training.train_starvla][INFO] - Step 2654, Loss: {'train/loss': 0.30140236020088196, 'train/grad_norm_pre_clip': 11.875018119812012, 'timing/data': 0.00921806413680315, 'timing/model': 1.1497464422136545, 'learning_rate/qwen_vl_interface': 7.134467465754612e-06, 'learning_rate/action_model': 7.134467465754612e-05, 'epoch': 1.09}) +[2026-06-17 20:57:04,349][starVLA.training.train_starvla][INFO] - Step 2655, Loss: {'train/loss': 0.28962549567222595, 'train/grad_norm_pre_clip': 13.680648803710938, 'timing/data': 0.007474234327673912, 'timing/model': 0.9336560405790806, 'learning_rate/qwen_vl_interface': 7.132482060341881e-06, 'learning_rate/action_model': 7.13248206034188e-05, 'epoch': 1.09}) +[2026-06-17 20:57:08,210][starVLA.training.train_starvla][INFO] - Step 2656, Loss: {'train/loss': 0.6347042322158813, 'train/grad_norm_pre_clip': 15.559800148010254, 'timing/data': 0.007332630455493927, 'timing/model': 0.9349062135443091, 'learning_rate/qwen_vl_interface': 7.1304962646885935e-06, 'learning_rate/action_model': 7.130496264688593e-05, 'epoch': 1.09}) +[2026-06-17 20:57:11,999][starVLA.training.train_starvla][INFO] - Step 2657, Loss: {'train/loss': 0.5953293442726135, 'train/grad_norm_pre_clip': 14.260773658752441, 'timing/data': 0.012675863690674305, 'timing/model': 0.9388982355594635, 'learning_rate/qwen_vl_interface': 7.128510079206408e-06, 'learning_rate/action_model': 7.128510079206407e-05, 'epoch': 1.09}) +[2026-06-17 20:57:15,728][starVLA.training.train_starvla][INFO] - Step 2658, Loss: {'train/loss': 0.41892802715301514, 'train/grad_norm_pre_clip': 10.829227447509766, 'timing/data': 0.015515762381255627, 'timing/model': 0.939926428720355, 'learning_rate/qwen_vl_interface': 7.126523504307061e-06, 'learning_rate/action_model': 7.126523504307061e-05, 'epoch': 1.09}) +[2026-06-17 20:57:19,613][starVLA.training.train_starvla][INFO] - Step 2659, Loss: {'train/loss': 0.787341833114624, 'train/grad_norm_pre_clip': 19.18325424194336, 'timing/data': 0.012731283903121948, 'timing/model': 0.9785822071135044, 'learning_rate/qwen_vl_interface': 7.124536540402371e-06, 'learning_rate/action_model': 7.124536540402372e-05, 'epoch': 1.09}) +[2026-06-17 20:57:23,461][starVLA.training.train_starvla][INFO] - Step 2660, Loss: {'train/loss': 0.7812336683273315, 'train/grad_norm_pre_clip': 15.614007949829102, 'timing/data': 0.008189154788851738, 'timing/model': 1.1642099563032389, 'learning_rate/qwen_vl_interface': 7.1225491879042405e-06, 'learning_rate/action_model': 7.12254918790424e-05, 'epoch': 1.09}) +[2026-06-17 20:57:27,137][starVLA.training.train_starvla][INFO] - Step 2661, Loss: {'train/loss': 0.4148375391960144, 'train/grad_norm_pre_clip': 11.174495697021484, 'timing/data': 0.01173487026244402, 'timing/model': 0.9470205940306187, 'learning_rate/qwen_vl_interface': 7.120561447224649e-06, 'learning_rate/action_model': 7.120561447224649e-05, 'epoch': 1.1}) +[2026-06-17 20:57:30,937][starVLA.training.train_starvla][INFO] - Step 2662, Loss: {'train/loss': 0.5501872301101685, 'train/grad_norm_pre_clip': 12.506881713867188, 'timing/data': 0.010149669833481312, 'timing/model': 0.9364323932677507, 'learning_rate/qwen_vl_interface': 7.118573318775656e-06, 'learning_rate/action_model': 7.118573318775655e-05, 'epoch': 1.1}) +[2026-06-17 20:57:34,723][starVLA.training.train_starvla][INFO] - Step 2663, Loss: {'train/loss': 0.6730248332023621, 'train/grad_norm_pre_clip': 15.297537803649902, 'timing/data': 0.01175733096897602, 'timing/model': 0.9540465604513884, 'learning_rate/qwen_vl_interface': 7.116584802969403e-06, 'learning_rate/action_model': 7.116584802969402e-05, 'epoch': 1.1}) +[2026-06-17 20:57:38,539][starVLA.training.train_starvla][INFO] - Step 2664, Loss: {'train/loss': 0.49246376752853394, 'train/grad_norm_pre_clip': 12.265924453735352, 'timing/data': 0.011804517358541489, 'timing/model': 0.9454693086445332, 'learning_rate/qwen_vl_interface': 7.114595900218112e-06, 'learning_rate/action_model': 7.114595900218112e-05, 'epoch': 1.1}) +[2026-06-17 20:57:42,377][starVLA.training.train_starvla][INFO] - Step 2665, Loss: {'train/loss': 0.6001343727111816, 'train/grad_norm_pre_clip': 15.65439224243164, 'timing/data': 0.011500674299895763, 'timing/model': 0.9351420737802982, 'learning_rate/qwen_vl_interface': 7.112606610934084e-06, 'learning_rate/action_model': 7.112606610934084e-05, 'epoch': 1.1}) +[2026-06-17 20:57:46,237][starVLA.training.train_starvla][INFO] - Step 2666, Loss: {'train/loss': 0.5831394791603088, 'train/grad_norm_pre_clip': 16.146652221679688, 'timing/data': 0.013447307050228119, 'timing/model': 1.1436882596462965, 'learning_rate/qwen_vl_interface': 7.110616935529702e-06, 'learning_rate/action_model': 7.110616935529702e-05, 'epoch': 1.1}) +[2026-06-17 20:57:50,117][starVLA.training.train_starvla][INFO] - Step 2667, Loss: {'train/loss': 0.38825830817222595, 'train/grad_norm_pre_clip': 11.969378471374512, 'timing/data': 0.008258995600044727, 'timing/model': 1.1212503649294376, 'learning_rate/qwen_vl_interface': 7.108626874417426e-06, 'learning_rate/action_model': 7.108626874417426e-05, 'epoch': 1.1}) +[2026-06-17 20:57:53,783][starVLA.training.train_starvla][INFO] - Step 2668, Loss: {'train/loss': 0.41506218910217285, 'train/grad_norm_pre_clip': 13.06753158569336, 'timing/data': 0.009538772515952587, 'timing/model': 0.9491401948034763, 'learning_rate/qwen_vl_interface': 7.106636428009801e-06, 'learning_rate/action_model': 7.1066364280098e-05, 'epoch': 1.1}) +[2026-06-17 20:57:57,680][starVLA.training.train_starvla][INFO] - Step 2669, Loss: {'train/loss': 0.5536454319953918, 'train/grad_norm_pre_clip': 15.224435806274414, 'timing/data': 0.011138337664306164, 'timing/model': 0.9418843919411302, 'learning_rate/qwen_vl_interface': 7.104645596719445e-06, 'learning_rate/action_model': 7.104645596719445e-05, 'epoch': 1.1}) +[2026-06-17 20:58:01,494][starVLA.training.train_starvla][INFO] - Step 2670, Loss: {'train/loss': 0.5624261498451233, 'train/grad_norm_pre_clip': 15.751809120178223, 'timing/data': 0.014367090538144112, 'timing/model': 0.9457876095548272, 'learning_rate/qwen_vl_interface': 7.102654380959063e-06, 'learning_rate/action_model': 7.102654380959063e-05, 'epoch': 1.1}) +[2026-06-17 20:58:05,297][starVLA.training.train_starvla][INFO] - Step 2671, Loss: {'train/loss': 0.7100566029548645, 'train/grad_norm_pre_clip': 20.011686325073242, 'timing/data': 0.008156301453709602, 'timing/model': 0.9461884368211031, 'learning_rate/qwen_vl_interface': 7.100662781141432e-06, 'learning_rate/action_model': 7.100662781141432e-05, 'epoch': 1.1}) +[2026-06-17 20:58:09,169][starVLA.training.train_starvla][INFO] - Step 2672, Loss: {'train/loss': 0.536772608757019, 'train/grad_norm_pre_clip': 13.018285751342773, 'timing/data': 0.01260738167911768, 'timing/model': 0.9902777150273323, 'learning_rate/qwen_vl_interface': 7.0986707976794155e-06, 'learning_rate/action_model': 7.098670797679415e-05, 'epoch': 1.1}) +[2026-06-17 20:58:13,020][starVLA.training.train_starvla][INFO] - Step 2673, Loss: {'train/loss': 0.35631635785102844, 'train/grad_norm_pre_clip': 11.482083320617676, 'timing/data': 0.008370591327548027, 'timing/model': 1.1188108855858445, 'learning_rate/qwen_vl_interface': 7.096678430985955e-06, 'learning_rate/action_model': 7.096678430985954e-05, 'epoch': 1.1}) +[2026-06-17 20:58:16,843][starVLA.training.train_starvla][INFO] - Step 2674, Loss: {'train/loss': 0.33612680435180664, 'train/grad_norm_pre_clip': 9.491063117980957, 'timing/data': 0.0066274311393499374, 'timing/model': 1.1062348866835237, 'learning_rate/qwen_vl_interface': 7.094685681474066e-06, 'learning_rate/action_model': 7.094685681474065e-05, 'epoch': 1.1}) +[2026-06-17 20:58:20,508][starVLA.training.train_starvla][INFO] - Step 2675, Loss: {'train/loss': 0.3496558666229248, 'train/grad_norm_pre_clip': 13.812936782836914, 'timing/data': 0.016493397764861584, 'timing/model': 0.9438722124323249, 'learning_rate/qwen_vl_interface': 7.092692549556853e-06, 'learning_rate/action_model': 7.092692549556853e-05, 'epoch': 1.1}) +[2026-06-17 20:58:24,356][starVLA.training.train_starvla][INFO] - Step 2676, Loss: {'train/loss': 0.47249072790145874, 'train/grad_norm_pre_clip': 15.111239433288574, 'timing/data': 0.016716026701033115, 'timing/model': 0.9285729490220547, 'learning_rate/qwen_vl_interface': 7.0906990356474895e-06, 'learning_rate/action_model': 7.090699035647489e-05, 'epoch': 1.1}) +[2026-06-17 20:58:28,141][starVLA.training.train_starvla][INFO] - Step 2677, Loss: {'train/loss': 0.3134465515613556, 'train/grad_norm_pre_clip': 13.290312767028809, 'timing/data': 0.009247926995158195, 'timing/model': 0.9373295120894909, 'learning_rate/qwen_vl_interface': 7.088705140159238e-06, 'learning_rate/action_model': 7.088705140159237e-05, 'epoch': 1.1}) +[2026-06-17 20:58:31,922][starVLA.training.train_starvla][INFO] - Step 2678, Loss: {'train/loss': 0.7006435394287109, 'train/grad_norm_pre_clip': 16.652118682861328, 'timing/data': 0.00900995172560215, 'timing/model': 0.9399806866422296, 'learning_rate/qwen_vl_interface': 7.086710863505428e-06, 'learning_rate/action_model': 7.086710863505428e-05, 'epoch': 1.1}) +[2026-06-17 20:58:35,840][starVLA.training.train_starvla][INFO] - Step 2679, Loss: {'train/loss': 0.9442158341407776, 'train/grad_norm_pre_clip': 15.226632118225098, 'timing/data': 0.008186226710677147, 'timing/model': 0.9320931667461991, 'learning_rate/qwen_vl_interface': 7.084716206099485e-06, 'learning_rate/action_model': 7.084716206099484e-05, 'epoch': 1.1}) +[2026-06-17 20:58:39,747][starVLA.training.train_starvla][INFO] - Step 2680, Loss: {'train/loss': 0.6547375917434692, 'train/grad_norm_pre_clip': 15.765892028808594, 'timing/data': 0.01139084342867136, 'timing/model': 1.1165153812617064, 'learning_rate/qwen_vl_interface': 7.082721168354894e-06, 'learning_rate/action_model': 7.082721168354894e-05, 'epoch': 1.1}) +[2026-06-17 20:58:43,538][starVLA.training.train_starvla][INFO] - Step 2681, Loss: {'train/loss': 0.6161041855812073, 'train/grad_norm_pre_clip': 15.752046585083008, 'timing/data': 0.011812679469585419, 'timing/model': 0.9566255901008844, 'learning_rate/qwen_vl_interface': 7.080725750685235e-06, 'learning_rate/action_model': 7.080725750685234e-05, 'epoch': 1.1}) +[2026-06-17 20:58:47,359][starVLA.training.train_starvla][INFO] - Step 2682, Loss: {'train/loss': 0.5456626415252686, 'train/grad_norm_pre_clip': 12.670323371887207, 'timing/data': 0.012771892361342907, 'timing/model': 0.939504967071116, 'learning_rate/qwen_vl_interface': 7.078729953504157e-06, 'learning_rate/action_model': 7.078729953504156e-05, 'epoch': 1.1}) +[2026-06-17 20:58:51,134][starVLA.training.train_starvla][INFO] - Step 2683, Loss: {'train/loss': 0.5072819590568542, 'train/grad_norm_pre_clip': 16.699268341064453, 'timing/data': 0.009241580031812191, 'timing/model': 0.9434618409723043, 'learning_rate/qwen_vl_interface': 7.076733777225391e-06, 'learning_rate/action_model': 7.07673377722539e-05, 'epoch': 1.1}) +[2026-06-17 20:58:55,033][starVLA.training.train_starvla][INFO] - Step 2684, Loss: {'train/loss': 0.561782956123352, 'train/grad_norm_pre_clip': 13.433502197265625, 'timing/data': 0.012911218218505383, 'timing/model': 0.94947275146842, 'learning_rate/qwen_vl_interface': 7.074737222262749e-06, 'learning_rate/action_model': 7.074737222262748e-05, 'epoch': 1.1}) +[2026-06-17 20:58:58,944][starVLA.training.train_starvla][INFO] - Step 2685, Loss: {'train/loss': 0.32427462935447693, 'train/grad_norm_pre_clip': 12.64307975769043, 'timing/data': 0.009423505514860153, 'timing/model': 0.9315455825999379, 'learning_rate/qwen_vl_interface': 7.072740289030115e-06, 'learning_rate/action_model': 7.072740289030115e-05, 'epoch': 1.11}) +[2026-06-17 20:59:02,784][starVLA.training.train_starvla][INFO] - Step 2686, Loss: {'train/loss': 0.5998716950416565, 'train/grad_norm_pre_clip': 27.070524215698242, 'timing/data': 0.009708702564239502, 'timing/model': 0.9465502230450511, 'learning_rate/qwen_vl_interface': 7.07074297794146e-06, 'learning_rate/action_model': 7.07074297794146e-05, 'epoch': 1.11}) +[2026-06-17 20:59:06,630][starVLA.training.train_starvla][INFO] - Step 2687, Loss: {'train/loss': 0.24774512648582458, 'train/grad_norm_pre_clip': 9.653600692749023, 'timing/data': 0.011080434545874596, 'timing/model': 1.1166400462388992, 'learning_rate/qwen_vl_interface': 7.068745289410826e-06, 'learning_rate/action_model': 7.068745289410827e-05, 'epoch': 1.11}) +[2026-06-17 20:59:10,332][starVLA.training.train_starvla][INFO] - Step 2688, Loss: {'train/loss': 0.7930959463119507, 'train/grad_norm_pre_clip': 17.17305564880371, 'timing/data': 0.008901799097657204, 'timing/model': 0.9784939577803016, 'learning_rate/qwen_vl_interface': 7.066747223852337e-06, 'learning_rate/action_model': 7.066747223852337e-05, 'epoch': 1.11}) +[2026-06-17 20:59:14,133][starVLA.training.train_starvla][INFO] - Step 2689, Loss: {'train/loss': 0.7914638519287109, 'train/grad_norm_pre_clip': 18.123064041137695, 'timing/data': 0.01336503867059946, 'timing/model': 0.9514880012720823, 'learning_rate/qwen_vl_interface': 7.064748781680192e-06, 'learning_rate/action_model': 7.064748781680192e-05, 'epoch': 1.11}) +[2026-06-17 20:59:17,959][starVLA.training.train_starvla][INFO] - Step 2690, Loss: {'train/loss': 0.7020682692527771, 'train/grad_norm_pre_clip': 17.050947189331055, 'timing/data': 0.0073663173243403435, 'timing/model': 0.9420415349304676, 'learning_rate/qwen_vl_interface': 7.062749963308673e-06, 'learning_rate/action_model': 7.062749963308672e-05, 'epoch': 1.11}) +[2026-06-17 20:59:21,843][starVLA.training.train_starvla][INFO] - Step 2691, Loss: {'train/loss': 0.7573898434638977, 'train/grad_norm_pre_clip': 15.64720344543457, 'timing/data': 0.007843442261219025, 'timing/model': 0.9465933088213205, 'learning_rate/qwen_vl_interface': 7.060750769152135e-06, 'learning_rate/action_model': 7.060750769152135e-05, 'epoch': 1.11}) +[2026-06-17 20:59:25,738][starVLA.training.train_starvla][INFO] - Step 2692, Loss: {'train/loss': 0.4310973286628723, 'train/grad_norm_pre_clip': 12.371410369873047, 'timing/data': 0.00861366931349039, 'timing/model': 0.9440887020900846, 'learning_rate/qwen_vl_interface': 7.058751199625014e-06, 'learning_rate/action_model': 7.058751199625015e-05, 'epoch': 1.11}) +[2026-06-17 20:59:29,568][starVLA.training.train_starvla][INFO] - Step 2693, Loss: {'train/loss': 0.4429813027381897, 'train/grad_norm_pre_clip': 12.569934844970703, 'timing/data': 0.011516405269503593, 'timing/model': 1.134569882415235, 'learning_rate/qwen_vl_interface': 7.056751255141824e-06, 'learning_rate/action_model': 7.056751255141824e-05, 'epoch': 1.11}) +[2026-06-17 20:59:33,461][starVLA.training.train_starvla][INFO] - Step 2694, Loss: {'train/loss': 0.4870467483997345, 'train/grad_norm_pre_clip': 14.181534767150879, 'timing/data': 0.007774055004119873, 'timing/model': 1.1672579580917954, 'learning_rate/qwen_vl_interface': 7.054750936117155e-06, 'learning_rate/action_model': 7.054750936117156e-05, 'epoch': 1.11}) +[2026-06-17 20:59:37,147][starVLA.training.train_starvla][INFO] - Step 2695, Loss: {'train/loss': 0.580284595489502, 'train/grad_norm_pre_clip': 14.99917984008789, 'timing/data': 0.008717075921595097, 'timing/model': 0.9370260229334235, 'learning_rate/qwen_vl_interface': 7.0527502429656745e-06, 'learning_rate/action_model': 7.052750242965674e-05, 'epoch': 1.11}) +[2026-06-17 20:59:41,201][starVLA.training.train_starvla][INFO] - Step 2696, Loss: {'train/loss': 0.38575416803359985, 'train/grad_norm_pre_clip': 12.310787200927734, 'timing/data': 0.01224917359650135, 'timing/model': 1.0222899010404944, 'learning_rate/qwen_vl_interface': 7.050749176102128e-06, 'learning_rate/action_model': 7.050749176102129e-05, 'epoch': 1.11}) +[2026-06-17 20:59:45,042][starVLA.training.train_starvla][INFO] - Step 2697, Loss: {'train/loss': 0.8158658146858215, 'train/grad_norm_pre_clip': 18.065933227539062, 'timing/data': 0.00811677984893322, 'timing/model': 0.9382745223119855, 'learning_rate/qwen_vl_interface': 7.048747735941342e-06, 'learning_rate/action_model': 7.048747735941341e-05, 'epoch': 1.11}) +[2026-06-17 20:59:48,840][starVLA.training.train_starvla][INFO] - Step 2698, Loss: {'train/loss': 0.4902063012123108, 'train/grad_norm_pre_clip': 11.723286628723145, 'timing/data': 0.00794224627315998, 'timing/model': 0.9405945185571909, 'learning_rate/qwen_vl_interface': 7.046745922898212e-06, 'learning_rate/action_model': 7.046745922898211e-05, 'epoch': 1.11}) +[2026-06-17 20:59:52,711][starVLA.training.train_starvla][INFO] - Step 2699, Loss: {'train/loss': 0.37171635031700134, 'train/grad_norm_pre_clip': 12.939224243164062, 'timing/data': 0.008619148284196854, 'timing/model': 0.9689206006005406, 'learning_rate/qwen_vl_interface': 7.04474373738772e-06, 'learning_rate/action_model': 7.04474373738772e-05, 'epoch': 1.11}) +[2026-06-17 21:02:04,557][starVLA.training.train_starvla][INFO] - Step 2700, Loss: {'train/loss': 0.5049605369567871, 'train/grad_norm_pre_clip': 11.594496726989746, 'eval/loss': 0.5392148494720459, 'eval/latency_6/loss': 0.5392148494720459, 'eval/demon_fire_precision': 0.9271341463414634, 'eval/demon_fire_recall': 0.9828700711053652, 'eval/demon_fire_f1': 0.9541888923752745, 'eval/demon_fire_rate_model': 0.9126321647189761, 'eval/demon_fire_rate_teacher': 0.860879243183083, 'eval/demon_left_precision': 0.6671619613670133, 'eval/demon_left_recall': 0.787719298245614, 'eval/demon_left_f1': 0.7224456958970235, 'eval/demon_left_rate_model': 0.1872565386755704, 'eval/demon_left_rate_teacher': 0.15859766277128548, 'eval/demon_right_precision': 0.840080971659919, 'eval/demon_right_recall': 0.6374807987711214, 'eval/demon_right_f1': 0.7248908296943232, 'eval/demon_right_rate_model': 0.13745130773511408, 'eval/demon_right_rate_teacher': 0.18113522537562604, 'eval/demon_move_f1': 0.7236682627956733, 'eval/demon_cc_f1': 0.8709572676145806, 'eval/latency_6/demon_fire_precision': 0.9271341463414634, 'eval/latency_6/demon_fire_recall': 0.9828700711053652, 'eval/latency_6/demon_fire_f1': 0.9541888923752745, 'eval/latency_6/demon_fire_rate_model': 0.9126321647189761, 'eval/latency_6/demon_fire_rate_teacher': 0.860879243183083, 'eval/latency_6/demon_left_precision': 0.6671619613670133, 'eval/latency_6/demon_left_recall': 0.787719298245614, 'eval/latency_6/demon_left_f1': 0.7224456958970235, 'eval/latency_6/demon_left_rate_model': 0.1872565386755704, 'eval/latency_6/demon_left_rate_teacher': 0.15859766277128548, 'eval/latency_6/demon_right_precision': 0.840080971659919, 'eval/latency_6/demon_right_recall': 0.6374807987711214, 'eval/latency_6/demon_right_f1': 0.7248908296943232, 'eval/latency_6/demon_right_rate_model': 0.13745130773511408, 'eval/latency_6/demon_right_rate_teacher': 0.18113522537562604, 'eval/latency_6/demon_move_f1': 0.7236682627956733, 'eval/latency_6/demon_cc_f1': 0.8709572676145806, 'timing/data': 0.012108745984733105, 'timing/model': 2.5130177857354283, 'learning_rate/qwen_vl_interface': 7.042741179824919e-06, 'learning_rate/action_model': 7.042741179824918e-05, 'epoch': 1.11}) +[2026-06-17 21:02:04,562][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_2700_state +[2026-06-17 21:02:04,567][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 21:02:11,320][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_2700_state/model.safetensors +[2026-06-17 21:02:26,498][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_2700_state/optimizer.bin +[2026-06-17 21:02:26,500][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_2700_state/sampler.bin +[2026-06-17 21:02:26,500][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_2700_state/sampler_1.bin +[2026-06-17 21:02:26,502][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_2700_state/random_states_0.pkl +[2026-06-17 21:02:26,505][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2700 +[2026-06-17 21:02:26,506][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 21:02:26,517][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 21:02:30,459][starVLA.training.train_starvla][INFO] - Step 2701, Loss: {'train/loss': 0.6279919743537903, 'train/grad_norm_pre_clip': 17.869792938232422, 'timing/data': 0.009488827548921108, 'timing/model': 0.9338885499164462, 'learning_rate/qwen_vl_interface': 7.040738250624944e-06, 'learning_rate/action_model': 7.040738250624944e-05, 'epoch': 1.11}) +[2026-06-17 21:02:34,410][starVLA.training.train_starvla][INFO] - Step 2702, Loss: {'train/loss': 0.3915930986404419, 'train/grad_norm_pre_clip': 12.980097770690918, 'timing/data': 0.009490682743489742, 'timing/model': 1.0274949576705694, 'learning_rate/qwen_vl_interface': 7.038734950203001e-06, 'learning_rate/action_model': 7.038734950203001e-05, 'epoch': 1.11}) +[2026-06-17 21:02:39,708][starVLA.training.train_starvla][INFO] - Step 2703, Loss: {'train/loss': 0.5463554263114929, 'train/grad_norm_pre_clip': 12.700937271118164, 'timing/data': 0.008108040317893028, 'timing/model': 1.2523645563051105, 'learning_rate/qwen_vl_interface': 7.036731278974379e-06, 'learning_rate/action_model': 7.036731278974378e-05, 'epoch': 1.11}) +[2026-06-17 21:02:43,444][starVLA.training.train_starvla][INFO] - Step 2704, Loss: {'train/loss': 0.24722756445407867, 'train/grad_norm_pre_clip': 8.20485782623291, 'timing/data': 0.015153389424085617, 'timing/model': 0.996272886171937, 'learning_rate/qwen_vl_interface': 7.034727237354442e-06, 'learning_rate/action_model': 7.034727237354441e-05, 'epoch': 1.11}) +[2026-06-17 21:02:48,014][starVLA.training.train_starvla][INFO] - Step 2705, Loss: {'train/loss': 0.5056607723236084, 'train/grad_norm_pre_clip': 13.105181694030762, 'timing/data': 0.013471906073391438, 'timing/model': 1.6362945009022951, 'learning_rate/qwen_vl_interface': 7.0327228257586244e-06, 'learning_rate/action_model': 7.032722825758625e-05, 'epoch': 1.11}) +[2026-06-17 21:02:51,876][starVLA.training.train_starvla][INFO] - Step 2706, Loss: {'train/loss': 0.3640364408493042, 'train/grad_norm_pre_clip': 11.351690292358398, 'timing/data': 0.010927863419055939, 'timing/model': 0.966712643392384, 'learning_rate/qwen_vl_interface': 7.030718044602451e-06, 'learning_rate/action_model': 7.03071804460245e-05, 'epoch': 1.11}) +[2026-06-17 21:02:55,765][starVLA.training.train_starvla][INFO] - Step 2707, Loss: {'train/loss': 0.30211248993873596, 'train/grad_norm_pre_clip': 9.70699405670166, 'timing/data': 0.008131026290357113, 'timing/model': 0.9570490894839168, 'learning_rate/qwen_vl_interface': 7.028712894301507e-06, 'learning_rate/action_model': 7.028712894301506e-05, 'epoch': 1.11}) +[2026-06-17 21:02:59,622][starVLA.training.train_starvla][INFO] - Step 2708, Loss: {'train/loss': 0.643736720085144, 'train/grad_norm_pre_clip': 15.11912727355957, 'timing/data': 0.011246341280639172, 'timing/model': 0.953864986076951, 'learning_rate/qwen_vl_interface': 7.026707375271469e-06, 'learning_rate/action_model': 7.026707375271468e-05, 'epoch': 1.11}) +[2026-06-17 21:03:03,487][starVLA.training.train_starvla][INFO] - Step 2709, Loss: {'train/loss': 0.5259268879890442, 'train/grad_norm_pre_clip': 14.332551956176758, 'timing/data': 0.008837884292006493, 'timing/model': 1.140005523338914, 'learning_rate/qwen_vl_interface': 7.024701487928079e-06, 'learning_rate/action_model': 7.024701487928078e-05, 'epoch': 1.12}) +[2026-06-17 21:03:07,312][starVLA.training.train_starvla][INFO] - Step 2710, Loss: {'train/loss': 0.46181777119636536, 'train/grad_norm_pre_clip': 14.931375503540039, 'timing/data': 0.007849674671888351, 'timing/model': 1.1125812763348222, 'learning_rate/qwen_vl_interface': 7.02269523268716e-06, 'learning_rate/action_model': 7.022695232687159e-05, 'epoch': 1.12}) +[2026-06-17 21:03:10,941][starVLA.training.train_starvla][INFO] - Step 2711, Loss: {'train/loss': 0.6133574843406677, 'train/grad_norm_pre_clip': 19.685285568237305, 'timing/data': 0.008054656907916069, 'timing/model': 0.9466421194374561, 'learning_rate/qwen_vl_interface': 7.020688609964612e-06, 'learning_rate/action_model': 7.020688609964612e-05, 'epoch': 1.12}) +[2026-06-17 21:03:14,857][starVLA.training.train_starvla][INFO] - Step 2712, Loss: {'train/loss': 0.733284592628479, 'train/grad_norm_pre_clip': 17.122024536132812, 'timing/data': 0.009910761378705502, 'timing/model': 0.9558539418503642, 'learning_rate/qwen_vl_interface': 7.018681620176412e-06, 'learning_rate/action_model': 7.018681620176412e-05, 'epoch': 1.12}) +[2026-06-17 21:03:18,736][starVLA.training.train_starvla][INFO] - Step 2713, Loss: {'train/loss': 0.4682603180408478, 'train/grad_norm_pre_clip': 12.260610580444336, 'timing/data': 0.008098883554339409, 'timing/model': 0.9379999917000532, 'learning_rate/qwen_vl_interface': 7.016674263738606e-06, 'learning_rate/action_model': 7.016674263738607e-05, 'epoch': 1.12}) +[2026-06-17 21:03:22,548][starVLA.training.train_starvla][INFO] - Step 2714, Loss: {'train/loss': 0.6985719799995422, 'train/grad_norm_pre_clip': 19.830049514770508, 'timing/data': 0.010972727090120316, 'timing/model': 0.9648084174841642, 'learning_rate/qwen_vl_interface': 7.014666541067325e-06, 'learning_rate/action_model': 7.014666541067325e-05, 'epoch': 1.12}) +[2026-06-17 21:03:26,444][starVLA.training.train_starvla][INFO] - Step 2715, Loss: {'train/loss': 0.6154729127883911, 'train/grad_norm_pre_clip': 11.246760368347168, 'timing/data': 0.010213006287813187, 'timing/model': 0.9386123409494758, 'learning_rate/qwen_vl_interface': 7.012658452578774e-06, 'learning_rate/action_model': 7.012658452578773e-05, 'epoch': 1.12}) +[2026-06-17 21:03:30,324][starVLA.training.train_starvla][INFO] - Step 2716, Loss: {'train/loss': 0.6461266279220581, 'train/grad_norm_pre_clip': 16.338390350341797, 'timing/data': 0.009446853771805763, 'timing/model': 1.1285194354131818, 'learning_rate/qwen_vl_interface': 7.010649998689225e-06, 'learning_rate/action_model': 7.010649998689224e-05, 'epoch': 1.12}) +[2026-06-17 21:03:34,005][starVLA.training.train_starvla][INFO] - Step 2717, Loss: {'train/loss': 0.430753231048584, 'train/grad_norm_pre_clip': 13.283415794372559, 'timing/data': 0.007976507768034935, 'timing/model': 0.9494041707366705, 'learning_rate/qwen_vl_interface': 7.0086411798150385e-06, 'learning_rate/action_model': 7.008641179815038e-05, 'epoch': 1.12}) +[2026-06-17 21:03:37,843][starVLA.training.train_starvla][INFO] - Step 2718, Loss: {'train/loss': 0.20275309681892395, 'train/grad_norm_pre_clip': 8.690258979797363, 'timing/data': 0.011577827855944633, 'timing/model': 0.950675287283957, 'learning_rate/qwen_vl_interface': 7.006631996372642e-06, 'learning_rate/action_model': 7.006631996372643e-05, 'epoch': 1.12}) +[2026-06-17 21:03:41,685][starVLA.training.train_starvla][INFO] - Step 2719, Loss: {'train/loss': 0.3217296898365021, 'train/grad_norm_pre_clip': 8.564326286315918, 'timing/data': 0.013290884904563427, 'timing/model': 0.99847995582968, 'learning_rate/qwen_vl_interface': 7.004622448778545e-06, 'learning_rate/action_model': 7.004622448778544e-05, 'epoch': 1.12}) +[2026-06-17 21:03:45,603][starVLA.training.train_starvla][INFO] - Step 2720, Loss: {'train/loss': 0.5418648719787598, 'train/grad_norm_pre_clip': 18.570322036743164, 'timing/data': 0.016336766071617603, 'timing/model': 0.9544983003288507, 'learning_rate/qwen_vl_interface': 7.0026125374493244e-06, 'learning_rate/action_model': 7.002612537449324e-05, 'epoch': 1.12}) +[2026-06-17 21:03:49,430][starVLA.training.train_starvla][INFO] - Step 2721, Loss: {'train/loss': 0.3129581809043884, 'train/grad_norm_pre_clip': 13.157129287719727, 'timing/data': 0.009236647747457027, 'timing/model': 0.9630771642550826, 'learning_rate/qwen_vl_interface': 7.000602262801639e-06, 'learning_rate/action_model': 7.000602262801638e-05, 'epoch': 1.12}) +[2026-06-17 21:03:53,263][starVLA.training.train_starvla][INFO] - Step 2722, Loss: {'train/loss': 0.5743989944458008, 'train/grad_norm_pre_clip': 15.1569242477417, 'timing/data': 0.007320373319089413, 'timing/model': 0.9295907886698842, 'learning_rate/qwen_vl_interface': 6.9985916252522225e-06, 'learning_rate/action_model': 6.998591625252222e-05, 'epoch': 1.12}) +[2026-06-17 21:03:57,153][starVLA.training.train_starvla][INFO] - Step 2723, Loss: {'train/loss': 0.467185378074646, 'train/grad_norm_pre_clip': 11.903372764587402, 'timing/data': 0.010861341841518879, 'timing/model': 1.1613992936909199, 'learning_rate/qwen_vl_interface': 6.9965806252178785e-06, 'learning_rate/action_model': 6.996580625217877e-05, 'epoch': 1.12}) +[2026-06-17 21:04:00,972][starVLA.training.train_starvla][INFO] - Step 2724, Loss: {'train/loss': 0.6104596257209778, 'train/grad_norm_pre_clip': 15.43928337097168, 'timing/data': 0.009720755741000175, 'timing/model': 0.9367614947259426, 'learning_rate/qwen_vl_interface': 6.994569263115491e-06, 'learning_rate/action_model': 6.994569263115491e-05, 'epoch': 1.12}) +[2026-06-17 21:04:04,926][starVLA.training.train_starvla][INFO] - Step 2725, Loss: {'train/loss': 0.4103904664516449, 'train/grad_norm_pre_clip': 14.57745361328125, 'timing/data': 0.009675329551100731, 'timing/model': 0.9580521183088422, 'learning_rate/qwen_vl_interface': 6.992557539362019e-06, 'learning_rate/action_model': 6.99255753936202e-05, 'epoch': 1.12}) +[2026-06-17 21:04:08,869][starVLA.training.train_starvla][INFO] - Step 2726, Loss: {'train/loss': 0.4722016453742981, 'train/grad_norm_pre_clip': 13.412215232849121, 'timing/data': 0.009469793178141117, 'timing/model': 0.938619033433497, 'learning_rate/qwen_vl_interface': 6.990545454374495e-06, 'learning_rate/action_model': 6.990545454374494e-05, 'epoch': 1.12}) +[2026-06-17 21:04:12,805][starVLA.training.train_starvla][INFO] - Step 2727, Loss: {'train/loss': 0.8705394864082336, 'train/grad_norm_pre_clip': 15.752523422241211, 'timing/data': 0.012990524061024189, 'timing/model': 0.9518103608861566, 'learning_rate/qwen_vl_interface': 6.988533008570024e-06, 'learning_rate/action_model': 6.988533008570024e-05, 'epoch': 1.12}) +[2026-06-17 21:04:17,517][starVLA.training.train_starvla][INFO] - Step 2728, Loss: {'train/loss': 0.6511788368225098, 'train/grad_norm_pre_clip': 14.402941703796387, 'timing/data': 0.013944740407168865, 'timing/model': 1.840512415394187, 'learning_rate/qwen_vl_interface': 6.98652020236579e-06, 'learning_rate/action_model': 6.986520202365789e-05, 'epoch': 1.12}) +[2026-06-17 21:04:21,410][starVLA.training.train_starvla][INFO] - Step 2729, Loss: {'train/loss': 0.4283192753791809, 'train/grad_norm_pre_clip': 12.565753936767578, 'timing/data': 0.012357302010059357, 'timing/model': 1.1951934471726418, 'learning_rate/qwen_vl_interface': 6.98450703617905e-06, 'learning_rate/action_model': 6.98450703617905e-05, 'epoch': 1.12}) +[2026-06-17 21:04:25,400][starVLA.training.train_starvla][INFO] - Step 2730, Loss: {'train/loss': 0.535971999168396, 'train/grad_norm_pre_clip': 14.52122974395752, 'timing/data': 0.015858846716582775, 'timing/model': 1.2480354188010097, 'learning_rate/qwen_vl_interface': 6.982493510427134e-06, 'learning_rate/action_model': 6.982493510427133e-05, 'epoch': 1.12}) +[2026-06-17 21:04:29,025][starVLA.training.train_starvla][INFO] - Step 2731, Loss: {'train/loss': 0.43960586190223694, 'train/grad_norm_pre_clip': 14.24446964263916, 'timing/data': 0.009394761174917221, 'timing/model': 0.9375612707808614, 'learning_rate/qwen_vl_interface': 6.980479625527449e-06, 'learning_rate/action_model': 6.980479625527448e-05, 'epoch': 1.12}) +[2026-06-17 21:04:32,841][starVLA.training.train_starvla][INFO] - Step 2732, Loss: {'train/loss': 0.2961767315864563, 'train/grad_norm_pre_clip': 9.199909210205078, 'timing/data': 0.01405581459403038, 'timing/model': 0.9438209002837539, 'learning_rate/qwen_vl_interface': 6.978465381897477e-06, 'learning_rate/action_model': 6.978465381897477e-05, 'epoch': 1.12}) +[2026-06-17 21:04:36,782][starVLA.training.train_starvla][INFO] - Step 2733, Loss: {'train/loss': 0.5311141014099121, 'train/grad_norm_pre_clip': 16.03943634033203, 'timing/data': 0.008321081288158894, 'timing/model': 1.0187466526404023, 'learning_rate/qwen_vl_interface': 6.97645077995477e-06, 'learning_rate/action_model': 6.97645077995477e-05, 'epoch': 1.13}) +[2026-06-17 21:04:40,621][starVLA.training.train_starvla][INFO] - Step 2734, Loss: {'train/loss': 0.5761635303497314, 'train/grad_norm_pre_clip': 15.642821311950684, 'timing/data': 0.014620487578213215, 'timing/model': 0.9521376192569733, 'learning_rate/qwen_vl_interface': 6.97443582011696e-06, 'learning_rate/action_model': 6.974435820116959e-05, 'epoch': 1.13}) +[2026-06-17 21:04:45,617][starVLA.training.train_starvla][INFO] - Step 2735, Loss: {'train/loss': 0.39562296867370605, 'train/grad_norm_pre_clip': 10.943021774291992, 'timing/data': 0.00902566034346819, 'timing/model': 0.956937875598669, 'learning_rate/qwen_vl_interface': 6.972420502801748e-06, 'learning_rate/action_model': 6.972420502801747e-05, 'epoch': 1.13}) +[2026-06-17 21:04:50,659][starVLA.training.train_starvla][INFO] - Step 2736, Loss: {'train/loss': 0.3649344742298126, 'train/grad_norm_pre_clip': 10.164204597473145, 'timing/data': 0.009905708022415638, 'timing/model': 1.166736806742847, 'learning_rate/qwen_vl_interface': 6.970404828426911e-06, 'learning_rate/action_model': 6.97040482842691e-05, 'epoch': 1.13}) +[2026-06-17 21:04:55,335][starVLA.training.train_starvla][INFO] - Step 2737, Loss: {'train/loss': 0.5037146806716919, 'train/grad_norm_pre_clip': 15.917962074279785, 'timing/data': 0.01857653073966503, 'timing/model': 1.1076261140406132, 'learning_rate/qwen_vl_interface': 6.9683887974103044e-06, 'learning_rate/action_model': 6.968388797410305e-05, 'epoch': 1.13}) +[2026-06-17 21:04:59,096][starVLA.training.train_starvla][INFO] - Step 2738, Loss: {'train/loss': 0.5339922904968262, 'train/grad_norm_pre_clip': 20.627099990844727, 'timing/data': 0.008668024092912674, 'timing/model': 0.9489352442324162, 'learning_rate/qwen_vl_interface': 6.966372410169848e-06, 'learning_rate/action_model': 6.966372410169848e-05, 'epoch': 1.13}) +[2026-06-17 21:05:02,959][starVLA.training.train_starvla][INFO] - Step 2739, Loss: {'train/loss': 0.4220905601978302, 'train/grad_norm_pre_clip': 13.161906242370605, 'timing/data': 0.009534758515655994, 'timing/model': 0.9479346796870232, 'learning_rate/qwen_vl_interface': 6.964355667123548e-06, 'learning_rate/action_model': 6.964355667123547e-05, 'epoch': 1.13}) +[2026-06-17 21:05:07,159][starVLA.training.train_starvla][INFO] - Step 2740, Loss: {'train/loss': 0.5346229076385498, 'train/grad_norm_pre_clip': 14.546857833862305, 'timing/data': 0.008478264324367046, 'timing/model': 0.9471490401774645, 'learning_rate/qwen_vl_interface': 6.96233856868947e-06, 'learning_rate/action_model': 6.962338568689469e-05, 'epoch': 1.13}) +[2026-06-17 21:05:11,591][starVLA.training.train_starvla][INFO] - Step 2741, Loss: {'train/loss': 0.4310263395309448, 'train/grad_norm_pre_clip': 65.98021697998047, 'timing/data': 0.012470738030970097, 'timing/model': 1.4248132379725575, 'learning_rate/qwen_vl_interface': 6.960321115285767e-06, 'learning_rate/action_model': 6.960321115285767e-05, 'epoch': 1.13}) +[2026-06-17 21:05:15,414][starVLA.training.train_starvla][INFO] - Step 2742, Loss: {'train/loss': 0.4381304681301117, 'train/grad_norm_pre_clip': 12.623966217041016, 'timing/data': 0.009718396700918674, 'timing/model': 0.9624414155259728, 'learning_rate/qwen_vl_interface': 6.958303307330655e-06, 'learning_rate/action_model': 6.958303307330655e-05, 'epoch': 1.13}) +[2026-06-17 21:05:19,243][starVLA.training.train_starvla][INFO] - Step 2743, Loss: {'train/loss': 0.270672082901001, 'train/grad_norm_pre_clip': 9.702361106872559, 'timing/data': 0.0103523014113307, 'timing/model': 1.154061671346426, 'learning_rate/qwen_vl_interface': 6.956285145242429e-06, 'learning_rate/action_model': 6.956285145242428e-05, 'epoch': 1.13}) +[2026-06-17 21:05:22,923][starVLA.training.train_starvla][INFO] - Step 2744, Loss: {'train/loss': 0.5925761461257935, 'train/grad_norm_pre_clip': 15.773521423339844, 'timing/data': 0.007701519876718521, 'timing/model': 0.9518896657973528, 'learning_rate/qwen_vl_interface': 6.954266629439456e-06, 'learning_rate/action_model': 6.954266629439457e-05, 'epoch': 1.13}) +[2026-06-17 21:05:26,804][starVLA.training.train_starvla][INFO] - Step 2745, Loss: {'train/loss': 0.3064122200012207, 'train/grad_norm_pre_clip': 10.793279647827148, 'timing/data': 0.013750802725553513, 'timing/model': 0.951341581530869, 'learning_rate/qwen_vl_interface': 6.952247760340176e-06, 'learning_rate/action_model': 6.952247760340176e-05, 'epoch': 1.13}) +[2026-06-17 21:05:30,644][starVLA.training.train_starvla][INFO] - Step 2746, Loss: {'train/loss': 0.2488202452659607, 'train/grad_norm_pre_clip': 10.404622077941895, 'timing/data': 0.011710962280631065, 'timing/model': 0.9352919468656182, 'learning_rate/qwen_vl_interface': 6.950228538363108e-06, 'learning_rate/action_model': 6.950228538363107e-05, 'epoch': 1.13}) +[2026-06-17 21:05:34,473][starVLA.training.train_starvla][INFO] - Step 2747, Loss: {'train/loss': 0.6313942670822144, 'train/grad_norm_pre_clip': 16.312366485595703, 'timing/data': 0.01318364031612873, 'timing/model': 0.9358087321743369, 'learning_rate/qwen_vl_interface': 6.948208963926829e-06, 'learning_rate/action_model': 6.94820896392683e-05, 'epoch': 1.13}) +[2026-06-17 21:05:38,849][starVLA.training.train_starvla][INFO] - Step 2748, Loss: {'train/loss': 0.6944953203201294, 'train/grad_norm_pre_clip': 13.335784912109375, 'timing/data': 0.008228417485952377, 'timing/model': 0.9531173082068563, 'learning_rate/qwen_vl_interface': 6.946189037450007e-06, 'learning_rate/action_model': 6.946189037450007e-05, 'epoch': 1.13}) +[2026-06-17 21:05:42,710][starVLA.training.train_starvla][INFO] - Step 2749, Loss: {'train/loss': 0.4772535562515259, 'train/grad_norm_pre_clip': 14.828155517578125, 'timing/data': 0.008345548994839191, 'timing/model': 1.1273047644644976, 'learning_rate/qwen_vl_interface': 6.9441687593513715e-06, 'learning_rate/action_model': 6.944168759351372e-05, 'epoch': 1.13}) +[2026-06-17 21:05:46,934][starVLA.training.train_starvla][INFO] - Step 2750, Loss: {'train/loss': 0.46377646923065186, 'train/grad_norm_pre_clip': 15.432621002197266, 'timing/data': 0.007999737747013569, 'timing/model': 1.1266559213399887, 'learning_rate/qwen_vl_interface': 6.942148130049732e-06, 'learning_rate/action_model': 6.942148130049731e-05, 'epoch': 1.13}) +[2026-06-17 21:05:50,596][starVLA.training.train_starvla][INFO] - Step 2751, Loss: {'train/loss': 0.29458826780319214, 'train/grad_norm_pre_clip': 11.70615291595459, 'timing/data': 0.01554618775844574, 'timing/model': 0.9568870877847075, 'learning_rate/qwen_vl_interface': 6.940127149963961e-06, 'learning_rate/action_model': 6.94012714996396e-05, 'epoch': 1.13}) +[2026-06-17 21:05:54,434][starVLA.training.train_starvla][INFO] - Step 2752, Loss: {'train/loss': 0.39655157923698425, 'train/grad_norm_pre_clip': 10.387289047241211, 'timing/data': 0.01795517187565565, 'timing/model': 0.942256873473525, 'learning_rate/qwen_vl_interface': 6.9381058195130146e-06, 'learning_rate/action_model': 6.938105819513014e-05, 'epoch': 1.13}) +[2026-06-17 21:05:58,256][starVLA.training.train_starvla][INFO] - Step 2753, Loss: {'train/loss': 0.2647649645805359, 'train/grad_norm_pre_clip': 9.903324127197266, 'timing/data': 0.009164408780634403, 'timing/model': 0.9457578985020518, 'learning_rate/qwen_vl_interface': 6.936084139115915e-06, 'learning_rate/action_model': 6.936084139115915e-05, 'epoch': 1.13}) +[2026-06-17 21:06:02,132][starVLA.training.train_starvla][INFO] - Step 2754, Loss: {'train/loss': 0.4725175201892853, 'train/grad_norm_pre_clip': 17.548744201660156, 'timing/data': 0.009603863582015038, 'timing/model': 0.936353144235909, 'learning_rate/qwen_vl_interface': 6.934062109191759e-06, 'learning_rate/action_model': 6.934062109191759e-05, 'epoch': 1.13}) +[2026-06-17 21:06:05,948][starVLA.training.train_starvla][INFO] - Step 2755, Loss: {'train/loss': 0.4657040238380432, 'train/grad_norm_pre_clip': 14.791050910949707, 'timing/data': 0.008053398691117764, 'timing/model': 0.9405305916443467, 'learning_rate/qwen_vl_interface': 6.9320397301597155e-06, 'learning_rate/action_model': 6.932039730159715e-05, 'epoch': 1.13}) +[2026-06-17 21:06:09,868][starVLA.training.train_starvla][INFO] - Step 2756, Loss: {'train/loss': 0.5040023326873779, 'train/grad_norm_pre_clip': 13.996115684509277, 'timing/data': 0.011187615804374218, 'timing/model': 1.1182401748374104, 'learning_rate/qwen_vl_interface': 6.930017002439027e-06, 'learning_rate/action_model': 6.930017002439027e-05, 'epoch': 1.13}) +[2026-06-17 21:06:13,712][starVLA.training.train_starvla][INFO] - Step 2757, Loss: {'train/loss': 0.5853930711746216, 'train/grad_norm_pre_clip': 16.664934158325195, 'timing/data': 0.010660930536687374, 'timing/model': 1.120021522976458, 'learning_rate/qwen_vl_interface': 6.927993926449005e-06, 'learning_rate/action_model': 6.927993926449005e-05, 'epoch': 1.13}) +[2026-06-17 21:06:17,375][starVLA.training.train_starvla][INFO] - Step 2758, Loss: {'train/loss': 0.652328372001648, 'train/grad_norm_pre_clip': 15.679666519165039, 'timing/data': 0.009986319579184055, 'timing/model': 0.9679171834141016, 'learning_rate/qwen_vl_interface': 6.925970502609036e-06, 'learning_rate/action_model': 6.925970502609036e-05, 'epoch': 1.14}) +[2026-06-17 21:06:21,269][starVLA.training.train_starvla][INFO] - Step 2759, Loss: {'train/loss': 0.27083709836006165, 'train/grad_norm_pre_clip': 11.653534889221191, 'timing/data': 0.009311896748840809, 'timing/model': 0.9391971826553345, 'learning_rate/qwen_vl_interface': 6.92394673133858e-06, 'learning_rate/action_model': 6.92394673133858e-05, 'epoch': 1.14}) +[2026-06-17 21:06:25,221][starVLA.training.train_starvla][INFO] - Step 2760, Loss: {'train/loss': 0.5825748443603516, 'train/grad_norm_pre_clip': 12.575682640075684, 'timing/data': 0.009884090162813663, 'timing/model': 0.9479102985933423, 'learning_rate/qwen_vl_interface': 6.921922613057165e-06, 'learning_rate/action_model': 6.921922613057164e-05, 'epoch': 1.14}) +[2026-06-17 21:06:29,145][starVLA.training.train_starvla][INFO] - Step 2761, Loss: {'train/loss': 0.2655249536037445, 'train/grad_norm_pre_clip': 10.597867965698242, 'timing/data': 0.013808926567435265, 'timing/model': 0.9843198759481311, 'learning_rate/qwen_vl_interface': 6.9198981481843915e-06, 'learning_rate/action_model': 6.919898148184392e-05, 'epoch': 1.14}) +[2026-06-17 21:06:33,051][starVLA.training.train_starvla][INFO] - Step 2762, Loss: {'train/loss': 0.3485337197780609, 'train/grad_norm_pre_clip': 14.143500328063965, 'timing/data': 0.010298178531229496, 'timing/model': 0.9483090005815029, 'learning_rate/qwen_vl_interface': 6.917873337139936e-06, 'learning_rate/action_model': 6.917873337139937e-05, 'epoch': 1.14}) +[2026-06-17 21:06:36,878][starVLA.training.train_starvla][INFO] - Step 2763, Loss: {'train/loss': 0.5980409979820251, 'train/grad_norm_pre_clip': 15.360118865966797, 'timing/data': 0.015409416519105434, 'timing/model': 1.118188038468361, 'learning_rate/qwen_vl_interface': 6.915848180343545e-06, 'learning_rate/action_model': 6.915848180343545e-05, 'epoch': 1.14}) +[2026-06-17 21:06:40,642][starVLA.training.train_starvla][INFO] - Step 2764, Loss: {'train/loss': 0.26361605525016785, 'train/grad_norm_pre_clip': 17.144960403442383, 'timing/data': 0.008761905133724213, 'timing/model': 0.930570712313056, 'learning_rate/qwen_vl_interface': 6.913822678215031e-06, 'learning_rate/action_model': 6.913822678215031e-05, 'epoch': 1.14}) +[2026-06-17 21:06:44,496][starVLA.training.train_starvla][INFO] - Step 2765, Loss: {'train/loss': 0.3343089520931244, 'train/grad_norm_pre_clip': 13.135348320007324, 'timing/data': 0.008562588132917881, 'timing/model': 0.9458979330956936, 'learning_rate/qwen_vl_interface': 6.911796831174287e-06, 'learning_rate/action_model': 6.911796831174286e-05, 'epoch': 1.14}) +[2026-06-17 21:06:48,637][starVLA.training.train_starvla][INFO] - Step 2766, Loss: {'train/loss': 0.6209124326705933, 'train/grad_norm_pre_clip': 16.079002380371094, 'timing/data': 0.014299733564257622, 'timing/model': 0.9381515653803945, 'learning_rate/qwen_vl_interface': 6.909770639641271e-06, 'learning_rate/action_model': 6.909770639641271e-05, 'epoch': 1.14}) +[2026-06-17 21:06:52,533][starVLA.training.train_starvla][INFO] - Step 2767, Loss: {'train/loss': 0.25394976139068604, 'train/grad_norm_pre_clip': 12.204739570617676, 'timing/data': 0.008510810323059559, 'timing/model': 0.9321679193526506, 'learning_rate/qwen_vl_interface': 6.907744104036015e-06, 'learning_rate/action_model': 6.907744104036015e-05, 'epoch': 1.14}) +[2026-06-17 21:06:56,370][starVLA.training.train_starvla][INFO] - Step 2768, Loss: {'train/loss': 0.5242547392845154, 'train/grad_norm_pre_clip': 14.223392486572266, 'timing/data': 0.009797205217182636, 'timing/model': 0.9431166462600231, 'learning_rate/qwen_vl_interface': 6.905717224778623e-06, 'learning_rate/action_model': 6.905717224778622e-05, 'epoch': 1.14}) +[2026-06-17 21:07:00,451][starVLA.training.train_starvla][INFO] - Step 2769, Loss: {'train/loss': 0.7666263580322266, 'train/grad_norm_pre_clip': 17.02672004699707, 'timing/data': 0.008615045808255672, 'timing/model': 1.0606932248920202, 'learning_rate/qwen_vl_interface': 6.903690002289268e-06, 'learning_rate/action_model': 6.903690002289267e-05, 'epoch': 1.14}) +[2026-06-17 21:07:04,503][starVLA.training.train_starvla][INFO] - Step 2770, Loss: {'train/loss': 0.43568485975265503, 'train/grad_norm_pre_clip': 13.223392486572266, 'timing/data': 0.016453931108117104, 'timing/model': 1.1479651127010584, 'learning_rate/qwen_vl_interface': 6.901662436988199e-06, 'learning_rate/action_model': 6.901662436988199e-05, 'epoch': 1.14}) +[2026-06-17 21:07:09,552][starVLA.training.train_starvla][INFO] - Step 2771, Loss: {'train/loss': 0.7522707581520081, 'train/grad_norm_pre_clip': 18.351335525512695, 'timing/data': 0.009658590890467167, 'timing/model': 1.0100718820467591, 'learning_rate/qwen_vl_interface': 6.899634529295726e-06, 'learning_rate/action_model': 6.899634529295726e-05, 'epoch': 1.14}) +[2026-06-17 21:07:14,324][starVLA.training.train_starvla][INFO] - Step 2772, Loss: {'train/loss': 0.33172982931137085, 'train/grad_norm_pre_clip': 14.27540397644043, 'timing/data': 0.010334305465221405, 'timing/model': 0.9716474218294024, 'learning_rate/qwen_vl_interface': 6.897606279632241e-06, 'learning_rate/action_model': 6.897606279632241e-05, 'epoch': 1.14}) +[2026-06-17 21:07:18,246][starVLA.training.train_starvla][INFO] - Step 2773, Loss: {'train/loss': 0.3440510332584381, 'train/grad_norm_pre_clip': 12.467267036437988, 'timing/data': 0.01317367609590292, 'timing/model': 0.932328375056386, 'learning_rate/qwen_vl_interface': 6.895577688418202e-06, 'learning_rate/action_model': 6.895577688418202e-05, 'epoch': 1.14}) +[2026-06-17 21:07:22,189][starVLA.training.train_starvla][INFO] - Step 2774, Loss: {'train/loss': 0.48191604018211365, 'train/grad_norm_pre_clip': 20.379091262817383, 'timing/data': 0.010753173381090164, 'timing/model': 0.9460310712456703, 'learning_rate/qwen_vl_interface': 6.893548756074136e-06, 'learning_rate/action_model': 6.893548756074136e-05, 'epoch': 1.14}) +[2026-06-17 21:07:26,354][starVLA.training.train_starvla][INFO] - Step 2775, Loss: {'train/loss': 0.4918588697910309, 'train/grad_norm_pre_clip': 12.288512229919434, 'timing/data': 0.015186138451099396, 'timing/model': 1.2132791625335813, 'learning_rate/qwen_vl_interface': 6.891519483020647e-06, 'learning_rate/action_model': 6.891519483020647e-05, 'epoch': 1.14}) +[2026-06-17 21:07:30,220][starVLA.training.train_starvla][INFO] - Step 2776, Loss: {'train/loss': 0.43827006220817566, 'train/grad_norm_pre_clip': 31.9364013671875, 'timing/data': 0.00925984513014555, 'timing/model': 1.1387773659080267, 'learning_rate/qwen_vl_interface': 6.8894898696784e-06, 'learning_rate/action_model': 6.889489869678399e-05, 'epoch': 1.14}) +[2026-06-17 21:07:34,053][starVLA.training.train_starvla][INFO] - Step 2777, Loss: {'train/loss': 0.2613162100315094, 'train/grad_norm_pre_clip': 45.475494384765625, 'timing/data': 0.015321540646255016, 'timing/model': 1.1026847837492824, 'learning_rate/qwen_vl_interface': 6.887459916468141e-06, 'learning_rate/action_model': 6.88745991646814e-05, 'epoch': 1.14}) +[2026-06-17 21:07:37,842][starVLA.training.train_starvla][INFO] - Step 2778, Loss: {'train/loss': 0.45147213339805603, 'train/grad_norm_pre_clip': 12.117523193359375, 'timing/data': 0.016272936016321182, 'timing/model': 0.9397737355902791, 'learning_rate/qwen_vl_interface': 6.885429623810678e-06, 'learning_rate/action_model': 6.885429623810677e-05, 'epoch': 1.14}) +[2026-06-17 21:07:42,221][starVLA.training.train_starvla][INFO] - Step 2779, Loss: {'train/loss': 0.34809133410453796, 'train/grad_norm_pre_clip': 18.67215919494629, 'timing/data': 0.015182511880993843, 'timing/model': 0.9598118485882878, 'learning_rate/qwen_vl_interface': 6.883398992126894e-06, 'learning_rate/action_model': 6.883398992126894e-05, 'epoch': 1.14}) +[2026-06-17 21:07:46,210][starVLA.training.train_starvla][INFO] - Step 2780, Loss: {'train/loss': 0.5777134895324707, 'train/grad_norm_pre_clip': 16.619050979614258, 'timing/data': 0.010544763877987862, 'timing/model': 1.031012357212603, 'learning_rate/qwen_vl_interface': 6.8813680218377425e-06, 'learning_rate/action_model': 6.881368021837741e-05, 'epoch': 1.14}) +[2026-06-17 21:07:50,120][starVLA.training.train_starvla][INFO] - Step 2781, Loss: {'train/loss': 0.5488799214363098, 'train/grad_norm_pre_clip': 14.202579498291016, 'timing/data': 0.012631470337510109, 'timing/model': 0.9492901731282473, 'learning_rate/qwen_vl_interface': 6.8793367133642445e-06, 'learning_rate/action_model': 6.879336713364243e-05, 'epoch': 1.14}) +[2026-06-17 21:07:54,049][starVLA.training.train_starvla][INFO] - Step 2782, Loss: {'train/loss': 0.382678359746933, 'train/grad_norm_pre_clip': 13.669967651367188, 'timing/data': 0.01093872357159853, 'timing/model': 0.9752865191549063, 'learning_rate/qwen_vl_interface': 6.877305067127492e-06, 'learning_rate/action_model': 6.877305067127492e-05, 'epoch': 1.15}) +[2026-06-17 21:07:57,878][starVLA.training.train_starvla][INFO] - Step 2783, Loss: {'train/loss': 0.4418010711669922, 'train/grad_norm_pre_clip': 16.6168270111084, 'timing/data': 0.012935811653733253, 'timing/model': 1.1118162162601948, 'learning_rate/qwen_vl_interface': 6.875273083548648e-06, 'learning_rate/action_model': 6.875273083548648e-05, 'epoch': 1.15}) +[2026-06-17 21:08:01,644][starVLA.training.train_starvla][INFO] - Step 2784, Loss: {'train/loss': 0.41913488507270813, 'train/grad_norm_pre_clip': 13.442895889282227, 'timing/data': 0.014088453724980354, 'timing/model': 1.0709028290584683, 'learning_rate/qwen_vl_interface': 6.873240763048945e-06, 'learning_rate/action_model': 6.873240763048945e-05, 'epoch': 1.15}) +[2026-06-17 21:08:05,282][starVLA.training.train_starvla][INFO] - Step 2785, Loss: {'train/loss': 0.2535129487514496, 'train/grad_norm_pre_clip': 13.869853973388672, 'timing/data': 0.013790618628263474, 'timing/model': 0.9423092156648636, 'learning_rate/qwen_vl_interface': 6.871208106049685e-06, 'learning_rate/action_model': 6.871208106049685e-05, 'epoch': 1.15}) +[2026-06-17 21:08:09,142][starVLA.training.train_starvla][INFO] - Step 2786, Loss: {'train/loss': 0.2958325147628784, 'train/grad_norm_pre_clip': 11.234458923339844, 'timing/data': 0.011256012134253979, 'timing/model': 0.9503028318285942, 'learning_rate/qwen_vl_interface': 6.869175112972239e-06, 'learning_rate/action_model': 6.86917511297224e-05, 'epoch': 1.15}) +[2026-06-17 21:08:12,950][starVLA.training.train_starvla][INFO] - Step 2787, Loss: {'train/loss': 0.40813320875167847, 'train/grad_norm_pre_clip': 11.080483436584473, 'timing/data': 0.008015031926333904, 'timing/model': 0.9396618464961648, 'learning_rate/qwen_vl_interface': 6.867141784238052e-06, 'learning_rate/action_model': 6.867141784238051e-05, 'epoch': 1.15}) +[2026-06-17 21:08:16,740][starVLA.training.train_starvla][INFO] - Step 2788, Loss: {'train/loss': 0.3211553990840912, 'train/grad_norm_pre_clip': 13.570402145385742, 'timing/data': 0.0077206166461110115, 'timing/model': 0.9425987293943763, 'learning_rate/qwen_vl_interface': 6.865108120268632e-06, 'learning_rate/action_model': 6.865108120268632e-05, 'epoch': 1.15}) +[2026-06-17 21:08:20,619][starVLA.training.train_starvla][INFO] - Step 2789, Loss: {'train/loss': 0.2524155378341675, 'train/grad_norm_pre_clip': 10.315844535827637, 'timing/data': 0.009343231096863747, 'timing/model': 0.9647496491670609, 'learning_rate/qwen_vl_interface': 6.863074121485561e-06, 'learning_rate/action_model': 6.86307412148556e-05, 'epoch': 1.15}) +[2026-06-17 21:08:24,501][starVLA.training.train_starvla][INFO] - Step 2790, Loss: {'train/loss': 0.30683237314224243, 'train/grad_norm_pre_clip': 9.69581127166748, 'timing/data': 0.012722081504762173, 'timing/model': 1.1550706988200545, 'learning_rate/qwen_vl_interface': 6.861039788310489e-06, 'learning_rate/action_model': 6.861039788310489e-05, 'epoch': 1.15}) +[2026-06-17 21:08:29,114][starVLA.training.train_starvla][INFO] - Step 2791, Loss: {'train/loss': 0.2648845911026001, 'train/grad_norm_pre_clip': 12.589022636413574, 'timing/data': 0.01341354101896286, 'timing/model': 1.911062704399228, 'learning_rate/qwen_vl_interface': 6.859005121165135e-06, 'learning_rate/action_model': 6.859005121165135e-05, 'epoch': 1.15}) +[2026-06-17 21:08:33,036][starVLA.training.train_starvla][INFO] - Step 2792, Loss: {'train/loss': 0.6078765988349915, 'train/grad_norm_pre_clip': 14.509445190429688, 'timing/data': 0.015154963359236717, 'timing/model': 0.9465232342481613, 'learning_rate/qwen_vl_interface': 6.856970120471289e-06, 'learning_rate/action_model': 6.856970120471289e-05, 'epoch': 1.15}) +[2026-06-17 21:08:36,871][starVLA.training.train_starvla][INFO] - Step 2793, Loss: {'train/loss': 0.4367009699344635, 'train/grad_norm_pre_clip': 18.824853897094727, 'timing/data': 0.013401600532233715, 'timing/model': 0.9795224675908685, 'learning_rate/qwen_vl_interface': 6.854934786650805e-06, 'learning_rate/action_model': 6.854934786650806e-05, 'epoch': 1.15}) +[2026-06-17 21:08:42,312][starVLA.training.train_starvla][INFO] - Step 2794, Loss: {'train/loss': 0.34844788908958435, 'train/grad_norm_pre_clip': 11.48686695098877, 'timing/data': 0.0005496200174093246, 'timing/model': 0.9742017090320587, 'learning_rate/qwen_vl_interface': 6.852899120125615e-06, 'learning_rate/action_model': 6.852899120125616e-05, 'epoch': 1.15}) +[2026-06-17 21:08:46,184][starVLA.training.train_starvla][INFO] - Step 2795, Loss: {'train/loss': 0.3936348259449005, 'train/grad_norm_pre_clip': 15.877286911010742, 'timing/data': 0.0005028983578085899, 'timing/model': 0.9558846540749073, 'learning_rate/qwen_vl_interface': 6.850863121317712e-06, 'learning_rate/action_model': 6.850863121317712e-05, 'epoch': 1.15}) +[2026-06-17 21:08:51,285][starVLA.training.train_starvla][INFO] - Step 2796, Loss: {'train/loss': 0.39123257994651794, 'train/grad_norm_pre_clip': 12.766463279724121, 'timing/data': 0.0004875408485531807, 'timing/model': 1.7272295281291008, 'learning_rate/qwen_vl_interface': 6.848826790649162e-06, 'learning_rate/action_model': 6.848826790649163e-05, 'epoch': 1.15}) +[2026-06-17 21:08:55,102][starVLA.training.train_starvla][INFO] - Step 2797, Loss: {'train/loss': 0.6824360489845276, 'train/grad_norm_pre_clip': 19.313440322875977, 'timing/data': 0.000561840832233429, 'timing/model': 1.0937138525769114, 'learning_rate/qwen_vl_interface': 6.846790128542098e-06, 'learning_rate/action_model': 6.846790128542098e-05, 'epoch': 1.15}) +[2026-06-17 21:08:58,789][starVLA.training.train_starvla][INFO] - Step 2798, Loss: {'train/loss': 0.15396660566329956, 'train/grad_norm_pre_clip': 11.347379684448242, 'timing/data': 0.000498657114803791, 'timing/model': 0.9566916935145855, 'learning_rate/qwen_vl_interface': 6.8447531354187214e-06, 'learning_rate/action_model': 6.844753135418722e-05, 'epoch': 1.15}) +[2026-06-17 21:09:02,830][starVLA.training.train_starvla][INFO] - Step 2799, Loss: {'train/loss': 0.4409857392311096, 'train/grad_norm_pre_clip': 12.905760765075684, 'timing/data': 0.0008518248796463013, 'timing/model': 0.9865861516445875, 'learning_rate/qwen_vl_interface': 6.842715811701305e-06, 'learning_rate/action_model': 6.842715811701304e-05, 'epoch': 1.15}) +[2026-06-17 21:11:18,029][starVLA.training.train_starvla][INFO] - Step 2800, Loss: {'train/loss': 0.46320220828056335, 'train/grad_norm_pre_clip': 14.894729614257812, 'eval/loss': 0.5267155766487122, 'eval/latency_6/loss': 0.5267155766487122, 'eval/demon_fire_precision': 0.9288537549407114, 'eval/demon_fire_recall': 0.9873949579831933, 'eval/demon_fire_f1': 0.9572301425661914, 'eval/demon_fire_rate_model': 0.9151363383416806, 'eval/demon_fire_rate_teacher': 0.860879243183083, 'eval/demon_left_precision': 0.7749077490774908, 'eval/demon_left_recall': 0.7368421052631579, 'eval/demon_left_f1': 0.7553956834532375, 'eval/demon_left_rate_model': 0.15080690038953812, 'eval/demon_left_rate_teacher': 0.15859766277128548, 'eval/demon_right_precision': 0.8599605522682445, 'eval/demon_right_recall': 0.6697388632872504, 'eval/demon_right_f1': 0.7530224525043179, 'eval/demon_right_rate_model': 0.14106844741235391, 'eval/demon_right_rate_teacher': 0.18113522537562604, 'eval/demon_move_f1': 0.7542090679787776, 'eval/demon_cc_f1': 0.8857048489191363, 'eval/latency_6/demon_fire_precision': 0.9288537549407114, 'eval/latency_6/demon_fire_recall': 0.9873949579831933, 'eval/latency_6/demon_fire_f1': 0.9572301425661914, 'eval/latency_6/demon_fire_rate_model': 0.9151363383416806, 'eval/latency_6/demon_fire_rate_teacher': 0.860879243183083, 'eval/latency_6/demon_left_precision': 0.7749077490774908, 'eval/latency_6/demon_left_recall': 0.7368421052631579, 'eval/latency_6/demon_left_f1': 0.7553956834532375, 'eval/latency_6/demon_left_rate_model': 0.15080690038953812, 'eval/latency_6/demon_left_rate_teacher': 0.15859766277128548, 'eval/latency_6/demon_right_precision': 0.8599605522682445, 'eval/latency_6/demon_right_recall': 0.6697388632872504, 'eval/latency_6/demon_right_f1': 0.7530224525043179, 'eval/latency_6/demon_right_rate_model': 0.14106844741235391, 'eval/latency_6/demon_right_rate_teacher': 0.18113522537562604, 'eval/latency_6/demon_move_f1': 0.7542090679787776, 'eval/latency_6/demon_cc_f1': 0.8857048489191363, 'timing/data': 0.000505576841533184, 'timing/model': 1.049967565573752, 'learning_rate/qwen_vl_interface': 6.840678157812184e-06, 'learning_rate/action_model': 6.840678157812184e-05, 'epoch': 1.15}) +[2026-06-17 21:11:18,033][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_2800_state +[2026-06-17 21:11:18,038][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 21:11:25,053][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_2800_state/model.safetensors +[2026-06-17 21:11:40,593][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_2800_state/optimizer.bin +[2026-06-17 21:11:40,594][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_2800_state/sampler.bin +[2026-06-17 21:11:40,594][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_2800_state/sampler_1.bin +[2026-06-17 21:11:40,610][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_2800_state/random_states_0.pkl +[2026-06-17 21:11:40,613][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2800 +[2026-06-17 21:11:40,614][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 21:11:40,624][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 21:11:44,512][starVLA.training.train_starvla][INFO] - Step 2801, Loss: {'train/loss': 0.6329188346862793, 'train/grad_norm_pre_clip': 17.475379943847656, 'timing/data': 0.0005379589274525642, 'timing/model': 0.9442662047222257, 'learning_rate/qwen_vl_interface': 6.8386401741737725e-06, 'learning_rate/action_model': 6.838640174173773e-05, 'epoch': 1.15}) +[2026-06-17 21:11:48,301][starVLA.training.train_starvla][INFO] - Step 2802, Loss: {'train/loss': 0.46624040603637695, 'train/grad_norm_pre_clip': 15.806324005126953, 'timing/data': 0.0004983572289347649, 'timing/model': 0.9377516098320484, 'learning_rate/qwen_vl_interface': 6.836601861208542e-06, 'learning_rate/action_model': 6.836601861208541e-05, 'epoch': 1.15}) +[2026-06-17 21:11:52,297][starVLA.training.train_starvla][INFO] - Step 2803, Loss: {'train/loss': 0.3770565390586853, 'train/grad_norm_pre_clip': 12.381306648254395, 'timing/data': 0.0005446337163448334, 'timing/model': 0.9345337580889463, 'learning_rate/qwen_vl_interface': 6.834563219339039e-06, 'learning_rate/action_model': 6.834563219339038e-05, 'epoch': 1.15}) +[2026-06-17 21:11:56,235][starVLA.training.train_starvla][INFO] - Step 2804, Loss: {'train/loss': 0.21845877170562744, 'train/grad_norm_pre_clip': 7.949276924133301, 'timing/data': 0.0010589761659502983, 'timing/model': 0.9425554368644953, 'learning_rate/qwen_vl_interface': 6.832524248987874e-06, 'learning_rate/action_model': 6.832524248987874e-05, 'epoch': 1.15}) +[2026-06-17 21:12:00,179][starVLA.training.train_starvla][INFO] - Step 2805, Loss: {'train/loss': 0.4491962790489197, 'train/grad_norm_pre_clip': 14.281288146972656, 'timing/data': 0.0006498768925666809, 'timing/model': 1.1305311592295766, 'learning_rate/qwen_vl_interface': 6.8304849505777314e-06, 'learning_rate/action_model': 6.830484950577732e-05, 'epoch': 1.15}) +[2026-06-17 21:12:04,029][starVLA.training.train_starvla][INFO] - Step 2806, Loss: {'train/loss': 0.5527629256248474, 'train/grad_norm_pre_clip': 24.337759017944336, 'timing/data': 0.0005866987630724907, 'timing/model': 1.0631679873913527, 'learning_rate/qwen_vl_interface': 6.828445324531357e-06, 'learning_rate/action_model': 6.828445324531357e-05, 'epoch': 1.16}) +[2026-06-17 21:12:07,758][starVLA.training.train_starvla][INFO] - Step 2807, Loss: {'train/loss': 0.7225316762924194, 'train/grad_norm_pre_clip': 17.72164535522461, 'timing/data': 0.0005271192640066147, 'timing/model': 0.9401480043306947, 'learning_rate/qwen_vl_interface': 6.826405371271567e-06, 'learning_rate/action_model': 6.826405371271567e-05, 'epoch': 1.16}) +[2026-06-17 21:12:11,578][starVLA.training.train_starvla][INFO] - Step 2808, Loss: {'train/loss': 0.08695437759160995, 'train/grad_norm_pre_clip': 5.0244221687316895, 'timing/data': 0.0007358118891716003, 'timing/model': 0.9453485626727343, 'learning_rate/qwen_vl_interface': 6.82436509122125e-06, 'learning_rate/action_model': 6.82436509122125e-05, 'epoch': 1.16}) +[2026-06-17 21:12:15,732][starVLA.training.train_starvla][INFO] - Step 2809, Loss: {'train/loss': 0.5973081588745117, 'train/grad_norm_pre_clip': 16.017253875732422, 'timing/data': 0.0005019139498472214, 'timing/model': 1.0658477395772934, 'learning_rate/qwen_vl_interface': 6.822324484803353e-06, 'learning_rate/action_model': 6.822324484803353e-05, 'epoch': 1.16}) +[2026-06-17 21:12:19,644][starVLA.training.train_starvla][INFO] - Step 2810, Loss: {'train/loss': 0.39424559473991394, 'train/grad_norm_pre_clip': 15.56843090057373, 'timing/data': 0.0009042173624038696, 'timing/model': 0.9503745194524527, 'learning_rate/qwen_vl_interface': 6.8202835524409e-06, 'learning_rate/action_model': 6.8202835524409e-05, 'epoch': 1.16}) +[2026-06-17 21:12:23,658][starVLA.training.train_starvla][INFO] - Step 2811, Loss: {'train/loss': 0.20679175853729248, 'train/grad_norm_pre_clip': 7.398420333862305, 'timing/data': 0.0005898568779230118, 'timing/model': 1.2742790738120675, 'learning_rate/qwen_vl_interface': 6.818242294556977e-06, 'learning_rate/action_model': 6.818242294556977e-05, 'epoch': 1.16}) +[2026-06-17 21:12:28,424][starVLA.training.train_starvla][INFO] - Step 2812, Loss: {'train/loss': 0.4417095482349396, 'train/grad_norm_pre_clip': 15.243715286254883, 'timing/data': 0.0005191043019294739, 'timing/model': 1.1490393169224262, 'learning_rate/qwen_vl_interface': 6.816200711574738e-06, 'learning_rate/action_model': 6.816200711574738e-05, 'epoch': 1.16}) +[2026-06-17 21:12:34,137][starVLA.training.train_starvla][INFO] - Step 2813, Loss: {'train/loss': 0.3227182924747467, 'train/grad_norm_pre_clip': 11.688444137573242, 'timing/data': 0.0007856069132685661, 'timing/model': 0.9606201890856028, 'learning_rate/qwen_vl_interface': 6.814158803917406e-06, 'learning_rate/action_model': 6.814158803917406e-05, 'epoch': 1.16}) +[2026-06-17 21:12:39,107][starVLA.training.train_starvla][INFO] - Step 2814, Loss: {'train/loss': 0.5003814101219177, 'train/grad_norm_pre_clip': 14.618614196777344, 'timing/data': 0.00048756878823041916, 'timing/model': 1.4555676253512502, 'learning_rate/qwen_vl_interface': 6.812116572008271e-06, 'learning_rate/action_model': 6.812116572008271e-05, 'epoch': 1.16}) +[2026-06-17 21:12:42,937][starVLA.training.train_starvla][INFO] - Step 2815, Loss: {'train/loss': 0.636465311050415, 'train/grad_norm_pre_clip': 17.16199493408203, 'timing/data': 0.0007409397512674332, 'timing/model': 0.9350498532876372, 'learning_rate/qwen_vl_interface': 6.8100740162706906e-06, 'learning_rate/action_model': 6.81007401627069e-05, 'epoch': 1.16}) +[2026-06-17 21:12:46,831][starVLA.training.train_starvla][INFO] - Step 2816, Loss: {'train/loss': 0.5346202254295349, 'train/grad_norm_pre_clip': 22.87928009033203, 'timing/data': 0.0008607171475887299, 'timing/model': 0.9534468771889806, 'learning_rate/qwen_vl_interface': 6.808031137128087e-06, 'learning_rate/action_model': 6.808031137128086e-05, 'epoch': 1.16}) +[2026-06-17 21:12:51,209][starVLA.training.train_starvla][INFO] - Step 2817, Loss: {'train/loss': 0.25910133123397827, 'train/grad_norm_pre_clip': 12.916552543640137, 'timing/data': 0.000696948729455471, 'timing/model': 0.9594140900298953, 'learning_rate/qwen_vl_interface': 6.805987935003951e-06, 'learning_rate/action_model': 6.805987935003951e-05, 'epoch': 1.16}) +[2026-06-17 21:12:55,137][starVLA.training.train_starvla][INFO] - Step 2818, Loss: {'train/loss': 0.3962250351905823, 'train/grad_norm_pre_clip': 14.947317123413086, 'timing/data': 0.0006292164325714111, 'timing/model': 0.9347259486094117, 'learning_rate/qwen_vl_interface': 6.803944410321843e-06, 'learning_rate/action_model': 6.803944410321844e-05, 'epoch': 1.16}) +[2026-06-17 21:12:59,062][starVLA.training.train_starvla][INFO] - Step 2819, Loss: {'train/loss': 0.5906595587730408, 'train/grad_norm_pre_clip': 16.66092872619629, 'timing/data': 0.0004999022930860519, 'timing/model': 1.1130276480689645, 'learning_rate/qwen_vl_interface': 6.801900563505388e-06, 'learning_rate/action_model': 6.801900563505388e-05, 'epoch': 1.16}) +[2026-06-17 21:13:02,813][starVLA.training.train_starvla][INFO] - Step 2820, Loss: {'train/loss': 0.69418865442276, 'train/grad_norm_pre_clip': 18.517471313476562, 'timing/data': 0.0008686631917953491, 'timing/model': 1.010527839884162, 'learning_rate/qwen_vl_interface': 6.7998563949782755e-06, 'learning_rate/action_model': 6.799856394978276e-05, 'epoch': 1.16}) +[2026-06-17 21:13:06,701][starVLA.training.train_starvla][INFO] - Step 2821, Loss: {'train/loss': 0.965081512928009, 'train/grad_norm_pre_clip': 21.893680572509766, 'timing/data': 0.0008887723088264465, 'timing/model': 1.0044969748705626, 'learning_rate/qwen_vl_interface': 6.797811905164263e-06, 'learning_rate/action_model': 6.797811905164263e-05, 'epoch': 1.16}) +[2026-06-17 21:13:10,502][starVLA.training.train_starvla][INFO] - Step 2822, Loss: {'train/loss': 0.27631211280822754, 'train/grad_norm_pre_clip': 12.355326652526855, 'timing/data': 0.0011491989716887474, 'timing/model': 0.9529229570180178, 'learning_rate/qwen_vl_interface': 6.795767094487179e-06, 'learning_rate/action_model': 6.795767094487179e-05, 'epoch': 1.16}) +[2026-06-17 21:13:14,419][starVLA.training.train_starvla][INFO] - Step 2823, Loss: {'train/loss': 0.6444131731987, 'train/grad_norm_pre_clip': 11.725242614746094, 'timing/data': 0.0005239741876721382, 'timing/model': 0.9607010567560792, 'learning_rate/qwen_vl_interface': 6.793721963370913e-06, 'learning_rate/action_model': 6.793721963370913e-05, 'epoch': 1.16}) +[2026-06-17 21:13:18,360][starVLA.training.train_starvla][INFO] - Step 2824, Loss: {'train/loss': 0.5485339760780334, 'train/grad_norm_pre_clip': 15.313952445983887, 'timing/data': 0.000520472414791584, 'timing/model': 0.9996247971430421, 'learning_rate/qwen_vl_interface': 6.7916765122394215e-06, 'learning_rate/action_model': 6.791676512239421e-05, 'epoch': 1.16}) +[2026-06-17 21:13:22,210][starVLA.training.train_starvla][INFO] - Step 2825, Loss: {'train/loss': 0.5547009110450745, 'train/grad_norm_pre_clip': 107.4358139038086, 'timing/data': 0.0008720299229025841, 'timing/model': 1.1221811519935727, 'learning_rate/qwen_vl_interface': 6.7896307415167305e-06, 'learning_rate/action_model': 6.78963074151673e-05, 'epoch': 1.16}) +[2026-06-17 21:13:27,125][starVLA.training.train_starvla][INFO] - Step 2826, Loss: {'train/loss': 0.4417215883731842, 'train/grad_norm_pre_clip': 14.075243949890137, 'timing/data': 0.0005288654938340187, 'timing/model': 1.090369806624949, 'learning_rate/qwen_vl_interface': 6.787584651626929e-06, 'learning_rate/action_model': 6.78758465162693e-05, 'epoch': 1.16}) +[2026-06-17 21:13:30,840][starVLA.training.train_starvla][INFO] - Step 2827, Loss: {'train/loss': 0.34945785999298096, 'train/grad_norm_pre_clip': 9.123360633850098, 'timing/data': 0.0008827159181237221, 'timing/model': 0.9481086805462837, 'learning_rate/qwen_vl_interface': 6.785538242994177e-06, 'learning_rate/action_model': 6.785538242994177e-05, 'epoch': 1.16}) +[2026-06-17 21:13:34,810][starVLA.training.train_starvla][INFO] - Step 2828, Loss: {'train/loss': 0.3046529293060303, 'train/grad_norm_pre_clip': 33.220821380615234, 'timing/data': 0.0004878118634223938, 'timing/model': 0.95566051825881, 'learning_rate/qwen_vl_interface': 6.783491516042694e-06, 'learning_rate/action_model': 6.783491516042693e-05, 'epoch': 1.16}) +[2026-06-17 21:13:39,121][starVLA.training.train_starvla][INFO] - Step 2829, Loss: {'train/loss': 0.4344029128551483, 'train/grad_norm_pre_clip': 14.242867469787598, 'timing/data': 0.0005209669470787048, 'timing/model': 1.368106801994145, 'learning_rate/qwen_vl_interface': 6.781444471196768e-06, 'learning_rate/action_model': 6.781444471196769e-05, 'epoch': 1.16}) +[2026-06-17 21:13:42,973][starVLA.training.train_starvla][INFO] - Step 2830, Loss: {'train/loss': 0.6288957595825195, 'train/grad_norm_pre_clip': 21.182952880859375, 'timing/data': 0.001067989505827427, 'timing/model': 0.9386790310963988, 'learning_rate/qwen_vl_interface': 6.7793971088807575e-06, 'learning_rate/action_model': 6.779397108880757e-05, 'epoch': 1.16}) +[2026-06-17 21:13:46,816][starVLA.training.train_starvla][INFO] - Step 2831, Loss: {'train/loss': 0.5035553574562073, 'train/grad_norm_pre_clip': 15.371317863464355, 'timing/data': 0.0005177995190024376, 'timing/model': 0.95978441555053, 'learning_rate/qwen_vl_interface': 6.77734942951908e-06, 'learning_rate/action_model': 6.77734942951908e-05, 'epoch': 1.17}) +[2026-06-17 21:13:50,618][starVLA.training.train_starvla][INFO] - Step 2832, Loss: {'train/loss': 0.4570074677467346, 'train/grad_norm_pre_clip': 20.02349853515625, 'timing/data': 0.0005068574100732803, 'timing/model': 1.1110349977388978, 'learning_rate/qwen_vl_interface': 6.7753014335362224e-06, 'learning_rate/action_model': 6.775301433536222e-05, 'epoch': 1.17}) +[2026-06-17 21:13:54,245][starVLA.training.train_starvla][INFO] - Step 2833, Loss: {'train/loss': 0.7595822811126709, 'train/grad_norm_pre_clip': 22.164783477783203, 'timing/data': 0.00048712827265262604, 'timing/model': 0.9341656928882003, 'learning_rate/qwen_vl_interface': 6.773253121356735e-06, 'learning_rate/action_model': 6.773253121356735e-05, 'epoch': 1.17}) +[2026-06-17 21:13:58,103][starVLA.training.train_starvla][INFO] - Step 2834, Loss: {'train/loss': 0.31967610120773315, 'train/grad_norm_pre_clip': 11.546655654907227, 'timing/data': 0.0005301050841808319, 'timing/model': 0.9659051178023219, 'learning_rate/qwen_vl_interface': 6.771204493405239e-06, 'learning_rate/action_model': 6.771204493405239e-05, 'epoch': 1.17}) +[2026-06-17 21:14:01,883][starVLA.training.train_starvla][INFO] - Step 2835, Loss: {'train/loss': 0.3290439248085022, 'train/grad_norm_pre_clip': 45.905303955078125, 'timing/data': 0.0006893817335367203, 'timing/model': 0.9475291455164552, 'learning_rate/qwen_vl_interface': 6.769155550106413e-06, 'learning_rate/action_model': 6.769155550106413e-05, 'epoch': 1.17}) +[2026-06-17 21:14:05,898][starVLA.training.train_starvla][INFO] - Step 2836, Loss: {'train/loss': 0.5802449584007263, 'train/grad_norm_pre_clip': 18.612520217895508, 'timing/data': 0.0007459921762347221, 'timing/model': 1.049013452604413, 'learning_rate/qwen_vl_interface': 6.767106291885006e-06, 'learning_rate/action_model': 6.767106291885006e-05, 'epoch': 1.17}) +[2026-06-17 21:14:09,711][starVLA.training.train_starvla][INFO] - Step 2837, Loss: {'train/loss': 0.5355852246284485, 'train/grad_norm_pre_clip': 13.268917083740234, 'timing/data': 0.00047381874173879623, 'timing/model': 0.9460308281704783, 'learning_rate/qwen_vl_interface': 6.765056719165833e-06, 'learning_rate/action_model': 6.765056719165834e-05, 'epoch': 1.17}) +[2026-06-17 21:14:13,626][starVLA.training.train_starvla][INFO] - Step 2838, Loss: {'train/loss': 0.5886155366897583, 'train/grad_norm_pre_clip': 15.736111640930176, 'timing/data': 0.00048022251576185226, 'timing/model': 1.1306922240182757, 'learning_rate/qwen_vl_interface': 6.763006832373772e-06, 'learning_rate/action_model': 6.763006832373771e-05, 'epoch': 1.17}) +[2026-06-17 21:14:17,537][starVLA.training.train_starvla][INFO] - Step 2839, Loss: {'train/loss': 0.8196046352386475, 'train/grad_norm_pre_clip': 15.113750457763672, 'timing/data': 0.0007264707237482071, 'timing/model': 1.151999800466001, 'learning_rate/qwen_vl_interface': 6.760956631933766e-06, 'learning_rate/action_model': 6.760956631933765e-05, 'epoch': 1.17}) +[2026-06-17 21:14:21,204][starVLA.training.train_starvla][INFO] - Step 2840, Loss: {'train/loss': 0.22822992503643036, 'train/grad_norm_pre_clip': 10.435829162597656, 'timing/data': 0.0004997393116354942, 'timing/model': 0.9478301787748933, 'learning_rate/qwen_vl_interface': 6.758906118270826e-06, 'learning_rate/action_model': 6.758906118270825e-05, 'epoch': 1.17}) +[2026-06-17 21:14:25,022][starVLA.training.train_starvla][INFO] - Step 2841, Loss: {'train/loss': 0.5780895948410034, 'train/grad_norm_pre_clip': 16.53462791442871, 'timing/data': 0.0006102630868554115, 'timing/model': 0.9530003508552909, 'learning_rate/qwen_vl_interface': 6.756855291810024e-06, 'learning_rate/action_model': 6.756855291810024e-05, 'epoch': 1.17}) +[2026-06-17 21:14:28,860][starVLA.training.train_starvla][INFO] - Step 2842, Loss: {'train/loss': 0.4220767617225647, 'train/grad_norm_pre_clip': 13.36276912689209, 'timing/data': 0.000928165391087532, 'timing/model': 0.9710348322987556, 'learning_rate/qwen_vl_interface': 6.754804152976495e-06, 'learning_rate/action_model': 6.754804152976495e-05, 'epoch': 1.17}) +[2026-06-17 21:14:32,649][starVLA.training.train_starvla][INFO] - Step 2843, Loss: {'train/loss': 0.44850972294807434, 'train/grad_norm_pre_clip': 67.70096588134766, 'timing/data': 0.0010937293991446495, 'timing/model': 0.9492397531867027, 'learning_rate/qwen_vl_interface': 6.752752702195448e-06, 'learning_rate/action_model': 6.752752702195448e-05, 'epoch': 1.17}) +[2026-06-17 21:14:37,840][starVLA.training.train_starvla][INFO] - Step 2844, Loss: {'train/loss': 0.6035137176513672, 'train/grad_norm_pre_clip': 15.847620010375977, 'timing/data': 0.0006802547723054886, 'timing/model': 0.9324508924037218, 'learning_rate/qwen_vl_interface': 6.750700939892151e-06, 'learning_rate/action_model': 6.75070093989215e-05, 'epoch': 1.17}) +[2026-06-17 21:14:42,607][starVLA.training.train_starvla][INFO] - Step 2845, Loss: {'train/loss': 0.5752127766609192, 'train/grad_norm_pre_clip': 32.43122863769531, 'timing/data': 0.0008574863895773888, 'timing/model': 1.9262333856895566, 'learning_rate/qwen_vl_interface': 6.748648866491931e-06, 'learning_rate/action_model': 6.748648866491932e-05, 'epoch': 1.17}) +[2026-06-17 21:14:47,180][starVLA.training.train_starvla][INFO] - Step 2846, Loss: {'train/loss': 0.5629975199699402, 'train/grad_norm_pre_clip': 16.339632034301758, 'timing/data': 0.0013275807723402977, 'timing/model': 1.8212264636531472, 'learning_rate/qwen_vl_interface': 6.7465964824201915e-06, 'learning_rate/action_model': 6.746596482420192e-05, 'epoch': 1.17}) +[2026-06-17 21:14:51,022][starVLA.training.train_starvla][INFO] - Step 2847, Loss: {'train/loss': 0.5507550239562988, 'train/grad_norm_pre_clip': 14.909719467163086, 'timing/data': 0.0005044387653470039, 'timing/model': 0.9512292351573706, 'learning_rate/qwen_vl_interface': 6.744543788102388e-06, 'learning_rate/action_model': 6.744543788102388e-05, 'epoch': 1.17}) +[2026-06-17 21:14:54,938][starVLA.training.train_starvla][INFO] - Step 2848, Loss: {'train/loss': 0.40349969267845154, 'train/grad_norm_pre_clip': 11.69502067565918, 'timing/data': 0.0006809402257204056, 'timing/model': 0.9367747390642762, 'learning_rate/qwen_vl_interface': 6.742490783964051e-06, 'learning_rate/action_model': 6.742490783964051e-05, 'epoch': 1.17}) +[2026-06-17 21:14:58,909][starVLA.training.train_starvla][INFO] - Step 2849, Loss: {'train/loss': 0.6299389600753784, 'train/grad_norm_pre_clip': 17.451953887939453, 'timing/data': 0.0004823785275220871, 'timing/model': 0.9563390305265784, 'learning_rate/qwen_vl_interface': 6.740437470430766e-06, 'learning_rate/action_model': 6.740437470430766e-05, 'epoch': 1.17}) +[2026-06-17 21:15:04,005][starVLA.training.train_starvla][INFO] - Step 2850, Loss: {'train/loss': 0.29302796721458435, 'train/grad_norm_pre_clip': 14.556279182434082, 'timing/data': 0.00047976337373256683, 'timing/model': 0.9585405848920345, 'learning_rate/qwen_vl_interface': 6.7383838479281915e-06, 'learning_rate/action_model': 6.738383847928191e-05, 'epoch': 1.17}) +[2026-06-17 21:15:08,296][starVLA.training.train_starvla][INFO] - Step 2851, Loss: {'train/loss': 0.508306086063385, 'train/grad_norm_pre_clip': 16.467578887939453, 'timing/data': 0.0006433306261897087, 'timing/model': 1.4753485061228275, 'learning_rate/qwen_vl_interface': 6.736329916882044e-06, 'learning_rate/action_model': 6.736329916882044e-05, 'epoch': 1.17}) +[2026-06-17 21:15:12,269][starVLA.training.train_starvla][INFO] - Step 2852, Loss: {'train/loss': 0.451413631439209, 'train/grad_norm_pre_clip': 18.877370834350586, 'timing/data': 0.0007252916693687439, 'timing/model': 1.23561601806432, 'learning_rate/qwen_vl_interface': 6.734275677718105e-06, 'learning_rate/action_model': 6.734275677718105e-05, 'epoch': 1.17}) +[2026-06-17 21:15:16,107][starVLA.training.train_starvla][INFO] - Step 2853, Loss: {'train/loss': 0.27447766065597534, 'train/grad_norm_pre_clip': 13.035025596618652, 'timing/data': 0.0008795121684670448, 'timing/model': 1.0672649024054408, 'learning_rate/qwen_vl_interface': 6.73222113086222e-06, 'learning_rate/action_model': 6.73222113086222e-05, 'epoch': 1.17}) +[2026-06-17 21:15:20,469][starVLA.training.train_starvla][INFO] - Step 2854, Loss: {'train/loss': 0.47776472568511963, 'train/grad_norm_pre_clip': 17.357303619384766, 'timing/data': 0.0005121910944581032, 'timing/model': 0.9410215560346842, 'learning_rate/qwen_vl_interface': 6.7301662767403e-06, 'learning_rate/action_model': 6.7301662767403e-05, 'epoch': 1.17}) +[2026-06-17 21:15:25,533][starVLA.training.train_starvla][INFO] - Step 2855, Loss: {'train/loss': 0.4184015691280365, 'train/grad_norm_pre_clip': 15.335326194763184, 'timing/data': 0.0005920631811022758, 'timing/model': 0.9370819749310613, 'learning_rate/qwen_vl_interface': 6.728111115778319e-06, 'learning_rate/action_model': 6.728111115778318e-05, 'epoch': 1.18}) +[2026-06-17 21:15:29,452][starVLA.training.train_starvla][INFO] - Step 2856, Loss: {'train/loss': 0.34381359815597534, 'train/grad_norm_pre_clip': 11.8434419631958, 'timing/data': 0.000810086727142334, 'timing/model': 0.9511957271024585, 'learning_rate/qwen_vl_interface': 6.7260556484023135e-06, 'learning_rate/action_model': 6.726055648402313e-05, 'epoch': 1.18}) +[2026-06-17 21:15:33,353][starVLA.training.train_starvla][INFO] - Step 2857, Loss: {'train/loss': 0.6859642863273621, 'train/grad_norm_pre_clip': 26.22684669494629, 'timing/data': 0.0007973052561283112, 'timing/model': 0.9489164715632796, 'learning_rate/qwen_vl_interface': 6.723999875038381e-06, 'learning_rate/action_model': 6.72399987503838e-05, 'epoch': 1.18}) +[2026-06-17 21:15:37,371][starVLA.training.train_starvla][INFO] - Step 2858, Loss: {'train/loss': 0.5221205949783325, 'train/grad_norm_pre_clip': 17.26454734802246, 'timing/data': 0.000601181760430336, 'timing/model': 1.148911447264254, 'learning_rate/qwen_vl_interface': 6.721943796112691e-06, 'learning_rate/action_model': 6.721943796112691e-05, 'epoch': 1.18}) +[2026-06-17 21:15:41,182][starVLA.training.train_starvla][INFO] - Step 2859, Loss: {'train/loss': 0.378388375043869, 'train/grad_norm_pre_clip': 106.1505126953125, 'timing/data': 0.00105209369212389, 'timing/model': 0.945369721390307, 'learning_rate/qwen_vl_interface': 6.719887412051468e-06, 'learning_rate/action_model': 6.719887412051468e-05, 'epoch': 1.18}) +[2026-06-17 21:15:45,106][starVLA.training.train_starvla][INFO] - Step 2860, Loss: {'train/loss': 0.506867527961731, 'train/grad_norm_pre_clip': 15.761494636535645, 'timing/data': 0.0005113314837217331, 'timing/model': 0.9431976843625307, 'learning_rate/qwen_vl_interface': 6.7178307232810015e-06, 'learning_rate/action_model': 6.717830723281e-05, 'epoch': 1.18}) +[2026-06-17 21:15:49,052][starVLA.training.train_starvla][INFO] - Step 2861, Loss: {'train/loss': 0.6539866924285889, 'train/grad_norm_pre_clip': 13.569498062133789, 'timing/data': 0.0006386982277035713, 'timing/model': 0.9586989721283317, 'learning_rate/qwen_vl_interface': 6.715773730227646e-06, 'learning_rate/action_model': 6.715773730227646e-05, 'epoch': 1.18}) +[2026-06-17 21:15:52,993][starVLA.training.train_starvla][INFO] - Step 2862, Loss: {'train/loss': 0.5466868281364441, 'train/grad_norm_pre_clip': 14.594202041625977, 'timing/data': 0.0006027929484844208, 'timing/model': 0.9506239108741283, 'learning_rate/qwen_vl_interface': 6.7137164333178216e-06, 'learning_rate/action_model': 6.713716433317821e-05, 'epoch': 1.18}) +[2026-06-17 21:15:56,881][starVLA.training.train_starvla][INFO] - Step 2863, Loss: {'train/loss': 0.558985710144043, 'train/grad_norm_pre_clip': 16.98383331298828, 'timing/data': 0.0004903096705675125, 'timing/model': 0.944643254391849, 'learning_rate/qwen_vl_interface': 6.711658832978004e-06, 'learning_rate/action_model': 6.711658832978003e-05, 'epoch': 1.18}) +[2026-06-17 21:16:00,673][starVLA.training.train_starvla][INFO] - Step 2864, Loss: {'train/loss': 0.33113181591033936, 'train/grad_norm_pre_clip': 10.781977653503418, 'timing/data': 0.0006593819707632065, 'timing/model': 0.9562301579862833, 'learning_rate/qwen_vl_interface': 6.709600929634737e-06, 'learning_rate/action_model': 6.709600929634738e-05, 'epoch': 1.18}) +[2026-06-17 21:16:04,651][starVLA.training.train_starvla][INFO] - Step 2865, Loss: {'train/loss': 0.3625759780406952, 'train/grad_norm_pre_clip': 10.879810333251953, 'timing/data': 0.0005258889868855476, 'timing/model': 1.1158322673290968, 'learning_rate/qwen_vl_interface': 6.707542723714629e-06, 'learning_rate/action_model': 6.707542723714629e-05, 'epoch': 1.18}) +[2026-06-17 21:16:08,366][starVLA.training.train_starvla][INFO] - Step 2866, Loss: {'train/loss': 0.4808080196380615, 'train/grad_norm_pre_clip': 13.677573204040527, 'timing/data': 0.0005130451172590256, 'timing/model': 0.9397119442000985, 'learning_rate/qwen_vl_interface': 6.7054842156443435e-06, 'learning_rate/action_model': 6.705484215644343e-05, 'epoch': 1.18}) +[2026-06-17 21:16:12,289][starVLA.training.train_starvla][INFO] - Step 2867, Loss: {'train/loss': 0.4143489599227905, 'train/grad_norm_pre_clip': 14.74560832977295, 'timing/data': 0.0006996802985668182, 'timing/model': 0.952622995711863, 'learning_rate/qwen_vl_interface': 6.703425405850614e-06, 'learning_rate/action_model': 6.703425405850613e-05, 'epoch': 1.18}) +[2026-06-17 21:16:16,072][starVLA.training.train_starvla][INFO] - Step 2868, Loss: {'train/loss': 0.2718716561794281, 'train/grad_norm_pre_clip': 10.06153678894043, 'timing/data': 0.0004837429150938988, 'timing/model': 0.9400349324569106, 'learning_rate/qwen_vl_interface': 6.701366294760236e-06, 'learning_rate/action_model': 6.701366294760235e-05, 'epoch': 1.18}) +[2026-06-17 21:16:19,862][starVLA.training.train_starvla][INFO] - Step 2869, Loss: {'train/loss': 0.3747067153453827, 'train/grad_norm_pre_clip': 10.081637382507324, 'timing/data': 0.0005602529272437096, 'timing/model': 0.9370639128610492, 'learning_rate/qwen_vl_interface': 6.6993068828000605e-06, 'learning_rate/action_model': 6.69930688280006e-05, 'epoch': 1.18}) +[2026-06-17 21:16:23,734][starVLA.training.train_starvla][INFO] - Step 2870, Loss: {'train/loss': 0.7670732140541077, 'train/grad_norm_pre_clip': 15.153573989868164, 'timing/data': 0.0010482091456651688, 'timing/model': 0.9450221378356218, 'learning_rate/qwen_vl_interface': 6.697247170397011e-06, 'learning_rate/action_model': 6.697247170397011e-05, 'epoch': 1.18}) +[2026-06-17 21:16:27,797][starVLA.training.train_starvla][INFO] - Step 2871, Loss: {'train/loss': 0.5618113279342651, 'train/grad_norm_pre_clip': 14.907132148742676, 'timing/data': 0.0004993118345737457, 'timing/model': 1.1204870641231537, 'learning_rate/qwen_vl_interface': 6.6951871579780645e-06, 'learning_rate/action_model': 6.695187157978064e-05, 'epoch': 1.18}) +[2026-06-17 21:16:31,623][starVLA.training.train_starvla][INFO] - Step 2872, Loss: {'train/loss': 0.7289509773254395, 'train/grad_norm_pre_clip': 15.95230770111084, 'timing/data': 0.0011202739551663399, 'timing/model': 1.126775898039341, 'learning_rate/qwen_vl_interface': 6.693126845970267e-06, 'learning_rate/action_model': 6.693126845970267e-05, 'epoch': 1.18}) +[2026-06-17 21:16:35,329][starVLA.training.train_starvla][INFO] - Step 2873, Loss: {'train/loss': 0.6245613694190979, 'train/grad_norm_pre_clip': 12.706944465637207, 'timing/data': 0.0008661700412631035, 'timing/model': 0.9658905463293195, 'learning_rate/qwen_vl_interface': 6.691066234800719e-06, 'learning_rate/action_model': 6.691066234800719e-05, 'epoch': 1.18}) +[2026-06-17 21:16:39,261][starVLA.training.train_starvla][INFO] - Step 2874, Loss: {'train/loss': 0.5706769227981567, 'train/grad_norm_pre_clip': 16.546689987182617, 'timing/data': 0.0005091158673167229, 'timing/model': 0.9394233338534832, 'learning_rate/qwen_vl_interface': 6.689005324896591e-06, 'learning_rate/action_model': 6.689005324896591e-05, 'epoch': 1.18}) +[2026-06-17 21:16:43,113][starVLA.training.train_starvla][INFO] - Step 2875, Loss: {'train/loss': 0.5449662804603577, 'train/grad_norm_pre_clip': 13.06955337524414, 'timing/data': 0.0004787677899003029, 'timing/model': 0.9530955897644162, 'learning_rate/qwen_vl_interface': 6.686944116685111e-06, 'learning_rate/action_model': 6.686944116685111e-05, 'epoch': 1.18}) +[2026-06-17 21:16:47,080][starVLA.training.train_starvla][INFO] - Step 2876, Loss: {'train/loss': 0.759515643119812, 'train/grad_norm_pre_clip': 17.0318603515625, 'timing/data': 0.0007651681080460548, 'timing/model': 1.0178927825763822, 'learning_rate/qwen_vl_interface': 6.684882610593568e-06, 'learning_rate/action_model': 6.684882610593567e-05, 'epoch': 1.18}) +[2026-06-17 21:16:51,025][starVLA.training.train_starvla][INFO] - Step 2877, Loss: {'train/loss': 0.3201311528682709, 'train/grad_norm_pre_clip': 10.787520408630371, 'timing/data': 0.0004916265606880188, 'timing/model': 0.9695054860785604, 'learning_rate/qwen_vl_interface': 6.682820807049317e-06, 'learning_rate/action_model': 6.682820807049317e-05, 'epoch': 1.18}) +[2026-06-17 21:16:54,894][starVLA.training.train_starvla][INFO] - Step 2878, Loss: {'train/loss': 0.6289908289909363, 'train/grad_norm_pre_clip': 16.839834213256836, 'timing/data': 0.0007810285314917564, 'timing/model': 1.1546144103631377, 'learning_rate/qwen_vl_interface': 6.680758706479772e-06, 'learning_rate/action_model': 6.680758706479772e-05, 'epoch': 1.18}) +[2026-06-17 21:16:58,787][starVLA.training.train_starvla][INFO] - Step 2879, Loss: {'train/loss': 0.38057196140289307, 'train/grad_norm_pre_clip': 12.617778778076172, 'timing/data': 0.0007617007941007614, 'timing/model': 1.0003737239167094, 'learning_rate/qwen_vl_interface': 6.678696309312405e-06, 'learning_rate/action_model': 6.678696309312406e-05, 'epoch': 1.19}) +[2026-06-17 21:17:05,303][starVLA.training.train_starvla][INFO] - Step 2880, Loss: {'train/loss': 0.40797609090805054, 'train/grad_norm_pre_clip': 11.83492660522461, 'timing/data': 0.0005581611767411232, 'timing/model': 0.9533251076936722, 'learning_rate/qwen_vl_interface': 6.676633615974758e-06, 'learning_rate/action_model': 6.676633615974757e-05, 'epoch': 1.19}) +[2026-06-17 21:17:10,593][starVLA.training.train_starvla][INFO] - Step 2881, Loss: {'train/loss': 0.32735416293144226, 'train/grad_norm_pre_clip': 11.244450569152832, 'timing/data': 0.0008596498519182205, 'timing/model': 0.9790755212306976, 'learning_rate/qwen_vl_interface': 6.6745706268944265e-06, 'learning_rate/action_model': 6.674570626894426e-05, 'epoch': 1.19}) +[2026-06-17 21:17:14,454][starVLA.training.train_starvla][INFO] - Step 2882, Loss: {'train/loss': 0.6097719669342041, 'train/grad_norm_pre_clip': 11.473709106445312, 'timing/data': 0.0004855804145336151, 'timing/model': 0.9373268811032176, 'learning_rate/qwen_vl_interface': 6.672507342499071e-06, 'learning_rate/action_model': 6.67250734249907e-05, 'epoch': 1.19}) +[2026-06-17 21:17:19,290][starVLA.training.train_starvla][INFO] - Step 2883, Loss: {'train/loss': 0.282230943441391, 'train/grad_norm_pre_clip': 8.265486717224121, 'timing/data': 0.0008157370612025261, 'timing/model': 1.929769086651504, 'learning_rate/qwen_vl_interface': 6.670443763216412e-06, 'learning_rate/action_model': 6.670443763216411e-05, 'epoch': 1.19}) +[2026-06-17 21:17:23,125][starVLA.training.train_starvla][INFO] - Step 2884, Loss: {'train/loss': 0.23707985877990723, 'train/grad_norm_pre_clip': 11.235856056213379, 'timing/data': 0.0004798080772161484, 'timing/model': 0.9567015031352639, 'learning_rate/qwen_vl_interface': 6.668379889474233e-06, 'learning_rate/action_model': 6.668379889474234e-05, 'epoch': 1.19}) +[2026-06-17 21:17:26,972][starVLA.training.train_starvla][INFO] - Step 2885, Loss: {'train/loss': 0.5250265002250671, 'train/grad_norm_pre_clip': 13.729832649230957, 'timing/data': 0.00089213065803051, 'timing/model': 1.1026412388309836, 'learning_rate/qwen_vl_interface': 6.666315721700378e-06, 'learning_rate/action_model': 6.666315721700377e-05, 'epoch': 1.19}) +[2026-06-17 21:17:30,622][starVLA.training.train_starvla][INFO] - Step 2886, Loss: {'train/loss': 0.6745214462280273, 'train/grad_norm_pre_clip': 15.566774368286133, 'timing/data': 0.0004924256354570389, 'timing/model': 0.9473113715648651, 'learning_rate/qwen_vl_interface': 6.664251260322748e-06, 'learning_rate/action_model': 6.664251260322749e-05, 'epoch': 1.19}) +[2026-06-17 21:17:34,502][starVLA.training.train_starvla][INFO] - Step 2887, Loss: {'train/loss': 0.6376500725746155, 'train/grad_norm_pre_clip': 256.5135192871094, 'timing/data': 0.0012567602097988129, 'timing/model': 0.9417271232232451, 'learning_rate/qwen_vl_interface': 6.66218650576931e-06, 'learning_rate/action_model': 6.66218650576931e-05, 'epoch': 1.19}) +[2026-06-17 21:17:38,412][starVLA.training.train_starvla][INFO] - Step 2888, Loss: {'train/loss': 0.5541343092918396, 'train/grad_norm_pre_clip': 16.39876937866211, 'timing/data': 0.0004862472414970398, 'timing/model': 1.0317238681018353, 'learning_rate/qwen_vl_interface': 6.66012145846809e-06, 'learning_rate/action_model': 6.660121458468091e-05, 'epoch': 1.19}) +[2026-06-17 21:17:42,242][starVLA.training.train_starvla][INFO] - Step 2889, Loss: {'train/loss': 0.7843946814537048, 'train/grad_norm_pre_clip': 25.45974349975586, 'timing/data': 0.0007810238748788834, 'timing/model': 0.9441531412303448, 'learning_rate/qwen_vl_interface': 6.658056118847176e-06, 'learning_rate/action_model': 6.658056118847175e-05, 'epoch': 1.19}) +[2026-06-17 21:17:46,135][starVLA.training.train_starvla][INFO] - Step 2890, Loss: {'train/loss': 0.3646608591079712, 'train/grad_norm_pre_clip': 14.658889770507812, 'timing/data': 0.0008360547944903374, 'timing/model': 0.9415422175079584, 'learning_rate/qwen_vl_interface': 6.655990487334713e-06, 'learning_rate/action_model': 6.655990487334712e-05, 'epoch': 1.19}) +[2026-06-17 21:17:50,110][starVLA.training.train_starvla][INFO] - Step 2891, Loss: {'train/loss': 0.8927189111709595, 'train/grad_norm_pre_clip': 24.283945083618164, 'timing/data': 0.0009001949802041054, 'timing/model': 1.123112021945417, 'learning_rate/qwen_vl_interface': 6.653924564358907e-06, 'learning_rate/action_model': 6.653924564358907e-05, 'epoch': 1.19}) +[2026-06-17 21:17:54,055][starVLA.training.train_starvla][INFO] - Step 2892, Loss: {'train/loss': 0.538288950920105, 'train/grad_norm_pre_clip': 14.378230094909668, 'timing/data': 0.001278483308851719, 'timing/model': 1.1476537706330419, 'learning_rate/qwen_vl_interface': 6.6518583503480294e-06, 'learning_rate/action_model': 6.651858350348029e-05, 'epoch': 1.19}) +[2026-06-17 21:17:58,348][starVLA.training.train_starvla][INFO] - Step 2893, Loss: {'train/loss': 0.6492434740066528, 'train/grad_norm_pre_clip': 15.625347137451172, 'timing/data': 0.000658196397125721, 'timing/model': 0.9390495177358389, 'learning_rate/qwen_vl_interface': 6.649791845730408e-06, 'learning_rate/action_model': 6.649791845730408e-05, 'epoch': 1.19}) +[2026-06-17 21:18:02,178][starVLA.training.train_starvla][INFO] - Step 2894, Loss: {'train/loss': 0.4393499195575714, 'train/grad_norm_pre_clip': 11.991840362548828, 'timing/data': 0.0006788689643144608, 'timing/model': 0.9373505664989352, 'learning_rate/qwen_vl_interface': 6.64772505093443e-06, 'learning_rate/action_model': 6.647725050934429e-05, 'epoch': 1.19}) +[2026-06-17 21:18:05,996][starVLA.training.train_starvla][INFO] - Step 2895, Loss: {'train/loss': 0.8153045773506165, 'train/grad_norm_pre_clip': 18.673181533813477, 'timing/data': 0.0004959246143698692, 'timing/model': 0.9560029692947865, 'learning_rate/qwen_vl_interface': 6.645657966388543e-06, 'learning_rate/action_model': 6.645657966388543e-05, 'epoch': 1.19}) +[2026-06-17 21:18:09,825][starVLA.training.train_starvla][INFO] - Step 2896, Loss: {'train/loss': 0.3858267664909363, 'train/grad_norm_pre_clip': 11.753814697265625, 'timing/data': 0.0004912810400128365, 'timing/model': 0.9695866256952286, 'learning_rate/qwen_vl_interface': 6.643590592521261e-06, 'learning_rate/action_model': 6.643590592521261e-05, 'epoch': 1.19}) +[2026-06-17 21:18:13,819][starVLA.training.train_starvla][INFO] - Step 2897, Loss: {'train/loss': 0.3558225929737091, 'train/grad_norm_pre_clip': 9.51724624633789, 'timing/data': 0.0005608461797237396, 'timing/model': 1.061373993754387, 'learning_rate/qwen_vl_interface': 6.641522929761148e-06, 'learning_rate/action_model': 6.641522929761148e-05, 'epoch': 1.19}) +[2026-06-17 21:18:17,764][starVLA.training.train_starvla][INFO] - Step 2898, Loss: {'train/loss': 0.5406818389892578, 'train/grad_norm_pre_clip': 15.66488265991211, 'timing/data': 0.0005148956552147865, 'timing/model': 1.1124524688348174, 'learning_rate/qwen_vl_interface': 6.639454978536835e-06, 'learning_rate/action_model': 6.639454978536834e-05, 'epoch': 1.19}) +[2026-06-17 21:18:21,446][starVLA.training.train_starvla][INFO] - Step 2899, Loss: {'train/loss': 0.3229222893714905, 'train/grad_norm_pre_clip': 10.877102851867676, 'timing/data': 0.0005119927227497101, 'timing/model': 0.9840482845902443, 'learning_rate/qwen_vl_interface': 6.63738673927701e-06, 'learning_rate/action_model': 6.63738673927701e-05, 'epoch': 1.19}) +[2026-06-17 21:20:35,656][starVLA.training.train_starvla][INFO] - Step 2900, Loss: {'train/loss': 0.6576026082038879, 'train/grad_norm_pre_clip': 14.39520263671875, 'eval/loss': 0.5159841775894165, 'eval/latency_6/loss': 0.5159841775894165, 'eval/demon_fire_precision': 0.9598301763553233, 'eval/demon_fire_recall': 0.9499030381383322, 'eval/demon_fire_f1': 0.9548408057179987, 'eval/demon_fire_rate_model': 0.8519755147468002, 'eval/demon_fire_rate_teacher': 0.860879243183083, 'eval/demon_left_precision': 0.7544483985765125, 'eval/demon_left_recall': 0.743859649122807, 'eval/demon_left_f1': 0.7491166077738516, 'eval/demon_left_rate_model': 0.1563717306622148, 'eval/demon_left_rate_teacher': 0.15859766277128548, 'eval/demon_right_precision': 0.7694704049844237, 'eval/demon_right_recall': 0.7588325652841782, 'eval/demon_right_f1': 0.7641144624903325, 'eval/demon_right_rate_model': 0.17863105175292154, 'eval/demon_right_rate_teacher': 0.18113522537562604, 'eval/demon_move_f1': 0.756615535132092, 'eval/demon_cc_f1': 0.8852620189629572, 'eval/latency_6/demon_fire_precision': 0.9598301763553233, 'eval/latency_6/demon_fire_recall': 0.9499030381383322, 'eval/latency_6/demon_fire_f1': 0.9548408057179987, 'eval/latency_6/demon_fire_rate_model': 0.8519755147468002, 'eval/latency_6/demon_fire_rate_teacher': 0.860879243183083, 'eval/latency_6/demon_left_precision': 0.7544483985765125, 'eval/latency_6/demon_left_recall': 0.743859649122807, 'eval/latency_6/demon_left_f1': 0.7491166077738516, 'eval/latency_6/demon_left_rate_model': 0.1563717306622148, 'eval/latency_6/demon_left_rate_teacher': 0.15859766277128548, 'eval/latency_6/demon_right_precision': 0.7694704049844237, 'eval/latency_6/demon_right_recall': 0.7588325652841782, 'eval/latency_6/demon_right_f1': 0.7641144624903325, 'eval/latency_6/demon_right_rate_model': 0.17863105175292154, 'eval/latency_6/demon_right_rate_teacher': 0.18113522537562604, 'eval/latency_6/demon_move_f1': 0.756615535132092, 'eval/latency_6/demon_cc_f1': 0.8852620189629572, 'timing/data': 0.0004843287169933319, 'timing/model': 0.95983359310776, 'learning_rate/qwen_vl_interface': 6.63531821241042e-06, 'learning_rate/action_model': 6.63531821241042e-05, 'epoch': 1.19}) +[2026-06-17 21:20:35,671][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_2900_state +[2026-06-17 21:20:35,706][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 21:20:42,672][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_2900_state/model.safetensors +[2026-06-17 21:20:58,394][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_2900_state/optimizer.bin +[2026-06-17 21:20:58,396][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_2900_state/sampler.bin +[2026-06-17 21:20:58,396][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_2900_state/sampler_1.bin +[2026-06-17 21:20:58,398][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_2900_state/random_states_0.pkl +[2026-06-17 21:20:58,402][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2900 +[2026-06-17 21:20:58,403][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 21:20:58,416][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 21:21:02,420][starVLA.training.train_starvla][INFO] - Step 2901, Loss: {'train/loss': 0.8738110661506653, 'train/grad_norm_pre_clip': 15.308098793029785, 'timing/data': 0.0005025723949074745, 'timing/model': 1.1878471542149782, 'learning_rate/qwen_vl_interface': 6.633249398365874e-06, 'learning_rate/action_model': 6.633249398365873e-05, 'epoch': 1.19}) +[2026-06-17 21:21:06,105][starVLA.training.train_starvla][INFO] - Step 2902, Loss: {'train/loss': 0.4936445355415344, 'train/grad_norm_pre_clip': 16.01038932800293, 'timing/data': 0.0005633039399981499, 'timing/model': 0.9573777783662081, 'learning_rate/qwen_vl_interface': 6.6311802975722395e-06, 'learning_rate/action_model': 6.63118029757224e-05, 'epoch': 1.19}) +[2026-06-17 21:21:10,065][starVLA.training.train_starvla][INFO] - Step 2903, Loss: {'train/loss': 0.4988837242126465, 'train/grad_norm_pre_clip': 13.974599838256836, 'timing/data': 0.0007313499227166176, 'timing/model': 0.942397209815681, 'learning_rate/qwen_vl_interface': 6.629110910458444e-06, 'learning_rate/action_model': 6.629110910458444e-05, 'epoch': 1.19}) +[2026-06-17 21:21:14,011][starVLA.training.train_starvla][INFO] - Step 2904, Loss: {'train/loss': 0.5760480165481567, 'train/grad_norm_pre_clip': 17.64309310913086, 'timing/data': 0.0007526176050305367, 'timing/model': 1.0304463189095259, 'learning_rate/qwen_vl_interface': 6.627041237453471e-06, 'learning_rate/action_model': 6.627041237453471e-05, 'epoch': 1.2}) +[2026-06-17 21:21:17,841][starVLA.training.train_starvla][INFO] - Step 2905, Loss: {'train/loss': 0.4178585410118103, 'train/grad_norm_pre_clip': 12.45000171661377, 'timing/data': 0.0008650822564959526, 'timing/model': 0.942078129388392, 'learning_rate/qwen_vl_interface': 6.624971278986369e-06, 'learning_rate/action_model': 6.62497127898637e-05, 'epoch': 1.2}) +[2026-06-17 21:21:21,737][starVLA.training.train_starvla][INFO] - Step 2906, Loss: {'train/loss': 0.6189725995063782, 'train/grad_norm_pre_clip': 20.8236083984375, 'timing/data': 0.0011628605425357819, 'timing/model': 0.9710216606035829, 'learning_rate/qwen_vl_interface': 6.622901035486241e-06, 'learning_rate/action_model': 6.622901035486241e-05, 'epoch': 1.2}) +[2026-06-17 21:21:25,671][starVLA.training.train_starvla][INFO] - Step 2907, Loss: {'train/loss': 0.46181124448776245, 'train/grad_norm_pre_clip': 15.609818458557129, 'timing/data': 0.0005552396178245544, 'timing/model': 1.1615265514701605, 'learning_rate/qwen_vl_interface': 6.620830507382251e-06, 'learning_rate/action_model': 6.620830507382251e-05, 'epoch': 1.2}) +[2026-06-17 21:21:29,523][starVLA.training.train_starvla][INFO] - Step 2908, Loss: {'train/loss': 0.15108540654182434, 'train/grad_norm_pre_clip': 5.88153600692749, 'timing/data': 0.0006128270179033279, 'timing/model': 1.1320712324231863, 'learning_rate/qwen_vl_interface': 6.618759695103623e-06, 'learning_rate/action_model': 6.618759695103622e-05, 'epoch': 1.2}) +[2026-06-17 21:21:33,292][starVLA.training.train_starvla][INFO] - Step 2909, Loss: {'train/loss': 0.23968517780303955, 'train/grad_norm_pre_clip': 9.060271263122559, 'timing/data': 0.0010832268744707108, 'timing/model': 0.944317203015089, 'learning_rate/qwen_vl_interface': 6.616688599079635e-06, 'learning_rate/action_model': 6.616688599079634e-05, 'epoch': 1.2}) +[2026-06-17 21:21:37,132][starVLA.training.train_starvla][INFO] - Step 2910, Loss: {'train/loss': 0.5778695940971375, 'train/grad_norm_pre_clip': 16.496978759765625, 'timing/data': 0.0012335972860455513, 'timing/model': 0.9417470954358578, 'learning_rate/qwen_vl_interface': 6.614617219739632e-06, 'learning_rate/action_model': 6.614617219739632e-05, 'epoch': 1.2}) +[2026-06-17 21:21:41,087][starVLA.training.train_starvla][INFO] - Step 2911, Loss: {'train/loss': 0.22107890248298645, 'train/grad_norm_pre_clip': 8.080233573913574, 'timing/data': 0.0007747896015644073, 'timing/model': 0.9433171823620796, 'learning_rate/qwen_vl_interface': 6.612545557513008e-06, 'learning_rate/action_model': 6.612545557513008e-05, 'epoch': 1.2}) +[2026-06-17 21:21:44,850][starVLA.training.train_starvla][INFO] - Step 2912, Loss: {'train/loss': 0.34698063135147095, 'train/grad_norm_pre_clip': 13.422057151794434, 'timing/data': 0.000838269479572773, 'timing/model': 0.926940337754786, 'learning_rate/qwen_vl_interface': 6.610473612829227e-06, 'learning_rate/action_model': 6.610473612829226e-05, 'epoch': 1.2}) +[2026-06-17 21:21:48,641][starVLA.training.train_starvla][INFO] - Step 2913, Loss: {'train/loss': 0.5075669288635254, 'train/grad_norm_pre_clip': 16.210025787353516, 'timing/data': 0.0008590416982769966, 'timing/model': 0.931704118847847, 'learning_rate/qwen_vl_interface': 6.608401386117801e-06, 'learning_rate/action_model': 6.6084013861178e-05, 'epoch': 1.2}) +[2026-06-17 21:21:52,569][starVLA.training.train_starvla][INFO] - Step 2914, Loss: {'train/loss': 0.6132041215896606, 'train/grad_norm_pre_clip': 18.480655670166016, 'timing/data': 0.0005771378055214882, 'timing/model': 1.1168292872607708, 'learning_rate/qwen_vl_interface': 6.606328877808306e-06, 'learning_rate/action_model': 6.606328877808305e-05, 'epoch': 1.2}) +[2026-06-17 21:21:56,250][starVLA.training.train_starvla][INFO] - Step 2915, Loss: {'train/loss': 0.3751585781574249, 'train/grad_norm_pre_clip': 10.545202255249023, 'timing/data': 0.0011859918013215065, 'timing/model': 0.9291947260499, 'learning_rate/qwen_vl_interface': 6.604256088330374e-06, 'learning_rate/action_model': 6.604256088330374e-05, 'epoch': 1.2}) +[2026-06-17 21:22:00,065][starVLA.training.train_starvla][INFO] - Step 2916, Loss: {'train/loss': 0.33771756291389465, 'train/grad_norm_pre_clip': 12.244376182556152, 'timing/data': 0.0011828355491161346, 'timing/model': 0.9309273418039083, 'learning_rate/qwen_vl_interface': 6.602183018113698e-06, 'learning_rate/action_model': 6.602183018113698e-05, 'epoch': 1.2}) +[2026-06-17 21:22:03,934][starVLA.training.train_starvla][INFO] - Step 2917, Loss: {'train/loss': 0.38360050320625305, 'train/grad_norm_pre_clip': 15.97031021118164, 'timing/data': 0.000641796737909317, 'timing/model': 1.0556348655372858, 'learning_rate/qwen_vl_interface': 6.600109667588029e-06, 'learning_rate/action_model': 6.60010966758803e-05, 'epoch': 1.2}) +[2026-06-17 21:22:07,841][starVLA.training.train_starvla][INFO] - Step 2918, Loss: {'train/loss': 0.27270445227622986, 'train/grad_norm_pre_clip': 9.600098609924316, 'timing/data': 0.0008674012497067451, 'timing/model': 0.9496097592636943, 'learning_rate/qwen_vl_interface': 6.598036037183172e-06, 'learning_rate/action_model': 6.598036037183172e-05, 'epoch': 1.2}) +[2026-06-17 21:22:11,741][starVLA.training.train_starvla][INFO] - Step 2919, Loss: {'train/loss': 0.4026924669742584, 'train/grad_norm_pre_clip': 12.787277221679688, 'timing/data': 0.000869976356625557, 'timing/model': 0.9436376597732306, 'learning_rate/qwen_vl_interface': 6.595962127328993e-06, 'learning_rate/action_model': 6.595962127328993e-05, 'epoch': 1.2}) +[2026-06-17 21:22:15,772][starVLA.training.train_starvla][INFO] - Step 2920, Loss: {'train/loss': 0.5167973041534424, 'train/grad_norm_pre_clip': 19.809371948242188, 'timing/data': 0.0004582703113555908, 'timing/model': 1.1422531111165881, 'learning_rate/qwen_vl_interface': 6.593887938455418e-06, 'learning_rate/action_model': 6.593887938455418e-05, 'epoch': 1.2}) +[2026-06-17 21:22:19,488][starVLA.training.train_starvla][INFO] - Step 2921, Loss: {'train/loss': 0.35473132133483887, 'train/grad_norm_pre_clip': 15.293627738952637, 'timing/data': 0.0007985066622495651, 'timing/model': 1.0044388603419065, 'learning_rate/qwen_vl_interface': 6.591813470992428e-06, 'learning_rate/action_model': 6.591813470992427e-05, 'epoch': 1.2}) +[2026-06-17 21:22:23,328][starVLA.training.train_starvla][INFO] - Step 2922, Loss: {'train/loss': 0.462067574262619, 'train/grad_norm_pre_clip': 16.78624153137207, 'timing/data': 0.0007620593532919884, 'timing/model': 0.9487133184447885, 'learning_rate/qwen_vl_interface': 6.58973872537006e-06, 'learning_rate/action_model': 6.589738725370059e-05, 'epoch': 1.2}) +[2026-06-17 21:22:27,278][starVLA.training.train_starvla][INFO] - Step 2923, Loss: {'train/loss': 0.3132143020629883, 'train/grad_norm_pre_clip': 11.47230339050293, 'timing/data': 0.0004906654357910156, 'timing/model': 0.9355934727936983, 'learning_rate/qwen_vl_interface': 6.587663702018413e-06, 'learning_rate/action_model': 6.587663702018412e-05, 'epoch': 1.2}) +[2026-06-17 21:22:31,125][starVLA.training.train_starvla][INFO] - Step 2924, Loss: {'train/loss': 0.5616931915283203, 'train/grad_norm_pre_clip': 17.460596084594727, 'timing/data': 0.0004882439970970154, 'timing/model': 0.943523564375937, 'learning_rate/qwen_vl_interface': 6.585588401367641e-06, 'learning_rate/action_model': 6.58558840136764e-05, 'epoch': 1.2}) +[2026-06-17 21:22:34,988][starVLA.training.train_starvla][INFO] - Step 2925, Loss: {'train/loss': 0.634172260761261, 'train/grad_norm_pre_clip': 14.956354141235352, 'timing/data': 0.0005512339994311333, 'timing/model': 0.9391341246664524, 'learning_rate/qwen_vl_interface': 6.583512823847957e-06, 'learning_rate/action_model': 6.583512823847958e-05, 'epoch': 1.2}) +[2026-06-17 21:22:38,998][starVLA.training.train_starvla][INFO] - Step 2926, Loss: {'train/loss': 0.25127819180488586, 'train/grad_norm_pre_clip': 11.890583038330078, 'timing/data': 0.0005224170163273811, 'timing/model': 1.0219400394707918, 'learning_rate/qwen_vl_interface': 6.5814369698896285e-06, 'learning_rate/action_model': 6.581436969889628e-05, 'epoch': 1.2}) +[2026-06-17 21:22:42,915][starVLA.training.train_starvla][INFO] - Step 2927, Loss: {'train/loss': 0.493612140417099, 'train/grad_norm_pre_clip': 17.748275756835938, 'timing/data': 0.00045368634164333344, 'timing/model': 1.097011641599238, 'learning_rate/qwen_vl_interface': 6.5793608399229855e-06, 'learning_rate/action_model': 6.579360839922985e-05, 'epoch': 1.2}) +[2026-06-17 21:22:46,550][starVLA.training.train_starvla][INFO] - Step 2928, Loss: {'train/loss': 0.5601719617843628, 'train/grad_norm_pre_clip': 16.20380401611328, 'timing/data': 0.0007188012823462486, 'timing/model': 0.9334489144384861, 'learning_rate/qwen_vl_interface': 6.577284434378408e-06, 'learning_rate/action_model': 6.577284434378408e-05, 'epoch': 1.21}) +[2026-06-17 21:22:50,433][starVLA.training.train_starvla][INFO] - Step 2929, Loss: {'train/loss': 0.5970913767814636, 'train/grad_norm_pre_clip': 19.310338973999023, 'timing/data': 0.0012414464727044106, 'timing/model': 0.9444296043366194, 'learning_rate/qwen_vl_interface': 6.5752077536863404e-06, 'learning_rate/action_model': 6.57520775368634e-05, 'epoch': 1.21}) +[2026-06-17 21:22:54,335][starVLA.training.train_starvla][INFO] - Step 2930, Loss: {'train/loss': 0.5748531222343445, 'train/grad_norm_pre_clip': 15.56003475189209, 'timing/data': 0.0008414527401328087, 'timing/model': 0.9381616655737162, 'learning_rate/qwen_vl_interface': 6.573130798277277e-06, 'learning_rate/action_model': 6.573130798277277e-05, 'epoch': 1.21}) +[2026-06-17 21:22:58,212][starVLA.training.train_starvla][INFO] - Step 2931, Loss: {'train/loss': 0.6506778597831726, 'train/grad_norm_pre_clip': 17.488704681396484, 'timing/data': 0.0005779564380645752, 'timing/model': 1.0312649421393871, 'learning_rate/qwen_vl_interface': 6.571053568581778e-06, 'learning_rate/action_model': 6.571053568581777e-05, 'epoch': 1.21}) +[2026-06-17 21:23:01,993][starVLA.training.train_starvla][INFO] - Step 2932, Loss: {'train/loss': 0.27592816948890686, 'train/grad_norm_pre_clip': 14.254302978515625, 'timing/data': 0.0005315821617841721, 'timing/model': 0.9423697367310524, 'learning_rate/qwen_vl_interface': 6.568976065030453e-06, 'learning_rate/action_model': 6.568976065030453e-05, 'epoch': 1.21}) +[2026-06-17 21:23:05,905][starVLA.training.train_starvla][INFO] - Step 2933, Loss: {'train/loss': 0.6033678650856018, 'train/grad_norm_pre_clip': 18.19500732421875, 'timing/data': 0.0008529983460903168, 'timing/model': 1.170602428726852, 'learning_rate/qwen_vl_interface': 6.566898288053968e-06, 'learning_rate/action_model': 6.566898288053968e-05, 'epoch': 1.21}) +[2026-06-17 21:23:09,585][starVLA.training.train_starvla][INFO] - Step 2934, Loss: {'train/loss': 0.35588762164115906, 'train/grad_norm_pre_clip': 15.012079238891602, 'timing/data': 0.0005022604018449783, 'timing/model': 0.9470426123589277, 'learning_rate/qwen_vl_interface': 6.564820238083053e-06, 'learning_rate/action_model': 6.564820238083052e-05, 'epoch': 1.21}) +[2026-06-17 21:23:13,748][starVLA.training.train_starvla][INFO] - Step 2935, Loss: {'train/loss': 0.6022909879684448, 'train/grad_norm_pre_clip': 17.089391708374023, 'timing/data': 0.0008217180147767067, 'timing/model': 0.9362843725830317, 'learning_rate/qwen_vl_interface': 6.562741915548486e-06, 'learning_rate/action_model': 6.562741915548486e-05, 'epoch': 1.21}) +[2026-06-17 21:23:17,538][starVLA.training.train_starvla][INFO] - Step 2936, Loss: {'train/loss': 0.5504143834114075, 'train/grad_norm_pre_clip': 15.683003425598145, 'timing/data': 0.0008378541097044945, 'timing/model': 0.9429575717076659, 'learning_rate/qwen_vl_interface': 6.5606633208811096e-06, 'learning_rate/action_model': 6.560663320881109e-05, 'epoch': 1.21}) +[2026-06-17 21:23:21,336][starVLA.training.train_starvla][INFO] - Step 2937, Loss: {'train/loss': 0.4509899914264679, 'train/grad_norm_pre_clip': 13.307401657104492, 'timing/data': 0.0006786128506064415, 'timing/model': 0.9399909861385822, 'learning_rate/qwen_vl_interface': 6.558584454511814e-06, 'learning_rate/action_model': 6.558584454511814e-05, 'epoch': 1.21}) +[2026-06-17 21:23:25,426][starVLA.training.train_starvla][INFO] - Step 2938, Loss: {'train/loss': 0.6866790652275085, 'train/grad_norm_pre_clip': 17.852237701416016, 'timing/data': 0.0008272277191281319, 'timing/model': 0.9609822435304523, 'learning_rate/qwen_vl_interface': 6.556505316871555e-06, 'learning_rate/action_model': 6.556505316871554e-05, 'epoch': 1.21}) +[2026-06-17 21:23:29,216][starVLA.training.train_starvla][INFO] - Step 2939, Loss: {'train/loss': 0.5436277389526367, 'train/grad_norm_pre_clip': 14.263370513916016, 'timing/data': 0.0008264165371656418, 'timing/model': 1.1040359567850828, 'learning_rate/qwen_vl_interface': 6.554425908391337e-06, 'learning_rate/action_model': 6.554425908391336e-05, 'epoch': 1.21}) +[2026-06-17 21:23:33,020][starVLA.training.train_starvla][INFO] - Step 2940, Loss: {'train/loss': 0.5186020731925964, 'train/grad_norm_pre_clip': 13.99057674407959, 'timing/data': 0.001097838394343853, 'timing/model': 1.1330618392676115, 'learning_rate/qwen_vl_interface': 6.552346229502224e-06, 'learning_rate/action_model': 6.552346229502224e-05, 'epoch': 1.21}) +[2026-06-17 21:23:36,692][starVLA.training.train_starvla][INFO] - Step 2941, Loss: {'train/loss': 0.4745814800262451, 'train/grad_norm_pre_clip': 17.375038146972656, 'timing/data': 0.0009982865303754807, 'timing/model': 0.9853636715561152, 'learning_rate/qwen_vl_interface': 6.550266280635337e-06, 'learning_rate/action_model': 6.550266280635336e-05, 'epoch': 1.21}) +[2026-06-17 21:23:40,530][starVLA.training.train_starvla][INFO] - Step 2942, Loss: {'train/loss': 0.6054795384407043, 'train/grad_norm_pre_clip': 113.57764434814453, 'timing/data': 0.0004889825358986855, 'timing/model': 0.9467493910342455, 'learning_rate/qwen_vl_interface': 6.54818606222185e-06, 'learning_rate/action_model': 6.548186062221849e-05, 'epoch': 1.21}) +[2026-06-17 21:23:44,437][starVLA.training.train_starvla][INFO] - Step 2943, Loss: {'train/loss': 0.383680135011673, 'train/grad_norm_pre_clip': 12.11689567565918, 'timing/data': 0.0007032342255115509, 'timing/model': 0.9404706582427025, 'learning_rate/qwen_vl_interface': 6.546105574692997e-06, 'learning_rate/action_model': 6.546105574692997e-05, 'epoch': 1.21}) +[2026-06-17 21:23:48,294][starVLA.training.train_starvla][INFO] - Step 2944, Loss: {'train/loss': 0.5747882723808289, 'train/grad_norm_pre_clip': 15.14034652709961, 'timing/data': 0.00047567766159772873, 'timing/model': 0.9512977646663785, 'learning_rate/qwen_vl_interface': 6.544024818480061e-06, 'learning_rate/action_model': 6.544024818480062e-05, 'epoch': 1.21}) +[2026-06-17 21:23:52,353][starVLA.training.train_starvla][INFO] - Step 2945, Loss: {'train/loss': 0.3843385875225067, 'train/grad_norm_pre_clip': 14.646462440490723, 'timing/data': 0.0006216121837496758, 'timing/model': 0.9458318706601858, 'learning_rate/qwen_vl_interface': 6.5419437940143905e-06, 'learning_rate/action_model': 6.54194379401439e-05, 'epoch': 1.21}) +[2026-06-17 21:23:56,269][starVLA.training.train_starvla][INFO] - Step 2946, Loss: {'train/loss': 0.9142245054244995, 'train/grad_norm_pre_clip': 17.98699378967285, 'timing/data': 0.0006967941299080849, 'timing/model': 1.1906294506043196, 'learning_rate/qwen_vl_interface': 6.5398625017273795e-06, 'learning_rate/action_model': 6.539862501727379e-05, 'epoch': 1.21}) +[2026-06-17 21:24:00,219][starVLA.training.train_starvla][INFO] - Step 2947, Loss: {'train/loss': 0.46498873829841614, 'train/grad_norm_pre_clip': 15.568264961242676, 'timing/data': 0.0010449746623635292, 'timing/model': 1.220096324570477, 'learning_rate/qwen_vl_interface': 6.537780942050486e-06, 'learning_rate/action_model': 6.537780942050486e-05, 'epoch': 1.21}) +[2026-06-17 21:24:04,249][starVLA.training.train_starvla][INFO] - Step 2948, Loss: {'train/loss': 0.4548201560974121, 'train/grad_norm_pre_clip': 13.710372924804688, 'timing/data': 0.0006019491702318192, 'timing/model': 1.0044029457494617, 'learning_rate/qwen_vl_interface': 6.535699115415216e-06, 'learning_rate/action_model': 6.535699115415216e-05, 'epoch': 1.21}) +[2026-06-17 21:24:08,340][starVLA.training.train_starvla][INFO] - Step 2949, Loss: {'train/loss': 0.27135366201400757, 'train/grad_norm_pre_clip': 11.424477577209473, 'timing/data': 0.0004917709156870842, 'timing/model': 1.1155833099037409, 'learning_rate/qwen_vl_interface': 6.533617022253137e-06, 'learning_rate/action_model': 6.533617022253137e-05, 'epoch': 1.21}) +[2026-06-17 21:24:12,334][starVLA.training.train_starvla][INFO] - Step 2950, Loss: {'train/loss': 0.34064093232154846, 'train/grad_norm_pre_clip': 11.65855598449707, 'timing/data': 0.000500895082950592, 'timing/model': 0.9622810902073979, 'learning_rate/qwen_vl_interface': 6.531534662995866e-06, 'learning_rate/action_model': 6.531534662995866e-05, 'epoch': 1.21}) +[2026-06-17 21:24:16,596][starVLA.training.train_starvla][INFO] - Step 2951, Loss: {'train/loss': 0.6010385751724243, 'train/grad_norm_pre_clip': 23.33205223083496, 'timing/data': 0.00048686377704143524, 'timing/model': 0.9595054788514972, 'learning_rate/qwen_vl_interface': 6.529452038075082e-06, 'learning_rate/action_model': 6.529452038075081e-05, 'epoch': 1.21}) +[2026-06-17 21:24:20,534][starVLA.training.train_starvla][INFO] - Step 2952, Loss: {'train/loss': 0.3575168251991272, 'train/grad_norm_pre_clip': 14.314725875854492, 'timing/data': 0.0005346117541193962, 'timing/model': 1.2230024598538876, 'learning_rate/qwen_vl_interface': 6.5273691479225124e-06, 'learning_rate/action_model': 6.527369147922511e-05, 'epoch': 1.22}) +[2026-06-17 21:24:24,428][starVLA.training.train_starvla][INFO] - Step 2953, Loss: {'train/loss': 0.7314159870147705, 'train/grad_norm_pre_clip': 16.359262466430664, 'timing/data': 0.0006183851510286331, 'timing/model': 1.13540246989578, 'learning_rate/qwen_vl_interface': 6.525285992969945e-06, 'learning_rate/action_model': 6.525285992969945e-05, 'epoch': 1.22}) +[2026-06-17 21:24:28,103][starVLA.training.train_starvla][INFO] - Step 2954, Loss: {'train/loss': 0.4992169439792633, 'train/grad_norm_pre_clip': 17.109310150146484, 'timing/data': 0.000506293959915638, 'timing/model': 0.948929968290031, 'learning_rate/qwen_vl_interface': 6.523202573649219e-06, 'learning_rate/action_model': 6.523202573649219e-05, 'epoch': 1.22}) +[2026-06-17 21:24:31,948][starVLA.training.train_starvla][INFO] - Step 2955, Loss: {'train/loss': 0.5145772695541382, 'train/grad_norm_pre_clip': 17.081308364868164, 'timing/data': 0.0005535418167710304, 'timing/model': 0.9425347410142422, 'learning_rate/qwen_vl_interface': 6.5211188903922294e-06, 'learning_rate/action_model': 6.52111889039223e-05, 'epoch': 1.22}) +[2026-06-17 21:24:35,818][starVLA.training.train_starvla][INFO] - Step 2956, Loss: {'train/loss': 0.609072208404541, 'train/grad_norm_pre_clip': 15.90911865234375, 'timing/data': 0.0009275656193494797, 'timing/model': 0.9667897168546915, 'learning_rate/qwen_vl_interface': 6.519034943630926e-06, 'learning_rate/action_model': 6.519034943630927e-05, 'epoch': 1.22}) +[2026-06-17 21:24:39,643][starVLA.training.train_starvla][INFO] - Step 2957, Loss: {'train/loss': 0.5552238821983337, 'train/grad_norm_pre_clip': 17.612056732177734, 'timing/data': 0.0006730342283844948, 'timing/model': 0.9580270675942302, 'learning_rate/qwen_vl_interface': 6.5169507337973125e-06, 'learning_rate/action_model': 6.516950733797312e-05, 'epoch': 1.22}) +[2026-06-17 21:24:43,609][starVLA.training.train_starvla][INFO] - Step 2958, Loss: {'train/loss': 0.3031414747238159, 'train/grad_norm_pre_clip': 9.3928804397583, 'timing/data': 0.0010761385783553123, 'timing/model': 1.0170497680082917, 'learning_rate/qwen_vl_interface': 6.51486626132345e-06, 'learning_rate/action_model': 6.51486626132345e-05, 'epoch': 1.22}) +[2026-06-17 21:24:47,797][starVLA.training.train_starvla][INFO] - Step 2959, Loss: {'train/loss': 0.5188992023468018, 'train/grad_norm_pre_clip': 14.523110389709473, 'timing/data': 0.0005333181470632553, 'timing/model': 1.2218604506924748, 'learning_rate/qwen_vl_interface': 6.512781526641449e-06, 'learning_rate/action_model': 6.512781526641449e-05, 'epoch': 1.22}) +[2026-06-17 21:24:51,466][starVLA.training.train_starvla][INFO] - Step 2960, Loss: {'train/loss': 0.6640981435775757, 'train/grad_norm_pre_clip': 16.99715232849121, 'timing/data': 0.0005668299272656441, 'timing/model': 0.942618016153574, 'learning_rate/qwen_vl_interface': 6.51069653018348e-06, 'learning_rate/action_model': 6.51069653018348e-05, 'epoch': 1.22}) +[2026-06-17 21:24:58,665][starVLA.training.train_starvla][INFO] - Step 2961, Loss: {'train/loss': 0.37674862146377563, 'train/grad_norm_pre_clip': 11.632874488830566, 'timing/data': 0.0005079172551631927, 'timing/model': 1.289202393963933, 'learning_rate/qwen_vl_interface': 6.508611272381763e-06, 'learning_rate/action_model': 6.508611272381763e-05, 'epoch': 1.22}) +[2026-06-17 21:25:02,806][starVLA.training.train_starvla][INFO] - Step 2962, Loss: {'train/loss': 0.6768978238105774, 'train/grad_norm_pre_clip': 18.369516372680664, 'timing/data': 0.0008354531601071358, 'timing/model': 1.001057787798345, 'learning_rate/qwen_vl_interface': 6.506525753668576e-06, 'learning_rate/action_model': 6.506525753668576e-05, 'epoch': 1.22}) +[2026-06-17 21:25:06,878][starVLA.training.train_starvla][INFO] - Step 2963, Loss: {'train/loss': 0.4593188762664795, 'train/grad_norm_pre_clip': 15.832598686218262, 'timing/data': 0.0005324222147464752, 'timing/model': 0.9449994442984462, 'learning_rate/qwen_vl_interface': 6.504439974476246e-06, 'learning_rate/action_model': 6.504439974476246e-05, 'epoch': 1.22}) +[2026-06-17 21:25:10,785][starVLA.training.train_starvla][INFO] - Step 2964, Loss: {'train/loss': 0.5918047428131104, 'train/grad_norm_pre_clip': 18.835018157958984, 'timing/data': 0.0005229506641626358, 'timing/model': 1.0182640505954623, 'learning_rate/qwen_vl_interface': 6.50235393523716e-06, 'learning_rate/action_model': 6.502353935237159e-05, 'epoch': 1.22}) +[2026-06-17 21:25:14,634][starVLA.training.train_starvla][INFO] - Step 2965, Loss: {'train/loss': 0.6762977838516235, 'train/grad_norm_pre_clip': 15.666234016418457, 'timing/data': 0.0012537799775600433, 'timing/model': 0.9378060540184379, 'learning_rate/qwen_vl_interface': 6.500267636383758e-06, 'learning_rate/action_model': 6.500267636383758e-05, 'epoch': 1.22}) +[2026-06-17 21:25:18,529][starVLA.training.train_starvla][INFO] - Step 2966, Loss: {'train/loss': 0.5497393608093262, 'train/grad_norm_pre_clip': 18.69844627380371, 'timing/data': 0.0005707601085305214, 'timing/model': 1.1637182394042611, 'learning_rate/qwen_vl_interface': 6.498181078348527e-06, 'learning_rate/action_model': 6.498181078348527e-05, 'epoch': 1.22}) +[2026-06-17 21:25:22,258][starVLA.training.train_starvla][INFO] - Step 2967, Loss: {'train/loss': 0.5659355521202087, 'train/grad_norm_pre_clip': 14.84549617767334, 'timing/data': 0.000563683919608593, 'timing/model': 0.9360834239050746, 'learning_rate/qwen_vl_interface': 6.496094261564018e-06, 'learning_rate/action_model': 6.496094261564018e-05, 'epoch': 1.22}) +[2026-06-17 21:25:26,103][starVLA.training.train_starvla][INFO] - Step 2968, Loss: {'train/loss': 0.34788233041763306, 'train/grad_norm_pre_clip': 11.612044334411621, 'timing/data': 0.0005020489916205406, 'timing/model': 0.9323982363566756, 'learning_rate/qwen_vl_interface': 6.494007186462824e-06, 'learning_rate/action_model': 6.494007186462823e-05, 'epoch': 1.22}) +[2026-06-17 21:25:29,876][starVLA.training.train_starvla][INFO] - Step 2969, Loss: {'train/loss': 0.5430740714073181, 'train/grad_norm_pre_clip': 16.92195701599121, 'timing/data': 0.000494467094540596, 'timing/model': 0.9416053323075175, 'learning_rate/qwen_vl_interface': 6.491919853477604e-06, 'learning_rate/action_model': 6.491919853477604e-05, 'epoch': 1.22}) +[2026-06-17 21:25:33,655][starVLA.training.train_starvla][INFO] - Step 2970, Loss: {'train/loss': 0.4799991250038147, 'train/grad_norm_pre_clip': 16.142593383789062, 'timing/data': 0.00047469884157180786, 'timing/model': 0.9392985934391618, 'learning_rate/qwen_vl_interface': 6.4898322630410615e-06, 'learning_rate/action_model': 6.489832263041062e-05, 'epoch': 1.22}) +[2026-06-17 21:25:37,445][starVLA.training.train_starvla][INFO] - Step 2971, Loss: {'train/loss': 0.35288241505622864, 'train/grad_norm_pre_clip': 11.91097640991211, 'timing/data': 0.0006892206147313118, 'timing/model': 0.9500660095363855, 'learning_rate/qwen_vl_interface': 6.487744415585958e-06, 'learning_rate/action_model': 6.487744415585958e-05, 'epoch': 1.22}) +[2026-06-17 21:25:41,396][starVLA.training.train_starvla][INFO] - Step 2972, Loss: {'train/loss': 0.2578747570514679, 'train/grad_norm_pre_clip': 9.587414741516113, 'timing/data': 0.0005315011367201805, 'timing/model': 1.1519634742289782, 'learning_rate/qwen_vl_interface': 6.485656311545103e-06, 'learning_rate/action_model': 6.485656311545103e-05, 'epoch': 1.22}) +[2026-06-17 21:25:45,052][starVLA.training.train_starvla][INFO] - Step 2973, Loss: {'train/loss': 0.48876380920410156, 'train/grad_norm_pre_clip': 15.832365036010742, 'timing/data': 0.0008358079940080643, 'timing/model': 0.9548042360693216, 'learning_rate/qwen_vl_interface': 6.483567951351366e-06, 'learning_rate/action_model': 6.483567951351365e-05, 'epoch': 1.22}) +[2026-06-17 21:25:48,938][starVLA.training.train_starvla][INFO] - Step 2974, Loss: {'train/loss': 0.6506384015083313, 'train/grad_norm_pre_clip': 17.067325592041016, 'timing/data': 0.0012850165367126465, 'timing/model': 0.9501072969287634, 'learning_rate/qwen_vl_interface': 6.481479335437667e-06, 'learning_rate/action_model': 6.481479335437667e-05, 'epoch': 1.22}) +[2026-06-17 21:25:52,806][starVLA.training.train_starvla][INFO] - Step 2975, Loss: {'train/loss': 0.620621383190155, 'train/grad_norm_pre_clip': 17.682207107543945, 'timing/data': 0.0012321341782808304, 'timing/model': 0.9548943908885121, 'learning_rate/qwen_vl_interface': 6.479390464236974e-06, 'learning_rate/action_model': 6.479390464236974e-05, 'epoch': 1.22}) +[2026-06-17 21:25:56,668][starVLA.training.train_starvla][INFO] - Step 2976, Loss: {'train/loss': 0.49658867716789246, 'train/grad_norm_pre_clip': 16.509729385375977, 'timing/data': 0.0005171364173293114, 'timing/model': 0.9405324934050441, 'learning_rate/qwen_vl_interface': 6.477301338182317e-06, 'learning_rate/action_model': 6.477301338182317e-05, 'epoch': 1.23}) +[2026-06-17 21:26:00,587][starVLA.training.train_starvla][INFO] - Step 2977, Loss: {'train/loss': 0.5311015248298645, 'train/grad_norm_pre_clip': 16.91837501525879, 'timing/data': 0.00046338699758052826, 'timing/model': 0.9514513164758682, 'learning_rate/qwen_vl_interface': 6.47521195770677e-06, 'learning_rate/action_model': 6.47521195770677e-05, 'epoch': 1.23}) +[2026-06-17 21:26:04,484][starVLA.training.train_starvla][INFO] - Step 2978, Loss: {'train/loss': 0.3152606189250946, 'train/grad_norm_pre_clip': 11.66230297088623, 'timing/data': 0.0004793331027030945, 'timing/model': 1.132870266214013, 'learning_rate/qwen_vl_interface': 6.473122323243467e-06, 'learning_rate/action_model': 6.473122323243467e-05, 'epoch': 1.23}) +[2026-06-17 21:26:12,297][starVLA.training.train_starvla][INFO] - Step 2979, Loss: {'train/loss': 0.4055863916873932, 'train/grad_norm_pre_clip': 12.454920768737793, 'timing/data': 0.0005361400544643402, 'timing/model': 1.1647208463400602, 'learning_rate/qwen_vl_interface': 6.471032435225591e-06, 'learning_rate/action_model': 6.47103243522559e-05, 'epoch': 1.23}) +[2026-06-17 21:26:16,094][starVLA.training.train_starvla][INFO] - Step 2980, Loss: {'train/loss': 0.3763681948184967, 'train/grad_norm_pre_clip': 14.062193870544434, 'timing/data': 0.0005814153701066971, 'timing/model': 0.949563211761415, 'learning_rate/qwen_vl_interface': 6.468942294086376e-06, 'learning_rate/action_model': 6.468942294086376e-05, 'epoch': 1.23}) +[2026-06-17 21:26:19,929][starVLA.training.train_starvla][INFO] - Step 2981, Loss: {'train/loss': 0.36890679597854614, 'train/grad_norm_pre_clip': 12.106365203857422, 'timing/data': 0.0005210023373365402, 'timing/model': 0.9491559257730842, 'learning_rate/qwen_vl_interface': 6.466851900259113e-06, 'learning_rate/action_model': 6.466851900259112e-05, 'epoch': 1.23}) +[2026-06-17 21:26:23,824][starVLA.training.train_starvla][INFO] - Step 2982, Loss: {'train/loss': 0.44933655858039856, 'train/grad_norm_pre_clip': 15.818441390991211, 'timing/data': 0.0005619125440716743, 'timing/model': 1.042266950942576, 'learning_rate/qwen_vl_interface': 6.464761254177139e-06, 'learning_rate/action_model': 6.464761254177138e-05, 'epoch': 1.23}) +[2026-06-17 21:26:27,637][starVLA.training.train_starvla][INFO] - Step 2983, Loss: {'train/loss': 0.3097952604293823, 'train/grad_norm_pre_clip': 7.91713809967041, 'timing/data': 0.0008721742779016495, 'timing/model': 0.9456628803163767, 'learning_rate/qwen_vl_interface': 6.462670356273852e-06, 'learning_rate/action_model': 6.462670356273852e-05, 'epoch': 1.23}) +[2026-06-17 21:26:31,915][starVLA.training.train_starvla][INFO] - Step 2984, Loss: {'train/loss': 0.5526478886604309, 'train/grad_norm_pre_clip': 15.16073989868164, 'timing/data': 0.0007528951391577721, 'timing/model': 0.957763203419745, 'learning_rate/qwen_vl_interface': 6.460579206982695e-06, 'learning_rate/action_model': 6.460579206982696e-05, 'epoch': 1.23}) +[2026-06-17 21:26:37,198][starVLA.training.train_starvla][INFO] - Step 2985, Loss: {'train/loss': 0.3873787820339203, 'train/grad_norm_pre_clip': 15.25680923461914, 'timing/data': 0.0005088774487376213, 'timing/model': 1.1589583531022072, 'learning_rate/qwen_vl_interface': 6.458487806737166e-06, 'learning_rate/action_model': 6.458487806737165e-05, 'epoch': 1.23}) +[2026-06-17 21:26:40,884][starVLA.training.train_starvla][INFO] - Step 2986, Loss: {'train/loss': 0.5174100399017334, 'train/grad_norm_pre_clip': 18.18183708190918, 'timing/data': 0.000703420490026474, 'timing/model': 0.9437379222363234, 'learning_rate/qwen_vl_interface': 6.456396155970814e-06, 'learning_rate/action_model': 6.456396155970814e-05, 'epoch': 1.23}) +[2026-06-17 21:26:44,698][starVLA.training.train_starvla][INFO] - Step 2987, Loss: {'train/loss': 0.3258005380630493, 'train/grad_norm_pre_clip': 13.217061042785645, 'timing/data': 0.0005116807296872139, 'timing/model': 0.9477222003042698, 'learning_rate/qwen_vl_interface': 6.4543042551172404e-06, 'learning_rate/action_model': 6.45430425511724e-05, 'epoch': 1.23}) +[2026-06-17 21:26:48,517][starVLA.training.train_starvla][INFO] - Step 2988, Loss: {'train/loss': 0.2596794068813324, 'train/grad_norm_pre_clip': 11.187617301940918, 'timing/data': 0.0005099428817629814, 'timing/model': 0.9447261886671185, 'learning_rate/qwen_vl_interface': 6.4522121046101e-06, 'learning_rate/action_model': 6.452212104610099e-05, 'epoch': 1.23}) +[2026-06-17 21:26:52,334][starVLA.training.train_starvla][INFO] - Step 2989, Loss: {'train/loss': 0.3408096432685852, 'train/grad_norm_pre_clip': 12.293547630310059, 'timing/data': 0.0010321000590920448, 'timing/model': 0.9479733519256115, 'learning_rate/qwen_vl_interface': 6.450119704883095e-06, 'learning_rate/action_model': 6.450119704883096e-05, 'epoch': 1.23}) +[2026-06-17 21:26:56,202][starVLA.training.train_starvla][INFO] - Step 2990, Loss: {'train/loss': 0.4756562411785126, 'train/grad_norm_pre_clip': 14.127138137817383, 'timing/data': 0.0004815906286239624, 'timing/model': 0.9531022720038891, 'learning_rate/qwen_vl_interface': 6.448027056369983e-06, 'learning_rate/action_model': 6.448027056369983e-05, 'epoch': 1.23}) +[2026-06-17 21:27:00,872][starVLA.training.train_starvla][INFO] - Step 2991, Loss: {'train/loss': 0.40475767850875854, 'train/grad_norm_pre_clip': 11.733205795288086, 'timing/data': 0.000691266730427742, 'timing/model': 1.9725756468251348, 'learning_rate/qwen_vl_interface': 6.445934159504576e-06, 'learning_rate/action_model': 6.445934159504576e-05, 'epoch': 1.23}) +[2026-06-17 21:27:04,730][starVLA.training.train_starvla][INFO] - Step 2992, Loss: {'train/loss': 0.6804742217063904, 'train/grad_norm_pre_clip': 18.38873291015625, 'timing/data': 0.0013121822848916054, 'timing/model': 1.1369765158742666, 'learning_rate/qwen_vl_interface': 6.44384101472073e-06, 'learning_rate/action_model': 6.443841014720729e-05, 'epoch': 1.23}) +[2026-06-17 21:27:08,509][starVLA.training.train_starvla][INFO] - Step 2993, Loss: {'train/loss': 0.19071853160858154, 'train/grad_norm_pre_clip': 9.898520469665527, 'timing/data': 0.0010787751525640488, 'timing/model': 0.9475395660847425, 'learning_rate/qwen_vl_interface': 6.441747622452357e-06, 'learning_rate/action_model': 6.441747622452356e-05, 'epoch': 1.23}) +[2026-06-17 21:27:12,448][starVLA.training.train_starvla][INFO] - Step 2994, Loss: {'train/loss': 0.5402243137359619, 'train/grad_norm_pre_clip': 16.630878448486328, 'timing/data': 0.00091602373868227, 'timing/model': 0.9544268185272813, 'learning_rate/qwen_vl_interface': 6.4396539831334195e-06, 'learning_rate/action_model': 6.439653983133419e-05, 'epoch': 1.23}) +[2026-06-17 21:27:16,732][starVLA.training.train_starvla][INFO] - Step 2995, Loss: {'train/loss': 0.5480314493179321, 'train/grad_norm_pre_clip': 15.654502868652344, 'timing/data': 0.0014193570241332054, 'timing/model': 0.9637485221028328, 'learning_rate/qwen_vl_interface': 6.437560097197928e-06, 'learning_rate/action_model': 6.437560097197927e-05, 'epoch': 1.23}) +[2026-06-17 21:27:20,543][starVLA.training.train_starvla][INFO] - Step 2996, Loss: {'train/loss': 0.8486688733100891, 'train/grad_norm_pre_clip': 19.864917755126953, 'timing/data': 0.0005488898605108261, 'timing/model': 0.9360785400494933, 'learning_rate/qwen_vl_interface': 6.4354659650799546e-06, 'learning_rate/action_model': 6.435465965079955e-05, 'epoch': 1.23}) +[2026-06-17 21:27:24,485][starVLA.training.train_starvla][INFO] - Step 2997, Loss: {'train/loss': 0.2784270644187927, 'train/grad_norm_pre_clip': 13.714485168457031, 'timing/data': 0.0005163028836250305, 'timing/model': 0.9438028251752257, 'learning_rate/qwen_vl_interface': 6.433371587213609e-06, 'learning_rate/action_model': 6.433371587213607e-05, 'epoch': 1.23}) +[2026-06-17 21:27:28,295][starVLA.training.train_starvla][INFO] - Step 2998, Loss: {'train/loss': 0.5668345093727112, 'train/grad_norm_pre_clip': 14.035099983215332, 'timing/data': 0.0005327453836798668, 'timing/model': 1.1285897688940167, 'learning_rate/qwen_vl_interface': 6.431276964033062e-06, 'learning_rate/action_model': 6.431276964033062e-05, 'epoch': 1.23}) +[2026-06-17 21:27:31,942][starVLA.training.train_starvla][INFO] - Step 2999, Loss: {'train/loss': 0.3047294020652771, 'train/grad_norm_pre_clip': 10.573722839355469, 'timing/data': 0.0008756043389439583, 'timing/model': 0.953235249966383, 'learning_rate/qwen_vl_interface': 6.429182095972526e-06, 'learning_rate/action_model': 6.429182095972526e-05, 'epoch': 1.23}) +[2026-06-17 21:29:48,403][starVLA.training.train_starvla][INFO] - Step 3000, Loss: {'train/loss': 0.6164078712463379, 'train/grad_norm_pre_clip': 18.48198127746582, 'eval/loss': 0.49767985939979553, 'eval/latency_6/loss': 0.49767985939979553, 'eval/demon_fire_precision': 0.931783420006118, 'eval/demon_fire_recall': 0.984486102133161, 'eval/demon_fire_f1': 0.9574100267169574, 'eval/demon_fire_rate_model': 0.9095715080690039, 'eval/demon_fire_rate_teacher': 0.860879243183083, 'eval/demon_left_precision': 0.852803738317757, 'eval/demon_left_recall': 0.6403508771929824, 'eval/demon_left_f1': 0.7314629258517035, 'eval/demon_left_rate_model': 0.11908736783528102, 'eval/demon_left_rate_teacher': 0.15859766277128548, 'eval/demon_right_precision': 0.797274275979557, 'eval/demon_right_recall': 0.7188940092165899, 'eval/demon_right_f1': 0.7560581583198707, 'eval/demon_right_rate_model': 0.16332776850306066, 'eval/demon_right_rate_teacher': 0.18113522537562604, 'eval/demon_move_f1': 0.7437605420857871, 'eval/demon_cc_f1': 0.8814495904952873, 'eval/latency_6/demon_fire_precision': 0.931783420006118, 'eval/latency_6/demon_fire_recall': 0.984486102133161, 'eval/latency_6/demon_fire_f1': 0.9574100267169574, 'eval/latency_6/demon_fire_rate_model': 0.9095715080690039, 'eval/latency_6/demon_fire_rate_teacher': 0.860879243183083, 'eval/latency_6/demon_left_precision': 0.852803738317757, 'eval/latency_6/demon_left_recall': 0.6403508771929824, 'eval/latency_6/demon_left_f1': 0.7314629258517035, 'eval/latency_6/demon_left_rate_model': 0.11908736783528102, 'eval/latency_6/demon_left_rate_teacher': 0.15859766277128548, 'eval/latency_6/demon_right_precision': 0.797274275979557, 'eval/latency_6/demon_right_recall': 0.7188940092165899, 'eval/latency_6/demon_right_f1': 0.7560581583198707, 'eval/latency_6/demon_right_rate_model': 0.16332776850306066, 'eval/latency_6/demon_right_rate_teacher': 0.18113522537562604, 'eval/latency_6/demon_move_f1': 0.7437605420857871, 'eval/latency_6/demon_cc_f1': 0.8814495904952873, 'timing/data': 0.0011954531073570251, 'timing/model': 0.9294073982164264, 'learning_rate/qwen_vl_interface': 6.427086983466275e-06, 'learning_rate/action_model': 6.427086983466275e-05, 'epoch': 1.23}) +[2026-06-17 21:29:48,407][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_3000_state +[2026-06-17 21:29:48,430][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 21:29:55,209][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_3000_state/model.safetensors +[2026-06-17 21:30:10,999][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_3000_state/optimizer.bin +[2026-06-17 21:30:11,001][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_3000_state/sampler.bin +[2026-06-17 21:30:11,001][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_3000_state/sampler_1.bin +[2026-06-17 21:30:11,005][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_3000_state/random_states_0.pkl +[2026-06-17 21:30:11,011][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3000 +[2026-06-17 21:30:11,012][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 21:30:11,033][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 21:30:15,101][starVLA.training.train_starvla][INFO] - Step 3001, Loss: {'train/loss': 0.41013264656066895, 'train/grad_norm_pre_clip': 18.44291877746582, 'timing/data': 0.0005450630560517311, 'timing/model': 1.1786062559112906, 'learning_rate/qwen_vl_interface': 6.4249916269486255e-06, 'learning_rate/action_model': 6.424991626948626e-05, 'epoch': 1.24}) +[2026-06-17 21:30:18,756][starVLA.training.train_starvla][INFO] - Step 3002, Loss: {'train/loss': 0.4487278163433075, 'train/grad_norm_pre_clip': 13.993809700012207, 'timing/data': 0.0005256142467260361, 'timing/model': 0.9759189039468765, 'learning_rate/qwen_vl_interface': 6.422896026853948e-06, 'learning_rate/action_model': 6.422896026853947e-05, 'epoch': 1.24}) +[2026-06-17 21:30:22,666][starVLA.training.train_starvla][INFO] - Step 3003, Loss: {'train/loss': 0.46928197145462036, 'train/grad_norm_pre_clip': 17.94007110595703, 'timing/data': 0.0005400925874710083, 'timing/model': 0.9474881226196885, 'learning_rate/qwen_vl_interface': 6.420800183616661e-06, 'learning_rate/action_model': 6.420800183616661e-05, 'epoch': 1.24}) +[2026-06-17 21:30:27,901][starVLA.training.train_starvla][INFO] - Step 3004, Loss: {'train/loss': 0.3012976050376892, 'train/grad_norm_pre_clip': 14.593070030212402, 'timing/data': 0.0005702357739210129, 'timing/model': 2.256999467499554, 'learning_rate/qwen_vl_interface': 6.4187040976712355e-06, 'learning_rate/action_model': 6.418704097671235e-05, 'epoch': 1.24}) +[2026-06-17 21:30:31,987][starVLA.training.train_starvla][INFO] - Step 3005, Loss: {'train/loss': 0.5753451585769653, 'train/grad_norm_pre_clip': 14.503114700317383, 'timing/data': 0.0009279055520892143, 'timing/model': 1.1332395523786545, 'learning_rate/qwen_vl_interface': 6.416607769452195e-06, 'learning_rate/action_model': 6.416607769452194e-05, 'epoch': 1.24}) +[2026-06-17 21:30:36,382][starVLA.training.train_starvla][INFO] - Step 3006, Loss: {'train/loss': 0.3752883970737457, 'train/grad_norm_pre_clip': 12.43881893157959, 'timing/data': 0.0005065454170107841, 'timing/model': 1.0041263634338975, 'learning_rate/qwen_vl_interface': 6.414511199394106e-06, 'learning_rate/action_model': 6.414511199394105e-05, 'epoch': 1.24}) +[2026-06-17 21:30:40,180][starVLA.training.train_starvla][INFO] - Step 3007, Loss: {'train/loss': 0.43929240107536316, 'train/grad_norm_pre_clip': 13.260099411010742, 'timing/data': 0.000514608807861805, 'timing/model': 1.1049473648890853, 'learning_rate/qwen_vl_interface': 6.4124143879315935e-06, 'learning_rate/action_model': 6.412414387931593e-05, 'epoch': 1.24}) +[2026-06-17 21:30:43,932][starVLA.training.train_starvla][INFO] - Step 3008, Loss: {'train/loss': 0.5284515023231506, 'train/grad_norm_pre_clip': 14.176173210144043, 'timing/data': 0.0005683284252882004, 'timing/model': 1.0703352987766266, 'learning_rate/qwen_vl_interface': 6.410317335499328e-06, 'learning_rate/action_model': 6.410317335499328e-05, 'epoch': 1.24}) +[2026-06-17 21:30:47,672][starVLA.training.train_starvla][INFO] - Step 3009, Loss: {'train/loss': 0.631062924861908, 'train/grad_norm_pre_clip': 18.158695220947266, 'timing/data': 0.0005131615325808525, 'timing/model': 0.9358139680698514, 'learning_rate/qwen_vl_interface': 6.40822004253203e-06, 'learning_rate/action_model': 6.40822004253203e-05, 'epoch': 1.24}) +[2026-06-17 21:30:51,507][starVLA.training.train_starvla][INFO] - Step 3010, Loss: {'train/loss': 0.5461748242378235, 'train/grad_norm_pre_clip': 18.548725128173828, 'timing/data': 0.0005181198939681053, 'timing/model': 0.992223066277802, 'learning_rate/qwen_vl_interface': 6.406122509464471e-06, 'learning_rate/action_model': 6.40612250946447e-05, 'epoch': 1.24}) +[2026-06-17 21:30:55,718][starVLA.training.train_starvla][INFO] - Step 3011, Loss: {'train/loss': 0.4425968527793884, 'train/grad_norm_pre_clip': 17.374279022216797, 'timing/data': 0.0005241641774773598, 'timing/model': 0.944661807268858, 'learning_rate/qwen_vl_interface': 6.404024736731472e-06, 'learning_rate/action_model': 6.404024736731472e-05, 'epoch': 1.24}) +[2026-06-17 21:30:59,458][starVLA.training.train_starvla][INFO] - Step 3012, Loss: {'train/loss': 0.4163150489330292, 'train/grad_norm_pre_clip': 13.28464412689209, 'timing/data': 0.0005159778520464897, 'timing/model': 0.9264241950586438, 'learning_rate/qwen_vl_interface': 6.401926724767903e-06, 'learning_rate/action_model': 6.401926724767902e-05, 'epoch': 1.24}) +[2026-06-17 21:31:03,250][starVLA.training.train_starvla][INFO] - Step 3013, Loss: {'train/loss': 0.44805797934532166, 'train/grad_norm_pre_clip': 18.825895309448242, 'timing/data': 0.0005210656672716141, 'timing/model': 0.9416553117334843, 'learning_rate/qwen_vl_interface': 6.399828474008683e-06, 'learning_rate/action_model': 6.399828474008682e-05, 'epoch': 1.24}) +[2026-06-17 21:31:07,140][starVLA.training.train_starvla][INFO] - Step 3014, Loss: {'train/loss': 0.26585716009140015, 'train/grad_norm_pre_clip': 11.726852416992188, 'timing/data': 0.0011550243943929672, 'timing/model': 1.1279489966109395, 'learning_rate/qwen_vl_interface': 6.397729984888785e-06, 'learning_rate/action_model': 6.397729984888785e-05, 'epoch': 1.24}) +[2026-06-17 21:31:10,817][starVLA.training.train_starvla][INFO] - Step 3015, Loss: {'train/loss': 0.23673197627067566, 'train/grad_norm_pre_clip': 29.012828826904297, 'timing/data': 0.0007107416167855263, 'timing/model': 0.961170862428844, 'learning_rate/qwen_vl_interface': 6.395631257843224e-06, 'learning_rate/action_model': 6.395631257843224e-05, 'epoch': 1.24}) +[2026-06-17 21:31:14,696][starVLA.training.train_starvla][INFO] - Step 3016, Loss: {'train/loss': 0.48192286491394043, 'train/grad_norm_pre_clip': 12.786506652832031, 'timing/data': 0.0006646150723099709, 'timing/model': 0.9473955249413848, 'learning_rate/qwen_vl_interface': 6.393532293307073e-06, 'learning_rate/action_model': 6.393532293307072e-05, 'epoch': 1.24}) +[2026-06-17 21:31:18,472][starVLA.training.train_starvla][INFO] - Step 3017, Loss: {'train/loss': 0.3002517521381378, 'train/grad_norm_pre_clip': 8.968772888183594, 'timing/data': 0.000532146543264389, 'timing/model': 0.9360204758122563, 'learning_rate/qwen_vl_interface': 6.391433091715447e-06, 'learning_rate/action_model': 6.391433091715446e-05, 'epoch': 1.24}) +[2026-06-17 21:31:22,523][starVLA.training.train_starvla][INFO] - Step 3018, Loss: {'train/loss': 0.37090399861335754, 'train/grad_norm_pre_clip': 12.54914379119873, 'timing/data': 0.000519835390150547, 'timing/model': 1.0743692750111222, 'learning_rate/qwen_vl_interface': 6.389333653503511e-06, 'learning_rate/action_model': 6.38933365350351e-05, 'epoch': 1.24}) +[2026-06-17 21:31:26,486][starVLA.training.train_starvla][INFO] - Step 3019, Loss: {'train/loss': 0.5208870768547058, 'train/grad_norm_pre_clip': 14.228286743164062, 'timing/data': 0.0009138612076640129, 'timing/model': 1.0037492671981454, 'learning_rate/qwen_vl_interface': 6.387233979106484e-06, 'learning_rate/action_model': 6.387233979106483e-05, 'epoch': 1.24}) +[2026-06-17 21:31:30,398][starVLA.training.train_starvla][INFO] - Step 3020, Loss: {'train/loss': 0.30172866582870483, 'train/grad_norm_pre_clip': 13.906923294067383, 'timing/data': 0.0008135586977005005, 'timing/model': 1.1610735850408673, 'learning_rate/qwen_vl_interface': 6.385134068959629e-06, 'learning_rate/action_model': 6.385134068959628e-05, 'epoch': 1.24}) +[2026-06-17 21:31:34,450][starVLA.training.train_starvla][INFO] - Step 3021, Loss: {'train/loss': 0.5352838039398193, 'train/grad_norm_pre_clip': 16.57447624206543, 'timing/data': 0.000498434528708458, 'timing/model': 1.2003406630828977, 'learning_rate/qwen_vl_interface': 6.3830339234982596e-06, 'learning_rate/action_model': 6.38303392349826e-05, 'epoch': 1.24}) +[2026-06-17 21:31:39,628][starVLA.training.train_starvla][INFO] - Step 3022, Loss: {'train/loss': 0.569233775138855, 'train/grad_norm_pre_clip': 20.139692306518555, 'timing/data': 0.0009570308029651642, 'timing/model': 0.9538269257172942, 'learning_rate/qwen_vl_interface': 6.38093354315774e-06, 'learning_rate/action_model': 6.38093354315774e-05, 'epoch': 1.24}) +[2026-06-17 21:31:44,640][starVLA.training.train_starvla][INFO] - Step 3023, Loss: {'train/loss': 0.735813558101654, 'train/grad_norm_pre_clip': 17.7681941986084, 'timing/data': 0.0006414679810404778, 'timing/model': 2.0757001638412476, 'learning_rate/qwen_vl_interface': 6.378832928373479e-06, 'learning_rate/action_model': 6.378832928373478e-05, 'epoch': 1.24}) +[2026-06-17 21:31:49,788][starVLA.training.train_starvla][INFO] - Step 3024, Loss: {'train/loss': 0.5746815800666809, 'train/grad_norm_pre_clip': 17.269174575805664, 'timing/data': 0.0007814811542630196, 'timing/model': 2.110474018380046, 'learning_rate/qwen_vl_interface': 6.376732079580936e-06, 'learning_rate/action_model': 6.376732079580936e-05, 'epoch': 1.24}) +[2026-06-17 21:31:54,292][starVLA.training.train_starvla][INFO] - Step 3025, Loss: {'train/loss': 0.6020112633705139, 'train/grad_norm_pre_clip': 21.567293167114258, 'timing/data': 0.0006411196663975716, 'timing/model': 1.0170115856453776, 'learning_rate/qwen_vl_interface': 6.374630997215623e-06, 'learning_rate/action_model': 6.374630997215623e-05, 'epoch': 1.25}) +[2026-06-17 21:31:58,124][starVLA.training.train_starvla][INFO] - Step 3026, Loss: {'train/loss': 0.5548563003540039, 'train/grad_norm_pre_clip': 16.896242141723633, 'timing/data': 0.0011811256408691406, 'timing/model': 0.9260441428050399, 'learning_rate/qwen_vl_interface': 6.372529681713093e-06, 'learning_rate/action_model': 6.372529681713094e-05, 'epoch': 1.25}) +[2026-06-17 21:32:01,962][starVLA.training.train_starvla][INFO] - Step 3027, Loss: {'train/loss': 0.5202497243881226, 'train/grad_norm_pre_clip': 22.504587173461914, 'timing/data': 0.0005786903202533722, 'timing/model': 1.151884880848229, 'learning_rate/qwen_vl_interface': 6.370428133508953e-06, 'learning_rate/action_model': 6.370428133508953e-05, 'epoch': 1.25}) +[2026-06-17 21:32:05,922][starVLA.training.train_starvla][INFO] - Step 3028, Loss: {'train/loss': 0.5931941270828247, 'train/grad_norm_pre_clip': 15.188316345214844, 'timing/data': 0.0005384432151913643, 'timing/model': 1.2477990873157978, 'learning_rate/qwen_vl_interface': 6.368326353038852e-06, 'learning_rate/action_model': 6.368326353038853e-05, 'epoch': 1.25}) +[2026-06-17 21:32:10,070][starVLA.training.train_starvla][INFO] - Step 3029, Loss: {'train/loss': 0.271419882774353, 'train/grad_norm_pre_clip': 9.76347541809082, 'timing/data': 0.0005327900871634483, 'timing/model': 1.0455434694886208, 'learning_rate/qwen_vl_interface': 6.366224340738499e-06, 'learning_rate/action_model': 6.366224340738499e-05, 'epoch': 1.25}) +[2026-06-17 21:32:13,913][starVLA.training.train_starvla][INFO] - Step 3030, Loss: {'train/loss': 0.2525481581687927, 'train/grad_norm_pre_clip': 10.277583122253418, 'timing/data': 0.0009126691147685051, 'timing/model': 0.9417073139920831, 'learning_rate/qwen_vl_interface': 6.3641220970436355e-06, 'learning_rate/action_model': 6.364122097043635e-05, 'epoch': 1.25}) +[2026-06-17 21:32:17,739][starVLA.training.train_starvla][INFO] - Step 3031, Loss: {'train/loss': 0.16287970542907715, 'train/grad_norm_pre_clip': 7.447549819946289, 'timing/data': 0.0007978780195116997, 'timing/model': 0.9428130378946662, 'learning_rate/qwen_vl_interface': 6.362019622390065e-06, 'learning_rate/action_model': 6.362019622390066e-05, 'epoch': 1.25}) +[2026-06-17 21:32:21,951][starVLA.training.train_starvla][INFO] - Step 3032, Loss: {'train/loss': 0.5015390515327454, 'train/grad_norm_pre_clip': 17.636585235595703, 'timing/data': 0.0008280230686068535, 'timing/model': 1.2207336844876409, 'learning_rate/qwen_vl_interface': 6.359916917213629e-06, 'learning_rate/action_model': 6.359916917213629e-05, 'epoch': 1.25}) +[2026-06-17 21:32:26,010][starVLA.training.train_starvla][INFO] - Step 3033, Loss: {'train/loss': 0.3453293740749359, 'train/grad_norm_pre_clip': 12.545279502868652, 'timing/data': 0.0005150903016328812, 'timing/model': 1.159739937633276, 'learning_rate/qwen_vl_interface': 6.357813981950223e-06, 'learning_rate/action_model': 6.357813981950223e-05, 'epoch': 1.25}) +[2026-06-17 21:32:30,233][starVLA.training.train_starvla][INFO] - Step 3034, Loss: {'train/loss': 0.36849474906921387, 'train/grad_norm_pre_clip': 17.763216018676758, 'timing/data': 0.000514226034283638, 'timing/model': 1.163585047237575, 'learning_rate/qwen_vl_interface': 6.355710817035784e-06, 'learning_rate/action_model': 6.355710817035783e-05, 'epoch': 1.25}) +[2026-06-17 21:32:33,991][starVLA.training.train_starvla][INFO] - Step 3035, Loss: {'train/loss': 0.35072633624076843, 'train/grad_norm_pre_clip': 12.78366756439209, 'timing/data': 0.0007249899208545685, 'timing/model': 0.949786001816392, 'learning_rate/qwen_vl_interface': 6.353607422906304e-06, 'learning_rate/action_model': 6.353607422906304e-05, 'epoch': 1.25}) +[2026-06-17 21:32:37,918][starVLA.training.train_starvla][INFO] - Step 3036, Loss: {'train/loss': 0.6145392656326294, 'train/grad_norm_pre_clip': 15.928062438964844, 'timing/data': 0.0005228221416473389, 'timing/model': 0.9564654650166631, 'learning_rate/qwen_vl_interface': 6.351503799997819e-06, 'learning_rate/action_model': 6.351503799997818e-05, 'epoch': 1.25}) +[2026-06-17 21:32:42,324][starVLA.training.train_starvla][INFO] - Step 3037, Loss: {'train/loss': 0.791824221611023, 'train/grad_norm_pre_clip': 18.13823890686035, 'timing/data': 0.0028662988916039467, 'timing/model': 1.522191945463419, 'learning_rate/qwen_vl_interface': 6.349399948746409e-06, 'learning_rate/action_model': 6.349399948746409e-05, 'epoch': 1.25}) +[2026-06-17 21:32:49,095][starVLA.training.train_starvla][INFO] - Step 3038, Loss: {'train/loss': 0.36079952120780945, 'train/grad_norm_pre_clip': 14.53407096862793, 'timing/data': 0.00054131168872118, 'timing/model': 1.8383587915450335, 'learning_rate/qwen_vl_interface': 6.347295869588208e-06, 'learning_rate/action_model': 6.347295869588208e-05, 'epoch': 1.25}) +[2026-06-17 21:32:54,558][starVLA.training.train_starvla][INFO] - Step 3039, Loss: {'train/loss': 0.3394927680492401, 'train/grad_norm_pre_clip': 14.983205795288086, 'timing/data': 0.00047899410128593445, 'timing/model': 0.9757716907188296, 'learning_rate/qwen_vl_interface': 6.345191562959392e-06, 'learning_rate/action_model': 6.345191562959391e-05, 'epoch': 1.25}) +[2026-06-17 21:33:00,454][starVLA.training.train_starvla][INFO] - Step 3040, Loss: {'train/loss': 0.43265628814697266, 'train/grad_norm_pre_clip': 16.382431030273438, 'timing/data': 0.0005188258364796638, 'timing/model': 1.1472409199923277, 'learning_rate/qwen_vl_interface': 6.343087029296187e-06, 'learning_rate/action_model': 6.343087029296186e-05, 'epoch': 1.25}) +[2026-06-17 21:33:05,603][starVLA.training.train_starvla][INFO] - Step 3041, Loss: {'train/loss': 0.24274739623069763, 'train/grad_norm_pre_clip': 9.486474990844727, 'timing/data': 0.0005275951698422432, 'timing/model': 0.9587679952383041, 'learning_rate/qwen_vl_interface': 6.340982269034863e-06, 'learning_rate/action_model': 6.340982269034863e-05, 'epoch': 1.25}) +[2026-06-17 21:33:09,434][starVLA.training.train_starvla][INFO] - Step 3042, Loss: {'train/loss': 0.3324945569038391, 'train/grad_norm_pre_clip': 12.156652450561523, 'timing/data': 0.0005120579153299332, 'timing/model': 0.9544759755954146, 'learning_rate/qwen_vl_interface': 6.3388772826117415e-06, 'learning_rate/action_model': 6.338877282611742e-05, 'epoch': 1.25}) +[2026-06-17 21:33:13,638][starVLA.training.train_starvla][INFO] - Step 3043, Loss: {'train/loss': 0.30706146359443665, 'train/grad_norm_pre_clip': 12.87749195098877, 'timing/data': 0.0008385805413126945, 'timing/model': 0.9522938309237361, 'learning_rate/qwen_vl_interface': 6.3367720704631895e-06, 'learning_rate/action_model': 6.33677207046319e-05, 'epoch': 1.25}) +[2026-06-17 21:33:18,052][starVLA.training.train_starvla][INFO] - Step 3044, Loss: {'train/loss': 0.2570318579673767, 'train/grad_norm_pre_clip': 11.193216323852539, 'timing/data': 0.0008897213265299797, 'timing/model': 0.9647724274545908, 'learning_rate/qwen_vl_interface': 6.334666633025616e-06, 'learning_rate/action_model': 6.334666633025615e-05, 'epoch': 1.25}) +[2026-06-17 21:33:22,072][starVLA.training.train_starvla][INFO] - Step 3045, Loss: {'train/loss': 0.3329048752784729, 'train/grad_norm_pre_clip': 17.56333351135254, 'timing/data': 0.00048801302909851074, 'timing/model': 0.9969138521701097, 'learning_rate/qwen_vl_interface': 6.332560970735482e-06, 'learning_rate/action_model': 6.332560970735482e-05, 'epoch': 1.25}) +[2026-06-17 21:33:25,991][starVLA.training.train_starvla][INFO] - Step 3046, Loss: {'train/loss': 0.4405423104763031, 'train/grad_norm_pre_clip': 13.940836906433105, 'timing/data': 0.0004789968952536583, 'timing/model': 0.9470501476898789, 'learning_rate/qwen_vl_interface': 6.3304550840292945e-06, 'learning_rate/action_model': 6.330455084029294e-05, 'epoch': 1.25}) +[2026-06-17 21:33:29,824][starVLA.training.train_starvla][INFO] - Step 3047, Loss: {'train/loss': 0.3108673691749573, 'train/grad_norm_pre_clip': 13.44497299194336, 'timing/data': 0.0014761565253138542, 'timing/model': 1.104712420143187, 'learning_rate/qwen_vl_interface': 6.328348973343605e-06, 'learning_rate/action_model': 6.328348973343605e-05, 'epoch': 1.25}) +[2026-06-17 21:33:33,503][starVLA.training.train_starvla][INFO] - Step 3048, Loss: {'train/loss': 0.6288846731185913, 'train/grad_norm_pre_clip': 26.251461029052734, 'timing/data': 0.0008689351379871368, 'timing/model': 0.9613626021891832, 'learning_rate/qwen_vl_interface': 6.326242639115014e-06, 'learning_rate/action_model': 6.326242639115015e-05, 'epoch': 1.25}) +[2026-06-17 21:33:37,328][starVLA.training.train_starvla][INFO] - Step 3049, Loss: {'train/loss': 0.49446991086006165, 'train/grad_norm_pre_clip': 14.922490119934082, 'timing/data': 0.0005071349442005157, 'timing/model': 0.9536217972636223, 'learning_rate/qwen_vl_interface': 6.324136081780166e-06, 'learning_rate/action_model': 6.324136081780165e-05, 'epoch': 1.26}) +[2026-06-17 21:33:41,224][starVLA.training.train_starvla][INFO] - Step 3050, Loss: {'train/loss': 0.44803011417388916, 'train/grad_norm_pre_clip': 14.987142562866211, 'timing/data': 0.0004936736077070236, 'timing/model': 0.950021774508059, 'learning_rate/qwen_vl_interface': 6.322029301775751e-06, 'learning_rate/action_model': 6.32202930177575e-05, 'epoch': 1.26}) +[2026-06-17 21:33:45,033][starVLA.training.train_starvla][INFO] - Step 3051, Loss: {'train/loss': 0.8238691687583923, 'train/grad_norm_pre_clip': 18.25057029724121, 'timing/data': 0.0010122321546077728, 'timing/model': 0.9592271950095892, 'learning_rate/qwen_vl_interface': 6.3199222995385105e-06, 'learning_rate/action_model': 6.31992229953851e-05, 'epoch': 1.26}) +[2026-06-17 21:33:48,937][starVLA.training.train_starvla][INFO] - Step 3052, Loss: {'train/loss': 0.4270901381969452, 'train/grad_norm_pre_clip': 16.283044815063477, 'timing/data': 0.0008854838088154793, 'timing/model': 0.9484502598643303, 'learning_rate/qwen_vl_interface': 6.317815075505223e-06, 'learning_rate/action_model': 6.317815075505222e-05, 'epoch': 1.26}) +[2026-06-17 21:33:52,830][starVLA.training.train_starvla][INFO] - Step 3053, Loss: {'train/loss': 0.7687143087387085, 'train/grad_norm_pre_clip': 15.76276969909668, 'timing/data': 0.0004973765462636948, 'timing/model': 1.111991772428155, 'learning_rate/qwen_vl_interface': 6.315707630112724e-06, 'learning_rate/action_model': 6.315707630112723e-05, 'epoch': 1.26}) +[2026-06-17 21:33:56,485][starVLA.training.train_starvla][INFO] - Step 3054, Loss: {'train/loss': 0.9478498101234436, 'train/grad_norm_pre_clip': 19.465702056884766, 'timing/data': 0.0004956815391778946, 'timing/model': 0.9407055582851171, 'learning_rate/qwen_vl_interface': 6.313599963797884e-06, 'learning_rate/action_model': 6.313599963797884e-05, 'epoch': 1.26}) +[2026-06-17 21:34:00,436][starVLA.training.train_starvla][INFO] - Step 3055, Loss: {'train/loss': 0.5272205471992493, 'train/grad_norm_pre_clip': 38.58015823364258, 'timing/data': 0.0008717654272913933, 'timing/model': 0.9489796543493867, 'learning_rate/qwen_vl_interface': 6.311492076997629e-06, 'learning_rate/action_model': 6.311492076997629e-05, 'epoch': 1.26}) +[2026-06-17 21:34:04,342][starVLA.training.train_starvla][INFO] - Step 3056, Loss: {'train/loss': 0.6331977248191833, 'train/grad_norm_pre_clip': 15.607842445373535, 'timing/data': 0.0012465128675103188, 'timing/model': 0.9444136470556259, 'learning_rate/qwen_vl_interface': 6.309383970148922e-06, 'learning_rate/action_model': 6.309383970148922e-05, 'epoch': 1.26}) +[2026-06-17 21:34:08,298][starVLA.training.train_starvla][INFO] - Step 3057, Loss: {'train/loss': 0.2874947488307953, 'train/grad_norm_pre_clip': 12.577709197998047, 'timing/data': 0.0009007425978779793, 'timing/model': 0.9496646746993065, 'learning_rate/qwen_vl_interface': 6.3072756436887805e-06, 'learning_rate/action_model': 6.30727564368878e-05, 'epoch': 1.26}) +[2026-06-17 21:34:12,227][starVLA.training.train_starvla][INFO] - Step 3058, Loss: {'train/loss': 0.5241079330444336, 'train/grad_norm_pre_clip': 17.251976013183594, 'timing/data': 0.0004679867997765541, 'timing/model': 0.9610412465408444, 'learning_rate/qwen_vl_interface': 6.30516709805426e-06, 'learning_rate/action_model': 6.30516709805426e-05, 'epoch': 1.26}) +[2026-06-17 21:34:16,093][starVLA.training.train_starvla][INFO] - Step 3059, Loss: {'train/loss': 0.38788872957229614, 'train/grad_norm_pre_clip': 13.233301162719727, 'timing/data': 0.0004987167194485664, 'timing/model': 1.1472605839371681, 'learning_rate/qwen_vl_interface': 6.303058333682464e-06, 'learning_rate/action_model': 6.303058333682463e-05, 'epoch': 1.26}) +[2026-06-17 21:34:19,927][starVLA.training.train_starvla][INFO] - Step 3060, Loss: {'train/loss': 0.4226738512516022, 'train/grad_norm_pre_clip': 18.275489807128906, 'timing/data': 0.0007837023586034775, 'timing/model': 1.132392518222332, 'learning_rate/qwen_vl_interface': 6.300949351010542e-06, 'learning_rate/action_model': 6.300949351010543e-05, 'epoch': 1.26}) +[2026-06-17 21:34:23,680][starVLA.training.train_starvla][INFO] - Step 3061, Loss: {'train/loss': 0.451989084482193, 'train/grad_norm_pre_clip': 23.49483299255371, 'timing/data': 0.0004988107830286026, 'timing/model': 0.9393511982634664, 'learning_rate/qwen_vl_interface': 6.298840150475689e-06, 'learning_rate/action_model': 6.298840150475689e-05, 'epoch': 1.26}) +[2026-06-17 21:34:27,632][starVLA.training.train_starvla][INFO] - Step 3062, Loss: {'train/loss': 0.4432382583618164, 'train/grad_norm_pre_clip': 14.263094902038574, 'timing/data': 0.0009203189983963966, 'timing/model': 0.9442165205255151, 'learning_rate/qwen_vl_interface': 6.296730732515145e-06, 'learning_rate/action_model': 6.296730732515145e-05, 'epoch': 1.26}) +[2026-06-17 21:34:31,500][starVLA.training.train_starvla][INFO] - Step 3063, Loss: {'train/loss': 0.8401166796684265, 'train/grad_norm_pre_clip': 28.123851776123047, 'timing/data': 0.0012605087831616402, 'timing/model': 0.9504215205088258, 'learning_rate/qwen_vl_interface': 6.294621097566193e-06, 'learning_rate/action_model': 6.294621097566193e-05, 'epoch': 1.26}) +[2026-06-17 21:34:35,343][starVLA.training.train_starvla][INFO] - Step 3064, Loss: {'train/loss': 0.299374520778656, 'train/grad_norm_pre_clip': 8.911101341247559, 'timing/data': 0.0010470831766724586, 'timing/model': 0.9528700420632958, 'learning_rate/qwen_vl_interface': 6.292511246066163e-06, 'learning_rate/action_model': 6.292511246066163e-05, 'epoch': 1.26}) +[2026-06-17 21:34:40,068][starVLA.training.train_starvla][INFO] - Step 3065, Loss: {'train/loss': 0.6528241038322449, 'train/grad_norm_pre_clip': 16.0792236328125, 'timing/data': 0.0005550337955355644, 'timing/model': 1.2358566643670201, 'learning_rate/qwen_vl_interface': 6.290401178452431e-06, 'learning_rate/action_model': 6.29040117845243e-05, 'epoch': 1.26}) +[2026-06-17 21:34:43,909][starVLA.training.train_starvla][INFO] - Step 3066, Loss: {'train/loss': 0.45989149808883667, 'train/grad_norm_pre_clip': 16.813966751098633, 'timing/data': 0.0008777426555752754, 'timing/model': 1.1141921766102314, 'learning_rate/qwen_vl_interface': 6.288290895162414e-06, 'learning_rate/action_model': 6.288290895162413e-05, 'epoch': 1.26}) +[2026-06-17 21:34:47,602][starVLA.training.train_starvla][INFO] - Step 3067, Loss: {'train/loss': 0.6028746962547302, 'train/grad_norm_pre_clip': 16.03483772277832, 'timing/data': 0.0007003676146268845, 'timing/model': 0.9570132018998265, 'learning_rate/qwen_vl_interface': 6.286180396633579e-06, 'learning_rate/action_model': 6.286180396633579e-05, 'epoch': 1.26}) +[2026-06-17 21:34:51,431][starVLA.training.train_starvla][INFO] - Step 3068, Loss: {'train/loss': 0.2659189701080322, 'train/grad_norm_pre_clip': 10.71731185913086, 'timing/data': 0.0005658846348524094, 'timing/model': 0.9547854848206043, 'learning_rate/qwen_vl_interface': 6.28406968330343e-06, 'learning_rate/action_model': 6.28406968330343e-05, 'epoch': 1.26}) +[2026-06-17 21:34:55,361][starVLA.training.train_starvla][INFO] - Step 3069, Loss: {'train/loss': 0.46985509991645813, 'train/grad_norm_pre_clip': 16.084156036376953, 'timing/data': 0.00048170052468776703, 'timing/model': 0.9855867121368647, 'learning_rate/qwen_vl_interface': 6.281958755609526e-06, 'learning_rate/action_model': 6.281958755609526e-05, 'epoch': 1.26}) +[2026-06-17 21:35:01,212][starVLA.training.train_starvla][INFO] - Step 3070, Loss: {'train/loss': 0.30772295594215393, 'train/grad_norm_pre_clip': 9.900139808654785, 'timing/data': 0.0006199358031153679, 'timing/model': 1.034510189667344, 'learning_rate/qwen_vl_interface': 6.279847613989459e-06, 'learning_rate/action_model': 6.279847613989459e-05, 'epoch': 1.26}) +[2026-06-17 21:35:05,207][starVLA.training.train_starvla][INFO] - Step 3071, Loss: {'train/loss': 0.6988213658332825, 'train/grad_norm_pre_clip': 21.263803482055664, 'timing/data': 0.0005590254440903664, 'timing/model': 0.9454381233081222, 'learning_rate/qwen_vl_interface': 6.2777362588808745e-06, 'learning_rate/action_model': 6.277736258880874e-05, 'epoch': 1.26}) +[2026-06-17 21:35:09,126][starVLA.training.train_starvla][INFO] - Step 3072, Loss: {'train/loss': 0.25104382634162903, 'train/grad_norm_pre_clip': 9.440469741821289, 'timing/data': 0.0005341051146388054, 'timing/model': 1.1376033313572407, 'learning_rate/qwen_vl_interface': 6.275624690721458e-06, 'learning_rate/action_model': 6.275624690721458e-05, 'epoch': 1.26}) +[2026-06-17 21:35:13,115][starVLA.training.train_starvla][INFO] - Step 3073, Loss: {'train/loss': 0.7334064245223999, 'train/grad_norm_pre_clip': 17.61086654663086, 'timing/data': 0.0010848818346858025, 'timing/model': 1.2194764725863934, 'learning_rate/qwen_vl_interface': 6.27351290994894e-06, 'learning_rate/action_model': 6.27351290994894e-05, 'epoch': 1.26}) +[2026-06-17 21:35:16,779][starVLA.training.train_starvla][INFO] - Step 3074, Loss: {'train/loss': 0.3352561295032501, 'train/grad_norm_pre_clip': 12.559755325317383, 'timing/data': 0.00047929398715496063, 'timing/model': 0.9512936007231474, 'learning_rate/qwen_vl_interface': 6.271400917001092e-06, 'learning_rate/action_model': 6.271400917001092e-05, 'epoch': 1.27}) +[2026-06-17 21:35:20,647][starVLA.training.train_starvla][INFO] - Step 3075, Loss: {'train/loss': 0.39012715220451355, 'train/grad_norm_pre_clip': 14.737607955932617, 'timing/data': 0.0008464092388749123, 'timing/model': 0.9505416424944997, 'learning_rate/qwen_vl_interface': 6.269288712315737e-06, 'learning_rate/action_model': 6.269288712315737e-05, 'epoch': 1.27}) +[2026-06-17 21:35:24,654][starVLA.training.train_starvla][INFO] - Step 3076, Loss: {'train/loss': 0.6465054750442505, 'train/grad_norm_pre_clip': 14.685717582702637, 'timing/data': 0.0005071936175227165, 'timing/model': 0.9322918774560094, 'learning_rate/qwen_vl_interface': 6.267176296330734e-06, 'learning_rate/action_model': 6.267176296330733e-05, 'epoch': 1.27}) +[2026-06-17 21:35:28,443][starVLA.training.train_starvla][INFO] - Step 3077, Loss: {'train/loss': 0.3168768286705017, 'train/grad_norm_pre_clip': 10.841219902038574, 'timing/data': 0.0005079843103885651, 'timing/model': 0.931020624935627, 'learning_rate/qwen_vl_interface': 6.265063669483992e-06, 'learning_rate/action_model': 6.265063669483991e-05, 'epoch': 1.27}) +[2026-06-17 21:35:32,237][starVLA.training.train_starvla][INFO] - Step 3078, Loss: {'train/loss': 0.5008914470672607, 'train/grad_norm_pre_clip': 16.747631072998047, 'timing/data': 0.000893264077603817, 'timing/model': 0.9281967608258128, 'learning_rate/qwen_vl_interface': 6.262950832213457e-06, 'learning_rate/action_model': 6.262950832213456e-05, 'epoch': 1.27}) +[2026-06-17 21:35:36,156][starVLA.training.train_starvla][INFO] - Step 3079, Loss: {'train/loss': 0.3117063343524933, 'train/grad_norm_pre_clip': 13.40570068359375, 'timing/data': 0.0007486129179596901, 'timing/model': 1.1793913813307881, 'learning_rate/qwen_vl_interface': 6.260837784957125e-06, 'learning_rate/action_model': 6.260837784957126e-05, 'epoch': 1.27}) +[2026-06-17 21:35:40,089][starVLA.training.train_starvla][INFO] - Step 3080, Loss: {'train/loss': 0.41254690289497375, 'train/grad_norm_pre_clip': 15.016861915588379, 'timing/data': 0.0008173175156116486, 'timing/model': 1.1228741109371185, 'learning_rate/qwen_vl_interface': 6.258724528153034e-06, 'learning_rate/action_model': 6.258724528153033e-05, 'epoch': 1.27}) +[2026-06-17 21:35:43,783][starVLA.training.train_starvla][INFO] - Step 3081, Loss: {'train/loss': 0.6295416355133057, 'train/grad_norm_pre_clip': 14.550302505493164, 'timing/data': 0.0011271052062511444, 'timing/model': 0.95871417876333, 'learning_rate/qwen_vl_interface': 6.256611062239261e-06, 'learning_rate/action_model': 6.25661106223926e-05, 'epoch': 1.27}) +[2026-06-17 21:35:47,722][starVLA.training.train_starvla][INFO] - Step 3082, Loss: {'train/loss': 0.3229854106903076, 'train/grad_norm_pre_clip': 10.694398880004883, 'timing/data': 0.00047264061868190765, 'timing/model': 1.0388965969905257, 'learning_rate/qwen_vl_interface': 6.254497387653933e-06, 'learning_rate/action_model': 6.254497387653932e-05, 'epoch': 1.27}) +[2026-06-17 21:35:51,638][starVLA.training.train_starvla][INFO] - Step 3083, Loss: {'train/loss': 0.41789379715919495, 'train/grad_norm_pre_clip': 11.938393592834473, 'timing/data': 0.0007576439529657364, 'timing/model': 0.9533015191555023, 'learning_rate/qwen_vl_interface': 6.252383504835213e-06, 'learning_rate/action_model': 6.252383504835213e-05, 'epoch': 1.27}) +[2026-06-17 21:35:55,512][starVLA.training.train_starvla][INFO] - Step 3084, Loss: {'train/loss': 0.2233750969171524, 'train/grad_norm_pre_clip': 9.304337501525879, 'timing/data': 0.0005165953189134598, 'timing/model': 0.9608168294653296, 'learning_rate/qwen_vl_interface': 6.250269414221317e-06, 'learning_rate/action_model': 6.250269414221317e-05, 'epoch': 1.27}) +[2026-06-17 21:35:59,320][starVLA.training.train_starvla][INFO] - Step 3085, Loss: {'train/loss': 0.5562928915023804, 'train/grad_norm_pre_clip': 13.740606307983398, 'timing/data': 0.0008533606305718422, 'timing/model': 0.9823956359177828, 'learning_rate/qwen_vl_interface': 6.248155116250494e-06, 'learning_rate/action_model': 6.248155116250493e-05, 'epoch': 1.27}) +[2026-06-17 21:36:03,608][starVLA.training.train_starvla][INFO] - Step 3086, Loss: {'train/loss': 0.6695191264152527, 'train/grad_norm_pre_clip': 18.30713653564453, 'timing/data': 0.000494878739118576, 'timing/model': 1.137482806108892, 'learning_rate/qwen_vl_interface': 6.246040611361042e-06, 'learning_rate/action_model': 6.246040611361041e-05, 'epoch': 1.27}) +[2026-06-17 21:36:10,429][starVLA.training.train_starvla][INFO] - Step 3087, Loss: {'train/loss': 0.5868626236915588, 'train/grad_norm_pre_clip': 16.049837112426758, 'timing/data': 0.0007558977231383324, 'timing/model': 1.062324314378202, 'learning_rate/qwen_vl_interface': 6.2439258999912976e-06, 'learning_rate/action_model': 6.243925899991298e-05, 'epoch': 1.27}) +[2026-06-17 21:36:14,578][starVLA.training.train_starvla][INFO] - Step 3088, Loss: {'train/loss': 0.37696537375450134, 'train/grad_norm_pre_clip': 16.50178337097168, 'timing/data': 0.0011274032294750214, 'timing/model': 1.049373091198504, 'learning_rate/qwen_vl_interface': 6.241810982579646e-06, 'learning_rate/action_model': 6.241810982579646e-05, 'epoch': 1.27}) +[2026-06-17 21:36:18,549][starVLA.training.train_starvla][INFO] - Step 3089, Loss: {'train/loss': 0.3984393775463104, 'train/grad_norm_pre_clip': 14.258155822753906, 'timing/data': 0.0012085353955626488, 'timing/model': 0.9561247238889337, 'learning_rate/qwen_vl_interface': 6.239695859564511e-06, 'learning_rate/action_model': 6.23969585956451e-05, 'epoch': 1.27}) +[2026-06-17 21:36:22,636][starVLA.training.train_starvla][INFO] - Step 3090, Loss: {'train/loss': 0.39749374985694885, 'train/grad_norm_pre_clip': 13.463759422302246, 'timing/data': 0.0011846832931041718, 'timing/model': 0.9764627255499363, 'learning_rate/qwen_vl_interface': 6.2375805313843564e-06, 'learning_rate/action_model': 6.237580531384356e-05, 'epoch': 1.27}) +[2026-06-17 21:36:26,643][starVLA.training.train_starvla][INFO] - Step 3091, Loss: {'train/loss': 0.5980367064476013, 'train/grad_norm_pre_clip': 18.942874908447266, 'timing/data': 0.0009412160143256187, 'timing/model': 0.9286496769636869, 'learning_rate/qwen_vl_interface': 6.2354649984777e-06, 'learning_rate/action_model': 6.235464998477699e-05, 'epoch': 1.27}) +[2026-06-17 21:36:30,526][starVLA.training.train_starvla][INFO] - Step 3092, Loss: {'train/loss': 0.25844693183898926, 'train/grad_norm_pre_clip': 11.843240737915039, 'timing/data': 0.0004907762631773949, 'timing/model': 0.993663433007896, 'learning_rate/qwen_vl_interface': 6.2333492612830835e-06, 'learning_rate/action_model': 6.233349261283083e-05, 'epoch': 1.27}) +[2026-06-17 21:36:34,472][starVLA.training.train_starvla][INFO] - Step 3093, Loss: {'train/loss': 0.356299489736557, 'train/grad_norm_pre_clip': 16.8057918548584, 'timing/data': 0.00047670304775238037, 'timing/model': 1.2156802471727133, 'learning_rate/qwen_vl_interface': 6.231233320239112e-06, 'learning_rate/action_model': 6.231233320239112e-05, 'epoch': 1.27}) +[2026-06-17 21:36:38,117][starVLA.training.train_starvla][INFO] - Step 3094, Loss: {'train/loss': 0.447139173746109, 'train/grad_norm_pre_clip': 14.616798400878906, 'timing/data': 0.0004907231777906418, 'timing/model': 0.942060086876154, 'learning_rate/qwen_vl_interface': 6.229117175784414e-06, 'learning_rate/action_model': 6.229117175784413e-05, 'epoch': 1.27}) +[2026-06-17 21:36:42,082][starVLA.training.train_starvla][INFO] - Step 3095, Loss: {'train/loss': 0.3869439363479614, 'train/grad_norm_pre_clip': 17.03475570678711, 'timing/data': 0.000988188199698925, 'timing/model': 0.9438782557845116, 'learning_rate/qwen_vl_interface': 6.227000828357672e-06, 'learning_rate/action_model': 6.227000828357672e-05, 'epoch': 1.27}) +[2026-06-17 21:36:45,944][starVLA.training.train_starvla][INFO] - Step 3096, Loss: {'train/loss': 0.29423174262046814, 'train/grad_norm_pre_clip': 15.128934860229492, 'timing/data': 0.0006287097930908203, 'timing/model': 0.9478671485558152, 'learning_rate/qwen_vl_interface': 6.2248842783976085e-06, 'learning_rate/action_model': 6.224884278397608e-05, 'epoch': 1.27}) +[2026-06-17 21:36:49,871][starVLA.training.train_starvla][INFO] - Step 3097, Loss: {'train/loss': 0.43145808577537537, 'train/grad_norm_pre_clip': 15.347152709960938, 'timing/data': 0.0005091242492198944, 'timing/model': 0.9389099897816777, 'learning_rate/qwen_vl_interface': 6.222767526342983e-06, 'learning_rate/action_model': 6.222767526342983e-05, 'epoch': 1.27}) +[2026-06-17 21:36:54,014][starVLA.training.train_starvla][INFO] - Step 3098, Loss: {'train/loss': 0.4889257848262787, 'train/grad_norm_pre_clip': 19.257278442382812, 'timing/data': 0.0009601237252354622, 'timing/model': 1.2169636208564043, 'learning_rate/qwen_vl_interface': 6.220650572632605e-06, 'learning_rate/action_model': 6.220650572632604e-05, 'epoch': 1.28}) +[2026-06-17 21:36:59,668][starVLA.training.train_starvla][INFO] - Step 3099, Loss: {'train/loss': 0.4641502797603607, 'train/grad_norm_pre_clip': 17.288925170898438, 'timing/data': 0.0006692158058285713, 'timing/model': 1.8044235445559025, 'learning_rate/qwen_vl_interface': 6.218533417705318e-06, 'learning_rate/action_model': 6.218533417705318e-05, 'epoch': 1.28}) +[2026-06-17 21:39:04,263][starVLA.training.train_starvla][INFO] - Step 3100, Loss: {'train/loss': 0.5959858298301697, 'train/grad_norm_pre_clip': 19.63875961303711, 'eval/loss': 0.48946431279182434, 'eval/latency_6/loss': 0.48946431279182434, 'eval/demon_fire_precision': 0.9380996595481275, 'eval/demon_fire_recall': 0.9796380090497737, 'eval/demon_fire_f1': 0.9584189723320158, 'eval/demon_fire_rate_model': 0.8989983305509182, 'eval/demon_fire_rate_teacher': 0.860879243183083, 'eval/demon_left_precision': 0.8842105263157894, 'eval/demon_left_recall': 0.5894736842105263, 'eval/demon_left_f1': 0.7073684210526315, 'eval/demon_left_rate_model': 0.10573177518085698, 'eval/demon_left_rate_teacher': 0.15859766277128548, 'eval/demon_right_precision': 0.7723704866562009, 'eval/demon_right_recall': 0.7557603686635944, 'eval/demon_right_f1': 0.7639751552795029, 'eval/demon_right_rate_model': 0.17723984418475236, 'eval/demon_right_rate_teacher': 0.18113522537562604, 'eval/demon_move_f1': 0.7356717881660673, 'eval/demon_cc_f1': 0.8786110446696758, 'eval/latency_6/demon_fire_precision': 0.9380996595481275, 'eval/latency_6/demon_fire_recall': 0.9796380090497737, 'eval/latency_6/demon_fire_f1': 0.9584189723320158, 'eval/latency_6/demon_fire_rate_model': 0.8989983305509182, 'eval/latency_6/demon_fire_rate_teacher': 0.860879243183083, 'eval/latency_6/demon_left_precision': 0.8842105263157894, 'eval/latency_6/demon_left_recall': 0.5894736842105263, 'eval/latency_6/demon_left_f1': 0.7073684210526315, 'eval/latency_6/demon_left_rate_model': 0.10573177518085698, 'eval/latency_6/demon_left_rate_teacher': 0.15859766277128548, 'eval/latency_6/demon_right_precision': 0.7723704866562009, 'eval/latency_6/demon_right_recall': 0.7557603686635944, 'eval/latency_6/demon_right_f1': 0.7639751552795029, 'eval/latency_6/demon_right_rate_model': 0.17723984418475236, 'eval/latency_6/demon_right_rate_teacher': 0.18113522537562604, 'eval/latency_6/demon_move_f1': 0.7356717881660673, 'eval/latency_6/demon_cc_f1': 0.8786110446696758, 'timing/data': 0.0006933314725756645, 'timing/model': 1.087134249508381, 'learning_rate/qwen_vl_interface': 6.216416062000012e-06, 'learning_rate/action_model': 6.216416062000013e-05, 'epoch': 1.28}) +[2026-06-17 21:39:04,266][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_3100_state +[2026-06-17 21:39:04,271][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 21:39:11,301][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_3100_state/model.safetensors +[2026-06-17 21:39:26,906][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_3100_state/optimizer.bin +[2026-06-17 21:39:26,907][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_3100_state/sampler.bin +[2026-06-17 21:39:26,907][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_3100_state/sampler_1.bin +[2026-06-17 21:39:26,918][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_3100_state/random_states_0.pkl +[2026-06-17 21:39:26,920][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3100 +[2026-06-17 21:39:26,921][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 21:39:26,931][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 21:39:30,874][starVLA.training.train_starvla][INFO] - Step 3101, Loss: {'train/loss': 0.5885071158409119, 'train/grad_norm_pre_clip': 12.377059936523438, 'timing/data': 0.0007500350475311279, 'timing/model': 1.1556497300043702, 'learning_rate/qwen_vl_interface': 6.214298505955615e-06, 'learning_rate/action_model': 6.214298505955614e-05, 'epoch': 1.28}) +[2026-06-17 21:39:34,629][starVLA.training.train_starvla][INFO] - Step 3102, Loss: {'train/loss': 0.33071744441986084, 'train/grad_norm_pre_clip': 16.06389617919922, 'timing/data': 0.0008437596261501312, 'timing/model': 1.0965131083503366, 'learning_rate/qwen_vl_interface': 6.2121807500111e-06, 'learning_rate/action_model': 6.212180750011099e-05, 'epoch': 1.28}) +[2026-06-17 21:39:38,225][starVLA.training.train_starvla][INFO] - Step 3103, Loss: {'train/loss': 0.24267584085464478, 'train/grad_norm_pre_clip': 10.049307823181152, 'timing/data': 0.0011664340272545815, 'timing/model': 0.939383308403194, 'learning_rate/qwen_vl_interface': 6.210062794605482e-06, 'learning_rate/action_model': 6.21006279460548e-05, 'epoch': 1.28}) +[2026-06-17 21:39:42,124][starVLA.training.train_starvla][INFO] - Step 3104, Loss: {'train/loss': 0.4414961636066437, 'train/grad_norm_pre_clip': 10.760610580444336, 'timing/data': 0.0007013129070401192, 'timing/model': 0.9533311454579234, 'learning_rate/qwen_vl_interface': 6.207944640177812e-06, 'learning_rate/action_model': 6.207944640177812e-05, 'epoch': 1.28}) +[2026-06-17 21:39:45,931][starVLA.training.train_starvla][INFO] - Step 3105, Loss: {'train/loss': 0.29808512330055237, 'train/grad_norm_pre_clip': 14.293020248413086, 'timing/data': 0.0011732401326298714, 'timing/model': 0.9429439669474959, 'learning_rate/qwen_vl_interface': 6.205826287167188e-06, 'learning_rate/action_model': 6.205826287167188e-05, 'epoch': 1.28}) +[2026-06-17 21:39:49,827][starVLA.training.train_starvla][INFO] - Step 3106, Loss: {'train/loss': 0.4600362181663513, 'train/grad_norm_pre_clip': 14.01563549041748, 'timing/data': 0.0005675200372934341, 'timing/model': 0.9480037884786725, 'learning_rate/qwen_vl_interface': 6.203707736012745e-06, 'learning_rate/action_model': 6.203707736012745e-05, 'epoch': 1.28}) +[2026-06-17 21:39:53,895][starVLA.training.train_starvla][INFO] - Step 3107, Loss: {'train/loss': 0.3724772334098816, 'train/grad_norm_pre_clip': 13.787768363952637, 'timing/data': 0.000547993928194046, 'timing/model': 1.016716075129807, 'learning_rate/qwen_vl_interface': 6.201588987153665e-06, 'learning_rate/action_model': 6.201588987153664e-05, 'epoch': 1.28}) +[2026-06-17 21:39:57,902][starVLA.training.train_starvla][INFO] - Step 3108, Loss: {'train/loss': 0.47501862049102783, 'train/grad_norm_pre_clip': 17.368425369262695, 'timing/data': 0.0005582412704825401, 'timing/model': 1.1693112831562757, 'learning_rate/qwen_vl_interface': 6.19947004102916e-06, 'learning_rate/action_model': 6.199470041029159e-05, 'epoch': 1.28}) +[2026-06-17 21:40:01,811][starVLA.training.train_starvla][INFO] - Step 3109, Loss: {'train/loss': 0.5469932556152344, 'train/grad_norm_pre_clip': 18.015363693237305, 'timing/data': 0.000532468780875206, 'timing/model': 1.1868663830682635, 'learning_rate/qwen_vl_interface': 6.197350898078496e-06, 'learning_rate/action_model': 6.197350898078495e-05, 'epoch': 1.28}) +[2026-06-17 21:40:07,594][starVLA.training.train_starvla][INFO] - Step 3110, Loss: {'train/loss': 0.4426034390926361, 'train/grad_norm_pre_clip': 13.510943412780762, 'timing/data': 0.0005100825801491737, 'timing/model': 0.9394549848511815, 'learning_rate/qwen_vl_interface': 6.195231558740972e-06, 'learning_rate/action_model': 6.195231558740971e-05, 'epoch': 1.28}) +[2026-06-17 21:40:12,370][starVLA.training.train_starvla][INFO] - Step 3111, Loss: {'train/loss': 0.5130043029785156, 'train/grad_norm_pre_clip': 18.936981201171875, 'timing/data': 0.0005137817934155464, 'timing/model': 0.9590882603079081, 'learning_rate/qwen_vl_interface': 6.1931120234559275e-06, 'learning_rate/action_model': 6.193112023455927e-05, 'epoch': 1.28}) +[2026-06-17 21:40:17,942][starVLA.training.train_starvla][INFO] - Step 3112, Loss: {'train/loss': 0.32970359921455383, 'train/grad_norm_pre_clip': 11.706500053405762, 'timing/data': 0.0012210030108690262, 'timing/model': 0.9425531169399619, 'learning_rate/qwen_vl_interface': 6.190992292662747e-06, 'learning_rate/action_model': 6.190992292662746e-05, 'epoch': 1.28}) +[2026-06-17 21:40:21,766][starVLA.training.train_starvla][INFO] - Step 3113, Loss: {'train/loss': 0.3557584583759308, 'train/grad_norm_pre_clip': 12.137504577636719, 'timing/data': 0.000539800152182579, 'timing/model': 0.9314935421571136, 'learning_rate/qwen_vl_interface': 6.18887236680085e-06, 'learning_rate/action_model': 6.18887236680085e-05, 'epoch': 1.28}) +[2026-06-17 21:40:25,791][starVLA.training.train_starvla][INFO] - Step 3114, Loss: {'train/loss': 0.5508637428283691, 'train/grad_norm_pre_clip': 17.32253646850586, 'timing/data': 0.0008682794868946075, 'timing/model': 0.9626338090747595, 'learning_rate/qwen_vl_interface': 6.186752246309702e-06, 'learning_rate/action_model': 6.186752246309702e-05, 'epoch': 1.28}) +[2026-06-17 21:40:29,665][starVLA.training.train_starvla][INFO] - Step 3115, Loss: {'train/loss': 0.4719947576522827, 'train/grad_norm_pre_clip': 16.40925407409668, 'timing/data': 0.0004959935322403908, 'timing/model': 1.1544676488265395, 'learning_rate/qwen_vl_interface': 6.184631931628806e-06, 'learning_rate/action_model': 6.184631931628806e-05, 'epoch': 1.28}) +[2026-06-17 21:40:33,570][starVLA.training.train_starvla][INFO] - Step 3116, Loss: {'train/loss': 0.5854586362838745, 'train/grad_norm_pre_clip': 19.612455368041992, 'timing/data': 0.0009792055934667587, 'timing/model': 1.1541462372988462, 'learning_rate/qwen_vl_interface': 6.1825114231977036e-06, 'learning_rate/action_model': 6.182511423197704e-05, 'epoch': 1.28}) +[2026-06-17 21:40:37,404][starVLA.training.train_starvla][INFO] - Step 3117, Loss: {'train/loss': 0.41102567315101624, 'train/grad_norm_pre_clip': 31.570133209228516, 'timing/data': 0.0005079545080661774, 'timing/model': 0.9349985923618078, 'learning_rate/qwen_vl_interface': 6.18039072145598e-06, 'learning_rate/action_model': 6.18039072145598e-05, 'epoch': 1.28}) +[2026-06-17 21:40:41,230][starVLA.training.train_starvla][INFO] - Step 3118, Loss: {'train/loss': 0.6564916372299194, 'train/grad_norm_pre_clip': 18.19565200805664, 'timing/data': 0.0005163745954632759, 'timing/model': 0.9391702134162188, 'learning_rate/qwen_vl_interface': 6.17826982684326e-06, 'learning_rate/action_model': 6.178269826843259e-05, 'epoch': 1.28}) +[2026-06-17 21:40:45,009][starVLA.training.train_starvla][INFO] - Step 3119, Loss: {'train/loss': 0.3545694351196289, 'train/grad_norm_pre_clip': 15.073020935058594, 'timing/data': 0.000568317249417305, 'timing/model': 0.9346463400870562, 'learning_rate/qwen_vl_interface': 6.176148739799205e-06, 'learning_rate/action_model': 6.176148739799205e-05, 'epoch': 1.28}) +[2026-06-17 21:40:48,771][starVLA.training.train_starvla][INFO] - Step 3120, Loss: {'train/loss': 0.2368459552526474, 'train/grad_norm_pre_clip': 9.72630786895752, 'timing/data': 0.0005736630409955978, 'timing/model': 0.9261140134185553, 'learning_rate/qwen_vl_interface': 6.1740274607635195e-06, 'learning_rate/action_model': 6.174027460763519e-05, 'epoch': 1.28}) +[2026-06-17 21:40:52,646][starVLA.training.train_starvla][INFO] - Step 3121, Loss: {'train/loss': 0.4875977337360382, 'train/grad_norm_pre_clip': 14.399184226989746, 'timing/data': 0.0005348008126020432, 'timing/model': 0.9508782671764493, 'learning_rate/qwen_vl_interface': 6.1719059901759474e-06, 'learning_rate/action_model': 6.171905990175947e-05, 'epoch': 1.28}) +[2026-06-17 21:40:56,548][starVLA.training.train_starvla][INFO] - Step 3122, Loss: {'train/loss': 0.41548094153404236, 'train/grad_norm_pre_clip': 15.941156387329102, 'timing/data': 0.0007804986089468002, 'timing/model': 1.1574940672144294, 'learning_rate/qwen_vl_interface': 6.169784328476272e-06, 'learning_rate/action_model': 6.169784328476271e-05, 'epoch': 1.29}) +[2026-06-17 21:41:00,147][starVLA.training.train_starvla][INFO] - Step 3123, Loss: {'train/loss': 0.41656965017318726, 'train/grad_norm_pre_clip': 13.984360694885254, 'timing/data': 0.0005044844001531601, 'timing/model': 0.9357805866748095, 'learning_rate/qwen_vl_interface': 6.167662476104314e-06, 'learning_rate/action_model': 6.167662476104314e-05, 'epoch': 1.29}) +[2026-06-17 21:41:03,933][starVLA.training.train_starvla][INFO] - Step 3124, Loss: {'train/loss': 0.2147369384765625, 'train/grad_norm_pre_clip': 10.728362083435059, 'timing/data': 0.0005385428667068481, 'timing/model': 0.93780583050102, 'learning_rate/qwen_vl_interface': 6.165540433499939e-06, 'learning_rate/action_model': 6.165540433499939e-05, 'epoch': 1.29}) +[2026-06-17 21:41:07,736][starVLA.training.train_starvla][INFO] - Step 3125, Loss: {'train/loss': 0.3607694208621979, 'train/grad_norm_pre_clip': 16.048322677612305, 'timing/data': 0.0007070284336805344, 'timing/model': 0.9426395138725638, 'learning_rate/qwen_vl_interface': 6.163418201103045e-06, 'learning_rate/action_model': 6.163418201103045e-05, 'epoch': 1.29}) +[2026-06-17 21:41:11,720][starVLA.training.train_starvla][INFO] - Step 3126, Loss: {'train/loss': 0.2615238428115845, 'train/grad_norm_pre_clip': 14.902793884277344, 'timing/data': 0.0005220966413617134, 'timing/model': 0.9604167826473713, 'learning_rate/qwen_vl_interface': 6.161295779353578e-06, 'learning_rate/action_model': 6.161295779353577e-05, 'epoch': 1.29}) +[2026-06-17 21:41:15,520][starVLA.training.train_starvla][INFO] - Step 3127, Loss: {'train/loss': 0.29277753829956055, 'train/grad_norm_pre_clip': 12.484980583190918, 'timing/data': 0.0007858248427510262, 'timing/model': 0.9310992285609245, 'learning_rate/qwen_vl_interface': 6.159173168691512e-06, 'learning_rate/action_model': 6.159173168691512e-05, 'epoch': 1.29}) +[2026-06-17 21:41:19,328][starVLA.training.train_starvla][INFO] - Step 3128, Loss: {'train/loss': 0.2351154237985611, 'train/grad_norm_pre_clip': 12.728191375732422, 'timing/data': 0.0005073482170701027, 'timing/model': 1.101638755761087, 'learning_rate/qwen_vl_interface': 6.157050369556871e-06, 'learning_rate/action_model': 6.157050369556872e-05, 'epoch': 1.29}) +[2026-06-17 21:41:23,062][starVLA.training.train_starvla][INFO] - Step 3129, Loss: {'train/loss': 0.5189117789268494, 'train/grad_norm_pre_clip': 15.17951488494873, 'timing/data': 0.0008224453777074814, 'timing/model': 1.0591438170522451, 'learning_rate/qwen_vl_interface': 6.154927382389714e-06, 'learning_rate/action_model': 6.154927382389714e-05, 'epoch': 1.29}) +[2026-06-17 21:41:26,693][starVLA.training.train_starvla][INFO] - Step 3130, Loss: {'train/loss': 0.25707903504371643, 'train/grad_norm_pre_clip': 12.554551124572754, 'timing/data': 0.0004996396601200104, 'timing/model': 0.9341671606525779, 'learning_rate/qwen_vl_interface': 6.152804207630135e-06, 'learning_rate/action_model': 6.152804207630136e-05, 'epoch': 1.29}) +[2026-06-17 21:41:30,496][starVLA.training.train_starvla][INFO] - Step 3131, Loss: {'train/loss': 0.5283496379852295, 'train/grad_norm_pre_clip': 17.5479679107666, 'timing/data': 0.0009055240079760551, 'timing/model': 0.9431841904297471, 'learning_rate/qwen_vl_interface': 6.150680845718274e-06, 'learning_rate/action_model': 6.150680845718274e-05, 'epoch': 1.29}) +[2026-06-17 21:41:34,330][starVLA.training.train_starvla][INFO] - Step 3132, Loss: {'train/loss': 0.3978661000728607, 'train/grad_norm_pre_clip': 11.701305389404297, 'timing/data': 0.0005097184330224991, 'timing/model': 0.9482998633757234, 'learning_rate/qwen_vl_interface': 6.148557297094303e-06, 'learning_rate/action_model': 6.148557297094303e-05, 'epoch': 1.29}) +[2026-06-17 21:41:38,474][starVLA.training.train_starvla][INFO] - Step 3133, Loss: {'train/loss': 0.4219891428947449, 'train/grad_norm_pre_clip': 16.060331344604492, 'timing/data': 0.0007019378244876862, 'timing/model': 0.9894295949488878, 'learning_rate/qwen_vl_interface': 6.1464335621984375e-06, 'learning_rate/action_model': 6.146433562198437e-05, 'epoch': 1.29}) +[2026-06-17 21:41:42,313][starVLA.training.train_starvla][INFO] - Step 3134, Loss: {'train/loss': 0.4489043951034546, 'train/grad_norm_pre_clip': 15.843893051147461, 'timing/data': 0.0007955655455589294, 'timing/model': 0.9542028037831187, 'learning_rate/qwen_vl_interface': 6.14430964147093e-06, 'learning_rate/action_model': 6.14430964147093e-05, 'epoch': 1.29}) +[2026-06-17 21:41:46,181][starVLA.training.train_starvla][INFO] - Step 3135, Loss: {'train/loss': 0.5090508460998535, 'train/grad_norm_pre_clip': 17.476953506469727, 'timing/data': 0.0004995381459593773, 'timing/model': 1.1375084454193711, 'learning_rate/qwen_vl_interface': 6.1421855353520735e-06, 'learning_rate/action_model': 6.142185535352073e-05, 'epoch': 1.29}) +[2026-06-17 21:41:50,087][starVLA.training.train_starvla][INFO] - Step 3136, Loss: {'train/loss': 0.4019108712673187, 'train/grad_norm_pre_clip': 15.601682662963867, 'timing/data': 0.0005192412063479424, 'timing/model': 1.1575906248763204, 'learning_rate/qwen_vl_interface': 6.140061244282194e-06, 'learning_rate/action_model': 6.140061244282194e-05, 'epoch': 1.29}) +[2026-06-17 21:41:54,047][starVLA.training.train_starvla][INFO] - Step 3137, Loss: {'train/loss': 0.5476713180541992, 'train/grad_norm_pre_clip': 18.08067512512207, 'timing/data': 0.0010009678080677986, 'timing/model': 0.9416622780263424, 'learning_rate/qwen_vl_interface': 6.137936768701659e-06, 'learning_rate/action_model': 6.137936768701659e-05, 'epoch': 1.29}) +[2026-06-17 21:41:58,108][starVLA.training.train_starvla][INFO] - Step 3138, Loss: {'train/loss': 0.6296753287315369, 'train/grad_norm_pre_clip': 18.10198402404785, 'timing/data': 0.0005015702918171883, 'timing/model': 0.953489069826901, 'learning_rate/qwen_vl_interface': 6.135812109050879e-06, 'learning_rate/action_model': 6.135812109050879e-05, 'epoch': 1.29}) +[2026-06-17 21:42:01,950][starVLA.training.train_starvla][INFO] - Step 3139, Loss: {'train/loss': 0.23282858729362488, 'train/grad_norm_pre_clip': 11.65040397644043, 'timing/data': 0.0007811663672327995, 'timing/model': 0.9547362336888909, 'learning_rate/qwen_vl_interface': 6.1336872657702936e-06, 'learning_rate/action_model': 6.133687265770293e-05, 'epoch': 1.29}) +[2026-06-17 21:42:07,683][starVLA.training.train_starvla][INFO] - Step 3140, Loss: {'train/loss': 0.7294315695762634, 'train/grad_norm_pre_clip': 17.677621841430664, 'timing/data': 0.0008855462074279785, 'timing/model': 1.102215969003737, 'learning_rate/qwen_vl_interface': 6.131562239300389e-06, 'learning_rate/action_model': 6.13156223930039e-05, 'epoch': 1.29}) +[2026-06-17 21:42:11,549][starVLA.training.train_starvla][INFO] - Step 3141, Loss: {'train/loss': 0.7409132122993469, 'train/grad_norm_pre_clip': 20.70964241027832, 'timing/data': 0.0010902825742959976, 'timing/model': 0.9610191313549876, 'learning_rate/qwen_vl_interface': 6.129437030081683e-06, 'learning_rate/action_model': 6.129437030081682e-05, 'epoch': 1.29}) +[2026-06-17 21:42:15,447][starVLA.training.train_starvla][INFO] - Step 3142, Loss: {'train/loss': 0.48537272214889526, 'train/grad_norm_pre_clip': 14.025146484375, 'timing/data': 0.0007015783339738846, 'timing/model': 1.144339413382113, 'learning_rate/qwen_vl_interface': 6.1273116385547336e-06, 'learning_rate/action_model': 6.127311638554734e-05, 'epoch': 1.29}) +[2026-06-17 21:42:19,134][starVLA.training.train_starvla][INFO] - Step 3143, Loss: {'train/loss': 0.5573550462722778, 'train/grad_norm_pre_clip': 20.394336700439453, 'timing/data': 0.0004903199151158333, 'timing/model': 0.950393483042717, 'learning_rate/qwen_vl_interface': 6.125186065160138e-06, 'learning_rate/action_model': 6.125186065160138e-05, 'epoch': 1.29}) +[2026-06-17 21:42:23,009][starVLA.training.train_starvla][INFO] - Step 3144, Loss: {'train/loss': 0.4673112630844116, 'train/grad_norm_pre_clip': 19.804697036743164, 'timing/data': 0.0006064446642994881, 'timing/model': 0.9478336479514837, 'learning_rate/qwen_vl_interface': 6.123060310338529e-06, 'learning_rate/action_model': 6.123060310338529e-05, 'epoch': 1.29}) +[2026-06-17 21:42:26,897][starVLA.training.train_starvla][INFO] - Step 3145, Loss: {'train/loss': 0.35323959589004517, 'train/grad_norm_pre_clip': 12.844369888305664, 'timing/data': 0.0005223499611020088, 'timing/model': 0.9449573354795575, 'learning_rate/qwen_vl_interface': 6.120934374530579e-06, 'learning_rate/action_model': 6.12093437453058e-05, 'epoch': 1.29}) +[2026-06-17 21:42:30,708][starVLA.training.train_starvla][INFO] - Step 3146, Loss: {'train/loss': 0.559391975402832, 'train/grad_norm_pre_clip': 17.951358795166016, 'timing/data': 0.0005574570968747139, 'timing/model': 0.9538491843268275, 'learning_rate/qwen_vl_interface': 6.118808258176996e-06, 'learning_rate/action_model': 6.118808258176995e-05, 'epoch': 1.3}) +[2026-06-17 21:42:34,553][starVLA.training.train_starvla][INFO] - Step 3147, Loss: {'train/loss': 0.29924729466438293, 'train/grad_norm_pre_clip': 12.52761459350586, 'timing/data': 0.0005155252292752266, 'timing/model': 0.9783468386158347, 'learning_rate/qwen_vl_interface': 6.116681961718525e-06, 'learning_rate/action_model': 6.116681961718524e-05, 'epoch': 1.3}) +[2026-06-17 21:42:38,425][starVLA.training.train_starvla][INFO] - Step 3148, Loss: {'train/loss': 0.5407240390777588, 'train/grad_norm_pre_clip': 677.0355224609375, 'timing/data': 0.0007167933508753777, 'timing/model': 1.1336763221770525, 'learning_rate/qwen_vl_interface': 6.114555485595952e-06, 'learning_rate/action_model': 6.114555485595952e-05, 'epoch': 1.3}) +[2026-06-17 21:42:42,263][starVLA.training.train_starvla][INFO] - Step 3149, Loss: {'train/loss': 0.4158081114292145, 'train/grad_norm_pre_clip': 18.345252990722656, 'timing/data': 0.0007280018180608749, 'timing/model': 1.0750241167843342, 'learning_rate/qwen_vl_interface': 6.112428830250096e-06, 'learning_rate/action_model': 6.112428830250095e-05, 'epoch': 1.3}) +[2026-06-17 21:42:45,951][starVLA.training.train_starvla][INFO] - Step 3150, Loss: {'train/loss': 0.7293533682823181, 'train/grad_norm_pre_clip': 17.191028594970703, 'timing/data': 0.0014411583542823792, 'timing/model': 0.9448931030929089, 'learning_rate/qwen_vl_interface': 6.110301996121816e-06, 'learning_rate/action_model': 6.110301996121816e-05, 'epoch': 1.3}) +[2026-06-17 21:42:49,835][starVLA.training.train_starvla][INFO] - Step 3151, Loss: {'train/loss': 0.31486475467681885, 'train/grad_norm_pre_clip': 10.50019359588623, 'timing/data': 0.0006051650270819664, 'timing/model': 0.9444378698244691, 'learning_rate/qwen_vl_interface': 6.108174983652007e-06, 'learning_rate/action_model': 6.108174983652006e-05, 'epoch': 1.3}) +[2026-06-17 21:42:53,745][starVLA.training.train_starvla][INFO] - Step 3152, Loss: {'train/loss': 0.24387887120246887, 'train/grad_norm_pre_clip': 11.120945930480957, 'timing/data': 0.0008255597203969955, 'timing/model': 0.9502695482224226, 'learning_rate/qwen_vl_interface': 6.106047793281598e-06, 'learning_rate/action_model': 6.106047793281598e-05, 'epoch': 1.3}) +[2026-06-17 21:42:57,706][starVLA.training.train_starvla][INFO] - Step 3153, Loss: {'train/loss': 0.5019606351852417, 'train/grad_norm_pre_clip': 13.203055381774902, 'timing/data': 0.0009087081998586655, 'timing/model': 1.0225685602054, 'learning_rate/qwen_vl_interface': 6.1039204254515626e-06, 'learning_rate/action_model': 6.103920425451563e-05, 'epoch': 1.3}) +[2026-06-17 21:43:01,567][starVLA.training.train_starvla][INFO] - Step 3154, Loss: {'train/loss': 0.6498358845710754, 'train/grad_norm_pre_clip': 22.359357833862305, 'timing/data': 0.0005273614078760147, 'timing/model': 0.9378198813647032, 'learning_rate/qwen_vl_interface': 6.101792880602901e-06, 'learning_rate/action_model': 6.101792880602901e-05, 'epoch': 1.3}) +[2026-06-17 21:43:05,580][starVLA.training.train_starvla][INFO] - Step 3155, Loss: {'train/loss': 0.468624085187912, 'train/grad_norm_pre_clip': 16.177555084228516, 'timing/data': 0.000686895102262497, 'timing/model': 1.139086346141994, 'learning_rate/qwen_vl_interface': 6.099665159176663e-06, 'learning_rate/action_model': 6.099665159176663e-05, 'epoch': 1.3}) +[2026-06-17 21:43:09,429][starVLA.training.train_starvla][INFO] - Step 3156, Loss: {'train/loss': 0.4427649974822998, 'train/grad_norm_pre_clip': 13.991239547729492, 'timing/data': 0.0005129696801304817, 'timing/model': 0.9687736351042986, 'learning_rate/qwen_vl_interface': 6.09753726161392e-06, 'learning_rate/action_model': 6.09753726161392e-05, 'epoch': 1.3}) +[2026-06-17 21:43:13,341][starVLA.training.train_starvla][INFO] - Step 3157, Loss: {'train/loss': 0.40154850482940674, 'train/grad_norm_pre_clip': 14.932501792907715, 'timing/data': 0.0008128443732857704, 'timing/model': 0.9416240947321057, 'learning_rate/qwen_vl_interface': 6.095409188355793e-06, 'learning_rate/action_model': 6.095409188355793e-05, 'epoch': 1.3}) +[2026-06-17 21:43:17,408][starVLA.training.train_starvla][INFO] - Step 3158, Loss: {'train/loss': 0.4702586829662323, 'train/grad_norm_pre_clip': 16.316072463989258, 'timing/data': 0.0014701969921588898, 'timing/model': 0.9507298106327653, 'learning_rate/qwen_vl_interface': 6.0932809398434315e-06, 'learning_rate/action_model': 6.093280939843431e-05, 'epoch': 1.3}) +[2026-06-17 21:43:21,299][starVLA.training.train_starvla][INFO] - Step 3159, Loss: {'train/loss': 0.3797597289085388, 'train/grad_norm_pre_clip': 14.664468765258789, 'timing/data': 0.0005488535389304161, 'timing/model': 0.9498945809900761, 'learning_rate/qwen_vl_interface': 6.091152516518022e-06, 'learning_rate/action_model': 6.0911525165180216e-05, 'epoch': 1.3}) +[2026-06-17 21:43:25,236][starVLA.training.train_starvla][INFO] - Step 3160, Loss: {'train/loss': 0.48296746611595154, 'train/grad_norm_pre_clip': 16.712848663330078, 'timing/data': 0.0004784930497407913, 'timing/model': 0.9486347325146198, 'learning_rate/qwen_vl_interface': 6.089023918820792e-06, 'learning_rate/action_model': 6.089023918820792e-05, 'epoch': 1.3}) +[2026-06-17 21:43:29,117][starVLA.training.train_starvla][INFO] - Step 3161, Loss: {'train/loss': 0.37605202198028564, 'train/grad_norm_pre_clip': 12.726932525634766, 'timing/data': 0.0004880838096141815, 'timing/model': 0.9528623707592487, 'learning_rate/qwen_vl_interface': 6.0868951471929986e-06, 'learning_rate/action_model': 6.086895147192998e-05, 'epoch': 1.3}) +[2026-06-17 21:43:33,018][starVLA.training.train_starvla][INFO] - Step 3162, Loss: {'train/loss': 0.3983761966228485, 'train/grad_norm_pre_clip': 18.174701690673828, 'timing/data': 0.0004990724846720695, 'timing/model': 1.1325707864016294, 'learning_rate/qwen_vl_interface': 6.084766202075944e-06, 'learning_rate/action_model': 6.0847662020759435e-05, 'epoch': 1.3}) +[2026-06-17 21:43:36,744][starVLA.training.train_starvla][INFO] - Step 3163, Loss: {'train/loss': 0.5307711362838745, 'train/grad_norm_pre_clip': 18.6099910736084, 'timing/data': 0.0009485222399234772, 'timing/model': 0.9487750520929694, 'learning_rate/qwen_vl_interface': 6.082637083910955e-06, 'learning_rate/action_model': 6.0826370839109546e-05, 'epoch': 1.3}) +[2026-06-17 21:43:40,571][starVLA.training.train_starvla][INFO] - Step 3164, Loss: {'train/loss': 0.43861791491508484, 'train/grad_norm_pre_clip': 15.275008201599121, 'timing/data': 0.0005373600870370865, 'timing/model': 0.9363695876672864, 'learning_rate/qwen_vl_interface': 6.080507793139403e-06, 'learning_rate/action_model': 6.080507793139403e-05, 'epoch': 1.3}) +[2026-06-17 21:43:44,436][starVLA.training.train_starvla][INFO] - Step 3165, Loss: {'train/loss': 0.4458616077899933, 'train/grad_norm_pre_clip': 16.33491325378418, 'timing/data': 0.0004912521690130234, 'timing/model': 0.9536919398233294, 'learning_rate/qwen_vl_interface': 6.0783783302026925e-06, 'learning_rate/action_model': 6.078378330202692e-05, 'epoch': 1.3}) +[2026-06-17 21:43:48,322][starVLA.training.train_starvla][INFO] - Step 3166, Loss: {'train/loss': 0.38746222853660583, 'train/grad_norm_pre_clip': 12.944314956665039, 'timing/data': 0.000506620854139328, 'timing/model': 0.9430392142385244, 'learning_rate/qwen_vl_interface': 6.076248695542262e-06, 'learning_rate/action_model': 6.076248695542261e-05, 'epoch': 1.3}) +[2026-06-17 21:43:52,283][starVLA.training.train_starvla][INFO] - Step 3167, Loss: {'train/loss': 0.18385373055934906, 'train/grad_norm_pre_clip': 10.453757286071777, 'timing/data': 0.000518500804901123, 'timing/model': 0.9464853564277291, 'learning_rate/qwen_vl_interface': 6.074118889599589e-06, 'learning_rate/action_model': 6.0741188895995884e-05, 'epoch': 1.3}) +[2026-06-17 21:43:56,271][starVLA.training.train_starvla][INFO] - Step 3168, Loss: {'train/loss': 0.6003785133361816, 'train/grad_norm_pre_clip': 15.475574493408203, 'timing/data': 0.0009320331737399101, 'timing/model': 1.288900530897081, 'learning_rate/qwen_vl_interface': 6.07198891281618e-06, 'learning_rate/action_model': 6.07198891281618e-05, 'epoch': 1.3}) +[2026-06-17 21:44:00,135][starVLA.training.train_starvla][INFO] - Step 3169, Loss: {'train/loss': 0.6311227083206177, 'train/grad_norm_pre_clip': 17.442995071411133, 'timing/data': 0.0005272319540381432, 'timing/model': 1.1106038996949792, 'learning_rate/qwen_vl_interface': 6.0698587656335895e-06, 'learning_rate/action_model': 6.069858765633589e-05, 'epoch': 1.3}) +[2026-06-17 21:44:03,880][starVLA.training.train_starvla][INFO] - Step 3170, Loss: {'train/loss': 0.4250849187374115, 'train/grad_norm_pre_clip': 16.52806282043457, 'timing/data': 0.0009152339771389961, 'timing/model': 0.951026713475585, 'learning_rate/qwen_vl_interface': 6.067728448493392e-06, 'learning_rate/action_model': 6.067728448493392e-05, 'epoch': 1.3}) +[2026-06-17 21:44:07,830][starVLA.training.train_starvla][INFO] - Step 3171, Loss: {'train/loss': 0.22734729945659637, 'train/grad_norm_pre_clip': 9.34834098815918, 'timing/data': 0.00048797670751810074, 'timing/model': 0.9463767604902387, 'learning_rate/qwen_vl_interface': 6.065597961837209e-06, 'learning_rate/action_model': 6.0655979618372086e-05, 'epoch': 1.31}) +[2026-06-17 21:44:11,824][starVLA.training.train_starvla][INFO] - Step 3172, Loss: {'train/loss': 0.39526328444480896, 'train/grad_norm_pre_clip': 11.284064292907715, 'timing/data': 0.0005007954314351082, 'timing/model': 0.9495745683088899, 'learning_rate/qwen_vl_interface': 6.063467306106688e-06, 'learning_rate/action_model': 6.0634673061066885e-05, 'epoch': 1.31}) +[2026-06-17 21:44:15,942][starVLA.training.train_starvla][INFO] - Step 3173, Loss: {'train/loss': 0.1281750649213791, 'train/grad_norm_pre_clip': 7.564669609069824, 'timing/data': 0.0004684245213866234, 'timing/model': 1.082771072164178, 'learning_rate/qwen_vl_interface': 6.061336481743522e-06, 'learning_rate/action_model': 6.061336481743522e-05, 'epoch': 1.31}) +[2026-06-17 21:44:21,357][starVLA.training.train_starvla][INFO] - Step 3174, Loss: {'train/loss': 0.39943695068359375, 'train/grad_norm_pre_clip': 17.32396125793457, 'timing/data': 0.00046883802860975266, 'timing/model': 0.9374531628564, 'learning_rate/qwen_vl_interface': 6.059205489189429e-06, 'learning_rate/action_model': 6.059205489189429e-05, 'epoch': 1.31}) +[2026-06-17 21:44:25,456][starVLA.training.train_starvla][INFO] - Step 3175, Loss: {'train/loss': 0.6563084125518799, 'train/grad_norm_pre_clip': 15.9493989944458, 'timing/data': 0.0005573462694883347, 'timing/model': 1.2023574644699693, 'learning_rate/qwen_vl_interface': 6.057074328886168e-06, 'learning_rate/action_model': 6.057074328886168e-05, 'epoch': 1.31}) +[2026-06-17 21:44:30,538][starVLA.training.train_starvla][INFO] - Step 3176, Loss: {'train/loss': 0.1629004329442978, 'train/grad_norm_pre_clip': 7.719609260559082, 'timing/data': 0.0007478715851902962, 'timing/model': 0.9884611200541258, 'learning_rate/qwen_vl_interface': 6.0549430012755295e-06, 'learning_rate/action_model': 6.0549430012755295e-05, 'epoch': 1.31}) +[2026-06-17 21:44:34,460][starVLA.training.train_starvla][INFO] - Step 3177, Loss: {'train/loss': 0.5275561809539795, 'train/grad_norm_pre_clip': 17.338390350341797, 'timing/data': 0.0006079766899347305, 'timing/model': 0.9491074131801724, 'learning_rate/qwen_vl_interface': 6.052811506799341e-06, 'learning_rate/action_model': 6.052811506799341e-05, 'epoch': 1.31}) +[2026-06-17 21:44:38,393][starVLA.training.train_starvla][INFO] - Step 3178, Loss: {'train/loss': 0.5661935210227966, 'train/grad_norm_pre_clip': 63.380428314208984, 'timing/data': 0.0004977509379386902, 'timing/model': 1.0182037921622396, 'learning_rate/qwen_vl_interface': 6.050679845899462e-06, 'learning_rate/action_model': 6.050679845899462e-05, 'epoch': 1.31}) +[2026-06-17 21:44:42,271][starVLA.training.train_starvla][INFO] - Step 3179, Loss: {'train/loss': 0.2512970268726349, 'train/grad_norm_pre_clip': 10.602323532104492, 'timing/data': 0.0012425445020198822, 'timing/model': 0.9399045584723353, 'learning_rate/qwen_vl_interface': 6.048548019017791e-06, 'learning_rate/action_model': 6.04854801901779e-05, 'epoch': 1.31}) +[2026-06-17 21:44:46,167][starVLA.training.train_starvla][INFO] - Step 3180, Loss: {'train/loss': 0.8910747766494751, 'train/grad_norm_pre_clip': 20.504730224609375, 'timing/data': 0.0011609746143221855, 'timing/model': 0.9366316320374608, 'learning_rate/qwen_vl_interface': 6.046416026596256e-06, 'learning_rate/action_model': 6.046416026596255e-05, 'epoch': 1.31}) +[2026-06-17 21:44:50,019][starVLA.training.train_starvla][INFO] - Step 3181, Loss: {'train/loss': 0.5833644866943359, 'train/grad_norm_pre_clip': 16.70555305480957, 'timing/data': 0.0004686787724494934, 'timing/model': 0.9409588752314448, 'learning_rate/qwen_vl_interface': 6.044283869076822e-06, 'learning_rate/action_model': 6.0442838690768224e-05, 'epoch': 1.31}) +[2026-06-17 21:44:53,863][starVLA.training.train_starvla][INFO] - Step 3182, Loss: {'train/loss': 0.28868815302848816, 'train/grad_norm_pre_clip': 11.690828323364258, 'timing/data': 0.0005116481333971024, 'timing/model': 1.1071912683546543, 'learning_rate/qwen_vl_interface': 6.042151546901487e-06, 'learning_rate/action_model': 6.042151546901487e-05, 'epoch': 1.31}) +[2026-06-17 21:44:57,780][starVLA.training.train_starvla][INFO] - Step 3183, Loss: {'train/loss': 0.6385957598686218, 'train/grad_norm_pre_clip': 22.408594131469727, 'timing/data': 0.00048455316573381424, 'timing/model': 0.9486938305199146, 'learning_rate/qwen_vl_interface': 6.0400190605122846e-06, 'learning_rate/action_model': 6.040019060512284e-05, 'epoch': 1.31}) +[2026-06-17 21:45:01,570][starVLA.training.train_starvla][INFO] - Step 3184, Loss: {'train/loss': 0.5125749111175537, 'train/grad_norm_pre_clip': 17.26374053955078, 'timing/data': 0.0010537514463067055, 'timing/model': 0.9365714332088828, 'learning_rate/qwen_vl_interface': 6.0378864103512805e-06, 'learning_rate/action_model': 6.03788641035128e-05, 'epoch': 1.31}) +[2026-06-17 21:45:05,395][starVLA.training.train_starvla][INFO] - Step 3185, Loss: {'train/loss': 0.43250763416290283, 'train/grad_norm_pre_clip': 19.72245216369629, 'timing/data': 0.0008753295987844467, 'timing/model': 0.9515078095719218, 'learning_rate/qwen_vl_interface': 6.035753596860573e-06, 'learning_rate/action_model': 6.0357535968605725e-05, 'epoch': 1.31}) +[2026-06-17 21:45:09,296][starVLA.training.train_starvla][INFO] - Step 3186, Loss: {'train/loss': 0.24676579236984253, 'train/grad_norm_pre_clip': 11.536092758178711, 'timing/data': 0.0005594249814748764, 'timing/model': 0.957385909743607, 'learning_rate/qwen_vl_interface': 6.033620620482303e-06, 'learning_rate/action_model': 6.033620620482303e-05, 'epoch': 1.31}) +[2026-06-17 21:45:13,206][starVLA.training.train_starvla][INFO] - Step 3187, Loss: {'train/loss': 0.37940412759780884, 'train/grad_norm_pre_clip': 17.298795700073242, 'timing/data': 0.0005133487284183502, 'timing/model': 0.9539006110280752, 'learning_rate/qwen_vl_interface': 6.031487481658633e-06, 'learning_rate/action_model': 6.031487481658633e-05, 'epoch': 1.31}) +[2026-06-17 21:45:17,089][starVLA.training.train_starvla][INFO] - Step 3188, Loss: {'train/loss': 0.5746874213218689, 'train/grad_norm_pre_clip': 16.630048751831055, 'timing/data': 0.0005569476634263992, 'timing/model': 1.1600247891619802, 'learning_rate/qwen_vl_interface': 6.029354180831769e-06, 'learning_rate/action_model': 6.029354180831769e-05, 'epoch': 1.31}) +[2026-06-17 21:45:20,825][starVLA.training.train_starvla][INFO] - Step 3189, Loss: {'train/loss': 0.3118356466293335, 'train/grad_norm_pre_clip': 13.94985294342041, 'timing/data': 0.0010455651208758354, 'timing/model': 1.0492345187813044, 'learning_rate/qwen_vl_interface': 6.027220718443943e-06, 'learning_rate/action_model': 6.027220718443943e-05, 'epoch': 1.31}) +[2026-06-17 21:45:24,463][starVLA.training.train_starvla][INFO] - Step 3190, Loss: {'train/loss': 0.32996639609336853, 'train/grad_norm_pre_clip': 14.984152793884277, 'timing/data': 0.0005441159009933472, 'timing/model': 0.9370525907725096, 'learning_rate/qwen_vl_interface': 6.025087094937426e-06, 'learning_rate/action_model': 6.025087094937426e-05, 'epoch': 1.31}) +[2026-06-17 21:45:28,326][starVLA.training.train_starvla][INFO] - Step 3191, Loss: {'train/loss': 0.6093375086784363, 'train/grad_norm_pre_clip': 17.876617431640625, 'timing/data': 0.0005028815940022469, 'timing/model': 0.929879130795598, 'learning_rate/qwen_vl_interface': 6.022953310754521e-06, 'learning_rate/action_model': 6.022953310754521e-05, 'epoch': 1.31}) +[2026-06-17 21:45:32,125][starVLA.training.train_starvla][INFO] - Step 3192, Loss: {'train/loss': 0.5864400863647461, 'train/grad_norm_pre_clip': 15.843550682067871, 'timing/data': 0.0005075531080365181, 'timing/model': 0.9527227440848947, 'learning_rate/qwen_vl_interface': 6.020819366337562e-06, 'learning_rate/action_model': 6.020819366337562e-05, 'epoch': 1.31}) +[2026-06-17 21:45:35,949][starVLA.training.train_starvla][INFO] - Step 3193, Loss: {'train/loss': 0.5344589948654175, 'train/grad_norm_pre_clip': 18.391759872436523, 'timing/data': 0.0007390649989247322, 'timing/model': 0.9395632529631257, 'learning_rate/qwen_vl_interface': 6.018685262128921e-06, 'learning_rate/action_model': 6.01868526212892e-05, 'epoch': 1.31}) +[2026-06-17 21:45:39,908][starVLA.training.train_starvla][INFO] - Step 3194, Loss: {'train/loss': 0.5521867275238037, 'train/grad_norm_pre_clip': 15.314698219299316, 'timing/data': 0.000690205954015255, 'timing/model': 0.9403324918821454, 'learning_rate/qwen_vl_interface': 6.016550998570995e-06, 'learning_rate/action_model': 6.0165509985709944e-05, 'epoch': 1.31}) +[2026-06-17 21:45:43,967][starVLA.training.train_starvla][INFO] - Step 3195, Loss: {'train/loss': 0.3800997734069824, 'train/grad_norm_pre_clip': 14.823081970214844, 'timing/data': 0.0008168872445821762, 'timing/model': 1.1860819701105356, 'learning_rate/qwen_vl_interface': 6.014416576106224e-06, 'learning_rate/action_model': 6.0144165761062244e-05, 'epoch': 1.32}) +[2026-06-17 21:45:48,164][starVLA.training.train_starvla][INFO] - Step 3196, Loss: {'train/loss': 0.5810840725898743, 'train/grad_norm_pre_clip': 16.223251342773438, 'timing/data': 0.0005304859951138496, 'timing/model': 1.398664304986596, 'learning_rate/qwen_vl_interface': 6.012281995177074e-06, 'learning_rate/action_model': 6.012281995177074e-05, 'epoch': 1.32}) +[2026-06-17 21:45:52,136][starVLA.training.train_starvla][INFO] - Step 3197, Loss: {'train/loss': 0.24940666556358337, 'train/grad_norm_pre_clip': 13.025561332702637, 'timing/data': 0.0005618399009108543, 'timing/model': 0.9738387241959572, 'learning_rate/qwen_vl_interface': 6.010147256226046e-06, 'learning_rate/action_model': 6.0101472562260454e-05, 'epoch': 1.32}) +[2026-06-17 21:45:56,823][starVLA.training.train_starvla][INFO] - Step 3198, Loss: {'train/loss': 0.25488176941871643, 'train/grad_norm_pre_clip': 13.535683631896973, 'timing/data': 0.0008662696927785873, 'timing/model': 0.9895280795171857, 'learning_rate/qwen_vl_interface': 6.008012359695673e-06, 'learning_rate/action_model': 6.0080123596956726e-05, 'epoch': 1.32}) +[2026-06-17 21:46:00,741][starVLA.training.train_starvla][INFO] - Step 3199, Loss: {'train/loss': 0.5108261108398438, 'train/grad_norm_pre_clip': 15.02078628540039, 'timing/data': 0.0008281506597995758, 'timing/model': 0.949789684265852, 'learning_rate/qwen_vl_interface': 6.005877306028522e-06, 'learning_rate/action_model': 6.005877306028522e-05, 'epoch': 1.32}) +[2026-06-17 21:48:09,596][starVLA.training.train_starvla][INFO] - Step 3200, Loss: {'train/loss': 0.270442932844162, 'train/grad_norm_pre_clip': 18.02357292175293, 'eval/loss': 0.48159804940223694, 'eval/latency_6/loss': 0.48159804940223694, 'eval/demon_fire_precision': 0.9525466624485922, 'eval/demon_fire_recall': 0.9731738849385908, 'eval/demon_fire_f1': 0.962749800159872, 'eval/demon_fire_rate_model': 0.8795214245965498, 'eval/demon_fire_rate_teacher': 0.860879243183083, 'eval/demon_left_precision': 0.8300220750551877, 'eval/demon_left_recall': 0.6596491228070176, 'eval/demon_left_f1': 0.7350928641251222, 'eval/demon_left_rate_model': 0.12604340567612687, 'eval/demon_left_rate_teacher': 0.15859766277128548, 'eval/demon_right_precision': 0.822141560798548, 'eval/demon_right_recall': 0.695852534562212, 'eval/demon_right_f1': 0.7537437603993343, 'eval/demon_right_rate_model': 0.15331107401224262, 'eval/demon_right_rate_teacher': 0.18113522537562604, 'eval/demon_move_f1': 0.7444183122622283, 'eval/demon_cc_f1': 0.8848905629341469, 'eval/latency_6/demon_fire_precision': 0.9525466624485922, 'eval/latency_6/demon_fire_recall': 0.9731738849385908, 'eval/latency_6/demon_fire_f1': 0.962749800159872, 'eval/latency_6/demon_fire_rate_model': 0.8795214245965498, 'eval/latency_6/demon_fire_rate_teacher': 0.860879243183083, 'eval/latency_6/demon_left_precision': 0.8300220750551877, 'eval/latency_6/demon_left_recall': 0.6596491228070176, 'eval/latency_6/demon_left_f1': 0.7350928641251222, 'eval/latency_6/demon_left_rate_model': 0.12604340567612687, 'eval/latency_6/demon_left_rate_teacher': 0.15859766277128548, 'eval/latency_6/demon_right_precision': 0.822141560798548, 'eval/latency_6/demon_right_recall': 0.695852534562212, 'eval/latency_6/demon_right_f1': 0.7537437603993343, 'eval/latency_6/demon_right_rate_model': 0.15331107401224262, 'eval/latency_6/demon_right_rate_teacher': 0.18113522537562604, 'eval/latency_6/demon_move_f1': 0.7444183122622283, 'eval/latency_6/demon_cc_f1': 0.8848905629341469, 'timing/data': 0.0013210959732532501, 'timing/model': 0.9539093682542443, 'learning_rate/qwen_vl_interface': 6.003742095667194e-06, 'learning_rate/action_model': 6.003742095667194e-05, 'epoch': 1.32}) +[2026-06-17 21:48:09,600][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_3200_state +[2026-06-17 21:48:09,605][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 21:48:16,447][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_3200_state/model.safetensors +[2026-06-17 21:48:32,015][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_3200_state/optimizer.bin +[2026-06-17 21:48:32,016][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_3200_state/sampler.bin +[2026-06-17 21:48:32,016][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_3200_state/sampler_1.bin +[2026-06-17 21:48:32,019][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_3200_state/random_states_0.pkl +[2026-06-17 21:48:32,022][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3200 +[2026-06-17 21:48:32,022][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 21:48:32,033][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 21:48:36,317][starVLA.training.train_starvla][INFO] - Step 3201, Loss: {'train/loss': 0.3629242777824402, 'train/grad_norm_pre_clip': 14.383623123168945, 'timing/data': 0.0008826050907373428, 'timing/model': 0.9499160870909691, 'learning_rate/qwen_vl_interface': 6.001606729054318e-06, 'learning_rate/action_model': 6.0016067290543176e-05, 'epoch': 1.32}) +[2026-06-17 21:48:40,159][starVLA.training.train_starvla][INFO] - Step 3202, Loss: {'train/loss': 0.4909989833831787, 'train/grad_norm_pre_clip': 13.947749137878418, 'timing/data': 0.000544271431863308, 'timing/model': 0.9528687726706266, 'learning_rate/qwen_vl_interface': 5.999471206632559e-06, 'learning_rate/action_model': 5.999471206632559e-05, 'epoch': 1.32}) +[2026-06-17 21:48:44,002][starVLA.training.train_starvla][INFO] - Step 3203, Loss: {'train/loss': 0.6052567362785339, 'train/grad_norm_pre_clip': 17.741003036499023, 'timing/data': 0.0005067074671387672, 'timing/model': 1.1393772084265947, 'learning_rate/qwen_vl_interface': 5.99733552884461e-06, 'learning_rate/action_model': 5.99733552884461e-05, 'epoch': 1.32}) +[2026-06-17 21:48:47,934][starVLA.training.train_starvla][INFO] - Step 3204, Loss: {'train/loss': 0.5164440870285034, 'train/grad_norm_pre_clip': 14.970710754394531, 'timing/data': 0.0005246438086032867, 'timing/model': 1.176294012926519, 'learning_rate/qwen_vl_interface': 5.995199696133204e-06, 'learning_rate/action_model': 5.9951996961332034e-05, 'epoch': 1.32}) +[2026-06-17 21:48:51,797][starVLA.training.train_starvla][INFO] - Step 3205, Loss: {'train/loss': 0.5626690983772278, 'train/grad_norm_pre_clip': 13.807400703430176, 'timing/data': 0.0007812930271029472, 'timing/model': 0.9732794370502234, 'learning_rate/qwen_vl_interface': 5.993063708941098e-06, 'learning_rate/action_model': 5.993063708941098e-05, 'epoch': 1.32}) +[2026-06-17 21:48:55,735][starVLA.training.train_starvla][INFO] - Step 3206, Loss: {'train/loss': 0.33302509784698486, 'train/grad_norm_pre_clip': 15.083915710449219, 'timing/data': 0.0008411463350057602, 'timing/model': 0.945501315407455, 'learning_rate/qwen_vl_interface': 5.990927567711085e-06, 'learning_rate/action_model': 5.990927567711085e-05, 'epoch': 1.32}) +[2026-06-17 21:48:59,693][starVLA.training.train_starvla][INFO] - Step 3207, Loss: {'train/loss': 0.3652240037918091, 'train/grad_norm_pre_clip': 12.797175407409668, 'timing/data': 0.0004970729351043701, 'timing/model': 1.018669062294066, 'learning_rate/qwen_vl_interface': 5.988791272885991e-06, 'learning_rate/action_model': 5.988791272885991e-05, 'epoch': 1.32}) +[2026-06-17 21:49:04,840][starVLA.training.train_starvla][INFO] - Step 3208, Loss: {'train/loss': 0.6442174911499023, 'train/grad_norm_pre_clip': 19.688663482666016, 'timing/data': 0.0004900079220533371, 'timing/model': 1.376675734296441, 'learning_rate/qwen_vl_interface': 5.986654824908671e-06, 'learning_rate/action_model': 5.9866548249086706e-05, 'epoch': 1.32}) +[2026-06-17 21:49:08,900][starVLA.training.train_starvla][INFO] - Step 3209, Loss: {'train/loss': 1.0887459516525269, 'train/grad_norm_pre_clip': 20.679502487182617, 'timing/data': 0.0007008016109466553, 'timing/model': 0.9468531785532832, 'learning_rate/qwen_vl_interface': 5.9845182242220125e-06, 'learning_rate/action_model': 5.984518224222012e-05, 'epoch': 1.32}) +[2026-06-17 21:49:12,950][starVLA.training.train_starvla][INFO] - Step 3210, Loss: {'train/loss': 0.4894905686378479, 'train/grad_norm_pre_clip': 12.926246643066406, 'timing/data': 0.0005800444632768631, 'timing/model': 1.2724467273801565, 'learning_rate/qwen_vl_interface': 5.982381471268935e-06, 'learning_rate/action_model': 5.9823814712689354e-05, 'epoch': 1.32}) +[2026-06-17 21:49:16,935][starVLA.training.train_starvla][INFO] - Step 3211, Loss: {'train/loss': 0.41875794529914856, 'train/grad_norm_pre_clip': 17.837589263916016, 'timing/data': 0.0008587725460529327, 'timing/model': 1.0926325535401702, 'learning_rate/qwen_vl_interface': 5.980244566492392e-06, 'learning_rate/action_model': 5.9802445664923925e-05, 'epoch': 1.32}) +[2026-06-17 21:49:20,700][starVLA.training.train_starvla][INFO] - Step 3212, Loss: {'train/loss': 0.37533456087112427, 'train/grad_norm_pre_clip': 14.690515518188477, 'timing/data': 0.001017250120639801, 'timing/model': 1.014287787489593, 'learning_rate/qwen_vl_interface': 5.978107510335365e-06, 'learning_rate/action_model': 5.9781075103353647e-05, 'epoch': 1.32}) +[2026-06-17 21:49:24,554][starVLA.training.train_starvla][INFO] - Step 3213, Loss: {'train/loss': 0.36836859583854675, 'train/grad_norm_pre_clip': 13.609248161315918, 'timing/data': 0.00046781450510025024, 'timing/model': 0.9703176273033023, 'learning_rate/qwen_vl_interface': 5.975970303240867e-06, 'learning_rate/action_model': 5.9759703032408664e-05, 'epoch': 1.32}) +[2026-06-17 21:49:28,398][starVLA.training.train_starvla][INFO] - Step 3214, Loss: {'train/loss': 0.4971754848957062, 'train/grad_norm_pre_clip': 14.687520980834961, 'timing/data': 0.0004956228658556938, 'timing/model': 0.9506116565316916, 'learning_rate/qwen_vl_interface': 5.973832945651944e-06, 'learning_rate/action_model': 5.9738329456519435e-05, 'epoch': 1.32}) +[2026-06-17 21:49:32,238][starVLA.training.train_starvla][INFO] - Step 3215, Loss: {'train/loss': 0.5684483051300049, 'train/grad_norm_pre_clip': 14.547798156738281, 'timing/data': 0.00048530660569667816, 'timing/model': 0.9758259765803814, 'learning_rate/qwen_vl_interface': 5.971695438011672e-06, 'learning_rate/action_model': 5.971695438011672e-05, 'epoch': 1.32}) +[2026-06-17 21:49:36,107][starVLA.training.train_starvla][INFO] - Step 3216, Loss: {'train/loss': 0.396179735660553, 'train/grad_norm_pre_clip': 11.770479202270508, 'timing/data': 0.0006736377254128456, 'timing/model': 0.9581509195268154, 'learning_rate/qwen_vl_interface': 5.969557780763159e-06, 'learning_rate/action_model': 5.9695577807631585e-05, 'epoch': 1.32}) +[2026-06-17 21:49:40,127][starVLA.training.train_starvla][INFO] - Step 3217, Loss: {'train/loss': 0.5161030292510986, 'train/grad_norm_pre_clip': 14.483216285705566, 'timing/data': 0.0008635884150862694, 'timing/model': 1.28932561352849, 'learning_rate/qwen_vl_interface': 5.9674199743495456e-06, 'learning_rate/action_model': 5.967419974349545e-05, 'epoch': 1.32}) +[2026-06-17 21:49:45,730][starVLA.training.train_starvla][INFO] - Step 3218, Loss: {'train/loss': 0.3516140282154083, 'train/grad_norm_pre_clip': 12.284894943237305, 'timing/data': 0.000498470850288868, 'timing/model': 2.87636645603925, 'learning_rate/qwen_vl_interface': 5.965282019213996e-06, 'learning_rate/action_model': 5.965282019213996e-05, 'epoch': 1.32}) +[2026-06-17 21:49:49,507][starVLA.training.train_starvla][INFO] - Step 3219, Loss: {'train/loss': 0.30727532505989075, 'train/grad_norm_pre_clip': 12.778914451599121, 'timing/data': 0.000483684241771698, 'timing/model': 0.9427733095362782, 'learning_rate/qwen_vl_interface': 5.963143915799718e-06, 'learning_rate/action_model': 5.9631439157997184e-05, 'epoch': 1.33}) +[2026-06-17 21:49:53,350][starVLA.training.train_starvla][INFO] - Step 3220, Loss: {'train/loss': 0.3871716260910034, 'train/grad_norm_pre_clip': 14.198630332946777, 'timing/data': 0.0005167350172996521, 'timing/model': 0.9775116294622421, 'learning_rate/qwen_vl_interface': 5.961005664549936e-06, 'learning_rate/action_model': 5.9610056645499354e-05, 'epoch': 1.33}) +[2026-06-17 21:49:57,192][starVLA.training.train_starvla][INFO] - Step 3221, Loss: {'train/loss': 0.4452037513256073, 'train/grad_norm_pre_clip': 11.820290565490723, 'timing/data': 0.0015224656090140343, 'timing/model': 0.956590598449111, 'learning_rate/qwen_vl_interface': 5.958867265907916e-06, 'learning_rate/action_model': 5.958867265907916e-05, 'epoch': 1.33}) +[2026-06-17 21:50:01,034][starVLA.training.train_starvla][INFO] - Step 3222, Loss: {'train/loss': 0.2691798806190491, 'train/grad_norm_pre_clip': 11.809674263000488, 'timing/data': 0.0005009807646274567, 'timing/model': 0.9593177251517773, 'learning_rate/qwen_vl_interface': 5.956728720316949e-06, 'learning_rate/action_model': 5.9567287203169485e-05, 'epoch': 1.33}) +[2026-06-17 21:50:05,204][starVLA.training.train_starvla][INFO] - Step 3223, Loss: {'train/loss': 0.4236714839935303, 'train/grad_norm_pre_clip': 16.1749267578125, 'timing/data': 0.0004992624744772911, 'timing/model': 1.1399597646668553, 'learning_rate/qwen_vl_interface': 5.954590028220357e-06, 'learning_rate/action_model': 5.954590028220357e-05, 'epoch': 1.33}) +[2026-06-17 21:50:09,061][starVLA.training.train_starvla][INFO] - Step 3224, Loss: {'train/loss': 0.6624792814254761, 'train/grad_norm_pre_clip': 16.79191780090332, 'timing/data': 0.0006611589342355728, 'timing/model': 1.1758539136499166, 'learning_rate/qwen_vl_interface': 5.952451190061496e-06, 'learning_rate/action_model': 5.9524511900614956e-05, 'epoch': 1.33}) +[2026-06-17 21:50:12,826][starVLA.training.train_starvla][INFO] - Step 3225, Loss: {'train/loss': 0.49120140075683594, 'train/grad_norm_pre_clip': 16.37496566772461, 'timing/data': 0.0005782060325145721, 'timing/model': 1.0633680243045092, 'learning_rate/qwen_vl_interface': 5.950312206283743e-06, 'learning_rate/action_model': 5.9503122062837434e-05, 'epoch': 1.33}) +[2026-06-17 21:50:16,466][starVLA.training.train_starvla][INFO] - Step 3226, Loss: {'train/loss': 0.5298931002616882, 'train/grad_norm_pre_clip': 19.91599464416504, 'timing/data': 0.0006967615336179733, 'timing/model': 0.9390810057520866, 'learning_rate/qwen_vl_interface': 5.948173077330522e-06, 'learning_rate/action_model': 5.948173077330522e-05, 'epoch': 1.33}) +[2026-06-17 21:50:20,452][starVLA.training.train_starvla][INFO] - Step 3227, Loss: {'train/loss': 0.3730125427246094, 'train/grad_norm_pre_clip': 12.802752494812012, 'timing/data': 0.0009918827563524246, 'timing/model': 0.9627337055280805, 'learning_rate/qwen_vl_interface': 5.9460338036452665e-06, 'learning_rate/action_model': 5.946033803645267e-05, 'epoch': 1.33}) +[2026-06-17 21:50:24,356][starVLA.training.train_starvla][INFO] - Step 3228, Loss: {'train/loss': 0.5203518271446228, 'train/grad_norm_pre_clip': 15.22038745880127, 'timing/data': 0.0008526630699634552, 'timing/model': 0.9630724266171455, 'learning_rate/qwen_vl_interface': 5.943894385671456e-06, 'learning_rate/action_model': 5.9438943856714556e-05, 'epoch': 1.33}) +[2026-06-17 21:50:28,240][starVLA.training.train_starvla][INFO] - Step 3229, Loss: {'train/loss': 0.4321904480457306, 'train/grad_norm_pre_clip': 16.55927085876465, 'timing/data': 0.000664987601339817, 'timing/model': 0.9440755806863308, 'learning_rate/qwen_vl_interface': 5.941754823852593e-06, 'learning_rate/action_model': 5.9417548238525924e-05, 'epoch': 1.33}) +[2026-06-17 21:50:32,146][starVLA.training.train_starvla][INFO] - Step 3230, Loss: {'train/loss': 0.642468273639679, 'train/grad_norm_pre_clip': 46.28580093383789, 'timing/data': 0.0004920288920402527, 'timing/model': 1.165654780343175, 'learning_rate/qwen_vl_interface': 5.939615118632208e-06, 'learning_rate/action_model': 5.939615118632208e-05, 'epoch': 1.33}) +[2026-06-17 21:50:35,941][starVLA.training.train_starvla][INFO] - Step 3231, Loss: {'train/loss': 0.7284753918647766, 'train/grad_norm_pre_clip': 18.012012481689453, 'timing/data': 0.00047370046377182007, 'timing/model': 1.0831572031602263, 'learning_rate/qwen_vl_interface': 5.937475270453872e-06, 'learning_rate/action_model': 5.937475270453872e-05, 'epoch': 1.33}) +[2026-06-17 21:50:39,615][starVLA.training.train_starvla][INFO] - Step 3232, Loss: {'train/loss': 0.48111745715141296, 'train/grad_norm_pre_clip': 12.410200119018555, 'timing/data': 0.0009921425953507423, 'timing/model': 0.9509631954133511, 'learning_rate/qwen_vl_interface': 5.935335279761169e-06, 'learning_rate/action_model': 5.935335279761168e-05, 'epoch': 1.33}) +[2026-06-17 21:50:43,433][starVLA.training.train_starvla][INFO] - Step 3233, Loss: {'train/loss': 0.3422350585460663, 'train/grad_norm_pre_clip': 14.310608863830566, 'timing/data': 0.0008364962413907051, 'timing/model': 0.9403769122436643, 'learning_rate/qwen_vl_interface': 5.933195146997725e-06, 'learning_rate/action_model': 5.9331951469977254e-05, 'epoch': 1.33}) +[2026-06-17 21:50:47,279][starVLA.training.train_starvla][INFO] - Step 3234, Loss: {'train/loss': 0.3848997950553894, 'train/grad_norm_pre_clip': 14.919148445129395, 'timing/data': 0.0005188416689634323, 'timing/model': 0.9391801841557026, 'learning_rate/qwen_vl_interface': 5.931054872607193e-06, 'learning_rate/action_model': 5.9310548726071925e-05, 'epoch': 1.33}) +[2026-06-17 21:50:51,243][starVLA.training.train_starvla][INFO] - Step 3235, Loss: {'train/loss': 0.19352319836616516, 'train/grad_norm_pre_clip': 10.614076614379883, 'timing/data': 0.0004920614883303642, 'timing/model': 0.9826978705823421, 'learning_rate/qwen_vl_interface': 5.928914457033252e-06, 'learning_rate/action_model': 5.928914457033252e-05, 'epoch': 1.33}) +[2026-06-17 21:50:55,120][starVLA.training.train_starvla][INFO] - Step 3236, Loss: {'train/loss': 0.6937823295593262, 'train/grad_norm_pre_clip': 17.41025733947754, 'timing/data': 0.0006638998165726662, 'timing/model': 0.9593725260347128, 'learning_rate/qwen_vl_interface': 5.926773900719616e-06, 'learning_rate/action_model': 5.9267739007196157e-05, 'epoch': 1.33}) +[2026-06-17 21:50:58,976][starVLA.training.train_starvla][INFO] - Step 3237, Loss: {'train/loss': 0.6922581791877747, 'train/grad_norm_pre_clip': 18.346435546875, 'timing/data': 0.0009004911407828331, 'timing/model': 1.1715265326201916, 'learning_rate/qwen_vl_interface': 5.92463320411002e-06, 'learning_rate/action_model': 5.92463320411002e-05, 'epoch': 1.33}) +[2026-06-17 21:51:02,661][starVLA.training.train_starvla][INFO] - Step 3238, Loss: {'train/loss': 0.4219963252544403, 'train/grad_norm_pre_clip': 16.819440841674805, 'timing/data': 0.0011481838300824165, 'timing/model': 0.9361436534672976, 'learning_rate/qwen_vl_interface': 5.922492367648236e-06, 'learning_rate/action_model': 5.922492367648236e-05, 'epoch': 1.33}) +[2026-06-17 21:51:06,574][starVLA.training.train_starvla][INFO] - Step 3239, Loss: {'train/loss': 0.33835235238075256, 'train/grad_norm_pre_clip': 11.600298881530762, 'timing/data': 0.0005437526851892471, 'timing/model': 0.9438693653792143, 'learning_rate/qwen_vl_interface': 5.920351391778059e-06, 'learning_rate/action_model': 5.920351391778059e-05, 'epoch': 1.33}) +[2026-06-17 21:51:10,417][starVLA.training.train_starvla][INFO] - Step 3240, Loss: {'train/loss': 0.47410503029823303, 'train/grad_norm_pre_clip': 15.597112655639648, 'timing/data': 0.0005047433078289032, 'timing/model': 0.9268088191747665, 'learning_rate/qwen_vl_interface': 5.918210276943318e-06, 'learning_rate/action_model': 5.918210276943318e-05, 'epoch': 1.33}) +[2026-06-17 21:51:14,234][starVLA.training.train_starvla][INFO] - Step 3241, Loss: {'train/loss': 0.23483705520629883, 'train/grad_norm_pre_clip': 10.943188667297363, 'timing/data': 0.0005139429122209549, 'timing/model': 0.9351420616731048, 'learning_rate/qwen_vl_interface': 5.916069023587867e-06, 'learning_rate/action_model': 5.916069023587867e-05, 'epoch': 1.33}) +[2026-06-17 21:51:18,103][starVLA.training.train_starvla][INFO] - Step 3242, Loss: {'train/loss': 0.26278480887413025, 'train/grad_norm_pre_clip': 11.502283096313477, 'timing/data': 0.00048244185745716095, 'timing/model': 0.9640248557552695, 'learning_rate/qwen_vl_interface': 5.913927632155592e-06, 'learning_rate/action_model': 5.913927632155591e-05, 'epoch': 1.33}) +[2026-06-17 21:51:22,008][starVLA.training.train_starvla][INFO] - Step 3243, Loss: {'train/loss': 0.5766803622245789, 'train/grad_norm_pre_clip': 13.105321884155273, 'timing/data': 0.0004769228398799896, 'timing/model': 1.2246269583702087, 'learning_rate/qwen_vl_interface': 5.911786103090402e-06, 'learning_rate/action_model': 5.911786103090402e-05, 'epoch': 1.33}) +[2026-06-17 21:51:25,708][starVLA.training.train_starvla][INFO] - Step 3244, Loss: {'train/loss': 0.5221003890037537, 'train/grad_norm_pre_clip': 15.451976776123047, 'timing/data': 0.0005520973354578018, 'timing/model': 0.9574528429657221, 'learning_rate/qwen_vl_interface': 5.909644436836243e-06, 'learning_rate/action_model': 5.909644436836244e-05, 'epoch': 1.34}) +[2026-06-17 21:51:30,508][starVLA.training.train_starvla][INFO] - Step 3245, Loss: {'train/loss': 0.6308331489562988, 'train/grad_norm_pre_clip': 16.278915405273438, 'timing/data': 0.0010829875245690346, 'timing/model': 1.0387721601873636, 'learning_rate/qwen_vl_interface': 5.907502633837082e-06, 'learning_rate/action_model': 5.907502633837082e-05, 'epoch': 1.34}) +[2026-06-17 21:51:34,491][starVLA.training.train_starvla][INFO] - Step 3246, Loss: {'train/loss': 0.3152559995651245, 'train/grad_norm_pre_clip': 10.68994140625, 'timing/data': 0.0007289424538612366, 'timing/model': 0.9482804350554943, 'learning_rate/qwen_vl_interface': 5.905360694536918e-06, 'learning_rate/action_model': 5.905360694536918e-05, 'epoch': 1.34}) +[2026-06-17 21:51:38,394][starVLA.training.train_starvla][INFO] - Step 3247, Loss: {'train/loss': 0.5301372408866882, 'train/grad_norm_pre_clip': 18.12436294555664, 'timing/data': 0.0005082320421934128, 'timing/model': 0.948346764780581, 'learning_rate/qwen_vl_interface': 5.903218619379776e-06, 'learning_rate/action_model': 5.903218619379776e-05, 'epoch': 1.34}) +[2026-06-17 21:51:42,304][starVLA.training.train_starvla][INFO] - Step 3248, Loss: {'train/loss': 0.36923477053642273, 'train/grad_norm_pre_clip': 14.885563850402832, 'timing/data': 0.0004976252093911171, 'timing/model': 0.9551883656531572, 'learning_rate/qwen_vl_interface': 5.901076408809712e-06, 'learning_rate/action_model': 5.9010764088097114e-05, 'epoch': 1.34}) +[2026-06-17 21:51:46,232][starVLA.training.train_starvla][INFO] - Step 3249, Loss: {'train/loss': 0.4715687930583954, 'train/grad_norm_pre_clip': 17.37801742553711, 'timing/data': 0.0005401158705353737, 'timing/model': 1.022272989153862, 'learning_rate/qwen_vl_interface': 5.898934063270806e-06, 'learning_rate/action_model': 5.8989340632708056e-05, 'epoch': 1.34}) +[2026-06-17 21:51:50,155][starVLA.training.train_starvla][INFO] - Step 3250, Loss: {'train/loss': 0.3089015781879425, 'train/grad_norm_pre_clip': 16.293527603149414, 'timing/data': 0.0006459178403019905, 'timing/model': 1.1280452683568, 'learning_rate/qwen_vl_interface': 5.896791583207173e-06, 'learning_rate/action_model': 5.896791583207173e-05, 'epoch': 1.34}) +[2026-06-17 21:51:53,847][starVLA.training.train_starvla][INFO] - Step 3251, Loss: {'train/loss': 0.2632468044757843, 'train/grad_norm_pre_clip': 12.478806495666504, 'timing/data': 0.0009265420958399773, 'timing/model': 0.9404694018885493, 'learning_rate/qwen_vl_interface': 5.894648969062947e-06, 'learning_rate/action_model': 5.894648969062947e-05, 'epoch': 1.34}) +[2026-06-17 21:51:57,825][starVLA.training.train_starvla][INFO] - Step 3252, Loss: {'train/loss': 0.36447250843048096, 'train/grad_norm_pre_clip': 14.280808448791504, 'timing/data': 0.0007958095520734787, 'timing/model': 0.9341571442782879, 'learning_rate/qwen_vl_interface': 5.892506221282297e-06, 'learning_rate/action_model': 5.892506221282297e-05, 'epoch': 1.34}) +[2026-06-17 21:52:01,710][starVLA.training.train_starvla][INFO] - Step 3253, Loss: {'train/loss': 0.42737042903900146, 'train/grad_norm_pre_clip': 21.0371150970459, 'timing/data': 0.001436612568795681, 'timing/model': 0.9535081125795841, 'learning_rate/qwen_vl_interface': 5.890363340309415e-06, 'learning_rate/action_model': 5.890363340309415e-05, 'epoch': 1.34}) +[2026-06-17 21:52:05,564][starVLA.training.train_starvla][INFO] - Step 3254, Loss: {'train/loss': 0.4460754096508026, 'train/grad_norm_pre_clip': 14.451915740966797, 'timing/data': 0.0005426853895187378, 'timing/model': 0.9493197957053781, 'learning_rate/qwen_vl_interface': 5.888220326588523e-06, 'learning_rate/action_model': 5.8882203265885226e-05, 'epoch': 1.34}) +[2026-06-17 21:52:11,742][starVLA.training.train_starvla][INFO] - Step 3255, Loss: {'train/loss': 0.5913481712341309, 'train/grad_norm_pre_clip': 18.5056095123291, 'timing/data': 0.0007484164088964462, 'timing/model': 3.2717519514262676, 'learning_rate/qwen_vl_interface': 5.8860771805638725e-06, 'learning_rate/action_model': 5.886077180563872e-05, 'epoch': 1.34}) +[2026-06-17 21:52:16,905][starVLA.training.train_starvla][INFO] - Step 3256, Loss: {'train/loss': 0.25118038058280945, 'train/grad_norm_pre_clip': 10.250362396240234, 'timing/data': 0.0006761299446225166, 'timing/model': 1.1442566737532616, 'learning_rate/qwen_vl_interface': 5.883933902679736e-06, 'learning_rate/action_model': 5.8839339026797355e-05, 'epoch': 1.34}) +[2026-06-17 21:52:20,750][starVLA.training.train_starvla][INFO] - Step 3257, Loss: {'train/loss': 0.733775794506073, 'train/grad_norm_pre_clip': 21.89203643798828, 'timing/data': 0.0007888833060860634, 'timing/model': 1.1159099051728845, 'learning_rate/qwen_vl_interface': 5.8817904933804214e-06, 'learning_rate/action_model': 5.881790493380421e-05, 'epoch': 1.34}) +[2026-06-17 21:52:24,558][starVLA.training.train_starvla][INFO] - Step 3258, Loss: {'train/loss': 0.6423841118812561, 'train/grad_norm_pre_clip': 17.48866081237793, 'timing/data': 0.0005055544897913933, 'timing/model': 0.9803490899503231, 'learning_rate/qwen_vl_interface': 5.8796469531102565e-06, 'learning_rate/action_model': 5.879646953110256e-05, 'epoch': 1.34}) +[2026-06-17 21:52:29,165][starVLA.training.train_starvla][INFO] - Step 3259, Loss: {'train/loss': 0.5648910403251648, 'train/grad_norm_pre_clip': 17.59286117553711, 'timing/data': 0.0007173316553235054, 'timing/model': 1.1359619237482548, 'learning_rate/qwen_vl_interface': 5.877503282313601e-06, 'learning_rate/action_model': 5.877503282313601e-05, 'epoch': 1.34}) +[2026-06-17 21:52:33,263][starVLA.training.train_starvla][INFO] - Step 3260, Loss: {'train/loss': 0.2789931893348694, 'train/grad_norm_pre_clip': 15.117586135864258, 'timing/data': 0.00045160669833421707, 'timing/model': 0.9366009812802076, 'learning_rate/qwen_vl_interface': 5.875359481434839e-06, 'learning_rate/action_model': 5.875359481434839e-05, 'epoch': 1.34}) +[2026-06-17 21:52:37,144][starVLA.training.train_starvla][INFO] - Step 3261, Loss: {'train/loss': 0.5051411986351013, 'train/grad_norm_pre_clip': 17.726146697998047, 'timing/data': 0.0008887620642781258, 'timing/model': 0.9504095371812582, 'learning_rate/qwen_vl_interface': 5.873215550918384e-06, 'learning_rate/action_model': 5.873215550918384e-05, 'epoch': 1.34}) +[2026-06-17 21:52:41,138][starVLA.training.train_starvla][INFO] - Step 3262, Loss: {'train/loss': 0.41310828924179077, 'train/grad_norm_pre_clip': 22.300321578979492, 'timing/data': 0.0009800409898161888, 'timing/model': 0.9402975738048553, 'learning_rate/qwen_vl_interface': 5.871071491208676e-06, 'learning_rate/action_model': 5.871071491208676e-05, 'epoch': 1.34}) +[2026-06-17 21:52:45,000][starVLA.training.train_starvla][INFO] - Step 3263, Loss: {'train/loss': 0.5074902772903442, 'train/grad_norm_pre_clip': 13.057191848754883, 'timing/data': 0.0005073705688118935, 'timing/model': 1.1442385725677013, 'learning_rate/qwen_vl_interface': 5.868927302750178e-06, 'learning_rate/action_model': 5.868927302750178e-05, 'epoch': 1.34}) +[2026-06-17 21:52:48,762][starVLA.training.train_starvla][INFO] - Step 3264, Loss: {'train/loss': 0.5543573498725891, 'train/grad_norm_pre_clip': 31.906963348388672, 'timing/data': 0.0007990850135684013, 'timing/model': 0.967944991774857, 'learning_rate/qwen_vl_interface': 5.866782985987386e-06, 'learning_rate/action_model': 5.866782985987386e-05, 'epoch': 1.34}) +[2026-06-17 21:52:52,654][starVLA.training.train_starvla][INFO] - Step 3265, Loss: {'train/loss': 0.4855780303478241, 'train/grad_norm_pre_clip': 16.68832015991211, 'timing/data': 0.0004779268056154251, 'timing/model': 0.9351347163319588, 'learning_rate/qwen_vl_interface': 5.864638541364815e-06, 'learning_rate/action_model': 5.8646385413648154e-05, 'epoch': 1.34}) +[2026-06-17 21:52:57,097][starVLA.training.train_starvla][INFO] - Step 3266, Loss: {'train/loss': 0.6718471050262451, 'train/grad_norm_pre_clip': 16.6146183013916, 'timing/data': 0.0008463244885206223, 'timing/model': 0.9490763749927282, 'learning_rate/qwen_vl_interface': 5.862493969327015e-06, 'learning_rate/action_model': 5.862493969327014e-05, 'epoch': 1.34}) +[2026-06-17 21:53:00,935][starVLA.training.train_starvla][INFO] - Step 3267, Loss: {'train/loss': 0.5042998790740967, 'train/grad_norm_pre_clip': 17.75407600402832, 'timing/data': 0.0005462188273668289, 'timing/model': 0.9561958331614733, 'learning_rate/qwen_vl_interface': 5.860349270318555e-06, 'learning_rate/action_model': 5.860349270318555e-05, 'epoch': 1.34}) +[2026-06-17 21:53:04,735][starVLA.training.train_starvla][INFO] - Step 3268, Loss: {'train/loss': 0.4698261618614197, 'train/grad_norm_pre_clip': 16.504898071289062, 'timing/data': 0.0010989373549818993, 'timing/model': 0.9350240007042885, 'learning_rate/qwen_vl_interface': 5.8582044447840346e-06, 'learning_rate/action_model': 5.858204444784034e-05, 'epoch': 1.35}) +[2026-06-17 21:53:08,642][starVLA.training.train_starvla][INFO] - Step 3269, Loss: {'train/loss': 0.5375756621360779, 'train/grad_norm_pre_clip': 17.146686553955078, 'timing/data': 0.0007317354902625084, 'timing/model': 0.9402152700349689, 'learning_rate/qwen_vl_interface': 5.8560594931680785e-06, 'learning_rate/action_model': 5.856059493168078e-05, 'epoch': 1.35}) +[2026-06-17 21:53:12,521][starVLA.training.train_starvla][INFO] - Step 3270, Loss: {'train/loss': 0.3707716464996338, 'train/grad_norm_pre_clip': 12.938192367553711, 'timing/data': 0.00045582279562950134, 'timing/model': 1.1371037187054753, 'learning_rate/qwen_vl_interface': 5.853914415915337e-06, 'learning_rate/action_model': 5.853914415915337e-05, 'epoch': 1.35}) +[2026-06-17 21:53:16,181][starVLA.training.train_starvla][INFO] - Step 3271, Loss: {'train/loss': 0.2901850938796997, 'train/grad_norm_pre_clip': 10.61671257019043, 'timing/data': 0.0007505845278501511, 'timing/model': 0.9479683618992567, 'learning_rate/qwen_vl_interface': 5.8517692134704875e-06, 'learning_rate/action_model': 5.851769213470487e-05, 'epoch': 1.35}) +[2026-06-17 21:53:20,059][starVLA.training.train_starvla][INFO] - Step 3272, Loss: {'train/loss': 0.4505525827407837, 'train/grad_norm_pre_clip': 14.84662914276123, 'timing/data': 0.0008544884622097015, 'timing/model': 0.9289152156561613, 'learning_rate/qwen_vl_interface': 5.849623886278232e-06, 'learning_rate/action_model': 5.849623886278231e-05, 'epoch': 1.35}) +[2026-06-17 21:53:23,908][starVLA.training.train_starvla][INFO] - Step 3273, Loss: {'train/loss': 0.3622506856918335, 'train/grad_norm_pre_clip': 13.640209197998047, 'timing/data': 0.0007931012660264969, 'timing/model': 0.9438134310767055, 'learning_rate/qwen_vl_interface': 5.8474784347833e-06, 'learning_rate/action_model': 5.8474784347833e-05, 'epoch': 1.35}) +[2026-06-17 21:53:27,732][starVLA.training.train_starvla][INFO] - Step 3274, Loss: {'train/loss': 0.5484092235565186, 'train/grad_norm_pre_clip': 13.415146827697754, 'timing/data': 0.0009556077420711517, 'timing/model': 0.9429276995360851, 'learning_rate/qwen_vl_interface': 5.845332859430446e-06, 'learning_rate/action_model': 5.845332859430446e-05, 'epoch': 1.35}) +[2026-06-17 21:53:31,637][starVLA.training.train_starvla][INFO] - Step 3275, Loss: {'train/loss': 0.4521912634372711, 'train/grad_norm_pre_clip': 18.389774322509766, 'timing/data': 0.0010248664766550064, 'timing/model': 0.9316116366535425, 'learning_rate/qwen_vl_interface': 5.843187160664449e-06, 'learning_rate/action_model': 5.8431871606644495e-05, 'epoch': 1.35}) +[2026-06-17 21:53:35,550][starVLA.training.train_starvla][INFO] - Step 3276, Loss: {'train/loss': 0.4611496925354004, 'train/grad_norm_pre_clip': 14.97203254699707, 'timing/data': 0.0005006100982427597, 'timing/model': 1.115042369812727, 'learning_rate/qwen_vl_interface': 5.841041338930115e-06, 'learning_rate/action_model': 5.8410413389301155e-05, 'epoch': 1.35}) +[2026-06-17 21:53:39,366][starVLA.training.train_starvla][INFO] - Step 3277, Loss: {'train/loss': 0.519183337688446, 'train/grad_norm_pre_clip': 24.298551559448242, 'timing/data': 0.0005671801045536995, 'timing/model': 1.0808926168829203, 'learning_rate/qwen_vl_interface': 5.8388953946722775e-06, 'learning_rate/action_model': 5.8388953946722766e-05, 'epoch': 1.35}) +[2026-06-17 21:53:43,033][starVLA.training.train_starvla][INFO] - Step 3278, Loss: {'train/loss': 0.2453768104314804, 'train/grad_norm_pre_clip': 14.511090278625488, 'timing/data': 0.0012736786156892776, 'timing/model': 0.9450943609699607, 'learning_rate/qwen_vl_interface': 5.836749328335789e-06, 'learning_rate/action_model': 5.8367493283357885e-05, 'epoch': 1.35}) +[2026-06-17 21:53:46,937][starVLA.training.train_starvla][INFO] - Step 3279, Loss: {'train/loss': 0.603676438331604, 'train/grad_norm_pre_clip': 16.177982330322266, 'timing/data': 0.0008506057783961296, 'timing/model': 0.9576634420081973, 'learning_rate/qwen_vl_interface': 5.8346031403655345e-06, 'learning_rate/action_model': 5.8346031403655345e-05, 'epoch': 1.35}) +[2026-06-17 21:53:50,814][starVLA.training.train_starvla][INFO] - Step 3280, Loss: {'train/loss': 0.31586283445358276, 'train/grad_norm_pre_clip': 14.317469596862793, 'timing/data': 0.0005285218358039856, 'timing/model': 0.9525509066879749, 'learning_rate/qwen_vl_interface': 5.832456831206418e-06, 'learning_rate/action_model': 5.8324568312064176e-05, 'epoch': 1.35}) +[2026-06-17 21:53:54,633][starVLA.training.train_starvla][INFO] - Step 3281, Loss: {'train/loss': 0.4307249188423157, 'train/grad_norm_pre_clip': 14.308441162109375, 'timing/data': 0.0011037765070796013, 'timing/model': 0.9454966224730015, 'learning_rate/qwen_vl_interface': 5.830310401303377e-06, 'learning_rate/action_model': 5.830310401303377e-05, 'epoch': 1.35}) +[2026-06-17 21:53:58,483][starVLA.training.train_starvla][INFO] - Step 3282, Loss: {'train/loss': 0.3924348056316376, 'train/grad_norm_pre_clip': 13.861021995544434, 'timing/data': 0.00047809258103370667, 'timing/model': 0.9428657563403249, 'learning_rate/qwen_vl_interface': 5.8281638511013625e-06, 'learning_rate/action_model': 5.828163851101363e-05, 'epoch': 1.35}) +[2026-06-17 21:54:02,381][starVLA.training.train_starvla][INFO] - Step 3283, Loss: {'train/loss': 0.27872899174690247, 'train/grad_norm_pre_clip': 11.338268280029297, 'timing/data': 0.0007854355499148369, 'timing/model': 1.1353744501248002, 'learning_rate/qwen_vl_interface': 5.8260171810453605e-06, 'learning_rate/action_model': 5.82601718104536e-05, 'epoch': 1.35}) +[2026-06-17 21:54:06,254][starVLA.training.train_starvla][INFO] - Step 3284, Loss: {'train/loss': 0.17610058188438416, 'train/grad_norm_pre_clip': 8.917950630187988, 'timing/data': 0.0005048681050539017, 'timing/model': 1.183230142109096, 'learning_rate/qwen_vl_interface': 5.823870391580376e-06, 'learning_rate/action_model': 5.823870391580376e-05, 'epoch': 1.35}) +[2026-06-17 21:54:09,937][starVLA.training.train_starvla][INFO] - Step 3285, Loss: {'train/loss': 0.14993786811828613, 'train/grad_norm_pre_clip': 6.665515899658203, 'timing/data': 0.0008703591302037239, 'timing/model': 0.9521675333380699, 'learning_rate/qwen_vl_interface': 5.821723483151444e-06, 'learning_rate/action_model': 5.821723483151443e-05, 'epoch': 1.35}) +[2026-06-17 21:54:13,768][starVLA.training.train_starvla][INFO] - Step 3286, Loss: {'train/loss': 0.3349694311618805, 'train/grad_norm_pre_clip': 14.654486656188965, 'timing/data': 0.0004834691062569618, 'timing/model': 0.9403652437031269, 'learning_rate/qwen_vl_interface': 5.819576456203617e-06, 'learning_rate/action_model': 5.8195764562036174e-05, 'epoch': 1.35}) +[2026-06-17 21:54:17,585][starVLA.training.train_starvla][INFO] - Step 3287, Loss: {'train/loss': 0.32267138361930847, 'train/grad_norm_pre_clip': 11.820915222167969, 'timing/data': 0.0005027167499065399, 'timing/model': 0.9407086502760649, 'learning_rate/qwen_vl_interface': 5.817429311181977e-06, 'learning_rate/action_model': 5.8174293111819764e-05, 'epoch': 1.35}) +[2026-06-17 21:54:21,484][starVLA.training.train_starvla][INFO] - Step 3288, Loss: {'train/loss': 0.6374735236167908, 'train/grad_norm_pre_clip': 17.65941619873047, 'timing/data': 0.0004996843636035919, 'timing/model': 0.9495492847636342, 'learning_rate/qwen_vl_interface': 5.815282048531633e-06, 'learning_rate/action_model': 5.815282048531633e-05, 'epoch': 1.35}) +[2026-06-17 21:54:25,323][starVLA.training.train_starvla][INFO] - Step 3289, Loss: {'train/loss': 0.5102300047874451, 'train/grad_norm_pre_clip': 18.49540138244629, 'timing/data': 0.00048709195107221603, 'timing/model': 0.9603530708700418, 'learning_rate/qwen_vl_interface': 5.81313466869771e-06, 'learning_rate/action_model': 5.8131346686977096e-05, 'epoch': 1.35}) +[2026-06-17 21:54:29,208][starVLA.training.train_starvla][INFO] - Step 3290, Loss: {'train/loss': 0.5575769543647766, 'train/grad_norm_pre_clip': 16.38475799560547, 'timing/data': 0.0004988266155123711, 'timing/model': 1.1266411356627941, 'learning_rate/qwen_vl_interface': 5.810987172125364e-06, 'learning_rate/action_model': 5.810987172125364e-05, 'epoch': 1.35}) +[2026-06-17 21:54:33,005][starVLA.training.train_starvla][INFO] - Step 3291, Loss: {'train/loss': 0.3449125587940216, 'train/grad_norm_pre_clip': 11.680288314819336, 'timing/data': 0.0005272515118122101, 'timing/model': 1.0609013754874468, 'learning_rate/qwen_vl_interface': 5.808839559259774e-06, 'learning_rate/action_model': 5.8088395592597746e-05, 'epoch': 1.35}) +[2026-06-17 21:54:36,741][starVLA.training.train_starvla][INFO] - Step 3292, Loss: {'train/loss': 0.4674276113510132, 'train/grad_norm_pre_clip': 12.678995132446289, 'timing/data': 0.0008399924263358116, 'timing/model': 0.9371153507381678, 'learning_rate/qwen_vl_interface': 5.806691830546143e-06, 'learning_rate/action_model': 5.8066918305461427e-05, 'epoch': 1.36}) +[2026-06-17 21:54:41,708][starVLA.training.train_starvla][INFO] - Step 3293, Loss: {'train/loss': 0.2032441347837448, 'train/grad_norm_pre_clip': 8.743352890014648, 'timing/data': 0.0004910370334982872, 'timing/model': 0.9452214660122991, 'learning_rate/qwen_vl_interface': 5.804543986429696e-06, 'learning_rate/action_model': 5.804543986429696e-05, 'epoch': 1.36}) +[2026-06-17 21:54:45,547][starVLA.training.train_starvla][INFO] - Step 3294, Loss: {'train/loss': 0.3668569028377533, 'train/grad_norm_pre_clip': 18.031333923339844, 'timing/data': 0.0005086911842226982, 'timing/model': 0.9405829012393951, 'learning_rate/qwen_vl_interface': 5.802396027355683e-06, 'learning_rate/action_model': 5.8023960273556835e-05, 'epoch': 1.36}) +[2026-06-17 21:54:49,619][starVLA.training.train_starvla][INFO] - Step 3295, Loss: {'train/loss': 0.5674256682395935, 'train/grad_norm_pre_clip': 16.546472549438477, 'timing/data': 0.0004720883443951607, 'timing/model': 1.0385243343189359, 'learning_rate/qwen_vl_interface': 5.800247953769381e-06, 'learning_rate/action_model': 5.800247953769381e-05, 'epoch': 1.36}) +[2026-06-17 21:54:53,543][starVLA.training.train_starvla][INFO] - Step 3296, Loss: {'train/loss': 0.6179044246673584, 'train/grad_norm_pre_clip': 18.442859649658203, 'timing/data': 0.0011159023270010948, 'timing/model': 1.1412210846319795, 'learning_rate/qwen_vl_interface': 5.798099766116084e-06, 'learning_rate/action_model': 5.798099766116084e-05, 'epoch': 1.36}) +[2026-06-17 21:54:57,865][starVLA.training.train_starvla][INFO] - Step 3297, Loss: {'train/loss': 0.31680062413215637, 'train/grad_norm_pre_clip': 14.009611129760742, 'timing/data': 0.0006642695516347885, 'timing/model': 1.1554390173405409, 'learning_rate/qwen_vl_interface': 5.7959514648411175e-06, 'learning_rate/action_model': 5.795951464841117e-05, 'epoch': 1.36}) +[2026-06-17 21:55:01,952][starVLA.training.train_starvla][INFO] - Step 3298, Loss: {'train/loss': 0.47607678174972534, 'train/grad_norm_pre_clip': 18.226545333862305, 'timing/data': 0.0005124583840370178, 'timing/model': 0.9962576786056161, 'learning_rate/qwen_vl_interface': 5.793803050389824e-06, 'learning_rate/action_model': 5.793803050389824e-05, 'epoch': 1.36}) +[2026-06-17 21:55:06,420][starVLA.training.train_starvla][INFO] - Step 3299, Loss: {'train/loss': 0.46364784240722656, 'train/grad_norm_pre_clip': 22.81195640563965, 'timing/data': 0.000504116527736187, 'timing/model': 0.9477273290976882, 'learning_rate/qwen_vl_interface': 5.7916545232075725e-06, 'learning_rate/action_model': 5.791654523207573e-05, 'epoch': 1.36}) +[2026-06-17 21:57:12,009][starVLA.training.train_starvla][INFO] - Step 3300, Loss: {'train/loss': 0.46895095705986023, 'train/grad_norm_pre_clip': 17.51749610900879, 'eval/loss': 0.4955894351005554, 'eval/latency_6/loss': 0.4955894351005554, 'eval/demon_fire_precision': 0.9333333333333333, 'eval/demon_fire_recall': 0.9864253393665159, 'eval/demon_fire_f1': 0.9591451917033312, 'eval/demon_fire_rate_model': 0.9098497495826378, 'eval/demon_fire_rate_teacher': 0.860879243183083, 'eval/demon_left_precision': 0.8787128712871287, 'eval/demon_left_recall': 0.6228070175438597, 'eval/demon_left_f1': 0.7289527720739221, 'eval/demon_left_rate_model': 0.112409571508069, 'eval/demon_left_rate_teacher': 0.15859766277128548, 'eval/demon_right_precision': 0.7753510140405616, 'eval/demon_right_recall': 0.7634408602150538, 'eval/demon_right_f1': 0.7693498452012385, 'eval/demon_right_rate_model': 0.1783528102392877, 'eval/demon_right_rate_teacher': 0.18113522537562604, 'eval/demon_move_f1': 0.7491513086375803, 'eval/demon_cc_f1': 0.8847446177495953, 'eval/latency_6/demon_fire_precision': 0.9333333333333333, 'eval/latency_6/demon_fire_recall': 0.9864253393665159, 'eval/latency_6/demon_fire_f1': 0.9591451917033312, 'eval/latency_6/demon_fire_rate_model': 0.9098497495826378, 'eval/latency_6/demon_fire_rate_teacher': 0.860879243183083, 'eval/latency_6/demon_left_precision': 0.8787128712871287, 'eval/latency_6/demon_left_recall': 0.6228070175438597, 'eval/latency_6/demon_left_f1': 0.7289527720739221, 'eval/latency_6/demon_left_rate_model': 0.112409571508069, 'eval/latency_6/demon_left_rate_teacher': 0.15859766277128548, 'eval/latency_6/demon_right_precision': 0.7753510140405616, 'eval/latency_6/demon_right_recall': 0.7634408602150538, 'eval/latency_6/demon_right_f1': 0.7693498452012385, 'eval/latency_6/demon_right_rate_model': 0.1783528102392877, 'eval/latency_6/demon_right_rate_teacher': 0.18113522537562604, 'eval/latency_6/demon_move_f1': 0.7491513086375803, 'eval/latency_6/demon_cc_f1': 0.8847446177495953, 'timing/data': 0.000873686745762825, 'timing/model': 0.9394604796543717, 'learning_rate/qwen_vl_interface': 5.789505883739756e-06, 'learning_rate/action_model': 5.7895058837397554e-05, 'epoch': 1.36}) +[2026-06-17 21:57:12,011][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_3300_state +[2026-06-17 21:57:12,016][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 21:57:18,489][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_3300_state/model.safetensors +[2026-06-17 21:57:33,495][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_3300_state/optimizer.bin +[2026-06-17 21:57:33,496][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_3300_state/sampler.bin +[2026-06-17 21:57:33,496][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_3300_state/sampler_1.bin +[2026-06-17 21:57:33,524][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_3300_state/random_states_0.pkl +[2026-06-17 21:57:33,527][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3300 +[2026-06-17 21:57:33,527][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 21:57:33,538][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 21:57:37,348][starVLA.training.train_starvla][INFO] - Step 3301, Loss: {'train/loss': 0.22195684909820557, 'train/grad_norm_pre_clip': 6.960445404052734, 'timing/data': 0.0005399351939558983, 'timing/model': 0.9250959847122431, 'learning_rate/qwen_vl_interface': 5.78735713243179e-06, 'learning_rate/action_model': 5.78735713243179e-05, 'epoch': 1.36}) +[2026-06-17 21:57:41,136][starVLA.training.train_starvla][INFO] - Step 3302, Loss: {'train/loss': 0.16666024923324585, 'train/grad_norm_pre_clip': 8.403613090515137, 'timing/data': 0.0010726666077971458, 'timing/model': 0.9325316734611988, 'learning_rate/qwen_vl_interface': 5.785208269729111e-06, 'learning_rate/action_model': 5.785208269729111e-05, 'epoch': 1.36}) +[2026-06-17 21:57:45,127][starVLA.training.train_starvla][INFO] - Step 3303, Loss: {'train/loss': 0.49317649006843567, 'train/grad_norm_pre_clip': 18.126192092895508, 'timing/data': 0.0006817001849412918, 'timing/model': 0.9342789873480797, 'learning_rate/qwen_vl_interface': 5.783059296077183e-06, 'learning_rate/action_model': 5.7830592960771826e-05, 'epoch': 1.36}) +[2026-06-17 21:57:49,093][starVLA.training.train_starvla][INFO] - Step 3304, Loss: {'train/loss': 0.8078691959381104, 'train/grad_norm_pre_clip': 21.22906494140625, 'timing/data': 0.0008612694218754768, 'timing/model': 0.9357142522931099, 'learning_rate/qwen_vl_interface': 5.780910211921488e-06, 'learning_rate/action_model': 5.7809102119214874e-05, 'epoch': 1.36}) +[2026-06-17 21:57:52,894][starVLA.training.train_starvla][INFO] - Step 3305, Loss: {'train/loss': 0.6821545958518982, 'train/grad_norm_pre_clip': 21.418119430541992, 'timing/data': 0.0004873909056186676, 'timing/model': 1.1205677725374699, 'learning_rate/qwen_vl_interface': 5.778761017707536e-06, 'learning_rate/action_model': 5.778761017707536e-05, 'epoch': 1.36}) +[2026-06-17 21:57:56,475][starVLA.training.train_starvla][INFO] - Step 3306, Loss: {'train/loss': 0.7414866089820862, 'train/grad_norm_pre_clip': 17.005704879760742, 'timing/data': 0.00048374850302934647, 'timing/model': 0.9280646108090878, 'learning_rate/qwen_vl_interface': 5.7766117138808545e-06, 'learning_rate/action_model': 5.776611713880854e-05, 'epoch': 1.36}) +[2026-06-17 21:58:00,457][starVLA.training.train_starvla][INFO] - Step 3307, Loss: {'train/loss': 0.14839018881320953, 'train/grad_norm_pre_clip': 7.195854663848877, 'timing/data': 0.0011123856529593468, 'timing/model': 0.9312974689528346, 'learning_rate/qwen_vl_interface': 5.774462300886999e-06, 'learning_rate/action_model': 5.774462300886999e-05, 'epoch': 1.36}) +[2026-06-17 21:58:04,382][starVLA.training.train_starvla][INFO] - Step 3308, Loss: {'train/loss': 0.307155042886734, 'train/grad_norm_pre_clip': 14.66586685180664, 'timing/data': 0.0006830915808677673, 'timing/model': 0.9925280893221498, 'learning_rate/qwen_vl_interface': 5.772312779171545e-06, 'learning_rate/action_model': 5.7723127791715445e-05, 'epoch': 1.36}) +[2026-06-17 21:58:08,193][starVLA.training.train_starvla][INFO] - Step 3309, Loss: {'train/loss': 0.3193104863166809, 'train/grad_norm_pre_clip': 15.307416915893555, 'timing/data': 0.0008834963664412498, 'timing/model': 0.9347269749268889, 'learning_rate/qwen_vl_interface': 5.770163149180089e-06, 'learning_rate/action_model': 5.770163149180089e-05, 'epoch': 1.36}) +[2026-06-17 21:58:12,158][starVLA.training.train_starvla][INFO] - Step 3310, Loss: {'train/loss': 0.4664367437362671, 'train/grad_norm_pre_clip': 16.585256576538086, 'timing/data': 0.0005187792703509331, 'timing/model': 0.9370794845744967, 'learning_rate/qwen_vl_interface': 5.768013411358252e-06, 'learning_rate/action_model': 5.768013411358252e-05, 'epoch': 1.36}) +[2026-06-17 21:58:16,194][starVLA.training.train_starvla][INFO] - Step 3311, Loss: {'train/loss': 0.37449219822883606, 'train/grad_norm_pre_clip': 15.573813438415527, 'timing/data': 0.0010422077029943466, 'timing/model': 1.2856053598225117, 'learning_rate/qwen_vl_interface': 5.765863566151677e-06, 'learning_rate/action_model': 5.7658635661516767e-05, 'epoch': 1.36}) +[2026-06-17 21:58:19,779][starVLA.training.train_starvla][INFO] - Step 3312, Loss: {'train/loss': 0.3886832296848297, 'train/grad_norm_pre_clip': 14.663418769836426, 'timing/data': 0.0004993602633476257, 'timing/model': 0.9260791651904583, 'learning_rate/qwen_vl_interface': 5.763713614006033e-06, 'learning_rate/action_model': 5.763713614006032e-05, 'epoch': 1.36}) +[2026-06-17 21:58:23,600][starVLA.training.train_starvla][INFO] - Step 3313, Loss: {'train/loss': 0.4580840468406677, 'train/grad_norm_pre_clip': 16.456554412841797, 'timing/data': 0.0005201837047934532, 'timing/model': 0.9327828902751207, 'learning_rate/qwen_vl_interface': 5.761563555367e-06, 'learning_rate/action_model': 5.761563555367e-05, 'epoch': 1.36}) +[2026-06-17 21:58:28,246][starVLA.training.train_starvla][INFO] - Step 3314, Loss: {'train/loss': 0.6024978756904602, 'train/grad_norm_pre_clip': 18.173978805541992, 'timing/data': 0.0005088169127702713, 'timing/model': 0.9356052540242672, 'learning_rate/qwen_vl_interface': 5.759413390680293e-06, 'learning_rate/action_model': 5.759413390680293e-05, 'epoch': 1.36}) +[2026-06-17 21:58:32,044][starVLA.training.train_starvla][INFO] - Step 3315, Loss: {'train/loss': 0.29046735167503357, 'train/grad_norm_pre_clip': 12.653482437133789, 'timing/data': 0.000906442292034626, 'timing/model': 0.9288337510079145, 'learning_rate/qwen_vl_interface': 5.757263120391643e-06, 'learning_rate/action_model': 5.757263120391643e-05, 'epoch': 1.36}) +[2026-06-17 21:58:35,946][starVLA.training.train_starvla][INFO] - Step 3316, Loss: {'train/loss': 0.39554232358932495, 'train/grad_norm_pre_clip': 16.7513370513916, 'timing/data': 0.0004949979484081268, 'timing/model': 0.9388982653617859, 'learning_rate/qwen_vl_interface': 5.7551127449468005e-06, 'learning_rate/action_model': 5.7551127449468004e-05, 'epoch': 1.36}) +[2026-06-17 21:58:40,002][starVLA.training.train_starvla][INFO] - Step 3317, Loss: {'train/loss': 0.7840402126312256, 'train/grad_norm_pre_clip': 19.022459030151367, 'timing/data': 0.0004807654768228531, 'timing/model': 1.3753678696230054, 'learning_rate/qwen_vl_interface': 5.752962264791543e-06, 'learning_rate/action_model': 5.752962264791543e-05, 'epoch': 1.37}) +[2026-06-17 21:58:43,753][starVLA.training.train_starvla][INFO] - Step 3318, Loss: {'train/loss': 0.20211128890514374, 'train/grad_norm_pre_clip': 15.947318077087402, 'timing/data': 0.0005412949249148369, 'timing/model': 1.0646286373957992, 'learning_rate/qwen_vl_interface': 5.750811680371666e-06, 'learning_rate/action_model': 5.750811680371666e-05, 'epoch': 1.37}) +[2026-06-17 21:58:48,121][starVLA.training.train_starvla][INFO] - Step 3319, Loss: {'train/loss': 0.486301451921463, 'train/grad_norm_pre_clip': 17.013076782226562, 'timing/data': 0.00047900620847940445, 'timing/model': 1.4835095033049583, 'learning_rate/qwen_vl_interface': 5.7486609921329914e-06, 'learning_rate/action_model': 5.748660992132992e-05, 'epoch': 1.37}) +[2026-06-17 21:58:52,315][starVLA.training.train_starvla][INFO] - Step 3320, Loss: {'train/loss': 0.49653252959251404, 'train/grad_norm_pre_clip': 17.266284942626953, 'timing/data': 0.0009352201595902443, 'timing/model': 0.9720920948311687, 'learning_rate/qwen_vl_interface': 5.746510200521354e-06, 'learning_rate/action_model': 5.746510200521355e-05, 'epoch': 1.37}) +[2026-06-17 21:58:56,511][starVLA.training.train_starvla][INFO] - Step 3321, Loss: {'train/loss': 0.48604729771614075, 'train/grad_norm_pre_clip': 14.554503440856934, 'timing/data': 0.0005194833502173424, 'timing/model': 0.9595721457153559, 'learning_rate/qwen_vl_interface': 5.7443593059826206e-06, 'learning_rate/action_model': 5.744359305982621e-05, 'epoch': 1.37}) +[2026-06-17 21:59:00,292][starVLA.training.train_starvla][INFO] - Step 3322, Loss: {'train/loss': 0.4020833671092987, 'train/grad_norm_pre_clip': 17.127382278442383, 'timing/data': 0.0005717650055885315, 'timing/model': 0.925070327706635, 'learning_rate/qwen_vl_interface': 5.742208308962669e-06, 'learning_rate/action_model': 5.742208308962669e-05, 'epoch': 1.37}) +[2026-06-17 21:59:04,120][starVLA.training.train_starvla][INFO] - Step 3323, Loss: {'train/loss': 0.44551998376846313, 'train/grad_norm_pre_clip': 14.62387752532959, 'timing/data': 0.0005526375025510788, 'timing/model': 0.94489311799407, 'learning_rate/qwen_vl_interface': 5.740057209907405e-06, 'learning_rate/action_model': 5.7400572099074046e-05, 'epoch': 1.37}) +[2026-06-17 21:59:08,087][starVLA.training.train_starvla][INFO] - Step 3324, Loss: {'train/loss': 0.645405113697052, 'train/grad_norm_pre_clip': 18.213815689086914, 'timing/data': 0.0005412967875599861, 'timing/model': 1.2362619126215577, 'learning_rate/qwen_vl_interface': 5.7379060092627545e-06, 'learning_rate/action_model': 5.7379060092627544e-05, 'epoch': 1.37}) +[2026-06-17 21:59:11,893][starVLA.training.train_starvla][INFO] - Step 3325, Loss: {'train/loss': 0.3334127366542816, 'train/grad_norm_pre_clip': 35.19811248779297, 'timing/data': 0.0011104186996817589, 'timing/model': 0.9351229798048735, 'learning_rate/qwen_vl_interface': 5.735754707474663e-06, 'learning_rate/action_model': 5.735754707474663e-05, 'epoch': 1.37}) +[2026-06-17 21:59:15,774][starVLA.training.train_starvla][INFO] - Step 3326, Loss: {'train/loss': 0.34793558716773987, 'train/grad_norm_pre_clip': 14.889629364013672, 'timing/data': 0.0011599492281675339, 'timing/model': 0.9444268411025405, 'learning_rate/qwen_vl_interface': 5.7336033049891e-06, 'learning_rate/action_model': 5.733603304989099e-05, 'epoch': 1.37}) +[2026-06-17 21:59:19,706][starVLA.training.train_starvla][INFO] - Step 3327, Loss: {'train/loss': 0.32262519001960754, 'train/grad_norm_pre_clip': 11.834392547607422, 'timing/data': 0.0005172528326511383, 'timing/model': 0.9403065079823136, 'learning_rate/qwen_vl_interface': 5.731451802252049e-06, 'learning_rate/action_model': 5.731451802252048e-05, 'epoch': 1.37}) +[2026-06-17 21:59:23,539][starVLA.training.train_starvla][INFO] - Step 3328, Loss: {'train/loss': 0.4060359001159668, 'train/grad_norm_pre_clip': 14.207514762878418, 'timing/data': 0.0005239173769950867, 'timing/model': 0.9602009458467364, 'learning_rate/qwen_vl_interface': 5.729300199709522e-06, 'learning_rate/action_model': 5.729300199709522e-05, 'epoch': 1.37}) +[2026-06-17 21:59:27,384][starVLA.training.train_starvla][INFO] - Step 3329, Loss: {'train/loss': 0.36308178305625916, 'train/grad_norm_pre_clip': 13.981610298156738, 'timing/data': 0.0007957033812999725, 'timing/model': 0.9402904268354177, 'learning_rate/qwen_vl_interface': 5.7271484978075475e-06, 'learning_rate/action_model': 5.7271484978075475e-05, 'epoch': 1.37}) +[2026-06-17 21:59:31,431][starVLA.training.train_starvla][INFO] - Step 3330, Loss: {'train/loss': 0.4694664180278778, 'train/grad_norm_pre_clip': 33.383243560791016, 'timing/data': 0.0008802637457847595, 'timing/model': 1.1896525947377086, 'learning_rate/qwen_vl_interface': 5.724996696992176e-06, 'learning_rate/action_model': 5.7249966969921764e-05, 'epoch': 1.37}) +[2026-06-17 21:59:35,231][starVLA.training.train_starvla][INFO] - Step 3331, Loss: {'train/loss': 0.5413987040519714, 'train/grad_norm_pre_clip': 16.822214126586914, 'timing/data': 0.0006220275536179543, 'timing/model': 1.1115911891683936, 'learning_rate/qwen_vl_interface': 5.72284479770948e-06, 'learning_rate/action_model': 5.72284479770948e-05, 'epoch': 1.37}) +[2026-06-17 21:59:38,849][starVLA.training.train_starvla][INFO] - Step 3332, Loss: {'train/loss': 0.4723588526248932, 'train/grad_norm_pre_clip': 16.96709442138672, 'timing/data': 0.000514383427798748, 'timing/model': 0.9305323055014014, 'learning_rate/qwen_vl_interface': 5.720692800405547e-06, 'learning_rate/action_model': 5.720692800405547e-05, 'epoch': 1.37}) +[2026-06-17 21:59:42,650][starVLA.training.train_starvla][INFO] - Step 3333, Loss: {'train/loss': 0.45434731245040894, 'train/grad_norm_pre_clip': 13.787335395812988, 'timing/data': 0.0005095712840557098, 'timing/model': 0.9350886894389987, 'learning_rate/qwen_vl_interface': 5.71854070552649e-06, 'learning_rate/action_model': 5.7185407055264904e-05, 'epoch': 1.37}) +[2026-06-17 21:59:46,462][starVLA.training.train_starvla][INFO] - Step 3334, Loss: {'train/loss': 0.26759782433509827, 'train/grad_norm_pre_clip': 9.109481811523438, 'timing/data': 0.0006080633029341698, 'timing/model': 0.9421100933104753, 'learning_rate/qwen_vl_interface': 5.71638851351844e-06, 'learning_rate/action_model': 5.71638851351844e-05, 'epoch': 1.37}) +[2026-06-17 21:59:50,448][starVLA.training.train_starvla][INFO] - Step 3335, Loss: {'train/loss': 0.6709237694740295, 'train/grad_norm_pre_clip': 22.310041427612305, 'timing/data': 0.0005657002329826355, 'timing/model': 0.9344666982069612, 'learning_rate/qwen_vl_interface': 5.714236224827551e-06, 'learning_rate/action_model': 5.714236224827551e-05, 'epoch': 1.37}) +[2026-06-17 21:59:55,114][starVLA.training.train_starvla][INFO] - Step 3336, Loss: {'train/loss': 0.2709970474243164, 'train/grad_norm_pre_clip': 17.00678062438965, 'timing/data': 0.0005106190219521523, 'timing/model': 1.5500246128067374, 'learning_rate/qwen_vl_interface': 5.712083839899991e-06, 'learning_rate/action_model': 5.712083839899991e-05, 'epoch': 1.37}) +[2026-06-17 21:59:59,087][starVLA.training.train_starvla][INFO] - Step 3337, Loss: {'train/loss': 0.4766753911972046, 'train/grad_norm_pre_clip': 16.399911880493164, 'timing/data': 0.0005128495395183563, 'timing/model': 1.1524207415059209, 'learning_rate/qwen_vl_interface': 5.709931359181956e-06, 'learning_rate/action_model': 5.709931359181956e-05, 'epoch': 1.37}) +[2026-06-17 22:00:03,094][starVLA.training.train_starvla][INFO] - Step 3338, Loss: {'train/loss': 0.5293921828269958, 'train/grad_norm_pre_clip': 15.355128288269043, 'timing/data': 0.0009667566046118736, 'timing/model': 1.3009499665349722, 'learning_rate/qwen_vl_interface': 5.707778783119654e-06, 'learning_rate/action_model': 5.707778783119654e-05, 'epoch': 1.37}) +[2026-06-17 22:00:06,831][starVLA.training.train_starvla][INFO] - Step 3339, Loss: {'train/loss': 0.6483402252197266, 'train/grad_norm_pre_clip': 17.171436309814453, 'timing/data': 0.00064101442694664, 'timing/model': 0.9725949624553323, 'learning_rate/qwen_vl_interface': 5.705626112159318e-06, 'learning_rate/action_model': 5.7056261121593177e-05, 'epoch': 1.37}) +[2026-06-17 22:00:10,739][starVLA.training.train_starvla][INFO] - Step 3340, Loss: {'train/loss': 0.5623124241828918, 'train/grad_norm_pre_clip': 34.48646545410156, 'timing/data': 0.0008980352431535721, 'timing/model': 0.9599367398768663, 'learning_rate/qwen_vl_interface': 5.703473346747199e-06, 'learning_rate/action_model': 5.703473346747199e-05, 'epoch': 1.37}) +[2026-06-17 22:00:14,615][starVLA.training.train_starvla][INFO] - Step 3341, Loss: {'train/loss': 0.5011716485023499, 'train/grad_norm_pre_clip': 14.373437881469727, 'timing/data': 0.0011878414079546928, 'timing/model': 0.9335527727380395, 'learning_rate/qwen_vl_interface': 5.701320487329567e-06, 'learning_rate/action_model': 5.701320487329567e-05, 'epoch': 1.38}) +[2026-06-17 22:00:18,381][starVLA.training.train_starvla][INFO] - Step 3342, Loss: {'train/loss': 0.5739957690238953, 'train/grad_norm_pre_clip': 16.339426040649414, 'timing/data': 0.0005177222192287445, 'timing/model': 0.9294370776042342, 'learning_rate/qwen_vl_interface': 5.699167534352712e-06, 'learning_rate/action_model': 5.699167534352712e-05, 'epoch': 1.38}) +[2026-06-17 22:00:22,203][starVLA.training.train_starvla][INFO] - Step 3343, Loss: {'train/loss': 0.8344119191169739, 'train/grad_norm_pre_clip': 16.622419357299805, 'timing/data': 0.0007205624133348465, 'timing/model': 0.9292141227051616, 'learning_rate/qwen_vl_interface': 5.697014488262945e-06, 'learning_rate/action_model': 5.697014488262945e-05, 'epoch': 1.38}) +[2026-06-17 22:00:26,023][starVLA.training.train_starvla][INFO] - Step 3344, Loss: {'train/loss': 0.17857369780540466, 'train/grad_norm_pre_clip': 11.09972858428955, 'timing/data': 0.001317792572081089, 'timing/model': 1.153285427019, 'learning_rate/qwen_vl_interface': 5.694861349506591e-06, 'learning_rate/action_model': 5.6948613495065903e-05, 'epoch': 1.38}) +[2026-06-17 22:00:29,645][starVLA.training.train_starvla][INFO] - Step 3345, Loss: {'train/loss': 0.3998832702636719, 'train/grad_norm_pre_clip': 15.609018325805664, 'timing/data': 0.0005530938506126404, 'timing/model': 0.930492764338851, 'learning_rate/qwen_vl_interface': 5.692708118530003e-06, 'learning_rate/action_model': 5.6927081185300025e-05, 'epoch': 1.38}) +[2026-06-17 22:00:33,546][starVLA.training.train_starvla][INFO] - Step 3346, Loss: {'train/loss': 0.5303728580474854, 'train/grad_norm_pre_clip': 17.538177490234375, 'timing/data': 0.0008572395890951157, 'timing/model': 0.953378644771874, 'learning_rate/qwen_vl_interface': 5.6905547957795446e-06, 'learning_rate/action_model': 5.690554795779545e-05, 'epoch': 1.38}) +[2026-06-17 22:00:37,311][starVLA.training.train_starvla][INFO] - Step 3347, Loss: {'train/loss': 0.6865032315254211, 'train/grad_norm_pre_clip': 15.597681045532227, 'timing/data': 0.0005523981526494026, 'timing/model': 0.9311777502298355, 'learning_rate/qwen_vl_interface': 5.688401381701604e-06, 'learning_rate/action_model': 5.688401381701603e-05, 'epoch': 1.38}) +[2026-06-17 22:00:41,130][starVLA.training.train_starvla][INFO] - Step 3348, Loss: {'train/loss': 0.5455910563468933, 'train/grad_norm_pre_clip': 17.844104766845703, 'timing/data': 0.0005078772082924843, 'timing/model': 0.938669616356492, 'learning_rate/qwen_vl_interface': 5.686247876742583e-06, 'learning_rate/action_model': 5.686247876742583e-05, 'epoch': 1.38}) +[2026-06-17 22:00:45,023][starVLA.training.train_starvla][INFO] - Step 3349, Loss: {'train/loss': 0.5958088040351868, 'train/grad_norm_pre_clip': 18.276018142700195, 'timing/data': 0.0005498528480529785, 'timing/model': 0.9338980009779334, 'learning_rate/qwen_vl_interface': 5.684094281348908e-06, 'learning_rate/action_model': 5.684094281348908e-05, 'epoch': 1.38}) +[2026-06-17 22:00:49,140][starVLA.training.train_starvla][INFO] - Step 3350, Loss: {'train/loss': 0.28645673394203186, 'train/grad_norm_pre_clip': 10.72225570678711, 'timing/data': 0.0006850305944681168, 'timing/model': 1.2640919974073768, 'learning_rate/qwen_vl_interface': 5.681940595967023e-06, 'learning_rate/action_model': 5.681940595967024e-05, 'epoch': 1.38}) +[2026-06-17 22:00:52,939][starVLA.training.train_starvla][INFO] - Step 3351, Loss: {'train/loss': 0.4290629029273987, 'train/grad_norm_pre_clip': 12.987688064575195, 'timing/data': 0.0011149998754262924, 'timing/model': 1.1124454913660884, 'learning_rate/qwen_vl_interface': 5.679786821043388e-06, 'learning_rate/action_model': 5.679786821043388e-05, 'epoch': 1.38}) +[2026-06-17 22:00:56,563][starVLA.training.train_starvla][INFO] - Step 3352, Loss: {'train/loss': 0.39314690232276917, 'train/grad_norm_pre_clip': 15.177228927612305, 'timing/data': 0.0008169850334525108, 'timing/model': 0.9320059427991509, 'learning_rate/qwen_vl_interface': 5.677632957024482e-06, 'learning_rate/action_model': 5.6776329570244826e-05, 'epoch': 1.38}) +[2026-06-17 22:01:00,691][starVLA.training.train_starvla][INFO] - Step 3353, Loss: {'train/loss': 0.8235259652137756, 'train/grad_norm_pre_clip': 21.471553802490234, 'timing/data': 0.0008316840976476669, 'timing/model': 0.9434304060414433, 'learning_rate/qwen_vl_interface': 5.675479004356803e-06, 'learning_rate/action_model': 5.675479004356803e-05, 'epoch': 1.38}) +[2026-06-17 22:01:04,503][starVLA.training.train_starvla][INFO] - Step 3354, Loss: {'train/loss': 0.33314257860183716, 'train/grad_norm_pre_clip': 13.111519813537598, 'timing/data': 0.0005154265090823174, 'timing/model': 0.9318834189325571, 'learning_rate/qwen_vl_interface': 5.673324963486871e-06, 'learning_rate/action_model': 5.67332496348687e-05, 'epoch': 1.38}) +[2026-06-17 22:01:08,393][starVLA.training.train_starvla][INFO] - Step 3355, Loss: {'train/loss': 0.25713029503822327, 'train/grad_norm_pre_clip': 11.092573165893555, 'timing/data': 0.0005126995965838432, 'timing/model': 0.934185354039073, 'learning_rate/qwen_vl_interface': 5.671170834861215e-06, 'learning_rate/action_model': 5.671170834861215e-05, 'epoch': 1.38}) +[2026-06-17 22:01:12,230][starVLA.training.train_starvla][INFO] - Step 3356, Loss: {'train/loss': 0.2700342535972595, 'train/grad_norm_pre_clip': 10.788923263549805, 'timing/data': 0.0014120535925030708, 'timing/model': 0.9388302192091942, 'learning_rate/qwen_vl_interface': 5.6690166189263925e-06, 'learning_rate/action_model': 5.669016618926393e-05, 'epoch': 1.38}) +[2026-06-17 22:01:16,099][starVLA.training.train_starvla][INFO] - Step 3357, Loss: {'train/loss': 0.4813408851623535, 'train/grad_norm_pre_clip': 20.323162078857422, 'timing/data': 0.0006116311997175217, 'timing/model': 1.1830174280330539, 'learning_rate/qwen_vl_interface': 5.666862316128977e-06, 'learning_rate/action_model': 5.666862316128977e-05, 'epoch': 1.38}) +[2026-06-17 22:01:21,465][starVLA.training.train_starvla][INFO] - Step 3358, Loss: {'train/loss': 0.3958357572555542, 'train/grad_norm_pre_clip': 14.418578147888184, 'timing/data': 0.0005442602559924126, 'timing/model': 2.6433397009968758, 'learning_rate/qwen_vl_interface': 5.664707926915551e-06, 'learning_rate/action_model': 5.66470792691555e-05, 'epoch': 1.38}) +[2026-06-17 22:01:25,136][starVLA.training.train_starvla][INFO] - Step 3359, Loss: {'train/loss': 0.2999207079410553, 'train/grad_norm_pre_clip': 11.927539825439453, 'timing/data': 0.0005376953631639481, 'timing/model': 0.945470692589879, 'learning_rate/qwen_vl_interface': 5.662553451732729e-06, 'learning_rate/action_model': 5.6625534517327285e-05, 'epoch': 1.38}) +[2026-06-17 22:01:29,033][starVLA.training.train_starvla][INFO] - Step 3360, Loss: {'train/loss': 0.34255456924438477, 'train/grad_norm_pre_clip': 15.869475364685059, 'timing/data': 0.0005349181592464447, 'timing/model': 0.9451853604987264, 'learning_rate/qwen_vl_interface': 5.660398891027129e-06, 'learning_rate/action_model': 5.660398891027129e-05, 'epoch': 1.38}) +[2026-06-17 22:01:32,840][starVLA.training.train_starvla][INFO] - Step 3361, Loss: {'train/loss': 0.35695263743400574, 'train/grad_norm_pre_clip': 12.916611671447754, 'timing/data': 0.000539126805961132, 'timing/model': 0.9307602634653449, 'learning_rate/qwen_vl_interface': 5.658244245245399e-06, 'learning_rate/action_model': 5.658244245245399e-05, 'epoch': 1.38}) +[2026-06-17 22:01:36,657][starVLA.training.train_starvla][INFO] - Step 3362, Loss: {'train/loss': 0.31073206663131714, 'train/grad_norm_pre_clip': 12.892255783081055, 'timing/data': 0.0008948585018515587, 'timing/model': 0.934915523044765, 'learning_rate/qwen_vl_interface': 5.6560895148341985e-06, 'learning_rate/action_model': 5.656089514834199e-05, 'epoch': 1.38}) +[2026-06-17 22:01:40,448][starVLA.training.train_starvla][INFO] - Step 3363, Loss: {'train/loss': 0.22645704448223114, 'train/grad_norm_pre_clip': 9.125191688537598, 'timing/data': 0.0008070878684520721, 'timing/model': 0.931610937230289, 'learning_rate/qwen_vl_interface': 5.653934700240204e-06, 'learning_rate/action_model': 5.6539347002402034e-05, 'epoch': 1.38}) +[2026-06-17 22:01:44,251][starVLA.training.train_starvla][INFO] - Step 3364, Loss: {'train/loss': 0.5107120275497437, 'train/grad_norm_pre_clip': 14.915477752685547, 'timing/data': 0.000504019670188427, 'timing/model': 1.1218958673998713, 'learning_rate/qwen_vl_interface': 5.651779801910111e-06, 'learning_rate/action_model': 5.651779801910111e-05, 'epoch': 1.38}) +[2026-06-17 22:01:47,975][starVLA.training.train_starvla][INFO] - Step 3365, Loss: {'train/loss': 0.39376354217529297, 'train/grad_norm_pre_clip': 11.953062057495117, 'timing/data': 0.0008170483633875847, 'timing/model': 1.0432560117915273, 'learning_rate/qwen_vl_interface': 5.649624820290631e-06, 'learning_rate/action_model': 5.649624820290631e-05, 'epoch': 1.39}) +[2026-06-17 22:01:51,568][starVLA.training.train_starvla][INFO] - Step 3366, Loss: {'train/loss': 0.5059072375297546, 'train/grad_norm_pre_clip': 13.835002899169922, 'timing/data': 0.000597560778260231, 'timing/model': 0.9306113943457603, 'learning_rate/qwen_vl_interface': 5.647469755828501e-06, 'learning_rate/action_model': 5.647469755828501e-05, 'epoch': 1.39}) +[2026-06-17 22:01:55,309][starVLA.training.train_starvla][INFO] - Step 3367, Loss: {'train/loss': 0.303264319896698, 'train/grad_norm_pre_clip': 11.090461730957031, 'timing/data': 0.000732039101421833, 'timing/model': 0.9306717943400145, 'learning_rate/qwen_vl_interface': 5.645314608970457e-06, 'learning_rate/action_model': 5.6453146089704566e-05, 'epoch': 1.39}) +[2026-06-17 22:01:59,161][starVLA.training.train_starvla][INFO] - Step 3368, Loss: {'train/loss': 0.3137814402580261, 'train/grad_norm_pre_clip': 12.88454532623291, 'timing/data': 0.0004976419731974602, 'timing/model': 0.929715046659112, 'learning_rate/qwen_vl_interface': 5.643159380163273e-06, 'learning_rate/action_model': 5.643159380163273e-05, 'epoch': 1.39}) +[2026-06-17 22:02:02,959][starVLA.training.train_starvla][INFO] - Step 3369, Loss: {'train/loss': 0.3904559016227722, 'train/grad_norm_pre_clip': 13.873736381530762, 'timing/data': 0.0012334585189819336, 'timing/model': 0.9290956612676382, 'learning_rate/qwen_vl_interface': 5.641004069853725e-06, 'learning_rate/action_model': 5.641004069853725e-05, 'epoch': 1.39}) +[2026-06-17 22:02:06,731][starVLA.training.train_starvla][INFO] - Step 3370, Loss: {'train/loss': 0.3579692244529724, 'train/grad_norm_pre_clip': 11.300124168395996, 'timing/data': 0.0005117375403642654, 'timing/model': 1.1072420356795192, 'learning_rate/qwen_vl_interface': 5.638848678488613e-06, 'learning_rate/action_model': 5.638848678488613e-05, 'epoch': 1.39}) +[2026-06-17 22:02:10,444][starVLA.training.train_starvla][INFO] - Step 3371, Loss: {'train/loss': 0.5377157926559448, 'train/grad_norm_pre_clip': 18.976680755615234, 'timing/data': 0.0005043288692831993, 'timing/model': 1.0408426560461521, 'learning_rate/qwen_vl_interface': 5.636693206514752e-06, 'learning_rate/action_model': 5.636693206514752e-05, 'epoch': 1.39}) +[2026-06-17 22:02:14,047][starVLA.training.train_starvla][INFO] - Step 3372, Loss: {'train/loss': 0.1992635428905487, 'train/grad_norm_pre_clip': 12.357730865478516, 'timing/data': 0.0005662878975272179, 'timing/model': 0.9223742187023163, 'learning_rate/qwen_vl_interface': 5.634537654378973e-06, 'learning_rate/action_model': 5.634537654378973e-05, 'epoch': 1.39}) +[2026-06-17 22:02:17,783][starVLA.training.train_starvla][INFO] - Step 3373, Loss: {'train/loss': 0.323460191488266, 'train/grad_norm_pre_clip': 12.70779800415039, 'timing/data': 0.0005401009693741798, 'timing/model': 0.9289889987558126, 'learning_rate/qwen_vl_interface': 5.632382022528123e-06, 'learning_rate/action_model': 5.632382022528123e-05, 'epoch': 1.39}) +[2026-06-17 22:02:21,526][starVLA.training.train_starvla][INFO] - Step 3374, Loss: {'train/loss': 0.3469974994659424, 'train/grad_norm_pre_clip': 12.093579292297363, 'timing/data': 0.0005278093740344048, 'timing/model': 0.9275830611586571, 'learning_rate/qwen_vl_interface': 5.6302263114090705e-06, 'learning_rate/action_model': 5.63022631140907e-05, 'epoch': 1.39}) +[2026-06-17 22:02:25,270][starVLA.training.train_starvla][INFO] - Step 3375, Loss: {'train/loss': 0.48829299211502075, 'train/grad_norm_pre_clip': 18.111572265625, 'timing/data': 0.0004950845614075661, 'timing/model': 0.9239485282450914, 'learning_rate/qwen_vl_interface': 5.628070521468693e-06, 'learning_rate/action_model': 5.628070521468692e-05, 'epoch': 1.39}) +[2026-06-17 22:02:28,958][starVLA.training.train_starvla][INFO] - Step 3376, Loss: {'train/loss': 0.38410961627960205, 'train/grad_norm_pre_clip': 18.360557556152344, 'timing/data': 0.0005135107785463333, 'timing/model': 0.9228288372978568, 'learning_rate/qwen_vl_interface': 5.625914653153889e-06, 'learning_rate/action_model': 5.625914653153889e-05, 'epoch': 1.39}) +[2026-06-17 22:02:32,764][starVLA.training.train_starvla][INFO] - Step 3377, Loss: {'train/loss': 0.49991410970687866, 'train/grad_norm_pre_clip': 15.974690437316895, 'timing/data': 0.0007224204018712044, 'timing/model': 1.124625182710588, 'learning_rate/qwen_vl_interface': 5.623758706911574e-06, 'learning_rate/action_model': 5.6237587069115737e-05, 'epoch': 1.39}) +[2026-06-17 22:02:36,361][starVLA.training.train_starvla][INFO] - Step 3378, Loss: {'train/loss': 0.6099255681037903, 'train/grad_norm_pre_clip': 18.57029914855957, 'timing/data': 0.0009480481967329979, 'timing/model': 0.9252964351326227, 'learning_rate/qwen_vl_interface': 5.621602683188677e-06, 'learning_rate/action_model': 5.621602683188677e-05, 'epoch': 1.39}) +[2026-06-17 22:02:40,159][starVLA.training.train_starvla][INFO] - Step 3379, Loss: {'train/loss': 0.32178524136543274, 'train/grad_norm_pre_clip': 14.252447128295898, 'timing/data': 0.0013372451066970825, 'timing/model': 0.940522488206625, 'learning_rate/qwen_vl_interface': 5.619446582432143e-06, 'learning_rate/action_model': 5.619446582432143e-05, 'epoch': 1.39}) +[2026-06-17 22:02:43,898][starVLA.training.train_starvla][INFO] - Step 3380, Loss: {'train/loss': 0.2399822622537613, 'train/grad_norm_pre_clip': 36.050045013427734, 'timing/data': 0.0005438569933176041, 'timing/model': 0.9317602105438709, 'learning_rate/qwen_vl_interface': 5.617290405088936e-06, 'learning_rate/action_model': 5.617290405088936e-05, 'epoch': 1.39}) +[2026-06-17 22:02:47,664][starVLA.training.train_starvla][INFO] - Step 3381, Loss: {'train/loss': 0.21395263075828552, 'train/grad_norm_pre_clip': 10.783390998840332, 'timing/data': 0.0012813983485102654, 'timing/model': 0.9279141584411263, 'learning_rate/qwen_vl_interface': 5.615134151606033e-06, 'learning_rate/action_model': 5.615134151606033e-05, 'epoch': 1.39}) +[2026-06-17 22:02:51,443][starVLA.training.train_starvla][INFO] - Step 3382, Loss: {'train/loss': 0.41662976145744324, 'train/grad_norm_pre_clip': 13.241400718688965, 'timing/data': 0.0012075314298272133, 'timing/model': 0.9435178469866514, 'learning_rate/qwen_vl_interface': 5.612977822430425e-06, 'learning_rate/action_model': 5.612977822430425e-05, 'epoch': 1.39}) +[2026-06-17 22:02:55,230][starVLA.training.train_starvla][INFO] - Step 3383, Loss: {'train/loss': 0.46220266819000244, 'train/grad_norm_pre_clip': 17.931697845458984, 'timing/data': 0.000509314239025116, 'timing/model': 1.103320418857038, 'learning_rate/qwen_vl_interface': 5.610821418009127e-06, 'learning_rate/action_model': 5.610821418009127e-05, 'epoch': 1.39}) +[2026-06-17 22:02:58,832][starVLA.training.train_starvla][INFO] - Step 3384, Loss: {'train/loss': 0.2797935903072357, 'train/grad_norm_pre_clip': 14.676048278808594, 'timing/data': 0.000576515682041645, 'timing/model': 0.9332706183195114, 'learning_rate/qwen_vl_interface': 5.608664938789159e-06, 'learning_rate/action_model': 5.608664938789159e-05, 'epoch': 1.39}) +[2026-06-17 22:03:02,604][starVLA.training.train_starvla][INFO] - Step 3385, Loss: {'train/loss': 0.29002347588539124, 'train/grad_norm_pre_clip': 16.124284744262695, 'timing/data': 0.0005349349230527878, 'timing/model': 0.9306454928591847, 'learning_rate/qwen_vl_interface': 5.606508385217564e-06, 'learning_rate/action_model': 5.6065083852175636e-05, 'epoch': 1.39}) +[2026-06-17 22:03:06,389][starVLA.training.train_starvla][INFO] - Step 3386, Loss: {'train/loss': 0.41903501749038696, 'train/grad_norm_pre_clip': 16.281253814697266, 'timing/data': 0.0005343984812498093, 'timing/model': 0.9448823723942041, 'learning_rate/qwen_vl_interface': 5.604351757741396e-06, 'learning_rate/action_model': 5.6043517577413964e-05, 'epoch': 1.39}) +[2026-06-17 22:03:10,243][starVLA.training.train_starvla][INFO] - Step 3387, Loss: {'train/loss': 0.3838043212890625, 'train/grad_norm_pre_clip': 13.42078685760498, 'timing/data': 0.0006817858666181564, 'timing/model': 0.9471740378066897, 'learning_rate/qwen_vl_interface': 5.602195056807727e-06, 'learning_rate/action_model': 5.602195056807727e-05, 'epoch': 1.39}) +[2026-06-17 22:03:14,104][starVLA.training.train_starvla][INFO] - Step 3388, Loss: {'train/loss': 0.5957997441291809, 'train/grad_norm_pre_clip': 19.548566818237305, 'timing/data': 0.0005366681143641472, 'timing/model': 0.9333221316337585, 'learning_rate/qwen_vl_interface': 5.600038282863647e-06, 'learning_rate/action_model': 5.6000382828636466e-05, 'epoch': 1.39}) +[2026-06-17 22:03:17,932][starVLA.training.train_starvla][INFO] - Step 3389, Loss: {'train/loss': 0.3519044518470764, 'train/grad_norm_pre_clip': 13.107592582702637, 'timing/data': 0.0006451504305005074, 'timing/model': 0.9305549031123519, 'learning_rate/qwen_vl_interface': 5.597881436356252e-06, 'learning_rate/action_model': 5.597881436356252e-05, 'epoch': 1.4}) +[2026-06-17 22:03:21,722][starVLA.training.train_starvla][INFO] - Step 3390, Loss: {'train/loss': 0.7156472206115723, 'train/grad_norm_pre_clip': 17.76431655883789, 'timing/data': 0.0005626380443572998, 'timing/model': 1.1273447517305613, 'learning_rate/qwen_vl_interface': 5.595724517732661e-06, 'learning_rate/action_model': 5.595724517732661e-05, 'epoch': 1.4}) +[2026-06-17 22:03:25,357][starVLA.training.train_starvla][INFO] - Step 3391, Loss: {'train/loss': 0.4477138817310333, 'train/grad_norm_pre_clip': 19.07063865661621, 'timing/data': 0.0004975674673914909, 'timing/model': 0.937216310761869, 'learning_rate/qwen_vl_interface': 5.5935675274400046e-06, 'learning_rate/action_model': 5.593567527440005e-05, 'epoch': 1.4}) +[2026-06-17 22:03:29,416][starVLA.training.train_starvla][INFO] - Step 3392, Loss: {'train/loss': 0.4612554609775543, 'train/grad_norm_pre_clip': 15.251870155334473, 'timing/data': 0.00055682472884655, 'timing/model': 0.9394140224903822, 'learning_rate/qwen_vl_interface': 5.591410465925431e-06, 'learning_rate/action_model': 5.5914104659254306e-05, 'epoch': 1.4}) +[2026-06-17 22:03:33,306][starVLA.training.train_starvla][INFO] - Step 3393, Loss: {'train/loss': 0.45960482954978943, 'train/grad_norm_pre_clip': 16.59115982055664, 'timing/data': 0.0006342437118291855, 'timing/model': 0.9594354685395956, 'learning_rate/qwen_vl_interface': 5.589253333636101e-06, 'learning_rate/action_model': 5.589253333636101e-05, 'epoch': 1.4}) +[2026-06-17 22:03:37,134][starVLA.training.train_starvla][INFO] - Step 3394, Loss: {'train/loss': 0.43473517894744873, 'train/grad_norm_pre_clip': 18.808618545532227, 'timing/data': 0.0012327441945672035, 'timing/model': 0.9310261998325586, 'learning_rate/qwen_vl_interface': 5.587096131019189e-06, 'learning_rate/action_model': 5.5870961310191896e-05, 'epoch': 1.4}) +[2026-06-17 22:03:40,913][starVLA.training.train_starvla][INFO] - Step 3395, Loss: {'train/loss': 0.6239139437675476, 'train/grad_norm_pre_clip': 17.490272521972656, 'timing/data': 0.0005489056929945946, 'timing/model': 0.936576334759593, 'learning_rate/qwen_vl_interface': 5.5849388585218865e-06, 'learning_rate/action_model': 5.5849388585218866e-05, 'epoch': 1.4}) +[2026-06-17 22:03:44,763][starVLA.training.train_starvla][INFO] - Step 3396, Loss: {'train/loss': 0.6529374122619629, 'train/grad_norm_pre_clip': 18.365697860717773, 'timing/data': 0.0005793208256363869, 'timing/model': 1.1278323642909527, 'learning_rate/qwen_vl_interface': 5.582781516591398e-06, 'learning_rate/action_model': 5.5827815165913977e-05, 'epoch': 1.4}) +[2026-06-17 22:03:48,575][starVLA.training.train_starvla][INFO] - Step 3397, Loss: {'train/loss': 0.3212575614452362, 'train/grad_norm_pre_clip': 14.128260612487793, 'timing/data': 0.0005264785140752792, 'timing/model': 1.1318518202751875, 'learning_rate/qwen_vl_interface': 5.580624105674945e-06, 'learning_rate/action_model': 5.580624105674944e-05, 'epoch': 1.4}) +[2026-06-17 22:03:52,244][starVLA.training.train_starvla][INFO] - Step 3398, Loss: {'train/loss': 0.4044998586177826, 'train/grad_norm_pre_clip': 13.04615592956543, 'timing/data': 0.0008702250197529793, 'timing/model': 0.9358458342030644, 'learning_rate/qwen_vl_interface': 5.578466626219756e-06, 'learning_rate/action_model': 5.578466626219756e-05, 'epoch': 1.4}) +[2026-06-17 22:03:56,174][starVLA.training.train_starvla][INFO] - Step 3399, Loss: {'train/loss': 0.4744359850883484, 'train/grad_norm_pre_clip': 20.658721923828125, 'timing/data': 0.0005274452269077301, 'timing/model': 1.001354799605906, 'learning_rate/qwen_vl_interface': 5.576309078673083e-06, 'learning_rate/action_model': 5.576309078673083e-05, 'epoch': 1.4}) +[2026-06-17 22:06:08,390][starVLA.training.train_starvla][INFO] - Step 3400, Loss: {'train/loss': 0.6379689574241638, 'train/grad_norm_pre_clip': 17.412391662597656, 'eval/loss': 0.4808656573295593, 'eval/latency_6/loss': 0.4808656573295593, 'eval/demon_fire_precision': 0.9560756652773325, 'eval/demon_fire_recall': 0.9638009049773756, 'eval/demon_fire_f1': 0.9599227426364075, 'eval/demon_fire_rate_model': 0.8678352810239288, 'eval/demon_fire_rate_teacher': 0.860879243183083, 'eval/demon_left_precision': 0.755700325732899, 'eval/demon_left_recall': 0.8140350877192982, 'eval/demon_left_f1': 0.7837837837837838, 'eval/demon_left_rate_model': 0.17084028937117418, 'eval/demon_left_rate_teacher': 0.15859766277128548, 'eval/demon_right_precision': 0.7908082408874801, 'eval/demon_right_recall': 0.7665130568356375, 'eval/demon_right_f1': 0.7784711388455539, 'eval/demon_right_rate_model': 0.17557039510294936, 'eval/demon_right_rate_teacher': 0.18113522537562604, 'eval/demon_move_f1': 0.7811274613146688, 'eval/demon_cc_f1': 0.8982421109078172, 'eval/latency_6/demon_fire_precision': 0.9560756652773325, 'eval/latency_6/demon_fire_recall': 0.9638009049773756, 'eval/latency_6/demon_fire_f1': 0.9599227426364075, 'eval/latency_6/demon_fire_rate_model': 0.8678352810239288, 'eval/latency_6/demon_fire_rate_teacher': 0.860879243183083, 'eval/latency_6/demon_left_precision': 0.755700325732899, 'eval/latency_6/demon_left_recall': 0.8140350877192982, 'eval/latency_6/demon_left_f1': 0.7837837837837838, 'eval/latency_6/demon_left_rate_model': 0.17084028937117418, 'eval/latency_6/demon_left_rate_teacher': 0.15859766277128548, 'eval/latency_6/demon_right_precision': 0.7908082408874801, 'eval/latency_6/demon_right_recall': 0.7665130568356375, 'eval/latency_6/demon_right_f1': 0.7784711388455539, 'eval/latency_6/demon_right_rate_model': 0.17557039510294936, 'eval/latency_6/demon_right_rate_teacher': 0.18113522537562604, 'eval/latency_6/demon_move_f1': 0.7811274613146688, 'eval/latency_6/demon_cc_f1': 0.8982421109078172, 'timing/data': 0.000526069663465023, 'timing/model': 0.9458176903426647, 'learning_rate/qwen_vl_interface': 5.5741514634821885e-06, 'learning_rate/action_model': 5.574151463482188e-05, 'epoch': 1.4}) +[2026-06-17 22:06:08,393][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_3400_state +[2026-06-17 22:06:08,398][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 22:06:15,187][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_3400_state/model.safetensors +[2026-06-17 22:06:30,533][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_3400_state/optimizer.bin +[2026-06-17 22:06:30,534][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_3400_state/sampler.bin +[2026-06-17 22:06:30,537][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_3400_state/sampler_1.bin +[2026-06-17 22:06:30,539][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_3400_state/random_states_0.pkl +[2026-06-17 22:06:30,543][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3400 +[2026-06-17 22:06:30,543][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 22:06:30,553][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 22:06:34,545][starVLA.training.train_starvla][INFO] - Step 3401, Loss: {'train/loss': 0.5091054439544678, 'train/grad_norm_pre_clip': 17.40814971923828, 'timing/data': 0.0009660655632615089, 'timing/model': 0.9270346574485302, 'learning_rate/qwen_vl_interface': 5.571993781094345e-06, 'learning_rate/action_model': 5.5719937810943444e-05, 'epoch': 1.4}) +[2026-06-17 22:06:38,442][starVLA.training.train_starvla][INFO] - Step 3402, Loss: {'train/loss': 0.46904638409614563, 'train/grad_norm_pre_clip': 14.459294319152832, 'timing/data': 0.0005269674584269524, 'timing/model': 0.9612642535939813, 'learning_rate/qwen_vl_interface': 5.569836031956845e-06, 'learning_rate/action_model': 5.5698360319568454e-05, 'epoch': 1.4}) +[2026-06-17 22:06:42,238][starVLA.training.train_starvla][INFO] - Step 3403, Loss: {'train/loss': 0.6072692275047302, 'train/grad_norm_pre_clip': 18.73257827758789, 'timing/data': 0.0010962309315800667, 'timing/model': 0.9546101512387395, 'learning_rate/qwen_vl_interface': 5.5676782165169915e-06, 'learning_rate/action_model': 5.567678216516991e-05, 'epoch': 1.4}) +[2026-06-17 22:06:46,043][starVLA.training.train_starvla][INFO] - Step 3404, Loss: {'train/loss': 0.18254823982715607, 'train/grad_norm_pre_clip': 9.122823715209961, 'timing/data': 0.0008350647985935211, 'timing/model': 0.9394053975120187, 'learning_rate/qwen_vl_interface': 5.565520335222101e-06, 'learning_rate/action_model': 5.565520335222101e-05, 'epoch': 1.4}) +[2026-06-17 22:06:49,934][starVLA.training.train_starvla][INFO] - Step 3405, Loss: {'train/loss': 0.5492792725563049, 'train/grad_norm_pre_clip': 21.15859031677246, 'timing/data': 0.0005065305158495903, 'timing/model': 1.1517018973827362, 'learning_rate/qwen_vl_interface': 5.563362388519507e-06, 'learning_rate/action_model': 5.5633623885195066e-05, 'epoch': 1.4}) +[2026-06-17 22:06:53,544][starVLA.training.train_starvla][INFO] - Step 3406, Loss: {'train/loss': 0.2643945813179016, 'train/grad_norm_pre_clip': 12.901114463806152, 'timing/data': 0.0005666408687829971, 'timing/model': 0.9374800454825163, 'learning_rate/qwen_vl_interface': 5.56120437685655e-06, 'learning_rate/action_model': 5.56120437685655e-05, 'epoch': 1.4}) +[2026-06-17 22:06:57,454][starVLA.training.train_starvla][INFO] - Step 3407, Loss: {'train/loss': 0.2517440617084503, 'train/grad_norm_pre_clip': 11.728949546813965, 'timing/data': 0.0008726464584469795, 'timing/model': 0.9551529856398702, 'learning_rate/qwen_vl_interface': 5.559046300680593e-06, 'learning_rate/action_model': 5.5590463006805925e-05, 'epoch': 1.4}) +[2026-06-17 22:07:01,340][starVLA.training.train_starvla][INFO] - Step 3408, Loss: {'train/loss': 0.31607773900032043, 'train/grad_norm_pre_clip': 11.879988670349121, 'timing/data': 0.0004946338012814522, 'timing/model': 0.9574950207024813, 'learning_rate/qwen_vl_interface': 5.556888160439006e-06, 'learning_rate/action_model': 5.556888160439006e-05, 'epoch': 1.4}) +[2026-06-17 22:07:05,297][starVLA.training.train_starvla][INFO] - Step 3409, Loss: {'train/loss': 0.3332369327545166, 'train/grad_norm_pre_clip': 99.52397155761719, 'timing/data': 0.000690847635269165, 'timing/model': 0.9685022542253137, 'learning_rate/qwen_vl_interface': 5.5547299565791725e-06, 'learning_rate/action_model': 5.5547299565791723e-05, 'epoch': 1.4}) +[2026-06-17 22:07:09,147][starVLA.training.train_starvla][INFO] - Step 3410, Loss: {'train/loss': 0.47505983710289, 'train/grad_norm_pre_clip': 19.0761661529541, 'timing/data': 0.0007437728345394135, 'timing/model': 0.9607014879584312, 'learning_rate/qwen_vl_interface': 5.55257168954849e-06, 'learning_rate/action_model': 5.5525716895484905e-05, 'epoch': 1.4}) +[2026-06-17 22:07:13,110][starVLA.training.train_starvla][INFO] - Step 3411, Loss: {'train/loss': 0.65376877784729, 'train/grad_norm_pre_clip': 15.34033203125, 'timing/data': 0.0008163135498762131, 'timing/model': 1.1170832822099328, 'learning_rate/qwen_vl_interface': 5.550413359794371e-06, 'learning_rate/action_model': 5.550413359794371e-05, 'epoch': 1.4}) +[2026-06-17 22:07:16,997][starVLA.training.train_starvla][INFO] - Step 3412, Loss: {'train/loss': 0.4654794931411743, 'train/grad_norm_pre_clip': 16.174230575561523, 'timing/data': 0.0004753163084387779, 'timing/model': 1.08494739420712, 'learning_rate/qwen_vl_interface': 5.548254967764239e-06, 'learning_rate/action_model': 5.5482549677642394e-05, 'epoch': 1.4}) +[2026-06-17 22:07:20,637][starVLA.training.train_starvla][INFO] - Step 3413, Loss: {'train/loss': 0.4999723732471466, 'train/grad_norm_pre_clip': 19.253128051757812, 'timing/data': 0.0010773297399282455, 'timing/model': 0.9427287383005023, 'learning_rate/qwen_vl_interface': 5.546096513905532e-06, 'learning_rate/action_model': 5.546096513905532e-05, 'epoch': 1.4}) +[2026-06-17 22:07:24,490][starVLA.training.train_starvla][INFO] - Step 3414, Loss: {'train/loss': 0.31171223521232605, 'train/grad_norm_pre_clip': 12.21005630493164, 'timing/data': 0.0006906520575284958, 'timing/model': 0.9495019316673279, 'learning_rate/qwen_vl_interface': 5.5439379986657e-06, 'learning_rate/action_model': 5.543937998665699e-05, 'epoch': 1.41}) +[2026-06-17 22:07:29,550][starVLA.training.train_starvla][INFO] - Step 3415, Loss: {'train/loss': 0.24368663132190704, 'train/grad_norm_pre_clip': 12.267449378967285, 'timing/data': 0.000494697131216526, 'timing/model': 0.9397892383858562, 'learning_rate/qwen_vl_interface': 5.5417794224922025e-06, 'learning_rate/action_model': 5.5417794224922023e-05, 'epoch': 1.41}) +[2026-06-17 22:07:34,296][starVLA.training.train_starvla][INFO] - Step 3416, Loss: {'train/loss': 0.48533451557159424, 'train/grad_norm_pre_clip': 19.820161819458008, 'timing/data': 0.0004958081990480423, 'timing/model': 0.9556275038048625, 'learning_rate/qwen_vl_interface': 5.539620785832517e-06, 'learning_rate/action_model': 5.5396207858325166e-05, 'epoch': 1.41}) +[2026-06-17 22:07:38,462][starVLA.training.train_starvla][INFO] - Step 3417, Loss: {'train/loss': 0.5859791040420532, 'train/grad_norm_pre_clip': 18.496109008789062, 'timing/data': 0.0004807310178875923, 'timing/model': 0.9355667643249035, 'learning_rate/qwen_vl_interface': 5.537462089134133e-06, 'learning_rate/action_model': 5.537462089134132e-05, 'epoch': 1.41}) +[2026-06-17 22:07:42,394][starVLA.training.train_starvla][INFO] - Step 3418, Loss: {'train/loss': 0.721659779548645, 'train/grad_norm_pre_clip': 24.685749053955078, 'timing/data': 0.000802924856543541, 'timing/model': 1.217939080670476, 'learning_rate/qwen_vl_interface': 5.535303332844546e-06, 'learning_rate/action_model': 5.5353033328445455e-05, 'epoch': 1.41}) +[2026-06-17 22:07:47,205][starVLA.training.train_starvla][INFO] - Step 3419, Loss: {'train/loss': 0.7319403290748596, 'train/grad_norm_pre_clip': 18.55719566345215, 'timing/data': 0.000741075724363327, 'timing/model': 2.1256322069093585, 'learning_rate/qwen_vl_interface': 5.533144517411273e-06, 'learning_rate/action_model': 5.533144517411273e-05, 'epoch': 1.41}) +[2026-06-17 22:07:51,402][starVLA.training.train_starvla][INFO] - Step 3420, Loss: {'train/loss': 0.3485015332698822, 'train/grad_norm_pre_clip': 11.901966094970703, 'timing/data': 0.0006290273740887642, 'timing/model': 1.4848914351314306, 'learning_rate/qwen_vl_interface': 5.530985643281834e-06, 'learning_rate/action_model': 5.530985643281834e-05, 'epoch': 1.41}) +[2026-06-17 22:07:55,239][starVLA.training.train_starvla][INFO] - Step 3421, Loss: {'train/loss': 0.5094232559204102, 'train/grad_norm_pre_clip': 14.76736831665039, 'timing/data': 0.0012398846447467804, 'timing/model': 0.9398992964997888, 'learning_rate/qwen_vl_interface': 5.528826710903774e-06, 'learning_rate/action_model': 5.528826710903774e-05, 'epoch': 1.41}) +[2026-06-17 22:07:59,124][starVLA.training.train_starvla][INFO] - Step 3422, Loss: {'train/loss': 0.38909825682640076, 'train/grad_norm_pre_clip': 13.01589584350586, 'timing/data': 0.0004665479063987732, 'timing/model': 0.9538639755919576, 'learning_rate/qwen_vl_interface': 5.526667720724632e-06, 'learning_rate/action_model': 5.526667720724632e-05, 'epoch': 1.41}) +[2026-06-17 22:08:02,990][starVLA.training.train_starvla][INFO] - Step 3423, Loss: {'train/loss': 0.21988478302955627, 'train/grad_norm_pre_clip': 9.744954109191895, 'timing/data': 0.0009472882375121117, 'timing/model': 0.9442811012268066, 'learning_rate/qwen_vl_interface': 5.5245086731919764e-06, 'learning_rate/action_model': 5.524508673191976e-05, 'epoch': 1.41}) +[2026-06-17 22:08:06,811][starVLA.training.train_starvla][INFO] - Step 3424, Loss: {'train/loss': 0.4281918406486511, 'train/grad_norm_pre_clip': 15.50815486907959, 'timing/data': 0.0004887301474809647, 'timing/model': 0.9552403353154659, 'learning_rate/qwen_vl_interface': 5.522349568753376e-06, 'learning_rate/action_model': 5.522349568753377e-05, 'epoch': 1.41}) +[2026-06-17 22:08:10,747][starVLA.training.train_starvla][INFO] - Step 3425, Loss: {'train/loss': 0.389238178730011, 'train/grad_norm_pre_clip': 15.272688865661621, 'timing/data': 0.0005180938169360161, 'timing/model': 1.177222984842956, 'learning_rate/qwen_vl_interface': 5.520190407856419e-06, 'learning_rate/action_model': 5.5201904078564185e-05, 'epoch': 1.41}) +[2026-06-17 22:08:14,856][starVLA.training.train_starvla][INFO] - Step 3426, Loss: {'train/loss': 0.4787568747997284, 'train/grad_norm_pre_clip': 17.58035659790039, 'timing/data': 0.0006726933643221855, 'timing/model': 1.270363688468933, 'learning_rate/qwen_vl_interface': 5.5180311909486974e-06, 'learning_rate/action_model': 5.518031190948697e-05, 'epoch': 1.41}) +[2026-06-17 22:08:18,659][starVLA.training.train_starvla][INFO] - Step 3427, Loss: {'train/loss': 0.36750030517578125, 'train/grad_norm_pre_clip': 15.085537910461426, 'timing/data': 0.001025637611746788, 'timing/model': 1.0061901872977614, 'learning_rate/qwen_vl_interface': 5.515871918477821e-06, 'learning_rate/action_model': 5.515871918477821e-05, 'epoch': 1.41}) +[2026-06-17 22:08:22,512][starVLA.training.train_starvla][INFO] - Step 3428, Loss: {'train/loss': 0.42521950602531433, 'train/grad_norm_pre_clip': 16.321001052856445, 'timing/data': 0.0005223248153924942, 'timing/model': 0.9519976638257504, 'learning_rate/qwen_vl_interface': 5.513712590891414e-06, 'learning_rate/action_model': 5.513712590891413e-05, 'epoch': 1.41}) +[2026-06-17 22:08:26,752][starVLA.training.train_starvla][INFO] - Step 3429, Loss: {'train/loss': 0.3479830324649811, 'train/grad_norm_pre_clip': 15.708857536315918, 'timing/data': 0.0005603982135653496, 'timing/model': 0.9391868412494659, 'learning_rate/qwen_vl_interface': 5.511553208637098e-06, 'learning_rate/action_model': 5.511553208637098e-05, 'epoch': 1.41}) +[2026-06-17 22:08:30,690][starVLA.training.train_starvla][INFO] - Step 3430, Loss: {'train/loss': 0.4939342439174652, 'train/grad_norm_pre_clip': 17.870960235595703, 'timing/data': 0.0016254596412181854, 'timing/model': 0.9392411643639207, 'learning_rate/qwen_vl_interface': 5.509393772162524e-06, 'learning_rate/action_model': 5.509393772162523e-05, 'epoch': 1.41}) +[2026-06-17 22:08:34,566][starVLA.training.train_starvla][INFO] - Step 3431, Loss: {'train/loss': 0.5204083323478699, 'train/grad_norm_pre_clip': 14.643505096435547, 'timing/data': 0.0004928316920995712, 'timing/model': 1.144625537097454, 'learning_rate/qwen_vl_interface': 5.50723428191534e-06, 'learning_rate/action_model': 5.50723428191534e-05, 'epoch': 1.41}) +[2026-06-17 22:08:38,341][starVLA.training.train_starvla][INFO] - Step 3432, Loss: {'train/loss': 0.34430527687072754, 'train/grad_norm_pre_clip': 13.027017593383789, 'timing/data': 0.0012723663821816444, 'timing/model': 1.048470919020474, 'learning_rate/qwen_vl_interface': 5.505074738343212e-06, 'learning_rate/action_model': 5.505074738343212e-05, 'epoch': 1.41}) +[2026-06-17 22:08:42,050][starVLA.training.train_starvla][INFO] - Step 3433, Loss: {'train/loss': 0.5774886608123779, 'train/grad_norm_pre_clip': 18.76243782043457, 'timing/data': 0.0010403692722320557, 'timing/model': 0.9383319728076458, 'learning_rate/qwen_vl_interface': 5.502915141893819e-06, 'learning_rate/action_model': 5.5029151418938184e-05, 'epoch': 1.41}) +[2026-06-17 22:08:45,988][starVLA.training.train_starvla][INFO] - Step 3434, Loss: {'train/loss': 0.28388506174087524, 'train/grad_norm_pre_clip': 12.097818374633789, 'timing/data': 0.0004952140152454376, 'timing/model': 1.1060364004224539, 'learning_rate/qwen_vl_interface': 5.5007554930148436e-06, 'learning_rate/action_model': 5.5007554930148434e-05, 'epoch': 1.41}) +[2026-06-17 22:08:49,807][starVLA.training.train_starvla][INFO] - Step 3435, Loss: {'train/loss': 0.32009780406951904, 'train/grad_norm_pre_clip': 13.44589614868164, 'timing/data': 0.0005318038165569305, 'timing/model': 0.9538705199956894, 'learning_rate/qwen_vl_interface': 5.498595792153985e-06, 'learning_rate/action_model': 5.498595792153984e-05, 'epoch': 1.41}) +[2026-06-17 22:08:53,615][starVLA.training.train_starvla][INFO] - Step 3436, Loss: {'train/loss': 0.4519582688808441, 'train/grad_norm_pre_clip': 17.153114318847656, 'timing/data': 0.00047555938363075256, 'timing/model': 0.9402027567848563, 'learning_rate/qwen_vl_interface': 5.49643603975895e-06, 'learning_rate/action_model': 5.4964360397589495e-05, 'epoch': 1.41}) +[2026-06-17 22:08:57,425][starVLA.training.train_starvla][INFO] - Step 3437, Loss: {'train/loss': 0.2256566286087036, 'train/grad_norm_pre_clip': 14.787625312805176, 'timing/data': 0.0007385751232504845, 'timing/model': 0.9598179822787642, 'learning_rate/qwen_vl_interface': 5.494276236277458e-06, 'learning_rate/action_model': 5.4942762362774584e-05, 'epoch': 1.41}) +[2026-06-17 22:09:01,232][starVLA.training.train_starvla][INFO] - Step 3438, Loss: {'train/loss': 0.4420138895511627, 'train/grad_norm_pre_clip': 14.17868423461914, 'timing/data': 0.0005850354209542274, 'timing/model': 1.1064322739839554, 'learning_rate/qwen_vl_interface': 5.492116382157242e-06, 'learning_rate/action_model': 5.492116382157242e-05, 'epoch': 1.42}) +[2026-06-17 22:09:04,908][starVLA.training.train_starvla][INFO] - Step 3439, Loss: {'train/loss': 0.46866005659103394, 'train/grad_norm_pre_clip': 17.239337921142578, 'timing/data': 0.0007042335346341133, 'timing/model': 0.9587141387164593, 'learning_rate/qwen_vl_interface': 5.489956477846038e-06, 'learning_rate/action_model': 5.489956477846038e-05, 'epoch': 1.42}) +[2026-06-17 22:09:08,911][starVLA.training.train_starvla][INFO] - Step 3440, Loss: {'train/loss': 0.46260517835617065, 'train/grad_norm_pre_clip': 16.454431533813477, 'timing/data': 0.0005139261484146118, 'timing/model': 1.0366423390805721, 'learning_rate/qwen_vl_interface': 5.4877965237915984e-06, 'learning_rate/action_model': 5.4877965237915986e-05, 'epoch': 1.42}) +[2026-06-17 22:09:12,743][starVLA.training.train_starvla][INFO] - Step 3441, Loss: {'train/loss': 0.4621516466140747, 'train/grad_norm_pre_clip': 16.547216415405273, 'timing/data': 0.0008854912593960762, 'timing/model': 0.944560931995511, 'learning_rate/qwen_vl_interface': 5.485636520441684e-06, 'learning_rate/action_model': 5.4856365204416836e-05, 'epoch': 1.42}) +[2026-06-17 22:09:16,567][starVLA.training.train_starvla][INFO] - Step 3442, Loss: {'train/loss': 0.22449623048305511, 'train/grad_norm_pre_clip': 7.166187763214111, 'timing/data': 0.0005219271406531334, 'timing/model': 0.9290595725178719, 'learning_rate/qwen_vl_interface': 5.483476468244063e-06, 'learning_rate/action_model': 5.483476468244063e-05, 'epoch': 1.42}) +[2026-06-17 22:09:20,451][starVLA.training.train_starvla][INFO] - Step 3443, Loss: {'train/loss': 0.30394068360328674, 'train/grad_norm_pre_clip': 10.218269348144531, 'timing/data': 0.0004921620711684227, 'timing/model': 0.9372951705008745, 'learning_rate/qwen_vl_interface': 5.481316367646522e-06, 'learning_rate/action_model': 5.4813163676465216e-05, 'epoch': 1.42}) +[2026-06-17 22:09:24,343][starVLA.training.train_starvla][INFO] - Step 3444, Loss: {'train/loss': 0.3026663362979889, 'train/grad_norm_pre_clip': 11.36663818359375, 'timing/data': 0.0012427931651473045, 'timing/model': 1.164722797460854, 'learning_rate/qwen_vl_interface': 5.479156219096847e-06, 'learning_rate/action_model': 5.479156219096847e-05, 'epoch': 1.42}) +[2026-06-17 22:09:28,013][starVLA.training.train_starvla][INFO] - Step 3445, Loss: {'train/loss': 0.17484335601329803, 'train/grad_norm_pre_clip': 10.36350154876709, 'timing/data': 0.0004958193749189377, 'timing/model': 0.9509956305846572, 'learning_rate/qwen_vl_interface': 5.476996023042843e-06, 'learning_rate/action_model': 5.476996023042843e-05, 'epoch': 1.42}) +[2026-06-17 22:09:31,905][starVLA.training.train_starvla][INFO] - Step 3446, Loss: {'train/loss': 0.3612081706523895, 'train/grad_norm_pre_clip': 12.822836875915527, 'timing/data': 0.0007411511614918709, 'timing/model': 0.9518631789833307, 'learning_rate/qwen_vl_interface': 5.474835779932317e-06, 'learning_rate/action_model': 5.474835779932317e-05, 'epoch': 1.42}) +[2026-06-17 22:09:35,748][starVLA.training.train_starvla][INFO] - Step 3447, Loss: {'train/loss': 0.4228287935256958, 'train/grad_norm_pre_clip': 15.149450302124023, 'timing/data': 0.0009744782000780106, 'timing/model': 0.9704807987436652, 'learning_rate/qwen_vl_interface': 5.472675490213093e-06, 'learning_rate/action_model': 5.472675490213093e-05, 'epoch': 1.42}) +[2026-06-17 22:09:39,891][starVLA.training.train_starvla][INFO] - Step 3448, Loss: {'train/loss': 0.34734389185905457, 'train/grad_norm_pre_clip': 14.437724113464355, 'timing/data': 0.0005060350522398949, 'timing/model': 1.010756015777588, 'learning_rate/qwen_vl_interface': 5.470515154333e-06, 'learning_rate/action_model': 5.470515154333e-05, 'epoch': 1.42}) +[2026-06-17 22:09:43,794][starVLA.training.train_starvla][INFO] - Step 3449, Loss: {'train/loss': 0.3300137519836426, 'train/grad_norm_pre_clip': 13.612967491149902, 'timing/data': 0.000522819347679615, 'timing/model': 0.9286103015765548, 'learning_rate/qwen_vl_interface': 5.468354772739879e-06, 'learning_rate/action_model': 5.468354772739879e-05, 'epoch': 1.42}) +[2026-06-17 22:09:47,631][starVLA.training.train_starvla][INFO] - Step 3450, Loss: {'train/loss': 0.2914973795413971, 'train/grad_norm_pre_clip': 15.587727546691895, 'timing/data': 0.0005477331578731537, 'timing/model': 0.9701775265857577, 'learning_rate/qwen_vl_interface': 5.466194345881578e-06, 'learning_rate/action_model': 5.4661943458815776e-05, 'epoch': 1.42}) +[2026-06-17 22:09:51,592][starVLA.training.train_starvla][INFO] - Step 3451, Loss: {'train/loss': 0.4973466098308563, 'train/grad_norm_pre_clip': 14.971394538879395, 'timing/data': 0.0005019847303628922, 'timing/model': 1.236732012592256, 'learning_rate/qwen_vl_interface': 5.4640338742059564e-06, 'learning_rate/action_model': 5.464033874205956e-05, 'epoch': 1.42}) +[2026-06-17 22:09:55,270][starVLA.training.train_starvla][INFO] - Step 3452, Loss: {'train/loss': 0.18167075514793396, 'train/grad_norm_pre_clip': 9.940832138061523, 'timing/data': 0.0009273933246731758, 'timing/model': 0.9701857743784785, 'learning_rate/qwen_vl_interface': 5.461873358160883e-06, 'learning_rate/action_model': 5.461873358160883e-05, 'epoch': 1.42}) +[2026-06-17 22:09:59,109][starVLA.training.train_starvla][INFO] - Step 3453, Loss: {'train/loss': 0.4695097506046295, 'train/grad_norm_pre_clip': 25.02959442138672, 'timing/data': 0.0014499565586447716, 'timing/model': 0.9397058738395572, 'learning_rate/qwen_vl_interface': 5.459712798194234e-06, 'learning_rate/action_model': 5.459712798194233e-05, 'epoch': 1.42}) +[2026-06-17 22:10:02,938][starVLA.training.train_starvla][INFO] - Step 3454, Loss: {'train/loss': 0.3331677317619324, 'train/grad_norm_pre_clip': 14.631519317626953, 'timing/data': 0.001238832250237465, 'timing/model': 0.9442293550819159, 'learning_rate/qwen_vl_interface': 5.457552194753896e-06, 'learning_rate/action_model': 5.457552194753895e-05, 'epoch': 1.42}) +[2026-06-17 22:10:06,942][starVLA.training.train_starvla][INFO] - Step 3455, Loss: {'train/loss': 0.4371945559978485, 'train/grad_norm_pre_clip': 17.338787078857422, 'timing/data': 0.0013980995863676071, 'timing/model': 0.9444939894601703, 'learning_rate/qwen_vl_interface': 5.455391548287764e-06, 'learning_rate/action_model': 5.455391548287764e-05, 'epoch': 1.42}) +[2026-06-17 22:10:10,907][starVLA.training.train_starvla][INFO] - Step 3456, Loss: {'train/loss': 0.3607757091522217, 'train/grad_norm_pre_clip': 12.158203125, 'timing/data': 0.0008343067020177841, 'timing/model': 1.0305925318971276, 'learning_rate/qwen_vl_interface': 5.453230859243745e-06, 'learning_rate/action_model': 5.4532308592437444e-05, 'epoch': 1.42}) +[2026-06-17 22:10:14,747][starVLA.training.train_starvla][INFO] - Step 3457, Loss: {'train/loss': 0.3905673325061798, 'train/grad_norm_pre_clip': 11.96945858001709, 'timing/data': 0.0004689199849963188, 'timing/model': 1.1210849517956376, 'learning_rate/qwen_vl_interface': 5.451070128069748e-06, 'learning_rate/action_model': 5.4510701280697485e-05, 'epoch': 1.42}) +[2026-06-17 22:10:18,612][starVLA.training.train_starvla][INFO] - Step 3458, Loss: {'train/loss': 0.2873934209346771, 'train/grad_norm_pre_clip': 14.755876541137695, 'timing/data': 0.0007168836891651154, 'timing/model': 1.1391653940081596, 'learning_rate/qwen_vl_interface': 5.4489093552136986e-06, 'learning_rate/action_model': 5.4489093552136984e-05, 'epoch': 1.42}) +[2026-06-17 22:10:22,270][starVLA.training.train_starvla][INFO] - Step 3459, Loss: {'train/loss': 0.5211480855941772, 'train/grad_norm_pre_clip': 41.04913330078125, 'timing/data': 0.000503290444612503, 'timing/model': 0.9373689945787191, 'learning_rate/qwen_vl_interface': 5.446748541123526e-06, 'learning_rate/action_model': 5.4467485411235254e-05, 'epoch': 1.42}) +[2026-06-17 22:10:26,090][starVLA.training.train_starvla][INFO] - Step 3460, Loss: {'train/loss': 0.48124539852142334, 'train/grad_norm_pre_clip': 21.735170364379883, 'timing/data': 0.0004560081288218498, 'timing/model': 0.9441062221303582, 'learning_rate/qwen_vl_interface': 5.444587686247168e-06, 'learning_rate/action_model': 5.4445876862471675e-05, 'epoch': 1.42}) +[2026-06-17 22:10:29,971][starVLA.training.train_starvla][INFO] - Step 3461, Loss: {'train/loss': 0.5900188684463501, 'train/grad_norm_pre_clip': 20.355974197387695, 'timing/data': 0.0005239034071564674, 'timing/model': 0.9425080977380276, 'learning_rate/qwen_vl_interface': 5.442426791032574e-06, 'learning_rate/action_model': 5.4424267910325746e-05, 'epoch': 1.42}) +[2026-06-17 22:10:33,785][starVLA.training.train_starvla][INFO] - Step 3462, Loss: {'train/loss': 0.21634264290332794, 'train/grad_norm_pre_clip': 15.483479499816895, 'timing/data': 0.00048179831355810165, 'timing/model': 0.9380710171535611, 'learning_rate/qwen_vl_interface': 5.4402658559277e-06, 'learning_rate/action_model': 5.4402658559276995e-05, 'epoch': 1.43}) +[2026-06-17 22:10:37,696][starVLA.training.train_starvla][INFO] - Step 3463, Loss: {'train/loss': 0.5418242812156677, 'train/grad_norm_pre_clip': 20.1090087890625, 'timing/data': 0.0005297819152474403, 'timing/model': 1.0269938008859754, 'learning_rate/qwen_vl_interface': 5.438104881380507e-06, 'learning_rate/action_model': 5.438104881380507e-05, 'epoch': 1.43}) +[2026-06-17 22:10:41,571][starVLA.training.train_starvla][INFO] - Step 3464, Loss: {'train/loss': 0.419925719499588, 'train/grad_norm_pre_clip': 14.511094093322754, 'timing/data': 0.0005431175231933594, 'timing/model': 1.1977895926684141, 'learning_rate/qwen_vl_interface': 5.435943867838971e-06, 'learning_rate/action_model': 5.435943867838971e-05, 'epoch': 1.43}) +[2026-06-17 22:10:45,447][starVLA.training.train_starvla][INFO] - Step 3465, Loss: {'train/loss': 0.4094923436641693, 'train/grad_norm_pre_clip': 13.388463973999023, 'timing/data': 0.0005396688356995583, 'timing/model': 1.1658497583121061, 'learning_rate/qwen_vl_interface': 5.433782815751072e-06, 'learning_rate/action_model': 5.433782815751072e-05, 'epoch': 1.43}) +[2026-06-17 22:10:49,135][starVLA.training.train_starvla][INFO] - Step 3466, Loss: {'train/loss': 0.3354906737804413, 'train/grad_norm_pre_clip': 18.361059188842773, 'timing/data': 0.0008131787180900574, 'timing/model': 0.9325132789090276, 'learning_rate/qwen_vl_interface': 5.431621725564797e-06, 'learning_rate/action_model': 5.4316217255647974e-05, 'epoch': 1.43}) +[2026-06-17 22:10:53,093][starVLA.training.train_starvla][INFO] - Step 3467, Loss: {'train/loss': 0.353537380695343, 'train/grad_norm_pre_clip': 14.303118705749512, 'timing/data': 0.000845620408654213, 'timing/model': 1.014410313218832, 'learning_rate/qwen_vl_interface': 5.429460597728142e-06, 'learning_rate/action_model': 5.429460597728142e-05, 'epoch': 1.43}) +[2026-06-17 22:10:56,973][starVLA.training.train_starvla][INFO] - Step 3468, Loss: {'train/loss': 0.40552428364753723, 'train/grad_norm_pre_clip': 13.518340110778809, 'timing/data': 0.0013277679681777954, 'timing/model': 0.9314877074211836, 'learning_rate/qwen_vl_interface': 5.4272994326891114e-06, 'learning_rate/action_model': 5.427299432689111e-05, 'epoch': 1.43}) +[2026-06-17 22:11:00,743][starVLA.training.train_starvla][INFO] - Step 3469, Loss: {'train/loss': 0.3049169182777405, 'train/grad_norm_pre_clip': 12.99841022491455, 'timing/data': 0.0008159112185239792, 'timing/model': 0.9396887654438615, 'learning_rate/qwen_vl_interface': 5.425138230895719e-06, 'learning_rate/action_model': 5.425138230895719e-05, 'epoch': 1.43}) +[2026-06-17 22:11:04,956][starVLA.training.train_starvla][INFO] - Step 3470, Loss: {'train/loss': 0.603685200214386, 'train/grad_norm_pre_clip': 11.69923210144043, 'timing/data': 0.000665583647787571, 'timing/model': 1.0135801108554006, 'learning_rate/qwen_vl_interface': 5.422976992795984e-06, 'learning_rate/action_model': 5.422976992795984e-05, 'epoch': 1.43}) +[2026-06-17 22:11:08,936][starVLA.training.train_starvla][INFO] - Step 3471, Loss: {'train/loss': 0.4014971852302551, 'train/grad_norm_pre_clip': 15.167527198791504, 'timing/data': 0.0005182800814509392, 'timing/model': 1.1605621492490172, 'learning_rate/qwen_vl_interface': 5.420815718837929e-06, 'learning_rate/action_model': 5.42081571883793e-05, 'epoch': 1.43}) +[2026-06-17 22:11:13,538][starVLA.training.train_starvla][INFO] - Step 3472, Loss: {'train/loss': 0.2550758123397827, 'train/grad_norm_pre_clip': 14.9605073928833, 'timing/data': 0.0009962040930986404, 'timing/model': 0.9418109124526381, 'learning_rate/qwen_vl_interface': 5.418654409469595e-06, 'learning_rate/action_model': 5.4186544094695946e-05, 'epoch': 1.43}) +[2026-06-17 22:11:17,491][starVLA.training.train_starvla][INFO] - Step 3473, Loss: {'train/loss': 0.19443738460540771, 'train/grad_norm_pre_clip': 15.637030601501465, 'timing/data': 0.0006368979811668396, 'timing/model': 1.0102188140153885, 'learning_rate/qwen_vl_interface': 5.416493065139019e-06, 'learning_rate/action_model': 5.4164930651390186e-05, 'epoch': 1.43}) +[2026-06-17 22:11:21,463][starVLA.training.train_starvla][INFO] - Step 3474, Loss: {'train/loss': 0.5445356369018555, 'train/grad_norm_pre_clip': 16.305986404418945, 'timing/data': 0.001005435362458229, 'timing/model': 0.943212284706533, 'learning_rate/qwen_vl_interface': 5.4143316862942516e-06, 'learning_rate/action_model': 5.4143316862942514e-05, 'epoch': 1.43}) +[2026-06-17 22:11:26,097][starVLA.training.train_starvla][INFO] - Step 3475, Loss: {'train/loss': 0.36448606848716736, 'train/grad_norm_pre_clip': 15.552651405334473, 'timing/data': 0.0005137920379638672, 'timing/model': 1.058727441355586, 'learning_rate/qwen_vl_interface': 5.412170273383348e-06, 'learning_rate/action_model': 5.412170273383348e-05, 'epoch': 1.43}) +[2026-06-17 22:11:29,946][starVLA.training.train_starvla][INFO] - Step 3476, Loss: {'train/loss': 0.34314659237861633, 'train/grad_norm_pre_clip': 18.791051864624023, 'timing/data': 0.000814034603536129, 'timing/model': 0.9376716986298561, 'learning_rate/qwen_vl_interface': 5.410008826854374e-06, 'learning_rate/action_model': 5.410008826854373e-05, 'epoch': 1.43}) +[2026-06-17 22:11:34,383][starVLA.training.train_starvla][INFO] - Step 3477, Loss: {'train/loss': 0.2532084882259369, 'train/grad_norm_pre_clip': 11.943257331848145, 'timing/data': 0.00046644173562526703, 'timing/model': 1.420670073479414, 'learning_rate/qwen_vl_interface': 5.407847347155395e-06, 'learning_rate/action_model': 5.407847347155395e-05, 'epoch': 1.43}) +[2026-06-17 22:11:38,313][starVLA.training.train_starvla][INFO] - Step 3478, Loss: {'train/loss': 0.2556169629096985, 'train/grad_norm_pre_clip': 12.21837329864502, 'timing/data': 0.0005306750535964966, 'timing/model': 1.0764015950262547, 'learning_rate/qwen_vl_interface': 5.4056858347344934e-06, 'learning_rate/action_model': 5.405685834734493e-05, 'epoch': 1.43}) +[2026-06-17 22:11:41,996][starVLA.training.train_starvla][INFO] - Step 3479, Loss: {'train/loss': 0.4654453992843628, 'train/grad_norm_pre_clip': 18.655664443969727, 'timing/data': 0.0005104821175336838, 'timing/model': 0.9499909449368715, 'learning_rate/qwen_vl_interface': 5.403524290039751e-06, 'learning_rate/action_model': 5.403524290039751e-05, 'epoch': 1.43}) +[2026-06-17 22:11:45,849][starVLA.training.train_starvla][INFO] - Step 3480, Loss: {'train/loss': 0.3844543993473053, 'train/grad_norm_pre_clip': 19.697914123535156, 'timing/data': 0.0008264239877462387, 'timing/model': 0.9587904550135136, 'learning_rate/qwen_vl_interface': 5.401362713519257e-06, 'learning_rate/action_model': 5.4013627135192565e-05, 'epoch': 1.43}) +[2026-06-17 22:11:49,824][starVLA.training.train_starvla][INFO] - Step 3481, Loss: {'train/loss': 0.32876524329185486, 'train/grad_norm_pre_clip': 16.746440887451172, 'timing/data': 0.0004932479932904243, 'timing/model': 1.0039261672645807, 'learning_rate/qwen_vl_interface': 5.399201105621111e-06, 'learning_rate/action_model': 5.39920110562111e-05, 'epoch': 1.43}) +[2026-06-17 22:11:53,675][starVLA.training.train_starvla][INFO] - Step 3482, Loss: {'train/loss': 0.21095900237560272, 'train/grad_norm_pre_clip': 10.081623077392578, 'timing/data': 0.0005149999633431435, 'timing/model': 0.9465088341385126, 'learning_rate/qwen_vl_interface': 5.397039466793412e-06, 'learning_rate/action_model': 5.3970394667934123e-05, 'epoch': 1.43}) +[2026-06-17 22:11:57,524][starVLA.training.train_starvla][INFO] - Step 3483, Loss: {'train/loss': 0.2556259036064148, 'train/grad_norm_pre_clip': 15.19985580444336, 'timing/data': 0.0005137985572218895, 'timing/model': 0.9493134813383222, 'learning_rate/qwen_vl_interface': 5.394877797484278e-06, 'learning_rate/action_model': 5.3948777974842777e-05, 'epoch': 1.43}) +[2026-06-17 22:12:01,329][starVLA.training.train_starvla][INFO] - Step 3484, Loss: {'train/loss': 0.4471185803413391, 'train/grad_norm_pre_clip': 16.213764190673828, 'timing/data': 0.0004606088623404503, 'timing/model': 1.102650166489184, 'learning_rate/qwen_vl_interface': 5.392716098141817e-06, 'learning_rate/action_model': 5.3927160981418165e-05, 'epoch': 1.43}) +[2026-06-17 22:12:05,162][starVLA.training.train_starvla][INFO] - Step 3485, Loss: {'train/loss': 0.2520848512649536, 'train/grad_norm_pre_clip': 12.596961975097656, 'timing/data': 0.0005859481170773506, 'timing/model': 1.1125556901097298, 'learning_rate/qwen_vl_interface': 5.390554369214157e-06, 'learning_rate/action_model': 5.390554369214157e-05, 'epoch': 1.43}) +[2026-06-17 22:12:08,839][starVLA.training.train_starvla][INFO] - Step 3486, Loss: {'train/loss': 0.2597111463546753, 'train/grad_norm_pre_clip': 12.408920288085938, 'timing/data': 0.0009016050025820732, 'timing/model': 0.9448380582034588, 'learning_rate/qwen_vl_interface': 5.388392611149425e-06, 'learning_rate/action_model': 5.388392611149424e-05, 'epoch': 1.43}) +[2026-06-17 22:12:12,729][starVLA.training.train_starvla][INFO] - Step 3487, Loss: {'train/loss': 0.5513834357261658, 'train/grad_norm_pre_clip': 20.697229385375977, 'timing/data': 0.0005258172750473022, 'timing/model': 0.9494096832349896, 'learning_rate/qwen_vl_interface': 5.386230824395754e-06, 'learning_rate/action_model': 5.386230824395754e-05, 'epoch': 1.44}) +[2026-06-17 22:12:16,915][starVLA.training.train_starvla][INFO] - Step 3488, Loss: {'train/loss': 0.6931952834129333, 'train/grad_norm_pre_clip': 24.92363739013672, 'timing/data': 0.0008188700303435326, 'timing/model': 1.0367617616429925, 'learning_rate/qwen_vl_interface': 5.384069009401288e-06, 'learning_rate/action_model': 5.384069009401288e-05, 'epoch': 1.44}) +[2026-06-17 22:12:20,778][starVLA.training.train_starvla][INFO] - Step 3489, Loss: {'train/loss': 0.422479510307312, 'train/grad_norm_pre_clip': 15.545700073242188, 'timing/data': 0.0005243150517344475, 'timing/model': 0.9422581717371941, 'learning_rate/qwen_vl_interface': 5.381907166614171e-06, 'learning_rate/action_model': 5.3819071666141704e-05, 'epoch': 1.44}) +[2026-06-17 22:12:24,697][starVLA.training.train_starvla][INFO] - Step 3490, Loss: {'train/loss': 0.5404694080352783, 'train/grad_norm_pre_clip': 21.10446548461914, 'timing/data': 0.0005391864106059074, 'timing/model': 0.940683038905263, 'learning_rate/qwen_vl_interface': 5.379745296482559e-06, 'learning_rate/action_model': 5.379745296482559e-05, 'epoch': 1.44}) +[2026-06-17 22:12:28,601][starVLA.training.train_starvla][INFO] - Step 3491, Loss: {'train/loss': 0.45858773589134216, 'train/grad_norm_pre_clip': 15.924957275390625, 'timing/data': 0.0008713575080037117, 'timing/model': 1.1882642358541489, 'learning_rate/qwen_vl_interface': 5.377583399454605e-06, 'learning_rate/action_model': 5.3775833994546044e-05, 'epoch': 1.44}) +[2026-06-17 22:12:32,401][starVLA.training.train_starvla][INFO] - Step 3492, Loss: {'train/loss': 0.5606024861335754, 'train/grad_norm_pre_clip': 14.245330810546875, 'timing/data': 0.001434890553355217, 'timing/model': 1.081742369569838, 'learning_rate/qwen_vl_interface': 5.375421475978476e-06, 'learning_rate/action_model': 5.375421475978476e-05, 'epoch': 1.44}) +[2026-06-17 22:12:36,184][starVLA.training.train_starvla][INFO] - Step 3493, Loss: {'train/loss': 0.42932969331741333, 'train/grad_norm_pre_clip': 20.73885154724121, 'timing/data': 0.0006573451682925224, 'timing/model': 0.9390612710267305, 'learning_rate/qwen_vl_interface': 5.37325952650234e-06, 'learning_rate/action_model': 5.37325952650234e-05, 'epoch': 1.44}) +[2026-06-17 22:12:40,158][starVLA.training.train_starvla][INFO] - Step 3494, Loss: {'train/loss': 0.3677811026573181, 'train/grad_norm_pre_clip': 13.452937126159668, 'timing/data': 0.0008751722052693367, 'timing/model': 0.9636717364192009, 'learning_rate/qwen_vl_interface': 5.371097551474371e-06, 'learning_rate/action_model': 5.371097551474371e-05, 'epoch': 1.44}) +[2026-06-17 22:12:44,177][starVLA.training.train_starvla][INFO] - Step 3495, Loss: {'train/loss': 0.1524825245141983, 'train/grad_norm_pre_clip': 11.218643188476562, 'timing/data': 0.0004463391378521919, 'timing/model': 0.9497379809617996, 'learning_rate/qwen_vl_interface': 5.36893555134275e-06, 'learning_rate/action_model': 5.3689355513427494e-05, 'epoch': 1.44}) +[2026-06-17 22:12:49,526][starVLA.training.train_starvla][INFO] - Step 3496, Loss: {'train/loss': 0.45119521021842957, 'train/grad_norm_pre_clip': 21.21554946899414, 'timing/data': 0.000560099259018898, 'timing/model': 1.0351044991984963, 'learning_rate/qwen_vl_interface': 5.366773526555661e-06, 'learning_rate/action_model': 5.366773526555661e-05, 'epoch': 1.44}) +[2026-06-17 22:12:53,492][starVLA.training.train_starvla][INFO] - Step 3497, Loss: {'train/loss': 0.058129165321588516, 'train/grad_norm_pre_clip': 3.6377837657928467, 'timing/data': 0.0005443422123789787, 'timing/model': 1.215492564253509, 'learning_rate/qwen_vl_interface': 5.364611477561294e-06, 'learning_rate/action_model': 5.3646114775612945e-05, 'epoch': 1.44}) +[2026-06-17 22:12:57,363][starVLA.training.train_starvla][INFO] - Step 3498, Loss: {'train/loss': 0.4720621407032013, 'train/grad_norm_pre_clip': 15.085831642150879, 'timing/data': 0.0008238023146986961, 'timing/model': 1.1073800744488835, 'learning_rate/qwen_vl_interface': 5.362449404807845e-06, 'learning_rate/action_model': 5.362449404807844e-05, 'epoch': 1.44}) +[2026-06-17 22:13:01,033][starVLA.training.train_starvla][INFO] - Step 3499, Loss: {'train/loss': 0.39444413781166077, 'train/grad_norm_pre_clip': 17.51860809326172, 'timing/data': 0.0008483557030558586, 'timing/model': 0.940207228064537, 'learning_rate/qwen_vl_interface': 5.360287308743512e-06, 'learning_rate/action_model': 5.3602873087435126e-05, 'epoch': 1.44}) +[2026-06-17 22:15:29,034][starVLA.training.train_starvla][INFO] - Step 3500, Loss: {'train/loss': 0.4683232605457306, 'train/grad_norm_pre_clip': 17.592628479003906, 'eval/loss': 0.4804536700248718, 'eval/latency_6/loss': 0.4804536700248718, 'eval/demon_fire_precision': 0.9501734468621886, 'eval/demon_fire_recall': 0.9738202973497091, 'eval/demon_fire_f1': 0.9618515562649641, 'eval/demon_fire_rate_model': 0.8823038397328882, 'eval/demon_fire_rate_teacher': 0.860879243183083, 'eval/demon_left_precision': 0.8174603174603174, 'eval/demon_left_recall': 0.7228070175438597, 'eval/demon_left_f1': 0.7672253258845437, 'eval/demon_left_rate_model': 0.14023372287145242, 'eval/demon_left_rate_teacher': 0.15859766277128548, 'eval/demon_right_precision': 0.8624229979466119, 'eval/demon_right_recall': 0.6451612903225806, 'eval/demon_right_f1': 0.7381370826010545, 'eval/demon_right_rate_model': 0.13550361713967723, 'eval/demon_right_rate_teacher': 0.18113522537562604, 'eval/demon_move_f1': 0.752681204242799, 'eval/demon_cc_f1': 0.8878328451395163, 'eval/latency_6/demon_fire_precision': 0.9501734468621886, 'eval/latency_6/demon_fire_recall': 0.9738202973497091, 'eval/latency_6/demon_fire_f1': 0.9618515562649641, 'eval/latency_6/demon_fire_rate_model': 0.8823038397328882, 'eval/latency_6/demon_fire_rate_teacher': 0.860879243183083, 'eval/latency_6/demon_left_precision': 0.8174603174603174, 'eval/latency_6/demon_left_recall': 0.7228070175438597, 'eval/latency_6/demon_left_f1': 0.7672253258845437, 'eval/latency_6/demon_left_rate_model': 0.14023372287145242, 'eval/latency_6/demon_left_rate_teacher': 0.15859766277128548, 'eval/latency_6/demon_right_precision': 0.8624229979466119, 'eval/latency_6/demon_right_recall': 0.6451612903225806, 'eval/latency_6/demon_right_f1': 0.7381370826010545, 'eval/latency_6/demon_right_rate_model': 0.13550361713967723, 'eval/latency_6/demon_right_rate_teacher': 0.18113522537562604, 'eval/latency_6/demon_move_f1': 0.752681204242799, 'eval/latency_6/demon_cc_f1': 0.8878328451395163, 'timing/data': 0.0005211299285292625, 'timing/model': 0.9557163408026099, 'learning_rate/qwen_vl_interface': 5.358125189816502e-06, 'learning_rate/action_model': 5.358125189816502e-05, 'epoch': 1.44}) +[2026-06-17 22:15:29,037][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_3500_state +[2026-06-17 22:15:29,042][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 22:15:35,909][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_3500_state/model.safetensors +[2026-06-17 22:15:51,942][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_3500_state/optimizer.bin +[2026-06-17 22:15:51,945][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_3500_state/sampler.bin +[2026-06-17 22:15:51,945][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_3500_state/sampler_1.bin +[2026-06-17 22:15:51,949][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_3500_state/random_states_0.pkl +[2026-06-17 22:15:51,952][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3500 +[2026-06-17 22:15:51,952][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 22:15:51,963][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 22:15:55,695][starVLA.training.train_starvla][INFO] - Step 3501, Loss: {'train/loss': 0.4471718668937683, 'train/grad_norm_pre_clip': 11.933906555175781, 'timing/data': 0.0006722547113895416, 'timing/model': 1.014040700159967, 'learning_rate/qwen_vl_interface': 5.355963048475024e-06, 'learning_rate/action_model': 5.3559630484750244e-05, 'epoch': 1.44}) +[2026-06-17 22:15:59,474][starVLA.training.train_starvla][INFO] - Step 3502, Loss: {'train/loss': 0.39203929901123047, 'train/grad_norm_pre_clip': 12.264721870422363, 'timing/data': 0.001095508225262165, 'timing/model': 0.9317485708743334, 'learning_rate/qwen_vl_interface': 5.3538008851672916e-06, 'learning_rate/action_model': 5.3538008851672916e-05, 'epoch': 1.44}) +[2026-06-17 22:16:03,296][starVLA.training.train_starvla][INFO] - Step 3503, Loss: {'train/loss': 0.2650336027145386, 'train/grad_norm_pre_clip': 7.984409809112549, 'timing/data': 0.0010348847135901451, 'timing/model': 0.9808849943801761, 'learning_rate/qwen_vl_interface': 5.351638700341525e-06, 'learning_rate/action_model': 5.351638700341525e-05, 'epoch': 1.44}) +[2026-06-17 22:16:07,214][starVLA.training.train_starvla][INFO] - Step 3504, Loss: {'train/loss': 0.4236561954021454, 'train/grad_norm_pre_clip': 134.71536254882812, 'timing/data': 0.0006688740104436874, 'timing/model': 0.9914701730012894, 'learning_rate/qwen_vl_interface': 5.349476494445947e-06, 'learning_rate/action_model': 5.3494764944459466e-05, 'epoch': 1.44}) +[2026-06-17 22:16:11,419][starVLA.training.train_starvla][INFO] - Step 3505, Loss: {'train/loss': 0.3860830068588257, 'train/grad_norm_pre_clip': 13.418534278869629, 'timing/data': 0.0005996720865368843, 'timing/model': 0.9946768255904317, 'learning_rate/qwen_vl_interface': 5.3473142679287824e-06, 'learning_rate/action_model': 5.3473142679287826e-05, 'epoch': 1.44}) +[2026-06-17 22:16:16,246][starVLA.training.train_starvla][INFO] - Step 3506, Loss: {'train/loss': 0.25873464345932007, 'train/grad_norm_pre_clip': 12.61628246307373, 'timing/data': 0.0014002257958054543, 'timing/model': 1.1340087335556746, 'learning_rate/qwen_vl_interface': 5.345152021238268e-06, 'learning_rate/action_model': 5.345152021238268e-05, 'epoch': 1.44}) +[2026-06-17 22:16:19,960][starVLA.training.train_starvla][INFO] - Step 3507, Loss: {'train/loss': 0.5278018116950989, 'train/grad_norm_pre_clip': 21.17217445373535, 'timing/data': 0.00050389114767313, 'timing/model': 0.9287362536415458, 'learning_rate/qwen_vl_interface': 5.342989754822635e-06, 'learning_rate/action_model': 5.342989754822635e-05, 'epoch': 1.44}) +[2026-06-17 22:16:24,582][starVLA.training.train_starvla][INFO] - Step 3508, Loss: {'train/loss': 0.58012455701828, 'train/grad_norm_pre_clip': 16.070688247680664, 'timing/data': 0.0011254986748099327, 'timing/model': 0.9321801047772169, 'learning_rate/qwen_vl_interface': 5.340827469130126e-06, 'learning_rate/action_model': 5.3408274691301255e-05, 'epoch': 1.44}) +[2026-06-17 22:16:28,384][starVLA.training.train_starvla][INFO] - Step 3509, Loss: {'train/loss': 0.3340908885002136, 'train/grad_norm_pre_clip': 13.278937339782715, 'timing/data': 0.0005148639902472496, 'timing/model': 0.9349501505494118, 'learning_rate/qwen_vl_interface': 5.338665164608985e-06, 'learning_rate/action_model': 5.3386651646089845e-05, 'epoch': 1.44}) +[2026-06-17 22:16:32,161][starVLA.training.train_starvla][INFO] - Step 3510, Loss: {'train/loss': 0.1757754236459732, 'train/grad_norm_pre_clip': 9.790262222290039, 'timing/data': 0.0007260926067829132, 'timing/model': 0.9317863127216697, 'learning_rate/qwen_vl_interface': 5.336502841707461e-06, 'learning_rate/action_model': 5.33650284170746e-05, 'epoch': 1.44}) +[2026-06-17 22:16:36,062][starVLA.training.train_starvla][INFO] - Step 3511, Loss: {'train/loss': 0.22727273404598236, 'train/grad_norm_pre_clip': 18.1258544921875, 'timing/data': 0.0015351278707385063, 'timing/model': 0.9501980282366276, 'learning_rate/qwen_vl_interface': 5.334340500873803e-06, 'learning_rate/action_model': 5.3343405008738034e-05, 'epoch': 1.45}) +[2026-06-17 22:16:41,839][starVLA.training.train_starvla][INFO] - Step 3512, Loss: {'train/loss': 0.4578080475330353, 'train/grad_norm_pre_clip': 13.418521881103516, 'timing/data': 0.0004941020160913467, 'timing/model': 1.1230349820107222, 'learning_rate/qwen_vl_interface': 5.33217814255627e-06, 'learning_rate/action_model': 5.33217814255627e-05, 'epoch': 1.45}) +[2026-06-17 22:16:45,569][starVLA.training.train_starvla][INFO] - Step 3513, Loss: {'train/loss': 0.3662320375442505, 'train/grad_norm_pre_clip': 13.810261726379395, 'timing/data': 0.0007420005276799202, 'timing/model': 0.9438569359481335, 'learning_rate/qwen_vl_interface': 5.330015767203118e-06, 'learning_rate/action_model': 5.330015767203118e-05, 'epoch': 1.45}) +[2026-06-17 22:16:49,562][starVLA.training.train_starvla][INFO] - Step 3514, Loss: {'train/loss': 0.4864376485347748, 'train/grad_norm_pre_clip': 17.588848114013672, 'timing/data': 0.0008388860151171684, 'timing/model': 0.9358573136851192, 'learning_rate/qwen_vl_interface': 5.3278533752626135e-06, 'learning_rate/action_model': 5.3278533752626136e-05, 'epoch': 1.45}) +[2026-06-17 22:16:53,418][starVLA.training.train_starvla][INFO] - Step 3515, Loss: {'train/loss': 0.31008708477020264, 'train/grad_norm_pre_clip': 15.391165733337402, 'timing/data': 0.0005527501925826073, 'timing/model': 0.9412009306252003, 'learning_rate/qwen_vl_interface': 5.325690967183019e-06, 'learning_rate/action_model': 5.325690967183019e-05, 'epoch': 1.45}) +[2026-06-17 22:16:57,209][starVLA.training.train_starvla][INFO] - Step 3516, Loss: {'train/loss': 0.3008783459663391, 'train/grad_norm_pre_clip': 13.361870765686035, 'timing/data': 0.0008640186861157417, 'timing/model': 0.9382429923862219, 'learning_rate/qwen_vl_interface': 5.323528543412607e-06, 'learning_rate/action_model': 5.323528543412607e-05, 'epoch': 1.45}) +[2026-06-17 22:17:00,994][starVLA.training.train_starvla][INFO] - Step 3517, Loss: {'train/loss': 0.47511032223701477, 'train/grad_norm_pre_clip': 21.06266212463379, 'timing/data': 0.0004907939583063126, 'timing/model': 0.9350759563967586, 'learning_rate/qwen_vl_interface': 5.321366104399648e-06, 'learning_rate/action_model': 5.321366104399648e-05, 'epoch': 1.45}) +[2026-06-17 22:17:04,812][starVLA.training.train_starvla][INFO] - Step 3518, Loss: {'train/loss': 0.45452362298965454, 'train/grad_norm_pre_clip': 15.849047660827637, 'timing/data': 0.0006446121260523796, 'timing/model': 1.121696723625064, 'learning_rate/qwen_vl_interface': 5.319203650592419e-06, 'learning_rate/action_model': 5.3192036505924194e-05, 'epoch': 1.45}) +[2026-06-17 22:17:08,710][starVLA.training.train_starvla][INFO] - Step 3519, Loss: {'train/loss': 0.37522247433662415, 'train/grad_norm_pre_clip': 15.6693754196167, 'timing/data': 0.0008009858429431915, 'timing/model': 1.1085828561335802, 'learning_rate/qwen_vl_interface': 5.317041182439199e-06, 'learning_rate/action_model': 5.3170411824391985e-05, 'epoch': 1.45}) +[2026-06-17 22:17:12,373][starVLA.training.train_starvla][INFO] - Step 3520, Loss: {'train/loss': 0.38509249687194824, 'train/grad_norm_pre_clip': 16.384471893310547, 'timing/data': 0.0005318159237504005, 'timing/model': 0.9381015868857503, 'learning_rate/qwen_vl_interface': 5.314878700388267e-06, 'learning_rate/action_model': 5.314878700388267e-05, 'epoch': 1.45}) +[2026-06-17 22:17:17,082][starVLA.training.train_starvla][INFO] - Step 3521, Loss: {'train/loss': 0.36798492074012756, 'train/grad_norm_pre_clip': 29.143407821655273, 'timing/data': 0.000515456311404705, 'timing/model': 1.02267786860466, 'learning_rate/qwen_vl_interface': 5.312716204887914e-06, 'learning_rate/action_model': 5.3127162048879145e-05, 'epoch': 1.45}) +[2026-06-17 22:17:20,997][starVLA.training.train_starvla][INFO] - Step 3522, Loss: {'train/loss': 0.3216104209423065, 'train/grad_norm_pre_clip': 19.952152252197266, 'timing/data': 0.000493939034640789, 'timing/model': 0.9311729753389955, 'learning_rate/qwen_vl_interface': 5.3105536963864226e-06, 'learning_rate/action_model': 5.310553696386422e-05, 'epoch': 1.45}) +[2026-06-17 22:17:24,880][starVLA.training.train_starvla][INFO] - Step 3523, Loss: {'train/loss': 0.2874010503292084, 'train/grad_norm_pre_clip': 14.277774810791016, 'timing/data': 0.0005029849708080292, 'timing/model': 0.9410331370308995, 'learning_rate/qwen_vl_interface': 5.3083911753320875e-06, 'learning_rate/action_model': 5.308391175332087e-05, 'epoch': 1.45}) +[2026-06-17 22:17:28,926][starVLA.training.train_starvla][INFO] - Step 3524, Loss: {'train/loss': 0.43342462182044983, 'train/grad_norm_pre_clip': 17.486398696899414, 'timing/data': 0.0008263066411018372, 'timing/model': 0.9359544375911355, 'learning_rate/qwen_vl_interface': 5.306228642173194e-06, 'learning_rate/action_model': 5.306228642173194e-05, 'epoch': 1.45}) +[2026-06-17 22:17:32,752][starVLA.training.train_starvla][INFO] - Step 3525, Loss: {'train/loss': 0.4885872006416321, 'train/grad_norm_pre_clip': 17.539718627929688, 'timing/data': 0.0009824670851230621, 'timing/model': 1.1345992619171739, 'learning_rate/qwen_vl_interface': 5.304066097358046e-06, 'learning_rate/action_model': 5.304066097358046e-05, 'epoch': 1.45}) +[2026-06-17 22:17:36,430][starVLA.training.train_starvla][INFO] - Step 3526, Loss: {'train/loss': 0.35243600606918335, 'train/grad_norm_pre_clip': 16.329015731811523, 'timing/data': 0.0005191899836063385, 'timing/model': 0.9471689527854323, 'learning_rate/qwen_vl_interface': 5.301903541334937e-06, 'learning_rate/action_model': 5.301903541334937e-05, 'epoch': 1.45}) +[2026-06-17 22:17:40,539][starVLA.training.train_starvla][INFO] - Step 3527, Loss: {'train/loss': 0.3921854496002197, 'train/grad_norm_pre_clip': 16.009521484375, 'timing/data': 0.0006007878109812737, 'timing/model': 1.172187983058393, 'learning_rate/qwen_vl_interface': 5.299740974552168e-06, 'learning_rate/action_model': 5.299740974552167e-05, 'epoch': 1.45}) +[2026-06-17 22:17:44,463][starVLA.training.train_starvla][INFO] - Step 3528, Loss: {'train/loss': 0.21880805492401123, 'train/grad_norm_pre_clip': 11.858284950256348, 'timing/data': 0.000508040189743042, 'timing/model': 0.9792613834142685, 'learning_rate/qwen_vl_interface': 5.297578397458042e-06, 'learning_rate/action_model': 5.297578397458042e-05, 'epoch': 1.45}) +[2026-06-17 22:17:48,440][starVLA.training.train_starvla][INFO] - Step 3529, Loss: {'train/loss': 0.42679163813591003, 'train/grad_norm_pre_clip': 15.112024307250977, 'timing/data': 0.0006224941462278366, 'timing/model': 0.9449747642502189, 'learning_rate/qwen_vl_interface': 5.295415810500862e-06, 'learning_rate/action_model': 5.295415810500862e-05, 'epoch': 1.45}) +[2026-06-17 22:17:52,808][starVLA.training.train_starvla][INFO] - Step 3530, Loss: {'train/loss': 0.613861083984375, 'train/grad_norm_pre_clip': 20.570255279541016, 'timing/data': 0.0015818430110812187, 'timing/model': 1.3023517150431871, 'learning_rate/qwen_vl_interface': 5.293253214128938e-06, 'learning_rate/action_model': 5.293253214128937e-05, 'epoch': 1.45}) +[2026-06-17 22:17:56,693][starVLA.training.train_starvla][INFO] - Step 3531, Loss: {'train/loss': 0.5835064649581909, 'train/grad_norm_pre_clip': 19.228778839111328, 'timing/data': 0.0007560811936855316, 'timing/model': 1.200274609029293, 'learning_rate/qwen_vl_interface': 5.291090608790572e-06, 'learning_rate/action_model': 5.291090608790572e-05, 'epoch': 1.45}) +[2026-06-17 22:18:00,525][starVLA.training.train_starvla][INFO] - Step 3532, Loss: {'train/loss': 0.4593656659126282, 'train/grad_norm_pre_clip': 17.0355167388916, 'timing/data': 0.0005370844155550003, 'timing/model': 1.106409426778555, 'learning_rate/qwen_vl_interface': 5.28892799493408e-06, 'learning_rate/action_model': 5.2889279949340796e-05, 'epoch': 1.45}) +[2026-06-17 22:18:04,293][starVLA.training.train_starvla][INFO] - Step 3533, Loss: {'train/loss': 0.41652557253837585, 'train/grad_norm_pre_clip': 18.034635543823242, 'timing/data': 0.0008512949571013451, 'timing/model': 1.0262099551036954, 'learning_rate/qwen_vl_interface': 5.286765373007774e-06, 'learning_rate/action_model': 5.2867653730077735e-05, 'epoch': 1.45}) +[2026-06-17 22:18:10,275][starVLA.training.train_starvla][INFO] - Step 3534, Loss: {'train/loss': 0.23187899589538574, 'train/grad_norm_pre_clip': 12.804116249084473, 'timing/data': 0.0009335549548268318, 'timing/model': 1.0025265393778682, 'learning_rate/qwen_vl_interface': 5.284602743459968e-06, 'learning_rate/action_model': 5.284602743459968e-05, 'epoch': 1.45}) +[2026-06-17 22:18:14,432][starVLA.training.train_starvla][INFO] - Step 3535, Loss: {'train/loss': 0.23214976489543915, 'train/grad_norm_pre_clip': 11.210755348205566, 'timing/data': 0.0007009329274296761, 'timing/model': 1.2445241510868073, 'learning_rate/qwen_vl_interface': 5.282440106738973e-06, 'learning_rate/action_model': 5.282440106738973e-05, 'epoch': 1.46}) +[2026-06-17 22:18:18,260][starVLA.training.train_starvla][INFO] - Step 3536, Loss: {'train/loss': 0.355363130569458, 'train/grad_norm_pre_clip': 15.777685165405273, 'timing/data': 0.0005243746563792229, 'timing/model': 0.9499573027715087, 'learning_rate/qwen_vl_interface': 5.280277463293111e-06, 'learning_rate/action_model': 5.280277463293111e-05, 'epoch': 1.46}) +[2026-06-17 22:18:22,121][starVLA.training.train_starvla][INFO] - Step 3537, Loss: {'train/loss': 0.3014909625053406, 'train/grad_norm_pre_clip': 15.01192569732666, 'timing/data': 0.0005274470895528793, 'timing/model': 0.941272871568799, 'learning_rate/qwen_vl_interface': 5.2781148135706975e-06, 'learning_rate/action_model': 5.278114813570697e-05, 'epoch': 1.46}) +[2026-06-17 22:18:25,971][starVLA.training.train_starvla][INFO] - Step 3538, Loss: {'train/loss': 0.5665447115898132, 'train/grad_norm_pre_clip': 18.222259521484375, 'timing/data': 0.0008423132821917534, 'timing/model': 1.154073833487928, 'learning_rate/qwen_vl_interface': 5.275952158020054e-06, 'learning_rate/action_model': 5.275952158020053e-05, 'epoch': 1.46}) +[2026-06-17 22:18:29,770][starVLA.training.train_starvla][INFO] - Step 3539, Loss: {'train/loss': 0.2807305157184601, 'train/grad_norm_pre_clip': 12.612674713134766, 'timing/data': 0.00132856797426939, 'timing/model': 1.0910336012020707, 'learning_rate/qwen_vl_interface': 5.273789497089497e-06, 'learning_rate/action_model': 5.2737894970894974e-05, 'epoch': 1.46}) +[2026-06-17 22:18:33,406][starVLA.training.train_starvla][INFO] - Step 3540, Loss: {'train/loss': 0.382038414478302, 'train/grad_norm_pre_clip': 14.682413101196289, 'timing/data': 0.0005397815257310867, 'timing/model': 0.9385033436119556, 'learning_rate/qwen_vl_interface': 5.271626831227355e-06, 'learning_rate/action_model': 5.2716268312273544e-05, 'epoch': 1.46}) +[2026-06-17 22:18:37,261][starVLA.training.train_starvla][INFO] - Step 3541, Loss: {'train/loss': 0.4306107461452484, 'train/grad_norm_pre_clip': 15.779243469238281, 'timing/data': 0.0005256952717900276, 'timing/model': 0.9362689927220345, 'learning_rate/qwen_vl_interface': 5.2694641608819465e-06, 'learning_rate/action_model': 5.269464160881946e-05, 'epoch': 1.46}) +[2026-06-17 22:18:41,001][starVLA.training.train_starvla][INFO] - Step 3542, Loss: {'train/loss': 0.48674848675727844, 'train/grad_norm_pre_clip': 16.42263412475586, 'timing/data': 0.0005348026752471924, 'timing/model': 0.9234780678525567, 'learning_rate/qwen_vl_interface': 5.267301486501595e-06, 'learning_rate/action_model': 5.2673014865015945e-05, 'epoch': 1.46}) +[2026-06-17 22:18:44,699][starVLA.training.train_starvla][INFO] - Step 3543, Loss: {'train/loss': 0.5422866940498352, 'train/grad_norm_pre_clip': 19.907419204711914, 'timing/data': 0.0005850018933415413, 'timing/model': 0.934788272716105, 'learning_rate/qwen_vl_interface': 5.265138808534628e-06, 'learning_rate/action_model': 5.265138808534628e-05, 'epoch': 1.46}) +[2026-06-17 22:18:48,455][starVLA.training.train_starvla][INFO] - Step 3544, Loss: {'train/loss': 0.33746588230133057, 'train/grad_norm_pre_clip': 16.73503303527832, 'timing/data': 0.0005153492093086243, 'timing/model': 0.9280626764521003, 'learning_rate/qwen_vl_interface': 5.26297612742937e-06, 'learning_rate/action_model': 5.26297612742937e-05, 'epoch': 1.46}) +[2026-06-17 22:18:52,378][starVLA.training.train_starvla][INFO] - Step 3545, Loss: {'train/loss': 0.4790106415748596, 'train/grad_norm_pre_clip': 21.377731323242188, 'timing/data': 0.0005197934806346893, 'timing/model': 1.170324512757361, 'learning_rate/qwen_vl_interface': 5.2608134436341476e-06, 'learning_rate/action_model': 5.260813443634147e-05, 'epoch': 1.46}) +[2026-06-17 22:18:56,017][starVLA.training.train_starvla][INFO] - Step 3546, Loss: {'train/loss': 0.3203848600387573, 'train/grad_norm_pre_clip': 15.493151664733887, 'timing/data': 0.0005448907613754272, 'timing/model': 0.9435440460219979, 'learning_rate/qwen_vl_interface': 5.258650757597285e-06, 'learning_rate/action_model': 5.258650757597286e-05, 'epoch': 1.46}) +[2026-06-17 22:18:59,834][starVLA.training.train_starvla][INFO] - Step 3547, Loss: {'train/loss': 0.14797808229923248, 'train/grad_norm_pre_clip': 10.498590469360352, 'timing/data': 0.0013711582869291306, 'timing/model': 0.9281705934554338, 'learning_rate/qwen_vl_interface': 5.256488069767113e-06, 'learning_rate/action_model': 5.256488069767113e-05, 'epoch': 1.46}) +[2026-06-17 22:19:03,562][starVLA.training.train_starvla][INFO] - Step 3548, Loss: {'train/loss': 0.6400555372238159, 'train/grad_norm_pre_clip': 21.648868560791016, 'timing/data': 0.0005467906594276428, 'timing/model': 0.924217770807445, 'learning_rate/qwen_vl_interface': 5.254325380591955e-06, 'learning_rate/action_model': 5.254325380591955e-05, 'epoch': 1.46}) +[2026-06-17 22:19:07,356][starVLA.training.train_starvla][INFO] - Step 3549, Loss: {'train/loss': 0.2973272502422333, 'train/grad_norm_pre_clip': 14.45621395111084, 'timing/data': 0.000752471387386322, 'timing/model': 0.9393692668527365, 'learning_rate/qwen_vl_interface': 5.252162690520142e-06, 'learning_rate/action_model': 5.252162690520142e-05, 'epoch': 1.46}) +[2026-06-17 22:19:11,120][starVLA.training.train_starvla][INFO] - Step 3550, Loss: {'train/loss': 0.45584386587142944, 'train/grad_norm_pre_clip': 16.824480056762695, 'timing/data': 0.0005727307870984077, 'timing/model': 0.9270788626745343, 'learning_rate/qwen_vl_interface': 5.2500000000000006e-06, 'learning_rate/action_model': 5.25e-05, 'epoch': 1.46}) +[2026-06-17 22:19:14,865][starVLA.training.train_starvla][INFO] - Step 3551, Loss: {'train/loss': 0.5076819062232971, 'train/grad_norm_pre_clip': 19.69386863708496, 'timing/data': 0.0005499189719557762, 'timing/model': 1.0867895390838385, 'learning_rate/qwen_vl_interface': 5.24783730947986e-06, 'learning_rate/action_model': 5.2478373094798594e-05, 'epoch': 1.46}) +[2026-06-17 22:19:18,579][starVLA.training.train_starvla][INFO] - Step 3552, Loss: {'train/loss': 0.23633205890655518, 'train/grad_norm_pre_clip': 20.188007354736328, 'timing/data': 0.0005362285301089287, 'timing/model': 1.0614089909940958, 'learning_rate/qwen_vl_interface': 5.245674619408045e-06, 'learning_rate/action_model': 5.245674619408045e-05, 'epoch': 1.46}) +[2026-06-17 22:19:22,164][starVLA.training.train_starvla][INFO] - Step 3553, Loss: {'train/loss': 0.5497193932533264, 'train/grad_norm_pre_clip': 18.712114334106445, 'timing/data': 0.0005182204768061638, 'timing/model': 0.9333658628165722, 'learning_rate/qwen_vl_interface': 5.243511930232889e-06, 'learning_rate/action_model': 5.243511930232888e-05, 'epoch': 1.46}) +[2026-06-17 22:19:26,092][starVLA.training.train_starvla][INFO] - Step 3554, Loss: {'train/loss': 0.42018213868141174, 'train/grad_norm_pre_clip': 19.374361038208008, 'timing/data': 0.0008785789832472801, 'timing/model': 1.0580679019913077, 'learning_rate/qwen_vl_interface': 5.2413492424027165e-06, 'learning_rate/action_model': 5.241349242402717e-05, 'epoch': 1.46}) +[2026-06-17 22:19:29,886][starVLA.training.train_starvla][INFO] - Step 3555, Loss: {'train/loss': 0.2685740292072296, 'train/grad_norm_pre_clip': 14.032652854919434, 'timing/data': 0.0005138535052537918, 'timing/model': 0.9272181587293744, 'learning_rate/qwen_vl_interface': 5.239186556365855e-06, 'learning_rate/action_model': 5.239186556365855e-05, 'epoch': 1.46}) +[2026-06-17 22:19:33,694][starVLA.training.train_starvla][INFO] - Step 3556, Loss: {'train/loss': 0.3955961763858795, 'train/grad_norm_pre_clip': 16.088848114013672, 'timing/data': 0.0005100024864077568, 'timing/model': 0.9298564931377769, 'learning_rate/qwen_vl_interface': 5.23702387257063e-06, 'learning_rate/action_model': 5.23702387257063e-05, 'epoch': 1.46}) +[2026-06-17 22:19:37,498][starVLA.training.train_starvla][INFO] - Step 3557, Loss: {'train/loss': 0.3742614686489105, 'train/grad_norm_pre_clip': 18.315227508544922, 'timing/data': 0.0005161762237548828, 'timing/model': 0.9354757759720087, 'learning_rate/qwen_vl_interface': 5.234861191465373e-06, 'learning_rate/action_model': 5.234861191465372e-05, 'epoch': 1.46}) +[2026-06-17 22:19:41,279][starVLA.training.train_starvla][INFO] - Step 3558, Loss: {'train/loss': 0.3240604102611542, 'train/grad_norm_pre_clip': 14.60206413269043, 'timing/data': 0.0004936037585139275, 'timing/model': 1.111124373972416, 'learning_rate/qwen_vl_interface': 5.232698513498407e-06, 'learning_rate/action_model': 5.2326985134984066e-05, 'epoch': 1.46}) +[2026-06-17 22:19:45,151][starVLA.training.train_starvla][INFO] - Step 3559, Loss: {'train/loss': 0.27540135383605957, 'train/grad_norm_pre_clip': 15.329913139343262, 'timing/data': 0.0005080839619040489, 'timing/model': 1.1766483336687088, 'learning_rate/qwen_vl_interface': 5.230535839118057e-06, 'learning_rate/action_model': 5.230535839118057e-05, 'epoch': 1.47}) +[2026-06-17 22:19:49,270][starVLA.training.train_starvla][INFO] - Step 3560, Loss: {'train/loss': 0.26989948749542236, 'train/grad_norm_pre_clip': 13.167135238647461, 'timing/data': 0.00048837810754776, 'timing/model': 0.9317721026018262, 'learning_rate/qwen_vl_interface': 5.2283731687726455e-06, 'learning_rate/action_model': 5.2283731687726454e-05, 'epoch': 1.47}) +[2026-06-17 22:19:53,131][starVLA.training.train_starvla][INFO] - Step 3561, Loss: {'train/loss': 0.4636743366718292, 'train/grad_norm_pre_clip': 20.14815330505371, 'timing/data': 0.0006819665431976318, 'timing/model': 0.9679764546453953, 'learning_rate/qwen_vl_interface': 5.226210502910504e-06, 'learning_rate/action_model': 5.226210502910504e-05, 'epoch': 1.47}) +[2026-06-17 22:19:56,945][starVLA.training.train_starvla][INFO] - Step 3562, Loss: {'train/loss': 0.5371519327163696, 'train/grad_norm_pre_clip': 16.077482223510742, 'timing/data': 0.0007850602269172668, 'timing/model': 0.9345834730193019, 'learning_rate/qwen_vl_interface': 5.224047841979949e-06, 'learning_rate/action_model': 5.2240478419799485e-05, 'epoch': 1.47}) +[2026-06-17 22:20:00,804][starVLA.training.train_starvla][INFO] - Step 3563, Loss: {'train/loss': 0.752975583076477, 'train/grad_norm_pre_clip': 21.904277801513672, 'timing/data': 0.000522182323038578, 'timing/model': 0.948085717856884, 'learning_rate/qwen_vl_interface': 5.2218851864293044e-06, 'learning_rate/action_model': 5.221885186429304e-05, 'epoch': 1.47}) +[2026-06-17 22:20:04,643][starVLA.training.train_starvla][INFO] - Step 3564, Loss: {'train/loss': 0.1945471167564392, 'train/grad_norm_pre_clip': 10.553526878356934, 'timing/data': 0.000894913449883461, 'timing/model': 0.9460012018680573, 'learning_rate/qwen_vl_interface': 5.21972253670689e-06, 'learning_rate/action_model': 5.21972253670689e-05, 'epoch': 1.47}) +[2026-06-17 22:20:08,557][starVLA.training.train_starvla][INFO] - Step 3565, Loss: {'train/loss': 0.5760637521743774, 'train/grad_norm_pre_clip': 20.601041793823242, 'timing/data': 0.0007491307333111763, 'timing/model': 1.189068192616105, 'learning_rate/qwen_vl_interface': 5.217559893261028e-06, 'learning_rate/action_model': 5.217559893261028e-05, 'epoch': 1.47}) +[2026-06-17 22:20:12,412][starVLA.training.train_starvla][INFO] - Step 3566, Loss: {'train/loss': 0.4228720963001251, 'train/grad_norm_pre_clip': 111.30741882324219, 'timing/data': 0.0010769953951239586, 'timing/model': 1.065620786510408, 'learning_rate/qwen_vl_interface': 5.2153972565400355e-06, 'learning_rate/action_model': 5.215397256540035e-05, 'epoch': 1.47}) +[2026-06-17 22:20:16,364][starVLA.training.train_starvla][INFO] - Step 3567, Loss: {'train/loss': 0.15162083506584167, 'train/grad_norm_pre_clip': 17.393543243408203, 'timing/data': 0.0005032382905483246, 'timing/model': 0.9889937927946448, 'learning_rate/qwen_vl_interface': 5.213234626992227e-06, 'learning_rate/action_model': 5.213234626992226e-05, 'epoch': 1.47}) +[2026-06-17 22:20:20,178][starVLA.training.train_starvla][INFO] - Step 3568, Loss: {'train/loss': 0.3178294003009796, 'train/grad_norm_pre_clip': 16.929367065429688, 'timing/data': 0.0008732294663786888, 'timing/model': 0.9428645865991712, 'learning_rate/qwen_vl_interface': 5.21107200506592e-06, 'learning_rate/action_model': 5.21107200506592e-05, 'epoch': 1.47}) +[2026-06-17 22:20:24,113][starVLA.training.train_starvla][INFO] - Step 3569, Loss: {'train/loss': 0.19786436855793, 'train/grad_norm_pre_clip': 10.107549667358398, 'timing/data': 0.0005665114149451256, 'timing/model': 0.9635233543813229, 'learning_rate/qwen_vl_interface': 5.208909391209429e-06, 'learning_rate/action_model': 5.208909391209429e-05, 'epoch': 1.47}) +[2026-06-17 22:20:27,938][starVLA.training.train_starvla][INFO] - Step 3570, Loss: {'train/loss': 0.5188209414482117, 'train/grad_norm_pre_clip': 20.143266677856445, 'timing/data': 0.0005035605281591415, 'timing/model': 0.9576144609600306, 'learning_rate/qwen_vl_interface': 5.206746785871066e-06, 'learning_rate/action_model': 5.2067467858710664e-05, 'epoch': 1.47}) +[2026-06-17 22:20:31,778][starVLA.training.train_starvla][INFO] - Step 3571, Loss: {'train/loss': 0.2625901997089386, 'train/grad_norm_pre_clip': 13.746374130249023, 'timing/data': 0.0004940927028656006, 'timing/model': 1.1253083255141973, 'learning_rate/qwen_vl_interface': 5.204584189499139e-06, 'learning_rate/action_model': 5.2045841894991385e-05, 'epoch': 1.47}) +[2026-06-17 22:20:35,659][starVLA.training.train_starvla][INFO] - Step 3572, Loss: {'train/loss': 0.17399466037750244, 'train/grad_norm_pre_clip': 55.282535552978516, 'timing/data': 0.0006072530522942543, 'timing/model': 1.1199758127331734, 'learning_rate/qwen_vl_interface': 5.202421602541959e-06, 'learning_rate/action_model': 5.2024216025419594e-05, 'epoch': 1.47}) +[2026-06-17 22:20:39,412][starVLA.training.train_starvla][INFO] - Step 3573, Loss: {'train/loss': 0.6914615631103516, 'train/grad_norm_pre_clip': 22.900218963623047, 'timing/data': 0.0005480591207742691, 'timing/model': 0.9432467138394713, 'learning_rate/qwen_vl_interface': 5.200259025447835e-06, 'learning_rate/action_model': 5.2002590254478345e-05, 'epoch': 1.47}) +[2026-06-17 22:20:43,289][starVLA.training.train_starvla][INFO] - Step 3574, Loss: {'train/loss': 0.7701318264007568, 'train/grad_norm_pre_clip': 24.3936710357666, 'timing/data': 0.0005835220217704773, 'timing/model': 0.9413319826126099, 'learning_rate/qwen_vl_interface': 5.198096458665064e-06, 'learning_rate/action_model': 5.198096458665064e-05, 'epoch': 1.47}) +[2026-06-17 22:20:47,172][starVLA.training.train_starvla][INFO] - Step 3575, Loss: {'train/loss': 0.49189209938049316, 'train/grad_norm_pre_clip': 22.342313766479492, 'timing/data': 0.0011077513918280602, 'timing/model': 0.9460452571511269, 'learning_rate/qwen_vl_interface': 5.195933902641955e-06, 'learning_rate/action_model': 5.195933902641955e-05, 'epoch': 1.47}) +[2026-06-17 22:20:51,016][starVLA.training.train_starvla][INFO] - Step 3576, Loss: {'train/loss': 0.25765755772590637, 'train/grad_norm_pre_clip': 14.747210502624512, 'timing/data': 0.0004933392629027367, 'timing/model': 0.9453552523627877, 'learning_rate/qwen_vl_interface': 5.193771357826807e-06, 'learning_rate/action_model': 5.193771357826806e-05, 'epoch': 1.47}) +[2026-06-17 22:20:55,031][starVLA.training.train_starvla][INFO] - Step 3577, Loss: {'train/loss': 0.4554794132709503, 'train/grad_norm_pre_clip': 18.361021041870117, 'timing/data': 0.0007326612249016762, 'timing/model': 1.052856800146401, 'learning_rate/qwen_vl_interface': 5.191608824667916e-06, 'learning_rate/action_model': 5.1916088246679165e-05, 'epoch': 1.47}) +[2026-06-17 22:20:58,904][starVLA.training.train_starvla][INFO] - Step 3578, Loss: {'train/loss': 0.35253122448921204, 'train/grad_norm_pre_clip': 14.841651916503906, 'timing/data': 0.0004767356440424919, 'timing/model': 1.1212641708552837, 'learning_rate/qwen_vl_interface': 5.189446303613578e-06, 'learning_rate/action_model': 5.1894463036135775e-05, 'epoch': 1.47}) +[2026-06-17 22:21:02,545][starVLA.training.train_starvla][INFO] - Step 3579, Loss: {'train/loss': 0.30846840143203735, 'train/grad_norm_pre_clip': 14.531551361083984, 'timing/data': 0.0010691368952393532, 'timing/model': 0.9556160848587751, 'learning_rate/qwen_vl_interface': 5.187283795112087e-06, 'learning_rate/action_model': 5.1872837951120866e-05, 'epoch': 1.47}) +[2026-06-17 22:21:06,475][starVLA.training.train_starvla][INFO] - Step 3580, Loss: {'train/loss': 0.41508153080940247, 'train/grad_norm_pre_clip': 16.69135856628418, 'timing/data': 0.0005153464153409004, 'timing/model': 0.9373693699017167, 'learning_rate/qwen_vl_interface': 5.185121299611734e-06, 'learning_rate/action_model': 5.1851212996117335e-05, 'epoch': 1.47}) +[2026-06-17 22:21:10,289][starVLA.training.train_starvla][INFO] - Step 3581, Loss: {'train/loss': 0.69371497631073, 'train/grad_norm_pre_clip': 21.474645614624023, 'timing/data': 0.0007628137245774269, 'timing/model': 0.9480841774493456, 'learning_rate/qwen_vl_interface': 5.182958817560803e-06, 'learning_rate/action_model': 5.1829588175608026e-05, 'epoch': 1.47}) +[2026-06-17 22:21:14,128][starVLA.training.train_starvla][INFO] - Step 3582, Loss: {'train/loss': 0.2458268105983734, 'train/grad_norm_pre_clip': 14.77949333190918, 'timing/data': 0.0005020620301365852, 'timing/model': 0.9523608917370439, 'learning_rate/qwen_vl_interface': 5.180796349407583e-06, 'learning_rate/action_model': 5.180796349407583e-05, 'epoch': 1.47}) +[2026-06-17 22:21:17,886][starVLA.training.train_starvla][INFO] - Step 3583, Loss: {'train/loss': 0.23659655451774597, 'train/grad_norm_pre_clip': 15.655710220336914, 'timing/data': 0.000496220774948597, 'timing/model': 0.926828564144671, 'learning_rate/qwen_vl_interface': 5.178633895600354e-06, 'learning_rate/action_model': 5.178633895600353e-05, 'epoch': 1.47}) +[2026-06-17 22:21:21,651][starVLA.training.train_starvla][INFO] - Step 3584, Loss: {'train/loss': 0.35261857509613037, 'train/grad_norm_pre_clip': 16.342205047607422, 'timing/data': 0.0004979055374860764, 'timing/model': 1.1031757136806846, 'learning_rate/qwen_vl_interface': 5.176471456587394e-06, 'learning_rate/action_model': 5.176471456587394e-05, 'epoch': 1.48}) +[2026-06-17 22:21:25,298][starVLA.training.train_starvla][INFO] - Step 3585, Loss: {'train/loss': 0.3746880888938904, 'train/grad_norm_pre_clip': 21.832679748535156, 'timing/data': 0.0005720006301999092, 'timing/model': 0.9386330731213093, 'learning_rate/qwen_vl_interface': 5.1743090328169805e-06, 'learning_rate/action_model': 5.1743090328169805e-05, 'epoch': 1.48}) +[2026-06-17 22:21:29,093][starVLA.training.train_starvla][INFO] - Step 3586, Loss: {'train/loss': 0.33576005697250366, 'train/grad_norm_pre_clip': 18.518402099609375, 'timing/data': 0.0005403049290180206, 'timing/model': 0.9234832571819425, 'learning_rate/qwen_vl_interface': 5.172146624737388e-06, 'learning_rate/action_model': 5.1721466247373875e-05, 'epoch': 1.48}) +[2026-06-17 22:21:32,885][starVLA.training.train_starvla][INFO] - Step 3587, Loss: {'train/loss': 0.6009187698364258, 'train/grad_norm_pre_clip': 22.422767639160156, 'timing/data': 0.0005120588466525078, 'timing/model': 0.9337312048301101, 'learning_rate/qwen_vl_interface': 5.169984232796883e-06, 'learning_rate/action_model': 5.169984232796883e-05, 'epoch': 1.48}) +[2026-06-17 22:21:36,854][starVLA.training.train_starvla][INFO] - Step 3588, Loss: {'train/loss': 0.22523416578769684, 'train/grad_norm_pre_clip': 10.590880393981934, 'timing/data': 0.0005160635337233543, 'timing/model': 0.9559268495067954, 'learning_rate/qwen_vl_interface': 5.1678218574437325e-06, 'learning_rate/action_model': 5.167821857443732e-05, 'epoch': 1.48}) +[2026-06-17 22:21:40,759][starVLA.training.train_starvla][INFO] - Step 3589, Loss: {'train/loss': 0.345732718706131, 'train/grad_norm_pre_clip': 19.586111068725586, 'timing/data': 0.0005208626389503479, 'timing/model': 0.9354686625301838, 'learning_rate/qwen_vl_interface': 5.165659499126198e-06, 'learning_rate/action_model': 5.165659499126198e-05, 'epoch': 1.48}) +[2026-06-17 22:21:44,621][starVLA.training.train_starvla][INFO] - Step 3590, Loss: {'train/loss': 0.4365431070327759, 'train/grad_norm_pre_clip': 19.24265480041504, 'timing/data': 0.0006485749036073685, 'timing/model': 0.9443104043602943, 'learning_rate/qwen_vl_interface': 5.1634971582925405e-06, 'learning_rate/action_model': 5.163497158292541e-05, 'epoch': 1.48}) +[2026-06-17 22:21:49,054][starVLA.training.train_starvla][INFO] - Step 3591, Loss: {'train/loss': 0.27558115124702454, 'train/grad_norm_pre_clip': 13.971090316772461, 'timing/data': 0.0004911636933684349, 'timing/model': 1.1478175949305296, 'learning_rate/qwen_vl_interface': 5.161334835391016e-06, 'learning_rate/action_model': 5.161334835391016e-05, 'epoch': 1.48}) +[2026-06-17 22:21:52,956][starVLA.training.train_starvla][INFO] - Step 3592, Loss: {'train/loss': 0.5630573630332947, 'train/grad_norm_pre_clip': 22.369733810424805, 'timing/data': 0.0007079262286424637, 'timing/model': 0.937669332139194, 'learning_rate/qwen_vl_interface': 5.159172530869875e-06, 'learning_rate/action_model': 5.159172530869874e-05, 'epoch': 1.48}) +[2026-06-17 22:21:56,986][starVLA.training.train_starvla][INFO] - Step 3593, Loss: {'train/loss': 0.34152185916900635, 'train/grad_norm_pre_clip': 33.72325897216797, 'timing/data': 0.0007183896377682686, 'timing/model': 1.0586662478744984, 'learning_rate/qwen_vl_interface': 5.157010245177366e-06, 'learning_rate/action_model': 5.1570102451773664e-05, 'epoch': 1.48}) +[2026-06-17 22:22:00,799][starVLA.training.train_starvla][INFO] - Step 3594, Loss: {'train/loss': 0.23152561485767365, 'train/grad_norm_pre_clip': 12.975801467895508, 'timing/data': 0.0010888716205954552, 'timing/model': 0.927033475600183, 'learning_rate/qwen_vl_interface': 5.1548479787617346e-06, 'learning_rate/action_model': 5.1548479787617346e-05, 'epoch': 1.48}) +[2026-06-17 22:22:05,076][starVLA.training.train_starvla][INFO] - Step 3595, Loss: {'train/loss': 0.5566421747207642, 'train/grad_norm_pre_clip': 25.925758361816406, 'timing/data': 0.0006653489544987679, 'timing/model': 0.9673528512939811, 'learning_rate/qwen_vl_interface': 5.152685732071219e-06, 'learning_rate/action_model': 5.1526857320712185e-05, 'epoch': 1.48}) +[2026-06-17 22:22:08,888][starVLA.training.train_starvla][INFO] - Step 3596, Loss: {'train/loss': 0.6738355755805969, 'train/grad_norm_pre_clip': 21.438549041748047, 'timing/data': 0.0011626286432147026, 'timing/model': 0.9321951437741518, 'learning_rate/qwen_vl_interface': 5.1505235055540545e-06, 'learning_rate/action_model': 5.1505235055540545e-05, 'epoch': 1.48}) +[2026-06-17 22:22:12,714][starVLA.training.train_starvla][INFO] - Step 3597, Loss: {'train/loss': 0.3317512273788452, 'train/grad_norm_pre_clip': 14.327864646911621, 'timing/data': 0.00048589427024126053, 'timing/model': 1.135380337946117, 'learning_rate/qwen_vl_interface': 5.148361299658477e-06, 'learning_rate/action_model': 5.148361299658476e-05, 'epoch': 1.48}) +[2026-06-17 22:22:16,464][starVLA.training.train_starvla][INFO] - Step 3598, Loss: {'train/loss': 0.2826627194881439, 'train/grad_norm_pre_clip': 11.358416557312012, 'timing/data': 0.001360367052257061, 'timing/model': 1.091978325508535, 'learning_rate/qwen_vl_interface': 5.1461991148327095e-06, 'learning_rate/action_model': 5.1461991148327095e-05, 'epoch': 1.48}) +[2026-06-17 22:22:20,070][starVLA.training.train_starvla][INFO] - Step 3599, Loss: {'train/loss': 0.6188678741455078, 'train/grad_norm_pre_clip': 21.544174194335938, 'timing/data': 0.000516778789460659, 'timing/model': 0.9385626660659909, 'learning_rate/qwen_vl_interface': 5.144036951524977e-06, 'learning_rate/action_model': 5.144036951524977e-05, 'epoch': 1.48}) +[2026-06-17 22:24:24,103][starVLA.training.train_starvla][INFO] - Step 3600, Loss: {'train/loss': 0.4213825464248657, 'train/grad_norm_pre_clip': 13.11412525177002, 'eval/loss': 0.4720050096511841, 'eval/latency_6/loss': 0.4720050096511841, 'eval/demon_fire_precision': 0.9431782703715267, 'eval/demon_fire_recall': 0.9764059469941823, 'eval/demon_fire_f1': 0.9595045259647451, 'eval/demon_fire_rate_model': 0.8912075681691708, 'eval/demon_fire_rate_teacher': 0.860879243183083, 'eval/demon_left_precision': 0.8176352705410822, 'eval/demon_left_recall': 0.7157894736842105, 'eval/demon_left_f1': 0.76333021515435, 'eval/demon_left_rate_model': 0.13884251530328326, 'eval/demon_left_rate_teacher': 0.15859766277128548, 'eval/demon_right_precision': 0.7901639344262295, 'eval/demon_right_recall': 0.7403993855606759, 'eval/demon_right_f1': 0.7644726407613006, 'eval/demon_right_rate_model': 0.16972732331663884, 'eval/demon_right_rate_teacher': 0.18113522537562604, 'eval/demon_move_f1': 0.7639014279578253, 'eval/demon_cc_f1': 0.8910559179374339, 'eval/latency_6/demon_fire_precision': 0.9431782703715267, 'eval/latency_6/demon_fire_recall': 0.9764059469941823, 'eval/latency_6/demon_fire_f1': 0.9595045259647451, 'eval/latency_6/demon_fire_rate_model': 0.8912075681691708, 'eval/latency_6/demon_fire_rate_teacher': 0.860879243183083, 'eval/latency_6/demon_left_precision': 0.8176352705410822, 'eval/latency_6/demon_left_recall': 0.7157894736842105, 'eval/latency_6/demon_left_f1': 0.76333021515435, 'eval/latency_6/demon_left_rate_model': 0.13884251530328326, 'eval/latency_6/demon_left_rate_teacher': 0.15859766277128548, 'eval/latency_6/demon_right_precision': 0.7901639344262295, 'eval/latency_6/demon_right_recall': 0.7403993855606759, 'eval/latency_6/demon_right_f1': 0.7644726407613006, 'eval/latency_6/demon_right_rate_model': 0.16972732331663884, 'eval/latency_6/demon_right_rate_teacher': 0.18113522537562604, 'eval/latency_6/demon_move_f1': 0.7639014279578253, 'eval/latency_6/demon_cc_f1': 0.8910559179374339, 'timing/data': 0.0004937341436743736, 'timing/model': 0.921559582464397, 'learning_rate/qwen_vl_interface': 5.141874810183498e-06, 'learning_rate/action_model': 5.141874810183498e-05, 'epoch': 1.48}) +[2026-06-17 22:24:24,106][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_3600_state +[2026-06-17 22:24:24,110][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 22:24:30,693][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_3600_state/model.safetensors +[2026-06-17 22:24:45,678][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_3600_state/optimizer.bin +[2026-06-17 22:24:45,681][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_3600_state/sampler.bin +[2026-06-17 22:24:45,681][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_3600_state/sampler_1.bin +[2026-06-17 22:24:45,683][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_3600_state/random_states_0.pkl +[2026-06-17 22:24:45,686][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3600 +[2026-06-17 22:24:45,687][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 22:24:45,698][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 22:24:49,276][starVLA.training.train_starvla][INFO] - Step 3601, Loss: {'train/loss': 0.2597575783729553, 'train/grad_norm_pre_clip': 13.22591781616211, 'timing/data': 0.0007028728723526001, 'timing/model': 0.9239943595603108, 'learning_rate/qwen_vl_interface': 5.139712691256489e-06, 'learning_rate/action_model': 5.1397126912564885e-05, 'epoch': 1.48}) +[2026-06-17 22:24:53,079][starVLA.training.train_starvla][INFO] - Step 3602, Loss: {'train/loss': 0.6055284738540649, 'train/grad_norm_pre_clip': 17.347230911254883, 'timing/data': 0.0005165431648492813, 'timing/model': 0.9247698439285159, 'learning_rate/qwen_vl_interface': 5.137550595192156e-06, 'learning_rate/action_model': 5.137550595192156e-05, 'epoch': 1.48}) +[2026-06-17 22:24:56,789][starVLA.training.train_starvla][INFO] - Step 3603, Loss: {'train/loss': 0.4314669668674469, 'train/grad_norm_pre_clip': 13.85477066040039, 'timing/data': 0.0005414886400103569, 'timing/model': 0.929494041018188, 'learning_rate/qwen_vl_interface': 5.135388522438707e-06, 'learning_rate/action_model': 5.1353885224387066e-05, 'epoch': 1.48}) +[2026-06-17 22:25:00,673][starVLA.training.train_starvla][INFO] - Step 3604, Loss: {'train/loss': 0.5615528225898743, 'train/grad_norm_pre_clip': 18.122377395629883, 'timing/data': 0.00047803856432437897, 'timing/model': 1.0508777648210526, 'learning_rate/qwen_vl_interface': 5.133226473444341e-06, 'learning_rate/action_model': 5.13322647344434e-05, 'epoch': 1.48}) +[2026-06-17 22:25:04,899][starVLA.training.train_starvla][INFO] - Step 3605, Loss: {'train/loss': 0.37850314378738403, 'train/grad_norm_pre_clip': 17.032821655273438, 'timing/data': 0.0007967455312609673, 'timing/model': 1.366011319682002, 'learning_rate/qwen_vl_interface': 5.1310644486572515e-06, 'learning_rate/action_model': 5.131064448657251e-05, 'epoch': 1.48}) +[2026-06-17 22:25:08,726][starVLA.training.train_starvla][INFO] - Step 3606, Loss: {'train/loss': 0.3125016689300537, 'train/grad_norm_pre_clip': 13.62771987915039, 'timing/data': 0.0011896099895238876, 'timing/model': 1.1340578319504857, 'learning_rate/qwen_vl_interface': 5.12890244852563e-06, 'learning_rate/action_model': 5.12890244852563e-05, 'epoch': 1.48}) +[2026-06-17 22:25:12,548][starVLA.training.train_starvla][INFO] - Step 3607, Loss: {'train/loss': 0.3808072507381439, 'train/grad_norm_pre_clip': 15.160266876220703, 'timing/data': 0.0005002683028578758, 'timing/model': 1.1206051418557763, 'learning_rate/qwen_vl_interface': 5.126740473497661e-06, 'learning_rate/action_model': 5.126740473497661e-05, 'epoch': 1.48}) +[2026-06-17 22:25:16,274][starVLA.training.train_starvla][INFO] - Step 3608, Loss: {'train/loss': 0.2513090670108795, 'train/grad_norm_pre_clip': 16.066476821899414, 'timing/data': 0.0010860394686460495, 'timing/model': 0.9334665182977915, 'learning_rate/qwen_vl_interface': 5.124578524021525e-06, 'learning_rate/action_model': 5.124578524021525e-05, 'epoch': 1.49}) +[2026-06-17 22:25:20,278][starVLA.training.train_starvla][INFO] - Step 3609, Loss: {'train/loss': 0.7242910861968994, 'train/grad_norm_pre_clip': 17.562278747558594, 'timing/data': 0.000512467697262764, 'timing/model': 0.9364545801654458, 'learning_rate/qwen_vl_interface': 5.122416600545396e-06, 'learning_rate/action_model': 5.122416600545397e-05, 'epoch': 1.49}) +[2026-06-17 22:25:24,086][starVLA.training.train_starvla][INFO] - Step 3610, Loss: {'train/loss': 0.6096658110618591, 'train/grad_norm_pre_clip': 18.872587203979492, 'timing/data': 0.0005166120827198029, 'timing/model': 0.9497804930433631, 'learning_rate/qwen_vl_interface': 5.1202547035174436e-06, 'learning_rate/action_model': 5.120254703517443e-05, 'epoch': 1.49}) +[2026-06-17 22:25:27,935][starVLA.training.train_starvla][INFO] - Step 3611, Loss: {'train/loss': 0.29007238149642944, 'train/grad_norm_pre_clip': 16.421344757080078, 'timing/data': 0.0005859509110450745, 'timing/model': 0.9315780932083726, 'learning_rate/qwen_vl_interface': 5.118092833385829e-06, 'learning_rate/action_model': 5.1180928333858293e-05, 'epoch': 1.49}) +[2026-06-17 22:25:31,937][starVLA.training.train_starvla][INFO] - Step 3612, Loss: {'train/loss': 0.34531235694885254, 'train/grad_norm_pre_clip': 17.643917083740234, 'timing/data': 0.0004968317225575447, 'timing/model': 0.9399714581668377, 'learning_rate/qwen_vl_interface': 5.115930990598713e-06, 'learning_rate/action_model': 5.115930990598713e-05, 'epoch': 1.49}) +[2026-06-17 22:25:35,966][starVLA.training.train_starvla][INFO] - Step 3613, Loss: {'train/loss': 0.17633134126663208, 'train/grad_norm_pre_clip': 10.417404174804688, 'timing/data': 0.0005219895392656326, 'timing/model': 1.1198891177773476, 'learning_rate/qwen_vl_interface': 5.113769175604248e-06, 'learning_rate/action_model': 5.113769175604247e-05, 'epoch': 1.49}) +[2026-06-17 22:25:39,705][starVLA.training.train_starvla][INFO] - Step 3614, Loss: {'train/loss': 0.41389110684394836, 'train/grad_norm_pre_clip': 15.043211936950684, 'timing/data': 0.0005312571302056313, 'timing/model': 1.053689050488174, 'learning_rate/qwen_vl_interface': 5.111607388850576e-06, 'learning_rate/action_model': 5.111607388850576e-05, 'epoch': 1.49}) +[2026-06-17 22:25:43,634][starVLA.training.train_starvla][INFO] - Step 3615, Loss: {'train/loss': 0.5721129179000854, 'train/grad_norm_pre_clip': 21.210317611694336, 'timing/data': 0.0005877334624528885, 'timing/model': 0.932864542119205, 'learning_rate/qwen_vl_interface': 5.1094456307858444e-06, 'learning_rate/action_model': 5.109445630785844e-05, 'epoch': 1.49}) +[2026-06-17 22:25:47,575][starVLA.training.train_starvla][INFO] - Step 3616, Loss: {'train/loss': 0.35205137729644775, 'train/grad_norm_pre_clip': 15.998970031738281, 'timing/data': 0.0005111154168844223, 'timing/model': 0.9360763384029269, 'learning_rate/qwen_vl_interface': 5.107283901858184e-06, 'learning_rate/action_model': 5.107283901858184e-05, 'epoch': 1.49}) +[2026-06-17 22:25:51,466][starVLA.training.train_starvla][INFO] - Step 3617, Loss: {'train/loss': 0.26273271441459656, 'train/grad_norm_pre_clip': 274.3888854980469, 'timing/data': 0.0008616605773568153, 'timing/model': 0.9343378413468599, 'learning_rate/qwen_vl_interface': 5.105122202515726e-06, 'learning_rate/action_model': 5.1051222025157255e-05, 'epoch': 1.49}) +[2026-06-17 22:25:55,314][starVLA.training.train_starvla][INFO] - Step 3618, Loss: {'train/loss': 0.396340548992157, 'train/grad_norm_pre_clip': 13.972553253173828, 'timing/data': 0.0006963647902011871, 'timing/model': 0.936075552366674, 'learning_rate/qwen_vl_interface': 5.102960533206588e-06, 'learning_rate/action_model': 5.1029605332065874e-05, 'epoch': 1.49}) +[2026-06-17 22:25:59,460][starVLA.training.train_starvla][INFO] - Step 3619, Loss: {'train/loss': 0.39803192019462585, 'train/grad_norm_pre_clip': 25.4565486907959, 'timing/data': 0.0008249469101428986, 'timing/model': 1.2709669638425112, 'learning_rate/qwen_vl_interface': 5.100798894378891e-06, 'learning_rate/action_model': 5.100798894378891e-05, 'epoch': 1.49}) +[2026-06-17 22:26:03,307][starVLA.training.train_starvla][INFO] - Step 3620, Loss: {'train/loss': 0.32897645235061646, 'train/grad_norm_pre_clip': 14.39555549621582, 'timing/data': 0.0008646976202726364, 'timing/model': 1.094583187252283, 'learning_rate/qwen_vl_interface': 5.098637286480744e-06, 'learning_rate/action_model': 5.098637286480744e-05, 'epoch': 1.49}) +[2026-06-17 22:26:07,162][starVLA.training.train_starvla][INFO] - Step 3621, Loss: {'train/loss': 0.38426777720451355, 'train/grad_norm_pre_clip': 100.49459075927734, 'timing/data': 0.0007698209956288338, 'timing/model': 1.1809425968676805, 'learning_rate/qwen_vl_interface': 5.096475709960252e-06, 'learning_rate/action_model': 5.096475709960251e-05, 'epoch': 1.49}) +[2026-06-17 22:26:11,489][starVLA.training.train_starvla][INFO] - Step 3622, Loss: {'train/loss': 0.36256685853004456, 'train/grad_norm_pre_clip': 17.917980194091797, 'timing/data': 0.00047300849109888077, 'timing/model': 1.317919890396297, 'learning_rate/qwen_vl_interface': 5.094314165265506e-06, 'learning_rate/action_model': 5.0943141652655056e-05, 'epoch': 1.49}) +[2026-06-17 22:26:15,370][starVLA.training.train_starvla][INFO] - Step 3623, Loss: {'train/loss': 0.3373390734195709, 'train/grad_norm_pre_clip': 20.086172103881836, 'timing/data': 0.00048458948731422424, 'timing/model': 0.9379925783723593, 'learning_rate/qwen_vl_interface': 5.092152652844606e-06, 'learning_rate/action_model': 5.092152652844606e-05, 'epoch': 1.49}) +[2026-06-17 22:26:19,340][starVLA.training.train_starvla][INFO] - Step 3624, Loss: {'train/loss': 0.3667601943016052, 'train/grad_norm_pre_clip': 28.089956283569336, 'timing/data': 0.0005034962669014931, 'timing/model': 0.9376338059082627, 'learning_rate/qwen_vl_interface': 5.089991173145628e-06, 'learning_rate/action_model': 5.0899911731456285e-05, 'epoch': 1.49}) +[2026-06-17 22:26:23,319][starVLA.training.train_starvla][INFO] - Step 3625, Loss: {'train/loss': 0.5175119042396545, 'train/grad_norm_pre_clip': 16.722909927368164, 'timing/data': 0.000530492514371872, 'timing/model': 0.9321438167244196, 'learning_rate/qwen_vl_interface': 5.087829726616653e-06, 'learning_rate/action_model': 5.087829726616653e-05, 'epoch': 1.49}) +[2026-06-17 22:26:27,332][starVLA.training.train_starvla][INFO] - Step 3626, Loss: {'train/loss': 0.6686853766441345, 'train/grad_norm_pre_clip': 21.323450088500977, 'timing/data': 0.0007241899147629738, 'timing/model': 0.9354207990691066, 'learning_rate/qwen_vl_interface': 5.08566831370575e-06, 'learning_rate/action_model': 5.0856683137057504e-05, 'epoch': 1.49}) +[2026-06-17 22:26:31,135][starVLA.training.train_starvla][INFO] - Step 3627, Loss: {'train/loss': 0.2973780333995819, 'train/grad_norm_pre_clip': 14.020673751831055, 'timing/data': 0.00048671476542949677, 'timing/model': 1.1566504472866654, 'learning_rate/qwen_vl_interface': 5.083506934860982e-06, 'learning_rate/action_model': 5.083506934860982e-05, 'epoch': 1.49}) +[2026-06-17 22:26:34,903][starVLA.training.train_starvla][INFO] - Step 3628, Loss: {'train/loss': 0.2681662440299988, 'train/grad_norm_pre_clip': 12.847142219543457, 'timing/data': 0.0009735478088259697, 'timing/model': 1.1092982552945614, 'learning_rate/qwen_vl_interface': 5.081345590530407e-06, 'learning_rate/action_model': 5.081345590530407e-05, 'epoch': 1.49}) +[2026-06-17 22:26:38,883][starVLA.training.train_starvla][INFO] - Step 3629, Loss: {'train/loss': 0.24650730192661285, 'train/grad_norm_pre_clip': 11.387208938598633, 'timing/data': 0.0004954002797603607, 'timing/model': 0.9423172492533922, 'learning_rate/qwen_vl_interface': 5.079184281162069e-06, 'learning_rate/action_model': 5.0791842811620694e-05, 'epoch': 1.49}) +[2026-06-17 22:26:42,701][starVLA.training.train_starvla][INFO] - Step 3630, Loss: {'train/loss': 0.5420501828193665, 'train/grad_norm_pre_clip': 22.098440170288086, 'timing/data': 0.000516524538397789, 'timing/model': 0.946857588365674, 'learning_rate/qwen_vl_interface': 5.077023007204018e-06, 'learning_rate/action_model': 5.077023007204018e-05, 'epoch': 1.49}) +[2026-06-17 22:26:46,787][starVLA.training.train_starvla][INFO] - Step 3631, Loss: {'train/loss': 0.5638125538825989, 'train/grad_norm_pre_clip': 16.805112838745117, 'timing/data': 0.0004938729107379913, 'timing/model': 0.9431911492720246, 'learning_rate/qwen_vl_interface': 5.074861769104281e-06, 'learning_rate/action_model': 5.074861769104281e-05, 'epoch': 1.49}) +[2026-06-17 22:26:50,679][starVLA.training.train_starvla][INFO] - Step 3632, Loss: {'train/loss': 0.1432146281003952, 'train/grad_norm_pre_clip': 8.662663459777832, 'timing/data': 0.000466163270175457, 'timing/model': 0.9601425416767597, 'learning_rate/qwen_vl_interface': 5.07270056731089e-06, 'learning_rate/action_model': 5.07270056731089e-05, 'epoch': 1.5}) +[2026-06-17 22:26:54,501][starVLA.training.train_starvla][INFO] - Step 3633, Loss: {'train/loss': 0.37123146653175354, 'train/grad_norm_pre_clip': 17.396223068237305, 'timing/data': 0.0005458071827888489, 'timing/model': 1.1381215574219823, 'learning_rate/qwen_vl_interface': 5.070539402271859e-06, 'learning_rate/action_model': 5.070539402271859e-05, 'epoch': 1.5}) +[2026-06-17 22:26:58,389][starVLA.training.train_starvla][INFO] - Step 3634, Loss: {'train/loss': 0.4934925436973572, 'train/grad_norm_pre_clip': 53.943485260009766, 'timing/data': 0.0005857525393366814, 'timing/model': 1.0476759131997824, 'learning_rate/qwen_vl_interface': 5.068378274435205e-06, 'learning_rate/action_model': 5.068378274435205e-05, 'epoch': 1.5}) +[2026-06-17 22:27:02,033][starVLA.training.train_starvla][INFO] - Step 3635, Loss: {'train/loss': 0.4083670377731323, 'train/grad_norm_pre_clip': 17.98409080505371, 'timing/data': 0.0012752339243888855, 'timing/model': 0.9403714369982481, 'learning_rate/qwen_vl_interface': 5.066217184248931e-06, 'learning_rate/action_model': 5.0662171842489305e-05, 'epoch': 1.5}) +[2026-06-17 22:27:05,940][starVLA.training.train_starvla][INFO] - Step 3636, Loss: {'train/loss': 0.29996824264526367, 'train/grad_norm_pre_clip': 13.820121765136719, 'timing/data': 0.0005103182047605515, 'timing/model': 0.9439423074945807, 'learning_rate/qwen_vl_interface': 5.0640561321610295e-06, 'learning_rate/action_model': 5.064056132161029e-05, 'epoch': 1.5}) +[2026-06-17 22:27:10,379][starVLA.training.train_starvla][INFO] - Step 3637, Loss: {'train/loss': 0.4476092457771301, 'train/grad_norm_pre_clip': 19.023967742919922, 'timing/data': 0.0004975404590368271, 'timing/model': 1.108105918392539, 'learning_rate/qwen_vl_interface': 5.061895118619493e-06, 'learning_rate/action_model': 5.061895118619493e-05, 'epoch': 1.5}) +[2026-06-17 22:27:14,259][starVLA.training.train_starvla][INFO] - Step 3638, Loss: {'train/loss': 0.648869514465332, 'train/grad_norm_pre_clip': 23.359638214111328, 'timing/data': 0.0005560163408517838, 'timing/model': 0.9810785809531808, 'learning_rate/qwen_vl_interface': 5.059734144072302e-06, 'learning_rate/action_model': 5.059734144072302e-05, 'epoch': 1.5}) +[2026-06-17 22:27:18,006][starVLA.training.train_starvla][INFO] - Step 3639, Loss: {'train/loss': 0.30400124192237854, 'train/grad_norm_pre_clip': 13.99942398071289, 'timing/data': 0.0006588995456695557, 'timing/model': 0.9323536641895771, 'learning_rate/qwen_vl_interface': 5.0575732089674275e-06, 'learning_rate/action_model': 5.057573208967428e-05, 'epoch': 1.5}) +[2026-06-17 22:27:21,900][starVLA.training.train_starvla][INFO] - Step 3640, Loss: {'train/loss': 0.46446871757507324, 'train/grad_norm_pre_clip': 18.030475616455078, 'timing/data': 0.0005145780742168427, 'timing/model': 1.1106962682679296, 'learning_rate/qwen_vl_interface': 5.055412313752832e-06, 'learning_rate/action_model': 5.055412313752832e-05, 'epoch': 1.5}) +[2026-06-17 22:27:25,642][starVLA.training.train_starvla][INFO] - Step 3641, Loss: {'train/loss': 0.35376620292663574, 'train/grad_norm_pre_clip': 63.22383117675781, 'timing/data': 0.000647946260869503, 'timing/model': 0.9291744977235794, 'learning_rate/qwen_vl_interface': 5.053251458876474e-06, 'learning_rate/action_model': 5.0532514588764744e-05, 'epoch': 1.5}) +[2026-06-17 22:27:29,510][starVLA.training.train_starvla][INFO] - Step 3642, Loss: {'train/loss': 0.3533208966255188, 'train/grad_norm_pre_clip': 16.81273651123047, 'timing/data': 0.0005259355530142784, 'timing/model': 0.9392489055171609, 'learning_rate/qwen_vl_interface': 5.0510906447863025e-06, 'learning_rate/action_model': 5.051090644786303e-05, 'epoch': 1.5}) +[2026-06-17 22:27:33,305][starVLA.training.train_starvla][INFO] - Step 3643, Loss: {'train/loss': 0.3660198450088501, 'train/grad_norm_pre_clip': 17.4091739654541, 'timing/data': 0.0004888568073511124, 'timing/model': 0.9371617697179317, 'learning_rate/qwen_vl_interface': 5.048929871930252e-06, 'learning_rate/action_model': 5.048929871930251e-05, 'epoch': 1.5}) +[2026-06-17 22:27:37,049][starVLA.training.train_starvla][INFO] - Step 3644, Loss: {'train/loss': 0.16961073875427246, 'train/grad_norm_pre_clip': 9.533167839050293, 'timing/data': 0.0004947688430547714, 'timing/model': 0.9220533017069101, 'learning_rate/qwen_vl_interface': 5.046769140756257e-06, 'learning_rate/action_model': 5.046769140756257e-05, 'epoch': 1.5}) +[2026-06-17 22:27:40,843][starVLA.training.train_starvla][INFO] - Step 3645, Loss: {'train/loss': 0.3629485070705414, 'train/grad_norm_pre_clip': 15.96521282196045, 'timing/data': 0.000651458278298378, 'timing/model': 0.924072508700192, 'learning_rate/qwen_vl_interface': 5.044608451712237e-06, 'learning_rate/action_model': 5.044608451712237e-05, 'epoch': 1.5}) +[2026-06-17 22:27:44,644][starVLA.training.train_starvla][INFO] - Step 3646, Loss: {'train/loss': 0.6816527843475342, 'train/grad_norm_pre_clip': 21.191333770751953, 'timing/data': 0.0005129855126142502, 'timing/model': 1.1143621541559696, 'learning_rate/qwen_vl_interface': 5.042447805246105e-06, 'learning_rate/action_model': 5.042447805246106e-05, 'epoch': 1.5}) +[2026-06-17 22:27:48,243][starVLA.training.train_starvla][INFO] - Step 3647, Loss: {'train/loss': 0.24961161613464355, 'train/grad_norm_pre_clip': 12.498215675354004, 'timing/data': 0.0005220901221036911, 'timing/model': 0.9243561746552587, 'learning_rate/qwen_vl_interface': 5.040287201805767e-06, 'learning_rate/action_model': 5.040287201805767e-05, 'epoch': 1.5}) +[2026-06-17 22:27:51,992][starVLA.training.train_starvla][INFO] - Step 3648, Loss: {'train/loss': 0.49459192156791687, 'train/grad_norm_pre_clip': 17.460391998291016, 'timing/data': 0.0005088336765766144, 'timing/model': 0.9252829616889358, 'learning_rate/qwen_vl_interface': 5.038126641839119e-06, 'learning_rate/action_model': 5.0381266418391184e-05, 'epoch': 1.5}) +[2026-06-17 22:27:55,738][starVLA.training.train_starvla][INFO] - Step 3649, Loss: {'train/loss': 0.25215020775794983, 'train/grad_norm_pre_clip': 10.91821002960205, 'timing/data': 0.0004763556644320488, 'timing/model': 0.9236262813210487, 'learning_rate/qwen_vl_interface': 5.0359661257940455e-06, 'learning_rate/action_model': 5.035966125794045e-05, 'epoch': 1.5}) +[2026-06-17 22:27:59,475][starVLA.training.train_starvla][INFO] - Step 3650, Loss: {'train/loss': 0.46784713864326477, 'train/grad_norm_pre_clip': 20.330995559692383, 'timing/data': 0.0005388455465435982, 'timing/model': 0.9232089687138796, 'learning_rate/qwen_vl_interface': 5.033805654118424e-06, 'learning_rate/action_model': 5.033805654118424e-05, 'epoch': 1.5}) +[2026-06-17 22:28:03,206][starVLA.training.train_starvla][INFO] - Step 3651, Loss: {'train/loss': 0.46346601843833923, 'train/grad_norm_pre_clip': 19.403385162353516, 'timing/data': 0.0004922142252326012, 'timing/model': 0.9226522985845804, 'learning_rate/qwen_vl_interface': 5.031645227260122e-06, 'learning_rate/action_model': 5.0316452272601224e-05, 'epoch': 1.5}) +[2026-06-17 22:28:06,978][starVLA.training.train_starvla][INFO] - Step 3652, Loss: {'train/loss': 0.29102176427841187, 'train/grad_norm_pre_clip': 14.194340705871582, 'timing/data': 0.0010032886639237404, 'timing/model': 0.9382487758994102, 'learning_rate/qwen_vl_interface': 5.029484845667001e-06, 'learning_rate/action_model': 5.029484845667001e-05, 'epoch': 1.5}) +[2026-06-17 22:28:10,745][starVLA.training.train_starvla][INFO] - Step 3653, Loss: {'train/loss': 0.24061836302280426, 'train/grad_norm_pre_clip': 15.252013206481934, 'timing/data': 0.0006625987589359283, 'timing/model': 1.1046606618911028, 'learning_rate/qwen_vl_interface': 5.027324509786908e-06, 'learning_rate/action_model': 5.027324509786908e-05, 'epoch': 1.5}) +[2026-06-17 22:28:14,354][starVLA.training.train_starvla][INFO] - Step 3654, Loss: {'train/loss': 0.19261853396892548, 'train/grad_norm_pre_clip': 13.320460319519043, 'timing/data': 0.0009007547050714493, 'timing/model': 0.9270692570134997, 'learning_rate/qwen_vl_interface': 5.025164220067683e-06, 'learning_rate/action_model': 5.025164220067683e-05, 'epoch': 1.5}) +[2026-06-17 22:28:18,137][starVLA.training.train_starvla][INFO] - Step 3655, Loss: {'train/loss': 0.3467543423175812, 'train/grad_norm_pre_clip': 17.334138870239258, 'timing/data': 0.0005285497754812241, 'timing/model': 0.9463505856692791, 'learning_rate/qwen_vl_interface': 5.023003976957159e-06, 'learning_rate/action_model': 5.023003976957159e-05, 'epoch': 1.5}) +[2026-06-17 22:28:21,904][starVLA.training.train_starvla][INFO] - Step 3656, Loss: {'train/loss': 0.20843830704689026, 'train/grad_norm_pre_clip': 16.49823760986328, 'timing/data': 0.0008103018626570702, 'timing/model': 0.9309565033763647, 'learning_rate/qwen_vl_interface': 5.020843780903155e-06, 'learning_rate/action_model': 5.020843780903155e-05, 'epoch': 1.5}) +[2026-06-17 22:28:25,646][starVLA.training.train_starvla][INFO] - Step 3657, Loss: {'train/loss': 0.6043719053268433, 'train/grad_norm_pre_clip': 38.3214225769043, 'timing/data': 0.000493612140417099, 'timing/model': 0.9207374658435583, 'learning_rate/qwen_vl_interface': 5.01868363235348e-06, 'learning_rate/action_model': 5.01868363235348e-05, 'epoch': 1.51}) +[2026-06-17 22:28:29,385][starVLA.training.train_starvla][INFO] - Step 3658, Loss: {'train/loss': 0.20451225340366364, 'train/grad_norm_pre_clip': 13.40600872039795, 'timing/data': 0.0005703819915652275, 'timing/model': 0.9276713635772467, 'learning_rate/qwen_vl_interface': 5.016523531755937e-06, 'learning_rate/action_model': 5.016523531755937e-05, 'epoch': 1.51}) +[2026-06-17 22:28:33,185][starVLA.training.train_starvla][INFO] - Step 3659, Loss: {'train/loss': 0.4869879484176636, 'train/grad_norm_pre_clip': 22.7031192779541, 'timing/data': 0.0005208533257246017, 'timing/model': 1.1392518756911159, 'learning_rate/qwen_vl_interface': 5.014363479558318e-06, 'learning_rate/action_model': 5.014363479558318e-05, 'epoch': 1.51}) +[2026-06-17 22:28:36,931][starVLA.training.train_starvla][INFO] - Step 3660, Loss: {'train/loss': 0.47896647453308105, 'train/grad_norm_pre_clip': 18.2872314453125, 'timing/data': 0.0005099056288599968, 'timing/model': 1.1016860576346517, 'learning_rate/qwen_vl_interface': 5.012203476208403e-06, 'learning_rate/action_model': 5.0122034762084025e-05, 'epoch': 1.51}) +[2026-06-17 22:28:40,553][starVLA.training.train_starvla][INFO] - Step 3661, Loss: {'train/loss': 0.41585150361061096, 'train/grad_norm_pre_clip': 18.941730499267578, 'timing/data': 0.0005121305584907532, 'timing/model': 0.9290670324116945, 'learning_rate/qwen_vl_interface': 5.010043522153965e-06, 'learning_rate/action_model': 5.0100435221539646e-05, 'epoch': 1.51}) +[2026-06-17 22:28:44,337][starVLA.training.train_starvla][INFO] - Step 3662, Loss: {'train/loss': 0.4504198729991913, 'train/grad_norm_pre_clip': 17.66890525817871, 'timing/data': 0.0004919338971376419, 'timing/model': 0.9293992510065436, 'learning_rate/qwen_vl_interface': 5.007883617842759e-06, 'learning_rate/action_model': 5.007883617842759e-05, 'epoch': 1.51}) +[2026-06-17 22:28:48,110][starVLA.training.train_starvla][INFO] - Step 3663, Loss: {'train/loss': 0.4034295678138733, 'train/grad_norm_pre_clip': 25.97365379333496, 'timing/data': 0.0010659517720341682, 'timing/model': 0.9300963645800948, 'learning_rate/qwen_vl_interface': 5.005723763722543e-06, 'learning_rate/action_model': 5.005723763722543e-05, 'epoch': 1.51}) +[2026-06-17 22:28:51,853][starVLA.training.train_starvla][INFO] - Step 3664, Loss: {'train/loss': 0.3418564200401306, 'train/grad_norm_pre_clip': 14.710110664367676, 'timing/data': 0.0006466424092650414, 'timing/model': 0.9248365322127938, 'learning_rate/qwen_vl_interface': 5.003563960241052e-06, 'learning_rate/action_model': 5.003563960241052e-05, 'epoch': 1.51}) +[2026-06-17 22:28:55,624][starVLA.training.train_starvla][INFO] - Step 3665, Loss: {'train/loss': 0.5925508737564087, 'train/grad_norm_pre_clip': 21.16629409790039, 'timing/data': 0.0007096165791153908, 'timing/model': 0.9234859561547637, 'learning_rate/qwen_vl_interface': 5.001404207846017e-06, 'learning_rate/action_model': 5.001404207846017e-05, 'epoch': 1.51}) +[2026-06-17 22:28:59,366][starVLA.training.train_starvla][INFO] - Step 3666, Loss: {'train/loss': 0.21119603514671326, 'train/grad_norm_pre_clip': 11.465987205505371, 'timing/data': 0.0006247395649552345, 'timing/model': 1.0916651152074337, 'learning_rate/qwen_vl_interface': 4.999244506985158e-06, 'learning_rate/action_model': 4.999244506985158e-05, 'epoch': 1.51}) +[2026-06-17 22:29:03,133][starVLA.training.train_starvla][INFO] - Step 3667, Loss: {'train/loss': 0.46261066198349, 'train/grad_norm_pre_clip': 17.841386795043945, 'timing/data': 0.0009478786960244179, 'timing/model': 1.0942476857453585, 'learning_rate/qwen_vl_interface': 4.997084858106182e-06, 'learning_rate/action_model': 4.997084858106183e-05, 'epoch': 1.51}) +[2026-06-17 22:29:06,762][starVLA.training.train_starvla][INFO] - Step 3668, Loss: {'train/loss': 0.2000337392091751, 'train/grad_norm_pre_clip': 10.762591361999512, 'timing/data': 0.0005159489810466766, 'timing/model': 0.9391623614355922, 'learning_rate/qwen_vl_interface': 4.994925261656789e-06, 'learning_rate/action_model': 4.994925261656789e-05, 'epoch': 1.51}) +[2026-06-17 22:29:10,572][starVLA.training.train_starvla][INFO] - Step 3669, Loss: {'train/loss': 0.4449312388896942, 'train/grad_norm_pre_clip': 19.785104751586914, 'timing/data': 0.000512721948325634, 'timing/model': 0.9329370865598321, 'learning_rate/qwen_vl_interface': 4.99276571808466e-06, 'learning_rate/action_model': 4.9927657180846605e-05, 'epoch': 1.51}) +[2026-06-17 22:29:14,342][starVLA.training.train_starvla][INFO] - Step 3670, Loss: {'train/loss': 0.5007556676864624, 'train/grad_norm_pre_clip': 20.59366798400879, 'timing/data': 0.0005076630041003227, 'timing/model': 0.9348587114363909, 'learning_rate/qwen_vl_interface': 4.9906062278374775e-06, 'learning_rate/action_model': 4.990606227837478e-05, 'epoch': 1.51}) +[2026-06-17 22:29:18,053][starVLA.training.train_starvla][INFO] - Step 3671, Loss: {'train/loss': 0.3096029758453369, 'train/grad_norm_pre_clip': 14.919677734375, 'timing/data': 0.0005316277965903282, 'timing/model': 0.9265636745840311, 'learning_rate/qwen_vl_interface': 4.988446791362903e-06, 'learning_rate/action_model': 4.988446791362903e-05, 'epoch': 1.51}) +[2026-06-17 22:29:21,837][starVLA.training.train_starvla][INFO] - Step 3672, Loss: {'train/loss': 0.33382952213287354, 'train/grad_norm_pre_clip': 29.172481536865234, 'timing/data': 0.00048048608005046844, 'timing/model': 0.9261894105002284, 'learning_rate/qwen_vl_interface': 4.986287409108589e-06, 'learning_rate/action_model': 4.9862874091085893e-05, 'epoch': 1.51}) +[2026-06-17 22:29:25,591][starVLA.training.train_starvla][INFO] - Step 3673, Loss: {'train/loss': 0.3986251652240753, 'train/grad_norm_pre_clip': 19.111759185791016, 'timing/data': 0.0008879369124770164, 'timing/model': 1.095219419337809, 'learning_rate/qwen_vl_interface': 4.984128081522179e-06, 'learning_rate/action_model': 4.984128081522178e-05, 'epoch': 1.51}) +[2026-06-17 22:29:29,169][starVLA.training.train_starvla][INFO] - Step 3674, Loss: {'train/loss': 0.4382581412792206, 'train/grad_norm_pre_clip': 16.77960205078125, 'timing/data': 0.0005071079358458519, 'timing/model': 0.9225202649831772, 'learning_rate/qwen_vl_interface': 4.981968809051304e-06, 'learning_rate/action_model': 4.981968809051304e-05, 'epoch': 1.51}) +[2026-06-17 22:29:32,926][starVLA.training.train_starvla][INFO] - Step 3675, Loss: {'train/loss': 0.2807754874229431, 'train/grad_norm_pre_clip': 16.368236541748047, 'timing/data': 0.0009888196364045143, 'timing/model': 0.9334880625829101, 'learning_rate/qwen_vl_interface': 4.979809592143583e-06, 'learning_rate/action_model': 4.979809592143583e-05, 'epoch': 1.51}) +[2026-06-17 22:29:36,626][starVLA.training.train_starvla][INFO] - Step 3676, Loss: {'train/loss': 0.22707927227020264, 'train/grad_norm_pre_clip': 13.869385719299316, 'timing/data': 0.0011281324550509453, 'timing/model': 0.9299598364159465, 'learning_rate/qwen_vl_interface': 4.977650431246624e-06, 'learning_rate/action_model': 4.977650431246624e-05, 'epoch': 1.51}) +[2026-06-17 22:29:40,483][starVLA.training.train_starvla][INFO] - Step 3677, Loss: {'train/loss': 0.41024187207221985, 'train/grad_norm_pre_clip': 21.08492088317871, 'timing/data': 0.001153123565018177, 'timing/model': 1.0112330885604024, 'learning_rate/qwen_vl_interface': 4.9754913268080255e-06, 'learning_rate/action_model': 4.975491326808025e-05, 'epoch': 1.51}) +[2026-06-17 22:29:44,275][starVLA.training.train_starvla][INFO] - Step 3678, Loss: {'train/loss': 0.46385908126831055, 'train/grad_norm_pre_clip': 23.640018463134766, 'timing/data': 0.0009869011119008064, 'timing/model': 0.9338667457923293, 'learning_rate/qwen_vl_interface': 4.973332279275369e-06, 'learning_rate/action_model': 4.973332279275369e-05, 'epoch': 1.51}) +[2026-06-17 22:29:48,065][starVLA.training.train_starvla][INFO] - Step 3679, Loss: {'train/loss': 0.32017946243286133, 'train/grad_norm_pre_clip': 13.640325546264648, 'timing/data': 0.0005925372242927551, 'timing/model': 1.110296887345612, 'learning_rate/qwen_vl_interface': 4.971173289096229e-06, 'learning_rate/action_model': 4.9711732890962286e-05, 'epoch': 1.51}) +[2026-06-17 22:29:51,773][starVLA.training.train_starvla][INFO] - Step 3680, Loss: {'train/loss': 0.14538569748401642, 'train/grad_norm_pre_clip': 21.588796615600586, 'timing/data': 0.0005306769162416458, 'timing/model': 1.049188831821084, 'learning_rate/qwen_vl_interface': 4.969014356718165e-06, 'learning_rate/action_model': 4.969014356718165e-05, 'epoch': 1.51}) +[2026-06-17 22:29:55,367][starVLA.training.train_starvla][INFO] - Step 3681, Loss: {'train/loss': 0.21903739869594574, 'train/grad_norm_pre_clip': 10.970229148864746, 'timing/data': 0.0005167797207832336, 'timing/model': 0.9273401545360684, 'learning_rate/qwen_vl_interface': 4.9668554825887285e-06, 'learning_rate/action_model': 4.966855482588728e-05, 'epoch': 1.52}) +[2026-06-17 22:29:59,351][starVLA.training.train_starvla][INFO] - Step 3682, Loss: {'train/loss': 0.2613365352153778, 'train/grad_norm_pre_clip': 13.982641220092773, 'timing/data': 0.0004929760470986366, 'timing/model': 0.9234315752983093, 'learning_rate/qwen_vl_interface': 4.964696667155455e-06, 'learning_rate/action_model': 4.964696667155454e-05, 'epoch': 1.52}) +[2026-06-17 22:30:03,113][starVLA.training.train_starvla][INFO] - Step 3683, Loss: {'train/loss': 0.2947157025337219, 'train/grad_norm_pre_clip': 16.151018142700195, 'timing/data': 0.0004997439682483673, 'timing/model': 0.9369522398337722, 'learning_rate/qwen_vl_interface': 4.96253791086587e-06, 'learning_rate/action_model': 4.96253791086587e-05, 'epoch': 1.52}) +[2026-06-17 22:30:06,870][starVLA.training.train_starvla][INFO] - Step 3684, Loss: {'train/loss': 0.43848007917404175, 'train/grad_norm_pre_clip': 16.09918212890625, 'timing/data': 0.0005615353584289551, 'timing/model': 0.9369836989790201, 'learning_rate/qwen_vl_interface': 4.960379214167483e-06, 'learning_rate/action_model': 4.960379214167483e-05, 'epoch': 1.52}) +[2026-06-17 22:30:10,641][starVLA.training.train_starvla][INFO] - Step 3685, Loss: {'train/loss': 0.8510409593582153, 'train/grad_norm_pre_clip': 28.265989303588867, 'timing/data': 0.0005339719355106354, 'timing/model': 0.9277054760605097, 'learning_rate/qwen_vl_interface': 4.9582205775077994e-06, 'learning_rate/action_model': 4.958220577507799e-05, 'epoch': 1.52}) +[2026-06-17 22:30:14,392][starVLA.training.train_starvla][INFO] - Step 3686, Loss: {'train/loss': 0.6259056925773621, 'train/grad_norm_pre_clip': 21.27304458618164, 'timing/data': 0.001234416849911213, 'timing/model': 1.106354552321136, 'learning_rate/qwen_vl_interface': 4.956062001334302e-06, 'learning_rate/action_model': 4.956062001334302e-05, 'epoch': 1.52}) +[2026-06-17 22:30:18,128][starVLA.training.train_starvla][INFO] - Step 3687, Loss: {'train/loss': 0.29729709029197693, 'train/grad_norm_pre_clip': 16.660242080688477, 'timing/data': 0.0005039991810917854, 'timing/model': 1.0851284768432379, 'learning_rate/qwen_vl_interface': 4.953903486094469e-06, 'learning_rate/action_model': 4.953903486094469e-05, 'epoch': 1.52}) +[2026-06-17 22:30:21,703][starVLA.training.train_starvla][INFO] - Step 3688, Loss: {'train/loss': 0.5264853835105896, 'train/grad_norm_pre_clip': 16.858856201171875, 'timing/data': 0.000497870147228241, 'timing/model': 0.9271718952804804, 'learning_rate/qwen_vl_interface': 4.951745032235762e-06, 'learning_rate/action_model': 4.951745032235762e-05, 'epoch': 1.52}) +[2026-06-17 22:30:25,493][starVLA.training.train_starvla][INFO] - Step 3689, Loss: {'train/loss': 0.23840592801570892, 'train/grad_norm_pre_clip': 13.587307929992676, 'timing/data': 0.0004760175943374634, 'timing/model': 0.9208156866952777, 'learning_rate/qwen_vl_interface': 4.949586640205631e-06, 'learning_rate/action_model': 4.94958664020563e-05, 'epoch': 1.52}) +[2026-06-17 22:30:29,222][starVLA.training.train_starvla][INFO] - Step 3690, Loss: {'train/loss': 0.4671027660369873, 'train/grad_norm_pre_clip': 24.146944046020508, 'timing/data': 0.00048159249126911163, 'timing/model': 0.91920998506248, 'learning_rate/qwen_vl_interface': 4.9474283104515124e-06, 'learning_rate/action_model': 4.9474283104515126e-05, 'epoch': 1.52}) +[2026-06-17 22:30:32,953][starVLA.training.train_starvla][INFO] - Step 3691, Loss: {'train/loss': 0.36579015851020813, 'train/grad_norm_pre_clip': 18.108747482299805, 'timing/data': 0.000630917027592659, 'timing/model': 0.9211732158437371, 'learning_rate/qwen_vl_interface': 4.945270043420829e-06, 'learning_rate/action_model': 4.945270043420828e-05, 'epoch': 1.52}) +[2026-06-17 22:30:36,724][starVLA.training.train_starvla][INFO] - Step 3692, Loss: {'train/loss': 0.5608919858932495, 'train/grad_norm_pre_clip': 18.961088180541992, 'timing/data': 0.0005266712978482246, 'timing/model': 0.9283607564866543, 'learning_rate/qwen_vl_interface': 4.9431118395609954e-06, 'learning_rate/action_model': 4.9431118395609954e-05, 'epoch': 1.52}) +[2026-06-17 22:30:40,448][starVLA.training.train_starvla][INFO] - Step 3693, Loss: {'train/loss': 0.47017139196395874, 'train/grad_norm_pre_clip': 18.401573181152344, 'timing/data': 0.0008278368040919304, 'timing/model': 1.0832588328048587, 'learning_rate/qwen_vl_interface': 4.940953699319407e-06, 'learning_rate/action_model': 4.940953699319407e-05, 'epoch': 1.52}) +[2026-06-17 22:30:44,105][starVLA.training.train_starvla][INFO] - Step 3694, Loss: {'train/loss': 0.5246106386184692, 'train/grad_norm_pre_clip': 25.67070198059082, 'timing/data': 0.0005132295191287994, 'timing/model': 1.0180429862812161, 'learning_rate/qwen_vl_interface': 4.93879562314345e-06, 'learning_rate/action_model': 4.93879562314345e-05, 'epoch': 1.52}) +[2026-06-17 22:30:47,674][starVLA.training.train_starvla][INFO] - Step 3695, Loss: {'train/loss': 0.3103967308998108, 'train/grad_norm_pre_clip': 20.634929656982422, 'timing/data': 0.0005204742774367332, 'timing/model': 0.9245197009295225, 'learning_rate/qwen_vl_interface': 4.936637611480495e-06, 'learning_rate/action_model': 4.9366376114804945e-05, 'epoch': 1.52}) +[2026-06-17 22:30:51,410][starVLA.training.train_starvla][INFO] - Step 3696, Loss: {'train/loss': 0.4109153747558594, 'train/grad_norm_pre_clip': 14.677155494689941, 'timing/data': 0.0005127135664224625, 'timing/model': 0.9317200556397438, 'learning_rate/qwen_vl_interface': 4.9344796647779e-06, 'learning_rate/action_model': 4.9344796647778995e-05, 'epoch': 1.52}) +[2026-06-17 22:30:55,161][starVLA.training.train_starvla][INFO] - Step 3697, Loss: {'train/loss': 0.46250879764556885, 'train/grad_norm_pre_clip': 28.38727378845215, 'timing/data': 0.0005276110023260117, 'timing/model': 0.927998230792582, 'learning_rate/qwen_vl_interface': 4.9323217834830105e-06, 'learning_rate/action_model': 4.9323217834830105e-05, 'epoch': 1.52}) +[2026-06-17 22:30:58,909][starVLA.training.train_starvla][INFO] - Step 3698, Loss: {'train/loss': 0.20201686024665833, 'train/grad_norm_pre_clip': 10.72284984588623, 'timing/data': 0.000544937327504158, 'timing/model': 0.9209679057821631, 'learning_rate/qwen_vl_interface': 4.930163968043155e-06, 'learning_rate/action_model': 4.930163968043155e-05, 'epoch': 1.52}) +[2026-06-17 22:31:02,657][starVLA.training.train_starvla][INFO] - Step 3699, Loss: {'train/loss': 0.1910725235939026, 'train/grad_norm_pre_clip': 12.077272415161133, 'timing/data': 0.0005055023357272148, 'timing/model': 1.0987704368308187, 'learning_rate/qwen_vl_interface': 4.928006218905656e-06, 'learning_rate/action_model': 4.928006218905656e-05, 'epoch': 1.52}) +[2026-06-17 22:33:02,317][starVLA.training.train_starvla][INFO] - Step 3700, Loss: {'train/loss': 0.3563312888145447, 'train/grad_norm_pre_clip': 15.064661026000977, 'eval/loss': 0.4694427251815796, 'eval/latency_6/loss': 0.4694427251815796, 'eval/demon_fire_precision': 0.9556477345245692, 'eval/demon_fire_recall': 0.9680025856496445, 'eval/demon_fire_f1': 0.9617854849068722, 'eval/demon_fire_rate_model': 0.8720089037284363, 'eval/demon_fire_rate_teacher': 0.860879243183083, 'eval/demon_left_precision': 0.799249530956848, 'eval/demon_left_recall': 0.7473684210526316, 'eval/demon_left_f1': 0.772438803263826, 'eval/demon_left_rate_model': 0.14830272676683362, 'eval/demon_left_rate_teacher': 0.15859766277128548, 'eval/demon_right_precision': 0.7790322580645161, 'eval/demon_right_recall': 0.7419354838709677, 'eval/demon_right_f1': 0.7600314712824547, 'eval/demon_right_rate_model': 0.17250973845297718, 'eval/demon_right_rate_teacher': 0.18113522537562604, 'eval/demon_move_f1': 0.7662351372731404, 'eval/demon_cc_f1': 0.8933854009526504, 'eval/latency_6/demon_fire_precision': 0.9556477345245692, 'eval/latency_6/demon_fire_recall': 0.9680025856496445, 'eval/latency_6/demon_fire_f1': 0.9617854849068722, 'eval/latency_6/demon_fire_rate_model': 0.8720089037284363, 'eval/latency_6/demon_fire_rate_teacher': 0.860879243183083, 'eval/latency_6/demon_left_precision': 0.799249530956848, 'eval/latency_6/demon_left_recall': 0.7473684210526316, 'eval/latency_6/demon_left_f1': 0.772438803263826, 'eval/latency_6/demon_left_rate_model': 0.14830272676683362, 'eval/latency_6/demon_left_rate_teacher': 0.15859766277128548, 'eval/latency_6/demon_right_precision': 0.7790322580645161, 'eval/latency_6/demon_right_recall': 0.7419354838709677, 'eval/latency_6/demon_right_f1': 0.7600314712824547, 'eval/latency_6/demon_right_rate_model': 0.17250973845297718, 'eval/latency_6/demon_right_rate_teacher': 0.18113522537562604, 'eval/latency_6/demon_move_f1': 0.7662351372731404, 'eval/latency_6/demon_cc_f1': 0.8933854009526504, 'timing/data': 0.0005226274952292442, 'timing/model': 1.0337798176333308, 'learning_rate/qwen_vl_interface': 4.9258485365178135e-06, 'learning_rate/action_model': 4.925848536517814e-05, 'epoch': 1.52}) +[2026-06-17 22:33:02,321][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_3700_state +[2026-06-17 22:33:02,325][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 22:33:09,093][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_3700_state/model.safetensors +[2026-06-17 22:33:24,278][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_3700_state/optimizer.bin +[2026-06-17 22:33:24,279][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_3700_state/sampler.bin +[2026-06-17 22:33:24,279][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_3700_state/sampler_1.bin +[2026-06-17 22:33:24,281][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_3700_state/random_states_0.pkl +[2026-06-17 22:33:24,285][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3700 +[2026-06-17 22:33:24,285][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 22:33:24,296][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 22:33:28,101][starVLA.training.train_starvla][INFO] - Step 3701, Loss: {'train/loss': 0.5756781101226807, 'train/grad_norm_pre_clip': 18.632314682006836, 'timing/data': 0.000528678297996521, 'timing/model': 1.137009521946311, 'learning_rate/qwen_vl_interface': 4.923690921326918e-06, 'learning_rate/action_model': 4.9236909213269175e-05, 'epoch': 1.52}) +[2026-06-17 22:33:31,873][starVLA.training.train_starvla][INFO] - Step 3702, Loss: {'train/loss': 0.36175256967544556, 'train/grad_norm_pre_clip': 17.02623748779297, 'timing/data': 0.0006970511749386787, 'timing/model': 1.0486132698133588, 'learning_rate/qwen_vl_interface': 4.921533373780244e-06, 'learning_rate/action_model': 4.921533373780244e-05, 'epoch': 1.52}) +[2026-06-17 22:33:35,454][starVLA.training.train_starvla][INFO] - Step 3703, Loss: {'train/loss': 0.2792694568634033, 'train/grad_norm_pre_clip': 15.28802490234375, 'timing/data': 0.0005081798881292343, 'timing/model': 0.9230754589661956, 'learning_rate/qwen_vl_interface': 4.919375894325057e-06, 'learning_rate/action_model': 4.9193758943250566e-05, 'epoch': 1.52}) +[2026-06-17 22:33:39,241][starVLA.training.train_starvla][INFO] - Step 3704, Loss: {'train/loss': 0.49584200978279114, 'train/grad_norm_pre_clip': 17.786540985107422, 'timing/data': 0.0004927786067128181, 'timing/model': 0.936530607752502, 'learning_rate/qwen_vl_interface': 4.917218483408603e-06, 'learning_rate/action_model': 4.9172184834086035e-05, 'epoch': 1.52}) +[2026-06-17 22:33:43,188][starVLA.training.train_starvla][INFO] - Step 3705, Loss: {'train/loss': 0.38988441228866577, 'train/grad_norm_pre_clip': 14.080002784729004, 'timing/data': 0.0005366262048482895, 'timing/model': 1.051906606182456, 'learning_rate/qwen_vl_interface': 4.915061141478114e-06, 'learning_rate/action_model': 4.915061141478113e-05, 'epoch': 1.53}) +[2026-06-17 22:33:47,056][starVLA.training.train_starvla][INFO] - Step 3706, Loss: {'train/loss': 0.529130756855011, 'train/grad_norm_pre_clip': 14.917317390441895, 'timing/data': 0.0005066888406872749, 'timing/model': 0.9568258766084909, 'learning_rate/qwen_vl_interface': 4.912903868980812e-06, 'learning_rate/action_model': 4.9129038689808115e-05, 'epoch': 1.53}) +[2026-06-17 22:33:50,873][starVLA.training.train_starvla][INFO] - Step 3707, Loss: {'train/loss': 0.38570499420166016, 'train/grad_norm_pre_clip': 18.326139450073242, 'timing/data': 0.0004898030310869217, 'timing/model': 1.1345423627644777, 'learning_rate/qwen_vl_interface': 4.9107466663639005e-06, 'learning_rate/action_model': 4.910746666363901e-05, 'epoch': 1.53}) +[2026-06-17 22:33:54,557][starVLA.training.train_starvla][INFO] - Step 3708, Loss: {'train/loss': 0.3308071792125702, 'train/grad_norm_pre_clip': 13.031070709228516, 'timing/data': 0.0005648229271173477, 'timing/model': 1.0351931536570191, 'learning_rate/qwen_vl_interface': 4.9085895340745705e-06, 'learning_rate/action_model': 4.9085895340745705e-05, 'epoch': 1.53}) +[2026-06-17 22:33:58,233][starVLA.training.train_starvla][INFO] - Step 3709, Loss: {'train/loss': 0.25834131240844727, 'train/grad_norm_pre_clip': 15.26726245880127, 'timing/data': 0.0004936987534165382, 'timing/model': 0.9293635003268719, 'learning_rate/qwen_vl_interface': 4.906432472559996e-06, 'learning_rate/action_model': 4.9064324725599955e-05, 'epoch': 1.53}) +[2026-06-17 22:34:02,039][starVLA.training.train_starvla][INFO] - Step 3710, Loss: {'train/loss': 0.38687705993652344, 'train/grad_norm_pre_clip': 11.9110689163208, 'timing/data': 0.0005615521222352982, 'timing/model': 0.9292144728824496, 'learning_rate/qwen_vl_interface': 4.90427548226734e-06, 'learning_rate/action_model': 4.9042754822673396e-05, 'epoch': 1.53}) +[2026-06-17 22:34:05,971][starVLA.training.train_starvla][INFO] - Step 3711, Loss: {'train/loss': 0.34227240085601807, 'train/grad_norm_pre_clip': 14.593366622924805, 'timing/data': 0.0008847983554005623, 'timing/model': 0.9382113935425878, 'learning_rate/qwen_vl_interface': 4.902118563643751e-06, 'learning_rate/action_model': 4.902118563643751e-05, 'epoch': 1.53}) +[2026-06-17 22:34:09,987][starVLA.training.train_starvla][INFO] - Step 3712, Loss: {'train/loss': 0.2641976475715637, 'train/grad_norm_pre_clip': 9.960774421691895, 'timing/data': 0.0004808129742741585, 'timing/model': 1.149659425020218, 'learning_rate/qwen_vl_interface': 4.899961717136356e-06, 'learning_rate/action_model': 4.899961717136355e-05, 'epoch': 1.53}) +[2026-06-17 22:34:13,769][starVLA.training.train_starvla][INFO] - Step 3713, Loss: {'train/loss': 0.21177911758422852, 'train/grad_norm_pre_clip': 13.020781517028809, 'timing/data': 0.0005410639569163322, 'timing/model': 0.935693709179759, 'learning_rate/qwen_vl_interface': 4.897804943192273e-06, 'learning_rate/action_model': 4.897804943192273e-05, 'epoch': 1.53}) +[2026-06-17 22:34:17,593][starVLA.training.train_starvla][INFO] - Step 3714, Loss: {'train/loss': 0.5307300090789795, 'train/grad_norm_pre_clip': 17.090927124023438, 'timing/data': 0.0005385158583521843, 'timing/model': 1.1039435137063265, 'learning_rate/qwen_vl_interface': 4.895648242258605e-06, 'learning_rate/action_model': 4.895648242258605e-05, 'epoch': 1.53}) +[2026-06-17 22:34:21,403][starVLA.training.train_starvla][INFO] - Step 3715, Loss: {'train/loss': 0.23883412778377533, 'train/grad_norm_pre_clip': 13.069647789001465, 'timing/data': 0.0004914300516247749, 'timing/model': 0.9281402165070176, 'learning_rate/qwen_vl_interface': 4.893491614782438e-06, 'learning_rate/action_model': 4.8934916147824375e-05, 'epoch': 1.53}) +[2026-06-17 22:34:25,167][starVLA.training.train_starvla][INFO] - Step 3716, Loss: {'train/loss': 0.13575708866119385, 'train/grad_norm_pre_clip': 8.966217994689941, 'timing/data': 0.0005075400695204735, 'timing/model': 0.9240002483129501, 'learning_rate/qwen_vl_interface': 4.891335061210841e-06, 'learning_rate/action_model': 4.891335061210841e-05, 'epoch': 1.53}) +[2026-06-17 22:34:28,913][starVLA.training.train_starvla][INFO] - Step 3717, Loss: {'train/loss': 0.5970906615257263, 'train/grad_norm_pre_clip': 18.29618263244629, 'timing/data': 0.0005063805729150772, 'timing/model': 0.9258305486291647, 'learning_rate/qwen_vl_interface': 4.889178581990874e-06, 'learning_rate/action_model': 4.8891785819908734e-05, 'epoch': 1.53}) +[2026-06-17 22:34:32,625][starVLA.training.train_starvla][INFO] - Step 3718, Loss: {'train/loss': 0.24669364094734192, 'train/grad_norm_pre_clip': 12.355777740478516, 'timing/data': 0.0005046594887971878, 'timing/model': 0.9256189996376634, 'learning_rate/qwen_vl_interface': 4.887022177569576e-06, 'learning_rate/action_model': 4.8870221775695754e-05, 'epoch': 1.53}) +[2026-06-17 22:34:36,384][starVLA.training.train_starvla][INFO] - Step 3719, Loss: {'train/loss': 0.21324293315410614, 'train/grad_norm_pre_clip': 10.919258117675781, 'timing/data': 0.0006099594756960869, 'timing/model': 0.9344191774725914, 'learning_rate/qwen_vl_interface': 4.88486584839397e-06, 'learning_rate/action_model': 4.8848658483939695e-05, 'epoch': 1.53}) +[2026-06-17 22:34:40,200][starVLA.training.train_starvla][INFO] - Step 3720, Loss: {'train/loss': 0.5730169415473938, 'train/grad_norm_pre_clip': 21.714380264282227, 'timing/data': 0.0006241528317332268, 'timing/model': 1.1452098712325096, 'learning_rate/qwen_vl_interface': 4.882709594911065e-06, 'learning_rate/action_model': 4.8827095949110655e-05, 'epoch': 1.53}) +[2026-06-17 22:34:43,803][starVLA.training.train_starvla][INFO] - Step 3721, Loss: {'train/loss': 0.356977641582489, 'train/grad_norm_pre_clip': 13.851105690002441, 'timing/data': 0.0008354401215910912, 'timing/model': 0.9241533000022173, 'learning_rate/qwen_vl_interface': 4.880553417567858e-06, 'learning_rate/action_model': 4.880553417567858e-05, 'epoch': 1.53}) +[2026-06-17 22:34:47,549][starVLA.training.train_starvla][INFO] - Step 3722, Loss: {'train/loss': 0.4126461148262024, 'train/grad_norm_pre_clip': 22.085121154785156, 'timing/data': 0.0005591604858636856, 'timing/model': 0.9287657747045159, 'learning_rate/qwen_vl_interface': 4.878397316811324e-06, 'learning_rate/action_model': 4.878397316811324e-05, 'epoch': 1.53}) +[2026-06-17 22:34:51,274][starVLA.training.train_starvla][INFO] - Step 3723, Loss: {'train/loss': 0.29698505997657776, 'train/grad_norm_pre_clip': 13.5269193649292, 'timing/data': 0.0006933547556400299, 'timing/model': 0.9289721641689539, 'learning_rate/qwen_vl_interface': 4.876241293088428e-06, 'learning_rate/action_model': 4.876241293088428e-05, 'epoch': 1.53}) +[2026-06-17 22:34:55,012][starVLA.training.train_starvla][INFO] - Step 3724, Loss: {'train/loss': 0.43103262782096863, 'train/grad_norm_pre_clip': 14.60080623626709, 'timing/data': 0.0004722941666841507, 'timing/model': 0.9243695512413979, 'learning_rate/qwen_vl_interface': 4.874085346846111e-06, 'learning_rate/action_model': 4.87408534684611e-05, 'epoch': 1.53}) +[2026-06-17 22:34:58,799][starVLA.training.train_starvla][INFO] - Step 3725, Loss: {'train/loss': 0.5080468058586121, 'train/grad_norm_pre_clip': 19.9405460357666, 'timing/data': 0.0005101580172777176, 'timing/model': 0.9227613266557455, 'learning_rate/qwen_vl_interface': 4.871929478531309e-06, 'learning_rate/action_model': 4.871929478531309e-05, 'epoch': 1.53}) +[2026-06-17 22:35:02,583][starVLA.training.train_starvla][INFO] - Step 3726, Loss: {'train/loss': 0.3087458610534668, 'train/grad_norm_pre_clip': 13.706260681152344, 'timing/data': 0.000495225191116333, 'timing/model': 0.9320717416703701, 'learning_rate/qwen_vl_interface': 4.869773688590932e-06, 'learning_rate/action_model': 4.8697736885909325e-05, 'epoch': 1.53}) +[2026-06-17 22:35:06,398][starVLA.training.train_starvla][INFO] - Step 3727, Loss: {'train/loss': 0.49546414613723755, 'train/grad_norm_pre_clip': 18.242477416992188, 'timing/data': 0.0007161227986216545, 'timing/model': 1.1268313582986593, 'learning_rate/qwen_vl_interface': 4.867617977471878e-06, 'learning_rate/action_model': 4.867617977471878e-05, 'epoch': 1.53}) +[2026-06-17 22:35:10,060][starVLA.training.train_starvla][INFO] - Step 3728, Loss: {'train/loss': 0.369050532579422, 'train/grad_norm_pre_clip': 21.119884490966797, 'timing/data': 0.0005260808393359184, 'timing/model': 0.9261759556829929, 'learning_rate/qwen_vl_interface': 4.865462345621029e-06, 'learning_rate/action_model': 4.865462345621029e-05, 'epoch': 1.53}) +[2026-06-17 22:35:13,810][starVLA.training.train_starvla][INFO] - Step 3729, Loss: {'train/loss': 0.35953739285469055, 'train/grad_norm_pre_clip': 16.186439514160156, 'timing/data': 0.0005173292011022568, 'timing/model': 0.9311012178659439, 'learning_rate/qwen_vl_interface': 4.86330679348525e-06, 'learning_rate/action_model': 4.863306793485249e-05, 'epoch': 1.53}) +[2026-06-17 22:35:17,607][starVLA.training.train_starvla][INFO] - Step 3730, Loss: {'train/loss': 0.5112195611000061, 'train/grad_norm_pre_clip': 21.739940643310547, 'timing/data': 0.0006540529429912567, 'timing/model': 0.963912958279252, 'learning_rate/qwen_vl_interface': 4.8611513215113894e-06, 'learning_rate/action_model': 4.861151321511389e-05, 'epoch': 1.54}) +[2026-06-17 22:35:21,336][starVLA.training.train_starvla][INFO] - Step 3731, Loss: {'train/loss': 0.39060544967651367, 'train/grad_norm_pre_clip': 14.740595817565918, 'timing/data': 0.0013123126700520515, 'timing/model': 0.9198539415374398, 'learning_rate/qwen_vl_interface': 4.858995930146276e-06, 'learning_rate/action_model': 4.858995930146275e-05, 'epoch': 1.54}) +[2026-06-17 22:35:25,094][starVLA.training.train_starvla][INFO] - Step 3732, Loss: {'train/loss': 0.5010607242584229, 'train/grad_norm_pre_clip': 17.592365264892578, 'timing/data': 0.0008697407320141792, 'timing/model': 0.9176136264577508, 'learning_rate/qwen_vl_interface': 4.856840619836729e-06, 'learning_rate/action_model': 4.856840619836728e-05, 'epoch': 1.54}) +[2026-06-17 22:35:28,832][starVLA.training.train_starvla][INFO] - Step 3733, Loss: {'train/loss': 0.4926770329475403, 'train/grad_norm_pre_clip': 17.87122344970703, 'timing/data': 0.0014996342360973358, 'timing/model': 1.0851409891620278, 'learning_rate/qwen_vl_interface': 4.854685391029543e-06, 'learning_rate/action_model': 4.854685391029543e-05, 'epoch': 1.54}) +[2026-06-17 22:35:32,559][starVLA.training.train_starvla][INFO] - Step 3734, Loss: {'train/loss': 0.1513662487268448, 'train/grad_norm_pre_clip': 19.327238082885742, 'timing/data': 0.0009672753512859344, 'timing/model': 1.0892544770613313, 'learning_rate/qwen_vl_interface': 4.852530244171502e-06, 'learning_rate/action_model': 4.852530244171502e-05, 'epoch': 1.54}) +[2026-06-17 22:35:36,127][starVLA.training.train_starvla][INFO] - Step 3735, Loss: {'train/loss': 0.22308559715747833, 'train/grad_norm_pre_clip': 12.147711753845215, 'timing/data': 0.001191101036965847, 'timing/model': 0.9253537813201547, 'learning_rate/qwen_vl_interface': 4.850375179709368e-06, 'learning_rate/action_model': 4.8503751797093674e-05, 'epoch': 1.54}) +[2026-06-17 22:35:39,873][starVLA.training.train_starvla][INFO] - Step 3736, Loss: {'train/loss': 0.6778589487075806, 'train/grad_norm_pre_clip': 22.548999786376953, 'timing/data': 0.001091289333999157, 'timing/model': 0.92974410392344, 'learning_rate/qwen_vl_interface': 4.84822019808989e-06, 'learning_rate/action_model': 4.84822019808989e-05, 'epoch': 1.54}) +[2026-06-17 22:35:43,615][starVLA.training.train_starvla][INFO] - Step 3737, Loss: {'train/loss': 0.5881619453430176, 'train/grad_norm_pre_clip': 19.409759521484375, 'timing/data': 0.0006364108994603157, 'timing/model': 0.9354001982137561, 'learning_rate/qwen_vl_interface': 4.846065299759799e-06, 'learning_rate/action_model': 4.846065299759799e-05, 'epoch': 1.54}) +[2026-06-17 22:35:47,468][starVLA.training.train_starvla][INFO] - Step 3738, Loss: {'train/loss': 0.357089102268219, 'train/grad_norm_pre_clip': 14.280830383300781, 'timing/data': 0.0016918415203690529, 'timing/model': 0.9283770015463233, 'learning_rate/qwen_vl_interface': 4.843910485165802e-06, 'learning_rate/action_model': 4.843910485165802e-05, 'epoch': 1.54}) +[2026-06-17 22:35:51,257][starVLA.training.train_starvla][INFO] - Step 3739, Loss: {'train/loss': 0.22536151111125946, 'train/grad_norm_pre_clip': 12.74912166595459, 'timing/data': 0.0005075288936495781, 'timing/model': 0.9248647037893534, 'learning_rate/qwen_vl_interface': 4.8417557547546015e-06, 'learning_rate/action_model': 4.8417557547546015e-05, 'epoch': 1.54}) +[2026-06-17 22:35:55,002][starVLA.training.train_starvla][INFO] - Step 3740, Loss: {'train/loss': 0.34057706594467163, 'train/grad_norm_pre_clip': 13.220735549926758, 'timing/data': 0.0005731713026762009, 'timing/model': 1.0932604875415564, 'learning_rate/qwen_vl_interface': 4.839601108972872e-06, 'learning_rate/action_model': 4.839601108972871e-05, 'epoch': 1.54}) +[2026-06-17 22:35:58,698][starVLA.training.train_starvla][INFO] - Step 3741, Loss: {'train/loss': 0.38094308972358704, 'train/grad_norm_pre_clip': 18.323060989379883, 'timing/data': 0.0004743514582514763, 'timing/model': 1.058517008088529, 'learning_rate/qwen_vl_interface': 4.837446548267275e-06, 'learning_rate/action_model': 4.837446548267274e-05, 'epoch': 1.54}) +[2026-06-17 22:36:02,269][starVLA.training.train_starvla][INFO] - Step 3742, Loss: {'train/loss': 0.23920859396457672, 'train/grad_norm_pre_clip': 12.591341972351074, 'timing/data': 0.0005625728517770767, 'timing/model': 0.9310897961258888, 'learning_rate/qwen_vl_interface': 4.835292073084449e-06, 'learning_rate/action_model': 4.8352920730844486e-05, 'epoch': 1.54}) +[2026-06-17 22:36:06,014][starVLA.training.train_starvla][INFO] - Step 3743, Loss: {'train/loss': 0.37360912561416626, 'train/grad_norm_pre_clip': 46.1711540222168, 'timing/data': 0.00048188213258981705, 'timing/model': 0.926255059428513, 'learning_rate/qwen_vl_interface': 4.833137683871025e-06, 'learning_rate/action_model': 4.833137683871025e-05, 'epoch': 1.54}) +[2026-06-17 22:36:09,907][starVLA.training.train_starvla][INFO] - Step 3744, Loss: {'train/loss': 0.3732623755931854, 'train/grad_norm_pre_clip': 13.471112251281738, 'timing/data': 0.0005098506808280945, 'timing/model': 0.9603956947103143, 'learning_rate/qwen_vl_interface': 4.830983381073609e-06, 'learning_rate/action_model': 4.830983381073608e-05, 'epoch': 1.54}) +[2026-06-17 22:36:13,636][starVLA.training.train_starvla][INFO] - Step 3745, Loss: {'train/loss': 0.4597373902797699, 'train/grad_norm_pre_clip': 15.070682525634766, 'timing/data': 0.0004928940907120705, 'timing/model': 0.9245539205148816, 'learning_rate/qwen_vl_interface': 4.828829165138787e-06, 'learning_rate/action_model': 4.828829165138787e-05, 'epoch': 1.54}) +[2026-06-17 22:36:17,375][starVLA.training.train_starvla][INFO] - Step 3746, Loss: {'train/loss': 0.47741565108299255, 'train/grad_norm_pre_clip': 15.349370956420898, 'timing/data': 0.0004764534533023834, 'timing/model': 0.9315228164196014, 'learning_rate/qwen_vl_interface': 4.826675036513132e-06, 'learning_rate/action_model': 4.826675036513132e-05, 'epoch': 1.54}) +[2026-06-17 22:36:21,094][starVLA.training.train_starvla][INFO] - Step 3747, Loss: {'train/loss': 0.388775110244751, 'train/grad_norm_pre_clip': 22.912235260009766, 'timing/data': 0.0005066422745585442, 'timing/model': 1.0839596930891275, 'learning_rate/qwen_vl_interface': 4.824520995643198e-06, 'learning_rate/action_model': 4.8245209956431984e-05, 'epoch': 1.54}) +[2026-06-17 22:36:24,654][starVLA.training.train_starvla][INFO] - Step 3748, Loss: {'train/loss': 0.3798586428165436, 'train/grad_norm_pre_clip': 14.602852821350098, 'timing/data': 0.0005388930439949036, 'timing/model': 0.9234955804422498, 'learning_rate/qwen_vl_interface': 4.82236704297552e-06, 'learning_rate/action_model': 4.82236704297552e-05, 'epoch': 1.54}) +[2026-06-17 22:36:28,439][starVLA.training.train_starvla][INFO] - Step 3749, Loss: {'train/loss': 0.7685432434082031, 'train/grad_norm_pre_clip': 24.320974349975586, 'timing/data': 0.0009684571996331215, 'timing/model': 0.9740846483036876, 'learning_rate/qwen_vl_interface': 4.820213178956613e-06, 'learning_rate/action_model': 4.820213178956612e-05, 'epoch': 1.54}) +[2026-06-17 22:36:32,111][starVLA.training.train_starvla][INFO] - Step 3750, Loss: {'train/loss': 0.5313965678215027, 'train/grad_norm_pre_clip': 20.579195022583008, 'timing/data': 0.0005200179293751717, 'timing/model': 0.928210437297821, 'learning_rate/qwen_vl_interface': 4.818059404032977e-06, 'learning_rate/action_model': 4.818059404032977e-05, 'epoch': 1.54}) +[2026-06-17 22:36:35,836][starVLA.training.train_starvla][INFO] - Step 3751, Loss: {'train/loss': 0.1667758822441101, 'train/grad_norm_pre_clip': 15.334234237670898, 'timing/data': 0.0005013961344957352, 'timing/model': 0.9255256876349449, 'learning_rate/qwen_vl_interface': 4.815905718651092e-06, 'learning_rate/action_model': 4.815905718651092e-05, 'epoch': 1.54}) +[2026-06-17 22:36:39,603][starVLA.training.train_starvla][INFO] - Step 3752, Loss: {'train/loss': 0.30684125423431396, 'train/grad_norm_pre_clip': 12.812989234924316, 'timing/data': 0.0004949597641825676, 'timing/model': 0.9323776615783572, 'learning_rate/qwen_vl_interface': 4.813752123257419e-06, 'learning_rate/action_model': 4.813752123257418e-05, 'epoch': 1.54}) +[2026-06-17 22:36:43,338][starVLA.training.train_starvla][INFO] - Step 3753, Loss: {'train/loss': 0.6972337961196899, 'train/grad_norm_pre_clip': 22.09318733215332, 'timing/data': 0.000546332448720932, 'timing/model': 1.0945418030023575, 'learning_rate/qwen_vl_interface': 4.811598618298398e-06, 'learning_rate/action_model': 4.811598618298398e-05, 'epoch': 1.54}) +[2026-06-17 22:36:47,030][starVLA.training.train_starvla][INFO] - Step 3754, Loss: {'train/loss': 0.195577934384346, 'train/grad_norm_pre_clip': 13.308309555053711, 'timing/data': 0.0004959329962730408, 'timing/model': 1.0518195079639554, 'learning_rate/qwen_vl_interface': 4.809445204220457e-06, 'learning_rate/action_model': 4.809445204220457e-05, 'epoch': 1.55}) +[2026-06-17 22:36:50,595][starVLA.training.train_starvla][INFO] - Step 3755, Loss: {'train/loss': 0.30072277784347534, 'train/grad_norm_pre_clip': 14.10778522491455, 'timing/data': 0.0005233399569988251, 'timing/model': 0.9269950296729803, 'learning_rate/qwen_vl_interface': 4.8072918814699985e-06, 'learning_rate/action_model': 4.8072918814699986e-05, 'epoch': 1.55}) +[2026-06-17 22:36:54,332][starVLA.training.train_starvla][INFO] - Step 3756, Loss: {'train/loss': 0.492090106010437, 'train/grad_norm_pre_clip': 17.08755874633789, 'timing/data': 0.0009369151666760445, 'timing/model': 0.9207978863269091, 'learning_rate/qwen_vl_interface': 4.80513865049341e-06, 'learning_rate/action_model': 4.80513865049341e-05, 'epoch': 1.55}) +[2026-06-17 22:36:58,069][starVLA.training.train_starvla][INFO] - Step 3757, Loss: {'train/loss': 0.6613365411758423, 'train/grad_norm_pre_clip': 19.30288314819336, 'timing/data': 0.0005357945337891579, 'timing/model': 0.9235701132565737, 'learning_rate/qwen_vl_interface': 4.802985511737057e-06, 'learning_rate/action_model': 4.802985511737057e-05, 'epoch': 1.55}) +[2026-06-17 22:37:01,823][starVLA.training.train_starvla][INFO] - Step 3758, Loss: {'train/loss': 0.24398548901081085, 'train/grad_norm_pre_clip': 13.465740203857422, 'timing/data': 0.0005294755101203918, 'timing/model': 0.9259400814771652, 'learning_rate/qwen_vl_interface': 4.80083246564729e-06, 'learning_rate/action_model': 4.8008324656472894e-05, 'epoch': 1.55}) +[2026-06-17 22:37:05,570][starVLA.training.train_starvla][INFO] - Step 3759, Loss: {'train/loss': 0.47684407234191895, 'train/grad_norm_pre_clip': 23.585668563842773, 'timing/data': 0.0005038734525442123, 'timing/model': 0.9292818941175938, 'learning_rate/qwen_vl_interface': 4.798679512670436e-06, 'learning_rate/action_model': 4.798679512670435e-05, 'epoch': 1.55}) +[2026-06-17 22:37:09,326][starVLA.training.train_starvla][INFO] - Step 3760, Loss: {'train/loss': 0.48302412033081055, 'train/grad_norm_pre_clip': 21.299358367919922, 'timing/data': 0.0006962819024920464, 'timing/model': 1.097298501059413, 'learning_rate/qwen_vl_interface': 4.796526653252802e-06, 'learning_rate/action_model': 4.796526653252802e-05, 'epoch': 1.55}) +[2026-06-17 22:37:12,917][starVLA.training.train_starvla][INFO] - Step 3761, Loss: {'train/loss': 0.3763713538646698, 'train/grad_norm_pre_clip': 15.326656341552734, 'timing/data': 0.0006020572036504745, 'timing/model': 0.9341053031384945, 'learning_rate/qwen_vl_interface': 4.794373887840683e-06, 'learning_rate/action_model': 4.794373887840682e-05, 'epoch': 1.55}) +[2026-06-17 22:37:16,667][starVLA.training.train_starvla][INFO] - Step 3762, Loss: {'train/loss': 0.29852139949798584, 'train/grad_norm_pre_clip': 15.396283149719238, 'timing/data': 0.0004871394485235214, 'timing/model': 0.930639524012804, 'learning_rate/qwen_vl_interface': 4.792221216880347e-06, 'learning_rate/action_model': 4.792221216880347e-05, 'epoch': 1.55}) +[2026-06-17 22:37:20,401][starVLA.training.train_starvla][INFO] - Step 3763, Loss: {'train/loss': 0.3976697623729706, 'train/grad_norm_pre_clip': 17.808181762695312, 'timing/data': 0.00048699695616960526, 'timing/model': 0.9199169110506773, 'learning_rate/qwen_vl_interface': 4.790068640818046e-06, 'learning_rate/action_model': 4.790068640818045e-05, 'epoch': 1.55}) +[2026-06-17 22:37:24,150][starVLA.training.train_starvla][INFO] - Step 3764, Loss: {'train/loss': 0.18960736691951752, 'train/grad_norm_pre_clip': 10.537110328674316, 'timing/data': 0.0005328003317117691, 'timing/model': 0.9250758998095989, 'learning_rate/qwen_vl_interface': 4.787916160100008e-06, 'learning_rate/action_model': 4.787916160100008e-05, 'epoch': 1.55}) +[2026-06-17 22:37:27,883][starVLA.training.train_starvla][INFO] - Step 3765, Loss: {'train/loss': 0.3096023499965668, 'train/grad_norm_pre_clip': 16.81307601928711, 'timing/data': 0.0004832511767745018, 'timing/model': 0.9243030697107315, 'learning_rate/qwen_vl_interface': 4.78576377517245e-06, 'learning_rate/action_model': 4.7857637751724504e-05, 'epoch': 1.55}) +[2026-06-17 22:37:31,617][starVLA.training.train_starvla][INFO] - Step 3766, Loss: {'train/loss': 0.2499910593032837, 'train/grad_norm_pre_clip': 11.305524826049805, 'timing/data': 0.0004967283457517624, 'timing/model': 0.9227195056155324, 'learning_rate/qwen_vl_interface': 4.783611486481561e-06, 'learning_rate/action_model': 4.7836114864815604e-05, 'epoch': 1.55}) +[2026-06-17 22:37:35,344][starVLA.training.train_starvla][INFO] - Step 3767, Loss: {'train/loss': 0.4804723858833313, 'train/grad_norm_pre_clip': 15.921183586120605, 'timing/data': 0.0008501429110765457, 'timing/model': 1.081739105284214, 'learning_rate/qwen_vl_interface': 4.781459294473511e-06, 'learning_rate/action_model': 4.781459294473511e-05, 'epoch': 1.55}) +[2026-06-17 22:37:38,920][starVLA.training.train_starvla][INFO] - Step 3768, Loss: {'train/loss': 0.24523548781871796, 'train/grad_norm_pre_clip': 12.112756729125977, 'timing/data': 0.0005571451038122177, 'timing/model': 0.926070699468255, 'learning_rate/qwen_vl_interface': 4.779307199594454e-06, 'learning_rate/action_model': 4.779307199594454e-05, 'epoch': 1.55}) +[2026-06-17 22:37:42,670][starVLA.training.train_starvla][INFO] - Step 3769, Loss: {'train/loss': 0.4877112805843353, 'train/grad_norm_pre_clip': 20.82228660583496, 'timing/data': 0.0009324066340923309, 'timing/model': 0.9210508493706584, 'learning_rate/qwen_vl_interface': 4.777155202290522e-06, 'learning_rate/action_model': 4.777155202290521e-05, 'epoch': 1.55}) +[2026-06-17 22:37:46,449][starVLA.training.train_starvla][INFO] - Step 3770, Loss: {'train/loss': 0.4651782512664795, 'train/grad_norm_pre_clip': 30.945743560791016, 'timing/data': 0.0008323034271597862, 'timing/model': 0.9222950702533126, 'learning_rate/qwen_vl_interface': 4.775003303007824e-06, 'learning_rate/action_model': 4.775003303007824e-05, 'epoch': 1.55}) +[2026-06-17 22:37:50,191][starVLA.training.train_starvla][INFO] - Step 3771, Loss: {'train/loss': 0.2630595266819, 'train/grad_norm_pre_clip': 12.036609649658203, 'timing/data': 0.0005544489249587059, 'timing/model': 0.9240624448284507, 'learning_rate/qwen_vl_interface': 4.772851502192453e-06, 'learning_rate/action_model': 4.772851502192452e-05, 'epoch': 1.55}) +[2026-06-17 22:37:53,922][starVLA.training.train_starvla][INFO] - Step 3772, Loss: {'train/loss': 0.4575098156929016, 'train/grad_norm_pre_clip': 16.193815231323242, 'timing/data': 0.00046862661838531494, 'timing/model': 0.9157490432262421, 'learning_rate/qwen_vl_interface': 4.770699800290478e-06, 'learning_rate/action_model': 4.770699800290478e-05, 'epoch': 1.55}) +[2026-06-17 22:37:57,645][starVLA.training.train_starvla][INFO] - Step 3773, Loss: {'train/loss': 0.7087868452072144, 'train/grad_norm_pre_clip': 27.87650489807129, 'timing/data': 0.0004965830594301224, 'timing/model': 0.9223202243447304, 'learning_rate/qwen_vl_interface': 4.768548197747953e-06, 'learning_rate/action_model': 4.768548197747953e-05, 'epoch': 1.55}) +[2026-06-17 22:38:01,382][starVLA.training.train_starvla][INFO] - Step 3774, Loss: {'train/loss': 0.3646427392959595, 'train/grad_norm_pre_clip': 17.55063819885254, 'timing/data': 0.0005374178290367126, 'timing/model': 1.1042456859722733, 'learning_rate/qwen_vl_interface': 4.766396695010903e-06, 'learning_rate/action_model': 4.766396695010903e-05, 'epoch': 1.55}) +[2026-06-17 22:38:04,952][starVLA.training.train_starvla][INFO] - Step 3775, Loss: {'train/loss': 0.21660858392715454, 'train/grad_norm_pre_clip': 12.247502326965332, 'timing/data': 0.000865597277879715, 'timing/model': 0.9241606956347823, 'learning_rate/qwen_vl_interface': 4.764245292525337e-06, 'learning_rate/action_model': 4.7642452925253376e-05, 'epoch': 1.55}) +[2026-06-17 22:38:08,689][starVLA.training.train_starvla][INFO] - Step 3776, Loss: {'train/loss': 0.11568265408277512, 'train/grad_norm_pre_clip': 8.706313133239746, 'timing/data': 0.0004952801391482353, 'timing/model': 0.9255649261176586, 'learning_rate/qwen_vl_interface': 4.762093990737247e-06, 'learning_rate/action_model': 4.762093990737246e-05, 'epoch': 1.55}) +[2026-06-17 22:38:12,451][starVLA.training.train_starvla][INFO] - Step 3777, Loss: {'train/loss': 0.2974953055381775, 'train/grad_norm_pre_clip': 11.0301513671875, 'timing/data': 0.0005423789843916893, 'timing/model': 0.9336136532947421, 'learning_rate/qwen_vl_interface': 4.759942790092596e-06, 'learning_rate/action_model': 4.759942790092596e-05, 'epoch': 1.55}) +[2026-06-17 22:38:16,177][starVLA.training.train_starvla][INFO] - Step 3778, Loss: {'train/loss': 0.2884121835231781, 'train/grad_norm_pre_clip': 17.253259658813477, 'timing/data': 0.0004909439012408257, 'timing/model': 0.9226166354492307, 'learning_rate/qwen_vl_interface': 4.757791691037332e-06, 'learning_rate/action_model': 4.757791691037332e-05, 'epoch': 1.56}) +[2026-06-17 22:38:19,906][starVLA.training.train_starvla][INFO] - Step 3779, Loss: {'train/loss': 0.37984776496887207, 'train/grad_norm_pre_clip': 16.190595626831055, 'timing/data': 0.0004975050687789917, 'timing/model': 0.9243335984647274, 'learning_rate/qwen_vl_interface': 4.755640694017381e-06, 'learning_rate/action_model': 4.755640694017381e-05, 'epoch': 1.56}) +[2026-06-17 22:38:23,641][starVLA.training.train_starvla][INFO] - Step 3780, Loss: {'train/loss': 0.6275853514671326, 'train/grad_norm_pre_clip': 24.378292083740234, 'timing/data': 0.0005076536908745766, 'timing/model': 1.0986899007111788, 'learning_rate/qwen_vl_interface': 4.753489799478647e-06, 'learning_rate/action_model': 4.7534897994786464e-05, 'epoch': 1.56}) +[2026-06-17 22:38:27,362][starVLA.training.train_starvla][INFO] - Step 3781, Loss: {'train/loss': 0.1898314356803894, 'train/grad_norm_pre_clip': 11.414901733398438, 'timing/data': 0.0004987828433513641, 'timing/model': 1.0830623665824533, 'learning_rate/qwen_vl_interface': 4.7513390078670105e-06, 'learning_rate/action_model': 4.75133900786701e-05, 'epoch': 1.56}) +[2026-06-17 22:38:30,913][starVLA.training.train_starvla][INFO] - Step 3782, Loss: {'train/loss': 0.37618181109428406, 'train/grad_norm_pre_clip': 14.0042724609375, 'timing/data': 0.0011839596554636955, 'timing/model': 0.9236221769824624, 'learning_rate/qwen_vl_interface': 4.749188319628333e-06, 'learning_rate/action_model': 4.749188319628333e-05, 'epoch': 1.56}) +[2026-06-17 22:38:34,643][starVLA.training.train_starvla][INFO] - Step 3783, Loss: {'train/loss': 0.24515178799629211, 'train/grad_norm_pre_clip': 11.314672470092773, 'timing/data': 0.0010372800752520561, 'timing/model': 0.9216085262596607, 'learning_rate/qwen_vl_interface': 4.747037735208457e-06, 'learning_rate/action_model': 4.7470377352084575e-05, 'epoch': 1.56}) +[2026-06-17 22:38:38,371][starVLA.training.train_starvla][INFO] - Step 3784, Loss: {'train/loss': 0.5646446347236633, 'train/grad_norm_pre_clip': 23.27646827697754, 'timing/data': 0.0004946086555719376, 'timing/model': 0.9212694633752108, 'learning_rate/qwen_vl_interface': 4.7448872550532006e-06, 'learning_rate/action_model': 4.7448872550532e-05, 'epoch': 1.56}) +[2026-06-17 22:38:42,089][starVLA.training.train_starvla][INFO] - Step 3785, Loss: {'train/loss': 0.3965902030467987, 'train/grad_norm_pre_clip': 12.047563552856445, 'timing/data': 0.00048327259719371796, 'timing/model': 0.9224583217874169, 'learning_rate/qwen_vl_interface': 4.74273687960836e-06, 'learning_rate/action_model': 4.74273687960836e-05, 'epoch': 1.56}) +[2026-06-17 22:38:45,842][starVLA.training.train_starvla][INFO] - Step 3786, Loss: {'train/loss': 0.2745603621006012, 'train/grad_norm_pre_clip': 14.126304626464844, 'timing/data': 0.0004778429865837097, 'timing/model': 0.9561613397672772, 'learning_rate/qwen_vl_interface': 4.7405866093197075e-06, 'learning_rate/action_model': 4.740586609319707e-05, 'epoch': 1.56}) +[2026-06-17 22:38:49,579][starVLA.training.train_starvla][INFO] - Step 3787, Loss: {'train/loss': 0.13506640493869781, 'train/grad_norm_pre_clip': 18.89043617248535, 'timing/data': 0.0011896640062332153, 'timing/model': 1.0869980659335852, 'learning_rate/qwen_vl_interface': 4.738436444633002e-06, 'learning_rate/action_model': 4.738436444633001e-05, 'epoch': 1.56}) +[2026-06-17 22:38:53,271][starVLA.training.train_starvla][INFO] - Step 3788, Loss: {'train/loss': 0.39491280913352966, 'train/grad_norm_pre_clip': 18.140859603881836, 'timing/data': 0.0004981495440006256, 'timing/model': 1.0532639399170876, 'learning_rate/qwen_vl_interface': 4.73628638599397e-06, 'learning_rate/action_model': 4.73628638599397e-05, 'epoch': 1.56}) +[2026-06-17 22:38:56,842][starVLA.training.train_starvla][INFO] - Step 3789, Loss: {'train/loss': 0.4186878800392151, 'train/grad_norm_pre_clip': 16.62413787841797, 'timing/data': 0.0011013532057404518, 'timing/model': 0.9235060550272465, 'learning_rate/qwen_vl_interface': 4.734136433848323e-06, 'learning_rate/action_model': 4.7341364338483224e-05, 'epoch': 1.56}) +[2026-06-17 22:39:00,573][starVLA.training.train_starvla][INFO] - Step 3790, Loss: {'train/loss': 0.3522791862487793, 'train/grad_norm_pre_clip': 18.66779136657715, 'timing/data': 0.0005150558426976204, 'timing/model': 0.9256701180711389, 'learning_rate/qwen_vl_interface': 4.731986588641748e-06, 'learning_rate/action_model': 4.7319865886417484e-05, 'epoch': 1.56}) +[2026-06-17 22:39:04,363][starVLA.training.train_starvla][INFO] - Step 3791, Loss: {'train/loss': 0.41980352997779846, 'train/grad_norm_pre_clip': 19.96640968322754, 'timing/data': 0.0010746307671070099, 'timing/model': 0.9226106330752373, 'learning_rate/qwen_vl_interface': 4.729836850819912e-06, 'learning_rate/action_model': 4.729836850819912e-05, 'epoch': 1.56}) +[2026-06-17 22:39:08,059][starVLA.training.train_starvla][INFO] - Step 3792, Loss: {'train/loss': 0.49481797218322754, 'train/grad_norm_pre_clip': 18.884916305541992, 'timing/data': 0.0005204500630497932, 'timing/model': 0.9252472119405866, 'learning_rate/qwen_vl_interface': 4.727687220828458e-06, 'learning_rate/action_model': 4.727687220828458e-05, 'epoch': 1.56}) +[2026-06-17 22:39:11,774][starVLA.training.train_starvla][INFO] - Step 3793, Loss: {'train/loss': 0.3622763156890869, 'train/grad_norm_pre_clip': 17.01400375366211, 'timing/data': 0.0005124611780047417, 'timing/model': 0.9253952307626605, 'learning_rate/qwen_vl_interface': 4.7255376991130005e-06, 'learning_rate/action_model': 4.7255376991130005e-05, 'epoch': 1.56}) +[2026-06-17 22:39:15,492][starVLA.training.train_starvla][INFO] - Step 3794, Loss: {'train/loss': 0.19412270188331604, 'train/grad_norm_pre_clip': 12.803213119506836, 'timing/data': 0.0004972284659743309, 'timing/model': 1.0827250825241208, 'learning_rate/qwen_vl_interface': 4.723388286119147e-06, 'learning_rate/action_model': 4.723388286119146e-05, 'epoch': 1.56}) +[2026-06-17 22:39:19,180][starVLA.training.train_starvla][INFO] - Step 3795, Loss: {'train/loss': 0.2527092695236206, 'train/grad_norm_pre_clip': 21.16674041748047, 'timing/data': 0.0009060585871338844, 'timing/model': 1.0278669437393546, 'learning_rate/qwen_vl_interface': 4.721238982292467e-06, 'learning_rate/action_model': 4.721238982292466e-05, 'epoch': 1.56}) +[2026-06-17 22:39:22,934][starVLA.training.train_starvla][INFO] - Step 3796, Loss: {'train/loss': 0.44379788637161255, 'train/grad_norm_pre_clip': 19.52619171142578, 'timing/data': 0.00051165372133255, 'timing/model': 0.9234927613288164, 'learning_rate/qwen_vl_interface': 4.719089788078514e-06, 'learning_rate/action_model': 4.719089788078514e-05, 'epoch': 1.56}) +[2026-06-17 22:39:26,629][starVLA.training.train_starvla][INFO] - Step 3797, Loss: {'train/loss': 0.20510518550872803, 'train/grad_norm_pre_clip': 13.530622482299805, 'timing/data': 0.0006762221455574036, 'timing/model': 0.919577076099813, 'learning_rate/qwen_vl_interface': 4.716940703922818e-06, 'learning_rate/action_model': 4.716940703922818e-05, 'epoch': 1.56}) +[2026-06-17 22:39:30,344][starVLA.training.train_starvla][INFO] - Step 3798, Loss: {'train/loss': 0.16919675469398499, 'train/grad_norm_pre_clip': 12.478038787841797, 'timing/data': 0.00048604607582092285, 'timing/model': 0.9202281888574362, 'learning_rate/qwen_vl_interface': 4.714791730270889e-06, 'learning_rate/action_model': 4.71479173027089e-05, 'epoch': 1.56}) +[2026-06-17 22:39:34,068][starVLA.training.train_starvla][INFO] - Step 3799, Loss: {'train/loss': 0.5681609511375427, 'train/grad_norm_pre_clip': 23.765193939208984, 'timing/data': 0.0004863925278186798, 'timing/model': 0.9239527257159352, 'learning_rate/qwen_vl_interface': 4.712642867568213e-06, 'learning_rate/action_model': 4.712642867568212e-05, 'epoch': 1.56}) +[2026-06-17 22:41:33,037][starVLA.training.train_starvla][INFO] - Step 3800, Loss: {'train/loss': 0.3813539743423462, 'train/grad_norm_pre_clip': 18.668954849243164, 'eval/loss': 0.47059711813926697, 'eval/latency_6/loss': 0.47059711813926697, 'eval/demon_fire_precision': 0.9477635283077885, 'eval/demon_fire_recall': 0.9793148028442146, 'eval/demon_fire_f1': 0.9632808774439676, 'eval/demon_fire_rate_model': 0.8895381190873678, 'eval/demon_fire_rate_teacher': 0.860879243183083, 'eval/demon_left_precision': 0.8947368421052632, 'eval/demon_left_recall': 0.5964912280701754, 'eval/demon_left_f1': 0.7157894736842105, 'eval/demon_left_rate_model': 0.10573177518085698, 'eval/demon_left_rate_teacher': 0.15859766277128548, 'eval/demon_right_precision': 0.8169491525423729, 'eval/demon_right_recall': 0.7403993855606759, 'eval/demon_right_f1': 0.7767929089443997, 'eval/demon_right_rate_model': 0.16416249304396216, 'eval/demon_right_rate_teacher': 0.18113522537562604, 'eval/demon_move_f1': 0.7462911913143051, 'eval/demon_cc_f1': 0.8859977563390475, 'eval/latency_6/demon_fire_precision': 0.9477635283077885, 'eval/latency_6/demon_fire_recall': 0.9793148028442146, 'eval/latency_6/demon_fire_f1': 0.9632808774439676, 'eval/latency_6/demon_fire_rate_model': 0.8895381190873678, 'eval/latency_6/demon_fire_rate_teacher': 0.860879243183083, 'eval/latency_6/demon_left_precision': 0.8947368421052632, 'eval/latency_6/demon_left_recall': 0.5964912280701754, 'eval/latency_6/demon_left_f1': 0.7157894736842105, 'eval/latency_6/demon_left_rate_model': 0.10573177518085698, 'eval/latency_6/demon_left_rate_teacher': 0.15859766277128548, 'eval/latency_6/demon_right_precision': 0.8169491525423729, 'eval/latency_6/demon_right_recall': 0.7403993855606759, 'eval/latency_6/demon_right_f1': 0.7767929089443997, 'eval/latency_6/demon_right_rate_model': 0.16416249304396216, 'eval/latency_6/demon_right_rate_teacher': 0.18113522537562604, 'eval/latency_6/demon_move_f1': 0.7462911913143051, 'eval/latency_6/demon_cc_f1': 0.8859977563390475, 'timing/data': 0.0005111033096909523, 'timing/model': 1.0901075406000018, 'learning_rate/qwen_vl_interface': 4.7104941162602436e-06, 'learning_rate/action_model': 4.710494116260244e-05, 'epoch': 1.56}) +[2026-06-17 22:41:33,041][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_3800_state +[2026-06-17 22:41:33,046][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 22:41:39,556][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_3800_state/model.safetensors +[2026-06-17 22:41:54,488][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_3800_state/optimizer.bin +[2026-06-17 22:41:54,491][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_3800_state/sampler.bin +[2026-06-17 22:41:54,491][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_3800_state/sampler_1.bin +[2026-06-17 22:41:54,494][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_3800_state/random_states_0.pkl +[2026-06-17 22:41:54,498][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3800 +[2026-06-17 22:41:54,498][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 22:41:54,509][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 22:41:58,339][starVLA.training.train_starvla][INFO] - Step 3801, Loss: {'train/loss': 0.5355597138404846, 'train/grad_norm_pre_clip': 19.022897720336914, 'timing/data': 0.0005834680050611496, 'timing/model': 0.926663726568222, 'learning_rate/qwen_vl_interface': 4.708345476792428e-06, 'learning_rate/action_model': 4.7083454767924276e-05, 'epoch': 1.56}) +[2026-06-17 22:42:02,098][starVLA.training.train_starvla][INFO] - Step 3802, Loss: {'train/loss': 0.2745433449745178, 'train/grad_norm_pre_clip': 23.640317916870117, 'timing/data': 0.0005967002362012863, 'timing/model': 1.113603862002492, 'learning_rate/qwen_vl_interface': 4.706196949610178e-06, 'learning_rate/action_model': 4.7061969496101776e-05, 'epoch': 1.57}) +[2026-06-17 22:42:05,860][starVLA.training.train_starvla][INFO] - Step 3803, Loss: {'train/loss': 0.5234358310699463, 'train/grad_norm_pre_clip': 21.676130294799805, 'timing/data': 0.0005841366946697235, 'timing/model': 1.1042587524279952, 'learning_rate/qwen_vl_interface': 4.7040485351588845e-06, 'learning_rate/action_model': 4.7040485351588845e-05, 'epoch': 1.57}) +[2026-06-17 22:42:09,461][starVLA.training.train_starvla][INFO] - Step 3804, Loss: {'train/loss': 0.22548408806324005, 'train/grad_norm_pre_clip': 13.851639747619629, 'timing/data': 0.0014256276190280914, 'timing/model': 0.9319878593087196, 'learning_rate/qwen_vl_interface': 4.701900233883916e-06, 'learning_rate/action_model': 4.7019002338839156e-05, 'epoch': 1.57}) +[2026-06-17 22:42:13,271][starVLA.training.train_starvla][INFO] - Step 3805, Loss: {'train/loss': 0.26749661564826965, 'train/grad_norm_pre_clip': 13.560432434082031, 'timing/data': 0.000499519519507885, 'timing/model': 0.9250229513272643, 'learning_rate/qwen_vl_interface': 4.69975204623062e-06, 'learning_rate/action_model': 4.69975204623062e-05, 'epoch': 1.57}) +[2026-06-17 22:42:17,025][starVLA.training.train_starvla][INFO] - Step 3806, Loss: {'train/loss': 0.25232356786727905, 'train/grad_norm_pre_clip': 36.40083694458008, 'timing/data': 0.0004893923178315163, 'timing/model': 0.9217649307101965, 'learning_rate/qwen_vl_interface': 4.697603972644319e-06, 'learning_rate/action_model': 4.697603972644318e-05, 'epoch': 1.57}) +[2026-06-17 22:42:20,760][starVLA.training.train_starvla][INFO] - Step 3807, Loss: {'train/loss': 0.25537002086639404, 'train/grad_norm_pre_clip': 15.661870956420898, 'timing/data': 0.0004894351586699486, 'timing/model': 0.9258002694696188, 'learning_rate/qwen_vl_interface': 4.695456013570306e-06, 'learning_rate/action_model': 4.695456013570306e-05, 'epoch': 1.57}) +[2026-06-17 22:42:24,502][starVLA.training.train_starvla][INFO] - Step 3808, Loss: {'train/loss': 0.638871967792511, 'train/grad_norm_pre_clip': 28.43515968322754, 'timing/data': 0.0004638694226741791, 'timing/model': 0.9323445362970233, 'learning_rate/qwen_vl_interface': 4.693308169453858e-06, 'learning_rate/action_model': 4.693308169453858e-05, 'epoch': 1.57}) +[2026-06-17 22:42:28,275][starVLA.training.train_starvla][INFO] - Step 3809, Loss: {'train/loss': 0.31279456615448, 'train/grad_norm_pre_clip': 23.096792221069336, 'timing/data': 0.00047498662024736404, 'timing/model': 1.1152319079264998, 'learning_rate/qwen_vl_interface': 4.691160440740226e-06, 'learning_rate/action_model': 4.691160440740226e-05, 'epoch': 1.57}) +[2026-06-17 22:42:31,897][starVLA.training.train_starvla][INFO] - Step 3810, Loss: {'train/loss': 0.5231496691703796, 'train/grad_norm_pre_clip': 22.04389190673828, 'timing/data': 0.0005088932812213898, 'timing/model': 0.9289292404428124, 'learning_rate/qwen_vl_interface': 4.689012827874637e-06, 'learning_rate/action_model': 4.689012827874637e-05, 'epoch': 1.57}) +[2026-06-17 22:42:35,689][starVLA.training.train_starvla][INFO] - Step 3811, Loss: {'train/loss': 0.46777787804603577, 'train/grad_norm_pre_clip': 47.441551208496094, 'timing/data': 0.0005776956677436829, 'timing/model': 0.9326187968254089, 'learning_rate/qwen_vl_interface': 4.6868653313022905e-06, 'learning_rate/action_model': 4.68686533130229e-05, 'epoch': 1.57}) +[2026-06-17 22:42:39,415][starVLA.training.train_starvla][INFO] - Step 3812, Loss: {'train/loss': 0.2365296483039856, 'train/grad_norm_pre_clip': 17.910829544067383, 'timing/data': 0.0005097817629575729, 'timing/model': 0.9282266069203615, 'learning_rate/qwen_vl_interface': 4.684717951468368e-06, 'learning_rate/action_model': 4.6847179514683684e-05, 'epoch': 1.57}) +[2026-06-17 22:42:43,173][starVLA.training.train_starvla][INFO] - Step 3813, Loss: {'train/loss': 0.28779342770576477, 'train/grad_norm_pre_clip': 20.04576301574707, 'timing/data': 0.0010048812255263329, 'timing/model': 0.921858049929142, 'learning_rate/qwen_vl_interface': 4.682570688818024e-06, 'learning_rate/action_model': 4.682570688818024e-05, 'epoch': 1.57}) +[2026-06-17 22:42:46,931][starVLA.training.train_starvla][INFO] - Step 3814, Loss: {'train/loss': 0.25486573576927185, 'train/grad_norm_pre_clip': 15.233105659484863, 'timing/data': 0.0005044648423790932, 'timing/model': 0.927388153038919, 'learning_rate/qwen_vl_interface': 4.680423543796385e-06, 'learning_rate/action_model': 4.680423543796385e-05, 'epoch': 1.57}) +[2026-06-17 22:42:50,672][starVLA.training.train_starvla][INFO] - Step 3815, Loss: {'train/loss': 0.27536386251449585, 'train/grad_norm_pre_clip': 16.57056999206543, 'timing/data': 0.0005284016951918602, 'timing/model': 1.09141815174371, 'learning_rate/qwen_vl_interface': 4.678276516848558e-06, 'learning_rate/action_model': 4.678276516848557e-05, 'epoch': 1.57}) +[2026-06-17 22:42:54,615][starVLA.training.train_starvla][INFO] - Step 3816, Loss: {'train/loss': 0.41256481409072876, 'train/grad_norm_pre_clip': 20.64964485168457, 'timing/data': 0.0004886612296104431, 'timing/model': 1.2652508355677128, 'learning_rate/qwen_vl_interface': 4.676129608419625e-06, 'learning_rate/action_model': 4.676129608419625e-05, 'epoch': 1.57}) +[2026-06-17 22:42:58,430][starVLA.training.train_starvla][INFO] - Step 3817, Loss: {'train/loss': 0.377670556306839, 'train/grad_norm_pre_clip': 21.638349533081055, 'timing/data': 0.0009869914501905441, 'timing/model': 0.9248456824570894, 'learning_rate/qwen_vl_interface': 4.6739828189546415e-06, 'learning_rate/action_model': 4.6739828189546417e-05, 'epoch': 1.57}) +[2026-06-17 22:43:02,427][starVLA.training.train_starvla][INFO] - Step 3818, Loss: {'train/loss': 0.3994631767272949, 'train/grad_norm_pre_clip': 16.975971221923828, 'timing/data': 0.0005094613879919052, 'timing/model': 0.9287228425964713, 'learning_rate/qwen_vl_interface': 4.671836148898641e-06, 'learning_rate/action_model': 4.6718361488986404e-05, 'epoch': 1.57}) +[2026-06-17 22:43:06,213][starVLA.training.train_starvla][INFO] - Step 3819, Loss: {'train/loss': 0.38128677010536194, 'train/grad_norm_pre_clip': 15.740691184997559, 'timing/data': 0.0005699098110198975, 'timing/model': 0.9279176946729422, 'learning_rate/qwen_vl_interface': 4.6696895986966254e-06, 'learning_rate/action_model': 4.669689598696625e-05, 'epoch': 1.57}) +[2026-06-17 22:43:09,974][starVLA.training.train_starvla][INFO] - Step 3820, Loss: {'train/loss': 0.25792592763900757, 'train/grad_norm_pre_clip': 13.923654556274414, 'timing/data': 0.0005102800205349922, 'timing/model': 0.9279077900573611, 'learning_rate/qwen_vl_interface': 4.667543168793583e-06, 'learning_rate/action_model': 4.667543168793583e-05, 'epoch': 1.57}) +[2026-06-17 22:43:13,737][starVLA.training.train_starvla][INFO] - Step 3821, Loss: {'train/loss': 0.33103835582733154, 'train/grad_norm_pre_clip': 16.12114906311035, 'timing/data': 0.0008728178218007088, 'timing/model': 0.9324519392102957, 'learning_rate/qwen_vl_interface': 4.665396859634468e-06, 'learning_rate/action_model': 4.665396859634468e-05, 'epoch': 1.57}) +[2026-06-17 22:43:17,485][starVLA.training.train_starvla][INFO] - Step 3822, Loss: {'train/loss': 0.3730083405971527, 'train/grad_norm_pre_clip': 20.02301597595215, 'timing/data': 0.0005833636969327927, 'timing/model': 1.097318772226572, 'learning_rate/qwen_vl_interface': 4.663250671664212e-06, 'learning_rate/action_model': 4.663250671664211e-05, 'epoch': 1.57}) +[2026-06-17 22:43:21,070][starVLA.training.train_starvla][INFO] - Step 3823, Loss: {'train/loss': 0.6275787949562073, 'train/grad_norm_pre_clip': 27.386001586914062, 'timing/data': 0.001338031142950058, 'timing/model': 0.9303529169410467, 'learning_rate/qwen_vl_interface': 4.6611046053277245e-06, 'learning_rate/action_model': 4.661104605327724e-05, 'epoch': 1.57}) +[2026-06-17 22:43:24,845][starVLA.training.train_starvla][INFO] - Step 3824, Loss: {'train/loss': 0.594836950302124, 'train/grad_norm_pre_clip': 19.953283309936523, 'timing/data': 0.0004920670762658119, 'timing/model': 0.9220370631664991, 'learning_rate/qwen_vl_interface': 4.658958661069886e-06, 'learning_rate/action_model': 4.6589586610698856e-05, 'epoch': 1.57}) +[2026-06-17 22:43:28,588][starVLA.training.train_starvla][INFO] - Step 3825, Loss: {'train/loss': 0.328175812959671, 'train/grad_norm_pre_clip': 24.159543991088867, 'timing/data': 0.0004910910502076149, 'timing/model': 0.9209461323916912, 'learning_rate/qwen_vl_interface': 4.6568128393355535e-06, 'learning_rate/action_model': 4.656812839335553e-05, 'epoch': 1.57}) +[2026-06-17 22:43:32,677][starVLA.training.train_starvla][INFO] - Step 3826, Loss: {'train/loss': 0.4122619032859802, 'train/grad_norm_pre_clip': 17.226682662963867, 'timing/data': 0.0008863601833581924, 'timing/model': 0.9287461675703526, 'learning_rate/qwen_vl_interface': 4.654667140569555e-06, 'learning_rate/action_model': 4.6546671405695544e-05, 'epoch': 1.57}) +[2026-06-17 22:43:36,494][starVLA.training.train_starvla][INFO] - Step 3827, Loss: {'train/loss': 0.4068205952644348, 'train/grad_norm_pre_clip': 18.0026912689209, 'timing/data': 0.0004899939522147179, 'timing/model': 0.9502113414928317, 'learning_rate/qwen_vl_interface': 4.652521565216701e-06, 'learning_rate/action_model': 4.652521565216701e-05, 'epoch': 1.58}) +[2026-06-17 22:43:40,306][starVLA.training.train_starvla][INFO] - Step 3828, Loss: {'train/loss': 0.6715698838233948, 'train/grad_norm_pre_clip': 19.039167404174805, 'timing/data': 0.0004972200840711594, 'timing/model': 0.925646754913032, 'learning_rate/qwen_vl_interface': 4.65037611372177e-06, 'learning_rate/action_model': 4.650376113721769e-05, 'epoch': 1.58}) +[2026-06-17 22:43:44,103][starVLA.training.train_starvla][INFO] - Step 3829, Loss: {'train/loss': 0.28211405873298645, 'train/grad_norm_pre_clip': 18.26418685913086, 'timing/data': 0.0005396539345383644, 'timing/model': 1.1295268284156919, 'learning_rate/qwen_vl_interface': 4.648230786529514e-06, 'learning_rate/action_model': 4.648230786529513e-05, 'epoch': 1.58}) +[2026-06-17 22:43:47,707][starVLA.training.train_starvla][INFO] - Step 3830, Loss: {'train/loss': 0.7470848560333252, 'train/grad_norm_pre_clip': 21.787294387817383, 'timing/data': 0.0005737049505114555, 'timing/model': 0.9267938127741218, 'learning_rate/qwen_vl_interface': 4.646085584084664e-06, 'learning_rate/action_model': 4.6460855840846637e-05, 'epoch': 1.58}) +[2026-06-17 22:43:51,501][starVLA.training.train_starvla][INFO] - Step 3831, Loss: {'train/loss': 0.23817914724349976, 'train/grad_norm_pre_clip': 11.545604705810547, 'timing/data': 0.0009380215778946877, 'timing/model': 0.9283606596291065, 'learning_rate/qwen_vl_interface': 4.643940506831923e-06, 'learning_rate/action_model': 4.6439405068319226e-05, 'epoch': 1.58}) +[2026-06-17 22:43:55,472][starVLA.training.train_starvla][INFO] - Step 3832, Loss: {'train/loss': 0.6248915195465088, 'train/grad_norm_pre_clip': 19.715797424316406, 'timing/data': 0.0005824314430356026, 'timing/model': 0.9312084969133139, 'learning_rate/qwen_vl_interface': 4.641795555215967e-06, 'learning_rate/action_model': 4.6417955552159676e-05, 'epoch': 1.58}) +[2026-06-17 22:43:59,258][starVLA.training.train_starvla][INFO] - Step 3833, Loss: {'train/loss': 0.46596595644950867, 'train/grad_norm_pre_clip': 42.28327941894531, 'timing/data': 0.0013059405609965324, 'timing/model': 0.9286413434892893, 'learning_rate/qwen_vl_interface': 4.639650729681445e-06, 'learning_rate/action_model': 4.6396507296814454e-05, 'epoch': 1.58}) +[2026-06-17 22:44:03,360][starVLA.training.train_starvla][INFO] - Step 3834, Loss: {'train/loss': 0.3339649438858032, 'train/grad_norm_pre_clip': 14.828472137451172, 'timing/data': 0.0005048131570219994, 'timing/model': 0.9455773523077369, 'learning_rate/qwen_vl_interface': 4.637506030672987e-06, 'learning_rate/action_model': 4.637506030672987e-05, 'epoch': 1.58}) +[2026-06-17 22:44:07,172][starVLA.training.train_starvla][INFO] - Step 3835, Loss: {'train/loss': 0.3999105393886566, 'train/grad_norm_pre_clip': 19.817659378051758, 'timing/data': 0.0008727628737688065, 'timing/model': 1.1355895521119237, 'learning_rate/qwen_vl_interface': 4.635361458635186e-06, 'learning_rate/action_model': 4.6353614586351864e-05, 'epoch': 1.58}) +[2026-06-17 22:44:11,571][starVLA.training.train_starvla][INFO] - Step 3836, Loss: {'train/loss': 0.2637918293476105, 'train/grad_norm_pre_clip': 13.456072807312012, 'timing/data': 0.0004890123382210732, 'timing/model': 1.324418194591999, 'learning_rate/qwen_vl_interface': 4.633217014012616e-06, 'learning_rate/action_model': 4.633217014012616e-05, 'epoch': 1.58}) +[2026-06-17 22:44:15,371][starVLA.training.train_starvla][INFO] - Step 3837, Loss: {'train/loss': 0.27566206455230713, 'train/grad_norm_pre_clip': 14.226622581481934, 'timing/data': 0.0012179184705018997, 'timing/model': 0.9240661468356848, 'learning_rate/qwen_vl_interface': 4.631072697249822e-06, 'learning_rate/action_model': 4.631072697249822e-05, 'epoch': 1.58}) +[2026-06-17 22:44:19,143][starVLA.training.train_starvla][INFO] - Step 3838, Loss: {'train/loss': 0.4209790825843811, 'train/grad_norm_pre_clip': 21.99518585205078, 'timing/data': 0.0009903935715556145, 'timing/model': 0.928363399580121, 'learning_rate/qwen_vl_interface': 4.628928508791325e-06, 'learning_rate/action_model': 4.628928508791324e-05, 'epoch': 1.58}) +[2026-06-17 22:44:22,909][starVLA.training.train_starvla][INFO] - Step 3839, Loss: {'train/loss': 0.4553246796131134, 'train/grad_norm_pre_clip': 19.330801010131836, 'timing/data': 0.0005351798608899117, 'timing/model': 0.9261595923453569, 'learning_rate/qwen_vl_interface': 4.626784449081617e-06, 'learning_rate/action_model': 4.626784449081617e-05, 'epoch': 1.58}) +[2026-06-17 22:44:26,667][starVLA.training.train_starvla][INFO] - Step 3840, Loss: {'train/loss': 0.37730491161346436, 'train/grad_norm_pre_clip': 18.86048698425293, 'timing/data': 0.0012436416000127792, 'timing/model': 0.9367606043815613, 'learning_rate/qwen_vl_interface': 4.62464051856516e-06, 'learning_rate/action_model': 4.6246405185651596e-05, 'epoch': 1.58}) +[2026-06-17 22:44:30,913][starVLA.training.train_starvla][INFO] - Step 3841, Loss: {'train/loss': 0.2598276734352112, 'train/grad_norm_pre_clip': 13.826258659362793, 'timing/data': 0.0005088578909635544, 'timing/model': 1.3421361222863197, 'learning_rate/qwen_vl_interface': 4.622496717686401e-06, 'learning_rate/action_model': 4.6224967176864005e-05, 'epoch': 1.58}) +[2026-06-17 22:44:34,757][starVLA.training.train_starvla][INFO] - Step 3842, Loss: {'train/loss': 0.22117093205451965, 'train/grad_norm_pre_clip': 15.637712478637695, 'timing/data': 0.0004997653886675835, 'timing/model': 1.1227823188528419, 'learning_rate/qwen_vl_interface': 4.620353046889745e-06, 'learning_rate/action_model': 4.620353046889745e-05, 'epoch': 1.58}) +[2026-06-17 22:44:38,540][starVLA.training.train_starvla][INFO] - Step 3843, Loss: {'train/loss': 0.5223242044448853, 'train/grad_norm_pre_clip': 16.74470329284668, 'timing/data': 0.00047815684229135513, 'timing/model': 1.0677938545122743, 'learning_rate/qwen_vl_interface': 4.6182095066195805e-06, 'learning_rate/action_model': 4.61820950661958e-05, 'epoch': 1.58}) +[2026-06-17 22:44:42,125][starVLA.training.train_starvla][INFO] - Step 3844, Loss: {'train/loss': 0.28359296917915344, 'train/grad_norm_pre_clip': 17.956527709960938, 'timing/data': 0.0010276082903146744, 'timing/model': 0.9232357535511255, 'learning_rate/qwen_vl_interface': 4.616066097320264e-06, 'learning_rate/action_model': 4.6160660973202636e-05, 'epoch': 1.58}) +[2026-06-17 22:44:46,299][starVLA.training.train_starvla][INFO] - Step 3845, Loss: {'train/loss': 0.22059190273284912, 'train/grad_norm_pre_clip': 13.29991340637207, 'timing/data': 0.0005252715200185776, 'timing/model': 0.9287069113925099, 'learning_rate/qwen_vl_interface': 4.613922819436129e-06, 'learning_rate/action_model': 4.613922819436128e-05, 'epoch': 1.58}) +[2026-06-17 22:44:50,145][starVLA.training.train_starvla][INFO] - Step 3846, Loss: {'train/loss': 0.203817218542099, 'train/grad_norm_pre_clip': 9.895610809326172, 'timing/data': 0.0009874114766716957, 'timing/model': 0.9230545004829764, 'learning_rate/qwen_vl_interface': 4.611779673411478e-06, 'learning_rate/action_model': 4.6117796734114785e-05, 'epoch': 1.58}) +[2026-06-17 22:44:53,843][starVLA.training.train_starvla][INFO] - Step 3847, Loss: {'train/loss': 0.3313656449317932, 'train/grad_norm_pre_clip': 107.938232421875, 'timing/data': 0.0005626752972602844, 'timing/model': 0.9379989514127374, 'learning_rate/qwen_vl_interface': 4.609636659690588e-06, 'learning_rate/action_model': 4.609636659690587e-05, 'epoch': 1.58}) +[2026-06-17 22:44:57,697][starVLA.training.train_starvla][INFO] - Step 3848, Loss: {'train/loss': 0.3566630780696869, 'train/grad_norm_pre_clip': 18.230480194091797, 'timing/data': 0.0008955132216215134, 'timing/model': 0.9989218711853027, 'learning_rate/qwen_vl_interface': 4.607493778717705e-06, 'learning_rate/action_model': 4.6074937787177055e-05, 'epoch': 1.58}) +[2026-06-17 22:45:01,476][starVLA.training.train_starvla][INFO] - Step 3849, Loss: {'train/loss': 0.3556603193283081, 'train/grad_norm_pre_clip': 15.092462539672852, 'timing/data': 0.00047994963824748993, 'timing/model': 1.098047073930502, 'learning_rate/qwen_vl_interface': 4.605351030937054e-06, 'learning_rate/action_model': 4.605351030937055e-05, 'epoch': 1.58}) +[2026-06-17 22:45:05,066][starVLA.training.train_starvla][INFO] - Step 3850, Loss: {'train/loss': 0.16351743042469025, 'train/grad_norm_pre_clip': 11.303326606750488, 'timing/data': 0.0005025872960686684, 'timing/model': 0.9304855419322848, 'learning_rate/qwen_vl_interface': 4.60320841679283e-06, 'learning_rate/action_model': 4.60320841679283e-05, 'epoch': 1.58}) +[2026-06-17 22:45:08,964][starVLA.training.train_starvla][INFO] - Step 3851, Loss: {'train/loss': 0.48146575689315796, 'train/grad_norm_pre_clip': 24.188251495361328, 'timing/data': 0.000504782423377037, 'timing/model': 0.9243152439594269, 'learning_rate/qwen_vl_interface': 4.6010659367291945e-06, 'learning_rate/action_model': 4.601065936729194e-05, 'epoch': 1.59}) +[2026-06-17 22:45:12,714][starVLA.training.train_starvla][INFO] - Step 3852, Loss: {'train/loss': 0.4340420663356781, 'train/grad_norm_pre_clip': 17.120849609375, 'timing/data': 0.0005362387746572495, 'timing/model': 0.9341911477968097, 'learning_rate/qwen_vl_interface': 4.598923591190289e-06, 'learning_rate/action_model': 4.598923591190289e-05, 'epoch': 1.59}) +[2026-06-17 22:45:16,542][starVLA.training.train_starvla][INFO] - Step 3853, Loss: {'train/loss': 0.19618895649909973, 'train/grad_norm_pre_clip': 13.657844543457031, 'timing/data': 0.0005222521722316742, 'timing/model': 0.9320408031344414, 'learning_rate/qwen_vl_interface': 4.596781380620225e-06, 'learning_rate/action_model': 4.5967813806202245e-05, 'epoch': 1.59}) +[2026-06-17 22:45:20,296][starVLA.training.train_starvla][INFO] - Step 3854, Loss: {'train/loss': 0.24094177782535553, 'train/grad_norm_pre_clip': 14.306059837341309, 'timing/data': 0.0006904499605298042, 'timing/model': 0.9351250939071178, 'learning_rate/qwen_vl_interface': 4.5946393054630845e-06, 'learning_rate/action_model': 4.5946393054630845e-05, 'epoch': 1.59}) +[2026-06-17 22:45:24,088][starVLA.training.train_starvla][INFO] - Step 3855, Loss: {'train/loss': 0.28939130902290344, 'train/grad_norm_pre_clip': 17.71518898010254, 'timing/data': 0.0005427151918411255, 'timing/model': 1.1107805790379643, 'learning_rate/qwen_vl_interface': 4.592497366162919e-06, 'learning_rate/action_model': 4.592497366162918e-05, 'epoch': 1.59}) +[2026-06-17 22:45:28,065][starVLA.training.train_starvla][INFO] - Step 3856, Loss: {'train/loss': 0.2457946240901947, 'train/grad_norm_pre_clip': 12.918474197387695, 'timing/data': 0.0004736408591270447, 'timing/model': 1.0678603844717145, 'learning_rate/qwen_vl_interface': 4.590355563163758e-06, 'learning_rate/action_model': 4.5903555631637574e-05, 'epoch': 1.59}) +[2026-06-17 22:45:31,722][starVLA.training.train_starvla][INFO] - Step 3857, Loss: {'train/loss': 0.20212145149707794, 'train/grad_norm_pre_clip': 22.11358642578125, 'timing/data': 0.000543537549674511, 'timing/model': 0.9308783942833543, 'learning_rate/qwen_vl_interface': 4.588213896909598e-06, 'learning_rate/action_model': 4.588213896909598e-05, 'epoch': 1.59}) +[2026-06-17 22:45:35,499][starVLA.training.train_starvla][INFO] - Step 3858, Loss: {'train/loss': 0.5237399935722351, 'train/grad_norm_pre_clip': 18.290973663330078, 'timing/data': 0.0012599602341651917, 'timing/model': 0.927812872454524, 'learning_rate/qwen_vl_interface': 4.58607236784441e-06, 'learning_rate/action_model': 4.58607236784441e-05, 'epoch': 1.59}) +[2026-06-17 22:45:39,338][starVLA.training.train_starvla][INFO] - Step 3859, Loss: {'train/loss': 0.35785380005836487, 'train/grad_norm_pre_clip': 18.227096557617188, 'timing/data': 0.0005068602040410042, 'timing/model': 0.9277325300499797, 'learning_rate/qwen_vl_interface': 4.583930976412133e-06, 'learning_rate/action_model': 4.583930976412133e-05, 'epoch': 1.59}) +[2026-06-17 22:45:43,106][starVLA.training.train_starvla][INFO] - Step 3860, Loss: {'train/loss': 0.3718978762626648, 'train/grad_norm_pre_clip': 16.45819664001465, 'timing/data': 0.0005843937397003174, 'timing/model': 0.9361468106508255, 'learning_rate/qwen_vl_interface': 4.5817897230566834e-06, 'learning_rate/action_model': 4.581789723056683e-05, 'epoch': 1.59}) +[2026-06-17 22:45:46,869][starVLA.training.train_starvla][INFO] - Step 3861, Loss: {'train/loss': 0.27172979712486267, 'train/grad_norm_pre_clip': 15.499232292175293, 'timing/data': 0.0008239196613430977, 'timing/model': 0.9353889022022486, 'learning_rate/qwen_vl_interface': 4.579648608221943e-06, 'learning_rate/action_model': 4.579648608221943e-05, 'epoch': 1.59}) +[2026-06-17 22:45:50,603][starVLA.training.train_starvla][INFO] - Step 3862, Loss: {'train/loss': 0.20334242284297943, 'train/grad_norm_pre_clip': 12.808773040771484, 'timing/data': 0.00046585220843553543, 'timing/model': 1.0950362021103501, 'learning_rate/qwen_vl_interface': 4.577507632351766e-06, 'learning_rate/action_model': 4.5775076323517664e-05, 'epoch': 1.59}) +[2026-06-17 22:45:54,328][starVLA.training.train_starvla][INFO] - Step 3863, Loss: {'train/loss': 0.24704459309577942, 'train/grad_norm_pre_clip': 14.667710304260254, 'timing/data': 0.0005472376942634583, 'timing/model': 1.073489934206009, 'learning_rate/qwen_vl_interface': 4.575366795889981e-06, 'learning_rate/action_model': 4.575366795889981e-05, 'epoch': 1.59}) +[2026-06-17 22:45:57,896][starVLA.training.train_starvla][INFO] - Step 3864, Loss: {'train/loss': 0.25121456384658813, 'train/grad_norm_pre_clip': 15.277888298034668, 'timing/data': 0.0005397778004407883, 'timing/model': 0.9279307760298252, 'learning_rate/qwen_vl_interface': 4.573226099280386e-06, 'learning_rate/action_model': 4.573226099280386e-05, 'epoch': 1.59}) +[2026-06-17 22:46:01,630][starVLA.training.train_starvla][INFO] - Step 3865, Loss: {'train/loss': 0.22624176740646362, 'train/grad_norm_pre_clip': 14.474953651428223, 'timing/data': 0.0005233464762568474, 'timing/model': 0.9283095095306635, 'learning_rate/qwen_vl_interface': 4.571085542966748e-06, 'learning_rate/action_model': 4.5710855429667485e-05, 'epoch': 1.59}) +[2026-06-17 22:46:05,369][starVLA.training.train_starvla][INFO] - Step 3866, Loss: {'train/loss': 0.24290762841701508, 'train/grad_norm_pre_clip': 14.453038215637207, 'timing/data': 0.0008441340178251266, 'timing/model': 0.9277484519407153, 'learning_rate/qwen_vl_interface': 4.568945127392807e-06, 'learning_rate/action_model': 4.5689451273928066e-05, 'epoch': 1.59}) +[2026-06-17 22:46:09,114][starVLA.training.train_starvla][INFO] - Step 3867, Loss: {'train/loss': 0.16091585159301758, 'train/grad_norm_pre_clip': 25.876192092895508, 'timing/data': 0.000512935221195221, 'timing/model': 0.9250897355377674, 'learning_rate/qwen_vl_interface': 4.566804853002275e-06, 'learning_rate/action_model': 4.566804853002275e-05, 'epoch': 1.59}) +[2026-06-17 22:46:12,882][starVLA.training.train_starvla][INFO] - Step 3868, Loss: {'train/loss': 0.2811483144760132, 'train/grad_norm_pre_clip': 14.142542839050293, 'timing/data': 0.0005248542875051498, 'timing/model': 0.9281137892976403, 'learning_rate/qwen_vl_interface': 4.564664720238833e-06, 'learning_rate/action_model': 4.564664720238833e-05, 'epoch': 1.59}) +[2026-06-17 22:46:16,604][starVLA.training.train_starvla][INFO] - Step 3869, Loss: {'train/loss': 0.23242422938346863, 'train/grad_norm_pre_clip': 12.219681739807129, 'timing/data': 0.0004880335181951523, 'timing/model': 1.0885684788227081, 'learning_rate/qwen_vl_interface': 4.5625247295461314e-06, 'learning_rate/action_model': 4.562524729546131e-05, 'epoch': 1.59}) +[2026-06-17 22:46:20,314][starVLA.training.train_starvla][INFO] - Step 3870, Loss: {'train/loss': 0.09063607454299927, 'train/grad_norm_pre_clip': 7.1302080154418945, 'timing/data': 0.0007953736931085587, 'timing/model': 1.0226104706525803, 'learning_rate/qwen_vl_interface': 4.560384881367791e-06, 'learning_rate/action_model': 4.5603848813677904e-05, 'epoch': 1.59}) +[2026-06-17 22:46:23,890][starVLA.training.train_starvla][INFO] - Step 3871, Loss: {'train/loss': 0.4321233630180359, 'train/grad_norm_pre_clip': 24.382823944091797, 'timing/data': 0.000521492213010788, 'timing/model': 0.9345091823488474, 'learning_rate/qwen_vl_interface': 4.5582451761474085e-06, 'learning_rate/action_model': 4.558245176147408e-05, 'epoch': 1.59}) +[2026-06-17 22:46:27,632][starVLA.training.train_starvla][INFO] - Step 3872, Loss: {'train/loss': 0.3736913800239563, 'train/grad_norm_pre_clip': 93.58953094482422, 'timing/data': 0.000748906284570694, 'timing/model': 0.9282841151580215, 'learning_rate/qwen_vl_interface': 4.556105614328545e-06, 'learning_rate/action_model': 4.5561056143285455e-05, 'epoch': 1.59}) +[2026-06-17 22:46:31,439][starVLA.training.train_starvla][INFO] - Step 3873, Loss: {'train/loss': 0.35387948155403137, 'train/grad_norm_pre_clip': 21.014925003051758, 'timing/data': 0.000500745140016079, 'timing/model': 0.9326006704941392, 'learning_rate/qwen_vl_interface': 4.553966196354734e-06, 'learning_rate/action_model': 4.553966196354734e-05, 'epoch': 1.59}) +[2026-06-17 22:46:35,187][starVLA.training.train_starvla][INFO] - Step 3874, Loss: {'train/loss': 0.9258375763893127, 'train/grad_norm_pre_clip': 51.33614730834961, 'timing/data': 0.0013083964586257935, 'timing/model': 0.9246560288593173, 'learning_rate/qwen_vl_interface': 4.55182692266948e-06, 'learning_rate/action_model': 4.551826922669479e-05, 'epoch': 1.59}) +[2026-06-17 22:46:38,925][starVLA.training.train_starvla][INFO] - Step 3875, Loss: {'train/loss': 0.4170862138271332, 'train/grad_norm_pre_clip': 23.835895538330078, 'timing/data': 0.001041666604578495, 'timing/model': 1.0898677157238126, 'learning_rate/qwen_vl_interface': 4.549687793716257e-06, 'learning_rate/action_model': 4.549687793716257e-05, 'epoch': 1.6}) +[2026-06-17 22:46:42,612][starVLA.training.train_starvla][INFO] - Step 3876, Loss: {'train/loss': 0.8052436709403992, 'train/grad_norm_pre_clip': 25.511459350585938, 'timing/data': 0.000502418726682663, 'timing/model': 1.0329263405874372, 'learning_rate/qwen_vl_interface': 4.547548809938507e-06, 'learning_rate/action_model': 4.5475488099385075e-05, 'epoch': 1.6}) +[2026-06-17 22:46:46,178][starVLA.training.train_starvla][INFO] - Step 3877, Loss: {'train/loss': 0.3937872052192688, 'train/grad_norm_pre_clip': 22.43211555480957, 'timing/data': 0.0004950221627950668, 'timing/model': 0.9244203036651015, 'learning_rate/qwen_vl_interface': 4.545409971779644e-06, 'learning_rate/action_model': 4.545409971779644e-05, 'epoch': 1.6}) +[2026-06-17 22:46:49,943][starVLA.training.train_starvla][INFO] - Step 3878, Loss: {'train/loss': 0.4466378092765808, 'train/grad_norm_pre_clip': 36.11844253540039, 'timing/data': 0.0005178041756153107, 'timing/model': 0.9360563298687339, 'learning_rate/qwen_vl_interface': 4.543271279683052e-06, 'learning_rate/action_model': 4.543271279683053e-05, 'epoch': 1.6}) +[2026-06-17 22:46:53,709][starVLA.training.train_starvla][INFO] - Step 3879, Loss: {'train/loss': 0.20125868916511536, 'train/grad_norm_pre_clip': 14.155927658081055, 'timing/data': 0.0005056848749518394, 'timing/model': 0.9378283694386482, 'learning_rate/qwen_vl_interface': 4.541132734092086e-06, 'learning_rate/action_model': 4.541132734092086e-05, 'epoch': 1.6}) +[2026-06-17 22:46:57,467][starVLA.training.train_starvla][INFO] - Step 3880, Loss: {'train/loss': 0.49998360872268677, 'train/grad_norm_pre_clip': 17.645530700683594, 'timing/data': 0.0004970841109752655, 'timing/model': 0.9287487640976906, 'learning_rate/qwen_vl_interface': 4.538994335450066e-06, 'learning_rate/action_model': 4.5389943354500664e-05, 'epoch': 1.6}) +[2026-06-17 22:47:01,153][starVLA.training.train_starvla][INFO] - Step 3881, Loss: {'train/loss': 0.3756103217601776, 'train/grad_norm_pre_clip': 31.778560638427734, 'timing/data': 0.000548848882317543, 'timing/model': 0.9271161612123251, 'learning_rate/qwen_vl_interface': 4.5368560842002836e-06, 'learning_rate/action_model': 4.5368560842002834e-05, 'epoch': 1.6}) +[2026-06-17 22:47:04,888][starVLA.training.train_starvla][INFO] - Step 3882, Loss: {'train/loss': 0.4453657567501068, 'train/grad_norm_pre_clip': 18.857532501220703, 'timing/data': 0.0004939818754792213, 'timing/model': 1.091974114999175, 'learning_rate/qwen_vl_interface': 4.534717980786005e-06, 'learning_rate/action_model': 4.5347179807860046e-05, 'epoch': 1.6}) +[2026-06-17 22:47:08,479][starVLA.training.train_starvla][INFO] - Step 3883, Loss: {'train/loss': 0.5002629160881042, 'train/grad_norm_pre_clip': 22.35215950012207, 'timing/data': 0.0005058282986283302, 'timing/model': 0.9287390494719148, 'learning_rate/qwen_vl_interface': 4.532580025650457e-06, 'learning_rate/action_model': 4.5325800256504574e-05, 'epoch': 1.6}) +[2026-06-17 22:47:12,223][starVLA.training.train_starvla][INFO] - Step 3884, Loss: {'train/loss': 0.33833047747612, 'train/grad_norm_pre_clip': 15.773468017578125, 'timing/data': 0.0007340079173445702, 'timing/model': 0.9299930147826672, 'learning_rate/qwen_vl_interface': 4.530442219236842e-06, 'learning_rate/action_model': 4.530442219236842e-05, 'epoch': 1.6}) +[2026-06-17 22:47:15,971][starVLA.training.train_starvla][INFO] - Step 3885, Loss: {'train/loss': 0.21739302575588226, 'train/grad_norm_pre_clip': 13.56212329864502, 'timing/data': 0.0005511688068509102, 'timing/model': 0.9256033375859261, 'learning_rate/qwen_vl_interface': 4.5283045619883285e-06, 'learning_rate/action_model': 4.5283045619883285e-05, 'epoch': 1.6}) +[2026-06-17 22:47:19,674][starVLA.training.train_starvla][INFO] - Step 3886, Loss: {'train/loss': 0.4430701732635498, 'train/grad_norm_pre_clip': 19.56785011291504, 'timing/data': 0.0005267169326543808, 'timing/model': 0.9222954325377941, 'learning_rate/qwen_vl_interface': 4.526167054348057e-06, 'learning_rate/action_model': 4.526167054348057e-05, 'epoch': 1.6}) +[2026-06-17 22:47:23,417][starVLA.training.train_starvla][INFO] - Step 3887, Loss: {'train/loss': 0.8933414220809937, 'train/grad_norm_pre_clip': 28.873788833618164, 'timing/data': 0.0005226172506809235, 'timing/model': 0.9254036778584123, 'learning_rate/qwen_vl_interface': 4.524029696759135e-06, 'learning_rate/action_model': 4.5240296967591354e-05, 'epoch': 1.6}) +[2026-06-17 22:47:27,156][starVLA.training.train_starvla][INFO] - Step 3888, Loss: {'train/loss': 0.5007359385490417, 'train/grad_norm_pre_clip': 21.836809158325195, 'timing/data': 0.0005507851019501686, 'timing/model': 1.0883706947788596, 'learning_rate/qwen_vl_interface': 4.5218924896646354e-06, 'learning_rate/action_model': 4.521892489664635e-05, 'epoch': 1.6}) +[2026-06-17 22:47:30,730][starVLA.training.train_starvla][INFO] - Step 3889, Loss: {'train/loss': 0.3675503134727478, 'train/grad_norm_pre_clip': 21.711811065673828, 'timing/data': 0.000497208908200264, 'timing/model': 0.9232457671314478, 'learning_rate/qwen_vl_interface': 4.519755433507609e-06, 'learning_rate/action_model': 4.5197554335076086e-05, 'epoch': 1.6}) +[2026-06-17 22:47:34,451][starVLA.training.train_starvla][INFO] - Step 3890, Loss: {'train/loss': 0.576716959476471, 'train/grad_norm_pre_clip': 27.272945404052734, 'timing/data': 0.0008673928678035736, 'timing/model': 0.9187306938692927, 'learning_rate/qwen_vl_interface': 4.517618528731066e-06, 'learning_rate/action_model': 4.517618528731065e-05, 'epoch': 1.6}) +[2026-06-17 22:47:38,189][starVLA.training.train_starvla][INFO] - Step 3891, Loss: {'train/loss': 0.21123863756656647, 'train/grad_norm_pre_clip': 16.806533813476562, 'timing/data': 0.0008825622498989105, 'timing/model': 0.9278384791687131, 'learning_rate/qwen_vl_interface': 4.515481775777989e-06, 'learning_rate/action_model': 4.515481775777988e-05, 'epoch': 1.6}) +[2026-06-17 22:47:41,980][starVLA.training.train_starvla][INFO] - Step 3892, Loss: {'train/loss': 0.37183815240859985, 'train/grad_norm_pre_clip': 15.01848316192627, 'timing/data': 0.00055715162307024, 'timing/model': 0.928125036880374, 'learning_rate/qwen_vl_interface': 4.51334517509133e-06, 'learning_rate/action_model': 4.5133451750913305e-05, 'epoch': 1.6}) +[2026-06-17 22:47:45,741][starVLA.training.train_starvla][INFO] - Step 3893, Loss: {'train/loss': 0.40376392006874084, 'train/grad_norm_pre_clip': 19.54037857055664, 'timing/data': 0.0005252854898571968, 'timing/model': 0.9300254806876183, 'learning_rate/qwen_vl_interface': 4.511208727114011e-06, 'learning_rate/action_model': 4.51120872711401e-05, 'epoch': 1.6}) +[2026-06-17 22:47:49,475][starVLA.training.train_starvla][INFO] - Step 3894, Loss: {'train/loss': 0.5612223148345947, 'train/grad_norm_pre_clip': 24.706823348999023, 'timing/data': 0.0004945565015077591, 'timing/model': 0.9239511210471392, 'learning_rate/qwen_vl_interface': 4.5090724322889175e-06, 'learning_rate/action_model': 4.5090724322889174e-05, 'epoch': 1.6}) +[2026-06-17 22:47:53,221][starVLA.training.train_starvla][INFO] - Step 3895, Loss: {'train/loss': 0.6080291271209717, 'train/grad_norm_pre_clip': 24.891817092895508, 'timing/data': 0.0005035325884819031, 'timing/model': 1.0949471713975072, 'learning_rate/qwen_vl_interface': 4.506936291058902e-06, 'learning_rate/action_model': 4.506936291058902e-05, 'epoch': 1.6}) +[2026-06-17 22:47:56,804][starVLA.training.train_starvla][INFO] - Step 3896, Loss: {'train/loss': 0.17028561234474182, 'train/grad_norm_pre_clip': 10.340194702148438, 'timing/data': 0.0005059121176600456, 'timing/model': 0.9253487875685096, 'learning_rate/qwen_vl_interface': 4.504800303866798e-06, 'learning_rate/action_model': 4.504800303866798e-05, 'epoch': 1.6}) +[2026-06-17 22:48:00,561][starVLA.training.train_starvla][INFO] - Step 3897, Loss: {'train/loss': 0.3153313398361206, 'train/grad_norm_pre_clip': 16.756710052490234, 'timing/data': 0.0005162348970770836, 'timing/model': 0.930685899220407, 'learning_rate/qwen_vl_interface': 4.50266447115539e-06, 'learning_rate/action_model': 4.5026644711553904e-05, 'epoch': 1.6}) +[2026-06-17 22:48:04,302][starVLA.training.train_starvla][INFO] - Step 3898, Loss: {'train/loss': 0.26606258749961853, 'train/grad_norm_pre_clip': 14.532946586608887, 'timing/data': 0.0007214834913611412, 'timing/model': 0.9255344104021788, 'learning_rate/qwen_vl_interface': 4.500528793367444e-06, 'learning_rate/action_model': 4.5005287933674435e-05, 'epoch': 1.6}) +[2026-06-17 22:48:08,066][starVLA.training.train_starvla][INFO] - Step 3899, Loss: {'train/loss': 0.3459794521331787, 'train/grad_norm_pre_clip': 17.965152740478516, 'timing/data': 0.0005128262564539909, 'timing/model': 0.935934160836041, 'learning_rate/qwen_vl_interface': 4.498393270945683e-06, 'learning_rate/action_model': 4.498393270945683e-05, 'epoch': 1.6}) +[2026-06-17 22:50:05,396][starVLA.training.train_starvla][INFO] - Step 3900, Loss: {'train/loss': 0.34380149841308594, 'train/grad_norm_pre_clip': 17.390483856201172, 'eval/loss': 0.4630642235279083, 'eval/latency_6/loss': 0.4630642235279083, 'eval/demon_fire_precision': 0.9401788467468394, 'eval/demon_fire_recall': 0.9854557207498384, 'eval/demon_fire_f1': 0.9622849928988482, 'eval/demon_fire_rate_model': 0.9023372287145242, 'eval/demon_fire_rate_teacher': 0.860879243183083, 'eval/demon_left_precision': 0.8216432865731463, 'eval/demon_left_recall': 0.7192982456140351, 'eval/demon_left_f1': 0.7670720299345183, 'eval/demon_left_rate_model': 0.13884251530328326, 'eval/demon_left_rate_teacher': 0.15859766277128548, 'eval/demon_right_precision': 0.8118811881188119, 'eval/demon_right_recall': 0.7557603686635944, 'eval/demon_right_f1': 0.782816229116945, 'eval/demon_right_rate_model': 0.1686143572621035, 'eval/demon_right_rate_teacher': 0.18113522537562604, 'eval/demon_move_f1': 0.7749441295257317, 'eval/demon_cc_f1': 0.8972151193361984, 'eval/latency_6/demon_fire_precision': 0.9401788467468394, 'eval/latency_6/demon_fire_recall': 0.9854557207498384, 'eval/latency_6/demon_fire_f1': 0.9622849928988482, 'eval/latency_6/demon_fire_rate_model': 0.9023372287145242, 'eval/latency_6/demon_fire_rate_teacher': 0.860879243183083, 'eval/latency_6/demon_left_precision': 0.8216432865731463, 'eval/latency_6/demon_left_recall': 0.7192982456140351, 'eval/latency_6/demon_left_f1': 0.7670720299345183, 'eval/latency_6/demon_left_rate_model': 0.13884251530328326, 'eval/latency_6/demon_left_rate_teacher': 0.15859766277128548, 'eval/latency_6/demon_right_precision': 0.8118811881188119, 'eval/latency_6/demon_right_recall': 0.7557603686635944, 'eval/latency_6/demon_right_f1': 0.782816229116945, 'eval/latency_6/demon_right_rate_model': 0.1686143572621035, 'eval/latency_6/demon_right_rate_teacher': 0.18113522537562604, 'eval/latency_6/demon_move_f1': 0.7749441295257317, 'eval/latency_6/demon_cc_f1': 0.8972151193361984, 'timing/data': 0.0004820600152015686, 'timing/model': 0.9264716608449817, 'learning_rate/qwen_vl_interface': 4.496257904332806e-06, 'learning_rate/action_model': 4.4962579043328064e-05, 'epoch': 1.61}) +[2026-06-17 22:50:05,401][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_3900_state +[2026-06-17 22:50:05,406][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 22:50:11,936][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_3900_state/model.safetensors +[2026-06-17 22:50:27,094][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_3900_state/optimizer.bin +[2026-06-17 22:50:27,094][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_3900_state/sampler.bin +[2026-06-17 22:50:27,095][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_3900_state/sampler_1.bin +[2026-06-17 22:50:27,097][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_3900_state/random_states_0.pkl +[2026-06-17 22:50:27,101][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3900 +[2026-06-17 22:50:27,101][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 22:50:27,111][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 22:50:30,864][starVLA.training.train_starvla][INFO] - Step 3901, Loss: {'train/loss': 0.2906436324119568, 'train/grad_norm_pre_clip': 16.017824172973633, 'timing/data': 0.0006130142137408257, 'timing/model': 0.9262411212548614, 'learning_rate/qwen_vl_interface': 4.494122693971478e-06, 'learning_rate/action_model': 4.4941226939714785e-05, 'epoch': 1.61}) +[2026-06-17 22:50:34,655][starVLA.training.train_starvla][INFO] - Step 3902, Loss: {'train/loss': 0.5528535842895508, 'train/grad_norm_pre_clip': 19.870222091674805, 'timing/data': 0.0005654366686940193, 'timing/model': 0.9414015142247081, 'learning_rate/qwen_vl_interface': 4.491987640304327e-06, 'learning_rate/action_model': 4.4919876403043265e-05, 'epoch': 1.61}) +[2026-06-17 22:50:38,437][starVLA.training.train_starvla][INFO] - Step 3903, Loss: {'train/loss': 0.20826441049575806, 'train/grad_norm_pre_clip': 10.846747398376465, 'timing/data': 0.0005639148876070976, 'timing/model': 1.1002655513584614, 'learning_rate/qwen_vl_interface': 4.489852743773955e-06, 'learning_rate/action_model': 4.489852743773955e-05, 'epoch': 1.61}) +[2026-06-17 22:50:42,269][starVLA.training.train_starvla][INFO] - Step 3904, Loss: {'train/loss': 0.47377750277519226, 'train/grad_norm_pre_clip': 23.766847610473633, 'timing/data': 0.0005251280963420868, 'timing/model': 1.1422620611265302, 'learning_rate/qwen_vl_interface': 4.487718004822928e-06, 'learning_rate/action_model': 4.487718004822928e-05, 'epoch': 1.61}) +[2026-06-17 22:50:45,988][starVLA.training.train_starvla][INFO] - Step 3905, Loss: {'train/loss': 0.4352637231349945, 'train/grad_norm_pre_clip': 21.200145721435547, 'timing/data': 0.0004926593974232674, 'timing/model': 0.9678246714174747, 'learning_rate/qwen_vl_interface': 4.485583423893778e-06, 'learning_rate/action_model': 4.4855834238937774e-05, 'epoch': 1.61}) +[2026-06-17 22:50:49,866][starVLA.training.train_starvla][INFO] - Step 3906, Loss: {'train/loss': 0.8071823120117188, 'train/grad_norm_pre_clip': 24.953405380249023, 'timing/data': 0.000755813904106617, 'timing/model': 0.9543922990560532, 'learning_rate/qwen_vl_interface': 4.483449001429005e-06, 'learning_rate/action_model': 4.483449001429005e-05, 'epoch': 1.61}) +[2026-06-17 22:50:53,728][starVLA.training.train_starvla][INFO] - Step 3907, Loss: {'train/loss': 0.556304395198822, 'train/grad_norm_pre_clip': 25.45701026916504, 'timing/data': 0.0005024559795856476, 'timing/model': 0.9368062317371368, 'learning_rate/qwen_vl_interface': 4.481314737871081e-06, 'learning_rate/action_model': 4.4813147378710804e-05, 'epoch': 1.61}) +[2026-06-17 22:50:57,475][starVLA.training.train_starvla][INFO] - Step 3908, Loss: {'train/loss': 0.1510913074016571, 'train/grad_norm_pre_clip': 14.134082794189453, 'timing/data': 0.0009250752627849579, 'timing/model': 0.9224491221830249, 'learning_rate/qwen_vl_interface': 4.47918063366244e-06, 'learning_rate/action_model': 4.4791806336624395e-05, 'epoch': 1.61}) +[2026-06-17 22:51:01,312][starVLA.training.train_starvla][INFO] - Step 3909, Loss: {'train/loss': 0.446932315826416, 'train/grad_norm_pre_clip': 21.77663230895996, 'timing/data': 0.0004972387105226517, 'timing/model': 0.9269457831978798, 'learning_rate/qwen_vl_interface': 4.477046689245481e-06, 'learning_rate/action_model': 4.477046689245481e-05, 'epoch': 1.61}) +[2026-06-17 22:51:05,105][starVLA.training.train_starvla][INFO] - Step 3910, Loss: {'train/loss': 0.6365484595298767, 'train/grad_norm_pre_clip': 17.400239944458008, 'timing/data': 0.0009279055520892143, 'timing/model': 1.1073944419622421, 'learning_rate/qwen_vl_interface': 4.474912905062575e-06, 'learning_rate/action_model': 4.474912905062575e-05, 'epoch': 1.61}) +[2026-06-17 22:51:08,938][starVLA.training.train_starvla][INFO] - Step 3911, Loss: {'train/loss': 0.46993887424468994, 'train/grad_norm_pre_clip': 20.881813049316406, 'timing/data': 0.0009093694388866425, 'timing/model': 1.073329179547727, 'learning_rate/qwen_vl_interface': 4.472779281556059e-06, 'learning_rate/action_model': 4.4727792815560585e-05, 'epoch': 1.61}) +[2026-06-17 22:51:12,525][starVLA.training.train_starvla][INFO] - Step 3912, Loss: {'train/loss': 0.588664710521698, 'train/grad_norm_pre_clip': 20.042457580566406, 'timing/data': 0.0004952158778905869, 'timing/model': 0.9305643737316132, 'learning_rate/qwen_vl_interface': 4.470645819168233e-06, 'learning_rate/action_model': 4.4706458191682324e-05, 'epoch': 1.61}) +[2026-06-17 22:51:16,299][starVLA.training.train_starvla][INFO] - Step 3913, Loss: {'train/loss': 0.6432405114173889, 'train/grad_norm_pre_clip': 21.229440689086914, 'timing/data': 0.0008473126217722893, 'timing/model': 0.9290481172502041, 'learning_rate/qwen_vl_interface': 4.468512518341367e-06, 'learning_rate/action_model': 4.4685125183413674e-05, 'epoch': 1.61}) +[2026-06-17 22:51:20,049][starVLA.training.train_starvla][INFO] - Step 3914, Loss: {'train/loss': 0.18563996255397797, 'train/grad_norm_pre_clip': 11.531039237976074, 'timing/data': 0.0005300315096974373, 'timing/model': 0.9261306403204799, 'learning_rate/qwen_vl_interface': 4.466379379517699e-06, 'learning_rate/action_model': 4.4663793795176985e-05, 'epoch': 1.61}) +[2026-06-17 22:51:23,773][starVLA.training.train_starvla][INFO] - Step 3915, Loss: {'train/loss': 0.5151606798171997, 'train/grad_norm_pre_clip': 35.489891052246094, 'timing/data': 0.0004902547225356102, 'timing/model': 0.9200164536014199, 'learning_rate/qwen_vl_interface': 4.464246403139429e-06, 'learning_rate/action_model': 4.464246403139428e-05, 'epoch': 1.61}) +[2026-06-17 22:51:27,527][starVLA.training.train_starvla][INFO] - Step 3916, Loss: {'train/loss': 0.4672035574913025, 'train/grad_norm_pre_clip': 18.894624710083008, 'timing/data': 0.0005195299163460732, 'timing/model': 0.9210615567862988, 'learning_rate/qwen_vl_interface': 4.462113589648723e-06, 'learning_rate/action_model': 4.462113589648723e-05, 'epoch': 1.61}) +[2026-06-17 22:51:31,455][starVLA.training.train_starvla][INFO] - Step 3917, Loss: {'train/loss': 0.42294976115226746, 'train/grad_norm_pre_clip': 20.120473861694336, 'timing/data': 0.0011693770065903664, 'timing/model': 1.243625403381884, 'learning_rate/qwen_vl_interface': 4.459980939487718e-06, 'learning_rate/action_model': 4.4599809394877176e-05, 'epoch': 1.61}) +[2026-06-17 22:51:35,270][starVLA.training.train_starvla][INFO] - Step 3918, Loss: {'train/loss': 0.6479729413986206, 'train/grad_norm_pre_clip': 25.433189392089844, 'timing/data': 0.000616561621427536, 'timing/model': 1.0819010557606816, 'learning_rate/qwen_vl_interface': 4.457848453098514e-06, 'learning_rate/action_model': 4.4578484530985144e-05, 'epoch': 1.61}) +[2026-06-17 22:51:38,849][starVLA.training.train_starvla][INFO] - Step 3919, Loss: {'train/loss': 0.6110005974769592, 'train/grad_norm_pre_clip': 19.486841201782227, 'timing/data': 0.000628029927611351, 'timing/model': 0.9304465856403112, 'learning_rate/qwen_vl_interface': 4.45571613092318e-06, 'learning_rate/action_model': 4.4557161309231794e-05, 'epoch': 1.61}) +[2026-06-17 22:51:42,725][starVLA.training.train_starvla][INFO] - Step 3920, Loss: {'train/loss': 0.2245493084192276, 'train/grad_norm_pre_clip': 11.689620971679688, 'timing/data': 0.000576547347009182, 'timing/model': 0.9392137257382274, 'learning_rate/qwen_vl_interface': 4.453583973403746e-06, 'learning_rate/action_model': 4.453583973403746e-05, 'epoch': 1.61}) +[2026-06-17 22:51:46,592][starVLA.training.train_starvla][INFO] - Step 3921, Loss: {'train/loss': 0.2340811789035797, 'train/grad_norm_pre_clip': 14.26278305053711, 'timing/data': 0.0007253503426909447, 'timing/model': 0.9430270101875067, 'learning_rate/qwen_vl_interface': 4.451451980982209e-06, 'learning_rate/action_model': 4.4514519809822086e-05, 'epoch': 1.61}) +[2026-06-17 22:51:50,353][starVLA.training.train_starvla][INFO] - Step 3922, Loss: {'train/loss': 0.4916786849498749, 'train/grad_norm_pre_clip': 19.076017379760742, 'timing/data': 0.0004645697772502899, 'timing/model': 0.9193664928898215, 'learning_rate/qwen_vl_interface': 4.4493201541005385e-06, 'learning_rate/action_model': 4.4493201541005384e-05, 'epoch': 1.61}) +[2026-06-17 22:51:54,099][starVLA.training.train_starvla][INFO] - Step 3923, Loss: {'train/loss': 0.21142534911632538, 'train/grad_norm_pre_clip': 12.089513778686523, 'timing/data': 0.000509391538798809, 'timing/model': 0.9265605388209224, 'learning_rate/qwen_vl_interface': 4.447188493200661e-06, 'learning_rate/action_model': 4.447188493200661e-05, 'epoch': 1.61}) +[2026-06-17 22:51:57,819][starVLA.training.train_starvla][INFO] - Step 3924, Loss: {'train/loss': 0.448739618062973, 'train/grad_norm_pre_clip': 27.31671142578125, 'timing/data': 0.00048445165157318115, 'timing/model': 1.083711737766862, 'learning_rate/qwen_vl_interface': 4.445056998724472e-06, 'learning_rate/action_model': 4.4450569987244716e-05, 'epoch': 1.62}) +[2026-06-17 22:52:01,497][starVLA.training.train_starvla][INFO] - Step 3925, Loss: {'train/loss': 0.48858842253685, 'train/grad_norm_pre_clip': 15.303542137145996, 'timing/data': 0.0005003027617931366, 'timing/model': 1.0237730536609888, 'learning_rate/qwen_vl_interface': 4.4429256711138335e-06, 'learning_rate/action_model': 4.442925671113833e-05, 'epoch': 1.62}) +[2026-06-17 22:52:05,087][starVLA.training.train_starvla][INFO] - Step 3926, Loss: {'train/loss': 0.2152031809091568, 'train/grad_norm_pre_clip': 12.598287582397461, 'timing/data': 0.0013722768053412437, 'timing/model': 0.9271108862012625, 'learning_rate/qwen_vl_interface': 4.440794510810572e-06, 'learning_rate/action_model': 4.440794510810572e-05, 'epoch': 1.62}) +[2026-06-17 22:52:08,910][starVLA.training.train_starvla][INFO] - Step 3927, Loss: {'train/loss': 0.8008384704589844, 'train/grad_norm_pre_clip': 21.285898208618164, 'timing/data': 0.0005027614533901215, 'timing/model': 0.9286944875493646, 'learning_rate/qwen_vl_interface': 4.438663518256481e-06, 'learning_rate/action_model': 4.438663518256481e-05, 'epoch': 1.62}) +[2026-06-17 22:52:12,639][starVLA.training.train_starvla][INFO] - Step 3928, Loss: {'train/loss': 0.18402697145938873, 'train/grad_norm_pre_clip': 11.852815628051758, 'timing/data': 0.0007318519055843353, 'timing/model': 0.9234972391277552, 'learning_rate/qwen_vl_interface': 4.436532693893311e-06, 'learning_rate/action_model': 4.436532693893311e-05, 'epoch': 1.62}) +[2026-06-17 22:52:16,382][starVLA.training.train_starvla][INFO] - Step 3929, Loss: {'train/loss': 0.2511642873287201, 'train/grad_norm_pre_clip': 14.671468734741211, 'timing/data': 0.000544816255569458, 'timing/model': 0.925445256754756, 'learning_rate/qwen_vl_interface': 4.434402038162793e-06, 'learning_rate/action_model': 4.434402038162793e-05, 'epoch': 1.62}) +[2026-06-17 22:52:20,133][starVLA.training.train_starvla][INFO] - Step 3930, Loss: {'train/loss': 0.1319885551929474, 'train/grad_norm_pre_clip': 8.978132247924805, 'timing/data': 0.0005262577906250954, 'timing/model': 1.0925312219187617, 'learning_rate/qwen_vl_interface': 4.432271551506609e-06, 'learning_rate/action_model': 4.4322715515066094e-05, 'epoch': 1.62}) +[2026-06-17 22:52:23,804][starVLA.training.train_starvla][INFO] - Step 3931, Loss: {'train/loss': 0.3558519780635834, 'train/grad_norm_pre_clip': 15.257170677185059, 'timing/data': 0.0008963486179709435, 'timing/model': 1.0242680059745908, 'learning_rate/qwen_vl_interface': 4.4301412343664125e-06, 'learning_rate/action_model': 4.4301412343664125e-05, 'epoch': 1.62}) +[2026-06-17 22:52:27,367][starVLA.training.train_starvla][INFO] - Step 3932, Loss: {'train/loss': 0.44876885414123535, 'train/grad_norm_pre_clip': 32.54928970336914, 'timing/data': 0.0005854489281773567, 'timing/model': 0.9300056425854564, 'learning_rate/qwen_vl_interface': 4.428011087183819e-06, 'learning_rate/action_model': 4.428011087183819e-05, 'epoch': 1.62}) +[2026-06-17 22:52:31,080][starVLA.training.train_starvla][INFO] - Step 3933, Loss: {'train/loss': 0.3263656497001648, 'train/grad_norm_pre_clip': 14.148134231567383, 'timing/data': 0.0005242228507995605, 'timing/model': 0.9327318100258708, 'learning_rate/qwen_vl_interface': 4.4258811104004125e-06, 'learning_rate/action_model': 4.425881110400413e-05, 'epoch': 1.62}) +[2026-06-17 22:52:34,790][starVLA.training.train_starvla][INFO] - Step 3934, Loss: {'train/loss': 0.3181900978088379, 'train/grad_norm_pre_clip': 12.994240760803223, 'timing/data': 0.0004857676103711128, 'timing/model': 0.9206741340458393, 'learning_rate/qwen_vl_interface': 4.4237513044577394e-06, 'learning_rate/action_model': 4.4237513044577386e-05, 'epoch': 1.62}) +[2026-06-17 22:52:38,557][starVLA.training.train_starvla][INFO] - Step 3935, Loss: {'train/loss': 0.6406479477882385, 'train/grad_norm_pre_clip': 51.289817810058594, 'timing/data': 0.0005865041166543961, 'timing/model': 0.9237571060657501, 'learning_rate/qwen_vl_interface': 4.421621669797309e-06, 'learning_rate/action_model': 4.421621669797308e-05, 'epoch': 1.62}) +[2026-06-17 22:52:42,241][starVLA.training.train_starvla][INFO] - Step 3936, Loss: {'train/loss': 0.5371453166007996, 'train/grad_norm_pre_clip': 24.29214096069336, 'timing/data': 0.0005429964512586594, 'timing/model': 0.9243890140205622, 'learning_rate/qwen_vl_interface': 4.419492206860599e-06, 'learning_rate/action_model': 4.419492206860598e-05, 'epoch': 1.62}) +[2026-06-17 22:52:46,000][starVLA.training.train_starvla][INFO] - Step 3937, Loss: {'train/loss': 0.22331734001636505, 'train/grad_norm_pre_clip': 13.754584312438965, 'timing/data': 0.0004910249263048172, 'timing/model': 1.0934107871726155, 'learning_rate/qwen_vl_interface': 4.417362916089048e-06, 'learning_rate/action_model': 4.417362916089047e-05, 'epoch': 1.62}) +[2026-06-17 22:52:49,598][starVLA.training.train_starvla][INFO] - Step 3938, Loss: {'train/loss': 0.22668197751045227, 'train/grad_norm_pre_clip': 10.627031326293945, 'timing/data': 0.0005036694929003716, 'timing/model': 0.9320944594219327, 'learning_rate/qwen_vl_interface': 4.415233797924059e-06, 'learning_rate/action_model': 4.415233797924059e-05, 'epoch': 1.62}) +[2026-06-17 22:52:53,342][starVLA.training.train_starvla][INFO] - Step 3939, Loss: {'train/loss': 0.46752414107322693, 'train/grad_norm_pre_clip': 28.340452194213867, 'timing/data': 0.0006858892738819122, 'timing/model': 0.9339479729533195, 'learning_rate/qwen_vl_interface': 4.413104852807002e-06, 'learning_rate/action_model': 4.413104852807001e-05, 'epoch': 1.62}) +[2026-06-17 22:52:57,092][starVLA.training.train_starvla][INFO] - Step 3940, Loss: {'train/loss': 0.3954971432685852, 'train/grad_norm_pre_clip': 15.485264778137207, 'timing/data': 0.0005783932283520699, 'timing/model': 0.930928299203515, 'learning_rate/qwen_vl_interface': 4.41097608117921e-06, 'learning_rate/action_model': 4.41097608117921e-05, 'epoch': 1.62}) +[2026-06-17 22:53:00,772][starVLA.training.train_starvla][INFO] - Step 3941, Loss: {'train/loss': 0.28960680961608887, 'train/grad_norm_pre_clip': 12.900845527648926, 'timing/data': 0.000498608686029911, 'timing/model': 0.9202184015884995, 'learning_rate/qwen_vl_interface': 4.40884748348198e-06, 'learning_rate/action_model': 4.40884748348198e-05, 'epoch': 1.62}) +[2026-06-17 22:53:04,554][starVLA.training.train_starvla][INFO] - Step 3942, Loss: {'train/loss': 0.3870070278644562, 'train/grad_norm_pre_clip': 15.1848783493042, 'timing/data': 0.0004891874268651009, 'timing/model': 0.923235640861094, 'learning_rate/qwen_vl_interface': 4.406719060156572e-06, 'learning_rate/action_model': 4.4067190601565716e-05, 'epoch': 1.62}) +[2026-06-17 22:53:08,320][starVLA.training.train_starvla][INFO] - Step 3943, Loss: {'train/loss': 0.4010469317436218, 'train/grad_norm_pre_clip': 16.95519256591797, 'timing/data': 0.0005227392539381981, 'timing/model': 1.103253803215921, 'learning_rate/qwen_vl_interface': 4.404590811644209e-06, 'learning_rate/action_model': 4.4045908116442087e-05, 'epoch': 1.62}) +[2026-06-17 22:53:11,919][starVLA.training.train_starvla][INFO] - Step 3944, Loss: {'train/loss': 0.4871382415294647, 'train/grad_norm_pre_clip': 13.34077262878418, 'timing/data': 0.0005160942673683167, 'timing/model': 0.9310692232102156, 'learning_rate/qwen_vl_interface': 4.40246273838608e-06, 'learning_rate/action_model': 4.40246273838608e-05, 'epoch': 1.62}) +[2026-06-17 22:53:15,736][starVLA.training.train_starvla][INFO] - Step 3945, Loss: {'train/loss': 0.3732661306858063, 'train/grad_norm_pre_clip': 13.390538215637207, 'timing/data': 0.000516129657626152, 'timing/model': 0.947772229090333, 'learning_rate/qwen_vl_interface': 4.400334840823339e-06, 'learning_rate/action_model': 4.400334840823339e-05, 'epoch': 1.62}) +[2026-06-17 22:53:19,482][starVLA.training.train_starvla][INFO] - Step 3946, Loss: {'train/loss': 0.27372977137565613, 'train/grad_norm_pre_clip': 16.013952255249023, 'timing/data': 0.0005571525543928146, 'timing/model': 0.9394053872674704, 'learning_rate/qwen_vl_interface': 4.398207119397098e-06, 'learning_rate/action_model': 4.398207119397098e-05, 'epoch': 1.62}) +[2026-06-17 22:53:23,241][starVLA.training.train_starvla][INFO] - Step 3947, Loss: {'train/loss': 0.5363034009933472, 'train/grad_norm_pre_clip': 17.826080322265625, 'timing/data': 0.000672096386551857, 'timing/model': 0.9381202226504683, 'learning_rate/qwen_vl_interface': 4.396079574548439e-06, 'learning_rate/action_model': 4.396079574548439e-05, 'epoch': 1.62}) +[2026-06-17 22:53:27,021][starVLA.training.train_starvla][INFO] - Step 3948, Loss: {'train/loss': 0.35889336466789246, 'train/grad_norm_pre_clip': 18.111412048339844, 'timing/data': 0.00048597343266010284, 'timing/model': 0.9221930652856827, 'learning_rate/qwen_vl_interface': 4.393952206718403e-06, 'learning_rate/action_model': 4.3939522067184026e-05, 'epoch': 1.63}) +[2026-06-17 22:53:30,791][starVLA.training.train_starvla][INFO] - Step 3949, Loss: {'train/loss': 0.27252432703971863, 'train/grad_norm_pre_clip': 16.482881546020508, 'timing/data': 0.0005365563556551933, 'timing/model': 0.9262731038033962, 'learning_rate/qwen_vl_interface': 4.391825016347996e-06, 'learning_rate/action_model': 4.391825016347996e-05, 'epoch': 1.63}) +[2026-06-17 22:53:34,561][starVLA.training.train_starvla][INFO] - Step 3950, Loss: {'train/loss': 0.41744187474250793, 'train/grad_norm_pre_clip': 17.0931396484375, 'timing/data': 0.0004926053807139397, 'timing/model': 1.1023618057370186, 'learning_rate/qwen_vl_interface': 4.3896980038781844e-06, 'learning_rate/action_model': 4.3896980038781846e-05, 'epoch': 1.63}) +[2026-06-17 22:53:38,152][starVLA.training.train_starvla][INFO] - Step 3951, Loss: {'train/loss': 0.21349041163921356, 'train/grad_norm_pre_clip': 11.29548168182373, 'timing/data': 0.0004955362528562546, 'timing/model': 0.9303938830271363, 'learning_rate/qwen_vl_interface': 4.3875711697499054e-06, 'learning_rate/action_model': 4.387571169749905e-05, 'epoch': 1.63}) +[2026-06-17 22:53:41,909][starVLA.training.train_starvla][INFO] - Step 3952, Loss: {'train/loss': 0.32451218366622925, 'train/grad_norm_pre_clip': 17.039398193359375, 'timing/data': 0.000516250729560852, 'timing/model': 0.9369290135800838, 'learning_rate/qwen_vl_interface': 4.38544451440405e-06, 'learning_rate/action_model': 4.3854445144040495e-05, 'epoch': 1.63}) +[2026-06-17 22:53:45,663][starVLA.training.train_starvla][INFO] - Step 3953, Loss: {'train/loss': 0.4744391143321991, 'train/grad_norm_pre_clip': 20.55076789855957, 'timing/data': 0.000515669584274292, 'timing/model': 0.9327741526067257, 'learning_rate/qwen_vl_interface': 4.383318038281475e-06, 'learning_rate/action_model': 4.383318038281475e-05, 'epoch': 1.63}) +[2026-06-17 22:53:49,422][starVLA.training.train_starvla][INFO] - Step 3954, Loss: {'train/loss': 0.41162723302841187, 'train/grad_norm_pre_clip': 17.592830657958984, 'timing/data': 0.0005596894770860672, 'timing/model': 0.924574445001781, 'learning_rate/qwen_vl_interface': 4.381191741823005e-06, 'learning_rate/action_model': 4.3811917418230046e-05, 'epoch': 1.63}) +[2026-06-17 22:53:53,172][starVLA.training.train_starvla][INFO] - Step 3955, Loss: {'train/loss': 0.46618273854255676, 'train/grad_norm_pre_clip': 16.325927734375, 'timing/data': 0.000510169193148613, 'timing/model': 0.9208667054772377, 'learning_rate/qwen_vl_interface': 4.379065625469422e-06, 'learning_rate/action_model': 4.3790656254694215e-05, 'epoch': 1.63}) +[2026-06-17 22:53:56,909][starVLA.training.train_starvla][INFO] - Step 3956, Loss: {'train/loss': 0.29455894231796265, 'train/grad_norm_pre_clip': 17.801105499267578, 'timing/data': 0.000532984733581543, 'timing/model': 1.086103918030858, 'learning_rate/qwen_vl_interface': 4.3769396896614724e-06, 'learning_rate/action_model': 4.376939689661472e-05, 'epoch': 1.63}) +[2026-06-17 22:54:00,648][starVLA.training.train_starvla][INFO] - Step 3957, Loss: {'train/loss': 0.4309855103492737, 'train/grad_norm_pre_clip': 16.037494659423828, 'timing/data': 0.0005091624334454536, 'timing/model': 1.0897999657317996, 'learning_rate/qwen_vl_interface': 4.374813934839862e-06, 'learning_rate/action_model': 4.374813934839862e-05, 'epoch': 1.63}) +[2026-06-17 22:54:04,250][starVLA.training.train_starvla][INFO] - Step 3958, Loss: {'train/loss': 0.1750679910182953, 'train/grad_norm_pre_clip': 10.467686653137207, 'timing/data': 0.0004895152524113655, 'timing/model': 0.9349076310172677, 'learning_rate/qwen_vl_interface': 4.372688361445267e-06, 'learning_rate/action_model': 4.3726883614452665e-05, 'epoch': 1.63}) +[2026-06-17 22:54:08,000][starVLA.training.train_starvla][INFO] - Step 3959, Loss: {'train/loss': 0.40609076619148254, 'train/grad_norm_pre_clip': 15.166036605834961, 'timing/data': 0.0006402963772416115, 'timing/model': 0.9274552585557103, 'learning_rate/qwen_vl_interface': 4.370562969918319e-06, 'learning_rate/action_model': 4.370562969918319e-05, 'epoch': 1.63}) +[2026-06-17 22:54:11,759][starVLA.training.train_starvla][INFO] - Step 3960, Loss: {'train/loss': 0.5062988996505737, 'train/grad_norm_pre_clip': 24.599382400512695, 'timing/data': 0.0005096802487969398, 'timing/model': 0.9296961016952991, 'learning_rate/qwen_vl_interface': 4.368437760699613e-06, 'learning_rate/action_model': 4.3684377606996124e-05, 'epoch': 1.63}) +[2026-06-17 22:54:15,502][starVLA.training.train_starvla][INFO] - Step 3961, Loss: {'train/loss': 0.5280776023864746, 'train/grad_norm_pre_clip': 20.811626434326172, 'timing/data': 0.0005142698064446449, 'timing/model': 0.9301929175853729, 'learning_rate/qwen_vl_interface': 4.366312734229706e-06, 'learning_rate/action_model': 4.366312734229706e-05, 'epoch': 1.63}) +[2026-06-17 22:54:19,253][starVLA.training.train_starvla][INFO] - Step 3962, Loss: {'train/loss': 0.20378172397613525, 'train/grad_norm_pre_clip': 12.670059204101562, 'timing/data': 0.00047649722546339035, 'timing/model': 0.9295626198872924, 'learning_rate/qwen_vl_interface': 4.364187890949122e-06, 'learning_rate/action_model': 4.364187890949122e-05, 'epoch': 1.63}) +[2026-06-17 22:54:23,018][starVLA.training.train_starvla][INFO] - Step 3963, Loss: {'train/loss': 0.39349931478500366, 'train/grad_norm_pre_clip': 17.74640464782715, 'timing/data': 0.0005642985925078392, 'timing/model': 1.1088756723329425, 'learning_rate/qwen_vl_interface': 4.362063231298342e-06, 'learning_rate/action_model': 4.3620632312983425e-05, 'epoch': 1.63}) +[2026-06-17 22:54:26,722][starVLA.training.train_starvla][INFO] - Step 3964, Loss: {'train/loss': 0.39702990651130676, 'train/grad_norm_pre_clip': 22.555252075195312, 'timing/data': 0.0004948917776346207, 'timing/model': 1.0556770879775286, 'learning_rate/qwen_vl_interface': 4.359938755717807e-06, 'learning_rate/action_model': 4.359938755717807e-05, 'epoch': 1.63}) +[2026-06-17 22:54:30,290][starVLA.training.train_starvla][INFO] - Step 3965, Loss: {'train/loss': 0.6143775582313538, 'train/grad_norm_pre_clip': 19.773147583007812, 'timing/data': 0.00049616489559412, 'timing/model': 0.9258060762658715, 'learning_rate/qwen_vl_interface': 4.357814464647929e-06, 'learning_rate/action_model': 4.357814464647929e-05, 'epoch': 1.63}) +[2026-06-17 22:54:34,012][starVLA.training.train_starvla][INFO] - Step 3966, Loss: {'train/loss': 0.20587590336799622, 'train/grad_norm_pre_clip': 12.809117317199707, 'timing/data': 0.0005024513229727745, 'timing/model': 0.9285446079447865, 'learning_rate/qwen_vl_interface': 4.355690358529071e-06, 'learning_rate/action_model': 4.3556903585290705e-05, 'epoch': 1.63}) +[2026-06-17 22:54:37,736][starVLA.training.train_starvla][INFO] - Step 3967, Loss: {'train/loss': 0.5566059947013855, 'train/grad_norm_pre_clip': 19.748733520507812, 'timing/data': 0.0004871264100074768, 'timing/model': 0.9252861477434635, 'learning_rate/qwen_vl_interface': 4.353566437801564e-06, 'learning_rate/action_model': 4.353566437801563e-05, 'epoch': 1.63}) +[2026-06-17 22:54:41,412][starVLA.training.train_starvla][INFO] - Step 3968, Loss: {'train/loss': 0.2590537667274475, 'train/grad_norm_pre_clip': 15.323945999145508, 'timing/data': 0.0004952345043420792, 'timing/model': 0.9216961367055774, 'learning_rate/qwen_vl_interface': 4.351442702905698e-06, 'learning_rate/action_model': 4.3514427029056976e-05, 'epoch': 1.63}) +[2026-06-17 22:54:45,155][starVLA.training.train_starvla][INFO] - Step 3969, Loss: {'train/loss': 0.15500669181346893, 'train/grad_norm_pre_clip': 10.845242500305176, 'timing/data': 0.0004760734736919403, 'timing/model': 0.9288233434781432, 'learning_rate/qwen_vl_interface': 4.349319154281727e-06, 'learning_rate/action_model': 4.349319154281728e-05, 'epoch': 1.63}) +[2026-06-17 22:54:48,929][starVLA.training.train_starvla][INFO] - Step 3970, Loss: {'train/loss': 0.428970605134964, 'train/grad_norm_pre_clip': 15.664134979248047, 'timing/data': 0.0013362979516386986, 'timing/model': 1.1010437859222293, 'learning_rate/qwen_vl_interface': 4.347195792369867e-06, 'learning_rate/action_model': 4.347195792369867e-05, 'epoch': 1.63}) +[2026-06-17 22:54:52,537][starVLA.training.train_starvla][INFO] - Step 3971, Loss: {'train/loss': 0.39150866866111755, 'train/grad_norm_pre_clip': 22.562702178955078, 'timing/data': 0.0005059270188212395, 'timing/model': 0.9306279998272657, 'learning_rate/qwen_vl_interface': 4.345072617610288e-06, 'learning_rate/action_model': 4.345072617610288e-05, 'epoch': 1.63}) +[2026-06-17 22:54:56,292][starVLA.training.train_starvla][INFO] - Step 3972, Loss: {'train/loss': 0.47457101941108704, 'train/grad_norm_pre_clip': 14.921411514282227, 'timing/data': 0.0005222596228122711, 'timing/model': 0.9316005893051624, 'learning_rate/qwen_vl_interface': 4.342949630443129e-06, 'learning_rate/action_model': 4.342949630443129e-05, 'epoch': 1.64}) +[2026-06-17 22:54:59,969][starVLA.training.train_starvla][INFO] - Step 3973, Loss: {'train/loss': 0.27531853318214417, 'train/grad_norm_pre_clip': 14.966320037841797, 'timing/data': 0.000510520301759243, 'timing/model': 0.932500559836626, 'learning_rate/qwen_vl_interface': 4.340826831308488e-06, 'learning_rate/action_model': 4.340826831308488e-05, 'epoch': 1.64}) +[2026-06-17 22:55:03,720][starVLA.training.train_starvla][INFO] - Step 3974, Loss: {'train/loss': 0.3265492022037506, 'train/grad_norm_pre_clip': 15.47339916229248, 'timing/data': 0.0005291895940899849, 'timing/model': 0.9301216723397374, 'learning_rate/qwen_vl_interface': 4.338704220646424e-06, 'learning_rate/action_model': 4.338704220646424e-05, 'epoch': 1.64}) +[2026-06-17 22:55:07,581][starVLA.training.train_starvla][INFO] - Step 3975, Loss: {'train/loss': 0.7900593876838684, 'train/grad_norm_pre_clip': 22.528343200683594, 'timing/data': 0.0008040471002459526, 'timing/model': 0.9434180548414588, 'learning_rate/qwen_vl_interface': 4.336581798896955e-06, 'learning_rate/action_model': 4.336581798896955e-05, 'epoch': 1.64}) +[2026-06-17 22:55:11,372][starVLA.training.train_starvla][INFO] - Step 3976, Loss: {'train/loss': 0.16022448241710663, 'train/grad_norm_pre_clip': 14.351628303527832, 'timing/data': 0.000795864500105381, 'timing/model': 1.1070214649662375, 'learning_rate/qwen_vl_interface': 4.334459566500062e-06, 'learning_rate/action_model': 4.334459566500062e-05, 'epoch': 1.64}) +[2026-06-17 22:55:15,113][starVLA.training.train_starvla][INFO] - Step 3977, Loss: {'train/loss': 0.623807430267334, 'train/grad_norm_pre_clip': 18.084299087524414, 'timing/data': 0.0005104318261146545, 'timing/model': 1.0535464389249682, 'learning_rate/qwen_vl_interface': 4.332337523895688e-06, 'learning_rate/action_model': 4.3323375238956876e-05, 'epoch': 1.64}) +[2026-06-17 22:55:18,743][starVLA.training.train_starvla][INFO] - Step 3978, Loss: {'train/loss': 0.5255985856056213, 'train/grad_norm_pre_clip': 18.163789749145508, 'timing/data': 0.0009628040716052055, 'timing/model': 0.9803904676809907, 'learning_rate/qwen_vl_interface': 4.330215671523731e-06, 'learning_rate/action_model': 4.330215671523731e-05, 'epoch': 1.64}) +[2026-06-17 22:55:22,574][starVLA.training.train_starvla][INFO] - Step 3979, Loss: {'train/loss': 0.29836294054985046, 'train/grad_norm_pre_clip': 15.993769645690918, 'timing/data': 0.0009828265756368637, 'timing/model': 0.927731336094439, 'learning_rate/qwen_vl_interface': 4.328094009824054e-06, 'learning_rate/action_model': 4.328094009824054e-05, 'epoch': 1.64}) +[2026-06-17 22:55:26,509][starVLA.training.train_starvla][INFO] - Step 3980, Loss: {'train/loss': 0.6844750642776489, 'train/grad_norm_pre_clip': 25.933856964111328, 'timing/data': 0.00047889910638332367, 'timing/model': 0.9294068720191717, 'learning_rate/qwen_vl_interface': 4.3259725392364825e-06, 'learning_rate/action_model': 4.325972539236482e-05, 'epoch': 1.64}) +[2026-06-17 22:55:30,270][starVLA.training.train_starvla][INFO] - Step 3981, Loss: {'train/loss': 0.4155990779399872, 'train/grad_norm_pre_clip': 17.082931518554688, 'timing/data': 0.0005491180345416069, 'timing/model': 0.9382353983819485, 'learning_rate/qwen_vl_interface': 4.323851260200797e-06, 'learning_rate/action_model': 4.323851260200796e-05, 'epoch': 1.64}) +[2026-06-17 22:55:34,064][starVLA.training.train_starvla][INFO] - Step 3982, Loss: {'train/loss': 0.5648546814918518, 'train/grad_norm_pre_clip': 15.035070419311523, 'timing/data': 0.0007116813212633133, 'timing/model': 0.9435772495344281, 'learning_rate/qwen_vl_interface': 4.321730173156743e-06, 'learning_rate/action_model': 4.321730173156743e-05, 'epoch': 1.64}) +[2026-06-17 22:55:37,853][starVLA.training.train_starvla][INFO] - Step 3983, Loss: {'train/loss': 0.49213674664497375, 'train/grad_norm_pre_clip': 16.363435745239258, 'timing/data': 0.0005758441984653473, 'timing/model': 1.104201128706336, 'learning_rate/qwen_vl_interface': 4.31960927854402e-06, 'learning_rate/action_model': 4.31960927854402e-05, 'epoch': 1.64}) +[2026-06-17 22:55:41,624][starVLA.training.train_starvla][INFO] - Step 3984, Loss: {'train/loss': 0.3079085350036621, 'train/grad_norm_pre_clip': 16.86333656311035, 'timing/data': 0.0005215136334300041, 'timing/model': 1.0858204048126936, 'learning_rate/qwen_vl_interface': 4.317488576802298e-06, 'learning_rate/action_model': 4.3174885768022986e-05, 'epoch': 1.64}) +[2026-06-17 22:55:45,297][starVLA.training.train_starvla][INFO] - Step 3985, Loss: {'train/loss': 0.18733282387256622, 'train/grad_norm_pre_clip': 12.321139335632324, 'timing/data': 0.0004830453544855118, 'timing/model': 0.9275706158950925, 'learning_rate/qwen_vl_interface': 4.315368068371197e-06, 'learning_rate/action_model': 4.315368068371197e-05, 'epoch': 1.64}) +[2026-06-17 22:55:49,059][starVLA.training.train_starvla][INFO] - Step 3986, Loss: {'train/loss': 0.2949979901313782, 'train/grad_norm_pre_clip': 12.484256744384766, 'timing/data': 0.0006917398422956467, 'timing/model': 0.9289569985121489, 'learning_rate/qwen_vl_interface': 4.313247753690298e-06, 'learning_rate/action_model': 4.3132477536902985e-05, 'epoch': 1.64}) +[2026-06-17 22:55:52,835][starVLA.training.train_starvla][INFO] - Step 3987, Loss: {'train/loss': 0.10175632685422897, 'train/grad_norm_pre_clip': 6.217721462249756, 'timing/data': 0.0005667321383953094, 'timing/model': 0.9262010343372822, 'learning_rate/qwen_vl_interface': 4.311127633199151e-06, 'learning_rate/action_model': 4.3111276331991504e-05, 'epoch': 1.64}) +[2026-06-17 22:55:56,573][starVLA.training.train_starvla][INFO] - Step 3988, Loss: {'train/loss': 0.38291648030281067, 'train/grad_norm_pre_clip': 20.899883270263672, 'timing/data': 0.0005085533484816551, 'timing/model': 0.9217467959970236, 'learning_rate/qwen_vl_interface': 4.309007707337255e-06, 'learning_rate/action_model': 4.309007707337255e-05, 'epoch': 1.64}) +[2026-06-17 22:56:00,306][starVLA.training.train_starvla][INFO] - Step 3989, Loss: {'train/loss': 0.2427932471036911, 'train/grad_norm_pre_clip': 12.565587043762207, 'timing/data': 0.00048795249313116074, 'timing/model': 0.9235943499952555, 'learning_rate/qwen_vl_interface': 4.3068879765440745e-06, 'learning_rate/action_model': 4.306887976544074e-05, 'epoch': 1.64}) +[2026-06-17 22:56:04,040][starVLA.training.train_starvla][INFO] - Step 3990, Loss: {'train/loss': 0.5935937166213989, 'train/grad_norm_pre_clip': 12.580354690551758, 'timing/data': 0.0005051735788583755, 'timing/model': 1.0845798309892416, 'learning_rate/qwen_vl_interface': 4.304768441259029e-06, 'learning_rate/action_model': 4.3047684412590285e-05, 'epoch': 1.64}) +[2026-06-17 22:56:07,742][starVLA.training.train_starvla][INFO] - Step 3991, Loss: {'train/loss': 0.3215090036392212, 'train/grad_norm_pre_clip': 15.668109893798828, 'timing/data': 0.0008271336555480957, 'timing/model': 1.0437710955739021, 'learning_rate/qwen_vl_interface': 4.302649101921505e-06, 'learning_rate/action_model': 4.302649101921505e-05, 'epoch': 1.64}) +[2026-06-17 22:56:11,329][starVLA.training.train_starvla][INFO] - Step 3992, Loss: {'train/loss': 0.40434226393699646, 'train/grad_norm_pre_clip': 15.218997955322266, 'timing/data': 0.0005038660019636154, 'timing/model': 0.9280555220320821, 'learning_rate/qwen_vl_interface': 4.300529958970841e-06, 'learning_rate/action_model': 4.300529958970841e-05, 'epoch': 1.64}) +[2026-06-17 22:56:15,059][starVLA.training.train_starvla][INFO] - Step 3993, Loss: {'train/loss': 0.45666229724884033, 'train/grad_norm_pre_clip': 27.33881950378418, 'timing/data': 0.0009976178407669067, 'timing/model': 0.924328688532114, 'learning_rate/qwen_vl_interface': 4.298411012846338e-06, 'learning_rate/action_model': 4.2984110128463376e-05, 'epoch': 1.64}) +[2026-06-17 22:56:18,803][starVLA.training.train_starvla][INFO] - Step 3994, Loss: {'train/loss': 0.22927020490169525, 'train/grad_norm_pre_clip': 13.459498405456543, 'timing/data': 0.0005061933770775795, 'timing/model': 0.9183147624135017, 'learning_rate/qwen_vl_interface': 4.296292263987255e-06, 'learning_rate/action_model': 4.2962922639872556e-05, 'epoch': 1.64}) +[2026-06-17 22:56:22,541][starVLA.training.train_starvla][INFO] - Step 3995, Loss: {'train/loss': 0.8365223407745361, 'train/grad_norm_pre_clip': 24.199203491210938, 'timing/data': 0.0004977546632289886, 'timing/model': 0.9260916234925389, 'learning_rate/qwen_vl_interface': 4.294173712832813e-06, 'learning_rate/action_model': 4.294173712832813e-05, 'epoch': 1.64}) +[2026-06-17 22:56:26,266][starVLA.training.train_starvla][INFO] - Step 3996, Loss: {'train/loss': 0.4291200637817383, 'train/grad_norm_pre_clip': 18.850025177001953, 'timing/data': 0.0005261925980448723, 'timing/model': 1.0866154059767723, 'learning_rate/qwen_vl_interface': 4.292055359822191e-06, 'learning_rate/action_model': 4.29205535982219e-05, 'epoch': 1.64}) +[2026-06-17 22:56:29,914][starVLA.training.train_starvla][INFO] - Step 3997, Loss: {'train/loss': 0.4855111241340637, 'train/grad_norm_pre_clip': 20.107017517089844, 'timing/data': 0.000492362305521965, 'timing/model': 1.0149304699152708, 'learning_rate/qwen_vl_interface': 4.289937205394519e-06, 'learning_rate/action_model': 4.2899372053945185e-05, 'epoch': 1.65}) +[2026-06-17 22:56:33,513][starVLA.training.train_starvla][INFO] - Step 3998, Loss: {'train/loss': 0.3071286082267761, 'train/grad_norm_pre_clip': 13.460348129272461, 'timing/data': 0.0011790962889790535, 'timing/model': 0.9307079995051026, 'learning_rate/qwen_vl_interface': 4.287819249988901e-06, 'learning_rate/action_model': 4.2878192499889007e-05, 'epoch': 1.65}) +[2026-06-17 22:56:37,243][starVLA.training.train_starvla][INFO] - Step 3999, Loss: {'train/loss': 0.3578760325908661, 'train/grad_norm_pre_clip': 30.67848014831543, 'timing/data': 0.0012921104207634926, 'timing/model': 0.9223705222830176, 'learning_rate/qwen_vl_interface': 4.285701494044387e-06, 'learning_rate/action_model': 4.285701494044387e-05, 'epoch': 1.65}) +[2026-06-17 22:58:31,057][starVLA.training.train_starvla][INFO] - Step 4000, Loss: {'train/loss': 0.34925445914268494, 'train/grad_norm_pre_clip': 16.41158676147461, 'eval/loss': 0.4455537796020508, 'eval/latency_6/loss': 0.4455537796020508, 'eval/demon_fire_precision': 0.9555414417275325, 'eval/demon_fire_recall': 0.9725274725274725, 'eval/demon_fire_f1': 0.9639596347909658, 'eval/demon_fire_rate_model': 0.8761825264329438, 'eval/demon_fire_rate_teacher': 0.860879243183083, 'eval/demon_left_precision': 0.8180039138943248, 'eval/demon_left_recall': 0.7333333333333333, 'eval/demon_left_f1': 0.7733580018501387, 'eval/demon_left_rate_model': 0.14218141346688926, 'eval/demon_left_rate_teacher': 0.15859766277128548, 'eval/demon_right_precision': 0.8150572831423896, 'eval/demon_right_recall': 0.7649769585253456, 'eval/demon_right_f1': 0.7892234548335976, 'eval/demon_right_rate_model': 0.17000556483027268, 'eval/demon_right_rate_teacher': 0.18113522537562604, 'eval/demon_move_f1': 0.7812907283418682, 'eval/demon_cc_f1': 0.9007779830157827, 'eval/latency_6/demon_fire_precision': 0.9555414417275325, 'eval/latency_6/demon_fire_recall': 0.9725274725274725, 'eval/latency_6/demon_fire_f1': 0.9639596347909658, 'eval/latency_6/demon_fire_rate_model': 0.8761825264329438, 'eval/latency_6/demon_fire_rate_teacher': 0.860879243183083, 'eval/latency_6/demon_left_precision': 0.8180039138943248, 'eval/latency_6/demon_left_recall': 0.7333333333333333, 'eval/latency_6/demon_left_f1': 0.7733580018501387, 'eval/latency_6/demon_left_rate_model': 0.14218141346688926, 'eval/latency_6/demon_left_rate_teacher': 0.15859766277128548, 'eval/latency_6/demon_right_precision': 0.8150572831423896, 'eval/latency_6/demon_right_recall': 0.7649769585253456, 'eval/latency_6/demon_right_f1': 0.7892234548335976, 'eval/latency_6/demon_right_rate_model': 0.17000556483027268, 'eval/latency_6/demon_right_rate_teacher': 0.18113522537562604, 'eval/latency_6/demon_move_f1': 0.7812907283418682, 'eval/latency_6/demon_cc_f1': 0.9007779830157827, 'timing/data': 0.0006701387465000153, 'timing/model': 0.9255981296300888, 'learning_rate/qwen_vl_interface': 4.283583937999991e-06, 'learning_rate/action_model': 4.2835839379999906e-05, 'epoch': 1.65}) +[2026-06-17 22:58:31,065][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_4000_state +[2026-06-17 22:58:31,070][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 22:58:37,658][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_4000_state/model.safetensors +[2026-06-17 22:58:53,574][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_4000_state/optimizer.bin +[2026-06-17 22:58:53,576][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_4000_state/sampler.bin +[2026-06-17 22:58:53,577][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_4000_state/sampler_1.bin +[2026-06-17 22:58:53,580][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_4000_state/random_states_0.pkl +[2026-06-17 22:58:53,583][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4000 +[2026-06-17 22:58:53,583][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 22:58:53,593][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 22:58:57,406][starVLA.training.train_starvla][INFO] - Step 4001, Loss: {'train/loss': 0.2840680778026581, 'train/grad_norm_pre_clip': 13.552545547485352, 'timing/data': 0.0006413329392671585, 'timing/model': 0.9294381029903889, 'learning_rate/qwen_vl_interface': 4.2814665822946825e-06, 'learning_rate/action_model': 4.281466582294682e-05, 'epoch': 1.65}) +[2026-06-17 22:59:01,257][starVLA.training.train_starvla][INFO] - Step 4002, Loss: {'train/loss': 0.3654455542564392, 'train/grad_norm_pre_clip': 18.535804748535156, 'timing/data': 0.0007833428680896759, 'timing/model': 0.9248537393286824, 'learning_rate/qwen_vl_interface': 4.279349427367395e-06, 'learning_rate/action_model': 4.279349427367395e-05, 'epoch': 1.65}) +[2026-06-17 22:59:05,150][starVLA.training.train_starvla][INFO] - Step 4003, Loss: {'train/loss': 0.5170198678970337, 'train/grad_norm_pre_clip': 33.08955001831055, 'timing/data': 0.0005591278895735741, 'timing/model': 0.9364361576735973, 'learning_rate/qwen_vl_interface': 4.277232473657018e-06, 'learning_rate/action_model': 4.277232473657018e-05, 'epoch': 1.65}) +[2026-06-17 22:59:08,930][starVLA.training.train_starvla][INFO] - Step 4004, Loss: {'train/loss': 0.4055134952068329, 'train/grad_norm_pre_clip': 14.134294509887695, 'timing/data': 0.0004909848794341087, 'timing/model': 0.9331147344782948, 'learning_rate/qwen_vl_interface': 4.2751157216023934e-06, 'learning_rate/action_model': 4.2751157216023936e-05, 'epoch': 1.65}) +[2026-06-17 22:59:12,729][starVLA.training.train_starvla][INFO] - Step 4005, Loss: {'train/loss': 0.3662708103656769, 'train/grad_norm_pre_clip': 14.72257137298584, 'timing/data': 0.0006996355950832367, 'timing/model': 1.1041788486763835, 'learning_rate/qwen_vl_interface': 4.272999171642329e-06, 'learning_rate/action_model': 4.272999171642329e-05, 'epoch': 1.65}) +[2026-06-17 22:59:16,307][starVLA.training.train_starvla][INFO] - Step 4006, Loss: {'train/loss': 0.4996856153011322, 'train/grad_norm_pre_clip': 23.4626407623291, 'timing/data': 0.0005243578925728798, 'timing/model': 0.9285396942868829, 'learning_rate/qwen_vl_interface': 4.270882824215588e-06, 'learning_rate/action_model': 4.270882824215588e-05, 'epoch': 1.65}) +[2026-06-17 22:59:20,046][starVLA.training.train_starvla][INFO] - Step 4007, Loss: {'train/loss': 0.4592549800872803, 'train/grad_norm_pre_clip': 16.04935073852539, 'timing/data': 0.0004894854500889778, 'timing/model': 0.9275523256510496, 'learning_rate/qwen_vl_interface': 4.268766679760891e-06, 'learning_rate/action_model': 4.268766679760891e-05, 'epoch': 1.65}) +[2026-06-17 22:59:23,804][starVLA.training.train_starvla][INFO] - Step 4008, Loss: {'train/loss': 0.19751133024692535, 'train/grad_norm_pre_clip': 11.70539379119873, 'timing/data': 0.0005159629508852959, 'timing/model': 0.9307694910094142, 'learning_rate/qwen_vl_interface': 4.266650738716916e-06, 'learning_rate/action_model': 4.266650738716916e-05, 'epoch': 1.65}) +[2026-06-17 22:59:27,549][starVLA.training.train_starvla][INFO] - Step 4009, Loss: {'train/loss': 0.3700033128261566, 'train/grad_norm_pre_clip': 16.698028564453125, 'timing/data': 0.0004873480647802353, 'timing/model': 0.9223785642534494, 'learning_rate/qwen_vl_interface': 4.264535001522302e-06, 'learning_rate/action_model': 4.2645350015223025e-05, 'epoch': 1.65}) +[2026-06-17 22:59:31,257][starVLA.training.train_starvla][INFO] - Step 4010, Loss: {'train/loss': 0.6841505169868469, 'train/grad_norm_pre_clip': 27.195735931396484, 'timing/data': 0.0004839915782213211, 'timing/model': 0.919566142372787, 'learning_rate/qwen_vl_interface': 4.262419468615644e-06, 'learning_rate/action_model': 4.2624194686156435e-05, 'epoch': 1.65}) +[2026-06-17 22:59:35,003][starVLA.training.train_starvla][INFO] - Step 4011, Loss: {'train/loss': 0.27700236439704895, 'train/grad_norm_pre_clip': 9.651826858520508, 'timing/data': 0.0005447911098599434, 'timing/model': 1.100893585011363, 'learning_rate/qwen_vl_interface': 4.2603041404354925e-06, 'learning_rate/action_model': 4.260304140435492e-05, 'epoch': 1.65}) +[2026-06-17 22:59:38,585][starVLA.training.train_starvla][INFO] - Step 4012, Loss: {'train/loss': 0.4827095866203308, 'train/grad_norm_pre_clip': 24.370445251464844, 'timing/data': 0.0004893019795417786, 'timing/model': 0.9217753876000643, 'learning_rate/qwen_vl_interface': 4.258189017420356e-06, 'learning_rate/action_model': 4.258189017420355e-05, 'epoch': 1.65}) +[2026-06-17 22:59:42,311][starVLA.training.train_starvla][INFO] - Step 4013, Loss: {'train/loss': 0.4375773072242737, 'train/grad_norm_pre_clip': 20.8713321685791, 'timing/data': 0.0005032038316130638, 'timing/model': 0.923967270180583, 'learning_rate/qwen_vl_interface': 4.2560741000087035e-06, 'learning_rate/action_model': 4.2560741000087034e-05, 'epoch': 1.65}) +[2026-06-17 22:59:46,029][starVLA.training.train_starvla][INFO] - Step 4014, Loss: {'train/loss': 0.26579591631889343, 'train/grad_norm_pre_clip': 15.108028411865234, 'timing/data': 0.0005069514736533165, 'timing/model': 0.9297404587268829, 'learning_rate/qwen_vl_interface': 4.253959388638961e-06, 'learning_rate/action_model': 4.2539593886389604e-05, 'epoch': 1.65}) +[2026-06-17 22:59:49,798][starVLA.training.train_starvla][INFO] - Step 4015, Loss: {'train/loss': 0.38695740699768066, 'train/grad_norm_pre_clip': 41.19184875488281, 'timing/data': 0.0005539180710911751, 'timing/model': 0.9245885889977217, 'learning_rate/qwen_vl_interface': 4.251844883749507e-06, 'learning_rate/action_model': 4.2518448837495066e-05, 'epoch': 1.65}) +[2026-06-17 22:59:53,561][starVLA.training.train_starvla][INFO] - Step 4016, Loss: {'train/loss': 0.28911229968070984, 'train/grad_norm_pre_clip': 12.925655364990234, 'timing/data': 0.0017068851739168167, 'timing/model': 0.9221797995269299, 'learning_rate/qwen_vl_interface': 4.249730585778684e-06, 'learning_rate/action_model': 4.2497305857786834e-05, 'epoch': 1.65}) +[2026-06-17 22:59:57,302][starVLA.training.train_starvla][INFO] - Step 4017, Loss: {'train/loss': 0.4397886097431183, 'train/grad_norm_pre_clip': 71.28910827636719, 'timing/data': 0.0005631735548377037, 'timing/model': 1.0979447346180677, 'learning_rate/qwen_vl_interface': 4.247616495164788e-06, 'learning_rate/action_model': 4.247616495164787e-05, 'epoch': 1.65}) +[2026-06-17 23:00:00,967][starVLA.training.train_starvla][INFO] - Step 4018, Loss: {'train/loss': 0.39187127351760864, 'train/grad_norm_pre_clip': 20.38765525817871, 'timing/data': 0.00047764554619789124, 'timing/model': 1.0273766769096255, 'learning_rate/qwen_vl_interface': 4.24550261234607e-06, 'learning_rate/action_model': 4.2455026123460694e-05, 'epoch': 1.65}) +[2026-06-17 23:00:04,554][starVLA.training.train_starvla][INFO] - Step 4019, Loss: {'train/loss': 0.2358422875404358, 'train/grad_norm_pre_clip': 14.33028793334961, 'timing/data': 0.0008080275729298592, 'timing/model': 0.9260296653956175, 'learning_rate/qwen_vl_interface': 4.2433889377607405e-06, 'learning_rate/action_model': 4.24338893776074e-05, 'epoch': 1.65}) +[2026-06-17 23:00:08,287][starVLA.training.train_starvla][INFO] - Step 4020, Loss: {'train/loss': 0.15058930218219757, 'train/grad_norm_pre_clip': 8.838974952697754, 'timing/data': 0.0005045915022492409, 'timing/model': 0.9256760720163584, 'learning_rate/qwen_vl_interface': 4.241275471846968e-06, 'learning_rate/action_model': 4.241275471846968e-05, 'epoch': 1.65}) +[2026-06-17 23:00:12,077][starVLA.training.train_starvla][INFO] - Step 4021, Loss: {'train/loss': 0.47722557187080383, 'train/grad_norm_pre_clip': 17.979387283325195, 'timing/data': 0.000944855622947216, 'timing/model': 0.9407066255807877, 'learning_rate/qwen_vl_interface': 4.239162215042876e-06, 'learning_rate/action_model': 4.239162215042876e-05, 'epoch': 1.66}) +[2026-06-17 23:00:15,836][starVLA.training.train_starvla][INFO] - Step 4022, Loss: {'train/loss': 0.41494181752204895, 'train/grad_norm_pre_clip': 20.39476203918457, 'timing/data': 0.0005568414926528931, 'timing/model': 0.9346074424684048, 'learning_rate/qwen_vl_interface': 4.237049167786544e-06, 'learning_rate/action_model': 4.2370491677865444e-05, 'epoch': 1.66}) +[2026-06-17 23:00:19,567][starVLA.training.train_starvla][INFO] - Step 4023, Loss: {'train/loss': 0.30864477157592773, 'train/grad_norm_pre_clip': 16.320470809936523, 'timing/data': 0.0005297446623444557, 'timing/model': 0.9221691703423858, 'learning_rate/qwen_vl_interface': 4.234936330516009e-06, 'learning_rate/action_model': 4.234936330516009e-05, 'epoch': 1.66}) +[2026-06-17 23:00:23,327][starVLA.training.train_starvla][INFO] - Step 4024, Loss: {'train/loss': 0.2911761701107025, 'train/grad_norm_pre_clip': 18.165645599365234, 'timing/data': 0.00047925394028425217, 'timing/model': 1.120212234556675, 'learning_rate/qwen_vl_interface': 4.2328237036692675e-06, 'learning_rate/action_model': 4.232823703669267e-05, 'epoch': 1.66}) +[2026-06-17 23:00:26,917][starVLA.training.train_starvla][INFO] - Step 4025, Loss: {'train/loss': 0.4343833029270172, 'train/grad_norm_pre_clip': 18.98440933227539, 'timing/data': 0.0005176588892936707, 'timing/model': 0.933565248735249, 'learning_rate/qwen_vl_interface': 4.2307112876842645e-06, 'learning_rate/action_model': 4.2307112876842645e-05, 'epoch': 1.66}) +[2026-06-17 23:00:30,687][starVLA.training.train_starvla][INFO] - Step 4026, Loss: {'train/loss': 0.10657592862844467, 'train/grad_norm_pre_clip': 7.632449626922607, 'timing/data': 0.0005514863878488541, 'timing/model': 0.9437686689198017, 'learning_rate/qwen_vl_interface': 4.228599082998908e-06, 'learning_rate/action_model': 4.228599082998908e-05, 'epoch': 1.66}) +[2026-06-17 23:00:34,498][starVLA.training.train_starvla][INFO] - Step 4027, Loss: {'train/loss': 0.3823226988315582, 'train/grad_norm_pre_clip': 17.243776321411133, 'timing/data': 0.0004896847531199455, 'timing/model': 0.9265428837388754, 'learning_rate/qwen_vl_interface': 4.226487090051062e-06, 'learning_rate/action_model': 4.2264870900510624e-05, 'epoch': 1.66}) +[2026-06-17 23:00:38,247][starVLA.training.train_starvla][INFO] - Step 4028, Loss: {'train/loss': 0.32333335280418396, 'train/grad_norm_pre_clip': 23.842226028442383, 'timing/data': 0.0004989895969629288, 'timing/model': 0.9251333894208074, 'learning_rate/qwen_vl_interface': 4.224375309278543e-06, 'learning_rate/action_model': 4.2243753092785434e-05, 'epoch': 1.66}) +[2026-06-17 23:00:42,013][starVLA.training.train_starvla][INFO] - Step 4029, Loss: {'train/loss': 0.3714837431907654, 'train/grad_norm_pre_clip': 9.262856483459473, 'timing/data': 0.0005006445571780205, 'timing/model': 0.9240562152117491, 'learning_rate/qwen_vl_interface': 4.222263741119127e-06, 'learning_rate/action_model': 4.222263741119126e-05, 'epoch': 1.66}) +[2026-06-17 23:00:45,786][starVLA.training.train_starvla][INFO] - Step 4030, Loss: {'train/loss': 0.41882920265197754, 'train/grad_norm_pre_clip': 14.48713207244873, 'timing/data': 0.0005076350644230843, 'timing/model': 1.1189125925302505, 'learning_rate/qwen_vl_interface': 4.220152386010541e-06, 'learning_rate/action_model': 4.2201523860105405e-05, 'epoch': 1.66}) +[2026-06-17 23:00:49,528][starVLA.training.train_starvla][INFO] - Step 4031, Loss: {'train/loss': 0.5832586884498596, 'train/grad_norm_pre_clip': 20.317886352539062, 'timing/data': 0.0005429647862911224, 'timing/model': 1.0912475865334272, 'learning_rate/qwen_vl_interface': 4.218041244390476e-06, 'learning_rate/action_model': 4.218041244390475e-05, 'epoch': 1.66}) +[2026-06-17 23:00:53,118][starVLA.training.train_starvla][INFO] - Step 4032, Loss: {'train/loss': 0.489154577255249, 'train/grad_norm_pre_clip': 16.8300838470459, 'timing/data': 0.0014499705284833908, 'timing/model': 0.9221926433965564, 'learning_rate/qwen_vl_interface': 4.215930316696571e-06, 'learning_rate/action_model': 4.215930316696571e-05, 'epoch': 1.66}) +[2026-06-17 23:00:56,895][starVLA.training.train_starvla][INFO] - Step 4033, Loss: {'train/loss': 0.5473682880401611, 'train/grad_norm_pre_clip': 19.580930709838867, 'timing/data': 0.0005752993747591972, 'timing/model': 0.9288496542721987, 'learning_rate/qwen_vl_interface': 4.213819603366424e-06, 'learning_rate/action_model': 4.213819603366424e-05, 'epoch': 1.66}) +[2026-06-17 23:01:00,697][starVLA.training.train_starvla][INFO] - Step 4034, Loss: {'train/loss': 0.14091652631759644, 'train/grad_norm_pre_clip': 12.35683536529541, 'timing/data': 0.0005027400329709053, 'timing/model': 0.9739541430026293, 'learning_rate/qwen_vl_interface': 4.211709104837587e-06, 'learning_rate/action_model': 4.211709104837587e-05, 'epoch': 1.66}) +[2026-06-17 23:01:04,449][starVLA.training.train_starvla][INFO] - Step 4035, Loss: {'train/loss': 0.33457061648368835, 'train/grad_norm_pre_clip': 14.833041191101074, 'timing/data': 0.0005138209089636803, 'timing/model': 0.9289714079350233, 'learning_rate/qwen_vl_interface': 4.2095988215475705e-06, 'learning_rate/action_model': 4.20959882154757e-05, 'epoch': 1.66}) +[2026-06-17 23:01:08,214][starVLA.training.train_starvla][INFO] - Step 4036, Loss: {'train/loss': 0.2153056263923645, 'train/grad_norm_pre_clip': 12.94868278503418, 'timing/data': 0.0012138551101088524, 'timing/model': 0.929945626296103, 'learning_rate/qwen_vl_interface': 4.207488753933838e-06, 'learning_rate/action_model': 4.207488753933838e-05, 'epoch': 1.66}) +[2026-06-17 23:01:11,979][starVLA.training.train_starvla][INFO] - Step 4037, Loss: {'train/loss': 0.17901206016540527, 'train/grad_norm_pre_clip': 15.825299263000488, 'timing/data': 0.0005837613716721535, 'timing/model': 1.1020281249657273, 'learning_rate/qwen_vl_interface': 4.205378902433807e-06, 'learning_rate/action_model': 4.205378902433807e-05, 'epoch': 1.66}) +[2026-06-17 23:01:15,686][starVLA.training.train_starvla][INFO] - Step 4038, Loss: {'train/loss': 0.7785171270370483, 'train/grad_norm_pre_clip': 33.24751281738281, 'timing/data': 0.00048563070595264435, 'timing/model': 1.0497444309294224, 'learning_rate/qwen_vl_interface': 4.203269267484856e-06, 'learning_rate/action_model': 4.203269267484856e-05, 'epoch': 1.66}) +[2026-06-17 23:01:19,250][starVLA.training.train_starvla][INFO] - Step 4039, Loss: {'train/loss': 0.23827330768108368, 'train/grad_norm_pre_clip': 14.350103378295898, 'timing/data': 0.0005811993032693863, 'timing/model': 0.924487641081214, 'learning_rate/qwen_vl_interface': 4.201159849524313e-06, 'learning_rate/action_model': 4.201159849524312e-05, 'epoch': 1.66}) +[2026-06-17 23:01:22,998][starVLA.training.train_starvla][INFO] - Step 4040, Loss: {'train/loss': 0.4180136024951935, 'train/grad_norm_pre_clip': 17.834449768066406, 'timing/data': 0.00048002786934375763, 'timing/model': 0.9228218561038375, 'learning_rate/qwen_vl_interface': 4.19905064898946e-06, 'learning_rate/action_model': 4.199050648989459e-05, 'epoch': 1.66}) +[2026-06-17 23:01:26,738][starVLA.training.train_starvla][INFO] - Step 4041, Loss: {'train/loss': 0.293843537569046, 'train/grad_norm_pre_clip': 15.361129760742188, 'timing/data': 0.0011728489771485329, 'timing/model': 0.9252557186409831, 'learning_rate/qwen_vl_interface': 4.196941666317537e-06, 'learning_rate/action_model': 4.196941666317537e-05, 'epoch': 1.66}) +[2026-06-17 23:01:30,421][starVLA.training.train_starvla][INFO] - Step 4042, Loss: {'train/loss': 0.5117283463478088, 'train/grad_norm_pre_clip': 17.263179779052734, 'timing/data': 0.0004838407039642334, 'timing/model': 0.928378077223897, 'learning_rate/qwen_vl_interface': 4.194832901945741e-06, 'learning_rate/action_model': 4.194832901945741e-05, 'epoch': 1.66}) +[2026-06-17 23:01:34,152][starVLA.training.train_starvla][INFO] - Step 4043, Loss: {'train/loss': 0.27938687801361084, 'train/grad_norm_pre_clip': 15.86734676361084, 'timing/data': 0.0015043634921312332, 'timing/model': 0.9290356589481235, 'learning_rate/qwen_vl_interface': 4.192724356311221e-06, 'learning_rate/action_model': 4.19272435631122e-05, 'epoch': 1.66}) +[2026-06-17 23:01:37,911][starVLA.training.train_starvla][INFO] - Step 4044, Loss: {'train/loss': 0.41325753927230835, 'train/grad_norm_pre_clip': 19.286569595336914, 'timing/data': 0.0005090860649943352, 'timing/model': 1.0961805107071996, 'learning_rate/qwen_vl_interface': 4.190616029851079e-06, 'learning_rate/action_model': 4.190616029851079e-05, 'epoch': 1.66}) +[2026-06-17 23:01:41,499][starVLA.training.train_starvla][INFO] - Step 4045, Loss: {'train/loss': 0.3285619616508484, 'train/grad_norm_pre_clip': 18.143089294433594, 'timing/data': 0.0005007050931453705, 'timing/model': 0.924122279509902, 'learning_rate/qwen_vl_interface': 4.1885079230023715e-06, 'learning_rate/action_model': 4.1885079230023715e-05, 'epoch': 1.67}) +[2026-06-17 23:01:45,246][starVLA.training.train_starvla][INFO] - Step 4046, Loss: {'train/loss': 0.5916572213172913, 'train/grad_norm_pre_clip': 20.328561782836914, 'timing/data': 0.0005021346732974052, 'timing/model': 0.927506766282022, 'learning_rate/qwen_vl_interface': 4.186400036202117e-06, 'learning_rate/action_model': 4.1864000362021175e-05, 'epoch': 1.67}) +[2026-06-17 23:01:48,951][starVLA.training.train_starvla][INFO] - Step 4047, Loss: {'train/loss': 0.34725576639175415, 'train/grad_norm_pre_clip': 18.15951919555664, 'timing/data': 0.0004947753623127937, 'timing/model': 0.9230520743876696, 'learning_rate/qwen_vl_interface': 4.184292369887279e-06, 'learning_rate/action_model': 4.1842923698872785e-05, 'epoch': 1.67}) +[2026-06-17 23:01:52,749][starVLA.training.train_starvla][INFO] - Step 4048, Loss: {'train/loss': 0.32852062582969666, 'train/grad_norm_pre_clip': 13.840885162353516, 'timing/data': 0.0005808798596262932, 'timing/model': 0.9247889779508114, 'learning_rate/qwen_vl_interface': 4.182184924494778e-06, 'learning_rate/action_model': 4.1821849244947773e-05, 'epoch': 1.67}) +[2026-06-17 23:01:56,512][starVLA.training.train_starvla][INFO] - Step 4049, Loss: {'train/loss': 0.2736840844154358, 'train/grad_norm_pre_clip': 10.919026374816895, 'timing/data': 0.0005063265562057495, 'timing/model': 0.9212175663560629, 'learning_rate/qwen_vl_interface': 4.1800777004614915e-06, 'learning_rate/action_model': 4.180077700461492e-05, 'epoch': 1.67}) +[2026-06-17 23:02:00,268][starVLA.training.train_starvla][INFO] - Step 4050, Loss: {'train/loss': 0.3737265467643738, 'train/grad_norm_pre_clip': 20.593358993530273, 'timing/data': 0.0005249427631497383, 'timing/model': 1.0952589279040694, 'learning_rate/qwen_vl_interface': 4.177970698224249e-06, 'learning_rate/action_model': 4.17797069822425e-05, 'epoch': 1.67}) +[2026-06-17 23:02:03,953][starVLA.training.train_starvla][INFO] - Step 4051, Loss: {'train/loss': 0.19373154640197754, 'train/grad_norm_pre_clip': 11.242236137390137, 'timing/data': 0.000497584231197834, 'timing/model': 1.038561349734664, 'learning_rate/qwen_vl_interface': 4.175863918219836e-06, 'learning_rate/action_model': 4.175863918219836e-05, 'epoch': 1.67}) +[2026-06-17 23:02:07,544][starVLA.training.train_starvla][INFO] - Step 4052, Loss: {'train/loss': 0.3442426323890686, 'train/grad_norm_pre_clip': 18.201318740844727, 'timing/data': 0.0005593365058302879, 'timing/model': 0.9354082671925426, 'learning_rate/qwen_vl_interface': 4.173757360884986e-06, 'learning_rate/action_model': 4.173757360884986e-05, 'epoch': 1.67}) +[2026-06-17 23:02:11,293][starVLA.training.train_starvla][INFO] - Step 4053, Loss: {'train/loss': 0.3829531669616699, 'train/grad_norm_pre_clip': 19.12437629699707, 'timing/data': 0.0005271760746836662, 'timing/model': 0.9297296218574047, 'learning_rate/qwen_vl_interface': 4.171651026656396e-06, 'learning_rate/action_model': 4.171651026656395e-05, 'epoch': 1.67}) +[2026-06-17 23:02:15,023][starVLA.training.train_starvla][INFO] - Step 4054, Loss: {'train/loss': 0.6715226769447327, 'train/grad_norm_pre_clip': 25.343734741210938, 'timing/data': 0.0009105335921049118, 'timing/model': 0.9272335702553391, 'learning_rate/qwen_vl_interface': 4.1695449159707066e-06, 'learning_rate/action_model': 4.169544915970707e-05, 'epoch': 1.67}) +[2026-06-17 23:02:18,776][starVLA.training.train_starvla][INFO] - Step 4055, Loss: {'train/loss': 0.22986379265785217, 'train/grad_norm_pre_clip': 13.964374542236328, 'timing/data': 0.0005064206197857857, 'timing/model': 0.9362993575632572, 'learning_rate/qwen_vl_interface': 4.167439029264519e-06, 'learning_rate/action_model': 4.167439029264519e-05, 'epoch': 1.67}) +[2026-06-17 23:02:22,677][starVLA.training.train_starvla][INFO] - Step 4056, Loss: {'train/loss': 0.38610610365867615, 'train/grad_norm_pre_clip': 16.96954345703125, 'timing/data': 0.0005710422992706299, 'timing/model': 0.9353586472570896, 'learning_rate/qwen_vl_interface': 4.165333366974385e-06, 'learning_rate/action_model': 4.165333366974385e-05, 'epoch': 1.67}) +[2026-06-17 23:02:26,478][starVLA.training.train_starvla][INFO] - Step 4057, Loss: {'train/loss': 0.26679345965385437, 'train/grad_norm_pre_clip': 17.08754539489746, 'timing/data': 0.0004907194525003433, 'timing/model': 1.1123440312221646, 'learning_rate/qwen_vl_interface': 4.163227929536812e-06, 'learning_rate/action_model': 4.163227929536812e-05, 'epoch': 1.67}) +[2026-06-17 23:02:30,302][starVLA.training.train_starvla][INFO] - Step 4058, Loss: {'train/loss': 0.2875842750072479, 'train/grad_norm_pre_clip': 11.69217300415039, 'timing/data': 0.0012610647827386856, 'timing/model': 1.088414098136127, 'learning_rate/qwen_vl_interface': 4.1611227173882596e-06, 'learning_rate/action_model': 4.1611227173882596e-05, 'epoch': 1.67}) +[2026-06-17 23:02:34,014][starVLA.training.train_starvla][INFO] - Step 4059, Loss: {'train/loss': 0.15923061966896057, 'train/grad_norm_pre_clip': 10.361520767211914, 'timing/data': 0.0006358008831739426, 'timing/model': 0.9377534976229072, 'learning_rate/qwen_vl_interface': 4.1590177309651365e-06, 'learning_rate/action_model': 4.159017730965136e-05, 'epoch': 1.67}) +[2026-06-17 23:02:37,833][starVLA.training.train_starvla][INFO] - Step 4060, Loss: {'train/loss': 0.19877570867538452, 'train/grad_norm_pre_clip': 11.135872840881348, 'timing/data': 0.0005127694457769394, 'timing/model': 0.9514772742986679, 'learning_rate/qwen_vl_interface': 4.156912970703815e-06, 'learning_rate/action_model': 4.1569129707038155e-05, 'epoch': 1.67}) +[2026-06-17 23:02:41,668][starVLA.training.train_starvla][INFO] - Step 4061, Loss: {'train/loss': 0.32180431485176086, 'train/grad_norm_pre_clip': 17.679237365722656, 'timing/data': 0.0005133207887411118, 'timing/model': 0.9350110637024045, 'learning_rate/qwen_vl_interface': 4.15480843704061e-06, 'learning_rate/action_model': 4.154808437040609e-05, 'epoch': 1.67}) +[2026-06-17 23:02:45,437][starVLA.training.train_starvla][INFO] - Step 4062, Loss: {'train/loss': 0.17294615507125854, 'train/grad_norm_pre_clip': 14.655984878540039, 'timing/data': 0.0007150303572416306, 'timing/model': 0.9315369194373488, 'learning_rate/qwen_vl_interface': 4.152704130411794e-06, 'learning_rate/action_model': 4.152704130411793e-05, 'epoch': 1.67}) +[2026-06-17 23:02:49,195][starVLA.training.train_starvla][INFO] - Step 4063, Loss: {'train/loss': 0.2982679307460785, 'train/grad_norm_pre_clip': 20.398303985595703, 'timing/data': 0.0004977835342288017, 'timing/model': 0.9246032563969493, 'learning_rate/qwen_vl_interface': 4.150600051253591e-06, 'learning_rate/action_model': 4.150600051253591e-05, 'epoch': 1.67}) +[2026-06-17 23:02:52,943][starVLA.training.train_starvla][INFO] - Step 4064, Loss: {'train/loss': 0.46579432487487793, 'train/grad_norm_pre_clip': 24.063262939453125, 'timing/data': 0.0005053281784057617, 'timing/model': 1.0956961195915937, 'learning_rate/qwen_vl_interface': 4.148496200002183e-06, 'learning_rate/action_model': 4.148496200002183e-05, 'epoch': 1.67}) +[2026-06-17 23:02:56,619][starVLA.training.train_starvla][INFO] - Step 4065, Loss: {'train/loss': 0.44150570034980774, 'train/grad_norm_pre_clip': 17.232463836669922, 'timing/data': 0.000510278157889843, 'timing/model': 1.0277397204190493, 'learning_rate/qwen_vl_interface': 4.146392577093697e-06, 'learning_rate/action_model': 4.1463925770936974e-05, 'epoch': 1.67}) +[2026-06-17 23:03:00,185][starVLA.training.train_starvla][INFO] - Step 4066, Loss: {'train/loss': 0.1947559118270874, 'train/grad_norm_pre_clip': 19.89833641052246, 'timing/data': 0.0007630493491888046, 'timing/model': 0.9239759407937527, 'learning_rate/qwen_vl_interface': 4.144289182964218e-06, 'learning_rate/action_model': 4.144289182964218e-05, 'epoch': 1.67}) +[2026-06-17 23:03:04,024][starVLA.training.train_starvla][INFO] - Step 4067, Loss: {'train/loss': 0.5746000409126282, 'train/grad_norm_pre_clip': 21.66286277770996, 'timing/data': 0.0005862535908818245, 'timing/model': 0.9429151387885213, 'learning_rate/qwen_vl_interface': 4.14218601804978e-06, 'learning_rate/action_model': 4.1421860180497796e-05, 'epoch': 1.67}) +[2026-06-17 23:03:07,801][starVLA.training.train_starvla][INFO] - Step 4068, Loss: {'train/loss': 0.5231717228889465, 'train/grad_norm_pre_clip': 16.968082427978516, 'timing/data': 0.0005103545263409615, 'timing/model': 0.9247169429436326, 'learning_rate/qwen_vl_interface': 4.1400830827863725e-06, 'learning_rate/action_model': 4.1400830827863725e-05, 'epoch': 1.67}) +[2026-06-17 23:03:11,531][starVLA.training.train_starvla][INFO] - Step 4069, Loss: {'train/loss': 0.24337269365787506, 'train/grad_norm_pre_clip': 15.600707054138184, 'timing/data': 0.0004999246448278427, 'timing/model': 0.9243090664967895, 'learning_rate/qwen_vl_interface': 4.137980377609937e-06, 'learning_rate/action_model': 4.137980377609937e-05, 'epoch': 1.67}) +[2026-06-17 23:03:15,271][starVLA.training.train_starvla][INFO] - Step 4070, Loss: {'train/loss': 0.364216148853302, 'train/grad_norm_pre_clip': 18.112831115722656, 'timing/data': 0.0004885857924818993, 'timing/model': 1.0889127431437373, 'learning_rate/qwen_vl_interface': 4.135877902956364e-06, 'learning_rate/action_model': 4.1358779029563636e-05, 'epoch': 1.68}) +[2026-06-17 23:03:18,959][starVLA.training.train_starvla][INFO] - Step 4071, Loss: {'train/loss': 0.35214555263519287, 'train/grad_norm_pre_clip': 11.988487243652344, 'timing/data': 0.0005206027999520302, 'timing/model': 1.0315974280238152, 'learning_rate/qwen_vl_interface': 4.133775659261503e-06, 'learning_rate/action_model': 4.133775659261502e-05, 'epoch': 1.68}) +[2026-06-17 23:03:22,585][starVLA.training.train_starvla][INFO] - Step 4072, Loss: {'train/loss': 0.5272614359855652, 'train/grad_norm_pre_clip': 17.202756881713867, 'timing/data': 0.0005908375605940819, 'timing/model': 0.9336546929553151, 'learning_rate/qwen_vl_interface': 4.131673646961149e-06, 'learning_rate/action_model': 4.131673646961149e-05, 'epoch': 1.68}) +[2026-06-17 23:03:26,311][starVLA.training.train_starvla][INFO] - Step 4073, Loss: {'train/loss': 0.5092296004295349, 'train/grad_norm_pre_clip': 14.464781761169434, 'timing/data': 0.0004990715533494949, 'timing/model': 0.9291771370917559, 'learning_rate/qwen_vl_interface': 4.129571866491051e-06, 'learning_rate/action_model': 4.1295718664910504e-05, 'epoch': 1.68}) +[2026-06-17 23:03:30,047][starVLA.training.train_starvla][INFO] - Step 4074, Loss: {'train/loss': 0.5479961037635803, 'train/grad_norm_pre_clip': 19.46027183532715, 'timing/data': 0.0005105277523398399, 'timing/model': 0.9198134234175086, 'learning_rate/qwen_vl_interface': 4.127470318286908e-06, 'learning_rate/action_model': 4.127470318286908e-05, 'epoch': 1.68}) +[2026-06-17 23:03:33,799][starVLA.training.train_starvla][INFO] - Step 4075, Loss: {'train/loss': 0.4601244628429413, 'train/grad_norm_pre_clip': 21.66744041442871, 'timing/data': 0.0005112718790769577, 'timing/model': 0.9256718307733536, 'learning_rate/qwen_vl_interface': 4.125369002784378e-06, 'learning_rate/action_model': 4.1253690027843776e-05, 'epoch': 1.68}) +[2026-06-17 23:03:37,513][starVLA.training.train_starvla][INFO] - Step 4076, Loss: {'train/loss': 0.2646341919898987, 'train/grad_norm_pre_clip': 14.143538475036621, 'timing/data': 0.0005094008520245552, 'timing/model': 0.9312393199652433, 'learning_rate/qwen_vl_interface': 4.123267920419064e-06, 'learning_rate/action_model': 4.123267920419064e-05, 'epoch': 1.68}) +[2026-06-17 23:03:41,281][starVLA.training.train_starvla][INFO] - Step 4077, Loss: {'train/loss': 0.11894156038761139, 'train/grad_norm_pre_clip': 6.532386302947998, 'timing/data': 0.000808090902864933, 'timing/model': 1.1124361855909228, 'learning_rate/qwen_vl_interface': 4.121167071626522e-06, 'learning_rate/action_model': 4.121167071626522e-05, 'epoch': 1.68}) +[2026-06-17 23:03:44,869][starVLA.training.train_starvla][INFO] - Step 4078, Loss: {'train/loss': 0.4973601698875427, 'train/grad_norm_pre_clip': 19.88508415222168, 'timing/data': 0.0008172402158379555, 'timing/model': 0.9324547694995999, 'learning_rate/qwen_vl_interface': 4.119066456842261e-06, 'learning_rate/action_model': 4.119066456842261e-05, 'epoch': 1.68}) +[2026-06-17 23:03:48,642][starVLA.training.train_starvla][INFO] - Step 4079, Loss: {'train/loss': 0.2900678217411041, 'train/grad_norm_pre_clip': 14.180620193481445, 'timing/data': 0.0008356580510735512, 'timing/model': 0.9361385051161051, 'learning_rate/qwen_vl_interface': 4.1169660765017416e-06, 'learning_rate/action_model': 4.1169660765017414e-05, 'epoch': 1.68}) +[2026-06-17 23:03:52,710][starVLA.training.train_starvla][INFO] - Step 4080, Loss: {'train/loss': 0.3358435034751892, 'train/grad_norm_pre_clip': 13.529279708862305, 'timing/data': 0.00047491583973169327, 'timing/model': 1.2080986145883799, 'learning_rate/qwen_vl_interface': 4.1148659310403725e-06, 'learning_rate/action_model': 4.1148659310403723e-05, 'epoch': 1.68}) +[2026-06-17 23:03:56,487][starVLA.training.train_starvla][INFO] - Step 4081, Loss: {'train/loss': 0.17031385004520416, 'train/grad_norm_pre_clip': 12.161844253540039, 'timing/data': 0.0006499402225017548, 'timing/model': 0.9581233132630587, 'learning_rate/qwen_vl_interface': 4.112766020893516e-06, 'learning_rate/action_model': 4.1127660208935164e-05, 'epoch': 1.68}) +[2026-06-17 23:04:00,725][starVLA.training.train_starvla][INFO] - Step 4082, Loss: {'train/loss': 0.3144401013851166, 'train/grad_norm_pre_clip': 15.534146308898926, 'timing/data': 0.0004952438175678253, 'timing/model': 0.9306729938834906, 'learning_rate/qwen_vl_interface': 4.11066634649649e-06, 'learning_rate/action_model': 4.1106663464964895e-05, 'epoch': 1.68}) +[2026-06-17 23:04:04,768][starVLA.training.train_starvla][INFO] - Step 4083, Loss: {'train/loss': 0.35808035731315613, 'train/grad_norm_pre_clip': 17.083757400512695, 'timing/data': 0.0012779990211129189, 'timing/model': 1.3655228801071644, 'learning_rate/qwen_vl_interface': 4.1085669082845545e-06, 'learning_rate/action_model': 4.108566908284555e-05, 'epoch': 1.68}) +[2026-06-17 23:04:08,463][starVLA.training.train_starvla][INFO] - Step 4084, Loss: {'train/loss': 0.5361676812171936, 'train/grad_norm_pre_clip': 18.131513595581055, 'timing/data': 0.0006740130484104156, 'timing/model': 0.929696380160749, 'learning_rate/qwen_vl_interface': 4.10646770669293e-06, 'learning_rate/action_model': 4.106467706692929e-05, 'epoch': 1.68}) +[2026-06-17 23:04:12,285][starVLA.training.train_starvla][INFO] - Step 4085, Loss: {'train/loss': 0.5312730073928833, 'train/grad_norm_pre_clip': 29.932350158691406, 'timing/data': 0.0006687017157673836, 'timing/model': 0.9281046763062477, 'learning_rate/qwen_vl_interface': 4.104368742156774e-06, 'learning_rate/action_model': 4.104368742156774e-05, 'epoch': 1.68}) +[2026-06-17 23:04:16,214][starVLA.training.train_starvla][INFO] - Step 4086, Loss: {'train/loss': 0.4035874009132385, 'train/grad_norm_pre_clip': 17.72421646118164, 'timing/data': 0.0012360019609332085, 'timing/model': 0.9477663971483707, 'learning_rate/qwen_vl_interface': 4.102270015111217e-06, 'learning_rate/action_model': 4.102270015111217e-05, 'epoch': 1.68}) +[2026-06-17 23:04:20,089][starVLA.training.train_starvla][INFO] - Step 4087, Loss: {'train/loss': 0.13007542490959167, 'train/grad_norm_pre_clip': 31.384557723999023, 'timing/data': 0.001010136678814888, 'timing/model': 0.954334988258779, 'learning_rate/qwen_vl_interface': 4.100171525991318e-06, 'learning_rate/action_model': 4.100171525991318e-05, 'epoch': 1.68}) +[2026-06-17 23:04:23,995][starVLA.training.train_starvla][INFO] - Step 4088, Loss: {'train/loss': 0.6428437829017639, 'train/grad_norm_pre_clip': 27.343055725097656, 'timing/data': 0.0008740201592445374, 'timing/model': 1.0451442748308182, 'learning_rate/qwen_vl_interface': 4.098073275232098e-06, 'learning_rate/action_model': 4.098073275232098e-05, 'epoch': 1.68}) +[2026-06-17 23:04:27,858][starVLA.training.train_starvla][INFO] - Step 4089, Loss: {'train/loss': 0.4776111841201782, 'train/grad_norm_pre_clip': 14.961384773254395, 'timing/data': 0.0005179168656468391, 'timing/model': 0.9452929748222232, 'learning_rate/qwen_vl_interface': 4.09597526326853e-06, 'learning_rate/action_model': 4.095975263268529e-05, 'epoch': 1.68}) +[2026-06-17 23:04:31,681][starVLA.training.train_starvla][INFO] - Step 4090, Loss: {'train/loss': 0.21427135169506073, 'train/grad_norm_pre_clip': 14.443646430969238, 'timing/data': 0.0005112839862704277, 'timing/model': 1.1455673882737756, 'learning_rate/qwen_vl_interface': 4.09387749053553e-06, 'learning_rate/action_model': 4.0938774905355296e-05, 'epoch': 1.68}) +[2026-06-17 23:04:35,298][starVLA.training.train_starvla][INFO] - Step 4091, Loss: {'train/loss': 0.2445824146270752, 'train/grad_norm_pre_clip': 15.294205665588379, 'timing/data': 0.0005739685148000717, 'timing/model': 0.9368300512433052, 'learning_rate/qwen_vl_interface': 4.0917799574679725e-06, 'learning_rate/action_model': 4.091779957467973e-05, 'epoch': 1.68}) +[2026-06-17 23:04:39,253][starVLA.training.train_starvla][INFO] - Step 4092, Loss: {'train/loss': 0.26832759380340576, 'train/grad_norm_pre_clip': 27.8370304107666, 'timing/data': 0.0005974611267447472, 'timing/model': 0.9549104431644082, 'learning_rate/qwen_vl_interface': 4.0896826645006725e-06, 'learning_rate/action_model': 4.089682664500673e-05, 'epoch': 1.68}) +[2026-06-17 23:04:43,350][starVLA.training.train_starvla][INFO] - Step 4093, Loss: {'train/loss': 0.28605324029922485, 'train/grad_norm_pre_clip': 15.656447410583496, 'timing/data': 0.0005516540259122849, 'timing/model': 0.94313685875386, 'learning_rate/qwen_vl_interface': 4.087585612068408e-06, 'learning_rate/action_model': 4.0875856120684074e-05, 'epoch': 1.68}) +[2026-06-17 23:04:47,210][starVLA.training.train_starvla][INFO] - Step 4094, Loss: {'train/loss': 0.33050814270973206, 'train/grad_norm_pre_clip': 15.838461875915527, 'timing/data': 0.0006730277091264725, 'timing/model': 0.944961765781045, 'learning_rate/qwen_vl_interface': 4.0854888006058955e-06, 'learning_rate/action_model': 4.085488800605896e-05, 'epoch': 1.69}) +[2026-06-17 23:04:51,258][starVLA.training.train_starvla][INFO] - Step 4095, Loss: {'train/loss': 0.3953428566455841, 'train/grad_norm_pre_clip': 12.855276107788086, 'timing/data': 0.0014042872935533524, 'timing/model': 0.9387493003159761, 'learning_rate/qwen_vl_interface': 4.083392230547808e-06, 'learning_rate/action_model': 4.083392230547808e-05, 'epoch': 1.69}) +[2026-06-17 23:04:55,038][starVLA.training.train_starvla][INFO] - Step 4096, Loss: {'train/loss': 0.5051931738853455, 'train/grad_norm_pre_clip': 21.25177764892578, 'timing/data': 0.0007502809166908264, 'timing/model': 1.1054004561156034, 'learning_rate/qwen_vl_interface': 4.081295902328765e-06, 'learning_rate/action_model': 4.081295902328765e-05, 'epoch': 1.69}) +[2026-06-17 23:04:58,806][starVLA.training.train_starvla][INFO] - Step 4097, Loss: {'train/loss': 0.26832789182662964, 'train/grad_norm_pre_clip': 13.683622360229492, 'timing/data': 0.0004907594993710518, 'timing/model': 1.0957987196743488, 'learning_rate/qwen_vl_interface': 4.079199816383341e-06, 'learning_rate/action_model': 4.0791998163833404e-05, 'epoch': 1.69}) +[2026-06-17 23:05:02,425][starVLA.training.train_starvla][INFO] - Step 4098, Loss: {'train/loss': 0.4849470257759094, 'train/grad_norm_pre_clip': 15.489602088928223, 'timing/data': 0.0005108611658215523, 'timing/model': 0.9577818252146244, 'learning_rate/qwen_vl_interface': 4.077103973146054e-06, 'learning_rate/action_model': 4.077103973146054e-05, 'epoch': 1.69}) +[2026-06-17 23:05:06,218][starVLA.training.train_starvla][INFO] - Step 4099, Loss: {'train/loss': 0.4484478533267975, 'train/grad_norm_pre_clip': 18.673912048339844, 'timing/data': 0.0005402695387601852, 'timing/model': 0.9305172497406602, 'learning_rate/qwen_vl_interface': 4.075008373051375e-06, 'learning_rate/action_model': 4.075008373051374e-05, 'epoch': 1.69}) +[2026-06-17 23:07:00,469][starVLA.training.train_starvla][INFO] - Step 4100, Loss: {'train/loss': 0.6004638075828552, 'train/grad_norm_pre_clip': 20.303791046142578, 'eval/loss': 0.44600918889045715, 'eval/latency_6/loss': 0.44600918889045715, 'eval/demon_fire_precision': 0.9602436678422571, 'eval/demon_fire_recall': 0.9680025856496445, 'eval/demon_fire_f1': 0.9641075164976661, 'eval/demon_fire_rate_model': 0.8678352810239288, 'eval/demon_fire_rate_teacher': 0.860879243183083, 'eval/demon_left_precision': 0.7809187279151943, 'eval/demon_left_recall': 0.775438596491228, 'eval/demon_left_f1': 0.778169014084507, 'eval/demon_left_rate_model': 0.15748469671675014, 'eval/demon_left_rate_teacher': 0.15859766277128548, 'eval/demon_right_precision': 0.8310104529616724, 'eval/demon_right_recall': 0.7327188940092166, 'eval/demon_right_f1': 0.7787755102040816, 'eval/demon_right_rate_model': 0.15971062882582082, 'eval/demon_right_rate_teacher': 0.18113522537562604, 'eval/demon_move_f1': 0.7784722621442943, 'eval/demon_cc_f1': 0.8997415429253357, 'eval/latency_6/demon_fire_precision': 0.9602436678422571, 'eval/latency_6/demon_fire_recall': 0.9680025856496445, 'eval/latency_6/demon_fire_f1': 0.9641075164976661, 'eval/latency_6/demon_fire_rate_model': 0.8678352810239288, 'eval/latency_6/demon_fire_rate_teacher': 0.860879243183083, 'eval/latency_6/demon_left_precision': 0.7809187279151943, 'eval/latency_6/demon_left_recall': 0.775438596491228, 'eval/latency_6/demon_left_f1': 0.778169014084507, 'eval/latency_6/demon_left_rate_model': 0.15748469671675014, 'eval/latency_6/demon_left_rate_teacher': 0.15859766277128548, 'eval/latency_6/demon_right_precision': 0.8310104529616724, 'eval/latency_6/demon_right_recall': 0.7327188940092166, 'eval/latency_6/demon_right_f1': 0.7787755102040816, 'eval/latency_6/demon_right_rate_model': 0.15971062882582082, 'eval/latency_6/demon_right_rate_teacher': 0.18113522537562604, 'eval/latency_6/demon_move_f1': 0.7784722621442943, 'eval/latency_6/demon_cc_f1': 0.8997415429253357, 'timing/data': 0.0005176141858100891, 'timing/model': 0.9273651158437133, 'learning_rate/qwen_vl_interface': 4.072913016533726e-06, 'learning_rate/action_model': 4.072913016533726e-05, 'epoch': 1.69}) +[2026-06-17 23:07:00,478][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_4100_state +[2026-06-17 23:07:00,482][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 23:07:07,184][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_4100_state/model.safetensors +[2026-06-17 23:07:22,644][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_4100_state/optimizer.bin +[2026-06-17 23:07:22,647][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_4100_state/sampler.bin +[2026-06-17 23:07:22,647][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_4100_state/sampler_1.bin +[2026-06-17 23:07:22,651][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_4100_state/random_states_0.pkl +[2026-06-17 23:07:22,653][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4100 +[2026-06-17 23:07:22,654][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 23:07:22,664][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 23:07:26,438][starVLA.training.train_starvla][INFO] - Step 4101, Loss: {'train/loss': 0.24769441783428192, 'train/grad_norm_pre_clip': 12.494535446166992, 'timing/data': 0.0005036052316427231, 'timing/model': 0.9317427026107907, 'learning_rate/qwen_vl_interface': 4.070817904027474e-06, 'learning_rate/action_model': 4.070817904027474e-05, 'epoch': 1.69}) +[2026-06-17 23:07:30,178][starVLA.training.train_starvla][INFO] - Step 4102, Loss: {'train/loss': 0.43468061089515686, 'train/grad_norm_pre_clip': 15.696834564208984, 'timing/data': 0.0005800016224384308, 'timing/model': 0.9244398958981037, 'learning_rate/qwen_vl_interface': 4.06872303596694e-06, 'learning_rate/action_model': 4.06872303596694e-05, 'epoch': 1.69}) +[2026-06-17 23:07:33,890][starVLA.training.train_starvla][INFO] - Step 4103, Loss: {'train/loss': 0.31769052147865295, 'train/grad_norm_pre_clip': 12.510366439819336, 'timing/data': 0.0004995418712496758, 'timing/model': 0.9255629172548652, 'learning_rate/qwen_vl_interface': 4.066628412786391e-06, 'learning_rate/action_model': 4.066628412786391e-05, 'epoch': 1.69}) +[2026-06-17 23:07:37,706][starVLA.training.train_starvla][INFO] - Step 4104, Loss: {'train/loss': 0.3882844150066376, 'train/grad_norm_pre_clip': 27.600072860717773, 'timing/data': 0.0007735136896371841, 'timing/model': 1.0010063983500004, 'learning_rate/qwen_vl_interface': 4.0645340349200466e-06, 'learning_rate/action_model': 4.0645340349200464e-05, 'epoch': 1.69}) +[2026-06-17 23:07:41,517][starVLA.training.train_starvla][INFO] - Step 4105, Loss: {'train/loss': 0.2046680748462677, 'train/grad_norm_pre_clip': 21.766008377075195, 'timing/data': 0.0005965782329440117, 'timing/model': 1.1397129371762276, 'learning_rate/qwen_vl_interface': 4.062439902802073e-06, 'learning_rate/action_model': 4.0624399028020725e-05, 'epoch': 1.69}) +[2026-06-17 23:07:45,128][starVLA.training.train_starvla][INFO] - Step 4106, Loss: {'train/loss': 0.45831456780433655, 'train/grad_norm_pre_clip': 17.211116790771484, 'timing/data': 0.0005542868748307228, 'timing/model': 0.9284291565418243, 'learning_rate/qwen_vl_interface': 4.060346016866584e-06, 'learning_rate/action_model': 4.060346016866584e-05, 'epoch': 1.69}) +[2026-06-17 23:07:49,333][starVLA.training.train_starvla][INFO] - Step 4107, Loss: {'train/loss': 0.3491979241371155, 'train/grad_norm_pre_clip': 18.89876365661621, 'timing/data': 0.0005079442635178566, 'timing/model': 0.9265973661094904, 'learning_rate/qwen_vl_interface': 4.058252377547645e-06, 'learning_rate/action_model': 4.058252377547645e-05, 'epoch': 1.69}) +[2026-06-17 23:07:53,045][starVLA.training.train_starvla][INFO] - Step 4108, Loss: {'train/loss': 0.341622531414032, 'train/grad_norm_pre_clip': 15.035357475280762, 'timing/data': 0.0005033183842897415, 'timing/model': 0.9306325679644942, 'learning_rate/qwen_vl_interface': 4.056158985279272e-06, 'learning_rate/action_model': 4.0561589852792715e-05, 'epoch': 1.69}) +[2026-06-17 23:07:56,813][starVLA.training.train_starvla][INFO] - Step 4109, Loss: {'train/loss': 0.30316320061683655, 'train/grad_norm_pre_clip': 26.426738739013672, 'timing/data': 0.000487644225358963, 'timing/model': 0.9421093268319964, 'learning_rate/qwen_vl_interface': 4.054065840495425e-06, 'learning_rate/action_model': 4.054065840495425e-05, 'epoch': 1.69}) +[2026-06-17 23:08:00,794][starVLA.training.train_starvla][INFO] - Step 4110, Loss: {'train/loss': 0.29665860533714294, 'train/grad_norm_pre_clip': 21.662076950073242, 'timing/data': 0.000550244003534317, 'timing/model': 0.9493652489036322, 'learning_rate/qwen_vl_interface': 4.0519729436300165e-06, 'learning_rate/action_model': 4.051972943630016e-05, 'epoch': 1.69}) +[2026-06-17 23:08:04,598][starVLA.training.train_starvla][INFO] - Step 4111, Loss: {'train/loss': 0.29363858699798584, 'train/grad_norm_pre_clip': 20.736047744750977, 'timing/data': 0.0009007807821035385, 'timing/model': 1.1195787619799376, 'learning_rate/qwen_vl_interface': 4.049880295116906e-06, 'learning_rate/action_model': 4.049880295116906e-05, 'epoch': 1.69}) +[2026-06-17 23:08:08,372][starVLA.training.train_starvla][INFO] - Step 4112, Loss: {'train/loss': 0.39111918210983276, 'train/grad_norm_pre_clip': 18.793479919433594, 'timing/data': 0.0007820269092917442, 'timing/model': 1.0837714727967978, 'learning_rate/qwen_vl_interface': 4.047787895389902e-06, 'learning_rate/action_model': 4.0477878953899016e-05, 'epoch': 1.69}) +[2026-06-17 23:08:12,025][starVLA.training.train_starvla][INFO] - Step 4113, Loss: {'train/loss': 0.44840890169143677, 'train/grad_norm_pre_clip': 17.85882568359375, 'timing/data': 0.000866452232003212, 'timing/model': 0.9566225297749043, 'learning_rate/qwen_vl_interface': 4.0456957448827615e-06, 'learning_rate/action_model': 4.045695744882761e-05, 'epoch': 1.69}) +[2026-06-17 23:08:16,194][starVLA.training.train_starvla][INFO] - Step 4114, Loss: {'train/loss': 0.2766916751861572, 'train/grad_norm_pre_clip': 15.582721710205078, 'timing/data': 0.0005783038213849068, 'timing/model': 1.1286239502951503, 'learning_rate/qwen_vl_interface': 4.043603844029188e-06, 'learning_rate/action_model': 4.0436038440291874e-05, 'epoch': 1.69}) +[2026-06-17 23:08:20,330][starVLA.training.train_starvla][INFO] - Step 4115, Loss: {'train/loss': 0.15362830460071564, 'train/grad_norm_pre_clip': 8.897679328918457, 'timing/data': 0.0005365591496229172, 'timing/model': 0.9344630874693394, 'learning_rate/qwen_vl_interface': 4.041512193262836e-06, 'learning_rate/action_model': 4.041512193262836e-05, 'epoch': 1.69}) +[2026-06-17 23:08:24,116][starVLA.training.train_starvla][INFO] - Step 4116, Loss: {'train/loss': 0.21400867402553558, 'train/grad_norm_pre_clip': 13.537972450256348, 'timing/data': 0.0013660173863172531, 'timing/model': 0.9455176834017038, 'learning_rate/qwen_vl_interface': 4.039420793017306e-06, 'learning_rate/action_model': 4.039420793017306e-05, 'epoch': 1.69}) +[2026-06-17 23:08:27,976][starVLA.training.train_starvla][INFO] - Step 4117, Loss: {'train/loss': 0.3995825946331024, 'train/grad_norm_pre_clip': 19.81159210205078, 'timing/data': 0.0005418434739112854, 'timing/model': 0.986320910975337, 'learning_rate/qwen_vl_interface': 4.03732964372615e-06, 'learning_rate/action_model': 4.037329643726149e-05, 'epoch': 1.69}) +[2026-06-17 23:08:31,847][starVLA.training.train_starvla][INFO] - Step 4118, Loss: {'train/loss': 0.48377761244773865, 'train/grad_norm_pre_clip': 22.408071517944336, 'timing/data': 0.000531916506588459, 'timing/model': 1.1217781584709883, 'learning_rate/qwen_vl_interface': 4.035238745822861e-06, 'learning_rate/action_model': 4.035238745822861e-05, 'epoch': 1.7}) +[2026-06-17 23:08:35,711][starVLA.training.train_starvla][INFO] - Step 4119, Loss: {'train/loss': 0.37082913517951965, 'train/grad_norm_pre_clip': 15.471388816833496, 'timing/data': 0.000555272214114666, 'timing/model': 1.0978880319744349, 'learning_rate/qwen_vl_interface': 4.033148099740889e-06, 'learning_rate/action_model': 4.0331480997408885e-05, 'epoch': 1.7}) +[2026-06-17 23:08:39,352][starVLA.training.train_starvla][INFO] - Step 4120, Loss: {'train/loss': 0.21052388846874237, 'train/grad_norm_pre_clip': 25.256772994995117, 'timing/data': 0.0005950341001152992, 'timing/model': 0.9344150247052312, 'learning_rate/qwen_vl_interface': 4.031057705913626e-06, 'learning_rate/action_model': 4.031057705913626e-05, 'epoch': 1.7}) +[2026-06-17 23:08:43,148][starVLA.training.train_starvla][INFO] - Step 4121, Loss: {'train/loss': 0.3987579941749573, 'train/grad_norm_pre_clip': 19.384811401367188, 'timing/data': 0.000551220029592514, 'timing/model': 0.9289208902046084, 'learning_rate/qwen_vl_interface': 4.028967564774411e-06, 'learning_rate/action_model': 4.0289675647744106e-05, 'epoch': 1.7}) +[2026-06-17 23:08:46,930][starVLA.training.train_starvla][INFO] - Step 4122, Loss: {'train/loss': 0.29865553975105286, 'train/grad_norm_pre_clip': 18.490821838378906, 'timing/data': 0.0006237654015421867, 'timing/model': 0.9344702456146479, 'learning_rate/qwen_vl_interface': 4.026877676756534e-06, 'learning_rate/action_model': 4.026877676756534e-05, 'epoch': 1.7}) +[2026-06-17 23:08:50,738][starVLA.training.train_starvla][INFO] - Step 4123, Loss: {'train/loss': 0.14486145973205566, 'train/grad_norm_pre_clip': 9.17011547088623, 'timing/data': 0.000528218224644661, 'timing/model': 0.9582166215404868, 'learning_rate/qwen_vl_interface': 4.0247880422932305e-06, 'learning_rate/action_model': 4.02478804229323e-05, 'epoch': 1.7}) +[2026-06-17 23:08:54,533][starVLA.training.train_starvla][INFO] - Step 4124, Loss: {'train/loss': 0.5035661458969116, 'train/grad_norm_pre_clip': 19.158180236816406, 'timing/data': 0.0009081307798624039, 'timing/model': 0.9323844434693456, 'learning_rate/qwen_vl_interface': 4.0226986618176855e-06, 'learning_rate/action_model': 4.022698661817686e-05, 'epoch': 1.7}) +[2026-06-17 23:08:58,321][starVLA.training.train_starvla][INFO] - Step 4125, Loss: {'train/loss': 0.4729968309402466, 'train/grad_norm_pre_clip': 18.905397415161133, 'timing/data': 0.00082358717918396, 'timing/model': 1.0951948976144195, 'learning_rate/qwen_vl_interface': 4.0206095357630256e-06, 'learning_rate/action_model': 4.020609535763026e-05, 'epoch': 1.7}) +[2026-06-17 23:09:02,052][starVLA.training.train_starvla][INFO] - Step 4126, Loss: {'train/loss': 0.3738323152065277, 'train/grad_norm_pre_clip': 12.363456726074219, 'timing/data': 0.001527520827949047, 'timing/model': 1.0573086496442556, 'learning_rate/qwen_vl_interface': 4.018520664562334e-06, 'learning_rate/action_model': 4.018520664562334e-05, 'epoch': 1.7}) +[2026-06-17 23:09:05,895][starVLA.training.train_starvla][INFO] - Step 4127, Loss: {'train/loss': 0.4416661858558655, 'train/grad_norm_pre_clip': 18.811023712158203, 'timing/data': 0.0007272288203239441, 'timing/model': 0.9541189754381776, 'learning_rate/qwen_vl_interface': 4.016432048648635e-06, 'learning_rate/action_model': 4.016432048648635e-05, 'epoch': 1.7}) +[2026-06-17 23:09:09,981][starVLA.training.train_starvla][INFO] - Step 4128, Loss: {'train/loss': 0.32221823930740356, 'train/grad_norm_pre_clip': 10.440792083740234, 'timing/data': 0.0007569137960672379, 'timing/model': 0.9405915373936296, 'learning_rate/qwen_vl_interface': 4.014343688454898e-06, 'learning_rate/action_model': 4.0143436884548976e-05, 'epoch': 1.7}) +[2026-06-17 23:09:13,855][starVLA.training.train_starvla][INFO] - Step 4129, Loss: {'train/loss': 0.2906947433948517, 'train/grad_norm_pre_clip': 15.539770126342773, 'timing/data': 0.0009107254445552826, 'timing/model': 0.9396557547152042, 'learning_rate/qwen_vl_interface': 4.012255584414044e-06, 'learning_rate/action_model': 4.0122555844140436e-05, 'epoch': 1.7}) +[2026-06-17 23:09:17,988][starVLA.training.train_starvla][INFO] - Step 4130, Loss: {'train/loss': 0.41622111201286316, 'train/grad_norm_pre_clip': 17.48879623413086, 'timing/data': 0.001000862568616867, 'timing/model': 0.9552501272410154, 'learning_rate/qwen_vl_interface': 4.010167736958939e-06, 'learning_rate/action_model': 4.010167736958938e-05, 'epoch': 1.7}) +[2026-06-17 23:09:21,800][starVLA.training.train_starvla][INFO] - Step 4131, Loss: {'train/loss': 0.31364190578460693, 'train/grad_norm_pre_clip': 14.450361251831055, 'timing/data': 0.0005025947466492653, 'timing/model': 1.1288957810029387, 'learning_rate/qwen_vl_interface': 4.008080146522397e-06, 'learning_rate/action_model': 4.008080146522396e-05, 'epoch': 1.7}) +[2026-06-17 23:09:25,576][starVLA.training.train_starvla][INFO] - Step 4132, Loss: {'train/loss': 0.34709763526916504, 'train/grad_norm_pre_clip': 18.67607879638672, 'timing/data': 0.0005342494696378708, 'timing/model': 1.096000112593174, 'learning_rate/qwen_vl_interface': 4.005992813537176e-06, 'learning_rate/action_model': 4.005992813537176e-05, 'epoch': 1.7}) +[2026-06-17 23:09:29,221][starVLA.training.train_starvla][INFO] - Step 4133, Loss: {'train/loss': 0.3745865821838379, 'train/grad_norm_pre_clip': 19.238916397094727, 'timing/data': 0.0005072653293609619, 'timing/model': 0.9338403753936291, 'learning_rate/qwen_vl_interface': 4.003905738435984e-06, 'learning_rate/action_model': 4.003905738435984e-05, 'epoch': 1.7}) +[2026-06-17 23:09:33,032][starVLA.training.train_starvla][INFO] - Step 4134, Loss: {'train/loss': 0.5545336008071899, 'train/grad_norm_pre_clip': 19.38599967956543, 'timing/data': 0.0007072687149047852, 'timing/model': 0.9506668141111732, 'learning_rate/qwen_vl_interface': 4.001818921651475e-06, 'learning_rate/action_model': 4.0018189216514746e-05, 'epoch': 1.7}) +[2026-06-17 23:09:36,887][starVLA.training.train_starvla][INFO] - Step 4135, Loss: {'train/loss': 0.5020857453346252, 'train/grad_norm_pre_clip': 21.89264488220215, 'timing/data': 0.0005183611065149307, 'timing/model': 0.9624267853796482, 'learning_rate/qwen_vl_interface': 3.999732363616244e-06, 'learning_rate/action_model': 3.999732363616244e-05, 'epoch': 1.7}) +[2026-06-17 23:09:40,681][starVLA.training.train_starvla][INFO] - Step 4136, Loss: {'train/loss': 0.08169914036989212, 'train/grad_norm_pre_clip': 5.8411865234375, 'timing/data': 0.0005044294521212578, 'timing/model': 0.9367762692272663, 'learning_rate/qwen_vl_interface': 3.9976460647628394e-06, 'learning_rate/action_model': 3.997646064762839e-05, 'epoch': 1.7}) +[2026-06-17 23:09:44,665][starVLA.training.train_starvla][INFO] - Step 4137, Loss: {'train/loss': 0.41833895444869995, 'train/grad_norm_pre_clip': 19.188861846923828, 'timing/data': 0.0006486829370260239, 'timing/model': 1.1898561250418425, 'learning_rate/qwen_vl_interface': 3.995560025523755e-06, 'learning_rate/action_model': 3.995560025523754e-05, 'epoch': 1.7}) +[2026-06-17 23:09:48,547][starVLA.training.train_starvla][INFO] - Step 4138, Loss: {'train/loss': 0.19194774329662323, 'train/grad_norm_pre_clip': 12.79898738861084, 'timing/data': 0.0005683833733201027, 'timing/model': 1.147534865885973, 'learning_rate/qwen_vl_interface': 3.993474246331426e-06, 'learning_rate/action_model': 3.993474246331426e-05, 'epoch': 1.7}) +[2026-06-17 23:09:52,221][starVLA.training.train_starvla][INFO] - Step 4139, Loss: {'train/loss': 0.43172985315322876, 'train/grad_norm_pre_clip': 19.12253189086914, 'timing/data': 0.0009090760722756386, 'timing/model': 0.9610376944765449, 'learning_rate/qwen_vl_interface': 3.991388727618237e-06, 'learning_rate/action_model': 3.991388727618237e-05, 'epoch': 1.7}) +[2026-06-17 23:09:56,074][starVLA.training.train_starvla][INFO] - Step 4140, Loss: {'train/loss': 0.3427753746509552, 'train/grad_norm_pre_clip': 16.55160903930664, 'timing/data': 0.0006989846006035805, 'timing/model': 0.9807316055521369, 'learning_rate/qwen_vl_interface': 3.989303469816521e-06, 'learning_rate/action_model': 3.9893034698165205e-05, 'epoch': 1.7}) +[2026-06-17 23:09:59,874][starVLA.training.train_starvla][INFO] - Step 4141, Loss: {'train/loss': 0.20050664246082306, 'train/grad_norm_pre_clip': 13.313606262207031, 'timing/data': 0.0006559016183018684, 'timing/model': 0.9322498077526689, 'learning_rate/qwen_vl_interface': 3.9872184733585526e-06, 'learning_rate/action_model': 3.987218473358552e-05, 'epoch': 1.7}) +[2026-06-17 23:10:03,846][starVLA.training.train_starvla][INFO] - Step 4142, Loss: {'train/loss': 0.2153911143541336, 'train/grad_norm_pre_clip': 11.917237281799316, 'timing/data': 0.0005285227671265602, 'timing/model': 0.9341270886361599, 'learning_rate/qwen_vl_interface': 3.985133738676552e-06, 'learning_rate/action_model': 3.985133738676552e-05, 'epoch': 1.71}) +[2026-06-17 23:10:07,784][starVLA.training.train_starvla][INFO] - Step 4143, Loss: {'train/loss': 0.2764815092086792, 'train/grad_norm_pre_clip': 18.148880004882812, 'timing/data': 0.0008525094017386436, 'timing/model': 0.9439356569200754, 'learning_rate/qwen_vl_interface': 3.983049266202688e-06, 'learning_rate/action_model': 3.983049266202687e-05, 'epoch': 1.71}) +[2026-06-17 23:10:11,623][starVLA.training.train_starvla][INFO] - Step 4144, Loss: {'train/loss': 0.24448055028915405, 'train/grad_norm_pre_clip': 17.152433395385742, 'timing/data': 0.0006372742354869843, 'timing/model': 1.134849245660007, 'learning_rate/qwen_vl_interface': 3.980965056369075e-06, 'learning_rate/action_model': 3.980965056369075e-05, 'epoch': 1.71}) +[2026-06-17 23:10:15,253][starVLA.training.train_starvla][INFO] - Step 4145, Loss: {'train/loss': 0.27721163630485535, 'train/grad_norm_pre_clip': 15.829916000366211, 'timing/data': 0.0008928347378969193, 'timing/model': 0.934633294120431, 'learning_rate/qwen_vl_interface': 3.978881109607772e-06, 'learning_rate/action_model': 3.978881109607772e-05, 'epoch': 1.71}) +[2026-06-17 23:10:19,032][starVLA.training.train_starvla][INFO] - Step 4146, Loss: {'train/loss': 0.2755272090435028, 'train/grad_norm_pre_clip': 22.486562728881836, 'timing/data': 0.0005515124648809433, 'timing/model': 0.9320321157574654, 'learning_rate/qwen_vl_interface': 3.976797426350783e-06, 'learning_rate/action_model': 3.976797426350783e-05, 'epoch': 1.71}) +[2026-06-17 23:10:22,999][starVLA.training.train_starvla][INFO] - Step 4147, Loss: {'train/loss': 0.35409024357795715, 'train/grad_norm_pre_clip': 23.36939239501953, 'timing/data': 0.0005033668130636215, 'timing/model': 0.9366866406053305, 'learning_rate/qwen_vl_interface': 3.974714007030055e-06, 'learning_rate/action_model': 3.974714007030056e-05, 'epoch': 1.71}) +[2026-06-17 23:10:26,870][starVLA.training.train_starvla][INFO] - Step 4148, Loss: {'train/loss': 0.36843356490135193, 'train/grad_norm_pre_clip': 26.06049919128418, 'timing/data': 0.0005085570737719536, 'timing/model': 0.9917157627642155, 'learning_rate/qwen_vl_interface': 3.972630852077488e-06, 'learning_rate/action_model': 3.972630852077488e-05, 'epoch': 1.71}) +[2026-06-17 23:10:30,717][starVLA.training.train_starvla][INFO] - Step 4149, Loss: {'train/loss': 0.5192263722419739, 'train/grad_norm_pre_clip': 15.515231132507324, 'timing/data': 0.0005186786875128746, 'timing/model': 0.958068679086864, 'learning_rate/qwen_vl_interface': 3.9705479619249195e-06, 'learning_rate/action_model': 3.9705479619249197e-05, 'epoch': 1.71}) +[2026-06-17 23:10:34,529][starVLA.training.train_starvla][INFO] - Step 4150, Loss: {'train/loss': 0.5552538633346558, 'train/grad_norm_pre_clip': 24.452951431274414, 'timing/data': 0.0005466071888804436, 'timing/model': 0.9332887390628457, 'learning_rate/qwen_vl_interface': 3.9684653370041345e-06, 'learning_rate/action_model': 3.968465337004134e-05, 'epoch': 1.71}) +[2026-06-17 23:10:38,352][starVLA.training.train_starvla][INFO] - Step 4151, Loss: {'train/loss': 0.2991583049297333, 'train/grad_norm_pre_clip': 20.249958038330078, 'timing/data': 0.0005605276674032211, 'timing/model': 1.1268801456317306, 'learning_rate/qwen_vl_interface': 3.966382977746866e-06, 'learning_rate/action_model': 3.9663829777468656e-05, 'epoch': 1.71}) +[2026-06-17 23:10:41,953][starVLA.training.train_starvla][INFO] - Step 4152, Loss: {'train/loss': 0.39674103260040283, 'train/grad_norm_pre_clip': 16.363361358642578, 'timing/data': 0.0005165962502360344, 'timing/model': 0.9282873421907425, 'learning_rate/qwen_vl_interface': 3.9643008845847845e-06, 'learning_rate/action_model': 3.964300884584785e-05, 'epoch': 1.71}) +[2026-06-17 23:10:45,770][starVLA.training.train_starvla][INFO] - Step 4153, Loss: {'train/loss': 0.40326669812202454, 'train/grad_norm_pre_clip': 20.36404800415039, 'timing/data': 0.0005067046731710434, 'timing/model': 0.9323190674185753, 'learning_rate/qwen_vl_interface': 3.962219057949517e-06, 'learning_rate/action_model': 3.9622190579495166e-05, 'epoch': 1.71}) +[2026-06-17 23:10:49,865][starVLA.training.train_starvla][INFO] - Step 4154, Loss: {'train/loss': 0.2109677940607071, 'train/grad_norm_pre_clip': 18.063217163085938, 'timing/data': 0.00108281709253788, 'timing/model': 0.9375244444236159, 'learning_rate/qwen_vl_interface': 3.960137498272619e-06, 'learning_rate/action_model': 3.9601374982726194e-05, 'epoch': 1.71}) +[2026-06-17 23:10:53,655][starVLA.training.train_starvla][INFO] - Step 4155, Loss: {'train/loss': 0.28469011187553406, 'train/grad_norm_pre_clip': 17.42725372314453, 'timing/data': 0.0014697182923555374, 'timing/model': 0.9328513462096453, 'learning_rate/qwen_vl_interface': 3.958056205985611e-06, 'learning_rate/action_model': 3.95805620598561e-05, 'epoch': 1.71}) +[2026-06-17 23:10:57,538][starVLA.training.train_starvla][INFO] - Step 4156, Loss: {'train/loss': 0.38588249683380127, 'train/grad_norm_pre_clip': 17.313491821289062, 'timing/data': 0.0006613973528146744, 'timing/model': 0.9414438782259822, 'learning_rate/qwen_vl_interface': 3.95597518151994e-06, 'learning_rate/action_model': 3.955975181519939e-05, 'epoch': 1.71}) +[2026-06-17 23:11:01,419][starVLA.training.train_starvla][INFO] - Step 4157, Loss: {'train/loss': 0.3980677127838135, 'train/grad_norm_pre_clip': 25.206628799438477, 'timing/data': 0.001024237833917141, 'timing/model': 1.1088185105472803, 'learning_rate/qwen_vl_interface': 3.953894425307005e-06, 'learning_rate/action_model': 3.9538944253070055e-05, 'epoch': 1.71}) +[2026-06-17 23:11:05,205][starVLA.training.train_starvla][INFO] - Step 4158, Loss: {'train/loss': 0.30542099475860596, 'train/grad_norm_pre_clip': 25.047021865844727, 'timing/data': 0.0007557841017842293, 'timing/model': 1.1145702134817839, 'learning_rate/qwen_vl_interface': 3.951813937778151e-06, 'learning_rate/action_model': 3.95181393777815e-05, 'epoch': 1.71}) +[2026-06-17 23:11:08,986][starVLA.training.train_starvla][INFO] - Step 4159, Loss: {'train/loss': 0.38629043102264404, 'train/grad_norm_pre_clip': 22.02943992614746, 'timing/data': 0.0005066171288490295, 'timing/model': 1.1019534152001143, 'learning_rate/qwen_vl_interface': 3.949733719364664e-06, 'learning_rate/action_model': 3.949733719364664e-05, 'epoch': 1.71}) +[2026-06-17 23:11:13,089][starVLA.training.train_starvla][INFO] - Step 4160, Loss: {'train/loss': 0.7265226244926453, 'train/grad_norm_pre_clip': 29.556642532348633, 'timing/data': 0.0005137352272868156, 'timing/model': 1.0220944471657276, 'learning_rate/qwen_vl_interface': 3.947653770497778e-06, 'learning_rate/action_model': 3.947653770497778e-05, 'epoch': 1.71}) +[2026-06-17 23:11:17,219][starVLA.training.train_starvla][INFO] - Step 4161, Loss: {'train/loss': 0.21448901295661926, 'train/grad_norm_pre_clip': 17.229124069213867, 'timing/data': 0.0005330862477421761, 'timing/model': 1.1556626809760928, 'learning_rate/qwen_vl_interface': 3.945574091608664e-06, 'learning_rate/action_model': 3.9455740916086636e-05, 'epoch': 1.71}) +[2026-06-17 23:11:22,192][starVLA.training.train_starvla][INFO] - Step 4162, Loss: {'train/loss': 0.3858092427253723, 'train/grad_norm_pre_clip': 23.850444793701172, 'timing/data': 0.0009116372093558311, 'timing/model': 2.0049650464206934, 'learning_rate/qwen_vl_interface': 3.9434946831284465e-06, 'learning_rate/action_model': 3.943494683128447e-05, 'epoch': 1.71}) +[2026-06-17 23:11:27,813][starVLA.training.train_starvla][INFO] - Step 4163, Loss: {'train/loss': 0.20377343893051147, 'train/grad_norm_pre_clip': 14.672298431396484, 'timing/data': 0.0006393343210220337, 'timing/model': 2.1206749910488725, 'learning_rate/qwen_vl_interface': 3.941415545488187e-06, 'learning_rate/action_model': 3.941415545488187e-05, 'epoch': 1.71}) +[2026-06-17 23:11:32,250][starVLA.training.train_starvla][INFO] - Step 4164, Loss: {'train/loss': 0.2892609238624573, 'train/grad_norm_pre_clip': 16.816747665405273, 'timing/data': 0.0008947718888521194, 'timing/model': 1.1197253996506333, 'learning_rate/qwen_vl_interface': 3.939336679118892e-06, 'learning_rate/action_model': 3.939336679118892e-05, 'epoch': 1.71}) +[2026-06-17 23:11:36,823][starVLA.training.train_starvla][INFO] - Step 4165, Loss: {'train/loss': 0.3384236991405487, 'train/grad_norm_pre_clip': 21.450462341308594, 'timing/data': 0.0005731796845793724, 'timing/model': 1.1044512400403619, 'learning_rate/qwen_vl_interface': 3.937258084451514e-06, 'learning_rate/action_model': 3.937258084451514e-05, 'epoch': 1.71}) +[2026-06-17 23:11:40,553][starVLA.training.train_starvla][INFO] - Step 4166, Loss: {'train/loss': 0.1930026412010193, 'train/grad_norm_pre_clip': 13.250581741333008, 'timing/data': 0.0006965622305870056, 'timing/model': 0.9690226837992668, 'learning_rate/qwen_vl_interface': 3.9351797619169485e-06, 'learning_rate/action_model': 3.935179761916949e-05, 'epoch': 1.71}) +[2026-06-17 23:11:44,964][starVLA.training.train_starvla][INFO] - Step 4167, Loss: {'train/loss': 0.17878195643424988, 'train/grad_norm_pre_clip': 13.334498405456543, 'timing/data': 0.0005156500265002251, 'timing/model': 0.9443656727671623, 'learning_rate/qwen_vl_interface': 3.933101711946034e-06, 'learning_rate/action_model': 3.9331017119460336e-05, 'epoch': 1.72}) +[2026-06-17 23:11:48,850][starVLA.training.train_starvla][INFO] - Step 4168, Loss: {'train/loss': 0.36092498898506165, 'train/grad_norm_pre_clip': 25.989036560058594, 'timing/data': 0.0009606722742319107, 'timing/model': 0.9406293900683522, 'learning_rate/qwen_vl_interface': 3.93102393496955e-06, 'learning_rate/action_model': 3.93102393496955e-05, 'epoch': 1.72}) +[2026-06-17 23:11:52,634][starVLA.training.train_starvla][INFO] - Step 4169, Loss: {'train/loss': 0.6559969186782837, 'train/grad_norm_pre_clip': 21.84210777282715, 'timing/data': 0.0005301712080836296, 'timing/model': 0.9434040104970336, 'learning_rate/qwen_vl_interface': 3.928946431418223e-06, 'learning_rate/action_model': 3.9289464314182225e-05, 'epoch': 1.72}) +[2026-06-17 23:11:56,379][starVLA.training.train_starvla][INFO] - Step 4170, Loss: {'train/loss': 0.38870304822921753, 'train/grad_norm_pre_clip': 15.315984725952148, 'timing/data': 0.0005018152296543121, 'timing/model': 0.9309720145538449, 'learning_rate/qwen_vl_interface': 3.926869201722723e-06, 'learning_rate/action_model': 3.926869201722723e-05, 'epoch': 1.72}) +[2026-06-17 23:12:00,167][starVLA.training.train_starvla][INFO] - Step 4171, Loss: {'train/loss': 0.41757771372795105, 'train/grad_norm_pre_clip': 19.196277618408203, 'timing/data': 0.0005664946511387825, 'timing/model': 1.1017167698591948, 'learning_rate/qwen_vl_interface': 3.9247922463136615e-06, 'learning_rate/action_model': 3.924792246313662e-05, 'epoch': 1.72}) +[2026-06-17 23:12:03,822][starVLA.training.train_starvla][INFO] - Step 4172, Loss: {'train/loss': 0.13996949791908264, 'train/grad_norm_pre_clip': 11.290095329284668, 'timing/data': 0.000530017539858818, 'timing/model': 0.9556236686185002, 'learning_rate/qwen_vl_interface': 3.922715565621593e-06, 'learning_rate/action_model': 3.9227155656215926e-05, 'epoch': 1.72}) +[2026-06-17 23:12:07,665][starVLA.training.train_starvla][INFO] - Step 4173, Loss: {'train/loss': 0.190794438123703, 'train/grad_norm_pre_clip': 13.662787437438965, 'timing/data': 0.0005517927929759026, 'timing/model': 0.9300918597728014, 'learning_rate/qwen_vl_interface': 3.9206391600770164e-06, 'learning_rate/action_model': 3.920639160077016e-05, 'epoch': 1.72}) +[2026-06-17 23:12:11,611][starVLA.training.train_starvla][INFO] - Step 4174, Loss: {'train/loss': 0.27303048968315125, 'train/grad_norm_pre_clip': 14.750785827636719, 'timing/data': 0.0005213664844632149, 'timing/model': 0.9331061039119959, 'learning_rate/qwen_vl_interface': 3.918563030110373e-06, 'learning_rate/action_model': 3.918563030110373e-05, 'epoch': 1.72}) +[2026-06-17 23:12:15,526][starVLA.training.train_starvla][INFO] - Step 4175, Loss: {'train/loss': 0.13401347398757935, 'train/grad_norm_pre_clip': 12.681404113769531, 'timing/data': 0.0004940228536725044, 'timing/model': 0.9577704621478915, 'learning_rate/qwen_vl_interface': 3.916487176152045e-06, 'learning_rate/action_model': 3.916487176152045e-05, 'epoch': 1.72}) +[2026-06-17 23:12:19,362][starVLA.training.train_starvla][INFO] - Step 4176, Loss: {'train/loss': 0.42282095551490784, 'train/grad_norm_pre_clip': 19.764894485473633, 'timing/data': 0.0013370141386985779, 'timing/model': 0.9243658799678087, 'learning_rate/qwen_vl_interface': 3.914411598632359e-06, 'learning_rate/action_model': 3.914411598632359e-05, 'epoch': 1.72}) +[2026-06-17 23:12:23,569][starVLA.training.train_starvla][INFO] - Step 4177, Loss: {'train/loss': 0.48642390966415405, 'train/grad_norm_pre_clip': 18.976850509643555, 'timing/data': 0.0005299970507621765, 'timing/model': 1.121576676145196, 'learning_rate/qwen_vl_interface': 3.912336297981588e-06, 'learning_rate/action_model': 3.912336297981588e-05, 'epoch': 1.72}) +[2026-06-17 23:12:27,401][starVLA.training.train_starvla][INFO] - Step 4178, Loss: {'train/loss': 0.16910500824451447, 'train/grad_norm_pre_clip': 9.471295356750488, 'timing/data': 0.0009648902341723442, 'timing/model': 1.0572125855833292, 'learning_rate/qwen_vl_interface': 3.910261274629941e-06, 'learning_rate/action_model': 3.910261274629941e-05, 'epoch': 1.72}) +[2026-06-17 23:12:31,018][starVLA.training.train_starvla][INFO] - Step 4179, Loss: {'train/loss': 0.20081759989261627, 'train/grad_norm_pre_clip': 16.42534637451172, 'timing/data': 0.0006314069032669067, 'timing/model': 0.950283613987267, 'learning_rate/qwen_vl_interface': 3.908186529007576e-06, 'learning_rate/action_model': 3.908186529007575e-05, 'epoch': 1.72}) +[2026-06-17 23:12:35,011][starVLA.training.train_starvla][INFO] - Step 4180, Loss: {'train/loss': 0.184159055352211, 'train/grad_norm_pre_clip': 21.112409591674805, 'timing/data': 0.0005075689405202866, 'timing/model': 0.958819312043488, 'learning_rate/qwen_vl_interface': 3.906112061544582e-06, 'learning_rate/action_model': 3.906112061544582e-05, 'epoch': 1.72}) +[2026-06-17 23:12:38,991][starVLA.training.train_starvla][INFO] - Step 4181, Loss: {'train/loss': 0.11687567085027695, 'train/grad_norm_pre_clip': 9.813530921936035, 'timing/data': 0.000522996298968792, 'timing/model': 0.9511378621682525, 'learning_rate/qwen_vl_interface': 3.904037872671008e-06, 'learning_rate/action_model': 3.904037872671008e-05, 'epoch': 1.72}) +[2026-06-17 23:12:42,925][starVLA.training.train_starvla][INFO] - Step 4182, Loss: {'train/loss': 0.9556969404220581, 'train/grad_norm_pre_clip': 24.82502555847168, 'timing/data': 0.0005048112943768501, 'timing/model': 0.9622946698218584, 'learning_rate/qwen_vl_interface': 3.90196396281683e-06, 'learning_rate/action_model': 3.9019639628168295e-05, 'epoch': 1.72}) +[2026-06-17 23:12:46,742][starVLA.training.train_starvla][INFO] - Step 4183, Loss: {'train/loss': 0.35971152782440186, 'train/grad_norm_pre_clip': 16.66686248779297, 'timing/data': 0.0010119173675775528, 'timing/model': 0.9353172089904547, 'learning_rate/qwen_vl_interface': 3.899890332411972e-06, 'learning_rate/action_model': 3.8998903324119724e-05, 'epoch': 1.72}) +[2026-06-17 23:12:50,613][starVLA.training.train_starvla][INFO] - Step 4184, Loss: {'train/loss': 0.5099827647209167, 'train/grad_norm_pre_clip': 21.439659118652344, 'timing/data': 0.0005456339567899704, 'timing/model': 1.1098705539479852, 'learning_rate/qwen_vl_interface': 3.897816981886302e-06, 'learning_rate/action_model': 3.897816981886302e-05, 'epoch': 1.72}) +[2026-06-17 23:12:54,377][starVLA.training.train_starvla][INFO] - Step 4185, Loss: {'train/loss': 0.6140116453170776, 'train/grad_norm_pre_clip': 21.821504592895508, 'timing/data': 0.0011584963649511337, 'timing/model': 1.0875654611736536, 'learning_rate/qwen_vl_interface': 3.895743911669627e-06, 'learning_rate/action_model': 3.895743911669627e-05, 'epoch': 1.72}) +[2026-06-17 23:12:58,317][starVLA.training.train_starvla][INFO] - Step 4186, Loss: {'train/loss': 0.23408563435077667, 'train/grad_norm_pre_clip': 16.05000114440918, 'timing/data': 0.0005677621811628342, 'timing/model': 1.0877796187996864, 'learning_rate/qwen_vl_interface': 3.893671122191698e-06, 'learning_rate/action_model': 3.8936711221916974e-05, 'epoch': 1.72}) +[2026-06-17 23:13:02,304][starVLA.training.train_starvla][INFO] - Step 4187, Loss: {'train/loss': 0.4170091152191162, 'train/grad_norm_pre_clip': 17.4150333404541, 'timing/data': 0.0009460821747779846, 'timing/model': 1.0626154830679297, 'learning_rate/qwen_vl_interface': 3.8915986138822e-06, 'learning_rate/action_model': 3.8915986138822e-05, 'epoch': 1.72}) +[2026-06-17 23:13:06,177][starVLA.training.train_starvla][INFO] - Step 4188, Loss: {'train/loss': 0.39837995171546936, 'train/grad_norm_pre_clip': 22.640352249145508, 'timing/data': 0.0010475050657987595, 'timing/model': 0.9830352608114481, 'learning_rate/qwen_vl_interface': 3.889526387170775e-06, 'learning_rate/action_model': 3.889526387170775e-05, 'epoch': 1.72}) +[2026-06-17 23:13:10,011][starVLA.training.train_starvla][INFO] - Step 4189, Loss: {'train/loss': 0.36874955892562866, 'train/grad_norm_pre_clip': 18.837677001953125, 'timing/data': 0.0005250293761491776, 'timing/model': 0.9353827321901917, 'learning_rate/qwen_vl_interface': 3.8874544424869924e-06, 'learning_rate/action_model': 3.8874544424869926e-05, 'epoch': 1.72}) +[2026-06-17 23:13:13,840][starVLA.training.train_starvla][INFO] - Step 4190, Loss: {'train/loss': 0.5044777989387512, 'train/grad_norm_pre_clip': 29.001506805419922, 'timing/data': 0.0005385791882872581, 'timing/model': 0.9353707460686564, 'learning_rate/qwen_vl_interface': 3.885382780260371e-06, 'learning_rate/action_model': 3.885382780260371e-05, 'epoch': 1.72}) +[2026-06-17 23:13:17,658][starVLA.training.train_starvla][INFO] - Step 4191, Loss: {'train/loss': 0.5643473863601685, 'train/grad_norm_pre_clip': 18.246156692504883, 'timing/data': 0.0006276201456785202, 'timing/model': 1.1064932914450765, 'learning_rate/qwen_vl_interface': 3.883311400920366e-06, 'learning_rate/action_model': 3.883311400920366e-05, 'epoch': 1.73}) +[2026-06-17 23:13:21,441][starVLA.training.train_starvla][INFO] - Step 4192, Loss: {'train/loss': 0.20666618645191193, 'train/grad_norm_pre_clip': 13.243708610534668, 'timing/data': 0.001133452169597149, 'timing/model': 1.0373682025820017, 'learning_rate/qwen_vl_interface': 3.88124030489638e-06, 'learning_rate/action_model': 3.88124030489638e-05, 'epoch': 1.73}) +[2026-06-17 23:13:25,073][starVLA.training.train_starvla][INFO] - Step 4193, Loss: {'train/loss': 0.46259963512420654, 'train/grad_norm_pre_clip': 13.350971221923828, 'timing/data': 0.0005408534780144691, 'timing/model': 0.9419108107686043, 'learning_rate/qwen_vl_interface': 3.8791694926177495e-06, 'learning_rate/action_model': 3.87916949261775e-05, 'epoch': 1.73}) +[2026-06-17 23:13:28,831][starVLA.training.train_starvla][INFO] - Step 4194, Loss: {'train/loss': 0.381397545337677, 'train/grad_norm_pre_clip': 14.170228958129883, 'timing/data': 0.0005248291417956352, 'timing/model': 0.9241700759157538, 'learning_rate/qwen_vl_interface': 3.877098964513759e-06, 'learning_rate/action_model': 3.877098964513759e-05, 'epoch': 1.73}) +[2026-06-17 23:13:32,688][starVLA.training.train_starvla][INFO] - Step 4195, Loss: {'train/loss': 0.3388550877571106, 'train/grad_norm_pre_clip': 11.993778228759766, 'timing/data': 0.0011577587574720383, 'timing/model': 0.939114541746676, 'learning_rate/qwen_vl_interface': 3.875028721013633e-06, 'learning_rate/action_model': 3.875028721013632e-05, 'epoch': 1.73}) +[2026-06-17 23:13:36,487][starVLA.training.train_starvla][INFO] - Step 4196, Loss: {'train/loss': 0.46718862652778625, 'train/grad_norm_pre_clip': 23.063434600830078, 'timing/data': 0.0005167471244931221, 'timing/model': 0.9305428899824619, 'learning_rate/qwen_vl_interface': 3.87295876254653e-06, 'learning_rate/action_model': 3.8729587625465304e-05, 'epoch': 1.73}) +[2026-06-17 23:13:40,323][starVLA.training.train_starvla][INFO] - Step 4197, Loss: {'train/loss': 0.4285019338130951, 'train/grad_norm_pre_clip': 16.269807815551758, 'timing/data': 0.000968199223279953, 'timing/model': 1.1427972642704844, 'learning_rate/qwen_vl_interface': 3.870889089541558e-06, 'learning_rate/action_model': 3.870889089541558e-05, 'epoch': 1.73}) +[2026-06-17 23:13:44,042][starVLA.training.train_starvla][INFO] - Step 4198, Loss: {'train/loss': 0.5263674259185791, 'train/grad_norm_pre_clip': 27.138198852539062, 'timing/data': 0.0006027426570653915, 'timing/model': 1.059891707263887, 'learning_rate/qwen_vl_interface': 3.868819702427762e-06, 'learning_rate/action_model': 3.8688197024277616e-05, 'epoch': 1.73}) +[2026-06-17 23:13:47,746][starVLA.training.train_starvla][INFO] - Step 4199, Loss: {'train/loss': 0.5695803165435791, 'train/grad_norm_pre_clip': 18.918441772460938, 'timing/data': 0.0005216225981712341, 'timing/model': 0.9382619494572282, 'learning_rate/qwen_vl_interface': 3.866750601634128e-06, 'learning_rate/action_model': 3.8667506016341277e-05, 'epoch': 1.73}) +[2026-06-17 23:15:59,717][starVLA.training.train_starvla][INFO] - Step 4200, Loss: {'train/loss': 0.24070265889167786, 'train/grad_norm_pre_clip': 16.472715377807617, 'eval/loss': 0.44197356700897217, 'eval/latency_6/loss': 0.44197356700897217, 'eval/demon_fire_precision': 0.9538120847832964, 'eval/demon_fire_recall': 0.9744667097608274, 'eval/demon_fire_f1': 0.9640287769784173, 'eval/demon_fire_rate_model': 0.8795214245965498, 'eval/demon_fire_rate_teacher': 0.860879243183083, 'eval/demon_left_precision': 0.7799642218246869, 'eval/demon_left_recall': 0.7649122807017544, 'eval/demon_left_f1': 0.7723649247121348, 'eval/demon_left_rate_model': 0.1555370061213133, 'eval/demon_left_rate_teacher': 0.15859766277128548, 'eval/demon_right_precision': 0.7912087912087912, 'eval/demon_right_recall': 0.7741935483870968, 'eval/demon_right_f1': 0.7826086956521738, 'eval/demon_right_rate_model': 0.17723984418475236, 'eval/demon_right_rate_teacher': 0.18113522537562604, 'eval/demon_move_f1': 0.7774868101821544, 'eval/demon_cc_f1': 0.8992983359733396, 'eval/latency_6/demon_fire_precision': 0.9538120847832964, 'eval/latency_6/demon_fire_recall': 0.9744667097608274, 'eval/latency_6/demon_fire_f1': 0.9640287769784173, 'eval/latency_6/demon_fire_rate_model': 0.8795214245965498, 'eval/latency_6/demon_fire_rate_teacher': 0.860879243183083, 'eval/latency_6/demon_left_precision': 0.7799642218246869, 'eval/latency_6/demon_left_recall': 0.7649122807017544, 'eval/latency_6/demon_left_f1': 0.7723649247121348, 'eval/latency_6/demon_left_rate_model': 0.1555370061213133, 'eval/latency_6/demon_left_rate_teacher': 0.15859766277128548, 'eval/latency_6/demon_right_precision': 0.7912087912087912, 'eval/latency_6/demon_right_recall': 0.7741935483870968, 'eval/latency_6/demon_right_f1': 0.7826086956521738, 'eval/latency_6/demon_right_rate_model': 0.17723984418475236, 'eval/latency_6/demon_right_rate_teacher': 0.18113522537562604, 'eval/latency_6/demon_move_f1': 0.7774868101821544, 'eval/latency_6/demon_cc_f1': 0.8992983359733396, 'timing/data': 0.0008375048637390137, 'timing/model': 0.9435838144272566, 'learning_rate/qwen_vl_interface': 3.864681787589581e-06, 'learning_rate/action_model': 3.8646817875895816e-05, 'epoch': 1.73}) +[2026-06-17 23:15:59,721][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_4200_state +[2026-06-17 23:15:59,726][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 23:16:07,387][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_4200_state/model.safetensors +[2026-06-17 23:16:23,086][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_4200_state/optimizer.bin +[2026-06-17 23:16:23,087][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_4200_state/sampler.bin +[2026-06-17 23:16:23,087][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_4200_state/sampler_1.bin +[2026-06-17 23:16:23,089][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_4200_state/random_states_0.pkl +[2026-06-17 23:16:23,091][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4200 +[2026-06-17 23:16:23,092][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 23:16:23,103][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 23:16:26,727][starVLA.training.train_starvla][INFO] - Step 4201, Loss: {'train/loss': 0.3836652338504791, 'train/grad_norm_pre_clip': 18.74369239807129, 'timing/data': 0.000524190254509449, 'timing/model': 0.9550050245597959, 'learning_rate/qwen_vl_interface': 3.862613260722991e-06, 'learning_rate/action_model': 3.862613260722992e-05, 'epoch': 1.73}) +[2026-06-17 23:16:30,506][starVLA.training.train_starvla][INFO] - Step 4202, Loss: {'train/loss': 0.2847767174243927, 'train/grad_norm_pre_clip': 14.703755378723145, 'timing/data': 0.000523061491549015, 'timing/model': 0.9356280202046037, 'learning_rate/qwen_vl_interface': 3.860545021463166e-06, 'learning_rate/action_model': 3.860545021463167e-05, 'epoch': 1.73}) +[2026-06-17 23:16:34,428][starVLA.training.train_starvla][INFO] - Step 4203, Loss: {'train/loss': 0.4991815686225891, 'train/grad_norm_pre_clip': 22.978382110595703, 'timing/data': 0.0005139568820595741, 'timing/model': 0.9336090004071593, 'learning_rate/qwen_vl_interface': 3.858477070238854e-06, 'learning_rate/action_model': 3.8584770702388534e-05, 'epoch': 1.73}) +[2026-06-17 23:16:38,442][starVLA.training.train_starvla][INFO] - Step 4204, Loss: {'train/loss': 0.35297495126724243, 'train/grad_norm_pre_clip': 18.771703720092773, 'timing/data': 0.0006346665322780609, 'timing/model': 0.9769288748502731, 'learning_rate/qwen_vl_interface': 3.856409407478741e-06, 'learning_rate/action_model': 3.856409407478741e-05, 'epoch': 1.73}) +[2026-06-17 23:16:44,724][starVLA.training.train_starvla][INFO] - Step 4205, Loss: {'train/loss': 0.39248403906822205, 'train/grad_norm_pre_clip': 18.347511291503906, 'timing/data': 0.0005093356594443321, 'timing/model': 2.2420016452670097, 'learning_rate/qwen_vl_interface': 3.854342033611456e-06, 'learning_rate/action_model': 3.854342033611456e-05, 'epoch': 1.73}) +[2026-06-17 23:16:48,725][starVLA.training.train_starvla][INFO] - Step 4206, Loss: {'train/loss': 0.27241820096969604, 'train/grad_norm_pre_clip': 14.988483428955078, 'timing/data': 0.0006488477811217308, 'timing/model': 1.1602730378508568, 'learning_rate/qwen_vl_interface': 3.852274949065571e-06, 'learning_rate/action_model': 3.852274949065571e-05, 'epoch': 1.73}) +[2026-06-17 23:16:52,435][starVLA.training.train_starvla][INFO] - Step 4207, Loss: {'train/loss': 0.2978936731815338, 'train/grad_norm_pre_clip': 43.472965240478516, 'timing/data': 0.0006313417106866837, 'timing/model': 0.9506449364125729, 'learning_rate/qwen_vl_interface': 3.850208154269593e-06, 'learning_rate/action_model': 3.8502081542695934e-05, 'epoch': 1.73}) +[2026-06-17 23:16:56,343][starVLA.training.train_starvla][INFO] - Step 4208, Loss: {'train/loss': 0.3097953200340271, 'train/grad_norm_pre_clip': 14.108243942260742, 'timing/data': 0.0005253385752439499, 'timing/model': 0.9330120515078306, 'learning_rate/qwen_vl_interface': 3.8481416496519725e-06, 'learning_rate/action_model': 3.848141649651972e-05, 'epoch': 1.73}) +[2026-06-17 23:17:00,737][starVLA.training.train_starvla][INFO] - Step 4209, Loss: {'train/loss': 0.46762678027153015, 'train/grad_norm_pre_clip': 17.403627395629883, 'timing/data': 0.0009922496974468231, 'timing/model': 0.9471296956762671, 'learning_rate/qwen_vl_interface': 3.846075435641093e-06, 'learning_rate/action_model': 3.846075435641093e-05, 'epoch': 1.73}) +[2026-06-17 23:17:06,478][starVLA.training.train_starvla][INFO] - Step 4210, Loss: {'train/loss': 0.19687533378601074, 'train/grad_norm_pre_clip': 14.338061332702637, 'timing/data': 0.0005196435377001762, 'timing/model': 0.9392171325162053, 'learning_rate/qwen_vl_interface': 3.844009512665289e-06, 'learning_rate/action_model': 3.844009512665289e-05, 'epoch': 1.73}) +[2026-06-17 23:17:12,690][starVLA.training.train_starvla][INFO] - Step 4211, Loss: {'train/loss': 0.2500689923763275, 'train/grad_norm_pre_clip': 12.325910568237305, 'timing/data': 0.0010330090299248695, 'timing/model': 0.9506133897230029, 'learning_rate/qwen_vl_interface': 3.841943881152826e-06, 'learning_rate/action_model': 3.841943881152826e-05, 'epoch': 1.73}) +[2026-06-17 23:17:16,537][starVLA.training.train_starvla][INFO] - Step 4212, Loss: {'train/loss': 0.3267459273338318, 'train/grad_norm_pre_clip': 18.306354522705078, 'timing/data': 0.0005035120993852615, 'timing/model': 1.1491287807002664, 'learning_rate/qwen_vl_interface': 3.83987854153191e-06, 'learning_rate/action_model': 3.83987854153191e-05, 'epoch': 1.73}) +[2026-06-17 23:17:21,030][starVLA.training.train_starvla][INFO] - Step 4213, Loss: {'train/loss': 0.477323055267334, 'train/grad_norm_pre_clip': 20.931299209594727, 'timing/data': 0.0005682799965143204, 'timing/model': 1.7746677594259381, 'learning_rate/qwen_vl_interface': 3.83781349423069e-06, 'learning_rate/action_model': 3.83781349423069e-05, 'epoch': 1.73}) +[2026-06-17 23:17:25,000][starVLA.training.train_starvla][INFO] - Step 4214, Loss: {'train/loss': 0.28279945254325867, 'train/grad_norm_pre_clip': 16.600135803222656, 'timing/data': 0.0007888916879892349, 'timing/model': 0.9681845009326935, 'learning_rate/qwen_vl_interface': 3.835748739677253e-06, 'learning_rate/action_model': 3.835748739677253e-05, 'epoch': 1.73}) +[2026-06-17 23:17:28,874][starVLA.training.train_starvla][INFO] - Step 4215, Loss: {'train/loss': 0.34695011377334595, 'train/grad_norm_pre_clip': 18.173673629760742, 'timing/data': 0.0007055001333355904, 'timing/model': 0.9315631994977593, 'learning_rate/qwen_vl_interface': 3.8336842782996245e-06, 'learning_rate/action_model': 3.8336842782996244e-05, 'epoch': 1.74}) +[2026-06-17 23:17:33,564][starVLA.training.train_starvla][INFO] - Step 4216, Loss: {'train/loss': 0.31488165259361267, 'train/grad_norm_pre_clip': 11.900585174560547, 'timing/data': 0.0009573949500918388, 'timing/model': 0.9395170249044895, 'learning_rate/qwen_vl_interface': 3.831620110525766e-06, 'learning_rate/action_model': 3.831620110525767e-05, 'epoch': 1.74}) +[2026-06-17 23:17:37,791][starVLA.training.train_starvla][INFO] - Step 4217, Loss: {'train/loss': 0.3192385733127594, 'train/grad_norm_pre_clip': 18.456363677978516, 'timing/data': 0.0005267439410090446, 'timing/model': 0.9398776767775416, 'learning_rate/qwen_vl_interface': 3.829556236783588e-06, 'learning_rate/action_model': 3.8295562367835884e-05, 'epoch': 1.74}) +[2026-06-17 23:17:42,207][starVLA.training.train_starvla][INFO] - Step 4218, Loss: {'train/loss': 0.532945990562439, 'train/grad_norm_pre_clip': 24.03860855102539, 'timing/data': 0.0006603552028536797, 'timing/model': 1.1681591281667352, 'learning_rate/qwen_vl_interface': 3.827492657500931e-06, 'learning_rate/action_model': 3.8274926575009305e-05, 'epoch': 1.74}) +[2026-06-17 23:17:46,740][starVLA.training.train_starvla][INFO] - Step 4219, Loss: {'train/loss': 0.3106115162372589, 'train/grad_norm_pre_clip': 16.93703842163086, 'timing/data': 0.0005482286214828491, 'timing/model': 1.3881854861974716, 'learning_rate/qwen_vl_interface': 3.8254293731055754e-06, 'learning_rate/action_model': 3.825429373105576e-05, 'epoch': 1.74}) +[2026-06-17 23:17:50,459][starVLA.training.train_starvla][INFO] - Step 4220, Loss: {'train/loss': 0.34487584233283997, 'train/grad_norm_pre_clip': 15.470205307006836, 'timing/data': 0.0010119015350937843, 'timing/model': 0.9505831757560372, 'learning_rate/qwen_vl_interface': 3.823366384025243e-06, 'learning_rate/action_model': 3.8233663840252426e-05, 'epoch': 1.74}) +[2026-06-17 23:17:54,373][starVLA.training.train_starvla][INFO] - Step 4221, Loss: {'train/loss': 0.44188711047172546, 'train/grad_norm_pre_clip': 16.799449920654297, 'timing/data': 0.0005339290946722031, 'timing/model': 1.017182701267302, 'learning_rate/qwen_vl_interface': 3.821303690687595e-06, 'learning_rate/action_model': 3.821303690687595e-05, 'epoch': 1.74}) +[2026-06-17 23:17:58,212][starVLA.training.train_starvla][INFO] - Step 4222, Loss: {'train/loss': 0.5248028039932251, 'train/grad_norm_pre_clip': 21.771018981933594, 'timing/data': 0.0006052283570170403, 'timing/model': 0.9339513992890716, 'learning_rate/qwen_vl_interface': 3.81924129352023e-06, 'learning_rate/action_model': 3.8192412935202304e-05, 'epoch': 1.74}) +[2026-06-17 23:18:01,999][starVLA.training.train_starvla][INFO] - Step 4223, Loss: {'train/loss': 0.3083885610103607, 'train/grad_norm_pre_clip': 17.64130210876465, 'timing/data': 0.0004995465278625488, 'timing/model': 0.9387622931972146, 'learning_rate/qwen_vl_interface': 3.817179192950682e-06, 'learning_rate/action_model': 3.8171791929506826e-05, 'epoch': 1.74}) +[2026-06-17 23:18:05,835][starVLA.training.train_starvla][INFO] - Step 4224, Loss: {'train/loss': 0.2154519259929657, 'train/grad_norm_pre_clip': 17.218372344970703, 'timing/data': 0.0009462190791964531, 'timing/model': 0.93426647875458, 'learning_rate/qwen_vl_interface': 3.815117389406432e-06, 'learning_rate/action_model': 3.8151173894064324e-05, 'epoch': 1.74}) +[2026-06-17 23:18:09,731][starVLA.training.train_starvla][INFO] - Step 4225, Loss: {'train/loss': 0.37185150384902954, 'train/grad_norm_pre_clip': 85.75996398925781, 'timing/data': 0.0008729947730898857, 'timing/model': 1.1249796878546476, 'learning_rate/qwen_vl_interface': 3.8130558833148906e-06, 'learning_rate/action_model': 3.813055883314891e-05, 'epoch': 1.74}) +[2026-06-17 23:18:13,405][starVLA.training.train_starvla][INFO] - Step 4226, Loss: {'train/loss': 0.31299203634262085, 'train/grad_norm_pre_clip': 26.710355758666992, 'timing/data': 0.0005904138088226318, 'timing/model': 0.943900995887816, 'learning_rate/qwen_vl_interface': 3.81099467510341e-06, 'learning_rate/action_model': 3.81099467510341e-05, 'epoch': 1.74}) +[2026-06-17 23:18:17,687][starVLA.training.train_starvla][INFO] - Step 4227, Loss: {'train/loss': 0.23335082828998566, 'train/grad_norm_pre_clip': 14.578043937683105, 'timing/data': 0.0005525536835193634, 'timing/model': 0.9343223953619599, 'learning_rate/qwen_vl_interface': 3.8089337651992815e-06, 'learning_rate/action_model': 3.808933765199281e-05, 'epoch': 1.74}) +[2026-06-17 23:18:21,802][starVLA.training.train_starvla][INFO] - Step 4228, Loss: {'train/loss': 0.49043118953704834, 'train/grad_norm_pre_clip': 125.31990051269531, 'timing/data': 0.0006599975749850273, 'timing/model': 1.0313761001452804, 'learning_rate/qwen_vl_interface': 3.8068731540297344e-06, 'learning_rate/action_model': 3.806873154029734e-05, 'epoch': 1.74}) +[2026-06-17 23:18:29,003][starVLA.training.train_starvla][INFO] - Step 4229, Loss: {'train/loss': 0.37283119559288025, 'train/grad_norm_pre_clip': 20.17281150817871, 'timing/data': 0.0010627694427967072, 'timing/model': 0.9797360934317112, 'learning_rate/qwen_vl_interface': 3.8048128420219366e-06, 'learning_rate/action_model': 3.804812842021936e-05, 'epoch': 1.74}) +[2026-06-17 23:18:33,025][starVLA.training.train_starvla][INFO] - Step 4230, Loss: {'train/loss': 0.826273500919342, 'train/grad_norm_pre_clip': 25.687257766723633, 'timing/data': 0.000568087212741375, 'timing/model': 0.935658291913569, 'learning_rate/qwen_vl_interface': 3.80275282960299e-06, 'learning_rate/action_model': 3.80275282960299e-05, 'epoch': 1.74}) +[2026-06-17 23:18:36,819][starVLA.training.train_starvla][INFO] - Step 4231, Loss: {'train/loss': 0.16700708866119385, 'train/grad_norm_pre_clip': 10.037450790405273, 'timing/data': 0.0008143046870827675, 'timing/model': 1.1205927282571793, 'learning_rate/qwen_vl_interface': 3.8006931171999394e-06, 'learning_rate/action_model': 3.800693117199939e-05, 'epoch': 1.74}) +[2026-06-17 23:18:40,638][starVLA.training.train_starvla][INFO] - Step 4232, Loss: {'train/loss': 0.16567657887935638, 'train/grad_norm_pre_clip': 9.627734184265137, 'timing/data': 0.0006055952981114388, 'timing/model': 1.1279979534447193, 'learning_rate/qwen_vl_interface': 3.798633705239766e-06, 'learning_rate/action_model': 3.7986337052397655e-05, 'epoch': 1.74}) +[2026-06-17 23:18:44,269][starVLA.training.train_starvla][INFO] - Step 4233, Loss: {'train/loss': 0.4004459083080292, 'train/grad_norm_pre_clip': 17.266489028930664, 'timing/data': 0.000514407642185688, 'timing/model': 0.9311700444668531, 'learning_rate/qwen_vl_interface': 3.796574594149387e-06, 'learning_rate/action_model': 3.796574594149387e-05, 'epoch': 1.74}) +[2026-06-17 23:18:48,125][starVLA.training.train_starvla][INFO] - Step 4234, Loss: {'train/loss': 0.49231094121932983, 'train/grad_norm_pre_clip': 19.39185333251953, 'timing/data': 0.0005047041922807693, 'timing/model': 0.9587008533999324, 'learning_rate/qwen_vl_interface': 3.794515784355657e-06, 'learning_rate/action_model': 3.7945157843556575e-05, 'epoch': 1.74}) +[2026-06-17 23:18:52,167][starVLA.training.train_starvla][INFO] - Step 4235, Loss: {'train/loss': 0.6131550669670105, 'train/grad_norm_pre_clip': 25.55134391784668, 'timing/data': 0.0010000597685575485, 'timing/model': 0.9751439280807972, 'learning_rate/qwen_vl_interface': 3.7924572762853727e-06, 'learning_rate/action_model': 3.7924572762853725e-05, 'epoch': 1.74}) +[2026-06-17 23:18:55,997][starVLA.training.train_starvla][INFO] - Step 4236, Loss: {'train/loss': 0.46794986724853516, 'train/grad_norm_pre_clip': 30.991737365722656, 'timing/data': 0.0005257679149508476, 'timing/model': 0.941763648763299, 'learning_rate/qwen_vl_interface': 3.7903990703652643e-06, 'learning_rate/action_model': 3.790399070365264e-05, 'epoch': 1.74}) +[2026-06-17 23:18:59,790][starVLA.training.train_starvla][INFO] - Step 4237, Loss: {'train/loss': 0.39884522557258606, 'train/grad_norm_pre_clip': 20.932859420776367, 'timing/data': 0.0005621481686830521, 'timing/model': 0.9322649724781513, 'learning_rate/qwen_vl_interface': 3.7883411670219976e-06, 'learning_rate/action_model': 3.788341167021998e-05, 'epoch': 1.74}) +[2026-06-17 23:19:03,642][starVLA.training.train_starvla][INFO] - Step 4238, Loss: {'train/loss': 0.1800665259361267, 'train/grad_norm_pre_clip': 14.19265365600586, 'timing/data': 0.0005039162933826447, 'timing/model': 1.1200367268174887, 'learning_rate/qwen_vl_interface': 3.7862835666821795e-06, 'learning_rate/action_model': 3.78628356668218e-05, 'epoch': 1.74}) +[2026-06-17 23:19:07,447][starVLA.training.train_starvla][INFO] - Step 4239, Loss: {'train/loss': 0.20850083231925964, 'train/grad_norm_pre_clip': 11.188638687133789, 'timing/data': 0.0005326811224222183, 'timing/model': 1.1004645954817533, 'learning_rate/qwen_vl_interface': 3.7842262697723543e-06, 'learning_rate/action_model': 3.784226269772354e-05, 'epoch': 1.74}) +[2026-06-17 23:19:11,146][starVLA.training.train_starvla][INFO] - Step 4240, Loss: {'train/loss': 0.7228461503982544, 'train/grad_norm_pre_clip': 25.63675880432129, 'timing/data': 0.0005001556128263474, 'timing/model': 0.9566613063216209, 'learning_rate/qwen_vl_interface': 3.782169276719e-06, 'learning_rate/action_model': 3.782169276719e-05, 'epoch': 1.75}) +[2026-06-17 23:19:15,049][starVLA.training.train_starvla][INFO] - Step 4241, Loss: {'train/loss': 0.42401123046875, 'train/grad_norm_pre_clip': 15.65144157409668, 'timing/data': 0.0008449023589491844, 'timing/model': 0.9383988110348582, 'learning_rate/qwen_vl_interface': 3.7801125879485355e-06, 'learning_rate/action_model': 3.7801125879485356e-05, 'epoch': 1.75}) +[2026-06-17 23:19:19,040][starVLA.training.train_starvla][INFO] - Step 4242, Loss: {'train/loss': 0.41440480947494507, 'train/grad_norm_pre_clip': 51.020530700683594, 'timing/data': 0.0005116118118166924, 'timing/model': 0.956067661754787, 'learning_rate/qwen_vl_interface': 3.77805620388731e-06, 'learning_rate/action_model': 3.7780562038873105e-05, 'epoch': 1.75}) +[2026-06-17 23:19:22,759][starVLA.training.train_starvla][INFO] - Step 4243, Loss: {'train/loss': 0.2817120850086212, 'train/grad_norm_pre_clip': 16.04772186279297, 'timing/data': 0.0005527017638087273, 'timing/model': 0.9377392614260316, 'learning_rate/qwen_vl_interface': 3.7760001249616204e-06, 'learning_rate/action_model': 3.77600012496162e-05, 'epoch': 1.75}) +[2026-06-17 23:19:26,590][starVLA.training.train_starvla][INFO] - Step 4244, Loss: {'train/loss': 0.44721463322639465, 'train/grad_norm_pre_clip': 19.075939178466797, 'timing/data': 0.000916079618036747, 'timing/model': 0.9426185032352805, 'learning_rate/qwen_vl_interface': 3.7739443515976898e-06, 'learning_rate/action_model': 3.7739443515976894e-05, 'epoch': 1.75}) +[2026-06-17 23:19:30,442][starVLA.training.train_starvla][INFO] - Step 4245, Loss: {'train/loss': 0.3102027177810669, 'train/grad_norm_pre_clip': 17.955045700073242, 'timing/data': 0.0013153636828064919, 'timing/model': 1.119431040249765, 'learning_rate/qwen_vl_interface': 3.7718888842216822e-06, 'learning_rate/action_model': 3.771888884221682e-05, 'epoch': 1.75}) +[2026-06-17 23:19:34,190][starVLA.training.train_starvla][INFO] - Step 4246, Loss: {'train/loss': 0.2550186812877655, 'train/grad_norm_pre_clip': 19.485666275024414, 'timing/data': 0.0005329092964529991, 'timing/model': 0.9406919106841087, 'learning_rate/qwen_vl_interface': 3.7698337232597002e-06, 'learning_rate/action_model': 3.7698337232597e-05, 'epoch': 1.75}) +[2026-06-17 23:19:38,065][starVLA.training.train_starvla][INFO] - Step 4247, Loss: {'train/loss': 0.5430612564086914, 'train/grad_norm_pre_clip': 25.090160369873047, 'timing/data': 0.0010112933814525604, 'timing/model': 0.9616364380344748, 'learning_rate/qwen_vl_interface': 3.7677788691377814e-06, 'learning_rate/action_model': 3.7677788691377817e-05, 'epoch': 1.75}) +[2026-06-17 23:19:42,006][starVLA.training.train_starvla][INFO] - Step 4248, Loss: {'train/loss': 0.4810972809791565, 'train/grad_norm_pre_clip': 19.582950592041016, 'timing/data': 0.0014311755076050758, 'timing/model': 1.0271742213517427, 'learning_rate/qwen_vl_interface': 3.7657243222818978e-06, 'learning_rate/action_model': 3.7657243222818976e-05, 'epoch': 1.75}) +[2026-06-17 23:19:46,088][starVLA.training.train_starvla][INFO] - Step 4249, Loss: {'train/loss': 0.15589754283428192, 'train/grad_norm_pre_clip': 13.106539726257324, 'timing/data': 0.0005465792492032051, 'timing/model': 0.9492218028753996, 'learning_rate/qwen_vl_interface': 3.7636700831179575e-06, 'learning_rate/action_model': 3.763670083117957e-05, 'epoch': 1.75}) +[2026-06-17 23:19:50,129][starVLA.training.train_starvla][INFO] - Step 4250, Loss: {'train/loss': 0.17218473553657532, 'train/grad_norm_pre_clip': 11.11037826538086, 'timing/data': 0.0005335677415132523, 'timing/model': 0.9633672945201397, 'learning_rate/qwen_vl_interface': 3.7616161520718096e-06, 'learning_rate/action_model': 3.76161615207181e-05, 'epoch': 1.75}) +[2026-06-17 23:19:54,078][starVLA.training.train_starvla][INFO] - Step 4251, Loss: {'train/loss': 0.21784473955631256, 'train/grad_norm_pre_clip': 14.242108345031738, 'timing/data': 0.0009672343730926514, 'timing/model': 1.1890479698777199, 'learning_rate/qwen_vl_interface': 3.7595625295692352e-06, 'learning_rate/action_model': 3.759562529569235e-05, 'epoch': 1.75}) +[2026-06-17 23:19:57,969][starVLA.training.train_starvla][INFO] - Step 4252, Loss: {'train/loss': 0.2019568681716919, 'train/grad_norm_pre_clip': 21.099069595336914, 'timing/data': 0.0012817401438951492, 'timing/model': 1.1489817518740892, 'learning_rate/qwen_vl_interface': 3.7575092160359523e-06, 'learning_rate/action_model': 3.757509216035952e-05, 'epoch': 1.75}) +[2026-06-17 23:20:01,699][starVLA.training.train_starvla][INFO] - Step 4253, Loss: {'train/loss': 0.30547264218330383, 'train/grad_norm_pre_clip': 14.022313117980957, 'timing/data': 0.0005002478137612343, 'timing/model': 0.9541546301916242, 'learning_rate/qwen_vl_interface': 3.7554562118976135e-06, 'learning_rate/action_model': 3.7554562118976136e-05, 'epoch': 1.75}) +[2026-06-17 23:20:05,561][starVLA.training.train_starvla][INFO] - Step 4254, Loss: {'train/loss': 0.16982263326644897, 'train/grad_norm_pre_clip': 14.371345520019531, 'timing/data': 0.0004846462979912758, 'timing/model': 0.9340879898518324, 'learning_rate/qwen_vl_interface': 3.7534035175798113e-06, 'learning_rate/action_model': 3.753403517579811e-05, 'epoch': 1.75}) +[2026-06-17 23:20:09,502][starVLA.training.train_starvla][INFO] - Step 4255, Loss: {'train/loss': 0.19362278282642365, 'train/grad_norm_pre_clip': 15.095061302185059, 'timing/data': 0.0005617858842015266, 'timing/model': 0.9731270037591457, 'learning_rate/qwen_vl_interface': 3.7513511335080706e-06, 'learning_rate/action_model': 3.75135113350807e-05, 'epoch': 1.75}) +[2026-06-17 23:20:13,526][starVLA.training.train_starvla][INFO] - Step 4256, Loss: {'train/loss': 0.5913488864898682, 'train/grad_norm_pre_clip': 25.201766967773438, 'timing/data': 0.0005140677094459534, 'timing/model': 1.0440660817548633, 'learning_rate/qwen_vl_interface': 3.74929906010785e-06, 'learning_rate/action_model': 3.74929906010785e-05, 'epoch': 1.75}) +[2026-06-17 23:20:17,507][starVLA.training.train_starvla][INFO] - Step 4257, Loss: {'train/loss': 0.3464624583721161, 'train/grad_norm_pre_clip': 14.539078712463379, 'timing/data': 0.0008317967876791954, 'timing/model': 0.9701885571703315, 'learning_rate/qwen_vl_interface': 3.747247297804553e-06, 'learning_rate/action_model': 3.7472472978045526e-05, 'epoch': 1.75}) +[2026-06-17 23:20:21,477][starVLA.training.train_starvla][INFO] - Step 4258, Loss: {'train/loss': 0.20655500888824463, 'train/grad_norm_pre_clip': 13.94876766204834, 'timing/data': 0.0004840046167373657, 'timing/model': 1.1504646055400372, 'learning_rate/qwen_vl_interface': 3.745195847023506e-06, 'learning_rate/action_model': 3.745195847023506e-05, 'epoch': 1.75}) +[2026-06-17 23:20:25,164][starVLA.training.train_starvla][INFO] - Step 4259, Loss: {'train/loss': 0.582555890083313, 'train/grad_norm_pre_clip': 28.01031494140625, 'timing/data': 0.0006924858316779137, 'timing/model': 0.9415485290810466, 'learning_rate/qwen_vl_interface': 3.7431447081899795e-06, 'learning_rate/action_model': 3.7431447081899796e-05, 'epoch': 1.75}) +[2026-06-17 23:20:29,063][starVLA.training.train_starvla][INFO] - Step 4260, Loss: {'train/loss': 0.24995264410972595, 'train/grad_norm_pre_clip': 17.545055389404297, 'timing/data': 0.0009759180247783661, 'timing/model': 0.9869501255452633, 'learning_rate/qwen_vl_interface': 3.7410938817291753e-06, 'learning_rate/action_model': 3.741093881729175e-05, 'epoch': 1.75}) +[2026-06-17 23:20:33,024][starVLA.training.train_starvla][INFO] - Step 4261, Loss: {'train/loss': 0.3889375627040863, 'train/grad_norm_pre_clip': 18.37800407409668, 'timing/data': 0.0012122727930545807, 'timing/model': 1.0255275126546621, 'learning_rate/qwen_vl_interface': 3.739043368066234e-06, 'learning_rate/action_model': 3.739043368066234e-05, 'epoch': 1.75}) +[2026-06-17 23:20:37,033][starVLA.training.train_starvla][INFO] - Step 4262, Loss: {'train/loss': 0.4809006154537201, 'train/grad_norm_pre_clip': 17.740524291992188, 'timing/data': 0.0008012382313609123, 'timing/model': 0.942924503237009, 'learning_rate/qwen_vl_interface': 3.7369931676262304e-06, 'learning_rate/action_model': 3.7369931676262304e-05, 'epoch': 1.75}) +[2026-06-17 23:20:40,940][starVLA.training.train_starvla][INFO] - Step 4263, Loss: {'train/loss': 0.1177520751953125, 'train/grad_norm_pre_clip': 20.10930824279785, 'timing/data': 0.0005072001367807388, 'timing/model': 0.9684732584282756, 'learning_rate/qwen_vl_interface': 3.7349432808341673e-06, 'learning_rate/action_model': 3.7349432808341674e-05, 'epoch': 1.75}) +[2026-06-17 23:20:45,008][starVLA.training.train_starvla][INFO] - Step 4264, Loss: {'train/loss': 0.4187706708908081, 'train/grad_norm_pre_clip': 25.873985290527344, 'timing/data': 0.0008120564743876457, 'timing/model': 1.111550078727305, 'learning_rate/qwen_vl_interface': 3.7328937081149952e-06, 'learning_rate/action_model': 3.732893708114995e-05, 'epoch': 1.76}) +[2026-06-17 23:20:48,837][starVLA.training.train_starvla][INFO] - Step 4265, Loss: {'train/loss': 0.2870471477508545, 'train/grad_norm_pre_clip': 16.588701248168945, 'timing/data': 0.0007749395444989204, 'timing/model': 1.1323222201317549, 'learning_rate/qwen_vl_interface': 3.7308444498935894e-06, 'learning_rate/action_model': 3.730844449893589e-05, 'epoch': 1.76}) +[2026-06-17 23:20:52,618][starVLA.training.train_starvla][INFO] - Step 4266, Loss: {'train/loss': 0.33891838788986206, 'train/grad_norm_pre_clip': 17.006044387817383, 'timing/data': 0.0014490783214569092, 'timing/model': 0.9587228409945965, 'learning_rate/qwen_vl_interface': 3.7287955065947634e-06, 'learning_rate/action_model': 3.728795506594763e-05, 'epoch': 1.76}) +[2026-06-17 23:20:56,523][starVLA.training.train_starvla][INFO] - Step 4267, Loss: {'train/loss': 0.4877412021160126, 'train/grad_norm_pre_clip': 20.922361373901367, 'timing/data': 0.0004689497873187065, 'timing/model': 0.9485789742320776, 'learning_rate/qwen_vl_interface': 3.7267468786432655e-06, 'learning_rate/action_model': 3.7267468786432654e-05, 'epoch': 1.76}) +[2026-06-17 23:21:00,368][starVLA.training.train_starvla][INFO] - Step 4268, Loss: {'train/loss': 0.12470583617687225, 'train/grad_norm_pre_clip': 26.43960952758789, 'timing/data': 0.0005270382389426231, 'timing/model': 0.942634210921824, 'learning_rate/qwen_vl_interface': 3.7246985664637787e-06, 'learning_rate/action_model': 3.724698566463778e-05, 'epoch': 1.76}) +[2026-06-17 23:21:04,247][starVLA.training.train_starvla][INFO] - Step 4269, Loss: {'train/loss': 0.4950520992279053, 'train/grad_norm_pre_clip': 17.518177032470703, 'timing/data': 0.0008002957329154015, 'timing/model': 0.9412355599924922, 'learning_rate/qwen_vl_interface': 3.722650570480922e-06, 'learning_rate/action_model': 3.722650570480922e-05, 'epoch': 1.76}) +[2026-06-17 23:21:08,135][starVLA.training.train_starvla][INFO] - Step 4270, Loss: {'train/loss': 0.2747654914855957, 'train/grad_norm_pre_clip': 16.6514835357666, 'timing/data': 0.0008815126493573189, 'timing/model': 0.9420825717970729, 'learning_rate/qwen_vl_interface': 3.7206028911192445e-06, 'learning_rate/action_model': 3.720602891119244e-05, 'epoch': 1.76}) +[2026-06-17 23:21:12,094][starVLA.training.train_starvla][INFO] - Step 4271, Loss: {'train/loss': 0.2970118820667267, 'train/grad_norm_pre_clip': 15.016721725463867, 'timing/data': 0.0004883008077740669, 'timing/model': 0.9662582036107779, 'learning_rate/qwen_vl_interface': 3.7185555288032315e-06, 'learning_rate/action_model': 3.7185555288032316e-05, 'epoch': 1.76}) +[2026-06-17 23:21:15,985][starVLA.training.train_starvla][INFO] - Step 4272, Loss: {'train/loss': 0.36011531949043274, 'train/grad_norm_pre_clip': 18.674747467041016, 'timing/data': 0.0004926426336169243, 'timing/model': 1.1536875097081065, 'learning_rate/qwen_vl_interface': 3.716508483957308e-06, 'learning_rate/action_model': 3.716508483957308e-05, 'epoch': 1.76}) +[2026-06-17 23:21:19,717][starVLA.training.train_starvla][INFO] - Step 4273, Loss: {'train/loss': 0.4085274934768677, 'train/grad_norm_pre_clip': 21.484804153442383, 'timing/data': 0.0009845597669482231, 'timing/model': 0.9522658167406917, 'learning_rate/qwen_vl_interface': 3.7144617570058245e-06, 'learning_rate/action_model': 3.7144617570058245e-05, 'epoch': 1.76}) +[2026-06-17 23:21:24,270][starVLA.training.train_starvla][INFO] - Step 4274, Loss: {'train/loss': 0.41972410678863525, 'train/grad_norm_pre_clip': 23.314119338989258, 'timing/data': 0.0005884971469640732, 'timing/model': 1.0239495132118464, 'learning_rate/qwen_vl_interface': 3.7124153483730706e-06, 'learning_rate/action_model': 3.712415348373071e-05, 'epoch': 1.76}) +[2026-06-17 23:21:29,754][starVLA.training.train_starvla][INFO] - Step 4275, Loss: {'train/loss': 0.6226038932800293, 'train/grad_norm_pre_clip': 27.726449966430664, 'timing/data': 0.0005465298891067505, 'timing/model': 0.9863486681133509, 'learning_rate/qwen_vl_interface': 3.71036925848327e-06, 'learning_rate/action_model': 3.71036925848327e-05, 'epoch': 1.76}) +[2026-06-17 23:21:33,922][starVLA.training.train_starvla][INFO] - Step 4276, Loss: {'train/loss': 0.18753062188625336, 'train/grad_norm_pre_clip': 14.042211532592773, 'timing/data': 0.0008560148999094963, 'timing/model': 0.954875273630023, 'learning_rate/qwen_vl_interface': 3.7083234877605804e-06, 'learning_rate/action_model': 3.7083234877605806e-05, 'epoch': 1.76}) +[2026-06-17 23:21:38,908][starVLA.training.train_starvla][INFO] - Step 4277, Loss: {'train/loss': 0.4058699309825897, 'train/grad_norm_pre_clip': 25.603870391845703, 'timing/data': 0.001115991733968258, 'timing/model': 2.0363160418346524, 'learning_rate/qwen_vl_interface': 3.70627803662909e-06, 'learning_rate/action_model': 3.70627803662909e-05, 'epoch': 1.76}) +[2026-06-17 23:21:43,878][starVLA.training.train_starvla][INFO] - Step 4278, Loss: {'train/loss': 0.2334422916173935, 'train/grad_norm_pre_clip': 20.96576690673828, 'timing/data': 0.0005447948351502419, 'timing/model': 1.1281581604853272, 'learning_rate/qwen_vl_interface': 3.7042329055128216e-06, 'learning_rate/action_model': 3.704232905512821e-05, 'epoch': 1.76}) +[2026-06-17 23:21:47,721][starVLA.training.train_starvla][INFO] - Step 4279, Loss: {'train/loss': 0.1907534897327423, 'train/grad_norm_pre_clip': 16.0316104888916, 'timing/data': 0.0009302468970417976, 'timing/model': 1.1257069511339068, 'learning_rate/qwen_vl_interface': 3.7021880948357375e-06, 'learning_rate/action_model': 3.702188094835737e-05, 'epoch': 1.76}) +[2026-06-17 23:21:51,513][starVLA.training.train_starvla][INFO] - Step 4280, Loss: {'train/loss': 0.5321280360221863, 'train/grad_norm_pre_clip': 23.968278884887695, 'timing/data': 0.000835946761071682, 'timing/model': 1.0297797853127122, 'learning_rate/qwen_vl_interface': 3.700143605021726e-06, 'learning_rate/action_model': 3.7001436050217256e-05, 'epoch': 1.76}) +[2026-06-17 23:21:55,382][starVLA.training.train_starvla][INFO] - Step 4281, Loss: {'train/loss': 0.43521827459335327, 'train/grad_norm_pre_clip': 21.771406173706055, 'timing/data': 0.0005283793434500694, 'timing/model': 0.9514800747856498, 'learning_rate/qwen_vl_interface': 3.6980994364946133e-06, 'learning_rate/action_model': 3.698099436494613e-05, 'epoch': 1.76}) +[2026-06-17 23:21:59,237][starVLA.training.train_starvla][INFO] - Step 4282, Loss: {'train/loss': 0.538804292678833, 'train/grad_norm_pre_clip': 22.093408584594727, 'timing/data': 0.000949932262301445, 'timing/model': 0.9463741285726428, 'learning_rate/qwen_vl_interface': 3.696055589678156e-06, 'learning_rate/action_model': 3.6960555896781554e-05, 'epoch': 1.76}) +[2026-06-17 23:22:03,189][starVLA.training.train_starvla][INFO] - Step 4283, Loss: {'train/loss': 0.319243848323822, 'train/grad_norm_pre_clip': 25.5304012298584, 'timing/data': 0.000500798225402832, 'timing/model': 1.0181269906461239, 'learning_rate/qwen_vl_interface': 3.694012064996049e-06, 'learning_rate/action_model': 3.694012064996049e-05, 'epoch': 1.76}) +[2026-06-17 23:22:07,092][starVLA.training.train_starvla][INFO] - Step 4284, Loss: {'train/loss': 0.3455958366394043, 'train/grad_norm_pre_clip': 20.19080352783203, 'timing/data': 0.0006545763462781906, 'timing/model': 0.9625488938763738, 'learning_rate/qwen_vl_interface': 3.6919688628719153e-06, 'learning_rate/action_model': 3.691968862871915e-05, 'epoch': 1.76}) +[2026-06-17 23:22:11,193][starVLA.training.train_starvla][INFO] - Step 4285, Loss: {'train/loss': 0.4794183075428009, 'train/grad_norm_pre_clip': 26.17764663696289, 'timing/data': 0.0005615754052996635, 'timing/model': 1.2107455441728234, 'learning_rate/qwen_vl_interface': 3.689925983729311e-06, 'learning_rate/action_model': 3.689925983729311e-05, 'epoch': 1.76}) +[2026-06-17 23:22:15,023][starVLA.training.train_starvla][INFO] - Step 4286, Loss: {'train/loss': 0.5505605936050415, 'train/grad_norm_pre_clip': 20.05280113220215, 'timing/data': 0.0006548389792442322, 'timing/model': 1.1138548096641898, 'learning_rate/qwen_vl_interface': 3.68788342799173e-06, 'learning_rate/action_model': 3.6878834279917294e-05, 'epoch': 1.76}) +[2026-06-17 23:22:18,814][starVLA.training.train_starvla][INFO] - Step 4287, Loss: {'train/loss': 0.22876830399036407, 'train/grad_norm_pre_clip': 16.7637996673584, 'timing/data': 0.0009105028584599495, 'timing/model': 1.032029907219112, 'learning_rate/qwen_vl_interface': 3.6858411960825953e-06, 'learning_rate/action_model': 3.685841196082595e-05, 'epoch': 1.76}) +[2026-06-17 23:22:22,783][starVLA.training.train_starvla][INFO] - Step 4288, Loss: {'train/loss': 0.5622425079345703, 'train/grad_norm_pre_clip': 25.46533203125, 'timing/data': 0.0005553616210818291, 'timing/model': 0.9383476236835122, 'learning_rate/qwen_vl_interface': 3.6837992884252636e-06, 'learning_rate/action_model': 3.683799288425264e-05, 'epoch': 1.77}) +[2026-06-17 23:22:26,597][starVLA.training.train_starvla][INFO] - Step 4289, Loss: {'train/loss': 0.5507761836051941, 'train/grad_norm_pre_clip': 28.08882713317871, 'timing/data': 0.0005605658516287804, 'timing/model': 0.9589751223102212, 'learning_rate/qwen_vl_interface': 3.6817577054430243e-06, 'learning_rate/action_model': 3.6817577054430244e-05, 'epoch': 1.77}) +[2026-06-17 23:22:30,349][starVLA.training.train_starvla][INFO] - Step 4290, Loss: {'train/loss': 0.4128873944282532, 'train/grad_norm_pre_clip': 17.7280216217041, 'timing/data': 0.0007693031802773476, 'timing/model': 0.962326368317008, 'learning_rate/qwen_vl_interface': 3.6797164475591003e-06, 'learning_rate/action_model': 3.6797164475591e-05, 'epoch': 1.77}) +[2026-06-17 23:22:34,240][starVLA.training.train_starvla][INFO] - Step 4291, Loss: {'train/loss': 0.39451050758361816, 'train/grad_norm_pre_clip': 20.320024490356445, 'timing/data': 0.0007849084213376045, 'timing/model': 0.9396133674308658, 'learning_rate/qwen_vl_interface': 3.677675515196648e-06, 'learning_rate/action_model': 3.677675515196648e-05, 'epoch': 1.77}) +[2026-06-17 23:22:38,232][starVLA.training.train_starvla][INFO] - Step 4292, Loss: {'train/loss': 0.29345837235450745, 'train/grad_norm_pre_clip': 17.99915885925293, 'timing/data': 0.0008122595027089119, 'timing/model': 1.1815934898331761, 'learning_rate/qwen_vl_interface': 3.6756349087787524e-06, 'learning_rate/action_model': 3.6756349087787524e-05, 'epoch': 1.77}) +[2026-06-17 23:22:41,972][starVLA.training.train_starvla][INFO] - Step 4293, Loss: {'train/loss': 0.32016074657440186, 'train/grad_norm_pre_clip': 17.231201171875, 'timing/data': 0.0010633403435349464, 'timing/model': 0.9918007645756006, 'learning_rate/qwen_vl_interface': 3.6735946287284325e-06, 'learning_rate/action_model': 3.6735946287284324e-05, 'epoch': 1.77}) +[2026-06-17 23:22:45,883][starVLA.training.train_starvla][INFO] - Step 4294, Loss: {'train/loss': 0.35997697710990906, 'train/grad_norm_pre_clip': 20.180370330810547, 'timing/data': 0.0006255647167563438, 'timing/model': 0.9952365886420012, 'learning_rate/qwen_vl_interface': 3.6715546754686444e-06, 'learning_rate/action_model': 3.6715546754686444e-05, 'epoch': 1.77}) +[2026-06-17 23:22:49,642][starVLA.training.train_starvla][INFO] - Step 4295, Loss: {'train/loss': 0.3668835461139679, 'train/grad_norm_pre_clip': 20.781036376953125, 'timing/data': 0.000673946924507618, 'timing/model': 0.9373888717964292, 'learning_rate/qwen_vl_interface': 3.6695150494222697e-06, 'learning_rate/action_model': 3.6695150494222695e-05, 'epoch': 1.77}) +[2026-06-17 23:22:53,551][starVLA.training.train_starvla][INFO] - Step 4296, Loss: {'train/loss': 0.20440542697906494, 'train/grad_norm_pre_clip': 47.53340148925781, 'timing/data': 0.0008404180407524109, 'timing/model': 0.9456517230719328, 'learning_rate/qwen_vl_interface': 3.6674757510121255e-06, 'learning_rate/action_model': 3.6674757510121255e-05, 'epoch': 1.77}) +[2026-06-17 23:22:57,504][starVLA.training.train_starvla][INFO] - Step 4297, Loss: {'train/loss': 0.4638082683086395, 'train/grad_norm_pre_clip': 20.33854866027832, 'timing/data': 0.0008801110088825226, 'timing/model': 0.9622187716886401, 'learning_rate/qwen_vl_interface': 3.6654367806609615e-06, 'learning_rate/action_model': 3.665436780660961e-05, 'epoch': 1.77}) +[2026-06-17 23:23:01,376][starVLA.training.train_starvla][INFO] - Step 4298, Loss: {'train/loss': 0.2835729122161865, 'train/grad_norm_pre_clip': 18.771453857421875, 'timing/data': 0.0005005793645977974, 'timing/model': 1.1530742775648832, 'learning_rate/qwen_vl_interface': 3.6633981387914597e-06, 'learning_rate/action_model': 3.66339813879146e-05, 'epoch': 1.77}) +[2026-06-17 23:23:05,212][starVLA.training.train_starvla][INFO] - Step 4299, Loss: {'train/loss': 0.42057275772094727, 'train/grad_norm_pre_clip': 46.5582275390625, 'timing/data': 0.0005214391276240349, 'timing/model': 1.1029188595712185, 'learning_rate/qwen_vl_interface': 3.6613598258262295e-06, 'learning_rate/action_model': 3.6613598258262296e-05, 'epoch': 1.77}) +[2026-06-17 23:25:15,147][starVLA.training.train_starvla][INFO] - Step 4300, Loss: {'train/loss': 0.37654775381088257, 'train/grad_norm_pre_clip': 19.904287338256836, 'eval/loss': 0.44883406162261963, 'eval/latency_6/loss': 0.44883406162261963, 'eval/demon_fire_precision': 0.954271197205462, 'eval/demon_fire_recall': 0.9712346477052359, 'eval/demon_fire_f1': 0.9626781995835335, 'eval/demon_fire_rate_model': 0.8761825264329438, 'eval/demon_fire_rate_teacher': 0.860879243183083, 'eval/demon_left_precision': 0.8775055679287305, 'eval/demon_left_recall': 0.6912280701754386, 'eval/demon_left_f1': 0.773307163886163, 'eval/demon_left_rate_model': 0.12493043962159155, 'eval/demon_left_rate_teacher': 0.15859766277128548, 'eval/demon_right_precision': 0.7874617737003058, 'eval/demon_right_recall': 0.7910906298003072, 'eval/demon_right_f1': 0.7892720306513409, 'eval/demon_right_rate_model': 0.18196994991652754, 'eval/demon_right_rate_teacher': 0.18113522537562604, 'eval/demon_move_f1': 0.7812895972687519, 'eval/demon_cc_f1': 0.8999939005897335, 'eval/latency_6/demon_fire_precision': 0.954271197205462, 'eval/latency_6/demon_fire_recall': 0.9712346477052359, 'eval/latency_6/demon_fire_f1': 0.9626781995835335, 'eval/latency_6/demon_fire_rate_model': 0.8761825264329438, 'eval/latency_6/demon_fire_rate_teacher': 0.860879243183083, 'eval/latency_6/demon_left_precision': 0.8775055679287305, 'eval/latency_6/demon_left_recall': 0.6912280701754386, 'eval/latency_6/demon_left_f1': 0.773307163886163, 'eval/latency_6/demon_left_rate_model': 0.12493043962159155, 'eval/latency_6/demon_left_rate_teacher': 0.15859766277128548, 'eval/latency_6/demon_right_precision': 0.7874617737003058, 'eval/latency_6/demon_right_recall': 0.7910906298003072, 'eval/latency_6/demon_right_f1': 0.7892720306513409, 'eval/latency_6/demon_right_rate_model': 0.18196994991652754, 'eval/latency_6/demon_right_rate_teacher': 0.18113522537562604, 'eval/latency_6/demon_move_f1': 0.7812895972687519, 'eval/latency_6/demon_cc_f1': 0.8999939005897335, 'timing/data': 0.0005243076011538506, 'timing/model': 0.9550046948716044, 'learning_rate/qwen_vl_interface': 3.6593218421878157e-06, 'learning_rate/action_model': 3.659321842187816e-05, 'epoch': 1.77}) +[2026-06-17 23:25:15,151][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_4300_state +[2026-06-17 23:25:15,170][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 23:25:22,036][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_4300_state/model.safetensors +[2026-06-17 23:25:37,578][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_4300_state/optimizer.bin +[2026-06-17 23:25:37,580][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_4300_state/sampler.bin +[2026-06-17 23:25:37,580][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_4300_state/sampler_1.bin +[2026-06-17 23:25:37,583][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_4300_state/random_states_0.pkl +[2026-06-17 23:25:37,587][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4300 +[2026-06-17 23:25:37,587][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 23:25:37,600][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 23:25:41,460][starVLA.training.train_starvla][INFO] - Step 4301, Loss: {'train/loss': 0.6211993098258972, 'train/grad_norm_pre_clip': 28.442873001098633, 'timing/data': 0.000712142325937748, 'timing/model': 1.1854218440130353, 'learning_rate/qwen_vl_interface': 3.657284188298697e-06, 'learning_rate/action_model': 3.657284188298697e-05, 'epoch': 1.77}) +[2026-06-17 23:25:45,073][starVLA.training.train_starvla][INFO] - Step 4302, Loss: {'train/loss': 0.17829638719558716, 'train/grad_norm_pre_clip': 12.070910453796387, 'timing/data': 0.0005324315279722214, 'timing/model': 0.933105131611228, 'learning_rate/qwen_vl_interface': 3.65524686458128e-06, 'learning_rate/action_model': 3.65524686458128e-05, 'epoch': 1.77}) +[2026-06-17 23:25:48,919][starVLA.training.train_starvla][INFO] - Step 4303, Loss: {'train/loss': 0.25708475708961487, 'train/grad_norm_pre_clip': 16.3730525970459, 'timing/data': 0.0005310224369168282, 'timing/model': 0.9476668126881123, 'learning_rate/qwen_vl_interface': 3.653209871457904e-06, 'learning_rate/action_model': 3.653209871457904e-05, 'epoch': 1.77}) +[2026-06-17 23:25:52,733][starVLA.training.train_starvla][INFO] - Step 4304, Loss: {'train/loss': 0.4776049852371216, 'train/grad_norm_pre_clip': 20.729312896728516, 'timing/data': 0.000519314780831337, 'timing/model': 0.9508717935532331, 'learning_rate/qwen_vl_interface': 3.6511732093508376e-06, 'learning_rate/action_model': 3.651173209350838e-05, 'epoch': 1.77}) +[2026-06-17 23:25:56,540][starVLA.training.train_starvla][INFO] - Step 4305, Loss: {'train/loss': 0.14018140733242035, 'train/grad_norm_pre_clip': 12.886419296264648, 'timing/data': 0.0008046776056289673, 'timing/model': 0.9497632877901196, 'learning_rate/qwen_vl_interface': 3.6491368786822885e-06, 'learning_rate/action_model': 3.6491368786822884e-05, 'epoch': 1.77}) +[2026-06-17 23:26:00,392][starVLA.training.train_starvla][INFO] - Step 4306, Loss: {'train/loss': 0.2860359847545624, 'train/grad_norm_pre_clip': 14.295350074768066, 'timing/data': 0.0005312804132699966, 'timing/model': 0.9727075407281518, 'learning_rate/qwen_vl_interface': 3.647100879874386e-06, 'learning_rate/action_model': 3.6471008798743855e-05, 'epoch': 1.77}) +[2026-06-17 23:26:04,302][starVLA.training.train_starvla][INFO] - Step 4307, Loss: {'train/loss': 0.4936329424381256, 'train/grad_norm_pre_clip': 24.262210845947266, 'timing/data': 0.0009311623871326447, 'timing/model': 1.1613928005099297, 'learning_rate/qwen_vl_interface': 3.645065213349195e-06, 'learning_rate/action_model': 3.645065213349195e-05, 'epoch': 1.77}) +[2026-06-17 23:26:08,029][starVLA.training.train_starvla][INFO] - Step 4308, Loss: {'train/loss': 0.5280616283416748, 'train/grad_norm_pre_clip': 20.254045486450195, 'timing/data': 0.0010061506181955338, 'timing/model': 0.9511506864801049, 'learning_rate/qwen_vl_interface': 3.6430298795287135e-06, 'learning_rate/action_model': 3.643029879528713e-05, 'epoch': 1.77}) +[2026-06-17 23:26:11,950][starVLA.training.train_starvla][INFO] - Step 4309, Loss: {'train/loss': 0.4951744079589844, 'train/grad_norm_pre_clip': 19.574323654174805, 'timing/data': 0.0008892519399523735, 'timing/model': 0.962270824238658, 'learning_rate/qwen_vl_interface': 3.640994878834866e-06, 'learning_rate/action_model': 3.640994878834866e-05, 'epoch': 1.77}) +[2026-06-17 23:26:15,835][starVLA.training.train_starvla][INFO] - Step 4310, Loss: {'train/loss': 0.46555086970329285, 'train/grad_norm_pre_clip': 18.013633728027344, 'timing/data': 0.0006905961781740189, 'timing/model': 0.9447595635429025, 'learning_rate/qwen_vl_interface': 3.6389602116895136e-06, 'learning_rate/action_model': 3.6389602116895134e-05, 'epoch': 1.77}) +[2026-06-17 23:26:19,693][starVLA.training.train_starvla][INFO] - Step 4311, Loss: {'train/loss': 0.2556358277797699, 'train/grad_norm_pre_clip': 16.70983123779297, 'timing/data': 0.0005171876400709152, 'timing/model': 0.9546652417629957, 'learning_rate/qwen_vl_interface': 3.6369258785144395e-06, 'learning_rate/action_model': 3.636925878514439e-05, 'epoch': 1.77}) +[2026-06-17 23:26:23,509][starVLA.training.train_starvla][INFO] - Step 4312, Loss: {'train/loss': 0.2888798713684082, 'train/grad_norm_pre_clip': 17.623106002807617, 'timing/data': 0.0005058636888861656, 'timing/model': 0.9455393655225635, 'learning_rate/qwen_vl_interface': 3.6348918797313693e-06, 'learning_rate/action_model': 3.634891879731369e-05, 'epoch': 1.77}) +[2026-06-17 23:26:27,465][starVLA.training.train_starvla][INFO] - Step 4313, Loss: {'train/loss': 0.19638298451900482, 'train/grad_norm_pre_clip': 13.149604797363281, 'timing/data': 0.0005664322525262833, 'timing/model': 1.2735898224636912, 'learning_rate/qwen_vl_interface': 3.6328582157619495e-06, 'learning_rate/action_model': 3.6328582157619496e-05, 'epoch': 1.78}) +[2026-06-17 23:26:31,273][starVLA.training.train_starvla][INFO] - Step 4314, Loss: {'train/loss': 0.3735100328922272, 'train/grad_norm_pre_clip': 16.476484298706055, 'timing/data': 0.0005910973995923996, 'timing/model': 1.0758033944293857, 'learning_rate/qwen_vl_interface': 3.6308248870277613e-06, 'learning_rate/action_model': 3.630824887027761e-05, 'epoch': 1.78}) +[2026-06-17 23:26:35,664][starVLA.training.train_starvla][INFO] - Step 4315, Loss: {'train/loss': 0.35454297065734863, 'train/grad_norm_pre_clip': 18.807126998901367, 'timing/data': 0.0006427336484193802, 'timing/model': 0.9377567414194345, 'learning_rate/qwen_vl_interface': 3.628791893950316e-06, 'learning_rate/action_model': 3.628791893950316e-05, 'epoch': 1.78}) +[2026-06-17 23:26:39,592][starVLA.training.train_starvla][INFO] - Step 4316, Loss: {'train/loss': 0.5346657633781433, 'train/grad_norm_pre_clip': 25.1375789642334, 'timing/data': 0.0005435282364487648, 'timing/model': 0.9509149268269539, 'learning_rate/qwen_vl_interface': 3.6267592369510563e-06, 'learning_rate/action_model': 3.6267592369510564e-05, 'epoch': 1.78}) +[2026-06-17 23:26:43,438][starVLA.training.train_starvla][INFO] - Step 4317, Loss: {'train/loss': 0.45858731865882874, 'train/grad_norm_pre_clip': 26.4456729888916, 'timing/data': 0.0005645947530865669, 'timing/model': 0.9380114246159792, 'learning_rate/qwen_vl_interface': 3.624726916451355e-06, 'learning_rate/action_model': 3.624726916451355e-05, 'epoch': 1.78}) +[2026-06-17 23:26:47,242][starVLA.training.train_starvla][INFO] - Step 4318, Loss: {'train/loss': 0.6208337545394897, 'train/grad_norm_pre_clip': 25.390228271484375, 'timing/data': 0.0005447342991828918, 'timing/model': 0.9340914078056812, 'learning_rate/qwen_vl_interface': 3.622694932872509e-06, 'learning_rate/action_model': 3.622694932872509e-05, 'epoch': 1.78}) +[2026-06-17 23:26:51,027][starVLA.training.train_starvla][INFO] - Step 4319, Loss: {'train/loss': 0.4051663279533386, 'train/grad_norm_pre_clip': 18.229616165161133, 'timing/data': 0.0005580270662903786, 'timing/model': 0.9292292557656765, 'learning_rate/qwen_vl_interface': 3.620663286635758e-06, 'learning_rate/action_model': 3.620663286635758e-05, 'epoch': 1.78}) +[2026-06-17 23:26:54,876][starVLA.training.train_starvla][INFO] - Step 4320, Loss: {'train/loss': 0.25863784551620483, 'train/grad_norm_pre_clip': 20.463058471679688, 'timing/data': 0.0005599036812782288, 'timing/model': 1.1140539953485131, 'learning_rate/qwen_vl_interface': 3.618631978162259e-06, 'learning_rate/action_model': 3.618631978162259e-05, 'epoch': 1.78}) +[2026-06-17 23:26:58,569][starVLA.training.train_starvla][INFO] - Step 4321, Loss: {'train/loss': 0.4182432293891907, 'train/grad_norm_pre_clip': 16.97140121459961, 'timing/data': 0.0005420111119747162, 'timing/model': 0.9432600801810622, 'learning_rate/qwen_vl_interface': 3.6166010078731073e-06, 'learning_rate/action_model': 3.6166010078731075e-05, 'epoch': 1.78}) +[2026-06-17 23:27:02,353][starVLA.training.train_starvla][INFO] - Step 4322, Loss: {'train/loss': 0.8728468418121338, 'train/grad_norm_pre_clip': 27.020265579223633, 'timing/data': 0.0011496003717184067, 'timing/model': 0.94118657335639, 'learning_rate/qwen_vl_interface': 3.614570376189323e-06, 'learning_rate/action_model': 3.614570376189323e-05, 'epoch': 1.78}) +[2026-06-17 23:27:06,283][starVLA.training.train_starvla][INFO] - Step 4323, Loss: {'train/loss': 0.23215164244174957, 'train/grad_norm_pre_clip': 14.317582130432129, 'timing/data': 0.0009367987513542175, 'timing/model': 0.9549738429486752, 'learning_rate/qwen_vl_interface': 3.6125400835318604e-06, 'learning_rate/action_model': 3.61254008353186e-05, 'epoch': 1.78}) +[2026-06-17 23:27:10,167][starVLA.training.train_starvla][INFO] - Step 4324, Loss: {'train/loss': 0.4823116660118103, 'train/grad_norm_pre_clip': 20.241853713989258, 'timing/data': 0.0005327388644218445, 'timing/model': 0.9322521276772022, 'learning_rate/qwen_vl_interface': 3.6105101303216016e-06, 'learning_rate/action_model': 3.610510130321601e-05, 'epoch': 1.78}) +[2026-06-17 23:27:14,202][starVLA.training.train_starvla][INFO] - Step 4325, Loss: {'train/loss': 0.172681525349617, 'train/grad_norm_pre_clip': 12.492897987365723, 'timing/data': 0.0009256303310394287, 'timing/model': 0.9319539647549391, 'learning_rate/qwen_vl_interface': 3.608480516979354e-06, 'learning_rate/action_model': 3.608480516979354e-05, 'epoch': 1.78}) +[2026-06-17 23:27:18,293][starVLA.training.train_starvla][INFO] - Step 4326, Loss: {'train/loss': 0.5073859691619873, 'train/grad_norm_pre_clip': 23.340057373046875, 'timing/data': 0.0010336004197597504, 'timing/model': 0.9541672682389617, 'learning_rate/qwen_vl_interface': 3.6064512439258646e-06, 'learning_rate/action_model': 3.606451243925865e-05, 'epoch': 1.78}) +[2026-06-17 23:27:22,134][starVLA.training.train_starvla][INFO] - Step 4327, Loss: {'train/loss': 0.531895101070404, 'train/grad_norm_pre_clip': 20.341890335083008, 'timing/data': 0.0005038995295763016, 'timing/model': 1.175543456338346, 'learning_rate/qwen_vl_interface': 3.6044223115818e-06, 'learning_rate/action_model': 3.6044223115818e-05, 'epoch': 1.78}) +[2026-06-17 23:27:26,262][starVLA.training.train_starvla][INFO] - Step 4328, Loss: {'train/loss': 0.6679656505584717, 'train/grad_norm_pre_clip': 27.161901473999023, 'timing/data': 0.0013466328382492065, 'timing/model': 1.096724265255034, 'learning_rate/qwen_vl_interface': 3.6023937203677606e-06, 'learning_rate/action_model': 3.6023937203677604e-05, 'epoch': 1.78}) +[2026-06-17 23:27:33,279][starVLA.training.train_starvla][INFO] - Step 4329, Loss: {'train/loss': 0.5261753797531128, 'train/grad_norm_pre_clip': 27.895511627197266, 'timing/data': 0.0005630366504192352, 'timing/model': 0.9437681278213859, 'learning_rate/qwen_vl_interface': 3.6003654707042746e-06, 'learning_rate/action_model': 3.6003654707042745e-05, 'epoch': 1.78}) +[2026-06-17 23:27:40,750][starVLA.training.train_starvla][INFO] - Step 4330, Loss: {'train/loss': 0.08257634192705154, 'train/grad_norm_pre_clip': 6.580824375152588, 'timing/data': 0.001334846019744873, 'timing/model': 1.128790246322751, 'learning_rate/qwen_vl_interface': 3.5983375630118033e-06, 'learning_rate/action_model': 3.598337563011803e-05, 'epoch': 1.78}) +[2026-06-17 23:27:44,613][starVLA.training.train_starvla][INFO] - Step 4331, Loss: {'train/loss': 0.21786090731620789, 'train/grad_norm_pre_clip': 14.269266128540039, 'timing/data': 0.0008346037939190865, 'timing/model': 0.9522686777636409, 'learning_rate/qwen_vl_interface': 3.596309997710733e-06, 'learning_rate/action_model': 3.596309997710733e-05, 'epoch': 1.78}) +[2026-06-17 23:27:48,439][starVLA.training.train_starvla][INFO] - Step 4332, Loss: {'train/loss': 0.3101132810115814, 'train/grad_norm_pre_clip': 18.622318267822266, 'timing/data': 0.0008550873026251793, 'timing/model': 0.9526954321190715, 'learning_rate/qwen_vl_interface': 3.5942827752213788e-06, 'learning_rate/action_model': 3.594282775221379e-05, 'epoch': 1.78}) +[2026-06-17 23:27:54,459][starVLA.training.train_starvla][INFO] - Step 4333, Loss: {'train/loss': 0.22970791161060333, 'train/grad_norm_pre_clip': 14.575244903564453, 'timing/data': 0.0005273995921015739, 'timing/model': 1.1588364532217383, 'learning_rate/qwen_vl_interface': 3.592255895963986e-06, 'learning_rate/action_model': 3.592255895963986e-05, 'epoch': 1.78}) +[2026-06-17 23:27:58,455][starVLA.training.train_starvla][INFO] - Step 4334, Loss: {'train/loss': 0.6402289271354675, 'train/grad_norm_pre_clip': 39.712799072265625, 'timing/data': 0.0009482074528932571, 'timing/model': 1.1238363021984696, 'learning_rate/qwen_vl_interface': 3.590229360358731e-06, 'learning_rate/action_model': 3.590229360358731e-05, 'epoch': 1.78}) +[2026-06-17 23:28:06,516][starVLA.training.train_starvla][INFO] - Step 4335, Loss: {'train/loss': 0.29048019647598267, 'train/grad_norm_pre_clip': 40.39910888671875, 'timing/data': 0.000492406077682972, 'timing/model': 2.3518991274759173, 'learning_rate/qwen_vl_interface': 3.5882031688257147e-06, 'learning_rate/action_model': 3.588203168825715e-05, 'epoch': 1.78}) +[2026-06-17 23:28:10,569][starVLA.training.train_starvla][INFO] - Step 4336, Loss: {'train/loss': 0.3881337642669678, 'train/grad_norm_pre_clip': 21.73329734802246, 'timing/data': 0.0006545828655362129, 'timing/model': 0.9969326928257942, 'learning_rate/qwen_vl_interface': 3.5861773217849694e-06, 'learning_rate/action_model': 3.5861773217849693e-05, 'epoch': 1.78}) +[2026-06-17 23:28:14,474][starVLA.training.train_starvla][INFO] - Step 4337, Loss: {'train/loss': 0.4109930098056793, 'train/grad_norm_pre_clip': 18.073835372924805, 'timing/data': 0.0006003836169838905, 'timing/model': 0.9422157555818558, 'learning_rate/qwen_vl_interface': 3.5841518196564566e-06, 'learning_rate/action_model': 3.584151819656457e-05, 'epoch': 1.79}) +[2026-06-17 23:28:18,626][starVLA.training.train_starvla][INFO] - Step 4338, Loss: {'train/loss': 0.6776083707809448, 'train/grad_norm_pre_clip': 23.86941909790039, 'timing/data': 0.001531444489955902, 'timing/model': 1.2387572983279824, 'learning_rate/qwen_vl_interface': 3.582126662860065e-06, 'learning_rate/action_model': 3.582126662860065e-05, 'epoch': 1.79}) +[2026-06-17 23:28:22,453][starVLA.training.train_starvla][INFO] - Step 4339, Loss: {'train/loss': 0.625874936580658, 'train/grad_norm_pre_clip': 23.422815322875977, 'timing/data': 0.0013416679576039314, 'timing/model': 0.9491468705236912, 'learning_rate/qwen_vl_interface': 3.5801018518156096e-06, 'learning_rate/action_model': 3.5801018518156096e-05, 'epoch': 1.79}) +[2026-06-17 23:28:26,320][starVLA.training.train_starvla][INFO] - Step 4340, Loss: {'train/loss': 0.25477123260498047, 'train/grad_norm_pre_clip': 14.087421417236328, 'timing/data': 0.0004953574389219284, 'timing/model': 1.136851011775434, 'learning_rate/qwen_vl_interface': 3.5780773869428366e-06, 'learning_rate/action_model': 3.5780773869428364e-05, 'epoch': 1.79}) +[2026-06-17 23:28:30,183][starVLA.training.train_starvla][INFO] - Step 4341, Loss: {'train/loss': 0.5767898559570312, 'train/grad_norm_pre_clip': 25.826465606689453, 'timing/data': 0.0005168160423636436, 'timing/model': 1.2000702191144228, 'learning_rate/qwen_vl_interface': 3.576053268661422e-06, 'learning_rate/action_model': 3.5760532686614216e-05, 'epoch': 1.79}) +[2026-06-17 23:28:35,891][starVLA.training.train_starvla][INFO] - Step 4342, Loss: {'train/loss': 0.3978913426399231, 'train/grad_norm_pre_clip': 16.592580795288086, 'timing/data': 0.0005105920135974884, 'timing/model': 2.2260144790634513, 'learning_rate/qwen_vl_interface': 3.5740294973909643e-06, 'learning_rate/action_model': 3.5740294973909644e-05, 'epoch': 1.79}) +[2026-06-17 23:28:39,825][starVLA.training.train_starvla][INFO] - Step 4343, Loss: {'train/loss': 0.3180164098739624, 'train/grad_norm_pre_clip': 20.215715408325195, 'timing/data': 0.00083893071860075, 'timing/model': 1.038843996822834, 'learning_rate/qwen_vl_interface': 3.5720060735509976e-06, 'learning_rate/action_model': 3.5720060735509975e-05, 'epoch': 1.79}) +[2026-06-17 23:28:43,774][starVLA.training.train_starvla][INFO] - Step 4344, Loss: {'train/loss': 0.7130606174468994, 'train/grad_norm_pre_clip': 22.376951217651367, 'timing/data': 0.0007174713537096977, 'timing/model': 1.021012301556766, 'learning_rate/qwen_vl_interface': 3.569982997560974e-06, 'learning_rate/action_model': 3.5699829975609736e-05, 'epoch': 1.79}) +[2026-06-17 23:28:47,538][starVLA.training.train_starvla][INFO] - Step 4345, Loss: {'train/loss': 0.1721280813217163, 'train/grad_norm_pre_clip': 13.038979530334473, 'timing/data': 0.0013488419353961945, 'timing/model': 0.9408355969935656, 'learning_rate/qwen_vl_interface': 3.567960269840286e-06, 'learning_rate/action_model': 3.567960269840286e-05, 'epoch': 1.79}) +[2026-06-17 23:28:51,664][starVLA.training.train_starvla][INFO] - Step 4346, Loss: {'train/loss': 0.329181045293808, 'train/grad_norm_pre_clip': 37.84311294555664, 'timing/data': 0.000550001859664917, 'timing/model': 0.9353712797164917, 'learning_rate/qwen_vl_interface': 3.5659378908082427e-06, 'learning_rate/action_model': 3.5659378908082426e-05, 'epoch': 1.79}) +[2026-06-17 23:28:56,170][starVLA.training.train_starvla][INFO] - Step 4347, Loss: {'train/loss': 0.35318493843078613, 'train/grad_norm_pre_clip': 16.799854278564453, 'timing/data': 0.000501999631524086, 'timing/model': 1.1385429138317704, 'learning_rate/qwen_vl_interface': 3.5639158608840862e-06, 'learning_rate/action_model': 3.563915860884086e-05, 'epoch': 1.79}) +[2026-06-17 23:28:59,872][starVLA.training.train_starvla][INFO] - Step 4348, Loss: {'train/loss': 0.11208892613649368, 'train/grad_norm_pre_clip': 11.22740364074707, 'timing/data': 0.0008952682837843895, 'timing/model': 0.9399598818272352, 'learning_rate/qwen_vl_interface': 3.5618941804869866e-06, 'learning_rate/action_model': 3.561894180486987e-05, 'epoch': 1.79}) +[2026-06-17 23:29:03,739][starVLA.training.train_starvla][INFO] - Step 4349, Loss: {'train/loss': 0.24267946183681488, 'train/grad_norm_pre_clip': 12.611400604248047, 'timing/data': 0.0008390862494707108, 'timing/model': 0.9486068356782198, 'learning_rate/qwen_vl_interface': 3.5598728500360395e-06, 'learning_rate/action_model': 3.55987285003604e-05, 'epoch': 1.79}) +[2026-06-17 23:29:07,541][starVLA.training.train_starvla][INFO] - Step 4350, Loss: {'train/loss': 0.14153723418712616, 'train/grad_norm_pre_clip': 15.328413963317871, 'timing/data': 0.0008048005402088165, 'timing/model': 0.9438397269695997, 'learning_rate/qwen_vl_interface': 3.557851869950272e-06, 'learning_rate/action_model': 3.557851869950272e-05, 'epoch': 1.79}) +[2026-06-17 23:29:11,440][starVLA.training.train_starvla][INFO] - Step 4351, Loss: {'train/loss': 0.1481829434633255, 'train/grad_norm_pre_clip': 12.230339050292969, 'timing/data': 0.0006009787321090698, 'timing/model': 0.9686647662892938, 'learning_rate/qwen_vl_interface': 3.555831240648628e-06, 'learning_rate/action_model': 3.5558312406486274e-05, 'epoch': 1.79}) +[2026-06-17 23:29:15,410][starVLA.training.train_starvla][INFO] - Step 4352, Loss: {'train/loss': 0.2840212881565094, 'train/grad_norm_pre_clip': 12.784622192382812, 'timing/data': 0.0005774442106485367, 'timing/model': 0.9559930060058832, 'learning_rate/qwen_vl_interface': 3.553810962549994e-06, 'learning_rate/action_model': 3.553810962549994e-05, 'epoch': 1.79}) +[2026-06-17 23:29:19,389][starVLA.training.train_starvla][INFO] - Step 4353, Loss: {'train/loss': 0.34159818291664124, 'train/grad_norm_pre_clip': 18.44951629638672, 'timing/data': 0.0008217254653573036, 'timing/model': 1.2095970017835498, 'learning_rate/qwen_vl_interface': 3.5517910360731723e-06, 'learning_rate/action_model': 3.551791036073172e-05, 'epoch': 1.79}) +[2026-06-17 23:29:23,246][starVLA.training.train_starvla][INFO] - Step 4354, Loss: {'train/loss': 0.4134491980075836, 'train/grad_norm_pre_clip': 18.50632095336914, 'timing/data': 0.0011982312425971031, 'timing/model': 1.1005900632590055, 'learning_rate/qwen_vl_interface': 3.549771461636896e-06, 'learning_rate/action_model': 3.549771461636896e-05, 'epoch': 1.79}) +[2026-06-17 23:29:27,056][starVLA.training.train_starvla][INFO] - Step 4355, Loss: {'train/loss': 0.2130746841430664, 'train/grad_norm_pre_clip': 13.124113082885742, 'timing/data': 0.00047708023339509964, 'timing/model': 0.9277947265654802, 'learning_rate/qwen_vl_interface': 3.5477522396598236e-06, 'learning_rate/action_model': 3.5477522396598235e-05, 'epoch': 1.79}) +[2026-06-17 23:29:30,908][starVLA.training.train_starvla][INFO] - Step 4356, Loss: {'train/loss': 0.35243040323257446, 'train/grad_norm_pre_clip': 15.132879257202148, 'timing/data': 0.0004943273961544037, 'timing/model': 0.9399875141680241, 'learning_rate/qwen_vl_interface': 3.5457333705605456e-06, 'learning_rate/action_model': 3.5457333705605456e-05, 'epoch': 1.79}) +[2026-06-17 23:29:34,795][starVLA.training.train_starvla][INFO] - Step 4357, Loss: {'train/loss': 0.2678810954093933, 'train/grad_norm_pre_clip': 19.76343536376953, 'timing/data': 0.0005137789994478226, 'timing/model': 1.0214804979041219, 'learning_rate/qwen_vl_interface': 3.543714854757574e-06, 'learning_rate/action_model': 3.5437148547575736e-05, 'epoch': 1.79}) +[2026-06-17 23:29:38,749][starVLA.training.train_starvla][INFO] - Step 4358, Loss: {'train/loss': 0.1723058819770813, 'train/grad_norm_pre_clip': 13.876762390136719, 'timing/data': 0.0006571663543581963, 'timing/model': 1.00837710686028, 'learning_rate/qwen_vl_interface': 3.5416966926693464e-06, 'learning_rate/action_model': 3.5416966926693465e-05, 'epoch': 1.79}) +[2026-06-17 23:29:42,644][starVLA.training.train_starvla][INFO] - Step 4359, Loss: {'train/loss': 0.4905996024608612, 'train/grad_norm_pre_clip': 22.441078186035156, 'timing/data': 0.0005009518936276436, 'timing/model': 0.9682759642601013, 'learning_rate/qwen_vl_interface': 3.539678884714235e-06, 'learning_rate/action_model': 3.539678884714235e-05, 'epoch': 1.79}) +[2026-06-17 23:29:47,479][starVLA.training.train_starvla][INFO] - Step 4360, Loss: {'train/loss': 0.25458064675331116, 'train/grad_norm_pre_clip': 21.47871208190918, 'timing/data': 0.0006650770083069801, 'timing/model': 1.1350130140781403, 'learning_rate/qwen_vl_interface': 3.5376614313105307e-06, 'learning_rate/action_model': 3.53766143131053e-05, 'epoch': 1.79}) +[2026-06-17 23:29:51,301][starVLA.training.train_starvla][INFO] - Step 4361, Loss: {'train/loss': 0.4051339030265808, 'train/grad_norm_pre_clip': 20.132905960083008, 'timing/data': 0.0008452450856566429, 'timing/model': 1.105907803401351, 'learning_rate/qwen_vl_interface': 3.5356443328764546e-06, 'learning_rate/action_model': 3.5356443328764546e-05, 'epoch': 1.8}) +[2026-06-17 23:29:54,941][starVLA.training.train_starvla][INFO] - Step 4362, Loss: {'train/loss': 0.25739508867263794, 'train/grad_norm_pre_clip': 24.51715850830078, 'timing/data': 0.0009079352021217346, 'timing/model': 0.9410609193146229, 'learning_rate/qwen_vl_interface': 3.5336275898301515e-06, 'learning_rate/action_model': 3.5336275898301514e-05, 'epoch': 1.8}) +[2026-06-17 23:29:58,838][starVLA.training.train_starvla][INFO] - Step 4363, Loss: {'train/loss': 0.4051344692707062, 'train/grad_norm_pre_clip': 22.439979553222656, 'timing/data': 0.0005192216485738754, 'timing/model': 0.9542853469029069, 'learning_rate/qwen_vl_interface': 3.531611202589697e-06, 'learning_rate/action_model': 3.531611202589697e-05, 'epoch': 1.8}) +[2026-06-17 23:30:02,726][starVLA.training.train_starvla][INFO] - Step 4364, Loss: {'train/loss': 0.37187451124191284, 'train/grad_norm_pre_clip': 26.51081085205078, 'timing/data': 0.00070924311876297, 'timing/model': 0.9621297689154744, 'learning_rate/qwen_vl_interface': 3.5295951715730902e-06, 'learning_rate/action_model': 3.52959517157309e-05, 'epoch': 1.8}) +[2026-06-17 23:30:06,608][starVLA.training.train_starvla][INFO] - Step 4365, Loss: {'train/loss': 0.30179163813591003, 'train/grad_norm_pre_clip': 13.781417846679688, 'timing/data': 0.0008453382179141045, 'timing/model': 0.9575430657714605, 'learning_rate/qwen_vl_interface': 3.527579497198255e-06, 'learning_rate/action_model': 3.527579497198255e-05, 'epoch': 1.8}) +[2026-06-17 23:30:10,465][starVLA.training.train_starvla][INFO] - Step 4366, Loss: {'train/loss': 0.11153702437877655, 'train/grad_norm_pre_clip': 26.135818481445312, 'timing/data': 0.0004800437018275261, 'timing/model': 0.937831474468112, 'learning_rate/qwen_vl_interface': 3.525564179883042e-06, 'learning_rate/action_model': 3.525564179883042e-05, 'epoch': 1.8}) +[2026-06-17 23:30:14,373][starVLA.training.train_starvla][INFO] - Step 4367, Loss: {'train/loss': 0.38771092891693115, 'train/grad_norm_pre_clip': 17.9008846282959, 'timing/data': 0.0007366985082626343, 'timing/model': 1.1582659529522061, 'learning_rate/qwen_vl_interface': 3.5235492200452308e-06, 'learning_rate/action_model': 3.523549220045231e-05, 'epoch': 1.8}) +[2026-06-17 23:30:18,132][starVLA.training.train_starvla][INFO] - Step 4368, Loss: {'train/loss': 0.33997607231140137, 'train/grad_norm_pre_clip': 24.91128921508789, 'timing/data': 0.0008198665454983711, 'timing/model': 1.0457409685477614, 'learning_rate/qwen_vl_interface': 3.521534618102525e-06, 'learning_rate/action_model': 3.5215346181025246e-05, 'epoch': 1.8}) +[2026-06-17 23:30:21,778][starVLA.training.train_starvla][INFO] - Step 4369, Loss: {'train/loss': 0.6953190565109253, 'train/grad_norm_pre_clip': 25.795734405517578, 'timing/data': 0.00048006046563386917, 'timing/model': 0.9495051512494683, 'learning_rate/qwen_vl_interface': 3.5195203744725516e-06, 'learning_rate/action_model': 3.5195203744725516e-05, 'epoch': 1.8}) +[2026-06-17 23:30:25,708][starVLA.training.train_starvla][INFO] - Step 4370, Loss: {'train/loss': 0.2552196681499481, 'train/grad_norm_pre_clip': 17.86458396911621, 'timing/data': 0.0005846796557307243, 'timing/model': 0.9694213755428791, 'learning_rate/qwen_vl_interface': 3.5175064895728665e-06, 'learning_rate/action_model': 3.5175064895728664e-05, 'epoch': 1.8}) +[2026-06-17 23:30:29,602][starVLA.training.train_starvla][INFO] - Step 4371, Loss: {'train/loss': 0.47849205136299133, 'train/grad_norm_pre_clip': 21.653060913085938, 'timing/data': 0.000501011498272419, 'timing/model': 0.956599660217762, 'learning_rate/qwen_vl_interface': 3.515492963820951e-06, 'learning_rate/action_model': 3.515492963820951e-05, 'epoch': 1.8}) +[2026-06-17 23:30:34,120][starVLA.training.train_starvla][INFO] - Step 4372, Loss: {'train/loss': 0.6622449159622192, 'train/grad_norm_pre_clip': 27.9461612701416, 'timing/data': 0.0005258237943053246, 'timing/model': 0.9603408575057983, 'learning_rate/qwen_vl_interface': 3.5134797976342123e-06, 'learning_rate/action_model': 3.513479797634212e-05, 'epoch': 1.8}) +[2026-06-17 23:30:38,045][starVLA.training.train_starvla][INFO] - Step 4373, Loss: {'train/loss': 0.19746962189674377, 'train/grad_norm_pre_clip': 12.164856910705566, 'timing/data': 0.0006852447986602783, 'timing/model': 1.1850052792578936, 'learning_rate/qwen_vl_interface': 3.5114669914299776e-06, 'learning_rate/action_model': 3.511466991429978e-05, 'epoch': 1.8}) +[2026-06-17 23:30:41,896][starVLA.training.train_starvla][INFO] - Step 4374, Loss: {'train/loss': 0.30659908056259155, 'train/grad_norm_pre_clip': 20.662784576416016, 'timing/data': 0.0004661167040467262, 'timing/model': 1.0508354343473911, 'learning_rate/qwen_vl_interface': 3.5094545456255065e-06, 'learning_rate/action_model': 3.509454545625506e-05, 'epoch': 1.8}) +[2026-06-17 23:30:45,533][starVLA.training.train_starvla][INFO] - Step 4375, Loss: {'train/loss': 0.3467208743095398, 'train/grad_norm_pre_clip': 15.691123962402344, 'timing/data': 0.0004887748509645462, 'timing/model': 0.9504295671358705, 'learning_rate/qwen_vl_interface': 3.5074424606379826e-06, 'learning_rate/action_model': 3.5074424606379824e-05, 'epoch': 1.8}) +[2026-06-17 23:30:49,336][starVLA.training.train_starvla][INFO] - Step 4376, Loss: {'train/loss': 0.32017505168914795, 'train/grad_norm_pre_clip': 18.733980178833008, 'timing/data': 0.0010148398578166962, 'timing/model': 0.9451820310205221, 'learning_rate/qwen_vl_interface': 3.505430736884511e-06, 'learning_rate/action_model': 3.5054307368845105e-05, 'epoch': 1.8}) +[2026-06-17 23:30:53,424][starVLA.training.train_starvla][INFO] - Step 4377, Loss: {'train/loss': 0.26655784249305725, 'train/grad_norm_pre_clip': 13.227075576782227, 'timing/data': 0.0004772236570715904, 'timing/model': 0.9651084439828992, 'learning_rate/qwen_vl_interface': 3.503419374782123e-06, 'learning_rate/action_model': 3.503419374782123e-05, 'epoch': 1.8}) +[2026-06-17 23:30:57,300][starVLA.training.train_starvla][INFO] - Step 4378, Loss: {'train/loss': 0.6848145723342896, 'train/grad_norm_pre_clip': 25.52642822265625, 'timing/data': 0.000516500324010849, 'timing/model': 0.9543457860127091, 'learning_rate/qwen_vl_interface': 3.50140837474778e-06, 'learning_rate/action_model': 3.50140837474778e-05, 'epoch': 1.8}) +[2026-06-17 23:31:01,208][starVLA.training.train_starvla][INFO] - Step 4379, Loss: {'train/loss': 0.388477623462677, 'train/grad_norm_pre_clip': 19.579233169555664, 'timing/data': 0.00044376589357852936, 'timing/model': 0.9478608630597591, 'learning_rate/qwen_vl_interface': 3.4993977371983625e-06, 'learning_rate/action_model': 3.499397737198362e-05, 'epoch': 1.8}) +[2026-06-17 23:31:05,105][starVLA.training.train_starvla][INFO] - Step 4380, Loss: {'train/loss': 0.19005882740020752, 'train/grad_norm_pre_clip': 12.040449142456055, 'timing/data': 0.0011534644290804863, 'timing/model': 1.138435736298561, 'learning_rate/qwen_vl_interface': 3.497387462550676e-06, 'learning_rate/action_model': 3.4973874625506757e-05, 'epoch': 1.8}) +[2026-06-17 23:31:08,802][starVLA.training.train_starvla][INFO] - Step 4381, Loss: {'train/loss': 0.7200816869735718, 'train/grad_norm_pre_clip': 25.760562896728516, 'timing/data': 0.0004697013646364212, 'timing/model': 0.9510429427027702, 'learning_rate/qwen_vl_interface': 3.4953775512214568e-06, 'learning_rate/action_model': 3.495377551221457e-05, 'epoch': 1.8}) +[2026-06-17 23:31:12,686][starVLA.training.train_starvla][INFO] - Step 4382, Loss: {'train/loss': 0.2931292653083801, 'train/grad_norm_pre_clip': 17.708616256713867, 'timing/data': 0.0007500164210796356, 'timing/model': 0.9509305572137237, 'learning_rate/qwen_vl_interface': 3.4933680036273583e-06, 'learning_rate/action_model': 3.493368003627358e-05, 'epoch': 1.8}) +[2026-06-17 23:31:16,533][starVLA.training.train_starvla][INFO] - Step 4383, Loss: {'train/loss': 0.30063506960868835, 'train/grad_norm_pre_clip': 15.949471473693848, 'timing/data': 0.0009155673906207085, 'timing/model': 0.9404249060899019, 'learning_rate/qwen_vl_interface': 3.491358820184963e-06, 'learning_rate/action_model': 3.491358820184963e-05, 'epoch': 1.8}) +[2026-06-17 23:31:20,376][starVLA.training.train_starvla][INFO] - Step 4384, Loss: {'train/loss': 0.28796613216400146, 'train/grad_norm_pre_clip': 16.569528579711914, 'timing/data': 0.000545046292245388, 'timing/model': 0.9384562158957124, 'learning_rate/qwen_vl_interface': 3.4893500013107758e-06, 'learning_rate/action_model': 3.4893500013107754e-05, 'epoch': 1.8}) +[2026-06-17 23:31:24,206][starVLA.training.train_starvla][INFO] - Step 4385, Loss: {'train/loss': 0.2285158634185791, 'train/grad_norm_pre_clip': 18.30234146118164, 'timing/data': 0.00048107001930475235, 'timing/model': 0.9550461322069168, 'learning_rate/qwen_vl_interface': 3.4873415474212287e-06, 'learning_rate/action_model': 3.487341547421228e-05, 'epoch': 1.81}) +[2026-06-17 23:31:28,091][starVLA.training.train_starvla][INFO] - Step 4386, Loss: {'train/loss': 0.4257031977176666, 'train/grad_norm_pre_clip': 21.35723876953125, 'timing/data': 0.0005051987245678902, 'timing/model': 1.1519040688872337, 'learning_rate/qwen_vl_interface': 3.4853334589326764e-06, 'learning_rate/action_model': 3.4853334589326765e-05, 'epoch': 1.81}) +[2026-06-17 23:31:31,728][starVLA.training.train_starvla][INFO] - Step 4387, Loss: {'train/loss': 0.443407267332077, 'train/grad_norm_pre_clip': 25.85114097595215, 'timing/data': 0.0006156163290143013, 'timing/model': 0.9396669445559382, 'learning_rate/qwen_vl_interface': 3.483325736261394e-06, 'learning_rate/action_model': 3.483325736261394e-05, 'epoch': 1.81}) +[2026-06-17 23:31:35,704][starVLA.training.train_starvla][INFO] - Step 4388, Loss: {'train/loss': 0.44757384061813354, 'train/grad_norm_pre_clip': 20.197799682617188, 'timing/data': 0.00046962033957242966, 'timing/model': 1.0368164898827672, 'learning_rate/qwen_vl_interface': 3.48131837982359e-06, 'learning_rate/action_model': 3.48131837982359e-05, 'epoch': 1.81}) +[2026-06-17 23:31:39,530][starVLA.training.train_starvla][INFO] - Step 4389, Loss: {'train/loss': 0.30098310112953186, 'train/grad_norm_pre_clip': 26.17951202392578, 'timing/data': 0.0005134344100952148, 'timing/model': 0.9515293342992663, 'learning_rate/qwen_vl_interface': 3.4793113900353885e-06, 'learning_rate/action_model': 3.479311390035389e-05, 'epoch': 1.81}) +[2026-06-17 23:31:43,351][starVLA.training.train_starvla][INFO] - Step 4390, Loss: {'train/loss': 0.3779342770576477, 'train/grad_norm_pre_clip': 18.86044692993164, 'timing/data': 0.0008254013955593109, 'timing/model': 0.9587868824601173, 'learning_rate/qwen_vl_interface': 3.4773047673128412e-06, 'learning_rate/action_model': 3.477304767312841e-05, 'epoch': 1.81}) +[2026-06-17 23:31:47,418][starVLA.training.train_starvla][INFO] - Step 4391, Loss: {'train/loss': 0.8507333993911743, 'train/grad_norm_pre_clip': 38.39686584472656, 'timing/data': 0.0007332265377044678, 'timing/model': 0.9558940324932337, 'learning_rate/qwen_vl_interface': 3.4752985120719225e-06, 'learning_rate/action_model': 3.475298512071922e-05, 'epoch': 1.81}) +[2026-06-17 23:31:51,233][starVLA.training.train_starvla][INFO] - Step 4392, Loss: {'train/loss': 0.17706337571144104, 'train/grad_norm_pre_clip': 15.46731185913086, 'timing/data': 0.0008225645869970322, 'timing/model': 0.9545342279598117, 'learning_rate/qwen_vl_interface': 3.4732926247285324e-06, 'learning_rate/action_model': 3.4732926247285325e-05, 'epoch': 1.81}) +[2026-06-17 23:31:55,050][starVLA.training.train_starvla][INFO] - Step 4393, Loss: {'train/loss': 0.4130719006061554, 'train/grad_norm_pre_clip': 26.778175354003906, 'timing/data': 0.0007318072021007538, 'timing/model': 1.1093551013618708, 'learning_rate/qwen_vl_interface': 3.471287105698494e-06, 'learning_rate/action_model': 3.471287105698494e-05, 'epoch': 1.81}) +[2026-06-17 23:31:58,680][starVLA.training.train_starvla][INFO] - Step 4394, Loss: {'train/loss': 0.49431079626083374, 'train/grad_norm_pre_clip': 19.376781463623047, 'timing/data': 0.0010409019887447357, 'timing/model': 0.9433445381000638, 'learning_rate/qwen_vl_interface': 3.4692819553975514e-06, 'learning_rate/action_model': 3.469281955397551e-05, 'epoch': 1.81}) +[2026-06-17 23:32:02,578][starVLA.training.train_starvla][INFO] - Step 4395, Loss: {'train/loss': 0.724984347820282, 'train/grad_norm_pre_clip': 23.77142333984375, 'timing/data': 0.0004943106323480606, 'timing/model': 0.9471718911081553, 'learning_rate/qwen_vl_interface': 3.4672771742413754e-06, 'learning_rate/action_model': 3.467277174241375e-05, 'epoch': 1.81}) +[2026-06-17 23:32:06,410][starVLA.training.train_starvla][INFO] - Step 4396, Loss: {'train/loss': 0.3742663860321045, 'train/grad_norm_pre_clip': 22.560260772705078, 'timing/data': 0.0005279853940010071, 'timing/model': 0.9663730217143893, 'learning_rate/qwen_vl_interface': 3.46527276264556e-06, 'learning_rate/action_model': 3.46527276264556e-05, 'epoch': 1.81}) +[2026-06-17 23:32:10,301][starVLA.training.train_starvla][INFO] - Step 4397, Loss: {'train/loss': 0.30207863450050354, 'train/grad_norm_pre_clip': 16.499984741210938, 'timing/data': 0.00048803724348545074, 'timing/model': 0.9576041614636779, 'learning_rate/qwen_vl_interface': 3.4632687210256218e-06, 'learning_rate/action_model': 3.463268721025621e-05, 'epoch': 1.81}) +[2026-06-17 23:32:14,159][starVLA.training.train_starvla][INFO] - Step 4398, Loss: {'train/loss': 0.1711818128824234, 'train/grad_norm_pre_clip': 11.909530639648438, 'timing/data': 0.00045667029917240143, 'timing/model': 0.9412343874573708, 'learning_rate/qwen_vl_interface': 3.461265049796999e-06, 'learning_rate/action_model': 3.461265049796999e-05, 'epoch': 1.81}) +[2026-06-17 23:32:18,058][starVLA.training.train_starvla][INFO] - Step 4399, Loss: {'train/loss': 0.11964979022741318, 'train/grad_norm_pre_clip': 12.980260848999023, 'timing/data': 0.0005226554349064827, 'timing/model': 1.124297964386642, 'learning_rate/qwen_vl_interface': 3.459261749375057e-06, 'learning_rate/action_model': 3.4592617493750567e-05, 'epoch': 1.81}) +[2026-06-17 23:34:33,470][starVLA.training.train_starvla][INFO] - Step 4400, Loss: {'train/loss': 0.3513103723526001, 'train/grad_norm_pre_clip': 18.779375076293945, 'eval/loss': 0.44205763936042786, 'eval/latency_6/loss': 0.44205763936042786, 'eval/demon_fire_precision': 0.9498275321417372, 'eval/demon_fire_recall': 0.9789915966386554, 'eval/demon_fire_f1': 0.9641890816488938, 'eval/demon_fire_rate_model': 0.8873121869782972, 'eval/demon_fire_rate_teacher': 0.860879243183083, 'eval/demon_left_precision': 0.9017199017199017, 'eval/demon_left_recall': 0.643859649122807, 'eval/demon_left_f1': 0.7512794268167861, 'eval/demon_left_rate_model': 0.1132442960489705, 'eval/demon_left_rate_teacher': 0.15859766277128548, 'eval/demon_right_precision': 0.7949921752738655, 'eval/demon_right_recall': 0.7803379416282642, 'eval/demon_right_f1': 0.7875968992248062, 'eval/demon_right_rate_model': 0.17779632721202004, 'eval/demon_right_rate_teacher': 0.18113522537562604, 'eval/demon_move_f1': 0.7694381630207962, 'eval/demon_cc_f1': 0.8961428192009131, 'eval/latency_6/demon_fire_precision': 0.9498275321417372, 'eval/latency_6/demon_fire_recall': 0.9789915966386554, 'eval/latency_6/demon_fire_f1': 0.9641890816488938, 'eval/latency_6/demon_fire_rate_model': 0.8873121869782972, 'eval/latency_6/demon_fire_rate_teacher': 0.860879243183083, 'eval/latency_6/demon_left_precision': 0.9017199017199017, 'eval/latency_6/demon_left_recall': 0.643859649122807, 'eval/latency_6/demon_left_f1': 0.7512794268167861, 'eval/latency_6/demon_left_rate_model': 0.1132442960489705, 'eval/latency_6/demon_left_rate_teacher': 0.15859766277128548, 'eval/latency_6/demon_right_precision': 0.7949921752738655, 'eval/latency_6/demon_right_recall': 0.7803379416282642, 'eval/latency_6/demon_right_f1': 0.7875968992248062, 'eval/latency_6/demon_right_rate_model': 0.17779632721202004, 'eval/latency_6/demon_right_rate_teacher': 0.18113522537562604, 'eval/latency_6/demon_move_f1': 0.7694381630207962, 'eval/latency_6/demon_cc_f1': 0.8961428192009131, 'timing/data': 0.0005523180589079857, 'timing/model': 1.2082346994429827, 'learning_rate/qwen_vl_interface': 3.457258820175082e-06, 'learning_rate/action_model': 3.457258820175082e-05, 'epoch': 1.81}) +[2026-06-17 23:34:33,472][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_4400_state +[2026-06-17 23:34:33,477][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 23:34:40,427][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_4400_state/model.safetensors +[2026-06-17 23:34:56,339][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_4400_state/optimizer.bin +[2026-06-17 23:34:56,340][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_4400_state/sampler.bin +[2026-06-17 23:34:56,340][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_4400_state/sampler_1.bin +[2026-06-17 23:34:56,342][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_4400_state/random_states_0.pkl +[2026-06-17 23:34:56,345][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4400 +[2026-06-17 23:34:56,346][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 23:34:56,356][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 23:35:00,260][starVLA.training.train_starvla][INFO] - Step 4401, Loss: {'train/loss': 0.577306866645813, 'train/grad_norm_pre_clip': 23.155109405517578, 'timing/data': 0.0009257793426513672, 'timing/model': 1.1720350813120604, 'learning_rate/qwen_vl_interface': 3.4552562626122817e-06, 'learning_rate/action_model': 3.455256262612281e-05, 'epoch': 1.81}) +[2026-06-17 23:35:04,156][starVLA.training.train_starvla][INFO] - Step 4402, Loss: {'train/loss': 0.3162050247192383, 'train/grad_norm_pre_clip': 19.989173889160156, 'timing/data': 0.0004996601492166519, 'timing/model': 1.1177167408168316, 'learning_rate/qwen_vl_interface': 3.453254077101789e-06, 'learning_rate/action_model': 3.453254077101789e-05, 'epoch': 1.81}) +[2026-06-17 23:35:07,972][starVLA.training.train_starvla][INFO] - Step 4403, Loss: {'train/loss': 0.5112490057945251, 'train/grad_norm_pre_clip': 23.93989372253418, 'timing/data': 0.000528654083609581, 'timing/model': 1.044320248067379, 'learning_rate/qwen_vl_interface': 3.45125226405866e-06, 'learning_rate/action_model': 3.4512522640586595e-05, 'epoch': 1.81}) +[2026-06-17 23:35:11,829][starVLA.training.train_starvla][INFO] - Step 4404, Loss: {'train/loss': 0.25330850481987, 'train/grad_norm_pre_clip': 27.328962326049805, 'timing/data': 0.0004925904795527458, 'timing/model': 0.9582155924290419, 'learning_rate/qwen_vl_interface': 3.449250823897872e-06, 'learning_rate/action_model': 3.449250823897872e-05, 'epoch': 1.81}) +[2026-06-17 23:35:15,795][starVLA.training.train_starvla][INFO] - Step 4405, Loss: {'train/loss': 0.3306490182876587, 'train/grad_norm_pre_clip': 19.128459930419922, 'timing/data': 0.0005525648593902588, 'timing/model': 0.9667807258665562, 'learning_rate/qwen_vl_interface': 3.447249757034328e-06, 'learning_rate/action_model': 3.4472497570343276e-05, 'epoch': 1.81}) +[2026-06-17 23:35:19,555][starVLA.training.train_starvla][INFO] - Step 4406, Loss: {'train/loss': 0.5765783190727234, 'train/grad_norm_pre_clip': 42.35252380371094, 'timing/data': 0.0005696667358279228, 'timing/model': 0.9419300761073828, 'learning_rate/qwen_vl_interface': 3.445249063882845e-06, 'learning_rate/action_model': 3.445249063882845e-05, 'epoch': 1.81}) +[2026-06-17 23:35:23,384][starVLA.training.train_starvla][INFO] - Step 4407, Loss: {'train/loss': 0.21933552622795105, 'train/grad_norm_pre_clip': 15.697938919067383, 'timing/data': 0.0005110017955303192, 'timing/model': 0.9457633327692747, 'learning_rate/qwen_vl_interface': 3.4432487448581762e-06, 'learning_rate/action_model': 3.443248744858176e-05, 'epoch': 1.81}) +[2026-06-17 23:35:27,224][starVLA.training.train_starvla][INFO] - Step 4408, Loss: {'train/loss': 0.28285840153694153, 'train/grad_norm_pre_clip': 24.62372589111328, 'timing/data': 0.000511443242430687, 'timing/model': 1.149747047573328, 'learning_rate/qwen_vl_interface': 3.4412488003749867e-06, 'learning_rate/action_model': 3.4412488003749866e-05, 'epoch': 1.81}) +[2026-06-17 23:35:30,859][starVLA.training.train_starvla][INFO] - Step 4409, Loss: {'train/loss': 0.32436665892601013, 'train/grad_norm_pre_clip': 20.415178298950195, 'timing/data': 0.0006188713014125824, 'timing/model': 0.9426868176087737, 'learning_rate/qwen_vl_interface': 3.4392492308478657e-06, 'learning_rate/action_model': 3.439249230847866e-05, 'epoch': 1.81}) +[2026-06-17 23:35:34,681][starVLA.training.train_starvla][INFO] - Step 4410, Loss: {'train/loss': 0.29766353964805603, 'train/grad_norm_pre_clip': 17.413883209228516, 'timing/data': 0.0005843527615070343, 'timing/model': 0.9357858411967754, 'learning_rate/qwen_vl_interface': 3.437250036691328e-06, 'learning_rate/action_model': 3.437250036691328e-05, 'epoch': 1.82}) +[2026-06-17 23:35:38,550][starVLA.training.train_starvla][INFO] - Step 4411, Loss: {'train/loss': 0.2321990728378296, 'train/grad_norm_pre_clip': 15.194572448730469, 'timing/data': 0.0005994848906993866, 'timing/model': 0.959565969184041, 'learning_rate/qwen_vl_interface': 3.4352512183198094e-06, 'learning_rate/action_model': 3.4352512183198095e-05, 'epoch': 1.82}) +[2026-06-17 23:35:42,437][starVLA.training.train_starvla][INFO] - Step 4412, Loss: {'train/loss': 0.4334507882595062, 'train/grad_norm_pre_clip': 20.761362075805664, 'timing/data': 0.0005051633343100548, 'timing/model': 0.9628497585654259, 'learning_rate/qwen_vl_interface': 3.4332527761476662e-06, 'learning_rate/action_model': 3.433252776147666e-05, 'epoch': 1.82}) +[2026-06-17 23:35:46,310][starVLA.training.train_starvla][INFO] - Step 4413, Loss: {'train/loss': 0.2154655009508133, 'train/grad_norm_pre_clip': 16.589988708496094, 'timing/data': 0.0005133179947733879, 'timing/model': 0.9512136299163103, 'learning_rate/qwen_vl_interface': 3.431254710589174e-06, 'learning_rate/action_model': 3.431254710589174e-05, 'epoch': 1.82}) +[2026-06-17 23:35:50,161][starVLA.training.train_starvla][INFO] - Step 4414, Loss: {'train/loss': 0.5442399978637695, 'train/grad_norm_pre_clip': 19.931934356689453, 'timing/data': 0.0006905198097229004, 'timing/model': 1.158097512088716, 'learning_rate/qwen_vl_interface': 3.4292570220585407e-06, 'learning_rate/action_model': 3.4292570220585407e-05, 'epoch': 1.82}) +[2026-06-17 23:35:53,866][starVLA.training.train_starvla][INFO] - Step 4415, Loss: {'train/loss': 0.9361594319343567, 'train/grad_norm_pre_clip': 32.3173713684082, 'timing/data': 0.0012132450938224792, 'timing/model': 0.9321839110925794, 'learning_rate/qwen_vl_interface': 3.4272597109698857e-06, 'learning_rate/action_model': 3.427259710969886e-05, 'epoch': 1.82}) +[2026-06-17 23:35:58,029][starVLA.training.train_starvla][INFO] - Step 4416, Loss: {'train/loss': 0.6626232266426086, 'train/grad_norm_pre_clip': 28.602336883544922, 'timing/data': 0.0007143942639231682, 'timing/model': 0.9354220377281308, 'learning_rate/qwen_vl_interface': 3.4252627777372533e-06, 'learning_rate/action_model': 3.425262777737253e-05, 'epoch': 1.82}) +[2026-06-17 23:36:01,837][starVLA.training.train_starvla][INFO] - Step 4417, Loss: {'train/loss': 0.3128770589828491, 'train/grad_norm_pre_clip': 27.494632720947266, 'timing/data': 0.0008940035477280617, 'timing/model': 0.9472290435805917, 'learning_rate/qwen_vl_interface': 3.4232662227746095e-06, 'learning_rate/action_model': 3.423266222774609e-05, 'epoch': 1.82}) +[2026-06-17 23:36:05,629][starVLA.training.train_starvla][INFO] - Step 4418, Loss: {'train/loss': 0.8147664666175842, 'train/grad_norm_pre_clip': 25.963150024414062, 'timing/data': 0.0005520889535546303, 'timing/model': 0.9453134713694453, 'learning_rate/qwen_vl_interface': 3.4212700464958448e-06, 'learning_rate/action_model': 3.421270046495845e-05, 'epoch': 1.82}) +[2026-06-17 23:36:09,595][starVLA.training.train_starvla][INFO] - Step 4419, Loss: {'train/loss': 0.20465511083602905, 'train/grad_norm_pre_clip': 13.666681289672852, 'timing/data': 0.0006579272449016571, 'timing/model': 0.9505761377513409, 'learning_rate/qwen_vl_interface': 3.4192742493147686e-06, 'learning_rate/action_model': 3.419274249314768e-05, 'epoch': 1.82}) +[2026-06-17 23:36:13,454][starVLA.training.train_starvla][INFO] - Step 4420, Loss: {'train/loss': 0.2845698893070221, 'train/grad_norm_pre_clip': 21.040691375732422, 'timing/data': 0.0005682716146111488, 'timing/model': 0.9367098864167929, 'learning_rate/qwen_vl_interface': 3.4172788316451065e-06, 'learning_rate/action_model': 3.417278831645107e-05, 'epoch': 1.82}) +[2026-06-17 23:36:17,349][starVLA.training.train_starvla][INFO] - Step 4421, Loss: {'train/loss': 0.3740372061729431, 'train/grad_norm_pre_clip': 29.840063095092773, 'timing/data': 0.0005249325186014175, 'timing/model': 1.1744231805205345, 'learning_rate/qwen_vl_interface': 3.4152837939005177e-06, 'learning_rate/action_model': 3.415283793900518e-05, 'epoch': 1.82}) +[2026-06-17 23:36:20,995][starVLA.training.train_starvla][INFO] - Step 4422, Loss: {'train/loss': 0.13127632439136505, 'train/grad_norm_pre_clip': 9.225142478942871, 'timing/data': 0.0005088811740279198, 'timing/model': 0.9389948798343539, 'learning_rate/qwen_vl_interface': 3.413289136494572e-06, 'learning_rate/action_model': 3.413289136494572e-05, 'epoch': 1.82}) +[2026-06-17 23:36:24,809][starVLA.training.train_starvla][INFO] - Step 4423, Loss: {'train/loss': 0.46471309661865234, 'train/grad_norm_pre_clip': 21.654176712036133, 'timing/data': 0.00047318730503320694, 'timing/model': 0.9473704434931278, 'learning_rate/qwen_vl_interface': 3.411294859840765e-06, 'learning_rate/action_model': 3.411294859840764e-05, 'epoch': 1.82}) +[2026-06-17 23:36:28,630][starVLA.training.train_starvla][INFO] - Step 4424, Loss: {'train/loss': 0.28013861179351807, 'train/grad_norm_pre_clip': 15.030591011047363, 'timing/data': 0.000476042740046978, 'timing/model': 0.9485465362668037, 'learning_rate/qwen_vl_interface': 3.409300964352511e-06, 'learning_rate/action_model': 3.409300964352511e-05, 'epoch': 1.82}) +[2026-06-17 23:36:32,509][starVLA.training.train_starvla][INFO] - Step 4425, Loss: {'train/loss': 0.37351375818252563, 'train/grad_norm_pre_clip': 20.642013549804688, 'timing/data': 0.00046825967729091644, 'timing/model': 0.9698161762207747, 'learning_rate/qwen_vl_interface': 3.407307450443148e-06, 'learning_rate/action_model': 3.407307450443148e-05, 'epoch': 1.82}) +[2026-06-17 23:36:36,346][starVLA.training.train_starvla][INFO] - Step 4426, Loss: {'train/loss': 0.3941340744495392, 'train/grad_norm_pre_clip': 20.456153869628906, 'timing/data': 0.0006195418536663055, 'timing/model': 0.9390210881829262, 'learning_rate/qwen_vl_interface': 3.4053143185259353e-06, 'learning_rate/action_model': 3.405314318525935e-05, 'epoch': 1.82}) +[2026-06-17 23:36:40,407][starVLA.training.train_starvla][INFO] - Step 4427, Loss: {'train/loss': 0.5229433178901672, 'train/grad_norm_pre_clip': 19.119468688964844, 'timing/data': 0.0006454372778534889, 'timing/model': 1.247035096399486, 'learning_rate/qwen_vl_interface': 3.4033215690140475e-06, 'learning_rate/action_model': 3.4033215690140476e-05, 'epoch': 1.82}) +[2026-06-17 23:36:45,376][starVLA.training.train_starvla][INFO] - Step 4428, Loss: {'train/loss': 0.4811062812805176, 'train/grad_norm_pre_clip': 24.093286514282227, 'timing/data': 0.001152566634118557, 'timing/model': 2.1860823137685657, 'learning_rate/qwen_vl_interface': 3.4013292023205852e-06, 'learning_rate/action_model': 3.401329202320585e-05, 'epoch': 1.82}) +[2026-06-17 23:36:49,087][starVLA.training.train_starvla][INFO] - Step 4429, Loss: {'train/loss': 0.2318754643201828, 'train/grad_norm_pre_clip': 14.934494972229004, 'timing/data': 0.0005261022597551346, 'timing/model': 0.9484660942107439, 'learning_rate/qwen_vl_interface': 3.3993372188585694e-06, 'learning_rate/action_model': 3.3993372188585694e-05, 'epoch': 1.82}) +[2026-06-17 23:36:54,391][starVLA.training.train_starvla][INFO] - Step 4430, Loss: {'train/loss': 0.2730904817581177, 'train/grad_norm_pre_clip': 21.423749923706055, 'timing/data': 0.000607355497777462, 'timing/model': 1.0125554949045181, 'learning_rate/qwen_vl_interface': 3.3973456190409397e-06, 'learning_rate/action_model': 3.39734561904094e-05, 'epoch': 1.82}) +[2026-06-17 23:36:58,443][starVLA.training.train_starvla][INFO] - Step 4431, Loss: {'train/loss': 0.3797411322593689, 'train/grad_norm_pre_clip': 14.856919288635254, 'timing/data': 0.0009441692382097244, 'timing/model': 0.9704651860520244, 'learning_rate/qwen_vl_interface': 3.395354403280555e-06, 'learning_rate/action_model': 3.395354403280555e-05, 'epoch': 1.82}) +[2026-06-17 23:37:02,631][starVLA.training.train_starvla][INFO] - Step 4432, Loss: {'train/loss': 0.3790794014930725, 'train/grad_norm_pre_clip': 18.62516212463379, 'timing/data': 0.0006775017827749252, 'timing/model': 0.9740349175408483, 'learning_rate/qwen_vl_interface': 3.3933635719902e-06, 'learning_rate/action_model': 3.3933635719901995e-05, 'epoch': 1.82}) +[2026-06-17 23:37:07,121][starVLA.training.train_starvla][INFO] - Step 4433, Loss: {'train/loss': 0.3430769145488739, 'train/grad_norm_pre_clip': 19.879364013671875, 'timing/data': 0.00048265885561704636, 'timing/model': 0.956350720487535, 'learning_rate/qwen_vl_interface': 3.3913731255825754e-06, 'learning_rate/action_model': 3.3913731255825754e-05, 'epoch': 1.82}) +[2026-06-17 23:37:10,958][starVLA.training.train_starvla][INFO] - Step 4434, Loss: {'train/loss': 0.3613927364349365, 'train/grad_norm_pre_clip': 18.76403045654297, 'timing/data': 0.000533214770257473, 'timing/model': 1.1238789791241288, 'learning_rate/qwen_vl_interface': 3.3893830644703003e-06, 'learning_rate/action_model': 3.3893830644703e-05, 'epoch': 1.83}) +[2026-06-17 23:37:14,712][starVLA.training.train_starvla][INFO] - Step 4435, Loss: {'train/loss': 0.541964590549469, 'train/grad_norm_pre_clip': 22.073471069335938, 'timing/data': 0.0005127498880028725, 'timing/model': 1.0345514826476574, 'learning_rate/qwen_vl_interface': 3.3873933890659174e-06, 'learning_rate/action_model': 3.3873933890659174e-05, 'epoch': 1.83}) +[2026-06-17 23:37:18,509][starVLA.training.train_starvla][INFO] - Step 4436, Loss: {'train/loss': 0.4766537547111511, 'train/grad_norm_pre_clip': 20.524749755859375, 'timing/data': 0.0007072705775499344, 'timing/model': 0.9388275761157274, 'learning_rate/qwen_vl_interface': 3.3854040997818897e-06, 'learning_rate/action_model': 3.38540409978189e-05, 'epoch': 1.83}) +[2026-06-17 23:37:22,404][starVLA.training.train_starvla][INFO] - Step 4437, Loss: {'train/loss': 0.25689855217933655, 'train/grad_norm_pre_clip': 14.937460899353027, 'timing/data': 0.001052049919962883, 'timing/model': 0.9461881099268794, 'learning_rate/qwen_vl_interface': 3.383415197030599e-06, 'learning_rate/action_model': 3.383415197030599e-05, 'epoch': 1.83}) +[2026-06-17 23:37:26,221][starVLA.training.train_starvla][INFO] - Step 4438, Loss: {'train/loss': 0.24172043800354004, 'train/grad_norm_pre_clip': 16.782981872558594, 'timing/data': 0.0004852320998907089, 'timing/model': 0.9427030207589269, 'learning_rate/qwen_vl_interface': 3.3814266812243473e-06, 'learning_rate/action_model': 3.3814266812243474e-05, 'epoch': 1.83}) +[2026-06-17 23:37:30,261][starVLA.training.train_starvla][INFO] - Step 4439, Loss: {'train/loss': 0.2540248930454254, 'train/grad_norm_pre_clip': 13.37639045715332, 'timing/data': 0.00048432406038045883, 'timing/model': 0.9380930541083217, 'learning_rate/qwen_vl_interface': 3.3794385527753522e-06, 'learning_rate/action_model': 3.379438552775352e-05, 'epoch': 1.83}) +[2026-06-17 23:37:34,409][starVLA.training.train_starvla][INFO] - Step 4440, Loss: {'train/loss': 0.3677849769592285, 'train/grad_norm_pre_clip': 46.583953857421875, 'timing/data': 0.0005266126245260239, 'timing/model': 1.1851616268977523, 'learning_rate/qwen_vl_interface': 3.3774508120957606e-06, 'learning_rate/action_model': 3.37745081209576e-05, 'epoch': 1.83}) +[2026-06-17 23:37:38,381][starVLA.training.train_starvla][INFO] - Step 4441, Loss: {'train/loss': 0.2526668310165405, 'train/grad_norm_pre_clip': 14.387742042541504, 'timing/data': 0.0005038557574152946, 'timing/model': 1.1471344586461782, 'learning_rate/qwen_vl_interface': 3.37546345959763e-06, 'learning_rate/action_model': 3.3754634595976295e-05, 'epoch': 1.83}) +[2026-06-17 23:37:43,186][starVLA.training.train_starvla][INFO] - Step 4442, Loss: {'train/loss': 0.4974716901779175, 'train/grad_norm_pre_clip': 21.206581115722656, 'timing/data': 0.0005725882947444916, 'timing/model': 0.9479466779157519, 'learning_rate/qwen_vl_interface': 3.3734764956929407e-06, 'learning_rate/action_model': 3.37347649569294e-05, 'epoch': 1.83}) +[2026-06-17 23:37:47,010][starVLA.training.train_starvla][INFO] - Step 4443, Loss: {'train/loss': 0.5200705528259277, 'train/grad_norm_pre_clip': 20.88853645324707, 'timing/data': 0.0012675123289227486, 'timing/model': 0.9365244852378964, 'learning_rate/qwen_vl_interface': 3.3714899207935937e-06, 'learning_rate/action_model': 3.371489920793594e-05, 'epoch': 1.83}) +[2026-06-17 23:37:50,783][starVLA.training.train_starvla][INFO] - Step 4444, Loss: {'train/loss': 0.2937155067920685, 'train/grad_norm_pre_clip': 11.787219047546387, 'timing/data': 0.0012001292780041695, 'timing/model': 0.9270954942330718, 'learning_rate/qwen_vl_interface': 3.369503735311407e-06, 'learning_rate/action_model': 3.369503735311407e-05, 'epoch': 1.83}) +[2026-06-17 23:37:54,633][starVLA.training.train_starvla][INFO] - Step 4445, Loss: {'train/loss': 0.4779803156852722, 'train/grad_norm_pre_clip': 23.861675262451172, 'timing/data': 0.0005498640239238739, 'timing/model': 0.9451762735843658, 'learning_rate/qwen_vl_interface': 3.3675179396581208e-06, 'learning_rate/action_model': 3.367517939658121e-05, 'epoch': 1.83}) +[2026-06-17 23:37:58,434][starVLA.training.train_starvla][INFO] - Step 4446, Loss: {'train/loss': 0.31507954001426697, 'train/grad_norm_pre_clip': 14.140121459960938, 'timing/data': 0.0009886296465992928, 'timing/model': 0.9370691645890474, 'learning_rate/qwen_vl_interface': 3.365532534245389e-06, 'learning_rate/action_model': 3.3655325342453894e-05, 'epoch': 1.83}) +[2026-06-17 23:38:02,269][starVLA.training.train_starvla][INFO] - Step 4447, Loss: {'train/loss': 0.23688943684101105, 'train/grad_norm_pre_clip': 26.365554809570312, 'timing/data': 0.0005119144916534424, 'timing/model': 0.9369989419355989, 'learning_rate/qwen_vl_interface': 3.3635475194847927e-06, 'learning_rate/action_model': 3.363547519484793e-05, 'epoch': 1.83}) +[2026-06-17 23:38:06,136][starVLA.training.train_starvla][INFO] - Step 4448, Loss: {'train/loss': 0.18120750784873962, 'train/grad_norm_pre_clip': 12.172539710998535, 'timing/data': 0.0009538447484374046, 'timing/model': 1.1416416252031922, 'learning_rate/qwen_vl_interface': 3.361562895787825e-06, 'learning_rate/action_model': 3.361562895787825e-05, 'epoch': 1.83}) +[2026-06-17 23:38:09,805][starVLA.training.train_starvla][INFO] - Step 4449, Loss: {'train/loss': 0.6495242118835449, 'train/grad_norm_pre_clip': 27.626501083374023, 'timing/data': 0.0009566852822899818, 'timing/model': 0.9509643921628594, 'learning_rate/qwen_vl_interface': 3.3595786635659e-06, 'learning_rate/action_model': 3.3595786635659e-05, 'epoch': 1.83}) +[2026-06-17 23:38:13,631][starVLA.training.train_starvla][INFO] - Step 4450, Loss: {'train/loss': 0.16882844269275665, 'train/grad_norm_pre_clip': 15.851667404174805, 'timing/data': 0.0005483701825141907, 'timing/model': 0.9447802202776074, 'learning_rate/qwen_vl_interface': 3.357594823230353e-06, 'learning_rate/action_model': 3.3575948232303525e-05, 'epoch': 1.83}) +[2026-06-17 23:38:17,440][starVLA.training.train_starvla][INFO] - Step 4451, Loss: {'train/loss': 0.7141250967979431, 'train/grad_norm_pre_clip': 27.845958709716797, 'timing/data': 0.0005032047629356384, 'timing/model': 0.9296440156176686, 'learning_rate/qwen_vl_interface': 3.3556113751924357e-06, 'learning_rate/action_model': 3.355611375192436e-05, 'epoch': 1.83}) +[2026-06-17 23:38:21,216][starVLA.training.train_starvla][INFO] - Step 4452, Loss: {'train/loss': 0.17663165926933289, 'train/grad_norm_pre_clip': 11.967289924621582, 'timing/data': 0.0004995148628950119, 'timing/model': 0.9496454261243343, 'learning_rate/qwen_vl_interface': 3.3536283198633183e-06, 'learning_rate/action_model': 3.353628319863318e-05, 'epoch': 1.83}) +[2026-06-17 23:38:25,055][starVLA.training.train_starvla][INFO] - Step 4453, Loss: {'train/loss': 0.38740819692611694, 'train/grad_norm_pre_clip': 21.58180046081543, 'timing/data': 0.0009667174890637398, 'timing/model': 0.9552990579977632, 'learning_rate/qwen_vl_interface': 3.3516456576540913e-06, 'learning_rate/action_model': 3.3516456576540914e-05, 'epoch': 1.83}) +[2026-06-17 23:38:29,119][starVLA.training.train_starvla][INFO] - Step 4454, Loss: {'train/loss': 0.4379150867462158, 'train/grad_norm_pre_clip': 21.423189163208008, 'timing/data': 0.0009687775745987892, 'timing/model': 1.125455716624856, 'learning_rate/qwen_vl_interface': 3.349663388975763e-06, 'learning_rate/action_model': 3.3496633889757625e-05, 'epoch': 1.83}) +[2026-06-17 23:38:32,956][starVLA.training.train_starvla][INFO] - Step 4455, Loss: {'train/loss': 0.4945513606071472, 'train/grad_norm_pre_clip': 21.424285888671875, 'timing/data': 0.00047416798770427704, 'timing/model': 1.120626906864345, 'learning_rate/qwen_vl_interface': 3.347681514239258e-06, 'learning_rate/action_model': 3.347681514239257e-05, 'epoch': 1.83}) +[2026-06-17 23:38:36,608][starVLA.training.train_starvla][INFO] - Step 4456, Loss: {'train/loss': 0.14927741885185242, 'train/grad_norm_pre_clip': 14.626655578613281, 'timing/data': 0.0005212044343352318, 'timing/model': 0.9570405967533588, 'learning_rate/qwen_vl_interface': 3.345700033855421e-06, 'learning_rate/action_model': 3.345700033855421e-05, 'epoch': 1.83}) +[2026-06-17 23:38:40,519][starVLA.training.train_starvla][INFO] - Step 4457, Loss: {'train/loss': 0.3360046446323395, 'train/grad_norm_pre_clip': 16.942256927490234, 'timing/data': 0.0005020732060074806, 'timing/model': 0.9654151713475585, 'learning_rate/qwen_vl_interface': 3.3437189482350173e-06, 'learning_rate/action_model': 3.343718948235017e-05, 'epoch': 1.83}) +[2026-06-17 23:38:44,403][starVLA.training.train_starvla][INFO] - Step 4458, Loss: {'train/loss': 0.2870638370513916, 'train/grad_norm_pre_clip': 19.297346115112305, 'timing/data': 0.000490955077111721, 'timing/model': 0.939962062984705, 'learning_rate/qwen_vl_interface': 3.341738257788727e-06, 'learning_rate/action_model': 3.3417382577887266e-05, 'epoch': 1.84}) +[2026-06-17 23:38:48,219][starVLA.training.train_starvla][INFO] - Step 4459, Loss: {'train/loss': 0.3145458698272705, 'train/grad_norm_pre_clip': 15.767646789550781, 'timing/data': 0.0005037151277065277, 'timing/model': 0.9466794040054083, 'learning_rate/qwen_vl_interface': 3.3397579629271487e-06, 'learning_rate/action_model': 3.3397579629271486e-05, 'epoch': 1.84}) +[2026-06-17 23:38:52,071][starVLA.training.train_starvla][INFO] - Step 4460, Loss: {'train/loss': 0.3245350420475006, 'train/grad_norm_pre_clip': 16.746543884277344, 'timing/data': 0.0013559097424149513, 'timing/model': 0.9598521599546075, 'learning_rate/qwen_vl_interface': 3.3377780640607997e-06, 'learning_rate/action_model': 3.3377780640608e-05, 'epoch': 1.84}) +[2026-06-17 23:38:55,886][starVLA.training.train_starvla][INFO] - Step 4461, Loss: {'train/loss': 0.3908276855945587, 'train/grad_norm_pre_clip': 15.719734191894531, 'timing/data': 0.0007355911657214165, 'timing/model': 1.1174926003441215, 'learning_rate/qwen_vl_interface': 3.3357985616001166e-06, 'learning_rate/action_model': 3.335798561600116e-05, 'epoch': 1.84}) +[2026-06-17 23:38:59,731][starVLA.training.train_starvla][INFO] - Step 4462, Loss: {'train/loss': 0.14158426225185394, 'train/grad_norm_pre_clip': 9.572898864746094, 'timing/data': 0.0005635088309645653, 'timing/model': 1.098846685141325, 'learning_rate/qwen_vl_interface': 3.3338194559554524e-06, 'learning_rate/action_model': 3.333819455955453e-05, 'epoch': 1.84}) +[2026-06-17 23:39:03,403][starVLA.training.train_starvla][INFO] - Step 4463, Loss: {'train/loss': 0.1623910665512085, 'train/grad_norm_pre_clip': 13.03492546081543, 'timing/data': 0.0005182996392250061, 'timing/model': 0.949089546687901, 'learning_rate/qwen_vl_interface': 3.3318407475370758e-06, 'learning_rate/action_model': 3.3318407475370755e-05, 'epoch': 1.84}) +[2026-06-17 23:39:07,272][starVLA.training.train_starvla][INFO] - Step 4464, Loss: {'train/loss': 0.44790393114089966, 'train/grad_norm_pre_clip': 20.691879272460938, 'timing/data': 0.0006164321675896645, 'timing/model': 0.9408920202404261, 'learning_rate/qwen_vl_interface': 3.329862436755176e-06, 'learning_rate/action_model': 3.329862436755176e-05, 'epoch': 1.84}) +[2026-06-17 23:39:11,095][starVLA.training.train_starvla][INFO] - Step 4465, Loss: {'train/loss': 0.5542320609092712, 'train/grad_norm_pre_clip': 23.536529541015625, 'timing/data': 0.0007102126255631447, 'timing/model': 0.9458146439865232, 'learning_rate/qwen_vl_interface': 3.3278845240198592e-06, 'learning_rate/action_model': 3.3278845240198595e-05, 'epoch': 1.84}) +[2026-06-17 23:39:14,929][starVLA.training.train_starvla][INFO] - Step 4466, Loss: {'train/loss': 0.4373462200164795, 'train/grad_norm_pre_clip': 22.346893310546875, 'timing/data': 0.0005060518160462379, 'timing/model': 1.0479075917974114, 'learning_rate/qwen_vl_interface': 3.325907009741149e-06, 'learning_rate/action_model': 3.3259070097411485e-05, 'epoch': 1.84}) +[2026-06-17 23:39:18,824][starVLA.training.train_starvla][INFO] - Step 4467, Loss: {'train/loss': 0.3239823281764984, 'train/grad_norm_pre_clip': 16.081567764282227, 'timing/data': 0.000493152067065239, 'timing/model': 0.9636121485382318, 'learning_rate/qwen_vl_interface': 3.323929894328987e-06, 'learning_rate/action_model': 3.323929894328987e-05, 'epoch': 1.84}) +[2026-06-17 23:39:22,606][starVLA.training.train_starvla][INFO] - Step 4468, Loss: {'train/loss': 0.18630751967430115, 'train/grad_norm_pre_clip': 20.42007064819336, 'timing/data': 0.0008929762989282608, 'timing/model': 1.1026159040629864, 'learning_rate/qwen_vl_interface': 3.321953178193228e-06, 'learning_rate/action_model': 3.321953178193228e-05, 'epoch': 1.84}) +[2026-06-17 23:39:26,624][starVLA.training.train_starvla][INFO] - Step 4469, Loss: {'train/loss': 0.22649478912353516, 'train/grad_norm_pre_clip': 16.615234375, 'timing/data': 0.0007872208952903748, 'timing/model': 0.9562062583863735, 'learning_rate/qwen_vl_interface': 3.319976861743653e-06, 'learning_rate/action_model': 3.319976861743653e-05, 'epoch': 1.84}) +[2026-06-17 23:39:30,704][starVLA.training.train_starvla][INFO] - Step 4470, Loss: {'train/loss': 0.4831182360649109, 'train/grad_norm_pre_clip': 22.780275344848633, 'timing/data': 0.00045743677765130997, 'timing/model': 1.1509199915453792, 'learning_rate/qwen_vl_interface': 3.3180009453899496e-06, 'learning_rate/action_model': 3.318000945389949e-05, 'epoch': 1.84}) +[2026-06-17 23:39:37,414][starVLA.training.train_starvla][INFO] - Step 4471, Loss: {'train/loss': 0.3469286859035492, 'train/grad_norm_pre_clip': 16.067569732666016, 'timing/data': 0.00054167490452528, 'timing/model': 3.841572212986648, 'learning_rate/qwen_vl_interface': 3.3160254295417287e-06, 'learning_rate/action_model': 3.3160254295417285e-05, 'epoch': 1.84}) +[2026-06-17 23:39:41,464][starVLA.training.train_starvla][INFO] - Step 4472, Loss: {'train/loss': 0.28371956944465637, 'train/grad_norm_pre_clip': 17.456911087036133, 'timing/data': 0.0007911650463938713, 'timing/model': 0.9610813306644559, 'learning_rate/qwen_vl_interface': 3.314050314608518e-06, 'learning_rate/action_model': 3.314050314608518e-05, 'epoch': 1.84}) +[2026-06-17 23:39:45,481][starVLA.training.train_starvla][INFO] - Step 4473, Loss: {'train/loss': 0.18156474828720093, 'train/grad_norm_pre_clip': 12.08492660522461, 'timing/data': 0.0004817964509129524, 'timing/model': 0.9508797265589237, 'learning_rate/qwen_vl_interface': 3.3120756009997606e-06, 'learning_rate/action_model': 3.312075600999761e-05, 'epoch': 1.84}) +[2026-06-17 23:39:49,339][starVLA.training.train_starvla][INFO] - Step 4474, Loss: {'train/loss': 0.2730938196182251, 'train/grad_norm_pre_clip': 17.316728591918945, 'timing/data': 0.0005149943754076958, 'timing/model': 1.1296615079045296, 'learning_rate/qwen_vl_interface': 3.310101289124817e-06, 'learning_rate/action_model': 3.310101289124817e-05, 'epoch': 1.84}) +[2026-06-17 23:39:53,251][starVLA.training.train_starvla][INFO] - Step 4475, Loss: {'train/loss': 0.3059599697589874, 'train/grad_norm_pre_clip': 16.895448684692383, 'timing/data': 0.0005079479888081551, 'timing/model': 1.0873035052791238, 'learning_rate/qwen_vl_interface': 3.3081273793929615e-06, 'learning_rate/action_model': 3.308127379392961e-05, 'epoch': 1.84}) +[2026-06-17 23:39:56,945][starVLA.training.train_starvla][INFO] - Step 4476, Loss: {'train/loss': 0.28606879711151123, 'train/grad_norm_pre_clip': 19.737878799438477, 'timing/data': 0.0006280811503529549, 'timing/model': 0.9338857494294643, 'learning_rate/qwen_vl_interface': 3.306153872213392e-06, 'learning_rate/action_model': 3.306153872213392e-05, 'epoch': 1.84}) +[2026-06-17 23:40:00,851][starVLA.training.train_starvla][INFO] - Step 4477, Loss: {'train/loss': 0.29515567421913147, 'train/grad_norm_pre_clip': 14.078635215759277, 'timing/data': 0.00047275424003601074, 'timing/model': 0.9315234571695328, 'learning_rate/qwen_vl_interface': 3.304180767995215e-06, 'learning_rate/action_model': 3.3041807679952146e-05, 'epoch': 1.84}) +[2026-06-17 23:40:04,802][starVLA.training.train_starvla][INFO] - Step 4478, Loss: {'train/loss': 0.05160090699791908, 'train/grad_norm_pre_clip': 6.760575771331787, 'timing/data': 0.0005399556830525398, 'timing/model': 0.9527224488556385, 'learning_rate/qwen_vl_interface': 3.3022080671474588e-06, 'learning_rate/action_model': 3.302208067147459e-05, 'epoch': 1.84}) +[2026-06-17 23:40:08,629][starVLA.training.train_starvla][INFO] - Step 4479, Loss: {'train/loss': 0.22498412430286407, 'train/grad_norm_pre_clip': 15.650276184082031, 'timing/data': 0.0008223243057727814, 'timing/model': 0.9389596525579691, 'learning_rate/qwen_vl_interface': 3.300235770079066e-06, 'learning_rate/action_model': 3.300235770079066e-05, 'epoch': 1.84}) +[2026-06-17 23:40:12,755][starVLA.training.train_starvla][INFO] - Step 4480, Loss: {'train/loss': 0.47691383957862854, 'train/grad_norm_pre_clip': 20.558277130126953, 'timing/data': 0.0005026021972298622, 'timing/model': 1.187566944397986, 'learning_rate/qwen_vl_interface': 3.298263877198896e-06, 'learning_rate/action_model': 3.298263877198896e-05, 'epoch': 1.84}) +[2026-06-17 23:40:16,645][starVLA.training.train_starvla][INFO] - Step 4481, Loss: {'train/loss': 0.33208465576171875, 'train/grad_norm_pre_clip': 23.477750778198242, 'timing/data': 0.0013841427862644196, 'timing/model': 1.1471746265888214, 'learning_rate/qwen_vl_interface': 3.296292388915726e-06, 'learning_rate/action_model': 3.296292388915726e-05, 'epoch': 1.84}) +[2026-06-17 23:40:20,387][starVLA.training.train_starvla][INFO] - Step 4482, Loss: {'train/loss': 0.1637880802154541, 'train/grad_norm_pre_clip': 12.694567680358887, 'timing/data': 0.0011576814576983452, 'timing/model': 0.9303012229502201, 'learning_rate/qwen_vl_interface': 3.294321305638243e-06, 'learning_rate/action_model': 3.294321305638243e-05, 'epoch': 1.84}) +[2026-06-17 23:40:25,704][starVLA.training.train_starvla][INFO] - Step 4483, Loss: {'train/loss': 0.2109866887331009, 'train/grad_norm_pre_clip': 13.93171501159668, 'timing/data': 0.0015789661556482315, 'timing/model': 0.9455922041088343, 'learning_rate/qwen_vl_interface': 3.292350627775061e-06, 'learning_rate/action_model': 3.292350627775061e-05, 'epoch': 1.85}) +[2026-06-17 23:40:29,563][starVLA.training.train_starvla][INFO] - Step 4484, Loss: {'train/loss': 0.19438999891281128, 'train/grad_norm_pre_clip': 16.26121711730957, 'timing/data': 0.0008172439411282539, 'timing/model': 0.9331953413784504, 'learning_rate/qwen_vl_interface': 3.2903803557346997e-06, 'learning_rate/action_model': 3.2903803557347e-05, 'epoch': 1.85}) +[2026-06-17 23:40:33,428][starVLA.training.train_starvla][INFO] - Step 4485, Loss: {'train/loss': 0.28398674726486206, 'train/grad_norm_pre_clip': 18.039932250976562, 'timing/data': 0.0005303351208567619, 'timing/model': 0.9465832430869341, 'learning_rate/qwen_vl_interface': 3.2884104899255986e-06, 'learning_rate/action_model': 3.288410489925599e-05, 'epoch': 1.85}) +[2026-06-17 23:40:37,314][starVLA.training.train_starvla][INFO] - Step 4486, Loss: {'train/loss': 0.21817491948604584, 'train/grad_norm_pre_clip': 15.653382301330566, 'timing/data': 0.0004872884601354599, 'timing/model': 0.9422620898112655, 'learning_rate/qwen_vl_interface': 3.2864410307561124e-06, 'learning_rate/action_model': 3.286441030756112e-05, 'epoch': 1.85}) +[2026-06-17 23:40:41,234][starVLA.training.train_starvla][INFO] - Step 4487, Loss: {'train/loss': 0.5758489966392517, 'train/grad_norm_pre_clip': 33.86777114868164, 'timing/data': 0.001003003679215908, 'timing/model': 0.9388275323435664, 'learning_rate/qwen_vl_interface': 3.284471978634515e-06, 'learning_rate/action_model': 3.284471978634515e-05, 'epoch': 1.85}) +[2026-06-17 23:40:45,163][starVLA.training.train_starvla][INFO] - Step 4488, Loss: {'train/loss': 0.26259559392929077, 'train/grad_norm_pre_clip': 16.449634552001953, 'timing/data': 0.00048455502837896347, 'timing/model': 1.1827903874218464, 'learning_rate/qwen_vl_interface': 3.2825033339689926e-06, 'learning_rate/action_model': 3.282503333968993e-05, 'epoch': 1.85}) +[2026-06-17 23:40:48,863][starVLA.training.train_starvla][INFO] - Step 4489, Loss: {'train/loss': 0.38852304220199585, 'train/grad_norm_pre_clip': 18.537601470947266, 'timing/data': 0.000708412379026413, 'timing/model': 0.9336561216041446, 'learning_rate/qwen_vl_interface': 3.2805350971676447e-06, 'learning_rate/action_model': 3.280535097167645e-05, 'epoch': 1.85}) +[2026-06-17 23:40:52,839][starVLA.training.train_starvla][INFO] - Step 4490, Loss: {'train/loss': 0.36318448185920715, 'train/grad_norm_pre_clip': 18.854734420776367, 'timing/data': 0.0005448441952466965, 'timing/model': 0.9560559373348951, 'learning_rate/qwen_vl_interface': 3.2785672686384893e-06, 'learning_rate/action_model': 3.278567268638489e-05, 'epoch': 1.85}) +[2026-06-17 23:40:56,641][starVLA.training.train_starvla][INFO] - Step 4491, Loss: {'train/loss': 0.4031451642513275, 'train/grad_norm_pre_clip': 18.446889877319336, 'timing/data': 0.001190229319036007, 'timing/model': 0.9329916220158339, 'learning_rate/qwen_vl_interface': 3.2765998487894622e-06, 'learning_rate/action_model': 3.276599848789462e-05, 'epoch': 1.85}) +[2026-06-17 23:41:00,483][starVLA.training.train_starvla][INFO] - Step 4492, Loss: {'train/loss': 0.19558583199977875, 'train/grad_norm_pre_clip': 17.250110626220703, 'timing/data': 0.0004933280870318413, 'timing/model': 0.9311558566987514, 'learning_rate/qwen_vl_interface': 3.2746328380284098e-06, 'learning_rate/action_model': 3.27463283802841e-05, 'epoch': 1.85}) +[2026-06-17 23:41:04,338][starVLA.training.train_starvla][INFO] - Step 4493, Loss: {'train/loss': 0.47713327407836914, 'train/grad_norm_pre_clip': 23.681289672851562, 'timing/data': 0.0008222311735153198, 'timing/model': 0.9388253092765808, 'learning_rate/qwen_vl_interface': 3.2726662367630953e-06, 'learning_rate/action_model': 3.272666236763095e-05, 'epoch': 1.85}) +[2026-06-17 23:41:08,292][starVLA.training.train_starvla][INFO] - Step 4494, Loss: {'train/loss': 0.38134995102882385, 'train/grad_norm_pre_clip': 29.828689575195312, 'timing/data': 0.0011931313201785088, 'timing/model': 1.0106414118781686, 'learning_rate/qwen_vl_interface': 3.270700045401197e-06, 'learning_rate/action_model': 3.2707000454011974e-05, 'epoch': 1.85}) +[2026-06-17 23:41:12,151][starVLA.training.train_starvla][INFO] - Step 4495, Loss: {'train/loss': 0.37454745173454285, 'train/grad_norm_pre_clip': 26.37689971923828, 'timing/data': 0.0004569143056869507, 'timing/model': 1.167968044988811, 'learning_rate/qwen_vl_interface': 3.26873426435031e-06, 'learning_rate/action_model': 3.2687342643503096e-05, 'epoch': 1.85}) +[2026-06-17 23:41:16,497][starVLA.training.train_starvla][INFO] - Step 4496, Loss: {'train/loss': 0.27181029319763184, 'train/grad_norm_pre_clip': 23.563093185424805, 'timing/data': 0.0008727042004466057, 'timing/model': 1.592174000106752, 'learning_rate/qwen_vl_interface': 3.266768894017942e-06, 'learning_rate/action_model': 3.266768894017942e-05, 'epoch': 1.85}) +[2026-06-17 23:41:20,368][starVLA.training.train_starvla][INFO] - Step 4497, Loss: {'train/loss': 0.10684522986412048, 'train/grad_norm_pre_clip': 9.598771095275879, 'timing/data': 0.00044727884232997894, 'timing/model': 0.9383013043552637, 'learning_rate/qwen_vl_interface': 3.264803934811514e-06, 'learning_rate/action_model': 3.264803934811514e-05, 'epoch': 1.85}) +[2026-06-17 23:41:24,221][starVLA.training.train_starvla][INFO] - Step 4498, Loss: {'train/loss': 0.3530239164829254, 'train/grad_norm_pre_clip': 19.520788192749023, 'timing/data': 0.0006082281470298767, 'timing/model': 0.932774493470788, 'learning_rate/qwen_vl_interface': 3.2628393871383706e-06, 'learning_rate/action_model': 3.26283938713837e-05, 'epoch': 1.85}) +[2026-06-17 23:41:28,039][starVLA.training.train_starvla][INFO] - Step 4499, Loss: {'train/loss': 0.3532528579235077, 'train/grad_norm_pre_clip': 21.92470932006836, 'timing/data': 0.00048436690121889114, 'timing/model': 0.9531229501590133, 'learning_rate/qwen_vl_interface': 3.2608752514057585e-06, 'learning_rate/action_model': 3.260875251405758e-05, 'epoch': 1.85}) +[2026-06-17 23:43:41,962][starVLA.training.train_starvla][INFO] - Step 4500, Loss: {'train/loss': 0.4444023370742798, 'train/grad_norm_pre_clip': 20.007549285888672, 'eval/loss': 0.4313433766365051, 'eval/latency_6/loss': 0.4313433766365051, 'eval/demon_fire_precision': 0.9489668127739511, 'eval/demon_fire_recall': 0.9796380090497737, 'eval/demon_fire_f1': 0.964058524173028, 'eval/demon_fire_rate_model': 0.8887033945464663, 'eval/demon_fire_rate_teacher': 0.860879243183083, 'eval/demon_left_precision': 0.9055299539170507, 'eval/demon_left_recall': 0.6894736842105263, 'eval/demon_left_f1': 0.7828685258964144, 'eval/demon_left_rate_model': 0.12075681691708402, 'eval/demon_left_rate_teacher': 0.15859766277128548, 'eval/demon_right_precision': 0.8133116883116883, 'eval/demon_right_recall': 0.7695852534562212, 'eval/demon_right_f1': 0.7908445146014207, 'eval/demon_right_rate_model': 0.17139677239844184, 'eval/demon_right_rate_teacher': 0.18113522537562604, 'eval/demon_move_f1': 0.7868565202489175, 'eval/demon_cc_f1': 0.9030479454504611, 'eval/latency_6/demon_fire_precision': 0.9489668127739511, 'eval/latency_6/demon_fire_recall': 0.9796380090497737, 'eval/latency_6/demon_fire_f1': 0.964058524173028, 'eval/latency_6/demon_fire_rate_model': 0.8887033945464663, 'eval/latency_6/demon_fire_rate_teacher': 0.860879243183083, 'eval/latency_6/demon_left_precision': 0.9055299539170507, 'eval/latency_6/demon_left_recall': 0.6894736842105263, 'eval/latency_6/demon_left_f1': 0.7828685258964144, 'eval/latency_6/demon_left_rate_model': 0.12075681691708402, 'eval/latency_6/demon_left_rate_teacher': 0.15859766277128548, 'eval/latency_6/demon_right_precision': 0.8133116883116883, 'eval/latency_6/demon_right_recall': 0.7695852534562212, 'eval/latency_6/demon_right_f1': 0.7908445146014207, 'eval/latency_6/demon_right_rate_model': 0.17139677239844184, 'eval/latency_6/demon_right_rate_teacher': 0.18113522537562604, 'eval/latency_6/demon_move_f1': 0.7868565202489175, 'eval/latency_6/demon_cc_f1': 0.9030479454504611, 'timing/data': 0.0006406642496585846, 'timing/model': 0.9932502107694745, 'learning_rate/qwen_vl_interface': 3.2589115280208473e-06, 'learning_rate/action_model': 3.258911528020847e-05, 'epoch': 1.85}) +[2026-06-17 23:43:41,966][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_4500_state +[2026-06-17 23:43:41,972][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 23:43:48,915][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_4500_state/model.safetensors +[2026-06-17 23:44:05,032][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_4500_state/optimizer.bin +[2026-06-17 23:44:05,032][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_4500_state/sampler.bin +[2026-06-17 23:44:05,034][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_4500_state/sampler_1.bin +[2026-06-17 23:44:05,037][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_4500_state/random_states_0.pkl +[2026-06-17 23:44:05,040][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4500 +[2026-06-17 23:44:05,040][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 23:44:05,050][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 23:44:09,141][starVLA.training.train_starvla][INFO] - Step 4501, Loss: {'train/loss': 0.33448970317840576, 'train/grad_norm_pre_clip': 20.421010971069336, 'timing/data': 0.0011045774444937706, 'timing/model': 0.9448262434452772, 'learning_rate/qwen_vl_interface': 3.2569482173907164e-06, 'learning_rate/action_model': 3.2569482173907165e-05, 'epoch': 1.85}) +[2026-06-17 23:44:13,104][starVLA.training.train_starvla][INFO] - Step 4502, Loss: {'train/loss': 0.3484240174293518, 'train/grad_norm_pre_clip': 23.369321823120117, 'timing/data': 0.000520290806889534, 'timing/model': 1.0450107036158442, 'learning_rate/qwen_vl_interface': 3.2549853199223668e-06, 'learning_rate/action_model': 3.254985319922367e-05, 'epoch': 1.85}) +[2026-06-17 23:44:16,936][starVLA.training.train_starvla][INFO] - Step 4503, Loss: {'train/loss': 0.1895979940891266, 'train/grad_norm_pre_clip': 12.064167022705078, 'timing/data': 0.000499330461025238, 'timing/model': 1.1285338681191206, 'learning_rate/qwen_vl_interface': 3.253022836022708e-06, 'learning_rate/action_model': 3.253022836022708e-05, 'epoch': 1.85}) +[2026-06-17 23:44:20,819][starVLA.training.train_starvla][INFO] - Step 4504, Loss: {'train/loss': 0.2349696159362793, 'train/grad_norm_pre_clip': 11.911871910095215, 'timing/data': 0.0007962537929415703, 'timing/model': 1.131578266620636, 'learning_rate/qwen_vl_interface': 3.2510607660985627e-06, 'learning_rate/action_model': 3.251060766098563e-05, 'epoch': 1.85}) +[2026-06-17 23:44:24,482][starVLA.training.train_starvla][INFO] - Step 4505, Loss: {'train/loss': 0.20795227587223053, 'train/grad_norm_pre_clip': 17.777246475219727, 'timing/data': 0.012233735993504524, 'timing/model': 0.9505101880058646, 'learning_rate/qwen_vl_interface': 3.249099110556672e-06, 'learning_rate/action_model': 3.249099110556672e-05, 'epoch': 1.85}) +[2026-06-17 23:44:28,656][starVLA.training.train_starvla][INFO] - Step 4506, Loss: {'train/loss': 0.5565048456192017, 'train/grad_norm_pre_clip': 21.721418380737305, 'timing/data': 0.009203140623867512, 'timing/model': 0.9436775473877788, 'learning_rate/qwen_vl_interface': 3.247137869803689e-06, 'learning_rate/action_model': 3.2471378698036886e-05, 'epoch': 1.85}) +[2026-06-17 23:44:32,528][starVLA.training.train_starvla][INFO] - Step 4507, Loss: {'train/loss': 0.3243751525878906, 'train/grad_norm_pre_clip': 17.792936325073242, 'timing/data': 0.012940249405801296, 'timing/model': 0.9327808953821659, 'learning_rate/qwen_vl_interface': 3.2451770442461813e-06, 'learning_rate/action_model': 3.245177044246181e-05, 'epoch': 1.86}) +[2026-06-17 23:44:36,329][starVLA.training.train_starvla][INFO] - Step 4508, Loss: {'train/loss': 0.23930802941322327, 'train/grad_norm_pre_clip': 16.967710494995117, 'timing/data': 0.013503575697541237, 'timing/model': 0.9366902876645327, 'learning_rate/qwen_vl_interface': 3.2432166342906286e-06, 'learning_rate/action_model': 3.2432166342906284e-05, 'epoch': 1.86}) +[2026-06-17 23:44:40,156][starVLA.training.train_starvla][INFO] - Step 4509, Loss: {'train/loss': 0.07738596200942993, 'train/grad_norm_pre_clip': 7.257208824157715, 'timing/data': 0.009289320558309555, 'timing/model': 0.9400856196880341, 'learning_rate/qwen_vl_interface': 3.241256640343427e-06, 'learning_rate/action_model': 3.241256640343427e-05, 'epoch': 1.86}) +[2026-06-17 23:44:44,018][starVLA.training.train_starvla][INFO] - Step 4510, Loss: {'train/loss': 0.3907657563686371, 'train/grad_norm_pre_clip': 21.551433563232422, 'timing/data': 0.008438104763627052, 'timing/model': 1.1228531021624804, 'learning_rate/qwen_vl_interface': 3.239297062810886e-06, 'learning_rate/action_model': 3.239297062810886e-05, 'epoch': 1.86}) +[2026-06-17 23:44:47,924][starVLA.training.train_starvla][INFO] - Step 4511, Loss: {'train/loss': 0.3853383958339691, 'train/grad_norm_pre_clip': 22.567668914794922, 'timing/data': 0.010736705735325813, 'timing/model': 1.1544498847797513, 'learning_rate/qwen_vl_interface': 3.2373379020992252e-06, 'learning_rate/action_model': 3.237337902099225e-05, 'epoch': 1.86}) +[2026-06-17 23:44:51,654][starVLA.training.train_starvla][INFO] - Step 4512, Loss: {'train/loss': 0.19132758677005768, 'train/grad_norm_pre_clip': 16.254743576049805, 'timing/data': 0.015443415381014347, 'timing/model': 0.9686899483203888, 'learning_rate/qwen_vl_interface': 3.2353791586145855e-06, 'learning_rate/action_model': 3.2353791586145855e-05, 'epoch': 1.86}) +[2026-06-17 23:44:55,626][starVLA.training.train_starvla][INFO] - Step 4513, Loss: {'train/loss': 0.21993935108184814, 'train/grad_norm_pre_clip': 13.970292091369629, 'timing/data': 0.013786853291094303, 'timing/model': 0.9478711020201445, 'learning_rate/qwen_vl_interface': 3.2334208327630158e-06, 'learning_rate/action_model': 3.233420832763016e-05, 'epoch': 1.86}) +[2026-06-17 23:44:59,727][starVLA.training.train_starvla][INFO] - Step 4514, Loss: {'train/loss': 0.26688152551651, 'train/grad_norm_pre_clip': 15.044757843017578, 'timing/data': 0.011162473820149899, 'timing/model': 0.9608975918963552, 'learning_rate/qwen_vl_interface': 3.231462924950475e-06, 'learning_rate/action_model': 3.2314629249504746e-05, 'epoch': 1.86}) +[2026-06-17 23:45:03,707][starVLA.training.train_starvla][INFO] - Step 4515, Loss: {'train/loss': 0.30386170744895935, 'train/grad_norm_pre_clip': 22.92747688293457, 'timing/data': 0.014593022875487804, 'timing/model': 0.9761305451393127, 'learning_rate/qwen_vl_interface': 3.2295054355828396e-06, 'learning_rate/action_model': 3.22950543558284e-05, 'epoch': 1.86}) +[2026-06-17 23:45:07,797][starVLA.training.train_starvla][INFO] - Step 4516, Loss: {'train/loss': 0.695749819278717, 'train/grad_norm_pre_clip': 25.31069564819336, 'timing/data': 0.016721614636480808, 'timing/model': 1.0152931157499552, 'learning_rate/qwen_vl_interface': 3.227548365065905e-06, 'learning_rate/action_model': 3.227548365065905e-05, 'epoch': 1.86}) +[2026-06-17 23:45:11,848][starVLA.training.train_starvla][INFO] - Step 4517, Loss: {'train/loss': 0.15755054354667664, 'train/grad_norm_pre_clip': 10.492339134216309, 'timing/data': 0.009766893461346626, 'timing/model': 1.3631500974297523, 'learning_rate/qwen_vl_interface': 3.2255917138053695e-06, 'learning_rate/action_model': 3.2255917138053696e-05, 'epoch': 1.86}) +[2026-06-17 23:45:15,800][starVLA.training.train_starvla][INFO] - Step 4518, Loss: {'train/loss': 0.4623856246471405, 'train/grad_norm_pre_clip': 25.076499938964844, 'timing/data': 0.00986940786242485, 'timing/model': 1.148434572853148, 'learning_rate/qwen_vl_interface': 3.2236354822068515e-06, 'learning_rate/action_model': 3.223635482206852e-05, 'epoch': 1.86}) +[2026-06-17 23:45:19,457][starVLA.training.train_starvla][INFO] - Step 4519, Loss: {'train/loss': 0.3116167485713959, 'train/grad_norm_pre_clip': 19.24042510986328, 'timing/data': 0.017580526880919933, 'timing/model': 0.9410610292106867, 'learning_rate/qwen_vl_interface': 3.2216796706758757e-06, 'learning_rate/action_model': 3.221679670675876e-05, 'epoch': 1.86}) +[2026-06-17 23:45:23,406][starVLA.training.train_starvla][INFO] - Step 4520, Loss: {'train/loss': 0.39283910393714905, 'train/grad_norm_pre_clip': 23.658445358276367, 'timing/data': 0.01311515737324953, 'timing/model': 0.9524677116423845, 'learning_rate/qwen_vl_interface': 3.219724279617888e-06, 'learning_rate/action_model': 3.219724279617888e-05, 'epoch': 1.86}) +[2026-06-17 23:45:27,316][starVLA.training.train_starvla][INFO] - Step 4521, Loss: {'train/loss': 0.3736283779144287, 'train/grad_norm_pre_clip': 21.991451263427734, 'timing/data': 0.012338410131633282, 'timing/model': 0.9462859993800521, 'learning_rate/qwen_vl_interface': 3.2177693094382424e-06, 'learning_rate/action_model': 3.2177693094382424e-05, 'epoch': 1.86}) +[2026-06-17 23:45:31,091][starVLA.training.train_starvla][INFO] - Step 4522, Loss: {'train/loss': 0.4666239619255066, 'train/grad_norm_pre_clip': 26.284069061279297, 'timing/data': 0.012350586242973804, 'timing/model': 0.9406884796917439, 'learning_rate/qwen_vl_interface': 3.2158147605422052e-06, 'learning_rate/action_model': 3.215814760542205e-05, 'epoch': 1.86}) +[2026-06-17 23:45:35,156][starVLA.training.train_starvla][INFO] - Step 4523, Loss: {'train/loss': 0.28636449575424194, 'train/grad_norm_pre_clip': 22.977336883544922, 'timing/data': 0.012387588620185852, 'timing/model': 0.9448721660301089, 'learning_rate/qwen_vl_interface': 3.2138606333349576e-06, 'learning_rate/action_model': 3.213860633334957e-05, 'epoch': 1.86}) +[2026-06-17 23:45:39,181][starVLA.training.train_starvla][INFO] - Step 4524, Loss: {'train/loss': 0.692043662071228, 'train/grad_norm_pre_clip': 26.126182556152344, 'timing/data': 0.019125432707369328, 'timing/model': 1.1779233505949378, 'learning_rate/qwen_vl_interface': 3.2119069282215913e-06, 'learning_rate/action_model': 3.211906928221591e-05, 'epoch': 1.86}) +[2026-06-17 23:45:43,097][starVLA.training.train_starvla][INFO] - Step 4525, Loss: {'train/loss': 0.2311796247959137, 'train/grad_norm_pre_clip': 15.037178993225098, 'timing/data': 0.009927111677825451, 'timing/model': 1.1318287244066596, 'learning_rate/qwen_vl_interface': 3.2099536456071118e-06, 'learning_rate/action_model': 3.209953645607112e-05, 'epoch': 1.86}) +[2026-06-17 23:45:48,117][starVLA.training.train_starvla][INFO] - Step 4526, Loss: {'train/loss': 0.4277903735637665, 'train/grad_norm_pre_clip': 24.474000930786133, 'timing/data': 0.010206438601016998, 'timing/model': 0.9574348218739033, 'learning_rate/qwen_vl_interface': 3.2080007858964347e-06, 'learning_rate/action_model': 3.2080007858964345e-05, 'epoch': 1.86}) +[2026-06-17 23:45:52,033][starVLA.training.train_starvla][INFO] - Step 4527, Loss: {'train/loss': 0.24361751973628998, 'train/grad_norm_pre_clip': 23.687152862548828, 'timing/data': 0.01219114288687706, 'timing/model': 0.9534537801519036, 'learning_rate/qwen_vl_interface': 3.206048349494396e-06, 'learning_rate/action_model': 3.206048349494396e-05, 'epoch': 1.86}) +[2026-06-17 23:45:55,910][starVLA.training.train_starvla][INFO] - Step 4528, Loss: {'train/loss': 0.32459017634391785, 'train/grad_norm_pre_clip': 20.725200653076172, 'timing/data': 0.014680774882435799, 'timing/model': 0.9391235811635852, 'learning_rate/qwen_vl_interface': 3.2040963368057323e-06, 'learning_rate/action_model': 3.204096336805732e-05, 'epoch': 1.86}) +[2026-06-17 23:45:59,793][starVLA.training.train_starvla][INFO] - Step 4529, Loss: {'train/loss': 0.2720956802368164, 'train/grad_norm_pre_clip': 15.990870475769043, 'timing/data': 0.00926436297595501, 'timing/model': 0.9456712389364839, 'learning_rate/qwen_vl_interface': 3.2021447482350994e-06, 'learning_rate/action_model': 3.202144748235099e-05, 'epoch': 1.86}) +[2026-06-17 23:46:03,835][starVLA.training.train_starvla][INFO] - Step 4530, Loss: {'train/loss': 0.40112632513046265, 'train/grad_norm_pre_clip': 24.535429000854492, 'timing/data': 0.008870274759829044, 'timing/model': 1.2308575240895152, 'learning_rate/qwen_vl_interface': 3.200193584187061e-06, 'learning_rate/action_model': 3.200193584187061e-05, 'epoch': 1.86}) +[2026-06-17 23:46:07,764][starVLA.training.train_starvla][INFO] - Step 4531, Loss: {'train/loss': 0.24009601771831512, 'train/grad_norm_pre_clip': 16.497013092041016, 'timing/data': 0.010937602259218693, 'timing/model': 1.147259222343564, 'learning_rate/qwen_vl_interface': 3.198242845066101e-06, 'learning_rate/action_model': 3.198242845066101e-05, 'epoch': 1.87}) +[2026-06-17 23:46:11,431][starVLA.training.train_starvla][INFO] - Step 4532, Loss: {'train/loss': 0.4052589237689972, 'train/grad_norm_pre_clip': 24.053184509277344, 'timing/data': 0.008434277959167957, 'timing/model': 0.9335228074342012, 'learning_rate/qwen_vl_interface': 3.1962925312766087e-06, 'learning_rate/action_model': 3.196292531276608e-05, 'epoch': 1.87}) +[2026-06-17 23:46:15,460][starVLA.training.train_starvla][INFO] - Step 4533, Loss: {'train/loss': 0.37375426292419434, 'train/grad_norm_pre_clip': 20.153696060180664, 'timing/data': 0.013500798493623734, 'timing/model': 0.9904796006157994, 'learning_rate/qwen_vl_interface': 3.19434264322288e-06, 'learning_rate/action_model': 3.19434264322288e-05, 'epoch': 1.87}) +[2026-06-17 23:46:19,469][starVLA.training.train_starvla][INFO] - Step 4534, Loss: {'train/loss': 0.155767023563385, 'train/grad_norm_pre_clip': 11.600201606750488, 'timing/data': 0.011876294389367104, 'timing/model': 0.9425939489156008, 'learning_rate/qwen_vl_interface': 3.1923931813091364e-06, 'learning_rate/action_model': 3.192393181309136e-05, 'epoch': 1.87}) +[2026-06-17 23:46:23,392][starVLA.training.train_starvla][INFO] - Step 4535, Loss: {'train/loss': 0.29205065965652466, 'train/grad_norm_pre_clip': 24.446680068969727, 'timing/data': 0.008805017918348312, 'timing/model': 0.9560245890170336, 'learning_rate/qwen_vl_interface': 3.1904441459395e-06, 'learning_rate/action_model': 3.1904441459395e-05, 'epoch': 1.87}) +[2026-06-17 23:46:27,272][starVLA.training.train_starvla][INFO] - Step 4536, Loss: {'train/loss': 1.03324556350708, 'train/grad_norm_pre_clip': 31.80904197692871, 'timing/data': 0.01261791493743658, 'timing/model': 0.9498612470924854, 'learning_rate/qwen_vl_interface': 3.1884955375180088e-06, 'learning_rate/action_model': 3.1884955375180085e-05, 'epoch': 1.87}) +[2026-06-17 23:46:31,231][starVLA.training.train_starvla][INFO] - Step 4537, Loss: {'train/loss': 0.24560652673244476, 'train/grad_norm_pre_clip': 14.334802627563477, 'timing/data': 0.00866279099136591, 'timing/model': 1.2626250395551324, 'learning_rate/qwen_vl_interface': 3.18654735644861e-06, 'learning_rate/action_model': 3.1865473564486095e-05, 'epoch': 1.87}) +[2026-06-17 23:46:36,008][starVLA.training.train_starvla][INFO] - Step 4538, Loss: {'train/loss': 0.3673509955406189, 'train/grad_norm_pre_clip': 18.46817398071289, 'timing/data': 0.012867070734500885, 'timing/model': 2.059397913515568, 'learning_rate/qwen_vl_interface': 3.184599603135165e-06, 'learning_rate/action_model': 3.1845996031351646e-05, 'epoch': 1.87}) +[2026-06-17 23:46:39,992][starVLA.training.train_starvla][INFO] - Step 4539, Loss: {'train/loss': 0.2393404245376587, 'train/grad_norm_pre_clip': 16.50078582763672, 'timing/data': 0.01110809575766325, 'timing/model': 0.9553764089941978, 'learning_rate/qwen_vl_interface': 3.1826522779814453e-06, 'learning_rate/action_model': 3.182652277981445e-05, 'epoch': 1.87}) +[2026-06-17 23:46:44,084][starVLA.training.train_starvla][INFO] - Step 4540, Loss: {'train/loss': 0.15027135610580444, 'train/grad_norm_pre_clip': 14.695881843566895, 'timing/data': 0.01411342155188322, 'timing/model': 1.0143430838361382, 'learning_rate/qwen_vl_interface': 3.1807053813911317e-06, 'learning_rate/action_model': 3.180705381391132e-05, 'epoch': 1.87}) +[2026-06-17 23:46:49,002][starVLA.training.train_starvla][INFO] - Step 4541, Loss: {'train/loss': 0.18429630994796753, 'train/grad_norm_pre_clip': 10.85822582244873, 'timing/data': 0.01259937509894371, 'timing/model': 2.0245716525241733, 'learning_rate/qwen_vl_interface': 3.17875891376782e-06, 'learning_rate/action_model': 3.1787589137678196e-05, 'epoch': 1.87}) +[2026-06-17 23:46:53,038][starVLA.training.train_starvla][INFO] - Step 4542, Loss: {'train/loss': 0.4823281764984131, 'train/grad_norm_pre_clip': 16.812511444091797, 'timing/data': 0.009932199493050575, 'timing/model': 0.9535335851833224, 'learning_rate/qwen_vl_interface': 3.176812875515014e-06, 'learning_rate/action_model': 3.176812875515014e-05, 'epoch': 1.87}) +[2026-06-17 23:46:56,860][starVLA.training.train_starvla][INFO] - Step 4543, Loss: {'train/loss': 0.35400351881980896, 'train/grad_norm_pre_clip': 22.336759567260742, 'timing/data': 0.014115986414253712, 'timing/model': 1.102464199066162, 'learning_rate/qwen_vl_interface': 3.1748672670361304e-06, 'learning_rate/action_model': 3.17486726703613e-05, 'epoch': 1.87}) +[2026-06-17 23:47:00,554][starVLA.training.train_starvla][INFO] - Step 4544, Loss: {'train/loss': 0.3666326701641083, 'train/grad_norm_pre_clip': 18.223093032836914, 'timing/data': 0.014288359321653843, 'timing/model': 0.9701218903064728, 'learning_rate/qwen_vl_interface': 3.172922088734493e-06, 'learning_rate/action_model': 3.172922088734493e-05, 'epoch': 1.87}) +[2026-06-17 23:47:04,504][starVLA.training.train_starvla][INFO] - Step 4545, Loss: {'train/loss': 0.24218954145908356, 'train/grad_norm_pre_clip': 19.165245056152344, 'timing/data': 0.013582926243543625, 'timing/model': 0.950856695882976, 'learning_rate/qwen_vl_interface': 3.1709773410133434e-06, 'learning_rate/action_model': 3.170977341013343e-05, 'epoch': 1.87}) +[2026-06-17 23:47:11,177][starVLA.training.train_starvla][INFO] - Step 4546, Loss: {'train/loss': 0.12061341851949692, 'train/grad_norm_pre_clip': 18.147571563720703, 'timing/data': 0.013091069646179676, 'timing/model': 1.0093434499576688, 'learning_rate/qwen_vl_interface': 3.16903302427583e-06, 'learning_rate/action_model': 3.16903302427583e-05, 'epoch': 1.87}) +[2026-06-17 23:47:16,416][starVLA.training.train_starvla][INFO] - Step 4547, Loss: {'train/loss': 0.21926981210708618, 'train/grad_norm_pre_clip': 15.960336685180664, 'timing/data': 0.012452092953026295, 'timing/model': 2.3644521720707417, 'learning_rate/qwen_vl_interface': 3.1670891389250073e-06, 'learning_rate/action_model': 3.167089138925007e-05, 'epoch': 1.87}) +[2026-06-17 23:47:20,436][starVLA.training.train_starvla][INFO] - Step 4548, Loss: {'train/loss': 0.21580442786216736, 'train/grad_norm_pre_clip': 16.11673927307129, 'timing/data': 0.008488190360367298, 'timing/model': 0.9338492639362812, 'learning_rate/qwen_vl_interface': 3.1651456853638464e-06, 'learning_rate/action_model': 3.165145685363846e-05, 'epoch': 1.87}) +[2026-06-17 23:47:24,389][starVLA.training.train_starvla][INFO] - Step 4549, Loss: {'train/loss': 0.2766207158565521, 'train/grad_norm_pre_clip': 15.832873344421387, 'timing/data': 0.011491320095956326, 'timing/model': 0.9568067537620664, 'learning_rate/qwen_vl_interface': 3.1632026639952303e-06, 'learning_rate/action_model': 3.16320266399523e-05, 'epoch': 1.87}) +[2026-06-17 23:47:28,331][starVLA.training.train_starvla][INFO] - Step 4550, Loss: {'train/loss': 0.22566655278205872, 'train/grad_norm_pre_clip': 14.58937931060791, 'timing/data': 0.018106689676642418, 'timing/model': 1.1334171118214726, 'learning_rate/qwen_vl_interface': 3.161260075221947e-06, 'learning_rate/action_model': 3.161260075221947e-05, 'epoch': 1.87}) +[2026-06-17 23:47:31,994][starVLA.training.train_starvla][INFO] - Step 4551, Loss: {'train/loss': 0.18290217220783234, 'train/grad_norm_pre_clip': 18.14949607849121, 'timing/data': 0.008328413590788841, 'timing/model': 0.9476864598691463, 'learning_rate/qwen_vl_interface': 3.1593179194467e-06, 'learning_rate/action_model': 3.1593179194467e-05, 'epoch': 1.87}) +[2026-06-17 23:47:35,889][starVLA.training.train_starvla][INFO] - Step 4552, Loss: {'train/loss': 0.30200985074043274, 'train/grad_norm_pre_clip': 17.44315528869629, 'timing/data': 0.012956211343407631, 'timing/model': 0.9875788064673543, 'learning_rate/qwen_vl_interface': 3.157376197072094e-06, 'learning_rate/action_model': 3.157376197072094e-05, 'epoch': 1.87}) +[2026-06-17 23:47:39,808][starVLA.training.train_starvla][INFO] - Step 4553, Loss: {'train/loss': 0.4197354018688202, 'train/grad_norm_pre_clip': 28.387493133544922, 'timing/data': 0.013585650362074375, 'timing/model': 0.974755022674799, 'learning_rate/qwen_vl_interface': 3.1554349085006557e-06, 'learning_rate/action_model': 3.155434908500656e-05, 'epoch': 1.87}) +[2026-06-17 23:47:43,741][starVLA.training.train_starvla][INFO] - Step 4554, Loss: {'train/loss': 0.3756987750530243, 'train/grad_norm_pre_clip': 19.396440505981445, 'timing/data': 0.013701040297746658, 'timing/model': 0.9318737052381039, 'learning_rate/qwen_vl_interface': 3.1534940541348152e-06, 'learning_rate/action_model': 3.153494054134815e-05, 'epoch': 1.87}) +[2026-06-17 23:47:47,601][starVLA.training.train_starvla][INFO] - Step 4555, Loss: {'train/loss': 0.3178732097148895, 'train/grad_norm_pre_clip': 17.375818252563477, 'timing/data': 0.010318852961063385, 'timing/model': 0.9464070154353976, 'learning_rate/qwen_vl_interface': 3.1515536343769133e-06, 'learning_rate/action_model': 3.151553634376913e-05, 'epoch': 1.88}) +[2026-06-17 23:47:51,455][starVLA.training.train_starvla][INFO] - Step 4556, Loss: {'train/loss': 0.5157852172851562, 'train/grad_norm_pre_clip': 22.416967391967773, 'timing/data': 0.015873730182647705, 'timing/model': 1.1331032859161496, 'learning_rate/qwen_vl_interface': 3.1496136496292e-06, 'learning_rate/action_model': 3.1496136496292e-05, 'epoch': 1.88}) +[2026-06-17 23:47:55,542][starVLA.training.train_starvla][INFO] - Step 4557, Loss: {'train/loss': 0.5581110119819641, 'train/grad_norm_pre_clip': 353.28741455078125, 'timing/data': 0.010309537872672081, 'timing/model': 1.2500966247171164, 'learning_rate/qwen_vl_interface': 3.1476741002938383e-06, 'learning_rate/action_model': 3.147674100293838e-05, 'epoch': 1.88}) +[2026-06-17 23:47:59,205][starVLA.training.train_starvla][INFO] - Step 4558, Loss: {'train/loss': 0.4129844903945923, 'train/grad_norm_pre_clip': 15.084836959838867, 'timing/data': 0.014800213277339935, 'timing/model': 0.9482941972091794, 'learning_rate/qwen_vl_interface': 3.145734986772897e-06, 'learning_rate/action_model': 3.1457349867728966e-05, 'epoch': 1.88}) +[2026-06-17 23:48:03,067][starVLA.training.train_starvla][INFO] - Step 4559, Loss: {'train/loss': 0.07256366312503815, 'train/grad_norm_pre_clip': 7.17071008682251, 'timing/data': 0.009830242954194546, 'timing/model': 0.9546354115009308, 'learning_rate/qwen_vl_interface': 3.1437963094683566e-06, 'learning_rate/action_model': 3.1437963094683564e-05, 'epoch': 1.88}) +[2026-06-17 23:48:06,909][starVLA.training.train_starvla][INFO] - Step 4560, Loss: {'train/loss': 0.3304436206817627, 'train/grad_norm_pre_clip': 18.65989112854004, 'timing/data': 0.007833078503608704, 'timing/model': 0.9473113473504782, 'learning_rate/qwen_vl_interface': 3.1418580687821103e-06, 'learning_rate/action_model': 3.14185806878211e-05, 'epoch': 1.88}) +[2026-06-17 23:48:11,089][starVLA.training.train_starvla][INFO] - Step 4561, Loss: {'train/loss': 0.22992031276226044, 'train/grad_norm_pre_clip': 14.687959671020508, 'timing/data': 0.011094201356172562, 'timing/model': 0.9471061453223228, 'learning_rate/qwen_vl_interface': 3.1399202651159532e-06, 'learning_rate/action_model': 3.1399202651159534e-05, 'epoch': 1.88}) +[2026-06-17 23:48:14,897][starVLA.training.train_starvla][INFO] - Step 4562, Loss: {'train/loss': 0.5219146609306335, 'train/grad_norm_pre_clip': 19.999473571777344, 'timing/data': 0.012802338227629662, 'timing/model': 0.9469674844294786, 'learning_rate/qwen_vl_interface': 3.1379828988715955e-06, 'learning_rate/action_model': 3.1379828988715956e-05, 'epoch': 1.88}) +[2026-06-17 23:48:18,830][starVLA.training.train_starvla][INFO] - Step 4563, Loss: {'train/loss': 0.5478607416152954, 'train/grad_norm_pre_clip': 22.90635871887207, 'timing/data': 0.012973073869943619, 'timing/model': 1.136487022973597, 'learning_rate/qwen_vl_interface': 3.136045970450654e-06, 'learning_rate/action_model': 3.136045970450654e-05, 'epoch': 1.88}) +[2026-06-17 23:48:22,643][starVLA.training.train_starvla][INFO] - Step 4564, Loss: {'train/loss': 0.29713600873947144, 'train/grad_norm_pre_clip': 18.806976318359375, 'timing/data': 0.013010925613343716, 'timing/model': 1.1314980620518327, 'learning_rate/qwen_vl_interface': 3.1341094802546595e-06, 'learning_rate/action_model': 3.1341094802546595e-05, 'epoch': 1.88}) +[2026-06-17 23:48:26,291][starVLA.training.train_starvla][INFO] - Step 4565, Loss: {'train/loss': 0.4797228276729584, 'train/grad_norm_pre_clip': 17.720983505249023, 'timing/data': 0.007144705392420292, 'timing/model': 0.9327518176287413, 'learning_rate/qwen_vl_interface': 3.132173428685048e-06, 'learning_rate/action_model': 3.132173428685048e-05, 'epoch': 1.88}) +[2026-06-17 23:48:30,375][starVLA.training.train_starvla][INFO] - Step 4566, Loss: {'train/loss': 0.45864635705947876, 'train/grad_norm_pre_clip': 30.52678108215332, 'timing/data': 0.008242211304605007, 'timing/model': 0.9411613466218114, 'learning_rate/qwen_vl_interface': 3.13023781614316e-06, 'learning_rate/action_model': 3.13023781614316e-05, 'epoch': 1.88}) +[2026-06-17 23:48:34,297][starVLA.training.train_starvla][INFO] - Step 4567, Loss: {'train/loss': 0.14445556700229645, 'train/grad_norm_pre_clip': 10.194186210632324, 'timing/data': 0.00855801347643137, 'timing/model': 0.992773650214076, 'learning_rate/qwen_vl_interface': 3.1283026430302553e-06, 'learning_rate/action_model': 3.128302643030255e-05, 'epoch': 1.88}) +[2026-06-17 23:48:38,071][starVLA.training.train_starvla][INFO] - Step 4568, Loss: {'train/loss': 0.2515086233615875, 'train/grad_norm_pre_clip': 16.98481559753418, 'timing/data': 0.011531113646924496, 'timing/model': 0.9397790906950831, 'learning_rate/qwen_vl_interface': 3.1263679097474958e-06, 'learning_rate/action_model': 3.126367909747496e-05, 'epoch': 1.88}) +[2026-06-17 23:48:41,938][starVLA.training.train_starvla][INFO] - Step 4569, Loss: {'train/loss': 0.6017351150512695, 'train/grad_norm_pre_clip': 24.68893051147461, 'timing/data': 0.012794007547199726, 'timing/model': 0.9279353842139244, 'learning_rate/qwen_vl_interface': 3.124433616695955e-06, 'learning_rate/action_model': 3.1244336166959545e-05, 'epoch': 1.88}) +[2026-06-17 23:48:45,731][starVLA.training.train_starvla][INFO] - Step 4570, Loss: {'train/loss': 0.4709498882293701, 'train/grad_norm_pre_clip': 14.456871032714844, 'timing/data': 0.013577445410192013, 'timing/model': 1.1298031862825155, 'learning_rate/qwen_vl_interface': 3.1224997642766115e-06, 'learning_rate/action_model': 3.122499764276611e-05, 'epoch': 1.88}) +[2026-06-17 23:48:49,304][starVLA.training.train_starvla][INFO] - Step 4571, Loss: {'train/loss': 0.36173519492149353, 'train/grad_norm_pre_clip': 20.239295959472656, 'timing/data': 0.008447807282209396, 'timing/model': 0.9241027170792222, 'learning_rate/qwen_vl_interface': 3.1205663528903563e-06, 'learning_rate/action_model': 3.1205663528903565e-05, 'epoch': 1.88}) +[2026-06-17 23:48:53,065][starVLA.training.train_starvla][INFO] - Step 4572, Loss: {'train/loss': 0.14370040595531464, 'train/grad_norm_pre_clip': 16.083528518676758, 'timing/data': 0.01921245828270912, 'timing/model': 0.9209079965949059, 'learning_rate/qwen_vl_interface': 3.118633382937987e-06, 'learning_rate/action_model': 3.118633382937987e-05, 'epoch': 1.88}) +[2026-06-17 23:48:56,812][starVLA.training.train_starvla][INFO] - Step 4573, Loss: {'train/loss': 0.6205083131790161, 'train/grad_norm_pre_clip': 21.086261749267578, 'timing/data': 0.01377289742231369, 'timing/model': 0.9286678535863757, 'learning_rate/qwen_vl_interface': 3.116700854820209e-06, 'learning_rate/action_model': 3.116700854820209e-05, 'epoch': 1.88}) +[2026-06-17 23:49:00,572][starVLA.training.train_starvla][INFO] - Step 4574, Loss: {'train/loss': 0.28281542658805847, 'train/grad_norm_pre_clip': 15.257224082946777, 'timing/data': 0.015063494443893433, 'timing/model': 0.9396872771903872, 'learning_rate/qwen_vl_interface': 3.1147687689376443e-06, 'learning_rate/action_model': 3.1147687689376444e-05, 'epoch': 1.88}) +[2026-06-17 23:49:05,053][starVLA.training.train_starvla][INFO] - Step 4575, Loss: {'train/loss': 0.25054118037223816, 'train/grad_norm_pre_clip': 12.180057525634766, 'timing/data': 0.011547092348337173, 'timing/model': 1.005759608000517, 'learning_rate/qwen_vl_interface': 3.1128371256908087e-06, 'learning_rate/action_model': 3.112837125690809e-05, 'epoch': 1.88}) +[2026-06-17 23:49:09,105][starVLA.training.train_starvla][INFO] - Step 4576, Loss: {'train/loss': 0.24598389863967896, 'train/grad_norm_pre_clip': 15.50216293334961, 'timing/data': 0.017737331800162792, 'timing/model': 1.1304273689165711, 'learning_rate/qwen_vl_interface': 3.1109059254801368e-06, 'learning_rate/action_model': 3.1109059254801366e-05, 'epoch': 1.88}) +[2026-06-17 23:49:12,882][starVLA.training.train_starvla][INFO] - Step 4577, Loss: {'train/loss': 0.4043623208999634, 'train/grad_norm_pre_clip': 17.00629425048828, 'timing/data': 0.009427756071090698, 'timing/model': 1.062160024419427, 'learning_rate/qwen_vl_interface': 3.108975168705966e-06, 'learning_rate/action_model': 3.1089751687059656e-05, 'epoch': 1.88}) +[2026-06-17 23:49:16,549][starVLA.training.train_starvla][INFO] - Step 4578, Loss: {'train/loss': 0.26523786783218384, 'train/grad_norm_pre_clip': 15.162641525268555, 'timing/data': 0.012887346558272839, 'timing/model': 0.9485876699909568, 'learning_rate/qwen_vl_interface': 3.107044855768549e-06, 'learning_rate/action_model': 3.107044855768549e-05, 'epoch': 1.88}) +[2026-06-17 23:49:20,349][starVLA.training.train_starvla][INFO] - Step 4579, Loss: {'train/loss': 0.5592754483222961, 'train/grad_norm_pre_clip': 19.845949172973633, 'timing/data': 0.011649239808321, 'timing/model': 0.9655829546973109, 'learning_rate/qwen_vl_interface': 3.1051149870680413e-06, 'learning_rate/action_model': 3.105114987068041e-05, 'epoch': 1.88}) +[2026-06-17 23:49:24,167][starVLA.training.train_starvla][INFO] - Step 4580, Loss: {'train/loss': 0.1955600380897522, 'train/grad_norm_pre_clip': 17.768049240112305, 'timing/data': 0.009459452703595161, 'timing/model': 0.9424814581871033, 'learning_rate/qwen_vl_interface': 3.103185563004503e-06, 'learning_rate/action_model': 3.103185563004503e-05, 'epoch': 1.89}) +[2026-06-17 23:49:28,092][starVLA.training.train_starvla][INFO] - Step 4581, Loss: {'train/loss': 0.4740455448627472, 'train/grad_norm_pre_clip': 21.221696853637695, 'timing/data': 0.010125326924026012, 'timing/model': 0.9216989576816559, 'learning_rate/qwen_vl_interface': 3.101256583977905e-06, 'learning_rate/action_model': 3.101256583977905e-05, 'epoch': 1.89}) +[2026-06-17 23:49:31,887][starVLA.training.train_starvla][INFO] - Step 4582, Loss: {'train/loss': 0.23122933506965637, 'train/grad_norm_pre_clip': 20.45616340637207, 'timing/data': 0.011436828412115574, 'timing/model': 0.938513501547277, 'learning_rate/qwen_vl_interface': 3.099328050388131e-06, 'learning_rate/action_model': 3.0993280503881305e-05, 'epoch': 1.89}) +[2026-06-17 23:49:35,749][starVLA.training.train_starvla][INFO] - Step 4583, Loss: {'train/loss': 0.23435308039188385, 'train/grad_norm_pre_clip': 41.40105056762695, 'timing/data': 0.01281552854925394, 'timing/model': 1.1466788770630956, 'learning_rate/qwen_vl_interface': 3.0973999626349667e-06, 'learning_rate/action_model': 3.097399962634966e-05, 'epoch': 1.89}) +[2026-06-17 23:49:39,628][starVLA.training.train_starvla][INFO] - Step 4584, Loss: {'train/loss': 0.15811339020729065, 'train/grad_norm_pre_clip': 9.970351219177246, 'timing/data': 0.007742632180452347, 'timing/model': 1.1189808789640665, 'learning_rate/qwen_vl_interface': 3.095472321118107e-06, 'learning_rate/action_model': 3.095472321118107e-05, 'epoch': 1.89}) +[2026-06-17 23:49:43,520][starVLA.training.train_starvla][INFO] - Step 4585, Loss: {'train/loss': 0.39292779564857483, 'train/grad_norm_pre_clip': 20.554109573364258, 'timing/data': 0.010887696407735348, 'timing/model': 1.1843508528545499, 'learning_rate/qwen_vl_interface': 3.093545126237153e-06, 'learning_rate/action_model': 3.0935451262371527e-05, 'epoch': 1.89}) +[2026-06-17 23:49:47,293][starVLA.training.train_starvla][INFO] - Step 4586, Loss: {'train/loss': 0.06941182166337967, 'train/grad_norm_pre_clip': 6.259547233581543, 'timing/data': 0.01017003320157528, 'timing/model': 0.9290079176425934, 'learning_rate/qwen_vl_interface': 3.091618378391615e-06, 'learning_rate/action_model': 3.0916183783916146e-05, 'epoch': 1.89}) +[2026-06-17 23:49:51,123][starVLA.training.train_starvla][INFO] - Step 4587, Loss: {'train/loss': 0.21593670547008514, 'train/grad_norm_pre_clip': 16.522117614746094, 'timing/data': 0.009321572259068489, 'timing/model': 0.9934704583138227, 'learning_rate/qwen_vl_interface': 3.0896920779809097e-06, 'learning_rate/action_model': 3.08969207798091e-05, 'epoch': 1.89}) +[2026-06-17 23:49:55,018][starVLA.training.train_starvla][INFO] - Step 4588, Loss: {'train/loss': 0.4623616933822632, 'train/grad_norm_pre_clip': 23.2169246673584, 'timing/data': 0.009897916577756405, 'timing/model': 1.0302556157112122, 'learning_rate/qwen_vl_interface': 3.0877662254043605e-06, 'learning_rate/action_model': 3.0877662254043606e-05, 'epoch': 1.89}) +[2026-06-17 23:49:58,972][starVLA.training.train_starvla][INFO] - Step 4589, Loss: {'train/loss': 0.33975717425346375, 'train/grad_norm_pre_clip': 23.46380043029785, 'timing/data': 0.008965086191892624, 'timing/model': 0.9372094590216875, 'learning_rate/qwen_vl_interface': 3.0858408210612e-06, 'learning_rate/action_model': 3.0858408210612e-05, 'epoch': 1.89}) +[2026-06-17 23:50:02,824][starVLA.training.train_starvla][INFO] - Step 4590, Loss: {'train/loss': 0.3560471534729004, 'train/grad_norm_pre_clip': 23.281124114990234, 'timing/data': 0.014393413439393044, 'timing/model': 1.1222217744216323, 'learning_rate/qwen_vl_interface': 3.0839158653505663e-06, 'learning_rate/action_model': 3.0839158653505664e-05, 'epoch': 1.89}) +[2026-06-17 23:50:07,047][starVLA.training.train_starvla][INFO] - Step 4591, Loss: {'train/loss': 0.2134988009929657, 'train/grad_norm_pre_clip': 17.37604331970215, 'timing/data': 0.014040333218872547, 'timing/model': 1.133499481715262, 'learning_rate/qwen_vl_interface': 3.081991358671504e-06, 'learning_rate/action_model': 3.081991358671504e-05, 'epoch': 1.89}) +[2026-06-17 23:50:10,693][starVLA.training.train_starvla][INFO] - Step 4592, Loss: {'train/loss': 0.20786142349243164, 'train/grad_norm_pre_clip': 15.964111328125, 'timing/data': 0.008985172025859356, 'timing/model': 0.9425422986969352, 'learning_rate/qwen_vl_interface': 3.080067301422965e-06, 'learning_rate/action_model': 3.080067301422965e-05, 'epoch': 1.89}) +[2026-06-17 23:50:14,538][starVLA.training.train_starvla][INFO] - Step 4593, Loss: {'train/loss': 0.48673078417778015, 'train/grad_norm_pre_clip': 26.35270881652832, 'timing/data': 0.013810894452035427, 'timing/model': 0.967530844733119, 'learning_rate/qwen_vl_interface': 3.07814369400381e-06, 'learning_rate/action_model': 3.07814369400381e-05, 'epoch': 1.89}) +[2026-06-17 23:50:18,326][starVLA.training.train_starvla][INFO] - Step 4594, Loss: {'train/loss': 0.4279611110687256, 'train/grad_norm_pre_clip': 21.448211669921875, 'timing/data': 0.008205256424844265, 'timing/model': 0.9279298800975084, 'learning_rate/qwen_vl_interface': 3.076220536812807e-06, 'learning_rate/action_model': 3.076220536812807e-05, 'epoch': 1.89}) +[2026-06-17 23:50:22,215][starVLA.training.train_starvla][INFO] - Step 4595, Loss: {'train/loss': 0.7253374457359314, 'train/grad_norm_pre_clip': 28.340490341186523, 'timing/data': 0.011636224575340748, 'timing/model': 0.9342180239036679, 'learning_rate/qwen_vl_interface': 3.0742978302486214e-06, 'learning_rate/action_model': 3.074297830248621e-05, 'epoch': 1.89}) +[2026-06-17 23:50:25,982][starVLA.training.train_starvla][INFO] - Step 4596, Loss: {'train/loss': 0.5594919919967651, 'train/grad_norm_pre_clip': 25.786657333374023, 'timing/data': 0.018548433668911457, 'timing/model': 1.0943215573206544, 'learning_rate/qwen_vl_interface': 3.0723755747098387e-06, 'learning_rate/action_model': 3.072375574709839e-05, 'epoch': 1.89}) +[2026-06-17 23:50:29,669][starVLA.training.train_starvla][INFO] - Step 4597, Loss: {'train/loss': 0.34563103318214417, 'train/grad_norm_pre_clip': 16.087770462036133, 'timing/data': 0.009604505263268948, 'timing/model': 1.045347879640758, 'learning_rate/qwen_vl_interface': 3.070453770594943e-06, 'learning_rate/action_model': 3.070453770594943e-05, 'epoch': 1.89}) +[2026-06-17 23:50:33,284][starVLA.training.train_starvla][INFO] - Step 4598, Loss: {'train/loss': 0.2175063192844391, 'train/grad_norm_pre_clip': 17.154470443725586, 'timing/data': 0.012468754313886166, 'timing/model': 0.9375202190130949, 'learning_rate/qwen_vl_interface': 3.0685324183023264e-06, 'learning_rate/action_model': 3.068532418302326e-05, 'epoch': 1.89}) +[2026-06-17 23:50:37,109][starVLA.training.train_starvla][INFO] - Step 4599, Loss: {'train/loss': 0.22640670835971832, 'train/grad_norm_pre_clip': 14.563224792480469, 'timing/data': 0.010363460518419743, 'timing/model': 0.954732509329915, 'learning_rate/qwen_vl_interface': 3.0666115182302863e-06, 'learning_rate/action_model': 3.0666115182302865e-05, 'epoch': 1.89}) +[2026-06-17 23:52:40,345][starVLA.training.train_starvla][INFO] - Step 4600, Loss: {'train/loss': 0.6049095392227173, 'train/grad_norm_pre_clip': 25.721582412719727, 'eval/loss': 0.42858174443244934, 'eval/latency_6/loss': 0.42858174443244934, 'eval/demon_fire_precision': 0.9560929048679605, 'eval/demon_fire_recall': 0.9712346477052359, 'eval/demon_fire_f1': 0.9636042969376302, 'eval/demon_fire_rate_model': 0.8745130773511408, 'eval/demon_fire_rate_teacher': 0.860879243183083, 'eval/demon_left_precision': 0.8637316561844863, 'eval/demon_left_recall': 0.7228070175438597, 'eval/demon_left_f1': 0.7870105062082139, 'eval/demon_left_rate_model': 0.1327212020033389, 'eval/demon_left_rate_teacher': 0.15859766277128548, 'eval/demon_right_precision': 0.8286189683860233, 'eval/demon_right_recall': 0.7649769585253456, 'eval/demon_right_f1': 0.7955271565495208, 'eval/demon_right_rate_model': 0.16722314969393434, 'eval/demon_right_rate_teacher': 0.18113522537562604, 'eval/demon_move_f1': 0.7912688313788674, 'eval/demon_cc_f1': 0.9045049342094593, 'eval/latency_6/demon_fire_precision': 0.9560929048679605, 'eval/latency_6/demon_fire_recall': 0.9712346477052359, 'eval/latency_6/demon_fire_f1': 0.9636042969376302, 'eval/latency_6/demon_fire_rate_model': 0.8745130773511408, 'eval/latency_6/demon_fire_rate_teacher': 0.860879243183083, 'eval/latency_6/demon_left_precision': 0.8637316561844863, 'eval/latency_6/demon_left_recall': 0.7228070175438597, 'eval/latency_6/demon_left_f1': 0.7870105062082139, 'eval/latency_6/demon_left_rate_model': 0.1327212020033389, 'eval/latency_6/demon_left_rate_teacher': 0.15859766277128548, 'eval/latency_6/demon_right_precision': 0.8286189683860233, 'eval/latency_6/demon_right_recall': 0.7649769585253456, 'eval/latency_6/demon_right_f1': 0.7955271565495208, 'eval/latency_6/demon_right_rate_model': 0.16722314969393434, 'eval/latency_6/demon_right_rate_teacher': 0.18113522537562604, 'eval/latency_6/demon_move_f1': 0.7912688313788674, 'eval/latency_6/demon_cc_f1': 0.9045049342094593, 'timing/data': 0.00834457203745842, 'timing/model': 1.0032809432595968, 'learning_rate/qwen_vl_interface': 3.064691070777027e-06, 'learning_rate/action_model': 3.064691070777027e-05, 'epoch': 1.89}) +[2026-06-17 23:52:40,348][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_4600_state +[2026-06-17 23:52:40,353][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 23:52:47,319][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_4600_state/model.safetensors +[2026-06-17 23:53:03,526][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_4600_state/optimizer.bin +[2026-06-17 23:53:03,527][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_4600_state/sampler.bin +[2026-06-17 23:53:03,527][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_4600_state/sampler_1.bin +[2026-06-17 23:53:03,530][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_4600_state/random_states_0.pkl +[2026-06-17 23:53:03,533][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4600 +[2026-06-17 23:53:03,533][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 23:53:03,545][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 23:53:08,618][starVLA.training.train_starvla][INFO] - Step 4601, Loss: {'train/loss': 0.2837333679199219, 'train/grad_norm_pre_clip': 14.509146690368652, 'timing/data': 0.009484999813139439, 'timing/model': 0.9535155240446329, 'learning_rate/qwen_vl_interface': 3.062771076340662e-06, 'learning_rate/action_model': 3.0627710763406617e-05, 'epoch': 1.89}) +[2026-06-17 23:53:12,397][starVLA.training.train_starvla][INFO] - Step 4602, Loss: {'train/loss': 0.19487938284873962, 'train/grad_norm_pre_clip': 14.648566246032715, 'timing/data': 0.00816529430449009, 'timing/model': 0.9416165696457028, 'learning_rate/qwen_vl_interface': 3.0608515353192046e-06, 'learning_rate/action_model': 3.060851535319205e-05, 'epoch': 1.89}) +[2026-06-17 23:53:16,182][starVLA.training.train_starvla][INFO] - Step 4603, Loss: {'train/loss': 0.30970117449760437, 'train/grad_norm_pre_clip': 22.904319763183594, 'timing/data': 0.011360952630639076, 'timing/model': 0.935212753713131, 'learning_rate/qwen_vl_interface': 3.0589324481105814e-06, 'learning_rate/action_model': 3.058932448110581e-05, 'epoch': 1.89}) +[2026-06-17 23:53:19,941][starVLA.training.train_starvla][INFO] - Step 4604, Loss: {'train/loss': 0.35866060853004456, 'train/grad_norm_pre_clip': 25.008840560913086, 'timing/data': 0.010392015799880028, 'timing/model': 0.9347064467146993, 'learning_rate/qwen_vl_interface': 3.0570138151126185e-06, 'learning_rate/action_model': 3.057013815112618e-05, 'epoch': 1.9}) +[2026-06-17 23:53:23,733][starVLA.training.train_starvla][INFO] - Step 4605, Loss: {'train/loss': 0.4165240526199341, 'train/grad_norm_pre_clip': 24.1569766998291, 'timing/data': 0.008369388058781624, 'timing/model': 1.1061685429885983, 'learning_rate/qwen_vl_interface': 3.0550956367230506e-06, 'learning_rate/action_model': 3.05509563672305e-05, 'epoch': 1.9}) +[2026-06-17 23:53:27,344][starVLA.training.train_starvla][INFO] - Step 4606, Loss: {'train/loss': 0.44478923082351685, 'train/grad_norm_pre_clip': 23.29998207092285, 'timing/data': 0.008620593696832657, 'timing/model': 0.9488376509398222, 'learning_rate/qwen_vl_interface': 3.0531779133395184e-06, 'learning_rate/action_model': 3.053177913339518e-05, 'epoch': 1.9}) +[2026-06-17 23:53:31,154][starVLA.training.train_starvla][INFO] - Step 4607, Loss: {'train/loss': 0.4625365138053894, 'train/grad_norm_pre_clip': 19.867034912109375, 'timing/data': 0.009220435284078121, 'timing/model': 0.9299467345699668, 'learning_rate/qwen_vl_interface': 3.0512606453595708e-06, 'learning_rate/action_model': 3.0512606453595704e-05, 'epoch': 1.9}) +[2026-06-17 23:53:35,255][starVLA.training.train_starvla][INFO] - Step 4608, Loss: {'train/loss': 0.4492320716381073, 'train/grad_norm_pre_clip': 18.160945892333984, 'timing/data': 0.01276431791484356, 'timing/model': 1.2128173857927322, 'learning_rate/qwen_vl_interface': 3.049343833180658e-06, 'learning_rate/action_model': 3.049343833180658e-05, 'epoch': 1.9}) +[2026-06-17 23:53:39,139][starVLA.training.train_starvla][INFO] - Step 4609, Loss: {'train/loss': 0.266777366399765, 'train/grad_norm_pre_clip': 12.928265571594238, 'timing/data': 0.009390784427523613, 'timing/model': 0.947677880525589, 'learning_rate/qwen_vl_interface': 3.0474274772001355e-06, 'learning_rate/action_model': 3.0474274772001355e-05, 'epoch': 1.9}) +[2026-06-17 23:53:43,002][starVLA.training.train_starvla][INFO] - Step 4610, Loss: {'train/loss': 0.27034792304039, 'train/grad_norm_pre_clip': 19.985366821289062, 'timing/data': 0.009573956951498985, 'timing/model': 0.9556250469759107, 'learning_rate/qwen_vl_interface': 3.045511577815265e-06, 'learning_rate/action_model': 3.045511577815265e-05, 'epoch': 1.9}) +[2026-06-17 23:53:47,570][starVLA.training.train_starvla][INFO] - Step 4611, Loss: {'train/loss': 0.36057496070861816, 'train/grad_norm_pre_clip': 20.973867416381836, 'timing/data': 0.01191068347543478, 'timing/model': 0.9781982311978936, 'learning_rate/qwen_vl_interface': 3.043596135423218e-06, 'learning_rate/action_model': 3.043596135423218e-05, 'epoch': 1.9}) +[2026-06-17 23:53:51,581][starVLA.training.train_starvla][INFO] - Step 4612, Loss: {'train/loss': 0.2540125846862793, 'train/grad_norm_pre_clip': 15.224906921386719, 'timing/data': 0.010414442978799343, 'timing/model': 1.1917411414906383, 'learning_rate/qwen_vl_interface': 3.041681150421067e-06, 'learning_rate/action_model': 3.041681150421067e-05, 'epoch': 1.9}) +[2026-06-17 23:53:55,442][starVLA.training.train_starvla][INFO] - Step 4613, Loss: {'train/loss': 0.207697793841362, 'train/grad_norm_pre_clip': 16.42664909362793, 'timing/data': 0.016388706862926483, 'timing/model': 0.9591964371502399, 'learning_rate/qwen_vl_interface': 3.039766623205791e-06, 'learning_rate/action_model': 3.039766623205791e-05, 'epoch': 1.9}) +[2026-06-17 23:53:59,222][starVLA.training.train_starvla][INFO] - Step 4614, Loss: {'train/loss': 0.16697627305984497, 'train/grad_norm_pre_clip': 13.001941680908203, 'timing/data': 0.009336203336715698, 'timing/model': 0.9375544469803572, 'learning_rate/qwen_vl_interface': 3.037852554174269e-06, 'learning_rate/action_model': 3.0378525541742692e-05, 'epoch': 1.9}) +[2026-06-17 23:54:03,050][starVLA.training.train_starvla][INFO] - Step 4615, Loss: {'train/loss': 0.2165495902299881, 'train/grad_norm_pre_clip': 14.407657623291016, 'timing/data': 0.007455453276634216, 'timing/model': 0.9423106275498867, 'learning_rate/qwen_vl_interface': 3.035938943723295e-06, 'learning_rate/action_model': 3.035938943723295e-05, 'epoch': 1.9}) +[2026-06-17 23:54:06,918][starVLA.training.train_starvla][INFO] - Step 4616, Loss: {'train/loss': 0.22536763548851013, 'train/grad_norm_pre_clip': 17.833707809448242, 'timing/data': 0.010915161110460758, 'timing/model': 0.9488566238433123, 'learning_rate/qwen_vl_interface': 3.03402579224956e-06, 'learning_rate/action_model': 3.03402579224956e-05, 'epoch': 1.9}) +[2026-06-17 23:54:10,750][starVLA.training.train_starvla][INFO] - Step 4617, Loss: {'train/loss': 0.4240797162055969, 'train/grad_norm_pre_clip': 43.72776412963867, 'timing/data': 0.01070582214742899, 'timing/model': 0.9329635547474027, 'learning_rate/qwen_vl_interface': 3.032113100149663e-06, 'learning_rate/action_model': 3.0321131001496628e-05, 'epoch': 1.9}) +[2026-06-17 23:54:14,601][starVLA.training.train_starvla][INFO] - Step 4618, Loss: {'train/loss': 0.400190532207489, 'train/grad_norm_pre_clip': 24.78498077392578, 'timing/data': 0.010408503003418446, 'timing/model': 1.1095963791012764, 'learning_rate/qwen_vl_interface': 3.0302008678201067e-06, 'learning_rate/action_model': 3.0302008678201067e-05, 'epoch': 1.9}) +[2026-06-17 23:54:18,429][starVLA.training.train_starvla][INFO] - Step 4619, Loss: {'train/loss': 0.2524891495704651, 'train/grad_norm_pre_clip': 20.558244705200195, 'timing/data': 0.011609576642513275, 'timing/model': 1.0791835216805339, 'learning_rate/qwen_vl_interface': 3.028289095657299e-06, 'learning_rate/action_model': 3.028289095657299e-05, 'epoch': 1.9}) +[2026-06-17 23:54:22,292][starVLA.training.train_starvla][INFO] - Step 4620, Loss: {'train/loss': 0.13636766374111176, 'train/grad_norm_pre_clip': 11.239784240722656, 'timing/data': 0.013434197753667831, 'timing/model': 0.9573344048112631, 'learning_rate/qwen_vl_interface': 3.026377784057552e-06, 'learning_rate/action_model': 3.026377784057552e-05, 'epoch': 1.9}) +[2026-06-17 23:54:26,170][starVLA.training.train_starvla][INFO] - Step 4621, Loss: {'train/loss': 0.19577209651470184, 'train/grad_norm_pre_clip': 12.190489768981934, 'timing/data': 0.013786484487354755, 'timing/model': 0.9930386813357472, 'learning_rate/qwen_vl_interface': 3.0244669334170822e-06, 'learning_rate/action_model': 3.0244669334170822e-05, 'epoch': 1.9}) +[2026-06-17 23:54:30,091][starVLA.training.train_starvla][INFO] - Step 4622, Loss: {'train/loss': 0.44719070196151733, 'train/grad_norm_pre_clip': 19.47646141052246, 'timing/data': 0.009458078071475029, 'timing/model': 0.925486359745264, 'learning_rate/qwen_vl_interface': 3.0225565441320156e-06, 'learning_rate/action_model': 3.0225565441320152e-05, 'epoch': 1.9}) +[2026-06-17 23:54:34,066][starVLA.training.train_starvla][INFO] - Step 4623, Loss: {'train/loss': 0.2051517814397812, 'train/grad_norm_pre_clip': 11.110398292541504, 'timing/data': 0.010121949948370457, 'timing/model': 0.9298598105087876, 'learning_rate/qwen_vl_interface': 3.0206466165983716e-06, 'learning_rate/action_model': 3.0206466165983716e-05, 'epoch': 1.9}) +[2026-06-17 23:54:37,939][starVLA.training.train_starvla][INFO] - Step 4624, Loss: {'train/loss': 0.3315119743347168, 'train/grad_norm_pre_clip': 15.776808738708496, 'timing/data': 0.010324439033865929, 'timing/model': 0.9325220566242933, 'learning_rate/qwen_vl_interface': 3.0187371512120833e-06, 'learning_rate/action_model': 3.0187371512120832e-05, 'epoch': 1.9}) +[2026-06-17 23:54:41,742][starVLA.training.train_starvla][INFO] - Step 4625, Loss: {'train/loss': 0.42828369140625, 'train/grad_norm_pre_clip': 19.934236526489258, 'timing/data': 0.009631703607738018, 'timing/model': 1.1349623063579202, 'learning_rate/qwen_vl_interface': 3.0168281483689825e-06, 'learning_rate/action_model': 3.0168281483689825e-05, 'epoch': 1.9}) +[2026-06-17 23:54:45,355][starVLA.training.train_starvla][INFO] - Step 4626, Loss: {'train/loss': 0.4716860353946686, 'train/grad_norm_pre_clip': 20.93502426147461, 'timing/data': 0.011400317773222923, 'timing/model': 0.9523652493953705, 'learning_rate/qwen_vl_interface': 3.014919608464812e-06, 'learning_rate/action_model': 3.0149196084648117e-05, 'epoch': 1.9}) +[2026-06-17 23:54:51,276][starVLA.training.train_starvla][INFO] - Step 4627, Loss: {'train/loss': 0.27538788318634033, 'train/grad_norm_pre_clip': 14.009163856506348, 'timing/data': 0.009339725598692894, 'timing/model': 0.9371988950297236, 'learning_rate/qwen_vl_interface': 3.0130115318952124e-06, 'learning_rate/action_model': 3.0130115318952122e-05, 'epoch': 1.9}) +[2026-06-17 23:54:55,038][starVLA.training.train_starvla][INFO] - Step 4628, Loss: {'train/loss': 0.41757795214653015, 'train/grad_norm_pre_clip': 21.453617095947266, 'timing/data': 0.013059394434094429, 'timing/model': 0.9428514176979661, 'learning_rate/qwen_vl_interface': 3.011103919055726e-06, 'learning_rate/action_model': 3.011103919055726e-05, 'epoch': 1.91}) +[2026-06-17 23:54:59,010][starVLA.training.train_starvla][INFO] - Step 4629, Loss: {'train/loss': 0.4036954343318939, 'train/grad_norm_pre_clip': 20.129898071289062, 'timing/data': 0.008170672692358494, 'timing/model': 0.9336156183853745, 'learning_rate/qwen_vl_interface': 3.009196770341809e-06, 'learning_rate/action_model': 3.009196770341809e-05, 'epoch': 1.91}) +[2026-06-17 23:55:02,876][starVLA.training.train_starvla][INFO] - Step 4630, Loss: {'train/loss': 0.3528454005718231, 'train/grad_norm_pre_clip': 18.163238525390625, 'timing/data': 0.010122916661202908, 'timing/model': 1.0099703008309007, 'learning_rate/qwen_vl_interface': 3.0072900861488125e-06, 'learning_rate/action_model': 3.0072900861488125e-05, 'epoch': 1.91}) +[2026-06-17 23:55:06,728][starVLA.training.train_starvla][INFO] - Step 4631, Loss: {'train/loss': 0.383758544921875, 'train/grad_norm_pre_clip': 18.5214900970459, 'timing/data': 0.009411013685166836, 'timing/model': 1.1232592416927218, 'learning_rate/qwen_vl_interface': 3.0053838668719944e-06, 'learning_rate/action_model': 3.0053838668719943e-05, 'epoch': 1.91}) +[2026-06-17 23:55:10,495][starVLA.training.train_starvla][INFO] - Step 4632, Loss: {'train/loss': 0.45035320520401, 'train/grad_norm_pre_clip': 28.29570770263672, 'timing/data': 0.010485373437404633, 'timing/model': 1.077127780765295, 'learning_rate/qwen_vl_interface': 3.003478112906515e-06, 'learning_rate/action_model': 3.0034781129065148e-05, 'epoch': 1.91}) +[2026-06-17 23:55:14,197][starVLA.training.train_starvla][INFO] - Step 4633, Loss: {'train/loss': 0.21461091935634613, 'train/grad_norm_pre_clip': 20.62442970275879, 'timing/data': 0.01060923095792532, 'timing/model': 0.9343684073537588, 'learning_rate/qwen_vl_interface': 3.0015728246474405e-06, 'learning_rate/action_model': 3.0015728246474405e-05, 'epoch': 1.91}) +[2026-06-17 23:55:18,044][starVLA.training.train_starvla][INFO] - Step 4634, Loss: {'train/loss': 0.19549639523029327, 'train/grad_norm_pre_clip': 13.406899452209473, 'timing/data': 0.012609601952135563, 'timing/model': 0.9315825793892145, 'learning_rate/qwen_vl_interface': 2.99966800248974e-06, 'learning_rate/action_model': 2.99966800248974e-05, 'epoch': 1.91}) +[2026-06-17 23:55:21,993][starVLA.training.train_starvla][INFO] - Step 4635, Loss: {'train/loss': 0.5194672346115112, 'train/grad_norm_pre_clip': 20.422542572021484, 'timing/data': 0.011884194798767567, 'timing/model': 0.9321341188624501, 'learning_rate/qwen_vl_interface': 2.9977636468282817e-06, 'learning_rate/action_model': 2.9977636468282816e-05, 'epoch': 1.91}) +[2026-06-17 23:55:25,804][starVLA.training.train_starvla][INFO] - Step 4636, Loss: {'train/loss': 0.31525301933288574, 'train/grad_norm_pre_clip': 14.809085845947266, 'timing/data': 0.015727759338915348, 'timing/model': 0.9671186534687877, 'learning_rate/qwen_vl_interface': 2.995859758057846e-06, 'learning_rate/action_model': 2.9958597580578456e-05, 'epoch': 1.91}) +[2026-06-17 23:55:29,592][starVLA.training.train_starvla][INFO] - Step 4637, Loss: {'train/loss': 0.28134188055992126, 'train/grad_norm_pre_clip': 18.211963653564453, 'timing/data': 0.00883872527629137, 'timing/model': 0.9432197846472263, 'learning_rate/qwen_vl_interface': 2.993956336573106e-06, 'learning_rate/action_model': 2.9939563365731062e-05, 'epoch': 1.91}) +[2026-06-17 23:55:33,408][starVLA.training.train_starvla][INFO] - Step 4638, Loss: {'train/loss': 0.3021741807460785, 'train/grad_norm_pre_clip': 19.24922752380371, 'timing/data': 0.009111881256103516, 'timing/model': 1.108467404730618, 'learning_rate/qwen_vl_interface': 2.9920533827686453e-06, 'learning_rate/action_model': 2.9920533827686453e-05, 'epoch': 1.91}) +[2026-06-17 23:55:37,216][starVLA.training.train_starvla][INFO] - Step 4639, Loss: {'train/loss': 0.6516150236129761, 'train/grad_norm_pre_clip': 31.158998489379883, 'timing/data': 0.010021175257861614, 'timing/model': 1.1114234710112214, 'learning_rate/qwen_vl_interface': 2.9901508970389477e-06, 'learning_rate/action_model': 2.9901508970389476e-05, 'epoch': 1.91}) +[2026-06-17 23:55:41,298][starVLA.training.train_starvla][INFO] - Step 4640, Loss: {'train/loss': 0.3206617832183838, 'train/grad_norm_pre_clip': 16.9686336517334, 'timing/data': 0.009759046137332916, 'timing/model': 1.3194683892652392, 'learning_rate/qwen_vl_interface': 2.988248879778401e-06, 'learning_rate/action_model': 2.988248879778401e-05, 'epoch': 1.91}) +[2026-06-17 23:55:45,200][starVLA.training.train_starvla][INFO] - Step 4641, Loss: {'train/loss': 0.4245038628578186, 'train/grad_norm_pre_clip': 17.964435577392578, 'timing/data': 0.010376647114753723, 'timing/model': 1.013659936375916, 'learning_rate/qwen_vl_interface': 2.986347331381297e-06, 'learning_rate/action_model': 2.9863473313812967e-05, 'epoch': 1.91}) +[2026-06-17 23:55:50,407][starVLA.training.train_starvla][INFO] - Step 4642, Loss: {'train/loss': 0.4058648645877838, 'train/grad_norm_pre_clip': 18.54164695739746, 'timing/data': 0.01990350801497698, 'timing/model': 2.314119314774871, 'learning_rate/qwen_vl_interface': 2.984446252241826e-06, 'learning_rate/action_model': 2.984446252241826e-05, 'epoch': 1.91}) +[2026-06-17 23:55:54,355][starVLA.training.train_starvla][INFO] - Step 4643, Loss: {'train/loss': 0.2935277819633484, 'train/grad_norm_pre_clip': 17.30779457092285, 'timing/data': 0.008720115758478642, 'timing/model': 0.9883928596973419, 'learning_rate/qwen_vl_interface': 2.982545642754083e-06, 'learning_rate/action_model': 2.9825456427540826e-05, 'epoch': 1.91}) +[2026-06-17 23:55:58,188][starVLA.training.train_starvla][INFO] - Step 4644, Loss: {'train/loss': 0.053111378103494644, 'train/grad_norm_pre_clip': 5.556413173675537, 'timing/data': 0.01010794099420309, 'timing/model': 1.002797127701342, 'learning_rate/qwen_vl_interface': 2.9806455033120697e-06, 'learning_rate/action_model': 2.9806455033120696e-05, 'epoch': 1.91}) +[2026-06-17 23:56:02,010][starVLA.training.train_starvla][INFO] - Step 4645, Loss: {'train/loss': 0.1659373641014099, 'train/grad_norm_pre_clip': 18.252086639404297, 'timing/data': 0.0096648670732975, 'timing/model': 1.1183689832687378, 'learning_rate/qwen_vl_interface': 2.978745834309685e-06, 'learning_rate/action_model': 2.9787458343096846e-05, 'epoch': 1.91}) +[2026-06-17 23:56:05,765][starVLA.training.train_starvla][INFO] - Step 4646, Loss: {'train/loss': 0.25357669591903687, 'train/grad_norm_pre_clip': 14.75735855102539, 'timing/data': 0.009222883731126785, 'timing/model': 1.0618467088788748, 'learning_rate/qwen_vl_interface': 2.9768466361407335e-06, 'learning_rate/action_model': 2.9768466361407334e-05, 'epoch': 1.91}) +[2026-06-17 23:56:09,479][starVLA.training.train_starvla][INFO] - Step 4647, Loss: {'train/loss': 0.17841562628746033, 'train/grad_norm_pre_clip': 14.878305435180664, 'timing/data': 0.008390617556869984, 'timing/model': 0.9589125951752067, 'learning_rate/qwen_vl_interface': 2.97494790919892e-06, 'learning_rate/action_model': 2.9749479091989197e-05, 'epoch': 1.91}) +[2026-06-17 23:56:13,274][starVLA.training.train_starvla][INFO] - Step 4648, Loss: {'train/loss': 0.4087241291999817, 'train/grad_norm_pre_clip': 29.640274047851562, 'timing/data': 0.008441371843218803, 'timing/model': 0.9377719396725297, 'learning_rate/qwen_vl_interface': 2.9730496538778524e-06, 'learning_rate/action_model': 2.9730496538778523e-05, 'epoch': 1.91}) +[2026-06-17 23:56:17,145][starVLA.training.train_starvla][INFO] - Step 4649, Loss: {'train/loss': 0.4475482702255249, 'train/grad_norm_pre_clip': 20.037233352661133, 'timing/data': 0.008491354994475842, 'timing/model': 0.9419498322531581, 'learning_rate/qwen_vl_interface': 2.9711518705710422e-06, 'learning_rate/action_model': 2.9711518705710424e-05, 'epoch': 1.91}) +[2026-06-17 23:56:21,161][starVLA.training.train_starvla][INFO] - Step 4650, Loss: {'train/loss': 0.4497013986110687, 'train/grad_norm_pre_clip': 25.1475887298584, 'timing/data': 0.010229823179543018, 'timing/model': 1.0834971824660897, 'learning_rate/qwen_vl_interface': 2.9692545596719014e-06, 'learning_rate/action_model': 2.969254559671901e-05, 'epoch': 1.91}) +[2026-06-17 23:56:25,131][starVLA.training.train_starvla][INFO] - Step 4651, Loss: {'train/loss': 0.1184300035238266, 'train/grad_norm_pre_clip': 10.01531982421875, 'timing/data': 0.013667278923094273, 'timing/model': 1.2049714382737875, 'learning_rate/qwen_vl_interface': 2.9673577215737454e-06, 'learning_rate/action_model': 2.9673577215737452e-05, 'epoch': 1.91}) +[2026-06-17 23:56:29,027][starVLA.training.train_starvla][INFO] - Step 4652, Loss: {'train/loss': 0.2516331672668457, 'train/grad_norm_pre_clip': 16.779621124267578, 'timing/data': 0.012037331238389015, 'timing/model': 1.0555949779227376, 'learning_rate/qwen_vl_interface': 2.9654613566697886e-06, 'learning_rate/action_model': 2.9654613566697887e-05, 'epoch': 1.91}) +[2026-06-17 23:56:32,677][starVLA.training.train_starvla][INFO] - Step 4653, Loss: {'train/loss': 0.2646535634994507, 'train/grad_norm_pre_clip': 14.616905212402344, 'timing/data': 0.009031660854816437, 'timing/model': 0.935590511187911, 'learning_rate/qwen_vl_interface': 2.9635654653531516e-06, 'learning_rate/action_model': 2.9635654653531515e-05, 'epoch': 1.92}) +[2026-06-17 23:56:36,598][starVLA.training.train_starvla][INFO] - Step 4654, Loss: {'train/loss': 0.4502418041229248, 'train/grad_norm_pre_clip': 20.846601486206055, 'timing/data': 0.010068242438137531, 'timing/model': 1.019795279018581, 'learning_rate/qwen_vl_interface': 2.961670048016852e-06, 'learning_rate/action_model': 2.9616700480168518e-05, 'epoch': 1.92}) +[2026-06-17 23:56:41,736][starVLA.training.train_starvla][INFO] - Step 4655, Loss: {'train/loss': 0.14295609295368195, 'train/grad_norm_pre_clip': 10.410457611083984, 'timing/data': 0.01070341281592846, 'timing/model': 0.9951850874349475, 'learning_rate/qwen_vl_interface': 2.9597751050538177e-06, 'learning_rate/action_model': 2.9597751050538176e-05, 'epoch': 1.92}) +[2026-06-17 23:56:47,036][starVLA.training.train_starvla][INFO] - Step 4656, Loss: {'train/loss': 0.20621339976787567, 'train/grad_norm_pre_clip': 16.278493881225586, 'timing/data': 0.010357492603361607, 'timing/model': 0.9424011800438166, 'learning_rate/qwen_vl_interface': 2.957880636856866e-06, 'learning_rate/action_model': 2.9578806368568655e-05, 'epoch': 1.92}) +[2026-06-17 23:56:50,775][starVLA.training.train_starvla][INFO] - Step 4657, Loss: {'train/loss': 0.2903294265270233, 'train/grad_norm_pre_clip': 14.689338684082031, 'timing/data': 0.010956952348351479, 'timing/model': 0.9460503282025456, 'learning_rate/qwen_vl_interface': 2.955986643818723e-06, 'learning_rate/action_model': 2.9559866438187228e-05, 'epoch': 1.92}) +[2026-06-17 23:56:54,599][starVLA.training.train_starvla][INFO] - Step 4658, Loss: {'train/loss': 0.5594257116317749, 'train/grad_norm_pre_clip': 26.241945266723633, 'timing/data': 0.013564259745180607, 'timing/model': 1.1081320950761437, 'learning_rate/qwen_vl_interface': 2.954093126332019e-06, 'learning_rate/action_model': 2.9540931263320194e-05, 'epoch': 1.92}) +[2026-06-17 23:56:58,253][starVLA.training.train_starvla][INFO] - Step 4659, Loss: {'train/loss': 0.2715602219104767, 'train/grad_norm_pre_clip': 24.234390258789062, 'timing/data': 0.010777723044157028, 'timing/model': 0.9396377122029662, 'learning_rate/qwen_vl_interface': 2.9522000847892807e-06, 'learning_rate/action_model': 2.9522000847892807e-05, 'epoch': 1.92}) +[2026-06-17 23:57:02,208][starVLA.training.train_starvla][INFO] - Step 4660, Loss: {'train/loss': 0.43177953362464905, 'train/grad_norm_pre_clip': 24.940196990966797, 'timing/data': 0.011784193105995655, 'timing/model': 1.0143974432721734, 'learning_rate/qwen_vl_interface': 2.9503075195829393e-06, 'learning_rate/action_model': 2.9503075195829393e-05, 'epoch': 1.92}) +[2026-06-17 23:57:06,286][starVLA.training.train_starvla][INFO] - Step 4661, Loss: {'train/loss': 0.17840658128261566, 'train/grad_norm_pre_clip': 13.05589485168457, 'timing/data': 0.01292254775762558, 'timing/model': 0.9947603642940521, 'learning_rate/qwen_vl_interface': 2.9484154311053203e-06, 'learning_rate/action_model': 2.94841543110532e-05, 'epoch': 1.92}) +[2026-06-17 23:57:14,151][starVLA.training.train_starvla][INFO] - Step 4662, Loss: {'train/loss': 0.3436559736728668, 'train/grad_norm_pre_clip': 24.353012084960938, 'timing/data': 0.008267266675829887, 'timing/model': 1.1832852447405457, 'learning_rate/qwen_vl_interface': 2.94652381974866e-06, 'learning_rate/action_model': 2.94652381974866e-05, 'epoch': 1.92}) +[2026-06-17 23:57:18,358][starVLA.training.train_starvla][INFO] - Step 4663, Loss: {'train/loss': 0.2957974672317505, 'train/grad_norm_pre_clip': 19.403654098510742, 'timing/data': 0.010777906514704227, 'timing/model': 1.0911760916933417, 'learning_rate/qwen_vl_interface': 2.9446326859050916e-06, 'learning_rate/action_model': 2.9446326859050916e-05, 'epoch': 1.92}) +[2026-06-17 23:57:22,514][starVLA.training.train_starvla][INFO] - Step 4664, Loss: {'train/loss': 0.16197900474071503, 'train/grad_norm_pre_clip': 13.691068649291992, 'timing/data': 0.011090201325714588, 'timing/model': 1.2231026599183679, 'learning_rate/qwen_vl_interface': 2.9427420299666472e-06, 'learning_rate/action_model': 2.942742029966647e-05, 'epoch': 1.92}) +[2026-06-17 23:57:27,481][starVLA.training.train_starvla][INFO] - Step 4665, Loss: {'train/loss': 0.33292534947395325, 'train/grad_norm_pre_clip': 22.334569931030273, 'timing/data': 0.014405439607799053, 'timing/model': 0.922751747071743, 'learning_rate/qwen_vl_interface': 2.940851852325263e-06, 'learning_rate/action_model': 2.9408518523252632e-05, 'epoch': 1.92}) +[2026-06-17 23:57:31,510][starVLA.training.train_starvla][INFO] - Step 4666, Loss: {'train/loss': 0.32587355375289917, 'train/grad_norm_pre_clip': 17.936479568481445, 'timing/data': 0.009166249074041843, 'timing/model': 1.0765565503388643, 'learning_rate/qwen_vl_interface': 2.9389621533727745e-06, 'learning_rate/action_model': 2.9389621533727747e-05, 'epoch': 1.92}) +[2026-06-17 23:57:37,177][starVLA.training.train_starvla][INFO] - Step 4667, Loss: {'train/loss': 0.17894110083580017, 'train/grad_norm_pre_clip': 14.43725872039795, 'timing/data': 0.00996050238609314, 'timing/model': 0.9485780745744705, 'learning_rate/qwen_vl_interface': 2.9370729335009185e-06, 'learning_rate/action_model': 2.9370729335009184e-05, 'epoch': 1.92}) +[2026-06-17 23:57:41,059][starVLA.training.train_starvla][INFO] - Step 4668, Loss: {'train/loss': 0.22812579572200775, 'train/grad_norm_pre_clip': 18.647037506103516, 'timing/data': 0.009494363330304623, 'timing/model': 0.9299601102247834, 'learning_rate/qwen_vl_interface': 2.93518419310133e-06, 'learning_rate/action_model': 2.93518419310133e-05, 'epoch': 1.92}) +[2026-06-17 23:57:44,934][starVLA.training.train_starvla][INFO] - Step 4669, Loss: {'train/loss': 0.48610371351242065, 'train/grad_norm_pre_clip': 19.62070655822754, 'timing/data': 0.008496730588376522, 'timing/model': 0.9352532615885139, 'learning_rate/qwen_vl_interface': 2.9332959325655537e-06, 'learning_rate/action_model': 2.9332959325655535e-05, 'epoch': 1.92}) +[2026-06-17 23:57:48,738][starVLA.training.train_starvla][INFO] - Step 4670, Loss: {'train/loss': 0.4928476810455322, 'train/grad_norm_pre_clip': 26.094881057739258, 'timing/data': 0.012097581289708614, 'timing/model': 0.9438977800309658, 'learning_rate/qwen_vl_interface': 2.9314081522850206e-06, 'learning_rate/action_model': 2.9314081522850205e-05, 'epoch': 1.92}) +[2026-06-17 23:57:52,755][starVLA.training.train_starvla][INFO] - Step 4671, Loss: {'train/loss': 0.4269036650657654, 'train/grad_norm_pre_clip': 22.88650131225586, 'timing/data': 0.009936188347637653, 'timing/model': 1.2763292584568262, 'learning_rate/qwen_vl_interface': 2.9295208526510726e-06, 'learning_rate/action_model': 2.9295208526510725e-05, 'epoch': 1.92}) +[2026-06-17 23:57:56,415][starVLA.training.train_starvla][INFO] - Step 4672, Loss: {'train/loss': 0.4645666182041168, 'train/grad_norm_pre_clip': 27.821338653564453, 'timing/data': 0.010078780353069305, 'timing/model': 0.9403215041384101, 'learning_rate/qwen_vl_interface': 2.9276340340549476e-06, 'learning_rate/action_model': 2.9276340340549473e-05, 'epoch': 1.92}) +[2026-06-17 23:58:00,221][starVLA.training.train_starvla][INFO] - Step 4673, Loss: {'train/loss': 0.6416792273521423, 'train/grad_norm_pre_clip': 28.602584838867188, 'timing/data': 0.011635356582701206, 'timing/model': 0.9432756751775742, 'learning_rate/qwen_vl_interface': 2.9257476968877875e-06, 'learning_rate/action_model': 2.9257476968877874e-05, 'epoch': 1.92}) +[2026-06-17 23:58:04,022][starVLA.training.train_starvla][INFO] - Step 4674, Loss: {'train/loss': 0.2646118104457855, 'train/grad_norm_pre_clip': 17.62158966064453, 'timing/data': 0.007147090509533882, 'timing/model': 0.9349736478179693, 'learning_rate/qwen_vl_interface': 2.923861841540631e-06, 'learning_rate/action_model': 2.923861841540631e-05, 'epoch': 1.92}) +[2026-06-17 23:58:07,907][starVLA.training.train_starvla][INFO] - Step 4675, Loss: {'train/loss': 0.5292564630508423, 'train/grad_norm_pre_clip': 21.5098876953125, 'timing/data': 0.010528475046157837, 'timing/model': 0.9372251546010375, 'learning_rate/qwen_vl_interface': 2.9219764684044196e-06, 'learning_rate/action_model': 2.9219764684044193e-05, 'epoch': 1.92}) +[2026-06-17 23:58:11,950][starVLA.training.train_starvla][INFO] - Step 4676, Loss: {'train/loss': 0.48185786604881287, 'train/grad_norm_pre_clip': 20.70829963684082, 'timing/data': 0.016688121482729912, 'timing/model': 1.0753703890368342, 'learning_rate/qwen_vl_interface': 2.920091577869988e-06, 'learning_rate/action_model': 2.9200915778699877e-05, 'epoch': 1.92}) +[2026-06-17 23:58:16,288][starVLA.training.train_starvla][INFO] - Step 4677, Loss: {'train/loss': 0.42153042554855347, 'train/grad_norm_pre_clip': 34.0714225769043, 'timing/data': 0.008384043350815773, 'timing/model': 1.2637331569567323, 'learning_rate/qwen_vl_interface': 2.918207170328081e-06, 'learning_rate/action_model': 2.918207170328081e-05, 'epoch': 1.93}) +[2026-06-17 23:58:22,401][starVLA.training.train_starvla][INFO] - Step 4678, Loss: {'train/loss': 0.40026527643203735, 'train/grad_norm_pre_clip': 24.07866668701172, 'timing/data': 0.010020822286605835, 'timing/model': 1.2459617042914033, 'learning_rate/qwen_vl_interface': 2.9163232461693373e-06, 'learning_rate/action_model': 2.9163232461693373e-05, 'epoch': 1.93}) +[2026-06-17 23:58:26,918][starVLA.training.train_starvla][INFO] - Step 4679, Loss: {'train/loss': 0.18215259909629822, 'train/grad_norm_pre_clip': 10.589676856994629, 'timing/data': 0.010009546764194965, 'timing/model': 0.9446355318650603, 'learning_rate/qwen_vl_interface': 2.9144398057842948e-06, 'learning_rate/action_model': 2.9144398057842948e-05, 'epoch': 1.93}) +[2026-06-17 23:58:30,741][starVLA.training.train_starvla][INFO] - Step 4680, Loss: {'train/loss': 0.2830045521259308, 'train/grad_norm_pre_clip': 16.841632843017578, 'timing/data': 0.012180449441075325, 'timing/model': 0.935389868915081, 'learning_rate/qwen_vl_interface': 2.912556849563394e-06, 'learning_rate/action_model': 2.9125568495633943e-05, 'epoch': 1.93}) +[2026-06-17 23:58:34,566][starVLA.training.train_starvla][INFO] - Step 4681, Loss: {'train/loss': 0.32275164127349854, 'train/grad_norm_pre_clip': 18.51831817626953, 'timing/data': 0.012993067502975464, 'timing/model': 0.9359928406774998, 'learning_rate/qwen_vl_interface': 2.910674377896974e-06, 'learning_rate/action_model': 2.910674377896974e-05, 'epoch': 1.93}) +[2026-06-17 23:58:38,473][starVLA.training.train_starvla][INFO] - Step 4682, Loss: {'train/loss': 0.4207262098789215, 'train/grad_norm_pre_clip': 20.589981079101562, 'timing/data': 0.008731663227081299, 'timing/model': 0.9308936586603522, 'learning_rate/qwen_vl_interface': 2.9087923911752723e-06, 'learning_rate/action_model': 2.9087923911752725e-05, 'epoch': 1.93}) +[2026-06-17 23:58:42,321][starVLA.training.train_starvla][INFO] - Step 4683, Loss: {'train/loss': 0.1382688283920288, 'train/grad_norm_pre_clip': 13.653253555297852, 'timing/data': 0.009005510248243809, 'timing/model': 0.942395213060081, 'learning_rate/qwen_vl_interface': 2.906910889788425e-06, 'learning_rate/action_model': 2.906910889788425e-05, 'epoch': 1.93}) +[2026-06-17 23:58:46,080][starVLA.training.train_starvla][INFO] - Step 4684, Loss: {'train/loss': 0.3443331718444824, 'train/grad_norm_pre_clip': 20.420021057128906, 'timing/data': 0.007892563007771969, 'timing/model': 1.1058323197066784, 'learning_rate/qwen_vl_interface': 2.905029874126476e-06, 'learning_rate/action_model': 2.905029874126476e-05, 'epoch': 1.93}) +[2026-06-17 23:58:49,820][starVLA.training.train_starvla][INFO] - Step 4685, Loss: {'train/loss': 0.6750425100326538, 'train/grad_norm_pre_clip': 27.008230209350586, 'timing/data': 0.008934021927416325, 'timing/model': 1.0721753342077136, 'learning_rate/qwen_vl_interface': 2.9031493445793558e-06, 'learning_rate/action_model': 2.9031493445793558e-05, 'epoch': 1.93}) +[2026-06-17 23:58:53,534][starVLA.training.train_starvla][INFO] - Step 4686, Loss: {'train/loss': 0.22405515611171722, 'train/grad_norm_pre_clip': 20.190017700195312, 'timing/data': 0.008994583040475845, 'timing/model': 0.989830800332129, 'learning_rate/qwen_vl_interface': 2.9012693015369007e-06, 'learning_rate/action_model': 2.9012693015369008e-05, 'epoch': 1.93}) +[2026-06-17 23:58:57,429][starVLA.training.train_starvla][INFO] - Step 4687, Loss: {'train/loss': 0.3427239656448364, 'train/grad_norm_pre_clip': 18.999404907226562, 'timing/data': 0.011518397368490696, 'timing/model': 0.9435670739039779, 'learning_rate/qwen_vl_interface': 2.8993897453888454e-06, 'learning_rate/action_model': 2.899389745388845e-05, 'epoch': 1.93}) +[2026-06-17 23:59:01,707][starVLA.training.train_starvla][INFO] - Step 4688, Loss: {'train/loss': 0.32116660475730896, 'train/grad_norm_pre_clip': 17.257190704345703, 'timing/data': 0.014614070765674114, 'timing/model': 1.1429889602586627, 'learning_rate/qwen_vl_interface': 2.8975106765248267e-06, 'learning_rate/action_model': 2.8975106765248267e-05, 'epoch': 1.93}) +[2026-06-17 23:59:05,559][starVLA.training.train_starvla][INFO] - Step 4689, Loss: {'train/loss': 0.35151317715644836, 'train/grad_norm_pre_clip': 18.99091339111328, 'timing/data': 0.00808771513402462, 'timing/model': 0.9635874917730689, 'learning_rate/qwen_vl_interface': 2.895632095334378e-06, 'learning_rate/action_model': 2.895632095334378e-05, 'epoch': 1.93}) +[2026-06-17 23:59:09,983][starVLA.training.train_starvla][INFO] - Step 4690, Loss: {'train/loss': 0.09272250533103943, 'train/grad_norm_pre_clip': 8.012937545776367, 'timing/data': 0.010154558345675468, 'timing/model': 1.0025328109040856, 'learning_rate/qwen_vl_interface': 2.8937540022069267e-06, 'learning_rate/action_model': 2.8937540022069265e-05, 'epoch': 1.93}) +[2026-06-17 23:59:13,844][starVLA.training.train_starvla][INFO] - Step 4691, Loss: {'train/loss': 0.30719518661499023, 'train/grad_norm_pre_clip': 14.433507919311523, 'timing/data': 0.011230246163904667, 'timing/model': 1.1092718485742807, 'learning_rate/qwen_vl_interface': 2.8918763975318076e-06, 'learning_rate/action_model': 2.8918763975318076e-05, 'epoch': 1.93}) +[2026-06-17 23:59:17,938][starVLA.training.train_starvla][INFO] - Step 4692, Loss: {'train/loss': 0.5680820345878601, 'train/grad_norm_pre_clip': 29.751962661743164, 'timing/data': 0.018210772424936295, 'timing/model': 0.9495121473446488, 'learning_rate/qwen_vl_interface': 2.889999281698248e-06, 'learning_rate/action_model': 2.889999281698248e-05, 'epoch': 1.93}) +[2026-06-17 23:59:21,974][starVLA.training.train_starvla][INFO] - Step 4693, Loss: {'train/loss': 0.19867150485515594, 'train/grad_norm_pre_clip': 15.4808349609375, 'timing/data': 0.010327468626201153, 'timing/model': 0.941399808973074, 'learning_rate/qwen_vl_interface': 2.8881226550953772e-06, 'learning_rate/action_model': 2.888122655095377e-05, 'epoch': 1.93}) +[2026-06-17 23:59:25,772][starVLA.training.train_starvla][INFO] - Step 4694, Loss: {'train/loss': 0.24516236782073975, 'train/grad_norm_pre_clip': 15.75011157989502, 'timing/data': 0.01427455898374319, 'timing/model': 0.9492003237828612, 'learning_rate/qwen_vl_interface': 2.8862465181122212e-06, 'learning_rate/action_model': 2.886246518112221e-05, 'epoch': 1.93}) +[2026-06-17 23:59:29,531][starVLA.training.train_starvla][INFO] - Step 4695, Loss: {'train/loss': 0.15349610149860382, 'train/grad_norm_pre_clip': 13.218103408813477, 'timing/data': 0.011734656058251858, 'timing/model': 0.9361946126446128, 'learning_rate/qwen_vl_interface': 2.884370871137706e-06, 'learning_rate/action_model': 2.884370871137706e-05, 'epoch': 1.93}) +[2026-06-17 23:59:33,397][starVLA.training.train_starvla][INFO] - Step 4696, Loss: {'train/loss': 0.26956963539123535, 'train/grad_norm_pre_clip': 22.471553802490234, 'timing/data': 0.007222577929496765, 'timing/model': 1.0335089759901166, 'learning_rate/qwen_vl_interface': 2.8824957145606548e-06, 'learning_rate/action_model': 2.8824957145606548e-05, 'epoch': 1.93}) +[2026-06-17 23:59:39,362][starVLA.training.train_starvla][INFO] - Step 4697, Loss: {'train/loss': 0.24385327100753784, 'train/grad_norm_pre_clip': 21.72149658203125, 'timing/data': 0.011157502420246601, 'timing/model': 2.0853198440745473, 'learning_rate/qwen_vl_interface': 2.8806210487697866e-06, 'learning_rate/action_model': 2.8806210487697867e-05, 'epoch': 1.93}) +[2026-06-17 23:59:43,350][starVLA.training.train_starvla][INFO] - Step 4698, Loss: {'train/loss': 0.36536315083503723, 'train/grad_norm_pre_clip': 21.702396392822266, 'timing/data': 0.009351784363389015, 'timing/model': 1.1041881013661623, 'learning_rate/qwen_vl_interface': 2.87874687415373e-06, 'learning_rate/action_model': 2.87874687415373e-05, 'epoch': 1.93}) +[2026-06-17 23:59:47,088][starVLA.training.train_starvla][INFO] - Step 4699, Loss: {'train/loss': 0.4239869713783264, 'train/grad_norm_pre_clip': 20.6095027923584, 'timing/data': 0.010003174655139446, 'timing/model': 0.9483399204909801, 'learning_rate/qwen_vl_interface': 2.876873191100995e-06, 'learning_rate/action_model': 2.8768731911009948e-05, 'epoch': 1.93}) +[2026-06-18 00:01:49,367][starVLA.training.train_starvla][INFO] - Step 4700, Loss: {'train/loss': 0.31864336133003235, 'train/grad_norm_pre_clip': 24.2591552734375, 'eval/loss': 0.4275997281074524, 'eval/latency_6/loss': 0.4275997281074524, 'eval/demon_fire_precision': 0.9509433962264151, 'eval/demon_fire_recall': 0.9773755656108597, 'eval/demon_fire_f1': 0.9639783232387632, 'eval/demon_fire_rate_model': 0.8848080133555927, 'eval/demon_fire_rate_teacher': 0.860879243183083, 'eval/demon_left_precision': 0.8336594911937377, 'eval/demon_left_recall': 0.7473684210526316, 'eval/demon_left_f1': 0.7881591119333949, 'eval/demon_left_rate_model': 0.14218141346688926, 'eval/demon_left_rate_teacher': 0.15859766277128548, 'eval/demon_right_precision': 0.8333333333333334, 'eval/demon_right_recall': 0.7603686635944701, 'eval/demon_right_f1': 0.7951807228915664, 'eval/demon_right_rate_model': 0.1652754590984975, 'eval/demon_right_rate_teacher': 0.18113522537562604, 'eval/demon_move_f1': 0.7916699174124806, 'eval/demon_cc_f1': 0.9048928480922539, 'eval/latency_6/demon_fire_precision': 0.9509433962264151, 'eval/latency_6/demon_fire_recall': 0.9773755656108597, 'eval/latency_6/demon_fire_f1': 0.9639783232387632, 'eval/latency_6/demon_fire_rate_model': 0.8848080133555927, 'eval/latency_6/demon_fire_rate_teacher': 0.860879243183083, 'eval/latency_6/demon_left_precision': 0.8336594911937377, 'eval/latency_6/demon_left_recall': 0.7473684210526316, 'eval/latency_6/demon_left_f1': 0.7881591119333949, 'eval/latency_6/demon_left_rate_model': 0.14218141346688926, 'eval/latency_6/demon_left_rate_teacher': 0.15859766277128548, 'eval/latency_6/demon_right_precision': 0.8333333333333334, 'eval/latency_6/demon_right_recall': 0.7603686635944701, 'eval/latency_6/demon_right_f1': 0.7951807228915664, 'eval/latency_6/demon_right_rate_model': 0.1652754590984975, 'eval/latency_6/demon_right_rate_teacher': 0.18113522537562604, 'eval/latency_6/demon_move_f1': 0.7916699174124806, 'eval/latency_6/demon_cc_f1': 0.9048928480922539, 'timing/data': 0.011244633235037327, 'timing/model': 0.9460603976622224, 'learning_rate/qwen_vl_interface': 2.8750000000000013e-06, 'learning_rate/action_model': 2.875000000000001e-05, 'epoch': 1.93}) +[2026-06-18 00:01:49,451][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_4700_state +[2026-06-18 00:01:49,468][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 00:01:56,237][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_4700_state/model.safetensors +[2026-06-18 00:02:06,746][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_4700_state/optimizer.bin +[2026-06-18 00:02:06,748][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_4700_state/sampler.bin +[2026-06-18 00:02:06,748][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_4700_state/sampler_1.bin +[2026-06-18 00:02:06,750][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_4700_state/random_states_0.pkl +[2026-06-18 00:02:06,754][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4700 +[2026-06-18 00:02:06,756][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 00:02:06,769][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 00:02:10,549][starVLA.training.train_starvla][INFO] - Step 4701, Loss: {'train/loss': 0.2838987708091736, 'train/grad_norm_pre_clip': 18.021116256713867, 'timing/data': 0.009256741963326931, 'timing/model': 1.0257262643426657, 'learning_rate/qwen_vl_interface': 2.8731273012390614e-06, 'learning_rate/action_model': 2.8731273012390615e-05, 'epoch': 1.94}) +[2026-06-18 00:02:14,502][starVLA.training.train_starvla][INFO] - Step 4702, Loss: {'train/loss': 0.4174768626689911, 'train/grad_norm_pre_clip': 20.053863525390625, 'timing/data': 0.01404742244631052, 'timing/model': 0.9998049912974238, 'learning_rate/qwen_vl_interface': 2.8712550952063912e-06, 'learning_rate/action_model': 2.871255095206391e-05, 'epoch': 1.94}) +[2026-06-18 00:02:18,543][starVLA.training.train_starvla][INFO] - Step 4703, Loss: {'train/loss': 0.42506906390190125, 'train/grad_norm_pre_clip': 31.333995819091797, 'timing/data': 0.007975144311785698, 'timing/model': 0.9494937751442194, 'learning_rate/qwen_vl_interface': 2.8693833822901012e-06, 'learning_rate/action_model': 2.8693833822901012e-05, 'epoch': 1.94}) +[2026-06-18 00:02:29,510][starVLA.training.train_starvla][INFO] - Step 4704, Loss: {'train/loss': 0.1119377538561821, 'train/grad_norm_pre_clip': 9.394465446472168, 'timing/data': 0.0153533099219203, 'timing/model': 4.028876123018563, 'learning_rate/qwen_vl_interface': 2.867512162878196e-06, 'learning_rate/action_model': 2.867512162878196e-05, 'epoch': 1.94}) +[2026-06-18 00:02:34,314][starVLA.training.train_starvla][INFO] - Step 4705, Loss: {'train/loss': 0.3505268394947052, 'train/grad_norm_pre_clip': 16.704208374023438, 'timing/data': 0.009583772160112858, 'timing/model': 1.0069246916100383, 'learning_rate/qwen_vl_interface': 2.8656414373585812e-06, 'learning_rate/action_model': 2.865641437358581e-05, 'epoch': 1.94}) +[2026-06-18 00:02:38,164][starVLA.training.train_starvla][INFO] - Step 4706, Loss: {'train/loss': 0.320498526096344, 'train/grad_norm_pre_clip': 22.97081756591797, 'timing/data': 0.008785317651927471, 'timing/model': 1.156879828311503, 'learning_rate/qwen_vl_interface': 2.8637712061190644e-06, 'learning_rate/action_model': 2.863771206119064e-05, 'epoch': 1.94}) +[2026-06-18 00:02:41,842][starVLA.training.train_starvla][INFO] - Step 4707, Loss: {'train/loss': 0.22605818510055542, 'train/grad_norm_pre_clip': 16.498018264770508, 'timing/data': 0.008218626491725445, 'timing/model': 0.9354961095377803, 'learning_rate/qwen_vl_interface': 2.861901469547344e-06, 'learning_rate/action_model': 2.861901469547344e-05, 'epoch': 1.94}) +[2026-06-18 00:02:45,659][starVLA.training.train_starvla][INFO] - Step 4708, Loss: {'train/loss': 0.4267886281013489, 'train/grad_norm_pre_clip': 23.55173110961914, 'timing/data': 0.012752826325595379, 'timing/model': 0.9211957436054945, 'learning_rate/qwen_vl_interface': 2.8600322280310185e-06, 'learning_rate/action_model': 2.8600322280310183e-05, 'epoch': 1.94}) +[2026-06-18 00:02:49,693][starVLA.training.train_starvla][INFO] - Step 4709, Loss: {'train/loss': 0.13983377814292908, 'train/grad_norm_pre_clip': 24.92067527770996, 'timing/data': 0.011777423322200775, 'timing/model': 1.0548949083313346, 'learning_rate/qwen_vl_interface': 2.8581634819575843e-06, 'learning_rate/action_model': 2.8581634819575843e-05, 'epoch': 1.94}) +[2026-06-18 00:02:53,448][starVLA.training.train_starvla][INFO] - Step 4710, Loss: {'train/loss': 0.42905867099761963, 'train/grad_norm_pre_clip': 24.05365562438965, 'timing/data': 0.010187739506363869, 'timing/model': 0.9457405591383576, 'learning_rate/qwen_vl_interface': 2.856295231714434e-06, 'learning_rate/action_model': 2.8562952317144337e-05, 'epoch': 1.94}) +[2026-06-18 00:02:57,278][starVLA.training.train_starvla][INFO] - Step 4711, Loss: {'train/loss': 0.34158220887184143, 'train/grad_norm_pre_clip': 21.879728317260742, 'timing/data': 0.00930449366569519, 'timing/model': 0.9428953547030687, 'learning_rate/qwen_vl_interface': 2.854427477688857e-06, 'learning_rate/action_model': 2.854427477688857e-05, 'epoch': 1.94}) +[2026-06-18 00:03:01,133][starVLA.training.train_starvla][INFO] - Step 4712, Loss: {'train/loss': 0.33905261754989624, 'train/grad_norm_pre_clip': 20.68967628479004, 'timing/data': 0.01095657143741846, 'timing/model': 1.119228401221335, 'learning_rate/qwen_vl_interface': 2.8525602202680424e-06, 'learning_rate/action_model': 2.8525602202680424e-05, 'epoch': 1.94}) +[2026-06-18 00:03:05,091][starVLA.training.train_starvla][INFO] - Step 4713, Loss: {'train/loss': 0.366373747587204, 'train/grad_norm_pre_clip': 21.588214874267578, 'timing/data': 0.009194831363856792, 'timing/model': 1.0764007149264216, 'learning_rate/qwen_vl_interface': 2.8506934598390736e-06, 'learning_rate/action_model': 2.8506934598390735e-05, 'epoch': 1.94}) +[2026-06-18 00:03:08,948][starVLA.training.train_starvla][INFO] - Step 4714, Loss: {'train/loss': 0.2869592308998108, 'train/grad_norm_pre_clip': 17.933122634887695, 'timing/data': 0.01071508601307869, 'timing/model': 0.9522510627284646, 'learning_rate/qwen_vl_interface': 2.848827196788932e-06, 'learning_rate/action_model': 2.848827196788932e-05, 'epoch': 1.94}) +[2026-06-18 00:03:12,783][starVLA.training.train_starvla][INFO] - Step 4715, Loss: {'train/loss': 0.2024872750043869, 'train/grad_norm_pre_clip': 14.451237678527832, 'timing/data': 0.009219626896083355, 'timing/model': 0.9528554165735841, 'learning_rate/qwen_vl_interface': 2.8469614315044975e-06, 'learning_rate/action_model': 2.8469614315044975e-05, 'epoch': 1.94}) +[2026-06-18 00:03:16,623][starVLA.training.train_starvla][INFO] - Step 4716, Loss: {'train/loss': 0.20665878057479858, 'train/grad_norm_pre_clip': 19.895286560058594, 'timing/data': 0.011185236275196075, 'timing/model': 0.9338561007753015, 'learning_rate/qwen_vl_interface': 2.8450961643725405e-06, 'learning_rate/action_model': 2.8450961643725405e-05, 'epoch': 1.94}) +[2026-06-18 00:03:20,395][starVLA.training.train_starvla][INFO] - Step 4717, Loss: {'train/loss': 0.42921188473701477, 'train/grad_norm_pre_clip': 25.515518188476562, 'timing/data': 0.007866467349231243, 'timing/model': 0.9297981206327677, 'learning_rate/qwen_vl_interface': 2.8432313957797415e-06, 'learning_rate/action_model': 2.8432313957797414e-05, 'epoch': 1.94}) +[2026-06-18 00:03:24,166][starVLA.training.train_starvla][INFO] - Step 4718, Loss: {'train/loss': 0.22612729668617249, 'train/grad_norm_pre_clip': 13.891596794128418, 'timing/data': 0.010119901970028877, 'timing/model': 0.9354070369154215, 'learning_rate/qwen_vl_interface': 2.8413671261126613e-06, 'learning_rate/action_model': 2.8413671261126613e-05, 'epoch': 1.94}) +[2026-06-18 00:03:28,572][starVLA.training.train_starvla][INFO] - Step 4719, Loss: {'train/loss': 0.23120325803756714, 'train/grad_norm_pre_clip': 18.54181671142578, 'timing/data': 0.00938668567687273, 'timing/model': 1.6288490062579513, 'learning_rate/qwen_vl_interface': 2.8395033557577668e-06, 'learning_rate/action_model': 2.8395033557577668e-05, 'epoch': 1.94}) +[2026-06-18 00:03:32,145][starVLA.training.train_starvla][INFO] - Step 4720, Loss: {'train/loss': 0.40783488750457764, 'train/grad_norm_pre_clip': 28.17211151123047, 'timing/data': 0.007163059897720814, 'timing/model': 0.9250370897352695, 'learning_rate/qwen_vl_interface': 2.8376400851014214e-06, 'learning_rate/action_model': 2.8376400851014214e-05, 'epoch': 1.94}) +[2026-06-18 00:03:36,006][starVLA.training.train_starvla][INFO] - Step 4721, Loss: {'train/loss': 0.5059443712234497, 'train/grad_norm_pre_clip': 28.570322036743164, 'timing/data': 0.007674713619053364, 'timing/model': 1.021762507967651, 'learning_rate/qwen_vl_interface': 2.8357773145298823e-06, 'learning_rate/action_model': 2.8357773145298823e-05, 'epoch': 1.94}) +[2026-06-18 00:03:40,026][starVLA.training.train_starvla][INFO] - Step 4722, Loss: {'train/loss': 0.2514386475086212, 'train/grad_norm_pre_clip': 19.199983596801758, 'timing/data': 0.014186366461217403, 'timing/model': 0.9567288737744093, 'learning_rate/qwen_vl_interface': 2.8339150444293055e-06, 'learning_rate/action_model': 2.8339150444293055e-05, 'epoch': 1.94}) +[2026-06-18 00:03:43,826][starVLA.training.train_starvla][INFO] - Step 4723, Loss: {'train/loss': 0.4673417806625366, 'train/grad_norm_pre_clip': 23.9283504486084, 'timing/data': 0.010513720102608204, 'timing/model': 0.941469008103013, 'learning_rate/qwen_vl_interface': 2.8320532751857365e-06, 'learning_rate/action_model': 2.8320532751857366e-05, 'epoch': 1.94}) +[2026-06-18 00:03:47,644][starVLA.training.train_starvla][INFO] - Step 4724, Loss: {'train/loss': 0.19828017055988312, 'train/grad_norm_pre_clip': 14.45198917388916, 'timing/data': 0.008834216743707657, 'timing/model': 0.9446203522384167, 'learning_rate/qwen_vl_interface': 2.830192007185127e-06, 'learning_rate/action_model': 2.830192007185127e-05, 'epoch': 1.94}) +[2026-06-18 00:03:51,540][starVLA.training.train_starvla][INFO] - Step 4725, Loss: {'train/loss': 0.5796076059341431, 'train/grad_norm_pre_clip': 30.126605987548828, 'timing/data': 0.010165863670408726, 'timing/model': 0.9314654218032956, 'learning_rate/qwen_vl_interface': 2.828331240813318e-06, 'learning_rate/action_model': 2.8283312408133176e-05, 'epoch': 1.94}) +[2026-06-18 00:03:55,370][starVLA.training.train_starvla][INFO] - Step 4726, Loss: {'train/loss': 0.3441277742385864, 'train/grad_norm_pre_clip': 23.204702377319336, 'timing/data': 0.00953445304185152, 'timing/model': 1.141128245741129, 'learning_rate/qwen_vl_interface': 2.826470976456049e-06, 'learning_rate/action_model': 2.826470976456049e-05, 'epoch': 1.95}) +[2026-06-18 00:03:59,262][starVLA.training.train_starvla][INFO] - Step 4727, Loss: {'train/loss': 0.2028883397579193, 'train/grad_norm_pre_clip': 16.212032318115234, 'timing/data': 0.009839534759521484, 'timing/model': 0.9887368017807603, 'learning_rate/qwen_vl_interface': 2.8246112144989548e-06, 'learning_rate/action_model': 2.8246112144989543e-05, 'epoch': 1.95}) +[2026-06-18 00:04:03,181][starVLA.training.train_starvla][INFO] - Step 4728, Loss: {'train/loss': 0.37391096353530884, 'train/grad_norm_pre_clip': 17.145187377929688, 'timing/data': 0.010143140330910683, 'timing/model': 0.937780330888927, 'learning_rate/qwen_vl_interface': 2.8227519553275644e-06, 'learning_rate/action_model': 2.8227519553275645e-05, 'epoch': 1.95}) +[2026-06-18 00:04:08,029][starVLA.training.train_starvla][INFO] - Step 4729, Loss: {'train/loss': 0.5113618969917297, 'train/grad_norm_pre_clip': 25.54183578491211, 'timing/data': 0.010348159819841385, 'timing/model': 1.7876764833927155, 'learning_rate/qwen_vl_interface': 2.8208931993273066e-06, 'learning_rate/action_model': 2.8208931993273062e-05, 'epoch': 1.95}) +[2026-06-18 00:04:12,650][starVLA.training.train_starvla][INFO] - Step 4730, Loss: {'train/loss': 0.22398072481155396, 'train/grad_norm_pre_clip': 15.076220512390137, 'timing/data': 0.011633067391812801, 'timing/model': 0.9524212991818786, 'learning_rate/qwen_vl_interface': 2.8190349468835006e-06, 'learning_rate/action_model': 2.8190349468835004e-05, 'epoch': 1.95}) +[2026-06-18 00:04:16,550][starVLA.training.train_starvla][INFO] - Step 4731, Loss: {'train/loss': 0.22151894867420197, 'train/grad_norm_pre_clip': 18.918607711791992, 'timing/data': 0.008371255360543728, 'timing/model': 0.9309489810839295, 'learning_rate/qwen_vl_interface': 2.817177198381369e-06, 'learning_rate/action_model': 2.817177198381369e-05, 'epoch': 1.95}) +[2026-06-18 00:04:20,348][starVLA.training.train_starvla][INFO] - Step 4732, Loss: {'train/loss': 0.3323204815387726, 'train/grad_norm_pre_clip': 17.37482261657715, 'timing/data': 0.011567825451493263, 'timing/model': 1.1055594449862838, 'learning_rate/qwen_vl_interface': 2.8153199542060206e-06, 'learning_rate/action_model': 2.8153199542060206e-05, 'epoch': 1.95}) +[2026-06-18 00:04:24,126][starVLA.training.train_starvla][INFO] - Step 4733, Loss: {'train/loss': 0.2483103722333908, 'train/grad_norm_pre_clip': 17.98722267150879, 'timing/data': 0.008043062873184681, 'timing/model': 1.1108223143965006, 'learning_rate/qwen_vl_interface': 2.813463214742465e-06, 'learning_rate/action_model': 2.813463214742465e-05, 'epoch': 1.95}) +[2026-06-18 00:04:27,733][starVLA.training.train_starvla][INFO] - Step 4734, Loss: {'train/loss': 0.2218089997768402, 'train/grad_norm_pre_clip': 15.295214653015137, 'timing/data': 0.008288969285786152, 'timing/model': 0.938470147550106, 'learning_rate/qwen_vl_interface': 2.8116069803756047e-06, 'learning_rate/action_model': 2.8116069803756047e-05, 'epoch': 1.95}) +[2026-06-18 00:04:31,492][starVLA.training.train_starvla][INFO] - Step 4735, Loss: {'train/loss': 0.23448655009269714, 'train/grad_norm_pre_clip': 15.537578582763672, 'timing/data': 0.009951533749699593, 'timing/model': 0.9252045480534434, 'learning_rate/qwen_vl_interface': 2.809751251490244e-06, 'learning_rate/action_model': 2.809751251490244e-05, 'epoch': 1.95}) +[2026-06-18 00:04:35,243][starVLA.training.train_starvla][INFO] - Step 4736, Loss: {'train/loss': 0.27102887630462646, 'train/grad_norm_pre_clip': 19.43317222595215, 'timing/data': 0.009944609366357327, 'timing/model': 0.9298971407115459, 'learning_rate/qwen_vl_interface': 2.807896028471076e-06, 'learning_rate/action_model': 2.807896028471076e-05, 'epoch': 1.95}) +[2026-06-18 00:04:38,984][starVLA.training.train_starvla][INFO] - Step 4737, Loss: {'train/loss': 0.4370696544647217, 'train/grad_norm_pre_clip': 21.768112182617188, 'timing/data': 0.009576494805514812, 'timing/model': 0.9275138340890408, 'learning_rate/qwen_vl_interface': 2.806041311702688e-06, 'learning_rate/action_model': 2.806041311702688e-05, 'epoch': 1.95}) +[2026-06-18 00:04:42,882][starVLA.training.train_starvla][INFO] - Step 4738, Loss: {'train/loss': 0.19332821667194366, 'train/grad_norm_pre_clip': 18.63182258605957, 'timing/data': 0.017537105828523636, 'timing/model': 0.9447886496782303, 'learning_rate/qwen_vl_interface': 2.8041871015695626e-06, 'learning_rate/action_model': 2.8041871015695626e-05, 'epoch': 1.95}) +[2026-06-18 00:04:46,837][starVLA.training.train_starvla][INFO] - Step 4739, Loss: {'train/loss': 0.42188727855682373, 'train/grad_norm_pre_clip': 19.3862247467041, 'timing/data': 0.00909576378762722, 'timing/model': 1.1987550584599376, 'learning_rate/qwen_vl_interface': 2.802333398456086e-06, 'learning_rate/action_model': 2.8023333984560856e-05, 'epoch': 1.95}) +[2026-06-18 00:04:50,614][starVLA.training.train_starvla][INFO] - Step 4740, Loss: {'train/loss': 0.6363980770111084, 'train/grad_norm_pre_clip': 25.981130599975586, 'timing/data': 0.010172771289944649, 'timing/model': 1.0834182212129235, 'learning_rate/qwen_vl_interface': 2.8004802027465296e-06, 'learning_rate/action_model': 2.8004802027465293e-05, 'epoch': 1.95}) +[2026-06-18 00:04:54,223][starVLA.training.train_starvla][INFO] - Step 4741, Loss: {'train/loss': 0.33734628558158875, 'train/grad_norm_pre_clip': 17.897647857666016, 'timing/data': 0.008857880719006062, 'timing/model': 0.9357942054048181, 'learning_rate/qwen_vl_interface': 2.7986275148250624e-06, 'learning_rate/action_model': 2.7986275148250622e-05, 'epoch': 1.95}) +[2026-06-18 00:04:58,046][starVLA.training.train_starvla][INFO] - Step 4742, Loss: {'train/loss': 0.22635850310325623, 'train/grad_norm_pre_clip': 26.45647430419922, 'timing/data': 0.010711590759456158, 'timing/model': 0.9622502755373716, 'learning_rate/qwen_vl_interface': 2.7967753350757498e-06, 'learning_rate/action_model': 2.7967753350757497e-05, 'epoch': 1.95}) +[2026-06-18 00:05:01,871][starVLA.training.train_starvla][INFO] - Step 4743, Loss: {'train/loss': 0.26321959495544434, 'train/grad_norm_pre_clip': 20.085092544555664, 'timing/data': 0.012112689204514027, 'timing/model': 0.9323321711272001, 'learning_rate/qwen_vl_interface': 2.7949236638825496e-06, 'learning_rate/action_model': 2.7949236638825495e-05, 'epoch': 1.95}) +[2026-06-18 00:05:05,572][starVLA.training.train_starvla][INFO] - Step 4744, Loss: {'train/loss': 0.5725016593933105, 'train/grad_norm_pre_clip': 27.451839447021484, 'timing/data': 0.011220390908420086, 'timing/model': 0.9344350080937147, 'learning_rate/qwen_vl_interface': 2.7930725016293155e-06, 'learning_rate/action_model': 2.793072501629315e-05, 'epoch': 1.95}) +[2026-06-18 00:05:09,359][starVLA.training.train_starvla][INFO] - Step 4745, Loss: {'train/loss': 0.24095112085342407, 'train/grad_norm_pre_clip': 17.34490203857422, 'timing/data': 0.008040404878556728, 'timing/model': 0.93227185215801, 'learning_rate/qwen_vl_interface': 2.7912218486997954e-06, 'learning_rate/action_model': 2.7912218486997954e-05, 'epoch': 1.95}) +[2026-06-18 00:05:13,230][starVLA.training.train_starvla][INFO] - Step 4746, Loss: {'train/loss': 0.26921918988227844, 'train/grad_norm_pre_clip': 14.905904769897461, 'timing/data': 0.008342872373759747, 'timing/model': 1.1625324562191963, 'learning_rate/qwen_vl_interface': 2.7893717054776327e-06, 'learning_rate/action_model': 2.7893717054776327e-05, 'epoch': 1.95}) +[2026-06-18 00:05:16,896][starVLA.training.train_starvla][INFO] - Step 4747, Loss: {'train/loss': 0.14782172441482544, 'train/grad_norm_pre_clip': 10.07502269744873, 'timing/data': 0.010521785356104374, 'timing/model': 0.9726369986310601, 'learning_rate/qwen_vl_interface': 2.7875220723463624e-06, 'learning_rate/action_model': 2.7875220723463625e-05, 'epoch': 1.95}) +[2026-06-18 00:05:20,766][starVLA.training.train_starvla][INFO] - Step 4748, Loss: {'train/loss': 0.4166068434715271, 'train/grad_norm_pre_clip': 21.23006820678711, 'timing/data': 0.007973480969667435, 'timing/model': 0.9711829544976354, 'learning_rate/qwen_vl_interface': 2.7856729496894165e-06, 'learning_rate/action_model': 2.7856729496894162e-05, 'epoch': 1.95}) +[2026-06-18 00:05:24,545][starVLA.training.train_starvla][INFO] - Step 4749, Loss: {'train/loss': 0.6331854462623596, 'train/grad_norm_pre_clip': 23.269012451171875, 'timing/data': 0.010167192667722702, 'timing/model': 0.9522299570962787, 'learning_rate/qwen_vl_interface': 2.783824337890118e-06, 'learning_rate/action_model': 2.783824337890118e-05, 'epoch': 1.95}) +[2026-06-18 00:05:28,484][starVLA.training.train_starvla][INFO] - Step 4750, Loss: {'train/loss': 0.3147451877593994, 'train/grad_norm_pre_clip': 22.341012954711914, 'timing/data': 0.009371173568069935, 'timing/model': 0.9660615744069219, 'learning_rate/qwen_vl_interface': 2.7819762373316916e-06, 'learning_rate/action_model': 2.7819762373316917e-05, 'epoch': 1.96}) +[2026-06-18 00:05:32,324][starVLA.training.train_starvla][INFO] - Step 4751, Loss: {'train/loss': 0.22783416509628296, 'train/grad_norm_pre_clip': 15.819014549255371, 'timing/data': 0.010467239655554295, 'timing/model': 0.9572639176622033, 'learning_rate/qwen_vl_interface': 2.7801286483972458e-06, 'learning_rate/action_model': 2.7801286483972457e-05, 'epoch': 1.96}) +[2026-06-18 00:05:36,154][starVLA.training.train_starvla][INFO] - Step 4752, Loss: {'train/loss': 0.4403440058231354, 'train/grad_norm_pre_clip': 20.89188003540039, 'timing/data': 0.0085029611364007, 'timing/model': 1.115385316312313, 'learning_rate/qwen_vl_interface': 2.778281571469786e-06, 'learning_rate/action_model': 2.7782815714697858e-05, 'epoch': 1.96}) +[2026-06-18 00:05:40,047][starVLA.training.train_starvla][INFO] - Step 4753, Loss: {'train/loss': 0.3706147074699402, 'train/grad_norm_pre_clip': 16.349475860595703, 'timing/data': 0.007140823639929295, 'timing/model': 1.1345725441351533, 'learning_rate/qwen_vl_interface': 2.776435006932219e-06, 'learning_rate/action_model': 2.7764350069322187e-05, 'epoch': 1.96}) +[2026-06-18 00:05:44,286][starVLA.training.train_starvla][INFO] - Step 4754, Loss: {'train/loss': 0.35923171043395996, 'train/grad_norm_pre_clip': 18.722335815429688, 'timing/data': 0.007593212649226189, 'timing/model': 0.9378422042354941, 'learning_rate/qwen_vl_interface': 2.7745889551673355e-06, 'learning_rate/action_model': 2.7745889551673353e-05, 'epoch': 1.96}) +[2026-06-18 00:05:48,181][starVLA.training.train_starvla][INFO] - Step 4755, Loss: {'train/loss': 0.5911875367164612, 'train/grad_norm_pre_clip': 28.94778060913086, 'timing/data': 0.011701125651597977, 'timing/model': 0.9392833393067122, 'learning_rate/qwen_vl_interface': 2.772743416557826e-06, 'learning_rate/action_model': 2.772743416557826e-05, 'epoch': 1.96}) +[2026-06-18 00:05:52,036][starVLA.training.train_starvla][INFO] - Step 4756, Loss: {'train/loss': 0.2098240852355957, 'train/grad_norm_pre_clip': 16.214187622070312, 'timing/data': 0.009617872536182404, 'timing/model': 0.9472243869677186, 'learning_rate/qwen_vl_interface': 2.770898391486272e-06, 'learning_rate/action_model': 2.7708983914862718e-05, 'epoch': 1.96}) +[2026-06-18 00:05:55,984][starVLA.training.train_starvla][INFO] - Step 4757, Loss: {'train/loss': 0.24067093431949615, 'train/grad_norm_pre_clip': 16.266952514648438, 'timing/data': 0.013878634199500084, 'timing/model': 0.9579107398167253, 'learning_rate/qwen_vl_interface': 2.7690538803351484e-06, 'learning_rate/action_model': 2.7690538803351484e-05, 'epoch': 1.96}) +[2026-06-18 00:05:59,991][starVLA.training.train_starvla][INFO] - Step 4758, Loss: {'train/loss': 0.5220880508422852, 'train/grad_norm_pre_clip': 22.630470275878906, 'timing/data': 0.0109111238270998, 'timing/model': 0.9927648259326816, 'learning_rate/qwen_vl_interface': 2.7672098834868253e-06, 'learning_rate/action_model': 2.7672098834868253e-05, 'epoch': 1.96}) +[2026-06-18 00:06:03,826][starVLA.training.train_starvla][INFO] - Step 4759, Loss: {'train/loss': 0.1978573203086853, 'train/grad_norm_pre_clip': 15.041050910949707, 'timing/data': 0.009152185171842575, 'timing/model': 1.1509913671761751, 'learning_rate/qwen_vl_interface': 2.7653664013235655e-06, 'learning_rate/action_model': 2.7653664013235653e-05, 'epoch': 1.96}) +[2026-06-18 00:06:07,663][starVLA.training.train_starvla][INFO] - Step 4760, Loss: {'train/loss': 0.35135307908058167, 'train/grad_norm_pre_clip': 20.94405746459961, 'timing/data': 0.010447394102811813, 'timing/model': 1.1209216192364693, 'learning_rate/qwen_vl_interface': 2.763523434227524e-06, 'learning_rate/action_model': 2.763523434227524e-05, 'epoch': 1.96}) +[2026-06-18 00:06:12,663][starVLA.training.train_starvla][INFO] - Step 4761, Loss: {'train/loss': 0.2445518523454666, 'train/grad_norm_pre_clip': 17.55801773071289, 'timing/data': 0.012230558320879936, 'timing/model': 0.9545693648979068, 'learning_rate/qwen_vl_interface': 2.76168098258075e-06, 'learning_rate/action_model': 2.76168098258075e-05, 'epoch': 1.96}) +[2026-06-18 00:06:16,545][starVLA.training.train_starvla][INFO] - Step 4762, Loss: {'train/loss': 0.5066222548484802, 'train/grad_norm_pre_clip': 23.888391494750977, 'timing/data': 0.010779483243823051, 'timing/model': 0.9519426543265581, 'learning_rate/qwen_vl_interface': 2.7598390467651865e-06, 'learning_rate/action_model': 2.7598390467651863e-05, 'epoch': 1.96}) +[2026-06-18 00:06:20,415][starVLA.training.train_starvla][INFO] - Step 4763, Loss: {'train/loss': 0.16480432450771332, 'train/grad_norm_pre_clip': 20.997804641723633, 'timing/data': 0.012805781327188015, 'timing/model': 0.931159327737987, 'learning_rate/qwen_vl_interface': 2.757997627162665e-06, 'learning_rate/action_model': 2.7579976271626652e-05, 'epoch': 1.96}) +[2026-06-18 00:06:24,253][starVLA.training.train_starvla][INFO] - Step 4764, Loss: {'train/loss': 0.37593722343444824, 'train/grad_norm_pre_clip': 21.04367446899414, 'timing/data': 0.008991850540041924, 'timing/model': 0.9381517302244902, 'learning_rate/qwen_vl_interface': 2.7561567241549203e-06, 'learning_rate/action_model': 2.7561567241549202e-05, 'epoch': 1.96}) +[2026-06-18 00:06:28,176][starVLA.training.train_starvla][INFO] - Step 4765, Loss: {'train/loss': 0.4459648132324219, 'train/grad_norm_pre_clip': 28.13572883605957, 'timing/data': 0.008258594200015068, 'timing/model': 0.9350124420598149, 'learning_rate/qwen_vl_interface': 2.7543163381235705e-06, 'learning_rate/action_model': 2.7543163381235702e-05, 'epoch': 1.96}) +[2026-06-18 00:06:32,123][starVLA.training.train_starvla][INFO] - Step 4766, Loss: {'train/loss': 0.1915811002254486, 'train/grad_norm_pre_clip': 11.792441368103027, 'timing/data': 0.009670908562839031, 'timing/model': 1.2134964251890779, 'learning_rate/qwen_vl_interface': 2.752476469450128e-06, 'learning_rate/action_model': 2.7524764694501283e-05, 'epoch': 1.96}) +[2026-06-18 00:06:35,945][starVLA.training.train_starvla][INFO] - Step 4767, Loss: {'train/loss': 0.29843422770500183, 'train/grad_norm_pre_clip': 19.08881378173828, 'timing/data': 0.009505050256848335, 'timing/model': 0.9401171803474426, 'learning_rate/qwen_vl_interface': 2.750637118515999e-06, 'learning_rate/action_model': 2.750637118515999e-05, 'epoch': 1.96}) +[2026-06-18 00:06:39,842][starVLA.training.train_starvla][INFO] - Step 4768, Loss: {'train/loss': 0.12614579498767853, 'train/grad_norm_pre_clip': 8.198199272155762, 'timing/data': 0.012869312427937984, 'timing/model': 1.0050945794209838, 'learning_rate/qwen_vl_interface': 2.7487982857024875e-06, 'learning_rate/action_model': 2.7487982857024876e-05, 'epoch': 1.96}) +[2026-06-18 00:06:43,636][starVLA.training.train_starvla][INFO] - Step 4769, Loss: {'train/loss': 0.29994869232177734, 'train/grad_norm_pre_clip': 21.02383041381836, 'timing/data': 0.011990435421466827, 'timing/model': 0.9410520559176803, 'learning_rate/qwen_vl_interface': 2.746959971390782e-06, 'learning_rate/action_model': 2.7469599713907817e-05, 'epoch': 1.96}) +[2026-06-18 00:06:47,589][starVLA.training.train_starvla][INFO] - Step 4770, Loss: {'train/loss': 0.35178273916244507, 'train/grad_norm_pre_clip': 17.487655639648438, 'timing/data': 0.009242934174835682, 'timing/model': 1.027506827376783, 'learning_rate/qwen_vl_interface': 2.745122175961967e-06, 'learning_rate/action_model': 2.7451221759619672e-05, 'epoch': 1.96}) +[2026-06-18 00:06:51,514][starVLA.training.train_starvla][INFO] - Step 4771, Loss: {'train/loss': 0.27092117071151733, 'train/grad_norm_pre_clip': 20.208450317382812, 'timing/data': 0.00928613729774952, 'timing/model': 1.035080169327557, 'learning_rate/qwen_vl_interface': 2.743284899797022e-06, 'learning_rate/action_model': 2.7432848997970218e-05, 'epoch': 1.96}) +[2026-06-18 00:06:55,303][starVLA.training.train_starvla][INFO] - Step 4772, Loss: {'train/loss': 0.32272616028785706, 'train/grad_norm_pre_clip': 21.622257232666016, 'timing/data': 0.012485705316066742, 'timing/model': 0.9365046378225088, 'learning_rate/qwen_vl_interface': 2.741448143276813e-06, 'learning_rate/action_model': 2.741448143276813e-05, 'epoch': 1.96}) +[2026-06-18 00:06:59,140][starVLA.training.train_starvla][INFO] - Step 4773, Loss: {'train/loss': 0.26905110478401184, 'train/grad_norm_pre_clip': 16.35706329345703, 'timing/data': 0.008261522278189659, 'timing/model': 1.1050177002325654, 'learning_rate/qwen_vl_interface': 2.7396119067821043e-06, 'learning_rate/action_model': 2.7396119067821043e-05, 'epoch': 1.96}) +[2026-06-18 00:07:02,791][starVLA.training.train_starvla][INFO] - Step 4774, Loss: {'train/loss': 0.32796046137809753, 'train/grad_norm_pre_clip': 31.015586853027344, 'timing/data': 0.009078491479158401, 'timing/model': 0.946081381291151, 'learning_rate/qwen_vl_interface': 2.737776190693548e-06, 'learning_rate/action_model': 2.7377761906935477e-05, 'epoch': 1.97}) +[2026-06-18 00:07:06,705][starVLA.training.train_starvla][INFO] - Step 4775, Loss: {'train/loss': 0.4831492304801941, 'train/grad_norm_pre_clip': 20.085111618041992, 'timing/data': 0.007629529573023319, 'timing/model': 0.9592431103810668, 'learning_rate/qwen_vl_interface': 2.7359409953916914e-06, 'learning_rate/action_model': 2.7359409953916913e-05, 'epoch': 1.97}) +[2026-06-18 00:07:10,597][starVLA.training.train_starvla][INFO] - Step 4776, Loss: {'train/loss': 0.3753229081630707, 'train/grad_norm_pre_clip': 17.066783905029297, 'timing/data': 0.007852967828512192, 'timing/model': 0.9497298523783684, 'learning_rate/qwen_vl_interface': 2.734106321256971e-06, 'learning_rate/action_model': 2.7341063212569705e-05, 'epoch': 1.97}) +[2026-06-18 00:07:14,423][starVLA.training.train_starvla][INFO] - Step 4777, Loss: {'train/loss': 0.4067647457122803, 'train/grad_norm_pre_clip': 25.44217872619629, 'timing/data': 0.008918199688196182, 'timing/model': 0.9441842017695308, 'learning_rate/qwen_vl_interface': 2.732272168669718e-06, 'learning_rate/action_model': 2.732272168669718e-05, 'epoch': 1.97}) +[2026-06-18 00:07:18,368][starVLA.training.train_starvla][INFO] - Step 4778, Loss: {'train/loss': 0.3236636519432068, 'train/grad_norm_pre_clip': 19.520673751831055, 'timing/data': 0.021188810467720032, 'timing/model': 0.9906428419053555, 'learning_rate/qwen_vl_interface': 2.7304385380101512e-06, 'learning_rate/action_model': 2.730438538010151e-05, 'epoch': 1.97}) +[2026-06-18 00:07:22,435][starVLA.training.train_starvla][INFO] - Step 4779, Loss: {'train/loss': 0.22822290658950806, 'train/grad_norm_pre_clip': 13.687422752380371, 'timing/data': 0.008282441645860672, 'timing/model': 1.1859884560108185, 'learning_rate/qwen_vl_interface': 2.7286054296583904e-06, 'learning_rate/action_model': 2.72860542965839e-05, 'epoch': 1.97}) +[2026-06-18 00:07:26,221][starVLA.training.train_starvla][INFO] - Step 4780, Loss: {'train/loss': 0.7385307550430298, 'train/grad_norm_pre_clip': 24.0316162109375, 'timing/data': 0.007576044648885727, 'timing/model': 1.105528554879129, 'learning_rate/qwen_vl_interface': 2.726772843994435e-06, 'learning_rate/action_model': 2.7267728439944352e-05, 'epoch': 1.97}) +[2026-06-18 00:07:29,837][starVLA.training.train_starvla][INFO] - Step 4781, Loss: {'train/loss': 0.21703888475894928, 'train/grad_norm_pre_clip': 15.13465690612793, 'timing/data': 0.007101851515471935, 'timing/model': 0.9286999264732003, 'learning_rate/qwen_vl_interface': 2.7249407813981834e-06, 'learning_rate/action_model': 2.724940781398183e-05, 'epoch': 1.97}) +[2026-06-18 00:07:33,642][starVLA.training.train_starvla][INFO] - Step 4782, Loss: {'train/loss': 0.32460761070251465, 'train/grad_norm_pre_clip': 20.80623435974121, 'timing/data': 0.01174873672425747, 'timing/model': 0.9490392059087753, 'learning_rate/qwen_vl_interface': 2.7231092422494255e-06, 'learning_rate/action_model': 2.7231092422494253e-05, 'epoch': 1.97}) +[2026-06-18 00:07:37,433][starVLA.training.train_starvla][INFO] - Step 4783, Loss: {'train/loss': 0.2350541651248932, 'train/grad_norm_pre_clip': 18.526979446411133, 'timing/data': 0.00806848332285881, 'timing/model': 0.9414176652207971, 'learning_rate/qwen_vl_interface': 2.7212782269278417e-06, 'learning_rate/action_model': 2.7212782269278413e-05, 'epoch': 1.97}) +[2026-06-18 00:07:41,256][starVLA.training.train_starvla][INFO] - Step 4784, Loss: {'train/loss': 0.17051474750041962, 'train/grad_norm_pre_clip': 11.726517677307129, 'timing/data': 0.009764393791556358, 'timing/model': 0.9301793053746223, 'learning_rate/qwen_vl_interface': 2.719447735813004e-06, 'learning_rate/action_model': 2.719447735813004e-05, 'epoch': 1.97}) +[2026-06-18 00:07:45,474][starVLA.training.train_starvla][INFO] - Step 4785, Loss: {'train/loss': 0.23682813346385956, 'train/grad_norm_pre_clip': 18.859989166259766, 'timing/data': 0.0086691714823246, 'timing/model': 0.9568329202011228, 'learning_rate/qwen_vl_interface': 2.7176177692843697e-06, 'learning_rate/action_model': 2.7176177692843697e-05, 'epoch': 1.97}) +[2026-06-18 00:07:49,355][starVLA.training.train_starvla][INFO] - Step 4786, Loss: {'train/loss': 0.18852964043617249, 'train/grad_norm_pre_clip': 17.627290725708008, 'timing/data': 0.009415953420102596, 'timing/model': 1.1221374813467264, 'learning_rate/qwen_vl_interface': 2.7157883277212985e-06, 'learning_rate/action_model': 2.7157883277212985e-05, 'epoch': 1.97}) +[2026-06-18 00:07:53,489][starVLA.training.train_starvla][INFO] - Step 4787, Loss: {'train/loss': 0.3447759747505188, 'train/grad_norm_pre_clip': 17.781600952148438, 'timing/data': 0.02349512930959463, 'timing/model': 1.3862643372267485, 'learning_rate/qwen_vl_interface': 2.7139594115030347e-06, 'learning_rate/action_model': 2.7139594115030348e-05, 'epoch': 1.97}) +[2026-06-18 00:07:58,521][starVLA.training.train_starvla][INFO] - Step 4788, Loss: {'train/loss': 0.6542167067527771, 'train/grad_norm_pre_clip': 29.518672943115234, 'timing/data': 0.010769165121018887, 'timing/model': 0.9437110600993037, 'learning_rate/qwen_vl_interface': 2.712131021008714e-06, 'learning_rate/action_model': 2.7121310210087137e-05, 'epoch': 1.97}) +[2026-06-18 00:08:02,698][starVLA.training.train_starvla][INFO] - Step 4789, Loss: {'train/loss': 0.653343141078949, 'train/grad_norm_pre_clip': 25.358980178833008, 'timing/data': 0.008560828864574432, 'timing/model': 0.9387522805482149, 'learning_rate/qwen_vl_interface': 2.7103031566173622e-06, 'learning_rate/action_model': 2.710303156617362e-05, 'epoch': 1.97}) +[2026-06-18 00:08:06,525][starVLA.training.train_starvla][INFO] - Step 4790, Loss: {'train/loss': 0.4416879415512085, 'train/grad_norm_pre_clip': 25.08553123474121, 'timing/data': 0.010947917588055134, 'timing/model': 0.946301719173789, 'learning_rate/qwen_vl_interface': 2.7084758187078995e-06, 'learning_rate/action_model': 2.7084758187078997e-05, 'epoch': 1.97}) +[2026-06-18 00:08:10,242][starVLA.training.train_starvla][INFO] - Step 4791, Loss: {'train/loss': 0.2724405825138092, 'train/grad_norm_pre_clip': 16.764938354492188, 'timing/data': 0.00915537029504776, 'timing/model': 0.9313313951715827, 'learning_rate/qwen_vl_interface': 2.7066490076591344e-06, 'learning_rate/action_model': 2.7066490076591343e-05, 'epoch': 1.97}) +[2026-06-18 00:08:14,181][starVLA.training.train_starvla][INFO] - Step 4792, Loss: {'train/loss': 0.23110099136829376, 'train/grad_norm_pre_clip': 22.96562957763672, 'timing/data': 0.008096541278064251, 'timing/model': 0.940689304843545, 'learning_rate/qwen_vl_interface': 2.704822723849765e-06, 'learning_rate/action_model': 2.7048227238497648e-05, 'epoch': 1.97}) +[2026-06-18 00:08:18,077][starVLA.training.train_starvla][INFO] - Step 4793, Loss: {'train/loss': 0.287361204624176, 'train/grad_norm_pre_clip': 17.305574417114258, 'timing/data': 0.013310783542692661, 'timing/model': 1.1151480972766876, 'learning_rate/qwen_vl_interface': 2.702996967658388e-06, 'learning_rate/action_model': 2.7029969676583876e-05, 'epoch': 1.97}) +[2026-06-18 00:08:21,759][starVLA.training.train_starvla][INFO] - Step 4794, Loss: {'train/loss': 0.47101932764053345, 'train/grad_norm_pre_clip': 28.56379508972168, 'timing/data': 0.006440897472202778, 'timing/model': 0.9281579293310642, 'learning_rate/qwen_vl_interface': 2.7011717394634767e-06, 'learning_rate/action_model': 2.701171739463477e-05, 'epoch': 1.97}) +[2026-06-18 00:08:25,600][starVLA.training.train_starvla][INFO] - Step 4795, Loss: {'train/loss': 0.55668705701828, 'train/grad_norm_pre_clip': 25.618005752563477, 'timing/data': 0.006579567678272724, 'timing/model': 0.9397917613387108, 'learning_rate/qwen_vl_interface': 2.699347039643408e-06, 'learning_rate/action_model': 2.699347039643408e-05, 'epoch': 1.97}) +[2026-06-18 00:08:29,349][starVLA.training.train_starvla][INFO] - Step 4796, Loss: {'train/loss': 0.20888274908065796, 'train/grad_norm_pre_clip': 17.047937393188477, 'timing/data': 0.006494076922535896, 'timing/model': 0.9341226667165756, 'learning_rate/qwen_vl_interface': 2.6975228685764398e-06, 'learning_rate/action_model': 2.69752286857644e-05, 'epoch': 1.97}) +[2026-06-18 00:08:33,224][starVLA.training.train_starvla][INFO] - Step 4797, Loss: {'train/loss': 0.14762462675571442, 'train/grad_norm_pre_clip': 12.463543891906738, 'timing/data': 0.00911020115017891, 'timing/model': 0.9378210594877601, 'learning_rate/qwen_vl_interface': 2.6956992266407307e-06, 'learning_rate/action_model': 2.6956992266407306e-05, 'epoch': 1.97}) +[2026-06-18 00:08:37,116][starVLA.training.train_starvla][INFO] - Step 4798, Loss: {'train/loss': 0.4452938139438629, 'train/grad_norm_pre_clip': 24.291580200195312, 'timing/data': 0.009076169691979885, 'timing/model': 0.9685583747923374, 'learning_rate/qwen_vl_interface': 2.693876114214321e-06, 'learning_rate/action_model': 2.6938761142143208e-05, 'epoch': 1.98}) +[2026-06-18 00:08:40,918][starVLA.training.train_starvla][INFO] - Step 4799, Loss: {'train/loss': 0.5463468432426453, 'train/grad_norm_pre_clip': 25.41434669494629, 'timing/data': 0.008804600685834885, 'timing/model': 1.1234081722795963, 'learning_rate/qwen_vl_interface': 2.6920535316751417e-06, 'learning_rate/action_model': 2.6920535316751416e-05, 'epoch': 1.98}) +[2026-06-18 00:10:42,778][starVLA.training.train_starvla][INFO] - Step 4800, Loss: {'train/loss': 0.22058185935020447, 'train/grad_norm_pre_clip': 12.233857154846191, 'eval/loss': 0.41549161076545715, 'eval/latency_6/loss': 0.41549161076545715, 'eval/demon_fire_precision': 0.9489508299404948, 'eval/demon_fire_recall': 0.9793148028442146, 'eval/demon_fire_f1': 0.9638937490058851, 'eval/demon_fire_rate_model': 0.8884251530328325, 'eval/demon_fire_rate_teacher': 0.860879243183083, 'eval/demon_left_precision': 0.8620689655172413, 'eval/demon_left_recall': 0.7456140350877193, 'eval/demon_left_f1': 0.799623706491063, 'eval/demon_left_rate_model': 0.13717306622148023, 'eval/demon_left_rate_teacher': 0.15859766277128548, 'eval/demon_right_precision': 0.8034591194968553, 'eval/demon_right_recall': 0.7849462365591398, 'eval/demon_right_f1': 0.794094794094794, 'eval/demon_right_rate_model': 0.17696160267111852, 'eval/demon_right_rate_teacher': 0.18113522537562604, 'eval/demon_move_f1': 0.7968592502929285, 'eval/demon_cc_f1': 0.9068656658744018, 'eval/latency_6/demon_fire_precision': 0.9489508299404948, 'eval/latency_6/demon_fire_recall': 0.9793148028442146, 'eval/latency_6/demon_fire_f1': 0.9638937490058851, 'eval/latency_6/demon_fire_rate_model': 0.8884251530328325, 'eval/latency_6/demon_fire_rate_teacher': 0.860879243183083, 'eval/latency_6/demon_left_precision': 0.8620689655172413, 'eval/latency_6/demon_left_recall': 0.7456140350877193, 'eval/latency_6/demon_left_f1': 0.799623706491063, 'eval/latency_6/demon_left_rate_model': 0.13717306622148023, 'eval/latency_6/demon_left_rate_teacher': 0.15859766277128548, 'eval/latency_6/demon_right_precision': 0.8034591194968553, 'eval/latency_6/demon_right_recall': 0.7849462365591398, 'eval/latency_6/demon_right_f1': 0.794094794094794, 'eval/latency_6/demon_right_rate_model': 0.17696160267111852, 'eval/latency_6/demon_right_rate_teacher': 0.18113522537562604, 'eval/latency_6/demon_move_f1': 0.7968592502929285, 'eval/latency_6/demon_cc_f1': 0.9068656658744018, 'timing/data': 0.01207013987004757, 'timing/model': 1.2188753690570593, 'learning_rate/qwen_vl_interface': 2.6902314794010164e-06, 'learning_rate/action_model': 2.6902314794010165e-05, 'epoch': 1.98}) +[2026-06-18 00:10:42,782][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_4800_state +[2026-06-18 00:10:42,786][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 00:10:49,402][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_4800_state/model.safetensors +[2026-06-18 00:11:05,543][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_4800_state/optimizer.bin +[2026-06-18 00:11:05,544][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_4800_state/sampler.bin +[2026-06-18 00:11:05,544][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_4800_state/sampler_1.bin +[2026-06-18 00:11:05,547][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_4800_state/random_states_0.pkl +[2026-06-18 00:11:05,550][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4800 +[2026-06-18 00:11:05,550][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 00:11:05,562][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 00:11:14,630][starVLA.training.train_starvla][INFO] - Step 4801, Loss: {'train/loss': 0.23377159237861633, 'train/grad_norm_pre_clip': 15.886404991149902, 'timing/data': 0.011426678858697414, 'timing/model': 1.0334529504179955, 'learning_rate/qwen_vl_interface': 2.6884099577696615e-06, 'learning_rate/action_model': 2.6884099577696615e-05, 'epoch': 1.98}) +[2026-06-18 00:11:20,186][starVLA.training.train_starvla][INFO] - Step 4802, Loss: {'train/loss': 0.25955942273139954, 'train/grad_norm_pre_clip': 18.833786010742188, 'timing/data': 0.009637315757572651, 'timing/model': 1.9860456325113773, 'learning_rate/qwen_vl_interface': 2.686588967158679e-06, 'learning_rate/action_model': 2.6865889671586792e-05, 'epoch': 1.98}) +[2026-06-18 00:11:24,654][starVLA.training.train_starvla][INFO] - Step 4803, Loss: {'train/loss': 0.2678718864917755, 'train/grad_norm_pre_clip': 44.686363220214844, 'timing/data': 0.00894605740904808, 'timing/model': 1.0823422372341156, 'learning_rate/qwen_vl_interface': 2.6847685079455614e-06, 'learning_rate/action_model': 2.6847685079455613e-05, 'epoch': 1.98}) +[2026-06-18 00:11:28,826][starVLA.training.train_starvla][INFO] - Step 4804, Loss: {'train/loss': 0.1775989681482315, 'train/grad_norm_pre_clip': 11.50242805480957, 'timing/data': 0.00968970824033022, 'timing/model': 0.9504136806353927, 'learning_rate/qwen_vl_interface': 2.6829485805076927e-06, 'learning_rate/action_model': 2.6829485805076925e-05, 'epoch': 1.98}) +[2026-06-18 00:11:32,645][starVLA.training.train_starvla][INFO] - Step 4805, Loss: {'train/loss': 0.4661370515823364, 'train/grad_norm_pre_clip': 34.07073974609375, 'timing/data': 0.007693564519286156, 'timing/model': 0.9454014142975211, 'learning_rate/qwen_vl_interface': 2.6811291852223452e-06, 'learning_rate/action_model': 2.6811291852223454e-05, 'epoch': 1.98}) +[2026-06-18 00:11:36,449][starVLA.training.train_starvla][INFO] - Step 4806, Loss: {'train/loss': 0.3697681725025177, 'train/grad_norm_pre_clip': 14.913125991821289, 'timing/data': 0.011168623343110085, 'timing/model': 0.9350171759724617, 'learning_rate/qwen_vl_interface': 2.679310322466683e-06, 'learning_rate/action_model': 2.6793103224666827e-05, 'epoch': 1.98}) +[2026-06-18 00:11:40,305][starVLA.training.train_starvla][INFO] - Step 4807, Loss: {'train/loss': 0.5115075707435608, 'train/grad_norm_pre_clip': 29.919418334960938, 'timing/data': 0.012837189249694347, 'timing/model': 0.942161557264626, 'learning_rate/qwen_vl_interface': 2.6774919926177563e-06, 'learning_rate/action_model': 2.677491992617756e-05, 'epoch': 1.98}) +[2026-06-18 00:11:44,124][starVLA.training.train_starvla][INFO] - Step 4808, Loss: {'train/loss': 0.3623667061328888, 'train/grad_norm_pre_clip': 20.45513343811035, 'timing/data': 0.013419434428215027, 'timing/model': 0.9633975904434919, 'learning_rate/qwen_vl_interface': 2.675674196052507e-06, 'learning_rate/action_model': 2.6756741960525072e-05, 'epoch': 1.98}) +[2026-06-18 00:11:47,923][starVLA.training.train_starvla][INFO] - Step 4809, Loss: {'train/loss': 0.15001942217350006, 'train/grad_norm_pre_clip': 13.22120189666748, 'timing/data': 0.011420289985835552, 'timing/model': 1.1255354452878237, 'learning_rate/qwen_vl_interface': 2.6738569331477677e-06, 'learning_rate/action_model': 2.6738569331477674e-05, 'epoch': 1.98}) +[2026-06-18 00:11:51,539][starVLA.training.train_starvla][INFO] - Step 4810, Loss: {'train/loss': 0.28436940908432007, 'train/grad_norm_pre_clip': 16.130775451660156, 'timing/data': 0.023160607554018497, 'timing/model': 0.9338284097611904, 'learning_rate/qwen_vl_interface': 2.672040204280258e-06, 'learning_rate/action_model': 2.672040204280258e-05, 'epoch': 1.98}) +[2026-06-18 00:11:55,428][starVLA.training.train_starvla][INFO] - Step 4811, Loss: {'train/loss': 0.24121764302253723, 'train/grad_norm_pre_clip': 19.559532165527344, 'timing/data': 0.018422060646116734, 'timing/model': 0.9477225169539452, 'learning_rate/qwen_vl_interface': 2.6702240098265864e-06, 'learning_rate/action_model': 2.6702240098265864e-05, 'epoch': 1.98}) +[2026-06-18 00:11:59,293][starVLA.training.train_starvla][INFO] - Step 4812, Loss: {'train/loss': 0.35594406723976135, 'train/grad_norm_pre_clip': 21.306324005126953, 'timing/data': 0.015673941001296043, 'timing/model': 0.9433256424963474, 'learning_rate/qwen_vl_interface': 2.6684083501632573e-06, 'learning_rate/action_model': 2.6684083501632574e-05, 'epoch': 1.98}) +[2026-06-18 00:12:03,099][starVLA.training.train_starvla][INFO] - Step 4813, Loss: {'train/loss': 0.4226943552494049, 'train/grad_norm_pre_clip': 23.724092483520508, 'timing/data': 0.010359758511185646, 'timing/model': 0.9305320139974356, 'learning_rate/qwen_vl_interface': 2.6665932256666525e-06, 'learning_rate/action_model': 2.6665932256666527e-05, 'epoch': 1.98}) +[2026-06-18 00:12:06,894][starVLA.training.train_starvla][INFO] - Step 4814, Loss: {'train/loss': 0.42194291949272156, 'train/grad_norm_pre_clip': 23.115535736083984, 'timing/data': 0.010684764012694359, 'timing/model': 0.9629413457587361, 'learning_rate/qwen_vl_interface': 2.6647786367130513e-06, 'learning_rate/action_model': 2.664778636713051e-05, 'epoch': 1.98}) +[2026-06-18 00:12:10,639][starVLA.training.train_starvla][INFO] - Step 4815, Loss: {'train/loss': 0.20877309143543243, 'train/grad_norm_pre_clip': 16.209688186645508, 'timing/data': 0.008268758654594421, 'timing/model': 1.0916572082787752, 'learning_rate/qwen_vl_interface': 2.6629645836786217e-06, 'learning_rate/action_model': 2.6629645836786217e-05, 'epoch': 1.98}) +[2026-06-18 00:12:14,507][starVLA.training.train_starvla][INFO] - Step 4816, Loss: {'train/loss': 0.4989505708217621, 'train/grad_norm_pre_clip': 36.34770965576172, 'timing/data': 0.0122215049341321, 'timing/model': 1.098653893917799, 'learning_rate/qwen_vl_interface': 2.661151066939417e-06, 'learning_rate/action_model': 2.6611510669394168e-05, 'epoch': 1.98}) +[2026-06-18 00:12:18,137][starVLA.training.train_starvla][INFO] - Step 4817, Loss: {'train/loss': 0.3404887318611145, 'train/grad_norm_pre_clip': 21.276308059692383, 'timing/data': 0.009890072047710419, 'timing/model': 0.9505039872601628, 'learning_rate/qwen_vl_interface': 2.6593380868713843e-06, 'learning_rate/action_model': 2.659338086871384e-05, 'epoch': 1.98}) +[2026-06-18 00:12:22,547][starVLA.training.train_starvla][INFO] - Step 4818, Loss: {'train/loss': 0.36700576543807983, 'train/grad_norm_pre_clip': 15.948452949523926, 'timing/data': 0.009214411489665508, 'timing/model': 0.9462523963302374, 'learning_rate/qwen_vl_interface': 2.6575256438503507e-06, 'learning_rate/action_model': 2.6575256438503504e-05, 'epoch': 1.98}) +[2026-06-18 00:12:26,422][starVLA.training.train_starvla][INFO] - Step 4819, Loss: {'train/loss': 0.29196086525917053, 'train/grad_norm_pre_clip': 18.396440505981445, 'timing/data': 0.01833090465515852, 'timing/model': 0.9502209974452853, 'learning_rate/qwen_vl_interface': 2.6557137382520413e-06, 'learning_rate/action_model': 2.6557137382520412e-05, 'epoch': 1.98}) +[2026-06-18 00:12:30,258][starVLA.training.train_starvla][INFO] - Step 4820, Loss: {'train/loss': 0.13765758275985718, 'train/grad_norm_pre_clip': 10.934337615966797, 'timing/data': 0.009183298796415329, 'timing/model': 0.9502094816416502, 'learning_rate/qwen_vl_interface': 2.6539023704520666e-06, 'learning_rate/action_model': 2.6539023704520664e-05, 'epoch': 1.98}) +[2026-06-18 00:12:34,350][starVLA.training.train_starvla][INFO] - Step 4821, Loss: {'train/loss': 0.220018669962883, 'train/grad_norm_pre_clip': 22.668495178222656, 'timing/data': 0.015457764267921448, 'timing/model': 0.9813796235248446, 'learning_rate/qwen_vl_interface': 2.6520915408259234e-06, 'learning_rate/action_model': 2.6520915408259235e-05, 'epoch': 1.98}) +[2026-06-18 00:12:38,536][starVLA.training.train_starvla][INFO] - Step 4822, Loss: {'train/loss': 0.34497514367103577, 'train/grad_norm_pre_clip': 19.930313110351562, 'timing/data': 0.00881569366902113, 'timing/model': 1.140987099148333, 'learning_rate/qwen_vl_interface': 2.650281249749e-06, 'learning_rate/action_model': 2.6502812497489998e-05, 'epoch': 1.98}) +[2026-06-18 00:12:42,246][starVLA.training.train_starvla][INFO] - Step 4823, Loss: {'train/loss': 0.3891594707965851, 'train/grad_norm_pre_clip': 23.119718551635742, 'timing/data': 0.0116520831361413, 'timing/model': 0.9754338292405009, 'learning_rate/qwen_vl_interface': 2.6484714975965704e-06, 'learning_rate/action_model': 2.6484714975965703e-05, 'epoch': 1.99}) +[2026-06-18 00:12:47,498][starVLA.training.train_starvla][INFO] - Step 4824, Loss: {'train/loss': 0.20736603438854218, 'train/grad_norm_pre_clip': 21.560771942138672, 'timing/data': 0.012333529070019722, 'timing/model': 1.3935760222375393, 'learning_rate/qwen_vl_interface': 2.646662284743799e-06, 'learning_rate/action_model': 2.646662284743799e-05, 'epoch': 1.99}) +[2026-06-18 00:12:51,472][starVLA.training.train_starvla][INFO] - Step 4825, Loss: {'train/loss': 0.2615950405597687, 'train/grad_norm_pre_clip': 16.306703567504883, 'timing/data': 0.010270049795508385, 'timing/model': 0.9630989599972963, 'learning_rate/qwen_vl_interface': 2.6448536115657353e-06, 'learning_rate/action_model': 2.6448536115657353e-05, 'epoch': 1.99}) +[2026-06-18 00:12:56,156][starVLA.training.train_starvla][INFO] - Step 4826, Loss: {'train/loss': 0.23731927573680878, 'train/grad_norm_pre_clip': 19.252777099609375, 'timing/data': 0.019123457372188568, 'timing/model': 0.9374374262988567, 'learning_rate/qwen_vl_interface': 2.6430454784373248e-06, 'learning_rate/action_model': 2.643045478437325e-05, 'epoch': 1.99}) +[2026-06-18 00:13:00,073][starVLA.training.train_starvla][INFO] - Step 4827, Loss: {'train/loss': 0.38799965381622314, 'train/grad_norm_pre_clip': 26.07054328918457, 'timing/data': 0.015714289620518684, 'timing/model': 0.9428535085171461, 'learning_rate/qwen_vl_interface': 2.6412378857333903e-06, 'learning_rate/action_model': 2.6412378857333903e-05, 'epoch': 1.99}) +[2026-06-18 00:13:03,911][starVLA.training.train_starvla][INFO] - Step 4828, Loss: {'train/loss': 0.14749252796173096, 'train/grad_norm_pre_clip': 13.739884376525879, 'timing/data': 0.009544729255139828, 'timing/model': 0.9468265725299716, 'learning_rate/qwen_vl_interface': 2.639430833828649e-06, 'learning_rate/action_model': 2.6394308338286488e-05, 'epoch': 1.99}) +[2026-06-18 00:13:07,724][starVLA.training.train_starvla][INFO] - Step 4829, Loss: {'train/loss': 0.1990315020084381, 'train/grad_norm_pre_clip': 10.310629844665527, 'timing/data': 0.008705229498445988, 'timing/model': 1.1171117052435875, 'learning_rate/qwen_vl_interface': 2.637624323097703e-06, 'learning_rate/action_model': 2.6376243230977026e-05, 'epoch': 1.99}) +[2026-06-18 00:13:11,346][starVLA.training.train_starvla][INFO] - Step 4830, Loss: {'train/loss': 0.2545531094074249, 'train/grad_norm_pre_clip': 17.755722045898438, 'timing/data': 0.008079715073108673, 'timing/model': 0.9465174339711666, 'learning_rate/qwen_vl_interface': 2.635818353915048e-06, 'learning_rate/action_model': 2.635818353915048e-05, 'epoch': 1.99}) +[2026-06-18 00:13:15,156][starVLA.training.train_starvla][INFO] - Step 4831, Loss: {'train/loss': 0.5880256295204163, 'train/grad_norm_pre_clip': 32.140586853027344, 'timing/data': 0.011630788445472717, 'timing/model': 0.9329653447493911, 'learning_rate/qwen_vl_interface': 2.634012926655063e-06, 'learning_rate/action_model': 2.6340129266550628e-05, 'epoch': 1.99}) +[2026-06-18 00:13:19,106][starVLA.training.train_starvla][INFO] - Step 4832, Loss: {'train/loss': 0.2502797245979309, 'train/grad_norm_pre_clip': 21.236974716186523, 'timing/data': 0.011992771178483963, 'timing/model': 0.9455516003072262, 'learning_rate/qwen_vl_interface': 2.632208041692009e-06, 'learning_rate/action_model': 2.6322080416920088e-05, 'epoch': 1.99}) +[2026-06-18 00:13:23,083][starVLA.training.train_starvla][INFO] - Step 4833, Loss: {'train/loss': 0.35278695821762085, 'train/grad_norm_pre_clip': 21.92698860168457, 'timing/data': 0.011385117657482624, 'timing/model': 1.1350836288183928, 'learning_rate/qwen_vl_interface': 2.630403699400047e-06, 'learning_rate/action_model': 2.6304036994000468e-05, 'epoch': 1.99}) +[2026-06-18 00:13:26,948][starVLA.training.train_starvla][INFO] - Step 4834, Loss: {'train/loss': 0.32189854979515076, 'train/grad_norm_pre_clip': 22.092538833618164, 'timing/data': 0.009724428877234459, 'timing/model': 0.9860526798292994, 'learning_rate/qwen_vl_interface': 2.628599900153217e-06, 'learning_rate/action_model': 2.628599900153217e-05, 'epoch': 1.99}) +[2026-06-18 00:13:30,869][starVLA.training.train_starvla][INFO] - Step 4835, Loss: {'train/loss': 0.6709581613540649, 'train/grad_norm_pre_clip': 29.86248779296875, 'timing/data': 0.010690351948142052, 'timing/model': 1.114620054140687, 'learning_rate/qwen_vl_interface': 2.6267966443254484e-06, 'learning_rate/action_model': 2.6267966443254483e-05, 'epoch': 1.99}) +[2026-06-18 00:13:36,098][starVLA.training.train_starvla][INFO] - Step 4836, Loss: {'train/loss': 0.3398350179195404, 'train/grad_norm_pre_clip': 27.854644775390625, 'timing/data': 0.012208295054733753, 'timing/model': 1.1434107795357704, 'learning_rate/qwen_vl_interface': 2.6249939322905576e-06, 'learning_rate/action_model': 2.6249939322905576e-05, 'epoch': 1.99}) +[2026-06-18 00:13:39,933][starVLA.training.train_starvla][INFO] - Step 4837, Loss: {'train/loss': 0.32093435525894165, 'train/grad_norm_pre_clip': 21.452239990234375, 'timing/data': 0.010412387549877167, 'timing/model': 0.9477009447291493, 'learning_rate/qwen_vl_interface': 2.6231917644222505e-06, 'learning_rate/action_model': 2.6231917644222504e-05, 'epoch': 1.99}) +[2026-06-18 00:13:46,320][starVLA.training.train_starvla][INFO] - Step 4838, Loss: {'train/loss': 0.6141061782836914, 'train/grad_norm_pre_clip': 26.6033878326416, 'timing/data': 0.012522052973508835, 'timing/model': 0.9643557835370302, 'learning_rate/qwen_vl_interface': 2.621390141094116e-06, 'learning_rate/action_model': 2.621390141094116e-05, 'epoch': 1.99}) +[2026-06-18 00:13:50,101][starVLA.training.train_starvla][INFO] - Step 4839, Loss: {'train/loss': 0.23144030570983887, 'train/grad_norm_pre_clip': 27.045917510986328, 'timing/data': 0.011020785197615623, 'timing/model': 0.9360167495906353, 'learning_rate/qwen_vl_interface': 2.6195890626796345e-06, 'learning_rate/action_model': 2.6195890626796345e-05, 'epoch': 1.99}) +[2026-06-18 00:13:53,907][starVLA.training.train_starvla][INFO] - Step 4840, Loss: {'train/loss': 0.14890822768211365, 'train/grad_norm_pre_clip': 14.037875175476074, 'timing/data': 0.007876808755099773, 'timing/model': 0.9515154296532273, 'learning_rate/qwen_vl_interface': 2.6177885295521706e-06, 'learning_rate/action_model': 2.6177885295521702e-05, 'epoch': 1.99}) +[2026-06-18 00:13:58,419][starVLA.training.train_starvla][INFO] - Step 4841, Loss: {'train/loss': 0.1992570012807846, 'train/grad_norm_pre_clip': 14.27541732788086, 'timing/data': 0.009170427918434143, 'timing/model': 1.6158023793250322, 'learning_rate/qwen_vl_interface': 2.6159885420849766e-06, 'learning_rate/action_model': 2.6159885420849766e-05, 'epoch': 1.99}) +[2026-06-18 00:14:02,357][starVLA.training.train_starvla][INFO] - Step 4842, Loss: {'train/loss': 0.439296156167984, 'train/grad_norm_pre_clip': 20.8687801361084, 'timing/data': 0.009698580019176006, 'timing/model': 1.2109187794849277, 'learning_rate/qwen_vl_interface': 2.6141891006511927e-06, 'learning_rate/action_model': 2.6141891006511925e-05, 'epoch': 1.99}) +[2026-06-18 00:14:06,263][starVLA.training.train_starvla][INFO] - Step 4843, Loss: {'train/loss': 0.3120911121368408, 'train/grad_norm_pre_clip': 22.352502822875977, 'timing/data': 0.009522473439574242, 'timing/model': 1.1390771539881825, 'learning_rate/qwen_vl_interface': 2.612390205623842e-06, 'learning_rate/action_model': 2.612390205623842e-05, 'epoch': 1.99}) +[2026-06-18 00:14:09,861][starVLA.training.train_starvla][INFO] - Step 4844, Loss: {'train/loss': 0.335805207490921, 'train/grad_norm_pre_clip': 29.487489700317383, 'timing/data': 0.007789604365825653, 'timing/model': 0.9302037917077541, 'learning_rate/qwen_vl_interface': 2.6105918573758425e-06, 'learning_rate/action_model': 2.6105918573758427e-05, 'epoch': 1.99}) +[2026-06-18 00:14:13,669][starVLA.training.train_starvla][INFO] - Step 4845, Loss: {'train/loss': 0.21479928493499756, 'train/grad_norm_pre_clip': 18.14927864074707, 'timing/data': 0.011147430166602135, 'timing/model': 0.9285212559625506, 'learning_rate/qwen_vl_interface': 2.608794056279991e-06, 'learning_rate/action_model': 2.6087940562799906e-05, 'epoch': 1.99}) +[2026-06-18 00:14:17,604][starVLA.training.train_starvla][INFO] - Step 4846, Loss: {'train/loss': 0.3276731073856354, 'train/grad_norm_pre_clip': 23.00541114807129, 'timing/data': 0.009421495720744133, 'timing/model': 0.9389123180881143, 'learning_rate/qwen_vl_interface': 2.6069968027089744e-06, 'learning_rate/action_model': 2.6069968027089742e-05, 'epoch': 1.99}) +[2026-06-18 00:14:21,419][starVLA.training.train_starvla][INFO] - Step 4847, Loss: {'train/loss': 0.4405258297920227, 'train/grad_norm_pre_clip': 19.75080680847168, 'timing/data': 0.012694668024778366, 'timing/model': 0.9633573750033975, 'learning_rate/qwen_vl_interface': 2.605200097035362e-06, 'learning_rate/action_model': 2.605200097035362e-05, 'epoch': 2.0}) +[2026-06-18 00:14:25,396][starVLA.training.train_starvla][INFO] - Step 4848, Loss: {'train/loss': 0.4886188209056854, 'train/grad_norm_pre_clip': 37.083213806152344, 'timing/data': 0.010211016982793808, 'timing/model': 1.0514686778187752, 'learning_rate/qwen_vl_interface': 2.6034039396316167e-06, 'learning_rate/action_model': 2.6034039396316167e-05, 'epoch': 2.0}) +[2026-06-18 00:14:29,320][starVLA.training.train_starvla][INFO] - Step 4849, Loss: {'train/loss': 0.2338639348745346, 'train/grad_norm_pre_clip': 18.349279403686523, 'timing/data': 0.013168378733098507, 'timing/model': 1.1358978636562824, 'learning_rate/qwen_vl_interface': 2.601608330870083e-06, 'learning_rate/action_model': 2.601608330870083e-05, 'epoch': 2.0}) +[2026-06-18 00:14:32,981][starVLA.training.train_starvla][INFO] - Step 4850, Loss: {'train/loss': 0.5401260256767273, 'train/grad_norm_pre_clip': 24.178686141967773, 'timing/data': 0.011330652981996536, 'timing/model': 0.9437287831678987, 'learning_rate/qwen_vl_interface': 2.5998132711229918e-06, 'learning_rate/action_model': 2.5998132711229915e-05, 'epoch': 2.0}) +[2026-06-18 00:14:38,705][starVLA.training.train_starvla][INFO] - Step 4851, Loss: {'train/loss': 0.31410714983940125, 'train/grad_norm_pre_clip': 19.294612884521484, 'timing/data': 0.01586511079221964, 'timing/model': 0.9555998733267188, 'learning_rate/qwen_vl_interface': 2.598018760762461e-06, 'learning_rate/action_model': 2.598018760762461e-05, 'epoch': 2.0}) +[2026-06-18 00:14:42,619][starVLA.training.train_starvla][INFO] - Step 4852, Loss: {'train/loss': 0.44634872674942017, 'train/grad_norm_pre_clip': 19.14067840576172, 'timing/data': 0.011612930335104465, 'timing/model': 0.9377861255779862, 'learning_rate/qwen_vl_interface': 2.5962248001604958e-06, 'learning_rate/action_model': 2.5962248001604956e-05, 'epoch': 2.0}) +[2026-06-18 00:14:46,528][starVLA.training.train_starvla][INFO] - Step 4853, Loss: {'train/loss': 0.17054347693920135, 'train/grad_norm_pre_clip': 15.597039222717285, 'timing/data': 0.0092422254383564, 'timing/model': 0.9475578842684627, 'learning_rate/qwen_vl_interface': 2.5944313896889837e-06, 'learning_rate/action_model': 2.594431389688984e-05, 'epoch': 2.0}) +[2026-06-18 00:14:50,378][starVLA.training.train_starvla][INFO] - Step 4854, Loss: {'train/loss': 0.41625961661338806, 'train/grad_norm_pre_clip': 29.92030906677246, 'timing/data': 0.018048872239887714, 'timing/model': 0.9371597561985254, 'learning_rate/qwen_vl_interface': 2.592638529719703e-06, 'learning_rate/action_model': 2.5926385297197033e-05, 'epoch': 2.0}) +[2026-06-18 00:14:54,869][starVLA.training.train_starvla][INFO] - Step 4855, Loss: {'train/loss': 0.2190559059381485, 'train/grad_norm_pre_clip': 18.420394897460938, 'timing/data': 0.009804883040487766, 'timing/model': 1.5799057502299547, 'learning_rate/qwen_vl_interface': 2.590846220624314e-06, 'learning_rate/action_model': 2.5908462206243138e-05, 'epoch': 2.0}) +[2026-06-18 00:14:58,737][starVLA.training.train_starvla][INFO] - Step 4856, Loss: {'train/loss': 0.3168827295303345, 'train/grad_norm_pre_clip': 19.137920379638672, 'timing/data': 0.00936324056237936, 'timing/model': 1.0973693942651153, 'learning_rate/qwen_vl_interface': 2.5890544627743652e-06, 'learning_rate/action_model': 2.589054462774365e-05, 'epoch': 2.0}) +[2026-06-18 00:15:02,454][starVLA.training.train_starvla][INFO] - Step 4857, Loss: {'train/loss': 0.26451727747917175, 'train/grad_norm_pre_clip': 19.49901008605957, 'timing/data': 0.008162853308022022, 'timing/model': 0.9346477594226599, 'learning_rate/qwen_vl_interface': 2.58726325654129e-06, 'learning_rate/action_model': 2.5872632565412898e-05, 'epoch': 2.0}) +[2026-06-18 00:15:06,491][starVLA.training.train_starvla][INFO] - Step 4858, Loss: {'train/loss': 0.5192708969116211, 'train/grad_norm_pre_clip': 21.671396255493164, 'timing/data': 0.01010863110423088, 'timing/model': 0.9540001209825277, 'learning_rate/qwen_vl_interface': 2.5854726022964043e-06, 'learning_rate/action_model': 2.5854726022964042e-05, 'epoch': 2.0}) +[2026-06-18 00:15:09,854][starVLA.training.train_starvla][INFO] - Step 4859, Loss: {'train/loss': 0.2770906388759613, 'train/grad_norm_pre_clip': 36.86980056762695, 'timing/data': 0.0006252732127904892, 'timing/model': 0.49332719296216965, 'learning_rate/qwen_vl_interface': 2.5836825004109173e-06, 'learning_rate/action_model': 2.5836825004109172e-05, 'epoch': 2.0}) +[2026-06-18 00:15:14,040][starVLA.training.train_starvla][INFO] - Step 4860, Loss: {'train/loss': 0.3999808728694916, 'train/grad_norm_pre_clip': 21.85989761352539, 'timing/data': 0.008713683113455772, 'timing/model': 0.9385048588737845, 'learning_rate/qwen_vl_interface': 2.5818929512559193e-06, 'learning_rate/action_model': 2.5818929512559193e-05, 'epoch': 2.0}) +[2026-06-18 00:15:17,886][starVLA.training.train_starvla][INFO] - Step 4861, Loss: {'train/loss': 0.10784187912940979, 'train/grad_norm_pre_clip': 9.410011291503906, 'timing/data': 0.009253300726413727, 'timing/model': 0.9412314239889383, 'learning_rate/qwen_vl_interface': 2.580103955202381e-06, 'learning_rate/action_model': 2.580103955202381e-05, 'epoch': 2.0}) +[2026-06-18 00:15:21,705][starVLA.training.train_starvla][INFO] - Step 4862, Loss: {'train/loss': 0.3151862323284149, 'train/grad_norm_pre_clip': 16.027019500732422, 'timing/data': 0.014079741202294827, 'timing/model': 1.1156062744557858, 'learning_rate/qwen_vl_interface': 2.5783155126211647e-06, 'learning_rate/action_model': 2.5783155126211644e-05, 'epoch': 2.0}) +[2026-06-18 00:15:25,327][starVLA.training.train_starvla][INFO] - Step 4863, Loss: {'train/loss': 0.22552679479122162, 'train/grad_norm_pre_clip': 15.636479377746582, 'timing/data': 0.01680732239037752, 'timing/model': 0.9361619921401143, 'learning_rate/qwen_vl_interface': 2.576527623883019e-06, 'learning_rate/action_model': 2.576527623883019e-05, 'epoch': 2.0}) +[2026-06-18 00:15:29,131][starVLA.training.train_starvla][INFO] - Step 4864, Loss: {'train/loss': 0.4252130389213562, 'train/grad_norm_pre_clip': 22.126569747924805, 'timing/data': 0.01059443037956953, 'timing/model': 0.9309462932869792, 'learning_rate/qwen_vl_interface': 2.5747402893585732e-06, 'learning_rate/action_model': 2.574740289358573e-05, 'epoch': 2.0}) +[2026-06-18 00:15:32,936][starVLA.training.train_starvla][INFO] - Step 4865, Loss: {'train/loss': 0.2225438952445984, 'train/grad_norm_pre_clip': 13.015035629272461, 'timing/data': 0.008332796394824982, 'timing/model': 0.9311403371393681, 'learning_rate/qwen_vl_interface': 2.5729535094183445e-06, 'learning_rate/action_model': 2.5729535094183444e-05, 'epoch': 2.0}) +[2026-06-18 00:15:36,744][starVLA.training.train_starvla][INFO] - Step 4866, Loss: {'train/loss': 0.4570710361003876, 'train/grad_norm_pre_clip': 23.63169288635254, 'timing/data': 0.01039785984903574, 'timing/model': 0.9391522789373994, 'learning_rate/qwen_vl_interface': 2.571167284432734e-06, 'learning_rate/action_model': 2.5711672844327338e-05, 'epoch': 2.0}) +[2026-06-18 00:15:40,561][starVLA.training.train_starvla][INFO] - Step 4867, Loss: {'train/loss': 0.3876130282878876, 'train/grad_norm_pre_clip': 17.23911476135254, 'timing/data': 0.015074514783918858, 'timing/model': 0.9315246334299445, 'learning_rate/qwen_vl_interface': 2.569381614772028e-06, 'learning_rate/action_model': 2.5693816147720278e-05, 'epoch': 2.0}) +[2026-06-18 00:15:44,404][starVLA.training.train_starvla][INFO] - Step 4868, Loss: {'train/loss': 0.0707230195403099, 'train/grad_norm_pre_clip': 5.912067890167236, 'timing/data': 0.015161995775997639, 'timing/model': 1.1261937897652388, 'learning_rate/qwen_vl_interface': 2.5675965008063976e-06, 'learning_rate/action_model': 2.5675965008063974e-05, 'epoch': 2.0}) +[2026-06-18 00:15:48,063][starVLA.training.train_starvla][INFO] - Step 4869, Loss: {'train/loss': 0.4105246961116791, 'train/grad_norm_pre_clip': 18.920591354370117, 'timing/data': 0.015025419183075428, 'timing/model': 0.9501253897324204, 'learning_rate/qwen_vl_interface': 2.565811942905899e-06, 'learning_rate/action_model': 2.5658119429058993e-05, 'epoch': 2.0}) +[2026-06-18 00:15:51,912][starVLA.training.train_starvla][INFO] - Step 4870, Loss: {'train/loss': 0.20595593750476837, 'train/grad_norm_pre_clip': 16.01044464111328, 'timing/data': 0.01027043629437685, 'timing/model': 0.9293934730812907, 'learning_rate/qwen_vl_interface': 2.5640279414404723e-06, 'learning_rate/action_model': 2.5640279414404722e-05, 'epoch': 2.0}) +[2026-06-18 00:15:55,666][starVLA.training.train_starvla][INFO] - Step 4871, Loss: {'train/loss': 0.5171462893486023, 'train/grad_norm_pre_clip': 26.951929092407227, 'timing/data': 0.00839314330369234, 'timing/model': 0.9250214798375964, 'learning_rate/qwen_vl_interface': 2.5622444967799444e-06, 'learning_rate/action_model': 2.5622444967799443e-05, 'epoch': 2.01}) +[2026-06-18 00:15:59,419][starVLA.training.train_starvla][INFO] - Step 4872, Loss: {'train/loss': 0.29683205485343933, 'train/grad_norm_pre_clip': 17.251497268676758, 'timing/data': 0.01028460543602705, 'timing/model': 0.9205046333372593, 'learning_rate/qwen_vl_interface': 2.5604616092940245e-06, 'learning_rate/action_model': 2.5604616092940242e-05, 'epoch': 2.01}) +[2026-06-18 00:16:03,191][starVLA.training.train_starvla][INFO] - Step 4873, Loss: {'train/loss': 0.1754256784915924, 'train/grad_norm_pre_clip': 17.992416381835938, 'timing/data': 0.007833949290215969, 'timing/model': 0.931255679577589, 'learning_rate/qwen_vl_interface': 2.5586792793523052e-06, 'learning_rate/action_model': 2.5586792793523052e-05, 'epoch': 2.01}) +[2026-06-18 00:16:06,977][starVLA.training.train_starvla][INFO] - Step 4874, Loss: {'train/loss': 0.3873739540576935, 'train/grad_norm_pre_clip': 362.6335754394531, 'timing/data': 0.01033813040703535, 'timing/model': 0.9261161331087351, 'learning_rate/qwen_vl_interface': 2.556897507324271e-06, 'learning_rate/action_model': 2.5568975073242705e-05, 'epoch': 2.01}) +[2026-06-18 00:16:10,911][starVLA.training.train_starvla][INFO] - Step 4875, Loss: {'train/loss': 0.143648162484169, 'train/grad_norm_pre_clip': 12.400696754455566, 'timing/data': 0.010038158856332302, 'timing/model': 1.2226917585358024, 'learning_rate/qwen_vl_interface': 2.5551162935792782e-06, 'learning_rate/action_model': 2.5551162935792784e-05, 'epoch': 2.01}) +[2026-06-18 00:16:14,571][starVLA.training.train_starvla][INFO] - Step 4876, Loss: {'train/loss': 0.32741621136665344, 'train/grad_norm_pre_clip': 25.817623138427734, 'timing/data': 0.010617614723742008, 'timing/model': 0.9406513664871454, 'learning_rate/qwen_vl_interface': 2.5533356384865777e-06, 'learning_rate/action_model': 2.5533356384865776e-05, 'epoch': 2.01}) +[2026-06-18 00:16:18,721][starVLA.training.train_starvla][INFO] - Step 4877, Loss: {'train/loss': 0.2899690270423889, 'train/grad_norm_pre_clip': 29.504915237426758, 'timing/data': 0.012411939911544323, 'timing/model': 1.0593210300430655, 'learning_rate/qwen_vl_interface': 2.5515555424153017e-06, 'learning_rate/action_model': 2.5515555424153016e-05, 'epoch': 2.01}) +[2026-06-18 00:16:22,576][starVLA.training.train_starvla][INFO] - Step 4878, Loss: {'train/loss': 0.18957161903381348, 'train/grad_norm_pre_clip': 14.012595176696777, 'timing/data': 0.011836802586913109, 'timing/model': 0.9458585130050778, 'learning_rate/qwen_vl_interface': 2.5497760057344663e-06, 'learning_rate/action_model': 2.549776005734466e-05, 'epoch': 2.01}) +[2026-06-18 00:16:26,458][starVLA.training.train_starvla][INFO] - Step 4879, Loss: {'train/loss': 0.48104408383369446, 'train/grad_norm_pre_clip': 27.75261878967285, 'timing/data': 0.012543018907308578, 'timing/model': 0.939922702498734, 'learning_rate/qwen_vl_interface': 2.5479970288129714e-06, 'learning_rate/action_model': 2.5479970288129714e-05, 'epoch': 2.01}) +[2026-06-18 00:16:30,340][starVLA.training.train_starvla][INFO] - Step 4880, Loss: {'train/loss': 0.4397627115249634, 'train/grad_norm_pre_clip': 23.424488067626953, 'timing/data': 0.007847343571484089, 'timing/model': 0.9412281839177012, 'learning_rate/qwen_vl_interface': 2.546218612019597e-06, 'learning_rate/action_model': 2.5462186120195967e-05, 'epoch': 2.01}) +[2026-06-18 00:16:34,384][starVLA.training.train_starvla][INFO] - Step 4881, Loss: {'train/loss': 0.3347131609916687, 'train/grad_norm_pre_clip': 21.858644485473633, 'timing/data': 0.01007922925055027, 'timing/model': 1.3536679418757558, 'learning_rate/qwen_vl_interface': 2.544440755723015e-06, 'learning_rate/action_model': 2.5444407557230147e-05, 'epoch': 2.01}) +[2026-06-18 00:16:38,254][starVLA.training.train_starvla][INFO] - Step 4882, Loss: {'train/loss': 0.12462904304265976, 'train/grad_norm_pre_clip': 10.520896911621094, 'timing/data': 0.011976122856140137, 'timing/model': 1.1321769701316953, 'learning_rate/qwen_vl_interface': 2.5426634602917756e-06, 'learning_rate/action_model': 2.5426634602917755e-05, 'epoch': 2.01}) +[2026-06-18 00:16:41,925][starVLA.training.train_starvla][INFO] - Step 4883, Loss: {'train/loss': 0.19114643335342407, 'train/grad_norm_pre_clip': 17.858844757080078, 'timing/data': 0.009794897399842739, 'timing/model': 0.9611454419791698, 'learning_rate/qwen_vl_interface': 2.540886726094315e-06, 'learning_rate/action_model': 2.540886726094315e-05, 'epoch': 2.01}) +[2026-06-18 00:16:45,755][starVLA.training.train_starvla][INFO] - Step 4884, Loss: {'train/loss': 0.20047807693481445, 'train/grad_norm_pre_clip': 20.182395935058594, 'timing/data': 0.00850957352668047, 'timing/model': 0.9565529329702258, 'learning_rate/qwen_vl_interface': 2.5391105534989503e-06, 'learning_rate/action_model': 2.53911055349895e-05, 'epoch': 2.01}) +[2026-06-18 00:16:49,588][starVLA.training.train_starvla][INFO] - Step 4885, Loss: {'train/loss': 0.5022411346435547, 'train/grad_norm_pre_clip': 23.285696029663086, 'timing/data': 0.011000469326972961, 'timing/model': 0.9427889687940478, 'learning_rate/qwen_vl_interface': 2.537334942873885e-06, 'learning_rate/action_model': 2.5373349428738846e-05, 'epoch': 2.01}) +[2026-06-18 00:16:53,504][starVLA.training.train_starvla][INFO] - Step 4886, Loss: {'train/loss': 0.22287435829639435, 'train/grad_norm_pre_clip': 16.629112243652344, 'timing/data': 0.008082002401351929, 'timing/model': 0.9514966001734138, 'learning_rate/qwen_vl_interface': 2.535559894587205e-06, 'learning_rate/action_model': 2.535559894587205e-05, 'epoch': 2.01}) +[2026-06-18 00:16:57,293][starVLA.training.train_starvla][INFO] - Step 4887, Loss: {'train/loss': 0.6991060376167297, 'train/grad_norm_pre_clip': 27.20548439025879, 'timing/data': 0.008569599129259586, 'timing/model': 0.9263238245621324, 'learning_rate/qwen_vl_interface': 2.533785409006878e-06, 'learning_rate/action_model': 2.533785409006878e-05, 'epoch': 2.01}) +[2026-06-18 00:17:01,016][starVLA.training.train_starvla][INFO] - Step 4888, Loss: {'train/loss': 0.3003704249858856, 'train/grad_norm_pre_clip': 18.089508056640625, 'timing/data': 0.010625957511365414, 'timing/model': 1.0824196301400661, 'learning_rate/qwen_vl_interface': 2.532011486500762e-06, 'learning_rate/action_model': 2.532011486500762e-05, 'epoch': 2.01}) +[2026-06-18 00:17:04,784][starVLA.training.train_starvla][INFO] - Step 4889, Loss: {'train/loss': 0.5446994304656982, 'train/grad_norm_pre_clip': 24.26805305480957, 'timing/data': 0.012240630574524403, 'timing/model': 1.0922830924391747, 'learning_rate/qwen_vl_interface': 2.530238127436588e-06, 'learning_rate/action_model': 2.5302381274365882e-05, 'epoch': 2.01}) +[2026-06-18 00:17:08,383][starVLA.training.train_starvla][INFO] - Step 4890, Loss: {'train/loss': 0.35756391286849976, 'train/grad_norm_pre_clip': 18.28455352783203, 'timing/data': 0.009119817987084389, 'timing/model': 0.9349456513300538, 'learning_rate/qwen_vl_interface': 2.528465332181977e-06, 'learning_rate/action_model': 2.5284653321819767e-05, 'epoch': 2.01}) +[2026-06-18 00:17:12,338][starVLA.training.train_starvla][INFO] - Step 4891, Loss: {'train/loss': 0.20878352224826813, 'train/grad_norm_pre_clip': 17.832075119018555, 'timing/data': 0.011281142011284828, 'timing/model': 0.9461398087441921, 'learning_rate/qwen_vl_interface': 2.526693101104429e-06, 'learning_rate/action_model': 2.526693101104429e-05, 'epoch': 2.01}) +[2026-06-18 00:17:16,098][starVLA.training.train_starvla][INFO] - Step 4892, Loss: {'train/loss': 0.41877636313438416, 'train/grad_norm_pre_clip': 27.511125564575195, 'timing/data': 0.008003599010407925, 'timing/model': 0.9171177241951227, 'learning_rate/qwen_vl_interface': 2.524921434571333e-06, 'learning_rate/action_model': 2.524921434571333e-05, 'epoch': 2.01}) +[2026-06-18 00:17:19,771][starVLA.training.train_starvla][INFO] - Step 4893, Loss: {'train/loss': 0.2796109914779663, 'train/grad_norm_pre_clip': 14.8623046875, 'timing/data': 0.01602451130747795, 'timing/model': 0.9259988600388169, 'learning_rate/qwen_vl_interface': 2.5231503329499585e-06, 'learning_rate/action_model': 2.5231503329499583e-05, 'epoch': 2.01}) +[2026-06-18 00:17:23,500][starVLA.training.train_starvla][INFO] - Step 4894, Loss: {'train/loss': 0.3512352406978607, 'train/grad_norm_pre_clip': 11.850303649902344, 'timing/data': 0.007414240390062332, 'timing/model': 0.935875890776515, 'learning_rate/qwen_vl_interface': 2.5213797966074506e-06, 'learning_rate/action_model': 2.5213797966074505e-05, 'epoch': 2.01}) +[2026-06-18 00:17:27,283][starVLA.training.train_starvla][INFO] - Step 4895, Loss: {'train/loss': 0.5408474802970886, 'train/grad_norm_pre_clip': 26.940120697021484, 'timing/data': 0.010004150681197643, 'timing/model': 1.120028511621058, 'learning_rate/qwen_vl_interface': 2.519609825910848e-06, 'learning_rate/action_model': 2.5196098259108482e-05, 'epoch': 2.01}) +[2026-06-18 00:17:30,869][starVLA.training.train_starvla][INFO] - Step 4896, Loss: {'train/loss': 0.26121628284454346, 'train/grad_norm_pre_clip': 17.536190032958984, 'timing/data': 0.009488663636147976, 'timing/model': 0.9287088708952069, 'learning_rate/qwen_vl_interface': 2.517840421227067e-06, 'learning_rate/action_model': 2.517840421227067e-05, 'epoch': 2.02}) +[2026-06-18 00:17:34,678][starVLA.training.train_starvla][INFO] - Step 4897, Loss: {'train/loss': 0.29962781071662903, 'train/grad_norm_pre_clip': 24.93260383605957, 'timing/data': 0.01019094418734312, 'timing/model': 0.9312608828768134, 'learning_rate/qwen_vl_interface': 2.5160715829229067e-06, 'learning_rate/action_model': 2.5160715829229066e-05, 'epoch': 2.02}) +[2026-06-18 00:17:38,438][starVLA.training.train_starvla][INFO] - Step 4898, Loss: {'train/loss': 0.21392744779586792, 'train/grad_norm_pre_clip': 23.846467971801758, 'timing/data': 0.009415261447429657, 'timing/model': 0.929758021607995, 'learning_rate/qwen_vl_interface': 2.5143033113650484e-06, 'learning_rate/action_model': 2.5143033113650483e-05, 'epoch': 2.02}) +[2026-06-18 00:17:42,281][starVLA.training.train_starvla][INFO] - Step 4899, Loss: {'train/loss': 0.35747870802879333, 'train/grad_norm_pre_clip': 18.466999053955078, 'timing/data': 0.00909543875604868, 'timing/model': 0.9653867892920971, 'learning_rate/qwen_vl_interface': 2.5125356069200572e-06, 'learning_rate/action_model': 2.5125356069200574e-05, 'epoch': 2.02}) +[2026-06-18 00:19:37,886][starVLA.training.train_starvla][INFO] - Step 4900, Loss: {'train/loss': 0.28202497959136963, 'train/grad_norm_pre_clip': 26.09381675720215, 'eval/loss': 0.4191906750202179, 'eval/latency_6/loss': 0.4191906750202179, 'eval/demon_fire_precision': 0.9547325102880658, 'eval/demon_fire_recall': 0.9747899159663865, 'eval/demon_fire_f1': 0.9646569646569646, 'eval/demon_fire_rate_model': 0.8789649415692822, 'eval/demon_fire_rate_teacher': 0.860879243183083, 'eval/demon_left_precision': 0.8278388278388278, 'eval/demon_left_recall': 0.7929824561403509, 'eval/demon_left_f1': 0.8100358422939069, 'eval/demon_left_rate_model': 0.15191986644407346, 'eval/demon_left_rate_teacher': 0.15859766277128548, 'eval/demon_right_precision': 0.8109375, 'eval/demon_right_recall': 0.7972350230414746, 'eval/demon_right_f1': 0.8040278853601859, 'eval/demon_right_rate_model': 0.17807456872565386, 'eval/demon_right_rate_teacher': 0.18113522537562604, 'eval/demon_move_f1': 0.8070318638270464, 'eval/demon_cc_f1': 0.9112621269505857, 'eval/latency_6/demon_fire_precision': 0.9547325102880658, 'eval/latency_6/demon_fire_recall': 0.9747899159663865, 'eval/latency_6/demon_fire_f1': 0.9646569646569646, 'eval/latency_6/demon_fire_rate_model': 0.8789649415692822, 'eval/latency_6/demon_fire_rate_teacher': 0.860879243183083, 'eval/latency_6/demon_left_precision': 0.8278388278388278, 'eval/latency_6/demon_left_recall': 0.7929824561403509, 'eval/latency_6/demon_left_f1': 0.8100358422939069, 'eval/latency_6/demon_left_rate_model': 0.15191986644407346, 'eval/latency_6/demon_left_rate_teacher': 0.15859766277128548, 'eval/latency_6/demon_right_precision': 0.8109375, 'eval/latency_6/demon_right_recall': 0.7972350230414746, 'eval/latency_6/demon_right_f1': 0.8040278853601859, 'eval/latency_6/demon_right_rate_model': 0.17807456872565386, 'eval/latency_6/demon_right_rate_teacher': 0.18113522537562604, 'eval/latency_6/demon_move_f1': 0.8070318638270464, 'eval/latency_6/demon_cc_f1': 0.9112621269505857, 'timing/data': 0.009446310810744762, 'timing/model': 0.9270577542483807, 'learning_rate/qwen_vl_interface': 2.5107684699543817e-06, 'learning_rate/action_model': 2.5107684699543816e-05, 'epoch': 2.02}) +[2026-06-18 00:19:37,892][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_4900_state +[2026-06-18 00:19:37,897][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 00:19:44,680][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_4900_state/model.safetensors +[2026-06-18 00:20:00,421][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_4900_state/optimizer.bin +[2026-06-18 00:20:00,422][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_4900_state/sampler.bin +[2026-06-18 00:20:00,422][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_4900_state/sampler_1.bin +[2026-06-18 00:20:00,425][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_4900_state/random_states_0.pkl +[2026-06-18 00:20:00,428][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4900 +[2026-06-18 00:20:00,428][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 00:20:00,440][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 00:20:04,536][starVLA.training.train_starvla][INFO] - Step 4901, Loss: {'train/loss': 0.3513677716255188, 'train/grad_norm_pre_clip': 24.884517669677734, 'timing/data': 0.00945610273629427, 'timing/model': 0.9335057819262147, 'learning_rate/qwen_vl_interface': 2.509001900834349e-06, 'learning_rate/action_model': 2.5090019008343492e-05, 'epoch': 2.02}) +[2026-06-18 00:20:08,711][starVLA.training.train_starvla][INFO] - Step 4902, Loss: {'train/loss': 0.31336459517478943, 'train/grad_norm_pre_clip': 23.6854190826416, 'timing/data': 0.010458293370902538, 'timing/model': 0.9571311427280307, 'learning_rate/qwen_vl_interface': 2.507235899926172e-06, 'learning_rate/action_model': 2.5072358999261718e-05, 'epoch': 2.02}) +[2026-06-18 00:20:12,795][starVLA.training.train_starvla][INFO] - Step 4903, Loss: {'train/loss': 0.39267948269844055, 'train/grad_norm_pre_clip': 21.044950485229492, 'timing/data': 0.009786512702703476, 'timing/model': 0.9610884413123131, 'learning_rate/qwen_vl_interface': 2.5054704675959445e-06, 'learning_rate/action_model': 2.5054704675959445e-05, 'epoch': 2.02}) +[2026-06-18 00:20:16,702][starVLA.training.train_starvla][INFO] - Step 4904, Loss: {'train/loss': 0.3331713080406189, 'train/grad_norm_pre_clip': 15.385454177856445, 'timing/data': 0.008746258914470673, 'timing/model': 1.1779949804767966, 'learning_rate/qwen_vl_interface': 2.5037056042096418e-06, 'learning_rate/action_model': 2.5037056042096417e-05, 'epoch': 2.02}) +[2026-06-18 00:20:20,312][starVLA.training.train_starvla][INFO] - Step 4905, Loss: {'train/loss': 0.3541485667228699, 'train/grad_norm_pre_clip': 58.39942169189453, 'timing/data': 0.011432242579758167, 'timing/model': 0.9306509010493755, 'learning_rate/qwen_vl_interface': 2.5019413101331212e-06, 'learning_rate/action_model': 2.5019413101331214e-05, 'epoch': 2.02}) +[2026-06-18 00:20:24,174][starVLA.training.train_starvla][INFO] - Step 4906, Loss: {'train/loss': 0.18894149363040924, 'train/grad_norm_pre_clip': 15.1376371383667, 'timing/data': 0.012704170309007168, 'timing/model': 0.930338972248137, 'learning_rate/qwen_vl_interface': 2.5001775857321254e-06, 'learning_rate/action_model': 2.5001775857321253e-05, 'epoch': 2.02}) +[2026-06-18 00:20:27,972][starVLA.training.train_starvla][INFO] - Step 4907, Loss: {'train/loss': 0.4226940870285034, 'train/grad_norm_pre_clip': 29.483644485473633, 'timing/data': 0.016902371309697628, 'timing/model': 0.9325247090309858, 'learning_rate/qwen_vl_interface': 2.4984144313722765e-06, 'learning_rate/action_model': 2.4984144313722766e-05, 'epoch': 2.02}) +[2026-06-18 00:20:31,931][starVLA.training.train_starvla][INFO] - Step 4908, Loss: {'train/loss': 0.1329881250858307, 'train/grad_norm_pre_clip': 16.23939323425293, 'timing/data': 0.010637159459292889, 'timing/model': 0.9360104780644178, 'learning_rate/qwen_vl_interface': 2.496651847419076e-06, 'learning_rate/action_model': 2.496651847419076e-05, 'epoch': 2.02}) +[2026-06-18 00:20:35,744][starVLA.training.train_starvla][INFO] - Step 4909, Loss: {'train/loss': 0.1741713285446167, 'train/grad_norm_pre_clip': 16.105728149414062, 'timing/data': 0.00887379702180624, 'timing/model': 0.9335318095982075, 'learning_rate/qwen_vl_interface': 2.4948898342379076e-06, 'learning_rate/action_model': 2.4948898342379075e-05, 'epoch': 2.02}) +[2026-06-18 00:20:39,573][starVLA.training.train_starvla][INFO] - Step 4910, Loss: {'train/loss': 0.35027673840522766, 'train/grad_norm_pre_clip': 26.31063461303711, 'timing/data': 0.010948664508759975, 'timing/model': 0.9318295530974865, 'learning_rate/qwen_vl_interface': 2.4931283921940434e-06, 'learning_rate/action_model': 2.4931283921940434e-05, 'epoch': 2.02}) +[2026-06-18 00:20:43,382][starVLA.training.train_starvla][INFO] - Step 4911, Loss: {'train/loss': 0.3656626343727112, 'train/grad_norm_pre_clip': 20.651342391967773, 'timing/data': 0.009802090004086494, 'timing/model': 1.1327952044084668, 'learning_rate/qwen_vl_interface': 2.4913675216526302e-06, 'learning_rate/action_model': 2.4913675216526304e-05, 'epoch': 2.02}) +[2026-06-18 00:20:46,998][starVLA.training.train_starvla][INFO] - Step 4912, Loss: {'train/loss': 0.23685963451862335, 'train/grad_norm_pre_clip': 16.462438583374023, 'timing/data': 0.009793966077268124, 'timing/model': 0.9439523760229349, 'learning_rate/qwen_vl_interface': 2.4896072229786996e-06, 'learning_rate/action_model': 2.4896072229786995e-05, 'epoch': 2.02}) +[2026-06-18 00:20:50,823][starVLA.training.train_starvla][INFO] - Step 4913, Loss: {'train/loss': 0.3845883905887604, 'train/grad_norm_pre_clip': 15.634177207946777, 'timing/data': 0.010134988464415073, 'timing/model': 0.9366987952962518, 'learning_rate/qwen_vl_interface': 2.4878474965371587e-06, 'learning_rate/action_model': 2.4878474965371586e-05, 'epoch': 2.02}) +[2026-06-18 00:20:54,775][starVLA.training.train_starvla][INFO] - Step 4914, Loss: {'train/loss': 0.2389332354068756, 'train/grad_norm_pre_clip': 19.51104736328125, 'timing/data': 0.011892705224454403, 'timing/model': 0.939986945129931, 'learning_rate/qwen_vl_interface': 2.4860883426928053e-06, 'learning_rate/action_model': 2.4860883426928053e-05, 'epoch': 2.02}) +[2026-06-18 00:20:58,694][starVLA.training.train_starvla][INFO] - Step 4915, Loss: {'train/loss': 0.3774605989456177, 'train/grad_norm_pre_clip': 21.797277450561523, 'timing/data': 0.008950880728662014, 'timing/model': 0.9476578198373318, 'learning_rate/qwen_vl_interface': 2.4843297618103127e-06, 'learning_rate/action_model': 2.4843297618103128e-05, 'epoch': 2.02}) +[2026-06-18 00:21:02,592][starVLA.training.train_starvla][INFO] - Step 4916, Loss: {'train/loss': 0.48005160689353943, 'train/grad_norm_pre_clip': 26.60361099243164, 'timing/data': 0.02448124159127474, 'timing/model': 0.9870750065892935, 'learning_rate/qwen_vl_interface': 2.482571754254237e-06, 'learning_rate/action_model': 2.4825717542542372e-05, 'epoch': 2.02}) +[2026-06-18 00:21:06,418][starVLA.training.train_starvla][INFO] - Step 4917, Loss: {'train/loss': 0.4133797585964203, 'train/grad_norm_pre_clip': 20.776254653930664, 'timing/data': 0.010498651303350925, 'timing/model': 1.1210221024230123, 'learning_rate/qwen_vl_interface': 2.4808143203890137e-06, 'learning_rate/action_model': 2.480814320389014e-05, 'epoch': 2.02}) +[2026-06-18 00:21:10,296][starVLA.training.train_starvla][INFO] - Step 4918, Loss: {'train/loss': 0.6251389980316162, 'train/grad_norm_pre_clip': 20.954544067382812, 'timing/data': 0.011262008920311928, 'timing/model': 1.146577519364655, 'learning_rate/qwen_vl_interface': 2.4790574605789627e-06, 'learning_rate/action_model': 2.4790574605789624e-05, 'epoch': 2.02}) +[2026-06-18 00:21:13,936][starVLA.training.train_starvla][INFO] - Step 4919, Loss: {'train/loss': 0.2063247114419937, 'train/grad_norm_pre_clip': 12.433439254760742, 'timing/data': 0.010729207657277584, 'timing/model': 0.9449323425069451, 'learning_rate/qwen_vl_interface': 2.4773011751882807e-06, 'learning_rate/action_model': 2.4773011751882804e-05, 'epoch': 2.02}) +[2026-06-18 00:21:17,753][starVLA.training.train_starvla][INFO] - Step 4920, Loss: {'train/loss': 0.3293110430240631, 'train/grad_norm_pre_clip': 19.824024200439453, 'timing/data': 0.013850996270775795, 'timing/model': 0.9281894797459245, 'learning_rate/qwen_vl_interface': 2.4755454645810473e-06, 'learning_rate/action_model': 2.4755454645810473e-05, 'epoch': 2.03}) +[2026-06-18 00:21:21,552][starVLA.training.train_starvla][INFO] - Step 4921, Loss: {'train/loss': 0.46641984581947327, 'train/grad_norm_pre_clip': 28.196533203125, 'timing/data': 0.008373452350497246, 'timing/model': 0.9323823098093271, 'learning_rate/qwen_vl_interface': 2.473790329121228e-06, 'learning_rate/action_model': 2.473790329121228e-05, 'epoch': 2.03}) +[2026-06-18 00:21:25,393][starVLA.training.train_starvla][INFO] - Step 4922, Loss: {'train/loss': 0.1703331470489502, 'train/grad_norm_pre_clip': 14.338003158569336, 'timing/data': 0.009848153218626976, 'timing/model': 0.9292179001495242, 'learning_rate/qwen_vl_interface': 2.4720357691726595e-06, 'learning_rate/action_model': 2.4720357691726594e-05, 'epoch': 2.03}) +[2026-06-18 00:21:29,200][starVLA.training.train_starvla][INFO] - Step 4923, Loss: {'train/loss': 0.19122040271759033, 'train/grad_norm_pre_clip': 17.17656898498535, 'timing/data': 0.010296687483787537, 'timing/model': 0.9330831225961447, 'learning_rate/qwen_vl_interface': 2.4702817850990645e-06, 'learning_rate/action_model': 2.470281785099064e-05, 'epoch': 2.03}) +[2026-06-18 00:21:33,010][starVLA.training.train_starvla][INFO] - Step 4924, Loss: {'train/loss': 0.3313007056713104, 'train/grad_norm_pre_clip': 17.870616912841797, 'timing/data': 0.011975889094173908, 'timing/model': 1.1353433579206467, 'learning_rate/qwen_vl_interface': 2.4685283772640452e-06, 'learning_rate/action_model': 2.468528377264045e-05, 'epoch': 2.03}) +[2026-06-18 00:21:36,801][starVLA.training.train_starvla][INFO] - Step 4925, Loss: {'train/loss': 0.08140279352664948, 'train/grad_norm_pre_clip': 8.06292724609375, 'timing/data': 0.008934317156672478, 'timing/model': 1.0989406192675233, 'learning_rate/qwen_vl_interface': 2.466775546031088e-06, 'learning_rate/action_model': 2.4667755460310878e-05, 'epoch': 2.03}) +[2026-06-18 00:21:40,408][starVLA.training.train_starvla][INFO] - Step 4926, Loss: {'train/loss': 0.36324602365493774, 'train/grad_norm_pre_clip': 19.314512252807617, 'timing/data': 0.009189982898533344, 'timing/model': 0.9368182476609945, 'learning_rate/qwen_vl_interface': 2.465023291763555e-06, 'learning_rate/action_model': 2.465023291763555e-05, 'epoch': 2.03}) +[2026-06-18 00:21:44,235][starVLA.training.train_starvla][INFO] - Step 4927, Loss: {'train/loss': 0.34588223695755005, 'train/grad_norm_pre_clip': 22.722307205200195, 'timing/data': 0.012832701206207275, 'timing/model': 0.9292927235364914, 'learning_rate/qwen_vl_interface': 2.4632716148246903e-06, 'learning_rate/action_model': 2.4632716148246903e-05, 'epoch': 2.03}) +[2026-06-18 00:21:48,051][starVLA.training.train_starvla][INFO] - Step 4928, Loss: {'train/loss': 0.40515321493148804, 'train/grad_norm_pre_clip': 24.83415985107422, 'timing/data': 0.009733164682984352, 'timing/model': 0.9295060094445944, 'learning_rate/qwen_vl_interface': 2.4615205155776185e-06, 'learning_rate/action_model': 2.4615205155776188e-05, 'epoch': 2.03}) +[2026-06-18 00:21:51,854][starVLA.training.train_starvla][INFO] - Step 4929, Loss: {'train/loss': 0.4432619512081146, 'train/grad_norm_pre_clip': 24.76892852783203, 'timing/data': 0.010470055975019932, 'timing/model': 0.9373705526813865, 'learning_rate/qwen_vl_interface': 2.4597699943853444e-06, 'learning_rate/action_model': 2.4597699943853442e-05, 'epoch': 2.03}) +[2026-06-18 00:21:55,680][starVLA.training.train_starvla][INFO] - Step 4930, Loss: {'train/loss': 0.14828258752822876, 'train/grad_norm_pre_clip': 11.252188682556152, 'timing/data': 0.014446821063756943, 'timing/model': 0.9315595962107182, 'learning_rate/qwen_vl_interface': 2.458020051610753e-06, 'learning_rate/action_model': 2.4580200516107528e-05, 'epoch': 2.03}) +[2026-06-18 00:21:59,514][starVLA.training.train_starvla][INFO] - Step 4931, Loss: {'train/loss': 0.29461976885795593, 'train/grad_norm_pre_clip': 20.479076385498047, 'timing/data': 0.01132930163294077, 'timing/model': 1.1374753899872303, 'learning_rate/qwen_vl_interface': 2.456270687616609e-06, 'learning_rate/action_model': 2.4562706876166087e-05, 'epoch': 2.03}) +[2026-06-18 00:22:03,111][starVLA.training.train_starvla][INFO] - Step 4932, Loss: {'train/loss': 0.17926155030727386, 'train/grad_norm_pre_clip': 107.59715270996094, 'timing/data': 0.011126317083835602, 'timing/model': 0.9304253049194813, 'learning_rate/qwen_vl_interface': 2.4545219027655577e-06, 'learning_rate/action_model': 2.4545219027655574e-05, 'epoch': 2.03}) +[2026-06-18 00:22:06,928][starVLA.training.train_starvla][INFO] - Step 4933, Loss: {'train/loss': 0.3039962351322174, 'train/grad_norm_pre_clip': 25.08295249938965, 'timing/data': 0.008179796859622002, 'timing/model': 0.9434575596824288, 'learning_rate/qwen_vl_interface': 2.452773697420124e-06, 'learning_rate/action_model': 2.452773697420124e-05, 'epoch': 2.03}) +[2026-06-18 00:22:10,722][starVLA.training.train_starvla][INFO] - Step 4934, Loss: {'train/loss': 0.3631121516227722, 'train/grad_norm_pre_clip': 25.87225914001465, 'timing/data': 0.011177323758602142, 'timing/model': 0.9607803132385015, 'learning_rate/qwen_vl_interface': 2.4510260719427127e-06, 'learning_rate/action_model': 2.4510260719427127e-05, 'epoch': 2.03}) +[2026-06-18 00:22:14,587][starVLA.training.train_starvla][INFO] - Step 4935, Loss: {'train/loss': 0.264825701713562, 'train/grad_norm_pre_clip': 20.067964553833008, 'timing/data': 0.0067555708810687065, 'timing/model': 0.9379018526524305, 'learning_rate/qwen_vl_interface': 2.4492790266956092e-06, 'learning_rate/action_model': 2.449279026695609e-05, 'epoch': 2.03}) +[2026-06-18 00:22:18,443][starVLA.training.train_starvla][INFO] - Step 4936, Loss: {'train/loss': 0.29588082432746887, 'train/grad_norm_pre_clip': 18.395368576049805, 'timing/data': 0.007162250578403473, 'timing/model': 0.9294512458145618, 'learning_rate/qwen_vl_interface': 2.4475325620409765e-06, 'learning_rate/action_model': 2.4475325620409764e-05, 'epoch': 2.03}) +[2026-06-18 00:22:22,272][starVLA.training.train_starvla][INFO] - Step 4937, Loss: {'train/loss': 0.33380910754203796, 'train/grad_norm_pre_clip': 21.843652725219727, 'timing/data': 0.009660206735134125, 'timing/model': 1.1423062831163406, 'learning_rate/qwen_vl_interface': 2.445786678340859e-06, 'learning_rate/action_model': 2.4457866783408588e-05, 'epoch': 2.03}) +[2026-06-18 00:22:26,093][starVLA.training.train_starvla][INFO] - Step 4938, Loss: {'train/loss': 0.2608117461204529, 'train/grad_norm_pre_clip': 20.112634658813477, 'timing/data': 0.013383644632995129, 'timing/model': 1.1344397133216262, 'learning_rate/qwen_vl_interface': 2.444041375957179e-06, 'learning_rate/action_model': 2.444041375957179e-05, 'epoch': 2.03}) +[2026-06-18 00:22:29,701][starVLA.training.train_starvla][INFO] - Step 4939, Loss: {'train/loss': 0.15763045847415924, 'train/grad_norm_pre_clip': 12.367403030395508, 'timing/data': 0.010386166162788868, 'timing/model': 0.9343423582613468, 'learning_rate/qwen_vl_interface': 2.4422966552517434e-06, 'learning_rate/action_model': 2.4422966552517433e-05, 'epoch': 2.03}) +[2026-06-18 00:22:33,496][starVLA.training.train_starvla][INFO] - Step 4940, Loss: {'train/loss': 0.3217644989490509, 'train/grad_norm_pre_clip': 19.650236129760742, 'timing/data': 0.007161954417824745, 'timing/model': 0.9201616756618023, 'learning_rate/qwen_vl_interface': 2.440552516586232e-06, 'learning_rate/action_model': 2.440552516586232e-05, 'epoch': 2.03}) +[2026-06-18 00:22:37,322][starVLA.training.train_starvla][INFO] - Step 4941, Loss: {'train/loss': 0.22685392200946808, 'train/grad_norm_pre_clip': 27.57233428955078, 'timing/data': 0.00776383001357317, 'timing/model': 0.9719179458916187, 'learning_rate/qwen_vl_interface': 2.4388089603222085e-06, 'learning_rate/action_model': 2.4388089603222087e-05, 'epoch': 2.03}) +[2026-06-18 00:22:41,238][starVLA.training.train_starvla][INFO] - Step 4942, Loss: {'train/loss': 0.32328373193740845, 'train/grad_norm_pre_clip': 20.055307388305664, 'timing/data': 0.019140096381306648, 'timing/model': 0.9500242732465267, 'learning_rate/qwen_vl_interface': 2.4370659868211095e-06, 'learning_rate/action_model': 2.4370659868211097e-05, 'epoch': 2.03}) +[2026-06-18 00:22:45,034][starVLA.training.train_starvla][INFO] - Step 4943, Loss: {'train/loss': 0.10501100867986679, 'train/grad_norm_pre_clip': 11.587410926818848, 'timing/data': 0.008761165663599968, 'timing/model': 0.9421258335933089, 'learning_rate/qwen_vl_interface': 2.4353235964442597e-06, 'learning_rate/action_model': 2.4353235964442598e-05, 'epoch': 2.03}) +[2026-06-18 00:22:48,912][starVLA.training.train_starvla][INFO] - Step 4944, Loss: {'train/loss': 0.21312081813812256, 'train/grad_norm_pre_clip': 17.589115142822266, 'timing/data': 0.012781958095729351, 'timing/model': 1.144859847612679, 'learning_rate/qwen_vl_interface': 2.433581789552857e-06, 'learning_rate/action_model': 2.433581789552857e-05, 'epoch': 2.04}) +[2026-06-18 00:22:52,773][starVLA.training.train_starvla][INFO] - Step 4945, Loss: {'train/loss': 0.3654301166534424, 'train/grad_norm_pre_clip': 22.49234962463379, 'timing/data': 0.007523643784224987, 'timing/model': 1.1350281918421388, 'learning_rate/qwen_vl_interface': 2.4318405665079806e-06, 'learning_rate/action_model': 2.4318405665079805e-05, 'epoch': 2.04}) +[2026-06-18 00:22:56,732][starVLA.training.train_starvla][INFO] - Step 4946, Loss: {'train/loss': 0.13201791048049927, 'train/grad_norm_pre_clip': 10.913992881774902, 'timing/data': 0.013814280740916729, 'timing/model': 1.118837321177125, 'learning_rate/qwen_vl_interface': 2.4300999276705868e-06, 'learning_rate/action_model': 2.4300999276705868e-05, 'epoch': 2.04}) +[2026-06-18 00:23:00,634][starVLA.training.train_starvla][INFO] - Step 4947, Loss: {'train/loss': 0.48317304253578186, 'train/grad_norm_pre_clip': 24.225263595581055, 'timing/data': 0.00994936190545559, 'timing/model': 0.9602119354531169, 'learning_rate/qwen_vl_interface': 2.428359873401512e-06, 'learning_rate/action_model': 2.4283598734015124e-05, 'epoch': 2.04}) +[2026-06-18 00:23:04,506][starVLA.training.train_starvla][INFO] - Step 4948, Loss: {'train/loss': 0.16609011590480804, 'train/grad_norm_pre_clip': 11.723118782043457, 'timing/data': 0.009889531880617142, 'timing/model': 0.9407357247546315, 'learning_rate/qwen_vl_interface': 2.4266204040614722e-06, 'learning_rate/action_model': 2.4266204040614724e-05, 'epoch': 2.04}) +[2026-06-18 00:23:08,346][starVLA.training.train_starvla][INFO] - Step 4949, Loss: {'train/loss': 0.32535475492477417, 'train/grad_norm_pre_clip': 19.701749801635742, 'timing/data': 0.00940485019236803, 'timing/model': 0.9350921818986535, 'learning_rate/qwen_vl_interface': 2.424881520011059e-06, 'learning_rate/action_model': 2.424881520011059e-05, 'epoch': 2.04}) +[2026-06-18 00:23:12,418][starVLA.training.train_starvla][INFO] - Step 4950, Loss: {'train/loss': 0.26124870777130127, 'train/grad_norm_pre_clip': 19.566707611083984, 'timing/data': 0.008061879314482212, 'timing/model': 0.9380822079256177, 'learning_rate/qwen_vl_interface': 2.4231432216107494e-06, 'learning_rate/action_model': 2.4231432216107494e-05, 'epoch': 2.04}) +[2026-06-18 00:23:16,199][starVLA.training.train_starvla][INFO] - Step 4951, Loss: {'train/loss': 0.3940833508968353, 'train/grad_norm_pre_clip': 33.32817459106445, 'timing/data': 0.009822697378695011, 'timing/model': 1.089473308995366, 'learning_rate/qwen_vl_interface': 2.4214055092208907e-06, 'learning_rate/action_model': 2.4214055092208907e-05, 'epoch': 2.04}) +[2026-06-18 00:23:19,810][starVLA.training.train_starvla][INFO] - Step 4952, Loss: {'train/loss': 0.19018681347370148, 'train/grad_norm_pre_clip': 15.190164566040039, 'timing/data': 0.009990330785512924, 'timing/model': 0.9363798061385751, 'learning_rate/qwen_vl_interface': 2.4196683832017134e-06, 'learning_rate/action_model': 2.4196683832017135e-05, 'epoch': 2.04}) +[2026-06-18 00:23:23,642][starVLA.training.train_starvla][INFO] - Step 4953, Loss: {'train/loss': 0.21245291829109192, 'train/grad_norm_pre_clip': 16.048511505126953, 'timing/data': 0.013091174885630608, 'timing/model': 0.928401262499392, 'learning_rate/qwen_vl_interface': 2.4179318439133245e-06, 'learning_rate/action_model': 2.417931843913324e-05, 'epoch': 2.04}) +[2026-06-18 00:23:27,429][starVLA.training.train_starvla][INFO] - Step 4954, Loss: {'train/loss': 0.08408303558826447, 'train/grad_norm_pre_clip': 8.418641090393066, 'timing/data': 0.013996453955769539, 'timing/model': 0.9274852471426129, 'learning_rate/qwen_vl_interface': 2.4161958917157126e-06, 'learning_rate/action_model': 2.4161958917157126e-05, 'epoch': 2.04}) +[2026-06-18 00:23:31,342][starVLA.training.train_starvla][INFO] - Step 4955, Loss: {'train/loss': 0.24789147078990936, 'train/grad_norm_pre_clip': 21.270109176635742, 'timing/data': 0.011247889138758183, 'timing/model': 1.0044297026470304, 'learning_rate/qwen_vl_interface': 2.4144605269687445e-06, 'learning_rate/action_model': 2.4144605269687446e-05, 'epoch': 2.04}) +[2026-06-18 00:23:35,194][starVLA.training.train_starvla][INFO] - Step 4956, Loss: {'train/loss': 0.35642892122268677, 'train/grad_norm_pre_clip': 21.976329803466797, 'timing/data': 0.009466630406677723, 'timing/model': 0.943017665296793, 'learning_rate/qwen_vl_interface': 2.412725750032156e-06, 'learning_rate/action_model': 2.4127257500321563e-05, 'epoch': 2.04}) +[2026-06-18 00:23:39,115][starVLA.training.train_starvla][INFO] - Step 4957, Loss: {'train/loss': 0.22686824202537537, 'train/grad_norm_pre_clip': 14.751423835754395, 'timing/data': 0.00997465755790472, 'timing/model': 0.9363872669637203, 'learning_rate/qwen_vl_interface': 2.4109915612655745e-06, 'learning_rate/action_model': 2.4109915612655745e-05, 'epoch': 2.04}) +[2026-06-18 00:23:42,937][starVLA.training.train_starvla][INFO] - Step 4958, Loss: {'train/loss': 0.42161425948143005, 'train/grad_norm_pre_clip': 23.32958984375, 'timing/data': 0.009147290140390396, 'timing/model': 1.141467702575028, 'learning_rate/qwen_vl_interface': 2.409257961028498e-06, 'learning_rate/action_model': 2.4092579610284978e-05, 'epoch': 2.04}) +[2026-06-18 00:23:46,599][starVLA.training.train_starvla][INFO] - Step 4959, Loss: {'train/loss': 0.5074229836463928, 'train/grad_norm_pre_clip': 25.5355224609375, 'timing/data': 0.010957534424960613, 'timing/model': 0.9620496015995741, 'learning_rate/qwen_vl_interface': 2.4075249496803026e-06, 'learning_rate/action_model': 2.4075249496803025e-05, 'epoch': 2.04}) +[2026-06-18 00:23:50,917][starVLA.training.train_starvla][INFO] - Step 4960, Loss: {'train/loss': 0.17041997611522675, 'train/grad_norm_pre_clip': 18.438392639160156, 'timing/data': 0.011163385584950447, 'timing/model': 0.9529912387952209, 'learning_rate/qwen_vl_interface': 2.405792527580243e-06, 'learning_rate/action_model': 2.405792527580243e-05, 'epoch': 2.04}) +[2026-06-18 00:23:54,843][starVLA.training.train_starvla][INFO] - Step 4961, Loss: {'train/loss': 0.4121142029762268, 'train/grad_norm_pre_clip': 20.562623977661133, 'timing/data': 0.01108032651245594, 'timing/model': 0.9410395380109549, 'learning_rate/qwen_vl_interface': 2.404060695087454e-06, 'learning_rate/action_model': 2.4040606950874536e-05, 'epoch': 2.04}) +[2026-06-18 00:23:58,931][starVLA.training.train_starvla][INFO] - Step 4962, Loss: {'train/loss': 0.44825315475463867, 'train/grad_norm_pre_clip': 25.612411499023438, 'timing/data': 0.014228302985429764, 'timing/model': 1.0528460731729865, 'learning_rate/qwen_vl_interface': 2.4023294525609437e-06, 'learning_rate/action_model': 2.4023294525609437e-05, 'epoch': 2.04}) +[2026-06-18 00:24:02,860][starVLA.training.train_starvla][INFO] - Step 4963, Loss: {'train/loss': 0.18404372036457062, 'train/grad_norm_pre_clip': 16.70220947265625, 'timing/data': 0.009913700632750988, 'timing/model': 1.0388746643438935, 'learning_rate/qwen_vl_interface': 2.400598800359604e-06, 'learning_rate/action_model': 2.4005988003596035e-05, 'epoch': 2.04}) +[2026-06-18 00:24:06,843][starVLA.training.train_starvla][INFO] - Step 4964, Loss: {'train/loss': 0.35585299134254456, 'train/grad_norm_pre_clip': 17.697195053100586, 'timing/data': 0.012298605404794216, 'timing/model': 1.1264608884230256, 'learning_rate/qwen_vl_interface': 2.398868738842198e-06, 'learning_rate/action_model': 2.3988687388421978e-05, 'epoch': 2.04}) +[2026-06-18 00:24:10,662][starVLA.training.train_starvla][INFO] - Step 4965, Loss: {'train/loss': 0.11157884448766708, 'train/grad_norm_pre_clip': 13.367021560668945, 'timing/data': 0.014298382215201855, 'timing/model': 1.1121078869327903, 'learning_rate/qwen_vl_interface': 2.3971392683673705e-06, 'learning_rate/action_model': 2.3971392683673706e-05, 'epoch': 2.04}) +[2026-06-18 00:24:14,463][starVLA.training.train_starvla][INFO] - Step 4966, Loss: {'train/loss': 0.2713622748851776, 'train/grad_norm_pre_clip': 19.91411590576172, 'timing/data': 0.010256670415401459, 'timing/model': 1.1197605626657605, 'learning_rate/qwen_vl_interface': 2.3954103892936425e-06, 'learning_rate/action_model': 2.3954103892936424e-05, 'epoch': 2.04}) +[2026-06-18 00:24:18,263][starVLA.training.train_starvla][INFO] - Step 4967, Loss: {'train/loss': 0.6379751563072205, 'train/grad_norm_pre_clip': 30.279449462890625, 'timing/data': 0.007470531389117241, 'timing/model': 0.9293654868379235, 'learning_rate/qwen_vl_interface': 2.393682101979411e-06, 'learning_rate/action_model': 2.3936821019794105e-05, 'epoch': 2.04}) +[2026-06-18 00:24:22,020][starVLA.training.train_starvla][INFO] - Step 4968, Loss: {'train/loss': 0.20924869179725647, 'train/grad_norm_pre_clip': 12.284195899963379, 'timing/data': 0.0068877460435032845, 'timing/model': 0.9290417293086648, 'learning_rate/qwen_vl_interface': 2.3919544067829546e-06, 'learning_rate/action_model': 2.3919544067829545e-05, 'epoch': 2.05}) +[2026-06-18 00:24:25,818][starVLA.training.train_starvla][INFO] - Step 4969, Loss: {'train/loss': 0.3523523211479187, 'train/grad_norm_pre_clip': 24.02288055419922, 'timing/data': 0.008354023098945618, 'timing/model': 0.928838400170207, 'learning_rate/qwen_vl_interface': 2.3902273040624267e-06, 'learning_rate/action_model': 2.390227304062427e-05, 'epoch': 2.05}) +[2026-06-18 00:24:29,552][starVLA.training.train_starvla][INFO] - Step 4970, Loss: {'train/loss': 0.2521687150001526, 'train/grad_norm_pre_clip': 26.1114501953125, 'timing/data': 0.008086832240223885, 'timing/model': 0.9247923763468862, 'learning_rate/qwen_vl_interface': 2.388500794175855e-06, 'learning_rate/action_model': 2.3885007941758548e-05, 'epoch': 2.05}) +[2026-06-18 00:24:33,417][starVLA.training.train_starvla][INFO] - Step 4971, Loss: {'train/loss': 0.37893709540367126, 'train/grad_norm_pre_clip': 20.452083587646484, 'timing/data': 0.00782976858317852, 'timing/model': 1.1170219723135233, 'learning_rate/qwen_vl_interface': 2.386774877481145e-06, 'learning_rate/action_model': 2.3867748774811452e-05, 'epoch': 2.05}) +[2026-06-18 00:24:37,159][starVLA.training.train_starvla][INFO] - Step 4972, Loss: {'train/loss': 0.28503262996673584, 'train/grad_norm_pre_clip': 24.403532028198242, 'timing/data': 0.011026831343770027, 'timing/model': 1.0673103518784046, 'learning_rate/qwen_vl_interface': 2.385049554336087e-06, 'learning_rate/action_model': 2.385049554336087e-05, 'epoch': 2.05}) +[2026-06-18 00:24:40,850][starVLA.training.train_starvla][INFO] - Step 4973, Loss: {'train/loss': 0.2046150118112564, 'train/grad_norm_pre_clip': 13.55854606628418, 'timing/data': 0.01691968273371458, 'timing/model': 0.9172780765220523, 'learning_rate/qwen_vl_interface': 2.3833248250983383e-06, 'learning_rate/action_model': 2.3833248250983383e-05, 'epoch': 2.05}) +[2026-06-18 00:24:44,617][starVLA.training.train_starvla][INFO] - Step 4974, Loss: {'train/loss': 0.11537522077560425, 'train/grad_norm_pre_clip': 8.943398475646973, 'timing/data': 0.009410873986780643, 'timing/model': 0.9280789466574788, 'learning_rate/qwen_vl_interface': 2.3816006901254396e-06, 'learning_rate/action_model': 2.3816006901254396e-05, 'epoch': 2.05}) +[2026-06-18 00:24:48,540][starVLA.training.train_starvla][INFO] - Step 4975, Loss: {'train/loss': 0.21330806612968445, 'train/grad_norm_pre_clip': 22.26313018798828, 'timing/data': 0.007501520216464996, 'timing/model': 0.9309964589774609, 'learning_rate/qwen_vl_interface': 2.3798771497748012e-06, 'learning_rate/action_model': 2.379877149774801e-05, 'epoch': 2.05}) +[2026-06-18 00:24:53,281][starVLA.training.train_starvla][INFO] - Step 4976, Loss: {'train/loss': 0.20580610632896423, 'train/grad_norm_pre_clip': 18.824840545654297, 'timing/data': 0.008782201446592808, 'timing/model': 0.9413501946255565, 'learning_rate/qwen_vl_interface': 2.3781542044037195e-06, 'learning_rate/action_model': 2.3781542044037193e-05, 'epoch': 2.05}) +[2026-06-18 00:24:57,332][starVLA.training.train_starvla][INFO] - Step 4977, Loss: {'train/loss': 0.23409460484981537, 'train/grad_norm_pre_clip': 20.523290634155273, 'timing/data': 0.011150524951517582, 'timing/model': 0.9505808018147945, 'learning_rate/qwen_vl_interface': 2.376431854369362e-06, 'learning_rate/action_model': 2.3764318543693618e-05, 'epoch': 2.05}) +[2026-06-18 00:25:01,230][starVLA.training.train_starvla][INFO] - Step 4978, Loss: {'train/loss': 0.5012985467910767, 'train/grad_norm_pre_clip': 21.197675704956055, 'timing/data': 0.007567697204649448, 'timing/model': 1.1249557742848992, 'learning_rate/qwen_vl_interface': 2.374710100028773e-06, 'learning_rate/action_model': 2.3747101000287728e-05, 'epoch': 2.05}) +[2026-06-18 00:25:04,923][starVLA.training.train_starvla][INFO] - Step 4979, Loss: {'train/loss': 0.1929728388786316, 'train/grad_norm_pre_clip': 17.640565872192383, 'timing/data': 0.009469461627304554, 'timing/model': 0.9466519728302956, 'learning_rate/qwen_vl_interface': 2.3729889417388736e-06, 'learning_rate/action_model': 2.3729889417388737e-05, 'epoch': 2.05}) +[2026-06-18 00:25:08,893][starVLA.training.train_starvla][INFO] - Step 4980, Loss: {'train/loss': 0.09970704466104507, 'train/grad_norm_pre_clip': 8.853973388671875, 'timing/data': 0.00874044094234705, 'timing/model': 0.9478530939668417, 'learning_rate/qwen_vl_interface': 2.3712683798564628e-06, 'learning_rate/action_model': 2.3712683798564625e-05, 'epoch': 2.05}) +[2026-06-18 00:25:12,646][starVLA.training.train_starvla][INFO] - Step 4981, Loss: {'train/loss': 0.553152859210968, 'train/grad_norm_pre_clip': 26.46734619140625, 'timing/data': 0.007445665076375008, 'timing/model': 0.9366363892331719, 'learning_rate/qwen_vl_interface': 2.3695484147382146e-06, 'learning_rate/action_model': 2.3695484147382147e-05, 'epoch': 2.05}) +[2026-06-18 00:25:16,809][starVLA.training.train_starvla][INFO] - Step 4982, Loss: {'train/loss': 0.200273334980011, 'train/grad_norm_pre_clip': 17.6896915435791, 'timing/data': 0.009667638689279556, 'timing/model': 0.9480688497424126, 'learning_rate/qwen_vl_interface': 2.367829046740678e-06, 'learning_rate/action_model': 2.3678290467406776e-05, 'epoch': 2.05}) +[2026-06-18 00:25:20,665][starVLA.training.train_starvla][INFO] - Step 4983, Loss: {'train/loss': 0.23003938794136047, 'train/grad_norm_pre_clip': 26.426668167114258, 'timing/data': 0.009480440989136696, 'timing/model': 0.9921428672969341, 'learning_rate/qwen_vl_interface': 2.3661102762202846e-06, 'learning_rate/action_model': 2.3661102762202843e-05, 'epoch': 2.05}) +[2026-06-18 00:25:24,648][starVLA.training.train_starvla][INFO] - Step 4984, Loss: {'train/loss': 0.4795767068862915, 'train/grad_norm_pre_clip': 22.515010833740234, 'timing/data': 0.022716141305863857, 'timing/model': 1.2051682425662875, 'learning_rate/qwen_vl_interface': 2.364392103533332e-06, 'learning_rate/action_model': 2.364392103533332e-05, 'epoch': 2.05}) +[2026-06-18 00:25:28,649][starVLA.training.train_starvla][INFO] - Step 4985, Loss: {'train/loss': 0.25741633772850037, 'train/grad_norm_pre_clip': 23.017425537109375, 'timing/data': 0.009766394272446632, 'timing/model': 1.1272372398525476, 'learning_rate/qwen_vl_interface': 2.362674529036002e-06, 'learning_rate/action_model': 2.362674529036002e-05, 'epoch': 2.05}) +[2026-06-18 00:25:32,427][starVLA.training.train_starvla][INFO] - Step 4986, Loss: {'train/loss': 0.21657171845436096, 'train/grad_norm_pre_clip': 19.900012969970703, 'timing/data': 0.007291172631084919, 'timing/model': 0.9556928621605039, 'learning_rate/qwen_vl_interface': 2.360957553084347e-06, 'learning_rate/action_model': 2.360957553084347e-05, 'epoch': 2.05}) +[2026-06-18 00:25:36,352][starVLA.training.train_starvla][INFO] - Step 4987, Loss: {'train/loss': 0.366566002368927, 'train/grad_norm_pre_clip': 30.342708587646484, 'timing/data': 0.014719152823090553, 'timing/model': 0.9307243190705776, 'learning_rate/qwen_vl_interface': 2.359241176034302e-06, 'learning_rate/action_model': 2.359241176034302e-05, 'epoch': 2.05}) +[2026-06-18 00:25:40,162][starVLA.training.train_starvla][INFO] - Step 4988, Loss: {'train/loss': 0.2769400179386139, 'train/grad_norm_pre_clip': 17.984920501708984, 'timing/data': 0.014068619348108768, 'timing/model': 0.9347644615918398, 'learning_rate/qwen_vl_interface': 2.3575253982416732e-06, 'learning_rate/action_model': 2.357525398241673e-05, 'epoch': 2.05}) +[2026-06-18 00:25:43,997][starVLA.training.train_starvla][INFO] - Step 4989, Loss: {'train/loss': 0.28803789615631104, 'train/grad_norm_pre_clip': 13.654999732971191, 'timing/data': 0.00864280853420496, 'timing/model': 0.9426296930760145, 'learning_rate/qwen_vl_interface': 2.3558102200621385e-06, 'learning_rate/action_model': 2.3558102200621386e-05, 'epoch': 2.05}) +[2026-06-18 00:25:47,817][starVLA.training.train_starvla][INFO] - Step 4990, Loss: {'train/loss': 0.3204518258571625, 'train/grad_norm_pre_clip': 26.679378509521484, 'timing/data': 0.008476135320961475, 'timing/model': 0.952734530903399, 'learning_rate/qwen_vl_interface': 2.3540956418512607e-06, 'learning_rate/action_model': 2.354095641851261e-05, 'epoch': 2.05}) +[2026-06-18 00:25:51,593][starVLA.training.train_starvla][INFO] - Step 4991, Loss: {'train/loss': 0.3594902455806732, 'train/grad_norm_pre_clip': 22.09880256652832, 'timing/data': 0.009737174957990646, 'timing/model': 1.1069757090881467, 'learning_rate/qwen_vl_interface': 2.352381663964473e-06, 'learning_rate/action_model': 2.352381663964473e-05, 'epoch': 2.05}) +[2026-06-18 00:25:55,388][starVLA.training.train_starvla][INFO] - Step 4992, Loss: {'train/loss': 0.31937411427497864, 'train/grad_norm_pre_clip': 20.85528564453125, 'timing/data': 0.0077399881556630135, 'timing/model': 1.0916314097121358, 'learning_rate/qwen_vl_interface': 2.3506682867570846e-06, 'learning_rate/action_model': 2.3506682867570844e-05, 'epoch': 2.05}) +[2026-06-18 00:25:58,967][starVLA.training.train_starvla][INFO] - Step 4993, Loss: {'train/loss': 0.340247243642807, 'train/grad_norm_pre_clip': 28.891571044921875, 'timing/data': 0.011047384701669216, 'timing/model': 0.9194310521706939, 'learning_rate/qwen_vl_interface': 2.3489555105842796e-06, 'learning_rate/action_model': 2.3489555105842794e-05, 'epoch': 2.06}) +[2026-06-18 00:26:02,729][starVLA.training.train_starvla][INFO] - Step 4994, Loss: {'train/loss': 0.5333713889122009, 'train/grad_norm_pre_clip': 29.07720947265625, 'timing/data': 0.012118939310312271, 'timing/model': 0.926663507707417, 'learning_rate/qwen_vl_interface': 2.347243335801119e-06, 'learning_rate/action_model': 2.347243335801119e-05, 'epoch': 2.06}) +[2026-06-18 00:26:06,497][starVLA.training.train_starvla][INFO] - Step 4995, Loss: {'train/loss': 0.09186328202486038, 'train/grad_norm_pre_clip': 25.015989303588867, 'timing/data': 0.011372937820851803, 'timing/model': 0.9296176610514522, 'learning_rate/qwen_vl_interface': 2.3455317627625385e-06, 'learning_rate/action_model': 2.3455317627625385e-05, 'epoch': 2.06}) +[2026-06-18 00:26:10,257][starVLA.training.train_starvla][INFO] - Step 4996, Loss: {'train/loss': 0.2277863621711731, 'train/grad_norm_pre_clip': 18.438024520874023, 'timing/data': 0.01309144962579012, 'timing/model': 0.9270122973248363, 'learning_rate/qwen_vl_interface': 2.3438207918233495e-06, 'learning_rate/action_model': 2.3438207918233494e-05, 'epoch': 2.06}) +[2026-06-18 00:26:14,052][starVLA.training.train_starvla][INFO] - Step 4997, Loss: {'train/loss': 0.1993667334318161, 'train/grad_norm_pre_clip': 16.854801177978516, 'timing/data': 0.01223122887313366, 'timing/model': 0.9336439538747072, 'learning_rate/qwen_vl_interface': 2.3421104233382366e-06, 'learning_rate/action_model': 2.3421104233382364e-05, 'epoch': 2.06}) +[2026-06-18 00:26:17,812][starVLA.training.train_starvla][INFO] - Step 4998, Loss: {'train/loss': 0.6427637934684753, 'train/grad_norm_pre_clip': 25.12492561340332, 'timing/data': 0.009732259437441826, 'timing/model': 1.0980516858398914, 'learning_rate/qwen_vl_interface': 2.340400657661762e-06, 'learning_rate/action_model': 2.340400657661762e-05, 'epoch': 2.06}) +[2026-06-18 00:26:21,398][starVLA.training.train_starvla][INFO] - Step 4999, Loss: {'train/loss': 0.15775227546691895, 'train/grad_norm_pre_clip': 14.611973762512207, 'timing/data': 0.008543442003428936, 'timing/model': 0.9269271465018392, 'learning_rate/qwen_vl_interface': 2.338691495148363e-06, 'learning_rate/action_model': 2.3386914951483626e-05, 'epoch': 2.06}) +[2026-06-18 00:28:14,598][starVLA.training.train_starvla][INFO] - Step 5000, Loss: {'train/loss': 0.44448861479759216, 'train/grad_norm_pre_clip': 94.9923324584961, 'eval/loss': 0.41911906003952026, 'eval/latency_6/loss': 0.41911906003952026, 'eval/demon_fire_precision': 0.9587991057170233, 'eval/demon_fire_recall': 0.9702650290885585, 'eval/demon_fire_f1': 0.9644979919678716, 'eval/demon_fire_rate_model': 0.8711741791875348, 'eval/demon_fire_rate_teacher': 0.860879243183083, 'eval/demon_left_precision': 0.8534653465346534, 'eval/demon_left_recall': 0.756140350877193, 'eval/demon_left_f1': 0.8018604651162791, 'eval/demon_left_rate_model': 0.14051196438508626, 'eval/demon_left_rate_teacher': 0.15859766277128548, 'eval/demon_right_precision': 0.8344594594594594, 'eval/demon_right_recall': 0.7588325652841782, 'eval/demon_right_f1': 0.7948511665325824, 'eval/demon_right_rate_model': 0.16471897607122982, 'eval/demon_right_rate_teacher': 0.18113522537562604, 'eval/demon_move_f1': 0.7983558158244307, 'eval/demon_cc_f1': 0.9078212348676375, 'eval/latency_6/demon_fire_precision': 0.9587991057170233, 'eval/latency_6/demon_fire_recall': 0.9702650290885585, 'eval/latency_6/demon_fire_f1': 0.9644979919678716, 'eval/latency_6/demon_fire_rate_model': 0.8711741791875348, 'eval/latency_6/demon_fire_rate_teacher': 0.860879243183083, 'eval/latency_6/demon_left_precision': 0.8534653465346534, 'eval/latency_6/demon_left_recall': 0.756140350877193, 'eval/latency_6/demon_left_f1': 0.8018604651162791, 'eval/latency_6/demon_left_rate_model': 0.14051196438508626, 'eval/latency_6/demon_left_rate_teacher': 0.15859766277128548, 'eval/latency_6/demon_right_precision': 0.8344594594594594, 'eval/latency_6/demon_right_recall': 0.7588325652841782, 'eval/latency_6/demon_right_f1': 0.7948511665325824, 'eval/latency_6/demon_right_rate_model': 0.16471897607122982, 'eval/latency_6/demon_right_rate_teacher': 0.18113522537562604, 'eval/latency_6/demon_move_f1': 0.7983558158244307, 'eval/latency_6/demon_cc_f1': 0.9078212348676375, 'timing/data': 0.01311995554715395, 'timing/model': 0.9392884382978082, 'learning_rate/qwen_vl_interface': 2.336982936152347e-06, 'learning_rate/action_model': 2.3369829361523468e-05, 'epoch': 2.06}) +[2026-06-18 00:28:14,624][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_5000_state +[2026-06-18 00:28:14,641][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 00:28:21,351][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_5000_state/model.safetensors +[2026-06-18 00:28:36,987][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_5000_state/optimizer.bin +[2026-06-18 00:28:36,990][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_5000_state/sampler.bin +[2026-06-18 00:28:36,990][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_5000_state/sampler_1.bin +[2026-06-18 00:28:36,992][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_5000_state/random_states_0.pkl +[2026-06-18 00:28:36,997][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 5000 +[2026-06-18 00:28:36,998][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 00:28:37,015][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 00:28:41,527][starVLA.training.train_starvla][INFO] - Step 5001, Loss: {'train/loss': 0.1320723444223404, 'train/grad_norm_pre_clip': 12.460779190063477, 'timing/data': 0.013567468151450157, 'timing/model': 1.1807776428759098, 'learning_rate/qwen_vl_interface': 2.3352749810279047e-06, 'learning_rate/action_model': 2.3352749810279047e-05, 'epoch': 2.06}) +[2026-06-18 00:28:45,290][starVLA.training.train_starvla][INFO] - Step 5002, Loss: {'train/loss': 0.03337091580033302, 'train/grad_norm_pre_clip': 4.667285919189453, 'timing/data': 0.00992328580468893, 'timing/model': 0.9472229443490505, 'learning_rate/qwen_vl_interface': 2.3335676301290972e-06, 'learning_rate/action_model': 2.3335676301290973e-05, 'epoch': 2.06}) +[2026-06-18 00:28:49,930][starVLA.training.train_starvla][INFO] - Step 5003, Loss: {'train/loss': 0.20781192183494568, 'train/grad_norm_pre_clip': 13.67798137664795, 'timing/data': 0.010297809727489948, 'timing/model': 1.0066273221746087, 'learning_rate/qwen_vl_interface': 2.331860883809857e-06, 'learning_rate/action_model': 2.3318608838098565e-05, 'epoch': 2.06}) +[2026-06-18 00:28:53,912][starVLA.training.train_starvla][INFO] - Step 5004, Loss: {'train/loss': 0.11871671676635742, 'train/grad_norm_pre_clip': 10.954611778259277, 'timing/data': 0.009768412448465824, 'timing/model': 0.9474705504253507, 'learning_rate/qwen_vl_interface': 2.330154742423993e-06, 'learning_rate/action_model': 2.330154742423993e-05, 'epoch': 2.06}) +[2026-06-18 00:28:57,939][starVLA.training.train_starvla][INFO] - Step 5005, Loss: {'train/loss': 0.2056513875722885, 'train/grad_norm_pre_clip': 16.72609519958496, 'timing/data': 0.014612891711294651, 'timing/model': 0.9676567101851106, 'learning_rate/qwen_vl_interface': 2.328449206325195e-06, 'learning_rate/action_model': 2.3284492063251947e-05, 'epoch': 2.06}) +[2026-06-18 00:29:01,747][starVLA.training.train_starvla][INFO] - Step 5006, Loss: {'train/loss': 0.3162921667098999, 'train/grad_norm_pre_clip': 29.89725112915039, 'timing/data': 0.013090033084154129, 'timing/model': 0.9614730915054679, 'learning_rate/qwen_vl_interface': 2.32674427586702e-06, 'learning_rate/action_model': 2.3267442758670198e-05, 'epoch': 2.06}) +[2026-06-18 00:29:05,547][starVLA.training.train_starvla][INFO] - Step 5007, Loss: {'train/loss': 0.3717346489429474, 'train/grad_norm_pre_clip': 27.42626190185547, 'timing/data': 0.009390419349074364, 'timing/model': 1.1050942204892635, 'learning_rate/qwen_vl_interface': 2.3250399514029018e-06, 'learning_rate/action_model': 2.3250399514029016e-05, 'epoch': 2.06}) +[2026-06-18 00:29:09,395][starVLA.training.train_starvla][INFO] - Step 5008, Loss: {'train/loss': 0.5050463676452637, 'train/grad_norm_pre_clip': 81.87944793701172, 'timing/data': 0.009727188386023045, 'timing/model': 1.1237972360104322, 'learning_rate/qwen_vl_interface': 2.32333623328615e-06, 'learning_rate/action_model': 2.3233362332861498e-05, 'epoch': 2.06}) +[2026-06-18 00:29:13,040][starVLA.training.train_starvla][INFO] - Step 5009, Loss: {'train/loss': 0.38576674461364746, 'train/grad_norm_pre_clip': 19.026100158691406, 'timing/data': 0.008357970975339413, 'timing/model': 0.9526431243866682, 'learning_rate/qwen_vl_interface': 2.3216331218699457e-06, 'learning_rate/action_model': 2.3216331218699456e-05, 'epoch': 2.06}) +[2026-06-18 00:29:16,850][starVLA.training.train_starvla][INFO] - Step 5010, Loss: {'train/loss': 0.32810550928115845, 'train/grad_norm_pre_clip': 28.628633499145508, 'timing/data': 0.007786695845425129, 'timing/model': 0.9703190512955189, 'learning_rate/qwen_vl_interface': 2.3199306175073465e-06, 'learning_rate/action_model': 2.3199306175073467e-05, 'epoch': 2.06}) +[2026-06-18 00:29:20,829][starVLA.training.train_starvla][INFO] - Step 5011, Loss: {'train/loss': 0.24813908338546753, 'train/grad_norm_pre_clip': 100.23653411865234, 'timing/data': 0.00948443915694952, 'timing/model': 0.9440157227218151, 'learning_rate/qwen_vl_interface': 2.318228720551283e-06, 'learning_rate/action_model': 2.318228720551283e-05, 'epoch': 2.06}) +[2026-06-18 00:29:24,568][starVLA.training.train_starvla][INFO] - Step 5012, Loss: {'train/loss': 0.5934338569641113, 'train/grad_norm_pre_clip': 23.043853759765625, 'timing/data': 0.008795811794698238, 'timing/model': 0.9321099799126387, 'learning_rate/qwen_vl_interface': 2.3165274313545606e-06, 'learning_rate/action_model': 2.3165274313545606e-05, 'epoch': 2.06}) +[2026-06-18 00:29:28,595][starVLA.training.train_starvla][INFO] - Step 5013, Loss: {'train/loss': 0.369789719581604, 'train/grad_norm_pre_clip': 22.549102783203125, 'timing/data': 0.009431743994355202, 'timing/model': 0.9429476838558912, 'learning_rate/qwen_vl_interface': 2.3148267502698584e-06, 'learning_rate/action_model': 2.3148267502698584e-05, 'epoch': 2.06}) +[2026-06-18 00:29:32,428][starVLA.training.train_starvla][INFO] - Step 5014, Loss: {'train/loss': 0.19784221053123474, 'train/grad_norm_pre_clip': 17.57727813720703, 'timing/data': 0.013608046807348728, 'timing/model': 1.130030003376305, 'learning_rate/qwen_vl_interface': 2.313126677649729e-06, 'learning_rate/action_model': 2.313126677649729e-05, 'epoch': 2.06}) +[2026-06-18 00:29:36,210][starVLA.training.train_starvla][INFO] - Step 5015, Loss: {'train/loss': 0.2292170226573944, 'train/grad_norm_pre_clip': 23.070281982421875, 'timing/data': 0.00897010788321495, 'timing/model': 1.0328561533242464, 'learning_rate/qwen_vl_interface': 2.3114272138465975e-06, 'learning_rate/action_model': 2.3114272138465977e-05, 'epoch': 2.06}) +[2026-06-18 00:29:42,726][starVLA.training.train_starvla][INFO] - Step 5016, Loss: {'train/loss': 0.2286897599697113, 'train/grad_norm_pre_clip': 21.202531814575195, 'timing/data': 0.018363741226494312, 'timing/model': 1.0553850140422583, 'learning_rate/qwen_vl_interface': 2.3097283592127706e-06, 'learning_rate/action_model': 2.3097283592127706e-05, 'epoch': 2.06}) +[2026-06-18 00:29:47,791][starVLA.training.train_starvla][INFO] - Step 5017, Loss: {'train/loss': 0.17470553517341614, 'train/grad_norm_pre_clip': 17.333303451538086, 'timing/data': 0.012374534271657467, 'timing/model': 1.0096168760210276, 'learning_rate/qwen_vl_interface': 2.3080301141004165e-06, 'learning_rate/action_model': 2.3080301141004165e-05, 'epoch': 2.07}) +[2026-06-18 00:29:51,649][starVLA.training.train_starvla][INFO] - Step 5018, Loss: {'train/loss': 0.13147059082984924, 'train/grad_norm_pre_clip': 18.355039596557617, 'timing/data': 0.007905942387878895, 'timing/model': 0.9420268703252077, 'learning_rate/qwen_vl_interface': 2.3063324788615843e-06, 'learning_rate/action_model': 2.306332478861584e-05, 'epoch': 2.07}) +[2026-06-18 00:29:56,216][starVLA.training.train_starvla][INFO] - Step 5019, Loss: {'train/loss': 0.4985242784023285, 'train/grad_norm_pre_clip': 32.78834533691406, 'timing/data': 0.012401255778968334, 'timing/model': 0.9976967396214604, 'learning_rate/qwen_vl_interface': 2.304635453848198e-06, 'learning_rate/action_model': 2.3046354538481983e-05, 'epoch': 2.07}) +[2026-06-18 00:30:00,457][starVLA.training.train_starvla][INFO] - Step 5020, Loss: {'train/loss': 0.2557651400566101, 'train/grad_norm_pre_clip': 14.744648933410645, 'timing/data': 0.011011536233127117, 'timing/model': 1.5505808526650071, 'learning_rate/qwen_vl_interface': 2.3029390394120517e-06, 'learning_rate/action_model': 2.3029390394120513e-05, 'epoch': 2.07}) +[2026-06-18 00:30:04,365][starVLA.training.train_starvla][INFO] - Step 5021, Loss: {'train/loss': 0.5208982825279236, 'train/grad_norm_pre_clip': 24.67177391052246, 'timing/data': 0.007835221476852894, 'timing/model': 1.1731771314516664, 'learning_rate/qwen_vl_interface': 2.301243235904813e-06, 'learning_rate/action_model': 2.3012432359048133e-05, 'epoch': 2.07}) +[2026-06-18 00:30:08,088][starVLA.training.train_starvla][INFO] - Step 5022, Loss: {'train/loss': 0.21185630559921265, 'train/grad_norm_pre_clip': 49.726593017578125, 'timing/data': 0.007776553742587566, 'timing/model': 0.9781955759972334, 'learning_rate/qwen_vl_interface': 2.299548043678026e-06, 'learning_rate/action_model': 2.299548043678026e-05, 'epoch': 2.07}) +[2026-06-18 00:30:12,099][starVLA.training.train_starvla][INFO] - Step 5023, Loss: {'train/loss': 0.5361214280128479, 'train/grad_norm_pre_clip': 24.665409088134766, 'timing/data': 0.009152072481811047, 'timing/model': 1.0441381745040417, 'learning_rate/qwen_vl_interface': 2.297853463083103e-06, 'learning_rate/action_model': 2.297853463083103e-05, 'epoch': 2.07}) +[2026-06-18 00:30:16,015][starVLA.training.train_starvla][INFO] - Step 5024, Loss: {'train/loss': 0.3146612048149109, 'train/grad_norm_pre_clip': 27.291749954223633, 'timing/data': 0.011645544320344925, 'timing/model': 0.9521743822842836, 'learning_rate/qwen_vl_interface': 2.296159494471335e-06, 'learning_rate/action_model': 2.2961594944713347e-05, 'epoch': 2.07}) +[2026-06-18 00:30:19,834][starVLA.training.train_starvla][INFO] - Step 5025, Loss: {'train/loss': 0.3150486648082733, 'train/grad_norm_pre_clip': 23.45756721496582, 'timing/data': 0.020581232383847237, 'timing/model': 0.9449022952467203, 'learning_rate/qwen_vl_interface': 2.294466138193881e-06, 'learning_rate/action_model': 2.2944661381938812e-05, 'epoch': 2.07}) +[2026-06-18 00:30:23,717][starVLA.training.train_starvla][INFO] - Step 5026, Loss: {'train/loss': 0.15854069590568542, 'train/grad_norm_pre_clip': 16.428131103515625, 'timing/data': 0.009779676795005798, 'timing/model': 0.9464868241921067, 'learning_rate/qwen_vl_interface': 2.2927733946017777e-06, 'learning_rate/action_model': 2.2927733946017778e-05, 'epoch': 2.07}) +[2026-06-18 00:30:27,660][starVLA.training.train_starvla][INFO] - Step 5027, Loss: {'train/loss': 0.31645604968070984, 'train/grad_norm_pre_clip': 27.977060317993164, 'timing/data': 0.008637621067464352, 'timing/model': 1.1718344939872622, 'learning_rate/qwen_vl_interface': 2.2910812640459314e-06, 'learning_rate/action_model': 2.2910812640459315e-05, 'epoch': 2.07}) +[2026-06-18 00:30:31,570][starVLA.training.train_starvla][INFO] - Step 5028, Loss: {'train/loss': 0.47388049960136414, 'train/grad_norm_pre_clip': 33.35702133178711, 'timing/data': 0.007927097380161285, 'timing/model': 1.1344562470912933, 'learning_rate/qwen_vl_interface': 2.289389746877123e-06, 'learning_rate/action_model': 2.2893897468771228e-05, 'epoch': 2.07}) +[2026-06-18 00:30:35,305][starVLA.training.train_starvla][INFO] - Step 5029, Loss: {'train/loss': 0.2699393630027771, 'train/grad_norm_pre_clip': 24.377731323242188, 'timing/data': 0.012983651831746101, 'timing/model': 0.9521067077293992, 'learning_rate/qwen_vl_interface': 2.2876988434460046e-06, 'learning_rate/action_model': 2.2876988434460046e-05, 'epoch': 2.07}) +[2026-06-18 00:30:39,280][starVLA.training.train_starvla][INFO] - Step 5030, Loss: {'train/loss': 0.14659391343593597, 'train/grad_norm_pre_clip': 23.918331146240234, 'timing/data': 0.01586620695888996, 'timing/model': 0.9331651544198394, 'learning_rate/qwen_vl_interface': 2.2860085541031048e-06, 'learning_rate/action_model': 2.2860085541031047e-05, 'epoch': 2.07}) +[2026-06-18 00:30:43,300][starVLA.training.train_starvla][INFO] - Step 5031, Loss: {'train/loss': 0.1731405258178711, 'train/grad_norm_pre_clip': 18.802946090698242, 'timing/data': 0.010626859031617641, 'timing/model': 0.9943664399906993, 'learning_rate/qwen_vl_interface': 2.284318879198823e-06, 'learning_rate/action_model': 2.284318879198823e-05, 'epoch': 2.07}) +[2026-06-18 00:30:47,131][starVLA.training.train_starvla][INFO] - Step 5032, Loss: {'train/loss': 0.4603676497936249, 'train/grad_norm_pre_clip': 27.31574058532715, 'timing/data': 0.00815690029412508, 'timing/model': 0.9269143212586641, 'learning_rate/qwen_vl_interface': 2.2826298190834266e-06, 'learning_rate/action_model': 2.2826298190834266e-05, 'epoch': 2.07}) +[2026-06-18 00:30:50,877][starVLA.training.train_starvla][INFO] - Step 5033, Loss: {'train/loss': 0.3851918578147888, 'train/grad_norm_pre_clip': 22.121810913085938, 'timing/data': 0.010385392233729362, 'timing/model': 0.9247921202331781, 'learning_rate/qwen_vl_interface': 2.2809413741070603e-06, 'learning_rate/action_model': 2.28094137410706e-05, 'epoch': 2.07}) +[2026-06-18 00:30:54,826][starVLA.training.train_starvla][INFO] - Step 5034, Loss: {'train/loss': 0.26131531596183777, 'train/grad_norm_pre_clip': 20.60383415222168, 'timing/data': 0.012081054970622063, 'timing/model': 1.2140890322625637, 'learning_rate/qwen_vl_interface': 2.2792535446197435e-06, 'learning_rate/action_model': 2.2792535446197436e-05, 'epoch': 2.07}) +[2026-06-18 00:30:58,400][starVLA.training.train_starvla][INFO] - Step 5035, Loss: {'train/loss': 0.3135257363319397, 'train/grad_norm_pre_clip': 22.748159408569336, 'timing/data': 0.01576903648674488, 'timing/model': 0.9232964282855392, 'learning_rate/qwen_vl_interface': 2.2775663309713643e-06, 'learning_rate/action_model': 2.2775663309713645e-05, 'epoch': 2.07}) +[2026-06-18 00:31:02,193][starVLA.training.train_starvla][INFO] - Step 5036, Loss: {'train/loss': 0.6287704706192017, 'train/grad_norm_pre_clip': 28.871042251586914, 'timing/data': 0.013184900395572186, 'timing/model': 0.9274306362494826, 'learning_rate/qwen_vl_interface': 2.2758797335116843e-06, 'learning_rate/action_model': 2.2758797335116844e-05, 'epoch': 2.07}) +[2026-06-18 00:31:05,949][starVLA.training.train_starvla][INFO] - Step 5037, Loss: {'train/loss': 0.3358403742313385, 'train/grad_norm_pre_clip': 20.52762794494629, 'timing/data': 0.009569761343300343, 'timing/model': 0.9276635637506843, 'learning_rate/qwen_vl_interface': 2.2741937525903324e-06, 'learning_rate/action_model': 2.2741937525903324e-05, 'epoch': 2.07}) +[2026-06-18 00:31:09,770][starVLA.training.train_starvla][INFO] - Step 5038, Loss: {'train/loss': 0.1054091602563858, 'train/grad_norm_pre_clip': 12.706064224243164, 'timing/data': 0.008124156855046749, 'timing/model': 0.982893262989819, 'learning_rate/qwen_vl_interface': 2.2725083885568198e-06, 'learning_rate/action_model': 2.2725083885568197e-05, 'epoch': 2.07}) +[2026-06-18 00:31:14,025][starVLA.training.train_starvla][INFO] - Step 5039, Loss: {'train/loss': 0.468988835811615, 'train/grad_norm_pre_clip': 20.335676193237305, 'timing/data': 0.013518898747861385, 'timing/model': 0.9295094916597009, 'learning_rate/qwen_vl_interface': 2.2708236417605217e-06, 'learning_rate/action_model': 2.2708236417605216e-05, 'epoch': 2.07}) +[2026-06-18 00:31:17,790][starVLA.training.train_starvla][INFO] - Step 5040, Loss: {'train/loss': 0.2415890395641327, 'train/grad_norm_pre_clip': 19.33271598815918, 'timing/data': 0.015964840538799763, 'timing/model': 0.9276684802025557, 'learning_rate/qwen_vl_interface': 2.2691395125506885e-06, 'learning_rate/action_model': 2.2691395125506885e-05, 'epoch': 2.07}) +[2026-06-18 00:31:21,569][starVLA.training.train_starvla][INFO] - Step 5041, Loss: {'train/loss': 0.370733380317688, 'train/grad_norm_pre_clip': 19.676162719726562, 'timing/data': 0.021670875139534473, 'timing/model': 1.0919538158923388, 'learning_rate/qwen_vl_interface': 2.2674560012764426e-06, 'learning_rate/action_model': 2.2674560012764423e-05, 'epoch': 2.08}) +[2026-06-18 00:31:25,191][starVLA.training.train_starvla][INFO] - Step 5042, Loss: {'train/loss': 0.39543867111206055, 'train/grad_norm_pre_clip': 25.681257247924805, 'timing/data': 0.012255308218300343, 'timing/model': 0.9343980401754379, 'learning_rate/qwen_vl_interface': 2.2657731082867764e-06, 'learning_rate/action_model': 2.265773108286776e-05, 'epoch': 2.08}) +[2026-06-18 00:31:29,101][starVLA.training.train_starvla][INFO] - Step 5043, Loss: {'train/loss': 0.21565894782543182, 'train/grad_norm_pre_clip': 19.378511428833008, 'timing/data': 0.008499698713421822, 'timing/model': 0.9492505844682455, 'learning_rate/qwen_vl_interface': 2.264090833930555e-06, 'learning_rate/action_model': 2.2640908339305553e-05, 'epoch': 2.08}) +[2026-06-18 00:31:32,861][starVLA.training.train_starvla][INFO] - Step 5044, Loss: {'train/loss': 0.21690022945404053, 'train/grad_norm_pre_clip': 19.06361961364746, 'timing/data': 0.012637135572731495, 'timing/model': 0.9279503226280212, 'learning_rate/qwen_vl_interface': 2.2624091785565165e-06, 'learning_rate/action_model': 2.2624091785565167e-05, 'epoch': 2.08}) +[2026-06-18 00:31:36,621][starVLA.training.train_starvla][INFO] - Step 5045, Loss: {'train/loss': 0.2954590618610382, 'train/grad_norm_pre_clip': 23.85148048400879, 'timing/data': 0.008558268658816814, 'timing/model': 0.9219127632677555, 'learning_rate/qwen_vl_interface': 2.2607281425132728e-06, 'learning_rate/action_model': 2.260728142513273e-05, 'epoch': 2.08}) +[2026-06-18 00:31:40,382][starVLA.training.train_starvla][INFO] - Step 5046, Loss: {'train/loss': 0.2591044306755066, 'train/grad_norm_pre_clip': 55.75186538696289, 'timing/data': 0.011577672325074673, 'timing/model': 0.9236822538077831, 'learning_rate/qwen_vl_interface': 2.2590477261493e-06, 'learning_rate/action_model': 2.2590477261493e-05, 'epoch': 2.08}) +[2026-06-18 00:31:44,180][starVLA.training.train_starvla][INFO] - Step 5047, Loss: {'train/loss': 0.2191484570503235, 'train/grad_norm_pre_clip': 23.891887664794922, 'timing/data': 0.01516345888376236, 'timing/model': 1.121295116841793, 'learning_rate/qwen_vl_interface': 2.257367929812953e-06, 'learning_rate/action_model': 2.2573679298129525e-05, 'epoch': 2.08}) +[2026-06-18 00:31:47,994][starVLA.training.train_starvla][INFO] - Step 5048, Loss: {'train/loss': 0.30491015315055847, 'train/grad_norm_pre_clip': 22.810495376586914, 'timing/data': 0.009592387825250626, 'timing/model': 1.1384942717850208, 'learning_rate/qwen_vl_interface': 2.2556887538524514e-06, 'learning_rate/action_model': 2.2556887538524512e-05, 'epoch': 2.08}) +[2026-06-18 00:31:51,605][starVLA.training.train_starvla][INFO] - Step 5049, Loss: {'train/loss': 0.15178754925727844, 'train/grad_norm_pre_clip': 16.602022171020508, 'timing/data': 0.01029762253165245, 'timing/model': 0.9266502633690834, 'learning_rate/qwen_vl_interface': 2.2540101986158953e-06, 'learning_rate/action_model': 2.2540101986158953e-05, 'epoch': 2.08}) +[2026-06-18 00:31:55,492][starVLA.training.train_starvla][INFO] - Step 5050, Loss: {'train/loss': 0.08835994452238083, 'train/grad_norm_pre_clip': 10.377274513244629, 'timing/data': 0.01028232928365469, 'timing/model': 0.9330509658902884, 'learning_rate/qwen_vl_interface': 2.2523322644512506e-06, 'learning_rate/action_model': 2.2523322644512504e-05, 'epoch': 2.08}) +[2026-06-18 00:31:59,304][starVLA.training.train_starvla][INFO] - Step 5051, Loss: {'train/loss': 0.48380133509635925, 'train/grad_norm_pre_clip': 27.80746841430664, 'timing/data': 0.009159065783023834, 'timing/model': 0.9299969132989645, 'learning_rate/qwen_vl_interface': 2.2506549517063497e-06, 'learning_rate/action_model': 2.2506549517063496e-05, 'epoch': 2.08}) +[2026-06-18 00:32:03,151][starVLA.training.train_starvla][INFO] - Step 5052, Loss: {'train/loss': 0.31019502878189087, 'train/grad_norm_pre_clip': 19.954286575317383, 'timing/data': 0.012053760699927807, 'timing/model': 0.9328752337023616, 'learning_rate/qwen_vl_interface': 2.248978260728907e-06, 'learning_rate/action_model': 2.2489782607289067e-05, 'epoch': 2.08}) +[2026-06-18 00:32:07,289][starVLA.training.train_starvla][INFO] - Step 5053, Loss: {'train/loss': 0.435371071100235, 'train/grad_norm_pre_clip': 24.458696365356445, 'timing/data': 0.0075933923944830894, 'timing/model': 0.939862391911447, 'learning_rate/qwen_vl_interface': 2.2473021918664987e-06, 'learning_rate/action_model': 2.247302191866499e-05, 'epoch': 2.08}) +[2026-06-18 00:32:11,598][starVLA.training.train_starvla][INFO] - Step 5054, Loss: {'train/loss': 0.3053855001926422, 'train/grad_norm_pre_clip': 55.458251953125, 'timing/data': 0.009179954417049885, 'timing/model': 1.1338942237198353, 'learning_rate/qwen_vl_interface': 2.245626745466578e-06, 'learning_rate/action_model': 2.2456267454665782e-05, 'epoch': 2.08}) +[2026-06-18 00:32:15,606][starVLA.training.train_starvla][INFO] - Step 5055, Loss: {'train/loss': 0.3613050878047943, 'train/grad_norm_pre_clip': 25.238576889038086, 'timing/data': 0.010403584688901901, 'timing/model': 1.3258914956822991, 'learning_rate/qwen_vl_interface': 2.2439519218764657e-06, 'learning_rate/action_model': 2.2439519218764656e-05, 'epoch': 2.08}) +[2026-06-18 00:32:19,211][starVLA.training.train_starvla][INFO] - Step 5056, Loss: {'train/loss': 0.3228357434272766, 'train/grad_norm_pre_clip': 22.22091293334961, 'timing/data': 0.008680537343025208, 'timing/model': 0.9289645571261644, 'learning_rate/qwen_vl_interface': 2.242277721443355e-06, 'learning_rate/action_model': 2.2422777214433545e-05, 'epoch': 2.08}) +[2026-06-18 00:32:23,027][starVLA.training.train_starvla][INFO] - Step 5057, Loss: {'train/loss': 0.10106240212917328, 'train/grad_norm_pre_clip': 13.041857719421387, 'timing/data': 0.008795293979346752, 'timing/model': 0.9390653083100915, 'learning_rate/qwen_vl_interface': 2.2406041445143086e-06, 'learning_rate/action_model': 2.2406041445143086e-05, 'epoch': 2.08}) +[2026-06-18 00:32:26,832][starVLA.training.train_starvla][INFO] - Step 5058, Loss: {'train/loss': 0.46071165800094604, 'train/grad_norm_pre_clip': 28.76123809814453, 'timing/data': 0.010591716505587101, 'timing/model': 0.9328788630664349, 'learning_rate/qwen_vl_interface': 2.2389311914362614e-06, 'learning_rate/action_model': 2.2389311914362616e-05, 'epoch': 2.08}) +[2026-06-18 00:32:30,545][starVLA.training.train_starvla][INFO] - Step 5059, Loss: {'train/loss': 0.5601353049278259, 'train/grad_norm_pre_clip': 33.332916259765625, 'timing/data': 0.013195994310081005, 'timing/model': 0.9240873893722892, 'learning_rate/qwen_vl_interface': 2.2372588625560185e-06, 'learning_rate/action_model': 2.2372588625560185e-05, 'epoch': 2.08}) +[2026-06-18 00:32:34,371][starVLA.training.train_starvla][INFO] - Step 5060, Loss: {'train/loss': 0.41683411598205566, 'train/grad_norm_pre_clip': 22.691986083984375, 'timing/data': 0.011409195140004158, 'timing/model': 0.9254867387935519, 'learning_rate/qwen_vl_interface': 2.2355871582202544e-06, 'learning_rate/action_model': 2.2355871582202544e-05, 'epoch': 2.08}) +[2026-06-18 00:32:38,132][starVLA.training.train_starvla][INFO] - Step 5061, Loss: {'train/loss': 0.30197739601135254, 'train/grad_norm_pre_clip': 21.54622459411621, 'timing/data': 0.00977545790374279, 'timing/model': 1.0973808942362666, 'learning_rate/qwen_vl_interface': 2.2339160787755175e-06, 'learning_rate/action_model': 2.2339160787755174e-05, 'epoch': 2.08}) +[2026-06-18 00:32:41,725][starVLA.training.train_starvla][INFO] - Step 5062, Loss: {'train/loss': 0.22623857855796814, 'train/grad_norm_pre_clip': 20.522151947021484, 'timing/data': 0.010382799431681633, 'timing/model': 0.9264003122225404, 'learning_rate/qwen_vl_interface': 2.23224562456822e-06, 'learning_rate/action_model': 2.23224562456822e-05, 'epoch': 2.08}) +[2026-06-18 00:32:45,533][starVLA.training.train_starvla][INFO] - Step 5063, Loss: {'train/loss': 0.2563215494155884, 'train/grad_norm_pre_clip': 22.846569061279297, 'timing/data': 0.013400444760918617, 'timing/model': 0.926529323682189, 'learning_rate/qwen_vl_interface': 2.2305757959446538e-06, 'learning_rate/action_model': 2.2305757959446537e-05, 'epoch': 2.08}) +[2026-06-18 00:32:49,262][starVLA.training.train_starvla][INFO] - Step 5064, Loss: {'train/loss': 0.3406026363372803, 'train/grad_norm_pre_clip': 29.689979553222656, 'timing/data': 0.009707850404083729, 'timing/model': 0.936364782974124, 'learning_rate/qwen_vl_interface': 2.228906593250976e-06, 'learning_rate/action_model': 2.228906593250976e-05, 'epoch': 2.08}) +[2026-06-18 00:32:53,099][starVLA.training.train_starvla][INFO] - Step 5065, Loss: {'train/loss': 0.11207670718431473, 'train/grad_norm_pre_clip': 17.899795532226562, 'timing/data': 0.0121805714443326, 'timing/model': 0.9388283528387547, 'learning_rate/qwen_vl_interface': 2.2272380168332105e-06, 'learning_rate/action_model': 2.2272380168332103e-05, 'epoch': 2.08}) +[2026-06-18 00:32:56,960][starVLA.training.train_starvla][INFO] - Step 5066, Loss: {'train/loss': 0.4642117917537689, 'train/grad_norm_pre_clip': 25.102983474731445, 'timing/data': 0.009155994281172752, 'timing/model': 0.9315175916999578, 'learning_rate/qwen_vl_interface': 2.225570067037255e-06, 'learning_rate/action_model': 2.2255700670372548e-05, 'epoch': 2.09}) +[2026-06-18 00:33:00,768][starVLA.training.train_starvla][INFO] - Step 5067, Loss: {'train/loss': 0.12544718384742737, 'train/grad_norm_pre_clip': 11.61736011505127, 'timing/data': 0.007845408283174038, 'timing/model': 1.1404772736132145, 'learning_rate/qwen_vl_interface': 2.223902744208881e-06, 'learning_rate/action_model': 2.2239027442088808e-05, 'epoch': 2.09}) +[2026-06-18 00:33:04,549][starVLA.training.train_starvla][INFO] - Step 5068, Loss: {'train/loss': 0.2424650341272354, 'train/grad_norm_pre_clip': 19.322607040405273, 'timing/data': 0.008029275573790073, 'timing/model': 1.1171076148748398, 'learning_rate/qwen_vl_interface': 2.2222360486937238e-06, 'learning_rate/action_model': 2.2222360486937235e-05, 'epoch': 2.09}) +[2026-06-18 00:33:08,153][starVLA.training.train_starvla][INFO] - Step 5069, Loss: {'train/loss': 0.30973124504089355, 'train/grad_norm_pre_clip': 20.709165573120117, 'timing/data': 0.010614801198244095, 'timing/model': 0.9331304617226124, 'learning_rate/qwen_vl_interface': 2.220569980837292e-06, 'learning_rate/action_model': 2.220569980837292e-05, 'epoch': 2.09}) +[2026-06-18 00:33:11,986][starVLA.training.train_starvla][INFO] - Step 5070, Loss: {'train/loss': 0.4890899956226349, 'train/grad_norm_pre_clip': 25.0147762298584, 'timing/data': 0.011659407056868076, 'timing/model': 0.9320106012746692, 'learning_rate/qwen_vl_interface': 2.2189045409849625e-06, 'learning_rate/action_model': 2.2189045409849628e-05, 'epoch': 2.09}) +[2026-06-18 00:33:15,798][starVLA.training.train_starvla][INFO] - Step 5071, Loss: {'train/loss': 0.3785175681114197, 'train/grad_norm_pre_clip': 23.64021873474121, 'timing/data': 0.01249592937529087, 'timing/model': 0.9328449033200741, 'learning_rate/qwen_vl_interface': 2.2172397294819843e-06, 'learning_rate/action_model': 2.2172397294819843e-05, 'epoch': 2.09}) +[2026-06-18 00:33:19,566][starVLA.training.train_starvla][INFO] - Step 5072, Loss: {'train/loss': 0.3909740447998047, 'train/grad_norm_pre_clip': 22.38180160522461, 'timing/data': 0.013127870857715607, 'timing/model': 0.9255834240466356, 'learning_rate/qwen_vl_interface': 2.215575546673473e-06, 'learning_rate/action_model': 2.215575546673473e-05, 'epoch': 2.09}) +[2026-06-18 00:33:23,385][starVLA.training.train_starvla][INFO] - Step 5073, Loss: {'train/loss': 0.1762264221906662, 'train/grad_norm_pre_clip': 21.34832763671875, 'timing/data': 0.00953244511038065, 'timing/model': 0.9382840869948268, 'learning_rate/qwen_vl_interface': 2.213911992904416e-06, 'learning_rate/action_model': 2.213911992904416e-05, 'epoch': 2.09}) +[2026-06-18 00:33:27,497][starVLA.training.train_starvla][INFO] - Step 5074, Loss: {'train/loss': 0.1930040717124939, 'train/grad_norm_pre_clip': 17.64702033996582, 'timing/data': 0.012364031746983528, 'timing/model': 1.1543748322874308, 'learning_rate/qwen_vl_interface': 2.2122490685196703e-06, 'learning_rate/action_model': 2.2122490685196704e-05, 'epoch': 2.09}) +[2026-06-18 00:33:31,522][starVLA.training.train_starvla][INFO] - Step 5075, Loss: {'train/loss': 0.35461780428886414, 'train/grad_norm_pre_clip': 19.11121368408203, 'timing/data': 0.013806352391839027, 'timing/model': 1.1481625204905868, 'learning_rate/qwen_vl_interface': 2.210586773863961e-06, 'learning_rate/action_model': 2.210586773863961e-05, 'epoch': 2.09}) +[2026-06-18 00:33:35,181][starVLA.training.train_starvla][INFO] - Step 5076, Loss: {'train/loss': 0.13201043009757996, 'train/grad_norm_pre_clip': 14.275843620300293, 'timing/data': 0.012496065348386765, 'timing/model': 0.9414135431870818, 'learning_rate/qwen_vl_interface': 2.208925109281884e-06, 'learning_rate/action_model': 2.208925109281884e-05, 'epoch': 2.09}) +[2026-06-18 00:33:39,093][starVLA.training.train_starvla][INFO] - Step 5077, Loss: {'train/loss': 0.197841078042984, 'train/grad_norm_pre_clip': 14.191001892089844, 'timing/data': 0.013218273408710957, 'timing/model': 0.9311575712636113, 'learning_rate/qwen_vl_interface': 2.2072640751179026e-06, 'learning_rate/action_model': 2.2072640751179025e-05, 'epoch': 2.09}) +[2026-06-18 00:33:42,985][starVLA.training.train_starvla][INFO] - Step 5078, Loss: {'train/loss': 0.3875839114189148, 'train/grad_norm_pre_clip': 25.55549430847168, 'timing/data': 0.015073778107762337, 'timing/model': 1.0161755224689841, 'learning_rate/qwen_vl_interface': 2.205603671716356e-06, 'learning_rate/action_model': 2.2056036717163557e-05, 'epoch': 2.09}) +[2026-06-18 00:33:46,818][starVLA.training.train_starvla][INFO] - Step 5079, Loss: {'train/loss': 0.4390667974948883, 'train/grad_norm_pre_clip': 24.439146041870117, 'timing/data': 0.009805054403841496, 'timing/model': 0.9579247571527958, 'learning_rate/qwen_vl_interface': 2.2039438994214424e-06, 'learning_rate/action_model': 2.2039438994214424e-05, 'epoch': 2.09}) +[2026-06-18 00:33:50,718][starVLA.training.train_starvla][INFO] - Step 5080, Loss: {'train/loss': 0.3447254002094269, 'train/grad_norm_pre_clip': 24.010522842407227, 'timing/data': 0.009352130815386772, 'timing/model': 0.9347525360062718, 'learning_rate/qwen_vl_interface': 2.202284758577235e-06, 'learning_rate/action_model': 2.2022847585772347e-05, 'epoch': 2.09}) +[2026-06-18 00:33:54,547][starVLA.training.train_starvla][INFO] - Step 5081, Loss: {'train/loss': 0.43816831707954407, 'train/grad_norm_pre_clip': 24.221403121948242, 'timing/data': 0.010004119016230106, 'timing/model': 1.1381931426003575, 'learning_rate/qwen_vl_interface': 2.2006262495276774e-06, 'learning_rate/action_model': 2.2006262495276775e-05, 'epoch': 2.09}) +[2026-06-18 00:33:58,191][starVLA.training.train_starvla][INFO] - Step 5082, Loss: {'train/loss': 0.5277028679847717, 'train/grad_norm_pre_clip': 25.18400001525879, 'timing/data': 0.01126108132302761, 'timing/model': 0.9313544742763042, 'learning_rate/qwen_vl_interface': 2.1989683726165795e-06, 'learning_rate/action_model': 2.1989683726165795e-05, 'epoch': 2.09}) +[2026-06-18 00:34:02,012][starVLA.training.train_starvla][INFO] - Step 5083, Loss: {'train/loss': 0.36226946115493774, 'train/grad_norm_pre_clip': 34.869136810302734, 'timing/data': 0.008006698451936245, 'timing/model': 0.9359950171783566, 'learning_rate/qwen_vl_interface': 2.197311128187623e-06, 'learning_rate/action_model': 2.1973111281876232e-05, 'epoch': 2.09}) +[2026-06-18 00:34:05,810][starVLA.training.train_starvla][INFO] - Step 5084, Loss: {'train/loss': 0.35595375299453735, 'train/grad_norm_pre_clip': 19.747873306274414, 'timing/data': 0.008181327022612095, 'timing/model': 0.9397267932072282, 'learning_rate/qwen_vl_interface': 2.1956545165843517e-06, 'learning_rate/action_model': 2.1956545165843517e-05, 'epoch': 2.09}) +[2026-06-18 00:34:09,784][starVLA.training.train_starvla][INFO] - Step 5085, Loss: {'train/loss': 0.30647915601730347, 'train/grad_norm_pre_clip': 48.549232482910156, 'timing/data': 0.010938303545117378, 'timing/model': 0.9307221388444304, 'learning_rate/qwen_vl_interface': 2.1939985381501866e-06, 'learning_rate/action_model': 2.1939985381501866e-05, 'epoch': 2.09}) +[2026-06-18 00:34:13,589][starVLA.training.train_starvla][INFO] - Step 5086, Loss: {'train/loss': 0.39007285237312317, 'train/grad_norm_pre_clip': 25.993812561035156, 'timing/data': 0.010412650182843208, 'timing/model': 0.9345673965290189, 'learning_rate/qwen_vl_interface': 2.1923431932284135e-06, 'learning_rate/action_model': 2.1923431932284136e-05, 'epoch': 2.09}) +[2026-06-18 00:34:17,415][starVLA.training.train_starvla][INFO] - Step 5087, Loss: {'train/loss': 0.3321625888347626, 'train/grad_norm_pre_clip': 18.661705017089844, 'timing/data': 0.007707586511969566, 'timing/model': 0.9398181736469269, 'learning_rate/qwen_vl_interface': 2.190688482162187e-06, 'learning_rate/action_model': 2.190688482162187e-05, 'epoch': 2.09}) +[2026-06-18 00:34:21,235][starVLA.training.train_starvla][INFO] - Step 5088, Loss: {'train/loss': 0.31228792667388916, 'train/grad_norm_pre_clip': 27.497310638427734, 'timing/data': 0.009854125790297985, 'timing/model': 1.119200131855905, 'learning_rate/qwen_vl_interface': 2.189034405294529e-06, 'learning_rate/action_model': 2.189034405294529e-05, 'epoch': 2.09}) +[2026-06-18 00:34:25,606][starVLA.training.train_starvla][INFO] - Step 5089, Loss: {'train/loss': 0.5159744620323181, 'train/grad_norm_pre_clip': 24.488088607788086, 'timing/data': 0.011630943976342678, 'timing/model': 1.0189384426921606, 'learning_rate/qwen_vl_interface': 2.1873809629683333e-06, 'learning_rate/action_model': 2.1873809629683333e-05, 'epoch': 2.09}) +[2026-06-18 00:34:29,579][starVLA.training.train_starvla][INFO] - Step 5090, Loss: {'train/loss': 0.3287370800971985, 'train/grad_norm_pre_clip': 17.495027542114258, 'timing/data': 0.009103300049901009, 'timing/model': 0.9396008998155594, 'learning_rate/qwen_vl_interface': 2.185728155526359e-06, 'learning_rate/action_model': 2.1857281555263588e-05, 'epoch': 2.1}) +[2026-06-18 00:34:33,426][starVLA.training.train_starvla][INFO] - Step 5091, Loss: {'train/loss': 0.1998119056224823, 'train/grad_norm_pre_clip': 69.49382019042969, 'timing/data': 0.009618747979402542, 'timing/model': 0.9565261900424957, 'learning_rate/qwen_vl_interface': 2.184075983311234e-06, 'learning_rate/action_model': 2.184075983311234e-05, 'epoch': 2.1}) +[2026-06-18 00:34:37,257][starVLA.training.train_starvla][INFO] - Step 5092, Loss: {'train/loss': 0.3858000934123993, 'train/grad_norm_pre_clip': 22.762617111206055, 'timing/data': 0.009415526874363422, 'timing/model': 0.9521017307415605, 'learning_rate/qwen_vl_interface': 2.18242444666546e-06, 'learning_rate/action_model': 2.18242444666546e-05, 'epoch': 2.1}) +[2026-06-18 00:34:41,122][starVLA.training.train_starvla][INFO] - Step 5093, Loss: {'train/loss': 0.30385804176330566, 'train/grad_norm_pre_clip': 21.313072204589844, 'timing/data': 0.008951525203883648, 'timing/model': 0.9342416264116764, 'learning_rate/qwen_vl_interface': 2.1807735459313974e-06, 'learning_rate/action_model': 2.180773545931397e-05, 'epoch': 2.1}) +[2026-06-18 00:34:45,030][starVLA.training.train_starvla][INFO] - Step 5094, Loss: {'train/loss': 0.5113511681556702, 'train/grad_norm_pre_clip': 29.185932159423828, 'timing/data': 0.01059348601847887, 'timing/model': 1.2068417333066463, 'learning_rate/qwen_vl_interface': 2.1791232814512815e-06, 'learning_rate/action_model': 2.1791232814512815e-05, 'epoch': 2.1}) +[2026-06-18 00:34:48,847][starVLA.training.train_starvla][INFO] - Step 5095, Loss: {'train/loss': 0.2608710825443268, 'train/grad_norm_pre_clip': 21.053621292114258, 'timing/data': 0.011764690279960632, 'timing/model': 1.119490234181285, 'learning_rate/qwen_vl_interface': 2.1774736535672117e-06, 'learning_rate/action_model': 2.1774736535672115e-05, 'epoch': 2.1}) +[2026-06-18 00:34:52,498][starVLA.training.train_starvla][INFO] - Step 5096, Loss: {'train/loss': 0.22426865994930267, 'train/grad_norm_pre_clip': 18.66468048095703, 'timing/data': 0.020851493813097477, 'timing/model': 0.9616283001378179, 'learning_rate/qwen_vl_interface': 2.1758246626211617e-06, 'learning_rate/action_model': 2.1758246626211617e-05, 'epoch': 2.1}) +[2026-06-18 00:34:56,325][starVLA.training.train_starvla][INFO] - Step 5097, Loss: {'train/loss': 0.2535609006881714, 'train/grad_norm_pre_clip': 17.801042556762695, 'timing/data': 0.00957859493792057, 'timing/model': 0.9409129433333874, 'learning_rate/qwen_vl_interface': 2.174176308954969e-06, 'learning_rate/action_model': 2.174176308954969e-05, 'epoch': 2.1}) +[2026-06-18 00:35:00,282][starVLA.training.train_starvla][INFO] - Step 5098, Loss: {'train/loss': 0.1955837607383728, 'train/grad_norm_pre_clip': 18.122467041015625, 'timing/data': 0.010033931583166122, 'timing/model': 0.9526048712432384, 'learning_rate/qwen_vl_interface': 2.172528592910335e-06, 'learning_rate/action_model': 2.1725285929103348e-05, 'epoch': 2.1}) +[2026-06-18 00:35:04,165][starVLA.training.train_starvla][INFO] - Step 5099, Loss: {'train/loss': 0.30343395471572876, 'train/grad_norm_pre_clip': 23.59137535095215, 'timing/data': 0.008521396666765213, 'timing/model': 0.9394047642126679, 'learning_rate/qwen_vl_interface': 2.170881514828834e-06, 'learning_rate/action_model': 2.170881514828834e-05, 'epoch': 2.1}) +[2026-06-18 00:36:58,775][starVLA.training.train_starvla][INFO] - Step 5100, Loss: {'train/loss': 0.4431000351905823, 'train/grad_norm_pre_clip': 28.436630249023438, 'eval/loss': 0.4136471450328827, 'eval/latency_6/loss': 0.4136471450328827, 'eval/demon_fire_precision': 0.9597186700767263, 'eval/demon_fire_recall': 0.9702650290885585, 'eval/demon_fire_f1': 0.9649630343940854, 'eval/demon_fire_rate_model': 0.8703394546466333, 'eval/demon_fire_rate_teacher': 0.860879243183083, 'eval/demon_left_precision': 0.8729508196721312, 'eval/demon_left_recall': 0.7473684210526316, 'eval/demon_left_f1': 0.8052930056710775, 'eval/demon_left_rate_model': 0.13578185865331108, 'eval/demon_left_rate_teacher': 0.15859766277128548, 'eval/demon_right_precision': 0.8523725834797891, 'eval/demon_right_recall': 0.7450076804915514, 'eval/demon_right_f1': 0.7950819672131146, 'eval/demon_right_rate_model': 0.15831942125765164, 'eval/demon_right_rate_teacher': 0.18113522537562604, 'eval/demon_move_f1': 0.8001874864420961, 'eval/demon_cc_f1': 0.9088194815637527, 'eval/latency_6/demon_fire_precision': 0.9597186700767263, 'eval/latency_6/demon_fire_recall': 0.9702650290885585, 'eval/latency_6/demon_fire_f1': 0.9649630343940854, 'eval/latency_6/demon_fire_rate_model': 0.8703394546466333, 'eval/latency_6/demon_fire_rate_teacher': 0.860879243183083, 'eval/latency_6/demon_left_precision': 0.8729508196721312, 'eval/latency_6/demon_left_recall': 0.7473684210526316, 'eval/latency_6/demon_left_f1': 0.8052930056710775, 'eval/latency_6/demon_left_rate_model': 0.13578185865331108, 'eval/latency_6/demon_left_rate_teacher': 0.15859766277128548, 'eval/latency_6/demon_right_precision': 0.8523725834797891, 'eval/latency_6/demon_right_recall': 0.7450076804915514, 'eval/latency_6/demon_right_f1': 0.7950819672131146, 'eval/latency_6/demon_right_rate_model': 0.15831942125765164, 'eval/latency_6/demon_right_rate_teacher': 0.18113522537562604, 'eval/latency_6/demon_move_f1': 0.8001874864420961, 'eval/latency_6/demon_cc_f1': 0.9088194815637527, 'timing/data': 0.010738364420831203, 'timing/model': 0.9685739427804947, 'learning_rate/qwen_vl_interface': 2.1692350750519093e-06, 'learning_rate/action_model': 2.1692350750519092e-05, 'epoch': 2.1}) +[2026-06-18 00:36:58,781][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_5100_state +[2026-06-18 00:36:58,785][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 00:37:05,368][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_5100_state/model.safetensors +[2026-06-18 00:37:21,338][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_5100_state/optimizer.bin +[2026-06-18 00:37:21,339][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_5100_state/sampler.bin +[2026-06-18 00:37:21,339][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_5100_state/sampler_1.bin +[2026-06-18 00:37:21,342][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_5100_state/random_states_0.pkl +[2026-06-18 00:37:21,345][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 5100 +[2026-06-18 00:37:21,345][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 00:37:21,356][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 00:37:26,945][starVLA.training.train_starvla][INFO] - Step 5101, Loss: {'train/loss': 0.5722018480300903, 'train/grad_norm_pre_clip': 24.327863693237305, 'timing/data': 0.007913188077509403, 'timing/model': 1.3613589778542519, 'learning_rate/qwen_vl_interface': 2.167589273920869e-06, 'learning_rate/action_model': 2.167589273920869e-05, 'epoch': 2.1}) +[2026-06-18 00:37:32,774][starVLA.training.train_starvla][INFO] - Step 5102, Loss: {'train/loss': 0.46901726722717285, 'train/grad_norm_pre_clip': 28.368900299072266, 'timing/data': 0.008197584189474583, 'timing/model': 1.3773322077468038, 'learning_rate/qwen_vl_interface': 2.1659441117768888e-06, 'learning_rate/action_model': 2.1659441117768886e-05, 'epoch': 2.1}) +[2026-06-18 00:37:36,860][starVLA.training.train_starvla][INFO] - Step 5103, Loss: {'train/loss': 0.4262498915195465, 'train/grad_norm_pre_clip': 24.722522735595703, 'timing/data': 0.010466398671269417, 'timing/model': 0.9334088312461972, 'learning_rate/qwen_vl_interface': 2.164299588961013e-06, 'learning_rate/action_model': 2.1642995889610128e-05, 'epoch': 2.1}) +[2026-06-18 00:37:40,699][starVLA.training.train_starvla][INFO] - Step 5104, Loss: {'train/loss': 0.2552696764469147, 'train/grad_norm_pre_clip': 18.088077545166016, 'timing/data': 0.009930256754159927, 'timing/model': 1.1579849533736706, 'learning_rate/qwen_vl_interface': 2.162655705814153e-06, 'learning_rate/action_model': 2.1626557058141527e-05, 'epoch': 2.1}) +[2026-06-18 00:37:44,340][starVLA.training.train_starvla][INFO] - Step 5105, Loss: {'train/loss': 0.6387892365455627, 'train/grad_norm_pre_clip': 27.213973999023438, 'timing/data': 0.010658478364348412, 'timing/model': 0.9270980665460229, 'learning_rate/qwen_vl_interface': 2.161012462677087e-06, 'learning_rate/action_model': 2.161012462677087e-05, 'epoch': 2.1}) +[2026-06-18 00:37:48,212][starVLA.training.train_starvla][INFO] - Step 5106, Loss: {'train/loss': 0.3427516520023346, 'train/grad_norm_pre_clip': 41.1403694152832, 'timing/data': 0.013913650065660477, 'timing/model': 0.9308131150901318, 'learning_rate/qwen_vl_interface': 2.1593698598904612e-06, 'learning_rate/action_model': 2.1593698598904614e-05, 'epoch': 2.1}) +[2026-06-18 00:37:51,973][starVLA.training.train_starvla][INFO] - Step 5107, Loss: {'train/loss': 0.10534708201885223, 'train/grad_norm_pre_clip': 7.548707962036133, 'timing/data': 0.007226032204926014, 'timing/model': 0.9355974616482854, 'learning_rate/qwen_vl_interface': 2.1577278977947896e-06, 'learning_rate/action_model': 2.1577278977947897e-05, 'epoch': 2.1}) +[2026-06-18 00:37:55,778][starVLA.training.train_starvla][INFO] - Step 5108, Loss: {'train/loss': 0.2653570771217346, 'train/grad_norm_pre_clip': 12.571352005004883, 'timing/data': 0.013796540908515453, 'timing/model': 0.9337963378056884, 'learning_rate/qwen_vl_interface': 2.1560865767304524e-06, 'learning_rate/action_model': 2.156086576730452e-05, 'epoch': 2.1}) +[2026-06-18 00:37:59,573][starVLA.training.train_starvla][INFO] - Step 5109, Loss: {'train/loss': 0.24952982366085052, 'train/grad_norm_pre_clip': 25.535663604736328, 'timing/data': 0.009192446246743202, 'timing/model': 0.9281186535954475, 'learning_rate/qwen_vl_interface': 2.154445897037696e-06, 'learning_rate/action_model': 2.154445897037696e-05, 'epoch': 2.1}) +[2026-06-18 00:38:03,495][starVLA.training.train_starvla][INFO] - Step 5110, Loss: {'train/loss': 0.292034775018692, 'train/grad_norm_pre_clip': 19.36526107788086, 'timing/data': 0.009831754490733147, 'timing/model': 1.2383373836055398, 'learning_rate/qwen_vl_interface': 2.1528058590566353e-06, 'learning_rate/action_model': 2.1528058590566353e-05, 'epoch': 2.1}) +[2026-06-18 00:38:07,158][starVLA.training.train_starvla][INFO] - Step 5111, Loss: {'train/loss': 0.2044694721698761, 'train/grad_norm_pre_clip': 34.95950698852539, 'timing/data': 0.00994095392525196, 'timing/model': 0.9398526111617684, 'learning_rate/qwen_vl_interface': 2.1511664631272543e-06, 'learning_rate/action_model': 2.1511664631272545e-05, 'epoch': 2.1}) +[2026-06-18 00:38:10,987][starVLA.training.train_starvla][INFO] - Step 5112, Loss: {'train/loss': 0.11138533800840378, 'train/grad_norm_pre_clip': 9.94297981262207, 'timing/data': 0.007349894382059574, 'timing/model': 0.9439418576657772, 'learning_rate/qwen_vl_interface': 2.1495277095894015e-06, 'learning_rate/action_model': 2.1495277095894014e-05, 'epoch': 2.1}) +[2026-06-18 00:38:14,847][starVLA.training.train_starvla][INFO] - Step 5113, Loss: {'train/loss': 0.468708336353302, 'train/grad_norm_pre_clip': 31.181428909301758, 'timing/data': 0.007622099481523037, 'timing/model': 0.9332094555720687, 'learning_rate/qwen_vl_interface': 2.147889598782787e-06, 'learning_rate/action_model': 2.147889598782787e-05, 'epoch': 2.1}) +[2026-06-18 00:38:18,666][starVLA.training.train_starvla][INFO] - Step 5114, Loss: {'train/loss': 0.20112363994121552, 'train/grad_norm_pre_clip': 16.85316276550293, 'timing/data': 0.009920830838382244, 'timing/model': 0.9359959978610277, 'learning_rate/qwen_vl_interface': 2.1462521310469987e-06, 'learning_rate/action_model': 2.1462521310469987e-05, 'epoch': 2.11}) +[2026-06-18 00:38:22,451][starVLA.training.train_starvla][INFO] - Step 5115, Loss: {'train/loss': 0.3204483091831207, 'train/grad_norm_pre_clip': 22.137474060058594, 'timing/data': 0.01053538080304861, 'timing/model': 0.9366169888526201, 'learning_rate/qwen_vl_interface': 2.144615306721484e-06, 'learning_rate/action_model': 2.144615306721484e-05, 'epoch': 2.11}) +[2026-06-18 00:38:26,234][starVLA.training.train_starvla][INFO] - Step 5116, Loss: {'train/loss': 0.49971622228622437, 'train/grad_norm_pre_clip': 22.090932846069336, 'timing/data': 0.007868302054703236, 'timing/model': 1.1024428494274616, 'learning_rate/qwen_vl_interface': 2.1429791261455575e-06, 'learning_rate/action_model': 2.1429791261455572e-05, 'epoch': 2.11}) +[2026-06-18 00:38:30,046][starVLA.training.train_starvla][INFO] - Step 5117, Loss: {'train/loss': 0.3621178865432739, 'train/grad_norm_pre_clip': 20.87897300720215, 'timing/data': 0.010807444341480732, 'timing/model': 1.1090882839635015, 'learning_rate/qwen_vl_interface': 2.1413435896584006e-06, 'learning_rate/action_model': 2.1413435896584005e-05, 'epoch': 2.11}) +[2026-06-18 00:38:33,799][starVLA.training.train_starvla][INFO] - Step 5118, Loss: {'train/loss': 0.5749935507774353, 'train/grad_norm_pre_clip': 24.34373664855957, 'timing/data': 0.010498572140932083, 'timing/model': 0.9398341132327914, 'learning_rate/qwen_vl_interface': 2.1397086975990633e-06, 'learning_rate/action_model': 2.139708697599063e-05, 'epoch': 2.11}) +[2026-06-18 00:38:37,605][starVLA.training.train_starvla][INFO] - Step 5119, Loss: {'train/loss': 0.3054121732711792, 'train/grad_norm_pre_clip': 20.838815689086914, 'timing/data': 0.014018882066011429, 'timing/model': 0.939374890178442, 'learning_rate/qwen_vl_interface': 2.1380744503064597e-06, 'learning_rate/action_model': 2.1380744503064598e-05, 'epoch': 2.11}) +[2026-06-18 00:38:41,454][starVLA.training.train_starvla][INFO] - Step 5120, Loss: {'train/loss': 0.3657311797142029, 'train/grad_norm_pre_clip': 18.05131721496582, 'timing/data': 0.016377278603613377, 'timing/model': 0.9364802027121186, 'learning_rate/qwen_vl_interface': 2.1364408481193714e-06, 'learning_rate/action_model': 2.1364408481193712e-05, 'epoch': 2.11}) +[2026-06-18 00:38:45,480][starVLA.training.train_starvla][INFO] - Step 5121, Loss: {'train/loss': 0.37264907360076904, 'train/grad_norm_pre_clip': 21.952245712280273, 'timing/data': 0.009816518984735012, 'timing/model': 0.9627206707373261, 'learning_rate/qwen_vl_interface': 2.1348078913764454e-06, 'learning_rate/action_model': 2.1348078913764452e-05, 'epoch': 2.11}) +[2026-06-18 00:38:49,658][starVLA.training.train_starvla][INFO] - Step 5122, Loss: {'train/loss': 0.30559027194976807, 'train/grad_norm_pre_clip': 21.340118408203125, 'timing/data': 0.0080649359151721, 'timing/model': 0.9262235704809427, 'learning_rate/qwen_vl_interface': 2.1331755804161957e-06, 'learning_rate/action_model': 2.133175580416196e-05, 'epoch': 2.11}) +[2026-06-18 00:38:53,709][starVLA.training.train_starvla][INFO] - Step 5123, Loss: {'train/loss': 0.3254771828651428, 'train/grad_norm_pre_clip': 23.104650497436523, 'timing/data': 0.01046445220708847, 'timing/model': 1.3831295054405928, 'learning_rate/qwen_vl_interface': 2.131543915577002e-06, 'learning_rate/action_model': 2.131543915577002e-05, 'epoch': 2.11}) +[2026-06-18 00:38:57,376][starVLA.training.train_starvla][INFO] - Step 5124, Loss: {'train/loss': 0.6778079271316528, 'train/grad_norm_pre_clip': 23.665721893310547, 'timing/data': 0.010365406051278114, 'timing/model': 0.933794267475605, 'learning_rate/qwen_vl_interface': 2.129912897197108e-06, 'learning_rate/action_model': 2.1299128971971084e-05, 'epoch': 2.11}) +[2026-06-18 00:39:01,192][starVLA.training.train_starvla][INFO] - Step 5125, Loss: {'train/loss': 0.1293366551399231, 'train/grad_norm_pre_clip': 11.767934799194336, 'timing/data': 0.017238170839846134, 'timing/model': 0.9400315955281258, 'learning_rate/qwen_vl_interface': 2.1282825256146305e-06, 'learning_rate/action_model': 2.12828252561463e-05, 'epoch': 2.11}) +[2026-06-18 00:39:04,938][starVLA.training.train_starvla][INFO] - Step 5126, Loss: {'train/loss': 0.4268245995044708, 'train/grad_norm_pre_clip': 21.9659423828125, 'timing/data': 0.009890060871839523, 'timing/model': 0.9302521049976349, 'learning_rate/qwen_vl_interface': 2.1266528011675453e-06, 'learning_rate/action_model': 2.126652801167545e-05, 'epoch': 2.11}) +[2026-06-18 00:39:08,716][starVLA.training.train_starvla][INFO] - Step 5127, Loss: {'train/loss': 0.35750260949134827, 'train/grad_norm_pre_clip': 38.79032516479492, 'timing/data': 0.008816437795758247, 'timing/model': 0.9362418912351131, 'learning_rate/qwen_vl_interface': 2.125023724193694e-06, 'learning_rate/action_model': 2.1250237241936936e-05, 'epoch': 2.11}) +[2026-06-18 00:39:12,554][starVLA.training.train_starvla][INFO] - Step 5128, Loss: {'train/loss': 0.630133867263794, 'train/grad_norm_pre_clip': 24.968196868896484, 'timing/data': 0.007537733763456345, 'timing/model': 0.9331076927483082, 'learning_rate/qwen_vl_interface': 2.1233952950307857e-06, 'learning_rate/action_model': 2.1233952950307855e-05, 'epoch': 2.11}) +[2026-06-18 00:39:16,342][starVLA.training.train_starvla][INFO] - Step 5129, Loss: {'train/loss': 0.5432563424110413, 'train/grad_norm_pre_clip': 36.83263397216797, 'timing/data': 0.01130925677716732, 'timing/model': 1.1249428763985634, 'learning_rate/qwen_vl_interface': 2.1217675140163984e-06, 'learning_rate/action_model': 2.1217675140163983e-05, 'epoch': 2.11}) +[2026-06-18 00:39:20,313][starVLA.training.train_starvla][INFO] - Step 5130, Loss: {'train/loss': 0.3919680714607239, 'train/grad_norm_pre_clip': 22.74714469909668, 'timing/data': 0.01034510601311922, 'timing/model': 1.2222601287066936, 'learning_rate/qwen_vl_interface': 2.1201403814879715e-06, 'learning_rate/action_model': 2.1201403814879717e-05, 'epoch': 2.11}) +[2026-06-18 00:39:23,907][starVLA.training.train_starvla][INFO] - Step 5131, Loss: {'train/loss': 0.4817780554294586, 'train/grad_norm_pre_clip': 24.753562927246094, 'timing/data': 0.011481794528663158, 'timing/model': 0.9254499226808548, 'learning_rate/qwen_vl_interface': 2.1185138977828122e-06, 'learning_rate/action_model': 2.118513897782812e-05, 'epoch': 2.11}) +[2026-06-18 00:39:27,691][starVLA.training.train_starvla][INFO] - Step 5132, Loss: {'train/loss': 0.21643298864364624, 'train/grad_norm_pre_clip': 16.72314453125, 'timing/data': 0.01029921043664217, 'timing/model': 0.9276400785893202, 'learning_rate/qwen_vl_interface': 2.116888063238091e-06, 'learning_rate/action_model': 2.116888063238091e-05, 'epoch': 2.11}) +[2026-06-18 00:39:31,466][starVLA.training.train_starvla][INFO] - Step 5133, Loss: {'train/loss': 0.3504697382450104, 'train/grad_norm_pre_clip': 21.023500442504883, 'timing/data': 0.00922221876680851, 'timing/model': 0.9294382454827428, 'learning_rate/qwen_vl_interface': 2.1152628781908455e-06, 'learning_rate/action_model': 2.1152628781908455e-05, 'epoch': 2.11}) +[2026-06-18 00:39:35,229][starVLA.training.train_starvla][INFO] - Step 5134, Loss: {'train/loss': 0.3106091618537903, 'train/grad_norm_pre_clip': 17.414033889770508, 'timing/data': 0.012138601392507553, 'timing/model': 0.9383409917354584, 'learning_rate/qwen_vl_interface': 2.1136383429779787e-06, 'learning_rate/action_model': 2.1136383429779788e-05, 'epoch': 2.11}) +[2026-06-18 00:39:39,033][starVLA.training.train_starvla][INFO] - Step 5135, Loss: {'train/loss': 0.1199382096529007, 'train/grad_norm_pre_clip': 10.283283233642578, 'timing/data': 0.0077909743413329124, 'timing/model': 0.9338809866458178, 'learning_rate/qwen_vl_interface': 2.1120144579362582e-06, 'learning_rate/action_model': 2.1120144579362582e-05, 'epoch': 2.11}) +[2026-06-18 00:39:42,843][starVLA.training.train_starvla][INFO] - Step 5136, Loss: {'train/loss': 0.18406644463539124, 'train/grad_norm_pre_clip': 16.42471694946289, 'timing/data': 0.013507252559065819, 'timing/model': 1.1357998857274652, 'learning_rate/qwen_vl_interface': 2.1103912234023175e-06, 'learning_rate/action_model': 2.1103912234023175e-05, 'epoch': 2.11}) +[2026-06-18 00:39:46,523][starVLA.training.train_starvla][INFO] - Step 5137, Loss: {'train/loss': 0.3952082097530365, 'train/grad_norm_pre_clip': 27.538217544555664, 'timing/data': 0.011259148828685284, 'timing/model': 0.972612464800477, 'learning_rate/qwen_vl_interface': 2.1087686397126544e-06, 'learning_rate/action_model': 2.1087686397126545e-05, 'epoch': 2.11}) +[2026-06-18 00:39:50,343][starVLA.training.train_starvla][INFO] - Step 5138, Loss: {'train/loss': 0.3454214632511139, 'train/grad_norm_pre_clip': 16.60793113708496, 'timing/data': 0.010307498276233673, 'timing/model': 0.9406943768262863, 'learning_rate/qwen_vl_interface': 2.107146707203632e-06, 'learning_rate/action_model': 2.1071467072036323e-05, 'epoch': 2.11}) +[2026-06-18 00:39:54,179][starVLA.training.train_starvla][INFO] - Step 5139, Loss: {'train/loss': 0.5116565227508545, 'train/grad_norm_pre_clip': 26.045124053955078, 'timing/data': 0.01177817489951849, 'timing/model': 0.9653735272586346, 'learning_rate/qwen_vl_interface': 2.105525426211478e-06, 'learning_rate/action_model': 2.1055254262114778e-05, 'epoch': 2.12}) +[2026-06-18 00:39:58,110][starVLA.training.train_starvla][INFO] - Step 5140, Loss: {'train/loss': 0.27919161319732666, 'train/grad_norm_pre_clip': 16.208425521850586, 'timing/data': 0.009066847153007984, 'timing/model': 0.9340755334123969, 'learning_rate/qwen_vl_interface': 2.1039047970722894e-06, 'learning_rate/action_model': 2.1039047970722893e-05, 'epoch': 2.12}) +[2026-06-18 00:40:01,912][starVLA.training.train_starvla][INFO] - Step 5141, Loss: {'train/loss': 0.29508572816848755, 'train/grad_norm_pre_clip': 36.04615020751953, 'timing/data': 0.010741977952420712, 'timing/model': 0.9303278122097254, 'learning_rate/qwen_vl_interface': 2.1022848201220204e-06, 'learning_rate/action_model': 2.1022848201220204e-05, 'epoch': 2.12}) +[2026-06-18 00:40:05,732][starVLA.training.train_starvla][INFO] - Step 5142, Loss: {'train/loss': 0.3238457143306732, 'train/grad_norm_pre_clip': 25.07427215576172, 'timing/data': 0.011112617328763008, 'timing/model': 0.9383785668760538, 'learning_rate/qwen_vl_interface': 2.1006654956964935e-06, 'learning_rate/action_model': 2.1006654956964936e-05, 'epoch': 2.12}) +[2026-06-18 00:40:09,537][starVLA.training.train_starvla][INFO] - Step 5143, Loss: {'train/loss': 0.33571696281433105, 'train/grad_norm_pre_clip': 23.175735473632812, 'timing/data': 0.012592561542987823, 'timing/model': 1.1091269347816706, 'learning_rate/qwen_vl_interface': 2.0990468241314004e-06, 'learning_rate/action_model': 2.0990468241314002e-05, 'epoch': 2.12}) +[2026-06-18 00:40:13,149][starVLA.training.train_starvla][INFO] - Step 5144, Loss: {'train/loss': 0.3232767879962921, 'train/grad_norm_pre_clip': 19.704978942871094, 'timing/data': 0.011538425460457802, 'timing/model': 0.9413669761270285, 'learning_rate/qwen_vl_interface': 2.0974288057622905e-06, 'learning_rate/action_model': 2.0974288057622903e-05, 'epoch': 2.12}) +[2026-06-18 00:40:16,927][starVLA.training.train_starvla][INFO] - Step 5145, Loss: {'train/loss': 0.24620549380779266, 'train/grad_norm_pre_clip': 20.569849014282227, 'timing/data': 0.00972510315477848, 'timing/model': 0.927794405259192, 'learning_rate/qwen_vl_interface': 2.095811440924584e-06, 'learning_rate/action_model': 2.0958114409245835e-05, 'epoch': 2.12}) +[2026-06-18 00:40:20,966][starVLA.training.train_starvla][INFO] - Step 5146, Loss: {'train/loss': 0.24178025126457214, 'train/grad_norm_pre_clip': 11.98999309539795, 'timing/data': 0.009423021227121353, 'timing/model': 0.928740544244647, 'learning_rate/qwen_vl_interface': 2.0941947299535554e-06, 'learning_rate/action_model': 2.0941947299535552e-05, 'epoch': 2.12}) +[2026-06-18 00:40:24,786][starVLA.training.train_starvla][INFO] - Step 5147, Loss: {'train/loss': 0.3105493187904358, 'train/grad_norm_pre_clip': 20.777095794677734, 'timing/data': 0.009746063500642776, 'timing/model': 0.9307537665590644, 'learning_rate/qwen_vl_interface': 2.0925786731843573e-06, 'learning_rate/action_model': 2.0925786731843572e-05, 'epoch': 2.12}) +[2026-06-18 00:40:28,658][starVLA.training.train_starvla][INFO] - Step 5148, Loss: {'train/loss': 0.1784544289112091, 'train/grad_norm_pre_clip': 12.835103988647461, 'timing/data': 0.008126617409288883, 'timing/model': 0.9392753560096025, 'learning_rate/qwen_vl_interface': 2.0909632709519976e-06, 'learning_rate/action_model': 2.090963270951998e-05, 'epoch': 2.12}) +[2026-06-18 00:40:32,443][starVLA.training.train_starvla][INFO] - Step 5149, Loss: {'train/loss': 0.35091036558151245, 'train/grad_norm_pre_clip': 28.72347640991211, 'timing/data': 0.011124171316623688, 'timing/model': 1.1099670575931668, 'learning_rate/qwen_vl_interface': 2.0893485235913514e-06, 'learning_rate/action_model': 2.0893485235913514e-05, 'epoch': 2.12}) +[2026-06-18 00:40:36,276][starVLA.training.train_starvla][INFO] - Step 5150, Loss: {'train/loss': 0.3695598542690277, 'train/grad_norm_pre_clip': 22.387428283691406, 'timing/data': 0.01033022440969944, 'timing/model': 1.1163472775369883, 'learning_rate/qwen_vl_interface': 2.0877344314371575e-06, 'learning_rate/action_model': 2.0877344314371573e-05, 'epoch': 2.12}) +[2026-06-18 00:40:39,910][starVLA.training.train_starvla][INFO] - Step 5151, Loss: {'train/loss': 0.2144833505153656, 'train/grad_norm_pre_clip': 14.749406814575195, 'timing/data': 0.01020005438476801, 'timing/model': 0.9464552626013756, 'learning_rate/qwen_vl_interface': 2.086120994824018e-06, 'learning_rate/action_model': 2.0861209948240178e-05, 'epoch': 2.12}) +[2026-06-18 00:40:43,713][starVLA.training.train_starvla][INFO] - Step 5152, Loss: {'train/loss': 0.43839243054389954, 'train/grad_norm_pre_clip': 24.376863479614258, 'timing/data': 0.008485371246933937, 'timing/model': 0.9378766622394323, 'learning_rate/qwen_vl_interface': 2.0845082140864012e-06, 'learning_rate/action_model': 2.084508214086401e-05, 'epoch': 2.12}) +[2026-06-18 00:40:47,589][starVLA.training.train_starvla][INFO] - Step 5153, Loss: {'train/loss': 0.23955319821834564, 'train/grad_norm_pre_clip': 17.665180206298828, 'timing/data': 0.008747987449169159, 'timing/model': 0.9337070463225245, 'learning_rate/qwen_vl_interface': 2.0828960895586365e-06, 'learning_rate/action_model': 2.0828960895586364e-05, 'epoch': 2.12}) +[2026-06-18 00:40:51,446][starVLA.training.train_starvla][INFO] - Step 5154, Loss: {'train/loss': 0.28288963437080383, 'train/grad_norm_pre_clip': 19.773643493652344, 'timing/data': 0.013276521116495132, 'timing/model': 0.9885468762367964, 'learning_rate/qwen_vl_interface': 2.0812846215749227e-06, 'learning_rate/action_model': 2.0812846215749227e-05, 'epoch': 2.12}) +[2026-06-18 00:40:55,254][starVLA.training.train_starvla][INFO] - Step 5155, Loss: {'train/loss': 0.3549181818962097, 'train/grad_norm_pre_clip': 22.929214477539062, 'timing/data': 0.011808409355580807, 'timing/model': 0.9308289885520935, 'learning_rate/qwen_vl_interface': 2.079673810469316e-06, 'learning_rate/action_model': 2.0796738104693158e-05, 'epoch': 2.12}) +[2026-06-18 00:40:59,115][starVLA.training.train_starvla][INFO] - Step 5156, Loss: {'train/loss': 0.3622537851333618, 'train/grad_norm_pre_clip': 24.573060989379883, 'timing/data': 0.009088381193578243, 'timing/model': 1.1643368816003203, 'learning_rate/qwen_vl_interface': 2.0780636565757392e-06, 'learning_rate/action_model': 2.0780636565757392e-05, 'epoch': 2.12}) +[2026-06-18 00:41:02,910][starVLA.training.train_starvla][INFO] - Step 5157, Loss: {'train/loss': 0.23633526265621185, 'train/grad_norm_pre_clip': 17.73253631591797, 'timing/data': 0.013267382979393005, 'timing/model': 1.1194138946011662, 'learning_rate/qwen_vl_interface': 2.0764541602279777e-06, 'learning_rate/action_model': 2.0764541602279778e-05, 'epoch': 2.12}) +[2026-06-18 00:41:06,654][starVLA.training.train_starvla][INFO] - Step 5158, Loss: {'train/loss': 0.4823639392852783, 'train/grad_norm_pre_clip': 25.0543212890625, 'timing/data': 0.011739388108253479, 'timing/model': 0.9343290943652391, 'learning_rate/qwen_vl_interface': 2.0748453217596856e-06, 'learning_rate/action_model': 2.0748453217596856e-05, 'epoch': 2.12}) +[2026-06-18 00:41:11,567][starVLA.training.train_starvla][INFO] - Step 5159, Loss: {'train/loss': 0.2161007970571518, 'train/grad_norm_pre_clip': 22.363462448120117, 'timing/data': 0.008825615979731083, 'timing/model': 0.9386502290144563, 'learning_rate/qwen_vl_interface': 2.0732371415043766e-06, 'learning_rate/action_model': 2.0732371415043765e-05, 'epoch': 2.12}) +[2026-06-18 00:41:15,433][starVLA.training.train_starvla][INFO] - Step 5160, Loss: {'train/loss': 0.12459185719490051, 'train/grad_norm_pre_clip': 41.03467559814453, 'timing/data': 0.009943056851625443, 'timing/model': 0.9309126259759068, 'learning_rate/qwen_vl_interface': 2.071629619795425e-06, 'learning_rate/action_model': 2.071629619795425e-05, 'epoch': 2.12}) +[2026-06-18 00:41:19,249][starVLA.training.train_starvla][INFO] - Step 5161, Loss: {'train/loss': 0.39283615350723267, 'train/grad_norm_pre_clip': 23.852352142333984, 'timing/data': 0.010773221962153912, 'timing/model': 0.9395531518384814, 'learning_rate/qwen_vl_interface': 2.070022756966071e-06, 'learning_rate/action_model': 2.0700227569660712e-05, 'epoch': 2.12}) +[2026-06-18 00:41:23,041][starVLA.training.train_starvla][INFO] - Step 5162, Loss: {'train/loss': 0.3263407349586487, 'train/grad_norm_pre_clip': 29.918350219726562, 'timing/data': 0.00987695436924696, 'timing/model': 0.9391369372606277, 'learning_rate/qwen_vl_interface': 2.068416553349424e-06, 'learning_rate/action_model': 2.068416553349424e-05, 'epoch': 2.12}) +[2026-06-18 00:41:26,884][starVLA.training.train_starvla][INFO] - Step 5163, Loss: {'train/loss': 0.22802965342998505, 'train/grad_norm_pre_clip': 18.66595458984375, 'timing/data': 0.009228784590959549, 'timing/model': 1.1234391573816538, 'learning_rate/qwen_vl_interface': 2.0668110092784484e-06, 'learning_rate/action_model': 2.0668110092784484e-05, 'epoch': 2.13}) +[2026-06-18 00:41:30,578][starVLA.training.train_starvla][INFO] - Step 5164, Loss: {'train/loss': 0.16942645609378815, 'train/grad_norm_pre_clip': 13.989538192749023, 'timing/data': 0.008759047836065292, 'timing/model': 0.9268775973469019, 'learning_rate/qwen_vl_interface': 2.0652061250859758e-06, 'learning_rate/action_model': 2.0652061250859756e-05, 'epoch': 2.13}) +[2026-06-18 00:41:34,445][starVLA.training.train_starvla][INFO] - Step 5165, Loss: {'train/loss': 0.22108791768550873, 'train/grad_norm_pre_clip': 16.17855453491211, 'timing/data': 0.016490327194333076, 'timing/model': 0.9381511844694614, 'learning_rate/qwen_vl_interface': 2.0636019011047005e-06, 'learning_rate/action_model': 2.0636019011047003e-05, 'epoch': 2.13}) +[2026-06-18 00:41:38,234][starVLA.training.train_starvla][INFO] - Step 5166, Loss: {'train/loss': 0.2772611975669861, 'train/grad_norm_pre_clip': 15.232880592346191, 'timing/data': 0.007767856121063232, 'timing/model': 0.9304651152342558, 'learning_rate/qwen_vl_interface': 2.0619983376671796e-06, 'learning_rate/action_model': 2.0619983376671793e-05, 'epoch': 2.13}) +[2026-06-18 00:41:42,019][starVLA.training.train_starvla][INFO] - Step 5167, Loss: {'train/loss': 0.08542446792125702, 'train/grad_norm_pre_clip': 8.933736801147461, 'timing/data': 0.009755505248904228, 'timing/model': 0.9401925625279546, 'learning_rate/qwen_vl_interface': 2.060395435105834e-06, 'learning_rate/action_model': 2.060395435105834e-05, 'epoch': 2.13}) +[2026-06-18 00:41:45,785][starVLA.training.train_starvla][INFO] - Step 5168, Loss: {'train/loss': 0.3462505638599396, 'train/grad_norm_pre_clip': 21.729248046875, 'timing/data': 0.012267785146832466, 'timing/model': 0.9302129512652755, 'learning_rate/qwen_vl_interface': 2.0587931937529465e-06, 'learning_rate/action_model': 2.0587931937529464e-05, 'epoch': 2.13}) +[2026-06-18 00:41:49,678][starVLA.training.train_starvla][INFO] - Step 5169, Loss: {'train/loss': 0.1421305239200592, 'train/grad_norm_pre_clip': 13.737390518188477, 'timing/data': 0.012437614612281322, 'timing/model': 1.1875745989382267, 'learning_rate/qwen_vl_interface': 2.0571916139406648e-06, 'learning_rate/action_model': 2.0571916139406647e-05, 'epoch': 2.13}) +[2026-06-18 00:41:53,480][starVLA.training.train_starvla][INFO] - Step 5170, Loss: {'train/loss': 0.19775022566318512, 'train/grad_norm_pre_clip': 15.88776683807373, 'timing/data': 0.010103445500135422, 'timing/model': 1.0985022550448775, 'learning_rate/qwen_vl_interface': 2.055590696000996e-06, 'learning_rate/action_model': 2.0555906960009962e-05, 'epoch': 2.13}) +[2026-06-18 00:41:57,189][starVLA.training.train_starvla][INFO] - Step 5171, Loss: {'train/loss': 0.47832608222961426, 'train/grad_norm_pre_clip': 26.330677032470703, 'timing/data': 0.00799610372632742, 'timing/model': 0.9336274210363626, 'learning_rate/qwen_vl_interface': 2.0539904402658144e-06, 'learning_rate/action_model': 2.0539904402658142e-05, 'epoch': 2.13}) +[2026-06-18 00:42:01,024][starVLA.training.train_starvla][INFO] - Step 5172, Loss: {'train/loss': 0.23115840554237366, 'train/grad_norm_pre_clip': 19.043087005615234, 'timing/data': 0.01543425116688013, 'timing/model': 0.927842378616333, 'learning_rate/qwen_vl_interface': 2.052390847066851e-06, 'learning_rate/action_model': 2.052390847066851e-05, 'epoch': 2.13}) +[2026-06-18 00:42:04,993][starVLA.training.train_starvla][INFO] - Step 5173, Loss: {'train/loss': 0.20931002497673035, 'train/grad_norm_pre_clip': 17.24704360961914, 'timing/data': 0.010834376327693462, 'timing/model': 1.048549816943705, 'learning_rate/qwen_vl_interface': 2.0507919167357093e-06, 'learning_rate/action_model': 2.050791916735709e-05, 'epoch': 2.13}) +[2026-06-18 00:42:08,901][starVLA.training.train_starvla][INFO] - Step 5174, Loss: {'train/loss': 0.1603703796863556, 'train/grad_norm_pre_clip': 19.13848114013672, 'timing/data': 0.011584409512579441, 'timing/model': 0.9520008526742458, 'learning_rate/qwen_vl_interface': 2.049193649603844e-06, 'learning_rate/action_model': 2.049193649603844e-05, 'epoch': 2.13}) +[2026-06-18 00:42:12,754][starVLA.training.train_starvla][INFO] - Step 5175, Loss: {'train/loss': 0.4078960120677948, 'train/grad_norm_pre_clip': 30.678850173950195, 'timing/data': 0.01113168615847826, 'timing/model': 0.9428820377215743, 'learning_rate/qwen_vl_interface': 2.047596046002578e-06, 'learning_rate/action_model': 2.0475960460025778e-05, 'epoch': 2.13}) +[2026-06-18 00:42:16,974][starVLA.training.train_starvla][INFO] - Step 5176, Loss: {'train/loss': 0.19392412900924683, 'train/grad_norm_pre_clip': 55.54151916503906, 'timing/data': 0.011153463274240494, 'timing/model': 1.253783917054534, 'learning_rate/qwen_vl_interface': 2.0459991062630994e-06, 'learning_rate/action_model': 2.045999106263099e-05, 'epoch': 2.13}) +[2026-06-18 00:42:20,810][starVLA.training.train_starvla][INFO] - Step 5177, Loss: {'train/loss': 0.1181454285979271, 'train/grad_norm_pre_clip': 11.858354568481445, 'timing/data': 0.013354460708796978, 'timing/model': 1.1171115348115563, 'learning_rate/qwen_vl_interface': 2.0444028307164532e-06, 'learning_rate/action_model': 2.0444028307164534e-05, 'epoch': 2.13}) +[2026-06-18 00:42:24,430][starVLA.training.train_starvla][INFO] - Step 5178, Loss: {'train/loss': 0.3510754108428955, 'train/grad_norm_pre_clip': 21.880613327026367, 'timing/data': 0.012422790750861168, 'timing/model': 0.9320645201951265, 'learning_rate/qwen_vl_interface': 2.0428072196935516e-06, 'learning_rate/action_model': 2.0428072196935514e-05, 'epoch': 2.13}) +[2026-06-18 00:42:28,245][starVLA.training.train_starvla][INFO] - Step 5179, Loss: {'train/loss': 0.21197009086608887, 'train/grad_norm_pre_clip': 36.471885681152344, 'timing/data': 0.010335186496376991, 'timing/model': 0.9292250005528331, 'learning_rate/qwen_vl_interface': 2.041212273525161e-06, 'learning_rate/action_model': 2.041212273525161e-05, 'epoch': 2.13}) +[2026-06-18 00:42:32,096][starVLA.training.train_starvla][INFO] - Step 5180, Loss: {'train/loss': 0.2953706979751587, 'train/grad_norm_pre_clip': 12.975761413574219, 'timing/data': 0.010228042490780354, 'timing/model': 0.93669059779495, 'learning_rate/qwen_vl_interface': 2.03961799254192e-06, 'learning_rate/action_model': 2.03961799254192e-05, 'epoch': 2.13}) +[2026-06-18 00:42:35,994][starVLA.training.train_starvla][INFO] - Step 5181, Loss: {'train/loss': 0.21885471045970917, 'train/grad_norm_pre_clip': 46.21001052856445, 'timing/data': 0.008297691121697426, 'timing/model': 0.9623972475528717, 'learning_rate/qwen_vl_interface': 2.0380243770743246e-06, 'learning_rate/action_model': 2.0380243770743247e-05, 'epoch': 2.13}) +[2026-06-18 00:42:39,978][starVLA.training.train_starvla][INFO] - Step 5182, Loss: {'train/loss': 0.2396285980939865, 'train/grad_norm_pre_clip': 15.063769340515137, 'timing/data': 0.010662086308002472, 'timing/model': 0.9427052652463317, 'learning_rate/qwen_vl_interface': 2.0364314274527318e-06, 'learning_rate/action_model': 2.0364314274527314e-05, 'epoch': 2.13}) +[2026-06-18 00:42:43,779][starVLA.training.train_starvla][INFO] - Step 5183, Loss: {'train/loss': 0.42363834381103516, 'train/grad_norm_pre_clip': 30.967947006225586, 'timing/data': 0.011617287993431091, 'timing/model': 1.0973873836919665, 'learning_rate/qwen_vl_interface': 2.034839144007362e-06, 'learning_rate/action_model': 2.034839144007362e-05, 'epoch': 2.13}) +[2026-06-18 00:42:47,489][starVLA.training.train_starvla][INFO] - Step 5184, Loss: {'train/loss': 0.2657261788845062, 'train/grad_norm_pre_clip': 27.05191421508789, 'timing/data': 0.010714561678469181, 'timing/model': 1.0064049996435642, 'learning_rate/qwen_vl_interface': 2.033247527068297e-06, 'learning_rate/action_model': 2.0332475270682972e-05, 'epoch': 2.13}) +[2026-06-18 00:42:51,286][starVLA.training.train_starvla][INFO] - Step 5185, Loss: {'train/loss': 0.16249163448810577, 'train/grad_norm_pre_clip': 12.311704635620117, 'timing/data': 0.010018599219620228, 'timing/model': 0.936968607828021, 'learning_rate/qwen_vl_interface': 2.0316565769654816e-06, 'learning_rate/action_model': 2.0316565769654816e-05, 'epoch': 2.13}) +[2026-06-18 00:42:55,510][starVLA.training.train_starvla][INFO] - Step 5186, Loss: {'train/loss': 0.37040191888809204, 'train/grad_norm_pre_clip': 19.6367130279541, 'timing/data': 0.013487215153872967, 'timing/model': 1.3695551184937358, 'learning_rate/qwen_vl_interface': 2.0300662940287193e-06, 'learning_rate/action_model': 2.0300662940287192e-05, 'epoch': 2.13}) +[2026-06-18 00:42:59,380][starVLA.training.train_starvla][INFO] - Step 5187, Loss: {'train/loss': 0.22487801313400269, 'train/grad_norm_pre_clip': 17.69275665283203, 'timing/data': 0.016644342802464962, 'timing/model': 0.9315549861639738, 'learning_rate/qwen_vl_interface': 2.0284766785876812e-06, 'learning_rate/action_model': 2.0284766785876812e-05, 'epoch': 2.14}) +[2026-06-18 00:43:03,258][starVLA.training.train_starvla][INFO] - Step 5188, Loss: {'train/loss': 0.5659797191619873, 'train/grad_norm_pre_clip': 31.846900939941406, 'timing/data': 0.021138020791113377, 'timing/model': 0.9327633054926991, 'learning_rate/qwen_vl_interface': 2.026887730971894e-06, 'learning_rate/action_model': 2.0268877309718938e-05, 'epoch': 2.14}) +[2026-06-18 00:43:07,064][starVLA.training.train_starvla][INFO] - Step 5189, Loss: {'train/loss': 0.23641899228096008, 'train/grad_norm_pre_clip': 15.747188568115234, 'timing/data': 0.00988038256764412, 'timing/model': 0.9433888783678412, 'learning_rate/qwen_vl_interface': 2.0252994515107475e-06, 'learning_rate/action_model': 2.0252994515107472e-05, 'epoch': 2.14}) +[2026-06-18 00:43:10,943][starVLA.training.train_starvla][INFO] - Step 5190, Loss: {'train/loss': 0.775763988494873, 'train/grad_norm_pre_clip': 36.333988189697266, 'timing/data': 0.009192820638418198, 'timing/model': 1.1100313141942024, 'learning_rate/qwen_vl_interface': 2.023711840533493e-06, 'learning_rate/action_model': 2.023711840533493e-05, 'epoch': 2.14}) +[2026-06-18 00:43:14,648][starVLA.training.train_starvla][INFO] - Step 5191, Loss: {'train/loss': 0.13039377331733704, 'train/grad_norm_pre_clip': 15.448236465454102, 'timing/data': 0.009632665663957596, 'timing/model': 0.9437424344941974, 'learning_rate/qwen_vl_interface': 2.0221248983692475e-06, 'learning_rate/action_model': 2.0221248983692474e-05, 'epoch': 2.14}) +[2026-06-18 00:43:18,501][starVLA.training.train_starvla][INFO] - Step 5192, Loss: {'train/loss': 0.4310692548751831, 'train/grad_norm_pre_clip': 23.743764877319336, 'timing/data': 0.011039650067687035, 'timing/model': 0.9407072188332677, 'learning_rate/qwen_vl_interface': 2.020538625346984e-06, 'learning_rate/action_model': 2.020538625346984e-05, 'epoch': 2.14}) +[2026-06-18 00:43:22,304][starVLA.training.train_starvla][INFO] - Step 5193, Loss: {'train/loss': 0.39452657103538513, 'train/grad_norm_pre_clip': 23.756031036376953, 'timing/data': 0.013517195358872414, 'timing/model': 0.9302651034668088, 'learning_rate/qwen_vl_interface': 2.0189530217955382e-06, 'learning_rate/action_model': 2.018953021795538e-05, 'epoch': 2.14}) +[2026-06-18 00:43:26,146][starVLA.training.train_starvla][INFO] - Step 5194, Loss: {'train/loss': 0.171744704246521, 'train/grad_norm_pre_clip': 17.726470947265625, 'timing/data': 0.016841014847159386, 'timing/model': 0.9420549245551229, 'learning_rate/qwen_vl_interface': 2.0173680880436076e-06, 'learning_rate/action_model': 2.0173680880436076e-05, 'epoch': 2.14}) +[2026-06-18 00:43:30,024][starVLA.training.train_starvla][INFO] - Step 5195, Loss: {'train/loss': 0.13868428766727448, 'train/grad_norm_pre_clip': 14.948253631591797, 'timing/data': 0.012679269537329674, 'timing/model': 0.9537583794444799, 'learning_rate/qwen_vl_interface': 2.015783824419751e-06, 'learning_rate/action_model': 2.0157838244197508e-05, 'epoch': 2.14}) +[2026-06-18 00:43:33,906][starVLA.training.train_starvla][INFO] - Step 5196, Loss: {'train/loss': 0.3313175439834595, 'train/grad_norm_pre_clip': 19.506141662597656, 'timing/data': 0.012341666966676712, 'timing/model': 1.146231454797089, 'learning_rate/qwen_vl_interface': 2.0142002312523877e-06, 'learning_rate/action_model': 2.014200231252388e-05, 'epoch': 2.14}) +[2026-06-18 00:43:37,742][starVLA.training.train_starvla][INFO] - Step 5197, Loss: {'train/loss': 0.36814790964126587, 'train/grad_norm_pre_clip': 15.570351600646973, 'timing/data': 0.009846038185060024, 'timing/model': 1.118465586565435, 'learning_rate/qwen_vl_interface': 2.0126173088697988e-06, 'learning_rate/action_model': 2.0126173088697987e-05, 'epoch': 2.14}) +[2026-06-18 00:43:41,477][starVLA.training.train_starvla][INFO] - Step 5198, Loss: {'train/loss': 0.30932027101516724, 'train/grad_norm_pre_clip': 21.290626525878906, 'timing/data': 0.018517205491662025, 'timing/model': 0.9277776842936873, 'learning_rate/qwen_vl_interface': 2.0110350576001244e-06, 'learning_rate/action_model': 2.0110350576001245e-05, 'epoch': 2.14}) +[2026-06-18 00:43:45,715][starVLA.training.train_starvla][INFO] - Step 5199, Loss: {'train/loss': 0.25093239545822144, 'train/grad_norm_pre_clip': 17.602449417114258, 'timing/data': 0.013093595393002033, 'timing/model': 0.9328974783420563, 'learning_rate/qwen_vl_interface': 2.0094534777713684e-06, 'learning_rate/action_model': 2.0094534777713684e-05, 'epoch': 2.14}) +[2026-06-18 00:45:42,284][starVLA.training.train_starvla][INFO] - Step 5200, Loss: {'train/loss': 0.46971189975738525, 'train/grad_norm_pre_clip': 19.955698013305664, 'eval/loss': 0.41556790471076965, 'eval/latency_6/loss': 0.41556790471076965, 'eval/demon_fire_precision': 0.9499217527386542, 'eval/demon_fire_recall': 0.9809308338720103, 'eval/demon_fire_f1': 0.9651772936873907, 'eval/demon_fire_rate_model': 0.8889816360601002, 'eval/demon_fire_rate_teacher': 0.860879243183083, 'eval/demon_left_precision': 0.8421052631578947, 'eval/demon_left_recall': 0.7859649122807018, 'eval/demon_left_f1': 0.8130671506352086, 'eval/demon_left_rate_model': 0.14802448525319978, 'eval/demon_left_rate_teacher': 0.15859766277128548, 'eval/demon_right_precision': 0.808282208588957, 'eval/demon_right_recall': 0.8095238095238095, 'eval/demon_right_f1': 0.8089025326170376, 'eval/demon_right_rate_model': 0.18141346688925988, 'eval/demon_right_rate_teacher': 0.18113522537562604, 'eval/demon_move_f1': 0.8109848416261232, 'eval/demon_cc_f1': 0.9130952841589247, 'eval/latency_6/demon_fire_precision': 0.9499217527386542, 'eval/latency_6/demon_fire_recall': 0.9809308338720103, 'eval/latency_6/demon_fire_f1': 0.9651772936873907, 'eval/latency_6/demon_fire_rate_model': 0.8889816360601002, 'eval/latency_6/demon_fire_rate_teacher': 0.860879243183083, 'eval/latency_6/demon_left_precision': 0.8421052631578947, 'eval/latency_6/demon_left_recall': 0.7859649122807018, 'eval/latency_6/demon_left_f1': 0.8130671506352086, 'eval/latency_6/demon_left_rate_model': 0.14802448525319978, 'eval/latency_6/demon_left_rate_teacher': 0.15859766277128548, 'eval/latency_6/demon_right_precision': 0.808282208588957, 'eval/latency_6/demon_right_recall': 0.8095238095238095, 'eval/latency_6/demon_right_f1': 0.8089025326170376, 'eval/latency_6/demon_right_rate_model': 0.18141346688925988, 'eval/latency_6/demon_right_rate_teacher': 0.18113522537562604, 'eval/latency_6/demon_move_f1': 0.8109848416261232, 'eval/latency_6/demon_cc_f1': 0.9130952841589247, 'timing/data': 0.013430988416075706, 'timing/model': 0.9316662698984146, 'learning_rate/qwen_vl_interface': 2.007872569711394e-06, 'learning_rate/action_model': 2.007872569711394e-05, 'epoch': 2.14}) +[2026-06-18 00:45:42,288][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_5200_state +[2026-06-18 00:45:42,293][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 00:45:48,934][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_5200_state/model.safetensors +[2026-06-18 00:46:04,169][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_5200_state/optimizer.bin +[2026-06-18 00:46:04,170][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_5200_state/sampler.bin +[2026-06-18 00:46:04,170][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_5200_state/sampler_1.bin +[2026-06-18 00:46:04,172][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_5200_state/random_states_0.pkl +[2026-06-18 00:46:04,175][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 5200 +[2026-06-18 00:46:04,175][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 00:46:04,186][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 00:46:08,114][starVLA.training.train_starvla][INFO] - Step 5201, Loss: {'train/loss': 0.4606790542602539, 'train/grad_norm_pre_clip': 27.629932403564453, 'timing/data': 0.010589199140667915, 'timing/model': 0.9342586118727922, 'learning_rate/qwen_vl_interface': 2.0062923337479223e-06, 'learning_rate/action_model': 2.006292333747922e-05, 'epoch': 2.14}) +[2026-06-18 00:46:11,999][starVLA.training.train_starvla][INFO] - Step 5202, Loss: {'train/loss': 0.34785979986190796, 'train/grad_norm_pre_clip': 25.535402297973633, 'timing/data': 0.007920058444142342, 'timing/model': 0.9448137739673257, 'learning_rate/qwen_vl_interface': 2.0047127702085437e-06, 'learning_rate/action_model': 2.0047127702085437e-05, 'epoch': 2.14}) +[2026-06-18 00:46:15,801][starVLA.training.train_starvla][INFO] - Step 5203, Loss: {'train/loss': 0.37875986099243164, 'train/grad_norm_pre_clip': 107.79981994628906, 'timing/data': 0.013918053358793259, 'timing/model': 1.0010339794680476, 'learning_rate/qwen_vl_interface': 2.003133879420698e-06, 'learning_rate/action_model': 2.003133879420698e-05, 'epoch': 2.14}) +[2026-06-18 00:46:19,632][starVLA.training.train_starvla][INFO] - Step 5204, Loss: {'train/loss': 0.1818762719631195, 'train/grad_norm_pre_clip': 18.515552520751953, 'timing/data': 0.009517556056380272, 'timing/model': 0.9220780916512012, 'learning_rate/qwen_vl_interface': 2.001555661711691e-06, 'learning_rate/action_model': 2.001555661711691e-05, 'epoch': 2.14}) +[2026-06-18 00:46:23,408][starVLA.training.train_starvla][INFO] - Step 5205, Loss: {'train/loss': 0.44466495513916016, 'train/grad_norm_pre_clip': 26.269824981689453, 'timing/data': 0.009085199795663357, 'timing/model': 1.1153737790882587, 'learning_rate/qwen_vl_interface': 1.999978117408693e-06, 'learning_rate/action_model': 1.999978117408693e-05, 'epoch': 2.14}) +[2026-06-18 00:46:27,041][starVLA.training.train_starvla][INFO] - Step 5206, Loss: {'train/loss': 0.3025268614292145, 'train/grad_norm_pre_clip': 21.497739791870117, 'timing/data': 0.00979379378259182, 'timing/model': 0.9359293654561043, 'learning_rate/qwen_vl_interface': 1.9984012468387285e-06, 'learning_rate/action_model': 1.9984012468387285e-05, 'epoch': 2.14}) +[2026-06-18 00:46:30,908][starVLA.training.train_starvla][INFO] - Step 5207, Loss: {'train/loss': 0.18205587565898895, 'train/grad_norm_pre_clip': 17.472108840942383, 'timing/data': 0.009232403710484505, 'timing/model': 1.0044666528701782, 'learning_rate/qwen_vl_interface': 1.9968250503286857e-06, 'learning_rate/action_model': 1.996825050328686e-05, 'epoch': 2.14}) +[2026-06-18 00:46:34,633][starVLA.training.train_starvla][INFO] - Step 5208, Loss: {'train/loss': 0.21876206994056702, 'train/grad_norm_pre_clip': 19.17534637451172, 'timing/data': 0.00749587919563055, 'timing/model': 0.9289436191320419, 'learning_rate/qwen_vl_interface': 1.9952495282053073e-06, 'learning_rate/action_model': 1.9952495282053073e-05, 'epoch': 2.14}) +[2026-06-18 00:46:38,424][starVLA.training.train_starvla][INFO] - Step 5209, Loss: {'train/loss': 0.254841148853302, 'train/grad_norm_pre_clip': 22.69886016845703, 'timing/data': 0.015824537724256516, 'timing/model': 0.9314887188374996, 'learning_rate/qwen_vl_interface': 1.993674680795206e-06, 'learning_rate/action_model': 1.9936746807952058e-05, 'epoch': 2.14}) +[2026-06-18 00:46:42,244][starVLA.training.train_starvla][INFO] - Step 5210, Loss: {'train/loss': 0.272480845451355, 'train/grad_norm_pre_clip': 18.78219223022461, 'timing/data': 0.011736728250980377, 'timing/model': 0.9290655618533492, 'learning_rate/qwen_vl_interface': 1.992100508424847e-06, 'learning_rate/action_model': 1.9921005084248467e-05, 'epoch': 2.14}) +[2026-06-18 00:46:46,052][starVLA.training.train_starvla][INFO] - Step 5211, Loss: {'train/loss': 0.672991931438446, 'train/grad_norm_pre_clip': 25.40650177001953, 'timing/data': 0.010510042309761047, 'timing/model': 1.1325827557593584, 'learning_rate/qwen_vl_interface': 1.990527011420558e-06, 'learning_rate/action_model': 1.990527011420558e-05, 'epoch': 2.15}) +[2026-06-18 00:46:49,895][starVLA.training.train_starvla][INFO] - Step 5212, Loss: {'train/loss': 0.3192422688007355, 'train/grad_norm_pre_clip': 22.639896392822266, 'timing/data': 0.011454463936388493, 'timing/model': 1.1627548364922404, 'learning_rate/qwen_vl_interface': 1.9889541901085274e-06, 'learning_rate/action_model': 1.9889541901085273e-05, 'epoch': 2.15}) +[2026-06-18 00:46:53,579][starVLA.training.train_starvla][INFO] - Step 5213, Loss: {'train/loss': 0.14807851612567902, 'train/grad_norm_pre_clip': 13.037774085998535, 'timing/data': 0.014507899060845375, 'timing/model': 0.9338268302381039, 'learning_rate/qwen_vl_interface': 1.987382044814802e-06, 'learning_rate/action_model': 1.987382044814802e-05, 'epoch': 2.15}) +[2026-06-18 00:46:57,821][starVLA.training.train_starvla][INFO] - Step 5214, Loss: {'train/loss': 0.17686620354652405, 'train/grad_norm_pre_clip': 11.011870384216309, 'timing/data': 0.01750793308019638, 'timing/model': 0.9332034895196557, 'learning_rate/qwen_vl_interface': 1.9858105758652895e-06, 'learning_rate/action_model': 1.9858105758652892e-05, 'epoch': 2.15}) +[2026-06-18 00:47:01,886][starVLA.training.train_starvla][INFO] - Step 5215, Loss: {'train/loss': 0.273656964302063, 'train/grad_norm_pre_clip': 17.626455307006836, 'timing/data': 0.00720567349344492, 'timing/model': 0.9384490046650171, 'learning_rate/qwen_vl_interface': 1.9842397835857555e-06, 'learning_rate/action_model': 1.9842397835857554e-05, 'epoch': 2.15}) +[2026-06-18 00:47:05,681][starVLA.training.train_starvla][INFO] - Step 5216, Loss: {'train/loss': 0.2850792706012726, 'train/grad_norm_pre_clip': 24.809865951538086, 'timing/data': 0.009665797464549541, 'timing/model': 0.9341007545590401, 'learning_rate/qwen_vl_interface': 1.982669668301831e-06, 'learning_rate/action_model': 1.982669668301831e-05, 'epoch': 2.15}) +[2026-06-18 00:47:09,499][starVLA.training.train_starvla][INFO] - Step 5217, Loss: {'train/loss': 0.5433300733566284, 'train/grad_norm_pre_clip': 26.64320945739746, 'timing/data': 0.011804874986410141, 'timing/model': 0.9333877908065915, 'learning_rate/qwen_vl_interface': 1.9811002303389993e-06, 'learning_rate/action_model': 1.9811002303389993e-05, 'epoch': 2.15}) +[2026-06-18 00:47:13,368][starVLA.training.train_starvla][INFO] - Step 5218, Loss: {'train/loss': 0.1486731916666031, 'train/grad_norm_pre_clip': 18.147974014282227, 'timing/data': 0.011192603968083858, 'timing/model': 1.1692132269963622, 'learning_rate/qwen_vl_interface': 1.9795314700226068e-06, 'learning_rate/action_model': 1.9795314700226066e-05, 'epoch': 2.15}) +[2026-06-18 00:47:17,191][starVLA.training.train_starvla][INFO] - Step 5219, Loss: {'train/loss': 0.47058776021003723, 'train/grad_norm_pre_clip': 23.766420364379883, 'timing/data': 0.010480198077857494, 'timing/model': 1.140356868505478, 'learning_rate/qwen_vl_interface': 1.9779633876778587e-06, 'learning_rate/action_model': 1.9779633876778586e-05, 'epoch': 2.15}) +[2026-06-18 00:47:20,798][starVLA.training.train_starvla][INFO] - Step 5220, Loss: {'train/loss': 0.14426493644714355, 'train/grad_norm_pre_clip': 19.34302520751953, 'timing/data': 0.014313510619103909, 'timing/model': 0.9283786062151194, 'learning_rate/qwen_vl_interface': 1.9763959836298225e-06, 'learning_rate/action_model': 1.9763959836298225e-05, 'epoch': 2.15}) +[2026-06-18 00:47:24,634][starVLA.training.train_starvla][INFO] - Step 5221, Loss: {'train/loss': 0.27535340189933777, 'train/grad_norm_pre_clip': 17.557254791259766, 'timing/data': 0.013375100679695606, 'timing/model': 0.9312715604901314, 'learning_rate/qwen_vl_interface': 1.9748292582034228e-06, 'learning_rate/action_model': 1.9748292582034228e-05, 'epoch': 2.15}) +[2026-06-18 00:47:28,637][starVLA.training.train_starvla][INFO] - Step 5222, Loss: {'train/loss': 0.27930378913879395, 'train/grad_norm_pre_clip': 20.938602447509766, 'timing/data': 0.018644669093191624, 'timing/model': 1.1090133488178253, 'learning_rate/qwen_vl_interface': 1.9732632117234405e-06, 'learning_rate/action_model': 1.9732632117234404e-05, 'epoch': 2.15}) +[2026-06-18 00:47:32,735][starVLA.training.train_starvla][INFO] - Step 5223, Loss: {'train/loss': 0.254524290561676, 'train/grad_norm_pre_clip': 18.561845779418945, 'timing/data': 0.01402366440743208, 'timing/model': 1.1210100715979934, 'learning_rate/qwen_vl_interface': 1.9716978445145195e-06, 'learning_rate/action_model': 1.9716978445145193e-05, 'epoch': 2.15}) +[2026-06-18 00:47:36,694][starVLA.training.train_starvla][INFO] - Step 5224, Loss: {'train/loss': 0.12857328355312347, 'train/grad_norm_pre_clip': 13.854686737060547, 'timing/data': 0.010757542215287685, 'timing/model': 0.9310028227046132, 'learning_rate/qwen_vl_interface': 1.9701331569011634e-06, 'learning_rate/action_model': 1.9701331569011633e-05, 'epoch': 2.15}) +[2026-06-18 00:47:40,557][starVLA.training.train_starvla][INFO] - Step 5225, Loss: {'train/loss': 0.2502220571041107, 'train/grad_norm_pre_clip': 23.476072311401367, 'timing/data': 0.012793179601430893, 'timing/model': 1.1331474389880896, 'learning_rate/qwen_vl_interface': 1.968569149207734e-06, 'learning_rate/action_model': 1.968569149207734e-05, 'epoch': 2.15}) +[2026-06-18 00:47:44,162][starVLA.training.train_starvla][INFO] - Step 5226, Loss: {'train/loss': 0.3321683704853058, 'train/grad_norm_pre_clip': 24.659128189086914, 'timing/data': 0.009832818061113358, 'timing/model': 0.9345378940925002, 'learning_rate/qwen_vl_interface': 1.9670058217584505e-06, 'learning_rate/action_model': 1.9670058217584504e-05, 'epoch': 2.15}) +[2026-06-18 00:47:48,576][starVLA.training.train_starvla][INFO] - Step 5227, Loss: {'train/loss': 0.20577023923397064, 'train/grad_norm_pre_clip': 12.43785572052002, 'timing/data': 0.010033526457846165, 'timing/model': 1.246629430912435, 'learning_rate/qwen_vl_interface': 1.965443174877393e-06, 'learning_rate/action_model': 1.965443174877393e-05, 'epoch': 2.15}) +[2026-06-18 00:47:53,598][starVLA.training.train_starvla][INFO] - Step 5228, Loss: {'train/loss': 0.13241475820541382, 'train/grad_norm_pre_clip': 15.152078628540039, 'timing/data': 0.010409054346382618, 'timing/model': 1.7222414268180728, 'learning_rate/qwen_vl_interface': 1.9638812088885003e-06, 'learning_rate/action_model': 1.9638812088885e-05, 'epoch': 2.15}) +[2026-06-18 00:47:57,447][starVLA.training.train_starvla][INFO] - Step 5229, Loss: {'train/loss': 0.2808164060115814, 'train/grad_norm_pre_clip': 29.73945426940918, 'timing/data': 0.008810175582766533, 'timing/model': 0.9579554200172424, 'learning_rate/qwen_vl_interface': 1.9623199241155684e-06, 'learning_rate/action_model': 1.9623199241155683e-05, 'epoch': 2.15}) +[2026-06-18 00:48:01,286][starVLA.training.train_starvla][INFO] - Step 5230, Loss: {'train/loss': 0.2738308906555176, 'train/grad_norm_pre_clip': 23.9989070892334, 'timing/data': 0.012461223639547825, 'timing/model': 0.9228529008105397, 'learning_rate/qwen_vl_interface': 1.9607593208822536e-06, 'learning_rate/action_model': 1.9607593208822535e-05, 'epoch': 2.15}) +[2026-06-18 00:48:05,183][starVLA.training.train_starvla][INFO] - Step 5231, Loss: {'train/loss': 0.33047401905059814, 'train/grad_norm_pre_clip': 20.101558685302734, 'timing/data': 0.009595262818038464, 'timing/model': 0.9409153563901782, 'learning_rate/qwen_vl_interface': 1.9591993995120715e-06, 'learning_rate/action_model': 1.9591993995120712e-05, 'epoch': 2.15}) +[2026-06-18 00:48:08,993][starVLA.training.train_starvla][INFO] - Step 5232, Loss: {'train/loss': 0.530097246170044, 'train/grad_norm_pre_clip': 25.48806381225586, 'timing/data': 0.01104902196675539, 'timing/model': 1.1288478607311845, 'learning_rate/qwen_vl_interface': 1.9576401603283944e-06, 'learning_rate/action_model': 1.9576401603283944e-05, 'epoch': 2.15}) +[2026-06-18 00:48:12,581][starVLA.training.train_starvla][INFO] - Step 5233, Loss: {'train/loss': 0.16722238063812256, 'train/grad_norm_pre_clip': 18.461986541748047, 'timing/data': 0.010014891624450684, 'timing/model': 0.9252782044932246, 'learning_rate/qwen_vl_interface': 1.956081603654455e-06, 'learning_rate/action_model': 1.956081603654455e-05, 'epoch': 2.15}) +[2026-06-18 00:48:16,416][starVLA.training.train_starvla][INFO] - Step 5234, Loss: {'train/loss': 0.3947002589702606, 'train/grad_norm_pre_clip': 19.023771286010742, 'timing/data': 0.009679140523076057, 'timing/model': 0.9454497499391437, 'learning_rate/qwen_vl_interface': 1.9545237298133423e-06, 'learning_rate/action_model': 1.9545237298133424e-05, 'epoch': 2.15}) +[2026-06-18 00:48:20,205][starVLA.training.train_starvla][INFO] - Step 5235, Loss: {'train/loss': 0.17652170360088348, 'train/grad_norm_pre_clip': 18.399919509887695, 'timing/data': 0.011251824907958508, 'timing/model': 0.9303303407505155, 'learning_rate/qwen_vl_interface': 1.9529665391280096e-06, 'learning_rate/action_model': 1.95296653912801e-05, 'epoch': 2.15}) +[2026-06-18 00:48:24,192][starVLA.training.train_starvla][INFO] - Step 5236, Loss: {'train/loss': 0.2412160038948059, 'train/grad_norm_pre_clip': 29.925615310668945, 'timing/data': 0.01067731250077486, 'timing/model': 0.9305041059851646, 'learning_rate/qwen_vl_interface': 1.9514100319212594e-06, 'learning_rate/action_model': 1.9514100319212593e-05, 'epoch': 2.16}) +[2026-06-18 00:48:27,993][starVLA.training.train_starvla][INFO] - Step 5237, Loss: {'train/loss': 0.26724791526794434, 'train/grad_norm_pre_clip': 14.342777252197266, 'timing/data': 0.01765322405844927, 'timing/model': 0.9304208746179938, 'learning_rate/qwen_vl_interface': 1.949854208515757e-06, 'learning_rate/action_model': 1.9498542085157573e-05, 'epoch': 2.16}) +[2026-06-18 00:48:31,847][starVLA.training.train_starvla][INFO] - Step 5238, Loss: {'train/loss': 0.3600807785987854, 'train/grad_norm_pre_clip': 22.196937561035156, 'timing/data': 0.015419325791299343, 'timing/model': 1.1626363769173622, 'learning_rate/qwen_vl_interface': 1.9482990692340305e-06, 'learning_rate/action_model': 1.9482990692340305e-05, 'epoch': 2.16}) +[2026-06-18 00:48:35,924][starVLA.training.train_starvla][INFO] - Step 5239, Loss: {'train/loss': 0.2037455290555954, 'train/grad_norm_pre_clip': 14.480596542358398, 'timing/data': 0.011563273146748543, 'timing/model': 1.1598716946318746, 'learning_rate/qwen_vl_interface': 1.94674461439846e-06, 'learning_rate/action_model': 1.9467446143984598e-05, 'epoch': 2.16}) +[2026-06-18 00:48:39,543][starVLA.training.train_starvla][INFO] - Step 5240, Loss: {'train/loss': 0.22468066215515137, 'train/grad_norm_pre_clip': 16.408248901367188, 'timing/data': 0.008864170871675014, 'timing/model': 0.9411126291379333, 'learning_rate/qwen_vl_interface': 1.9451908443312863e-06, 'learning_rate/action_model': 1.9451908443312865e-05, 'epoch': 2.16}) +[2026-06-18 00:48:43,529][starVLA.training.train_starvla][INFO] - Step 5241, Loss: {'train/loss': 0.7456706166267395, 'train/grad_norm_pre_clip': 26.13546371459961, 'timing/data': 0.00949313584715128, 'timing/model': 0.9667487349361181, 'learning_rate/qwen_vl_interface': 1.943637759354604e-06, 'learning_rate/action_model': 1.9436377593546042e-05, 'epoch': 2.16}) +[2026-06-18 00:48:47,383][starVLA.training.train_starvla][INFO] - Step 5242, Loss: {'train/loss': 0.29698172211647034, 'train/grad_norm_pre_clip': 18.866926193237305, 'timing/data': 0.009336188435554504, 'timing/model': 0.9319992689415812, 'learning_rate/qwen_vl_interface': 1.942085359790373e-06, 'learning_rate/action_model': 1.942085359790373e-05, 'epoch': 2.16}) +[2026-06-18 00:48:51,279][starVLA.training.train_starvla][INFO] - Step 5243, Loss: {'train/loss': 0.20852316915988922, 'train/grad_norm_pre_clip': 20.62445831298828, 'timing/data': 0.009951220825314522, 'timing/model': 0.9442181624472141, 'learning_rate/qwen_vl_interface': 1.940533645960407e-06, 'learning_rate/action_model': 1.940533645960407e-05, 'epoch': 2.16}) +[2026-06-18 00:48:55,080][starVLA.training.train_starvla][INFO] - Step 5244, Loss: {'train/loss': 0.13565726578235626, 'train/grad_norm_pre_clip': 11.799288749694824, 'timing/data': 0.009348340332508087, 'timing/model': 0.9292029961943626, 'learning_rate/qwen_vl_interface': 1.9389826181863767e-06, 'learning_rate/action_model': 1.9389826181863767e-05, 'epoch': 2.16}) +[2026-06-18 00:48:58,829][starVLA.training.train_starvla][INFO] - Step 5245, Loss: {'train/loss': 0.24027997255325317, 'train/grad_norm_pre_clip': 15.241341590881348, 'timing/data': 0.01323420088738203, 'timing/model': 1.084548600949347, 'learning_rate/qwen_vl_interface': 1.937432276789813e-06, 'learning_rate/action_model': 1.9374322767898127e-05, 'epoch': 2.16}) +[2026-06-18 00:49:02,558][starVLA.training.train_starvla][INFO] - Step 5246, Loss: {'train/loss': 0.36321860551834106, 'train/grad_norm_pre_clip': 24.191022872924805, 'timing/data': 0.010960540734231472, 'timing/model': 1.0633982494473457, 'learning_rate/qwen_vl_interface': 1.935882622092103e-06, 'learning_rate/action_model': 1.935882622092103e-05, 'epoch': 2.16}) +[2026-06-18 00:49:06,149][starVLA.training.train_starvla][INFO] - Step 5247, Loss: {'train/loss': 0.35420891642570496, 'train/grad_norm_pre_clip': 15.6850004196167, 'timing/data': 0.008009319193661213, 'timing/model': 0.9298320030793548, 'learning_rate/qwen_vl_interface': 1.9343336544144907e-06, 'learning_rate/action_model': 1.9343336544144906e-05, 'epoch': 2.16}) +[2026-06-18 00:49:09,940][starVLA.training.train_starvla][INFO] - Step 5248, Loss: {'train/loss': 0.3280491530895233, 'train/grad_norm_pre_clip': 19.62546157836914, 'timing/data': 0.008284356445074081, 'timing/model': 0.9298700699582696, 'learning_rate/qwen_vl_interface': 1.9327853740780787e-06, 'learning_rate/action_model': 1.932785374078079e-05, 'epoch': 2.16}) +[2026-06-18 00:49:13,736][starVLA.training.train_starvla][INFO] - Step 5249, Loss: {'train/loss': 0.5355358123779297, 'train/grad_norm_pre_clip': 22.241735458374023, 'timing/data': 0.008762444369494915, 'timing/model': 0.9343006480485201, 'learning_rate/qwen_vl_interface': 1.9312377814038314e-06, 'learning_rate/action_model': 1.9312377814038312e-05, 'epoch': 2.16}) +[2026-06-18 00:49:17,444][starVLA.training.train_starvla][INFO] - Step 5250, Loss: {'train/loss': 0.12531012296676636, 'train/grad_norm_pre_clip': 11.51242733001709, 'timing/data': 0.010984964668750763, 'timing/model': 0.9265831848606467, 'learning_rate/qwen_vl_interface': 1.9296908767125612e-06, 'learning_rate/action_model': 1.929690876712561e-05, 'epoch': 2.16}) +[2026-06-18 00:49:21,227][starVLA.training.train_starvla][INFO] - Step 5251, Loss: {'train/loss': 0.2259030044078827, 'train/grad_norm_pre_clip': 16.91155433654785, 'timing/data': 0.010425373911857605, 'timing/model': 0.934564733877778, 'learning_rate/qwen_vl_interface': 1.928144660324945e-06, 'learning_rate/action_model': 1.928144660324945e-05, 'epoch': 2.16}) +[2026-06-18 00:49:25,120][starVLA.training.train_starvla][INFO] - Step 5252, Loss: {'train/loss': 0.153031587600708, 'train/grad_norm_pre_clip': 16.802326202392578, 'timing/data': 0.008184188045561314, 'timing/model': 1.1872832523658872, 'learning_rate/qwen_vl_interface': 1.9265991325615135e-06, 'learning_rate/action_model': 1.9265991325615135e-05, 'epoch': 2.16}) +[2026-06-18 00:49:28,739][starVLA.training.train_starvla][INFO] - Step 5253, Loss: {'train/loss': 0.2361692637205124, 'train/grad_norm_pre_clip': 18.867097854614258, 'timing/data': 0.009060189127922058, 'timing/model': 0.9328965749591589, 'learning_rate/qwen_vl_interface': 1.9250542937426604e-06, 'learning_rate/action_model': 1.9250542937426602e-05, 'epoch': 2.16}) +[2026-06-18 00:49:32,515][starVLA.training.train_starvla][INFO] - Step 5254, Loss: {'train/loss': 0.4904700517654419, 'train/grad_norm_pre_clip': 25.911447525024414, 'timing/data': 0.010156084783375263, 'timing/model': 0.931292668916285, 'learning_rate/qwen_vl_interface': 1.9235101441886307e-06, 'learning_rate/action_model': 1.923510144188631e-05, 'epoch': 2.16}) +[2026-06-18 00:49:36,240][starVLA.training.train_starvla][INFO] - Step 5255, Loss: {'train/loss': 0.16836027801036835, 'train/grad_norm_pre_clip': 17.580734252929688, 'timing/data': 0.01016897615045309, 'timing/model': 0.9339393628761172, 'learning_rate/qwen_vl_interface': 1.921966684219525e-06, 'learning_rate/action_model': 1.921966684219525e-05, 'epoch': 2.16}) +[2026-06-18 00:49:40,018][starVLA.training.train_starvla][INFO] - Step 5256, Loss: {'train/loss': 0.25472331047058105, 'train/grad_norm_pre_clip': 18.38391876220703, 'timing/data': 0.012565082870423794, 'timing/model': 0.9349381532520056, 'learning_rate/qwen_vl_interface': 1.920423914155308e-06, 'learning_rate/action_model': 1.920423914155308e-05, 'epoch': 2.16}) +[2026-06-18 00:49:43,893][starVLA.training.train_starvla][INFO] - Step 5257, Loss: {'train/loss': 0.25687897205352783, 'train/grad_norm_pre_clip': 24.461000442504883, 'timing/data': 0.010050431825220585, 'timing/model': 0.930535945110023, 'learning_rate/qwen_vl_interface': 1.9188818343157962e-06, 'learning_rate/action_model': 1.9188818343157964e-05, 'epoch': 2.16}) +[2026-06-18 00:49:47,690][starVLA.training.train_starvla][INFO] - Step 5258, Loss: {'train/loss': 0.1558842808008194, 'train/grad_norm_pre_clip': 14.216766357421875, 'timing/data': 0.016055641695857048, 'timing/model': 1.1296375002712011, 'learning_rate/qwen_vl_interface': 1.9173404450206654e-06, 'learning_rate/action_model': 1.9173404450206654e-05, 'epoch': 2.16}) +[2026-06-18 00:49:51,544][starVLA.training.train_starvla][INFO] - Step 5259, Loss: {'train/loss': 0.24334858357906342, 'train/grad_norm_pre_clip': 61.20888137817383, 'timing/data': 0.013910921290516853, 'timing/model': 1.1526785008609295, 'learning_rate/qwen_vl_interface': 1.9157997465894452e-06, 'learning_rate/action_model': 1.9157997465894454e-05, 'epoch': 2.16}) +[2026-06-18 00:49:56,297][starVLA.training.train_starvla][INFO] - Step 5260, Loss: {'train/loss': 0.2831804156303406, 'train/grad_norm_pre_clip': 14.716756820678711, 'timing/data': 0.008987813256680965, 'timing/model': 1.1466869041323662, 'learning_rate/qwen_vl_interface': 1.9142597393415264e-06, 'learning_rate/action_model': 1.914259739341526e-05, 'epoch': 2.17}) +[2026-06-18 00:50:00,103][starVLA.training.train_starvla][INFO] - Step 5261, Loss: {'train/loss': 0.18126362562179565, 'train/grad_norm_pre_clip': 16.230892181396484, 'timing/data': 0.01040674839168787, 'timing/model': 0.9299419922754169, 'learning_rate/qwen_vl_interface': 1.9127204235961524e-06, 'learning_rate/action_model': 1.9127204235961525e-05, 'epoch': 2.17}) +[2026-06-18 00:50:03,925][starVLA.training.train_starvla][INFO] - Step 5262, Loss: {'train/loss': 0.42091599106788635, 'train/grad_norm_pre_clip': 25.985673904418945, 'timing/data': 0.008312218822538853, 'timing/model': 0.9374735308811069, 'learning_rate/qwen_vl_interface': 1.911181799672426e-06, 'learning_rate/action_model': 1.9111817996724264e-05, 'epoch': 2.17}) +[2026-06-18 00:50:07,708][starVLA.training.train_starvla][INFO] - Step 5263, Loss: {'train/loss': 0.32315659523010254, 'train/grad_norm_pre_clip': 18.91147804260254, 'timing/data': 0.00889529474079609, 'timing/model': 0.941605162806809, 'learning_rate/qwen_vl_interface': 1.909643867889305e-06, 'learning_rate/action_model': 1.909643867889305e-05, 'epoch': 2.17}) +[2026-06-18 00:50:11,520][starVLA.training.train_starvla][INFO] - Step 5264, Loss: {'train/loss': 0.25530189275741577, 'train/grad_norm_pre_clip': 21.198463439941406, 'timing/data': 0.011486194096505642, 'timing/model': 0.9395378259941936, 'learning_rate/qwen_vl_interface': 1.908106628565605e-06, 'learning_rate/action_model': 1.9081066285656047e-05, 'epoch': 2.17}) +[2026-06-18 00:50:15,336][starVLA.training.train_starvla][INFO] - Step 5265, Loss: {'train/loss': 0.32188236713409424, 'train/grad_norm_pre_clip': 21.660064697265625, 'timing/data': 0.009973831474781036, 'timing/model': 1.1430053543299437, 'learning_rate/qwen_vl_interface': 1.9065700820199965e-06, 'learning_rate/action_model': 1.9065700820199963e-05, 'epoch': 2.17}) +[2026-06-18 00:50:19,125][starVLA.training.train_starvla][INFO] - Step 5266, Loss: {'train/loss': 0.1592358946800232, 'train/grad_norm_pre_clip': 13.975918769836426, 'timing/data': 0.008786838501691818, 'timing/model': 1.1104514338076115, 'learning_rate/qwen_vl_interface': 1.9050342285710063e-06, 'learning_rate/action_model': 1.9050342285710062e-05, 'epoch': 2.17}) +[2026-06-18 00:50:22,741][starVLA.training.train_starvla][INFO] - Step 5267, Loss: {'train/loss': 0.3908091187477112, 'train/grad_norm_pre_clip': 67.16117858886719, 'timing/data': 0.00991025660187006, 'timing/model': 0.9591133361682296, 'learning_rate/qwen_vl_interface': 1.9034990685370217e-06, 'learning_rate/action_model': 1.9034990685370217e-05, 'epoch': 2.17}) +[2026-06-18 00:50:26,570][starVLA.training.train_starvla][INFO] - Step 5268, Loss: {'train/loss': 0.3005369007587433, 'train/grad_norm_pre_clip': 19.155031204223633, 'timing/data': 0.010377810336649418, 'timing/model': 0.9263726472854614, 'learning_rate/qwen_vl_interface': 1.901964602236282e-06, 'learning_rate/action_model': 1.901964602236282e-05, 'epoch': 2.17}) +[2026-06-18 00:50:30,368][starVLA.training.train_starvla][INFO] - Step 5269, Loss: {'train/loss': 0.32672634720802307, 'train/grad_norm_pre_clip': 26.23712158203125, 'timing/data': 0.007151096127927303, 'timing/model': 0.9385590422898531, 'learning_rate/qwen_vl_interface': 1.900430829986882e-06, 'learning_rate/action_model': 1.900430829986882e-05, 'epoch': 2.17}) +[2026-06-18 00:50:34,153][starVLA.training.train_starvla][INFO] - Step 5270, Loss: {'train/loss': 0.2534712255001068, 'train/grad_norm_pre_clip': 18.961851119995117, 'timing/data': 0.007451347075402737, 'timing/model': 0.9273476349189878, 'learning_rate/qwen_vl_interface': 1.898897752106773e-06, 'learning_rate/action_model': 1.898897752106773e-05, 'epoch': 2.17}) +[2026-06-18 00:50:38,148][starVLA.training.train_starvla][INFO] - Step 5271, Loss: {'train/loss': 0.24303457140922546, 'train/grad_norm_pre_clip': 37.83977508544922, 'timing/data': 0.008112735114991665, 'timing/model': 0.92459069006145, 'learning_rate/qwen_vl_interface': 1.8973653689137683e-06, 'learning_rate/action_model': 1.8973653689137684e-05, 'epoch': 2.17}) +[2026-06-18 00:50:41,977][starVLA.training.train_starvla][INFO] - Step 5272, Loss: {'train/loss': 0.254852294921875, 'train/grad_norm_pre_clip': 18.925579071044922, 'timing/data': 0.011185159906744957, 'timing/model': 1.12056377902627, 'learning_rate/qwen_vl_interface': 1.8958336807255298e-06, 'learning_rate/action_model': 1.8958336807255297e-05, 'epoch': 2.17}) +[2026-06-18 00:50:45,577][starVLA.training.train_starvla][INFO] - Step 5273, Loss: {'train/loss': 0.19391073286533356, 'train/grad_norm_pre_clip': 17.091144561767578, 'timing/data': 0.011578422039747238, 'timing/model': 0.9300493337213993, 'learning_rate/qwen_vl_interface': 1.8943026878595785e-06, 'learning_rate/action_model': 1.8943026878595786e-05, 'epoch': 2.17}) +[2026-06-18 00:50:49,406][starVLA.training.train_starvla][INFO] - Step 5274, Loss: {'train/loss': 0.3260992467403412, 'train/grad_norm_pre_clip': 23.90595054626465, 'timing/data': 0.015498783439397812, 'timing/model': 0.9349707802757621, 'learning_rate/qwen_vl_interface': 1.8927723906332913e-06, 'learning_rate/action_model': 1.8927723906332912e-05, 'epoch': 2.17}) +[2026-06-18 00:50:53,253][starVLA.training.train_starvla][INFO] - Step 5275, Loss: {'train/loss': 0.1877126842737198, 'train/grad_norm_pre_clip': 24.04793357849121, 'timing/data': 0.00870781484991312, 'timing/model': 0.9265567269176245, 'learning_rate/qwen_vl_interface': 1.8912427893638996e-06, 'learning_rate/action_model': 1.8912427893638996e-05, 'epoch': 2.17}) +[2026-06-18 00:50:57,213][starVLA.training.train_starvla][INFO] - Step 5276, Loss: {'train/loss': 0.21981923282146454, 'train/grad_norm_pre_clip': 22.146032333374023, 'timing/data': 0.009729650802910328, 'timing/model': 1.0351449865847826, 'learning_rate/qwen_vl_interface': 1.8897138843684918e-06, 'learning_rate/action_model': 1.8897138843684917e-05, 'epoch': 2.17}) +[2026-06-18 00:51:01,496][starVLA.training.train_starvla][INFO] - Step 5277, Loss: {'train/loss': 0.4068413972854614, 'train/grad_norm_pre_clip': 24.76203155517578, 'timing/data': 0.011358494870364666, 'timing/model': 0.9559213928878307, 'learning_rate/qwen_vl_interface': 1.8881856759640122e-06, 'learning_rate/action_model': 1.8881856759640122e-05, 'epoch': 2.17}) +[2026-06-18 00:51:05,469][starVLA.training.train_starvla][INFO] - Step 5278, Loss: {'train/loss': 0.26863834261894226, 'train/grad_norm_pre_clip': 23.47315788269043, 'timing/data': 0.008920027874410152, 'timing/model': 0.9376088520511985, 'learning_rate/qwen_vl_interface': 1.8866581644672589e-06, 'learning_rate/action_model': 1.8866581644672586e-05, 'epoch': 2.17}) +[2026-06-18 00:51:09,278][starVLA.training.train_starvla][INFO] - Step 5279, Loss: {'train/loss': 0.10277967900037766, 'train/grad_norm_pre_clip': 10.292378425598145, 'timing/data': 0.00810981635004282, 'timing/model': 1.1409044358879328, 'learning_rate/qwen_vl_interface': 1.8851313501948872e-06, 'learning_rate/action_model': 1.885131350194887e-05, 'epoch': 2.17}) +[2026-06-18 00:51:12,899][starVLA.training.train_starvla][INFO] - Step 5280, Loss: {'train/loss': 0.31592047214508057, 'train/grad_norm_pre_clip': 23.69475746154785, 'timing/data': 0.009489217773079872, 'timing/model': 0.9293786808848381, 'learning_rate/qwen_vl_interface': 1.8836052334634072e-06, 'learning_rate/action_model': 1.883605233463407e-05, 'epoch': 2.17}) +[2026-06-18 00:51:16,698][starVLA.training.train_starvla][INFO] - Step 5281, Loss: {'train/loss': 0.23091118037700653, 'train/grad_norm_pre_clip': 21.507747650146484, 'timing/data': 0.008839888498187065, 'timing/model': 0.9313531937077641, 'learning_rate/qwen_vl_interface': 1.8820798145891834e-06, 'learning_rate/action_model': 1.8820798145891832e-05, 'epoch': 2.17}) +[2026-06-18 00:51:20,496][starVLA.training.train_starvla][INFO] - Step 5282, Loss: {'train/loss': 0.0911884605884552, 'train/grad_norm_pre_clip': 12.298067092895508, 'timing/data': 0.014200651086866856, 'timing/model': 0.9306454220786691, 'learning_rate/qwen_vl_interface': 1.88055509388844e-06, 'learning_rate/action_model': 1.88055509388844e-05, 'epoch': 2.17}) +[2026-06-18 00:51:24,297][starVLA.training.train_starvla][INFO] - Step 5283, Loss: {'train/loss': 0.19608275592327118, 'train/grad_norm_pre_clip': 25.00274658203125, 'timing/data': 0.012984415516257286, 'timing/model': 0.9311083881184459, 'learning_rate/qwen_vl_interface': 1.8790310716772525e-06, 'learning_rate/action_model': 1.8790310716772524e-05, 'epoch': 2.17}) +[2026-06-18 00:51:28,086][starVLA.training.train_starvla][INFO] - Step 5284, Loss: {'train/loss': 0.6756616830825806, 'train/grad_norm_pre_clip': 31.660297393798828, 'timing/data': 0.010767380706965923, 'timing/model': 0.9291861383244395, 'learning_rate/qwen_vl_interface': 1.8775077482715496e-06, 'learning_rate/action_model': 1.8775077482715494e-05, 'epoch': 2.18}) +[2026-06-18 00:51:31,874][starVLA.training.train_starvla][INFO] - Step 5285, Loss: {'train/loss': 0.26356106996536255, 'train/grad_norm_pre_clip': 31.046876907348633, 'timing/data': 0.012023691087961197, 'timing/model': 1.1295776814222336, 'learning_rate/qwen_vl_interface': 1.8759851239871176e-06, 'learning_rate/action_model': 1.8759851239871175e-05, 'epoch': 2.18}) +[2026-06-18 00:51:35,660][starVLA.training.train_starvla][INFO] - Step 5286, Loss: {'train/loss': 0.13025352358818054, 'train/grad_norm_pre_clip': 16.716094970703125, 'timing/data': 0.008989115245640278, 'timing/model': 1.1392771657556295, 'learning_rate/qwen_vl_interface': 1.8744631991396023e-06, 'learning_rate/action_model': 1.8744631991396023e-05, 'epoch': 2.18}) +[2026-06-18 00:51:39,249][starVLA.training.train_starvla][INFO] - Step 5287, Loss: {'train/loss': 0.18998637795448303, 'train/grad_norm_pre_clip': 17.704458236694336, 'timing/data': 0.010121162980794907, 'timing/model': 0.929308271035552, 'learning_rate/qwen_vl_interface': 1.8729419740444977e-06, 'learning_rate/action_model': 1.8729419740444977e-05, 'epoch': 2.18}) +[2026-06-18 00:51:43,030][starVLA.training.train_starvla][INFO] - Step 5288, Loss: {'train/loss': 0.8182227611541748, 'train/grad_norm_pre_clip': 35.385475158691406, 'timing/data': 0.010280716232955456, 'timing/model': 0.9291651686653495, 'learning_rate/qwen_vl_interface': 1.8714214490171554e-06, 'learning_rate/action_model': 1.8714214490171552e-05, 'epoch': 2.18}) +[2026-06-18 00:51:46,853][starVLA.training.train_starvla][INFO] - Step 5289, Loss: {'train/loss': 0.34555983543395996, 'train/grad_norm_pre_clip': 21.819093704223633, 'timing/data': 0.010110432282090187, 'timing/model': 0.9331310568377376, 'learning_rate/qwen_vl_interface': 1.8699016243727826e-06, 'learning_rate/action_model': 1.8699016243727827e-05, 'epoch': 2.18}) +[2026-06-18 00:51:50,603][starVLA.training.train_starvla][INFO] - Step 5290, Loss: {'train/loss': 0.19994953274726868, 'train/grad_norm_pre_clip': 14.464644432067871, 'timing/data': 0.00968774314969778, 'timing/model': 0.9314002590253949, 'learning_rate/qwen_vl_interface': 1.8683825004264402e-06, 'learning_rate/action_model': 1.86838250042644e-05, 'epoch': 2.18}) +[2026-06-18 00:51:54,358][starVLA.training.train_starvla][INFO] - Step 5291, Loss: {'train/loss': 0.3129419982433319, 'train/grad_norm_pre_clip': 24.60395050048828, 'timing/data': 0.008739426732063293, 'timing/model': 0.9298547422513366, 'learning_rate/qwen_vl_interface': 1.8668640774930444e-06, 'learning_rate/action_model': 1.8668640774930442e-05, 'epoch': 2.18}) +[2026-06-18 00:51:58,150][starVLA.training.train_starvla][INFO] - Step 5292, Loss: {'train/loss': 0.49638572335243225, 'train/grad_norm_pre_clip': 27.844097137451172, 'timing/data': 0.010213012807071209, 'timing/model': 1.1243192963302135, 'learning_rate/qwen_vl_interface': 1.8653463558873651e-06, 'learning_rate/action_model': 1.8653463558873652e-05, 'epoch': 2.18}) +[2026-06-18 00:52:01,905][starVLA.training.train_starvla][INFO] - Step 5293, Loss: {'train/loss': 0.25420206785202026, 'train/grad_norm_pre_clip': 69.8195571899414, 'timing/data': 0.01108888816088438, 'timing/model': 1.095783855766058, 'learning_rate/qwen_vl_interface': 1.863829335924029e-06, 'learning_rate/action_model': 1.8638293359240288e-05, 'epoch': 2.18}) +[2026-06-18 00:52:05,560][starVLA.training.train_starvla][INFO] - Step 5294, Loss: {'train/loss': 0.2814639210700989, 'train/grad_norm_pre_clip': 21.416166305541992, 'timing/data': 0.013066117651760578, 'timing/model': 0.9418024625629187, 'learning_rate/qwen_vl_interface': 1.8623130179175148e-06, 'learning_rate/action_model': 1.8623130179175146e-05, 'epoch': 2.18}) +[2026-06-18 00:52:09,425][starVLA.training.train_starvla][INFO] - Step 5295, Loss: {'train/loss': 0.3530118763446808, 'train/grad_norm_pre_clip': 26.715240478515625, 'timing/data': 0.010612879879772663, 'timing/model': 0.9857507450506091, 'learning_rate/qwen_vl_interface': 1.8607974021821573e-06, 'learning_rate/action_model': 1.8607974021821572e-05, 'epoch': 2.18}) +[2026-06-18 00:52:13,989][starVLA.training.train_starvla][INFO] - Step 5296, Loss: {'train/loss': 0.20716790854930878, 'train/grad_norm_pre_clip': 18.51349449157715, 'timing/data': 0.009386716410517693, 'timing/model': 1.0650159930810332, 'learning_rate/qwen_vl_interface': 1.8592824890321443e-06, 'learning_rate/action_model': 1.8592824890321442e-05, 'epoch': 2.18}) +[2026-06-18 00:52:17,806][starVLA.training.train_starvla][INFO] - Step 5297, Loss: {'train/loss': 0.5099676847457886, 'train/grad_norm_pre_clip': 28.677026748657227, 'timing/data': 0.009044724516570568, 'timing/model': 0.9376254444941878, 'learning_rate/qwen_vl_interface': 1.8577682787815223e-06, 'learning_rate/action_model': 1.8577682787815225e-05, 'epoch': 2.18}) +[2026-06-18 00:52:21,574][starVLA.training.train_starvla][INFO] - Step 5298, Loss: {'train/loss': 0.299689382314682, 'train/grad_norm_pre_clip': 21.667226791381836, 'timing/data': 0.013448404148221016, 'timing/model': 0.929009698331356, 'learning_rate/qwen_vl_interface': 1.856254771744185e-06, 'learning_rate/action_model': 1.856254771744185e-05, 'epoch': 2.18}) +[2026-06-18 00:52:25,459][starVLA.training.train_starvla][INFO] - Step 5299, Loss: {'train/loss': 0.5675879120826721, 'train/grad_norm_pre_clip': 19.69680404663086, 'timing/data': 0.007706096395850182, 'timing/model': 1.1630485318601131, 'learning_rate/qwen_vl_interface': 1.8547419682338838e-06, 'learning_rate/action_model': 1.8547419682338836e-05, 'epoch': 2.18}) +[2026-06-18 00:54:17,505][starVLA.training.train_starvla][INFO] - Step 5300, Loss: {'train/loss': 0.25373420119285583, 'train/grad_norm_pre_clip': 193.2663116455078, 'eval/loss': 0.411575585603714, 'eval/latency_6/loss': 0.411575585603714, 'eval/demon_fire_precision': 0.9490306441525954, 'eval/demon_fire_recall': 0.9809308338720103, 'eval/demon_fire_f1': 0.9647171010807374, 'eval/demon_fire_rate_model': 0.8898163606010017, 'eval/demon_fire_rate_teacher': 0.860879243183083, 'eval/demon_left_precision': 0.8729508196721312, 'eval/demon_left_recall': 0.7473684210526316, 'eval/demon_left_f1': 0.8052930056710775, 'eval/demon_left_rate_model': 0.13578185865331108, 'eval/demon_left_rate_teacher': 0.15859766277128548, 'eval/demon_right_precision': 0.8316666666666667, 'eval/demon_right_recall': 0.7665130568356375, 'eval/demon_right_f1': 0.7977617905675459, 'eval/demon_right_rate_model': 0.1669449081803005, 'eval/demon_right_rate_teacher': 0.18113522537562604, 'eval/demon_move_f1': 0.8015273981193117, 'eval/demon_cc_f1': 0.9091845431605864, 'eval/latency_6/demon_fire_precision': 0.9490306441525954, 'eval/latency_6/demon_fire_recall': 0.9809308338720103, 'eval/latency_6/demon_fire_f1': 0.9647171010807374, 'eval/latency_6/demon_fire_rate_model': 0.8898163606010017, 'eval/latency_6/demon_fire_rate_teacher': 0.860879243183083, 'eval/latency_6/demon_left_precision': 0.8729508196721312, 'eval/latency_6/demon_left_recall': 0.7473684210526316, 'eval/latency_6/demon_left_f1': 0.8052930056710775, 'eval/latency_6/demon_left_rate_model': 0.13578185865331108, 'eval/latency_6/demon_left_rate_teacher': 0.15859766277128548, 'eval/latency_6/demon_right_precision': 0.8316666666666667, 'eval/latency_6/demon_right_recall': 0.7665130568356375, 'eval/latency_6/demon_right_f1': 0.7977617905675459, 'eval/latency_6/demon_right_rate_model': 0.1669449081803005, 'eval/latency_6/demon_right_rate_teacher': 0.18113522537562604, 'eval/latency_6/demon_move_f1': 0.8015273981193117, 'eval/latency_6/demon_cc_f1': 0.9091845431605864, 'timing/data': 0.010629121214151382, 'timing/model': 0.9332863576710224, 'learning_rate/qwen_vl_interface': 1.8532298685642275e-06, 'learning_rate/action_model': 1.8532298685642274e-05, 'epoch': 2.18}) +[2026-06-18 00:54:17,541][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_5300_state +[2026-06-18 00:54:17,545][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 00:54:23,989][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_5300_state/model.safetensors +[2026-06-18 00:54:39,012][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_5300_state/optimizer.bin +[2026-06-18 00:54:39,013][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_5300_state/sampler.bin +[2026-06-18 00:54:39,013][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_5300_state/sampler_1.bin +[2026-06-18 00:54:39,015][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_5300_state/random_states_0.pkl +[2026-06-18 00:54:39,019][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 5300 +[2026-06-18 00:54:39,019][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 00:54:39,030][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 00:54:42,944][starVLA.training.train_starvla][INFO] - Step 5301, Loss: {'train/loss': 0.31373530626296997, 'train/grad_norm_pre_clip': 18.36298179626465, 'timing/data': 0.010105873458087444, 'timing/model': 1.1370230056345463, 'learning_rate/qwen_vl_interface': 1.8517184730486745e-06, 'learning_rate/action_model': 1.8517184730486746e-05, 'epoch': 2.18}) +[2026-06-18 00:54:46,709][starVLA.training.train_starvla][INFO] - Step 5302, Loss: {'train/loss': 0.5082043409347534, 'train/grad_norm_pre_clip': 25.16673469543457, 'timing/data': 0.01046757586300373, 'timing/model': 1.0992039274424314, 'learning_rate/qwen_vl_interface': 1.8502077820005408e-06, 'learning_rate/action_model': 1.8502077820005407e-05, 'epoch': 2.18}) +[2026-06-18 00:54:51,164][starVLA.training.train_starvla][INFO] - Step 5303, Loss: {'train/loss': 0.14153043925762177, 'train/grad_norm_pre_clip': 10.403196334838867, 'timing/data': 0.009626822546124458, 'timing/model': 0.9265746371820569, 'learning_rate/qwen_vl_interface': 1.8486977957329888e-06, 'learning_rate/action_model': 1.8486977957329886e-05, 'epoch': 2.18}) +[2026-06-18 00:54:54,993][starVLA.training.train_starvla][INFO] - Step 5304, Loss: {'train/loss': 0.28553271293640137, 'train/grad_norm_pre_clip': 18.152095794677734, 'timing/data': 0.008693790063261986, 'timing/model': 0.9281336730346084, 'learning_rate/qwen_vl_interface': 1.847188514559045e-06, 'learning_rate/action_model': 1.847188514559045e-05, 'epoch': 2.18}) +[2026-06-18 00:54:58,816][starVLA.training.train_starvla][INFO] - Step 5305, Loss: {'train/loss': 0.16985349357128143, 'train/grad_norm_pre_clip': 14.283449172973633, 'timing/data': 0.00875138770788908, 'timing/model': 0.9191862856969237, 'learning_rate/qwen_vl_interface': 1.8456799387915839e-06, 'learning_rate/action_model': 1.8456799387915838e-05, 'epoch': 2.18}) +[2026-06-18 00:55:02,633][starVLA.training.train_starvla][INFO] - Step 5306, Loss: {'train/loss': 0.4633523225784302, 'train/grad_norm_pre_clip': 24.38340187072754, 'timing/data': 0.009617181494832039, 'timing/model': 0.9272853443399072, 'learning_rate/qwen_vl_interface': 1.8441720687433345e-06, 'learning_rate/action_model': 1.8441720687433344e-05, 'epoch': 2.18}) +[2026-06-18 00:55:06,463][starVLA.training.train_starvla][INFO] - Step 5307, Loss: {'train/loss': 0.15334953367710114, 'train/grad_norm_pre_clip': 12.775619506835938, 'timing/data': 0.007224336266517639, 'timing/model': 0.9285762999206781, 'learning_rate/qwen_vl_interface': 1.8426649047268808e-06, 'learning_rate/action_model': 1.8426649047268807e-05, 'epoch': 2.18}) +[2026-06-18 00:55:10,242][starVLA.training.train_starvla][INFO] - Step 5308, Loss: {'train/loss': 0.2989012598991394, 'train/grad_norm_pre_clip': 20.176435470581055, 'timing/data': 0.007230821996927261, 'timing/model': 1.1125672599300742, 'learning_rate/qwen_vl_interface': 1.8411584470546582e-06, 'learning_rate/action_model': 1.841158447054658e-05, 'epoch': 2.18}) +[2026-06-18 00:55:13,827][starVLA.training.train_starvla][INFO] - Step 5309, Loss: {'train/loss': 0.3848634362220764, 'train/grad_norm_pre_clip': 21.787832260131836, 'timing/data': 0.007127095013856888, 'timing/model': 0.9286227859556675, 'learning_rate/qwen_vl_interface': 1.8396526960389575e-06, 'learning_rate/action_model': 1.8396526960389575e-05, 'epoch': 2.19}) +[2026-06-18 00:55:17,608][starVLA.training.train_starvla][INFO] - Step 5310, Loss: {'train/loss': 0.2589951157569885, 'train/grad_norm_pre_clip': 20.54120635986328, 'timing/data': 0.007412443868815899, 'timing/model': 0.9305825987830758, 'learning_rate/qwen_vl_interface': 1.8381476519919215e-06, 'learning_rate/action_model': 1.8381476519919212e-05, 'epoch': 2.19}) +[2026-06-18 00:55:21,383][starVLA.training.train_starvla][INFO] - Step 5311, Loss: {'train/loss': 0.22476091980934143, 'train/grad_norm_pre_clip': 15.726524353027344, 'timing/data': 0.006926036439836025, 'timing/model': 0.9350976925343275, 'learning_rate/qwen_vl_interface': 1.8366433152255511e-06, 'learning_rate/action_model': 1.8366433152255512e-05, 'epoch': 2.19}) +[2026-06-18 00:55:25,210][starVLA.training.train_starvla][INFO] - Step 5312, Loss: {'train/loss': 0.09330511838197708, 'train/grad_norm_pre_clip': 10.421571731567383, 'timing/data': 0.010537202470004559, 'timing/model': 0.9756727144122124, 'learning_rate/qwen_vl_interface': 1.8351396860516931e-06, 'learning_rate/action_model': 1.835139686051693e-05, 'epoch': 2.19}) +[2026-06-18 00:55:29,114][starVLA.training.train_starvla][INFO] - Step 5313, Loss: {'train/loss': 0.2703469395637512, 'train/grad_norm_pre_clip': 20.11842918395996, 'timing/data': 0.00841622892767191, 'timing/model': 0.9313999535515904, 'learning_rate/qwen_vl_interface': 1.8336367647820527e-06, 'learning_rate/action_model': 1.8336367647820528e-05, 'epoch': 2.19}) +[2026-06-18 00:55:32,894][starVLA.training.train_starvla][INFO] - Step 5314, Loss: {'train/loss': 0.22363802790641785, 'train/grad_norm_pre_clip': 20.7542781829834, 'timing/data': 0.008373611606657505, 'timing/model': 0.9248012695461512, 'learning_rate/qwen_vl_interface': 1.832134551728186e-06, 'learning_rate/action_model': 1.832134551728186e-05, 'epoch': 2.19}) +[2026-06-18 00:55:36,686][starVLA.training.train_starvla][INFO] - Step 5315, Loss: {'train/loss': 0.2527894675731659, 'train/grad_norm_pre_clip': 18.929683685302734, 'timing/data': 0.008649611845612526, 'timing/model': 1.1379950046539307, 'learning_rate/qwen_vl_interface': 1.8306330472015051e-06, 'learning_rate/action_model': 1.8306330472015053e-05, 'epoch': 2.19}) +[2026-06-18 00:55:40,393][starVLA.training.train_starvla][INFO] - Step 5316, Loss: {'train/loss': 0.31214845180511475, 'train/grad_norm_pre_clip': 24.88189697265625, 'timing/data': 0.010671328753232956, 'timing/model': 0.9422257775440812, 'learning_rate/qwen_vl_interface': 1.8291322515132757e-06, 'learning_rate/action_model': 1.8291322515132756e-05, 'epoch': 2.19}) +[2026-06-18 00:55:44,174][starVLA.training.train_starvla][INFO] - Step 5317, Loss: {'train/loss': 0.12691767513751984, 'train/grad_norm_pre_clip': 12.337936401367188, 'timing/data': 0.011565159074962139, 'timing/model': 0.9346267879009247, 'learning_rate/qwen_vl_interface': 1.8276321649746076e-06, 'learning_rate/action_model': 1.8276321649746076e-05, 'epoch': 2.19}) +[2026-06-18 00:55:47,974][starVLA.training.train_starvla][INFO] - Step 5318, Loss: {'train/loss': 0.11390597373247147, 'train/grad_norm_pre_clip': 29.886611938476562, 'timing/data': 0.007943637669086456, 'timing/model': 0.9280838929116726, 'learning_rate/qwen_vl_interface': 1.826132787896476e-06, 'learning_rate/action_model': 1.826132787896476e-05, 'epoch': 2.19}) +[2026-06-18 00:55:51,759][starVLA.training.train_starvla][INFO] - Step 5319, Loss: {'train/loss': 0.270035982131958, 'train/grad_norm_pre_clip': 16.59271812438965, 'timing/data': 0.009818251244723797, 'timing/model': 0.9305121377110481, 'learning_rate/qwen_vl_interface': 1.8246341205897018e-06, 'learning_rate/action_model': 1.8246341205897017e-05, 'epoch': 2.19}) +[2026-06-18 00:55:55,545][starVLA.training.train_starvla][INFO] - Step 5320, Loss: {'train/loss': 0.28307804465293884, 'train/grad_norm_pre_clip': 21.60126495361328, 'timing/data': 0.01279409322887659, 'timing/model': 0.9307380309328437, 'learning_rate/qwen_vl_interface': 1.8231361633649606e-06, 'learning_rate/action_model': 1.8231361633649606e-05, 'epoch': 2.19}) +[2026-06-18 00:55:59,321][starVLA.training.train_starvla][INFO] - Step 5321, Loss: {'train/loss': 0.09848528355360031, 'train/grad_norm_pre_clip': 10.894244194030762, 'timing/data': 0.009728966280817986, 'timing/model': 1.1135409325361252, 'learning_rate/qwen_vl_interface': 1.8216389165327797e-06, 'learning_rate/action_model': 1.8216389165327798e-05, 'epoch': 2.19}) +[2026-06-18 00:56:03,098][starVLA.training.train_starvla][INFO] - Step 5322, Loss: {'train/loss': 0.09587232023477554, 'train/grad_norm_pre_clip': 10.489283561706543, 'timing/data': 0.010968047194182873, 'timing/model': 1.1077622100710869, 'learning_rate/qwen_vl_interface': 1.8201423804035406e-06, 'learning_rate/action_model': 1.8201423804035405e-05, 'epoch': 2.19}) +[2026-06-18 00:56:06,683][starVLA.training.train_starvla][INFO] - Step 5323, Loss: {'train/loss': 0.3639560639858246, 'train/grad_norm_pre_clip': 17.567964553833008, 'timing/data': 0.009580312296748161, 'timing/model': 0.9287211513146758, 'learning_rate/qwen_vl_interface': 1.8186465552874769e-06, 'learning_rate/action_model': 1.8186465552874766e-05, 'epoch': 2.19}) +[2026-06-18 00:56:10,470][starVLA.training.train_starvla][INFO] - Step 5324, Loss: {'train/loss': 0.600166916847229, 'train/grad_norm_pre_clip': 33.15830612182617, 'timing/data': 0.009569618850946426, 'timing/model': 0.929788782261312, 'learning_rate/qwen_vl_interface': 1.8171514414946745e-06, 'learning_rate/action_model': 1.8171514414946745e-05, 'epoch': 2.19}) +[2026-06-18 00:56:14,333][starVLA.training.train_starvla][INFO] - Step 5325, Loss: {'train/loss': 0.2619425654411316, 'train/grad_norm_pre_clip': 22.320058822631836, 'timing/data': 0.009545344859361649, 'timing/model': 0.9488196894526482, 'learning_rate/qwen_vl_interface': 1.815657039335072e-06, 'learning_rate/action_model': 1.815657039335072e-05, 'epoch': 2.19}) +[2026-06-18 00:56:18,182][starVLA.training.train_starvla][INFO] - Step 5326, Loss: {'train/loss': 0.2745751142501831, 'train/grad_norm_pre_clip': 16.91657257080078, 'timing/data': 0.00828870851546526, 'timing/model': 0.9380817962810397, 'learning_rate/qwen_vl_interface': 1.8141633491184612e-06, 'learning_rate/action_model': 1.814163349118461e-05, 'epoch': 2.19}) +[2026-06-18 00:56:22,010][starVLA.training.train_starvla][INFO] - Step 5327, Loss: {'train/loss': 0.2733880877494812, 'train/grad_norm_pre_clip': 24.649755477905273, 'timing/data': 0.010687907226383686, 'timing/model': 0.956472959369421, 'learning_rate/qwen_vl_interface': 1.8126703711544848e-06, 'learning_rate/action_model': 1.812670371154485e-05, 'epoch': 2.19}) +[2026-06-18 00:56:25,877][starVLA.training.train_starvla][INFO] - Step 5328, Loss: {'train/loss': 0.16333071887493134, 'train/grad_norm_pre_clip': 20.577407836914062, 'timing/data': 0.009305083192884922, 'timing/model': 1.1511638956144452, 'learning_rate/qwen_vl_interface': 1.8111781057526381e-06, 'learning_rate/action_model': 1.811178105752638e-05, 'epoch': 2.19}) +[2026-06-18 00:56:29,603][starVLA.training.train_starvla][INFO] - Step 5329, Loss: {'train/loss': 0.17077897489070892, 'train/grad_norm_pre_clip': 15.625513076782227, 'timing/data': 0.010125869885087013, 'timing/model': 1.057319306768477, 'learning_rate/qwen_vl_interface': 1.8096865532222714e-06, 'learning_rate/action_model': 1.8096865532222714e-05, 'epoch': 2.19}) +[2026-06-18 00:56:33,210][starVLA.training.train_starvla][INFO] - Step 5330, Loss: {'train/loss': 0.30622535943984985, 'train/grad_norm_pre_clip': 19.027729034423828, 'timing/data': 0.00816381350159645, 'timing/model': 0.9347759773954749, 'learning_rate/qwen_vl_interface': 1.8081957138725861e-06, 'learning_rate/action_model': 1.808195713872586e-05, 'epoch': 2.19}) +[2026-06-18 00:56:37,020][starVLA.training.train_starvla][INFO] - Step 5331, Loss: {'train/loss': 0.2659517228603363, 'train/grad_norm_pre_clip': 20.33763313293457, 'timing/data': 0.011409912258386612, 'timing/model': 0.9591381521895528, 'learning_rate/qwen_vl_interface': 1.8067055880126311e-06, 'learning_rate/action_model': 1.806705588012631e-05, 'epoch': 2.19}) +[2026-06-18 00:56:40,857][starVLA.training.train_starvla][INFO] - Step 5332, Loss: {'train/loss': 0.11441608518362045, 'train/grad_norm_pre_clip': 14.173458099365234, 'timing/data': 0.009822197258472443, 'timing/model': 0.9290141994133592, 'learning_rate/qwen_vl_interface': 1.8052161759513114e-06, 'learning_rate/action_model': 1.8052161759513115e-05, 'epoch': 2.19}) +[2026-06-18 00:56:44,653][starVLA.training.train_starvla][INFO] - Step 5333, Loss: {'train/loss': 0.14945092797279358, 'train/grad_norm_pre_clip': 22.1532039642334, 'timing/data': 0.009407003410160542, 'timing/model': 0.9288788102567196, 'learning_rate/qwen_vl_interface': 1.8037274779973868e-06, 'learning_rate/action_model': 1.8037274779973867e-05, 'epoch': 2.2}) +[2026-06-18 00:56:48,411][starVLA.training.train_starvla][INFO] - Step 5334, Loss: {'train/loss': 0.16806814074516296, 'train/grad_norm_pre_clip': 19.230195999145508, 'timing/data': 0.009808383882045746, 'timing/model': 0.9298980589956045, 'learning_rate/qwen_vl_interface': 1.8022394944594643e-06, 'learning_rate/action_model': 1.802239494459464e-05, 'epoch': 2.2}) +[2026-06-18 00:56:52,204][starVLA.training.train_starvla][INFO] - Step 5335, Loss: {'train/loss': 0.14557504653930664, 'train/grad_norm_pre_clip': 12.373559951782227, 'timing/data': 0.008883684873580933, 'timing/model': 1.1283524129539728, 'learning_rate/qwen_vl_interface': 1.8007522256460062e-06, 'learning_rate/action_model': 1.8007522256460062e-05, 'epoch': 2.2}) +[2026-06-18 00:56:55,817][starVLA.training.train_starvla][INFO] - Step 5336, Loss: {'train/loss': 0.2102544754743576, 'train/grad_norm_pre_clip': 15.96548843383789, 'timing/data': 0.00968876387923956, 'timing/model': 0.9360096007585526, 'learning_rate/qwen_vl_interface': 1.7992656718653201e-06, 'learning_rate/action_model': 1.79926567186532e-05, 'epoch': 2.2}) +[2026-06-18 00:56:59,635][starVLA.training.train_starvla][INFO] - Step 5337, Loss: {'train/loss': 0.3274797201156616, 'train/grad_norm_pre_clip': 28.1644229888916, 'timing/data': 0.014540557749569416, 'timing/model': 0.9287520544603467, 'learning_rate/qwen_vl_interface': 1.797779833425575e-06, 'learning_rate/action_model': 1.797779833425575e-05, 'epoch': 2.2}) +[2026-06-18 00:57:03,384][starVLA.training.train_starvla][INFO] - Step 5338, Loss: {'train/loss': 0.17392565310001373, 'train/grad_norm_pre_clip': 18.204988479614258, 'timing/data': 0.011712513864040375, 'timing/model': 0.93381417542696, 'learning_rate/qwen_vl_interface': 1.7962947106347846e-06, 'learning_rate/action_model': 1.7962947106347845e-05, 'epoch': 2.2}) +[2026-06-18 00:57:07,394][starVLA.training.train_starvla][INFO] - Step 5339, Loss: {'train/loss': 0.5256865620613098, 'train/grad_norm_pre_clip': 29.411605834960938, 'timing/data': 0.009534548036754131, 'timing/model': 0.9337553381919861, 'learning_rate/qwen_vl_interface': 1.794810303800818e-06, 'learning_rate/action_model': 1.794810303800818e-05, 'epoch': 2.2}) +[2026-06-18 00:57:11,226][starVLA.training.train_starvla][INFO] - Step 5340, Loss: {'train/loss': 0.4118290841579437, 'train/grad_norm_pre_clip': 28.58262825012207, 'timing/data': 0.0072742849588394165, 'timing/model': 0.9340017950162292, 'learning_rate/qwen_vl_interface': 1.7933266132313924e-06, 'learning_rate/action_model': 1.7933266132313923e-05, 'epoch': 2.2}) +[2026-06-18 00:57:15,074][starVLA.training.train_starvla][INFO] - Step 5341, Loss: {'train/loss': 0.07555250823497772, 'train/grad_norm_pre_clip': 13.568737030029297, 'timing/data': 0.015493017621338367, 'timing/model': 1.1414939928799868, 'learning_rate/qwen_vl_interface': 1.7918436392340803e-06, 'learning_rate/action_model': 1.7918436392340802e-05, 'epoch': 2.2}) +[2026-06-18 00:57:18,870][starVLA.training.train_starvla][INFO] - Step 5342, Loss: {'train/loss': 0.105963796377182, 'train/grad_norm_pre_clip': 11.711179733276367, 'timing/data': 0.0090562142431736, 'timing/model': 1.123838547617197, 'learning_rate/qwen_vl_interface': 1.7903613821163022e-06, 'learning_rate/action_model': 1.790361382116302e-05, 'epoch': 2.2}) +[2026-06-18 00:57:22,481][starVLA.training.train_starvla][INFO] - Step 5343, Loss: {'train/loss': 0.3457641303539276, 'train/grad_norm_pre_clip': 30.90522003173828, 'timing/data': 0.01212856825441122, 'timing/model': 0.940788023173809, 'learning_rate/qwen_vl_interface': 1.7888798421853313e-06, 'learning_rate/action_model': 1.7888798421853313e-05, 'epoch': 2.2}) +[2026-06-18 00:57:26,295][starVLA.training.train_starvla][INFO] - Step 5344, Loss: {'train/loss': 0.2684004008769989, 'train/grad_norm_pre_clip': 16.810487747192383, 'timing/data': 0.009099295362830162, 'timing/model': 0.9378916975110769, 'learning_rate/qwen_vl_interface': 1.7873990197482962e-06, 'learning_rate/action_model': 1.787399019748296e-05, 'epoch': 2.2}) +[2026-06-18 00:57:30,087][starVLA.training.train_starvla][INFO] - Step 5345, Loss: {'train/loss': 0.20998725295066833, 'train/grad_norm_pre_clip': 18.484621047973633, 'timing/data': 0.009873646311461926, 'timing/model': 0.9321988746523857, 'learning_rate/qwen_vl_interface': 1.785918915112169e-06, 'learning_rate/action_model': 1.785918915112169e-05, 'epoch': 2.2}) +[2026-06-18 00:57:33,871][starVLA.training.train_starvla][INFO] - Step 5346, Loss: {'train/loss': 0.283870667219162, 'train/grad_norm_pre_clip': 22.07157325744629, 'timing/data': 0.012940119951963425, 'timing/model': 0.9312961269170046, 'learning_rate/qwen_vl_interface': 1.7844395285837773e-06, 'learning_rate/action_model': 1.7844395285837774e-05, 'epoch': 2.2}) +[2026-06-18 00:57:37,687][starVLA.training.train_starvla][INFO] - Step 5347, Loss: {'train/loss': 0.12816599011421204, 'train/grad_norm_pre_clip': 15.899602890014648, 'timing/data': 0.009223897010087967, 'timing/model': 0.9287872202694416, 'learning_rate/qwen_vl_interface': 1.7829608604698003e-06, 'learning_rate/action_model': 1.7829608604698e-05, 'epoch': 2.2}) +[2026-06-18 00:57:41,529][starVLA.training.train_starvla][INFO] - Step 5348, Loss: {'train/loss': 0.048076871782541275, 'train/grad_norm_pre_clip': 5.9634575843811035, 'timing/data': 0.015130921266973019, 'timing/model': 1.124459221959114, 'learning_rate/qwen_vl_interface': 1.7814829110767685e-06, 'learning_rate/action_model': 1.7814829110767685e-05, 'epoch': 2.2}) +[2026-06-18 00:57:45,313][starVLA.training.train_starvla][INFO] - Step 5349, Loss: {'train/loss': 0.397970974445343, 'train/grad_norm_pre_clip': 22.057140350341797, 'timing/data': 0.011060275137424469, 'timing/model': 1.113108397461474, 'learning_rate/qwen_vl_interface': 1.7800056807110627e-06, 'learning_rate/action_model': 1.7800056807110626e-05, 'epoch': 2.2}) +[2026-06-18 00:57:48,920][starVLA.training.train_starvla][INFO] - Step 5350, Loss: {'train/loss': 0.5994634032249451, 'train/grad_norm_pre_clip': 33.363155364990234, 'timing/data': 0.01467436458915472, 'timing/model': 0.9323638528585434, 'learning_rate/qwen_vl_interface': 1.7785291696789103e-06, 'learning_rate/action_model': 1.7785291696789102e-05, 'epoch': 2.2}) +[2026-06-18 00:57:52,772][starVLA.training.train_starvla][INFO] - Step 5351, Loss: {'train/loss': 0.13447628915309906, 'train/grad_norm_pre_clip': 11.058548927307129, 'timing/data': 0.008038224652409554, 'timing/model': 0.9327043918892741, 'learning_rate/qwen_vl_interface': 1.7770533782863983e-06, 'learning_rate/action_model': 1.7770533782863983e-05, 'epoch': 2.2}) +[2026-06-18 00:57:56,566][starVLA.training.train_starvla][INFO] - Step 5352, Loss: {'train/loss': 0.22855007648468018, 'train/grad_norm_pre_clip': 19.882749557495117, 'timing/data': 0.012242896482348442, 'timing/model': 0.9283848470076919, 'learning_rate/qwen_vl_interface': 1.7755783068394578e-06, 'learning_rate/action_model': 1.7755783068394577e-05, 'epoch': 2.2}) +[2026-06-18 00:58:00,279][starVLA.training.train_starvla][INFO] - Step 5353, Loss: {'train/loss': 0.44999510049819946, 'train/grad_norm_pre_clip': 26.181198120117188, 'timing/data': 0.007263108156621456, 'timing/model': 0.9325306536629796, 'learning_rate/qwen_vl_interface': 1.7741039556438733e-06, 'learning_rate/action_model': 1.7741039556438733e-05, 'epoch': 2.2}) +[2026-06-18 00:58:04,059][starVLA.training.train_starvla][INFO] - Step 5354, Loss: {'train/loss': 0.04613976553082466, 'train/grad_norm_pre_clip': 7.170132160186768, 'timing/data': 0.00969027727842331, 'timing/model': 0.9310349933803082, 'learning_rate/qwen_vl_interface': 1.7726303250052785e-06, 'learning_rate/action_model': 1.7726303250052784e-05, 'epoch': 2.2}) +[2026-06-18 00:58:07,854][starVLA.training.train_starvla][INFO] - Step 5355, Loss: {'train/loss': 0.37150463461875916, 'train/grad_norm_pre_clip': 72.96387481689453, 'timing/data': 0.009715515188872814, 'timing/model': 1.108152768574655, 'learning_rate/qwen_vl_interface': 1.7711574152291591e-06, 'learning_rate/action_model': 1.771157415229159e-05, 'epoch': 2.2}) +[2026-06-18 00:58:11,451][starVLA.training.train_starvla][INFO] - Step 5356, Loss: {'train/loss': 0.15373069047927856, 'train/grad_norm_pre_clip': 21.83112335205078, 'timing/data': 0.010235686786472797, 'timing/model': 0.9316946025937796, 'learning_rate/qwen_vl_interface': 1.7696852266208512e-06, 'learning_rate/action_model': 1.7696852266208512e-05, 'epoch': 2.2}) +[2026-06-18 00:58:15,210][starVLA.training.train_starvla][INFO] - Step 5357, Loss: {'train/loss': 0.4439290761947632, 'train/grad_norm_pre_clip': 25.04725456237793, 'timing/data': 0.01034857239574194, 'timing/model': 0.9278210559859872, 'learning_rate/qwen_vl_interface': 1.7682137594855408e-06, 'learning_rate/action_model': 1.768213759485541e-05, 'epoch': 2.21}) +[2026-06-18 00:58:18,949][starVLA.training.train_starvla][INFO] - Step 5358, Loss: {'train/loss': 0.4073401093482971, 'train/grad_norm_pre_clip': 19.755876541137695, 'timing/data': 0.009028845466673374, 'timing/model': 0.9287123717367649, 'learning_rate/qwen_vl_interface': 1.7667430141282649e-06, 'learning_rate/action_model': 1.7667430141282648e-05, 'epoch': 2.21}) +[2026-06-18 00:58:22,710][starVLA.training.train_starvla][INFO] - Step 5359, Loss: {'train/loss': 0.3403368592262268, 'train/grad_norm_pre_clip': 20.86366844177246, 'timing/data': 0.010380271822214127, 'timing/model': 0.9276062026619911, 'learning_rate/qwen_vl_interface': 1.7652729908539101e-06, 'learning_rate/action_model': 1.76527299085391e-05, 'epoch': 2.21}) +[2026-06-18 00:58:26,501][starVLA.training.train_starvla][INFO] - Step 5360, Loss: {'train/loss': 0.12009169161319733, 'train/grad_norm_pre_clip': 24.363332748413086, 'timing/data': 0.013968246057629585, 'timing/model': 0.9324908778071404, 'learning_rate/qwen_vl_interface': 1.7638036899672137e-06, 'learning_rate/action_model': 1.7638036899672136e-05, 'epoch': 2.21}) +[2026-06-18 00:58:30,253][starVLA.training.train_starvla][INFO] - Step 5361, Loss: {'train/loss': 0.1596314162015915, 'train/grad_norm_pre_clip': 17.339799880981445, 'timing/data': 0.012725036591291428, 'timing/model': 0.931165567599237, 'learning_rate/qwen_vl_interface': 1.7623351117727628e-06, 'learning_rate/action_model': 1.7623351117727626e-05, 'epoch': 2.21}) +[2026-06-18 00:58:34,066][starVLA.training.train_starvla][INFO] - Step 5362, Loss: {'train/loss': 0.30743208527565, 'train/grad_norm_pre_clip': 19.328901290893555, 'timing/data': 0.00935477577149868, 'timing/model': 1.153101067058742, 'learning_rate/qwen_vl_interface': 1.7608672565749973e-06, 'learning_rate/action_model': 1.760867256574997e-05, 'epoch': 2.21}) +[2026-06-18 00:58:37,702][starVLA.training.train_starvla][INFO] - Step 5363, Loss: {'train/loss': 0.2664019763469696, 'train/grad_norm_pre_clip': 16.49259376525879, 'timing/data': 0.012836931273341179, 'timing/model': 0.9362917840480804, 'learning_rate/qwen_vl_interface': 1.7594001246782039e-06, 'learning_rate/action_model': 1.7594001246782037e-05, 'epoch': 2.21}) +[2026-06-18 00:58:41,754][starVLA.training.train_starvla][INFO] - Step 5364, Loss: {'train/loss': 0.3245560824871063, 'train/grad_norm_pre_clip': 23.244653701782227, 'timing/data': 0.009027419611811638, 'timing/model': 0.9681904446333647, 'learning_rate/qwen_vl_interface': 1.7579337163865221e-06, 'learning_rate/action_model': 1.757933716386522e-05, 'epoch': 2.21}) +[2026-06-18 00:58:46,744][starVLA.training.train_starvla][INFO] - Step 5365, Loss: {'train/loss': 0.20414181053638458, 'train/grad_norm_pre_clip': 22.785310745239258, 'timing/data': 0.02282350603491068, 'timing/model': 1.9009736850857735, 'learning_rate/qwen_vl_interface': 1.7564680320039356e-06, 'learning_rate/action_model': 1.7564680320039358e-05, 'epoch': 2.21}) +[2026-06-18 00:58:50,976][starVLA.training.train_starvla][INFO] - Step 5366, Loss: {'train/loss': 0.4523327350616455, 'train/grad_norm_pre_clip': 19.422748565673828, 'timing/data': 0.010095329955220222, 'timing/model': 0.9291970254853368, 'learning_rate/qwen_vl_interface': 1.7550030718342867e-06, 'learning_rate/action_model': 1.7550030718342867e-05, 'epoch': 2.21}) +[2026-06-18 00:58:54,784][starVLA.training.train_starvla][INFO] - Step 5367, Loss: {'train/loss': 0.1542607694864273, 'train/grad_norm_pre_clip': 18.70265007019043, 'timing/data': 0.007855952717363834, 'timing/model': 0.9319935888051987, 'learning_rate/qwen_vl_interface': 1.753538836181262e-06, 'learning_rate/action_model': 1.753538836181262e-05, 'epoch': 2.21}) +[2026-06-18 00:58:58,550][starVLA.training.train_starvla][INFO] - Step 5368, Loss: {'train/loss': 0.4137963652610779, 'train/grad_norm_pre_clip': 40.70286178588867, 'timing/data': 0.008535835891962051, 'timing/model': 1.1026396919041872, 'learning_rate/qwen_vl_interface': 1.7520753253483987e-06, 'learning_rate/action_model': 1.7520753253483986e-05, 'epoch': 2.21}) +[2026-06-18 00:59:02,349][starVLA.training.train_starvla][INFO] - Step 5369, Loss: {'train/loss': 0.17933352291584015, 'train/grad_norm_pre_clip': 25.87956428527832, 'timing/data': 0.010141883976757526, 'timing/model': 1.1431397683918476, 'learning_rate/qwen_vl_interface': 1.7506125396390847e-06, 'learning_rate/action_model': 1.7506125396390847e-05, 'epoch': 2.21}) +[2026-06-18 00:59:05,968][starVLA.training.train_starvla][INFO] - Step 5370, Loss: {'train/loss': 0.15824827551841736, 'train/grad_norm_pre_clip': 30.04611587524414, 'timing/data': 0.010148960165679455, 'timing/model': 0.9334357185289264, 'learning_rate/qwen_vl_interface': 1.7491504793565564e-06, 'learning_rate/action_model': 1.7491504793565562e-05, 'epoch': 2.21}) +[2026-06-18 00:59:09,800][starVLA.training.train_starvla][INFO] - Step 5371, Loss: {'train/loss': 0.4163896143436432, 'train/grad_norm_pre_clip': 21.769145965576172, 'timing/data': 0.010014291852712631, 'timing/model': 0.9415052905678749, 'learning_rate/qwen_vl_interface': 1.7476891448039012e-06, 'learning_rate/action_model': 1.7476891448039013e-05, 'epoch': 2.21}) +[2026-06-18 00:59:13,578][starVLA.training.train_starvla][INFO] - Step 5372, Loss: {'train/loss': 0.2945071756839752, 'train/grad_norm_pre_clip': 22.777114868164062, 'timing/data': 0.010409507900476456, 'timing/model': 0.9361110078170896, 'learning_rate/qwen_vl_interface': 1.7462285362840534e-06, 'learning_rate/action_model': 1.746228536284053e-05, 'epoch': 2.21}) +[2026-06-18 00:59:17,376][starVLA.training.train_starvla][INFO] - Step 5373, Loss: {'train/loss': 0.3175473213195801, 'train/grad_norm_pre_clip': 21.05174446105957, 'timing/data': 0.008139773271977901, 'timing/model': 0.9317317372187972, 'learning_rate/qwen_vl_interface': 1.7447686540998027e-06, 'learning_rate/action_model': 1.7447686540998026e-05, 'epoch': 2.21}) +[2026-06-18 00:59:21,153][starVLA.training.train_starvla][INFO] - Step 5374, Loss: {'train/loss': 0.2946820855140686, 'train/grad_norm_pre_clip': 27.118188858032227, 'timing/data': 0.015886218287050724, 'timing/model': 0.9309361763298512, 'learning_rate/qwen_vl_interface': 1.743309498553781e-06, 'learning_rate/action_model': 1.743309498553781e-05, 'epoch': 2.21}) +[2026-06-18 00:59:24,970][starVLA.training.train_starvla][INFO] - Step 5375, Loss: {'train/loss': 0.4377462863922119, 'train/grad_norm_pre_clip': 73.40506744384766, 'timing/data': 0.008753366768360138, 'timing/model': 1.1090114526450634, 'learning_rate/qwen_vl_interface': 1.7418510699484728e-06, 'learning_rate/action_model': 1.7418510699484727e-05, 'epoch': 2.21}) +[2026-06-18 00:59:28,721][starVLA.training.train_starvla][INFO] - Step 5376, Loss: {'train/loss': 0.3594476878643036, 'train/grad_norm_pre_clip': 24.982837677001953, 'timing/data': 0.009329132735729218, 'timing/model': 1.0843139542266726, 'learning_rate/qwen_vl_interface': 1.7403933685862118e-06, 'learning_rate/action_model': 1.7403933685862118e-05, 'epoch': 2.21}) +[2026-06-18 00:59:32,323][starVLA.training.train_starvla][INFO] - Step 5377, Loss: {'train/loss': 0.4688689112663269, 'train/grad_norm_pre_clip': 21.903335571289062, 'timing/data': 0.011143071576952934, 'timing/model': 0.9304929925128818, 'learning_rate/qwen_vl_interface': 1.7389363947691837e-06, 'learning_rate/action_model': 1.7389363947691838e-05, 'epoch': 2.21}) +[2026-06-18 00:59:36,175][starVLA.training.train_starvla][INFO] - Step 5378, Loss: {'train/loss': 0.12229184061288834, 'train/grad_norm_pre_clip': 13.226962089538574, 'timing/data': 0.0092394994571805, 'timing/model': 0.9279830809682608, 'learning_rate/qwen_vl_interface': 1.7374801487994198e-06, 'learning_rate/action_model': 1.7374801487994198e-05, 'epoch': 2.21}) +[2026-06-18 00:59:40,034][starVLA.training.train_starvla][INFO] - Step 5379, Loss: {'train/loss': 0.6727468967437744, 'train/grad_norm_pre_clip': 27.406164169311523, 'timing/data': 0.009487934410572052, 'timing/model': 0.929079432040453, 'learning_rate/qwen_vl_interface': 1.7360246309787982e-06, 'learning_rate/action_model': 1.736024630978798e-05, 'epoch': 2.21}) +[2026-06-18 00:59:43,766][starVLA.training.train_starvla][INFO] - Step 5380, Loss: {'train/loss': 0.1526748239994049, 'train/grad_norm_pre_clip': 14.75095272064209, 'timing/data': 0.008197398856282234, 'timing/model': 0.9390650177374482, 'learning_rate/qwen_vl_interface': 1.7345698416090534e-06, 'learning_rate/action_model': 1.7345698416090533e-05, 'epoch': 2.21}) +[2026-06-18 00:59:47,515][starVLA.training.train_starvla][INFO] - Step 5381, Loss: {'train/loss': 0.13062383234500885, 'train/grad_norm_pre_clip': 17.902936935424805, 'timing/data': 0.00845000334084034, 'timing/model': 0.9280446348711848, 'learning_rate/qwen_vl_interface': 1.7331157809917632e-06, 'learning_rate/action_model': 1.733115780991763e-05, 'epoch': 2.22}) +[2026-06-18 00:59:51,276][starVLA.training.train_starvla][INFO] - Step 5382, Loss: {'train/loss': 0.18168140947818756, 'train/grad_norm_pre_clip': 19.516756057739258, 'timing/data': 0.010063677094876766, 'timing/model': 1.1047902265563607, 'learning_rate/qwen_vl_interface': 1.7316624494283557e-06, 'learning_rate/action_model': 1.7316624494283557e-05, 'epoch': 2.22}) +[2026-06-18 00:59:54,874][starVLA.training.train_starvla][INFO] - Step 5383, Loss: {'train/loss': 0.40347203612327576, 'train/grad_norm_pre_clip': 19.27679443359375, 'timing/data': 0.009767976589500904, 'timing/model': 0.9257514169439673, 'learning_rate/qwen_vl_interface': 1.7302098472201085e-06, 'learning_rate/action_model': 1.7302098472201084e-05, 'epoch': 2.22}) +[2026-06-18 00:59:58,690][starVLA.training.train_starvla][INFO] - Step 5384, Loss: {'train/loss': 0.18960076570510864, 'train/grad_norm_pre_clip': 15.790565490722656, 'timing/data': 0.010801165364682674, 'timing/model': 0.9466795036569238, 'learning_rate/qwen_vl_interface': 1.728757974668147e-06, 'learning_rate/action_model': 1.728757974668147e-05, 'epoch': 2.22}) +[2026-06-18 01:00:02,675][starVLA.training.train_starvla][INFO] - Step 5385, Loss: {'train/loss': 0.3785860538482666, 'train/grad_norm_pre_clip': 23.481464385986328, 'timing/data': 0.009279431775212288, 'timing/model': 1.0457122111693025, 'learning_rate/qwen_vl_interface': 1.727306832073447e-06, 'learning_rate/action_model': 1.7273068320734472e-05, 'epoch': 2.22}) +[2026-06-18 01:00:06,904][starVLA.training.train_starvla][INFO] - Step 5386, Loss: {'train/loss': 0.39534449577331543, 'train/grad_norm_pre_clip': 30.877737045288086, 'timing/data': 0.00936842616647482, 'timing/model': 0.9715227093547583, 'learning_rate/qwen_vl_interface': 1.7258564197368308e-06, 'learning_rate/action_model': 1.725856419736831e-05, 'epoch': 2.22}) +[2026-06-18 01:00:10,716][starVLA.training.train_starvla][INFO] - Step 5387, Loss: {'train/loss': 0.33629393577575684, 'train/grad_norm_pre_clip': 23.149736404418945, 'timing/data': 0.012558039277791977, 'timing/model': 0.9485056716948748, 'learning_rate/qwen_vl_interface': 1.7244067379589708e-06, 'learning_rate/action_model': 1.7244067379589707e-05, 'epoch': 2.22}) +[2026-06-18 01:00:14,599][starVLA.training.train_starvla][INFO] - Step 5388, Loss: {'train/loss': 0.33925744891166687, 'train/grad_norm_pre_clip': 24.664730072021484, 'timing/data': 0.0072405654937028885, 'timing/model': 1.1531027117744088, 'learning_rate/qwen_vl_interface': 1.7229577870403886e-06, 'learning_rate/action_model': 1.7229577870403886e-05, 'epoch': 2.22}) +[2026-06-18 01:00:18,401][starVLA.training.train_starvla][INFO] - Step 5389, Loss: {'train/loss': 0.37857145071029663, 'train/grad_norm_pre_clip': 21.436683654785156, 'timing/data': 0.010229053907096386, 'timing/model': 1.1134622264653444, 'learning_rate/qwen_vl_interface': 1.7215095672814517e-06, 'learning_rate/action_model': 1.7215095672814516e-05, 'epoch': 2.22}) +[2026-06-18 01:00:22,015][starVLA.training.train_starvla][INFO] - Step 5390, Loss: {'train/loss': 0.22614632546901703, 'train/grad_norm_pre_clip': 19.768455505371094, 'timing/data': 0.009453034028410912, 'timing/model': 0.9312077239155769, 'learning_rate/qwen_vl_interface': 1.7200620789823769e-06, 'learning_rate/action_model': 1.7200620789823767e-05, 'epoch': 2.22}) +[2026-06-18 01:00:25,862][starVLA.training.train_starvla][INFO] - Step 5391, Loss: {'train/loss': 0.15098543465137482, 'train/grad_norm_pre_clip': 15.455679893493652, 'timing/data': 0.00988088920712471, 'timing/model': 0.9279749887064099, 'learning_rate/qwen_vl_interface': 1.7186153224432331e-06, 'learning_rate/action_model': 1.718615322443233e-05, 'epoch': 2.22}) +[2026-06-18 01:00:29,613][starVLA.training.train_starvla][INFO] - Step 5392, Loss: {'train/loss': 0.37483200430870056, 'train/grad_norm_pre_clip': 21.926725387573242, 'timing/data': 0.008734041824936867, 'timing/model': 0.9284808998927474, 'learning_rate/qwen_vl_interface': 1.7171692979639342e-06, 'learning_rate/action_model': 1.717169297963934e-05, 'epoch': 2.22}) +[2026-06-18 01:00:33,363][starVLA.training.train_starvla][INFO] - Step 5393, Loss: {'train/loss': 0.21370111405849457, 'train/grad_norm_pre_clip': 16.608182907104492, 'timing/data': 0.009958282113075256, 'timing/model': 0.929930878803134, 'learning_rate/qwen_vl_interface': 1.7157240058442397e-06, 'learning_rate/action_model': 1.71572400584424e-05, 'epoch': 2.22}) +[2026-06-18 01:00:37,158][starVLA.training.train_starvla][INFO] - Step 5394, Loss: {'train/loss': 0.6803737878799438, 'train/grad_norm_pre_clip': 31.640310287475586, 'timing/data': 0.009084030985832214, 'timing/model': 0.9301955560222268, 'learning_rate/qwen_vl_interface': 1.7142794463837608e-06, 'learning_rate/action_model': 1.714279446383761e-05, 'epoch': 2.22}) +[2026-06-18 01:00:40,947][starVLA.training.train_starvla][INFO] - Step 5395, Loss: {'train/loss': 0.26107579469680786, 'train/grad_norm_pre_clip': 21.939380645751953, 'timing/data': 0.008855740539729595, 'timing/model': 1.1305855140089989, 'learning_rate/qwen_vl_interface': 1.712835619881959e-06, 'learning_rate/action_model': 1.7128356198819588e-05, 'epoch': 2.22}) +[2026-06-18 01:00:44,711][starVLA.training.train_starvla][INFO] - Step 5396, Loss: {'train/loss': 0.20617936551570892, 'train/grad_norm_pre_clip': 20.67285919189453, 'timing/data': 0.008213218301534653, 'timing/model': 1.1087236599996686, 'learning_rate/qwen_vl_interface': 1.7113925266381383e-06, 'learning_rate/action_model': 1.7113925266381383e-05, 'epoch': 2.22}) +[2026-06-18 01:00:48,333][starVLA.training.train_starvla][INFO] - Step 5397, Loss: {'train/loss': 0.3160523772239685, 'train/grad_norm_pre_clip': 17.977602005004883, 'timing/data': 0.010985393077135086, 'timing/model': 0.933402362279594, 'learning_rate/qwen_vl_interface': 1.7099501669514569e-06, 'learning_rate/action_model': 1.7099501669514568e-05, 'epoch': 2.22}) +[2026-06-18 01:00:52,139][starVLA.training.train_starvla][INFO] - Step 5398, Loss: {'train/loss': 0.2886649966239929, 'train/grad_norm_pre_clip': 24.854965209960938, 'timing/data': 0.009268945083022118, 'timing/model': 0.9298003390431404, 'learning_rate/qwen_vl_interface': 1.7085085411209118e-06, 'learning_rate/action_model': 1.7085085411209118e-05, 'epoch': 2.22}) +[2026-06-18 01:00:55,955][starVLA.training.train_starvla][INFO] - Step 5399, Loss: {'train/loss': 0.21325519680976868, 'train/grad_norm_pre_clip': 26.716609954833984, 'timing/data': 0.010178714990615845, 'timing/model': 0.9345988510176539, 'learning_rate/qwen_vl_interface': 1.7070676494453587e-06, 'learning_rate/action_model': 1.7070676494453586e-05, 'epoch': 2.22}) +[2026-06-18 01:02:48,228][starVLA.training.train_starvla][INFO] - Step 5400, Loss: {'train/loss': 0.13792717456817627, 'train/grad_norm_pre_clip': 17.7698974609375, 'eval/loss': 0.411810964345932, 'eval/latency_6/loss': 0.411810964345932, 'eval/demon_fire_precision': 0.9556821779044001, 'eval/demon_fire_recall': 0.975759534583064, 'eval/demon_fire_f1': 0.9656165040780424, 'eval/demon_fire_rate_model': 0.8789649415692822, 'eval/demon_fire_rate_teacher': 0.860879243183083, 'eval/demon_left_precision': 0.8705179282868526, 'eval/demon_left_recall': 0.7666666666666667, 'eval/demon_left_f1': 0.8152985074626866, 'eval/demon_left_rate_model': 0.13967723984418476, 'eval/demon_left_rate_teacher': 0.15859766277128548, 'eval/demon_right_precision': 0.8111111111111111, 'eval/demon_right_recall': 0.7849462365591398, 'eval/demon_right_f1': 0.7978142076502732, 'eval/demon_right_rate_model': 0.17529215358931552, 'eval/demon_right_rate_teacher': 0.18113522537562604, 'eval/demon_move_f1': 0.8065563575564799, 'eval/demon_cc_f1': 0.9116791149849323, 'eval/latency_6/demon_fire_precision': 0.9556821779044001, 'eval/latency_6/demon_fire_recall': 0.975759534583064, 'eval/latency_6/demon_fire_f1': 0.9656165040780424, 'eval/latency_6/demon_fire_rate_model': 0.8789649415692822, 'eval/latency_6/demon_fire_rate_teacher': 0.860879243183083, 'eval/latency_6/demon_left_precision': 0.8705179282868526, 'eval/latency_6/demon_left_recall': 0.7666666666666667, 'eval/latency_6/demon_left_f1': 0.8152985074626866, 'eval/latency_6/demon_left_rate_model': 0.13967723984418476, 'eval/latency_6/demon_left_rate_teacher': 0.15859766277128548, 'eval/latency_6/demon_right_precision': 0.8111111111111111, 'eval/latency_6/demon_right_recall': 0.7849462365591398, 'eval/latency_6/demon_right_f1': 0.7978142076502732, 'eval/latency_6/demon_right_rate_model': 0.17529215358931552, 'eval/latency_6/demon_right_rate_teacher': 0.18113522537562604, 'eval/latency_6/demon_move_f1': 0.8065563575564799, 'eval/latency_6/demon_cc_f1': 0.9116791149849323, 'timing/data': 0.014001147821545601, 'timing/model': 0.925774117000401, 'learning_rate/qwen_vl_interface': 1.705627492223494e-06, 'learning_rate/action_model': 1.705627492223494e-05, 'epoch': 2.22}) +[2026-06-18 01:02:48,240][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_5400_state +[2026-06-18 01:02:48,245][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 01:02:55,524][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_5400_state/model.safetensors +[2026-06-18 01:03:12,174][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_5400_state/optimizer.bin +[2026-06-18 01:03:12,175][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_5400_state/sampler.bin +[2026-06-18 01:03:12,175][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_5400_state/sampler_1.bin +[2026-06-18 01:03:12,177][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_5400_state/random_states_0.pkl +[2026-06-18 01:03:12,181][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 5400 +[2026-06-18 01:03:12,181][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 01:03:12,193][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 01:03:16,234][starVLA.training.train_starvla][INFO] - Step 5401, Loss: {'train/loss': 0.38371655344963074, 'train/grad_norm_pre_clip': 20.065567016601562, 'timing/data': 0.00945009384304285, 'timing/model': 0.9275897648185492, 'learning_rate/qwen_vl_interface': 1.7041880697538644e-06, 'learning_rate/action_model': 1.704188069753864e-05, 'epoch': 2.22}) +[2026-06-18 01:03:20,014][starVLA.training.train_starvla][INFO] - Step 5402, Loss: {'train/loss': 0.10125037282705307, 'train/grad_norm_pre_clip': 9.999334335327148, 'timing/data': 0.017570078372955322, 'timing/model': 0.9280337421223521, 'learning_rate/qwen_vl_interface': 1.702749382334863e-06, 'learning_rate/action_model': 1.7027493823348632e-05, 'epoch': 2.22}) +[2026-06-18 01:03:23,763][starVLA.training.train_starvla][INFO] - Step 5403, Loss: {'train/loss': 0.4918188154697418, 'train/grad_norm_pre_clip': 53.95643615722656, 'timing/data': 0.00888229999691248, 'timing/model': 0.9273877367377281, 'learning_rate/qwen_vl_interface': 1.7013114302647318e-06, 'learning_rate/action_model': 1.7013114302647318e-05, 'epoch': 2.22}) +[2026-06-18 01:03:27,549][starVLA.training.train_starvla][INFO] - Step 5404, Loss: {'train/loss': 0.20304453372955322, 'train/grad_norm_pre_clip': 21.493083953857422, 'timing/data': 0.01344335824251175, 'timing/model': 1.1149949058890343, 'learning_rate/qwen_vl_interface': 1.69987421384156e-06, 'learning_rate/action_model': 1.69987421384156e-05, 'epoch': 2.22}) +[2026-06-18 01:03:31,331][starVLA.training.train_starvla][INFO] - Step 5405, Loss: {'train/loss': 0.2240181863307953, 'train/grad_norm_pre_clip': 35.592559814453125, 'timing/data': 0.016621382907032967, 'timing/model': 1.1034519830718637, 'learning_rate/qwen_vl_interface': 1.6984377333632815e-06, 'learning_rate/action_model': 1.6984377333632813e-05, 'epoch': 2.22}) +[2026-06-18 01:03:34,958][starVLA.training.train_starvla][INFO] - Step 5406, Loss: {'train/loss': 0.35316064953804016, 'train/grad_norm_pre_clip': 24.154399871826172, 'timing/data': 0.011424030177295208, 'timing/model': 0.9377306215465069, 'learning_rate/qwen_vl_interface': 1.697001989127686e-06, 'learning_rate/action_model': 1.697001989127686e-05, 'epoch': 2.23}) +[2026-06-18 01:03:39,289][starVLA.training.train_starvla][INFO] - Step 5407, Loss: {'train/loss': 0.23658299446105957, 'train/grad_norm_pre_clip': 22.38395118713379, 'timing/data': 0.008334971033036709, 'timing/model': 1.4894728362560272, 'learning_rate/qwen_vl_interface': 1.6955669814324003e-06, 'learning_rate/action_model': 1.6955669814324e-05, 'epoch': 2.23}) +[2026-06-18 01:03:43,110][starVLA.training.train_starvla][INFO] - Step 5408, Loss: {'train/loss': 0.18468555808067322, 'train/grad_norm_pre_clip': 10.239696502685547, 'timing/data': 0.008367269299924374, 'timing/model': 0.9317783853039145, 'learning_rate/qwen_vl_interface': 1.6941327105749034e-06, 'learning_rate/action_model': 1.694132710574903e-05, 'epoch': 2.23}) +[2026-06-18 01:03:46,928][starVLA.training.train_starvla][INFO] - Step 5409, Loss: {'train/loss': 0.4127647280693054, 'train/grad_norm_pre_clip': 22.356456756591797, 'timing/data': 0.010108808055520058, 'timing/model': 0.9461608557030559, 'learning_rate/qwen_vl_interface': 1.6926991768525205e-06, 'learning_rate/action_model': 1.6926991768525205e-05, 'epoch': 2.23}) +[2026-06-18 01:03:50,880][starVLA.training.train_starvla][INFO] - Step 5410, Loss: {'train/loss': 0.15083637833595276, 'train/grad_norm_pre_clip': 47.604248046875, 'timing/data': 0.008982155472040176, 'timing/model': 0.9452159889042377, 'learning_rate/qwen_vl_interface': 1.6912663805624278e-06, 'learning_rate/action_model': 1.6912663805624275e-05, 'epoch': 2.23}) +[2026-06-18 01:03:54,691][starVLA.training.train_starvla][INFO] - Step 5411, Loss: {'train/loss': 0.44916343688964844, 'train/grad_norm_pre_clip': 36.52544021606445, 'timing/data': 0.00990249216556549, 'timing/model': 1.1408167714253068, 'learning_rate/qwen_vl_interface': 1.6898343220016446e-06, 'learning_rate/action_model': 1.6898343220016447e-05, 'epoch': 2.23}) +[2026-06-18 01:03:58,469][starVLA.training.train_starvla][INFO] - Step 5412, Loss: {'train/loss': 0.6954081058502197, 'train/grad_norm_pre_clip': 26.15437889099121, 'timing/data': 0.011066332459449768, 'timing/model': 1.108152637258172, 'learning_rate/qwen_vl_interface': 1.6884030014670353e-06, 'learning_rate/action_model': 1.688403001467035e-05, 'epoch': 2.23}) +[2026-06-18 01:04:02,073][starVLA.training.train_starvla][INFO] - Step 5413, Loss: {'train/loss': 0.19896911084651947, 'train/grad_norm_pre_clip': 16.04610252380371, 'timing/data': 0.0101458840072155, 'timing/model': 0.9297281457111239, 'learning_rate/qwen_vl_interface': 1.6869724192553168e-06, 'learning_rate/action_model': 1.686972419255317e-05, 'epoch': 2.23}) +[2026-06-18 01:04:05,905][starVLA.training.train_starvla][INFO] - Step 5414, Loss: {'train/loss': 0.1028163731098175, 'train/grad_norm_pre_clip': 8.892380714416504, 'timing/data': 0.010350596159696579, 'timing/model': 0.9377127466723323, 'learning_rate/qwen_vl_interface': 1.6855425756630506e-06, 'learning_rate/action_model': 1.6855425756630506e-05, 'epoch': 2.23}) +[2026-06-18 01:04:09,779][starVLA.training.train_starvla][INFO] - Step 5415, Loss: {'train/loss': 0.06261307746171951, 'train/grad_norm_pre_clip': 8.395060539245605, 'timing/data': 0.009857569821178913, 'timing/model': 0.9341703513637185, 'learning_rate/qwen_vl_interface': 1.684113470986644e-06, 'learning_rate/action_model': 1.6841134709866438e-05, 'epoch': 2.23}) +[2026-06-18 01:04:13,571][starVLA.training.train_starvla][INFO] - Step 5416, Loss: {'train/loss': 0.3599529266357422, 'train/grad_norm_pre_clip': 18.422456741333008, 'timing/data': 0.008980994112789631, 'timing/model': 0.9397860486060381, 'learning_rate/qwen_vl_interface': 1.682685105522351e-06, 'learning_rate/action_model': 1.682685105522351e-05, 'epoch': 2.23}) +[2026-06-18 01:04:17,377][starVLA.training.train_starvla][INFO] - Step 5417, Loss: {'train/loss': 0.11458130925893784, 'train/grad_norm_pre_clip': 13.002789497375488, 'timing/data': 0.009997718967497349, 'timing/model': 0.9330285964533687, 'learning_rate/qwen_vl_interface': 1.6812574795662751e-06, 'learning_rate/action_model': 1.681257479566275e-05, 'epoch': 2.23}) +[2026-06-18 01:04:21,801][starVLA.training.train_starvla][INFO] - Step 5418, Loss: {'train/loss': 0.1337076872587204, 'train/grad_norm_pre_clip': 11.165437698364258, 'timing/data': 0.008035844191908836, 'timing/model': 1.2169558107852936, 'learning_rate/qwen_vl_interface': 1.6798305934143635e-06, 'learning_rate/action_model': 1.6798305934143633e-05, 'epoch': 2.23}) +[2026-06-18 01:04:25,422][starVLA.training.train_starvla][INFO] - Step 5419, Loss: {'train/loss': 0.393257200717926, 'train/grad_norm_pre_clip': 26.8278751373291, 'timing/data': 0.009993341751396656, 'timing/model': 0.9309008792042732, 'learning_rate/qwen_vl_interface': 1.6784044473624118e-06, 'learning_rate/action_model': 1.6784044473624118e-05, 'epoch': 2.23}) +[2026-06-18 01:04:29,298][starVLA.training.train_starvla][INFO] - Step 5420, Loss: {'train/loss': 0.28422898054122925, 'train/grad_norm_pre_clip': 18.23834228515625, 'timing/data': 0.014553293585777283, 'timing/model': 0.9460187833756208, 'learning_rate/qwen_vl_interface': 1.6769790417060611e-06, 'learning_rate/action_model': 1.676979041706061e-05, 'epoch': 2.23}) +[2026-06-18 01:04:33,068][starVLA.training.train_starvla][INFO] - Step 5421, Loss: {'train/loss': 0.2358112633228302, 'train/grad_norm_pre_clip': 20.913867950439453, 'timing/data': 0.009381350129842758, 'timing/model': 0.9595355167984962, 'learning_rate/qwen_vl_interface': 1.6755543767407998e-06, 'learning_rate/action_model': 1.6755543767407998e-05, 'epoch': 2.23}) +[2026-06-18 01:04:36,912][starVLA.training.train_starvla][INFO] - Step 5422, Loss: {'train/loss': 0.15442806482315063, 'train/grad_norm_pre_clip': 13.185089111328125, 'timing/data': 0.012368199415504932, 'timing/model': 0.9407941307872534, 'learning_rate/qwen_vl_interface': 1.6741304527619624e-06, 'learning_rate/action_model': 1.674130452761962e-05, 'epoch': 2.23}) +[2026-06-18 01:04:40,724][starVLA.training.train_starvla][INFO] - Step 5423, Loss: {'train/loss': 0.22421398758888245, 'train/grad_norm_pre_clip': 17.288562774658203, 'timing/data': 0.018019720911979675, 'timing/model': 0.9354174435138702, 'learning_rate/qwen_vl_interface': 1.672707270064729e-06, 'learning_rate/action_model': 1.6727072700647288e-05, 'epoch': 2.23}) +[2026-06-18 01:04:44,545][starVLA.training.train_starvla][INFO] - Step 5424, Loss: {'train/loss': 0.11660399287939072, 'train/grad_norm_pre_clip': 11.7708101272583, 'timing/data': 0.008758767507970333, 'timing/model': 1.1363347275182605, 'learning_rate/qwen_vl_interface': 1.6712848289441288e-06, 'learning_rate/action_model': 1.6712848289441287e-05, 'epoch': 2.23}) +[2026-06-18 01:04:48,432][starVLA.training.train_starvla][INFO] - Step 5425, Loss: {'train/loss': 0.40130916237831116, 'train/grad_norm_pre_clip': 23.821901321411133, 'timing/data': 0.00823653768748045, 'timing/model': 1.1327240206301212, 'learning_rate/qwen_vl_interface': 1.669863129695036e-06, 'learning_rate/action_model': 1.669863129695036e-05, 'epoch': 2.23}) +[2026-06-18 01:04:52,036][starVLA.training.train_starvla][INFO] - Step 5426, Loss: {'train/loss': 0.11316490173339844, 'train/grad_norm_pre_clip': 11.128836631774902, 'timing/data': 0.008108830079436302, 'timing/model': 0.928617101162672, 'learning_rate/qwen_vl_interface': 1.6684421726121675e-06, 'learning_rate/action_model': 1.6684421726121674e-05, 'epoch': 2.23}) +[2026-06-18 01:04:55,844][starVLA.training.train_starvla][INFO] - Step 5427, Loss: {'train/loss': 0.29527777433395386, 'train/grad_norm_pre_clip': 25.540355682373047, 'timing/data': 0.007014656439423561, 'timing/model': 0.9306498300284147, 'learning_rate/qwen_vl_interface': 1.6670219579900888e-06, 'learning_rate/action_model': 1.6670219579900887e-05, 'epoch': 2.23}) +[2026-06-18 01:04:59,622][starVLA.training.train_starvla][INFO] - Step 5428, Loss: {'train/loss': 0.4115476608276367, 'train/grad_norm_pre_clip': 28.272533416748047, 'timing/data': 0.009737699292600155, 'timing/model': 0.926106221973896, 'learning_rate/qwen_vl_interface': 1.665602486123215e-06, 'learning_rate/action_model': 1.665602486123215e-05, 'epoch': 2.23}) +[2026-06-18 01:05:03,389][starVLA.training.train_starvla][INFO] - Step 5429, Loss: {'train/loss': 0.30415377020835876, 'train/grad_norm_pre_clip': 25.297115325927734, 'timing/data': 0.009241550229489803, 'timing/model': 0.9289237158372998, 'learning_rate/qwen_vl_interface': 1.6641837573058034e-06, 'learning_rate/action_model': 1.6641837573058035e-05, 'epoch': 2.23}) +[2026-06-18 01:05:07,200][starVLA.training.train_starvla][INFO] - Step 5430, Loss: {'train/loss': 0.2246265709400177, 'train/grad_norm_pre_clip': 22.513629913330078, 'timing/data': 0.008458886295557022, 'timing/model': 0.9324073726311326, 'learning_rate/qwen_vl_interface': 1.6627657718319568e-06, 'learning_rate/action_model': 1.6627657718319568e-05, 'epoch': 2.24}) +[2026-06-18 01:05:10,997][starVLA.training.train_starvla][INFO] - Step 5431, Loss: {'train/loss': 0.17713171243667603, 'train/grad_norm_pre_clip': 16.60449981689453, 'timing/data': 0.010677304118871689, 'timing/model': 1.1298380931839347, 'learning_rate/qwen_vl_interface': 1.6613485299956265e-06, 'learning_rate/action_model': 1.6613485299956262e-05, 'epoch': 2.24}) +[2026-06-18 01:05:14,834][starVLA.training.train_starvla][INFO] - Step 5432, Loss: {'train/loss': 0.12769293785095215, 'train/grad_norm_pre_clip': 13.393864631652832, 'timing/data': 0.014718121849000454, 'timing/model': 1.1125440634787083, 'learning_rate/qwen_vl_interface': 1.6599320320906064e-06, 'learning_rate/action_model': 1.6599320320906066e-05, 'epoch': 2.24}) +[2026-06-18 01:05:18,450][starVLA.training.train_starvla][INFO] - Step 5433, Loss: {'train/loss': 0.30357325077056885, 'train/grad_norm_pre_clip': 61.569007873535156, 'timing/data': 0.020130647346377373, 'timing/model': 0.9340527672320604, 'learning_rate/qwen_vl_interface': 1.65851627841054e-06, 'learning_rate/action_model': 1.65851627841054e-05, 'epoch': 2.24}) +[2026-06-18 01:05:22,240][starVLA.training.train_starvla][INFO] - Step 5434, Loss: {'train/loss': 0.06956003606319427, 'train/grad_norm_pre_clip': 7.480128288269043, 'timing/data': 0.013605349697172642, 'timing/model': 0.9306803988292813, 'learning_rate/qwen_vl_interface': 1.657101269248913e-06, 'learning_rate/action_model': 1.657101269248913e-05, 'epoch': 2.24}) +[2026-06-18 01:05:26,073][starVLA.training.train_starvla][INFO] - Step 5435, Loss: {'train/loss': 0.14176368713378906, 'train/grad_norm_pre_clip': 18.57962989807129, 'timing/data': 0.01416577585041523, 'timing/model': 0.9358803648501635, 'learning_rate/qwen_vl_interface': 1.65568700489906e-06, 'learning_rate/action_model': 1.65568700489906e-05, 'epoch': 2.24}) +[2026-06-18 01:05:29,812][starVLA.training.train_starvla][INFO] - Step 5436, Loss: {'train/loss': 0.17799077928066254, 'train/grad_norm_pre_clip': 18.381181716918945, 'timing/data': 0.012210809625685215, 'timing/model': 0.9300302006304264, 'learning_rate/qwen_vl_interface': 1.6542734856541577e-06, 'learning_rate/action_model': 1.6542734856541576e-05, 'epoch': 2.24}) +[2026-06-18 01:05:33,586][starVLA.training.train_starvla][INFO] - Step 5437, Loss: {'train/loss': 0.10205744206905365, 'train/grad_norm_pre_clip': 9.602765083312988, 'timing/data': 0.013786425814032555, 'timing/model': 0.9298666305840015, 'learning_rate/qwen_vl_interface': 1.652860711807231e-06, 'learning_rate/action_model': 1.652860711807231e-05, 'epoch': 2.24}) +[2026-06-18 01:05:37,399][starVLA.training.train_starvla][INFO] - Step 5438, Loss: {'train/loss': 0.2801167368888855, 'train/grad_norm_pre_clip': 18.62275505065918, 'timing/data': 0.0091458884999156, 'timing/model': 1.1400173818692565, 'learning_rate/qwen_vl_interface': 1.6514486836511484e-06, 'learning_rate/action_model': 1.6514486836511484e-05, 'epoch': 2.24}) +[2026-06-18 01:05:40,993][starVLA.training.train_starvla][INFO] - Step 5439, Loss: {'train/loss': 0.25921618938446045, 'train/grad_norm_pre_clip': 24.329303741455078, 'timing/data': 0.01001875288784504, 'timing/model': 0.9264728650450706, 'learning_rate/qwen_vl_interface': 1.6500374014786288e-06, 'learning_rate/action_model': 1.650037401478629e-05, 'epoch': 2.24}) +[2026-06-18 01:05:44,902][starVLA.training.train_starvla][INFO] - Step 5440, Loss: {'train/loss': 0.09338446706533432, 'train/grad_norm_pre_clip': 10.127209663391113, 'timing/data': 0.011980576440691948, 'timing/model': 0.9434949494898319, 'learning_rate/qwen_vl_interface': 1.6486268655822277e-06, 'learning_rate/action_model': 1.6486268655822277e-05, 'epoch': 2.24}) +[2026-06-18 01:05:48,767][starVLA.training.train_starvla][INFO] - Step 5441, Loss: {'train/loss': 0.27527669072151184, 'train/grad_norm_pre_clip': 28.991052627563477, 'timing/data': 0.01130915991961956, 'timing/model': 0.9356111399829388, 'learning_rate/qwen_vl_interface': 1.6472170762543515e-06, 'learning_rate/action_model': 1.6472170762543514e-05, 'epoch': 2.24}) +[2026-06-18 01:05:52,646][starVLA.training.train_starvla][INFO] - Step 5442, Loss: {'train/loss': 0.24357809126377106, 'train/grad_norm_pre_clip': 16.97589874267578, 'timing/data': 0.0127017293125391, 'timing/model': 0.9560750899836421, 'learning_rate/qwen_vl_interface': 1.6458080337872534e-06, 'learning_rate/action_model': 1.6458080337872535e-05, 'epoch': 2.24}) +[2026-06-18 01:05:56,547][starVLA.training.train_starvla][INFO] - Step 5443, Loss: {'train/loss': 0.4282224476337433, 'train/grad_norm_pre_clip': 28.850196838378906, 'timing/data': 0.011032056994736195, 'timing/model': 0.9727764101698995, 'learning_rate/qwen_vl_interface': 1.644399738473029e-06, 'learning_rate/action_model': 1.644399738473029e-05, 'epoch': 2.24}) +[2026-06-18 01:06:00,347][starVLA.training.train_starvla][INFO] - Step 5444, Loss: {'train/loss': 0.24335446953773499, 'train/grad_norm_pre_clip': 18.83257293701172, 'timing/data': 0.010524381883442402, 'timing/model': 0.9349550735205412, 'learning_rate/qwen_vl_interface': 1.6429921906036175e-06, 'learning_rate/action_model': 1.6429921906036176e-05, 'epoch': 2.24}) +[2026-06-18 01:06:04,239][starVLA.training.train_starvla][INFO] - Step 5445, Loss: {'train/loss': 0.3818713426589966, 'train/grad_norm_pre_clip': 23.657047271728516, 'timing/data': 0.012587891891598701, 'timing/model': 1.1821639025583863, 'learning_rate/qwen_vl_interface': 1.6415853904708068e-06, 'learning_rate/action_model': 1.6415853904708066e-05, 'epoch': 2.24}) +[2026-06-18 01:06:08,032][starVLA.training.train_starvla][INFO] - Step 5446, Loss: {'train/loss': 0.1575254201889038, 'train/grad_norm_pre_clip': 14.543416976928711, 'timing/data': 0.010684949345886707, 'timing/model': 1.052968961186707, 'learning_rate/qwen_vl_interface': 1.6401793383662274e-06, 'learning_rate/action_model': 1.6401793383662273e-05, 'epoch': 2.24}) +[2026-06-18 01:06:12,004][starVLA.training.train_starvla][INFO] - Step 5447, Loss: {'train/loss': 0.3598018288612366, 'train/grad_norm_pre_clip': 25.051013946533203, 'timing/data': 0.010705282911658287, 'timing/model': 0.9610171103850007, 'learning_rate/qwen_vl_interface': 1.6387740345813553e-06, 'learning_rate/action_model': 1.6387740345813552e-05, 'epoch': 2.24}) +[2026-06-18 01:06:17,526][starVLA.training.train_starvla][INFO] - Step 5448, Loss: {'train/loss': 0.45723825693130493, 'train/grad_norm_pre_clip': 27.679983139038086, 'timing/data': 0.008894280530512333, 'timing/model': 0.9686435451731086, 'learning_rate/qwen_vl_interface': 1.6373694794075115e-06, 'learning_rate/action_model': 1.6373694794075115e-05, 'epoch': 2.24}) +[2026-06-18 01:06:22,727][starVLA.training.train_starvla][INFO] - Step 5449, Loss: {'train/loss': 0.08377889543771744, 'train/grad_norm_pre_clip': 7.9575910568237305, 'timing/data': 0.015204554423689842, 'timing/model': 1.050404790788889, 'learning_rate/qwen_vl_interface': 1.6359656731358626e-06, 'learning_rate/action_model': 1.6359656731358626e-05, 'epoch': 2.24}) +[2026-06-18 01:06:26,832][starVLA.training.train_starvla][INFO] - Step 5450, Loss: {'train/loss': 0.11101658642292023, 'train/grad_norm_pre_clip': 18.39976692199707, 'timing/data': 0.018721663393080235, 'timing/model': 0.9413387132808566, 'learning_rate/qwen_vl_interface': 1.634562616057418e-06, 'learning_rate/action_model': 1.634562616057418e-05, 'epoch': 2.24}) +[2026-06-18 01:06:31,100][starVLA.training.train_starvla][INFO] - Step 5451, Loss: {'train/loss': 0.5757570862770081, 'train/grad_norm_pre_clip': 37.08452224731445, 'timing/data': 0.013283248990774155, 'timing/model': 1.5175019102171063, 'learning_rate/qwen_vl_interface': 1.6331603084630338e-06, 'learning_rate/action_model': 1.6331603084630338e-05, 'epoch': 2.24}) +[2026-06-18 01:06:34,962][starVLA.training.train_starvla][INFO] - Step 5452, Loss: {'train/loss': 0.5760576725006104, 'train/grad_norm_pre_clip': 29.15654754638672, 'timing/data': 0.013019544072449207, 'timing/model': 1.1851151743903756, 'learning_rate/qwen_vl_interface': 1.6317587506434076e-06, 'learning_rate/action_model': 1.6317587506434076e-05, 'epoch': 2.24}) +[2026-06-18 01:06:38,573][starVLA.training.train_starvla][INFO] - Step 5453, Loss: {'train/loss': 0.24286553263664246, 'train/grad_norm_pre_clip': 17.55689239501953, 'timing/data': 0.009915690869092941, 'timing/model': 0.931989997625351, 'learning_rate/qwen_vl_interface': 1.6303579428890881e-06, 'learning_rate/action_model': 1.630357942889088e-05, 'epoch': 2.24}) +[2026-06-18 01:06:42,490][starVLA.training.train_starvla][INFO] - Step 5454, Loss: {'train/loss': 0.19105124473571777, 'train/grad_norm_pre_clip': 18.027212142944336, 'timing/data': 0.009255082346498966, 'timing/model': 0.9417119044810534, 'learning_rate/qwen_vl_interface': 1.6289578854904626e-06, 'learning_rate/action_model': 1.6289578854904625e-05, 'epoch': 2.25}) +[2026-06-18 01:06:46,335][starVLA.training.train_starvla][INFO] - Step 5455, Loss: {'train/loss': 0.2241438925266266, 'train/grad_norm_pre_clip': 16.200157165527344, 'timing/data': 0.01163399126380682, 'timing/model': 0.9420028449967504, 'learning_rate/qwen_vl_interface': 1.6275585787377627e-06, 'learning_rate/action_model': 1.6275585787377626e-05, 'epoch': 2.25}) +[2026-06-18 01:06:50,148][starVLA.training.train_starvla][INFO] - Step 5456, Loss: {'train/loss': 0.16644299030303955, 'train/grad_norm_pre_clip': 16.36444091796875, 'timing/data': 0.008504858240485191, 'timing/model': 0.9403901770710945, 'learning_rate/qwen_vl_interface': 1.6261600229210662e-06, 'learning_rate/action_model': 1.626160022921066e-05, 'epoch': 2.25}) +[2026-06-18 01:06:54,006][starVLA.training.train_starvla][INFO] - Step 5457, Loss: {'train/loss': 0.17756733298301697, 'train/grad_norm_pre_clip': 18.96253204345703, 'timing/data': 0.012319623492658138, 'timing/model': 0.9388944003731012, 'learning_rate/qwen_vl_interface': 1.6247622183302972e-06, 'learning_rate/action_model': 1.624762218330297e-05, 'epoch': 2.25}) +[2026-06-18 01:06:57,890][starVLA.training.train_starvla][INFO] - Step 5458, Loss: {'train/loss': 0.406585156917572, 'train/grad_norm_pre_clip': 22.863527297973633, 'timing/data': 0.008537317626178265, 'timing/model': 1.2001121947541833, 'learning_rate/qwen_vl_interface': 1.6233651652552218e-06, 'learning_rate/action_model': 1.6233651652552217e-05, 'epoch': 2.25}) +[2026-06-18 01:07:01,674][starVLA.training.train_starvla][INFO] - Step 5459, Loss: {'train/loss': 0.2902590036392212, 'train/grad_norm_pre_clip': 28.6180419921875, 'timing/data': 0.011499534361064434, 'timing/model': 1.0961409006267786, 'learning_rate/qwen_vl_interface': 1.6219688639854507e-06, 'learning_rate/action_model': 1.6219688639854506e-05, 'epoch': 2.25}) +[2026-06-18 01:07:05,292][starVLA.training.train_starvla][INFO] - Step 5460, Loss: {'train/loss': 0.11105791479349136, 'train/grad_norm_pre_clip': 13.990339279174805, 'timing/data': 0.00974862277507782, 'timing/model': 0.9266134509816766, 'learning_rate/qwen_vl_interface': 1.620573314810436e-06, 'learning_rate/action_model': 1.620573314810436e-05, 'epoch': 2.25}) +[2026-06-18 01:07:09,104][starVLA.training.train_starvla][INFO] - Step 5461, Loss: {'train/loss': 0.207938551902771, 'train/grad_norm_pre_clip': 19.496488571166992, 'timing/data': 0.010228424333035946, 'timing/model': 0.9302919264882803, 'learning_rate/qwen_vl_interface': 1.6191785180194797e-06, 'learning_rate/action_model': 1.6191785180194796e-05, 'epoch': 2.25}) +[2026-06-18 01:07:12,900][starVLA.training.train_starvla][INFO] - Step 5462, Loss: {'train/loss': 0.40495994687080383, 'train/grad_norm_pre_clip': 22.30583381652832, 'timing/data': 0.010604138486087322, 'timing/model': 0.9329721285030246, 'learning_rate/qwen_vl_interface': 1.6177844739017236e-06, 'learning_rate/action_model': 1.6177844739017236e-05, 'epoch': 2.25}) +[2026-06-18 01:07:16,712][starVLA.training.train_starvla][INFO] - Step 5463, Loss: {'train/loss': 0.36058855056762695, 'train/grad_norm_pre_clip': 22.000675201416016, 'timing/data': 0.009753049351274967, 'timing/model': 0.975350234657526, 'learning_rate/qwen_vl_interface': 1.6163911827461546e-06, 'learning_rate/action_model': 1.6163911827461546e-05, 'epoch': 2.25}) +[2026-06-18 01:07:20,638][starVLA.training.train_starvla][INFO] - Step 5464, Loss: {'train/loss': 0.16253720223903656, 'train/grad_norm_pre_clip': 15.20980167388916, 'timing/data': 0.008819468319416046, 'timing/model': 0.943608264438808, 'learning_rate/qwen_vl_interface': 1.6149986448416032e-06, 'learning_rate/action_model': 1.614998644841603e-05, 'epoch': 2.25}) +[2026-06-18 01:07:24,525][starVLA.training.train_starvla][INFO] - Step 5465, Loss: {'train/loss': 0.25343143939971924, 'train/grad_norm_pre_clip': 95.00440979003906, 'timing/data': 0.013904882594943047, 'timing/model': 1.1943701542913914, 'learning_rate/qwen_vl_interface': 1.6136068604767444e-06, 'learning_rate/action_model': 1.6136068604767442e-05, 'epoch': 2.25}) +[2026-06-18 01:07:28,158][starVLA.training.train_starvla][INFO] - Step 5466, Loss: {'train/loss': 0.3382277190685272, 'train/grad_norm_pre_clip': 21.420534133911133, 'timing/data': 0.010936920531094074, 'timing/model': 0.9289821051061153, 'learning_rate/qwen_vl_interface': 1.6122158299400962e-06, 'learning_rate/action_model': 1.612215829940096e-05, 'epoch': 2.25}) +[2026-06-18 01:07:32,176][starVLA.training.train_starvla][INFO] - Step 5467, Loss: {'train/loss': 0.3824384808540344, 'train/grad_norm_pre_clip': 24.62209701538086, 'timing/data': 0.010559281334280968, 'timing/model': 0.9332425892353058, 'learning_rate/qwen_vl_interface': 1.6108255535200193e-06, 'learning_rate/action_model': 1.6108255535200193e-05, 'epoch': 2.25}) +[2026-06-18 01:07:35,936][starVLA.training.train_starvla][INFO] - Step 5468, Loss: {'train/loss': 0.25765448808670044, 'train/grad_norm_pre_clip': 23.465927124023438, 'timing/data': 0.014422914944589138, 'timing/model': 0.9349810183048248, 'learning_rate/qwen_vl_interface': 1.6094360315047236e-06, 'learning_rate/action_model': 1.6094360315047237e-05, 'epoch': 2.25}) +[2026-06-18 01:07:39,777][starVLA.training.train_starvla][INFO] - Step 5469, Loss: {'train/loss': 0.34717485308647156, 'train/grad_norm_pre_clip': 15.392230033874512, 'timing/data': 0.0071179065853357315, 'timing/model': 0.9379553170874715, 'learning_rate/qwen_vl_interface': 1.608047264182255e-06, 'learning_rate/action_model': 1.608047264182255e-05, 'epoch': 2.25}) +[2026-06-18 01:07:43,644][starVLA.training.train_starvla][INFO] - Step 5470, Loss: {'train/loss': 0.21771404147148132, 'train/grad_norm_pre_clip': 19.185104370117188, 'timing/data': 0.008096134290099144, 'timing/model': 0.9452610611915588, 'learning_rate/qwen_vl_interface': 1.6066592518405077e-06, 'learning_rate/action_model': 1.6066592518405076e-05, 'epoch': 2.25}) +[2026-06-18 01:07:47,470][starVLA.training.train_starvla][INFO] - Step 5471, Loss: {'train/loss': 0.4297693073749542, 'train/grad_norm_pre_clip': 28.166879653930664, 'timing/data': 0.009971808642148972, 'timing/model': 1.1513683795928955, 'learning_rate/qwen_vl_interface': 1.605271994767216e-06, 'learning_rate/action_model': 1.605271994767216e-05, 'epoch': 2.25}) +[2026-06-18 01:07:51,850][starVLA.training.train_starvla][INFO] - Step 5472, Loss: {'train/loss': 0.17571017146110535, 'train/grad_norm_pre_clip': 14.138364791870117, 'timing/data': 0.01678662747144699, 'timing/model': 1.3939998000860214, 'learning_rate/qwen_vl_interface': 1.6038854932499635e-06, 'learning_rate/action_model': 1.6038854932499633e-05, 'epoch': 2.25}) +[2026-06-18 01:07:55,528][starVLA.training.train_starvla][INFO] - Step 5473, Loss: {'train/loss': 0.4643492102622986, 'train/grad_norm_pre_clip': 26.946365356445312, 'timing/data': 0.011946184560656548, 'timing/model': 0.9333756742998958, 'learning_rate/qwen_vl_interface': 1.6024997475761727e-06, 'learning_rate/action_model': 1.602499747576173e-05, 'epoch': 2.25}) +[2026-06-18 01:07:59,372][starVLA.training.train_starvla][INFO] - Step 5474, Loss: {'train/loss': 0.16975758969783783, 'train/grad_norm_pre_clip': 16.98272132873535, 'timing/data': 0.009558661840856075, 'timing/model': 0.9275015937164426, 'learning_rate/qwen_vl_interface': 1.6011147580331065e-06, 'learning_rate/action_model': 1.6011147580331066e-05, 'epoch': 2.25}) +[2026-06-18 01:08:03,169][starVLA.training.train_starvla][INFO] - Step 5475, Loss: {'train/loss': 0.22337575256824493, 'train/grad_norm_pre_clip': 23.891035079956055, 'timing/data': 0.008126785047352314, 'timing/model': 0.9373866803944111, 'learning_rate/qwen_vl_interface': 1.5997305249078788e-06, 'learning_rate/action_model': 1.5997305249078787e-05, 'epoch': 2.25}) +[2026-06-18 01:08:06,972][starVLA.training.train_starvla][INFO] - Step 5476, Loss: {'train/loss': 0.2520270049571991, 'train/grad_norm_pre_clip': 17.233156204223633, 'timing/data': 0.010861030779778957, 'timing/model': 0.9436963461339474, 'learning_rate/qwen_vl_interface': 1.5983470484874412e-06, 'learning_rate/action_model': 1.598347048487441e-05, 'epoch': 2.25}) +[2026-06-18 01:08:10,765][starVLA.training.train_starvla][INFO] - Step 5477, Loss: {'train/loss': 0.3351730406284332, 'train/grad_norm_pre_clip': 28.788288116455078, 'timing/data': 0.009610096924006939, 'timing/model': 0.9308991134166718, 'learning_rate/qwen_vl_interface': 1.5969643290585893e-06, 'learning_rate/action_model': 1.5969643290585892e-05, 'epoch': 2.25}) +[2026-06-18 01:08:14,610][starVLA.training.train_starvla][INFO] - Step 5478, Loss: {'train/loss': 0.16694465279579163, 'train/grad_norm_pre_clip': 15.027379035949707, 'timing/data': 0.009420652873814106, 'timing/model': 1.1579669676721096, 'learning_rate/qwen_vl_interface': 1.5955823669079632e-06, 'learning_rate/action_model': 1.595582366907963e-05, 'epoch': 2.25}) +[2026-06-18 01:08:18,411][starVLA.training.train_starvla][INFO] - Step 5479, Loss: {'train/loss': 0.5343554615974426, 'train/grad_norm_pre_clip': 26.190759658813477, 'timing/data': 0.010542839765548706, 'timing/model': 1.1165307993069291, 'learning_rate/qwen_vl_interface': 1.594201162322045e-06, 'learning_rate/action_model': 1.594201162322045e-05, 'epoch': 2.26}) +[2026-06-18 01:08:22,023][starVLA.training.train_starvla][INFO] - Step 5480, Loss: {'train/loss': 0.26547956466674805, 'train/grad_norm_pre_clip': 22.048999786376953, 'timing/data': 0.013879166916012764, 'timing/model': 0.9331686478108168, 'learning_rate/qwen_vl_interface': 1.592820715587159e-06, 'learning_rate/action_model': 1.5928207155871593e-05, 'epoch': 2.26}) +[2026-06-18 01:08:26,198][starVLA.training.train_starvla][INFO] - Step 5481, Loss: {'train/loss': 0.3033466935157776, 'train/grad_norm_pre_clip': 19.1571044921875, 'timing/data': 0.010896307416260242, 'timing/model': 0.9314534729346633, 'learning_rate/qwen_vl_interface': 1.5914410269894743e-06, 'learning_rate/action_model': 1.591441026989474e-05, 'epoch': 2.26}) +[2026-06-18 01:08:29,990][starVLA.training.train_starvla][INFO] - Step 5482, Loss: {'train/loss': 0.1756352186203003, 'train/grad_norm_pre_clip': 46.836952209472656, 'timing/data': 0.011645453982055187, 'timing/model': 0.9295256333425641, 'learning_rate/qwen_vl_interface': 1.5900620968150015e-06, 'learning_rate/action_model': 1.5900620968150015e-05, 'epoch': 2.26}) +[2026-06-18 01:08:33,753][starVLA.training.train_starvla][INFO] - Step 5483, Loss: {'train/loss': 0.6826966404914856, 'train/grad_norm_pre_clip': 24.325891494750977, 'timing/data': 0.009597398340702057, 'timing/model': 0.9351614024490118, 'learning_rate/qwen_vl_interface': 1.5886839253495944e-06, 'learning_rate/action_model': 1.5886839253495943e-05, 'epoch': 2.26}) +[2026-06-18 01:08:37,840][starVLA.training.train_starvla][INFO] - Step 5484, Loss: {'train/loss': 0.8221473693847656, 'train/grad_norm_pre_clip': 33.08857345581055, 'timing/data': 0.011028740555047989, 'timing/model': 1.052890312857926, 'learning_rate/qwen_vl_interface': 1.5873065128789495e-06, 'learning_rate/action_model': 1.5873065128789493e-05, 'epoch': 2.26}) +[2026-06-18 01:08:42,511][starVLA.training.train_starvla][INFO] - Step 5485, Loss: {'train/loss': 0.4721921384334564, 'train/grad_norm_pre_clip': 28.700153350830078, 'timing/data': 0.008266372606158257, 'timing/model': 1.171907925978303, 'learning_rate/qwen_vl_interface': 1.5859298596886038e-06, 'learning_rate/action_model': 1.585929859688604e-05, 'epoch': 2.26}) +[2026-06-18 01:08:46,268][starVLA.training.train_starvla][INFO] - Step 5486, Loss: {'train/loss': 0.3823454976081848, 'train/grad_norm_pre_clip': 31.178829193115234, 'timing/data': 0.013551738113164902, 'timing/model': 0.9654355803504586, 'learning_rate/qwen_vl_interface': 1.5845539660639429e-06, 'learning_rate/action_model': 1.5845539660639426e-05, 'epoch': 2.26}) +[2026-06-18 01:08:50,275][starVLA.training.train_starvla][INFO] - Step 5487, Loss: {'train/loss': 0.5000396966934204, 'train/grad_norm_pre_clip': 26.27849006652832, 'timing/data': 0.014540988951921463, 'timing/model': 0.9307567784562707, 'learning_rate/qwen_vl_interface': 1.5831788322901901e-06, 'learning_rate/action_model': 1.58317883229019e-05, 'epoch': 2.26}) +[2026-06-18 01:08:54,044][starVLA.training.train_starvla][INFO] - Step 5488, Loss: {'train/loss': 0.19776935875415802, 'train/grad_norm_pre_clip': 21.823339462280273, 'timing/data': 0.010040657594799995, 'timing/model': 0.9315019054338336, 'learning_rate/qwen_vl_interface': 1.5818044586524092e-06, 'learning_rate/action_model': 1.5818044586524092e-05, 'epoch': 2.26}) +[2026-06-18 01:08:57,902][starVLA.training.train_starvla][INFO] - Step 5489, Loss: {'train/loss': 0.20974305272102356, 'train/grad_norm_pre_clip': 19.43398666381836, 'timing/data': 0.0087833097204566, 'timing/model': 0.9659357499331236, 'learning_rate/qwen_vl_interface': 1.5804308454355102e-06, 'learning_rate/action_model': 1.5804308454355102e-05, 'epoch': 2.26}) +[2026-06-18 01:09:01,769][starVLA.training.train_starvla][INFO] - Step 5490, Loss: {'train/loss': 0.2684885263442993, 'train/grad_norm_pre_clip': 25.539966583251953, 'timing/data': 0.008955615572631359, 'timing/model': 0.9330813670530915, 'learning_rate/qwen_vl_interface': 1.5790579929242475e-06, 'learning_rate/action_model': 1.5790579929242475e-05, 'epoch': 2.26}) +[2026-06-18 01:09:05,591][starVLA.training.train_starvla][INFO] - Step 5491, Loss: {'train/loss': 0.2482248693704605, 'train/grad_norm_pre_clip': 25.210660934448242, 'timing/data': 0.0172213027253747, 'timing/model': 0.9340398516505957, 'learning_rate/qwen_vl_interface': 1.577685901403212e-06, 'learning_rate/action_model': 1.577685901403212e-05, 'epoch': 2.26}) +[2026-06-18 01:09:09,393][starVLA.training.train_starvla][INFO] - Step 5492, Loss: {'train/loss': 0.32996752858161926, 'train/grad_norm_pre_clip': 59.866973876953125, 'timing/data': 0.012290611863136292, 'timing/model': 1.122548141516745, 'learning_rate/qwen_vl_interface': 1.5763145711568414e-06, 'learning_rate/action_model': 1.5763145711568413e-05, 'epoch': 2.26}) +[2026-06-18 01:09:13,012][starVLA.training.train_starvla][INFO] - Step 5493, Loss: {'train/loss': 0.35097649693489075, 'train/grad_norm_pre_clip': 23.73944854736328, 'timing/data': 0.010956252925097942, 'timing/model': 0.9313974743708968, 'learning_rate/qwen_vl_interface': 1.5749440024694129e-06, 'learning_rate/action_model': 1.574944002469413e-05, 'epoch': 2.26}) +[2026-06-18 01:09:16,826][starVLA.training.train_starvla][INFO] - Step 5494, Loss: {'train/loss': 0.2840988039970398, 'train/grad_norm_pre_clip': 16.209423065185547, 'timing/data': 0.009392805397510529, 'timing/model': 0.9308898719027638, 'learning_rate/qwen_vl_interface': 1.5735741956250476e-06, 'learning_rate/action_model': 1.5735741956250475e-05, 'epoch': 2.26}) +[2026-06-18 01:09:20,597][starVLA.training.train_starvla][INFO] - Step 5495, Loss: {'train/loss': 0.09620679169893265, 'train/grad_norm_pre_clip': 10.452868461608887, 'timing/data': 0.01407526433467865, 'timing/model': 0.9275918137282133, 'learning_rate/qwen_vl_interface': 1.5722051509077068e-06, 'learning_rate/action_model': 1.572205150907707e-05, 'epoch': 2.26}) +[2026-06-18 01:09:24,436][starVLA.training.train_starvla][INFO] - Step 5496, Loss: {'train/loss': 0.23830297589302063, 'train/grad_norm_pre_clip': 22.351531982421875, 'timing/data': 0.011238726787269115, 'timing/model': 0.933891006745398, 'learning_rate/qwen_vl_interface': 1.570836868601196e-06, 'learning_rate/action_model': 1.570836868601196e-05, 'epoch': 2.26}) +[2026-06-18 01:09:28,321][starVLA.training.train_starvla][INFO] - Step 5497, Loss: {'train/loss': 0.3838105797767639, 'train/grad_norm_pre_clip': 25.52385902404785, 'timing/data': 0.014472104609012604, 'timing/model': 0.9389657769352198, 'learning_rate/qwen_vl_interface': 1.5694693489891616e-06, 'learning_rate/action_model': 1.5694693489891617e-05, 'epoch': 2.26}) +[2026-06-18 01:09:33,202][starVLA.training.train_starvla][INFO] - Step 5498, Loss: {'train/loss': 0.42125123739242554, 'train/grad_norm_pre_clip': 41.729347229003906, 'timing/data': 0.008443544618785381, 'timing/model': 1.1688340744003654, 'learning_rate/qwen_vl_interface': 1.5681025923550914e-06, 'learning_rate/action_model': 1.568102592355091e-05, 'epoch': 2.26}) +[2026-06-18 01:09:37,121][starVLA.training.train_starvla][INFO] - Step 5499, Loss: {'train/loss': 0.42060619592666626, 'train/grad_norm_pre_clip': 24.12847137451172, 'timing/data': 0.008280085399746895, 'timing/model': 1.1164087317883968, 'learning_rate/qwen_vl_interface': 1.5667365989823161e-06, 'learning_rate/action_model': 1.566736598982316e-05, 'epoch': 2.26}) +[2026-06-18 01:11:37,113][starVLA.training.train_starvla][INFO] - Step 5500, Loss: {'train/loss': 0.13149771094322205, 'train/grad_norm_pre_clip': 15.63851547241211, 'eval/loss': 0.41326525807380676, 'eval/latency_6/loss': 0.41326525807380676, 'eval/demon_fire_precision': 0.954775458570525, 'eval/demon_fire_recall': 0.975759534583064, 'eval/demon_fire_f1': 0.9651534526854221, 'eval/demon_fire_rate_model': 0.8797996661101837, 'eval/demon_fire_rate_teacher': 0.860879243183083, 'eval/demon_left_precision': 0.8824742268041237, 'eval/demon_left_recall': 0.7508771929824561, 'eval/demon_left_f1': 0.8113744075829383, 'eval/demon_left_rate_model': 0.13494713411240958, 'eval/demon_left_rate_teacher': 0.15859766277128548, 'eval/demon_right_precision': 0.809375, 'eval/demon_right_recall': 0.7956989247311828, 'eval/demon_right_f1': 0.8024786986831912, 'eval/demon_right_rate_model': 0.17807456872565386, 'eval/demon_right_rate_teacher': 0.18113522537562604, 'eval/demon_move_f1': 0.8069265531330647, 'eval/demon_cc_f1': 0.9115318938764477, 'eval/latency_6/demon_fire_precision': 0.954775458570525, 'eval/latency_6/demon_fire_recall': 0.975759534583064, 'eval/latency_6/demon_fire_f1': 0.9651534526854221, 'eval/latency_6/demon_fire_rate_model': 0.8797996661101837, 'eval/latency_6/demon_fire_rate_teacher': 0.860879243183083, 'eval/latency_6/demon_left_precision': 0.8824742268041237, 'eval/latency_6/demon_left_recall': 0.7508771929824561, 'eval/latency_6/demon_left_f1': 0.8113744075829383, 'eval/latency_6/demon_left_rate_model': 0.13494713411240958, 'eval/latency_6/demon_left_rate_teacher': 0.15859766277128548, 'eval/latency_6/demon_right_precision': 0.809375, 'eval/latency_6/demon_right_recall': 0.7956989247311828, 'eval/latency_6/demon_right_f1': 0.8024786986831912, 'eval/latency_6/demon_right_rate_model': 0.17807456872565386, 'eval/latency_6/demon_right_rate_teacher': 0.18113522537562604, 'eval/latency_6/demon_move_f1': 0.8069265531330647, 'eval/latency_6/demon_cc_f1': 0.9115318938764477, 'timing/data': 0.008403879590332508, 'timing/model': 0.9322847090661526, 'learning_rate/qwen_vl_interface': 1.5653713691540053e-06, 'learning_rate/action_model': 1.5653713691540054e-05, 'epoch': 2.26}) +[2026-06-18 01:11:37,115][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_5500_state +[2026-06-18 01:11:37,135][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 01:11:43,748][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_5500_state/model.safetensors +[2026-06-18 01:11:59,407][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_5500_state/optimizer.bin +[2026-06-18 01:11:59,410][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_5500_state/sampler.bin +[2026-06-18 01:11:59,410][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_5500_state/sampler_1.bin +[2026-06-18 01:11:59,412][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_5500_state/random_states_0.pkl +[2026-06-18 01:11:59,416][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 5500 +[2026-06-18 01:11:59,416][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 01:11:59,427][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 01:12:03,295][starVLA.training.train_starvla][INFO] - Step 5501, Loss: {'train/loss': 0.3436428904533386, 'train/grad_norm_pre_clip': 35.60099411010742, 'timing/data': 0.011484150774776936, 'timing/model': 1.1492322543635964, 'learning_rate/qwen_vl_interface': 1.5640069031531774e-06, 'learning_rate/action_model': 1.5640069031531772e-05, 'epoch': 2.26}) +[2026-06-18 01:12:06,961][starVLA.training.train_starvla][INFO] - Step 5502, Loss: {'train/loss': 0.2692725360393524, 'train/grad_norm_pre_clip': 17.767925262451172, 'timing/data': 0.011489932425320148, 'timing/model': 0.961303099989891, 'learning_rate/qwen_vl_interface': 1.5626432012626835e-06, 'learning_rate/action_model': 1.5626432012626836e-05, 'epoch': 2.26}) +[2026-06-18 01:12:10,932][starVLA.training.train_starvla][INFO] - Step 5503, Loss: {'train/loss': 0.3122205138206482, 'train/grad_norm_pre_clip': 24.724668502807617, 'timing/data': 0.008813533000648022, 'timing/model': 0.9420470492914319, 'learning_rate/qwen_vl_interface': 1.5612802637652202e-06, 'learning_rate/action_model': 1.5612802637652203e-05, 'epoch': 2.27}) +[2026-06-18 01:12:14,744][starVLA.training.train_starvla][INFO] - Step 5504, Loss: {'train/loss': 0.2267666906118393, 'train/grad_norm_pre_clip': 49.6044921875, 'timing/data': 0.011395446956157684, 'timing/model': 0.9331687716767192, 'learning_rate/qwen_vl_interface': 1.5599180909433292e-06, 'learning_rate/action_model': 1.5599180909433292e-05, 'epoch': 2.27}) +[2026-06-18 01:12:18,502][starVLA.training.train_starvla][INFO] - Step 5505, Loss: {'train/loss': 0.07848943769931793, 'train/grad_norm_pre_clip': 7.803404808044434, 'timing/data': 0.012197902426123619, 'timing/model': 0.9316668529063463, 'learning_rate/qwen_vl_interface': 1.5585566830793883e-06, 'learning_rate/action_model': 1.558556683079388e-05, 'epoch': 2.27}) +[2026-06-18 01:12:22,297][starVLA.training.train_starvla][INFO] - Step 5506, Loss: {'train/loss': 0.2101878821849823, 'train/grad_norm_pre_clip': 24.74462890625, 'timing/data': 0.009617173112928867, 'timing/model': 0.93166867736727, 'learning_rate/qwen_vl_interface': 1.5571960404556204e-06, 'learning_rate/action_model': 1.5571960404556203e-05, 'epoch': 2.27}) +[2026-06-18 01:12:26,122][starVLA.training.train_starvla][INFO] - Step 5507, Loss: {'train/loss': 0.3783796429634094, 'train/grad_norm_pre_clip': 22.017784118652344, 'timing/data': 0.019676264375448227, 'timing/model': 1.1328481705859303, 'learning_rate/qwen_vl_interface': 1.5558361633540848e-06, 'learning_rate/action_model': 1.5558361633540848e-05, 'epoch': 2.27}) +[2026-06-18 01:12:29,742][starVLA.training.train_starvla][INFO] - Step 5508, Loss: {'train/loss': 0.45755535364151, 'train/grad_norm_pre_clip': 26.35930824279785, 'timing/data': 0.01045374944806099, 'timing/model': 0.9323258753865957, 'learning_rate/qwen_vl_interface': 1.5544770520566884e-06, 'learning_rate/action_model': 1.5544770520566885e-05, 'epoch': 2.27}) +[2026-06-18 01:12:33,525][starVLA.training.train_starvla][INFO] - Step 5509, Loss: {'train/loss': 0.37104082107543945, 'train/grad_norm_pre_clip': 23.668642044067383, 'timing/data': 0.016000057570636272, 'timing/model': 0.9322741935029626, 'learning_rate/qwen_vl_interface': 1.553118706845176e-06, 'learning_rate/action_model': 1.5531187068451757e-05, 'epoch': 2.27}) +[2026-06-18 01:12:37,371][starVLA.training.train_starvla][INFO] - Step 5510, Loss: {'train/loss': 0.32382529973983765, 'train/grad_norm_pre_clip': 21.550798416137695, 'timing/data': 0.007175063714385033, 'timing/model': 0.9361705826595426, 'learning_rate/qwen_vl_interface': 1.5517611280011333e-06, 'learning_rate/action_model': 1.551761128001133e-05, 'epoch': 2.27}) +[2026-06-18 01:12:41,216][starVLA.training.train_starvla][INFO] - Step 5511, Loss: {'train/loss': 0.17942291498184204, 'train/grad_norm_pre_clip': 11.945369720458984, 'timing/data': 0.01245593186467886, 'timing/model': 0.932456131093204, 'learning_rate/qwen_vl_interface': 1.5504043158059886e-06, 'learning_rate/action_model': 1.5504043158059885e-05, 'epoch': 2.27}) +[2026-06-18 01:12:45,015][starVLA.training.train_starvla][INFO] - Step 5512, Loss: {'train/loss': 0.3502919375896454, 'train/grad_norm_pre_clip': 23.978069305419922, 'timing/data': 0.009553776122629642, 'timing/model': 0.9302502162754536, 'learning_rate/qwen_vl_interface': 1.54904827054101e-06, 'learning_rate/action_model': 1.54904827054101e-05, 'epoch': 2.27}) +[2026-06-18 01:12:48,890][starVLA.training.train_starvla][INFO] - Step 5513, Loss: {'train/loss': 0.22415779531002045, 'train/grad_norm_pre_clip': 18.174755096435547, 'timing/data': 0.013756693340837955, 'timing/model': 1.137713729403913, 'learning_rate/qwen_vl_interface': 1.5476929924873074e-06, 'learning_rate/action_model': 1.5476929924873073e-05, 'epoch': 2.27}) +[2026-06-18 01:12:53,626][starVLA.training.train_starvla][INFO] - Step 5514, Loss: {'train/loss': 0.5045932531356812, 'train/grad_norm_pre_clip': 37.77540588378906, 'timing/data': 0.00909426063299179, 'timing/model': 0.953845857642591, 'learning_rate/qwen_vl_interface': 1.5463384819258302e-06, 'learning_rate/action_model': 1.54633848192583e-05, 'epoch': 2.27}) +[2026-06-18 01:12:57,692][starVLA.training.train_starvla][INFO] - Step 5515, Loss: {'train/loss': 0.255596786737442, 'train/grad_norm_pre_clip': 23.315561294555664, 'timing/data': 0.009947845712304115, 'timing/model': 1.1850428003817797, 'learning_rate/qwen_vl_interface': 1.544984739137373e-06, 'learning_rate/action_model': 1.544984739137373e-05, 'epoch': 2.27}) +[2026-06-18 01:13:01,913][starVLA.training.train_starvla][INFO] - Step 5516, Loss: {'train/loss': 0.05186809226870537, 'train/grad_norm_pre_clip': 7.573668479919434, 'timing/data': 0.012987270951271057, 'timing/model': 0.9597749700769782, 'learning_rate/qwen_vl_interface': 1.543631764402566e-06, 'learning_rate/action_model': 1.5436317644025658e-05, 'epoch': 2.27}) +[2026-06-18 01:13:05,734][starVLA.training.train_starvla][INFO] - Step 5517, Loss: {'train/loss': 0.4218573570251465, 'train/grad_norm_pre_clip': 24.951948165893555, 'timing/data': 0.008339514955878258, 'timing/model': 0.938437326811254, 'learning_rate/qwen_vl_interface': 1.5422795580018813e-06, 'learning_rate/action_model': 1.542279558001881e-05, 'epoch': 2.27}) +[2026-06-18 01:13:09,627][starVLA.training.train_starvla][INFO] - Step 5518, Loss: {'train/loss': 0.2601926624774933, 'train/grad_norm_pre_clip': 22.52280044555664, 'timing/data': 0.010533643886446953, 'timing/model': 0.9530413616448641, 'learning_rate/qwen_vl_interface': 1.5409281202156329e-06, 'learning_rate/action_model': 1.540928120215633e-05, 'epoch': 2.27}) +[2026-06-18 01:13:13,460][starVLA.training.train_starvla][INFO] - Step 5519, Loss: {'train/loss': 0.4734601080417633, 'train/grad_norm_pre_clip': 26.15658950805664, 'timing/data': 0.0123616186901927, 'timing/model': 0.9497531438246369, 'learning_rate/qwen_vl_interface': 1.5395774513239777e-06, 'learning_rate/action_model': 1.539577451323978e-05, 'epoch': 2.27}) +[2026-06-18 01:13:17,459][starVLA.training.train_starvla][INFO] - Step 5520, Loss: {'train/loss': 0.14145860075950623, 'train/grad_norm_pre_clip': 16.283823013305664, 'timing/data': 0.016726276837289333, 'timing/model': 1.1231373213231564, 'learning_rate/qwen_vl_interface': 1.5382275516069112e-06, 'learning_rate/action_model': 1.5382275516069113e-05, 'epoch': 2.27}) +[2026-06-18 01:13:21,205][starVLA.training.train_starvla][INFO] - Step 5521, Loss: {'train/loss': 0.22163018584251404, 'train/grad_norm_pre_clip': 13.878335952758789, 'timing/data': 0.012297128327190876, 'timing/model': 1.0178272854536772, 'learning_rate/qwen_vl_interface': 1.5368784213442653e-06, 'learning_rate/action_model': 1.5368784213442654e-05, 'epoch': 2.27}) +[2026-06-18 01:13:25,051][starVLA.training.train_starvla][INFO] - Step 5522, Loss: {'train/loss': 0.17253945767879486, 'train/grad_norm_pre_clip': 16.150066375732422, 'timing/data': 0.017340773716568947, 'timing/model': 0.9594711354002357, 'learning_rate/qwen_vl_interface': 1.5355300608157172e-06, 'learning_rate/action_model': 1.5355300608157173e-05, 'epoch': 2.27}) +[2026-06-18 01:13:28,981][starVLA.training.train_starvla][INFO] - Step 5523, Loss: {'train/loss': 0.32440561056137085, 'train/grad_norm_pre_clip': 22.734811782836914, 'timing/data': 0.01212645135819912, 'timing/model': 0.9668804593384266, 'learning_rate/qwen_vl_interface': 1.5341824703007854e-06, 'learning_rate/action_model': 1.5341824703007854e-05, 'epoch': 2.27}) +[2026-06-18 01:13:32,809][starVLA.training.train_starvla][INFO] - Step 5524, Loss: {'train/loss': 0.3402218818664551, 'train/grad_norm_pre_clip': 26.061283111572266, 'timing/data': 0.009159366600215435, 'timing/model': 0.9452540362253785, 'learning_rate/qwen_vl_interface': 1.5328356500788263e-06, 'learning_rate/action_model': 1.5328356500788264e-05, 'epoch': 2.27}) +[2026-06-18 01:13:36,614][starVLA.training.train_starvla][INFO] - Step 5525, Loss: {'train/loss': 0.31070616841316223, 'train/grad_norm_pre_clip': 15.799114227294922, 'timing/data': 0.00804527010768652, 'timing/model': 0.947526266798377, 'learning_rate/qwen_vl_interface': 1.5314896004290366e-06, 'learning_rate/action_model': 1.5314896004290365e-05, 'epoch': 2.27}) +[2026-06-18 01:13:40,411][starVLA.training.train_starvla][INFO] - Step 5526, Loss: {'train/loss': 0.47461560368537903, 'train/grad_norm_pre_clip': 23.180246353149414, 'timing/data': 0.008044187910854816, 'timing/model': 1.106463055126369, 'learning_rate/qwen_vl_interface': 1.530144321630454e-06, 'learning_rate/action_model': 1.530144321630454e-05, 'epoch': 2.27}) +[2026-06-18 01:13:44,199][starVLA.training.train_starvla][INFO] - Step 5527, Loss: {'train/loss': 0.24527285993099213, 'train/grad_norm_pre_clip': 16.187562942504883, 'timing/data': 0.00922598596662283, 'timing/model': 1.1119739646092057, 'learning_rate/qwen_vl_interface': 1.5287998139619564e-06, 'learning_rate/action_model': 1.5287998139619562e-05, 'epoch': 2.28}) +[2026-06-18 01:13:47,833][starVLA.training.train_starvla][INFO] - Step 5528, Loss: {'train/loss': 0.14699789881706238, 'train/grad_norm_pre_clip': 12.64822006225586, 'timing/data': 0.011216312646865845, 'timing/model': 0.93744957447052, 'learning_rate/qwen_vl_interface': 1.5274560777022613e-06, 'learning_rate/action_model': 1.5274560777022613e-05, 'epoch': 2.28}) +[2026-06-18 01:13:51,645][starVLA.training.train_starvla][INFO] - Step 5529, Loss: {'train/loss': 0.24265143275260925, 'train/grad_norm_pre_clip': 13.319592475891113, 'timing/data': 0.00955337192863226, 'timing/model': 0.9588142381981015, 'learning_rate/qwen_vl_interface': 1.5261131131299274e-06, 'learning_rate/action_model': 1.5261131131299272e-05, 'epoch': 2.28}) +[2026-06-18 01:13:55,502][starVLA.training.train_starvla][INFO] - Step 5530, Loss: {'train/loss': 0.3577382266521454, 'train/grad_norm_pre_clip': 26.23959732055664, 'timing/data': 0.009532558731734753, 'timing/model': 0.9530413346365094, 'learning_rate/qwen_vl_interface': 1.5247709205233518e-06, 'learning_rate/action_model': 1.5247709205233518e-05, 'epoch': 2.28}) +[2026-06-18 01:13:59,364][starVLA.training.train_starvla][INFO] - Step 5531, Loss: {'train/loss': 0.278052419424057, 'train/grad_norm_pre_clip': 22.980621337890625, 'timing/data': 0.009338708594441414, 'timing/model': 0.9684680104255676, 'learning_rate/qwen_vl_interface': 1.5234295001607732e-06, 'learning_rate/action_model': 1.5234295001607732e-05, 'epoch': 2.28}) +[2026-06-18 01:14:03,238][starVLA.training.train_starvla][INFO] - Step 5532, Loss: {'train/loss': 0.09260043501853943, 'train/grad_norm_pre_clip': 8.31368637084961, 'timing/data': 0.0080458028241992, 'timing/model': 0.933267462067306, 'learning_rate/qwen_vl_interface': 1.5220888523202687e-06, 'learning_rate/action_model': 1.5220888523202686e-05, 'epoch': 2.28}) +[2026-06-18 01:14:07,148][starVLA.training.train_starvla][INFO] - Step 5533, Loss: {'train/loss': 0.5085386633872986, 'train/grad_norm_pre_clip': 24.82141876220703, 'timing/data': 0.010007013566792011, 'timing/model': 1.1775770708918571, 'learning_rate/qwen_vl_interface': 1.5207489772797557e-06, 'learning_rate/action_model': 1.5207489772797556e-05, 'epoch': 2.28}) +[2026-06-18 01:14:10,824][starVLA.training.train_starvla][INFO] - Step 5534, Loss: {'train/loss': 0.13090376555919647, 'train/grad_norm_pre_clip': 13.773388862609863, 'timing/data': 0.009706760756671429, 'timing/model': 0.9539364706724882, 'learning_rate/qwen_vl_interface': 1.5194098753169933e-06, 'learning_rate/action_model': 1.5194098753169933e-05, 'epoch': 2.28}) +[2026-06-18 01:14:14,645][starVLA.training.train_starvla][INFO] - Step 5535, Loss: {'train/loss': 0.14413826167583466, 'train/grad_norm_pre_clip': 15.989876747131348, 'timing/data': 0.013181236572563648, 'timing/model': 0.9356067590415478, 'learning_rate/qwen_vl_interface': 1.5180715467095799e-06, 'learning_rate/action_model': 1.5180715467095798e-05, 'epoch': 2.28}) +[2026-06-18 01:14:18,490][starVLA.training.train_starvla][INFO] - Step 5536, Loss: {'train/loss': 0.24486178159713745, 'train/grad_norm_pre_clip': 20.36451530456543, 'timing/data': 0.016067917458713055, 'timing/model': 0.943646720610559, 'learning_rate/qwen_vl_interface': 1.5167339917349482e-06, 'learning_rate/action_model': 1.5167339917349482e-05, 'epoch': 2.28}) +[2026-06-18 01:14:22,236][starVLA.training.train_starvla][INFO] - Step 5537, Loss: {'train/loss': 0.47778385877609253, 'train/grad_norm_pre_clip': 25.04164695739746, 'timing/data': 0.008704856969416142, 'timing/model': 0.9299929393455386, 'learning_rate/qwen_vl_interface': 1.5153972106703778e-06, 'learning_rate/action_model': 1.5153972106703779e-05, 'epoch': 2.28}) +[2026-06-18 01:14:26,071][starVLA.training.train_starvla][INFO] - Step 5538, Loss: {'train/loss': 0.158633753657341, 'train/grad_norm_pre_clip': 15.016270637512207, 'timing/data': 0.012277927249670029, 'timing/model': 0.9279932295903563, 'learning_rate/qwen_vl_interface': 1.5140612037929844e-06, 'learning_rate/action_model': 1.5140612037929844e-05, 'epoch': 2.28}) +[2026-06-18 01:14:30,028][starVLA.training.train_starvla][INFO] - Step 5539, Loss: {'train/loss': 0.1995885968208313, 'train/grad_norm_pre_clip': 16.903301239013672, 'timing/data': 0.014109919779002666, 'timing/model': 1.1668758476153016, 'learning_rate/qwen_vl_interface': 1.512725971379724e-06, 'learning_rate/action_model': 1.5127259713797239e-05, 'epoch': 2.28}) +[2026-06-18 01:14:33,832][starVLA.training.train_starvla][INFO] - Step 5540, Loss: {'train/loss': 0.284618079662323, 'train/grad_norm_pre_clip': 24.939455032348633, 'timing/data': 0.014197209849953651, 'timing/model': 1.0905422857031226, 'learning_rate/qwen_vl_interface': 1.5113915137073905e-06, 'learning_rate/action_model': 1.5113915137073906e-05, 'epoch': 2.28}) +[2026-06-18 01:14:37,475][starVLA.training.train_starvla][INFO] - Step 5541, Loss: {'train/loss': 0.17712675034999847, 'train/grad_norm_pre_clip': 18.337337493896484, 'timing/data': 0.012364229187369347, 'timing/model': 0.9348444612696767, 'learning_rate/qwen_vl_interface': 1.5100578310526198e-06, 'learning_rate/action_model': 1.5100578310526197e-05, 'epoch': 2.28}) +[2026-06-18 01:14:41,510][starVLA.training.train_starvla][INFO] - Step 5542, Loss: {'train/loss': 0.2689046859741211, 'train/grad_norm_pre_clip': 14.903437614440918, 'timing/data': 0.008633268997073174, 'timing/model': 0.9501547766849399, 'learning_rate/qwen_vl_interface': 1.508724923691885e-06, 'learning_rate/action_model': 1.5087249236918848e-05, 'epoch': 2.28}) +[2026-06-18 01:14:45,503][starVLA.training.train_starvla][INFO] - Step 5543, Loss: {'train/loss': 0.28530773520469666, 'train/grad_norm_pre_clip': 24.817630767822266, 'timing/data': 0.014139075763523579, 'timing/model': 1.0220050998032093, 'learning_rate/qwen_vl_interface': 1.5073927919014986e-06, 'learning_rate/action_model': 1.5073927919014985e-05, 'epoch': 2.28}) +[2026-06-18 01:14:49,428][starVLA.training.train_starvla][INFO] - Step 5544, Loss: {'train/loss': 0.30302268266677856, 'train/grad_norm_pre_clip': 26.94271469116211, 'timing/data': 0.011917948722839355, 'timing/model': 0.9737341571599245, 'learning_rate/qwen_vl_interface': 1.506061435957614e-06, 'learning_rate/action_model': 1.506061435957614e-05, 'epoch': 2.28}) +[2026-06-18 01:14:53,442][starVLA.training.train_starvla][INFO] - Step 5545, Loss: {'train/loss': 0.2522331774234772, 'train/grad_norm_pre_clip': 21.444583892822266, 'timing/data': 0.009200791828334332, 'timing/model': 1.0361808575689793, 'learning_rate/qwen_vl_interface': 1.5047308561362225e-06, 'learning_rate/action_model': 1.5047308561362225e-05, 'epoch': 2.28}) +[2026-06-18 01:14:57,260][starVLA.training.train_starvla][INFO] - Step 5546, Loss: {'train/loss': 0.04570770636200905, 'train/grad_norm_pre_clip': 6.889963626861572, 'timing/data': 0.00944796483963728, 'timing/model': 1.1083303894847631, 'learning_rate/qwen_vl_interface': 1.5034010527131548e-06, 'learning_rate/action_model': 1.5034010527131549e-05, 'epoch': 2.28}) +[2026-06-18 01:15:00,941][starVLA.training.train_starvla][INFO] - Step 5547, Loss: {'train/loss': 0.5643250942230225, 'train/grad_norm_pre_clip': 28.122581481933594, 'timing/data': 0.01946009323000908, 'timing/model': 0.9469409566372633, 'learning_rate/qwen_vl_interface': 1.5020720259640792e-06, 'learning_rate/action_model': 1.5020720259640792e-05, 'epoch': 2.28}) +[2026-06-18 01:15:05,106][starVLA.training.train_starvla][INFO] - Step 5548, Loss: {'train/loss': 0.17620548605918884, 'train/grad_norm_pre_clip': 17.907732009887695, 'timing/data': 0.01109307911247015, 'timing/model': 0.964238871820271, 'learning_rate/qwen_vl_interface': 1.5007437761645064e-06, 'learning_rate/action_model': 1.5007437761645063e-05, 'epoch': 2.28}) +[2026-06-18 01:15:08,991][starVLA.training.train_starvla][INFO] - Step 5549, Loss: {'train/loss': 0.3053317070007324, 'train/grad_norm_pre_clip': 25.492727279663086, 'timing/data': 0.008709322661161423, 'timing/model': 0.9434931399300694, 'learning_rate/qwen_vl_interface': 1.4994163035897845e-06, 'learning_rate/action_model': 1.4994163035897846e-05, 'epoch': 2.28}) +[2026-06-18 01:15:12,912][starVLA.training.train_starvla][INFO] - Step 5550, Loss: {'train/loss': 0.30173006653785706, 'train/grad_norm_pre_clip': 20.548051834106445, 'timing/data': 0.009517766535282135, 'timing/model': 0.9503513434901834, 'learning_rate/qwen_vl_interface': 1.4980896085150976e-06, 'learning_rate/action_model': 1.4980896085150975e-05, 'epoch': 2.28}) +[2026-06-18 01:15:16,845][starVLA.training.train_starvla][INFO] - Step 5551, Loss: {'train/loss': 0.42748337984085083, 'train/grad_norm_pre_clip': 27.489028930664062, 'timing/data': 0.00837622582912445, 'timing/model': 0.9447073033079505, 'learning_rate/qwen_vl_interface': 1.4967636912154707e-06, 'learning_rate/action_model': 1.4967636912154706e-05, 'epoch': 2.29}) +[2026-06-18 01:15:21,275][starVLA.training.train_starvla][INFO] - Step 5552, Loss: {'train/loss': 0.10395632684230804, 'train/grad_norm_pre_clip': 13.082822799682617, 'timing/data': 0.011622747406363487, 'timing/model': 1.4882311578840017, 'learning_rate/qwen_vl_interface': 1.4954385519657701e-06, 'learning_rate/action_model': 1.49543855196577e-05, 'epoch': 2.29}) +[2026-06-18 01:15:25,149][starVLA.training.train_starvla][INFO] - Step 5553, Loss: {'train/loss': 0.11261013895273209, 'train/grad_norm_pre_clip': 13.362397193908691, 'timing/data': 0.009384450502693653, 'timing/model': 1.1332554230466485, 'learning_rate/qwen_vl_interface': 1.4941141910406984e-06, 'learning_rate/action_model': 1.4941141910406984e-05, 'epoch': 2.29}) +[2026-06-18 01:15:28,884][starVLA.training.train_starvla][INFO] - Step 5554, Loss: {'train/loss': 0.2812878489494324, 'train/grad_norm_pre_clip': 17.62181854248047, 'timing/data': 0.009537642821669579, 'timing/model': 0.9434270868077874, 'learning_rate/qwen_vl_interface': 1.4927906087147966e-06, 'learning_rate/action_model': 1.4927906087147967e-05, 'epoch': 2.29}) +[2026-06-18 01:15:32,957][starVLA.training.train_starvla][INFO] - Step 5555, Loss: {'train/loss': 0.35659894347190857, 'train/grad_norm_pre_clip': 20.685592651367188, 'timing/data': 0.009516360238194466, 'timing/model': 1.0895844316110015, 'learning_rate/qwen_vl_interface': 1.491467805262444e-06, 'learning_rate/action_model': 1.4914678052624439e-05, 'epoch': 2.29}) +[2026-06-18 01:15:37,349][starVLA.training.train_starvla][INFO] - Step 5556, Loss: {'train/loss': 0.2997216284275055, 'train/grad_norm_pre_clip': 17.362686157226562, 'timing/data': 0.008035301230847836, 'timing/model': 0.9383277371525764, 'learning_rate/qwen_vl_interface': 1.4901457809578598e-06, 'learning_rate/action_model': 1.4901457809578597e-05, 'epoch': 2.29}) +[2026-06-18 01:15:41,805][starVLA.training.train_starvla][INFO] - Step 5557, Loss: {'train/loss': 0.1950259804725647, 'train/grad_norm_pre_clip': 24.704486846923828, 'timing/data': 0.012435819022357464, 'timing/model': 0.9541793894022703, 'learning_rate/qwen_vl_interface': 1.4888245360751014e-06, 'learning_rate/action_model': 1.4888245360751013e-05, 'epoch': 2.29}) +[2026-06-18 01:15:45,871][starVLA.training.train_starvla][INFO] - Step 5558, Loss: {'train/loss': 0.35749873518943787, 'train/grad_norm_pre_clip': 24.082162857055664, 'timing/data': 0.008183146826922894, 'timing/model': 1.2074572341516614, 'learning_rate/qwen_vl_interface': 1.4875040708880631e-06, 'learning_rate/action_model': 1.487504070888063e-05, 'epoch': 2.29}) +[2026-06-18 01:15:49,847][starVLA.training.train_starvla][INFO] - Step 5559, Loss: {'train/loss': 0.46382927894592285, 'train/grad_norm_pre_clip': 22.132417678833008, 'timing/data': 0.010984392836689949, 'timing/model': 1.166614598594606, 'learning_rate/qwen_vl_interface': 1.4861843856704802e-06, 'learning_rate/action_model': 1.4861843856704801e-05, 'epoch': 2.29}) +[2026-06-18 01:15:53,670][starVLA.training.train_starvla][INFO] - Step 5560, Loss: {'train/loss': 0.29021140933036804, 'train/grad_norm_pre_clip': 21.55061912536621, 'timing/data': 0.008923622779548168, 'timing/model': 1.1229807445779443, 'learning_rate/qwen_vl_interface': 1.4848654806959243e-06, 'learning_rate/action_model': 1.4848654806959242e-05, 'epoch': 2.29}) +[2026-06-18 01:15:57,339][starVLA.training.train_starvla][INFO] - Step 5561, Loss: {'train/loss': 0.28531116247177124, 'train/grad_norm_pre_clip': 21.69508934020996, 'timing/data': 0.011481263674795628, 'timing/model': 0.9390454348176718, 'learning_rate/qwen_vl_interface': 1.483547356237805e-06, 'learning_rate/action_model': 1.4835473562378049e-05, 'epoch': 2.29}) +[2026-06-18 01:16:01,239][starVLA.training.train_starvla][INFO] - Step 5562, Loss: {'train/loss': 0.27355721592903137, 'train/grad_norm_pre_clip': 14.209111213684082, 'timing/data': 0.008714440278708935, 'timing/model': 0.967422392219305, 'learning_rate/qwen_vl_interface': 1.4822300125693709e-06, 'learning_rate/action_model': 1.4822300125693708e-05, 'epoch': 2.29}) +[2026-06-18 01:16:05,236][starVLA.training.train_starvla][INFO] - Step 5563, Loss: {'train/loss': 0.2023213505744934, 'train/grad_norm_pre_clip': 16.935096740722656, 'timing/data': 0.011965841054916382, 'timing/model': 0.9432314308360219, 'learning_rate/qwen_vl_interface': 1.480913449963712e-06, 'learning_rate/action_model': 1.480913449963712e-05, 'epoch': 2.29}) +[2026-06-18 01:16:09,137][starVLA.training.train_starvla][INFO] - Step 5564, Loss: {'train/loss': 0.24206194281578064, 'train/grad_norm_pre_clip': 22.34685707092285, 'timing/data': 0.010241067036986351, 'timing/model': 0.9448322672396898, 'learning_rate/qwen_vl_interface': 1.4795976686937488e-06, 'learning_rate/action_model': 1.4795976686937487e-05, 'epoch': 2.29}) +[2026-06-18 01:16:13,047][starVLA.training.train_starvla][INFO] - Step 5565, Loss: {'train/loss': 0.19147741794586182, 'train/grad_norm_pre_clip': 16.995235443115234, 'timing/data': 0.008172263391315937, 'timing/model': 0.9604474622756243, 'learning_rate/qwen_vl_interface': 1.4782826690322453e-06, 'learning_rate/action_model': 1.4782826690322451e-05, 'epoch': 2.29}) +[2026-06-18 01:16:16,957][starVLA.training.train_starvla][INFO] - Step 5566, Loss: {'train/loss': 0.23130233585834503, 'train/grad_norm_pre_clip': 19.363534927368164, 'timing/data': 0.009477196261286736, 'timing/model': 1.1579631101340055, 'learning_rate/qwen_vl_interface': 1.4769684512518045e-06, 'learning_rate/action_model': 1.4769684512518045e-05, 'epoch': 2.29}) +[2026-06-18 01:16:20,820][starVLA.training.train_starvla][INFO] - Step 5567, Loss: {'train/loss': 0.06105984374880791, 'train/grad_norm_pre_clip': 9.610785484313965, 'timing/data': 0.010832524858415127, 'timing/model': 1.066472064703703, 'learning_rate/qwen_vl_interface': 1.4756550156248633e-06, 'learning_rate/action_model': 1.4756550156248632e-05, 'epoch': 2.29}) +[2026-06-18 01:16:24,623][starVLA.training.train_starvla][INFO] - Step 5568, Loss: {'train/loss': 0.44550058245658875, 'train/grad_norm_pre_clip': 28.787641525268555, 'timing/data': 0.012630593031644821, 'timing/model': 0.9336195169016719, 'learning_rate/qwen_vl_interface': 1.4743423624237e-06, 'learning_rate/action_model': 1.4743423624236998e-05, 'epoch': 2.29}) +[2026-06-18 01:16:28,591][starVLA.training.train_starvla][INFO] - Step 5569, Loss: {'train/loss': 0.1861807256937027, 'train/grad_norm_pre_clip': 18.426504135131836, 'timing/data': 0.011492704041302204, 'timing/model': 1.0133573804050684, 'learning_rate/qwen_vl_interface': 1.4730304919204253e-06, 'learning_rate/action_model': 1.4730304919204252e-05, 'epoch': 2.29}) +[2026-06-18 01:16:32,429][starVLA.training.train_starvla][INFO] - Step 5570, Loss: {'train/loss': 0.44337305426597595, 'train/grad_norm_pre_clip': 29.908485412597656, 'timing/data': 0.008390805684030056, 'timing/model': 0.9323954172432423, 'learning_rate/qwen_vl_interface': 1.4717194043869953e-06, 'learning_rate/action_model': 1.4717194043869953e-05, 'epoch': 2.29}) +[2026-06-18 01:16:36,209][starVLA.training.train_starvla][INFO] - Step 5571, Loss: {'train/loss': 0.29891452193260193, 'train/grad_norm_pre_clip': 21.915143966674805, 'timing/data': 0.016653590835630894, 'timing/model': 0.9541865978389978, 'learning_rate/qwen_vl_interface': 1.4704091000951982e-06, 'learning_rate/action_model': 1.4704091000951981e-05, 'epoch': 2.29}) +[2026-06-18 01:16:40,034][starVLA.training.train_starvla][INFO] - Step 5572, Loss: {'train/loss': 0.3138561248779297, 'train/grad_norm_pre_clip': 27.099613189697266, 'timing/data': 0.015159621834754944, 'timing/model': 0.9618242559954524, 'learning_rate/qwen_vl_interface': 1.4690995793166615e-06, 'learning_rate/action_model': 1.4690995793166616e-05, 'epoch': 2.29}) +[2026-06-18 01:16:44,016][starVLA.training.train_starvla][INFO] - Step 5573, Loss: {'train/loss': 0.4535198211669922, 'train/grad_norm_pre_clip': 34.862281799316406, 'timing/data': 0.008197884075343609, 'timing/model': 1.18624671921134, 'learning_rate/qwen_vl_interface': 1.4677908423228506e-06, 'learning_rate/action_model': 1.4677908423228507e-05, 'epoch': 2.29}) +[2026-06-18 01:16:47,633][starVLA.training.train_starvla][INFO] - Step 5574, Loss: {'train/loss': 0.3323729932308197, 'train/grad_norm_pre_clip': 30.516109466552734, 'timing/data': 0.007263964973390102, 'timing/model': 0.9334871377795935, 'learning_rate/qwen_vl_interface': 1.4664828893850677e-06, 'learning_rate/action_model': 1.4664828893850676e-05, 'epoch': 2.29}) +[2026-06-18 01:16:51,484][starVLA.training.train_starvla][INFO] - Step 5575, Loss: {'train/loss': 0.20647937059402466, 'train/grad_norm_pre_clip': 25.046863555908203, 'timing/data': 0.008868475444614887, 'timing/model': 0.9399755150079727, 'learning_rate/qwen_vl_interface': 1.4651757207744539e-06, 'learning_rate/action_model': 1.4651757207744537e-05, 'epoch': 2.29}) +[2026-06-18 01:16:55,245][starVLA.training.train_starvla][INFO] - Step 5576, Loss: {'train/loss': 0.2735026776790619, 'train/grad_norm_pre_clip': 16.07518196105957, 'timing/data': 0.01144009456038475, 'timing/model': 0.9268676917999983, 'learning_rate/qwen_vl_interface': 1.4638693367619837e-06, 'learning_rate/action_model': 1.4638693367619834e-05, 'epoch': 2.3}) +[2026-06-18 01:16:59,059][starVLA.training.train_starvla][INFO] - Step 5577, Loss: {'train/loss': 0.35459813475608826, 'train/grad_norm_pre_clip': 19.99862289428711, 'timing/data': 0.01269495114684105, 'timing/model': 0.936271850951016, 'learning_rate/qwen_vl_interface': 1.4625637376184768e-06, 'learning_rate/action_model': 1.4625637376184768e-05, 'epoch': 2.3}) +[2026-06-18 01:17:02,969][starVLA.training.train_starvla][INFO] - Step 5578, Loss: {'train/loss': 0.09851977974176407, 'train/grad_norm_pre_clip': 11.100074768066406, 'timing/data': 0.013361524790525436, 'timing/model': 0.9860287802293897, 'learning_rate/qwen_vl_interface': 1.4612589236145811e-06, 'learning_rate/action_model': 1.461258923614581e-05, 'epoch': 2.3}) +[2026-06-18 01:17:06,832][starVLA.training.train_starvla][INFO] - Step 5579, Loss: {'train/loss': 0.41234251856803894, 'train/grad_norm_pre_clip': 19.430418014526367, 'timing/data': 0.00814883317798376, 'timing/model': 1.1387410247698426, 'learning_rate/qwen_vl_interface': 1.4599548950207878e-06, 'learning_rate/action_model': 1.4599548950207878e-05, 'epoch': 2.3}) +[2026-06-18 01:17:10,760][starVLA.training.train_starvla][INFO] - Step 5580, Loss: {'train/loss': 0.29393327236175537, 'train/grad_norm_pre_clip': 22.790143966674805, 'timing/data': 0.01773146353662014, 'timing/model': 1.1719527589157224, 'learning_rate/qwen_vl_interface': 1.4586516521074214e-06, 'learning_rate/action_model': 1.4586516521074214e-05, 'epoch': 2.3}) +[2026-06-18 01:17:15,077][starVLA.training.train_starvla][INFO] - Step 5581, Loss: {'train/loss': 0.23816946148872375, 'train/grad_norm_pre_clip': 16.832111358642578, 'timing/data': 0.008875945582985878, 'timing/model': 0.9329896857962012, 'learning_rate/qwen_vl_interface': 1.457349195144649e-06, 'learning_rate/action_model': 1.457349195144649e-05, 'epoch': 2.3}) +[2026-06-18 01:17:22,118][starVLA.training.train_starvla][INFO] - Step 5582, Loss: {'train/loss': 0.14510664343833923, 'train/grad_norm_pre_clip': 15.712085723876953, 'timing/data': 0.011027351953089237, 'timing/model': 0.9898778358474374, 'learning_rate/qwen_vl_interface': 1.4560475244024702e-06, 'learning_rate/action_model': 1.4560475244024702e-05, 'epoch': 2.3}) +[2026-06-18 01:17:28,943][starVLA.training.train_starvla][INFO] - Step 5583, Loss: {'train/loss': 0.3703859746456146, 'train/grad_norm_pre_clip': 20.056289672851562, 'timing/data': 0.008098999038338661, 'timing/model': 0.9385706353932619, 'learning_rate/qwen_vl_interface': 1.4547466401507213e-06, 'learning_rate/action_model': 1.4547466401507211e-05, 'epoch': 2.3}) +[2026-06-18 01:17:32,873][starVLA.training.train_starvla][INFO] - Step 5584, Loss: {'train/loss': 0.24721220135688782, 'train/grad_norm_pre_clip': 23.87916374206543, 'timing/data': 0.009715382941067219, 'timing/model': 1.0064549120143056, 'learning_rate/qwen_vl_interface': 1.4534465426590763e-06, 'learning_rate/action_model': 1.4534465426590763e-05, 'epoch': 2.3}) +[2026-06-18 01:17:36,955][starVLA.training.train_starvla][INFO] - Step 5585, Loss: {'train/loss': 0.283452570438385, 'train/grad_norm_pre_clip': 21.186424255371094, 'timing/data': 0.009494557045400143, 'timing/model': 0.9765819199383259, 'learning_rate/qwen_vl_interface': 1.4521472321970496e-06, 'learning_rate/action_model': 1.4521472321970494e-05, 'epoch': 2.3}) +[2026-06-18 01:17:40,801][starVLA.training.train_starvla][INFO] - Step 5586, Loss: {'train/loss': 0.3236774504184723, 'train/grad_norm_pre_clip': 26.353256225585938, 'timing/data': 0.01273010578006506, 'timing/model': 1.1154169654473662, 'learning_rate/qwen_vl_interface': 1.450848709033988e-06, 'learning_rate/action_model': 1.4508487090339878e-05, 'epoch': 2.3}) +[2026-06-18 01:17:44,597][starVLA.training.train_starvla][INFO] - Step 5587, Loss: {'train/loss': 0.410366952419281, 'train/grad_norm_pre_clip': 428.2485656738281, 'timing/data': 0.011478802189230919, 'timing/model': 1.109310525469482, 'learning_rate/qwen_vl_interface': 1.4495509734390769e-06, 'learning_rate/action_model': 1.4495509734390769e-05, 'epoch': 2.3}) +[2026-06-18 01:17:48,533][starVLA.training.train_starvla][INFO] - Step 5588, Loss: {'train/loss': 0.024940188974142075, 'train/grad_norm_pre_clip': 3.266254186630249, 'timing/data': 0.009871136397123337, 'timing/model': 0.9264964945614338, 'learning_rate/qwen_vl_interface': 1.448254025681339e-06, 'learning_rate/action_model': 1.448254025681339e-05, 'epoch': 2.3}) +[2026-06-18 01:17:52,434][starVLA.training.train_starvla][INFO] - Step 5589, Loss: {'train/loss': 0.49423810839653015, 'train/grad_norm_pre_clip': 29.720611572265625, 'timing/data': 0.011056991294026375, 'timing/model': 0.9495212752372026, 'learning_rate/qwen_vl_interface': 1.4469578660296327e-06, 'learning_rate/action_model': 1.4469578660296326e-05, 'epoch': 2.3}) +[2026-06-18 01:17:56,261][starVLA.training.train_starvla][INFO] - Step 5590, Loss: {'train/loss': 0.16007809340953827, 'train/grad_norm_pre_clip': 13.912927627563477, 'timing/data': 0.010163231752812862, 'timing/model': 0.9320831559598446, 'learning_rate/qwen_vl_interface': 1.4456624947526524e-06, 'learning_rate/action_model': 1.4456624947526524e-05, 'epoch': 2.3}) +[2026-06-18 01:18:00,134][starVLA.training.train_starvla][INFO] - Step 5591, Loss: {'train/loss': 0.11612942814826965, 'train/grad_norm_pre_clip': 14.698037147521973, 'timing/data': 0.009596363641321659, 'timing/model': 0.9468832714483142, 'learning_rate/qwen_vl_interface': 1.4443679121189306e-06, 'learning_rate/action_model': 1.4443679121189305e-05, 'epoch': 2.3}) +[2026-06-18 01:18:04,259][starVLA.training.train_starvla][INFO] - Step 5592, Loss: {'train/loss': 0.49195635318756104, 'train/grad_norm_pre_clip': 25.04534149169922, 'timing/data': 0.006880726665258408, 'timing/model': 1.1845924500375986, 'learning_rate/qwen_vl_interface': 1.4430741183968361e-06, 'learning_rate/action_model': 1.443074118396836e-05, 'epoch': 2.3}) +[2026-06-18 01:18:08,138][starVLA.training.train_starvla][INFO] - Step 5593, Loss: {'train/loss': 0.2438686490058899, 'train/grad_norm_pre_clip': 14.90936279296875, 'timing/data': 0.008879770524799824, 'timing/model': 1.1194676561281085, 'learning_rate/qwen_vl_interface': 1.441781113854573e-06, 'learning_rate/action_model': 1.4417811138545728e-05, 'epoch': 2.3}) +[2026-06-18 01:18:11,994][starVLA.training.train_starvla][INFO] - Step 5594, Loss: {'train/loss': 0.07738712430000305, 'train/grad_norm_pre_clip': 27.27702522277832, 'timing/data': 0.010309417732059956, 'timing/model': 1.0403458159416914, 'learning_rate/qwen_vl_interface': 1.4404888987601828e-06, 'learning_rate/action_model': 1.4404888987601828e-05, 'epoch': 2.3}) +[2026-06-18 01:18:16,845][starVLA.training.train_starvla][INFO] - Step 5595, Loss: {'train/loss': 0.0757758766412735, 'train/grad_norm_pre_clip': 9.483055114746094, 'timing/data': 0.009586993604898453, 'timing/model': 1.0786337684839964, 'learning_rate/qwen_vl_interface': 1.4391974733815417e-06, 'learning_rate/action_model': 1.4391974733815416e-05, 'epoch': 2.3}) +[2026-06-18 01:18:21,135][starVLA.training.train_starvla][INFO] - Step 5596, Loss: {'train/loss': 0.26266923546791077, 'train/grad_norm_pre_clip': 23.11581039428711, 'timing/data': 0.009614530950784683, 'timing/model': 0.9477409841492772, 'learning_rate/qwen_vl_interface': 1.437906837986368e-06, 'learning_rate/action_model': 1.4379068379863677e-05, 'epoch': 2.3}) +[2026-06-18 01:18:25,001][starVLA.training.train_starvla][INFO] - Step 5597, Loss: {'train/loss': 0.6887539029121399, 'train/grad_norm_pre_clip': 25.117868423461914, 'timing/data': 0.016735466197133064, 'timing/model': 0.9489536033943295, 'learning_rate/qwen_vl_interface': 1.436616992842207e-06, 'learning_rate/action_model': 1.4366169928422068e-05, 'epoch': 2.3}) +[2026-06-18 01:18:28,874][starVLA.training.train_starvla][INFO] - Step 5598, Loss: {'train/loss': 0.0788462907075882, 'train/grad_norm_pre_clip': 15.33568286895752, 'timing/data': 0.009829696267843246, 'timing/model': 0.9340826254338026, 'learning_rate/qwen_vl_interface': 1.4353279382164459e-06, 'learning_rate/action_model': 1.4353279382164458e-05, 'epoch': 2.3}) +[2026-06-18 01:18:35,166][starVLA.training.train_starvla][INFO] - Step 5599, Loss: {'train/loss': 0.2564658224582672, 'train/grad_norm_pre_clip': 15.988140106201172, 'timing/data': 0.01436552032828331, 'timing/model': 0.9855209114030004, 'learning_rate/qwen_vl_interface': 1.4340396743763084e-06, 'learning_rate/action_model': 1.4340396743763085e-05, 'epoch': 2.3}) +[2026-06-18 01:20:56,515][starVLA.training.train_starvla][INFO] - Step 5600, Loss: {'train/loss': 0.22494538128376007, 'train/grad_norm_pre_clip': 23.533935546875, 'eval/loss': 0.41420748829841614, 'eval/latency_6/loss': 0.41420748829841614, 'eval/demon_fire_precision': 0.9556119213696893, 'eval/demon_fire_recall': 0.9741435035552682, 'eval/demon_fire_f1': 0.9647887323943662, 'eval/demon_fire_rate_model': 0.8775737340011129, 'eval/demon_fire_rate_teacher': 0.860879243183083, 'eval/demon_left_precision': 0.8919491525423728, 'eval/demon_left_recall': 0.7385964912280701, 'eval/demon_left_f1': 0.8080614203454894, 'eval/demon_left_rate_model': 0.13132999443516974, 'eval/demon_left_rate_teacher': 0.15859766277128548, 'eval/demon_right_precision': 0.8539130434782609, 'eval/demon_right_recall': 0.7542242703533026, 'eval/demon_right_f1': 0.800978792822186, 'eval/demon_right_rate_model': 0.15998887033945464, 'eval/demon_right_rate_teacher': 0.18113522537562604, 'eval/demon_move_f1': 0.8045201065838377, 'eval/demon_cc_f1': 0.9103815408331228, 'eval/latency_6/demon_fire_precision': 0.9556119213696893, 'eval/latency_6/demon_fire_recall': 0.9741435035552682, 'eval/latency_6/demon_fire_f1': 0.9647887323943662, 'eval/latency_6/demon_fire_rate_model': 0.8775737340011129, 'eval/latency_6/demon_fire_rate_teacher': 0.860879243183083, 'eval/latency_6/demon_left_precision': 0.8919491525423728, 'eval/latency_6/demon_left_recall': 0.7385964912280701, 'eval/latency_6/demon_left_f1': 0.8080614203454894, 'eval/latency_6/demon_left_rate_model': 0.13132999443516974, 'eval/latency_6/demon_left_rate_teacher': 0.15859766277128548, 'eval/latency_6/demon_right_precision': 0.8539130434782609, 'eval/latency_6/demon_right_recall': 0.7542242703533026, 'eval/latency_6/demon_right_f1': 0.800978792822186, 'eval/latency_6/demon_right_rate_model': 0.15998887033945464, 'eval/latency_6/demon_right_rate_teacher': 0.18113522537562604, 'eval/latency_6/demon_move_f1': 0.8045201065838377, 'eval/latency_6/demon_cc_f1': 0.9103815408331228, 'timing/data': 0.01329725794494152, 'timing/model': 1.1279978323727846, 'learning_rate/qwen_vl_interface': 1.4327522015888535e-06, 'learning_rate/action_model': 1.4327522015888534e-05, 'epoch': 2.31}) +[2026-06-18 01:20:56,518][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_5600_state +[2026-06-18 01:20:56,523][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 01:21:03,479][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_5600_state/model.safetensors +[2026-06-18 01:21:19,449][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_5600_state/optimizer.bin +[2026-06-18 01:21:19,451][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_5600_state/sampler.bin +[2026-06-18 01:21:19,452][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_5600_state/sampler_1.bin +[2026-06-18 01:21:19,453][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_5600_state/random_states_0.pkl +[2026-06-18 01:21:19,457][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 5600 +[2026-06-18 01:21:19,458][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 01:21:19,468][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 01:21:23,372][starVLA.training.train_starvla][INFO] - Step 5601, Loss: {'train/loss': 0.1686324179172516, 'train/grad_norm_pre_clip': 12.8673095703125, 'timing/data': 0.010028202086687088, 'timing/model': 0.9388925619423389, 'learning_rate/qwen_vl_interface': 1.4314655201209756e-06, 'learning_rate/action_model': 1.4314655201209757e-05, 'epoch': 2.31}) +[2026-06-18 01:21:27,264][starVLA.training.train_starvla][INFO] - Step 5602, Loss: {'train/loss': 0.11533485352993011, 'train/grad_norm_pre_clip': 16.62322235107422, 'timing/data': 0.011375529691576958, 'timing/model': 1.1766295908018947, 'learning_rate/qwen_vl_interface': 1.4301796302394019e-06, 'learning_rate/action_model': 1.4301796302394018e-05, 'epoch': 2.31}) +[2026-06-18 01:21:31,103][starVLA.training.train_starvla][INFO] - Step 5603, Loss: {'train/loss': 0.4247087240219116, 'train/grad_norm_pre_clip': 34.14077377319336, 'timing/data': 0.009787904098629951, 'timing/model': 1.1226474167779088, 'learning_rate/qwen_vl_interface': 1.4288945322107025e-06, 'learning_rate/action_model': 1.4288945322107026e-05, 'epoch': 2.31}) +[2026-06-18 01:21:34,759][starVLA.training.train_starvla][INFO] - Step 5604, Loss: {'train/loss': 0.09584908187389374, 'train/grad_norm_pre_clip': 14.121007919311523, 'timing/data': 0.012053125537931919, 'timing/model': 0.9428657852113247, 'learning_rate/qwen_vl_interface': 1.4276102263012781e-06, 'learning_rate/action_model': 1.427610226301278e-05, 'epoch': 2.31}) +[2026-06-18 01:21:38,657][starVLA.training.train_starvla][INFO] - Step 5605, Loss: {'train/loss': 0.056868817657232285, 'train/grad_norm_pre_clip': 7.782130241394043, 'timing/data': 0.01065114140510559, 'timing/model': 0.9397855969145894, 'learning_rate/qwen_vl_interface': 1.426326712777367e-06, 'learning_rate/action_model': 1.4263267127773669e-05, 'epoch': 2.31}) +[2026-06-18 01:21:42,500][starVLA.training.train_starvla][INFO] - Step 5606, Loss: {'train/loss': 0.16791634261608124, 'train/grad_norm_pre_clip': 16.726093292236328, 'timing/data': 0.009351825341582298, 'timing/model': 0.9412687076255679, 'learning_rate/qwen_vl_interface': 1.425043991905043e-06, 'learning_rate/action_model': 1.425043991905043e-05, 'epoch': 2.31}) +[2026-06-18 01:21:46,240][starVLA.training.train_starvla][INFO] - Step 5607, Loss: {'train/loss': 0.16752813756465912, 'train/grad_norm_pre_clip': 16.61635971069336, 'timing/data': 0.008771066553890705, 'timing/model': 0.9439497292041779, 'learning_rate/qwen_vl_interface': 1.4237620639502149e-06, 'learning_rate/action_model': 1.4237620639502147e-05, 'epoch': 2.31}) +[2026-06-18 01:21:50,063][starVLA.training.train_starvla][INFO] - Step 5608, Loss: {'train/loss': 0.3238469660282135, 'train/grad_norm_pre_clip': 21.881874084472656, 'timing/data': 0.00913184229284525, 'timing/model': 0.9362269528210163, 'learning_rate/qwen_vl_interface': 1.4224809291786281e-06, 'learning_rate/action_model': 1.4224809291786282e-05, 'epoch': 2.31}) +[2026-06-18 01:21:53,907][starVLA.training.train_starvla][INFO] - Step 5609, Loss: {'train/loss': 0.4144216775894165, 'train/grad_norm_pre_clip': 26.991228103637695, 'timing/data': 0.017677778378129005, 'timing/model': 1.1231463132426143, 'learning_rate/qwen_vl_interface': 1.4212005878558622e-06, 'learning_rate/action_model': 1.4212005878558622e-05, 'epoch': 2.31}) +[2026-06-18 01:21:57,565][starVLA.training.train_starvla][INFO] - Step 5610, Loss: {'train/loss': 0.2514757513999939, 'train/grad_norm_pre_clip': 19.49363136291504, 'timing/data': 0.009277512319386005, 'timing/model': 0.9374790601432323, 'learning_rate/qwen_vl_interface': 1.4199210402473368e-06, 'learning_rate/action_model': 1.4199210402473367e-05, 'epoch': 2.31}) +[2026-06-18 01:22:01,461][starVLA.training.train_starvla][INFO] - Step 5611, Loss: {'train/loss': 0.19096815586090088, 'train/grad_norm_pre_clip': 21.0294132232666, 'timing/data': 0.007752818055450916, 'timing/model': 0.9637515209615231, 'learning_rate/qwen_vl_interface': 1.4186422866183e-06, 'learning_rate/action_model': 1.4186422866182999e-05, 'epoch': 2.31}) +[2026-06-18 01:22:05,333][starVLA.training.train_starvla][INFO] - Step 5612, Loss: {'train/loss': 0.2903067469596863, 'train/grad_norm_pre_clip': 27.66799545288086, 'timing/data': 0.008654084987938404, 'timing/model': 0.94366639200598, 'learning_rate/qwen_vl_interface': 1.41736432723384e-06, 'learning_rate/action_model': 1.41736432723384e-05, 'epoch': 2.31}) +[2026-06-18 01:22:09,205][starVLA.training.train_starvla][INFO] - Step 5613, Loss: {'train/loss': 0.30786582827568054, 'train/grad_norm_pre_clip': 24.743349075317383, 'timing/data': 0.012351347133517265, 'timing/model': 0.9470346681773663, 'learning_rate/qwen_vl_interface': 1.4160871623588774e-06, 'learning_rate/action_model': 1.4160871623588772e-05, 'epoch': 2.31}) +[2026-06-18 01:22:13,026][starVLA.training.train_starvla][INFO] - Step 5614, Loss: {'train/loss': 0.3815520107746124, 'train/grad_norm_pre_clip': 29.16483497619629, 'timing/data': 0.012909621931612492, 'timing/model': 0.953562811948359, 'learning_rate/qwen_vl_interface': 1.414810792258173e-06, 'learning_rate/action_model': 1.414810792258173e-05, 'epoch': 2.31}) +[2026-06-18 01:22:16,834][starVLA.training.train_starvla][INFO] - Step 5615, Loss: {'train/loss': 0.2166035920381546, 'train/grad_norm_pre_clip': 28.77290153503418, 'timing/data': 0.019955061376094818, 'timing/model': 1.098134153522551, 'learning_rate/qwen_vl_interface': 1.4135352171963186e-06, 'learning_rate/action_model': 1.4135352171963185e-05, 'epoch': 2.31}) +[2026-06-18 01:22:20,721][starVLA.training.train_starvla][INFO] - Step 5616, Loss: {'train/loss': 0.38418278098106384, 'train/grad_norm_pre_clip': 25.525392532348633, 'timing/data': 0.010492905043065548, 'timing/model': 1.171145892702043, 'learning_rate/qwen_vl_interface': 1.412260437437741e-06, 'learning_rate/action_model': 1.412260437437741e-05, 'epoch': 2.31}) +[2026-06-18 01:22:24,345][starVLA.training.train_starvla][INFO] - Step 5617, Loss: {'train/loss': 0.3845272660255432, 'train/grad_norm_pre_clip': 27.818492889404297, 'timing/data': 0.009154237806797028, 'timing/model': 0.9482760103419423, 'learning_rate/qwen_vl_interface': 1.4109864532467051e-06, 'learning_rate/action_model': 1.4109864532467052e-05, 'epoch': 2.31}) +[2026-06-18 01:22:28,223][starVLA.training.train_starvla][INFO] - Step 5618, Loss: {'train/loss': 0.3002318739891052, 'train/grad_norm_pre_clip': 26.373245239257812, 'timing/data': 0.011964653618633747, 'timing/model': 0.957676843740046, 'learning_rate/qwen_vl_interface': 1.4097132648873082e-06, 'learning_rate/action_model': 1.4097132648873083e-05, 'epoch': 2.31}) +[2026-06-18 01:22:34,479][starVLA.training.train_starvla][INFO] - Step 5619, Loss: {'train/loss': 0.22247706353664398, 'train/grad_norm_pre_clip': 31.98872947692871, 'timing/data': 0.00699504092335701, 'timing/model': 0.9350659558549523, 'learning_rate/qwen_vl_interface': 1.4084408726234838e-06, 'learning_rate/action_model': 1.4084408726234838e-05, 'epoch': 2.31}) +[2026-06-18 01:22:38,315][starVLA.training.train_starvla][INFO] - Step 5620, Loss: {'train/loss': 0.5703288912773132, 'train/grad_norm_pre_clip': 26.9278507232666, 'timing/data': 0.01074051670730114, 'timing/model': 0.9438070701435208, 'learning_rate/qwen_vl_interface': 1.407169276719e-06, 'learning_rate/action_model': 1.4071692767190001e-05, 'epoch': 2.31}) +[2026-06-18 01:22:42,212][starVLA.training.train_starvla][INFO] - Step 5621, Loss: {'train/loss': 0.24173744022846222, 'train/grad_norm_pre_clip': 26.927413940429688, 'timing/data': 0.011905038729310036, 'timing/model': 0.9626100370660424, 'learning_rate/qwen_vl_interface': 1.4058984774374606e-06, 'learning_rate/action_model': 1.4058984774374607e-05, 'epoch': 2.31}) +[2026-06-18 01:22:46,285][starVLA.training.train_starvla][INFO] - Step 5622, Loss: {'train/loss': 0.2842715382575989, 'train/grad_norm_pre_clip': 26.291080474853516, 'timing/data': 0.013650999404489994, 'timing/model': 1.2677919091656804, 'learning_rate/qwen_vl_interface': 1.404628475042303e-06, 'learning_rate/action_model': 1.404628475042303e-05, 'epoch': 2.31}) +[2026-06-18 01:22:50,138][starVLA.training.train_starvla][INFO] - Step 5623, Loss: {'train/loss': 0.3591379225254059, 'train/grad_norm_pre_clip': 19.511825561523438, 'timing/data': 0.0085517643019557, 'timing/model': 1.1248311130329967, 'learning_rate/qwen_vl_interface': 1.4033592697967998e-06, 'learning_rate/action_model': 1.4033592697967998e-05, 'epoch': 2.31}) +[2026-06-18 01:22:53,871][starVLA.training.train_starvla][INFO] - Step 5624, Loss: {'train/loss': 0.7312245965003967, 'train/grad_norm_pre_clip': 21.105562210083008, 'timing/data': 0.011816414073109627, 'timing/model': 0.9270984530448914, 'learning_rate/qwen_vl_interface': 1.4020908619640602e-06, 'learning_rate/action_model': 1.4020908619640602e-05, 'epoch': 2.32}) +[2026-06-18 01:22:57,737][starVLA.training.train_starvla][INFO] - Step 5625, Loss: {'train/loss': 0.30167344212532043, 'train/grad_norm_pre_clip': 18.22806167602539, 'timing/data': 0.009407592006027699, 'timing/model': 0.940564320422709, 'learning_rate/qwen_vl_interface': 1.4008232518070248e-06, 'learning_rate/action_model': 1.4008232518070247e-05, 'epoch': 2.32}) +[2026-06-18 01:23:01,580][starVLA.training.train_starvla][INFO] - Step 5626, Loss: {'train/loss': 0.16926556825637817, 'train/grad_norm_pre_clip': 19.93672752380371, 'timing/data': 0.014858809299767017, 'timing/model': 0.9375526998192072, 'learning_rate/qwen_vl_interface': 1.3995564395884712e-06, 'learning_rate/action_model': 1.399556439588471e-05, 'epoch': 2.32}) +[2026-06-18 01:23:05,464][starVLA.training.train_starvla][INFO] - Step 5627, Loss: {'train/loss': 0.27283692359924316, 'train/grad_norm_pre_clip': 22.230018615722656, 'timing/data': 0.011160926893353462, 'timing/model': 0.9611728265881538, 'learning_rate/qwen_vl_interface': 1.3982904255710094e-06, 'learning_rate/action_model': 1.3982904255710094e-05, 'epoch': 2.32}) +[2026-06-18 01:23:09,334][starVLA.training.train_starvla][INFO] - Step 5628, Loss: {'train/loss': 0.5501684546470642, 'train/grad_norm_pre_clip': 23.30158805847168, 'timing/data': 0.011193277314305305, 'timing/model': 0.9481494156643748, 'learning_rate/qwen_vl_interface': 1.3970252100170892e-06, 'learning_rate/action_model': 1.3970252100170892e-05, 'epoch': 2.32}) +[2026-06-18 01:23:13,166][starVLA.training.train_starvla][INFO] - Step 5629, Loss: {'train/loss': 0.1490294635295868, 'train/grad_norm_pre_clip': 16.052656173706055, 'timing/data': 0.010066607035696507, 'timing/model': 1.1121151773259044, 'learning_rate/qwen_vl_interface': 1.3957607931889885e-06, 'learning_rate/action_model': 1.3957607931889884e-05, 'epoch': 2.32}) +[2026-06-18 01:23:16,934][starVLA.training.train_starvla][INFO] - Step 5630, Loss: {'train/loss': 0.27521923184394836, 'train/grad_norm_pre_clip': 20.90594482421875, 'timing/data': 0.008730479516088963, 'timing/model': 0.9352658428251743, 'learning_rate/qwen_vl_interface': 1.3944971753488238e-06, 'learning_rate/action_model': 1.3944971753488236e-05, 'epoch': 2.32}) +[2026-06-18 01:23:20,733][starVLA.training.train_starvla][INFO] - Step 5631, Loss: {'train/loss': 0.12491972744464874, 'train/grad_norm_pre_clip': 14.759881019592285, 'timing/data': 0.009517745114862919, 'timing/model': 0.944972506724298, 'learning_rate/qwen_vl_interface': 1.3932343567585408e-06, 'learning_rate/action_model': 1.3932343567585407e-05, 'epoch': 2.32}) +[2026-06-18 01:23:24,529][starVLA.training.train_starvla][INFO] - Step 5632, Loss: {'train/loss': 0.3389407694339752, 'train/grad_norm_pre_clip': 27.06069564819336, 'timing/data': 0.012960156425833702, 'timing/model': 0.9399624625220895, 'learning_rate/qwen_vl_interface': 1.3919723376799268e-06, 'learning_rate/action_model': 1.3919723376799269e-05, 'epoch': 2.32}) +[2026-06-18 01:23:28,394][starVLA.training.train_starvla][INFO] - Step 5633, Loss: {'train/loss': 0.4988402724266052, 'train/grad_norm_pre_clip': 28.361454010009766, 'timing/data': 0.007226482965052128, 'timing/model': 0.9384046914055943, 'learning_rate/qwen_vl_interface': 1.3907111183745992e-06, 'learning_rate/action_model': 1.3907111183745991e-05, 'epoch': 2.32}) +[2026-06-18 01:23:32,272][starVLA.training.train_starvla][INFO] - Step 5634, Loss: {'train/loss': 0.26836997270584106, 'train/grad_norm_pre_clip': 24.56356430053711, 'timing/data': 0.012378123588860035, 'timing/model': 0.98764942958951, 'learning_rate/qwen_vl_interface': 1.3894506991040089e-06, 'learning_rate/action_model': 1.3894506991040087e-05, 'epoch': 2.32}) +[2026-06-18 01:23:36,141][starVLA.training.train_starvla][INFO] - Step 5635, Loss: {'train/loss': 0.24170075356960297, 'train/grad_norm_pre_clip': 21.51143455505371, 'timing/data': 0.008857334963977337, 'timing/model': 0.9362498139962554, 'learning_rate/qwen_vl_interface': 1.3881910801294433e-06, 'learning_rate/action_model': 1.3881910801294431e-05, 'epoch': 2.32}) +[2026-06-18 01:23:40,165][starVLA.training.train_starvla][INFO] - Step 5636, Loss: {'train/loss': 0.07662368565797806, 'train/grad_norm_pre_clip': 22.609975814819336, 'timing/data': 0.010573682375252247, 'timing/model': 1.3046745620667934, 'learning_rate/qwen_vl_interface': 1.3869322617120223e-06, 'learning_rate/action_model': 1.3869322617120222e-05, 'epoch': 2.32}) +[2026-06-18 01:23:43,874][starVLA.training.train_starvla][INFO] - Step 5637, Loss: {'train/loss': 0.1883489489555359, 'train/grad_norm_pre_clip': 24.688678741455078, 'timing/data': 0.0089921485632658, 'timing/model': 0.9694695631042123, 'learning_rate/qwen_vl_interface': 1.3856742441127e-06, 'learning_rate/action_model': 1.3856742441126999e-05, 'epoch': 2.32}) +[2026-06-18 01:23:48,053][starVLA.training.train_starvla][INFO] - Step 5638, Loss: {'train/loss': 0.39958328008651733, 'train/grad_norm_pre_clip': 17.133766174316406, 'timing/data': 0.013370459899306297, 'timing/model': 0.9319829260930419, 'learning_rate/qwen_vl_interface': 1.384417027592264e-06, 'learning_rate/action_model': 1.3844170275922638e-05, 'epoch': 2.32}) +[2026-06-18 01:23:51,881][starVLA.training.train_starvla][INFO] - Step 5639, Loss: {'train/loss': 0.12983277440071106, 'train/grad_norm_pre_clip': 14.079122543334961, 'timing/data': 0.012551503255963326, 'timing/model': 0.9423699788749218, 'learning_rate/qwen_vl_interface': 1.3831606124113397e-06, 'learning_rate/action_model': 1.3831606124113398e-05, 'epoch': 2.32}) +[2026-06-18 01:23:57,268][starVLA.training.train_starvla][INFO] - Step 5640, Loss: {'train/loss': 0.2933224141597748, 'train/grad_norm_pre_clip': 21.26166534423828, 'timing/data': 0.00856383703649044, 'timing/model': 0.9365039383992553, 'learning_rate/qwen_vl_interface': 1.3819049988303809e-06, 'learning_rate/action_model': 1.381904998830381e-05, 'epoch': 2.32}) +[2026-06-18 01:24:02,594][starVLA.training.train_starvla][INFO] - Step 5641, Loss: {'train/loss': 0.20948903262615204, 'train/grad_norm_pre_clip': 21.110431671142578, 'timing/data': 0.007743672467768192, 'timing/model': 0.9662215430289507, 'learning_rate/qwen_vl_interface': 1.3806501871096774e-06, 'learning_rate/action_model': 1.3806501871096774e-05, 'epoch': 2.32}) +[2026-06-18 01:24:07,931][starVLA.training.train_starvla][INFO] - Step 5642, Loss: {'train/loss': 0.3657858669757843, 'train/grad_norm_pre_clip': 27.36187171936035, 'timing/data': 0.008514279499650002, 'timing/model': 1.180314702913165, 'learning_rate/qwen_vl_interface': 1.3793961775093528e-06, 'learning_rate/action_model': 1.3793961775093528e-05, 'epoch': 2.32}) +[2026-06-18 01:24:11,854][starVLA.training.train_starvla][INFO] - Step 5643, Loss: {'train/loss': 0.23427706956863403, 'train/grad_norm_pre_clip': 21.736698150634766, 'timing/data': 0.011853097938001156, 'timing/model': 1.197089919820428, 'learning_rate/qwen_vl_interface': 1.378142970289367e-06, 'learning_rate/action_model': 1.378142970289367e-05, 'epoch': 2.32}) +[2026-06-18 01:24:15,537][starVLA.training.train_starvla][INFO] - Step 5644, Loss: {'train/loss': 0.13452386856079102, 'train/grad_norm_pre_clip': 11.462242126464844, 'timing/data': 0.0083607854321599, 'timing/model': 0.9329532533884048, 'learning_rate/qwen_vl_interface': 1.3768905657095109e-06, 'learning_rate/action_model': 1.3768905657095107e-05, 'epoch': 2.32}) +[2026-06-18 01:24:19,334][starVLA.training.train_starvla][INFO] - Step 5645, Loss: {'train/loss': 0.1962907761335373, 'train/grad_norm_pre_clip': 23.393117904663086, 'timing/data': 0.01038888469338417, 'timing/model': 0.9389151241630316, 'learning_rate/qwen_vl_interface': 1.3756389640294066e-06, 'learning_rate/action_model': 1.3756389640294064e-05, 'epoch': 2.32}) +[2026-06-18 01:24:23,420][starVLA.training.train_starvla][INFO] - Step 5646, Loss: {'train/loss': 0.51509028673172, 'train/grad_norm_pre_clip': 30.831111907958984, 'timing/data': 0.009002885781228542, 'timing/model': 1.0356561904773116, 'learning_rate/qwen_vl_interface': 1.3743881655085131e-06, 'learning_rate/action_model': 1.3743881655085131e-05, 'epoch': 2.32}) +[2026-06-18 01:24:27,395][starVLA.training.train_starvla][INFO] - Step 5647, Loss: {'train/loss': 0.14800624549388885, 'train/grad_norm_pre_clip': 13.229291915893555, 'timing/data': 0.008782606571912766, 'timing/model': 1.0667976411059499, 'learning_rate/qwen_vl_interface': 1.3731381704061248e-06, 'learning_rate/action_model': 1.3731381704061246e-05, 'epoch': 2.32}) +[2026-06-18 01:24:31,345][starVLA.training.train_starvla][INFO] - Step 5648, Loss: {'train/loss': 0.07785008102655411, 'train/grad_norm_pre_clip': 9.59192180633545, 'timing/data': 0.009624977596104145, 'timing/model': 0.9351541958749294, 'learning_rate/qwen_vl_interface': 1.3718889789813647e-06, 'learning_rate/action_model': 1.3718889789813646e-05, 'epoch': 2.32}) +[2026-06-18 01:24:35,264][starVLA.training.train_starvla][INFO] - Step 5649, Loss: {'train/loss': 0.27559345960617065, 'train/grad_norm_pre_clip': 21.355533599853516, 'timing/data': 0.009922168217599392, 'timing/model': 1.1350997015833855, 'learning_rate/qwen_vl_interface': 1.370640591493193e-06, 'learning_rate/action_model': 1.370640591493193e-05, 'epoch': 2.33}) +[2026-06-18 01:24:40,324][starVLA.training.train_starvla][INFO] - Step 5650, Loss: {'train/loss': 0.4237297475337982, 'train/grad_norm_pre_clip': 24.083633422851562, 'timing/data': 0.01223796233534813, 'timing/model': 2.340857825241983, 'learning_rate/qwen_vl_interface': 1.3693930082004006e-06, 'learning_rate/action_model': 1.3693930082004006e-05, 'epoch': 2.33}) +[2026-06-18 01:24:44,047][starVLA.training.train_starvla][INFO] - Step 5651, Loss: {'train/loss': 0.29246285557746887, 'train/grad_norm_pre_clip': 20.921064376831055, 'timing/data': 0.00965805258601904, 'timing/model': 0.9370816387236118, 'learning_rate/qwen_vl_interface': 1.3681462293616131e-06, 'learning_rate/action_model': 1.3681462293616131e-05, 'epoch': 2.33}) +[2026-06-18 01:24:49,264][starVLA.training.train_starvla][INFO] - Step 5652, Loss: {'train/loss': 0.0900476798415184, 'train/grad_norm_pre_clip': 10.4496488571167, 'timing/data': 0.009033223614096642, 'timing/model': 0.9338851543143392, 'learning_rate/qwen_vl_interface': 1.3669002552352898e-06, 'learning_rate/action_model': 1.3669002552352897e-05, 'epoch': 2.33}) +[2026-06-18 01:24:53,277][starVLA.training.train_starvla][INFO] - Step 5653, Loss: {'train/loss': 0.32040682435035706, 'train/grad_norm_pre_clip': 24.390682220458984, 'timing/data': 0.009023199789226055, 'timing/model': 0.9371018288657069, 'learning_rate/qwen_vl_interface': 1.3656550860797223e-06, 'learning_rate/action_model': 1.3656550860797223e-05, 'epoch': 2.33}) +[2026-06-18 01:24:57,290][starVLA.training.train_starvla][INFO] - Step 5654, Loss: {'train/loss': 0.3695535957813263, 'train/grad_norm_pre_clip': 26.542152404785156, 'timing/data': 0.009761975146830082, 'timing/model': 0.9228151589632034, 'learning_rate/qwen_vl_interface': 1.3644107221530344e-06, 'learning_rate/action_model': 1.3644107221530345e-05, 'epoch': 2.33}) +[2026-06-18 01:25:01,365][starVLA.training.train_starvla][INFO] - Step 5655, Loss: {'train/loss': 0.2143472284078598, 'train/grad_norm_pre_clip': 18.616127014160156, 'timing/data': 0.010691261850297451, 'timing/model': 0.9858423667028546, 'learning_rate/qwen_vl_interface': 1.3631671637131856e-06, 'learning_rate/action_model': 1.3631671637131855e-05, 'epoch': 2.33}) +[2026-06-18 01:25:05,378][starVLA.training.train_starvla][INFO] - Step 5656, Loss: {'train/loss': 0.35464945435523987, 'train/grad_norm_pre_clip': 24.56283950805664, 'timing/data': 0.00889295432716608, 'timing/model': 1.1417333940044045, 'learning_rate/qwen_vl_interface': 1.3619244110179655e-06, 'learning_rate/action_model': 1.3619244110179654e-05, 'epoch': 2.33}) +[2026-06-18 01:25:09,044][starVLA.training.train_starvla][INFO] - Step 5657, Loss: {'train/loss': 0.20990267395973206, 'train/grad_norm_pre_clip': 15.835752487182617, 'timing/data': 0.01429845206439495, 'timing/model': 0.9485027370974422, 'learning_rate/qwen_vl_interface': 1.360682464324998e-06, 'learning_rate/action_model': 1.360682464324998e-05, 'epoch': 2.33}) +[2026-06-18 01:25:12,899][starVLA.training.train_starvla][INFO] - Step 5658, Loss: {'train/loss': 0.3145499527454376, 'train/grad_norm_pre_clip': 28.274782180786133, 'timing/data': 0.010128415189683437, 'timing/model': 0.9485607789829373, 'learning_rate/qwen_vl_interface': 1.3594413238917425e-06, 'learning_rate/action_model': 1.3594413238917425e-05, 'epoch': 2.33}) +[2026-06-18 01:25:16,674][starVLA.training.train_starvla][INFO] - Step 5659, Loss: {'train/loss': 0.25465062260627747, 'train/grad_norm_pre_clip': 24.163169860839844, 'timing/data': 0.0194346122443676, 'timing/model': 0.96302595641464, 'learning_rate/qwen_vl_interface': 1.3582009899754864e-06, 'learning_rate/action_model': 1.3582009899754864e-05, 'epoch': 2.33}) +[2026-06-18 01:25:20,701][starVLA.training.train_starvla][INFO] - Step 5660, Loss: {'train/loss': 0.116446353495121, 'train/grad_norm_pre_clip': 15.415627479553223, 'timing/data': 0.009129846468567848, 'timing/model': 1.0235712127760053, 'learning_rate/qwen_vl_interface': 1.3569614628333515e-06, 'learning_rate/action_model': 1.3569614628333516e-05, 'epoch': 2.33}) +[2026-06-18 01:25:24,507][starVLA.training.train_starvla][INFO] - Step 5661, Loss: {'train/loss': 0.28821009397506714, 'train/grad_norm_pre_clip': 85.61030578613281, 'timing/data': 0.01020835805684328, 'timing/model': 0.9440800743177533, 'learning_rate/qwen_vl_interface': 1.3557227427222954e-06, 'learning_rate/action_model': 1.3557227427222954e-05, 'epoch': 2.33}) +[2026-06-18 01:25:28,285][starVLA.training.train_starvla][INFO] - Step 5662, Loss: {'train/loss': 0.9068854451179504, 'train/grad_norm_pre_clip': 40.27109909057617, 'timing/data': 0.01150301843881607, 'timing/model': 1.0985313821583986, 'learning_rate/qwen_vl_interface': 1.3544848298991056e-06, 'learning_rate/action_model': 1.3544848298991056e-05, 'epoch': 2.33}) +[2026-06-18 01:25:32,055][starVLA.training.train_starvla][INFO] - Step 5663, Loss: {'train/loss': 0.16830618679523468, 'train/grad_norm_pre_clip': 18.81759262084961, 'timing/data': 0.007924751378595829, 'timing/model': 1.1015454344451427, 'learning_rate/qwen_vl_interface': 1.3532477246204037e-06, 'learning_rate/action_model': 1.3532477246204036e-05, 'epoch': 2.33}) +[2026-06-18 01:25:35,700][starVLA.training.train_starvla][INFO] - Step 5664, Loss: {'train/loss': 0.06201707944273949, 'train/grad_norm_pre_clip': 10.643728256225586, 'timing/data': 0.013188915327191353, 'timing/model': 0.9497711816802621, 'learning_rate/qwen_vl_interface': 1.3520114271426382e-06, 'learning_rate/action_model': 1.3520114271426381e-05, 'epoch': 2.33}) +[2026-06-18 01:25:39,541][starVLA.training.train_starvla][INFO] - Step 5665, Loss: {'train/loss': 0.29437291622161865, 'train/grad_norm_pre_clip': 27.818010330200195, 'timing/data': 0.008275847882032394, 'timing/model': 0.9463092442601919, 'learning_rate/qwen_vl_interface': 1.3507759377221001e-06, 'learning_rate/action_model': 1.3507759377221003e-05, 'epoch': 2.33}) +[2026-06-18 01:25:43,461][starVLA.training.train_starvla][INFO] - Step 5666, Loss: {'train/loss': 0.22468921542167664, 'train/grad_norm_pre_clip': 18.524276733398438, 'timing/data': 0.011301805265247822, 'timing/model': 0.9377880385145545, 'learning_rate/qwen_vl_interface': 1.3495412566149058e-06, 'learning_rate/action_model': 1.3495412566149057e-05, 'epoch': 2.33}) +[2026-06-18 01:25:47,280][starVLA.training.train_starvla][INFO] - Step 5667, Loss: {'train/loss': 0.4035525619983673, 'train/grad_norm_pre_clip': 26.309185028076172, 'timing/data': 0.008298326283693314, 'timing/model': 0.9369282461702824, 'learning_rate/qwen_vl_interface': 1.3483073840770067e-06, 'learning_rate/action_model': 1.3483073840770066e-05, 'epoch': 2.33}) +[2026-06-18 01:25:51,132][starVLA.training.train_starvla][INFO] - Step 5668, Loss: {'train/loss': 0.3747514486312866, 'train/grad_norm_pre_clip': 25.848005294799805, 'timing/data': 0.011019796133041382, 'timing/model': 0.92549685575068, 'learning_rate/qwen_vl_interface': 1.3470743203641846e-06, 'learning_rate/action_model': 1.3470743203641844e-05, 'epoch': 2.33}) +[2026-06-18 01:25:54,958][starVLA.training.train_starvla][INFO] - Step 5669, Loss: {'train/loss': 0.4372090995311737, 'train/grad_norm_pre_clip': 35.78919219970703, 'timing/data': 0.012042366899549961, 'timing/model': 1.1143136350437999, 'learning_rate/qwen_vl_interface': 1.3458420657320555e-06, 'learning_rate/action_model': 1.3458420657320555e-05, 'epoch': 2.33}) +[2026-06-18 01:25:59,198][starVLA.training.train_starvla][INFO] - Step 5670, Loss: {'train/loss': 0.4207254648208618, 'train/grad_norm_pre_clip': 33.14781188964844, 'timing/data': 0.012827089056372643, 'timing/model': 1.1675378354266286, 'learning_rate/qwen_vl_interface': 1.344610620436067e-06, 'learning_rate/action_model': 1.344610620436067e-05, 'epoch': 2.33}) +[2026-06-18 01:26:02,936][starVLA.training.train_starvla][INFO] - Step 5671, Loss: {'train/loss': 0.3252023160457611, 'train/grad_norm_pre_clip': 22.498973846435547, 'timing/data': 0.010815713554620743, 'timing/model': 0.9344777101650834, 'learning_rate/qwen_vl_interface': 1.3433799847314989e-06, 'learning_rate/action_model': 1.3433799847314988e-05, 'epoch': 2.33}) +[2026-06-18 01:26:06,742][starVLA.training.train_starvla][INFO] - Step 5672, Loss: {'train/loss': 0.16001002490520477, 'train/grad_norm_pre_clip': 16.051481246948242, 'timing/data': 0.00942759495228529, 'timing/model': 0.9569053836166859, 'learning_rate/qwen_vl_interface': 1.3421501588734655e-06, 'learning_rate/action_model': 1.3421501588734656e-05, 'epoch': 2.33}) +[2026-06-18 01:26:10,570][starVLA.training.train_starvla][INFO] - Step 5673, Loss: {'train/loss': 0.1646699607372284, 'train/grad_norm_pre_clip': 31.667268753051758, 'timing/data': 0.008134504780173302, 'timing/model': 0.9419392421841621, 'learning_rate/qwen_vl_interface': 1.3409211431169084e-06, 'learning_rate/action_model': 1.3409211431169083e-05, 'epoch': 2.34}) +[2026-06-18 01:26:14,393][starVLA.training.train_starvla][INFO] - Step 5674, Loss: {'train/loss': 0.19277499616146088, 'train/grad_norm_pre_clip': 15.207199096679688, 'timing/data': 0.010000734589993954, 'timing/model': 0.9387485533952713, 'learning_rate/qwen_vl_interface': 1.3396929377166042e-06, 'learning_rate/action_model': 1.3396929377166043e-05, 'epoch': 2.34}) +[2026-06-18 01:26:18,697][starVLA.training.train_starvla][INFO] - Step 5675, Loss: {'train/loss': 0.09004314243793488, 'train/grad_norm_pre_clip': 9.92332935333252, 'timing/data': 0.008927726186811924, 'timing/model': 0.9546015271916986, 'learning_rate/qwen_vl_interface': 1.338465542927161e-06, 'learning_rate/action_model': 1.3384655429271608e-05, 'epoch': 2.34}) +[2026-06-18 01:26:22,502][starVLA.training.train_starvla][INFO] - Step 5676, Loss: {'train/loss': 0.4290083050727844, 'train/grad_norm_pre_clip': 20.50975227355957, 'timing/data': 0.01005119550973177, 'timing/model': 1.131774908863008, 'learning_rate/qwen_vl_interface': 1.3372389590030204e-06, 'learning_rate/action_model': 1.3372389590030204e-05, 'epoch': 2.34}) +[2026-06-18 01:26:26,140][starVLA.training.train_starvla][INFO] - Step 5677, Loss: {'train/loss': 0.4200243651866913, 'train/grad_norm_pre_clip': 25.67389488220215, 'timing/data': 0.0104835769161582, 'timing/model': 0.9251552121713758, 'learning_rate/qwen_vl_interface': 1.3360131861984551e-06, 'learning_rate/action_model': 1.3360131861984552e-05, 'epoch': 2.34}) +[2026-06-18 01:26:30,028][starVLA.training.train_starvla][INFO] - Step 5678, Loss: {'train/loss': 0.09441020339727402, 'train/grad_norm_pre_clip': 11.319979667663574, 'timing/data': 0.011832521297037601, 'timing/model': 0.9297631047666073, 'learning_rate/qwen_vl_interface': 1.3347882247675662e-06, 'learning_rate/action_model': 1.3347882247675661e-05, 'epoch': 2.34}) +[2026-06-18 01:26:33,816][starVLA.training.train_starvla][INFO] - Step 5679, Loss: {'train/loss': 0.17106880247592926, 'train/grad_norm_pre_clip': 15.746265411376953, 'timing/data': 0.010810804553329945, 'timing/model': 0.9310192614793777, 'learning_rate/qwen_vl_interface': 1.3335640749642923e-06, 'learning_rate/action_model': 1.3335640749642923e-05, 'epoch': 2.34}) +[2026-06-18 01:26:37,631][starVLA.training.train_starvla][INFO] - Step 5680, Loss: {'train/loss': 0.11875096708536148, 'train/grad_norm_pre_clip': 15.555254936218262, 'timing/data': 0.01025476772338152, 'timing/model': 0.9426975334063172, 'learning_rate/qwen_vl_interface': 1.3323407370423993e-06, 'learning_rate/action_model': 1.3323407370423994e-05, 'epoch': 2.34}) +[2026-06-18 01:26:41,590][starVLA.training.train_starvla][INFO] - Step 5681, Loss: {'train/loss': 0.3907582759857178, 'train/grad_norm_pre_clip': 22.516714096069336, 'timing/data': 0.0101472781971097, 'timing/model': 1.0365294571965933, 'learning_rate/qwen_vl_interface': 1.3311182112554877e-06, 'learning_rate/action_model': 1.3311182112554876e-05, 'epoch': 2.34}) +[2026-06-18 01:26:45,471][starVLA.training.train_starvla][INFO] - Step 5682, Loss: {'train/loss': 0.24844741821289062, 'train/grad_norm_pre_clip': 27.202022552490234, 'timing/data': 0.01066528633236885, 'timing/model': 0.9361429577693343, 'learning_rate/qwen_vl_interface': 1.3298964978569873e-06, 'learning_rate/action_model': 1.3298964978569872e-05, 'epoch': 2.34}) +[2026-06-18 01:26:49,412][starVLA.training.train_starvla][INFO] - Step 5683, Loss: {'train/loss': 0.3600224554538727, 'train/grad_norm_pre_clip': 33.69240188598633, 'timing/data': 0.012013555504381657, 'timing/model': 1.1466598184779286, 'learning_rate/qwen_vl_interface': 1.3286755971001618e-06, 'learning_rate/action_model': 1.3286755971001618e-05, 'epoch': 2.34}) +[2026-06-18 01:26:53,290][starVLA.training.train_starvla][INFO] - Step 5684, Loss: {'train/loss': 0.16598571836948395, 'train/grad_norm_pre_clip': 14.424273490905762, 'timing/data': 0.011812767945230007, 'timing/model': 1.0278484374284744, 'learning_rate/qwen_vl_interface': 1.327455509238104e-06, 'learning_rate/action_model': 1.3274555092381039e-05, 'epoch': 2.34}) +[2026-06-18 01:26:57,159][starVLA.training.train_starvla][INFO] - Step 5685, Loss: {'train/loss': 0.3270266652107239, 'train/grad_norm_pre_clip': 20.510513305664062, 'timing/data': 0.013172022998332977, 'timing/model': 0.9366993019357324, 'learning_rate/qwen_vl_interface': 1.32623623452374e-06, 'learning_rate/action_model': 1.32623623452374e-05, 'epoch': 2.34}) +[2026-06-18 01:27:01,951][starVLA.training.train_starvla][INFO] - Step 5686, Loss: {'train/loss': 0.15540216863155365, 'train/grad_norm_pre_clip': 16.977197647094727, 'timing/data': 0.009605363011360168, 'timing/model': 0.9452879913151264, 'learning_rate/qwen_vl_interface': 1.325017773209827e-06, 'learning_rate/action_model': 1.325017773209827e-05, 'epoch': 2.34}) +[2026-06-18 01:27:06,654][starVLA.training.train_starvla][INFO] - Step 5687, Loss: {'train/loss': 0.1694711595773697, 'train/grad_norm_pre_clip': 21.932849884033203, 'timing/data': 0.008146274834871292, 'timing/model': 1.7414065161719918, 'learning_rate/qwen_vl_interface': 1.323800125548953e-06, 'learning_rate/action_model': 1.323800125548953e-05, 'epoch': 2.34}) +[2026-06-18 01:27:10,607][starVLA.training.train_starvla][INFO] - Step 5688, Loss: {'train/loss': 0.47191548347473145, 'train/grad_norm_pre_clip': 26.062255859375, 'timing/data': 0.008525174111127853, 'timing/model': 0.9485580530017614, 'learning_rate/qwen_vl_interface': 1.3225832917935376e-06, 'learning_rate/action_model': 1.3225832917935377e-05, 'epoch': 2.34}) +[2026-06-18 01:27:14,440][starVLA.training.train_starvla][INFO] - Step 5689, Loss: {'train/loss': 0.4704574942588806, 'train/grad_norm_pre_clip': 30.21595001220703, 'timing/data': 0.008207432925701141, 'timing/model': 1.1379850022494793, 'learning_rate/qwen_vl_interface': 1.3213672721958309e-06, 'learning_rate/action_model': 1.321367272195831e-05, 'epoch': 2.34}) +[2026-06-18 01:27:18,303][starVLA.training.train_starvla][INFO] - Step 5690, Loss: {'train/loss': 0.11038637161254883, 'train/grad_norm_pre_clip': 19.96001625061035, 'timing/data': 0.008651469834148884, 'timing/model': 1.0930124502629042, 'learning_rate/qwen_vl_interface': 1.3201520670079172e-06, 'learning_rate/action_model': 1.3201520670079171e-05, 'epoch': 2.34}) +[2026-06-18 01:27:21,945][starVLA.training.train_starvla][INFO] - Step 5691, Loss: {'train/loss': 0.06562742590904236, 'train/grad_norm_pre_clip': 7.99509334564209, 'timing/data': 0.008838113397359848, 'timing/model': 0.956122076138854, 'learning_rate/qwen_vl_interface': 1.31893767648171e-06, 'learning_rate/action_model': 1.3189376764817099e-05, 'epoch': 2.34}) +[2026-06-18 01:27:29,468][starVLA.training.train_starvla][INFO] - Step 5692, Loss: {'train/loss': 0.27226293087005615, 'train/grad_norm_pre_clip': 21.418277740478516, 'timing/data': 0.008979663252830505, 'timing/model': 2.5035380953922868, 'learning_rate/qwen_vl_interface': 1.3177241008689513e-06, 'learning_rate/action_model': 1.3177241008689511e-05, 'epoch': 2.34}) +[2026-06-18 01:27:34,689][starVLA.training.train_starvla][INFO] - Step 5693, Loss: {'train/loss': 0.09893954545259476, 'train/grad_norm_pre_clip': 10.949848175048828, 'timing/data': 0.008821211755275726, 'timing/model': 0.9559597931802273, 'learning_rate/qwen_vl_interface': 1.316511340421216e-06, 'learning_rate/action_model': 1.316511340421216e-05, 'epoch': 2.34}) +[2026-06-18 01:27:38,596][starVLA.training.train_starvla][INFO] - Step 5694, Loss: {'train/loss': 0.28480446338653564, 'train/grad_norm_pre_clip': 21.66518211364746, 'timing/data': 0.009293468669056892, 'timing/model': 0.9423788236454129, 'learning_rate/qwen_vl_interface': 1.315299395389914e-06, 'learning_rate/action_model': 1.315299395389914e-05, 'epoch': 2.34}) +[2026-06-18 01:27:42,446][starVLA.training.train_starvla][INFO] - Step 5695, Loss: {'train/loss': 0.2670658528804779, 'train/grad_norm_pre_clip': 27.06138038635254, 'timing/data': 0.010292497463524342, 'timing/model': 0.9581289179623127, 'learning_rate/qwen_vl_interface': 1.3140882660262818e-06, 'learning_rate/action_model': 1.3140882660262818e-05, 'epoch': 2.34}) +[2026-06-18 01:27:46,261][starVLA.training.train_starvla][INFO] - Step 5696, Loss: {'train/loss': 0.3200598359107971, 'train/grad_norm_pre_clip': 20.785503387451172, 'timing/data': 0.00876626931130886, 'timing/model': 1.128827410750091, 'learning_rate/qwen_vl_interface': 1.3128779525813865e-06, 'learning_rate/action_model': 1.3128779525813864e-05, 'epoch': 2.34}) +[2026-06-18 01:27:50,038][starVLA.training.train_starvla][INFO] - Step 5697, Loss: {'train/loss': 0.2803534269332886, 'train/grad_norm_pre_clip': 19.898426055908203, 'timing/data': 0.009443200193345547, 'timing/model': 1.0559651544317603, 'learning_rate/qwen_vl_interface': 1.3116684553061288e-06, 'learning_rate/action_model': 1.3116684553061287e-05, 'epoch': 2.35}) +[2026-06-18 01:27:53,743][starVLA.training.train_starvla][INFO] - Step 5698, Loss: {'train/loss': 0.3452152609825134, 'train/grad_norm_pre_clip': 19.3347225189209, 'timing/data': 0.011041834950447083, 'timing/model': 0.9438745360821486, 'learning_rate/qwen_vl_interface': 1.310459774451238e-06, 'learning_rate/action_model': 1.3104597744512379e-05, 'epoch': 2.35}) +[2026-06-18 01:27:57,671][starVLA.training.train_starvla][INFO] - Step 5699, Loss: {'train/loss': 0.07240311056375504, 'train/grad_norm_pre_clip': 9.404631614685059, 'timing/data': 0.008831741288304329, 'timing/model': 0.9378568679094315, 'learning_rate/qwen_vl_interface': 1.3092519102672744e-06, 'learning_rate/action_model': 1.3092519102672742e-05, 'epoch': 2.35}) +[2026-06-18 01:30:12,554][starVLA.training.train_starvla][INFO] - Step 5700, Loss: {'train/loss': 0.14764006435871124, 'train/grad_norm_pre_clip': 23.01576805114746, 'eval/loss': 0.41747745871543884, 'eval/latency_6/loss': 0.41747745871543884, 'eval/demon_fire_precision': 0.9559150015857913, 'eval/demon_fire_recall': 0.9741435035552682, 'eval/demon_fire_f1': 0.9649431727229071, 'eval/demon_fire_rate_model': 0.8772954924874792, 'eval/demon_fire_rate_teacher': 0.860879243183083, 'eval/demon_left_precision': 0.9008620689655172, 'eval/demon_left_recall': 0.7333333333333333, 'eval/demon_left_f1': 0.8085106382978723, 'eval/demon_left_rate_model': 0.12910406232609906, 'eval/demon_left_rate_teacher': 0.15859766277128548, 'eval/demon_right_precision': 0.8391959798994975, 'eval/demon_right_recall': 0.7695852534562212, 'eval/demon_right_f1': 0.8028846153846153, 'eval/demon_right_rate_model': 0.166110183639399, 'eval/demon_right_rate_teacher': 0.18113522537562604, 'eval/demon_move_f1': 0.8056976268412438, 'eval/demon_cc_f1': 0.9109297909103861, 'eval/latency_6/demon_fire_precision': 0.9559150015857913, 'eval/latency_6/demon_fire_recall': 0.9741435035552682, 'eval/latency_6/demon_fire_f1': 0.9649431727229071, 'eval/latency_6/demon_fire_rate_model': 0.8772954924874792, 'eval/latency_6/demon_fire_rate_teacher': 0.860879243183083, 'eval/latency_6/demon_left_precision': 0.9008620689655172, 'eval/latency_6/demon_left_recall': 0.7333333333333333, 'eval/latency_6/demon_left_f1': 0.8085106382978723, 'eval/latency_6/demon_left_rate_model': 0.12910406232609906, 'eval/latency_6/demon_left_rate_teacher': 0.15859766277128548, 'eval/latency_6/demon_right_precision': 0.8391959798994975, 'eval/latency_6/demon_right_recall': 0.7695852534562212, 'eval/latency_6/demon_right_f1': 0.8028846153846153, 'eval/latency_6/demon_right_rate_model': 0.166110183639399, 'eval/latency_6/demon_right_rate_teacher': 0.18113522537562604, 'eval/latency_6/demon_move_f1': 0.8056976268412438, 'eval/latency_6/demon_cc_f1': 0.9109297909103861, 'timing/data': 0.010041480883955956, 'timing/model': 0.9488658029586077, 'learning_rate/qwen_vl_interface': 1.3080448630046312e-06, 'learning_rate/action_model': 1.308044863004631e-05, 'epoch': 2.35}) +[2026-06-18 01:30:12,556][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_5700_state +[2026-06-18 01:30:12,563][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 01:30:19,755][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_5700_state/model.safetensors +[2026-06-18 01:30:35,701][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_5700_state/optimizer.bin +[2026-06-18 01:30:35,702][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_5700_state/sampler.bin +[2026-06-18 01:30:35,702][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_5700_state/sampler_1.bin +[2026-06-18 01:30:35,706][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_5700_state/random_states_0.pkl +[2026-06-18 01:30:35,713][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 5700 +[2026-06-18 01:30:35,713][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 01:30:35,724][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 01:30:39,677][starVLA.training.train_starvla][INFO] - Step 5701, Loss: {'train/loss': 0.36776867508888245, 'train/grad_norm_pre_clip': 22.67054557800293, 'timing/data': 0.009225360117852688, 'timing/model': 0.9424296068027616, 'learning_rate/qwen_vl_interface': 1.3068386329135287e-06, 'learning_rate/action_model': 1.3068386329135287e-05, 'epoch': 2.35}) +[2026-06-18 01:30:43,483][starVLA.training.train_starvla][INFO] - Step 5702, Loss: {'train/loss': 0.16292311251163483, 'train/grad_norm_pre_clip': 18.852081298828125, 'timing/data': 0.009566922672092915, 'timing/model': 0.9203283283859491, 'learning_rate/qwen_vl_interface': 1.3056332202440211e-06, 'learning_rate/action_model': 1.305633220244021e-05, 'epoch': 2.35}) +[2026-06-18 01:30:47,261][starVLA.training.train_starvla][INFO] - Step 5703, Loss: {'train/loss': 0.27360063791275024, 'train/grad_norm_pre_clip': 22.909645080566406, 'timing/data': 0.007877410389482975, 'timing/model': 0.9393206210806966, 'learning_rate/qwen_vl_interface': 1.3044286252459914e-06, 'learning_rate/action_model': 1.3044286252459915e-05, 'epoch': 2.35}) +[2026-06-18 01:30:51,077][starVLA.training.train_starvla][INFO] - Step 5704, Loss: {'train/loss': 0.32142767310142517, 'train/grad_norm_pre_clip': 77.6683578491211, 'timing/data': 0.009061587043106556, 'timing/model': 0.961292203515768, 'learning_rate/qwen_vl_interface': 1.3032248481691523e-06, 'learning_rate/action_model': 1.3032248481691523e-05, 'epoch': 2.35}) +[2026-06-18 01:30:54,933][starVLA.training.train_starvla][INFO] - Step 5705, Loss: {'train/loss': 0.30509987473487854, 'train/grad_norm_pre_clip': 23.20659637451172, 'timing/data': 0.014464394189417362, 'timing/model': 1.109913925640285, 'learning_rate/qwen_vl_interface': 1.302021889263051e-06, 'learning_rate/action_model': 1.3020218892630508e-05, 'epoch': 2.35}) +[2026-06-18 01:30:58,779][starVLA.training.train_starvla][INFO] - Step 5706, Loss: {'train/loss': 0.31609517335891724, 'train/grad_norm_pre_clip': 23.76882553100586, 'timing/data': 0.00908617116510868, 'timing/model': 1.100214073434472, 'learning_rate/qwen_vl_interface': 1.3008197487770593e-06, 'learning_rate/action_model': 1.3008197487770593e-05, 'epoch': 2.35}) +[2026-06-18 01:31:02,486][starVLA.training.train_starvla][INFO] - Step 5707, Loss: {'train/loss': 0.21056349575519562, 'train/grad_norm_pre_clip': 24.28955841064453, 'timing/data': 0.010104629211127758, 'timing/model': 0.9370444510132074, 'learning_rate/qwen_vl_interface': 1.2996184269603835e-06, 'learning_rate/action_model': 1.2996184269603833e-05, 'epoch': 2.35}) +[2026-06-18 01:31:06,309][starVLA.training.train_starvla][INFO] - Step 5708, Loss: {'train/loss': 0.32840803265571594, 'train/grad_norm_pre_clip': 20.619129180908203, 'timing/data': 0.011580600403249264, 'timing/model': 0.9428603919222951, 'learning_rate/qwen_vl_interface': 1.2984179240620567e-06, 'learning_rate/action_model': 1.2984179240620567e-05, 'epoch': 2.35}) +[2026-06-18 01:31:10,157][starVLA.training.train_starvla][INFO] - Step 5709, Loss: {'train/loss': 0.34887468814849854, 'train/grad_norm_pre_clip': 24.93169403076172, 'timing/data': 0.011654006317257881, 'timing/model': 0.9411514438688755, 'learning_rate/qwen_vl_interface': 1.2972182403309474e-06, 'learning_rate/action_model': 1.2972182403309474e-05, 'epoch': 2.35}) +[2026-06-18 01:31:14,044][starVLA.training.train_starvla][INFO] - Step 5710, Loss: {'train/loss': 0.0964718759059906, 'train/grad_norm_pre_clip': 12.432846069335938, 'timing/data': 0.008765777572989464, 'timing/model': 0.9442091705277562, 'learning_rate/qwen_vl_interface': 1.2960193760157507e-06, 'learning_rate/action_model': 1.2960193760157507e-05, 'epoch': 2.35}) +[2026-06-18 01:31:19,076][starVLA.training.train_starvla][INFO] - Step 5711, Loss: {'train/loss': 0.3481109142303467, 'train/grad_norm_pre_clip': 20.360565185546875, 'timing/data': 0.010260299779474735, 'timing/model': 2.0317991375923157, 'learning_rate/qwen_vl_interface': 1.294821331364992e-06, 'learning_rate/action_model': 1.294821331364992e-05, 'epoch': 2.35}) +[2026-06-18 01:31:23,238][starVLA.training.train_starvla][INFO] - Step 5712, Loss: {'train/loss': 0.15280140936374664, 'train/grad_norm_pre_clip': 9.845346450805664, 'timing/data': 0.011792588047683239, 'timing/model': 1.2650115080177784, 'learning_rate/qwen_vl_interface': 1.2936241066270278e-06, 'learning_rate/action_model': 1.2936241066270278e-05, 'epoch': 2.35}) +[2026-06-18 01:31:27,133][starVLA.training.train_starvla][INFO] - Step 5713, Loss: {'train/loss': 0.30312883853912354, 'train/grad_norm_pre_clip': 21.25393295288086, 'timing/data': 0.013871537521481514, 'timing/model': 1.0585030931979418, 'learning_rate/qwen_vl_interface': 1.2924277020500427e-06, 'learning_rate/action_model': 1.2924277020500428e-05, 'epoch': 2.35}) +[2026-06-18 01:31:31,032][starVLA.training.train_starvla][INFO] - Step 5714, Loss: {'train/loss': 0.2978026866912842, 'train/grad_norm_pre_clip': 28.296245574951172, 'timing/data': 0.021238505840301514, 'timing/model': 0.9302931018173695, 'learning_rate/qwen_vl_interface': 1.291232117882054e-06, 'learning_rate/action_model': 1.2912321178820538e-05, 'epoch': 2.35}) +[2026-06-18 01:31:34,861][starVLA.training.train_starvla][INFO] - Step 5715, Loss: {'train/loss': 0.07513714581727982, 'train/grad_norm_pre_clip': 9.57261848449707, 'timing/data': 0.009906437247991562, 'timing/model': 0.9436374232172966, 'learning_rate/qwen_vl_interface': 1.2900373543709065e-06, 'learning_rate/action_model': 1.2900373543709065e-05, 'epoch': 2.35}) +[2026-06-18 01:31:38,801][starVLA.training.train_starvla][INFO] - Step 5716, Loss: {'train/loss': 0.1933889240026474, 'train/grad_norm_pre_clip': 17.557863235473633, 'timing/data': 0.011384033598005772, 'timing/model': 0.9463809011504054, 'learning_rate/qwen_vl_interface': 1.288843411764277e-06, 'learning_rate/action_model': 1.288843411764277e-05, 'epoch': 2.35}) +[2026-06-18 01:31:42,749][starVLA.training.train_starvla][INFO] - Step 5717, Loss: {'train/loss': 0.3624386191368103, 'train/grad_norm_pre_clip': 23.403404235839844, 'timing/data': 0.009777878411114216, 'timing/model': 0.9433247921988368, 'learning_rate/qwen_vl_interface': 1.2876502903096709e-06, 'learning_rate/action_model': 1.2876502903096708e-05, 'epoch': 2.35}) +[2026-06-18 01:31:46,649][starVLA.training.train_starvla][INFO] - Step 5718, Loss: {'train/loss': 0.2997986376285553, 'train/grad_norm_pre_clip': 16.316436767578125, 'timing/data': 0.012625476345419884, 'timing/model': 1.1588933523744345, 'learning_rate/qwen_vl_interface': 1.286457990254423e-06, 'learning_rate/action_model': 1.286457990254423e-05, 'epoch': 2.35}) +[2026-06-18 01:31:50,596][starVLA.training.train_starvla][INFO] - Step 5719, Loss: {'train/loss': 0.2880082428455353, 'train/grad_norm_pre_clip': 34.26191329956055, 'timing/data': 0.011342553421854973, 'timing/model': 1.1555969268083572, 'learning_rate/qwen_vl_interface': 1.2852665118456971e-06, 'learning_rate/action_model': 1.2852665118456972e-05, 'epoch': 2.35}) +[2026-06-18 01:31:54,237][starVLA.training.train_starvla][INFO] - Step 5720, Loss: {'train/loss': 0.24343706667423248, 'train/grad_norm_pre_clip': 29.59337043762207, 'timing/data': 0.010327747091650963, 'timing/model': 0.94424207508564, 'learning_rate/qwen_vl_interface': 1.2840758553304927e-06, 'learning_rate/action_model': 1.2840758553304927e-05, 'epoch': 2.35}) +[2026-06-18 01:31:58,196][starVLA.training.train_starvla][INFO] - Step 5721, Loss: {'train/loss': 0.11997086554765701, 'train/grad_norm_pre_clip': 24.27623176574707, 'timing/data': 0.010262897238135338, 'timing/model': 0.9547527153044939, 'learning_rate/qwen_vl_interface': 1.2828860209556293e-06, 'learning_rate/action_model': 1.2828860209556292e-05, 'epoch': 2.35}) +[2026-06-18 01:32:02,059][starVLA.training.train_starvla][INFO] - Step 5722, Loss: {'train/loss': 0.38184675574302673, 'train/grad_norm_pre_clip': 30.156307220458984, 'timing/data': 0.0093741649761796, 'timing/model': 0.9368053358048201, 'learning_rate/qwen_vl_interface': 1.2816970089677617e-06, 'learning_rate/action_model': 1.2816970089677615e-05, 'epoch': 2.36}) +[2026-06-18 01:32:06,086][starVLA.training.train_starvla][INFO] - Step 5723, Loss: {'train/loss': 0.3756182789802551, 'train/grad_norm_pre_clip': 26.88908576965332, 'timing/data': 0.010790746659040451, 'timing/model': 1.1084974193945527, 'learning_rate/qwen_vl_interface': 1.280508819613375e-06, 'learning_rate/action_model': 1.280508819613375e-05, 'epoch': 2.36}) +[2026-06-18 01:32:09,934][starVLA.training.train_starvla][INFO] - Step 5724, Loss: {'train/loss': 0.29110535979270935, 'train/grad_norm_pre_clip': 16.457847595214844, 'timing/data': 0.012333798222243786, 'timing/model': 0.9451503157615662, 'learning_rate/qwen_vl_interface': 1.2793214531387812e-06, 'learning_rate/action_model': 1.2793214531387812e-05, 'epoch': 2.36}) +[2026-06-18 01:32:13,705][starVLA.training.train_starvla][INFO] - Step 5725, Loss: {'train/loss': 0.38451772928237915, 'train/grad_norm_pre_clip': 23.249454498291016, 'timing/data': 0.0096932053565979, 'timing/model': 0.9538902081549168, 'learning_rate/qwen_vl_interface': 1.2781349097901246e-06, 'learning_rate/action_model': 1.2781349097901247e-05, 'epoch': 2.36}) +[2026-06-18 01:32:17,569][starVLA.training.train_starvla][INFO] - Step 5726, Loss: {'train/loss': 0.2454323023557663, 'train/grad_norm_pre_clip': 16.580772399902344, 'timing/data': 0.008034567348659039, 'timing/model': 1.094110424630344, 'learning_rate/qwen_vl_interface': 1.2769491898133723e-06, 'learning_rate/action_model': 1.2769491898133722e-05, 'epoch': 2.36}) +[2026-06-18 01:32:21,395][starVLA.training.train_starvla][INFO] - Step 5727, Loss: {'train/loss': 0.0960451140999794, 'train/grad_norm_pre_clip': 10.33266544342041, 'timing/data': 0.00901055708527565, 'timing/model': 0.9512230549007654, 'learning_rate/qwen_vl_interface': 1.2757642934543293e-06, 'learning_rate/action_model': 1.2757642934543293e-05, 'epoch': 2.36}) +[2026-06-18 01:32:25,246][starVLA.training.train_starvla][INFO] - Step 5728, Loss: {'train/loss': 0.30329105257987976, 'train/grad_norm_pre_clip': 19.589954376220703, 'timing/data': 0.017133360728621483, 'timing/model': 0.9398934673517942, 'learning_rate/qwen_vl_interface': 1.274580220958625e-06, 'learning_rate/action_model': 1.274580220958625e-05, 'epoch': 2.36}) +[2026-06-18 01:32:29,095][starVLA.training.train_starvla][INFO] - Step 5729, Loss: {'train/loss': 0.23455359041690826, 'train/grad_norm_pre_clip': 19.00262451171875, 'timing/data': 0.010944453068077564, 'timing/model': 0.9321703650057316, 'learning_rate/qwen_vl_interface': 1.2733969725717184e-06, 'learning_rate/action_model': 1.2733969725717184e-05, 'epoch': 2.36}) +[2026-06-18 01:32:32,854][starVLA.training.train_starvla][INFO] - Step 5730, Loss: {'train/loss': 0.41102638840675354, 'train/grad_norm_pre_clip': 24.220928192138672, 'timing/data': 0.010712326504290104, 'timing/model': 0.9371612602844834, 'learning_rate/qwen_vl_interface': 1.272214548538899e-06, 'learning_rate/action_model': 1.2722145485388989e-05, 'epoch': 2.36}) +[2026-06-18 01:32:36,696][starVLA.training.train_starvla][INFO] - Step 5731, Loss: {'train/loss': 0.08684471994638443, 'train/grad_norm_pre_clip': 11.286360740661621, 'timing/data': 0.012009908445179462, 'timing/model': 1.0176540091633797, 'learning_rate/qwen_vl_interface': 1.2710329491052834e-06, 'learning_rate/action_model': 1.2710329491052833e-05, 'epoch': 2.36}) +[2026-06-18 01:32:40,525][starVLA.training.train_starvla][INFO] - Step 5732, Loss: {'train/loss': 0.2541637420654297, 'train/grad_norm_pre_clip': 19.848020553588867, 'timing/data': 0.010221876204013824, 'timing/model': 1.1284079477190971, 'learning_rate/qwen_vl_interface': 1.2698521745158188e-06, 'learning_rate/action_model': 1.2698521745158187e-05, 'epoch': 2.36}) +[2026-06-18 01:32:44,268][starVLA.training.train_starvla][INFO] - Step 5733, Loss: {'train/loss': 0.2799132466316223, 'train/grad_norm_pre_clip': 25.715843200683594, 'timing/data': 0.007376241497695446, 'timing/model': 1.0763570675626397, 'learning_rate/qwen_vl_interface': 1.2686722250152804e-06, 'learning_rate/action_model': 1.2686722250152806e-05, 'epoch': 2.36}) +[2026-06-18 01:32:47,935][starVLA.training.train_starvla][INFO] - Step 5734, Loss: {'train/loss': 0.15032930672168732, 'train/grad_norm_pre_clip': 15.451969146728516, 'timing/data': 0.010608607903122902, 'timing/model': 0.9379700925201178, 'learning_rate/qwen_vl_interface': 1.2674931008482763e-06, 'learning_rate/action_model': 1.2674931008482763e-05, 'epoch': 2.36}) +[2026-06-18 01:32:51,742][starVLA.training.train_starvla][INFO] - Step 5735, Loss: {'train/loss': 0.24150559306144714, 'train/grad_norm_pre_clip': 22.266036987304688, 'timing/data': 0.008113289251923561, 'timing/model': 0.9504945278167725, 'learning_rate/qwen_vl_interface': 1.2663148022592365e-06, 'learning_rate/action_model': 1.2663148022592363e-05, 'epoch': 2.36}) +[2026-06-18 01:32:55,574][starVLA.training.train_starvla][INFO] - Step 5736, Loss: {'train/loss': 0.2836528718471527, 'train/grad_norm_pre_clip': 18.707172393798828, 'timing/data': 0.0069824038073420525, 'timing/model': 0.9402248356491327, 'learning_rate/qwen_vl_interface': 1.2651373294924249e-06, 'learning_rate/action_model': 1.2651373294924248e-05, 'epoch': 2.36}) +[2026-06-18 01:32:59,603][starVLA.training.train_starvla][INFO] - Step 5737, Loss: {'train/loss': 0.15098083019256592, 'train/grad_norm_pre_clip': 14.764067649841309, 'timing/data': 0.00868504773825407, 'timing/model': 0.9573812894523144, 'learning_rate/qwen_vl_interface': 1.2639606827919313e-06, 'learning_rate/action_model': 1.2639606827919311e-05, 'epoch': 2.36}) +[2026-06-18 01:33:04,713][starVLA.training.train_starvla][INFO] - Step 5738, Loss: {'train/loss': 0.5077522993087769, 'train/grad_norm_pre_clip': 26.03592300415039, 'timing/data': 0.011359432712197304, 'timing/model': 1.9960863515734673, 'learning_rate/qwen_vl_interface': 1.2627848624016788e-06, 'learning_rate/action_model': 1.2627848624016788e-05, 'epoch': 2.36}) +[2026-06-18 01:33:08,734][starVLA.training.train_starvla][INFO] - Step 5739, Loss: {'train/loss': 0.25180649757385254, 'train/grad_norm_pre_clip': 21.442529678344727, 'timing/data': 0.00855047907680273, 'timing/model': 1.1791398646309972, 'learning_rate/qwen_vl_interface': 1.261609868565416e-06, 'learning_rate/action_model': 1.2616098685654159e-05, 'epoch': 2.36}) +[2026-06-18 01:33:12,476][starVLA.training.train_starvla][INFO] - Step 5740, Loss: {'train/loss': 0.20650911331176758, 'train/grad_norm_pre_clip': 24.5535945892334, 'timing/data': 0.009443514049053192, 'timing/model': 0.9386701863259077, 'learning_rate/qwen_vl_interface': 1.2604357015267165e-06, 'learning_rate/action_model': 1.2604357015267166e-05, 'epoch': 2.36}) +[2026-06-18 01:33:18,562][starVLA.training.train_starvla][INFO] - Step 5741, Loss: {'train/loss': 0.19139331579208374, 'train/grad_norm_pre_clip': 20.059370040893555, 'timing/data': 0.009020118042826653, 'timing/model': 1.7807120410725474, 'learning_rate/qwen_vl_interface': 1.2592623615289906e-06, 'learning_rate/action_model': 1.2592623615289906e-05, 'epoch': 2.36}) +[2026-06-18 01:33:22,367][starVLA.training.train_starvla][INFO] - Step 5742, Loss: {'train/loss': 0.36567652225494385, 'train/grad_norm_pre_clip': 23.516895294189453, 'timing/data': 0.009824787266552448, 'timing/model': 0.9355557486414909, 'learning_rate/qwen_vl_interface': 1.258089848815471e-06, 'learning_rate/action_model': 1.2580898488154708e-05, 'epoch': 2.36}) +[2026-06-18 01:33:28,256][starVLA.training.train_starvla][INFO] - Step 5743, Loss: {'train/loss': 0.25758567452430725, 'train/grad_norm_pre_clip': 15.868306159973145, 'timing/data': 0.008597095496952534, 'timing/model': 0.9475622735917568, 'learning_rate/qwen_vl_interface': 1.256918163629221e-06, 'learning_rate/action_model': 1.256918163629221e-05, 'epoch': 2.36}) +[2026-06-18 01:33:32,153][starVLA.training.train_starvla][INFO] - Step 5744, Loss: {'train/loss': 0.304315447807312, 'train/grad_norm_pre_clip': 24.303630828857422, 'timing/data': 0.009377927519381046, 'timing/model': 0.978168411180377, 'learning_rate/qwen_vl_interface': 1.2557473062131332e-06, 'learning_rate/action_model': 1.2557473062131331e-05, 'epoch': 2.36}) +[2026-06-18 01:33:36,010][starVLA.training.train_starvla][INFO] - Step 5745, Loss: {'train/loss': 0.34236401319503784, 'train/grad_norm_pre_clip': 34.846031188964844, 'timing/data': 0.008087072521448135, 'timing/model': 1.1145923249423504, 'learning_rate/qwen_vl_interface': 1.2545772768099263e-06, 'learning_rate/action_model': 1.2545772768099263e-05, 'epoch': 2.36}) +[2026-06-18 01:33:39,805][starVLA.training.train_starvla][INFO] - Step 5746, Loss: {'train/loss': 0.1884651631116867, 'train/grad_norm_pre_clip': 22.62679100036621, 'timing/data': 0.00805982667952776, 'timing/model': 1.118377267383039, 'learning_rate/qwen_vl_interface': 1.2534080756621488e-06, 'learning_rate/action_model': 1.2534080756621488e-05, 'epoch': 2.37}) +[2026-06-18 01:33:43,499][starVLA.training.train_starvla][INFO] - Step 5747, Loss: {'train/loss': 0.4228360056877136, 'train/grad_norm_pre_clip': 15.966788291931152, 'timing/data': 0.011443699710071087, 'timing/model': 0.9811493549495935, 'learning_rate/qwen_vl_interface': 1.2522397030121782e-06, 'learning_rate/action_model': 1.2522397030121783e-05, 'epoch': 2.37}) +[2026-06-18 01:33:47,388][starVLA.training.train_starvla][INFO] - Step 5748, Loss: {'train/loss': 0.17495892941951752, 'train/grad_norm_pre_clip': 20.01668930053711, 'timing/data': 0.013063712976872921, 'timing/model': 0.9396498240530491, 'learning_rate/qwen_vl_interface': 1.2510721591022194e-06, 'learning_rate/action_model': 1.2510721591022192e-05, 'epoch': 2.37}) +[2026-06-18 01:33:51,224][starVLA.training.train_starvla][INFO] - Step 5749, Loss: {'train/loss': 0.282640278339386, 'train/grad_norm_pre_clip': 21.37794303894043, 'timing/data': 0.010223648510873318, 'timing/model': 0.9446631213650107, 'learning_rate/qwen_vl_interface': 1.2499054441743047e-06, 'learning_rate/action_model': 1.2499054441743047e-05, 'epoch': 2.37}) +[2026-06-18 01:33:55,000][starVLA.training.train_starvla][INFO] - Step 5750, Loss: {'train/loss': 0.2163916379213333, 'train/grad_norm_pre_clip': 18.888978958129883, 'timing/data': 0.009444255381822586, 'timing/model': 0.934370594099164, 'learning_rate/qwen_vl_interface': 1.2487395584702955e-06, 'learning_rate/action_model': 1.2487395584702954e-05, 'epoch': 2.37}) +[2026-06-18 01:33:58,886][starVLA.training.train_starvla][INFO] - Step 5751, Loss: {'train/loss': 0.029629819095134735, 'train/grad_norm_pre_clip': 4.600975513458252, 'timing/data': 0.012997264973819256, 'timing/model': 0.9942607060074806, 'learning_rate/qwen_vl_interface': 1.2475745022318803e-06, 'learning_rate/action_model': 1.2475745022318803e-05, 'epoch': 2.37}) +[2026-06-18 01:34:02,780][starVLA.training.train_starvla][INFO] - Step 5752, Loss: {'train/loss': 0.13569258153438568, 'train/grad_norm_pre_clip': 15.538008689880371, 'timing/data': 0.008810924366116524, 'timing/model': 1.1319434577599168, 'learning_rate/qwen_vl_interface': 1.2464102757005788e-06, 'learning_rate/action_model': 1.2464102757005788e-05, 'epoch': 2.37}) +[2026-06-18 01:34:06,613][starVLA.training.train_starvla][INFO] - Step 5753, Loss: {'train/loss': 0.29394087195396423, 'train/grad_norm_pre_clip': 28.676856994628906, 'timing/data': 0.01015438325703144, 'timing/model': 1.149077270179987, 'learning_rate/qwen_vl_interface': 1.2452468791177357e-06, 'learning_rate/action_model': 1.2452468791177357e-05, 'epoch': 2.37}) +[2026-06-18 01:34:10,292][starVLA.training.train_starvla][INFO] - Step 5754, Loss: {'train/loss': 0.12375771254301071, 'train/grad_norm_pre_clip': 12.407323837280273, 'timing/data': 0.0109605323523283, 'timing/model': 0.935677601955831, 'learning_rate/qwen_vl_interface': 1.2440843127245235e-06, 'learning_rate/action_model': 1.2440843127245235e-05, 'epoch': 2.37}) +[2026-06-18 01:34:14,153][starVLA.training.train_starvla][INFO] - Step 5755, Loss: {'train/loss': 0.15996885299682617, 'train/grad_norm_pre_clip': 17.80986785888672, 'timing/data': 0.011111265048384666, 'timing/model': 0.9465554878115654, 'learning_rate/qwen_vl_interface': 1.2429225767619427e-06, 'learning_rate/action_model': 1.2429225767619427e-05, 'epoch': 2.37}) +[2026-06-18 01:34:17,969][starVLA.training.train_starvla][INFO] - Step 5756, Loss: {'train/loss': 0.27340614795684814, 'train/grad_norm_pre_clip': 142.24728393554688, 'timing/data': 0.01327627431601286, 'timing/model': 0.9266101093962789, 'learning_rate/qwen_vl_interface': 1.2417616714708246e-06, 'learning_rate/action_model': 1.2417616714708247e-05, 'epoch': 2.37}) +[2026-06-18 01:34:21,741][starVLA.training.train_starvla][INFO] - Step 5757, Loss: {'train/loss': 0.24433015286922455, 'train/grad_norm_pre_clip': 19.629844665527344, 'timing/data': 0.009529275819659233, 'timing/model': 0.9448086563497782, 'learning_rate/qwen_vl_interface': 1.2406015970918247e-06, 'learning_rate/action_model': 1.2406015970918247e-05, 'epoch': 2.37}) +[2026-06-18 01:34:25,579][starVLA.training.train_starvla][INFO] - Step 5758, Loss: {'train/loss': 0.29503172636032104, 'train/grad_norm_pre_clip': 54.31428909301758, 'timing/data': 0.019441833719611168, 'timing/model': 0.9389624465256929, 'learning_rate/qwen_vl_interface': 1.2394423538654296e-06, 'learning_rate/action_model': 1.2394423538654296e-05, 'epoch': 2.37}) +[2026-06-18 01:34:29,440][starVLA.training.train_starvla][INFO] - Step 5759, Loss: {'train/loss': 0.14017535746097565, 'train/grad_norm_pre_clip': 16.568817138671875, 'timing/data': 0.010404294356703758, 'timing/model': 1.11096853017807, 'learning_rate/qwen_vl_interface': 1.2382839420319478e-06, 'learning_rate/action_model': 1.2382839420319478e-05, 'epoch': 2.37}) +[2026-06-18 01:34:33,276][starVLA.training.train_starvla][INFO] - Step 5760, Loss: {'train/loss': 0.08907358348369598, 'train/grad_norm_pre_clip': 12.155125617980957, 'timing/data': 0.007584967650473118, 'timing/model': 0.9413471249863505, 'learning_rate/qwen_vl_interface': 1.2371263618315225e-06, 'learning_rate/action_model': 1.2371263618315224e-05, 'epoch': 2.37}) +[2026-06-18 01:34:37,103][starVLA.training.train_starvla][INFO] - Step 5761, Loss: {'train/loss': 0.36154747009277344, 'train/grad_norm_pre_clip': 25.220569610595703, 'timing/data': 0.011801499873399734, 'timing/model': 0.9531747885048389, 'learning_rate/qwen_vl_interface': 1.2359696135041202e-06, 'learning_rate/action_model': 1.2359696135041201e-05, 'epoch': 2.37}) +[2026-06-18 01:34:40,910][starVLA.training.train_starvla][INFO] - Step 5762, Loss: {'train/loss': 0.2900465726852417, 'train/grad_norm_pre_clip': 22.611736297607422, 'timing/data': 0.01037670485675335, 'timing/model': 0.9336128830909729, 'learning_rate/qwen_vl_interface': 1.2348136972895369e-06, 'learning_rate/action_model': 1.2348136972895368e-05, 'epoch': 2.37}) +[2026-06-18 01:34:44,952][starVLA.training.train_starvla][INFO] - Step 5763, Loss: {'train/loss': 0.12291561812162399, 'train/grad_norm_pre_clip': 24.046985626220703, 'timing/data': 0.008630550466477871, 'timing/model': 0.9844212513417006, 'learning_rate/qwen_vl_interface': 1.2336586134273945e-06, 'learning_rate/action_model': 1.2336586134273945e-05, 'epoch': 2.37}) +[2026-06-18 01:34:48,872][starVLA.training.train_starvla][INFO] - Step 5764, Loss: {'train/loss': 0.073679119348526, 'train/grad_norm_pre_clip': 9.605425834655762, 'timing/data': 0.009444593451917171, 'timing/model': 0.9528520787134767, 'learning_rate/qwen_vl_interface': 1.2325043621571428e-06, 'learning_rate/action_model': 1.2325043621571427e-05, 'epoch': 2.37}) +[2026-06-18 01:34:52,733][starVLA.training.train_starvla][INFO] - Step 5765, Loss: {'train/loss': 0.4870447814464569, 'train/grad_norm_pre_clip': 18.120433807373047, 'timing/data': 0.009138424880802631, 'timing/model': 0.9336321912705898, 'learning_rate/qwen_vl_interface': 1.2313509437180603e-06, 'learning_rate/action_model': 1.2313509437180604e-05, 'epoch': 2.37}) +[2026-06-18 01:34:57,121][starVLA.training.train_starvla][INFO] - Step 5766, Loss: {'train/loss': 0.10731141269207001, 'train/grad_norm_pre_clip': 11.947714805603027, 'timing/data': 0.0083144661039114, 'timing/model': 1.1662297388538718, 'learning_rate/qwen_vl_interface': 1.2301983583492502e-06, 'learning_rate/action_model': 1.2301983583492502e-05, 'epoch': 2.37}) +[2026-06-18 01:35:00,932][starVLA.training.train_starvla][INFO] - Step 5767, Loss: {'train/loss': 0.04696929082274437, 'train/grad_norm_pre_clip': 7.005545616149902, 'timing/data': 0.010342782363295555, 'timing/model': 0.9372997544705868, 'learning_rate/qwen_vl_interface': 1.229046606289648e-06, 'learning_rate/action_model': 1.229046606289648e-05, 'epoch': 2.37}) +[2026-06-18 01:35:04,872][starVLA.training.train_starvla][INFO] - Step 5768, Loss: {'train/loss': 0.5946626663208008, 'train/grad_norm_pre_clip': 36.015228271484375, 'timing/data': 0.00921802781522274, 'timing/model': 0.9398057898506522, 'learning_rate/qwen_vl_interface': 1.2278956877780102e-06, 'learning_rate/action_model': 1.2278956877780101e-05, 'epoch': 2.37}) +[2026-06-18 01:35:10,047][starVLA.training.train_starvla][INFO] - Step 5769, Loss: {'train/loss': 0.2014821171760559, 'train/grad_norm_pre_clip': 24.858821868896484, 'timing/data': 0.008344155736267567, 'timing/model': 0.937853884883225, 'learning_rate/qwen_vl_interface': 1.2267456030529245e-06, 'learning_rate/action_model': 1.2267456030529245e-05, 'epoch': 2.37}) +[2026-06-18 01:35:14,261][starVLA.training.train_starvla][INFO] - Step 5770, Loss: {'train/loss': 0.5696631073951721, 'train/grad_norm_pre_clip': 30.339614868164062, 'timing/data': 0.011737584136426449, 'timing/model': 0.9340818030759692, 'learning_rate/qwen_vl_interface': 1.225596352352803e-06, 'learning_rate/action_model': 1.225596352352803e-05, 'epoch': 2.38}) +[2026-06-18 01:35:18,101][starVLA.training.train_starvla][INFO] - Step 5771, Loss: {'train/loss': 0.5300648808479309, 'train/grad_norm_pre_clip': 41.15816116333008, 'timing/data': 0.009513571858406067, 'timing/model': 0.9380489587783813, 'learning_rate/qwen_vl_interface': 1.2244479359158895e-06, 'learning_rate/action_model': 1.2244479359158895e-05, 'epoch': 2.38}) +[2026-06-18 01:35:21,875][starVLA.training.train_starvla][INFO] - Step 5772, Loss: {'train/loss': 0.5590597987174988, 'train/grad_norm_pre_clip': 28.785768508911133, 'timing/data': 0.010242803953588009, 'timing/model': 1.0988361416384578, 'learning_rate/qwen_vl_interface': 1.2233003539802512e-06, 'learning_rate/action_model': 1.2233003539802513e-05, 'epoch': 2.38}) +[2026-06-18 01:35:25,715][starVLA.training.train_starvla][INFO] - Step 5773, Loss: {'train/loss': 0.06874187290668488, 'train/grad_norm_pre_clip': 10.474905967712402, 'timing/data': 0.00861311238259077, 'timing/model': 1.1198454396799207, 'learning_rate/qwen_vl_interface': 1.2221536067837807e-06, 'learning_rate/action_model': 1.2221536067837808e-05, 'epoch': 2.38}) +[2026-06-18 01:35:29,318][starVLA.training.train_starvla][INFO] - Step 5774, Loss: {'train/loss': 0.2786858081817627, 'train/grad_norm_pre_clip': 25.87810707092285, 'timing/data': 0.008695684373378754, 'timing/model': 0.9333183169364929, 'learning_rate/qwen_vl_interface': 1.2210076945642025e-06, 'learning_rate/action_model': 1.2210076945642024e-05, 'epoch': 2.38}) +[2026-06-18 01:35:33,203][starVLA.training.train_starvla][INFO] - Step 5775, Loss: {'train/loss': 0.4336608946323395, 'train/grad_norm_pre_clip': 20.944904327392578, 'timing/data': 0.011850342154502869, 'timing/model': 0.9414934553205967, 'learning_rate/qwen_vl_interface': 1.219862617559065e-06, 'learning_rate/action_model': 1.219862617559065e-05, 'epoch': 2.38}) +[2026-06-18 01:35:37,161][starVLA.training.train_starvla][INFO] - Step 5776, Loss: {'train/loss': 0.2251659482717514, 'train/grad_norm_pre_clip': 17.261474609375, 'timing/data': 0.014658345840871334, 'timing/model': 0.9852433605119586, 'learning_rate/qwen_vl_interface': 1.2187183760057431e-06, 'learning_rate/action_model': 1.218718376005743e-05, 'epoch': 2.38}) +[2026-06-18 01:35:40,979][starVLA.training.train_starvla][INFO] - Step 5777, Loss: {'train/loss': 0.4619118273258209, 'train/grad_norm_pre_clip': 30.682886123657227, 'timing/data': 0.009901651181280613, 'timing/model': 0.9352926900610328, 'learning_rate/qwen_vl_interface': 1.2175749701414393e-06, 'learning_rate/action_model': 1.2175749701414394e-05, 'epoch': 2.38}) +[2026-06-18 01:35:44,901][starVLA.training.train_starvla][INFO] - Step 5778, Loss: {'train/loss': 0.21697241067886353, 'train/grad_norm_pre_clip': 16.85349464416504, 'timing/data': 0.010583903640508652, 'timing/model': 0.9360115323215723, 'learning_rate/qwen_vl_interface': 1.216432400203184e-06, 'learning_rate/action_model': 1.2164324002031841e-05, 'epoch': 2.38}) +[2026-06-18 01:35:48,723][starVLA.training.train_starvla][INFO] - Step 5779, Loss: {'train/loss': 0.30289483070373535, 'train/grad_norm_pre_clip': 27.880016326904297, 'timing/data': 0.01143051590770483, 'timing/model': 1.0967866005375981, 'learning_rate/qwen_vl_interface': 1.215290666427832e-06, 'learning_rate/action_model': 1.2152906664278318e-05, 'epoch': 2.38}) +[2026-06-18 01:35:52,497][starVLA.training.train_starvla][INFO] - Step 5780, Loss: {'train/loss': 0.33992186188697815, 'train/grad_norm_pre_clip': 16.94907569885254, 'timing/data': 0.013085170648992062, 'timing/model': 1.094383616000414, 'learning_rate/qwen_vl_interface': 1.214149769052067e-06, 'learning_rate/action_model': 1.214149769052067e-05, 'epoch': 2.38}) +[2026-06-18 01:35:56,139][starVLA.training.train_starvla][INFO] - Step 5781, Loss: {'train/loss': 0.3946991264820099, 'train/grad_norm_pre_clip': 61.25743103027344, 'timing/data': 0.011061817407608032, 'timing/model': 0.9355266476050019, 'learning_rate/qwen_vl_interface': 1.2130097083123971e-06, 'learning_rate/action_model': 1.213009708312397e-05, 'epoch': 2.38}) +[2026-06-18 01:35:59,966][starVLA.training.train_starvla][INFO] - Step 5782, Loss: {'train/loss': 0.3255012631416321, 'train/grad_norm_pre_clip': 20.14971351623535, 'timing/data': 0.009571724571287632, 'timing/model': 0.9820938613265753, 'learning_rate/qwen_vl_interface': 1.2118704844451587e-06, 'learning_rate/action_model': 1.2118704844451586e-05, 'epoch': 2.38}) +[2026-06-18 01:36:03,785][starVLA.training.train_starvla][INFO] - Step 5783, Loss: {'train/loss': 0.2808428704738617, 'train/grad_norm_pre_clip': 27.44172477722168, 'timing/data': 0.009779758751392365, 'timing/model': 0.9386141151189804, 'learning_rate/qwen_vl_interface': 1.2107320976865138e-06, 'learning_rate/action_model': 1.2107320976865139e-05, 'epoch': 2.38}) +[2026-06-18 01:36:07,649][starVLA.training.train_starvla][INFO] - Step 5784, Loss: {'train/loss': 0.38492053747177124, 'train/grad_norm_pre_clip': 24.396928787231445, 'timing/data': 0.009093686006963253, 'timing/model': 0.9390597492456436, 'learning_rate/qwen_vl_interface': 1.209594548272451e-06, 'learning_rate/action_model': 1.209594548272451e-05, 'epoch': 2.38}) +[2026-06-18 01:36:11,442][starVLA.training.train_starvla][INFO] - Step 5785, Loss: {'train/loss': 0.15091904997825623, 'train/grad_norm_pre_clip': 19.05229377746582, 'timing/data': 0.009781780652701855, 'timing/model': 0.9373249392956495, 'learning_rate/qwen_vl_interface': 1.2084578364387869e-06, 'learning_rate/action_model': 1.208457836438787e-05, 'epoch': 2.38}) +[2026-06-18 01:36:15,219][starVLA.training.train_starvla][INFO] - Step 5786, Loss: {'train/loss': 0.0987904891371727, 'train/grad_norm_pre_clip': 15.749860763549805, 'timing/data': 0.010644357651472092, 'timing/model': 1.0977544467896223, 'learning_rate/qwen_vl_interface': 1.2073219624211635e-06, 'learning_rate/action_model': 1.2073219624211635e-05, 'epoch': 2.38}) +[2026-06-18 01:36:18,818][starVLA.training.train_starvla][INFO] - Step 5787, Loss: {'train/loss': 0.12758494913578033, 'train/grad_norm_pre_clip': 17.37699317932129, 'timing/data': 0.007002509199082851, 'timing/model': 0.9385118698701262, 'learning_rate/qwen_vl_interface': 1.2061869264550458e-06, 'learning_rate/action_model': 1.2061869264550457e-05, 'epoch': 2.38}) +[2026-06-18 01:36:22,638][starVLA.training.train_starvla][INFO] - Step 5788, Loss: {'train/loss': 0.2754366397857666, 'train/grad_norm_pre_clip': 22.196609497070312, 'timing/data': 0.01113695465028286, 'timing/model': 0.937345371581614, 'learning_rate/qwen_vl_interface': 1.2050527287757288e-06, 'learning_rate/action_model': 1.2050527287757289e-05, 'epoch': 2.38}) +[2026-06-18 01:36:26,445][starVLA.training.train_starvla][INFO] - Step 5789, Loss: {'train/loss': 0.13613125681877136, 'train/grad_norm_pre_clip': 17.551877975463867, 'timing/data': 0.009726526215672493, 'timing/model': 0.9491366427391768, 'learning_rate/qwen_vl_interface': 1.203919369618335e-06, 'learning_rate/action_model': 1.2039193696183348e-05, 'epoch': 2.38}) +[2026-06-18 01:36:30,313][starVLA.training.train_starvla][INFO] - Step 5790, Loss: {'train/loss': 0.41430002450942993, 'train/grad_norm_pre_clip': 32.06388473510742, 'timing/data': 0.013589393347501755, 'timing/model': 0.956978926435113, 'learning_rate/qwen_vl_interface': 1.202786849217809e-06, 'learning_rate/action_model': 1.202786849217809e-05, 'epoch': 2.38}) +[2026-06-18 01:36:34,107][starVLA.training.train_starvla][INFO] - Step 5791, Loss: {'train/loss': 0.25434741377830505, 'train/grad_norm_pre_clip': 26.158384323120117, 'timing/data': 0.01578586269170046, 'timing/model': 0.9469977598637342, 'learning_rate/qwen_vl_interface': 1.2016551678089247e-06, 'learning_rate/action_model': 1.2016551678089246e-05, 'epoch': 2.38}) +[2026-06-18 01:36:37,934][starVLA.training.train_starvla][INFO] - Step 5792, Loss: {'train/loss': 0.12024669349193573, 'train/grad_norm_pre_clip': 13.732637405395508, 'timing/data': 0.009332866407930851, 'timing/model': 1.141689347103238, 'learning_rate/qwen_vl_interface': 1.2005243256262803e-06, 'learning_rate/action_model': 1.2005243256262802e-05, 'epoch': 2.38}) +[2026-06-18 01:36:41,951][starVLA.training.train_starvla][INFO] - Step 5793, Loss: {'train/loss': 0.17703533172607422, 'train/grad_norm_pre_clip': 15.822315216064453, 'timing/data': 0.012403745204210281, 'timing/model': 1.3441798686981201, 'learning_rate/qwen_vl_interface': 1.1993943229043003e-06, 'learning_rate/action_model': 1.1993943229043003e-05, 'epoch': 2.38}) +[2026-06-18 01:36:45,997][starVLA.training.train_starvla][INFO] - Step 5794, Loss: {'train/loss': 0.5291035175323486, 'train/grad_norm_pre_clip': 25.143930435180664, 'timing/data': 0.009311352856457233, 'timing/model': 1.0255576139315963, 'learning_rate/qwen_vl_interface': 1.1982651598772358e-06, 'learning_rate/action_model': 1.1982651598772356e-05, 'epoch': 2.39}) +[2026-06-18 01:36:50,657][starVLA.training.train_starvla][INFO] - Step 5795, Loss: {'train/loss': 0.12150276452302933, 'train/grad_norm_pre_clip': 15.719765663146973, 'timing/data': 0.009061087854206562, 'timing/model': 0.9429958071559668, 'learning_rate/qwen_vl_interface': 1.197136836779164e-06, 'learning_rate/action_model': 1.1971368367791639e-05, 'epoch': 2.39}) +[2026-06-18 01:36:54,660][starVLA.training.train_starvla][INFO] - Step 5796, Loss: {'train/loss': 0.22303856909275055, 'train/grad_norm_pre_clip': 19.706886291503906, 'timing/data': 0.012564553879201412, 'timing/model': 1.0216515334323049, 'learning_rate/qwen_vl_interface': 1.1960093538439866e-06, 'learning_rate/action_model': 1.1960093538439864e-05, 'epoch': 2.39}) +[2026-06-18 01:36:58,674][starVLA.training.train_starvla][INFO] - Step 5797, Loss: {'train/loss': 0.28206107020378113, 'train/grad_norm_pre_clip': 25.581026077270508, 'timing/data': 0.009675952605903149, 'timing/model': 0.9825856583192945, 'learning_rate/qwen_vl_interface': 1.1948827113054323e-06, 'learning_rate/action_model': 1.1948827113054323e-05, 'epoch': 2.39}) +[2026-06-18 01:37:02,580][starVLA.training.train_starvla][INFO] - Step 5798, Loss: {'train/loss': 0.10522335767745972, 'train/grad_norm_pre_clip': 17.272014617919922, 'timing/data': 0.011403531767427921, 'timing/model': 0.9454352268949151, 'learning_rate/qwen_vl_interface': 1.1937569093970558e-06, 'learning_rate/action_model': 1.1937569093970558e-05, 'epoch': 2.39}) +[2026-06-18 01:37:06,461][starVLA.training.train_starvla][INFO] - Step 5799, Loss: {'train/loss': 0.1415122002363205, 'train/grad_norm_pre_clip': 14.472550392150879, 'timing/data': 0.008533901534974575, 'timing/model': 1.1221149479970336, 'learning_rate/qwen_vl_interface': 1.1926319483522353e-06, 'learning_rate/action_model': 1.1926319483522353e-05, 'epoch': 2.39}) +[2026-06-18 01:39:14,886][starVLA.training.train_starvla][INFO] - Step 5800, Loss: {'train/loss': 0.19067247211933136, 'train/grad_norm_pre_clip': 17.53011703491211, 'eval/loss': 0.4208742082118988, 'eval/latency_6/loss': 0.4208742082118988, 'eval/demon_fire_precision': 0.950547730829421, 'eval/demon_fire_recall': 0.9815772462831286, 'eval/demon_fire_f1': 0.9658133248529178, 'eval/demon_fire_rate_model': 0.8889816360601002, 'eval/demon_fire_rate_teacher': 0.860879243183083, 'eval/demon_left_precision': 0.8951781970649895, 'eval/demon_left_recall': 0.7491228070175439, 'eval/demon_left_f1': 0.8156638013371537, 'eval/demon_left_rate_model': 0.1327212020033389, 'eval/demon_left_rate_teacher': 0.15859766277128548, 'eval/demon_right_precision': 0.8360927152317881, 'eval/demon_right_recall': 0.7757296466973886, 'eval/demon_right_f1': 0.8047808764940239, 'eval/demon_right_rate_model': 0.16805787423483584, 'eval/demon_right_rate_teacher': 0.18113522537562604, 'eval/demon_move_f1': 0.8102223389155887, 'eval/demon_cc_f1': 0.9132032496373594, 'eval/latency_6/demon_fire_precision': 0.950547730829421, 'eval/latency_6/demon_fire_recall': 0.9815772462831286, 'eval/latency_6/demon_fire_f1': 0.9658133248529178, 'eval/latency_6/demon_fire_rate_model': 0.8889816360601002, 'eval/latency_6/demon_fire_rate_teacher': 0.860879243183083, 'eval/latency_6/demon_left_precision': 0.8951781970649895, 'eval/latency_6/demon_left_recall': 0.7491228070175439, 'eval/latency_6/demon_left_f1': 0.8156638013371537, 'eval/latency_6/demon_left_rate_model': 0.1327212020033389, 'eval/latency_6/demon_left_rate_teacher': 0.15859766277128548, 'eval/latency_6/demon_right_precision': 0.8360927152317881, 'eval/latency_6/demon_right_recall': 0.7757296466973886, 'eval/latency_6/demon_right_f1': 0.8047808764940239, 'eval/latency_6/demon_right_rate_model': 0.16805787423483584, 'eval/latency_6/demon_right_rate_teacher': 0.18113522537562604, 'eval/latency_6/demon_move_f1': 0.8102223389155887, 'eval/latency_6/demon_cc_f1': 0.9132032496373594, 'timing/data': 0.009006233885884285, 'timing/model': 1.1086227959021926, 'learning_rate/qwen_vl_interface': 1.1915078284041793e-06, 'learning_rate/action_model': 1.1915078284041793e-05, 'epoch': 2.39}) +[2026-06-18 01:39:14,891][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_5800_state +[2026-06-18 01:39:14,896][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 01:39:21,906][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_5800_state/model.safetensors +[2026-06-18 01:39:37,669][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_5800_state/optimizer.bin +[2026-06-18 01:39:37,671][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_5800_state/sampler.bin +[2026-06-18 01:39:37,672][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_5800_state/sampler_1.bin +[2026-06-18 01:39:37,675][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_5800_state/random_states_0.pkl +[2026-06-18 01:39:37,682][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 5800 +[2026-06-18 01:39:37,689][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 01:39:37,703][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 01:39:41,495][starVLA.training.train_starvla][INFO] - Step 5801, Loss: {'train/loss': 0.28769198060035706, 'train/grad_norm_pre_clip': 17.351701736450195, 'timing/data': 0.009239976294338703, 'timing/model': 0.9305779291316867, 'learning_rate/qwen_vl_interface': 1.1903845497859185e-06, 'learning_rate/action_model': 1.1903845497859184e-05, 'epoch': 2.39}) +[2026-06-18 01:39:45,277][starVLA.training.train_starvla][INFO] - Step 5802, Loss: {'train/loss': 0.4584957957267761, 'train/grad_norm_pre_clip': 25.97316551208496, 'timing/data': 0.016113510355353355, 'timing/model': 1.1125750001519918, 'learning_rate/qwen_vl_interface': 1.1892621127303067e-06, 'learning_rate/action_model': 1.1892621127303066e-05, 'epoch': 2.39}) +[2026-06-18 01:39:48,944][starVLA.training.train_starvla][INFO] - Step 5803, Loss: {'train/loss': 0.23784175515174866, 'train/grad_norm_pre_clip': 20.955842971801758, 'timing/data': 0.012744702398777008, 'timing/model': 0.9242478460073471, 'learning_rate/qwen_vl_interface': 1.1881405174700292e-06, 'learning_rate/action_model': 1.1881405174700291e-05, 'epoch': 2.39}) +[2026-06-18 01:39:52,845][starVLA.training.train_starvla][INFO] - Step 5804, Loss: {'train/loss': 0.1064140796661377, 'train/grad_norm_pre_clip': 16.522876739501953, 'timing/data': 0.007419995963573456, 'timing/model': 0.9556992445141077, 'learning_rate/qwen_vl_interface': 1.1870197642375935e-06, 'learning_rate/action_model': 1.1870197642375935e-05, 'epoch': 2.39}) +[2026-06-18 01:39:56,693][starVLA.training.train_starvla][INFO] - Step 5805, Loss: {'train/loss': 0.18400049209594727, 'train/grad_norm_pre_clip': 19.61565589904785, 'timing/data': 0.009961146861314774, 'timing/model': 0.9490306545048952, 'learning_rate/qwen_vl_interface': 1.1858998532653326e-06, 'learning_rate/action_model': 1.1858998532653325e-05, 'epoch': 2.39}) +[2026-06-18 01:40:00,566][starVLA.training.train_starvla][INFO] - Step 5806, Loss: {'train/loss': 0.24981524050235748, 'train/grad_norm_pre_clip': 16.17668342590332, 'timing/data': 0.010981518775224686, 'timing/model': 0.9377624606713653, 'learning_rate/qwen_vl_interface': 1.1847807847854054e-06, 'learning_rate/action_model': 1.1847807847854054e-05, 'epoch': 2.39}) +[2026-06-18 01:40:04,486][starVLA.training.train_starvla][INFO] - Step 5807, Loss: {'train/loss': 0.36710795760154724, 'train/grad_norm_pre_clip': 23.183486938476562, 'timing/data': 0.010229230858385563, 'timing/model': 1.0065397815778852, 'learning_rate/qwen_vl_interface': 1.1836625590297959e-06, 'learning_rate/action_model': 1.1836625590297958e-05, 'epoch': 2.39}) +[2026-06-18 01:40:08,376][starVLA.training.train_starvla][INFO] - Step 5808, Loss: {'train/loss': 0.484134316444397, 'train/grad_norm_pre_clip': 29.438093185424805, 'timing/data': 0.009443951770663261, 'timing/model': 0.9338404107838869, 'learning_rate/qwen_vl_interface': 1.1825451762303134e-06, 'learning_rate/action_model': 1.1825451762303133e-05, 'epoch': 2.39}) +[2026-06-18 01:40:12,248][starVLA.training.train_starvla][INFO] - Step 5809, Loss: {'train/loss': 0.39694327116012573, 'train/grad_norm_pre_clip': 24.59103012084961, 'timing/data': 0.008369202725589275, 'timing/model': 1.0887647438794374, 'learning_rate/qwen_vl_interface': 1.1814286366185923e-06, 'learning_rate/action_model': 1.1814286366185922e-05, 'epoch': 2.39}) +[2026-06-18 01:40:16,012][starVLA.training.train_starvla][INFO] - Step 5810, Loss: {'train/loss': 0.363625705242157, 'train/grad_norm_pre_clip': 18.537357330322266, 'timing/data': 0.008610869757831097, 'timing/model': 0.9957599844783545, 'learning_rate/qwen_vl_interface': 1.1803129404260932e-06, 'learning_rate/action_model': 1.1803129404260931e-05, 'epoch': 2.39}) +[2026-06-18 01:40:19,853][starVLA.training.train_starvla][INFO] - Step 5811, Loss: {'train/loss': 0.20752719044685364, 'train/grad_norm_pre_clip': 17.25438117980957, 'timing/data': 0.014557618647813797, 'timing/model': 0.9398894356563687, 'learning_rate/qwen_vl_interface': 1.1791980878841008e-06, 'learning_rate/action_model': 1.1791980878841008e-05, 'epoch': 2.39}) +[2026-06-18 01:40:23,782][starVLA.training.train_starvla][INFO] - Step 5812, Loss: {'train/loss': 0.3995406925678253, 'train/grad_norm_pre_clip': 23.469707489013672, 'timing/data': 0.008256466127932072, 'timing/model': 0.9653440257534385, 'learning_rate/qwen_vl_interface': 1.1780840792237246e-06, 'learning_rate/action_model': 1.1780840792237246e-05, 'epoch': 2.39}) +[2026-06-18 01:40:27,562][starVLA.training.train_starvla][INFO] - Step 5813, Loss: {'train/loss': 0.13742992281913757, 'train/grad_norm_pre_clip': 22.64896011352539, 'timing/data': 0.010792691260576248, 'timing/model': 0.9394207000732422, 'learning_rate/qwen_vl_interface': 1.1769709146759002e-06, 'learning_rate/action_model': 1.1769709146759e-05, 'epoch': 2.39}) +[2026-06-18 01:40:31,353][starVLA.training.train_starvla][INFO] - Step 5814, Loss: {'train/loss': 0.07235656678676605, 'train/grad_norm_pre_clip': 9.53342342376709, 'timing/data': 0.008070211857557297, 'timing/model': 0.9401127705350518, 'learning_rate/qwen_vl_interface': 1.175858594471388e-06, 'learning_rate/action_model': 1.175858594471388e-05, 'epoch': 2.39}) +[2026-06-18 01:40:35,355][starVLA.training.train_starvla][INFO] - Step 5815, Loss: {'train/loss': 0.33028554916381836, 'train/grad_norm_pre_clip': 29.94350814819336, 'timing/data': 0.008496853522956371, 'timing/model': 1.2297968519851565, 'learning_rate/qwen_vl_interface': 1.174747118840775e-06, 'learning_rate/action_model': 1.1747471188407749e-05, 'epoch': 2.39}) +[2026-06-18 01:40:39,220][starVLA.training.train_starvla][INFO] - Step 5816, Loss: {'train/loss': 0.4295353293418884, 'train/grad_norm_pre_clip': 30.951494216918945, 'timing/data': 0.008961313404142857, 'timing/model': 1.1099277380853891, 'learning_rate/qwen_vl_interface': 1.173636488014468e-06, 'learning_rate/action_model': 1.173636488014468e-05, 'epoch': 2.39}) +[2026-06-18 01:40:42,879][starVLA.training.train_starvla][INFO] - Step 5817, Loss: {'train/loss': 0.2107706367969513, 'train/grad_norm_pre_clip': 21.9733943939209, 'timing/data': 0.009175353683531284, 'timing/model': 0.9366424418985844, 'learning_rate/qwen_vl_interface': 1.1725267022227024e-06, 'learning_rate/action_model': 1.1725267022227025e-05, 'epoch': 2.39}) +[2026-06-18 01:40:46,734][starVLA.training.train_starvla][INFO] - Step 5818, Loss: {'train/loss': 0.12185989320278168, 'train/grad_norm_pre_clip': 23.84204864501953, 'timing/data': 0.008452994748950005, 'timing/model': 0.9350630976259708, 'learning_rate/qwen_vl_interface': 1.17141776169554e-06, 'learning_rate/action_model': 1.17141776169554e-05, 'epoch': 2.39}) +[2026-06-18 01:40:50,568][starVLA.training.train_starvla][INFO] - Step 5819, Loss: {'train/loss': 0.7210473418235779, 'train/grad_norm_pre_clip': 27.828529357910156, 'timing/data': 0.011140930466353893, 'timing/model': 0.9403457446023822, 'learning_rate/qwen_vl_interface': 1.170309666662864e-06, 'learning_rate/action_model': 1.170309666662864e-05, 'epoch': 2.4}) +[2026-06-18 01:40:54,339][starVLA.training.train_starvla][INFO] - Step 5820, Loss: {'train/loss': 0.3393707573413849, 'train/grad_norm_pre_clip': 49.070655822753906, 'timing/data': 0.010413878597319126, 'timing/model': 0.9553459668532014, 'learning_rate/qwen_vl_interface': 1.1692024173543858e-06, 'learning_rate/action_model': 1.1692024173543859e-05, 'epoch': 2.4}) +[2026-06-18 01:40:58,150][starVLA.training.train_starvla][INFO] - Step 5821, Loss: {'train/loss': 0.2234945446252823, 'train/grad_norm_pre_clip': 22.985397338867188, 'timing/data': 0.007962999865412712, 'timing/model': 0.9420205466449261, 'learning_rate/qwen_vl_interface': 1.168096013999635e-06, 'learning_rate/action_model': 1.168096013999635e-05, 'epoch': 2.4}) +[2026-06-18 01:41:02,018][starVLA.training.train_starvla][INFO] - Step 5822, Loss: {'train/loss': 0.2905001640319824, 'train/grad_norm_pre_clip': 25.52766227722168, 'timing/data': 0.007927224040031433, 'timing/model': 1.1148905865848064, 'learning_rate/qwen_vl_interface': 1.1669904568279742e-06, 'learning_rate/action_model': 1.1669904568279741e-05, 'epoch': 2.4}) +[2026-06-18 01:41:05,653][starVLA.training.train_starvla][INFO] - Step 5823, Loss: {'train/loss': 0.174648255109787, 'train/grad_norm_pre_clip': 15.415782928466797, 'timing/data': 0.010548622347414494, 'timing/model': 0.9431702690199018, 'learning_rate/qwen_vl_interface': 1.165885746068585e-06, 'learning_rate/action_model': 1.1658857460685848e-05, 'epoch': 2.4}) +[2026-06-18 01:41:09,584][starVLA.training.train_starvla][INFO] - Step 5824, Loss: {'train/loss': 0.2559892237186432, 'train/grad_norm_pre_clip': 16.17649269104004, 'timing/data': 0.011722953990101814, 'timing/model': 0.9996617455035448, 'learning_rate/qwen_vl_interface': 1.1647818819504754e-06, 'learning_rate/action_model': 1.1647818819504756e-05, 'epoch': 2.4}) +[2026-06-18 01:41:13,467][starVLA.training.train_starvla][INFO] - Step 5825, Loss: {'train/loss': 0.3612935543060303, 'train/grad_norm_pre_clip': 21.93726348876953, 'timing/data': 0.00879417173564434, 'timing/model': 0.9864584626629949, 'learning_rate/qwen_vl_interface': 1.1636788647024776e-06, 'learning_rate/action_model': 1.1636788647024776e-05, 'epoch': 2.4}) +[2026-06-18 01:41:17,366][starVLA.training.train_starvla][INFO] - Step 5826, Loss: {'train/loss': 0.4021979868412018, 'train/grad_norm_pre_clip': 30.52043914794922, 'timing/data': 0.008749488741159439, 'timing/model': 0.9524265956133604, 'learning_rate/qwen_vl_interface': 1.1625766945532485e-06, 'learning_rate/action_model': 1.1625766945532484e-05, 'epoch': 2.4}) +[2026-06-18 01:41:21,290][starVLA.training.train_starvla][INFO] - Step 5827, Loss: {'train/loss': 0.29600533843040466, 'train/grad_norm_pre_clip': 34.78334426879883, 'timing/data': 0.008745066821575165, 'timing/model': 0.9394383914768696, 'learning_rate/qwen_vl_interface': 1.1614753717312683e-06, 'learning_rate/action_model': 1.1614753717312681e-05, 'epoch': 2.4}) +[2026-06-18 01:41:25,132][starVLA.training.train_starvla][INFO] - Step 5828, Loss: {'train/loss': 0.19822344183921814, 'train/grad_norm_pre_clip': 16.38813591003418, 'timing/data': 0.008018884807825089, 'timing/model': 1.1077257506549358, 'learning_rate/qwen_vl_interface': 1.1603748964648423e-06, 'learning_rate/action_model': 1.1603748964648422e-05, 'epoch': 2.4}) +[2026-06-18 01:41:29,159][starVLA.training.train_starvla][INFO] - Step 5829, Loss: {'train/loss': 0.4703082740306854, 'train/grad_norm_pre_clip': 32.60211181640625, 'timing/data': 0.0139036625623703, 'timing/model': 1.1858751764521003, 'learning_rate/qwen_vl_interface': 1.1592752689821024e-06, 'learning_rate/action_model': 1.1592752689821024e-05, 'epoch': 2.4}) +[2026-06-18 01:41:32,983][starVLA.training.train_starvla][INFO] - Step 5830, Loss: {'train/loss': 0.21467863023281097, 'train/grad_norm_pre_clip': 13.362860679626465, 'timing/data': 0.00786499958485365, 'timing/model': 0.9378684777766466, 'learning_rate/qwen_vl_interface': 1.1581764895110007e-06, 'learning_rate/action_model': 1.1581764895110007e-05, 'epoch': 2.4}) +[2026-06-18 01:41:36,885][starVLA.training.train_starvla][INFO] - Step 5831, Loss: {'train/loss': 0.21969372034072876, 'train/grad_norm_pre_clip': 23.813926696777344, 'timing/data': 0.018555104732513428, 'timing/model': 0.9383537312969565, 'learning_rate/qwen_vl_interface': 1.1570785582793153e-06, 'learning_rate/action_model': 1.1570785582793154e-05, 'epoch': 2.4}) +[2026-06-18 01:41:40,686][starVLA.training.train_starvla][INFO] - Step 5832, Loss: {'train/loss': 0.2124575674533844, 'train/grad_norm_pre_clip': 19.820358276367188, 'timing/data': 0.008696974255144596, 'timing/model': 0.9304822375997901, 'learning_rate/qwen_vl_interface': 1.1559814755146484e-06, 'learning_rate/action_model': 1.1559814755146483e-05, 'epoch': 2.4}) +[2026-06-18 01:41:44,476][starVLA.training.train_starvla][INFO] - Step 5833, Loss: {'train/loss': 0.3166787028312683, 'train/grad_norm_pre_clip': 28.106775283813477, 'timing/data': 0.012357540428638458, 'timing/model': 0.9388113208115101, 'learning_rate/qwen_vl_interface': 1.154885241444428e-06, 'learning_rate/action_model': 1.1548852414444277e-05, 'epoch': 2.4}) +[2026-06-18 01:41:48,400][starVLA.training.train_starvla][INFO] - Step 5834, Loss: {'train/loss': 0.10868826508522034, 'train/grad_norm_pre_clip': 14.600340843200684, 'timing/data': 0.012974469922482967, 'timing/model': 1.0339104980230331, 'learning_rate/qwen_vl_interface': 1.1537898562959039e-06, 'learning_rate/action_model': 1.1537898562959037e-05, 'epoch': 2.4}) +[2026-06-18 01:41:52,234][starVLA.training.train_starvla][INFO] - Step 5835, Loss: {'train/loss': 0.6074311137199402, 'train/grad_norm_pre_clip': 29.909080505371094, 'timing/data': 0.008230636827647686, 'timing/model': 1.1260327948257327, 'learning_rate/qwen_vl_interface': 1.152695320296149e-06, 'learning_rate/action_model': 1.1526953202961488e-05, 'epoch': 2.4}) +[2026-06-18 01:41:56,034][starVLA.training.train_starvla][INFO] - Step 5836, Loss: {'train/loss': 0.12910115718841553, 'train/grad_norm_pre_clip': 20.998828887939453, 'timing/data': 0.011005080305039883, 'timing/model': 1.1112351585179567, 'learning_rate/qwen_vl_interface': 1.1516016336720635e-06, 'learning_rate/action_model': 1.1516016336720633e-05, 'epoch': 2.4}) +[2026-06-18 01:41:59,691][starVLA.training.train_starvla][INFO] - Step 5837, Loss: {'train/loss': 0.1528848111629486, 'train/grad_norm_pre_clip': 17.79932975769043, 'timing/data': 0.015091774053871632, 'timing/model': 0.9461069526150823, 'learning_rate/qwen_vl_interface': 1.1505087966503697e-06, 'learning_rate/action_model': 1.1505087966503697e-05, 'epoch': 2.4}) +[2026-06-18 01:42:03,570][starVLA.training.train_starvla][INFO] - Step 5838, Loss: {'train/loss': 0.057528119534254074, 'train/grad_norm_pre_clip': 8.872235298156738, 'timing/data': 0.012045600451529026, 'timing/model': 0.9634853769093752, 'learning_rate/qwen_vl_interface': 1.1494168094576138e-06, 'learning_rate/action_model': 1.1494168094576138e-05, 'epoch': 2.4}) +[2026-06-18 01:42:07,453][starVLA.training.train_starvla][INFO] - Step 5839, Loss: {'train/loss': 0.07959695160388947, 'train/grad_norm_pre_clip': 11.569756507873535, 'timing/data': 0.00765970628708601, 'timing/model': 0.9450977919623256, 'learning_rate/qwen_vl_interface': 1.1483256723201656e-06, 'learning_rate/action_model': 1.1483256723201655e-05, 'epoch': 2.4}) +[2026-06-18 01:42:12,157][starVLA.training.train_starvla][INFO] - Step 5840, Loss: {'train/loss': 0.2795281410217285, 'train/grad_norm_pre_clip': 24.5988712310791, 'timing/data': 0.007828433066606522, 'timing/model': 0.9631314296275377, 'learning_rate/qwen_vl_interface': 1.1472353854642187e-06, 'learning_rate/action_model': 1.1472353854642187e-05, 'epoch': 2.4}) +[2026-06-18 01:42:17,105][starVLA.training.train_starvla][INFO] - Step 5841, Loss: {'train/loss': 0.32620054483413696, 'train/grad_norm_pre_clip': 22.242774963378906, 'timing/data': 0.010114455595612526, 'timing/model': 0.9311254350468516, 'learning_rate/qwen_vl_interface': 1.1461459491157914e-06, 'learning_rate/action_model': 1.1461459491157914e-05, 'epoch': 2.4}) +[2026-06-18 01:42:20,927][starVLA.training.train_starvla][INFO] - Step 5842, Loss: {'train/loss': 0.26009640097618103, 'train/grad_norm_pre_clip': 22.837921142578125, 'timing/data': 0.011174321174621582, 'timing/model': 1.125607960857451, 'learning_rate/qwen_vl_interface': 1.145057363500725e-06, 'learning_rate/action_model': 1.1450573635007251e-05, 'epoch': 2.4}) +[2026-06-18 01:42:24,753][starVLA.training.train_starvla][INFO] - Step 5843, Loss: {'train/loss': 0.2669173777103424, 'train/grad_norm_pre_clip': 17.344186782836914, 'timing/data': 0.011687121354043484, 'timing/model': 0.9356304202228785, 'learning_rate/qwen_vl_interface': 1.1439696288446845e-06, 'learning_rate/action_model': 1.1439696288446844e-05, 'epoch': 2.41}) +[2026-06-18 01:42:28,575][starVLA.training.train_starvla][INFO] - Step 5844, Loss: {'train/loss': 0.3515397906303406, 'train/grad_norm_pre_clip': 28.14845085144043, 'timing/data': 0.00947919487953186, 'timing/model': 0.9371316758915782, 'learning_rate/qwen_vl_interface': 1.1428827453731575e-06, 'learning_rate/action_model': 1.1428827453731574e-05, 'epoch': 2.41}) +[2026-06-18 01:42:32,450][starVLA.training.train_starvla][INFO] - Step 5845, Loss: {'train/loss': 0.07462135702371597, 'train/grad_norm_pre_clip': 15.870665550231934, 'timing/data': 0.008689767681062222, 'timing/model': 0.9470478091388941, 'learning_rate/qwen_vl_interface': 1.1417967133114575e-06, 'learning_rate/action_model': 1.1417967133114573e-05, 'epoch': 2.41}) +[2026-06-18 01:42:40,286][starVLA.training.train_starvla][INFO] - Step 5846, Loss: {'train/loss': 0.46471744775772095, 'train/grad_norm_pre_clip': 32.74592590332031, 'timing/data': 0.008253085426986217, 'timing/model': 2.0186549676582217, 'learning_rate/qwen_vl_interface': 1.140711532884718e-06, 'learning_rate/action_model': 1.140711532884718e-05, 'epoch': 2.41}) +[2026-06-18 01:42:44,160][starVLA.training.train_starvla][INFO] - Step 5847, Loss: {'train/loss': 0.16192083060741425, 'train/grad_norm_pre_clip': 15.692750930786133, 'timing/data': 0.007261640392243862, 'timing/model': 0.9437187854200602, 'learning_rate/qwen_vl_interface': 1.1396272043179004e-06, 'learning_rate/action_model': 1.1396272043179005e-05, 'epoch': 2.41}) +[2026-06-18 01:42:48,241][starVLA.training.train_starvla][INFO] - Step 5848, Loss: {'train/loss': 0.3549126386642456, 'train/grad_norm_pre_clip': 22.7482967376709, 'timing/data': 0.01021695788949728, 'timing/model': 0.939093622379005, 'learning_rate/qwen_vl_interface': 1.138543727835788e-06, 'learning_rate/action_model': 1.138543727835788e-05, 'epoch': 2.41}) +[2026-06-18 01:42:52,220][starVLA.training.train_starvla][INFO] - Step 5849, Loss: {'train/loss': 0.04936746507883072, 'train/grad_norm_pre_clip': 7.592046737670898, 'timing/data': 0.010352926328778267, 'timing/model': 1.1103982301428914, 'learning_rate/qwen_vl_interface': 1.1374611036629832e-06, 'learning_rate/action_model': 1.1374611036629831e-05, 'epoch': 2.41}) +[2026-06-18 01:42:55,830][starVLA.training.train_starvla][INFO] - Step 5850, Loss: {'train/loss': 0.3058415949344635, 'train/grad_norm_pre_clip': 24.10179901123047, 'timing/data': 0.009344101883471012, 'timing/model': 0.9363916823640466, 'learning_rate/qwen_vl_interface': 1.136379332023916e-06, 'learning_rate/action_model': 1.1363793320239161e-05, 'epoch': 2.41}) +[2026-06-18 01:42:59,627][starVLA.training.train_starvla][INFO] - Step 5851, Loss: {'train/loss': 0.27065008878707886, 'train/grad_norm_pre_clip': 18.404645919799805, 'timing/data': 0.00956039223819971, 'timing/model': 0.9321491476148367, 'learning_rate/qwen_vl_interface': 1.1352984131428413e-06, 'learning_rate/action_model': 1.1352984131428413e-05, 'epoch': 2.41}) +[2026-06-18 01:43:03,606][starVLA.training.train_starvla][INFO] - Step 5852, Loss: {'train/loss': 0.0906340703368187, 'train/grad_norm_pre_clip': 12.089275360107422, 'timing/data': 0.008970426395535469, 'timing/model': 1.0256320042535663, 'learning_rate/qwen_vl_interface': 1.1342183472438329e-06, 'learning_rate/action_model': 1.1342183472438328e-05, 'epoch': 2.41}) +[2026-06-18 01:43:07,619][starVLA.training.train_starvla][INFO] - Step 5853, Loss: {'train/loss': 0.3276035785675049, 'train/grad_norm_pre_clip': 22.056711196899414, 'timing/data': 0.01093726884573698, 'timing/model': 1.0353073216974735, 'learning_rate/qwen_vl_interface': 1.1331391345507905e-06, 'learning_rate/action_model': 1.1331391345507905e-05, 'epoch': 2.41}) +[2026-06-18 01:43:11,471][starVLA.training.train_starvla][INFO] - Step 5854, Loss: {'train/loss': 0.10919462144374847, 'train/grad_norm_pre_clip': 17.148340225219727, 'timing/data': 0.009359602816402912, 'timing/model': 0.9456566330045462, 'learning_rate/qwen_vl_interface': 1.1320607752874341e-06, 'learning_rate/action_model': 1.1320607752874341e-05, 'epoch': 2.41}) +[2026-06-18 01:43:15,351][starVLA.training.train_starvla][INFO] - Step 5855, Loss: {'train/loss': 0.10237350314855576, 'train/grad_norm_pre_clip': 11.88775634765625, 'timing/data': 0.010589072480797768, 'timing/model': 1.11615121178329, 'learning_rate/qwen_vl_interface': 1.1309832696773104e-06, 'learning_rate/action_model': 1.1309832696773104e-05, 'epoch': 2.41}) +[2026-06-18 01:43:19,259][starVLA.training.train_starvla][INFO] - Step 5856, Loss: {'train/loss': 0.46281716227531433, 'train/grad_norm_pre_clip': 30.976911544799805, 'timing/data': 0.0094473697245121, 'timing/model': 1.112916654907167, 'learning_rate/qwen_vl_interface': 1.1299066179437876e-06, 'learning_rate/action_model': 1.1299066179437876e-05, 'epoch': 2.41}) +[2026-06-18 01:43:22,968][starVLA.training.train_starvla][INFO] - Step 5857, Loss: {'train/loss': 0.18225327134132385, 'train/grad_norm_pre_clip': 16.069583892822266, 'timing/data': 0.008398501202464104, 'timing/model': 0.9519375581294298, 'learning_rate/qwen_vl_interface': 1.1288308203100565e-06, 'learning_rate/action_model': 1.1288308203100564e-05, 'epoch': 2.41}) +[2026-06-18 01:43:27,157][starVLA.training.train_starvla][INFO] - Step 5858, Loss: {'train/loss': 0.10851725190877914, 'train/grad_norm_pre_clip': 12.140972137451172, 'timing/data': 0.009788252413272858, 'timing/model': 0.9307888364419341, 'learning_rate/qwen_vl_interface': 1.12775587699913e-06, 'learning_rate/action_model': 1.12775587699913e-05, 'epoch': 2.41}) +[2026-06-18 01:43:30,953][starVLA.training.train_starvla][INFO] - Step 5859, Loss: {'train/loss': 0.5255767107009888, 'train/grad_norm_pre_clip': 36.545257568359375, 'timing/data': 0.008792010135948658, 'timing/model': 0.9331729570403695, 'learning_rate/qwen_vl_interface': 1.126681788233847e-06, 'learning_rate/action_model': 1.1266817882338468e-05, 'epoch': 2.41}) +[2026-06-18 01:43:34,733][starVLA.training.train_starvla][INFO] - Step 5860, Loss: {'train/loss': 0.19365355372428894, 'train/grad_norm_pre_clip': 19.195005416870117, 'timing/data': 0.011679084971547127, 'timing/model': 0.9477470368146896, 'learning_rate/qwen_vl_interface': 1.1256085542368655e-06, 'learning_rate/action_model': 1.1256085542368655e-05, 'epoch': 2.41}) +[2026-06-18 01:43:38,643][starVLA.training.train_starvla][INFO] - Step 5861, Loss: {'train/loss': 0.35377123951911926, 'train/grad_norm_pre_clip': 24.99639129638672, 'timing/data': 0.011639428324997425, 'timing/model': 0.9447476752102375, 'learning_rate/qwen_vl_interface': 1.1245361752306676e-06, 'learning_rate/action_model': 1.1245361752306677e-05, 'epoch': 2.41}) +[2026-06-18 01:43:42,436][starVLA.training.train_starvla][INFO] - Step 5862, Loss: {'train/loss': 0.1840781420469284, 'train/grad_norm_pre_clip': 15.6964693069458, 'timing/data': 0.008841774426400661, 'timing/model': 1.1086910059675574, 'learning_rate/qwen_vl_interface': 1.1234646514375625e-06, 'learning_rate/action_model': 1.1234646514375624e-05, 'epoch': 2.41}) +[2026-06-18 01:43:46,197][starVLA.training.train_starvla][INFO] - Step 5863, Loss: {'train/loss': 0.27775654196739197, 'train/grad_norm_pre_clip': 30.098691940307617, 'timing/data': 0.008150780573487282, 'timing/model': 1.084568103775382, 'learning_rate/qwen_vl_interface': 1.122393983079674e-06, 'learning_rate/action_model': 1.122393983079674e-05, 'epoch': 2.41}) +[2026-06-18 01:43:49,829][starVLA.training.train_starvla][INFO] - Step 5864, Loss: {'train/loss': 0.30634474754333496, 'train/grad_norm_pre_clip': 24.941843032836914, 'timing/data': 0.010883158072829247, 'timing/model': 0.9296093629673123, 'learning_rate/qwen_vl_interface': 1.1213241703789534e-06, 'learning_rate/action_model': 1.1213241703789534e-05, 'epoch': 2.41}) +[2026-06-18 01:43:53,766][starVLA.training.train_starvla][INFO] - Step 5865, Loss: {'train/loss': 0.19929586350917816, 'train/grad_norm_pre_clip': 29.07549476623535, 'timing/data': 0.013065043836832047, 'timing/model': 0.9413256142288446, 'learning_rate/qwen_vl_interface': 1.1202552135571759e-06, 'learning_rate/action_model': 1.1202552135571759e-05, 'epoch': 2.41}) +[2026-06-18 01:43:57,700][starVLA.training.train_starvla][INFO] - Step 5866, Loss: {'train/loss': 0.21920493245124817, 'train/grad_norm_pre_clip': 22.317344665527344, 'timing/data': 0.006896705366671085, 'timing/model': 0.9516161307692528, 'learning_rate/qwen_vl_interface': 1.1191871128359367e-06, 'learning_rate/action_model': 1.1191871128359365e-05, 'epoch': 2.41}) +[2026-06-18 01:44:01,461][starVLA.training.train_starvla][INFO] - Step 5867, Loss: {'train/loss': 0.15220008790493011, 'train/grad_norm_pre_clip': 18.81779670715332, 'timing/data': 0.014561330899596214, 'timing/model': 0.9312820127233863, 'learning_rate/qwen_vl_interface': 1.1181198684366548e-06, 'learning_rate/action_model': 1.1181198684366548e-05, 'epoch': 2.42}) +[2026-06-18 01:44:05,470][starVLA.training.train_starvla][INFO] - Step 5868, Loss: {'train/loss': 0.22944875061511993, 'train/grad_norm_pre_clip': 19.923017501831055, 'timing/data': 0.008033243007957935, 'timing/model': 1.0260438900440931, 'learning_rate/qwen_vl_interface': 1.117053480580568e-06, 'learning_rate/action_model': 1.1170534805805679e-05, 'epoch': 2.42}) +[2026-06-18 01:44:11,598][starVLA.training.train_starvla][INFO] - Step 5869, Loss: {'train/loss': 0.15905018150806427, 'train/grad_norm_pre_clip': 17.74827003479004, 'timing/data': 0.009850431233644485, 'timing/model': 1.2226165849715471, 'learning_rate/qwen_vl_interface': 1.1159879494887429e-06, 'learning_rate/action_model': 1.115987949488743e-05, 'epoch': 2.42}) +[2026-06-18 01:44:15,943][starVLA.training.train_starvla][INFO] - Step 5870, Loss: {'train/loss': 0.34653714299201965, 'train/grad_norm_pre_clip': 21.897010803222656, 'timing/data': 0.008076862432062626, 'timing/model': 0.9350670920684934, 'learning_rate/qwen_vl_interface': 1.1149232753820645e-06, 'learning_rate/action_model': 1.1149232753820647e-05, 'epoch': 2.42}) +[2026-06-18 01:44:19,798][starVLA.training.train_starvla][INFO] - Step 5871, Loss: {'train/loss': 0.23688200116157532, 'train/grad_norm_pre_clip': 19.784076690673828, 'timing/data': 0.009669305756688118, 'timing/model': 0.9545383239164948, 'learning_rate/qwen_vl_interface': 1.1138594584812406e-06, 'learning_rate/action_model': 1.1138594584812406e-05, 'epoch': 2.42}) +[2026-06-18 01:44:23,591][starVLA.training.train_starvla][INFO] - Step 5872, Loss: {'train/loss': 0.3555428981781006, 'train/grad_norm_pre_clip': 27.509788513183594, 'timing/data': 0.01069083996117115, 'timing/model': 0.9413038827478886, 'learning_rate/qwen_vl_interface': 1.1127964990068017e-06, 'learning_rate/action_model': 1.1127964990068017e-05, 'epoch': 2.42}) +[2026-06-18 01:44:27,456][starVLA.training.train_starvla][INFO] - Step 5873, Loss: {'train/loss': 0.1857936978340149, 'train/grad_norm_pre_clip': 20.148874282836914, 'timing/data': 0.010001463815569878, 'timing/model': 0.9424117803573608, 'learning_rate/qwen_vl_interface': 1.1117343971791005e-06, 'learning_rate/action_model': 1.1117343971791006e-05, 'epoch': 2.42}) +[2026-06-18 01:44:31,350][starVLA.training.train_starvla][INFO] - Step 5874, Loss: {'train/loss': 0.259113073348999, 'train/grad_norm_pre_clip': 22.74431800842285, 'timing/data': 0.008962402120232582, 'timing/model': 0.9342889217659831, 'learning_rate/qwen_vl_interface': 1.1106731532183114e-06, 'learning_rate/action_model': 1.1106731532183114e-05, 'epoch': 2.42}) +[2026-06-18 01:44:35,146][starVLA.training.train_starvla][INFO] - Step 5875, Loss: {'train/loss': 0.504662036895752, 'train/grad_norm_pre_clip': 40.491119384765625, 'timing/data': 0.010756824165582657, 'timing/model': 1.1127195470035076, 'learning_rate/qwen_vl_interface': 1.1096127673444309e-06, 'learning_rate/action_model': 1.109612767344431e-05, 'epoch': 2.42}) +[2026-06-18 01:44:39,032][starVLA.training.train_starvla][INFO] - Step 5876, Loss: {'train/loss': 0.1262488067150116, 'train/grad_norm_pre_clip': 14.112089157104492, 'timing/data': 0.010311811231076717, 'timing/model': 1.111997497268021, 'learning_rate/qwen_vl_interface': 1.108553239777281e-06, 'learning_rate/action_model': 1.108553239777281e-05, 'epoch': 2.42}) +[2026-06-18 01:44:42,698][starVLA.training.train_starvla][INFO] - Step 5877, Loss: {'train/loss': 0.28576958179473877, 'train/grad_norm_pre_clip': 27.80377769470215, 'timing/data': 0.009748089127242565, 'timing/model': 0.9550890857353806, 'learning_rate/qwen_vl_interface': 1.1074945707365004e-06, 'learning_rate/action_model': 1.1074945707365003e-05, 'epoch': 2.42}) +[2026-06-18 01:44:46,542][starVLA.training.train_starvla][INFO] - Step 5878, Loss: {'train/loss': 0.23945170640945435, 'train/grad_norm_pre_clip': 20.952821731567383, 'timing/data': 0.013698850758373737, 'timing/model': 0.9518941296264529, 'learning_rate/qwen_vl_interface': 1.1064367604415522e-06, 'learning_rate/action_model': 1.1064367604415521e-05, 'epoch': 2.42}) +[2026-06-18 01:44:50,377][starVLA.training.train_starvla][INFO] - Step 5879, Loss: {'train/loss': 0.14654597640037537, 'train/grad_norm_pre_clip': 13.154553413391113, 'timing/data': 0.014148655347526073, 'timing/model': 0.9356937492266297, 'learning_rate/qwen_vl_interface': 1.105379809111721e-06, 'learning_rate/action_model': 1.105379809111721e-05, 'epoch': 2.42}) +[2026-06-18 01:44:54,178][starVLA.training.train_starvla][INFO] - Step 5880, Loss: {'train/loss': 0.2840264141559601, 'train/grad_norm_pre_clip': 25.92938804626465, 'timing/data': 0.011922715231776237, 'timing/model': 0.9314549583941698, 'learning_rate/qwen_vl_interface': 1.1043237169661167e-06, 'learning_rate/action_model': 1.1043237169661166e-05, 'epoch': 2.42}) +[2026-06-18 01:44:57,996][starVLA.training.train_starvla][INFO] - Step 5881, Loss: {'train/loss': 0.46798914670944214, 'train/grad_norm_pre_clip': 23.175634384155273, 'timing/data': 0.010077254846692085, 'timing/model': 0.9384258734062314, 'learning_rate/qwen_vl_interface': 1.103268484223667e-06, 'learning_rate/action_model': 1.1032684842236669e-05, 'epoch': 2.42}) +[2026-06-18 01:45:01,842][starVLA.training.train_starvla][INFO] - Step 5882, Loss: {'train/loss': 0.2858019173145294, 'train/grad_norm_pre_clip': 30.14257049560547, 'timing/data': 0.017702873796224594, 'timing/model': 1.120860118418932, 'learning_rate/qwen_vl_interface': 1.102214111103123e-06, 'learning_rate/action_model': 1.102214111103123e-05, 'epoch': 2.42}) +[2026-06-18 01:45:05,711][starVLA.training.train_starvla][INFO] - Step 5883, Loss: {'train/loss': 0.38109731674194336, 'train/grad_norm_pre_clip': 25.729066848754883, 'timing/data': 0.014542968943715096, 'timing/model': 1.1117156939581037, 'learning_rate/qwen_vl_interface': 1.1011605978230552e-06, 'learning_rate/action_model': 1.1011605978230551e-05, 'epoch': 2.42}) +[2026-06-18 01:45:09,431][starVLA.training.train_starvla][INFO] - Step 5884, Loss: {'train/loss': 0.5015595555305481, 'train/grad_norm_pre_clip': 26.542957305908203, 'timing/data': 0.010106416419148445, 'timing/model': 0.9715333199128509, 'learning_rate/qwen_vl_interface': 1.100107944601861e-06, 'learning_rate/action_model': 1.100107944601861e-05, 'epoch': 2.42}) +[2026-06-18 01:45:13,370][starVLA.training.train_starvla][INFO] - Step 5885, Loss: {'train/loss': 0.317230761051178, 'train/grad_norm_pre_clip': 20.4400577545166, 'timing/data': 0.01012630295008421, 'timing/model': 0.9364491626620293, 'learning_rate/qwen_vl_interface': 1.0990561516577552e-06, 'learning_rate/action_model': 1.0990561516577552e-05, 'epoch': 2.42}) +[2026-06-18 01:45:17,291][starVLA.training.train_starvla][INFO] - Step 5886, Loss: {'train/loss': 0.1221328154206276, 'train/grad_norm_pre_clip': 24.977603912353516, 'timing/data': 0.008538597263395786, 'timing/model': 1.003178121522069, 'learning_rate/qwen_vl_interface': 1.0980052192087753e-06, 'learning_rate/action_model': 1.0980052192087754e-05, 'epoch': 2.42}) +[2026-06-18 01:45:21,097][starVLA.training.train_starvla][INFO] - Step 5887, Loss: {'train/loss': 0.22236377000808716, 'train/grad_norm_pre_clip': 19.53268814086914, 'timing/data': 0.011449435725808144, 'timing/model': 0.9724451908841729, 'learning_rate/qwen_vl_interface': 1.096955147472782e-06, 'learning_rate/action_model': 1.0969551474727821e-05, 'epoch': 2.42}) +[2026-06-18 01:45:24,958][starVLA.training.train_starvla][INFO] - Step 5888, Loss: {'train/loss': 0.25983431935310364, 'train/grad_norm_pre_clip': 23.45456886291504, 'timing/data': 0.01006771158427, 'timing/model': 0.9332617344334722, 'learning_rate/qwen_vl_interface': 1.095905936667455e-06, 'learning_rate/action_model': 1.0959059366674552e-05, 'epoch': 2.42}) +[2026-06-18 01:45:28,738][starVLA.training.train_starvla][INFO] - Step 5889, Loss: {'train/loss': 0.2410920113325119, 'train/grad_norm_pre_clip': 22.585420608520508, 'timing/data': 0.010347641073167324, 'timing/model': 1.0986621594056487, 'learning_rate/qwen_vl_interface': 1.0948575870102977e-06, 'learning_rate/action_model': 1.0948575870102978e-05, 'epoch': 2.42}) +[2026-06-18 01:45:32,504][starVLA.training.train_starvla][INFO] - Step 5890, Loss: {'train/loss': 0.240388885140419, 'train/grad_norm_pre_clip': 17.268402099609375, 'timing/data': 0.009099372662603855, 'timing/model': 1.024705197662115, 'learning_rate/qwen_vl_interface': 1.0938100987186325e-06, 'learning_rate/action_model': 1.0938100987186323e-05, 'epoch': 2.42}) +[2026-06-18 01:45:36,339][starVLA.training.train_starvla][INFO] - Step 5891, Loss: {'train/loss': 0.441714346408844, 'train/grad_norm_pre_clip': 26.788591384887695, 'timing/data': 0.008074485696852207, 'timing/model': 0.940068376250565, 'learning_rate/qwen_vl_interface': 1.092763472009608e-06, 'learning_rate/action_model': 1.092763472009608e-05, 'epoch': 2.42}) +[2026-06-18 01:45:40,142][starVLA.training.train_starvla][INFO] - Step 5892, Loss: {'train/loss': 0.371451199054718, 'train/grad_norm_pre_clip': 27.732954025268555, 'timing/data': 0.009531867690384388, 'timing/model': 0.9449815340340137, 'learning_rate/qwen_vl_interface': 1.0917177071001884e-06, 'learning_rate/action_model': 1.0917177071001884e-05, 'epoch': 2.43}) +[2026-06-18 01:45:44,059][starVLA.training.train_starvla][INFO] - Step 5893, Loss: {'train/loss': 0.3577524721622467, 'train/grad_norm_pre_clip': 21.747833251953125, 'timing/data': 0.013893907889723778, 'timing/model': 0.943706170655787, 'learning_rate/qwen_vl_interface': 1.0906728042071627e-06, 'learning_rate/action_model': 1.0906728042071627e-05, 'epoch': 2.43}) +[2026-06-18 01:45:48,005][starVLA.training.train_starvla][INFO] - Step 5894, Loss: {'train/loss': 0.06104535982012749, 'train/grad_norm_pre_clip': 8.763802528381348, 'timing/data': 0.010410914197564125, 'timing/model': 0.9852862693369389, 'learning_rate/qwen_vl_interface': 1.089628763547139e-06, 'learning_rate/action_model': 1.0896287635471389e-05, 'epoch': 2.43}) +[2026-06-18 01:45:51,879][starVLA.training.train_starvla][INFO] - Step 5895, Loss: {'train/loss': 0.2408805787563324, 'train/grad_norm_pre_clip': 217.4981689453125, 'timing/data': 0.010687123984098434, 'timing/model': 1.0021799067035317, 'learning_rate/qwen_vl_interface': 1.0885855853365505e-06, 'learning_rate/action_model': 1.0885855853365503e-05, 'epoch': 2.43}) +[2026-06-18 01:45:55,817][starVLA.training.train_starvla][INFO] - Step 5896, Loss: {'train/loss': 0.16434484720230103, 'train/grad_norm_pre_clip': 11.078182220458984, 'timing/data': 0.007974879816174507, 'timing/model': 1.1065363017842174, 'learning_rate/qwen_vl_interface': 1.0875432697916492e-06, 'learning_rate/action_model': 1.087543269791649e-05, 'epoch': 2.43}) +[2026-06-18 01:45:59,466][starVLA.training.train_starvla][INFO] - Step 5897, Loss: {'train/loss': 0.28176361322402954, 'train/grad_norm_pre_clip': 28.670085906982422, 'timing/data': 0.01443004235625267, 'timing/model': 0.9381710095331073, 'learning_rate/qwen_vl_interface': 1.0865018171285047e-06, 'learning_rate/action_model': 1.0865018171285047e-05, 'epoch': 2.43}) +[2026-06-18 01:46:03,283][starVLA.training.train_starvla][INFO] - Step 5898, Loss: {'train/loss': 0.41366875171661377, 'train/grad_norm_pre_clip': 46.184207916259766, 'timing/data': 0.013369116932153702, 'timing/model': 0.9295680588111281, 'learning_rate/qwen_vl_interface': 1.0854612275630146e-06, 'learning_rate/action_model': 1.0854612275630146e-05, 'epoch': 2.43}) +[2026-06-18 01:46:07,217][starVLA.training.train_starvla][INFO] - Step 5899, Loss: {'train/loss': 0.2984192669391632, 'train/grad_norm_pre_clip': 24.41347312927246, 'timing/data': 0.011120641604065895, 'timing/model': 1.0413587763905525, 'learning_rate/qwen_vl_interface': 1.0844215013108936e-06, 'learning_rate/action_model': 1.0844215013108936e-05, 'epoch': 2.43}) +[2026-06-18 01:48:17,640][starVLA.training.train_starvla][INFO] - Step 5900, Loss: {'train/loss': 0.25781452655792236, 'train/grad_norm_pre_clip': 22.69462776184082, 'eval/loss': 0.4232144057750702, 'eval/latency_6/loss': 0.4232144057750702, 'eval/demon_fire_precision': 0.9488459139114162, 'eval/demon_fire_recall': 0.9831932773109243, 'eval/demon_fire_f1': 0.9657142857142857, 'eval/demon_fire_rate_model': 0.8920422927100723, 'eval/demon_fire_rate_teacher': 0.860879243183083, 'eval/demon_left_precision': 0.9017094017094017, 'eval/demon_left_recall': 0.7403508771929824, 'eval/demon_left_f1': 0.8131021194605009, 'eval/demon_left_rate_model': 0.1302170283806344, 'eval/demon_left_rate_teacher': 0.15859766277128548, 'eval/demon_right_precision': 0.8233438485804416, 'eval/demon_right_recall': 0.8018433179723502, 'eval/demon_right_f1': 0.8124513618677042, 'eval/demon_right_rate_model': 0.17640511964385086, 'eval/demon_right_rate_teacher': 0.18113522537562604, 'eval/demon_move_f1': 0.8127767406641025, 'eval/demon_cc_f1': 0.9141126103004904, 'eval/latency_6/demon_fire_precision': 0.9488459139114162, 'eval/latency_6/demon_fire_recall': 0.9831932773109243, 'eval/latency_6/demon_fire_f1': 0.9657142857142857, 'eval/latency_6/demon_fire_rate_model': 0.8920422927100723, 'eval/latency_6/demon_fire_rate_teacher': 0.860879243183083, 'eval/latency_6/demon_left_precision': 0.9017094017094017, 'eval/latency_6/demon_left_recall': 0.7403508771929824, 'eval/latency_6/demon_left_f1': 0.8131021194605009, 'eval/latency_6/demon_left_rate_model': 0.1302170283806344, 'eval/latency_6/demon_left_rate_teacher': 0.15859766277128548, 'eval/latency_6/demon_right_precision': 0.8233438485804416, 'eval/latency_6/demon_right_recall': 0.8018433179723502, 'eval/latency_6/demon_right_f1': 0.8124513618677042, 'eval/latency_6/demon_right_rate_model': 0.17640511964385086, 'eval/latency_6/demon_right_rate_teacher': 0.18113522537562604, 'eval/latency_6/demon_move_f1': 0.8127767406641025, 'eval/latency_6/demon_cc_f1': 0.9141126103004904, 'timing/data': 0.009127517230808735, 'timing/model': 0.9387392196804285, 'learning_rate/qwen_vl_interface': 1.0833826385876772e-06, 'learning_rate/action_model': 1.0833826385876771e-05, 'epoch': 2.43}) +[2026-06-18 01:48:17,657][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_5900_state +[2026-06-18 01:48:17,662][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 01:48:24,526][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_5900_state/model.safetensors +[2026-06-18 01:48:39,980][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_5900_state/optimizer.bin +[2026-06-18 01:48:39,981][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_5900_state/sampler.bin +[2026-06-18 01:48:39,981][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_5900_state/sampler_1.bin +[2026-06-18 01:48:39,983][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_5900_state/random_states_0.pkl +[2026-06-18 01:48:39,986][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 5900 +[2026-06-18 01:48:39,986][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 01:48:40,001][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 01:48:43,892][starVLA.training.train_starvla][INFO] - Step 5901, Loss: {'train/loss': 0.17408226430416107, 'train/grad_norm_pre_clip': 15.511656761169434, 'timing/data': 0.011869976297020912, 'timing/model': 0.9298980757594109, 'learning_rate/qwen_vl_interface': 1.0823446396087239e-06, 'learning_rate/action_model': 1.0823446396087239e-05, 'epoch': 2.43}) +[2026-06-18 01:48:47,623][starVLA.training.train_starvla][INFO] - Step 5902, Loss: {'train/loss': 0.18588724732398987, 'train/grad_norm_pre_clip': 17.011131286621094, 'timing/data': 0.009952343069016933, 'timing/model': 0.9255311638116837, 'learning_rate/qwen_vl_interface': 1.0813075045892116e-06, 'learning_rate/action_model': 1.0813075045892116e-05, 'epoch': 2.43}) +[2026-06-18 01:48:51,441][starVLA.training.train_starvla][INFO] - Step 5903, Loss: {'train/loss': 0.6289963126182556, 'train/grad_norm_pre_clip': 26.963621139526367, 'timing/data': 0.013815376907587051, 'timing/model': 0.9352672128006816, 'learning_rate/qwen_vl_interface': 1.0802712337441383e-06, 'learning_rate/action_model': 1.0802712337441384e-05, 'epoch': 2.43}) +[2026-06-18 01:48:55,348][starVLA.training.train_starvla][INFO] - Step 5904, Loss: {'train/loss': 0.2672678232192993, 'train/grad_norm_pre_clip': 16.464691162109375, 'timing/data': 0.00813063234090805, 'timing/model': 1.1174474088475108, 'learning_rate/qwen_vl_interface': 1.0792358272883262e-06, 'learning_rate/action_model': 1.0792358272883262e-05, 'epoch': 2.43}) +[2026-06-18 01:48:59,211][starVLA.training.train_starvla][INFO] - Step 5905, Loss: {'train/loss': 0.15320931375026703, 'train/grad_norm_pre_clip': 14.18381118774414, 'timing/data': 0.009241491556167603, 'timing/model': 1.1042200094088912, 'learning_rate/qwen_vl_interface': 1.0782012854364145e-06, 'learning_rate/action_model': 1.0782012854364145e-05, 'epoch': 2.43}) +[2026-06-18 01:49:02,846][starVLA.training.train_starvla][INFO] - Step 5906, Loss: {'train/loss': 0.40019774436950684, 'train/grad_norm_pre_clip': 30.60441017150879, 'timing/data': 0.007765393704175949, 'timing/model': 0.9432493718340993, 'learning_rate/qwen_vl_interface': 1.0771676084028652e-06, 'learning_rate/action_model': 1.0771676084028652e-05, 'epoch': 2.43}) +[2026-06-18 01:49:06,700][starVLA.training.train_starvla][INFO] - Step 5907, Loss: {'train/loss': 0.3035663366317749, 'train/grad_norm_pre_clip': 25.262250900268555, 'timing/data': 0.010165994986891747, 'timing/model': 0.9565404420718551, 'learning_rate/qwen_vl_interface': 1.0761347964019611e-06, 'learning_rate/action_model': 1.076134796401961e-05, 'epoch': 2.43}) +[2026-06-18 01:49:10,782][starVLA.training.train_starvla][INFO] - Step 5908, Loss: {'train/loss': 0.4518480896949768, 'train/grad_norm_pre_clip': 34.01101303100586, 'timing/data': 0.010780555196106434, 'timing/model': 0.9437666228041053, 'learning_rate/qwen_vl_interface': 1.075102849647804e-06, 'learning_rate/action_model': 1.0751028496478038e-05, 'epoch': 2.43}) +[2026-06-18 01:49:14,673][starVLA.training.train_starvla][INFO] - Step 5909, Loss: {'train/loss': 0.25204625725746155, 'train/grad_norm_pre_clip': 13.021217346191406, 'timing/data': 0.008349979296326637, 'timing/model': 0.9383372087031603, 'learning_rate/qwen_vl_interface': 1.0740717683543193e-06, 'learning_rate/action_model': 1.0740717683543191e-05, 'epoch': 2.43}) +[2026-06-18 01:49:18,690][starVLA.training.train_starvla][INFO] - Step 5910, Loss: {'train/loss': 0.09696639329195023, 'train/grad_norm_pre_clip': 11.668656349182129, 'timing/data': 0.009941442869603634, 'timing/model': 1.009182465262711, 'learning_rate/qwen_vl_interface': 1.0730415527352518e-06, 'learning_rate/action_model': 1.0730415527352517e-05, 'epoch': 2.43}) +[2026-06-18 01:49:22,530][starVLA.training.train_starvla][INFO] - Step 5911, Loss: {'train/loss': 0.3030427098274231, 'train/grad_norm_pre_clip': 19.680892944335938, 'timing/data': 0.007754288613796234, 'timing/model': 1.126583338715136, 'learning_rate/qwen_vl_interface': 1.0720122030041638e-06, 'learning_rate/action_model': 1.0720122030041637e-05, 'epoch': 2.43}) +[2026-06-18 01:49:26,302][starVLA.training.train_starvla][INFO] - Step 5912, Loss: {'train/loss': 0.17437715828418732, 'train/grad_norm_pre_clip': 23.89513397216797, 'timing/data': 0.009373822249472141, 'timing/model': 0.9581612842157483, 'learning_rate/qwen_vl_interface': 1.0709837193744412e-06, 'learning_rate/action_model': 1.070983719374441e-05, 'epoch': 2.43}) +[2026-06-18 01:49:30,206][starVLA.training.train_starvla][INFO] - Step 5913, Loss: {'train/loss': 0.11649265885353088, 'train/grad_norm_pre_clip': 10.999476432800293, 'timing/data': 0.008006613701581955, 'timing/model': 0.978284397162497, 'learning_rate/qwen_vl_interface': 1.069956102059291e-06, 'learning_rate/action_model': 1.069956102059291e-05, 'epoch': 2.43}) +[2026-06-18 01:49:33,955][starVLA.training.train_starvla][INFO] - Step 5914, Loss: {'train/loss': 0.14207780361175537, 'train/grad_norm_pre_clip': 18.60808753967285, 'timing/data': 0.008938969112932682, 'timing/model': 0.93885463103652, 'learning_rate/qwen_vl_interface': 1.0689293512717387e-06, 'learning_rate/action_model': 1.0689293512717385e-05, 'epoch': 2.43}) +[2026-06-18 01:49:37,760][starVLA.training.train_starvla][INFO] - Step 5915, Loss: {'train/loss': 0.1974763572216034, 'train/grad_norm_pre_clip': 18.629762649536133, 'timing/data': 0.009694489650428295, 'timing/model': 0.9424397237598896, 'learning_rate/qwen_vl_interface': 1.0679034672246322e-06, 'learning_rate/action_model': 1.0679034672246322e-05, 'epoch': 2.43}) +[2026-06-18 01:49:41,649][starVLA.training.train_starvla][INFO] - Step 5916, Loss: {'train/loss': 0.37292057275772095, 'train/grad_norm_pre_clip': 21.755163192749023, 'timing/data': 0.00793525855988264, 'timing/model': 0.9420745829120278, 'learning_rate/qwen_vl_interface': 1.0668784501306348e-06, 'learning_rate/action_model': 1.0668784501306348e-05, 'epoch': 2.44}) +[2026-06-18 01:49:45,519][starVLA.training.train_starvla][INFO] - Step 5917, Loss: {'train/loss': 0.41101065278053284, 'train/grad_norm_pre_clip': 19.922792434692383, 'timing/data': 0.010991145856678486, 'timing/model': 1.1200790349394083, 'learning_rate/qwen_vl_interface': 1.0658543002022377e-06, 'learning_rate/action_model': 1.0658543002022375e-05, 'epoch': 2.44}) +[2026-06-18 01:49:52,051][starVLA.training.train_starvla][INFO] - Step 5918, Loss: {'train/loss': 0.31198498606681824, 'train/grad_norm_pre_clip': 22.662494659423828, 'timing/data': 0.012432863935828209, 'timing/model': 2.3009375743567944, 'learning_rate/qwen_vl_interface': 1.0648310176517454e-06, 'learning_rate/action_model': 1.0648310176517453e-05, 'epoch': 2.44}) +[2026-06-18 01:49:55,834][starVLA.training.train_starvla][INFO] - Step 5919, Loss: {'train/loss': 0.15250526368618011, 'train/grad_norm_pre_clip': 19.482831954956055, 'timing/data': 0.007166869007050991, 'timing/model': 0.9375641886144876, 'learning_rate/qwen_vl_interface': 1.0638086026912873e-06, 'learning_rate/action_model': 1.0638086026912872e-05, 'epoch': 2.44}) +[2026-06-18 01:49:59,679][starVLA.training.train_starvla][INFO] - Step 5920, Loss: {'train/loss': 0.38762205839157104, 'train/grad_norm_pre_clip': 30.127201080322266, 'timing/data': 0.006796188652515411, 'timing/model': 0.9296283526346087, 'learning_rate/qwen_vl_interface': 1.0627870555328102e-06, 'learning_rate/action_model': 1.06278705553281e-05, 'epoch': 2.44}) +[2026-06-18 01:50:03,563][starVLA.training.train_starvla][INFO] - Step 5921, Loss: {'train/loss': 0.18335586786270142, 'train/grad_norm_pre_clip': 16.325414657592773, 'timing/data': 0.009357227012515068, 'timing/model': 0.9335270654410124, 'learning_rate/qwen_vl_interface': 1.061766376388082e-06, 'learning_rate/action_model': 1.0617663763880821e-05, 'epoch': 2.44}) +[2026-06-18 01:50:07,454][starVLA.training.train_starvla][INFO] - Step 5922, Loss: {'train/loss': 0.1349368542432785, 'train/grad_norm_pre_clip': 16.189069747924805, 'timing/data': 0.008386817760765553, 'timing/model': 0.9302018051967025, 'learning_rate/qwen_vl_interface': 1.0607465654686908e-06, 'learning_rate/action_model': 1.0607465654686907e-05, 'epoch': 2.44}) +[2026-06-18 01:50:11,335][starVLA.training.train_starvla][INFO] - Step 5923, Loss: {'train/loss': 0.20917347073554993, 'train/grad_norm_pre_clip': 30.54671287536621, 'timing/data': 0.007482163608074188, 'timing/model': 0.9460657155141234, 'learning_rate/qwen_vl_interface': 1.0597276229860434e-06, 'learning_rate/action_model': 1.0597276229860434e-05, 'epoch': 2.44}) +[2026-06-18 01:50:15,152][starVLA.training.train_starvla][INFO] - Step 5924, Loss: {'train/loss': 0.3435204029083252, 'train/grad_norm_pre_clip': 92.99592590332031, 'timing/data': 0.00684871431440115, 'timing/model': 1.1001907996833324, 'learning_rate/qwen_vl_interface': 1.0587095491513709e-06, 'learning_rate/action_model': 1.0587095491513709e-05, 'epoch': 2.44}) +[2026-06-18 01:50:19,062][starVLA.training.train_starvla][INFO] - Step 5925, Loss: {'train/loss': 0.352191299200058, 'train/grad_norm_pre_clip': 31.069217681884766, 'timing/data': 0.0101001663133502, 'timing/model': 1.1816863724961877, 'learning_rate/qwen_vl_interface': 1.0576923441757174e-06, 'learning_rate/action_model': 1.0576923441757176e-05, 'epoch': 2.44}) +[2026-06-18 01:50:22,728][starVLA.training.train_starvla][INFO] - Step 5926, Loss: {'train/loss': 0.19818980991840363, 'train/grad_norm_pre_clip': 22.031070709228516, 'timing/data': 0.01975392922759056, 'timing/model': 0.9376577557995915, 'learning_rate/qwen_vl_interface': 1.0566760082699515e-06, 'learning_rate/action_model': 1.0566760082699514e-05, 'epoch': 2.44}) +[2026-06-18 01:50:26,534][starVLA.training.train_starvla][INFO] - Step 5927, Loss: {'train/loss': 0.12061568349599838, 'train/grad_norm_pre_clip': 11.178719520568848, 'timing/data': 0.008835743181407452, 'timing/model': 0.9469734206795692, 'learning_rate/qwen_vl_interface': 1.0556605416447626e-06, 'learning_rate/action_model': 1.0556605416447625e-05, 'epoch': 2.44}) +[2026-06-18 01:50:30,423][starVLA.training.train_starvla][INFO] - Step 5928, Loss: {'train/loss': 0.2958444356918335, 'train/grad_norm_pre_clip': 18.971921920776367, 'timing/data': 0.0098762521520257, 'timing/model': 1.0323584154248238, 'learning_rate/qwen_vl_interface': 1.054645944510656e-06, 'learning_rate/action_model': 1.054645944510656e-05, 'epoch': 2.44}) +[2026-06-18 01:50:34,203][starVLA.training.train_starvla][INFO] - Step 5929, Loss: {'train/loss': 0.1166582778096199, 'train/grad_norm_pre_clip': 18.00560188293457, 'timing/data': 0.011835099197924137, 'timing/model': 0.9406205844134092, 'learning_rate/qwen_vl_interface': 1.0536322170779605e-06, 'learning_rate/action_model': 1.0536322170779606e-05, 'epoch': 2.44}) +[2026-06-18 01:50:38,048][starVLA.training.train_starvla][INFO] - Step 5930, Loss: {'train/loss': 0.10576671361923218, 'train/grad_norm_pre_clip': 10.000107765197754, 'timing/data': 0.009483748115599155, 'timing/model': 0.9448936134576797, 'learning_rate/qwen_vl_interface': 1.0526193595568203e-06, 'learning_rate/action_model': 1.0526193595568203e-05, 'epoch': 2.44}) +[2026-06-18 01:50:41,857][starVLA.training.train_starvla][INFO] - Step 5931, Loss: {'train/loss': 0.2782087028026581, 'train/grad_norm_pre_clip': 30.116968154907227, 'timing/data': 0.012878642417490482, 'timing/model': 1.09572343993932, 'learning_rate/qwen_vl_interface': 1.0516073721572043e-06, 'learning_rate/action_model': 1.0516073721572044e-05, 'epoch': 2.44}) +[2026-06-18 01:50:45,533][starVLA.training.train_starvla][INFO] - Step 5932, Loss: {'train/loss': 0.6464401483535767, 'train/grad_norm_pre_clip': 26.717103958129883, 'timing/data': 0.013196191750466824, 'timing/model': 0.9381877537816763, 'learning_rate/qwen_vl_interface': 1.0505962550888982e-06, 'learning_rate/action_model': 1.0505962550888982e-05, 'epoch': 2.44}) +[2026-06-18 01:50:49,343][starVLA.training.train_starvla][INFO] - Step 5933, Loss: {'train/loss': 0.17879824340343475, 'train/grad_norm_pre_clip': 16.58527946472168, 'timing/data': 0.011466313153505325, 'timing/model': 0.9447395158931613, 'learning_rate/qwen_vl_interface': 1.0495860085615065e-06, 'learning_rate/action_model': 1.0495860085615064e-05, 'epoch': 2.44}) +[2026-06-18 01:50:53,136][starVLA.training.train_starvla][INFO] - Step 5934, Loss: {'train/loss': 0.1054447740316391, 'train/grad_norm_pre_clip': 14.897335052490234, 'timing/data': 0.007749181240797043, 'timing/model': 0.9324282584711909, 'learning_rate/qwen_vl_interface': 1.0485766327844558e-06, 'learning_rate/action_model': 1.0485766327844557e-05, 'epoch': 2.44}) +[2026-06-18 01:50:57,005][starVLA.training.train_starvla][INFO] - Step 5935, Loss: {'train/loss': 0.25547298789024353, 'train/grad_norm_pre_clip': 18.463266372680664, 'timing/data': 0.009680577553808689, 'timing/model': 0.9710039254277945, 'learning_rate/qwen_vl_interface': 1.0475681279669892e-06, 'learning_rate/action_model': 1.0475681279669891e-05, 'epoch': 2.44}) +[2026-06-18 01:51:00,845][starVLA.training.train_starvla][INFO] - Step 5936, Loss: {'train/loss': 0.1660880744457245, 'train/grad_norm_pre_clip': 18.852842330932617, 'timing/data': 0.009898152202367783, 'timing/model': 0.9582477593794465, 'learning_rate/qwen_vl_interface': 1.0465604943181717e-06, 'learning_rate/action_model': 1.0465604943181717e-05, 'epoch': 2.44}) +[2026-06-18 01:51:04,751][starVLA.training.train_starvla][INFO] - Step 5937, Loss: {'train/loss': 0.12382593005895615, 'train/grad_norm_pre_clip': 11.50207805633545, 'timing/data': 0.009983204305171967, 'timing/model': 0.9717410570010543, 'learning_rate/qwen_vl_interface': 1.0455537320468861e-06, 'learning_rate/action_model': 1.045553732046886e-05, 'epoch': 2.44}) +[2026-06-18 01:51:09,190][starVLA.training.train_starvla][INFO] - Step 5938, Loss: {'train/loss': 0.16929258406162262, 'train/grad_norm_pre_clip': 33.912357330322266, 'timing/data': 0.01000795979052782, 'timing/model': 1.313811275176704, 'learning_rate/qwen_vl_interface': 1.0445478413618382e-06, 'learning_rate/action_model': 1.0445478413618382e-05, 'epoch': 2.44}) +[2026-06-18 01:51:12,985][starVLA.training.train_starvla][INFO] - Step 5939, Loss: {'train/loss': 0.27741777896881104, 'train/grad_norm_pre_clip': 23.811473846435547, 'timing/data': 0.010279340669512749, 'timing/model': 0.9526913119480014, 'learning_rate/qwen_vl_interface': 1.0435428224715472e-06, 'learning_rate/action_model': 1.0435428224715471e-05, 'epoch': 2.44}) +[2026-06-18 01:51:18,007][starVLA.training.train_starvla][INFO] - Step 5940, Loss: {'train/loss': 0.13636550307273865, 'train/grad_norm_pre_clip': 14.688335418701172, 'timing/data': 0.010530982166528702, 'timing/model': 1.124105317518115, 'learning_rate/qwen_vl_interface': 1.0425386755843547e-06, 'learning_rate/action_model': 1.0425386755843547e-05, 'epoch': 2.45}) +[2026-06-18 01:51:21,997][starVLA.training.train_starvla][INFO] - Step 5941, Loss: {'train/loss': 0.3535241186618805, 'train/grad_norm_pre_clip': 19.231708526611328, 'timing/data': 0.010980205610394478, 'timing/model': 1.0059474725276232, 'learning_rate/qwen_vl_interface': 1.0415354009084219e-06, 'learning_rate/action_model': 1.041535400908422e-05, 'epoch': 2.45}) +[2026-06-18 01:51:25,835][starVLA.training.train_starvla][INFO] - Step 5942, Loss: {'train/loss': 0.12471948564052582, 'train/grad_norm_pre_clip': 13.186601638793945, 'timing/data': 0.010408538393676281, 'timing/model': 0.9446311146020889, 'learning_rate/qwen_vl_interface': 1.0405329986517301e-06, 'learning_rate/action_model': 1.0405329986517301e-05, 'epoch': 2.45}) +[2026-06-18 01:51:29,671][starVLA.training.train_starvla][INFO] - Step 5943, Loss: {'train/loss': 0.3327616751194, 'train/grad_norm_pre_clip': 19.151992797851562, 'timing/data': 0.009749729186296463, 'timing/model': 0.9771422706544399, 'learning_rate/qwen_vl_interface': 1.0395314690220776e-06, 'learning_rate/action_model': 1.0395314690220776e-05, 'epoch': 2.45}) +[2026-06-18 01:51:33,616][starVLA.training.train_starvla][INFO] - Step 5944, Loss: {'train/loss': 0.13786938786506653, 'train/grad_norm_pre_clip': 15.660743713378906, 'timing/data': 0.008498266339302063, 'timing/model': 1.1235104380175471, 'learning_rate/qwen_vl_interface': 1.0385308122270815e-06, 'learning_rate/action_model': 1.0385308122270814e-05, 'epoch': 2.45}) +[2026-06-18 01:51:37,419][starVLA.training.train_starvla][INFO] - Step 5945, Loss: {'train/loss': 0.18630433082580566, 'train/grad_norm_pre_clip': 18.96603012084961, 'timing/data': 0.007844513282179832, 'timing/model': 1.1157149709761143, 'learning_rate/qwen_vl_interface': 1.0375310284741787e-06, 'learning_rate/action_model': 1.0375310284741788e-05, 'epoch': 2.45}) +[2026-06-18 01:51:41,043][starVLA.training.train_starvla][INFO] - Step 5946, Loss: {'train/loss': 0.2411368489265442, 'train/grad_norm_pre_clip': 23.892560958862305, 'timing/data': 0.009149407967925072, 'timing/model': 0.938768282532692, 'learning_rate/qwen_vl_interface': 1.0365321179706276e-06, 'learning_rate/action_model': 1.0365321179706276e-05, 'epoch': 2.45}) +[2026-06-18 01:51:44,911][starVLA.training.train_starvla][INFO] - Step 5947, Loss: {'train/loss': 0.17432034015655518, 'train/grad_norm_pre_clip': 14.18889331817627, 'timing/data': 0.00954683031886816, 'timing/model': 0.9553153924643993, 'learning_rate/qwen_vl_interface': 1.035534080923503e-06, 'learning_rate/action_model': 1.0355340809235029e-05, 'epoch': 2.45}) +[2026-06-18 01:51:48,780][starVLA.training.train_starvla][INFO] - Step 5948, Loss: {'train/loss': 0.2528649568557739, 'train/grad_norm_pre_clip': 23.791454315185547, 'timing/data': 0.012051357887685299, 'timing/model': 0.9466339619830251, 'learning_rate/qwen_vl_interface': 1.0345369175396972e-06, 'learning_rate/action_model': 1.0345369175396972e-05, 'epoch': 2.45}) +[2026-06-18 01:51:52,627][starVLA.training.train_starvla][INFO] - Step 5949, Loss: {'train/loss': 0.2963605523109436, 'train/grad_norm_pre_clip': 26.76679039001465, 'timing/data': 0.00858671311289072, 'timing/model': 0.9669135250151157, 'learning_rate/qwen_vl_interface': 1.0335406280259255e-06, 'learning_rate/action_model': 1.0335406280259254e-05, 'epoch': 2.45}) +[2026-06-18 01:51:56,526][starVLA.training.train_starvla][INFO] - Step 5950, Loss: {'train/loss': 0.11426694691181183, 'train/grad_norm_pre_clip': 13.45801830291748, 'timing/data': 0.010406331159174442, 'timing/model': 0.972817437723279, 'learning_rate/qwen_vl_interface': 1.0325452125887177e-06, 'learning_rate/action_model': 1.0325452125887178e-05, 'epoch': 2.45}) +[2026-06-18 01:52:00,314][starVLA.training.train_starvla][INFO] - Step 5951, Loss: {'train/loss': 0.15710973739624023, 'train/grad_norm_pre_clip': 24.031221389770508, 'timing/data': 0.011690022423863411, 'timing/model': 1.1059682928025723, 'learning_rate/qwen_vl_interface': 1.031550671434426e-06, 'learning_rate/action_model': 1.031550671434426e-05, 'epoch': 2.45}) +[2026-06-18 01:52:04,085][starVLA.training.train_starvla][INFO] - Step 5952, Loss: {'train/loss': 0.11059208959341049, 'train/grad_norm_pre_clip': 14.836243629455566, 'timing/data': 0.007952823303639889, 'timing/model': 1.0739388819783926, 'learning_rate/qwen_vl_interface': 1.0305570047692197e-06, 'learning_rate/action_model': 1.0305570047692195e-05, 'epoch': 2.45}) +[2026-06-18 01:52:07,844][starVLA.training.train_starvla][INFO] - Step 5953, Loss: {'train/loss': 0.3013869524002075, 'train/grad_norm_pre_clip': 22.430795669555664, 'timing/data': 0.008596248924732208, 'timing/model': 0.938435273244977, 'learning_rate/qwen_vl_interface': 1.0295642127990863e-06, 'learning_rate/action_model': 1.0295642127990863e-05, 'epoch': 2.45}) +[2026-06-18 01:52:11,642][starVLA.training.train_starvla][INFO] - Step 5954, Loss: {'train/loss': 0.19840961694717407, 'train/grad_norm_pre_clip': 22.21062469482422, 'timing/data': 0.00951511412858963, 'timing/model': 0.942624170333147, 'learning_rate/qwen_vl_interface': 1.0285722957298326e-06, 'learning_rate/action_model': 1.0285722957298326e-05, 'epoch': 2.45}) +[2026-06-18 01:52:15,517][starVLA.training.train_starvla][INFO] - Step 5955, Loss: {'train/loss': 0.24909472465515137, 'train/grad_norm_pre_clip': 22.719228744506836, 'timing/data': 0.011638443917036057, 'timing/model': 0.9489428270608187, 'learning_rate/qwen_vl_interface': 1.027581253767085e-06, 'learning_rate/action_model': 1.027581253767085e-05, 'epoch': 2.45}) +[2026-06-18 01:52:19,254][starVLA.training.train_starvla][INFO] - Step 5956, Loss: {'train/loss': 0.2323204129934311, 'train/grad_norm_pre_clip': 24.798778533935547, 'timing/data': 0.013171122409403324, 'timing/model': 0.9279597299173474, 'learning_rate/qwen_vl_interface': 1.0265910871162858e-06, 'learning_rate/action_model': 1.0265910871162858e-05, 'epoch': 2.45}) +[2026-06-18 01:52:23,461][starVLA.training.train_starvla][INFO] - Step 5957, Loss: {'train/loss': 0.39078089594841003, 'train/grad_norm_pre_clip': 29.945873260498047, 'timing/data': 0.009479165077209473, 'timing/model': 0.952191486954689, 'learning_rate/qwen_vl_interface': 1.0256017959827008e-06, 'learning_rate/action_model': 1.0256017959827008e-05, 'epoch': 2.45}) +[2026-06-18 01:52:27,436][starVLA.training.train_starvla][INFO] - Step 5958, Loss: {'train/loss': 0.10080316662788391, 'train/grad_norm_pre_clip': 19.452438354492188, 'timing/data': 0.0007475968450307846, 'timing/model': 1.1933807283639908, 'learning_rate/qwen_vl_interface': 1.0246133805714084e-06, 'learning_rate/action_model': 1.0246133805714084e-05, 'epoch': 2.45}) +[2026-06-18 01:52:31,262][starVLA.training.train_starvla][INFO] - Step 5959, Loss: {'train/loss': 0.4139106273651123, 'train/grad_norm_pre_clip': 29.867748260498047, 'timing/data': 0.0004556868225336075, 'timing/model': 0.9323588395491242, 'learning_rate/qwen_vl_interface': 1.0236258410873076e-06, 'learning_rate/action_model': 1.0236258410873076e-05, 'epoch': 2.45}) +[2026-06-18 01:52:35,076][starVLA.training.train_starvla][INFO] - Step 5960, Loss: {'train/loss': 0.1707758903503418, 'train/grad_norm_pre_clip': 14.295672416687012, 'timing/data': 0.00047217216342687607, 'timing/model': 0.9376198695972562, 'learning_rate/qwen_vl_interface': 1.0226391777351188e-06, 'learning_rate/action_model': 1.0226391777351189e-05, 'epoch': 2.45}) +[2026-06-18 01:52:38,876][starVLA.training.train_starvla][INFO] - Step 5961, Loss: {'train/loss': 0.1681181937456131, 'train/grad_norm_pre_clip': 12.919300079345703, 'timing/data': 0.0008892016485333443, 'timing/model': 0.9348305966705084, 'learning_rate/qwen_vl_interface': 1.0216533907193777e-06, 'learning_rate/action_model': 1.0216533907193776e-05, 'epoch': 2.45}) +[2026-06-18 01:52:42,905][starVLA.training.train_starvla][INFO] - Step 5962, Loss: {'train/loss': 0.09947459399700165, 'train/grad_norm_pre_clip': 12.970563888549805, 'timing/data': 0.000843273475766182, 'timing/model': 0.955910081975162, 'learning_rate/qwen_vl_interface': 1.020668480244438e-06, 'learning_rate/action_model': 1.020668480244438e-05, 'epoch': 2.45}) +[2026-06-18 01:52:48,778][starVLA.training.train_starvla][INFO] - Step 5963, Loss: {'train/loss': 0.8258091807365417, 'train/grad_norm_pre_clip': 28.15309715270996, 'timing/data': 0.0008029937744140625, 'timing/model': 1.1084920894354582, 'learning_rate/qwen_vl_interface': 1.0196844465144735e-06, 'learning_rate/action_model': 1.0196844465144734e-05, 'epoch': 2.45}) +[2026-06-18 01:52:52,796][starVLA.training.train_starvla][INFO] - Step 5964, Loss: {'train/loss': 0.3434281349182129, 'train/grad_norm_pre_clip': 28.594778060913086, 'timing/data': 0.0006195297464728355, 'timing/model': 1.1646391935646534, 'learning_rate/qwen_vl_interface': 1.0187012897334755e-06, 'learning_rate/action_model': 1.0187012897334755e-05, 'epoch': 2.46}) +[2026-06-18 01:52:56,696][starVLA.training.train_starvla][INFO] - Step 5965, Loss: {'train/loss': 0.17597314715385437, 'train/grad_norm_pre_clip': 31.553918838500977, 'timing/data': 0.0008297516033053398, 'timing/model': 1.139982858672738, 'learning_rate/qwen_vl_interface': 1.017719010105253e-06, 'learning_rate/action_model': 1.017719010105253e-05, 'epoch': 2.46}) +[2026-06-18 01:53:00,406][starVLA.training.train_starvla][INFO] - Step 5966, Loss: {'train/loss': 0.2252863198518753, 'train/grad_norm_pre_clip': 16.26576042175293, 'timing/data': 0.0013977615162730217, 'timing/model': 0.9482415849342942, 'learning_rate/qwen_vl_interface': 1.0167376078334337e-06, 'learning_rate/action_model': 1.0167376078334337e-05, 'epoch': 2.46}) +[2026-06-18 01:53:06,166][starVLA.training.train_starvla][INFO] - Step 5967, Loss: {'train/loss': 0.15476831793785095, 'train/grad_norm_pre_clip': 13.61666488647461, 'timing/data': 0.0005907062441110611, 'timing/model': 1.0883700428530574, 'learning_rate/qwen_vl_interface': 1.0157570831214634e-06, 'learning_rate/action_model': 1.0157570831214634e-05, 'epoch': 2.46}) +[2026-06-18 01:53:11,323][starVLA.training.train_starvla][INFO] - Step 5968, Loss: {'train/loss': 0.19751128554344177, 'train/grad_norm_pre_clip': 15.324895858764648, 'timing/data': 0.0007016044110059738, 'timing/model': 0.984161782078445, 'learning_rate/qwen_vl_interface': 1.0147774361726066e-06, 'learning_rate/action_model': 1.0147774361726064e-05, 'epoch': 2.46}) +[2026-06-18 01:53:15,528][starVLA.training.train_starvla][INFO] - Step 5969, Loss: {'train/loss': 0.7736860513687134, 'train/grad_norm_pre_clip': 42.43584442138672, 'timing/data': 0.0010112719610333443, 'timing/model': 0.970327346585691, 'learning_rate/qwen_vl_interface': 1.013798667189944e-06, 'learning_rate/action_model': 1.013798667189944e-05, 'epoch': 2.46}) +[2026-06-18 01:53:19,679][starVLA.training.train_starvla][INFO] - Step 5970, Loss: {'train/loss': 0.5216038227081299, 'train/grad_norm_pre_clip': 33.34956741333008, 'timing/data': 0.00044263992458581924, 'timing/model': 1.0132897989824414, 'learning_rate/qwen_vl_interface': 1.0128207763763755e-06, 'learning_rate/action_model': 1.0128207763763755e-05, 'epoch': 2.46}) +[2026-06-18 01:53:23,590][starVLA.training.train_starvla][INFO] - Step 5971, Loss: {'train/loss': 0.28296083211898804, 'train/grad_norm_pre_clip': 19.59513282775879, 'timing/data': 0.0004753880202770233, 'timing/model': 1.1500754049047828, 'learning_rate/qwen_vl_interface': 1.0118437639346204e-06, 'learning_rate/action_model': 1.0118437639346204e-05, 'epoch': 2.46}) +[2026-06-18 01:53:27,545][starVLA.training.train_starvla][INFO] - Step 5972, Loss: {'train/loss': 0.260935515165329, 'train/grad_norm_pre_clip': 29.73064422607422, 'timing/data': 0.0005288189277052879, 'timing/model': 0.9469583611935377, 'learning_rate/qwen_vl_interface': 1.0108676300672144e-06, 'learning_rate/action_model': 1.0108676300672144e-05, 'epoch': 2.46}) +[2026-06-18 01:53:31,643][starVLA.training.train_starvla][INFO] - Step 5973, Loss: {'train/loss': 0.31314513087272644, 'train/grad_norm_pre_clip': 24.804838180541992, 'timing/data': 0.0004877513274550438, 'timing/model': 0.9906775252893567, 'learning_rate/qwen_vl_interface': 1.0098923749765086e-06, 'learning_rate/action_model': 1.0098923749765086e-05, 'epoch': 2.46}) +[2026-06-18 01:53:35,620][starVLA.training.train_starvla][INFO] - Step 5974, Loss: {'train/loss': 0.34516438841819763, 'train/grad_norm_pre_clip': 29.668960571289062, 'timing/data': 0.0004924032837152481, 'timing/model': 1.0049123903736472, 'learning_rate/qwen_vl_interface': 1.008917998864676e-06, 'learning_rate/action_model': 1.008917998864676e-05, 'epoch': 2.46}) +[2026-06-18 01:53:39,427][starVLA.training.train_starvla][INFO] - Step 5975, Loss: {'train/loss': 0.25498342514038086, 'train/grad_norm_pre_clip': 24.336332321166992, 'timing/data': 0.0008015967905521393, 'timing/model': 0.9328156225383282, 'learning_rate/qwen_vl_interface': 1.0079445019337063e-06, 'learning_rate/action_model': 1.0079445019337063e-05, 'epoch': 2.46}) +[2026-06-18 01:53:43,230][starVLA.training.train_starvla][INFO] - Step 5976, Loss: {'train/loss': 0.31998109817504883, 'train/grad_norm_pre_clip': 26.64333152770996, 'timing/data': 0.0008846372365951538, 'timing/model': 0.939924968406558, 'learning_rate/qwen_vl_interface': 1.0069718843854058e-06, 'learning_rate/action_model': 1.0069718843854058e-05, 'epoch': 2.46}) +[2026-06-18 01:53:47,054][starVLA.training.train_starvla][INFO] - Step 5977, Loss: {'train/loss': 0.1132105365395546, 'train/grad_norm_pre_clip': 10.688946723937988, 'timing/data': 0.0004720129072666168, 'timing/model': 1.1255603777244687, 'learning_rate/qwen_vl_interface': 1.0060001464214007e-06, 'learning_rate/action_model': 1.0060001464214006e-05, 'epoch': 2.46}) +[2026-06-18 01:53:51,119][starVLA.training.train_starvla][INFO] - Step 5978, Loss: {'train/loss': 0.16132594645023346, 'train/grad_norm_pre_clip': 19.521989822387695, 'timing/data': 0.0005379840731620789, 'timing/model': 1.2648252425715327, 'learning_rate/qwen_vl_interface': 1.00502928824313e-06, 'learning_rate/action_model': 1.00502928824313e-05, 'epoch': 2.46}) +[2026-06-18 01:53:54,853][starVLA.training.train_starvla][INFO] - Step 5979, Loss: {'train/loss': 0.40090763568878174, 'train/grad_norm_pre_clip': 28.683496475219727, 'timing/data': 0.0010982397943735123, 'timing/model': 0.9854193730279803, 'learning_rate/qwen_vl_interface': 1.0040593100518567e-06, 'learning_rate/action_model': 1.0040593100518566e-05, 'epoch': 2.46}) +[2026-06-18 01:53:58,885][starVLA.training.train_starvla][INFO] - Step 5980, Loss: {'train/loss': 0.08850297331809998, 'train/grad_norm_pre_clip': 10.862360000610352, 'timing/data': 0.0005881814286112785, 'timing/model': 1.0081071695312858, 'learning_rate/qwen_vl_interface': 1.003090212048657e-06, 'learning_rate/action_model': 1.003090212048657e-05, 'epoch': 2.46}) +[2026-06-18 01:54:02,843][starVLA.training.train_starvla][INFO] - Step 5981, Loss: {'train/loss': 0.23387813568115234, 'train/grad_norm_pre_clip': 24.108097076416016, 'timing/data': 0.0006277831271290779, 'timing/model': 0.9788369666785002, 'learning_rate/qwen_vl_interface': 1.0021219944344263e-06, 'learning_rate/action_model': 1.0021219944344262e-05, 'epoch': 2.46}) +[2026-06-18 01:54:06,769][starVLA.training.train_starvla][INFO] - Step 5982, Loss: {'train/loss': 0.16769830882549286, 'train/grad_norm_pre_clip': 27.2768611907959, 'timing/data': 0.001079280860722065, 'timing/model': 0.9744231840595603, 'learning_rate/qwen_vl_interface': 1.0011546574098773e-06, 'learning_rate/action_model': 1.0011546574098772e-05, 'epoch': 2.46}) +[2026-06-18 01:54:10,834][starVLA.training.train_starvla][INFO] - Step 5983, Loss: {'train/loss': 0.3167479634284973, 'train/grad_norm_pre_clip': 21.69350814819336, 'timing/data': 0.0005265148356556892, 'timing/model': 0.9850442036986351, 'learning_rate/qwen_vl_interface': 1.0001882011755396e-06, 'learning_rate/action_model': 1.0001882011755395e-05, 'epoch': 2.46}) +[2026-06-18 01:54:15,193][starVLA.training.train_starvla][INFO] - Step 5984, Loss: {'train/loss': 0.3338569104671478, 'train/grad_norm_pre_clip': 25.144601821899414, 'timing/data': 0.0006859460845589638, 'timing/model': 1.1588862715288997, 'learning_rate/qwen_vl_interface': 9.992226259317605e-07, 'learning_rate/action_model': 9.992226259317605e-06, 'epoch': 2.46}) +[2026-06-18 01:54:19,054][starVLA.training.train_starvla][INFO] - Step 5985, Loss: {'train/loss': 0.3488413691520691, 'train/grad_norm_pre_clip': 16.792200088500977, 'timing/data': 0.0005724439397454262, 'timing/model': 1.1464164024218917, 'learning_rate/qwen_vl_interface': 9.982579318787044e-07, 'learning_rate/action_model': 9.982579318787044e-06, 'epoch': 2.46}) +[2026-06-18 01:54:24,134][starVLA.training.train_starvla][INFO] - Step 5986, Loss: {'train/loss': 0.21468885242938995, 'train/grad_norm_pre_clip': 15.978556632995605, 'timing/data': 0.0010381853207945824, 'timing/model': 0.9336383389309049, 'learning_rate/qwen_vl_interface': 9.97294119216356e-07, 'learning_rate/action_model': 9.97294119216356e-06, 'epoch': 2.46}) +[2026-06-18 01:54:28,442][starVLA.training.train_starvla][INFO] - Step 5987, Loss: {'train/loss': 0.2198568731546402, 'train/grad_norm_pre_clip': 19.164384841918945, 'timing/data': 0.0007933424785733223, 'timing/model': 0.9463268807157874, 'learning_rate/qwen_vl_interface': 9.96331188144512e-07, 'learning_rate/action_model': 9.96331188144512e-06, 'epoch': 2.46}) +[2026-06-18 01:54:32,238][starVLA.training.train_starvla][INFO] - Step 5988, Loss: {'train/loss': 0.17219653725624084, 'train/grad_norm_pre_clip': 16.229705810546875, 'timing/data': 0.0008062031120061874, 'timing/model': 0.9432265786454082, 'learning_rate/qwen_vl_interface': 9.953691388627887e-07, 'learning_rate/action_model': 9.953691388627886e-06, 'epoch': 2.46}) +[2026-06-18 01:54:37,777][starVLA.training.train_starvla][INFO] - Step 5989, Loss: {'train/loss': 0.2843078672885895, 'train/grad_norm_pre_clip': 80.38867950439453, 'timing/data': 0.00047862250357866287, 'timing/model': 0.9977197702974081, 'learning_rate/qwen_vl_interface': 9.94407971570622e-07, 'learning_rate/action_model': 9.94407971570622e-06, 'epoch': 2.47}) +[2026-06-18 01:54:41,688][starVLA.training.train_starvla][INFO] - Step 5990, Loss: {'train/loss': 0.5206028819084167, 'train/grad_norm_pre_clip': 35.98549270629883, 'timing/data': 0.0005860431119799614, 'timing/model': 1.0038487799465656, 'learning_rate/qwen_vl_interface': 9.934476864672618e-07, 'learning_rate/action_model': 9.934476864672617e-06, 'epoch': 2.47}) +[2026-06-18 01:54:46,001][starVLA.training.train_starvla][INFO] - Step 5991, Loss: {'train/loss': 0.06746704876422882, 'train/grad_norm_pre_clip': 8.627477645874023, 'timing/data': 0.0011801701039075851, 'timing/model': 1.2470909440889955, 'learning_rate/qwen_vl_interface': 9.924882837517781e-07, 'learning_rate/action_model': 9.924882837517782e-06, 'epoch': 2.47}) +[2026-06-18 01:54:49,788][starVLA.training.train_starvla][INFO] - Step 5992, Loss: {'train/loss': 0.21420085430145264, 'train/grad_norm_pre_clip': 15.868128776550293, 'timing/data': 0.0005048923194408417, 'timing/model': 1.1066931253299117, 'learning_rate/qwen_vl_interface': 9.91529763623052e-07, 'learning_rate/action_model': 9.91529763623052e-06, 'epoch': 2.47}) +[2026-06-18 01:54:53,716][starVLA.training.train_starvla][INFO] - Step 5993, Loss: {'train/loss': 0.24750149250030518, 'train/grad_norm_pre_clip': 23.014450073242188, 'timing/data': 0.0007661916315555573, 'timing/model': 1.0608706949278712, 'learning_rate/qwen_vl_interface': 9.905721262797894e-07, 'learning_rate/action_model': 9.905721262797892e-06, 'epoch': 2.47}) +[2026-06-18 01:54:58,006][starVLA.training.train_starvla][INFO] - Step 5994, Loss: {'train/loss': 0.4003960192203522, 'train/grad_norm_pre_clip': 25.273998260498047, 'timing/data': 0.0012142853811383247, 'timing/model': 1.3990990482270718, 'learning_rate/qwen_vl_interface': 9.896153719205082e-07, 'learning_rate/action_model': 9.89615371920508e-06, 'epoch': 2.47}) +[2026-06-18 01:55:01,841][starVLA.training.train_starvla][INFO] - Step 5995, Loss: {'train/loss': 0.44457581639289856, 'train/grad_norm_pre_clip': 25.798717498779297, 'timing/data': 0.0008608438074588776, 'timing/model': 0.9503510603681207, 'learning_rate/qwen_vl_interface': 9.886595007435448e-07, 'learning_rate/action_model': 9.886595007435447e-06, 'epoch': 2.47}) +[2026-06-18 01:55:05,662][starVLA.training.train_starvla][INFO] - Step 5996, Loss: {'train/loss': 0.3391149938106537, 'train/grad_norm_pre_clip': 26.449886322021484, 'timing/data': 0.0008811485022306442, 'timing/model': 0.9632064932957292, 'learning_rate/qwen_vl_interface': 9.877045129470524e-07, 'learning_rate/action_model': 9.877045129470524e-06, 'epoch': 2.47}) +[2026-06-18 01:55:09,641][starVLA.training.train_starvla][INFO] - Step 5997, Loss: {'train/loss': 0.12526419758796692, 'train/grad_norm_pre_clip': 21.226106643676758, 'timing/data': 0.000981646589934826, 'timing/model': 0.9383555259555578, 'learning_rate/qwen_vl_interface': 9.86750408729001e-07, 'learning_rate/action_model': 9.867504087290009e-06, 'epoch': 2.47}) +[2026-06-18 01:55:13,584][starVLA.training.train_starvla][INFO] - Step 5998, Loss: {'train/loss': 0.4055776298046112, 'train/grad_norm_pre_clip': 32.71632385253906, 'timing/data': 0.0004931418225169182, 'timing/model': 1.2009512223303318, 'learning_rate/qwen_vl_interface': 9.857971882871773e-07, 'learning_rate/action_model': 9.857971882871772e-06, 'epoch': 2.47}) +[2026-06-18 01:55:17,539][starVLA.training.train_starvla][INFO] - Step 5999, Loss: {'train/loss': 0.27974122762680054, 'train/grad_norm_pre_clip': 26.066194534301758, 'timing/data': 0.0009618271142244339, 'timing/model': 0.9345848523080349, 'learning_rate/qwen_vl_interface': 9.848448518191838e-07, 'learning_rate/action_model': 9.848448518191836e-06, 'epoch': 2.47}) +[2026-06-18 01:57:20,190][starVLA.training.train_starvla][INFO] - Step 6000, Loss: {'train/loss': 0.2097894698381424, 'train/grad_norm_pre_clip': 25.996444702148438, 'eval/loss': 0.4205079674720764, 'eval/latency_6/loss': 0.4205079674720764, 'eval/demon_fire_precision': 0.9502191609267376, 'eval/demon_fire_recall': 0.9809308338720103, 'eval/demon_fire_f1': 0.9653307888040713, 'eval/demon_fire_rate_model': 0.8887033945464663, 'eval/demon_fire_rate_teacher': 0.860879243183083, 'eval/demon_left_precision': 0.8722109533468559, 'eval/demon_left_recall': 0.7543859649122807, 'eval/demon_left_f1': 0.8090310442144873, 'eval/demon_left_rate_model': 0.13717306622148023, 'eval/demon_left_rate_teacher': 0.15859766277128548, 'eval/demon_right_precision': 0.8471986417657046, 'eval/demon_right_recall': 0.7665130568356375, 'eval/demon_right_f1': 0.8048387096774194, 'eval/demon_right_rate_model': 0.16388425153032832, 'eval/demon_right_rate_teacher': 0.18113522537562604, 'eval/demon_move_f1': 0.8069348769459534, 'eval/demon_cc_f1': 0.9116457994144459, 'eval/latency_6/demon_fire_precision': 0.9502191609267376, 'eval/latency_6/demon_fire_recall': 0.9809308338720103, 'eval/latency_6/demon_fire_f1': 0.9653307888040713, 'eval/latency_6/demon_fire_rate_model': 0.8887033945464663, 'eval/latency_6/demon_fire_rate_teacher': 0.860879243183083, 'eval/latency_6/demon_left_precision': 0.8722109533468559, 'eval/latency_6/demon_left_recall': 0.7543859649122807, 'eval/latency_6/demon_left_f1': 0.8090310442144873, 'eval/latency_6/demon_left_rate_model': 0.13717306622148023, 'eval/latency_6/demon_left_rate_teacher': 0.15859766277128548, 'eval/latency_6/demon_right_precision': 0.8471986417657046, 'eval/latency_6/demon_right_recall': 0.7665130568356375, 'eval/latency_6/demon_right_f1': 0.8048387096774194, 'eval/latency_6/demon_right_rate_model': 0.16388425153032832, 'eval/latency_6/demon_right_rate_teacher': 0.18113522537562604, 'eval/latency_6/demon_move_f1': 0.8069348769459534, 'eval/latency_6/demon_cc_f1': 0.9116457994144459, 'timing/data': 0.0005167014896869659, 'timing/model': 0.9717174675315619, 'learning_rate/qwen_vl_interface': 9.838933995224439e-07, 'learning_rate/action_model': 9.83893399522444e-06, 'epoch': 2.47}) +[2026-06-18 01:57:27,865][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_6000_state +[2026-06-18 01:57:27,871][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 01:57:34,811][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_6000_state/model.safetensors +[2026-06-18 01:57:49,889][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_6000_state/optimizer.bin +[2026-06-18 01:57:49,897][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_6000_state/sampler.bin +[2026-06-18 01:57:49,897][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_6000_state/sampler_1.bin +[2026-06-18 01:57:49,899][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_6000_state/random_states_0.pkl +[2026-06-18 01:57:49,903][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 6000 +[2026-06-18 01:57:49,903][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 01:57:49,914][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 01:57:53,734][starVLA.training.train_starvla][INFO] - Step 6001, Loss: {'train/loss': 0.5716740489006042, 'train/grad_norm_pre_clip': 38.0250244140625, 'timing/data': 0.0005577290430665016, 'timing/model': 1.1138105224817991, 'learning_rate/qwen_vl_interface': 9.829428315941921e-07, 'learning_rate/action_model': 9.829428315941922e-06, 'epoch': 2.47}) +[2026-06-18 01:57:57,334][starVLA.training.train_starvla][INFO] - Step 6002, Loss: {'train/loss': 0.11894302815198898, 'train/grad_norm_pre_clip': 16.864208221435547, 'timing/data': 0.0005243699997663498, 'timing/model': 0.9405355649068952, 'learning_rate/qwen_vl_interface': 9.819931482314823e-07, 'learning_rate/action_model': 9.819931482314823e-06, 'epoch': 2.47}) +[2026-06-18 01:58:01,162][starVLA.training.train_starvla][INFO] - Step 6003, Loss: {'train/loss': 0.34964603185653687, 'train/grad_norm_pre_clip': 17.502836227416992, 'timing/data': 0.0006801551207900047, 'timing/model': 0.9727291241288185, 'learning_rate/qwen_vl_interface': 9.81044349631185e-07, 'learning_rate/action_model': 9.81044349631185e-06, 'epoch': 2.47}) +[2026-06-18 01:58:05,085][starVLA.training.train_starvla][INFO] - Step 6004, Loss: {'train/loss': 0.5860810875892639, 'train/grad_norm_pre_clip': 26.109397888183594, 'timing/data': 0.0011672675609588623, 'timing/model': 0.9332197038456798, 'learning_rate/qwen_vl_interface': 9.800964359899886e-07, 'learning_rate/action_model': 9.800964359899886e-06, 'epoch': 2.47}) +[2026-06-18 01:58:10,816][starVLA.training.train_starvla][INFO] - Step 6005, Loss: {'train/loss': 0.19284383952617645, 'train/grad_norm_pre_clip': 12.904117584228516, 'timing/data': 0.0011931108310818672, 'timing/model': 2.8474405594170094, 'learning_rate/qwen_vl_interface': 9.791494075043963e-07, 'learning_rate/action_model': 9.791494075043963e-06, 'epoch': 2.47}) +[2026-06-18 01:58:15,624][starVLA.training.train_starvla][INFO] - Step 6006, Loss: {'train/loss': 0.17089904844760895, 'train/grad_norm_pre_clip': 23.11959457397461, 'timing/data': 0.0008900649845600128, 'timing/model': 1.9435411961749196, 'learning_rate/qwen_vl_interface': 9.78203264370727e-07, 'learning_rate/action_model': 9.78203264370727e-06, 'epoch': 2.47}) +[2026-06-18 01:58:19,439][starVLA.training.train_starvla][INFO] - Step 6007, Loss: {'train/loss': 0.43248704075813293, 'train/grad_norm_pre_clip': 28.20568084716797, 'timing/data': 0.0004968782886862755, 'timing/model': 1.1078862510621548, 'learning_rate/qwen_vl_interface': 9.772580067851162e-07, 'learning_rate/action_model': 9.772580067851161e-06, 'epoch': 2.47}) +[2026-06-18 01:58:23,211][starVLA.training.train_starvla][INFO] - Step 6008, Loss: {'train/loss': 0.10159312933683395, 'train/grad_norm_pre_clip': 14.86026382446289, 'timing/data': 0.0005220277234911919, 'timing/model': 0.9948975322768092, 'learning_rate/qwen_vl_interface': 9.763136349435195e-07, 'learning_rate/action_model': 9.763136349435194e-06, 'epoch': 2.47}) +[2026-06-18 01:58:27,158][starVLA.training.train_starvla][INFO] - Step 6009, Loss: {'train/loss': 0.17815826833248138, 'train/grad_norm_pre_clip': 20.115856170654297, 'timing/data': 0.0005101505666971207, 'timing/model': 0.9497186038643122, 'learning_rate/qwen_vl_interface': 9.753701490417048e-07, 'learning_rate/action_model': 9.753701490417047e-06, 'epoch': 2.47}) +[2026-06-18 01:58:32,309][starVLA.training.train_starvla][INFO] - Step 6010, Loss: {'train/loss': 0.17645147442817688, 'train/grad_norm_pre_clip': 15.010480880737305, 'timing/data': 0.0005094390362501144, 'timing/model': 0.9500820776447654, 'learning_rate/qwen_vl_interface': 9.744275492752584e-07, 'learning_rate/action_model': 9.744275492752583e-06, 'epoch': 2.47}) +[2026-06-18 01:58:36,203][starVLA.training.train_starvla][INFO] - Step 6011, Loss: {'train/loss': 0.3214733600616455, 'train/grad_norm_pre_clip': 32.665470123291016, 'timing/data': 0.0014887526631355286, 'timing/model': 1.0132166193798184, 'learning_rate/qwen_vl_interface': 9.734858358395808e-07, 'learning_rate/action_model': 9.734858358395807e-06, 'epoch': 2.47}) +[2026-06-18 01:58:40,281][starVLA.training.train_starvla][INFO] - Step 6012, Loss: {'train/loss': 0.16303782165050507, 'train/grad_norm_pre_clip': 22.794044494628906, 'timing/data': 0.0005141329020261765, 'timing/model': 0.934870476834476, 'learning_rate/qwen_vl_interface': 9.72545008929892e-07, 'learning_rate/action_model': 9.725450089298921e-06, 'epoch': 2.47}) +[2026-06-18 01:58:44,119][starVLA.training.train_starvla][INFO] - Step 6013, Loss: {'train/loss': 0.3948977589607239, 'train/grad_norm_pre_clip': 24.93317222595215, 'timing/data': 0.0009923437610268593, 'timing/model': 1.1111602447926998, 'learning_rate/qwen_vl_interface': 9.71605068741225e-07, 'learning_rate/action_model': 9.716050687412251e-06, 'epoch': 2.48}) +[2026-06-18 01:58:47,806][starVLA.training.train_starvla][INFO] - Step 6014, Loss: {'train/loss': 0.2662239670753479, 'train/grad_norm_pre_clip': 20.0578670501709, 'timing/data': 0.0004954654723405838, 'timing/model': 0.9554419573396444, 'learning_rate/qwen_vl_interface': 9.706660154684312e-07, 'learning_rate/action_model': 9.706660154684312e-06, 'epoch': 2.48}) +[2026-06-18 01:58:51,674][starVLA.training.train_starvla][INFO] - Step 6015, Loss: {'train/loss': 0.25879615545272827, 'train/grad_norm_pre_clip': 22.897809982299805, 'timing/data': 0.00138956680893898, 'timing/model': 0.9459845144301653, 'learning_rate/qwen_vl_interface': 9.697278493061773e-07, 'learning_rate/action_model': 9.697278493061773e-06, 'epoch': 2.48}) +[2026-06-18 01:58:55,537][starVLA.training.train_starvla][INFO] - Step 6016, Loss: {'train/loss': 0.12851834297180176, 'train/grad_norm_pre_clip': 12.658975601196289, 'timing/data': 0.000509616918861866, 'timing/model': 0.9616951905190945, 'learning_rate/qwen_vl_interface': 9.687905704489461e-07, 'learning_rate/action_model': 9.687905704489462e-06, 'epoch': 2.48}) +[2026-06-18 01:58:59,372][starVLA.training.train_starvla][INFO] - Step 6017, Loss: {'train/loss': 0.1557226926088333, 'train/grad_norm_pre_clip': 17.625869750976562, 'timing/data': 0.0005643879994750023, 'timing/model': 0.9871587101370096, 'learning_rate/qwen_vl_interface': 9.678541790910361e-07, 'learning_rate/action_model': 9.678541790910361e-06, 'epoch': 2.48}) +[2026-06-18 01:59:03,181][starVLA.training.train_starvla][INFO] - Step 6018, Loss: {'train/loss': 0.20364418625831604, 'train/grad_norm_pre_clip': 17.2221622467041, 'timing/data': 0.0005163028836250305, 'timing/model': 0.9413697263225913, 'learning_rate/qwen_vl_interface': 9.669186754265619e-07, 'learning_rate/action_model': 9.669186754265618e-06, 'epoch': 2.48}) +[2026-06-18 01:59:07,077][starVLA.training.train_starvla][INFO] - Step 6019, Loss: {'train/loss': 0.2870171368122101, 'train/grad_norm_pre_clip': 21.299734115600586, 'timing/data': 0.0005392814055085182, 'timing/model': 1.1208727983757854, 'learning_rate/qwen_vl_interface': 9.65984059649457e-07, 'learning_rate/action_model': 9.65984059649457e-06, 'epoch': 2.48}) +[2026-06-18 01:59:10,994][starVLA.training.train_starvla][INFO] - Step 6020, Loss: {'train/loss': 0.35148122906684875, 'train/grad_norm_pre_clip': 24.686607360839844, 'timing/data': 0.0005305539816617966, 'timing/model': 1.0801331037655473, 'learning_rate/qwen_vl_interface': 9.650503319534652e-07, 'learning_rate/action_model': 9.650503319534651e-06, 'epoch': 2.48}) +[2026-06-18 01:59:14,701][starVLA.training.train_starvla][INFO] - Step 6021, Loss: {'train/loss': 0.2260323166847229, 'train/grad_norm_pre_clip': 18.006147384643555, 'timing/data': 0.0005096839740872383, 'timing/model': 0.9529393529519439, 'learning_rate/qwen_vl_interface': 9.641174925321494e-07, 'learning_rate/action_model': 9.641174925321494e-06, 'epoch': 2.48}) +[2026-06-18 01:59:18,606][starVLA.training.train_starvla][INFO] - Step 6022, Loss: {'train/loss': 0.15360619127750397, 'train/grad_norm_pre_clip': 17.39208221435547, 'timing/data': 0.0004908423870801926, 'timing/model': 0.9362110560759902, 'learning_rate/qwen_vl_interface': 9.631855415788904e-07, 'learning_rate/action_model': 9.631855415788904e-06, 'epoch': 2.48}) +[2026-06-18 01:59:22,454][starVLA.training.train_starvla][INFO] - Step 6023, Loss: {'train/loss': 0.4499891698360443, 'train/grad_norm_pre_clip': 21.579578399658203, 'timing/data': 0.0006632516160607338, 'timing/model': 0.9497474329546094, 'learning_rate/qwen_vl_interface': 9.622544792868812e-07, 'learning_rate/action_model': 9.622544792868812e-06, 'epoch': 2.48}) +[2026-06-18 01:59:26,383][starVLA.training.train_starvla][INFO] - Step 6024, Loss: {'train/loss': 0.21614600718021393, 'train/grad_norm_pre_clip': 31.703351974487305, 'timing/data': 0.0013332022354006767, 'timing/model': 1.0549064399674535, 'learning_rate/qwen_vl_interface': 9.613243058491334e-07, 'learning_rate/action_model': 9.613243058491334e-06, 'epoch': 2.48}) +[2026-06-18 01:59:30,674][starVLA.training.train_starvla][INFO] - Step 6025, Loss: {'train/loss': 0.26475128531455994, 'train/grad_norm_pre_clip': 29.43528938293457, 'timing/data': 0.00048680417239665985, 'timing/model': 0.9880275297909975, 'learning_rate/qwen_vl_interface': 9.603950214584695e-07, 'learning_rate/action_model': 9.603950214584696e-06, 'epoch': 2.48}) +[2026-06-18 01:59:34,617][starVLA.training.train_starvla][INFO] - Step 6026, Loss: {'train/loss': 0.1440328061580658, 'train/grad_norm_pre_clip': 12.370399475097656, 'timing/data': 0.0004773261025547981, 'timing/model': 1.0943572875112295, 'learning_rate/qwen_vl_interface': 9.59466626307535e-07, 'learning_rate/action_model': 9.594666263075349e-06, 'epoch': 2.48}) +[2026-06-18 01:59:38,439][starVLA.training.train_starvla][INFO] - Step 6027, Loss: {'train/loss': 0.2378886193037033, 'train/grad_norm_pre_clip': 29.534988403320312, 'timing/data': 0.0005082981660962105, 'timing/model': 1.1058008084073663, 'learning_rate/qwen_vl_interface': 9.585391205887851e-07, 'learning_rate/action_model': 9.585391205887851e-06, 'epoch': 2.48}) +[2026-06-18 01:59:42,094][starVLA.training.train_starvla][INFO] - Step 6028, Loss: {'train/loss': 0.2073361873626709, 'train/grad_norm_pre_clip': 15.558218002319336, 'timing/data': 0.00051131471991539, 'timing/model': 0.9506705449894071, 'learning_rate/qwen_vl_interface': 9.57612504494494e-07, 'learning_rate/action_model': 9.57612504494494e-06, 'epoch': 2.48}) +[2026-06-18 01:59:45,948][starVLA.training.train_starvla][INFO] - Step 6029, Loss: {'train/loss': 0.16609454154968262, 'train/grad_norm_pre_clip': 23.2015380859375, 'timing/data': 0.0011195922270417213, 'timing/model': 0.9519126517698169, 'learning_rate/qwen_vl_interface': 9.566867782167494e-07, 'learning_rate/action_model': 9.566867782167494e-06, 'epoch': 2.48}) +[2026-06-18 01:59:49,786][starVLA.training.train_starvla][INFO] - Step 6030, Loss: {'train/loss': 0.11584902554750443, 'train/grad_norm_pre_clip': 12.98339557647705, 'timing/data': 0.0007819794118404388, 'timing/model': 0.9330323562026024, 'learning_rate/qwen_vl_interface': 9.55761941947455e-07, 'learning_rate/action_model': 9.557619419474551e-06, 'epoch': 2.48}) +[2026-06-18 01:59:53,726][starVLA.training.train_starvla][INFO] - Step 6031, Loss: {'train/loss': 0.23432226479053497, 'train/grad_norm_pre_clip': 18.22445297241211, 'timing/data': 0.0005162348970770836, 'timing/model': 0.9599628187716007, 'learning_rate/qwen_vl_interface': 9.548379958783308e-07, 'learning_rate/action_model': 9.548379958783308e-06, 'epoch': 2.48}) +[2026-06-18 01:59:57,577][starVLA.training.train_starvla][INFO] - Step 6032, Loss: {'train/loss': 0.13108016550540924, 'train/grad_norm_pre_clip': 14.089715957641602, 'timing/data': 0.0005113007500767708, 'timing/model': 0.9487238936126232, 'learning_rate/qwen_vl_interface': 9.539149402009114e-07, 'learning_rate/action_model': 9.539149402009115e-06, 'epoch': 2.48}) +[2026-06-18 02:00:01,576][starVLA.training.train_starvla][INFO] - Step 6033, Loss: {'train/loss': 0.21980434656143188, 'train/grad_norm_pre_clip': 19.884000778198242, 'timing/data': 0.000931568443775177, 'timing/model': 1.21971935313195, 'learning_rate/qwen_vl_interface': 9.529927751065493e-07, 'learning_rate/action_model': 9.529927751065492e-06, 'epoch': 2.48}) +[2026-06-18 02:00:05,262][starVLA.training.train_starvla][INFO] - Step 6034, Loss: {'train/loss': 0.11499001830816269, 'train/grad_norm_pre_clip': 15.660388946533203, 'timing/data': 0.0005657868459820747, 'timing/model': 0.9368165349587798, 'learning_rate/qwen_vl_interface': 9.520715007864082e-07, 'learning_rate/action_model': 9.52071500786408e-06, 'epoch': 2.48}) +[2026-06-18 02:00:09,203][starVLA.training.train_starvla][INFO] - Step 6035, Loss: {'train/loss': 0.2938084900379181, 'train/grad_norm_pre_clip': 22.977128982543945, 'timing/data': 0.0005785999819636345, 'timing/model': 0.963459481485188, 'learning_rate/qwen_vl_interface': 9.511511174314689e-07, 'learning_rate/action_model': 9.511511174314688e-06, 'epoch': 2.48}) +[2026-06-18 02:00:13,055][starVLA.training.train_starvla][INFO] - Step 6036, Loss: {'train/loss': 0.30675917863845825, 'train/grad_norm_pre_clip': 20.16529083251953, 'timing/data': 0.0008435528725385666, 'timing/model': 0.9570504371076822, 'learning_rate/qwen_vl_interface': 9.502316252325283e-07, 'learning_rate/action_model': 9.502316252325283e-06, 'epoch': 2.48}) +[2026-06-18 02:00:17,059][starVLA.training.train_starvla][INFO] - Step 6037, Loss: {'train/loss': 0.5496087670326233, 'train/grad_norm_pre_clip': 29.60202407836914, 'timing/data': 0.0011231591925024986, 'timing/model': 1.0793629409745336, 'learning_rate/qwen_vl_interface': 9.493130243801993e-07, 'learning_rate/action_model': 9.493130243801992e-06, 'epoch': 2.49}) +[2026-06-18 02:00:20,987][starVLA.training.train_starvla][INFO] - Step 6038, Loss: {'train/loss': 0.339776873588562, 'train/grad_norm_pre_clip': 28.188993453979492, 'timing/data': 0.0015245825052261353, 'timing/model': 0.9398571280762553, 'learning_rate/qwen_vl_interface': 9.483953150649091e-07, 'learning_rate/action_model': 9.483953150649091e-06, 'epoch': 2.49}) +[2026-06-18 02:00:24,871][starVLA.training.train_starvla][INFO] - Step 6039, Loss: {'train/loss': 0.06725479662418365, 'train/grad_norm_pre_clip': 9.634100914001465, 'timing/data': 0.0005283039063215256, 'timing/model': 1.1139101572334766, 'learning_rate/qwen_vl_interface': 9.474784974768982e-07, 'learning_rate/action_model': 9.474784974768982e-06, 'epoch': 2.49}) +[2026-06-18 02:00:28,698][starVLA.training.train_starvla][INFO] - Step 6040, Loss: {'train/loss': 0.24665693938732147, 'train/grad_norm_pre_clip': 21.514909744262695, 'timing/data': 0.0010688155889511108, 'timing/model': 1.1039758753031492, 'learning_rate/qwen_vl_interface': 9.465625718062236e-07, 'learning_rate/action_model': 9.465625718062235e-06, 'epoch': 2.49}) +[2026-06-18 02:00:33,516][starVLA.training.train_starvla][INFO] - Step 6041, Loss: {'train/loss': 0.09522711485624313, 'train/grad_norm_pre_clip': 12.301290512084961, 'timing/data': 0.0007800590246915817, 'timing/model': 0.9674284160137177, 'learning_rate/qwen_vl_interface': 9.456475382427591e-07, 'learning_rate/action_model': 9.45647538242759e-06, 'epoch': 2.49}) +[2026-06-18 02:00:37,577][starVLA.training.train_starvla][INFO] - Step 6042, Loss: {'train/loss': 0.281335711479187, 'train/grad_norm_pre_clip': 27.43004035949707, 'timing/data': 0.000806380994617939, 'timing/model': 0.9660889534279704, 'learning_rate/qwen_vl_interface': 9.447333969761915e-07, 'learning_rate/action_model': 9.447333969761915e-06, 'epoch': 2.49}) +[2026-06-18 02:00:41,434][starVLA.training.train_starvla][INFO] - Step 6043, Loss: {'train/loss': 0.09287643432617188, 'train/grad_norm_pre_clip': 18.10582160949707, 'timing/data': 0.0004847962409257889, 'timing/model': 0.951752089895308, 'learning_rate/qwen_vl_interface': 9.438201481960235e-07, 'learning_rate/action_model': 9.438201481960234e-06, 'epoch': 2.49}) +[2026-06-18 02:00:45,330][starVLA.training.train_starvla][INFO] - Step 6044, Loss: {'train/loss': 0.1145549938082695, 'train/grad_norm_pre_clip': 15.710851669311523, 'timing/data': 0.0008979151025414467, 'timing/model': 0.9340142048895359, 'learning_rate/qwen_vl_interface': 9.429077920915724e-07, 'learning_rate/action_model': 9.429077920915724e-06, 'epoch': 2.49}) +[2026-06-18 02:00:49,178][starVLA.training.train_starvla][INFO] - Step 6045, Loss: {'train/loss': 0.16609740257263184, 'train/grad_norm_pre_clip': 25.344141006469727, 'timing/data': 0.0005793962627649307, 'timing/model': 0.9522361503913999, 'learning_rate/qwen_vl_interface': 9.419963288519711e-07, 'learning_rate/action_model': 9.41996328851971e-06, 'epoch': 2.49}) +[2026-06-18 02:00:53,199][starVLA.training.train_starvla][INFO] - Step 6046, Loss: {'train/loss': 0.17210540175437927, 'train/grad_norm_pre_clip': 16.511075973510742, 'timing/data': 0.0008484534919261932, 'timing/model': 1.166008172556758, 'learning_rate/qwen_vl_interface': 9.410857586661659e-07, 'learning_rate/action_model': 9.410857586661659e-06, 'epoch': 2.49}) +[2026-06-18 02:00:57,184][starVLA.training.train_starvla][INFO] - Step 6047, Loss: {'train/loss': 0.5923895239830017, 'train/grad_norm_pre_clip': 26.128313064575195, 'timing/data': 0.000834382139146328, 'timing/model': 1.15966472402215, 'learning_rate/qwen_vl_interface': 9.401760817229197e-07, 'learning_rate/action_model': 9.401760817229196e-06, 'epoch': 2.49}) +[2026-06-18 02:01:01,006][starVLA.training.train_starvla][INFO] - Step 6048, Loss: {'train/loss': 0.13939331471920013, 'train/grad_norm_pre_clip': 20.033720016479492, 'timing/data': 0.0005199415609240532, 'timing/model': 0.9593083085492253, 'learning_rate/qwen_vl_interface': 9.392672982108091e-07, 'learning_rate/action_model': 9.392672982108091e-06, 'epoch': 2.49}) +[2026-06-18 02:01:06,803][starVLA.training.train_starvla][INFO] - Step 6049, Loss: {'train/loss': 0.2536972165107727, 'train/grad_norm_pre_clip': 23.788314819335938, 'timing/data': 0.0007434003055095673, 'timing/model': 0.9587140707299113, 'learning_rate/qwen_vl_interface': 9.383594083182265e-07, 'learning_rate/action_model': 9.383594083182265e-06, 'epoch': 2.49}) +[2026-06-18 02:01:10,802][starVLA.training.train_starvla][INFO] - Step 6050, Loss: {'train/loss': 0.20388299226760864, 'train/grad_norm_pre_clip': 14.995878219604492, 'timing/data': 0.0006727492436766624, 'timing/model': 0.9620372736826539, 'learning_rate/qwen_vl_interface': 9.374524122333769e-07, 'learning_rate/action_model': 9.374524122333769e-06, 'epoch': 2.49}) +[2026-06-18 02:01:16,255][starVLA.training.train_starvla][INFO] - Step 6051, Loss: {'train/loss': 0.30072569847106934, 'train/grad_norm_pre_clip': 29.606094360351562, 'timing/data': 0.0005184048786759377, 'timing/model': 0.9726620204746723, 'learning_rate/qwen_vl_interface': 9.365463101442842e-07, 'learning_rate/action_model': 9.365463101442842e-06, 'epoch': 2.49}) +[2026-06-18 02:01:20,141][starVLA.training.train_starvla][INFO] - Step 6052, Loss: {'train/loss': 0.3262060582637787, 'train/grad_norm_pre_clip': 26.88736343383789, 'timing/data': 0.0009638369083404541, 'timing/model': 0.9511307589709759, 'learning_rate/qwen_vl_interface': 9.356411022387823e-07, 'learning_rate/action_model': 9.356411022387824e-06, 'epoch': 2.49}) +[2026-06-18 02:01:24,045][starVLA.training.train_starvla][INFO] - Step 6053, Loss: {'train/loss': 0.15752433240413666, 'train/grad_norm_pre_clip': 19.807161331176758, 'timing/data': 0.0006036087870597839, 'timing/model': 1.1315100528299809, 'learning_rate/qwen_vl_interface': 9.347367887045236e-07, 'learning_rate/action_model': 9.347367887045236e-06, 'epoch': 2.49}) +[2026-06-18 02:01:27,809][starVLA.training.train_starvla][INFO] - Step 6054, Loss: {'train/loss': 0.3871605098247528, 'train/grad_norm_pre_clip': 24.828292846679688, 'timing/data': 0.0004665600135922432, 'timing/model': 0.950416088104248, 'learning_rate/qwen_vl_interface': 9.338333697289697e-07, 'learning_rate/action_model': 9.338333697289696e-06, 'epoch': 2.49}) +[2026-06-18 02:01:31,636][starVLA.training.train_starvla][INFO] - Step 6055, Loss: {'train/loss': 0.10485810041427612, 'train/grad_norm_pre_clip': 12.701339721679688, 'timing/data': 0.0005654608830809593, 'timing/model': 0.949520961381495, 'learning_rate/qwen_vl_interface': 9.329308454994044e-07, 'learning_rate/action_model': 9.329308454994043e-06, 'epoch': 2.49}) +[2026-06-18 02:01:35,488][starVLA.training.train_starvla][INFO] - Step 6056, Loss: {'train/loss': 0.6882261633872986, 'train/grad_norm_pre_clip': 51.541648864746094, 'timing/data': 0.0005954932421445847, 'timing/model': 1.0079065365716815, 'learning_rate/qwen_vl_interface': 9.320292162029192e-07, 'learning_rate/action_model': 9.320292162029191e-06, 'epoch': 2.49}) +[2026-06-18 02:01:39,332][starVLA.training.train_starvla][INFO] - Step 6057, Loss: {'train/loss': 0.16177701950073242, 'train/grad_norm_pre_clip': 18.776752471923828, 'timing/data': 0.0005809301510453224, 'timing/model': 0.9511758396402001, 'learning_rate/qwen_vl_interface': 9.311284820264245e-07, 'learning_rate/action_model': 9.311284820264244e-06, 'epoch': 2.49}) +[2026-06-18 02:01:43,162][starVLA.training.train_starvla][INFO] - Step 6058, Loss: {'train/loss': 0.2373019903898239, 'train/grad_norm_pre_clip': 19.62498664855957, 'timing/data': 0.00047613680362701416, 'timing/model': 0.9360174797475338, 'learning_rate/qwen_vl_interface': 9.302286431566428e-07, 'learning_rate/action_model': 9.302286431566428e-06, 'epoch': 2.49}) +[2026-06-18 02:01:47,263][starVLA.training.train_starvla][INFO] - Step 6059, Loss: {'train/loss': 0.21594947576522827, 'train/grad_norm_pre_clip': 29.383602142333984, 'timing/data': 0.0005638357251882553, 'timing/model': 1.3790216417983174, 'learning_rate/qwen_vl_interface': 9.293296997801113e-07, 'learning_rate/action_model': 9.293296997801112e-06, 'epoch': 2.49}) +[2026-06-18 02:01:51,374][starVLA.training.train_starvla][INFO] - Step 6060, Loss: {'train/loss': 0.06809581816196442, 'train/grad_norm_pre_clip': 9.080880165100098, 'timing/data': 0.00048203859478235245, 'timing/model': 1.1246319198980927, 'learning_rate/qwen_vl_interface': 9.284316520831825e-07, 'learning_rate/action_model': 9.284316520831825e-06, 'epoch': 2.49}) +[2026-06-18 02:01:55,095][starVLA.training.train_starvla][INFO] - Step 6061, Loss: {'train/loss': 0.09906932711601257, 'train/grad_norm_pre_clip': 10.339885711669922, 'timing/data': 0.001293259672820568, 'timing/model': 1.0066414838656783, 'learning_rate/qwen_vl_interface': 9.275345002520225e-07, 'learning_rate/action_model': 9.275345002520225e-06, 'epoch': 2.49}) +[2026-06-18 02:01:59,013][starVLA.training.train_starvla][INFO] - Step 6062, Loss: {'train/loss': 0.247167706489563, 'train/grad_norm_pre_clip': 22.314847946166992, 'timing/data': 0.0006410805508494377, 'timing/model': 0.9548867428675294, 'learning_rate/qwen_vl_interface': 9.266382444726117e-07, 'learning_rate/action_model': 9.266382444726117e-06, 'epoch': 2.5}) +[2026-06-18 02:02:02,944][starVLA.training.train_starvla][INFO] - Step 6063, Loss: {'train/loss': 0.18929515779018402, 'train/grad_norm_pre_clip': 13.667037963867188, 'timing/data': 0.0009618280455470085, 'timing/model': 0.9431185973808169, 'learning_rate/qwen_vl_interface': 9.257428849307446e-07, 'learning_rate/action_model': 9.257428849307445e-06, 'epoch': 2.5}) +[2026-06-18 02:02:06,980][starVLA.training.train_starvla][INFO] - Step 6064, Loss: {'train/loss': 0.14266064763069153, 'train/grad_norm_pre_clip': 13.858805656433105, 'timing/data': 0.0006785821169614792, 'timing/model': 0.9516066564247012, 'learning_rate/qwen_vl_interface': 9.248484218120307e-07, 'learning_rate/action_model': 9.248484218120307e-06, 'epoch': 2.5}) +[2026-06-18 02:02:11,037][starVLA.training.train_starvla][INFO] - Step 6065, Loss: {'train/loss': 0.2104218602180481, 'train/grad_norm_pre_clip': 15.364334106445312, 'timing/data': 0.0006997715681791306, 'timing/model': 0.9789919797331095, 'learning_rate/qwen_vl_interface': 9.239548553018925e-07, 'learning_rate/action_model': 9.239548553018924e-06, 'epoch': 2.5}) +[2026-06-18 02:02:15,242][starVLA.training.train_starvla][INFO] - Step 6066, Loss: {'train/loss': 0.2514009177684784, 'train/grad_norm_pre_clip': 22.823884963989258, 'timing/data': 0.0005193976685404778, 'timing/model': 1.3881736192852259, 'learning_rate/qwen_vl_interface': 9.230621855855687e-07, 'learning_rate/action_model': 9.230621855855687e-06, 'epoch': 2.5}) +[2026-06-18 02:02:19,128][starVLA.training.train_starvla][INFO] - Step 6067, Loss: {'train/loss': 0.2063649296760559, 'train/grad_norm_pre_clip': 14.168527603149414, 'timing/data': 0.0005001109093427658, 'timing/model': 1.1626646146178246, 'learning_rate/qwen_vl_interface': 9.221704128481107e-07, 'learning_rate/action_model': 9.221704128481106e-06, 'epoch': 2.5}) +[2026-06-18 02:02:22,960][starVLA.training.train_starvla][INFO] - Step 6068, Loss: {'train/loss': 0.29134610295295715, 'train/grad_norm_pre_clip': 28.33797264099121, 'timing/data': 0.0008210893720388412, 'timing/model': 0.9680183567106724, 'learning_rate/qwen_vl_interface': 9.212795372743815e-07, 'learning_rate/action_model': 9.212795372743813e-06, 'epoch': 2.5}) +[2026-06-18 02:02:26,869][starVLA.training.train_starvla][INFO] - Step 6069, Loss: {'train/loss': 0.296369731426239, 'train/grad_norm_pre_clip': 36.544189453125, 'timing/data': 0.0006076851859688759, 'timing/model': 0.9453242495656013, 'learning_rate/qwen_vl_interface': 9.203895590490617e-07, 'learning_rate/action_model': 9.203895590490616e-06, 'epoch': 2.5}) +[2026-06-18 02:02:31,067][starVLA.training.train_starvla][INFO] - Step 6070, Loss: {'train/loss': 0.12010939419269562, 'train/grad_norm_pre_clip': 13.458623886108398, 'timing/data': 0.00047334562987089157, 'timing/model': 0.9445308651775122, 'learning_rate/qwen_vl_interface': 9.195004783566464e-07, 'learning_rate/action_model': 9.195004783566463e-06, 'epoch': 2.5}) +[2026-06-18 02:02:34,988][starVLA.training.train_starvla][INFO] - Step 6071, Loss: {'train/loss': 0.27314168214797974, 'train/grad_norm_pre_clip': 21.224613189697266, 'timing/data': 0.0014904867857694626, 'timing/model': 0.9924168549478054, 'learning_rate/qwen_vl_interface': 9.186122953814414e-07, 'learning_rate/action_model': 9.186122953814413e-06, 'epoch': 2.5}) +[2026-06-18 02:02:38,892][starVLA.training.train_starvla][INFO] - Step 6072, Loss: {'train/loss': 0.18739546835422516, 'train/grad_norm_pre_clip': 17.321794509887695, 'timing/data': 0.0005332315340638161, 'timing/model': 0.9338536961004138, 'learning_rate/qwen_vl_interface': 9.177250103075686e-07, 'learning_rate/action_model': 9.177250103075686e-06, 'epoch': 2.5}) +[2026-06-18 02:02:42,767][starVLA.training.train_starvla][INFO] - Step 6073, Loss: {'train/loss': 0.22629335522651672, 'train/grad_norm_pre_clip': 19.64783477783203, 'timing/data': 0.0005323905497789383, 'timing/model': 1.1495771147310734, 'learning_rate/qwen_vl_interface': 9.168386233189627e-07, 'learning_rate/action_model': 9.168386233189627e-06, 'epoch': 2.5}) +[2026-06-18 02:02:46,582][starVLA.training.train_starvla][INFO] - Step 6074, Loss: {'train/loss': 0.2962583005428314, 'train/grad_norm_pre_clip': 21.161836624145508, 'timing/data': 0.0005182735621929169, 'timing/model': 0.9508002940565348, 'learning_rate/qwen_vl_interface': 9.159531345993726e-07, 'learning_rate/action_model': 9.159531345993725e-06, 'epoch': 2.5}) +[2026-06-18 02:02:50,570][starVLA.training.train_starvla][INFO] - Step 6075, Loss: {'train/loss': 0.39370155334472656, 'train/grad_norm_pre_clip': 20.920635223388672, 'timing/data': 0.0005140043795108795, 'timing/model': 0.9590571112930775, 'learning_rate/qwen_vl_interface': 9.150685443323619e-07, 'learning_rate/action_model': 9.150685443323618e-06, 'epoch': 2.5}) +[2026-06-18 02:02:56,040][starVLA.training.train_starvla][INFO] - Step 6076, Loss: {'train/loss': 0.1491689234972, 'train/grad_norm_pre_clip': 19.366336822509766, 'timing/data': 0.0004968922585248947, 'timing/model': 0.981153691187501, 'learning_rate/qwen_vl_interface': 9.141848527013059e-07, 'learning_rate/action_model': 9.141848527013059e-06, 'epoch': 2.5}) +[2026-06-18 02:02:59,975][starVLA.training.train_starvla][INFO] - Step 6077, Loss: {'train/loss': 0.25009775161743164, 'train/grad_norm_pre_clip': 14.106189727783203, 'timing/data': 0.0005123605951666832, 'timing/model': 0.9628889886662364, 'learning_rate/qwen_vl_interface': 9.13302059889396e-07, 'learning_rate/action_model': 9.13302059889396e-06, 'epoch': 2.5}) +[2026-06-18 02:03:04,268][starVLA.training.train_starvla][INFO] - Step 6078, Loss: {'train/loss': 0.32155513763427734, 'train/grad_norm_pre_clip': 25.856517791748047, 'timing/data': 0.0007467512041330338, 'timing/model': 0.9806723706424236, 'learning_rate/qwen_vl_interface': 9.12420166079635e-07, 'learning_rate/action_model': 9.12420166079635e-06, 'epoch': 2.5}) +[2026-06-18 02:03:08,458][starVLA.training.train_starvla][INFO] - Step 6079, Loss: {'train/loss': 0.24147319793701172, 'train/grad_norm_pre_clip': 16.04082489013672, 'timing/data': 0.0008698832243680954, 'timing/model': 1.184767640195787, 'learning_rate/qwen_vl_interface': 9.115391714548418e-07, 'learning_rate/action_model': 9.115391714548418e-06, 'epoch': 2.5}) +[2026-06-18 02:03:12,554][starVLA.training.train_starvla][INFO] - Step 6080, Loss: {'train/loss': 0.5520311594009399, 'train/grad_norm_pre_clip': 25.693252563476562, 'timing/data': 0.0007593855261802673, 'timing/model': 1.2364089591428638, 'learning_rate/qwen_vl_interface': 9.106590761976454e-07, 'learning_rate/action_model': 9.106590761976455e-06, 'epoch': 2.5}) +[2026-06-18 02:03:16,278][starVLA.training.train_starvla][INFO] - Step 6081, Loss: {'train/loss': 0.15386734902858734, 'train/grad_norm_pre_clip': 19.71634292602539, 'timing/data': 0.0005773426964879036, 'timing/model': 0.9449202939867973, 'learning_rate/qwen_vl_interface': 9.097798804904934e-07, 'learning_rate/action_model': 9.097798804904933e-06, 'epoch': 2.5}) +[2026-06-18 02:03:20,336][starVLA.training.train_starvla][INFO] - Step 6082, Loss: {'train/loss': 0.4011631906032562, 'train/grad_norm_pre_clip': 26.898103713989258, 'timing/data': 0.00048078596591949463, 'timing/model': 0.9649083083495498, 'learning_rate/qwen_vl_interface': 9.089015845156422e-07, 'learning_rate/action_model': 9.089015845156422e-06, 'epoch': 2.5}) +[2026-06-18 02:03:24,194][starVLA.training.train_starvla][INFO] - Step 6083, Loss: {'train/loss': 0.2872796058654785, 'train/grad_norm_pre_clip': 19.340375900268555, 'timing/data': 0.000732039101421833, 'timing/model': 0.9493781644850969, 'learning_rate/qwen_vl_interface': 9.080241884551627e-07, 'learning_rate/action_model': 9.080241884551627e-06, 'epoch': 2.5}) +[2026-06-18 02:03:28,058][starVLA.training.train_starvla][INFO] - Step 6084, Loss: {'train/loss': 0.36603209376335144, 'train/grad_norm_pre_clip': 35.4208984375, 'timing/data': 0.0004921602085232735, 'timing/model': 0.9394266307353973, 'learning_rate/qwen_vl_interface': 9.071476924909417e-07, 'learning_rate/action_model': 9.071476924909416e-06, 'epoch': 2.5}) +[2026-06-18 02:03:31,936][starVLA.training.train_starvla][INFO] - Step 6085, Loss: {'train/loss': 0.3687342405319214, 'train/grad_norm_pre_clip': 23.77177619934082, 'timing/data': 0.0008067535236477852, 'timing/model': 0.9343153545632958, 'learning_rate/qwen_vl_interface': 9.062720968046773e-07, 'learning_rate/action_model': 9.062720968046773e-06, 'epoch': 2.5}) +[2026-06-18 02:03:35,892][starVLA.training.train_starvla][INFO] - Step 6086, Loss: {'train/loss': 0.3364461064338684, 'train/grad_norm_pre_clip': 18.395917892456055, 'timing/data': 0.0005844570696353912, 'timing/model': 1.2203545942902565, 'learning_rate/qwen_vl_interface': 9.053974015778817e-07, 'learning_rate/action_model': 9.053974015778817e-06, 'epoch': 2.51}) +[2026-06-18 02:03:39,735][starVLA.training.train_starvla][INFO] - Step 6087, Loss: {'train/loss': 0.15338361263275146, 'train/grad_norm_pre_clip': 23.312082290649414, 'timing/data': 0.0011476818472146988, 'timing/model': 1.1213114717975259, 'learning_rate/qwen_vl_interface': 9.045236069918788e-07, 'learning_rate/action_model': 9.045236069918788e-06, 'epoch': 2.51}) +[2026-06-18 02:03:43,435][starVLA.training.train_starvla][INFO] - Step 6088, Loss: {'train/loss': 0.21898582577705383, 'train/grad_norm_pre_clip': 19.050540924072266, 'timing/data': 0.0008266875520348549, 'timing/model': 0.9478539479896426, 'learning_rate/qwen_vl_interface': 9.036507132278085e-07, 'learning_rate/action_model': 9.036507132278086e-06, 'epoch': 2.51}) +[2026-06-18 02:03:47,256][starVLA.training.train_starvla][INFO] - Step 6089, Loss: {'train/loss': 0.2873992919921875, 'train/grad_norm_pre_clip': 26.47690773010254, 'timing/data': 0.000814768485724926, 'timing/model': 0.9624705780297518, 'learning_rate/qwen_vl_interface': 9.02778720466622e-07, 'learning_rate/action_model': 9.02778720466622e-06, 'epoch': 2.51}) +[2026-06-18 02:03:51,190][starVLA.training.train_starvla][INFO] - Step 6090, Loss: {'train/loss': 0.1498582363128662, 'train/grad_norm_pre_clip': 16.12041664123535, 'timing/data': 0.0009441720321774483, 'timing/model': 0.9565732292830944, 'learning_rate/qwen_vl_interface': 9.019076288890852e-07, 'learning_rate/action_model': 9.019076288890852e-06, 'epoch': 2.51}) +[2026-06-18 02:03:55,169][starVLA.training.train_starvla][INFO] - Step 6091, Loss: {'train/loss': 0.3586668074131012, 'train/grad_norm_pre_clip': 35.78258514404297, 'timing/data': 0.0007244311273097992, 'timing/model': 0.9429308725520968, 'learning_rate/qwen_vl_interface': 9.010374386757755e-07, 'learning_rate/action_model': 9.010374386757754e-06, 'epoch': 2.51}) +[2026-06-18 02:04:01,556][starVLA.training.train_starvla][INFO] - Step 6092, Loss: {'train/loss': 0.27994105219841003, 'train/grad_norm_pre_clip': 16.759904861450195, 'timing/data': 0.0004681255668401718, 'timing/model': 1.1839303839951754, 'learning_rate/qwen_vl_interface': 9.001681500070842e-07, 'learning_rate/action_model': 9.001681500070842e-06, 'epoch': 2.51}) +[2026-06-18 02:04:05,462][starVLA.training.train_starvla][INFO] - Step 6093, Loss: {'train/loss': 0.1128464937210083, 'train/grad_norm_pre_clip': 14.584527015686035, 'timing/data': 0.0005185166373848915, 'timing/model': 1.1383304502815008, 'learning_rate/qwen_vl_interface': 8.992997630632161e-07, 'learning_rate/action_model': 8.99299763063216e-06, 'epoch': 2.51}) +[2026-06-18 02:04:09,145][starVLA.training.train_starvla][INFO] - Step 6094, Loss: {'train/loss': 0.387493759393692, 'train/grad_norm_pre_clip': 21.97829818725586, 'timing/data': 0.00070175901055336, 'timing/model': 0.9500020015984774, 'learning_rate/qwen_vl_interface': 8.984322780241879e-07, 'learning_rate/action_model': 8.984322780241878e-06, 'epoch': 2.51}) +[2026-06-18 02:04:13,780][starVLA.training.train_starvla][INFO] - Step 6095, Loss: {'train/loss': 0.08424314111471176, 'train/grad_norm_pre_clip': 14.085793495178223, 'timing/data': 0.0010348744690418243, 'timing/model': 0.9888808727264404, 'learning_rate/qwen_vl_interface': 8.975656950698325e-07, 'learning_rate/action_model': 8.975656950698324e-06, 'epoch': 2.51}) +[2026-06-18 02:04:17,701][starVLA.training.train_starvla][INFO] - Step 6096, Loss: {'train/loss': 0.20405487716197968, 'train/grad_norm_pre_clip': 19.136425018310547, 'timing/data': 0.0005400665104389191, 'timing/model': 0.9656412219628692, 'learning_rate/qwen_vl_interface': 8.967000143797913e-07, 'learning_rate/action_model': 8.967000143797913e-06, 'epoch': 2.51}) +[2026-06-18 02:04:21,577][starVLA.training.train_starvla][INFO] - Step 6097, Loss: {'train/loss': 0.18805356323719025, 'train/grad_norm_pre_clip': 17.53729248046875, 'timing/data': 0.000528920441865921, 'timing/model': 0.9564200025051832, 'learning_rate/qwen_vl_interface': 8.958352361335214e-07, 'learning_rate/action_model': 8.958352361335213e-06, 'epoch': 2.51}) +[2026-06-18 02:04:25,506][starVLA.training.train_starvla][INFO] - Step 6098, Loss: {'train/loss': 0.6817898750305176, 'train/grad_norm_pre_clip': 30.225011825561523, 'timing/data': 0.0009766584262251854, 'timing/model': 0.9540520394220948, 'learning_rate/qwen_vl_interface': 8.949713605102917e-07, 'learning_rate/action_model': 8.949713605102916e-06, 'epoch': 2.51}) +[2026-06-18 02:04:29,315][starVLA.training.train_starvla][INFO] - Step 6099, Loss: {'train/loss': 0.09538041800260544, 'train/grad_norm_pre_clip': 15.513266563415527, 'timing/data': 0.0011953162029385567, 'timing/model': 1.1173211066052318, 'learning_rate/qwen_vl_interface': 8.941083876891857e-07, 'learning_rate/action_model': 8.941083876891856e-06, 'epoch': 2.51}) +[2026-06-18 02:06:40,172][starVLA.training.train_starvla][INFO] - Step 6100, Loss: {'train/loss': 0.3578997850418091, 'train/grad_norm_pre_clip': 18.679269790649414, 'eval/loss': 0.4162292182445526, 'eval/latency_6/loss': 0.4162292182445526, 'eval/demon_fire_precision': 0.962700964630225, 'eval/demon_fire_recall': 0.9676793794440853, 'eval/demon_fire_f1': 0.965183752417795, 'eval/demon_fire_rate_model': 0.8653311074012242, 'eval/demon_fire_rate_teacher': 0.860879243183083, 'eval/demon_left_precision': 0.8818737270875764, 'eval/demon_left_recall': 0.7596491228070176, 'eval/demon_left_f1': 0.8162111215834119, 'eval/demon_left_rate_model': 0.13661658319421258, 'eval/demon_left_rate_teacher': 0.15859766277128548, 'eval/demon_right_precision': 0.839607201309329, 'eval/demon_right_recall': 0.7880184331797235, 'eval/demon_right_f1': 0.8129952456418383, 'eval/demon_right_rate_model': 0.17000556483027268, 'eval/demon_right_rate_teacher': 0.18113522537562604, 'eval/demon_move_f1': 0.8146031836126251, 'eval/demon_cc_f1': 0.9144713495888133, 'eval/latency_6/demon_fire_precision': 0.962700964630225, 'eval/latency_6/demon_fire_recall': 0.9676793794440853, 'eval/latency_6/demon_fire_f1': 0.965183752417795, 'eval/latency_6/demon_fire_rate_model': 0.8653311074012242, 'eval/latency_6/demon_fire_rate_teacher': 0.860879243183083, 'eval/latency_6/demon_left_precision': 0.8818737270875764, 'eval/latency_6/demon_left_recall': 0.7596491228070176, 'eval/latency_6/demon_left_f1': 0.8162111215834119, 'eval/latency_6/demon_left_rate_model': 0.13661658319421258, 'eval/latency_6/demon_left_rate_teacher': 0.15859766277128548, 'eval/latency_6/demon_right_precision': 0.839607201309329, 'eval/latency_6/demon_right_recall': 0.7880184331797235, 'eval/latency_6/demon_right_f1': 0.8129952456418383, 'eval/latency_6/demon_right_rate_model': 0.17000556483027268, 'eval/latency_6/demon_right_rate_teacher': 0.18113522537562604, 'eval/latency_6/demon_move_f1': 0.8146031836126251, 'eval/latency_6/demon_cc_f1': 0.9144713495888133, 'timing/data': 0.0005045458674430847, 'timing/model': 1.0876592081040144, 'learning_rate/qwen_vl_interface': 8.932463178490991e-07, 'learning_rate/action_model': 8.93246317849099e-06, 'epoch': 2.51}) +[2026-06-18 02:06:40,175][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_6100_state +[2026-06-18 02:06:40,180][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 02:06:46,850][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_6100_state/model.safetensors +[2026-06-18 02:07:02,635][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_6100_state/optimizer.bin +[2026-06-18 02:07:02,636][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_6100_state/sampler.bin +[2026-06-18 02:07:02,636][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_6100_state/sampler_1.bin +[2026-06-18 02:07:02,638][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_6100_state/random_states_0.pkl +[2026-06-18 02:07:02,642][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 6100 +[2026-06-18 02:07:02,642][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 02:07:02,652][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 02:07:06,484][starVLA.training.train_starvla][INFO] - Step 6101, Loss: {'train/loss': 0.1438182145357132, 'train/grad_norm_pre_clip': 15.371240615844727, 'timing/data': 0.0005440497770905495, 'timing/model': 1.1432487918063998, 'learning_rate/qwen_vl_interface': 8.923851511687379e-07, 'learning_rate/action_model': 8.923851511687378e-06, 'epoch': 2.51}) +[2026-06-18 02:07:10,360][starVLA.training.train_starvla][INFO] - Step 6102, Loss: {'train/loss': 0.3217248320579529, 'train/grad_norm_pre_clip': 35.091251373291016, 'timing/data': 0.000895993784070015, 'timing/model': 1.196517076343298, 'learning_rate/qwen_vl_interface': 8.915248878266228e-07, 'learning_rate/action_model': 8.915248878266227e-06, 'epoch': 2.51}) +[2026-06-18 02:07:13,989][starVLA.training.train_starvla][INFO] - Step 6103, Loss: {'train/loss': 0.13117262721061707, 'train/grad_norm_pre_clip': 16.773052215576172, 'timing/data': 0.0007278472185134888, 'timing/model': 0.9368379525840282, 'learning_rate/qwen_vl_interface': 8.906655280010895e-07, 'learning_rate/action_model': 8.906655280010896e-06, 'epoch': 2.51}) +[2026-06-18 02:07:17,819][starVLA.training.train_starvla][INFO] - Step 6104, Loss: {'train/loss': 0.6182451248168945, 'train/grad_norm_pre_clip': 22.866308212280273, 'timing/data': 0.0005637165158987045, 'timing/model': 0.928498893044889, 'learning_rate/qwen_vl_interface': 8.898070718702831e-07, 'learning_rate/action_model': 8.898070718702831e-06, 'epoch': 2.51}) +[2026-06-18 02:07:21,637][starVLA.training.train_starvla][INFO] - Step 6105, Loss: {'train/loss': 0.1853909194469452, 'train/grad_norm_pre_clip': 24.966032028198242, 'timing/data': 0.001003309153020382, 'timing/model': 0.9341377252712846, 'learning_rate/qwen_vl_interface': 8.889495196121628e-07, 'learning_rate/action_model': 8.889495196121627e-06, 'epoch': 2.51}) +[2026-06-18 02:07:25,460][starVLA.training.train_starvla][INFO] - Step 6106, Loss: {'train/loss': 0.43157488107681274, 'train/grad_norm_pre_clip': 31.131044387817383, 'timing/data': 0.0010927114635705948, 'timing/model': 0.9328422248363495, 'learning_rate/qwen_vl_interface': 8.880928714044994e-07, 'learning_rate/action_model': 8.880928714044993e-06, 'epoch': 2.51}) +[2026-06-18 02:07:29,382][starVLA.training.train_starvla][INFO] - Step 6107, Loss: {'train/loss': 0.2001880258321762, 'train/grad_norm_pre_clip': 48.54427719116211, 'timing/data': 0.0009269211441278458, 'timing/model': 1.0273865777999163, 'learning_rate/qwen_vl_interface': 8.872371274248773e-07, 'learning_rate/action_model': 8.872371274248772e-06, 'epoch': 2.51}) +[2026-06-18 02:07:33,199][starVLA.training.train_starvla][INFO] - Step 6108, Loss: {'train/loss': 0.4433818459510803, 'train/grad_norm_pre_clip': 25.28106117248535, 'timing/data': 0.000529835931956768, 'timing/model': 0.9345252923667431, 'learning_rate/qwen_vl_interface': 8.863822878506931e-07, 'learning_rate/action_model': 8.86382287850693e-06, 'epoch': 2.51}) +[2026-06-18 02:07:36,975][starVLA.training.train_starvla][INFO] - Step 6109, Loss: {'train/loss': 0.36222967505455017, 'train/grad_norm_pre_clip': 32.1054573059082, 'timing/data': 0.001149606890976429, 'timing/model': 1.1111240880563855, 'learning_rate/qwen_vl_interface': 8.855283528591566e-07, 'learning_rate/action_model': 8.855283528591565e-06, 'epoch': 2.51}) +[2026-06-18 02:07:40,597][starVLA.training.train_starvla][INFO] - Step 6110, Loss: {'train/loss': 0.2308909296989441, 'train/grad_norm_pre_clip': 19.773332595825195, 'timing/data': 0.00119088776409626, 'timing/model': 0.9514506682753563, 'learning_rate/qwen_vl_interface': 8.846753226272884e-07, 'learning_rate/action_model': 8.846753226272883e-06, 'epoch': 2.52}) +[2026-06-18 02:07:44,546][starVLA.training.train_starvla][INFO] - Step 6111, Loss: {'train/loss': 0.41713401675224304, 'train/grad_norm_pre_clip': 30.416900634765625, 'timing/data': 0.0005763936787843704, 'timing/model': 0.9485226692631841, 'learning_rate/qwen_vl_interface': 8.838231973319235e-07, 'learning_rate/action_model': 8.838231973319235e-06, 'epoch': 2.52}) +[2026-06-18 02:07:48,716][starVLA.training.train_starvla][INFO] - Step 6112, Loss: {'train/loss': 0.08402231335639954, 'train/grad_norm_pre_clip': 18.183095932006836, 'timing/data': 0.0005293162539601326, 'timing/model': 0.9322870802134275, 'learning_rate/qwen_vl_interface': 8.829719771497072e-07, 'learning_rate/action_model': 8.829719771497071e-06, 'epoch': 2.52}) +[2026-06-18 02:07:52,490][starVLA.training.train_starvla][INFO] - Step 6113, Loss: {'train/loss': 0.4577401578426361, 'train/grad_norm_pre_clip': 26.24000358581543, 'timing/data': 0.0006192130967974663, 'timing/model': 0.9398557329550385, 'learning_rate/qwen_vl_interface': 8.821216622570999e-07, 'learning_rate/action_model': 8.821216622570998e-06, 'epoch': 2.52}) +[2026-06-18 02:07:56,577][starVLA.training.train_starvla][INFO] - Step 6114, Loss: {'train/loss': 0.21545527875423431, 'train/grad_norm_pre_clip': 20.30788230895996, 'timing/data': 0.0005158344283699989, 'timing/model': 1.1730539733543992, 'learning_rate/qwen_vl_interface': 8.812722528303727e-07, 'learning_rate/action_model': 8.812722528303727e-06, 'epoch': 2.52}) +[2026-06-18 02:08:00,548][starVLA.training.train_starvla][INFO] - Step 6115, Loss: {'train/loss': 0.29584401845932007, 'train/grad_norm_pre_clip': 23.746553421020508, 'timing/data': 0.0005499338731169701, 'timing/model': 1.0952448872849345, 'learning_rate/qwen_vl_interface': 8.804237490456076e-07, 'learning_rate/action_model': 8.804237490456075e-06, 'epoch': 2.52}) +[2026-06-18 02:08:04,362][starVLA.training.train_starvla][INFO] - Step 6116, Loss: {'train/loss': 0.3961039185523987, 'train/grad_norm_pre_clip': 20.733592987060547, 'timing/data': 0.0011828718706965446, 'timing/model': 1.1215244950726628, 'learning_rate/qwen_vl_interface': 8.795761510787007e-07, 'learning_rate/action_model': 8.795761510787006e-06, 'epoch': 2.52}) +[2026-06-18 02:08:07,977][starVLA.training.train_starvla][INFO] - Step 6117, Loss: {'train/loss': 0.18811586499214172, 'train/grad_norm_pre_clip': 22.104114532470703, 'timing/data': 0.0010668374598026276, 'timing/model': 0.9325782926753163, 'learning_rate/qwen_vl_interface': 8.78729459105361e-07, 'learning_rate/action_model': 8.78729459105361e-06, 'epoch': 2.52}) +[2026-06-18 02:08:11,840][starVLA.training.train_starvla][INFO] - Step 6118, Loss: {'train/loss': 0.6761263012886047, 'train/grad_norm_pre_clip': 38.24412155151367, 'timing/data': 0.0006880881264805794, 'timing/model': 0.9519937485456467, 'learning_rate/qwen_vl_interface': 8.778836733011079e-07, 'learning_rate/action_model': 8.778836733011079e-06, 'epoch': 2.52}) +[2026-06-18 02:08:15,811][starVLA.training.train_starvla][INFO] - Step 6119, Loss: {'train/loss': 0.5052437782287598, 'train/grad_norm_pre_clip': 27.970287322998047, 'timing/data': 0.0010157478973269463, 'timing/model': 0.9445103583857417, 'learning_rate/qwen_vl_interface': 8.770387938412753e-07, 'learning_rate/action_model': 8.770387938412753e-06, 'epoch': 2.52}) +[2026-06-18 02:08:19,734][starVLA.training.train_starvla][INFO] - Step 6120, Loss: {'train/loss': 0.2300139218568802, 'train/grad_norm_pre_clip': 33.044219970703125, 'timing/data': 0.0009518731385469437, 'timing/model': 0.9283345071598887, 'learning_rate/qwen_vl_interface': 8.76194820901004e-07, 'learning_rate/action_model': 8.76194820901004e-06, 'epoch': 2.52}) +[2026-06-18 02:08:23,650][starVLA.training.train_starvla][INFO] - Step 6121, Loss: {'train/loss': 0.1551501750946045, 'train/grad_norm_pre_clip': 17.345766067504883, 'timing/data': 0.0005339896306395531, 'timing/model': 0.930384530685842, 'learning_rate/qwen_vl_interface': 8.753517546552545e-07, 'learning_rate/action_model': 8.753517546552544e-06, 'epoch': 2.52}) +[2026-06-18 02:08:27,823][starVLA.training.train_starvla][INFO] - Step 6122, Loss: {'train/loss': 0.34621357917785645, 'train/grad_norm_pre_clip': 24.06951904296875, 'timing/data': 0.0012192269787192345, 'timing/model': 1.2530700154602528, 'learning_rate/qwen_vl_interface': 8.74509595278793e-07, 'learning_rate/action_model': 8.74509595278793e-06, 'epoch': 2.52}) +[2026-06-18 02:08:31,732][starVLA.training.train_starvla][INFO] - Step 6123, Loss: {'train/loss': 0.20204727351665497, 'train/grad_norm_pre_clip': 21.59384536743164, 'timing/data': 0.0005686609074473381, 'timing/model': 1.2097326163202524, 'learning_rate/qwen_vl_interface': 8.736683429462003e-07, 'learning_rate/action_model': 8.736683429462004e-06, 'epoch': 2.52}) +[2026-06-18 02:08:35,378][starVLA.training.train_starvla][INFO] - Step 6124, Loss: {'train/loss': 0.523404598236084, 'train/grad_norm_pre_clip': 27.66706657409668, 'timing/data': 0.0007782196626067162, 'timing/model': 0.9336243970319629, 'learning_rate/qwen_vl_interface': 8.728279978318696e-07, 'learning_rate/action_model': 8.728279978318695e-06, 'epoch': 2.52}) +[2026-06-18 02:08:39,228][starVLA.training.train_starvla][INFO] - Step 6125, Loss: {'train/loss': 0.577102541923523, 'train/grad_norm_pre_clip': 85.59044647216797, 'timing/data': 0.0005654413253068924, 'timing/model': 0.9345538187772036, 'learning_rate/qwen_vl_interface': 8.719885601100052e-07, 'learning_rate/action_model': 8.719885601100051e-06, 'epoch': 2.52}) +[2026-06-18 02:08:43,091][starVLA.training.train_starvla][INFO] - Step 6126, Loss: {'train/loss': 0.21013306081295013, 'train/grad_norm_pre_clip': 17.15836524963379, 'timing/data': 0.0005504721775650978, 'timing/model': 0.935897939838469, 'learning_rate/qwen_vl_interface': 8.71150029954623e-07, 'learning_rate/action_model': 8.71150029954623e-06, 'epoch': 2.52}) +[2026-06-18 02:08:46,943][starVLA.training.train_starvla][INFO] - Step 6127, Loss: {'train/loss': 0.3902018964290619, 'train/grad_norm_pre_clip': 24.210891723632812, 'timing/data': 0.0012485338374972343, 'timing/model': 0.9335224889218807, 'learning_rate/qwen_vl_interface': 8.703124075395504e-07, 'learning_rate/action_model': 8.703124075395504e-06, 'epoch': 2.52}) +[2026-06-18 02:08:50,789][starVLA.training.train_starvla][INFO] - Step 6128, Loss: {'train/loss': 0.1620253324508667, 'train/grad_norm_pre_clip': 16.295835494995117, 'timing/data': 0.0008208267390727997, 'timing/model': 0.9303674921393394, 'learning_rate/qwen_vl_interface': 8.694756930384305e-07, 'learning_rate/action_model': 8.694756930384306e-06, 'epoch': 2.52}) +[2026-06-18 02:08:54,569][starVLA.training.train_starvla][INFO] - Step 6129, Loss: {'train/loss': 0.14391541481018066, 'train/grad_norm_pre_clip': 13.563715934753418, 'timing/data': 0.0008733756840229034, 'timing/model': 0.9263148810714483, 'learning_rate/qwen_vl_interface': 8.686398866247121e-07, 'learning_rate/action_model': 8.68639886624712e-06, 'epoch': 2.52}) +[2026-06-18 02:08:58,408][starVLA.training.train_starvla][INFO] - Step 6130, Loss: {'train/loss': 0.26323753595352173, 'train/grad_norm_pre_clip': 21.910579681396484, 'timing/data': 0.0007245615124702454, 'timing/model': 1.1485768277198076, 'learning_rate/qwen_vl_interface': 8.6780498847166e-07, 'learning_rate/action_model': 8.678049884716599e-06, 'epoch': 2.52}) +[2026-06-18 02:09:02,176][starVLA.training.train_starvla][INFO] - Step 6131, Loss: {'train/loss': 0.20464850962162018, 'train/grad_norm_pre_clip': 13.32289981842041, 'timing/data': 0.0007368270307779312, 'timing/model': 0.9357424331828952, 'learning_rate/qwen_vl_interface': 8.669709987523479e-07, 'learning_rate/action_model': 8.669709987523477e-06, 'epoch': 2.52}) +[2026-06-18 02:09:06,072][starVLA.training.train_starvla][INFO] - Step 6132, Loss: {'train/loss': 0.3087352514266968, 'train/grad_norm_pre_clip': 19.063058853149414, 'timing/data': 0.0007526008412241936, 'timing/model': 0.9312688782811165, 'learning_rate/qwen_vl_interface': 8.661379176396651e-07, 'learning_rate/action_model': 8.66137917639665e-06, 'epoch': 2.52}) +[2026-06-18 02:09:09,932][starVLA.training.train_starvla][INFO] - Step 6133, Loss: {'train/loss': 0.1584155410528183, 'train/grad_norm_pre_clip': 29.201255798339844, 'timing/data': 0.0005305111408233643, 'timing/model': 0.9370395746082067, 'learning_rate/qwen_vl_interface': 8.653057453063092e-07, 'learning_rate/action_model': 8.653057453063091e-06, 'epoch': 2.52}) +[2026-06-18 02:09:13,885][starVLA.training.train_starvla][INFO] - Step 6134, Loss: {'train/loss': 0.4074380695819855, 'train/grad_norm_pre_clip': 32.89026641845703, 'timing/data': 0.0005391985177993774, 'timing/model': 0.9868789929896593, 'learning_rate/qwen_vl_interface': 8.644744819247908e-07, 'learning_rate/action_model': 8.644744819247908e-06, 'epoch': 2.52}) +[2026-06-18 02:09:17,725][starVLA.training.train_starvla][INFO] - Step 6135, Loss: {'train/loss': 0.10073649138212204, 'train/grad_norm_pre_clip': 13.1987886428833, 'timing/data': 0.0005362685769796371, 'timing/model': 0.9366986909881234, 'learning_rate/qwen_vl_interface': 8.636441276674308e-07, 'learning_rate/action_model': 8.636441276674308e-06, 'epoch': 2.53}) +[2026-06-18 02:09:21,670][starVLA.training.train_starvla][INFO] - Step 6136, Loss: {'train/loss': 0.36370590329170227, 'train/grad_norm_pre_clip': 31.764894485473633, 'timing/data': 0.0005145352333784103, 'timing/model': 0.9559316923841834, 'learning_rate/qwen_vl_interface': 8.628146827063632e-07, 'learning_rate/action_model': 8.628146827063633e-06, 'epoch': 2.53}) +[2026-06-18 02:09:25,487][starVLA.training.train_starvla][INFO] - Step 6137, Loss: {'train/loss': 0.3428371548652649, 'train/grad_norm_pre_clip': 30.637563705444336, 'timing/data': 0.0007489286363124847, 'timing/model': 1.1266530798748136, 'learning_rate/qwen_vl_interface': 8.619861472135332e-07, 'learning_rate/action_model': 8.619861472135332e-06, 'epoch': 2.53}) +[2026-06-18 02:09:29,132][starVLA.training.train_starvla][INFO] - Step 6138, Loss: {'train/loss': 0.2937967777252197, 'train/grad_norm_pre_clip': 20.665910720825195, 'timing/data': 0.000558946281671524, 'timing/model': 0.9288302380591631, 'learning_rate/qwen_vl_interface': 8.611585213606964e-07, 'learning_rate/action_model': 8.611585213606964e-06, 'epoch': 2.53}) +[2026-06-18 02:09:33,196][starVLA.training.train_starvla][INFO] - Step 6139, Loss: {'train/loss': 0.14249517023563385, 'train/grad_norm_pre_clip': 13.998712539672852, 'timing/data': 0.0010681608691811562, 'timing/model': 1.0319358222186565, 'learning_rate/qwen_vl_interface': 8.603318053194206e-07, 'learning_rate/action_model': 8.603318053194205e-06, 'epoch': 2.53}) +[2026-06-18 02:09:37,070][starVLA.training.train_starvla][INFO] - Step 6140, Loss: {'train/loss': 0.06907356530427933, 'train/grad_norm_pre_clip': 17.085344314575195, 'timing/data': 0.0009580086916685104, 'timing/model': 0.9379823431372643, 'learning_rate/qwen_vl_interface': 8.595059992610846e-07, 'learning_rate/action_model': 8.595059992610846e-06, 'epoch': 2.53}) +[2026-06-18 02:09:40,877][starVLA.training.train_starvla][INFO] - Step 6141, Loss: {'train/loss': 0.39283546805381775, 'train/grad_norm_pre_clip': 33.957828521728516, 'timing/data': 0.0005208738148212433, 'timing/model': 0.9292188119143248, 'learning_rate/qwen_vl_interface': 8.586811033568796e-07, 'learning_rate/action_model': 8.586811033568797e-06, 'epoch': 2.53}) +[2026-06-18 02:09:44,663][starVLA.training.train_starvla][INFO] - Step 6142, Loss: {'train/loss': 0.33042436838150024, 'train/grad_norm_pre_clip': 27.305484771728516, 'timing/data': 0.0011643823236227036, 'timing/model': 0.9235608223825693, 'learning_rate/qwen_vl_interface': 8.578571177778076e-07, 'learning_rate/action_model': 8.578571177778075e-06, 'epoch': 2.53}) +[2026-06-18 02:09:48,449][starVLA.training.train_starvla][INFO] - Step 6143, Loss: {'train/loss': 0.08612283319234848, 'train/grad_norm_pre_clip': 9.201446533203125, 'timing/data': 0.0005455370992422104, 'timing/model': 0.9350770683959126, 'learning_rate/qwen_vl_interface': 8.570340426946805e-07, 'learning_rate/action_model': 8.570340426946804e-06, 'epoch': 2.53}) +[2026-06-18 02:09:52,264][starVLA.training.train_starvla][INFO] - Step 6144, Loss: {'train/loss': 0.24213774502277374, 'train/grad_norm_pre_clip': 28.261611938476562, 'timing/data': 0.0006936406716704369, 'timing/model': 1.1484422581270337, 'learning_rate/qwen_vl_interface': 8.562118782781241e-07, 'learning_rate/action_model': 8.56211878278124e-06, 'epoch': 2.53}) +[2026-06-18 02:09:55,882][starVLA.training.train_starvla][INFO] - Step 6145, Loss: {'train/loss': 0.17903073132038116, 'train/grad_norm_pre_clip': 15.892131805419922, 'timing/data': 0.0005139047279953957, 'timing/model': 0.9306662315502763, 'learning_rate/qwen_vl_interface': 8.553906246985722e-07, 'learning_rate/action_model': 8.553906246985721e-06, 'epoch': 2.53}) +[2026-06-18 02:09:59,698][starVLA.training.train_starvla][INFO] - Step 6146, Loss: {'train/loss': 0.1409406065940857, 'train/grad_norm_pre_clip': 12.810019493103027, 'timing/data': 0.0009988434612751007, 'timing/model': 0.934815289452672, 'learning_rate/qwen_vl_interface': 8.545702821262733e-07, 'learning_rate/action_model': 8.545702821262733e-06, 'epoch': 2.53}) +[2026-06-18 02:10:03,580][starVLA.training.train_starvla][INFO] - Step 6147, Loss: {'train/loss': 0.20012527704238892, 'train/grad_norm_pre_clip': 16.15472984313965, 'timing/data': 0.000775504857301712, 'timing/model': 0.9475583164021373, 'learning_rate/qwen_vl_interface': 8.537508507312844e-07, 'learning_rate/action_model': 8.537508507312844e-06, 'epoch': 2.53}) +[2026-06-18 02:10:07,366][starVLA.training.train_starvla][INFO] - Step 6148, Loss: {'train/loss': 0.06302755326032639, 'train/grad_norm_pre_clip': 9.078158378601074, 'timing/data': 0.0006119208410382271, 'timing/model': 0.9297082368284464, 'learning_rate/qwen_vl_interface': 8.52932330683476e-07, 'learning_rate/action_model': 8.529323306834759e-06, 'epoch': 2.53}) +[2026-06-18 02:10:11,186][starVLA.training.train_starvla][INFO] - Step 6149, Loss: {'train/loss': 0.04301350191235542, 'train/grad_norm_pre_clip': 5.646575450897217, 'timing/data': 0.0008795801550149918, 'timing/model': 0.931969235651195, 'learning_rate/qwen_vl_interface': 8.52114722152525e-07, 'learning_rate/action_model': 8.521147221525249e-06, 'epoch': 2.53}) +[2026-06-18 02:10:15,029][starVLA.training.train_starvla][INFO] - Step 6150, Loss: {'train/loss': 0.19151122868061066, 'train/grad_norm_pre_clip': 12.973834037780762, 'timing/data': 0.0007910002022981644, 'timing/model': 0.9520261669531465, 'learning_rate/qwen_vl_interface': 8.51298025307925e-07, 'learning_rate/action_model': 8.51298025307925e-06, 'epoch': 2.53}) +[2026-06-18 02:10:18,970][starVLA.training.train_starvla][INFO] - Step 6151, Loss: {'train/loss': 0.2482275664806366, 'train/grad_norm_pre_clip': 16.48836898803711, 'timing/data': 0.0009229807183146477, 'timing/model': 1.1880305726081133, 'learning_rate/qwen_vl_interface': 8.504822403189783e-07, 'learning_rate/action_model': 8.504822403189782e-06, 'epoch': 2.53}) +[2026-06-18 02:10:22,558][starVLA.training.train_starvla][INFO] - Step 6152, Loss: {'train/loss': 0.13818123936653137, 'train/grad_norm_pre_clip': 18.981924057006836, 'timing/data': 0.0005252351984381676, 'timing/model': 0.9284712970256805, 'learning_rate/qwen_vl_interface': 8.496673673547967e-07, 'learning_rate/action_model': 8.496673673547968e-06, 'epoch': 2.53}) +[2026-06-18 02:10:26,582][starVLA.training.train_starvla][INFO] - Step 6153, Loss: {'train/loss': 0.6397998332977295, 'train/grad_norm_pre_clip': 29.88199806213379, 'timing/data': 0.0009046588093042374, 'timing/model': 0.9346731789410114, 'learning_rate/qwen_vl_interface': 8.488534065843046e-07, 'learning_rate/action_model': 8.488534065843045e-06, 'epoch': 2.53}) +[2026-06-18 02:10:30,350][starVLA.training.train_starvla][INFO] - Step 6154, Loss: {'train/loss': 0.149345263838768, 'train/grad_norm_pre_clip': 19.293188095092773, 'timing/data': 0.0011660708114504814, 'timing/model': 0.9239437310025096, 'learning_rate/qwen_vl_interface': 8.480403581762378e-07, 'learning_rate/action_model': 8.480403581762378e-06, 'epoch': 2.53}) +[2026-06-18 02:10:34,117][starVLA.training.train_starvla][INFO] - Step 6155, Loss: {'train/loss': 0.13609831035137177, 'train/grad_norm_pre_clip': 14.438968658447266, 'timing/data': 0.0005232710391283035, 'timing/model': 0.9329582806676626, 'learning_rate/qwen_vl_interface': 8.472282222991411e-07, 'learning_rate/action_model': 8.47228222299141e-06, 'epoch': 2.53}) +[2026-06-18 02:10:38,001][starVLA.training.train_starvla][INFO] - Step 6156, Loss: {'train/loss': 0.05062771961092949, 'train/grad_norm_pre_clip': 6.980656147003174, 'timing/data': 0.0005883229896426201, 'timing/model': 0.9924495499581099, 'learning_rate/qwen_vl_interface': 8.464169991213705e-07, 'learning_rate/action_model': 8.464169991213706e-06, 'epoch': 2.53}) +[2026-06-18 02:10:41,828][starVLA.training.train_starvla][INFO] - Step 6157, Loss: {'train/loss': 0.09725387394428253, 'train/grad_norm_pre_clip': 9.92677116394043, 'timing/data': 0.0005752155557274818, 'timing/model': 0.9304103162139654, 'learning_rate/qwen_vl_interface': 8.456066888110964e-07, 'learning_rate/action_model': 8.456066888110964e-06, 'epoch': 2.53}) +[2026-06-18 02:10:45,858][starVLA.training.train_starvla][INFO] - Step 6158, Loss: {'train/loss': 0.2905484735965729, 'train/grad_norm_pre_clip': 22.77989959716797, 'timing/data': 0.0008149202913045883, 'timing/model': 1.3559589628130198, 'learning_rate/qwen_vl_interface': 8.447972915362941e-07, 'learning_rate/action_model': 8.44797291536294e-06, 'epoch': 2.53}) +[2026-06-18 02:10:49,506][starVLA.training.train_starvla][INFO] - Step 6159, Loss: {'train/loss': 0.32758206129074097, 'train/grad_norm_pre_clip': 22.562646865844727, 'timing/data': 0.00126645527780056, 'timing/model': 0.9372634561732411, 'learning_rate/qwen_vl_interface': 8.439888074647536e-07, 'learning_rate/action_model': 8.439888074647536e-06, 'epoch': 2.54}) +[2026-06-18 02:10:53,359][starVLA.training.train_starvla][INFO] - Step 6160, Loss: {'train/loss': 0.4079849421977997, 'train/grad_norm_pre_clip': 25.065107345581055, 'timing/data': 0.0007523652166128159, 'timing/model': 0.9334276868030429, 'learning_rate/qwen_vl_interface': 8.431812367640732e-07, 'learning_rate/action_model': 8.431812367640731e-06, 'epoch': 2.54}) +[2026-06-18 02:10:57,229][starVLA.training.train_starvla][INFO] - Step 6161, Loss: {'train/loss': 0.24476087093353271, 'train/grad_norm_pre_clip': 18.519912719726562, 'timing/data': 0.0007274048402905464, 'timing/model': 0.9321294268593192, 'learning_rate/qwen_vl_interface': 8.423745796016659e-07, 'learning_rate/action_model': 8.42374579601666e-06, 'epoch': 2.54}) +[2026-06-18 02:11:01,042][starVLA.training.train_starvla][INFO] - Step 6162, Loss: {'train/loss': 0.1487293392419815, 'train/grad_norm_pre_clip': 25.92266082763672, 'timing/data': 0.0005332697182893753, 'timing/model': 0.9291390180587769, 'learning_rate/qwen_vl_interface': 8.41568836144751e-07, 'learning_rate/action_model': 8.41568836144751e-06, 'epoch': 2.54}) +[2026-06-18 02:11:04,994][starVLA.training.train_starvla][INFO] - Step 6163, Loss: {'train/loss': 0.17741701006889343, 'train/grad_norm_pre_clip': 22.29172134399414, 'timing/data': 0.0010883389040827751, 'timing/model': 1.0534414472058415, 'learning_rate/qwen_vl_interface': 8.407640065603601e-07, 'learning_rate/action_model': 8.4076400656036e-06, 'epoch': 2.54}) +[2026-06-18 02:11:08,842][starVLA.training.train_starvla][INFO] - Step 6164, Loss: {'train/loss': 0.3052365481853485, 'train/grad_norm_pre_clip': 26.977840423583984, 'timing/data': 0.0006010178476572037, 'timing/model': 0.9316218914464116, 'learning_rate/qwen_vl_interface': 8.399600910153336e-07, 'learning_rate/action_model': 8.399600910153336e-06, 'epoch': 2.54}) +[2026-06-18 02:11:12,657][starVLA.training.train_starvla][INFO] - Step 6165, Loss: {'train/loss': 0.21215666830539703, 'train/grad_norm_pre_clip': 21.78032112121582, 'timing/data': 0.0006272373721003532, 'timing/model': 1.1396921565756202, 'learning_rate/qwen_vl_interface': 8.391570896763266e-07, 'learning_rate/action_model': 8.391570896763267e-06, 'epoch': 2.54}) +[2026-06-18 02:11:16,256][starVLA.training.train_starvla][INFO] - Step 6166, Loss: {'train/loss': 0.076134093105793, 'train/grad_norm_pre_clip': 14.229711532592773, 'timing/data': 0.0005343127995729446, 'timing/model': 0.9287383304908872, 'learning_rate/qwen_vl_interface': 8.38355002709802e-07, 'learning_rate/action_model': 8.38355002709802e-06, 'epoch': 2.54}) +[2026-06-18 02:11:20,137][starVLA.training.train_starvla][INFO] - Step 6167, Loss: {'train/loss': 0.2379637360572815, 'train/grad_norm_pre_clip': 19.55650520324707, 'timing/data': 0.0006511043757200241, 'timing/model': 0.9333221781998873, 'learning_rate/qwen_vl_interface': 8.375538302820314e-07, 'learning_rate/action_model': 8.375538302820315e-06, 'epoch': 2.54}) +[2026-06-18 02:11:23,986][starVLA.training.train_starvla][INFO] - Step 6168, Loss: {'train/loss': 0.12162329256534576, 'train/grad_norm_pre_clip': 11.855049133300781, 'timing/data': 0.0009070876985788345, 'timing/model': 0.9408691748976707, 'learning_rate/qwen_vl_interface': 8.367535725591003e-07, 'learning_rate/action_model': 8.367535725591003e-06, 'epoch': 2.54}) +[2026-06-18 02:11:27,855][starVLA.training.train_starvla][INFO] - Step 6169, Loss: {'train/loss': 0.30361026525497437, 'train/grad_norm_pre_clip': 25.272422790527344, 'timing/data': 0.0006232131272554398, 'timing/model': 0.9326295228675008, 'learning_rate/qwen_vl_interface': 8.359542297069024e-07, 'learning_rate/action_model': 8.359542297069024e-06, 'epoch': 2.54}) +[2026-06-18 02:11:31,782][starVLA.training.train_starvla][INFO] - Step 6170, Loss: {'train/loss': 0.3883855640888214, 'train/grad_norm_pre_clip': 23.547250747680664, 'timing/data': 0.0005179615691304207, 'timing/model': 0.9310516770929098, 'learning_rate/qwen_vl_interface': 8.351558018911423e-07, 'learning_rate/action_model': 8.351558018911423e-06, 'epoch': 2.54}) +[2026-06-18 02:11:35,591][starVLA.training.train_starvla][INFO] - Step 6171, Loss: {'train/loss': 0.21457551419734955, 'train/grad_norm_pre_clip': 30.264965057373047, 'timing/data': 0.0005173021927475929, 'timing/model': 0.9280336890369654, 'learning_rate/qwen_vl_interface': 8.343582892773352e-07, 'learning_rate/action_model': 8.343582892773351e-06, 'epoch': 2.54}) +[2026-06-18 02:11:39,416][starVLA.training.train_starvla][INFO] - Step 6172, Loss: {'train/loss': 0.26267534494400024, 'train/grad_norm_pre_clip': 39.103492736816406, 'timing/data': 0.0006787832826375961, 'timing/model': 1.1356154484674335, 'learning_rate/qwen_vl_interface': 8.335616920308061e-07, 'learning_rate/action_model': 8.33561692030806e-06, 'epoch': 2.54}) +[2026-06-18 02:11:43,019][starVLA.training.train_starvla][INFO] - Step 6173, Loss: {'train/loss': 0.1288159191608429, 'train/grad_norm_pre_clip': 12.917820930480957, 'timing/data': 0.0009756889194250107, 'timing/model': 0.9328371109440923, 'learning_rate/qwen_vl_interface': 8.327660103166907e-07, 'learning_rate/action_model': 8.327660103166906e-06, 'epoch': 2.54}) +[2026-06-18 02:11:46,800][starVLA.training.train_starvla][INFO] - Step 6174, Loss: {'train/loss': 0.23365943133831024, 'train/grad_norm_pre_clip': 17.16411018371582, 'timing/data': 0.0006655948236584663, 'timing/model': 0.9290293110534549, 'learning_rate/qwen_vl_interface': 8.319712442999335e-07, 'learning_rate/action_model': 8.319712442999335e-06, 'epoch': 2.54}) +[2026-06-18 02:11:50,594][starVLA.training.train_starvla][INFO] - Step 6175, Loss: {'train/loss': 0.4482404887676239, 'train/grad_norm_pre_clip': 31.23241424560547, 'timing/data': 0.0007238304242491722, 'timing/model': 0.9280767347663641, 'learning_rate/qwen_vl_interface': 8.311773941452924e-07, 'learning_rate/action_model': 8.311773941452925e-06, 'epoch': 2.54}) +[2026-06-18 02:11:54,428][starVLA.training.train_starvla][INFO] - Step 6176, Loss: {'train/loss': 0.2685105800628662, 'train/grad_norm_pre_clip': 28.380287170410156, 'timing/data': 0.0006839996203780174, 'timing/model': 0.928305690176785, 'learning_rate/qwen_vl_interface': 8.303844600173326e-07, 'learning_rate/action_model': 8.303844600173325e-06, 'epoch': 2.54}) +[2026-06-18 02:11:58,238][starVLA.training.train_starvla][INFO] - Step 6177, Loss: {'train/loss': 0.23022401332855225, 'train/grad_norm_pre_clip': 18.482276916503906, 'timing/data': 0.0005269208922982216, 'timing/model': 0.9273308869451284, 'learning_rate/qwen_vl_interface': 8.295924420804297e-07, 'learning_rate/action_model': 8.295924420804296e-06, 'epoch': 2.54}) +[2026-06-18 02:12:02,036][starVLA.training.train_starvla][INFO] - Step 6178, Loss: {'train/loss': 0.21474984288215637, 'train/grad_norm_pre_clip': 22.30497169494629, 'timing/data': 0.0006250878795981407, 'timing/model': 0.9357411991804838, 'learning_rate/qwen_vl_interface': 8.288013404987691e-07, 'learning_rate/action_model': 8.28801340498769e-06, 'epoch': 2.54}) +[2026-06-18 02:12:05,808][starVLA.training.train_starvla][INFO] - Step 6179, Loss: {'train/loss': 0.6718132495880127, 'train/grad_norm_pre_clip': 27.347272872924805, 'timing/data': 0.0007035760208964348, 'timing/model': 1.1058936333283782, 'learning_rate/qwen_vl_interface': 8.28011155436349e-07, 'learning_rate/action_model': 8.28011155436349e-06, 'epoch': 2.54}) +[2026-06-18 02:12:09,597][starVLA.training.train_starvla][INFO] - Step 6180, Loss: {'train/loss': 0.13605788350105286, 'train/grad_norm_pre_clip': 23.438007354736328, 'timing/data': 0.0005611954256892204, 'timing/model': 0.9440869120880961, 'learning_rate/qwen_vl_interface': 8.272218870569748e-07, 'learning_rate/action_model': 8.272218870569748e-06, 'epoch': 2.54}) +[2026-06-18 02:12:13,490][starVLA.training.train_starvla][INFO] - Step 6181, Loss: {'train/loss': 0.31151026487350464, 'train/grad_norm_pre_clip': 22.226177215576172, 'timing/data': 0.0008502397686243057, 'timing/model': 0.9930803934112191, 'learning_rate/qwen_vl_interface': 8.264335355242625e-07, 'learning_rate/action_model': 8.264335355242625e-06, 'epoch': 2.54}) +[2026-06-18 02:12:17,520][starVLA.training.train_starvla][INFO] - Step 6182, Loss: {'train/loss': 0.19536198675632477, 'train/grad_norm_pre_clip': 15.576650619506836, 'timing/data': 0.000545731745660305, 'timing/model': 0.9332442367449403, 'learning_rate/qwen_vl_interface': 8.256461010016376e-07, 'learning_rate/action_model': 8.256461010016376e-06, 'epoch': 2.54}) +[2026-06-18 02:12:21,494][starVLA.training.train_starvla][INFO] - Step 6183, Loss: {'train/loss': 0.19902221858501434, 'train/grad_norm_pre_clip': 17.45587921142578, 'timing/data': 0.0005247220396995544, 'timing/model': 0.9365704245865345, 'learning_rate/qwen_vl_interface': 8.248595836523368e-07, 'learning_rate/action_model': 8.248595836523368e-06, 'epoch': 2.55}) +[2026-06-18 02:12:25,328][starVLA.training.train_starvla][INFO] - Step 6184, Loss: {'train/loss': 0.08797378093004227, 'train/grad_norm_pre_clip': 22.802928924560547, 'timing/data': 0.0005586938932538033, 'timing/model': 0.9495106898248196, 'learning_rate/qwen_vl_interface': 8.240739836394068e-07, 'learning_rate/action_model': 8.240739836394068e-06, 'epoch': 2.55}) +[2026-06-18 02:12:29,263][starVLA.training.train_starvla][INFO] - Step 6185, Loss: {'train/loss': 0.26778867840766907, 'train/grad_norm_pre_clip': 22.780324935913086, 'timing/data': 0.0005936361849308014, 'timing/model': 0.944076239131391, 'learning_rate/qwen_vl_interface': 8.232893011257019e-07, 'learning_rate/action_model': 8.232893011257018e-06, 'epoch': 2.55}) +[2026-06-18 02:12:33,141][starVLA.training.train_starvla][INFO] - Step 6186, Loss: {'train/loss': 0.05046606436371803, 'train/grad_norm_pre_clip': 7.615905284881592, 'timing/data': 0.0005648629739880562, 'timing/model': 1.1434483602643013, 'learning_rate/qwen_vl_interface': 8.225055362738886e-07, 'learning_rate/action_model': 8.225055362738885e-06, 'epoch': 2.55}) +[2026-06-18 02:12:37,237][starVLA.training.train_starvla][INFO] - Step 6187, Loss: {'train/loss': 0.3033629059791565, 'train/grad_norm_pre_clip': 62.194156646728516, 'timing/data': 0.0010241121053695679, 'timing/model': 0.9494206206873059, 'learning_rate/qwen_vl_interface': 8.217226892464411e-07, 'learning_rate/action_model': 8.21722689246441e-06, 'epoch': 2.55}) +[2026-06-18 02:12:41,087][starVLA.training.train_starvla][INFO] - Step 6188, Loss: {'train/loss': 0.09846008569002151, 'train/grad_norm_pre_clip': 15.828210830688477, 'timing/data': 0.0007313508540391922, 'timing/model': 0.9448578394949436, 'learning_rate/qwen_vl_interface': 8.209407602056457e-07, 'learning_rate/action_model': 8.209407602056457e-06, 'epoch': 2.55}) +[2026-06-18 02:12:45,101][starVLA.training.train_starvla][INFO] - Step 6189, Loss: {'train/loss': 0.24767638742923737, 'train/grad_norm_pre_clip': 35.204925537109375, 'timing/data': 0.0005691805854439735, 'timing/model': 0.9540264504030347, 'learning_rate/qwen_vl_interface': 8.201597493135962e-07, 'learning_rate/action_model': 8.201597493135961e-06, 'epoch': 2.55}) +[2026-06-18 02:12:49,769][starVLA.training.train_starvla][INFO] - Step 6190, Loss: {'train/loss': 0.19626155495643616, 'train/grad_norm_pre_clip': 14.396011352539062, 'timing/data': 0.0005510663613677025, 'timing/model': 0.9389254786074162, 'learning_rate/qwen_vl_interface': 8.193796567321982e-07, 'learning_rate/action_model': 8.193796567321982e-06, 'epoch': 2.55}) +[2026-06-18 02:12:53,616][starVLA.training.train_starvla][INFO] - Step 6191, Loss: {'train/loss': 0.282890260219574, 'train/grad_norm_pre_clip': 26.69935417175293, 'timing/data': 0.0009102588519454002, 'timing/model': 0.9418716309592128, 'learning_rate/qwen_vl_interface': 8.186004826231644e-07, 'learning_rate/action_model': 8.186004826231644e-06, 'epoch': 2.55}) +[2026-06-18 02:12:57,530][starVLA.training.train_starvla][INFO] - Step 6192, Loss: {'train/loss': 0.336885541677475, 'train/grad_norm_pre_clip': 22.160266876220703, 'timing/data': 0.0006092945113778114, 'timing/model': 0.9527357947081327, 'learning_rate/qwen_vl_interface': 8.178222271480192e-07, 'learning_rate/action_model': 8.178222271480191e-06, 'epoch': 2.55}) +[2026-06-18 02:13:01,536][starVLA.training.train_starvla][INFO] - Step 6193, Loss: {'train/loss': 0.145350843667984, 'train/grad_norm_pre_clip': 15.618144035339355, 'timing/data': 0.00115900207310915, 'timing/model': 1.1337410937994719, 'learning_rate/qwen_vl_interface': 8.170448904680943e-07, 'learning_rate/action_model': 8.170448904680943e-06, 'epoch': 2.55}) +[2026-06-18 02:13:05,427][starVLA.training.train_starvla][INFO] - Step 6194, Loss: {'train/loss': 0.14844807982444763, 'train/grad_norm_pre_clip': 16.607210159301758, 'timing/data': 0.0009052390232682228, 'timing/model': 0.9387470660731196, 'learning_rate/qwen_vl_interface': 8.162684727445347e-07, 'learning_rate/action_model': 8.162684727445346e-06, 'epoch': 2.55}) +[2026-06-18 02:13:09,395][starVLA.training.train_starvla][INFO] - Step 6195, Loss: {'train/loss': 0.19422310590744019, 'train/grad_norm_pre_clip': 23.095748901367188, 'timing/data': 0.0005666045472025871, 'timing/model': 0.9387797741219401, 'learning_rate/qwen_vl_interface': 8.154929741382922e-07, 'learning_rate/action_model': 8.154929741382922e-06, 'epoch': 2.55}) +[2026-06-18 02:13:13,243][starVLA.training.train_starvla][INFO] - Step 6196, Loss: {'train/loss': 0.03758365288376808, 'train/grad_norm_pre_clip': 5.860504150390625, 'timing/data': 0.0009056441485881805, 'timing/model': 0.9331469470635056, 'learning_rate/qwen_vl_interface': 8.147183948101266e-07, 'learning_rate/action_model': 8.147183948101265e-06, 'epoch': 2.55}) +[2026-06-18 02:13:17,085][starVLA.training.train_starvla][INFO] - Step 6197, Loss: {'train/loss': 0.17148683965206146, 'train/grad_norm_pre_clip': 16.195390701293945, 'timing/data': 0.0004999246448278427, 'timing/model': 0.9546059425920248, 'learning_rate/qwen_vl_interface': 8.139447349206111e-07, 'learning_rate/action_model': 8.139447349206111e-06, 'epoch': 2.55}) +[2026-06-18 02:13:20,941][starVLA.training.train_starvla][INFO] - Step 6198, Loss: {'train/loss': 0.08082683384418488, 'train/grad_norm_pre_clip': 9.217607498168945, 'timing/data': 0.0007301019504666328, 'timing/model': 0.9358111303299665, 'learning_rate/qwen_vl_interface': 8.131719946301255e-07, 'learning_rate/action_model': 8.131719946301255e-06, 'epoch': 2.55}) +[2026-06-18 02:13:24,797][starVLA.training.train_starvla][INFO] - Step 6199, Loss: {'train/loss': 0.2336788773536682, 'train/grad_norm_pre_clip': 22.710704803466797, 'timing/data': 0.0004526190459728241, 'timing/model': 0.9705724334344268, 'learning_rate/qwen_vl_interface': 8.124001740988595e-07, 'learning_rate/action_model': 8.124001740988595e-06, 'epoch': 2.55}) +[2026-06-18 02:15:34,983][starVLA.training.train_starvla][INFO] - Step 6200, Loss: {'train/loss': 0.10741084069013596, 'train/grad_norm_pre_clip': 13.737504959106445, 'eval/loss': 0.42000120878219604, 'eval/latency_6/loss': 0.42000120878219604, 'eval/demon_fire_precision': 0.9538171536286523, 'eval/demon_fire_recall': 0.9812540400775694, 'eval/demon_fire_f1': 0.9673410865062928, 'eval/demon_fire_rate_model': 0.8856427378964942, 'eval/demon_fire_rate_teacher': 0.860879243183083, 'eval/demon_left_precision': 0.897489539748954, 'eval/demon_left_recall': 0.7526315789473684, 'eval/demon_left_f1': 0.8187022900763358, 'eval/demon_left_rate_model': 0.13299944351697274, 'eval/demon_left_rate_teacher': 0.15859766277128548, 'eval/demon_right_precision': 0.8357963875205254, 'eval/demon_right_recall': 0.7818740399385561, 'eval/demon_right_f1': 0.8079365079365081, 'eval/demon_right_rate_model': 0.169449081803005, 'eval/demon_right_rate_teacher': 0.18113522537562604, 'eval/demon_move_f1': 0.8133193990064219, 'eval/demon_cc_f1': 0.9153386376142321, 'eval/latency_6/demon_fire_precision': 0.9538171536286523, 'eval/latency_6/demon_fire_recall': 0.9812540400775694, 'eval/latency_6/demon_fire_f1': 0.9673410865062928, 'eval/latency_6/demon_fire_rate_model': 0.8856427378964942, 'eval/latency_6/demon_fire_rate_teacher': 0.860879243183083, 'eval/latency_6/demon_left_precision': 0.897489539748954, 'eval/latency_6/demon_left_recall': 0.7526315789473684, 'eval/latency_6/demon_left_f1': 0.8187022900763358, 'eval/latency_6/demon_left_rate_model': 0.13299944351697274, 'eval/latency_6/demon_left_rate_teacher': 0.15859766277128548, 'eval/latency_6/demon_right_precision': 0.8357963875205254, 'eval/latency_6/demon_right_recall': 0.7818740399385561, 'eval/latency_6/demon_right_f1': 0.8079365079365081, 'eval/latency_6/demon_right_rate_model': 0.169449081803005, 'eval/latency_6/demon_right_rate_teacher': 0.18113522537562604, 'eval/latency_6/demon_move_f1': 0.8133193990064219, 'eval/latency_6/demon_cc_f1': 0.9153386376142321, 'timing/data': 0.0005479278042912483, 'timing/model': 1.1253377059474587, 'learning_rate/qwen_vl_interface': 8.116292734868126e-07, 'learning_rate/action_model': 8.116292734868125e-06, 'epoch': 2.55}) +[2026-06-18 02:15:35,015][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_6200_state +[2026-06-18 02:15:35,019][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 02:15:41,966][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_6200_state/model.safetensors +[2026-06-18 02:15:57,388][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_6200_state/optimizer.bin +[2026-06-18 02:15:57,390][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_6200_state/sampler.bin +[2026-06-18 02:15:57,390][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_6200_state/sampler_1.bin +[2026-06-18 02:15:57,392][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_6200_state/random_states_0.pkl +[2026-06-18 02:15:57,396][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 6200 +[2026-06-18 02:15:57,396][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 02:15:57,411][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 02:16:01,216][starVLA.training.train_starvla][INFO] - Step 6201, Loss: {'train/loss': 0.17797894775867462, 'train/grad_norm_pre_clip': 28.07616424560547, 'timing/data': 0.000699988566339016, 'timing/model': 0.9457474509254098, 'learning_rate/qwen_vl_interface': 8.108592929537928e-07, 'learning_rate/action_model': 8.108592929537928e-06, 'epoch': 2.55}) +[2026-06-18 02:16:05,006][starVLA.training.train_starvla][INFO] - Step 6202, Loss: {'train/loss': 0.19911275804042816, 'train/grad_norm_pre_clip': 17.549036026000977, 'timing/data': 0.0004995204508304596, 'timing/model': 1.1235422054305673, 'learning_rate/qwen_vl_interface': 8.100902326594186e-07, 'learning_rate/action_model': 8.100902326594186e-06, 'epoch': 2.55}) +[2026-06-18 02:16:08,622][starVLA.training.train_starvla][INFO] - Step 6203, Loss: {'train/loss': 0.17109674215316772, 'train/grad_norm_pre_clip': 14.978654861450195, 'timing/data': 0.0005413927137851715, 'timing/model': 0.9374674838036299, 'learning_rate/qwen_vl_interface': 8.093220927631171e-07, 'learning_rate/action_model': 8.09322092763117e-06, 'epoch': 2.55}) +[2026-06-18 02:16:12,403][starVLA.training.train_starvla][INFO] - Step 6204, Loss: {'train/loss': 0.14468581974506378, 'train/grad_norm_pre_clip': 15.94179630279541, 'timing/data': 0.0005027847364544868, 'timing/model': 0.9354895688593388, 'learning_rate/qwen_vl_interface': 8.085548734241233e-07, 'learning_rate/action_model': 8.085548734241232e-06, 'epoch': 2.55}) +[2026-06-18 02:16:16,126][starVLA.training.train_starvla][INFO] - Step 6205, Loss: {'train/loss': 0.35881176590919495, 'train/grad_norm_pre_clip': 24.497194290161133, 'timing/data': 0.0009006224572658539, 'timing/model': 0.9316531280055642, 'learning_rate/qwen_vl_interface': 8.077885748014843e-07, 'learning_rate/action_model': 8.077885748014841e-06, 'epoch': 2.55}) +[2026-06-18 02:16:19,915][starVLA.training.train_starvla][INFO] - Step 6206, Loss: {'train/loss': 0.48530372977256775, 'train/grad_norm_pre_clip': 34.20496368408203, 'timing/data': 0.00046173762530088425, 'timing/model': 0.9330397350713611, 'learning_rate/qwen_vl_interface': 8.070231970540532e-07, 'learning_rate/action_model': 8.070231970540532e-06, 'epoch': 2.55}) +[2026-06-18 02:16:23,818][starVLA.training.train_starvla][INFO] - Step 6207, Loss: {'train/loss': 0.40438976883888245, 'train/grad_norm_pre_clip': 38.93421936035156, 'timing/data': 0.0008466979488730431, 'timing/model': 1.0512956539168954, 'learning_rate/qwen_vl_interface': 8.062587403404934e-07, 'learning_rate/action_model': 8.062587403404934e-06, 'epoch': 2.56}) +[2026-06-18 02:16:27,715][starVLA.training.train_starvla][INFO] - Step 6208, Loss: {'train/loss': 0.17175380885601044, 'train/grad_norm_pre_clip': 19.23310089111328, 'timing/data': 0.0008141091093420982, 'timing/model': 1.1289625857025385, 'learning_rate/qwen_vl_interface': 8.054952048192793e-07, 'learning_rate/action_model': 8.054952048192794e-06, 'epoch': 2.56}) +[2026-06-18 02:16:31,307][starVLA.training.train_starvla][INFO] - Step 6209, Loss: {'train/loss': 0.18960264325141907, 'train/grad_norm_pre_clip': 16.668535232543945, 'timing/data': 0.0004951506853103638, 'timing/model': 0.9351514428853989, 'learning_rate/qwen_vl_interface': 8.047325906486926e-07, 'learning_rate/action_model': 8.047325906486925e-06, 'epoch': 2.56}) +[2026-06-18 02:16:35,197][starVLA.training.train_starvla][INFO] - Step 6210, Loss: {'train/loss': 0.2392900437116623, 'train/grad_norm_pre_clip': 18.48844337463379, 'timing/data': 0.000483841635286808, 'timing/model': 1.0175767187029123, 'learning_rate/qwen_vl_interface': 8.039708979868229e-07, 'learning_rate/action_model': 8.039708979868227e-06, 'epoch': 2.56}) +[2026-06-18 02:16:39,030][starVLA.training.train_starvla][INFO] - Step 6211, Loss: {'train/loss': 0.25779274106025696, 'train/grad_norm_pre_clip': 30.67722511291504, 'timing/data': 0.000456174835562706, 'timing/model': 0.9458705317229033, 'learning_rate/qwen_vl_interface': 8.032101269915687e-07, 'learning_rate/action_model': 8.032101269915687e-06, 'epoch': 2.56}) +[2026-06-18 02:16:43,067][starVLA.training.train_starvla][INFO] - Step 6212, Loss: {'train/loss': 0.19511252641677856, 'train/grad_norm_pre_clip': 16.4842472076416, 'timing/data': 0.0005049044266343117, 'timing/model': 0.9392188536003232, 'learning_rate/qwen_vl_interface': 8.024502778206411e-07, 'learning_rate/action_model': 8.02450277820641e-06, 'epoch': 2.56}) +[2026-06-18 02:16:46,931][starVLA.training.train_starvla][INFO] - Step 6213, Loss: {'train/loss': 0.216363787651062, 'train/grad_norm_pre_clip': 16.429067611694336, 'timing/data': 0.0010096384212374687, 'timing/model': 0.9343040008097887, 'learning_rate/qwen_vl_interface': 8.016913506315566e-07, 'learning_rate/action_model': 8.016913506315565e-06, 'epoch': 2.56}) +[2026-06-18 02:16:50,766][starVLA.training.train_starvla][INFO] - Step 6214, Loss: {'train/loss': 0.2552700638771057, 'train/grad_norm_pre_clip': 25.17131805419922, 'timing/data': 0.00044192373752593994, 'timing/model': 0.9957140805199742, 'learning_rate/qwen_vl_interface': 8.009333455816416e-07, 'learning_rate/action_model': 8.009333455816417e-06, 'epoch': 2.56}) +[2026-06-18 02:16:54,618][starVLA.training.train_starvla][INFO] - Step 6215, Loss: {'train/loss': 0.3136281371116638, 'train/grad_norm_pre_clip': 24.98201560974121, 'timing/data': 0.001086430624127388, 'timing/model': 1.1249058796092868, 'learning_rate/qwen_vl_interface': 8.001762628280318e-07, 'learning_rate/action_model': 8.001762628280319e-06, 'epoch': 2.56}) +[2026-06-18 02:16:58,327][starVLA.training.train_starvla][INFO] - Step 6216, Loss: {'train/loss': 0.1524585634469986, 'train/grad_norm_pre_clip': 14.754108428955078, 'timing/data': 0.0008270451799035072, 'timing/model': 0.9360964335501194, 'learning_rate/qwen_vl_interface': 7.994201025276708e-07, 'learning_rate/action_model': 7.994201025276708e-06, 'epoch': 2.56}) +[2026-06-18 02:17:02,234][starVLA.training.train_starvla][INFO] - Step 6217, Loss: {'train/loss': 0.14408189058303833, 'train/grad_norm_pre_clip': 14.268255233764648, 'timing/data': 0.0005442295223474503, 'timing/model': 0.9506133003160357, 'learning_rate/qwen_vl_interface': 7.986648648373115e-07, 'learning_rate/action_model': 7.986648648373115e-06, 'epoch': 2.56}) +[2026-06-18 02:17:06,158][starVLA.training.train_starvla][INFO] - Step 6218, Loss: {'train/loss': 0.2520436942577362, 'train/grad_norm_pre_clip': 63.4285888671875, 'timing/data': 0.0006981333717703819, 'timing/model': 0.9701116560027003, 'learning_rate/qwen_vl_interface': 7.979105499135157e-07, 'learning_rate/action_model': 7.979105499135155e-06, 'epoch': 2.56}) +[2026-06-18 02:17:10,039][starVLA.training.train_starvla][INFO] - Step 6219, Loss: {'train/loss': 0.14401771128177643, 'train/grad_norm_pre_clip': 16.83627700805664, 'timing/data': 0.0005015544593334198, 'timing/model': 0.9606451261788607, 'learning_rate/qwen_vl_interface': 7.971571579126533e-07, 'learning_rate/action_model': 7.971571579126534e-06, 'epoch': 2.56}) +[2026-06-18 02:17:13,906][starVLA.training.train_starvla][INFO] - Step 6220, Loss: {'train/loss': 0.17950941622257233, 'train/grad_norm_pre_clip': 15.960359573364258, 'timing/data': 0.0005164090543985367, 'timing/model': 0.9486222555860877, 'learning_rate/qwen_vl_interface': 7.964046889909038e-07, 'learning_rate/action_model': 7.964046889909038e-06, 'epoch': 2.56}) +[2026-06-18 02:17:17,724][starVLA.training.train_starvla][INFO] - Step 6221, Loss: {'train/loss': 0.21189281344413757, 'train/grad_norm_pre_clip': 21.965389251708984, 'timing/data': 0.0005237814038991928, 'timing/model': 0.9488181276246905, 'learning_rate/qwen_vl_interface': 7.956531433042543e-07, 'learning_rate/action_model': 7.956531433042542e-06, 'epoch': 2.56}) +[2026-06-18 02:17:21,532][starVLA.training.train_starvla][INFO] - Step 6222, Loss: {'train/loss': 0.33704692125320435, 'train/grad_norm_pre_clip': 36.31865692138672, 'timing/data': 0.0015262113884091377, 'timing/model': 1.1376758366823196, 'learning_rate/qwen_vl_interface': 7.94902521008501e-07, 'learning_rate/action_model': 7.94902521008501e-06, 'epoch': 2.56}) +[2026-06-18 02:17:25,263][starVLA.training.train_starvla][INFO] - Step 6223, Loss: {'train/loss': 0.06198917329311371, 'train/grad_norm_pre_clip': 9.970109939575195, 'timing/data': 0.0011607268825173378, 'timing/model': 0.9602951863780618, 'learning_rate/qwen_vl_interface': 7.941528222592501e-07, 'learning_rate/action_model': 7.941528222592501e-06, 'epoch': 2.56}) +[2026-06-18 02:17:29,163][starVLA.training.train_starvla][INFO] - Step 6224, Loss: {'train/loss': 0.28214097023010254, 'train/grad_norm_pre_clip': 22.525224685668945, 'timing/data': 0.0004770280793309212, 'timing/model': 0.9449671199545264, 'learning_rate/qwen_vl_interface': 7.934040472119132e-07, 'learning_rate/action_model': 7.934040472119133e-06, 'epoch': 2.56}) +[2026-06-18 02:17:33,065][starVLA.training.train_starvla][INFO] - Step 6225, Loss: {'train/loss': 0.10182145982980728, 'train/grad_norm_pre_clip': 17.408432006835938, 'timing/data': 0.0007164850831031799, 'timing/model': 0.9734168155118823, 'learning_rate/qwen_vl_interface': 7.926561960217127e-07, 'learning_rate/action_model': 7.926561960217126e-06, 'epoch': 2.56}) +[2026-06-18 02:17:37,022][starVLA.training.train_starvla][INFO] - Step 6226, Loss: {'train/loss': 0.2447291612625122, 'train/grad_norm_pre_clip': 16.75576400756836, 'timing/data': 0.0004920279607176781, 'timing/model': 0.9327820353209972, 'learning_rate/qwen_vl_interface': 7.919092688436783e-07, 'learning_rate/action_model': 7.919092688436783e-06, 'epoch': 2.56}) +[2026-06-18 02:17:40,862][starVLA.training.train_starvla][INFO] - Step 6227, Loss: {'train/loss': 0.22067084908485413, 'train/grad_norm_pre_clip': 25.72585105895996, 'timing/data': 0.0011123530566692352, 'timing/model': 0.9834087705239654, 'learning_rate/qwen_vl_interface': 7.911632658326505e-07, 'learning_rate/action_model': 7.911632658326505e-06, 'epoch': 2.56}) +[2026-06-18 02:17:44,693][starVLA.training.train_starvla][INFO] - Step 6228, Loss: {'train/loss': 0.03706016391515732, 'train/grad_norm_pre_clip': 9.686408996582031, 'timing/data': 0.0004529142752289772, 'timing/model': 0.942062265239656, 'learning_rate/qwen_vl_interface': 7.904181871432757e-07, 'learning_rate/action_model': 7.904181871432757e-06, 'epoch': 2.56}) +[2026-06-18 02:17:48,523][starVLA.training.train_starvla][INFO] - Step 6229, Loss: {'train/loss': 0.2550075352191925, 'train/grad_norm_pre_clip': 12.993687629699707, 'timing/data': 0.0008141724392771721, 'timing/model': 1.0958772590383887, 'learning_rate/qwen_vl_interface': 7.896740329300091e-07, 'learning_rate/action_model': 7.89674032930009e-06, 'epoch': 2.56}) +[2026-06-18 02:17:52,134][starVLA.training.train_starvla][INFO] - Step 6230, Loss: {'train/loss': 0.23680761456489563, 'train/grad_norm_pre_clip': 22.645374298095703, 'timing/data': 0.0010464489459991455, 'timing/model': 0.9397841645404696, 'learning_rate/qwen_vl_interface': 7.889308033471156e-07, 'learning_rate/action_model': 7.889308033471155e-06, 'epoch': 2.56}) +[2026-06-18 02:17:55,930][starVLA.training.train_starvla][INFO] - Step 6231, Loss: {'train/loss': 0.1904284507036209, 'train/grad_norm_pre_clip': 17.393238067626953, 'timing/data': 0.0007880404591560364, 'timing/model': 0.9314562967047095, 'learning_rate/qwen_vl_interface': 7.881884985486662e-07, 'learning_rate/action_model': 7.881884985486662e-06, 'epoch': 2.56}) +[2026-06-18 02:17:59,717][starVLA.training.train_starvla][INFO] - Step 6232, Loss: {'train/loss': 0.04517071321606636, 'train/grad_norm_pre_clip': 7.539131164550781, 'timing/data': 0.00046124961227178574, 'timing/model': 0.9449904039502144, 'learning_rate/qwen_vl_interface': 7.874471186885427e-07, 'learning_rate/action_model': 7.874471186885427e-06, 'epoch': 2.57}) +[2026-06-18 02:18:03,525][starVLA.training.train_starvla][INFO] - Step 6233, Loss: {'train/loss': 0.16311213374137878, 'train/grad_norm_pre_clip': 14.672579765319824, 'timing/data': 0.0006847726181149483, 'timing/model': 0.9468377269804478, 'learning_rate/qwen_vl_interface': 7.867066639204326e-07, 'learning_rate/action_model': 7.867066639204327e-06, 'epoch': 2.57}) +[2026-06-18 02:18:07,441][starVLA.training.train_starvla][INFO] - Step 6234, Loss: {'train/loss': 0.27132999897003174, 'train/grad_norm_pre_clip': 22.8431453704834, 'timing/data': 0.0004860358312726021, 'timing/model': 0.9585233209654689, 'learning_rate/qwen_vl_interface': 7.859671343978348e-07, 'learning_rate/action_model': 7.859671343978348e-06, 'epoch': 2.57}) +[2026-06-18 02:18:11,373][starVLA.training.train_starvla][INFO] - Step 6235, Loss: {'train/loss': 0.1995914876461029, 'train/grad_norm_pre_clip': 21.473957061767578, 'timing/data': 0.001185363158583641, 'timing/model': 0.9371536476537585, 'learning_rate/qwen_vl_interface': 7.852285302740528e-07, 'learning_rate/action_model': 7.852285302740527e-06, 'epoch': 2.57}) +[2026-06-18 02:18:15,336][starVLA.training.train_starvla][INFO] - Step 6236, Loss: {'train/loss': 0.22994428873062134, 'train/grad_norm_pre_clip': 26.818119049072266, 'timing/data': 0.00046209990978240967, 'timing/model': 1.130284946411848, 'learning_rate/qwen_vl_interface': 7.844908517022005e-07, 'learning_rate/action_model': 7.844908517022005e-06, 'epoch': 2.57}) +[2026-06-18 02:18:19,164][starVLA.training.train_starvla][INFO] - Step 6237, Loss: {'train/loss': 0.2775772511959076, 'train/grad_norm_pre_clip': 21.165868759155273, 'timing/data': 0.0007258616387844086, 'timing/model': 0.9311972800642252, 'learning_rate/qwen_vl_interface': 7.837540988352001e-07, 'learning_rate/action_model': 7.837540988352e-06, 'epoch': 2.57}) +[2026-06-18 02:18:24,277][starVLA.training.train_starvla][INFO] - Step 6238, Loss: {'train/loss': 0.09853019565343857, 'train/grad_norm_pre_clip': 9.82391357421875, 'timing/data': 0.0005444269627332687, 'timing/model': 0.9378203777596354, 'learning_rate/qwen_vl_interface': 7.83018271825782e-07, 'learning_rate/action_model': 7.83018271825782e-06, 'epoch': 2.57}) +[2026-06-18 02:18:28,129][starVLA.training.train_starvla][INFO] - Step 6239, Loss: {'train/loss': 0.18520089983940125, 'train/grad_norm_pre_clip': 21.356996536254883, 'timing/data': 0.0010527968406677246, 'timing/model': 0.9330401299521327, 'learning_rate/qwen_vl_interface': 7.822833708264821e-07, 'learning_rate/action_model': 7.822833708264821e-06, 'epoch': 2.57}) +[2026-06-18 02:18:31,929][starVLA.training.train_starvla][INFO] - Step 6240, Loss: {'train/loss': 0.1125488206744194, 'train/grad_norm_pre_clip': 10.842778205871582, 'timing/data': 0.0007011480629444122, 'timing/model': 0.9284875225275755, 'learning_rate/qwen_vl_interface': 7.81549395989646e-07, 'learning_rate/action_model': 7.81549395989646e-06, 'epoch': 2.57}) +[2026-06-18 02:18:35,725][starVLA.training.train_starvla][INFO] - Step 6241, Loss: {'train/loss': 0.09507118165493011, 'train/grad_norm_pre_clip': 11.753393173217773, 'timing/data': 0.00045034009963274, 'timing/model': 0.9541249088943005, 'learning_rate/qwen_vl_interface': 7.808163474674291e-07, 'learning_rate/action_model': 7.80816347467429e-06, 'epoch': 2.57}) +[2026-06-18 02:18:39,520][starVLA.training.train_starvla][INFO] - Step 6242, Loss: {'train/loss': 0.22746339440345764, 'train/grad_norm_pre_clip': 19.399635314941406, 'timing/data': 0.0010360628366470337, 'timing/model': 0.9429254466667771, 'learning_rate/qwen_vl_interface': 7.800842254117922e-07, 'learning_rate/action_model': 7.80084225411792e-06, 'epoch': 2.57}) +[2026-06-18 02:18:43,321][starVLA.training.train_starvla][INFO] - Step 6243, Loss: {'train/loss': 0.3908092975616455, 'train/grad_norm_pre_clip': 27.696075439453125, 'timing/data': 0.0006712367758154869, 'timing/model': 1.1255478486418724, 'learning_rate/qwen_vl_interface': 7.793530299745061e-07, 'learning_rate/action_model': 7.793530299745061e-06, 'epoch': 2.57}) +[2026-06-18 02:18:47,048][starVLA.training.train_starvla][INFO] - Step 6244, Loss: {'train/loss': 0.23199418187141418, 'train/grad_norm_pre_clip': 20.883499145507812, 'timing/data': 0.0004788525402545929, 'timing/model': 0.9890477042645216, 'learning_rate/qwen_vl_interface': 7.786227613071464e-07, 'learning_rate/action_model': 7.786227613071464e-06, 'epoch': 2.57}) +[2026-06-18 02:18:50,991][starVLA.training.train_starvla][INFO] - Step 6245, Loss: {'train/loss': 0.2913872003555298, 'train/grad_norm_pre_clip': 19.76423454284668, 'timing/data': 0.0006413711234927177, 'timing/model': 0.9996168334037066, 'learning_rate/qwen_vl_interface': 7.778934195610998e-07, 'learning_rate/action_model': 7.778934195610998e-06, 'epoch': 2.57}) +[2026-06-18 02:18:54,861][starVLA.training.train_starvla][INFO] - Step 6246, Loss: {'train/loss': 0.14890159666538239, 'train/grad_norm_pre_clip': 13.399557113647461, 'timing/data': 0.0006898157298564911, 'timing/model': 0.9550911914557219, 'learning_rate/qwen_vl_interface': 7.771650048875598e-07, 'learning_rate/action_model': 7.771650048875597e-06, 'epoch': 2.57}) +[2026-06-18 02:19:00,002][starVLA.training.train_starvla][INFO] - Step 6247, Loss: {'train/loss': 0.15743263065814972, 'train/grad_norm_pre_clip': 23.69878578186035, 'timing/data': 0.0005502747371792793, 'timing/model': 0.9326006732881069, 'learning_rate/qwen_vl_interface': 7.764375174375276e-07, 'learning_rate/action_model': 7.764375174375276e-06, 'epoch': 2.57}) +[2026-06-18 02:19:03,849][starVLA.training.train_starvla][INFO] - Step 6248, Loss: {'train/loss': 0.34129709005355835, 'train/grad_norm_pre_clip': 26.510412216186523, 'timing/data': 0.0014022132381796837, 'timing/model': 0.9659069487825036, 'learning_rate/qwen_vl_interface': 7.757109573618116e-07, 'learning_rate/action_model': 7.757109573618115e-06, 'epoch': 2.57}) +[2026-06-18 02:19:07,955][starVLA.training.train_starvla][INFO] - Step 6249, Loss: {'train/loss': 0.27491068840026855, 'train/grad_norm_pre_clip': 27.43045425415039, 'timing/data': 0.0010302979499101639, 'timing/model': 1.3212281558662653, 'learning_rate/qwen_vl_interface': 7.749853248110287e-07, 'learning_rate/action_model': 7.749853248110287e-06, 'epoch': 2.57}) +[2026-06-18 02:19:11,909][starVLA.training.train_starvla][INFO] - Step 6250, Loss: {'train/loss': 0.2033684104681015, 'train/grad_norm_pre_clip': 19.802005767822266, 'timing/data': 0.0006634313613176346, 'timing/model': 1.1163589777424932, 'learning_rate/qwen_vl_interface': 7.742606199356032e-07, 'learning_rate/action_model': 7.74260619935603e-06, 'epoch': 2.57}) +[2026-06-18 02:19:15,783][starVLA.training.train_starvla][INFO] - Step 6251, Loss: {'train/loss': 0.01797325722873211, 'train/grad_norm_pre_clip': 2.580198049545288, 'timing/data': 0.0004598153755068779, 'timing/model': 0.940858606249094, 'learning_rate/qwen_vl_interface': 7.735368428857665e-07, 'learning_rate/action_model': 7.735368428857665e-06, 'epoch': 2.57}) +[2026-06-18 02:19:19,704][starVLA.training.train_starvla][INFO] - Step 6252, Loss: {'train/loss': 0.5548270344734192, 'train/grad_norm_pre_clip': 35.86130142211914, 'timing/data': 0.0015550442039966583, 'timing/model': 0.9581643063575029, 'learning_rate/qwen_vl_interface': 7.728139938115605e-07, 'learning_rate/action_model': 7.728139938115604e-06, 'epoch': 2.57}) +[2026-06-18 02:19:24,164][starVLA.training.train_starvla][INFO] - Step 6253, Loss: {'train/loss': 0.3413074016571045, 'train/grad_norm_pre_clip': 18.808940887451172, 'timing/data': 0.0007287925109267235, 'timing/model': 0.9461488565430045, 'learning_rate/qwen_vl_interface': 7.720920728628306e-07, 'learning_rate/action_model': 7.720920728628306e-06, 'epoch': 2.57}) +[2026-06-18 02:19:28,081][starVLA.training.train_starvla][INFO] - Step 6254, Loss: {'train/loss': 0.22178061306476593, 'train/grad_norm_pre_clip': 16.201507568359375, 'timing/data': 0.0004764087498188019, 'timing/model': 0.9449496204033494, 'learning_rate/qwen_vl_interface': 7.713710801892327e-07, 'learning_rate/action_model': 7.713710801892326e-06, 'epoch': 2.57}) +[2026-06-18 02:19:31,916][starVLA.training.train_starvla][INFO] - Step 6255, Loss: {'train/loss': 0.5092328786849976, 'train/grad_norm_pre_clip': 24.433578491210938, 'timing/data': 0.00045344047248363495, 'timing/model': 0.9397675720974803, 'learning_rate/qwen_vl_interface': 7.70651015940228e-07, 'learning_rate/action_model': 7.70651015940228e-06, 'epoch': 2.57}) +[2026-06-18 02:19:35,816][starVLA.training.train_starvla][INFO] - Step 6256, Loss: {'train/loss': 0.0625828206539154, 'train/grad_norm_pre_clip': 8.815007209777832, 'timing/data': 0.0008615627884864807, 'timing/model': 1.1212326372042298, 'learning_rate/qwen_vl_interface': 7.699318802650891e-07, 'learning_rate/action_model': 7.69931880265089e-06, 'epoch': 2.58}) +[2026-06-18 02:19:39,655][starVLA.training.train_starvla][INFO] - Step 6257, Loss: {'train/loss': 0.24148264527320862, 'train/grad_norm_pre_clip': 18.411792755126953, 'timing/data': 0.0005066925659775734, 'timing/model': 1.1428087670356035, 'learning_rate/qwen_vl_interface': 7.692136733128918e-07, 'learning_rate/action_model': 7.692136733128917e-06, 'epoch': 2.58}) +[2026-06-18 02:19:43,323][starVLA.training.train_starvla][INFO] - Step 6258, Loss: {'train/loss': 0.12506315112113953, 'train/grad_norm_pre_clip': 27.39794158935547, 'timing/data': 0.0004826448857784271, 'timing/model': 0.9430647240951657, 'learning_rate/qwen_vl_interface': 7.684963952325207e-07, 'learning_rate/action_model': 7.684963952325206e-06, 'epoch': 2.58}) +[2026-06-18 02:19:47,150][starVLA.training.train_starvla][INFO] - Step 6259, Loss: {'train/loss': 0.5023782253265381, 'train/grad_norm_pre_clip': 42.7435417175293, 'timing/data': 0.000489954836666584, 'timing/model': 0.9335048766806722, 'learning_rate/qwen_vl_interface': 7.6778004617267e-07, 'learning_rate/action_model': 7.6778004617267e-06, 'epoch': 2.58}) +[2026-06-18 02:19:51,644][starVLA.training.train_starvla][INFO] - Step 6260, Loss: {'train/loss': 0.30694904923439026, 'train/grad_norm_pre_clip': 22.335683822631836, 'timing/data': 0.0010546371340751648, 'timing/model': 0.9284235704690218, 'learning_rate/qwen_vl_interface': 7.670646262818387e-07, 'learning_rate/action_model': 7.670646262818387e-06, 'epoch': 2.58}) +[2026-06-18 02:19:55,441][starVLA.training.train_starvla][INFO] - Step 6261, Loss: {'train/loss': 0.08432327210903168, 'train/grad_norm_pre_clip': 13.581619262695312, 'timing/data': 0.024905587546527386, 'timing/model': 0.9316524537280202, 'learning_rate/qwen_vl_interface': 7.663501357083342e-07, 'learning_rate/action_model': 7.66350135708334e-06, 'epoch': 2.58}) +[2026-06-18 02:19:59,324][starVLA.training.train_starvla][INFO] - Step 6262, Loss: {'train/loss': 0.26099321246147156, 'train/grad_norm_pre_clip': 16.51962661743164, 'timing/data': 0.030571285635232925, 'timing/model': 0.9345347750931978, 'learning_rate/qwen_vl_interface': 7.656365746002715e-07, 'learning_rate/action_model': 7.656365746002716e-06, 'epoch': 2.58}) +[2026-06-18 02:20:03,143][starVLA.training.train_starvla][INFO] - Step 6263, Loss: {'train/loss': 0.5283418297767639, 'train/grad_norm_pre_clip': 28.416458129882812, 'timing/data': 0.018678082153201103, 'timing/model': 1.135894862934947, 'learning_rate/qwen_vl_interface': 7.649239431055718e-07, 'learning_rate/action_model': 7.649239431055717e-06, 'epoch': 2.58}) +[2026-06-18 02:20:07,070][starVLA.training.train_starvla][INFO] - Step 6264, Loss: {'train/loss': 0.24612998962402344, 'train/grad_norm_pre_clip': 30.395893096923828, 'timing/data': 0.030277266167104244, 'timing/model': 1.2001368654891849, 'learning_rate/qwen_vl_interface': 7.642122413719652e-07, 'learning_rate/action_model': 7.642122413719651e-06, 'epoch': 2.58}) +[2026-06-18 02:20:10,839][starVLA.training.train_starvla][INFO] - Step 6265, Loss: {'train/loss': 0.32114270329475403, 'train/grad_norm_pre_clip': 26.21355628967285, 'timing/data': 0.022599115036427975, 'timing/model': 0.9393677450716496, 'learning_rate/qwen_vl_interface': 7.635014695469877e-07, 'learning_rate/action_model': 7.635014695469878e-06, 'epoch': 2.58}) +[2026-06-18 02:20:14,653][starVLA.training.train_starvla][INFO] - Step 6266, Loss: {'train/loss': 0.24223704636096954, 'train/grad_norm_pre_clip': 18.46766471862793, 'timing/data': 0.02136178594082594, 'timing/model': 0.9410879043862224, 'learning_rate/qwen_vl_interface': 7.627916277779832e-07, 'learning_rate/action_model': 7.627916277779831e-06, 'epoch': 2.58}) +[2026-06-18 02:20:18,637][starVLA.training.train_starvla][INFO] - Step 6267, Loss: {'train/loss': 0.27951595187187195, 'train/grad_norm_pre_clip': 20.571861267089844, 'timing/data': 0.02157275564968586, 'timing/model': 0.9292186368256807, 'learning_rate/qwen_vl_interface': 7.62082716212103e-07, 'learning_rate/action_model': 7.62082716212103e-06, 'epoch': 2.58}) +[2026-06-18 02:20:22,636][starVLA.training.train_starvla][INFO] - Step 6268, Loss: {'train/loss': 0.1679152250289917, 'train/grad_norm_pre_clip': 14.570737838745117, 'timing/data': 0.02633353602141142, 'timing/model': 0.9296197136864066, 'learning_rate/qwen_vl_interface': 7.613747349963048e-07, 'learning_rate/action_model': 7.613747349963048e-06, 'epoch': 2.58}) +[2026-06-18 02:20:26,425][starVLA.training.train_starvla][INFO] - Step 6269, Loss: {'train/loss': 0.14533841609954834, 'train/grad_norm_pre_clip': 12.904332160949707, 'timing/data': 0.02885937225073576, 'timing/model': 0.9357259059324861, 'learning_rate/qwen_vl_interface': 7.606676842773531e-07, 'learning_rate/action_model': 7.60667684277353e-06, 'epoch': 2.58}) +[2026-06-18 02:20:30,198][starVLA.training.train_starvla][INFO] - Step 6270, Loss: {'train/loss': 0.3443780243396759, 'train/grad_norm_pre_clip': 17.493778228759766, 'timing/data': 0.02114016469568014, 'timing/model': 1.103133162483573, 'learning_rate/qwen_vl_interface': 7.599615642018218e-07, 'learning_rate/action_model': 7.599615642018218e-06, 'epoch': 2.58}) +[2026-06-18 02:20:33,869][starVLA.training.train_starvla][INFO] - Step 6271, Loss: {'train/loss': 0.2312401831150055, 'train/grad_norm_pre_clip': 21.597497940063477, 'timing/data': 0.02493153139948845, 'timing/model': 0.9368561189621687, 'learning_rate/qwen_vl_interface': 7.592563749160908e-07, 'learning_rate/action_model': 7.592563749160908e-06, 'epoch': 2.58}) +[2026-06-18 02:20:37,910][starVLA.training.train_starvla][INFO] - Step 6272, Loss: {'train/loss': 0.19789086282253265, 'train/grad_norm_pre_clip': 16.671497344970703, 'timing/data': 0.022074482403695583, 'timing/model': 1.0258581275120378, 'learning_rate/qwen_vl_interface': 7.585521165663445e-07, 'learning_rate/action_model': 7.585521165663445e-06, 'epoch': 2.58}) +[2026-06-18 02:20:41,731][starVLA.training.train_starvla][INFO] - Step 6273, Loss: {'train/loss': 0.2268335521221161, 'train/grad_norm_pre_clip': 21.435617446899414, 'timing/data': 0.02049992699176073, 'timing/model': 0.9266698844730854, 'learning_rate/qwen_vl_interface': 7.578487892985769e-07, 'learning_rate/action_model': 7.578487892985769e-06, 'epoch': 2.58}) +[2026-06-18 02:20:45,608][starVLA.training.train_starvla][INFO] - Step 6274, Loss: {'train/loss': 0.384670227766037, 'train/grad_norm_pre_clip': 26.68500518798828, 'timing/data': 0.02654128149151802, 'timing/model': 0.9463480738922954, 'learning_rate/qwen_vl_interface': 7.571463932585901e-07, 'learning_rate/action_model': 7.571463932585901e-06, 'epoch': 2.58}) +[2026-06-18 02:20:49,621][starVLA.training.train_starvla][INFO] - Step 6275, Loss: {'train/loss': 0.25503504276275635, 'train/grad_norm_pre_clip': 23.29729461669922, 'timing/data': 0.02990440186113119, 'timing/model': 0.9911881014704704, 'learning_rate/qwen_vl_interface': 7.564449285919897e-07, 'learning_rate/action_model': 7.564449285919897e-06, 'epoch': 2.58}) +[2026-06-18 02:20:53,551][starVLA.training.train_starvla][INFO] - Step 6276, Loss: {'train/loss': 0.10854391008615494, 'train/grad_norm_pre_clip': 11.851996421813965, 'timing/data': 0.03094175923615694, 'timing/model': 0.943066930398345, 'learning_rate/qwen_vl_interface': 7.557443954441924e-07, 'learning_rate/action_model': 7.557443954441923e-06, 'epoch': 2.58}) +[2026-06-18 02:20:57,421][starVLA.training.train_starvla][INFO] - Step 6277, Loss: {'train/loss': 0.18354934453964233, 'train/grad_norm_pre_clip': 18.979934692382812, 'timing/data': 0.023145589977502823, 'timing/model': 1.1128897108137608, 'learning_rate/qwen_vl_interface': 7.550447939604161e-07, 'learning_rate/action_model': 7.550447939604161e-06, 'epoch': 2.58}) +[2026-06-18 02:21:01,081][starVLA.training.train_starvla][INFO] - Step 6278, Loss: {'train/loss': 0.5353859066963196, 'train/grad_norm_pre_clip': 23.378164291381836, 'timing/data': 0.023528452962636948, 'timing/model': 0.963941958732903, 'learning_rate/qwen_vl_interface': 7.543461242856916e-07, 'learning_rate/action_model': 7.543461242856916e-06, 'epoch': 2.58}) +[2026-06-18 02:21:06,427][starVLA.training.train_starvla][INFO] - Step 6279, Loss: {'train/loss': 0.18179264664649963, 'train/grad_norm_pre_clip': 20.526926040649414, 'timing/data': 0.02317905519157648, 'timing/model': 2.3101786114275455, 'learning_rate/qwen_vl_interface': 7.536483865648531e-07, 'learning_rate/action_model': 7.536483865648531e-06, 'epoch': 2.58}) +[2026-06-18 02:21:10,553][starVLA.training.train_starvla][INFO] - Step 6280, Loss: {'train/loss': 0.25830748677253723, 'train/grad_norm_pre_clip': 58.89250946044922, 'timing/data': 0.02318658586591482, 'timing/model': 0.940150267444551, 'learning_rate/qwen_vl_interface': 7.52951580942542e-07, 'learning_rate/action_model': 7.5295158094254195e-06, 'epoch': 2.59}) +[2026-06-18 02:21:14,536][starVLA.training.train_starvla][INFO] - Step 6281, Loss: {'train/loss': 0.0955120399594307, 'train/grad_norm_pre_clip': 10.773231506347656, 'timing/data': 0.026931488886475563, 'timing/model': 0.9506739201024175, 'learning_rate/qwen_vl_interface': 7.522557075632078e-07, 'learning_rate/action_model': 7.522557075632078e-06, 'epoch': 2.59}) +[2026-06-18 02:21:18,582][starVLA.training.train_starvla][INFO] - Step 6282, Loss: {'train/loss': 0.18663229048252106, 'train/grad_norm_pre_clip': 13.553313255310059, 'timing/data': 0.03062955942004919, 'timing/model': 1.1876443140208721, 'learning_rate/qwen_vl_interface': 7.515607665711051e-07, 'learning_rate/action_model': 7.515607665711051e-06, 'epoch': 2.59}) +[2026-06-18 02:21:22,436][starVLA.training.train_starvla][INFO] - Step 6283, Loss: {'train/loss': 0.12416023015975952, 'train/grad_norm_pre_clip': 17.997323989868164, 'timing/data': 0.02691403403878212, 'timing/model': 1.1082898871973157, 'learning_rate/qwen_vl_interface': 7.508667581102959e-07, 'learning_rate/action_model': 7.508667581102958e-06, 'epoch': 2.59}) +[2026-06-18 02:21:26,252][starVLA.training.train_starvla][INFO] - Step 6284, Loss: {'train/loss': 0.21604228019714355, 'train/grad_norm_pre_clip': 23.686885833740234, 'timing/data': 0.03286201041191816, 'timing/model': 1.1035579666495323, 'learning_rate/qwen_vl_interface': 7.50173682324648e-07, 'learning_rate/action_model': 7.501736823246481e-06, 'epoch': 2.59}) +[2026-06-18 02:21:30,085][starVLA.training.train_starvla][INFO] - Step 6285, Loss: {'train/loss': 0.37317973375320435, 'train/grad_norm_pre_clip': 30.17986297607422, 'timing/data': 0.02576520573347807, 'timing/model': 1.020307233557105, 'learning_rate/qwen_vl_interface': 7.494815393578394e-07, 'learning_rate/action_model': 7.494815393578393e-06, 'epoch': 2.59}) +[2026-06-18 02:21:34,029][starVLA.training.train_starvla][INFO] - Step 6286, Loss: {'train/loss': 0.2897878587245941, 'train/grad_norm_pre_clip': 24.43921661376953, 'timing/data': 0.027264684438705444, 'timing/model': 0.9369243560358882, 'learning_rate/qwen_vl_interface': 7.487903293533501e-07, 'learning_rate/action_model': 7.4879032935335e-06, 'epoch': 2.59}) +[2026-06-18 02:21:37,834][starVLA.training.train_starvla][INFO] - Step 6287, Loss: {'train/loss': 0.16451969742774963, 'train/grad_norm_pre_clip': 17.763870239257812, 'timing/data': 0.020750150084495544, 'timing/model': 0.9287868058308959, 'learning_rate/qwen_vl_interface': 7.481000524544686e-07, 'learning_rate/action_model': 7.481000524544686e-06, 'epoch': 2.59}) +[2026-06-18 02:21:41,643][starVLA.training.train_starvla][INFO] - Step 6288, Loss: {'train/loss': 0.06689835339784622, 'train/grad_norm_pre_clip': 11.970272064208984, 'timing/data': 0.01968424953520298, 'timing/model': 0.9391515776515007, 'learning_rate/qwen_vl_interface': 7.474107088042898e-07, 'learning_rate/action_model': 7.474107088042898e-06, 'epoch': 2.59}) +[2026-06-18 02:21:45,439][starVLA.training.train_starvla][INFO] - Step 6289, Loss: {'train/loss': 0.24825192987918854, 'train/grad_norm_pre_clip': 25.92310333251953, 'timing/data': 0.019848186522722244, 'timing/model': 0.9317050240933895, 'learning_rate/qwen_vl_interface': 7.467222985457167e-07, 'learning_rate/action_model': 7.4672229854571675e-06, 'epoch': 2.59}) +[2026-06-18 02:21:49,242][starVLA.training.train_starvla][INFO] - Step 6290, Loss: {'train/loss': 0.5119460821151733, 'train/grad_norm_pre_clip': 31.069223403930664, 'timing/data': 0.023264575749635696, 'timing/model': 1.0960265258327127, 'learning_rate/qwen_vl_interface': 7.46034821821458e-07, 'learning_rate/action_model': 7.46034821821458e-06, 'epoch': 2.59}) +[2026-06-18 02:21:53,048][starVLA.training.train_starvla][INFO] - Step 6291, Loss: {'train/loss': 0.1293214112520218, 'train/grad_norm_pre_clip': 13.769978523254395, 'timing/data': 0.029208440333604813, 'timing/model': 1.075568050146103, 'learning_rate/qwen_vl_interface': 7.453482787740255e-07, 'learning_rate/action_model': 7.4534827877402545e-06, 'epoch': 2.59}) +[2026-06-18 02:21:56,699][starVLA.training.train_starvla][INFO] - Step 6292, Loss: {'train/loss': 0.13325048983097076, 'train/grad_norm_pre_clip': 19.707012176513672, 'timing/data': 0.026012624613940716, 'timing/model': 0.9422262283042073, 'learning_rate/qwen_vl_interface': 7.446626695457427e-07, 'learning_rate/action_model': 7.446626695457427e-06, 'epoch': 2.59}) +[2026-06-18 02:22:00,877][starVLA.training.train_starvla][INFO] - Step 6293, Loss: {'train/loss': 0.4452400207519531, 'train/grad_norm_pre_clip': 26.7034969329834, 'timing/data': 0.02852143906056881, 'timing/model': 0.9389120927080512, 'learning_rate/qwen_vl_interface': 7.439779942787367e-07, 'learning_rate/action_model': 7.439779942787367e-06, 'epoch': 2.59}) +[2026-06-18 02:22:04,724][starVLA.training.train_starvla][INFO] - Step 6294, Loss: {'train/loss': 0.0998404398560524, 'train/grad_norm_pre_clip': 13.680944442749023, 'timing/data': 0.029165604151785374, 'timing/model': 0.9472716506570578, 'learning_rate/qwen_vl_interface': 7.432942531149409e-07, 'learning_rate/action_model': 7.432942531149409e-06, 'epoch': 2.59}) +[2026-06-18 02:22:08,562][starVLA.training.train_starvla][INFO] - Step 6295, Loss: {'train/loss': 0.23658868670463562, 'train/grad_norm_pre_clip': 27.114965438842773, 'timing/data': 0.027253509499132633, 'timing/model': 0.9660208355635405, 'learning_rate/qwen_vl_interface': 7.42611446196096e-07, 'learning_rate/action_model': 7.426114461960959e-06, 'epoch': 2.59}) +[2026-06-18 02:22:12,369][starVLA.training.train_starvla][INFO] - Step 6296, Loss: {'train/loss': 0.13401247560977936, 'train/grad_norm_pre_clip': 14.511664390563965, 'timing/data': 0.02642970997840166, 'timing/model': 0.9313421826809645, 'learning_rate/qwen_vl_interface': 7.419295736637483e-07, 'learning_rate/action_model': 7.419295736637483e-06, 'epoch': 2.59}) +[2026-06-18 02:22:16,232][starVLA.training.train_starvla][INFO] - Step 6297, Loss: {'train/loss': 0.3398050367832184, 'train/grad_norm_pre_clip': 21.00459098815918, 'timing/data': 0.026417779736220837, 'timing/model': 1.113821747712791, 'learning_rate/qwen_vl_interface': 7.412486356592512e-07, 'learning_rate/action_model': 7.412486356592511e-06, 'epoch': 2.59}) +[2026-06-18 02:22:19,868][starVLA.training.train_starvla][INFO] - Step 6298, Loss: {'train/loss': 0.210926353931427, 'train/grad_norm_pre_clip': 18.49420166015625, 'timing/data': 0.029826410114765167, 'timing/model': 0.9293524576351047, 'learning_rate/qwen_vl_interface': 7.405686323237623e-07, 'learning_rate/action_model': 7.405686323237623e-06, 'epoch': 2.59}) +[2026-06-18 02:22:23,633][starVLA.training.train_starvla][INFO] - Step 6299, Loss: {'train/loss': 0.3333794176578522, 'train/grad_norm_pre_clip': 24.976898193359375, 'timing/data': 0.025495686568319798, 'timing/model': 0.9351337803527713, 'learning_rate/qwen_vl_interface': 7.398895637982504e-07, 'learning_rate/action_model': 7.398895637982503e-06, 'epoch': 2.59}) +[2026-06-18 02:24:32,258][starVLA.training.train_starvla][INFO] - Step 6300, Loss: {'train/loss': 0.09969823062419891, 'train/grad_norm_pre_clip': 14.303960800170898, 'eval/loss': 0.4166637063026428, 'eval/latency_6/loss': 0.4166637063026428, 'eval/demon_fire_precision': 0.956370534302877, 'eval/demon_fire_recall': 0.9776987718164188, 'eval/demon_fire_f1': 0.9669170529007513, 'eval/demon_fire_rate_model': 0.8800779076238174, 'eval/demon_fire_rate_teacher': 0.860879243183083, 'eval/demon_left_precision': 0.8615984405458089, 'eval/demon_left_recall': 0.775438596491228, 'eval/demon_left_f1': 0.8162511542012928, 'eval/demon_left_rate_model': 0.14273789649415694, 'eval/demon_left_rate_teacher': 0.15859766277128548, 'eval/demon_right_precision': 0.8352553542009885, 'eval/demon_right_recall': 0.7788018433179723, 'eval/demon_right_f1': 0.8060413354531001, 'eval/demon_right_rate_model': 0.16889259877573734, 'eval/demon_right_rate_teacher': 0.18113522537562604, 'eval/demon_move_f1': 0.8111462448271964, 'eval/demon_cc_f1': 0.9142460877889158, 'eval/latency_6/demon_fire_precision': 0.956370534302877, 'eval/latency_6/demon_fire_recall': 0.9776987718164188, 'eval/latency_6/demon_fire_f1': 0.9669170529007513, 'eval/latency_6/demon_fire_rate_model': 0.8800779076238174, 'eval/latency_6/demon_fire_rate_teacher': 0.860879243183083, 'eval/latency_6/demon_left_precision': 0.8615984405458089, 'eval/latency_6/demon_left_recall': 0.775438596491228, 'eval/latency_6/demon_left_f1': 0.8162511542012928, 'eval/latency_6/demon_left_rate_model': 0.14273789649415694, 'eval/latency_6/demon_left_rate_teacher': 0.15859766277128548, 'eval/latency_6/demon_right_precision': 0.8352553542009885, 'eval/latency_6/demon_right_recall': 0.7788018433179723, 'eval/latency_6/demon_right_f1': 0.8060413354531001, 'eval/latency_6/demon_right_rate_model': 0.16889259877573734, 'eval/latency_6/demon_right_rate_teacher': 0.18113522537562604, 'eval/latency_6/demon_move_f1': 0.8111462448271964, 'eval/latency_6/demon_cc_f1': 0.9142460877889158, 'timing/data': 0.03134086448699236, 'timing/model': 0.9331677360460162, 'learning_rate/qwen_vl_interface': 7.392114302234838e-07, 'learning_rate/action_model': 7.3921143022348385e-06, 'epoch': 2.59}) +[2026-06-18 02:24:32,268][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_6300_state +[2026-06-18 02:24:32,273][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 02:24:39,092][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_6300_state/model.safetensors +[2026-06-18 02:24:54,566][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_6300_state/optimizer.bin +[2026-06-18 02:24:54,567][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_6300_state/sampler.bin +[2026-06-18 02:24:54,567][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_6300_state/sampler_1.bin +[2026-06-18 02:24:54,569][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_6300_state/random_states_0.pkl +[2026-06-18 02:24:54,573][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 6300 +[2026-06-18 02:24:54,573][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 02:24:54,583][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 02:24:58,471][starVLA.training.train_starvla][INFO] - Step 6301, Loss: {'train/loss': 0.2619851529598236, 'train/grad_norm_pre_clip': 19.955005645751953, 'timing/data': 0.03475914150476456, 'timing/model': 0.9354873737320304, 'learning_rate/qwen_vl_interface': 7.385342317400418e-07, 'learning_rate/action_model': 7.385342317400417e-06, 'epoch': 2.59}) +[2026-06-18 02:25:05,561][starVLA.training.train_starvla][INFO] - Step 6302, Loss: {'train/loss': 0.09386761486530304, 'train/grad_norm_pre_clip': 9.999011993408203, 'timing/data': 0.03711267188191414, 'timing/model': 1.097212865948677, 'learning_rate/qwen_vl_interface': 7.378579684883069e-07, 'learning_rate/action_model': 7.378579684883069e-06, 'epoch': 2.59}) +[2026-06-18 02:25:11,519][starVLA.training.train_starvla][INFO] - Step 6303, Loss: {'train/loss': 0.2396346628665924, 'train/grad_norm_pre_clip': 21.52444839477539, 'timing/data': 0.025820251554250717, 'timing/model': 0.9577507460489869, 'learning_rate/qwen_vl_interface': 7.371826406084714e-07, 'learning_rate/action_model': 7.371826406084714e-06, 'epoch': 2.59}) +[2026-06-18 02:25:15,392][starVLA.training.train_starvla][INFO] - Step 6304, Loss: {'train/loss': 0.3949248790740967, 'train/grad_norm_pre_clip': 26.6998348236084, 'timing/data': 0.02735507022589445, 'timing/model': 0.937165305018425, 'learning_rate/qwen_vl_interface': 7.365082482405315e-07, 'learning_rate/action_model': 7.365082482405315e-06, 'epoch': 2.59}) +[2026-06-18 02:25:19,235][starVLA.training.train_starvla][INFO] - Step 6305, Loss: {'train/loss': 0.24512799084186554, 'train/grad_norm_pre_clip': 23.552846908569336, 'timing/data': 0.04183019232004881, 'timing/model': 0.9386192532256246, 'learning_rate/qwen_vl_interface': 7.358347915242877e-07, 'learning_rate/action_model': 7.358347915242876e-06, 'epoch': 2.6}) +[2026-06-18 02:25:23,026][starVLA.training.train_starvla][INFO] - Step 6306, Loss: {'train/loss': 0.34206658601760864, 'train/grad_norm_pre_clip': 24.48470115661621, 'timing/data': 0.021945087239146233, 'timing/model': 1.1229987377300858, 'learning_rate/qwen_vl_interface': 7.351622705993481e-07, 'learning_rate/action_model': 7.351622705993481e-06, 'epoch': 2.6}) +[2026-06-18 02:25:26,640][starVLA.training.train_starvla][INFO] - Step 6307, Loss: {'train/loss': 0.09090424329042435, 'train/grad_norm_pre_clip': 11.91565990447998, 'timing/data': 0.023423975333571434, 'timing/model': 0.9270670898258686, 'learning_rate/qwen_vl_interface': 7.344906856051288e-07, 'learning_rate/action_model': 7.344906856051288e-06, 'epoch': 2.6}) +[2026-06-18 02:25:30,455][starVLA.training.train_starvla][INFO] - Step 6308, Loss: {'train/loss': 0.3336290419101715, 'train/grad_norm_pre_clip': 27.325763702392578, 'timing/data': 0.03222379833459854, 'timing/model': 0.9318666150793433, 'learning_rate/qwen_vl_interface': 7.338200366808493e-07, 'learning_rate/action_model': 7.338200366808492e-06, 'epoch': 2.6}) +[2026-06-18 02:25:34,235][starVLA.training.train_starvla][INFO] - Step 6309, Loss: {'train/loss': 0.1743260622024536, 'train/grad_norm_pre_clip': 18.24211883544922, 'timing/data': 0.022621233016252518, 'timing/model': 0.9315601121634245, 'learning_rate/qwen_vl_interface': 7.331503239655361e-07, 'learning_rate/action_model': 7.33150323965536e-06, 'epoch': 2.6}) +[2026-06-18 02:25:38,141][starVLA.training.train_starvla][INFO] - Step 6310, Loss: {'train/loss': 0.272786021232605, 'train/grad_norm_pre_clip': 24.94366455078125, 'timing/data': 0.028892493806779385, 'timing/model': 0.9548763306811452, 'learning_rate/qwen_vl_interface': 7.324815475980207e-07, 'learning_rate/action_model': 7.324815475980207e-06, 'epoch': 2.6}) +[2026-06-18 02:25:41,973][starVLA.training.train_starvla][INFO] - Step 6311, Loss: {'train/loss': 0.18899555504322052, 'train/grad_norm_pre_clip': 15.092766761779785, 'timing/data': 0.02798144891858101, 'timing/model': 0.9349259696900845, 'learning_rate/qwen_vl_interface': 7.318137077169416e-07, 'learning_rate/action_model': 7.318137077169415e-06, 'epoch': 2.6}) +[2026-06-18 02:25:45,981][starVLA.training.train_starvla][INFO] - Step 6312, Loss: {'train/loss': 0.24458086490631104, 'train/grad_norm_pre_clip': 33.01985549926758, 'timing/data': 0.034441206604242325, 'timing/model': 0.9875517906621099, 'learning_rate/qwen_vl_interface': 7.311468044607425e-07, 'learning_rate/action_model': 7.311468044607425e-06, 'epoch': 2.6}) +[2026-06-18 02:25:49,909][starVLA.training.train_starvla][INFO] - Step 6313, Loss: {'train/loss': 0.6714113354682922, 'train/grad_norm_pre_clip': 44.328826904296875, 'timing/data': 0.03303501661866903, 'timing/model': 1.1119030835106969, 'learning_rate/qwen_vl_interface': 7.304808379676739e-07, 'learning_rate/action_model': 7.3048083796767394e-06, 'epoch': 2.6}) +[2026-06-18 02:25:53,540][starVLA.training.train_starvla][INFO] - Step 6314, Loss: {'train/loss': 0.16676458716392517, 'train/grad_norm_pre_clip': 16.19478416442871, 'timing/data': 0.02320060972124338, 'timing/model': 0.9374549742788076, 'learning_rate/qwen_vl_interface': 7.298158083757903e-07, 'learning_rate/action_model': 7.298158083757903e-06, 'epoch': 2.6}) +[2026-06-18 02:25:57,334][starVLA.training.train_starvla][INFO] - Step 6315, Loss: {'train/loss': 0.2715276777744293, 'train/grad_norm_pre_clip': 27.25460433959961, 'timing/data': 0.024177036248147488, 'timing/model': 0.9516680771484971, 'learning_rate/qwen_vl_interface': 7.291517158229536e-07, 'learning_rate/action_model': 7.291517158229535e-06, 'epoch': 2.6}) +[2026-06-18 02:26:01,162][starVLA.training.train_starvla][INFO] - Step 6316, Loss: {'train/loss': 0.5634167194366455, 'train/grad_norm_pre_clip': 24.206130981445312, 'timing/data': 0.03050078544765711, 'timing/model': 0.9363576686009765, 'learning_rate/qwen_vl_interface': 7.284885604468312e-07, 'learning_rate/action_model': 7.284885604468311e-06, 'epoch': 2.6}) +[2026-06-18 02:26:05,065][starVLA.training.train_starvla][INFO] - Step 6317, Loss: {'train/loss': 0.19653183221817017, 'train/grad_norm_pre_clip': 19.086833953857422, 'timing/data': 0.03185615316033363, 'timing/model': 0.9423448545858264, 'learning_rate/qwen_vl_interface': 7.278263423848948e-07, 'learning_rate/action_model': 7.278263423848948e-06, 'epoch': 2.6}) +[2026-06-18 02:26:09,171][starVLA.training.train_starvla][INFO] - Step 6318, Loss: {'train/loss': 0.23761728405952454, 'train/grad_norm_pre_clip': 28.834943771362305, 'timing/data': 0.02960795257240534, 'timing/model': 0.9481091508641839, 'learning_rate/qwen_vl_interface': 7.271650617744239e-07, 'learning_rate/action_model': 7.271650617744238e-06, 'epoch': 2.6}) +[2026-06-18 02:26:12,924][starVLA.training.train_starvla][INFO] - Step 6319, Loss: {'train/loss': 0.17158158123493195, 'train/grad_norm_pre_clip': 22.12982940673828, 'timing/data': 0.02597891166806221, 'timing/model': 0.9347755117341876, 'learning_rate/qwen_vl_interface': 7.265047187525033e-07, 'learning_rate/action_model': 7.265047187525033e-06, 'epoch': 2.6}) +[2026-06-18 02:26:16,645][starVLA.training.train_starvla][INFO] - Step 6320, Loss: {'train/loss': 0.08732113242149353, 'train/grad_norm_pre_clip': 11.696883201599121, 'timing/data': 0.031554016284644604, 'timing/model': 1.0673364084213972, 'learning_rate/qwen_vl_interface': 7.258453134560212e-07, 'learning_rate/action_model': 7.258453134560211e-06, 'epoch': 2.6}) +[2026-06-18 02:26:20,245][starVLA.training.train_starvla][INFO] - Step 6321, Loss: {'train/loss': 0.21945613622665405, 'train/grad_norm_pre_clip': 23.676546096801758, 'timing/data': 0.023860552348196507, 'timing/model': 0.928895341232419, 'learning_rate/qwen_vl_interface': 7.251868460216743e-07, 'learning_rate/action_model': 7.251868460216743e-06, 'epoch': 2.6}) +[2026-06-18 02:26:24,073][starVLA.training.train_starvla][INFO] - Step 6322, Loss: {'train/loss': 0.27773624658584595, 'train/grad_norm_pre_clip': 26.488510131835938, 'timing/data': 0.028551236726343632, 'timing/model': 0.9377562198787928, 'learning_rate/qwen_vl_interface': 7.245293165859629e-07, 'learning_rate/action_model': 7.245293165859629e-06, 'epoch': 2.6}) +[2026-06-18 02:26:27,875][starVLA.training.train_starvla][INFO] - Step 6323, Loss: {'train/loss': 0.26377686858177185, 'train/grad_norm_pre_clip': 31.682289123535156, 'timing/data': 0.03274958115071058, 'timing/model': 0.9341534608975053, 'learning_rate/qwen_vl_interface': 7.238727252851935e-07, 'learning_rate/action_model': 7.238727252851935e-06, 'epoch': 2.6}) +[2026-06-18 02:26:31,624][starVLA.training.train_starvla][INFO] - Step 6324, Loss: {'train/loss': 0.2959331274032593, 'train/grad_norm_pre_clip': 23.970050811767578, 'timing/data': 0.026923482306301594, 'timing/model': 0.9689122838899493, 'learning_rate/qwen_vl_interface': 7.23217072255479e-07, 'learning_rate/action_model': 7.23217072255479e-06, 'epoch': 2.6}) +[2026-06-18 02:26:35,467][starVLA.training.train_starvla][INFO] - Step 6325, Loss: {'train/loss': 0.18007420003414154, 'train/grad_norm_pre_clip': 20.923837661743164, 'timing/data': 0.03241855464875698, 'timing/model': 0.9535508127883077, 'learning_rate/qwen_vl_interface': 7.225623576327361e-07, 'learning_rate/action_model': 7.22562357632736e-06, 'epoch': 2.6}) +[2026-06-18 02:26:39,230][starVLA.training.train_starvla][INFO] - Step 6326, Loss: {'train/loss': 0.08774624764919281, 'train/grad_norm_pre_clip': 25.441200256347656, 'timing/data': 0.03772482089698315, 'timing/model': 1.0972194308415055, 'learning_rate/qwen_vl_interface': 7.219085815526887e-07, 'learning_rate/action_model': 7.219085815526887e-06, 'epoch': 2.6}) +[2026-06-18 02:26:43,009][starVLA.training.train_starvla][INFO] - Step 6327, Loss: {'train/loss': 0.3222394585609436, 'train/grad_norm_pre_clip': 30.359540939331055, 'timing/data': 0.028063423931598663, 'timing/model': 1.098119887523353, 'learning_rate/qwen_vl_interface': 7.212557441508641e-07, 'learning_rate/action_model': 7.212557441508641e-06, 'epoch': 2.6}) +[2026-06-18 02:26:46,636][starVLA.training.train_starvla][INFO] - Step 6328, Loss: {'train/loss': 0.1577615588903427, 'train/grad_norm_pre_clip': 15.78316593170166, 'timing/data': 0.02273132000118494, 'timing/model': 0.9367827698588371, 'learning_rate/qwen_vl_interface': 7.206038455625972e-07, 'learning_rate/action_model': 7.206038455625971e-06, 'epoch': 2.6}) +[2026-06-18 02:26:50,462][starVLA.training.train_starvla][INFO] - Step 6329, Loss: {'train/loss': 0.13274835050106049, 'train/grad_norm_pre_clip': 15.868226051330566, 'timing/data': 0.021384410560131073, 'timing/model': 0.929345203563571, 'learning_rate/qwen_vl_interface': 7.199528859230271e-07, 'learning_rate/action_model': 7.199528859230272e-06, 'epoch': 2.61}) +[2026-06-18 02:26:54,248][starVLA.training.train_starvla][INFO] - Step 6330, Loss: {'train/loss': 0.2835105061531067, 'train/grad_norm_pre_clip': 29.033737182617188, 'timing/data': 0.024277053773403168, 'timing/model': 0.931479156948626, 'learning_rate/qwen_vl_interface': 7.19302865367098e-07, 'learning_rate/action_model': 7.19302865367098e-06, 'epoch': 2.61}) +[2026-06-18 02:26:58,026][starVLA.training.train_starvla][INFO] - Step 6331, Loss: {'train/loss': 0.3814477324485779, 'train/grad_norm_pre_clip': 42.17695617675781, 'timing/data': 0.03470160160213709, 'timing/model': 0.9324933215975761, 'learning_rate/qwen_vl_interface': 7.186537840295599e-07, 'learning_rate/action_model': 7.186537840295598e-06, 'epoch': 2.61}) +[2026-06-18 02:27:01,831][starVLA.training.train_starvla][INFO] - Step 6332, Loss: {'train/loss': 0.26390540599823, 'train/grad_norm_pre_clip': 24.354990005493164, 'timing/data': 0.024947152473032475, 'timing/model': 0.9530402375385165, 'learning_rate/qwen_vl_interface': 7.18005642044969e-07, 'learning_rate/action_model': 7.180056420449689e-06, 'epoch': 2.61}) +[2026-06-18 02:27:05,679][starVLA.training.train_starvla][INFO] - Step 6333, Loss: {'train/loss': 0.06675494462251663, 'train/grad_norm_pre_clip': 11.432626724243164, 'timing/data': 0.023154789581894875, 'timing/model': 1.140513015910983, 'learning_rate/qwen_vl_interface': 7.173584395476852e-07, 'learning_rate/action_model': 7.173584395476851e-06, 'epoch': 2.61}) +[2026-06-18 02:27:09,369][starVLA.training.train_starvla][INFO] - Step 6334, Loss: {'train/loss': 0.4602217972278595, 'train/grad_norm_pre_clip': 29.86871910095215, 'timing/data': 0.023838416673243046, 'timing/model': 0.9359753206372261, 'learning_rate/qwen_vl_interface': 7.16712176671874e-07, 'learning_rate/action_model': 7.16712176671874e-06, 'epoch': 2.61}) +[2026-06-18 02:27:13,147][starVLA.training.train_starvla][INFO] - Step 6335, Loss: {'train/loss': 0.19218038022518158, 'train/grad_norm_pre_clip': 24.669429779052734, 'timing/data': 0.0269906185567379, 'timing/model': 0.9296325007453561, 'learning_rate/qwen_vl_interface': 7.160668535515059e-07, 'learning_rate/action_model': 7.1606685355150585e-06, 'epoch': 2.61}) +[2026-06-18 02:27:16,853][starVLA.training.train_starvla][INFO] - Step 6336, Loss: {'train/loss': 0.29364070296287537, 'train/grad_norm_pre_clip': 28.23504066467285, 'timing/data': 0.027126071974635124, 'timing/model': 0.9234817856922746, 'learning_rate/qwen_vl_interface': 7.154224703203584e-07, 'learning_rate/action_model': 7.154224703203584e-06, 'epoch': 2.61}) +[2026-06-18 02:27:20,595][starVLA.training.train_starvla][INFO] - Step 6337, Loss: {'train/loss': 0.2810079753398895, 'train/grad_norm_pre_clip': 27.971397399902344, 'timing/data': 0.03386780060827732, 'timing/model': 0.9249671120196581, 'learning_rate/qwen_vl_interface': 7.147790271120122e-07, 'learning_rate/action_model': 7.147790271120122e-06, 'epoch': 2.61}) +[2026-06-18 02:27:24,402][starVLA.training.train_starvla][INFO] - Step 6338, Loss: {'train/loss': 0.09828398376703262, 'train/grad_norm_pre_clip': 18.108396530151367, 'timing/data': 0.024351119063794613, 'timing/model': 0.9581969687715173, 'learning_rate/qwen_vl_interface': 7.141365240598545e-07, 'learning_rate/action_model': 7.141365240598545e-06, 'epoch': 2.61}) +[2026-06-18 02:27:28,211][starVLA.training.train_starvla][INFO] - Step 6339, Loss: {'train/loss': 0.0468943752348423, 'train/grad_norm_pre_clip': 14.09546184539795, 'timing/data': 0.027392110787332058, 'timing/model': 1.092136949300766, 'learning_rate/qwen_vl_interface': 7.134949612970749e-07, 'learning_rate/action_model': 7.134949612970748e-06, 'epoch': 2.61}) +[2026-06-18 02:27:32,123][starVLA.training.train_starvla][INFO] - Step 6340, Loss: {'train/loss': 0.5347493290901184, 'train/grad_norm_pre_clip': 27.99785041809082, 'timing/data': 0.021953686140477657, 'timing/model': 1.0866233836859465, 'learning_rate/qwen_vl_interface': 7.128543389566718e-07, 'learning_rate/action_model': 7.128543389566718e-06, 'epoch': 2.61}) +[2026-06-18 02:27:35,732][starVLA.training.train_starvla][INFO] - Step 6341, Loss: {'train/loss': 0.3756895959377289, 'train/grad_norm_pre_clip': 33.05034255981445, 'timing/data': 0.03826172649860382, 'timing/model': 0.9243721580132842, 'learning_rate/qwen_vl_interface': 7.122146571714464e-07, 'learning_rate/action_model': 7.122146571714463e-06, 'epoch': 2.61}) +[2026-06-18 02:27:39,516][starVLA.training.train_starvla][INFO] - Step 6342, Loss: {'train/loss': 0.07668894529342651, 'train/grad_norm_pre_clip': 10.169820785522461, 'timing/data': 0.0254839975386858, 'timing/model': 0.9227329175919294, 'learning_rate/qwen_vl_interface': 7.115759160740058e-07, 'learning_rate/action_model': 7.115759160740057e-06, 'epoch': 2.61}) +[2026-06-18 02:27:43,338][starVLA.training.train_starvla][INFO] - Step 6343, Loss: {'train/loss': 0.3391647934913635, 'train/grad_norm_pre_clip': 32.440399169921875, 'timing/data': 0.04040519613772631, 'timing/model': 0.9226137353107333, 'learning_rate/qwen_vl_interface': 7.109381157967607e-07, 'learning_rate/action_model': 7.109381157967607e-06, 'epoch': 2.61}) +[2026-06-18 02:27:47,073][starVLA.training.train_starvla][INFO] - Step 6344, Loss: {'train/loss': 0.2109631896018982, 'train/grad_norm_pre_clip': 27.73456573486328, 'timing/data': 0.040337870828807354, 'timing/model': 0.9219957562163472, 'learning_rate/qwen_vl_interface': 7.103012564719289e-07, 'learning_rate/action_model': 7.103012564719288e-06, 'epoch': 2.61}) +[2026-06-18 02:27:50,812][starVLA.training.train_starvla][INFO] - Step 6345, Loss: {'train/loss': 0.18256407976150513, 'train/grad_norm_pre_clip': 16.14691162109375, 'timing/data': 0.023764572106301785, 'timing/model': 0.9202330205589533, 'learning_rate/qwen_vl_interface': 7.096653382315315e-07, 'learning_rate/action_model': 7.096653382315315e-06, 'epoch': 2.61}) +[2026-06-18 02:27:54,558][starVLA.training.train_starvla][INFO] - Step 6346, Loss: {'train/loss': 0.04690966010093689, 'train/grad_norm_pre_clip': 10.50180435180664, 'timing/data': 0.023629692383110523, 'timing/model': 1.0914639001712203, 'learning_rate/qwen_vl_interface': 7.090303612073946e-07, 'learning_rate/action_model': 7.090303612073946e-06, 'epoch': 2.61}) +[2026-06-18 02:27:58,308][starVLA.training.train_starvla][INFO] - Step 6347, Loss: {'train/loss': 0.1986931413412094, 'train/grad_norm_pre_clip': 22.446502685546875, 'timing/data': 0.025315284729003906, 'timing/model': 1.0939173316583037, 'learning_rate/qwen_vl_interface': 7.083963255311508e-07, 'learning_rate/action_model': 7.083963255311509e-06, 'epoch': 2.61}) +[2026-06-18 02:28:01,960][starVLA.training.train_starvla][INFO] - Step 6348, Loss: {'train/loss': 0.3257748782634735, 'train/grad_norm_pre_clip': 13.44936752319336, 'timing/data': 0.04993482772260904, 'timing/model': 0.9624393600970507, 'learning_rate/qwen_vl_interface': 7.07763231334236e-07, 'learning_rate/action_model': 7.077632313342359e-06, 'epoch': 2.61}) +[2026-06-18 02:28:05,747][starVLA.training.train_starvla][INFO] - Step 6349, Loss: {'train/loss': 0.2859381437301636, 'train/grad_norm_pre_clip': 26.2069091796875, 'timing/data': 0.02696585562080145, 'timing/model': 0.9331782646477222, 'learning_rate/qwen_vl_interface': 7.071310787478902e-07, 'learning_rate/action_model': 7.071310787478902e-06, 'epoch': 2.61}) +[2026-06-18 02:28:09,509][starVLA.training.train_starvla][INFO] - Step 6350, Loss: {'train/loss': 0.14955300092697144, 'train/grad_norm_pre_clip': 22.282100677490234, 'timing/data': 0.025770751759409904, 'timing/model': 0.9304885547608137, 'learning_rate/qwen_vl_interface': 7.0649986790316e-07, 'learning_rate/action_model': 7.0649986790315995e-06, 'epoch': 2.61}) +[2026-06-18 02:28:13,253][starVLA.training.train_starvla][INFO] - Step 6351, Loss: {'train/loss': 0.15276449918746948, 'train/grad_norm_pre_clip': 20.177906036376953, 'timing/data': 0.02421610616147518, 'timing/model': 0.9303744425997138, 'learning_rate/qwen_vl_interface': 7.058695989308965e-07, 'learning_rate/action_model': 7.058695989308965e-06, 'epoch': 2.61}) +[2026-06-18 02:28:17,007][starVLA.training.train_starvla][INFO] - Step 6352, Loss: {'train/loss': 0.346159428358078, 'train/grad_norm_pre_clip': 34.44335174560547, 'timing/data': 0.025844584219157696, 'timing/model': 0.9216116378083825, 'learning_rate/qwen_vl_interface': 7.052402719617554e-07, 'learning_rate/action_model': 7.052402719617554e-06, 'epoch': 2.61}) +[2026-06-18 02:28:20,768][starVLA.training.train_starvla][INFO] - Step 6353, Loss: {'train/loss': 0.3547080159187317, 'train/grad_norm_pre_clip': 21.928646087646484, 'timing/data': 0.03050547745078802, 'timing/model': 1.1034440929070115, 'learning_rate/qwen_vl_interface': 7.046118871261952e-07, 'learning_rate/action_model': 7.0461188712619514e-06, 'epoch': 2.62}) +[2026-06-18 02:28:24,363][starVLA.training.train_starvla][INFO] - Step 6354, Loss: {'train/loss': 0.11270607262849808, 'train/grad_norm_pre_clip': 18.79287338256836, 'timing/data': 0.02392367646098137, 'timing/model': 0.9245334882289171, 'learning_rate/qwen_vl_interface': 7.039844445544828e-07, 'learning_rate/action_model': 7.039844445544827e-06, 'epoch': 2.62}) +[2026-06-18 02:28:28,159][starVLA.training.train_starvla][INFO] - Step 6355, Loss: {'train/loss': 0.2598029375076294, 'train/grad_norm_pre_clip': 21.62021827697754, 'timing/data': 0.02295225765556097, 'timing/model': 0.9263410922139883, 'learning_rate/qwen_vl_interface': 7.033579443766864e-07, 'learning_rate/action_model': 7.033579443766864e-06, 'epoch': 2.62}) +[2026-06-18 02:28:31,921][starVLA.training.train_starvla][INFO] - Step 6356, Loss: {'train/loss': 0.31726720929145813, 'train/grad_norm_pre_clip': 16.352821350097656, 'timing/data': 0.030535656958818436, 'timing/model': 0.9214622676372528, 'learning_rate/qwen_vl_interface': 7.027323867226807e-07, 'learning_rate/action_model': 7.0273238672268065e-06, 'epoch': 2.62}) +[2026-06-18 02:28:35,690][starVLA.training.train_starvla][INFO] - Step 6357, Loss: {'train/loss': 0.3806457817554474, 'train/grad_norm_pre_clip': 26.638315200805664, 'timing/data': 0.028365127742290497, 'timing/model': 0.9286326989531517, 'learning_rate/qwen_vl_interface': 7.021077717221441e-07, 'learning_rate/action_model': 7.0210777172214415e-06, 'epoch': 2.62}) +[2026-06-18 02:28:39,514][starVLA.training.train_starvla][INFO] - Step 6358, Loss: {'train/loss': 0.6085841655731201, 'train/grad_norm_pre_clip': 35.32137680053711, 'timing/data': 0.02573191374540329, 'timing/model': 0.9285098193213344, 'learning_rate/qwen_vl_interface': 7.014840995045602e-07, 'learning_rate/action_model': 7.014840995045602e-06, 'epoch': 2.62}) +[2026-06-18 02:28:43,336][starVLA.training.train_starvla][INFO] - Step 6359, Loss: {'train/loss': 0.05887888744473457, 'train/grad_norm_pre_clip': 10.221372604370117, 'timing/data': 0.029628939926624298, 'timing/model': 0.9298634128645062, 'learning_rate/qwen_vl_interface': 7.008613701992172e-07, 'learning_rate/action_model': 7.008613701992171e-06, 'epoch': 2.62}) +[2026-06-18 02:28:47,170][starVLA.training.train_starvla][INFO] - Step 6360, Loss: {'train/loss': 0.2497963011264801, 'train/grad_norm_pre_clip': 27.323593139648438, 'timing/data': 0.02532642800360918, 'timing/model': 1.126831597648561, 'learning_rate/qwen_vl_interface': 7.002395839352063e-07, 'learning_rate/action_model': 7.002395839352063e-06, 'epoch': 2.62}) +[2026-06-18 02:28:50,781][starVLA.training.train_starvla][INFO] - Step 6361, Loss: {'train/loss': 0.04900810867547989, 'train/grad_norm_pre_clip': 11.550860404968262, 'timing/data': 0.026033961214125156, 'timing/model': 0.9234898667782545, 'learning_rate/qwen_vl_interface': 6.996187408414268e-07, 'learning_rate/action_model': 6.996187408414268e-06, 'epoch': 2.62}) +[2026-06-18 02:28:54,580][starVLA.training.train_starvla][INFO] - Step 6362, Loss: {'train/loss': 0.14414313435554504, 'train/grad_norm_pre_clip': 22.771812438964844, 'timing/data': 0.028144773095846176, 'timing/model': 0.9310361556708813, 'learning_rate/qwen_vl_interface': 6.98998841046578e-07, 'learning_rate/action_model': 6.989988410465779e-06, 'epoch': 2.62}) +[2026-06-18 02:28:58,387][starVLA.training.train_starvla][INFO] - Step 6363, Loss: {'train/loss': 0.2273438572883606, 'train/grad_norm_pre_clip': 27.752193450927734, 'timing/data': 0.03073969017714262, 'timing/model': 0.9301272835582495, 'learning_rate/qwen_vl_interface': 6.983798846791667e-07, 'learning_rate/action_model': 6.983798846791667e-06, 'epoch': 2.62}) +[2026-06-18 02:29:02,330][starVLA.training.train_starvla][INFO] - Step 6364, Loss: {'train/loss': 0.2789716422557831, 'train/grad_norm_pre_clip': 94.67273712158203, 'timing/data': 0.034561810083687305, 'timing/model': 0.9318024981766939, 'learning_rate/qwen_vl_interface': 6.977618718675023e-07, 'learning_rate/action_model': 6.977618718675024e-06, 'epoch': 2.62}) +[2026-06-18 02:29:06,142][starVLA.training.train_starvla][INFO] - Step 6365, Loss: {'train/loss': 0.29584312438964844, 'train/grad_norm_pre_clip': 34.922061920166016, 'timing/data': 0.026542714796960354, 'timing/model': 0.9401521245017648, 'learning_rate/qwen_vl_interface': 6.971448027397007e-07, 'learning_rate/action_model': 6.971448027397007e-06, 'epoch': 2.62}) +[2026-06-18 02:29:09,966][starVLA.training.train_starvla][INFO] - Step 6366, Loss: {'train/loss': 0.27119576930999756, 'train/grad_norm_pre_clip': 20.894800186157227, 'timing/data': 0.024069787934422493, 'timing/model': 1.1296941954642534, 'learning_rate/qwen_vl_interface': 6.965286774236814e-07, 'learning_rate/action_model': 6.965286774236813e-06, 'epoch': 2.62}) +[2026-06-18 02:29:13,774][starVLA.training.train_starvla][INFO] - Step 6367, Loss: {'train/loss': 0.2989159822463989, 'train/grad_norm_pre_clip': 20.70746421813965, 'timing/data': 0.035318031907081604, 'timing/model': 1.1075348714366555, 'learning_rate/qwen_vl_interface': 6.959134960471664e-07, 'learning_rate/action_model': 6.959134960471664e-06, 'epoch': 2.62}) +[2026-06-18 02:29:17,437][starVLA.training.train_starvla][INFO] - Step 6368, Loss: {'train/loss': 0.0919877216219902, 'train/grad_norm_pre_clip': 10.044295310974121, 'timing/data': 0.028365736827254295, 'timing/model': 0.9663368612527847, 'learning_rate/qwen_vl_interface': 6.952992587376836e-07, 'learning_rate/action_model': 6.952992587376836e-06, 'epoch': 2.62}) +[2026-06-18 02:29:21,387][starVLA.training.train_starvla][INFO] - Step 6369, Loss: {'train/loss': 0.2537638247013092, 'train/grad_norm_pre_clip': 23.42285919189453, 'timing/data': 0.03939960245043039, 'timing/model': 0.9288874100893736, 'learning_rate/qwen_vl_interface': 6.946859656225662e-07, 'learning_rate/action_model': 6.946859656225661e-06, 'epoch': 2.62}) +[2026-06-18 02:29:26,480][starVLA.training.train_starvla][INFO] - Step 6370, Loss: {'train/loss': 0.5415539741516113, 'train/grad_norm_pre_clip': 42.56366729736328, 'timing/data': 0.02719303034245968, 'timing/model': 0.9342765677720308, 'learning_rate/qwen_vl_interface': 6.940736168289497e-07, 'learning_rate/action_model': 6.940736168289497e-06, 'epoch': 2.62}) +[2026-06-18 02:29:30,223][starVLA.training.train_starvla][INFO] - Step 6371, Loss: {'train/loss': 0.29622969031333923, 'train/grad_norm_pre_clip': 23.70034408569336, 'timing/data': 0.025904349982738495, 'timing/model': 0.9454393917694688, 'learning_rate/qwen_vl_interface': 6.934622124837751e-07, 'learning_rate/action_model': 6.93462212483775e-06, 'epoch': 2.62}) +[2026-06-18 02:29:34,122][starVLA.training.train_starvla][INFO] - Step 6372, Loss: {'train/loss': 0.10181980580091476, 'train/grad_norm_pre_clip': 16.486862182617188, 'timing/data': 0.022839680314064026, 'timing/model': 0.9290012419223785, 'learning_rate/qwen_vl_interface': 6.928517527137864e-07, 'learning_rate/action_model': 6.9285175271378635e-06, 'epoch': 2.62}) +[2026-06-18 02:29:37,921][starVLA.training.train_starvla][INFO] - Step 6373, Loss: {'train/loss': 0.07550937682390213, 'train/grad_norm_pre_clip': 20.9593448638916, 'timing/data': 0.023087947629392147, 'timing/model': 1.1204468877986073, 'learning_rate/qwen_vl_interface': 6.922422376455333e-07, 'learning_rate/action_model': 6.9224223764553326e-06, 'epoch': 2.62}) +[2026-06-18 02:29:42,089][starVLA.training.train_starvla][INFO] - Step 6374, Loss: {'train/loss': 0.1288691610097885, 'train/grad_norm_pre_clip': 15.14033031463623, 'timing/data': 0.029994605109095573, 'timing/model': 1.067836930975318, 'learning_rate/qwen_vl_interface': 6.916336674053688e-07, 'learning_rate/action_model': 6.9163366740536875e-06, 'epoch': 2.62}) +[2026-06-18 02:29:45,865][starVLA.training.train_starvla][INFO] - Step 6375, Loss: {'train/loss': 0.26530003547668457, 'train/grad_norm_pre_clip': 25.049955368041992, 'timing/data': 0.028873258270323277, 'timing/model': 0.9393260590732098, 'learning_rate/qwen_vl_interface': 6.9102604211945e-07, 'learning_rate/action_model': 6.910260421194501e-06, 'epoch': 2.62}) +[2026-06-18 02:29:49,843][starVLA.training.train_starvla][INFO] - Step 6376, Loss: {'train/loss': 0.1423834264278412, 'train/grad_norm_pre_clip': 14.697630882263184, 'timing/data': 0.026725122705101967, 'timing/model': 0.9367115506902337, 'learning_rate/qwen_vl_interface': 6.904193619137381e-07, 'learning_rate/action_model': 6.904193619137381e-06, 'epoch': 2.62}) +[2026-06-18 02:29:53,644][starVLA.training.train_starvla][INFO] - Step 6377, Loss: {'train/loss': 0.3510568141937256, 'train/grad_norm_pre_clip': 27.568117141723633, 'timing/data': 0.023494308814406395, 'timing/model': 0.9345176247879863, 'learning_rate/qwen_vl_interface': 6.898136269139989e-07, 'learning_rate/action_model': 6.8981362691399885e-06, 'epoch': 2.63}) +[2026-06-18 02:29:57,454][starVLA.training.train_starvla][INFO] - Step 6378, Loss: {'train/loss': 0.3249463140964508, 'train/grad_norm_pre_clip': 26.02243995666504, 'timing/data': 0.02815578132867813, 'timing/model': 0.9570128479972482, 'learning_rate/qwen_vl_interface': 6.89208837245802e-07, 'learning_rate/action_model': 6.892088372458019e-06, 'epoch': 2.63}) +[2026-06-18 02:30:01,279][starVLA.training.train_starvla][INFO] - Step 6379, Loss: {'train/loss': 0.09846331179141998, 'train/grad_norm_pre_clip': 25.522783279418945, 'timing/data': 0.02032059896737337, 'timing/model': 0.9625092139467597, 'learning_rate/qwen_vl_interface': 6.886049930345204e-07, 'learning_rate/action_model': 6.8860499303452035e-06, 'epoch': 2.63}) +[2026-06-18 02:30:05,157][starVLA.training.train_starvla][INFO] - Step 6380, Loss: {'train/loss': 0.3216811716556549, 'train/grad_norm_pre_clip': 31.348230361938477, 'timing/data': 0.026052472181618214, 'timing/model': 1.0985564747825265, 'learning_rate/qwen_vl_interface': 6.880020944053326e-07, 'learning_rate/action_model': 6.880020944053325e-06, 'epoch': 2.63}) +[2026-06-18 02:30:08,822][starVLA.training.train_starvla][INFO] - Step 6381, Loss: {'train/loss': 0.53721684217453, 'train/grad_norm_pre_clip': 35.60065841674805, 'timing/data': 0.02432029601186514, 'timing/model': 0.9635949563235044, 'learning_rate/qwen_vl_interface': 6.87400141483219e-07, 'learning_rate/action_model': 6.87400141483219e-06, 'epoch': 2.63}) +[2026-06-18 02:30:12,841][starVLA.training.train_starvla][INFO] - Step 6382, Loss: {'train/loss': 0.12406701594591141, 'train/grad_norm_pre_clip': 17.171022415161133, 'timing/data': 0.023122576996684074, 'timing/model': 1.0558001967146993, 'learning_rate/qwen_vl_interface': 6.867991343929655e-07, 'learning_rate/action_model': 6.867991343929656e-06, 'epoch': 2.63}) +[2026-06-18 02:30:16,742][starVLA.training.train_starvla][INFO] - Step 6383, Loss: {'train/loss': 0.42697998881340027, 'train/grad_norm_pre_clip': 23.681026458740234, 'timing/data': 0.036297029815614223, 'timing/model': 0.9273305423557758, 'learning_rate/qwen_vl_interface': 6.861990732591622e-07, 'learning_rate/action_model': 6.861990732591622e-06, 'epoch': 2.63}) +[2026-06-18 02:30:20,530][starVLA.training.train_starvla][INFO] - Step 6384, Loss: {'train/loss': 0.14816202223300934, 'train/grad_norm_pre_clip': 19.510271072387695, 'timing/data': 0.028791514225304127, 'timing/model': 0.9303813567385077, 'learning_rate/qwen_vl_interface': 6.855999582062017e-07, 'learning_rate/action_model': 6.855999582062017e-06, 'epoch': 2.63}) +[2026-06-18 02:30:24,310][starVLA.training.train_starvla][INFO] - Step 6385, Loss: {'train/loss': 0.24430067837238312, 'train/grad_norm_pre_clip': 21.22911834716797, 'timing/data': 0.0324622830376029, 'timing/model': 0.9175752373412251, 'learning_rate/qwen_vl_interface': 6.850017893582818e-07, 'learning_rate/action_model': 6.850017893582818e-06, 'epoch': 2.63}) +[2026-06-18 02:30:28,282][starVLA.training.train_starvla][INFO] - Step 6386, Loss: {'train/loss': 0.33170363306999207, 'train/grad_norm_pre_clip': 24.9914608001709, 'timing/data': 0.024916470050811768, 'timing/model': 1.0877625728026032, 'learning_rate/qwen_vl_interface': 6.844045668394024e-07, 'learning_rate/action_model': 6.844045668394024e-06, 'epoch': 2.63}) +[2026-06-18 02:30:32,062][starVLA.training.train_starvla][INFO] - Step 6387, Loss: {'train/loss': 0.38898178935050964, 'train/grad_norm_pre_clip': 22.950519561767578, 'timing/data': 0.02956721279770136, 'timing/model': 1.087312982417643, 'learning_rate/qwen_vl_interface': 6.838082907733694e-07, 'learning_rate/action_model': 6.838082907733694e-06, 'epoch': 2.63}) +[2026-06-18 02:30:35,705][starVLA.training.train_starvla][INFO] - Step 6388, Loss: {'train/loss': 0.2818298935890198, 'train/grad_norm_pre_clip': 29.13721466064453, 'timing/data': 0.030417163856327534, 'timing/model': 0.9543961491435766, 'learning_rate/qwen_vl_interface': 6.832129612837912e-07, 'learning_rate/action_model': 6.832129612837912e-06, 'epoch': 2.63}) +[2026-06-18 02:30:39,614][starVLA.training.train_starvla][INFO] - Step 6389, Loss: {'train/loss': 0.38362762331962585, 'train/grad_norm_pre_clip': 29.388153076171875, 'timing/data': 0.036365194246172905, 'timing/model': 0.943310379050672, 'learning_rate/qwen_vl_interface': 6.826185784940801e-07, 'learning_rate/action_model': 6.826185784940801e-06, 'epoch': 2.63}) +[2026-06-18 02:30:43,439][starVLA.training.train_starvla][INFO] - Step 6390, Loss: {'train/loss': 0.18144813179969788, 'train/grad_norm_pre_clip': 26.70421028137207, 'timing/data': 0.027827651239931583, 'timing/model': 0.9477257486432791, 'learning_rate/qwen_vl_interface': 6.820251425274526e-07, 'learning_rate/action_model': 6.820251425274527e-06, 'epoch': 2.63}) +[2026-06-18 02:30:47,356][starVLA.training.train_starvla][INFO] - Step 6391, Loss: {'train/loss': 0.24587085843086243, 'train/grad_norm_pre_clip': 37.37468719482422, 'timing/data': 0.030099556781351566, 'timing/model': 0.9948049280792475, 'learning_rate/qwen_vl_interface': 6.81432653506928e-07, 'learning_rate/action_model': 6.814326535069279e-06, 'epoch': 2.63}) +[2026-06-18 02:30:51,154][starVLA.training.train_starvla][INFO] - Step 6392, Loss: {'train/loss': 0.07790842652320862, 'train/grad_norm_pre_clip': 11.851519584655762, 'timing/data': 0.02372425328940153, 'timing/model': 0.927358353510499, 'learning_rate/qwen_vl_interface': 6.808411115553305e-07, 'learning_rate/action_model': 6.808411115553305e-06, 'epoch': 2.63}) +[2026-06-18 02:30:55,046][starVLA.training.train_starvla][INFO] - Step 6393, Loss: {'train/loss': 0.0677015408873558, 'train/grad_norm_pre_clip': 8.870430946350098, 'timing/data': 0.027009938843548298, 'timing/model': 1.1549444701522589, 'learning_rate/qwen_vl_interface': 6.802505167952871e-07, 'learning_rate/action_model': 6.80250516795287e-06, 'epoch': 2.63}) +[2026-06-18 02:30:58,825][starVLA.training.train_starvla][INFO] - Step 6394, Loss: {'train/loss': 0.11685757339000702, 'train/grad_norm_pre_clip': 15.170235633850098, 'timing/data': 0.028626401908695698, 'timing/model': 1.0902592819184065, 'learning_rate/qwen_vl_interface': 6.796608693492293e-07, 'learning_rate/action_model': 6.7966086934922925e-06, 'epoch': 2.63}) +[2026-06-18 02:31:02,439][starVLA.training.train_starvla][INFO] - Step 6395, Loss: {'train/loss': 0.23175282776355743, 'train/grad_norm_pre_clip': 22.421308517456055, 'timing/data': 0.02756612002849579, 'timing/model': 0.9341098321601748, 'learning_rate/qwen_vl_interface': 6.790721693393912e-07, 'learning_rate/action_model': 6.790721693393912e-06, 'epoch': 2.63}) +[2026-06-18 02:31:06,239][starVLA.training.train_starvla][INFO] - Step 6396, Loss: {'train/loss': 0.2889275550842285, 'train/grad_norm_pre_clip': 25.189428329467773, 'timing/data': 0.021436425857245922, 'timing/model': 0.9333382453769445, 'learning_rate/qwen_vl_interface': 6.784844168878108e-07, 'learning_rate/action_model': 6.784844168878107e-06, 'epoch': 2.63}) +[2026-06-18 02:31:10,052][starVLA.training.train_starvla][INFO] - Step 6397, Loss: {'train/loss': 0.23624414205551147, 'train/grad_norm_pre_clip': 20.954923629760742, 'timing/data': 0.019911075942218304, 'timing/model': 0.9426141632720828, 'learning_rate/qwen_vl_interface': 6.778976121163289e-07, 'learning_rate/action_model': 6.778976121163288e-06, 'epoch': 2.63}) +[2026-06-18 02:31:13,853][starVLA.training.train_starvla][INFO] - Step 6398, Loss: {'train/loss': 0.46037933230400085, 'train/grad_norm_pre_clip': 28.948314666748047, 'timing/data': 0.02889166958630085, 'timing/model': 0.9328409265726805, 'learning_rate/qwen_vl_interface': 6.773117551465927e-07, 'learning_rate/action_model': 6.773117551465927e-06, 'epoch': 2.63}) +[2026-06-18 02:31:17,715][starVLA.training.train_starvla][INFO] - Step 6399, Loss: {'train/loss': 0.27817732095718384, 'train/grad_norm_pre_clip': 23.70603370666504, 'timing/data': 0.023834509775042534, 'timing/model': 0.9939000559970737, 'learning_rate/qwen_vl_interface': 6.767268461000498e-07, 'learning_rate/action_model': 6.7672684610004975e-06, 'epoch': 2.63}) +[2026-06-18 02:33:21,448][starVLA.training.train_starvla][INFO] - Step 6400, Loss: {'train/loss': 0.13887162506580353, 'train/grad_norm_pre_clip': 22.200454711914062, 'eval/loss': 0.41504406929016113, 'eval/latency_6/loss': 0.41504406929016113, 'eval/demon_fire_precision': 0.9588647959183674, 'eval/demon_fire_recall': 0.9718810601163542, 'eval/demon_fire_f1': 0.9653290529695023, 'eval/demon_fire_rate_model': 0.8725653867557039, 'eval/demon_fire_rate_teacher': 0.860879243183083, 'eval/demon_left_precision': 0.875, 'eval/demon_left_recall': 0.7859649122807018, 'eval/demon_left_f1': 0.8280961182994455, 'eval/demon_left_rate_model': 0.1424596549805231, 'eval/demon_left_rate_teacher': 0.15859766277128548, 'eval/demon_right_precision': 0.8220472440944881, 'eval/demon_right_recall': 0.8018433179723502, 'eval/demon_right_f1': 0.8118195956454122, 'eval/demon_right_rate_model': 0.1766833611574847, 'eval/demon_right_rate_teacher': 0.18113522537562604, 'eval/demon_move_f1': 0.8199578569724288, 'eval/demon_cc_f1': 0.9165786911925877, 'eval/latency_6/demon_fire_precision': 0.9588647959183674, 'eval/latency_6/demon_fire_recall': 0.9718810601163542, 'eval/latency_6/demon_fire_f1': 0.9653290529695023, 'eval/latency_6/demon_fire_rate_model': 0.8725653867557039, 'eval/latency_6/demon_fire_rate_teacher': 0.860879243183083, 'eval/latency_6/demon_left_precision': 0.875, 'eval/latency_6/demon_left_recall': 0.7859649122807018, 'eval/latency_6/demon_left_f1': 0.8280961182994455, 'eval/latency_6/demon_left_rate_model': 0.1424596549805231, 'eval/latency_6/demon_left_rate_teacher': 0.15859766277128548, 'eval/latency_6/demon_right_precision': 0.8220472440944881, 'eval/latency_6/demon_right_recall': 0.8018433179723502, 'eval/latency_6/demon_right_f1': 0.8118195956454122, 'eval/latency_6/demon_right_rate_model': 0.1766833611574847, 'eval/latency_6/demon_right_rate_teacher': 0.18113522537562604, 'eval/latency_6/demon_move_f1': 0.8199578569724288, 'eval/latency_6/demon_cc_f1': 0.9165786911925877, 'timing/data': 0.028051720932126045, 'timing/model': 1.1080097993835807, 'learning_rate/qwen_vl_interface': 6.761428850979542e-07, 'learning_rate/action_model': 6.7614288509795415e-06, 'epoch': 2.63}) +[2026-06-18 02:33:21,451][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_6400_state +[2026-06-18 02:33:21,456][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 02:33:28,143][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_6400_state/model.safetensors +[2026-06-18 02:33:43,407][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_6400_state/optimizer.bin +[2026-06-18 02:33:43,410][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_6400_state/sampler.bin +[2026-06-18 02:33:43,410][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_6400_state/sampler_1.bin +[2026-06-18 02:33:43,412][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_6400_state/random_states_0.pkl +[2026-06-18 02:33:43,417][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 6400 +[2026-06-18 02:33:43,417][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 02:33:43,431][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 02:33:47,490][starVLA.training.train_starvla][INFO] - Step 6401, Loss: {'train/loss': 0.35360288619995117, 'train/grad_norm_pre_clip': 49.99785232543945, 'timing/data': 0.027503417804837227, 'timing/model': 0.9747433066368103, 'learning_rate/qwen_vl_interface': 6.755598722613579e-07, 'learning_rate/action_model': 6.755598722613579e-06, 'epoch': 2.63}) +[2026-06-18 02:33:51,338][starVLA.training.train_starvla][INFO] - Step 6402, Loss: {'train/loss': 0.49895909428596497, 'train/grad_norm_pre_clip': 49.99078369140625, 'timing/data': 0.06294858362525702, 'timing/model': 0.9318295512348413, 'learning_rate/qwen_vl_interface': 6.749778077111241e-07, 'learning_rate/action_model': 6.749778077111241e-06, 'epoch': 2.64}) +[2026-06-18 02:33:55,166][starVLA.training.train_starvla][INFO] - Step 6403, Loss: {'train/loss': 0.1702704131603241, 'train/grad_norm_pre_clip': 20.057870864868164, 'timing/data': 0.02599814161658287, 'timing/model': 1.1408978719264269, 'learning_rate/qwen_vl_interface': 6.743966915679128e-07, 'learning_rate/action_model': 6.743966915679129e-06, 'epoch': 2.64}) +[2026-06-18 02:33:58,798][starVLA.training.train_starvla][INFO] - Step 6404, Loss: {'train/loss': 0.2566671073436737, 'train/grad_norm_pre_clip': 26.53080940246582, 'timing/data': 0.024238656274974346, 'timing/model': 0.9273738255724311, 'learning_rate/qwen_vl_interface': 6.738165239521909e-07, 'learning_rate/action_model': 6.7381652395219094e-06, 'epoch': 2.64}) +[2026-06-18 02:34:02,743][starVLA.training.train_starvla][INFO] - Step 6405, Loss: {'train/loss': 0.3048606216907501, 'train/grad_norm_pre_clip': 31.119840621948242, 'timing/data': 0.023049820214509964, 'timing/model': 0.932645520195365, 'learning_rate/qwen_vl_interface': 6.732373049842284e-07, 'learning_rate/action_model': 6.7323730498422846e-06, 'epoch': 2.64}) +[2026-06-18 02:34:06,532][starVLA.training.train_starvla][INFO] - Step 6406, Loss: {'train/loss': 0.1663503497838974, 'train/grad_norm_pre_clip': 18.798599243164062, 'timing/data': 0.023914665915071964, 'timing/model': 0.9287156630307436, 'learning_rate/qwen_vl_interface': 6.726590347840972e-07, 'learning_rate/action_model': 6.726590347840972e-06, 'epoch': 2.64}) +[2026-06-18 02:34:10,738][starVLA.training.train_starvla][INFO] - Step 6407, Loss: {'train/loss': 0.18612580001354218, 'train/grad_norm_pre_clip': 18.05853843688965, 'timing/data': 0.030244958586990833, 'timing/model': 0.9412356968969107, 'learning_rate/qwen_vl_interface': 6.720817134716736e-07, 'learning_rate/action_model': 6.720817134716736e-06, 'epoch': 2.64}) +[2026-06-18 02:34:14,754][starVLA.training.train_starvla][INFO] - Step 6408, Loss: {'train/loss': 0.2289818674325943, 'train/grad_norm_pre_clip': 25.758493423461914, 'timing/data': 0.02979794703423977, 'timing/model': 0.9357138834893703, 'learning_rate/qwen_vl_interface': 6.71505341166636e-07, 'learning_rate/action_model': 6.71505341166636e-06, 'epoch': 2.64}) +[2026-06-18 02:34:18,560][starVLA.training.train_starvla][INFO] - Step 6409, Loss: {'train/loss': 0.12628549337387085, 'train/grad_norm_pre_clip': 14.064216613769531, 'timing/data': 0.02945967298001051, 'timing/model': 1.0938100339844823, 'learning_rate/qwen_vl_interface': 6.709299179884689e-07, 'learning_rate/action_model': 6.709299179884689e-06, 'epoch': 2.64}) +[2026-06-18 02:34:22,329][starVLA.training.train_starvla][INFO] - Step 6410, Loss: {'train/loss': 0.14246633648872375, 'train/grad_norm_pre_clip': 16.81100845336914, 'timing/data': 0.0265547726303339, 'timing/model': 1.1049318425357342, 'learning_rate/qwen_vl_interface': 6.703554440564569e-07, 'learning_rate/action_model': 6.703554440564569e-06, 'epoch': 2.64}) +[2026-06-18 02:34:25,935][starVLA.training.train_starvla][INFO] - Step 6411, Loss: {'train/loss': 0.20735827088356018, 'train/grad_norm_pre_clip': 22.962791442871094, 'timing/data': 0.02523153182119131, 'timing/model': 0.9340553488582373, 'learning_rate/qwen_vl_interface': 6.697819194896889e-07, 'learning_rate/action_model': 6.697819194896889e-06, 'epoch': 2.64}) +[2026-06-18 02:34:29,901][starVLA.training.train_starvla][INFO] - Step 6412, Loss: {'train/loss': 0.17117953300476074, 'train/grad_norm_pre_clip': 19.72449493408203, 'timing/data': 0.023190844804048538, 'timing/model': 0.9700634609907866, 'learning_rate/qwen_vl_interface': 6.692093444070571e-07, 'learning_rate/action_model': 6.692093444070572e-06, 'epoch': 2.64}) +[2026-06-18 02:34:33,818][starVLA.training.train_starvla][INFO] - Step 6413, Loss: {'train/loss': 0.22196969389915466, 'train/grad_norm_pre_clip': 17.719711303710938, 'timing/data': 0.023038676008582115, 'timing/model': 0.9477758593857288, 'learning_rate/qwen_vl_interface': 6.686377189272584e-07, 'learning_rate/action_model': 6.686377189272583e-06, 'epoch': 2.64}) +[2026-06-18 02:34:37,780][starVLA.training.train_starvla][INFO] - Step 6414, Loss: {'train/loss': 0.45491379499435425, 'train/grad_norm_pre_clip': 25.457246780395508, 'timing/data': 0.032731784507632256, 'timing/model': 1.0975591382011771, 'learning_rate/qwen_vl_interface': 6.680670431687898e-07, 'learning_rate/action_model': 6.680670431687898e-06, 'epoch': 2.64}) +[2026-06-18 02:34:41,645][starVLA.training.train_starvla][INFO] - Step 6415, Loss: {'train/loss': 0.3373378813266754, 'train/grad_norm_pre_clip': 18.381969451904297, 'timing/data': 0.02359289675951004, 'timing/model': 0.927112408913672, 'learning_rate/qwen_vl_interface': 6.674973172499536e-07, 'learning_rate/action_model': 6.674973172499535e-06, 'epoch': 2.64}) +[2026-06-18 02:34:45,474][starVLA.training.train_starvla][INFO] - Step 6416, Loss: {'train/loss': 0.24510328471660614, 'train/grad_norm_pre_clip': 23.767805099487305, 'timing/data': 0.02911985758692026, 'timing/model': 1.1001171441748738, 'learning_rate/qwen_vl_interface': 6.669285412888544e-07, 'learning_rate/action_model': 6.669285412888544e-06, 'epoch': 2.64}) +[2026-06-18 02:34:49,358][starVLA.training.train_starvla][INFO] - Step 6417, Loss: {'train/loss': 0.16466473042964935, 'train/grad_norm_pre_clip': 17.116342544555664, 'timing/data': 0.03302685730159283, 'timing/model': 1.1743013113737106, 'learning_rate/qwen_vl_interface': 6.663607154034003e-07, 'learning_rate/action_model': 6.663607154034003e-06, 'epoch': 2.64}) +[2026-06-18 02:34:53,094][starVLA.training.train_starvla][INFO] - Step 6418, Loss: {'train/loss': 0.2968069911003113, 'train/grad_norm_pre_clip': 25.95533561706543, 'timing/data': 0.026197285391390324, 'timing/model': 0.9475070312619209, 'learning_rate/qwen_vl_interface': 6.657938397113019e-07, 'learning_rate/action_model': 6.657938397113018e-06, 'epoch': 2.64}) +[2026-06-18 02:34:56,871][starVLA.training.train_starvla][INFO] - Step 6419, Loss: {'train/loss': 0.17605474591255188, 'train/grad_norm_pre_clip': 21.528989791870117, 'timing/data': 0.02602983172982931, 'timing/model': 0.9222879754379392, 'learning_rate/qwen_vl_interface': 6.652279143300734e-07, 'learning_rate/action_model': 6.652279143300734e-06, 'epoch': 2.64}) +[2026-06-18 02:35:00,703][starVLA.training.train_starvla][INFO] - Step 6420, Loss: {'train/loss': 0.055834244936704636, 'train/grad_norm_pre_clip': 10.69500732421875, 'timing/data': 0.03801098093390465, 'timing/model': 0.9381456803530455, 'learning_rate/qwen_vl_interface': 6.64662939377032e-07, 'learning_rate/action_model': 6.646629393770321e-06, 'epoch': 2.64}) +[2026-06-18 02:35:04,600][starVLA.training.train_starvla][INFO] - Step 6421, Loss: {'train/loss': 0.2800779342651367, 'train/grad_norm_pre_clip': 23.31509017944336, 'timing/data': 0.026999786496162415, 'timing/model': 0.986458033323288, 'learning_rate/qwen_vl_interface': 6.64098914969297e-07, 'learning_rate/action_model': 6.6409891496929695e-06, 'epoch': 2.64}) +[2026-06-18 02:35:08,559][starVLA.training.train_starvla][INFO] - Step 6422, Loss: {'train/loss': 0.3369191288948059, 'train/grad_norm_pre_clip': 20.542278289794922, 'timing/data': 0.027233845554292202, 'timing/model': 0.9690735628828406, 'learning_rate/qwen_vl_interface': 6.635358412237912e-07, 'learning_rate/action_model': 6.635358412237912e-06, 'epoch': 2.64}) +[2026-06-18 02:35:12,570][starVLA.training.train_starvla][INFO] - Step 6423, Loss: {'train/loss': 0.13345244526863098, 'train/grad_norm_pre_clip': 18.04114532470703, 'timing/data': 0.026278948411345482, 'timing/model': 1.1355544095858932, 'learning_rate/qwen_vl_interface': 6.629737182572418e-07, 'learning_rate/action_model': 6.629737182572418e-06, 'epoch': 2.64}) +[2026-06-18 02:35:16,559][starVLA.training.train_starvla][INFO] - Step 6424, Loss: {'train/loss': 0.08163898438215256, 'train/grad_norm_pre_clip': 12.137442588806152, 'timing/data': 0.026177569292485714, 'timing/model': 1.1666940031573176, 'learning_rate/qwen_vl_interface': 6.62412546186176e-07, 'learning_rate/action_model': 6.62412546186176e-06, 'epoch': 2.64}) +[2026-06-18 02:35:20,304][starVLA.training.train_starvla][INFO] - Step 6425, Loss: {'train/loss': 0.17245961725711823, 'train/grad_norm_pre_clip': 17.680940628051758, 'timing/data': 0.03313530329614878, 'timing/model': 0.9355335161089897, 'learning_rate/qwen_vl_interface': 6.61852325126926e-07, 'learning_rate/action_model': 6.61852325126926e-06, 'epoch': 2.64}) +[2026-06-18 02:35:24,923][starVLA.training.train_starvla][INFO] - Step 6426, Loss: {'train/loss': 0.05350678414106369, 'train/grad_norm_pre_clip': 7.7885260581970215, 'timing/data': 0.026691759005188942, 'timing/model': 1.6777974124997854, 'learning_rate/qwen_vl_interface': 6.61293055195625e-07, 'learning_rate/action_model': 6.6129305519562504e-06, 'epoch': 2.65}) +[2026-06-18 02:35:28,815][starVLA.training.train_starvla][INFO] - Step 6427, Loss: {'train/loss': 0.17177565395832062, 'train/grad_norm_pre_clip': 20.961910247802734, 'timing/data': 0.023589862510561943, 'timing/model': 0.9561241883784533, 'learning_rate/qwen_vl_interface': 6.607347365082127e-07, 'learning_rate/action_model': 6.607347365082127e-06, 'epoch': 2.65}) +[2026-06-18 02:35:32,754][starVLA.training.train_starvla][INFO] - Step 6428, Loss: {'train/loss': 0.10528291761875153, 'train/grad_norm_pre_clip': 21.063566207885742, 'timing/data': 0.027942231856286526, 'timing/model': 0.9376160493120551, 'learning_rate/qwen_vl_interface': 6.601773691804273e-07, 'learning_rate/action_model': 6.6017736918042735e-06, 'epoch': 2.65}) +[2026-06-18 02:35:36,781][starVLA.training.train_starvla][INFO] - Step 6429, Loss: {'train/loss': 0.23730820417404175, 'train/grad_norm_pre_clip': 27.474943161010742, 'timing/data': 0.025004804134368896, 'timing/model': 0.9965191511437297, 'learning_rate/qwen_vl_interface': 6.596209533278122e-07, 'learning_rate/action_model': 6.596209533278122e-06, 'epoch': 2.65}) +[2026-06-18 02:35:40,657][starVLA.training.train_starvla][INFO] - Step 6430, Loss: {'train/loss': 0.38267725706100464, 'train/grad_norm_pre_clip': 29.173641204833984, 'timing/data': 0.028600391000509262, 'timing/model': 1.1139967134222388, 'learning_rate/qwen_vl_interface': 6.590654890657125e-07, 'learning_rate/action_model': 6.590654890657125e-06, 'epoch': 2.65}) +[2026-06-18 02:35:44,377][starVLA.training.train_starvla][INFO] - Step 6431, Loss: {'train/loss': 0.14563459157943726, 'train/grad_norm_pre_clip': 15.323533058166504, 'timing/data': 0.028567065484821796, 'timing/model': 0.9752146312966943, 'learning_rate/qwen_vl_interface': 6.585109765092772e-07, 'learning_rate/action_model': 6.585109765092772e-06, 'epoch': 2.65}) +[2026-06-18 02:35:48,796][starVLA.training.train_starvla][INFO] - Step 6432, Loss: {'train/loss': 0.14979802072048187, 'train/grad_norm_pre_clip': 16.248546600341797, 'timing/data': 0.02621354442089796, 'timing/model': 0.9380022203549743, 'learning_rate/qwen_vl_interface': 6.579574157734575e-07, 'learning_rate/action_model': 6.579574157734575e-06, 'epoch': 2.65}) +[2026-06-18 02:35:52,688][starVLA.training.train_starvla][INFO] - Step 6433, Loss: {'train/loss': 0.10822466015815735, 'train/grad_norm_pre_clip': 11.759658813476562, 'timing/data': 0.024665435776114464, 'timing/model': 1.0008118711411953, 'learning_rate/qwen_vl_interface': 6.574048069730062e-07, 'learning_rate/action_model': 6.574048069730063e-06, 'epoch': 2.65}) +[2026-06-18 02:35:56,827][starVLA.training.train_starvla][INFO] - Step 6434, Loss: {'train/loss': 0.26029887795448303, 'train/grad_norm_pre_clip': 29.427331924438477, 'timing/data': 0.03445988707244396, 'timing/model': 0.9581091422587633, 'learning_rate/qwen_vl_interface': 6.568531502224808e-07, 'learning_rate/action_model': 6.568531502224808e-06, 'epoch': 2.65}) +[2026-06-18 02:36:00,631][starVLA.training.train_starvla][INFO] - Step 6435, Loss: {'train/loss': 0.11816337704658508, 'train/grad_norm_pre_clip': 11.670855522155762, 'timing/data': 0.03199942223727703, 'timing/model': 0.9359882976859808, 'learning_rate/qwen_vl_interface': 6.563024456362395e-07, 'learning_rate/action_model': 6.563024456362395e-06, 'epoch': 2.65}) +[2026-06-18 02:36:04,469][starVLA.training.train_starvla][INFO] - Step 6436, Loss: {'train/loss': 0.21363723278045654, 'train/grad_norm_pre_clip': 30.94194793701172, 'timing/data': 0.023480684496462345, 'timing/model': 0.9366130344569683, 'learning_rate/qwen_vl_interface': 6.557526933284443e-07, 'learning_rate/action_model': 6.557526933284443e-06, 'epoch': 2.65}) +[2026-06-18 02:36:08,287][starVLA.training.train_starvla][INFO] - Step 6437, Loss: {'train/loss': 0.19012047350406647, 'train/grad_norm_pre_clip': 22.12481117248535, 'timing/data': 0.033385652117431164, 'timing/model': 1.1481250990182161, 'learning_rate/qwen_vl_interface': 6.552038934130595e-07, 'learning_rate/action_model': 6.552038934130594e-06, 'epoch': 2.65}) +[2026-06-18 02:36:11,905][starVLA.training.train_starvla][INFO] - Step 6438, Loss: {'train/loss': 0.09764792770147324, 'train/grad_norm_pre_clip': 14.632408142089844, 'timing/data': 0.026374852284789085, 'timing/model': 0.9380764178931713, 'learning_rate/qwen_vl_interface': 6.546560460038514e-07, 'learning_rate/action_model': 6.546560460038514e-06, 'epoch': 2.65}) +[2026-06-18 02:36:15,722][starVLA.training.train_starvla][INFO] - Step 6439, Loss: {'train/loss': 0.26380524039268494, 'train/grad_norm_pre_clip': 26.12213134765625, 'timing/data': 0.04587210062891245, 'timing/model': 0.94603574834764, 'learning_rate/qwen_vl_interface': 6.5410915121439e-07, 'learning_rate/action_model': 6.5410915121439e-06, 'epoch': 2.65}) +[2026-06-18 02:36:19,566][starVLA.training.train_starvla][INFO] - Step 6440, Loss: {'train/loss': 0.22435706853866577, 'train/grad_norm_pre_clip': 29.42584800720215, 'timing/data': 0.02315570693463087, 'timing/model': 0.9310587393119931, 'learning_rate/qwen_vl_interface': 6.535632091580474e-07, 'learning_rate/action_model': 6.535632091580474e-06, 'epoch': 2.65}) +[2026-06-18 02:36:23,498][starVLA.training.train_starvla][INFO] - Step 6441, Loss: {'train/loss': 0.160194531083107, 'train/grad_norm_pre_clip': 18.13250732421875, 'timing/data': 0.0332498624920845, 'timing/model': 0.9486618284136057, 'learning_rate/qwen_vl_interface': 6.530182199479965e-07, 'learning_rate/action_model': 6.5301821994799645e-06, 'epoch': 2.65}) +[2026-06-18 02:36:27,370][starVLA.training.train_starvla][INFO] - Step 6442, Loss: {'train/loss': 0.12266561388969421, 'train/grad_norm_pre_clip': 16.442331314086914, 'timing/data': 0.029588395729660988, 'timing/model': 1.0367443887516856, 'learning_rate/qwen_vl_interface': 6.524741836972162e-07, 'learning_rate/action_model': 6.524741836972162e-06, 'epoch': 2.65}) +[2026-06-18 02:36:31,447][starVLA.training.train_starvla][INFO] - Step 6443, Loss: {'train/loss': 0.19328589737415314, 'train/grad_norm_pre_clip': 25.765634536743164, 'timing/data': 0.03148859366774559, 'timing/model': 1.156791820190847, 'learning_rate/qwen_vl_interface': 6.519311005184848e-07, 'learning_rate/action_model': 6.5193110051848476e-06, 'epoch': 2.65}) +[2026-06-18 02:36:35,218][starVLA.training.train_starvla][INFO] - Step 6444, Loss: {'train/loss': 0.3371129333972931, 'train/grad_norm_pre_clip': 24.338499069213867, 'timing/data': 0.022627527825534344, 'timing/model': 1.1054347679018974, 'learning_rate/qwen_vl_interface': 6.51388970524383e-07, 'learning_rate/action_model': 6.513889705243829e-06, 'epoch': 2.65}) +[2026-06-18 02:36:38,938][starVLA.training.train_starvla][INFO] - Step 6445, Loss: {'train/loss': 0.25973567366600037, 'train/grad_norm_pre_clip': 29.68569564819336, 'timing/data': 0.026173151098191738, 'timing/model': 0.9358889926224947, 'learning_rate/qwen_vl_interface': 6.508477938272963e-07, 'learning_rate/action_model': 6.508477938272962e-06, 'epoch': 2.65}) +[2026-06-18 02:36:42,743][starVLA.training.train_starvla][INFO] - Step 6446, Loss: {'train/loss': 0.14056850969791412, 'train/grad_norm_pre_clip': 24.361637115478516, 'timing/data': 0.030517171137034893, 'timing/model': 0.9214854771271348, 'learning_rate/qwen_vl_interface': 6.503075705394113e-07, 'learning_rate/action_model': 6.5030757053941125e-06, 'epoch': 2.65}) +[2026-06-18 02:36:46,495][starVLA.training.train_starvla][INFO] - Step 6447, Loss: {'train/loss': 0.17193518579006195, 'train/grad_norm_pre_clip': 17.31550407409668, 'timing/data': 0.03045818116515875, 'timing/model': 0.9293457325547934, 'learning_rate/qwen_vl_interface': 6.497683007727166e-07, 'learning_rate/action_model': 6.497683007727166e-06, 'epoch': 2.65}) +[2026-06-18 02:36:50,323][starVLA.training.train_starvla][INFO] - Step 6448, Loss: {'train/loss': 0.16488610208034515, 'train/grad_norm_pre_clip': 18.275840759277344, 'timing/data': 0.031139124184846878, 'timing/model': 0.9439117591828108, 'learning_rate/qwen_vl_interface': 6.492299846390024e-07, 'learning_rate/action_model': 6.492299846390023e-06, 'epoch': 2.65}) +[2026-06-18 02:36:54,179][starVLA.training.train_starvla][INFO] - Step 6449, Loss: {'train/loss': 0.22129026055335999, 'train/grad_norm_pre_clip': 19.252302169799805, 'timing/data': 0.03664121404290199, 'timing/model': 0.9425098039209843, 'learning_rate/qwen_vl_interface': 6.486926222498633e-07, 'learning_rate/action_model': 6.486926222498633e-06, 'epoch': 2.65}) +[2026-06-18 02:36:59,113][starVLA.training.train_starvla][INFO] - Step 6450, Loss: {'train/loss': 0.2917337119579315, 'train/grad_norm_pre_clip': 21.1396541595459, 'timing/data': 0.031969718635082245, 'timing/model': 1.1153278881683946, 'learning_rate/qwen_vl_interface': 6.481562137166952e-07, 'learning_rate/action_model': 6.481562137166952e-06, 'epoch': 2.66}) +[2026-06-18 02:37:03,325][starVLA.training.train_starvla][INFO] - Step 6451, Loss: {'train/loss': 0.5810045003890991, 'train/grad_norm_pre_clip': 35.79084014892578, 'timing/data': 0.022569192573428154, 'timing/model': 1.1426345882937312, 'learning_rate/qwen_vl_interface': 6.476207591506954e-07, 'learning_rate/action_model': 6.476207591506954e-06, 'epoch': 2.66}) +[2026-06-18 02:37:07,280][starVLA.training.train_starvla][INFO] - Step 6452, Loss: {'train/loss': 0.10506150126457214, 'train/grad_norm_pre_clip': 14.358011245727539, 'timing/data': 0.026425150223076344, 'timing/model': 0.9739161590114236, 'learning_rate/qwen_vl_interface': 6.470862586628645e-07, 'learning_rate/action_model': 6.470862586628646e-06, 'epoch': 2.66}) +[2026-06-18 02:37:11,278][starVLA.training.train_starvla][INFO] - Step 6453, Loss: {'train/loss': 0.05799919739365578, 'train/grad_norm_pre_clip': 12.255207061767578, 'timing/data': 0.02440018765628338, 'timing/model': 1.0343605121597648, 'learning_rate/qwen_vl_interface': 6.465527123640056e-07, 'learning_rate/action_model': 6.465527123640056e-06, 'epoch': 2.66}) +[2026-06-18 02:37:15,746][starVLA.training.train_starvla][INFO] - Step 6454, Loss: {'train/loss': 0.3460538387298584, 'train/grad_norm_pre_clip': 24.655858993530273, 'timing/data': 0.02865634486079216, 'timing/model': 0.9376590298488736, 'learning_rate/qwen_vl_interface': 6.460201203647229e-07, 'learning_rate/action_model': 6.460201203647229e-06, 'epoch': 2.66}) +[2026-06-18 02:37:19,565][starVLA.training.train_starvla][INFO] - Step 6455, Loss: {'train/loss': 0.02665538713335991, 'train/grad_norm_pre_clip': 4.010086536407471, 'timing/data': 0.027148699387907982, 'timing/model': 0.9993585040792823, 'learning_rate/qwen_vl_interface': 6.454884827754224e-07, 'learning_rate/action_model': 6.4548848277542235e-06, 'epoch': 2.66}) +[2026-06-18 02:37:23,510][starVLA.training.train_starvla][INFO] - Step 6456, Loss: {'train/loss': 0.5462679266929626, 'train/grad_norm_pre_clip': 33.467159271240234, 'timing/data': 0.022503981366753578, 'timing/model': 0.9730862500146031, 'learning_rate/qwen_vl_interface': 6.449577997063156e-07, 'learning_rate/action_model': 6.449577997063155e-06, 'epoch': 2.66}) +[2026-06-18 02:37:27,755][starVLA.training.train_starvla][INFO] - Step 6457, Loss: {'train/loss': 0.21931064128875732, 'train/grad_norm_pre_clip': 21.555500030517578, 'timing/data': 0.026925851590931416, 'timing/model': 1.5023446353152394, 'learning_rate/qwen_vl_interface': 6.444280712674112e-07, 'learning_rate/action_model': 6.444280712674111e-06, 'epoch': 2.66}) +[2026-06-18 02:37:31,371][starVLA.training.train_starvla][INFO] - Step 6458, Loss: {'train/loss': 0.11657402664422989, 'train/grad_norm_pre_clip': 13.04385757446289, 'timing/data': 0.02878235001116991, 'timing/model': 0.9306959928944707, 'learning_rate/qwen_vl_interface': 6.438992975685237e-07, 'learning_rate/action_model': 6.438992975685237e-06, 'epoch': 2.66}) +[2026-06-18 02:37:35,183][starVLA.training.train_starvla][INFO] - Step 6459, Loss: {'train/loss': 0.2347753793001175, 'train/grad_norm_pre_clip': 29.905738830566406, 'timing/data': 0.028951875865459442, 'timing/model': 0.9262247430160642, 'learning_rate/qwen_vl_interface': 6.433714787192677e-07, 'learning_rate/action_model': 6.4337147871926766e-06, 'epoch': 2.66}) +[2026-06-18 02:37:39,056][starVLA.training.train_starvla][INFO] - Step 6460, Loss: {'train/loss': 0.07383710891008377, 'train/grad_norm_pre_clip': 11.531010627746582, 'timing/data': 0.027175319381058216, 'timing/model': 0.9890298098325729, 'learning_rate/qwen_vl_interface': 6.428446148290615e-07, 'learning_rate/action_model': 6.428446148290615e-06, 'epoch': 2.66}) +[2026-06-18 02:37:42,945][starVLA.training.train_starvla][INFO] - Step 6461, Loss: {'train/loss': 0.10122812539339066, 'train/grad_norm_pre_clip': 16.3652286529541, 'timing/data': 0.02298036776483059, 'timing/model': 0.9393406948074698, 'learning_rate/qwen_vl_interface': 6.423187060071248e-07, 'learning_rate/action_model': 6.423187060071248e-06, 'epoch': 2.66}) +[2026-06-18 02:37:46,860][starVLA.training.train_starvla][INFO] - Step 6462, Loss: {'train/loss': 0.47253140807151794, 'train/grad_norm_pre_clip': 46.69554901123047, 'timing/data': 0.021244263276457787, 'timing/model': 0.94928118865937, 'learning_rate/qwen_vl_interface': 6.417937523624778e-07, 'learning_rate/action_model': 6.417937523624778e-06, 'epoch': 2.66}) +[2026-06-18 02:37:50,897][starVLA.training.train_starvla][INFO] - Step 6463, Loss: {'train/loss': 0.05841031298041344, 'train/grad_norm_pre_clip': 8.617966651916504, 'timing/data': 0.025119825266301632, 'timing/model': 1.218754610978067, 'learning_rate/qwen_vl_interface': 6.412697540039442e-07, 'learning_rate/action_model': 6.412697540039442e-06, 'epoch': 2.66}) +[2026-06-18 02:37:54,970][starVLA.training.train_starvla][INFO] - Step 6464, Loss: {'train/loss': 0.09427915513515472, 'train/grad_norm_pre_clip': 12.569806098937988, 'timing/data': 0.03799245785921812, 'timing/model': 1.1321852393448353, 'learning_rate/qwen_vl_interface': 6.407467110401505e-07, 'learning_rate/action_model': 6.407467110401505e-06, 'epoch': 2.66}) +[2026-06-18 02:37:58,657][starVLA.training.train_starvla][INFO] - Step 6465, Loss: {'train/loss': 0.03318110108375549, 'train/grad_norm_pre_clip': 5.320865631103516, 'timing/data': 0.023854424245655537, 'timing/model': 0.9644432328641415, 'learning_rate/qwen_vl_interface': 6.402246235795228e-07, 'learning_rate/action_model': 6.402246235795227e-06, 'epoch': 2.66}) +[2026-06-18 02:38:02,491][starVLA.training.train_starvla][INFO] - Step 6466, Loss: {'train/loss': 0.1843309998512268, 'train/grad_norm_pre_clip': 23.377513885498047, 'timing/data': 0.022732886485755444, 'timing/model': 0.9371173204854131, 'learning_rate/qwen_vl_interface': 6.397034917302913e-07, 'learning_rate/action_model': 6.397034917302913e-06, 'epoch': 2.66}) +[2026-06-18 02:38:06,256][starVLA.training.train_starvla][INFO] - Step 6467, Loss: {'train/loss': 0.18429993093013763, 'train/grad_norm_pre_clip': 26.40433120727539, 'timing/data': 0.023232316598296165, 'timing/model': 0.9259289493784308, 'learning_rate/qwen_vl_interface': 6.391833156004871e-07, 'learning_rate/action_model': 6.39183315600487e-06, 'epoch': 2.66}) +[2026-06-18 02:38:09,999][starVLA.training.train_starvla][INFO] - Step 6468, Loss: {'train/loss': 0.23844563961029053, 'train/grad_norm_pre_clip': 15.582320213317871, 'timing/data': 0.026574671268463135, 'timing/model': 0.928280551917851, 'learning_rate/qwen_vl_interface': 6.386640952979427e-07, 'learning_rate/action_model': 6.386640952979426e-06, 'epoch': 2.66}) +[2026-06-18 02:38:13,809][starVLA.training.train_starvla][INFO] - Step 6469, Loss: {'train/loss': 0.2533954679965973, 'train/grad_norm_pre_clip': 50.80048370361328, 'timing/data': 0.024921122007071972, 'timing/model': 0.9314390141516924, 'learning_rate/qwen_vl_interface': 6.381458309302937e-07, 'learning_rate/action_model': 6.381458309302937e-06, 'epoch': 2.66}) +[2026-06-18 02:38:17,547][starVLA.training.train_starvla][INFO] - Step 6470, Loss: {'train/loss': 0.25051775574684143, 'train/grad_norm_pre_clip': 20.62150764465332, 'timing/data': 0.025880074128508568, 'timing/model': 1.0856759138405323, 'learning_rate/qwen_vl_interface': 6.376285226049762e-07, 'learning_rate/action_model': 6.376285226049762e-06, 'epoch': 2.66}) +[2026-06-18 02:38:21,334][starVLA.training.train_starvla][INFO] - Step 6471, Loss: {'train/loss': 0.057329948991537094, 'train/grad_norm_pre_clip': 6.73486328125, 'timing/data': 0.02601795457303524, 'timing/model': 1.0989732565358281, 'learning_rate/qwen_vl_interface': 6.371121704292292e-07, 'learning_rate/action_model': 6.371121704292292e-06, 'epoch': 2.66}) +[2026-06-18 02:38:24,919][starVLA.training.train_starvla][INFO] - Step 6472, Loss: {'train/loss': 0.26240405440330505, 'train/grad_norm_pre_clip': 20.949678421020508, 'timing/data': 0.024841267615556717, 'timing/model': 0.9191438835114241, 'learning_rate/qwen_vl_interface': 6.365967745100926e-07, 'learning_rate/action_model': 6.3659677451009256e-06, 'epoch': 2.66}) +[2026-06-18 02:38:28,671][starVLA.training.train_starvla][INFO] - Step 6473, Loss: {'train/loss': 0.21473459899425507, 'train/grad_norm_pre_clip': 19.644920349121094, 'timing/data': 0.02565211057662964, 'timing/model': 0.9203720577061176, 'learning_rate/qwen_vl_interface': 6.360823349544092e-07, 'learning_rate/action_model': 6.360823349544092e-06, 'epoch': 2.66}) +[2026-06-18 02:38:32,522][starVLA.training.train_starvla][INFO] - Step 6474, Loss: {'train/loss': 0.12914294004440308, 'train/grad_norm_pre_clip': 15.677816390991211, 'timing/data': 0.04847572464495897, 'timing/model': 1.008998659439385, 'learning_rate/qwen_vl_interface': 6.355688518688222e-07, 'learning_rate/action_model': 6.355688518688221e-06, 'epoch': 2.66}) +[2026-06-18 02:38:36,248][starVLA.training.train_starvla][INFO] - Step 6475, Loss: {'train/loss': 0.48445457220077515, 'train/grad_norm_pre_clip': 22.593307495117188, 'timing/data': 0.023104630410671234, 'timing/model': 0.9210573965683579, 'learning_rate/qwen_vl_interface': 6.350563253597781e-07, 'learning_rate/action_model': 6.350563253597781e-06, 'epoch': 2.67}) +[2026-06-18 02:38:40,052][starVLA.training.train_starvla][INFO] - Step 6476, Loss: {'train/loss': 0.2147664725780487, 'train/grad_norm_pre_clip': 21.05459213256836, 'timing/data': 0.023300585336983204, 'timing/model': 0.9219064759090543, 'learning_rate/qwen_vl_interface': 6.345447555335236e-07, 'learning_rate/action_model': 6.3454475553352345e-06, 'epoch': 2.67}) +[2026-06-18 02:38:43,801][starVLA.training.train_starvla][INFO] - Step 6477, Loss: {'train/loss': 0.3312637507915497, 'train/grad_norm_pre_clip': 30.746971130371094, 'timing/data': 0.022748553194105625, 'timing/model': 1.089615685865283, 'learning_rate/qwen_vl_interface': 6.340341424961072e-07, 'learning_rate/action_model': 6.340341424961071e-06, 'epoch': 2.67}) +[2026-06-18 02:38:47,393][starVLA.training.train_starvla][INFO] - Step 6478, Loss: {'train/loss': 0.39680996537208557, 'train/grad_norm_pre_clip': 27.132423400878906, 'timing/data': 0.024887466803193092, 'timing/model': 0.9266340900212526, 'learning_rate/qwen_vl_interface': 6.335244863533802e-07, 'learning_rate/action_model': 6.3352448635338016e-06, 'epoch': 2.67}) +[2026-06-18 02:38:51,134][starVLA.training.train_starvla][INFO] - Step 6479, Loss: {'train/loss': 0.13487324118614197, 'train/grad_norm_pre_clip': 12.584023475646973, 'timing/data': 0.02789365593343973, 'timing/model': 0.926027949899435, 'learning_rate/qwen_vl_interface': 6.330157872109953e-07, 'learning_rate/action_model': 6.330157872109953e-06, 'epoch': 2.67}) +[2026-06-18 02:38:55,041][starVLA.training.train_starvla][INFO] - Step 6480, Loss: {'train/loss': 0.2904353141784668, 'train/grad_norm_pre_clip': 23.806108474731445, 'timing/data': 0.030526496469974518, 'timing/model': 0.9394826041534543, 'learning_rate/qwen_vl_interface': 6.325080451744053e-07, 'learning_rate/action_model': 6.325080451744053e-06, 'epoch': 2.67}) +[2026-06-18 02:38:59,097][starVLA.training.train_starvla][INFO] - Step 6481, Loss: {'train/loss': 0.11817418038845062, 'train/grad_norm_pre_clip': 18.361818313598633, 'timing/data': 0.021996742114424706, 'timing/model': 0.940319444052875, 'learning_rate/qwen_vl_interface': 6.320012603488663e-07, 'learning_rate/action_model': 6.320012603488662e-06, 'epoch': 2.67}) +[2026-06-18 02:39:02,971][starVLA.training.train_starvla][INFO] - Step 6482, Loss: {'train/loss': 0.12697087228298187, 'train/grad_norm_pre_clip': 19.260326385498047, 'timing/data': 0.032114082016050816, 'timing/model': 1.0415963977575302, 'learning_rate/qwen_vl_interface': 6.314954328394351e-07, 'learning_rate/action_model': 6.3149543283943515e-06, 'epoch': 2.67}) +[2026-06-18 02:39:08,205][starVLA.training.train_starvla][INFO] - Step 6483, Loss: {'train/loss': 0.2019912302494049, 'train/grad_norm_pre_clip': 22.996440887451172, 'timing/data': 0.03331667836755514, 'timing/model': 0.9756601816043258, 'learning_rate/qwen_vl_interface': 6.309905627509707e-07, 'learning_rate/action_model': 6.309905627509706e-06, 'epoch': 2.67}) +[2026-06-18 02:39:12,363][starVLA.training.train_starvla][INFO] - Step 6484, Loss: {'train/loss': 0.33432766795158386, 'train/grad_norm_pre_clip': 35.476993560791016, 'timing/data': 0.02469109743833542, 'timing/model': 1.2841107305139303, 'learning_rate/qwen_vl_interface': 6.304866501881327e-07, 'learning_rate/action_model': 6.304866501881327e-06, 'epoch': 2.67}) +[2026-06-18 02:39:16,191][starVLA.training.train_starvla][INFO] - Step 6485, Loss: {'train/loss': 0.08792103081941605, 'train/grad_norm_pre_clip': 12.269376754760742, 'timing/data': 0.025387680158019066, 'timing/model': 0.9683062145486474, 'learning_rate/qwen_vl_interface': 6.299836952553831e-07, 'learning_rate/action_model': 6.299836952553831e-06, 'epoch': 2.67}) +[2026-06-18 02:39:20,105][starVLA.training.train_starvla][INFO] - Step 6486, Loss: {'train/loss': 0.3557318449020386, 'train/grad_norm_pre_clip': 24.614274978637695, 'timing/data': 0.03939014486968517, 'timing/model': 0.9494793154299259, 'learning_rate/qwen_vl_interface': 6.294816980569845e-07, 'learning_rate/action_model': 6.294816980569845e-06, 'epoch': 2.67}) +[2026-06-18 02:39:23,938][starVLA.training.train_starvla][INFO] - Step 6487, Loss: {'train/loss': 0.21686531603336334, 'train/grad_norm_pre_clip': 26.49107551574707, 'timing/data': 0.025824910029768944, 'timing/model': 0.9387805834412575, 'learning_rate/qwen_vl_interface': 6.289806586970021e-07, 'learning_rate/action_model': 6.289806586970021e-06, 'epoch': 2.67}) +[2026-06-18 02:39:27,875][starVLA.training.train_starvla][INFO] - Step 6488, Loss: {'train/loss': 0.1473337709903717, 'train/grad_norm_pre_clip': 16.35830307006836, 'timing/data': 0.03184133768081665, 'timing/model': 0.9890219699591398, 'learning_rate/qwen_vl_interface': 6.284805772793011e-07, 'learning_rate/action_model': 6.284805772793011e-06, 'epoch': 2.67}) +[2026-06-18 02:39:31,758][starVLA.training.train_starvla][INFO] - Step 6489, Loss: {'train/loss': 0.3177265524864197, 'train/grad_norm_pre_clip': 21.444904327392578, 'timing/data': 0.02217945083975792, 'timing/model': 0.9448149185627699, 'learning_rate/qwen_vl_interface': 6.279814539075499e-07, 'learning_rate/action_model': 6.2798145390754985e-06, 'epoch': 2.67}) +[2026-06-18 02:39:35,594][starVLA.training.train_starvla][INFO] - Step 6490, Loss: {'train/loss': 0.20898285508155823, 'train/grad_norm_pre_clip': 19.649362564086914, 'timing/data': 0.04218406789004803, 'timing/model': 1.1140314554795623, 'learning_rate/qwen_vl_interface': 6.274832886852172e-07, 'learning_rate/action_model': 6.2748328868521724e-06, 'epoch': 2.67}) +[2026-06-18 02:39:39,432][starVLA.training.train_starvla][INFO] - Step 6491, Loss: {'train/loss': 0.27085283398628235, 'train/grad_norm_pre_clip': 22.70564842224121, 'timing/data': 0.031724466010928154, 'timing/model': 1.140916993841529, 'learning_rate/qwen_vl_interface': 6.269860817155721e-07, 'learning_rate/action_model': 6.269860817155721e-06, 'epoch': 2.67}) +[2026-06-18 02:39:43,266][starVLA.training.train_starvla][INFO] - Step 6492, Loss: {'train/loss': 0.20084865391254425, 'train/grad_norm_pre_clip': 21.5546817779541, 'timing/data': 0.024654516018927097, 'timing/model': 0.934912926517427, 'learning_rate/qwen_vl_interface': 6.264898331016865e-07, 'learning_rate/action_model': 6.264898331016865e-06, 'epoch': 2.67}) +[2026-06-18 02:39:47,194][starVLA.training.train_starvla][INFO] - Step 6493, Loss: {'train/loss': 0.2685442566871643, 'train/grad_norm_pre_clip': 19.41765022277832, 'timing/data': 0.027263139374554157, 'timing/model': 0.9503378728404641, 'learning_rate/qwen_vl_interface': 6.259945429464343e-07, 'learning_rate/action_model': 6.259945429464342e-06, 'epoch': 2.67}) +[2026-06-18 02:39:51,160][starVLA.training.train_starvla][INFO] - Step 6494, Loss: {'train/loss': 0.4010581970214844, 'train/grad_norm_pre_clip': 24.13553810119629, 'timing/data': 0.028986845165491104, 'timing/model': 0.9485047329217196, 'learning_rate/qwen_vl_interface': 6.255002113524885e-07, 'learning_rate/action_model': 6.255002113524886e-06, 'epoch': 2.67}) +[2026-06-18 02:39:54,966][starVLA.training.train_starvla][INFO] - Step 6495, Loss: {'train/loss': 0.13356083631515503, 'train/grad_norm_pre_clip': 15.11150074005127, 'timing/data': 0.026732674799859524, 'timing/model': 0.9314601570367813, 'learning_rate/qwen_vl_interface': 6.250068384223255e-07, 'learning_rate/action_model': 6.250068384223255e-06, 'epoch': 2.67}) +[2026-06-18 02:40:00,837][starVLA.training.train_starvla][INFO] - Step 6496, Loss: {'train/loss': 0.42352738976478577, 'train/grad_norm_pre_clip': 42.62508773803711, 'timing/data': 0.024485675618052483, 'timing/model': 2.8579842299222946, 'learning_rate/qwen_vl_interface': 6.245144242582214e-07, 'learning_rate/action_model': 6.245144242582214e-06, 'epoch': 2.67}) +[2026-06-18 02:40:04,989][starVLA.training.train_starvla][INFO] - Step 6497, Loss: {'train/loss': 0.4217962622642517, 'train/grad_norm_pre_clip': 27.353275299072266, 'timing/data': 0.022758671082556248, 'timing/model': 1.3269967185333371, 'learning_rate/qwen_vl_interface': 6.240229689622546e-07, 'learning_rate/action_model': 6.240229689622545e-06, 'epoch': 2.67}) +[2026-06-18 02:40:08,938][starVLA.training.train_starvla][INFO] - Step 6498, Loss: {'train/loss': 0.17292603850364685, 'train/grad_norm_pre_clip': 18.402332305908203, 'timing/data': 0.027441627345979214, 'timing/model': 1.1256555682048202, 'learning_rate/qwen_vl_interface': 6.235324726363042e-07, 'learning_rate/action_model': 6.235324726363041e-06, 'epoch': 2.67}) +[2026-06-18 02:40:12,600][starVLA.training.train_starvla][INFO] - Step 6499, Loss: {'train/loss': 0.20461930334568024, 'train/grad_norm_pre_clip': 27.648210525512695, 'timing/data': 0.022006486542522907, 'timing/model': 0.9698538463562727, 'learning_rate/qwen_vl_interface': 6.230429353820503e-07, 'learning_rate/action_model': 6.230429353820503e-06, 'epoch': 2.68}) +[2026-06-18 02:42:21,183][starVLA.training.train_starvla][INFO] - Step 6500, Loss: {'train/loss': 0.218216210603714, 'train/grad_norm_pre_clip': 19.40958595275879, 'eval/loss': 0.41687628626823425, 'eval/latency_6/loss': 0.41687628626823425, 'eval/demon_fire_precision': 0.960025583626479, 'eval/demon_fire_recall': 0.9702650290885585, 'eval/demon_fire_f1': 0.9651181482076836, 'eval/demon_fire_rate_model': 0.8700612131329994, 'eval/demon_fire_rate_teacher': 0.860879243183083, 'eval/demon_left_precision': 0.8740157480314961, 'eval/demon_left_recall': 0.7789473684210526, 'eval/demon_left_f1': 0.8237476808905381, 'eval/demon_left_rate_model': 0.14134668892598776, 'eval/demon_left_rate_teacher': 0.15859766277128548, 'eval/demon_right_precision': 0.8327868852459016, 'eval/demon_right_recall': 0.7803379416282642, 'eval/demon_right_f1': 0.8057097541633624, 'eval/demon_right_rate_model': 0.16972732331663884, 'eval/demon_right_rate_teacher': 0.18113522537562604, 'eval/demon_move_f1': 0.8147287175269502, 'eval/demon_cc_f1': 0.9144775758163808, 'eval/latency_6/demon_fire_precision': 0.960025583626479, 'eval/latency_6/demon_fire_recall': 0.9702650290885585, 'eval/latency_6/demon_fire_f1': 0.9651181482076836, 'eval/latency_6/demon_fire_rate_model': 0.8700612131329994, 'eval/latency_6/demon_fire_rate_teacher': 0.860879243183083, 'eval/latency_6/demon_left_precision': 0.8740157480314961, 'eval/latency_6/demon_left_recall': 0.7789473684210526, 'eval/latency_6/demon_left_f1': 0.8237476808905381, 'eval/latency_6/demon_left_rate_model': 0.14134668892598776, 'eval/latency_6/demon_left_rate_teacher': 0.15859766277128548, 'eval/latency_6/demon_right_precision': 0.8327868852459016, 'eval/latency_6/demon_right_recall': 0.7803379416282642, 'eval/latency_6/demon_right_f1': 0.8057097541633624, 'eval/latency_6/demon_right_rate_model': 0.16972732331663884, 'eval/latency_6/demon_right_rate_teacher': 0.18113522537562604, 'eval/latency_6/demon_move_f1': 0.8147287175269502, 'eval/latency_6/demon_cc_f1': 0.9144775758163808, 'timing/data': 0.02081548422574997, 'timing/model': 0.9339548358693719, 'learning_rate/qwen_vl_interface': 6.225543573009747e-07, 'learning_rate/action_model': 6.225543573009748e-06, 'epoch': 2.68}) +[2026-06-18 02:42:21,185][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_6500_state +[2026-06-18 02:42:21,206][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 02:42:27,969][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_6500_state/model.safetensors +[2026-06-18 02:42:43,172][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_6500_state/optimizer.bin +[2026-06-18 02:42:43,181][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_6500_state/sampler.bin +[2026-06-18 02:42:43,181][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_6500_state/sampler_1.bin +[2026-06-18 02:42:43,184][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_6500_state/random_states_0.pkl +[2026-06-18 02:42:43,188][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 6500 +[2026-06-18 02:42:43,188][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 02:42:43,205][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 02:42:46,847][starVLA.training.train_starvla][INFO] - Step 6501, Loss: {'train/loss': 0.21201351284980774, 'train/grad_norm_pre_clip': 17.23957633972168, 'timing/data': 0.030780979432165623, 'timing/model': 0.9324065009132028, 'learning_rate/qwen_vl_interface': 6.2206673849436e-07, 'learning_rate/action_model': 6.2206673849436e-06, 'epoch': 2.68}) +[2026-06-18 02:42:50,834][starVLA.training.train_starvla][INFO] - Step 6502, Loss: {'train/loss': 0.28981783986091614, 'train/grad_norm_pre_clip': 26.866178512573242, 'timing/data': 0.0336623452603817, 'timing/model': 0.9432571018114686, 'learning_rate/qwen_vl_interface': 6.215800790632905e-07, 'learning_rate/action_model': 6.215800790632904e-06, 'epoch': 2.68}) +[2026-06-18 02:42:54,757][starVLA.training.train_starvla][INFO] - Step 6503, Loss: {'train/loss': 0.5206361413002014, 'train/grad_norm_pre_clip': 33.584014892578125, 'timing/data': 0.029186404310166836, 'timing/model': 0.9303718786686659, 'learning_rate/qwen_vl_interface': 6.210943791086506e-07, 'learning_rate/action_model': 6.210943791086506e-06, 'epoch': 2.68}) +[2026-06-18 02:42:58,553][starVLA.training.train_starvla][INFO] - Step 6504, Loss: {'train/loss': 0.4969516396522522, 'train/grad_norm_pre_clip': 25.424955368041992, 'timing/data': 0.022364115342497826, 'timing/model': 0.947734042070806, 'learning_rate/qwen_vl_interface': 6.206096387311278e-07, 'learning_rate/action_model': 6.206096387311277e-06, 'epoch': 2.68}) +[2026-06-18 02:43:02,464][starVLA.training.train_starvla][INFO] - Step 6505, Loss: {'train/loss': 0.23199020326137543, 'train/grad_norm_pre_clip': 15.715106010437012, 'timing/data': 0.029218411073088646, 'timing/model': 0.9327701162546873, 'learning_rate/qwen_vl_interface': 6.201258580312078e-07, 'learning_rate/action_model': 6.201258580312077e-06, 'epoch': 2.68}) +[2026-06-18 02:43:06,395][starVLA.training.train_starvla][INFO] - Step 6506, Loss: {'train/loss': 0.33811482787132263, 'train/grad_norm_pre_clip': 31.117368698120117, 'timing/data': 0.024595863185822964, 'timing/model': 0.9543332699686289, 'learning_rate/qwen_vl_interface': 6.196430371091792e-07, 'learning_rate/action_model': 6.196430371091792e-06, 'epoch': 2.68}) +[2026-06-18 02:43:10,284][starVLA.training.train_starvla][INFO] - Step 6507, Loss: {'train/loss': 0.27544066309928894, 'train/grad_norm_pre_clip': 19.387802124023438, 'timing/data': 0.02178104128688574, 'timing/model': 1.1506695291027427, 'learning_rate/qwen_vl_interface': 6.191611760651312e-07, 'learning_rate/action_model': 6.191611760651312e-06, 'epoch': 2.68}) +[2026-06-18 02:43:14,248][starVLA.training.train_starvla][INFO] - Step 6508, Loss: {'train/loss': 0.33823174238204956, 'train/grad_norm_pre_clip': 20.540246963500977, 'timing/data': 0.023511042818427086, 'timing/model': 0.9247491396963596, 'learning_rate/qwen_vl_interface': 6.18680274998955e-07, 'learning_rate/action_model': 6.186802749989549e-06, 'epoch': 2.68}) +[2026-06-18 02:43:18,040][starVLA.training.train_starvla][INFO] - Step 6509, Loss: {'train/loss': 0.05645471066236496, 'train/grad_norm_pre_clip': 6.543392658233643, 'timing/data': 0.021864648908376694, 'timing/model': 0.9239857541397214, 'learning_rate/qwen_vl_interface': 6.182003340103413e-07, 'learning_rate/action_model': 6.1820033401034135e-06, 'epoch': 2.68}) +[2026-06-18 02:43:21,784][starVLA.training.train_starvla][INFO] - Step 6510, Loss: {'train/loss': 0.3384696841239929, 'train/grad_norm_pre_clip': 25.4531192779541, 'timing/data': 0.025041548535227776, 'timing/model': 0.9217352122068405, 'learning_rate/qwen_vl_interface': 6.177213531987819e-07, 'learning_rate/action_model': 6.177213531987819e-06, 'epoch': 2.68}) +[2026-06-18 02:43:25,583][starVLA.training.train_starvla][INFO] - Step 6511, Loss: {'train/loss': 0.03377183899283409, 'train/grad_norm_pre_clip': 6.589432239532471, 'timing/data': 0.03022481221705675, 'timing/model': 0.9377398556098342, 'learning_rate/qwen_vl_interface': 6.172433326635714e-07, 'learning_rate/action_model': 6.172433326635713e-06, 'epoch': 2.68}) +[2026-06-18 02:43:31,082][starVLA.training.train_starvla][INFO] - Step 6512, Loss: {'train/loss': 0.16359825432300568, 'train/grad_norm_pre_clip': 15.759936332702637, 'timing/data': 0.02695674356073141, 'timing/model': 2.380935446359217, 'learning_rate/qwen_vl_interface': 6.167662725038025e-07, 'learning_rate/action_model': 6.1676627250380255e-06, 'epoch': 2.68}) +[2026-06-18 02:43:35,003][starVLA.training.train_starvla][INFO] - Step 6513, Loss: {'train/loss': 0.19633929431438446, 'train/grad_norm_pre_clip': 15.250068664550781, 'timing/data': 0.026555541902780533, 'timing/model': 1.0985955782234669, 'learning_rate/qwen_vl_interface': 6.16290172818371e-07, 'learning_rate/action_model': 6.16290172818371e-06, 'epoch': 2.68}) +[2026-06-18 02:43:38,657][starVLA.training.train_starvla][INFO] - Step 6514, Loss: {'train/loss': 0.11853974312543869, 'train/grad_norm_pre_clip': 16.809425354003906, 'timing/data': 0.024458455853164196, 'timing/model': 0.9284433359280229, 'learning_rate/qwen_vl_interface': 6.158150337059725e-07, 'learning_rate/action_model': 6.158150337059725e-06, 'epoch': 2.68}) +[2026-06-18 02:43:43,735][starVLA.training.train_starvla][INFO] - Step 6515, Loss: {'train/loss': 0.24290741980075836, 'train/grad_norm_pre_clip': 21.166030883789062, 'timing/data': 0.03182661160826683, 'timing/model': 0.9346997747197747, 'learning_rate/qwen_vl_interface': 6.153408552651045e-07, 'learning_rate/action_model': 6.153408552651045e-06, 'epoch': 2.68}) +[2026-06-18 02:43:47,635][starVLA.training.train_starvla][INFO] - Step 6516, Loss: {'train/loss': 0.44807058572769165, 'train/grad_norm_pre_clip': 25.876253128051758, 'timing/data': 0.03301279619336128, 'timing/model': 0.9455003347247839, 'learning_rate/qwen_vl_interface': 6.14867637594064e-07, 'learning_rate/action_model': 6.14867637594064e-06, 'epoch': 2.68}) +[2026-06-18 02:43:51,532][starVLA.training.train_starvla][INFO] - Step 6517, Loss: {'train/loss': 0.07228630781173706, 'train/grad_norm_pre_clip': 9.709284782409668, 'timing/data': 0.03504992835223675, 'timing/model': 0.9428360583260655, 'learning_rate/qwen_vl_interface': 6.143953807909497e-07, 'learning_rate/action_model': 6.143953807909496e-06, 'epoch': 2.68}) +[2026-06-18 02:43:55,493][starVLA.training.train_starvla][INFO] - Step 6518, Loss: {'train/loss': 0.07830337435007095, 'train/grad_norm_pre_clip': 27.686723709106445, 'timing/data': 0.022275686264038086, 'timing/model': 1.0422412352636456, 'learning_rate/qwen_vl_interface': 6.139240849536616e-07, 'learning_rate/action_model': 6.1392408495366156e-06, 'epoch': 2.68}) +[2026-06-18 02:43:59,425][starVLA.training.train_starvla][INFO] - Step 6519, Loss: {'train/loss': 0.02704264223575592, 'train/grad_norm_pre_clip': 4.424079895019531, 'timing/data': 0.024649986065924168, 'timing/model': 1.1155690615996718, 'learning_rate/qwen_vl_interface': 6.134537501798993e-07, 'learning_rate/action_model': 6.1345375017989925e-06, 'epoch': 2.68}) +[2026-06-18 02:44:03,221][starVLA.training.train_starvla][INFO] - Step 6520, Loss: {'train/loss': 0.3835887312889099, 'train/grad_norm_pre_clip': 30.509645462036133, 'timing/data': 0.029024073854088783, 'timing/model': 1.118144447915256, 'learning_rate/qwen_vl_interface': 6.129843765671634e-07, 'learning_rate/action_model': 6.129843765671633e-06, 'epoch': 2.68}) +[2026-06-18 02:44:06,839][starVLA.training.train_starvla][INFO] - Step 6521, Loss: {'train/loss': 0.11363652348518372, 'train/grad_norm_pre_clip': 14.949549674987793, 'timing/data': 0.025314574129879475, 'timing/model': 0.9359099203720689, 'learning_rate/qwen_vl_interface': 6.125159642127559e-07, 'learning_rate/action_model': 6.125159642127558e-06, 'epoch': 2.68}) +[2026-06-18 02:44:10,891][starVLA.training.train_starvla][INFO] - Step 6522, Loss: {'train/loss': 0.10336533933877945, 'train/grad_norm_pre_clip': 25.125320434570312, 'timing/data': 0.028223847970366478, 'timing/model': 0.9371044533327222, 'learning_rate/qwen_vl_interface': 6.120485132137793e-07, 'learning_rate/action_model': 6.120485132137794e-06, 'epoch': 2.68}) +[2026-06-18 02:44:14,778][starVLA.training.train_starvla][INFO] - Step 6523, Loss: {'train/loss': 0.09311278909444809, 'train/grad_norm_pre_clip': 19.403968811035156, 'timing/data': 0.026392100378870964, 'timing/model': 0.9256129218265414, 'learning_rate/qwen_vl_interface': 6.115820236671369e-07, 'learning_rate/action_model': 6.115820236671369e-06, 'epoch': 2.69}) +[2026-06-18 02:44:18,671][starVLA.training.train_starvla][INFO] - Step 6524, Loss: {'train/loss': 0.17195701599121094, 'train/grad_norm_pre_clip': 15.811602592468262, 'timing/data': 0.0386483957991004, 'timing/model': 0.9305813172832131, 'learning_rate/qwen_vl_interface': 6.111164956695316e-07, 'learning_rate/action_model': 6.111164956695316e-06, 'epoch': 2.69}) +[2026-06-18 02:44:22,833][starVLA.training.train_starvla][INFO] - Step 6525, Loss: {'train/loss': 0.1075364276766777, 'train/grad_norm_pre_clip': 14.79423713684082, 'timing/data': 0.03455042839050293, 'timing/model': 0.9593660281971097, 'learning_rate/qwen_vl_interface': 6.106519293174684e-07, 'learning_rate/action_model': 6.106519293174684e-06, 'epoch': 2.69}) +[2026-06-18 02:44:31,942][starVLA.training.train_starvla][INFO] - Step 6526, Loss: {'train/loss': 0.16690880060195923, 'train/grad_norm_pre_clip': 26.207454681396484, 'timing/data': 0.04193305317312479, 'timing/model': 3.4646275816485286, 'learning_rate/qwen_vl_interface': 6.101883247072525e-07, 'learning_rate/action_model': 6.1018832470725245e-06, 'epoch': 2.69}) +[2026-06-18 02:44:35,600][starVLA.training.train_starvla][INFO] - Step 6527, Loss: {'train/loss': 0.13938403129577637, 'train/grad_norm_pre_clip': 17.71700096130371, 'timing/data': 0.032661029137670994, 'timing/model': 0.9397519500926137, 'learning_rate/qwen_vl_interface': 6.097256819349899e-07, 'learning_rate/action_model': 6.0972568193498985e-06, 'epoch': 2.69}) +[2026-06-18 02:44:40,656][starVLA.training.train_starvla][INFO] - Step 6528, Loss: {'train/loss': 0.41652679443359375, 'train/grad_norm_pre_clip': 34.620086669921875, 'timing/data': 0.028464125469326973, 'timing/model': 0.9389848662540317, 'learning_rate/qwen_vl_interface': 6.092640010965864e-07, 'learning_rate/action_model': 6.092640010965864e-06, 'epoch': 2.69}) +[2026-06-18 02:44:44,360][starVLA.training.train_starvla][INFO] - Step 6529, Loss: {'train/loss': 0.16609162092208862, 'train/grad_norm_pre_clip': 20.105375289916992, 'timing/data': 0.030565875582396984, 'timing/model': 0.9297592891380191, 'learning_rate/qwen_vl_interface': 6.088032822877488e-07, 'learning_rate/action_model': 6.088032822877488e-06, 'epoch': 2.69}) +[2026-06-18 02:44:48,142][starVLA.training.train_starvla][INFO] - Step 6530, Loss: {'train/loss': 0.05075022950768471, 'train/grad_norm_pre_clip': 7.935551166534424, 'timing/data': 0.032960034906864166, 'timing/model': 0.9392914613708854, 'learning_rate/qwen_vl_interface': 6.083435256039854e-07, 'learning_rate/action_model': 6.083435256039854e-06, 'epoch': 2.69}) +[2026-06-18 02:44:52,047][starVLA.training.train_starvla][INFO] - Step 6531, Loss: {'train/loss': 0.20483611524105072, 'train/grad_norm_pre_clip': 16.23586654663086, 'timing/data': 0.03206085879355669, 'timing/model': 0.9426618544384837, 'learning_rate/qwen_vl_interface': 6.078847311406033e-07, 'learning_rate/action_model': 6.078847311406032e-06, 'epoch': 2.69}) +[2026-06-18 02:44:55,952][starVLA.training.train_starvla][INFO] - Step 6532, Loss: {'train/loss': 0.23151659965515137, 'train/grad_norm_pre_clip': 26.45369529724121, 'timing/data': 0.02775725442916155, 'timing/model': 1.1045530727133155, 'learning_rate/qwen_vl_interface': 6.07426898992712e-07, 'learning_rate/action_model': 6.074268989927119e-06, 'epoch': 2.69}) +[2026-06-18 02:44:59,817][starVLA.training.train_starvla][INFO] - Step 6533, Loss: {'train/loss': 0.4847971796989441, 'train/grad_norm_pre_clip': 34.313724517822266, 'timing/data': 0.02694892417639494, 'timing/model': 1.1123125590384007, 'learning_rate/qwen_vl_interface': 6.069700292552202e-07, 'learning_rate/action_model': 6.069700292552202e-06, 'epoch': 2.69}) +[2026-06-18 02:45:03,445][starVLA.training.train_starvla][INFO] - Step 6534, Loss: {'train/loss': 0.0765487551689148, 'train/grad_norm_pre_clip': 10.163873672485352, 'timing/data': 0.028575352393090725, 'timing/model': 0.9452041042968631, 'learning_rate/qwen_vl_interface': 6.065141220228372e-07, 'learning_rate/action_model': 6.065141220228372e-06, 'epoch': 2.69}) +[2026-06-18 02:45:07,353][starVLA.training.train_starvla][INFO] - Step 6535, Loss: {'train/loss': 0.2600145936012268, 'train/grad_norm_pre_clip': 23.2391300201416, 'timing/data': 0.027231226675212383, 'timing/model': 0.9470833782106638, 'learning_rate/qwen_vl_interface': 6.060591773900732e-07, 'learning_rate/action_model': 6.060591773900733e-06, 'epoch': 2.69}) +[2026-06-18 02:45:11,164][starVLA.training.train_starvla][INFO] - Step 6536, Loss: {'train/loss': 0.15927262604236603, 'train/grad_norm_pre_clip': 17.49606704711914, 'timing/data': 0.029515701346099377, 'timing/model': 0.9358079414814711, 'learning_rate/qwen_vl_interface': 6.056051954512388e-07, 'learning_rate/action_model': 6.0560519545123875e-06, 'epoch': 2.69}) +[2026-06-18 02:45:15,084][starVLA.training.train_starvla][INFO] - Step 6537, Loss: {'train/loss': 0.1344269961118698, 'train/grad_norm_pre_clip': 15.060538291931152, 'timing/data': 0.03677043039351702, 'timing/model': 0.9384418334811926, 'learning_rate/qwen_vl_interface': 6.051521763004455e-07, 'learning_rate/action_model': 6.051521763004456e-06, 'epoch': 2.69}) +[2026-06-18 02:45:18,989][starVLA.training.train_starvla][INFO] - Step 6538, Loss: {'train/loss': 0.19585970044136047, 'train/grad_norm_pre_clip': 25.42500114440918, 'timing/data': 0.03623265586793423, 'timing/model': 0.9420324312523007, 'learning_rate/qwen_vl_interface': 6.047001200316038e-07, 'learning_rate/action_model': 6.047001200316037e-06, 'epoch': 2.69}) +[2026-06-18 02:45:22,888][starVLA.training.train_starvla][INFO] - Step 6539, Loss: {'train/loss': 0.30123454332351685, 'train/grad_norm_pre_clip': 39.79175567626953, 'timing/data': 0.02375673968344927, 'timing/model': 1.1659767320379615, 'learning_rate/qwen_vl_interface': 6.042490267384255e-07, 'learning_rate/action_model': 6.0424902673842545e-06, 'epoch': 2.69}) +[2026-06-18 02:45:26,704][starVLA.training.train_starvla][INFO] - Step 6540, Loss: {'train/loss': 0.31704631447792053, 'train/grad_norm_pre_clip': 47.664390563964844, 'timing/data': 0.027551569044589996, 'timing/model': 0.9506119005382061, 'learning_rate/qwen_vl_interface': 6.03798896514423e-07, 'learning_rate/action_model': 6.0379889651442295e-06, 'epoch': 2.69}) +[2026-06-18 02:45:30,448][starVLA.training.train_starvla][INFO] - Step 6541, Loss: {'train/loss': 0.14422418177127838, 'train/grad_norm_pre_clip': 18.77229881286621, 'timing/data': 0.03195756208151579, 'timing/model': 0.9144220305606723, 'learning_rate/qwen_vl_interface': 6.033497294529093e-07, 'learning_rate/action_model': 6.033497294529093e-06, 'epoch': 2.69}) +[2026-06-18 02:45:34,158][starVLA.training.train_starvla][INFO] - Step 6542, Loss: {'train/loss': 0.11216370761394501, 'train/grad_norm_pre_clip': 21.466184616088867, 'timing/data': 0.03261403553187847, 'timing/model': 0.9247371563687921, 'learning_rate/qwen_vl_interface': 6.029015256469967e-07, 'learning_rate/action_model': 6.0290152564699666e-06, 'epoch': 2.69}) +[2026-06-18 02:45:37,932][starVLA.training.train_starvla][INFO] - Step 6543, Loss: {'train/loss': 0.14129143953323364, 'train/grad_norm_pre_clip': 15.248085975646973, 'timing/data': 0.030000035651028156, 'timing/model': 0.931554451584816, 'learning_rate/qwen_vl_interface': 6.024542851895976e-07, 'learning_rate/action_model': 6.0245428518959756e-06, 'epoch': 2.69}) +[2026-06-18 02:45:41,863][starVLA.training.train_starvla][INFO] - Step 6544, Loss: {'train/loss': 0.1192665547132492, 'train/grad_norm_pre_clip': 20.48094940185547, 'timing/data': 0.03864549659192562, 'timing/model': 0.9513406865298748, 'learning_rate/qwen_vl_interface': 6.020080081734267e-07, 'learning_rate/action_model': 6.020080081734266e-06, 'epoch': 2.69}) +[2026-06-18 02:45:45,763][starVLA.training.train_starvla][INFO] - Step 6545, Loss: {'train/loss': 0.11564327776432037, 'train/grad_norm_pre_clip': 19.871862411499023, 'timing/data': 0.031075497157871723, 'timing/model': 0.9683956522494555, 'learning_rate/qwen_vl_interface': 6.015626946909974e-07, 'learning_rate/action_model': 6.015626946909973e-06, 'epoch': 2.69}) +[2026-06-18 02:45:49,598][starVLA.training.train_starvla][INFO] - Step 6546, Loss: {'train/loss': 0.21233490109443665, 'train/grad_norm_pre_clip': 22.933401107788086, 'timing/data': 0.024306777864694595, 'timing/model': 1.12234160117805, 'learning_rate/qwen_vl_interface': 6.011183448346233e-07, 'learning_rate/action_model': 6.011183448346233e-06, 'epoch': 2.69}) +[2026-06-18 02:45:53,307][starVLA.training.train_starvla][INFO] - Step 6547, Loss: {'train/loss': 0.26877260208129883, 'train/grad_norm_pre_clip': 25.55668830871582, 'timing/data': 0.025050866417586803, 'timing/model': 0.9497216297313571, 'learning_rate/qwen_vl_interface': 6.006749586964183e-07, 'learning_rate/action_model': 6.006749586964183e-06, 'epoch': 2.69}) +[2026-06-18 02:45:57,514][starVLA.training.train_starvla][INFO] - Step 6548, Loss: {'train/loss': 0.09253069013357162, 'train/grad_norm_pre_clip': 13.51911449432373, 'timing/data': 0.02622292935848236, 'timing/model': 0.9560113810002804, 'learning_rate/qwen_vl_interface': 6.002325363682974e-07, 'learning_rate/action_model': 6.0023253636829736e-06, 'epoch': 2.7}) +[2026-06-18 02:46:01,320][starVLA.training.train_starvla][INFO] - Step 6549, Loss: {'train/loss': 0.3282078206539154, 'train/grad_norm_pre_clip': 19.295787811279297, 'timing/data': 0.021269788034260273, 'timing/model': 0.9820924764499068, 'learning_rate/qwen_vl_interface': 5.997910779419756e-07, 'learning_rate/action_model': 5.9979107794197565e-06, 'epoch': 2.7}) +[2026-06-18 02:46:05,362][starVLA.training.train_starvla][INFO] - Step 6550, Loss: {'train/loss': 0.23047664761543274, 'train/grad_norm_pre_clip': 25.016281127929688, 'timing/data': 0.03560935705900192, 'timing/model': 0.9507954977452755, 'learning_rate/qwen_vl_interface': 5.993505835089661e-07, 'learning_rate/action_model': 5.993505835089661e-06, 'epoch': 2.7}) +[2026-06-18 02:46:10,334][starVLA.training.train_starvla][INFO] - Step 6551, Loss: {'train/loss': 0.17236779630184174, 'train/grad_norm_pre_clip': 20.688079833984375, 'timing/data': 0.022744635120034218, 'timing/model': 2.0251848809421062, 'learning_rate/qwen_vl_interface': 5.98911053160586e-07, 'learning_rate/action_model': 5.989110531605859e-06, 'epoch': 2.7}) +[2026-06-18 02:46:14,712][starVLA.training.train_starvla][INFO] - Step 6552, Loss: {'train/loss': 0.13341282308101654, 'train/grad_norm_pre_clip': 21.761308670043945, 'timing/data': 0.0229913592338562, 'timing/model': 1.0907357316464186, 'learning_rate/qwen_vl_interface': 5.984724869879494e-07, 'learning_rate/action_model': 5.984724869879493e-06, 'epoch': 2.7}) +[2026-06-18 02:46:18,688][starVLA.training.train_starvla][INFO] - Step 6553, Loss: {'train/loss': 0.1989014595746994, 'train/grad_norm_pre_clip': 21.916902542114258, 'timing/data': 0.028504138812422752, 'timing/model': 1.0898249996826053, 'learning_rate/qwen_vl_interface': 5.980348850819711e-07, 'learning_rate/action_model': 5.98034885081971e-06, 'epoch': 2.7}) +[2026-06-18 02:46:22,270][starVLA.training.train_starvla][INFO] - Step 6554, Loss: {'train/loss': 0.11926186084747314, 'train/grad_norm_pre_clip': 13.84957218170166, 'timing/data': 0.03170089330524206, 'timing/model': 0.9206539178267121, 'learning_rate/qwen_vl_interface': 5.975982475333664e-07, 'learning_rate/action_model': 5.975982475333663e-06, 'epoch': 2.7}) +[2026-06-18 02:46:26,079][starVLA.training.train_starvla][INFO] - Step 6555, Loss: {'train/loss': 0.0936252772808075, 'train/grad_norm_pre_clip': 13.203910827636719, 'timing/data': 0.026550592854619026, 'timing/model': 0.9410360408946872, 'learning_rate/qwen_vl_interface': 5.971625744326516e-07, 'learning_rate/action_model': 5.971625744326515e-06, 'epoch': 2.7}) +[2026-06-18 02:46:29,901][starVLA.training.train_starvla][INFO] - Step 6556, Loss: {'train/loss': 0.16945014894008636, 'train/grad_norm_pre_clip': 23.036806106567383, 'timing/data': 0.027001825161278248, 'timing/model': 0.949495155364275, 'learning_rate/qwen_vl_interface': 5.967278658701421e-07, 'learning_rate/action_model': 5.967278658701421e-06, 'epoch': 2.7}) +[2026-06-18 02:46:34,503][starVLA.training.train_starvla][INFO] - Step 6557, Loss: {'train/loss': 0.160495787858963, 'train/grad_norm_pre_clip': 20.23673439025879, 'timing/data': 0.02651855070143938, 'timing/model': 0.9501366633921862, 'learning_rate/qwen_vl_interface': 5.962941219359527e-07, 'learning_rate/action_model': 5.962941219359527e-06, 'epoch': 2.7}) +[2026-06-18 02:46:38,508][starVLA.training.train_starvla][INFO] - Step 6558, Loss: {'train/loss': 0.21528735756874084, 'train/grad_norm_pre_clip': 25.252281188964844, 'timing/data': 0.024913846515119076, 'timing/model': 0.9430685034021735, 'learning_rate/qwen_vl_interface': 5.958613427199998e-07, 'learning_rate/action_model': 5.958613427199997e-06, 'epoch': 2.7}) +[2026-06-18 02:46:43,268][starVLA.training.train_starvla][INFO] - Step 6559, Loss: {'train/loss': 0.3267170190811157, 'train/grad_norm_pre_clip': 28.517066955566406, 'timing/data': 0.024789493530988693, 'timing/model': 1.2777934856712818, 'learning_rate/qwen_vl_interface': 5.954295283119985e-07, 'learning_rate/action_model': 5.954295283119985e-06, 'epoch': 2.7}) +[2026-06-18 02:46:47,136][starVLA.training.train_starvla][INFO] - Step 6560, Loss: {'train/loss': 0.356863409280777, 'train/grad_norm_pre_clip': 29.157392501831055, 'timing/data': 0.0281536178663373, 'timing/model': 1.070170528255403, 'learning_rate/qwen_vl_interface': 5.949986788014643e-07, 'learning_rate/action_model': 5.949986788014643e-06, 'epoch': 2.7}) +[2026-06-18 02:46:50,975][starVLA.training.train_starvla][INFO] - Step 6561, Loss: {'train/loss': 0.40025094151496887, 'train/grad_norm_pre_clip': 25.001558303833008, 'timing/data': 0.02366217877715826, 'timing/model': 0.940362467430532, 'learning_rate/qwen_vl_interface': 5.94568794277713e-07, 'learning_rate/action_model': 5.94568794277713e-06, 'epoch': 2.7}) +[2026-06-18 02:46:54,869][starVLA.training.train_starvla][INFO] - Step 6562, Loss: {'train/loss': 0.25536176562309265, 'train/grad_norm_pre_clip': 29.221553802490234, 'timing/data': 0.028130030259490013, 'timing/model': 0.9429266471415758, 'learning_rate/qwen_vl_interface': 5.941398748298603e-07, 'learning_rate/action_model': 5.941398748298603e-06, 'epoch': 2.7}) +[2026-06-18 02:46:58,721][starVLA.training.train_starvla][INFO] - Step 6563, Loss: {'train/loss': 0.37065014243125916, 'train/grad_norm_pre_clip': 33.0805549621582, 'timing/data': 0.02636503055691719, 'timing/model': 0.9351806193590164, 'learning_rate/qwen_vl_interface': 5.937119205468217e-07, 'learning_rate/action_model': 5.9371192054682165e-06, 'epoch': 2.7}) +[2026-06-18 02:47:02,467][starVLA.training.train_starvla][INFO] - Step 6564, Loss: {'train/loss': 0.391716867685318, 'train/grad_norm_pre_clip': 33.44061279296875, 'timing/data': 0.02613589260727167, 'timing/model': 0.9376235771924257, 'learning_rate/qwen_vl_interface': 5.93284931517312e-07, 'learning_rate/action_model': 5.93284931517312e-06, 'epoch': 2.7}) +[2026-06-18 02:47:06,362][starVLA.training.train_starvla][INFO] - Step 6565, Loss: {'train/loss': 0.05449775606393814, 'train/grad_norm_pre_clip': 10.81749439239502, 'timing/data': 0.02287624590098858, 'timing/model': 0.931218777783215, 'learning_rate/qwen_vl_interface': 5.928589078298468e-07, 'learning_rate/action_model': 5.928589078298468e-06, 'epoch': 2.7}) +[2026-06-18 02:47:10,131][starVLA.training.train_starvla][INFO] - Step 6566, Loss: {'train/loss': 0.4442332983016968, 'train/grad_norm_pre_clip': 26.588397979736328, 'timing/data': 0.022433833219110966, 'timing/model': 1.0984726417809725, 'learning_rate/qwen_vl_interface': 5.924338495727416e-07, 'learning_rate/action_model': 5.9243384957274155e-06, 'epoch': 2.7}) +[2026-06-18 02:47:13,737][starVLA.training.train_starvla][INFO] - Step 6567, Loss: {'train/loss': 0.29243767261505127, 'train/grad_norm_pre_clip': 22.353700637817383, 'timing/data': 0.026639857329428196, 'timing/model': 0.9271841654554009, 'learning_rate/qwen_vl_interface': 5.92009756834111e-07, 'learning_rate/action_model': 5.9200975683411094e-06, 'epoch': 2.7}) +[2026-06-18 02:47:17,909][starVLA.training.train_starvla][INFO] - Step 6568, Loss: {'train/loss': 0.18582960963249207, 'train/grad_norm_pre_clip': 24.586305618286133, 'timing/data': 0.027167816646397114, 'timing/model': 0.9593112822622061, 'learning_rate/qwen_vl_interface': 5.915866297018697e-07, 'learning_rate/action_model': 5.915866297018696e-06, 'epoch': 2.7}) +[2026-06-18 02:47:21,725][starVLA.training.train_starvla][INFO] - Step 6569, Loss: {'train/loss': 0.12771815061569214, 'train/grad_norm_pre_clip': 21.25370979309082, 'timing/data': 0.022608849219977856, 'timing/model': 0.9828642336651683, 'learning_rate/qwen_vl_interface': 5.911644682637329e-07, 'learning_rate/action_model': 5.911644682637329e-06, 'epoch': 2.7}) +[2026-06-18 02:47:25,579][starVLA.training.train_starvla][INFO] - Step 6570, Loss: {'train/loss': 0.2948276102542877, 'train/grad_norm_pre_clip': 26.737096786499023, 'timing/data': 0.02623080462217331, 'timing/model': 0.9382093073800206, 'learning_rate/qwen_vl_interface': 5.907432726072149e-07, 'learning_rate/action_model': 5.9074327260721485e-06, 'epoch': 2.7}) +[2026-06-18 02:47:29,641][starVLA.training.train_starvla][INFO] - Step 6571, Loss: {'train/loss': 0.36427900195121765, 'train/grad_norm_pre_clip': 28.215848922729492, 'timing/data': 0.02417147345840931, 'timing/model': 0.9326138533651829, 'learning_rate/qwen_vl_interface': 5.903230428196305e-07, 'learning_rate/action_model': 5.9032304281963045e-06, 'epoch': 2.7}) +[2026-06-18 02:47:33,631][starVLA.training.train_starvla][INFO] - Step 6572, Loss: {'train/loss': 0.26650118827819824, 'train/grad_norm_pre_clip': 28.239707946777344, 'timing/data': 0.02878453303128481, 'timing/model': 1.1069983467459679, 'learning_rate/qwen_vl_interface': 5.899037789880929e-07, 'learning_rate/action_model': 5.899037789880929e-06, 'epoch': 2.71}) +[2026-06-18 02:47:37,453][starVLA.training.train_starvla][INFO] - Step 6573, Loss: {'train/loss': 0.16149312257766724, 'train/grad_norm_pre_clip': 16.376649856567383, 'timing/data': 0.02187723759561777, 'timing/model': 1.1044872356578708, 'learning_rate/qwen_vl_interface': 5.894854811995166e-07, 'learning_rate/action_model': 5.894854811995165e-06, 'epoch': 2.71}) +[2026-06-18 02:47:42,414][starVLA.training.train_starvla][INFO] - Step 6574, Loss: {'train/loss': 0.07236961275339127, 'train/grad_norm_pre_clip': 9.354570388793945, 'timing/data': 0.028407092206180096, 'timing/model': 1.034580216743052, 'learning_rate/qwen_vl_interface': 5.890681495406152e-07, 'learning_rate/action_model': 5.890681495406151e-06, 'epoch': 2.71}) +[2026-06-18 02:47:46,245][starVLA.training.train_starvla][INFO] - Step 6575, Loss: {'train/loss': 0.15698032081127167, 'train/grad_norm_pre_clip': 26.310291290283203, 'timing/data': 0.025068250484764576, 'timing/model': 0.9319348819553852, 'learning_rate/qwen_vl_interface': 5.886517840979015e-07, 'learning_rate/action_model': 5.886517840979016e-06, 'epoch': 2.71}) +[2026-06-18 02:47:50,266][starVLA.training.train_starvla][INFO] - Step 6576, Loss: {'train/loss': 0.19078391790390015, 'train/grad_norm_pre_clip': 41.859336853027344, 'timing/data': 0.024591808207333088, 'timing/model': 0.9408403867855668, 'learning_rate/qwen_vl_interface': 5.882363849576889e-07, 'learning_rate/action_model': 5.882363849576889e-06, 'epoch': 2.71}) +[2026-06-18 02:47:54,128][starVLA.training.train_starvla][INFO] - Step 6577, Loss: {'train/loss': 0.13541710376739502, 'train/grad_norm_pre_clip': 15.310381889343262, 'timing/data': 0.031336892396211624, 'timing/model': 0.9436866817995906, 'learning_rate/qwen_vl_interface': 5.8782195220609e-07, 'learning_rate/action_model': 5.878219522060899e-06, 'epoch': 2.71}) +[2026-06-18 02:47:57,955][starVLA.training.train_starvla][INFO] - Step 6578, Loss: {'train/loss': 0.1245686337351799, 'train/grad_norm_pre_clip': 21.907384872436523, 'timing/data': 0.023468149825930595, 'timing/model': 0.924772965721786, 'learning_rate/qwen_vl_interface': 5.874084859290172e-07, 'learning_rate/action_model': 5.874084859290171e-06, 'epoch': 2.71}) +[2026-06-18 02:48:02,288][starVLA.training.train_starvla][INFO] - Step 6579, Loss: {'train/loss': 0.14583392441272736, 'train/grad_norm_pre_clip': 22.03123664855957, 'timing/data': 0.030700959265232086, 'timing/model': 1.1592796510085464, 'learning_rate/qwen_vl_interface': 5.869959862121822e-07, 'learning_rate/action_model': 5.869959862121822e-06, 'epoch': 2.71}) +[2026-06-18 02:48:06,117][starVLA.training.train_starvla][INFO] - Step 6580, Loss: {'train/loss': 0.07241798937320709, 'train/grad_norm_pre_clip': 12.048698425292969, 'timing/data': 0.030435962602496147, 'timing/model': 1.1124254493042827, 'learning_rate/qwen_vl_interface': 5.86584453141097e-07, 'learning_rate/action_model': 5.86584453141097e-06, 'epoch': 2.71}) +[2026-06-18 02:48:09,777][starVLA.training.train_starvla][INFO] - Step 6581, Loss: {'train/loss': 0.22601497173309326, 'train/grad_norm_pre_clip': 24.25567054748535, 'timing/data': 0.024819187819957733, 'timing/model': 0.9615150606259704, 'learning_rate/qwen_vl_interface': 5.861738868010723e-07, 'learning_rate/action_model': 5.861738868010723e-06, 'epoch': 2.71}) +[2026-06-18 02:48:13,663][starVLA.training.train_starvla][INFO] - Step 6582, Loss: {'train/loss': 0.2618340253829956, 'train/grad_norm_pre_clip': 25.648452758789062, 'timing/data': 0.023020402528345585, 'timing/model': 0.9412243729457259, 'learning_rate/qwen_vl_interface': 5.857642872772198e-07, 'learning_rate/action_model': 5.8576428727721986e-06, 'epoch': 2.71}) +[2026-06-18 02:48:17,639][starVLA.training.train_starvla][INFO] - Step 6583, Loss: {'train/loss': 0.167883962392807, 'train/grad_norm_pre_clip': 15.236822128295898, 'timing/data': 0.02893845271319151, 'timing/model': 0.9347809897735715, 'learning_rate/qwen_vl_interface': 5.853556546544491e-07, 'learning_rate/action_model': 5.8535565465444904e-06, 'epoch': 2.71}) +[2026-06-18 02:48:21,514][starVLA.training.train_starvla][INFO] - Step 6584, Loss: {'train/loss': 0.2198391705751419, 'train/grad_norm_pre_clip': 18.04432487487793, 'timing/data': 0.031108926981687546, 'timing/model': 0.9471072629094124, 'learning_rate/qwen_vl_interface': 5.849479890174708e-07, 'learning_rate/action_model': 5.8494798901747075e-06, 'epoch': 2.71}) +[2026-06-18 02:48:25,417][starVLA.training.train_starvla][INFO] - Step 6585, Loss: {'train/loss': 0.21511125564575195, 'train/grad_norm_pre_clip': 35.0693473815918, 'timing/data': 0.02872767671942711, 'timing/model': 0.9318636152893305, 'learning_rate/qwen_vl_interface': 5.845412904507948e-07, 'learning_rate/action_model': 5.845412904507948e-06, 'epoch': 2.71}) +[2026-06-18 02:48:29,215][starVLA.training.train_starvla][INFO] - Step 6586, Loss: {'train/loss': 0.3066578507423401, 'train/grad_norm_pre_clip': 25.155344009399414, 'timing/data': 0.031784070655703545, 'timing/model': 1.1081402665004134, 'learning_rate/qwen_vl_interface': 5.841355590387291e-07, 'learning_rate/action_model': 5.841355590387291e-06, 'epoch': 2.71}) +[2026-06-18 02:48:32,854][starVLA.training.train_starvla][INFO] - Step 6587, Loss: {'train/loss': 0.3214683532714844, 'train/grad_norm_pre_clip': 16.715286254882812, 'timing/data': 0.02933281473815441, 'timing/model': 0.9516098368912935, 'learning_rate/qwen_vl_interface': 5.837307948653821e-07, 'learning_rate/action_model': 5.837307948653821e-06, 'epoch': 2.71}) +[2026-06-18 02:48:36,650][starVLA.training.train_starvla][INFO] - Step 6588, Loss: {'train/loss': 0.4812048673629761, 'train/grad_norm_pre_clip': 43.31862258911133, 'timing/data': 0.027855181135237217, 'timing/model': 0.9287247266620398, 'learning_rate/qwen_vl_interface': 5.833269980146631e-07, 'learning_rate/action_model': 5.83326998014663e-06, 'epoch': 2.71}) +[2026-06-18 02:48:40,418][starVLA.training.train_starvla][INFO] - Step 6589, Loss: {'train/loss': 0.11288826912641525, 'train/grad_norm_pre_clip': 13.651822090148926, 'timing/data': 0.03339049406349659, 'timing/model': 0.957386551424861, 'learning_rate/qwen_vl_interface': 5.829241685702791e-07, 'learning_rate/action_model': 5.82924168570279e-06, 'epoch': 2.71}) +[2026-06-18 02:48:44,242][starVLA.training.train_starvla][INFO] - Step 6590, Loss: {'train/loss': 0.11938858032226562, 'train/grad_norm_pre_clip': 18.429807662963867, 'timing/data': 0.03902567643672228, 'timing/model': 0.960696361027658, 'learning_rate/qwen_vl_interface': 5.825223066157365e-07, 'learning_rate/action_model': 5.8252230661573645e-06, 'epoch': 2.71}) +[2026-06-18 02:48:48,034][starVLA.training.train_starvla][INFO] - Step 6591, Loss: {'train/loss': 0.33167675137519836, 'train/grad_norm_pre_clip': 46.40742492675781, 'timing/data': 0.022268233820796013, 'timing/model': 0.9359580986201763, 'learning_rate/qwen_vl_interface': 5.821214122343423e-07, 'learning_rate/action_model': 5.821214122343422e-06, 'epoch': 2.71}) +[2026-06-18 02:48:51,833][starVLA.training.train_starvla][INFO] - Step 6592, Loss: {'train/loss': 0.10634343326091766, 'train/grad_norm_pre_clip': 12.884231567382812, 'timing/data': 0.023366198875010014, 'timing/model': 0.9230442363768816, 'learning_rate/qwen_vl_interface': 5.817214855092024e-07, 'learning_rate/action_model': 5.817214855092024e-06, 'epoch': 2.71}) +[2026-06-18 02:48:55,624][starVLA.training.train_starvla][INFO] - Step 6593, Loss: {'train/loss': 0.14730127155780792, 'train/grad_norm_pre_clip': 20.031179428100586, 'timing/data': 0.02171473205089569, 'timing/model': 1.1282702395692468, 'learning_rate/qwen_vl_interface': 5.813225265232216e-07, 'learning_rate/action_model': 5.813225265232215e-06, 'epoch': 2.71}) +[2026-06-18 02:48:59,246][starVLA.training.train_starvla][INFO] - Step 6594, Loss: {'train/loss': 0.1436777114868164, 'train/grad_norm_pre_clip': 21.37388801574707, 'timing/data': 0.031246593222022057, 'timing/model': 0.9318651640787721, 'learning_rate/qwen_vl_interface': 5.809245353591048e-07, 'learning_rate/action_model': 5.8092453535910485e-06, 'epoch': 2.71}) +[2026-06-18 02:49:03,041][starVLA.training.train_starvla][INFO] - Step 6595, Loss: {'train/loss': 0.11574720591306686, 'train/grad_norm_pre_clip': 12.381826400756836, 'timing/data': 0.023969488218426704, 'timing/model': 0.9306598454713821, 'learning_rate/qwen_vl_interface': 5.805275120993561e-07, 'learning_rate/action_model': 5.805275120993561e-06, 'epoch': 2.71}) +[2026-06-18 02:49:06,812][starVLA.training.train_starvla][INFO] - Step 6596, Loss: {'train/loss': 0.591596245765686, 'train/grad_norm_pre_clip': 36.237815856933594, 'timing/data': 0.023906724527478218, 'timing/model': 0.9335018368437886, 'learning_rate/qwen_vl_interface': 5.801314568262787e-07, 'learning_rate/action_model': 5.801314568262787e-06, 'epoch': 2.72}) +[2026-06-18 02:49:10,590][starVLA.training.train_starvla][INFO] - Step 6597, Loss: {'train/loss': 0.2503370940685272, 'train/grad_norm_pre_clip': 22.1282901763916, 'timing/data': 0.02343730814754963, 'timing/model': 0.9435072559863329, 'learning_rate/qwen_vl_interface': 5.797363696219751e-07, 'learning_rate/action_model': 5.797363696219751e-06, 'epoch': 2.72}) +[2026-06-18 02:49:14,434][starVLA.training.train_starvla][INFO] - Step 6598, Loss: {'train/loss': 0.1427098661661148, 'train/grad_norm_pre_clip': 20.45860481262207, 'timing/data': 0.027055297046899796, 'timing/model': 0.9665940217673779, 'learning_rate/qwen_vl_interface': 5.793422505683474e-07, 'learning_rate/action_model': 5.793422505683474e-06, 'epoch': 2.72}) +[2026-06-18 02:49:18,261][starVLA.training.train_starvla][INFO] - Step 6599, Loss: {'train/loss': 0.2727276086807251, 'train/grad_norm_pre_clip': 25.072200775146484, 'timing/data': 0.02653734851628542, 'timing/model': 1.1383425015956163, 'learning_rate/qwen_vl_interface': 5.789490997470978e-07, 'learning_rate/action_model': 5.789490997470978e-06, 'epoch': 2.72}) +[2026-06-18 02:51:13,540][starVLA.training.train_starvla][INFO] - Step 6600, Loss: {'train/loss': 0.18074963986873627, 'train/grad_norm_pre_clip': 58.91825485229492, 'eval/loss': 0.42115089297294617, 'eval/latency_6/loss': 0.42115089297294617, 'eval/demon_fire_precision': 0.9548468582254499, 'eval/demon_fire_recall': 0.9773755656108597, 'eval/demon_fire_f1': 0.9659798754192621, 'eval/demon_fire_rate_model': 0.8811908736783528, 'eval/demon_fire_rate_teacher': 0.860879243183083, 'eval/demon_left_precision': 0.8918918918918919, 'eval/demon_left_recall': 0.7526315789473684, 'eval/demon_left_f1': 0.8163653663177926, 'eval/demon_left_rate_model': 0.13383416805787424, 'eval/demon_left_rate_teacher': 0.15859766277128548, 'eval/demon_right_precision': 0.8388704318936877, 'eval/demon_right_recall': 0.7757296466973886, 'eval/demon_right_f1': 0.8060654429369513, 'eval/demon_right_rate_model': 0.16750139120756816, 'eval/demon_right_rate_teacher': 0.18113522537562604, 'eval/demon_move_f1': 0.8112154046273718, 'eval/demon_cc_f1': 0.9136857180998091, 'eval/latency_6/demon_fire_precision': 0.9548468582254499, 'eval/latency_6/demon_fire_recall': 0.9773755656108597, 'eval/latency_6/demon_fire_f1': 0.9659798754192621, 'eval/latency_6/demon_fire_rate_model': 0.8811908736783528, 'eval/latency_6/demon_fire_rate_teacher': 0.860879243183083, 'eval/latency_6/demon_left_precision': 0.8918918918918919, 'eval/latency_6/demon_left_recall': 0.7526315789473684, 'eval/latency_6/demon_left_f1': 0.8163653663177926, 'eval/latency_6/demon_left_rate_model': 0.13383416805787424, 'eval/latency_6/demon_left_rate_teacher': 0.15859766277128548, 'eval/latency_6/demon_right_precision': 0.8388704318936877, 'eval/latency_6/demon_right_recall': 0.7757296466973886, 'eval/latency_6/demon_right_f1': 0.8060654429369513, 'eval/latency_6/demon_right_rate_model': 0.16750139120756816, 'eval/latency_6/demon_right_rate_teacher': 0.18113522537562604, 'eval/latency_6/demon_move_f1': 0.8112154046273718, 'eval/latency_6/demon_cc_f1': 0.9136857180998091, 'timing/data': 0.0374677749350667, 'timing/model': 1.1720308465883136, 'learning_rate/qwen_vl_interface': 5.785569172397254e-07, 'learning_rate/action_model': 5.785569172397254e-06, 'epoch': 2.72}) +[2026-06-18 02:51:13,541][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_6600_state +[2026-06-18 02:51:13,546][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 02:51:20,088][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_6600_state/model.safetensors +[2026-06-18 02:51:35,434][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_6600_state/optimizer.bin +[2026-06-18 02:51:35,434][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_6600_state/sampler.bin +[2026-06-18 02:51:35,434][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_6600_state/sampler_1.bin +[2026-06-18 02:51:35,438][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_6600_state/random_states_0.pkl +[2026-06-18 02:51:35,442][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 6600 +[2026-06-18 02:51:35,442][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 02:51:35,452][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 02:51:39,240][starVLA.training.train_starvla][INFO] - Step 6601, Loss: {'train/loss': 0.10558752715587616, 'train/grad_norm_pre_clip': 15.83371353149414, 'timing/data': 0.02633175253868103, 'timing/model': 1.0964927915483713, 'learning_rate/qwen_vl_interface': 5.781657031275306e-07, 'learning_rate/action_model': 5.781657031275306e-06, 'epoch': 2.72}) +[2026-06-18 02:51:43,006][starVLA.training.train_starvla][INFO] - Step 6602, Loss: {'train/loss': 0.22552917897701263, 'train/grad_norm_pre_clip': 23.429582595825195, 'timing/data': 0.0263356389477849, 'timing/model': 1.1113043213263154, 'learning_rate/qwen_vl_interface': 5.777754574916129e-07, 'learning_rate/action_model': 5.7777545749161285e-06, 'epoch': 2.72}) +[2026-06-18 02:51:46,585][starVLA.training.train_starvla][INFO] - Step 6603, Loss: {'train/loss': 0.2099209874868393, 'train/grad_norm_pre_clip': 21.93840789794922, 'timing/data': 0.026451142504811287, 'timing/model': 0.9271299261599779, 'learning_rate/qwen_vl_interface': 5.773861804128701e-07, 'learning_rate/action_model': 5.773861804128701e-06, 'epoch': 2.72}) +[2026-06-18 02:51:50,392][starVLA.training.train_starvla][INFO] - Step 6604, Loss: {'train/loss': 0.07987776398658752, 'train/grad_norm_pre_clip': 15.003292083740234, 'timing/data': 0.02995868306607008, 'timing/model': 0.9248964441940188, 'learning_rate/qwen_vl_interface': 5.769978719720011e-07, 'learning_rate/action_model': 5.76997871972001e-06, 'epoch': 2.72}) +[2026-06-18 02:51:54,228][starVLA.training.train_starvla][INFO] - Step 6605, Loss: {'train/loss': 0.21014511585235596, 'train/grad_norm_pre_clip': 19.83855438232422, 'timing/data': 0.03229731973260641, 'timing/model': 0.9206528533250093, 'learning_rate/qwen_vl_interface': 5.766105322495002e-07, 'learning_rate/action_model': 5.766105322495002e-06, 'epoch': 2.72}) +[2026-06-18 02:51:57,909][starVLA.training.train_starvla][INFO] - Step 6606, Loss: {'train/loss': 0.06646772474050522, 'train/grad_norm_pre_clip': 12.13346004486084, 'timing/data': 0.0269218347966671, 'timing/model': 0.9194483645260334, 'learning_rate/qwen_vl_interface': 5.762241613256654e-07, 'learning_rate/action_model': 5.762241613256654e-06, 'epoch': 2.72}) +[2026-06-18 02:52:01,666][starVLA.training.train_starvla][INFO] - Step 6607, Loss: {'train/loss': 0.4243430495262146, 'train/grad_norm_pre_clip': 24.418869018554688, 'timing/data': 0.02762338612228632, 'timing/model': 0.9269736679270864, 'learning_rate/qwen_vl_interface': 5.758387592805916e-07, 'learning_rate/action_model': 5.758387592805915e-06, 'epoch': 2.72}) +[2026-06-18 02:52:05,486][starVLA.training.train_starvla][INFO] - Step 6608, Loss: {'train/loss': 0.535759449005127, 'train/grad_norm_pre_clip': 61.15325164794922, 'timing/data': 0.029852600768208504, 'timing/model': 1.1368457851931453, 'learning_rate/qwen_vl_interface': 5.754543261941719e-07, 'learning_rate/action_model': 5.754543261941718e-06, 'epoch': 2.72}) +[2026-06-18 02:52:09,086][starVLA.training.train_starvla][INFO] - Step 6609, Loss: {'train/loss': 0.3205089271068573, 'train/grad_norm_pre_clip': 36.86861038208008, 'timing/data': 0.026097235269844532, 'timing/model': 0.9348409734666348, 'learning_rate/qwen_vl_interface': 5.750708621461011e-07, 'learning_rate/action_model': 5.75070862146101e-06, 'epoch': 2.72}) +[2026-06-18 02:52:13,083][starVLA.training.train_starvla][INFO] - Step 6610, Loss: {'train/loss': 0.30615532398223877, 'train/grad_norm_pre_clip': 25.33241844177246, 'timing/data': 0.026835757307708263, 'timing/model': 0.936364246532321, 'learning_rate/qwen_vl_interface': 5.746883672158705e-07, 'learning_rate/action_model': 5.746883672158705e-06, 'epoch': 2.72}) +[2026-06-18 02:52:16,918][starVLA.training.train_starvla][INFO] - Step 6611, Loss: {'train/loss': 0.11016473174095154, 'train/grad_norm_pre_clip': 16.43182945251465, 'timing/data': 0.025411161594092846, 'timing/model': 0.92115169018507, 'learning_rate/qwen_vl_interface': 5.743068414827726e-07, 'learning_rate/action_model': 5.743068414827726e-06, 'epoch': 2.72}) +[2026-06-18 02:52:20,680][starVLA.training.train_starvla][INFO] - Step 6612, Loss: {'train/loss': 0.4416254460811615, 'train/grad_norm_pre_clip': 35.16314697265625, 'timing/data': 0.041638909839093685, 'timing/model': 0.913028983399272, 'learning_rate/qwen_vl_interface': 5.739262850258968e-07, 'learning_rate/action_model': 5.7392628502589684e-06, 'epoch': 2.72}) +[2026-06-18 02:52:24,475][starVLA.training.train_starvla][INFO] - Step 6613, Loss: {'train/loss': 0.1742684245109558, 'train/grad_norm_pre_clip': 22.122472763061523, 'timing/data': 0.04508699383586645, 'timing/model': 0.9236004846170545, 'learning_rate/qwen_vl_interface': 5.735466979241352e-07, 'learning_rate/action_model': 5.735466979241352e-06, 'epoch': 2.72}) +[2026-06-18 02:52:28,250][starVLA.training.train_starvla][INFO] - Step 6614, Loss: {'train/loss': 0.45047375559806824, 'train/grad_norm_pre_clip': 31.95514678955078, 'timing/data': 0.03186899796128273, 'timing/model': 1.0973598919808865, 'learning_rate/qwen_vl_interface': 5.731680802561744e-07, 'learning_rate/action_model': 5.731680802561745e-06, 'epoch': 2.72}) +[2026-06-18 02:52:31,866][starVLA.training.train_starvla][INFO] - Step 6615, Loss: {'train/loss': 0.24447017908096313, 'train/grad_norm_pre_clip': 22.423959732055664, 'timing/data': 0.02831332851201296, 'timing/model': 0.9313405556604266, 'learning_rate/qwen_vl_interface': 5.727904321005031e-07, 'learning_rate/action_model': 5.727904321005031e-06, 'epoch': 2.72}) +[2026-06-18 02:52:35,624][starVLA.training.train_starvla][INFO] - Step 6616, Loss: {'train/loss': 0.6071609854698181, 'train/grad_norm_pre_clip': 28.23429298400879, 'timing/data': 0.036045171320438385, 'timing/model': 0.9281845251098275, 'learning_rate/qwen_vl_interface': 5.724137535354077e-07, 'learning_rate/action_model': 5.724137535354077e-06, 'epoch': 2.72}) +[2026-06-18 02:52:39,400][starVLA.training.train_starvla][INFO] - Step 6617, Loss: {'train/loss': 0.39867734909057617, 'train/grad_norm_pre_clip': 30.34807777404785, 'timing/data': 0.023651384748518467, 'timing/model': 0.9290076578035951, 'learning_rate/qwen_vl_interface': 5.720380446389748e-07, 'learning_rate/action_model': 5.720380446389748e-06, 'epoch': 2.72}) +[2026-06-18 02:52:43,173][starVLA.training.train_starvla][INFO] - Step 6618, Loss: {'train/loss': 0.4987790286540985, 'train/grad_norm_pre_clip': 32.13547897338867, 'timing/data': 0.03875171113759279, 'timing/model': 0.9264503680169582, 'learning_rate/qwen_vl_interface': 5.716633054890897e-07, 'learning_rate/action_model': 5.716633054890896e-06, 'epoch': 2.72}) +[2026-06-18 02:52:46,982][starVLA.training.train_starvla][INFO] - Step 6619, Loss: {'train/loss': 0.13240499794483185, 'train/grad_norm_pre_clip': 11.621646881103516, 'timing/data': 0.032114146277308464, 'timing/model': 0.9351000562310219, 'learning_rate/qwen_vl_interface': 5.712895361634345e-07, 'learning_rate/action_model': 5.712895361634345e-06, 'epoch': 2.72}) +[2026-06-18 02:52:50,795][starVLA.training.train_starvla][INFO] - Step 6620, Loss: {'train/loss': 0.06099232658743858, 'train/grad_norm_pre_clip': 9.135544776916504, 'timing/data': 0.028626921586692333, 'timing/model': 0.9288278464227915, 'learning_rate/qwen_vl_interface': 5.709167367394936e-07, 'learning_rate/action_model': 5.709167367394936e-06, 'epoch': 2.73}) +[2026-06-18 02:52:54,572][starVLA.training.train_starvla][INFO] - Step 6621, Loss: {'train/loss': 0.2733958959579468, 'train/grad_norm_pre_clip': 23.5087947845459, 'timing/data': 0.029104351066052914, 'timing/model': 1.1196066504344344, 'learning_rate/qwen_vl_interface': 5.705449072945479e-07, 'learning_rate/action_model': 5.705449072945479e-06, 'epoch': 2.73}) +[2026-06-18 02:52:58,185][starVLA.training.train_starvla][INFO] - Step 6622, Loss: {'train/loss': 0.04548943415284157, 'train/grad_norm_pre_clip': 6.9959635734558105, 'timing/data': 0.024854875169694424, 'timing/model': 0.9340596394613385, 'learning_rate/qwen_vl_interface': 5.70174047905678e-07, 'learning_rate/action_model': 5.701740479056779e-06, 'epoch': 2.73}) +[2026-06-18 02:53:01,951][starVLA.training.train_starvla][INFO] - Step 6623, Loss: {'train/loss': 0.16408826410770416, 'train/grad_norm_pre_clip': 14.235967636108398, 'timing/data': 0.030405775643885136, 'timing/model': 0.93389599211514, 'learning_rate/qwen_vl_interface': 5.69804158649764e-07, 'learning_rate/action_model': 5.6980415864976405e-06, 'epoch': 2.73}) +[2026-06-18 02:53:05,710][starVLA.training.train_starvla][INFO] - Step 6624, Loss: {'train/loss': 0.3555949330329895, 'train/grad_norm_pre_clip': 23.845443725585938, 'timing/data': 0.025854237377643585, 'timing/model': 0.9294759090989828, 'learning_rate/qwen_vl_interface': 5.69435239603484e-07, 'learning_rate/action_model': 5.69435239603484e-06, 'epoch': 2.73}) +[2026-06-18 02:53:09,486][starVLA.training.train_starvla][INFO] - Step 6625, Loss: {'train/loss': 0.32651540637016296, 'train/grad_norm_pre_clip': 32.90363311767578, 'timing/data': 0.025058744475245476, 'timing/model': 0.9294314160943031, 'learning_rate/qwen_vl_interface': 5.690672908433149e-07, 'learning_rate/action_model': 5.690672908433149e-06, 'epoch': 2.73}) +[2026-06-18 02:53:13,412][starVLA.training.train_starvla][INFO] - Step 6626, Loss: {'train/loss': 0.10254041850566864, 'train/grad_norm_pre_clip': 16.855817794799805, 'timing/data': 0.024180615320801735, 'timing/model': 0.933436743915081, 'learning_rate/qwen_vl_interface': 5.68700312445534e-07, 'learning_rate/action_model': 5.687003124455339e-06, 'epoch': 2.73}) +[2026-06-18 02:53:17,240][starVLA.training.train_starvla][INFO] - Step 6627, Loss: {'train/loss': 0.2695763111114502, 'train/grad_norm_pre_clip': 19.555253982543945, 'timing/data': 0.029511064291000366, 'timing/model': 1.1302902344614267, 'learning_rate/qwen_vl_interface': 5.683343044862151e-07, 'learning_rate/action_model': 5.68334304486215e-06, 'epoch': 2.73}) +[2026-06-18 02:53:21,050][starVLA.training.train_starvla][INFO] - Step 6628, Loss: {'train/loss': 0.05061573162674904, 'train/grad_norm_pre_clip': 8.095004081726074, 'timing/data': 0.031177987344563007, 'timing/model': 1.136426049284637, 'learning_rate/qwen_vl_interface': 5.679692670412324e-07, 'learning_rate/action_model': 5.679692670412324e-06, 'epoch': 2.73}) +[2026-06-18 02:53:24,640][starVLA.training.train_starvla][INFO] - Step 6629, Loss: {'train/loss': 0.3369077742099762, 'train/grad_norm_pre_clip': 24.65428924560547, 'timing/data': 0.0236115213483572, 'timing/model': 0.9271937496960163, 'learning_rate/qwen_vl_interface': 5.67605200186259e-07, 'learning_rate/action_model': 5.67605200186259e-06, 'epoch': 2.73}) +[2026-06-18 02:53:28,531][starVLA.training.train_starvla][INFO] - Step 6630, Loss: {'train/loss': 0.15020808577537537, 'train/grad_norm_pre_clip': 15.290911674499512, 'timing/data': 0.037561274133622646, 'timing/model': 0.9436140572652221, 'learning_rate/qwen_vl_interface': 5.67242103996766e-07, 'learning_rate/action_model': 5.67242103996766e-06, 'epoch': 2.73}) +[2026-06-18 02:53:32,419][starVLA.training.train_starvla][INFO] - Step 6631, Loss: {'train/loss': 0.08323874324560165, 'train/grad_norm_pre_clip': 13.05225658416748, 'timing/data': 0.030167797580361366, 'timing/model': 0.9426975455135107, 'learning_rate/qwen_vl_interface': 5.668799785480242e-07, 'learning_rate/action_model': 5.668799785480242e-06, 'epoch': 2.73}) +[2026-06-18 02:53:36,252][starVLA.training.train_starvla][INFO] - Step 6632, Loss: {'train/loss': 0.15346337854862213, 'train/grad_norm_pre_clip': 14.725411415100098, 'timing/data': 0.030547965317964554, 'timing/model': 0.9409855101257563, 'learning_rate/qwen_vl_interface': 5.665188239151019e-07, 'learning_rate/action_model': 5.665188239151018e-06, 'epoch': 2.73}) +[2026-06-18 02:53:40,035][starVLA.training.train_starvla][INFO] - Step 6633, Loss: {'train/loss': 0.3708057105541229, 'train/grad_norm_pre_clip': 30.868854522705078, 'timing/data': 0.02522166632115841, 'timing/model': 0.932651424780488, 'learning_rate/qwen_vl_interface': 5.661586401728671e-07, 'learning_rate/action_model': 5.6615864017286705e-06, 'epoch': 2.73}) +[2026-06-18 02:53:43,798][starVLA.training.train_starvla][INFO] - Step 6634, Loss: {'train/loss': 0.1989322155714035, 'train/grad_norm_pre_clip': 23.480634689331055, 'timing/data': 0.02618226408958435, 'timing/model': 1.095666947774589, 'learning_rate/qwen_vl_interface': 5.657994273959857e-07, 'learning_rate/action_model': 5.657994273959856e-06, 'epoch': 2.73}) +[2026-06-18 02:53:47,409][starVLA.training.train_starvla][INFO] - Step 6635, Loss: {'train/loss': 0.15312211215496063, 'train/grad_norm_pre_clip': 84.18727111816406, 'timing/data': 0.027556639164686203, 'timing/model': 0.9239663165062666, 'learning_rate/qwen_vl_interface': 5.654411856589233e-07, 'learning_rate/action_model': 5.654411856589233e-06, 'epoch': 2.73}) +[2026-06-18 02:53:51,240][starVLA.training.train_starvla][INFO] - Step 6636, Loss: {'train/loss': 0.12237868458032608, 'train/grad_norm_pre_clip': 23.586776733398438, 'timing/data': 0.04722103849053383, 'timing/model': 0.9193633124232292, 'learning_rate/qwen_vl_interface': 5.650839150359444e-07, 'learning_rate/action_model': 5.650839150359444e-06, 'epoch': 2.73}) +[2026-06-18 02:53:54,962][starVLA.training.train_starvla][INFO] - Step 6637, Loss: {'train/loss': 0.2418215125799179, 'train/grad_norm_pre_clip': 21.93340492248535, 'timing/data': 0.028010236099362373, 'timing/model': 0.9327403241768479, 'learning_rate/qwen_vl_interface': 5.647276156011114e-07, 'learning_rate/action_model': 5.6472761560111135e-06, 'epoch': 2.73}) +[2026-06-18 02:53:58,740][starVLA.training.train_starvla][INFO] - Step 6638, Loss: {'train/loss': 0.09219637513160706, 'train/grad_norm_pre_clip': 12.207569122314453, 'timing/data': 0.023284821771085262, 'timing/model': 0.9335595211014152, 'learning_rate/qwen_vl_interface': 5.643722874282846e-07, 'learning_rate/action_model': 5.643722874282846e-06, 'epoch': 2.73}) +[2026-06-18 02:54:02,552][starVLA.training.train_starvla][INFO] - Step 6639, Loss: {'train/loss': 0.20890352129936218, 'train/grad_norm_pre_clip': 27.985675811767578, 'timing/data': 0.025468753650784492, 'timing/model': 0.9333030553534627, 'learning_rate/qwen_vl_interface': 5.640179305911253e-07, 'learning_rate/action_model': 5.640179305911253e-06, 'epoch': 2.73}) +[2026-06-18 02:54:06,402][starVLA.training.train_starvla][INFO] - Step 6640, Loss: {'train/loss': 0.27672451734542847, 'train/grad_norm_pre_clip': 21.549938201904297, 'timing/data': 0.023537468165159225, 'timing/model': 1.1836365135386586, 'learning_rate/qwen_vl_interface': 5.636645451630908e-07, 'learning_rate/action_model': 5.6366454516309084e-06, 'epoch': 2.73}) +[2026-06-18 02:54:10,241][starVLA.training.train_starvla][INFO] - Step 6641, Loss: {'train/loss': 0.08092436194419861, 'train/grad_norm_pre_clip': 8.267370223999023, 'timing/data': 0.03013024292886257, 'timing/model': 1.109902175143361, 'learning_rate/qwen_vl_interface': 5.63312131217439e-07, 'learning_rate/action_model': 5.633121312174389e-06, 'epoch': 2.73}) +[2026-06-18 02:54:13,934][starVLA.training.train_starvla][INFO] - Step 6642, Loss: {'train/loss': 0.1524927318096161, 'train/grad_norm_pre_clip': 17.415273666381836, 'timing/data': 0.023374758660793304, 'timing/model': 0.9415230732411146, 'learning_rate/qwen_vl_interface': 5.629606888272253e-07, 'learning_rate/action_model': 5.629606888272252e-06, 'epoch': 2.73}) +[2026-06-18 02:54:17,971][starVLA.training.train_starvla][INFO] - Step 6643, Loss: {'train/loss': 0.14337365329265594, 'train/grad_norm_pre_clip': 14.4020357131958, 'timing/data': 0.024555726908147335, 'timing/model': 1.1640019277110696, 'learning_rate/qwen_vl_interface': 5.626102180653046e-07, 'learning_rate/action_model': 5.626102180653045e-06, 'epoch': 2.73}) +[2026-06-18 02:54:21,815][starVLA.training.train_starvla][INFO] - Step 6644, Loss: {'train/loss': 0.18473289906978607, 'train/grad_norm_pre_clip': 23.915672302246094, 'timing/data': 0.03190773446112871, 'timing/model': 0.9448471227660775, 'learning_rate/qwen_vl_interface': 5.62260719004329e-07, 'learning_rate/action_model': 5.6226071900432895e-06, 'epoch': 2.73}) +[2026-06-18 02:54:25,618][starVLA.training.train_starvla][INFO] - Step 6645, Loss: {'train/loss': 0.260287344455719, 'train/grad_norm_pre_clip': 33.849090576171875, 'timing/data': 0.022618712857365608, 'timing/model': 0.9449011059477925, 'learning_rate/qwen_vl_interface': 5.619121917167504e-07, 'learning_rate/action_model': 5.619121917167505e-06, 'epoch': 2.74}) +[2026-06-18 02:54:29,433][starVLA.training.train_starvla][INFO] - Step 6646, Loss: {'train/loss': 0.16329847276210785, 'train/grad_norm_pre_clip': 19.823854446411133, 'timing/data': 0.021670294925570488, 'timing/model': 0.939578709192574, 'learning_rate/qwen_vl_interface': 5.615646362748199e-07, 'learning_rate/action_model': 5.6156463627482e-06, 'epoch': 2.74}) +[2026-06-18 02:54:33,362][starVLA.training.train_starvla][INFO] - Step 6647, Loss: {'train/loss': 0.2298668920993805, 'train/grad_norm_pre_clip': 20.070632934570312, 'timing/data': 0.026138578541576862, 'timing/model': 1.2043543439358473, 'learning_rate/qwen_vl_interface': 5.612180527505854e-07, 'learning_rate/action_model': 5.612180527505854e-06, 'epoch': 2.74}) +[2026-06-18 02:54:38,039][starVLA.training.train_starvla][INFO] - Step 6648, Loss: {'train/loss': 0.12850870192050934, 'train/grad_norm_pre_clip': 14.70435619354248, 'timing/data': 0.02647059876471758, 'timing/model': 0.9310132106766105, 'learning_rate/qwen_vl_interface': 5.608724412158931e-07, 'learning_rate/action_model': 5.608724412158931e-06, 'epoch': 2.74}) +[2026-06-18 02:54:41,911][starVLA.training.train_starvla][INFO] - Step 6649, Loss: {'train/loss': 0.10097168385982513, 'train/grad_norm_pre_clip': 11.578935623168945, 'timing/data': 0.02728383243083954, 'timing/model': 0.9714712630957365, 'learning_rate/qwen_vl_interface': 5.605278017423897e-07, 'learning_rate/action_model': 5.605278017423897e-06, 'epoch': 2.74}) +[2026-06-18 02:54:45,700][starVLA.training.train_starvla][INFO] - Step 6650, Loss: {'train/loss': 0.09830259531736374, 'train/grad_norm_pre_clip': 14.194430351257324, 'timing/data': 0.027300044894218445, 'timing/model': 0.929647782817483, 'learning_rate/qwen_vl_interface': 5.601841344015193e-07, 'learning_rate/action_model': 5.601841344015193e-06, 'epoch': 2.74}) +[2026-06-18 02:54:50,052][starVLA.training.train_starvla][INFO] - Step 6651, Loss: {'train/loss': 0.11919591575860977, 'train/grad_norm_pre_clip': 20.772865295410156, 'timing/data': 0.02294764295220375, 'timing/model': 1.129315808415413, 'learning_rate/qwen_vl_interface': 5.598414392645239e-07, 'learning_rate/action_model': 5.598414392645239e-06, 'epoch': 2.74}) +[2026-06-18 02:54:53,899][starVLA.training.train_starvla][INFO] - Step 6652, Loss: {'train/loss': 0.28187575936317444, 'train/grad_norm_pre_clip': 32.014625549316406, 'timing/data': 0.02270070929080248, 'timing/model': 0.9251577639952302, 'learning_rate/qwen_vl_interface': 5.594997164024454e-07, 'learning_rate/action_model': 5.594997164024454e-06, 'epoch': 2.74}) +[2026-06-18 02:54:57,819][starVLA.training.train_starvla][INFO] - Step 6653, Loss: {'train/loss': 0.38165411353111267, 'train/grad_norm_pre_clip': 26.59482192993164, 'timing/data': 0.029506118036806583, 'timing/model': 1.0337309772148728, 'learning_rate/qwen_vl_interface': 5.591589658861228e-07, 'learning_rate/action_model': 5.591589658861228e-06, 'epoch': 2.74}) +[2026-06-18 02:55:01,640][starVLA.training.train_starvla][INFO] - Step 6654, Loss: {'train/loss': 0.5945584774017334, 'train/grad_norm_pre_clip': 89.44234466552734, 'timing/data': 0.02372540719807148, 'timing/model': 1.1242632092908025, 'learning_rate/qwen_vl_interface': 5.588191877861945e-07, 'learning_rate/action_model': 5.588191877861944e-06, 'epoch': 2.74}) +[2026-06-18 02:55:05,341][starVLA.training.train_starvla][INFO] - Step 6655, Loss: {'train/loss': 0.10599244385957718, 'train/grad_norm_pre_clip': 14.706655502319336, 'timing/data': 0.0344581063836813, 'timing/model': 0.9429193865507841, 'learning_rate/qwen_vl_interface': 5.584803821730961e-07, 'learning_rate/action_model': 5.58480382173096e-06, 'epoch': 2.74}) +[2026-06-18 02:55:09,396][starVLA.training.train_starvla][INFO] - Step 6656, Loss: {'train/loss': 0.203990638256073, 'train/grad_norm_pre_clip': 25.048805236816406, 'timing/data': 0.028348345309495926, 'timing/model': 1.0188960786908865, 'learning_rate/qwen_vl_interface': 5.581425491170627e-07, 'learning_rate/action_model': 5.581425491170626e-06, 'epoch': 2.74}) +[2026-06-18 02:55:13,332][starVLA.training.train_starvla][INFO] - Step 6657, Loss: {'train/loss': 0.25110092759132385, 'train/grad_norm_pre_clip': 46.05693817138672, 'timing/data': 0.026370970532298088, 'timing/model': 0.9412785647436976, 'learning_rate/qwen_vl_interface': 5.578056886881274e-07, 'learning_rate/action_model': 5.578056886881274e-06, 'epoch': 2.74}) +[2026-06-18 02:55:17,244][starVLA.training.train_starvla][INFO] - Step 6658, Loss: {'train/loss': 0.09899067133665085, 'train/grad_norm_pre_clip': 18.737241744995117, 'timing/data': 0.02928811777383089, 'timing/model': 0.943844985216856, 'learning_rate/qwen_vl_interface': 5.574698009561223e-07, 'learning_rate/action_model': 5.574698009561223e-06, 'epoch': 2.74}) +[2026-06-18 02:55:21,263][starVLA.training.train_starvla][INFO] - Step 6659, Loss: {'train/loss': 0.17735426127910614, 'train/grad_norm_pre_clip': 19.56304931640625, 'timing/data': 0.028748891316354275, 'timing/model': 0.9879683321341872, 'learning_rate/qwen_vl_interface': 5.571348859906763e-07, 'learning_rate/action_model': 5.571348859906764e-06, 'epoch': 2.74}) +[2026-06-18 02:55:25,115][starVLA.training.train_starvla][INFO] - Step 6660, Loss: {'train/loss': 0.09680648148059845, 'train/grad_norm_pre_clip': 15.22742748260498, 'timing/data': 0.03724260255694389, 'timing/model': 1.1026857364922762, 'learning_rate/qwen_vl_interface': 5.568009438612186e-07, 'learning_rate/action_model': 5.568009438612186e-06, 'epoch': 2.74}) +[2026-06-18 02:55:28,901][starVLA.training.train_starvla][INFO] - Step 6661, Loss: {'train/loss': 0.2639347016811371, 'train/grad_norm_pre_clip': 20.746429443359375, 'timing/data': 0.03253956511616707, 'timing/model': 1.0930524095892906, 'learning_rate/qwen_vl_interface': 5.564679746369755e-07, 'learning_rate/action_model': 5.5646797463697556e-06, 'epoch': 2.74}) +[2026-06-18 02:55:32,500][starVLA.training.train_starvla][INFO] - Step 6662, Loss: {'train/loss': 0.2370058298110962, 'train/grad_norm_pre_clip': 17.279844284057617, 'timing/data': 0.03969892207533121, 'timing/model': 0.9243411729112267, 'learning_rate/qwen_vl_interface': 5.561359783869712e-07, 'learning_rate/action_model': 5.5613597838697116e-06, 'epoch': 2.74}) +[2026-06-18 02:55:36,299][starVLA.training.train_starvla][INFO] - Step 6663, Loss: {'train/loss': 0.33874863386154175, 'train/grad_norm_pre_clip': 20.324234008789062, 'timing/data': 0.03831614926457405, 'timing/model': 0.9273364003747702, 'learning_rate/qwen_vl_interface': 5.558049551800291e-07, 'learning_rate/action_model': 5.558049551800292e-06, 'epoch': 2.74}) +[2026-06-18 02:55:40,089][starVLA.training.train_starvla][INFO] - Step 6664, Loss: {'train/loss': 0.21345844864845276, 'train/grad_norm_pre_clip': 20.027482986450195, 'timing/data': 0.03135216422379017, 'timing/model': 0.9251063084229827, 'learning_rate/qwen_vl_interface': 5.554749050847718e-07, 'learning_rate/action_model': 5.5547490508477176e-06, 'epoch': 2.74}) +[2026-06-18 02:55:43,904][starVLA.training.train_starvla][INFO] - Step 6665, Loss: {'train/loss': 0.24466513097286224, 'train/grad_norm_pre_clip': 17.224258422851562, 'timing/data': 0.028248369693756104, 'timing/model': 0.9215550757944584, 'learning_rate/qwen_vl_interface': 5.55145828169618e-07, 'learning_rate/action_model': 5.5514582816961805e-06, 'epoch': 2.74}) +[2026-06-18 02:55:47,681][starVLA.training.train_starvla][INFO] - Step 6666, Loss: {'train/loss': 0.5120886564254761, 'train/grad_norm_pre_clip': 35.06526184082031, 'timing/data': 0.029438096098601818, 'timing/model': 0.9206540817394853, 'learning_rate/qwen_vl_interface': 5.548177245027859e-07, 'learning_rate/action_model': 5.548177245027859e-06, 'epoch': 2.74}) +[2026-06-18 02:55:51,478][starVLA.training.train_starvla][INFO] - Step 6667, Loss: {'train/loss': 0.23584415018558502, 'train/grad_norm_pre_clip': 22.960811614990234, 'timing/data': 0.02426688466221094, 'timing/model': 1.11608119122684, 'learning_rate/qwen_vl_interface': 5.544905941522913e-07, 'learning_rate/action_model': 5.544905941522913e-06, 'epoch': 2.74}) +[2026-06-18 02:55:55,238][starVLA.training.train_starvla][INFO] - Step 6668, Loss: {'train/loss': 0.07705969363451004, 'train/grad_norm_pre_clip': 10.017704010009766, 'timing/data': 0.025362852960824966, 'timing/model': 1.099446576088667, 'learning_rate/qwen_vl_interface': 5.541644371859498e-07, 'learning_rate/action_model': 5.541644371859498e-06, 'epoch': 2.74}) +[2026-06-18 02:55:58,842][starVLA.training.train_starvla][INFO] - Step 6669, Loss: {'train/loss': 0.167258158326149, 'train/grad_norm_pre_clip': 19.036544799804688, 'timing/data': 0.023357982747256756, 'timing/model': 0.9226890131831169, 'learning_rate/qwen_vl_interface': 5.53839253671373e-07, 'learning_rate/action_model': 5.5383925367137305e-06, 'epoch': 2.75}) +[2026-06-18 02:56:02,645][starVLA.training.train_starvla][INFO] - Step 6670, Loss: {'train/loss': 0.1210247352719307, 'train/grad_norm_pre_clip': 15.848149299621582, 'timing/data': 0.024056293070316315, 'timing/model': 0.9424562267959118, 'learning_rate/qwen_vl_interface': 5.535150436759725e-07, 'learning_rate/action_model': 5.535150436759725e-06, 'epoch': 2.75}) +[2026-06-18 02:56:06,537][starVLA.training.train_starvla][INFO] - Step 6671, Loss: {'train/loss': 0.2869108021259308, 'train/grad_norm_pre_clip': 33.43285369873047, 'timing/data': 0.030636228621006012, 'timing/model': 0.940306699834764, 'learning_rate/qwen_vl_interface': 5.531918072669566e-07, 'learning_rate/action_model': 5.531918072669565e-06, 'epoch': 2.75}) +[2026-06-18 02:56:10,343][starVLA.training.train_starvla][INFO] - Step 6672, Loss: {'train/loss': 0.19773991405963898, 'train/grad_norm_pre_clip': 24.90615463256836, 'timing/data': 0.02262208331376314, 'timing/model': 0.9283363372087479, 'learning_rate/qwen_vl_interface': 5.528695445113336e-07, 'learning_rate/action_model': 5.528695445113335e-06, 'epoch': 2.75}) +[2026-06-18 02:56:14,182][starVLA.training.train_starvla][INFO] - Step 6673, Loss: {'train/loss': 0.09400409460067749, 'train/grad_norm_pre_clip': 11.848692893981934, 'timing/data': 0.025428883731365204, 'timing/model': 0.9318049624562263, 'learning_rate/qwen_vl_interface': 5.525482554759079e-07, 'learning_rate/action_model': 5.5254825547590795e-06, 'epoch': 2.75}) +[2026-06-18 02:56:18,258][starVLA.training.train_starvla][INFO] - Step 6674, Loss: {'train/loss': 0.11511539667844772, 'train/grad_norm_pre_clip': 18.723440170288086, 'timing/data': 0.03204693831503391, 'timing/model': 1.1015924690291286, 'learning_rate/qwen_vl_interface': 5.522279402272834e-07, 'learning_rate/action_model': 5.522279402272834e-06, 'epoch': 2.75}) +[2026-06-18 02:56:22,093][starVLA.training.train_starvla][INFO] - Step 6675, Loss: {'train/loss': 0.12365639954805374, 'train/grad_norm_pre_clip': 29.39743995666504, 'timing/data': 0.02897947933524847, 'timing/model': 1.1536026513203979, 'learning_rate/qwen_vl_interface': 5.519085988318627e-07, 'learning_rate/action_model': 5.5190859883186265e-06, 'epoch': 2.75}) +[2026-06-18 02:56:25,921][starVLA.training.train_starvla][INFO] - Step 6676, Loss: {'train/loss': 0.3915109932422638, 'train/grad_norm_pre_clip': 37.1236457824707, 'timing/data': 0.024867326021194458, 'timing/model': 0.9247848866507411, 'learning_rate/qwen_vl_interface': 5.515902313558443e-07, 'learning_rate/action_model': 5.515902313558444e-06, 'epoch': 2.75}) +[2026-06-18 02:56:29,691][starVLA.training.train_starvla][INFO] - Step 6677, Loss: {'train/loss': 0.5855699777603149, 'train/grad_norm_pre_clip': 28.940370559692383, 'timing/data': 0.02703683450818062, 'timing/model': 0.9239655192941427, 'learning_rate/qwen_vl_interface': 5.512728378652266e-07, 'learning_rate/action_model': 5.512728378652266e-06, 'epoch': 2.75}) +[2026-06-18 02:56:33,436][starVLA.training.train_starvla][INFO] - Step 6678, Loss: {'train/loss': 0.377055287361145, 'train/grad_norm_pre_clip': 36.893375396728516, 'timing/data': 0.02750052697956562, 'timing/model': 0.9288101047277451, 'learning_rate/qwen_vl_interface': 5.509564184258055e-07, 'learning_rate/action_model': 5.509564184258055e-06, 'epoch': 2.75}) +[2026-06-18 02:56:37,268][starVLA.training.train_starvla][INFO] - Step 6679, Loss: {'train/loss': 0.057958465069532394, 'train/grad_norm_pre_clip': 10.168667793273926, 'timing/data': 0.033813479356467724, 'timing/model': 0.9740329086780548, 'learning_rate/qwen_vl_interface': 5.506409731031754e-07, 'learning_rate/action_model': 5.506409731031755e-06, 'epoch': 2.75}) +[2026-06-18 02:56:41,072][starVLA.training.train_starvla][INFO] - Step 6680, Loss: {'train/loss': 0.29901808500289917, 'train/grad_norm_pre_clip': 21.747997283935547, 'timing/data': 0.03606030624359846, 'timing/model': 1.1180215291678905, 'learning_rate/qwen_vl_interface': 5.503265019627289e-07, 'learning_rate/action_model': 5.503265019627289e-06, 'epoch': 2.75}) +[2026-06-18 02:56:44,846][starVLA.training.train_starvla][INFO] - Step 6681, Loss: {'train/loss': 0.28312838077545166, 'train/grad_norm_pre_clip': 23.08828353881836, 'timing/data': 0.029059313237667084, 'timing/model': 1.0612086430191994, 'learning_rate/qwen_vl_interface': 5.500130050696548e-07, 'learning_rate/action_model': 5.500130050696548e-06, 'epoch': 2.75}) +[2026-06-18 02:56:48,524][starVLA.training.train_starvla][INFO] - Step 6682, Loss: {'train/loss': 0.030841533094644547, 'train/grad_norm_pre_clip': 6.825750827789307, 'timing/data': 0.035004026256501675, 'timing/model': 0.954067817889154, 'learning_rate/qwen_vl_interface': 5.497004824889424e-07, 'learning_rate/action_model': 5.497004824889423e-06, 'epoch': 2.75}) +[2026-06-18 02:56:52,415][starVLA.training.train_starvla][INFO] - Step 6683, Loss: {'train/loss': 0.24301356077194214, 'train/grad_norm_pre_clip': 22.56678009033203, 'timing/data': 0.035024262964725494, 'timing/model': 0.9536449443548918, 'learning_rate/qwen_vl_interface': 5.493889342853779e-07, 'learning_rate/action_model': 5.493889342853778e-06, 'epoch': 2.75}) +[2026-06-18 02:56:56,220][starVLA.training.train_starvla][INFO] - Step 6684, Loss: {'train/loss': 0.4328441023826599, 'train/grad_norm_pre_clip': 26.03095817565918, 'timing/data': 0.029799912124872208, 'timing/model': 0.937985859811306, 'learning_rate/qwen_vl_interface': 5.490783605235451e-07, 'learning_rate/action_model': 5.49078360523545e-06, 'epoch': 2.75}) +[2026-06-18 02:57:00,055][starVLA.training.train_starvla][INFO] - Step 6685, Loss: {'train/loss': 0.08307957649230957, 'train/grad_norm_pre_clip': 15.949585914611816, 'timing/data': 0.030798343010246754, 'timing/model': 0.9474322535097599, 'learning_rate/qwen_vl_interface': 5.487687612678265e-07, 'learning_rate/action_model': 5.4876876126782645e-06, 'epoch': 2.75}) +[2026-06-18 02:57:03,853][starVLA.training.train_starvla][INFO] - Step 6686, Loss: {'train/loss': 0.03480272367596626, 'train/grad_norm_pre_clip': 6.173419952392578, 'timing/data': 0.03042236715555191, 'timing/model': 0.92682518158108, 'learning_rate/qwen_vl_interface': 5.484601365824032e-07, 'learning_rate/action_model': 5.4846013658240315e-06, 'epoch': 2.75}) +[2026-06-18 02:57:07,605][starVLA.training.train_starvla][INFO] - Step 6687, Loss: {'train/loss': 0.16597650945186615, 'train/grad_norm_pre_clip': 19.33797264099121, 'timing/data': 0.024582126177847385, 'timing/model': 1.0935505805537105, 'learning_rate/qwen_vl_interface': 5.481524865312518e-07, 'learning_rate/action_model': 5.481524865312518e-06, 'epoch': 2.75}) +[2026-06-18 02:57:11,410][starVLA.training.train_starvla][INFO] - Step 6688, Loss: {'train/loss': 0.2479867935180664, 'train/grad_norm_pre_clip': 19.335372924804688, 'timing/data': 0.0312978932633996, 'timing/model': 1.1070298040285707, 'learning_rate/qwen_vl_interface': 5.478458111781496e-07, 'learning_rate/action_model': 5.478458111781496e-06, 'epoch': 2.75}) +[2026-06-18 02:57:14,988][starVLA.training.train_starvla][INFO] - Step 6689, Loss: {'train/loss': 0.048711489886045456, 'train/grad_norm_pre_clip': 8.781466484069824, 'timing/data': 0.026736900210380554, 'timing/model': 0.9204802066087723, 'learning_rate/qwen_vl_interface': 5.475401105866712e-07, 'learning_rate/action_model': 5.4754011058667114e-06, 'epoch': 2.75}) +[2026-06-18 02:57:18,742][starVLA.training.train_starvla][INFO] - Step 6690, Loss: {'train/loss': 0.061399154365062714, 'train/grad_norm_pre_clip': 7.739278316497803, 'timing/data': 0.024676427245140076, 'timing/model': 0.920489483512938, 'learning_rate/qwen_vl_interface': 5.47235384820187e-07, 'learning_rate/action_model': 5.47235384820187e-06, 'epoch': 2.75}) +[2026-06-18 02:57:22,499][starVLA.training.train_starvla][INFO] - Step 6691, Loss: {'train/loss': 0.13222859799861908, 'train/grad_norm_pre_clip': 18.915834426879883, 'timing/data': 0.031051366589963436, 'timing/model': 0.9349024929106236, 'learning_rate/qwen_vl_interface': 5.469316339418681e-07, 'learning_rate/action_model': 5.469316339418681e-06, 'epoch': 2.75}) +[2026-06-18 02:57:26,213][starVLA.training.train_starvla][INFO] - Step 6692, Loss: {'train/loss': 0.19184474647045135, 'train/grad_norm_pre_clip': 30.293935775756836, 'timing/data': 0.03598839044570923, 'timing/model': 0.9226778037846088, 'learning_rate/qwen_vl_interface': 5.466288580146825e-07, 'learning_rate/action_model': 5.466288580146825e-06, 'epoch': 2.75}) +[2026-06-18 02:57:29,958][starVLA.training.train_starvla][INFO] - Step 6693, Loss: {'train/loss': 0.1773354709148407, 'train/grad_norm_pre_clip': 21.775075912475586, 'timing/data': 0.029416638426482677, 'timing/model': 0.9220642996951938, 'learning_rate/qwen_vl_interface': 5.463270571013954e-07, 'learning_rate/action_model': 5.463270571013954e-06, 'epoch': 2.76}) +[2026-06-18 02:57:33,739][starVLA.training.train_starvla][INFO] - Step 6694, Loss: {'train/loss': 0.14842984080314636, 'train/grad_norm_pre_clip': 16.70514488220215, 'timing/data': 0.03009657096117735, 'timing/model': 1.128266429528594, 'learning_rate/qwen_vl_interface': 5.460262312645709e-07, 'learning_rate/action_model': 5.460262312645709e-06, 'epoch': 2.76}) +[2026-06-18 02:57:37,413][starVLA.training.train_starvla][INFO] - Step 6695, Loss: {'train/loss': 0.22490884363651276, 'train/grad_norm_pre_clip': 16.41191864013672, 'timing/data': 0.023574039340019226, 'timing/model': 1.020218658261001, 'learning_rate/qwen_vl_interface': 5.457263805665707e-07, 'learning_rate/action_model': 5.457263805665707e-06, 'epoch': 2.76}) +[2026-06-18 02:57:40,986][starVLA.training.train_starvla][INFO] - Step 6696, Loss: {'train/loss': 0.38211962580680847, 'train/grad_norm_pre_clip': 29.7346134185791, 'timing/data': 0.030211763456463814, 'timing/model': 0.9249653574079275, 'learning_rate/qwen_vl_interface': 5.454275050695529e-07, 'learning_rate/action_model': 5.454275050695529e-06, 'epoch': 2.76}) +[2026-06-18 02:57:44,696][starVLA.training.train_starvla][INFO] - Step 6697, Loss: {'train/loss': 0.1286662220954895, 'train/grad_norm_pre_clip': 14.336174964904785, 'timing/data': 0.024184328503906727, 'timing/model': 0.9234765451401472, 'learning_rate/qwen_vl_interface': 5.451296048354764e-07, 'learning_rate/action_model': 5.4512960483547635e-06, 'epoch': 2.76}) +[2026-06-18 02:57:48,484][starVLA.training.train_starvla][INFO] - Step 6698, Loss: {'train/loss': 0.16072335839271545, 'train/grad_norm_pre_clip': 18.24897575378418, 'timing/data': 0.02227176446467638, 'timing/model': 0.9582425123080611, 'learning_rate/qwen_vl_interface': 5.448326799260951e-07, 'learning_rate/action_model': 5.448326799260952e-06, 'epoch': 2.76}) +[2026-06-18 02:57:52,463][starVLA.training.train_starvla][INFO] - Step 6699, Loss: {'train/loss': 0.17507918179035187, 'train/grad_norm_pre_clip': 16.79045295715332, 'timing/data': 0.02326454594731331, 'timing/model': 0.9382739355787635, 'learning_rate/qwen_vl_interface': 5.445367304029627e-07, 'learning_rate/action_model': 5.445367304029626e-06, 'epoch': 2.76}) +[2026-06-18 02:59:45,552][starVLA.training.train_starvla][INFO] - Step 6700, Loss: {'train/loss': 0.4108169376850128, 'train/grad_norm_pre_clip': 30.281797409057617, 'eval/loss': 0.4216782748699188, 'eval/latency_6/loss': 0.4216782748699188, 'eval/demon_fire_precision': 0.95141065830721, 'eval/demon_fire_recall': 0.9809308338720103, 'eval/demon_fire_f1': 0.9659452577975812, 'eval/demon_fire_rate_model': 0.887590428491931, 'eval/demon_fire_rate_teacher': 0.860879243183083, 'eval/demon_left_precision': 0.889795918367347, 'eval/demon_left_recall': 0.7649122807017544, 'eval/demon_left_f1': 0.8226415094339622, 'eval/demon_left_rate_model': 0.13633834168057873, 'eval/demon_left_rate_teacher': 0.15859766277128548, 'eval/demon_right_precision': 0.8244766505636071, 'eval/demon_right_recall': 0.7864823348694316, 'eval/demon_right_f1': 0.8050314465408805, 'eval/demon_right_rate_model': 0.17278797996661102, 'eval/demon_right_rate_teacher': 0.18113522537562604, 'eval/demon_move_f1': 0.8138364779874214, 'eval/demon_cc_f1': 0.9146593817303259, 'eval/latency_6/demon_fire_precision': 0.95141065830721, 'eval/latency_6/demon_fire_recall': 0.9809308338720103, 'eval/latency_6/demon_fire_f1': 0.9659452577975812, 'eval/latency_6/demon_fire_rate_model': 0.887590428491931, 'eval/latency_6/demon_fire_rate_teacher': 0.860879243183083, 'eval/latency_6/demon_left_precision': 0.889795918367347, 'eval/latency_6/demon_left_recall': 0.7649122807017544, 'eval/latency_6/demon_left_f1': 0.8226415094339622, 'eval/latency_6/demon_left_rate_model': 0.13633834168057873, 'eval/latency_6/demon_left_rate_teacher': 0.15859766277128548, 'eval/latency_6/demon_right_precision': 0.8244766505636071, 'eval/latency_6/demon_right_recall': 0.7864823348694316, 'eval/latency_6/demon_right_f1': 0.8050314465408805, 'eval/latency_6/demon_right_rate_model': 0.17278797996661102, 'eval/latency_6/demon_right_rate_teacher': 0.18113522537562604, 'eval/latency_6/demon_move_f1': 0.8138364779874214, 'eval/latency_6/demon_cc_f1': 0.9146593817303259, 'timing/data': 0.02874821238219738, 'timing/model': 1.14171670563519, 'learning_rate/qwen_vl_interface': 5.442417563274289e-07, 'learning_rate/action_model': 5.442417563274289e-06, 'epoch': 2.76}) +[2026-06-18 02:59:45,553][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_6700_state +[2026-06-18 02:59:45,558][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 02:59:52,025][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_6700_state/model.safetensors +[2026-06-18 03:00:07,127][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_6700_state/optimizer.bin +[2026-06-18 03:00:07,128][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_6700_state/sampler.bin +[2026-06-18 03:00:07,128][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_6700_state/sampler_1.bin +[2026-06-18 03:00:07,129][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_6700_state/random_states_0.pkl +[2026-06-18 03:00:07,133][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 6700 +[2026-06-18 03:00:07,133][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 03:00:07,145][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 03:00:12,110][starVLA.training.train_starvla][INFO] - Step 6701, Loss: {'train/loss': 0.316010981798172, 'train/grad_norm_pre_clip': 16.22330665588379, 'timing/data': 0.02753143571317196, 'timing/model': 2.0285691572353244, 'learning_rate/qwen_vl_interface': 5.439477577606429e-07, 'learning_rate/action_model': 5.4394775776064295e-06, 'epoch': 2.76}) +[2026-06-18 03:00:16,090][starVLA.training.train_starvla][INFO] - Step 6702, Loss: {'train/loss': 0.09962338954210281, 'train/grad_norm_pre_clip': 13.367707252502441, 'timing/data': 0.030296235345304012, 'timing/model': 0.9515260923653841, 'learning_rate/qwen_vl_interface': 5.436547347635506e-07, 'learning_rate/action_model': 5.436547347635506e-06, 'epoch': 2.76}) +[2026-06-18 03:00:19,918][starVLA.training.train_starvla][INFO] - Step 6703, Loss: {'train/loss': 0.24018515646457672, 'train/grad_norm_pre_clip': 15.272464752197266, 'timing/data': 0.024818926118314266, 'timing/model': 1.082700285129249, 'learning_rate/qwen_vl_interface': 5.433626873968962e-07, 'learning_rate/action_model': 5.433626873968962e-06, 'epoch': 2.76}) +[2026-06-18 03:00:23,617][starVLA.training.train_starvla][INFO] - Step 6704, Loss: {'train/loss': 0.19123518466949463, 'train/grad_norm_pre_clip': 28.06932830810547, 'timing/data': 0.03300714399665594, 'timing/model': 1.0459532774984837, 'learning_rate/qwen_vl_interface': 5.43071615721221e-07, 'learning_rate/action_model': 5.43071615721221e-06, 'epoch': 2.76}) +[2026-06-18 03:00:27,197][starVLA.training.train_starvla][INFO] - Step 6705, Loss: {'train/loss': 0.23522540926933289, 'train/grad_norm_pre_clip': 18.018352508544922, 'timing/data': 0.034708421677351, 'timing/model': 0.9219544501975179, 'learning_rate/qwen_vl_interface': 5.427815197968647e-07, 'learning_rate/action_model': 5.4278151979686465e-06, 'epoch': 2.76}) +[2026-06-18 03:00:31,047][starVLA.training.train_starvla][INFO] - Step 6706, Loss: {'train/loss': 0.20362859964370728, 'train/grad_norm_pre_clip': 24.985992431640625, 'timing/data': 0.025227708742022514, 'timing/model': 0.9210776425898075, 'learning_rate/qwen_vl_interface': 5.424923996839647e-07, 'learning_rate/action_model': 5.4249239968396466e-06, 'epoch': 2.76}) +[2026-06-18 03:00:34,786][starVLA.training.train_starvla][INFO] - Step 6707, Loss: {'train/loss': 0.2523551285266876, 'train/grad_norm_pre_clip': 14.254420280456543, 'timing/data': 0.02827169932425022, 'timing/model': 0.9277329109609127, 'learning_rate/qwen_vl_interface': 5.422042554424557e-07, 'learning_rate/action_model': 5.4220425544245565e-06, 'epoch': 2.76}) +[2026-06-18 03:00:38,473][starVLA.training.train_starvla][INFO] - Step 6708, Loss: {'train/loss': 0.06361131370067596, 'train/grad_norm_pre_clip': 14.919185638427734, 'timing/data': 0.025208299048244953, 'timing/model': 0.9219217393547297, 'learning_rate/qwen_vl_interface': 5.419170871320705e-07, 'learning_rate/action_model': 5.419170871320706e-06, 'epoch': 2.76}) +[2026-06-18 03:00:42,223][starVLA.training.train_starvla][INFO] - Step 6709, Loss: {'train/loss': 0.07116406410932541, 'train/grad_norm_pre_clip': 15.662165641784668, 'timing/data': 0.025650071911513805, 'timing/model': 0.9470646306872368, 'learning_rate/qwen_vl_interface': 5.41630894812339e-07, 'learning_rate/action_model': 5.41630894812339e-06, 'epoch': 2.76}) +[2026-06-18 03:00:46,026][starVLA.training.train_starvla][INFO] - Step 6710, Loss: {'train/loss': 0.2631612718105316, 'train/grad_norm_pre_clip': 22.610004425048828, 'timing/data': 0.02594457846134901, 'timing/model': 1.1315772533416748, 'learning_rate/qwen_vl_interface': 5.413456785425896e-07, 'learning_rate/action_model': 5.413456785425896e-06, 'epoch': 2.76}) +[2026-06-18 03:00:49,640][starVLA.training.train_starvla][INFO] - Step 6711, Loss: {'train/loss': 0.15412069857120514, 'train/grad_norm_pre_clip': 23.253427505493164, 'timing/data': 0.02226459886878729, 'timing/model': 0.9326091008260846, 'learning_rate/qwen_vl_interface': 5.410614383819474e-07, 'learning_rate/action_model': 5.4106143838194736e-06, 'epoch': 2.76}) +[2026-06-18 03:00:53,500][starVLA.training.train_starvla][INFO] - Step 6712, Loss: {'train/loss': 0.026511715725064278, 'train/grad_norm_pre_clip': 4.685103893280029, 'timing/data': 0.03510122653096914, 'timing/model': 0.9543171599507332, 'learning_rate/qwen_vl_interface': 5.407781743893364e-07, 'learning_rate/action_model': 5.407781743893364e-06, 'epoch': 2.76}) +[2026-06-18 03:00:57,349][starVLA.training.train_starvla][INFO] - Step 6713, Loss: {'train/loss': 0.21266844868659973, 'train/grad_norm_pre_clip': 19.766902923583984, 'timing/data': 0.029479230754077435, 'timing/model': 0.9574011424556375, 'learning_rate/qwen_vl_interface': 5.404958866234772e-07, 'learning_rate/action_model': 5.404958866234772e-06, 'epoch': 2.76}) +[2026-06-18 03:01:01,260][starVLA.training.train_starvla][INFO] - Step 6714, Loss: {'train/loss': 0.16073034703731537, 'train/grad_norm_pre_clip': 18.3336181640625, 'timing/data': 0.03643193747848272, 'timing/model': 0.9803078416734934, 'learning_rate/qwen_vl_interface': 5.402145751428884e-07, 'learning_rate/action_model': 5.402145751428884e-06, 'epoch': 2.76}) +[2026-06-18 03:01:08,509][starVLA.training.train_starvla][INFO] - Step 6715, Loss: {'train/loss': 0.1816975623369217, 'train/grad_norm_pre_clip': 20.467098236083984, 'timing/data': 0.03295171074569225, 'timing/model': 2.8150526490062475, 'learning_rate/qwen_vl_interface': 5.399342400058856e-07, 'learning_rate/action_model': 5.399342400058856e-06, 'epoch': 2.76}) +[2026-06-18 03:01:12,631][starVLA.training.train_starvla][INFO] - Step 6716, Loss: {'train/loss': 0.22754555940628052, 'train/grad_norm_pre_clip': 37.72997283935547, 'timing/data': 0.028510681353509426, 'timing/model': 1.1720208171755075, 'learning_rate/qwen_vl_interface': 5.396548812705838e-07, 'learning_rate/action_model': 5.396548812705839e-06, 'epoch': 2.76}) +[2026-06-18 03:01:17,023][starVLA.training.train_starvla][INFO] - Step 6717, Loss: {'train/loss': 0.07644330710172653, 'train/grad_norm_pre_clip': 17.045269012451172, 'timing/data': 0.029625612311065197, 'timing/model': 1.13331941049546, 'learning_rate/qwen_vl_interface': 5.393764989948933e-07, 'learning_rate/action_model': 5.393764989948933e-06, 'epoch': 2.76}) +[2026-06-18 03:01:20,671][starVLA.training.train_starvla][INFO] - Step 6718, Loss: {'train/loss': 0.2233235239982605, 'train/grad_norm_pre_clip': 22.094375610351562, 'timing/data': 0.029895924031734467, 'timing/model': 0.9377335840836167, 'learning_rate/qwen_vl_interface': 5.390990932365234e-07, 'learning_rate/action_model': 5.390990932365233e-06, 'epoch': 2.77}) +[2026-06-18 03:01:24,575][starVLA.training.train_starvla][INFO] - Step 6719, Loss: {'train/loss': 0.1724999099969864, 'train/grad_norm_pre_clip': 15.365031242370605, 'timing/data': 0.03001575917005539, 'timing/model': 0.999744956381619, 'learning_rate/qwen_vl_interface': 5.388226640529808e-07, 'learning_rate/action_model': 5.388226640529808e-06, 'epoch': 2.77}) +[2026-06-18 03:01:28,366][starVLA.training.train_starvla][INFO] - Step 6720, Loss: {'train/loss': 0.16620515286922455, 'train/grad_norm_pre_clip': 15.239943504333496, 'timing/data': 0.030627706088125706, 'timing/model': 0.9324316568672657, 'learning_rate/qwen_vl_interface': 5.385472115015697e-07, 'learning_rate/action_model': 5.385472115015696e-06, 'epoch': 2.77}) +[2026-06-18 03:01:32,244][starVLA.training.train_starvla][INFO] - Step 6721, Loss: {'train/loss': 0.2878045439720154, 'train/grad_norm_pre_clip': 24.61634635925293, 'timing/data': 0.03678462188690901, 'timing/model': 0.9630307387560606, 'learning_rate/qwen_vl_interface': 5.382727356393909e-07, 'learning_rate/action_model': 5.382727356393909e-06, 'epoch': 2.77}) +[2026-06-18 03:01:36,035][starVLA.training.train_starvla][INFO] - Step 6722, Loss: {'train/loss': 0.20040565729141235, 'train/grad_norm_pre_clip': 30.48243522644043, 'timing/data': 0.030855817720294, 'timing/model': 0.9375916961580515, 'learning_rate/qwen_vl_interface': 5.379992365233442e-07, 'learning_rate/action_model': 5.379992365233442e-06, 'epoch': 2.77}) +[2026-06-18 03:01:39,867][starVLA.training.train_starvla][INFO] - Step 6723, Loss: {'train/loss': 0.323775976896286, 'train/grad_norm_pre_clip': 27.251907348632812, 'timing/data': 0.03249739482998848, 'timing/model': 1.1120031578466296, 'learning_rate/qwen_vl_interface': 5.377267142101261e-07, 'learning_rate/action_model': 5.37726714210126e-06, 'epoch': 2.77}) +[2026-06-18 03:01:43,904][starVLA.training.train_starvla][INFO] - Step 6724, Loss: {'train/loss': 0.12242517620325089, 'train/grad_norm_pre_clip': 11.550263404846191, 'timing/data': 0.03352600988000631, 'timing/model': 1.2275410592556, 'learning_rate/qwen_vl_interface': 5.374551687562306e-07, 'learning_rate/action_model': 5.374551687562306e-06, 'epoch': 2.77}) +[2026-06-18 03:01:47,621][starVLA.training.train_starvla][INFO] - Step 6725, Loss: {'train/loss': 0.44604161381721497, 'train/grad_norm_pre_clip': 28.809871673583984, 'timing/data': 0.035332110710442066, 'timing/model': 0.9504868248477578, 'learning_rate/qwen_vl_interface': 5.371846002179494e-07, 'learning_rate/action_model': 5.371846002179493e-06, 'epoch': 2.77}) +[2026-06-18 03:01:51,443][starVLA.training.train_starvla][INFO] - Step 6726, Loss: {'train/loss': 0.07919298857450485, 'train/grad_norm_pre_clip': 14.195609092712402, 'timing/data': 0.026103368028998375, 'timing/model': 0.9453568384051323, 'learning_rate/qwen_vl_interface': 5.36915008651372e-07, 'learning_rate/action_model': 5.369150086513719e-06, 'epoch': 2.77}) +[2026-06-18 03:01:55,345][starVLA.training.train_starvla][INFO] - Step 6727, Loss: {'train/loss': 0.19886468350887299, 'train/grad_norm_pre_clip': 25.22860336303711, 'timing/data': 0.03884180262684822, 'timing/model': 0.9746753014624119, 'learning_rate/qwen_vl_interface': 5.366463941123853e-07, 'learning_rate/action_model': 5.366463941123853e-06, 'epoch': 2.77}) +[2026-06-18 03:01:59,381][starVLA.training.train_starvla][INFO] - Step 6728, Loss: {'train/loss': 0.22621123492717743, 'train/grad_norm_pre_clip': 22.512237548828125, 'timing/data': 0.0444176634773612, 'timing/model': 1.0233922749757767, 'learning_rate/qwen_vl_interface': 5.36378756656672e-07, 'learning_rate/action_model': 5.36378756656672e-06, 'epoch': 2.77}) +[2026-06-18 03:02:03,240][starVLA.training.train_starvla][INFO] - Step 6729, Loss: {'train/loss': 0.14996108412742615, 'train/grad_norm_pre_clip': 14.5125150680542, 'timing/data': 0.02890831232070923, 'timing/model': 0.9366604927927256, 'learning_rate/qwen_vl_interface': 5.361120963397151e-07, 'learning_rate/action_model': 5.361120963397152e-06, 'epoch': 2.77}) +[2026-06-18 03:02:07,172][starVLA.training.train_starvla][INFO] - Step 6730, Loss: {'train/loss': 0.12584203481674194, 'train/grad_norm_pre_clip': 11.810089111328125, 'timing/data': 0.025871917605400085, 'timing/model': 1.2134862197563052, 'learning_rate/qwen_vl_interface': 5.358464132167927e-07, 'learning_rate/action_model': 5.358464132167927e-06, 'epoch': 2.77}) +[2026-06-18 03:02:10,883][starVLA.training.train_starvla][INFO] - Step 6731, Loss: {'train/loss': 0.1259433478116989, 'train/grad_norm_pre_clip': 13.95617961883545, 'timing/data': 0.03682768065482378, 'timing/model': 0.9406970124691725, 'learning_rate/qwen_vl_interface': 5.355817073429818e-07, 'learning_rate/action_model': 5.355817073429818e-06, 'epoch': 2.77}) +[2026-06-18 03:02:14,742][starVLA.training.train_starvla][INFO] - Step 6732, Loss: {'train/loss': 0.21126338839530945, 'train/grad_norm_pre_clip': 16.123638153076172, 'timing/data': 0.035338470712304115, 'timing/model': 0.9466357538476586, 'learning_rate/qwen_vl_interface': 5.353179787731557e-07, 'learning_rate/action_model': 5.353179787731557e-06, 'epoch': 2.77}) +[2026-06-18 03:02:18,668][starVLA.training.train_starvla][INFO] - Step 6733, Loss: {'train/loss': 0.18420502543449402, 'train/grad_norm_pre_clip': 19.106433868408203, 'timing/data': 0.03290386684238911, 'timing/model': 0.928496296517551, 'learning_rate/qwen_vl_interface': 5.350552275619857e-07, 'learning_rate/action_model': 5.350552275619857e-06, 'epoch': 2.77}) +[2026-06-18 03:02:22,541][starVLA.training.train_starvla][INFO] - Step 6734, Loss: {'train/loss': 0.07144635915756226, 'train/grad_norm_pre_clip': 14.800702095031738, 'timing/data': 0.024283629842102528, 'timing/model': 0.9309478038921952, 'learning_rate/qwen_vl_interface': 5.347934537639405e-07, 'learning_rate/action_model': 5.347934537639405e-06, 'epoch': 2.77}) +[2026-06-18 03:02:26,346][starVLA.training.train_starvla][INFO] - Step 6735, Loss: {'train/loss': 0.24110034108161926, 'train/grad_norm_pre_clip': 34.29779815673828, 'timing/data': 0.0277018491178751, 'timing/model': 0.9389551021158695, 'learning_rate/qwen_vl_interface': 5.345326574332865e-07, 'learning_rate/action_model': 5.345326574332865e-06, 'epoch': 2.77}) +[2026-06-18 03:02:30,152][starVLA.training.train_starvla][INFO] - Step 6736, Loss: {'train/loss': 0.11157423257827759, 'train/grad_norm_pre_clip': 22.25733757019043, 'timing/data': 0.029263729229569435, 'timing/model': 0.9306910345330834, 'learning_rate/qwen_vl_interface': 5.342728386240867e-07, 'learning_rate/action_model': 5.3427283862408665e-06, 'epoch': 2.77}) +[2026-06-18 03:02:34,023][starVLA.training.train_starvla][INFO] - Step 6737, Loss: {'train/loss': 0.14605243504047394, 'train/grad_norm_pre_clip': 18.806528091430664, 'timing/data': 0.044377950951457024, 'timing/model': 1.1041453806683421, 'learning_rate/qwen_vl_interface': 5.340139973902013e-07, 'learning_rate/action_model': 5.340139973902013e-06, 'epoch': 2.77}) +[2026-06-18 03:02:37,631][starVLA.training.train_starvla][INFO] - Step 6738, Loss: {'train/loss': 0.3639153242111206, 'train/grad_norm_pre_clip': 26.862442016601562, 'timing/data': 0.020793842151761055, 'timing/model': 0.9230481823906302, 'learning_rate/qwen_vl_interface': 5.337561337852892e-07, 'learning_rate/action_model': 5.337561337852892e-06, 'epoch': 2.77}) +[2026-06-18 03:02:41,524][starVLA.training.train_starvla][INFO] - Step 6739, Loss: {'train/loss': 0.21735087037086487, 'train/grad_norm_pre_clip': 29.142486572265625, 'timing/data': 0.03575887903571129, 'timing/model': 0.9378973254933953, 'learning_rate/qwen_vl_interface': 5.334992478628054e-07, 'learning_rate/action_model': 5.334992478628054e-06, 'epoch': 2.77}) +[2026-06-18 03:02:45,357][starVLA.training.train_starvla][INFO] - Step 6740, Loss: {'train/loss': 0.2023826688528061, 'train/grad_norm_pre_clip': 16.004005432128906, 'timing/data': 0.03119217697530985, 'timing/model': 0.9422368844971061, 'learning_rate/qwen_vl_interface': 5.332433396760027e-07, 'learning_rate/action_model': 5.332433396760028e-06, 'epoch': 2.77}) +[2026-06-18 03:02:49,372][starVLA.training.train_starvla][INFO] - Step 6741, Loss: {'train/loss': 0.14997071027755737, 'train/grad_norm_pre_clip': 22.750391006469727, 'timing/data': 0.023767813108861446, 'timing/model': 0.9611092610284686, 'learning_rate/qwen_vl_interface': 5.329884092779312e-07, 'learning_rate/action_model': 5.329884092779312e-06, 'epoch': 2.77}) +[2026-06-18 03:02:53,520][starVLA.training.train_starvla][INFO] - Step 6742, Loss: {'train/loss': 0.1740187108516693, 'train/grad_norm_pre_clip': 23.071962356567383, 'timing/data': 0.02642827946692705, 'timing/model': 1.012810479849577, 'learning_rate/qwen_vl_interface': 5.327344567214386e-07, 'learning_rate/action_model': 5.327344567214385e-06, 'epoch': 2.78}) +[2026-06-18 03:02:57,318][starVLA.training.train_starvla][INFO] - Step 6743, Loss: {'train/loss': 0.3306325078010559, 'train/grad_norm_pre_clip': 17.450056076049805, 'timing/data': 0.021889379248023033, 'timing/model': 1.107844707556069, 'learning_rate/qwen_vl_interface': 5.324814820591683e-07, 'learning_rate/action_model': 5.324814820591682e-06, 'epoch': 2.78}) +[2026-06-18 03:03:01,201][starVLA.training.train_starvla][INFO] - Step 6744, Loss: {'train/loss': 0.2778932452201843, 'train/grad_norm_pre_clip': 20.2919979095459, 'timing/data': 0.02163649909198284, 'timing/model': 1.1847773101180792, 'learning_rate/qwen_vl_interface': 5.322294853435631e-07, 'learning_rate/action_model': 5.322294853435631e-06, 'epoch': 2.78}) +[2026-06-18 03:03:04,920][starVLA.training.train_starvla][INFO] - Step 6745, Loss: {'train/loss': 0.30469948053359985, 'train/grad_norm_pre_clip': 26.270021438598633, 'timing/data': 0.0350026274099946, 'timing/model': 0.9494529953226447, 'learning_rate/qwen_vl_interface': 5.319784666268626e-07, 'learning_rate/action_model': 5.319784666268626e-06, 'epoch': 2.78}) +[2026-06-18 03:03:08,760][starVLA.training.train_starvla][INFO] - Step 6746, Loss: {'train/loss': 0.41601771116256714, 'train/grad_norm_pre_clip': 24.028024673461914, 'timing/data': 0.031323366798460484, 'timing/model': 0.9431232465431094, 'learning_rate/qwen_vl_interface': 5.317284259611024e-07, 'learning_rate/action_model': 5.317284259611024e-06, 'epoch': 2.78}) +[2026-06-18 03:03:12,684][starVLA.training.train_starvla][INFO] - Step 6747, Loss: {'train/loss': 0.16193611919879913, 'train/grad_norm_pre_clip': 22.108272552490234, 'timing/data': 0.028932519257068634, 'timing/model': 1.0040120761841536, 'learning_rate/qwen_vl_interface': 5.314793633981169e-07, 'learning_rate/action_model': 5.314793633981169e-06, 'epoch': 2.78}) +[2026-06-18 03:03:16,594][starVLA.training.train_starvla][INFO] - Step 6748, Loss: {'train/loss': 0.2955531179904938, 'train/grad_norm_pre_clip': 26.57203483581543, 'timing/data': 0.031419348903000355, 'timing/model': 0.9343681149184704, 'learning_rate/qwen_vl_interface': 5.312312789895369e-07, 'learning_rate/action_model': 5.3123127898953685e-06, 'epoch': 2.78}) +[2026-06-18 03:03:20,452][starVLA.training.train_starvla][INFO] - Step 6749, Loss: {'train/loss': 0.4856055974960327, 'train/grad_norm_pre_clip': 26.64990997314453, 'timing/data': 0.025784486904740334, 'timing/model': 0.928212346509099, 'learning_rate/qwen_vl_interface': 5.3098417278679e-07, 'learning_rate/action_model': 5.3098417278679e-06, 'epoch': 2.78}) +[2026-06-18 03:03:24,427][starVLA.training.train_starvla][INFO] - Step 6750, Loss: {'train/loss': 0.1798638552427292, 'train/grad_norm_pre_clip': 21.196455001831055, 'timing/data': 0.021267710253596306, 'timing/model': 1.116940121166408, 'learning_rate/qwen_vl_interface': 5.307380448411022e-07, 'learning_rate/action_model': 5.307380448411022e-06, 'epoch': 2.78}) +[2026-06-18 03:03:28,305][starVLA.training.train_starvla][INFO] - Step 6751, Loss: {'train/loss': 0.3732079267501831, 'train/grad_norm_pre_clip': 32.234134674072266, 'timing/data': 0.030071144923567772, 'timing/model': 1.0993174901232123, 'learning_rate/qwen_vl_interface': 5.304928952034955e-07, 'learning_rate/action_model': 5.304928952034956e-06, 'epoch': 2.78}) +[2026-06-18 03:03:31,905][starVLA.training.train_starvla][INFO] - Step 6752, Loss: {'train/loss': 0.3231053054332733, 'train/grad_norm_pre_clip': 33.158424377441406, 'timing/data': 0.024639698676764965, 'timing/model': 0.9376536160707474, 'learning_rate/qwen_vl_interface': 5.302487239247905e-07, 'learning_rate/action_model': 5.302487239247904e-06, 'epoch': 2.78}) +[2026-06-18 03:03:35,881][starVLA.training.train_starvla][INFO] - Step 6753, Loss: {'train/loss': 0.24993620812892914, 'train/grad_norm_pre_clip': 27.252376556396484, 'timing/data': 0.024754063226282597, 'timing/model': 1.105005501769483, 'learning_rate/qwen_vl_interface': 5.300055310556034e-07, 'learning_rate/action_model': 5.300055310556034e-06, 'epoch': 2.78}) +[2026-06-18 03:03:40,086][starVLA.training.train_starvla][INFO] - Step 6754, Loss: {'train/loss': 0.1383529007434845, 'train/grad_norm_pre_clip': 21.88714599609375, 'timing/data': 0.027024230919778347, 'timing/model': 1.0159685630351305, 'learning_rate/qwen_vl_interface': 5.297633166463482e-07, 'learning_rate/action_model': 5.297633166463481e-06, 'epoch': 2.78}) +[2026-06-18 03:03:43,970][starVLA.training.train_starvla][INFO] - Step 6755, Loss: {'train/loss': 0.11971986293792725, 'train/grad_norm_pre_clip': 12.985864639282227, 'timing/data': 0.02729815896600485, 'timing/model': 0.9240099331364036, 'learning_rate/qwen_vl_interface': 5.295220807472373e-07, 'learning_rate/action_model': 5.295220807472374e-06, 'epoch': 2.78}) +[2026-06-18 03:03:47,853][starVLA.training.train_starvla][INFO] - Step 6756, Loss: {'train/loss': 0.24780704081058502, 'train/grad_norm_pre_clip': 33.66862487792969, 'timing/data': 0.026614383794367313, 'timing/model': 0.9316842574626207, 'learning_rate/qwen_vl_interface': 5.292818234082786e-07, 'learning_rate/action_model': 5.292818234082787e-06, 'epoch': 2.78}) +[2026-06-18 03:03:51,629][starVLA.training.train_starvla][INFO] - Step 6757, Loss: {'train/loss': 0.05102010443806648, 'train/grad_norm_pre_clip': 9.272419929504395, 'timing/data': 0.022982862778007984, 'timing/model': 1.0837128106504679, 'learning_rate/qwen_vl_interface': 5.290425446792772e-07, 'learning_rate/action_model': 5.2904254467927716e-06, 'epoch': 2.78}) +[2026-06-18 03:03:55,301][starVLA.training.train_starvla][INFO] - Step 6758, Loss: {'train/loss': 0.4483638107776642, 'train/grad_norm_pre_clip': 30.330913543701172, 'timing/data': 0.024451417848467827, 'timing/model': 0.9324537646025419, 'learning_rate/qwen_vl_interface': 5.288042446098364e-07, 'learning_rate/action_model': 5.2880424460983634e-06, 'epoch': 2.78}) +[2026-06-18 03:03:59,079][starVLA.training.train_starvla][INFO] - Step 6759, Loss: {'train/loss': 0.13019989430904388, 'train/grad_norm_pre_clip': 16.198509216308594, 'timing/data': 0.02951290737837553, 'timing/model': 0.9206775417551398, 'learning_rate/qwen_vl_interface': 5.28566923249356e-07, 'learning_rate/action_model': 5.28566923249356e-06, 'epoch': 2.78}) +[2026-06-18 03:04:02,789][starVLA.training.train_starvla][INFO] - Step 6760, Loss: {'train/loss': 0.026657259091734886, 'train/grad_norm_pre_clip': 5.550522327423096, 'timing/data': 0.02814633771777153, 'timing/model': 0.9370620520785451, 'learning_rate/qwen_vl_interface': 5.283305806470329e-07, 'learning_rate/action_model': 5.2833058064703286e-06, 'epoch': 2.78}) +[2026-06-18 03:04:06,541][starVLA.training.train_starvla][INFO] - Step 6761, Loss: {'train/loss': 0.18106193840503693, 'train/grad_norm_pre_clip': 22.672805786132812, 'timing/data': 0.02592128235846758, 'timing/model': 0.9344282653182745, 'learning_rate/qwen_vl_interface': 5.280952168518612e-07, 'learning_rate/action_model': 5.280952168518612e-06, 'epoch': 2.78}) +[2026-06-18 03:04:10,328][starVLA.training.train_starvla][INFO] - Step 6762, Loss: {'train/loss': 0.3164890706539154, 'train/grad_norm_pre_clip': 15.788132667541504, 'timing/data': 0.02288237027823925, 'timing/model': 0.9439909383654594, 'learning_rate/qwen_vl_interface': 5.278608319126318e-07, 'learning_rate/action_model': 5.278608319126318e-06, 'epoch': 2.78}) +[2026-06-18 03:04:14,128][starVLA.training.train_starvla][INFO] - Step 6763, Loss: {'train/loss': 0.463212251663208, 'train/grad_norm_pre_clip': 27.355072021484375, 'timing/data': 0.02054504957050085, 'timing/model': 1.1176787819713354, 'learning_rate/qwen_vl_interface': 5.276274258779332e-07, 'learning_rate/action_model': 5.2762742587793315e-06, 'epoch': 2.78}) +[2026-06-18 03:04:18,346][starVLA.training.train_starvla][INFO] - Step 6764, Loss: {'train/loss': 0.2742527723312378, 'train/grad_norm_pre_clip': 32.81208801269531, 'timing/data': 0.027669028379023075, 'timing/model': 1.4313180325552821, 'learning_rate/qwen_vl_interface': 5.273949987961508e-07, 'learning_rate/action_model': 5.273949987961508e-06, 'epoch': 2.78}) +[2026-06-18 03:04:22,059][starVLA.training.train_starvla][INFO] - Step 6765, Loss: {'train/loss': 0.3437284529209137, 'train/grad_norm_pre_clip': 22.569076538085938, 'timing/data': 0.0249588368460536, 'timing/model': 0.9518767753615975, 'learning_rate/qwen_vl_interface': 5.271635507154671e-07, 'learning_rate/action_model': 5.2716355071546706e-06, 'epoch': 2.78}) +[2026-06-18 03:04:26,074][starVLA.training.train_starvla][INFO] - Step 6766, Loss: {'train/loss': 0.10497397184371948, 'train/grad_norm_pre_clip': 14.382733345031738, 'timing/data': 0.03091542422771454, 'timing/model': 0.9362190328538418, 'learning_rate/qwen_vl_interface': 5.269330816838614e-07, 'learning_rate/action_model': 5.269330816838613e-06, 'epoch': 2.79}) +[2026-06-18 03:04:30,076][starVLA.training.train_starvla][INFO] - Step 6767, Loss: {'train/loss': 0.1585109531879425, 'train/grad_norm_pre_clip': 25.71820831298828, 'timing/data': 0.03332430496811867, 'timing/model': 0.9515270004048944, 'learning_rate/qwen_vl_interface': 5.267035917491095e-07, 'learning_rate/action_model': 5.267035917491094e-06, 'epoch': 2.79}) +[2026-06-18 03:04:33,865][starVLA.training.train_starvla][INFO] - Step 6768, Loss: {'train/loss': 0.05034378916025162, 'train/grad_norm_pre_clip': 10.059069633483887, 'timing/data': 0.026523754000663757, 'timing/model': 0.9423505691811442, 'learning_rate/qwen_vl_interface': 5.264750809587862e-07, 'learning_rate/action_model': 5.264750809587863e-06, 'epoch': 2.79}) +[2026-06-18 03:04:37,830][starVLA.training.train_starvla][INFO] - Step 6769, Loss: {'train/loss': 0.3261047303676605, 'train/grad_norm_pre_clip': 23.268909454345703, 'timing/data': 0.02470778487622738, 'timing/model': 0.9344219053164124, 'learning_rate/qwen_vl_interface': 5.262475493602609e-07, 'learning_rate/action_model': 5.26247549360261e-06, 'epoch': 2.79}) +[2026-06-18 03:04:41,628][starVLA.training.train_starvla][INFO] - Step 6770, Loss: {'train/loss': 0.03984029218554497, 'train/grad_norm_pre_clip': 6.113147735595703, 'timing/data': 0.03945579379796982, 'timing/model': 1.1248790640383959, 'learning_rate/qwen_vl_interface': 5.260209970007018e-07, 'learning_rate/action_model': 5.260209970007018e-06, 'epoch': 2.79}) +[2026-06-18 03:04:45,407][starVLA.training.train_starvla][INFO] - Step 6771, Loss: {'train/loss': 0.08225461840629578, 'train/grad_norm_pre_clip': 10.728279113769531, 'timing/data': 0.024158142507076263, 'timing/model': 1.1041940925642848, 'learning_rate/qwen_vl_interface': 5.257954239270732e-07, 'learning_rate/action_model': 5.257954239270732e-06, 'epoch': 2.79}) +[2026-06-18 03:04:49,117][starVLA.training.train_starvla][INFO] - Step 6772, Loss: {'train/loss': 0.38216322660446167, 'train/grad_norm_pre_clip': 21.052818298339844, 'timing/data': 0.0310371620580554, 'timing/model': 0.9998391475528479, 'learning_rate/qwen_vl_interface': 5.255708301861367e-07, 'learning_rate/action_model': 5.255708301861367e-06, 'epoch': 2.79}) +[2026-06-18 03:04:53,365][starVLA.training.train_starvla][INFO] - Step 6773, Loss: {'train/loss': 0.46119916439056396, 'train/grad_norm_pre_clip': 31.293533325195312, 'timing/data': 0.02696357946842909, 'timing/model': 0.9906012313440442, 'learning_rate/qwen_vl_interface': 5.253472158244506e-07, 'learning_rate/action_model': 5.253472158244505e-06, 'epoch': 2.79}) +[2026-06-18 03:04:57,234][starVLA.training.train_starvla][INFO] - Step 6774, Loss: {'train/loss': 0.09801948815584183, 'train/grad_norm_pre_clip': 12.519777297973633, 'timing/data': 0.02296258695423603, 'timing/model': 0.9266862291842699, 'learning_rate/qwen_vl_interface': 5.251245808883707e-07, 'learning_rate/action_model': 5.251245808883707e-06, 'epoch': 2.79}) +[2026-06-18 03:05:01,037][starVLA.training.train_starvla][INFO] - Step 6775, Loss: {'train/loss': 0.174063578248024, 'train/grad_norm_pre_clip': 17.49537467956543, 'timing/data': 0.04069513641297817, 'timing/model': 0.9497805070132017, 'learning_rate/qwen_vl_interface': 5.249029254240495e-07, 'learning_rate/action_model': 5.249029254240495e-06, 'epoch': 2.79}) +[2026-06-18 03:05:04,843][starVLA.training.train_starvla][INFO] - Step 6776, Loss: {'train/loss': 0.16262094676494598, 'train/grad_norm_pre_clip': 19.142040252685547, 'timing/data': 0.02393599320203066, 'timing/model': 0.9476487981155515, 'learning_rate/qwen_vl_interface': 5.246822494774364e-07, 'learning_rate/action_model': 5.246822494774363e-06, 'epoch': 2.79}) +[2026-06-18 03:05:08,837][starVLA.training.train_starvla][INFO] - Step 6777, Loss: {'train/loss': 0.10397084057331085, 'train/grad_norm_pre_clip': 12.542818069458008, 'timing/data': 0.028639364056289196, 'timing/model': 1.149343777447939, 'learning_rate/qwen_vl_interface': 5.244625530942774e-07, 'learning_rate/action_model': 5.244625530942774e-06, 'epoch': 2.79}) +[2026-06-18 03:05:12,594][starVLA.training.train_starvla][INFO] - Step 6778, Loss: {'train/loss': 0.2327347844839096, 'train/grad_norm_pre_clip': 21.56591033935547, 'timing/data': 0.029376442544162273, 'timing/model': 0.9330735979601741, 'learning_rate/qwen_vl_interface': 5.242438363201161e-07, 'learning_rate/action_model': 5.2424383632011605e-06, 'epoch': 2.79}) +[2026-06-18 03:05:16,502][starVLA.training.train_starvla][INFO] - Step 6779, Loss: {'train/loss': 0.3464067876338959, 'train/grad_norm_pre_clip': 24.75698471069336, 'timing/data': 0.022259806282818317, 'timing/model': 1.000090061686933, 'learning_rate/qwen_vl_interface': 5.240260992002928e-07, 'learning_rate/action_model': 5.240260992002928e-06, 'epoch': 2.79}) +[2026-06-18 03:05:20,367][starVLA.training.train_starvla][INFO] - Step 6780, Loss: {'train/loss': 0.2272525429725647, 'train/grad_norm_pre_clip': 24.261932373046875, 'timing/data': 0.032261268235743046, 'timing/model': 0.9326299019157887, 'learning_rate/qwen_vl_interface': 5.238093417799446e-07, 'learning_rate/action_model': 5.238093417799446e-06, 'epoch': 2.79}) +[2026-06-18 03:05:24,250][starVLA.training.train_starvla][INFO] - Step 6781, Loss: {'train/loss': 0.07013826072216034, 'train/grad_norm_pre_clip': 11.81576156616211, 'timing/data': 0.027398922480642796, 'timing/model': 0.9350238619372249, 'learning_rate/qwen_vl_interface': 5.235935641040048e-07, 'learning_rate/action_model': 5.235935641040048e-06, 'epoch': 2.79}) +[2026-06-18 03:05:28,033][starVLA.training.train_starvla][INFO] - Step 6782, Loss: {'train/loss': 0.2532467246055603, 'train/grad_norm_pre_clip': 33.43570327758789, 'timing/data': 0.024634061381220818, 'timing/model': 0.9277798887342215, 'learning_rate/qwen_vl_interface': 5.233787662172059e-07, 'learning_rate/action_model': 5.23378766217206e-06, 'epoch': 2.79}) +[2026-06-18 03:05:31,785][starVLA.training.train_starvla][INFO] - Step 6783, Loss: {'train/loss': 0.09816565364599228, 'train/grad_norm_pre_clip': 15.22420597076416, 'timing/data': 0.02841951511800289, 'timing/model': 0.9336376506835222, 'learning_rate/qwen_vl_interface': 5.231649481640741e-07, 'learning_rate/action_model': 5.231649481640741e-06, 'epoch': 2.79}) +[2026-06-18 03:05:35,626][starVLA.training.train_starvla][INFO] - Step 6784, Loss: {'train/loss': 0.05024076998233795, 'train/grad_norm_pre_clip': 6.9329833984375, 'timing/data': 0.03273965045809746, 'timing/model': 1.1542491065338254, 'learning_rate/qwen_vl_interface': 5.229521099889351e-07, 'learning_rate/action_model': 5.229521099889352e-06, 'epoch': 2.79}) +[2026-06-18 03:05:39,514][starVLA.training.train_starvla][INFO] - Step 6785, Loss: {'train/loss': 0.18195080757141113, 'train/grad_norm_pre_clip': 22.571882247924805, 'timing/data': 0.034332660026848316, 'timing/model': 0.9408651357516646, 'learning_rate/qwen_vl_interface': 5.227402517359105e-07, 'learning_rate/action_model': 5.2274025173591045e-06, 'epoch': 2.79}) +[2026-06-18 03:05:43,420][starVLA.training.train_starvla][INFO] - Step 6786, Loss: {'train/loss': 0.17484574019908905, 'train/grad_norm_pre_clip': 24.789348602294922, 'timing/data': 0.023699762299656868, 'timing/model': 0.9339582603424788, 'learning_rate/qwen_vl_interface': 5.225293734489182e-07, 'learning_rate/action_model': 5.225293734489181e-06, 'epoch': 2.79}) +[2026-06-18 03:05:47,267][starVLA.training.train_starvla][INFO] - Step 6787, Loss: {'train/loss': 0.8047081828117371, 'train/grad_norm_pre_clip': 38.81333541870117, 'timing/data': 0.023175486363470554, 'timing/model': 0.9381633047014475, 'learning_rate/qwen_vl_interface': 5.223194751716733e-07, 'learning_rate/action_model': 5.223194751716733e-06, 'epoch': 2.79}) +[2026-06-18 03:05:51,191][starVLA.training.train_starvla][INFO] - Step 6788, Loss: {'train/loss': 0.13033327460289001, 'train/grad_norm_pre_clip': 20.448423385620117, 'timing/data': 0.026742679998278618, 'timing/model': 0.9484854098409414, 'learning_rate/qwen_vl_interface': 5.221105569476893e-07, 'learning_rate/action_model': 5.221105569476892e-06, 'epoch': 2.79}) +[2026-06-18 03:05:54,996][starVLA.training.train_starvla][INFO] - Step 6789, Loss: {'train/loss': 0.3202362060546875, 'train/grad_norm_pre_clip': 31.90123748779297, 'timing/data': 0.02508531417697668, 'timing/model': 0.9320228602737188, 'learning_rate/qwen_vl_interface': 5.21902618820274e-07, 'learning_rate/action_model': 5.21902618820274e-06, 'epoch': 2.79}) +[2026-06-18 03:05:58,813][starVLA.training.train_starvla][INFO] - Step 6790, Loss: {'train/loss': 0.17024554312229156, 'train/grad_norm_pre_clip': 33.4395751953125, 'timing/data': 0.024963035248219967, 'timing/model': 1.1075858296826482, 'learning_rate/qwen_vl_interface': 5.216956608325331e-07, 'learning_rate/action_model': 5.2169566083253304e-06, 'epoch': 2.8}) +[2026-06-18 03:06:02,625][starVLA.training.train_starvla][INFO] - Step 6791, Loss: {'train/loss': 0.33911842107772827, 'train/grad_norm_pre_clip': 56.272029876708984, 'timing/data': 0.02500252053141594, 'timing/model': 1.1169094191864133, 'learning_rate/qwen_vl_interface': 5.2148968302737e-07, 'learning_rate/action_model': 5.2148968302737005e-06, 'epoch': 2.8}) +[2026-06-18 03:06:06,302][starVLA.training.train_starvla][INFO] - Step 6792, Loss: {'train/loss': 0.09951155632734299, 'train/grad_norm_pre_clip': 13.727999687194824, 'timing/data': 0.035033767111599445, 'timing/model': 0.9655409641563892, 'learning_rate/qwen_vl_interface': 5.212846854474834e-07, 'learning_rate/action_model': 5.212846854474833e-06, 'epoch': 2.8}) +[2026-06-18 03:06:10,164][starVLA.training.train_starvla][INFO] - Step 6793, Loss: {'train/loss': 0.34334126114845276, 'train/grad_norm_pre_clip': 20.610891342163086, 'timing/data': 0.0223246356472373, 'timing/model': 0.9490648591890931, 'learning_rate/qwen_vl_interface': 5.2108066813537e-07, 'learning_rate/action_model': 5.2108066813537e-06, 'epoch': 2.8}) +[2026-06-18 03:06:15,007][starVLA.training.train_starvla][INFO] - Step 6794, Loss: {'train/loss': 0.09995176643133163, 'train/grad_norm_pre_clip': 17.930282592773438, 'timing/data': 0.02455966453999281, 'timing/model': 0.9476483622565866, 'learning_rate/qwen_vl_interface': 5.208776311333223e-07, 'learning_rate/action_model': 5.208776311333223e-06, 'epoch': 2.8}) +[2026-06-18 03:06:19,177][starVLA.training.train_starvla][INFO] - Step 6795, Loss: {'train/loss': 0.40935611724853516, 'train/grad_norm_pre_clip': 26.665706634521484, 'timing/data': 0.027045621536672115, 'timing/model': 0.9365128120407462, 'learning_rate/qwen_vl_interface': 5.206755744834312e-07, 'learning_rate/action_model': 5.2067557448343125e-06, 'epoch': 2.8}) +[2026-06-18 03:06:23,068][starVLA.training.train_starvla][INFO] - Step 6796, Loss: {'train/loss': 0.11037878692150116, 'train/grad_norm_pre_clip': 12.630483627319336, 'timing/data': 0.022624925710260868, 'timing/model': 0.9457743214443326, 'learning_rate/qwen_vl_interface': 5.20474498227582e-07, 'learning_rate/action_model': 5.2047449822758196e-06, 'epoch': 2.8}) +[2026-06-18 03:06:26,936][starVLA.training.train_starvla][INFO] - Step 6797, Loss: {'train/loss': 0.13780659437179565, 'train/grad_norm_pre_clip': 18.434478759765625, 'timing/data': 0.025237560272216797, 'timing/model': 1.1062961462885141, 'learning_rate/qwen_vl_interface': 5.202744024074583e-07, 'learning_rate/action_model': 5.202744024074584e-06, 'epoch': 2.8}) +[2026-06-18 03:06:30,801][starVLA.training.train_starvla][INFO] - Step 6798, Loss: {'train/loss': 0.1449878066778183, 'train/grad_norm_pre_clip': 21.96210289001465, 'timing/data': 0.029974231496453285, 'timing/model': 1.1121932184323668, 'learning_rate/qwen_vl_interface': 5.200752870645407e-07, 'learning_rate/action_model': 5.200752870645406e-06, 'epoch': 2.8}) +[2026-06-18 03:06:34,433][starVLA.training.train_starvla][INFO] - Step 6799, Loss: {'train/loss': 0.07593747973442078, 'train/grad_norm_pre_clip': 14.295212745666504, 'timing/data': 0.025737482123076916, 'timing/model': 0.9282430009916425, 'learning_rate/qwen_vl_interface': 5.198771522401054e-07, 'learning_rate/action_model': 5.198771522401054e-06, 'epoch': 2.8}) +[2026-06-18 03:08:38,685][starVLA.training.train_starvla][INFO] - Step 6800, Loss: {'train/loss': 0.15307585895061493, 'train/grad_norm_pre_clip': 19.185224533081055, 'eval/loss': 0.4257333278656006, 'eval/latency_6/loss': 0.4257333278656006, 'eval/demon_fire_precision': 0.9513496547394853, 'eval/demon_fire_recall': 0.9796380090497737, 'eval/demon_fire_f1': 0.9652866242038216, 'eval/demon_fire_rate_model': 0.8864774624373957, 'eval/demon_fire_rate_teacher': 0.860879243183083, 'eval/demon_left_precision': 0.8790322580645161, 'eval/demon_left_recall': 0.7649122807017544, 'eval/demon_left_f1': 0.8180112570356473, 'eval/demon_left_rate_model': 0.13800779076238176, 'eval/demon_left_rate_teacher': 0.15859766277128548, 'eval/demon_right_precision': 0.8327868852459016, 'eval/demon_right_recall': 0.7803379416282642, 'eval/demon_right_f1': 0.8057097541633624, 'eval/demon_right_rate_model': 0.16972732331663884, 'eval/demon_right_rate_teacher': 0.18113522537562604, 'eval/demon_move_f1': 0.8118605055995048, 'eval/demon_cc_f1': 0.9134966045766484, 'eval/latency_6/demon_fire_precision': 0.9513496547394853, 'eval/latency_6/demon_fire_recall': 0.9796380090497737, 'eval/latency_6/demon_fire_f1': 0.9652866242038216, 'eval/latency_6/demon_fire_rate_model': 0.8864774624373957, 'eval/latency_6/demon_fire_rate_teacher': 0.860879243183083, 'eval/latency_6/demon_left_precision': 0.8790322580645161, 'eval/latency_6/demon_left_recall': 0.7649122807017544, 'eval/latency_6/demon_left_f1': 0.8180112570356473, 'eval/latency_6/demon_left_rate_model': 0.13800779076238176, 'eval/latency_6/demon_left_rate_teacher': 0.15859766277128548, 'eval/latency_6/demon_right_precision': 0.8327868852459016, 'eval/latency_6/demon_right_recall': 0.7803379416282642, 'eval/latency_6/demon_right_f1': 0.8057097541633624, 'eval/latency_6/demon_right_rate_model': 0.16972732331663884, 'eval/latency_6/demon_right_rate_teacher': 0.18113522537562604, 'eval/latency_6/demon_move_f1': 0.8118605055995048, 'eval/latency_6/demon_cc_f1': 0.9134966045766484, 'timing/data': 0.031202366575598717, 'timing/model': 1.0435837544500828, 'learning_rate/qwen_vl_interface': 5.196799979752265e-07, 'learning_rate/action_model': 5.196799979752264e-06, 'epoch': 2.8}) +[2026-06-18 03:08:38,697][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_6800_state +[2026-06-18 03:08:38,716][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 03:08:45,213][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_6800_state/model.safetensors +[2026-06-18 03:08:59,958][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_6800_state/optimizer.bin +[2026-06-18 03:08:59,959][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_6800_state/sampler.bin +[2026-06-18 03:08:59,960][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_6800_state/sampler_1.bin +[2026-06-18 03:08:59,963][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_6800_state/random_states_0.pkl +[2026-06-18 03:08:59,967][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 6800 +[2026-06-18 03:08:59,967][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 03:08:59,977][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 03:09:03,574][starVLA.training.train_starvla][INFO] - Step 6801, Loss: {'train/loss': 0.148140549659729, 'train/grad_norm_pre_clip': 14.899393081665039, 'timing/data': 0.02870267629623413, 'timing/model': 0.931578267365694, 'learning_rate/qwen_vl_interface': 5.194838243107742e-07, 'learning_rate/action_model': 5.194838243107742e-06, 'epoch': 2.8}) +[2026-06-18 03:09:07,385][starVLA.training.train_starvla][INFO] - Step 6802, Loss: {'train/loss': 0.07316528260707855, 'train/grad_norm_pre_clip': 12.892205238342285, 'timing/data': 0.034011200070381165, 'timing/model': 0.935685339383781, 'learning_rate/qwen_vl_interface': 5.192886312874148e-07, 'learning_rate/action_model': 5.1928863128741475e-06, 'epoch': 2.8}) +[2026-06-18 03:09:11,211][starVLA.training.train_starvla][INFO] - Step 6803, Loss: {'train/loss': 0.22106698155403137, 'train/grad_norm_pre_clip': 31.163925170898438, 'timing/data': 0.026901685632765293, 'timing/model': 0.9463162841275334, 'learning_rate/qwen_vl_interface': 5.190944189456133e-07, 'learning_rate/action_model': 5.190944189456132e-06, 'epoch': 2.8}) +[2026-06-18 03:09:15,078][starVLA.training.train_starvla][INFO] - Step 6804, Loss: {'train/loss': 0.32872363924980164, 'train/grad_norm_pre_clip': 27.790273666381836, 'timing/data': 0.03102889098227024, 'timing/model': 0.9353864630684257, 'learning_rate/qwen_vl_interface': 5.189011873256295e-07, 'learning_rate/action_model': 5.189011873256295e-06, 'epoch': 2.8}) +[2026-06-18 03:09:18,857][starVLA.training.train_starvla][INFO] - Step 6805, Loss: {'train/loss': 0.45536676049232483, 'train/grad_norm_pre_clip': 31.619516372680664, 'timing/data': 0.023193396627902985, 'timing/model': 0.9306711973622441, 'learning_rate/qwen_vl_interface': 5.187089364675195e-07, 'learning_rate/action_model': 5.187089364675194e-06, 'epoch': 2.8}) +[2026-06-18 03:09:22,639][starVLA.training.train_starvla][INFO] - Step 6806, Loss: {'train/loss': 0.1974039524793625, 'train/grad_norm_pre_clip': 23.03952980041504, 'timing/data': 0.022961136884987354, 'timing/model': 1.0949083054438233, 'learning_rate/qwen_vl_interface': 5.18517666411139e-07, 'learning_rate/action_model': 5.18517666411139e-06, 'epoch': 2.8}) +[2026-06-18 03:09:26,438][starVLA.training.train_starvla][INFO] - Step 6807, Loss: {'train/loss': 0.2892540693283081, 'train/grad_norm_pre_clip': 29.2117977142334, 'timing/data': 0.023530589416623116, 'timing/model': 1.138812916353345, 'learning_rate/qwen_vl_interface': 5.18327377196137e-07, 'learning_rate/action_model': 5.18327377196137e-06, 'epoch': 2.8}) +[2026-06-18 03:09:30,249][starVLA.training.train_starvla][INFO] - Step 6808, Loss: {'train/loss': 0.05993424355983734, 'train/grad_norm_pre_clip': 10.756499290466309, 'timing/data': 0.02400433924049139, 'timing/model': 0.9514460610225797, 'learning_rate/qwen_vl_interface': 5.181380688619617e-07, 'learning_rate/action_model': 5.181380688619617e-06, 'epoch': 2.8}) +[2026-06-18 03:09:34,022][starVLA.training.train_starvla][INFO] - Step 6809, Loss: {'train/loss': 0.18320852518081665, 'train/grad_norm_pre_clip': 20.442285537719727, 'timing/data': 0.028653152287006378, 'timing/model': 0.9330281652510166, 'learning_rate/qwen_vl_interface': 5.179497414478561e-07, 'learning_rate/action_model': 5.179497414478562e-06, 'epoch': 2.8}) +[2026-06-18 03:09:37,778][starVLA.training.train_starvla][INFO] - Step 6810, Loss: {'train/loss': 0.30615249276161194, 'train/grad_norm_pre_clip': 19.720081329345703, 'timing/data': 0.025169013999402523, 'timing/model': 0.9252211470156908, 'learning_rate/qwen_vl_interface': 5.177623949928608e-07, 'learning_rate/action_model': 5.177623949928608e-06, 'epoch': 2.8}) +[2026-06-18 03:09:41,586][starVLA.training.train_starvla][INFO] - Step 6811, Loss: {'train/loss': 0.201470747590065, 'train/grad_norm_pre_clip': 38.005958557128906, 'timing/data': 0.025452492758631706, 'timing/model': 0.9298904435709119, 'learning_rate/qwen_vl_interface': 5.175760295358137e-07, 'learning_rate/action_model': 5.175760295358136e-06, 'epoch': 2.8}) +[2026-06-18 03:09:45,862][starVLA.training.train_starvla][INFO] - Step 6812, Loss: {'train/loss': 0.1535712331533432, 'train/grad_norm_pre_clip': 18.490253448486328, 'timing/data': 0.029178636148571968, 'timing/model': 1.151942909695208, 'learning_rate/qwen_vl_interface': 5.173906451153477e-07, 'learning_rate/action_model': 5.173906451153477e-06, 'epoch': 2.8}) +[2026-06-18 03:09:49,651][starVLA.training.train_starvla][INFO] - Step 6813, Loss: {'train/loss': 0.07974261045455933, 'train/grad_norm_pre_clip': 19.470613479614258, 'timing/data': 0.02835810463875532, 'timing/model': 1.0861201845109463, 'learning_rate/qwen_vl_interface': 5.172062417698931e-07, 'learning_rate/action_model': 5.1720624176989304e-06, 'epoch': 2.8}) +[2026-06-18 03:09:53,476][starVLA.training.train_starvla][INFO] - Step 6814, Loss: {'train/loss': 0.4114159941673279, 'train/grad_norm_pre_clip': 19.12592315673828, 'timing/data': 0.031955499202013016, 'timing/model': 1.1439428627490997, 'learning_rate/qwen_vl_interface': 5.170228195376778e-07, 'learning_rate/action_model': 5.170228195376778e-06, 'epoch': 2.8}) +[2026-06-18 03:09:57,135][starVLA.training.train_starvla][INFO] - Step 6815, Loss: {'train/loss': 0.3566586375236511, 'train/grad_norm_pre_clip': 25.694255828857422, 'timing/data': 0.02796315774321556, 'timing/model': 1.0006080362945795, 'learning_rate/qwen_vl_interface': 5.168403784567247e-07, 'learning_rate/action_model': 5.1684037845672475e-06, 'epoch': 2.81}) +[2026-06-18 03:10:01,175][starVLA.training.train_starvla][INFO] - Step 6816, Loss: {'train/loss': 0.3226962387561798, 'train/grad_norm_pre_clip': 24.646038055419922, 'timing/data': 0.026348325423896313, 'timing/model': 1.111456099897623, 'learning_rate/qwen_vl_interface': 5.166589185648543e-07, 'learning_rate/action_model': 5.1665891856485425e-06, 'epoch': 2.81}) +[2026-06-18 03:10:04,994][starVLA.training.train_starvla][INFO] - Step 6817, Loss: {'train/loss': 0.18809831142425537, 'train/grad_norm_pre_clip': 17.645998001098633, 'timing/data': 0.02992907166481018, 'timing/model': 0.9372220048680902, 'learning_rate/qwen_vl_interface': 5.164784398996836e-07, 'learning_rate/action_model': 5.164784398996836e-06, 'epoch': 2.81}) +[2026-06-18 03:10:08,828][starVLA.training.train_starvla][INFO] - Step 6818, Loss: {'train/loss': 0.22761915624141693, 'train/grad_norm_pre_clip': 21.978229522705078, 'timing/data': 0.03014133032411337, 'timing/model': 0.9341022176668048, 'learning_rate/qwen_vl_interface': 5.162989424986253e-07, 'learning_rate/action_model': 5.162989424986253e-06, 'epoch': 2.81}) +[2026-06-18 03:10:12,588][starVLA.training.train_starvla][INFO] - Step 6819, Loss: {'train/loss': 0.3447797894477844, 'train/grad_norm_pre_clip': 142.9425811767578, 'timing/data': 0.025181119330227375, 'timing/model': 0.9287130730226636, 'learning_rate/qwen_vl_interface': 5.161204263988904e-07, 'learning_rate/action_model': 5.161204263988904e-06, 'epoch': 2.81}) +[2026-06-18 03:10:16,380][starVLA.training.train_starvla][INFO] - Step 6820, Loss: {'train/loss': 0.11798987537622452, 'train/grad_norm_pre_clip': 9.840780258178711, 'timing/data': 0.033271634951233864, 'timing/model': 1.0891651809215546, 'learning_rate/qwen_vl_interface': 5.159428916374845e-07, 'learning_rate/action_model': 5.159428916374845e-06, 'epoch': 2.81}) +[2026-06-18 03:10:19,957][starVLA.training.train_starvla][INFO] - Step 6821, Loss: {'train/loss': 0.22164830565452576, 'train/grad_norm_pre_clip': 58.76213836669922, 'timing/data': 0.02540496736764908, 'timing/model': 0.9202682469040155, 'learning_rate/qwen_vl_interface': 5.157663382512118e-07, 'learning_rate/action_model': 5.157663382512119e-06, 'epoch': 2.81}) +[2026-06-18 03:10:24,030][starVLA.training.train_starvla][INFO] - Step 6822, Loss: {'train/loss': 0.13812248408794403, 'train/grad_norm_pre_clip': 15.930437088012695, 'timing/data': 0.024955451488494873, 'timing/model': 0.92185686994344, 'learning_rate/qwen_vl_interface': 5.155907662766711e-07, 'learning_rate/action_model': 5.155907662766711e-06, 'epoch': 2.81}) +[2026-06-18 03:10:27,723][starVLA.training.train_starvla][INFO] - Step 6823, Loss: {'train/loss': 0.2727770507335663, 'train/grad_norm_pre_clip': 16.1368465423584, 'timing/data': 0.024823187850415707, 'timing/model': 0.9262678883969784, 'learning_rate/qwen_vl_interface': 5.15416175750259e-07, 'learning_rate/action_model': 5.154161757502589e-06, 'epoch': 2.81}) +[2026-06-18 03:10:31,802][starVLA.training.train_starvla][INFO] - Step 6824, Loss: {'train/loss': 0.10528383404016495, 'train/grad_norm_pre_clip': 18.717201232910156, 'timing/data': 0.02326629962772131, 'timing/model': 1.1638973476365209, 'learning_rate/qwen_vl_interface': 5.15242566708168e-07, 'learning_rate/action_model': 5.15242566708168e-06, 'epoch': 2.81}) +[2026-06-18 03:10:35,778][starVLA.training.train_starvla][INFO] - Step 6825, Loss: {'train/loss': 0.10715750604867935, 'train/grad_norm_pre_clip': 13.312713623046875, 'timing/data': 0.022607619874179363, 'timing/model': 0.9641679022461176, 'learning_rate/qwen_vl_interface': 5.150699391863884e-07, 'learning_rate/action_model': 5.150699391863884e-06, 'epoch': 2.81}) +[2026-06-18 03:10:39,644][starVLA.training.train_starvla][INFO] - Step 6826, Loss: {'train/loss': 0.23663359880447388, 'train/grad_norm_pre_clip': 30.04425048828125, 'timing/data': 0.026710359379649162, 'timing/model': 1.1661876179277897, 'learning_rate/qwen_vl_interface': 5.148982932207052e-07, 'learning_rate/action_model': 5.148982932207052e-06, 'epoch': 2.81}) +[2026-06-18 03:10:43,398][starVLA.training.train_starvla][INFO] - Step 6827, Loss: {'train/loss': 0.2305336594581604, 'train/grad_norm_pre_clip': 19.021574020385742, 'timing/data': 0.0311479726806283, 'timing/model': 1.085015187971294, 'learning_rate/qwen_vl_interface': 5.147276288467008e-07, 'learning_rate/action_model': 5.147276288467008e-06, 'epoch': 2.81}) +[2026-06-18 03:10:47,012][starVLA.training.train_starvla][INFO] - Step 6828, Loss: {'train/loss': 0.08388049155473709, 'train/grad_norm_pre_clip': 12.721896171569824, 'timing/data': 0.044669424183666706, 'timing/model': 0.9271116834133863, 'learning_rate/qwen_vl_interface': 5.145579460997544e-07, 'learning_rate/action_model': 5.145579460997544e-06, 'epoch': 2.81}) +[2026-06-18 03:10:50,787][starVLA.training.train_starvla][INFO] - Step 6829, Loss: {'train/loss': 0.10169124603271484, 'train/grad_norm_pre_clip': 18.72171401977539, 'timing/data': 0.046687533147633076, 'timing/model': 0.9269443219527602, 'learning_rate/qwen_vl_interface': 5.143892450150411e-07, 'learning_rate/action_model': 5.143892450150411e-06, 'epoch': 2.81}) +[2026-06-18 03:10:54,571][starVLA.training.train_starvla][INFO] - Step 6830, Loss: {'train/loss': 0.4910818636417389, 'train/grad_norm_pre_clip': 37.373924255371094, 'timing/data': 0.03947345074266195, 'timing/model': 0.929801981896162, 'learning_rate/qwen_vl_interface': 5.142215256275338e-07, 'learning_rate/action_model': 5.142215256275338e-06, 'epoch': 2.81}) +[2026-06-18 03:10:58,355][starVLA.training.train_starvla][INFO] - Step 6831, Loss: {'train/loss': 0.1404852718114853, 'train/grad_norm_pre_clip': 18.408124923706055, 'timing/data': 0.03823923319578171, 'timing/model': 0.9214231921359897, 'learning_rate/qwen_vl_interface': 5.140547879719993e-07, 'learning_rate/action_model': 5.140547879719993e-06, 'epoch': 2.81}) +[2026-06-18 03:11:02,163][starVLA.training.train_starvla][INFO] - Step 6832, Loss: {'train/loss': 0.1191440224647522, 'train/grad_norm_pre_clip': 18.48851203918457, 'timing/data': 0.04655888397246599, 'timing/model': 0.9256773898378015, 'learning_rate/qwen_vl_interface': 5.138890320830038e-07, 'learning_rate/action_model': 5.138890320830038e-06, 'epoch': 2.81}) +[2026-06-18 03:11:05,925][starVLA.training.train_starvla][INFO] - Step 6833, Loss: {'train/loss': 0.2859783172607422, 'train/grad_norm_pre_clip': 28.503936767578125, 'timing/data': 0.026589836925268173, 'timing/model': 1.1030363403260708, 'learning_rate/qwen_vl_interface': 5.137242579949081e-07, 'learning_rate/action_model': 5.137242579949081e-06, 'epoch': 2.81}) +[2026-06-18 03:11:09,502][starVLA.training.train_starvla][INFO] - Step 6834, Loss: {'train/loss': 0.2052546739578247, 'train/grad_norm_pre_clip': 19.420063018798828, 'timing/data': 0.02293169591575861, 'timing/model': 0.9246824095025659, 'learning_rate/qwen_vl_interface': 5.135604657418702e-07, 'learning_rate/action_model': 5.135604657418702e-06, 'epoch': 2.81}) +[2026-06-18 03:11:13,293][starVLA.training.train_starvla][INFO] - Step 6835, Loss: {'train/loss': 0.35351911187171936, 'train/grad_norm_pre_clip': 33.8953971862793, 'timing/data': 0.022353217005729675, 'timing/model': 0.9259011028334498, 'learning_rate/qwen_vl_interface': 5.133976553578443e-07, 'learning_rate/action_model': 5.1339765535784425e-06, 'epoch': 2.81}) +[2026-06-18 03:11:17,022][starVLA.training.train_starvla][INFO] - Step 6836, Loss: {'train/loss': 0.8479215502738953, 'train/grad_norm_pre_clip': 39.5529670715332, 'timing/data': 0.03081634361296892, 'timing/model': 0.9264131812378764, 'learning_rate/qwen_vl_interface': 5.132358268765813e-07, 'learning_rate/action_model': 5.132358268765813e-06, 'epoch': 2.81}) +[2026-06-18 03:11:20,785][starVLA.training.train_starvla][INFO] - Step 6837, Loss: {'train/loss': 0.1927390992641449, 'train/grad_norm_pre_clip': 22.527734756469727, 'timing/data': 0.026139957830309868, 'timing/model': 0.9227584907785058, 'learning_rate/qwen_vl_interface': 5.130749803316284e-07, 'learning_rate/action_model': 5.130749803316284e-06, 'epoch': 2.81}) +[2026-06-18 03:11:24,634][starVLA.training.train_starvla][INFO] - Step 6838, Loss: {'train/loss': 0.09262841194868088, 'train/grad_norm_pre_clip': 13.427694320678711, 'timing/data': 0.02492502424865961, 'timing/model': 0.9268990093842149, 'learning_rate/qwen_vl_interface': 5.129151157563294e-07, 'learning_rate/action_model': 5.129151157563293e-06, 'epoch': 2.81}) +[2026-06-18 03:11:28,396][starVLA.training.train_starvla][INFO] - Step 6839, Loss: {'train/loss': 0.26317456364631653, 'train/grad_norm_pre_clip': 31.559375762939453, 'timing/data': 0.025199145078659058, 'timing/model': 0.9211037904024124, 'learning_rate/qwen_vl_interface': 5.127562331838241e-07, 'learning_rate/action_model': 5.12756233183824e-06, 'epoch': 2.82}) +[2026-06-18 03:11:32,676][starVLA.training.train_starvla][INFO] - Step 6840, Loss: {'train/loss': 0.08688365668058395, 'train/grad_norm_pre_clip': 12.826254844665527, 'timing/data': 0.024817341938614845, 'timing/model': 1.152701591141522, 'learning_rate/qwen_vl_interface': 5.125983326470492e-07, 'learning_rate/action_model': 5.125983326470492e-06, 'epoch': 2.82}) +[2026-06-18 03:11:36,471][starVLA.training.train_starvla][INFO] - Step 6841, Loss: {'train/loss': 0.20049768686294556, 'train/grad_norm_pre_clip': 17.441808700561523, 'timing/data': 0.02620418183505535, 'timing/model': 0.9331340901553631, 'learning_rate/qwen_vl_interface': 5.124414141787372e-07, 'learning_rate/action_model': 5.124414141787372e-06, 'epoch': 2.82}) +[2026-06-18 03:11:40,264][starVLA.training.train_starvla][INFO] - Step 6842, Loss: {'train/loss': 0.10987992584705353, 'train/grad_norm_pre_clip': 12.216513633728027, 'timing/data': 0.02459467016160488, 'timing/model': 0.9323580143973231, 'learning_rate/qwen_vl_interface': 5.122854778114185e-07, 'learning_rate/action_model': 5.122854778114185e-06, 'epoch': 2.82}) +[2026-06-18 03:11:44,116][starVLA.training.train_starvla][INFO] - Step 6843, Loss: {'train/loss': 0.11159573495388031, 'train/grad_norm_pre_clip': 18.828594207763672, 'timing/data': 0.027041982859373093, 'timing/model': 0.998299808241427, 'learning_rate/qwen_vl_interface': 5.12130523577418e-07, 'learning_rate/action_model': 5.12130523577418e-06, 'epoch': 2.82}) +[2026-06-18 03:11:47,972][starVLA.training.train_starvla][INFO] - Step 6844, Loss: {'train/loss': 0.23027218878269196, 'train/grad_norm_pre_clip': 31.202987670898438, 'timing/data': 0.0261303149163723, 'timing/model': 1.0068691773340106, 'learning_rate/qwen_vl_interface': 5.119765515088579e-07, 'learning_rate/action_model': 5.1197655150885785e-06, 'epoch': 2.82}) +[2026-06-18 03:11:51,952][starVLA.training.train_starvla][INFO] - Step 6845, Loss: {'train/loss': 0.0541706420481205, 'train/grad_norm_pre_clip': 8.558329582214355, 'timing/data': 0.02920929715037346, 'timing/model': 1.113356634043157, 'learning_rate/qwen_vl_interface': 5.11823561637657e-07, 'learning_rate/action_model': 5.1182356163765705e-06, 'epoch': 2.82}) +[2026-06-18 03:11:55,722][starVLA.training.train_starvla][INFO] - Step 6846, Loss: {'train/loss': 0.5154818892478943, 'train/grad_norm_pre_clip': 26.1479434967041, 'timing/data': 0.028928568586707115, 'timing/model': 1.101652255281806, 'learning_rate/qwen_vl_interface': 5.116715539955306e-07, 'learning_rate/action_model': 5.116715539955305e-06, 'epoch': 2.82}) +[2026-06-18 03:11:59,800][starVLA.training.train_starvla][INFO] - Step 6847, Loss: {'train/loss': 0.24351796507835388, 'train/grad_norm_pre_clip': 20.842151641845703, 'timing/data': 0.0470550861209631, 'timing/model': 1.1068383529782295, 'learning_rate/qwen_vl_interface': 5.115205286139895e-07, 'learning_rate/action_model': 5.115205286139895e-06, 'epoch': 2.82}) +[2026-06-18 03:12:03,408][starVLA.training.train_starvla][INFO] - Step 6848, Loss: {'train/loss': 0.19586800038814545, 'train/grad_norm_pre_clip': 14.71487045288086, 'timing/data': 0.026483358815312386, 'timing/model': 0.9412999246269464, 'learning_rate/qwen_vl_interface': 5.11370485524342e-07, 'learning_rate/action_model': 5.1137048552434205e-06, 'epoch': 2.82}) +[2026-06-18 03:12:07,288][starVLA.training.train_starvla][INFO] - Step 6849, Loss: {'train/loss': 0.16733390092849731, 'train/grad_norm_pre_clip': 26.1590576171875, 'timing/data': 0.03063226956874132, 'timing/model': 0.9371722396463156, 'learning_rate/qwen_vl_interface': 5.112214247576915e-07, 'learning_rate/action_model': 5.1122142475769155e-06, 'epoch': 2.82}) +[2026-06-18 03:12:11,197][starVLA.training.train_starvla][INFO] - Step 6850, Loss: {'train/loss': 0.0917162373661995, 'train/grad_norm_pre_clip': 22.864286422729492, 'timing/data': 0.03646452724933624, 'timing/model': 0.9378921231254935, 'learning_rate/qwen_vl_interface': 5.110733463449384e-07, 'learning_rate/action_model': 5.110733463449385e-06, 'epoch': 2.82}) +[2026-06-18 03:12:14,969][starVLA.training.train_starvla][INFO] - Step 6851, Loss: {'train/loss': 0.22743837535381317, 'train/grad_norm_pre_clip': 23.639808654785156, 'timing/data': 0.024676376953721046, 'timing/model': 0.9305880088359118, 'learning_rate/qwen_vl_interface': 5.109262503167808e-07, 'learning_rate/action_model': 5.109262503167808e-06, 'epoch': 2.82}) +[2026-06-18 03:12:18,737][starVLA.training.train_starvla][INFO] - Step 6852, Loss: {'train/loss': 0.1679365187883377, 'train/grad_norm_pre_clip': 19.324230194091797, 'timing/data': 0.029644319787621498, 'timing/model': 0.9283500704914331, 'learning_rate/qwen_vl_interface': 5.107801367037106e-07, 'learning_rate/action_model': 5.107801367037107e-06, 'epoch': 2.82}) +[2026-06-18 03:12:22,531][starVLA.training.train_starvla][INFO] - Step 6853, Loss: {'train/loss': 0.1913735717535019, 'train/grad_norm_pre_clip': 39.95998764038086, 'timing/data': 0.029678909108042717, 'timing/model': 1.1298615671694279, 'learning_rate/qwen_vl_interface': 5.106350055360173e-07, 'learning_rate/action_model': 5.106350055360173e-06, 'epoch': 2.82}) +[2026-06-18 03:12:26,128][starVLA.training.train_starvla][INFO] - Step 6854, Loss: {'train/loss': 0.048423368483781815, 'train/grad_norm_pre_clip': 7.667468070983887, 'timing/data': 0.02552506048232317, 'timing/model': 0.9334902707487345, 'learning_rate/qwen_vl_interface': 5.104908568437875e-07, 'learning_rate/action_model': 5.104908568437875e-06, 'epoch': 2.82}) +[2026-06-18 03:12:29,900][starVLA.training.train_starvla][INFO] - Step 6855, Loss: {'train/loss': 0.26868289709091187, 'train/grad_norm_pre_clip': 21.98764419555664, 'timing/data': 0.025392952375113964, 'timing/model': 0.9254803964868188, 'learning_rate/qwen_vl_interface': 5.103476906569034e-07, 'learning_rate/action_model': 5.103476906569033e-06, 'epoch': 2.82}) +[2026-06-18 03:12:33,644][starVLA.training.train_starvla][INFO] - Step 6856, Loss: {'train/loss': 0.19304779171943665, 'train/grad_norm_pre_clip': 16.216646194458008, 'timing/data': 0.023842130787670612, 'timing/model': 0.9225068064406514, 'learning_rate/qwen_vl_interface': 5.102055070050428e-07, 'learning_rate/action_model': 5.102055070050427e-06, 'epoch': 2.82}) +[2026-06-18 03:12:37,448][starVLA.training.train_starvla][INFO] - Step 6857, Loss: {'train/loss': 0.09106811136007309, 'train/grad_norm_pre_clip': 10.67423152923584, 'timing/data': 0.030755355022847652, 'timing/model': 0.9292409466579556, 'learning_rate/qwen_vl_interface': 5.100643059176803e-07, 'learning_rate/action_model': 5.100643059176802e-06, 'epoch': 2.82}) +[2026-06-18 03:12:41,392][starVLA.training.train_starvla][INFO] - Step 6858, Loss: {'train/loss': 0.11746417731046677, 'train/grad_norm_pre_clip': 10.154999732971191, 'timing/data': 0.03926186170428991, 'timing/model': 0.9620892638340592, 'learning_rate/qwen_vl_interface': 5.099240874240883e-07, 'learning_rate/action_model': 5.099240874240883e-06, 'epoch': 2.82}) +[2026-06-18 03:12:45,203][starVLA.training.train_starvla][INFO] - Step 6859, Loss: {'train/loss': 0.1702372431755066, 'train/grad_norm_pre_clip': 19.144819259643555, 'timing/data': 0.023682480677962303, 'timing/model': 0.9330149507150054, 'learning_rate/qwen_vl_interface': 5.097848515533331e-07, 'learning_rate/action_model': 5.0978485155333305e-06, 'epoch': 2.82}) +[2026-06-18 03:12:49,066][starVLA.training.train_starvla][INFO] - Step 6860, Loss: {'train/loss': 0.045603878796100616, 'train/grad_norm_pre_clip': 9.147188186645508, 'timing/data': 0.029417191632092, 'timing/model': 1.1687543652951717, 'learning_rate/qwen_vl_interface': 5.096465983342791e-07, 'learning_rate/action_model': 5.096465983342791e-06, 'epoch': 2.82}) +[2026-06-18 03:12:53,585][starVLA.training.train_starvla][INFO] - Step 6861, Loss: {'train/loss': 0.21772953867912292, 'train/grad_norm_pre_clip': 11.417862892150879, 'timing/data': 0.031132922507822514, 'timing/model': 0.9456583661958575, 'learning_rate/qwen_vl_interface': 5.095093277955859e-07, 'learning_rate/action_model': 5.095093277955859e-06, 'epoch': 2.82}) +[2026-06-18 03:12:57,712][starVLA.training.train_starvla][INFO] - Step 6862, Loss: {'train/loss': 0.18409670889377594, 'train/grad_norm_pre_clip': 20.721071243286133, 'timing/data': 0.032582362182438374, 'timing/model': 1.0256958240643144, 'learning_rate/qwen_vl_interface': 5.093730399657101e-07, 'learning_rate/action_model': 5.0937303996571e-06, 'epoch': 2.82}) +[2026-06-18 03:13:01,561][starVLA.training.train_starvla][INFO] - Step 6863, Loss: {'train/loss': 0.16088345646858215, 'train/grad_norm_pre_clip': 20.700037002563477, 'timing/data': 0.02586876694113016, 'timing/model': 0.956822270527482, 'learning_rate/qwen_vl_interface': 5.092377348729043e-07, 'learning_rate/action_model': 5.092377348729043e-06, 'epoch': 2.83}) +[2026-06-18 03:13:05,411][starVLA.training.train_starvla][INFO] - Step 6864, Loss: {'train/loss': 0.12961630523204803, 'train/grad_norm_pre_clip': 21.219221115112305, 'timing/data': 0.0280405068770051, 'timing/model': 0.9403156451880932, 'learning_rate/qwen_vl_interface': 5.091034125452171e-07, 'learning_rate/action_model': 5.091034125452171e-06, 'epoch': 2.83}) +[2026-06-18 03:13:09,185][starVLA.training.train_starvla][INFO] - Step 6865, Loss: {'train/loss': 0.2701542377471924, 'train/grad_norm_pre_clip': 23.67800521850586, 'timing/data': 0.025133050978183746, 'timing/model': 0.9300581496208906, 'learning_rate/qwen_vl_interface': 5.089700730104944e-07, 'learning_rate/action_model': 5.089700730104944e-06, 'epoch': 2.83}) +[2026-06-18 03:13:13,078][starVLA.training.train_starvla][INFO] - Step 6866, Loss: {'train/loss': 0.2404424399137497, 'train/grad_norm_pre_clip': 16.250431060791016, 'timing/data': 0.03088035900145769, 'timing/model': 1.1121221128851175, 'learning_rate/qwen_vl_interface': 5.08837716296377e-07, 'learning_rate/action_model': 5.0883771629637696e-06, 'epoch': 2.83}) +[2026-06-18 03:13:16,926][starVLA.training.train_starvla][INFO] - Step 6867, Loss: {'train/loss': 0.14418230950832367, 'train/grad_norm_pre_clip': 23.406049728393555, 'timing/data': 0.024000201374292374, 'timing/model': 1.1208537686616182, 'learning_rate/qwen_vl_interface': 5.087063424303026e-07, 'learning_rate/action_model': 5.087063424303026e-06, 'epoch': 2.83}) +[2026-06-18 03:13:20,686][starVLA.training.train_starvla][INFO] - Step 6868, Loss: {'train/loss': 0.28564906120300293, 'train/grad_norm_pre_clip': 27.030872344970703, 'timing/data': 0.02413714211434126, 'timing/model': 0.9322370067238808, 'learning_rate/qwen_vl_interface': 5.085759514395052e-07, 'learning_rate/action_model': 5.085759514395051e-06, 'epoch': 2.83}) +[2026-06-18 03:13:24,453][starVLA.training.train_starvla][INFO] - Step 6869, Loss: {'train/loss': 0.10747741162776947, 'train/grad_norm_pre_clip': 14.026884078979492, 'timing/data': 0.025672603398561478, 'timing/model': 0.9378690170124173, 'learning_rate/qwen_vl_interface': 5.084465433510151e-07, 'learning_rate/action_model': 5.0844654335101515e-06, 'epoch': 2.83}) +[2026-06-18 03:13:28,209][starVLA.training.train_starvla][INFO] - Step 6870, Loss: {'train/loss': 0.3823317587375641, 'train/grad_norm_pre_clip': 31.30290412902832, 'timing/data': 0.03324510809034109, 'timing/model': 0.9289763998240232, 'learning_rate/qwen_vl_interface': 5.083181181916588e-07, 'learning_rate/action_model': 5.083181181916588e-06, 'epoch': 2.83}) +[2026-06-18 03:13:31,961][starVLA.training.train_starvla][INFO] - Step 6871, Loss: {'train/loss': 0.23715338110923767, 'train/grad_norm_pre_clip': 34.41651916503906, 'timing/data': 0.031197356060147285, 'timing/model': 0.9301877887919545, 'learning_rate/qwen_vl_interface': 5.081906759880585e-07, 'learning_rate/action_model': 5.081906759880585e-06, 'epoch': 2.83}) +[2026-06-18 03:13:35,743][starVLA.training.train_starvla][INFO] - Step 6872, Loss: {'train/loss': 0.31257888674736023, 'train/grad_norm_pre_clip': 18.302583694458008, 'timing/data': 0.024778297170996666, 'timing/model': 0.9438599776476622, 'learning_rate/qwen_vl_interface': 5.080642167666338e-07, 'learning_rate/action_model': 5.080642167666337e-06, 'epoch': 2.83}) +[2026-06-18 03:13:39,614][starVLA.training.train_starvla][INFO] - Step 6873, Loss: {'train/loss': 0.32819586992263794, 'train/grad_norm_pre_clip': 27.352149963378906, 'timing/data': 0.030242854729294777, 'timing/model': 1.1486513689160347, 'learning_rate/qwen_vl_interface': 5.079387405535995e-07, 'learning_rate/action_model': 5.079387405535995e-06, 'epoch': 2.83}) +[2026-06-18 03:13:43,195][starVLA.training.train_starvla][INFO] - Step 6874, Loss: {'train/loss': 0.1566193848848343, 'train/grad_norm_pre_clip': 15.299357414245605, 'timing/data': 0.025232932530343533, 'timing/model': 0.9200945245102048, 'learning_rate/qwen_vl_interface': 5.07814247374967e-07, 'learning_rate/action_model': 5.078142473749671e-06, 'epoch': 2.83}) +[2026-06-18 03:13:46,959][starVLA.training.train_starvla][INFO] - Step 6875, Loss: {'train/loss': 0.19151704013347626, 'train/grad_norm_pre_clip': 34.907875061035156, 'timing/data': 0.02320124302059412, 'timing/model': 0.9325415743514895, 'learning_rate/qwen_vl_interface': 5.07690737256544e-07, 'learning_rate/action_model': 5.07690737256544e-06, 'epoch': 2.83}) +[2026-06-18 03:13:50,646][starVLA.training.train_starvla][INFO] - Step 6876, Loss: {'train/loss': 0.15037189424037933, 'train/grad_norm_pre_clip': 20.348018646240234, 'timing/data': 0.027186197228729725, 'timing/model': 0.9253791775554419, 'learning_rate/qwen_vl_interface': 5.075682102239336e-07, 'learning_rate/action_model': 5.075682102239335e-06, 'epoch': 2.83}) +[2026-06-18 03:13:54,392][starVLA.training.train_starvla][INFO] - Step 6877, Loss: {'train/loss': 0.3868703246116638, 'train/grad_norm_pre_clip': 30.10742950439453, 'timing/data': 0.024496115744113922, 'timing/model': 0.9251040397211909, 'learning_rate/qwen_vl_interface': 5.074466663025369e-07, 'learning_rate/action_model': 5.074466663025369e-06, 'epoch': 2.83}) +[2026-06-18 03:13:58,189][starVLA.training.train_starvla][INFO] - Step 6878, Loss: {'train/loss': 0.16611000895500183, 'train/grad_norm_pre_clip': 26.451929092407227, 'timing/data': 0.031002460978925228, 'timing/model': 0.9382490115240216, 'learning_rate/qwen_vl_interface': 5.07326105517549e-07, 'learning_rate/action_model': 5.07326105517549e-06, 'epoch': 2.83}) +[2026-06-18 03:14:02,053][starVLA.training.train_starvla][INFO] - Step 6879, Loss: {'train/loss': 0.266167551279068, 'train/grad_norm_pre_clip': 25.38318634033203, 'timing/data': 0.02884687762707472, 'timing/model': 1.1865718802437186, 'learning_rate/qwen_vl_interface': 5.072065278939628e-07, 'learning_rate/action_model': 5.072065278939627e-06, 'epoch': 2.83}) +[2026-06-18 03:14:05,859][starVLA.training.train_starvla][INFO] - Step 6880, Loss: {'train/loss': 0.15516169369220734, 'train/grad_norm_pre_clip': 10.831933975219727, 'timing/data': 0.030008472502231598, 'timing/model': 1.0993771543726325, 'learning_rate/qwen_vl_interface': 5.070879334565671e-07, 'learning_rate/action_model': 5.070879334565671e-06, 'epoch': 2.83}) +[2026-06-18 03:14:09,602][starVLA.training.train_starvla][INFO] - Step 6881, Loss: {'train/loss': 0.09236938506364822, 'train/grad_norm_pre_clip': 15.052770614624023, 'timing/data': 0.022777151316404343, 'timing/model': 0.9857323439791799, 'learning_rate/qwen_vl_interface': 5.069703222299464e-07, 'learning_rate/action_model': 5.069703222299463e-06, 'epoch': 2.83}) +[2026-06-18 03:14:13,397][starVLA.training.train_starvla][INFO] - Step 6882, Loss: {'train/loss': 0.0418785996735096, 'train/grad_norm_pre_clip': 8.269347190856934, 'timing/data': 0.030633294954895973, 'timing/model': 0.9300878895446658, 'learning_rate/qwen_vl_interface': 5.068536942384813e-07, 'learning_rate/action_model': 5.068536942384812e-06, 'epoch': 2.83}) +[2026-06-18 03:14:17,289][starVLA.training.train_starvla][INFO] - Step 6883, Loss: {'train/loss': 0.20804496109485626, 'train/grad_norm_pre_clip': 26.601638793945312, 'timing/data': 0.02944505214691162, 'timing/model': 0.9376021288335323, 'learning_rate/qwen_vl_interface': 5.067380495063494e-07, 'learning_rate/action_model': 5.067380495063494e-06, 'epoch': 2.83}) +[2026-06-18 03:14:21,112][starVLA.training.train_starvla][INFO] - Step 6884, Loss: {'train/loss': 0.17577806115150452, 'train/grad_norm_pre_clip': 28.531917572021484, 'timing/data': 0.02684345468878746, 'timing/model': 0.9209431260824203, 'learning_rate/qwen_vl_interface': 5.066233880575239e-07, 'learning_rate/action_model': 5.0662338805752395e-06, 'epoch': 2.83}) +[2026-06-18 03:14:24,876][starVLA.training.train_starvla][INFO] - Step 6885, Loss: {'train/loss': 0.09744640439748764, 'train/grad_norm_pre_clip': 14.490395545959473, 'timing/data': 0.022974397987127304, 'timing/model': 0.9249827610328794, 'learning_rate/qwen_vl_interface': 5.065097099157743e-07, 'learning_rate/action_model': 5.065097099157743e-06, 'epoch': 2.83}) +[2026-06-18 03:14:28,641][starVLA.training.train_starvla][INFO] - Step 6886, Loss: {'train/loss': 0.25273576378822327, 'train/grad_norm_pre_clip': 22.576040267944336, 'timing/data': 0.024600734934210777, 'timing/model': 1.114352266304195, 'learning_rate/qwen_vl_interface': 5.063970151046654e-07, 'learning_rate/action_model': 5.0639701510466545e-06, 'epoch': 2.83}) +[2026-06-18 03:14:32,282][starVLA.training.train_starvla][INFO] - Step 6887, Loss: {'train/loss': 0.24016502499580383, 'train/grad_norm_pre_clip': 22.111047744750977, 'timing/data': 0.02806820347905159, 'timing/model': 0.93110456969589, 'learning_rate/qwen_vl_interface': 5.062853036475604e-07, 'learning_rate/action_model': 5.062853036475603e-06, 'epoch': 2.83}) +[2026-06-18 03:14:36,052][starVLA.training.train_starvla][INFO] - Step 6888, Loss: {'train/loss': 0.17281512916088104, 'train/grad_norm_pre_clip': 23.444622039794922, 'timing/data': 0.029072197154164314, 'timing/model': 0.9268273105844855, 'learning_rate/qwen_vl_interface': 5.061745755676161e-07, 'learning_rate/action_model': 5.061745755676161e-06, 'epoch': 2.84}) +[2026-06-18 03:14:39,854][starVLA.training.train_starvla][INFO] - Step 6889, Loss: {'train/loss': 0.036111146211624146, 'train/grad_norm_pre_clip': 5.870105743408203, 'timing/data': 0.02835894748568535, 'timing/model': 0.9318407149985433, 'learning_rate/qwen_vl_interface': 5.060648308877866e-07, 'learning_rate/action_model': 5.0606483088778656e-06, 'epoch': 2.84}) +[2026-06-18 03:14:43,662][starVLA.training.train_starvla][INFO] - Step 6890, Loss: {'train/loss': 0.4672173261642456, 'train/grad_norm_pre_clip': 29.610212326049805, 'timing/data': 0.023682687431573868, 'timing/model': 0.9344238424673676, 'learning_rate/qwen_vl_interface': 5.059560696308229e-07, 'learning_rate/action_model': 5.059560696308228e-06, 'epoch': 2.84}) +[2026-06-18 03:14:49,312][starVLA.training.train_starvla][INFO] - Step 6891, Loss: {'train/loss': 0.1856905221939087, 'train/grad_norm_pre_clip': 15.254290580749512, 'timing/data': 0.026690020225942135, 'timing/model': 0.9443766502663493, 'learning_rate/qwen_vl_interface': 5.0584829181927e-07, 'learning_rate/action_model': 5.058482918192699e-06, 'epoch': 2.84}) +[2026-06-18 03:14:53,588][starVLA.training.train_starvla][INFO] - Step 6892, Loss: {'train/loss': 0.4253227114677429, 'train/grad_norm_pre_clip': 30.797428131103516, 'timing/data': 0.023585965856909752, 'timing/model': 0.9298947155475616, 'learning_rate/qwen_vl_interface': 5.05741497475472e-07, 'learning_rate/action_model': 5.05741497475472e-06, 'epoch': 2.84}) +[2026-06-18 03:14:57,373][starVLA.training.train_starvla][INFO] - Step 6893, Loss: {'train/loss': 0.1172291561961174, 'train/grad_norm_pre_clip': 19.75454330444336, 'timing/data': 0.021907382644712925, 'timing/model': 1.1275108391419053, 'learning_rate/qwen_vl_interface': 5.056356866215663e-07, 'learning_rate/action_model': 5.056356866215662e-06, 'epoch': 2.84}) +[2026-06-18 03:15:00,973][starVLA.training.train_starvla][INFO] - Step 6894, Loss: {'train/loss': 0.22158807516098022, 'train/grad_norm_pre_clip': 18.058387756347656, 'timing/data': 0.03014425002038479, 'timing/model': 0.9336957633495331, 'learning_rate/qwen_vl_interface': 5.055308592794876e-07, 'learning_rate/action_model': 5.0553085927948756e-06, 'epoch': 2.84}) +[2026-06-18 03:15:04,774][starVLA.training.train_starvla][INFO] - Step 6895, Loss: {'train/loss': 0.1416577249765396, 'train/grad_norm_pre_clip': 28.891441345214844, 'timing/data': 0.02202549483627081, 'timing/model': 0.9377240026369691, 'learning_rate/qwen_vl_interface': 5.054270154709674e-07, 'learning_rate/action_model': 5.054270154709673e-06, 'epoch': 2.84}) +[2026-06-18 03:15:08,655][starVLA.training.train_starvla][INFO] - Step 6896, Loss: {'train/loss': 0.1558087021112442, 'train/grad_norm_pre_clip': 21.72624397277832, 'timing/data': 0.029274427331984043, 'timing/model': 0.9292280096560717, 'learning_rate/qwen_vl_interface': 5.053241552175322e-07, 'learning_rate/action_model': 5.053241552175321e-06, 'epoch': 2.84}) +[2026-06-18 03:15:13,130][starVLA.training.train_starvla][INFO] - Step 6897, Loss: {'train/loss': 0.0488990880548954, 'train/grad_norm_pre_clip': 13.779932022094727, 'timing/data': 0.024189777672290802, 'timing/model': 1.2589975846931338, 'learning_rate/qwen_vl_interface': 5.05222278540505e-07, 'learning_rate/action_model': 5.05222278540505e-06, 'epoch': 2.84}) +[2026-06-18 03:15:17,294][starVLA.training.train_starvla][INFO] - Step 6898, Loss: {'train/loss': 0.11940116435289383, 'train/grad_norm_pre_clip': 14.432372093200684, 'timing/data': 0.02954200003296137, 'timing/model': 1.235870273783803, 'learning_rate/qwen_vl_interface': 5.051213854610051e-07, 'learning_rate/action_model': 5.0512138546100514e-06, 'epoch': 2.84}) +[2026-06-18 03:15:21,281][starVLA.training.train_starvla][INFO] - Step 6899, Loss: {'train/loss': 0.385162353515625, 'train/grad_norm_pre_clip': 26.39931869506836, 'timing/data': 0.02238349709659815, 'timing/model': 1.218160292133689, 'learning_rate/qwen_vl_interface': 5.050214759999478e-07, 'learning_rate/action_model': 5.050214759999478e-06, 'epoch': 2.84}) +[2026-06-18 03:17:16,966][starVLA.training.train_starvla][INFO] - Step 6900, Loss: {'train/loss': 0.38986027240753174, 'train/grad_norm_pre_clip': 22.075037002563477, 'eval/loss': 0.4302213788032532, 'eval/latency_6/loss': 0.4302213788032532, 'eval/demon_fire_precision': 0.9562460367786937, 'eval/demon_fire_recall': 0.9747899159663865, 'eval/demon_fire_f1': 0.9654289372599232, 'eval/demon_fire_rate_model': 0.8775737340011129, 'eval/demon_fire_rate_teacher': 0.860879243183083, 'eval/demon_left_precision': 0.8842975206611571, 'eval/demon_left_recall': 0.7508771929824561, 'eval/demon_left_f1': 0.8121442125237192, 'eval/demon_left_rate_model': 0.13466889259877574, 'eval/demon_left_rate_teacher': 0.15859766277128548, 'eval/demon_right_precision': 0.8017883755588674, 'eval/demon_right_recall': 0.82642089093702, 'eval/demon_right_f1': 0.8139183055975794, 'eval/demon_right_rate_model': 0.18670005564830272, 'eval/demon_right_rate_teacher': 0.18113522537562604, 'eval/demon_move_f1': 0.8130312590606492, 'eval/demon_cc_f1': 0.9140301487433902, 'eval/latency_6/demon_fire_precision': 0.9562460367786937, 'eval/latency_6/demon_fire_recall': 0.9747899159663865, 'eval/latency_6/demon_fire_f1': 0.9654289372599232, 'eval/latency_6/demon_fire_rate_model': 0.8775737340011129, 'eval/latency_6/demon_fire_rate_teacher': 0.860879243183083, 'eval/latency_6/demon_left_precision': 0.8842975206611571, 'eval/latency_6/demon_left_recall': 0.7508771929824561, 'eval/latency_6/demon_left_f1': 0.8121442125237192, 'eval/latency_6/demon_left_rate_model': 0.13466889259877574, 'eval/latency_6/demon_left_rate_teacher': 0.15859766277128548, 'eval/latency_6/demon_right_precision': 0.8017883755588674, 'eval/latency_6/demon_right_recall': 0.82642089093702, 'eval/latency_6/demon_right_f1': 0.8139183055975794, 'eval/latency_6/demon_right_rate_model': 0.18670005564830272, 'eval/latency_6/demon_right_rate_teacher': 0.18113522537562604, 'eval/latency_6/demon_move_f1': 0.8130312590606492, 'eval/latency_6/demon_cc_f1': 0.9140301487433902, 'timing/data': 0.031923980452120304, 'timing/model': 1.1279855174943805, 'learning_rate/qwen_vl_interface': 5.049225501780441e-07, 'learning_rate/action_model': 5.049225501780441e-06, 'epoch': 2.84}) +[2026-06-18 03:17:16,970][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_6900_state +[2026-06-18 03:17:16,975][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 03:17:23,524][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_6900_state/model.safetensors +[2026-06-18 03:17:39,181][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_6900_state/optimizer.bin +[2026-06-18 03:17:39,184][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_6900_state/sampler.bin +[2026-06-18 03:17:39,184][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_6900_state/sampler_1.bin +[2026-06-18 03:17:39,186][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_6900_state/random_states_0.pkl +[2026-06-18 03:17:39,190][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 6900 +[2026-06-18 03:17:39,190][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 03:17:39,200][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 03:17:43,118][starVLA.training.train_starvla][INFO] - Step 6901, Loss: {'train/loss': 0.11278873682022095, 'train/grad_norm_pre_clip': 14.53880786895752, 'timing/data': 0.03094231989234686, 'timing/model': 1.1641761716455221, 'learning_rate/qwen_vl_interface': 5.048246080158016e-07, 'learning_rate/action_model': 5.048246080158016e-06, 'epoch': 2.84}) +[2026-06-18 03:17:46,875][starVLA.training.train_starvla][INFO] - Step 6902, Loss: {'train/loss': 0.1446571946144104, 'train/grad_norm_pre_clip': 19.113597869873047, 'timing/data': 0.017041601240634918, 'timing/model': 1.1058533629402518, 'learning_rate/qwen_vl_interface': 5.047276495335243e-07, 'learning_rate/action_model': 5.047276495335242e-06, 'epoch': 2.84}) +[2026-06-18 03:17:50,506][starVLA.training.train_starvla][INFO] - Step 6903, Loss: {'train/loss': 0.11481443047523499, 'train/grad_norm_pre_clip': 17.987855911254883, 'timing/data': 0.01799482572823763, 'timing/model': 0.9286401588469744, 'learning_rate/qwen_vl_interface': 5.046316747513115e-07, 'learning_rate/action_model': 5.046316747513115e-06, 'epoch': 2.84}) +[2026-06-18 03:17:54,752][starVLA.training.train_starvla][INFO] - Step 6904, Loss: {'train/loss': 0.13941146433353424, 'train/grad_norm_pre_clip': 17.60103416442871, 'timing/data': 0.025295342318713665, 'timing/model': 0.9300330961123109, 'learning_rate/qwen_vl_interface': 5.045366836890585e-07, 'learning_rate/action_model': 5.045366836890584e-06, 'epoch': 2.84}) +[2026-06-18 03:17:58,795][starVLA.training.train_starvla][INFO] - Step 6905, Loss: {'train/loss': 0.28754934668540955, 'train/grad_norm_pre_clip': 193.76751708984375, 'timing/data': 0.019985000602900982, 'timing/model': 0.9334026109427214, 'learning_rate/qwen_vl_interface': 5.044426763664572e-07, 'learning_rate/action_model': 5.044426763664571e-06, 'epoch': 2.84}) +[2026-06-18 03:18:02,709][starVLA.training.train_starvla][INFO] - Step 6906, Loss: {'train/loss': 0.11445131152868271, 'train/grad_norm_pre_clip': 20.004438400268555, 'timing/data': 0.020838672295212746, 'timing/model': 0.9997972846031189, 'learning_rate/qwen_vl_interface': 5.043496528029954e-07, 'learning_rate/action_model': 5.043496528029954e-06, 'epoch': 2.84}) +[2026-06-18 03:18:06,558][starVLA.training.train_starvla][INFO] - Step 6907, Loss: {'train/loss': 0.20353423058986664, 'train/grad_norm_pre_clip': 17.0676212310791, 'timing/data': 0.01873766351491213, 'timing/model': 0.9305844409391284, 'learning_rate/qwen_vl_interface': 5.042576130179573e-07, 'learning_rate/action_model': 5.042576130179572e-06, 'epoch': 2.84}) +[2026-06-18 03:18:10,343][starVLA.training.train_starvla][INFO] - Step 6908, Loss: {'train/loss': 0.6352726221084595, 'train/grad_norm_pre_clip': 23.60393524169922, 'timing/data': 0.019073765724897385, 'timing/model': 1.1244599958881736, 'learning_rate/qwen_vl_interface': 5.041665570304227e-07, 'learning_rate/action_model': 5.041665570304226e-06, 'epoch': 2.84}) +[2026-06-18 03:18:13,960][starVLA.training.train_starvla][INFO] - Step 6909, Loss: {'train/loss': 0.17196109890937805, 'train/grad_norm_pre_clip': 17.06495475769043, 'timing/data': 0.02707911841571331, 'timing/model': 0.926052413880825, 'learning_rate/qwen_vl_interface': 5.040764848592675e-07, 'learning_rate/action_model': 5.040764848592675e-06, 'epoch': 2.84}) +[2026-06-18 03:18:18,192][starVLA.training.train_starvla][INFO] - Step 6910, Loss: {'train/loss': 0.16993629932403564, 'train/grad_norm_pre_clip': 18.671035766601562, 'timing/data': 0.01697054412215948, 'timing/model': 0.9284629113972187, 'learning_rate/qwen_vl_interface': 5.039873965231644e-07, 'learning_rate/action_model': 5.039873965231643e-06, 'epoch': 2.84}) +[2026-06-18 03:18:22,625][starVLA.training.train_starvla][INFO] - Step 6911, Loss: {'train/loss': 0.21178928017616272, 'train/grad_norm_pre_clip': 26.65641212463379, 'timing/data': 0.017602252773940563, 'timing/model': 1.4605578975751996, 'learning_rate/qwen_vl_interface': 5.0389929204058e-07, 'learning_rate/action_model': 5.0389929204058e-06, 'epoch': 2.84}) +[2026-06-18 03:18:26,455][starVLA.training.train_starvla][INFO] - Step 6912, Loss: {'train/loss': 0.059729088097810745, 'train/grad_norm_pre_clip': 10.785419464111328, 'timing/data': 0.02523021772503853, 'timing/model': 0.9556169807910919, 'learning_rate/qwen_vl_interface': 5.038121714297802e-07, 'learning_rate/action_model': 5.038121714297802e-06, 'epoch': 2.85}) +[2026-06-18 03:18:30,385][starVLA.training.train_starvla][INFO] - Step 6913, Loss: {'train/loss': 0.23639622330665588, 'train/grad_norm_pre_clip': 41.5189094543457, 'timing/data': 0.01673542521893978, 'timing/model': 0.9522547265514731, 'learning_rate/qwen_vl_interface': 5.037260347088236e-07, 'learning_rate/action_model': 5.037260347088236e-06, 'epoch': 2.85}) +[2026-06-18 03:18:34,245][starVLA.training.train_starvla][INFO] - Step 6914, Loss: {'train/loss': 0.1918039321899414, 'train/grad_norm_pre_clip': 21.593311309814453, 'timing/data': 0.01911361049860716, 'timing/model': 1.131315354257822, 'learning_rate/qwen_vl_interface': 5.03640881895568e-07, 'learning_rate/action_model': 5.03640881895568e-06, 'epoch': 2.85}) +[2026-06-18 03:18:37,890][starVLA.training.train_starvla][INFO] - Step 6915, Loss: {'train/loss': 0.20004339516162872, 'train/grad_norm_pre_clip': 16.887832641601562, 'timing/data': 0.02130876760929823, 'timing/model': 0.9452640181407332, 'learning_rate/qwen_vl_interface': 5.035567130076642e-07, 'learning_rate/action_model': 5.035567130076642e-06, 'epoch': 2.85}) +[2026-06-18 03:18:41,765][starVLA.training.train_starvla][INFO] - Step 6916, Loss: {'train/loss': 0.05042605474591255, 'train/grad_norm_pre_clip': 8.803319931030273, 'timing/data': 0.024708038195967674, 'timing/model': 0.9335238384082913, 'learning_rate/qwen_vl_interface': 5.034735280625617e-07, 'learning_rate/action_model': 5.034735280625616e-06, 'epoch': 2.85}) +[2026-06-18 03:18:45,643][starVLA.training.train_starvla][INFO] - Step 6917, Loss: {'train/loss': 0.3218329846858978, 'train/grad_norm_pre_clip': 30.2275333404541, 'timing/data': 0.028746825642883778, 'timing/model': 0.9351968578994274, 'learning_rate/qwen_vl_interface': 5.033913270775039e-07, 'learning_rate/action_model': 5.033913270775039e-06, 'epoch': 2.85}) +[2026-06-18 03:18:49,561][starVLA.training.train_starvla][INFO] - Step 6918, Loss: {'train/loss': 0.07527508586645126, 'train/grad_norm_pre_clip': 11.861593246459961, 'timing/data': 0.019424689933657646, 'timing/model': 0.9624749887734652, 'learning_rate/qwen_vl_interface': 5.03310110069532e-07, 'learning_rate/action_model': 5.03310110069532e-06, 'epoch': 2.85}) +[2026-06-18 03:18:53,460][starVLA.training.train_starvla][INFO] - Step 6919, Loss: {'train/loss': 0.059297140687704086, 'train/grad_norm_pre_clip': 11.384870529174805, 'timing/data': 0.014796541072428226, 'timing/model': 0.9345285641029477, 'learning_rate/qwen_vl_interface': 5.032298770554818e-07, 'learning_rate/action_model': 5.032298770554818e-06, 'epoch': 2.85}) +[2026-06-18 03:18:57,582][starVLA.training.train_starvla][INFO] - Step 6920, Loss: {'train/loss': 0.3500435948371887, 'train/grad_norm_pre_clip': 31.73916244506836, 'timing/data': 0.028539998456835747, 'timing/model': 0.9307814491912723, 'learning_rate/qwen_vl_interface': 5.031506280519855e-07, 'learning_rate/action_model': 5.031506280519855e-06, 'epoch': 2.85}) +[2026-06-18 03:19:01,779][starVLA.training.train_starvla][INFO] - Step 6921, Loss: {'train/loss': 0.18934020400047302, 'train/grad_norm_pre_clip': 14.91193962097168, 'timing/data': 0.020273640751838684, 'timing/model': 1.4890666417777538, 'learning_rate/qwen_vl_interface': 5.030723630754723e-07, 'learning_rate/action_model': 5.030723630754723e-06, 'epoch': 2.85}) +[2026-06-18 03:19:05,443][starVLA.training.train_starvla][INFO] - Step 6922, Loss: {'train/loss': 0.17993299663066864, 'train/grad_norm_pre_clip': 12.018721580505371, 'timing/data': 0.015966554172337055, 'timing/model': 0.9372775657102466, 'learning_rate/qwen_vl_interface': 5.02995082142166e-07, 'learning_rate/action_model': 5.029950821421659e-06, 'epoch': 2.85}) +[2026-06-18 03:19:09,246][starVLA.training.train_starvla][INFO] - Step 6923, Loss: {'train/loss': 0.18319182097911835, 'train/grad_norm_pre_clip': 26.907773971557617, 'timing/data': 0.016453455202281475, 'timing/model': 0.9330458026379347, 'learning_rate/qwen_vl_interface': 5.029187852680869e-07, 'learning_rate/action_model': 5.029187852680868e-06, 'epoch': 2.85}) +[2026-06-18 03:19:13,071][starVLA.training.train_starvla][INFO] - Step 6924, Loss: {'train/loss': 0.30634140968322754, 'train/grad_norm_pre_clip': 28.97015380859375, 'timing/data': 0.01813990343362093, 'timing/model': 0.9384185001254082, 'learning_rate/qwen_vl_interface': 5.028434724690521e-07, 'learning_rate/action_model': 5.028434724690521e-06, 'epoch': 2.85}) +[2026-06-18 03:19:17,003][starVLA.training.train_starvla][INFO] - Step 6925, Loss: {'train/loss': 0.20040126144886017, 'train/grad_norm_pre_clip': 18.787731170654297, 'timing/data': 0.017118041403591633, 'timing/model': 0.9610043922439218, 'learning_rate/qwen_vl_interface': 5.027691437606734e-07, 'learning_rate/action_model': 5.027691437606733e-06, 'epoch': 2.85}) +[2026-06-18 03:19:21,071][starVLA.training.train_starvla][INFO] - Step 6926, Loss: {'train/loss': 0.529777467250824, 'train/grad_norm_pre_clip': 29.94642448425293, 'timing/data': 0.01699434407055378, 'timing/model': 0.9613892994821072, 'learning_rate/qwen_vl_interface': 5.026957991583597e-07, 'learning_rate/action_model': 5.0269579915835965e-06, 'epoch': 2.85}) +[2026-06-18 03:19:24,844][starVLA.training.train_starvla][INFO] - Step 6927, Loss: {'train/loss': 0.14070674777030945, 'train/grad_norm_pre_clip': 19.9715576171875, 'timing/data': 0.021449114196002483, 'timing/model': 1.1070528933778405, 'learning_rate/qwen_vl_interface': 5.026234386773145e-07, 'learning_rate/action_model': 5.026234386773145e-06, 'epoch': 2.85}) +[2026-06-18 03:19:28,898][starVLA.training.train_starvla][INFO] - Step 6928, Loss: {'train/loss': 0.1381765902042389, 'train/grad_norm_pre_clip': 21.035837173461914, 'timing/data': 0.022089075297117233, 'timing/model': 1.0923290709033608, 'learning_rate/qwen_vl_interface': 5.025520623325398e-07, 'learning_rate/action_model': 5.025520623325397e-06, 'epoch': 2.85}) +[2026-06-18 03:19:32,499][starVLA.training.train_starvla][INFO] - Step 6929, Loss: {'train/loss': 0.1673138290643692, 'train/grad_norm_pre_clip': 23.464580535888672, 'timing/data': 0.0318740913644433, 'timing/model': 0.9227222222834826, 'learning_rate/qwen_vl_interface': 5.024816701388302e-07, 'learning_rate/action_model': 5.024816701388302e-06, 'epoch': 2.85}) +[2026-06-18 03:19:36,377][starVLA.training.train_starvla][INFO] - Step 6930, Loss: {'train/loss': 0.2859867513179779, 'train/grad_norm_pre_clip': 29.617948532104492, 'timing/data': 0.022835329174995422, 'timing/model': 0.9343676595017314, 'learning_rate/qwen_vl_interface': 5.024122621107795e-07, 'learning_rate/action_model': 5.024122621107795e-06, 'epoch': 2.85}) +[2026-06-18 03:19:40,227][starVLA.training.train_starvla][INFO] - Step 6931, Loss: {'train/loss': 0.10705214738845825, 'train/grad_norm_pre_clip': 17.17848777770996, 'timing/data': 0.01794192846864462, 'timing/model': 0.9296607533469796, 'learning_rate/qwen_vl_interface': 5.023438382627749e-07, 'learning_rate/action_model': 5.0234383826277486e-06, 'epoch': 2.85}) +[2026-06-18 03:19:44,501][starVLA.training.train_starvla][INFO] - Step 6932, Loss: {'train/loss': 0.44907450675964355, 'train/grad_norm_pre_clip': 29.157974243164062, 'timing/data': 0.02058440539985895, 'timing/model': 0.9333577584475279, 'learning_rate/qwen_vl_interface': 5.022763986090018e-07, 'learning_rate/action_model': 5.022763986090018e-06, 'epoch': 2.85}) +[2026-06-18 03:19:48,534][starVLA.training.train_starvla][INFO] - Step 6933, Loss: {'train/loss': 0.29219114780426025, 'train/grad_norm_pre_clip': 28.051551818847656, 'timing/data': 0.015097388066351414, 'timing/model': 0.9819162404164672, 'learning_rate/qwen_vl_interface': 5.022099431634398e-07, 'learning_rate/action_model': 5.022099431634398e-06, 'epoch': 2.85}) +[2026-06-18 03:19:52,495][starVLA.training.train_starvla][INFO] - Step 6934, Loss: {'train/loss': 0.24399815499782562, 'train/grad_norm_pre_clip': 20.223651885986328, 'timing/data': 0.018569663166999817, 'timing/model': 1.2309281462803483, 'learning_rate/qwen_vl_interface': 5.021444719398656e-07, 'learning_rate/action_model': 5.021444719398656e-06, 'epoch': 2.85}) +[2026-06-18 03:19:56,189][starVLA.training.train_starvla][INFO] - Step 6935, Loss: {'train/loss': 0.13938438892364502, 'train/grad_norm_pre_clip': 17.640483856201172, 'timing/data': 0.019558938220143318, 'timing/model': 0.9377326751127839, 'learning_rate/qwen_vl_interface': 5.020799849518513e-07, 'learning_rate/action_model': 5.020799849518513e-06, 'epoch': 2.85}) +[2026-06-18 03:19:59,990][starVLA.training.train_starvla][INFO] - Step 6936, Loss: {'train/loss': 0.28780046105384827, 'train/grad_norm_pre_clip': 19.73263168334961, 'timing/data': 0.016694147139787674, 'timing/model': 0.9348067296668887, 'learning_rate/qwen_vl_interface': 5.020164822127652e-07, 'learning_rate/action_model': 5.020164822127652e-06, 'epoch': 2.86}) +[2026-06-18 03:20:03,763][starVLA.training.train_starvla][INFO] - Step 6937, Loss: {'train/loss': 0.5127778649330139, 'train/grad_norm_pre_clip': 29.40180778503418, 'timing/data': 0.017167496494948864, 'timing/model': 0.9467056849971414, 'learning_rate/qwen_vl_interface': 5.019539637357715e-07, 'learning_rate/action_model': 5.019539637357714e-06, 'epoch': 2.86}) +[2026-06-18 03:20:07,666][starVLA.training.train_starvla][INFO] - Step 6938, Loss: {'train/loss': 0.29811891913414, 'train/grad_norm_pre_clip': 30.201108932495117, 'timing/data': 0.017575977370142937, 'timing/model': 0.9444489926099777, 'learning_rate/qwen_vl_interface': 5.0189242953383e-07, 'learning_rate/action_model': 5.018924295338299e-06, 'epoch': 2.86}) +[2026-06-18 03:20:11,678][starVLA.training.train_starvla][INFO] - Step 6939, Loss: {'train/loss': 0.047220051288604736, 'train/grad_norm_pre_clip': 9.714190483093262, 'timing/data': 0.01894689816981554, 'timing/model': 1.0098561830818653, 'learning_rate/qwen_vl_interface': 5.018318796196967e-07, 'learning_rate/action_model': 5.018318796196967e-06, 'epoch': 2.86}) +[2026-06-18 03:20:20,154][starVLA.training.train_starvla][INFO] - Step 6940, Loss: {'train/loss': 0.24423730373382568, 'train/grad_norm_pre_clip': 22.997180938720703, 'timing/data': 0.017249817959964275, 'timing/model': 1.1216625375673175, 'learning_rate/qwen_vl_interface': 5.017723140059243e-07, 'learning_rate/action_model': 5.017723140059242e-06, 'epoch': 2.86}) +[2026-06-18 03:20:24,268][starVLA.training.train_starvla][INFO] - Step 6941, Loss: {'train/loss': 0.454786479473114, 'train/grad_norm_pre_clip': 30.374691009521484, 'timing/data': 0.01818903349339962, 'timing/model': 1.1020307810977101, 'learning_rate/qwen_vl_interface': 5.017137327048606e-07, 'learning_rate/action_model': 5.017137327048606e-06, 'epoch': 2.86}) +[2026-06-18 03:20:27,986][starVLA.training.train_starvla][INFO] - Step 6942, Loss: {'train/loss': 0.2371126115322113, 'train/grad_norm_pre_clip': 23.17139434814453, 'timing/data': 0.015466687269508839, 'timing/model': 0.938621056266129, 'learning_rate/qwen_vl_interface': 5.016561357286498e-07, 'learning_rate/action_model': 5.016561357286498e-06, 'epoch': 2.86}) +[2026-06-18 03:20:31,797][starVLA.training.train_starvla][INFO] - Step 6943, Loss: {'train/loss': 0.3045452833175659, 'train/grad_norm_pre_clip': 29.640165328979492, 'timing/data': 0.01652098260819912, 'timing/model': 0.9560199603438377, 'learning_rate/qwen_vl_interface': 5.015995230892312e-07, 'learning_rate/action_model': 5.015995230892311e-06, 'epoch': 2.86}) +[2026-06-18 03:20:36,609][starVLA.training.train_starvla][INFO] - Step 6944, Loss: {'train/loss': 0.38614293932914734, 'train/grad_norm_pre_clip': 28.02828598022461, 'timing/data': 0.020066925324499607, 'timing/model': 0.990303510800004, 'learning_rate/qwen_vl_interface': 5.015438947983409e-07, 'learning_rate/action_model': 5.015438947983409e-06, 'epoch': 2.86}) +[2026-06-18 03:20:40,623][starVLA.training.train_starvla][INFO] - Step 6945, Loss: {'train/loss': 0.20369453728199005, 'train/grad_norm_pre_clip': 22.55359649658203, 'timing/data': 0.01714162342250347, 'timing/model': 0.9800364710390568, 'learning_rate/qwen_vl_interface': 5.014892508675108e-07, 'learning_rate/action_model': 5.014892508675107e-06, 'epoch': 2.86}) +[2026-06-18 03:20:44,424][starVLA.training.train_starvla][INFO] - Step 6946, Loss: {'train/loss': 0.33033397793769836, 'train/grad_norm_pre_clip': 37.123592376708984, 'timing/data': 0.020949804224073887, 'timing/model': 0.9403723496943712, 'learning_rate/qwen_vl_interface': 5.014355913080683e-07, 'learning_rate/action_model': 5.014355913080683e-06, 'epoch': 2.86}) +[2026-06-18 03:20:48,347][starVLA.training.train_starvla][INFO] - Step 6947, Loss: {'train/loss': 0.247615784406662, 'train/grad_norm_pre_clip': 26.337472915649414, 'timing/data': 0.019709170795977116, 'timing/model': 1.1689515439793468, 'learning_rate/qwen_vl_interface': 5.013829161311377e-07, 'learning_rate/action_model': 5.013829161311377e-06, 'epoch': 2.86}) +[2026-06-18 03:20:52,017][starVLA.training.train_starvla][INFO] - Step 6948, Loss: {'train/loss': 0.18311776220798492, 'train/grad_norm_pre_clip': 17.320863723754883, 'timing/data': 0.01756280567497015, 'timing/model': 0.9532314734533429, 'learning_rate/qwen_vl_interface': 5.013312253476383e-07, 'learning_rate/action_model': 5.0133122534763826e-06, 'epoch': 2.86}) +[2026-06-18 03:20:55,835][starVLA.training.train_starvla][INFO] - Step 6949, Loss: {'train/loss': 0.2151903212070465, 'train/grad_norm_pre_clip': 27.605802536010742, 'timing/data': 0.02129824459552765, 'timing/model': 0.9361187592148781, 'learning_rate/qwen_vl_interface': 5.012805189682861e-07, 'learning_rate/action_model': 5.01280518968286e-06, 'epoch': 2.86}) +[2026-06-18 03:20:59,795][starVLA.training.train_starvla][INFO] - Step 6950, Loss: {'train/loss': 0.17201679944992065, 'train/grad_norm_pre_clip': 20.891796112060547, 'timing/data': 0.01834118738770485, 'timing/model': 0.9506910014897585, 'learning_rate/qwen_vl_interface': 5.012307970035915e-07, 'learning_rate/action_model': 5.012307970035915e-06, 'epoch': 2.86}) +[2026-06-18 03:21:03,641][starVLA.training.train_starvla][INFO] - Step 6951, Loss: {'train/loss': 0.1433108150959015, 'train/grad_norm_pre_clip': 17.53257179260254, 'timing/data': 0.015578323975205421, 'timing/model': 0.9472299553453922, 'learning_rate/qwen_vl_interface': 5.011820594638626e-07, 'learning_rate/action_model': 5.0118205946386256e-06, 'epoch': 2.86}) +[2026-06-18 03:21:07,561][starVLA.training.train_starvla][INFO] - Step 6952, Loss: {'train/loss': 0.24111764132976532, 'train/grad_norm_pre_clip': 33.62348175048828, 'timing/data': 0.017392397858202457, 'timing/model': 0.9442907189950347, 'learning_rate/qwen_vl_interface': 5.011343063592032e-07, 'learning_rate/action_model': 5.011343063592032e-06, 'epoch': 2.86}) +[2026-06-18 03:21:11,455][starVLA.training.train_starvla][INFO] - Step 6953, Loss: {'train/loss': 0.12940247356891632, 'train/grad_norm_pre_clip': 21.4910831451416, 'timing/data': 0.016637971624732018, 'timing/model': 0.9575482541695237, 'learning_rate/qwen_vl_interface': 5.010875376995121e-07, 'learning_rate/action_model': 5.01087537699512e-06, 'epoch': 2.86}) +[2026-06-18 03:21:15,434][starVLA.training.train_starvla][INFO] - Step 6954, Loss: {'train/loss': 0.035639651119709015, 'train/grad_norm_pre_clip': 6.684773921966553, 'timing/data': 0.01921813376247883, 'timing/model': 1.13110286090523, 'learning_rate/qwen_vl_interface': 5.010417534944839e-07, 'learning_rate/action_model': 5.010417534944838e-06, 'epoch': 2.86}) +[2026-06-18 03:21:19,413][starVLA.training.train_starvla][INFO] - Step 6955, Loss: {'train/loss': 0.23068657517433167, 'train/grad_norm_pre_clip': 20.8626651763916, 'timing/data': 0.03222733363509178, 'timing/model': 0.9419643571600318, 'learning_rate/qwen_vl_interface': 5.009969537536106e-07, 'learning_rate/action_model': 5.009969537536106e-06, 'epoch': 2.86}) +[2026-06-18 03:21:23,462][starVLA.training.train_starvla][INFO] - Step 6956, Loss: {'train/loss': 0.17349226772785187, 'train/grad_norm_pre_clip': 20.386974334716797, 'timing/data': 0.015570040792226791, 'timing/model': 1.1786297466605902, 'learning_rate/qwen_vl_interface': 5.00953138486179e-07, 'learning_rate/action_model': 5.00953138486179e-06, 'epoch': 2.86}) +[2026-06-18 03:21:31,192][starVLA.training.train_starvla][INFO] - Step 6957, Loss: {'train/loss': 0.1678725630044937, 'train/grad_norm_pre_clip': 23.931333541870117, 'timing/data': 0.017205326817929745, 'timing/model': 2.31925418227911, 'learning_rate/qwen_vl_interface': 5.009103077012723e-07, 'learning_rate/action_model': 5.009103077012723e-06, 'epoch': 2.86}) +[2026-06-18 03:21:35,244][starVLA.training.train_starvla][INFO] - Step 6958, Loss: {'train/loss': 0.25981950759887695, 'train/grad_norm_pre_clip': 23.23189926147461, 'timing/data': 0.022515103220939636, 'timing/model': 0.9476779401302338, 'learning_rate/qwen_vl_interface': 5.008684614077684e-07, 'learning_rate/action_model': 5.008684614077684e-06, 'epoch': 2.86}) +[2026-06-18 03:21:39,192][starVLA.training.train_starvla][INFO] - Step 6959, Loss: {'train/loss': 0.16687926650047302, 'train/grad_norm_pre_clip': 17.739322662353516, 'timing/data': 0.016132796183228493, 'timing/model': 0.9732546163722873, 'learning_rate/qwen_vl_interface': 5.008275996143429e-07, 'learning_rate/action_model': 5.0082759961434295e-06, 'epoch': 2.86}) +[2026-06-18 03:21:43,170][starVLA.training.train_starvla][INFO] - Step 6960, Loss: {'train/loss': 0.3174627423286438, 'train/grad_norm_pre_clip': 22.90357208251953, 'timing/data': 0.01776816602796316, 'timing/model': 1.205341724678874, 'learning_rate/qwen_vl_interface': 5.007877223294668e-07, 'learning_rate/action_model': 5.0078772232946675e-06, 'epoch': 2.86}) +[2026-06-18 03:21:47,070][starVLA.training.train_starvla][INFO] - Step 6961, Loss: {'train/loss': 0.4133867621421814, 'train/grad_norm_pre_clip': 27.806283950805664, 'timing/data': 0.017097266390919685, 'timing/model': 1.1275564031675458, 'learning_rate/qwen_vl_interface': 5.007488295614058e-07, 'learning_rate/action_model': 5.007488295614058e-06, 'epoch': 2.87}) +[2026-06-18 03:21:50,729][starVLA.training.train_starvla][INFO] - Step 6962, Loss: {'train/loss': 0.0945790708065033, 'train/grad_norm_pre_clip': 12.017046928405762, 'timing/data': 0.018223939463496208, 'timing/model': 0.970029610209167, 'learning_rate/qwen_vl_interface': 5.007109213182231e-07, 'learning_rate/action_model': 5.007109213182231e-06, 'epoch': 2.87}) +[2026-06-18 03:21:54,676][starVLA.training.train_starvla][INFO] - Step 6963, Loss: {'train/loss': 0.3145025372505188, 'train/grad_norm_pre_clip': 28.392152786254883, 'timing/data': 0.02237587422132492, 'timing/model': 0.9276718879118562, 'learning_rate/qwen_vl_interface': 5.006739976077769e-07, 'learning_rate/action_model': 5.0067399760777695e-06, 'epoch': 2.87}) +[2026-06-18 03:21:58,549][starVLA.training.train_starvla][INFO] - Step 6964, Loss: {'train/loss': 0.10395947098731995, 'train/grad_norm_pre_clip': 12.06390380859375, 'timing/data': 0.01870215218514204, 'timing/model': 0.9394109137356281, 'learning_rate/qwen_vl_interface': 5.006380584377213e-07, 'learning_rate/action_model': 5.006380584377213e-06, 'epoch': 2.87}) +[2026-06-18 03:22:02,361][starVLA.training.train_starvla][INFO] - Step 6965, Loss: {'train/loss': 0.5057916045188904, 'train/grad_norm_pre_clip': 33.78729248046875, 'timing/data': 0.02070347871631384, 'timing/model': 0.9460076186805964, 'learning_rate/qwen_vl_interface': 5.006031038155068e-07, 'learning_rate/action_model': 5.006031038155067e-06, 'epoch': 2.87}) +[2026-06-18 03:22:06,518][starVLA.training.train_starvla][INFO] - Step 6966, Loss: {'train/loss': 0.0463775098323822, 'train/grad_norm_pre_clip': 18.26039695739746, 'timing/data': 0.020475550554692745, 'timing/model': 1.0109687680378556, 'learning_rate/qwen_vl_interface': 5.00569133748379e-07, 'learning_rate/action_model': 5.00569133748379e-06, 'epoch': 2.87}) +[2026-06-18 03:22:10,414][starVLA.training.train_starvla][INFO] - Step 6967, Loss: {'train/loss': 0.2591184675693512, 'train/grad_norm_pre_clip': 34.81706619262695, 'timing/data': 0.019963608123362064, 'timing/model': 1.1198903862386942, 'learning_rate/qwen_vl_interface': 5.005361482433811e-07, 'learning_rate/action_model': 5.005361482433811e-06, 'epoch': 2.87}) +[2026-06-18 03:22:14,268][starVLA.training.train_starvla][INFO] - Step 6968, Loss: {'train/loss': 0.12917061150074005, 'train/grad_norm_pre_clip': 17.123865127563477, 'timing/data': 0.02188031654804945, 'timing/model': 1.1013617552816868, 'learning_rate/qwen_vl_interface': 5.005041473073501e-07, 'learning_rate/action_model': 5.005041473073501e-06, 'epoch': 2.87}) +[2026-06-18 03:22:18,002][starVLA.training.train_starvla][INFO] - Step 6969, Loss: {'train/loss': 0.38694867491722107, 'train/grad_norm_pre_clip': 31.196962356567383, 'timing/data': 0.024057041853666306, 'timing/model': 1.0487879319116473, 'learning_rate/qwen_vl_interface': 5.004731309469201e-07, 'learning_rate/action_model': 5.004731309469201e-06, 'epoch': 2.87}) +[2026-06-18 03:22:21,866][starVLA.training.train_starvla][INFO] - Step 6970, Loss: {'train/loss': 0.41822731494903564, 'train/grad_norm_pre_clip': 19.699827194213867, 'timing/data': 0.0204691207036376, 'timing/model': 0.9274167204275727, 'learning_rate/qwen_vl_interface': 5.004430991685201e-07, 'learning_rate/action_model': 5.0044309916852016e-06, 'epoch': 2.87}) +[2026-06-18 03:22:25,742][starVLA.training.train_starvla][INFO] - Step 6971, Loss: {'train/loss': 0.586479902267456, 'train/grad_norm_pre_clip': 32.242191314697266, 'timing/data': 0.025590266101062298, 'timing/model': 0.931037568487227, 'learning_rate/qwen_vl_interface': 5.004140519783773e-07, 'learning_rate/action_model': 5.0041405197837725e-06, 'epoch': 2.87}) +[2026-06-18 03:22:29,469][starVLA.training.train_starvla][INFO] - Step 6972, Loss: {'train/loss': 0.15889953076839447, 'train/grad_norm_pre_clip': 23.049110412597656, 'timing/data': 0.01920801494270563, 'timing/model': 0.9264266276732087, 'learning_rate/qwen_vl_interface': 5.003859893825117e-07, 'learning_rate/action_model': 5.003859893825117e-06, 'epoch': 2.87}) +[2026-06-18 03:22:33,317][starVLA.training.train_starvla][INFO] - Step 6973, Loss: {'train/loss': 0.11677203327417374, 'train/grad_norm_pre_clip': 17.364513397216797, 'timing/data': 0.017211144790053368, 'timing/model': 0.9430217267945409, 'learning_rate/qwen_vl_interface': 5.003589113867419e-07, 'learning_rate/action_model': 5.003589113867419e-06, 'epoch': 2.87}) +[2026-06-18 03:22:37,079][starVLA.training.train_starvla][INFO] - Step 6974, Loss: {'train/loss': 0.26500070095062256, 'train/grad_norm_pre_clip': 23.600446701049805, 'timing/data': 0.01751760859042406, 'timing/model': 1.0905478140339255, 'learning_rate/qwen_vl_interface': 5.003328179966799e-07, 'learning_rate/action_model': 5.003328179966798e-06, 'epoch': 2.87}) +[2026-06-18 03:22:40,670][starVLA.training.train_starvla][INFO] - Step 6975, Loss: {'train/loss': 0.2935308814048767, 'train/grad_norm_pre_clip': 29.521778106689453, 'timing/data': 0.020136632025241852, 'timing/model': 0.9278430975973606, 'learning_rate/qwen_vl_interface': 5.003077092177363e-07, 'learning_rate/action_model': 5.003077092177363e-06, 'epoch': 2.87}) +[2026-06-18 03:22:44,456][starVLA.training.train_starvla][INFO] - Step 6976, Loss: {'train/loss': 0.15565063059329987, 'train/grad_norm_pre_clip': 21.041507720947266, 'timing/data': 0.019039218313992023, 'timing/model': 0.9356031566858292, 'learning_rate/qwen_vl_interface': 5.002835850551146e-07, 'learning_rate/action_model': 5.002835850551146e-06, 'epoch': 2.87}) +[2026-06-18 03:22:48,228][starVLA.training.train_starvla][INFO] - Step 6977, Loss: {'train/loss': 0.15343093872070312, 'train/grad_norm_pre_clip': 20.705533981323242, 'timing/data': 0.021285721100866795, 'timing/model': 0.9339129887521267, 'learning_rate/qwen_vl_interface': 5.002604455138174e-07, 'learning_rate/action_model': 5.002604455138173e-06, 'epoch': 2.87}) +[2026-06-18 03:22:52,196][starVLA.training.train_starvla][INFO] - Step 6978, Loss: {'train/loss': 0.49280646443367004, 'train/grad_norm_pre_clip': 30.66311264038086, 'timing/data': 0.02082916535437107, 'timing/model': 0.9634197289124131, 'learning_rate/qwen_vl_interface': 5.002382905986407e-07, 'learning_rate/action_model': 5.002382905986407e-06, 'epoch': 2.87}) +[2026-06-18 03:22:56,100][starVLA.training.train_starvla][INFO] - Step 6979, Loss: {'train/loss': 0.11233419179916382, 'train/grad_norm_pre_clip': 13.640366554260254, 'timing/data': 0.017833275720477104, 'timing/model': 0.9493839368224144, 'learning_rate/qwen_vl_interface': 5.002171203141771e-07, 'learning_rate/action_model': 5.0021712031417704e-06, 'epoch': 2.87}) +[2026-06-18 03:22:59,963][starVLA.training.train_starvla][INFO] - Step 6980, Loss: {'train/loss': 0.038865137845277786, 'train/grad_norm_pre_clip': 9.396570205688477, 'timing/data': 0.01593388244509697, 'timing/model': 1.1324654081836343, 'learning_rate/qwen_vl_interface': 5.001969346648154e-07, 'learning_rate/action_model': 5.001969346648154e-06, 'epoch': 2.87}) +[2026-06-18 03:23:03,783][starVLA.training.train_starvla][INFO] - Step 6981, Loss: {'train/loss': 0.8236573338508606, 'train/grad_norm_pre_clip': 38.48140335083008, 'timing/data': 0.017677735537290573, 'timing/model': 1.0605843188241124, 'learning_rate/qwen_vl_interface': 5.0017773365474e-07, 'learning_rate/action_model': 5.001777336547399e-06, 'epoch': 2.87}) +[2026-06-18 03:23:07,426][starVLA.training.train_starvla][INFO] - Step 6982, Loss: {'train/loss': 0.0689147561788559, 'train/grad_norm_pre_clip': 12.53982162475586, 'timing/data': 0.019469138234853745, 'timing/model': 0.9366939114406705, 'learning_rate/qwen_vl_interface': 5.001595172879317e-07, 'learning_rate/action_model': 5.001595172879316e-06, 'epoch': 2.87}) +[2026-06-18 03:23:11,258][starVLA.training.train_starvla][INFO] - Step 6983, Loss: {'train/loss': 0.07619145512580872, 'train/grad_norm_pre_clip': 12.180106163024902, 'timing/data': 0.020271819084882736, 'timing/model': 0.9447022210806608, 'learning_rate/qwen_vl_interface': 5.001422855681669e-07, 'learning_rate/action_model': 5.001422855681669e-06, 'epoch': 2.87}) +[2026-06-18 03:23:15,175][starVLA.training.train_starvla][INFO] - Step 6984, Loss: {'train/loss': 0.3128972053527832, 'train/grad_norm_pre_clip': 26.653669357299805, 'timing/data': 0.023625277914106846, 'timing/model': 0.9416228840127587, 'learning_rate/qwen_vl_interface': 5.001260384990169e-07, 'learning_rate/action_model': 5.001260384990169e-06, 'epoch': 2.87}) +[2026-06-18 03:23:19,078][starVLA.training.train_starvla][INFO] - Step 6985, Loss: {'train/loss': 0.15900735557079315, 'train/grad_norm_pre_clip': 17.082242965698242, 'timing/data': 0.016808154992759228, 'timing/model': 0.9699576981365681, 'learning_rate/qwen_vl_interface': 5.001107760838504e-07, 'learning_rate/action_model': 5.001107760838504e-06, 'epoch': 2.88}) +[2026-06-18 03:23:22,891][starVLA.training.train_starvla][INFO] - Step 6986, Loss: {'train/loss': 0.20579665899276733, 'train/grad_norm_pre_clip': 19.23370361328125, 'timing/data': 0.022067134268581867, 'timing/model': 0.934716172516346, 'learning_rate/qwen_vl_interface': 5.000964983258306e-07, 'learning_rate/action_model': 5.000964983258306e-06, 'epoch': 2.88}) +[2026-06-18 03:23:26,722][starVLA.training.train_starvla][INFO] - Step 6987, Loss: {'train/loss': 0.11210796982049942, 'train/grad_norm_pre_clip': 31.77440643310547, 'timing/data': 0.01741624064743519, 'timing/model': 1.1134494580328465, 'learning_rate/qwen_vl_interface': 5.000832052279185e-07, 'learning_rate/action_model': 5.000832052279185e-06, 'epoch': 2.88}) +[2026-06-18 03:23:30,513][starVLA.training.train_starvla][INFO] - Step 6988, Loss: {'train/loss': 0.06067021191120148, 'train/grad_norm_pre_clip': 12.678610801696777, 'timing/data': 0.018277877010405064, 'timing/model': 1.098602233454585, 'learning_rate/qwen_vl_interface': 5.000708967928686e-07, 'learning_rate/action_model': 5.000708967928685e-06, 'epoch': 2.88}) +[2026-06-18 03:23:34,171][starVLA.training.train_starvla][INFO] - Step 6989, Loss: {'train/loss': 0.5368696451187134, 'train/grad_norm_pre_clip': 31.292875289916992, 'timing/data': 0.019369469955563545, 'timing/model': 0.9323428301140666, 'learning_rate/qwen_vl_interface': 5.00059573023233e-07, 'learning_rate/action_model': 5.0005957302323305e-06, 'epoch': 2.88}) +[2026-06-18 03:23:38,039][starVLA.training.train_starvla][INFO] - Step 6990, Loss: {'train/loss': 0.08708013594150543, 'train/grad_norm_pre_clip': 21.17766571044922, 'timing/data': 0.01929300930351019, 'timing/model': 0.952251747250557, 'learning_rate/qwen_vl_interface': 5.000492339213594e-07, 'learning_rate/action_model': 5.000492339213594e-06, 'epoch': 2.88}) +[2026-06-18 03:23:42,170][starVLA.training.train_starvla][INFO] - Step 6991, Loss: {'train/loss': 0.09413211047649384, 'train/grad_norm_pre_clip': 12.681808471679688, 'timing/data': 0.017046768218278885, 'timing/model': 0.9402496926486492, 'learning_rate/qwen_vl_interface': 5.000398794893907e-07, 'learning_rate/action_model': 5.000398794893906e-06, 'epoch': 2.88}) +[2026-06-18 03:23:46,163][starVLA.training.train_starvla][INFO] - Step 6992, Loss: {'train/loss': 0.10705568641424179, 'train/grad_norm_pre_clip': 16.763748168945312, 'timing/data': 0.020485898479819298, 'timing/model': 0.9344122102484107, 'learning_rate/qwen_vl_interface': 5.000315097292659e-07, 'learning_rate/action_model': 5.000315097292659e-06, 'epoch': 2.88}) +[2026-06-18 03:23:50,305][starVLA.training.train_starvla][INFO] - Step 6993, Loss: {'train/loss': 0.32913434505462646, 'train/grad_norm_pre_clip': 25.151620864868164, 'timing/data': 0.02210639789700508, 'timing/model': 1.0268750311806798, 'learning_rate/qwen_vl_interface': 5.000241246427207e-07, 'learning_rate/action_model': 5.000241246427207e-06, 'epoch': 2.88}) +[2026-06-18 03:23:54,133][starVLA.training.train_starvla][INFO] - Step 6994, Loss: {'train/loss': 0.056779488921165466, 'train/grad_norm_pre_clip': 8.470596313476562, 'timing/data': 0.014580763876438141, 'timing/model': 1.1261769607663155, 'learning_rate/qwen_vl_interface': 5.000177242312855e-07, 'learning_rate/action_model': 5.000177242312855e-06, 'epoch': 2.88}) +[2026-06-18 03:23:58,009][starVLA.training.train_starvla][INFO] - Step 6995, Loss: {'train/loss': 0.3647416830062866, 'train/grad_norm_pre_clip': 26.07598114013672, 'timing/data': 0.026109409518539906, 'timing/model': 1.04976349696517, 'learning_rate/qwen_vl_interface': 5.00012308496287e-07, 'learning_rate/action_model': 5.000123084962871e-06, 'epoch': 2.88}) +[2026-06-18 03:24:01,678][starVLA.training.train_starvla][INFO] - Step 6996, Loss: {'train/loss': 0.12994197010993958, 'train/grad_norm_pre_clip': 16.82327651977539, 'timing/data': 0.016023543663322926, 'timing/model': 0.934438219293952, 'learning_rate/qwen_vl_interface': 5.000078774388487e-07, 'learning_rate/action_model': 5.000078774388487e-06, 'epoch': 2.88}) +[2026-06-18 03:24:05,471][starVLA.training.train_starvla][INFO] - Step 6997, Loss: {'train/loss': 0.3962334990501404, 'train/grad_norm_pre_clip': 36.25251770019531, 'timing/data': 0.017978379502892494, 'timing/model': 0.9368715677410364, 'learning_rate/qwen_vl_interface': 5.000044310598879e-07, 'learning_rate/action_model': 5.00004431059888e-06, 'epoch': 2.88}) +[2026-06-18 03:24:09,374][starVLA.training.train_starvla][INFO] - Step 6998, Loss: {'train/loss': 0.11165750026702881, 'train/grad_norm_pre_clip': 18.191089630126953, 'timing/data': 0.018600533716380596, 'timing/model': 0.9337194748222828, 'learning_rate/qwen_vl_interface': 5.000019693601204e-07, 'learning_rate/action_model': 5.000019693601204e-06, 'epoch': 2.88}) +[2026-06-18 03:24:13,259][starVLA.training.train_starvla][INFO] - Step 6999, Loss: {'train/loss': 0.16202081739902496, 'train/grad_norm_pre_clip': 12.423044204711914, 'timing/data': 0.019662877544760704, 'timing/model': 0.9338086377829313, 'learning_rate/qwen_vl_interface': 5.000004923400556e-07, 'learning_rate/action_model': 5.0000049234005565e-06, 'epoch': 2.88}) +[2026-06-18 03:26:23,091][starVLA.training.train_starvla][INFO] - Step 7000, Loss: {'train/loss': 0.09867047518491745, 'train/grad_norm_pre_clip': 19.724416732788086, 'eval/loss': 0.4238983392715454, 'eval/latency_6/loss': 0.4238983392715454, 'eval/demon_fire_precision': 0.9583862770012707, 'eval/demon_fire_recall': 0.9751131221719457, 'eval/demon_fire_f1': 0.9666773470041653, 'eval/demon_fire_rate_model': 0.8759042849193099, 'eval/demon_fire_rate_teacher': 0.860879243183083, 'eval/demon_left_precision': 0.8934426229508197, 'eval/demon_left_recall': 0.7649122807017544, 'eval/demon_left_f1': 0.8241965973534972, 'eval/demon_left_rate_model': 0.13578185865331108, 'eval/demon_left_rate_teacher': 0.15859766277128548, 'eval/demon_right_precision': 0.820919175911252, 'eval/demon_right_recall': 0.7956989247311828, 'eval/demon_right_f1': 0.8081123244929797, 'eval/demon_right_rate_model': 0.17557039510294936, 'eval/demon_right_rate_teacher': 0.18113522537562604, 'eval/demon_move_f1': 0.8161544609232384, 'eval/demon_cc_f1': 0.9159963471166157, 'eval/latency_6/demon_fire_precision': 0.9583862770012707, 'eval/latency_6/demon_fire_recall': 0.9751131221719457, 'eval/latency_6/demon_fire_f1': 0.9666773470041653, 'eval/latency_6/demon_fire_rate_model': 0.8759042849193099, 'eval/latency_6/demon_fire_rate_teacher': 0.860879243183083, 'eval/latency_6/demon_left_precision': 0.8934426229508197, 'eval/latency_6/demon_left_recall': 0.7649122807017544, 'eval/latency_6/demon_left_f1': 0.8241965973534972, 'eval/latency_6/demon_left_rate_model': 0.13578185865331108, 'eval/latency_6/demon_left_rate_teacher': 0.15859766277128548, 'eval/latency_6/demon_right_precision': 0.820919175911252, 'eval/latency_6/demon_right_recall': 0.7956989247311828, 'eval/latency_6/demon_right_f1': 0.8081123244929797, 'eval/latency_6/demon_right_rate_model': 0.17557039510294936, 'eval/latency_6/demon_right_rate_teacher': 0.18113522537562604, 'eval/latency_6/demon_move_f1': 0.8161544609232384, 'eval/latency_6/demon_cc_f1': 0.9159963471166157, 'timing/data': 0.016417388804256916, 'timing/model': 2.2512913728132844, 'learning_rate/qwen_vl_interface': 5e-07, 'learning_rate/action_model': 4.9999999999999996e-06, 'epoch': 2.88}) +[2026-06-18 03:26:23,095][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_7000_state +[2026-06-18 03:26:23,100][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 03:26:29,561][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_7000_state/model.safetensors +[2026-06-18 03:26:44,452][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_7000_state/optimizer.bin +[2026-06-18 03:26:44,452][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_7000_state/sampler.bin +[2026-06-18 03:26:44,452][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_7000_state/sampler_1.bin +[2026-06-18 03:26:44,455][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints/steps_7000_state/random_states_0.pkl +[2026-06-18 03:26:44,458][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 7000 +[2026-06-18 03:26:44,458][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 03:26:44,469][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 03:26:44,469][starVLA.training.train_starvla][INFO] - Training complete. Final training state is saved under /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_6_200ep/checkpoints +[2026-06-18 03:26:44,717][starVLA.training.train_starvla][INFO] - ... and that's all, folks! diff --git a/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_6_200ep/summary.jsonl b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_6_200ep/summary.jsonl new file mode 100644 index 0000000000000000000000000000000000000000..67f4020ac3871f475553f86f0c342379cb905502 --- /dev/null +++ b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_6_200ep/summary.jsonl @@ -0,0 +1,70 @@ +{"steps": 100} +{"steps": 200} +{"steps": 300} +{"steps": 400} +{"steps": 500} +{"steps": 600} +{"steps": 700} +{"steps": 800} +{"steps": 900} +{"steps": 1000} +{"steps": 1100} +{"steps": 1200} +{"steps": 1300} +{"steps": 1400} +{"steps": 1500} +{"steps": 1600} +{"steps": 1700} +{"steps": 1800} +{"steps": 1900} +{"steps": 2000} +{"steps": 2100} +{"steps": 2200} +{"steps": 2300} +{"steps": 2400} +{"steps": 2500} +{"steps": 2600} +{"steps": 2700} +{"steps": 2800} +{"steps": 2900} +{"steps": 3000} +{"steps": 3100} +{"steps": 3200} +{"steps": 3300} +{"steps": 3400} +{"steps": 3500} +{"steps": 3600} +{"steps": 3700} +{"steps": 3800} +{"steps": 3900} +{"steps": 4000} +{"steps": 4100} +{"steps": 4200} +{"steps": 4300} +{"steps": 4400} +{"steps": 4500} +{"steps": 4600} +{"steps": 4700} +{"steps": 4800} +{"steps": 4900} +{"steps": 5000} +{"steps": 5100} +{"steps": 5200} +{"steps": 5300} +{"steps": 5400} +{"steps": 5500} +{"steps": 5600} +{"steps": 5700} +{"steps": 5800} +{"steps": 5900} +{"steps": 6000} +{"steps": 6100} +{"steps": 6200} +{"steps": 6300} +{"steps": 6400} +{"steps": 6500} +{"steps": 6600} +{"steps": 6700} +{"steps": 6800} +{"steps": 6900} +{"steps": 7000} diff --git a/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_8_200ep/checkpoints/steps_7000_model.safetensors b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_8_200ep/checkpoints/steps_7000_model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..f85c968b09ce5080223da723c2668d232b0bce1d --- /dev/null +++ b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_8_200ep/checkpoints/steps_7000_model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:0d52d791412c41f17f607087c988d99267ef09ac097022b3fc3d97ebbb5b0b00 +size 9138230516 diff --git a/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_8_200ep/config.full.yaml b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_8_200ep/config.full.yaml new file mode 100644 index 0000000000000000000000000000000000000000..db226a325888835e917320b3e55f308fe0c9e11d --- /dev/null +++ b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_8_200ep/config.full.yaml @@ -0,0 +1,209 @@ +framework: + name: QwenOFT + qwenvl: + base_vlm: /workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + attn_implementation: flash_attention_2 + enable_gradient_checkpointing: true + action_model: + action_model_type: MLP + action_dim: 7 + action_hidden_dim: 2560 + future_action_window_size: 0 + past_action_window_size: 0 + loss_type: discrete_ce + state_dim: 7 + action_horizon: 1 + action_env_dim: 6 +datasets: + vla_data: + dataset_py: lerobot_datasets + include_state: true + data_root_dir: /workspace/starVLA/data/demon_attack_fix_latency_8_200ep + data_mix: demon_attack_train__bridge + eval_data_mix: demon_attack_train__bridge__val + custom_mixtures_path: null + action_type: discrete + sequential_step_sampling: false + shuffle: true + action_balance: + enabled: false + strategy: balanced_epoch + action_key: action_id + target_flap_fraction: 0.3 + noop_id: 0 + flap_id: 1 + latency_curriculum: + enabled: false + strategy: exclusive + latencies: null + phase_steps: null + per_device_batch_size: 32 + load_all_data_for_training: true + obs_image_size: null + video_backend: torchvision_av +dataset: + source_hf: '' + config_name: null + source_subdir: null + converted_name: demon_attack_train + single_source_hf: '' + mixed_source_hf: '' + single_converted_name: demon_attack_train + mixed_converted_name: demon_attack_mixed_latency_train + single_latency_filter: null + mixed_latency_filter: null + force_download: false + setup_force: false + skip_verification: false + verify_rows: 200 + max_episodes: null + episodes_per_latency: null + latency_filter: null + debug_subset: + enabled: false + max_episodes: 5 + suffix: debug +base_model: + repo_id: Qwen/Qwen3-VL-4B-Instruct +initialization: + checkpoint_local_dir: playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_hf_repo_id: StarVLA/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_filename: checkpoints/steps_5000_pytorch_model.pt +trainer: + max_train_steps: 7000 + num_warmup_steps: 100 + save_interval: 100 + eval_interval: 100 + eval_num_batches: 100 + per_latency_eval_num_batches: null + eval_action_classification: true + cc_f1_tolerance: 1 + learning_rate: + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + action_model: 0.0001 + lr_scheduler_type: cosine_with_min_lr + scheduler_specific_kwargs: + min_lr: 1.0e-06 + freeze_modules: '' + loss_scale: + vla: 1.0 + vlm: 0.1 + max_grad_norm: 1.0 + weight_decay: 0.0 + logging_frequency: 1 + gradient_clipping: 1.0 + gradient_accumulation_steps: 4 + distributed_backend: none + is_resume: false + pretrained_checkpoint: null + resume_step: 0 + reload_modules: null + optimizer: + name: AdamW + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + weight_decay: 1.0e-08 + fused: false + save_format: pt +workspace_dir: WORKSPACE_DIR +run_root_dir: /workspace/starVLA/results/Checkpoints +seed: 42 +wandb_entity: dongqianyu99-zhejiang-university +wandb_project: starVLA_rl_games +auth: + env_file: null + hf_token_env: HF_TOKEN + wandb_api_key_env: WANDB_API_KEY +paths: + run_root_dir: results/Checkpoints + dataset_local_dir: playground/Datasets/rl_games + dataset_cache_dir: null + base_model_dir: playground/Pretrained_models/Qwen3-VL-4B-Instruct + accelerate_config: starVLA/config/deepseeds/deepspeed_zero2.yaml +launch: + use_accelerate: true + gpus: null + num_processes: 1 + dry_run: false +conda: + enabled: true + env_name: null +rl_games: + model_alias: openvla + env_eval: + image_size: 224 + frameskip: 4 + seed: 42 + fixed_episode_seeds: true + latency_seed_stride: 0 + task_seed_stride: 0 + task_description: 'You are playing Demon Attack from a single game image. Choose + exactly one action from: NOOP, FIRE, RIGHT, LEFT, RIGHTFIRE, LEFTFIRE.' + demon_attack: + noop_max: 30 + enabled: true + distributed_mode: none + vectorized: + enabled: false + batch_size: 1 + latency: + prompt_map_path: /workspace/starVLA/data/demon_attack_fix_latency_8_200ep/demon_attack_train__bridge/latency_prompt_map.json + mode: single + values: + - 0 + mid_train: + enabled: false + interval_steps: 100 + latencies: + - 0 + num_episodes: 5 + max_steps_per_episode: 3600 + post_train: + enabled: false + latencies: + - 0 + - 1 + - 2 + - 3 + - 4 + num_episodes: 5 + max_steps_per_episode: 3600 + task: demon_attack + initialization_mode: bridge + action_carrier: bridge +model: openvla +env: demon_attack +init: bridge +bridge_base_model: + repo_id: + openvla: Qwen/Qwen3-VL-4B-Instruct + pi0: StarVLA/Qwen2.5-VL-3B-Instruct-Action + pi05: Qwen/Qwen3-VL-4B-Instruct + gr00t: Qwen/Qwen3-VL-4B-Instruct + local_dir: + openvla: playground/Pretrained_models/Qwen3-VL-4B-Instruct + pi0: playground/Pretrained_models/Qwen2.5-VL-3B-Instruct-Action + pi05: playground/Pretrained_models/Qwen3-VL-4B-Instruct + gr00t: playground/Pretrained_models/Qwen3-VL-4B-Instruct +mode: single +checkpoint: + load: auto + hf_repo_id: null + save_best_model: false + save_pt_file: false + local: + keep_last_n: 1 + sync: + enabled: false + repo_id: null + keep_last_n: 0 + sync_every_n_checkpoints: 1 + resume_policy: local_latest +run_id: demon_attack_fix_latency_8_200ep +output_dir: /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_8_200ep +config_yaml: null +is_debug: false +version_id: '0.21' diff --git a/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_8_200ep/config.yaml b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_8_200ep/config.yaml new file mode 100644 index 0000000000000000000000000000000000000000..9ac7ffa30a5a72745509d14e7364adbf5f747bca --- /dev/null +++ b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_8_200ep/config.yaml @@ -0,0 +1,75 @@ +checkpoint: + local: + keep_last_n: 1 + save_best_model: false + save_pt_file: false + sync: + enabled: false + keep_last_n: 0 + repo_id: null +datasets: + vla_data: + data_mix: demon_attack_train__bridge + dataset_py: lerobot_datasets + eval_data_mix: demon_attack_train__bridge__val + latency_curriculum: + enabled: false + obs_image_size: null + per_device_batch_size: 32 +framework: + action_model: + action_dim: 7 + action_env_dim: 6 + action_hidden_dim: 2560 + action_horizon: 1 + action_model_type: MLP + loss_type: discrete_ce + name: QwenOFT + qwenvl: + attn_implementation: flash_attention_2 + base_vlm: /workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + enable_gradient_checkpointing: true +output_dir: /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_8_200ep +rl_games: + env_eval: + enabled: true + mid_train: + enabled: false + interval_steps: 100 + task: demon_attack +run_id: demon_attack_fix_latency_8_200ep +run_root_dir: /workspace/starVLA/results/Checkpoints +seed: 42 +trainer: + cc_f1_tolerance: 1 + distributed_backend: none + eval_action_classification: true + eval_interval: 100 + eval_num_batches: 100 + freeze_modules: '' + gradient_accumulation_steps: 4 + gradient_clipping: 1.0 + is_resume: false + learning_rate: + action_model: 0.0001 + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + logging_frequency: 1 + lr_scheduler_type: cosine_with_min_lr + max_train_steps: 7000 + num_warmup_steps: 100 + optimizer: + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + fused: false + weight_decay: 1.0e-08 + per_latency_eval_num_batches: null + pretrained_checkpoint: null + reload_modules: null + save_interval: 100 + scheduler_specific_kwargs: + min_lr: 1.0e-06 +wandb_entity: dongqianyu99-zhejiang-university +wandb_project: starVLA_rl_games diff --git a/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_8_200ep/dataset_statistics.json b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_8_200ep/dataset_statistics.json new file mode 100644 index 0000000000000000000000000000000000000000..7d744d1929781dd3eca6bf4c1d10cd64b6109e9d --- /dev/null +++ b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_8_200ep/dataset_statistics.json @@ -0,0 +1,127 @@ +{ + "new_embodiment": { + "action": { + "mean": [ + 0.08639267086982727, + 0.627116858959198, + 0.006486095488071442, + 0.005628371611237526, + 0.148070827126503, + 0.12630519270896912, + 0.0 + ], + "std": [ + 0.28125613927841187, + 0.48417720198631287, + 0.08015890419483185, + 0.07480393350124359, + 0.35502511262893677, + 0.33193621039390564, + 0.0 + ], + "max": [ + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 1.0, + 1.0, + 0.0, + 0.0, + 1.0, + 1.0, + 0.0 + ], + "mask": [ + true, + true, + true, + true, + true, + true, + true + ] + }, + "state": { + "mean": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ] + }, + "num_transitions": 282142, + "num_trajectories": 180 + } +} \ No newline at end of file diff --git a/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_8_200ep/dataset_statistics_eval.json b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_8_200ep/dataset_statistics_eval.json new file mode 100644 index 0000000000000000000000000000000000000000..1e9026e03bfcd7b5bb756a8a120f0c3e4f4b0433 --- /dev/null +++ b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_8_200ep/dataset_statistics_eval.json @@ -0,0 +1,127 @@ +{ + "new_embodiment": { + "action": { + "mean": [ + 0.09224364906549454, + 0.6070592999458313, + 0.007906598970293999, + 0.0057144262827932835, + 0.1609153002500534, + 0.12616074085235596, + 0.0 + ], + "std": [ + 0.2894008159637451, + 0.48834389448165894, + 0.08855940401554108, + 0.07536990940570831, + 0.36744311451911926, + 0.33201834559440613, + 0.0 + ], + "max": [ + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 1.0, + 1.0, + 0.0, + 0.0, + 1.0, + 1.0, + 0.0 + ], + "mask": [ + true, + true, + true, + true, + true, + true, + true + ] + }, + "state": { + "mean": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ] + }, + "num_transitions": 40599, + "num_trajectories": 20 + } +} \ No newline at end of file diff --git a/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_8_200ep/hydra/.hydra/config.yaml b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_8_200ep/hydra/.hydra/config.yaml new file mode 100644 index 0000000000000000000000000000000000000000..8203389d3b98816f114d4161dcfe00717f2528b3 --- /dev/null +++ b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_8_200ep/hydra/.hydra/config.yaml @@ -0,0 +1,207 @@ +framework: + qwenvl: + base_vlm: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/playground/Pretrained_models/Qwen3-VL-4B-Instruct + attn_implementation: flash_attention_2 + enable_gradient_checkpointing: true + action_model: + state_dim: 7 + loss_type: discrete_ce + action_horizon: 1 + future_action_window_size: 0 + past_action_window_size: 0 + action_dim: 7 + action_env_dim: 6 + name: QwenOFT +datasets: + vla_data: + dataset_py: lerobot_datasets + include_state: true + data_root_dir: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/data/demon_attack_fix_latency_8_200ep + data_mix: demon_attack_train__bridge + eval_data_mix: demon_attack_train__bridge__val + custom_mixtures_path: null + action_type: discrete + sequential_step_sampling: false + shuffle: true + action_balance: + enabled: false + strategy: balanced_epoch + action_key: action_id + target_flap_fraction: 0.3 + noop_id: 0 + flap_id: 1 + latency_curriculum: + enabled: false + strategy: exclusive + latencies: null + phase_steps: null + per_device_batch_size: 32 + load_all_data_for_training: true + obs_image_size: null + video_backend: torchvision_av +dataset: + source_hf: ${dataset.single_source_hf} + config_name: null + source_subdir: null + converted_name: ${dataset.single_converted_name} + single_source_hf: '' + mixed_source_hf: '' + single_converted_name: demon_attack_train + mixed_converted_name: demon_attack_mixed_latency_train + single_latency_filter: null + mixed_latency_filter: null + force_download: false + setup_force: false + skip_verification: false + verify_rows: 200 + max_episodes: null + episodes_per_latency: null + latency_filter: ${dataset.single_latency_filter} + debug_subset: + enabled: false + max_episodes: 5 + suffix: debug +base_model: + repo_id: ${bridge_base_model.repo_id.${model}} +initialization: + checkpoint_local_dir: playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_hf_repo_id: StarVLA/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_filename: checkpoints/steps_5000_pytorch_model.pt +trainer: + max_train_steps: 7000 + num_warmup_steps: 100 + save_interval: 100 + eval_interval: 100 + eval_num_batches: 100 + per_latency_eval_num_batches: null + eval_action_classification: true + cc_f1_tolerance: 1 + learning_rate: + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + action_model: 0.0001 + lr_scheduler_type: cosine_with_min_lr + scheduler_specific_kwargs: + min_lr: 1.0e-06 + freeze_modules: '' + loss_scale: + vla: 1.0 + vlm: 0.1 + max_grad_norm: 1.0 + weight_decay: 0.0 + logging_frequency: 1 + gradient_clipping: 1.0 + gradient_accumulation_steps: 4 + distributed_backend: none + is_resume: false + pretrained_checkpoint: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt + resume_step: 0 + reload_modules: null + optimizer: + name: AdamW + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + weight_decay: 1.0e-08 + fused: false + save_format: pt +workspace_dir: WORKSPACE_DIR +run_root_dir: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints +seed: 42 +wandb_entity: dongqianyu99-zhejiang-university +wandb_project: starVLA_rl_games +auth: + env_file: null + hf_token_env: HF_TOKEN + wandb_api_key_env: WANDB_API_KEY +paths: + run_root_dir: results/Checkpoints + dataset_local_dir: playground/Datasets/rl_games + dataset_cache_dir: null + base_model_dir: ${bridge_base_model.local_dir.${model}} + accelerate_config: starVLA/config/deepseeds/deepspeed_zero2.yaml +launch: + use_accelerate: true + gpus: null + num_processes: 1 + dry_run: false +conda: + enabled: true + env_name: null +rl_games: + model_alias: openvla + env_eval: + image_size: 224 + frameskip: 4 + seed: 42 + fixed_episode_seeds: true + latency_seed_stride: 0 + task_seed_stride: 0 + task_description: 'You are playing Demon Attack from a single game image. Choose + exactly one action from: NOOP, FIRE, RIGHT, LEFT, RIGHTFIRE, LEFTFIRE.' + demon_attack: + noop_max: 30 + enabled: true + distributed_mode: none + vectorized: + enabled: false + batch_size: 1 + latency: + prompt_map_path: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/data/demon_attack_fix_latency_8_200ep/demon_attack_train__bridge/latency_prompt_map.json + mode: single + values: + - 0 + mid_train: + enabled: false + interval_steps: 100 + latencies: + - 0 + num_episodes: 5 + max_steps_per_episode: 3600 + post_train: + enabled: false + latencies: + - 0 + - 1 + - 2 + - 3 + - 4 + num_episodes: 5 + max_steps_per_episode: 3600 + task: demon_attack + initialization_mode: bridge + action_carrier: bridge +model: openvla +env: demon_attack +init: bridge +bridge_base_model: + repo_id: + openvla: Qwen/Qwen3-VL-4B-Instruct + pi0: StarVLA/Qwen2.5-VL-3B-Instruct-Action + pi05: Qwen/Qwen3-VL-4B-Instruct + gr00t: Qwen/Qwen3-VL-4B-Instruct + local_dir: + openvla: playground/Pretrained_models/Qwen3-VL-4B-Instruct + pi0: playground/Pretrained_models/Qwen2.5-VL-3B-Instruct-Action + pi05: playground/Pretrained_models/Qwen3-VL-4B-Instruct + gr00t: playground/Pretrained_models/Qwen3-VL-4B-Instruct +mode: single +checkpoint: + load: auto + hf_repo_id: null + save_best_model: false + save_pt_file: false + local: + keep_last_n: 1 + sync: + enabled: false + repo_id: null + keep_last_n: 0 + sync_every_n_checkpoints: 1 + resume_policy: local_latest +run_id: demon_attack_fix_latency_8_200ep +output_dir: null +config_yaml: null +is_debug: false +version_id: '0.21' diff --git a/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_8_200ep/hydra/.hydra/hydra.yaml b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_8_200ep/hydra/.hydra/hydra.yaml new file mode 100644 index 0000000000000000000000000000000000000000..35c52f317fdf623270fcb6075d204dda5f0f9994 --- /dev/null +++ b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_8_200ep/hydra/.hydra/hydra.yaml @@ -0,0 +1,259 @@ +hydra: + run: + dir: ${run_root_dir}/${run_id}/hydra + sweep: + dir: multirun/${now:%Y-%m-%d}/${now:%H-%M-%S} + subdir: ${hydra.job.num} + launcher: + _target_: hydra._internal.core_plugins.basic_launcher.BasicLauncher + sweeper: + _target_: hydra._internal.core_plugins.basic_sweeper.BasicSweeper + max_batch_size: null + params: null + help: + app_name: ${hydra.job.name} + header: '${hydra.help.app_name} is powered by Hydra. + + ' + footer: 'Powered by Hydra (https://hydra.cc) + + Use --hydra-help to view Hydra specific help + + ' + template: '${hydra.help.header} + + == Configuration groups == + + Compose your configuration from those groups (group=option) + + + $APP_CONFIG_GROUPS + + + == Config == + + Override anything in the config (foo.bar=value) + + + $CONFIG + + + ${hydra.help.footer} + + ' + hydra_help: + template: 'Hydra (${hydra.runtime.version}) + + See https://hydra.cc for more info. + + + == Flags == + + $FLAGS_HELP + + + == Configuration groups == + + Compose your configuration from those groups (For example, append hydra/job_logging=disabled + to command line) + + + $HYDRA_CONFIG_GROUPS + + + Use ''--cfg hydra'' to Show the Hydra config. + + ' + hydra_help: ??? + hydra_logging: + version: 1 + formatters: + simple: + format: '[%(asctime)s][HYDRA] %(message)s' + handlers: + console: + class: logging.StreamHandler + formatter: simple + stream: ext://sys.stdout + root: + level: INFO + handlers: + - console + loggers: + logging_example: + level: DEBUG + disable_existing_loggers: false + job_logging: + version: 1 + formatters: + simple: + format: '[%(asctime)s][%(name)s][%(levelname)s] - %(message)s' + handlers: + console: + class: logging.StreamHandler + formatter: simple + stream: ext://sys.stdout + file: + class: logging.FileHandler + formatter: simple + filename: ${hydra.runtime.output_dir}/${hydra.job.name}.log + root: + level: INFO + handlers: + - console + - file + disable_existing_loggers: false + env: {} + mode: RUN + searchpath: [] + callbacks: {} + output_subdir: .hydra + overrides: + hydra: + - hydra.mode=RUN + task: + - model=openvla + - env=demon_attack + - init=bridge + - mode=single + - run_id=demon_attack_fix_latency_8_200ep + - run_root_dir=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints + - seed=42 + - wandb_entity=dongqianyu99-zhejiang-university + - wandb_project=starVLA_rl_games + - rl_games.env_eval.enabled=true + - checkpoint.sync.enabled=false + - checkpoint.sync.keep_last_n=0 + - checkpoint.local.keep_last_n=1 + - checkpoint.save_best_model=false + - checkpoint.save_pt_file=false + - trainer.is_resume=false + - trainer.pretrained_checkpoint=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt + - trainer.resume_step=0 + - trainer.max_train_steps=7000 + - trainer.num_warmup_steps=100 + - trainer.save_interval=100 + - trainer.eval_interval=100 + - trainer.eval_num_batches=100 + - trainer.logging_frequency=1 + - trainer.gradient_accumulation_steps=4 + - trainer.distributed_backend=none + - trainer.learning_rate.base=2e-05 + - trainer.learning_rate.qwen_vl_interface=1e-05 + - trainer.learning_rate.action_model=0.0001 + - trainer.lr_scheduler_type=cosine_with_min_lr + - trainer.scheduler_specific_kwargs.min_lr=1e-06 + - trainer.loss_scale.vla=1.0 + - trainer.loss_scale.vlm=0.1 + - trainer.max_grad_norm=1.0 + - trainer.weight_decay=0.0 + - trainer.gradient_clipping=1.0 + - trainer.optimizer.name=AdamW + - trainer.optimizer.betas=[0.9,0.95] + - trainer.optimizer.eps=1e-08 + - trainer.optimizer.weight_decay=1e-08 + - trainer.optimizer.fused=false + - trainer.save_format=pt + - framework.name=QwenOFT + - framework.qwenvl.attn_implementation=flash_attention_2 + - framework.qwenvl.enable_gradient_checkpointing=true + - framework.action_model.action_dim=7 + - framework.action_model.action_env_dim=6 + - framework.action_model.state_dim=7 + - framework.action_model.loss_type=discrete_ce + - framework.action_model.action_horizon=1 + - framework.action_model.future_action_window_size=0 + - framework.action_model.past_action_window_size=0 + - datasets.vla_data.include_state=true + - datasets.vla_data.action_type=discrete + - datasets.vla_data.sequential_step_sampling=false + - datasets.vla_data.shuffle=true + - datasets.vla_data.action_balance.enabled=false + - datasets.vla_data.action_balance.strategy=balanced_epoch + - datasets.vla_data.action_balance.action_key=action_id + - datasets.vla_data.action_balance.target_flap_fraction=0.3 + - datasets.vla_data.action_balance.noop_id=0 + - datasets.vla_data.action_balance.flap_id=1 + - datasets.vla_data.latency_curriculum.enabled=false + - datasets.vla_data.latency_curriculum.strategy=exclusive + - datasets.vla_data.per_device_batch_size=32 + - datasets.vla_data.load_all_data_for_training=true + - datasets.vla_data.video_backend=torchvision_av + - datasets.vla_data.data_root_dir=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/data/demon_attack_fix_latency_8_200ep + - datasets.vla_data.data_mix=demon_attack_train__bridge + - datasets.vla_data.eval_data_mix=demon_attack_train__bridge__val + - framework.qwenvl.base_vlm=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/playground/Pretrained_models/Qwen3-VL-4B-Instruct + - rl_games.task=demon_attack + - rl_games.model_alias=openvla + - rl_games.initialization_mode=bridge + - rl_games.action_carrier=bridge + - rl_games.env_eval.distributed_mode=none + - rl_games.env_eval.latency.mode=single + - rl_games.env_eval.frameskip=4 + - rl_games.env_eval.image_size=224 + - rl_games.env_eval.seed=42 + - rl_games.env_eval.fixed_episode_seeds=true + - rl_games.env_eval.latency_seed_stride=0 + - rl_games.env_eval.task_seed_stride=0 + - 'rl_games.env_eval.task_description=''You are playing Demon Attack from a single + game image. Choose exactly one action from: NOOP, FIRE, RIGHT, LEFT, RIGHTFIRE, + LEFTFIRE.''' + - rl_games.env_eval.latency.values=[0] + - rl_games.env_eval.latency.prompt_map_path=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/data/demon_attack_fix_latency_8_200ep/demon_attack_train__bridge/latency_prompt_map.json + - rl_games.env_eval.mid_train.enabled=false + - rl_games.env_eval.mid_train.interval_steps=100 + - rl_games.env_eval.mid_train.num_episodes=5 + - rl_games.env_eval.mid_train.max_steps_per_episode=3600 + - rl_games.env_eval.mid_train.latencies=[0] + - rl_games.env_eval.post_train.enabled=false + - rl_games.env_eval.post_train.num_episodes=5 + - rl_games.env_eval.post_train.max_steps_per_episode=3600 + - rl_games.env_eval.post_train.latencies=[0,1,2,3,4] + job: + name: train_starvla_hydra + chdir: false + override_dirname: 'checkpoint.local.keep_last_n=1,checkpoint.save_best_model=false,checkpoint.save_pt_file=false,checkpoint.sync.enabled=false,checkpoint.sync.keep_last_n=0,datasets.vla_data.action_balance.action_key=action_id,datasets.vla_data.action_balance.enabled=false,datasets.vla_data.action_balance.flap_id=1,datasets.vla_data.action_balance.noop_id=0,datasets.vla_data.action_balance.strategy=balanced_epoch,datasets.vla_data.action_balance.target_flap_fraction=0.3,datasets.vla_data.action_type=discrete,datasets.vla_data.data_mix=demon_attack_train__bridge,datasets.vla_data.data_root_dir=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/data/demon_attack_fix_latency_8_200ep,datasets.vla_data.eval_data_mix=demon_attack_train__bridge__val,datasets.vla_data.include_state=true,datasets.vla_data.latency_curriculum.enabled=false,datasets.vla_data.latency_curriculum.strategy=exclusive,datasets.vla_data.load_all_data_for_training=true,datasets.vla_data.per_device_batch_size=32,datasets.vla_data.sequential_step_sampling=false,datasets.vla_data.shuffle=true,datasets.vla_data.video_backend=torchvision_av,env=demon_attack,framework.action_model.action_dim=7,framework.action_model.action_env_dim=6,framework.action_model.action_horizon=1,framework.action_model.future_action_window_size=0,framework.action_model.loss_type=discrete_ce,framework.action_model.past_action_window_size=0,framework.action_model.state_dim=7,framework.name=QwenOFT,framework.qwenvl.attn_implementation=flash_attention_2,framework.qwenvl.base_vlm=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/playground/Pretrained_models/Qwen3-VL-4B-Instruct,framework.qwenvl.enable_gradient_checkpointing=true,init=bridge,mode=single,model=openvla,rl_games.action_carrier=bridge,rl_games.env_eval.distributed_mode=none,rl_games.env_eval.enabled=true,rl_games.env_eval.fixed_episode_seeds=true,rl_games.env_eval.frameskip=4,rl_games.env_eval.image_size=224,rl_games.env_eval.latency.mode=single,rl_games.env_eval.latency.prompt_map_path=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/data/demon_attack_fix_latency_8_200ep/demon_attack_train__bridge/latency_prompt_map.json,rl_games.env_eval.latency.values=[0],rl_games.env_eval.latency_seed_stride=0,rl_games.env_eval.mid_train.enabled=false,rl_games.env_eval.mid_train.interval_steps=100,rl_games.env_eval.mid_train.latencies=[0],rl_games.env_eval.mid_train.max_steps_per_episode=3600,rl_games.env_eval.mid_train.num_episodes=5,rl_games.env_eval.post_train.enabled=false,rl_games.env_eval.post_train.latencies=[0,1,2,3,4],rl_games.env_eval.post_train.max_steps_per_episode=3600,rl_games.env_eval.post_train.num_episodes=5,rl_games.env_eval.seed=42,rl_games.env_eval.task_description=''You + are playing Demon Attack from a single game image. Choose exactly one action + from: NOOP, FIRE, RIGHT, LEFT, RIGHTFIRE, LEFTFIRE.'',rl_games.env_eval.task_seed_stride=0,rl_games.initialization_mode=bridge,rl_games.model_alias=openvla,rl_games.task=demon_attack,run_id=demon_attack_fix_latency_8_200ep,run_root_dir=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints,seed=42,trainer.distributed_backend=none,trainer.eval_interval=100,trainer.eval_num_batches=100,trainer.gradient_accumulation_steps=4,trainer.gradient_clipping=1.0,trainer.is_resume=false,trainer.learning_rate.action_model=0.0001,trainer.learning_rate.base=2e-05,trainer.learning_rate.qwen_vl_interface=1e-05,trainer.logging_frequency=1,trainer.loss_scale.vla=1.0,trainer.loss_scale.vlm=0.1,trainer.lr_scheduler_type=cosine_with_min_lr,trainer.max_grad_norm=1.0,trainer.max_train_steps=7000,trainer.num_warmup_steps=100,trainer.optimizer.betas=[0.9,0.95],trainer.optimizer.eps=1e-08,trainer.optimizer.fused=false,trainer.optimizer.name=AdamW,trainer.optimizer.weight_decay=1e-08,trainer.pretrained_checkpoint=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt,trainer.resume_step=0,trainer.save_format=pt,trainer.save_interval=100,trainer.scheduler_specific_kwargs.min_lr=1e-06,trainer.weight_decay=0.0,wandb_entity=dongqianyu99-zhejiang-university,wandb_project=starVLA_rl_games' + id: ??? + num: ??? + config_name: train + env_set: {} + env_copy: [] + config: + override_dirname: + kv_sep: '=' + item_sep: ',' + exclude_keys: [] + runtime: + version: 1.3.2 + version_base: '1.1' + cwd: /inspire/hdd/project/spatialintelligence/public/lzj/starVLA + config_sources: + - path: hydra.conf + schema: pkg + provider: hydra + - path: /inspire/hdd/project/spatialintelligence/public/lzj/starVLA/examples/rl_games/config + schema: file + provider: main + - path: '' + schema: structured + provider: schema + output_dir: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/hydra + choices: + cross_task_setup: null + checkpoint: default + mode: single + init: bridge + env: demon_attack + model: openvla + hydra/env: default + hydra/callbacks: null + hydra/job_logging: default + hydra/hydra_logging: default + hydra/hydra_help: default + hydra/help: default + hydra/sweeper: basic + hydra/launcher: basic + hydra/output: default + verbose: false diff --git a/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_8_200ep/hydra/.hydra/overrides.yaml b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_8_200ep/hydra/.hydra/overrides.yaml new file mode 100644 index 0000000000000000000000000000000000000000..c5c4df76b8595efa7602e333cc11c64c638cf6e7 --- /dev/null +++ b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_8_200ep/hydra/.hydra/overrides.yaml @@ -0,0 +1,97 @@ +- model=openvla +- env=demon_attack +- init=bridge +- mode=single +- run_id=demon_attack_fix_latency_8_200ep +- run_root_dir=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints +- seed=42 +- wandb_entity=dongqianyu99-zhejiang-university +- wandb_project=starVLA_rl_games +- rl_games.env_eval.enabled=true +- checkpoint.sync.enabled=false +- checkpoint.sync.keep_last_n=0 +- checkpoint.local.keep_last_n=1 +- checkpoint.save_best_model=false +- checkpoint.save_pt_file=false +- trainer.is_resume=false +- trainer.pretrained_checkpoint=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt +- trainer.resume_step=0 +- trainer.max_train_steps=7000 +- trainer.num_warmup_steps=100 +- trainer.save_interval=100 +- trainer.eval_interval=100 +- trainer.eval_num_batches=100 +- trainer.logging_frequency=1 +- trainer.gradient_accumulation_steps=4 +- trainer.distributed_backend=none +- trainer.learning_rate.base=2e-05 +- trainer.learning_rate.qwen_vl_interface=1e-05 +- trainer.learning_rate.action_model=0.0001 +- trainer.lr_scheduler_type=cosine_with_min_lr +- trainer.scheduler_specific_kwargs.min_lr=1e-06 +- trainer.loss_scale.vla=1.0 +- trainer.loss_scale.vlm=0.1 +- trainer.max_grad_norm=1.0 +- trainer.weight_decay=0.0 +- trainer.gradient_clipping=1.0 +- trainer.optimizer.name=AdamW +- trainer.optimizer.betas=[0.9,0.95] +- trainer.optimizer.eps=1e-08 +- trainer.optimizer.weight_decay=1e-08 +- trainer.optimizer.fused=false +- trainer.save_format=pt +- framework.name=QwenOFT +- framework.qwenvl.attn_implementation=flash_attention_2 +- framework.qwenvl.enable_gradient_checkpointing=true +- framework.action_model.action_dim=7 +- framework.action_model.action_env_dim=6 +- framework.action_model.state_dim=7 +- framework.action_model.loss_type=discrete_ce +- framework.action_model.action_horizon=1 +- framework.action_model.future_action_window_size=0 +- framework.action_model.past_action_window_size=0 +- datasets.vla_data.include_state=true +- datasets.vla_data.action_type=discrete +- datasets.vla_data.sequential_step_sampling=false +- datasets.vla_data.shuffle=true +- datasets.vla_data.action_balance.enabled=false +- datasets.vla_data.action_balance.strategy=balanced_epoch +- datasets.vla_data.action_balance.action_key=action_id +- datasets.vla_data.action_balance.target_flap_fraction=0.3 +- datasets.vla_data.action_balance.noop_id=0 +- datasets.vla_data.action_balance.flap_id=1 +- datasets.vla_data.latency_curriculum.enabled=false +- datasets.vla_data.latency_curriculum.strategy=exclusive +- datasets.vla_data.per_device_batch_size=32 +- datasets.vla_data.load_all_data_for_training=true +- datasets.vla_data.video_backend=torchvision_av +- datasets.vla_data.data_root_dir=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/data/demon_attack_fix_latency_8_200ep +- datasets.vla_data.data_mix=demon_attack_train__bridge +- datasets.vla_data.eval_data_mix=demon_attack_train__bridge__val +- framework.qwenvl.base_vlm=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/playground/Pretrained_models/Qwen3-VL-4B-Instruct +- rl_games.task=demon_attack +- rl_games.model_alias=openvla +- rl_games.initialization_mode=bridge +- rl_games.action_carrier=bridge +- rl_games.env_eval.distributed_mode=none +- rl_games.env_eval.latency.mode=single +- rl_games.env_eval.frameskip=4 +- rl_games.env_eval.image_size=224 +- rl_games.env_eval.seed=42 +- rl_games.env_eval.fixed_episode_seeds=true +- rl_games.env_eval.latency_seed_stride=0 +- rl_games.env_eval.task_seed_stride=0 +- 'rl_games.env_eval.task_description=''You are playing Demon Attack from a single + game image. Choose exactly one action from: NOOP, FIRE, RIGHT, LEFT, RIGHTFIRE, + LEFTFIRE.''' +- rl_games.env_eval.latency.values=[0] +- rl_games.env_eval.latency.prompt_map_path=/inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/data/demon_attack_fix_latency_8_200ep/demon_attack_train__bridge/latency_prompt_map.json +- rl_games.env_eval.mid_train.enabled=false +- rl_games.env_eval.mid_train.interval_steps=100 +- rl_games.env_eval.mid_train.num_episodes=5 +- rl_games.env_eval.mid_train.max_steps_per_episode=3600 +- rl_games.env_eval.mid_train.latencies=[0] +- rl_games.env_eval.post_train.enabled=false +- rl_games.env_eval.post_train.num_episodes=5 +- rl_games.env_eval.post_train.max_steps_per_episode=3600 +- rl_games.env_eval.post_train.latencies=[0,1,2,3,4] diff --git a/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_8_200ep/hydra/train_starvla_hydra.log b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_8_200ep/hydra/train_starvla_hydra.log new file mode 100644 index 0000000000000000000000000000000000000000..6a275b077cfba2563528d038426c8ffe67d63eb3 --- /dev/null +++ b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_8_200ep/hydra/train_starvla_hydra.log @@ -0,0 +1,7752 @@ +[2026-06-17 16:47:00,378][starVLA.training.train_starvla][INFO] - VLA Training :: Warming Up +[2026-06-17 16:47:00,390][starVLA.training.train_starvla][INFO] - ✅ Configuration wrapped for access tracking +[2026-06-17 16:49:02,837][starVLA.training.train_starvla][INFO] - Preloaded model checkpoint before Accelerator/DeepSpeed initialization: /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt +[2026-06-17 16:49:02,980][starVLA.training.train_starvla][INFO] - Creating VLA Dataset with Mixture `demon_attack_train__bridge` +[2026-06-17 16:49:13,830][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Behavior (data_config): ['BEHAVIOR_challenge'] +[2026-06-17 16:49:13,830][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from DOMINO: ['robotwin'] +[2026-06-17 16:49:13,830][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from DOMINO (data_config): [] +[2026-06-17 16:49:13,830][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from DOMINO (data_config): ['domino', 'domino_clean', 'domino_random', 'domino_cotrain'] +[2026-06-17 16:49:13,830][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Franka: ['custom_robot_config', 'demo_sim_franka_delta_joints', 'SO101'] +[2026-06-17 16:49:13,830][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Franka (data_config): [] +[2026-06-17 16:49:13,830][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Franka (data_config): ['custom_dataset', 'custom_dataset_2', 'demo_sim_pick_place', 'SO101_pick'] +[2026-06-17 16:49:13,830][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from LIBERO: ['libero_franka'] +[2026-06-17 16:49:13,831][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from LIBERO (data_config): [] +[2026-06-17 16:49:13,831][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from LIBERO (data_config): ['libero_all', 'libero_goal', 'multi_robot'] +[2026-06-17 16:49:13,831][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from RoboChallenge_table30v2: ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-17 16:49:13,831][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from RoboChallenge_table30v2 (data_config): ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-17 16:49:13,831][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from RoboChallenge_table30v2 (data_config): ['robochallenge_table30v2_shred_paper', 'robochallenge_table30v2_ur5_all', 'robochallenge_table30v2_arx5_all', 'robochallenge_table30v2_dosw1_all'] +[2026-06-17 16:49:13,831][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_365: ['panda_omron_robocasa365'] +[2026-06-17 16:49:13,831][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_365 (data_config): [] +[2026-06-17 16:49:13,831][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_365 (data_config): ['robocasa365_open_drawer_target_human', 'robocasa365_atomic_target_human_all', 'robocasa365_composite_target_human_all', 'robocasa365_target_human_all'] +[2026-06-17 16:49:13,831][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_tabletop: ['fourier_gr1_arms_waist'] +[2026-06-17 16:49:13,831][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_tabletop (data_config): [] +[2026-06-17 16:49:13,831][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_tabletop (data_config): ['fourier_gr1_unified_1000'] +[2026-06-17 16:49:13,834][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robotwin: ['robotwin', 'robotwin50', 'arx_x5'] +[2026-06-17 16:49:13,834][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robotwin (data_config): [] +[2026-06-17 16:49:13,834][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robotwin (data_config): ['robotwin_all', 'robotwin_all_50', 'robotwin', 'robotwin_task1', 'robotwin_task2', 'arx_x5'] +[2026-06-17 16:49:13,834][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from SimplerEnv: ['oxe_droid', 'oxe_bridge', 'oxe_rt1'] +[2026-06-17 16:49:13,834][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from SimplerEnv (data_config): [] +[2026-06-17 16:49:13,834][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from SimplerEnv (data_config): ['bridge', 'bridge_rt_1'] +[2026-06-17 16:49:13,834][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from VLA-Arena: ['vla_arena_franka'] +[2026-06-17 16:49:13,835][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from VLA-Arena (data_config): [] +[2026-06-17 16:49:13,835][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from VLA-Arena (data_config): ['vla_arena_L0_S', 'vla_arena_L0_M', 'vla_arena_L0_L'] +[2026-06-17 16:49:13,835][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from rl_games: ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-17 16:49:13,835][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from rl_games (data_config): ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-17 16:49:13,835][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from rl_games (data_config): ['flappy_train', 'flappy_train__bridge', 'flappy_mixed_latency_train', 'flappy_mixed_latency_train__bridge', 'demon_attack_train', 'demon_attack_train__bridge', 'demon_attack_mixed_latency_train', 'demon_attack_mixed_latency_train__bridge', 'deadly_corridor_train', 'deadly_corridor_train__bridge', 'deadly_corridor_mixed_latency_train', 'deadly_corridor_mixed_latency_train__bridge'] +[2026-06-17 16:49:13,980][starVLA.training.train_starvla][INFO] - Creating VLA Eval Dataset with Mixture `demon_attack_train__bridge__val` +[2026-06-17 16:49:14,225][starVLA.training.train_starvla][INFO] - 📝 Full config saved at /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/config.full.yaml +[2026-06-17 16:49:14,234][starVLA.training.train_starvla][INFO] - 📊 Accessed config snapshot saved at /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/config.yaml +[2026-06-17 16:49:14,234][starVLA.training.train_starvla][INFO] - No pretrained checkpoint provided. Starting training from scratch. +[2026-06-17 16:49:26,270][starVLA.training.train_starvla][INFO] - ***** Training Configuration ***** +[2026-06-17 16:49:26,271][starVLA.training.train_starvla][INFO] - Total optimization steps = 7000 +[2026-06-17 16:49:26,271][starVLA.training.train_starvla][INFO] - Per device batch size = 32 +[2026-06-17 16:49:26,271][starVLA.training.train_starvla][INFO] - Gradient accumulation steps = 4 +[2026-06-17 16:49:26,271][starVLA.training.train_starvla][INFO] - Total batch size = 128 +[2026-06-17 16:50:33,602][starVLA.training.train_starvla][INFO] - Step 1, Loss: {'train/loss': 1.7706451416015625, 'train/grad_norm_pre_clip': 10.501188278198242, 'timing/data': 0.029601264744997025, 'timing/model': 0.8526543499901891, 'learning_rate/qwen_vl_interface': 1.0000000000000001e-07, 'learning_rate/action_model': 1.0000000000000002e-06, 'epoch': 0.0}) +[2026-06-17 16:50:36,489][starVLA.training.train_starvla][INFO] - Step 2, Loss: {'train/loss': 1.7709919214248657, 'train/grad_norm_pre_clip': 10.471224784851074, 'timing/data': 0.029320566914975643, 'timing/model': 0.7740110885351896, 'learning_rate/qwen_vl_interface': 2.0000000000000002e-07, 'learning_rate/action_model': 2.0000000000000003e-06, 'epoch': 0.0}) +[2026-06-17 16:50:39,386][starVLA.training.train_starvla][INFO] - Step 3, Loss: {'train/loss': 1.748056173324585, 'train/grad_norm_pre_clip': 10.00935173034668, 'timing/data': 0.03240472264587879, 'timing/model': 0.7173677999526262, 'learning_rate/qwen_vl_interface': 3.0000000000000004e-07, 'learning_rate/action_model': 3e-06, 'epoch': 0.0}) +[2026-06-17 16:50:42,297][starVLA.training.train_starvla][INFO] - Step 4, Loss: {'train/loss': 1.7392796277999878, 'train/grad_norm_pre_clip': 9.715324401855469, 'timing/data': 0.031238388270139694, 'timing/model': 0.7194917211309075, 'learning_rate/qwen_vl_interface': 4.0000000000000003e-07, 'learning_rate/action_model': 4.000000000000001e-06, 'epoch': 0.0}) +[2026-06-17 16:50:45,200][starVLA.training.train_starvla][INFO] - Step 5, Loss: {'train/loss': 1.7169194221496582, 'train/grad_norm_pre_clip': 9.26742172241211, 'timing/data': 0.028499457985162735, 'timing/model': 0.8611328592523932, 'learning_rate/qwen_vl_interface': 5.000000000000001e-07, 'learning_rate/action_model': 5e-06, 'epoch': 0.0}) +[2026-06-17 16:50:48,272][starVLA.training.train_starvla][INFO] - Step 6, Loss: {'train/loss': 1.6680948734283447, 'train/grad_norm_pre_clip': 8.94197940826416, 'timing/data': 0.03296315670013428, 'timing/model': 0.7255999492481351, 'learning_rate/qwen_vl_interface': 6.000000000000001e-07, 'learning_rate/action_model': 6e-06, 'epoch': 0.0}) +[2026-06-17 16:50:51,266][starVLA.training.train_starvla][INFO] - Step 7, Loss: {'train/loss': 1.6626064777374268, 'train/grad_norm_pre_clip': 6.27971887588501, 'timing/data': 0.025117154233157635, 'timing/model': 0.7566402535885572, 'learning_rate/qwen_vl_interface': 7.000000000000001e-07, 'learning_rate/action_model': 7.000000000000001e-06, 'epoch': 0.0}) +[2026-06-17 16:50:54,131][starVLA.training.train_starvla][INFO] - Step 8, Loss: {'train/loss': 1.6289708614349365, 'train/grad_norm_pre_clip': 5.531403064727783, 'timing/data': 0.024970143102109432, 'timing/model': 0.7261716611683369, 'learning_rate/qwen_vl_interface': 8.000000000000001e-07, 'learning_rate/action_model': 8.000000000000001e-06, 'epoch': 0.0}) +[2026-06-17 16:50:57,797][starVLA.training.train_starvla][INFO] - Step 9, Loss: {'train/loss': 1.4975364208221436, 'train/grad_norm_pre_clip': 8.122950553894043, 'timing/data': 0.036932346411049366, 'timing/model': 0.7347576888278127, 'learning_rate/qwen_vl_interface': 9.000000000000001e-07, 'learning_rate/action_model': 9e-06, 'epoch': 0.0}) +[2026-06-17 16:51:00,626][starVLA.training.train_starvla][INFO] - Step 10, Loss: {'train/loss': 1.4734073877334595, 'train/grad_norm_pre_clip': 6.3642401695251465, 'timing/data': 0.03173237107694149, 'timing/model': 0.7575634494423866, 'learning_rate/qwen_vl_interface': 1.0000000000000002e-06, 'learning_rate/action_model': 1e-05, 'epoch': 0.0}) +[2026-06-17 16:51:03,395][starVLA.training.train_starvla][INFO] - Step 11, Loss: {'train/loss': 1.3404481410980225, 'train/grad_norm_pre_clip': 7.337996482849121, 'timing/data': 0.02794785425066948, 'timing/model': 0.7335999626666307, 'learning_rate/qwen_vl_interface': 1.1e-06, 'learning_rate/action_model': 1.1000000000000001e-05, 'epoch': 0.0}) +[2026-06-17 16:51:06,229][starVLA.training.train_starvla][INFO] - Step 12, Loss: {'train/loss': 1.2259531021118164, 'train/grad_norm_pre_clip': 7.595338344573975, 'timing/data': 0.024603286758065224, 'timing/model': 0.7647427758201957, 'learning_rate/qwen_vl_interface': 1.2000000000000002e-06, 'learning_rate/action_model': 1.2e-05, 'epoch': 0.01}) +[2026-06-17 16:51:09,067][starVLA.training.train_starvla][INFO] - Step 13, Loss: {'train/loss': 1.3598991632461548, 'train/grad_norm_pre_clip': 4.151146411895752, 'timing/data': 0.023873220197856426, 'timing/model': 0.7658372595906258, 'learning_rate/qwen_vl_interface': 1.3e-06, 'learning_rate/action_model': 1.3000000000000001e-05, 'epoch': 0.01}) +[2026-06-17 16:51:12,368][starVLA.training.train_starvla][INFO] - Step 14, Loss: {'train/loss': 1.066351294517517, 'train/grad_norm_pre_clip': 5.595298767089844, 'timing/data': 0.0272498382255435, 'timing/model': 0.7313494514673948, 'learning_rate/qwen_vl_interface': 1.4000000000000001e-06, 'learning_rate/action_model': 1.4000000000000001e-05, 'epoch': 0.01}) +[2026-06-17 16:51:15,392][starVLA.training.train_starvla][INFO] - Step 15, Loss: {'train/loss': 1.1346312761306763, 'train/grad_norm_pre_clip': 3.1100661754608154, 'timing/data': 0.025738482363522053, 'timing/model': 0.7807462010532618, 'learning_rate/qwen_vl_interface': 1.5e-06, 'learning_rate/action_model': 1.5e-05, 'epoch': 0.01}) +[2026-06-17 16:51:18,176][starVLA.training.train_starvla][INFO] - Step 16, Loss: {'train/loss': 1.4672237634658813, 'train/grad_norm_pre_clip': 3.9741551876068115, 'timing/data': 0.027263060212135315, 'timing/model': 0.7268552910536528, 'learning_rate/qwen_vl_interface': 1.6000000000000001e-06, 'learning_rate/action_model': 1.6000000000000003e-05, 'epoch': 0.01}) +[2026-06-17 16:51:21,108][starVLA.training.train_starvla][INFO] - Step 17, Loss: {'train/loss': 1.102254033088684, 'train/grad_norm_pre_clip': 2.123138666152954, 'timing/data': 0.031810810789465904, 'timing/model': 0.8142671445384622, 'learning_rate/qwen_vl_interface': 1.7000000000000002e-06, 'learning_rate/action_model': 1.7000000000000003e-05, 'epoch': 0.01}) +[2026-06-17 16:51:23,947][starVLA.training.train_starvla][INFO] - Step 18, Loss: {'train/loss': 1.2643918991088867, 'train/grad_norm_pre_clip': 3.8922712802886963, 'timing/data': 0.027239062823355198, 'timing/model': 0.7380293663591146, 'learning_rate/qwen_vl_interface': 1.8000000000000001e-06, 'learning_rate/action_model': 1.8e-05, 'epoch': 0.01}) +[2026-06-17 16:51:26,718][starVLA.training.train_starvla][INFO] - Step 19, Loss: {'train/loss': 1.2230453491210938, 'train/grad_norm_pre_clip': 3.8448901176452637, 'timing/data': 0.02625865489244461, 'timing/model': 0.7197519661858678, 'learning_rate/qwen_vl_interface': 1.9000000000000002e-06, 'learning_rate/action_model': 1.9e-05, 'epoch': 0.01}) +[2026-06-17 16:51:29,529][starVLA.training.train_starvla][INFO] - Step 20, Loss: {'train/loss': 1.1854044198989868, 'train/grad_norm_pre_clip': 3.785252094268799, 'timing/data': 0.021723823621869087, 'timing/model': 0.723364569246769, 'learning_rate/qwen_vl_interface': 2.0000000000000003e-06, 'learning_rate/action_model': 2e-05, 'epoch': 0.01}) +[2026-06-17 16:51:32,330][starVLA.training.train_starvla][INFO] - Step 21, Loss: {'train/loss': 1.1581473350524902, 'train/grad_norm_pre_clip': 2.5595133304595947, 'timing/data': 0.0255190534517169, 'timing/model': 0.7261861516162753, 'learning_rate/qwen_vl_interface': 2.1000000000000002e-06, 'learning_rate/action_model': 2.1e-05, 'epoch': 0.01}) +[2026-06-17 16:51:35,292][starVLA.training.train_starvla][INFO] - Step 22, Loss: {'train/loss': 1.2143038511276245, 'train/grad_norm_pre_clip': 2.0785787105560303, 'timing/data': 0.03300856798887253, 'timing/model': 0.8190599251538515, 'learning_rate/qwen_vl_interface': 2.2e-06, 'learning_rate/action_model': 2.2000000000000003e-05, 'epoch': 0.01}) +[2026-06-17 16:51:38,258][starVLA.training.train_starvla][INFO] - Step 23, Loss: {'train/loss': 1.410003662109375, 'train/grad_norm_pre_clip': 3.8435022830963135, 'timing/data': 0.029239867813885212, 'timing/model': 0.7318229433149099, 'learning_rate/qwen_vl_interface': 2.3000000000000004e-06, 'learning_rate/action_model': 2.3000000000000003e-05, 'epoch': 0.01}) +[2026-06-17 16:51:41,122][starVLA.training.train_starvla][INFO] - Step 24, Loss: {'train/loss': 1.1114853620529175, 'train/grad_norm_pre_clip': 2.4849441051483154, 'timing/data': 0.03045951295644045, 'timing/model': 0.747879022732377, 'learning_rate/qwen_vl_interface': 2.4000000000000003e-06, 'learning_rate/action_model': 2.4e-05, 'epoch': 0.01}) +[2026-06-17 16:51:43,906][starVLA.training.train_starvla][INFO] - Step 25, Loss: {'train/loss': 1.1250360012054443, 'train/grad_norm_pre_clip': 1.364992618560791, 'timing/data': 0.03268952202051878, 'timing/model': 0.7201054394245148, 'learning_rate/qwen_vl_interface': 2.5e-06, 'learning_rate/action_model': 2.5e-05, 'epoch': 0.01}) +[2026-06-17 16:51:47,034][starVLA.training.train_starvla][INFO] - Step 26, Loss: {'train/loss': 1.3115497827529907, 'train/grad_norm_pre_clip': 2.6402814388275146, 'timing/data': 0.035334113985300064, 'timing/model': 1.0236512133851647, 'learning_rate/qwen_vl_interface': 2.6e-06, 'learning_rate/action_model': 2.6000000000000002e-05, 'epoch': 0.01}) +[2026-06-17 16:51:50,079][starVLA.training.train_starvla][INFO] - Step 27, Loss: {'train/loss': 1.0026147365570068, 'train/grad_norm_pre_clip': 2.2175185680389404, 'timing/data': 0.031075784005224705, 'timing/model': 0.733429946936667, 'learning_rate/qwen_vl_interface': 2.7000000000000004e-06, 'learning_rate/action_model': 2.7000000000000002e-05, 'epoch': 0.01}) +[2026-06-17 16:51:54,432][starVLA.training.train_starvla][INFO] - Step 28, Loss: {'train/loss': 1.1294872760772705, 'train/grad_norm_pre_clip': 3.7951316833496094, 'timing/data': 0.024647177197039127, 'timing/model': 1.1539144683629274, 'learning_rate/qwen_vl_interface': 2.8000000000000003e-06, 'learning_rate/action_model': 2.8000000000000003e-05, 'epoch': 0.01}) +[2026-06-17 16:51:57,384][starVLA.training.train_starvla][INFO] - Step 29, Loss: {'train/loss': 1.2800651788711548, 'train/grad_norm_pre_clip': 2.3158786296844482, 'timing/data': 0.03814413212239742, 'timing/model': 0.8027452705428004, 'learning_rate/qwen_vl_interface': 2.9e-06, 'learning_rate/action_model': 2.9e-05, 'epoch': 0.01}) +[2026-06-17 16:52:00,207][starVLA.training.train_starvla][INFO] - Step 30, Loss: {'train/loss': 1.1163538694381714, 'train/grad_norm_pre_clip': 3.900789260864258, 'timing/data': 0.028111924417316914, 'timing/model': 0.7324039191007614, 'learning_rate/qwen_vl_interface': 3e-06, 'learning_rate/action_model': 3e-05, 'epoch': 0.01}) +[2026-06-17 16:52:03,034][starVLA.training.train_starvla][INFO] - Step 31, Loss: {'train/loss': 1.1346664428710938, 'train/grad_norm_pre_clip': 2.97534441947937, 'timing/data': 0.023693744093179703, 'timing/model': 0.72920913528651, 'learning_rate/qwen_vl_interface': 3.1000000000000004e-06, 'learning_rate/action_model': 3.1e-05, 'epoch': 0.01}) +[2026-06-17 16:52:05,816][starVLA.training.train_starvla][INFO] - Step 32, Loss: {'train/loss': 0.9602581858634949, 'train/grad_norm_pre_clip': 4.013017177581787, 'timing/data': 0.02431095950305462, 'timing/model': 0.7364874901250005, 'learning_rate/qwen_vl_interface': 3.2000000000000003e-06, 'learning_rate/action_model': 3.2000000000000005e-05, 'epoch': 0.01}) +[2026-06-17 16:52:08,619][starVLA.training.train_starvla][INFO] - Step 33, Loss: {'train/loss': 1.1013460159301758, 'train/grad_norm_pre_clip': 3.035651922225952, 'timing/data': 0.024873871356248856, 'timing/model': 0.7260614167898893, 'learning_rate/qwen_vl_interface': 3.3000000000000006e-06, 'learning_rate/action_model': 3.3e-05, 'epoch': 0.01}) +[2026-06-17 16:52:11,412][starVLA.training.train_starvla][INFO] - Step 34, Loss: {'train/loss': 1.119347095489502, 'train/grad_norm_pre_clip': 1.113670825958252, 'timing/data': 0.02542338240891695, 'timing/model': 0.7194147668778896, 'learning_rate/qwen_vl_interface': 3.4000000000000005e-06, 'learning_rate/action_model': 3.4000000000000007e-05, 'epoch': 0.02}) +[2026-06-17 16:52:14,394][starVLA.training.train_starvla][INFO] - Step 35, Loss: {'train/loss': 1.1125162839889526, 'train/grad_norm_pre_clip': 3.612804889678955, 'timing/data': 0.020651904866099358, 'timing/model': 0.7230930086225271, 'learning_rate/qwen_vl_interface': 3.5e-06, 'learning_rate/action_model': 3.5e-05, 'epoch': 0.02}) +[2026-06-17 16:52:17,482][starVLA.training.train_starvla][INFO] - Step 36, Loss: {'train/loss': 0.9159272313117981, 'train/grad_norm_pre_clip': 1.6176224946975708, 'timing/data': 0.031216845847666264, 'timing/model': 0.7366118719801307, 'learning_rate/qwen_vl_interface': 3.6000000000000003e-06, 'learning_rate/action_model': 3.6e-05, 'epoch': 0.02}) +[2026-06-17 16:52:20,429][starVLA.training.train_starvla][INFO] - Step 37, Loss: {'train/loss': 1.2731739282608032, 'train/grad_norm_pre_clip': 4.767294406890869, 'timing/data': 0.03198453690856695, 'timing/model': 0.7332100328058004, 'learning_rate/qwen_vl_interface': 3.7e-06, 'learning_rate/action_model': 3.7e-05, 'epoch': 0.02}) +[2026-06-17 16:52:23,207][starVLA.training.train_starvla][INFO] - Step 38, Loss: {'train/loss': 1.2809669971466064, 'train/grad_norm_pre_clip': 3.836834192276001, 'timing/data': 0.023180969059467316, 'timing/model': 0.7317688148468733, 'learning_rate/qwen_vl_interface': 3.8000000000000005e-06, 'learning_rate/action_model': 3.8e-05, 'epoch': 0.02}) +[2026-06-17 16:52:25,990][starVLA.training.train_starvla][INFO] - Step 39, Loss: {'train/loss': 1.1531766653060913, 'train/grad_norm_pre_clip': 1.8751051425933838, 'timing/data': 0.03375665284693241, 'timing/model': 0.7239629030227661, 'learning_rate/qwen_vl_interface': 3.900000000000001e-06, 'learning_rate/action_model': 3.9000000000000006e-05, 'epoch': 0.02}) +[2026-06-17 16:52:28,751][starVLA.training.train_starvla][INFO] - Step 40, Loss: {'train/loss': 1.2601617574691772, 'train/grad_norm_pre_clip': 4.09926700592041, 'timing/data': 0.025944167748093605, 'timing/model': 0.7235765056684613, 'learning_rate/qwen_vl_interface': 4.000000000000001e-06, 'learning_rate/action_model': 4e-05, 'epoch': 0.02}) +[2026-06-17 16:52:31,508][starVLA.training.train_starvla][INFO] - Step 41, Loss: {'train/loss': 1.5539882183074951, 'train/grad_norm_pre_clip': 3.888711452484131, 'timing/data': 0.030633635818958282, 'timing/model': 0.7153975171968341, 'learning_rate/qwen_vl_interface': 4.1e-06, 'learning_rate/action_model': 4.1e-05, 'epoch': 0.02}) +[2026-06-17 16:52:34,266][starVLA.training.train_starvla][INFO] - Step 42, Loss: {'train/loss': 1.3577853441238403, 'train/grad_norm_pre_clip': 3.5629093647003174, 'timing/data': 0.027670550160109997, 'timing/model': 0.7270123548805714, 'learning_rate/qwen_vl_interface': 4.2000000000000004e-06, 'learning_rate/action_model': 4.2e-05, 'epoch': 0.02}) +[2026-06-17 16:52:37,054][starVLA.training.train_starvla][INFO] - Step 43, Loss: {'train/loss': 1.2172343730926514, 'train/grad_norm_pre_clip': 2.6080570220947266, 'timing/data': 0.02236223593354225, 'timing/model': 0.7367955120280385, 'learning_rate/qwen_vl_interface': 4.3e-06, 'learning_rate/action_model': 4.3e-05, 'epoch': 0.02}) +[2026-06-17 16:52:39,837][starVLA.training.train_starvla][INFO] - Step 44, Loss: {'train/loss': 1.0432745218276978, 'train/grad_norm_pre_clip': 1.4872320890426636, 'timing/data': 0.02472889143973589, 'timing/model': 0.7250993698835373, 'learning_rate/qwen_vl_interface': 4.4e-06, 'learning_rate/action_model': 4.4000000000000006e-05, 'epoch': 0.02}) +[2026-06-17 16:52:42,683][starVLA.training.train_starvla][INFO] - Step 45, Loss: {'train/loss': 1.260026216506958, 'train/grad_norm_pre_clip': 2.725874662399292, 'timing/data': 0.030814820900559425, 'timing/model': 0.724619003944099, 'learning_rate/qwen_vl_interface': 4.5e-06, 'learning_rate/action_model': 4.5e-05, 'epoch': 0.02}) +[2026-06-17 16:52:46,243][starVLA.training.train_starvla][INFO] - Step 46, Loss: {'train/loss': 0.9931522011756897, 'train/grad_norm_pre_clip': 3.1129794120788574, 'timing/data': 0.026229390874505043, 'timing/model': 1.1322420565411448, 'learning_rate/qwen_vl_interface': 4.600000000000001e-06, 'learning_rate/action_model': 4.600000000000001e-05, 'epoch': 0.02}) +[2026-06-17 16:52:49,131][starVLA.training.train_starvla][INFO] - Step 47, Loss: {'train/loss': 1.3299049139022827, 'train/grad_norm_pre_clip': 2.2204933166503906, 'timing/data': 0.03238649107515812, 'timing/model': 0.7325131697580218, 'learning_rate/qwen_vl_interface': 4.7e-06, 'learning_rate/action_model': 4.7e-05, 'epoch': 0.02}) +[2026-06-17 16:52:51,943][starVLA.training.train_starvla][INFO] - Step 48, Loss: {'train/loss': 1.2486422061920166, 'train/grad_norm_pre_clip': 3.0399599075317383, 'timing/data': 0.03505549393594265, 'timing/model': 0.7271939860656857, 'learning_rate/qwen_vl_interface': 4.800000000000001e-06, 'learning_rate/action_model': 4.8e-05, 'epoch': 0.02}) +[2026-06-17 16:52:54,733][starVLA.training.train_starvla][INFO] - Step 49, Loss: {'train/loss': 1.0946167707443237, 'train/grad_norm_pre_clip': 3.52089262008667, 'timing/data': 0.026491126976907253, 'timing/model': 0.7172187296673656, 'learning_rate/qwen_vl_interface': 4.9000000000000005e-06, 'learning_rate/action_model': 4.9e-05, 'epoch': 0.02}) +[2026-06-17 16:52:57,514][starVLA.training.train_starvla][INFO] - Step 50, Loss: {'train/loss': 0.9755496382713318, 'train/grad_norm_pre_clip': 2.9668710231781006, 'timing/data': 0.027611764147877693, 'timing/model': 0.7231206893920898, 'learning_rate/qwen_vl_interface': 5e-06, 'learning_rate/action_model': 5e-05, 'epoch': 0.02}) +[2026-06-17 16:53:00,280][starVLA.training.train_starvla][INFO] - Step 51, Loss: {'train/loss': 1.154510259628296, 'train/grad_norm_pre_clip': 2.8211820125579834, 'timing/data': 0.042263331823050976, 'timing/model': 0.7215421386063099, 'learning_rate/qwen_vl_interface': 5.1e-06, 'learning_rate/action_model': 5.1000000000000006e-05, 'epoch': 0.02}) +[2026-06-17 16:53:03,049][starVLA.training.train_starvla][INFO] - Step 52, Loss: {'train/loss': 1.1366773843765259, 'train/grad_norm_pre_clip': 2.824643135070801, 'timing/data': 0.022583323530852795, 'timing/model': 0.7248558811843395, 'learning_rate/qwen_vl_interface': 5.2e-06, 'learning_rate/action_model': 5.2000000000000004e-05, 'epoch': 0.02}) +[2026-06-17 16:53:05,841][starVLA.training.train_starvla][INFO] - Step 53, Loss: {'train/loss': 1.2550700902938843, 'train/grad_norm_pre_clip': 2.882507085800171, 'timing/data': 0.024656918831169605, 'timing/model': 0.7398048313334584, 'learning_rate/qwen_vl_interface': 5.300000000000001e-06, 'learning_rate/action_model': 5.300000000000001e-05, 'epoch': 0.02}) +[2026-06-17 16:53:08,633][starVLA.training.train_starvla][INFO] - Step 54, Loss: {'train/loss': 0.946902871131897, 'train/grad_norm_pre_clip': 1.6131391525268555, 'timing/data': 0.025632305070757866, 'timing/model': 0.7277397885918617, 'learning_rate/qwen_vl_interface': 5.400000000000001e-06, 'learning_rate/action_model': 5.4000000000000005e-05, 'epoch': 0.02}) +[2026-06-17 16:53:11,550][starVLA.training.train_starvla][INFO] - Step 55, Loss: {'train/loss': 1.5570963621139526, 'train/grad_norm_pre_clip': 4.974882125854492, 'timing/data': 0.024148625321686268, 'timing/model': 0.7367219952866435, 'learning_rate/qwen_vl_interface': 5.500000000000001e-06, 'learning_rate/action_model': 5.500000000000001e-05, 'epoch': 0.02}) +[2026-06-17 16:53:14,587][starVLA.training.train_starvla][INFO] - Step 56, Loss: {'train/loss': 1.036300539970398, 'train/grad_norm_pre_clip': 1.4296882152557373, 'timing/data': 0.023023661226034164, 'timing/model': 0.9907616097480059, 'learning_rate/qwen_vl_interface': 5.600000000000001e-06, 'learning_rate/action_model': 5.6000000000000006e-05, 'epoch': 0.03}) +[2026-06-17 16:53:17,431][starVLA.training.train_starvla][INFO] - Step 57, Loss: {'train/loss': 1.061751127243042, 'train/grad_norm_pre_clip': 2.7719624042510986, 'timing/data': 0.025629770010709763, 'timing/model': 0.7192149208858609, 'learning_rate/qwen_vl_interface': 5.7e-06, 'learning_rate/action_model': 5.6999999999999996e-05, 'epoch': 0.03}) +[2026-06-17 16:53:20,883][starVLA.training.train_starvla][INFO] - Step 58, Loss: {'train/loss': 0.8654072880744934, 'train/grad_norm_pre_clip': 3.4700865745544434, 'timing/data': 0.025445151142776012, 'timing/model': 1.3728579515591264, 'learning_rate/qwen_vl_interface': 5.8e-06, 'learning_rate/action_model': 5.8e-05, 'epoch': 0.03}) +[2026-06-17 16:53:23,693][starVLA.training.train_starvla][INFO] - Step 59, Loss: {'train/loss': 0.8913629055023193, 'train/grad_norm_pre_clip': 3.5111582279205322, 'timing/data': 0.03066546656191349, 'timing/model': 0.7200293773785233, 'learning_rate/qwen_vl_interface': 5.9e-06, 'learning_rate/action_model': 5.9e-05, 'epoch': 0.03}) +[2026-06-17 16:53:26,460][starVLA.training.train_starvla][INFO] - Step 60, Loss: {'train/loss': 0.9385161995887756, 'train/grad_norm_pre_clip': 4.302988052368164, 'timing/data': 0.02098312508314848, 'timing/model': 0.7206988418474793, 'learning_rate/qwen_vl_interface': 6e-06, 'learning_rate/action_model': 6e-05, 'epoch': 0.03}) +[2026-06-17 16:53:29,437][starVLA.training.train_starvla][INFO] - Step 61, Loss: {'train/loss': 1.4248359203338623, 'train/grad_norm_pre_clip': 7.134669303894043, 'timing/data': 0.0235025892034173, 'timing/model': 0.7271013502031565, 'learning_rate/qwen_vl_interface': 6.1e-06, 'learning_rate/action_model': 6.1e-05, 'epoch': 0.03}) +[2026-06-17 16:53:32,238][starVLA.training.train_starvla][INFO] - Step 62, Loss: {'train/loss': 1.4362313747406006, 'train/grad_norm_pre_clip': 8.758091926574707, 'timing/data': 0.025834555737674236, 'timing/model': 0.7187005653977394, 'learning_rate/qwen_vl_interface': 6.200000000000001e-06, 'learning_rate/action_model': 6.2e-05, 'epoch': 0.03}) +[2026-06-17 16:53:35,011][starVLA.training.train_starvla][INFO] - Step 63, Loss: {'train/loss': 0.759178876876831, 'train/grad_norm_pre_clip': 2.9651505947113037, 'timing/data': 0.03164312057197094, 'timing/model': 0.7212927853688598, 'learning_rate/qwen_vl_interface': 6.300000000000001e-06, 'learning_rate/action_model': 6.3e-05, 'epoch': 0.03}) +[2026-06-17 16:53:37,769][starVLA.training.train_starvla][INFO] - Step 64, Loss: {'train/loss': 1.0406651496887207, 'train/grad_norm_pre_clip': 5.8394455909729, 'timing/data': 0.029463209211826324, 'timing/model': 0.7190063251182437, 'learning_rate/qwen_vl_interface': 6.4000000000000006e-06, 'learning_rate/action_model': 6.400000000000001e-05, 'epoch': 0.03}) +[2026-06-17 16:53:40,555][starVLA.training.train_starvla][INFO] - Step 65, Loss: {'train/loss': 1.1041650772094727, 'train/grad_norm_pre_clip': 3.1547772884368896, 'timing/data': 0.03706400562077761, 'timing/model': 0.7220612987875938, 'learning_rate/qwen_vl_interface': 6.5000000000000004e-06, 'learning_rate/action_model': 6.500000000000001e-05, 'epoch': 0.03}) +[2026-06-17 16:53:43,343][starVLA.training.train_starvla][INFO] - Step 66, Loss: {'train/loss': 1.161285400390625, 'train/grad_norm_pre_clip': 3.6275055408477783, 'timing/data': 0.03767707385122776, 'timing/model': 0.7248354451730847, 'learning_rate/qwen_vl_interface': 6.600000000000001e-06, 'learning_rate/action_model': 6.6e-05, 'epoch': 0.03}) +[2026-06-17 16:53:46,108][starVLA.training.train_starvla][INFO] - Step 67, Loss: {'train/loss': 1.3564006090164185, 'train/grad_norm_pre_clip': 6.024158000946045, 'timing/data': 0.030276711098849773, 'timing/model': 0.7176249325275421, 'learning_rate/qwen_vl_interface': 6.700000000000001e-06, 'learning_rate/action_model': 6.7e-05, 'epoch': 0.03}) +[2026-06-17 16:53:49,438][starVLA.training.train_starvla][INFO] - Step 68, Loss: {'train/loss': 1.4272520542144775, 'train/grad_norm_pre_clip': 6.599739074707031, 'timing/data': 0.02255597524344921, 'timing/model': 0.9983228277415037, 'learning_rate/qwen_vl_interface': 6.800000000000001e-06, 'learning_rate/action_model': 6.800000000000001e-05, 'epoch': 0.03}) +[2026-06-17 16:53:52,395][starVLA.training.train_starvla][INFO] - Step 69, Loss: {'train/loss': 1.2720152139663696, 'train/grad_norm_pre_clip': 14.420050621032715, 'timing/data': 0.029875027015805244, 'timing/model': 0.7197791123762727, 'learning_rate/qwen_vl_interface': 6.9e-06, 'learning_rate/action_model': 6.9e-05, 'epoch': 0.03}) +[2026-06-17 16:53:55,185][starVLA.training.train_starvla][INFO] - Step 70, Loss: {'train/loss': 1.1781619787216187, 'train/grad_norm_pre_clip': 12.254569053649902, 'timing/data': 0.03953599091619253, 'timing/model': 0.7235891642048955, 'learning_rate/qwen_vl_interface': 7e-06, 'learning_rate/action_model': 7e-05, 'epoch': 0.03}) +[2026-06-17 16:53:58,039][starVLA.training.train_starvla][INFO] - Step 71, Loss: {'train/loss': 1.1372320652008057, 'train/grad_norm_pre_clip': 9.834624290466309, 'timing/data': 0.03271935600787401, 'timing/model': 0.7185926847159863, 'learning_rate/qwen_vl_interface': 7.100000000000001e-06, 'learning_rate/action_model': 7.1e-05, 'epoch': 0.03}) +[2026-06-17 16:54:00,794][starVLA.training.train_starvla][INFO] - Step 72, Loss: {'train/loss': 1.0040960311889648, 'train/grad_norm_pre_clip': 3.326932907104492, 'timing/data': 0.02551164198666811, 'timing/model': 0.7227516267448664, 'learning_rate/qwen_vl_interface': 7.2000000000000005e-06, 'learning_rate/action_model': 7.2e-05, 'epoch': 0.03}) +[2026-06-17 16:54:03,550][starVLA.training.train_starvla][INFO] - Step 73, Loss: {'train/loss': 0.9446784257888794, 'train/grad_norm_pre_clip': 6.963712215423584, 'timing/data': 0.025332373566925526, 'timing/model': 0.7188880192115903, 'learning_rate/qwen_vl_interface': 7.3e-06, 'learning_rate/action_model': 7.3e-05, 'epoch': 0.03}) +[2026-06-17 16:54:06,295][starVLA.training.train_starvla][INFO] - Step 74, Loss: {'train/loss': 1.0574002265930176, 'train/grad_norm_pre_clip': 5.073915481567383, 'timing/data': 0.025778493843972683, 'timing/model': 0.7221459671854973, 'learning_rate/qwen_vl_interface': 7.4e-06, 'learning_rate/action_model': 7.4e-05, 'epoch': 0.03}) +[2026-06-17 16:54:09,053][starVLA.training.train_starvla][INFO] - Step 75, Loss: {'train/loss': 1.546154499053955, 'train/grad_norm_pre_clip': 8.014444351196289, 'timing/data': 0.025345558300614357, 'timing/model': 0.7163250092417002, 'learning_rate/qwen_vl_interface': 7.500000000000001e-06, 'learning_rate/action_model': 7.500000000000001e-05, 'epoch': 0.03}) +[2026-06-17 16:54:11,816][starVLA.training.train_starvla][INFO] - Step 76, Loss: {'train/loss': 1.186887264251709, 'train/grad_norm_pre_clip': 3.975128650665283, 'timing/data': 0.023500679060816765, 'timing/model': 0.7305270805954933, 'learning_rate/qwen_vl_interface': 7.600000000000001e-06, 'learning_rate/action_model': 7.6e-05, 'epoch': 0.03}) +[2026-06-17 16:54:14,580][starVLA.training.train_starvla][INFO] - Step 77, Loss: {'train/loss': 1.2954317331314087, 'train/grad_norm_pre_clip': 3.9389171600341797, 'timing/data': 0.023991214111447334, 'timing/model': 0.7258690353482962, 'learning_rate/qwen_vl_interface': 7.7e-06, 'learning_rate/action_model': 7.7e-05, 'epoch': 0.03}) +[2026-06-17 16:54:17,347][starVLA.training.train_starvla][INFO] - Step 78, Loss: {'train/loss': 0.9112586379051208, 'train/grad_norm_pre_clip': 3.709578275680542, 'timing/data': 0.025748204439878464, 'timing/model': 0.7221182566136122, 'learning_rate/qwen_vl_interface': 7.800000000000002e-06, 'learning_rate/action_model': 7.800000000000001e-05, 'epoch': 0.04}) +[2026-06-17 16:54:20,109][starVLA.training.train_starvla][INFO] - Step 79, Loss: {'train/loss': 1.1712567806243896, 'train/grad_norm_pre_clip': 4.100214958190918, 'timing/data': 0.02543486375361681, 'timing/model': 0.7184589700773358, 'learning_rate/qwen_vl_interface': 7.9e-06, 'learning_rate/action_model': 7.900000000000001e-05, 'epoch': 0.04}) +[2026-06-17 16:54:22,880][starVLA.training.train_starvla][INFO] - Step 80, Loss: {'train/loss': 0.8430335521697998, 'train/grad_norm_pre_clip': 9.97946834564209, 'timing/data': 0.027652968652546406, 'timing/model': 0.7204494085162878, 'learning_rate/qwen_vl_interface': 8.000000000000001e-06, 'learning_rate/action_model': 8e-05, 'epoch': 0.04}) +[2026-06-17 16:54:25,667][starVLA.training.train_starvla][INFO] - Step 81, Loss: {'train/loss': 1.154937982559204, 'train/grad_norm_pre_clip': 6.5588178634643555, 'timing/data': 0.036242326721549034, 'timing/model': 0.7206394830718637, 'learning_rate/qwen_vl_interface': 8.1e-06, 'learning_rate/action_model': 8.1e-05, 'epoch': 0.04}) +[2026-06-17 16:54:28,440][starVLA.training.train_starvla][INFO] - Step 82, Loss: {'train/loss': 1.1954641342163086, 'train/grad_norm_pre_clip': 7.899061679840088, 'timing/data': 0.023409122601151466, 'timing/model': 0.7264528069645166, 'learning_rate/qwen_vl_interface': 8.2e-06, 'learning_rate/action_model': 8.2e-05, 'epoch': 0.04}) +[2026-06-17 16:54:31,204][starVLA.training.train_starvla][INFO] - Step 83, Loss: {'train/loss': 1.148545265197754, 'train/grad_norm_pre_clip': 5.033731937408447, 'timing/data': 0.023093003779649734, 'timing/model': 0.7245652498677373, 'learning_rate/qwen_vl_interface': 8.3e-06, 'learning_rate/action_model': 8.3e-05, 'epoch': 0.04}) +[2026-06-17 16:54:33,973][starVLA.training.train_starvla][INFO] - Step 84, Loss: {'train/loss': 1.1724677085876465, 'train/grad_norm_pre_clip': 6.58311128616333, 'timing/data': 0.028566899709403515, 'timing/model': 0.7235123077407479, 'learning_rate/qwen_vl_interface': 8.400000000000001e-06, 'learning_rate/action_model': 8.4e-05, 'epoch': 0.04}) +[2026-06-17 16:54:36,733][starVLA.training.train_starvla][INFO] - Step 85, Loss: {'train/loss': 1.1231030225753784, 'train/grad_norm_pre_clip': 5.79580020904541, 'timing/data': 0.024661773815751076, 'timing/model': 0.7196396244689822, 'learning_rate/qwen_vl_interface': 8.5e-06, 'learning_rate/action_model': 8.5e-05, 'epoch': 0.04}) +[2026-06-17 16:54:39,580][starVLA.training.train_starvla][INFO] - Step 86, Loss: {'train/loss': 1.2088043689727783, 'train/grad_norm_pre_clip': 4.926159381866455, 'timing/data': 0.030642904341220856, 'timing/model': 0.7272398509085178, 'learning_rate/qwen_vl_interface': 8.6e-06, 'learning_rate/action_model': 8.6e-05, 'epoch': 0.04}) +[2026-06-17 16:54:42,776][starVLA.training.train_starvla][INFO] - Step 87, Loss: {'train/loss': 0.9830068945884705, 'train/grad_norm_pre_clip': 4.796361923217773, 'timing/data': 0.026407822035253048, 'timing/model': 0.7172936843708158, 'learning_rate/qwen_vl_interface': 8.700000000000001e-06, 'learning_rate/action_model': 8.7e-05, 'epoch': 0.04}) +[2026-06-17 16:54:45,716][starVLA.training.train_starvla][INFO] - Step 88, Loss: {'train/loss': 0.9015918970108032, 'train/grad_norm_pre_clip': 8.292099952697754, 'timing/data': 0.028206069953739643, 'timing/model': 0.7867845604196191, 'learning_rate/qwen_vl_interface': 8.8e-06, 'learning_rate/action_model': 8.800000000000001e-05, 'epoch': 0.04}) +[2026-06-17 16:54:48,575][starVLA.training.train_starvla][INFO] - Step 89, Loss: {'train/loss': 1.590255618095398, 'train/grad_norm_pre_clip': 5.444117546081543, 'timing/data': 0.0270810192450881, 'timing/model': 0.8083304166793823, 'learning_rate/qwen_vl_interface': 8.900000000000001e-06, 'learning_rate/action_model': 8.900000000000001e-05, 'epoch': 0.04}) +[2026-06-17 16:54:51,601][starVLA.training.train_starvla][INFO] - Step 90, Loss: {'train/loss': 1.0619428157806396, 'train/grad_norm_pre_clip': 2.9918553829193115, 'timing/data': 0.03284035716205835, 'timing/model': 0.741046354174614, 'learning_rate/qwen_vl_interface': 9e-06, 'learning_rate/action_model': 9e-05, 'epoch': 0.04}) +[2026-06-17 16:54:54,784][starVLA.training.train_starvla][INFO] - Step 91, Loss: {'train/loss': 0.953579306602478, 'train/grad_norm_pre_clip': 5.560227870941162, 'timing/data': 0.027246586047112942, 'timing/model': 0.834573095664382, 'learning_rate/qwen_vl_interface': 9.100000000000001e-06, 'learning_rate/action_model': 9.1e-05, 'epoch': 0.04}) +[2026-06-17 16:54:58,046][starVLA.training.train_starvla][INFO] - Step 92, Loss: {'train/loss': 0.9913548231124878, 'train/grad_norm_pre_clip': 4.145427227020264, 'timing/data': 0.02885161153972149, 'timing/model': 0.7471087872982025, 'learning_rate/qwen_vl_interface': 9.200000000000002e-06, 'learning_rate/action_model': 9.200000000000001e-05, 'epoch': 0.04}) +[2026-06-17 16:55:00,950][starVLA.training.train_starvla][INFO] - Step 93, Loss: {'train/loss': 1.2079920768737793, 'train/grad_norm_pre_clip': 4.398559093475342, 'timing/data': 0.030732505954802036, 'timing/model': 0.7854943815618753, 'learning_rate/qwen_vl_interface': 9.3e-06, 'learning_rate/action_model': 9.300000000000001e-05, 'epoch': 0.04}) +[2026-06-17 16:55:03,758][starVLA.training.train_starvla][INFO] - Step 94, Loss: {'train/loss': 1.146898865699768, 'train/grad_norm_pre_clip': 6.485319137573242, 'timing/data': 0.028168197721242905, 'timing/model': 0.7330620298162103, 'learning_rate/qwen_vl_interface': 9.4e-06, 'learning_rate/action_model': 9.4e-05, 'epoch': 0.04}) +[2026-06-17 16:55:06,788][starVLA.training.train_starvla][INFO] - Step 95, Loss: {'train/loss': 1.1099656820297241, 'train/grad_norm_pre_clip': 4.097105026245117, 'timing/data': 0.029807508923113346, 'timing/model': 0.7313853790983558, 'learning_rate/qwen_vl_interface': 9.5e-06, 'learning_rate/action_model': 9.5e-05, 'epoch': 0.04}) +[2026-06-17 16:55:10,259][starVLA.training.train_starvla][INFO] - Step 96, Loss: {'train/loss': 0.9554011821746826, 'train/grad_norm_pre_clip': 3.8506662845611572, 'timing/data': 0.025080794468522072, 'timing/model': 1.0955455787479877, 'learning_rate/qwen_vl_interface': 9.600000000000001e-06, 'learning_rate/action_model': 9.6e-05, 'epoch': 0.04}) +[2026-06-17 16:55:13,079][starVLA.training.train_starvla][INFO] - Step 97, Loss: {'train/loss': 0.7866567373275757, 'train/grad_norm_pre_clip': 5.632067680358887, 'timing/data': 0.025301484391093254, 'timing/model': 0.730174127034843, 'learning_rate/qwen_vl_interface': 9.7e-06, 'learning_rate/action_model': 9.7e-05, 'epoch': 0.04}) +[2026-06-17 16:55:16,077][starVLA.training.train_starvla][INFO] - Step 98, Loss: {'train/loss': 0.9312888979911804, 'train/grad_norm_pre_clip': 9.942238807678223, 'timing/data': 0.038427213206887245, 'timing/model': 0.7274939073249698, 'learning_rate/qwen_vl_interface': 9.800000000000001e-06, 'learning_rate/action_model': 9.8e-05, 'epoch': 0.04}) +[2026-06-17 16:55:19,370][starVLA.training.train_starvla][INFO] - Step 99, Loss: {'train/loss': 1.007576823234558, 'train/grad_norm_pre_clip': 5.718916416168213, 'timing/data': 0.03494237083941698, 'timing/model': 0.9893253827467561, 'learning_rate/qwen_vl_interface': 9.9e-06, 'learning_rate/action_model': 9.900000000000001e-05, 'epoch': 0.04}) +[2026-06-17 16:56:53,845][root][INFO] - gcc -pthread -B /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/include -fPIC -O2 -isystem /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/include -fPIC -c /tmp/tmpfohf10ib/test.c -o /tmp/tmpfohf10ib/test.o +[2026-06-17 16:56:53,883][root][INFO] - gcc -pthread -B /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/compiler_compat /tmp/tmpfohf10ib/test.o -laio -o /tmp/tmpfohf10ib/a.out +[2026-06-17 16:56:54,526][root][INFO] - gcc -pthread -B /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/include -fPIC -O2 -isystem /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/include -fPIC -c /tmp/tmp924y817f/test.c -o /tmp/tmp924y817f/test.o +[2026-06-17 16:56:54,541][root][INFO] - gcc -pthread -B /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/compiler_compat /tmp/tmp924y817f/test.o -L/usr/local/cuda -L/usr/local/cuda/lib64 -lcufile -o /tmp/tmp924y817f/a.out +[2026-06-17 16:56:54,584][root][INFO] - gcc -pthread -B /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -O2 -isystem /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/include -fPIC -O2 -isystem /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/include -fPIC -c /tmp/tmpm34jv9mu/test.c -o /tmp/tmpm34jv9mu/test.o +[2026-06-17 16:56:54,596][root][INFO] - gcc -pthread -B /inspire/hdd/project/spatialintelligence/public/lzj/miniconda3/envs/starvla_openvla/compiler_compat /tmp/tmpm34jv9mu/test.o -laio -o /tmp/tmpm34jv9mu/a.out +[2026-06-17 16:58:28,510][starVLA.training.train_starvla][INFO] - Step 100, Loss: {'train/loss': 0.7543548345565796, 'train/grad_norm_pre_clip': 4.842778205871582, 'eval/loss': 1.0844460725784302, 'eval/latency_8/loss': 1.0844460725784302, 'eval/demon_fire_precision': 0.8956594323873122, 'eval/demon_fire_recall': 1.0, 'eval/demon_fire_f1': 0.944958168207838, 'eval/demon_fire_rate_model': 1.0, 'eval/demon_fire_rate_teacher': 0.8956594323873122, 'eval/demon_left_precision': 0.3325213154689403, 'eval/demon_left_recall': 0.5845824411134903, 'eval/demon_left_f1': 0.42391304347826086, 'eval/demon_left_rate_model': 0.22843628269337785, 'eval/demon_left_rate_teacher': 0.12993878686700056, 'eval/demon_right_precision': 0.9333333333333333, 'eval/demon_right_recall': 0.021212121212121213, 'eval/demon_right_f1': 0.04148148148148149, 'eval/demon_right_rate_model': 0.004173622704507512, 'eval/demon_right_rate_teacher': 0.18363939899833054, 'eval/demon_move_f1': 0.2326972624798712, 'eval/demon_cc_f1': 0.4926102845105575, 'eval/latency_8/demon_fire_precision': 0.8956594323873122, 'eval/latency_8/demon_fire_recall': 1.0, 'eval/latency_8/demon_fire_f1': 0.944958168207838, 'eval/latency_8/demon_fire_rate_model': 1.0, 'eval/latency_8/demon_fire_rate_teacher': 0.8956594323873122, 'eval/latency_8/demon_left_precision': 0.3325213154689403, 'eval/latency_8/demon_left_recall': 0.5845824411134903, 'eval/latency_8/demon_left_f1': 0.42391304347826086, 'eval/latency_8/demon_left_rate_model': 0.22843628269337785, 'eval/latency_8/demon_left_rate_teacher': 0.12993878686700056, 'eval/latency_8/demon_right_precision': 0.9333333333333333, 'eval/latency_8/demon_right_recall': 0.021212121212121213, 'eval/latency_8/demon_right_f1': 0.04148148148148149, 'eval/latency_8/demon_right_rate_model': 0.004173622704507512, 'eval/latency_8/demon_right_rate_teacher': 0.18363939899833054, 'eval/latency_8/demon_move_f1': 0.2326972624798712, 'eval/latency_8/demon_cc_f1': 0.4926102845105575, 'timing/data': 0.02338568028062582, 'timing/model': 0.728020547889173, 'learning_rate/qwen_vl_interface': 1e-05, 'learning_rate/action_model': 0.0001, 'epoch': 0.05}) +[2026-06-17 16:58:28,514][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_100_state +[2026-06-17 16:58:28,535][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 16:58:35,133][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_100_state/model.safetensors +[2026-06-17 16:58:49,076][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_100_state/optimizer.bin +[2026-06-17 16:58:49,078][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_100_state/sampler.bin +[2026-06-17 16:58:49,078][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_100_state/sampler_1.bin +[2026-06-17 16:58:49,081][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_100_state/random_states_0.pkl +[2026-06-17 16:58:49,085][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 100 +[2026-06-17 16:58:49,085][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 16:58:49,097][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 16:58:53,001][starVLA.training.train_starvla][INFO] - Step 101, Loss: {'train/loss': 1.5898029804229736, 'train/grad_norm_pre_clip': 16.334686279296875, 'timing/data': 0.025937805883586407, 'timing/model': 0.9759051306173205, 'learning_rate/qwen_vl_interface': 9.999999507659945e-06, 'learning_rate/action_model': 9.999999507659945e-05, 'epoch': 0.05}) +[2026-06-17 16:58:56,855][starVLA.training.train_starvla][INFO] - Step 102, Loss: {'train/loss': 0.9343233704566956, 'train/grad_norm_pre_clip': 10.502344131469727, 'timing/data': 0.022222265601158142, 'timing/model': 0.9242323338985443, 'learning_rate/qwen_vl_interface': 9.99999803063988e-06, 'learning_rate/action_model': 9.99999803063988e-05, 'epoch': 0.05}) +[2026-06-17 16:59:00,711][starVLA.training.train_starvla][INFO] - Step 103, Loss: {'train/loss': 1.1617443561553955, 'train/grad_norm_pre_clip': 7.649282932281494, 'timing/data': 0.025660287588834763, 'timing/model': 0.9238929282873869, 'learning_rate/qwen_vl_interface': 9.999995568940113e-06, 'learning_rate/action_model': 9.999995568940113e-05, 'epoch': 0.05}) +[2026-06-17 16:59:04,639][starVLA.training.train_starvla][INFO] - Step 104, Loss: {'train/loss': 1.0631194114685059, 'train/grad_norm_pre_clip': 5.381468296051025, 'timing/data': 0.03941654320806265, 'timing/model': 1.240578499622643, 'learning_rate/qwen_vl_interface': 9.999992122561152e-06, 'learning_rate/action_model': 9.999992122561152e-05, 'epoch': 0.05}) +[2026-06-17 16:59:08,254][starVLA.training.train_starvla][INFO] - Step 105, Loss: {'train/loss': 0.8274486064910889, 'train/grad_norm_pre_clip': 3.5930163860321045, 'timing/data': 0.026487653143703938, 'timing/model': 0.9352363646030426, 'learning_rate/qwen_vl_interface': 9.999987691503714e-06, 'learning_rate/action_model': 9.999987691503714e-05, 'epoch': 0.05}) +[2026-06-17 16:59:12,101][starVLA.training.train_starvla][INFO] - Step 106, Loss: {'train/loss': 1.083579659461975, 'train/grad_norm_pre_clip': 3.983243465423584, 'timing/data': 0.021484115161001682, 'timing/model': 0.9354280615225434, 'learning_rate/qwen_vl_interface': 9.999982275768716e-06, 'learning_rate/action_model': 9.999982275768715e-05, 'epoch': 0.05}) +[2026-06-17 16:59:15,861][starVLA.training.train_starvla][INFO] - Step 107, Loss: {'train/loss': 0.7833734154701233, 'train/grad_norm_pre_clip': 5.251889705657959, 'timing/data': 0.02502588927745819, 'timing/model': 0.9338478529825807, 'learning_rate/qwen_vl_interface': 9.99997587535728e-06, 'learning_rate/action_model': 9.99997587535728e-05, 'epoch': 0.05}) +[2026-06-17 16:59:19,710][starVLA.training.train_starvla][INFO] - Step 108, Loss: {'train/loss': 1.0652998685836792, 'train/grad_norm_pre_clip': 8.104029655456543, 'timing/data': 0.025378097780048847, 'timing/model': 0.9365443009883165, 'learning_rate/qwen_vl_interface': 9.999968490270735e-06, 'learning_rate/action_model': 9.999968490270735e-05, 'epoch': 0.05}) +[2026-06-17 16:59:23,556][starVLA.training.train_starvla][INFO] - Step 109, Loss: {'train/loss': 0.8594313859939575, 'train/grad_norm_pre_clip': 5.808663845062256, 'timing/data': 0.024165303446352482, 'timing/model': 0.9547256510704756, 'learning_rate/qwen_vl_interface': 9.99996012051061e-06, 'learning_rate/action_model': 9.99996012051061e-05, 'epoch': 0.05}) +[2026-06-17 16:59:27,481][starVLA.training.train_starvla][INFO] - Step 110, Loss: {'train/loss': 1.0070099830627441, 'train/grad_norm_pre_clip': 3.0175118446350098, 'timing/data': 0.02545677125453949, 'timing/model': 1.115612043067813, 'learning_rate/qwen_vl_interface': 9.999950766078641e-06, 'learning_rate/action_model': 9.99995076607864e-05, 'epoch': 0.05}) +[2026-06-17 16:59:31,291][starVLA.training.train_starvla][INFO] - Step 111, Loss: {'train/loss': 0.9697561264038086, 'train/grad_norm_pre_clip': 6.726690769195557, 'timing/data': 0.026898115873336792, 'timing/model': 1.1191463489085436, 'learning_rate/qwen_vl_interface': 9.999940426976768e-06, 'learning_rate/action_model': 9.999940426976768e-05, 'epoch': 0.05}) +[2026-06-17 16:59:34,982][starVLA.training.train_starvla][INFO] - Step 112, Loss: {'train/loss': 0.8583449125289917, 'train/grad_norm_pre_clip': 7.460453987121582, 'timing/data': 0.03413231950253248, 'timing/model': 0.9282870050519705, 'learning_rate/qwen_vl_interface': 9.999929103207133e-06, 'learning_rate/action_model': 9.999929103207132e-05, 'epoch': 0.05}) +[2026-06-17 16:59:38,835][starVLA.training.train_starvla][INFO] - Step 113, Loss: {'train/loss': 0.724861741065979, 'train/grad_norm_pre_clip': 11.90494155883789, 'timing/data': 0.02510082721710205, 'timing/model': 0.941569889895618, 'learning_rate/qwen_vl_interface': 9.999916794772083e-06, 'learning_rate/action_model': 9.999916794772082e-05, 'epoch': 0.05}) +[2026-06-17 16:59:42,709][starVLA.training.train_starvla][INFO] - Step 114, Loss: {'train/loss': 0.9113349318504333, 'train/grad_norm_pre_clip': 3.5168137550354004, 'timing/data': 0.02267881017178297, 'timing/model': 0.9667765228077769, 'learning_rate/qwen_vl_interface': 9.99990350167417e-06, 'learning_rate/action_model': 9.999903501674169e-05, 'epoch': 0.05}) +[2026-06-17 16:59:46,458][starVLA.training.train_starvla][INFO] - Step 115, Loss: {'train/loss': 0.8894466161727905, 'train/grad_norm_pre_clip': 7.768228054046631, 'timing/data': 0.031079129315912724, 'timing/model': 0.9361083898693323, 'learning_rate/qwen_vl_interface': 9.99988922391615e-06, 'learning_rate/action_model': 9.99988922391615e-05, 'epoch': 0.05}) +[2026-06-17 16:59:50,236][starVLA.training.train_starvla][INFO] - Step 116, Loss: {'train/loss': 0.8160854578018188, 'train/grad_norm_pre_clip': 7.129879951477051, 'timing/data': 0.02546916250139475, 'timing/model': 0.9253812180832028, 'learning_rate/qwen_vl_interface': 9.999873961500985e-06, 'learning_rate/action_model': 9.999873961500984e-05, 'epoch': 0.05}) +[2026-06-17 16:59:54,194][starVLA.training.train_starvla][INFO] - Step 117, Loss: {'train/loss': 1.260056972503662, 'train/grad_norm_pre_clip': 10.719533920288086, 'timing/data': 0.02463266160339117, 'timing/model': 1.1476994575932622, 'learning_rate/qwen_vl_interface': 9.999857714431835e-06, 'learning_rate/action_model': 9.999857714431834e-05, 'epoch': 0.05}) +[2026-06-17 16:59:57,932][starVLA.training.train_starvla][INFO] - Step 118, Loss: {'train/loss': 1.2440710067749023, 'train/grad_norm_pre_clip': 9.867912292480469, 'timing/data': 0.022591757588088512, 'timing/model': 0.9345322586596012, 'learning_rate/qwen_vl_interface': 9.999840482712069e-06, 'learning_rate/action_model': 9.999840482712068e-05, 'epoch': 0.05}) +[2026-06-17 17:00:02,240][starVLA.training.train_starvla][INFO] - Step 119, Loss: {'train/loss': 1.4833418130874634, 'train/grad_norm_pre_clip': 9.220990180969238, 'timing/data': 0.02064269222319126, 'timing/model': 1.3073936635628343, 'learning_rate/qwen_vl_interface': 9.99982226634526e-06, 'learning_rate/action_model': 9.99982226634526e-05, 'epoch': 0.05}) +[2026-06-17 17:00:06,290][starVLA.training.train_starvla][INFO] - Step 120, Loss: {'train/loss': 0.7498841881752014, 'train/grad_norm_pre_clip': 3.609933614730835, 'timing/data': 0.026831193827092648, 'timing/model': 0.9432017402723432, 'learning_rate/qwen_vl_interface': 9.999803065335185e-06, 'learning_rate/action_model': 9.999803065335185e-05, 'epoch': 0.05}) +[2026-06-17 17:00:10,225][starVLA.training.train_starvla][INFO] - Step 121, Loss: {'train/loss': 1.0379763841629028, 'train/grad_norm_pre_clip': 4.44509220123291, 'timing/data': 0.02737036533653736, 'timing/model': 1.0553962662816048, 'learning_rate/qwen_vl_interface': 9.999782879685825e-06, 'learning_rate/action_model': 9.999782879685824e-05, 'epoch': 0.05}) +[2026-06-17 17:00:13,999][starVLA.training.train_starvla][INFO] - Step 122, Loss: {'train/loss': 0.9960706233978271, 'train/grad_norm_pre_clip': 7.745736122131348, 'timing/data': 0.02160944789648056, 'timing/model': 0.9306833986192942, 'learning_rate/qwen_vl_interface': 9.99976170940136e-06, 'learning_rate/action_model': 9.99976170940136e-05, 'epoch': 0.06}) +[2026-06-17 17:00:17,791][starVLA.training.train_starvla][INFO] - Step 123, Loss: {'train/loss': 0.888119101524353, 'train/grad_norm_pre_clip': 5.779280185699463, 'timing/data': 0.024365835823118687, 'timing/model': 0.9335398096591234, 'learning_rate/qwen_vl_interface': 9.999739554486184e-06, 'learning_rate/action_model': 9.999739554486182e-05, 'epoch': 0.06}) +[2026-06-17 17:00:21,583][starVLA.training.train_starvla][INFO] - Step 124, Loss: {'train/loss': 0.9354068040847778, 'train/grad_norm_pre_clip': 4.398255348205566, 'timing/data': 0.029623021371662617, 'timing/model': 1.1454575015231967, 'learning_rate/qwen_vl_interface': 9.999716414944887e-06, 'learning_rate/action_model': 9.999716414944886e-05, 'epoch': 0.06}) +[2026-06-17 17:00:25,288][starVLA.training.train_starvla][INFO] - Step 125, Loss: {'train/loss': 0.7934780120849609, 'train/grad_norm_pre_clip': 5.310899257659912, 'timing/data': 0.03002410475164652, 'timing/model': 0.9321966078132391, 'learning_rate/qwen_vl_interface': 9.999692290782264e-06, 'learning_rate/action_model': 9.999692290782263e-05, 'epoch': 0.06}) +[2026-06-17 17:00:29,397][starVLA.training.train_starvla][INFO] - Step 126, Loss: {'train/loss': 0.9918950796127319, 'train/grad_norm_pre_clip': 5.722708702087402, 'timing/data': 0.02643804159015417, 'timing/model': 1.1386448135599494, 'learning_rate/qwen_vl_interface': 9.999667182003321e-06, 'learning_rate/action_model': 9.999667182003321e-05, 'epoch': 0.06}) +[2026-06-17 17:00:33,210][starVLA.training.train_starvla][INFO] - Step 127, Loss: {'train/loss': 0.8491494655609131, 'train/grad_norm_pre_clip': 6.17737340927124, 'timing/data': 0.03077522572129965, 'timing/model': 0.9236585143953562, 'learning_rate/qwen_vl_interface': 9.99964108861326e-06, 'learning_rate/action_model': 9.99964108861326e-05, 'epoch': 0.06}) +[2026-06-17 17:00:37,018][starVLA.training.train_starvla][INFO] - Step 128, Loss: {'train/loss': 1.0266410112380981, 'train/grad_norm_pre_clip': 6.67377233505249, 'timing/data': 0.026439505629241467, 'timing/model': 0.9286442156881094, 'learning_rate/qwen_vl_interface': 9.99961401061749e-06, 'learning_rate/action_model': 9.99961401061749e-05, 'epoch': 0.06}) +[2026-06-17 17:00:40,893][starVLA.training.train_starvla][INFO] - Step 129, Loss: {'train/loss': 0.8669521808624268, 'train/grad_norm_pre_clip': 4.738900661468506, 'timing/data': 0.02971648331731558, 'timing/model': 0.998434160836041, 'learning_rate/qwen_vl_interface': 9.999585948021623e-06, 'learning_rate/action_model': 9.999585948021624e-05, 'epoch': 0.06}) +[2026-06-17 17:00:44,695][starVLA.training.train_starvla][INFO] - Step 130, Loss: {'train/loss': 1.0359669923782349, 'train/grad_norm_pre_clip': 4.955963134765625, 'timing/data': 0.025839018635451794, 'timing/model': 1.141200021840632, 'learning_rate/qwen_vl_interface': 9.99955690083148e-06, 'learning_rate/action_model': 9.99955690083148e-05, 'epoch': 0.06}) +[2026-06-17 17:00:48,798][starVLA.training.train_starvla][INFO] - Step 131, Loss: {'train/loss': 1.5837973356246948, 'train/grad_norm_pre_clip': 9.450804710388184, 'timing/data': 0.022887110710144043, 'timing/model': 1.1243979819118977, 'learning_rate/qwen_vl_interface': 9.99952686905308e-06, 'learning_rate/action_model': 9.999526869053081e-05, 'epoch': 0.06}) +[2026-06-17 17:00:52,420][starVLA.training.train_starvla][INFO] - Step 132, Loss: {'train/loss': 1.0195914506912231, 'train/grad_norm_pre_clip': 3.9511799812316895, 'timing/data': 0.03416185732930899, 'timing/model': 0.9423529161140323, 'learning_rate/qwen_vl_interface': 9.99949585269265e-06, 'learning_rate/action_model': 9.99949585269265e-05, 'epoch': 0.06}) +[2026-06-17 17:00:56,221][starVLA.training.train_starvla][INFO] - Step 133, Loss: {'train/loss': 0.9217588901519775, 'train/grad_norm_pre_clip': 4.484586238861084, 'timing/data': 0.031125039793550968, 'timing/model': 0.9227052414789796, 'learning_rate/qwen_vl_interface': 9.99946385175662e-06, 'learning_rate/action_model': 9.999463851756619e-05, 'epoch': 0.06}) +[2026-06-17 17:01:00,098][starVLA.training.train_starvla][INFO] - Step 134, Loss: {'train/loss': 0.862670361995697, 'train/grad_norm_pre_clip': 9.514795303344727, 'timing/data': 0.022540700621902943, 'timing/model': 0.9201890230178833, 'learning_rate/qwen_vl_interface': 9.999430866251623e-06, 'learning_rate/action_model': 9.999430866251622e-05, 'epoch': 0.06}) +[2026-06-17 17:01:03,911][starVLA.training.train_starvla][INFO] - Step 135, Loss: {'train/loss': 1.2761770486831665, 'train/grad_norm_pre_clip': 12.812026977539062, 'timing/data': 0.029387877322733402, 'timing/model': 0.949050810188055, 'learning_rate/qwen_vl_interface': 9.999396896184494e-06, 'learning_rate/action_model': 9.999396896184494e-05, 'epoch': 0.06}) +[2026-06-17 17:01:07,840][starVLA.training.train_starvla][INFO] - Step 136, Loss: {'train/loss': 0.9416841268539429, 'train/grad_norm_pre_clip': 12.838310241699219, 'timing/data': 0.025170687586069107, 'timing/model': 0.9215857349336147, 'learning_rate/qwen_vl_interface': 9.99936194156228e-06, 'learning_rate/action_model': 9.99936194156228e-05, 'epoch': 0.06}) +[2026-06-17 17:01:11,654][starVLA.training.train_starvla][INFO] - Step 137, Loss: {'train/loss': 1.1648935079574585, 'train/grad_norm_pre_clip': 8.33169937133789, 'timing/data': 0.027329007163643837, 'timing/model': 1.136832058429718, 'learning_rate/qwen_vl_interface': 9.999326002392225e-06, 'learning_rate/action_model': 9.999326002392224e-05, 'epoch': 0.06}) +[2026-06-17 17:01:15,553][starVLA.training.train_starvla][INFO] - Step 138, Loss: {'train/loss': 0.90981125831604, 'train/grad_norm_pre_clip': 4.712766647338867, 'timing/data': 0.03753833007067442, 'timing/model': 1.100743643939495, 'learning_rate/qwen_vl_interface': 9.999289078681779e-06, 'learning_rate/action_model': 9.999289078681778e-05, 'epoch': 0.06}) +[2026-06-17 17:01:19,157][starVLA.training.train_starvla][INFO] - Step 139, Loss: {'train/loss': 1.1284765005111694, 'train/grad_norm_pre_clip': 4.374481678009033, 'timing/data': 0.022299306467175484, 'timing/model': 0.9628916578367352, 'learning_rate/qwen_vl_interface': 9.999251170438594e-06, 'learning_rate/action_model': 9.999251170438594e-05, 'epoch': 0.06}) +[2026-06-17 17:01:22,979][starVLA.training.train_starvla][INFO] - Step 140, Loss: {'train/loss': 1.332549810409546, 'train/grad_norm_pre_clip': 5.953297138214111, 'timing/data': 0.04040268249809742, 'timing/model': 0.9284633919596672, 'learning_rate/qwen_vl_interface': 9.999212277670533e-06, 'learning_rate/action_model': 9.999212277670533e-05, 'epoch': 0.06}) +[2026-06-17 17:01:26,796][starVLA.training.train_starvla][INFO] - Step 141, Loss: {'train/loss': 1.1750380992889404, 'train/grad_norm_pre_clip': 4.765677452087402, 'timing/data': 0.025060288608074188, 'timing/model': 0.925862186588347, 'learning_rate/qwen_vl_interface': 9.999172400385658e-06, 'learning_rate/action_model': 9.999172400385658e-05, 'epoch': 0.06}) +[2026-06-17 17:01:30,589][starVLA.training.train_starvla][INFO] - Step 142, Loss: {'train/loss': 0.9972698092460632, 'train/grad_norm_pre_clip': 5.845907211303711, 'timing/data': 0.03506391029804945, 'timing/model': 0.9292498677968979, 'learning_rate/qwen_vl_interface': 9.999131538592232e-06, 'learning_rate/action_model': 9.999131538592232e-05, 'epoch': 0.06}) +[2026-06-17 17:01:34,386][starVLA.training.train_starvla][INFO] - Step 143, Loss: {'train/loss': 0.9566828012466431, 'train/grad_norm_pre_clip': 2.775258779525757, 'timing/data': 0.022675834596157074, 'timing/model': 0.9281343808397651, 'learning_rate/qwen_vl_interface': 9.999089692298729e-06, 'learning_rate/action_model': 9.99908969229873e-05, 'epoch': 0.06}) +[2026-06-17 17:01:38,216][starVLA.training.train_starvla][INFO] - Step 144, Loss: {'train/loss': 0.8987962603569031, 'train/grad_norm_pre_clip': 3.230325222015381, 'timing/data': 0.03026542440056801, 'timing/model': 1.1415729029104114, 'learning_rate/qwen_vl_interface': 9.999046861513822e-06, 'learning_rate/action_model': 9.999046861513821e-05, 'epoch': 0.07}) +[2026-06-17 17:01:41,821][starVLA.training.train_starvla][INFO] - Step 145, Loss: {'train/loss': 0.8038846850395203, 'train/grad_norm_pre_clip': 3.9154763221740723, 'timing/data': 0.025329352356493473, 'timing/model': 0.9333441909402609, 'learning_rate/qwen_vl_interface': 9.99900304624639e-06, 'learning_rate/action_model': 9.99900304624639e-05, 'epoch': 0.07}) +[2026-06-17 17:01:45,614][starVLA.training.train_starvla][INFO] - Step 146, Loss: {'train/loss': 0.8691929578781128, 'train/grad_norm_pre_clip': 4.038172721862793, 'timing/data': 0.03133841045200825, 'timing/model': 0.9244142919778824, 'learning_rate/qwen_vl_interface': 9.998958246505518e-06, 'learning_rate/action_model': 9.998958246505516e-05, 'epoch': 0.07}) +[2026-06-17 17:01:49,378][starVLA.training.train_starvla][INFO] - Step 147, Loss: {'train/loss': 0.705109715461731, 'train/grad_norm_pre_clip': 4.677451133728027, 'timing/data': 0.03192002605646849, 'timing/model': 0.9306015707552433, 'learning_rate/qwen_vl_interface': 9.998912462300489e-06, 'learning_rate/action_model': 9.998912462300488e-05, 'epoch': 0.07}) +[2026-06-17 17:01:53,169][starVLA.training.train_starvla][INFO] - Step 148, Loss: {'train/loss': 0.46453797817230225, 'train/grad_norm_pre_clip': 2.5881025791168213, 'timing/data': 0.027160272002220154, 'timing/model': 0.9323278004303575, 'learning_rate/qwen_vl_interface': 9.998865693640797e-06, 'learning_rate/action_model': 9.998865693640796e-05, 'epoch': 0.07}) +[2026-06-17 17:01:57,186][starVLA.training.train_starvla][INFO] - Step 149, Loss: {'train/loss': 0.8593137860298157, 'train/grad_norm_pre_clip': 3.9826290607452393, 'timing/data': 0.02602084632962942, 'timing/model': 0.9575360240414739, 'learning_rate/qwen_vl_interface': 9.998817940536138e-06, 'learning_rate/action_model': 9.998817940536137e-05, 'epoch': 0.07}) +[2026-06-17 17:02:01,824][starVLA.training.train_starvla][INFO] - Step 150, Loss: {'train/loss': 0.7335131168365479, 'train/grad_norm_pre_clip': 4.428955554962158, 'timing/data': 0.02523856982588768, 'timing/model': 1.1261635543778539, 'learning_rate/qwen_vl_interface': 9.99876920299641e-06, 'learning_rate/action_model': 9.998769202996408e-05, 'epoch': 0.07}) +[2026-06-17 17:02:05,583][starVLA.training.train_starvla][INFO] - Step 151, Loss: {'train/loss': 0.7758022546768188, 'train/grad_norm_pre_clip': 3.7396762371063232, 'timing/data': 0.022786525078117847, 'timing/model': 1.089228754863143, 'learning_rate/qwen_vl_interface': 9.998719481031715e-06, 'learning_rate/action_model': 9.998719481031715e-05, 'epoch': 0.07}) +[2026-06-17 17:02:09,189][starVLA.training.train_starvla][INFO] - Step 152, Loss: {'train/loss': 1.2231779098510742, 'train/grad_norm_pre_clip': 5.636488914489746, 'timing/data': 0.024970908649265766, 'timing/model': 0.9294594954699278, 'learning_rate/qwen_vl_interface': 9.998668774652363e-06, 'learning_rate/action_model': 9.998668774652362e-05, 'epoch': 0.07}) +[2026-06-17 17:02:13,064][starVLA.training.train_starvla][INFO] - Step 153, Loss: {'train/loss': 1.081765055656433, 'train/grad_norm_pre_clip': 5.74921989440918, 'timing/data': 0.029493297450244427, 'timing/model': 0.9336253060027957, 'learning_rate/qwen_vl_interface': 9.998617083868863e-06, 'learning_rate/action_model': 9.998617083868863e-05, 'epoch': 0.07}) +[2026-06-17 17:02:16,878][starVLA.training.train_starvla][INFO] - Step 154, Loss: {'train/loss': 1.184003233909607, 'train/grad_norm_pre_clip': 5.12278413772583, 'timing/data': 0.028495749458670616, 'timing/model': 0.9203105932101607, 'learning_rate/qwen_vl_interface': 9.998564408691933e-06, 'learning_rate/action_model': 9.998564408691932e-05, 'epoch': 0.07}) +[2026-06-17 17:02:20,637][starVLA.training.train_starvla][INFO] - Step 155, Loss: {'train/loss': 0.8247854113578796, 'train/grad_norm_pre_clip': 5.144400596618652, 'timing/data': 0.02466469258069992, 'timing/model': 0.9293115762993693, 'learning_rate/qwen_vl_interface': 9.998510749132491e-06, 'learning_rate/action_model': 9.998510749132491e-05, 'epoch': 0.07}) +[2026-06-17 17:02:24,463][starVLA.training.train_starvla][INFO] - Step 156, Loss: {'train/loss': 1.32381010055542, 'train/grad_norm_pre_clip': 5.453355312347412, 'timing/data': 0.021128094755113125, 'timing/model': 0.9440859667956829, 'learning_rate/qwen_vl_interface': 9.998456105201659e-06, 'learning_rate/action_model': 9.99845610520166e-05, 'epoch': 0.07}) +[2026-06-17 17:02:28,397][starVLA.training.train_starvla][INFO] - Step 157, Loss: {'train/loss': 0.8327551484107971, 'train/grad_norm_pre_clip': 4.339565277099609, 'timing/data': 0.02662545721977949, 'timing/model': 1.150766889564693, 'learning_rate/qwen_vl_interface': 9.99840047691077e-06, 'learning_rate/action_model': 9.99840047691077e-05, 'epoch': 0.07}) +[2026-06-17 17:02:32,268][starVLA.training.train_starvla][INFO] - Step 158, Loss: {'train/loss': 0.8773046135902405, 'train/grad_norm_pre_clip': 4.346442699432373, 'timing/data': 0.029037730768322945, 'timing/model': 1.1115119755268097, 'learning_rate/qwen_vl_interface': 9.998343864271352e-06, 'learning_rate/action_model': 9.998343864271351e-05, 'epoch': 0.07}) +[2026-06-17 17:02:36,108][starVLA.training.train_starvla][INFO] - Step 159, Loss: {'train/loss': 0.9201170802116394, 'train/grad_norm_pre_clip': 1.9397869110107422, 'timing/data': 0.02912031300365925, 'timing/model': 0.9429482640698552, 'learning_rate/qwen_vl_interface': 9.99828626729514e-06, 'learning_rate/action_model': 9.998286267295139e-05, 'epoch': 0.07}) +[2026-06-17 17:02:39,929][starVLA.training.train_starvla][INFO] - Step 160, Loss: {'train/loss': 0.9077041149139404, 'train/grad_norm_pre_clip': 4.990501403808594, 'timing/data': 0.0296003557741642, 'timing/model': 0.9375065360218287, 'learning_rate/qwen_vl_interface': 9.998227685994077e-06, 'learning_rate/action_model': 9.998227685994075e-05, 'epoch': 0.07}) +[2026-06-17 17:02:43,711][starVLA.training.train_starvla][INFO] - Step 161, Loss: {'train/loss': 0.7404306530952454, 'train/grad_norm_pre_clip': 3.4023754596710205, 'timing/data': 0.02602569479495287, 'timing/model': 0.9388697715476155, 'learning_rate/qwen_vl_interface': 9.998168120380303e-06, 'learning_rate/action_model': 9.998168120380303e-05, 'epoch': 0.07}) +[2026-06-17 17:02:47,556][starVLA.training.train_starvla][INFO] - Step 162, Loss: {'train/loss': 0.816785454750061, 'train/grad_norm_pre_clip': 2.4337782859802246, 'timing/data': 0.023812499828636646, 'timing/model': 0.9820818863809109, 'learning_rate/qwen_vl_interface': 9.99810757046617e-06, 'learning_rate/action_model': 9.99810757046617e-05, 'epoch': 0.07}) +[2026-06-17 17:02:51,437][starVLA.training.train_starvla][INFO] - Step 163, Loss: {'train/loss': 0.42926645278930664, 'train/grad_norm_pre_clip': 2.628319263458252, 'timing/data': 0.025607488118112087, 'timing/model': 0.9359625214710832, 'learning_rate/qwen_vl_interface': 9.998046036264229e-06, 'learning_rate/action_model': 9.99804603626423e-05, 'epoch': 0.07}) +[2026-06-17 17:02:56,573][starVLA.training.train_starvla][INFO] - Step 164, Loss: {'train/loss': 0.8537225127220154, 'train/grad_norm_pre_clip': 3.398932695388794, 'timing/data': 0.037816367112100124, 'timing/model': 1.4771185861900449, 'learning_rate/qwen_vl_interface': 9.997983517787236e-06, 'learning_rate/action_model': 9.997983517787236e-05, 'epoch': 0.07}) +[2026-06-17 17:03:00,614][starVLA.training.train_starvla][INFO] - Step 165, Loss: {'train/loss': 0.6778502464294434, 'train/grad_norm_pre_clip': 4.832521438598633, 'timing/data': 0.03271812945604324, 'timing/model': 0.9494970785453916, 'learning_rate/qwen_vl_interface': 9.99792001504815e-06, 'learning_rate/action_model': 9.997920015048149e-05, 'epoch': 0.07}) +[2026-06-17 17:03:05,068][starVLA.training.train_starvla][INFO] - Step 166, Loss: {'train/loss': 0.7636123895645142, 'train/grad_norm_pre_clip': 4.413714408874512, 'timing/data': 0.043854775838553905, 'timing/model': 1.2850585049018264, 'learning_rate/qwen_vl_interface': 9.997855528060136e-06, 'learning_rate/action_model': 9.997855528060135e-05, 'epoch': 0.08}) +[2026-06-17 17:03:09,353][starVLA.training.train_starvla][INFO] - Step 167, Loss: {'train/loss': 0.48866772651672363, 'train/grad_norm_pre_clip': 2.7347333431243896, 'timing/data': 0.03476206958293915, 'timing/model': 0.9607968796044588, 'learning_rate/qwen_vl_interface': 9.997790056836562e-06, 'learning_rate/action_model': 9.997790056836562e-05, 'epoch': 0.08}) +[2026-06-17 17:03:13,257][starVLA.training.train_starvla][INFO] - Step 168, Loss: {'train/loss': 1.5312590599060059, 'train/grad_norm_pre_clip': 8.321100234985352, 'timing/data': 0.03614567872136831, 'timing/model': 0.974909070879221, 'learning_rate/qwen_vl_interface': 9.997723601390998e-06, 'learning_rate/action_model': 9.997723601390998e-05, 'epoch': 0.08}) +[2026-06-17 17:03:17,195][starVLA.training.train_starvla][INFO] - Step 169, Loss: {'train/loss': 1.0009032487869263, 'train/grad_norm_pre_clip': 8.353460311889648, 'timing/data': 0.025171110406517982, 'timing/model': 0.9854855993762612, 'learning_rate/qwen_vl_interface': 9.997656161737227e-06, 'learning_rate/action_model': 9.997656161737226e-05, 'epoch': 0.08}) +[2026-06-17 17:03:21,094][starVLA.training.train_starvla][INFO] - Step 170, Loss: {'train/loss': 1.0383329391479492, 'train/grad_norm_pre_clip': 12.810720443725586, 'timing/data': 0.02298352774232626, 'timing/model': 1.185335018672049, 'learning_rate/qwen_vl_interface': 9.997587737889222e-06, 'learning_rate/action_model': 9.997587737889221e-05, 'epoch': 0.08}) +[2026-06-17 17:03:25,244][starVLA.training.train_starvla][INFO] - Step 171, Loss: {'train/loss': 0.9072451591491699, 'train/grad_norm_pre_clip': 4.4451375007629395, 'timing/data': 0.030239499174058437, 'timing/model': 1.2302178898826241, 'learning_rate/qwen_vl_interface': 9.997518329861172e-06, 'learning_rate/action_model': 9.997518329861171e-05, 'epoch': 0.08}) +[2026-06-17 17:03:28,924][starVLA.training.train_starvla][INFO] - Step 172, Loss: {'train/loss': 1.2188067436218262, 'train/grad_norm_pre_clip': 6.8742570877075195, 'timing/data': 0.03971507865935564, 'timing/model': 0.9574633995071054, 'learning_rate/qwen_vl_interface': 9.997447937667462e-06, 'learning_rate/action_model': 9.997447937667462e-05, 'epoch': 0.08}) +[2026-06-17 17:03:32,779][starVLA.training.train_starvla][INFO] - Step 173, Loss: {'train/loss': 1.381616473197937, 'train/grad_norm_pre_clip': 7.338604927062988, 'timing/data': 0.023277362808585167, 'timing/model': 0.9372393749654293, 'learning_rate/qwen_vl_interface': 9.997376561322686e-06, 'learning_rate/action_model': 9.997376561322686e-05, 'epoch': 0.08}) +[2026-06-17 17:03:36,640][starVLA.training.train_starvla][INFO] - Step 174, Loss: {'train/loss': 1.4123679399490356, 'train/grad_norm_pre_clip': 11.073025703430176, 'timing/data': 0.02840580139309168, 'timing/model': 0.9373990437015891, 'learning_rate/qwen_vl_interface': 9.997304200841642e-06, 'learning_rate/action_model': 9.997304200841641e-05, 'epoch': 0.08}) +[2026-06-17 17:03:40,520][starVLA.training.train_starvla][INFO] - Step 175, Loss: {'train/loss': 1.115614652633667, 'train/grad_norm_pre_clip': 9.09473991394043, 'timing/data': 0.0335758738219738, 'timing/model': 0.9514469923451543, 'learning_rate/qwen_vl_interface': 9.997230856239327e-06, 'learning_rate/action_model': 9.997230856239327e-05, 'epoch': 0.08}) +[2026-06-17 17:03:44,422][starVLA.training.train_starvla][INFO] - Step 176, Loss: {'train/loss': 1.190015435218811, 'train/grad_norm_pre_clip': 10.672455787658691, 'timing/data': 0.04171501751989126, 'timing/model': 0.9330113455653191, 'learning_rate/qwen_vl_interface': 9.997156527530948e-06, 'learning_rate/action_model': 9.997156527530948e-05, 'epoch': 0.08}) +[2026-06-17 17:03:48,236][starVLA.training.train_starvla][INFO] - Step 177, Loss: {'train/loss': 0.9167132377624512, 'train/grad_norm_pre_clip': 8.007118225097656, 'timing/data': 0.033181704580783844, 'timing/model': 1.1259289477020502, 'learning_rate/qwen_vl_interface': 9.997081214731915e-06, 'learning_rate/action_model': 9.997081214731914e-05, 'epoch': 0.08}) +[2026-06-17 17:03:51,825][starVLA.training.train_starvla][INFO] - Step 178, Loss: {'train/loss': 0.874774694442749, 'train/grad_norm_pre_clip': 4.527942180633545, 'timing/data': 0.023866698145866394, 'timing/model': 0.9222810557112098, 'learning_rate/qwen_vl_interface': 9.997004917857834e-06, 'learning_rate/action_model': 9.997004917857834e-05, 'epoch': 0.08}) +[2026-06-17 17:03:55,626][starVLA.training.train_starvla][INFO] - Step 179, Loss: {'train/loss': 0.7688978910446167, 'train/grad_norm_pre_clip': 5.303214073181152, 'timing/data': 0.03546289820224047, 'timing/model': 0.9480981174856424, 'learning_rate/qwen_vl_interface': 9.996927636924529e-06, 'learning_rate/action_model': 9.996927636924528e-05, 'epoch': 0.08}) +[2026-06-17 17:03:59,417][starVLA.training.train_starvla][INFO] - Step 180, Loss: {'train/loss': 1.0227410793304443, 'train/grad_norm_pre_clip': 6.088117599487305, 'timing/data': 0.028956666588783264, 'timing/model': 0.9466894483193755, 'learning_rate/qwen_vl_interface': 9.996849371948016e-06, 'learning_rate/action_model': 9.996849371948016e-05, 'epoch': 0.08}) +[2026-06-17 17:04:03,388][starVLA.training.train_starvla][INFO] - Step 181, Loss: {'train/loss': 0.8828924894332886, 'train/grad_norm_pre_clip': 4.831934928894043, 'timing/data': 0.025632301345467567, 'timing/model': 1.0823263265192509, 'learning_rate/qwen_vl_interface': 9.996770122944518e-06, 'learning_rate/action_model': 9.996770122944518e-05, 'epoch': 0.08}) +[2026-06-17 17:04:07,363][starVLA.training.train_starvla][INFO] - Step 182, Loss: {'train/loss': 1.1239588260650635, 'train/grad_norm_pre_clip': 5.287752628326416, 'timing/data': 0.03205002564936876, 'timing/model': 1.004053176380694, 'learning_rate/qwen_vl_interface': 9.99668988993047e-06, 'learning_rate/action_model': 9.99668988993047e-05, 'epoch': 0.08}) +[2026-06-17 17:04:11,201][starVLA.training.train_starvla][INFO] - Step 183, Loss: {'train/loss': 1.314661979675293, 'train/grad_norm_pre_clip': 6.089856147766113, 'timing/data': 0.023437353782355785, 'timing/model': 0.9172131093218923, 'learning_rate/qwen_vl_interface': 9.996608672922496e-06, 'learning_rate/action_model': 9.996608672922496e-05, 'epoch': 0.08}) +[2026-06-17 17:04:15,009][starVLA.training.train_starvla][INFO] - Step 184, Loss: {'train/loss': 0.5634939074516296, 'train/grad_norm_pre_clip': 2.8963229656219482, 'timing/data': 0.03372273966670036, 'timing/model': 1.1518806479871273, 'learning_rate/qwen_vl_interface': 9.996526471937438e-06, 'learning_rate/action_model': 9.996526471937439e-05, 'epoch': 0.08}) +[2026-06-17 17:04:18,581][starVLA.training.train_starvla][INFO] - Step 185, Loss: {'train/loss': 1.0516234636306763, 'train/grad_norm_pre_clip': 11.143449783325195, 'timing/data': 0.027123664505779743, 'timing/model': 0.9285811325535178, 'learning_rate/qwen_vl_interface': 9.996443286992337e-06, 'learning_rate/action_model': 9.996443286992336e-05, 'epoch': 0.08}) +[2026-06-17 17:04:22,420][starVLA.training.train_starvla][INFO] - Step 186, Loss: {'train/loss': 0.9858307838439941, 'train/grad_norm_pre_clip': 11.373809814453125, 'timing/data': 0.026796751655638218, 'timing/model': 0.9294761950150132, 'learning_rate/qwen_vl_interface': 9.996359118104432e-06, 'learning_rate/action_model': 9.996359118104432e-05, 'epoch': 0.08}) +[2026-06-17 17:04:26,217][starVLA.training.train_starvla][INFO] - Step 187, Loss: {'train/loss': 1.0798851251602173, 'train/grad_norm_pre_clip': 3.1296162605285645, 'timing/data': 0.02192532643675804, 'timing/model': 0.9239545352756977, 'learning_rate/qwen_vl_interface': 9.996273965291177e-06, 'learning_rate/action_model': 9.996273965291176e-05, 'epoch': 0.08}) +[2026-06-17 17:04:30,027][starVLA.training.train_starvla][INFO] - Step 188, Loss: {'train/loss': 0.9740068316459656, 'train/grad_norm_pre_clip': 4.203335285186768, 'timing/data': 0.03337189648300409, 'timing/model': 0.9200144968926907, 'learning_rate/qwen_vl_interface': 9.996187828570221e-06, 'learning_rate/action_model': 9.996187828570221e-05, 'epoch': 0.09}) +[2026-06-17 17:04:33,850][starVLA.training.train_starvla][INFO] - Step 189, Loss: {'train/loss': 1.0663830041885376, 'train/grad_norm_pre_clip': 13.765149116516113, 'timing/data': 0.0463921008631587, 'timing/model': 0.9390733111649752, 'learning_rate/qwen_vl_interface': 9.996100707959422e-06, 'learning_rate/action_model': 9.996100707959421e-05, 'epoch': 0.09}) +[2026-06-17 17:04:37,691][starVLA.training.train_starvla][INFO] - Step 190, Loss: {'train/loss': 0.9330272078514099, 'train/grad_norm_pre_clip': 4.993135929107666, 'timing/data': 0.028146673925220966, 'timing/model': 1.161330934613943, 'learning_rate/qwen_vl_interface': 9.996012603476836e-06, 'learning_rate/action_model': 9.996012603476837e-05, 'epoch': 0.09}) +[2026-06-17 17:04:41,426][starVLA.training.train_starvla][INFO] - Step 191, Loss: {'train/loss': 0.708927571773529, 'train/grad_norm_pre_clip': 6.004505157470703, 'timing/data': 0.023916607722640038, 'timing/model': 1.1286058844998479, 'learning_rate/qwen_vl_interface': 9.995923515140734e-06, 'learning_rate/action_model': 9.995923515140732e-05, 'epoch': 0.09}) +[2026-06-17 17:04:44,953][starVLA.training.train_starvla][INFO] - Step 192, Loss: {'train/loss': 1.2340742349624634, 'train/grad_norm_pre_clip': 5.845513343811035, 'timing/data': 0.03238625451922417, 'timing/model': 0.9107435578480363, 'learning_rate/qwen_vl_interface': 9.995833442969578e-06, 'learning_rate/action_model': 9.995833442969578e-05, 'epoch': 0.09}) +[2026-06-17 17:04:48,658][starVLA.training.train_starvla][INFO] - Step 193, Loss: {'train/loss': 1.0340029001235962, 'train/grad_norm_pre_clip': 2.892679452896118, 'timing/data': 0.02737837564200163, 'timing/model': 0.916898213326931, 'learning_rate/qwen_vl_interface': 9.995742386982045e-06, 'learning_rate/action_model': 9.995742386982045e-05, 'epoch': 0.09}) +[2026-06-17 17:04:52,381][starVLA.training.train_starvla][INFO] - Step 194, Loss: {'train/loss': 1.005603551864624, 'train/grad_norm_pre_clip': 4.920250415802002, 'timing/data': 0.031930577009916306, 'timing/model': 0.9182883445173502, 'learning_rate/qwen_vl_interface': 9.995650347197006e-06, 'learning_rate/action_model': 9.995650347197005e-05, 'epoch': 0.09}) +[2026-06-17 17:04:56,165][starVLA.training.train_starvla][INFO] - Step 195, Loss: {'train/loss': 0.6058819890022278, 'train/grad_norm_pre_clip': 3.7053372859954834, 'timing/data': 0.02889227494597435, 'timing/model': 0.9271109970286489, 'learning_rate/qwen_vl_interface': 9.995557323633544e-06, 'learning_rate/action_model': 9.995557323633543e-05, 'epoch': 0.09}) +[2026-06-17 17:05:00,088][starVLA.training.train_starvla][INFO] - Step 196, Loss: {'train/loss': 0.531725287437439, 'train/grad_norm_pre_clip': 1.634358286857605, 'timing/data': 0.028083345852792263, 'timing/model': 0.9320190418511629, 'learning_rate/qwen_vl_interface': 9.995463316310942e-06, 'learning_rate/action_model': 9.995463316310943e-05, 'epoch': 0.09}) +[2026-06-17 17:05:03,975][starVLA.training.train_starvla][INFO] - Step 197, Loss: {'train/loss': 1.2750605344772339, 'train/grad_norm_pre_clip': 5.508572578430176, 'timing/data': 0.026659896597266197, 'timing/model': 1.194590455852449, 'learning_rate/qwen_vl_interface': 9.995368325248688e-06, 'learning_rate/action_model': 9.995368325248688e-05, 'epoch': 0.09}) +[2026-06-17 17:05:07,655][starVLA.training.train_starvla][INFO] - Step 198, Loss: {'train/loss': 1.0760148763656616, 'train/grad_norm_pre_clip': 5.406312465667725, 'timing/data': 0.025407692417502403, 'timing/model': 0.9309107307344675, 'learning_rate/qwen_vl_interface': 9.995272350466476e-06, 'learning_rate/action_model': 9.995272350466476e-05, 'epoch': 0.09}) +[2026-06-17 17:05:11,482][starVLA.training.train_starvla][INFO] - Step 199, Loss: {'train/loss': 1.2119426727294922, 'train/grad_norm_pre_clip': 4.651337623596191, 'timing/data': 0.02840764168649912, 'timing/model': 0.9368569133803248, 'learning_rate/qwen_vl_interface': 9.995175391984199e-06, 'learning_rate/action_model': 9.995175391984198e-05, 'epoch': 0.09}) +[2026-06-17 17:07:15,184][starVLA.training.train_starvla][INFO] - Step 200, Loss: {'train/loss': 1.2651450634002686, 'train/grad_norm_pre_clip': 6.045853137969971, 'eval/loss': 0.9264984130859375, 'eval/latency_8/loss': 0.9264984130859375, 'eval/demon_fire_precision': 0.8956594323873122, 'eval/demon_fire_recall': 1.0, 'eval/demon_fire_f1': 0.944958168207838, 'eval/demon_fire_rate_model': 1.0, 'eval/demon_fire_rate_teacher': 0.8956594323873122, 'eval/demon_left_precision': 0.6649214659685864, 'eval/demon_left_recall': 0.27194860813704497, 'eval/demon_left_f1': 0.3860182370820669, 'eval/demon_left_rate_model': 0.053144129104062326, 'eval/demon_left_rate_teacher': 0.12993878686700056, 'eval/demon_right_precision': 0.8487084870848709, 'eval/demon_right_recall': 0.3484848484848485, 'eval/demon_right_f1': 0.49409237379162185, 'eval/demon_right_rate_model': 0.07540345019476906, 'eval/demon_right_rate_teacher': 0.18363939899833054, 'eval/demon_move_f1': 0.44005530543684435, 'eval/demon_cc_f1': 0.702984720633999, 'eval/latency_8/demon_fire_precision': 0.8956594323873122, 'eval/latency_8/demon_fire_recall': 1.0, 'eval/latency_8/demon_fire_f1': 0.944958168207838, 'eval/latency_8/demon_fire_rate_model': 1.0, 'eval/latency_8/demon_fire_rate_teacher': 0.8956594323873122, 'eval/latency_8/demon_left_precision': 0.6649214659685864, 'eval/latency_8/demon_left_recall': 0.27194860813704497, 'eval/latency_8/demon_left_f1': 0.3860182370820669, 'eval/latency_8/demon_left_rate_model': 0.053144129104062326, 'eval/latency_8/demon_left_rate_teacher': 0.12993878686700056, 'eval/latency_8/demon_right_precision': 0.8487084870848709, 'eval/latency_8/demon_right_recall': 0.3484848484848485, 'eval/latency_8/demon_right_f1': 0.49409237379162185, 'eval/latency_8/demon_right_rate_model': 0.07540345019476906, 'eval/latency_8/demon_right_rate_teacher': 0.18363939899833054, 'eval/latency_8/demon_move_f1': 0.44005530543684435, 'eval/latency_8/demon_cc_f1': 0.702984720633999, 'timing/data': 0.03200296685099602, 'timing/model': 0.935087195597589, 'learning_rate/qwen_vl_interface': 9.995077449821957e-06, 'learning_rate/action_model': 9.995077449821956e-05, 'epoch': 0.09}) +[2026-06-17 17:07:15,188][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_200_state +[2026-06-17 17:07:15,193][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 17:07:21,559][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_200_state/model.safetensors +[2026-06-17 17:07:36,622][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_200_state/optimizer.bin +[2026-06-17 17:07:36,623][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_200_state/sampler.bin +[2026-06-17 17:07:36,624][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_200_state/sampler_1.bin +[2026-06-17 17:07:36,626][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_200_state/random_states_0.pkl +[2026-06-17 17:07:36,633][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 200 +[2026-06-17 17:07:36,633][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 17:07:36,644][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 17:07:40,264][starVLA.training.train_starvla][INFO] - Step 201, Loss: {'train/loss': 0.651111364364624, 'train/grad_norm_pre_clip': 3.2719240188598633, 'timing/data': 0.03091495670378208, 'timing/model': 0.9511235440149903, 'learning_rate/qwen_vl_interface': 9.994978524000053e-06, 'learning_rate/action_model': 9.994978524000052e-05, 'epoch': 0.09}) +[2026-06-17 17:07:44,069][starVLA.training.train_starvla][INFO] - Step 202, Loss: {'train/loss': 0.9694337248802185, 'train/grad_norm_pre_clip': 3.6837854385375977, 'timing/data': 0.035794852301478386, 'timing/model': 0.9194798776879907, 'learning_rate/qwen_vl_interface': 9.994878614538996e-06, 'learning_rate/action_model': 9.994878614538996e-05, 'epoch': 0.09}) +[2026-06-17 17:07:47,811][starVLA.training.train_starvla][INFO] - Step 203, Loss: {'train/loss': 0.6917529702186584, 'train/grad_norm_pre_clip': 3.087275266647339, 'timing/data': 0.03302135318517685, 'timing/model': 0.9138309778645635, 'learning_rate/qwen_vl_interface': 9.994777721459496e-06, 'learning_rate/action_model': 9.994777721459497e-05, 'epoch': 0.09}) +[2026-06-17 17:07:51,657][starVLA.training.train_starvla][INFO] - Step 204, Loss: {'train/loss': 0.7833438515663147, 'train/grad_norm_pre_clip': 2.8246026039123535, 'timing/data': 0.03631255961954594, 'timing/model': 0.9283434422686696, 'learning_rate/qwen_vl_interface': 9.994675844782468e-06, 'learning_rate/action_model': 9.994675844782468e-05, 'epoch': 0.09}) +[2026-06-17 17:07:55,475][starVLA.training.train_starvla][INFO] - Step 205, Loss: {'train/loss': 0.9543095827102661, 'train/grad_norm_pre_clip': 8.690110206604004, 'timing/data': 0.0257629556581378, 'timing/model': 0.9154878137633204, 'learning_rate/qwen_vl_interface': 9.994572984529033e-06, 'learning_rate/action_model': 9.994572984529034e-05, 'epoch': 0.09}) +[2026-06-17 17:07:59,280][starVLA.training.train_starvla][INFO] - Step 206, Loss: {'train/loss': 0.7837172746658325, 'train/grad_norm_pre_clip': 1.9642490148544312, 'timing/data': 0.027647619135677814, 'timing/model': 1.1475174631923437, 'learning_rate/qwen_vl_interface': 9.994469140720513e-06, 'learning_rate/action_model': 9.994469140720512e-05, 'epoch': 0.09}) +[2026-06-17 17:08:02,891][starVLA.training.train_starvla][INFO] - Step 207, Loss: {'train/loss': 0.5989338159561157, 'train/grad_norm_pre_clip': 2.382883071899414, 'timing/data': 0.036568897776305676, 'timing/model': 0.913650419563055, 'learning_rate/qwen_vl_interface': 9.994364313378435e-06, 'learning_rate/action_model': 9.994364313378433e-05, 'epoch': 0.09}) +[2026-06-17 17:08:06,880][starVLA.training.train_starvla][INFO] - Step 208, Loss: {'train/loss': 1.2258434295654297, 'train/grad_norm_pre_clip': 4.0157060623168945, 'timing/data': 0.034316795878112316, 'timing/model': 0.94995994027704, 'learning_rate/qwen_vl_interface': 9.99425850252453e-06, 'learning_rate/action_model': 9.994258502524529e-05, 'epoch': 0.09}) +[2026-06-17 17:08:10,910][starVLA.training.train_starvla][INFO] - Step 209, Loss: {'train/loss': 0.6841045618057251, 'train/grad_norm_pre_clip': 1.9965896606445312, 'timing/data': 0.036977252922952175, 'timing/model': 0.9211498145014048, 'learning_rate/qwen_vl_interface': 9.994151708180731e-06, 'learning_rate/action_model': 9.99415170818073e-05, 'epoch': 0.09}) +[2026-06-17 17:08:14,781][starVLA.training.train_starvla][INFO] - Step 210, Loss: {'train/loss': 0.7817380428314209, 'train/grad_norm_pre_clip': 2.6032168865203857, 'timing/data': 0.03742763493210077, 'timing/model': 0.9731526747345924, 'learning_rate/qwen_vl_interface': 9.994043930369179e-06, 'learning_rate/action_model': 9.994043930369178e-05, 'epoch': 0.1}) +[2026-06-17 17:08:18,574][starVLA.training.train_starvla][INFO] - Step 211, Loss: {'train/loss': 0.7410285472869873, 'train/grad_norm_pre_clip': 3.068448543548584, 'timing/data': 0.03555453196167946, 'timing/model': 0.9804745353758335, 'learning_rate/qwen_vl_interface': 9.993935169112213e-06, 'learning_rate/action_model': 9.993935169112213e-05, 'epoch': 0.1}) +[2026-06-17 17:08:22,509][starVLA.training.train_starvla][INFO] - Step 212, Loss: {'train/loss': 0.7888737916946411, 'train/grad_norm_pre_clip': 3.2092151641845703, 'timing/data': 0.026248915120959282, 'timing/model': 0.9228303423151374, 'learning_rate/qwen_vl_interface': 9.993825424432385e-06, 'learning_rate/action_model': 9.993825424432385e-05, 'epoch': 0.1}) +[2026-06-17 17:08:26,254][starVLA.training.train_starvla][INFO] - Step 213, Loss: {'train/loss': 0.7341142892837524, 'train/grad_norm_pre_clip': 2.513472318649292, 'timing/data': 0.02309218980371952, 'timing/model': 1.0870754653587937, 'learning_rate/qwen_vl_interface': 9.99371469635244e-06, 'learning_rate/action_model': 9.99371469635244e-05, 'epoch': 0.1}) +[2026-06-17 17:08:29,983][starVLA.training.train_starvla][INFO] - Step 214, Loss: {'train/loss': 0.9426327347755432, 'train/grad_norm_pre_clip': 3.4478139877319336, 'timing/data': 0.033804855309426785, 'timing/model': 0.9130524843931198, 'learning_rate/qwen_vl_interface': 9.993602984895336e-06, 'learning_rate/action_model': 9.993602984895335e-05, 'epoch': 0.1}) +[2026-06-17 17:08:33,855][starVLA.training.train_starvla][INFO] - Step 215, Loss: {'train/loss': 0.9493904709815979, 'train/grad_norm_pre_clip': 3.7310523986816406, 'timing/data': 0.03239962924271822, 'timing/model': 0.9141982588917017, 'learning_rate/qwen_vl_interface': 9.993490290084225e-06, 'learning_rate/action_model': 9.993490290084225e-05, 'epoch': 0.1}) +[2026-06-17 17:08:37,665][starVLA.training.train_starvla][INFO] - Step 216, Loss: {'train/loss': 0.7409861087799072, 'train/grad_norm_pre_clip': 2.678138256072998, 'timing/data': 0.03272071201354265, 'timing/model': 0.9859361909329891, 'learning_rate/qwen_vl_interface': 9.993376611942477e-06, 'learning_rate/action_model': 9.993376611942477e-05, 'epoch': 0.1}) +[2026-06-17 17:08:41,670][starVLA.training.train_starvla][INFO] - Step 217, Loss: {'train/loss': 0.853801429271698, 'train/grad_norm_pre_clip': 3.753117084503174, 'timing/data': 0.030469301156699657, 'timing/model': 1.165960387326777, 'learning_rate/qwen_vl_interface': 9.99326195049365e-06, 'learning_rate/action_model': 9.99326195049365e-05, 'epoch': 0.1}) +[2026-06-17 17:08:46,884][starVLA.training.train_starvla][INFO] - Step 218, Loss: {'train/loss': 1.1362992525100708, 'train/grad_norm_pre_clip': 4.525938034057617, 'timing/data': 0.02250234130769968, 'timing/model': 1.0618868498131633, 'learning_rate/qwen_vl_interface': 9.99314630576152e-06, 'learning_rate/action_model': 9.993146305761519e-05, 'epoch': 0.1}) +[2026-06-17 17:08:50,719][starVLA.training.train_starvla][INFO] - Step 219, Loss: {'train/loss': 0.8866468667984009, 'train/grad_norm_pre_clip': 5.574400901794434, 'timing/data': 0.026005552150309086, 'timing/model': 1.1269537834450603, 'learning_rate/qwen_vl_interface': 9.993029677770055e-06, 'learning_rate/action_model': 9.993029677770055e-05, 'epoch': 0.1}) +[2026-06-17 17:08:54,534][starVLA.training.train_starvla][INFO] - Step 220, Loss: {'train/loss': 0.5908495187759399, 'train/grad_norm_pre_clip': 3.9045469760894775, 'timing/data': 0.025869054719805717, 'timing/model': 1.1336283748969436, 'learning_rate/qwen_vl_interface': 9.992912066543433e-06, 'learning_rate/action_model': 9.992912066543433e-05, 'epoch': 0.1}) +[2026-06-17 17:08:58,147][starVLA.training.train_starvla][INFO] - Step 221, Loss: {'train/loss': 0.651671290397644, 'train/grad_norm_pre_clip': 3.287043809890747, 'timing/data': 0.037834168411791325, 'timing/model': 0.9287336282432079, 'learning_rate/qwen_vl_interface': 9.992793472106039e-06, 'learning_rate/action_model': 9.992793472106038e-05, 'epoch': 0.1}) +[2026-06-17 17:09:02,059][starVLA.training.train_starvla][INFO] - Step 222, Loss: {'train/loss': 0.8295011520385742, 'train/grad_norm_pre_clip': 3.6087422370910645, 'timing/data': 0.04449065402150154, 'timing/model': 0.9954243479296565, 'learning_rate/qwen_vl_interface': 9.992673894482452e-06, 'learning_rate/action_model': 9.992673894482451e-05, 'epoch': 0.1}) +[2026-06-17 17:09:05,949][starVLA.training.train_starvla][INFO] - Step 223, Loss: {'train/loss': 1.1744277477264404, 'train/grad_norm_pre_clip': 5.717631816864014, 'timing/data': 0.023232958279550076, 'timing/model': 0.9773485995829105, 'learning_rate/qwen_vl_interface': 9.992553333697464e-06, 'learning_rate/action_model': 9.992553333697463e-05, 'epoch': 0.1}) +[2026-06-17 17:09:09,831][starVLA.training.train_starvla][INFO] - Step 224, Loss: {'train/loss': 0.6104741096496582, 'train/grad_norm_pre_clip': 2.85506534576416, 'timing/data': 0.031804571859538555, 'timing/model': 0.9432391310110688, 'learning_rate/qwen_vl_interface': 9.992431789776067e-06, 'learning_rate/action_model': 9.992431789776066e-05, 'epoch': 0.1}) +[2026-06-17 17:09:14,125][starVLA.training.train_starvla][INFO] - Step 225, Loss: {'train/loss': 0.8964409232139587, 'train/grad_norm_pre_clip': 6.554437637329102, 'timing/data': 0.02620566450059414, 'timing/model': 0.9547341903671622, 'learning_rate/qwen_vl_interface': 9.992309262743458e-06, 'learning_rate/action_model': 9.992309262743457e-05, 'epoch': 0.1}) +[2026-06-17 17:09:18,141][starVLA.training.train_starvla][INFO] - Step 226, Loss: {'train/loss': 0.7084860801696777, 'train/grad_norm_pre_clip': 3.138686180114746, 'timing/data': 0.0354560436680913, 'timing/model': 1.1419553086161613, 'learning_rate/qwen_vl_interface': 9.992185752625035e-06, 'learning_rate/action_model': 9.992185752625033e-05, 'epoch': 0.1}) +[2026-06-17 17:09:21,945][starVLA.training.train_starvla][INFO] - Step 227, Loss: {'train/loss': 0.7490381002426147, 'train/grad_norm_pre_clip': 2.4810032844543457, 'timing/data': 0.031499896198511124, 'timing/model': 0.9386561280116439, 'learning_rate/qwen_vl_interface': 9.992061259446402e-06, 'learning_rate/action_model': 9.992061259446401e-05, 'epoch': 0.1}) +[2026-06-17 17:09:29,882][starVLA.training.train_starvla][INFO] - Step 228, Loss: {'train/loss': 1.0076669454574585, 'train/grad_norm_pre_clip': 3.4236011505126953, 'timing/data': 0.039458719082176685, 'timing/model': 2.2989331800490618, 'learning_rate/qwen_vl_interface': 9.991935783233368e-06, 'learning_rate/action_model': 9.991935783233367e-05, 'epoch': 0.1}) +[2026-06-17 17:09:34,115][starVLA.training.train_starvla][INFO] - Step 229, Loss: {'train/loss': 1.2170400619506836, 'train/grad_norm_pre_clip': 6.763067722320557, 'timing/data': 0.029089885763823986, 'timing/model': 0.9522676542401314, 'learning_rate/qwen_vl_interface': 9.991809324011943e-06, 'learning_rate/action_model': 9.991809324011943e-05, 'epoch': 0.1}) +[2026-06-17 17:09:38,124][starVLA.training.train_starvla][INFO] - Step 230, Loss: {'train/loss': 0.9480496644973755, 'train/grad_norm_pre_clip': 10.136597633361816, 'timing/data': 0.03599440213292837, 'timing/model': 0.9448880758136511, 'learning_rate/qwen_vl_interface': 9.991681881808343e-06, 'learning_rate/action_model': 9.991681881808343e-05, 'epoch': 0.1}) +[2026-06-17 17:09:41,940][starVLA.training.train_starvla][INFO] - Step 231, Loss: {'train/loss': 0.6972334384918213, 'train/grad_norm_pre_clip': 4.241311550140381, 'timing/data': 0.029964961111545563, 'timing/model': 0.9245953084900975, 'learning_rate/qwen_vl_interface': 9.991553456648986e-06, 'learning_rate/action_model': 9.991553456648985e-05, 'epoch': 0.1}) +[2026-06-17 17:09:45,742][starVLA.training.train_starvla][INFO] - Step 232, Loss: {'train/loss': 0.6809777021408081, 'train/grad_norm_pre_clip': 6.368584632873535, 'timing/data': 0.021048011258244514, 'timing/model': 1.1356482338160276, 'learning_rate/qwen_vl_interface': 9.991424048560497e-06, 'learning_rate/action_model': 9.991424048560496e-05, 'epoch': 0.11}) +[2026-06-17 17:09:49,559][starVLA.training.train_starvla][INFO] - Step 233, Loss: {'train/loss': 0.938387393951416, 'train/grad_norm_pre_clip': 6.773221969604492, 'timing/data': 0.0340112904086709, 'timing/model': 1.1451853243634105, 'learning_rate/qwen_vl_interface': 9.991293657569698e-06, 'learning_rate/action_model': 9.991293657569698e-05, 'epoch': 0.11}) +[2026-06-17 17:09:53,266][starVLA.training.train_starvla][INFO] - Step 234, Loss: {'train/loss': 0.9827600121498108, 'train/grad_norm_pre_clip': 4.344334602355957, 'timing/data': 0.023797519505023956, 'timing/model': 0.9615388242527843, 'learning_rate/qwen_vl_interface': 9.991162283703624e-06, 'learning_rate/action_model': 9.991162283703624e-05, 'epoch': 0.11}) +[2026-06-17 17:09:57,364][starVLA.training.train_starvla][INFO] - Step 235, Loss: {'train/loss': 1.039889931678772, 'train/grad_norm_pre_clip': 3.4684898853302, 'timing/data': 0.02632494643330574, 'timing/model': 0.9311132654547691, 'learning_rate/qwen_vl_interface': 9.991029926989506e-06, 'learning_rate/action_model': 9.991029926989506e-05, 'epoch': 0.11}) +[2026-06-17 17:10:01,264][starVLA.training.train_starvla][INFO] - Step 236, Loss: {'train/loss': 0.9838354587554932, 'train/grad_norm_pre_clip': 5.617554664611816, 'timing/data': 0.025946469977498055, 'timing/model': 0.9375103311613202, 'learning_rate/qwen_vl_interface': 9.990896587454783e-06, 'learning_rate/action_model': 9.990896587454783e-05, 'epoch': 0.11}) +[2026-06-17 17:10:05,155][starVLA.training.train_starvla][INFO] - Step 237, Loss: {'train/loss': 1.2309257984161377, 'train/grad_norm_pre_clip': 7.665692329406738, 'timing/data': 0.029229438863694668, 'timing/model': 0.9560976615175605, 'learning_rate/qwen_vl_interface': 9.990762265127096e-06, 'learning_rate/action_model': 9.990762265127096e-05, 'epoch': 0.11}) +[2026-06-17 17:10:09,036][starVLA.training.train_starvla][INFO] - Step 238, Loss: {'train/loss': 0.807411789894104, 'train/grad_norm_pre_clip': 4.849579334259033, 'timing/data': 0.031614295206964016, 'timing/model': 0.9513693433254957, 'learning_rate/qwen_vl_interface': 9.990626960034291e-06, 'learning_rate/action_model': 9.990626960034291e-05, 'epoch': 0.11}) +[2026-06-17 17:10:12,943][starVLA.training.train_starvla][INFO] - Step 239, Loss: {'train/loss': 0.865035355091095, 'train/grad_norm_pre_clip': 3.9293267726898193, 'timing/data': 0.031430477276444435, 'timing/model': 1.1767245354130864, 'learning_rate/qwen_vl_interface': 9.990490672204417e-06, 'learning_rate/action_model': 9.990490672204415e-05, 'epoch': 0.11}) +[2026-06-17 17:10:16,628][starVLA.training.train_starvla][INFO] - Step 240, Loss: {'train/loss': 1.0957273244857788, 'train/grad_norm_pre_clip': 6.233311653137207, 'timing/data': 0.024423500522971153, 'timing/model': 0.9289984432980418, 'learning_rate/qwen_vl_interface': 9.99035340166572e-06, 'learning_rate/action_model': 9.990353401665721e-05, 'epoch': 0.11}) +[2026-06-17 17:10:20,617][starVLA.training.train_starvla][INFO] - Step 241, Loss: {'train/loss': 0.8329953551292419, 'train/grad_norm_pre_clip': 6.630755424499512, 'timing/data': 0.026198972947895527, 'timing/model': 0.9310733377933502, 'learning_rate/qwen_vl_interface': 9.990215148446667e-06, 'learning_rate/action_model': 9.990215148446666e-05, 'epoch': 0.11}) +[2026-06-17 17:10:24,533][starVLA.training.train_starvla][INFO] - Step 242, Loss: {'train/loss': 1.1867934465408325, 'train/grad_norm_pre_clip': 6.300502300262451, 'timing/data': 0.0516902981325984, 'timing/model': 0.9906630953773856, 'learning_rate/qwen_vl_interface': 9.990075912575912e-06, 'learning_rate/action_model': 9.990075912575911e-05, 'epoch': 0.11}) +[2026-06-17 17:10:28,370][starVLA.training.train_starvla][INFO] - Step 243, Loss: {'train/loss': 0.9746268391609192, 'train/grad_norm_pre_clip': 4.207128047943115, 'timing/data': 0.030760099180042744, 'timing/model': 0.939868894405663, 'learning_rate/qwen_vl_interface': 9.98993569408232e-06, 'learning_rate/action_model': 9.989935694082321e-05, 'epoch': 0.11}) +[2026-06-17 17:10:32,299][starVLA.training.train_starvla][INFO] - Step 244, Loss: {'train/loss': 1.0374490022659302, 'train/grad_norm_pre_clip': 4.48752498626709, 'timing/data': 0.03133358992636204, 'timing/model': 1.0595298511907458, 'learning_rate/qwen_vl_interface': 9.989794492994958e-06, 'learning_rate/action_model': 9.989794492994958e-05, 'epoch': 0.11}) +[2026-06-17 17:10:36,043][starVLA.training.train_starvla][INFO] - Step 245, Loss: {'train/loss': 1.0009969472885132, 'train/grad_norm_pre_clip': 3.869844436645508, 'timing/data': 0.031207533553242683, 'timing/model': 0.9293464664369822, 'learning_rate/qwen_vl_interface': 9.989652309343097e-06, 'learning_rate/action_model': 9.989652309343098e-05, 'epoch': 0.11}) +[2026-06-17 17:10:39,906][starVLA.training.train_starvla][INFO] - Step 246, Loss: {'train/loss': 0.7981864809989929, 'train/grad_norm_pre_clip': 3.420990228652954, 'timing/data': 0.046630519442260265, 'timing/model': 1.167687065899372, 'learning_rate/qwen_vl_interface': 9.989509143156213e-06, 'learning_rate/action_model': 9.989509143156213e-05, 'epoch': 0.11}) +[2026-06-17 17:10:43,801][starVLA.training.train_starvla][INFO] - Step 247, Loss: {'train/loss': 0.9583539366722107, 'train/grad_norm_pre_clip': 6.659523963928223, 'timing/data': 0.025459157302975655, 'timing/model': 1.2713823737576604, 'learning_rate/qwen_vl_interface': 9.989364994463983e-06, 'learning_rate/action_model': 9.989364994463983e-05, 'epoch': 0.11}) +[2026-06-17 17:10:47,621][starVLA.training.train_starvla][INFO] - Step 248, Loss: {'train/loss': 0.9666725993156433, 'train/grad_norm_pre_clip': 8.975481986999512, 'timing/data': 0.023687773384153843, 'timing/model': 0.9559865156188607, 'learning_rate/qwen_vl_interface': 9.989219863296291e-06, 'learning_rate/action_model': 9.989219863296291e-05, 'epoch': 0.11}) +[2026-06-17 17:10:51,347][starVLA.training.train_starvla][INFO] - Step 249, Loss: {'train/loss': 0.9343419075012207, 'train/grad_norm_pre_clip': 5.520371437072754, 'timing/data': 0.02850277628749609, 'timing/model': 0.9334230124950409, 'learning_rate/qwen_vl_interface': 9.98907374968322e-06, 'learning_rate/action_model': 9.989073749683219e-05, 'epoch': 0.11}) +[2026-06-17 17:10:55,145][starVLA.training.train_starvla][INFO] - Step 250, Loss: {'train/loss': 0.9442943334579468, 'train/grad_norm_pre_clip': 5.8508219718933105, 'timing/data': 0.02979801967740059, 'timing/model': 0.9447812624275684, 'learning_rate/qwen_vl_interface': 9.988926653655062e-06, 'learning_rate/action_model': 9.988926653655062e-05, 'epoch': 0.11}) +[2026-06-17 17:10:58,951][starVLA.training.train_starvla][INFO] - Step 251, Loss: {'train/loss': 1.032581090927124, 'train/grad_norm_pre_clip': 4.334722995758057, 'timing/data': 0.022714409977197647, 'timing/model': 0.9252245016396046, 'learning_rate/qwen_vl_interface': 9.988778575242309e-06, 'learning_rate/action_model': 9.988778575242309e-05, 'epoch': 0.11}) +[2026-06-17 17:11:02,924][starVLA.training.train_starvla][INFO] - Step 252, Loss: {'train/loss': 0.5927929282188416, 'train/grad_norm_pre_clip': 1.566695213317871, 'timing/data': 0.026648794300854206, 'timing/model': 1.2451399993151426, 'learning_rate/qwen_vl_interface': 9.98862951447566e-06, 'learning_rate/action_model': 9.988629514475659e-05, 'epoch': 0.11}) +[2026-06-17 17:11:06,794][starVLA.training.train_starvla][INFO] - Step 253, Loss: {'train/loss': 0.7694016098976135, 'train/grad_norm_pre_clip': 3.366910457611084, 'timing/data': 0.03068669233471155, 'timing/model': 1.1868381379172206, 'learning_rate/qwen_vl_interface': 9.98847947138601e-06, 'learning_rate/action_model': 9.98847947138601e-05, 'epoch': 0.11}) +[2026-06-17 17:11:10,733][starVLA.training.train_starvla][INFO] - Step 254, Loss: {'train/loss': 0.9909772276878357, 'train/grad_norm_pre_clip': 4.596222400665283, 'timing/data': 0.032789913937449455, 'timing/model': 1.022451457567513, 'learning_rate/qwen_vl_interface': 9.98832844600447e-06, 'learning_rate/action_model': 9.98832844600447e-05, 'epoch': 0.12}) +[2026-06-17 17:11:14,551][starVLA.training.train_starvla][INFO] - Step 255, Loss: {'train/loss': 0.6839221715927124, 'train/grad_norm_pre_clip': 4.358016014099121, 'timing/data': 0.024611156433820724, 'timing/model': 0.9347417699173093, 'learning_rate/qwen_vl_interface': 9.988176438362344e-06, 'learning_rate/action_model': 9.988176438362343e-05, 'epoch': 0.12}) +[2026-06-17 17:11:19,247][starVLA.training.train_starvla][INFO] - Step 256, Loss: {'train/loss': 0.966532826423645, 'train/grad_norm_pre_clip': 4.695639610290527, 'timing/data': 0.023514882661402225, 'timing/model': 0.9735558656975627, 'learning_rate/qwen_vl_interface': 9.988023448491143e-06, 'learning_rate/action_model': 9.988023448491143e-05, 'epoch': 0.12}) +[2026-06-17 17:11:23,099][starVLA.training.train_starvla][INFO] - Step 257, Loss: {'train/loss': 0.964828610420227, 'train/grad_norm_pre_clip': 3.987260580062866, 'timing/data': 0.0215574000030756, 'timing/model': 0.9428414702415466, 'learning_rate/qwen_vl_interface': 9.987869476422582e-06, 'learning_rate/action_model': 9.987869476422583e-05, 'epoch': 0.12}) +[2026-06-17 17:11:27,077][starVLA.training.train_starvla][INFO] - Step 258, Loss: {'train/loss': 0.9411910772323608, 'train/grad_norm_pre_clip': 3.004032611846924, 'timing/data': 0.03883321676403284, 'timing/model': 1.0937496712431312, 'learning_rate/qwen_vl_interface': 9.987714522188583e-06, 'learning_rate/action_model': 9.987714522188582e-05, 'epoch': 0.12}) +[2026-06-17 17:11:31,002][starVLA.training.train_starvla][INFO] - Step 259, Loss: {'train/loss': 1.2808388471603394, 'train/grad_norm_pre_clip': 3.1141200065612793, 'timing/data': 0.031801595352590084, 'timing/model': 1.1443678019568324, 'learning_rate/qwen_vl_interface': 9.987558585821264e-06, 'learning_rate/action_model': 9.987558585821263e-05, 'epoch': 0.12}) +[2026-06-17 17:11:34,814][starVLA.training.train_starvla][INFO] - Step 260, Loss: {'train/loss': 0.6085233688354492, 'train/grad_norm_pre_clip': 4.173033714294434, 'timing/data': 0.02481565997004509, 'timing/model': 1.068015893921256, 'learning_rate/qwen_vl_interface': 9.987401667352952e-06, 'learning_rate/action_model': 9.987401667352951e-05, 'epoch': 0.12}) +[2026-06-17 17:11:38,431][starVLA.training.train_starvla][INFO] - Step 261, Loss: {'train/loss': 0.7725027799606323, 'train/grad_norm_pre_clip': 3.6265504360198975, 'timing/data': 0.027642236091196537, 'timing/model': 0.9379025101661682, 'learning_rate/qwen_vl_interface': 9.987243766816175e-06, 'learning_rate/action_model': 9.987243766816176e-05, 'epoch': 0.12}) +[2026-06-17 17:11:42,292][starVLA.training.train_starvla][INFO] - Step 262, Loss: {'train/loss': 1.0114773511886597, 'train/grad_norm_pre_clip': 4.147220134735107, 'timing/data': 0.02766015101224184, 'timing/model': 0.9351179441437125, 'learning_rate/qwen_vl_interface': 9.987084884243671e-06, 'learning_rate/action_model': 9.987084884243672e-05, 'epoch': 0.12}) +[2026-06-17 17:11:46,529][starVLA.training.train_starvla][INFO] - Step 263, Loss: {'train/loss': 1.0052627325057983, 'train/grad_norm_pre_clip': 7.752010345458984, 'timing/data': 0.03427985589951277, 'timing/model': 0.9157214164733887, 'learning_rate/qwen_vl_interface': 9.986925019668374e-06, 'learning_rate/action_model': 9.986925019668373e-05, 'epoch': 0.12}) +[2026-06-17 17:11:50,227][starVLA.training.train_starvla][INFO] - Step 264, Loss: {'train/loss': 1.0456193685531616, 'train/grad_norm_pre_clip': 4.381565093994141, 'timing/data': 0.0269120829179883, 'timing/model': 0.9151296820491552, 'learning_rate/qwen_vl_interface': 9.98676417312342e-06, 'learning_rate/action_model': 9.986764173123419e-05, 'epoch': 0.12}) +[2026-06-17 17:11:54,010][starVLA.training.train_starvla][INFO] - Step 265, Loss: {'train/loss': 0.9833176732063293, 'train/grad_norm_pre_clip': 7.7460126876831055, 'timing/data': 0.026506458409130573, 'timing/model': 0.9293711483478546, 'learning_rate/qwen_vl_interface': 9.986602344642157e-06, 'learning_rate/action_model': 9.986602344642156e-05, 'epoch': 0.12}) +[2026-06-17 17:11:58,093][starVLA.training.train_starvla][INFO] - Step 266, Loss: {'train/loss': 0.6226627230644226, 'train/grad_norm_pre_clip': 4.332113742828369, 'timing/data': 0.03201712295413017, 'timing/model': 1.1630975939333439, 'learning_rate/qwen_vl_interface': 9.986439534258132e-06, 'learning_rate/action_model': 9.986439534258131e-05, 'epoch': 0.12}) +[2026-06-17 17:12:01,691][starVLA.training.train_starvla][INFO] - Step 267, Loss: {'train/loss': 0.6667315363883972, 'train/grad_norm_pre_clip': 2.2532942295074463, 'timing/data': 0.02317868359386921, 'timing/model': 0.954698477871716, 'learning_rate/qwen_vl_interface': 9.986275742005093e-06, 'learning_rate/action_model': 9.986275742005093e-05, 'epoch': 0.12}) +[2026-06-17 17:12:05,601][starVLA.training.train_starvla][INFO] - Step 268, Loss: {'train/loss': 0.8512907028198242, 'train/grad_norm_pre_clip': 3.441711187362671, 'timing/data': 0.022877476178109646, 'timing/model': 1.039456557482481, 'learning_rate/qwen_vl_interface': 9.986110967916997e-06, 'learning_rate/action_model': 9.986110967916996e-05, 'epoch': 0.12}) +[2026-06-17 17:12:09,373][starVLA.training.train_starvla][INFO] - Step 269, Loss: {'train/loss': 1.1458810567855835, 'train/grad_norm_pre_clip': 7.176022529602051, 'timing/data': 0.02856258861720562, 'timing/model': 0.9174938974902034, 'learning_rate/qwen_vl_interface': 9.985945212028002e-06, 'learning_rate/action_model': 9.985945212028001e-05, 'epoch': 0.12}) +[2026-06-17 17:12:13,351][starVLA.training.train_starvla][INFO] - Step 270, Loss: {'train/loss': 0.9204984307289124, 'train/grad_norm_pre_clip': 5.065110683441162, 'timing/data': 0.028510434553027153, 'timing/model': 0.9203806230798364, 'learning_rate/qwen_vl_interface': 9.985778474372467e-06, 'learning_rate/action_model': 9.985778474372466e-05, 'epoch': 0.12}) +[2026-06-17 17:12:17,076][starVLA.training.train_starvla][INFO] - Step 271, Loss: {'train/loss': 0.8603728413581848, 'train/grad_norm_pre_clip': 5.15943717956543, 'timing/data': 0.03490919526666403, 'timing/model': 0.9147714702412486, 'learning_rate/qwen_vl_interface': 9.98561075498496e-06, 'learning_rate/action_model': 9.98561075498496e-05, 'epoch': 0.12}) +[2026-06-17 17:12:20,860][starVLA.training.train_starvla][INFO] - Step 272, Loss: {'train/loss': 0.7173033356666565, 'train/grad_norm_pre_clip': 4.244190692901611, 'timing/data': 0.03843666147440672, 'timing/model': 1.1371709601953626, 'learning_rate/qwen_vl_interface': 9.985442053900246e-06, 'learning_rate/action_model': 9.985442053900246e-05, 'epoch': 0.12}) +[2026-06-17 17:12:24,746][starVLA.training.train_starvla][INFO] - Step 273, Loss: {'train/loss': 0.9056156873703003, 'train/grad_norm_pre_clip': 3.571042537689209, 'timing/data': 0.02518503461033106, 'timing/model': 1.052925449796021, 'learning_rate/qwen_vl_interface': 9.9852723711533e-06, 'learning_rate/action_model': 9.9852723711533e-05, 'epoch': 0.12}) +[2026-06-17 17:12:28,378][starVLA.training.train_starvla][INFO] - Step 274, Loss: {'train/loss': 0.6736643314361572, 'train/grad_norm_pre_clip': 4.16842794418335, 'timing/data': 0.026495429687201977, 'timing/model': 0.9281013440340757, 'learning_rate/qwen_vl_interface': 9.985101706779297e-06, 'learning_rate/action_model': 9.985101706779296e-05, 'epoch': 0.12}) +[2026-06-17 17:12:32,475][starVLA.training.train_starvla][INFO] - Step 275, Loss: {'train/loss': 0.6932697296142578, 'train/grad_norm_pre_clip': 4.567576885223389, 'timing/data': 0.029150795191526413, 'timing/model': 0.969568750821054, 'learning_rate/qwen_vl_interface': 9.984930060813613e-06, 'learning_rate/action_model': 9.984930060813613e-05, 'epoch': 0.12}) +[2026-06-17 17:12:36,462][starVLA.training.train_starvla][INFO] - Step 276, Loss: {'train/loss': 1.2169214487075806, 'train/grad_norm_pre_clip': 5.930419445037842, 'timing/data': 0.03137199021875858, 'timing/model': 1.0598305892199278, 'learning_rate/qwen_vl_interface': 9.984757433291832e-06, 'learning_rate/action_model': 9.984757433291832e-05, 'epoch': 0.13}) +[2026-06-17 17:12:40,338][starVLA.training.train_starvla][INFO] - Step 277, Loss: {'train/loss': 0.7862438559532166, 'train/grad_norm_pre_clip': 6.230163097381592, 'timing/data': 0.02836125437170267, 'timing/model': 0.9239116031676531, 'learning_rate/qwen_vl_interface': 9.984583824249742e-06, 'learning_rate/action_model': 9.984583824249742e-05, 'epoch': 0.13}) +[2026-06-17 17:12:44,177][starVLA.training.train_starvla][INFO] - Step 278, Loss: {'train/loss': 1.190360426902771, 'train/grad_norm_pre_clip': 3.7393181324005127, 'timing/data': 0.033901472575962543, 'timing/model': 0.9313365630805492, 'learning_rate/qwen_vl_interface': 9.98440923372333e-06, 'learning_rate/action_model': 9.984409233723328e-05, 'epoch': 0.13}) +[2026-06-17 17:12:48,156][starVLA.training.train_starvla][INFO] - Step 279, Loss: {'train/loss': 0.7879639267921448, 'train/grad_norm_pre_clip': 3.6147053241729736, 'timing/data': 0.03616566862910986, 'timing/model': 1.1270244922488928, 'learning_rate/qwen_vl_interface': 9.984233661748788e-06, 'learning_rate/action_model': 9.984233661748788e-05, 'epoch': 0.13}) +[2026-06-17 17:12:51,967][starVLA.training.train_starvla][INFO] - Step 280, Loss: {'train/loss': 0.8895232677459717, 'train/grad_norm_pre_clip': 4.752045154571533, 'timing/data': 0.036155262030661106, 'timing/model': 1.1049402058124542, 'learning_rate/qwen_vl_interface': 9.984057108362516e-06, 'learning_rate/action_model': 9.984057108362515e-05, 'epoch': 0.13}) +[2026-06-17 17:12:55,555][starVLA.training.train_starvla][INFO] - Step 281, Loss: {'train/loss': 0.733230710029602, 'train/grad_norm_pre_clip': 6.2727203369140625, 'timing/data': 0.025171847082674503, 'timing/model': 0.9258027011528611, 'learning_rate/qwen_vl_interface': 9.98387957360111e-06, 'learning_rate/action_model': 9.98387957360111e-05, 'epoch': 0.13}) +[2026-06-17 17:12:59,350][starVLA.training.train_starvla][INFO] - Step 282, Loss: {'train/loss': 1.1520531177520752, 'train/grad_norm_pre_clip': 4.345653057098389, 'timing/data': 0.03127405606210232, 'timing/model': 0.9279010379686952, 'learning_rate/qwen_vl_interface': 9.983701057501375e-06, 'learning_rate/action_model': 9.983701057501375e-05, 'epoch': 0.13}) +[2026-06-17 17:13:03,349][starVLA.training.train_starvla][INFO] - Step 283, Loss: {'train/loss': 1.2243098020553589, 'train/grad_norm_pre_clip': 6.267591953277588, 'timing/data': 0.030564951710402966, 'timing/model': 0.9180944645777345, 'learning_rate/qwen_vl_interface': 9.983521560100317e-06, 'learning_rate/action_model': 9.983521560100316e-05, 'epoch': 0.13}) +[2026-06-17 17:13:07,092][starVLA.training.train_starvla][INFO] - Step 284, Loss: {'train/loss': 0.785282552242279, 'train/grad_norm_pre_clip': 4.030513763427734, 'timing/data': 0.021517991088330746, 'timing/model': 0.9214688241481781, 'learning_rate/qwen_vl_interface': 9.983341081435145e-06, 'learning_rate/action_model': 9.983341081435145e-05, 'epoch': 0.13}) +[2026-06-17 17:13:10,873][starVLA.training.train_starvla][INFO] - Step 285, Loss: {'train/loss': 1.010623574256897, 'train/grad_norm_pre_clip': 3.221442461013794, 'timing/data': 0.03608911205083132, 'timing/model': 0.9310257500037551, 'learning_rate/qwen_vl_interface': 9.983159621543275e-06, 'learning_rate/action_model': 9.983159621543275e-05, 'epoch': 0.13}) +[2026-06-17 17:13:14,710][starVLA.training.train_starvla][INFO] - Step 286, Loss: {'train/loss': 0.9144361019134521, 'train/grad_norm_pre_clip': 4.003036022186279, 'timing/data': 0.023015113547444344, 'timing/model': 1.1964144622907043, 'learning_rate/qwen_vl_interface': 9.982977180462324e-06, 'learning_rate/action_model': 9.982977180462323e-05, 'epoch': 0.13}) +[2026-06-17 17:13:18,451][starVLA.training.train_starvla][INFO] - Step 287, Loss: {'train/loss': 0.7900763154029846, 'train/grad_norm_pre_clip': 3.2234737873077393, 'timing/data': 0.03226288873702288, 'timing/model': 1.0406081462278962, 'learning_rate/qwen_vl_interface': 9.982793758230107e-06, 'learning_rate/action_model': 9.982793758230108e-05, 'epoch': 0.13}) +[2026-06-17 17:13:22,031][starVLA.training.train_starvla][INFO] - Step 288, Loss: {'train/loss': 0.9141961336135864, 'train/grad_norm_pre_clip': 14.036602020263672, 'timing/data': 0.023499086499214172, 'timing/model': 0.929848263040185, 'learning_rate/qwen_vl_interface': 9.982609354884653e-06, 'learning_rate/action_model': 9.982609354884653e-05, 'epoch': 0.13}) +[2026-06-17 17:13:25,775][starVLA.training.train_starvla][INFO] - Step 289, Loss: {'train/loss': 0.988535463809967, 'train/grad_norm_pre_clip': 4.6313676834106445, 'timing/data': 0.02287174016237259, 'timing/model': 0.9276257827877998, 'learning_rate/qwen_vl_interface': 9.982423970464187e-06, 'learning_rate/action_model': 9.982423970464186e-05, 'epoch': 0.13}) +[2026-06-17 17:13:29,810][starVLA.training.train_starvla][INFO] - Step 290, Loss: {'train/loss': 0.6320818066596985, 'train/grad_norm_pre_clip': 3.7309296131134033, 'timing/data': 0.03564706537872553, 'timing/model': 0.9532683501020074, 'learning_rate/qwen_vl_interface': 9.98223760500714e-06, 'learning_rate/action_model': 9.982237605007139e-05, 'epoch': 0.13}) +[2026-06-17 17:13:33,926][starVLA.training.train_starvla][INFO] - Step 291, Loss: {'train/loss': 0.8906859755516052, 'train/grad_norm_pre_clip': 4.682072639465332, 'timing/data': 0.018832235597074032, 'timing/model': 1.2215995835140347, 'learning_rate/qwen_vl_interface': 9.982050258552146e-06, 'learning_rate/action_model': 9.982050258552145e-05, 'epoch': 0.13}) +[2026-06-17 17:13:37,948][starVLA.training.train_starvla][INFO] - Step 292, Loss: {'train/loss': 0.9005151987075806, 'train/grad_norm_pre_clip': 2.40053391456604, 'timing/data': 0.028470100834965706, 'timing/model': 1.2652763845399022, 'learning_rate/qwen_vl_interface': 9.981861931138039e-06, 'learning_rate/action_model': 9.981861931138039e-05, 'epoch': 0.13}) +[2026-06-17 17:13:41,966][starVLA.training.train_starvla][INFO] - Step 293, Loss: {'train/loss': 0.4645010828971863, 'train/grad_norm_pre_clip': 3.7266390323638916, 'timing/data': 0.03497991804033518, 'timing/model': 1.1185078769922256, 'learning_rate/qwen_vl_interface': 9.981672622803863e-06, 'learning_rate/action_model': 9.981672622803862e-05, 'epoch': 0.13}) +[2026-06-17 17:13:45,741][starVLA.training.train_starvla][INFO] - Step 294, Loss: {'train/loss': 0.8811702728271484, 'train/grad_norm_pre_clip': 4.3516082763671875, 'timing/data': 0.04933747462928295, 'timing/model': 0.9458341207355261, 'learning_rate/qwen_vl_interface': 9.981482333588863e-06, 'learning_rate/action_model': 9.981482333588862e-05, 'epoch': 0.13}) +[2026-06-17 17:13:49,842][starVLA.training.train_starvla][INFO] - Step 295, Loss: {'train/loss': 0.8327733874320984, 'train/grad_norm_pre_clip': 3.1621923446655273, 'timing/data': 0.03799679782241583, 'timing/model': 0.9395340131595731, 'learning_rate/qwen_vl_interface': 9.981291063532481e-06, 'learning_rate/action_model': 9.981291063532481e-05, 'epoch': 0.13}) +[2026-06-17 17:13:53,828][starVLA.training.train_starvla][INFO] - Step 296, Loss: {'train/loss': 0.6435325145721436, 'train/grad_norm_pre_clip': 4.883779525756836, 'timing/data': 0.026979735121130943, 'timing/model': 1.0674710031598806, 'learning_rate/qwen_vl_interface': 9.981098812674372e-06, 'learning_rate/action_model': 9.981098812674371e-05, 'epoch': 0.13}) +[2026-06-17 17:13:57,861][starVLA.training.train_starvla][INFO] - Step 297, Loss: {'train/loss': 0.7978837490081787, 'train/grad_norm_pre_clip': 3.1170480251312256, 'timing/data': 0.024845833890140057, 'timing/model': 1.004202926531434, 'learning_rate/qwen_vl_interface': 9.980905581054388e-06, 'learning_rate/action_model': 9.980905581054387e-05, 'epoch': 0.13}) +[2026-06-17 17:14:01,723][starVLA.training.train_starvla][INFO] - Step 298, Loss: {'train/loss': 0.6835452914237976, 'train/grad_norm_pre_clip': 4.703733444213867, 'timing/data': 0.03304796479642391, 'timing/model': 0.9652825174853206, 'learning_rate/qwen_vl_interface': 9.980711368712586e-06, 'learning_rate/action_model': 9.980711368712585e-05, 'epoch': 0.14}) +[2026-06-17 17:14:05,733][starVLA.training.train_starvla][INFO] - Step 299, Loss: {'train/loss': 1.0061529874801636, 'train/grad_norm_pre_clip': 3.5473716259002686, 'timing/data': 0.0336869265884161, 'timing/model': 1.1707260943949223, 'learning_rate/qwen_vl_interface': 9.980516175689228e-06, 'learning_rate/action_model': 9.980516175689227e-05, 'epoch': 0.14}) +[2026-06-17 17:16:23,948][starVLA.training.train_starvla][INFO] - Step 300, Loss: {'train/loss': 0.8646201491355896, 'train/grad_norm_pre_clip': 4.282968997955322, 'eval/loss': 0.898021399974823, 'eval/latency_8/loss': 0.898021399974823, 'eval/demon_fire_precision': 0.8956594323873122, 'eval/demon_fire_recall': 1.0, 'eval/demon_fire_f1': 0.944958168207838, 'eval/demon_fire_rate_model': 1.0, 'eval/demon_fire_rate_teacher': 0.8956594323873122, 'eval/demon_left_precision': 0.49643705463182897, 'eval/demon_left_recall': 0.4475374732334047, 'eval/demon_left_f1': 0.47072072072072074, 'eval/demon_left_rate_model': 0.11713967723984418, 'eval/demon_left_rate_teacher': 0.12993878686700056, 'eval/demon_right_precision': 0.7635327635327636, 'eval/demon_right_recall': 0.40606060606060607, 'eval/demon_right_f1': 0.5301681503461919, 'eval/demon_right_rate_model': 0.09766277128547579, 'eval/demon_right_rate_teacher': 0.18363939899833054, 'eval/demon_move_f1': 0.5004444355334563, 'eval/demon_cc_f1': 0.7461346402737855, 'eval/latency_8/demon_fire_precision': 0.8956594323873122, 'eval/latency_8/demon_fire_recall': 1.0, 'eval/latency_8/demon_fire_f1': 0.944958168207838, 'eval/latency_8/demon_fire_rate_model': 1.0, 'eval/latency_8/demon_fire_rate_teacher': 0.8956594323873122, 'eval/latency_8/demon_left_precision': 0.49643705463182897, 'eval/latency_8/demon_left_recall': 0.4475374732334047, 'eval/latency_8/demon_left_f1': 0.47072072072072074, 'eval/latency_8/demon_left_rate_model': 0.11713967723984418, 'eval/latency_8/demon_left_rate_teacher': 0.12993878686700056, 'eval/latency_8/demon_right_precision': 0.7635327635327636, 'eval/latency_8/demon_right_recall': 0.40606060606060607, 'eval/latency_8/demon_right_f1': 0.5301681503461919, 'eval/latency_8/demon_right_rate_model': 0.09766277128547579, 'eval/latency_8/demon_right_rate_teacher': 0.18363939899833054, 'eval/latency_8/demon_move_f1': 0.5004444355334563, 'eval/latency_8/demon_cc_f1': 0.7461346402737855, 'timing/data': 0.025219228118658066, 'timing/model': 0.919900719076395, 'learning_rate/qwen_vl_interface': 9.980320002024774e-06, 'learning_rate/action_model': 9.980320002024774e-05, 'epoch': 0.14}) +[2026-06-17 17:16:23,950][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_300_state +[2026-06-17 17:16:23,976][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 17:16:31,323][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_300_state/model.safetensors +[2026-06-17 17:16:47,586][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_300_state/optimizer.bin +[2026-06-17 17:16:47,587][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_300_state/sampler.bin +[2026-06-17 17:16:47,588][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_300_state/sampler_1.bin +[2026-06-17 17:16:47,592][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_300_state/random_states_0.pkl +[2026-06-17 17:16:47,599][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 300 +[2026-06-17 17:16:47,599][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 17:16:47,611][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 17:16:51,463][starVLA.training.train_starvla][INFO] - Step 301, Loss: {'train/loss': 0.7330392599105835, 'train/grad_norm_pre_clip': 3.628857135772705, 'timing/data': 0.04621297866106033, 'timing/model': 1.1524449354037642, 'learning_rate/qwen_vl_interface': 9.980122847759896e-06, 'learning_rate/action_model': 9.980122847759896e-05, 'epoch': 0.14}) +[2026-06-17 17:16:55,085][starVLA.training.train_starvla][INFO] - Step 302, Loss: {'train/loss': 0.7645901441574097, 'train/grad_norm_pre_clip': 4.276193141937256, 'timing/data': 0.02703179232776165, 'timing/model': 0.9271572111174464, 'learning_rate/qwen_vl_interface': 9.97992471293546e-06, 'learning_rate/action_model': 9.97992471293546e-05, 'epoch': 0.14}) +[2026-06-17 17:16:58,930][starVLA.training.train_starvla][INFO] - Step 303, Loss: {'train/loss': 0.6267547607421875, 'train/grad_norm_pre_clip': 5.268565654754639, 'timing/data': 0.027346170507371426, 'timing/model': 0.9160186322405934, 'learning_rate/qwen_vl_interface': 9.979725597592543e-06, 'learning_rate/action_model': 9.979725597592541e-05, 'epoch': 0.14}) +[2026-06-17 17:17:02,921][starVLA.training.train_starvla][INFO] - Step 304, Loss: {'train/loss': 0.8182425498962402, 'train/grad_norm_pre_clip': 4.938477993011475, 'timing/data': 0.02945271972566843, 'timing/model': 0.9280220558866858, 'learning_rate/qwen_vl_interface': 9.979525501772418e-06, 'learning_rate/action_model': 9.979525501772417e-05, 'epoch': 0.14}) +[2026-06-17 17:17:07,682][starVLA.training.train_starvla][INFO] - Step 305, Loss: {'train/loss': 0.5255504250526428, 'train/grad_norm_pre_clip': 3.814145803451538, 'timing/data': 0.030757143162190914, 'timing/model': 1.4751157406717539, 'learning_rate/qwen_vl_interface': 9.97932442551657e-06, 'learning_rate/action_model': 9.979324425516569e-05, 'epoch': 0.14}) +[2026-06-17 17:17:11,708][starVLA.training.train_starvla][INFO] - Step 306, Loss: {'train/loss': 0.6768946051597595, 'train/grad_norm_pre_clip': 4.001811504364014, 'timing/data': 0.03195853438228369, 'timing/model': 1.049328327178955, 'learning_rate/qwen_vl_interface': 9.979122368866678e-06, 'learning_rate/action_model': 9.979122368866679e-05, 'epoch': 0.14}) +[2026-06-17 17:17:15,757][starVLA.training.train_starvla][INFO] - Step 307, Loss: {'train/loss': 1.214182734489441, 'train/grad_norm_pre_clip': 8.519322395324707, 'timing/data': 0.037678672932088375, 'timing/model': 1.1473044157028198, 'learning_rate/qwen_vl_interface': 9.978919331864631e-06, 'learning_rate/action_model': 9.97891933186463e-05, 'epoch': 0.14}) +[2026-06-17 17:17:19,706][starVLA.training.train_starvla][INFO] - Step 308, Loss: {'train/loss': 1.120665192604065, 'train/grad_norm_pre_clip': 5.391540050506592, 'timing/data': 0.0267525902017951, 'timing/model': 1.176160154864192, 'learning_rate/qwen_vl_interface': 9.978715314552518e-06, 'learning_rate/action_model': 9.978715314552518e-05, 'epoch': 0.14}) +[2026-06-17 17:17:23,441][starVLA.training.train_starvla][INFO] - Step 309, Loss: {'train/loss': 0.7914242148399353, 'train/grad_norm_pre_clip': 5.035274028778076, 'timing/data': 0.02592929918318987, 'timing/model': 0.9307073624804616, 'learning_rate/qwen_vl_interface': 9.978510316972631e-06, 'learning_rate/action_model': 9.978510316972631e-05, 'epoch': 0.14}) +[2026-06-17 17:17:27,486][starVLA.training.train_starvla][INFO] - Step 310, Loss: {'train/loss': 0.660724401473999, 'train/grad_norm_pre_clip': 3.4815919399261475, 'timing/data': 0.03512313310056925, 'timing/model': 1.1693344227969646, 'learning_rate/qwen_vl_interface': 9.978304339167467e-06, 'learning_rate/action_model': 9.978304339167467e-05, 'epoch': 0.14}) +[2026-06-17 17:17:31,387][starVLA.training.train_starvla][INFO] - Step 311, Loss: {'train/loss': 0.766678512096405, 'train/grad_norm_pre_clip': 3.295236110687256, 'timing/data': 0.03642946481704712, 'timing/model': 0.9335213517770171, 'learning_rate/qwen_vl_interface': 9.978097381179726e-06, 'learning_rate/action_model': 9.978097381179726e-05, 'epoch': 0.14}) +[2026-06-17 17:17:37,576][starVLA.training.train_starvla][INFO] - Step 312, Loss: {'train/loss': 0.6869518756866455, 'train/grad_norm_pre_clip': 4.5514445304870605, 'timing/data': 0.032890611328184605, 'timing/model': 2.376090247184038, 'learning_rate/qwen_vl_interface': 9.977889443052312e-06, 'learning_rate/action_model': 9.977889443052311e-05, 'epoch': 0.14}) +[2026-06-17 17:17:41,829][starVLA.training.train_starvla][INFO] - Step 313, Loss: {'train/loss': 0.5208292007446289, 'train/grad_norm_pre_clip': 4.8269877433776855, 'timing/data': 0.03606298007071018, 'timing/model': 0.9323463765904307, 'learning_rate/qwen_vl_interface': 9.977680524828327e-06, 'learning_rate/action_model': 9.977680524828327e-05, 'epoch': 0.14}) +[2026-06-17 17:17:48,565][starVLA.training.train_starvla][INFO] - Step 314, Loss: {'train/loss': 0.7432939410209656, 'train/grad_norm_pre_clip': 5.20766544342041, 'timing/data': 0.02581117022782564, 'timing/model': 3.9662386644631624, 'learning_rate/qwen_vl_interface': 9.977470626551083e-06, 'learning_rate/action_model': 9.977470626551083e-05, 'epoch': 0.14}) +[2026-06-17 17:17:52,231][starVLA.training.train_starvla][INFO] - Step 315, Loss: {'train/loss': 0.8148714303970337, 'train/grad_norm_pre_clip': 4.81931734085083, 'timing/data': 0.029411496594548225, 'timing/model': 0.937316432595253, 'learning_rate/qwen_vl_interface': 9.977259748264089e-06, 'learning_rate/action_model': 9.97725974826409e-05, 'epoch': 0.14}) +[2026-06-17 17:17:56,255][starVLA.training.train_starvla][INFO] - Step 316, Loss: {'train/loss': 0.9932589530944824, 'train/grad_norm_pre_clip': 7.946063995361328, 'timing/data': 0.027415352873504162, 'timing/model': 0.9604831049218774, 'learning_rate/qwen_vl_interface': 9.977047890011065e-06, 'learning_rate/action_model': 9.977047890011065e-05, 'epoch': 0.14}) +[2026-06-17 17:18:00,067][starVLA.training.train_starvla][INFO] - Step 317, Loss: {'train/loss': 1.092773675918579, 'train/grad_norm_pre_clip': 5.286280632019043, 'timing/data': 0.027993513271212578, 'timing/model': 0.9200014257803559, 'learning_rate/qwen_vl_interface': 9.976835051835928e-06, 'learning_rate/action_model': 9.976835051835926e-05, 'epoch': 0.14}) +[2026-06-17 17:18:04,141][starVLA.training.train_starvla][INFO] - Step 318, Loss: {'train/loss': 0.889939546585083, 'train/grad_norm_pre_clip': 4.632460594177246, 'timing/data': 0.024681183509528637, 'timing/model': 0.9353187652304769, 'learning_rate/qwen_vl_interface': 9.976621233782795e-06, 'learning_rate/action_model': 9.976621233782795e-05, 'epoch': 0.14}) +[2026-06-17 17:18:08,207][starVLA.training.train_starvla][INFO] - Step 319, Loss: {'train/loss': 0.8247949481010437, 'train/grad_norm_pre_clip': 4.296199321746826, 'timing/data': 0.02323373779654503, 'timing/model': 1.2026890218257904, 'learning_rate/qwen_vl_interface': 9.976406435895995e-06, 'learning_rate/action_model': 9.976406435895995e-05, 'epoch': 0.14}) +[2026-06-17 17:18:12,093][starVLA.training.train_starvla][INFO] - Step 320, Loss: {'train/loss': 0.731284499168396, 'train/grad_norm_pre_clip': 4.25765323638916, 'timing/data': 0.0236396174877882, 'timing/model': 1.0050970884039998, 'learning_rate/qwen_vl_interface': 9.976190658220056e-06, 'learning_rate/action_model': 9.976190658220056e-05, 'epoch': 0.15}) +[2026-06-17 17:18:16,215][starVLA.training.train_starvla][INFO] - Step 321, Loss: {'train/loss': 0.9471758604049683, 'train/grad_norm_pre_clip': 8.697300910949707, 'timing/data': 0.03391302190721035, 'timing/model': 1.175527478568256, 'learning_rate/qwen_vl_interface': 9.975973900799709e-06, 'learning_rate/action_model': 9.975973900799708e-05, 'epoch': 0.15}) +[2026-06-17 17:18:20,235][starVLA.training.train_starvla][INFO] - Step 322, Loss: {'train/loss': 0.8279477953910828, 'train/grad_norm_pre_clip': 3.6485207080841064, 'timing/data': 0.026858902536332607, 'timing/model': 0.9294358901679516, 'learning_rate/qwen_vl_interface': 9.975756163679884e-06, 'learning_rate/action_model': 9.975756163679885e-05, 'epoch': 0.15}) +[2026-06-17 17:18:24,506][starVLA.training.train_starvla][INFO] - Step 323, Loss: {'train/loss': 0.6665688753128052, 'train/grad_norm_pre_clip': 5.966402530670166, 'timing/data': 0.034755161963403225, 'timing/model': 1.0351025322452188, 'learning_rate/qwen_vl_interface': 9.975537446905722e-06, 'learning_rate/action_model': 9.975537446905723e-05, 'epoch': 0.15}) +[2026-06-17 17:18:28,626][starVLA.training.train_starvla][INFO] - Step 324, Loss: {'train/loss': 1.1799819469451904, 'train/grad_norm_pre_clip': 7.383425235748291, 'timing/data': 0.04119277186691761, 'timing/model': 0.9362059663981199, 'learning_rate/qwen_vl_interface': 9.975317750522565e-06, 'learning_rate/action_model': 9.975317750522564e-05, 'epoch': 0.15}) +[2026-06-17 17:18:32,586][starVLA.training.train_starvla][INFO] - Step 325, Loss: {'train/loss': 0.756934404373169, 'train/grad_norm_pre_clip': 3.9144582748413086, 'timing/data': 0.03675275668501854, 'timing/model': 0.9215642549097538, 'learning_rate/qwen_vl_interface': 9.97509707457595e-06, 'learning_rate/action_model': 9.97509707457595e-05, 'epoch': 0.15}) +[2026-06-17 17:18:36,383][starVLA.training.train_starvla][INFO] - Step 326, Loss: {'train/loss': 0.7194247841835022, 'train/grad_norm_pre_clip': 4.350706577301025, 'timing/data': 0.02559006493538618, 'timing/model': 0.9263155218213797, 'learning_rate/qwen_vl_interface': 9.974875419111631e-06, 'learning_rate/action_model': 9.97487541911163e-05, 'epoch': 0.15}) +[2026-06-17 17:18:40,164][starVLA.training.train_starvla][INFO] - Step 327, Loss: {'train/loss': 0.527134120464325, 'train/grad_norm_pre_clip': 3.9970881938934326, 'timing/data': 0.03665694501250982, 'timing/model': 1.106784425675869, 'learning_rate/qwen_vl_interface': 9.974652784175552e-06, 'learning_rate/action_model': 9.974652784175551e-05, 'epoch': 0.15}) +[2026-06-17 17:18:43,932][starVLA.training.train_starvla][INFO] - Step 328, Loss: {'train/loss': 1.1918643712997437, 'train/grad_norm_pre_clip': 7.853901386260986, 'timing/data': 0.023743238300085068, 'timing/model': 1.1172696640715003, 'learning_rate/qwen_vl_interface': 9.974429169813865e-06, 'learning_rate/action_model': 9.974429169813864e-05, 'epoch': 0.15}) +[2026-06-17 17:18:47,793][starVLA.training.train_starvla][INFO] - Step 329, Loss: {'train/loss': 1.030035376548767, 'train/grad_norm_pre_clip': 4.285116195678711, 'timing/data': 0.02053381595760584, 'timing/model': 0.9255218924954534, 'learning_rate/qwen_vl_interface': 9.974204576072928e-06, 'learning_rate/action_model': 9.974204576072928e-05, 'epoch': 0.15}) +[2026-06-17 17:18:51,606][starVLA.training.train_starvla][INFO] - Step 330, Loss: {'train/loss': 0.6396985054016113, 'train/grad_norm_pre_clip': 4.456879615783691, 'timing/data': 0.023198827169835567, 'timing/model': 0.9286836609244347, 'learning_rate/qwen_vl_interface': 9.973979002999299e-06, 'learning_rate/action_model': 9.973979002999299e-05, 'epoch': 0.15}) +[2026-06-17 17:18:55,404][starVLA.training.train_starvla][INFO] - Step 331, Loss: {'train/loss': 0.5559616088867188, 'train/grad_norm_pre_clip': 2.6249592304229736, 'timing/data': 0.036274404264986515, 'timing/model': 0.9300747476518154, 'learning_rate/qwen_vl_interface': 9.97375245063974e-06, 'learning_rate/action_model': 9.97375245063974e-05, 'epoch': 0.15}) +[2026-06-17 17:18:59,264][starVLA.training.train_starvla][INFO] - Step 332, Loss: {'train/loss': 0.6634922027587891, 'train/grad_norm_pre_clip': 4.009986400604248, 'timing/data': 0.03128939773887396, 'timing/model': 0.9836399955675006, 'learning_rate/qwen_vl_interface': 9.973524919041215e-06, 'learning_rate/action_model': 9.973524919041215e-05, 'epoch': 0.15}) +[2026-06-17 17:19:03,150][starVLA.training.train_starvla][INFO] - Step 333, Loss: {'train/loss': 0.6280734539031982, 'train/grad_norm_pre_clip': 4.561916351318359, 'timing/data': 0.029164384119212627, 'timing/model': 0.9599881777539849, 'learning_rate/qwen_vl_interface': 9.97329640825089e-06, 'learning_rate/action_model': 9.973296408250891e-05, 'epoch': 0.15}) +[2026-06-17 17:19:06,995][starVLA.training.train_starvla][INFO] - Step 334, Loss: {'train/loss': 0.9728977680206299, 'train/grad_norm_pre_clip': 5.028264999389648, 'timing/data': 0.03060372732579708, 'timing/model': 1.136940031312406, 'learning_rate/qwen_vl_interface': 9.97306691831614e-06, 'learning_rate/action_model': 9.97306691831614e-05, 'epoch': 0.15}) +[2026-06-17 17:19:10,838][starVLA.training.train_starvla][INFO] - Step 335, Loss: {'train/loss': 0.5723496675491333, 'train/grad_norm_pre_clip': 5.781803131103516, 'timing/data': 0.024513129144906998, 'timing/model': 1.0808355621993542, 'learning_rate/qwen_vl_interface': 9.972836449284533e-06, 'learning_rate/action_model': 9.972836449284532e-05, 'epoch': 0.15}) +[2026-06-17 17:19:14,600][starVLA.training.train_starvla][INFO] - Step 336, Loss: {'train/loss': 0.6834913492202759, 'train/grad_norm_pre_clip': 3.9579365253448486, 'timing/data': 0.031966375187039375, 'timing/model': 1.030623773112893, 'learning_rate/qwen_vl_interface': 9.97260500120385e-06, 'learning_rate/action_model': 9.97260500120385e-05, 'epoch': 0.15}) +[2026-06-17 17:19:18,520][starVLA.training.train_starvla][INFO] - Step 337, Loss: {'train/loss': 0.8376589417457581, 'train/grad_norm_pre_clip': 4.5163655281066895, 'timing/data': 0.02707612421363592, 'timing/model': 1.045139729976654, 'learning_rate/qwen_vl_interface': 9.972372574122067e-06, 'learning_rate/action_model': 9.972372574122068e-05, 'epoch': 0.15}) +[2026-06-17 17:19:22,543][starVLA.training.train_starvla][INFO] - Step 338, Loss: {'train/loss': 0.6024271249771118, 'train/grad_norm_pre_clip': 4.793778419494629, 'timing/data': 0.0276954285800457, 'timing/model': 0.9310879101976752, 'learning_rate/qwen_vl_interface': 9.972139168087368e-06, 'learning_rate/action_model': 9.972139168087369e-05, 'epoch': 0.15}) +[2026-06-17 17:19:26,303][starVLA.training.train_starvla][INFO] - Step 339, Loss: {'train/loss': 0.9297447204589844, 'train/grad_norm_pre_clip': 3.2236034870147705, 'timing/data': 0.0227573923766613, 'timing/model': 0.9847829546779394, 'learning_rate/qwen_vl_interface': 9.97190478314814e-06, 'learning_rate/action_model': 9.97190478314814e-05, 'epoch': 0.15}) +[2026-06-17 17:19:32,191][starVLA.training.train_starvla][INFO] - Step 340, Loss: {'train/loss': 0.5842101573944092, 'train/grad_norm_pre_clip': 2.2838733196258545, 'timing/data': 0.029247375205159187, 'timing/model': 1.08561294618994, 'learning_rate/qwen_vl_interface': 9.971669419352968e-06, 'learning_rate/action_model': 9.971669419352966e-05, 'epoch': 0.15}) +[2026-06-17 17:19:36,010][starVLA.training.train_starvla][INFO] - Step 341, Loss: {'train/loss': 0.7298845648765564, 'train/grad_norm_pre_clip': 5.266915798187256, 'timing/data': 0.02839541621506214, 'timing/model': 1.1157978577539325, 'learning_rate/qwen_vl_interface': 9.971433076750644e-06, 'learning_rate/action_model': 9.971433076750643e-05, 'epoch': 0.15}) +[2026-06-17 17:19:39,906][starVLA.training.train_starvla][INFO] - Step 342, Loss: {'train/loss': 0.7121683955192566, 'train/grad_norm_pre_clip': 4.542931079864502, 'timing/data': 0.028708884492516518, 'timing/model': 1.0272944774478674, 'learning_rate/qwen_vl_interface': 9.971195755390166e-06, 'learning_rate/action_model': 9.971195755390164e-05, 'epoch': 0.16}) +[2026-06-17 17:19:43,913][starVLA.training.train_starvla][INFO] - Step 343, Loss: {'train/loss': 1.0991522073745728, 'train/grad_norm_pre_clip': 6.878576278686523, 'timing/data': 0.03484859969466925, 'timing/model': 0.9361335719004273, 'learning_rate/qwen_vl_interface': 9.970957455320723e-06, 'learning_rate/action_model': 9.970957455320723e-05, 'epoch': 0.16}) +[2026-06-17 17:19:47,668][starVLA.training.train_starvla][INFO] - Step 344, Loss: {'train/loss': 0.6932663321495056, 'train/grad_norm_pre_clip': 5.106848239898682, 'timing/data': 0.027819213457405567, 'timing/model': 0.9260231656953692, 'learning_rate/qwen_vl_interface': 9.970718176591723e-06, 'learning_rate/action_model': 9.970718176591722e-05, 'epoch': 0.16}) +[2026-06-17 17:19:51,602][starVLA.training.train_starvla][INFO] - Step 345, Loss: {'train/loss': 1.1583694219589233, 'train/grad_norm_pre_clip': 5.779850482940674, 'timing/data': 0.0309079485014081, 'timing/model': 0.9320754026994109, 'learning_rate/qwen_vl_interface': 9.970477919252763e-06, 'learning_rate/action_model': 9.970477919252763e-05, 'epoch': 0.16}) +[2026-06-17 17:19:55,390][starVLA.training.train_starvla][INFO] - Step 346, Loss: {'train/loss': 0.9147231578826904, 'train/grad_norm_pre_clip': 3.796452760696411, 'timing/data': 0.030968972481787205, 'timing/model': 0.9248890252783895, 'learning_rate/qwen_vl_interface': 9.970236683353652e-06, 'learning_rate/action_model': 9.970236683353652e-05, 'epoch': 0.16}) +[2026-06-17 17:19:59,201][starVLA.training.train_starvla][INFO] - Step 347, Loss: {'train/loss': 0.7585973143577576, 'train/grad_norm_pre_clip': 3.3212530612945557, 'timing/data': 0.033421147614717484, 'timing/model': 1.1592222098261118, 'learning_rate/qwen_vl_interface': 9.969994468944399e-06, 'learning_rate/action_model': 9.969994468944398e-05, 'epoch': 0.16}) +[2026-06-17 17:20:03,040][starVLA.training.train_starvla][INFO] - Step 348, Loss: {'train/loss': 0.8612465858459473, 'train/grad_norm_pre_clip': 5.111877918243408, 'timing/data': 0.030735693871974945, 'timing/model': 1.1765286726877093, 'learning_rate/qwen_vl_interface': 9.969751276075212e-06, 'learning_rate/action_model': 9.969751276075212e-05, 'epoch': 0.16}) +[2026-06-17 17:20:06,789][starVLA.training.train_starvla][INFO] - Step 349, Loss: {'train/loss': 0.8416014313697815, 'train/grad_norm_pre_clip': 5.573671817779541, 'timing/data': 0.027031725272536278, 'timing/model': 0.9310129899531603, 'learning_rate/qwen_vl_interface': 9.969507104796506e-06, 'learning_rate/action_model': 9.969507104796505e-05, 'epoch': 0.16}) +[2026-06-17 17:20:10,652][starVLA.training.train_starvla][INFO] - Step 350, Loss: {'train/loss': 1.1072386503219604, 'train/grad_norm_pre_clip': 7.262192726135254, 'timing/data': 0.02735781855881214, 'timing/model': 0.9506485713645816, 'learning_rate/qwen_vl_interface': 9.969261955158898e-06, 'learning_rate/action_model': 9.969261955158898e-05, 'epoch': 0.16}) +[2026-06-17 17:20:14,784][starVLA.training.train_starvla][INFO] - Step 351, Loss: {'train/loss': 0.7438064813613892, 'train/grad_norm_pre_clip': 4.669114112854004, 'timing/data': 0.024219680577516556, 'timing/model': 1.1835169298574328, 'learning_rate/qwen_vl_interface': 9.96901582721321e-06, 'learning_rate/action_model': 9.96901582721321e-05, 'epoch': 0.16}) +[2026-06-17 17:20:19,330][starVLA.training.train_starvla][INFO] - Step 352, Loss: {'train/loss': 0.7668755650520325, 'train/grad_norm_pre_clip': 5.9232025146484375, 'timing/data': 0.03447040170431137, 'timing/model': 1.1812279159203172, 'learning_rate/qwen_vl_interface': 9.968768721010463e-06, 'learning_rate/action_model': 9.968768721010463e-05, 'epoch': 0.16}) +[2026-06-17 17:20:23,192][starVLA.training.train_starvla][INFO] - Step 353, Loss: {'train/loss': 1.0041403770446777, 'train/grad_norm_pre_clip': 4.158671855926514, 'timing/data': 0.03502717148512602, 'timing/model': 0.9467489346861839, 'learning_rate/qwen_vl_interface': 9.968520636601883e-06, 'learning_rate/action_model': 9.968520636601883e-05, 'epoch': 0.16}) +[2026-06-17 17:20:27,428][starVLA.training.train_starvla][INFO] - Step 354, Loss: {'train/loss': 0.6926277279853821, 'train/grad_norm_pre_clip': 6.068181037902832, 'timing/data': 0.026135388761758804, 'timing/model': 1.1026920340955257, 'learning_rate/qwen_vl_interface': 9.968271574038899e-06, 'learning_rate/action_model': 9.968271574038898e-05, 'epoch': 0.16}) +[2026-06-17 17:20:31,315][starVLA.training.train_starvla][INFO] - Step 355, Loss: {'train/loss': 1.0002142190933228, 'train/grad_norm_pre_clip': 4.8841657638549805, 'timing/data': 0.02685353998094797, 'timing/model': 1.1121281571686268, 'learning_rate/qwen_vl_interface': 9.968021533373138e-06, 'learning_rate/action_model': 9.968021533373138e-05, 'epoch': 0.16}) +[2026-06-17 17:20:34,938][starVLA.training.train_starvla][INFO] - Step 356, Loss: {'train/loss': 0.8235604166984558, 'train/grad_norm_pre_clip': 4.7306694984436035, 'timing/data': 0.025842929258942604, 'timing/model': 0.959497575648129, 'learning_rate/qwen_vl_interface': 9.967770514656439e-06, 'learning_rate/action_model': 9.967770514656438e-05, 'epoch': 0.16}) +[2026-06-17 17:20:38,724][starVLA.training.train_starvla][INFO] - Step 357, Loss: {'train/loss': 0.8336331248283386, 'train/grad_norm_pre_clip': 3.896153688430786, 'timing/data': 0.02835164964199066, 'timing/model': 0.9143688259646297, 'learning_rate/qwen_vl_interface': 9.967518517940832e-06, 'learning_rate/action_model': 9.967518517940832e-05, 'epoch': 0.16}) +[2026-06-17 17:20:42,492][starVLA.training.train_starvla][INFO] - Step 358, Loss: {'train/loss': 0.793217658996582, 'train/grad_norm_pre_clip': 6.75840950012207, 'timing/data': 0.03188958019018173, 'timing/model': 0.9237069394439459, 'learning_rate/qwen_vl_interface': 9.967265543278563e-06, 'learning_rate/action_model': 9.967265543278562e-05, 'epoch': 0.16}) +[2026-06-17 17:20:46,314][starVLA.training.train_starvla][INFO] - Step 359, Loss: {'train/loss': 0.8156213760375977, 'train/grad_norm_pre_clip': 4.951179504394531, 'timing/data': 0.033682710491120815, 'timing/model': 0.949722315184772, 'learning_rate/qwen_vl_interface': 9.96701159072207e-06, 'learning_rate/action_model': 9.96701159072207e-05, 'epoch': 0.16}) +[2026-06-17 17:20:50,039][starVLA.training.train_starvla][INFO] - Step 360, Loss: {'train/loss': 0.6575568318367004, 'train/grad_norm_pre_clip': 4.673176288604736, 'timing/data': 0.027756298892199993, 'timing/model': 0.9123746510595083, 'learning_rate/qwen_vl_interface': 9.966756660323998e-06, 'learning_rate/action_model': 9.966756660323999e-05, 'epoch': 0.16}) +[2026-06-17 17:20:53,824][starVLA.training.train_starvla][INFO] - Step 361, Loss: {'train/loss': 0.611059308052063, 'train/grad_norm_pre_clip': 4.414820194244385, 'timing/data': 0.02277651522308588, 'timing/model': 1.1060521993786097, 'learning_rate/qwen_vl_interface': 9.966500752137195e-06, 'learning_rate/action_model': 9.966500752137196e-05, 'epoch': 0.16}) +[2026-06-17 17:20:57,413][starVLA.training.train_starvla][INFO] - Step 362, Loss: {'train/loss': 0.623139500617981, 'train/grad_norm_pre_clip': 5.315093517303467, 'timing/data': 0.024018069729208946, 'timing/model': 0.921709137968719, 'learning_rate/qwen_vl_interface': 9.966243866214712e-06, 'learning_rate/action_model': 9.966243866214712e-05, 'epoch': 0.16}) +[2026-06-17 17:21:01,202][starVLA.training.train_starvla][INFO] - Step 363, Loss: {'train/loss': 0.8025220036506653, 'train/grad_norm_pre_clip': 3.4865219593048096, 'timing/data': 0.02485031448304653, 'timing/model': 0.9451560247689486, 'learning_rate/qwen_vl_interface': 9.965986002609799e-06, 'learning_rate/action_model': 9.965986002609799e-05, 'epoch': 0.16}) +[2026-06-17 17:21:05,169][starVLA.training.train_starvla][INFO] - Step 364, Loss: {'train/loss': 0.897312581539154, 'train/grad_norm_pre_clip': 5.78231954574585, 'timing/data': 0.02621718030422926, 'timing/model': 0.989603248424828, 'learning_rate/qwen_vl_interface': 9.965727161375915e-06, 'learning_rate/action_model': 9.965727161375914e-05, 'epoch': 0.17}) +[2026-06-17 17:21:09,219][starVLA.training.train_starvla][INFO] - Step 365, Loss: {'train/loss': 1.0747939348220825, 'train/grad_norm_pre_clip': 5.973972320556641, 'timing/data': 0.023461196571588516, 'timing/model': 1.0597887597978115, 'learning_rate/qwen_vl_interface': 9.965467342566715e-06, 'learning_rate/action_model': 9.965467342566715e-05, 'epoch': 0.17}) +[2026-06-17 17:21:13,124][starVLA.training.train_starvla][INFO] - Step 366, Loss: {'train/loss': 0.9464852809906006, 'train/grad_norm_pre_clip': 4.537799835205078, 'timing/data': 0.031039707362651825, 'timing/model': 0.9664713097736239, 'learning_rate/qwen_vl_interface': 9.965206546236061e-06, 'learning_rate/action_model': 9.96520654623606e-05, 'epoch': 0.17}) +[2026-06-17 17:21:16,969][starVLA.training.train_starvla][INFO] - Step 367, Loss: {'train/loss': 0.8277245759963989, 'train/grad_norm_pre_clip': 3.920024871826172, 'timing/data': 0.034642698243260384, 'timing/model': 0.9306772835552692, 'learning_rate/qwen_vl_interface': 9.964944772438014e-06, 'learning_rate/action_model': 9.964944772438013e-05, 'epoch': 0.17}) +[2026-06-17 17:21:20,996][starVLA.training.train_starvla][INFO] - Step 368, Loss: {'train/loss': 1.2008644342422485, 'train/grad_norm_pre_clip': 5.054434299468994, 'timing/data': 0.036793128587305546, 'timing/model': 1.224060021340847, 'learning_rate/qwen_vl_interface': 9.964682021226844e-06, 'learning_rate/action_model': 9.964682021226844e-05, 'epoch': 0.17}) +[2026-06-17 17:21:24,742][starVLA.training.train_starvla][INFO] - Step 369, Loss: {'train/loss': 0.6639100909233093, 'train/grad_norm_pre_clip': 3.8152987957000732, 'timing/data': 0.043902765959501266, 'timing/model': 0.9272079570218921, 'learning_rate/qwen_vl_interface': 9.964418292657019e-06, 'learning_rate/action_model': 9.964418292657018e-05, 'epoch': 0.17}) +[2026-06-17 17:21:28,705][starVLA.training.train_starvla][INFO] - Step 370, Loss: {'train/loss': 0.8675001263618469, 'train/grad_norm_pre_clip': 3.406525135040283, 'timing/data': 0.028925112448632717, 'timing/model': 0.9605666166171432, 'learning_rate/qwen_vl_interface': 9.96415358678321e-06, 'learning_rate/action_model': 9.964153586783209e-05, 'epoch': 0.17}) +[2026-06-17 17:21:32,588][starVLA.training.train_starvla][INFO] - Step 371, Loss: {'train/loss': 1.0834107398986816, 'train/grad_norm_pre_clip': 4.883136749267578, 'timing/data': 0.03972330503165722, 'timing/model': 0.9831223571673036, 'learning_rate/qwen_vl_interface': 9.963887903660285e-06, 'learning_rate/action_model': 9.963887903660285e-05, 'epoch': 0.17}) +[2026-06-17 17:21:36,318][starVLA.training.train_starvla][INFO] - Step 372, Loss: {'train/loss': 0.8522610068321228, 'train/grad_norm_pre_clip': 7.460484504699707, 'timing/data': 0.02825932204723358, 'timing/model': 0.9177779704332352, 'learning_rate/qwen_vl_interface': 9.96362124334333e-06, 'learning_rate/action_model': 9.963621243343329e-05, 'epoch': 0.17}) +[2026-06-17 17:21:40,100][starVLA.training.train_starvla][INFO] - Step 373, Loss: {'train/loss': 0.6211402416229248, 'train/grad_norm_pre_clip': 7.574045181274414, 'timing/data': 0.03416411578655243, 'timing/model': 0.9200962986797094, 'learning_rate/qwen_vl_interface': 9.963353605887617e-06, 'learning_rate/action_model': 9.963353605887616e-05, 'epoch': 0.17}) +[2026-06-17 17:21:43,812][starVLA.training.train_starvla][INFO] - Step 374, Loss: {'train/loss': 0.8248274326324463, 'train/grad_norm_pre_clip': 5.060171604156494, 'timing/data': 0.035113343968987465, 'timing/model': 1.0815627556294203, 'learning_rate/qwen_vl_interface': 9.963084991348628e-06, 'learning_rate/action_model': 9.963084991348628e-05, 'epoch': 0.17}) +[2026-06-17 17:21:47,528][starVLA.training.train_starvla][INFO] - Step 375, Loss: {'train/loss': 0.9337460994720459, 'train/grad_norm_pre_clip': 3.17506742477417, 'timing/data': 0.042077734135091305, 'timing/model': 1.0600780611857772, 'learning_rate/qwen_vl_interface': 9.962815399782051e-06, 'learning_rate/action_model': 9.96281539978205e-05, 'epoch': 0.17}) +[2026-06-17 17:21:51,135][starVLA.training.train_starvla][INFO] - Step 376, Loss: {'train/loss': 1.0194002389907837, 'train/grad_norm_pre_clip': 4.198518753051758, 'timing/data': 0.03363378345966339, 'timing/model': 0.9330947585403919, 'learning_rate/qwen_vl_interface': 9.96254483124377e-06, 'learning_rate/action_model': 9.96254483124377e-05, 'epoch': 0.17}) +[2026-06-17 17:21:55,320][starVLA.training.train_starvla][INFO] - Step 377, Loss: {'train/loss': 0.8810290098190308, 'train/grad_norm_pre_clip': 3.410569667816162, 'timing/data': 0.043968020007014275, 'timing/model': 0.9297114489600062, 'learning_rate/qwen_vl_interface': 9.962273285789876e-06, 'learning_rate/action_model': 9.962273285789875e-05, 'epoch': 0.17}) +[2026-06-17 17:21:59,140][starVLA.training.train_starvla][INFO] - Step 378, Loss: {'train/loss': 0.7175341248512268, 'train/grad_norm_pre_clip': 6.419093608856201, 'timing/data': 0.035350143909454346, 'timing/model': 0.9230573046952486, 'learning_rate/qwen_vl_interface': 9.962000763476658e-06, 'learning_rate/action_model': 9.962000763476658e-05, 'epoch': 0.17}) +[2026-06-17 17:22:04,463][starVLA.training.train_starvla][INFO] - Step 379, Loss: {'train/loss': 0.983984112739563, 'train/grad_norm_pre_clip': 5.564039707183838, 'timing/data': 0.028532189317047596, 'timing/model': 0.9307117471471429, 'learning_rate/qwen_vl_interface': 9.96172726436061e-06, 'learning_rate/action_model': 9.96172726436061e-05, 'epoch': 0.17}) +[2026-06-17 17:22:11,226][starVLA.training.train_starvla][INFO] - Step 380, Loss: {'train/loss': 0.9358300566673279, 'train/grad_norm_pre_clip': 4.134192943572998, 'timing/data': 0.03718474041670561, 'timing/model': 0.9455074304714799, 'learning_rate/qwen_vl_interface': 9.961452788498432e-06, 'learning_rate/action_model': 9.961452788498433e-05, 'epoch': 0.17}) +[2026-06-17 17:22:15,105][starVLA.training.train_starvla][INFO] - Step 381, Loss: {'train/loss': 0.7759968042373657, 'train/grad_norm_pre_clip': 5.356113910675049, 'timing/data': 0.029964773915708065, 'timing/model': 1.1757855843752623, 'learning_rate/qwen_vl_interface': 9.96117733594702e-06, 'learning_rate/action_model': 9.96117733594702e-05, 'epoch': 0.17}) +[2026-06-17 17:22:19,494][starVLA.training.train_starvla][INFO] - Step 382, Loss: {'train/loss': 0.6763355731964111, 'train/grad_norm_pre_clip': 4.614033222198486, 'timing/data': 0.04356030374765396, 'timing/model': 1.1780015751719475, 'learning_rate/qwen_vl_interface': 9.960900906763478e-06, 'learning_rate/action_model': 9.960900906763477e-05, 'epoch': 0.17}) +[2026-06-17 17:22:23,250][starVLA.training.train_starvla][INFO] - Step 383, Loss: {'train/loss': 1.1512985229492188, 'train/grad_norm_pre_clip': 5.305505275726318, 'timing/data': 0.04044636059552431, 'timing/model': 0.9364484939724207, 'learning_rate/qwen_vl_interface': 9.960623501005108e-06, 'learning_rate/action_model': 9.960623501005107e-05, 'epoch': 0.17}) +[2026-06-17 17:22:27,194][starVLA.training.train_starvla][INFO] - Step 384, Loss: {'train/loss': 1.0417873859405518, 'train/grad_norm_pre_clip': 5.01210355758667, 'timing/data': 0.03655970562249422, 'timing/model': 0.9809656785801053, 'learning_rate/qwen_vl_interface': 9.960345118729416e-06, 'learning_rate/action_model': 9.960345118729417e-05, 'epoch': 0.17}) +[2026-06-17 17:22:30,974][starVLA.training.train_starvla][INFO] - Step 385, Loss: {'train/loss': 1.000802993774414, 'train/grad_norm_pre_clip': 5.414013385772705, 'timing/data': 0.036159940995275974, 'timing/model': 0.9123864974826574, 'learning_rate/qwen_vl_interface': 9.960065759994116e-06, 'learning_rate/action_model': 9.960065759994114e-05, 'epoch': 0.17}) +[2026-06-17 17:22:34,882][starVLA.training.train_starvla][INFO] - Step 386, Loss: {'train/loss': 0.38707077503204346, 'train/grad_norm_pre_clip': 3.201580047607422, 'timing/data': 0.03382032550871372, 'timing/model': 0.9297816101461649, 'learning_rate/qwen_vl_interface': 9.959785424857113e-06, 'learning_rate/action_model': 9.959785424857112e-05, 'epoch': 0.18}) +[2026-06-17 17:22:39,289][starVLA.training.train_starvla][INFO] - Step 387, Loss: {'train/loss': 0.7850532531738281, 'train/grad_norm_pre_clip': 7.83217191696167, 'timing/data': 0.041326708160340786, 'timing/model': 0.9349915571510792, 'learning_rate/qwen_vl_interface': 9.959504113376523e-06, 'learning_rate/action_model': 9.959504113376523e-05, 'epoch': 0.18}) +[2026-06-17 17:22:43,424][starVLA.training.train_starvla][INFO] - Step 388, Loss: {'train/loss': 1.3969285488128662, 'train/grad_norm_pre_clip': 5.445092678070068, 'timing/data': 0.04133303835988045, 'timing/model': 1.4132082322612405, 'learning_rate/qwen_vl_interface': 9.959221825610664e-06, 'learning_rate/action_model': 9.959221825610663e-05, 'epoch': 0.18}) +[2026-06-17 17:22:47,084][starVLA.training.train_starvla][INFO] - Step 389, Loss: {'train/loss': 0.7650207877159119, 'train/grad_norm_pre_clip': 4.266180038452148, 'timing/data': 0.03687900863587856, 'timing/model': 0.9461669065058231, 'learning_rate/qwen_vl_interface': 9.958938561618053e-06, 'learning_rate/action_model': 9.958938561618053e-05, 'epoch': 0.18}) +[2026-06-17 17:22:51,547][starVLA.training.train_starvla][INFO] - Step 390, Loss: {'train/loss': 0.7732303738594055, 'train/grad_norm_pre_clip': 6.102575778961182, 'timing/data': 0.030443070456385612, 'timing/model': 0.9300146698951721, 'learning_rate/qwen_vl_interface': 9.958654321457412e-06, 'learning_rate/action_model': 9.958654321457412e-05, 'epoch': 0.18}) +[2026-06-17 17:22:57,300][starVLA.training.train_starvla][INFO] - Step 391, Loss: {'train/loss': 1.026200532913208, 'train/grad_norm_pre_clip': 6.077122211456299, 'timing/data': 0.0476110577583313, 'timing/model': 0.9535144725814462, 'learning_rate/qwen_vl_interface': 9.958369105187661e-06, 'learning_rate/action_model': 9.958369105187661e-05, 'epoch': 0.18}) +[2026-06-17 17:23:01,223][starVLA.training.train_starvla][INFO] - Step 392, Loss: {'train/loss': 0.530823290348053, 'train/grad_norm_pre_clip': 3.8457212448120117, 'timing/data': 0.04034200590103865, 'timing/model': 0.9604027112945914, 'learning_rate/qwen_vl_interface': 9.958082912867932e-06, 'learning_rate/action_model': 9.958082912867931e-05, 'epoch': 0.18}) +[2026-06-17 17:23:05,457][starVLA.training.train_starvla][INFO] - Step 393, Loss: {'train/loss': 0.6050905585289001, 'train/grad_norm_pre_clip': 2.8770663738250732, 'timing/data': 0.05117490515112877, 'timing/model': 1.2798622511327267, 'learning_rate/qwen_vl_interface': 9.957795744557545e-06, 'learning_rate/action_model': 9.957795744557546e-05, 'epoch': 0.18}) +[2026-06-17 17:23:10,397][starVLA.training.train_starvla][INFO] - Step 394, Loss: {'train/loss': 0.9196128845214844, 'train/grad_norm_pre_clip': 4.312438488006592, 'timing/data': 0.02966941799968481, 'timing/model': 0.9199183415621519, 'learning_rate/qwen_vl_interface': 9.957507600316036e-06, 'learning_rate/action_model': 9.957507600316036e-05, 'epoch': 0.18}) +[2026-06-17 17:23:14,465][starVLA.training.train_starvla][INFO] - Step 395, Loss: {'train/loss': 1.001197338104248, 'train/grad_norm_pre_clip': 5.076587200164795, 'timing/data': 0.03444174490869045, 'timing/model': 1.2516635032370687, 'learning_rate/qwen_vl_interface': 9.957218480203137e-06, 'learning_rate/action_model': 9.957218480203136e-05, 'epoch': 0.18}) +[2026-06-17 17:23:18,245][starVLA.training.train_starvla][INFO] - Step 396, Loss: {'train/loss': 0.9211125373840332, 'train/grad_norm_pre_clip': 4.731624126434326, 'timing/data': 0.042159467935562134, 'timing/model': 0.9263175558298826, 'learning_rate/qwen_vl_interface': 9.95692838427878e-06, 'learning_rate/action_model': 9.956928384278778e-05, 'epoch': 0.18}) +[2026-06-17 17:23:22,033][starVLA.training.train_starvla][INFO] - Step 397, Loss: {'train/loss': 0.6296528577804565, 'train/grad_norm_pre_clip': 3.3549604415893555, 'timing/data': 0.047846932895481586, 'timing/model': 0.9185139993205667, 'learning_rate/qwen_vl_interface': 9.956637312603105e-06, 'learning_rate/action_model': 9.956637312603104e-05, 'epoch': 0.18}) +[2026-06-17 17:23:25,766][starVLA.training.train_starvla][INFO] - Step 398, Loss: {'train/loss': 0.9296965003013611, 'train/grad_norm_pre_clip': 5.057521343231201, 'timing/data': 0.03924562782049179, 'timing/model': 0.9167149122804403, 'learning_rate/qwen_vl_interface': 9.95634526523645e-06, 'learning_rate/action_model': 9.956345265236449e-05, 'epoch': 0.18}) +[2026-06-17 17:23:29,503][starVLA.training.train_starvla][INFO] - Step 399, Loss: {'train/loss': 0.8671145439147949, 'train/grad_norm_pre_clip': 3.9015090465545654, 'timing/data': 0.035635572858154774, 'timing/model': 0.9180338000878692, 'learning_rate/qwen_vl_interface': 9.956052242239358e-06, 'learning_rate/action_model': 9.956052242239358e-05, 'epoch': 0.18}) +[2026-06-17 17:25:50,852][starVLA.training.train_starvla][INFO] - Step 400, Loss: {'train/loss': 0.8741209506988525, 'train/grad_norm_pre_clip': 3.2489752769470215, 'eval/loss': 0.8057727217674255, 'eval/latency_8/loss': 0.8057727217674255, 'eval/demon_fire_precision': 0.9085435779816514, 'eval/demon_fire_recall': 0.9844672258465362, 'eval/demon_fire_f1': 0.9449828537349039, 'eval/demon_fire_rate_model': 0.9705063995548135, 'eval/demon_fire_rate_teacher': 0.8956594323873122, 'eval/demon_left_precision': 0.6742424242424242, 'eval/demon_left_recall': 0.1905781584582441, 'eval/demon_left_f1': 0.2971619365609349, 'eval/demon_left_rate_model': 0.03672787979966611, 'eval/demon_left_rate_teacher': 0.12993878686700056, 'eval/demon_right_precision': 0.49884526558891457, 'eval/demon_right_recall': 0.6545454545454545, 'eval/demon_right_f1': 0.5661861074705111, 'eval/demon_right_rate_model': 0.2409571508069004, 'eval/demon_right_rate_teacher': 0.18363939899833054, 'eval/demon_move_f1': 0.431674022015723, 'eval/demon_cc_f1': 0.6965131687294883, 'eval/latency_8/demon_fire_precision': 0.9085435779816514, 'eval/latency_8/demon_fire_recall': 0.9844672258465362, 'eval/latency_8/demon_fire_f1': 0.9449828537349039, 'eval/latency_8/demon_fire_rate_model': 0.9705063995548135, 'eval/latency_8/demon_fire_rate_teacher': 0.8956594323873122, 'eval/latency_8/demon_left_precision': 0.6742424242424242, 'eval/latency_8/demon_left_recall': 0.1905781584582441, 'eval/latency_8/demon_left_f1': 0.2971619365609349, 'eval/latency_8/demon_left_rate_model': 0.03672787979966611, 'eval/latency_8/demon_left_rate_teacher': 0.12993878686700056, 'eval/latency_8/demon_right_precision': 0.49884526558891457, 'eval/latency_8/demon_right_recall': 0.6545454545454545, 'eval/latency_8/demon_right_f1': 0.5661861074705111, 'eval/latency_8/demon_right_rate_model': 0.2409571508069004, 'eval/latency_8/demon_right_rate_teacher': 0.18363939899833054, 'eval/latency_8/demon_move_f1': 0.431674022015723, 'eval/latency_8/demon_cc_f1': 0.6965131687294883, 'timing/data': 0.047065118327736855, 'timing/model': 1.0789818791672587, 'learning_rate/qwen_vl_interface': 9.955758243672572e-06, 'learning_rate/action_model': 9.955758243672572e-05, 'epoch': 0.18}) +[2026-06-17 17:25:50,860][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_400_state +[2026-06-17 17:25:50,866][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 17:25:58,319][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_400_state/model.safetensors +[2026-06-17 17:26:10,199][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_400_state/optimizer.bin +[2026-06-17 17:26:10,200][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_400_state/sampler.bin +[2026-06-17 17:26:10,202][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_400_state/sampler_1.bin +[2026-06-17 17:26:10,206][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_400_state/random_states_0.pkl +[2026-06-17 17:26:10,226][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 400 +[2026-06-17 17:26:10,226][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 17:26:10,238][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 17:26:14,103][starVLA.training.train_starvla][INFO] - Step 401, Loss: {'train/loss': 0.6750887632369995, 'train/grad_norm_pre_clip': 3.1769258975982666, 'timing/data': 0.032406688667833805, 'timing/model': 0.9316114913672209, 'learning_rate/qwen_vl_interface': 9.955463269597038e-06, 'learning_rate/action_model': 9.955463269597038e-05, 'epoch': 0.18}) +[2026-06-17 17:26:17,908][starVLA.training.train_starvla][INFO] - Step 402, Loss: {'train/loss': 0.7725513577461243, 'train/grad_norm_pre_clip': 4.294989585876465, 'timing/data': 0.030473003163933754, 'timing/model': 0.9368572384119034, 'learning_rate/qwen_vl_interface': 9.955167320073905e-06, 'learning_rate/action_model': 9.955167320073905e-05, 'epoch': 0.18}) +[2026-06-17 17:26:21,850][starVLA.training.train_starvla][INFO] - Step 403, Loss: {'train/loss': 0.8152421712875366, 'train/grad_norm_pre_clip': 3.9877138137817383, 'timing/data': 0.034133180975914, 'timing/model': 1.1686169495806098, 'learning_rate/qwen_vl_interface': 9.954870395164524e-06, 'learning_rate/action_model': 9.954870395164524e-05, 'epoch': 0.18}) +[2026-06-17 17:26:25,715][starVLA.training.train_starvla][INFO] - Step 404, Loss: {'train/loss': 0.9109913110733032, 'train/grad_norm_pre_clip': 5.441704750061035, 'timing/data': 0.04955615010112524, 'timing/model': 1.1559299938380718, 'learning_rate/qwen_vl_interface': 9.954572494930448e-06, 'learning_rate/action_model': 9.954572494930448e-05, 'epoch': 0.18}) +[2026-06-17 17:26:29,346][starVLA.training.train_starvla][INFO] - Step 405, Loss: {'train/loss': 0.7736164927482605, 'train/grad_norm_pre_clip': 5.241401672363281, 'timing/data': 0.03091532364487648, 'timing/model': 0.9526021014899015, 'learning_rate/qwen_vl_interface': 9.95427361943343e-06, 'learning_rate/action_model': 9.954273619433431e-05, 'epoch': 0.18}) +[2026-06-17 17:26:33,232][starVLA.training.train_starvla][INFO] - Step 406, Loss: {'train/loss': 1.017911434173584, 'train/grad_norm_pre_clip': 7.220222473144531, 'timing/data': 0.03391896095126867, 'timing/model': 0.9660465028136969, 'learning_rate/qwen_vl_interface': 9.95397376873543e-06, 'learning_rate/action_model': 9.95397376873543e-05, 'epoch': 0.18}) +[2026-06-17 17:26:37,054][starVLA.training.train_starvla][INFO] - Step 407, Loss: {'train/loss': 0.6565631031990051, 'train/grad_norm_pre_clip': 4.742918968200684, 'timing/data': 0.030071494169533253, 'timing/model': 0.938498442992568, 'learning_rate/qwen_vl_interface': 9.953672942898606e-06, 'learning_rate/action_model': 9.953672942898605e-05, 'epoch': 0.18}) +[2026-06-17 17:26:40,841][starVLA.training.train_starvla][INFO] - Step 408, Loss: {'train/loss': 0.7817212343215942, 'train/grad_norm_pre_clip': 6.247310638427734, 'timing/data': 0.03252769913524389, 'timing/model': 0.9263899773359299, 'learning_rate/qwen_vl_interface': 9.953371141985318e-06, 'learning_rate/action_model': 9.953371141985319e-05, 'epoch': 0.19}) +[2026-06-17 17:26:44,794][starVLA.training.train_starvla][INFO] - Step 409, Loss: {'train/loss': 0.6745108962059021, 'train/grad_norm_pre_clip': 4.498972415924072, 'timing/data': 0.029655613005161285, 'timing/model': 0.9520173026248813, 'learning_rate/qwen_vl_interface': 9.953068366058133e-06, 'learning_rate/action_model': 9.953068366058133e-05, 'epoch': 0.19}) +[2026-06-17 17:26:49,204][starVLA.training.train_starvla][INFO] - Step 410, Loss: {'train/loss': 0.8237974643707275, 'train/grad_norm_pre_clip': 4.584784984588623, 'timing/data': 0.03007110022008419, 'timing/model': 1.325450192205608, 'learning_rate/qwen_vl_interface': 9.952764615179813e-06, 'learning_rate/action_model': 9.952764615179812e-05, 'epoch': 0.19}) +[2026-06-17 17:26:52,983][starVLA.training.train_starvla][INFO] - Step 411, Loss: {'train/loss': 0.6792530417442322, 'train/grad_norm_pre_clip': 5.211867332458496, 'timing/data': 0.029096905142068863, 'timing/model': 1.108970164321363, 'learning_rate/qwen_vl_interface': 9.95245988941333e-06, 'learning_rate/action_model': 9.95245988941333e-05, 'epoch': 0.19}) +[2026-06-17 17:26:56,740][starVLA.training.train_starvla][INFO] - Step 412, Loss: {'train/loss': 0.9469475150108337, 'train/grad_norm_pre_clip': 5.82603645324707, 'timing/data': 0.034845915623009205, 'timing/model': 1.101806695573032, 'learning_rate/qwen_vl_interface': 9.952154188821852e-06, 'learning_rate/action_model': 9.952154188821851e-05, 'epoch': 0.19}) +[2026-06-17 17:27:00,684][starVLA.training.train_starvla][INFO] - Step 413, Loss: {'train/loss': 0.6379960179328918, 'train/grad_norm_pre_clip': 4.7788896560668945, 'timing/data': 0.03962336853146553, 'timing/model': 1.08694355096668, 'learning_rate/qwen_vl_interface': 9.95184751346875e-06, 'learning_rate/action_model': 9.95184751346875e-05, 'epoch': 0.19}) +[2026-06-17 17:27:04,735][starVLA.training.train_starvla][INFO] - Step 414, Loss: {'train/loss': 0.7020741105079651, 'train/grad_norm_pre_clip': 4.132023334503174, 'timing/data': 0.03764025308191776, 'timing/model': 0.9273822857066989, 'learning_rate/qwen_vl_interface': 9.951539863417598e-06, 'learning_rate/action_model': 9.951539863417599e-05, 'epoch': 0.19}) +[2026-06-17 17:27:08,482][starVLA.training.train_starvla][INFO] - Step 415, Loss: {'train/loss': 0.6873689889907837, 'train/grad_norm_pre_clip': 4.180981159210205, 'timing/data': 0.029880760237574577, 'timing/model': 0.9243769366294146, 'learning_rate/qwen_vl_interface': 9.951231238732174e-06, 'learning_rate/action_model': 9.951231238732173e-05, 'epoch': 0.19}) +[2026-06-17 17:27:12,384][starVLA.training.train_starvla][INFO] - Step 416, Loss: {'train/loss': 0.7362355589866638, 'train/grad_norm_pre_clip': 5.429872035980225, 'timing/data': 0.03697365615516901, 'timing/model': 0.9780452521517873, 'learning_rate/qwen_vl_interface': 9.950921639476455e-06, 'learning_rate/action_model': 9.950921639476455e-05, 'epoch': 0.19}) +[2026-06-17 17:27:16,139][starVLA.training.train_starvla][INFO] - Step 417, Loss: {'train/loss': 0.6163225173950195, 'train/grad_norm_pre_clip': 6.14151668548584, 'timing/data': 0.03028641641139984, 'timing/model': 1.1088202651590109, 'learning_rate/qwen_vl_interface': 9.950611065714623e-06, 'learning_rate/action_model': 9.950611065714623e-05, 'epoch': 0.19}) +[2026-06-17 17:27:20,045][starVLA.training.train_starvla][INFO] - Step 418, Loss: {'train/loss': 0.5774702429771423, 'train/grad_norm_pre_clip': 2.9436538219451904, 'timing/data': 0.03792892396450043, 'timing/model': 0.9141017021611333, 'learning_rate/qwen_vl_interface': 9.950299517511058e-06, 'learning_rate/action_model': 9.950299517511058e-05, 'epoch': 0.19}) +[2026-06-17 17:27:23,853][starVLA.training.train_starvla][INFO] - Step 419, Loss: {'train/loss': 0.8236280679702759, 'train/grad_norm_pre_clip': 6.358705520629883, 'timing/data': 0.03295441437512636, 'timing/model': 0.9388602739199996, 'learning_rate/qwen_vl_interface': 9.949986994930346e-06, 'learning_rate/action_model': 9.949986994930346e-05, 'epoch': 0.19}) +[2026-06-17 17:27:27,750][starVLA.training.train_starvla][INFO] - Step 420, Loss: {'train/loss': 0.7185373902320862, 'train/grad_norm_pre_clip': 3.8607840538024902, 'timing/data': 0.03741237334907055, 'timing/model': 0.959916546009481, 'learning_rate/qwen_vl_interface': 9.949673498037272e-06, 'learning_rate/action_model': 9.949673498037272e-05, 'epoch': 0.19}) +[2026-06-17 17:27:31,695][starVLA.training.train_starvla][INFO] - Step 421, Loss: {'train/loss': 1.4242292642593384, 'train/grad_norm_pre_clip': 6.077369689941406, 'timing/data': 0.03529071621596813, 'timing/model': 0.93898905813694, 'learning_rate/qwen_vl_interface': 9.949359026896826e-06, 'learning_rate/action_model': 9.949359026896826e-05, 'epoch': 0.19}) +[2026-06-17 17:27:35,702][starVLA.training.train_starvla][INFO] - Step 422, Loss: {'train/loss': 0.44004616141319275, 'train/grad_norm_pre_clip': 2.7649245262145996, 'timing/data': 0.029251674190163612, 'timing/model': 0.9709891332313418, 'learning_rate/qwen_vl_interface': 9.949043581574196e-06, 'learning_rate/action_model': 9.949043581574196e-05, 'epoch': 0.19}) +[2026-06-17 17:27:39,806][starVLA.training.train_starvla][INFO] - Step 423, Loss: {'train/loss': 0.642599880695343, 'train/grad_norm_pre_clip': 3.6868767738342285, 'timing/data': 0.0365588441491127, 'timing/model': 1.1648134011775255, 'learning_rate/qwen_vl_interface': 9.948727162134775e-06, 'learning_rate/action_model': 9.948727162134775e-05, 'epoch': 0.19}) +[2026-06-17 17:27:43,725][starVLA.training.train_starvla][INFO] - Step 424, Loss: {'train/loss': 0.8697016835212708, 'train/grad_norm_pre_clip': 4.210402011871338, 'timing/data': 0.029556014575064182, 'timing/model': 1.2053766455501318, 'learning_rate/qwen_vl_interface': 9.948409768644157e-06, 'learning_rate/action_model': 9.948409768644156e-05, 'epoch': 0.19}) +[2026-06-17 17:27:47,436][starVLA.training.train_starvla][INFO] - Step 425, Loss: {'train/loss': 0.7549946904182434, 'train/grad_norm_pre_clip': 5.374624252319336, 'timing/data': 0.03572412393987179, 'timing/model': 0.975761896930635, 'learning_rate/qwen_vl_interface': 9.948091401168139e-06, 'learning_rate/action_model': 9.948091401168138e-05, 'epoch': 0.19}) +[2026-06-17 17:27:51,340][starVLA.training.train_starvla][INFO] - Step 426, Loss: {'train/loss': 0.6452598571777344, 'train/grad_norm_pre_clip': 4.611030578613281, 'timing/data': 0.03153874911367893, 'timing/model': 0.9698656881228089, 'learning_rate/qwen_vl_interface': 9.947772059772718e-06, 'learning_rate/action_model': 9.947772059772717e-05, 'epoch': 0.19}) +[2026-06-17 17:27:55,241][starVLA.training.train_starvla][INFO] - Step 427, Loss: {'train/loss': 0.7611490488052368, 'train/grad_norm_pre_clip': 6.017287254333496, 'timing/data': 0.04239284247159958, 'timing/model': 0.9689630921930075, 'learning_rate/qwen_vl_interface': 9.947451744524094e-06, 'learning_rate/action_model': 9.947451744524093e-05, 'epoch': 0.19}) +[2026-06-17 17:27:59,273][starVLA.training.train_starvla][INFO] - Step 428, Loss: {'train/loss': 1.152490496635437, 'train/grad_norm_pre_clip': 7.309767246246338, 'timing/data': 0.03146904334425926, 'timing/model': 0.9628229020163417, 'learning_rate/qwen_vl_interface': 9.947130455488668e-06, 'learning_rate/action_model': 9.947130455488666e-05, 'epoch': 0.19}) +[2026-06-17 17:28:03,182][starVLA.training.train_starvla][INFO] - Step 429, Loss: {'train/loss': 0.6680091619491577, 'train/grad_norm_pre_clip': 4.178415298461914, 'timing/data': 0.04306782968342304, 'timing/model': 0.9930282793939114, 'learning_rate/qwen_vl_interface': 9.946808192733045e-06, 'learning_rate/action_model': 9.946808192733044e-05, 'epoch': 0.19}) +[2026-06-17 17:28:07,182][starVLA.training.train_starvla][INFO] - Step 430, Loss: {'train/loss': 1.1246800422668457, 'train/grad_norm_pre_clip': 8.219797134399414, 'timing/data': 0.03318210784345865, 'timing/model': 1.1746765691787004, 'learning_rate/qwen_vl_interface': 9.946484956324028e-06, 'learning_rate/action_model': 9.946484956324028e-05, 'epoch': 0.2}) +[2026-06-17 17:28:11,043][starVLA.training.train_starvla][INFO] - Step 431, Loss: {'train/loss': 0.8172895312309265, 'train/grad_norm_pre_clip': 4.308009147644043, 'timing/data': 0.03731259144842625, 'timing/model': 1.1180056985467672, 'learning_rate/qwen_vl_interface': 9.946160746328628e-06, 'learning_rate/action_model': 9.946160746328627e-05, 'epoch': 0.2}) +[2026-06-17 17:28:14,962][starVLA.training.train_starvla][INFO] - Step 432, Loss: {'train/loss': 0.6945856809616089, 'train/grad_norm_pre_clip': 13.096145629882812, 'timing/data': 0.041644202545285225, 'timing/model': 0.9637267896905541, 'learning_rate/qwen_vl_interface': 9.945835562814052e-06, 'learning_rate/action_model': 9.945835562814051e-05, 'epoch': 0.2}) +[2026-06-17 17:28:18,773][starVLA.training.train_starvla][INFO] - Step 433, Loss: {'train/loss': 1.0182991027832031, 'train/grad_norm_pre_clip': 5.38156795501709, 'timing/data': 0.03433513455092907, 'timing/model': 0.917589234188199, 'learning_rate/qwen_vl_interface': 9.94550940584771e-06, 'learning_rate/action_model': 9.945509405847709e-05, 'epoch': 0.2}) +[2026-06-17 17:28:22,570][starVLA.training.train_starvla][INFO] - Step 434, Loss: {'train/loss': 0.8413509130477905, 'train/grad_norm_pre_clip': 9.453369140625, 'timing/data': 0.04305427800863981, 'timing/model': 0.9355884743854403, 'learning_rate/qwen_vl_interface': 9.945182275497215e-06, 'learning_rate/action_model': 9.945182275497215e-05, 'epoch': 0.2}) +[2026-06-17 17:28:26,470][starVLA.training.train_starvla][INFO] - Step 435, Loss: {'train/loss': 0.6535040140151978, 'train/grad_norm_pre_clip': 4.6538286209106445, 'timing/data': 0.05154014192521572, 'timing/model': 0.9560294905677438, 'learning_rate/qwen_vl_interface': 9.944854171830382e-06, 'learning_rate/action_model': 9.944854171830383e-05, 'epoch': 0.2}) +[2026-06-17 17:28:30,396][starVLA.training.train_starvla][INFO] - Step 436, Loss: {'train/loss': 0.7675542831420898, 'train/grad_norm_pre_clip': 5.596768856048584, 'timing/data': 0.06469852197915316, 'timing/model': 0.9197567272931337, 'learning_rate/qwen_vl_interface': 9.94452509491523e-06, 'learning_rate/action_model': 9.94452509491523e-05, 'epoch': 0.2}) +[2026-06-17 17:28:34,291][starVLA.training.train_starvla][INFO] - Step 437, Loss: {'train/loss': 0.7688392996788025, 'train/grad_norm_pre_clip': 5.405617713928223, 'timing/data': 0.04771342780441046, 'timing/model': 1.197605318389833, 'learning_rate/qwen_vl_interface': 9.944195044819973e-06, 'learning_rate/action_model': 9.944195044819972e-05, 'epoch': 0.2}) +[2026-06-17 17:28:37,958][starVLA.training.train_starvla][INFO] - Step 438, Loss: {'train/loss': 0.7586554884910583, 'train/grad_norm_pre_clip': 3.733948230743408, 'timing/data': 0.03144486900418997, 'timing/model': 0.9727051146328449, 'learning_rate/qwen_vl_interface': 9.94386402161303e-06, 'learning_rate/action_model': 9.943864021613029e-05, 'epoch': 0.2}) +[2026-06-17 17:28:41,750][starVLA.training.train_starvla][INFO] - Step 439, Loss: {'train/loss': 0.569461464881897, 'train/grad_norm_pre_clip': 4.287115097045898, 'timing/data': 0.03195943310856819, 'timing/model': 0.9393262146040797, 'learning_rate/qwen_vl_interface': 9.943532025363026e-06, 'learning_rate/action_model': 9.943532025363025e-05, 'epoch': 0.2}) +[2026-06-17 17:28:45,730][starVLA.training.train_starvla][INFO] - Step 440, Loss: {'train/loss': 0.7714090347290039, 'train/grad_norm_pre_clip': 4.842307090759277, 'timing/data': 0.04829521384090185, 'timing/model': 0.9250934682786465, 'learning_rate/qwen_vl_interface': 9.943199056138781e-06, 'learning_rate/action_model': 9.943199056138782e-05, 'epoch': 0.2}) +[2026-06-17 17:28:49,687][starVLA.training.train_starvla][INFO] - Step 441, Loss: {'train/loss': 0.6394466757774353, 'train/grad_norm_pre_clip': 3.412909984588623, 'timing/data': 0.04105713963508606, 'timing/model': 0.9341903990134597, 'learning_rate/qwen_vl_interface': 9.942865114009325e-06, 'learning_rate/action_model': 9.942865114009324e-05, 'epoch': 0.2}) +[2026-06-17 17:28:53,854][starVLA.training.train_starvla][INFO] - Step 442, Loss: {'train/loss': 0.49604979157447815, 'train/grad_norm_pre_clip': 4.319591045379639, 'timing/data': 0.03824280668050051, 'timing/model': 0.9396830471232533, 'learning_rate/qwen_vl_interface': 9.94253019904388e-06, 'learning_rate/action_model': 9.942530199043879e-05, 'epoch': 0.2}) +[2026-06-17 17:28:57,647][starVLA.training.train_starvla][INFO] - Step 443, Loss: {'train/loss': 0.7567146420478821, 'train/grad_norm_pre_clip': 4.504629135131836, 'timing/data': 0.048425992019474506, 'timing/model': 0.925147489644587, 'learning_rate/qwen_vl_interface': 9.942194311311874e-06, 'learning_rate/action_model': 9.942194311311874e-05, 'epoch': 0.2}) +[2026-06-17 17:29:01,521][starVLA.training.train_starvla][INFO] - Step 444, Loss: {'train/loss': 0.5540914535522461, 'train/grad_norm_pre_clip': 4.1840033531188965, 'timing/data': 0.04203930962830782, 'timing/model': 1.1309764329344034, 'learning_rate/qwen_vl_interface': 9.941857450882939e-06, 'learning_rate/action_model': 9.941857450882937e-05, 'epoch': 0.2}) +[2026-06-17 17:29:05,737][starVLA.training.train_starvla][INFO] - Step 445, Loss: {'train/loss': 0.8765805959701538, 'train/grad_norm_pre_clip': 5.039889812469482, 'timing/data': 0.04121503420174122, 'timing/model': 1.4771166816353798, 'learning_rate/qwen_vl_interface': 9.941519617826906e-06, 'learning_rate/action_model': 9.941519617826906e-05, 'epoch': 0.2}) +[2026-06-17 17:29:09,590][starVLA.training.train_starvla][INFO] - Step 446, Loss: {'train/loss': 0.5660098791122437, 'train/grad_norm_pre_clip': 4.0954999923706055, 'timing/data': 0.048812770284712315, 'timing/model': 0.9524769866839051, 'learning_rate/qwen_vl_interface': 9.941180812213807e-06, 'learning_rate/action_model': 9.941180812213806e-05, 'epoch': 0.2}) +[2026-06-17 17:29:13,356][starVLA.training.train_starvla][INFO] - Step 447, Loss: {'train/loss': 0.908157467842102, 'train/grad_norm_pre_clip': 4.775209903717041, 'timing/data': 0.03133551590144634, 'timing/model': 0.9226698968559504, 'learning_rate/qwen_vl_interface': 9.940841034113878e-06, 'learning_rate/action_model': 9.940841034113877e-05, 'epoch': 0.2}) +[2026-06-17 17:29:17,289][starVLA.training.train_starvla][INFO] - Step 448, Loss: {'train/loss': 0.7691943645477295, 'train/grad_norm_pre_clip': 5.07075834274292, 'timing/data': 0.04043281264603138, 'timing/model': 1.0361457839608192, 'learning_rate/qwen_vl_interface': 9.940500283597557e-06, 'learning_rate/action_model': 9.940500283597556e-05, 'epoch': 0.2}) +[2026-06-17 17:29:21,113][starVLA.training.train_starvla][INFO] - Step 449, Loss: {'train/loss': 0.8286601305007935, 'train/grad_norm_pre_clip': 5.5830864906311035, 'timing/data': 0.04156787972897291, 'timing/model': 0.9187833080068231, 'learning_rate/qwen_vl_interface': 9.940158560735478e-06, 'learning_rate/action_model': 9.940158560735477e-05, 'epoch': 0.2}) +[2026-06-17 17:29:24,860][starVLA.training.train_starvla][INFO] - Step 450, Loss: {'train/loss': 0.7178119421005249, 'train/grad_norm_pre_clip': 7.815565586090088, 'timing/data': 0.03693009912967682, 'timing/model': 0.9212744450196624, 'learning_rate/qwen_vl_interface': 9.939815865598482e-06, 'learning_rate/action_model': 9.93981586559848e-05, 'epoch': 0.2}) +[2026-06-17 17:29:28,597][starVLA.training.train_starvla][INFO] - Step 451, Loss: {'train/loss': 0.7329911589622498, 'train/grad_norm_pre_clip': 5.173798084259033, 'timing/data': 0.038439005613327026, 'timing/model': 1.0941686453297734, 'learning_rate/qwen_vl_interface': 9.939472198257611e-06, 'learning_rate/action_model': 9.93947219825761e-05, 'epoch': 0.2}) +[2026-06-17 17:29:32,158][starVLA.training.train_starvla][INFO] - Step 452, Loss: {'train/loss': 0.7692018151283264, 'train/grad_norm_pre_clip': 6.546518325805664, 'timing/data': 0.03382633626461029, 'timing/model': 0.9220614247024059, 'learning_rate/qwen_vl_interface': 9.939127558784108e-06, 'learning_rate/action_model': 9.939127558784107e-05, 'epoch': 0.21}) +[2026-06-17 17:29:36,012][starVLA.training.train_starvla][INFO] - Step 453, Loss: {'train/loss': 0.9483132362365723, 'train/grad_norm_pre_clip': 6.448631286621094, 'timing/data': 0.04844248294830322, 'timing/model': 1.0018398575484753, 'learning_rate/qwen_vl_interface': 9.938781947249417e-06, 'learning_rate/action_model': 9.938781947249416e-05, 'epoch': 0.21}) +[2026-06-17 17:29:39,887][starVLA.training.train_starvla][INFO] - Step 454, Loss: {'train/loss': 0.5356472730636597, 'train/grad_norm_pre_clip': 5.0748162269592285, 'timing/data': 0.03825025353580713, 'timing/model': 0.9499963782727718, 'learning_rate/qwen_vl_interface': 9.93843536372518e-06, 'learning_rate/action_model': 9.93843536372518e-05, 'epoch': 0.21}) +[2026-06-17 17:29:43,897][starVLA.training.train_starvla][INFO] - Step 455, Loss: {'train/loss': 0.9528737664222717, 'train/grad_norm_pre_clip': 6.223448753356934, 'timing/data': 0.033821647986769676, 'timing/model': 1.013457590714097, 'learning_rate/qwen_vl_interface': 9.938087808283248e-06, 'learning_rate/action_model': 9.938087808283248e-05, 'epoch': 0.21}) +[2026-06-17 17:29:47,729][starVLA.training.train_starvla][INFO] - Step 456, Loss: {'train/loss': 0.8048961162567139, 'train/grad_norm_pre_clip': 14.620152473449707, 'timing/data': 0.029971505515277386, 'timing/model': 0.9311868753284216, 'learning_rate/qwen_vl_interface': 9.937739280995672e-06, 'learning_rate/action_model': 9.937739280995672e-05, 'epoch': 0.21}) +[2026-06-17 17:29:51,691][starVLA.training.train_starvla][INFO] - Step 457, Loss: {'train/loss': 0.6727051734924316, 'train/grad_norm_pre_clip': 4.537511825561523, 'timing/data': 0.03802405949681997, 'timing/model': 0.9415349187329412, 'learning_rate/qwen_vl_interface': 9.937389781934697e-06, 'learning_rate/action_model': 9.937389781934696e-05, 'epoch': 0.21}) +[2026-06-17 17:29:55,630][starVLA.training.train_starvla][INFO] - Step 458, Loss: {'train/loss': 0.8074131011962891, 'train/grad_norm_pre_clip': 4.859424591064453, 'timing/data': 0.043039931915700436, 'timing/model': 1.1985668633133173, 'learning_rate/qwen_vl_interface': 9.937039311172776e-06, 'learning_rate/action_model': 9.937039311172776e-05, 'epoch': 0.21}) +[2026-06-17 17:29:59,390][starVLA.training.train_starvla][INFO] - Step 459, Loss: {'train/loss': 0.8903455138206482, 'train/grad_norm_pre_clip': 9.588226318359375, 'timing/data': 0.0344678470864892, 'timing/model': 1.067989938892424, 'learning_rate/qwen_vl_interface': 9.936687868782562e-06, 'learning_rate/action_model': 9.936687868782561e-05, 'epoch': 0.21}) +[2026-06-17 17:30:03,699][starVLA.training.train_starvla][INFO] - Step 460, Loss: {'train/loss': 0.9315411448478699, 'train/grad_norm_pre_clip': 6.088647365570068, 'timing/data': 0.04523774143308401, 'timing/model': 0.9558549961075187, 'learning_rate/qwen_vl_interface': 9.93633545483691e-06, 'learning_rate/action_model': 9.93633545483691e-05, 'epoch': 0.21}) +[2026-06-17 17:30:07,728][starVLA.training.train_starvla][INFO] - Step 461, Loss: {'train/loss': 1.0218989849090576, 'train/grad_norm_pre_clip': 7.955115795135498, 'timing/data': 0.03828148916363716, 'timing/model': 1.0907315015792847, 'learning_rate/qwen_vl_interface': 9.935982069408877e-06, 'learning_rate/action_model': 9.935982069408877e-05, 'epoch': 0.21}) +[2026-06-17 17:30:12,643][starVLA.training.train_starvla][INFO] - Step 462, Loss: {'train/loss': 1.0277544260025024, 'train/grad_norm_pre_clip': 5.03715705871582, 'timing/data': 0.038149282336235046, 'timing/model': 1.0115727577358484, 'learning_rate/qwen_vl_interface': 9.935627712571716e-06, 'learning_rate/action_model': 9.935627712571715e-05, 'epoch': 0.21}) +[2026-06-17 17:30:16,606][starVLA.training.train_starvla][INFO] - Step 463, Loss: {'train/loss': 0.9900122880935669, 'train/grad_norm_pre_clip': 15.479377746582031, 'timing/data': 0.04096699878573418, 'timing/model': 1.0145415691658854, 'learning_rate/qwen_vl_interface': 9.93527238439889e-06, 'learning_rate/action_model': 9.935272384398889e-05, 'epoch': 0.21}) +[2026-06-17 17:30:20,478][starVLA.training.train_starvla][INFO] - Step 464, Loss: {'train/loss': 0.9021983742713928, 'train/grad_norm_pre_clip': 7.226650238037109, 'timing/data': 0.04073507059365511, 'timing/model': 1.1410120707005262, 'learning_rate/qwen_vl_interface': 9.934916084964056e-06, 'learning_rate/action_model': 9.934916084964056e-05, 'epoch': 0.21}) +[2026-06-17 17:30:24,361][starVLA.training.train_starvla][INFO] - Step 465, Loss: {'train/loss': 0.7235106825828552, 'train/grad_norm_pre_clip': 4.5958571434021, 'timing/data': 0.03822650574147701, 'timing/model': 1.1489618327468634, 'learning_rate/qwen_vl_interface': 9.934558814341077e-06, 'learning_rate/action_model': 9.934558814341077e-05, 'epoch': 0.21}) +[2026-06-17 17:30:27,981][starVLA.training.train_starvla][INFO] - Step 466, Loss: {'train/loss': 0.611947774887085, 'train/grad_norm_pre_clip': 5.792108058929443, 'timing/data': 0.04226187337189913, 'timing/model': 0.9269887516275048, 'learning_rate/qwen_vl_interface': 9.934200572604015e-06, 'learning_rate/action_model': 9.934200572604015e-05, 'epoch': 0.21}) +[2026-06-17 17:30:31,806][starVLA.training.train_starvla][INFO] - Step 467, Loss: {'train/loss': 0.8215526938438416, 'train/grad_norm_pre_clip': 5.567013740539551, 'timing/data': 0.04788574110716581, 'timing/model': 0.9237459767609835, 'learning_rate/qwen_vl_interface': 9.933841359827133e-06, 'learning_rate/action_model': 9.933841359827133e-05, 'epoch': 0.21}) +[2026-06-17 17:30:35,641][starVLA.training.train_starvla][INFO] - Step 468, Loss: {'train/loss': 0.6639488935470581, 'train/grad_norm_pre_clip': 4.35675048828125, 'timing/data': 0.043643650598824024, 'timing/model': 0.9258408937603235, 'learning_rate/qwen_vl_interface': 9.933481176084899e-06, 'learning_rate/action_model': 9.933481176084899e-05, 'epoch': 0.21}) +[2026-06-17 17:30:39,536][starVLA.training.train_starvla][INFO] - Step 469, Loss: {'train/loss': 0.8460965156555176, 'train/grad_norm_pre_clip': 6.094531059265137, 'timing/data': 0.034358629025518894, 'timing/model': 0.9442650284618139, 'learning_rate/qwen_vl_interface': 9.933120021451976e-06, 'learning_rate/action_model': 9.933120021451976e-05, 'epoch': 0.21}) +[2026-06-17 17:30:43,363][starVLA.training.train_starvla][INFO] - Step 470, Loss: {'train/loss': 0.9673988223075867, 'train/grad_norm_pre_clip': 4.996490001678467, 'timing/data': 0.0405094139277935, 'timing/model': 0.9334905846044421, 'learning_rate/qwen_vl_interface': 9.932757896003235e-06, 'learning_rate/action_model': 9.932757896003235e-05, 'epoch': 0.21}) +[2026-06-17 17:30:47,412][starVLA.training.train_starvla][INFO] - Step 471, Loss: {'train/loss': 0.991205632686615, 'train/grad_norm_pre_clip': 6.863566875457764, 'timing/data': 0.05156328342854977, 'timing/model': 1.101862638257444, 'learning_rate/qwen_vl_interface': 9.932394799813742e-06, 'learning_rate/action_model': 9.932394799813741e-05, 'epoch': 0.21}) +[2026-06-17 17:30:51,107][starVLA.training.train_starvla][INFO] - Step 472, Loss: {'train/loss': 1.0196607112884521, 'train/grad_norm_pre_clip': 5.614438056945801, 'timing/data': 0.0401200270280242, 'timing/model': 0.929660283960402, 'learning_rate/qwen_vl_interface': 9.932030732958768e-06, 'learning_rate/action_model': 9.932030732958769e-05, 'epoch': 0.21}) +[2026-06-17 17:30:54,934][starVLA.training.train_starvla][INFO] - Step 473, Loss: {'train/loss': 0.6764654517173767, 'train/grad_norm_pre_clip': 4.531271934509277, 'timing/data': 0.044229683466255665, 'timing/model': 0.9323750976473093, 'learning_rate/qwen_vl_interface': 9.931665695513786e-06, 'learning_rate/action_model': 9.931665695513786e-05, 'epoch': 0.21}) +[2026-06-17 17:30:58,738][starVLA.training.train_starvla][INFO] - Step 474, Loss: {'train/loss': 0.5390903949737549, 'train/grad_norm_pre_clip': 5.2397284507751465, 'timing/data': 0.046637705527246, 'timing/model': 0.9237894555553794, 'learning_rate/qwen_vl_interface': 9.931299687554466e-06, 'learning_rate/action_model': 9.931299687554466e-05, 'epoch': 0.22}) +[2026-06-17 17:31:02,541][starVLA.training.train_starvla][INFO] - Step 475, Loss: {'train/loss': 0.5596988201141357, 'train/grad_norm_pre_clip': 5.372400283813477, 'timing/data': 0.03946536127477884, 'timing/model': 0.9347472647204995, 'learning_rate/qwen_vl_interface': 9.930932709156686e-06, 'learning_rate/action_model': 9.930932709156685e-05, 'epoch': 0.22}) +[2026-06-17 17:31:06,426][starVLA.training.train_starvla][INFO] - Step 476, Loss: {'train/loss': 0.7910317182540894, 'train/grad_norm_pre_clip': 6.875648021697998, 'timing/data': 0.02989883255213499, 'timing/model': 0.9994534309953451, 'learning_rate/qwen_vl_interface': 9.930564760396518e-06, 'learning_rate/action_model': 9.930564760396518e-05, 'epoch': 0.22}) +[2026-06-17 17:31:10,342][starVLA.training.train_starvla][INFO] - Step 477, Loss: {'train/loss': 0.5902106165885925, 'train/grad_norm_pre_clip': 6.801040172576904, 'timing/data': 0.05828937795013189, 'timing/model': 0.9305731980130076, 'learning_rate/qwen_vl_interface': 9.930195841350237e-06, 'learning_rate/action_model': 9.930195841350236e-05, 'epoch': 0.22}) +[2026-06-17 17:31:14,717][starVLA.training.train_starvla][INFO] - Step 478, Loss: {'train/loss': 0.911474347114563, 'train/grad_norm_pre_clip': 7.477874755859375, 'timing/data': 0.03987873438745737, 'timing/model': 1.41075630299747, 'learning_rate/qwen_vl_interface': 9.929825952094323e-06, 'learning_rate/action_model': 9.929825952094323e-05, 'epoch': 0.22}) +[2026-06-17 17:31:18,722][starVLA.training.train_starvla][INFO] - Step 479, Loss: {'train/loss': 1.1860040426254272, 'train/grad_norm_pre_clip': 7.742685794830322, 'timing/data': 0.04099519923329353, 'timing/model': 0.9467692822217941, 'learning_rate/qwen_vl_interface': 9.929455092705453e-06, 'learning_rate/action_model': 9.929455092705453e-05, 'epoch': 0.22}) +[2026-06-17 17:31:22,697][starVLA.training.train_starvla][INFO] - Step 480, Loss: {'train/loss': 0.6410323977470398, 'train/grad_norm_pre_clip': 6.338620662689209, 'timing/data': 0.03750328999012709, 'timing/model': 1.0354947466403246, 'learning_rate/qwen_vl_interface': 9.929083263260508e-06, 'learning_rate/action_model': 9.929083263260507e-05, 'epoch': 0.22}) +[2026-06-17 17:31:27,226][starVLA.training.train_starvla][INFO] - Step 481, Loss: {'train/loss': 0.5773641467094421, 'train/grad_norm_pre_clip': 5.759979724884033, 'timing/data': 0.037984734401106834, 'timing/model': 1.3212547283619642, 'learning_rate/qwen_vl_interface': 9.928710463836566e-06, 'learning_rate/action_model': 9.928710463836565e-05, 'epoch': 0.22}) +[2026-06-17 17:31:31,541][starVLA.training.train_starvla][INFO] - Step 482, Loss: {'train/loss': 1.007555603981018, 'train/grad_norm_pre_clip': 8.560484886169434, 'timing/data': 0.04286307841539383, 'timing/model': 0.9476488707587123, 'learning_rate/qwen_vl_interface': 9.928336694510912e-06, 'learning_rate/action_model': 9.928336694510911e-05, 'epoch': 0.22}) +[2026-06-17 17:31:35,434][starVLA.training.train_starvla][INFO] - Step 483, Loss: {'train/loss': 0.7177924513816833, 'train/grad_norm_pre_clip': 6.016103267669678, 'timing/data': 0.034464484080672264, 'timing/model': 0.9365608254447579, 'learning_rate/qwen_vl_interface': 9.927961955361025e-06, 'learning_rate/action_model': 9.927961955361025e-05, 'epoch': 0.22}) +[2026-06-17 17:31:39,455][starVLA.training.train_starvla][INFO] - Step 484, Loss: {'train/loss': 0.7577826380729675, 'train/grad_norm_pre_clip': 5.997891902923584, 'timing/data': 0.03899056185036898, 'timing/model': 0.94691411126405, 'learning_rate/qwen_vl_interface': 9.927586246464593e-06, 'learning_rate/action_model': 9.927586246464593e-05, 'epoch': 0.22}) +[2026-06-17 17:31:43,674][starVLA.training.train_starvla][INFO] - Step 485, Loss: {'train/loss': 1.076748251914978, 'train/grad_norm_pre_clip': 7.512033462524414, 'timing/data': 0.03967588394880295, 'timing/model': 1.3603175301104784, 'learning_rate/qwen_vl_interface': 9.927209567899499e-06, 'learning_rate/action_model': 9.927209567899498e-05, 'epoch': 0.22}) +[2026-06-17 17:31:52,066][starVLA.training.train_starvla][INFO] - Step 486, Loss: {'train/loss': 0.9990481734275818, 'train/grad_norm_pre_clip': 7.061592102050781, 'timing/data': 0.050193917006254196, 'timing/model': 1.065041652880609, 'learning_rate/qwen_vl_interface': 9.926831919743827e-06, 'learning_rate/action_model': 9.926831919743827e-05, 'epoch': 0.22}) +[2026-06-17 17:31:55,858][starVLA.training.train_starvla][INFO] - Step 487, Loss: {'train/loss': 0.871735155582428, 'train/grad_norm_pre_clip': 5.84354305267334, 'timing/data': 0.037341320887207985, 'timing/model': 0.9196862764656544, 'learning_rate/qwen_vl_interface': 9.926453302075866e-06, 'learning_rate/action_model': 9.926453302075867e-05, 'epoch': 0.22}) +[2026-06-17 17:31:59,582][starVLA.training.train_starvla][INFO] - Step 488, Loss: {'train/loss': 0.6685148477554321, 'train/grad_norm_pre_clip': 10.506103515625, 'timing/data': 0.03606363944709301, 'timing/model': 0.9159593367949128, 'learning_rate/qwen_vl_interface': 9.926073714974103e-06, 'learning_rate/action_model': 9.926073714974104e-05, 'epoch': 0.22}) +[2026-06-17 17:32:03,483][starVLA.training.train_starvla][INFO] - Step 489, Loss: {'train/loss': 0.7541882395744324, 'train/grad_norm_pre_clip': 8.745824813842773, 'timing/data': 0.03334754891693592, 'timing/model': 0.9386292397975922, 'learning_rate/qwen_vl_interface': 9.925693158517228e-06, 'learning_rate/action_model': 9.925693158517228e-05, 'epoch': 0.22}) +[2026-06-17 17:32:07,409][starVLA.training.train_starvla][INFO] - Step 490, Loss: {'train/loss': 0.6715680956840515, 'train/grad_norm_pre_clip': 7.652861595153809, 'timing/data': 0.033940987661480904, 'timing/model': 0.9602126814424992, 'learning_rate/qwen_vl_interface': 9.92531163278413e-06, 'learning_rate/action_model': 9.92531163278413e-05, 'epoch': 0.22}) +[2026-06-17 17:32:11,228][starVLA.training.train_starvla][INFO] - Step 491, Loss: {'train/loss': 0.8362925052642822, 'train/grad_norm_pre_clip': 7.356116771697998, 'timing/data': 0.03079922031611204, 'timing/model': 1.1798216048628092, 'learning_rate/qwen_vl_interface': 9.924929137853899e-06, 'learning_rate/action_model': 9.924929137853898e-05, 'epoch': 0.22}) +[2026-06-17 17:32:15,232][starVLA.training.train_starvla][INFO] - Step 492, Loss: {'train/loss': 0.6689654588699341, 'train/grad_norm_pre_clip': 8.415125846862793, 'timing/data': 0.04484341852366924, 'timing/model': 1.1191406268626451, 'learning_rate/qwen_vl_interface': 9.92454567380583e-06, 'learning_rate/action_model': 9.924545673805829e-05, 'epoch': 0.22}) +[2026-06-17 17:32:18,810][starVLA.training.train_starvla][INFO] - Step 493, Loss: {'train/loss': 1.118538498878479, 'train/grad_norm_pre_clip': 6.733096122741699, 'timing/data': 0.036823016591370106, 'timing/model': 0.9297899715602398, 'learning_rate/qwen_vl_interface': 9.924161240719409e-06, 'learning_rate/action_model': 9.924161240719408e-05, 'epoch': 0.22}) +[2026-06-17 17:32:22,646][starVLA.training.train_starvla][INFO] - Step 494, Loss: {'train/loss': 0.6819413900375366, 'train/grad_norm_pre_clip': 9.327770233154297, 'timing/data': 0.025187798775732517, 'timing/model': 0.9295427426695824, 'learning_rate/qwen_vl_interface': 9.923775838674336e-06, 'learning_rate/action_model': 9.923775838674335e-05, 'epoch': 0.22}) +[2026-06-17 17:32:26,643][starVLA.training.train_starvla][INFO] - Step 495, Loss: {'train/loss': 0.9508312344551086, 'train/grad_norm_pre_clip': 8.344594955444336, 'timing/data': 0.05034577287733555, 'timing/model': 1.0573700070381165, 'learning_rate/qwen_vl_interface': 9.9233894677505e-06, 'learning_rate/action_model': 9.9233894677505e-05, 'epoch': 0.22}) +[2026-06-17 17:32:30,888][starVLA.training.train_starvla][INFO] - Step 496, Loss: {'train/loss': 0.8289564847946167, 'train/grad_norm_pre_clip': 7.361480236053467, 'timing/data': 0.12067120615392923, 'timing/model': 0.9557217545807362, 'learning_rate/qwen_vl_interface': 9.923002128028e-06, 'learning_rate/action_model': 9.923002128028e-05, 'epoch': 0.23}) +[2026-06-17 17:32:36,093][starVLA.training.train_starvla][INFO] - Step 497, Loss: {'train/loss': 0.8984991908073425, 'train/grad_norm_pre_clip': 8.447112083435059, 'timing/data': 0.013392878696322441, 'timing/model': 0.9742618296295404, 'learning_rate/qwen_vl_interface': 9.922613819587131e-06, 'learning_rate/action_model': 9.92261381958713e-05, 'epoch': 0.23}) +[2026-06-17 17:32:40,076][starVLA.training.train_starvla][INFO] - Step 498, Loss: {'train/loss': 1.172607183456421, 'train/grad_norm_pre_clip': 10.792457580566406, 'timing/data': 0.03773793950676918, 'timing/model': 1.1581230517476797, 'learning_rate/qwen_vl_interface': 9.922224542508389e-06, 'learning_rate/action_model': 9.922224542508388e-05, 'epoch': 0.23}) +[2026-06-17 17:32:43,807][starVLA.training.train_starvla][INFO] - Step 499, Loss: {'train/loss': 0.701245903968811, 'train/grad_norm_pre_clip': 7.710873126983643, 'timing/data': 0.04116189759224653, 'timing/model': 1.026972884312272, 'learning_rate/qwen_vl_interface': 9.92183429687247e-06, 'learning_rate/action_model': 9.92183429687247e-05, 'epoch': 0.23}) +[2026-06-17 17:34:47,707][starVLA.training.train_starvla][INFO] - Step 500, Loss: {'train/loss': 0.5917872786521912, 'train/grad_norm_pre_clip': 7.482654571533203, 'eval/loss': 0.7479068636894226, 'eval/latency_8/loss': 0.7479068636894226, 'eval/demon_fire_precision': 0.9012622720897616, 'eval/demon_fire_recall': 0.9981360671015843, 'eval/demon_fire_f1': 0.9472287735849058, 'eval/demon_fire_rate_model': 0.9919309961046188, 'eval/demon_fire_rate_teacher': 0.8956594323873122, 'eval/demon_left_precision': 0.7336956521739131, 'eval/demon_left_recall': 0.2890792291220557, 'eval/demon_left_f1': 0.4147465437788019, 'eval/demon_left_rate_model': 0.051196438508625486, 'eval/demon_left_rate_teacher': 0.12993878686700056, 'eval/demon_right_precision': 0.5980253878702397, 'eval/demon_right_recall': 0.6424242424242425, 'eval/demon_right_f1': 0.6194302410518626, 'eval/demon_right_rate_model': 0.19727323316638842, 'eval/demon_right_rate_teacher': 0.18363939899833054, 'eval/demon_move_f1': 0.5170883924153322, 'eval/demon_cc_f1': 0.7580527442264774, 'eval/latency_8/demon_fire_precision': 0.9012622720897616, 'eval/latency_8/demon_fire_recall': 0.9981360671015843, 'eval/latency_8/demon_fire_f1': 0.9472287735849058, 'eval/latency_8/demon_fire_rate_model': 0.9919309961046188, 'eval/latency_8/demon_fire_rate_teacher': 0.8956594323873122, 'eval/latency_8/demon_left_precision': 0.7336956521739131, 'eval/latency_8/demon_left_recall': 0.2890792291220557, 'eval/latency_8/demon_left_f1': 0.4147465437788019, 'eval/latency_8/demon_left_rate_model': 0.051196438508625486, 'eval/latency_8/demon_left_rate_teacher': 0.12993878686700056, 'eval/latency_8/demon_right_precision': 0.5980253878702397, 'eval/latency_8/demon_right_recall': 0.6424242424242425, 'eval/latency_8/demon_right_f1': 0.6194302410518626, 'eval/latency_8/demon_right_rate_model': 0.19727323316638842, 'eval/latency_8/demon_right_rate_teacher': 0.18363939899833054, 'eval/latency_8/demon_move_f1': 0.5170883924153322, 'eval/latency_8/demon_cc_f1': 0.7580527442264774, 'timing/data': 0.03176603093743324, 'timing/model': 0.9341326933354139, 'learning_rate/qwen_vl_interface': 9.921443082760276e-06, 'learning_rate/action_model': 9.921443082760276e-05, 'epoch': 0.23}) +[2026-06-17 17:34:47,720][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_500_state +[2026-06-17 17:34:47,743][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 17:34:54,986][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_500_state/model.safetensors +[2026-06-17 17:35:12,352][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_500_state/optimizer.bin +[2026-06-17 17:35:12,354][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_500_state/sampler.bin +[2026-06-17 17:35:12,354][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_500_state/sampler_1.bin +[2026-06-17 17:35:12,385][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_500_state/random_states_0.pkl +[2026-06-17 17:35:12,398][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 500 +[2026-06-17 17:35:12,398][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 17:35:12,409][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 17:35:16,262][starVLA.training.train_starvla][INFO] - Step 501, Loss: {'train/loss': 0.6248898506164551, 'train/grad_norm_pre_clip': 7.231699466705322, 'timing/data': 0.04342610388994217, 'timing/model': 1.1565159549936652, 'learning_rate/qwen_vl_interface': 9.921050900252904e-06, 'learning_rate/action_model': 9.921050900252903e-05, 'epoch': 0.23}) +[2026-06-17 17:35:19,927][starVLA.training.train_starvla][INFO] - Step 502, Loss: {'train/loss': 0.6782177686691284, 'train/grad_norm_pre_clip': 7.1470489501953125, 'timing/data': 0.036612873896956444, 'timing/model': 0.9299765918403864, 'learning_rate/qwen_vl_interface': 9.920657749431654e-06, 'learning_rate/action_model': 9.920657749431654e-05, 'epoch': 0.23}) +[2026-06-17 17:35:23,743][starVLA.training.train_starvla][INFO] - Step 503, Loss: {'train/loss': 0.7043214440345764, 'train/grad_norm_pre_clip': 6.030272960662842, 'timing/data': 0.03491188958287239, 'timing/model': 0.9343927288427949, 'learning_rate/qwen_vl_interface': 9.920263630378026e-06, 'learning_rate/action_model': 9.920263630378024e-05, 'epoch': 0.23}) +[2026-06-17 17:35:27,625][starVLA.training.train_starvla][INFO] - Step 504, Loss: {'train/loss': 1.0367780923843384, 'train/grad_norm_pre_clip': 8.037848472595215, 'timing/data': 0.04278515558689833, 'timing/model': 0.9580939598381519, 'learning_rate/qwen_vl_interface': 9.919868543173722e-06, 'learning_rate/action_model': 9.919868543173721e-05, 'epoch': 0.23}) +[2026-06-17 17:35:31,757][starVLA.training.train_starvla][INFO] - Step 505, Loss: {'train/loss': 0.674971878528595, 'train/grad_norm_pre_clip': 4.857687473297119, 'timing/data': 0.03410847391933203, 'timing/model': 0.9831294165924191, 'learning_rate/qwen_vl_interface': 9.919472487900645e-06, 'learning_rate/action_model': 9.919472487900644e-05, 'epoch': 0.23}) +[2026-06-17 17:35:35,686][starVLA.training.train_starvla][INFO] - Step 506, Loss: {'train/loss': 0.4569275975227356, 'train/grad_norm_pre_clip': 4.348249435424805, 'timing/data': 0.05042279418557882, 'timing/model': 1.2148026349022985, 'learning_rate/qwen_vl_interface': 9.919075464640895e-06, 'learning_rate/action_model': 9.919075464640895e-05, 'epoch': 0.23}) +[2026-06-17 17:35:39,445][starVLA.training.train_starvla][INFO] - Step 507, Loss: {'train/loss': 1.1872316598892212, 'train/grad_norm_pre_clip': 7.845139503479004, 'timing/data': 0.03241368755698204, 'timing/model': 1.1029564877972007, 'learning_rate/qwen_vl_interface': 9.91867747347678e-06, 'learning_rate/action_model': 9.918677473476779e-05, 'epoch': 0.23}) +[2026-06-17 17:35:43,084][starVLA.training.train_starvla][INFO] - Step 508, Loss: {'train/loss': 0.9517298340797424, 'train/grad_norm_pre_clip': 6.4457688331604, 'timing/data': 0.0371427983045578, 'timing/model': 0.922424434684217, 'learning_rate/qwen_vl_interface': 9.918278514490798e-06, 'learning_rate/action_model': 9.918278514490799e-05, 'epoch': 0.23}) +[2026-06-17 17:35:46,931][starVLA.training.train_starvla][INFO] - Step 509, Loss: {'train/loss': 1.3792678117752075, 'train/grad_norm_pre_clip': 9.709563255310059, 'timing/data': 0.035078239627182484, 'timing/model': 0.9956637909635901, 'learning_rate/qwen_vl_interface': 9.917878587765658e-06, 'learning_rate/action_model': 9.917878587765658e-05, 'epoch': 0.23}) +[2026-06-17 17:35:50,726][starVLA.training.train_starvla][INFO] - Step 510, Loss: {'train/loss': 0.8138735890388489, 'train/grad_norm_pre_clip': 5.808284282684326, 'timing/data': 0.056835862807929516, 'timing/model': 0.9208984402939677, 'learning_rate/qwen_vl_interface': 9.917477693384264e-06, 'learning_rate/action_model': 9.917477693384264e-05, 'epoch': 0.23}) +[2026-06-17 17:35:54,571][starVLA.training.train_starvla][INFO] - Step 511, Loss: {'train/loss': 0.7258014678955078, 'train/grad_norm_pre_clip': 7.763469219207764, 'timing/data': 0.050318583846092224, 'timing/model': 0.9337484678253531, 'learning_rate/qwen_vl_interface': 9.917075831429721e-06, 'learning_rate/action_model': 9.917075831429722e-05, 'epoch': 0.23}) +[2026-06-17 17:35:59,664][starVLA.training.train_starvla][INFO] - Step 512, Loss: {'train/loss': 0.7841972708702087, 'train/grad_norm_pre_clip': 8.35644245147705, 'timing/data': 0.030450697988271713, 'timing/model': 0.9258797401562333, 'learning_rate/qwen_vl_interface': 9.916673001985339e-06, 'learning_rate/action_model': 9.916673001985338e-05, 'epoch': 0.23}) +[2026-06-17 17:36:03,506][starVLA.training.train_starvla][INFO] - Step 513, Loss: {'train/loss': 0.862701952457428, 'train/grad_norm_pre_clip': 8.518974304199219, 'timing/data': 0.04568464122712612, 'timing/model': 0.9261006433516741, 'learning_rate/qwen_vl_interface': 9.91626920513462e-06, 'learning_rate/action_model': 9.916269205134619e-05, 'epoch': 0.23}) +[2026-06-17 17:36:07,315][starVLA.training.train_starvla][INFO] - Step 514, Loss: {'train/loss': 1.0958647727966309, 'train/grad_norm_pre_clip': 10.030999183654785, 'timing/data': 0.031511420384049416, 'timing/model': 1.1252205865457654, 'learning_rate/qwen_vl_interface': 9.915864440961272e-06, 'learning_rate/action_model': 9.915864440961272e-05, 'epoch': 0.23}) +[2026-06-17 17:36:10,902][starVLA.training.train_starvla][INFO] - Step 515, Loss: {'train/loss': 0.589553713798523, 'train/grad_norm_pre_clip': 10.128153800964355, 'timing/data': 0.032895022071897984, 'timing/model': 0.9407085310667753, 'learning_rate/qwen_vl_interface': 9.915458709549205e-06, 'learning_rate/action_model': 9.915458709549206e-05, 'epoch': 0.23}) +[2026-06-17 17:36:14,691][starVLA.training.train_starvla][INFO] - Step 516, Loss: {'train/loss': 0.7055256962776184, 'train/grad_norm_pre_clip': 10.798931121826172, 'timing/data': 0.03448851406574249, 'timing/model': 0.9369338424876332, 'learning_rate/qwen_vl_interface': 9.91505201098253e-06, 'learning_rate/action_model': 9.91505201098253e-05, 'epoch': 0.23}) +[2026-06-17 17:36:18,503][starVLA.training.train_starvla][INFO] - Step 517, Loss: {'train/loss': 0.9768505692481995, 'train/grad_norm_pre_clip': 12.303003311157227, 'timing/data': 0.04664493724703789, 'timing/model': 0.9457243420183659, 'learning_rate/qwen_vl_interface': 9.914644345345551e-06, 'learning_rate/action_model': 9.914644345345552e-05, 'epoch': 0.23}) +[2026-06-17 17:36:22,314][starVLA.training.train_starvla][INFO] - Step 518, Loss: {'train/loss': 0.944082498550415, 'train/grad_norm_pre_clip': 8.143214225769043, 'timing/data': 0.03876797296106815, 'timing/model': 0.9352360600605607, 'learning_rate/qwen_vl_interface': 9.914235712722782e-06, 'learning_rate/action_model': 9.914235712722781e-05, 'epoch': 0.24}) +[2026-06-17 17:36:27,230][starVLA.training.train_starvla][INFO] - Step 519, Loss: {'train/loss': 0.7286660671234131, 'train/grad_norm_pre_clip': 6.401454925537109, 'timing/data': 0.038963102735579014, 'timing/model': 0.9885737607255578, 'learning_rate/qwen_vl_interface': 9.913826113198927e-06, 'learning_rate/action_model': 9.913826113198927e-05, 'epoch': 0.24}) +[2026-06-17 17:36:31,389][starVLA.training.train_starvla][INFO] - Step 520, Loss: {'train/loss': 0.938266932964325, 'train/grad_norm_pre_clip': 8.613590240478516, 'timing/data': 0.044223763048648834, 'timing/model': 1.1909165540710092, 'learning_rate/qwen_vl_interface': 9.913415546858904e-06, 'learning_rate/action_model': 9.913415546858904e-05, 'epoch': 0.24}) +[2026-06-17 17:36:35,201][starVLA.training.train_starvla][INFO] - Step 521, Loss: {'train/loss': 0.973863422870636, 'train/grad_norm_pre_clip': 7.236850738525391, 'timing/data': 0.039756580255925655, 'timing/model': 1.108281803317368, 'learning_rate/qwen_vl_interface': 9.913004013787818e-06, 'learning_rate/action_model': 9.913004013787819e-05, 'epoch': 0.24}) +[2026-06-17 17:36:41,211][starVLA.training.train_starvla][INFO] - Step 522, Loss: {'train/loss': 0.7263181805610657, 'train/grad_norm_pre_clip': 6.877974987030029, 'timing/data': 0.04332998115569353, 'timing/model': 0.9392548343166709, 'learning_rate/qwen_vl_interface': 9.912591514070983e-06, 'learning_rate/action_model': 9.912591514070983e-05, 'epoch': 0.24}) +[2026-06-17 17:36:45,140][starVLA.training.train_starvla][INFO] - Step 523, Loss: {'train/loss': 0.4799287021160126, 'train/grad_norm_pre_clip': 5.5994086265563965, 'timing/data': 0.041629779152572155, 'timing/model': 0.9241221519187093, 'learning_rate/qwen_vl_interface': 9.912178047793912e-06, 'learning_rate/action_model': 9.91217804779391e-05, 'epoch': 0.24}) +[2026-06-17 17:36:49,027][starVLA.training.train_starvla][INFO] - Step 524, Loss: {'train/loss': 0.5747804641723633, 'train/grad_norm_pre_clip': 5.094475269317627, 'timing/data': 0.05187175516039133, 'timing/model': 0.9909928282722831, 'learning_rate/qwen_vl_interface': 9.911763615042311e-06, 'learning_rate/action_model': 9.91176361504231e-05, 'epoch': 0.24}) +[2026-06-17 17:36:52,941][starVLA.training.train_starvla][INFO] - Step 525, Loss: {'train/loss': 0.5184204578399658, 'train/grad_norm_pre_clip': 4.703683376312256, 'timing/data': 0.05388987623155117, 'timing/model': 0.9359118156135082, 'learning_rate/qwen_vl_interface': 9.911348215902099e-06, 'learning_rate/action_model': 9.911348215902099e-05, 'epoch': 0.24}) +[2026-06-17 17:36:56,797][starVLA.training.train_starvla][INFO] - Step 526, Loss: {'train/loss': 0.5504316687583923, 'train/grad_norm_pre_clip': 6.782219886779785, 'timing/data': 0.0334931593388319, 'timing/model': 1.169248217716813, 'learning_rate/qwen_vl_interface': 9.910931850459385e-06, 'learning_rate/action_model': 9.910931850459385e-05, 'epoch': 0.24}) +[2026-06-17 17:37:00,969][starVLA.training.train_starvla][INFO] - Step 527, Loss: {'train/loss': 0.8465979695320129, 'train/grad_norm_pre_clip': 4.632782459259033, 'timing/data': 0.05502353701740503, 'timing/model': 1.1690504392609, 'learning_rate/qwen_vl_interface': 9.910514518800485e-06, 'learning_rate/action_model': 9.910514518800484e-05, 'epoch': 0.24}) +[2026-06-17 17:37:04,968][starVLA.training.train_starvla][INFO] - Step 528, Loss: {'train/loss': 0.5185458064079285, 'train/grad_norm_pre_clip': 4.545886993408203, 'timing/data': 0.053541598841547966, 'timing/model': 0.9666371960192919, 'learning_rate/qwen_vl_interface': 9.910096221011908e-06, 'learning_rate/action_model': 9.910096221011907e-05, 'epoch': 0.24}) +[2026-06-17 17:37:08,966][starVLA.training.train_starvla][INFO] - Step 529, Loss: {'train/loss': 0.5760292410850525, 'train/grad_norm_pre_clip': 5.6724324226379395, 'timing/data': 0.03731849975883961, 'timing/model': 0.9351602997630835, 'learning_rate/qwen_vl_interface': 9.90967695718037e-06, 'learning_rate/action_model': 9.90967695718037e-05, 'epoch': 0.24}) +[2026-06-17 17:37:13,184][starVLA.training.train_starvla][INFO] - Step 530, Loss: {'train/loss': 0.4867135286331177, 'train/grad_norm_pre_clip': 4.231477737426758, 'timing/data': 0.04899997264146805, 'timing/model': 0.9487899402156472, 'learning_rate/qwen_vl_interface': 9.909256727392786e-06, 'learning_rate/action_model': 9.909256727392784e-05, 'epoch': 0.24}) +[2026-06-17 17:37:17,158][starVLA.training.train_starvla][INFO] - Step 531, Loss: {'train/loss': 0.9598336219787598, 'train/grad_norm_pre_clip': 6.781742572784424, 'timing/data': 0.05265645869076252, 'timing/model': 1.0313145592808723, 'learning_rate/qwen_vl_interface': 9.908835531736269e-06, 'learning_rate/action_model': 9.908835531736267e-05, 'epoch': 0.24}) +[2026-06-17 17:37:20,996][starVLA.training.train_starvla][INFO] - Step 532, Loss: {'train/loss': 0.6593585014343262, 'train/grad_norm_pre_clip': 4.4570417404174805, 'timing/data': 0.04047936853021383, 'timing/model': 0.9362524449825287, 'learning_rate/qwen_vl_interface': 9.908413370298131e-06, 'learning_rate/action_model': 9.90841337029813e-05, 'epoch': 0.24}) +[2026-06-17 17:37:24,835][starVLA.training.train_starvla][INFO] - Step 533, Loss: {'train/loss': 0.9632790684700012, 'train/grad_norm_pre_clip': 6.564311981201172, 'timing/data': 0.035919335670769215, 'timing/model': 1.1757565131410956, 'learning_rate/qwen_vl_interface': 9.90799024316589e-06, 'learning_rate/action_model': 9.90799024316589e-05, 'epoch': 0.24}) +[2026-06-17 17:37:28,507][starVLA.training.train_starvla][INFO] - Step 534, Loss: {'train/loss': 0.7827633023262024, 'train/grad_norm_pre_clip': 20.99552345275879, 'timing/data': 0.04005155712366104, 'timing/model': 0.9246560027822852, 'learning_rate/qwen_vl_interface': 9.90756615042726e-06, 'learning_rate/action_model': 9.907566150427259e-05, 'epoch': 0.24}) +[2026-06-17 17:37:32,766][starVLA.training.train_starvla][INFO] - Step 535, Loss: {'train/loss': 0.8104928135871887, 'train/grad_norm_pre_clip': 6.243985176086426, 'timing/data': 0.030645313672721386, 'timing/model': 0.916343386285007, 'learning_rate/qwen_vl_interface': 9.907141092170154e-06, 'learning_rate/action_model': 9.907141092170153e-05, 'epoch': 0.24}) +[2026-06-17 17:37:36,750][starVLA.training.train_starvla][INFO] - Step 536, Loss: {'train/loss': 0.7229812145233154, 'train/grad_norm_pre_clip': 6.786951065063477, 'timing/data': 0.04706982243806124, 'timing/model': 1.0733132679015398, 'learning_rate/qwen_vl_interface': 9.906715068482689e-06, 'learning_rate/action_model': 9.90671506848269e-05, 'epoch': 0.24}) +[2026-06-17 17:37:40,771][starVLA.training.train_starvla][INFO] - Step 537, Loss: {'train/loss': 0.5910030603408813, 'train/grad_norm_pre_clip': 5.486731052398682, 'timing/data': 0.04167252033948898, 'timing/model': 0.9388139666989446, 'learning_rate/qwen_vl_interface': 9.90628807945318e-06, 'learning_rate/action_model': 9.90628807945318e-05, 'epoch': 0.24}) +[2026-06-17 17:37:44,798][starVLA.training.train_starvla][INFO] - Step 538, Loss: {'train/loss': 0.58237224817276, 'train/grad_norm_pre_clip': 10.088744163513184, 'timing/data': 0.05444419663399458, 'timing/model': 1.0046537201851606, 'learning_rate/qwen_vl_interface': 9.90586012517014e-06, 'learning_rate/action_model': 9.90586012517014e-05, 'epoch': 0.24}) +[2026-06-17 17:37:48,676][starVLA.training.train_starvla][INFO] - Step 539, Loss: {'train/loss': 1.0429368019104004, 'train/grad_norm_pre_clip': 5.952744007110596, 'timing/data': 0.04316085856407881, 'timing/model': 0.9595030304044485, 'learning_rate/qwen_vl_interface': 9.905431205722288e-06, 'learning_rate/action_model': 9.905431205722288e-05, 'epoch': 0.24}) +[2026-06-17 17:37:52,663][starVLA.training.train_starvla][INFO] - Step 540, Loss: {'train/loss': 0.6618033647537231, 'train/grad_norm_pre_clip': 5.9869771003723145, 'timing/data': 0.042327526956796646, 'timing/model': 1.1656853733584285, 'learning_rate/qwen_vl_interface': 9.905001321198536e-06, 'learning_rate/action_model': 9.905001321198536e-05, 'epoch': 0.24}) +[2026-06-17 17:37:56,373][starVLA.training.train_starvla][INFO] - Step 541, Loss: {'train/loss': 0.8649980425834656, 'train/grad_norm_pre_clip': 5.98203182220459, 'timing/data': 0.0387400658801198, 'timing/model': 0.9414307456463575, 'learning_rate/qwen_vl_interface': 9.904570471688003e-06, 'learning_rate/action_model': 9.904570471688002e-05, 'epoch': 0.25}) +[2026-06-17 17:38:03,910][starVLA.training.train_starvla][INFO] - Step 542, Loss: {'train/loss': 0.529218852519989, 'train/grad_norm_pre_clip': 6.01348876953125, 'timing/data': 0.04581612907350063, 'timing/model': 2.387455867603421, 'learning_rate/qwen_vl_interface': 9.90413865728e-06, 'learning_rate/action_model': 9.90413865728e-05, 'epoch': 0.25}) +[2026-06-17 17:38:08,048][starVLA.training.train_starvla][INFO] - Step 543, Loss: {'train/loss': 0.7431979179382324, 'train/grad_norm_pre_clip': 5.794236183166504, 'timing/data': 0.09879851806908846, 'timing/model': 0.9399862512946129, 'learning_rate/qwen_vl_interface': 9.903705878064047e-06, 'learning_rate/action_model': 9.903705878064046e-05, 'epoch': 0.25}) +[2026-06-17 17:38:12,503][starVLA.training.train_starvla][INFO] - Step 544, Loss: {'train/loss': 0.6433907151222229, 'train/grad_norm_pre_clip': 5.059603214263916, 'timing/data': 0.05070774722844362, 'timing/model': 0.945545538328588, 'learning_rate/qwen_vl_interface': 9.903272134129858e-06, 'learning_rate/action_model': 9.903272134129858e-05, 'epoch': 0.25}) +[2026-06-17 17:38:16,425][starVLA.training.train_starvla][INFO] - Step 545, Loss: {'train/loss': 0.7002990245819092, 'train/grad_norm_pre_clip': 5.637619972229004, 'timing/data': 0.04506062529981136, 'timing/model': 0.9812745740637183, 'learning_rate/qwen_vl_interface': 9.90283742556735e-06, 'learning_rate/action_model': 9.902837425567349e-05, 'epoch': 0.25}) +[2026-06-17 17:38:20,369][starVLA.training.train_starvla][INFO] - Step 546, Loss: {'train/loss': 0.9018740653991699, 'train/grad_norm_pre_clip': 7.5014872550964355, 'timing/data': 0.04295926075428724, 'timing/model': 0.9311156421899796, 'learning_rate/qwen_vl_interface': 9.902401752466635e-06, 'learning_rate/action_model': 9.902401752466634e-05, 'epoch': 0.25}) +[2026-06-17 17:38:24,267][starVLA.training.train_starvla][INFO] - Step 547, Loss: {'train/loss': 0.8389154076576233, 'train/grad_norm_pre_clip': 5.430192470550537, 'timing/data': 0.04748482536524534, 'timing/model': 1.1189116323366761, 'learning_rate/qwen_vl_interface': 9.90196511491803e-06, 'learning_rate/action_model': 9.901965114918029e-05, 'epoch': 0.25}) +[2026-06-17 17:38:28,001][starVLA.training.train_starvla][INFO] - Step 548, Loss: {'train/loss': 0.6056092381477356, 'train/grad_norm_pre_clip': 7.646155834197998, 'timing/data': 0.048581275157630444, 'timing/model': 0.9460678296163678, 'learning_rate/qwen_vl_interface': 9.901527513012053e-06, 'learning_rate/action_model': 9.901527513012052e-05, 'epoch': 0.25}) +[2026-06-17 17:38:31,986][starVLA.training.train_starvla][INFO] - Step 549, Loss: {'train/loss': 1.0245884656906128, 'train/grad_norm_pre_clip': 5.787375450134277, 'timing/data': 0.04852334875613451, 'timing/model': 1.0459062419831753, 'learning_rate/qwen_vl_interface': 9.901088946839415e-06, 'learning_rate/action_model': 9.901088946839414e-05, 'epoch': 0.25}) +[2026-06-17 17:38:35,777][starVLA.training.train_starvla][INFO] - Step 550, Loss: {'train/loss': 0.6816079616546631, 'train/grad_norm_pre_clip': 4.3042216300964355, 'timing/data': 0.04769984073936939, 'timing/model': 0.9340382535010576, 'learning_rate/qwen_vl_interface': 9.900649416491035e-06, 'learning_rate/action_model': 9.900649416491035e-05, 'epoch': 0.25}) +[2026-06-17 17:38:39,514][starVLA.training.train_starvla][INFO] - Step 551, Loss: {'train/loss': 0.6463829874992371, 'train/grad_norm_pre_clip': 5.429555892944336, 'timing/data': 0.03835930861532688, 'timing/model': 0.9198885979130864, 'learning_rate/qwen_vl_interface': 9.900208922058026e-06, 'learning_rate/action_model': 9.900208922058025e-05, 'epoch': 0.25}) +[2026-06-17 17:38:43,345][starVLA.training.train_starvla][INFO] - Step 552, Loss: {'train/loss': 0.7498326301574707, 'train/grad_norm_pre_clip': 8.154397964477539, 'timing/data': 0.03412019181996584, 'timing/model': 0.9224416995421052, 'learning_rate/qwen_vl_interface': 9.899767463631704e-06, 'learning_rate/action_model': 9.899767463631704e-05, 'epoch': 0.25}) +[2026-06-17 17:38:47,155][starVLA.training.train_starvla][INFO] - Step 553, Loss: {'train/loss': 0.56733238697052, 'train/grad_norm_pre_clip': 5.222784042358398, 'timing/data': 0.036217572167515755, 'timing/model': 1.1317254127934575, 'learning_rate/qwen_vl_interface': 9.899325041303582e-06, 'learning_rate/action_model': 9.899325041303583e-05, 'epoch': 0.25}) +[2026-06-17 17:38:50,982][starVLA.training.train_starvla][INFO] - Step 554, Loss: {'train/loss': 0.8945274949073792, 'train/grad_norm_pre_clip': 10.607558250427246, 'timing/data': 0.038115501403808594, 'timing/model': 1.1521836267784238, 'learning_rate/qwen_vl_interface': 9.89888165516538e-06, 'learning_rate/action_model': 9.898881655165379e-05, 'epoch': 0.25}) +[2026-06-17 17:38:54,620][starVLA.training.train_starvla][INFO] - Step 555, Loss: {'train/loss': 1.04304039478302, 'train/grad_norm_pre_clip': 7.603922367095947, 'timing/data': 0.03559336345642805, 'timing/model': 0.9291367279365659, 'learning_rate/qwen_vl_interface': 9.898437305309004e-06, 'learning_rate/action_model': 9.898437305309003e-05, 'epoch': 0.25}) +[2026-06-17 17:38:58,538][starVLA.training.train_starvla][INFO] - Step 556, Loss: {'train/loss': 0.8462494015693665, 'train/grad_norm_pre_clip': 5.240739822387695, 'timing/data': 0.03232760075479746, 'timing/model': 0.9263171525672078, 'learning_rate/qwen_vl_interface': 9.897991991826575e-06, 'learning_rate/action_model': 9.897991991826574e-05, 'epoch': 0.25}) +[2026-06-17 17:39:02,333][starVLA.training.train_starvla][INFO] - Step 557, Loss: {'train/loss': 0.919994592666626, 'train/grad_norm_pre_clip': 9.633550643920898, 'timing/data': 0.04165509436279535, 'timing/model': 0.9342674911022186, 'learning_rate/qwen_vl_interface': 9.897545714810404e-06, 'learning_rate/action_model': 9.897545714810403e-05, 'epoch': 0.25}) +[2026-06-17 17:39:06,497][starVLA.training.train_starvla][INFO] - Step 558, Loss: {'train/loss': 0.5765483379364014, 'train/grad_norm_pre_clip': 7.453366279602051, 'timing/data': 0.038357765413820744, 'timing/model': 1.0258145648986101, 'learning_rate/qwen_vl_interface': 9.897098474353004e-06, 'learning_rate/action_model': 9.897098474353004e-05, 'epoch': 0.25}) +[2026-06-17 17:39:10,346][starVLA.training.train_starvla][INFO] - Step 559, Loss: {'train/loss': 0.7969745397567749, 'train/grad_norm_pre_clip': 6.619673252105713, 'timing/data': 0.03487431164830923, 'timing/model': 0.9438579520210624, 'learning_rate/qwen_vl_interface': 9.896650270547092e-06, 'learning_rate/action_model': 9.89665027054709e-05, 'epoch': 0.25}) +[2026-06-17 17:39:14,302][starVLA.training.train_starvla][INFO] - Step 560, Loss: {'train/loss': 0.7362651824951172, 'train/grad_norm_pre_clip': 4.847813606262207, 'timing/data': 0.0321739362552762, 'timing/model': 1.233014757744968, 'learning_rate/qwen_vl_interface': 9.896201103485577e-06, 'learning_rate/action_model': 9.896201103485577e-05, 'epoch': 0.25}) +[2026-06-17 17:39:18,189][starVLA.training.train_starvla][INFO] - Step 561, Loss: {'train/loss': 0.6464364528656006, 'train/grad_norm_pre_clip': 4.3213653564453125, 'timing/data': 0.046742032282054424, 'timing/model': 1.102704762481153, 'learning_rate/qwen_vl_interface': 9.895750973261576e-06, 'learning_rate/action_model': 9.895750973261574e-05, 'epoch': 0.25}) +[2026-06-17 17:39:21,852][starVLA.training.train_starvla][INFO] - Step 562, Loss: {'train/loss': 0.7215541005134583, 'train/grad_norm_pre_clip': 6.865474224090576, 'timing/data': 0.03540449123829603, 'timing/model': 0.9776966329663992, 'learning_rate/qwen_vl_interface': 9.895299879968398e-06, 'learning_rate/action_model': 9.895299879968398e-05, 'epoch': 0.25}) +[2026-06-17 17:39:25,736][starVLA.training.train_starvla][INFO] - Step 563, Loss: {'train/loss': 0.9507941603660583, 'train/grad_norm_pre_clip': 6.361757755279541, 'timing/data': 0.03915489371865988, 'timing/model': 0.9416319187730551, 'learning_rate/qwen_vl_interface': 9.894847823699555e-06, 'learning_rate/action_model': 9.894847823699555e-05, 'epoch': 0.26}) +[2026-06-17 17:39:29,676][starVLA.training.train_starvla][INFO] - Step 564, Loss: {'train/loss': 0.46982693672180176, 'train/grad_norm_pre_clip': 4.897209167480469, 'timing/data': 0.033857799135148525, 'timing/model': 0.9982236986979842, 'learning_rate/qwen_vl_interface': 9.894394804548763e-06, 'learning_rate/action_model': 9.894394804548762e-05, 'epoch': 0.26}) +[2026-06-17 17:39:33,516][starVLA.training.train_starvla][INFO] - Step 565, Loss: {'train/loss': 0.4714868664741516, 'train/grad_norm_pre_clip': 5.328598499298096, 'timing/data': 0.04912085831165314, 'timing/model': 0.9310629451647401, 'learning_rate/qwen_vl_interface': 9.893940822609928e-06, 'learning_rate/action_model': 9.893940822609927e-05, 'epoch': 0.26}) +[2026-06-17 17:39:37,565][starVLA.training.train_starvla][INFO] - Step 566, Loss: {'train/loss': 0.7431813478469849, 'train/grad_norm_pre_clip': 6.164735794067383, 'timing/data': 0.03307241294533014, 'timing/model': 1.2583803310990334, 'learning_rate/qwen_vl_interface': 9.893485877977163e-06, 'learning_rate/action_model': 9.893485877977163e-05, 'epoch': 0.26}) +[2026-06-17 17:39:41,563][starVLA.training.train_starvla][INFO] - Step 567, Loss: {'train/loss': 0.8227452635765076, 'train/grad_norm_pre_clip': 9.650123596191406, 'timing/data': 0.039482973515987396, 'timing/model': 1.1690918756648898, 'learning_rate/qwen_vl_interface': 9.89302997074478e-06, 'learning_rate/action_model': 9.89302997074478e-05, 'epoch': 0.26}) +[2026-06-17 17:39:45,193][starVLA.training.train_starvla][INFO] - Step 568, Loss: {'train/loss': 0.5375389456748962, 'train/grad_norm_pre_clip': 5.862095832824707, 'timing/data': 0.036918207071721554, 'timing/model': 0.9229331547394395, 'learning_rate/qwen_vl_interface': 9.892573101007288e-06, 'learning_rate/action_model': 9.892573101007289e-05, 'epoch': 0.26}) +[2026-06-17 17:39:49,033][starVLA.training.train_starvla][INFO] - Step 569, Loss: {'train/loss': 0.37722334265708923, 'train/grad_norm_pre_clip': 5.273642063140869, 'timing/data': 0.029245274141430855, 'timing/model': 0.9158829916268587, 'learning_rate/qwen_vl_interface': 9.892115268859398e-06, 'learning_rate/action_model': 9.892115268859397e-05, 'epoch': 0.26}) +[2026-06-17 17:39:52,841][starVLA.training.train_starvla][INFO] - Step 570, Loss: {'train/loss': 0.913394570350647, 'train/grad_norm_pre_clip': 7.414515972137451, 'timing/data': 0.04000077489763498, 'timing/model': 0.9332795673981309, 'learning_rate/qwen_vl_interface': 9.891656474396015e-06, 'learning_rate/action_model': 9.891656474396015e-05, 'epoch': 0.26}) +[2026-06-17 17:39:56,627][starVLA.training.train_starvla][INFO] - Step 571, Loss: {'train/loss': 0.6248905062675476, 'train/grad_norm_pre_clip': 5.667761325836182, 'timing/data': 0.04580111987888813, 'timing/model': 0.9281330266967416, 'learning_rate/qwen_vl_interface': 9.891196717712253e-06, 'learning_rate/action_model': 9.891196717712252e-05, 'epoch': 0.26}) +[2026-06-17 17:40:00,442][starVLA.training.train_starvla][INFO] - Step 572, Loss: {'train/loss': 0.7076776623725891, 'train/grad_norm_pre_clip': 8.658390998840332, 'timing/data': 0.044331807643175125, 'timing/model': 0.9257752988487482, 'learning_rate/qwen_vl_interface': 9.890735998903414e-06, 'learning_rate/action_model': 9.890735998903414e-05, 'epoch': 0.26}) +[2026-06-17 17:40:04,246][starVLA.training.train_starvla][INFO] - Step 573, Loss: {'train/loss': 0.8521766662597656, 'train/grad_norm_pre_clip': 21.281957626342773, 'timing/data': 0.03917940892279148, 'timing/model': 0.9240732956677675, 'learning_rate/qwen_vl_interface': 9.890274318065012e-06, 'learning_rate/action_model': 9.890274318065011e-05, 'epoch': 0.26}) +[2026-06-17 17:40:07,990][starVLA.training.train_starvla][INFO] - Step 574, Loss: {'train/loss': 0.5154096484184265, 'train/grad_norm_pre_clip': 7.293431282043457, 'timing/data': 0.03233656194061041, 'timing/model': 1.095424273982644, 'learning_rate/qwen_vl_interface': 9.889811675292748e-06, 'learning_rate/action_model': 9.889811675292749e-05, 'epoch': 0.26}) +[2026-06-17 17:40:12,089][starVLA.training.train_starvla][INFO] - Step 575, Loss: {'train/loss': 0.7179945111274719, 'train/grad_norm_pre_clip': 6.786635875701904, 'timing/data': 0.03647941164672375, 'timing/model': 0.9660639930516481, 'learning_rate/qwen_vl_interface': 9.889348070682533e-06, 'learning_rate/action_model': 9.889348070682533e-05, 'epoch': 0.26}) +[2026-06-17 17:40:15,877][starVLA.training.train_starvla][INFO] - Step 576, Loss: {'train/loss': 0.6663915514945984, 'train/grad_norm_pre_clip': 9.887137413024902, 'timing/data': 0.03378441743552685, 'timing/model': 0.9511447818949819, 'learning_rate/qwen_vl_interface': 9.888883504330469e-06, 'learning_rate/action_model': 9.888883504330468e-05, 'epoch': 0.26}) +[2026-06-17 17:40:19,677][starVLA.training.train_starvla][INFO] - Step 577, Loss: {'train/loss': 0.676989734172821, 'train/grad_norm_pre_clip': 6.006119251251221, 'timing/data': 0.028348361141979694, 'timing/model': 0.9206372359767556, 'learning_rate/qwen_vl_interface': 9.888417976332865e-06, 'learning_rate/action_model': 9.888417976332865e-05, 'epoch': 0.26}) +[2026-06-17 17:40:23,609][starVLA.training.train_starvla][INFO] - Step 578, Loss: {'train/loss': 0.5891871452331543, 'train/grad_norm_pre_clip': 5.843216419219971, 'timing/data': 0.03415017668157816, 'timing/model': 1.0068220468237996, 'learning_rate/qwen_vl_interface': 9.887951486786222e-06, 'learning_rate/action_model': 9.887951486786221e-05, 'epoch': 0.26}) +[2026-06-17 17:40:27,397][starVLA.training.train_starvla][INFO] - Step 579, Loss: {'train/loss': 0.7164285778999329, 'train/grad_norm_pre_clip': 8.130685806274414, 'timing/data': 0.04522308614104986, 'timing/model': 0.9204907938838005, 'learning_rate/qwen_vl_interface': 9.887484035787246e-06, 'learning_rate/action_model': 9.887484035787246e-05, 'epoch': 0.26}) +[2026-06-17 17:40:31,295][starVLA.training.train_starvla][INFO] - Step 580, Loss: {'train/loss': 0.6269329786300659, 'train/grad_norm_pre_clip': 8.115367889404297, 'timing/data': 0.0454942537471652, 'timing/model': 0.9711972959339619, 'learning_rate/qwen_vl_interface': 9.887015623432837e-06, 'learning_rate/action_model': 9.887015623432837e-05, 'epoch': 0.26}) +[2026-06-17 17:40:35,091][starVLA.training.train_starvla][INFO] - Step 581, Loss: {'train/loss': 0.9655051827430725, 'train/grad_norm_pre_clip': 8.720455169677734, 'timing/data': 0.03525359835475683, 'timing/model': 1.0847536604851484, 'learning_rate/qwen_vl_interface': 9.886546249820101e-06, 'learning_rate/action_model': 9.886546249820102e-05, 'epoch': 0.26}) +[2026-06-17 17:40:38,919][starVLA.training.train_starvla][INFO] - Step 582, Loss: {'train/loss': 0.655616044998169, 'train/grad_norm_pre_clip': 7.178527355194092, 'timing/data': 0.031778255477547646, 'timing/model': 0.9493877673521638, 'learning_rate/qwen_vl_interface': 9.886075915046339e-06, 'learning_rate/action_model': 9.886075915046338e-05, 'epoch': 0.26}) +[2026-06-17 17:40:42,786][starVLA.training.train_starvla][INFO] - Step 583, Loss: {'train/loss': 0.8156666159629822, 'train/grad_norm_pre_clip': 9.665033340454102, 'timing/data': 0.03904097434133291, 'timing/model': 0.9405003916472197, 'learning_rate/qwen_vl_interface': 9.885604619209052e-06, 'learning_rate/action_model': 9.885604619209051e-05, 'epoch': 0.26}) +[2026-06-17 17:40:46,513][starVLA.training.train_starvla][INFO] - Step 584, Loss: {'train/loss': 0.7059577703475952, 'train/grad_norm_pre_clip': 59.96413040161133, 'timing/data': 0.03774147480726242, 'timing/model': 0.9218594804406166, 'learning_rate/qwen_vl_interface': 9.885132362405935e-06, 'learning_rate/action_model': 9.885132362405936e-05, 'epoch': 0.26}) +[2026-06-17 17:40:50,240][starVLA.training.train_starvla][INFO] - Step 585, Loss: {'train/loss': 0.5224744081497192, 'train/grad_norm_pre_clip': 5.916935443878174, 'timing/data': 0.033977555111050606, 'timing/model': 0.9245776431635022, 'learning_rate/qwen_vl_interface': 9.884659144734896e-06, 'learning_rate/action_model': 9.884659144734896e-05, 'epoch': 0.27}) +[2026-06-17 17:40:54,384][starVLA.training.train_starvla][INFO] - Step 586, Loss: {'train/loss': 0.8848727941513062, 'train/grad_norm_pre_clip': 9.434165000915527, 'timing/data': 0.04065632168203592, 'timing/model': 1.0048405984416604, 'learning_rate/qwen_vl_interface': 9.884184966294028e-06, 'learning_rate/action_model': 9.884184966294028e-05, 'epoch': 0.27}) +[2026-06-17 17:40:58,301][starVLA.training.train_starvla][INFO] - Step 587, Loss: {'train/loss': 0.4984329342842102, 'train/grad_norm_pre_clip': 5.292819023132324, 'timing/data': 0.032231626100838184, 'timing/model': 1.175279633142054, 'learning_rate/qwen_vl_interface': 9.883709827181632e-06, 'learning_rate/action_model': 9.88370982718163e-05, 'epoch': 0.27}) +[2026-06-17 17:41:02,204][starVLA.training.train_starvla][INFO] - Step 588, Loss: {'train/loss': 0.8774500489234924, 'train/grad_norm_pre_clip': 5.609436511993408, 'timing/data': 0.03468833025544882, 'timing/model': 1.1214840551838279, 'learning_rate/qwen_vl_interface': 9.8832337274962e-06, 'learning_rate/action_model': 9.883233727496198e-05, 'epoch': 0.27}) +[2026-06-17 17:41:07,122][starVLA.training.train_starvla][INFO] - Step 589, Loss: {'train/loss': 0.6554385423660278, 'train/grad_norm_pre_clip': 7.840298652648926, 'timing/data': 0.041561516001820564, 'timing/model': 1.343355848453939, 'learning_rate/qwen_vl_interface': 9.88275666733643e-06, 'learning_rate/action_model': 9.882756667336429e-05, 'epoch': 0.27}) +[2026-06-17 17:41:11,012][starVLA.training.train_starvla][INFO] - Step 590, Loss: {'train/loss': 0.6295888423919678, 'train/grad_norm_pre_clip': 6.5529022216796875, 'timing/data': 0.04218473471701145, 'timing/model': 0.9766221698373556, 'learning_rate/qwen_vl_interface': 9.88227864680122e-06, 'learning_rate/action_model': 9.88227864680122e-05, 'epoch': 0.27}) +[2026-06-17 17:41:14,940][starVLA.training.train_starvla][INFO] - Step 591, Loss: {'train/loss': 0.6858916282653809, 'train/grad_norm_pre_clip': 7.125792026519775, 'timing/data': 0.040940048173069954, 'timing/model': 0.9361549289897084, 'learning_rate/qwen_vl_interface': 9.88179966598966e-06, 'learning_rate/action_model': 9.88179966598966e-05, 'epoch': 0.27}) +[2026-06-17 17:41:18,963][starVLA.training.train_starvla][INFO] - Step 592, Loss: {'train/loss': 0.6698889136314392, 'train/grad_norm_pre_clip': 6.255506992340088, 'timing/data': 0.034776924178004265, 'timing/model': 0.957616520114243, 'learning_rate/qwen_vl_interface': 9.881319725001046e-06, 'learning_rate/action_model': 9.881319725001046e-05, 'epoch': 0.27}) +[2026-06-17 17:41:22,794][starVLA.training.train_starvla][INFO] - Step 593, Loss: {'train/loss': 0.9736312627792358, 'train/grad_norm_pre_clip': 9.807056427001953, 'timing/data': 0.04146189894527197, 'timing/model': 0.9564484693109989, 'learning_rate/qwen_vl_interface': 9.88083882393487e-06, 'learning_rate/action_model': 9.880838823934869e-05, 'epoch': 0.27}) +[2026-06-17 17:41:26,754][starVLA.training.train_starvla][INFO] - Step 594, Loss: {'train/loss': 0.7006383538246155, 'train/grad_norm_pre_clip': 6.739472389221191, 'timing/data': 0.03386352304369211, 'timing/model': 1.1356918523088098, 'learning_rate/qwen_vl_interface': 9.880356962890823e-06, 'learning_rate/action_model': 9.880356962890822e-05, 'epoch': 0.27}) +[2026-06-17 17:41:30,664][starVLA.training.train_starvla][INFO] - Step 595, Loss: {'train/loss': 1.0124036073684692, 'train/grad_norm_pre_clip': 7.477741241455078, 'timing/data': 0.03324860706925392, 'timing/model': 1.1461297189816833, 'learning_rate/qwen_vl_interface': 9.879874141968792e-06, 'learning_rate/action_model': 9.879874141968792e-05, 'epoch': 0.27}) +[2026-06-17 17:41:34,592][starVLA.training.train_starvla][INFO] - Step 596, Loss: {'train/loss': 0.6187232732772827, 'train/grad_norm_pre_clip': 5.353180885314941, 'timing/data': 0.029373318888247013, 'timing/model': 0.9453992145135999, 'learning_rate/qwen_vl_interface': 9.879390361268873e-06, 'learning_rate/action_model': 9.879390361268873e-05, 'epoch': 0.27}) +[2026-06-17 17:41:38,446][starVLA.training.train_starvla][INFO] - Step 597, Loss: {'train/loss': 0.765397310256958, 'train/grad_norm_pre_clip': 8.457540512084961, 'timing/data': 0.03163151815533638, 'timing/model': 0.9334969222545624, 'learning_rate/qwen_vl_interface': 9.878905620891349e-06, 'learning_rate/action_model': 9.878905620891349e-05, 'epoch': 0.27}) +[2026-06-17 17:41:42,378][starVLA.training.train_starvla][INFO] - Step 598, Loss: {'train/loss': 0.5820069909095764, 'train/grad_norm_pre_clip': 7.797126770019531, 'timing/data': 0.041940401308238506, 'timing/model': 0.9383600754663348, 'learning_rate/qwen_vl_interface': 9.87841992093671e-06, 'learning_rate/action_model': 9.87841992093671e-05, 'epoch': 0.27}) +[2026-06-17 17:41:46,274][starVLA.training.train_starvla][INFO] - Step 599, Loss: {'train/loss': 0.6122732162475586, 'train/grad_norm_pre_clip': 7.695893287658691, 'timing/data': 0.041770593263208866, 'timing/model': 0.9439626559615135, 'learning_rate/qwen_vl_interface': 9.877933261505642e-06, 'learning_rate/action_model': 9.877933261505641e-05, 'epoch': 0.27}) +[2026-06-17 17:43:53,607][starVLA.training.train_starvla][INFO] - Step 600, Loss: {'train/loss': 0.7993944883346558, 'train/grad_norm_pre_clip': 10.879646301269531, 'eval/loss': 0.7425737977027893, 'eval/latency_8/loss': 0.7425737977027893, 'eval/demon_fire_precision': 0.8990748528174937, 'eval/demon_fire_recall': 0.9962721342031687, 'eval/demon_fire_f1': 0.945181255526083, 'eval/demon_fire_rate_model': 0.9924874791318865, 'eval/demon_fire_rate_teacher': 0.8956594323873122, 'eval/demon_left_precision': 0.5296703296703297, 'eval/demon_left_recall': 0.5160599571734475, 'eval/demon_left_f1': 0.5227765726681128, 'eval/demon_left_rate_model': 0.12659988870339456, 'eval/demon_left_rate_teacher': 0.12993878686700056, 'eval/demon_right_precision': 0.9699248120300752, 'eval/demon_right_recall': 0.19545454545454546, 'eval/demon_right_f1': 0.32534678436317777, 'eval/demon_right_rate_model': 0.037006121313299944, 'eval/demon_right_rate_teacher': 0.18363939899833054, 'eval/demon_move_f1': 0.4240616785156453, 'eval/demon_cc_f1': 0.6905872511040223, 'eval/latency_8/demon_fire_precision': 0.8990748528174937, 'eval/latency_8/demon_fire_recall': 0.9962721342031687, 'eval/latency_8/demon_fire_f1': 0.945181255526083, 'eval/latency_8/demon_fire_rate_model': 0.9924874791318865, 'eval/latency_8/demon_fire_rate_teacher': 0.8956594323873122, 'eval/latency_8/demon_left_precision': 0.5296703296703297, 'eval/latency_8/demon_left_recall': 0.5160599571734475, 'eval/latency_8/demon_left_f1': 0.5227765726681128, 'eval/latency_8/demon_left_rate_model': 0.12659988870339456, 'eval/latency_8/demon_left_rate_teacher': 0.12993878686700056, 'eval/latency_8/demon_right_precision': 0.9699248120300752, 'eval/latency_8/demon_right_recall': 0.19545454545454546, 'eval/latency_8/demon_right_f1': 0.32534678436317777, 'eval/latency_8/demon_right_rate_model': 0.037006121313299944, 'eval/latency_8/demon_right_rate_teacher': 0.18363939899833054, 'eval/latency_8/demon_move_f1': 0.4240616785156453, 'eval/latency_8/demon_cc_f1': 0.6905872511040223, 'timing/data': 0.04566219728440046, 'timing/model': 0.9296453511342406, 'learning_rate/qwen_vl_interface': 9.877445642699026e-06, 'learning_rate/action_model': 9.877445642699026e-05, 'epoch': 0.27}) +[2026-06-17 17:43:53,659][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_600_state +[2026-06-17 17:43:53,665][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 17:44:01,313][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_600_state/model.safetensors +[2026-06-17 17:44:13,980][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_600_state/optimizer.bin +[2026-06-17 17:44:13,981][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_600_state/sampler.bin +[2026-06-17 17:44:13,981][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_600_state/sampler_1.bin +[2026-06-17 17:44:13,983][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_600_state/random_states_0.pkl +[2026-06-17 17:44:14,001][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 600 +[2026-06-17 17:44:14,001][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 17:44:14,014][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 17:44:18,288][starVLA.training.train_starvla][INFO] - Step 601, Loss: {'train/loss': 0.5958181619644165, 'train/grad_norm_pre_clip': 6.517453193664551, 'timing/data': 0.04599198047071695, 'timing/model': 0.9602379575371742, 'learning_rate/qwen_vl_interface': 9.876957064617951e-06, 'learning_rate/action_model': 9.876957064617951e-05, 'epoch': 0.27}) +[2026-06-17 17:44:22,238][starVLA.training.train_starvla][INFO] - Step 602, Loss: {'train/loss': 1.0229905843734741, 'train/grad_norm_pre_clip': 9.321106910705566, 'timing/data': 0.040504978969693184, 'timing/model': 0.9676679922267795, 'learning_rate/qwen_vl_interface': 9.876467527363696e-06, 'learning_rate/action_model': 9.876467527363697e-05, 'epoch': 0.27}) +[2026-06-17 17:44:26,046][starVLA.training.train_starvla][INFO] - Step 603, Loss: {'train/loss': 0.9610300660133362, 'train/grad_norm_pre_clip': 9.572052955627441, 'timing/data': 0.054562315344810486, 'timing/model': 1.1266417214646935, 'learning_rate/qwen_vl_interface': 9.875977031037746e-06, 'learning_rate/action_model': 9.875977031037745e-05, 'epoch': 0.27}) +[2026-06-17 17:44:29,718][starVLA.training.train_starvla][INFO] - Step 604, Loss: {'train/loss': 0.638806164264679, 'train/grad_norm_pre_clip': 7.844567775726318, 'timing/data': 0.04242536146193743, 'timing/model': 1.0192188657820225, 'learning_rate/qwen_vl_interface': 9.875485575741779e-06, 'learning_rate/action_model': 9.875485575741779e-05, 'epoch': 0.27}) +[2026-06-17 17:44:33,279][starVLA.training.train_starvla][INFO] - Step 605, Loss: {'train/loss': 0.5452414155006409, 'train/grad_norm_pre_clip': 6.90620756149292, 'timing/data': 0.04858262836933136, 'timing/model': 0.9165229490026832, 'learning_rate/qwen_vl_interface': 9.874993161577674e-06, 'learning_rate/action_model': 9.874993161577675e-05, 'epoch': 0.27}) +[2026-06-17 17:44:37,020][starVLA.training.train_starvla][INFO] - Step 606, Loss: {'train/loss': 0.9557796716690063, 'train/grad_norm_pre_clip': 6.9356489181518555, 'timing/data': 0.03773725684732199, 'timing/model': 0.9221946336328983, 'learning_rate/qwen_vl_interface': 9.874499788647512e-06, 'learning_rate/action_model': 9.874499788647513e-05, 'epoch': 0.27}) +[2026-06-17 17:44:40,891][starVLA.training.train_starvla][INFO] - Step 607, Loss: {'train/loss': 0.7886481881141663, 'train/grad_norm_pre_clip': 7.817336082458496, 'timing/data': 0.03685434814542532, 'timing/model': 0.9323961706832051, 'learning_rate/qwen_vl_interface': 9.874005457053565e-06, 'learning_rate/action_model': 9.874005457053565e-05, 'epoch': 0.28}) +[2026-06-17 17:44:44,640][starVLA.training.train_starvla][INFO] - Step 608, Loss: {'train/loss': 0.6027791500091553, 'train/grad_norm_pre_clip': 7.322215557098389, 'timing/data': 0.04186581540852785, 'timing/model': 0.9152749767526984, 'learning_rate/qwen_vl_interface': 9.873510166898315e-06, 'learning_rate/action_model': 9.873510166898314e-05, 'epoch': 0.28}) +[2026-06-17 17:44:48,432][starVLA.training.train_starvla][INFO] - Step 609, Loss: {'train/loss': 0.8348042964935303, 'train/grad_norm_pre_clip': 7.022736072540283, 'timing/data': 0.052131119184195995, 'timing/model': 0.9271794874221087, 'learning_rate/qwen_vl_interface': 9.87301391828443e-06, 'learning_rate/action_model': 9.873013918284428e-05, 'epoch': 0.28}) +[2026-06-17 17:44:52,243][starVLA.training.train_starvla][INFO] - Step 610, Loss: {'train/loss': 1.098297357559204, 'train/grad_norm_pre_clip': 6.604660987854004, 'timing/data': 0.03482500556856394, 'timing/model': 1.1046114694327116, 'learning_rate/qwen_vl_interface': 9.872516711314784e-06, 'learning_rate/action_model': 9.872516711314784e-05, 'epoch': 0.28}) +[2026-06-17 17:44:56,046][starVLA.training.train_starvla][INFO] - Step 611, Loss: {'train/loss': 0.8059119582176208, 'train/grad_norm_pre_clip': 6.084564685821533, 'timing/data': 0.0411835890263319, 'timing/model': 0.9358674110844731, 'learning_rate/qwen_vl_interface': 9.872018546092452e-06, 'learning_rate/action_model': 9.872018546092451e-05, 'epoch': 0.28}) +[2026-06-17 17:44:59,942][starVLA.training.train_starvla][INFO] - Step 612, Loss: {'train/loss': 0.6071312427520752, 'train/grad_norm_pre_clip': 6.811075687408447, 'timing/data': 0.043998329900205135, 'timing/model': 0.9374107755720615, 'learning_rate/qwen_vl_interface': 9.871519422720699e-06, 'learning_rate/action_model': 9.871519422720698e-05, 'epoch': 0.28}) +[2026-06-17 17:45:03,844][starVLA.training.train_starvla][INFO] - Step 613, Loss: {'train/loss': 0.9749133586883545, 'train/grad_norm_pre_clip': 7.441717624664307, 'timing/data': 0.034192739985883236, 'timing/model': 0.9739266680553555, 'learning_rate/qwen_vl_interface': 9.871019341303e-06, 'learning_rate/action_model': 9.871019341303e-05, 'epoch': 0.28}) +[2026-06-17 17:45:08,178][starVLA.training.train_starvla][INFO] - Step 614, Loss: {'train/loss': 0.46530526876449585, 'train/grad_norm_pre_clip': 4.955909729003906, 'timing/data': 0.03135163150727749, 'timing/model': 0.991637859493494, 'learning_rate/qwen_vl_interface': 9.870518301943018e-06, 'learning_rate/action_model': 9.870518301943018e-05, 'epoch': 0.28}) +[2026-06-17 17:45:12,238][starVLA.training.train_starvla][INFO] - Step 615, Loss: {'train/loss': 0.5036803483963013, 'train/grad_norm_pre_clip': 4.049597263336182, 'timing/data': 0.03274123929440975, 'timing/model': 0.9445313168689609, 'learning_rate/qwen_vl_interface': 9.870016304744617e-06, 'learning_rate/action_model': 9.870016304744616e-05, 'epoch': 0.28}) +[2026-06-17 17:45:16,321][starVLA.training.train_starvla][INFO] - Step 616, Loss: {'train/loss': 0.9330711364746094, 'train/grad_norm_pre_clip': 10.025527000427246, 'timing/data': 0.027792188338935375, 'timing/model': 1.1262542540207505, 'learning_rate/qwen_vl_interface': 9.869513349811868e-06, 'learning_rate/action_model': 9.869513349811868e-05, 'epoch': 0.28}) +[2026-06-17 17:45:20,149][starVLA.training.train_starvla][INFO] - Step 617, Loss: {'train/loss': 1.1828254461288452, 'train/grad_norm_pre_clip': 24.96915626525879, 'timing/data': 0.02885212656110525, 'timing/model': 1.1285606110468507, 'learning_rate/qwen_vl_interface': 9.86900943724903e-06, 'learning_rate/action_model': 9.869009437249031e-05, 'epoch': 0.28}) +[2026-06-17 17:45:24,160][starVLA.training.train_starvla][INFO] - Step 618, Loss: {'train/loss': 0.8477551937103271, 'train/grad_norm_pre_clip': 11.368253707885742, 'timing/data': 0.03891291934996843, 'timing/model': 1.2680062679573894, 'learning_rate/qwen_vl_interface': 9.868504567160566e-06, 'learning_rate/action_model': 9.868504567160565e-05, 'epoch': 0.28}) +[2026-06-17 17:45:28,301][starVLA.training.train_starvla][INFO] - Step 619, Loss: {'train/loss': 1.1648777723312378, 'train/grad_norm_pre_clip': 8.333535194396973, 'timing/data': 0.046512178145349026, 'timing/model': 0.9477130444720387, 'learning_rate/qwen_vl_interface': 9.867998739651134e-06, 'learning_rate/action_model': 9.867998739651134e-05, 'epoch': 0.28}) +[2026-06-17 17:45:32,204][starVLA.training.train_starvla][INFO] - Step 620, Loss: {'train/loss': 1.0462933778762817, 'train/grad_norm_pre_clip': 9.197711944580078, 'timing/data': 0.030190762132406235, 'timing/model': 0.9596674777567387, 'learning_rate/qwen_vl_interface': 9.867491954825595e-06, 'learning_rate/action_model': 9.867491954825595e-05, 'epoch': 0.28}) +[2026-06-17 17:45:36,049][starVLA.training.train_starvla][INFO] - Step 621, Loss: {'train/loss': 0.8372715711593628, 'train/grad_norm_pre_clip': 8.44796371459961, 'timing/data': 0.04132381360977888, 'timing/model': 0.9446207350119948, 'learning_rate/qwen_vl_interface': 9.866984212789007e-06, 'learning_rate/action_model': 9.866984212789006e-05, 'epoch': 0.28}) +[2026-06-17 17:45:39,954][starVLA.training.train_starvla][INFO] - Step 622, Loss: {'train/loss': 0.6489406824111938, 'train/grad_norm_pre_clip': 9.585726737976074, 'timing/data': 0.027340522035956383, 'timing/model': 0.9360069260001183, 'learning_rate/qwen_vl_interface': 9.86647551364662e-06, 'learning_rate/action_model': 9.86647551364662e-05, 'epoch': 0.28}) +[2026-06-17 17:45:44,320][starVLA.training.train_starvla][INFO] - Step 623, Loss: {'train/loss': 0.923021674156189, 'train/grad_norm_pre_clip': 16.89824676513672, 'timing/data': 0.05440633464604616, 'timing/model': 1.580644248984754, 'learning_rate/qwen_vl_interface': 9.865965857503893e-06, 'learning_rate/action_model': 9.865965857503894e-05, 'epoch': 0.28}) +[2026-06-17 17:45:49,087][starVLA.training.train_starvla][INFO] - Step 624, Loss: {'train/loss': 0.8534873127937317, 'train/grad_norm_pre_clip': 9.124197959899902, 'timing/data': 0.03501562401652336, 'timing/model': 0.947640098631382, 'learning_rate/qwen_vl_interface': 9.865455244466478e-06, 'learning_rate/action_model': 9.865455244466477e-05, 'epoch': 0.28}) +[2026-06-17 17:45:53,990][starVLA.training.train_starvla][INFO] - Step 625, Loss: {'train/loss': 0.9730201363563538, 'train/grad_norm_pre_clip': 9.470258712768555, 'timing/data': 0.0346039617434144, 'timing/model': 1.9757278207689524, 'learning_rate/qwen_vl_interface': 9.864943674640223e-06, 'learning_rate/action_model': 9.864943674640222e-05, 'epoch': 0.28}) +[2026-06-17 17:45:59,412][starVLA.training.train_starvla][INFO] - Step 626, Loss: {'train/loss': 0.6935924291610718, 'train/grad_norm_pre_clip': 8.206338882446289, 'timing/data': 0.04340555798262358, 'timing/model': 0.9360880954191089, 'learning_rate/qwen_vl_interface': 9.864431148131178e-06, 'learning_rate/action_model': 9.864431148131178e-05, 'epoch': 0.28}) +[2026-06-17 17:46:03,266][starVLA.training.train_starvla][INFO] - Step 627, Loss: {'train/loss': 0.7311216592788696, 'train/grad_norm_pre_clip': 7.65675163269043, 'timing/data': 0.032727669924497604, 'timing/model': 0.930775853805244, 'learning_rate/qwen_vl_interface': 9.863917665045591e-06, 'learning_rate/action_model': 9.863917665045592e-05, 'epoch': 0.28}) +[2026-06-17 17:46:07,347][starVLA.training.train_starvla][INFO] - Step 628, Loss: {'train/loss': 0.5665274858474731, 'train/grad_norm_pre_clip': 7.790591239929199, 'timing/data': 0.03512255847454071, 'timing/model': 0.9389084475114942, 'learning_rate/qwen_vl_interface': 9.863403225489909e-06, 'learning_rate/action_model': 9.863403225489909e-05, 'epoch': 0.28}) +[2026-06-17 17:46:11,144][starVLA.training.train_starvla][INFO] - Step 629, Loss: {'train/loss': 0.9805911779403687, 'train/grad_norm_pre_clip': 8.665994644165039, 'timing/data': 0.03633613046258688, 'timing/model': 0.9259910797700286, 'learning_rate/qwen_vl_interface': 9.862887829570771e-06, 'learning_rate/action_model': 9.862887829570772e-05, 'epoch': 0.29}) +[2026-06-17 17:46:14,960][starVLA.training.train_starvla][INFO] - Step 630, Loss: {'train/loss': 0.7066987752914429, 'train/grad_norm_pre_clip': 6.766645908355713, 'timing/data': 0.0339306341484189, 'timing/model': 1.1468699956312776, 'learning_rate/qwen_vl_interface': 9.862371477395024e-06, 'learning_rate/action_model': 9.862371477395025e-05, 'epoch': 0.29}) +[2026-06-17 17:46:18,530][starVLA.training.train_starvla][INFO] - Step 631, Loss: {'train/loss': 0.7903378009796143, 'train/grad_norm_pre_clip': 9.69349193572998, 'timing/data': 0.03263674210757017, 'timing/model': 0.9333431981503963, 'learning_rate/qwen_vl_interface': 9.861854169069708e-06, 'learning_rate/action_model': 9.861854169069707e-05, 'epoch': 0.29}) +[2026-06-17 17:46:22,628][starVLA.training.train_starvla][INFO] - Step 632, Loss: {'train/loss': 0.5424765944480896, 'train/grad_norm_pre_clip': 7.126346588134766, 'timing/data': 0.03725737053900957, 'timing/model': 1.0387520417571068, 'learning_rate/qwen_vl_interface': 9.861335904702058e-06, 'learning_rate/action_model': 9.861335904702058e-05, 'epoch': 0.29}) +[2026-06-17 17:46:26,749][starVLA.training.train_starvla][INFO] - Step 633, Loss: {'train/loss': 0.6913629770278931, 'train/grad_norm_pre_clip': 6.909070014953613, 'timing/data': 0.04392281547188759, 'timing/model': 0.942312803119421, 'learning_rate/qwen_vl_interface': 9.860816684399514e-06, 'learning_rate/action_model': 9.860816684399514e-05, 'epoch': 0.29}) +[2026-06-17 17:46:33,957][starVLA.training.train_starvla][INFO] - Step 634, Loss: {'train/loss': 0.8502694368362427, 'train/grad_norm_pre_clip': 7.293895244598389, 'timing/data': 0.0421022055670619, 'timing/model': 0.9432794880121946, 'learning_rate/qwen_vl_interface': 9.86029650826971e-06, 'learning_rate/action_model': 9.86029650826971e-05, 'epoch': 0.29}) +[2026-06-17 17:46:38,346][starVLA.training.train_starvla][INFO] - Step 635, Loss: {'train/loss': 0.9392120242118835, 'train/grad_norm_pre_clip': 7.647538185119629, 'timing/data': 0.0487827779725194, 'timing/model': 0.9333495730534196, 'learning_rate/qwen_vl_interface': 9.859775376420478e-06, 'learning_rate/action_model': 9.859775376420478e-05, 'epoch': 0.29}) +[2026-06-17 17:46:42,390][starVLA.training.train_starvla][INFO] - Step 636, Loss: {'train/loss': 0.6585305333137512, 'train/grad_norm_pre_clip': 23.385971069335938, 'timing/data': 0.034506771713495255, 'timing/model': 1.218255527317524, 'learning_rate/qwen_vl_interface': 9.85925328895985e-06, 'learning_rate/action_model': 9.85925328895985e-05, 'epoch': 0.29}) +[2026-06-17 17:46:46,445][starVLA.training.train_starvla][INFO] - Step 637, Loss: {'train/loss': 0.7575176954269409, 'train/grad_norm_pre_clip': 7.619141578674316, 'timing/data': 0.043775844387710094, 'timing/model': 1.2735113687813282, 'learning_rate/qwen_vl_interface': 9.858730245996057e-06, 'learning_rate/action_model': 9.858730245996057e-05, 'epoch': 0.29}) +[2026-06-17 17:46:50,372][starVLA.training.train_starvla][INFO] - Step 638, Loss: {'train/loss': 0.5805845260620117, 'train/grad_norm_pre_clip': 5.605257034301758, 'timing/data': 0.034679693169891834, 'timing/model': 1.0967925228178501, 'learning_rate/qwen_vl_interface': 9.858206247637522e-06, 'learning_rate/action_model': 9.858206247637523e-05, 'epoch': 0.29}) +[2026-06-17 17:46:56,956][starVLA.training.train_starvla][INFO] - Step 639, Loss: {'train/loss': 0.9436478018760681, 'train/grad_norm_pre_clip': 12.251842498779297, 'timing/data': 0.04427632316946983, 'timing/model': 0.9323913082480431, 'learning_rate/qwen_vl_interface': 9.857681293992876e-06, 'learning_rate/action_model': 9.857681293992877e-05, 'epoch': 0.29}) +[2026-06-17 17:47:01,394][starVLA.training.train_starvla][INFO] - Step 640, Loss: {'train/loss': 0.8823858499526978, 'train/grad_norm_pre_clip': 9.892046928405762, 'timing/data': 0.03898444585502148, 'timing/model': 0.9631613055244088, 'learning_rate/qwen_vl_interface': 9.85715538517094e-06, 'learning_rate/action_model': 9.857155385170938e-05, 'epoch': 0.29}) +[2026-06-17 17:47:06,289][starVLA.training.train_starvla][INFO] - Step 641, Loss: {'train/loss': 0.7530630826950073, 'train/grad_norm_pre_clip': 8.178603172302246, 'timing/data': 0.03397664614021778, 'timing/model': 0.951720411889255, 'learning_rate/qwen_vl_interface': 9.856628521280733e-06, 'learning_rate/action_model': 9.856628521280733e-05, 'epoch': 0.29}) +[2026-06-17 17:47:10,409][starVLA.training.train_starvla][INFO] - Step 642, Loss: {'train/loss': 0.8012725114822388, 'train/grad_norm_pre_clip': 11.388922691345215, 'timing/data': 0.039755732752382755, 'timing/model': 1.0205124355852604, 'learning_rate/qwen_vl_interface': 9.856100702431477e-06, 'learning_rate/action_model': 9.856100702431477e-05, 'epoch': 0.29}) +[2026-06-17 17:47:14,437][starVLA.training.train_starvla][INFO] - Step 643, Loss: {'train/loss': 0.4773035943508148, 'train/grad_norm_pre_clip': 6.18652868270874, 'timing/data': 0.03738114982843399, 'timing/model': 1.239390468224883, 'learning_rate/qwen_vl_interface': 9.85557192873259e-06, 'learning_rate/action_model': 9.85557192873259e-05, 'epoch': 0.29}) +[2026-06-17 17:47:18,427][starVLA.training.train_starvla][INFO] - Step 644, Loss: {'train/loss': 0.6484634280204773, 'train/grad_norm_pre_clip': 6.3455915451049805, 'timing/data': 0.03369605261832476, 'timing/model': 1.2175363609567285, 'learning_rate/qwen_vl_interface': 9.855042200293684e-06, 'learning_rate/action_model': 9.855042200293685e-05, 'epoch': 0.29}) +[2026-06-17 17:47:22,243][starVLA.training.train_starvla][INFO] - Step 645, Loss: {'train/loss': 0.8605671525001526, 'train/grad_norm_pre_clip': 10.171309471130371, 'timing/data': 0.03937931917607784, 'timing/model': 0.9808085514232516, 'learning_rate/qwen_vl_interface': 9.854511517224578e-06, 'learning_rate/action_model': 9.854511517224578e-05, 'epoch': 0.29}) +[2026-06-17 17:47:26,221][starVLA.training.train_starvla][INFO] - Step 646, Loss: {'train/loss': 0.8600398302078247, 'train/grad_norm_pre_clip': 7.864853858947754, 'timing/data': 0.037612926214933395, 'timing/model': 0.9539128728210926, 'learning_rate/qwen_vl_interface': 9.853979879635278e-06, 'learning_rate/action_model': 9.853979879635278e-05, 'epoch': 0.29}) +[2026-06-17 17:47:30,115][starVLA.training.train_starvla][INFO] - Step 647, Loss: {'train/loss': 0.5842248797416687, 'train/grad_norm_pre_clip': 5.288127422332764, 'timing/data': 0.03941345680505037, 'timing/model': 0.9309949707239866, 'learning_rate/qwen_vl_interface': 9.853447287635995e-06, 'learning_rate/action_model': 9.853447287635995e-05, 'epoch': 0.29}) +[2026-06-17 17:47:34,378][starVLA.training.train_starvla][INFO] - Step 648, Loss: {'train/loss': 0.7919163107872009, 'train/grad_norm_pre_clip': 9.718945503234863, 'timing/data': 0.029947512783110142, 'timing/model': 0.9743104483932257, 'learning_rate/qwen_vl_interface': 9.852913741337136e-06, 'learning_rate/action_model': 9.852913741337136e-05, 'epoch': 0.29}) +[2026-06-17 17:47:38,328][starVLA.training.train_starvla][INFO] - Step 649, Loss: {'train/loss': 0.7724986672401428, 'train/grad_norm_pre_clip': 9.126934051513672, 'timing/data': 0.039387524127960205, 'timing/model': 0.9897149037569761, 'learning_rate/qwen_vl_interface': 9.852379240849306e-06, 'learning_rate/action_model': 9.852379240849305e-05, 'epoch': 0.29}) +[2026-06-17 17:47:42,236][starVLA.training.train_starvla][INFO] - Step 650, Loss: {'train/loss': 0.7276791930198669, 'train/grad_norm_pre_clip': 7.001543045043945, 'timing/data': 0.04645045567303896, 'timing/model': 1.1818903358653188, 'learning_rate/qwen_vl_interface': 9.851843786283306e-06, 'learning_rate/action_model': 9.851843786283306e-05, 'epoch': 0.29}) +[2026-06-17 17:47:46,103][starVLA.training.train_starvla][INFO] - Step 651, Loss: {'train/loss': 0.7780675292015076, 'train/grad_norm_pre_clip': 9.337589263916016, 'timing/data': 0.03673319797962904, 'timing/model': 1.1410296941176057, 'learning_rate/qwen_vl_interface': 9.851307377750137e-06, 'learning_rate/action_model': 9.851307377750137e-05, 'epoch': 0.3}) +[2026-06-17 17:47:49,809][starVLA.training.train_starvla][INFO] - Step 652, Loss: {'train/loss': 0.5186216235160828, 'train/grad_norm_pre_clip': 8.909284591674805, 'timing/data': 0.045833293348550797, 'timing/model': 0.9579856358468533, 'learning_rate/qwen_vl_interface': 9.850770015360999e-06, 'learning_rate/action_model': 9.850770015360999e-05, 'epoch': 0.3}) +[2026-06-17 17:47:53,597][starVLA.training.train_starvla][INFO] - Step 653, Loss: {'train/loss': 0.5737001299858093, 'train/grad_norm_pre_clip': 5.949669361114502, 'timing/data': 0.030106225050985813, 'timing/model': 0.9193526795133948, 'learning_rate/qwen_vl_interface': 9.850231699227284e-06, 'learning_rate/action_model': 9.850231699227284e-05, 'epoch': 0.3}) +[2026-06-17 17:47:57,449][starVLA.training.train_starvla][INFO] - Step 654, Loss: {'train/loss': 0.525147557258606, 'train/grad_norm_pre_clip': 7.223473072052002, 'timing/data': 0.037110960111021996, 'timing/model': 0.9263817351311445, 'learning_rate/qwen_vl_interface': 9.84969242946059e-06, 'learning_rate/action_model': 9.84969242946059e-05, 'epoch': 0.3}) +[2026-06-17 17:48:01,199][starVLA.training.train_starvla][INFO] - Step 655, Loss: {'train/loss': 0.8316909670829773, 'train/grad_norm_pre_clip': 6.303562164306641, 'timing/data': 0.03594698663800955, 'timing/model': 0.9369025994092226, 'learning_rate/qwen_vl_interface': 9.849152206172704e-06, 'learning_rate/action_model': 9.849152206172704e-05, 'epoch': 0.3}) +[2026-06-17 17:48:05,118][starVLA.training.train_starvla][INFO] - Step 656, Loss: {'train/loss': 0.6951183676719666, 'train/grad_norm_pre_clip': 5.678740978240967, 'timing/data': 0.034512417390942574, 'timing/model': 0.9316598102450371, 'learning_rate/qwen_vl_interface': 9.84861102947562e-06, 'learning_rate/action_model': 9.848611029475619e-05, 'epoch': 0.3}) +[2026-06-17 17:48:08,936][starVLA.training.train_starvla][INFO] - Step 657, Loss: {'train/loss': 0.4892505705356598, 'train/grad_norm_pre_clip': 7.5279340744018555, 'timing/data': 0.032183642499148846, 'timing/model': 1.149035363458097, 'learning_rate/qwen_vl_interface': 9.848068899481516e-06, 'learning_rate/action_model': 9.848068899481516e-05, 'epoch': 0.3}) +[2026-06-17 17:48:12,809][starVLA.training.train_starvla][INFO] - Step 658, Loss: {'train/loss': 0.705328106880188, 'train/grad_norm_pre_clip': 7.351140022277832, 'timing/data': 0.03804670553654432, 'timing/model': 1.1106695588678122, 'learning_rate/qwen_vl_interface': 9.847525816302784e-06, 'learning_rate/action_model': 9.847525816302784e-05, 'epoch': 0.3}) +[2026-06-17 17:48:16,540][starVLA.training.train_starvla][INFO] - Step 659, Loss: {'train/loss': 0.5480383038520813, 'train/grad_norm_pre_clip': 7.4845757484436035, 'timing/data': 0.05437207501381636, 'timing/model': 0.9392766784876585, 'learning_rate/qwen_vl_interface': 9.846981780052005e-06, 'learning_rate/action_model': 9.846981780052003e-05, 'epoch': 0.3}) +[2026-06-17 17:48:20,350][starVLA.training.train_starvla][INFO] - Step 660, Loss: {'train/loss': 0.7803357243537903, 'train/grad_norm_pre_clip': 9.348943710327148, 'timing/data': 0.043119678273797035, 'timing/model': 0.937707687728107, 'learning_rate/qwen_vl_interface': 9.846436790841955e-06, 'learning_rate/action_model': 9.846436790841955e-05, 'epoch': 0.3}) +[2026-06-17 17:48:24,452][starVLA.training.train_starvla][INFO] - Step 661, Loss: {'train/loss': 0.8969769477844238, 'train/grad_norm_pre_clip': 7.6873908042907715, 'timing/data': 0.0350324222818017, 'timing/model': 0.9350204262882471, 'learning_rate/qwen_vl_interface': 9.84589084878561e-06, 'learning_rate/action_model': 9.845890848785611e-05, 'epoch': 0.3}) +[2026-06-17 17:48:28,258][starVLA.training.train_starvla][INFO] - Step 662, Loss: {'train/loss': 0.7839111685752869, 'train/grad_norm_pre_clip': 7.887874603271484, 'timing/data': 0.04392290860414505, 'timing/model': 0.9211890967562795, 'learning_rate/qwen_vl_interface': 9.84534395399615e-06, 'learning_rate/action_model': 9.84534395399615e-05, 'epoch': 0.3}) +[2026-06-17 17:48:32,328][starVLA.training.train_starvla][INFO] - Step 663, Loss: {'train/loss': 0.8437802791595459, 'train/grad_norm_pre_clip': 9.532790184020996, 'timing/data': 0.03207192197442055, 'timing/model': 0.93830830976367, 'learning_rate/qwen_vl_interface': 9.844796106586942e-06, 'learning_rate/action_model': 9.844796106586941e-05, 'epoch': 0.3}) +[2026-06-17 17:48:36,409][starVLA.training.train_starvla][INFO] - Step 664, Loss: {'train/loss': 0.8454863429069519, 'train/grad_norm_pre_clip': 6.379021167755127, 'timing/data': 0.04410244897007942, 'timing/model': 1.241515664383769, 'learning_rate/qwen_vl_interface': 9.844247306671557e-06, 'learning_rate/action_model': 9.844247306671555e-05, 'epoch': 0.3}) +[2026-06-17 17:48:40,249][starVLA.training.train_starvla][INFO] - Step 665, Loss: {'train/loss': 0.7905277609825134, 'train/grad_norm_pre_clip': 8.00151538848877, 'timing/data': 0.040666467510163784, 'timing/model': 1.0945979990065098, 'learning_rate/qwen_vl_interface': 9.843697554363762e-06, 'learning_rate/action_model': 9.843697554363762e-05, 'epoch': 0.3}) +[2026-06-17 17:48:44,014][starVLA.training.train_starvla][INFO] - Step 666, Loss: {'train/loss': 0.6743372678756714, 'train/grad_norm_pre_clip': 6.6181464195251465, 'timing/data': 0.03650410659611225, 'timing/model': 1.0521254586055875, 'learning_rate/qwen_vl_interface': 9.84314684977752e-06, 'learning_rate/action_model': 9.84314684977752e-05, 'epoch': 0.3}) +[2026-06-17 17:48:48,204][starVLA.training.train_starvla][INFO] - Step 667, Loss: {'train/loss': 0.6602128148078918, 'train/grad_norm_pre_clip': 7.091075897216797, 'timing/data': 0.04299178998917341, 'timing/model': 1.204908979125321, 'learning_rate/qwen_vl_interface': 9.842595193026993e-06, 'learning_rate/action_model': 9.842595193026993e-05, 'epoch': 0.3}) +[2026-06-17 17:48:52,290][starVLA.training.train_starvla][INFO] - Step 668, Loss: {'train/loss': 0.7681955695152283, 'train/grad_norm_pre_clip': 8.44912052154541, 'timing/data': 0.04611356649547815, 'timing/model': 1.0829747263342142, 'learning_rate/qwen_vl_interface': 9.842042584226545e-06, 'learning_rate/action_model': 9.842042584226544e-05, 'epoch': 0.3}) +[2026-06-17 17:48:56,349][starVLA.training.train_starvla][INFO] - Step 669, Loss: {'train/loss': 0.7560688257217407, 'train/grad_norm_pre_clip': 7.9052629470825195, 'timing/data': 0.04115327540785074, 'timing/model': 0.9259291179478168, 'learning_rate/qwen_vl_interface': 9.841489023490723e-06, 'learning_rate/action_model': 9.841489023490723e-05, 'epoch': 0.3}) +[2026-06-17 17:49:00,242][starVLA.training.train_starvla][INFO] - Step 670, Loss: {'train/loss': 0.8763030171394348, 'train/grad_norm_pre_clip': 8.788037300109863, 'timing/data': 0.03622131422162056, 'timing/model': 0.9355668015778065, 'learning_rate/qwen_vl_interface': 9.840934510934288e-06, 'learning_rate/action_model': 9.840934510934288e-05, 'epoch': 0.3}) +[2026-06-17 17:49:04,332][starVLA.training.train_starvla][INFO] - Step 671, Loss: {'train/loss': 1.1338974237442017, 'train/grad_norm_pre_clip': 8.00225830078125, 'timing/data': 0.03229174017906189, 'timing/model': 1.0410831859335303, 'learning_rate/qwen_vl_interface': 9.840379046672189e-06, 'learning_rate/action_model': 9.840379046672189e-05, 'epoch': 0.3}) +[2026-06-17 17:49:08,177][starVLA.training.train_starvla][INFO] - Step 672, Loss: {'train/loss': 1.0588129758834839, 'train/grad_norm_pre_clip': 9.850021362304688, 'timing/data': 0.03273338917642832, 'timing/model': 1.165885215625167, 'learning_rate/qwen_vl_interface': 9.839822630819573e-06, 'learning_rate/action_model': 9.839822630819572e-05, 'epoch': 0.3}) +[2026-06-17 17:49:11,995][starVLA.training.train_starvla][INFO] - Step 673, Loss: {'train/loss': 0.6799504160881042, 'train/grad_norm_pre_clip': 7.904260635375977, 'timing/data': 0.03494112193584442, 'timing/model': 1.0158962104469538, 'learning_rate/qwen_vl_interface': 9.839265263491789e-06, 'learning_rate/action_model': 9.839265263491789e-05, 'epoch': 0.31}) +[2026-06-17 17:49:15,858][starVLA.training.train_starvla][INFO] - Step 674, Loss: {'train/loss': 0.7573612332344055, 'train/grad_norm_pre_clip': 6.975850582122803, 'timing/data': 0.04123515170067549, 'timing/model': 0.939298459328711, 'learning_rate/qwen_vl_interface': 9.838706944804376e-06, 'learning_rate/action_model': 9.838706944804374e-05, 'epoch': 0.31}) +[2026-06-17 17:49:19,746][starVLA.training.train_starvla][INFO] - Step 675, Loss: {'train/loss': 0.9731947183609009, 'train/grad_norm_pre_clip': 8.053873062133789, 'timing/data': 0.03031099960207939, 'timing/model': 0.9352039797231555, 'learning_rate/qwen_vl_interface': 9.838147674873076e-06, 'learning_rate/action_model': 9.838147674873075e-05, 'epoch': 0.31}) +[2026-06-17 17:49:23,492][starVLA.training.train_starvla][INFO] - Step 676, Loss: {'train/loss': 0.6787304878234863, 'train/grad_norm_pre_clip': 5.837159633636475, 'timing/data': 0.03543484024703503, 'timing/model': 0.936734838411212, 'learning_rate/qwen_vl_interface': 9.837587453813826e-06, 'learning_rate/action_model': 9.837587453813824e-05, 'epoch': 0.31}) +[2026-06-17 17:49:27,603][starVLA.training.train_starvla][INFO] - Step 677, Loss: {'train/loss': 0.5702409744262695, 'train/grad_norm_pre_clip': 4.944546222686768, 'timing/data': 0.033831908367574215, 'timing/model': 0.9317794535309076, 'learning_rate/qwen_vl_interface': 9.83702628174276e-06, 'learning_rate/action_model': 9.837026281742759e-05, 'epoch': 0.31}) +[2026-06-17 17:49:34,173][starVLA.training.train_starvla][INFO] - Step 678, Loss: {'train/loss': 0.7856200933456421, 'train/grad_norm_pre_clip': 6.879722595214844, 'timing/data': 0.050320494920015335, 'timing/model': 1.2668427946045995, 'learning_rate/qwen_vl_interface': 9.83646415877621e-06, 'learning_rate/action_model': 9.83646415877621e-05, 'epoch': 0.31}) +[2026-06-17 17:49:38,484][starVLA.training.train_starvla][INFO] - Step 679, Loss: {'train/loss': 0.6568477153778076, 'train/grad_norm_pre_clip': 5.961904048919678, 'timing/data': 0.042666202411055565, 'timing/model': 0.9611053401604295, 'learning_rate/qwen_vl_interface': 9.835901085030703e-06, 'learning_rate/action_model': 9.835901085030703e-05, 'epoch': 0.31}) +[2026-06-17 17:49:43,115][starVLA.training.train_starvla][INFO] - Step 680, Loss: {'train/loss': 0.5873878002166748, 'train/grad_norm_pre_clip': 7.058852672576904, 'timing/data': 0.05370876472443342, 'timing/model': 1.0214697811752558, 'learning_rate/qwen_vl_interface': 9.83533706062297e-06, 'learning_rate/action_model': 9.835337060622969e-05, 'epoch': 0.31}) +[2026-06-17 17:49:46,841][starVLA.training.train_starvla][INFO] - Step 681, Loss: {'train/loss': 0.7329170107841492, 'train/grad_norm_pre_clip': 7.064957618713379, 'timing/data': 0.03462590463459492, 'timing/model': 0.9243021309375763, 'learning_rate/qwen_vl_interface': 9.834772085669927e-06, 'learning_rate/action_model': 9.834772085669927e-05, 'epoch': 0.31}) +[2026-06-17 17:49:50,596][starVLA.training.train_starvla][INFO] - Step 682, Loss: {'train/loss': 0.6243849396705627, 'train/grad_norm_pre_clip': 5.99198579788208, 'timing/data': 0.03078940138220787, 'timing/model': 0.9337908690795302, 'learning_rate/qwen_vl_interface': 9.834206160288698e-06, 'learning_rate/action_model': 9.834206160288698e-05, 'epoch': 0.31}) +[2026-06-17 17:49:55,040][starVLA.training.train_starvla][INFO] - Step 683, Loss: {'train/loss': 0.6622759699821472, 'train/grad_norm_pre_clip': 5.963551044464111, 'timing/data': 0.03799922298640013, 'timing/model': 0.9483742155134678, 'learning_rate/qwen_vl_interface': 9.8336392845966e-06, 'learning_rate/action_model': 9.833639284596599e-05, 'epoch': 0.31}) +[2026-06-17 17:49:58,991][starVLA.training.train_starvla][INFO] - Step 684, Loss: {'train/loss': 0.6625086665153503, 'train/grad_norm_pre_clip': 7.70699405670166, 'timing/data': 0.040886351838707924, 'timing/model': 1.2571588810533285, 'learning_rate/qwen_vl_interface': 9.833071458711146e-06, 'learning_rate/action_model': 9.833071458711146e-05, 'epoch': 0.31}) +[2026-06-17 17:50:03,061][starVLA.training.train_starvla][INFO] - Step 685, Loss: {'train/loss': 0.6999891400337219, 'train/grad_norm_pre_clip': 5.581284523010254, 'timing/data': 0.03231269959360361, 'timing/model': 1.2663206169381738, 'learning_rate/qwen_vl_interface': 9.832502682750048e-06, 'learning_rate/action_model': 9.832502682750047e-05, 'epoch': 0.31}) +[2026-06-17 17:50:06,767][starVLA.training.train_starvla][INFO] - Step 686, Loss: {'train/loss': 0.8978208303451538, 'train/grad_norm_pre_clip': 11.328206062316895, 'timing/data': 0.031999681144952774, 'timing/model': 0.9622537912800908, 'learning_rate/qwen_vl_interface': 9.831932956831212e-06, 'learning_rate/action_model': 9.831932956831211e-05, 'epoch': 0.31}) +[2026-06-17 17:50:10,863][starVLA.training.train_starvla][INFO] - Step 687, Loss: {'train/loss': 0.7465354800224304, 'train/grad_norm_pre_clip': 8.460742950439453, 'timing/data': 0.03513033036142588, 'timing/model': 1.1120273731648922, 'learning_rate/qwen_vl_interface': 9.831362281072743e-06, 'learning_rate/action_model': 9.831362281072742e-05, 'epoch': 0.31}) +[2026-06-17 17:50:15,521][starVLA.training.train_starvla][INFO] - Step 688, Loss: {'train/loss': 0.7506153583526611, 'train/grad_norm_pre_clip': 7.435019016265869, 'timing/data': 0.05920288432389498, 'timing/model': 1.1270975591614842, 'learning_rate/qwen_vl_interface': 9.830790655592944e-06, 'learning_rate/action_model': 9.830790655592943e-05, 'epoch': 0.31}) +[2026-06-17 17:50:19,253][starVLA.training.train_starvla][INFO] - Step 689, Loss: {'train/loss': 0.9458156824111938, 'train/grad_norm_pre_clip': 7.55910587310791, 'timing/data': 0.054866706021130085, 'timing/model': 0.9236098807305098, 'learning_rate/qwen_vl_interface': 9.830218080510313e-06, 'learning_rate/action_model': 9.830218080510312e-05, 'epoch': 0.31}) +[2026-06-17 17:50:23,064][starVLA.training.train_starvla][INFO] - Step 690, Loss: {'train/loss': 0.6608853936195374, 'train/grad_norm_pre_clip': 7.596677780151367, 'timing/data': 0.030015932396054268, 'timing/model': 0.9410891588777304, 'learning_rate/qwen_vl_interface': 9.829644555943545e-06, 'learning_rate/action_model': 9.829644555943545e-05, 'epoch': 0.31}) +[2026-06-17 17:50:27,040][starVLA.training.train_starvla][INFO] - Step 691, Loss: {'train/loss': 0.719590961933136, 'train/grad_norm_pre_clip': 5.295495510101318, 'timing/data': 0.03336478490382433, 'timing/model': 1.1932272743433714, 'learning_rate/qwen_vl_interface': 9.82907008201153e-06, 'learning_rate/action_model': 9.829070082011531e-05, 'epoch': 0.31}) +[2026-06-17 17:50:30,912][starVLA.training.train_starvla][INFO] - Step 692, Loss: {'train/loss': 0.8153260946273804, 'train/grad_norm_pre_clip': 7.150230407714844, 'timing/data': 0.05073979403823614, 'timing/model': 1.1200808649882674, 'learning_rate/qwen_vl_interface': 9.828494658833366e-06, 'learning_rate/action_model': 9.828494658833365e-05, 'epoch': 0.31}) +[2026-06-17 17:50:35,035][starVLA.training.train_starvla][INFO] - Step 693, Loss: {'train/loss': 0.6387926340103149, 'train/grad_norm_pre_clip': 15.129267692565918, 'timing/data': 0.03545678220689297, 'timing/model': 0.9316690117120743, 'learning_rate/qwen_vl_interface': 9.82791828652833e-06, 'learning_rate/action_model': 9.827918286528329e-05, 'epoch': 0.31}) +[2026-06-17 17:50:39,046][starVLA.training.train_starvla][INFO] - Step 694, Loss: {'train/loss': 1.0150066614151, 'train/grad_norm_pre_clip': 5.435070514678955, 'timing/data': 0.039860439486801624, 'timing/model': 1.11043565813452, 'learning_rate/qwen_vl_interface': 9.827340965215903e-06, 'learning_rate/action_model': 9.827340965215904e-05, 'epoch': 0.31}) +[2026-06-17 17:50:42,956][starVLA.training.train_starvla][INFO] - Step 695, Loss: {'train/loss': 0.673716127872467, 'train/grad_norm_pre_clip': 5.181984901428223, 'timing/data': 0.04948409087955952, 'timing/model': 0.969500201754272, 'learning_rate/qwen_vl_interface': 9.826762695015773e-06, 'learning_rate/action_model': 9.826762695015772e-05, 'epoch': 0.32}) +[2026-06-17 17:50:46,749][starVLA.training.train_starvla][INFO] - Step 696, Loss: {'train/loss': 1.0562018156051636, 'train/grad_norm_pre_clip': 12.71316909790039, 'timing/data': 0.035045746713876724, 'timing/model': 0.9277145219966769, 'learning_rate/qwen_vl_interface': 9.82618347604781e-06, 'learning_rate/action_model': 9.82618347604781e-05, 'epoch': 0.32}) +[2026-06-17 17:50:50,546][starVLA.training.train_starvla][INFO] - Step 697, Loss: {'train/loss': 0.6293742060661316, 'train/grad_norm_pre_clip': 8.225744247436523, 'timing/data': 0.04532413091510534, 'timing/model': 0.9278229493647814, 'learning_rate/qwen_vl_interface': 9.825603308432089e-06, 'learning_rate/action_model': 9.825603308432088e-05, 'epoch': 0.32}) +[2026-06-17 17:50:54,386][starVLA.training.train_starvla][INFO] - Step 698, Loss: {'train/loss': 0.3589782416820526, 'train/grad_norm_pre_clip': 3.730485439300537, 'timing/data': 0.031138554215431213, 'timing/model': 1.131465562619269, 'learning_rate/qwen_vl_interface': 9.825022192288878e-06, 'learning_rate/action_model': 9.825022192288877e-05, 'epoch': 0.32}) +[2026-06-17 17:50:58,149][starVLA.training.train_starvla][INFO] - Step 699, Loss: {'train/loss': 0.5699548721313477, 'train/grad_norm_pre_clip': 5.156219005584717, 'timing/data': 0.031391989439725876, 'timing/model': 1.0147388353943825, 'learning_rate/qwen_vl_interface': 9.824440127738643e-06, 'learning_rate/action_model': 9.824440127738642e-05, 'epoch': 0.32}) +[2026-06-17 17:53:03,404][starVLA.training.train_starvla][INFO] - Step 700, Loss: {'train/loss': 0.6186475157737732, 'train/grad_norm_pre_clip': 6.229691505432129, 'eval/loss': 0.6661040782928467, 'eval/latency_8/loss': 0.6661040782928467, 'eval/demon_fire_precision': 0.90304396843292, 'eval/demon_fire_recall': 0.9953401677539608, 'eval/demon_fire_f1': 0.9469484261859021, 'eval/demon_fire_rate_model': 0.9872008903728436, 'eval/demon_fire_rate_teacher': 0.8956594323873122, 'eval/demon_left_precision': 0.6987577639751553, 'eval/demon_left_recall': 0.4817987152034261, 'eval/demon_left_f1': 0.5703422053231939, 'eval/demon_left_rate_model': 0.0895937673900946, 'eval/demon_left_rate_teacher': 0.12993878686700056, 'eval/demon_right_precision': 0.7806004618937644, 'eval/demon_right_recall': 0.5121212121212121, 'eval/demon_right_f1': 0.6184812442817932, 'eval/demon_right_rate_model': 0.1204785754034502, 'eval/demon_right_rate_teacher': 0.18363939899833054, 'eval/demon_move_f1': 0.5944117248024936, 'eval/demon_cc_f1': 0.803911976282974, 'eval/latency_8/demon_fire_precision': 0.90304396843292, 'eval/latency_8/demon_fire_recall': 0.9953401677539608, 'eval/latency_8/demon_fire_f1': 0.9469484261859021, 'eval/latency_8/demon_fire_rate_model': 0.9872008903728436, 'eval/latency_8/demon_fire_rate_teacher': 0.8956594323873122, 'eval/latency_8/demon_left_precision': 0.6987577639751553, 'eval/latency_8/demon_left_recall': 0.4817987152034261, 'eval/latency_8/demon_left_f1': 0.5703422053231939, 'eval/latency_8/demon_left_rate_model': 0.0895937673900946, 'eval/latency_8/demon_left_rate_teacher': 0.12993878686700056, 'eval/latency_8/demon_right_precision': 0.7806004618937644, 'eval/latency_8/demon_right_recall': 0.5121212121212121, 'eval/latency_8/demon_right_f1': 0.6184812442817932, 'eval/latency_8/demon_right_rate_model': 0.1204785754034502, 'eval/latency_8/demon_right_rate_teacher': 0.18363939899833054, 'eval/latency_8/demon_move_f1': 0.5944117248024936, 'eval/latency_8/demon_cc_f1': 0.803911976282974, 'timing/data': 0.043181718327105045, 'timing/model': 0.9600902907550335, 'learning_rate/qwen_vl_interface': 9.823857114902047e-06, 'learning_rate/action_model': 9.823857114902047e-05, 'epoch': 0.32}) +[2026-06-17 17:53:09,141][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_700_state +[2026-06-17 17:53:09,162][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 17:53:15,998][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_700_state/model.safetensors +[2026-06-17 17:53:31,450][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_700_state/optimizer.bin +[2026-06-17 17:53:31,452][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_700_state/sampler.bin +[2026-06-17 17:53:31,452][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_700_state/sampler_1.bin +[2026-06-17 17:53:31,454][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_700_state/random_states_0.pkl +[2026-06-17 17:53:31,459][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 700 +[2026-06-17 17:53:31,459][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 17:53:31,470][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 17:53:35,488][starVLA.training.train_starvla][INFO] - Step 701, Loss: {'train/loss': 0.4356882572174072, 'train/grad_norm_pre_clip': 3.934309720993042, 'timing/data': 0.047573065385222435, 'timing/model': 1.1492906650528312, 'learning_rate/qwen_vl_interface': 9.823273153899951e-06, 'learning_rate/action_model': 9.82327315389995e-05, 'epoch': 0.32}) +[2026-06-17 17:53:39,109][starVLA.training.train_starvla][INFO] - Step 702, Loss: {'train/loss': 0.31860315799713135, 'train/grad_norm_pre_clip': 6.415828704833984, 'timing/data': 0.04981561005115509, 'timing/model': 0.9124468835070729, 'learning_rate/qwen_vl_interface': 9.822688244853407e-06, 'learning_rate/action_model': 9.822688244853408e-05, 'epoch': 0.32}) +[2026-06-17 17:53:42,876][starVLA.training.train_starvla][INFO] - Step 703, Loss: {'train/loss': 0.9691923260688782, 'train/grad_norm_pre_clip': 9.27061939239502, 'timing/data': 0.0008788378909230232, 'timing/model': 0.9268390974029899, 'learning_rate/qwen_vl_interface': 9.822102387883672e-06, 'learning_rate/action_model': 9.822102387883671e-05, 'epoch': 0.32}) +[2026-06-17 17:53:46,610][starVLA.training.train_starvla][INFO] - Step 704, Loss: {'train/loss': 0.48349782824516296, 'train/grad_norm_pre_clip': 6.426707744598389, 'timing/data': 0.0005121221765875816, 'timing/model': 0.9158940277993679, 'learning_rate/qwen_vl_interface': 9.82151558311219e-06, 'learning_rate/action_model': 9.821515583112191e-05, 'epoch': 0.32}) +[2026-06-17 17:53:50,331][starVLA.training.train_starvla][INFO] - Step 705, Loss: {'train/loss': 0.5769080519676208, 'train/grad_norm_pre_clip': 5.232637882232666, 'timing/data': 0.000482817180454731, 'timing/model': 0.9139158511534333, 'learning_rate/qwen_vl_interface': 9.82092783066061e-06, 'learning_rate/action_model': 9.82092783066061e-05, 'epoch': 0.32}) +[2026-06-17 17:53:54,121][starVLA.training.train_starvla][INFO] - Step 706, Loss: {'train/loss': 0.5580736398696899, 'train/grad_norm_pre_clip': 6.253907680511475, 'timing/data': 0.00048438645899295807, 'timing/model': 0.9156147297471762, 'learning_rate/qwen_vl_interface': 9.820339130650771e-06, 'learning_rate/action_model': 9.820339130650771e-05, 'epoch': 0.32}) +[2026-06-17 17:53:58,078][starVLA.training.train_starvla][INFO] - Step 707, Loss: {'train/loss': 0.898709774017334, 'train/grad_norm_pre_clip': 8.70368766784668, 'timing/data': 0.0016809171065688133, 'timing/model': 1.080370657145977, 'learning_rate/qwen_vl_interface': 9.819749483204715e-06, 'learning_rate/action_model': 9.819749483204715e-05, 'epoch': 0.32}) +[2026-06-17 17:54:01,979][starVLA.training.train_starvla][INFO] - Step 708, Loss: {'train/loss': 0.9988599419593811, 'train/grad_norm_pre_clip': 10.274560928344727, 'timing/data': 0.0007959539070725441, 'timing/model': 1.1588945286348462, 'learning_rate/qwen_vl_interface': 9.81915888844467e-06, 'learning_rate/action_model': 9.81915888844467e-05, 'epoch': 0.32}) +[2026-06-17 17:54:05,704][starVLA.training.train_starvla][INFO] - Step 709, Loss: {'train/loss': 0.5874448418617249, 'train/grad_norm_pre_clip': 7.522789001464844, 'timing/data': 0.0006299624219536781, 'timing/model': 0.9195712180808187, 'learning_rate/qwen_vl_interface': 9.818567346493072e-06, 'learning_rate/action_model': 9.818567346493071e-05, 'epoch': 0.32}) +[2026-06-17 17:54:09,954][starVLA.training.train_starvla][INFO] - Step 710, Loss: {'train/loss': 0.7720767855644226, 'train/grad_norm_pre_clip': 14.127690315246582, 'timing/data': 0.0005631893873214722, 'timing/model': 0.9940557321533561, 'learning_rate/qwen_vl_interface': 9.817974857472549e-06, 'learning_rate/action_model': 9.817974857472548e-05, 'epoch': 0.32}) +[2026-06-17 17:54:13,709][starVLA.training.train_starvla][INFO] - Step 711, Loss: {'train/loss': 0.622443675994873, 'train/grad_norm_pre_clip': 7.6642045974731445, 'timing/data': 0.0004935981705784798, 'timing/model': 0.9213348971679807, 'learning_rate/qwen_vl_interface': 9.817381421505922e-06, 'learning_rate/action_model': 9.817381421505921e-05, 'epoch': 0.32}) +[2026-06-17 17:54:17,656][starVLA.training.train_starvla][INFO] - Step 712, Loss: {'train/loss': 0.7891333103179932, 'train/grad_norm_pre_clip': 7.564293384552002, 'timing/data': 0.00048073288053274155, 'timing/model': 0.9198091477155685, 'learning_rate/qwen_vl_interface': 9.81678703871621e-06, 'learning_rate/action_model': 9.81678703871621e-05, 'epoch': 0.32}) +[2026-06-17 17:54:21,985][starVLA.training.train_starvla][INFO] - Step 713, Loss: {'train/loss': 0.6373291015625, 'train/grad_norm_pre_clip': 7.679399490356445, 'timing/data': 0.0005159089341759682, 'timing/model': 1.3045684779062867, 'learning_rate/qwen_vl_interface': 9.816191709226631e-06, 'learning_rate/action_model': 9.816191709226632e-05, 'epoch': 0.32}) +[2026-06-17 17:54:25,960][starVLA.training.train_starvla][INFO] - Step 714, Loss: {'train/loss': 0.9278069138526917, 'train/grad_norm_pre_clip': 12.472145080566406, 'timing/data': 0.0005251290276646614, 'timing/model': 1.079163913615048, 'learning_rate/qwen_vl_interface': 9.815595433160599e-06, 'learning_rate/action_model': 9.815595433160598e-05, 'epoch': 0.32}) +[2026-06-17 17:54:29,636][starVLA.training.train_starvla][INFO] - Step 715, Loss: {'train/loss': 0.8849992156028748, 'train/grad_norm_pre_clip': 9.507534980773926, 'timing/data': 0.0005508437752723694, 'timing/model': 1.0348113235086203, 'learning_rate/qwen_vl_interface': 9.81499821064172e-06, 'learning_rate/action_model': 9.81499821064172e-05, 'epoch': 0.32}) +[2026-06-17 17:54:33,401][starVLA.training.train_starvla][INFO] - Step 716, Loss: {'train/loss': 0.673940122127533, 'train/grad_norm_pre_clip': 8.352227210998535, 'timing/data': 0.0004832344129681587, 'timing/model': 0.9189108097925782, 'learning_rate/qwen_vl_interface': 9.814400041793799e-06, 'learning_rate/action_model': 9.814400041793798e-05, 'epoch': 0.32}) +[2026-06-17 17:54:37,728][starVLA.training.train_starvla][INFO] - Step 717, Loss: {'train/loss': 0.6355995535850525, 'train/grad_norm_pre_clip': 8.846630096435547, 'timing/data': 0.0008604815229773521, 'timing/model': 0.9219831181690097, 'learning_rate/qwen_vl_interface': 9.813800926740839e-06, 'learning_rate/action_model': 9.813800926740838e-05, 'epoch': 0.33}) +[2026-06-17 17:54:41,680][starVLA.training.train_starvla][INFO] - Step 718, Loss: {'train/loss': 0.9888008832931519, 'train/grad_norm_pre_clip': 8.165975570678711, 'timing/data': 0.0004825349897146225, 'timing/model': 0.9419459067285061, 'learning_rate/qwen_vl_interface': 9.813200865607034e-06, 'learning_rate/action_model': 9.813200865607034e-05, 'epoch': 0.33}) +[2026-06-17 17:54:45,544][starVLA.training.train_starvla][INFO] - Step 719, Loss: {'train/loss': 0.5204682350158691, 'train/grad_norm_pre_clip': 6.717031002044678, 'timing/data': 0.0008931681513786316, 'timing/model': 1.0077281529083848, 'learning_rate/qwen_vl_interface': 9.812599858516782e-06, 'learning_rate/action_model': 9.812599858516782e-05, 'epoch': 0.33}) +[2026-06-17 17:54:49,381][starVLA.training.train_starvla][INFO] - Step 720, Loss: {'train/loss': 0.637413740158081, 'train/grad_norm_pre_clip': 9.444459915161133, 'timing/data': 0.0011339737102389336, 'timing/model': 0.9594129603356123, 'learning_rate/qwen_vl_interface': 9.81199790559467e-06, 'learning_rate/action_model': 9.811997905594668e-05, 'epoch': 0.33}) +[2026-06-17 17:54:53,338][starVLA.training.train_starvla][INFO] - Step 721, Loss: {'train/loss': 0.7864510416984558, 'train/grad_norm_pre_clip': 7.5140581130981445, 'timing/data': 0.000522601418197155, 'timing/model': 1.0864586373791099, 'learning_rate/qwen_vl_interface': 9.811395006965481e-06, 'learning_rate/action_model': 9.81139500696548e-05, 'epoch': 0.33}) +[2026-06-17 17:54:57,140][starVLA.training.train_starvla][INFO] - Step 722, Loss: {'train/loss': 0.43677186965942383, 'train/grad_norm_pre_clip': 4.709321022033691, 'timing/data': 0.00046135764569044113, 'timing/model': 0.9178544441238046, 'learning_rate/qwen_vl_interface': 9.810791162754199e-06, 'learning_rate/action_model': 9.810791162754198e-05, 'epoch': 0.33}) +[2026-06-17 17:55:00,901][starVLA.training.train_starvla][INFO] - Step 723, Loss: {'train/loss': 0.6814545392990112, 'train/grad_norm_pre_clip': 5.864939212799072, 'timing/data': 0.0005276510491967201, 'timing/model': 0.9166800566017628, 'learning_rate/qwen_vl_interface': 9.810186373086002e-06, 'learning_rate/action_model': 9.810186373086002e-05, 'epoch': 0.33}) +[2026-06-17 17:55:05,001][starVLA.training.train_starvla][INFO] - Step 724, Loss: {'train/loss': 0.8048694729804993, 'train/grad_norm_pre_clip': 8.990056037902832, 'timing/data': 0.0009629139676690102, 'timing/model': 1.2553031025454402, 'learning_rate/qwen_vl_interface': 9.809580638086264e-06, 'learning_rate/action_model': 9.809580638086263e-05, 'epoch': 0.33}) +[2026-06-17 17:55:08,773][starVLA.training.train_starvla][INFO] - Step 725, Loss: {'train/loss': 0.8842331767082214, 'train/grad_norm_pre_clip': 9.05534839630127, 'timing/data': 0.0006249053403735161, 'timing/model': 0.9520353581756353, 'learning_rate/qwen_vl_interface': 9.80897395788055e-06, 'learning_rate/action_model': 9.808973957880551e-05, 'epoch': 0.33}) +[2026-06-17 17:55:12,533][starVLA.training.train_starvla][INFO] - Step 726, Loss: {'train/loss': 0.7113960385322571, 'train/grad_norm_pre_clip': 6.971153736114502, 'timing/data': 0.0006042281165719032, 'timing/model': 0.9287983765825629, 'learning_rate/qwen_vl_interface': 9.808366332594632e-06, 'learning_rate/action_model': 9.808366332594632e-05, 'epoch': 0.33}) +[2026-06-17 17:55:16,441][starVLA.training.train_starvla][INFO] - Step 727, Loss: {'train/loss': 0.5901730060577393, 'train/grad_norm_pre_clip': 7.397254943847656, 'timing/data': 0.0009563695639371872, 'timing/model': 1.1020236546173692, 'learning_rate/qwen_vl_interface': 9.807757762354467e-06, 'learning_rate/action_model': 9.807757762354467e-05, 'epoch': 0.33}) +[2026-06-17 17:55:20,349][starVLA.training.train_starvla][INFO] - Step 728, Loss: {'train/loss': 0.934159517288208, 'train/grad_norm_pre_clip': 8.792916297912598, 'timing/data': 0.0004816371947526932, 'timing/model': 0.9175572516396642, 'learning_rate/qwen_vl_interface': 9.807148247286215e-06, 'learning_rate/action_model': 9.807148247286215e-05, 'epoch': 0.33}) +[2026-06-17 17:55:24,300][starVLA.training.train_starvla][INFO] - Step 729, Loss: {'train/loss': 0.7000395059585571, 'train/grad_norm_pre_clip': 8.03683090209961, 'timing/data': 0.0005273446440696716, 'timing/model': 0.9574595345184207, 'learning_rate/qwen_vl_interface': 9.806537787516226e-06, 'learning_rate/action_model': 9.806537787516226e-05, 'epoch': 0.33}) +[2026-06-17 17:55:28,087][starVLA.training.train_starvla][INFO] - Step 730, Loss: {'train/loss': 0.6084926724433899, 'train/grad_norm_pre_clip': 6.660386562347412, 'timing/data': 0.0005151471123099327, 'timing/model': 0.9217397319152951, 'learning_rate/qwen_vl_interface': 9.80592638317105e-06, 'learning_rate/action_model': 9.80592638317105e-05, 'epoch': 0.33}) +[2026-06-17 17:55:31,845][starVLA.training.train_starvla][INFO] - Step 731, Loss: {'train/loss': 0.4844559133052826, 'train/grad_norm_pre_clip': 8.124923706054688, 'timing/data': 0.0008503943681716919, 'timing/model': 0.9520822977647185, 'learning_rate/qwen_vl_interface': 9.805314034377434e-06, 'learning_rate/action_model': 9.805314034377434e-05, 'epoch': 0.33}) +[2026-06-17 17:55:35,867][starVLA.training.train_starvla][INFO] - Step 732, Loss: {'train/loss': 0.5271980166435242, 'train/grad_norm_pre_clip': 7.16727876663208, 'timing/data': 0.0010198978707194328, 'timing/model': 1.0376702127978206, 'learning_rate/qwen_vl_interface': 9.804700741262318e-06, 'learning_rate/action_model': 9.804700741262317e-05, 'epoch': 0.33}) +[2026-06-17 17:55:39,723][starVLA.training.train_starvla][INFO] - Step 733, Loss: {'train/loss': 0.9801598191261292, 'train/grad_norm_pre_clip': 8.62700366973877, 'timing/data': 0.0005345391109585762, 'timing/model': 1.078796224668622, 'learning_rate/qwen_vl_interface': 9.804086503952835e-06, 'learning_rate/action_model': 9.804086503952835e-05, 'epoch': 0.33}) +[2026-06-17 17:55:43,461][starVLA.training.train_starvla][INFO] - Step 734, Loss: {'train/loss': 0.6442859768867493, 'train/grad_norm_pre_clip': 6.736212253570557, 'timing/data': 0.00047570280730724335, 'timing/model': 1.0851323856040835, 'learning_rate/qwen_vl_interface': 9.803471322576319e-06, 'learning_rate/action_model': 9.803471322576319e-05, 'epoch': 0.33}) +[2026-06-17 17:55:47,449][starVLA.training.train_starvla][INFO] - Step 735, Loss: {'train/loss': 0.619758129119873, 'train/grad_norm_pre_clip': 14.489802360534668, 'timing/data': 0.0005382271483540535, 'timing/model': 0.9152394020929933, 'learning_rate/qwen_vl_interface': 9.8028551972603e-06, 'learning_rate/action_model': 9.802855197260299e-05, 'epoch': 0.33}) +[2026-06-17 17:55:51,402][starVLA.training.train_starvla][INFO] - Step 736, Loss: {'train/loss': 0.7964027523994446, 'train/grad_norm_pre_clip': 8.843795776367188, 'timing/data': 0.0005554221570491791, 'timing/model': 0.9209053590893745, 'learning_rate/qwen_vl_interface': 9.802238128132498e-06, 'learning_rate/action_model': 9.802238128132498e-05, 'epoch': 0.33}) +[2026-06-17 17:55:55,394][starVLA.training.train_starvla][INFO] - Step 737, Loss: {'train/loss': 0.6052858829498291, 'train/grad_norm_pre_clip': 5.614561080932617, 'timing/data': 0.0008225739002227783, 'timing/model': 1.0452658152207732, 'learning_rate/qwen_vl_interface': 9.801620115320836e-06, 'learning_rate/action_model': 9.801620115320835e-05, 'epoch': 0.33}) +[2026-06-17 17:55:59,280][starVLA.training.train_starvla][INFO] - Step 738, Loss: {'train/loss': 0.7386295795440674, 'train/grad_norm_pre_clip': 9.4053316116333, 'timing/data': 0.0004620524123311043, 'timing/model': 0.9317263858392835, 'learning_rate/qwen_vl_interface': 9.801001158953425e-06, 'learning_rate/action_model': 9.801001158953423e-05, 'epoch': 0.33}) +[2026-06-17 17:56:03,148][starVLA.training.train_starvla][INFO] - Step 739, Loss: {'train/loss': 0.5959746241569519, 'train/grad_norm_pre_clip': 7.055870532989502, 'timing/data': 0.0008214805275201797, 'timing/model': 0.9214708730578423, 'learning_rate/qwen_vl_interface': 9.800381259158575e-06, 'learning_rate/action_model': 9.800381259158574e-05, 'epoch': 0.34}) +[2026-06-17 17:56:08,175][starVLA.training.train_starvla][INFO] - Step 740, Loss: {'train/loss': 0.5607131719589233, 'train/grad_norm_pre_clip': 7.172462463378906, 'timing/data': 0.0010045496746897697, 'timing/model': 1.3943267222493887, 'learning_rate/qwen_vl_interface': 9.799760416064794e-06, 'learning_rate/action_model': 9.799760416064793e-05, 'epoch': 0.34}) +[2026-06-17 17:56:12,264][starVLA.training.train_starvla][INFO] - Step 741, Loss: {'train/loss': 0.7947391271591187, 'train/grad_norm_pre_clip': 8.402301788330078, 'timing/data': 0.0004945863038301468, 'timing/model': 1.1758829122409225, 'learning_rate/qwen_vl_interface': 9.799138629800785e-06, 'learning_rate/action_model': 9.799138629800785e-05, 'epoch': 0.34}) +[2026-06-17 17:56:16,148][starVLA.training.train_starvla][INFO] - Step 742, Loss: {'train/loss': 0.65827876329422, 'train/grad_norm_pre_clip': 8.447066307067871, 'timing/data': 0.0006911130622029305, 'timing/model': 0.9192547779530287, 'learning_rate/qwen_vl_interface': 9.798515900495441e-06, 'learning_rate/action_model': 9.79851590049544e-05, 'epoch': 0.34}) +[2026-06-17 17:56:19,893][starVLA.training.train_starvla][INFO] - Step 743, Loss: {'train/loss': 0.4816821813583374, 'train/grad_norm_pre_clip': 8.233684539794922, 'timing/data': 0.0004979213699698448, 'timing/model': 0.9267998849973083, 'learning_rate/qwen_vl_interface': 9.797892228277857e-06, 'learning_rate/action_model': 9.797892228277857e-05, 'epoch': 0.34}) +[2026-06-17 17:56:23,689][starVLA.training.train_starvla][INFO] - Step 744, Loss: {'train/loss': 0.7062957882881165, 'train/grad_norm_pre_clip': 8.599984169006348, 'timing/data': 0.0007055206224322319, 'timing/model': 0.9197400398552418, 'learning_rate/qwen_vl_interface': 9.797267613277319e-06, 'learning_rate/action_model': 9.797267613277319e-05, 'epoch': 0.34}) +[2026-06-17 17:56:27,431][starVLA.training.train_starvla][INFO] - Step 745, Loss: {'train/loss': 0.6109890937805176, 'train/grad_norm_pre_clip': 8.126129150390625, 'timing/data': 0.0009225569665431976, 'timing/model': 0.9286580653861165, 'learning_rate/qwen_vl_interface': 9.796642055623314e-06, 'learning_rate/action_model': 9.796642055623314e-05, 'epoch': 0.34}) +[2026-06-17 17:56:31,232][starVLA.training.train_starvla][INFO] - Step 746, Loss: {'train/loss': 0.5950889587402344, 'train/grad_norm_pre_clip': 5.453620433807373, 'timing/data': 0.0004157908260822296, 'timing/model': 0.9236929025501013, 'learning_rate/qwen_vl_interface': 9.796015555445518e-06, 'learning_rate/action_model': 9.796015555445517e-05, 'epoch': 0.34}) +[2026-06-17 17:56:35,313][starVLA.training.train_starvla][INFO] - Step 747, Loss: {'train/loss': 0.6464309692382812, 'train/grad_norm_pre_clip': 6.935819149017334, 'timing/data': 0.0005163801833987236, 'timing/model': 1.0855476362630725, 'learning_rate/qwen_vl_interface': 9.795388112873806e-06, 'learning_rate/action_model': 9.795388112873806e-05, 'epoch': 0.34}) +[2026-06-17 17:56:38,958][starVLA.training.train_starvla][INFO] - Step 748, Loss: {'train/loss': 0.6312641501426697, 'train/grad_norm_pre_clip': 8.46097183227539, 'timing/data': 0.0008683754131197929, 'timing/model': 0.9195624310523272, 'learning_rate/qwen_vl_interface': 9.794759728038246e-06, 'learning_rate/action_model': 9.794759728038246e-05, 'epoch': 0.34}) +[2026-06-17 17:56:42,709][starVLA.training.train_starvla][INFO] - Step 749, Loss: {'train/loss': 0.8055840134620667, 'train/grad_norm_pre_clip': 9.49205493927002, 'timing/data': 0.0009874263778328896, 'timing/model': 0.9195816786959767, 'learning_rate/qwen_vl_interface': 9.794130401069105e-06, 'learning_rate/action_model': 9.794130401069104e-05, 'epoch': 0.34}) +[2026-06-17 17:56:46,389][starVLA.training.train_starvla][INFO] - Step 750, Loss: {'train/loss': 0.5146870017051697, 'train/grad_norm_pre_clip': 9.374643325805664, 'timing/data': 0.001323966309428215, 'timing/model': 0.9585726112127304, 'learning_rate/qwen_vl_interface': 9.79350013209684e-06, 'learning_rate/action_model': 9.79350013209684e-05, 'epoch': 0.34}) +[2026-06-17 17:56:50,380][starVLA.training.train_starvla][INFO] - Step 751, Loss: {'train/loss': 0.7068083882331848, 'train/grad_norm_pre_clip': 7.81228494644165, 'timing/data': 0.0008232872933149338, 'timing/model': 1.051947308704257, 'learning_rate/qwen_vl_interface': 9.792868921252112e-06, 'learning_rate/action_model': 9.792868921252111e-05, 'epoch': 0.34}) +[2026-06-17 17:56:54,517][starVLA.training.train_starvla][INFO] - Step 752, Loss: {'train/loss': 0.6607747077941895, 'train/grad_norm_pre_clip': 7.088450908660889, 'timing/data': 0.0004987418651580811, 'timing/model': 0.9309023162350059, 'learning_rate/qwen_vl_interface': 9.792236768665765e-06, 'learning_rate/action_model': 9.792236768665764e-05, 'epoch': 0.34}) +[2026-06-17 17:56:58,284][starVLA.training.train_starvla][INFO] - Step 753, Loss: {'train/loss': 0.6515534520149231, 'train/grad_norm_pre_clip': 6.582904815673828, 'timing/data': 0.0012603290379047394, 'timing/model': 1.0851045306771994, 'learning_rate/qwen_vl_interface': 9.79160367446885e-06, 'learning_rate/action_model': 9.79160367446885e-05, 'epoch': 0.34}) +[2026-06-17 17:57:01,941][starVLA.training.train_starvla][INFO] - Step 754, Loss: {'train/loss': 0.6863051652908325, 'train/grad_norm_pre_clip': 8.265974998474121, 'timing/data': 0.0009943554177880287, 'timing/model': 0.9930893331766129, 'learning_rate/qwen_vl_interface': 9.790969638792606e-06, 'learning_rate/action_model': 9.790969638792606e-05, 'epoch': 0.34}) +[2026-06-17 17:57:05,706][starVLA.training.train_starvla][INFO] - Step 755, Loss: {'train/loss': 0.8110078573226929, 'train/grad_norm_pre_clip': 9.031436920166016, 'timing/data': 0.0008057160302996635, 'timing/model': 0.922291417606175, 'learning_rate/qwen_vl_interface': 9.790334661768471e-06, 'learning_rate/action_model': 9.79033466176847e-05, 'epoch': 0.34}) +[2026-06-17 17:57:10,235][starVLA.training.train_starvla][INFO] - Step 756, Loss: {'train/loss': 0.5871059894561768, 'train/grad_norm_pre_clip': 8.75625228881836, 'timing/data': 0.0004974072799086571, 'timing/model': 1.1039118887856603, 'learning_rate/qwen_vl_interface': 9.78969874352807e-06, 'learning_rate/action_model': 9.78969874352807e-05, 'epoch': 0.34}) +[2026-06-17 17:57:14,028][starVLA.training.train_starvla][INFO] - Step 757, Loss: {'train/loss': 0.5767825841903687, 'train/grad_norm_pre_clip': 6.394834995269775, 'timing/data': 0.0014693103730678558, 'timing/model': 0.9184696208685637, 'learning_rate/qwen_vl_interface': 9.78906188420324e-06, 'learning_rate/action_model': 9.78906188420324e-05, 'epoch': 0.34}) +[2026-06-17 17:57:17,774][starVLA.training.train_starvla][INFO] - Step 758, Loss: {'train/loss': 0.8908157348632812, 'train/grad_norm_pre_clip': 9.796215057373047, 'timing/data': 0.0005373451858758926, 'timing/model': 0.9194480665028095, 'learning_rate/qwen_vl_interface': 9.788424083925996e-06, 'learning_rate/action_model': 9.788424083925995e-05, 'epoch': 0.34}) +[2026-06-17 17:57:21,605][starVLA.training.train_starvla][INFO] - Step 759, Loss: {'train/loss': 0.9100385904312134, 'train/grad_norm_pre_clip': 8.82223129272461, 'timing/data': 0.0004826243966817856, 'timing/model': 1.0911718858405948, 'learning_rate/qwen_vl_interface': 9.787785342828554e-06, 'learning_rate/action_model': 9.787785342828553e-05, 'epoch': 0.34}) +[2026-06-17 17:57:25,473][starVLA.training.train_starvla][INFO] - Step 760, Loss: {'train/loss': 0.6492778658866882, 'train/grad_norm_pre_clip': 6.943143367767334, 'timing/data': 0.0009199418127536774, 'timing/model': 1.2055303314700723, 'learning_rate/qwen_vl_interface': 9.78714566104333e-06, 'learning_rate/action_model': 9.787145661043329e-05, 'epoch': 0.34}) +[2026-06-17 17:57:29,379][starVLA.training.train_starvla][INFO] - Step 761, Loss: {'train/loss': 0.6631228923797607, 'train/grad_norm_pre_clip': 7.617739200592041, 'timing/data': 0.0005025425925850868, 'timing/model': 1.051938796415925, 'learning_rate/qwen_vl_interface': 9.786505038702926e-06, 'learning_rate/action_model': 9.786505038702925e-05, 'epoch': 0.35}) +[2026-06-17 17:57:33,242][starVLA.training.train_starvla][INFO] - Step 762, Loss: {'train/loss': 0.9172842502593994, 'train/grad_norm_pre_clip': 8.090072631835938, 'timing/data': 0.0009445194154977798, 'timing/model': 0.9169520540162921, 'learning_rate/qwen_vl_interface': 9.785863475940147e-06, 'learning_rate/action_model': 9.785863475940145e-05, 'epoch': 0.35}) +[2026-06-17 17:57:37,035][starVLA.training.train_starvla][INFO] - Step 763, Loss: {'train/loss': 0.8715047836303711, 'train/grad_norm_pre_clip': 9.071260452270508, 'timing/data': 0.0004623010754585266, 'timing/model': 0.9155370881780982, 'learning_rate/qwen_vl_interface': 9.785220972887988e-06, 'learning_rate/action_model': 9.785220972887989e-05, 'epoch': 0.35}) +[2026-06-17 17:57:40,773][starVLA.training.train_starvla][INFO] - Step 764, Loss: {'train/loss': 0.4941635727882385, 'train/grad_norm_pre_clip': 7.555044651031494, 'timing/data': 0.0005836933851242065, 'timing/model': 0.923110930249095, 'learning_rate/qwen_vl_interface': 9.784577529679643e-06, 'learning_rate/action_model': 9.784577529679643e-05, 'epoch': 0.35}) +[2026-06-17 17:57:44,540][starVLA.training.train_starvla][INFO] - Step 765, Loss: {'train/loss': 0.5854892730712891, 'train/grad_norm_pre_clip': 10.584430694580078, 'timing/data': 0.0004733670502901077, 'timing/model': 0.9241355648264289, 'learning_rate/qwen_vl_interface': 9.783933146448495e-06, 'learning_rate/action_model': 9.783933146448495e-05, 'epoch': 0.35}) +[2026-06-17 17:57:48,657][starVLA.training.train_starvla][INFO] - Step 766, Loss: {'train/loss': 1.0437796115875244, 'train/grad_norm_pre_clip': 7.245242595672607, 'timing/data': 0.0005539814010262489, 'timing/model': 1.4889641217887402, 'learning_rate/qwen_vl_interface': 9.783287823328127e-06, 'learning_rate/action_model': 9.783287823328127e-05, 'epoch': 0.35}) +[2026-06-17 17:57:52,513][starVLA.training.train_starvla][INFO] - Step 767, Loss: {'train/loss': 0.8964293599128723, 'train/grad_norm_pre_clip': 14.16864013671875, 'timing/data': 0.0005544340237975121, 'timing/model': 0.9207356693223119, 'learning_rate/qwen_vl_interface': 9.782641560452317e-06, 'learning_rate/action_model': 9.782641560452316e-05, 'epoch': 0.35}) +[2026-06-17 17:57:56,271][starVLA.training.train_starvla][INFO] - Step 768, Loss: {'train/loss': 0.8151401281356812, 'train/grad_norm_pre_clip': 8.848067283630371, 'timing/data': 0.0005120644345879555, 'timing/model': 0.9266916392371058, 'learning_rate/qwen_vl_interface': 9.781994357955032e-06, 'learning_rate/action_model': 9.781994357955032e-05, 'epoch': 0.35}) +[2026-06-17 17:58:00,064][starVLA.training.train_starvla][INFO] - Step 769, Loss: {'train/loss': 0.7334364652633667, 'train/grad_norm_pre_clip': 6.462275981903076, 'timing/data': 0.0004784446209669113, 'timing/model': 0.9186996854841709, 'learning_rate/qwen_vl_interface': 9.78134621597044e-06, 'learning_rate/action_model': 9.781346215970441e-05, 'epoch': 0.35}) +[2026-06-17 17:58:03,833][starVLA.training.train_starvla][INFO] - Step 770, Loss: {'train/loss': 0.6827256679534912, 'train/grad_norm_pre_clip': 10.957036018371582, 'timing/data': 0.0010048560798168182, 'timing/model': 0.9197844471782446, 'learning_rate/qwen_vl_interface': 9.780697134632903e-06, 'learning_rate/action_model': 9.780697134632902e-05, 'epoch': 0.35}) +[2026-06-17 17:58:07,730][starVLA.training.train_starvla][INFO] - Step 771, Loss: {'train/loss': 0.5827991366386414, 'train/grad_norm_pre_clip': 6.081429958343506, 'timing/data': 0.0005040932446718216, 'timing/model': 0.9231990948319435, 'learning_rate/qwen_vl_interface': 9.780047114076974e-06, 'learning_rate/action_model': 9.780047114076973e-05, 'epoch': 0.35}) +[2026-06-17 17:58:11,526][starVLA.training.train_starvla][INFO] - Step 772, Loss: {'train/loss': 0.6341924667358398, 'train/grad_norm_pre_clip': 7.051452159881592, 'timing/data': 0.0007258821278810501, 'timing/model': 0.9487144071608782, 'learning_rate/qwen_vl_interface': 9.779396154437404e-06, 'learning_rate/action_model': 9.779396154437403e-05, 'epoch': 0.35}) +[2026-06-17 17:58:15,425][starVLA.training.train_starvla][INFO] - Step 773, Loss: {'train/loss': 1.0252755880355835, 'train/grad_norm_pre_clip': 8.877922058105469, 'timing/data': 0.0007460964843630791, 'timing/model': 1.0928810788318515, 'learning_rate/qwen_vl_interface': 9.778744255849138e-06, 'learning_rate/action_model': 9.778744255849138e-05, 'epoch': 0.35}) +[2026-06-17 17:58:19,055][starVLA.training.train_starvla][INFO] - Step 774, Loss: {'train/loss': 0.8225877285003662, 'train/grad_norm_pre_clip': 18.429157257080078, 'timing/data': 0.0005516074597835541, 'timing/model': 0.9449715334922075, 'learning_rate/qwen_vl_interface': 9.778091418447313e-06, 'learning_rate/action_model': 9.778091418447312e-05, 'epoch': 0.35}) +[2026-06-17 17:58:22,768][starVLA.training.train_starvla][INFO] - Step 775, Loss: {'train/loss': 0.45352452993392944, 'train/grad_norm_pre_clip': 11.75423812866211, 'timing/data': 0.0012201517820358276, 'timing/model': 0.9164557559415698, 'learning_rate/qwen_vl_interface': 9.777437642367265e-06, 'learning_rate/action_model': 9.777437642367264e-05, 'epoch': 0.35}) +[2026-06-17 17:58:26,610][starVLA.training.train_starvla][INFO] - Step 776, Loss: {'train/loss': 0.704639732837677, 'train/grad_norm_pre_clip': 7.859114170074463, 'timing/data': 0.0005096960812807083, 'timing/model': 0.9158728029578924, 'learning_rate/qwen_vl_interface': 9.776782927744523e-06, 'learning_rate/action_model': 9.776782927744522e-05, 'epoch': 0.35}) +[2026-06-17 17:58:30,344][starVLA.training.train_starvla][INFO] - Step 777, Loss: {'train/loss': 0.4575115144252777, 'train/grad_norm_pre_clip': 6.483867645263672, 'timing/data': 0.0008201925083994865, 'timing/model': 0.913255057297647, 'learning_rate/qwen_vl_interface': 9.776127274714808e-06, 'learning_rate/action_model': 9.776127274714807e-05, 'epoch': 0.35}) +[2026-06-17 17:58:34,069][starVLA.training.train_starvla][INFO] - Step 778, Loss: {'train/loss': 0.48433077335357666, 'train/grad_norm_pre_clip': 8.100536346435547, 'timing/data': 0.0004808390513062477, 'timing/model': 0.9137467918917537, 'learning_rate/qwen_vl_interface': 9.775470683414038e-06, 'learning_rate/action_model': 9.775470683414038e-05, 'epoch': 0.35}) +[2026-06-17 17:58:38,493][starVLA.training.train_starvla][INFO] - Step 779, Loss: {'train/loss': 0.5721318125724792, 'train/grad_norm_pre_clip': 6.9990925788879395, 'timing/data': 0.0005025733262300491, 'timing/model': 1.041396927088499, 'learning_rate/qwen_vl_interface': 9.774813153978327e-06, 'learning_rate/action_model': 9.774813153978327e-05, 'epoch': 0.35}) +[2026-06-17 17:58:42,505][starVLA.training.train_starvla][INFO] - Step 780, Loss: {'train/loss': 0.3941633701324463, 'train/grad_norm_pre_clip': 4.629100322723389, 'timing/data': 0.0006777262315154076, 'timing/model': 1.169942794367671, 'learning_rate/qwen_vl_interface': 9.77415468654398e-06, 'learning_rate/action_model': 9.77415468654398e-05, 'epoch': 0.35}) +[2026-06-17 17:58:46,206][starVLA.training.train_starvla][INFO] - Step 781, Loss: {'train/loss': 0.6574785113334656, 'train/grad_norm_pre_clip': 7.034188270568848, 'timing/data': 0.00047768186777830124, 'timing/model': 0.9154828544706106, 'learning_rate/qwen_vl_interface': 9.773495281247497e-06, 'learning_rate/action_model': 9.773495281247497e-05, 'epoch': 0.35}) +[2026-06-17 17:58:50,122][starVLA.training.train_starvla][INFO] - Step 782, Loss: {'train/loss': 0.6801296472549438, 'train/grad_norm_pre_clip': 6.911197185516357, 'timing/data': 0.0005005709826946259, 'timing/model': 0.9276263117790222, 'learning_rate/qwen_vl_interface': 9.772834938225577e-06, 'learning_rate/action_model': 9.772834938225576e-05, 'epoch': 0.35}) +[2026-06-17 17:58:53,957][starVLA.training.train_starvla][INFO] - Step 783, Loss: {'train/loss': 0.7398033738136292, 'train/grad_norm_pre_clip': 14.254843711853027, 'timing/data': 0.0005249753594398499, 'timing/model': 0.9225643146783113, 'learning_rate/qwen_vl_interface': 9.772173657615107e-06, 'learning_rate/action_model': 9.772173657615106e-05, 'epoch': 0.36}) +[2026-06-17 17:58:57,777][starVLA.training.train_starvla][INFO] - Step 784, Loss: {'train/loss': 0.8581510186195374, 'train/grad_norm_pre_clip': 9.588611602783203, 'timing/data': 0.00048337969928979874, 'timing/model': 0.9256024798378348, 'learning_rate/qwen_vl_interface': 9.77151143955317e-06, 'learning_rate/action_model': 9.77151143955317e-05, 'epoch': 0.36}) +[2026-06-17 17:59:01,620][starVLA.training.train_starvla][INFO] - Step 785, Loss: {'train/loss': 0.7459021210670471, 'train/grad_norm_pre_clip': 8.227182388305664, 'timing/data': 0.00047156307846307755, 'timing/model': 0.9232060331851244, 'learning_rate/qwen_vl_interface': 9.770848284177048e-06, 'learning_rate/action_model': 9.770848284177047e-05, 'epoch': 0.36}) +[2026-06-17 17:59:05,419][starVLA.training.train_starvla][INFO] - Step 786, Loss: {'train/loss': 0.6648886799812317, 'train/grad_norm_pre_clip': 7.0366740226745605, 'timing/data': 0.0005011521279811859, 'timing/model': 1.1177676832303405, 'learning_rate/qwen_vl_interface': 9.77018419162421e-06, 'learning_rate/action_model': 9.77018419162421e-05, 'epoch': 0.36}) +[2026-06-17 17:59:09,047][starVLA.training.train_starvla][INFO] - Step 787, Loss: {'train/loss': 0.6806815266609192, 'train/grad_norm_pre_clip': 10.440805435180664, 'timing/data': 0.0005208337679505348, 'timing/model': 0.9195854282006621, 'learning_rate/qwen_vl_interface': 9.769519162032328e-06, 'learning_rate/action_model': 9.769519162032327e-05, 'epoch': 0.36}) +[2026-06-17 17:59:13,000][starVLA.training.train_starvla][INFO] - Step 788, Loss: {'train/loss': 0.569603681564331, 'train/grad_norm_pre_clip': 9.790355682373047, 'timing/data': 0.0005058739334344864, 'timing/model': 1.0115131763741374, 'learning_rate/qwen_vl_interface': 9.768853195539259e-06, 'learning_rate/action_model': 9.768853195539258e-05, 'epoch': 0.36}) +[2026-06-17 17:59:16,767][starVLA.training.train_starvla][INFO] - Step 789, Loss: {'train/loss': 0.8741990923881531, 'train/grad_norm_pre_clip': 11.33269214630127, 'timing/data': 0.000922677107155323, 'timing/model': 0.9322663461789489, 'learning_rate/qwen_vl_interface': 9.76818629228306e-06, 'learning_rate/action_model': 9.768186292283059e-05, 'epoch': 0.36}) +[2026-06-17 17:59:20,514][starVLA.training.train_starvla][INFO] - Step 790, Loss: {'train/loss': 0.47311651706695557, 'train/grad_norm_pre_clip': 6.895906925201416, 'timing/data': 0.0008427668362855911, 'timing/model': 0.916488160379231, 'learning_rate/qwen_vl_interface': 9.767518452401981e-06, 'learning_rate/action_model': 9.76751845240198e-05, 'epoch': 0.36}) +[2026-06-17 17:59:24,689][starVLA.training.train_starvla][INFO] - Step 791, Loss: {'train/loss': 0.48539403080940247, 'train/grad_norm_pre_clip': 6.106931686401367, 'timing/data': 0.0007981797680258751, 'timing/model': 0.9178976500406861, 'learning_rate/qwen_vl_interface': 9.766849676034465e-06, 'learning_rate/action_model': 9.766849676034464e-05, 'epoch': 0.36}) +[2026-06-17 17:59:28,502][starVLA.training.train_starvla][INFO] - Step 792, Loss: {'train/loss': 0.7891933917999268, 'train/grad_norm_pre_clip': 8.830467224121094, 'timing/data': 0.0008978378027677536, 'timing/model': 1.1254924600943923, 'learning_rate/qwen_vl_interface': 9.766179963319152e-06, 'learning_rate/action_model': 9.76617996331915e-05, 'epoch': 0.36}) +[2026-06-17 17:59:32,422][starVLA.training.train_starvla][INFO] - Step 793, Loss: {'train/loss': 0.4686957895755768, 'train/grad_norm_pre_clip': 9.530550003051758, 'timing/data': 0.000888570211827755, 'timing/model': 1.2075857864692807, 'learning_rate/qwen_vl_interface': 9.765509314394872e-06, 'learning_rate/action_model': 9.765509314394872e-05, 'epoch': 0.36}) +[2026-06-17 17:59:35,992][starVLA.training.train_starvla][INFO] - Step 794, Loss: {'train/loss': 0.3379831314086914, 'train/grad_norm_pre_clip': 6.290445804595947, 'timing/data': 0.000791679136455059, 'timing/model': 0.9160450045019388, 'learning_rate/qwen_vl_interface': 9.764837729400652e-06, 'learning_rate/action_model': 9.764837729400653e-05, 'epoch': 0.36}) +[2026-06-17 17:59:39,835][starVLA.training.train_starvla][INFO] - Step 795, Loss: {'train/loss': 0.6879630088806152, 'train/grad_norm_pre_clip': 7.0313262939453125, 'timing/data': 0.0008280379697680473, 'timing/model': 0.918851075693965, 'learning_rate/qwen_vl_interface': 9.764165208475714e-06, 'learning_rate/action_model': 9.764165208475714e-05, 'epoch': 0.36}) +[2026-06-17 17:59:43,848][starVLA.training.train_starvla][INFO] - Step 796, Loss: {'train/loss': 0.8832681179046631, 'train/grad_norm_pre_clip': 7.193333148956299, 'timing/data': 0.0004971073940396309, 'timing/model': 1.0666344799101353, 'learning_rate/qwen_vl_interface': 9.76349175175947e-06, 'learning_rate/action_model': 9.763491751759469e-05, 'epoch': 0.36}) +[2026-06-17 17:59:48,212][starVLA.training.train_starvla][INFO] - Step 797, Loss: {'train/loss': 0.8536110520362854, 'train/grad_norm_pre_clip': 9.81742000579834, 'timing/data': 0.0005070017650723457, 'timing/model': 1.4856982175260782, 'learning_rate/qwen_vl_interface': 9.76281735939153e-06, 'learning_rate/action_model': 9.762817359391529e-05, 'epoch': 0.36}) +[2026-06-17 17:59:52,122][starVLA.training.train_starvla][INFO] - Step 798, Loss: {'train/loss': 0.8278123736381531, 'train/grad_norm_pre_clip': 8.656675338745117, 'timing/data': 0.0011218199506402016, 'timing/model': 0.94157490786165, 'learning_rate/qwen_vl_interface': 9.762142031511695e-06, 'learning_rate/action_model': 9.762142031511695e-05, 'epoch': 0.36}) +[2026-06-17 17:59:55,959][starVLA.training.train_starvla][INFO] - Step 799, Loss: {'train/loss': 0.40186411142349243, 'train/grad_norm_pre_clip': 4.352676868438721, 'timing/data': 0.0004987819120287895, 'timing/model': 1.1265597194433212, 'learning_rate/qwen_vl_interface': 9.76146576825996e-06, 'learning_rate/action_model': 9.761465768259959e-05, 'epoch': 0.36}) +[2026-06-17 18:01:49,480][starVLA.training.train_starvla][INFO] - Step 800, Loss: {'train/loss': 0.8750394582748413, 'train/grad_norm_pre_clip': 18.158828735351562, 'eval/loss': 0.6459647417068481, 'eval/latency_8/loss': 0.6459647417068481, 'eval/demon_fire_precision': 0.9032805429864253, 'eval/demon_fire_recall': 0.9922336129232681, 'eval/demon_fire_f1': 0.9456698741672834, 'eval/demon_fire_rate_model': 0.9838619922092376, 'eval/demon_fire_rate_teacher': 0.8956594323873122, 'eval/demon_left_precision': 0.7282608695652174, 'eval/demon_left_recall': 0.430406852248394, 'eval/demon_left_f1': 0.5410497981157469, 'eval/demon_left_rate_model': 0.07679465776293823, 'eval/demon_left_rate_teacher': 0.12993878686700056, 'eval/demon_right_precision': 0.6945898778359512, 'eval/demon_right_recall': 0.603030303030303, 'eval/demon_right_f1': 0.6455798864557989, 'eval/demon_right_rate_model': 0.15943238731218698, 'eval/demon_right_rate_teacher': 0.18363939899833054, 'eval/demon_move_f1': 0.5933148422857729, 'eval/demon_cc_f1': 0.802664998373593, 'eval/latency_8/demon_fire_precision': 0.9032805429864253, 'eval/latency_8/demon_fire_recall': 0.9922336129232681, 'eval/latency_8/demon_fire_f1': 0.9456698741672834, 'eval/latency_8/demon_fire_rate_model': 0.9838619922092376, 'eval/latency_8/demon_fire_rate_teacher': 0.8956594323873122, 'eval/latency_8/demon_left_precision': 0.7282608695652174, 'eval/latency_8/demon_left_recall': 0.430406852248394, 'eval/latency_8/demon_left_f1': 0.5410497981157469, 'eval/latency_8/demon_left_rate_model': 0.07679465776293823, 'eval/latency_8/demon_left_rate_teacher': 0.12993878686700056, 'eval/latency_8/demon_right_precision': 0.6945898778359512, 'eval/latency_8/demon_right_recall': 0.603030303030303, 'eval/latency_8/demon_right_f1': 0.6455798864557989, 'eval/latency_8/demon_right_rate_model': 0.15943238731218698, 'eval/latency_8/demon_right_rate_teacher': 0.18363939899833054, 'eval/latency_8/demon_move_f1': 0.5933148422857729, 'eval/latency_8/demon_cc_f1': 0.802664998373593, 'timing/data': 0.0004695439711213112, 'timing/model': 1.1177263697609305, 'learning_rate/qwen_vl_interface': 9.760788569776518e-06, 'learning_rate/action_model': 9.760788569776517e-05, 'epoch': 0.36}) +[2026-06-17 18:01:49,482][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_800_state +[2026-06-17 18:01:49,489][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 18:01:56,171][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_800_state/model.safetensors +[2026-06-17 18:02:11,493][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_800_state/optimizer.bin +[2026-06-17 18:02:11,494][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_800_state/sampler.bin +[2026-06-17 18:02:11,494][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_800_state/sampler_1.bin +[2026-06-17 18:02:11,497][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_800_state/random_states_0.pkl +[2026-06-17 18:02:11,502][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 800 +[2026-06-17 18:02:11,502][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 18:02:11,515][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 18:02:15,616][starVLA.training.train_starvla][INFO] - Step 801, Loss: {'train/loss': 0.6730498671531677, 'train/grad_norm_pre_clip': 9.521907806396484, 'timing/data': 0.0007628863677382469, 'timing/model': 1.397158614359796, 'learning_rate/qwen_vl_interface': 9.760110436201752e-06, 'learning_rate/action_model': 9.76011043620175e-05, 'epoch': 0.36}) +[2026-06-17 18:02:20,607][starVLA.training.train_starvla][INFO] - Step 802, Loss: {'train/loss': 0.39366960525512695, 'train/grad_norm_pre_clip': 6.3655571937561035, 'timing/data': 0.001301320269703865, 'timing/model': 1.2373609282076359, 'learning_rate/qwen_vl_interface': 9.759431367676238e-06, 'learning_rate/action_model': 9.759431367676237e-05, 'epoch': 0.36}) +[2026-06-17 18:02:24,944][starVLA.training.train_starvla][INFO] - Step 803, Loss: {'train/loss': 0.6163610219955444, 'train/grad_norm_pre_clip': 8.795836448669434, 'timing/data': 0.0009859176352620125, 'timing/model': 0.947496211156249, 'learning_rate/qwen_vl_interface': 9.758751364340751e-06, 'learning_rate/action_model': 9.75875136434075e-05, 'epoch': 0.36}) +[2026-06-17 18:02:28,710][starVLA.training.train_starvla][INFO] - Step 804, Loss: {'train/loss': 0.5595424175262451, 'train/grad_norm_pre_clip': 9.060240745544434, 'timing/data': 0.0004904093220829964, 'timing/model': 0.915378232486546, 'learning_rate/qwen_vl_interface': 9.758070426336252e-06, 'learning_rate/action_model': 9.758070426336252e-05, 'epoch': 0.36}) +[2026-06-17 18:02:32,432][starVLA.training.train_starvla][INFO] - Step 805, Loss: {'train/loss': 0.7403808832168579, 'train/grad_norm_pre_clip': 20.216655731201172, 'timing/data': 0.0004979278892278671, 'timing/model': 0.9175240118056536, 'learning_rate/qwen_vl_interface': 9.757388553803905e-06, 'learning_rate/action_model': 9.757388553803905e-05, 'epoch': 0.37}) +[2026-06-17 18:02:36,176][starVLA.training.train_starvla][INFO] - Step 806, Loss: {'train/loss': 0.7540433406829834, 'train/grad_norm_pre_clip': 6.929010391235352, 'timing/data': 0.0008656177669763565, 'timing/model': 0.9249671278521419, 'learning_rate/qwen_vl_interface': 9.75670574688506e-06, 'learning_rate/action_model': 9.75670574688506e-05, 'epoch': 0.37}) +[2026-06-17 18:02:39,913][starVLA.training.train_starvla][INFO] - Step 807, Loss: {'train/loss': 0.6123911142349243, 'train/grad_norm_pre_clip': 7.819106101989746, 'timing/data': 0.0004958333447575569, 'timing/model': 0.9178594732657075, 'learning_rate/qwen_vl_interface': 9.756022005721265e-06, 'learning_rate/action_model': 9.756022005721265e-05, 'epoch': 0.37}) +[2026-06-17 18:02:43,657][starVLA.training.train_starvla][INFO] - Step 808, Loss: {'train/loss': 0.7194092273712158, 'train/grad_norm_pre_clip': 8.282392501831055, 'timing/data': 0.0005214205011725426, 'timing/model': 1.1048137778416276, 'learning_rate/qwen_vl_interface': 9.755337330454257e-06, 'learning_rate/action_model': 9.755337330454257e-05, 'epoch': 0.37}) +[2026-06-17 18:02:47,226][starVLA.training.train_starvla][INFO] - Step 809, Loss: {'train/loss': 0.7572612166404724, 'train/grad_norm_pre_clip': 9.81238079071045, 'timing/data': 0.0004991134628653526, 'timing/model': 0.9325254857540131, 'learning_rate/qwen_vl_interface': 9.754651721225975e-06, 'learning_rate/action_model': 9.754651721225975e-05, 'epoch': 0.37}) +[2026-06-17 18:02:50,965][starVLA.training.train_starvla][INFO] - Step 810, Loss: {'train/loss': 0.6254233121871948, 'train/grad_norm_pre_clip': 9.337925910949707, 'timing/data': 0.0010243169963359833, 'timing/model': 0.9311063643544912, 'learning_rate/qwen_vl_interface': 9.753965178178543e-06, 'learning_rate/action_model': 9.753965178178543e-05, 'epoch': 0.37}) +[2026-06-17 18:02:54,745][starVLA.training.train_starvla][INFO] - Step 811, Loss: {'train/loss': 0.5220128297805786, 'train/grad_norm_pre_clip': 6.849663257598877, 'timing/data': 0.000544433481991291, 'timing/model': 0.9232159499078989, 'learning_rate/qwen_vl_interface': 9.753277701454285e-06, 'learning_rate/action_model': 9.753277701454284e-05, 'epoch': 0.37}) +[2026-06-17 18:02:58,487][starVLA.training.train_starvla][INFO] - Step 812, Loss: {'train/loss': 0.6343428492546082, 'train/grad_norm_pre_clip': 86.91362762451172, 'timing/data': 0.0004898151382803917, 'timing/model': 0.9170160787180066, 'learning_rate/qwen_vl_interface': 9.752589291195711e-06, 'learning_rate/action_model': 9.75258929119571e-05, 'epoch': 0.37}) +[2026-06-17 18:03:02,289][starVLA.training.train_starvla][INFO] - Step 813, Loss: {'train/loss': 0.6536598205566406, 'train/grad_norm_pre_clip': 6.953794479370117, 'timing/data': 0.0004922021180391312, 'timing/model': 0.9215521449223161, 'learning_rate/qwen_vl_interface': 9.751899947545534e-06, 'learning_rate/action_model': 9.751899947545533e-05, 'epoch': 0.37}) +[2026-06-17 18:03:06,079][starVLA.training.train_starvla][INFO] - Step 814, Loss: {'train/loss': 0.624879777431488, 'train/grad_norm_pre_clip': 9.685491561889648, 'timing/data': 0.0004915865138173103, 'timing/model': 0.9549369048327208, 'learning_rate/qwen_vl_interface': 9.751209670646653e-06, 'learning_rate/action_model': 9.751209670646652e-05, 'epoch': 0.37}) +[2026-06-17 18:03:09,972][starVLA.training.train_starvla][INFO] - Step 815, Loss: {'train/loss': 0.615044116973877, 'train/grad_norm_pre_clip': 8.941838264465332, 'timing/data': 0.0005111461505293846, 'timing/model': 1.1394194411113858, 'learning_rate/qwen_vl_interface': 9.750518460642162e-06, 'learning_rate/action_model': 9.75051846064216e-05, 'epoch': 0.37}) +[2026-06-17 18:03:13,623][starVLA.training.train_starvla][INFO] - Step 816, Loss: {'train/loss': 0.6217762231826782, 'train/grad_norm_pre_clip': 10.125825881958008, 'timing/data': 0.0005350802093744278, 'timing/model': 0.9135726317763329, 'learning_rate/qwen_vl_interface': 9.749826317675352e-06, 'learning_rate/action_model': 9.749826317675352e-05, 'epoch': 0.37}) +[2026-06-17 18:03:17,369][starVLA.training.train_starvla][INFO] - Step 817, Loss: {'train/loss': 0.7489773631095886, 'train/grad_norm_pre_clip': 9.026307106018066, 'timing/data': 0.0005660839378833771, 'timing/model': 0.942827851511538, 'learning_rate/qwen_vl_interface': 9.749133241889706e-06, 'learning_rate/action_model': 9.749133241889705e-05, 'epoch': 0.37}) +[2026-06-17 18:03:21,364][starVLA.training.train_starvla][INFO] - Step 818, Loss: {'train/loss': 0.6331118941307068, 'train/grad_norm_pre_clip': 9.526021957397461, 'timing/data': 0.0005340538918972015, 'timing/model': 0.9867144767194986, 'learning_rate/qwen_vl_interface': 9.748439233428897e-06, 'learning_rate/action_model': 9.748439233428896e-05, 'epoch': 0.37}) +[2026-06-17 18:03:25,880][starVLA.training.train_starvla][INFO] - Step 819, Loss: {'train/loss': 0.7405736446380615, 'train/grad_norm_pre_clip': 7.380589008331299, 'timing/data': 0.000533723272383213, 'timing/model': 0.9470347203314304, 'learning_rate/qwen_vl_interface': 9.747744292436794e-06, 'learning_rate/action_model': 9.747744292436793e-05, 'epoch': 0.37}) +[2026-06-17 18:03:29,581][starVLA.training.train_starvla][INFO] - Step 820, Loss: {'train/loss': 0.5035690069198608, 'train/grad_norm_pre_clip': 8.073586463928223, 'timing/data': 0.00045386888086795807, 'timing/model': 0.9168211277574301, 'learning_rate/qwen_vl_interface': 9.747048419057458e-06, 'learning_rate/action_model': 9.747048419057458e-05, 'epoch': 0.37}) +[2026-06-17 18:03:33,450][starVLA.training.train_starvla][INFO] - Step 821, Loss: {'train/loss': 0.6045594811439514, 'train/grad_norm_pre_clip': 7.806750297546387, 'timing/data': 0.00047945789992809296, 'timing/model': 0.9144935375079513, 'learning_rate/qwen_vl_interface': 9.746351613435148e-06, 'learning_rate/action_model': 9.746351613435147e-05, 'epoch': 0.37}) +[2026-06-17 18:03:37,151][starVLA.training.train_starvla][INFO] - Step 822, Loss: {'train/loss': 0.8557933568954468, 'train/grad_norm_pre_clip': 7.875974178314209, 'timing/data': 0.000460255891084671, 'timing/model': 1.0790098262950778, 'learning_rate/qwen_vl_interface': 9.74565387571431e-06, 'learning_rate/action_model': 9.745653875714309e-05, 'epoch': 0.37}) +[2026-06-17 18:03:40,948][starVLA.training.train_starvla][INFO] - Step 823, Loss: {'train/loss': 0.5905309319496155, 'train/grad_norm_pre_clip': 7.543687343597412, 'timing/data': 0.0008704643696546555, 'timing/model': 0.9158010724931955, 'learning_rate/qwen_vl_interface': 9.744955206039585e-06, 'learning_rate/action_model': 9.744955206039585e-05, 'epoch': 0.37}) +[2026-06-17 18:03:44,783][starVLA.training.train_starvla][INFO] - Step 824, Loss: {'train/loss': 0.6314837336540222, 'train/grad_norm_pre_clip': 8.916513442993164, 'timing/data': 0.0005285404622554779, 'timing/model': 1.0114416358992457, 'learning_rate/qwen_vl_interface': 9.74425560455581e-06, 'learning_rate/action_model': 9.74425560455581e-05, 'epoch': 0.37}) +[2026-06-17 18:03:48,524][starVLA.training.train_starvla][INFO] - Step 825, Loss: {'train/loss': 0.8159478902816772, 'train/grad_norm_pre_clip': 9.800771713256836, 'timing/data': 0.0004765978083014488, 'timing/model': 0.921227247454226, 'learning_rate/qwen_vl_interface': 9.743555071408011e-06, 'learning_rate/action_model': 9.743555071408011e-05, 'epoch': 0.37}) +[2026-06-17 18:03:52,239][starVLA.training.train_starvla][INFO] - Step 826, Loss: {'train/loss': 0.6790257096290588, 'train/grad_norm_pre_clip': 8.172898292541504, 'timing/data': 0.0006830347701907158, 'timing/model': 0.928095986135304, 'learning_rate/qwen_vl_interface': 9.742853606741412e-06, 'learning_rate/action_model': 9.742853606741411e-05, 'epoch': 0.37}) +[2026-06-17 18:03:56,194][starVLA.training.train_starvla][INFO] - Step 827, Loss: {'train/loss': 0.5395795106887817, 'train/grad_norm_pre_clip': 8.094789505004883, 'timing/data': 0.0005709119141101837, 'timing/model': 0.937251684255898, 'learning_rate/qwen_vl_interface': 9.742151210701423e-06, 'learning_rate/action_model': 9.742151210701424e-05, 'epoch': 0.38}) +[2026-06-17 18:03:59,915][starVLA.training.train_starvla][INFO] - Step 828, Loss: {'train/loss': 0.4323296844959259, 'train/grad_norm_pre_clip': 11.227173805236816, 'timing/data': 0.0013858098536729813, 'timing/model': 1.081892647780478, 'learning_rate/qwen_vl_interface': 9.741447883433656e-06, 'learning_rate/action_model': 9.741447883433656e-05, 'epoch': 0.38}) +[2026-06-17 18:04:03,444][starVLA.training.train_starvla][INFO] - Step 829, Loss: {'train/loss': 1.1094834804534912, 'train/grad_norm_pre_clip': 14.467849731445312, 'timing/data': 0.0012950794771313667, 'timing/model': 0.9174303663894534, 'learning_rate/qwen_vl_interface': 9.74074362508391e-06, 'learning_rate/action_model': 9.740743625083911e-05, 'epoch': 0.38}) +[2026-06-17 18:04:09,777][starVLA.training.train_starvla][INFO] - Step 830, Loss: {'train/loss': 0.5389866232872009, 'train/grad_norm_pre_clip': 7.910501003265381, 'timing/data': 0.000481962226331234, 'timing/model': 0.9374648248776793, 'learning_rate/qwen_vl_interface': 9.74003843579818e-06, 'learning_rate/action_model': 9.740038435798178e-05, 'epoch': 0.38}) +[2026-06-17 18:04:13,703][starVLA.training.train_starvla][INFO] - Step 831, Loss: {'train/loss': 0.6222675442695618, 'train/grad_norm_pre_clip': 13.385581016540527, 'timing/data': 0.000506778247654438, 'timing/model': 0.9161623157560825, 'learning_rate/qwen_vl_interface': 9.739332315722647e-06, 'learning_rate/action_model': 9.739332315722646e-05, 'epoch': 0.38}) +[2026-06-17 18:04:17,832][starVLA.training.train_starvla][INFO] - Step 832, Loss: {'train/loss': 0.6566609740257263, 'train/grad_norm_pre_clip': 15.49096393585205, 'timing/data': 0.0005604410544037819, 'timing/model': 0.9308681273832917, 'learning_rate/qwen_vl_interface': 9.738625265003695e-06, 'learning_rate/action_model': 9.738625265003695e-05, 'epoch': 0.38}) +[2026-06-17 18:04:21,816][starVLA.training.train_starvla][INFO] - Step 833, Loss: {'train/loss': 0.7002370953559875, 'train/grad_norm_pre_clip': 12.813475608825684, 'timing/data': 0.0007085176184773445, 'timing/model': 0.9675314174965024, 'learning_rate/qwen_vl_interface': 9.7379172837879e-06, 'learning_rate/action_model': 9.737917283787898e-05, 'epoch': 0.38}) +[2026-06-17 18:04:25,605][starVLA.training.train_starvla][INFO] - Step 834, Loss: {'train/loss': 0.7198255658149719, 'train/grad_norm_pre_clip': 9.153203964233398, 'timing/data': 0.0004863329231739044, 'timing/model': 0.9168459856882691, 'learning_rate/qwen_vl_interface': 9.737208372222018e-06, 'learning_rate/action_model': 9.737208372222017e-05, 'epoch': 0.38}) +[2026-06-17 18:04:29,385][starVLA.training.train_starvla][INFO] - Step 835, Loss: {'train/loss': 0.6200343370437622, 'train/grad_norm_pre_clip': 7.8015313148498535, 'timing/data': 0.0009558228775858879, 'timing/model': 1.0918514169752598, 'learning_rate/qwen_vl_interface': 9.736498530453015e-06, 'learning_rate/action_model': 9.736498530453014e-05, 'epoch': 0.38}) +[2026-06-17 18:04:33,556][starVLA.training.train_starvla][INFO] - Step 836, Loss: {'train/loss': 0.6565304398536682, 'train/grad_norm_pre_clip': 16.963050842285156, 'timing/data': 0.0005048662424087524, 'timing/model': 0.9235494546592236, 'learning_rate/qwen_vl_interface': 9.735787758628036e-06, 'learning_rate/action_model': 9.735787758628035e-05, 'epoch': 0.38}) +[2026-06-17 18:04:37,894][starVLA.training.train_starvla][INFO] - Step 837, Loss: {'train/loss': 0.6756252646446228, 'train/grad_norm_pre_clip': 15.806543350219727, 'timing/data': 0.0007202457636594772, 'timing/model': 1.3717230204492807, 'learning_rate/qwen_vl_interface': 9.73507605689443e-06, 'learning_rate/action_model': 9.735076056894429e-05, 'epoch': 0.38}) +[2026-06-17 18:04:42,628][starVLA.training.train_starvla][INFO] - Step 838, Loss: {'train/loss': 0.6808942556381226, 'train/grad_norm_pre_clip': 12.212807655334473, 'timing/data': 0.008752075955271721, 'timing/model': 0.9805126320570707, 'learning_rate/qwen_vl_interface': 9.73436342539973e-06, 'learning_rate/action_model': 9.734363425399728e-05, 'epoch': 0.38}) +[2026-06-17 18:04:46,459][starVLA.training.train_starvla][INFO] - Step 839, Loss: {'train/loss': 0.607302725315094, 'train/grad_norm_pre_clip': 7.313112735748291, 'timing/data': 0.02155644167214632, 'timing/model': 0.9381977124139667, 'learning_rate/qwen_vl_interface': 9.733649864291667e-06, 'learning_rate/action_model': 9.733649864291666e-05, 'epoch': 0.38}) +[2026-06-17 18:04:51,438][starVLA.training.train_starvla][INFO] - Step 840, Loss: {'train/loss': 0.7086578011512756, 'train/grad_norm_pre_clip': 7.748918056488037, 'timing/data': 0.02394327986985445, 'timing/model': 2.181073747575283, 'learning_rate/qwen_vl_interface': 9.732935373718162e-06, 'learning_rate/action_model': 9.732935373718161e-05, 'epoch': 0.38}) +[2026-06-17 18:04:56,380][starVLA.training.train_starvla][INFO] - Step 841, Loss: {'train/loss': 0.8504124879837036, 'train/grad_norm_pre_clip': 14.485611915588379, 'timing/data': 0.015259969979524612, 'timing/model': 1.3389838393777609, 'learning_rate/qwen_vl_interface': 9.73221995382733e-06, 'learning_rate/action_model': 9.73221995382733e-05, 'epoch': 0.38}) +[2026-06-17 18:05:00,171][starVLA.training.train_starvla][INFO] - Step 842, Loss: {'train/loss': 0.7449910044670105, 'train/grad_norm_pre_clip': 11.662778854370117, 'timing/data': 0.01970648020505905, 'timing/model': 1.0147881684824824, 'learning_rate/qwen_vl_interface': 9.73150360476748e-06, 'learning_rate/action_model': 9.73150360476748e-05, 'epoch': 0.38}) +[2026-06-17 18:05:04,326][starVLA.training.train_starvla][INFO] - Step 843, Loss: {'train/loss': 0.6676594018936157, 'train/grad_norm_pre_clip': 9.922039985656738, 'timing/data': 0.023284301161766052, 'timing/model': 1.0667818142101169, 'learning_rate/qwen_vl_interface': 9.73078632668711e-06, 'learning_rate/action_model': 9.73078632668711e-05, 'epoch': 0.38}) +[2026-06-17 18:05:08,324][starVLA.training.train_starvla][INFO] - Step 844, Loss: {'train/loss': 1.0281848907470703, 'train/grad_norm_pre_clip': 11.280805587768555, 'timing/data': 0.017793025821447372, 'timing/model': 1.0516482191160321, 'learning_rate/qwen_vl_interface': 9.730068119734912e-06, 'learning_rate/action_model': 9.730068119734912e-05, 'epoch': 0.38}) +[2026-06-17 18:05:12,164][starVLA.training.train_starvla][INFO] - Step 845, Loss: {'train/loss': 0.8248879313468933, 'train/grad_norm_pre_clip': 8.029438972473145, 'timing/data': 0.017902128398418427, 'timing/model': 0.9528571022674441, 'learning_rate/qwen_vl_interface': 9.729348984059773e-06, 'learning_rate/action_model': 9.729348984059774e-05, 'epoch': 0.38}) +[2026-06-17 18:05:15,989][starVLA.training.train_starvla][INFO] - Step 846, Loss: {'train/loss': 0.9521430134773254, 'train/grad_norm_pre_clip': 15.48859691619873, 'timing/data': 0.015378294512629509, 'timing/model': 0.9747189097106457, 'learning_rate/qwen_vl_interface': 9.728628919810769e-06, 'learning_rate/action_model': 9.728628919810768e-05, 'epoch': 0.38}) +[2026-06-17 18:05:21,717][starVLA.training.train_starvla][INFO] - Step 847, Loss: {'train/loss': 0.6146680116653442, 'train/grad_norm_pre_clip': 11.77815055847168, 'timing/data': 0.02164333686232567, 'timing/model': 0.9794258754700422, 'learning_rate/qwen_vl_interface': 9.727907927137171e-06, 'learning_rate/action_model': 9.727907927137171e-05, 'epoch': 0.38}) +[2026-06-17 18:05:25,517][starVLA.training.train_starvla][INFO] - Step 848, Loss: {'train/loss': 0.4335663318634033, 'train/grad_norm_pre_clip': 9.307130813598633, 'timing/data': 0.015612329356372356, 'timing/model': 1.1025075512006879, 'learning_rate/qwen_vl_interface': 9.72718600618844e-06, 'learning_rate/action_model': 9.727186006188441e-05, 'epoch': 0.38}) +[2026-06-17 18:05:29,278][starVLA.training.train_starvla][INFO] - Step 849, Loss: {'train/loss': 0.6999741792678833, 'train/grad_norm_pre_clip': 9.404025077819824, 'timing/data': 0.026648941449820995, 'timing/model': 0.9214473152533174, 'learning_rate/qwen_vl_interface': 9.726463157114233e-06, 'learning_rate/action_model': 9.726463157114234e-05, 'epoch': 0.39}) +[2026-06-17 18:05:33,121][starVLA.training.train_starvla][INFO] - Step 850, Loss: {'train/loss': 0.8087890148162842, 'train/grad_norm_pre_clip': 10.69273567199707, 'timing/data': 0.025187479332089424, 'timing/model': 0.9333511535078287, 'learning_rate/qwen_vl_interface': 9.725739380064399e-06, 'learning_rate/action_model': 9.725739380064398e-05, 'epoch': 0.39}) +[2026-06-17 18:05:37,607][starVLA.training.train_starvla][INFO] - Step 851, Loss: {'train/loss': 0.6402234435081482, 'train/grad_norm_pre_clip': 7.298975944519043, 'timing/data': 0.018614938482642174, 'timing/model': 1.6473061805590987, 'learning_rate/qwen_vl_interface': 9.725014675188972e-06, 'learning_rate/action_model': 9.725014675188973e-05, 'epoch': 0.39}) +[2026-06-17 18:05:42,936][starVLA.training.train_starvla][INFO] - Step 852, Loss: {'train/loss': 0.5836312770843506, 'train/grad_norm_pre_clip': 8.988438606262207, 'timing/data': 0.020602744072675705, 'timing/model': 0.9989244397729635, 'learning_rate/qwen_vl_interface': 9.72428904263819e-06, 'learning_rate/action_model': 9.72428904263819e-05, 'epoch': 0.39}) +[2026-06-17 18:05:46,958][starVLA.training.train_starvla][INFO] - Step 853, Loss: {'train/loss': 0.7629855275154114, 'train/grad_norm_pre_clip': 8.267891883850098, 'timing/data': 0.016173547133803368, 'timing/model': 0.9927663570269942, 'learning_rate/qwen_vl_interface': 9.723562482562474e-06, 'learning_rate/action_model': 9.723562482562473e-05, 'epoch': 0.39}) +[2026-06-17 18:05:50,959][starVLA.training.train_starvla][INFO] - Step 854, Loss: {'train/loss': 0.7659901976585388, 'train/grad_norm_pre_clip': 10.024163246154785, 'timing/data': 0.024586686864495277, 'timing/model': 1.1937069352716208, 'learning_rate/qwen_vl_interface': 9.72283499511244e-06, 'learning_rate/action_model': 9.722834995112439e-05, 'epoch': 0.39}) +[2026-06-17 18:05:54,761][starVLA.training.train_starvla][INFO] - Step 855, Loss: {'train/loss': 0.8748337030410767, 'train/grad_norm_pre_clip': 13.218222618103027, 'timing/data': 0.025901375338435173, 'timing/model': 1.1024044249206781, 'learning_rate/qwen_vl_interface': 9.722106580438901e-06, 'learning_rate/action_model': 9.7221065804389e-05, 'epoch': 0.39}) +[2026-06-17 18:05:58,750][starVLA.training.train_starvla][INFO] - Step 856, Loss: {'train/loss': 0.7817081212997437, 'train/grad_norm_pre_clip': 10.401293754577637, 'timing/data': 0.021442417986690998, 'timing/model': 1.080906929448247, 'learning_rate/qwen_vl_interface': 9.721377238692855e-06, 'learning_rate/action_model': 9.721377238692854e-05, 'epoch': 0.39}) +[2026-06-17 18:06:02,696][starVLA.training.train_starvla][INFO] - Step 857, Loss: {'train/loss': 0.6592310667037964, 'train/grad_norm_pre_clip': 7.510798931121826, 'timing/data': 0.019667436368763447, 'timing/model': 0.9284693272784352, 'learning_rate/qwen_vl_interface': 9.720646970025494e-06, 'learning_rate/action_model': 9.720646970025494e-05, 'epoch': 0.39}) +[2026-06-17 18:06:07,057][starVLA.training.train_starvla][INFO] - Step 858, Loss: {'train/loss': 0.5664501190185547, 'train/grad_norm_pre_clip': 8.001293182373047, 'timing/data': 0.02323335502296686, 'timing/model': 1.2048362120985985, 'learning_rate/qwen_vl_interface': 9.719915774588209e-06, 'learning_rate/action_model': 9.719915774588209e-05, 'epoch': 0.39}) +[2026-06-17 18:06:10,992][starVLA.training.train_starvla][INFO] - Step 859, Loss: {'train/loss': 0.45852360129356384, 'train/grad_norm_pre_clip': 6.246254920959473, 'timing/data': 0.02268958557397127, 'timing/model': 0.9374626707285643, 'learning_rate/qwen_vl_interface': 9.719183652532572e-06, 'learning_rate/action_model': 9.719183652532572e-05, 'epoch': 0.39}) +[2026-06-17 18:06:17,700][starVLA.training.train_starvla][INFO] - Step 860, Loss: {'train/loss': 0.6666615605354309, 'train/grad_norm_pre_clip': 9.397308349609375, 'timing/data': 0.017779560759663582, 'timing/model': 3.5983553491532803, 'learning_rate/qwen_vl_interface': 9.718450604010355e-06, 'learning_rate/action_model': 9.718450604010355e-05, 'epoch': 0.39}) +[2026-06-17 18:06:22,015][starVLA.training.train_starvla][INFO] - Step 861, Loss: {'train/loss': 0.8777643442153931, 'train/grad_norm_pre_clip': 7.8277716636657715, 'timing/data': 0.01972116157412529, 'timing/model': 1.163158681243658, 'learning_rate/qwen_vl_interface': 9.717716629173521e-06, 'learning_rate/action_model': 9.71771662917352e-05, 'epoch': 0.39}) +[2026-06-17 18:06:25,674][starVLA.training.train_starvla][INFO] - Step 862, Loss: {'train/loss': 1.0229674577713013, 'train/grad_norm_pre_clip': 11.35114860534668, 'timing/data': 0.014341176487505436, 'timing/model': 0.9363465057685971, 'learning_rate/qwen_vl_interface': 9.71698172817422e-06, 'learning_rate/action_model': 9.716981728174219e-05, 'epoch': 0.39}) +[2026-06-17 18:06:29,539][starVLA.training.train_starvla][INFO] - Step 863, Loss: {'train/loss': 0.6864290833473206, 'train/grad_norm_pre_clip': 6.584697723388672, 'timing/data': 0.019457013346254826, 'timing/model': 0.9471276123076677, 'learning_rate/qwen_vl_interface': 9.7162459011648e-06, 'learning_rate/action_model': 9.7162459011648e-05, 'epoch': 0.39}) +[2026-06-17 18:06:33,334][starVLA.training.train_starvla][INFO] - Step 864, Loss: {'train/loss': 0.9466326832771301, 'train/grad_norm_pre_clip': 9.886838912963867, 'timing/data': 0.01662042737007141, 'timing/model': 0.9310134723782539, 'learning_rate/qwen_vl_interface': 9.7155091482978e-06, 'learning_rate/action_model': 9.7155091482978e-05, 'epoch': 0.39}) +[2026-06-17 18:06:37,168][starVLA.training.train_starvla][INFO] - Step 865, Loss: {'train/loss': 0.4657396674156189, 'train/grad_norm_pre_clip': 6.186907768249512, 'timing/data': 0.015127222053706646, 'timing/model': 0.9301190758123994, 'learning_rate/qwen_vl_interface': 9.714771469725947e-06, 'learning_rate/action_model': 9.714771469725947e-05, 'epoch': 0.39}) +[2026-06-17 18:06:41,019][starVLA.training.train_starvla][INFO] - Step 866, Loss: {'train/loss': 0.523290753364563, 'train/grad_norm_pre_clip': 7.3881659507751465, 'timing/data': 0.016907647252082825, 'timing/model': 0.9518261896446347, 'learning_rate/qwen_vl_interface': 9.714032865602167e-06, 'learning_rate/action_model': 9.714032865602166e-05, 'epoch': 0.39}) +[2026-06-17 18:06:45,118][starVLA.training.train_starvla][INFO] - Step 867, Loss: {'train/loss': 0.6276309490203857, 'train/grad_norm_pre_clip': 6.854348659515381, 'timing/data': 0.02236923575401306, 'timing/model': 1.1574069010093808, 'learning_rate/qwen_vl_interface': 9.713293336079568e-06, 'learning_rate/action_model': 9.713293336079568e-05, 'epoch': 0.39}) +[2026-06-17 18:06:48,870][starVLA.training.train_starvla][INFO] - Step 868, Loss: {'train/loss': 0.4664338231086731, 'train/grad_norm_pre_clip': 6.746013641357422, 'timing/data': 0.015943252481520176, 'timing/model': 1.059201030060649, 'learning_rate/qwen_vl_interface': 9.712552881311459e-06, 'learning_rate/action_model': 9.712552881311458e-05, 'epoch': 0.39}) +[2026-06-17 18:06:52,548][starVLA.training.train_starvla][INFO] - Step 869, Loss: {'train/loss': 0.7956885099411011, 'train/grad_norm_pre_clip': 8.384596824645996, 'timing/data': 0.02415869291871786, 'timing/model': 0.9427260058000684, 'learning_rate/qwen_vl_interface': 9.711811501451336e-06, 'learning_rate/action_model': 9.711811501451335e-05, 'epoch': 0.39}) +[2026-06-17 18:06:56,603][starVLA.training.train_starvla][INFO] - Step 870, Loss: {'train/loss': 0.7082111239433289, 'train/grad_norm_pre_clip': 19.5551700592041, 'timing/data': 0.02780719194561243, 'timing/model': 0.9443605653941631, 'learning_rate/qwen_vl_interface': 9.711069196652886e-06, 'learning_rate/action_model': 9.711069196652886e-05, 'epoch': 0.39}) +[2026-06-17 18:07:00,423][starVLA.training.train_starvla][INFO] - Step 871, Loss: {'train/loss': 0.5937615633010864, 'train/grad_norm_pre_clip': 6.221062183380127, 'timing/data': 0.021746832877397537, 'timing/model': 0.9248489560559392, 'learning_rate/qwen_vl_interface': 9.710325967069992e-06, 'learning_rate/action_model': 9.710325967069991e-05, 'epoch': 0.4}) +[2026-06-17 18:07:04,254][starVLA.training.train_starvla][INFO] - Step 872, Loss: {'train/loss': 0.4985549747943878, 'train/grad_norm_pre_clip': 11.533035278320312, 'timing/data': 0.02276435773819685, 'timing/model': 0.9762654630467296, 'learning_rate/qwen_vl_interface': 9.709581812856725e-06, 'learning_rate/action_model': 9.709581812856726e-05, 'epoch': 0.4}) +[2026-06-17 18:07:08,253][starVLA.training.train_starvla][INFO] - Step 873, Loss: {'train/loss': 0.5740264654159546, 'train/grad_norm_pre_clip': 21.12799644470215, 'timing/data': 0.01800005789846182, 'timing/model': 0.9616426099091768, 'learning_rate/qwen_vl_interface': 9.708836734167351e-06, 'learning_rate/action_model': 9.708836734167351e-05, 'epoch': 0.4}) +[2026-06-17 18:07:12,215][starVLA.training.train_starvla][INFO] - Step 874, Loss: {'train/loss': 0.6580607295036316, 'train/grad_norm_pre_clip': 8.947256088256836, 'timing/data': 0.01974219735711813, 'timing/model': 1.1876511638984084, 'learning_rate/qwen_vl_interface': 9.708090731156321e-06, 'learning_rate/action_model': 9.708090731156321e-05, 'epoch': 0.4}) +[2026-06-17 18:07:16,042][starVLA.training.train_starvla][INFO] - Step 875, Loss: {'train/loss': 0.6015031933784485, 'train/grad_norm_pre_clip': 7.280036926269531, 'timing/data': 0.019984985701739788, 'timing/model': 1.1087624533101916, 'learning_rate/qwen_vl_interface': 9.707343803978288e-06, 'learning_rate/action_model': 9.707343803978288e-05, 'epoch': 0.4}) +[2026-06-17 18:07:19,748][starVLA.training.train_starvla][INFO] - Step 876, Loss: {'train/loss': 0.6242761611938477, 'train/grad_norm_pre_clip': 7.516002655029297, 'timing/data': 0.018198788166046143, 'timing/model': 0.9468631185591221, 'learning_rate/qwen_vl_interface': 9.706595952788088e-06, 'learning_rate/action_model': 9.706595952788088e-05, 'epoch': 0.4}) +[2026-06-17 18:07:23,650][starVLA.training.train_starvla][INFO] - Step 877, Loss: {'train/loss': 0.41522926092147827, 'train/grad_norm_pre_clip': 7.020659446716309, 'timing/data': 0.018752642907202244, 'timing/model': 0.9297390142455697, 'learning_rate/qwen_vl_interface': 9.705847177740751e-06, 'learning_rate/action_model': 9.705847177740751e-05, 'epoch': 0.4}) +[2026-06-17 18:07:27,473][starVLA.training.train_starvla][INFO] - Step 878, Loss: {'train/loss': 0.8405797481536865, 'train/grad_norm_pre_clip': 9.956923484802246, 'timing/data': 0.01843676157295704, 'timing/model': 0.9320381600409746, 'learning_rate/qwen_vl_interface': 9.7050974789915e-06, 'learning_rate/action_model': 9.7050974789915e-05, 'epoch': 0.4}) +[2026-06-17 18:07:31,377][starVLA.training.train_starvla][INFO] - Step 879, Loss: {'train/loss': 0.4285762310028076, 'train/grad_norm_pre_clip': 11.411514282226562, 'timing/data': 0.025514205917716026, 'timing/model': 0.9880287060514092, 'learning_rate/qwen_vl_interface': 9.704346856695747e-06, 'learning_rate/action_model': 9.704346856695746e-05, 'epoch': 0.4}) +[2026-06-17 18:07:35,219][starVLA.training.train_starvla][INFO] - Step 880, Loss: {'train/loss': 0.5763726234436035, 'train/grad_norm_pre_clip': 41.7138671875, 'timing/data': 0.016681062057614326, 'timing/model': 0.9469652902334929, 'learning_rate/qwen_vl_interface': 9.703595311009098e-06, 'learning_rate/action_model': 9.703595311009097e-05, 'epoch': 0.4}) +[2026-06-17 18:07:39,020][starVLA.training.train_starvla][INFO] - Step 881, Loss: {'train/loss': 0.6398542523384094, 'train/grad_norm_pre_clip': 9.150994300842285, 'timing/data': 0.016734600998461246, 'timing/model': 1.0965084917843342, 'learning_rate/qwen_vl_interface': 9.702842842087348e-06, 'learning_rate/action_model': 9.702842842087348e-05, 'epoch': 0.4}) +[2026-06-17 18:07:42,780][starVLA.training.train_starvla][INFO] - Step 882, Loss: {'train/loss': 0.6850093603134155, 'train/grad_norm_pre_clip': 13.321680068969727, 'timing/data': 0.020836351439356804, 'timing/model': 1.0971455574035645, 'learning_rate/qwen_vl_interface': 9.702089450086486e-06, 'learning_rate/action_model': 9.702089450086486e-05, 'epoch': 0.4}) +[2026-06-17 18:07:46,425][starVLA.training.train_starvla][INFO] - Step 883, Loss: {'train/loss': 0.44715186953544617, 'train/grad_norm_pre_clip': 8.016764640808105, 'timing/data': 0.019405117258429527, 'timing/model': 0.9508073450997472, 'learning_rate/qwen_vl_interface': 9.70133513516269e-06, 'learning_rate/action_model': 9.701335135162688e-05, 'epoch': 0.4}) +[2026-06-17 18:07:50,269][starVLA.training.train_starvla][INFO] - Step 884, Loss: {'train/loss': 0.6892074942588806, 'train/grad_norm_pre_clip': 17.101150512695312, 'timing/data': 0.02385473996400833, 'timing/model': 0.9268355770036578, 'learning_rate/qwen_vl_interface': 9.70057989747233e-06, 'learning_rate/action_model': 9.700579897472329e-05, 'epoch': 0.4}) +[2026-06-17 18:07:54,043][starVLA.training.train_starvla][INFO] - Step 885, Loss: {'train/loss': 0.4550643563270569, 'train/grad_norm_pre_clip': 8.95749282836914, 'timing/data': 0.02169721946120262, 'timing/model': 0.9325082348659635, 'learning_rate/qwen_vl_interface': 9.699823737171969e-06, 'learning_rate/action_model': 9.699823737171969e-05, 'epoch': 0.4}) +[2026-06-17 18:07:57,897][starVLA.training.train_starvla][INFO] - Step 886, Loss: {'train/loss': 0.7985920906066895, 'train/grad_norm_pre_clip': 10.40897274017334, 'timing/data': 0.01601369120180607, 'timing/model': 0.9272087682038546, 'learning_rate/qwen_vl_interface': 9.699066654418358e-06, 'learning_rate/action_model': 9.699066654418358e-05, 'epoch': 0.4}) +[2026-06-17 18:08:01,654][starVLA.training.train_starvla][INFO] - Step 887, Loss: {'train/loss': 0.6170615553855896, 'train/grad_norm_pre_clip': 9.873024940490723, 'timing/data': 0.01643526740372181, 'timing/model': 1.1036295481026173, 'learning_rate/qwen_vl_interface': 9.698308649368444e-06, 'learning_rate/action_model': 9.698308649368444e-05, 'epoch': 0.4}) +[2026-06-17 18:08:05,453][starVLA.training.train_starvla][INFO] - Step 888, Loss: {'train/loss': 0.2656681537628174, 'train/grad_norm_pre_clip': 4.147326946258545, 'timing/data': 0.022029563784599304, 'timing/model': 1.0516715217381716, 'learning_rate/qwen_vl_interface': 9.69754972217936e-06, 'learning_rate/action_model': 9.69754972217936e-05, 'epoch': 0.4}) +[2026-06-17 18:08:09,330][starVLA.training.train_starvla][INFO] - Step 889, Loss: {'train/loss': 0.7632008194923401, 'train/grad_norm_pre_clip': 10.959945678710938, 'timing/data': 0.018141500651836395, 'timing/model': 1.1626295112073421, 'learning_rate/qwen_vl_interface': 9.696789873008432e-06, 'learning_rate/action_model': 9.696789873008433e-05, 'epoch': 0.4}) +[2026-06-17 18:08:13,127][starVLA.training.train_starvla][INFO] - Step 890, Loss: {'train/loss': 0.5788105130195618, 'train/grad_norm_pre_clip': 7.8456645011901855, 'timing/data': 0.01819274853914976, 'timing/model': 0.9349044281989336, 'learning_rate/qwen_vl_interface': 9.69602910201318e-06, 'learning_rate/action_model': 9.696029102013179e-05, 'epoch': 0.4}) +[2026-06-17 18:08:16,940][starVLA.training.train_starvla][INFO] - Step 891, Loss: {'train/loss': 0.8371201753616333, 'train/grad_norm_pre_clip': 13.57686996459961, 'timing/data': 0.016418895684182644, 'timing/model': 0.9351361365988851, 'learning_rate/qwen_vl_interface': 9.695267409351308e-06, 'learning_rate/action_model': 9.695267409351307e-05, 'epoch': 0.4}) +[2026-06-17 18:08:22,035][starVLA.training.train_starvla][INFO] - Step 892, Loss: {'train/loss': 0.68171626329422, 'train/grad_norm_pre_clip': 8.572842597961426, 'timing/data': 0.01809894759207964, 'timing/model': 0.9308729823678732, 'learning_rate/qwen_vl_interface': 9.694504795180722e-06, 'learning_rate/action_model': 9.694504795180721e-05, 'epoch': 0.4}) +[2026-06-17 18:08:27,613][starVLA.training.train_starvla][INFO] - Step 893, Loss: {'train/loss': 0.6388157606124878, 'train/grad_norm_pre_clip': 8.927919387817383, 'timing/data': 0.020950318314135075, 'timing/model': 2.7553956424817443, 'learning_rate/qwen_vl_interface': 9.693741259659507e-06, 'learning_rate/action_model': 9.693741259659507e-05, 'epoch': 0.41}) +[2026-06-17 18:08:32,503][starVLA.training.train_starvla][INFO] - Step 894, Loss: {'train/loss': 0.7785860896110535, 'train/grad_norm_pre_clip': 11.657029151916504, 'timing/data': 0.015428533777594566, 'timing/model': 1.1068208757787943, 'learning_rate/qwen_vl_interface': 9.692976802945947e-06, 'learning_rate/action_model': 9.692976802945948e-05, 'epoch': 0.41}) +[2026-06-17 18:08:36,206][starVLA.training.train_starvla][INFO] - Step 895, Loss: {'train/loss': 0.6682780385017395, 'train/grad_norm_pre_clip': 9.261200904846191, 'timing/data': 0.01758422516286373, 'timing/model': 0.9652079921215773, 'learning_rate/qwen_vl_interface': 9.692211425198517e-06, 'learning_rate/action_model': 9.692211425198517e-05, 'epoch': 0.41}) +[2026-06-17 18:08:40,166][starVLA.training.train_starvla][INFO] - Step 896, Loss: {'train/loss': 0.6199880838394165, 'train/grad_norm_pre_clip': 9.303013801574707, 'timing/data': 0.016568563878536224, 'timing/model': 0.92381726577878, 'learning_rate/qwen_vl_interface': 9.691445126575879e-06, 'learning_rate/action_model': 9.691445126575877e-05, 'epoch': 0.41}) +[2026-06-17 18:08:44,972][starVLA.training.train_starvla][INFO] - Step 897, Loss: {'train/loss': 0.8922607898712158, 'train/grad_norm_pre_clip': 10.1297025680542, 'timing/data': 0.020247526466846466, 'timing/model': 0.919704801402986, 'learning_rate/qwen_vl_interface': 9.690677907236884e-06, 'learning_rate/action_model': 9.690677907236884e-05, 'epoch': 0.41}) +[2026-06-17 18:08:48,764][starVLA.training.train_starvla][INFO] - Step 898, Loss: {'train/loss': 1.1467896699905396, 'train/grad_norm_pre_clip': 10.8146333694458, 'timing/data': 0.029169362969696522, 'timing/model': 0.9287787042558193, 'learning_rate/qwen_vl_interface': 9.689909767340582e-06, 'learning_rate/action_model': 9.689909767340583e-05, 'epoch': 0.41}) +[2026-06-17 18:08:52,606][starVLA.training.train_starvla][INFO] - Step 899, Loss: {'train/loss': 0.5132322907447815, 'train/grad_norm_pre_clip': 6.599815845489502, 'timing/data': 0.016857415437698364, 'timing/model': 0.9462926620617509, 'learning_rate/qwen_vl_interface': 9.689140707046209e-06, 'learning_rate/action_model': 9.689140707046209e-05, 'epoch': 0.41}) +[2026-06-17 18:10:56,665][starVLA.training.train_starvla][INFO] - Step 900, Loss: {'train/loss': 0.41351017355918884, 'train/grad_norm_pre_clip': 5.86962366104126, 'eval/loss': 0.6254054307937622, 'eval/latency_8/loss': 0.6254054307937622, 'eval/demon_fire_precision': 0.9144092219020173, 'eval/demon_fire_recall': 0.9857098477788133, 'eval/demon_fire_f1': 0.9487217820301989, 'eval/demon_fire_rate_model': 0.9654980523094046, 'eval/demon_fire_rate_teacher': 0.8956594323873122, 'eval/demon_left_precision': 0.776824034334764, 'eval/demon_left_recall': 0.3875802997858672, 'eval/demon_left_f1': 0.5171428571428571, 'eval/demon_left_rate_model': 0.06483027267668336, 'eval/demon_left_rate_teacher': 0.12993878686700056, 'eval/demon_right_precision': 0.5929526123936817, 'eval/demon_right_recall': 0.7393939393939394, 'eval/demon_right_f1': 0.6581254214430209, 'eval/demon_right_rate_model': 0.22899276572064553, 'eval/demon_right_rate_teacher': 0.18363939899833054, 'eval/demon_move_f1': 0.587634139292939, 'eval/demon_cc_f1': 0.8010531663678913, 'eval/latency_8/demon_fire_precision': 0.9144092219020173, 'eval/latency_8/demon_fire_recall': 0.9857098477788133, 'eval/latency_8/demon_fire_f1': 0.9487217820301989, 'eval/latency_8/demon_fire_rate_model': 0.9654980523094046, 'eval/latency_8/demon_fire_rate_teacher': 0.8956594323873122, 'eval/latency_8/demon_left_precision': 0.776824034334764, 'eval/latency_8/demon_left_recall': 0.3875802997858672, 'eval/latency_8/demon_left_f1': 0.5171428571428571, 'eval/latency_8/demon_left_rate_model': 0.06483027267668336, 'eval/latency_8/demon_left_rate_teacher': 0.12993878686700056, 'eval/latency_8/demon_right_precision': 0.5929526123936817, 'eval/latency_8/demon_right_recall': 0.7393939393939394, 'eval/latency_8/demon_right_f1': 0.6581254214430209, 'eval/latency_8/demon_right_rate_model': 0.22899276572064553, 'eval/latency_8/demon_right_rate_teacher': 0.18363939899833054, 'eval/latency_8/demon_move_f1': 0.587634139292939, 'eval/latency_8/demon_cc_f1': 0.8010531663678913, 'timing/data': 0.018390721641480923, 'timing/model': 1.2296191910281777, 'learning_rate/qwen_vl_interface': 9.68837072651319e-06, 'learning_rate/action_model': 9.688370726513188e-05, 'epoch': 0.41}) +[2026-06-17 18:10:56,668][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_900_state +[2026-06-17 18:10:56,673][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 18:11:02,996][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_900_state/model.safetensors +[2026-06-17 18:11:18,986][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_900_state/optimizer.bin +[2026-06-17 18:11:18,988][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_900_state/sampler.bin +[2026-06-17 18:11:18,988][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_900_state/sampler_1.bin +[2026-06-17 18:11:18,990][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_900_state/random_states_0.pkl +[2026-06-17 18:11:18,993][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 900 +[2026-06-17 18:11:18,993][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 18:11:19,007][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 18:11:23,413][starVLA.training.train_starvla][INFO] - Step 901, Loss: {'train/loss': 0.36362573504447937, 'train/grad_norm_pre_clip': 7.330333709716797, 'timing/data': 0.0159523393958807, 'timing/model': 1.1243489561602473, 'learning_rate/qwen_vl_interface': 9.687599825901141e-06, 'learning_rate/action_model': 9.687599825901142e-05, 'epoch': 0.41}) +[2026-06-17 18:11:27,342][starVLA.training.train_starvla][INFO] - Step 902, Loss: {'train/loss': 0.7837556600570679, 'train/grad_norm_pre_clip': 10.771125793457031, 'timing/data': 0.017368494533002377, 'timing/model': 1.1358004938811064, 'learning_rate/qwen_vl_interface': 9.686828005369876e-06, 'learning_rate/action_model': 9.686828005369875e-05, 'epoch': 0.41}) +[2026-06-17 18:11:31,355][starVLA.training.train_starvla][INFO] - Step 903, Loss: {'train/loss': 0.768515408039093, 'train/grad_norm_pre_clip': 12.10499382019043, 'timing/data': 0.016947604715824127, 'timing/model': 1.114054842852056, 'learning_rate/qwen_vl_interface': 9.68605526507939e-06, 'learning_rate/action_model': 9.686055265079389e-05, 'epoch': 0.41}) +[2026-06-17 18:11:34,972][starVLA.training.train_starvla][INFO] - Step 904, Loss: {'train/loss': 0.6687931418418884, 'train/grad_norm_pre_clip': 9.240504264831543, 'timing/data': 0.01771559100598097, 'timing/model': 0.9252638770267367, 'learning_rate/qwen_vl_interface': 9.685281605189874e-06, 'learning_rate/action_model': 9.685281605189874e-05, 'epoch': 0.41}) +[2026-06-17 18:11:38,773][starVLA.training.train_starvla][INFO] - Step 905, Loss: {'train/loss': 0.6510202884674072, 'train/grad_norm_pre_clip': 9.163254737854004, 'timing/data': 0.01845554355531931, 'timing/model': 0.9286143481731415, 'learning_rate/qwen_vl_interface': 9.68450702586171e-06, 'learning_rate/action_model': 9.684507025861709e-05, 'epoch': 0.41}) +[2026-06-17 18:11:42,581][starVLA.training.train_starvla][INFO] - Step 906, Loss: {'train/loss': 0.7533791065216064, 'train/grad_norm_pre_clip': 8.56877613067627, 'timing/data': 0.0184791823849082, 'timing/model': 0.9285155851393938, 'learning_rate/qwen_vl_interface': 9.683731527255467e-06, 'learning_rate/action_model': 9.683731527255467e-05, 'epoch': 0.41}) +[2026-06-17 18:11:46,597][starVLA.training.train_starvla][INFO] - Step 907, Loss: {'train/loss': 0.28143393993377686, 'train/grad_norm_pre_clip': 6.27931547164917, 'timing/data': 0.018776346929371357, 'timing/model': 1.101857186295092, 'learning_rate/qwen_vl_interface': 9.682955109531906e-06, 'learning_rate/action_model': 9.682955109531905e-05, 'epoch': 0.41}) +[2026-06-17 18:11:50,455][starVLA.training.train_starvla][INFO] - Step 908, Loss: {'train/loss': 0.6337200999259949, 'train/grad_norm_pre_clip': 7.748345375061035, 'timing/data': 0.025424981489777565, 'timing/model': 0.918683641590178, 'learning_rate/qwen_vl_interface': 9.682177772851983e-06, 'learning_rate/action_model': 9.682177772851982e-05, 'epoch': 0.41}) +[2026-06-17 18:11:54,650][starVLA.training.train_starvla][INFO] - Step 909, Loss: {'train/loss': 1.016440749168396, 'train/grad_norm_pre_clip': 11.008329391479492, 'timing/data': 0.018831598572432995, 'timing/model': 1.1079650362953544, 'learning_rate/qwen_vl_interface': 9.681399517376837e-06, 'learning_rate/action_model': 9.681399517376837e-05, 'epoch': 0.41}) +[2026-06-17 18:11:58,418][starVLA.training.train_starvla][INFO] - Step 910, Loss: {'train/loss': 0.7464098930358887, 'train/grad_norm_pre_clip': 11.298884391784668, 'timing/data': 0.019350592978298664, 'timing/model': 1.0324579067528248, 'learning_rate/qwen_vl_interface': 9.680620343267803e-06, 'learning_rate/action_model': 9.680620343267802e-05, 'epoch': 0.41}) +[2026-06-17 18:12:02,073][starVLA.training.train_starvla][INFO] - Step 911, Loss: {'train/loss': 0.6765888333320618, 'train/grad_norm_pre_clip': 7.675837516784668, 'timing/data': 0.016511888243258, 'timing/model': 0.9343523252755404, 'learning_rate/qwen_vl_interface': 9.679840250686403e-06, 'learning_rate/action_model': 9.679840250686404e-05, 'epoch': 0.41}) +[2026-06-17 18:12:06,574][starVLA.training.train_starvla][INFO] - Step 912, Loss: {'train/loss': 0.7872914671897888, 'train/grad_norm_pre_clip': 9.999872207641602, 'timing/data': 0.017918512225151062, 'timing/model': 0.9260613201186061, 'learning_rate/qwen_vl_interface': 9.679059239794355e-06, 'learning_rate/action_model': 9.679059239794355e-05, 'epoch': 0.41}) +[2026-06-17 18:12:10,432][starVLA.training.train_starvla][INFO] - Step 913, Loss: {'train/loss': 0.6012996435165405, 'train/grad_norm_pre_clip': 7.186434268951416, 'timing/data': 0.024105556309223175, 'timing/model': 0.9429850578308105, 'learning_rate/qwen_vl_interface': 9.67827731075356e-06, 'learning_rate/action_model': 9.67827731075356e-05, 'epoch': 0.41}) +[2026-06-17 18:12:14,243][starVLA.training.train_starvla][INFO] - Step 914, Loss: {'train/loss': 0.7278757095336914, 'train/grad_norm_pre_clip': 7.952214241027832, 'timing/data': 0.01835851464420557, 'timing/model': 0.9312662249431014, 'learning_rate/qwen_vl_interface': 9.677494463726112e-06, 'learning_rate/action_model': 9.677494463726112e-05, 'epoch': 0.41}) +[2026-06-17 18:12:18,534][starVLA.training.train_starvla][INFO] - Step 915, Loss: {'train/loss': 0.909963071346283, 'train/grad_norm_pre_clip': 8.542952537536621, 'timing/data': 0.021868865936994553, 'timing/model': 1.1469000363722444, 'learning_rate/qwen_vl_interface': 9.676710698874299e-06, 'learning_rate/action_model': 9.676710698874298e-05, 'epoch': 0.42}) +[2026-06-17 18:12:22,239][starVLA.training.train_starvla][INFO] - Step 916, Loss: {'train/loss': 0.4720735549926758, 'train/grad_norm_pre_clip': 13.185142517089844, 'timing/data': 0.016040127724409103, 'timing/model': 1.0479638054966927, 'learning_rate/qwen_vl_interface': 9.675926016360595e-06, 'learning_rate/action_model': 9.675926016360595e-05, 'epoch': 0.42}) +[2026-06-17 18:12:25,963][starVLA.training.train_starvla][INFO] - Step 917, Loss: {'train/loss': 0.48402684926986694, 'train/grad_norm_pre_clip': 6.313776969909668, 'timing/data': 0.02256004512310028, 'timing/model': 0.9392747385427356, 'learning_rate/qwen_vl_interface': 9.675140416347664e-06, 'learning_rate/action_model': 9.675140416347663e-05, 'epoch': 0.42}) +[2026-06-17 18:12:29,749][starVLA.training.train_starvla][INFO] - Step 918, Loss: {'train/loss': 0.9017409682273865, 'train/grad_norm_pre_clip': 8.184906959533691, 'timing/data': 0.023419253528118134, 'timing/model': 0.9195242375135422, 'learning_rate/qwen_vl_interface': 9.674353898998362e-06, 'learning_rate/action_model': 9.674353898998363e-05, 'epoch': 0.42}) +[2026-06-17 18:12:33,552][starVLA.training.train_starvla][INFO] - Step 919, Loss: {'train/loss': 0.9693425297737122, 'train/grad_norm_pre_clip': 10.177619934082031, 'timing/data': 0.023353767581284046, 'timing/model': 0.9287536768242717, 'learning_rate/qwen_vl_interface': 9.673566464475738e-06, 'learning_rate/action_model': 9.673566464475739e-05, 'epoch': 0.42}) +[2026-06-17 18:12:37,330][starVLA.training.train_starvla][INFO] - Step 920, Loss: {'train/loss': 0.484427273273468, 'train/grad_norm_pre_clip': 7.342180252075195, 'timing/data': 0.02846459113061428, 'timing/model': 0.9260082868859172, 'learning_rate/qwen_vl_interface': 9.672778112943026e-06, 'learning_rate/action_model': 9.672778112943026e-05, 'epoch': 0.42}) +[2026-06-17 18:12:41,035][starVLA.training.train_starvla][INFO] - Step 921, Loss: {'train/loss': 0.44243884086608887, 'train/grad_norm_pre_clip': 6.42047643661499, 'timing/data': 0.016959020867943764, 'timing/model': 0.9371568337082863, 'learning_rate/qwen_vl_interface': 9.671988844563652e-06, 'learning_rate/action_model': 9.671988844563652e-05, 'epoch': 0.42}) +[2026-06-17 18:12:44,801][starVLA.training.train_starvla][INFO] - Step 922, Loss: {'train/loss': 0.8486599326133728, 'train/grad_norm_pre_clip': 8.656761169433594, 'timing/data': 0.014878200367093086, 'timing/model': 1.0919284652918577, 'learning_rate/qwen_vl_interface': 9.671198659501233e-06, 'learning_rate/action_model': 9.671198659501231e-05, 'epoch': 0.42}) +[2026-06-17 18:12:48,686][starVLA.training.train_starvla][INFO] - Step 923, Loss: {'train/loss': 0.5900742411613464, 'train/grad_norm_pre_clip': 7.567383289337158, 'timing/data': 0.01907426305115223, 'timing/model': 1.1929585048928857, 'learning_rate/qwen_vl_interface': 9.670407557919573e-06, 'learning_rate/action_model': 9.670407557919572e-05, 'epoch': 0.42}) +[2026-06-17 18:12:52,445][starVLA.training.train_starvla][INFO] - Step 924, Loss: {'train/loss': 0.6906646490097046, 'train/grad_norm_pre_clip': 6.8424973487854, 'timing/data': 0.019038246013224125, 'timing/model': 0.9262015903368592, 'learning_rate/qwen_vl_interface': 9.669615539982668e-06, 'learning_rate/action_model': 9.669615539982668e-05, 'epoch': 0.42}) +[2026-06-17 18:12:56,250][starVLA.training.train_starvla][INFO] - Step 925, Loss: {'train/loss': 0.6673170328140259, 'train/grad_norm_pre_clip': 6.629215717315674, 'timing/data': 0.019289951771497726, 'timing/model': 0.9407828832045197, 'learning_rate/qwen_vl_interface': 9.668822605854708e-06, 'learning_rate/action_model': 9.668822605854706e-05, 'epoch': 0.42}) +[2026-06-17 18:12:59,938][starVLA.training.train_starvla][INFO] - Step 926, Loss: {'train/loss': 0.5273532867431641, 'train/grad_norm_pre_clip': 6.112695693969727, 'timing/data': 0.017287607304751873, 'timing/model': 0.9221503930166364, 'learning_rate/qwen_vl_interface': 9.668028755700067e-06, 'learning_rate/action_model': 9.668028755700067e-05, 'epoch': 0.42}) +[2026-06-17 18:13:04,262][starVLA.training.train_starvla][INFO] - Step 927, Loss: {'train/loss': 0.5175095200538635, 'train/grad_norm_pre_clip': 7.405338764190674, 'timing/data': 0.018748891539871693, 'timing/model': 0.9363911803811789, 'learning_rate/qwen_vl_interface': 9.66723398968331e-06, 'learning_rate/action_model': 9.66723398968331e-05, 'epoch': 0.42}) +[2026-06-17 18:13:08,229][starVLA.training.train_starvla][INFO] - Step 928, Loss: {'train/loss': 0.651284396648407, 'train/grad_norm_pre_clip': 8.375910758972168, 'timing/data': 0.018885428085923195, 'timing/model': 1.1582592390477657, 'learning_rate/qwen_vl_interface': 9.666438307969194e-06, 'learning_rate/action_model': 9.666438307969195e-05, 'epoch': 0.42}) +[2026-06-17 18:13:11,848][starVLA.training.train_starvla][INFO] - Step 929, Loss: {'train/loss': 0.5953468084335327, 'train/grad_norm_pre_clip': 7.352015018463135, 'timing/data': 0.01741806138306856, 'timing/model': 0.9329419396817684, 'learning_rate/qwen_vl_interface': 9.665641710722666e-06, 'learning_rate/action_model': 9.665641710722665e-05, 'epoch': 0.42}) +[2026-06-17 18:13:15,723][starVLA.training.train_starvla][INFO] - Step 930, Loss: {'train/loss': 0.46965134143829346, 'train/grad_norm_pre_clip': 6.434617042541504, 'timing/data': 0.016443362459540367, 'timing/model': 1.0101449005305767, 'learning_rate/qwen_vl_interface': 9.664844198108858e-06, 'learning_rate/action_model': 9.664844198108858e-05, 'epoch': 0.42}) +[2026-06-17 18:13:19,537][starVLA.training.train_starvla][INFO] - Step 931, Loss: {'train/loss': 0.6003779768943787, 'train/grad_norm_pre_clip': 8.378975868225098, 'timing/data': 0.01796029880642891, 'timing/model': 0.9274844275787473, 'learning_rate/qwen_vl_interface': 9.6640457702931e-06, 'learning_rate/action_model': 9.664045770293099e-05, 'epoch': 0.42}) +[2026-06-17 18:13:23,389][starVLA.training.train_starvla][INFO] - Step 932, Loss: {'train/loss': 0.5515559911727905, 'train/grad_norm_pre_clip': 6.512157440185547, 'timing/data': 0.016525221057236195, 'timing/model': 0.9414752339944243, 'learning_rate/qwen_vl_interface': 9.663246427440901e-06, 'learning_rate/action_model': 9.6632464274409e-05, 'epoch': 0.42}) +[2026-06-17 18:13:27,195][starVLA.training.train_starvla][INFO] - Step 933, Loss: {'train/loss': 0.8038106560707092, 'train/grad_norm_pre_clip': 10.580986022949219, 'timing/data': 0.01919927541166544, 'timing/model': 0.9692581808194518, 'learning_rate/qwen_vl_interface': 9.662446169717969e-06, 'learning_rate/action_model': 9.662446169717969e-05, 'epoch': 0.42}) +[2026-06-17 18:13:31,203][starVLA.training.train_starvla][INFO] - Step 934, Loss: {'train/loss': 0.4990372657775879, 'train/grad_norm_pre_clip': 6.423547267913818, 'timing/data': 0.02222712431102991, 'timing/model': 0.9854276282712817, 'learning_rate/qwen_vl_interface': 9.6616449972902e-06, 'learning_rate/action_model': 9.6616449972902e-05, 'epoch': 0.42}) +[2026-06-17 18:13:35,037][starVLA.training.train_starvla][INFO] - Step 935, Loss: {'train/loss': 0.5454661250114441, 'train/grad_norm_pre_clip': 7.0780510902404785, 'timing/data': 0.01906762830913067, 'timing/model': 1.151416408829391, 'learning_rate/qwen_vl_interface': 9.660842910323675e-06, 'learning_rate/action_model': 9.660842910323674e-05, 'epoch': 0.42}) +[2026-06-17 18:13:38,638][starVLA.training.train_starvla][INFO] - Step 936, Loss: {'train/loss': 0.6849232316017151, 'train/grad_norm_pre_clip': 10.858199119567871, 'timing/data': 0.01747030671685934, 'timing/model': 0.9346114257350564, 'learning_rate/qwen_vl_interface': 9.660039908984667e-06, 'learning_rate/action_model': 9.660039908984666e-05, 'epoch': 0.42}) +[2026-06-17 18:13:42,423][starVLA.training.train_starvla][INFO] - Step 937, Loss: {'train/loss': 0.7142319083213806, 'train/grad_norm_pre_clip': 12.050726890563965, 'timing/data': 0.016133882105350494, 'timing/model': 0.9316211296245456, 'learning_rate/qwen_vl_interface': 9.659235993439643e-06, 'learning_rate/action_model': 9.659235993439643e-05, 'epoch': 0.43}) +[2026-06-17 18:13:46,237][starVLA.training.train_starvla][INFO] - Step 938, Loss: {'train/loss': 0.6898485422134399, 'train/grad_norm_pre_clip': 9.016413688659668, 'timing/data': 0.01480826735496521, 'timing/model': 0.9294561352580786, 'learning_rate/qwen_vl_interface': 9.658431163855251e-06, 'learning_rate/action_model': 9.658431163855251e-05, 'epoch': 0.43}) +[2026-06-17 18:13:50,073][starVLA.training.train_starvla][INFO] - Step 939, Loss: {'train/loss': 0.7449737787246704, 'train/grad_norm_pre_clip': 10.749085426330566, 'timing/data': 0.014945554547011852, 'timing/model': 0.9897337779402733, 'learning_rate/qwen_vl_interface': 9.657625420398334e-06, 'learning_rate/action_model': 9.657625420398333e-05, 'epoch': 0.43}) +[2026-06-17 18:13:53,974][starVLA.training.train_starvla][INFO] - Step 940, Loss: {'train/loss': 0.820034921169281, 'train/grad_norm_pre_clip': 8.603322982788086, 'timing/data': 0.020009859465062618, 'timing/model': 0.9341157386079431, 'learning_rate/qwen_vl_interface': 9.656818763235927e-06, 'learning_rate/action_model': 9.656818763235926e-05, 'epoch': 0.43}) +[2026-06-17 18:13:57,844][starVLA.training.train_starvla][INFO] - Step 941, Loss: {'train/loss': 0.49961942434310913, 'train/grad_norm_pre_clip': 9.127460479736328, 'timing/data': 0.0164102790877223, 'timing/model': 1.1516734259203076, 'learning_rate/qwen_vl_interface': 9.656011192535248e-06, 'learning_rate/action_model': 9.656011192535247e-05, 'epoch': 0.43}) +[2026-06-17 18:14:01,695][starVLA.training.train_starvla][INFO] - Step 942, Loss: {'train/loss': 0.785970151424408, 'train/grad_norm_pre_clip': 12.239809036254883, 'timing/data': 0.017283178865909576, 'timing/model': 1.1874976940453053, 'learning_rate/qwen_vl_interface': 9.655202708463706e-06, 'learning_rate/action_model': 9.655202708463707e-05, 'epoch': 0.43}) +[2026-06-17 18:14:05,329][starVLA.training.train_starvla][INFO] - Step 943, Loss: {'train/loss': 0.6140856742858887, 'train/grad_norm_pre_clip': 7.112648010253906, 'timing/data': 0.017748608253896236, 'timing/model': 0.9349543731659651, 'learning_rate/qwen_vl_interface': 9.654393311188905e-06, 'learning_rate/action_model': 9.654393311188905e-05, 'epoch': 0.43}) +[2026-06-17 18:14:09,234][starVLA.training.train_starvla][INFO] - Step 944, Loss: {'train/loss': 0.9003284573554993, 'train/grad_norm_pre_clip': 11.792802810668945, 'timing/data': 0.018487948924303055, 'timing/model': 0.9413625728338957, 'learning_rate/qwen_vl_interface': 9.653583000878631e-06, 'learning_rate/action_model': 9.65358300087863e-05, 'epoch': 0.43}) +[2026-06-17 18:14:13,043][starVLA.training.train_starvla][INFO] - Step 945, Loss: {'train/loss': 0.6095222234725952, 'train/grad_norm_pre_clip': 6.848658561706543, 'timing/data': 0.026751278899610043, 'timing/model': 0.931904380209744, 'learning_rate/qwen_vl_interface': 9.65277177770086e-06, 'learning_rate/action_model': 9.65277177770086e-05, 'epoch': 0.43}) +[2026-06-17 18:14:16,965][starVLA.training.train_starvla][INFO] - Step 946, Loss: {'train/loss': 0.7590174674987793, 'train/grad_norm_pre_clip': 9.27921199798584, 'timing/data': 0.017555068247020245, 'timing/model': 0.9789844993501902, 'learning_rate/qwen_vl_interface': 9.651959641823764e-06, 'learning_rate/action_model': 9.651959641823763e-05, 'epoch': 0.43}) +[2026-06-17 18:14:20,853][starVLA.training.train_starvla][INFO] - Step 947, Loss: {'train/loss': 0.6074627041816711, 'train/grad_norm_pre_clip': 10.55531120300293, 'timing/data': 0.01600811257958412, 'timing/model': 0.9270975328981876, 'learning_rate/qwen_vl_interface': 9.651146593415696e-06, 'learning_rate/action_model': 9.651146593415695e-05, 'epoch': 0.43}) +[2026-06-17 18:14:24,744][starVLA.training.train_starvla][INFO] - Step 948, Loss: {'train/loss': 0.6668967008590698, 'train/grad_norm_pre_clip': 11.493094444274902, 'timing/data': 0.015275681391358376, 'timing/model': 1.120818736962974, 'learning_rate/qwen_vl_interface': 9.650332632645206e-06, 'learning_rate/action_model': 9.650332632645205e-05, 'epoch': 0.43}) +[2026-06-17 18:14:28,530][starVLA.training.train_starvla][INFO] - Step 949, Loss: {'train/loss': 0.7137200236320496, 'train/grad_norm_pre_clip': 8.901779174804688, 'timing/data': 0.017173903062939644, 'timing/model': 1.0827939854934812, 'learning_rate/qwen_vl_interface': 9.649517759681022e-06, 'learning_rate/action_model': 9.649517759681023e-05, 'epoch': 0.43}) +[2026-06-17 18:14:32,137][starVLA.training.train_starvla][INFO] - Step 950, Loss: {'train/loss': 0.3926110565662384, 'train/grad_norm_pre_clip': 7.893772602081299, 'timing/data': 0.014620277099311352, 'timing/model': 0.9313724050298333, 'learning_rate/qwen_vl_interface': 9.648701974692075e-06, 'learning_rate/action_model': 9.648701974692075e-05, 'epoch': 0.43}) +[2026-06-17 18:14:35,954][starVLA.training.train_starvla][INFO] - Step 951, Loss: {'train/loss': 0.9148054122924805, 'train/grad_norm_pre_clip': 17.95453453063965, 'timing/data': 0.014435709454119205, 'timing/model': 0.9714341908693314, 'learning_rate/qwen_vl_interface': 9.647885277847476e-06, 'learning_rate/action_model': 9.647885277847476e-05, 'epoch': 0.43}) +[2026-06-17 18:14:40,160][starVLA.training.train_starvla][INFO] - Step 952, Loss: {'train/loss': 0.809696614742279, 'train/grad_norm_pre_clip': 10.380236625671387, 'timing/data': 0.014625578187406063, 'timing/model': 1.3531912015751004, 'learning_rate/qwen_vl_interface': 9.647067669316525e-06, 'learning_rate/action_model': 9.647067669316525e-05, 'epoch': 0.43}) +[2026-06-17 18:14:44,626][starVLA.training.train_starvla][INFO] - Step 953, Loss: {'train/loss': 0.3892554044723511, 'train/grad_norm_pre_clip': 7.197964191436768, 'timing/data': 0.06431842967867851, 'timing/model': 0.9235784122720361, 'learning_rate/qwen_vl_interface': 9.646249149268716e-06, 'learning_rate/action_model': 9.646249149268715e-05, 'epoch': 0.43}) +[2026-06-17 18:14:48,706][starVLA.training.train_starvla][INFO] - Step 954, Loss: {'train/loss': 0.5622983574867249, 'train/grad_norm_pre_clip': 9.849225997924805, 'timing/data': 0.016538359224796295, 'timing/model': 1.0259996494278312, 'learning_rate/qwen_vl_interface': 9.645429717873728e-06, 'learning_rate/action_model': 9.645429717873727e-05, 'epoch': 0.43}) +[2026-06-17 18:14:53,234][starVLA.training.train_starvla][INFO] - Step 955, Loss: {'train/loss': 0.314299076795578, 'train/grad_norm_pre_clip': 10.450840950012207, 'timing/data': 0.011056927964091301, 'timing/model': 1.3092932477593422, 'learning_rate/qwen_vl_interface': 9.644609375301429e-06, 'learning_rate/action_model': 9.64460937530143e-05, 'epoch': 0.43}) +[2026-06-17 18:14:56,934][starVLA.training.train_starvla][INFO] - Step 956, Loss: {'train/loss': 0.8524511456489563, 'train/grad_norm_pre_clip': 11.135619163513184, 'timing/data': 0.01075028907507658, 'timing/model': 0.9343320271000266, 'learning_rate/qwen_vl_interface': 9.643788121721877e-06, 'learning_rate/action_model': 9.643788121721876e-05, 'epoch': 0.43}) +[2026-06-17 18:15:00,825][starVLA.training.train_starvla][INFO] - Step 957, Loss: {'train/loss': 0.6688921451568604, 'train/grad_norm_pre_clip': 6.916136741638184, 'timing/data': 0.011336209252476692, 'timing/model': 0.9271145649254322, 'learning_rate/qwen_vl_interface': 9.64296595730532e-06, 'learning_rate/action_model': 9.64296595730532e-05, 'epoch': 0.43}) +[2026-06-17 18:15:04,654][starVLA.training.train_starvla][INFO] - Step 958, Loss: {'train/loss': 0.7605348229408264, 'train/grad_norm_pre_clip': 10.057212829589844, 'timing/data': 0.010386044159531593, 'timing/model': 0.9516928987577558, 'learning_rate/qwen_vl_interface': 9.642142882222193e-06, 'learning_rate/action_model': 9.642142882222193e-05, 'epoch': 0.43}) +[2026-06-17 18:15:08,561][starVLA.training.train_starvla][INFO] - Step 959, Loss: {'train/loss': 0.5194166898727417, 'train/grad_norm_pre_clip': 5.15210485458374, 'timing/data': 0.009461537934839725, 'timing/model': 0.9397849412634969, 'learning_rate/qwen_vl_interface': 9.641318896643123e-06, 'learning_rate/action_model': 9.641318896643122e-05, 'epoch': 0.44}) +[2026-06-17 18:15:15,900][starVLA.training.train_starvla][INFO] - Step 960, Loss: {'train/loss': 0.6605274677276611, 'train/grad_norm_pre_clip': 6.634211540222168, 'timing/data': 0.009299023076891899, 'timing/model': 0.9312041206285357, 'learning_rate/qwen_vl_interface': 9.640494000738917e-06, 'learning_rate/action_model': 9.640494000738917e-05, 'epoch': 0.44}) +[2026-06-17 18:15:20,183][starVLA.training.train_starvla][INFO] - Step 961, Loss: {'train/loss': 0.5452788472175598, 'train/grad_norm_pre_clip': 7.25220251083374, 'timing/data': 0.009227776899933815, 'timing/model': 0.9495822740718722, 'learning_rate/qwen_vl_interface': 9.63966819468058e-06, 'learning_rate/action_model': 9.63966819468058e-05, 'epoch': 0.44}) +[2026-06-17 18:15:24,074][starVLA.training.train_starvla][INFO] - Step 962, Loss: {'train/loss': 0.6746368408203125, 'train/grad_norm_pre_clip': 18.33432388305664, 'timing/data': 0.009740598499774933, 'timing/model': 1.1145410854369402, 'learning_rate/qwen_vl_interface': 9.638841478639306e-06, 'learning_rate/action_model': 9.638841478639306e-05, 'epoch': 0.44}) +[2026-06-17 18:15:27,837][starVLA.training.train_starvla][INFO] - Step 963, Loss: {'train/loss': 0.5394930839538574, 'train/grad_norm_pre_clip': 7.168151378631592, 'timing/data': 0.008847295306622982, 'timing/model': 0.9255999512970448, 'learning_rate/qwen_vl_interface': 9.638013852786468e-06, 'learning_rate/action_model': 9.638013852786467e-05, 'epoch': 0.44}) +[2026-06-17 18:15:31,630][starVLA.training.train_starvla][INFO] - Step 964, Loss: {'train/loss': 0.8604101538658142, 'train/grad_norm_pre_clip': 8.992447853088379, 'timing/data': 0.010010289028286934, 'timing/model': 0.9454473694786429, 'learning_rate/qwen_vl_interface': 9.637185317293638e-06, 'learning_rate/action_model': 9.637185317293638e-05, 'epoch': 0.44}) +[2026-06-17 18:15:35,406][starVLA.training.train_starvla][INFO] - Step 965, Loss: {'train/loss': 0.5754542350769043, 'train/grad_norm_pre_clip': 5.980429649353027, 'timing/data': 0.009137826040387154, 'timing/model': 0.9265712993219495, 'learning_rate/qwen_vl_interface': 9.63635587233257e-06, 'learning_rate/action_model': 9.63635587233257e-05, 'epoch': 0.44}) +[2026-06-17 18:15:39,260][starVLA.training.train_starvla][INFO] - Step 966, Loss: {'train/loss': 1.0311334133148193, 'train/grad_norm_pre_clip': 7.700397491455078, 'timing/data': 0.008990787900984287, 'timing/model': 0.9332826845347881, 'learning_rate/qwen_vl_interface': 9.63552551807521e-06, 'learning_rate/action_model': 9.63552551807521e-05, 'epoch': 0.44}) +[2026-06-17 18:15:43,088][starVLA.training.train_starvla][INFO] - Step 967, Loss: {'train/loss': 0.5817481279373169, 'train/grad_norm_pre_clip': 10.238496780395508, 'timing/data': 0.007425754331052303, 'timing/model': 0.9353994186967611, 'learning_rate/qwen_vl_interface': 9.634694254693691e-06, 'learning_rate/action_model': 9.634694254693692e-05, 'epoch': 0.44}) +[2026-06-17 18:15:46,952][starVLA.training.train_starvla][INFO] - Step 968, Loss: {'train/loss': 0.3797428607940674, 'train/grad_norm_pre_clip': 7.596118450164795, 'timing/data': 0.0068960040807724, 'timing/model': 0.9385036053135991, 'learning_rate/qwen_vl_interface': 9.633862082360336e-06, 'learning_rate/action_model': 9.633862082360335e-05, 'epoch': 0.44}) +[2026-06-17 18:15:50,726][starVLA.training.train_starvla][INFO] - Step 969, Loss: {'train/loss': 0.5476930141448975, 'train/grad_norm_pre_clip': 7.141269207000732, 'timing/data': 0.011688889935612679, 'timing/model': 1.1162433167919517, 'learning_rate/qwen_vl_interface': 9.633029001247654e-06, 'learning_rate/action_model': 9.633029001247653e-05, 'epoch': 0.44}) +[2026-06-17 18:15:54,310][starVLA.training.train_starvla][INFO] - Step 970, Loss: {'train/loss': 0.5638940334320068, 'train/grad_norm_pre_clip': 7.797810077667236, 'timing/data': 0.0066378116607666016, 'timing/model': 0.929416835308075, 'learning_rate/qwen_vl_interface': 9.63219501152834e-06, 'learning_rate/action_model': 9.632195011528341e-05, 'epoch': 0.44}) +[2026-06-17 18:15:58,266][starVLA.training.train_starvla][INFO] - Step 971, Loss: {'train/loss': 0.4392535984516144, 'train/grad_norm_pre_clip': 7.310395240783691, 'timing/data': 0.010832278057932854, 'timing/model': 0.9601743128150702, 'learning_rate/qwen_vl_interface': 9.63136011337529e-06, 'learning_rate/action_model': 9.63136011337529e-05, 'epoch': 0.44}) +[2026-06-17 18:16:02,923][starVLA.training.train_starvla][INFO] - Step 972, Loss: {'train/loss': 0.693986177444458, 'train/grad_norm_pre_clip': 5.330872058868408, 'timing/data': 0.008663312532007694, 'timing/model': 0.9300291957333684, 'learning_rate/qwen_vl_interface': 9.63052430696157e-06, 'learning_rate/action_model': 9.63052430696157e-05, 'epoch': 0.44}) +[2026-06-17 18:16:06,749][starVLA.training.train_starvla][INFO] - Step 973, Loss: {'train/loss': 0.7382805347442627, 'train/grad_norm_pre_clip': 8.667280197143555, 'timing/data': 0.009900754317641258, 'timing/model': 0.936150255613029, 'learning_rate/qwen_vl_interface': 9.629687592460449e-06, 'learning_rate/action_model': 9.62968759246045e-05, 'epoch': 0.44}) +[2026-06-17 18:16:10,721][starVLA.training.train_starvla][INFO] - Step 974, Loss: {'train/loss': 0.5016738772392273, 'train/grad_norm_pre_clip': 8.140827178955078, 'timing/data': 0.008474036119878292, 'timing/model': 1.0394212268292904, 'learning_rate/qwen_vl_interface': 9.628849970045378e-06, 'learning_rate/action_model': 9.628849970045378e-05, 'epoch': 0.44}) +[2026-06-17 18:16:14,562][starVLA.training.train_starvla][INFO] - Step 975, Loss: {'train/loss': 0.6619216799736023, 'train/grad_norm_pre_clip': 7.897912502288818, 'timing/data': 0.013346545398235321, 'timing/model': 0.9397982722148299, 'learning_rate/qwen_vl_interface': 9.628011439889997e-06, 'learning_rate/action_model': 9.628011439889996e-05, 'epoch': 0.44}) +[2026-06-17 18:16:18,727][starVLA.training.train_starvla][INFO] - Step 976, Loss: {'train/loss': 0.5814478993415833, 'train/grad_norm_pre_clip': 8.700553894042969, 'timing/data': 0.008359243161976337, 'timing/model': 1.1352857379242778, 'learning_rate/qwen_vl_interface': 9.62717200216813e-06, 'learning_rate/action_model': 9.62717200216813e-05, 'epoch': 0.44}) +[2026-06-17 18:16:22,358][starVLA.training.train_starvla][INFO] - Step 977, Loss: {'train/loss': 0.6582270264625549, 'train/grad_norm_pre_clip': 11.811670303344727, 'timing/data': 0.00845563504844904, 'timing/model': 0.9422887666150928, 'learning_rate/qwen_vl_interface': 9.6263316570538e-06, 'learning_rate/action_model': 9.6263316570538e-05, 'epoch': 0.44}) +[2026-06-17 18:16:26,265][starVLA.training.train_starvla][INFO] - Step 978, Loss: {'train/loss': 0.30895981192588806, 'train/grad_norm_pre_clip': 6.119709491729736, 'timing/data': 0.008816743269562721, 'timing/model': 0.9309883061796427, 'learning_rate/qwen_vl_interface': 9.625490404721208e-06, 'learning_rate/action_model': 9.625490404721208e-05, 'epoch': 0.44}) +[2026-06-17 18:16:30,097][starVLA.training.train_starvla][INFO] - Step 979, Loss: {'train/loss': 0.5757326483726501, 'train/grad_norm_pre_clip': 7.445973873138428, 'timing/data': 0.010588816367089748, 'timing/model': 0.919792509637773, 'learning_rate/qwen_vl_interface': 9.624648245344747e-06, 'learning_rate/action_model': 9.624648245344746e-05, 'epoch': 0.44}) +[2026-06-17 18:16:33,953][starVLA.training.train_starvla][INFO] - Step 980, Loss: {'train/loss': 0.6285503506660461, 'train/grad_norm_pre_clip': 7.64484167098999, 'timing/data': 0.012371450662612915, 'timing/model': 0.9383951863273978, 'learning_rate/qwen_vl_interface': 9.623805179098996e-06, 'learning_rate/action_model': 9.623805179098996e-05, 'epoch': 0.44}) +[2026-06-17 18:16:37,813][starVLA.training.train_starvla][INFO] - Step 981, Loss: {'train/loss': 0.7144647836685181, 'train/grad_norm_pre_clip': 13.131680488586426, 'timing/data': 0.008809790946543217, 'timing/model': 0.956586230546236, 'learning_rate/qwen_vl_interface': 9.622961206158726e-06, 'learning_rate/action_model': 9.622961206158726e-05, 'epoch': 0.45}) +[2026-06-17 18:16:41,827][starVLA.training.train_starvla][INFO] - Step 982, Loss: {'train/loss': 1.1738425493240356, 'train/grad_norm_pre_clip': 11.761366844177246, 'timing/data': 0.009058710187673569, 'timing/model': 1.1764617143198848, 'learning_rate/qwen_vl_interface': 9.622116326698894e-06, 'learning_rate/action_model': 9.622116326698893e-05, 'epoch': 0.45}) +[2026-06-17 18:16:45,832][starVLA.training.train_starvla][INFO] - Step 983, Loss: {'train/loss': 0.5543969869613647, 'train/grad_norm_pre_clip': 10.132704734802246, 'timing/data': 0.008220924064517021, 'timing/model': 1.0890585351735353, 'learning_rate/qwen_vl_interface': 9.621270540894639e-06, 'learning_rate/action_model': 9.621270540894639e-05, 'epoch': 0.45}) +[2026-06-17 18:16:49,454][starVLA.training.train_starvla][INFO] - Step 984, Loss: {'train/loss': 0.8583669662475586, 'train/grad_norm_pre_clip': 10.28867244720459, 'timing/data': 0.011584372259676456, 'timing/model': 0.9617811162024736, 'learning_rate/qwen_vl_interface': 9.6204238489213e-06, 'learning_rate/action_model': 9.6204238489213e-05, 'epoch': 0.45}) +[2026-06-17 18:16:53,259][starVLA.training.train_starvla][INFO] - Step 985, Loss: {'train/loss': 0.6303651332855225, 'train/grad_norm_pre_clip': 11.337493896484375, 'timing/data': 0.012132493779063225, 'timing/model': 0.9366273256018758, 'learning_rate/qwen_vl_interface': 9.619576250954393e-06, 'learning_rate/action_model': 9.619576250954393e-05, 'epoch': 0.45}) +[2026-06-17 18:16:57,027][starVLA.training.train_starvla][INFO] - Step 986, Loss: {'train/loss': 0.7580450177192688, 'train/grad_norm_pre_clip': 10.357295989990234, 'timing/data': 0.010108367539942265, 'timing/model': 0.9331122171133757, 'learning_rate/qwen_vl_interface': 9.618727747169628e-06, 'learning_rate/action_model': 9.618727747169627e-05, 'epoch': 0.45}) +[2026-06-17 18:17:01,052][starVLA.training.train_starvla][INFO] - Step 987, Loss: {'train/loss': 0.5839371085166931, 'train/grad_norm_pre_clip': 12.79024887084961, 'timing/data': 0.010895654559135437, 'timing/model': 0.9304524976760149, 'learning_rate/qwen_vl_interface': 9.617878337742902e-06, 'learning_rate/action_model': 9.617878337742902e-05, 'epoch': 0.45}) +[2026-06-17 18:17:04,869][starVLA.training.train_starvla][INFO] - Step 988, Loss: {'train/loss': 0.8477616310119629, 'train/grad_norm_pre_clip': 14.53891372680664, 'timing/data': 0.008444525301456451, 'timing/model': 0.9411174822598696, 'learning_rate/qwen_vl_interface': 9.617028022850294e-06, 'learning_rate/action_model': 9.617028022850293e-05, 'epoch': 0.45}) +[2026-06-17 18:17:11,537][starVLA.training.train_starvla][INFO] - Step 989, Loss: {'train/loss': 0.6294940114021301, 'train/grad_norm_pre_clip': 8.196122169494629, 'timing/data': 0.010690158233046532, 'timing/model': 1.4589165337383747, 'learning_rate/qwen_vl_interface': 9.616176802668078e-06, 'learning_rate/action_model': 9.616176802668078e-05, 'epoch': 0.45}) +[2026-06-17 18:17:16,472][starVLA.training.train_starvla][INFO] - Step 990, Loss: {'train/loss': 0.907326877117157, 'train/grad_norm_pre_clip': 13.938570976257324, 'timing/data': 0.008804365992546082, 'timing/model': 1.132856173440814, 'learning_rate/qwen_vl_interface': 9.615324677372712e-06, 'learning_rate/action_model': 9.615324677372713e-05, 'epoch': 0.45}) +[2026-06-17 18:17:20,143][starVLA.training.train_starvla][INFO] - Step 991, Loss: {'train/loss': 0.5904631614685059, 'train/grad_norm_pre_clip': 8.79735279083252, 'timing/data': 0.013172630220651627, 'timing/model': 0.9557731095701456, 'learning_rate/qwen_vl_interface': 9.614471647140844e-06, 'learning_rate/action_model': 9.614471647140844e-05, 'epoch': 0.45}) +[2026-06-17 18:17:24,033][starVLA.training.train_starvla][INFO] - Step 992, Loss: {'train/loss': 0.7564828395843506, 'train/grad_norm_pre_clip': 10.956961631774902, 'timing/data': 0.010354451835155487, 'timing/model': 0.9474958516657352, 'learning_rate/qwen_vl_interface': 9.613617712149309e-06, 'learning_rate/action_model': 9.613617712149308e-05, 'epoch': 0.45}) +[2026-06-17 18:17:27,925][starVLA.training.train_starvla][INFO] - Step 993, Loss: {'train/loss': 0.4747808575630188, 'train/grad_norm_pre_clip': 11.267008781433105, 'timing/data': 0.007621581666171551, 'timing/model': 0.9459832115098834, 'learning_rate/qwen_vl_interface': 9.612762872575124e-06, 'learning_rate/action_model': 9.612762872575124e-05, 'epoch': 0.45}) +[2026-06-17 18:17:31,748][starVLA.training.train_starvla][INFO] - Step 994, Loss: {'train/loss': 0.4679788053035736, 'train/grad_norm_pre_clip': 8.532275199890137, 'timing/data': 0.01123211719095707, 'timing/model': 0.9658319875597954, 'learning_rate/qwen_vl_interface': 9.611907128595503e-06, 'learning_rate/action_model': 9.611907128595502e-05, 'epoch': 0.45}) +[2026-06-17 18:17:35,795][starVLA.training.train_starvla][INFO] - Step 995, Loss: {'train/loss': 0.9952800869941711, 'train/grad_norm_pre_clip': 11.501570701599121, 'timing/data': 0.011425716802477837, 'timing/model': 0.9523580614477396, 'learning_rate/qwen_vl_interface': 9.61105048038784e-06, 'learning_rate/action_model': 9.611050480387839e-05, 'epoch': 0.45}) +[2026-06-17 18:17:40,121][starVLA.training.train_starvla][INFO] - Step 996, Loss: {'train/loss': 0.8569221496582031, 'train/grad_norm_pre_clip': 9.398528099060059, 'timing/data': 0.008987090550363064, 'timing/model': 1.1263734493404627, 'learning_rate/qwen_vl_interface': 9.610192928129716e-06, 'learning_rate/action_model': 9.610192928129717e-05, 'epoch': 0.45}) +[2026-06-17 18:17:43,730][starVLA.training.train_starvla][INFO] - Step 997, Loss: {'train/loss': 0.45318764448165894, 'train/grad_norm_pre_clip': 6.082344055175781, 'timing/data': 0.009218203835189342, 'timing/model': 0.9464103197678924, 'learning_rate/qwen_vl_interface': 9.609334471998911e-06, 'learning_rate/action_model': 9.609334471998912e-05, 'epoch': 0.45}) +[2026-06-17 18:17:47,570][starVLA.training.train_starvla][INFO] - Step 998, Loss: {'train/loss': 0.7584030628204346, 'train/grad_norm_pre_clip': 9.436923027038574, 'timing/data': 0.008780443109571934, 'timing/model': 0.9409324051812291, 'learning_rate/qwen_vl_interface': 9.608475112173379e-06, 'learning_rate/action_model': 9.608475112173379e-05, 'epoch': 0.45}) +[2026-06-17 18:17:51,315][starVLA.training.train_starvla][INFO] - Step 999, Loss: {'train/loss': 1.051980972290039, 'train/grad_norm_pre_clip': 8.664022445678711, 'timing/data': 0.011095857247710228, 'timing/model': 0.9328373447060585, 'learning_rate/qwen_vl_interface': 9.607614848831264e-06, 'learning_rate/action_model': 9.607614848831264e-05, 'epoch': 0.45}) +[2026-06-17 18:19:57,418][starVLA.training.train_starvla][INFO] - Step 1000, Loss: {'train/loss': 0.880964994430542, 'train/grad_norm_pre_clip': 11.238966941833496, 'eval/loss': 0.5911813378334045, 'eval/latency_8/loss': 0.5911813378334045, 'eval/demon_fire_precision': 0.9069833191970597, 'eval/demon_fire_recall': 0.9965827896862379, 'eval/demon_fire_f1': 0.9496743635287151, 'eval/demon_fire_rate_model': 0.9841402337228714, 'eval/demon_fire_rate_teacher': 0.8956594323873122, 'eval/demon_left_precision': 0.7427652733118971, 'eval/demon_left_recall': 0.49464668094218417, 'eval/demon_left_f1': 0.5938303341902313, 'eval/demon_left_rate_model': 0.08653311074012243, 'eval/demon_left_rate_teacher': 0.12993878686700056, 'eval/demon_right_precision': 0.8218527315914489, 'eval/demon_right_recall': 0.5242424242424243, 'eval/demon_right_f1': 0.6401480111008325, 'eval/demon_right_rate_model': 0.11713967723984418, 'eval/demon_right_rate_teacher': 0.18363939899833054, 'eval/demon_move_f1': 0.6169891726455319, 'eval/demon_cc_f1': 0.8174429517770394, 'eval/latency_8/demon_fire_precision': 0.9069833191970597, 'eval/latency_8/demon_fire_recall': 0.9965827896862379, 'eval/latency_8/demon_fire_f1': 0.9496743635287151, 'eval/latency_8/demon_fire_rate_model': 0.9841402337228714, 'eval/latency_8/demon_fire_rate_teacher': 0.8956594323873122, 'eval/latency_8/demon_left_precision': 0.7427652733118971, 'eval/latency_8/demon_left_recall': 0.49464668094218417, 'eval/latency_8/demon_left_f1': 0.5938303341902313, 'eval/latency_8/demon_left_rate_model': 0.08653311074012243, 'eval/latency_8/demon_left_rate_teacher': 0.12993878686700056, 'eval/latency_8/demon_right_precision': 0.8218527315914489, 'eval/latency_8/demon_right_recall': 0.5242424242424243, 'eval/latency_8/demon_right_f1': 0.6401480111008325, 'eval/latency_8/demon_right_rate_model': 0.11713967723984418, 'eval/latency_8/demon_right_rate_teacher': 0.18363939899833054, 'eval/latency_8/demon_move_f1': 0.6169891726455319, 'eval/latency_8/demon_cc_f1': 0.8174429517770394, 'timing/data': 0.010836305096745491, 'timing/model': 1.139152755960822, 'learning_rate/qwen_vl_interface': 9.606753682150904e-06, 'learning_rate/action_model': 9.606753682150903e-05, 'epoch': 0.45}) +[2026-06-17 18:19:57,422][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_1000_state +[2026-06-17 18:19:57,427][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 18:20:04,121][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_1000_state/model.safetensors +[2026-06-17 18:20:20,455][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_1000_state/optimizer.bin +[2026-06-17 18:20:20,456][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_1000_state/sampler.bin +[2026-06-17 18:20:20,456][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_1000_state/sampler_1.bin +[2026-06-17 18:20:20,458][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_1000_state/random_states_0.pkl +[2026-06-17 18:20:20,466][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1000 +[2026-06-17 18:20:20,466][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 18:20:20,477][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 18:20:26,724][starVLA.training.train_starvla][INFO] - Step 1001, Loss: {'train/loss': 0.7475582361221313, 'train/grad_norm_pre_clip': 8.88976001739502, 'timing/data': 0.009679845534265041, 'timing/model': 1.7340631112456322, 'learning_rate/qwen_vl_interface': 9.605891612310815e-06, 'learning_rate/action_model': 9.605891612310814e-05, 'epoch': 0.45}) +[2026-06-17 18:20:30,648][starVLA.training.train_starvla][INFO] - Step 1002, Loss: {'train/loss': 0.8765884637832642, 'train/grad_norm_pre_clip': 8.668197631835938, 'timing/data': 0.01027267798781395, 'timing/model': 0.9407614329829812, 'learning_rate/qwen_vl_interface': 9.60502863948971e-06, 'learning_rate/action_model': 9.60502863948971e-05, 'epoch': 0.45}) +[2026-06-17 18:20:35,103][starVLA.training.train_starvla][INFO] - Step 1003, Loss: {'train/loss': 0.5540534257888794, 'train/grad_norm_pre_clip': 7.484533309936523, 'timing/data': 0.009895345196127892, 'timing/model': 1.0530436504632235, 'learning_rate/qwen_vl_interface': 9.604164763866479e-06, 'learning_rate/action_model': 9.60416476386648e-05, 'epoch': 0.46}) +[2026-06-17 18:20:39,039][starVLA.training.train_starvla][INFO] - Step 1004, Loss: {'train/loss': 0.6831117868423462, 'train/grad_norm_pre_clip': 11.677815437316895, 'timing/data': 0.014766600914299488, 'timing/model': 0.9400232266634703, 'learning_rate/qwen_vl_interface': 9.60329998562021e-06, 'learning_rate/action_model': 9.60329998562021e-05, 'epoch': 0.46}) +[2026-06-17 18:20:43,068][starVLA.training.train_starvla][INFO] - Step 1005, Loss: {'train/loss': 0.7502045631408691, 'train/grad_norm_pre_clip': 8.8183012008667, 'timing/data': 0.008839928545057774, 'timing/model': 1.215136961080134, 'learning_rate/qwen_vl_interface': 9.602434304930168e-06, 'learning_rate/action_model': 9.602434304930168e-05, 'epoch': 0.46}) +[2026-06-17 18:20:46,720][starVLA.training.train_starvla][INFO] - Step 1006, Loss: {'train/loss': 0.688569188117981, 'train/grad_norm_pre_clip': 7.827206134796143, 'timing/data': 0.009035482071340084, 'timing/model': 0.9559053564444184, 'learning_rate/qwen_vl_interface': 9.601567721975811e-06, 'learning_rate/action_model': 9.601567721975812e-05, 'epoch': 0.46}) +[2026-06-17 18:20:50,532][starVLA.training.train_starvla][INFO] - Step 1007, Loss: {'train/loss': 0.4669635593891144, 'train/grad_norm_pre_clip': 5.726832389831543, 'timing/data': 0.00880802795290947, 'timing/model': 0.9205302046611905, 'learning_rate/qwen_vl_interface': 9.600700236936785e-06, 'learning_rate/action_model': 9.600700236936785e-05, 'epoch': 0.46}) +[2026-06-17 18:20:54,351][starVLA.training.train_starvla][INFO] - Step 1008, Loss: {'train/loss': 0.5975379347801208, 'train/grad_norm_pre_clip': 7.913700580596924, 'timing/data': 0.011464547365903854, 'timing/model': 0.9212949192151427, 'learning_rate/qwen_vl_interface': 9.599831849992916e-06, 'learning_rate/action_model': 9.599831849992916e-05, 'epoch': 0.46}) +[2026-06-17 18:20:58,132][starVLA.training.train_starvla][INFO] - Step 1009, Loss: {'train/loss': 0.6889867782592773, 'train/grad_norm_pre_clip': 7.772773742675781, 'timing/data': 0.012622345238924026, 'timing/model': 0.92140350677073, 'learning_rate/qwen_vl_interface': 9.598962561324225e-06, 'learning_rate/action_model': 9.598962561324225e-05, 'epoch': 0.46}) +[2026-06-17 18:21:01,869][starVLA.training.train_starvla][INFO] - Step 1010, Loss: {'train/loss': 0.5487449765205383, 'train/grad_norm_pre_clip': 8.7658052444458, 'timing/data': 0.008767061866819859, 'timing/model': 0.9194877874106169, 'learning_rate/qwen_vl_interface': 9.598092371110916e-06, 'learning_rate/action_model': 9.598092371110916e-05, 'epoch': 0.46}) +[2026-06-17 18:21:05,952][starVLA.training.train_starvla][INFO] - Step 1011, Loss: {'train/loss': 0.6825294494628906, 'train/grad_norm_pre_clip': 11.144585609436035, 'timing/data': 0.017765452153980732, 'timing/model': 1.2598953004926443, 'learning_rate/qwen_vl_interface': 9.59722127953338e-06, 'learning_rate/action_model': 9.597221279533379e-05, 'epoch': 0.46}) +[2026-06-17 18:21:09,705][starVLA.training.train_starvla][INFO] - Step 1012, Loss: {'train/loss': 0.6514360904693604, 'train/grad_norm_pre_clip': 6.572537422180176, 'timing/data': 0.009804815985262394, 'timing/model': 1.0633518621325493, 'learning_rate/qwen_vl_interface': 9.596349286772194e-06, 'learning_rate/action_model': 9.596349286772193e-05, 'epoch': 0.46}) +[2026-06-17 18:21:13,331][starVLA.training.train_starvla][INFO] - Step 1013, Loss: {'train/loss': 0.729049801826477, 'train/grad_norm_pre_clip': 7.314629077911377, 'timing/data': 0.008159313350915909, 'timing/model': 0.938025550916791, 'learning_rate/qwen_vl_interface': 9.595476393008123e-06, 'learning_rate/action_model': 9.595476393008123e-05, 'epoch': 0.46}) +[2026-06-17 18:21:17,143][starVLA.training.train_starvla][INFO] - Step 1014, Loss: {'train/loss': 0.7123389840126038, 'train/grad_norm_pre_clip': 9.34593677520752, 'timing/data': 0.009466647170484066, 'timing/model': 0.9549518898129463, 'learning_rate/qwen_vl_interface': 9.59460259842212e-06, 'learning_rate/action_model': 9.59460259842212e-05, 'epoch': 0.46}) +[2026-06-17 18:21:20,936][starVLA.training.train_starvla][INFO] - Step 1015, Loss: {'train/loss': 0.7574317455291748, 'train/grad_norm_pre_clip': 7.093087673187256, 'timing/data': 0.009877863340079784, 'timing/model': 0.9484272943809628, 'learning_rate/qwen_vl_interface': 9.593727903195323e-06, 'learning_rate/action_model': 9.593727903195324e-05, 'epoch': 0.46}) +[2026-06-17 18:21:24,796][starVLA.training.train_starvla][INFO] - Step 1016, Loss: {'train/loss': 0.7374762296676636, 'train/grad_norm_pre_clip': 8.576026916503906, 'timing/data': 0.009535104036331177, 'timing/model': 0.9520739279687405, 'learning_rate/qwen_vl_interface': 9.59285230750906e-06, 'learning_rate/action_model': 9.592852307509059e-05, 'epoch': 0.46}) +[2026-06-17 18:21:28,705][starVLA.training.train_starvla][INFO] - Step 1017, Loss: {'train/loss': 0.4999196231365204, 'train/grad_norm_pre_clip': 5.898338794708252, 'timing/data': 0.009065727703273296, 'timing/model': 0.9666748158633709, 'learning_rate/qwen_vl_interface': 9.591975811544839e-06, 'learning_rate/action_model': 9.591975811544839e-05, 'epoch': 0.46}) +[2026-06-17 18:21:32,531][starVLA.training.train_starvla][INFO] - Step 1018, Loss: {'train/loss': 0.7065161466598511, 'train/grad_norm_pre_clip': 10.044458389282227, 'timing/data': 0.009029748849570751, 'timing/model': 1.1524062128737569, 'learning_rate/qwen_vl_interface': 9.59109841548436e-06, 'learning_rate/action_model': 9.591098415484359e-05, 'epoch': 0.46}) +[2026-06-17 18:21:36,194][starVLA.training.train_starvla][INFO] - Step 1019, Loss: {'train/loss': 0.6645189523696899, 'train/grad_norm_pre_clip': 14.116353034973145, 'timing/data': 0.009332207962870598, 'timing/model': 0.935522154904902, 'learning_rate/qwen_vl_interface': 9.590220119509508e-06, 'learning_rate/action_model': 9.590220119509508e-05, 'epoch': 0.46}) +[2026-06-17 18:21:40,089][starVLA.training.train_starvla][INFO] - Step 1020, Loss: {'train/loss': 0.5008758902549744, 'train/grad_norm_pre_clip': 9.597599029541016, 'timing/data': 0.013131384737789631, 'timing/model': 0.9974776161834598, 'learning_rate/qwen_vl_interface': 9.589340923802355e-06, 'learning_rate/action_model': 9.589340923802354e-05, 'epoch': 0.46}) +[2026-06-17 18:21:43,914][starVLA.training.train_starvla][INFO] - Step 1021, Loss: {'train/loss': 0.7075192332267761, 'train/grad_norm_pre_clip': 8.794426918029785, 'timing/data': 0.010509603656828403, 'timing/model': 0.9255017936229706, 'learning_rate/qwen_vl_interface': 9.58846082854516e-06, 'learning_rate/action_model': 9.58846082854516e-05, 'epoch': 0.46}) +[2026-06-17 18:21:47,676][starVLA.training.train_starvla][INFO] - Step 1022, Loss: {'train/loss': 0.41105806827545166, 'train/grad_norm_pre_clip': 7.988426208496094, 'timing/data': 0.010017964988946915, 'timing/model': 0.9271741891279817, 'learning_rate/qwen_vl_interface': 9.587579833920366e-06, 'learning_rate/action_model': 9.587579833920365e-05, 'epoch': 0.46}) +[2026-06-17 18:21:51,688][starVLA.training.train_starvla][INFO] - Step 1023, Loss: {'train/loss': 0.8892638087272644, 'train/grad_norm_pre_clip': 10.371673583984375, 'timing/data': 0.00979685503989458, 'timing/model': 1.1108621573075652, 'learning_rate/qwen_vl_interface': 9.586697940110605e-06, 'learning_rate/action_model': 9.586697940110605e-05, 'epoch': 0.46}) +[2026-06-17 18:21:55,473][starVLA.training.train_starvla][INFO] - Step 1024, Loss: {'train/loss': 0.6009030342102051, 'train/grad_norm_pre_clip': 8.253682136535645, 'timing/data': 0.01086475234478712, 'timing/model': 0.926336876116693, 'learning_rate/qwen_vl_interface': 9.585815147298696e-06, 'learning_rate/action_model': 9.585815147298696e-05, 'epoch': 0.46}) +[2026-06-17 18:21:59,330][starVLA.training.train_starvla][INFO] - Step 1025, Loss: {'train/loss': 0.6936966776847839, 'train/grad_norm_pre_clip': 7.83296012878418, 'timing/data': 0.01020889077335596, 'timing/model': 1.0957781057804823, 'learning_rate/qwen_vl_interface': 9.58493145566764e-06, 'learning_rate/action_model': 9.58493145566764e-05, 'epoch': 0.47}) +[2026-06-17 18:22:03,004][starVLA.training.train_starvla][INFO] - Step 1026, Loss: {'train/loss': 0.9213857650756836, 'train/grad_norm_pre_clip': 14.997621536254883, 'timing/data': 0.008484520949423313, 'timing/model': 0.9558212952688336, 'learning_rate/qwen_vl_interface': 9.584046865400627e-06, 'learning_rate/action_model': 9.584046865400627e-05, 'epoch': 0.47}) +[2026-06-17 18:22:06,893][starVLA.training.train_starvla][INFO] - Step 1027, Loss: {'train/loss': 0.5868425965309143, 'train/grad_norm_pre_clip': 10.45674991607666, 'timing/data': 0.013475751504302025, 'timing/model': 1.0029419790953398, 'learning_rate/qwen_vl_interface': 9.583161376681039e-06, 'learning_rate/action_model': 9.583161376681038e-05, 'epoch': 0.47}) +[2026-06-17 18:22:12,086][starVLA.training.train_starvla][INFO] - Step 1028, Loss: {'train/loss': 0.9278814792633057, 'train/grad_norm_pre_clip': 12.66737174987793, 'timing/data': 0.0072988709434866905, 'timing/model': 2.3445635866373777, 'learning_rate/qwen_vl_interface': 9.582274989692432e-06, 'learning_rate/action_model': 9.582274989692432e-05, 'epoch': 0.47}) +[2026-06-17 18:22:15,952][starVLA.training.train_starvla][INFO] - Step 1029, Loss: {'train/loss': 0.4366472661495209, 'train/grad_norm_pre_clip': 7.601690292358398, 'timing/data': 0.008134182542562485, 'timing/model': 0.9371721614152193, 'learning_rate/qwen_vl_interface': 9.58138770461856e-06, 'learning_rate/action_model': 9.58138770461856e-05, 'epoch': 0.47}) +[2026-06-17 18:22:20,777][starVLA.training.train_starvla][INFO] - Step 1030, Loss: {'train/loss': 0.7941094636917114, 'train/grad_norm_pre_clip': 11.496967315673828, 'timing/data': 0.016083110123872757, 'timing/model': 0.9229836380109191, 'learning_rate/qwen_vl_interface': 9.580499521643355e-06, 'learning_rate/action_model': 9.580499521643355e-05, 'epoch': 0.47}) +[2026-06-17 18:22:24,718][starVLA.training.train_starvla][INFO] - Step 1031, Loss: {'train/loss': 0.5742377042770386, 'train/grad_norm_pre_clip': 11.144124031066895, 'timing/data': 0.008444460108876228, 'timing/model': 1.248415439389646, 'learning_rate/qwen_vl_interface': 9.579610440950939e-06, 'learning_rate/action_model': 9.579610440950937e-05, 'epoch': 0.47}) +[2026-06-17 18:22:28,558][starVLA.training.train_starvla][INFO] - Step 1032, Loss: {'train/loss': 0.7961602210998535, 'train/grad_norm_pre_clip': 11.369492530822754, 'timing/data': 0.007448069751262665, 'timing/model': 1.1519796084612608, 'learning_rate/qwen_vl_interface': 9.57872046272562e-06, 'learning_rate/action_model': 9.57872046272562e-05, 'epoch': 0.47}) +[2026-06-17 18:22:32,294][starVLA.training.train_starvla][INFO] - Step 1033, Loss: {'train/loss': 0.6960513591766357, 'train/grad_norm_pre_clip': 9.419293403625488, 'timing/data': 0.008818301372230053, 'timing/model': 0.9421230088919401, 'learning_rate/qwen_vl_interface': 9.577829587151892e-06, 'learning_rate/action_model': 9.577829587151891e-05, 'epoch': 0.47}) +[2026-06-17 18:22:36,161][starVLA.training.train_starvla][INFO] - Step 1034, Loss: {'train/loss': 0.6277047991752625, 'train/grad_norm_pre_clip': 11.67927360534668, 'timing/data': 0.011002961546182632, 'timing/model': 0.9318341873586178, 'learning_rate/qwen_vl_interface': 9.576937814414433e-06, 'learning_rate/action_model': 9.576937814414432e-05, 'epoch': 0.47}) +[2026-06-17 18:22:40,045][starVLA.training.train_starvla][INFO] - Step 1035, Loss: {'train/loss': 0.6835285425186157, 'train/grad_norm_pre_clip': 15.576642990112305, 'timing/data': 0.01034310832619667, 'timing/model': 0.9396688677370548, 'learning_rate/qwen_vl_interface': 9.576045144698109e-06, 'learning_rate/action_model': 9.576045144698108e-05, 'epoch': 0.47}) +[2026-06-17 18:22:43,893][starVLA.training.train_starvla][INFO] - Step 1036, Loss: {'train/loss': 0.7571761012077332, 'train/grad_norm_pre_clip': 9.836336135864258, 'timing/data': 0.01879575103521347, 'timing/model': 0.9361659856513143, 'learning_rate/qwen_vl_interface': 9.57515157818797e-06, 'learning_rate/action_model': 9.57515157818797e-05, 'epoch': 0.47}) +[2026-06-17 18:22:47,823][starVLA.training.train_starvla][INFO] - Step 1037, Loss: {'train/loss': 0.6289429664611816, 'train/grad_norm_pre_clip': 10.176475524902344, 'timing/data': 0.01942725758999586, 'timing/model': 0.9459737278521061, 'learning_rate/qwen_vl_interface': 9.574257115069256e-06, 'learning_rate/action_model': 9.574257115069256e-05, 'epoch': 0.47}) +[2026-06-17 18:22:51,652][starVLA.training.train_starvla][INFO] - Step 1038, Loss: {'train/loss': 0.5815643668174744, 'train/grad_norm_pre_clip': 7.7223334312438965, 'timing/data': 0.012654883787035942, 'timing/model': 1.172405006363988, 'learning_rate/qwen_vl_interface': 9.573361755527389e-06, 'learning_rate/action_model': 9.573361755527389e-05, 'epoch': 0.47}) +[2026-06-17 18:22:55,421][starVLA.training.train_starvla][INFO] - Step 1039, Loss: {'train/loss': 0.4477402865886688, 'train/grad_norm_pre_clip': 6.679872035980225, 'timing/data': 0.011710936203598976, 'timing/model': 0.9536763932555914, 'learning_rate/qwen_vl_interface': 9.572465499747978e-06, 'learning_rate/action_model': 9.572465499747978e-05, 'epoch': 0.47}) +[2026-06-17 18:22:59,328][starVLA.training.train_starvla][INFO] - Step 1040, Loss: {'train/loss': 0.5753839612007141, 'train/grad_norm_pre_clip': 16.791845321655273, 'timing/data': 0.01129039004445076, 'timing/model': 0.9462921125814319, 'learning_rate/qwen_vl_interface': 9.571568347916819e-06, 'learning_rate/action_model': 9.571568347916819e-05, 'epoch': 0.47}) +[2026-06-17 18:23:03,149][starVLA.training.train_starvla][INFO] - Step 1041, Loss: {'train/loss': 0.6582265496253967, 'train/grad_norm_pre_clip': 17.940305709838867, 'timing/data': 0.009714018553495407, 'timing/model': 0.9459687024354935, 'learning_rate/qwen_vl_interface': 9.570670300219889e-06, 'learning_rate/action_model': 9.570670300219889e-05, 'epoch': 0.47}) +[2026-06-17 18:23:06,979][starVLA.training.train_starvla][INFO] - Step 1042, Loss: {'train/loss': 0.5100281834602356, 'train/grad_norm_pre_clip': 7.837742328643799, 'timing/data': 0.00976728554815054, 'timing/model': 0.9387053260579705, 'learning_rate/qwen_vl_interface': 9.56977135684336e-06, 'learning_rate/action_model': 9.569771356843359e-05, 'epoch': 0.47}) +[2026-06-17 18:23:11,471][starVLA.training.train_starvla][INFO] - Step 1043, Loss: {'train/loss': 0.455185204744339, 'train/grad_norm_pre_clip': 6.85669469833374, 'timing/data': 0.01076594553887844, 'timing/model': 0.9356624959036708, 'learning_rate/qwen_vl_interface': 9.568871517973576e-06, 'learning_rate/action_model': 9.568871517973577e-05, 'epoch': 0.47}) +[2026-06-17 18:23:15,520][starVLA.training.train_starvla][INFO] - Step 1044, Loss: {'train/loss': 0.743887722492218, 'train/grad_norm_pre_clip': 9.096345901489258, 'timing/data': 0.012053754180669785, 'timing/model': 0.9451431529596448, 'learning_rate/qwen_vl_interface': 9.567970783797081e-06, 'learning_rate/action_model': 9.567970783797082e-05, 'epoch': 0.47}) +[2026-06-17 18:23:19,651][starVLA.training.train_starvla][INFO] - Step 1045, Loss: {'train/loss': 0.8176392912864685, 'train/grad_norm_pre_clip': 9.410555839538574, 'timing/data': 0.018806468695402145, 'timing/model': 1.1611516438424587, 'learning_rate/qwen_vl_interface': 9.567069154500597e-06, 'learning_rate/action_model': 9.567069154500596e-05, 'epoch': 0.47}) +[2026-06-17 18:23:23,274][starVLA.training.train_starvla][INFO] - Step 1046, Loss: {'train/loss': 0.5267195105552673, 'train/grad_norm_pre_clip': 10.279006004333496, 'timing/data': 0.009527349844574928, 'timing/model': 0.9373404886573553, 'learning_rate/qwen_vl_interface': 9.56616663027103e-06, 'learning_rate/action_model': 9.566166630271029e-05, 'epoch': 0.47}) +[2026-06-17 18:23:27,096][starVLA.training.train_starvla][INFO] - Step 1047, Loss: {'train/loss': 0.42268407344818115, 'train/grad_norm_pre_clip': 7.3365864753723145, 'timing/data': 0.013024978339672089, 'timing/model': 0.9339991584420204, 'learning_rate/qwen_vl_interface': 9.56526321129548e-06, 'learning_rate/action_model': 9.565263211295479e-05, 'epoch': 0.47}) +[2026-06-17 18:23:30,988][starVLA.training.train_starvla][INFO] - Step 1048, Loss: {'train/loss': 0.7179428339004517, 'train/grad_norm_pre_clip': 24.85431671142578, 'timing/data': 0.013687408529222012, 'timing/model': 0.9401671886444092, 'learning_rate/qwen_vl_interface': 9.56435889776122e-06, 'learning_rate/action_model': 9.564358897761218e-05, 'epoch': 0.48}) +[2026-06-17 18:23:34,826][starVLA.training.train_starvla][INFO] - Step 1049, Loss: {'train/loss': 0.3067333698272705, 'train/grad_norm_pre_clip': 5.868814945220947, 'timing/data': 0.015128735452890396, 'timing/model': 0.9473490845412016, 'learning_rate/qwen_vl_interface': 9.563453689855717e-06, 'learning_rate/action_model': 9.563453689855716e-05, 'epoch': 0.48}) +[2026-06-17 18:23:38,702][starVLA.training.train_starvla][INFO] - Step 1050, Loss: {'train/loss': 1.0669264793395996, 'train/grad_norm_pre_clip': 9.116944313049316, 'timing/data': 0.012557689100503922, 'timing/model': 0.9429809469729662, 'learning_rate/qwen_vl_interface': 9.562547587766624e-06, 'learning_rate/action_model': 9.562547587766623e-05, 'epoch': 0.48}) +[2026-06-17 18:23:42,570][starVLA.training.train_starvla][INFO] - Step 1051, Loss: {'train/loss': 0.5306606888771057, 'train/grad_norm_pre_clip': 5.933287143707275, 'timing/data': 0.0070256683975458145, 'timing/model': 1.1866845060139894, 'learning_rate/qwen_vl_interface': 9.561640591681775e-06, 'learning_rate/action_model': 9.561640591681774e-05, 'epoch': 0.48}) +[2026-06-17 18:23:46,371][starVLA.training.train_starvla][INFO] - Step 1052, Loss: {'train/loss': 0.7495563626289368, 'train/grad_norm_pre_clip': 9.188133239746094, 'timing/data': 0.012076443061232567, 'timing/model': 1.1005702363327146, 'learning_rate/qwen_vl_interface': 9.560732701789193e-06, 'learning_rate/action_model': 9.560732701789193e-05, 'epoch': 0.48}) +[2026-06-17 18:23:50,088][starVLA.training.train_starvla][INFO] - Step 1053, Loss: {'train/loss': 0.4998188018798828, 'train/grad_norm_pre_clip': 9.171638488769531, 'timing/data': 0.015182098373770714, 'timing/model': 0.9343176865950227, 'learning_rate/qwen_vl_interface': 9.55982391827708e-06, 'learning_rate/action_model': 9.559823918277081e-05, 'epoch': 0.48}) +[2026-06-17 18:23:53,928][starVLA.training.train_starvla][INFO] - Step 1054, Loss: {'train/loss': 0.37157633900642395, 'train/grad_norm_pre_clip': 10.351776123046875, 'timing/data': 0.009449864737689495, 'timing/model': 0.9402948739007115, 'learning_rate/qwen_vl_interface': 9.558914241333835e-06, 'learning_rate/action_model': 9.558914241333835e-05, 'epoch': 0.48}) +[2026-06-17 18:23:57,849][starVLA.training.train_starvla][INFO] - Step 1055, Loss: {'train/loss': 0.5931588411331177, 'train/grad_norm_pre_clip': 9.657787322998047, 'timing/data': 0.014097585342824459, 'timing/model': 0.9414453897625208, 'learning_rate/qwen_vl_interface': 9.55800367114803e-06, 'learning_rate/action_model': 9.55800367114803e-05, 'epoch': 0.48}) +[2026-06-17 18:24:01,665][starVLA.training.train_starvla][INFO] - Step 1056, Loss: {'train/loss': 1.0514549016952515, 'train/grad_norm_pre_clip': 13.956914901733398, 'timing/data': 0.015228234231472015, 'timing/model': 0.9893528716638684, 'learning_rate/qwen_vl_interface': 9.557092207908428e-06, 'learning_rate/action_model': 9.557092207908428e-05, 'epoch': 0.48}) +[2026-06-17 18:24:05,580][starVLA.training.train_starvla][INFO] - Step 1057, Loss: {'train/loss': 0.779583215713501, 'train/grad_norm_pre_clip': 9.362643241882324, 'timing/data': 0.012810132466256618, 'timing/model': 0.9352734973654151, 'learning_rate/qwen_vl_interface': 9.556179851803978e-06, 'learning_rate/action_model': 9.556179851803978e-05, 'epoch': 0.48}) +[2026-06-17 18:24:09,409][starVLA.training.train_starvla][INFO] - Step 1058, Loss: {'train/loss': 0.34999197721481323, 'train/grad_norm_pre_clip': 7.570390701293945, 'timing/data': 0.009427190758287907, 'timing/model': 1.125637256540358, 'learning_rate/qwen_vl_interface': 9.555266603023809e-06, 'learning_rate/action_model': 9.555266603023808e-05, 'epoch': 0.48}) +[2026-06-17 18:24:13,036][starVLA.training.train_starvla][INFO] - Step 1059, Loss: {'train/loss': 0.2996980547904968, 'train/grad_norm_pre_clip': 5.296506881713867, 'timing/data': 0.010432744398713112, 'timing/model': 0.9441237114369869, 'learning_rate/qwen_vl_interface': 9.554352461757243e-06, 'learning_rate/action_model': 9.554352461757242e-05, 'epoch': 0.48}) +[2026-06-17 18:24:16,978][starVLA.training.train_starvla][INFO] - Step 1060, Loss: {'train/loss': 0.6241380572319031, 'train/grad_norm_pre_clip': 8.661337852478027, 'timing/data': 0.018278068862855434, 'timing/model': 1.0017061438411474, 'learning_rate/qwen_vl_interface': 9.553437428193778e-06, 'learning_rate/action_model': 9.553437428193778e-05, 'epoch': 0.48}) +[2026-06-17 18:24:20,842][starVLA.training.train_starvla][INFO] - Step 1061, Loss: {'train/loss': 0.624268651008606, 'train/grad_norm_pre_clip': 6.709102630615234, 'timing/data': 0.010179138742387295, 'timing/model': 0.9587531825527549, 'learning_rate/qwen_vl_interface': 9.552521502523104e-06, 'learning_rate/action_model': 9.552521502523102e-05, 'epoch': 0.48}) +[2026-06-17 18:24:24,663][starVLA.training.train_starvla][INFO] - Step 1062, Loss: {'train/loss': 0.751436173915863, 'train/grad_norm_pre_clip': 12.39321231842041, 'timing/data': 0.008445776998996735, 'timing/model': 0.9387811562046409, 'learning_rate/qwen_vl_interface': 9.551604684935093e-06, 'learning_rate/action_model': 9.551604684935092e-05, 'epoch': 0.48}) +[2026-06-17 18:24:28,507][starVLA.training.train_starvla][INFO] - Step 1063, Loss: {'train/loss': 0.40521350502967834, 'train/grad_norm_pre_clip': 8.948616981506348, 'timing/data': 0.00959772802889347, 'timing/model': 0.9290039325132966, 'learning_rate/qwen_vl_interface': 9.550686975619803e-06, 'learning_rate/action_model': 9.550686975619802e-05, 'epoch': 0.48}) +[2026-06-17 18:24:32,369][starVLA.training.train_starvla][INFO] - Step 1064, Loss: {'train/loss': 0.744363009929657, 'train/grad_norm_pre_clip': 10.760454177856445, 'timing/data': 0.00802670605480671, 'timing/model': 0.938265641219914, 'learning_rate/qwen_vl_interface': 9.549768374767473e-06, 'learning_rate/action_model': 9.549768374767473e-05, 'epoch': 0.48}) +[2026-06-17 18:24:36,201][starVLA.training.train_starvla][INFO] - Step 1065, Loss: {'train/loss': 0.929901123046875, 'train/grad_norm_pre_clip': 11.735779762268066, 'timing/data': 0.00994946900755167, 'timing/model': 1.1383859580382705, 'learning_rate/qwen_vl_interface': 9.548848882568533e-06, 'learning_rate/action_model': 9.548848882568532e-05, 'epoch': 0.48}) +[2026-06-17 18:24:39,830][starVLA.training.train_starvla][INFO] - Step 1066, Loss: {'train/loss': 0.47972893714904785, 'train/grad_norm_pre_clip': 7.479805946350098, 'timing/data': 0.01353576872497797, 'timing/model': 0.9433456333354115, 'learning_rate/qwen_vl_interface': 9.547928499213593e-06, 'learning_rate/action_model': 9.547928499213593e-05, 'epoch': 0.48}) +[2026-06-17 18:24:43,671][starVLA.training.train_starvla][INFO] - Step 1067, Loss: {'train/loss': 0.2635422348976135, 'train/grad_norm_pre_clip': 7.986931800842285, 'timing/data': 0.00861435942351818, 'timing/model': 0.9824727158993483, 'learning_rate/qwen_vl_interface': 9.547007224893452e-06, 'learning_rate/action_model': 9.547007224893452e-05, 'epoch': 0.48}) +[2026-06-17 18:24:47,569][starVLA.training.train_starvla][INFO] - Step 1068, Loss: {'train/loss': 0.9330437183380127, 'train/grad_norm_pre_clip': 14.117315292358398, 'timing/data': 0.011205828748643398, 'timing/model': 0.9390099961310625, 'learning_rate/qwen_vl_interface': 9.546085059799088e-06, 'learning_rate/action_model': 9.546085059799088e-05, 'epoch': 0.48}) +[2026-06-17 18:24:51,423][starVLA.training.train_starvla][INFO] - Step 1069, Loss: {'train/loss': 0.5410863757133484, 'train/grad_norm_pre_clip': 8.209811210632324, 'timing/data': 0.008438695222139359, 'timing/model': 0.971807555295527, 'learning_rate/qwen_vl_interface': 9.545162004121672e-06, 'learning_rate/action_model': 9.54516200412167e-05, 'epoch': 0.48}) +[2026-06-17 18:24:55,314][starVLA.training.train_starvla][INFO] - Step 1070, Loss: {'train/loss': 0.9759494066238403, 'train/grad_norm_pre_clip': 9.97948169708252, 'timing/data': 0.010886897332966328, 'timing/model': 0.9403980011120439, 'learning_rate/qwen_vl_interface': 9.544238058052546e-06, 'learning_rate/action_model': 9.544238058052545e-05, 'epoch': 0.49}) +[2026-06-17 18:24:59,305][starVLA.training.train_starvla][INFO] - Step 1071, Loss: {'train/loss': 0.6097188591957092, 'train/grad_norm_pre_clip': 8.5418701171875, 'timing/data': 0.01277326699346304, 'timing/model': 1.2935926243662834, 'learning_rate/qwen_vl_interface': 9.54331322178325e-06, 'learning_rate/action_model': 9.543313221783251e-05, 'epoch': 0.49}) +[2026-06-17 18:25:04,747][starVLA.training.train_starvla][INFO] - Step 1072, Loss: {'train/loss': 0.6478705406188965, 'train/grad_norm_pre_clip': 8.442536354064941, 'timing/data': 0.013351473957300186, 'timing/model': 2.7612137710675597, 'learning_rate/qwen_vl_interface': 9.542387495505505e-06, 'learning_rate/action_model': 9.542387495505506e-05, 'epoch': 0.49}) +[2026-06-17 18:25:08,565][starVLA.training.train_starvla][INFO] - Step 1073, Loss: {'train/loss': 0.8555784821510315, 'train/grad_norm_pre_clip': 8.947160720825195, 'timing/data': 0.015132677741348743, 'timing/model': 0.9345071678981185, 'learning_rate/qwen_vl_interface': 9.541460879411214e-06, 'learning_rate/action_model': 9.541460879411215e-05, 'epoch': 0.49}) +[2026-06-17 18:25:12,457][starVLA.training.train_starvla][INFO] - Step 1074, Loss: {'train/loss': 0.3715474605560303, 'train/grad_norm_pre_clip': 7.884502410888672, 'timing/data': 0.011266715824604034, 'timing/model': 0.9478120831772685, 'learning_rate/qwen_vl_interface': 9.540533373692466e-06, 'learning_rate/action_model': 9.540533373692466e-05, 'epoch': 0.49}) +[2026-06-17 18:25:16,417][starVLA.training.train_starvla][INFO] - Step 1075, Loss: {'train/loss': 0.527553141117096, 'train/grad_norm_pre_clip': 9.178173065185547, 'timing/data': 0.010546193458139896, 'timing/model': 0.9992153886705637, 'learning_rate/qwen_vl_interface': 9.539604978541531e-06, 'learning_rate/action_model': 9.539604978541531e-05, 'epoch': 0.49}) +[2026-06-17 18:25:20,221][starVLA.training.train_starvla][INFO] - Step 1076, Loss: {'train/loss': 0.5812079310417175, 'train/grad_norm_pre_clip': 8.795599937438965, 'timing/data': 0.011245914734899998, 'timing/model': 0.9238004479557276, 'learning_rate/qwen_vl_interface': 9.538675694150868e-06, 'learning_rate/action_model': 9.538675694150868e-05, 'epoch': 0.49}) +[2026-06-17 18:25:24,297][starVLA.training.train_starvla][INFO] - Step 1077, Loss: {'train/loss': 0.9811933040618896, 'train/grad_norm_pre_clip': 11.201196670532227, 'timing/data': 0.011963206343352795, 'timing/model': 0.9370144130662084, 'learning_rate/qwen_vl_interface': 9.53774552071312e-06, 'learning_rate/action_model': 9.537745520713119e-05, 'epoch': 0.49}) +[2026-06-17 18:25:28,144][starVLA.training.train_starvla][INFO] - Step 1078, Loss: {'train/loss': 0.48232269287109375, 'train/grad_norm_pre_clip': 7.723759174346924, 'timing/data': 0.01341941487044096, 'timing/model': 1.1025598580017686, 'learning_rate/qwen_vl_interface': 9.53681445842111e-06, 'learning_rate/action_model': 9.53681445842111e-05, 'epoch': 0.49}) +[2026-06-17 18:25:31,860][starVLA.training.train_starvla][INFO] - Step 1079, Loss: {'train/loss': 0.705865740776062, 'train/grad_norm_pre_clip': 8.197707176208496, 'timing/data': 0.012175774201750755, 'timing/model': 0.9505740525200963, 'learning_rate/qwen_vl_interface': 9.53588250746785e-06, 'learning_rate/action_model': 9.53588250746785e-05, 'epoch': 0.49}) +[2026-06-17 18:25:36,309][starVLA.training.train_starvla][INFO] - Step 1080, Loss: {'train/loss': 0.32972007989883423, 'train/grad_norm_pre_clip': 10.073203086853027, 'timing/data': 0.012805919162929058, 'timing/model': 1.0222808001562953, 'learning_rate/qwen_vl_interface': 9.534949668046537e-06, 'learning_rate/action_model': 9.534949668046536e-05, 'epoch': 0.49}) +[2026-06-17 18:25:40,233][starVLA.training.train_starvla][INFO] - Step 1081, Loss: {'train/loss': 0.837813675403595, 'train/grad_norm_pre_clip': 9.895467758178711, 'timing/data': 0.009759250096976757, 'timing/model': 0.9373435219749808, 'learning_rate/qwen_vl_interface': 9.534015940350545e-06, 'learning_rate/action_model': 9.534015940350544e-05, 'epoch': 0.49}) +[2026-06-17 18:25:44,139][starVLA.training.train_starvla][INFO] - Step 1082, Loss: {'train/loss': 0.6830216646194458, 'train/grad_norm_pre_clip': 8.65279483795166, 'timing/data': 0.012132653035223484, 'timing/model': 0.9362272098660469, 'learning_rate/qwen_vl_interface': 9.533081324573437e-06, 'learning_rate/action_model': 9.533081324573437e-05, 'epoch': 0.49}) +[2026-06-17 18:25:50,407][starVLA.training.train_starvla][INFO] - Step 1083, Loss: {'train/loss': 0.6809508800506592, 'train/grad_norm_pre_clip': 7.4043684005737305, 'timing/data': 0.017265353351831436, 'timing/model': 0.9434887645766139, 'learning_rate/qwen_vl_interface': 9.532145820908965e-06, 'learning_rate/action_model': 9.532145820908964e-05, 'epoch': 0.49}) +[2026-06-17 18:25:54,305][starVLA.training.train_starvla][INFO] - Step 1084, Loss: {'train/loss': 0.5788789391517639, 'train/grad_norm_pre_clip': 8.068167686462402, 'timing/data': 0.013844872824847698, 'timing/model': 0.9319320246577263, 'learning_rate/qwen_vl_interface': 9.531209429551056e-06, 'learning_rate/action_model': 9.531209429551054e-05, 'epoch': 0.49}) +[2026-06-17 18:25:58,137][starVLA.training.train_starvla][INFO] - Step 1085, Loss: {'train/loss': 0.5110392570495605, 'train/grad_norm_pre_clip': 8.359236717224121, 'timing/data': 0.009582020342350006, 'timing/model': 1.1238235691562295, 'learning_rate/qwen_vl_interface': 9.530272150693823e-06, 'learning_rate/action_model': 9.530272150693824e-05, 'epoch': 0.49}) +[2026-06-17 18:26:01,856][starVLA.training.train_starvla][INFO] - Step 1086, Loss: {'train/loss': 0.46326032280921936, 'train/grad_norm_pre_clip': 7.346348762512207, 'timing/data': 0.01121953222900629, 'timing/model': 0.9771292172372341, 'learning_rate/qwen_vl_interface': 9.52933398453157e-06, 'learning_rate/action_model': 9.52933398453157e-05, 'epoch': 0.49}) +[2026-06-17 18:26:05,778][starVLA.training.train_starvla][INFO] - Step 1087, Loss: {'train/loss': 0.7322747111320496, 'train/grad_norm_pre_clip': 8.890585899353027, 'timing/data': 0.013127594254910946, 'timing/model': 0.9657603828236461, 'learning_rate/qwen_vl_interface': 9.528394931258777e-06, 'learning_rate/action_model': 9.528394931258776e-05, 'epoch': 0.49}) +[2026-06-17 18:26:09,793][starVLA.training.train_starvla][INFO] - Step 1088, Loss: {'train/loss': 0.41819578409194946, 'train/grad_norm_pre_clip': 5.892707824707031, 'timing/data': 0.009704863652586937, 'timing/model': 1.0291138719767332, 'learning_rate/qwen_vl_interface': 9.527454991070108e-06, 'learning_rate/action_model': 9.527454991070108e-05, 'epoch': 0.49}) +[2026-06-17 18:26:13,645][starVLA.training.train_starvla][INFO] - Step 1089, Loss: {'train/loss': 0.3965023458003998, 'train/grad_norm_pre_clip': 5.368531703948975, 'timing/data': 0.010924272239208221, 'timing/model': 0.9262350732460618, 'learning_rate/qwen_vl_interface': 9.52651416416042e-06, 'learning_rate/action_model': 9.526514164160421e-05, 'epoch': 0.49}) +[2026-06-17 18:26:17,513][starVLA.training.train_starvla][INFO] - Step 1090, Loss: {'train/loss': 0.508263885974884, 'train/grad_norm_pre_clip': 8.475714683532715, 'timing/data': 0.010703012347221375, 'timing/model': 0.9546488290652633, 'learning_rate/qwen_vl_interface': 9.525572450724743e-06, 'learning_rate/action_model': 9.525572450724743e-05, 'epoch': 0.49}) +[2026-06-17 18:26:21,368][starVLA.training.train_starvla][INFO] - Step 1091, Loss: {'train/loss': 0.7098384499549866, 'train/grad_norm_pre_clip': 8.480835914611816, 'timing/data': 0.009625863283872604, 'timing/model': 1.1502611935138702, 'learning_rate/qwen_vl_interface': 9.524629850958297e-06, 'learning_rate/action_model': 9.524629850958297e-05, 'epoch': 0.49}) +[2026-06-17 18:26:25,396][starVLA.training.train_starvla][INFO] - Step 1092, Loss: {'train/loss': 0.7719241380691528, 'train/grad_norm_pre_clip': 9.489577293395996, 'timing/data': 0.011226391419768333, 'timing/model': 1.3126613348722458, 'learning_rate/qwen_vl_interface': 9.523686365056482e-06, 'learning_rate/action_model': 9.523686365056481e-05, 'epoch': 0.5}) +[2026-06-17 18:26:29,288][starVLA.training.train_starvla][INFO] - Step 1093, Loss: {'train/loss': 0.6947730779647827, 'train/grad_norm_pre_clip': 8.576032638549805, 'timing/data': 0.01600291207432747, 'timing/model': 1.0070212995633483, 'learning_rate/qwen_vl_interface': 9.522741993214886e-06, 'learning_rate/action_model': 9.522741993214886e-05, 'epoch': 0.5}) +[2026-06-17 18:26:33,333][starVLA.training.train_starvla][INFO] - Step 1094, Loss: {'train/loss': 0.7791102528572083, 'train/grad_norm_pre_clip': 10.214259147644043, 'timing/data': 0.012678404338657856, 'timing/model': 0.9871484264731407, 'learning_rate/qwen_vl_interface': 9.521796735629275e-06, 'learning_rate/action_model': 9.521796735629275e-05, 'epoch': 0.5}) +[2026-06-17 18:26:37,629][starVLA.training.train_starvla][INFO] - Step 1095, Loss: {'train/loss': 0.5465579032897949, 'train/grad_norm_pre_clip': 9.230467796325684, 'timing/data': 0.015000124461948872, 'timing/model': 0.9337933175265789, 'learning_rate/qwen_vl_interface': 9.520850592495603e-06, 'learning_rate/action_model': 9.520850592495604e-05, 'epoch': 0.5}) +[2026-06-17 18:26:41,928][starVLA.training.train_starvla][INFO] - Step 1096, Loss: {'train/loss': 0.7128174304962158, 'train/grad_norm_pre_clip': 8.50817584991455, 'timing/data': 0.009046340361237526, 'timing/model': 0.9399802898988128, 'learning_rate/qwen_vl_interface': 9.519903564010012e-06, 'learning_rate/action_model': 9.51990356401001e-05, 'epoch': 0.5}) +[2026-06-17 18:26:45,876][starVLA.training.train_starvla][INFO] - Step 1097, Loss: {'train/loss': 0.7310518026351929, 'train/grad_norm_pre_clip': 12.869308471679688, 'timing/data': 0.016983909532427788, 'timing/model': 0.9875115472823381, 'learning_rate/qwen_vl_interface': 9.518955650368816e-06, 'learning_rate/action_model': 9.518955650368814e-05, 'epoch': 0.5}) +[2026-06-17 18:26:49,764][starVLA.training.train_starvla][INFO] - Step 1098, Loss: {'train/loss': 0.3889378607273102, 'train/grad_norm_pre_clip': 16.275598526000977, 'timing/data': 0.012208196334540844, 'timing/model': 1.1367360716685653, 'learning_rate/qwen_vl_interface': 9.51800685176852e-06, 'learning_rate/action_model': 9.518006851768518e-05, 'epoch': 0.5}) +[2026-06-17 18:26:53,582][starVLA.training.train_starvla][INFO] - Step 1099, Loss: {'train/loss': 0.49151918292045593, 'train/grad_norm_pre_clip': 18.847469329833984, 'timing/data': 0.0172945698723197, 'timing/model': 0.973809496499598, 'learning_rate/qwen_vl_interface': 9.51705716840581e-06, 'learning_rate/action_model': 9.517057168405808e-05, 'epoch': 0.5}) +[2026-06-17 18:29:08,437][starVLA.training.train_starvla][INFO] - Step 1100, Loss: {'train/loss': 0.3616883456707001, 'train/grad_norm_pre_clip': 7.519779682159424, 'eval/loss': 0.6533539295196533, 'eval/latency_8/loss': 0.6533539295196533, 'eval/demon_fire_precision': 0.9211606096131302, 'eval/demon_fire_recall': 0.976390183286735, 'eval/demon_fire_f1': 0.947971648318504, 'eval/demon_fire_rate_model': 0.9493600445186422, 'eval/demon_fire_rate_teacher': 0.8956594323873122, 'eval/demon_left_precision': 0.812807881773399, 'eval/demon_left_recall': 0.3533190578158458, 'eval/demon_left_f1': 0.4925373134328358, 'eval/demon_left_rate_model': 0.05648302726766834, 'eval/demon_left_rate_teacher': 0.12993878686700056, 'eval/demon_right_precision': 0.8895027624309392, 'eval/demon_right_recall': 0.48787878787878786, 'eval/demon_right_f1': 0.6301369863013698, 'eval/demon_right_rate_model': 0.10072342793544797, 'eval/demon_right_rate_teacher': 0.18363939899833054, 'eval/demon_move_f1': 0.5613371498671028, 'eval/demon_cc_f1': 0.7856342813355827, 'eval/latency_8/demon_fire_precision': 0.9211606096131302, 'eval/latency_8/demon_fire_recall': 0.976390183286735, 'eval/latency_8/demon_fire_f1': 0.947971648318504, 'eval/latency_8/demon_fire_rate_model': 0.9493600445186422, 'eval/latency_8/demon_fire_rate_teacher': 0.8956594323873122, 'eval/latency_8/demon_left_precision': 0.812807881773399, 'eval/latency_8/demon_left_recall': 0.3533190578158458, 'eval/latency_8/demon_left_f1': 0.4925373134328358, 'eval/latency_8/demon_left_rate_model': 0.05648302726766834, 'eval/latency_8/demon_left_rate_teacher': 0.12993878686700056, 'eval/latency_8/demon_right_precision': 0.8895027624309392, 'eval/latency_8/demon_right_recall': 0.48787878787878786, 'eval/latency_8/demon_right_f1': 0.6301369863013698, 'eval/latency_8/demon_right_rate_model': 0.10072342793544797, 'eval/latency_8/demon_right_rate_teacher': 0.18363939899833054, 'eval/latency_8/demon_move_f1': 0.5613371498671028, 'eval/latency_8/demon_cc_f1': 0.7856342813355827, 'timing/data': 0.011932656168937683, 'timing/model': 0.9376943875104189, 'learning_rate/qwen_vl_interface': 9.516106600477557e-06, 'learning_rate/action_model': 9.516106600477557e-05, 'epoch': 0.5}) +[2026-06-17 18:29:08,442][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_1100_state +[2026-06-17 18:29:08,468][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 18:29:16,106][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_1100_state/model.safetensors +[2026-06-17 18:29:32,647][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_1100_state/optimizer.bin +[2026-06-17 18:29:32,649][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_1100_state/sampler.bin +[2026-06-17 18:29:32,649][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_1100_state/sampler_1.bin +[2026-06-17 18:29:32,652][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_1100_state/random_states_0.pkl +[2026-06-17 18:29:32,658][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1100 +[2026-06-17 18:29:32,659][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 18:29:32,670][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 18:29:38,048][starVLA.training.train_starvla][INFO] - Step 1101, Loss: {'train/loss': 0.33847668766975403, 'train/grad_norm_pre_clip': 5.507523059844971, 'timing/data': 0.01904628612101078, 'timing/model': 1.3411445338279009, 'learning_rate/qwen_vl_interface': 9.515155148180816e-06, 'learning_rate/action_model': 9.515155148180816e-05, 'epoch': 0.5}) +[2026-06-17 18:29:41,957][starVLA.training.train_starvla][INFO] - Step 1102, Loss: {'train/loss': 0.6108195781707764, 'train/grad_norm_pre_clip': 10.300132751464844, 'timing/data': 0.012808695435523987, 'timing/model': 1.0836451835930347, 'learning_rate/qwen_vl_interface': 9.514202811712825e-06, 'learning_rate/action_model': 9.514202811712825e-05, 'epoch': 0.5}) +[2026-06-17 18:29:45,997][starVLA.training.train_starvla][INFO] - Step 1103, Loss: {'train/loss': 0.8248576521873474, 'train/grad_norm_pre_clip': 13.757004737854004, 'timing/data': 0.023165968246757984, 'timing/model': 0.9724566619843245, 'learning_rate/qwen_vl_interface': 9.513249591271e-06, 'learning_rate/action_model': 9.513249591271e-05, 'epoch': 0.5}) +[2026-06-17 18:29:50,045][starVLA.training.train_starvla][INFO] - Step 1104, Loss: {'train/loss': 0.6533250212669373, 'train/grad_norm_pre_clip': 10.502876281738281, 'timing/data': 0.021368484012782574, 'timing/model': 1.1095198225229979, 'learning_rate/qwen_vl_interface': 9.512295487052949e-06, 'learning_rate/action_model': 9.512295487052949e-05, 'epoch': 0.5}) +[2026-06-17 18:29:54,214][starVLA.training.train_starvla][INFO] - Step 1105, Loss: {'train/loss': 1.0232865810394287, 'train/grad_norm_pre_clip': 11.650334358215332, 'timing/data': 0.02957483660429716, 'timing/model': 0.9623971506953239, 'learning_rate/qwen_vl_interface': 9.511340499256455e-06, 'learning_rate/action_model': 9.511340499256455e-05, 'epoch': 0.5}) +[2026-06-17 18:29:58,310][starVLA.training.train_starvla][INFO] - Step 1106, Loss: {'train/loss': 1.0645817518234253, 'train/grad_norm_pre_clip': 13.355420112609863, 'timing/data': 0.02499205805361271, 'timing/model': 1.016523684374988, 'learning_rate/qwen_vl_interface': 9.510384628079493e-06, 'learning_rate/action_model': 9.510384628079493e-05, 'epoch': 0.5}) +[2026-06-17 18:30:02,278][starVLA.training.train_starvla][INFO] - Step 1107, Loss: {'train/loss': 0.9224132895469666, 'train/grad_norm_pre_clip': 8.324745178222656, 'timing/data': 0.023073710501194, 'timing/model': 1.2139586554840207, 'learning_rate/qwen_vl_interface': 9.50942787372021e-06, 'learning_rate/action_model': 9.50942787372021e-05, 'epoch': 0.5}) +[2026-06-17 18:30:06,017][starVLA.training.train_starvla][INFO] - Step 1108, Loss: {'train/loss': 0.6253032088279724, 'train/grad_norm_pre_clip': 9.475937843322754, 'timing/data': 0.021076057106256485, 'timing/model': 0.9870031047612429, 'learning_rate/qwen_vl_interface': 9.508470236376948e-06, 'learning_rate/action_model': 9.508470236376948e-05, 'epoch': 0.5}) +[2026-06-17 18:30:10,006][starVLA.training.train_starvla][INFO] - Step 1109, Loss: {'train/loss': 0.5166079998016357, 'train/grad_norm_pre_clip': 10.856111526489258, 'timing/data': 0.018516055308282375, 'timing/model': 0.9627718403935432, 'learning_rate/qwen_vl_interface': 9.507511716248224e-06, 'learning_rate/action_model': 9.507511716248224e-05, 'epoch': 0.5}) +[2026-06-17 18:30:13,973][starVLA.training.train_starvla][INFO] - Step 1110, Loss: {'train/loss': 0.6386361718177795, 'train/grad_norm_pre_clip': 12.35433578491211, 'timing/data': 0.018567820079624653, 'timing/model': 0.9503044970333576, 'learning_rate/qwen_vl_interface': 9.506552313532739e-06, 'learning_rate/action_model': 9.506552313532739e-05, 'epoch': 0.5}) +[2026-06-17 18:30:18,052][starVLA.training.train_starvla][INFO] - Step 1111, Loss: {'train/loss': 0.7042030692100525, 'train/grad_norm_pre_clip': 14.526089668273926, 'timing/data': 0.01850089803338051, 'timing/model': 0.9565238934010267, 'learning_rate/qwen_vl_interface': 9.505592028429377e-06, 'learning_rate/action_model': 9.505592028429377e-05, 'epoch': 0.5}) +[2026-06-17 18:30:22,083][starVLA.training.train_starvla][INFO] - Step 1112, Loss: {'train/loss': 0.5922951102256775, 'train/grad_norm_pre_clip': 7.66984748840332, 'timing/data': 0.016208999790251255, 'timing/model': 1.0004009837284684, 'learning_rate/qwen_vl_interface': 9.504630861137212e-06, 'learning_rate/action_model': 9.504630861137212e-05, 'epoch': 0.5}) +[2026-06-17 18:30:26,144][starVLA.training.train_starvla][INFO] - Step 1113, Loss: {'train/loss': 0.5013249516487122, 'train/grad_norm_pre_clip': 10.02994155883789, 'timing/data': 0.01708744466304779, 'timing/model': 0.9773514419794083, 'learning_rate/qwen_vl_interface': 9.503668811855489e-06, 'learning_rate/action_model': 9.503668811855488e-05, 'epoch': 0.5}) +[2026-06-17 18:30:30,065][starVLA.training.train_starvla][INFO] - Step 1114, Loss: {'train/loss': 0.7350661754608154, 'train/grad_norm_pre_clip': 10.983148574829102, 'timing/data': 0.028922772966325283, 'timing/model': 1.1741843204945326, 'learning_rate/qwen_vl_interface': 9.502705880783644e-06, 'learning_rate/action_model': 9.502705880783645e-05, 'epoch': 0.51}) +[2026-06-17 18:30:33,879][starVLA.training.train_starvla][INFO] - Step 1115, Loss: {'train/loss': 0.5424509644508362, 'train/grad_norm_pre_clip': 10.799084663391113, 'timing/data': 0.017640995793044567, 'timing/model': 0.9601239273324609, 'learning_rate/qwen_vl_interface': 9.501742068121296e-06, 'learning_rate/action_model': 9.501742068121296e-05, 'epoch': 0.51}) +[2026-06-17 18:30:38,699][starVLA.training.train_starvla][INFO] - Step 1116, Loss: {'train/loss': 0.597030520439148, 'train/grad_norm_pre_clip': 7.301507949829102, 'timing/data': 0.020973085425794125, 'timing/model': 0.968811271712184, 'learning_rate/qwen_vl_interface': 9.500777374068241e-06, 'learning_rate/action_model': 9.50077737406824e-05, 'epoch': 0.51}) +[2026-06-17 18:30:42,867][starVLA.training.train_starvla][INFO] - Step 1117, Loss: {'train/loss': 0.5173047184944153, 'train/grad_norm_pre_clip': 7.5654473304748535, 'timing/data': 0.02216928731650114, 'timing/model': 0.9871010147035122, 'learning_rate/qwen_vl_interface': 9.49981179882446e-06, 'learning_rate/action_model': 9.499811798824461e-05, 'epoch': 0.51}) +[2026-06-17 18:30:47,086][starVLA.training.train_starvla][INFO] - Step 1118, Loss: {'train/loss': 0.6981330513954163, 'train/grad_norm_pre_clip': 10.385801315307617, 'timing/data': 0.016745788045227528, 'timing/model': 1.0256337570026517, 'learning_rate/qwen_vl_interface': 9.498845342590124e-06, 'learning_rate/action_model': 9.498845342590124e-05, 'epoch': 0.51}) +[2026-06-17 18:30:51,128][starVLA.training.train_starvla][INFO] - Step 1119, Loss: {'train/loss': 0.733293354511261, 'train/grad_norm_pre_clip': 8.813386917114258, 'timing/data': 0.0173711059615016, 'timing/model': 0.971533115953207, 'learning_rate/qwen_vl_interface': 9.497878005565575e-06, 'learning_rate/action_model': 9.497878005565574e-05, 'epoch': 0.51}) +[2026-06-17 18:30:55,085][starVLA.training.train_starvla][INFO] - Step 1120, Loss: {'train/loss': 0.9849003553390503, 'train/grad_norm_pre_clip': 10.043532371520996, 'timing/data': 0.03219054639339447, 'timing/model': 1.196785761974752, 'learning_rate/qwen_vl_interface': 9.496909787951344e-06, 'learning_rate/action_model': 9.496909787951343e-05, 'epoch': 0.51}) +[2026-06-17 18:30:59,206][starVLA.training.train_starvla][INFO] - Step 1121, Loss: {'train/loss': 0.6539350748062134, 'train/grad_norm_pre_clip': 7.922177791595459, 'timing/data': 0.017289210110902786, 'timing/model': 1.3146779043599963, 'learning_rate/qwen_vl_interface': 9.495940689948144e-06, 'learning_rate/action_model': 9.495940689948144e-05, 'epoch': 0.51}) +[2026-06-17 18:31:02,965][starVLA.training.train_starvla][INFO] - Step 1122, Loss: {'train/loss': 0.45196038484573364, 'train/grad_norm_pre_clip': 6.867295265197754, 'timing/data': 0.042479791678488255, 'timing/model': 0.9770807204768062, 'learning_rate/qwen_vl_interface': 9.49497071175687e-06, 'learning_rate/action_model': 9.49497071175687e-05, 'epoch': 0.51}) +[2026-06-17 18:31:07,005][starVLA.training.train_starvla][INFO] - Step 1123, Loss: {'train/loss': 0.5526411533355713, 'train/grad_norm_pre_clip': 9.353302001953125, 'timing/data': 0.018749536015093327, 'timing/model': 0.9528441810980439, 'learning_rate/qwen_vl_interface': 9.493999853578601e-06, 'learning_rate/action_model': 9.493999853578602e-05, 'epoch': 0.51}) +[2026-06-17 18:31:10,938][starVLA.training.train_starvla][INFO] - Step 1124, Loss: {'train/loss': 0.4549534320831299, 'train/grad_norm_pre_clip': 9.315533638000488, 'timing/data': 0.01830470096319914, 'timing/model': 0.9575955960899591, 'learning_rate/qwen_vl_interface': 9.493028115614595e-06, 'learning_rate/action_model': 9.493028115614595e-05, 'epoch': 0.51}) +[2026-06-17 18:31:15,228][starVLA.training.train_starvla][INFO] - Step 1125, Loss: {'train/loss': 0.5581459999084473, 'train/grad_norm_pre_clip': 7.265738487243652, 'timing/data': 0.02039529476314783, 'timing/model': 1.0375347137451172, 'learning_rate/qwen_vl_interface': 9.492055498066295e-06, 'learning_rate/action_model': 9.492055498066294e-05, 'epoch': 0.51}) +[2026-06-17 18:31:19,319][starVLA.training.train_starvla][INFO] - Step 1126, Loss: {'train/loss': 0.7329920530319214, 'train/grad_norm_pre_clip': 8.69403076171875, 'timing/data': 0.014321812428534031, 'timing/model': 0.9737971164286137, 'learning_rate/qwen_vl_interface': 9.491082001135326e-06, 'learning_rate/action_model': 9.491082001135325e-05, 'epoch': 0.51}) +[2026-06-17 18:31:23,665][starVLA.training.train_starvla][INFO] - Step 1127, Loss: {'train/loss': 0.835171103477478, 'train/grad_norm_pre_clip': 10.115296363830566, 'timing/data': 0.0206632474437356, 'timing/model': 1.4811259377747774, 'learning_rate/qwen_vl_interface': 9.490107625023492e-06, 'learning_rate/action_model': 9.490107625023492e-05, 'epoch': 0.51}) +[2026-06-17 18:31:29,561][starVLA.training.train_starvla][INFO] - Step 1128, Loss: {'train/loss': 0.5359387993812561, 'train/grad_norm_pre_clip': 10.56324291229248, 'timing/data': 0.023524307645857334, 'timing/model': 1.2906051278114319, 'learning_rate/qwen_vl_interface': 9.489132369932788e-06, 'learning_rate/action_model': 9.489132369932788e-05, 'epoch': 0.51}) +[2026-06-17 18:31:33,292][starVLA.training.train_starvla][INFO] - Step 1129, Loss: {'train/loss': 0.8732390999794006, 'train/grad_norm_pre_clip': 9.008009910583496, 'timing/data': 0.020650474354624748, 'timing/model': 0.9818650782108307, 'learning_rate/qwen_vl_interface': 9.488156236065378e-06, 'learning_rate/action_model': 9.488156236065379e-05, 'epoch': 0.51}) +[2026-06-17 18:31:37,289][starVLA.training.train_starvla][INFO] - Step 1130, Loss: {'train/loss': 0.4725387692451477, 'train/grad_norm_pre_clip': 7.472344398498535, 'timing/data': 0.019698589108884335, 'timing/model': 0.9930565468966961, 'learning_rate/qwen_vl_interface': 9.487179223623626e-06, 'learning_rate/action_model': 9.487179223623624e-05, 'epoch': 0.51}) +[2026-06-17 18:31:41,349][starVLA.training.train_starvla][INFO] - Step 1131, Loss: {'train/loss': 0.8090580105781555, 'train/grad_norm_pre_clip': 9.278334617614746, 'timing/data': 0.03242773003876209, 'timing/model': 0.9689358780160546, 'learning_rate/qwen_vl_interface': 9.486201332810058e-06, 'learning_rate/action_model': 9.486201332810057e-05, 'epoch': 0.51}) +[2026-06-17 18:31:45,308][starVLA.training.train_starvla][INFO] - Step 1132, Loss: {'train/loss': 0.35801461338996887, 'train/grad_norm_pre_clip': 7.27053165435791, 'timing/data': 0.028870140202343464, 'timing/model': 0.955548994243145, 'learning_rate/qwen_vl_interface': 9.485222563827395e-06, 'learning_rate/action_model': 9.485222563827395e-05, 'epoch': 0.51}) +[2026-06-17 18:31:49,251][starVLA.training.train_starvla][INFO] - Step 1133, Loss: {'train/loss': 0.655205488204956, 'train/grad_norm_pre_clip': 9.208113670349121, 'timing/data': 0.02766102645546198, 'timing/model': 0.9505308354273438, 'learning_rate/qwen_vl_interface': 9.484242916878537e-06, 'learning_rate/action_model': 9.484242916878536e-05, 'epoch': 0.51}) +[2026-06-17 18:31:53,300][starVLA.training.train_starvla][INFO] - Step 1134, Loss: {'train/loss': 0.749168872833252, 'train/grad_norm_pre_clip': 10.11844253540039, 'timing/data': 0.022011044435203075, 'timing/model': 1.2694920441135764, 'learning_rate/qwen_vl_interface': 9.483262392166567e-06, 'learning_rate/action_model': 9.483262392166566e-05, 'epoch': 0.51}) +[2026-06-17 18:31:57,058][starVLA.training.train_starvla][INFO] - Step 1135, Loss: {'train/loss': 0.5194036364555359, 'train/grad_norm_pre_clip': 8.35924243927002, 'timing/data': 0.022636176086962223, 'timing/model': 0.9854677151888609, 'learning_rate/qwen_vl_interface': 9.482280989894748e-06, 'learning_rate/action_model': 9.482280989894748e-05, 'epoch': 0.51}) +[2026-06-17 18:32:01,065][starVLA.training.train_starvla][INFO] - Step 1136, Loss: {'train/loss': 0.8522390723228455, 'train/grad_norm_pre_clip': 10.006744384765625, 'timing/data': 0.01956187654286623, 'timing/model': 0.9576637791469693, 'learning_rate/qwen_vl_interface': 9.481298710266525e-06, 'learning_rate/action_model': 9.481298710266526e-05, 'epoch': 0.52}) +[2026-06-17 18:32:05,150][starVLA.training.train_starvla][INFO] - Step 1137, Loss: {'train/loss': 0.7372667789459229, 'train/grad_norm_pre_clip': 9.416379928588867, 'timing/data': 0.021471748128533363, 'timing/model': 0.9702357249334455, 'learning_rate/qwen_vl_interface': 9.480315553485527e-06, 'learning_rate/action_model': 9.480315553485527e-05, 'epoch': 0.52}) +[2026-06-17 18:32:09,271][starVLA.training.train_starvla][INFO] - Step 1138, Loss: {'train/loss': 0.542996346950531, 'train/grad_norm_pre_clip': 8.20773696899414, 'timing/data': 0.015484919771552086, 'timing/model': 0.9604156399145722, 'learning_rate/qwen_vl_interface': 9.479331519755563e-06, 'learning_rate/action_model': 9.479331519755562e-05, 'epoch': 0.52}) +[2026-06-17 18:32:13,332][starVLA.training.train_starvla][INFO] - Step 1139, Loss: {'train/loss': 0.7730063796043396, 'train/grad_norm_pre_clip': 9.358641624450684, 'timing/data': 0.025534729473292828, 'timing/model': 0.9607700891792774, 'learning_rate/qwen_vl_interface': 9.478346609280622e-06, 'learning_rate/action_model': 9.478346609280623e-05, 'epoch': 0.52}) +[2026-06-17 18:32:17,456][starVLA.training.train_starvla][INFO] - Step 1140, Loss: {'train/loss': 0.594295084476471, 'train/grad_norm_pre_clip': 8.36976432800293, 'timing/data': 0.019388574175536633, 'timing/model': 0.9673143420368433, 'learning_rate/qwen_vl_interface': 9.477360822264883e-06, 'learning_rate/action_model': 9.477360822264882e-05, 'epoch': 0.52}) +[2026-06-17 18:32:22,206][starVLA.training.train_starvla][INFO] - Step 1141, Loss: {'train/loss': 0.6228934526443481, 'train/grad_norm_pre_clip': 9.160529136657715, 'timing/data': 0.019646497443318367, 'timing/model': 1.3035468999296427, 'learning_rate/qwen_vl_interface': 9.476374158912692e-06, 'learning_rate/action_model': 9.476374158912691e-05, 'epoch': 0.52}) +[2026-06-17 18:32:29,002][starVLA.training.train_starvla][INFO] - Step 1142, Loss: {'train/loss': 0.6123937964439392, 'train/grad_norm_pre_clip': 10.034276962280273, 'timing/data': 0.03141653072088957, 'timing/model': 3.1154362400993705, 'learning_rate/qwen_vl_interface': 9.475386619428593e-06, 'learning_rate/action_model': 9.475386619428594e-05, 'epoch': 0.52}) +[2026-06-17 18:32:33,107][starVLA.training.train_starvla][INFO] - Step 1143, Loss: {'train/loss': 0.25574052333831787, 'train/grad_norm_pre_clip': 5.4640607833862305, 'timing/data': 0.018815300427377224, 'timing/model': 0.9735728427767754, 'learning_rate/qwen_vl_interface': 9.4743982040173e-06, 'learning_rate/action_model': 9.4743982040173e-05, 'epoch': 0.52}) +[2026-06-17 18:32:37,154][starVLA.training.train_starvla][INFO] - Step 1144, Loss: {'train/loss': 0.36586689949035645, 'train/grad_norm_pre_clip': 7.049795627593994, 'timing/data': 0.024153745733201504, 'timing/model': 0.9662799639627337, 'learning_rate/qwen_vl_interface': 9.473408912883714e-06, 'learning_rate/action_model': 9.473408912883714e-05, 'epoch': 0.52}) +[2026-06-17 18:32:41,246][starVLA.training.train_starvla][INFO] - Step 1145, Loss: {'train/loss': 0.6104177236557007, 'train/grad_norm_pre_clip': 6.347857475280762, 'timing/data': 0.017597775906324387, 'timing/model': 0.9476426057517529, 'learning_rate/qwen_vl_interface': 9.472418746232916e-06, 'learning_rate/action_model': 9.472418746232916e-05, 'epoch': 0.52}) +[2026-06-17 18:32:45,166][starVLA.training.train_starvla][INFO] - Step 1146, Loss: {'train/loss': 0.6278440356254578, 'train/grad_norm_pre_clip': 9.572169303894043, 'timing/data': 0.0225137360394001, 'timing/model': 0.9610180864110589, 'learning_rate/qwen_vl_interface': 9.471427704270168e-06, 'learning_rate/action_model': 9.471427704270169e-05, 'epoch': 0.52}) +[2026-06-17 18:32:49,154][starVLA.training.train_starvla][INFO] - Step 1147, Loss: {'train/loss': 0.49074670672416687, 'train/grad_norm_pre_clip': 10.314706802368164, 'timing/data': 0.017291217111051083, 'timing/model': 1.2327290112152696, 'learning_rate/qwen_vl_interface': 9.470435787200914e-06, 'learning_rate/action_model': 9.470435787200915e-05, 'epoch': 0.52}) +[2026-06-17 18:32:53,233][starVLA.training.train_starvla][INFO] - Step 1148, Loss: {'train/loss': 0.6015156507492065, 'train/grad_norm_pre_clip': 11.556538581848145, 'timing/data': 0.01626154500991106, 'timing/model': 1.3203813275322318, 'learning_rate/qwen_vl_interface': 9.469442995230781e-06, 'learning_rate/action_model': 9.469442995230782e-05, 'epoch': 0.52}) +[2026-06-17 18:32:57,041][starVLA.training.train_starvla][INFO] - Step 1149, Loss: {'train/loss': 0.60774165391922, 'train/grad_norm_pre_clip': 17.19608497619629, 'timing/data': 0.025060557760298252, 'timing/model': 0.9507628111168742, 'learning_rate/qwen_vl_interface': 9.468449328565574e-06, 'learning_rate/action_model': 9.468449328565573e-05, 'epoch': 0.52}) +[2026-06-17 18:33:01,165][starVLA.training.train_starvla][INFO] - Step 1150, Loss: {'train/loss': 0.5158730149269104, 'train/grad_norm_pre_clip': 7.424891948699951, 'timing/data': 0.017173059284687042, 'timing/model': 0.9509118013083935, 'learning_rate/qwen_vl_interface': 9.467454787411282e-06, 'learning_rate/action_model': 9.467454787411281e-05, 'epoch': 0.52}) +[2026-06-17 18:33:06,000][starVLA.training.train_starvla][INFO] - Step 1151, Loss: {'train/loss': 0.8214851021766663, 'train/grad_norm_pre_clip': 16.729049682617188, 'timing/data': 0.02010011300444603, 'timing/model': 1.2898192312568426, 'learning_rate/qwen_vl_interface': 9.466459371974076e-06, 'learning_rate/action_model': 9.466459371974076e-05, 'epoch': 0.52}) +[2026-06-17 18:33:10,063][starVLA.training.train_starvla][INFO] - Step 1152, Loss: {'train/loss': 0.5762759447097778, 'train/grad_norm_pre_clip': 10.256261825561523, 'timing/data': 0.015689289197325706, 'timing/model': 0.9892050102353096, 'learning_rate/qwen_vl_interface': 9.465463082460304e-06, 'learning_rate/action_model': 9.465463082460304e-05, 'epoch': 0.52}) +[2026-06-17 18:33:14,140][starVLA.training.train_starvla][INFO] - Step 1153, Loss: {'train/loss': 0.6714441776275635, 'train/grad_norm_pre_clip': 10.376453399658203, 'timing/data': 0.013670376501977444, 'timing/model': 0.9638488898053765, 'learning_rate/qwen_vl_interface': 9.4644659190765e-06, 'learning_rate/action_model': 9.464465919076499e-05, 'epoch': 0.52}) +[2026-06-17 18:33:18,146][starVLA.training.train_starvla][INFO] - Step 1154, Loss: {'train/loss': 0.46809521317481995, 'train/grad_norm_pre_clip': 10.331721305847168, 'timing/data': 0.016757291741669178, 'timing/model': 1.160468122921884, 'learning_rate/qwen_vl_interface': 9.463467882029374e-06, 'learning_rate/action_model': 9.463467882029373e-05, 'epoch': 0.52}) +[2026-06-17 18:33:22,161][starVLA.training.train_starvla][INFO] - Step 1155, Loss: {'train/loss': 0.9964137673377991, 'train/grad_norm_pre_clip': 10.148785591125488, 'timing/data': 0.01598531659692526, 'timing/model': 1.2083068182691932, 'learning_rate/qwen_vl_interface': 9.462468971525822e-06, 'learning_rate/action_model': 9.462468971525821e-05, 'epoch': 0.52}) +[2026-06-17 18:33:25,894][starVLA.training.train_starvla][INFO] - Step 1156, Loss: {'train/loss': 0.6191458702087402, 'train/grad_norm_pre_clip': 10.299897193908691, 'timing/data': 0.01601272076368332, 'timing/model': 0.9620046280324459, 'learning_rate/qwen_vl_interface': 9.461469187772921e-06, 'learning_rate/action_model': 9.46146918777292e-05, 'epoch': 0.52}) +[2026-06-17 18:33:30,832][starVLA.training.train_starvla][INFO] - Step 1157, Loss: {'train/loss': 0.36520349979400635, 'train/grad_norm_pre_clip': 8.152572631835938, 'timing/data': 0.020214151591062546, 'timing/model': 1.0449660355225205, 'learning_rate/qwen_vl_interface': 9.460468530977924e-06, 'learning_rate/action_model': 9.460468530977924e-05, 'epoch': 0.52}) +[2026-06-17 18:33:35,788][starVLA.training.train_starvla][INFO] - Step 1158, Loss: {'train/loss': 0.4076728820800781, 'train/grad_norm_pre_clip': 7.646739959716797, 'timing/data': 0.030910196714103222, 'timing/model': 1.0515839979052544, 'learning_rate/qwen_vl_interface': 9.459467001348271e-06, 'learning_rate/action_model': 9.45946700134827e-05, 'epoch': 0.53}) +[2026-06-17 18:33:41,093][starVLA.training.train_starvla][INFO] - Step 1159, Loss: {'train/loss': 0.6255385279655457, 'train/grad_norm_pre_clip': 8.000931739807129, 'timing/data': 0.018021496944129467, 'timing/model': 0.9864425184205174, 'learning_rate/qwen_vl_interface': 9.458464599091578e-06, 'learning_rate/action_model': 9.458464599091578e-05, 'epoch': 0.53}) +[2026-06-17 18:33:45,586][starVLA.training.train_starvla][INFO] - Step 1160, Loss: {'train/loss': 0.752677857875824, 'train/grad_norm_pre_clip': 8.643502235412598, 'timing/data': 0.019087995402514935, 'timing/model': 1.0030359448865056, 'learning_rate/qwen_vl_interface': 9.457461324415647e-06, 'learning_rate/action_model': 9.457461324415646e-05, 'epoch': 0.53}) +[2026-06-17 18:33:49,842][starVLA.training.train_starvla][INFO] - Step 1161, Loss: {'train/loss': 0.5323376655578613, 'train/grad_norm_pre_clip': 11.95937442779541, 'timing/data': 0.026215746998786926, 'timing/model': 1.3487868271768093, 'learning_rate/qwen_vl_interface': 9.456457177528454e-06, 'learning_rate/action_model': 9.456457177528453e-05, 'epoch': 0.53}) +[2026-06-17 18:33:53,659][starVLA.training.train_starvla][INFO] - Step 1162, Loss: {'train/loss': 0.8860962390899658, 'train/grad_norm_pre_clip': 11.324984550476074, 'timing/data': 0.022178606130182743, 'timing/model': 0.9458511983975768, 'learning_rate/qwen_vl_interface': 9.455452158638163e-06, 'learning_rate/action_model': 9.455452158638163e-05, 'epoch': 0.53}) +[2026-06-17 18:33:57,667][starVLA.training.train_starvla][INFO] - Step 1163, Loss: {'train/loss': 1.0292037725448608, 'train/grad_norm_pre_clip': 11.029400825500488, 'timing/data': 0.015214798040688038, 'timing/model': 0.9516240991652012, 'learning_rate/qwen_vl_interface': 9.454446267953114e-06, 'learning_rate/action_model': 9.454446267953114e-05, 'epoch': 0.53}) +[2026-06-17 18:34:01,582][starVLA.training.train_starvla][INFO] - Step 1164, Loss: {'train/loss': 0.4891303777694702, 'train/grad_norm_pre_clip': 8.606972694396973, 'timing/data': 0.025551450438797474, 'timing/model': 0.9643725454807281, 'learning_rate/qwen_vl_interface': 9.453439505681829e-06, 'learning_rate/action_model': 9.453439505681829e-05, 'epoch': 0.53}) +[2026-06-17 18:34:05,614][starVLA.training.train_starvla][INFO] - Step 1165, Loss: {'train/loss': 0.5815707445144653, 'train/grad_norm_pre_clip': 8.220635414123535, 'timing/data': 0.016690718941390514, 'timing/model': 0.9580123955383897, 'learning_rate/qwen_vl_interface': 9.452431872033012e-06, 'learning_rate/action_model': 9.452431872033012e-05, 'epoch': 0.53}) +[2026-06-17 18:34:09,787][starVLA.training.train_starvla][INFO] - Step 1166, Loss: {'train/loss': 0.660988986492157, 'train/grad_norm_pre_clip': 8.553061485290527, 'timing/data': 0.01835931744426489, 'timing/model': 0.9995749238878489, 'learning_rate/qwen_vl_interface': 9.451423367215546e-06, 'learning_rate/action_model': 9.451423367215545e-05, 'epoch': 0.53}) +[2026-06-17 18:34:13,749][starVLA.training.train_starvla][INFO] - Step 1167, Loss: {'train/loss': 0.866309642791748, 'train/grad_norm_pre_clip': 9.452434539794922, 'timing/data': 0.01743363682180643, 'timing/model': 1.2102534053847194, 'learning_rate/qwen_vl_interface': 9.450413991438494e-06, 'learning_rate/action_model': 9.450413991438494e-05, 'epoch': 0.53}) +[2026-06-17 18:34:17,681][starVLA.training.train_starvla][INFO] - Step 1168, Loss: {'train/loss': 0.5894049406051636, 'train/grad_norm_pre_clip': 6.979860305786133, 'timing/data': 0.02073678281158209, 'timing/model': 1.1635690787807107, 'learning_rate/qwen_vl_interface': 9.449403744911104e-06, 'learning_rate/action_model': 9.449403744911103e-05, 'epoch': 0.53}) +[2026-06-17 18:34:21,375][starVLA.training.train_starvla][INFO] - Step 1169, Loss: {'train/loss': 0.6792687773704529, 'train/grad_norm_pre_clip': 11.419695854187012, 'timing/data': 0.018733840435743332, 'timing/model': 0.9534874223172665, 'learning_rate/qwen_vl_interface': 9.448392627842797e-06, 'learning_rate/action_model': 9.448392627842795e-05, 'epoch': 0.53}) +[2026-06-17 18:34:25,463][starVLA.training.train_starvla][INFO] - Step 1170, Loss: {'train/loss': 0.8602890372276306, 'train/grad_norm_pre_clip': 9.683786392211914, 'timing/data': 0.01700152177363634, 'timing/model': 0.9469133215025067, 'learning_rate/qwen_vl_interface': 9.44738064044318e-06, 'learning_rate/action_model': 9.44738064044318e-05, 'epoch': 0.53}) +[2026-06-17 18:34:29,367][starVLA.training.train_starvla][INFO] - Step 1171, Loss: {'train/loss': 0.6970208287239075, 'train/grad_norm_pre_clip': 10.288442611694336, 'timing/data': 0.017647690139710903, 'timing/model': 0.9526062980294228, 'learning_rate/qwen_vl_interface': 9.44636778292204e-06, 'learning_rate/action_model': 9.44636778292204e-05, 'epoch': 0.53}) +[2026-06-17 18:34:33,421][starVLA.training.train_starvla][INFO] - Step 1172, Loss: {'train/loss': 0.8132308125495911, 'train/grad_norm_pre_clip': 9.112649917602539, 'timing/data': 0.019754359498620033, 'timing/model': 1.034701719880104, 'learning_rate/qwen_vl_interface': 9.445354055489345e-06, 'learning_rate/action_model': 9.445354055489345e-05, 'epoch': 0.53}) +[2026-06-17 18:34:37,489][starVLA.training.train_starvla][INFO] - Step 1173, Loss: {'train/loss': 0.5947706699371338, 'train/grad_norm_pre_clip': 9.681230545043945, 'timing/data': 0.02855458203703165, 'timing/model': 0.9621296068653464, 'learning_rate/qwen_vl_interface': 9.444339458355239e-06, 'learning_rate/action_model': 9.444339458355239e-05, 'epoch': 0.53}) +[2026-06-17 18:34:41,470][starVLA.training.train_starvla][INFO] - Step 1174, Loss: {'train/loss': 0.3765299320220947, 'train/grad_norm_pre_clip': 6.841820240020752, 'timing/data': 0.01778178382664919, 'timing/model': 1.2141342144459486, 'learning_rate/qwen_vl_interface': 9.443323991730049e-06, 'learning_rate/action_model': 9.443323991730049e-05, 'epoch': 0.53}) +[2026-06-17 18:34:45,619][starVLA.training.train_starvla][INFO] - Step 1175, Loss: {'train/loss': 0.7989054918289185, 'train/grad_norm_pre_clip': 8.738821029663086, 'timing/data': 0.017961801029741764, 'timing/model': 1.2063115267083049, 'learning_rate/qwen_vl_interface': 9.442307655824284e-06, 'learning_rate/action_model': 9.442307655824285e-05, 'epoch': 0.53}) +[2026-06-17 18:34:49,400][starVLA.training.train_starvla][INFO] - Step 1176, Loss: {'train/loss': 0.5326129794120789, 'train/grad_norm_pre_clip': 6.305724143981934, 'timing/data': 0.021294616162776947, 'timing/model': 1.0201978888362646, 'learning_rate/qwen_vl_interface': 9.441290450848631e-06, 'learning_rate/action_model': 9.44129045084863e-05, 'epoch': 0.53}) +[2026-06-17 18:34:53,407][starVLA.training.train_starvla][INFO] - Step 1177, Loss: {'train/loss': 0.7785617709159851, 'train/grad_norm_pre_clip': 9.361796379089355, 'timing/data': 0.027223662473261356, 'timing/model': 0.9799936516210437, 'learning_rate/qwen_vl_interface': 9.440272377013957e-06, 'learning_rate/action_model': 9.440272377013956e-05, 'epoch': 0.53}) +[2026-06-17 18:34:57,369][starVLA.training.train_starvla][INFO] - Step 1178, Loss: {'train/loss': 0.5660188794136047, 'train/grad_norm_pre_clip': 8.456925392150879, 'timing/data': 0.01599791180342436, 'timing/model': 0.9599744360893965, 'learning_rate/qwen_vl_interface': 9.43925343453131e-06, 'learning_rate/action_model': 9.43925343453131e-05, 'epoch': 0.53}) +[2026-06-17 18:35:01,221][starVLA.training.train_starvla][INFO] - Step 1179, Loss: {'train/loss': 0.8683072328567505, 'train/grad_norm_pre_clip': 8.196243286132812, 'timing/data': 0.02145387325435877, 'timing/model': 0.9622788410633802, 'learning_rate/qwen_vl_interface': 9.438233623611918e-06, 'learning_rate/action_model': 9.438233623611918e-05, 'epoch': 0.53}) +[2026-06-17 18:35:05,171][starVLA.training.train_starvla][INFO] - Step 1180, Loss: {'train/loss': 0.6938720941543579, 'train/grad_norm_pre_clip': 8.886358261108398, 'timing/data': 0.016825849190354347, 'timing/model': 0.9474182669073343, 'learning_rate/qwen_vl_interface': 9.43721294446719e-06, 'learning_rate/action_model': 9.437212944467191e-05, 'epoch': 0.54}) +[2026-06-17 18:35:09,136][starVLA.training.train_starvla][INFO] - Step 1181, Loss: {'train/loss': 0.608116865158081, 'train/grad_norm_pre_clip': 9.153258323669434, 'timing/data': 0.01693579088896513, 'timing/model': 1.1515455711632967, 'learning_rate/qwen_vl_interface': 9.436191397308713e-06, 'learning_rate/action_model': 9.436191397308713e-05, 'epoch': 0.54}) +[2026-06-17 18:35:12,913][starVLA.training.train_starvla][INFO] - Step 1182, Loss: {'train/loss': 0.36475029587745667, 'train/grad_norm_pre_clip': 4.789762020111084, 'timing/data': 0.025177364237606525, 'timing/model': 0.9743384635075927, 'learning_rate/qwen_vl_interface': 9.435168982348254e-06, 'learning_rate/action_model': 9.435168982348255e-05, 'epoch': 0.54}) +[2026-06-17 18:35:16,910][starVLA.training.train_starvla][INFO] - Step 1183, Loss: {'train/loss': 0.6254277229309082, 'train/grad_norm_pre_clip': 8.645029067993164, 'timing/data': 0.020763871259987354, 'timing/model': 0.9654258200898767, 'learning_rate/qwen_vl_interface': 9.434145699797763e-06, 'learning_rate/action_model': 9.434145699797764e-05, 'epoch': 0.54}) +[2026-06-17 18:35:20,870][starVLA.training.train_starvla][INFO] - Step 1184, Loss: {'train/loss': 0.8525277376174927, 'train/grad_norm_pre_clip': 11.567166328430176, 'timing/data': 0.018322416581213474, 'timing/model': 0.9520504614338279, 'learning_rate/qwen_vl_interface': 9.433121549869365e-06, 'learning_rate/action_model': 9.433121549869365e-05, 'epoch': 0.54}) +[2026-06-17 18:35:24,794][starVLA.training.train_starvla][INFO] - Step 1185, Loss: {'train/loss': 0.49379926919937134, 'train/grad_norm_pre_clip': 8.691034317016602, 'timing/data': 0.03223821148276329, 'timing/model': 0.9542661625891924, 'learning_rate/qwen_vl_interface': 9.43209653277537e-06, 'learning_rate/action_model': 9.432096532775369e-05, 'epoch': 0.54}) +[2026-06-17 18:35:28,795][starVLA.training.train_starvla][INFO] - Step 1186, Loss: {'train/loss': 0.4567393958568573, 'train/grad_norm_pre_clip': 6.07633638381958, 'timing/data': 0.01875412743538618, 'timing/model': 0.9747403962537646, 'learning_rate/qwen_vl_interface': 9.431070648728262e-06, 'learning_rate/action_model': 9.431070648728261e-05, 'epoch': 0.54}) +[2026-06-17 18:35:32,753][starVLA.training.train_starvla][INFO] - Step 1187, Loss: {'train/loss': 0.3841542601585388, 'train/grad_norm_pre_clip': 22.643383026123047, 'timing/data': 0.01604408212006092, 'timing/model': 0.9510932695120573, 'learning_rate/qwen_vl_interface': 9.430043897940709e-06, 'learning_rate/action_model': 9.43004389794071e-05, 'epoch': 0.54}) +[2026-06-17 18:35:38,186][starVLA.training.train_starvla][INFO] - Step 1188, Loss: {'train/loss': 0.5375238656997681, 'train/grad_norm_pre_clip': 7.4065752029418945, 'timing/data': 0.022870565764606, 'timing/model': 2.605087839998305, 'learning_rate/qwen_vl_interface': 9.429016280625559e-06, 'learning_rate/action_model': 9.429016280625559e-05, 'epoch': 0.54}) +[2026-06-17 18:35:42,004][starVLA.training.train_starvla][INFO] - Step 1189, Loss: {'train/loss': 0.6181452870368958, 'train/grad_norm_pre_clip': 8.225444793701172, 'timing/data': 0.02116247732192278, 'timing/model': 0.9599500065669417, 'learning_rate/qwen_vl_interface': 9.427987796995838e-06, 'learning_rate/action_model': 9.427987796995837e-05, 'epoch': 0.54}) +[2026-06-17 18:35:46,059][starVLA.training.train_starvla][INFO] - Step 1190, Loss: {'train/loss': 0.6332988142967224, 'train/grad_norm_pre_clip': 8.228320121765137, 'timing/data': 0.02026366163045168, 'timing/model': 0.9666875330731273, 'learning_rate/qwen_vl_interface': 9.42695844726475e-06, 'learning_rate/action_model': 9.426958447264749e-05, 'epoch': 0.54}) +[2026-06-17 18:35:50,205][starVLA.training.train_starvla][INFO] - Step 1191, Loss: {'train/loss': 0.7817263603210449, 'train/grad_norm_pre_clip': 8.89819049835205, 'timing/data': 0.03379776980727911, 'timing/model': 0.9609013125300407, 'learning_rate/qwen_vl_interface': 9.42592823164568e-06, 'learning_rate/action_model': 9.42592823164568e-05, 'epoch': 0.54}) +[2026-06-17 18:35:54,374][starVLA.training.train_starvla][INFO] - Step 1192, Loss: {'train/loss': 0.5343652963638306, 'train/grad_norm_pre_clip': 7.518853187561035, 'timing/data': 0.01779912505298853, 'timing/model': 0.9591061836108565, 'learning_rate/qwen_vl_interface': 9.424897150352197e-06, 'learning_rate/action_model': 9.424897150352197e-05, 'epoch': 0.54}) +[2026-06-17 18:35:58,348][starVLA.training.train_starvla][INFO] - Step 1193, Loss: {'train/loss': 0.7421295046806335, 'train/grad_norm_pre_clip': 8.813298225402832, 'timing/data': 0.017912345938384533, 'timing/model': 0.9588524969294667, 'learning_rate/qwen_vl_interface': 9.42386520359804e-06, 'learning_rate/action_model': 9.42386520359804e-05, 'epoch': 0.54}) +[2026-06-17 18:36:02,346][starVLA.training.train_starvla][INFO] - Step 1194, Loss: {'train/loss': 0.6322791576385498, 'train/grad_norm_pre_clip': 7.081727504730225, 'timing/data': 0.017242242582142353, 'timing/model': 1.2417631102725863, 'learning_rate/qwen_vl_interface': 9.422832391597136e-06, 'learning_rate/action_model': 9.422832391597136e-05, 'epoch': 0.54}) +[2026-06-17 18:36:06,458][starVLA.training.train_starvla][INFO] - Step 1195, Loss: {'train/loss': 0.6212091445922852, 'train/grad_norm_pre_clip': 9.596805572509766, 'timing/data': 0.01910348329693079, 'timing/model': 1.2680352926254272, 'learning_rate/qwen_vl_interface': 9.421798714563586e-06, 'learning_rate/action_model': 9.421798714563586e-05, 'epoch': 0.54}) +[2026-06-17 18:36:10,271][starVLA.training.train_starvla][INFO] - Step 1196, Loss: {'train/loss': 0.6301575303077698, 'train/grad_norm_pre_clip': 8.932416915893555, 'timing/data': 0.01912926696240902, 'timing/model': 0.9779027998447418, 'learning_rate/qwen_vl_interface': 9.420764172711676e-06, 'learning_rate/action_model': 9.420764172711676e-05, 'epoch': 0.54}) +[2026-06-17 18:36:14,379][starVLA.training.train_starvla][INFO] - Step 1197, Loss: {'train/loss': 0.8701814413070679, 'train/grad_norm_pre_clip': 14.670866966247559, 'timing/data': 0.02164609730243683, 'timing/model': 0.9876504437997937, 'learning_rate/qwen_vl_interface': 9.419728766255863e-06, 'learning_rate/action_model': 9.419728766255862e-05, 'epoch': 0.54}) +[2026-06-17 18:36:18,649][starVLA.training.train_starvla][INFO] - Step 1198, Loss: {'train/loss': 0.6182146668434143, 'train/grad_norm_pre_clip': 7.412882328033447, 'timing/data': 0.02605604100972414, 'timing/model': 0.9815577995032072, 'learning_rate/qwen_vl_interface': 9.418692495410789e-06, 'learning_rate/action_model': 9.418692495410789e-05, 'epoch': 0.54}) +[2026-06-17 18:36:22,870][starVLA.training.train_starvla][INFO] - Step 1199, Loss: {'train/loss': 0.6773450374603271, 'train/grad_norm_pre_clip': 10.008533477783203, 'timing/data': 0.020253026857972145, 'timing/model': 0.9601737149059772, 'learning_rate/qwen_vl_interface': 9.417655360391277e-06, 'learning_rate/action_model': 9.417655360391277e-05, 'epoch': 0.54}) +[2026-06-17 18:38:35,568][starVLA.training.train_starvla][INFO] - Step 1200, Loss: {'train/loss': 0.7160528302192688, 'train/grad_norm_pre_clip': 11.615374565124512, 'eval/loss': 0.6434138417243958, 'eval/latency_8/loss': 0.6434138417243958, 'eval/demon_fire_precision': 0.9419773095623987, 'eval/demon_fire_recall': 0.9027648337993166, 'eval/demon_fire_f1': 0.9219543147208122, 'eval/demon_fire_rate_model': 0.8583750695603785, 'eval/demon_fire_rate_teacher': 0.8956594323873122, 'eval/demon_left_precision': 0.8817204301075269, 'eval/demon_left_recall': 0.3511777301927195, 'eval/demon_left_f1': 0.5022970903522205, 'eval/demon_left_rate_model': 0.05175292153589316, 'eval/demon_left_rate_teacher': 0.12993878686700056, 'eval/demon_right_precision': 0.751937984496124, 'eval/demon_right_recall': 0.5878787878787879, 'eval/demon_right_f1': 0.6598639455782314, 'eval/demon_right_rate_model': 0.14357262103505844, 'eval/demon_right_rate_teacher': 0.18363939899833054, 'eval/demon_move_f1': 0.5810805179652259, 'eval/demon_cc_f1': 0.7839839723990674, 'eval/latency_8/demon_fire_precision': 0.9419773095623987, 'eval/latency_8/demon_fire_recall': 0.9027648337993166, 'eval/latency_8/demon_fire_f1': 0.9219543147208122, 'eval/latency_8/demon_fire_rate_model': 0.8583750695603785, 'eval/latency_8/demon_fire_rate_teacher': 0.8956594323873122, 'eval/latency_8/demon_left_precision': 0.8817204301075269, 'eval/latency_8/demon_left_recall': 0.3511777301927195, 'eval/latency_8/demon_left_f1': 0.5022970903522205, 'eval/latency_8/demon_left_rate_model': 0.05175292153589316, 'eval/latency_8/demon_left_rate_teacher': 0.12993878686700056, 'eval/latency_8/demon_right_precision': 0.751937984496124, 'eval/latency_8/demon_right_recall': 0.5878787878787879, 'eval/latency_8/demon_right_f1': 0.6598639455782314, 'eval/latency_8/demon_right_rate_model': 0.14357262103505844, 'eval/latency_8/demon_right_rate_teacher': 0.18363939899833054, 'eval/latency_8/demon_move_f1': 0.5810805179652259, 'eval/latency_8/demon_cc_f1': 0.7839839723990674, 'timing/data': 0.019271302968263626, 'timing/model': 0.9973196471109986, 'learning_rate/qwen_vl_interface': 9.416617361412325e-06, 'learning_rate/action_model': 9.416617361412325e-05, 'epoch': 0.54}) +[2026-06-17 18:38:35,636][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_1200_state +[2026-06-17 18:38:35,641][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 18:38:43,169][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_1200_state/model.safetensors +[2026-06-17 18:39:01,009][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_1200_state/optimizer.bin +[2026-06-17 18:39:01,010][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_1200_state/sampler.bin +[2026-06-17 18:39:01,010][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_1200_state/sampler_1.bin +[2026-06-17 18:39:01,012][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_1200_state/random_states_0.pkl +[2026-06-17 18:39:01,016][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1200 +[2026-06-17 18:39:01,016][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 18:39:01,028][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 18:39:05,311][starVLA.training.train_starvla][INFO] - Step 1201, Loss: {'train/loss': 0.4343717694282532, 'train/grad_norm_pre_clip': 8.660582542419434, 'timing/data': 0.015622847713530064, 'timing/model': 1.0082202916964889, 'learning_rate/qwen_vl_interface': 9.415578498689106e-06, 'learning_rate/action_model': 9.415578498689106e-05, 'epoch': 0.54}) +[2026-06-17 18:39:09,284][starVLA.training.train_starvla][INFO] - Step 1202, Loss: {'train/loss': 0.5969285368919373, 'train/grad_norm_pre_clip': 48.11097717285156, 'timing/data': 0.017322976142168045, 'timing/model': 0.9436497958377004, 'learning_rate/qwen_vl_interface': 9.414538772436987e-06, 'learning_rate/action_model': 9.414538772436986e-05, 'epoch': 0.55}) +[2026-06-17 18:39:13,413][starVLA.training.train_starvla][INFO] - Step 1203, Loss: {'train/loss': 0.5571543574333191, 'train/grad_norm_pre_clip': 9.147674560546875, 'timing/data': 0.038409183733165264, 'timing/model': 1.2824665820226073, 'learning_rate/qwen_vl_interface': 9.413498182871496e-06, 'learning_rate/action_model': 9.413498182871496e-05, 'epoch': 0.55}) +[2026-06-17 18:39:17,090][starVLA.training.train_starvla][INFO] - Step 1204, Loss: {'train/loss': 0.6198427081108093, 'train/grad_norm_pre_clip': 12.034276008605957, 'timing/data': 0.020855464041233063, 'timing/model': 0.9462952902540565, 'learning_rate/qwen_vl_interface': 9.412456730208352e-06, 'learning_rate/action_model': 9.412456730208352e-05, 'epoch': 0.55}) +[2026-06-17 18:39:20,988][starVLA.training.train_starvla][INFO] - Step 1205, Loss: {'train/loss': 0.56919264793396, 'train/grad_norm_pre_clip': 9.367598533630371, 'timing/data': 0.018941177055239677, 'timing/model': 0.9534842558205128, 'learning_rate/qwen_vl_interface': 9.41141441466345e-06, 'learning_rate/action_model': 9.41141441466345e-05, 'epoch': 0.55}) +[2026-06-17 18:39:25,190][starVLA.training.train_starvla][INFO] - Step 1206, Loss: {'train/loss': 0.9474202990531921, 'train/grad_norm_pre_clip': 11.533488273620605, 'timing/data': 0.025574488565325737, 'timing/model': 0.9556112261489034, 'learning_rate/qwen_vl_interface': 9.41037123645286e-06, 'learning_rate/action_model': 9.41037123645286e-05, 'epoch': 0.55}) +[2026-06-17 18:39:29,217][starVLA.training.train_starvla][INFO] - Step 1207, Loss: {'train/loss': 0.4869188070297241, 'train/grad_norm_pre_clip': 9.366238594055176, 'timing/data': 0.06478771287947893, 'timing/model': 1.0086820973083377, 'learning_rate/qwen_vl_interface': 9.409327195792838e-06, 'learning_rate/action_model': 9.409327195792838e-05, 'epoch': 0.55}) +[2026-06-17 18:39:33,404][starVLA.training.train_starvla][INFO] - Step 1208, Loss: {'train/loss': 0.6647945642471313, 'train/grad_norm_pre_clip': 11.449868202209473, 'timing/data': 0.01902030035853386, 'timing/model': 0.9606437608599663, 'learning_rate/qwen_vl_interface': 9.408282292899812e-06, 'learning_rate/action_model': 9.408282292899812e-05, 'epoch': 0.55}) +[2026-06-17 18:39:40,851][starVLA.training.train_starvla][INFO] - Step 1209, Loss: {'train/loss': 0.5661048293113708, 'train/grad_norm_pre_clip': 7.157716274261475, 'timing/data': 0.016940240748226643, 'timing/model': 1.794766267761588, 'learning_rate/qwen_vl_interface': 9.407236527990394e-06, 'learning_rate/action_model': 9.407236527990393e-05, 'epoch': 0.55}) +[2026-06-17 18:39:47,713][starVLA.training.train_starvla][INFO] - Step 1210, Loss: {'train/loss': 0.4907541871070862, 'train/grad_norm_pre_clip': 6.851428985595703, 'timing/data': 0.022281873039901257, 'timing/model': 2.485912710428238, 'learning_rate/qwen_vl_interface': 9.406189901281369e-06, 'learning_rate/action_model': 9.406189901281368e-05, 'epoch': 0.55}) +[2026-06-17 18:39:51,604][starVLA.training.train_starvla][INFO] - Step 1211, Loss: {'train/loss': 0.939028799533844, 'train/grad_norm_pre_clip': 12.390317916870117, 'timing/data': 0.018511264584958553, 'timing/model': 0.9607451483607292, 'learning_rate/qwen_vl_interface': 9.405142412989705e-06, 'learning_rate/action_model': 9.405142412989704e-05, 'epoch': 0.55}) +[2026-06-17 18:39:55,590][starVLA.training.train_starvla][INFO] - Step 1212, Loss: {'train/loss': 0.9403783679008484, 'train/grad_norm_pre_clip': 15.14531421661377, 'timing/data': 0.03333662077784538, 'timing/model': 0.9488612543791533, 'learning_rate/qwen_vl_interface': 9.404094063332546e-06, 'learning_rate/action_model': 9.404094063332546e-05, 'epoch': 0.55}) +[2026-06-17 18:39:59,542][starVLA.training.train_starvla][INFO] - Step 1213, Loss: {'train/loss': 0.8829084038734436, 'train/grad_norm_pre_clip': 11.172270774841309, 'timing/data': 0.020366120152175426, 'timing/model': 0.9502584533765912, 'learning_rate/qwen_vl_interface': 9.40304485252722e-06, 'learning_rate/action_model': 9.403044852527219e-05, 'epoch': 0.55}) +[2026-06-17 18:40:03,468][starVLA.training.train_starvla][INFO] - Step 1214, Loss: {'train/loss': 0.8196842670440674, 'train/grad_norm_pre_clip': 11.425325393676758, 'timing/data': 0.03258517012000084, 'timing/model': 0.9506555730476975, 'learning_rate/qwen_vl_interface': 9.401994780791226e-06, 'learning_rate/action_model': 9.401994780791226e-05, 'epoch': 0.55}) +[2026-06-17 18:40:07,707][starVLA.training.train_starvla][INFO] - Step 1215, Loss: {'train/loss': 0.9415087103843689, 'train/grad_norm_pre_clip': 12.689708709716797, 'timing/data': 0.017115630209445953, 'timing/model': 1.2414469560608268, 'learning_rate/qwen_vl_interface': 9.400943848342246e-06, 'learning_rate/action_model': 9.400943848342246e-05, 'epoch': 0.55}) +[2026-06-17 18:40:11,693][starVLA.training.train_starvla][INFO] - Step 1216, Loss: {'train/loss': 0.542946457862854, 'train/grad_norm_pre_clip': 10.122377395629883, 'timing/data': 0.018013288266956806, 'timing/model': 1.234140930697322, 'learning_rate/qwen_vl_interface': 9.39989205539814e-06, 'learning_rate/action_model': 9.39989205539814e-05, 'epoch': 0.55}) +[2026-06-17 18:40:15,366][starVLA.training.train_starvla][INFO] - Step 1217, Loss: {'train/loss': 0.5800166726112366, 'train/grad_norm_pre_clip': 8.819860458374023, 'timing/data': 0.014589261263608932, 'timing/model': 0.9513575723394752, 'learning_rate/qwen_vl_interface': 9.398839402176946e-06, 'learning_rate/action_model': 9.398839402176946e-05, 'epoch': 0.55}) +[2026-06-17 18:40:19,389][starVLA.training.train_starvla][INFO] - Step 1218, Loss: {'train/loss': 0.4731109142303467, 'train/grad_norm_pre_clip': 15.913904190063477, 'timing/data': 0.028503751382231712, 'timing/model': 0.9588631391525269, 'learning_rate/qwen_vl_interface': 9.39778588889688e-06, 'learning_rate/action_model': 9.39778588889688e-05, 'epoch': 0.55}) +[2026-06-17 18:40:23,401][starVLA.training.train_starvla][INFO] - Step 1219, Loss: {'train/loss': 0.5666634440422058, 'train/grad_norm_pre_clip': 8.792806625366211, 'timing/data': 0.01695644110441208, 'timing/model': 0.9651374230161309, 'learning_rate/qwen_vl_interface': 9.396731515776335e-06, 'learning_rate/action_model': 9.396731515776333e-05, 'epoch': 0.55}) +[2026-06-17 18:40:27,433][starVLA.training.train_starvla][INFO] - Step 1220, Loss: {'train/loss': 0.346642404794693, 'train/grad_norm_pre_clip': 5.7005085945129395, 'timing/data': 0.031071672216057777, 'timing/model': 0.9611534206196666, 'learning_rate/qwen_vl_interface': 9.395676283033884e-06, 'learning_rate/action_model': 9.395676283033885e-05, 'epoch': 0.55}) +[2026-06-17 18:40:31,483][starVLA.training.train_starvla][INFO] - Step 1221, Loss: {'train/loss': 0.5937401652336121, 'train/grad_norm_pre_clip': 8.507105827331543, 'timing/data': 0.01901158969849348, 'timing/model': 0.9982982445508242, 'learning_rate/qwen_vl_interface': 9.39462019088828e-06, 'learning_rate/action_model': 9.394620190888279e-05, 'epoch': 0.55}) +[2026-06-17 18:40:35,479][starVLA.training.train_starvla][INFO] - Step 1222, Loss: {'train/loss': 0.6367838978767395, 'train/grad_norm_pre_clip': 9.942267417907715, 'timing/data': 0.017676771618425846, 'timing/model': 1.280022400431335, 'learning_rate/qwen_vl_interface': 9.393563239558449e-06, 'learning_rate/action_model': 9.39356323955845e-05, 'epoch': 0.55}) +[2026-06-17 18:40:39,633][starVLA.training.train_starvla][INFO] - Step 1223, Loss: {'train/loss': 0.7716109752655029, 'train/grad_norm_pre_clip': 9.754006385803223, 'timing/data': 0.020969530567526817, 'timing/model': 1.349707828834653, 'learning_rate/qwen_vl_interface': 9.392505429263501e-06, 'learning_rate/action_model': 9.392505429263501e-05, 'epoch': 0.55}) +[2026-06-17 18:40:43,425][starVLA.training.train_starvla][INFO] - Step 1224, Loss: {'train/loss': 0.8315038681030273, 'train/grad_norm_pre_clip': 10.593161582946777, 'timing/data': 0.017219231463968754, 'timing/model': 0.971348661929369, 'learning_rate/qwen_vl_interface': 9.39144676022272e-06, 'learning_rate/action_model': 9.39144676022272e-05, 'epoch': 0.56}) +[2026-06-17 18:40:47,460][starVLA.training.train_starvla][INFO] - Step 1225, Loss: {'train/loss': 0.7858233451843262, 'train/grad_norm_pre_clip': 10.179128646850586, 'timing/data': 0.018959471955895424, 'timing/model': 0.9519681539386511, 'learning_rate/qwen_vl_interface': 9.390387232655568e-06, 'learning_rate/action_model': 9.390387232655569e-05, 'epoch': 0.56}) +[2026-06-17 18:40:51,463][starVLA.training.train_starvla][INFO] - Step 1226, Loss: {'train/loss': 0.5555593967437744, 'train/grad_norm_pre_clip': 12.050130844116211, 'timing/data': 0.024891387671232224, 'timing/model': 0.9878821289166808, 'learning_rate/qwen_vl_interface': 9.38932684678169e-06, 'learning_rate/action_model': 9.38932684678169e-05, 'epoch': 0.56}) +[2026-06-17 18:40:55,608][starVLA.training.train_starvla][INFO] - Step 1227, Loss: {'train/loss': 0.7010114789009094, 'train/grad_norm_pre_clip': 10.046082496643066, 'timing/data': 0.018342996947467327, 'timing/model': 0.9820122187957168, 'learning_rate/qwen_vl_interface': 9.388265602820901e-06, 'learning_rate/action_model': 9.3882656028209e-05, 'epoch': 0.56}) +[2026-06-17 18:41:00,920][starVLA.training.train_starvla][INFO] - Step 1228, Loss: {'train/loss': 0.523157000541687, 'train/grad_norm_pre_clip': 7.739991188049316, 'timing/data': 0.024820969440042973, 'timing/model': 0.9708938160911202, 'learning_rate/qwen_vl_interface': 9.387203500993199e-06, 'learning_rate/action_model': 9.387203500993199e-05, 'epoch': 0.56}) +[2026-06-17 18:41:05,054][starVLA.training.train_starvla][INFO] - Step 1229, Loss: {'train/loss': 0.6461007595062256, 'train/grad_norm_pre_clip': 7.6909050941467285, 'timing/data': 0.02055140770971775, 'timing/model': 1.2066879887133837, 'learning_rate/qwen_vl_interface': 9.38614054151876e-06, 'learning_rate/action_model': 9.38614054151876e-05, 'epoch': 0.56}) +[2026-06-17 18:41:09,029][starVLA.training.train_starvla][INFO] - Step 1230, Loss: {'train/loss': 0.5225548148155212, 'train/grad_norm_pre_clip': 8.596776962280273, 'timing/data': 0.018912285566329956, 'timing/model': 1.1950803389772773, 'learning_rate/qwen_vl_interface': 9.385076724617937e-06, 'learning_rate/action_model': 9.385076724617935e-05, 'epoch': 0.56}) +[2026-06-17 18:41:12,885][starVLA.training.train_starvla][INFO] - Step 1231, Loss: {'train/loss': 0.73329097032547, 'train/grad_norm_pre_clip': 11.814499855041504, 'timing/data': 0.019084857776761055, 'timing/model': 1.008283719420433, 'learning_rate/qwen_vl_interface': 9.384012050511258e-06, 'learning_rate/action_model': 9.384012050511258e-05, 'epoch': 0.56}) +[2026-06-17 18:41:16,882][starVLA.training.train_starvla][INFO] - Step 1232, Loss: {'train/loss': 0.9096447825431824, 'train/grad_norm_pre_clip': 10.682780265808105, 'timing/data': 0.014815901406109333, 'timing/model': 0.967183068394661, 'learning_rate/qwen_vl_interface': 9.382946519419433e-06, 'learning_rate/action_model': 9.382946519419434e-05, 'epoch': 0.56}) +[2026-06-17 18:41:21,064][starVLA.training.train_starvla][INFO] - Step 1233, Loss: {'train/loss': 0.4541122019290924, 'train/grad_norm_pre_clip': 8.345114707946777, 'timing/data': 0.030533015727996826, 'timing/model': 0.9715467300266027, 'learning_rate/qwen_vl_interface': 9.381880131563347e-06, 'learning_rate/action_model': 9.381880131563346e-05, 'epoch': 0.56}) +[2026-06-17 18:41:25,052][starVLA.training.train_starvla][INFO] - Step 1234, Loss: {'train/loss': 0.48584142327308655, 'train/grad_norm_pre_clip': 9.24991512298584, 'timing/data': 0.020938253961503506, 'timing/model': 1.0166807351633906, 'learning_rate/qwen_vl_interface': 9.380812887164065e-06, 'learning_rate/action_model': 9.380812887164065e-05, 'epoch': 0.56}) +[2026-06-17 18:41:29,078][starVLA.training.train_starvla][INFO] - Step 1235, Loss: {'train/loss': 0.6225655674934387, 'train/grad_norm_pre_clip': 8.388025283813477, 'timing/data': 0.022126474417746067, 'timing/model': 0.9928580485284328, 'learning_rate/qwen_vl_interface': 9.379744786442824e-06, 'learning_rate/action_model': 9.379744786442824e-05, 'epoch': 0.56}) +[2026-06-17 18:41:33,601][starVLA.training.train_starvla][INFO] - Step 1236, Loss: {'train/loss': 0.7972363829612732, 'train/grad_norm_pre_clip': 17.34258460998535, 'timing/data': 0.018177014775574207, 'timing/model': 1.2529313787817955, 'learning_rate/qwen_vl_interface': 9.378675829621047e-06, 'learning_rate/action_model': 9.378675829621046e-05, 'epoch': 0.56}) +[2026-06-17 18:41:37,343][starVLA.training.train_starvla][INFO] - Step 1237, Loss: {'train/loss': 0.5371098518371582, 'train/grad_norm_pre_clip': 8.533910751342773, 'timing/data': 0.02205691486597061, 'timing/model': 0.9439093442633748, 'learning_rate/qwen_vl_interface': 9.377606016920328e-06, 'learning_rate/action_model': 9.377606016920327e-05, 'epoch': 0.56}) +[2026-06-17 18:41:41,550][starVLA.training.train_starvla][INFO] - Step 1238, Loss: {'train/loss': 0.43343162536621094, 'train/grad_norm_pre_clip': 6.926949501037598, 'timing/data': 0.023173149675130844, 'timing/model': 1.0750675974413753, 'learning_rate/qwen_vl_interface': 9.376535348562438e-06, 'learning_rate/action_model': 9.376535348562439e-05, 'epoch': 0.56}) +[2026-06-17 18:41:45,772][starVLA.training.train_starvla][INFO] - Step 1239, Loss: {'train/loss': 0.4225563704967499, 'train/grad_norm_pre_clip': 5.477136611938477, 'timing/data': 0.021976285614073277, 'timing/model': 0.9376035612076521, 'learning_rate/qwen_vl_interface': 9.375463824769332e-06, 'learning_rate/action_model': 9.375463824769333e-05, 'epoch': 0.56}) +[2026-06-17 18:41:49,877][starVLA.training.train_starvla][INFO] - Step 1240, Loss: {'train/loss': 0.41662904620170593, 'train/grad_norm_pre_clip': 8.384719848632812, 'timing/data': 0.016166132874786854, 'timing/model': 0.94317536521703, 'learning_rate/qwen_vl_interface': 9.374391445763135e-06, 'learning_rate/action_model': 9.374391445763135e-05, 'epoch': 0.56}) +[2026-06-17 18:41:53,914][starVLA.training.train_starvla][INFO] - Step 1241, Loss: {'train/loss': 0.5148962736129761, 'train/grad_norm_pre_clip': 7.626434803009033, 'timing/data': 0.032774013467133045, 'timing/model': 1.0131957912817597, 'learning_rate/qwen_vl_interface': 9.373318211766155e-06, 'learning_rate/action_model': 9.373318211766155e-05, 'epoch': 0.56}) +[2026-06-17 18:41:58,211][starVLA.training.train_starvla][INFO] - Step 1242, Loss: {'train/loss': 0.4020664393901825, 'train/grad_norm_pre_clip': 6.560671329498291, 'timing/data': 0.02389808278530836, 'timing/model': 1.545111852698028, 'learning_rate/qwen_vl_interface': 9.37224412300087e-06, 'learning_rate/action_model': 9.37224412300087e-05, 'epoch': 0.56}) +[2026-06-17 18:42:02,226][starVLA.training.train_starvla][INFO] - Step 1243, Loss: {'train/loss': 0.8669010996818542, 'train/grad_norm_pre_clip': 10.533722877502441, 'timing/data': 0.023797334171831608, 'timing/model': 1.1919312262907624, 'learning_rate/qwen_vl_interface': 9.371169179689945e-06, 'learning_rate/action_model': 9.371169179689945e-05, 'epoch': 0.56}) +[2026-06-17 18:42:05,954][starVLA.training.train_starvla][INFO] - Step 1244, Loss: {'train/loss': 0.47078636288642883, 'train/grad_norm_pre_clip': 6.782387733459473, 'timing/data': 0.019254056736826897, 'timing/model': 0.9820886766538024, 'learning_rate/qwen_vl_interface': 9.370093382056213e-06, 'learning_rate/action_model': 9.370093382056214e-05, 'epoch': 0.56}) +[2026-06-17 18:42:10,084][starVLA.training.train_starvla][INFO] - Step 1245, Loss: {'train/loss': 0.7735251784324646, 'train/grad_norm_pre_clip': 8.853731155395508, 'timing/data': 0.01855424325913191, 'timing/model': 0.9623264521360397, 'learning_rate/qwen_vl_interface': 9.369016730322692e-06, 'learning_rate/action_model': 9.36901673032269e-05, 'epoch': 0.56}) +[2026-06-17 18:42:14,460][starVLA.training.train_starvla][INFO] - Step 1246, Loss: {'train/loss': 0.36958110332489014, 'train/grad_norm_pre_clip': 6.716615200042725, 'timing/data': 0.015516746789216995, 'timing/model': 0.9654960343614221, 'learning_rate/qwen_vl_interface': 9.367939224712566e-06, 'learning_rate/action_model': 9.367939224712566e-05, 'epoch': 0.57}) +[2026-06-17 18:42:19,462][starVLA.training.train_starvla][INFO] - Step 1247, Loss: {'train/loss': 0.6531537175178528, 'train/grad_norm_pre_clip': 7.3857622146606445, 'timing/data': 0.024016451090574265, 'timing/model': 0.9603086840361357, 'learning_rate/qwen_vl_interface': 9.366860865449213e-06, 'learning_rate/action_model': 9.366860865449212e-05, 'epoch': 0.57}) +[2026-06-17 18:42:23,645][starVLA.training.train_starvla][INFO] - Step 1248, Loss: {'train/loss': 0.8078650236129761, 'train/grad_norm_pre_clip': 12.100936889648438, 'timing/data': 0.018190906383097172, 'timing/model': 0.9596708975732327, 'learning_rate/qwen_vl_interface': 9.365781652756168e-06, 'learning_rate/action_model': 9.365781652756168e-05, 'epoch': 0.57}) +[2026-06-17 18:42:27,642][starVLA.training.train_starvla][INFO] - Step 1249, Loss: {'train/loss': 0.5613731741905212, 'train/grad_norm_pre_clip': 9.073746681213379, 'timing/data': 0.020423111505806446, 'timing/model': 1.2245990317314863, 'learning_rate/qwen_vl_interface': 9.36470158685716e-06, 'learning_rate/action_model': 9.36470158685716e-05, 'epoch': 0.57}) +[2026-06-17 18:42:31,646][starVLA.training.train_starvla][INFO] - Step 1250, Loss: {'train/loss': 0.5187091827392578, 'train/grad_norm_pre_clip': 7.316363334655762, 'timing/data': 0.025606569834053516, 'timing/model': 1.1956475609913468, 'learning_rate/qwen_vl_interface': 9.363620667976085e-06, 'learning_rate/action_model': 9.363620667976085e-05, 'epoch': 0.57}) +[2026-06-17 18:42:35,414][starVLA.training.train_starvla][INFO] - Step 1251, Loss: {'train/loss': 0.5699875950813293, 'train/grad_norm_pre_clip': 30.803239822387695, 'timing/data': 0.02162762638181448, 'timing/model': 0.9604836609214544, 'learning_rate/qwen_vl_interface': 9.362538896337019e-06, 'learning_rate/action_model': 9.362538896337019e-05, 'epoch': 0.57}) +[2026-06-17 18:42:39,910][starVLA.training.train_starvla][INFO] - Step 1252, Loss: {'train/loss': 0.1830883026123047, 'train/grad_norm_pre_clip': 7.969141483306885, 'timing/data': 0.02967823389917612, 'timing/model': 1.039568493142724, 'learning_rate/qwen_vl_interface': 9.361456272164215e-06, 'learning_rate/action_model': 9.361456272164214e-05, 'epoch': 0.57}) +[2026-06-17 18:42:43,852][starVLA.training.train_starvla][INFO] - Step 1253, Loss: {'train/loss': 0.35463428497314453, 'train/grad_norm_pre_clip': 7.502564430236816, 'timing/data': 0.017995391972362995, 'timing/model': 0.9834518088027835, 'learning_rate/qwen_vl_interface': 9.3603727956821e-06, 'learning_rate/action_model': 9.3603727956821e-05, 'epoch': 0.57}) +[2026-06-17 18:42:47,841][starVLA.training.train_starvla][INFO] - Step 1254, Loss: {'train/loss': 0.6811406016349792, 'train/grad_norm_pre_clip': 12.159625053405762, 'timing/data': 0.022071179002523422, 'timing/model': 0.9695991724729538, 'learning_rate/qwen_vl_interface': 9.359288467115283e-06, 'learning_rate/action_model': 9.359288467115283e-05, 'epoch': 0.57}) +[2026-06-17 18:42:51,907][starVLA.training.train_starvla][INFO] - Step 1255, Loss: {'train/loss': 0.5059299468994141, 'train/grad_norm_pre_clip': 10.191669464111328, 'timing/data': 0.0182638606056571, 'timing/model': 1.0213861931115389, 'learning_rate/qwen_vl_interface': 9.358203286688544e-06, 'learning_rate/action_model': 9.358203286688544e-05, 'epoch': 0.57}) +[2026-06-17 18:42:55,942][starVLA.training.train_starvla][INFO] - Step 1256, Loss: {'train/loss': 0.7670318484306335, 'train/grad_norm_pre_clip': 11.074259757995605, 'timing/data': 0.028559592552483082, 'timing/model': 1.2699429178610444, 'learning_rate/qwen_vl_interface': 9.357117254626844e-06, 'learning_rate/action_model': 9.357117254626844e-05, 'epoch': 0.57}) +[2026-06-17 18:42:59,760][starVLA.training.train_starvla][INFO] - Step 1257, Loss: {'train/loss': 0.6489508748054504, 'train/grad_norm_pre_clip': 10.313992500305176, 'timing/data': 0.01660486590117216, 'timing/model': 0.9635985130444169, 'learning_rate/qwen_vl_interface': 9.356030371155316e-06, 'learning_rate/action_model': 9.356030371155316e-05, 'epoch': 0.57}) +[2026-06-17 18:43:03,760][starVLA.training.train_starvla][INFO] - Step 1258, Loss: {'train/loss': 0.6019071936607361, 'train/grad_norm_pre_clip': 10.688467979431152, 'timing/data': 0.0262615280225873, 'timing/model': 0.9768004296347499, 'learning_rate/qwen_vl_interface': 9.354942636499275e-06, 'learning_rate/action_model': 9.354942636499276e-05, 'epoch': 0.57}) +[2026-06-17 18:43:07,743][starVLA.training.train_starvla][INFO] - Step 1259, Loss: {'train/loss': 0.8018307089805603, 'train/grad_norm_pre_clip': 13.73625373840332, 'timing/data': 0.0195171982049942, 'timing/model': 0.9753978261724114, 'learning_rate/qwen_vl_interface': 9.35385405088421e-06, 'learning_rate/action_model': 9.353854050884208e-05, 'epoch': 0.57}) +[2026-06-17 18:43:12,239][starVLA.training.train_starvla][INFO] - Step 1260, Loss: {'train/loss': 0.855717122554779, 'train/grad_norm_pre_clip': 9.231237411499023, 'timing/data': 0.022120515815913677, 'timing/model': 1.1820591362193227, 'learning_rate/qwen_vl_interface': 9.352764614535782e-06, 'learning_rate/action_model': 9.352764614535781e-05, 'epoch': 0.57}) +[2026-06-17 18:43:16,294][starVLA.training.train_starvla][INFO] - Step 1261, Loss: {'train/loss': 0.5462197661399841, 'train/grad_norm_pre_clip': 10.284564971923828, 'timing/data': 0.024082419462502003, 'timing/model': 1.012795795686543, 'learning_rate/qwen_vl_interface': 9.351674327679837e-06, 'learning_rate/action_model': 9.351674327679836e-05, 'epoch': 0.57}) +[2026-06-17 18:43:23,513][starVLA.training.train_starvla][INFO] - Step 1262, Loss: {'train/loss': 0.801207423210144, 'train/grad_norm_pre_clip': 16.520111083984375, 'timing/data': 0.01998744998127222, 'timing/model': 0.9769765613600612, 'learning_rate/qwen_vl_interface': 9.350583190542387e-06, 'learning_rate/action_model': 9.350583190542387e-05, 'epoch': 0.57}) +[2026-06-17 18:43:27,924][starVLA.training.train_starvla][INFO] - Step 1263, Loss: {'train/loss': 0.5374189615249634, 'train/grad_norm_pre_clip': 9.60189151763916, 'timing/data': 0.016239079646766186, 'timing/model': 1.294438000768423, 'learning_rate/qwen_vl_interface': 9.34949120334963e-06, 'learning_rate/action_model': 9.34949120334963e-05, 'epoch': 0.57}) +[2026-06-17 18:43:31,658][starVLA.training.train_starvla][INFO] - Step 1264, Loss: {'train/loss': 0.6580865979194641, 'train/grad_norm_pre_clip': 12.055009841918945, 'timing/data': 0.02110653556883335, 'timing/model': 0.9556828588247299, 'learning_rate/qwen_vl_interface': 9.348398366327938e-06, 'learning_rate/action_model': 9.348398366327937e-05, 'epoch': 0.57}) +[2026-06-17 18:43:35,729][starVLA.training.train_starvla][INFO] - Step 1265, Loss: {'train/loss': 0.8642747402191162, 'train/grad_norm_pre_clip': 11.119536399841309, 'timing/data': 0.017878448590636253, 'timing/model': 0.9663463179022074, 'learning_rate/qwen_vl_interface': 9.347304679703852e-06, 'learning_rate/action_model': 9.347304679703851e-05, 'epoch': 0.57}) +[2026-06-17 18:43:39,993][starVLA.training.train_starvla][INFO] - Step 1266, Loss: {'train/loss': 0.4954146444797516, 'train/grad_norm_pre_clip': 8.338380813598633, 'timing/data': 0.024898244999349117, 'timing/model': 1.187305822968483, 'learning_rate/qwen_vl_interface': 9.346210143704099e-06, 'learning_rate/action_model': 9.346210143704098e-05, 'epoch': 0.57}) +[2026-06-17 18:43:44,057][starVLA.training.train_starvla][INFO] - Step 1267, Loss: {'train/loss': 0.5876318216323853, 'train/grad_norm_pre_clip': 9.460658073425293, 'timing/data': 0.016846194863319397, 'timing/model': 0.9938032506033778, 'learning_rate/qwen_vl_interface': 9.345114758555573e-06, 'learning_rate/action_model': 9.345114758555573e-05, 'epoch': 0.57}) +[2026-06-17 18:43:48,461][starVLA.training.train_starvla][INFO] - Step 1268, Loss: {'train/loss': 0.44040611386299133, 'train/grad_norm_pre_clip': 15.064709663391113, 'timing/data': 0.019935956224799156, 'timing/model': 0.9763121856376529, 'learning_rate/qwen_vl_interface': 9.344018524485352e-06, 'learning_rate/action_model': 9.344018524485351e-05, 'epoch': 0.58}) +[2026-06-17 18:43:52,818][starVLA.training.train_starvla][INFO] - Step 1269, Loss: {'train/loss': 0.4998946189880371, 'train/grad_norm_pre_clip': 8.140180587768555, 'timing/data': 0.019310555420815945, 'timing/model': 1.1690156618133187, 'learning_rate/qwen_vl_interface': 9.342921441720686e-06, 'learning_rate/action_model': 9.342921441720685e-05, 'epoch': 0.58}) +[2026-06-17 18:43:56,924][starVLA.training.train_starvla][INFO] - Step 1270, Loss: {'train/loss': 0.6742874383926392, 'train/grad_norm_pre_clip': 9.372014045715332, 'timing/data': 0.023744996637105942, 'timing/model': 1.3425570596009493, 'learning_rate/qwen_vl_interface': 9.341823510489e-06, 'learning_rate/action_model': 9.341823510489e-05, 'epoch': 0.58}) +[2026-06-17 18:44:00,679][starVLA.training.train_starvla][INFO] - Step 1271, Loss: {'train/loss': 0.3700692355632782, 'train/grad_norm_pre_clip': 7.7887187004089355, 'timing/data': 0.017897476442158222, 'timing/model': 1.0025107245892286, 'learning_rate/qwen_vl_interface': 9.340724731017898e-06, 'learning_rate/action_model': 9.340724731017898e-05, 'epoch': 0.58}) +[2026-06-17 18:44:04,561][starVLA.training.train_starvla][INFO] - Step 1272, Loss: {'train/loss': 0.7427737712860107, 'train/grad_norm_pre_clip': 9.128988265991211, 'timing/data': 0.015738610178232193, 'timing/model': 0.9447905598208308, 'learning_rate/qwen_vl_interface': 9.339625103535158e-06, 'learning_rate/action_model': 9.339625103535156e-05, 'epoch': 0.58}) +[2026-06-17 18:44:08,555][starVLA.training.train_starvla][INFO] - Step 1273, Loss: {'train/loss': 0.33497950434684753, 'train/grad_norm_pre_clip': 7.529979228973389, 'timing/data': 0.016417263075709343, 'timing/model': 0.9426549728959799, 'learning_rate/qwen_vl_interface': 9.338524628268734e-06, 'learning_rate/action_model': 9.338524628268733e-05, 'epoch': 0.58}) +[2026-06-17 18:44:12,533][starVLA.training.train_starvla][INFO] - Step 1274, Loss: {'train/loss': 0.5835530161857605, 'train/grad_norm_pre_clip': 8.38261890411377, 'timing/data': 0.026808726601302624, 'timing/model': 0.9575253752991557, 'learning_rate/qwen_vl_interface': 9.337423305446753e-06, 'learning_rate/action_model': 9.337423305446753e-05, 'epoch': 0.58}) +[2026-06-17 18:44:16,986][starVLA.training.train_starvla][INFO] - Step 1275, Loss: {'train/loss': 0.7606559991836548, 'train/grad_norm_pre_clip': 13.986807823181152, 'timing/data': 0.023805412463843822, 'timing/model': 1.2149256216362119, 'learning_rate/qwen_vl_interface': 9.336321135297525e-06, 'learning_rate/action_model': 9.336321135297525e-05, 'epoch': 0.58}) +[2026-06-17 18:44:20,932][starVLA.training.train_starvla][INFO] - Step 1276, Loss: {'train/loss': 0.6272917985916138, 'train/grad_norm_pre_clip': 8.27763557434082, 'timing/data': 0.021346847526729107, 'timing/model': 1.2315034484490752, 'learning_rate/qwen_vl_interface': 9.335218118049526e-06, 'learning_rate/action_model': 9.335218118049526e-05, 'epoch': 0.58}) +[2026-06-17 18:44:25,288][starVLA.training.train_starvla][INFO] - Step 1277, Loss: {'train/loss': 0.6070529222488403, 'train/grad_norm_pre_clip': 9.228611946105957, 'timing/data': 0.02455082442611456, 'timing/model': 1.301289750263095, 'learning_rate/qwen_vl_interface': 9.334114253931417e-06, 'learning_rate/action_model': 9.334114253931416e-05, 'epoch': 0.58}) +[2026-06-17 18:44:29,044][starVLA.training.train_starvla][INFO] - Step 1278, Loss: {'train/loss': 0.6637396812438965, 'train/grad_norm_pre_clip': 7.253528118133545, 'timing/data': 0.029115745797753334, 'timing/model': 0.9451320767402649, 'learning_rate/qwen_vl_interface': 9.333009543172027e-06, 'learning_rate/action_model': 9.333009543172027e-05, 'epoch': 0.58}) +[2026-06-17 18:44:33,020][starVLA.training.train_starvla][INFO] - Step 1279, Loss: {'train/loss': 0.7917562127113342, 'train/grad_norm_pre_clip': 10.094331741333008, 'timing/data': 0.018735921010375023, 'timing/model': 0.9577444801107049, 'learning_rate/qwen_vl_interface': 9.331903986000365e-06, 'learning_rate/action_model': 9.331903986000365e-05, 'epoch': 0.58}) +[2026-06-17 18:44:36,975][starVLA.training.train_starvla][INFO] - Step 1280, Loss: {'train/loss': 0.6617935299873352, 'train/grad_norm_pre_clip': 10.105521202087402, 'timing/data': 0.019002617336809635, 'timing/model': 0.9536510352045298, 'learning_rate/qwen_vl_interface': 9.330797582645615e-06, 'learning_rate/action_model': 9.330797582645615e-05, 'epoch': 0.58}) +[2026-06-17 18:44:40,891][starVLA.training.train_starvla][INFO] - Step 1281, Loss: {'train/loss': 0.566601037979126, 'train/grad_norm_pre_clip': 8.80603313446045, 'timing/data': 0.030230792239308357, 'timing/model': 0.9641060475260019, 'learning_rate/qwen_vl_interface': 9.329690333337137e-06, 'learning_rate/action_model': 9.329690333337137e-05, 'epoch': 0.58}) +[2026-06-17 18:44:44,762][starVLA.training.train_starvla][INFO] - Step 1282, Loss: {'train/loss': 0.730780303478241, 'train/grad_norm_pre_clip': 8.476170539855957, 'timing/data': 0.023216556757688522, 'timing/model': 0.9857375677675009, 'learning_rate/qwen_vl_interface': 9.328582238304461e-06, 'learning_rate/action_model': 9.328582238304461e-05, 'epoch': 0.58}) +[2026-06-17 18:44:48,728][starVLA.training.train_starvla][INFO] - Step 1283, Loss: {'train/loss': 0.42294901609420776, 'train/grad_norm_pre_clip': 6.166253566741943, 'timing/data': 0.01751885563135147, 'timing/model': 1.1261096224188805, 'learning_rate/qwen_vl_interface': 9.327473297777297e-06, 'learning_rate/action_model': 9.327473297777297e-05, 'epoch': 0.58}) +[2026-06-17 18:44:52,378][starVLA.training.train_starvla][INFO] - Step 1284, Loss: {'train/loss': 0.8429898619651794, 'train/grad_norm_pre_clip': 9.447173118591309, 'timing/data': 0.027325918897986412, 'timing/model': 0.9463086118921638, 'learning_rate/qwen_vl_interface': 9.326363511985533e-06, 'learning_rate/action_model': 9.326363511985533e-05, 'epoch': 0.58}) +[2026-06-17 18:44:56,345][starVLA.training.train_starvla][INFO] - Step 1285, Loss: {'train/loss': 0.4864683151245117, 'train/grad_norm_pre_clip': 7.48323392868042, 'timing/data': 0.02098178956657648, 'timing/model': 0.9423638721928, 'learning_rate/qwen_vl_interface': 9.325252881159226e-06, 'learning_rate/action_model': 9.325252881159226e-05, 'epoch': 0.58}) +[2026-06-17 18:45:00,218][starVLA.training.train_starvla][INFO] - Step 1286, Loss: {'train/loss': 0.8835374116897583, 'train/grad_norm_pre_clip': 10.650928497314453, 'timing/data': 0.024476464837789536, 'timing/model': 0.9508879687637091, 'learning_rate/qwen_vl_interface': 9.324141405528612e-06, 'learning_rate/action_model': 9.324141405528612e-05, 'epoch': 0.58}) +[2026-06-17 18:45:04,065][starVLA.training.train_starvla][INFO] - Step 1287, Loss: {'train/loss': 0.6332574486732483, 'train/grad_norm_pre_clip': 7.999398231506348, 'timing/data': 0.01824228558689356, 'timing/model': 0.9556073267012835, 'learning_rate/qwen_vl_interface': 9.3230290853241e-06, 'learning_rate/action_model': 9.3230290853241e-05, 'epoch': 0.58}) +[2026-06-17 18:45:07,962][starVLA.training.train_starvla][INFO] - Step 1288, Loss: {'train/loss': 0.5261324048042297, 'train/grad_norm_pre_clip': 9.089219093322754, 'timing/data': 0.01598718948662281, 'timing/model': 0.9532978041097522, 'learning_rate/qwen_vl_interface': 9.321915920776278e-06, 'learning_rate/action_model': 9.321915920776278e-05, 'epoch': 0.58}) +[2026-06-17 18:45:11,868][starVLA.training.train_starvla][INFO] - Step 1289, Loss: {'train/loss': 0.7903612852096558, 'train/grad_norm_pre_clip': 9.918360710144043, 'timing/data': 0.017378992401063442, 'timing/model': 1.1557915704324841, 'learning_rate/qwen_vl_interface': 9.320801912115901e-06, 'learning_rate/action_model': 9.320801912115901e-05, 'epoch': 0.58}) +[2026-06-17 18:45:15,824][starVLA.training.train_starvla][INFO] - Step 1290, Loss: {'train/loss': 0.4454125761985779, 'train/grad_norm_pre_clip': 8.004730224609375, 'timing/data': 0.013578197918832302, 'timing/model': 1.1247535524889827, 'learning_rate/qwen_vl_interface': 9.319687059573909e-06, 'learning_rate/action_model': 9.319687059573909e-05, 'epoch': 0.59}) +[2026-06-17 18:45:19,466][starVLA.training.train_starvla][INFO] - Step 1291, Loss: {'train/loss': 0.4980611801147461, 'train/grad_norm_pre_clip': 8.303634643554688, 'timing/data': 0.02351730316877365, 'timing/model': 0.9308768892660737, 'learning_rate/qwen_vl_interface': 9.318571363381408e-06, 'learning_rate/action_model': 9.318571363381408e-05, 'epoch': 0.59}) +[2026-06-17 18:45:23,377][starVLA.training.train_starvla][INFO] - Step 1292, Loss: {'train/loss': 0.3357372283935547, 'train/grad_norm_pre_clip': 11.136943817138672, 'timing/data': 0.017956345342099667, 'timing/model': 0.9631450576707721, 'learning_rate/qwen_vl_interface': 9.317454823769686e-06, 'learning_rate/action_model': 9.317454823769687e-05, 'epoch': 0.59}) +[2026-06-17 18:45:27,358][starVLA.training.train_starvla][INFO] - Step 1293, Loss: {'train/loss': 0.49280261993408203, 'train/grad_norm_pre_clip': 7.867564678192139, 'timing/data': 0.02808730397373438, 'timing/model': 0.9859162084758282, 'learning_rate/qwen_vl_interface': 9.316337440970206e-06, 'learning_rate/action_model': 9.316337440970206e-05, 'epoch': 0.59}) +[2026-06-17 18:45:31,270][starVLA.training.train_starvla][INFO] - Step 1294, Loss: {'train/loss': 1.144805669784546, 'train/grad_norm_pre_clip': 17.900806427001953, 'timing/data': 0.01831452827900648, 'timing/model': 0.9510761937126517, 'learning_rate/qwen_vl_interface': 9.315219215214596e-06, 'learning_rate/action_model': 9.315219215214595e-05, 'epoch': 0.59}) +[2026-06-17 18:45:35,386][starVLA.training.train_starvla][INFO] - Step 1295, Loss: {'train/loss': 0.9200522303581238, 'train/grad_norm_pre_clip': 20.881132125854492, 'timing/data': 0.02212437614798546, 'timing/model': 1.0044203186407685, 'learning_rate/qwen_vl_interface': 9.314100146734668e-06, 'learning_rate/action_model': 9.314100146734668e-05, 'epoch': 0.59}) +[2026-06-17 18:45:39,333][starVLA.training.train_starvla][INFO] - Step 1296, Loss: {'train/loss': 0.46029311418533325, 'train/grad_norm_pre_clip': 7.1330246925354, 'timing/data': 0.015470602549612522, 'timing/model': 1.1671423139050603, 'learning_rate/qwen_vl_interface': 9.312980235762408e-06, 'learning_rate/action_model': 9.312980235762408e-05, 'epoch': 0.59}) +[2026-06-17 18:45:43,179][starVLA.training.train_starvla][INFO] - Step 1297, Loss: {'train/loss': 0.505607545375824, 'train/grad_norm_pre_clip': 9.726651191711426, 'timing/data': 0.019878799095749855, 'timing/model': 1.1447926796972752, 'learning_rate/qwen_vl_interface': 9.311859482529971e-06, 'learning_rate/action_model': 9.31185948252997e-05, 'epoch': 0.59}) +[2026-06-17 18:45:46,895][starVLA.training.train_starvla][INFO] - Step 1298, Loss: {'train/loss': 0.6783835291862488, 'train/grad_norm_pre_clip': 11.644810676574707, 'timing/data': 0.02473765704780817, 'timing/model': 0.9735695160925388, 'learning_rate/qwen_vl_interface': 9.310737887269694e-06, 'learning_rate/action_model': 9.310737887269694e-05, 'epoch': 0.59}) +[2026-06-17 18:45:51,497][starVLA.training.train_starvla][INFO] - Step 1299, Loss: {'train/loss': 0.7393192052841187, 'train/grad_norm_pre_clip': 10.793664932250977, 'timing/data': 0.026349982246756554, 'timing/model': 0.9558358518406749, 'learning_rate/qwen_vl_interface': 9.309615450214084e-06, 'learning_rate/action_model': 9.309615450214084e-05, 'epoch': 0.59}) +[2026-06-17 18:48:14,662][starVLA.training.train_starvla][INFO] - Step 1300, Loss: {'train/loss': 0.5699867606163025, 'train/grad_norm_pre_clip': 7.183804988861084, 'eval/loss': 0.5667157173156738, 'eval/latency_8/loss': 0.5667157173156738, 'eval/demon_fire_precision': 0.9211143695014663, 'eval/demon_fire_recall': 0.9757688723205965, 'eval/demon_fire_f1': 0.9476542464926837, 'eval/demon_fire_rate_model': 0.9488035614913745, 'eval/demon_fire_rate_teacher': 0.8956594323873122, 'eval/demon_left_precision': 0.7637540453074434, 'eval/demon_left_recall': 0.5053533190578159, 'eval/demon_left_f1': 0.6082474226804124, 'eval/demon_left_rate_model': 0.08597662771285476, 'eval/demon_left_rate_teacher': 0.12993878686700056, 'eval/demon_right_precision': 0.6068476977567887, 'eval/demon_right_recall': 0.7787878787878788, 'eval/demon_right_f1': 0.6821499668214998, 'eval/demon_right_rate_model': 0.23567056204785755, 'eval/demon_right_rate_teacher': 0.18363939899833054, 'eval/demon_move_f1': 0.645198694750956, 'eval/demon_cc_f1': 0.8308138403420025, 'eval/latency_8/demon_fire_precision': 0.9211143695014663, 'eval/latency_8/demon_fire_recall': 0.9757688723205965, 'eval/latency_8/demon_fire_f1': 0.9476542464926837, 'eval/latency_8/demon_fire_rate_model': 0.9488035614913745, 'eval/latency_8/demon_fire_rate_teacher': 0.8956594323873122, 'eval/latency_8/demon_left_precision': 0.7637540453074434, 'eval/latency_8/demon_left_recall': 0.5053533190578159, 'eval/latency_8/demon_left_f1': 0.6082474226804124, 'eval/latency_8/demon_left_rate_model': 0.08597662771285476, 'eval/latency_8/demon_left_rate_teacher': 0.12993878686700056, 'eval/latency_8/demon_right_precision': 0.6068476977567887, 'eval/latency_8/demon_right_recall': 0.7787878787878788, 'eval/latency_8/demon_right_f1': 0.6821499668214998, 'eval/latency_8/demon_right_rate_model': 0.23567056204785755, 'eval/latency_8/demon_right_rate_teacher': 0.18363939899833054, 'eval/latency_8/demon_move_f1': 0.645198694750956, 'eval/latency_8/demon_cc_f1': 0.8308138403420025, 'timing/data': 0.023726669140160084, 'timing/model': 0.9660188229754567, 'learning_rate/qwen_vl_interface': 9.308492171595821e-06, 'learning_rate/action_model': 9.30849217159582e-05, 'epoch': 0.59}) +[2026-06-17 18:48:14,667][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_1300_state +[2026-06-17 18:48:14,672][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 18:48:21,739][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_1300_state/model.safetensors +[2026-06-17 18:48:39,981][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_1300_state/optimizer.bin +[2026-06-17 18:48:39,982][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_1300_state/sampler.bin +[2026-06-17 18:48:39,982][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_1300_state/sampler_1.bin +[2026-06-17 18:48:39,985][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_1300_state/random_states_0.pkl +[2026-06-17 18:48:39,990][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1300 +[2026-06-17 18:48:39,990][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 18:48:40,003][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 18:48:44,585][starVLA.training.train_starvla][INFO] - Step 1301, Loss: {'train/loss': 0.49749526381492615, 'train/grad_norm_pre_clip': 7.932337760925293, 'timing/data': 0.01915423758327961, 'timing/model': 1.4343860866501927, 'learning_rate/qwen_vl_interface': 9.307368051647765e-06, 'learning_rate/action_model': 9.307368051647765e-05, 'epoch': 0.59}) +[2026-06-17 18:48:48,575][starVLA.training.train_starvla][INFO] - Step 1302, Loss: {'train/loss': 0.7759802341461182, 'train/grad_norm_pre_clip': 9.718587875366211, 'timing/data': 0.020391782745718956, 'timing/model': 0.9787918534129858, 'learning_rate/qwen_vl_interface': 9.306243090602945e-06, 'learning_rate/action_model': 9.306243090602946e-05, 'epoch': 0.59}) +[2026-06-17 18:48:52,481][starVLA.training.train_starvla][INFO] - Step 1303, Loss: {'train/loss': 0.6909197568893433, 'train/grad_norm_pre_clip': 11.146767616271973, 'timing/data': 0.01973326224833727, 'timing/model': 0.9359370972961187, 'learning_rate/qwen_vl_interface': 9.30511728869457e-06, 'learning_rate/action_model': 9.305117288694569e-05, 'epoch': 0.59}) +[2026-06-17 18:48:56,971][starVLA.training.train_starvla][INFO] - Step 1304, Loss: {'train/loss': 0.6906651854515076, 'train/grad_norm_pre_clip': 17.126514434814453, 'timing/data': 0.018347864039242268, 'timing/model': 1.4993915371596813, 'learning_rate/qwen_vl_interface': 9.303990646156015e-06, 'learning_rate/action_model': 9.303990646156014e-05, 'epoch': 0.59}) +[2026-06-17 18:49:01,072][starVLA.training.train_starvla][INFO] - Step 1305, Loss: {'train/loss': 0.7609407901763916, 'train/grad_norm_pre_clip': 10.10381031036377, 'timing/data': 0.017059817910194397, 'timing/model': 1.1327461283653975, 'learning_rate/qwen_vl_interface': 9.302863163220837e-06, 'learning_rate/action_model': 9.302863163220837e-05, 'epoch': 0.59}) +[2026-06-17 18:49:08,035][starVLA.training.train_starvla][INFO] - Step 1306, Loss: {'train/loss': 0.7462593913078308, 'train/grad_norm_pre_clip': 24.481924057006836, 'timing/data': 0.017378827556967735, 'timing/model': 0.9798948857933283, 'learning_rate/qwen_vl_interface': 9.301734840122765e-06, 'learning_rate/action_model': 9.301734840122766e-05, 'epoch': 0.59}) +[2026-06-17 18:49:14,592][starVLA.training.train_starvla][INFO] - Step 1307, Loss: {'train/loss': 0.6642906665802002, 'train/grad_norm_pre_clip': 9.043831825256348, 'timing/data': 0.019342999905347824, 'timing/model': 0.9545270260423422, 'learning_rate/qwen_vl_interface': 9.300605677095701e-06, 'learning_rate/action_model': 9.3006056770957e-05, 'epoch': 0.59}) +[2026-06-17 18:49:19,970][starVLA.training.train_starvla][INFO] - Step 1308, Loss: {'train/loss': 0.48478761315345764, 'train/grad_norm_pre_clip': 8.4385404586792, 'timing/data': 0.01924269925802946, 'timing/model': 0.9579718364402652, 'learning_rate/qwen_vl_interface': 9.299475674373722e-06, 'learning_rate/action_model': 9.299475674373721e-05, 'epoch': 0.59}) +[2026-06-17 18:49:23,898][starVLA.training.train_starvla][INFO] - Step 1309, Loss: {'train/loss': 0.6198195815086365, 'train/grad_norm_pre_clip': 9.195028305053711, 'timing/data': 0.01946841087192297, 'timing/model': 0.9718673434108496, 'learning_rate/qwen_vl_interface': 9.298344832191076e-06, 'learning_rate/action_model': 9.298344832191076e-05, 'epoch': 0.59}) +[2026-06-17 18:49:28,041][starVLA.training.train_starvla][INFO] - Step 1310, Loss: {'train/loss': 0.445101797580719, 'train/grad_norm_pre_clip': 9.697922706604004, 'timing/data': 0.021923585794866085, 'timing/model': 0.9288411261513829, 'learning_rate/qwen_vl_interface': 9.297213150782192e-06, 'learning_rate/action_model': 9.297213150782192e-05, 'epoch': 0.59}) +[2026-06-17 18:49:31,945][starVLA.training.train_starvla][INFO] - Step 1311, Loss: {'train/loss': 0.405523419380188, 'train/grad_norm_pre_clip': 7.83071231842041, 'timing/data': 0.025212284177541733, 'timing/model': 1.1345022339373827, 'learning_rate/qwen_vl_interface': 9.296080630381666e-06, 'learning_rate/action_model': 9.296080630381666e-05, 'epoch': 0.59}) +[2026-06-17 18:49:35,895][starVLA.training.train_starvla][INFO] - Step 1312, Loss: {'train/loss': 0.7632311582565308, 'train/grad_norm_pre_clip': 16.444904327392578, 'timing/data': 0.02487017586827278, 'timing/model': 1.11944860778749, 'learning_rate/qwen_vl_interface': 9.294947271224271e-06, 'learning_rate/action_model': 9.294947271224272e-05, 'epoch': 0.6}) +[2026-06-17 18:49:39,557][starVLA.training.train_starvla][INFO] - Step 1313, Loss: {'train/loss': 0.8950140476226807, 'train/grad_norm_pre_clip': 12.941789627075195, 'timing/data': 0.019597874023020267, 'timing/model': 0.9403367759659886, 'learning_rate/qwen_vl_interface': 9.293813073544957e-06, 'learning_rate/action_model': 9.293813073544956e-05, 'epoch': 0.6}) +[2026-06-17 18:49:43,461][starVLA.training.train_starvla][INFO] - Step 1314, Loss: {'train/loss': 0.46613383293151855, 'train/grad_norm_pre_clip': 8.590618133544922, 'timing/data': 0.01695441547781229, 'timing/model': 0.9333980483934283, 'learning_rate/qwen_vl_interface': 9.292678037578838e-06, 'learning_rate/action_model': 9.292678037578837e-05, 'epoch': 0.6}) +[2026-06-17 18:49:47,493][starVLA.training.train_starvla][INFO] - Step 1315, Loss: {'train/loss': 0.37302178144454956, 'train/grad_norm_pre_clip': 8.189010620117188, 'timing/data': 0.018514269031584263, 'timing/model': 0.9799051275476813, 'learning_rate/qwen_vl_interface': 9.291542163561214e-06, 'learning_rate/action_model': 9.291542163561213e-05, 'epoch': 0.6}) +[2026-06-17 18:49:51,452][starVLA.training.train_starvla][INFO] - Step 1316, Loss: {'train/loss': 0.6508708596229553, 'train/grad_norm_pre_clip': 20.524263381958008, 'timing/data': 0.015640605241060257, 'timing/model': 0.9761633081361651, 'learning_rate/qwen_vl_interface': 9.29040545172755e-06, 'learning_rate/action_model': 9.290405451727549e-05, 'epoch': 0.6}) +[2026-06-17 18:49:55,434][starVLA.training.train_starvla][INFO] - Step 1317, Loss: {'train/loss': 0.8270636200904846, 'train/grad_norm_pre_clip': 10.697782516479492, 'timing/data': 0.018905766308307648, 'timing/model': 0.942862318828702, 'learning_rate/qwen_vl_interface': 9.289267902313486e-06, 'learning_rate/action_model': 9.289267902313487e-05, 'epoch': 0.6}) +[2026-06-17 18:49:59,296][starVLA.training.train_starvla][INFO] - Step 1318, Loss: {'train/loss': 0.5553449988365173, 'train/grad_norm_pre_clip': 9.626011848449707, 'timing/data': 0.017387602478265762, 'timing/model': 1.1592066893354058, 'learning_rate/qwen_vl_interface': 9.288129515554843e-06, 'learning_rate/action_model': 9.288129515554843e-05, 'epoch': 0.6}) +[2026-06-17 18:50:03,005][starVLA.training.train_starvla][INFO] - Step 1319, Loss: {'train/loss': 0.6093887686729431, 'train/grad_norm_pre_clip': 9.694363594055176, 'timing/data': 0.021085425280034542, 'timing/model': 0.9531191429123282, 'learning_rate/qwen_vl_interface': 9.286990291687605e-06, 'learning_rate/action_model': 9.286990291687604e-05, 'epoch': 0.6}) +[2026-06-17 18:50:08,374][starVLA.training.train_starvla][INFO] - Step 1320, Loss: {'train/loss': 0.9283015727996826, 'train/grad_norm_pre_clip': 12.502461433410645, 'timing/data': 0.01607460528612137, 'timing/model': 0.9546191673725843, 'learning_rate/qwen_vl_interface': 9.285850230947934e-06, 'learning_rate/action_model': 9.285850230947934e-05, 'epoch': 0.6}) +[2026-06-17 18:50:15,055][starVLA.training.train_starvla][INFO] - Step 1321, Loss: {'train/loss': 0.608127772808075, 'train/grad_norm_pre_clip': 8.488632202148438, 'timing/data': 0.022838703356683254, 'timing/model': 2.719403005205095, 'learning_rate/qwen_vl_interface': 9.28470933357217e-06, 'learning_rate/action_model': 9.28470933357217e-05, 'epoch': 0.6}) +[2026-06-17 18:50:19,499][starVLA.training.train_starvla][INFO] - Step 1322, Loss: {'train/loss': 0.804043173789978, 'train/grad_norm_pre_clip': 13.89418888092041, 'timing/data': 0.02800597343593836, 'timing/model': 0.9611145881935954, 'learning_rate/qwen_vl_interface': 9.283567599796819e-06, 'learning_rate/action_model': 9.283567599796818e-05, 'epoch': 0.6}) +[2026-06-17 18:50:23,676][starVLA.training.train_starvla][INFO] - Step 1323, Loss: {'train/loss': 0.7535951137542725, 'train/grad_norm_pre_clip': 11.639204025268555, 'timing/data': 0.019489635713398457, 'timing/model': 0.9535103077068925, 'learning_rate/qwen_vl_interface': 9.282425029858561e-06, 'learning_rate/action_model': 9.282425029858561e-05, 'epoch': 0.6}) +[2026-06-17 18:50:27,815][starVLA.training.train_starvla][INFO] - Step 1324, Loss: {'train/loss': 0.5687041878700256, 'train/grad_norm_pre_clip': 10.97565746307373, 'timing/data': 0.023072737269103527, 'timing/model': 0.9682030463591218, 'learning_rate/qwen_vl_interface': 9.281281623994258e-06, 'learning_rate/action_model': 9.281281623994258e-05, 'epoch': 0.6}) +[2026-06-17 18:50:31,756][starVLA.training.train_starvla][INFO] - Step 1325, Loss: {'train/loss': 0.5178908705711365, 'train/grad_norm_pre_clip': 9.889577865600586, 'timing/data': 0.0280534652993083, 'timing/model': 1.1532542407512665, 'learning_rate/qwen_vl_interface': 9.280137382440936e-06, 'learning_rate/action_model': 9.280137382440937e-05, 'epoch': 0.6}) +[2026-06-17 18:50:35,490][starVLA.training.train_starvla][INFO] - Step 1326, Loss: {'train/loss': 0.7545201182365417, 'train/grad_norm_pre_clip': 10.60137939453125, 'timing/data': 0.01596697047352791, 'timing/model': 0.96629349142313, 'learning_rate/qwen_vl_interface': 9.278992305435798e-06, 'learning_rate/action_model': 9.278992305435797e-05, 'epoch': 0.6}) +[2026-06-17 18:50:39,761][starVLA.training.train_starvla][INFO] - Step 1327, Loss: {'train/loss': 0.8522046208381653, 'train/grad_norm_pre_clip': 10.94310188293457, 'timing/data': 0.01657437440007925, 'timing/model': 0.9407215844839811, 'learning_rate/qwen_vl_interface': 9.27784639321622e-06, 'learning_rate/action_model': 9.27784639321622e-05, 'epoch': 0.6}) +[2026-06-17 18:50:43,884][starVLA.training.train_starvla][INFO] - Step 1328, Loss: {'train/loss': 0.529771625995636, 'train/grad_norm_pre_clip': 35.47600173950195, 'timing/data': 0.023076102137565613, 'timing/model': 1.0251364493742585, 'learning_rate/qwen_vl_interface': 9.27669964601975e-06, 'learning_rate/action_model': 9.27669964601975e-05, 'epoch': 0.6}) +[2026-06-17 18:50:47,947][starVLA.training.train_starvla][INFO] - Step 1329, Loss: {'train/loss': 0.4466734230518341, 'train/grad_norm_pre_clip': 7.134219169616699, 'timing/data': 0.02366587519645691, 'timing/model': 0.9549584211781621, 'learning_rate/qwen_vl_interface': 9.27555206408411e-06, 'learning_rate/action_model': 9.275552064084111e-05, 'epoch': 0.6}) +[2026-06-17 18:50:51,871][starVLA.training.train_starvla][INFO] - Step 1330, Loss: {'train/loss': 0.602537989616394, 'train/grad_norm_pre_clip': 9.387918472290039, 'timing/data': 0.0188309196382761, 'timing/model': 0.9573378283530474, 'learning_rate/qwen_vl_interface': 9.274403647647197e-06, 'learning_rate/action_model': 9.274403647647197e-05, 'epoch': 0.6}) +[2026-06-17 18:50:55,899][starVLA.training.train_starvla][INFO] - Step 1331, Loss: {'train/loss': 0.6204323172569275, 'train/grad_norm_pre_clip': 8.808233261108398, 'timing/data': 0.022854979149997234, 'timing/model': 1.1820374745875597, 'learning_rate/qwen_vl_interface': 9.273254396947079e-06, 'learning_rate/action_model': 9.273254396947078e-05, 'epoch': 0.6}) +[2026-06-17 18:50:59,788][starVLA.training.train_starvla][INFO] - Step 1332, Loss: {'train/loss': 0.6500241756439209, 'train/grad_norm_pre_clip': 12.118424415588379, 'timing/data': 0.016699878498911858, 'timing/model': 1.1767807379364967, 'learning_rate/qwen_vl_interface': 9.27210431222199e-06, 'learning_rate/action_model': 9.27210431222199e-05, 'epoch': 0.6}) +[2026-06-17 18:51:03,584][starVLA.training.train_starvla][INFO] - Step 1333, Loss: {'train/loss': 0.33832940459251404, 'train/grad_norm_pre_clip': 7.213555812835693, 'timing/data': 0.021809018217027187, 'timing/model': 0.9760092366486788, 'learning_rate/qwen_vl_interface': 9.270953393710353e-06, 'learning_rate/action_model': 9.270953393710352e-05, 'epoch': 0.6}) +[2026-06-17 18:51:07,657][starVLA.training.train_starvla][INFO] - Step 1334, Loss: {'train/loss': 0.4332800805568695, 'train/grad_norm_pre_clip': 6.337229251861572, 'timing/data': 0.025715170428156853, 'timing/model': 0.9518023980781436, 'learning_rate/qwen_vl_interface': 9.269801641650751e-06, 'learning_rate/action_model': 9.26980164165075e-05, 'epoch': 0.61}) +[2026-06-17 18:51:11,716][starVLA.training.train_starvla][INFO] - Step 1335, Loss: {'train/loss': 0.8911654949188232, 'train/grad_norm_pre_clip': 11.930140495300293, 'timing/data': 0.015742489136755466, 'timing/model': 1.0368357803672552, 'learning_rate/qwen_vl_interface': 9.268649056281941e-06, 'learning_rate/action_model': 9.268649056281941e-05, 'epoch': 0.61}) +[2026-06-17 18:51:15,698][starVLA.training.train_starvla][INFO] - Step 1336, Loss: {'train/loss': 0.5879519581794739, 'train/grad_norm_pre_clip': 7.81928825378418, 'timing/data': 0.017031916417181492, 'timing/model': 1.014624871313572, 'learning_rate/qwen_vl_interface': 9.267495637842859e-06, 'learning_rate/action_model': 9.267495637842857e-05, 'epoch': 0.61}) +[2026-06-17 18:51:19,597][starVLA.training.train_starvla][INFO] - Step 1337, Loss: {'train/loss': 0.4707023501396179, 'train/grad_norm_pre_clip': 9.258867263793945, 'timing/data': 0.019753824919462204, 'timing/model': 0.9574252283200622, 'learning_rate/qwen_vl_interface': 9.266341386572607e-06, 'learning_rate/action_model': 9.266341386572605e-05, 'epoch': 0.61}) +[2026-06-17 18:51:23,566][starVLA.training.train_starvla][INFO] - Step 1338, Loss: {'train/loss': 0.46693676710128784, 'train/grad_norm_pre_clip': 7.543463706970215, 'timing/data': 0.019508762285113335, 'timing/model': 1.274098975583911, 'learning_rate/qwen_vl_interface': 9.265186302710465e-06, 'learning_rate/action_model': 9.265186302710465e-05, 'epoch': 0.61}) +[2026-06-17 18:51:27,528][starVLA.training.train_starvla][INFO] - Step 1339, Loss: {'train/loss': 0.7659977674484253, 'train/grad_norm_pre_clip': 12.605265617370605, 'timing/data': 0.01890902128070593, 'timing/model': 1.1696155741810799, 'learning_rate/qwen_vl_interface': 9.26403038649588e-06, 'learning_rate/action_model': 9.26403038649588e-05, 'epoch': 0.61}) +[2026-06-17 18:51:31,252][starVLA.training.train_starvla][INFO] - Step 1340, Loss: {'train/loss': 0.6410497426986694, 'train/grad_norm_pre_clip': 10.60281753540039, 'timing/data': 0.028489618562161922, 'timing/model': 0.958725756034255, 'learning_rate/qwen_vl_interface': 9.262873638168478e-06, 'learning_rate/action_model': 9.262873638168478e-05, 'epoch': 0.61}) +[2026-06-17 18:51:35,141][starVLA.training.train_starvla][INFO] - Step 1341, Loss: {'train/loss': 0.3562973439693451, 'train/grad_norm_pre_clip': 6.366120338439941, 'timing/data': 0.01752978190779686, 'timing/model': 0.9857609877362847, 'learning_rate/qwen_vl_interface': 9.261716057968054e-06, 'learning_rate/action_model': 9.261716057968053e-05, 'epoch': 0.61}) +[2026-06-17 18:51:39,037][starVLA.training.train_starvla][INFO] - Step 1342, Loss: {'train/loss': 0.6985427141189575, 'train/grad_norm_pre_clip': 11.72138786315918, 'timing/data': 0.018531818874180317, 'timing/model': 0.939214208163321, 'learning_rate/qwen_vl_interface': 9.260557646134572e-06, 'learning_rate/action_model': 9.260557646134572e-05, 'epoch': 0.61}) +[2026-06-17 18:51:42,898][starVLA.training.train_starvla][INFO] - Step 1343, Loss: {'train/loss': 0.5474070310592651, 'train/grad_norm_pre_clip': 12.044633865356445, 'timing/data': 0.01872180588543415, 'timing/model': 0.9593686424195766, 'learning_rate/qwen_vl_interface': 9.259398402908175e-06, 'learning_rate/action_model': 9.259398402908174e-05, 'epoch': 0.61}) +[2026-06-17 18:51:46,898][starVLA.training.train_starvla][INFO] - Step 1344, Loss: {'train/loss': 0.48144203424453735, 'train/grad_norm_pre_clip': 8.856830596923828, 'timing/data': 0.020895509980618954, 'timing/model': 0.9608201887458563, 'learning_rate/qwen_vl_interface': 9.258238328529177e-06, 'learning_rate/action_model': 9.258238328529176e-05, 'epoch': 0.61}) +[2026-06-17 18:51:50,708][starVLA.training.train_starvla][INFO] - Step 1345, Loss: {'train/loss': 0.7201328873634338, 'train/grad_norm_pre_clip': 13.562320709228516, 'timing/data': 0.02519835438579321, 'timing/model': 1.0981444446370006, 'learning_rate/qwen_vl_interface': 9.257077423238058e-06, 'learning_rate/action_model': 9.257077423238058e-05, 'epoch': 0.61}) +[2026-06-17 18:51:54,424][starVLA.training.train_starvla][INFO] - Step 1346, Loss: {'train/loss': 0.6983603835105896, 'train/grad_norm_pre_clip': 20.529903411865234, 'timing/data': 0.022543172352015972, 'timing/model': 0.9616576321423054, 'learning_rate/qwen_vl_interface': 9.25591568727548e-06, 'learning_rate/action_model': 9.255915687275478e-05, 'epoch': 0.61}) +[2026-06-17 18:51:58,427][starVLA.training.train_starvla][INFO] - Step 1347, Loss: {'train/loss': 0.8190453052520752, 'train/grad_norm_pre_clip': 17.136985778808594, 'timing/data': 0.01622739154845476, 'timing/model': 0.9457937870174646, 'learning_rate/qwen_vl_interface': 9.254753120882265e-06, 'learning_rate/action_model': 9.254753120882266e-05, 'epoch': 0.61}) +[2026-06-17 18:52:02,954][starVLA.training.train_starvla][INFO] - Step 1348, Loss: {'train/loss': 0.6652540564537048, 'train/grad_norm_pre_clip': 8.867384910583496, 'timing/data': 0.019547116942703724, 'timing/model': 0.9474648553878069, 'learning_rate/qwen_vl_interface': 9.25358972429942e-06, 'learning_rate/action_model': 9.253589724299421e-05, 'epoch': 0.61}) +[2026-06-17 18:52:07,033][starVLA.training.train_starvla][INFO] - Step 1349, Loss: {'train/loss': 0.6257208585739136, 'train/grad_norm_pre_clip': 13.116484642028809, 'timing/data': 0.016671564429998398, 'timing/model': 0.9652271885424852, 'learning_rate/qwen_vl_interface': 9.25242549776812e-06, 'learning_rate/action_model': 9.252425497768119e-05, 'epoch': 0.61}) +[2026-06-17 18:52:10,957][starVLA.training.train_starvla][INFO] - Step 1350, Loss: {'train/loss': 0.6172620058059692, 'train/grad_norm_pre_clip': 10.538941383361816, 'timing/data': 0.01788461022078991, 'timing/model': 0.9562525190412998, 'learning_rate/qwen_vl_interface': 9.251260441529706e-06, 'learning_rate/action_model': 9.251260441529705e-05, 'epoch': 0.61}) +[2026-06-17 18:52:14,886][starVLA.training.train_starvla][INFO] - Step 1351, Loss: {'train/loss': 0.8508821725845337, 'train/grad_norm_pre_clip': 14.06534194946289, 'timing/data': 0.02057604491710663, 'timing/model': 1.1746142329648137, 'learning_rate/qwen_vl_interface': 9.250094555825697e-06, 'learning_rate/action_model': 9.250094555825696e-05, 'epoch': 0.61}) +[2026-06-17 18:52:18,772][starVLA.training.train_starvla][INFO] - Step 1352, Loss: {'train/loss': 0.685493528842926, 'train/grad_norm_pre_clip': 11.423233985900879, 'timing/data': 0.024796858429908752, 'timing/model': 1.1165540255606174, 'learning_rate/qwen_vl_interface': 9.248927840897782e-06, 'learning_rate/action_model': 9.248927840897781e-05, 'epoch': 0.61}) +[2026-06-17 18:52:22,492][starVLA.training.train_starvla][INFO] - Step 1353, Loss: {'train/loss': 0.5753088593482971, 'train/grad_norm_pre_clip': 14.489197731018066, 'timing/data': 0.016190236434340477, 'timing/model': 0.9649577457457781, 'learning_rate/qwen_vl_interface': 9.247760296987822e-06, 'learning_rate/action_model': 9.247760296987821e-05, 'epoch': 0.61}) +[2026-06-17 18:52:26,374][starVLA.training.train_starvla][INFO] - Step 1354, Loss: {'train/loss': 0.545487105846405, 'train/grad_norm_pre_clip': 13.573796272277832, 'timing/data': 0.0166015662252903, 'timing/model': 0.9523101942613721, 'learning_rate/qwen_vl_interface': 9.246591924337852e-06, 'learning_rate/action_model': 9.246591924337852e-05, 'epoch': 0.61}) +[2026-06-17 18:52:30,352][starVLA.training.train_starvla][INFO] - Step 1355, Loss: {'train/loss': 0.39903724193573, 'train/grad_norm_pre_clip': 15.05154037475586, 'timing/data': 0.016482054255902767, 'timing/model': 0.9512799782678485, 'learning_rate/qwen_vl_interface': 9.245422723190076e-06, 'learning_rate/action_model': 9.245422723190074e-05, 'epoch': 0.61}) +[2026-06-17 18:52:34,222][starVLA.training.train_starvla][INFO] - Step 1356, Loss: {'train/loss': 0.6715035438537598, 'train/grad_norm_pre_clip': 11.716306686401367, 'timing/data': 0.028706148266792297, 'timing/model': 0.9605755722150207, 'learning_rate/qwen_vl_interface': 9.244252693786868e-06, 'learning_rate/action_model': 9.244252693786869e-05, 'epoch': 0.62}) +[2026-06-17 18:52:38,105][starVLA.training.train_starvla][INFO] - Step 1357, Loss: {'train/loss': 0.4218667447566986, 'train/grad_norm_pre_clip': 8.429854393005371, 'timing/data': 0.016694163903594017, 'timing/model': 0.953134074807167, 'learning_rate/qwen_vl_interface': 9.24308183637078e-06, 'learning_rate/action_model': 9.24308183637078e-05, 'epoch': 0.62}) +[2026-06-17 18:52:42,020][starVLA.training.train_starvla][INFO] - Step 1358, Loss: {'train/loss': 0.4865943491458893, 'train/grad_norm_pre_clip': 9.318730354309082, 'timing/data': 0.02420964650809765, 'timing/model': 1.1841344004496932, 'learning_rate/qwen_vl_interface': 9.24191015118453e-06, 'learning_rate/action_model': 9.24191015118453e-05, 'epoch': 0.62}) +[2026-06-17 18:52:46,018][starVLA.training.train_starvla][INFO] - Step 1359, Loss: {'train/loss': 0.466521680355072, 'train/grad_norm_pre_clip': 9.451844215393066, 'timing/data': 0.01926696114242077, 'timing/model': 1.245369864627719, 'learning_rate/qwen_vl_interface': 9.24073763847101e-06, 'learning_rate/action_model': 9.24073763847101e-05, 'epoch': 0.62}) +[2026-06-17 18:52:49,767][starVLA.training.train_starvla][INFO] - Step 1360, Loss: {'train/loss': 0.6232548952102661, 'train/grad_norm_pre_clip': 11.723714828491211, 'timing/data': 0.01616563554853201, 'timing/model': 1.0036998121067882, 'learning_rate/qwen_vl_interface': 9.239564298473284e-06, 'learning_rate/action_model': 9.239564298473284e-05, 'epoch': 0.62}) +[2026-06-17 18:52:53,702][starVLA.training.train_starvla][INFO] - Step 1361, Loss: {'train/loss': 0.37100011110305786, 'train/grad_norm_pre_clip': 7.384647369384766, 'timing/data': 0.014480763114988804, 'timing/model': 1.0204390734434128, 'learning_rate/qwen_vl_interface': 9.238390131434586e-06, 'learning_rate/action_model': 9.238390131434586e-05, 'epoch': 0.62}) +[2026-06-17 18:52:57,603][starVLA.training.train_starvla][INFO] - Step 1362, Loss: {'train/loss': 0.5088173151016235, 'train/grad_norm_pre_clip': 9.815051078796387, 'timing/data': 0.020253166556358337, 'timing/model': 0.9695124151185155, 'learning_rate/qwen_vl_interface': 9.237215137598322e-06, 'learning_rate/action_model': 9.237215137598322e-05, 'epoch': 0.62}) +[2026-06-17 18:53:01,791][starVLA.training.train_starvla][INFO] - Step 1363, Loss: {'train/loss': 0.44530582427978516, 'train/grad_norm_pre_clip': 9.409825325012207, 'timing/data': 0.025216491892933846, 'timing/model': 1.0191137110814452, 'learning_rate/qwen_vl_interface': 9.236039317208068e-06, 'learning_rate/action_model': 9.236039317208068e-05, 'epoch': 0.62}) +[2026-06-17 18:53:05,856][starVLA.training.train_starvla][INFO] - Step 1364, Loss: {'train/loss': 0.5088773369789124, 'train/grad_norm_pre_clip': 9.676620483398438, 'timing/data': 0.01744952704757452, 'timing/model': 0.9545990163460374, 'learning_rate/qwen_vl_interface': 9.234862670507577e-06, 'learning_rate/action_model': 9.234862670507576e-05, 'epoch': 0.62}) +[2026-06-17 18:53:09,854][starVLA.training.train_starvla][INFO] - Step 1365, Loss: {'train/loss': 0.6411895751953125, 'train/grad_norm_pre_clip': 12.180266380310059, 'timing/data': 0.01746336091309786, 'timing/model': 1.153181068599224, 'learning_rate/qwen_vl_interface': 9.233685197740765e-06, 'learning_rate/action_model': 9.233685197740764e-05, 'epoch': 0.62}) +[2026-06-17 18:53:13,674][starVLA.training.train_starvla][INFO] - Step 1366, Loss: {'train/loss': 0.4459880590438843, 'train/grad_norm_pre_clip': 7.985641956329346, 'timing/data': 0.018788525834679604, 'timing/model': 1.062664563767612, 'learning_rate/qwen_vl_interface': 9.232506899151725e-06, 'learning_rate/action_model': 9.232506899151725e-05, 'epoch': 0.62}) +[2026-06-17 18:53:18,517][starVLA.training.train_starvla][INFO] - Step 1367, Loss: {'train/loss': 0.3962714374065399, 'train/grad_norm_pre_clip': 8.514754295349121, 'timing/data': 0.017493645660579205, 'timing/model': 1.036078816279769, 'learning_rate/qwen_vl_interface': 9.23132777498472e-06, 'learning_rate/action_model': 9.231327774984719e-05, 'epoch': 0.62}) +[2026-06-17 18:53:24,396][starVLA.training.train_starvla][INFO] - Step 1368, Loss: {'train/loss': 0.6828067302703857, 'train/grad_norm_pre_clip': 17.723400115966797, 'timing/data': 0.014302176423370838, 'timing/model': 2.874324096366763, 'learning_rate/qwen_vl_interface': 9.230147825484182e-06, 'learning_rate/action_model': 9.230147825484183e-05, 'epoch': 0.62}) +[2026-06-17 18:53:28,856][starVLA.training.train_starvla][INFO] - Step 1369, Loss: {'train/loss': 0.7035924196243286, 'train/grad_norm_pre_clip': 15.551107406616211, 'timing/data': 0.018493236042559147, 'timing/model': 0.9379184748977423, 'learning_rate/qwen_vl_interface': 9.228967050894717e-06, 'learning_rate/action_model': 9.228967050894716e-05, 'epoch': 0.62}) +[2026-06-17 18:53:32,837][starVLA.training.train_starvla][INFO] - Step 1370, Loss: {'train/loss': 0.47077417373657227, 'train/grad_norm_pre_clip': 15.074206352233887, 'timing/data': 0.01752591785043478, 'timing/model': 0.9372854139655828, 'learning_rate/qwen_vl_interface': 9.227785451461103e-06, 'learning_rate/action_model': 9.227785451461102e-05, 'epoch': 0.62}) +[2026-06-17 18:53:36,765][starVLA.training.train_starvla][INFO] - Step 1371, Loss: {'train/loss': 0.719040036201477, 'train/grad_norm_pre_clip': 15.604427337646484, 'timing/data': 0.017270208336412907, 'timing/model': 0.9642743188887835, 'learning_rate/qwen_vl_interface': 9.226603027428281e-06, 'learning_rate/action_model': 9.226603027428281e-05, 'epoch': 0.62}) +[2026-06-17 18:53:40,735][starVLA.training.train_starvla][INFO] - Step 1372, Loss: {'train/loss': 0.7549265623092651, 'train/grad_norm_pre_clip': 12.670431137084961, 'timing/data': 0.03097931481897831, 'timing/model': 1.1945182895287871, 'learning_rate/qwen_vl_interface': 9.225419779041376e-06, 'learning_rate/action_model': 9.225419779041376e-05, 'epoch': 0.62}) +[2026-06-17 18:53:44,466][starVLA.training.train_starvla][INFO] - Step 1373, Loss: {'train/loss': 0.4873102009296417, 'train/grad_norm_pre_clip': 8.073246955871582, 'timing/data': 0.016355267725884914, 'timing/model': 0.9951679715886712, 'learning_rate/qwen_vl_interface': 9.224235706545672e-06, 'learning_rate/action_model': 9.224235706545671e-05, 'epoch': 0.62}) +[2026-06-17 18:53:48,386][starVLA.training.train_starvla][INFO] - Step 1374, Loss: {'train/loss': 0.634763777256012, 'train/grad_norm_pre_clip': 8.642723083496094, 'timing/data': 0.014952532015740871, 'timing/model': 0.9759578835219145, 'learning_rate/qwen_vl_interface': 9.22305081018663e-06, 'learning_rate/action_model': 9.223050810186629e-05, 'epoch': 0.62}) +[2026-06-17 18:53:52,743][starVLA.training.train_starvla][INFO] - Step 1375, Loss: {'train/loss': 0.6570032835006714, 'train/grad_norm_pre_clip': 39.76142120361328, 'timing/data': 0.017036299221217632, 'timing/model': 0.9718616595491767, 'learning_rate/qwen_vl_interface': 9.221865090209878e-06, 'learning_rate/action_model': 9.221865090209878e-05, 'epoch': 0.62}) +[2026-06-17 18:53:56,827][starVLA.training.train_starvla][INFO] - Step 1376, Loss: {'train/loss': 0.49591130018234253, 'train/grad_norm_pre_clip': 8.183086395263672, 'timing/data': 0.023967611603438854, 'timing/model': 1.0074292430654168, 'learning_rate/qwen_vl_interface': 9.22067854686122e-06, 'learning_rate/action_model': 9.22067854686122e-05, 'epoch': 0.62}) +[2026-06-17 18:54:00,671][starVLA.training.train_starvla][INFO] - Step 1377, Loss: {'train/loss': 0.5969192981719971, 'train/grad_norm_pre_clip': 12.888269424438477, 'timing/data': 0.028646042570471764, 'timing/model': 0.9516958314925432, 'learning_rate/qwen_vl_interface': 9.219491180386626e-06, 'learning_rate/action_model': 9.219491180386626e-05, 'epoch': 0.62}) +[2026-06-17 18:54:04,735][starVLA.training.train_starvla][INFO] - Step 1378, Loss: {'train/loss': 0.5296701788902283, 'train/grad_norm_pre_clip': 9.316316604614258, 'timing/data': 0.018537077121436596, 'timing/model': 1.1894729733467102, 'learning_rate/qwen_vl_interface': 9.21830299103224e-06, 'learning_rate/action_model': 9.21830299103224e-05, 'epoch': 0.63}) +[2026-06-17 18:54:08,705][starVLA.training.train_starvla][INFO] - Step 1379, Loss: {'train/loss': 0.6599416136741638, 'train/grad_norm_pre_clip': 10.209883689880371, 'timing/data': 0.016884994693100452, 'timing/model': 1.1691158628091216, 'learning_rate/qwen_vl_interface': 9.217113979044373e-06, 'learning_rate/action_model': 9.217113979044373e-05, 'epoch': 0.63}) +[2026-06-17 18:54:12,498][starVLA.training.train_starvla][INFO] - Step 1380, Loss: {'train/loss': 0.6600009202957153, 'train/grad_norm_pre_clip': 9.88028621673584, 'timing/data': 0.019006245769560337, 'timing/model': 0.9665628271177411, 'learning_rate/qwen_vl_interface': 9.215924144669509e-06, 'learning_rate/action_model': 9.215924144669509e-05, 'epoch': 0.63}) +[2026-06-17 18:54:16,506][starVLA.training.train_starvla][INFO] - Step 1381, Loss: {'train/loss': 1.0920615196228027, 'train/grad_norm_pre_clip': 15.948775291442871, 'timing/data': 0.021977702155709267, 'timing/model': 0.9561716290190816, 'learning_rate/qwen_vl_interface': 9.214733488154304e-06, 'learning_rate/action_model': 9.214733488154304e-05, 'epoch': 0.63}) +[2026-06-17 18:54:20,425][starVLA.training.train_starvla][INFO] - Step 1382, Loss: {'train/loss': 0.5654029846191406, 'train/grad_norm_pre_clip': 7.982056617736816, 'timing/data': 0.02121326792985201, 'timing/model': 0.9537099227309227, 'learning_rate/qwen_vl_interface': 9.213542009745579e-06, 'learning_rate/action_model': 9.213542009745578e-05, 'epoch': 0.63}) +[2026-06-17 18:54:24,323][starVLA.training.train_starvla][INFO] - Step 1383, Loss: {'train/loss': 0.6382080912590027, 'train/grad_norm_pre_clip': 11.7328519821167, 'timing/data': 0.019301888532936573, 'timing/model': 0.965952655300498, 'learning_rate/qwen_vl_interface': 9.21234970969033e-06, 'learning_rate/action_model': 9.21234970969033e-05, 'epoch': 0.63}) +[2026-06-17 18:54:28,146][starVLA.training.train_starvla][INFO] - Step 1384, Loss: {'train/loss': 0.5352569818496704, 'train/grad_norm_pre_clip': 8.561467170715332, 'timing/data': 0.01432257704436779, 'timing/model': 0.9357070224359632, 'learning_rate/qwen_vl_interface': 9.211156588235724e-06, 'learning_rate/action_model': 9.211156588235725e-05, 'epoch': 0.63}) +[2026-06-17 18:54:32,166][starVLA.training.train_starvla][INFO] - Step 1385, Loss: {'train/loss': 0.501464307308197, 'train/grad_norm_pre_clip': 9.916474342346191, 'timing/data': 0.02129257470369339, 'timing/model': 1.1822170093655586, 'learning_rate/qwen_vl_interface': 9.209962645629095e-06, 'learning_rate/action_model': 9.209962645629095e-05, 'epoch': 0.63}) +[2026-06-17 18:54:37,702][starVLA.training.train_starvla][INFO] - Step 1386, Loss: {'train/loss': 0.5897305607795715, 'train/grad_norm_pre_clip': 9.170343399047852, 'timing/data': 0.01795777026563883, 'timing/model': 1.2393028242513537, 'learning_rate/qwen_vl_interface': 9.208767882117947e-06, 'learning_rate/action_model': 9.208767882117947e-05, 'epoch': 0.63}) +[2026-06-17 18:54:42,858][starVLA.training.train_starvla][INFO] - Step 1387, Loss: {'train/loss': 0.44641315937042236, 'train/grad_norm_pre_clip': 35.02400588989258, 'timing/data': 0.018449293449521065, 'timing/model': 0.9860043125227094, 'learning_rate/qwen_vl_interface': 9.207572297949958e-06, 'learning_rate/action_model': 9.207572297949958e-05, 'epoch': 0.63}) +[2026-06-17 18:54:47,544][starVLA.training.train_starvla][INFO] - Step 1388, Loss: {'train/loss': 0.7582046985626221, 'train/grad_norm_pre_clip': 11.7866849899292, 'timing/data': 0.024150077253580093, 'timing/model': 0.949377503246069, 'learning_rate/qwen_vl_interface': 9.206375893372973e-06, 'learning_rate/action_model': 9.206375893372973e-05, 'epoch': 0.63}) +[2026-06-17 18:54:51,492][starVLA.training.train_starvla][INFO] - Step 1389, Loss: {'train/loss': 0.7592102289199829, 'train/grad_norm_pre_clip': 8.969517707824707, 'timing/data': 0.0217563658952713, 'timing/model': 1.0284032439813018, 'learning_rate/qwen_vl_interface': 9.20517866863501e-06, 'learning_rate/action_model': 9.205178668635009e-05, 'epoch': 0.63}) +[2026-06-17 18:54:55,370][starVLA.training.train_starvla][INFO] - Step 1390, Loss: {'train/loss': 0.4599817097187042, 'train/grad_norm_pre_clip': 8.944368362426758, 'timing/data': 0.015183246694505215, 'timing/model': 0.9529417874291539, 'learning_rate/qwen_vl_interface': 9.20398062398425e-06, 'learning_rate/action_model': 9.20398062398425e-05, 'epoch': 0.63}) +[2026-06-17 18:54:59,249][starVLA.training.train_starvla][INFO] - Step 1391, Loss: {'train/loss': 0.9147488474845886, 'train/grad_norm_pre_clip': 14.23866081237793, 'timing/data': 0.013857631012797356, 'timing/model': 0.9428741997107863, 'learning_rate/qwen_vl_interface': 9.202781759669053e-06, 'learning_rate/action_model': 9.202781759669052e-05, 'epoch': 0.63}) +[2026-06-17 18:55:03,170][starVLA.training.train_starvla][INFO] - Step 1392, Loss: {'train/loss': 0.4057828187942505, 'train/grad_norm_pre_clip': 7.518554210662842, 'timing/data': 0.015239779837429523, 'timing/model': 1.1653867037966847, 'learning_rate/qwen_vl_interface': 9.201582075937943e-06, 'learning_rate/action_model': 9.201582075937944e-05, 'epoch': 0.63}) +[2026-06-17 18:55:06,947][starVLA.training.train_starvla][INFO] - Step 1393, Loss: {'train/loss': 0.4499381184577942, 'train/grad_norm_pre_clip': 8.068962097167969, 'timing/data': 0.015657073818147182, 'timing/model': 0.939803003333509, 'learning_rate/qwen_vl_interface': 9.200381573039619e-06, 'learning_rate/action_model': 9.200381573039619e-05, 'epoch': 0.63}) +[2026-06-17 18:55:10,891][starVLA.training.train_starvla][INFO] - Step 1394, Loss: {'train/loss': 0.55025714635849, 'train/grad_norm_pre_clip': 8.299420356750488, 'timing/data': 0.029527834616601467, 'timing/model': 0.9573923135176301, 'learning_rate/qwen_vl_interface': 9.199180251222942e-06, 'learning_rate/action_model': 9.199180251222942e-05, 'epoch': 0.63}) +[2026-06-17 18:55:14,777][starVLA.training.train_starvla][INFO] - Step 1395, Loss: {'train/loss': 0.5395464301109314, 'train/grad_norm_pre_clip': 8.693365097045898, 'timing/data': 0.01954442262649536, 'timing/model': 1.0087292790412903, 'learning_rate/qwen_vl_interface': 9.197978110736951e-06, 'learning_rate/action_model': 9.19797811073695e-05, 'epoch': 0.63}) +[2026-06-17 18:55:18,793][starVLA.training.train_starvla][INFO] - Step 1396, Loss: {'train/loss': 0.7543473839759827, 'train/grad_norm_pre_clip': 11.312298774719238, 'timing/data': 0.018848346546292305, 'timing/model': 0.9727959418669343, 'learning_rate/qwen_vl_interface': 9.196775151830849e-06, 'learning_rate/action_model': 9.196775151830849e-05, 'epoch': 0.63}) +[2026-06-17 18:55:24,026][starVLA.training.train_starvla][INFO] - Step 1397, Loss: {'train/loss': 0.5772563815116882, 'train/grad_norm_pre_clip': 12.281712532043457, 'timing/data': 0.01758239045739174, 'timing/model': 1.0319156078621745, 'learning_rate/qwen_vl_interface': 9.19557137475401e-06, 'learning_rate/action_model': 9.19557137475401e-05, 'epoch': 0.63}) +[2026-06-17 18:55:28,006][starVLA.training.train_starvla][INFO] - Step 1398, Loss: {'train/loss': 0.4590992331504822, 'train/grad_norm_pre_clip': 9.000333786010742, 'timing/data': 0.01803346909582615, 'timing/model': 1.271267756819725, 'learning_rate/qwen_vl_interface': 9.19436677975598e-06, 'learning_rate/action_model': 9.19436677975598e-05, 'epoch': 0.63}) +[2026-06-17 18:55:31,960][starVLA.training.train_starvla][INFO] - Step 1399, Loss: {'train/loss': 0.4734288156032562, 'train/grad_norm_pre_clip': 7.822129726409912, 'timing/data': 0.02590304333716631, 'timing/model': 1.2094028191640973, 'learning_rate/qwen_vl_interface': 9.193161367086472e-06, 'learning_rate/action_model': 9.193161367086472e-05, 'epoch': 0.63}) +[2026-06-17 18:57:45,446][starVLA.training.train_starvla][INFO] - Step 1400, Loss: {'train/loss': 0.7036783695220947, 'train/grad_norm_pre_clip': 10.523850440979004, 'eval/loss': 0.5318992733955383, 'eval/latency_8/loss': 0.5318992733955383, 'eval/demon_fire_precision': 0.912568306010929, 'eval/demon_fire_recall': 0.9857098477788133, 'eval/demon_fire_f1': 0.9477299880525687, 'eval/demon_fire_rate_model': 0.9674457429048414, 'eval/demon_fire_rate_teacher': 0.8956594323873122, 'eval/demon_left_precision': 0.8551401869158879, 'eval/demon_left_recall': 0.39186295503211993, 'eval/demon_left_f1': 0.5374449339207048, 'eval/demon_left_rate_model': 0.05954368391764051, 'eval/demon_left_rate_teacher': 0.12993878686700056, 'eval/demon_right_precision': 0.7607573149741824, 'eval/demon_right_recall': 0.6696969696969697, 'eval/demon_right_f1': 0.7123287671232877, 'eval/demon_right_rate_model': 0.16165831942125766, 'eval/demon_right_rate_teacher': 0.18363939899833054, 'eval/demon_move_f1': 0.6248868505219962, 'eval/demon_cc_f1': 0.820550583636078, 'eval/latency_8/demon_fire_precision': 0.912568306010929, 'eval/latency_8/demon_fire_recall': 0.9857098477788133, 'eval/latency_8/demon_fire_f1': 0.9477299880525687, 'eval/latency_8/demon_fire_rate_model': 0.9674457429048414, 'eval/latency_8/demon_fire_rate_teacher': 0.8956594323873122, 'eval/latency_8/demon_left_precision': 0.8551401869158879, 'eval/latency_8/demon_left_recall': 0.39186295503211993, 'eval/latency_8/demon_left_f1': 0.5374449339207048, 'eval/latency_8/demon_left_rate_model': 0.05954368391764051, 'eval/latency_8/demon_left_rate_teacher': 0.12993878686700056, 'eval/latency_8/demon_right_precision': 0.7607573149741824, 'eval/latency_8/demon_right_recall': 0.6696969696969697, 'eval/latency_8/demon_right_f1': 0.7123287671232877, 'eval/latency_8/demon_right_rate_model': 0.16165831942125766, 'eval/latency_8/demon_right_rate_teacher': 0.18363939899833054, 'eval/latency_8/demon_move_f1': 0.6248868505219962, 'eval/latency_8/demon_cc_f1': 0.820550583636078, 'timing/data': 0.023364592343568802, 'timing/model': 0.9551044683903456, 'learning_rate/qwen_vl_interface': 9.191955136995371e-06, 'learning_rate/action_model': 9.191955136995371e-05, 'epoch': 0.64}) +[2026-06-17 18:57:45,448][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_1400_state +[2026-06-17 18:57:45,469][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 18:57:52,200][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_1400_state/model.safetensors +[2026-06-17 18:58:09,003][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_1400_state/optimizer.bin +[2026-06-17 18:58:09,004][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_1400_state/sampler.bin +[2026-06-17 18:58:09,004][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_1400_state/sampler_1.bin +[2026-06-17 18:58:09,006][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_1400_state/random_states_0.pkl +[2026-06-17 18:58:09,010][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1400 +[2026-06-17 18:58:09,010][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 18:58:09,021][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 18:58:13,264][starVLA.training.train_starvla][INFO] - Step 1401, Loss: {'train/loss': 0.5208767652511597, 'train/grad_norm_pre_clip': 7.009433746337891, 'timing/data': 0.015685846097767353, 'timing/model': 1.2732165409252048, 'learning_rate/qwen_vl_interface': 9.190748089732727e-06, 'learning_rate/action_model': 9.190748089732728e-05, 'epoch': 0.64}) +[2026-06-17 18:58:17,096][starVLA.training.train_starvla][INFO] - Step 1402, Loss: {'train/loss': 0.8244460225105286, 'train/grad_norm_pre_clip': 13.363081932067871, 'timing/data': 0.016338948160409927, 'timing/model': 1.0078973583877087, 'learning_rate/qwen_vl_interface': 9.189540225548763e-06, 'learning_rate/action_model': 9.189540225548764e-05, 'epoch': 0.64}) +[2026-06-17 18:58:21,233][starVLA.training.train_starvla][INFO] - Step 1403, Loss: {'train/loss': 0.41196128726005554, 'train/grad_norm_pre_clip': 7.8809003829956055, 'timing/data': 0.02202665340155363, 'timing/model': 0.9380419291555882, 'learning_rate/qwen_vl_interface': 9.188331544693872e-06, 'learning_rate/action_model': 9.188331544693873e-05, 'epoch': 0.64}) +[2026-06-17 18:58:25,201][starVLA.training.train_starvla][INFO] - Step 1404, Loss: {'train/loss': 0.35267558693885803, 'train/grad_norm_pre_clip': 7.227898597717285, 'timing/data': 0.023118695244193077, 'timing/model': 0.9536150461062789, 'learning_rate/qwen_vl_interface': 9.187122047418614e-06, 'learning_rate/action_model': 9.187122047418614e-05, 'epoch': 0.64}) +[2026-06-17 18:58:29,323][starVLA.training.train_starvla][INFO] - Step 1405, Loss: {'train/loss': 1.0819815397262573, 'train/grad_norm_pre_clip': 11.640055656433105, 'timing/data': 0.02004800457507372, 'timing/model': 0.9447848023846745, 'learning_rate/qwen_vl_interface': 9.18591173397372e-06, 'learning_rate/action_model': 9.185911733973719e-05, 'epoch': 0.64}) +[2026-06-17 18:58:33,900][starVLA.training.train_starvla][INFO] - Step 1406, Loss: {'train/loss': 0.4743027687072754, 'train/grad_norm_pre_clip': 9.04546070098877, 'timing/data': 0.023306122049689293, 'timing/model': 1.635461138561368, 'learning_rate/qwen_vl_interface': 9.184700604610087e-06, 'learning_rate/action_model': 9.184700604610085e-05, 'epoch': 0.64}) +[2026-06-17 18:58:37,907][starVLA.training.train_starvla][INFO] - Step 1407, Loss: {'train/loss': 0.6808618903160095, 'train/grad_norm_pre_clip': 11.209479331970215, 'timing/data': 0.019157144241034985, 'timing/model': 1.2792665008455515, 'learning_rate/qwen_vl_interface': 9.183488659578786e-06, 'learning_rate/action_model': 9.183488659578785e-05, 'epoch': 0.64}) +[2026-06-17 18:58:43,155][starVLA.training.train_starvla][INFO] - Step 1408, Loss: {'train/loss': 0.6517224907875061, 'train/grad_norm_pre_clip': 10.731805801391602, 'timing/data': 0.01824841368943453, 'timing/model': 1.1328497529029846, 'learning_rate/qwen_vl_interface': 9.18227589913105e-06, 'learning_rate/action_model': 9.18227589913105e-05, 'epoch': 0.64}) +[2026-06-17 18:58:46,935][starVLA.training.train_starvla][INFO] - Step 1409, Loss: {'train/loss': 0.27915462851524353, 'train/grad_norm_pre_clip': 6.175706386566162, 'timing/data': 0.017987646162509918, 'timing/model': 0.9447701442986727, 'learning_rate/qwen_vl_interface': 9.181062323518293e-06, 'learning_rate/action_model': 9.181062323518292e-05, 'epoch': 0.64}) +[2026-06-17 18:58:51,296][starVLA.training.train_starvla][INFO] - Step 1410, Loss: {'train/loss': 0.5654329657554626, 'train/grad_norm_pre_clip': 11.567643165588379, 'timing/data': 0.023952162824571133, 'timing/model': 1.205211359076202, 'learning_rate/qwen_vl_interface': 9.179847932992084e-06, 'learning_rate/action_model': 9.179847932992083e-05, 'epoch': 0.64}) +[2026-06-17 18:58:55,294][starVLA.training.train_starvla][INFO] - Step 1411, Loss: {'train/loss': 0.4967595040798187, 'train/grad_norm_pre_clip': 10.060009956359863, 'timing/data': 0.01934912707656622, 'timing/model': 0.9522199165076017, 'learning_rate/qwen_vl_interface': 9.17863272780417e-06, 'learning_rate/action_model': 9.178632727804168e-05, 'epoch': 0.64}) +[2026-06-17 18:58:59,786][starVLA.training.train_starvla][INFO] - Step 1412, Loss: {'train/loss': 0.4781234860420227, 'train/grad_norm_pre_clip': 7.241232872009277, 'timing/data': 0.024638216942548752, 'timing/model': 0.976730233989656, 'learning_rate/qwen_vl_interface': 9.177416708206463e-06, 'learning_rate/action_model': 9.177416708206463e-05, 'epoch': 0.64}) +[2026-06-17 18:59:03,727][starVLA.training.train_starvla][INFO] - Step 1413, Loss: {'train/loss': 0.6258000731468201, 'train/grad_norm_pre_clip': 9.63209342956543, 'timing/data': 0.019073442555963993, 'timing/model': 0.96255396399647, 'learning_rate/qwen_vl_interface': 9.176199874451049e-06, 'learning_rate/action_model': 9.176199874451048e-05, 'epoch': 0.64}) +[2026-06-17 18:59:07,783][starVLA.training.train_starvla][INFO] - Step 1414, Loss: {'train/loss': 0.4444068372249603, 'train/grad_norm_pre_clip': 7.889469146728516, 'timing/data': 0.017382663674652576, 'timing/model': 1.1852858560159802, 'learning_rate/qwen_vl_interface': 9.174982226790175e-06, 'learning_rate/action_model': 9.174982226790175e-05, 'epoch': 0.64}) +[2026-06-17 18:59:12,191][starVLA.training.train_starvla][INFO] - Step 1415, Loss: {'train/loss': 0.431230753660202, 'train/grad_norm_pre_clip': 9.26608657836914, 'timing/data': 0.01767665520310402, 'timing/model': 1.118851613253355, 'learning_rate/qwen_vl_interface': 9.173763765476262e-06, 'learning_rate/action_model': 9.17376376547626e-05, 'epoch': 0.64}) +[2026-06-17 18:59:16,820][starVLA.training.train_starvla][INFO] - Step 1416, Loss: {'train/loss': 0.7679185271263123, 'train/grad_norm_pre_clip': 12.439024925231934, 'timing/data': 0.07688588835299015, 'timing/model': 1.4898874601349235, 'learning_rate/qwen_vl_interface': 9.172544490761897e-06, 'learning_rate/action_model': 9.172544490761897e-05, 'epoch': 0.64}) +[2026-06-17 18:59:21,128][starVLA.training.train_starvla][INFO] - Step 1417, Loss: {'train/loss': 0.5212735533714294, 'train/grad_norm_pre_clip': 9.363842010498047, 'timing/data': 0.026750760152935982, 'timing/model': 1.0647740419954062, 'learning_rate/qwen_vl_interface': 9.171324402899841e-06, 'learning_rate/action_model': 9.17132440289984e-05, 'epoch': 0.64}) +[2026-06-17 18:59:26,022][starVLA.training.train_starvla][INFO] - Step 1418, Loss: {'train/loss': 0.7753446102142334, 'train/grad_norm_pre_clip': 14.572571754455566, 'timing/data': 0.0408535934984684, 'timing/model': 0.9531886400654912, 'learning_rate/qwen_vl_interface': 9.170103502143014e-06, 'learning_rate/action_model': 9.170103502143013e-05, 'epoch': 0.64}) +[2026-06-17 18:59:30,021][starVLA.training.train_starvla][INFO] - Step 1419, Loss: {'train/loss': 0.28156185150146484, 'train/grad_norm_pre_clip': 10.560225486755371, 'timing/data': 0.042314319871366024, 'timing/model': 1.1362724350765347, 'learning_rate/qwen_vl_interface': 9.168881788744513e-06, 'learning_rate/action_model': 9.168881788744513e-05, 'epoch': 0.64}) +[2026-06-17 18:59:33,853][starVLA.training.train_starvla][INFO] - Step 1420, Loss: {'train/loss': 0.5815755724906921, 'train/grad_norm_pre_clip': 9.946484565734863, 'timing/data': 0.04655069764703512, 'timing/model': 1.1389876687899232, 'learning_rate/qwen_vl_interface': 9.1676592629576e-06, 'learning_rate/action_model': 9.167659262957601e-05, 'epoch': 0.64}) +[2026-06-17 18:59:37,710][starVLA.training.train_starvla][INFO] - Step 1421, Loss: {'train/loss': 0.47122812271118164, 'train/grad_norm_pre_clip': 9.148128509521484, 'timing/data': 0.03604455292224884, 'timing/model': 1.095511277206242, 'learning_rate/qwen_vl_interface': 9.16643592503571e-06, 'learning_rate/action_model': 9.166435925035709e-05, 'epoch': 0.64}) +[2026-06-17 18:59:41,725][starVLA.training.train_starvla][INFO] - Step 1422, Loss: {'train/loss': 0.6217724680900574, 'train/grad_norm_pre_clip': 10.382245063781738, 'timing/data': 0.03845772985368967, 'timing/model': 1.1228793822228909, 'learning_rate/qwen_vl_interface': 9.165211775232436e-06, 'learning_rate/action_model': 9.165211775232435e-05, 'epoch': 0.65}) +[2026-06-17 18:59:45,593][starVLA.training.train_starvla][INFO] - Step 1423, Loss: {'train/loss': 0.6587346792221069, 'train/grad_norm_pre_clip': 9.803017616271973, 'timing/data': 0.030481671914458275, 'timing/model': 0.9297274686396122, 'learning_rate/qwen_vl_interface': 9.163986813801548e-06, 'learning_rate/action_model': 9.163986813801548e-05, 'epoch': 0.65}) +[2026-06-17 18:59:49,544][starVLA.training.train_starvla][INFO] - Step 1424, Loss: {'train/loss': 0.42852476239204407, 'train/grad_norm_pre_clip': 9.164671897888184, 'timing/data': 0.034095993265509605, 'timing/model': 0.9740476030856371, 'learning_rate/qwen_vl_interface': 9.162761040996979e-06, 'learning_rate/action_model': 9.162761040996978e-05, 'epoch': 0.65}) +[2026-06-17 18:59:53,480][starVLA.training.train_starvla][INFO] - Step 1425, Loss: {'train/loss': 0.4996167719364166, 'train/grad_norm_pre_clip': 8.868529319763184, 'timing/data': 0.041194869205355644, 'timing/model': 0.9633772950619459, 'learning_rate/qwen_vl_interface': 9.161534457072839e-06, 'learning_rate/action_model': 9.161534457072839e-05, 'epoch': 0.65}) +[2026-06-17 18:59:57,747][starVLA.training.train_starvla][INFO] - Step 1426, Loss: {'train/loss': 0.5654808282852173, 'train/grad_norm_pre_clip': 7.891298294067383, 'timing/data': 0.03397826850414276, 'timing/model': 1.080892475321889, 'learning_rate/qwen_vl_interface': 9.160307062283397e-06, 'learning_rate/action_model': 9.160307062283397e-05, 'epoch': 0.65}) +[2026-06-17 19:00:01,984][starVLA.training.train_starvla][INFO] - Step 1427, Loss: {'train/loss': 0.3537735044956207, 'train/grad_norm_pre_clip': 7.265630722045898, 'timing/data': 0.029866215772926807, 'timing/model': 1.2486063241958618, 'learning_rate/qwen_vl_interface': 9.159078856883093e-06, 'learning_rate/action_model': 9.159078856883093e-05, 'epoch': 0.65}) +[2026-06-17 19:00:07,726][starVLA.training.train_starvla][INFO] - Step 1428, Loss: {'train/loss': 0.6535328030586243, 'train/grad_norm_pre_clip': 8.82296371459961, 'timing/data': 0.034305719658732414, 'timing/model': 0.9655686775222421, 'learning_rate/qwen_vl_interface': 9.157849841126535e-06, 'learning_rate/action_model': 9.157849841126535e-05, 'epoch': 0.65}) +[2026-06-17 19:00:14,407][starVLA.training.train_starvla][INFO] - Step 1429, Loss: {'train/loss': 0.6848065257072449, 'train/grad_norm_pre_clip': 12.141302108764648, 'timing/data': 0.07455655001103878, 'timing/model': 3.7003315975889564, 'learning_rate/qwen_vl_interface': 9.156620015268502e-06, 'learning_rate/action_model': 9.1566200152685e-05, 'epoch': 0.65}) +[2026-06-17 19:00:18,478][starVLA.training.train_starvla][INFO] - Step 1430, Loss: {'train/loss': 0.594576358795166, 'train/grad_norm_pre_clip': 8.610223770141602, 'timing/data': 0.03241680469363928, 'timing/model': 0.9948083544149995, 'learning_rate/qwen_vl_interface': 9.155389379563934e-06, 'learning_rate/action_model': 9.155389379563933e-05, 'epoch': 0.65}) +[2026-06-17 19:00:24,156][starVLA.training.train_starvla][INFO] - Step 1431, Loss: {'train/loss': 0.35917389392852783, 'train/grad_norm_pre_clip': 8.684590339660645, 'timing/data': 0.03843617718666792, 'timing/model': 0.9412225894629955, 'learning_rate/qwen_vl_interface': 9.154157934267946e-06, 'learning_rate/action_model': 9.154157934267944e-05, 'epoch': 0.65}) +[2026-06-17 19:00:27,989][starVLA.training.train_starvla][INFO] - Step 1432, Loss: {'train/loss': 0.7123365998268127, 'train/grad_norm_pre_clip': 11.866963386535645, 'timing/data': 0.04201985988765955, 'timing/model': 0.9440329754725099, 'learning_rate/qwen_vl_interface': 9.152925679635817e-06, 'learning_rate/action_model': 9.152925679635817e-05, 'epoch': 0.65}) +[2026-06-17 19:00:32,138][starVLA.training.train_starvla][INFO] - Step 1433, Loss: {'train/loss': 0.367777019739151, 'train/grad_norm_pre_clip': 9.197443008422852, 'timing/data': 0.03398236259818077, 'timing/model': 0.9754964485764503, 'learning_rate/qwen_vl_interface': 9.151692615922995e-06, 'learning_rate/action_model': 9.151692615922994e-05, 'epoch': 0.65}) +[2026-06-17 19:00:36,443][starVLA.training.train_starvla][INFO] - Step 1434, Loss: {'train/loss': 0.2597830593585968, 'train/grad_norm_pre_clip': 6.650777339935303, 'timing/data': 0.025994778610765934, 'timing/model': 1.267483071424067, 'learning_rate/qwen_vl_interface': 9.150458743385095e-06, 'learning_rate/action_model': 9.150458743385094e-05, 'epoch': 0.65}) +[2026-06-17 19:00:40,619][starVLA.training.train_starvla][INFO] - Step 1435, Loss: {'train/loss': 0.4636216163635254, 'train/grad_norm_pre_clip': 6.74692440032959, 'timing/data': 0.028024673461914062, 'timing/model': 0.9563063783571124, 'learning_rate/qwen_vl_interface': 9.1492240622779e-06, 'learning_rate/action_model': 9.1492240622779e-05, 'epoch': 0.65}) +[2026-06-17 19:00:44,883][starVLA.training.train_starvla][INFO] - Step 1436, Loss: {'train/loss': 0.5570080876350403, 'train/grad_norm_pre_clip': 11.381263732910156, 'timing/data': 0.04011013079434633, 'timing/model': 0.9777064668014646, 'learning_rate/qwen_vl_interface': 9.147988572857362e-06, 'learning_rate/action_model': 9.147988572857362e-05, 'epoch': 0.65}) +[2026-06-17 19:00:48,960][starVLA.training.train_starvla][INFO] - Step 1437, Loss: {'train/loss': 0.7904922962188721, 'train/grad_norm_pre_clip': 15.239867210388184, 'timing/data': 0.03781087324023247, 'timing/model': 0.9772320045158267, 'learning_rate/qwen_vl_interface': 9.146752275379599e-06, 'learning_rate/action_model': 9.146752275379598e-05, 'epoch': 0.65}) +[2026-06-17 19:00:53,558][starVLA.training.train_starvla][INFO] - Step 1438, Loss: {'train/loss': 0.6683393716812134, 'train/grad_norm_pre_clip': 10.451435089111328, 'timing/data': 0.03374236263334751, 'timing/model': 0.9878623886033893, 'learning_rate/qwen_vl_interface': 9.145515170100894e-06, 'learning_rate/action_model': 9.145515170100894e-05, 'epoch': 0.65}) +[2026-06-17 19:00:57,442][starVLA.training.train_starvla][INFO] - Step 1439, Loss: {'train/loss': 0.5135670304298401, 'train/grad_norm_pre_clip': 10.11021900177002, 'timing/data': 0.033415867015719414, 'timing/model': 0.9344765301793814, 'learning_rate/qwen_vl_interface': 9.144277257277705e-06, 'learning_rate/action_model': 9.144277257277703e-05, 'epoch': 0.65}) +[2026-06-17 19:01:01,945][starVLA.training.train_starvla][INFO] - Step 1440, Loss: {'train/loss': 0.5591143369674683, 'train/grad_norm_pre_clip': 9.81113338470459, 'timing/data': 0.03257217351347208, 'timing/model': 1.4621914932504296, 'learning_rate/qwen_vl_interface': 9.14303853716665e-06, 'learning_rate/action_model': 9.143038537166649e-05, 'epoch': 0.65}) +[2026-06-17 19:01:06,703][starVLA.training.train_starvla][INFO] - Step 1441, Loss: {'train/loss': 0.5117140412330627, 'train/grad_norm_pre_clip': 9.538607597351074, 'timing/data': 0.0285197701305151, 'timing/model': 1.1867945697158575, 'learning_rate/qwen_vl_interface': 9.141799010024516e-06, 'learning_rate/action_model': 9.141799010024515e-05, 'epoch': 0.65}) +[2026-06-17 19:01:12,720][starVLA.training.train_starvla][INFO] - Step 1442, Loss: {'train/loss': 0.5053094029426575, 'train/grad_norm_pre_clip': 11.039132118225098, 'timing/data': 0.0322749949991703, 'timing/model': 2.588680971413851, 'learning_rate/qwen_vl_interface': 9.140558676108258e-06, 'learning_rate/action_model': 9.140558676108259e-05, 'epoch': 0.65}) +[2026-06-17 19:01:16,707][starVLA.training.train_starvla][INFO] - Step 1443, Loss: {'train/loss': 0.43351054191589355, 'train/grad_norm_pre_clip': 8.00684642791748, 'timing/data': 0.03718293458223343, 'timing/model': 0.9722394598647952, 'learning_rate/qwen_vl_interface': 9.139317535675002e-06, 'learning_rate/action_model': 9.139317535675002e-05, 'epoch': 0.65}) +[2026-06-17 19:01:21,818][starVLA.training.train_starvla][INFO] - Step 1444, Loss: {'train/loss': 0.3860587477684021, 'train/grad_norm_pre_clip': 9.344342231750488, 'timing/data': 0.036869822070002556, 'timing/model': 1.695002212189138, 'learning_rate/qwen_vl_interface': 9.138075588982035e-06, 'learning_rate/action_model': 9.138075588982036e-05, 'epoch': 0.66}) +[2026-06-17 19:01:25,962][starVLA.training.train_starvla][INFO] - Step 1445, Loss: {'train/loss': 0.7415388822555542, 'train/grad_norm_pre_clip': 11.997838973999023, 'timing/data': 0.03753607254475355, 'timing/model': 1.0039685331285, 'learning_rate/qwen_vl_interface': 9.136832836286816e-06, 'learning_rate/action_model': 9.136832836286816e-05, 'epoch': 0.66}) +[2026-06-17 19:01:29,962][starVLA.training.train_starvla][INFO] - Step 1446, Loss: {'train/loss': 0.4709194004535675, 'train/grad_norm_pre_clip': 11.217597007751465, 'timing/data': 0.04180415719747543, 'timing/model': 0.9509724518284202, 'learning_rate/qwen_vl_interface': 9.135589277846967e-06, 'learning_rate/action_model': 9.135589277846966e-05, 'epoch': 0.66}) +[2026-06-17 19:01:33,995][starVLA.training.train_starvla][INFO] - Step 1447, Loss: {'train/loss': 0.5015730857849121, 'train/grad_norm_pre_clip': 9.702656745910645, 'timing/data': 0.03926393389701843, 'timing/model': 1.156483800150454, 'learning_rate/qwen_vl_interface': 9.134344913920278e-06, 'learning_rate/action_model': 9.134344913920278e-05, 'epoch': 0.66}) +[2026-06-17 19:01:37,897][starVLA.training.train_starvla][INFO] - Step 1448, Loss: {'train/loss': 0.5485682487487793, 'train/grad_norm_pre_clip': 10.658528327941895, 'timing/data': 0.038844987750053406, 'timing/model': 1.1524683218449354, 'learning_rate/qwen_vl_interface': 9.133099744764711e-06, 'learning_rate/action_model': 9.133099744764711e-05, 'epoch': 0.66}) +[2026-06-17 19:01:43,976][starVLA.training.train_starvla][INFO] - Step 1449, Loss: {'train/loss': 0.4843407869338989, 'train/grad_norm_pre_clip': 8.789820671081543, 'timing/data': 0.04494677949696779, 'timing/model': 3.291196874342859, 'learning_rate/qwen_vl_interface': 9.131853770638386e-06, 'learning_rate/action_model': 9.131853770638386e-05, 'epoch': 0.66}) +[2026-06-17 19:01:48,454][starVLA.training.train_starvla][INFO] - Step 1450, Loss: {'train/loss': 0.4864922761917114, 'train/grad_norm_pre_clip': 10.045019149780273, 'timing/data': 0.035361635498702526, 'timing/model': 1.0821416983380914, 'learning_rate/qwen_vl_interface': 9.130606991799599e-06, 'learning_rate/action_model': 9.130606991799599e-05, 'epoch': 0.66}) +[2026-06-17 19:01:54,173][starVLA.training.train_starvla][INFO] - Step 1451, Loss: {'train/loss': 0.4845578968524933, 'train/grad_norm_pre_clip': 9.131589889526367, 'timing/data': 0.03288425877690315, 'timing/model': 1.9738105805590749, 'learning_rate/qwen_vl_interface': 9.129359408506808e-06, 'learning_rate/action_model': 9.129359408506808e-05, 'epoch': 0.66}) +[2026-06-17 19:02:00,547][starVLA.training.train_starvla][INFO] - Step 1452, Loss: {'train/loss': 0.7720586061477661, 'train/grad_norm_pre_clip': 16.80610466003418, 'timing/data': 0.02958780713379383, 'timing/model': 0.9344928599894047, 'learning_rate/qwen_vl_interface': 9.128111021018636e-06, 'learning_rate/action_model': 9.128111021018636e-05, 'epoch': 0.66}) +[2026-06-17 19:02:04,375][starVLA.training.train_starvla][INFO] - Step 1453, Loss: {'train/loss': 0.5355772972106934, 'train/grad_norm_pre_clip': 11.226356506347656, 'timing/data': 0.037344967015087605, 'timing/model': 0.9445639969781041, 'learning_rate/qwen_vl_interface': 9.126861829593876e-06, 'learning_rate/action_model': 9.126861829593876e-05, 'epoch': 0.66}) +[2026-06-17 19:02:08,444][starVLA.training.train_starvla][INFO] - Step 1454, Loss: {'train/loss': 0.3809255063533783, 'train/grad_norm_pre_clip': 8.866630554199219, 'timing/data': 0.030379251576960087, 'timing/model': 1.1412546895444393, 'learning_rate/qwen_vl_interface': 9.125611834491487e-06, 'learning_rate/action_model': 9.125611834491486e-05, 'epoch': 0.66}) +[2026-06-17 19:02:12,399][starVLA.training.train_starvla][INFO] - Step 1455, Loss: {'train/loss': 0.6820939779281616, 'train/grad_norm_pre_clip': 13.223885536193848, 'timing/data': 0.05163256544619799, 'timing/model': 1.224956426769495, 'learning_rate/qwen_vl_interface': 9.124361035970594e-06, 'learning_rate/action_model': 9.124361035970594e-05, 'epoch': 0.66}) +[2026-06-17 19:02:16,163][starVLA.training.train_starvla][INFO] - Step 1456, Loss: {'train/loss': 0.7024369239807129, 'train/grad_norm_pre_clip': 15.259535789489746, 'timing/data': 0.039030082523822784, 'timing/model': 0.9683716148138046, 'learning_rate/qwen_vl_interface': 9.123109434290491e-06, 'learning_rate/action_model': 9.123109434290492e-05, 'epoch': 0.66}) +[2026-06-17 19:02:20,074][starVLA.training.train_starvla][INFO] - Step 1457, Loss: {'train/loss': 0.5825302600860596, 'train/grad_norm_pre_clip': 10.042708396911621, 'timing/data': 0.04523615725338459, 'timing/model': 0.9541771505028009, 'learning_rate/qwen_vl_interface': 9.121857029710632e-06, 'learning_rate/action_model': 9.121857029710633e-05, 'epoch': 0.66}) +[2026-06-17 19:02:24,033][starVLA.training.train_starvla][INFO] - Step 1458, Loss: {'train/loss': 0.7425650358200073, 'train/grad_norm_pre_clip': 17.36577606201172, 'timing/data': 0.03233472537249327, 'timing/model': 0.9640406258404255, 'learning_rate/qwen_vl_interface': 9.120603822490647e-06, 'learning_rate/action_model': 9.120603822490647e-05, 'epoch': 0.66}) +[2026-06-17 19:02:27,861][starVLA.training.train_starvla][INFO] - Step 1459, Loss: {'train/loss': 0.47579026222229004, 'train/grad_norm_pre_clip': 9.223834037780762, 'timing/data': 0.030248582363128662, 'timing/model': 0.9431201433762908, 'learning_rate/qwen_vl_interface': 9.119349812890323e-06, 'learning_rate/action_model': 9.119349812890323e-05, 'epoch': 0.66}) +[2026-06-17 19:02:31,802][starVLA.training.train_starvla][INFO] - Step 1460, Loss: {'train/loss': 0.6134588122367859, 'train/grad_norm_pre_clip': 13.808404922485352, 'timing/data': 0.029660390689969063, 'timing/model': 1.037356123328209, 'learning_rate/qwen_vl_interface': 9.11809500116962e-06, 'learning_rate/action_model': 9.11809500116962e-05, 'epoch': 0.66}) +[2026-06-17 19:02:35,720][starVLA.training.train_starvla][INFO] - Step 1461, Loss: {'train/loss': 0.6731106042861938, 'train/grad_norm_pre_clip': 13.363826751708984, 'timing/data': 0.031376988627016544, 'timing/model': 1.2051201118156314, 'learning_rate/qwen_vl_interface': 9.116839387588662e-06, 'learning_rate/action_model': 9.116839387588661e-05, 'epoch': 0.66}) +[2026-06-17 19:02:39,390][starVLA.training.train_starvla][INFO] - Step 1462, Loss: {'train/loss': 0.815495491027832, 'train/grad_norm_pre_clip': 27.08327865600586, 'timing/data': 0.04070372320711613, 'timing/model': 0.9546516370028257, 'learning_rate/qwen_vl_interface': 9.115582972407735e-06, 'learning_rate/action_model': 9.115582972407736e-05, 'epoch': 0.66}) +[2026-06-17 19:02:43,357][starVLA.training.train_starvla][INFO] - Step 1463, Loss: {'train/loss': 0.7157092094421387, 'train/grad_norm_pre_clip': 11.525412559509277, 'timing/data': 0.035801815800368786, 'timing/model': 0.9347989531233907, 'learning_rate/qwen_vl_interface': 9.114325755887301e-06, 'learning_rate/action_model': 9.114325755887301e-05, 'epoch': 0.66}) +[2026-06-17 19:02:47,257][starVLA.training.train_starvla][INFO] - Step 1464, Loss: {'train/loss': 0.662070631980896, 'train/grad_norm_pre_clip': 13.867731094360352, 'timing/data': 0.033740587532520294, 'timing/model': 0.9546125838533044, 'learning_rate/qwen_vl_interface': 9.113067738287978e-06, 'learning_rate/action_model': 9.113067738287979e-05, 'epoch': 0.66}) +[2026-06-17 19:02:51,271][starVLA.training.train_starvla][INFO] - Step 1465, Loss: {'train/loss': 0.4236011207103729, 'train/grad_norm_pre_clip': 9.746589660644531, 'timing/data': 0.04088407754898071, 'timing/model': 0.9863546183332801, 'learning_rate/qwen_vl_interface': 9.111808919870557e-06, 'learning_rate/action_model': 9.111808919870557e-05, 'epoch': 0.66}) +[2026-06-17 19:02:55,259][starVLA.training.train_starvla][INFO] - Step 1466, Loss: {'train/loss': 0.5496517419815063, 'train/grad_norm_pre_clip': 13.15880012512207, 'timing/data': 0.03209583833813667, 'timing/model': 1.0795207917690277, 'learning_rate/qwen_vl_interface': 9.11054930089599e-06, 'learning_rate/action_model': 9.110549300895991e-05, 'epoch': 0.67}) +[2026-06-17 19:02:59,184][starVLA.training.train_starvla][INFO] - Step 1467, Loss: {'train/loss': 0.7686849236488342, 'train/grad_norm_pre_clip': 14.294820785522461, 'timing/data': 0.03297414164990187, 'timing/model': 0.9538775784894824, 'learning_rate/qwen_vl_interface': 9.109288881625402e-06, 'learning_rate/action_model': 9.109288881625402e-05, 'epoch': 0.67}) +[2026-06-17 19:03:03,123][starVLA.training.train_starvla][INFO] - Step 1468, Loss: {'train/loss': 0.7524044513702393, 'train/grad_norm_pre_clip': 12.073234558105469, 'timing/data': 0.03895490989089012, 'timing/model': 1.1873280331492424, 'learning_rate/qwen_vl_interface': 9.108027662320073e-06, 'learning_rate/action_model': 9.108027662320073e-05, 'epoch': 0.67}) +[2026-06-17 19:03:06,831][starVLA.training.train_starvla][INFO] - Step 1469, Loss: {'train/loss': 0.672125518321991, 'train/grad_norm_pre_clip': 12.382474899291992, 'timing/data': 0.031670065596699715, 'timing/model': 0.9617087924852967, 'learning_rate/qwen_vl_interface': 9.10676564324146e-06, 'learning_rate/action_model': 9.10676564324146e-05, 'epoch': 0.67}) +[2026-06-17 19:03:10,738][starVLA.training.train_starvla][INFO] - Step 1470, Loss: {'train/loss': 0.3616987466812134, 'train/grad_norm_pre_clip': 10.365785598754883, 'timing/data': 0.03746810182929039, 'timing/model': 0.9771925359964371, 'learning_rate/qwen_vl_interface': 9.105502824651179e-06, 'learning_rate/action_model': 9.105502824651178e-05, 'epoch': 0.67}) +[2026-06-17 19:03:16,147][starVLA.training.train_starvla][INFO] - Step 1471, Loss: {'train/loss': 0.41531240940093994, 'train/grad_norm_pre_clip': 9.004203796386719, 'timing/data': 0.03355951141566038, 'timing/model': 1.0021159127354622, 'learning_rate/qwen_vl_interface': 9.104239206811012e-06, 'learning_rate/action_model': 9.104239206811011e-05, 'epoch': 0.67}) +[2026-06-17 19:03:20,179][starVLA.training.train_starvla][INFO] - Step 1472, Loss: {'train/loss': 0.4612323045730591, 'train/grad_norm_pre_clip': 9.496814727783203, 'timing/data': 0.04349437728524208, 'timing/model': 0.9325491664931178, 'learning_rate/qwen_vl_interface': 9.102974789982911e-06, 'learning_rate/action_model': 9.102974789982912e-05, 'epoch': 0.67}) +[2026-06-17 19:03:24,133][starVLA.training.train_starvla][INFO] - Step 1473, Loss: {'train/loss': 0.47276178002357483, 'train/grad_norm_pre_clip': 12.735939025878906, 'timing/data': 0.03913608007133007, 'timing/model': 0.9547730488702655, 'learning_rate/qwen_vl_interface': 9.10170957442899e-06, 'learning_rate/action_model': 9.101709574428991e-05, 'epoch': 0.67}) +[2026-06-17 19:03:28,050][starVLA.training.train_starvla][INFO] - Step 1474, Loss: {'train/loss': 0.796228289604187, 'train/grad_norm_pre_clip': 17.26262855529785, 'timing/data': 0.030989428982138634, 'timing/model': 0.9436754956841469, 'learning_rate/qwen_vl_interface': 9.10044356041153e-06, 'learning_rate/action_model': 9.10044356041153e-05, 'epoch': 0.67}) +[2026-06-17 19:03:31,930][starVLA.training.train_starvla][INFO] - Step 1475, Loss: {'train/loss': 0.6043306589126587, 'train/grad_norm_pre_clip': 13.991570472717285, 'timing/data': 0.03085277695208788, 'timing/model': 1.1568430606275797, 'learning_rate/qwen_vl_interface': 9.099176748192976e-06, 'learning_rate/action_model': 9.099176748192976e-05, 'epoch': 0.67}) +[2026-06-17 19:03:35,746][starVLA.training.train_starvla][INFO] - Step 1476, Loss: {'train/loss': 0.322742760181427, 'train/grad_norm_pre_clip': 9.595170021057129, 'timing/data': 0.03824992571026087, 'timing/model': 1.086840140633285, 'learning_rate/qwen_vl_interface': 9.097909138035943e-06, 'learning_rate/action_model': 9.097909138035941e-05, 'epoch': 0.67}) +[2026-06-17 19:03:39,692][starVLA.training.train_starvla][INFO] - Step 1477, Loss: {'train/loss': 0.5020407438278198, 'train/grad_norm_pre_clip': 10.488414764404297, 'timing/data': 0.03608717955648899, 'timing/model': 0.95116410497576, 'learning_rate/qwen_vl_interface': 9.096640730203202e-06, 'learning_rate/action_model': 9.096640730203202e-05, 'epoch': 0.67}) +[2026-06-17 19:03:43,718][starVLA.training.train_starvla][INFO] - Step 1478, Loss: {'train/loss': 1.001105785369873, 'train/grad_norm_pre_clip': 15.212852478027344, 'timing/data': 0.03514937125146389, 'timing/model': 1.0036819046363235, 'learning_rate/qwen_vl_interface': 9.095371524957698e-06, 'learning_rate/action_model': 9.095371524957698e-05, 'epoch': 0.67}) +[2026-06-17 19:03:47,616][starVLA.training.train_starvla][INFO] - Step 1479, Loss: {'train/loss': 0.5433183312416077, 'train/grad_norm_pre_clip': 9.64801025390625, 'timing/data': 0.0288544287905097, 'timing/model': 0.9571993546560407, 'learning_rate/qwen_vl_interface': 9.094101522562541e-06, 'learning_rate/action_model': 9.09410152256254e-05, 'epoch': 0.67}) +[2026-06-17 19:03:51,634][starVLA.training.train_starvla][INFO] - Step 1480, Loss: {'train/loss': 0.4192759096622467, 'train/grad_norm_pre_clip': 8.144013404846191, 'timing/data': 0.028393038548529148, 'timing/model': 1.0306381676346064, 'learning_rate/qwen_vl_interface': 9.092830723281e-06, 'learning_rate/action_model': 9.092830723281e-05, 'epoch': 0.67}) +[2026-06-17 19:03:55,465][starVLA.training.train_starvla][INFO] - Step 1481, Loss: {'train/loss': 0.7216877937316895, 'train/grad_norm_pre_clip': 11.572998046875, 'timing/data': 0.02985554374754429, 'timing/model': 1.1526529481634498, 'learning_rate/qwen_vl_interface': 9.091559127376519e-06, 'learning_rate/action_model': 9.091559127376517e-05, 'epoch': 0.67}) +[2026-06-17 19:04:03,270][starVLA.training.train_starvla][INFO] - Step 1482, Loss: {'train/loss': 0.6996365189552307, 'train/grad_norm_pre_clip': 11.046492576599121, 'timing/data': 0.036562913097441196, 'timing/model': 1.2201673816889524, 'learning_rate/qwen_vl_interface': 9.090286735112693e-06, 'learning_rate/action_model': 9.090286735112693e-05, 'epoch': 0.67}) +[2026-06-17 19:04:07,103][starVLA.training.train_starvla][INFO] - Step 1483, Loss: {'train/loss': 0.5125395655632019, 'train/grad_norm_pre_clip': 9.250471115112305, 'timing/data': 0.0322023369371891, 'timing/model': 1.009375847876072, 'learning_rate/qwen_vl_interface': 9.089013546753296e-06, 'learning_rate/action_model': 9.089013546753295e-05, 'epoch': 0.67}) +[2026-06-17 19:04:11,069][starVLA.training.train_starvla][INFO] - Step 1484, Loss: {'train/loss': 0.7488415241241455, 'train/grad_norm_pre_clip': 26.011396408081055, 'timing/data': 0.04133136197924614, 'timing/model': 0.9472696268931031, 'learning_rate/qwen_vl_interface': 9.08773956256226e-06, 'learning_rate/action_model': 9.087739562562259e-05, 'epoch': 0.67}) +[2026-06-17 19:04:15,105][starVLA.training.train_starvla][INFO] - Step 1485, Loss: {'train/loss': 0.4416014850139618, 'train/grad_norm_pre_clip': 9.358305931091309, 'timing/data': 0.043366990983486176, 'timing/model': 1.0191243616864085, 'learning_rate/qwen_vl_interface': 9.086464782803682e-06, 'learning_rate/action_model': 9.086464782803682e-05, 'epoch': 0.67}) +[2026-06-17 19:04:19,201][starVLA.training.train_starvla][INFO] - Step 1486, Loss: {'train/loss': 0.6206875443458557, 'train/grad_norm_pre_clip': 10.370339393615723, 'timing/data': 0.03405959438532591, 'timing/model': 0.9436177955940366, 'learning_rate/qwen_vl_interface': 9.085189207741827e-06, 'learning_rate/action_model': 9.085189207741827e-05, 'epoch': 0.67}) +[2026-06-17 19:04:23,234][starVLA.training.train_starvla][INFO] - Step 1487, Loss: {'train/loss': 0.47879043221473694, 'train/grad_norm_pre_clip': 8.833651542663574, 'timing/data': 0.03243510331958532, 'timing/model': 1.043965583667159, 'learning_rate/qwen_vl_interface': 9.083912837641123e-06, 'learning_rate/action_model': 9.083912837641123e-05, 'epoch': 0.67}) +[2026-06-17 19:04:27,389][starVLA.training.train_starvla][INFO] - Step 1488, Loss: {'train/loss': 0.6028919219970703, 'train/grad_norm_pre_clip': 16.947795867919922, 'timing/data': 0.03363776579499245, 'timing/model': 1.4187048031017184, 'learning_rate/qwen_vl_interface': 9.082635672766162e-06, 'learning_rate/action_model': 9.082635672766161e-05, 'epoch': 0.68}) +[2026-06-17 19:04:31,295][starVLA.training.train_starvla][INFO] - Step 1489, Loss: {'train/loss': 0.5815122127532959, 'train/grad_norm_pre_clip': 10.515406608581543, 'timing/data': 0.0474082101136446, 'timing/model': 1.1417336268350482, 'learning_rate/qwen_vl_interface': 9.081357713381702e-06, 'learning_rate/action_model': 9.081357713381701e-05, 'epoch': 0.68}) +[2026-06-17 19:04:37,046][starVLA.training.train_starvla][INFO] - Step 1490, Loss: {'train/loss': 0.5758499503135681, 'train/grad_norm_pre_clip': 9.567614555358887, 'timing/data': 0.040762449614703655, 'timing/model': 0.9355794284492731, 'learning_rate/qwen_vl_interface': 9.080078959752664e-06, 'learning_rate/action_model': 9.080078959752664e-05, 'epoch': 0.68}) +[2026-06-17 19:04:40,883][starVLA.training.train_starvla][INFO] - Step 1491, Loss: {'train/loss': 0.7556575536727905, 'train/grad_norm_pre_clip': 13.896450996398926, 'timing/data': 0.036736806854605675, 'timing/model': 0.9482134180143476, 'learning_rate/qwen_vl_interface': 9.078799412144137e-06, 'learning_rate/action_model': 9.078799412144137e-05, 'epoch': 0.68}) +[2026-06-17 19:04:44,896][starVLA.training.train_starvla][INFO] - Step 1492, Loss: {'train/loss': 0.8388419151306152, 'train/grad_norm_pre_clip': 19.56723403930664, 'timing/data': 0.03679715562611818, 'timing/model': 1.0188697976991534, 'learning_rate/qwen_vl_interface': 9.077519070821373e-06, 'learning_rate/action_model': 9.077519070821373e-05, 'epoch': 0.68}) +[2026-06-17 19:04:48,756][starVLA.training.train_starvla][INFO] - Step 1493, Loss: {'train/loss': 0.6196243166923523, 'train/grad_norm_pre_clip': 10.682366371154785, 'timing/data': 0.03188535664230585, 'timing/model': 0.9319627108052373, 'learning_rate/qwen_vl_interface': 9.076237936049787e-06, 'learning_rate/action_model': 9.076237936049786e-05, 'epoch': 0.68}) +[2026-06-17 19:04:52,534][starVLA.training.train_starvla][INFO] - Step 1494, Loss: {'train/loss': 0.6778905391693115, 'train/grad_norm_pre_clip': 17.525794982910156, 'timing/data': 0.04165446572005749, 'timing/model': 0.9434642577543855, 'learning_rate/qwen_vl_interface': 9.074956008094958e-06, 'learning_rate/action_model': 9.074956008094958e-05, 'epoch': 0.68}) +[2026-06-17 19:04:56,345][starVLA.training.train_starvla][INFO] - Step 1495, Loss: {'train/loss': 0.6242114305496216, 'train/grad_norm_pre_clip': 9.881769180297852, 'timing/data': 0.03233405761420727, 'timing/model': 1.1036615893244743, 'learning_rate/qwen_vl_interface': 9.073673287222634e-06, 'learning_rate/action_model': 9.073673287222633e-05, 'epoch': 0.68}) +[2026-06-17 19:05:00,057][starVLA.training.train_starvla][INFO] - Step 1496, Loss: {'train/loss': 0.6336351037025452, 'train/grad_norm_pre_clip': 10.619243621826172, 'timing/data': 0.03188537713140249, 'timing/model': 1.0138531848788261, 'learning_rate/qwen_vl_interface': 9.072389773698723e-06, 'learning_rate/action_model': 9.072389773698723e-05, 'epoch': 0.68}) +[2026-06-17 19:05:04,043][starVLA.training.train_starvla][INFO] - Step 1497, Loss: {'train/loss': 0.8930846452713013, 'train/grad_norm_pre_clip': 12.722841262817383, 'timing/data': 0.03650292381644249, 'timing/model': 0.9558273814618587, 'learning_rate/qwen_vl_interface': 9.071105467789299e-06, 'learning_rate/action_model': 9.071105467789298e-05, 'epoch': 0.68}) +[2026-06-17 19:05:07,904][starVLA.training.train_starvla][INFO] - Step 1498, Loss: {'train/loss': 0.596053957939148, 'train/grad_norm_pre_clip': 8.109673500061035, 'timing/data': 0.03390622418373823, 'timing/model': 0.9585102656856179, 'learning_rate/qwen_vl_interface': 9.0698203697606e-06, 'learning_rate/action_model': 9.0698203697606e-05, 'epoch': 0.68}) +[2026-06-17 19:05:11,851][starVLA.training.train_starvla][INFO] - Step 1499, Loss: {'train/loss': 0.6208012700080872, 'train/grad_norm_pre_clip': 10.046563148498535, 'timing/data': 0.027334040962159634, 'timing/model': 0.9636501930654049, 'learning_rate/qwen_vl_interface': 9.068534479879026e-06, 'learning_rate/action_model': 9.068534479879024e-05, 'epoch': 0.68}) +[2026-06-17 19:07:27,425][starVLA.training.train_starvla][INFO] - Step 1500, Loss: {'train/loss': 0.613971471786499, 'train/grad_norm_pre_clip': 11.20779037475586, 'eval/loss': 0.5199655294418335, 'eval/latency_8/loss': 0.5199655294418335, 'eval/demon_fire_precision': 0.9282550930026572, 'eval/demon_fire_recall': 0.9767008387698043, 'eval/demon_fire_f1': 0.9518619436875566, 'eval/demon_fire_rate_model': 0.9424040066777963, 'eval/demon_fire_rate_teacher': 0.8956594323873122, 'eval/demon_left_precision': 0.7461300309597523, 'eval/demon_left_recall': 0.5160599571734475, 'eval/demon_left_f1': 0.6101265822784809, 'eval/demon_left_rate_model': 0.08987200890372844, 'eval/demon_left_rate_teacher': 0.12993878686700056, 'eval/demon_right_precision': 0.7198142414860681, 'eval/demon_right_recall': 0.7045454545454546, 'eval/demon_right_f1': 0.7120980091883615, 'eval/demon_right_rate_model': 0.17974401780745689, 'eval/demon_right_rate_teacher': 0.18363939899833054, 'eval/demon_move_f1': 0.6611122957334212, 'eval/demon_cc_f1': 0.8409145544563029, 'eval/latency_8/demon_fire_precision': 0.9282550930026572, 'eval/latency_8/demon_fire_recall': 0.9767008387698043, 'eval/latency_8/demon_fire_f1': 0.9518619436875566, 'eval/latency_8/demon_fire_rate_model': 0.9424040066777963, 'eval/latency_8/demon_fire_rate_teacher': 0.8956594323873122, 'eval/latency_8/demon_left_precision': 0.7461300309597523, 'eval/latency_8/demon_left_recall': 0.5160599571734475, 'eval/latency_8/demon_left_f1': 0.6101265822784809, 'eval/latency_8/demon_left_rate_model': 0.08987200890372844, 'eval/latency_8/demon_left_rate_teacher': 0.12993878686700056, 'eval/latency_8/demon_right_precision': 0.7198142414860681, 'eval/latency_8/demon_right_recall': 0.7045454545454546, 'eval/latency_8/demon_right_f1': 0.7120980091883615, 'eval/latency_8/demon_right_rate_model': 0.17974401780745689, 'eval/latency_8/demon_right_rate_teacher': 0.18363939899833054, 'eval/latency_8/demon_move_f1': 0.6611122957334212, 'eval/latency_8/demon_cc_f1': 0.8409145544563029, 'timing/data': 0.04330671764910221, 'timing/model': 0.9364703372120857, 'learning_rate/qwen_vl_interface': 9.067247798411147e-06, 'learning_rate/action_model': 9.067247798411147e-05, 'epoch': 0.68}) +[2026-06-17 19:07:27,427][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_1500_state +[2026-06-17 19:07:27,432][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 19:07:34,301][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_1500_state/model.safetensors +[2026-06-17 19:07:50,710][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_1500_state/optimizer.bin +[2026-06-17 19:07:50,713][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_1500_state/sampler.bin +[2026-06-17 19:07:50,713][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_1500_state/sampler_1.bin +[2026-06-17 19:07:50,714][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_1500_state/random_states_0.pkl +[2026-06-17 19:07:50,718][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1500 +[2026-06-17 19:07:50,718][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 19:07:50,729][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 19:07:54,753][starVLA.training.train_starvla][INFO] - Step 1501, Loss: {'train/loss': 0.6186231970787048, 'train/grad_norm_pre_clip': 11.262228965759277, 'timing/data': 0.04341842886060476, 'timing/model': 0.938599850051105, 'learning_rate/qwen_vl_interface': 9.065960325623692e-06, 'learning_rate/action_model': 9.065960325623692e-05, 'epoch': 0.68}) +[2026-06-17 19:07:58,597][starVLA.training.train_starvla][INFO] - Step 1502, Loss: {'train/loss': 0.6881040334701538, 'train/grad_norm_pre_clip': 11.208597183227539, 'timing/data': 0.034677525982260704, 'timing/model': 0.9409911017864943, 'learning_rate/qwen_vl_interface': 9.064672061783555e-06, 'learning_rate/action_model': 9.064672061783554e-05, 'epoch': 0.68}) +[2026-06-17 19:08:02,473][starVLA.training.train_starvla][INFO] - Step 1503, Loss: {'train/loss': 0.546019971370697, 'train/grad_norm_pre_clip': 8.931676864624023, 'timing/data': 0.04214709252119064, 'timing/model': 0.9312936244532466, 'learning_rate/qwen_vl_interface': 9.063383007157795e-06, 'learning_rate/action_model': 9.063383007157796e-05, 'epoch': 0.68}) +[2026-06-17 19:08:08,873][starVLA.training.train_starvla][INFO] - Step 1504, Loss: {'train/loss': 0.6352592706680298, 'train/grad_norm_pre_clip': 19.188535690307617, 'timing/data': 0.02706361562013626, 'timing/model': 3.680924654006958, 'learning_rate/qwen_vl_interface': 9.062093162013635e-06, 'learning_rate/action_model': 9.062093162013634e-05, 'epoch': 0.68}) +[2026-06-17 19:08:12,701][starVLA.training.train_starvla][INFO] - Step 1505, Loss: {'train/loss': 1.0347517728805542, 'train/grad_norm_pre_clip': 15.677579879760742, 'timing/data': 0.04683843348175287, 'timing/model': 1.049513672478497, 'learning_rate/qwen_vl_interface': 9.06080252661846e-06, 'learning_rate/action_model': 9.060802526618458e-05, 'epoch': 0.68}) +[2026-06-17 19:08:16,608][starVLA.training.train_starvla][INFO] - Step 1506, Loss: {'train/loss': 0.5392317771911621, 'train/grad_norm_pre_clip': 9.852616310119629, 'timing/data': 0.037856968119740486, 'timing/model': 0.9888157295063138, 'learning_rate/qwen_vl_interface': 9.059511101239818e-06, 'learning_rate/action_model': 9.059511101239818e-05, 'epoch': 0.68}) +[2026-06-17 19:08:20,539][starVLA.training.train_starvla][INFO] - Step 1507, Loss: {'train/loss': 0.580036461353302, 'train/grad_norm_pre_clip': 9.825765609741211, 'timing/data': 0.05097529198974371, 'timing/model': 0.9460214963182807, 'learning_rate/qwen_vl_interface': 9.058218886145427e-06, 'learning_rate/action_model': 9.058218886145427e-05, 'epoch': 0.68}) +[2026-06-17 19:08:24,364][starVLA.training.train_starvla][INFO] - Step 1508, Loss: {'train/loss': 0.4519323408603668, 'train/grad_norm_pre_clip': 10.557341575622559, 'timing/data': 0.029249416664242744, 'timing/model': 0.9422696307301521, 'learning_rate/qwen_vl_interface': 9.056925881603165e-06, 'learning_rate/action_model': 9.056925881603166e-05, 'epoch': 0.68}) +[2026-06-17 19:08:28,308][starVLA.training.train_starvla][INFO] - Step 1509, Loss: {'train/loss': 0.5673554539680481, 'train/grad_norm_pre_clip': 14.130592346191406, 'timing/data': 0.03153287526220083, 'timing/model': 0.944591180421412, 'learning_rate/qwen_vl_interface': 9.05563208788107e-06, 'learning_rate/action_model': 9.05563208788107e-05, 'epoch': 0.68}) +[2026-06-17 19:08:32,488][starVLA.training.train_starvla][INFO] - Step 1510, Loss: {'train/loss': 0.539404034614563, 'train/grad_norm_pre_clip': 11.293949127197266, 'timing/data': 0.03373570740222931, 'timing/model': 0.9849561201408505, 'learning_rate/qwen_vl_interface': 9.054337505247349e-06, 'learning_rate/action_model': 9.054337505247348e-05, 'epoch': 0.69}) +[2026-06-17 19:08:37,039][starVLA.training.train_starvla][INFO] - Step 1511, Loss: {'train/loss': 0.3588119149208069, 'train/grad_norm_pre_clip': 13.768948554992676, 'timing/data': 0.029976909048855305, 'timing/model': 1.1390979131683707, 'learning_rate/qwen_vl_interface': 9.05304213397037e-06, 'learning_rate/action_model': 9.053042133970369e-05, 'epoch': 0.69}) +[2026-06-17 19:08:41,264][starVLA.training.train_starvla][INFO] - Step 1512, Loss: {'train/loss': 0.4435155391693115, 'train/grad_norm_pre_clip': 10.204058647155762, 'timing/data': 0.04048214387148619, 'timing/model': 0.97593186609447, 'learning_rate/qwen_vl_interface': 9.051745974318663e-06, 'learning_rate/action_model': 9.051745974318662e-05, 'epoch': 0.69}) +[2026-06-17 19:08:45,248][starVLA.training.train_starvla][INFO] - Step 1513, Loss: {'train/loss': 0.4194866120815277, 'train/grad_norm_pre_clip': 10.201981544494629, 'timing/data': 0.045790961012244225, 'timing/model': 0.9548757299780846, 'learning_rate/qwen_vl_interface': 9.050449026560923e-06, 'learning_rate/action_model': 9.050449026560924e-05, 'epoch': 0.69}) +[2026-06-17 19:08:49,146][starVLA.training.train_starvla][INFO] - Step 1514, Loss: {'train/loss': 0.3889693021774292, 'train/grad_norm_pre_clip': 10.086341857910156, 'timing/data': 0.04091112595051527, 'timing/model': 0.9407745227217674, 'learning_rate/qwen_vl_interface': 9.049151290966013e-06, 'learning_rate/action_model': 9.049151290966012e-05, 'epoch': 0.69}) +[2026-06-17 19:08:53,557][starVLA.training.train_starvla][INFO] - Step 1515, Loss: {'train/loss': 0.4202503263950348, 'train/grad_norm_pre_clip': 10.83552074432373, 'timing/data': 0.03859707247465849, 'timing/model': 0.9557629693299532, 'learning_rate/qwen_vl_interface': 9.04785276780295e-06, 'learning_rate/action_model': 9.04785276780295e-05, 'epoch': 0.69}) +[2026-06-17 19:08:57,601][starVLA.training.train_starvla][INFO] - Step 1516, Loss: {'train/loss': 0.6097668409347534, 'train/grad_norm_pre_clip': 11.66299057006836, 'timing/data': 0.03210418298840523, 'timing/model': 0.9455556981265545, 'learning_rate/qwen_vl_interface': 9.046553457340924e-06, 'learning_rate/action_model': 9.046553457340924e-05, 'epoch': 0.69}) +[2026-06-17 19:09:03,430][starVLA.training.train_starvla][INFO] - Step 1517, Loss: {'train/loss': 0.5665950179100037, 'train/grad_norm_pre_clip': 11.982457160949707, 'timing/data': 0.029693915508687496, 'timing/model': 0.9553077258169651, 'learning_rate/qwen_vl_interface': 9.04525335984928e-06, 'learning_rate/action_model': 9.04525335984928e-05, 'epoch': 0.69}) +[2026-06-17 19:09:07,446][starVLA.training.train_starvla][INFO] - Step 1518, Loss: {'train/loss': 0.26692819595336914, 'train/grad_norm_pre_clip': 6.688277244567871, 'timing/data': 0.0358551312237978, 'timing/model': 1.2095503909513354, 'learning_rate/qwen_vl_interface': 9.043952475597532e-06, 'learning_rate/action_model': 9.043952475597532e-05, 'epoch': 0.69}) +[2026-06-17 19:09:12,343][starVLA.training.train_starvla][INFO] - Step 1519, Loss: {'train/loss': 0.8685727715492249, 'train/grad_norm_pre_clip': 16.1449031829834, 'timing/data': 0.017067544162273407, 'timing/model': 0.9887990346178412, 'learning_rate/qwen_vl_interface': 9.04265080485535e-06, 'learning_rate/action_model': 9.042650804855351e-05, 'epoch': 0.69}) +[2026-06-17 19:09:17,870][starVLA.training.train_starvla][INFO] - Step 1520, Loss: {'train/loss': 0.530457615852356, 'train/grad_norm_pre_clip': 17.24138641357422, 'timing/data': 0.01863066852092743, 'timing/model': 0.9529511574655771, 'learning_rate/qwen_vl_interface': 9.041348347892578e-06, 'learning_rate/action_model': 9.041348347892579e-05, 'epoch': 0.69}) +[2026-06-17 19:09:21,898][starVLA.training.train_starvla][INFO] - Step 1521, Loss: {'train/loss': 0.8070368766784668, 'train/grad_norm_pre_clip': 17.721052169799805, 'timing/data': 0.02279889304190874, 'timing/model': 0.9644307093694806, 'learning_rate/qwen_vl_interface': 9.040045104979213e-06, 'learning_rate/action_model': 9.040045104979214e-05, 'epoch': 0.69}) +[2026-06-17 19:09:25,763][starVLA.training.train_starvla][INFO] - Step 1522, Loss: {'train/loss': 0.9339216947555542, 'train/grad_norm_pre_clip': 16.199514389038086, 'timing/data': 0.020222040824592113, 'timing/model': 0.9487398127093911, 'learning_rate/qwen_vl_interface': 9.03874107638542e-06, 'learning_rate/action_model': 9.038741076385419e-05, 'epoch': 0.69}) +[2026-06-17 19:09:29,732][starVLA.training.train_starvla][INFO] - Step 1523, Loss: {'train/loss': 0.6232343316078186, 'train/grad_norm_pre_clip': 11.500589370727539, 'timing/data': 0.018452879041433334, 'timing/model': 0.9691652227193117, 'learning_rate/qwen_vl_interface': 9.037436262381525e-06, 'learning_rate/action_model': 9.037436262381525e-05, 'epoch': 0.69}) +[2026-06-17 19:09:33,681][starVLA.training.train_starvla][INFO] - Step 1524, Loss: {'train/loss': 0.4908335208892822, 'train/grad_norm_pre_clip': 11.762452125549316, 'timing/data': 0.019776002503931522, 'timing/model': 1.1667383769527078, 'learning_rate/qwen_vl_interface': 9.036130663238018e-06, 'learning_rate/action_model': 9.036130663238017e-05, 'epoch': 0.69}) +[2026-06-17 19:09:37,540][starVLA.training.train_starvla][INFO] - Step 1525, Loss: {'train/loss': 0.568227231502533, 'train/grad_norm_pre_clip': 15.92747688293457, 'timing/data': 0.0189999807626009, 'timing/model': 1.1172418538480997, 'learning_rate/qwen_vl_interface': 9.034824279225547e-06, 'learning_rate/action_model': 9.034824279225547e-05, 'epoch': 0.69}) +[2026-06-17 19:09:41,261][starVLA.training.train_starvla][INFO] - Step 1526, Loss: {'train/loss': 0.5133009552955627, 'train/grad_norm_pre_clip': 13.190919876098633, 'timing/data': 0.021176166832447052, 'timing/model': 0.9319226481020451, 'learning_rate/qwen_vl_interface': 9.033517110614933e-06, 'learning_rate/action_model': 9.033517110614933e-05, 'epoch': 0.69}) +[2026-06-17 19:09:45,317][starVLA.training.train_starvla][INFO] - Step 1527, Loss: {'train/loss': 0.6016645431518555, 'train/grad_norm_pre_clip': 13.828235626220703, 'timing/data': 0.02200111746788025, 'timing/model': 0.9629673436284065, 'learning_rate/qwen_vl_interface': 9.032209157677151e-06, 'learning_rate/action_model': 9.032209157677151e-05, 'epoch': 0.69}) +[2026-06-17 19:09:49,530][starVLA.training.train_starvla][INFO] - Step 1528, Loss: {'train/loss': 0.6861265301704407, 'train/grad_norm_pre_clip': 13.253881454467773, 'timing/data': 0.027753976173698902, 'timing/model': 0.9611497521400452, 'learning_rate/qwen_vl_interface': 9.03090042068334e-06, 'learning_rate/action_model': 9.030900420683339e-05, 'epoch': 0.69}) +[2026-06-17 19:09:53,556][starVLA.training.train_starvla][INFO] - Step 1529, Loss: {'train/loss': 0.29187867045402527, 'train/grad_norm_pre_clip': 8.218562126159668, 'timing/data': 0.024920961819589138, 'timing/model': 0.9576698923483491, 'learning_rate/qwen_vl_interface': 9.029590899904803e-06, 'learning_rate/action_model': 9.029590899904803e-05, 'epoch': 0.69}) +[2026-06-17 19:09:59,014][starVLA.training.train_starvla][INFO] - Step 1530, Loss: {'train/loss': 0.6131033897399902, 'train/grad_norm_pre_clip': 14.347278594970703, 'timing/data': 0.020284338854253292, 'timing/model': 0.9557837871834636, 'learning_rate/qwen_vl_interface': 9.028280595613006e-06, 'learning_rate/action_model': 9.028280595613005e-05, 'epoch': 0.69}) +[2026-06-17 19:10:03,192][starVLA.training.train_starvla][INFO] - Step 1531, Loss: {'train/loss': 0.6888622045516968, 'train/grad_norm_pre_clip': 12.913582801818848, 'timing/data': 0.014270562678575516, 'timing/model': 1.177521805278957, 'learning_rate/qwen_vl_interface': 9.026969508079575e-06, 'learning_rate/action_model': 9.026969508079575e-05, 'epoch': 0.69}) +[2026-06-17 19:10:06,999][starVLA.training.train_starvla][INFO] - Step 1532, Loss: {'train/loss': 0.6842855215072632, 'train/grad_norm_pre_clip': 11.531420707702637, 'timing/data': 0.01710470300167799, 'timing/model': 0.9619189854711294, 'learning_rate/qwen_vl_interface': 9.025657637576302e-06, 'learning_rate/action_model': 9.025657637576302e-05, 'epoch': 0.7}) +[2026-06-17 19:10:10,909][starVLA.training.train_starvla][INFO] - Step 1533, Loss: {'train/loss': 0.5502923727035522, 'train/grad_norm_pre_clip': 12.1748046875, 'timing/data': 0.015359138138592243, 'timing/model': 0.9581080805510283, 'learning_rate/qwen_vl_interface': 9.024344984375138e-06, 'learning_rate/action_model': 9.024344984375137e-05, 'epoch': 0.7}) +[2026-06-17 19:10:14,755][starVLA.training.train_starvla][INFO] - Step 1534, Loss: {'train/loss': 0.5516810417175293, 'train/grad_norm_pre_clip': 17.79512596130371, 'timing/data': 0.018128447234630585, 'timing/model': 0.9682623418048024, 'learning_rate/qwen_vl_interface': 9.023031548748197e-06, 'learning_rate/action_model': 9.023031548748196e-05, 'epoch': 0.7}) +[2026-06-17 19:10:18,675][starVLA.training.train_starvla][INFO] - Step 1535, Loss: {'train/loss': 0.6034072637557983, 'train/grad_norm_pre_clip': 11.740407943725586, 'timing/data': 0.017596201971173286, 'timing/model': 0.9530282476916909, 'learning_rate/qwen_vl_interface': 9.021717330967755e-06, 'learning_rate/action_model': 9.021717330967754e-05, 'epoch': 0.7}) +[2026-06-17 19:10:22,704][starVLA.training.train_starvla][INFO] - Step 1536, Loss: {'train/loss': 0.5072395205497742, 'train/grad_norm_pre_clip': 12.601751327514648, 'timing/data': 0.017694968730211258, 'timing/model': 1.0360454069450498, 'learning_rate/qwen_vl_interface': 9.020402331306254e-06, 'learning_rate/action_model': 9.020402331306253e-05, 'epoch': 0.7}) +[2026-06-17 19:10:26,727][starVLA.training.train_starvla][INFO] - Step 1537, Loss: {'train/loss': 0.9375707507133484, 'train/grad_norm_pre_clip': 16.382492065429688, 'timing/data': 0.01816362701356411, 'timing/model': 1.169492264278233, 'learning_rate/qwen_vl_interface': 9.01908655003629e-06, 'learning_rate/action_model': 9.01908655003629e-05, 'epoch': 0.7}) +[2026-06-17 19:10:30,664][starVLA.training.train_starvla][INFO] - Step 1538, Loss: {'train/loss': 0.906390368938446, 'train/grad_norm_pre_clip': 14.603869438171387, 'timing/data': 0.02112750057131052, 'timing/model': 1.1633160272613168, 'learning_rate/qwen_vl_interface': 9.017769987430629e-06, 'learning_rate/action_model': 9.017769987430628e-05, 'epoch': 0.7}) +[2026-06-17 19:10:34,445][starVLA.training.train_starvla][INFO] - Step 1539, Loss: {'train/loss': 0.5232291221618652, 'train/grad_norm_pre_clip': 9.993614196777344, 'timing/data': 0.023424060083925724, 'timing/model': 0.9715630626305938, 'learning_rate/qwen_vl_interface': 9.016452643762197e-06, 'learning_rate/action_model': 9.016452643762197e-05, 'epoch': 0.7}) +[2026-06-17 19:10:39,705][starVLA.training.train_starvla][INFO] - Step 1540, Loss: {'train/loss': 0.22759605944156647, 'train/grad_norm_pre_clip': 6.464902877807617, 'timing/data': 0.02525664120912552, 'timing/model': 1.0042521534487605, 'learning_rate/qwen_vl_interface': 9.015134519304078e-06, 'learning_rate/action_model': 9.015134519304078e-05, 'epoch': 0.7}) +[2026-06-17 19:10:43,567][starVLA.training.train_starvla][INFO] - Step 1541, Loss: {'train/loss': 0.2469635009765625, 'train/grad_norm_pre_clip': 7.1554694175720215, 'timing/data': 0.02169360127300024, 'timing/model': 0.9429255900904536, 'learning_rate/qwen_vl_interface': 9.013815614329521e-06, 'learning_rate/action_model': 9.013815614329522e-05, 'epoch': 0.7}) +[2026-06-17 19:10:48,367][starVLA.training.train_starvla][INFO] - Step 1542, Loss: {'train/loss': 0.3996507525444031, 'train/grad_norm_pre_clip': 9.016101837158203, 'timing/data': 0.01952529978007078, 'timing/model': 0.9453073479235172, 'learning_rate/qwen_vl_interface': 9.01249592911194e-06, 'learning_rate/action_model': 9.012495929111938e-05, 'epoch': 0.7}) +[2026-06-17 19:10:56,254][starVLA.training.train_starvla][INFO] - Step 1543, Loss: {'train/loss': 0.3315352201461792, 'train/grad_norm_pre_clip': 6.749128341674805, 'timing/data': 0.018627685494720936, 'timing/model': 0.955641089938581, 'learning_rate/qwen_vl_interface': 9.0111754639249e-06, 'learning_rate/action_model': 9.0111754639249e-05, 'epoch': 0.7}) +[2026-06-17 19:11:01,268][starVLA.training.train_starvla][INFO] - Step 1544, Loss: {'train/loss': 0.5396968722343445, 'train/grad_norm_pre_clip': 18.637901306152344, 'timing/data': 0.017009706236422062, 'timing/model': 1.1708875624462962, 'learning_rate/qwen_vl_interface': 9.009854219042142e-06, 'learning_rate/action_model': 9.009854219042142e-05, 'epoch': 0.7}) +[2026-06-17 19:11:05,225][starVLA.training.train_starvla][INFO] - Step 1545, Loss: {'train/loss': 0.46286314725875854, 'train/grad_norm_pre_clip': 9.329090118408203, 'timing/data': 0.025778518058359623, 'timing/model': 1.1706012263894081, 'learning_rate/qwen_vl_interface': 9.008532194737556e-06, 'learning_rate/action_model': 9.008532194737557e-05, 'epoch': 0.7}) +[2026-06-17 19:11:08,906][starVLA.training.train_starvla][INFO] - Step 1546, Loss: {'train/loss': 0.7196484804153442, 'train/grad_norm_pre_clip': 13.229348182678223, 'timing/data': 0.01744471676647663, 'timing/model': 0.9586540134623647, 'learning_rate/qwen_vl_interface': 9.007209391285205e-06, 'learning_rate/action_model': 9.007209391285205e-05, 'epoch': 0.7}) +[2026-06-17 19:11:12,829][starVLA.training.train_starvla][INFO] - Step 1547, Loss: {'train/loss': 0.5815209150314331, 'train/grad_norm_pre_clip': 11.168227195739746, 'timing/data': 0.01929518859833479, 'timing/model': 0.9526659715920687, 'learning_rate/qwen_vl_interface': 9.005885808959301e-06, 'learning_rate/action_model': 9.005885808959302e-05, 'epoch': 0.7}) +[2026-06-17 19:11:16,745][starVLA.training.train_starvla][INFO] - Step 1548, Loss: {'train/loss': 0.7497962713241577, 'train/grad_norm_pre_clip': 11.83431339263916, 'timing/data': 0.015746857970952988, 'timing/model': 0.9408294623717666, 'learning_rate/qwen_vl_interface': 9.00456144803423e-06, 'learning_rate/action_model': 9.00456144803423e-05, 'epoch': 0.7}) +[2026-06-17 19:11:20,794][starVLA.training.train_starvla][INFO] - Step 1549, Loss: {'train/loss': 0.5226316452026367, 'train/grad_norm_pre_clip': 11.120975494384766, 'timing/data': 0.022066302597522736, 'timing/model': 1.0336373504251242, 'learning_rate/qwen_vl_interface': 9.00323630878453e-06, 'learning_rate/action_model': 9.00323630878453e-05, 'epoch': 0.7}) +[2026-06-17 19:11:24,636][starVLA.training.train_starvla][INFO] - Step 1550, Loss: {'train/loss': 0.6937568783760071, 'train/grad_norm_pre_clip': 11.37796688079834, 'timing/data': 0.02002300973981619, 'timing/model': 0.9423486813902855, 'learning_rate/qwen_vl_interface': 9.001910391484903e-06, 'learning_rate/action_model': 9.001910391484904e-05, 'epoch': 0.7}) +[2026-06-17 19:11:28,536][starVLA.training.train_starvla][INFO] - Step 1551, Loss: {'train/loss': 0.4371257722377777, 'train/grad_norm_pre_clip': 13.667546272277832, 'timing/data': 0.026212538592517376, 'timing/model': 1.1555641712620854, 'learning_rate/qwen_vl_interface': 9.000583696410216e-06, 'learning_rate/action_model': 9.000583696410217e-05, 'epoch': 0.7}) +[2026-06-17 19:11:32,563][starVLA.training.train_starvla][INFO] - Step 1552, Loss: {'train/loss': 0.4062013626098633, 'train/grad_norm_pre_clip': 8.84801959991455, 'timing/data': 0.015615648590028286, 'timing/model': 1.0665778256952763, 'learning_rate/qwen_vl_interface': 8.999256223835493e-06, 'learning_rate/action_model': 8.999256223835494e-05, 'epoch': 0.7}) +[2026-06-17 19:11:36,244][starVLA.training.train_starvla][INFO] - Step 1553, Loss: {'train/loss': 0.5192159414291382, 'train/grad_norm_pre_clip': 11.811066627502441, 'timing/data': 0.013852923177182674, 'timing/model': 0.9350594123825431, 'learning_rate/qwen_vl_interface': 8.997927974035922e-06, 'learning_rate/action_model': 8.997927974035921e-05, 'epoch': 0.7}) +[2026-06-17 19:11:40,136][starVLA.training.train_starvla][INFO] - Step 1554, Loss: {'train/loss': 0.4934268891811371, 'train/grad_norm_pre_clip': 9.296794891357422, 'timing/data': 0.017463031224906445, 'timing/model': 0.9338776497170329, 'learning_rate/qwen_vl_interface': 8.996598947286846e-06, 'learning_rate/action_model': 8.996598947286846e-05, 'epoch': 0.71}) +[2026-06-17 19:11:44,081][starVLA.training.train_starvla][INFO] - Step 1555, Loss: {'train/loss': 0.4803770184516907, 'train/grad_norm_pre_clip': 9.246665954589844, 'timing/data': 0.019188889302313328, 'timing/model': 0.9428777610883117, 'learning_rate/qwen_vl_interface': 8.99526914386378e-06, 'learning_rate/action_model': 8.995269143863779e-05, 'epoch': 0.71}) +[2026-06-17 19:11:48,265][starVLA.training.train_starvla][INFO] - Step 1556, Loss: {'train/loss': 0.5242716073989868, 'train/grad_norm_pre_clip': 8.835833549499512, 'timing/data': 0.01639892626553774, 'timing/model': 1.2386884177103639, 'learning_rate/qwen_vl_interface': 8.993938564042386e-06, 'learning_rate/action_model': 8.993938564042387e-05, 'epoch': 0.71}) +[2026-06-17 19:11:52,293][starVLA.training.train_starvla][INFO] - Step 1557, Loss: {'train/loss': 0.6744031310081482, 'train/grad_norm_pre_clip': 10.418194770812988, 'timing/data': 0.020247981883585453, 'timing/model': 1.2231334866955876, 'learning_rate/qwen_vl_interface': 8.992607208098502e-06, 'learning_rate/action_model': 8.992607208098503e-05, 'epoch': 0.71}) +[2026-06-17 19:11:56,195][starVLA.training.train_starvla][INFO] - Step 1558, Loss: {'train/loss': 0.7478982210159302, 'train/grad_norm_pre_clip': 17.638450622558594, 'timing/data': 0.019952974282205105, 'timing/model': 1.1676423558965325, 'learning_rate/qwen_vl_interface': 8.991275076308117e-06, 'learning_rate/action_model': 8.991275076308117e-05, 'epoch': 0.71}) +[2026-06-17 19:11:59,938][starVLA.training.train_starvla][INFO] - Step 1559, Loss: {'train/loss': 0.6288660764694214, 'train/grad_norm_pre_clip': 14.155074119567871, 'timing/data': 0.01664679776877165, 'timing/model': 0.9362110579386353, 'learning_rate/qwen_vl_interface': 8.989942168947382e-06, 'learning_rate/action_model': 8.989942168947381e-05, 'epoch': 0.71}) +[2026-06-17 19:12:03,991][starVLA.training.train_starvla][INFO] - Step 1560, Loss: {'train/loss': 0.6428390741348267, 'train/grad_norm_pre_clip': 32.168365478515625, 'timing/data': 0.02529796212911606, 'timing/model': 1.0218385672196746, 'learning_rate/qwen_vl_interface': 8.98860848629261e-06, 'learning_rate/action_model': 8.988608486292611e-05, 'epoch': 0.71}) +[2026-06-17 19:12:07,951][starVLA.training.train_starvla][INFO] - Step 1561, Loss: {'train/loss': 0.48764514923095703, 'train/grad_norm_pre_clip': 10.02969741821289, 'timing/data': 0.016940265893936157, 'timing/model': 0.9824103098362684, 'learning_rate/qwen_vl_interface': 8.987274028620276e-06, 'learning_rate/action_model': 8.987274028620276e-05, 'epoch': 0.71}) +[2026-06-17 19:12:11,854][starVLA.training.train_starvla][INFO] - Step 1562, Loss: {'train/loss': 0.44860321283340454, 'train/grad_norm_pre_clip': 16.243818283081055, 'timing/data': 0.017419714480638504, 'timing/model': 0.9527813317254186, 'learning_rate/qwen_vl_interface': 8.985938796207017e-06, 'learning_rate/action_model': 8.985938796207017e-05, 'epoch': 0.71}) +[2026-06-17 19:12:15,781][starVLA.training.train_starvla][INFO] - Step 1563, Loss: {'train/loss': 0.5505526065826416, 'train/grad_norm_pre_clip': 10.935898780822754, 'timing/data': 0.016041005961596966, 'timing/model': 0.9479470048099756, 'learning_rate/qwen_vl_interface': 8.984602789329623e-06, 'learning_rate/action_model': 8.984602789329623e-05, 'epoch': 0.71}) +[2026-06-17 19:12:19,912][starVLA.training.train_starvla][INFO] - Step 1564, Loss: {'train/loss': 0.42255663871765137, 'train/grad_norm_pre_clip': 9.048784255981445, 'timing/data': 0.021192071959376335, 'timing/model': 1.2142655085772276, 'learning_rate/qwen_vl_interface': 8.983266008265052e-06, 'learning_rate/action_model': 8.983266008265052e-05, 'epoch': 0.71}) +[2026-06-17 19:12:23,939][starVLA.training.train_starvla][INFO] - Step 1565, Loss: {'train/loss': 0.6311625838279724, 'train/grad_norm_pre_clip': 13.326197624206543, 'timing/data': 0.019967232830822468, 'timing/model': 0.9563776776194572, 'learning_rate/qwen_vl_interface': 8.981928453290421e-06, 'learning_rate/action_model': 8.981928453290421e-05, 'epoch': 0.71}) +[2026-06-17 19:12:27,929][starVLA.training.train_starvla][INFO] - Step 1566, Loss: {'train/loss': 0.40943723917007446, 'train/grad_norm_pre_clip': 9.242264747619629, 'timing/data': 0.014920697547495365, 'timing/model': 0.9643785571679473, 'learning_rate/qwen_vl_interface': 8.980590124683007e-06, 'learning_rate/action_model': 8.980590124683007e-05, 'epoch': 0.71}) +[2026-06-17 19:12:32,011][starVLA.training.train_starvla][INFO] - Step 1567, Loss: {'train/loss': 0.5279263257980347, 'train/grad_norm_pre_clip': 11.897163391113281, 'timing/data': 0.018226878717541695, 'timing/model': 0.9878419497981668, 'learning_rate/qwen_vl_interface': 8.979251022720244e-06, 'learning_rate/action_model': 8.979251022720244e-05, 'epoch': 0.71}) +[2026-06-17 19:12:36,019][starVLA.training.train_starvla][INFO] - Step 1568, Loss: {'train/loss': 0.418746680021286, 'train/grad_norm_pre_clip': 7.710789680480957, 'timing/data': 0.019896412268280983, 'timing/model': 1.0795310446992517, 'learning_rate/qwen_vl_interface': 8.977911147679733e-06, 'learning_rate/action_model': 8.977911147679733e-05, 'epoch': 0.71}) +[2026-06-17 19:12:39,903][starVLA.training.train_starvla][INFO] - Step 1569, Loss: {'train/loss': 0.5003363490104675, 'train/grad_norm_pre_clip': 11.7823486328125, 'timing/data': 0.018409239128232002, 'timing/model': 0.9600970633327961, 'learning_rate/qwen_vl_interface': 8.976570499839229e-06, 'learning_rate/action_model': 8.976570499839228e-05, 'epoch': 0.71}) +[2026-06-17 19:12:43,820][starVLA.training.train_starvla][INFO] - Step 1570, Loss: {'train/loss': 0.6410119533538818, 'train/grad_norm_pre_clip': 13.563565254211426, 'timing/data': 0.0187088493257761, 'timing/model': 1.147286900319159, 'learning_rate/qwen_vl_interface': 8.97522907947665e-06, 'learning_rate/action_model': 8.975229079476649e-05, 'epoch': 0.71}) +[2026-06-17 19:12:47,650][starVLA.training.train_starvla][INFO] - Step 1571, Loss: {'train/loss': 0.6762887239456177, 'train/grad_norm_pre_clip': 14.508910179138184, 'timing/data': 0.01671392098069191, 'timing/model': 0.9468090692535043, 'learning_rate/qwen_vl_interface': 8.973886886870075e-06, 'learning_rate/action_model': 8.973886886870074e-05, 'epoch': 0.71}) +[2026-06-17 19:12:51,602][starVLA.training.train_starvla][INFO] - Step 1572, Loss: {'train/loss': 0.5335994362831116, 'train/grad_norm_pre_clip': 12.615301132202148, 'timing/data': 0.015421431511640549, 'timing/model': 0.9468189515173435, 'learning_rate/qwen_vl_interface': 8.972543922297739e-06, 'learning_rate/action_model': 8.972543922297739e-05, 'epoch': 0.71}) +[2026-06-17 19:12:55,403][starVLA.training.train_starvla][INFO] - Step 1573, Loss: {'train/loss': 0.592805802822113, 'train/grad_norm_pre_clip': 10.552142143249512, 'timing/data': 0.017669294960796833, 'timing/model': 0.9481223467737436, 'learning_rate/qwen_vl_interface': 8.971200186038046e-06, 'learning_rate/action_model': 8.971200186038045e-05, 'epoch': 0.71}) +[2026-06-17 19:12:59,300][starVLA.training.train_starvla][INFO] - Step 1574, Loss: {'train/loss': 0.3197122812271118, 'train/grad_norm_pre_clip': 13.229927062988281, 'timing/data': 0.020009778439998627, 'timing/model': 0.9350281693041325, 'learning_rate/qwen_vl_interface': 8.969855678369547e-06, 'learning_rate/action_model': 8.969855678369546e-05, 'epoch': 0.71}) +[2026-06-17 19:13:04,194][starVLA.training.train_starvla][INFO] - Step 1575, Loss: {'train/loss': 0.6378356218338013, 'train/grad_norm_pre_clip': 12.318028450012207, 'timing/data': 0.01672327145934105, 'timing/model': 0.942487389780581, 'learning_rate/qwen_vl_interface': 8.968510399570964e-06, 'learning_rate/action_model': 8.968510399570965e-05, 'epoch': 0.71}) +[2026-06-17 19:13:08,155][starVLA.training.train_starvla][INFO] - Step 1576, Loss: {'train/loss': 0.6746498942375183, 'train/grad_norm_pre_clip': 12.206473350524902, 'timing/data': 0.01515465322881937, 'timing/model': 0.9505748599767685, 'learning_rate/qwen_vl_interface': 8.967164349921173e-06, 'learning_rate/action_model': 8.967164349921174e-05, 'epoch': 0.71}) +[2026-06-17 19:13:12,157][starVLA.training.train_starvla][INFO] - Step 1577, Loss: {'train/loss': 0.6058599948883057, 'train/grad_norm_pre_clip': 14.691169738769531, 'timing/data': 0.02045970782637596, 'timing/model': 1.1913824584335089, 'learning_rate/qwen_vl_interface': 8.965817529699216e-06, 'learning_rate/action_model': 8.965817529699215e-05, 'epoch': 0.72}) +[2026-06-17 19:13:15,929][starVLA.training.train_starvla][INFO] - Step 1578, Loss: {'train/loss': 0.7549248933792114, 'train/grad_norm_pre_clip': 12.079384803771973, 'timing/data': 0.016892614774405956, 'timing/model': 0.9479554984718561, 'learning_rate/qwen_vl_interface': 8.964469939184283e-06, 'learning_rate/action_model': 8.964469939184282e-05, 'epoch': 0.72}) +[2026-06-17 19:13:19,859][starVLA.training.train_starvla][INFO] - Step 1579, Loss: {'train/loss': 0.567970335483551, 'train/grad_norm_pre_clip': 10.5096435546875, 'timing/data': 0.019224144518375397, 'timing/model': 0.9333437066525221, 'learning_rate/qwen_vl_interface': 8.963121578655736e-06, 'learning_rate/action_model': 8.963121578655736e-05, 'epoch': 0.72}) +[2026-06-17 19:13:23,765][starVLA.training.train_starvla][INFO] - Step 1580, Loss: {'train/loss': 0.862199604511261, 'train/grad_norm_pre_clip': 15.848336219787598, 'timing/data': 0.015242663212120533, 'timing/model': 0.9374100863933563, 'learning_rate/qwen_vl_interface': 8.96177244839309e-06, 'learning_rate/action_model': 8.961772448393092e-05, 'epoch': 0.72}) +[2026-06-17 19:13:27,889][starVLA.training.train_starvla][INFO] - Step 1581, Loss: {'train/loss': 0.655072033405304, 'train/grad_norm_pre_clip': 24.06252098083496, 'timing/data': 0.01758211199194193, 'timing/model': 0.9473317814990878, 'learning_rate/qwen_vl_interface': 8.960422548676024e-06, 'learning_rate/action_model': 8.960422548676022e-05, 'epoch': 0.72}) +[2026-06-17 19:13:32,227][starVLA.training.train_starvla][INFO] - Step 1582, Loss: {'train/loss': 0.6500668525695801, 'train/grad_norm_pre_clip': 11.267401695251465, 'timing/data': 0.03315509855747223, 'timing/model': 1.2979579186066985, 'learning_rate/qwen_vl_interface': 8.959071879784368e-06, 'learning_rate/action_model': 8.959071879784368e-05, 'epoch': 0.72}) +[2026-06-17 19:13:36,219][starVLA.training.train_starvla][INFO] - Step 1583, Loss: {'train/loss': 0.604292094707489, 'train/grad_norm_pre_clip': 10.608092308044434, 'timing/data': 0.02777300402522087, 'timing/model': 1.1688033286482096, 'learning_rate/qwen_vl_interface': 8.95772044199812e-06, 'learning_rate/action_model': 8.95772044199812e-05, 'epoch': 0.72}) +[2026-06-17 19:13:40,356][starVLA.training.train_starvla][INFO] - Step 1584, Loss: {'train/loss': 0.6013665199279785, 'train/grad_norm_pre_clip': 17.340150833129883, 'timing/data': 0.030904093757271767, 'timing/model': 1.3789131958037615, 'learning_rate/qwen_vl_interface': 8.956368235597437e-06, 'learning_rate/action_model': 8.956368235597436e-05, 'epoch': 0.72}) +[2026-06-17 19:13:43,999][starVLA.training.train_starvla][INFO] - Step 1585, Loss: {'train/loss': 0.4769745469093323, 'train/grad_norm_pre_clip': 10.226858139038086, 'timing/data': 0.016228530555963516, 'timing/model': 0.9362514866515994, 'learning_rate/qwen_vl_interface': 8.955015260862627e-06, 'learning_rate/action_model': 8.955015260862627e-05, 'epoch': 0.72}) +[2026-06-17 19:13:47,850][starVLA.training.train_starvla][INFO] - Step 1586, Loss: {'train/loss': 0.5057962536811829, 'train/grad_norm_pre_clip': 11.978236198425293, 'timing/data': 0.014455738477408886, 'timing/model': 0.9485875507816672, 'learning_rate/qwen_vl_interface': 8.95366151807417e-06, 'learning_rate/action_model': 8.95366151807417e-05, 'epoch': 0.72}) +[2026-06-17 19:13:53,318][starVLA.training.train_starvla][INFO] - Step 1587, Loss: {'train/loss': 0.8507885932922363, 'train/grad_norm_pre_clip': 15.275687217712402, 'timing/data': 0.022565895691514015, 'timing/model': 2.2915697740390897, 'learning_rate/qwen_vl_interface': 8.952307007512694e-06, 'learning_rate/action_model': 8.952307007512694e-05, 'epoch': 0.72}) +[2026-06-17 19:13:57,322][starVLA.training.train_starvla][INFO] - Step 1588, Loss: {'train/loss': 0.36201339960098267, 'train/grad_norm_pre_clip': 7.299171447753906, 'timing/data': 0.017956005409359932, 'timing/model': 0.9457335248589516, 'learning_rate/qwen_vl_interface': 8.950951729458991e-06, 'learning_rate/action_model': 8.950951729458991e-05, 'epoch': 0.72}) +[2026-06-17 19:14:01,392][starVLA.training.train_starvla][INFO] - Step 1589, Loss: {'train/loss': 0.6039499044418335, 'train/grad_norm_pre_clip': 11.41566276550293, 'timing/data': 0.026625079102814198, 'timing/model': 0.9491363624110818, 'learning_rate/qwen_vl_interface': 8.949595684194012e-06, 'learning_rate/action_model': 8.949595684194012e-05, 'epoch': 0.72}) +[2026-06-17 19:14:05,413][starVLA.training.train_starvla][INFO] - Step 1590, Loss: {'train/loss': 0.46605542302131653, 'train/grad_norm_pre_clip': 8.43032455444336, 'timing/data': 0.016477730125188828, 'timing/model': 1.1653237203136086, 'learning_rate/qwen_vl_interface': 8.948238871998866e-06, 'learning_rate/action_model': 8.948238871998866e-05, 'epoch': 0.72}) +[2026-06-17 19:14:09,340][starVLA.training.train_starvla][INFO] - Step 1591, Loss: {'train/loss': 0.4253644347190857, 'train/grad_norm_pre_clip': 9.270458221435547, 'timing/data': 0.015902435407042503, 'timing/model': 1.1804156145080924, 'learning_rate/qwen_vl_interface': 8.946881293154825e-06, 'learning_rate/action_model': 8.946881293154825e-05, 'epoch': 0.72}) +[2026-06-17 19:14:13,183][starVLA.training.train_starvla][INFO] - Step 1592, Loss: {'train/loss': 0.5939371585845947, 'train/grad_norm_pre_clip': 12.284524917602539, 'timing/data': 0.016082558780908585, 'timing/model': 0.9563787970691919, 'learning_rate/qwen_vl_interface': 8.945522947943312e-06, 'learning_rate/action_model': 8.945522947943313e-05, 'epoch': 0.72}) +[2026-06-17 19:14:17,227][starVLA.training.train_starvla][INFO] - Step 1593, Loss: {'train/loss': 0.4022333025932312, 'train/grad_norm_pre_clip': 9.832914352416992, 'timing/data': 0.020543985068798065, 'timing/model': 0.9660015273839235, 'learning_rate/qwen_vl_interface': 8.944163836645916e-06, 'learning_rate/action_model': 8.944163836645915e-05, 'epoch': 0.72}) +[2026-06-17 19:14:21,263][starVLA.training.train_starvla][INFO] - Step 1594, Loss: {'train/loss': 0.6114341020584106, 'train/grad_norm_pre_clip': 14.474481582641602, 'timing/data': 0.015731774270534515, 'timing/model': 0.9518491672351956, 'learning_rate/qwen_vl_interface': 8.942803959544381e-06, 'learning_rate/action_model': 8.942803959544381e-05, 'epoch': 0.72}) +[2026-06-17 19:14:25,279][starVLA.training.train_starvla][INFO] - Step 1595, Loss: {'train/loss': 0.7247986197471619, 'train/grad_norm_pre_clip': 12.343317031860352, 'timing/data': 0.024746038019657135, 'timing/model': 1.000747874379158, 'learning_rate/qwen_vl_interface': 8.941443316920613e-06, 'learning_rate/action_model': 8.941443316920612e-05, 'epoch': 0.72}) +[2026-06-17 19:14:29,163][starVLA.training.train_starvla][INFO] - Step 1596, Loss: {'train/loss': 0.4864855408668518, 'train/grad_norm_pre_clip': 12.187376976013184, 'timing/data': 0.014119110070168972, 'timing/model': 0.950596496462822, 'learning_rate/qwen_vl_interface': 8.940081909056671e-06, 'learning_rate/action_model': 8.940081909056671e-05, 'epoch': 0.72}) +[2026-06-17 19:14:33,198][starVLA.training.train_starvla][INFO] - Step 1597, Loss: {'train/loss': 0.4921150803565979, 'train/grad_norm_pre_clip': 10.279955863952637, 'timing/data': 0.019434507004916668, 'timing/model': 1.173991491086781, 'learning_rate/qwen_vl_interface': 8.93871973623478e-06, 'learning_rate/action_model': 8.93871973623478e-05, 'epoch': 0.72}) +[2026-06-17 19:14:36,900][starVLA.training.train_starvla][INFO] - Step 1598, Loss: {'train/loss': 0.31066253781318665, 'train/grad_norm_pre_clip': 9.132173538208008, 'timing/data': 0.01744637731462717, 'timing/model': 0.9860278386622667, 'learning_rate/qwen_vl_interface': 8.937356798737317e-06, 'learning_rate/action_model': 8.937356798737317e-05, 'epoch': 0.72}) +[2026-06-17 19:14:40,795][starVLA.training.train_starvla][INFO] - Step 1599, Loss: {'train/loss': 0.6312627196311951, 'train/grad_norm_pre_clip': 12.856993675231934, 'timing/data': 0.015801007859408855, 'timing/model': 0.9437816012650728, 'learning_rate/qwen_vl_interface': 8.935993096846825e-06, 'learning_rate/action_model': 8.935993096846825e-05, 'epoch': 0.73}) +[2026-06-17 19:17:29,494][starVLA.training.train_starvla][INFO] - Step 1600, Loss: {'train/loss': 0.6806759238243103, 'train/grad_norm_pre_clip': 12.347911834716797, 'eval/loss': 0.5037348866462708, 'eval/latency_8/loss': 0.5037348866462708, 'eval/demon_fire_precision': 0.921090803597331, 'eval/demon_fire_recall': 0.9863311587449518, 'eval/demon_fire_f1': 0.9525952595259525, 'eval/demon_fire_rate_model': 0.9590984974958264, 'eval/demon_fire_rate_teacher': 0.8956594323873122, 'eval/demon_left_precision': 0.7593123209169055, 'eval/demon_left_recall': 0.5674518201284796, 'eval/demon_left_f1': 0.6495098039215687, 'eval/demon_left_rate_model': 0.09710628825820812, 'eval/demon_left_rate_teacher': 0.12993878686700056, 'eval/demon_right_precision': 0.7537190082644628, 'eval/demon_right_recall': 0.6909090909090909, 'eval/demon_right_f1': 0.7209486166007905, 'eval/demon_right_rate_model': 0.16833611574846968, 'eval/demon_right_rate_teacher': 0.18363939899833054, 'eval/demon_move_f1': 0.6852292102611796, 'eval/demon_cc_f1': 0.8527734751293606, 'eval/latency_8/demon_fire_precision': 0.921090803597331, 'eval/latency_8/demon_fire_recall': 0.9863311587449518, 'eval/latency_8/demon_fire_f1': 0.9525952595259525, 'eval/latency_8/demon_fire_rate_model': 0.9590984974958264, 'eval/latency_8/demon_fire_rate_teacher': 0.8956594323873122, 'eval/latency_8/demon_left_precision': 0.7593123209169055, 'eval/latency_8/demon_left_recall': 0.5674518201284796, 'eval/latency_8/demon_left_f1': 0.6495098039215687, 'eval/latency_8/demon_left_rate_model': 0.09710628825820812, 'eval/latency_8/demon_left_rate_teacher': 0.12993878686700056, 'eval/latency_8/demon_right_precision': 0.7537190082644628, 'eval/latency_8/demon_right_recall': 0.6909090909090909, 'eval/latency_8/demon_right_f1': 0.7209486166007905, 'eval/latency_8/demon_right_rate_model': 0.16833611574846968, 'eval/latency_8/demon_right_rate_teacher': 0.18363939899833054, 'eval/latency_8/demon_move_f1': 0.6852292102611796, 'eval/latency_8/demon_cc_f1': 0.8527734751293606, 'timing/data': 0.03159778565168381, 'timing/model': 0.9814248494803905, 'learning_rate/qwen_vl_interface': 8.934628630845994e-06, 'learning_rate/action_model': 8.934628630845995e-05, 'epoch': 0.73}) +[2026-06-17 19:17:29,495][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_1600_state +[2026-06-17 19:17:29,501][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 19:17:36,916][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_1600_state/model.safetensors +[2026-06-17 19:17:55,289][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_1600_state/optimizer.bin +[2026-06-17 19:17:55,290][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_1600_state/sampler.bin +[2026-06-17 19:17:55,290][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_1600_state/sampler_1.bin +[2026-06-17 19:17:55,353][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_1600_state/random_states_0.pkl +[2026-06-17 19:17:55,358][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1600 +[2026-06-17 19:17:55,358][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 19:17:55,369][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 19:17:59,657][starVLA.training.train_starvla][INFO] - Step 1601, Loss: {'train/loss': 0.8515563607215881, 'train/grad_norm_pre_clip': 10.166105270385742, 'timing/data': 0.026346057653427124, 'timing/model': 0.9754515886306763, 'learning_rate/qwen_vl_interface': 8.933263401017685e-06, 'learning_rate/action_model': 8.933263401017684e-05, 'epoch': 0.73}) +[2026-06-17 19:18:03,638][starVLA.training.train_starvla][INFO] - Step 1602, Loss: {'train/loss': 0.31924504041671753, 'train/grad_norm_pre_clip': 16.974138259887695, 'timing/data': 0.019227655604481697, 'timing/model': 0.9685374815016985, 'learning_rate/qwen_vl_interface': 8.93189740764491e-06, 'learning_rate/action_model': 8.931897407644909e-05, 'epoch': 0.73}) +[2026-06-17 19:18:07,577][starVLA.training.train_starvla][INFO] - Step 1603, Loss: {'train/loss': 0.746873140335083, 'train/grad_norm_pre_clip': 18.85123634338379, 'timing/data': 0.018562478944659233, 'timing/model': 0.9602388916537166, 'learning_rate/qwen_vl_interface': 8.93053065101084e-06, 'learning_rate/action_model': 8.93053065101084e-05, 'epoch': 0.73}) +[2026-06-17 19:18:13,042][starVLA.training.train_starvla][INFO] - Step 1604, Loss: {'train/loss': 0.33509886264801025, 'train/grad_norm_pre_clip': 8.565834999084473, 'timing/data': 0.018278883770108223, 'timing/model': 1.0236886292696, 'learning_rate/qwen_vl_interface': 8.929163131398804e-06, 'learning_rate/action_model': 8.929163131398805e-05, 'epoch': 0.73}) +[2026-06-17 19:18:16,974][starVLA.training.train_starvla][INFO] - Step 1605, Loss: {'train/loss': 0.37830764055252075, 'train/grad_norm_pre_clip': 9.074661254882812, 'timing/data': 0.020048484206199646, 'timing/model': 1.183132209815085, 'learning_rate/qwen_vl_interface': 8.927794849092294e-06, 'learning_rate/action_model': 8.927794849092293e-05, 'epoch': 0.73}) +[2026-06-17 19:18:21,192][starVLA.training.train_starvla][INFO] - Step 1606, Loss: {'train/loss': 1.054720401763916, 'train/grad_norm_pre_clip': 16.54907989501953, 'timing/data': 0.020747259259223938, 'timing/model': 1.0095229046419263, 'learning_rate/qwen_vl_interface': 8.926425804374953e-06, 'learning_rate/action_model': 8.926425804374954e-05, 'epoch': 0.73}) +[2026-06-17 19:18:25,300][starVLA.training.train_starvla][INFO] - Step 1607, Loss: {'train/loss': 0.5798258781433105, 'train/grad_norm_pre_clip': 11.157544136047363, 'timing/data': 0.02835104800760746, 'timing/model': 0.9872621204704046, 'learning_rate/qwen_vl_interface': 8.925055997530587e-06, 'learning_rate/action_model': 8.925055997530587e-05, 'epoch': 0.73}) +[2026-06-17 19:18:29,300][starVLA.training.train_starvla][INFO] - Step 1608, Loss: {'train/loss': 0.797475278377533, 'train/grad_norm_pre_clip': 13.8285551071167, 'timing/data': 0.016067538410425186, 'timing/model': 1.031744183972478, 'learning_rate/qwen_vl_interface': 8.92368542884316e-06, 'learning_rate/action_model': 8.923685428843159e-05, 'epoch': 0.73}) +[2026-06-17 19:18:33,990][starVLA.training.train_starvla][INFO] - Step 1609, Loss: {'train/loss': 0.5179608464241028, 'train/grad_norm_pre_clip': 11.540178298950195, 'timing/data': 0.018377872183918953, 'timing/model': 1.0293862773105502, 'learning_rate/qwen_vl_interface': 8.922314098596789e-06, 'learning_rate/action_model': 8.922314098596789e-05, 'epoch': 0.73}) +[2026-06-17 19:18:38,137][starVLA.training.train_starvla][INFO] - Step 1610, Loss: {'train/loss': 0.8328882455825806, 'train/grad_norm_pre_clip': 14.112509727478027, 'timing/data': 0.02266879193484783, 'timing/model': 1.0144906789064407, 'learning_rate/qwen_vl_interface': 8.920942007075752e-06, 'learning_rate/action_model': 8.920942007075752e-05, 'epoch': 0.73}) +[2026-06-17 19:18:42,506][starVLA.training.train_starvla][INFO] - Step 1611, Loss: {'train/loss': 0.6556991934776306, 'train/grad_norm_pre_clip': 16.349578857421875, 'timing/data': 0.026679988019168377, 'timing/model': 0.942319636233151, 'learning_rate/qwen_vl_interface': 8.91956915456449e-06, 'learning_rate/action_model': 8.91956915456449e-05, 'epoch': 0.73}) +[2026-06-17 19:18:46,641][starVLA.training.train_starvla][INFO] - Step 1612, Loss: {'train/loss': 0.4328177273273468, 'train/grad_norm_pre_clip': 10.309394836425781, 'timing/data': 0.01845618337392807, 'timing/model': 1.3852765001356602, 'learning_rate/qwen_vl_interface': 8.918195541347593e-06, 'learning_rate/action_model': 8.918195541347593e-05, 'epoch': 0.73}) +[2026-06-17 19:18:50,346][starVLA.training.train_starvla][INFO] - Step 1613, Loss: {'train/loss': 0.6413378119468689, 'train/grad_norm_pre_clip': 14.295117378234863, 'timing/data': 0.016112130135297775, 'timing/model': 0.9407743252813816, 'learning_rate/qwen_vl_interface': 8.916821167709811e-06, 'learning_rate/action_model': 8.916821167709812e-05, 'epoch': 0.73}) +[2026-06-17 19:18:55,125][starVLA.training.train_starvla][INFO] - Step 1614, Loss: {'train/loss': 0.6713201403617859, 'train/grad_norm_pre_clip': 10.552128791809082, 'timing/data': 0.010605629533529282, 'timing/model': 0.9793895268812776, 'learning_rate/qwen_vl_interface': 8.915446033936057e-06, 'learning_rate/action_model': 8.915446033936056e-05, 'epoch': 0.73}) +[2026-06-17 19:18:59,324][starVLA.training.train_starvla][INFO] - Step 1615, Loss: {'train/loss': 0.6771613359451294, 'train/grad_norm_pre_clip': 11.211377143859863, 'timing/data': 0.01019325666129589, 'timing/model': 1.0605869283899665, 'learning_rate/qwen_vl_interface': 8.914070140311395e-06, 'learning_rate/action_model': 8.914070140311395e-05, 'epoch': 0.73}) +[2026-06-17 19:19:05,249][starVLA.training.train_starvla][INFO] - Step 1616, Loss: {'train/loss': 0.5870270133018494, 'train/grad_norm_pre_clip': 13.17249584197998, 'timing/data': 0.013324325904250145, 'timing/model': 2.5446675261482596, 'learning_rate/qwen_vl_interface': 8.912693487121053e-06, 'learning_rate/action_model': 8.912693487121052e-05, 'epoch': 0.73}) +[2026-06-17 19:19:09,504][starVLA.training.train_starvla][INFO] - Step 1617, Loss: {'train/loss': 0.5324090719223022, 'train/grad_norm_pre_clip': 11.101066589355469, 'timing/data': 0.008971565403044224, 'timing/model': 0.9644876671954989, 'learning_rate/qwen_vl_interface': 8.911316074650407e-06, 'learning_rate/action_model': 8.911316074650406e-05, 'epoch': 0.73}) +[2026-06-17 19:19:13,882][starVLA.training.train_starvla][INFO] - Step 1618, Loss: {'train/loss': 0.5498351454734802, 'train/grad_norm_pre_clip': 12.03203296661377, 'timing/data': 0.011590555310249329, 'timing/model': 1.4296139748767018, 'learning_rate/qwen_vl_interface': 8.909937903185e-06, 'learning_rate/action_model': 8.909937903185e-05, 'epoch': 0.73}) +[2026-06-17 19:19:17,960][starVLA.training.train_starvla][INFO] - Step 1619, Loss: {'train/loss': 0.5228874087333679, 'train/grad_norm_pre_clip': 11.018058776855469, 'timing/data': 0.010903933085501194, 'timing/model': 1.3004915099591017, 'learning_rate/qwen_vl_interface': 8.908558973010528e-06, 'learning_rate/action_model': 8.908558973010526e-05, 'epoch': 0.73}) +[2026-06-17 19:19:21,792][starVLA.training.train_starvla][INFO] - Step 1620, Loss: {'train/loss': 0.44759929180145264, 'train/grad_norm_pre_clip': 9.18529224395752, 'timing/data': 0.009904189966619015, 'timing/model': 0.9657731121405959, 'learning_rate/qwen_vl_interface': 8.907179284412841e-06, 'learning_rate/action_model': 8.90717928441284e-05, 'epoch': 0.73}) +[2026-06-17 19:19:25,957][starVLA.training.train_starvla][INFO] - Step 1621, Loss: {'train/loss': 0.47521376609802246, 'train/grad_norm_pre_clip': 9.92251968383789, 'timing/data': 0.017105973325669765, 'timing/model': 0.9479929320514202, 'learning_rate/qwen_vl_interface': 8.905798837677956e-06, 'learning_rate/action_model': 8.905798837677956e-05, 'epoch': 0.74}) +[2026-06-17 19:19:30,205][starVLA.training.train_starvla][INFO] - Step 1622, Loss: {'train/loss': 0.33932867646217346, 'train/grad_norm_pre_clip': 12.142393112182617, 'timing/data': 0.008409741334617138, 'timing/model': 0.9494841955602169, 'learning_rate/qwen_vl_interface': 8.904417633092037e-06, 'learning_rate/action_model': 8.904417633092037e-05, 'epoch': 0.74}) +[2026-06-17 19:19:34,241][starVLA.training.train_starvla][INFO] - Step 1623, Loss: {'train/loss': 0.4570288062095642, 'train/grad_norm_pre_clip': 8.46911907196045, 'timing/data': 0.00983275007456541, 'timing/model': 0.9725105725228786, 'learning_rate/qwen_vl_interface': 8.903035670941411e-06, 'learning_rate/action_model': 8.903035670941412e-05, 'epoch': 0.74}) +[2026-06-17 19:19:38,326][starVLA.training.train_starvla][INFO] - Step 1624, Loss: {'train/loss': 0.9278330206871033, 'train/grad_norm_pre_clip': 15.73954963684082, 'timing/data': 0.013190706260502338, 'timing/model': 0.9516984084621072, 'learning_rate/qwen_vl_interface': 8.901652951512559e-06, 'learning_rate/action_model': 8.90165295151256e-05, 'epoch': 0.74}) +[2026-06-17 19:19:42,169][starVLA.training.train_starvla][INFO] - Step 1625, Loss: {'train/loss': 0.3596341609954834, 'train/grad_norm_pre_clip': 12.507951736450195, 'timing/data': 0.00950914341956377, 'timing/model': 1.1680135438218713, 'learning_rate/qwen_vl_interface': 8.900269475092122e-06, 'learning_rate/action_model': 8.900269475092122e-05, 'epoch': 0.74}) +[2026-06-17 19:19:46,169][starVLA.training.train_starvla][INFO] - Step 1626, Loss: {'train/loss': 0.7540789246559143, 'train/grad_norm_pre_clip': 16.669221878051758, 'timing/data': 0.010814775712788105, 'timing/model': 0.9378991201519966, 'learning_rate/qwen_vl_interface': 8.898885241966894e-06, 'learning_rate/action_model': 8.898885241966893e-05, 'epoch': 0.74}) +[2026-06-17 19:19:52,440][starVLA.training.train_starvla][INFO] - Step 1627, Loss: {'train/loss': 0.3308961093425751, 'train/grad_norm_pre_clip': 7.333948135375977, 'timing/data': 0.012016933411359787, 'timing/model': 1.0095793595537543, 'learning_rate/qwen_vl_interface': 8.89750025242383e-06, 'learning_rate/action_model': 8.89750025242383e-05, 'epoch': 0.74}) +[2026-06-17 19:19:57,605][starVLA.training.train_starvla][INFO] - Step 1628, Loss: {'train/loss': 0.4992524981498718, 'train/grad_norm_pre_clip': 10.122760772705078, 'timing/data': 0.007780665531754494, 'timing/model': 0.9293976612389088, 'learning_rate/qwen_vl_interface': 8.896114506750037e-06, 'learning_rate/action_model': 8.896114506750036e-05, 'epoch': 0.74}) +[2026-06-17 19:20:01,366][starVLA.training.train_starvla][INFO] - Step 1629, Loss: {'train/loss': 0.32490652799606323, 'train/grad_norm_pre_clip': 8.488422393798828, 'timing/data': 0.014628083445131779, 'timing/model': 0.9507593689486384, 'learning_rate/qwen_vl_interface': 8.894728005232785e-06, 'learning_rate/action_model': 8.894728005232785e-05, 'epoch': 0.74}) +[2026-06-17 19:20:05,340][starVLA.training.train_starvla][INFO] - Step 1630, Loss: {'train/loss': 0.508447527885437, 'train/grad_norm_pre_clip': 10.64821720123291, 'timing/data': 0.009843884967267513, 'timing/model': 0.9890754977241158, 'learning_rate/qwen_vl_interface': 8.893340748159494e-06, 'learning_rate/action_model': 8.893340748159493e-05, 'epoch': 0.74}) +[2026-06-17 19:20:09,261][starVLA.training.train_starvla][INFO] - Step 1631, Loss: {'train/loss': 0.5801011919975281, 'train/grad_norm_pre_clip': 10.578108787536621, 'timing/data': 0.008155765011906624, 'timing/model': 1.1139612644910812, 'learning_rate/qwen_vl_interface': 8.891952735817746e-06, 'learning_rate/action_model': 8.891952735817745e-05, 'epoch': 0.74}) +[2026-06-17 19:20:13,228][starVLA.training.train_starvla][INFO] - Step 1632, Loss: {'train/loss': 0.9372602701187134, 'train/grad_norm_pre_clip': 14.114424705505371, 'timing/data': 0.00871026050299406, 'timing/model': 1.1598251815885305, 'learning_rate/qwen_vl_interface': 8.890563968495278e-06, 'learning_rate/action_model': 8.890563968495278e-05, 'epoch': 0.74}) +[2026-06-17 19:20:16,886][starVLA.training.train_starvla][INFO] - Step 1633, Loss: {'train/loss': 0.7099791765213013, 'train/grad_norm_pre_clip': 14.593103408813477, 'timing/data': 0.013574003241956234, 'timing/model': 0.9571116100996733, 'learning_rate/qwen_vl_interface': 8.88917444647998e-06, 'learning_rate/action_model': 8.889174446479981e-05, 'epoch': 0.74}) +[2026-06-17 19:20:20,752][starVLA.training.train_starvla][INFO] - Step 1634, Loss: {'train/loss': 0.28925853967666626, 'train/grad_norm_pre_clip': 7.795897006988525, 'timing/data': 0.011501009576022625, 'timing/model': 0.9332131911069155, 'learning_rate/qwen_vl_interface': 8.887784170059905e-06, 'learning_rate/action_model': 8.887784170059905e-05, 'epoch': 0.74}) +[2026-06-17 19:20:24,536][starVLA.training.train_starvla][INFO] - Step 1635, Loss: {'train/loss': 0.23894333839416504, 'train/grad_norm_pre_clip': 8.0914945602417, 'timing/data': 0.010620208457112312, 'timing/model': 0.9302469892427325, 'learning_rate/qwen_vl_interface': 8.886393139523258e-06, 'learning_rate/action_model': 8.886393139523257e-05, 'epoch': 0.74}) +[2026-06-17 19:20:28,388][starVLA.training.train_starvla][INFO] - Step 1636, Loss: {'train/loss': 0.5289191603660583, 'train/grad_norm_pre_clip': 12.01162052154541, 'timing/data': 0.007246080785989761, 'timing/model': 1.0092632416635752, 'learning_rate/qwen_vl_interface': 8.885001355158397e-06, 'learning_rate/action_model': 8.885001355158397e-05, 'epoch': 0.74}) +[2026-06-17 19:20:32,213][starVLA.training.train_starvla][INFO] - Step 1637, Loss: {'train/loss': 0.41057929396629333, 'train/grad_norm_pre_clip': 12.412062644958496, 'timing/data': 0.011092407628893852, 'timing/model': 0.9467425746843219, 'learning_rate/qwen_vl_interface': 8.883608817253847e-06, 'learning_rate/action_model': 8.883608817253847e-05, 'epoch': 0.74}) +[2026-06-17 19:20:36,029][starVLA.training.train_starvla][INFO] - Step 1638, Loss: {'train/loss': 0.44142380356788635, 'train/grad_norm_pre_clip': 11.771855354309082, 'timing/data': 0.009062105789780617, 'timing/model': 1.1122476821765304, 'learning_rate/qwen_vl_interface': 8.882215526098276e-06, 'learning_rate/action_model': 8.882215526098277e-05, 'epoch': 0.74}) +[2026-06-17 19:20:39,658][starVLA.training.train_starvla][INFO] - Step 1639, Loss: {'train/loss': 0.529079020023346, 'train/grad_norm_pre_clip': 11.734702110290527, 'timing/data': 0.0078803189098835, 'timing/model': 0.9435573564842343, 'learning_rate/qwen_vl_interface': 8.880821481980521e-06, 'learning_rate/action_model': 8.880821481980521e-05, 'epoch': 0.74}) +[2026-06-17 19:20:43,498][starVLA.training.train_starvla][INFO] - Step 1640, Loss: {'train/loss': 0.46135008335113525, 'train/grad_norm_pre_clip': 10.36882209777832, 'timing/data': 0.01095590926706791, 'timing/model': 0.9473668439313769, 'learning_rate/qwen_vl_interface': 8.879426685189565e-06, 'learning_rate/action_model': 8.879426685189565e-05, 'epoch': 0.74}) +[2026-06-17 19:20:47,448][starVLA.training.train_starvla][INFO] - Step 1641, Loss: {'train/loss': 0.532548725605011, 'train/grad_norm_pre_clip': 13.322072982788086, 'timing/data': 0.008437059819698334, 'timing/model': 0.9461723538115621, 'learning_rate/qwen_vl_interface': 8.878031136014552e-06, 'learning_rate/action_model': 8.878031136014551e-05, 'epoch': 0.74}) +[2026-06-17 19:20:51,402][starVLA.training.train_starvla][INFO] - Step 1642, Loss: {'train/loss': 0.6402937769889832, 'train/grad_norm_pre_clip': 11.17558765411377, 'timing/data': 0.020326769910752773, 'timing/model': 1.0037106117233634, 'learning_rate/qwen_vl_interface': 8.87663483474478e-06, 'learning_rate/action_model': 8.87663483474478e-05, 'epoch': 0.74}) +[2026-06-17 19:20:55,579][starVLA.training.train_starvla][INFO] - Step 1643, Loss: {'train/loss': 0.46458297967910767, 'train/grad_norm_pre_clip': 13.332399368286133, 'timing/data': 0.012614171020686626, 'timing/model': 1.109747238457203, 'learning_rate/qwen_vl_interface': 8.875237781669704e-06, 'learning_rate/action_model': 8.875237781669703e-05, 'epoch': 0.75}) +[2026-06-17 19:20:59,629][starVLA.training.train_starvla][INFO] - Step 1644, Loss: {'train/loss': 0.5964093804359436, 'train/grad_norm_pre_clip': 15.202569007873535, 'timing/data': 0.010406656190752983, 'timing/model': 0.9712449852377176, 'learning_rate/qwen_vl_interface': 8.873839977078933e-06, 'learning_rate/action_model': 8.873839977078934e-05, 'epoch': 0.75}) +[2026-06-17 19:21:03,589][starVLA.training.train_starvla][INFO] - Step 1645, Loss: {'train/loss': 0.4729786217212677, 'train/grad_norm_pre_clip': 10.913355827331543, 'timing/data': 0.010451399721205235, 'timing/model': 1.1725004315376282, 'learning_rate/qwen_vl_interface': 8.87244142126224e-06, 'learning_rate/action_model': 8.87244142126224e-05, 'epoch': 0.75}) +[2026-06-17 19:21:07,545][starVLA.training.train_starvla][INFO] - Step 1646, Loss: {'train/loss': 0.647286593914032, 'train/grad_norm_pre_clip': 13.54976749420166, 'timing/data': 0.010362032800912857, 'timing/model': 1.0737688019871712, 'learning_rate/qwen_vl_interface': 8.871042114509539e-06, 'learning_rate/action_model': 8.87104211450954e-05, 'epoch': 0.75}) +[2026-06-17 19:21:11,435][starVLA.training.train_starvla][INFO] - Step 1647, Loss: {'train/loss': 0.6686872839927673, 'train/grad_norm_pre_clip': 12.436131477355957, 'timing/data': 0.010170919820666313, 'timing/model': 0.941015200689435, 'learning_rate/qwen_vl_interface': 8.869642057110912e-06, 'learning_rate/action_model': 8.869642057110911e-05, 'epoch': 0.75}) +[2026-06-17 19:21:15,457][starVLA.training.train_starvla][INFO] - Step 1648, Loss: {'train/loss': 0.45816120505332947, 'train/grad_norm_pre_clip': 9.665910720825195, 'timing/data': 0.007990805432200432, 'timing/model': 0.9515653569251299, 'learning_rate/qwen_vl_interface': 8.868241249356592e-06, 'learning_rate/action_model': 8.868241249356592e-05, 'epoch': 0.75}) +[2026-06-17 19:21:19,531][starVLA.training.train_starvla][INFO] - Step 1649, Loss: {'train/loss': 0.580765426158905, 'train/grad_norm_pre_clip': 11.820067405700684, 'timing/data': 0.008478566072881222, 'timing/model': 0.9713569777086377, 'learning_rate/qwen_vl_interface': 8.866839691536968e-06, 'learning_rate/action_model': 8.866839691536968e-05, 'epoch': 0.75}) +[2026-06-17 19:21:23,501][starVLA.training.train_starvla][INFO] - Step 1650, Loss: {'train/loss': 0.3585803210735321, 'train/grad_norm_pre_clip': 10.307296752929688, 'timing/data': 0.007438261993229389, 'timing/model': 1.0166800245642662, 'learning_rate/qwen_vl_interface': 8.865437383942584e-06, 'learning_rate/action_model': 8.865437383942583e-05, 'epoch': 0.75}) +[2026-06-17 19:21:27,462][starVLA.training.train_starvla][INFO] - Step 1651, Loss: {'train/loss': 0.639685332775116, 'train/grad_norm_pre_clip': 12.073705673217773, 'timing/data': 0.008655212819576263, 'timing/model': 1.182703422382474, 'learning_rate/qwen_vl_interface': 8.86403432686414e-06, 'learning_rate/action_model': 8.86403432686414e-05, 'epoch': 0.75}) +[2026-06-17 19:21:31,388][starVLA.training.train_starvla][INFO] - Step 1652, Loss: {'train/loss': 0.4148397743701935, 'train/grad_norm_pre_clip': 10.069751739501953, 'timing/data': 0.01025493536144495, 'timing/model': 1.125471050851047, 'learning_rate/qwen_vl_interface': 8.86263052059249e-06, 'learning_rate/action_model': 8.862630520592489e-05, 'epoch': 0.75}) +[2026-06-17 19:21:36,050][starVLA.training.train_starvla][INFO] - Step 1653, Loss: {'train/loss': 0.6289593577384949, 'train/grad_norm_pre_clip': 12.714227676391602, 'timing/data': 0.007816588506102562, 'timing/model': 0.9444253286346793, 'learning_rate/qwen_vl_interface': 8.861225965418647e-06, 'learning_rate/action_model': 8.861225965418646e-05, 'epoch': 0.75}) +[2026-06-17 19:21:39,944][starVLA.training.train_starvla][INFO] - Step 1654, Loss: {'train/loss': 0.5240848064422607, 'train/grad_norm_pre_clip': 13.297887802124023, 'timing/data': 0.010380471125245094, 'timing/model': 0.9519814383238554, 'learning_rate/qwen_vl_interface': 8.859820661633775e-06, 'learning_rate/action_model': 8.859820661633775e-05, 'epoch': 0.75}) +[2026-06-17 19:21:43,912][starVLA.training.train_starvla][INFO] - Step 1655, Loss: {'train/loss': 0.5933722853660583, 'train/grad_norm_pre_clip': 13.221792221069336, 'timing/data': 0.00894620455801487, 'timing/model': 0.9558763485401869, 'learning_rate/qwen_vl_interface': 8.858414609529193e-06, 'learning_rate/action_model': 8.858414609529193e-05, 'epoch': 0.75}) +[2026-06-17 19:21:47,892][starVLA.training.train_starvla][INFO] - Step 1656, Loss: {'train/loss': 0.29153144359588623, 'train/grad_norm_pre_clip': 7.9380083084106445, 'timing/data': 0.011369437910616398, 'timing/model': 0.9552110470831394, 'learning_rate/qwen_vl_interface': 8.857007809396384e-06, 'learning_rate/action_model': 8.857007809396383e-05, 'epoch': 0.75}) +[2026-06-17 19:21:51,770][starVLA.training.train_starvla][INFO] - Step 1657, Loss: {'train/loss': 0.4055163264274597, 'train/grad_norm_pre_clip': 9.670073509216309, 'timing/data': 0.013608449138700962, 'timing/model': 0.9505270114168525, 'learning_rate/qwen_vl_interface': 8.855600261526973e-06, 'learning_rate/action_model': 8.855600261526973e-05, 'epoch': 0.75}) +[2026-06-17 19:21:55,740][starVLA.training.train_starvla][INFO] - Step 1658, Loss: {'train/loss': 0.5668144822120667, 'train/grad_norm_pre_clip': 13.35162353515625, 'timing/data': 0.013932157307863235, 'timing/model': 1.185550726018846, 'learning_rate/qwen_vl_interface': 8.854191966212747e-06, 'learning_rate/action_model': 8.854191966212748e-05, 'epoch': 0.75}) +[2026-06-17 19:21:59,668][starVLA.training.train_starvla][INFO] - Step 1659, Loss: {'train/loss': 0.7097863554954529, 'train/grad_norm_pre_clip': 14.757196426391602, 'timing/data': 0.008629638701677322, 'timing/model': 1.1575100235641003, 'learning_rate/qwen_vl_interface': 8.852782923745648e-06, 'learning_rate/action_model': 8.852782923745648e-05, 'epoch': 0.75}) +[2026-06-17 19:22:03,406][starVLA.training.train_starvla][INFO] - Step 1660, Loss: {'train/loss': 0.38720595836639404, 'train/grad_norm_pre_clip': 9.01358413696289, 'timing/data': 0.01048762071877718, 'timing/model': 0.9685571920126677, 'learning_rate/qwen_vl_interface': 8.851373134417773e-06, 'learning_rate/action_model': 8.851373134417773e-05, 'epoch': 0.75}) +[2026-06-17 19:22:07,346][starVLA.training.train_starvla][INFO] - Step 1661, Loss: {'train/loss': 0.6730192303657532, 'train/grad_norm_pre_clip': 12.073131561279297, 'timing/data': 0.008217396214604378, 'timing/model': 0.9418014297261834, 'learning_rate/qwen_vl_interface': 8.849962598521373e-06, 'learning_rate/action_model': 8.849962598521373e-05, 'epoch': 0.75}) +[2026-06-17 19:22:11,421][starVLA.training.train_starvla][INFO] - Step 1662, Loss: {'train/loss': 0.49482277035713196, 'train/grad_norm_pre_clip': 10.899750709533691, 'timing/data': 0.009884759783744812, 'timing/model': 0.9612135365605354, 'learning_rate/qwen_vl_interface': 8.848551316348852e-06, 'learning_rate/action_model': 8.848551316348853e-05, 'epoch': 0.75}) +[2026-06-17 19:22:15,499][starVLA.training.train_starvla][INFO] - Step 1663, Loss: {'train/loss': 0.37269553542137146, 'train/grad_norm_pre_clip': 11.227705001831055, 'timing/data': 0.01109408587217331, 'timing/model': 1.0411124089732766, 'learning_rate/qwen_vl_interface': 8.847139288192771e-06, 'learning_rate/action_model': 8.84713928819277e-05, 'epoch': 0.75}) +[2026-06-17 19:22:19,909][starVLA.training.train_starvla][INFO] - Step 1664, Loss: {'train/loss': 0.3081716001033783, 'train/grad_norm_pre_clip': 8.705068588256836, 'timing/data': 0.00799026433378458, 'timing/model': 1.0312310233712196, 'learning_rate/qwen_vl_interface': 8.845726514345845e-06, 'learning_rate/action_model': 8.845726514345843e-05, 'epoch': 0.75}) +[2026-06-17 19:22:23,755][starVLA.training.train_starvla][INFO] - Step 1665, Loss: {'train/loss': 0.5915195345878601, 'train/grad_norm_pre_clip': 17.160112380981445, 'timing/data': 0.008759291842579842, 'timing/model': 1.1332434974610806, 'learning_rate/qwen_vl_interface': 8.844312995100943e-06, 'learning_rate/action_model': 8.844312995100942e-05, 'epoch': 0.76}) +[2026-06-17 19:22:27,545][starVLA.training.train_starvla][INFO] - Step 1666, Loss: {'train/loss': 0.44512802362442017, 'train/grad_norm_pre_clip': 11.185513496398926, 'timing/data': 0.01035973709076643, 'timing/model': 0.9624510556459427, 'learning_rate/qwen_vl_interface': 8.842898730751088e-06, 'learning_rate/action_model': 8.842898730751087e-05, 'epoch': 0.76}) +[2026-06-17 19:22:31,446][starVLA.training.train_starvla][INFO] - Step 1667, Loss: {'train/loss': 0.5055920481681824, 'train/grad_norm_pre_clip': 12.743724822998047, 'timing/data': 0.007918397895991802, 'timing/model': 0.9421928049996495, 'learning_rate/qwen_vl_interface': 8.841483721589462e-06, 'learning_rate/action_model': 8.84148372158946e-05, 'epoch': 0.76}) +[2026-06-17 19:22:35,367][starVLA.training.train_starvla][INFO] - Step 1668, Loss: {'train/loss': 0.7168143391609192, 'train/grad_norm_pre_clip': 12.020048141479492, 'timing/data': 0.00888404157012701, 'timing/model': 0.9472783459350467, 'learning_rate/qwen_vl_interface': 8.840067967909396e-06, 'learning_rate/action_model': 8.840067967909394e-05, 'epoch': 0.76}) +[2026-06-17 19:22:40,624][starVLA.training.train_starvla][INFO] - Step 1669, Loss: {'train/loss': 0.5068541169166565, 'train/grad_norm_pre_clip': 13.694132804870605, 'timing/data': 0.011810515075922012, 'timing/model': 2.1348138572648168, 'learning_rate/qwen_vl_interface': 8.838651470004376e-06, 'learning_rate/action_model': 8.838651470004375e-05, 'epoch': 0.76}) +[2026-06-17 19:22:44,690][starVLA.training.train_starvla][INFO] - Step 1670, Loss: {'train/loss': 0.9614094495773315, 'train/grad_norm_pre_clip': 14.400470733642578, 'timing/data': 0.013033542782068253, 'timing/model': 1.0469288676977158, 'learning_rate/qwen_vl_interface': 8.837234228168045e-06, 'learning_rate/action_model': 8.837234228168044e-05, 'epoch': 0.76}) +[2026-06-17 19:22:48,571][starVLA.training.train_starvla][INFO] - Step 1671, Loss: {'train/loss': 0.5898627638816833, 'train/grad_norm_pre_clip': 11.103316307067871, 'timing/data': 0.008171499706804752, 'timing/model': 1.1625691559165716, 'learning_rate/qwen_vl_interface': 8.835816242694197e-06, 'learning_rate/action_model': 8.835816242694197e-05, 'epoch': 0.76}) +[2026-06-17 19:22:53,359][starVLA.training.train_starvla][INFO] - Step 1672, Loss: {'train/loss': 0.6392443776130676, 'train/grad_norm_pre_clip': 12.889688491821289, 'timing/data': 0.013674275018274784, 'timing/model': 1.1636105934157968, 'learning_rate/qwen_vl_interface': 8.834397513876786e-06, 'learning_rate/action_model': 8.834397513876785e-05, 'epoch': 0.76}) +[2026-06-17 19:22:58,644][starVLA.training.train_starvla][INFO] - Step 1673, Loss: {'train/loss': 0.7667280435562134, 'train/grad_norm_pre_clip': 18.60952377319336, 'timing/data': 0.01512170024216175, 'timing/model': 0.9782440261915326, 'learning_rate/qwen_vl_interface': 8.832978042009913e-06, 'learning_rate/action_model': 8.832978042009912e-05, 'epoch': 0.76}) +[2026-06-17 19:23:02,659][starVLA.training.train_starvla][INFO] - Step 1674, Loss: {'train/loss': 0.3656501770019531, 'train/grad_norm_pre_clip': 11.126131057739258, 'timing/data': 0.013095392845571041, 'timing/model': 0.9787280904129148, 'learning_rate/qwen_vl_interface': 8.831557827387835e-06, 'learning_rate/action_model': 8.831557827387834e-05, 'epoch': 0.76}) +[2026-06-17 19:23:06,827][starVLA.training.train_starvla][INFO] - Step 1675, Loss: {'train/loss': 0.5302479863166809, 'train/grad_norm_pre_clip': 12.255619049072266, 'timing/data': 0.011579392477869987, 'timing/model': 0.9670429611578584, 'learning_rate/qwen_vl_interface': 8.830136870304966e-06, 'learning_rate/action_model': 8.830136870304965e-05, 'epoch': 0.76}) +[2026-06-17 19:23:12,080][starVLA.training.train_starvla][INFO] - Step 1676, Loss: {'train/loss': 0.5274090766906738, 'train/grad_norm_pre_clip': 12.40454387664795, 'timing/data': 0.009171932935714722, 'timing/model': 2.21623363904655, 'learning_rate/qwen_vl_interface': 8.828715171055871e-06, 'learning_rate/action_model': 8.828715171055871e-05, 'epoch': 0.76}) +[2026-06-17 19:23:16,199][starVLA.training.train_starvla][INFO] - Step 1677, Loss: {'train/loss': 0.37129876017570496, 'train/grad_norm_pre_clip': 8.856006622314453, 'timing/data': 0.01498955674469471, 'timing/model': 1.093517235480249, 'learning_rate/qwen_vl_interface': 8.827292729935271e-06, 'learning_rate/action_model': 8.827292729935271e-05, 'epoch': 0.76}) +[2026-06-17 19:23:20,143][starVLA.training.train_starvla][INFO] - Step 1678, Loss: {'train/loss': 0.5104807615280151, 'train/grad_norm_pre_clip': 12.501173973083496, 'timing/data': 0.013356940820813179, 'timing/model': 1.1553995022550225, 'learning_rate/qwen_vl_interface': 8.825869547238038e-06, 'learning_rate/action_model': 8.825869547238038e-05, 'epoch': 0.76}) +[2026-06-17 19:23:23,991][starVLA.training.train_starvla][INFO] - Step 1679, Loss: {'train/loss': 0.4394843280315399, 'train/grad_norm_pre_clip': 10.588200569152832, 'timing/data': 0.010485941544175148, 'timing/model': 1.1061502527445555, 'learning_rate/qwen_vl_interface': 8.824445623259202e-06, 'learning_rate/action_model': 8.8244456232592e-05, 'epoch': 0.76}) +[2026-06-17 19:23:27,638][starVLA.training.train_starvla][INFO] - Step 1680, Loss: {'train/loss': 0.5593590140342712, 'train/grad_norm_pre_clip': 9.734119415283203, 'timing/data': 0.012398519553244114, 'timing/model': 0.9436757350340486, 'learning_rate/qwen_vl_interface': 8.82302095829394e-06, 'learning_rate/action_model': 8.82302095829394e-05, 'epoch': 0.76}) +[2026-06-17 19:23:31,450][starVLA.training.train_starvla][INFO] - Step 1681, Loss: {'train/loss': 0.5255821943283081, 'train/grad_norm_pre_clip': 11.945897102355957, 'timing/data': 0.009843478910624981, 'timing/model': 0.931306947953999, 'learning_rate/qwen_vl_interface': 8.82159555263759e-06, 'learning_rate/action_model': 8.821595552637589e-05, 'epoch': 0.76}) +[2026-06-17 19:23:35,283][starVLA.training.train_starvla][INFO] - Step 1682, Loss: {'train/loss': 0.3646583557128906, 'train/grad_norm_pre_clip': 11.75226879119873, 'timing/data': 0.011531729251146317, 'timing/model': 0.9308146852999926, 'learning_rate/qwen_vl_interface': 8.820169406585638e-06, 'learning_rate/action_model': 8.820169406585637e-05, 'epoch': 0.76}) +[2026-06-17 19:23:39,095][starVLA.training.train_starvla][INFO] - Step 1683, Loss: {'train/loss': 0.4283479154109955, 'train/grad_norm_pre_clip': 12.33448314666748, 'timing/data': 0.016136356629431248, 'timing/model': 0.9474715311080217, 'learning_rate/qwen_vl_interface': 8.818742520433727e-06, 'learning_rate/action_model': 8.818742520433726e-05, 'epoch': 0.76}) +[2026-06-17 19:23:42,999][starVLA.training.train_starvla][INFO] - Step 1684, Loss: {'train/loss': 0.45557355880737305, 'train/grad_norm_pre_clip': 11.410942077636719, 'timing/data': 0.01033534575253725, 'timing/model': 0.9471693746745586, 'learning_rate/qwen_vl_interface': 8.81731489447765e-06, 'learning_rate/action_model': 8.81731489447765e-05, 'epoch': 0.76}) +[2026-06-17 19:23:46,869][starVLA.training.train_starvla][INFO] - Step 1685, Loss: {'train/loss': 0.5599388480186462, 'train/grad_norm_pre_clip': 11.29541015625, 'timing/data': 0.008456098847091198, 'timing/model': 1.1176623729988933, 'learning_rate/qwen_vl_interface': 8.815886529013356e-06, 'learning_rate/action_model': 8.815886529013356e-05, 'epoch': 0.76}) +[2026-06-17 19:23:50,610][starVLA.training.train_starvla][INFO] - Step 1686, Loss: {'train/loss': 0.33214670419692993, 'train/grad_norm_pre_clip': 9.970211029052734, 'timing/data': 0.01985722966492176, 'timing/model': 0.9480901844799519, 'learning_rate/qwen_vl_interface': 8.814457424336951e-06, 'learning_rate/action_model': 8.81445742433695e-05, 'epoch': 0.76}) +[2026-06-17 19:23:54,447][starVLA.training.train_starvla][INFO] - Step 1687, Loss: {'train/loss': 0.4984961152076721, 'train/grad_norm_pre_clip': 12.027312278747559, 'timing/data': 0.012083031237125397, 'timing/model': 0.9485162440687418, 'learning_rate/qwen_vl_interface': 8.813027580744685e-06, 'learning_rate/action_model': 8.813027580744684e-05, 'epoch': 0.77}) +[2026-06-17 19:23:58,281][starVLA.training.train_starvla][INFO] - Step 1688, Loss: {'train/loss': 0.4605695605278015, 'train/grad_norm_pre_clip': 9.502540588378906, 'timing/data': 0.007489501498639584, 'timing/model': 0.9458923572674394, 'learning_rate/qwen_vl_interface': 8.811596998532965e-06, 'learning_rate/action_model': 8.811596998532965e-05, 'epoch': 0.77}) +[2026-06-17 19:24:02,507][starVLA.training.train_starvla][INFO] - Step 1689, Loss: {'train/loss': 0.5603214502334595, 'train/grad_norm_pre_clip': 11.791793823242188, 'timing/data': 0.008489668369293213, 'timing/model': 0.9781245039775968, 'learning_rate/qwen_vl_interface': 8.810165677998356e-06, 'learning_rate/action_model': 8.810165677998356e-05, 'epoch': 0.77}) +[2026-06-17 19:24:06,497][starVLA.training.train_starvla][INFO] - Step 1690, Loss: {'train/loss': 0.38188809156417847, 'train/grad_norm_pre_clip': 15.293526649475098, 'timing/data': 0.011681039817631245, 'timing/model': 1.021079775877297, 'learning_rate/qwen_vl_interface': 8.808733619437572e-06, 'learning_rate/action_model': 8.808733619437572e-05, 'epoch': 0.77}) +[2026-06-17 19:24:10,809][starVLA.training.train_starvla][INFO] - Step 1691, Loss: {'train/loss': 0.7410538196563721, 'train/grad_norm_pre_clip': 16.223527908325195, 'timing/data': 0.011948060244321823, 'timing/model': 1.0020777052268386, 'learning_rate/qwen_vl_interface': 8.80730082314748e-06, 'learning_rate/action_model': 8.80730082314748e-05, 'epoch': 0.77}) +[2026-06-17 19:24:14,999][starVLA.training.train_starvla][INFO] - Step 1692, Loss: {'train/loss': 0.43690580129623413, 'train/grad_norm_pre_clip': 11.796026229858398, 'timing/data': 0.029855151660740376, 'timing/model': 1.3706563990563154, 'learning_rate/qwen_vl_interface': 8.805867289425098e-06, 'learning_rate/action_model': 8.805867289425098e-05, 'epoch': 0.77}) +[2026-06-17 19:24:19,300][starVLA.training.train_starvla][INFO] - Step 1693, Loss: {'train/loss': 0.6126070618629456, 'train/grad_norm_pre_clip': 13.354169845581055, 'timing/data': 0.02141923736780882, 'timing/model': 0.9532121457159519, 'learning_rate/qwen_vl_interface': 8.804433018567601e-06, 'learning_rate/action_model': 8.8044330185676e-05, 'epoch': 0.77}) +[2026-06-17 19:24:26,557][starVLA.training.train_starvla][INFO] - Step 1694, Loss: {'train/loss': 0.5447920560836792, 'train/grad_norm_pre_clip': 15.430962562561035, 'timing/data': 0.01730577740818262, 'timing/model': 0.953431929461658, 'learning_rate/qwen_vl_interface': 8.802998010872316e-06, 'learning_rate/action_model': 8.802998010872316e-05, 'epoch': 0.77}) +[2026-06-17 19:24:30,542][starVLA.training.train_starvla][INFO] - Step 1695, Loss: {'train/loss': 0.5537857413291931, 'train/grad_norm_pre_clip': 10.630192756652832, 'timing/data': 0.020577792078256607, 'timing/model': 0.9431394580751657, 'learning_rate/qwen_vl_interface': 8.801562266636719e-06, 'learning_rate/action_model': 8.801562266636719e-05, 'epoch': 0.77}) +[2026-06-17 19:24:34,542][starVLA.training.train_starvla][INFO] - Step 1696, Loss: {'train/loss': 0.44922107458114624, 'train/grad_norm_pre_clip': 11.297480583190918, 'timing/data': 0.022274465300142765, 'timing/model': 0.967157643288374, 'learning_rate/qwen_vl_interface': 8.800125786158442e-06, 'learning_rate/action_model': 8.800125786158442e-05, 'epoch': 0.77}) +[2026-06-17 19:24:38,873][starVLA.training.train_starvla][INFO] - Step 1697, Loss: {'train/loss': 0.4882783889770508, 'train/grad_norm_pre_clip': 10.755199432373047, 'timing/data': 0.01958101149648428, 'timing/model': 1.1468076696619391, 'learning_rate/qwen_vl_interface': 8.798688569735268e-06, 'learning_rate/action_model': 8.798688569735268e-05, 'epoch': 0.77}) +[2026-06-17 19:24:43,066][starVLA.training.train_starvla][INFO] - Step 1698, Loss: {'train/loss': 0.4058774709701538, 'train/grad_norm_pre_clip': 9.944844245910645, 'timing/data': 0.022780202329158783, 'timing/model': 1.2791397105902433, 'learning_rate/qwen_vl_interface': 8.797250617665139e-06, 'learning_rate/action_model': 8.797250617665138e-05, 'epoch': 0.77}) +[2026-06-17 19:24:47,295][starVLA.training.train_starvla][INFO] - Step 1699, Loss: {'train/loss': 0.4277741014957428, 'train/grad_norm_pre_clip': 12.353479385375977, 'timing/data': 0.015706229023635387, 'timing/model': 1.3591651795431972, 'learning_rate/qwen_vl_interface': 8.795811930246136e-06, 'learning_rate/action_model': 8.795811930246136e-05, 'epoch': 0.77}) +[2026-06-17 19:27:10,737][starVLA.training.train_starvla][INFO] - Step 1700, Loss: {'train/loss': 0.34701162576675415, 'train/grad_norm_pre_clip': 11.411392211914062, 'eval/loss': 0.48145875334739685, 'eval/latency_8/loss': 0.48145875334739685, 'eval/demon_fire_precision': 0.9254385964912281, 'eval/demon_fire_recall': 0.983224603914259, 'eval/demon_fire_f1': 0.9534568459105286, 'eval/demon_fire_rate_model': 0.9515859766277128, 'eval/demon_fire_rate_teacher': 0.8956594323873122, 'eval/demon_left_precision': 0.7670454545454546, 'eval/demon_left_recall': 0.5781584582441114, 'eval/demon_left_f1': 0.6593406593406594, 'eval/demon_left_rate_model': 0.09794101279910963, 'eval/demon_left_rate_teacher': 0.12993878686700056, 'eval/demon_right_precision': 0.7373572593800979, 'eval/demon_right_recall': 0.6848484848484848, 'eval/demon_right_f1': 0.7101335428122546, 'eval/demon_right_rate_model': 0.17056204785754034, 'eval/demon_right_rate_teacher': 0.18363939899833054, 'eval/demon_move_f1': 0.684737101076457, 'eval/demon_cc_f1': 0.8530276300587352, 'eval/latency_8/demon_fire_precision': 0.9254385964912281, 'eval/latency_8/demon_fire_recall': 0.983224603914259, 'eval/latency_8/demon_fire_f1': 0.9534568459105286, 'eval/latency_8/demon_fire_rate_model': 0.9515859766277128, 'eval/latency_8/demon_fire_rate_teacher': 0.8956594323873122, 'eval/latency_8/demon_left_precision': 0.7670454545454546, 'eval/latency_8/demon_left_recall': 0.5781584582441114, 'eval/latency_8/demon_left_f1': 0.6593406593406594, 'eval/latency_8/demon_left_rate_model': 0.09794101279910963, 'eval/latency_8/demon_left_rate_teacher': 0.12993878686700056, 'eval/latency_8/demon_right_precision': 0.7373572593800979, 'eval/latency_8/demon_right_recall': 0.6848484848484848, 'eval/latency_8/demon_right_f1': 0.7101335428122546, 'eval/latency_8/demon_right_rate_model': 0.17056204785754034, 'eval/latency_8/demon_right_rate_teacher': 0.18363939899833054, 'eval/latency_8/demon_move_f1': 0.684737101076457, 'eval/latency_8/demon_cc_f1': 0.8530276300587352, 'timing/data': 0.016794917173683643, 'timing/model': 0.9655164917930961, 'learning_rate/qwen_vl_interface': 8.794372507776508e-06, 'learning_rate/action_model': 8.794372507776508e-05, 'epoch': 0.77}) +[2026-06-17 19:27:10,740][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_1700_state +[2026-06-17 19:27:10,760][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 19:27:18,178][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_1700_state/model.safetensors +[2026-06-17 19:27:33,859][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_1700_state/optimizer.bin +[2026-06-17 19:27:33,860][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_1700_state/sampler.bin +[2026-06-17 19:27:33,860][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_1700_state/sampler_1.bin +[2026-06-17 19:27:33,862][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_1700_state/random_states_0.pkl +[2026-06-17 19:27:33,870][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1700 +[2026-06-17 19:27:33,870][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 19:27:33,884][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 19:27:37,809][starVLA.training.train_starvla][INFO] - Step 1701, Loss: {'train/loss': 0.6010273694992065, 'train/grad_norm_pre_clip': 13.560348510742188, 'timing/data': 0.01741977035999298, 'timing/model': 1.20616165548563, 'learning_rate/qwen_vl_interface': 8.792932350554642e-06, 'learning_rate/action_model': 8.792932350554643e-05, 'epoch': 0.77}) +[2026-06-17 19:27:41,448][starVLA.training.train_starvla][INFO] - Step 1702, Loss: {'train/loss': 0.3396061062812805, 'train/grad_norm_pre_clip': 8.871964454650879, 'timing/data': 0.01975980680435896, 'timing/model': 0.9383359281346202, 'learning_rate/qwen_vl_interface': 8.791491458879089e-06, 'learning_rate/action_model': 8.791491458879089e-05, 'epoch': 0.77}) +[2026-06-17 19:27:45,328][starVLA.training.train_starvla][INFO] - Step 1703, Loss: {'train/loss': 0.4081478714942932, 'train/grad_norm_pre_clip': 12.279550552368164, 'timing/data': 0.019183742813766003, 'timing/model': 0.9323022477328777, 'learning_rate/qwen_vl_interface': 8.790049833048545e-06, 'learning_rate/action_model': 8.790049833048545e-05, 'epoch': 0.77}) +[2026-06-17 19:27:49,152][starVLA.training.train_starvla][INFO] - Step 1704, Loss: {'train/loss': 0.7616398334503174, 'train/grad_norm_pre_clip': 16.751127243041992, 'timing/data': 0.017490596510469913, 'timing/model': 1.0116655845195055, 'learning_rate/qwen_vl_interface': 8.788607473361862e-06, 'learning_rate/action_model': 8.788607473361862e-05, 'epoch': 0.77}) +[2026-06-17 19:27:53,529][starVLA.training.train_starvla][INFO] - Step 1705, Loss: {'train/loss': 0.26456135511398315, 'train/grad_norm_pre_clip': 15.730010032653809, 'timing/data': 0.02417451050132513, 'timing/model': 1.26644941046834, 'learning_rate/qwen_vl_interface': 8.787164380118044e-06, 'learning_rate/action_model': 8.787164380118043e-05, 'epoch': 0.77}) +[2026-06-17 19:27:57,427][starVLA.training.train_starvla][INFO] - Step 1706, Loss: {'train/loss': 0.41754066944122314, 'train/grad_norm_pre_clip': 11.241536140441895, 'timing/data': 0.030380381271243095, 'timing/model': 0.9314618753269315, 'learning_rate/qwen_vl_interface': 8.785720553616241e-06, 'learning_rate/action_model': 8.78572055361624e-05, 'epoch': 0.77}) +[2026-06-17 19:28:01,296][starVLA.training.train_starvla][INFO] - Step 1707, Loss: {'train/loss': 0.8377298712730408, 'train/grad_norm_pre_clip': 17.3155517578125, 'timing/data': 0.016484285704791546, 'timing/model': 1.0324394386261702, 'learning_rate/qwen_vl_interface': 8.784275994155762e-06, 'learning_rate/action_model': 8.784275994155762e-05, 'epoch': 0.77}) +[2026-06-17 19:28:05,211][starVLA.training.train_starvla][INFO] - Step 1708, Loss: {'train/loss': 0.48774629831314087, 'train/grad_norm_pre_clip': 10.651981353759766, 'timing/data': 0.018980220891535282, 'timing/model': 1.1743903309106827, 'learning_rate/qwen_vl_interface': 8.782830702036067e-06, 'learning_rate/action_model': 8.782830702036067e-05, 'epoch': 0.77}) +[2026-06-17 19:28:09,322][starVLA.training.train_starvla][INFO] - Step 1709, Loss: {'train/loss': 0.46712252497673035, 'train/grad_norm_pre_clip': 17.163105010986328, 'timing/data': 0.02051689848303795, 'timing/model': 1.0534422881901264, 'learning_rate/qwen_vl_interface': 8.781384677556767e-06, 'learning_rate/action_model': 8.781384677556767e-05, 'epoch': 0.78}) +[2026-06-17 19:28:13,476][starVLA.training.train_starvla][INFO] - Step 1710, Loss: {'train/loss': 0.5509873628616333, 'train/grad_norm_pre_clip': 58.721927642822266, 'timing/data': 0.020429099909961224, 'timing/model': 1.1169316656887531, 'learning_rate/qwen_vl_interface': 8.779937921017624e-06, 'learning_rate/action_model': 8.779937921017623e-05, 'epoch': 0.78}) +[2026-06-17 19:28:17,684][starVLA.training.train_starvla][INFO] - Step 1711, Loss: {'train/loss': 0.7076743841171265, 'train/grad_norm_pre_clip': 13.667374610900879, 'timing/data': 0.015405605547130108, 'timing/model': 1.0080138463526964, 'learning_rate/qwen_vl_interface': 8.77849043271855e-06, 'learning_rate/action_model': 8.778490432718551e-05, 'epoch': 0.78}) +[2026-06-17 19:28:21,926][starVLA.training.train_starvla][INFO] - Step 1712, Loss: {'train/loss': 0.44688600301742554, 'train/grad_norm_pre_clip': 10.638540267944336, 'timing/data': 0.017650562338531017, 'timing/model': 1.0331643037497997, 'learning_rate/qwen_vl_interface': 8.777042212959615e-06, 'learning_rate/action_model': 8.777042212959614e-05, 'epoch': 0.78}) +[2026-06-17 19:28:26,221][starVLA.training.train_starvla][INFO] - Step 1713, Loss: {'train/loss': 0.37210187315940857, 'train/grad_norm_pre_clip': 10.38437271118164, 'timing/data': 0.03606557846069336, 'timing/model': 0.9894207948818803, 'learning_rate/qwen_vl_interface': 8.775593262041032e-06, 'learning_rate/action_model': 8.77559326204103e-05, 'epoch': 0.78}) +[2026-06-17 19:28:36,500][starVLA.training.train_starvla][INFO] - Step 1714, Loss: {'train/loss': 0.6898131966590881, 'train/grad_norm_pre_clip': 17.591903686523438, 'timing/data': 0.018910538405179977, 'timing/model': 3.9186845188960433, 'learning_rate/qwen_vl_interface': 8.774143580263172e-06, 'learning_rate/action_model': 8.77414358026317e-05, 'epoch': 0.78}) +[2026-06-17 19:28:41,390][starVLA.training.train_starvla][INFO] - Step 1715, Loss: {'train/loss': 0.6515472531318665, 'train/grad_norm_pre_clip': 14.756634712219238, 'timing/data': 0.017978385090827942, 'timing/model': 1.2207596888765693, 'learning_rate/qwen_vl_interface': 8.772693167926554e-06, 'learning_rate/action_model': 8.772693167926555e-05, 'epoch': 0.78}) +[2026-06-17 19:28:46,923][starVLA.training.train_starvla][INFO] - Step 1716, Loss: {'train/loss': 0.4250640869140625, 'train/grad_norm_pre_clip': 11.300737380981445, 'timing/data': 0.06538404989987612, 'timing/model': 1.0133447013795376, 'learning_rate/qwen_vl_interface': 8.771242025331855e-06, 'learning_rate/action_model': 8.771242025331854e-05, 'epoch': 0.78}) +[2026-06-17 19:28:52,958][starVLA.training.train_starvla][INFO] - Step 1717, Loss: {'train/loss': 0.7572570443153381, 'train/grad_norm_pre_clip': 19.717924118041992, 'timing/data': 0.01661392580717802, 'timing/model': 1.223371827043593, 'learning_rate/qwen_vl_interface': 8.769790152779893e-06, 'learning_rate/action_model': 8.769790152779892e-05, 'epoch': 0.78}) +[2026-06-17 19:28:57,429][starVLA.training.train_starvla][INFO] - Step 1718, Loss: {'train/loss': 0.6592830419540405, 'train/grad_norm_pre_clip': 10.98624038696289, 'timing/data': 0.023700885474681854, 'timing/model': 1.253750454634428, 'learning_rate/qwen_vl_interface': 8.768337550571644e-06, 'learning_rate/action_model': 8.768337550571645e-05, 'epoch': 0.78}) +[2026-06-17 19:29:01,293][starVLA.training.train_starvla][INFO] - Step 1719, Loss: {'train/loss': 0.4890749752521515, 'train/grad_norm_pre_clip': 10.045905113220215, 'timing/data': 0.021924628876149654, 'timing/model': 1.0109632099047303, 'learning_rate/qwen_vl_interface': 8.766884219008238e-06, 'learning_rate/action_model': 8.766884219008238e-05, 'epoch': 0.78}) +[2026-06-17 19:29:05,365][starVLA.training.train_starvla][INFO] - Step 1720, Loss: {'train/loss': 0.5424520969390869, 'train/grad_norm_pre_clip': 12.037482261657715, 'timing/data': 0.022003063932061195, 'timing/model': 0.9617337612435222, 'learning_rate/qwen_vl_interface': 8.765430158390948e-06, 'learning_rate/action_model': 8.765430158390947e-05, 'epoch': 0.78}) +[2026-06-17 19:29:09,407][starVLA.training.train_starvla][INFO] - Step 1721, Loss: {'train/loss': 0.5122485756874084, 'train/grad_norm_pre_clip': 12.1205415725708, 'timing/data': 0.018823960795998573, 'timing/model': 1.2133783856406808, 'learning_rate/qwen_vl_interface': 8.763975369021202e-06, 'learning_rate/action_model': 8.763975369021203e-05, 'epoch': 0.78}) +[2026-06-17 19:29:13,064][starVLA.training.train_starvla][INFO] - Step 1722, Loss: {'train/loss': 0.7459492683410645, 'train/grad_norm_pre_clip': 15.544689178466797, 'timing/data': 0.01670998800545931, 'timing/model': 0.937208347953856, 'learning_rate/qwen_vl_interface': 8.762519851200581e-06, 'learning_rate/action_model': 8.762519851200581e-05, 'epoch': 0.78}) +[2026-06-17 19:29:17,599][starVLA.training.train_starvla][INFO] - Step 1723, Loss: {'train/loss': 0.3973042070865631, 'train/grad_norm_pre_clip': 9.45601749420166, 'timing/data': 0.023812281899154186, 'timing/model': 1.0848368220031261, 'learning_rate/qwen_vl_interface': 8.761063605230818e-06, 'learning_rate/action_model': 8.761063605230818e-05, 'epoch': 0.78}) +[2026-06-17 19:29:21,507][starVLA.training.train_starvla][INFO] - Step 1724, Loss: {'train/loss': 0.5346980690956116, 'train/grad_norm_pre_clip': 12.671451568603516, 'timing/data': 0.023076990619301796, 'timing/model': 1.038325371220708, 'learning_rate/qwen_vl_interface': 8.759606631413788e-06, 'learning_rate/action_model': 8.759606631413788e-05, 'epoch': 0.78}) +[2026-06-17 19:29:25,493][starVLA.training.train_starvla][INFO] - Step 1725, Loss: {'train/loss': 0.5439514517784119, 'train/grad_norm_pre_clip': 13.249937057495117, 'timing/data': 0.015862195752561092, 'timing/model': 1.072244375012815, 'learning_rate/qwen_vl_interface': 8.75814893005153e-06, 'learning_rate/action_model': 8.75814893005153e-05, 'epoch': 0.78}) +[2026-06-17 19:29:29,457][starVLA.training.train_starvla][INFO] - Step 1726, Loss: {'train/loss': 0.35876500606536865, 'train/grad_norm_pre_clip': 7.828178882598877, 'timing/data': 0.014886686578392982, 'timing/model': 0.9450614107772708, 'learning_rate/qwen_vl_interface': 8.756690501446222e-06, 'learning_rate/action_model': 8.756690501446221e-05, 'epoch': 0.78}) +[2026-06-17 19:29:33,722][starVLA.training.train_starvla][INFO] - Step 1727, Loss: {'train/loss': 0.4953975975513458, 'train/grad_norm_pre_clip': 9.569025039672852, 'timing/data': 0.020788123831152916, 'timing/model': 1.452342220582068, 'learning_rate/qwen_vl_interface': 8.755231345900199e-06, 'learning_rate/action_model': 8.755231345900198e-05, 'epoch': 0.78}) +[2026-06-17 19:29:37,875][starVLA.training.train_starvla][INFO] - Step 1728, Loss: {'train/loss': 0.5046775341033936, 'train/grad_norm_pre_clip': 10.933467864990234, 'timing/data': 0.020481493324041367, 'timing/model': 1.4074127869680524, 'learning_rate/qwen_vl_interface': 8.753771463715947e-06, 'learning_rate/action_model': 8.753771463715947e-05, 'epoch': 0.78}) +[2026-06-17 19:29:41,946][starVLA.training.train_starvla][INFO] - Step 1729, Loss: {'train/loss': 0.5553337335586548, 'train/grad_norm_pre_clip': 12.651989936828613, 'timing/data': 0.023799443617463112, 'timing/model': 1.055769370868802, 'learning_rate/qwen_vl_interface': 8.7523108551961e-06, 'learning_rate/action_model': 8.7523108551961e-05, 'epoch': 0.78}) +[2026-06-17 19:29:46,477][starVLA.training.train_starvla][INFO] - Step 1730, Loss: {'train/loss': 0.41994526982307434, 'train/grad_norm_pre_clip': 8.707886695861816, 'timing/data': 0.019389753229916096, 'timing/model': 0.9595985421910882, 'learning_rate/qwen_vl_interface': 8.750849520643445e-06, 'learning_rate/action_model': 8.750849520643444e-05, 'epoch': 0.78}) +[2026-06-17 19:29:50,375][starVLA.training.train_starvla][INFO] - Step 1731, Loss: {'train/loss': 0.5807240605354309, 'train/grad_norm_pre_clip': 13.18557357788086, 'timing/data': 0.018554357811808586, 'timing/model': 0.9669721834361553, 'learning_rate/qwen_vl_interface': 8.749387460360917e-06, 'learning_rate/action_model': 8.749387460360916e-05, 'epoch': 0.79}) +[2026-06-17 19:29:55,713][starVLA.training.train_starvla][INFO] - Step 1732, Loss: {'train/loss': 0.35425153374671936, 'train/grad_norm_pre_clip': 10.626500129699707, 'timing/data': 0.017918864265084267, 'timing/model': 1.0394623661413789, 'learning_rate/qwen_vl_interface': 8.747924674651602e-06, 'learning_rate/action_model': 8.747924674651602e-05, 'epoch': 0.79}) +[2026-06-17 19:29:59,698][starVLA.training.train_starvla][INFO] - Step 1733, Loss: {'train/loss': 0.4223497211933136, 'train/grad_norm_pre_clip': 10.802268981933594, 'timing/data': 0.02916016336530447, 'timing/model': 0.9916812730953097, 'learning_rate/qwen_vl_interface': 8.74646116381874e-06, 'learning_rate/action_model': 8.74646116381874e-05, 'epoch': 0.79}) +[2026-06-17 19:30:04,992][starVLA.training.train_starvla][INFO] - Step 1734, Loss: {'train/loss': 0.5532606840133667, 'train/grad_norm_pre_clip': 15.990100860595703, 'timing/data': 0.022503969259560108, 'timing/model': 1.4618265787139535, 'learning_rate/qwen_vl_interface': 8.744996928165714e-06, 'learning_rate/action_model': 8.744996928165714e-05, 'epoch': 0.79}) +[2026-06-17 19:30:09,833][starVLA.training.train_starvla][INFO] - Step 1735, Loss: {'train/loss': 0.5098622441291809, 'train/grad_norm_pre_clip': 11.739043235778809, 'timing/data': 0.017290029674768448, 'timing/model': 1.1902040038257837, 'learning_rate/qwen_vl_interface': 8.743531967996066e-06, 'learning_rate/action_model': 8.743531967996065e-05, 'epoch': 0.79}) +[2026-06-17 19:30:15,051][starVLA.training.train_starvla][INFO] - Step 1736, Loss: {'train/loss': 0.38233888149261475, 'train/grad_norm_pre_clip': 10.53720474243164, 'timing/data': 0.018641184084117413, 'timing/model': 0.958705447614193, 'learning_rate/qwen_vl_interface': 8.742066283613481e-06, 'learning_rate/action_model': 8.74206628361348e-05, 'epoch': 0.79}) +[2026-06-17 19:30:19,525][starVLA.training.train_starvla][INFO] - Step 1737, Loss: {'train/loss': 0.7525413036346436, 'train/grad_norm_pre_clip': 14.482279777526855, 'timing/data': 0.01609979197382927, 'timing/model': 1.1716166734695435, 'learning_rate/qwen_vl_interface': 8.740599875321797e-06, 'learning_rate/action_model': 8.740599875321797e-05, 'epoch': 0.79}) +[2026-06-17 19:30:23,638][starVLA.training.train_starvla][INFO] - Step 1738, Loss: {'train/loss': 0.5837605595588684, 'train/grad_norm_pre_clip': 11.246123313903809, 'timing/data': 0.016718877479434013, 'timing/model': 1.1291156634688377, 'learning_rate/qwen_vl_interface': 8.739132743425004e-06, 'learning_rate/action_model': 8.739132743425002e-05, 'epoch': 0.79}) +[2026-06-17 19:30:27,597][starVLA.training.train_starvla][INFO] - Step 1739, Loss: {'train/loss': 0.468316912651062, 'train/grad_norm_pre_clip': 14.651280403137207, 'timing/data': 0.018951124511659145, 'timing/model': 0.9498420348390937, 'learning_rate/qwen_vl_interface': 8.737664888227239e-06, 'learning_rate/action_model': 8.737664888227239e-05, 'epoch': 0.79}) +[2026-06-17 19:30:31,480][starVLA.training.train_starvla][INFO] - Step 1740, Loss: {'train/loss': 0.5646380186080933, 'train/grad_norm_pre_clip': 14.222908973693848, 'timing/data': 0.021267525851726532, 'timing/model': 0.9551987694576383, 'learning_rate/qwen_vl_interface': 8.736196310032787e-06, 'learning_rate/action_model': 8.736196310032787e-05, 'epoch': 0.79}) +[2026-06-17 19:30:36,755][starVLA.training.train_starvla][INFO] - Step 1741, Loss: {'train/loss': 0.31384897232055664, 'train/grad_norm_pre_clip': 7.887302398681641, 'timing/data': 0.018702581524848938, 'timing/model': 1.2733677057549357, 'learning_rate/qwen_vl_interface': 8.734727009146092e-06, 'learning_rate/action_model': 8.734727009146092e-05, 'epoch': 0.79}) +[2026-06-17 19:30:40,716][starVLA.training.train_starvla][INFO] - Step 1742, Loss: {'train/loss': 0.5399030447006226, 'train/grad_norm_pre_clip': 10.564420700073242, 'timing/data': 0.020375282503664494, 'timing/model': 1.0518854362890124, 'learning_rate/qwen_vl_interface': 8.733256985871737e-06, 'learning_rate/action_model': 8.733256985871737e-05, 'epoch': 0.79}) +[2026-06-17 19:30:47,022][starVLA.training.train_starvla][INFO] - Step 1743, Loss: {'train/loss': 0.7036039233207703, 'train/grad_norm_pre_clip': 16.365373611450195, 'timing/data': 0.022038848139345646, 'timing/model': 1.9351823134347796, 'learning_rate/qwen_vl_interface': 8.73178624051446e-06, 'learning_rate/action_model': 8.73178624051446e-05, 'epoch': 0.79}) +[2026-06-17 19:30:50,841][starVLA.training.train_starvla][INFO] - Step 1744, Loss: {'train/loss': 0.4368973970413208, 'train/grad_norm_pre_clip': 12.20178508758545, 'timing/data': 0.01986855175346136, 'timing/model': 0.9375061234459281, 'learning_rate/qwen_vl_interface': 8.73031477337915e-06, 'learning_rate/action_model': 8.73031477337915e-05, 'epoch': 0.79}) +[2026-06-17 19:30:54,714][starVLA.training.train_starvla][INFO] - Step 1745, Loss: {'train/loss': 0.43791458010673523, 'train/grad_norm_pre_clip': 9.248172760009766, 'timing/data': 0.019363234750926495, 'timing/model': 0.9351552799344063, 'learning_rate/qwen_vl_interface': 8.728842584770842e-06, 'learning_rate/action_model': 8.728842584770842e-05, 'epoch': 0.79}) +[2026-06-17 19:30:58,568][starVLA.training.train_starvla][INFO] - Step 1746, Loss: {'train/loss': 0.5194989442825317, 'train/grad_norm_pre_clip': 17.066679000854492, 'timing/data': 0.014556167647242546, 'timing/model': 0.9375500855967402, 'learning_rate/qwen_vl_interface': 8.727369674994722e-06, 'learning_rate/action_model': 8.727369674994722e-05, 'epoch': 0.79}) +[2026-06-17 19:31:02,451][starVLA.training.train_starvla][INFO] - Step 1747, Loss: {'train/loss': 0.42891213297843933, 'train/grad_norm_pre_clip': 11.040400505065918, 'timing/data': 0.016847223043441772, 'timing/model': 0.9461022336035967, 'learning_rate/qwen_vl_interface': 8.725896044356128e-06, 'learning_rate/action_model': 8.725896044356128e-05, 'epoch': 0.79}) +[2026-06-17 19:31:06,647][starVLA.training.train_starvla][INFO] - Step 1748, Loss: {'train/loss': 0.6925404667854309, 'train/grad_norm_pre_clip': 12.599359512329102, 'timing/data': 0.01735623087733984, 'timing/model': 1.2319653620943427, 'learning_rate/qwen_vl_interface': 8.724421693160544e-06, 'learning_rate/action_model': 8.724421693160543e-05, 'epoch': 0.79}) +[2026-06-17 19:31:10,304][starVLA.training.train_starvla][INFO] - Step 1749, Loss: {'train/loss': 0.3238964378833771, 'train/grad_norm_pre_clip': 11.753783226013184, 'timing/data': 0.01979354489594698, 'timing/model': 0.9478035094216466, 'learning_rate/qwen_vl_interface': 8.722946621713603e-06, 'learning_rate/action_model': 8.722946621713602e-05, 'epoch': 0.79}) +[2026-06-17 19:31:14,236][starVLA.training.train_starvla][INFO] - Step 1750, Loss: {'train/loss': 0.2580054700374603, 'train/grad_norm_pre_clip': 7.564176082611084, 'timing/data': 0.01641346514225006, 'timing/model': 0.9991421522572637, 'learning_rate/qwen_vl_interface': 8.721470830321091e-06, 'learning_rate/action_model': 8.72147083032109e-05, 'epoch': 0.79}) +[2026-06-17 19:31:18,892][starVLA.training.train_starvla][INFO] - Step 1751, Loss: {'train/loss': 0.30076709389686584, 'train/grad_norm_pre_clip': 9.316171646118164, 'timing/data': 0.015660203993320465, 'timing/model': 1.0850483123213053, 'learning_rate/qwen_vl_interface': 8.719994319288939e-06, 'learning_rate/action_model': 8.71999431928894e-05, 'epoch': 0.79}) +[2026-06-17 19:31:23,611][starVLA.training.train_starvla][INFO] - Step 1752, Loss: {'train/loss': 0.23460881412029266, 'train/grad_norm_pre_clip': 7.5497727394104, 'timing/data': 0.020565999671816826, 'timing/model': 1.1483861012384295, 'learning_rate/qwen_vl_interface': 8.718517088923233e-06, 'learning_rate/action_model': 8.718517088923233e-05, 'epoch': 0.79}) +[2026-06-17 19:31:27,649][starVLA.training.train_starvla][INFO] - Step 1753, Loss: {'train/loss': 0.6476698517799377, 'train/grad_norm_pre_clip': 24.087322235107422, 'timing/data': 0.017426051199436188, 'timing/model': 0.9392541702836752, 'learning_rate/qwen_vl_interface': 8.7170391395302e-06, 'learning_rate/action_model': 8.7170391395302e-05, 'epoch': 0.8}) +[2026-06-17 19:31:31,986][starVLA.training.train_starvla][INFO] - Step 1754, Loss: {'train/loss': 0.562947690486908, 'train/grad_norm_pre_clip': 16.041484832763672, 'timing/data': 0.016608193516731262, 'timing/model': 1.1752027608454227, 'learning_rate/qwen_vl_interface': 8.715560471416223e-06, 'learning_rate/action_model': 8.715560471416224e-05, 'epoch': 0.8}) +[2026-06-17 19:31:39,780][starVLA.training.train_starvla][INFO] - Step 1755, Loss: {'train/loss': 0.5943331718444824, 'train/grad_norm_pre_clip': 13.836188316345215, 'timing/data': 0.024007716216146946, 'timing/model': 1.1494548451155424, 'learning_rate/qwen_vl_interface': 8.714081084887833e-06, 'learning_rate/action_model': 8.714081084887833e-05, 'epoch': 0.8}) +[2026-06-17 19:31:43,701][starVLA.training.train_starvla][INFO] - Step 1756, Loss: {'train/loss': 0.3538358807563782, 'train/grad_norm_pre_clip': 11.443962097167969, 'timing/data': 0.017716875299811363, 'timing/model': 1.108398549258709, 'learning_rate/qwen_vl_interface': 8.712600980251706e-06, 'learning_rate/action_model': 8.712600980251705e-05, 'epoch': 0.8}) +[2026-06-17 19:31:47,938][starVLA.training.train_starvla][INFO] - Step 1757, Loss: {'train/loss': 0.1801348179578781, 'train/grad_norm_pre_clip': 9.420065879821777, 'timing/data': 0.025213874876499176, 'timing/model': 1.0228325258940458, 'learning_rate/qwen_vl_interface': 8.71112015781467e-06, 'learning_rate/action_model': 8.71112015781467e-05, 'epoch': 0.8}) +[2026-06-17 19:31:51,949][starVLA.training.train_starvla][INFO] - Step 1758, Loss: {'train/loss': 0.5570087432861328, 'train/grad_norm_pre_clip': 20.81780242919922, 'timing/data': 0.02288422640413046, 'timing/model': 1.0041742557659745, 'learning_rate/qwen_vl_interface': 8.709638617883699e-06, 'learning_rate/action_model': 8.709638617883699e-05, 'epoch': 0.8}) +[2026-06-17 19:31:55,871][starVLA.training.train_starvla][INFO] - Step 1759, Loss: {'train/loss': 0.715053915977478, 'train/grad_norm_pre_clip': 15.73930835723877, 'timing/data': 0.017177571542561054, 'timing/model': 0.9614084484055638, 'learning_rate/qwen_vl_interface': 8.708156360765922e-06, 'learning_rate/action_model': 8.708156360765922e-05, 'epoch': 0.8}) +[2026-06-17 19:31:59,799][starVLA.training.train_starvla][INFO] - Step 1760, Loss: {'train/loss': 1.0797035694122314, 'train/grad_norm_pre_clip': 22.058958053588867, 'timing/data': 0.014929431490600109, 'timing/model': 0.9558219835162163, 'learning_rate/qwen_vl_interface': 8.706673386768609e-06, 'learning_rate/action_model': 8.706673386768609e-05, 'epoch': 0.8}) +[2026-06-17 19:32:03,931][starVLA.training.train_starvla][INFO] - Step 1761, Loss: {'train/loss': 0.6480212211608887, 'train/grad_norm_pre_clip': 13.784128189086914, 'timing/data': 0.01835337746888399, 'timing/model': 1.220817600376904, 'learning_rate/qwen_vl_interface': 8.705189696199184e-06, 'learning_rate/action_model': 8.705189696199183e-05, 'epoch': 0.8}) +[2026-06-17 19:32:07,872][starVLA.training.train_starvla][INFO] - Step 1762, Loss: {'train/loss': 0.45623838901519775, 'train/grad_norm_pre_clip': 16.886676788330078, 'timing/data': 0.019102315418422222, 'timing/model': 1.1534929517656565, 'learning_rate/qwen_vl_interface': 8.703705289365216e-06, 'learning_rate/action_model': 8.703705289365216e-05, 'epoch': 0.8}) +[2026-06-17 19:32:11,726][starVLA.training.train_starvla][INFO] - Step 1763, Loss: {'train/loss': 0.5321680903434753, 'train/grad_norm_pre_clip': 12.542940139770508, 'timing/data': 0.016136890277266502, 'timing/model': 0.9554788125678897, 'learning_rate/qwen_vl_interface': 8.702220166574426e-06, 'learning_rate/action_model': 8.702220166574425e-05, 'epoch': 0.8}) +[2026-06-17 19:32:15,806][starVLA.training.train_starvla][INFO] - Step 1764, Loss: {'train/loss': 0.6192342042922974, 'train/grad_norm_pre_clip': 13.241076469421387, 'timing/data': 0.019965064711868763, 'timing/model': 1.0230453265830874, 'learning_rate/qwen_vl_interface': 8.700734328134681e-06, 'learning_rate/action_model': 8.70073432813468e-05, 'epoch': 0.8}) +[2026-06-17 19:32:19,769][starVLA.training.train_starvla][INFO] - Step 1765, Loss: {'train/loss': 0.5507609844207764, 'train/grad_norm_pre_clip': 13.249366760253906, 'timing/data': 0.014965126290917397, 'timing/model': 0.9659456443041563, 'learning_rate/qwen_vl_interface': 8.699247774353997e-06, 'learning_rate/action_model': 8.699247774353996e-05, 'epoch': 0.8}) +[2026-06-17 19:32:23,727][starVLA.training.train_starvla][INFO] - Step 1766, Loss: {'train/loss': 0.5553311109542847, 'train/grad_norm_pre_clip': 16.512836456298828, 'timing/data': 0.017143884673714638, 'timing/model': 0.9681530669331551, 'learning_rate/qwen_vl_interface': 8.697760505540536e-06, 'learning_rate/action_model': 8.697760505540536e-05, 'epoch': 0.8}) +[2026-06-17 19:32:27,842][starVLA.training.train_starvla][INFO] - Step 1767, Loss: {'train/loss': 0.4874674081802368, 'train/grad_norm_pre_clip': 17.495092391967773, 'timing/data': 0.01877598464488983, 'timing/model': 1.0039047002792358, 'learning_rate/qwen_vl_interface': 8.696272522002614e-06, 'learning_rate/action_model': 8.696272522002615e-05, 'epoch': 0.8}) +[2026-06-17 19:32:32,089][starVLA.training.train_starvla][INFO] - Step 1768, Loss: {'train/loss': 0.6758621335029602, 'train/grad_norm_pre_clip': 18.006816864013672, 'timing/data': 0.019444514997303486, 'timing/model': 1.2450339002534747, 'learning_rate/qwen_vl_interface': 8.69478382404869e-06, 'learning_rate/action_model': 8.69478382404869e-05, 'epoch': 0.8}) +[2026-06-17 19:32:35,804][starVLA.training.train_starvla][INFO] - Step 1769, Loss: {'train/loss': 0.6161113977432251, 'train/grad_norm_pre_clip': 13.46755313873291, 'timing/data': 0.014598991721868515, 'timing/model': 0.9596160324290395, 'learning_rate/qwen_vl_interface': 8.693294411987371e-06, 'learning_rate/action_model': 8.693294411987371e-05, 'epoch': 0.8}) +[2026-06-17 19:32:39,771][starVLA.training.train_starvla][INFO] - Step 1770, Loss: {'train/loss': 0.5241342186927795, 'train/grad_norm_pre_clip': 17.54084014892578, 'timing/data': 0.021402121521532536, 'timing/model': 0.95325209479779, 'learning_rate/qwen_vl_interface': 8.691804286127416e-06, 'learning_rate/action_model': 8.691804286127415e-05, 'epoch': 0.8}) +[2026-06-17 19:32:43,757][starVLA.training.train_starvla][INFO] - Step 1771, Loss: {'train/loss': 0.5285483002662659, 'train/grad_norm_pre_clip': 12.182327270507812, 'timing/data': 0.016665118746459484, 'timing/model': 0.9587045647203922, 'learning_rate/qwen_vl_interface': 8.690313446777729e-06, 'learning_rate/action_model': 8.690313446777728e-05, 'epoch': 0.8}) +[2026-06-17 19:32:47,858][starVLA.training.train_starvla][INFO] - Step 1772, Loss: {'train/loss': 0.40475913882255554, 'train/grad_norm_pre_clip': 8.587983131408691, 'timing/data': 0.019783210940659046, 'timing/model': 0.9681737711653113, 'learning_rate/qwen_vl_interface': 8.688821894247363e-06, 'learning_rate/action_model': 8.688821894247362e-05, 'epoch': 0.8}) +[2026-06-17 19:32:51,911][starVLA.training.train_starvla][INFO] - Step 1773, Loss: {'train/loss': 0.6554169058799744, 'train/grad_norm_pre_clip': 11.249154090881348, 'timing/data': 0.015106488950550556, 'timing/model': 0.9774528155103326, 'learning_rate/qwen_vl_interface': 8.687329628845517e-06, 'learning_rate/action_model': 8.687329628845516e-05, 'epoch': 0.8}) +[2026-06-17 19:32:55,804][starVLA.training.train_starvla][INFO] - Step 1774, Loss: {'train/loss': 0.4431188404560089, 'train/grad_norm_pre_clip': 11.908595085144043, 'timing/data': 0.01992290373891592, 'timing/model': 1.1513331327587366, 'learning_rate/qwen_vl_interface': 8.68583665088154e-06, 'learning_rate/action_model': 8.68583665088154e-05, 'epoch': 0.8}) +[2026-06-17 19:33:03,413][starVLA.training.train_starvla][INFO] - Step 1775, Loss: {'train/loss': 0.8854631185531616, 'train/grad_norm_pre_clip': 15.628212928771973, 'timing/data': 0.016289085149765015, 'timing/model': 2.1832125280052423, 'learning_rate/qwen_vl_interface': 8.684342960664928e-06, 'learning_rate/action_model': 8.684342960664928e-05, 'epoch': 0.81}) +[2026-06-17 19:33:12,510][starVLA.training.train_starvla][INFO] - Step 1776, Loss: {'train/loss': 0.628178596496582, 'train/grad_norm_pre_clip': 22.54464340209961, 'timing/data': 0.01794765144586563, 'timing/model': 3.0185319362208247, 'learning_rate/qwen_vl_interface': 8.682848558505327e-06, 'learning_rate/action_model': 8.682848558505326e-05, 'epoch': 0.81}) +[2026-06-17 19:33:16,493][starVLA.training.train_starvla][INFO] - Step 1777, Loss: {'train/loss': 0.4567422568798065, 'train/grad_norm_pre_clip': 11.268007278442383, 'timing/data': 0.01764621026813984, 'timing/model': 1.0256028827279806, 'learning_rate/qwen_vl_interface': 8.681353444712525e-06, 'learning_rate/action_model': 8.681353444712524e-05, 'epoch': 0.81}) +[2026-06-17 19:33:20,477][starVLA.training.train_starvla][INFO] - Step 1778, Loss: {'train/loss': 0.7786363363265991, 'train/grad_norm_pre_clip': 17.146203994750977, 'timing/data': 0.017935623414814472, 'timing/model': 0.9611354554072022, 'learning_rate/qwen_vl_interface': 8.67985761959646e-06, 'learning_rate/action_model': 8.67985761959646e-05, 'epoch': 0.81}) +[2026-06-17 19:33:25,300][starVLA.training.train_starvla][INFO] - Step 1779, Loss: {'train/loss': 0.6082913279533386, 'train/grad_norm_pre_clip': 12.893696784973145, 'timing/data': 0.01793033257126808, 'timing/model': 1.7014144817367196, 'learning_rate/qwen_vl_interface': 8.678361083467221e-06, 'learning_rate/action_model': 8.67836108346722e-05, 'epoch': 0.81}) +[2026-06-17 19:33:29,348][starVLA.training.train_starvla][INFO] - Step 1780, Loss: {'train/loss': 0.5787237882614136, 'train/grad_norm_pre_clip': 13.595258712768555, 'timing/data': 0.021882987581193447, 'timing/model': 0.9468206884339452, 'learning_rate/qwen_vl_interface': 8.67686383663504e-06, 'learning_rate/action_model': 8.67686383663504e-05, 'epoch': 0.81}) +[2026-06-17 19:33:33,512][starVLA.training.train_starvla][INFO] - Step 1781, Loss: {'train/loss': 0.45833101868629456, 'train/grad_norm_pre_clip': 9.331507682800293, 'timing/data': 0.019395368173718452, 'timing/model': 1.2327403500676155, 'learning_rate/qwen_vl_interface': 8.6753658794103e-06, 'learning_rate/action_model': 8.675365879410298e-05, 'epoch': 0.81}) +[2026-06-17 19:33:37,543][starVLA.training.train_starvla][INFO] - Step 1782, Loss: {'train/loss': 0.7156612277030945, 'train/grad_norm_pre_clip': 13.69554615020752, 'timing/data': 0.015620985999703407, 'timing/model': 1.2659373814240098, 'learning_rate/qwen_vl_interface': 8.673867212103525e-06, 'learning_rate/action_model': 8.673867212103524e-05, 'epoch': 0.81}) +[2026-06-17 19:33:41,242][starVLA.training.train_starvla][INFO] - Step 1783, Loss: {'train/loss': 0.5811876654624939, 'train/grad_norm_pre_clip': 12.326804161071777, 'timing/data': 0.02209315448999405, 'timing/model': 0.9495256245136261, 'learning_rate/qwen_vl_interface': 8.672367835025395e-06, 'learning_rate/action_model': 8.672367835025393e-05, 'epoch': 0.81}) +[2026-06-17 19:33:45,126][starVLA.training.train_starvla][INFO] - Step 1784, Loss: {'train/loss': 0.8231874108314514, 'train/grad_norm_pre_clip': 14.467794418334961, 'timing/data': 0.017133121378719807, 'timing/model': 0.9537879060953856, 'learning_rate/qwen_vl_interface': 8.670867748486727e-06, 'learning_rate/action_model': 8.670867748486726e-05, 'epoch': 0.81}) +[2026-06-17 19:33:49,069][starVLA.training.train_starvla][INFO] - Step 1785, Loss: {'train/loss': 0.615861713886261, 'train/grad_norm_pre_clip': 12.686483383178711, 'timing/data': 0.01591577846556902, 'timing/model': 0.9666490266099572, 'learning_rate/qwen_vl_interface': 8.669366952798494e-06, 'learning_rate/action_model': 8.669366952798494e-05, 'epoch': 0.81}) +[2026-06-17 19:33:53,106][starVLA.training.train_starvla][INFO] - Step 1786, Loss: {'train/loss': 0.5088584423065186, 'train/grad_norm_pre_clip': 9.942249298095703, 'timing/data': 0.015045476146042347, 'timing/model': 1.0269031375646591, 'learning_rate/qwen_vl_interface': 8.667865448271814e-06, 'learning_rate/action_model': 8.667865448271814e-05, 'epoch': 0.81}) +[2026-06-17 19:33:56,991][starVLA.training.train_starvla][INFO] - Step 1787, Loss: {'train/loss': 0.4456138014793396, 'train/grad_norm_pre_clip': 9.195240020751953, 'timing/data': 0.01652657613158226, 'timing/model': 0.9837471945211291, 'learning_rate/qwen_vl_interface': 8.666363235217949e-06, 'learning_rate/action_model': 8.66636323521795e-05, 'epoch': 0.81}) +[2026-06-17 19:34:00,964][starVLA.training.train_starvla][INFO] - Step 1788, Loss: {'train/loss': 0.5260900855064392, 'train/grad_norm_pre_clip': 9.67344856262207, 'timing/data': 0.017566632479429245, 'timing/model': 1.1225452600046992, 'learning_rate/qwen_vl_interface': 8.66486031394831e-06, 'learning_rate/action_model': 8.664860313948309e-05, 'epoch': 0.81}) +[2026-06-17 19:34:04,757][starVLA.training.train_starvla][INFO] - Step 1789, Loss: {'train/loss': 0.5985718369483948, 'train/grad_norm_pre_clip': 10.432036399841309, 'timing/data': 0.021052406169474125, 'timing/model': 0.9620470963418484, 'learning_rate/qwen_vl_interface': 8.66335668477445e-06, 'learning_rate/action_model': 8.66335668477445e-05, 'epoch': 0.81}) +[2026-06-17 19:34:08,763][starVLA.training.train_starvla][INFO] - Step 1790, Loss: {'train/loss': 0.7580090165138245, 'train/grad_norm_pre_clip': 12.882495880126953, 'timing/data': 0.01514718122780323, 'timing/model': 0.95202702190727, 'learning_rate/qwen_vl_interface': 8.661852348008079e-06, 'learning_rate/action_model': 8.661852348008078e-05, 'epoch': 0.81}) +[2026-06-17 19:34:12,859][starVLA.training.train_starvla][INFO] - Step 1791, Loss: {'train/loss': 0.43429553508758545, 'train/grad_norm_pre_clip': 10.227407455444336, 'timing/data': 0.01751650869846344, 'timing/model': 0.9456104086712003, 'learning_rate/qwen_vl_interface': 8.660347303961044e-06, 'learning_rate/action_model': 8.660347303961044e-05, 'epoch': 0.81}) +[2026-06-17 19:34:16,804][starVLA.training.train_starvla][INFO] - Step 1792, Loss: {'train/loss': 0.8131080865859985, 'train/grad_norm_pre_clip': 13.877609252929688, 'timing/data': 0.016861782409250736, 'timing/model': 1.0031386371701956, 'learning_rate/qwen_vl_interface': 8.658841552945342e-06, 'learning_rate/action_model': 8.658841552945342e-05, 'epoch': 0.81}) +[2026-06-17 19:34:20,634][starVLA.training.train_starvla][INFO] - Step 1793, Loss: {'train/loss': 0.5282003879547119, 'train/grad_norm_pre_clip': 9.435281753540039, 'timing/data': 0.021320728585124016, 'timing/model': 0.9341070409864187, 'learning_rate/qwen_vl_interface': 8.65733509527312e-06, 'learning_rate/action_model': 8.65733509527312e-05, 'epoch': 0.81}) +[2026-06-17 19:34:24,587][starVLA.training.train_starvla][INFO] - Step 1794, Loss: {'train/loss': 0.6434544324874878, 'train/grad_norm_pre_clip': 13.699750900268555, 'timing/data': 0.018472720868885517, 'timing/model': 0.9402584750205278, 'learning_rate/qwen_vl_interface': 8.655827931256667e-06, 'learning_rate/action_model': 8.655827931256667e-05, 'epoch': 0.81}) +[2026-06-17 19:34:30,844][starVLA.training.train_starvla][INFO] - Step 1795, Loss: {'train/loss': 0.27773505449295044, 'train/grad_norm_pre_clip': 7.014104843139648, 'timing/data': 0.016572603955864906, 'timing/model': 1.12231470271945, 'learning_rate/qwen_vl_interface': 8.654320061208418e-06, 'learning_rate/action_model': 8.654320061208417e-05, 'epoch': 0.81}) +[2026-06-17 19:34:34,541][starVLA.training.train_starvla][INFO] - Step 1796, Loss: {'train/loss': 0.41264796257019043, 'train/grad_norm_pre_clip': 8.861059188842773, 'timing/data': 0.020961899310350418, 'timing/model': 0.968859588727355, 'learning_rate/qwen_vl_interface': 8.652811485440956e-06, 'learning_rate/action_model': 8.652811485440956e-05, 'epoch': 0.81}) +[2026-06-17 19:34:38,443][starVLA.training.train_starvla][INFO] - Step 1797, Loss: {'train/loss': 0.4342549741268158, 'train/grad_norm_pre_clip': 9.949051856994629, 'timing/data': 0.02418831828981638, 'timing/model': 0.9532005423679948, 'learning_rate/qwen_vl_interface': 8.651302204267013e-06, 'learning_rate/action_model': 8.651302204267013e-05, 'epoch': 0.82}) +[2026-06-17 19:34:42,391][starVLA.training.train_starvla][INFO] - Step 1798, Loss: {'train/loss': 0.34099751710891724, 'train/grad_norm_pre_clip': 8.431924819946289, 'timing/data': 0.02454777155071497, 'timing/model': 1.0096374340355396, 'learning_rate/qwen_vl_interface': 8.649792217999461e-06, 'learning_rate/action_model': 8.649792217999462e-05, 'epoch': 0.82}) +[2026-06-17 19:34:46,234][starVLA.training.train_starvla][INFO] - Step 1799, Loss: {'train/loss': 0.5698236227035522, 'train/grad_norm_pre_clip': 11.781293869018555, 'timing/data': 0.030074832029640675, 'timing/model': 0.961247431114316, 'learning_rate/qwen_vl_interface': 8.648281526951327e-06, 'learning_rate/action_model': 8.648281526951327e-05, 'epoch': 0.82}) +[2026-06-17 19:37:25,212][starVLA.training.train_starvla][INFO] - Step 1800, Loss: {'train/loss': 0.29639750719070435, 'train/grad_norm_pre_clip': 8.742448806762695, 'eval/loss': 0.4793308675289154, 'eval/latency_8/loss': 0.4793308675289154, 'eval/demon_fire_precision': 0.9078575466365179, 'eval/demon_fire_recall': 0.997825411618515, 'eval/demon_fire_f1': 0.9507177741601303, 'eval/demon_fire_rate_model': 0.9844184752365053, 'eval/demon_fire_rate_teacher': 0.8956594323873122, 'eval/demon_left_precision': 0.7699386503067485, 'eval/demon_left_recall': 0.5374732334047109, 'eval/demon_left_f1': 0.6330390920554855, 'eval/demon_left_rate_model': 0.09070673344462994, 'eval/demon_left_rate_teacher': 0.12993878686700056, 'eval/demon_right_precision': 0.7220588235294118, 'eval/demon_right_recall': 0.7439393939393939, 'eval/demon_right_f1': 0.7328358208955223, 'eval/demon_right_rate_model': 0.18920422927100725, 'eval/demon_right_rate_teacher': 0.18363939899833054, 'eval/demon_move_f1': 0.682937456475504, 'eval/demon_cc_f1': 0.8506550217888345, 'eval/latency_8/demon_fire_precision': 0.9078575466365179, 'eval/latency_8/demon_fire_recall': 0.997825411618515, 'eval/latency_8/demon_fire_f1': 0.9507177741601303, 'eval/latency_8/demon_fire_rate_model': 0.9844184752365053, 'eval/latency_8/demon_fire_rate_teacher': 0.8956594323873122, 'eval/latency_8/demon_left_precision': 0.7699386503067485, 'eval/latency_8/demon_left_recall': 0.5374732334047109, 'eval/latency_8/demon_left_f1': 0.6330390920554855, 'eval/latency_8/demon_left_rate_model': 0.09070673344462994, 'eval/latency_8/demon_left_rate_teacher': 0.12993878686700056, 'eval/latency_8/demon_right_precision': 0.7220588235294118, 'eval/latency_8/demon_right_recall': 0.7439393939393939, 'eval/latency_8/demon_right_f1': 0.7328358208955223, 'eval/latency_8/demon_right_rate_model': 0.18920422927100725, 'eval/latency_8/demon_right_rate_teacher': 0.18363939899833054, 'eval/latency_8/demon_move_f1': 0.682937456475504, 'eval/latency_8/demon_cc_f1': 0.8506550217888345, 'timing/data': 0.016767947003245354, 'timing/model': 0.950977542437613, 'learning_rate/qwen_vl_interface': 8.646770131435774e-06, 'learning_rate/action_model': 8.646770131435773e-05, 'epoch': 0.82}) +[2026-06-17 19:37:25,217][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_1800_state +[2026-06-17 19:37:25,222][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 19:37:33,752][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_1800_state/model.safetensors +[2026-06-17 19:37:51,882][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_1800_state/optimizer.bin +[2026-06-17 19:37:51,883][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_1800_state/sampler.bin +[2026-06-17 19:37:51,883][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_1800_state/sampler_1.bin +[2026-06-17 19:37:51,885][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_1800_state/random_states_0.pkl +[2026-06-17 19:37:51,890][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1800 +[2026-06-17 19:37:51,890][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 19:37:51,906][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 19:37:55,855][starVLA.training.train_starvla][INFO] - Step 1801, Loss: {'train/loss': 0.464942991733551, 'train/grad_norm_pre_clip': 11.223834991455078, 'timing/data': 0.019975120201706886, 'timing/model': 0.9619147395715117, 'learning_rate/qwen_vl_interface': 8.645258031766116e-06, 'learning_rate/action_model': 8.645258031766116e-05, 'epoch': 0.82}) +[2026-06-17 19:37:59,807][starVLA.training.train_starvla][INFO] - Step 1802, Loss: {'train/loss': 0.47031229734420776, 'train/grad_norm_pre_clip': 12.586195945739746, 'timing/data': 0.016327060759067535, 'timing/model': 1.034796110354364, 'learning_rate/qwen_vl_interface': 8.643745228255816e-06, 'learning_rate/action_model': 8.643745228255817e-05, 'epoch': 0.82}) +[2026-06-17 19:38:03,846][starVLA.training.train_starvla][INFO] - Step 1803, Loss: {'train/loss': 0.44349586963653564, 'train/grad_norm_pre_clip': 8.97705364227295, 'timing/data': 0.0155949043110013, 'timing/model': 1.2420523213222623, 'learning_rate/qwen_vl_interface': 8.64223172121848e-06, 'learning_rate/action_model': 8.642231721218479e-05, 'epoch': 0.82}) +[2026-06-17 19:38:07,666][starVLA.training.train_starvla][INFO] - Step 1804, Loss: {'train/loss': 0.673679769039154, 'train/grad_norm_pre_clip': 14.731324195861816, 'timing/data': 0.017677180469036102, 'timing/model': 0.9663198608905077, 'learning_rate/qwen_vl_interface': 8.640717510967857e-06, 'learning_rate/action_model': 8.640717510967856e-05, 'epoch': 0.82}) +[2026-06-17 19:38:11,702][starVLA.training.train_starvla][INFO] - Step 1805, Loss: {'train/loss': 0.48634907603263855, 'train/grad_norm_pre_clip': 12.574528694152832, 'timing/data': 0.01669382583349943, 'timing/model': 0.9556366568431258, 'learning_rate/qwen_vl_interface': 8.639202597817844e-06, 'learning_rate/action_model': 8.639202597817844e-05, 'epoch': 0.82}) +[2026-06-17 19:38:15,814][starVLA.training.train_starvla][INFO] - Step 1806, Loss: {'train/loss': 0.4562625288963318, 'train/grad_norm_pre_clip': 12.148768424987793, 'timing/data': 0.021715098060667515, 'timing/model': 0.9605988748371601, 'learning_rate/qwen_vl_interface': 8.637686982082487e-06, 'learning_rate/action_model': 8.637686982082487e-05, 'epoch': 0.82}) +[2026-06-17 19:38:19,906][starVLA.training.train_starvla][INFO] - Step 1807, Loss: {'train/loss': 0.6205843687057495, 'train/grad_norm_pre_clip': 13.766467094421387, 'timing/data': 0.01789653953164816, 'timing/model': 0.9605897990986705, 'learning_rate/qwen_vl_interface': 8.636170664075972e-06, 'learning_rate/action_model': 8.636170664075972e-05, 'epoch': 0.82}) +[2026-06-17 19:38:24,122][starVLA.training.train_starvla][INFO] - Step 1808, Loss: {'train/loss': 0.4363933801651001, 'train/grad_norm_pre_clip': 9.499794006347656, 'timing/data': 0.02099280245602131, 'timing/model': 1.046459723263979, 'learning_rate/qwen_vl_interface': 8.634653644112636e-06, 'learning_rate/action_model': 8.634653644112635e-05, 'epoch': 0.82}) +[2026-06-17 19:38:28,205][starVLA.training.train_starvla][INFO] - Step 1809, Loss: {'train/loss': 0.6017752885818481, 'train/grad_norm_pre_clip': 17.787874221801758, 'timing/data': 0.01517021656036377, 'timing/model': 1.1251143282279372, 'learning_rate/qwen_vl_interface': 8.633135922506957e-06, 'learning_rate/action_model': 8.633135922506957e-05, 'epoch': 0.82}) +[2026-06-17 19:38:32,339][starVLA.training.train_starvla][INFO] - Step 1810, Loss: {'train/loss': 0.7032851576805115, 'train/grad_norm_pre_clip': 12.186564445495605, 'timing/data': 0.016914579086005688, 'timing/model': 1.1946245124563575, 'learning_rate/qwen_vl_interface': 8.63161749957356e-06, 'learning_rate/action_model': 8.63161749957356e-05, 'epoch': 0.82}) +[2026-06-17 19:38:36,054][starVLA.training.train_starvla][INFO] - Step 1811, Loss: {'train/loss': 0.4292762875556946, 'train/grad_norm_pre_clip': 11.570817947387695, 'timing/data': 0.016223998740315437, 'timing/model': 0.9914480708539486, 'learning_rate/qwen_vl_interface': 8.630098375627216e-06, 'learning_rate/action_model': 8.630098375627217e-05, 'epoch': 0.82}) +[2026-06-17 19:38:40,701][starVLA.training.train_starvla][INFO] - Step 1812, Loss: {'train/loss': 0.525100588798523, 'train/grad_norm_pre_clip': 13.551593780517578, 'timing/data': 0.01633928343653679, 'timing/model': 0.9813811704516411, 'learning_rate/qwen_vl_interface': 8.628578550982846e-06, 'learning_rate/action_model': 8.628578550982845e-05, 'epoch': 0.82}) +[2026-06-17 19:38:46,491][starVLA.training.train_starvla][INFO] - Step 1813, Loss: {'train/loss': 0.4036763608455658, 'train/grad_norm_pre_clip': 10.578851699829102, 'timing/data': 0.01728702336549759, 'timing/model': 1.1319520073011518, 'learning_rate/qwen_vl_interface': 8.627058025955504e-06, 'learning_rate/action_model': 8.627058025955503e-05, 'epoch': 0.82}) +[2026-06-17 19:38:50,577][starVLA.training.train_starvla][INFO] - Step 1814, Loss: {'train/loss': 0.36799949407577515, 'train/grad_norm_pre_clip': 9.9661226272583, 'timing/data': 0.021349508315324783, 'timing/model': 1.034488215111196, 'learning_rate/qwen_vl_interface': 8.6255368008604e-06, 'learning_rate/action_model': 8.625536800860398e-05, 'epoch': 0.82}) +[2026-06-17 19:38:55,076][starVLA.training.train_starvla][INFO] - Step 1815, Loss: {'train/loss': 0.49147894978523254, 'train/grad_norm_pre_clip': 12.08566951751709, 'timing/data': 0.018357116729021072, 'timing/model': 0.9480033256113529, 'learning_rate/qwen_vl_interface': 8.624014876012883e-06, 'learning_rate/action_model': 8.624014876012882e-05, 'epoch': 0.82}) +[2026-06-17 19:39:00,119][starVLA.training.train_starvla][INFO] - Step 1816, Loss: {'train/loss': 0.36230406165122986, 'train/grad_norm_pre_clip': 9.521923065185547, 'timing/data': 0.022988660261034966, 'timing/model': 1.2332838587462902, 'learning_rate/qwen_vl_interface': 8.622492251728453e-06, 'learning_rate/action_model': 8.622492251728452e-05, 'epoch': 0.82}) +[2026-06-17 19:39:04,017][starVLA.training.train_starvla][INFO] - Step 1817, Loss: {'train/loss': 0.6637018322944641, 'train/grad_norm_pre_clip': 14.444040298461914, 'timing/data': 0.023143800906836987, 'timing/model': 1.175719971768558, 'learning_rate/qwen_vl_interface': 8.62096892832275e-06, 'learning_rate/action_model': 8.620968928322749e-05, 'epoch': 0.82}) +[2026-06-17 19:39:07,824][starVLA.training.train_starvla][INFO] - Step 1818, Loss: {'train/loss': 0.8939952254295349, 'train/grad_norm_pre_clip': 19.68836212158203, 'timing/data': 0.016092448495328426, 'timing/model': 1.080290149897337, 'learning_rate/qwen_vl_interface': 8.619444906111562e-06, 'learning_rate/action_model': 8.619444906111561e-05, 'epoch': 0.82}) +[2026-06-17 19:39:12,220][starVLA.training.train_starvla][INFO] - Step 1819, Loss: {'train/loss': 0.6018722057342529, 'train/grad_norm_pre_clip': 14.04946517944336, 'timing/data': 0.01809620577841997, 'timing/model': 1.4152216017246246, 'learning_rate/qwen_vl_interface': 8.617920185410817e-06, 'learning_rate/action_model': 8.617920185410817e-05, 'epoch': 0.83}) +[2026-06-17 19:39:16,560][starVLA.training.train_starvla][INFO] - Step 1820, Loss: {'train/loss': 0.36547520756721497, 'train/grad_norm_pre_clip': 10.75198745727539, 'timing/data': 0.025640465319156647, 'timing/model': 0.9390239426866174, 'learning_rate/qwen_vl_interface': 8.616394766536594e-06, 'learning_rate/action_model': 8.616394766536593e-05, 'epoch': 0.83}) +[2026-06-17 19:39:20,697][starVLA.training.train_starvla][INFO] - Step 1821, Loss: {'train/loss': 0.5121593475341797, 'train/grad_norm_pre_clip': 13.547622680664062, 'timing/data': 0.01795478630810976, 'timing/model': 0.9536820705980062, 'learning_rate/qwen_vl_interface': 8.614868649805115e-06, 'learning_rate/action_model': 8.614868649805114e-05, 'epoch': 0.83}) +[2026-06-17 19:39:24,558][starVLA.training.train_starvla][INFO] - Step 1822, Loss: {'train/loss': 0.6571230888366699, 'train/grad_norm_pre_clip': 12.644962310791016, 'timing/data': 0.0184863293543458, 'timing/model': 0.9517686497420073, 'learning_rate/qwen_vl_interface': 8.613341835532743e-06, 'learning_rate/action_model': 8.613341835532743e-05, 'epoch': 0.83}) +[2026-06-17 19:39:28,714][starVLA.training.train_starvla][INFO] - Step 1823, Loss: {'train/loss': 0.5288223624229431, 'train/grad_norm_pre_clip': 13.899255752563477, 'timing/data': 0.026395654305815697, 'timing/model': 1.2080258652567863, 'learning_rate/qwen_vl_interface': 8.611814324035989e-06, 'learning_rate/action_model': 8.61181432403599e-05, 'epoch': 0.83}) +[2026-06-17 19:39:32,620][starVLA.training.train_starvla][INFO] - Step 1824, Loss: {'train/loss': 0.47496816515922546, 'train/grad_norm_pre_clip': 11.746537208557129, 'timing/data': 0.01635032333433628, 'timing/model': 1.1522511690855026, 'learning_rate/qwen_vl_interface': 8.610286115631508e-06, 'learning_rate/action_model': 8.610286115631509e-05, 'epoch': 0.83}) +[2026-06-17 19:39:36,461][starVLA.training.train_starvla][INFO] - Step 1825, Loss: {'train/loss': 0.5064708590507507, 'train/grad_norm_pre_clip': 13.569694519042969, 'timing/data': 0.020882537588477135, 'timing/model': 0.9459912972524762, 'learning_rate/qwen_vl_interface': 8.608757210636101e-06, 'learning_rate/action_model': 8.608757210636101e-05, 'epoch': 0.83}) +[2026-06-17 19:39:41,944][starVLA.training.train_starvla][INFO] - Step 1826, Loss: {'train/loss': 0.49532097578048706, 'train/grad_norm_pre_clip': 12.775235176086426, 'timing/data': 0.04929273575544357, 'timing/model': 2.2426703851670027, 'learning_rate/qwen_vl_interface': 8.60722760936671e-06, 'learning_rate/action_model': 8.60722760936671e-05, 'epoch': 0.83}) +[2026-06-17 19:39:47,268][starVLA.training.train_starvla][INFO] - Step 1827, Loss: {'train/loss': 0.5711959600448608, 'train/grad_norm_pre_clip': 86.35070037841797, 'timing/data': 0.02060140948742628, 'timing/model': 0.9665602697059512, 'learning_rate/qwen_vl_interface': 8.605697312140421e-06, 'learning_rate/action_model': 8.605697312140422e-05, 'epoch': 0.83}) +[2026-06-17 19:39:51,762][starVLA.training.train_starvla][INFO] - Step 1828, Loss: {'train/loss': 0.2101280391216278, 'train/grad_norm_pre_clip': 8.047981262207031, 'timing/data': 0.01968629565089941, 'timing/model': 1.2169101582840085, 'learning_rate/qwen_vl_interface': 8.604166319274472e-06, 'learning_rate/action_model': 8.604166319274471e-05, 'epoch': 0.83}) +[2026-06-17 19:39:55,957][starVLA.training.train_starvla][INFO] - Step 1829, Loss: {'train/loss': 0.42851680517196655, 'train/grad_norm_pre_clip': 11.532029151916504, 'timing/data': 0.01811841130256653, 'timing/model': 0.9615855738520622, 'learning_rate/qwen_vl_interface': 8.602634631086233e-06, 'learning_rate/action_model': 8.602634631086233e-05, 'epoch': 0.83}) +[2026-06-17 19:40:00,128][starVLA.training.train_starvla][INFO] - Step 1830, Loss: {'train/loss': 0.41088956594467163, 'train/grad_norm_pre_clip': 9.949432373046875, 'timing/data': 0.01714202854782343, 'timing/model': 1.1738692773506045, 'learning_rate/qwen_vl_interface': 8.601102247893227e-06, 'learning_rate/action_model': 8.601102247893226e-05, 'epoch': 0.83}) +[2026-06-17 19:40:04,079][starVLA.training.train_starvla][INFO] - Step 1831, Loss: {'train/loss': 0.31921103596687317, 'train/grad_norm_pre_clip': 11.844959259033203, 'timing/data': 0.01739359088242054, 'timing/model': 0.9589054426178336, 'learning_rate/qwen_vl_interface': 8.599569170013121e-06, 'learning_rate/action_model': 8.59956917001312e-05, 'epoch': 0.83}) +[2026-06-17 19:40:08,287][starVLA.training.train_starvla][INFO] - Step 1832, Loss: {'train/loss': 0.3402961790561676, 'train/grad_norm_pre_clip': 16.30607032775879, 'timing/data': 0.01808985508978367, 'timing/model': 0.9450571499764919, 'learning_rate/qwen_vl_interface': 8.59803539776372e-06, 'learning_rate/action_model': 8.59803539776372e-05, 'epoch': 0.83}) +[2026-06-17 19:40:12,267][starVLA.training.train_starvla][INFO] - Step 1833, Loss: {'train/loss': 0.6345818042755127, 'train/grad_norm_pre_clip': 15.249022483825684, 'timing/data': 0.017556359991431236, 'timing/model': 0.9761773347854614, 'learning_rate/qwen_vl_interface': 8.596500931462979e-06, 'learning_rate/action_model': 8.596500931462979e-05, 'epoch': 0.83}) +[2026-06-17 19:40:16,347][starVLA.training.train_starvla][INFO] - Step 1834, Loss: {'train/loss': 0.6408151984214783, 'train/grad_norm_pre_clip': 17.487489700317383, 'timing/data': 0.018995029851794243, 'timing/model': 0.9904775442555547, 'learning_rate/qwen_vl_interface': 8.594965771428993e-06, 'learning_rate/action_model': 8.594965771428993e-05, 'epoch': 0.83}) +[2026-06-17 19:40:20,221][starVLA.training.train_starvla][INFO] - Step 1835, Loss: {'train/loss': 0.49779850244522095, 'train/grad_norm_pre_clip': 11.17201042175293, 'timing/data': 0.02161539439111948, 'timing/model': 0.9472177755087614, 'learning_rate/qwen_vl_interface': 8.593429917980006e-06, 'learning_rate/action_model': 8.593429917980006e-05, 'epoch': 0.83}) +[2026-06-17 19:40:24,288][starVLA.training.train_starvla][INFO] - Step 1836, Loss: {'train/loss': 0.6143172979354858, 'train/grad_norm_pre_clip': 21.031444549560547, 'timing/data': 0.019543781876564026, 'timing/model': 1.17862156778574, 'learning_rate/qwen_vl_interface': 8.591893371434397e-06, 'learning_rate/action_model': 8.591893371434396e-05, 'epoch': 0.83}) +[2026-06-17 19:40:28,313][starVLA.training.train_starvla][INFO] - Step 1837, Loss: {'train/loss': 0.4466857314109802, 'train/grad_norm_pre_clip': 12.709916114807129, 'timing/data': 0.017444211058318615, 'timing/model': 1.1872209617868066, 'learning_rate/qwen_vl_interface': 8.590356132110695e-06, 'learning_rate/action_model': 8.590356132110695e-05, 'epoch': 0.83}) +[2026-06-17 19:40:31,997][starVLA.training.train_starvla][INFO] - Step 1838, Loss: {'train/loss': 0.2976016402244568, 'train/grad_norm_pre_clip': 10.305893898010254, 'timing/data': 0.025679239071905613, 'timing/model': 0.9622978558763862, 'learning_rate/qwen_vl_interface': 8.588818200327575e-06, 'learning_rate/action_model': 8.588818200327575e-05, 'epoch': 0.83}) +[2026-06-17 19:40:35,981][starVLA.training.train_starvla][INFO] - Step 1839, Loss: {'train/loss': 0.3337966203689575, 'train/grad_norm_pre_clip': 14.853381156921387, 'timing/data': 0.018489323556423187, 'timing/model': 0.972018962725997, 'learning_rate/qwen_vl_interface': 8.587279576403848e-06, 'learning_rate/action_model': 8.587279576403847e-05, 'epoch': 0.83}) +[2026-06-17 19:40:40,105][starVLA.training.train_starvla][INFO] - Step 1840, Loss: {'train/loss': 0.769778847694397, 'train/grad_norm_pre_clip': 25.565732955932617, 'timing/data': 0.016430933959782124, 'timing/model': 0.9708422580733895, 'learning_rate/qwen_vl_interface': 8.585740260658475e-06, 'learning_rate/action_model': 8.585740260658475e-05, 'epoch': 0.83}) +[2026-06-17 19:40:44,004][starVLA.training.train_starvla][INFO] - Step 1841, Loss: {'train/loss': 0.4886800944805145, 'train/grad_norm_pre_clip': 12.42597484588623, 'timing/data': 0.017620956525206566, 'timing/model': 0.972862733528018, 'learning_rate/qwen_vl_interface': 8.584200253410557e-06, 'learning_rate/action_model': 8.584200253410556e-05, 'epoch': 0.84}) +[2026-06-17 19:40:47,976][starVLA.training.train_starvla][INFO] - Step 1842, Loss: {'train/loss': 0.5747369527816772, 'train/grad_norm_pre_clip': 13.052863121032715, 'timing/data': 0.016995015554130077, 'timing/model': 0.9434899166226387, 'learning_rate/qwen_vl_interface': 8.582659554979336e-06, 'learning_rate/action_model': 8.582659554979336e-05, 'epoch': 0.84}) +[2026-06-17 19:40:51,907][starVLA.training.train_starvla][INFO] - Step 1843, Loss: {'train/loss': 0.5944592952728271, 'train/grad_norm_pre_clip': 13.159497261047363, 'timing/data': 0.021436434239149094, 'timing/model': 1.1695447247475386, 'learning_rate/qwen_vl_interface': 8.581118165684204e-06, 'learning_rate/action_model': 8.581118165684204e-05, 'epoch': 0.84}) +[2026-06-17 19:40:56,001][starVLA.training.train_starvla][INFO] - Step 1844, Loss: {'train/loss': 0.9860829710960388, 'train/grad_norm_pre_clip': 21.341068267822266, 'timing/data': 0.02611362375319004, 'timing/model': 1.1809524139389396, 'learning_rate/qwen_vl_interface': 8.579576085844695e-06, 'learning_rate/action_model': 8.579576085844694e-05, 'epoch': 0.84}) +[2026-06-17 19:41:01,758][starVLA.training.train_starvla][INFO] - Step 1845, Loss: {'train/loss': 0.6667736172676086, 'train/grad_norm_pre_clip': 13.647018432617188, 'timing/data': 0.01443133782595396, 'timing/model': 0.9686998659744859, 'learning_rate/qwen_vl_interface': 8.578033315780474e-06, 'learning_rate/action_model': 8.578033315780475e-05, 'epoch': 0.84}) +[2026-06-17 19:41:05,797][starVLA.training.train_starvla][INFO] - Step 1846, Loss: {'train/loss': 0.4140014052391052, 'train/grad_norm_pre_clip': 18.119647979736328, 'timing/data': 0.019195619970560074, 'timing/model': 0.9589252267032862, 'learning_rate/qwen_vl_interface': 8.576489855811371e-06, 'learning_rate/action_model': 8.57648985581137e-05, 'epoch': 0.84}) +[2026-06-17 19:41:09,957][starVLA.training.train_starvla][INFO] - Step 1847, Loss: {'train/loss': 0.5149725675582886, 'train/grad_norm_pre_clip': 14.918567657470703, 'timing/data': 0.020094220526516438, 'timing/model': 0.9609399540349841, 'learning_rate/qwen_vl_interface': 8.57494570625734e-06, 'learning_rate/action_model': 8.57494570625734e-05, 'epoch': 0.84}) +[2026-06-17 19:41:13,933][starVLA.training.train_starvla][INFO] - Step 1848, Loss: {'train/loss': 0.4109797477722168, 'train/grad_norm_pre_clip': 10.121233940124512, 'timing/data': 0.016722269356250763, 'timing/model': 0.9998444141820073, 'learning_rate/qwen_vl_interface': 8.573400867438486e-06, 'learning_rate/action_model': 8.573400867438486e-05, 'epoch': 0.84}) +[2026-06-17 19:41:17,944][starVLA.training.train_starvla][INFO] - Step 1849, Loss: {'train/loss': 0.8174741864204407, 'train/grad_norm_pre_clip': 19.87327766418457, 'timing/data': 0.01570090651512146, 'timing/model': 0.9532631114125252, 'learning_rate/qwen_vl_interface': 8.571855339675057e-06, 'learning_rate/action_model': 8.571855339675057e-05, 'epoch': 0.84}) +[2026-06-17 19:41:22,070][starVLA.training.train_starvla][INFO] - Step 1850, Loss: {'train/loss': 0.30741286277770996, 'train/grad_norm_pre_clip': 9.463766098022461, 'timing/data': 0.021669543348252773, 'timing/model': 1.1585162477567792, 'learning_rate/qwen_vl_interface': 8.57030912328744e-06, 'learning_rate/action_model': 8.57030912328744e-05, 'epoch': 0.84}) +[2026-06-17 19:41:25,883][starVLA.training.train_starvla][INFO] - Step 1851, Loss: {'train/loss': 0.7103762626647949, 'train/grad_norm_pre_clip': 20.207435607910156, 'timing/data': 0.022834100760519505, 'timing/model': 0.9400436468422413, 'learning_rate/qwen_vl_interface': 8.568762218596172e-06, 'learning_rate/action_model': 8.568762218596171e-05, 'epoch': 0.84}) +[2026-06-17 19:41:29,836][starVLA.training.train_starvla][INFO] - Step 1852, Loss: {'train/loss': 0.45256996154785156, 'train/grad_norm_pre_clip': 12.38374137878418, 'timing/data': 0.01664821896702051, 'timing/model': 0.9735304638743401, 'learning_rate/qwen_vl_interface': 8.567214625921922e-06, 'learning_rate/action_model': 8.567214625921922e-05, 'epoch': 0.84}) +[2026-06-17 19:41:33,640][starVLA.training.train_starvla][INFO] - Step 1853, Loss: {'train/loss': 0.726607084274292, 'train/grad_norm_pre_clip': 18.50465202331543, 'timing/data': 0.022110498510301113, 'timing/model': 0.9521991806104779, 'learning_rate/qwen_vl_interface': 8.565666345585511e-06, 'learning_rate/action_model': 8.56566634558551e-05, 'epoch': 0.84}) +[2026-06-17 19:41:37,552][starVLA.training.train_starvla][INFO] - Step 1854, Loss: {'train/loss': 0.532194972038269, 'train/grad_norm_pre_clip': 12.174410820007324, 'timing/data': 0.020165910944342613, 'timing/model': 0.9748841570690274, 'learning_rate/qwen_vl_interface': 8.564117377907898e-06, 'learning_rate/action_model': 8.564117377907898e-05, 'epoch': 0.84}) +[2026-06-17 19:41:41,427][starVLA.training.train_starvla][INFO] - Step 1855, Loss: {'train/loss': 0.4405096173286438, 'train/grad_norm_pre_clip': 13.305999755859375, 'timing/data': 0.018446622416377068, 'timing/model': 0.9393091527745128, 'learning_rate/qwen_vl_interface': 8.56256772321019e-06, 'learning_rate/action_model': 8.562567723210188e-05, 'epoch': 0.84}) +[2026-06-17 19:41:45,252][starVLA.training.train_starvla][INFO] - Step 1856, Loss: {'train/loss': 0.3686264157295227, 'train/grad_norm_pre_clip': 19.346994400024414, 'timing/data': 0.0178358880802989, 'timing/model': 0.9655926870182157, 'learning_rate/qwen_vl_interface': 8.561017381813624e-06, 'learning_rate/action_model': 8.561017381813624e-05, 'epoch': 0.84}) +[2026-06-17 19:41:49,213][starVLA.training.train_starvla][INFO] - Step 1857, Loss: {'train/loss': 0.530232846736908, 'train/grad_norm_pre_clip': 11.686579704284668, 'timing/data': 0.024440222419798374, 'timing/model': 1.2389698782935739, 'learning_rate/qwen_vl_interface': 8.559466354039594e-06, 'learning_rate/action_model': 8.559466354039594e-05, 'epoch': 0.84}) +[2026-06-17 19:41:52,841][starVLA.training.train_starvla][INFO] - Step 1858, Loss: {'train/loss': 0.477181077003479, 'train/grad_norm_pre_clip': 10.23882007598877, 'timing/data': 0.021663443185389042, 'timing/model': 0.9379075383767486, 'learning_rate/qwen_vl_interface': 8.557914640209628e-06, 'learning_rate/action_model': 8.557914640209627e-05, 'epoch': 0.84}) +[2026-06-17 19:41:56,702][starVLA.training.train_starvla][INFO] - Step 1859, Loss: {'train/loss': 0.4221334755420685, 'train/grad_norm_pre_clip': 10.451040267944336, 'timing/data': 0.02090202458202839, 'timing/model': 0.9496730724349618, 'learning_rate/qwen_vl_interface': 8.556362240645397e-06, 'learning_rate/action_model': 8.556362240645397e-05, 'epoch': 0.84}) +[2026-06-17 19:42:00,601][starVLA.training.train_starvla][INFO] - Step 1860, Loss: {'train/loss': 0.41366422176361084, 'train/grad_norm_pre_clip': 9.54648494720459, 'timing/data': 0.01625937595963478, 'timing/model': 0.9905439764261246, 'learning_rate/qwen_vl_interface': 8.554809155668716e-06, 'learning_rate/action_model': 8.554809155668716e-05, 'epoch': 0.84}) +[2026-06-17 19:42:04,521][starVLA.training.train_starvla][INFO] - Step 1861, Loss: {'train/loss': 0.4894654154777527, 'train/grad_norm_pre_clip': 11.911439895629883, 'timing/data': 0.014568795450031757, 'timing/model': 0.9627210982143879, 'learning_rate/qwen_vl_interface': 8.553255385601541e-06, 'learning_rate/action_model': 8.55325538560154e-05, 'epoch': 0.84}) +[2026-06-17 19:42:08,361][starVLA.training.train_starvla][INFO] - Step 1862, Loss: {'train/loss': 0.36677443981170654, 'train/grad_norm_pre_clip': 9.424402236938477, 'timing/data': 0.014044482260942459, 'timing/model': 0.9350936617702246, 'learning_rate/qwen_vl_interface': 8.551700930765969e-06, 'learning_rate/action_model': 8.551700930765969e-05, 'epoch': 0.84}) +[2026-06-17 19:42:12,506][starVLA.training.train_starvla][INFO] - Step 1863, Loss: {'train/loss': 0.49447572231292725, 'train/grad_norm_pre_clip': 13.435236930847168, 'timing/data': 0.014415733516216278, 'timing/model': 1.1896796002984047, 'learning_rate/qwen_vl_interface': 8.550145791484243e-06, 'learning_rate/action_model': 8.550145791484242e-05, 'epoch': 0.85}) +[2026-06-17 19:42:16,543][starVLA.training.train_starvla][INFO] - Step 1864, Loss: {'train/loss': 0.21261468529701233, 'train/grad_norm_pre_clip': 7.8869733810424805, 'timing/data': 0.01757117174565792, 'timing/model': 1.2316229762509465, 'learning_rate/qwen_vl_interface': 8.548589968078743e-06, 'learning_rate/action_model': 8.548589968078743e-05, 'epoch': 0.85}) +[2026-06-17 19:42:20,299][starVLA.training.train_starvla][INFO] - Step 1865, Loss: {'train/loss': 0.3356868624687195, 'train/grad_norm_pre_clip': 9.354920387268066, 'timing/data': 0.01529895793646574, 'timing/model': 0.9457491459324956, 'learning_rate/qwen_vl_interface': 8.547033460871992e-06, 'learning_rate/action_model': 8.547033460871992e-05, 'epoch': 0.85}) +[2026-06-17 19:42:24,219][starVLA.training.train_starvla][INFO] - Step 1866, Loss: {'train/loss': 0.7135429978370667, 'train/grad_norm_pre_clip': 20.20719337463379, 'timing/data': 0.020161496475338936, 'timing/model': 0.9572774581611156, 'learning_rate/qwen_vl_interface': 8.545476270186658e-06, 'learning_rate/action_model': 8.545476270186659e-05, 'epoch': 0.85}) +[2026-06-17 19:42:28,170][starVLA.training.train_starvla][INFO] - Step 1867, Loss: {'train/loss': 0.6037951111793518, 'train/grad_norm_pre_clip': 13.836383819580078, 'timing/data': 0.023630560375750065, 'timing/model': 0.9579826854169369, 'learning_rate/qwen_vl_interface': 8.543918396345546e-06, 'learning_rate/action_model': 8.543918396345546e-05, 'epoch': 0.85}) +[2026-06-17 19:42:32,232][starVLA.training.train_starvla][INFO] - Step 1868, Loss: {'train/loss': 0.41731053590774536, 'train/grad_norm_pre_clip': 97.0545425415039, 'timing/data': 0.01937870215624571, 'timing/model': 0.964539984241128, 'learning_rate/qwen_vl_interface': 8.542359839671607e-06, 'learning_rate/action_model': 8.542359839671606e-05, 'epoch': 0.85}) +[2026-06-17 19:42:36,208][starVLA.training.train_starvla][INFO] - Step 1869, Loss: {'train/loss': 0.3276585042476654, 'train/grad_norm_pre_clip': 13.47590446472168, 'timing/data': 0.017175217159092426, 'timing/model': 0.9606841020286083, 'learning_rate/qwen_vl_interface': 8.540800600487931e-06, 'learning_rate/action_model': 8.54080060048793e-05, 'epoch': 0.85}) +[2026-06-17 19:42:40,313][starVLA.training.train_starvla][INFO] - Step 1870, Loss: {'train/loss': 0.4529588222503662, 'train/grad_norm_pre_clip': 9.88035774230957, 'timing/data': 0.016311186365783215, 'timing/model': 1.229915919713676, 'learning_rate/qwen_vl_interface': 8.539240679117748e-06, 'learning_rate/action_model': 8.539240679117748e-05, 'epoch': 0.85}) +[2026-06-17 19:42:44,318][starVLA.training.train_starvla][INFO] - Step 1871, Loss: {'train/loss': 0.36701858043670654, 'train/grad_norm_pre_clip': 20.107421875, 'timing/data': 0.019316645339131355, 'timing/model': 1.2319369381293654, 'learning_rate/qwen_vl_interface': 8.537680075884433e-06, 'learning_rate/action_model': 8.537680075884433e-05, 'epoch': 0.85}) +[2026-06-17 19:42:48,111][starVLA.training.train_starvla][INFO] - Step 1872, Loss: {'train/loss': 0.5659621357917786, 'train/grad_norm_pre_clip': 12.889091491699219, 'timing/data': 0.018100230023264885, 'timing/model': 0.9723610654473305, 'learning_rate/qwen_vl_interface': 8.5361187911115e-06, 'learning_rate/action_model': 8.5361187911115e-05, 'epoch': 0.85}) +[2026-06-17 19:42:52,068][starVLA.training.train_starvla][INFO] - Step 1873, Loss: {'train/loss': 0.5034977793693542, 'train/grad_norm_pre_clip': 15.622313499450684, 'timing/data': 0.02600529044866562, 'timing/model': 0.9782722145318985, 'learning_rate/qwen_vl_interface': 8.534556825122608e-06, 'learning_rate/action_model': 8.534556825122607e-05, 'epoch': 0.85}) +[2026-06-17 19:42:56,069][starVLA.training.train_starvla][INFO] - Step 1874, Loss: {'train/loss': 0.5916046500205994, 'train/grad_norm_pre_clip': 15.18750286102295, 'timing/data': 0.021552552469074726, 'timing/model': 0.9840767616406083, 'learning_rate/qwen_vl_interface': 8.532994178241551e-06, 'learning_rate/action_model': 8.53299417824155e-05, 'epoch': 0.85}) +[2026-06-17 19:43:00,032][starVLA.training.train_starvla][INFO] - Step 1875, Loss: {'train/loss': 0.5350342988967896, 'train/grad_norm_pre_clip': 14.315478324890137, 'timing/data': 0.016376922838389874, 'timing/model': 1.070789385586977, 'learning_rate/qwen_vl_interface': 8.531430850792267e-06, 'learning_rate/action_model': 8.531430850792266e-05, 'epoch': 0.85}) +[2026-06-17 19:43:04,165][starVLA.training.train_starvla][INFO] - Step 1876, Loss: {'train/loss': 0.6239984035491943, 'train/grad_norm_pre_clip': 15.826756477355957, 'timing/data': 0.019501172937452793, 'timing/model': 0.9518386097624898, 'learning_rate/qwen_vl_interface': 8.529866843098838e-06, 'learning_rate/action_model': 8.529866843098837e-05, 'epoch': 0.85}) +[2026-06-17 19:43:08,373][starVLA.training.train_starvla][INFO] - Step 1877, Loss: {'train/loss': 0.6894370913505554, 'train/grad_norm_pre_clip': 14.18786334991455, 'timing/data': 0.022242680191993713, 'timing/model': 1.3690064745023847, 'learning_rate/qwen_vl_interface': 8.528302155485482e-06, 'learning_rate/action_model': 8.528302155485481e-05, 'epoch': 0.85}) +[2026-06-17 19:43:12,335][starVLA.training.train_starvla][INFO] - Step 1878, Loss: {'train/loss': 0.6468438506126404, 'train/grad_norm_pre_clip': 11.793458938598633, 'timing/data': 0.03293832205235958, 'timing/model': 0.9858290543779731, 'learning_rate/qwen_vl_interface': 8.526736788276563e-06, 'learning_rate/action_model': 8.526736788276563e-05, 'epoch': 0.85}) +[2026-06-17 19:43:16,516][starVLA.training.train_starvla][INFO] - Step 1879, Loss: {'train/loss': 0.40053069591522217, 'train/grad_norm_pre_clip': 9.159090042114258, 'timing/data': 0.027254152111709118, 'timing/model': 0.9786283215507865, 'learning_rate/qwen_vl_interface': 8.525170741796578e-06, 'learning_rate/action_model': 8.525170741796579e-05, 'epoch': 0.85}) +[2026-06-17 19:43:20,599][starVLA.training.train_starvla][INFO] - Step 1880, Loss: {'train/loss': 0.3793984651565552, 'train/grad_norm_pre_clip': 10.141266822814941, 'timing/data': 0.020031407475471497, 'timing/model': 0.9845514185726643, 'learning_rate/qwen_vl_interface': 8.523604016370179e-06, 'learning_rate/action_model': 8.523604016370179e-05, 'epoch': 0.85}) +[2026-06-17 19:43:24,548][starVLA.training.train_starvla][INFO] - Step 1881, Loss: {'train/loss': 0.8727342486381531, 'train/grad_norm_pre_clip': 15.5309419631958, 'timing/data': 0.018021998926997185, 'timing/model': 0.9472537711262703, 'learning_rate/qwen_vl_interface': 8.522036612322143e-06, 'learning_rate/action_model': 8.522036612322142e-05, 'epoch': 0.85}) +[2026-06-17 19:43:28,633][starVLA.training.train_starvla][INFO] - Step 1882, Loss: {'train/loss': 0.3343065083026886, 'train/grad_norm_pre_clip': 8.898138046264648, 'timing/data': 0.02981063909828663, 'timing/model': 0.9621215602383018, 'learning_rate/qwen_vl_interface': 8.520468529977395e-06, 'learning_rate/action_model': 8.520468529977395e-05, 'epoch': 0.85}) +[2026-06-17 19:43:32,615][starVLA.training.train_starvla][INFO] - Step 1883, Loss: {'train/loss': 0.34837231040000916, 'train/grad_norm_pre_clip': 9.828405380249023, 'timing/data': 0.02264026738703251, 'timing/model': 1.211869434453547, 'learning_rate/qwen_vl_interface': 8.518899769661002e-06, 'learning_rate/action_model': 8.518899769661002e-05, 'epoch': 0.85}) +[2026-06-17 19:43:36,564][starVLA.training.train_starvla][INFO] - Step 1884, Loss: {'train/loss': 0.678667426109314, 'train/grad_norm_pre_clip': 14.868188858032227, 'timing/data': 0.015134441666305065, 'timing/model': 1.1986914752051234, 'learning_rate/qwen_vl_interface': 8.517330331698171e-06, 'learning_rate/action_model': 8.51733033169817e-05, 'epoch': 0.85}) +[2026-06-17 19:43:40,267][starVLA.training.train_starvla][INFO] - Step 1885, Loss: {'train/loss': 0.6069895625114441, 'train/grad_norm_pre_clip': 14.507763862609863, 'timing/data': 0.01784280128777027, 'timing/model': 0.9474205300211906, 'learning_rate/qwen_vl_interface': 8.515760216414244e-06, 'learning_rate/action_model': 8.515760216414244e-05, 'epoch': 0.86}) +[2026-06-17 19:43:44,193][starVLA.training.train_starvla][INFO] - Step 1886, Loss: {'train/loss': 0.48320358991622925, 'train/grad_norm_pre_clip': 9.225028038024902, 'timing/data': 0.023301051929593086, 'timing/model': 0.9576528184115887, 'learning_rate/qwen_vl_interface': 8.514189424134712e-06, 'learning_rate/action_model': 8.514189424134713e-05, 'epoch': 0.86}) +[2026-06-17 19:43:48,265][starVLA.training.train_starvla][INFO] - Step 1887, Loss: {'train/loss': 0.31902509927749634, 'train/grad_norm_pre_clip': 12.412614822387695, 'timing/data': 0.023977543227374554, 'timing/model': 1.0195045759901404, 'learning_rate/qwen_vl_interface': 8.512617955185199e-06, 'learning_rate/action_model': 8.5126179551852e-05, 'epoch': 0.86}) +[2026-06-17 19:43:52,378][starVLA.training.train_starvla][INFO] - Step 1888, Loss: {'train/loss': 0.5358617901802063, 'train/grad_norm_pre_clip': 13.921008110046387, 'timing/data': 0.01876614522188902, 'timing/model': 0.9600322796031833, 'learning_rate/qwen_vl_interface': 8.511045809891472e-06, 'learning_rate/action_model': 8.511045809891472e-05, 'epoch': 0.86}) +[2026-06-17 19:43:56,540][starVLA.training.train_starvla][INFO] - Step 1889, Loss: {'train/loss': 0.4170423746109009, 'train/grad_norm_pre_clip': 11.109292030334473, 'timing/data': 0.022860275581479073, 'timing/model': 1.0082305055111647, 'learning_rate/qwen_vl_interface': 8.509472988579443e-06, 'learning_rate/action_model': 8.509472988579443e-05, 'epoch': 0.86}) +[2026-06-17 19:44:00,525][starVLA.training.train_starvla][INFO] - Step 1890, Loss: {'train/loss': 0.8737415671348572, 'train/grad_norm_pre_clip': 16.246896743774414, 'timing/data': 0.018706152215600014, 'timing/model': 1.2483208030462265, 'learning_rate/qwen_vl_interface': 8.507899491575153e-06, 'learning_rate/action_model': 8.507899491575154e-05, 'epoch': 0.86}) +[2026-06-17 19:44:04,503][starVLA.training.train_starvla][INFO] - Step 1891, Loss: {'train/loss': 0.6651325225830078, 'train/grad_norm_pre_clip': 14.585928916931152, 'timing/data': 0.015825017355382442, 'timing/model': 1.1959941172972322, 'learning_rate/qwen_vl_interface': 8.506325319204795e-06, 'learning_rate/action_model': 8.506325319204796e-05, 'epoch': 0.86}) +[2026-06-17 19:44:08,255][starVLA.training.train_starvla][INFO] - Step 1892, Loss: {'train/loss': 0.8489975333213806, 'train/grad_norm_pre_clip': 13.787615776062012, 'timing/data': 0.018259715288877487, 'timing/model': 0.9823007043451071, 'learning_rate/qwen_vl_interface': 8.504750471794694e-06, 'learning_rate/action_model': 8.504750471794693e-05, 'epoch': 0.86}) +[2026-06-17 19:44:12,248][starVLA.training.train_starvla][INFO] - Step 1893, Loss: {'train/loss': 0.46095147728919983, 'train/grad_norm_pre_clip': 10.157220840454102, 'timing/data': 0.01846804842352867, 'timing/model': 0.9679943574592471, 'learning_rate/qwen_vl_interface': 8.503174949671315e-06, 'learning_rate/action_model': 8.503174949671315e-05, 'epoch': 0.86}) +[2026-06-17 19:44:16,331][starVLA.training.train_starvla][INFO] - Step 1894, Loss: {'train/loss': 0.49335774779319763, 'train/grad_norm_pre_clip': 10.628257751464844, 'timing/data': 0.021513638086616993, 'timing/model': 0.9593103677034378, 'learning_rate/qwen_vl_interface': 8.501598753161273e-06, 'learning_rate/action_model': 8.501598753161272e-05, 'epoch': 0.86}) +[2026-06-17 19:44:20,261][starVLA.training.train_starvla][INFO] - Step 1895, Loss: {'train/loss': 0.4311330020427704, 'train/grad_norm_pre_clip': 10.588197708129883, 'timing/data': 0.01822054572403431, 'timing/model': 0.9496288411319256, 'learning_rate/qwen_vl_interface': 8.500021882591308e-06, 'learning_rate/action_model': 8.500021882591308e-05, 'epoch': 0.86}) +[2026-06-17 19:44:24,295][starVLA.training.train_starvla][INFO] - Step 1896, Loss: {'train/loss': 0.33477962017059326, 'train/grad_norm_pre_clip': 10.620612144470215, 'timing/data': 0.017978300340473652, 'timing/model': 0.9563122289255261, 'learning_rate/qwen_vl_interface': 8.498444338288308e-06, 'learning_rate/action_model': 8.498444338288309e-05, 'epoch': 0.86}) +[2026-06-17 19:44:28,207][starVLA.training.train_starvla][INFO] - Step 1897, Loss: {'train/loss': 0.5247240662574768, 'train/grad_norm_pre_clip': 20.39752769470215, 'timing/data': 0.017981699667871, 'timing/model': 1.186917757615447, 'learning_rate/qwen_vl_interface': 8.496866120579304e-06, 'learning_rate/action_model': 8.496866120579303e-05, 'epoch': 0.86}) +[2026-06-17 19:44:31,907][starVLA.training.train_starvla][INFO] - Step 1898, Loss: {'train/loss': 0.4974710941314697, 'train/grad_norm_pre_clip': 12.188068389892578, 'timing/data': 0.015201843343675137, 'timing/model': 0.9478461612015963, 'learning_rate/qwen_vl_interface': 8.495287229791459e-06, 'learning_rate/action_model': 8.495287229791459e-05, 'epoch': 0.86}) +[2026-06-17 19:44:35,821][starVLA.training.train_starvla][INFO] - Step 1899, Loss: {'train/loss': 0.5931032299995422, 'train/grad_norm_pre_clip': 12.10145378112793, 'timing/data': 0.01951038185507059, 'timing/model': 0.9449326498433948, 'learning_rate/qwen_vl_interface': 8.493707666252078e-06, 'learning_rate/action_model': 8.493707666252078e-05, 'epoch': 0.86}) +[2026-06-17 19:46:52,178][starVLA.training.train_starvla][INFO] - Step 1900, Loss: {'train/loss': 0.5403759479522705, 'train/grad_norm_pre_clip': 14.983656883239746, 'eval/loss': 0.4613812565803528, 'eval/latency_8/loss': 0.4613812565803528, 'eval/demon_fire_precision': 0.9299552906110283, 'eval/demon_fire_recall': 0.9692451071761417, 'eval/demon_fire_f1': 0.9491937937328871, 'eval/demon_fire_rate_model': 0.9335002782415136, 'eval/demon_fire_rate_teacher': 0.8956594323873122, 'eval/demon_left_precision': 0.7368421052631579, 'eval/demon_left_recall': 0.6295503211991434, 'eval/demon_left_f1': 0.6789838337182448, 'eval/demon_left_rate_model': 0.11101836393989983, 'eval/demon_left_rate_teacher': 0.12993878686700056, 'eval/demon_right_precision': 0.8007662835249042, 'eval/demon_right_recall': 0.6333333333333333, 'eval/demon_right_f1': 0.7072758037225041, 'eval/demon_right_rate_model': 0.14524207011686144, 'eval/demon_right_rate_teacher': 0.18363939899833054, 'eval/demon_move_f1': 0.6931298187203745, 'eval/demon_cc_f1': 0.8544910280936318, 'eval/latency_8/demon_fire_precision': 0.9299552906110283, 'eval/latency_8/demon_fire_recall': 0.9692451071761417, 'eval/latency_8/demon_fire_f1': 0.9491937937328871, 'eval/latency_8/demon_fire_rate_model': 0.9335002782415136, 'eval/latency_8/demon_fire_rate_teacher': 0.8956594323873122, 'eval/latency_8/demon_left_precision': 0.7368421052631579, 'eval/latency_8/demon_left_recall': 0.6295503211991434, 'eval/latency_8/demon_left_f1': 0.6789838337182448, 'eval/latency_8/demon_left_rate_model': 0.11101836393989983, 'eval/latency_8/demon_left_rate_teacher': 0.12993878686700056, 'eval/latency_8/demon_right_precision': 0.8007662835249042, 'eval/latency_8/demon_right_recall': 0.6333333333333333, 'eval/latency_8/demon_right_f1': 0.7072758037225041, 'eval/latency_8/demon_right_rate_model': 0.14524207011686144, 'eval/latency_8/demon_right_rate_teacher': 0.18363939899833054, 'eval/latency_8/demon_move_f1': 0.6931298187203745, 'eval/latency_8/demon_cc_f1': 0.8544910280936318, 'timing/data': 0.019367748871445656, 'timing/model': 0.9509236328303814, 'learning_rate/qwen_vl_interface': 8.492127430288609e-06, 'learning_rate/action_model': 8.492127430288608e-05, 'epoch': 0.86}) +[2026-06-17 19:46:52,183][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_1900_state +[2026-06-17 19:46:52,189][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 19:46:59,829][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_1900_state/model.safetensors +[2026-06-17 19:47:16,925][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_1900_state/optimizer.bin +[2026-06-17 19:47:16,926][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_1900_state/sampler.bin +[2026-06-17 19:47:16,926][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_1900_state/sampler_1.bin +[2026-06-17 19:47:16,945][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_1900_state/random_states_0.pkl +[2026-06-17 19:47:16,948][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1900 +[2026-06-17 19:47:16,949][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 19:47:16,960][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 19:47:20,955][starVLA.training.train_starvla][INFO] - Step 1901, Loss: {'train/loss': 0.35661518573760986, 'train/grad_norm_pre_clip': 11.37977409362793, 'timing/data': 0.01736351754516363, 'timing/model': 0.9463710729032755, 'learning_rate/qwen_vl_interface': 8.490546522228633e-06, 'learning_rate/action_model': 8.490546522228633e-05, 'epoch': 0.86}) +[2026-06-17 19:47:25,202][starVLA.training.train_starvla][INFO] - Step 1902, Loss: {'train/loss': 0.2690247595310211, 'train/grad_norm_pre_clip': 9.38797664642334, 'timing/data': 0.019505536183714867, 'timing/model': 1.2300673248246312, 'learning_rate/qwen_vl_interface': 8.488964942399876e-06, 'learning_rate/action_model': 8.488964942399877e-05, 'epoch': 0.86}) +[2026-06-17 19:47:29,563][starVLA.training.train_starvla][INFO] - Step 1903, Loss: {'train/loss': 0.5280563235282898, 'train/grad_norm_pre_clip': 12.54301643371582, 'timing/data': 0.016007994301617146, 'timing/model': 0.9522312106564641, 'learning_rate/qwen_vl_interface': 8.487382691130203e-06, 'learning_rate/action_model': 8.487382691130202e-05, 'epoch': 0.86}) +[2026-06-17 19:47:37,267][starVLA.training.train_starvla][INFO] - Step 1904, Loss: {'train/loss': 0.7275264859199524, 'train/grad_norm_pre_clip': 13.280202865600586, 'timing/data': 0.01951044797897339, 'timing/model': 0.9459548629820347, 'learning_rate/qwen_vl_interface': 8.485799768747614e-06, 'learning_rate/action_model': 8.485799768747614e-05, 'epoch': 0.86}) +[2026-06-17 19:47:41,874][starVLA.training.train_starvla][INFO] - Step 1905, Loss: {'train/loss': 0.6977245807647705, 'train/grad_norm_pre_clip': 13.883769035339355, 'timing/data': 0.0197521960362792, 'timing/model': 0.9584458479657769, 'learning_rate/qwen_vl_interface': 8.484216175580248e-06, 'learning_rate/action_model': 8.484216175580249e-05, 'epoch': 0.86}) +[2026-06-17 19:47:45,831][starVLA.training.train_starvla][INFO] - Step 1906, Loss: {'train/loss': 0.7049065232276917, 'train/grad_norm_pre_clip': 17.484172821044922, 'timing/data': 0.01967179123312235, 'timing/model': 1.1989754494279623, 'learning_rate/qwen_vl_interface': 8.482631911956395e-06, 'learning_rate/action_model': 8.482631911956395e-05, 'epoch': 0.86}) +[2026-06-17 19:47:49,734][starVLA.training.train_starvla][INFO] - Step 1907, Loss: {'train/loss': 0.416103720664978, 'train/grad_norm_pre_clip': 10.942911148071289, 'timing/data': 0.022142454981803894, 'timing/model': 1.1483800997957587, 'learning_rate/qwen_vl_interface': 8.481046978204462e-06, 'learning_rate/action_model': 8.481046978204463e-05, 'epoch': 0.87}) +[2026-06-17 19:47:54,687][starVLA.training.train_starvla][INFO] - Step 1908, Loss: {'train/loss': 0.4211849272251129, 'train/grad_norm_pre_clip': 13.402945518493652, 'timing/data': 0.016827653162181377, 'timing/model': 1.034825368784368, 'learning_rate/qwen_vl_interface': 8.479461374653017e-06, 'learning_rate/action_model': 8.479461374653018e-05, 'epoch': 0.87}) +[2026-06-17 19:47:59,516][starVLA.training.train_starvla][INFO] - Step 1909, Loss: {'train/loss': 0.34601476788520813, 'train/grad_norm_pre_clip': 13.264643669128418, 'timing/data': 0.02739034127444029, 'timing/model': 1.0154444389045238, 'learning_rate/qwen_vl_interface': 8.477875101630752e-06, 'learning_rate/action_model': 8.477875101630753e-05, 'epoch': 0.87}) +[2026-06-17 19:48:03,597][starVLA.training.train_starvla][INFO] - Step 1910, Loss: {'train/loss': 0.32448911666870117, 'train/grad_norm_pre_clip': 10.1676664352417, 'timing/data': 0.017541205510497093, 'timing/model': 0.9443712830543518, 'learning_rate/qwen_vl_interface': 8.476288159466508e-06, 'learning_rate/action_model': 8.476288159466508e-05, 'epoch': 0.87}) +[2026-06-17 19:48:07,729][starVLA.training.train_starvla][INFO] - Step 1911, Loss: {'train/loss': 0.6667941212654114, 'train/grad_norm_pre_clip': 15.923517227172852, 'timing/data': 0.020822491496801376, 'timing/model': 1.0664999056607485, 'learning_rate/qwen_vl_interface': 8.474700548489255e-06, 'learning_rate/action_model': 8.474700548489255e-05, 'epoch': 0.87}) +[2026-06-17 19:48:12,305][starVLA.training.train_starvla][INFO] - Step 1912, Loss: {'train/loss': 0.4700000584125519, 'train/grad_norm_pre_clip': 11.994440078735352, 'timing/data': 0.019791454076766968, 'timing/model': 1.62449533585459, 'learning_rate/qwen_vl_interface': 8.473112269028108e-06, 'learning_rate/action_model': 8.473112269028107e-05, 'epoch': 0.87}) +[2026-06-17 19:48:16,264][starVLA.training.train_starvla][INFO] - Step 1913, Loss: {'train/loss': 0.5476561188697815, 'train/grad_norm_pre_clip': 10.858839988708496, 'timing/data': 0.016779212281107903, 'timing/model': 1.1279002148658037, 'learning_rate/qwen_vl_interface': 8.47152332141232e-06, 'learning_rate/action_model': 8.47152332141232e-05, 'epoch': 0.87}) +[2026-06-17 19:48:21,965][starVLA.training.train_starvla][INFO] - Step 1914, Loss: {'train/loss': 0.352655291557312, 'train/grad_norm_pre_clip': 10.161149978637695, 'timing/data': 0.016205632127821445, 'timing/model': 0.9292598906904459, 'learning_rate/qwen_vl_interface': 8.469933705971282e-06, 'learning_rate/action_model': 8.469933705971282e-05, 'epoch': 0.87}) +[2026-06-17 19:48:26,182][starVLA.training.train_starvla][INFO] - Step 1915, Loss: {'train/loss': 0.33578118681907654, 'train/grad_norm_pre_clip': 14.783851623535156, 'timing/data': 0.0178363136947155, 'timing/model': 1.16980988252908, 'learning_rate/qwen_vl_interface': 8.468343423034521e-06, 'learning_rate/action_model': 8.46834342303452e-05, 'epoch': 0.87}) +[2026-06-17 19:48:31,185][starVLA.training.train_starvla][INFO] - Step 1916, Loss: {'train/loss': 0.39637938141822815, 'train/grad_norm_pre_clip': 12.370139122009277, 'timing/data': 0.02421187236905098, 'timing/model': 1.161052767187357, 'learning_rate/qwen_vl_interface': 8.466752472931704e-06, 'learning_rate/action_model': 8.466752472931704e-05, 'epoch': 0.87}) +[2026-06-17 19:48:35,600][starVLA.training.train_starvla][INFO] - Step 1917, Loss: {'train/loss': 0.48932474851608276, 'train/grad_norm_pre_clip': 14.23735237121582, 'timing/data': 0.017750179395079613, 'timing/model': 0.947045985609293, 'learning_rate/qwen_vl_interface': 8.465160855992639e-06, 'learning_rate/action_model': 8.465160855992638e-05, 'epoch': 0.87}) +[2026-06-17 19:48:41,154][starVLA.training.train_starvla][INFO] - Step 1918, Loss: {'train/loss': 0.33438369631767273, 'train/grad_norm_pre_clip': 8.779909133911133, 'timing/data': 0.014308537356555462, 'timing/model': 0.9357759021222591, 'learning_rate/qwen_vl_interface': 8.46356857254727e-06, 'learning_rate/action_model': 8.46356857254727e-05, 'epoch': 0.87}) +[2026-06-17 19:48:44,968][starVLA.training.train_starvla][INFO] - Step 1919, Loss: {'train/loss': 0.7304581999778748, 'train/grad_norm_pre_clip': 14.261677742004395, 'timing/data': 0.01691486779600382, 'timing/model': 1.1062019048258662, 'learning_rate/qwen_vl_interface': 8.461975622925676e-06, 'learning_rate/action_model': 8.461975622925676e-05, 'epoch': 0.87}) +[2026-06-17 19:48:48,672][starVLA.training.train_starvla][INFO] - Step 1920, Loss: {'train/loss': 0.6097160577774048, 'train/grad_norm_pre_clip': 17.49424171447754, 'timing/data': 0.02184420730918646, 'timing/model': 0.9384141555055976, 'learning_rate/qwen_vl_interface': 8.460382007458081e-06, 'learning_rate/action_model': 8.460382007458081e-05, 'epoch': 0.87}) +[2026-06-17 19:48:52,533][starVLA.training.train_starvla][INFO] - Step 1921, Loss: {'train/loss': 0.6298062801361084, 'train/grad_norm_pre_clip': 19.247045516967773, 'timing/data': 0.01644384302198887, 'timing/model': 0.9418108835816383, 'learning_rate/qwen_vl_interface': 8.45878772647484e-06, 'learning_rate/action_model': 8.458787726474839e-05, 'epoch': 0.87}) +[2026-06-17 19:48:56,391][starVLA.training.train_starvla][INFO] - Step 1922, Loss: {'train/loss': 0.48089098930358887, 'train/grad_norm_pre_clip': 9.192206382751465, 'timing/data': 0.022425959818065166, 'timing/model': 0.9415911920368671, 'learning_rate/qwen_vl_interface': 8.457192780306452e-06, 'learning_rate/action_model': 8.457192780306451e-05, 'epoch': 0.87}) +[2026-06-17 19:49:00,285][starVLA.training.train_starvla][INFO] - Step 1923, Loss: {'train/loss': 0.6537708640098572, 'train/grad_norm_pre_clip': 18.081605911254883, 'timing/data': 0.016069823876023293, 'timing/model': 0.9412147356197238, 'learning_rate/qwen_vl_interface': 8.455597169283549e-06, 'learning_rate/action_model': 8.455597169283548e-05, 'epoch': 0.87}) +[2026-06-17 19:49:04,247][starVLA.training.train_starvla][INFO] - Step 1924, Loss: {'train/loss': 0.5449795126914978, 'train/grad_norm_pre_clip': 12.258747100830078, 'timing/data': 0.022171557880938053, 'timing/model': 0.9425057182088494, 'learning_rate/qwen_vl_interface': 8.454000893736901e-06, 'learning_rate/action_model': 8.4540008937369e-05, 'epoch': 0.87}) +[2026-06-17 19:49:08,178][starVLA.training.train_starvla][INFO] - Step 1925, Loss: {'train/loss': 0.4882257580757141, 'train/grad_norm_pre_clip': 14.125276565551758, 'timing/data': 0.015612740069627762, 'timing/model': 1.1594877867028117, 'learning_rate/qwen_vl_interface': 8.452403953997422e-06, 'learning_rate/action_model': 8.452403953997422e-05, 'epoch': 0.87}) +[2026-06-17 19:49:12,271][starVLA.training.train_starvla][INFO] - Step 1926, Loss: {'train/loss': 0.6138225793838501, 'train/grad_norm_pre_clip': 11.239089012145996, 'timing/data': 0.02963361795991659, 'timing/model': 1.2496862122789025, 'learning_rate/qwen_vl_interface': 8.450806350396158e-06, 'learning_rate/action_model': 8.450806350396158e-05, 'epoch': 0.87}) +[2026-06-17 19:49:16,179][starVLA.training.train_starvla][INFO] - Step 1927, Loss: {'train/loss': 0.7034215927124023, 'train/grad_norm_pre_clip': 14.0806884765625, 'timing/data': 0.023829247802495956, 'timing/model': 0.9802358848974109, 'learning_rate/qwen_vl_interface': 8.449208083264291e-06, 'learning_rate/action_model': 8.449208083264292e-05, 'epoch': 0.87}) +[2026-06-17 19:49:20,137][starVLA.training.train_starvla][INFO] - Step 1928, Loss: {'train/loss': 0.34385043382644653, 'train/grad_norm_pre_clip': 10.268692016601562, 'timing/data': 0.016787038184702396, 'timing/model': 0.9625981440767646, 'learning_rate/qwen_vl_interface': 8.447609152933149e-06, 'learning_rate/action_model': 8.447609152933149e-05, 'epoch': 0.87}) +[2026-06-17 19:49:24,093][starVLA.training.train_starvla][INFO] - Step 1929, Loss: {'train/loss': 0.47807732224464417, 'train/grad_norm_pre_clip': 11.330972671508789, 'timing/data': 0.02032283041626215, 'timing/model': 0.9533633133396506, 'learning_rate/qwen_vl_interface': 8.44600955973419e-06, 'learning_rate/action_model': 8.446009559734188e-05, 'epoch': 0.88}) +[2026-06-17 19:49:28,057][starVLA.training.train_starvla][INFO] - Step 1930, Loss: {'train/loss': 0.34732601046562195, 'train/grad_norm_pre_clip': 15.822198867797852, 'timing/data': 0.018635726533830166, 'timing/model': 0.9666373394429684, 'learning_rate/qwen_vl_interface': 8.444409303999005e-06, 'learning_rate/action_model': 8.444409303999005e-05, 'epoch': 0.88}) +[2026-06-17 19:49:35,952][starVLA.training.train_starvla][INFO] - Step 1931, Loss: {'train/loss': 0.377549946308136, 'train/grad_norm_pre_clip': 12.965399742126465, 'timing/data': 0.016589192673563957, 'timing/model': 0.9512200662866235, 'learning_rate/qwen_vl_interface': 8.442808386059337e-06, 'learning_rate/action_model': 8.442808386059337e-05, 'epoch': 0.88}) +[2026-06-17 19:49:41,150][starVLA.training.train_starvla][INFO] - Step 1932, Loss: {'train/loss': 0.3884139060974121, 'train/grad_norm_pre_clip': 12.167630195617676, 'timing/data': 0.01603926718235016, 'timing/model': 1.2141607953235507, 'learning_rate/qwen_vl_interface': 8.441206806247055e-06, 'learning_rate/action_model': 8.441206806247054e-05, 'epoch': 0.88}) +[2026-06-17 19:49:45,100][starVLA.training.train_starvla][INFO] - Step 1933, Loss: {'train/loss': 0.5975428819656372, 'train/grad_norm_pre_clip': 15.517001152038574, 'timing/data': 0.018430563621222973, 'timing/model': 1.2260907925665379, 'learning_rate/qwen_vl_interface': 8.439604564894167e-06, 'learning_rate/action_model': 8.439604564894167e-05, 'epoch': 0.88}) +[2026-06-17 19:49:48,863][starVLA.training.train_starvla][INFO] - Step 1934, Loss: {'train/loss': 0.3753339946269989, 'train/grad_norm_pre_clip': 13.666097640991211, 'timing/data': 0.01778452657163143, 'timing/model': 0.9769036173820496, 'learning_rate/qwen_vl_interface': 8.438001662332821e-06, 'learning_rate/action_model': 8.43800166233282e-05, 'epoch': 0.88}) +[2026-06-17 19:49:53,726][starVLA.training.train_starvla][INFO] - Step 1935, Loss: {'train/loss': 0.4763543903827667, 'train/grad_norm_pre_clip': 25.231891632080078, 'timing/data': 0.04156030621379614, 'timing/model': 1.7731116367504, 'learning_rate/qwen_vl_interface': 8.4363980988953e-06, 'learning_rate/action_model': 8.436398098895301e-05, 'epoch': 0.88}) +[2026-06-17 19:49:58,013][starVLA.training.train_starvla][INFO] - Step 1936, Loss: {'train/loss': 0.41807833313941956, 'train/grad_norm_pre_clip': 11.635200500488281, 'timing/data': 0.021902565844357014, 'timing/model': 0.993274625390768, 'learning_rate/qwen_vl_interface': 8.434793874914026e-06, 'learning_rate/action_model': 8.434793874914026e-05, 'epoch': 0.88}) +[2026-06-17 19:50:04,244][starVLA.training.train_starvla][INFO] - Step 1937, Loss: {'train/loss': 0.5781031250953674, 'train/grad_norm_pre_clip': 14.524701118469238, 'timing/data': 0.018232216127216816, 'timing/model': 1.0752301337197423, 'learning_rate/qwen_vl_interface': 8.433188990721553e-06, 'learning_rate/action_model': 8.433188990721552e-05, 'epoch': 0.88}) +[2026-06-17 19:50:08,445][starVLA.training.train_starvla][INFO] - Step 1938, Loss: {'train/loss': 0.573766827583313, 'train/grad_norm_pre_clip': 15.22601318359375, 'timing/data': 0.023059158585965633, 'timing/model': 1.0218065101653337, 'learning_rate/qwen_vl_interface': 8.431583446650577e-06, 'learning_rate/action_model': 8.431583446650578e-05, 'epoch': 0.88}) +[2026-06-17 19:50:12,617][starVLA.training.train_starvla][INFO] - Step 1939, Loss: {'train/loss': 0.25559088587760925, 'train/grad_norm_pre_clip': 9.64052677154541, 'timing/data': 0.01746526826173067, 'timing/model': 1.1946059288457036, 'learning_rate/qwen_vl_interface': 8.429977243033928e-06, 'learning_rate/action_model': 8.429977243033928e-05, 'epoch': 0.88}) +[2026-06-17 19:50:16,495][starVLA.training.train_starvla][INFO] - Step 1940, Loss: {'train/loss': 0.5739287734031677, 'train/grad_norm_pre_clip': 14.942243576049805, 'timing/data': 0.017677578143775463, 'timing/model': 1.0036218613386154, 'learning_rate/qwen_vl_interface': 8.428370380204577e-06, 'learning_rate/action_model': 8.428370380204577e-05, 'epoch': 0.88}) +[2026-06-17 19:50:23,463][starVLA.training.train_starvla][INFO] - Step 1941, Loss: {'train/loss': 0.28908050060272217, 'train/grad_norm_pre_clip': 10.905305862426758, 'timing/data': 0.018900803290307522, 'timing/model': 2.5878284154459834, 'learning_rate/qwen_vl_interface': 8.426762858495625e-06, 'learning_rate/action_model': 8.426762858495625e-05, 'epoch': 0.88}) +[2026-06-17 19:50:27,447][starVLA.training.train_starvla][INFO] - Step 1942, Loss: {'train/loss': 0.21577304601669312, 'train/grad_norm_pre_clip': 7.414656639099121, 'timing/data': 0.024170306511223316, 'timing/model': 0.9519391329959035, 'learning_rate/qwen_vl_interface': 8.425154678240315e-06, 'learning_rate/action_model': 8.425154678240314e-05, 'epoch': 0.88}) +[2026-06-17 19:50:31,666][starVLA.training.train_starvla][INFO] - Step 1943, Loss: {'train/loss': 0.6422543525695801, 'train/grad_norm_pre_clip': 15.330108642578125, 'timing/data': 0.029130304232239723, 'timing/model': 0.9711831593886018, 'learning_rate/qwen_vl_interface': 8.423545839772021e-06, 'learning_rate/action_model': 8.423545839772021e-05, 'epoch': 0.88}) +[2026-06-17 19:50:35,633][starVLA.training.train_starvla][INFO] - Step 1944, Loss: {'train/loss': 0.5879918336868286, 'train/grad_norm_pre_clip': 20.98714256286621, 'timing/data': 0.01330975629389286, 'timing/model': 0.9576135156676173, 'learning_rate/qwen_vl_interface': 8.421936343424264e-06, 'learning_rate/action_model': 8.421936343424262e-05, 'epoch': 0.88}) +[2026-06-17 19:50:39,546][starVLA.training.train_starvla][INFO] - Step 1945, Loss: {'train/loss': 0.4878559708595276, 'train/grad_norm_pre_clip': 12.543821334838867, 'timing/data': 0.0249939002096653, 'timing/model': 1.172875288873911, 'learning_rate/qwen_vl_interface': 8.420326189530685e-06, 'learning_rate/action_model': 8.420326189530685e-05, 'epoch': 0.88}) +[2026-06-17 19:50:43,511][starVLA.training.train_starvla][INFO] - Step 1946, Loss: {'train/loss': 0.6719159483909607, 'train/grad_norm_pre_clip': 18.399362564086914, 'timing/data': 0.019608305767178535, 'timing/model': 1.1889272350817919, 'learning_rate/qwen_vl_interface': 8.418715378425078e-06, 'learning_rate/action_model': 8.418715378425078e-05, 'epoch': 0.88}) +[2026-06-17 19:50:48,788][starVLA.training.train_starvla][INFO] - Step 1947, Loss: {'train/loss': 0.9505820870399475, 'train/grad_norm_pre_clip': 14.717804908752441, 'timing/data': 0.016645350493490696, 'timing/model': 1.024078406393528, 'learning_rate/qwen_vl_interface': 8.417103910441364e-06, 'learning_rate/action_model': 8.417103910441363e-05, 'epoch': 0.88}) +[2026-06-17 19:50:52,868][starVLA.training.train_starvla][INFO] - Step 1948, Loss: {'train/loss': 0.36453837156295776, 'train/grad_norm_pre_clip': 11.91460132598877, 'timing/data': 0.026816637255251408, 'timing/model': 1.0764658460393548, 'learning_rate/qwen_vl_interface': 8.415491785913599e-06, 'learning_rate/action_model': 8.4154917859136e-05, 'epoch': 0.88}) +[2026-06-17 19:50:56,951][starVLA.training.train_starvla][INFO] - Step 1949, Loss: {'train/loss': 0.505553662776947, 'train/grad_norm_pre_clip': 12.643284797668457, 'timing/data': 0.014984128065407276, 'timing/model': 0.9455306734889746, 'learning_rate/qwen_vl_interface': 8.413879005175983e-06, 'learning_rate/action_model': 8.413879005175983e-05, 'epoch': 0.88}) +[2026-06-17 19:51:00,958][starVLA.training.train_starvla][INFO] - Step 1950, Loss: {'train/loss': 0.8242735862731934, 'train/grad_norm_pre_clip': 25.441959381103516, 'timing/data': 0.01913042739033699, 'timing/model': 0.9660285543650389, 'learning_rate/qwen_vl_interface': 8.412265568562843e-06, 'learning_rate/action_model': 8.412265568562842e-05, 'epoch': 0.88}) +[2026-06-17 19:51:04,922][starVLA.training.train_starvla][INFO] - Step 1951, Loss: {'train/loss': 0.46884310245513916, 'train/grad_norm_pre_clip': 13.28360366821289, 'timing/data': 0.016892499290406704, 'timing/model': 0.9549406114965677, 'learning_rate/qwen_vl_interface': 8.41065147640865e-06, 'learning_rate/action_model': 8.41065147640865e-05, 'epoch': 0.89}) +[2026-06-17 19:51:08,950][starVLA.training.train_starvla][INFO] - Step 1952, Loss: {'train/loss': 0.3252173364162445, 'train/grad_norm_pre_clip': 10.978777885437012, 'timing/data': 0.017932909540832043, 'timing/model': 1.1831267094239593, 'learning_rate/qwen_vl_interface': 8.409036729048002e-06, 'learning_rate/action_model': 8.409036729048002e-05, 'epoch': 0.89}) +[2026-06-17 19:51:12,922][starVLA.training.train_starvla][INFO] - Step 1953, Loss: {'train/loss': 0.5243217349052429, 'train/grad_norm_pre_clip': 12.161138534545898, 'timing/data': 0.016198648139834404, 'timing/model': 1.1876200791448355, 'learning_rate/qwen_vl_interface': 8.407421326815644e-06, 'learning_rate/action_model': 8.407421326815644e-05, 'epoch': 0.89}) +[2026-06-17 19:51:16,659][starVLA.training.train_starvla][INFO] - Step 1954, Loss: {'train/loss': 0.321887344121933, 'train/grad_norm_pre_clip': 10.545791625976562, 'timing/data': 0.017408528365194798, 'timing/model': 0.96050214022398, 'learning_rate/qwen_vl_interface': 8.405805270046445e-06, 'learning_rate/action_model': 8.405805270046445e-05, 'epoch': 0.89}) +[2026-06-17 19:51:20,598][starVLA.training.train_starvla][INFO] - Step 1955, Loss: {'train/loss': 0.47308897972106934, 'train/grad_norm_pre_clip': 12.18034839630127, 'timing/data': 0.02134215459227562, 'timing/model': 0.9627931574359536, 'learning_rate/qwen_vl_interface': 8.404188559075419e-06, 'learning_rate/action_model': 8.404188559075419e-05, 'epoch': 0.89}) +[2026-06-17 19:51:27,055][starVLA.training.train_starvla][INFO] - Step 1956, Loss: {'train/loss': 0.6150911450386047, 'train/grad_norm_pre_clip': 11.310813903808594, 'timing/data': 0.018653728999197483, 'timing/model': 0.9440041556954384, 'learning_rate/qwen_vl_interface': 8.40257119423771e-06, 'learning_rate/action_model': 8.402571194237709e-05, 'epoch': 0.89}) +[2026-06-17 19:51:31,852][starVLA.training.train_starvla][INFO] - Step 1957, Loss: {'train/loss': 0.6988652348518372, 'train/grad_norm_pre_clip': 14.380536079406738, 'timing/data': 0.02093637641519308, 'timing/model': 0.9880363279953599, 'learning_rate/qwen_vl_interface': 8.4009531758686e-06, 'learning_rate/action_model': 8.4009531758686e-05, 'epoch': 0.89}) +[2026-06-17 19:51:35,825][starVLA.training.train_starvla][INFO] - Step 1958, Loss: {'train/loss': 0.6659327745437622, 'train/grad_norm_pre_clip': 16.189659118652344, 'timing/data': 0.020250876434147358, 'timing/model': 0.9806414702907205, 'learning_rate/qwen_vl_interface': 8.399334504303507e-06, 'learning_rate/action_model': 8.399334504303507e-05, 'epoch': 0.89}) +[2026-06-17 19:51:39,921][starVLA.training.train_starvla][INFO] - Step 1959, Loss: {'train/loss': 0.4392189383506775, 'train/grad_norm_pre_clip': 9.932012557983398, 'timing/data': 0.019672556780278683, 'timing/model': 1.2581754140555859, 'learning_rate/qwen_vl_interface': 8.397715179877982e-06, 'learning_rate/action_model': 8.397715179877982e-05, 'epoch': 0.89}) +[2026-06-17 19:51:43,623][starVLA.training.train_starvla][INFO] - Step 1960, Loss: {'train/loss': 0.4174255132675171, 'train/grad_norm_pre_clip': 9.86240005493164, 'timing/data': 0.019424105994403362, 'timing/model': 0.9655760759487748, 'learning_rate/qwen_vl_interface': 8.396095202927712e-06, 'learning_rate/action_model': 8.396095202927712e-05, 'epoch': 0.89}) +[2026-06-17 19:51:47,649][starVLA.training.train_starvla][INFO] - Step 1961, Loss: {'train/loss': 0.5938798785209656, 'train/grad_norm_pre_clip': 11.623804092407227, 'timing/data': 0.016757838428020477, 'timing/model': 0.956337409093976, 'learning_rate/qwen_vl_interface': 8.394474573788522e-06, 'learning_rate/action_model': 8.394474573788522e-05, 'epoch': 0.89}) +[2026-06-17 19:51:51,899][starVLA.training.train_starvla][INFO] - Step 1962, Loss: {'train/loss': 0.767311155796051, 'train/grad_norm_pre_clip': 19.25852394104004, 'timing/data': 0.01789749786257744, 'timing/model': 1.0299211256206036, 'learning_rate/qwen_vl_interface': 8.39285329279637e-06, 'learning_rate/action_model': 8.39285329279637e-05, 'epoch': 0.89}) +[2026-06-17 19:51:57,066][starVLA.training.train_starvla][INFO] - Step 1963, Loss: {'train/loss': 0.35665830969810486, 'train/grad_norm_pre_clip': 10.110804557800293, 'timing/data': 0.03418524656444788, 'timing/model': 1.0177990421652794, 'learning_rate/qwen_vl_interface': 8.391231360287348e-06, 'learning_rate/action_model': 8.391231360287347e-05, 'epoch': 0.89}) +[2026-06-17 19:52:01,072][starVLA.training.train_starvla][INFO] - Step 1964, Loss: {'train/loss': 0.7694088816642761, 'train/grad_norm_pre_clip': 17.617307662963867, 'timing/data': 0.02310949470847845, 'timing/model': 0.9475177777931094, 'learning_rate/qwen_vl_interface': 8.389608776597684e-06, 'learning_rate/action_model': 8.389608776597684e-05, 'epoch': 0.89}) +[2026-06-17 19:52:05,168][starVLA.training.train_starvla][INFO] - Step 1965, Loss: {'train/loss': 0.43941307067871094, 'train/grad_norm_pre_clip': 10.786429405212402, 'timing/data': 0.015332566574215889, 'timing/model': 0.9628540622070432, 'learning_rate/qwen_vl_interface': 8.387985542063743e-06, 'learning_rate/action_model': 8.387985542063743e-05, 'epoch': 0.89}) +[2026-06-17 19:52:09,239][starVLA.training.train_starvla][INFO] - Step 1966, Loss: {'train/loss': 0.7947632670402527, 'train/grad_norm_pre_clip': 14.348298072814941, 'timing/data': 0.016854163259267807, 'timing/model': 1.2097113709896803, 'learning_rate/qwen_vl_interface': 8.386361657022024e-06, 'learning_rate/action_model': 8.386361657022023e-05, 'epoch': 0.89}) +[2026-06-17 19:52:12,970][starVLA.training.train_starvla][INFO] - Step 1967, Loss: {'train/loss': 0.40414270758628845, 'train/grad_norm_pre_clip': 17.310619354248047, 'timing/data': 0.016802629455924034, 'timing/model': 0.9501828327775002, 'learning_rate/qwen_vl_interface': 8.384737121809156e-06, 'learning_rate/action_model': 8.384737121809156e-05, 'epoch': 0.89}) +[2026-06-17 19:52:17,747][starVLA.training.train_starvla][INFO] - Step 1968, Loss: {'train/loss': 0.3044346868991852, 'train/grad_norm_pre_clip': 11.262682914733887, 'timing/data': 0.021884070709347725, 'timing/model': 0.9478245712816715, 'learning_rate/qwen_vl_interface': 8.38311193676191e-06, 'learning_rate/action_model': 8.383111936761909e-05, 'epoch': 0.89}) +[2026-06-17 19:52:21,929][starVLA.training.train_starvla][INFO] - Step 1969, Loss: {'train/loss': 0.5621835589408875, 'train/grad_norm_pre_clip': 12.372114181518555, 'timing/data': 0.020976711064577103, 'timing/model': 0.962562900967896, 'learning_rate/qwen_vl_interface': 8.381486102217188e-06, 'learning_rate/action_model': 8.381486102217188e-05, 'epoch': 0.89}) +[2026-06-17 19:52:25,934][starVLA.training.train_starvla][INFO] - Step 1970, Loss: {'train/loss': 0.7806084752082825, 'train/grad_norm_pre_clip': 14.70527172088623, 'timing/data': 0.023972860537469387, 'timing/model': 0.9693652195855975, 'learning_rate/qwen_vl_interface': 8.379859618512029e-06, 'learning_rate/action_model': 8.379859618512029e-05, 'epoch': 0.89}) +[2026-06-17 19:52:30,046][starVLA.training.train_starvla][INFO] - Step 1971, Loss: {'train/loss': 0.4582410454750061, 'train/grad_norm_pre_clip': 12.090838432312012, 'timing/data': 0.020820160396397114, 'timing/model': 1.0450296280905604, 'learning_rate/qwen_vl_interface': 8.378232485983603e-06, 'learning_rate/action_model': 8.378232485983603e-05, 'epoch': 0.89}) +[2026-06-17 19:52:34,029][starVLA.training.train_starvla][INFO] - Step 1972, Loss: {'train/loss': 0.386128306388855, 'train/grad_norm_pre_clip': 9.188324928283691, 'timing/data': 0.017710551619529724, 'timing/model': 1.2201272314414382, 'learning_rate/qwen_vl_interface': 8.376604704969216e-06, 'learning_rate/action_model': 8.376604704969216e-05, 'epoch': 0.89}) +[2026-06-17 19:52:38,072][starVLA.training.train_starvla][INFO] - Step 1973, Loss: {'train/loss': 0.7416923642158508, 'train/grad_norm_pre_clip': 18.679481506347656, 'timing/data': 0.026648390106856823, 'timing/model': 1.2448857836425304, 'learning_rate/qwen_vl_interface': 8.374976275806307e-06, 'learning_rate/action_model': 8.374976275806308e-05, 'epoch': 0.9}) +[2026-06-17 19:52:41,914][starVLA.training.train_starvla][INFO] - Step 1974, Loss: {'train/loss': 0.5509013533592224, 'train/grad_norm_pre_clip': 11.979940414428711, 'timing/data': 0.021044819615781307, 'timing/model': 0.972243526019156, 'learning_rate/qwen_vl_interface': 8.373347198832457e-06, 'learning_rate/action_model': 8.373347198832457e-05, 'epoch': 0.9}) +[2026-06-17 19:52:45,890][starVLA.training.train_starvla][INFO] - Step 1975, Loss: {'train/loss': 0.8135175704956055, 'train/grad_norm_pre_clip': 13.419814109802246, 'timing/data': 0.022494145669043064, 'timing/model': 0.9503095149993896, 'learning_rate/qwen_vl_interface': 8.37171747438537e-06, 'learning_rate/action_model': 8.37171747438537e-05, 'epoch': 0.9}) +[2026-06-17 19:52:50,982][starVLA.training.train_starvla][INFO] - Step 1976, Loss: {'train/loss': 0.2732650637626648, 'train/grad_norm_pre_clip': 8.811452865600586, 'timing/data': 0.052709415555000305, 'timing/model': 0.9870267082005739, 'learning_rate/qwen_vl_interface': 8.370087102802892e-06, 'learning_rate/action_model': 8.370087102802892e-05, 'epoch': 0.9}) +[2026-06-17 19:52:55,262][starVLA.training.train_starvla][INFO] - Step 1977, Loss: {'train/loss': 0.686014711856842, 'train/grad_norm_pre_clip': 15.930383682250977, 'timing/data': 0.022376430220901966, 'timing/model': 1.1486705271527171, 'learning_rate/qwen_vl_interface': 8.368456084423e-06, 'learning_rate/action_model': 8.368456084422999e-05, 'epoch': 0.9}) +[2026-06-17 19:52:59,291][starVLA.training.train_starvla][INFO] - Step 1978, Loss: {'train/loss': 0.3131065368652344, 'train/grad_norm_pre_clip': 8.922074317932129, 'timing/data': 0.019144482910633087, 'timing/model': 0.9604733008891344, 'learning_rate/qwen_vl_interface': 8.366824419583806e-06, 'learning_rate/action_model': 8.366824419583806e-05, 'epoch': 0.9}) +[2026-06-17 19:53:03,364][starVLA.training.train_starvla][INFO] - Step 1979, Loss: {'train/loss': 0.5986509323120117, 'train/grad_norm_pre_clip': 15.905149459838867, 'timing/data': 0.016862211748957634, 'timing/model': 1.2853804668411613, 'learning_rate/qwen_vl_interface': 8.365192108623555e-06, 'learning_rate/action_model': 8.365192108623555e-05, 'epoch': 0.9}) +[2026-06-17 19:53:08,951][starVLA.training.train_starvla][INFO] - Step 1980, Loss: {'train/loss': 0.24738554656505585, 'train/grad_norm_pre_clip': 18.05672836303711, 'timing/data': 0.02237865701317787, 'timing/model': 1.2092741914093494, 'learning_rate/qwen_vl_interface': 8.363559151880631e-06, 'learning_rate/action_model': 8.36355915188063e-05, 'epoch': 0.9}) +[2026-06-17 19:53:12,761][starVLA.training.train_starvla][INFO] - Step 1981, Loss: {'train/loss': 0.30990588665008545, 'train/grad_norm_pre_clip': 10.272504806518555, 'timing/data': 0.019092273898422718, 'timing/model': 0.9475510725751519, 'learning_rate/qwen_vl_interface': 8.361925549693541e-06, 'learning_rate/action_model': 8.36192554969354e-05, 'epoch': 0.9}) +[2026-06-17 19:53:16,822][starVLA.training.train_starvla][INFO] - Step 1982, Loss: {'train/loss': 0.29653066396713257, 'train/grad_norm_pre_clip': 8.702327728271484, 'timing/data': 0.0203708503395319, 'timing/model': 0.9840067587792873, 'learning_rate/qwen_vl_interface': 8.360291302400937e-06, 'learning_rate/action_model': 8.360291302400937e-05, 'epoch': 0.9}) +[2026-06-17 19:53:22,810][starVLA.training.train_starvla][INFO] - Step 1983, Loss: {'train/loss': 0.6101768612861633, 'train/grad_norm_pre_clip': 11.872368812561035, 'timing/data': 0.0293718371540308, 'timing/model': 2.9625097773969173, 'learning_rate/qwen_vl_interface': 8.3586564103416e-06, 'learning_rate/action_model': 8.358656410341599e-05, 'epoch': 0.9}) +[2026-06-17 19:53:27,200][starVLA.training.train_starvla][INFO] - Step 1984, Loss: {'train/loss': 0.5876390337944031, 'train/grad_norm_pre_clip': 13.581437110900879, 'timing/data': 0.018070340156555176, 'timing/model': 0.9724262244999409, 'learning_rate/qwen_vl_interface': 8.357020873854445e-06, 'learning_rate/action_model': 8.357020873854444e-05, 'epoch': 0.9}) +[2026-06-17 19:53:31,736][starVLA.training.train_starvla][INFO] - Step 1985, Loss: {'train/loss': 0.5972431898117065, 'train/grad_norm_pre_clip': 13.2681245803833, 'timing/data': 0.02222437597811222, 'timing/model': 0.9626225437968969, 'learning_rate/qwen_vl_interface': 8.355384693278516e-06, 'learning_rate/action_model': 8.355384693278517e-05, 'epoch': 0.9}) +[2026-06-17 19:53:35,894][starVLA.training.train_starvla][INFO] - Step 1986, Loss: {'train/loss': 0.42872586846351624, 'train/grad_norm_pre_clip': 11.601781845092773, 'timing/data': 0.03409203514456749, 'timing/model': 1.3209187518805265, 'learning_rate/qwen_vl_interface': 8.353747868953002e-06, 'learning_rate/action_model': 8.353747868953003e-05, 'epoch': 0.9}) +[2026-06-17 19:53:39,663][starVLA.training.train_starvla][INFO] - Step 1987, Loss: {'train/loss': 0.5216928720474243, 'train/grad_norm_pre_clip': 13.568853378295898, 'timing/data': 0.019349846988916397, 'timing/model': 0.9533343696966767, 'learning_rate/qwen_vl_interface': 8.352110401217213e-06, 'learning_rate/action_model': 8.352110401217213e-05, 'epoch': 0.9}) +[2026-06-17 19:53:44,124][starVLA.training.train_starvla][INFO] - Step 1988, Loss: {'train/loss': 0.6948817372322083, 'train/grad_norm_pre_clip': 15.72152328491211, 'timing/data': 0.02192406915128231, 'timing/model': 1.2372165666893125, 'learning_rate/qwen_vl_interface': 8.350472290410602e-06, 'learning_rate/action_model': 8.350472290410602e-05, 'epoch': 0.9}) +[2026-06-17 19:53:48,021][starVLA.training.train_starvla][INFO] - Step 1989, Loss: {'train/loss': 0.3082098662853241, 'train/grad_norm_pre_clip': 8.72778606414795, 'timing/data': 0.02342881727963686, 'timing/model': 0.9571841740980744, 'learning_rate/qwen_vl_interface': 8.348833536872747e-06, 'learning_rate/action_model': 8.348833536872747e-05, 'epoch': 0.9}) +[2026-06-17 19:53:51,954][starVLA.training.train_starvla][INFO] - Step 1990, Loss: {'train/loss': 0.2934116721153259, 'train/grad_norm_pre_clip': 8.770090103149414, 'timing/data': 0.01720159687101841, 'timing/model': 0.9440976521000266, 'learning_rate/qwen_vl_interface': 8.347194140943366e-06, 'learning_rate/action_model': 8.347194140943366e-05, 'epoch': 0.9}) +[2026-06-17 19:53:56,010][starVLA.training.train_starvla][INFO] - Step 1991, Loss: {'train/loss': 0.6357899904251099, 'train/grad_norm_pre_clip': 15.025155067443848, 'timing/data': 0.02033276855945587, 'timing/model': 1.0350364753976464, 'learning_rate/qwen_vl_interface': 8.345554102962306e-06, 'learning_rate/action_model': 8.345554102962305e-05, 'epoch': 0.9}) +[2026-06-17 19:53:59,929][starVLA.training.train_starvla][INFO] - Step 1992, Loss: {'train/loss': 0.28046178817749023, 'train/grad_norm_pre_clip': 7.720657825469971, 'timing/data': 0.02291214559227228, 'timing/model': 1.1833360251039267, 'learning_rate/qwen_vl_interface': 8.34391342326955e-06, 'learning_rate/action_model': 8.343913423269549e-05, 'epoch': 0.9}) +[2026-06-17 19:54:03,925][starVLA.training.train_starvla][INFO] - Step 1993, Loss: {'train/loss': 0.3279240131378174, 'train/grad_norm_pre_clip': 11.124728202819824, 'timing/data': 0.015210069715976715, 'timing/model': 1.1819551698863506, 'learning_rate/qwen_vl_interface': 8.342272102205213e-06, 'learning_rate/action_model': 8.342272102205212e-05, 'epoch': 0.9}) +[2026-06-17 19:54:07,759][starVLA.training.train_starvla][INFO] - Step 1994, Loss: {'train/loss': 0.5107879042625427, 'train/grad_norm_pre_clip': 14.575014114379883, 'timing/data': 0.02217029221355915, 'timing/model': 0.9572087656706572, 'learning_rate/qwen_vl_interface': 8.340630140109539e-06, 'learning_rate/action_model': 8.340630140109539e-05, 'epoch': 0.9}) +[2026-06-17 19:54:11,816][starVLA.training.train_starvla][INFO] - Step 1995, Loss: {'train/loss': 0.4988403916358948, 'train/grad_norm_pre_clip': 11.386054992675781, 'timing/data': 0.015672706067562103, 'timing/model': 0.9642315292730927, 'learning_rate/qwen_vl_interface': 8.338987537322916e-06, 'learning_rate/action_model': 8.338987537322916e-05, 'epoch': 0.91}) +[2026-06-17 19:54:15,976][starVLA.training.train_starvla][INFO] - Step 1996, Loss: {'train/loss': 0.8318400979042053, 'train/grad_norm_pre_clip': 15.97411060333252, 'timing/data': 0.01782643049955368, 'timing/model': 0.9655107632279396, 'learning_rate/qwen_vl_interface': 8.337344294185848e-06, 'learning_rate/action_model': 8.337344294185849e-05, 'epoch': 0.91}) +[2026-06-17 19:54:19,981][starVLA.training.train_starvla][INFO] - Step 1997, Loss: {'train/loss': 0.6357366442680359, 'train/grad_norm_pre_clip': 14.447932243347168, 'timing/data': 0.01564173214137554, 'timing/model': 0.9538163468241692, 'learning_rate/qwen_vl_interface': 8.335700411038988e-06, 'learning_rate/action_model': 8.335700411038988e-05, 'epoch': 0.91}) +[2026-06-17 19:54:25,230][starVLA.training.train_starvla][INFO] - Step 1998, Loss: {'train/loss': 0.6211718916893005, 'train/grad_norm_pre_clip': 15.344414710998535, 'timing/data': 0.01712966151535511, 'timing/model': 0.9585508080199361, 'learning_rate/qwen_vl_interface': 8.334055888223111e-06, 'learning_rate/action_model': 8.334055888223111e-05, 'epoch': 0.91}) +[2026-06-17 19:54:29,180][starVLA.training.train_starvla][INFO] - Step 1999, Loss: {'train/loss': 0.3180904984474182, 'train/grad_norm_pre_clip': 10.384576797485352, 'timing/data': 0.017778396606445312, 'timing/model': 1.1926208985969424, 'learning_rate/qwen_vl_interface': 8.332410726079132e-06, 'learning_rate/action_model': 8.332410726079132e-05, 'epoch': 0.91}) +[2026-06-17 19:56:42,866][starVLA.training.train_starvla][INFO] - Step 2000, Loss: {'train/loss': 0.3659229278564453, 'train/grad_norm_pre_clip': 11.995805740356445, 'eval/loss': 0.4457814693450928, 'eval/latency_8/loss': 0.4457814693450928, 'eval/demon_fire_precision': 0.9301361752516282, 'eval/demon_fire_recall': 0.9760795278036657, 'eval/demon_fire_f1': 0.9525541912990753, 'eval/demon_fire_rate_model': 0.9398998330550918, 'eval/demon_fire_rate_teacher': 0.8956594323873122, 'eval/demon_left_precision': 0.748730964467005, 'eval/demon_left_recall': 0.6316916488222698, 'eval/demon_left_f1': 0.6852497096399536, 'eval/demon_left_rate_model': 0.10962715637173066, 'eval/demon_left_rate_teacher': 0.12993878686700056, 'eval/demon_right_precision': 0.7296898079763663, 'eval/demon_right_recall': 0.7484848484848485, 'eval/demon_right_f1': 0.7389678384442783, 'eval/demon_right_rate_model': 0.18836950473010572, 'eval/demon_right_rate_teacher': 0.18363939899833054, 'eval/demon_move_f1': 0.7121087740421159, 'eval/demon_cc_f1': 0.8649394601082588, 'eval/latency_8/demon_fire_precision': 0.9301361752516282, 'eval/latency_8/demon_fire_recall': 0.9760795278036657, 'eval/latency_8/demon_fire_f1': 0.9525541912990753, 'eval/latency_8/demon_fire_rate_model': 0.9398998330550918, 'eval/latency_8/demon_fire_rate_teacher': 0.8956594323873122, 'eval/latency_8/demon_left_precision': 0.748730964467005, 'eval/latency_8/demon_left_recall': 0.6316916488222698, 'eval/latency_8/demon_left_f1': 0.6852497096399536, 'eval/latency_8/demon_left_rate_model': 0.10962715637173066, 'eval/latency_8/demon_left_rate_teacher': 0.12993878686700056, 'eval/latency_8/demon_right_precision': 0.7296898079763663, 'eval/latency_8/demon_right_recall': 0.7484848484848485, 'eval/latency_8/demon_right_f1': 0.7389678384442783, 'eval/latency_8/demon_right_rate_model': 0.18836950473010572, 'eval/latency_8/demon_right_rate_teacher': 0.18363939899833054, 'eval/latency_8/demon_move_f1': 0.7121087740421159, 'eval/latency_8/demon_cc_f1': 0.8649394601082588, 'timing/data': 0.02213982306420803, 'timing/model': 1.2739678649231791, 'learning_rate/qwen_vl_interface': 8.330764924948093e-06, 'learning_rate/action_model': 8.330764924948093e-05, 'epoch': 0.91}) +[2026-06-17 19:56:42,884][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_2000_state +[2026-06-17 19:56:42,892][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 19:56:49,869][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_2000_state/model.safetensors +[2026-06-17 19:57:06,662][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_2000_state/optimizer.bin +[2026-06-17 19:57:06,664][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_2000_state/sampler.bin +[2026-06-17 19:57:06,664][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_2000_state/sampler_1.bin +[2026-06-17 19:57:06,667][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_2000_state/random_states_0.pkl +[2026-06-17 19:57:06,678][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2000 +[2026-06-17 19:57:06,678][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 19:57:06,692][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 19:57:10,678][starVLA.training.train_starvla][INFO] - Step 2001, Loss: {'train/loss': 0.32451263070106506, 'train/grad_norm_pre_clip': 9.638528823852539, 'timing/data': 0.021023050881922245, 'timing/model': 1.2211774699389935, 'learning_rate/qwen_vl_interface': 8.329118485171167e-06, 'learning_rate/action_model': 8.329118485171167e-05, 'epoch': 0.91}) +[2026-06-17 19:57:14,723][starVLA.training.train_starvla][INFO] - Step 2002, Loss: {'train/loss': 0.6009351015090942, 'train/grad_norm_pre_clip': 12.241329193115234, 'timing/data': 0.021676610223948956, 'timing/model': 1.2197826392948627, 'learning_rate/qwen_vl_interface': 8.327471407089667e-06, 'learning_rate/action_model': 8.327471407089668e-05, 'epoch': 0.91}) +[2026-06-17 19:57:18,504][starVLA.training.train_starvla][INFO] - Step 2003, Loss: {'train/loss': 0.5200113654136658, 'train/grad_norm_pre_clip': 11.851691246032715, 'timing/data': 0.026574131101369858, 'timing/model': 0.9256680924445391, 'learning_rate/qwen_vl_interface': 8.325823691045032e-06, 'learning_rate/action_model': 8.325823691045033e-05, 'epoch': 0.91}) +[2026-06-17 19:57:22,462][starVLA.training.train_starvla][INFO] - Step 2004, Loss: {'train/loss': 0.46426233649253845, 'train/grad_norm_pre_clip': 10.5828218460083, 'timing/data': 0.022347970865666866, 'timing/model': 0.9366346159949899, 'learning_rate/qwen_vl_interface': 8.32417533737884e-06, 'learning_rate/action_model': 8.324175337378839e-05, 'epoch': 0.91}) +[2026-06-17 19:57:26,478][starVLA.training.train_starvla][INFO] - Step 2005, Loss: {'train/loss': 0.7877135872840881, 'train/grad_norm_pre_clip': 19.392488479614258, 'timing/data': 0.025045360438525677, 'timing/model': 0.934936199337244, 'learning_rate/qwen_vl_interface': 8.322526346432789e-06, 'learning_rate/action_model': 8.322526346432789e-05, 'epoch': 0.91}) +[2026-06-17 19:57:30,427][starVLA.training.train_starvla][INFO] - Step 2006, Loss: {'train/loss': 0.38092589378356934, 'train/grad_norm_pre_clip': 8.750384330749512, 'timing/data': 0.015278303995728493, 'timing/model': 0.9460206916555762, 'learning_rate/qwen_vl_interface': 8.32087671854872e-06, 'learning_rate/action_model': 8.32087671854872e-05, 'epoch': 0.91}) +[2026-06-17 19:57:34,346][starVLA.training.train_starvla][INFO] - Step 2007, Loss: {'train/loss': 0.27817708253860474, 'train/grad_norm_pre_clip': 8.54544448852539, 'timing/data': 0.024932626634836197, 'timing/model': 0.9548824727535248, 'learning_rate/qwen_vl_interface': 8.319226454068604e-06, 'learning_rate/action_model': 8.319226454068604e-05, 'epoch': 0.91}) +[2026-06-17 19:57:38,567][starVLA.training.train_starvla][INFO] - Step 2008, Loss: {'train/loss': 0.9046276807785034, 'train/grad_norm_pre_clip': 15.181180000305176, 'timing/data': 0.024376722052693367, 'timing/model': 1.1783556146547198, 'learning_rate/qwen_vl_interface': 8.317575553334542e-06, 'learning_rate/action_model': 8.317575553334542e-05, 'epoch': 0.91}) +[2026-06-17 19:57:43,724][starVLA.training.train_starvla][INFO] - Step 2009, Loss: {'train/loss': 0.45926302671432495, 'train/grad_norm_pre_clip': 20.44705581665039, 'timing/data': 0.03706034738570452, 'timing/model': 1.0174607811495662, 'learning_rate/qwen_vl_interface': 8.315924016688767e-06, 'learning_rate/action_model': 8.315924016688766e-05, 'epoch': 0.91}) +[2026-06-17 19:57:48,645][starVLA.training.train_starvla][INFO] - Step 2010, Loss: {'train/loss': 0.4651030898094177, 'train/grad_norm_pre_clip': 12.605935096740723, 'timing/data': 0.023851096630096436, 'timing/model': 1.9430430056527257, 'learning_rate/qwen_vl_interface': 8.314271844473644e-06, 'learning_rate/action_model': 8.314271844473644e-05, 'epoch': 0.91}) +[2026-06-17 19:57:52,551][starVLA.training.train_starvla][INFO] - Step 2011, Loss: {'train/loss': 0.4564824104309082, 'train/grad_norm_pre_clip': 11.528231620788574, 'timing/data': 0.015276423655450344, 'timing/model': 0.9440945824608207, 'learning_rate/qwen_vl_interface': 8.312619037031668e-06, 'learning_rate/action_model': 8.312619037031668e-05, 'epoch': 0.91}) +[2026-06-17 19:57:56,434][starVLA.training.train_starvla][INFO] - Step 2012, Loss: {'train/loss': 0.36258983612060547, 'train/grad_norm_pre_clip': 10.774982452392578, 'timing/data': 0.017099203541874886, 'timing/model': 0.9578972766175866, 'learning_rate/qwen_vl_interface': 8.310965594705473e-06, 'learning_rate/action_model': 8.310965594705472e-05, 'epoch': 0.91}) +[2026-06-17 19:58:00,409][starVLA.training.train_starvla][INFO] - Step 2013, Loss: {'train/loss': 0.5153777003288269, 'train/grad_norm_pre_clip': 14.861459732055664, 'timing/data': 0.019564767368137836, 'timing/model': 0.989193495362997, 'learning_rate/qwen_vl_interface': 8.309311517837814e-06, 'learning_rate/action_model': 8.309311517837813e-05, 'epoch': 0.91}) +[2026-06-17 19:58:04,344][starVLA.training.train_starvla][INFO] - Step 2014, Loss: {'train/loss': 0.5249341130256653, 'train/grad_norm_pre_clip': 15.13383960723877, 'timing/data': 0.02362880203872919, 'timing/model': 1.2237999057397246, 'learning_rate/qwen_vl_interface': 8.307656806771587e-06, 'learning_rate/action_model': 8.307656806771588e-05, 'epoch': 0.91}) +[2026-06-17 19:58:08,368][starVLA.training.train_starvla][INFO] - Step 2015, Loss: {'train/loss': 0.6531472206115723, 'train/grad_norm_pre_clip': 14.2391939163208, 'timing/data': 0.016839204356074333, 'timing/model': 1.137249800376594, 'learning_rate/qwen_vl_interface': 8.306001461849814e-06, 'learning_rate/action_model': 8.306001461849814e-05, 'epoch': 0.91}) +[2026-06-17 19:58:12,118][starVLA.training.train_starvla][INFO] - Step 2016, Loss: {'train/loss': 0.5014057159423828, 'train/grad_norm_pre_clip': 13.630659103393555, 'timing/data': 0.015596424229443073, 'timing/model': 1.0003023575991392, 'learning_rate/qwen_vl_interface': 8.30434548341565e-06, 'learning_rate/action_model': 8.304345483415649e-05, 'epoch': 0.91}) +[2026-06-17 19:58:16,075][starVLA.training.train_starvla][INFO] - Step 2017, Loss: {'train/loss': 0.46510422229766846, 'train/grad_norm_pre_clip': 15.7127046585083, 'timing/data': 0.01871880330145359, 'timing/model': 0.9548003012314439, 'learning_rate/qwen_vl_interface': 8.30268887181238e-06, 'learning_rate/action_model': 8.302688871812378e-05, 'epoch': 0.92}) +[2026-06-17 19:58:20,173][starVLA.training.train_starvla][INFO] - Step 2018, Loss: {'train/loss': 0.36049556732177734, 'train/grad_norm_pre_clip': 10.25208568572998, 'timing/data': 0.0298324478790164, 'timing/model': 0.9900735225528479, 'learning_rate/qwen_vl_interface': 8.301031627383422e-06, 'learning_rate/action_model': 8.301031627383421e-05, 'epoch': 0.92}) +[2026-06-17 19:58:24,242][starVLA.training.train_starvla][INFO] - Step 2019, Loss: {'train/loss': 0.45329907536506653, 'train/grad_norm_pre_clip': 15.661718368530273, 'timing/data': 0.01953638531267643, 'timing/model': 0.9754564901813865, 'learning_rate/qwen_vl_interface': 8.299373750472323e-06, 'learning_rate/action_model': 8.299373750472324e-05, 'epoch': 0.92}) +[2026-06-17 19:58:28,256][starVLA.training.train_starvla][INFO] - Step 2020, Loss: {'train/loss': 0.22323793172836304, 'train/grad_norm_pre_clip': 6.842510223388672, 'timing/data': 0.01640887837857008, 'timing/model': 0.9472495010122657, 'learning_rate/qwen_vl_interface': 8.297715241422765e-06, 'learning_rate/action_model': 8.297715241422766e-05, 'epoch': 0.92}) +[2026-06-17 19:58:32,159][starVLA.training.train_starvla][INFO] - Step 2021, Loss: {'train/loss': 0.44229066371917725, 'train/grad_norm_pre_clip': 36.9635009765625, 'timing/data': 0.021073536016047, 'timing/model': 1.1743687940761447, 'learning_rate/qwen_vl_interface': 8.29605610057856e-06, 'learning_rate/action_model': 8.29605610057856e-05, 'epoch': 0.92}) +[2026-06-17 19:58:35,955][starVLA.training.train_starvla][INFO] - Step 2022, Loss: {'train/loss': 0.7234735488891602, 'train/grad_norm_pre_clip': 18.49201011657715, 'timing/data': 0.01691504195332527, 'timing/model': 0.9572473177686334, 'learning_rate/qwen_vl_interface': 8.294396328283647e-06, 'learning_rate/action_model': 8.294396328283645e-05, 'epoch': 0.92}) +[2026-06-17 19:58:39,938][starVLA.training.train_starvla][INFO] - Step 2023, Loss: {'train/loss': 0.570315957069397, 'train/grad_norm_pre_clip': 11.718428611755371, 'timing/data': 0.03226672671735287, 'timing/model': 0.9656328046694398, 'learning_rate/qwen_vl_interface': 8.292735924882097e-06, 'learning_rate/action_model': 8.292735924882097e-05, 'epoch': 0.92}) +[2026-06-17 19:58:43,799][starVLA.training.train_starvla][INFO] - Step 2024, Loss: {'train/loss': 0.48645445704460144, 'train/grad_norm_pre_clip': 9.448406219482422, 'timing/data': 0.01782365795224905, 'timing/model': 0.9543017596006393, 'learning_rate/qwen_vl_interface': 8.291074890718117e-06, 'learning_rate/action_model': 8.291074890718117e-05, 'epoch': 0.92}) +[2026-06-17 19:58:47,791][starVLA.training.train_starvla][INFO] - Step 2025, Loss: {'train/loss': 0.34749171137809753, 'train/grad_norm_pre_clip': 9.300333023071289, 'timing/data': 0.024618047289550304, 'timing/model': 0.9995399639010429, 'learning_rate/qwen_vl_interface': 8.28941322613604e-06, 'learning_rate/action_model': 8.289413226136039e-05, 'epoch': 0.92}) +[2026-06-17 19:58:51,670][starVLA.training.train_starvla][INFO] - Step 2026, Loss: {'train/loss': 0.27445268630981445, 'train/grad_norm_pre_clip': 9.601261138916016, 'timing/data': 0.022079874761402607, 'timing/model': 0.9451153259724379, 'learning_rate/qwen_vl_interface': 8.287750931480332e-06, 'learning_rate/action_model': 8.287750931480332e-05, 'epoch': 0.92}) +[2026-06-17 19:58:55,667][starVLA.training.train_starvla][INFO] - Step 2027, Loss: {'train/loss': 0.7298442125320435, 'train/grad_norm_pre_clip': 18.269376754760742, 'timing/data': 0.022284021601080894, 'timing/model': 0.9439014503732324, 'learning_rate/qwen_vl_interface': 8.286088007095584e-06, 'learning_rate/action_model': 8.286088007095584e-05, 'epoch': 0.92}) +[2026-06-17 19:58:59,653][starVLA.training.train_starvla][INFO] - Step 2028, Loss: {'train/loss': 0.3462049663066864, 'train/grad_norm_pre_clip': 9.432635307312012, 'timing/data': 0.01986867841333151, 'timing/model': 1.237570209428668, 'learning_rate/qwen_vl_interface': 8.284424453326528e-06, 'learning_rate/action_model': 8.284424453326528e-05, 'epoch': 0.92}) +[2026-06-17 19:59:03,495][starVLA.training.train_starvla][INFO] - Step 2029, Loss: {'train/loss': 0.8296192288398743, 'train/grad_norm_pre_clip': 18.30294418334961, 'timing/data': 0.022901320829987526, 'timing/model': 1.007535700686276, 'learning_rate/qwen_vl_interface': 8.282760270518017e-06, 'learning_rate/action_model': 8.282760270518017e-05, 'epoch': 0.92}) +[2026-06-17 19:59:07,353][starVLA.training.train_starvla][INFO] - Step 2030, Loss: {'train/loss': 0.5499134659767151, 'train/grad_norm_pre_clip': 11.673744201660156, 'timing/data': 0.015564989298582077, 'timing/model': 0.9408191004768014, 'learning_rate/qwen_vl_interface': 8.281095459015037e-06, 'learning_rate/action_model': 8.281095459015036e-05, 'epoch': 0.92}) +[2026-06-17 19:59:11,639][starVLA.training.train_starvla][INFO] - Step 2031, Loss: {'train/loss': 0.3164847195148468, 'train/grad_norm_pre_clip': 9.56779956817627, 'timing/data': 0.01579912193119526, 'timing/model': 1.358158327639103, 'learning_rate/qwen_vl_interface': 8.279430019162709e-06, 'learning_rate/action_model': 8.279430019162708e-05, 'epoch': 0.92}) +[2026-06-17 19:59:15,633][starVLA.training.train_starvla][INFO] - Step 2032, Loss: {'train/loss': 0.48281341791152954, 'train/grad_norm_pre_clip': 10.620349884033203, 'timing/data': 0.019362538121640682, 'timing/model': 0.968289447017014, 'learning_rate/qwen_vl_interface': 8.277763951306277e-06, 'learning_rate/action_model': 8.277763951306277e-05, 'epoch': 0.92}) +[2026-06-17 19:59:19,542][starVLA.training.train_starvla][INFO] - Step 2033, Loss: {'train/loss': 0.2649286985397339, 'train/grad_norm_pre_clip': 7.908664226531982, 'timing/data': 0.020578172989189625, 'timing/model': 0.9366526082158089, 'learning_rate/qwen_vl_interface': 8.27609725579112e-06, 'learning_rate/action_model': 8.27609725579112e-05, 'epoch': 0.92}) +[2026-06-17 19:59:26,360][starVLA.training.train_starvla][INFO] - Step 2034, Loss: {'train/loss': 0.6338521838188171, 'train/grad_norm_pre_clip': 11.574653625488281, 'timing/data': 0.019543616101145744, 'timing/model': 1.1773693105205894, 'learning_rate/qwen_vl_interface': 8.274429932962745e-06, 'learning_rate/action_model': 8.274429932962746e-05, 'epoch': 0.92}) +[2026-06-17 19:59:30,222][starVLA.training.train_starvla][INFO] - Step 2035, Loss: {'train/loss': 0.48600882291793823, 'train/grad_norm_pre_clip': 11.510836601257324, 'timing/data': 0.017942998558282852, 'timing/model': 1.1102573666721582, 'learning_rate/qwen_vl_interface': 8.272761983166793e-06, 'learning_rate/action_model': 8.272761983166792e-05, 'epoch': 0.92}) +[2026-06-17 19:59:33,931][starVLA.training.train_starvla][INFO] - Step 2036, Loss: {'train/loss': 0.2362176775932312, 'train/grad_norm_pre_clip': 9.657391548156738, 'timing/data': 0.022315382957458496, 'timing/model': 0.9449725253507495, 'learning_rate/qwen_vl_interface': 8.271093406749025e-06, 'learning_rate/action_model': 8.271093406749025e-05, 'epoch': 0.92}) +[2026-06-17 19:59:37,805][starVLA.training.train_starvla][INFO] - Step 2037, Loss: {'train/loss': 0.606562614440918, 'train/grad_norm_pre_clip': 11.009383201599121, 'timing/data': 0.019015520811080933, 'timing/model': 0.9632786074653268, 'learning_rate/qwen_vl_interface': 8.269424204055346e-06, 'learning_rate/action_model': 8.269424204055346e-05, 'epoch': 0.92}) +[2026-06-17 19:59:41,637][starVLA.training.train_starvla][INFO] - Step 2038, Loss: {'train/loss': 0.8872393369674683, 'train/grad_norm_pre_clip': 24.677162170410156, 'timing/data': 0.017934763804078102, 'timing/model': 0.9412165563553572, 'learning_rate/qwen_vl_interface': 8.26775437543178e-06, 'learning_rate/action_model': 8.26775437543178e-05, 'epoch': 0.92}) +[2026-06-17 19:59:45,517][starVLA.training.train_starvla][INFO] - Step 2039, Loss: {'train/loss': 0.3406446576118469, 'train/grad_norm_pre_clip': 9.820125579833984, 'timing/data': 0.027616274543106556, 'timing/model': 0.9528254680335522, 'learning_rate/qwen_vl_interface': 8.266083921224484e-06, 'learning_rate/action_model': 8.266083921224485e-05, 'epoch': 0.93}) +[2026-06-17 19:59:49,338][starVLA.training.train_starvla][INFO] - Step 2040, Loss: {'train/loss': 0.38497060537338257, 'train/grad_norm_pre_clip': 10.397991180419922, 'timing/data': 0.022802269086241722, 'timing/model': 0.9518219418823719, 'learning_rate/qwen_vl_interface': 8.264412841779747e-06, 'learning_rate/action_model': 8.264412841779746e-05, 'epoch': 0.93}) +[2026-06-17 19:59:53,210][starVLA.training.train_starvla][INFO] - Step 2041, Loss: {'train/loss': 0.35900652408599854, 'train/grad_norm_pre_clip': 9.768451690673828, 'timing/data': 0.02025850210338831, 'timing/model': 1.13946452178061, 'learning_rate/qwen_vl_interface': 8.262741137443984e-06, 'learning_rate/action_model': 8.262741137443985e-05, 'epoch': 0.93}) +[2026-06-17 19:59:57,249][starVLA.training.train_starvla][INFO] - Step 2042, Loss: {'train/loss': 0.3230065703392029, 'train/grad_norm_pre_clip': 9.396960258483887, 'timing/data': 0.015176471322774887, 'timing/model': 1.0858331052586436, 'learning_rate/qwen_vl_interface': 8.26106880856374e-06, 'learning_rate/action_model': 8.26106880856374e-05, 'epoch': 0.93}) +[2026-06-17 20:00:00,983][starVLA.training.train_starvla][INFO] - Step 2043, Loss: {'train/loss': 0.3716953992843628, 'train/grad_norm_pre_clip': 11.422203063964844, 'timing/data': 0.024617711082100868, 'timing/model': 0.9495995938777924, 'learning_rate/qwen_vl_interface': 8.259395855485693e-06, 'learning_rate/action_model': 8.259395855485694e-05, 'epoch': 0.93}) +[2026-06-17 20:00:04,960][starVLA.training.train_starvla][INFO] - Step 2044, Loss: {'train/loss': 0.31793731451034546, 'train/grad_norm_pre_clip': 11.013360023498535, 'timing/data': 0.013156112283468246, 'timing/model': 0.9466131748631597, 'learning_rate/qwen_vl_interface': 8.257722278556647e-06, 'learning_rate/action_model': 8.257722278556646e-05, 'epoch': 0.93}) +[2026-06-17 20:00:09,020][starVLA.training.train_starvla][INFO] - Step 2045, Loss: {'train/loss': 0.34518370032310486, 'train/grad_norm_pre_clip': 9.025081634521484, 'timing/data': 0.013983689248561859, 'timing/model': 1.0305156335234642, 'learning_rate/qwen_vl_interface': 8.256048078123536e-06, 'learning_rate/action_model': 8.256048078123535e-05, 'epoch': 0.93}) +[2026-06-17 20:00:12,944][starVLA.training.train_starvla][INFO] - Step 2046, Loss: {'train/loss': 0.33054542541503906, 'train/grad_norm_pre_clip': 15.89875602722168, 'timing/data': 0.01566418632864952, 'timing/model': 0.9381181644275784, 'learning_rate/qwen_vl_interface': 8.254373254533424e-06, 'learning_rate/action_model': 8.254373254533423e-05, 'epoch': 0.93}) +[2026-06-17 20:00:16,840][starVLA.training.train_starvla][INFO] - Step 2047, Loss: {'train/loss': 0.5928642749786377, 'train/grad_norm_pre_clip': 14.257157325744629, 'timing/data': 0.02575306035578251, 'timing/model': 0.945876426063478, 'learning_rate/qwen_vl_interface': 8.252697808133501e-06, 'learning_rate/action_model': 8.252697808133501e-05, 'epoch': 0.93}) +[2026-06-17 20:00:22,290][starVLA.training.train_starvla][INFO] - Step 2048, Loss: {'train/loss': 0.6243588924407959, 'train/grad_norm_pre_clip': 15.121397972106934, 'timing/data': 0.02527032420039177, 'timing/model': 1.3735995516180992, 'learning_rate/qwen_vl_interface': 8.251021739271095e-06, 'learning_rate/action_model': 8.251021739271094e-05, 'epoch': 0.93}) +[2026-06-17 20:00:26,105][starVLA.training.train_starvla][INFO] - Step 2049, Loss: {'train/loss': 0.5120324492454529, 'train/grad_norm_pre_clip': 15.803930282592773, 'timing/data': 0.023446930572390556, 'timing/model': 1.0250558638945222, 'learning_rate/qwen_vl_interface': 8.249345048293652e-06, 'learning_rate/action_model': 8.24934504829365e-05, 'epoch': 0.93}) +[2026-06-17 20:00:30,119][starVLA.training.train_starvla][INFO] - Step 2050, Loss: {'train/loss': 0.4379776418209076, 'train/grad_norm_pre_clip': 13.670842170715332, 'timing/data': 0.01822660770267248, 'timing/model': 1.0326461531221867, 'learning_rate/qwen_vl_interface': 8.247667735548753e-06, 'learning_rate/action_model': 8.247667735548752e-05, 'epoch': 0.93}) +[2026-06-17 20:00:34,164][starVLA.training.train_starvla][INFO] - Step 2051, Loss: {'train/loss': 0.5605365633964539, 'train/grad_norm_pre_clip': 17.180444717407227, 'timing/data': 0.015948518179357052, 'timing/model': 0.9657153012230992, 'learning_rate/qwen_vl_interface': 8.245989801384104e-06, 'learning_rate/action_model': 8.245989801384104e-05, 'epoch': 0.93}) +[2026-06-17 20:00:38,216][starVLA.training.train_starvla][INFO] - Step 2052, Loss: {'train/loss': 0.37548184394836426, 'train/grad_norm_pre_clip': 12.459065437316895, 'timing/data': 0.02136768400669098, 'timing/model': 0.952350021339953, 'learning_rate/qwen_vl_interface': 8.24431124614755e-06, 'learning_rate/action_model': 8.244311246147549e-05, 'epoch': 0.93}) +[2026-06-17 20:00:42,179][starVLA.training.train_starvla][INFO] - Step 2053, Loss: {'train/loss': 0.44815129041671753, 'train/grad_norm_pre_clip': 11.532601356506348, 'timing/data': 0.02074086293578148, 'timing/model': 0.9615562995895743, 'learning_rate/qwen_vl_interface': 8.24263207018705e-06, 'learning_rate/action_model': 8.24263207018705e-05, 'epoch': 0.93}) +[2026-06-17 20:00:46,113][starVLA.training.train_starvla][INFO] - Step 2054, Loss: {'train/loss': 0.3675394654273987, 'train/grad_norm_pre_clip': 12.6084623336792, 'timing/data': 0.020353170111775398, 'timing/model': 1.1927443062886596, 'learning_rate/qwen_vl_interface': 8.2409522738507e-06, 'learning_rate/action_model': 8.240952273850701e-05, 'epoch': 0.93}) +[2026-06-17 20:00:50,218][starVLA.training.train_starvla][INFO] - Step 2055, Loss: {'train/loss': 0.708965003490448, 'train/grad_norm_pre_clip': 14.498103141784668, 'timing/data': 0.017372251488268375, 'timing/model': 1.245109767653048, 'learning_rate/qwen_vl_interface': 8.23927185748673e-06, 'learning_rate/action_model': 8.239271857486729e-05, 'epoch': 0.93}) +[2026-06-17 20:00:55,713][starVLA.training.train_starvla][INFO] - Step 2056, Loss: {'train/loss': 0.3986157774925232, 'train/grad_norm_pre_clip': 10.277142524719238, 'timing/data': 0.03208199702203274, 'timing/model': 1.6099589867517352, 'learning_rate/qwen_vl_interface': 8.237590821443484e-06, 'learning_rate/action_model': 8.237590821443483e-05, 'epoch': 0.93}) +[2026-06-17 20:00:59,744][starVLA.training.train_starvla][INFO] - Step 2057, Loss: {'train/loss': 0.27118051052093506, 'train/grad_norm_pre_clip': 7.836912631988525, 'timing/data': 0.016144181601703167, 'timing/model': 0.9372390601783991, 'learning_rate/qwen_vl_interface': 8.235909166069447e-06, 'learning_rate/action_model': 8.235909166069447e-05, 'epoch': 0.93}) +[2026-06-17 20:01:04,033][starVLA.training.train_starvla][INFO] - Step 2058, Loss: {'train/loss': 0.20755240321159363, 'train/grad_norm_pre_clip': 9.606938362121582, 'timing/data': 0.014490732923150063, 'timing/model': 0.9332104222849011, 'learning_rate/qwen_vl_interface': 8.234226891713225e-06, 'learning_rate/action_model': 8.234226891713225e-05, 'epoch': 0.93}) +[2026-06-17 20:01:07,898][starVLA.training.train_starvla][INFO] - Step 2059, Loss: {'train/loss': 0.3762420117855072, 'train/grad_norm_pre_clip': 9.424849510192871, 'timing/data': 0.027195447124540806, 'timing/model': 0.9510115459561348, 'learning_rate/qwen_vl_interface': 8.232543998723559e-06, 'learning_rate/action_model': 8.232543998723559e-05, 'epoch': 0.93}) +[2026-06-17 20:01:11,732][starVLA.training.train_starvla][INFO] - Step 2060, Loss: {'train/loss': 0.5927695631980896, 'train/grad_norm_pre_clip': 38.518741607666016, 'timing/data': 0.02040009293705225, 'timing/model': 0.935110286809504, 'learning_rate/qwen_vl_interface': 8.230860487449314e-06, 'learning_rate/action_model': 8.230860487449313e-05, 'epoch': 0.93}) +[2026-06-17 20:01:15,636][starVLA.training.train_starvla][INFO] - Step 2061, Loss: {'train/loss': 0.3183657228946686, 'train/grad_norm_pre_clip': 10.571474075317383, 'timing/data': 0.021323061548173428, 'timing/model': 1.1378240324556828, 'learning_rate/qwen_vl_interface': 8.22917635823948e-06, 'learning_rate/action_model': 8.229176358239478e-05, 'epoch': 0.94}) +[2026-06-17 20:01:20,565][starVLA.training.train_starvla][INFO] - Step 2062, Loss: {'train/loss': 1.1581064462661743, 'train/grad_norm_pre_clip': 22.978473663330078, 'timing/data': 0.024116124026477337, 'timing/model': 2.15268350020051, 'learning_rate/qwen_vl_interface': 8.227491611443182e-06, 'learning_rate/action_model': 8.227491611443182e-05, 'epoch': 0.94}) +[2026-06-17 20:01:25,408][starVLA.training.train_starvla][INFO] - Step 2063, Loss: {'train/loss': 0.39891383051872253, 'train/grad_norm_pre_clip': 13.096367835998535, 'timing/data': 0.02416418492794037, 'timing/model': 2.0219396902248263, 'learning_rate/qwen_vl_interface': 8.225806247409667e-06, 'learning_rate/action_model': 8.225806247409668e-05, 'epoch': 0.94}) +[2026-06-17 20:01:29,347][starVLA.training.train_starvla][INFO] - Step 2064, Loss: {'train/loss': 0.7993646860122681, 'train/grad_norm_pre_clip': 18.30926513671875, 'timing/data': 0.01934842113405466, 'timing/model': 0.9542487831786275, 'learning_rate/qwen_vl_interface': 8.224120266488318e-06, 'learning_rate/action_model': 8.224120266488318e-05, 'epoch': 0.94}) +[2026-06-17 20:01:33,408][starVLA.training.train_starvla][INFO] - Step 2065, Loss: {'train/loss': 0.5858176350593567, 'train/grad_norm_pre_clip': 12.889198303222656, 'timing/data': 0.03243176266551018, 'timing/model': 0.949744114652276, 'learning_rate/qwen_vl_interface': 8.222433669028636e-06, 'learning_rate/action_model': 8.222433669028635e-05, 'epoch': 0.94}) +[2026-06-17 20:01:37,347][starVLA.training.train_starvla][INFO] - Step 2066, Loss: {'train/loss': 0.568679690361023, 'train/grad_norm_pre_clip': 12.087934494018555, 'timing/data': 0.02695891261100769, 'timing/model': 0.9579412965103984, 'learning_rate/qwen_vl_interface': 8.220746455380257e-06, 'learning_rate/action_model': 8.220746455380257e-05, 'epoch': 0.94}) +[2026-06-17 20:01:41,421][starVLA.training.train_starvla][INFO] - Step 2067, Loss: {'train/loss': 0.6929177045822144, 'train/grad_norm_pre_clip': 16.71147346496582, 'timing/data': 0.021894275210797787, 'timing/model': 1.0132109764963388, 'learning_rate/qwen_vl_interface': 8.219058625892939e-06, 'learning_rate/action_model': 8.21905862589294e-05, 'epoch': 0.94}) +[2026-06-17 20:01:45,439][starVLA.training.train_starvla][INFO] - Step 2068, Loss: {'train/loss': 0.6103934645652771, 'train/grad_norm_pre_clip': 15.228550910949707, 'timing/data': 0.02077651210129261, 'timing/model': 1.2215273389592767, 'learning_rate/qwen_vl_interface': 8.217370180916574e-06, 'learning_rate/action_model': 8.217370180916575e-05, 'epoch': 0.94}) +[2026-06-17 20:01:49,198][starVLA.training.train_starvla][INFO] - Step 2069, Loss: {'train/loss': 0.5688715577125549, 'train/grad_norm_pre_clip': 13.291532516479492, 'timing/data': 0.017888334579765797, 'timing/model': 0.9726568032056093, 'learning_rate/qwen_vl_interface': 8.215681120801179e-06, 'learning_rate/action_model': 8.215681120801179e-05, 'epoch': 0.94}) +[2026-06-17 20:01:53,262][starVLA.training.train_starvla][INFO] - Step 2070, Loss: {'train/loss': 0.5456376671791077, 'train/grad_norm_pre_clip': 11.324164390563965, 'timing/data': 0.025326833128929138, 'timing/model': 0.9725499609485269, 'learning_rate/qwen_vl_interface': 8.213991445896895e-06, 'learning_rate/action_model': 8.213991445896895e-05, 'epoch': 0.94}) +[2026-06-17 20:01:57,274][starVLA.training.train_starvla][INFO] - Step 2071, Loss: {'train/loss': 0.5248469114303589, 'train/grad_norm_pre_clip': 14.09900951385498, 'timing/data': 0.0178828751668334, 'timing/model': 0.9648204687982798, 'learning_rate/qwen_vl_interface': 8.212301156553995e-06, 'learning_rate/action_model': 8.212301156553995e-05, 'epoch': 0.94}) +[2026-06-17 20:02:01,365][starVLA.training.train_starvla][INFO] - Step 2072, Loss: {'train/loss': 0.47422441840171814, 'train/grad_norm_pre_clip': 13.264318466186523, 'timing/data': 0.02214006334543228, 'timing/model': 0.9689894625917077, 'learning_rate/qwen_vl_interface': 8.210610253122879e-06, 'learning_rate/action_model': 8.210610253122879e-05, 'epoch': 0.94}) +[2026-06-17 20:02:05,436][starVLA.training.train_starvla][INFO] - Step 2073, Loss: {'train/loss': 0.5858956575393677, 'train/grad_norm_pre_clip': 29.577219009399414, 'timing/data': 0.026555825024843216, 'timing/model': 0.9608523258939385, 'learning_rate/qwen_vl_interface': 8.20891873595407e-06, 'learning_rate/action_model': 8.20891873595407e-05, 'epoch': 0.94}) +[2026-06-17 20:02:09,661][starVLA.training.train_starvla][INFO] - Step 2074, Loss: {'train/loss': 0.6621809601783752, 'train/grad_norm_pre_clip': 12.54208755493164, 'timing/data': 0.0188785158097744, 'timing/model': 0.9731114348396659, 'learning_rate/qwen_vl_interface': 8.207226605398225e-06, 'learning_rate/action_model': 8.207226605398224e-05, 'epoch': 0.94}) +[2026-06-17 20:02:13,948][starVLA.training.train_starvla][INFO] - Step 2075, Loss: {'train/loss': 0.41737130284309387, 'train/grad_norm_pre_clip': 10.920236587524414, 'timing/data': 0.022109278477728367, 'timing/model': 1.1919223563745618, 'learning_rate/qwen_vl_interface': 8.20553386180612e-06, 'learning_rate/action_model': 8.205533861806119e-05, 'epoch': 0.94}) +[2026-06-17 20:02:17,658][starVLA.training.train_starvla][INFO] - Step 2076, Loss: {'train/loss': 0.3734334707260132, 'train/grad_norm_pre_clip': 7.742258548736572, 'timing/data': 0.01833068486303091, 'timing/model': 0.9661327572539449, 'learning_rate/qwen_vl_interface': 8.203840505528665e-06, 'learning_rate/action_model': 8.203840505528665e-05, 'epoch': 0.94}) +[2026-06-17 20:02:21,640][starVLA.training.train_starvla][INFO] - Step 2077, Loss: {'train/loss': 0.5628336071968079, 'train/grad_norm_pre_clip': 12.357866287231445, 'timing/data': 0.023099983111023903, 'timing/model': 0.9447316937148571, 'learning_rate/qwen_vl_interface': 8.202146536916899e-06, 'learning_rate/action_model': 8.202146536916899e-05, 'epoch': 0.94}) +[2026-06-17 20:02:25,690][starVLA.training.train_starvla][INFO] - Step 2078, Loss: {'train/loss': 0.5957022905349731, 'train/grad_norm_pre_clip': 12.844582557678223, 'timing/data': 0.02938446495682001, 'timing/model': 0.9802084807306528, 'learning_rate/qwen_vl_interface': 8.200451956321974e-06, 'learning_rate/action_model': 8.200451956321974e-05, 'epoch': 0.94}) +[2026-06-17 20:02:30,043][starVLA.training.train_starvla][INFO] - Step 2079, Loss: {'train/loss': 0.44995301961898804, 'train/grad_norm_pre_clip': 10.937350273132324, 'timing/data': 0.02979341708123684, 'timing/model': 1.021523516625166, 'learning_rate/qwen_vl_interface': 8.198756764095187e-06, 'learning_rate/action_model': 8.198756764095187e-05, 'epoch': 0.94}) +[2026-06-17 20:02:34,073][starVLA.training.train_starvla][INFO] - Step 2080, Loss: {'train/loss': 0.5853521823883057, 'train/grad_norm_pre_clip': 12.939047813415527, 'timing/data': 0.02602247428148985, 'timing/model': 0.9763899715617299, 'learning_rate/qwen_vl_interface': 8.19706096058795e-06, 'learning_rate/action_model': 8.197060960587949e-05, 'epoch': 0.94}) +[2026-06-17 20:02:38,142][starVLA.training.train_starvla][INFO] - Step 2081, Loss: {'train/loss': 0.552226722240448, 'train/grad_norm_pre_clip': 10.86215877532959, 'timing/data': 0.020447109825909138, 'timing/model': 1.240732440724969, 'learning_rate/qwen_vl_interface': 8.195364546151802e-06, 'learning_rate/action_model': 8.195364546151801e-05, 'epoch': 0.94}) +[2026-06-17 20:02:42,109][starVLA.training.train_starvla][INFO] - Step 2082, Loss: {'train/loss': 0.38519087433815, 'train/grad_norm_pre_clip': 9.719351768493652, 'timing/data': 0.023368703201413155, 'timing/model': 1.135732620023191, 'learning_rate/qwen_vl_interface': 8.193667521138416e-06, 'learning_rate/action_model': 8.193667521138416e-05, 'epoch': 0.94}) +[2026-06-17 20:02:45,745][starVLA.training.train_starvla][INFO] - Step 2083, Loss: {'train/loss': 0.3951100707054138, 'train/grad_norm_pre_clip': 12.655641555786133, 'timing/data': 0.021521017886698246, 'timing/model': 0.9300180664286017, 'learning_rate/qwen_vl_interface': 8.191969885899585e-06, 'learning_rate/action_model': 8.191969885899585e-05, 'epoch': 0.94}) +[2026-06-17 20:02:49,641][starVLA.training.train_starvla][INFO] - Step 2084, Loss: {'train/loss': 0.44508829712867737, 'train/grad_norm_pre_clip': 10.641653060913086, 'timing/data': 0.024269241839647293, 'timing/model': 0.9673544960096478, 'learning_rate/qwen_vl_interface': 8.190271640787232e-06, 'learning_rate/action_model': 8.190271640787232e-05, 'epoch': 0.95}) +[2026-06-17 20:02:53,449][starVLA.training.train_starvla][INFO] - Step 2085, Loss: {'train/loss': 0.5310801267623901, 'train/grad_norm_pre_clip': 10.643858909606934, 'timing/data': 0.014293941669166088, 'timing/model': 0.9451843006536365, 'learning_rate/qwen_vl_interface': 8.188572786153403e-06, 'learning_rate/action_model': 8.188572786153402e-05, 'epoch': 0.95}) +[2026-06-17 20:02:57,227][starVLA.training.train_starvla][INFO] - Step 2086, Loss: {'train/loss': 0.27372267842292786, 'train/grad_norm_pre_clip': 9.465593338012695, 'timing/data': 0.017686513252556324, 'timing/model': 0.9272762127220631, 'learning_rate/qwen_vl_interface': 8.186873322350274e-06, 'learning_rate/action_model': 8.186873322350273e-05, 'epoch': 0.95}) +[2026-06-17 20:03:01,057][starVLA.training.train_starvla][INFO] - Step 2087, Loss: {'train/loss': 0.8259663581848145, 'train/grad_norm_pre_clip': 16.37413215637207, 'timing/data': 0.02737010270357132, 'timing/model': 0.92963882163167, 'learning_rate/qwen_vl_interface': 8.185173249730143e-06, 'learning_rate/action_model': 8.185173249730142e-05, 'epoch': 0.95}) +[2026-06-17 20:03:04,857][starVLA.training.train_starvla][INFO] - Step 2088, Loss: {'train/loss': 0.4862990975379944, 'train/grad_norm_pre_clip': 11.454049110412598, 'timing/data': 0.016892016865313053, 'timing/model': 1.1100015882402658, 'learning_rate/qwen_vl_interface': 8.183472568645441e-06, 'learning_rate/action_model': 8.18347256864544e-05, 'epoch': 0.95}) +[2026-06-17 20:03:08,709][starVLA.training.train_starvla][INFO] - Step 2089, Loss: {'train/loss': 0.5917760133743286, 'train/grad_norm_pre_clip': 13.332686424255371, 'timing/data': 0.017990530468523502, 'timing/model': 1.1113949157297611, 'learning_rate/qwen_vl_interface': 8.181771279448718e-06, 'learning_rate/action_model': 8.181771279448718e-05, 'epoch': 0.95}) +[2026-06-17 20:03:12,376][starVLA.training.train_starvla][INFO] - Step 2090, Loss: {'train/loss': 0.46933165192604065, 'train/grad_norm_pre_clip': 12.80579948425293, 'timing/data': 0.01839896384626627, 'timing/model': 0.9368963092565536, 'learning_rate/qwen_vl_interface': 8.180069382492655e-06, 'learning_rate/action_model': 8.180069382492654e-05, 'epoch': 0.95}) +[2026-06-17 20:03:16,240][starVLA.training.train_starvla][INFO] - Step 2091, Loss: {'train/loss': 0.48461899161338806, 'train/grad_norm_pre_clip': 10.313931465148926, 'timing/data': 0.025918168015778065, 'timing/model': 0.9327464168891311, 'learning_rate/qwen_vl_interface': 8.178366878130056e-06, 'learning_rate/action_model': 8.178366878130056e-05, 'epoch': 0.95}) +[2026-06-17 20:03:20,069][starVLA.training.train_starvla][INFO] - Step 2092, Loss: {'train/loss': 0.3189295828342438, 'train/grad_norm_pre_clip': 10.570405960083008, 'timing/data': 0.017253431491553783, 'timing/model': 0.930114927701652, 'learning_rate/qwen_vl_interface': 8.176663766713852e-06, 'learning_rate/action_model': 8.176663766713852e-05, 'epoch': 0.95}) +[2026-06-17 20:03:23,856][starVLA.training.train_starvla][INFO] - Step 2093, Loss: {'train/loss': 0.3113906979560852, 'train/grad_norm_pre_clip': 11.851422309875488, 'timing/data': 0.01732301525771618, 'timing/model': 0.9295075479894876, 'learning_rate/qwen_vl_interface': 8.174960048597099e-06, 'learning_rate/action_model': 8.174960048597099e-05, 'epoch': 0.95}) +[2026-06-17 20:03:27,696][starVLA.training.train_starvla][INFO] - Step 2094, Loss: {'train/loss': 0.36994311213493347, 'train/grad_norm_pre_clip': 13.676867485046387, 'timing/data': 0.013004071079194546, 'timing/model': 0.9481723541393876, 'learning_rate/qwen_vl_interface': 8.173255724132982e-06, 'learning_rate/action_model': 8.17325572413298e-05, 'epoch': 0.95}) +[2026-06-17 20:03:31,612][starVLA.training.train_starvla][INFO] - Step 2095, Loss: {'train/loss': 0.37305858731269836, 'train/grad_norm_pre_clip': 10.856263160705566, 'timing/data': 0.026119421236217022, 'timing/model': 1.1617113323882222, 'learning_rate/qwen_vl_interface': 8.171550793674806e-06, 'learning_rate/action_model': 8.171550793674805e-05, 'epoch': 0.95}) +[2026-06-17 20:03:35,266][starVLA.training.train_starvla][INFO] - Step 2096, Loss: {'train/loss': 0.6800591349601746, 'train/grad_norm_pre_clip': 14.570619583129883, 'timing/data': 0.027061075903475285, 'timing/model': 0.9399201003834605, 'learning_rate/qwen_vl_interface': 8.169845257576008e-06, 'learning_rate/action_model': 8.169845257576007e-05, 'epoch': 0.95}) +[2026-06-17 20:03:39,638][starVLA.training.train_starvla][INFO] - Step 2097, Loss: {'train/loss': 0.37522685527801514, 'train/grad_norm_pre_clip': 9.81556224822998, 'timing/data': 0.020096836611628532, 'timing/model': 0.9333884716033936, 'learning_rate/qwen_vl_interface': 8.168139116190146e-06, 'learning_rate/action_model': 8.168139116190146e-05, 'epoch': 0.95}) +[2026-06-17 20:03:43,553][starVLA.training.train_starvla][INFO] - Step 2098, Loss: {'train/loss': 0.5488926768302917, 'train/grad_norm_pre_clip': 13.443108558654785, 'timing/data': 0.022327481769025326, 'timing/model': 0.9852956542745233, 'learning_rate/qwen_vl_interface': 8.166432369870905e-06, 'learning_rate/action_model': 8.166432369870904e-05, 'epoch': 0.95}) +[2026-06-17 20:03:47,630][starVLA.training.train_starvla][INFO] - Step 2099, Loss: {'train/loss': 0.6062847375869751, 'train/grad_norm_pre_clip': 14.343480110168457, 'timing/data': 0.023206221871078014, 'timing/model': 0.9679237548261881, 'learning_rate/qwen_vl_interface': 8.164725018972096e-06, 'learning_rate/action_model': 8.164725018972096e-05, 'epoch': 0.95}) +[2026-06-17 20:06:09,882][starVLA.training.train_starvla][INFO] - Step 2100, Loss: {'train/loss': 0.32257288694381714, 'train/grad_norm_pre_clip': 13.392032623291016, 'eval/loss': 0.48730987310409546, 'eval/latency_8/loss': 0.48730987310409546, 'eval/demon_fire_precision': 0.9172413793103448, 'eval/demon_fire_recall': 0.9916123019571296, 'eval/demon_fire_f1': 0.9529780564263323, 'eval/demon_fire_rate_model': 0.9682804674457429, 'eval/demon_fire_rate_teacher': 0.8956594323873122, 'eval/demon_left_precision': 0.8, 'eval/demon_left_recall': 0.4625267665952891, 'eval/demon_left_f1': 0.5861601085481682, 'eval/demon_left_rate_model': 0.07512520868113523, 'eval/demon_left_rate_teacher': 0.12993878686700056, 'eval/demon_right_precision': 0.8542094455852156, 'eval/demon_right_recall': 0.6303030303030303, 'eval/demon_right_f1': 0.7253705318221447, 'eval/demon_right_rate_model': 0.13550361713967723, 'eval/demon_right_rate_teacher': 0.18363939899833054, 'eval/demon_move_f1': 0.6557653201851565, 'eval/demon_cc_f1': 0.8389119526384547, 'eval/latency_8/demon_fire_precision': 0.9172413793103448, 'eval/latency_8/demon_fire_recall': 0.9916123019571296, 'eval/latency_8/demon_fire_f1': 0.9529780564263323, 'eval/latency_8/demon_fire_rate_model': 0.9682804674457429, 'eval/latency_8/demon_fire_rate_teacher': 0.8956594323873122, 'eval/latency_8/demon_left_precision': 0.8, 'eval/latency_8/demon_left_recall': 0.4625267665952891, 'eval/latency_8/demon_left_f1': 0.5861601085481682, 'eval/latency_8/demon_left_rate_model': 0.07512520868113523, 'eval/latency_8/demon_left_rate_teacher': 0.12993878686700056, 'eval/latency_8/demon_right_precision': 0.8542094455852156, 'eval/latency_8/demon_right_recall': 0.6303030303030303, 'eval/latency_8/demon_right_f1': 0.7253705318221447, 'eval/latency_8/demon_right_rate_model': 0.13550361713967723, 'eval/latency_8/demon_right_rate_teacher': 0.18363939899833054, 'eval/latency_8/demon_move_f1': 0.6557653201851565, 'eval/latency_8/demon_cc_f1': 0.8389119526384547, 'timing/data': 0.018112540245056152, 'timing/model': 0.9957860242575407, 'learning_rate/qwen_vl_interface': 8.163017063847654e-06, 'learning_rate/action_model': 8.163017063847654e-05, 'epoch': 0.95}) +[2026-06-17 20:06:09,885][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_2100_state +[2026-06-17 20:06:09,890][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 20:06:17,310][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_2100_state/model.safetensors +[2026-06-17 20:06:33,783][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_2100_state/optimizer.bin +[2026-06-17 20:06:33,784][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_2100_state/sampler.bin +[2026-06-17 20:06:33,784][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_2100_state/sampler_1.bin +[2026-06-17 20:06:33,786][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_2100_state/random_states_0.pkl +[2026-06-17 20:06:33,791][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2100 +[2026-06-17 20:06:33,791][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 20:06:33,803][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 20:06:37,699][starVLA.training.train_starvla][INFO] - Step 2101, Loss: {'train/loss': 0.46959859132766724, 'train/grad_norm_pre_clip': 12.73403549194336, 'timing/data': 0.019471270963549614, 'timing/model': 0.9364147959277034, 'learning_rate/qwen_vl_interface': 8.16130850485164e-06, 'learning_rate/action_model': 8.161308504851639e-05, 'epoch': 0.95}) +[2026-06-17 20:06:41,540][starVLA.training.train_starvla][INFO] - Step 2102, Loss: {'train/loss': 0.5111880898475647, 'train/grad_norm_pre_clip': 12.614185333251953, 'timing/data': 0.018724212422966957, 'timing/model': 0.9407851323485374, 'learning_rate/qwen_vl_interface': 8.15959934233824e-06, 'learning_rate/action_model': 8.15959934233824e-05, 'epoch': 0.95}) +[2026-06-17 20:06:45,396][starVLA.training.train_starvla][INFO] - Step 2103, Loss: {'train/loss': 0.7954796552658081, 'train/grad_norm_pre_clip': 13.241535186767578, 'timing/data': 0.024474075995385647, 'timing/model': 0.9437519134953618, 'learning_rate/qwen_vl_interface': 8.157889576661765e-06, 'learning_rate/action_model': 8.157889576661764e-05, 'epoch': 0.95}) +[2026-06-17 20:06:51,699][starVLA.training.train_starvla][INFO] - Step 2104, Loss: {'train/loss': 0.38033491373062134, 'train/grad_norm_pre_clip': 11.307154655456543, 'timing/data': 0.017434616573154926, 'timing/model': 1.1936289807781577, 'learning_rate/qwen_vl_interface': 8.156179208176652e-06, 'learning_rate/action_model': 8.156179208176652e-05, 'epoch': 0.95}) +[2026-06-17 20:06:55,411][starVLA.training.train_starvla][INFO] - Step 2105, Loss: {'train/loss': 0.7372405529022217, 'train/grad_norm_pre_clip': 14.226783752441406, 'timing/data': 0.021498664282262325, 'timing/model': 0.9322965843603015, 'learning_rate/qwen_vl_interface': 8.154468237237463e-06, 'learning_rate/action_model': 8.154468237237462e-05, 'epoch': 0.95}) +[2026-06-17 20:06:59,279][starVLA.training.train_starvla][INFO] - Step 2106, Loss: {'train/loss': 0.5833297967910767, 'train/grad_norm_pre_clip': 18.870922088623047, 'timing/data': 0.019199898466467857, 'timing/model': 0.9309541955590248, 'learning_rate/qwen_vl_interface': 8.152756664198882e-06, 'learning_rate/action_model': 8.152756664198882e-05, 'epoch': 0.96}) +[2026-06-17 20:07:03,110][starVLA.training.train_starvla][INFO] - Step 2107, Loss: {'train/loss': 0.4509040415287018, 'train/grad_norm_pre_clip': 15.131324768066406, 'timing/data': 0.019394969567656517, 'timing/model': 0.9433400612324476, 'learning_rate/qwen_vl_interface': 8.151044489415722e-06, 'learning_rate/action_model': 8.151044489415722e-05, 'epoch': 0.96}) +[2026-06-17 20:07:07,076][starVLA.training.train_starvla][INFO] - Step 2108, Loss: {'train/loss': 0.43937548995018005, 'train/grad_norm_pre_clip': 12.124479293823242, 'timing/data': 0.01757862139493227, 'timing/model': 1.039263604208827, 'learning_rate/qwen_vl_interface': 8.149331713242916e-06, 'learning_rate/action_model': 8.149331713242916e-05, 'epoch': 0.96}) +[2026-06-17 20:07:11,579][starVLA.training.train_starvla][INFO] - Step 2109, Loss: {'train/loss': 0.6782053112983704, 'train/grad_norm_pre_clip': 17.97813606262207, 'timing/data': 0.01742779277265072, 'timing/model': 1.2405394380912185, 'learning_rate/qwen_vl_interface': 8.147618336035526e-06, 'learning_rate/action_model': 8.147618336035526e-05, 'epoch': 0.96}) +[2026-06-17 20:07:15,908][starVLA.training.train_starvla][INFO] - Step 2110, Loss: {'train/loss': 0.36716681718826294, 'train/grad_norm_pre_clip': 12.387619972229004, 'timing/data': 0.019087649881839752, 'timing/model': 1.22213638946414, 'learning_rate/qwen_vl_interface': 8.14590435814874e-06, 'learning_rate/action_model': 8.14590435814874e-05, 'epoch': 0.96}) +[2026-06-17 20:07:19,822][starVLA.training.train_starvla][INFO] - Step 2111, Loss: {'train/loss': 0.5500445365905762, 'train/grad_norm_pre_clip': 14.76832389831543, 'timing/data': 0.014985733665525913, 'timing/model': 1.1930605778470635, 'learning_rate/qwen_vl_interface': 8.144189779937862e-06, 'learning_rate/action_model': 8.144189779937862e-05, 'epoch': 0.96}) +[2026-06-17 20:07:25,303][starVLA.training.train_starvla][INFO] - Step 2112, Loss: {'train/loss': 0.6708296537399292, 'train/grad_norm_pre_clip': 14.725958824157715, 'timing/data': 0.016164276748895645, 'timing/model': 2.7194194300100207, 'learning_rate/qwen_vl_interface': 8.142474601758331e-06, 'learning_rate/action_model': 8.142474601758329e-05, 'epoch': 0.96}) +[2026-06-17 20:07:29,347][starVLA.training.train_starvla][INFO] - Step 2113, Loss: {'train/loss': 0.5042256116867065, 'train/grad_norm_pre_clip': 14.4967679977417, 'timing/data': 0.017323301173746586, 'timing/model': 0.9799143271520734, 'learning_rate/qwen_vl_interface': 8.1407588239657e-06, 'learning_rate/action_model': 8.140758823965698e-05, 'epoch': 0.96}) +[2026-06-17 20:07:34,409][starVLA.training.train_starvla][INFO] - Step 2114, Loss: {'train/loss': 0.6804099678993225, 'train/grad_norm_pre_clip': 14.431635856628418, 'timing/data': 0.01899588480591774, 'timing/model': 0.9485076582059264, 'learning_rate/qwen_vl_interface': 8.139042446915653e-06, 'learning_rate/action_model': 8.139042446915654e-05, 'epoch': 0.96}) +[2026-06-17 20:07:38,718][starVLA.training.train_starvla][INFO] - Step 2115, Loss: {'train/loss': 0.46281349658966064, 'train/grad_norm_pre_clip': 9.360566139221191, 'timing/data': 0.05277613364160061, 'timing/model': 0.9744996270164847, 'learning_rate/qwen_vl_interface': 8.137325470964001e-06, 'learning_rate/action_model': 8.137325470964e-05, 'epoch': 0.96}) +[2026-06-17 20:07:43,352][starVLA.training.train_starvla][INFO] - Step 2116, Loss: {'train/loss': 0.6046963334083557, 'train/grad_norm_pre_clip': 13.045784950256348, 'timing/data': 0.03064396046102047, 'timing/model': 0.9916553311049938, 'learning_rate/qwen_vl_interface': 8.135607896466669e-06, 'learning_rate/action_model': 8.13560789646667e-05, 'epoch': 0.96}) +[2026-06-17 20:07:47,559][starVLA.training.train_starvla][INFO] - Step 2117, Loss: {'train/loss': 0.4775213897228241, 'train/grad_norm_pre_clip': 9.42640209197998, 'timing/data': 0.031427592039108276, 'timing/model': 1.2734470898285508, 'learning_rate/qwen_vl_interface': 8.133889723779718e-06, 'learning_rate/action_model': 8.133889723779717e-05, 'epoch': 0.96}) +[2026-06-17 20:07:52,964][starVLA.training.train_starvla][INFO] - Step 2118, Loss: {'train/loss': 0.38497650623321533, 'train/grad_norm_pre_clip': 10.160354614257812, 'timing/data': 0.03585419524461031, 'timing/model': 1.2586049735546112, 'learning_rate/qwen_vl_interface': 8.132170953259321e-06, 'learning_rate/action_model': 8.132170953259322e-05, 'epoch': 0.96}) +[2026-06-17 20:07:56,766][starVLA.training.train_starvla][INFO] - Step 2119, Loss: {'train/loss': 0.5244455933570862, 'train/grad_norm_pre_clip': 12.491273880004883, 'timing/data': 0.028422679752111435, 'timing/model': 0.9508112277835608, 'learning_rate/qwen_vl_interface': 8.130451585261786e-06, 'learning_rate/action_model': 8.130451585261786e-05, 'epoch': 0.96}) +[2026-06-17 20:08:00,881][starVLA.training.train_starvla][INFO] - Step 2120, Loss: {'train/loss': 0.5256935954093933, 'train/grad_norm_pre_clip': 10.038185119628906, 'timing/data': 0.03744933940470219, 'timing/model': 0.9933483367785811, 'learning_rate/qwen_vl_interface': 8.128731620143538e-06, 'learning_rate/action_model': 8.128731620143538e-05, 'epoch': 0.96}) +[2026-06-17 20:08:04,959][starVLA.training.train_starvla][INFO] - Step 2121, Loss: {'train/loss': 0.3792544901371002, 'train/grad_norm_pre_clip': 9.790133476257324, 'timing/data': 0.0357439611107111, 'timing/model': 0.9668717607855797, 'learning_rate/qwen_vl_interface': 8.127011058261129e-06, 'learning_rate/action_model': 8.127011058261128e-05, 'epoch': 0.96}) +[2026-06-17 20:08:09,093][starVLA.training.train_starvla][INFO] - Step 2122, Loss: {'train/loss': 0.5382254123687744, 'train/grad_norm_pre_clip': 14.039361000061035, 'timing/data': 0.0393596850335598, 'timing/model': 0.9802730744704604, 'learning_rate/qwen_vl_interface': 8.125289899971229e-06, 'learning_rate/action_model': 8.125289899971228e-05, 'epoch': 0.96}) +[2026-06-17 20:08:15,184][starVLA.training.train_starvla][INFO] - Step 2123, Loss: {'train/loss': 0.47338661551475525, 'train/grad_norm_pre_clip': 17.497581481933594, 'timing/data': 0.03702315501868725, 'timing/model': 0.9967996897175908, 'learning_rate/qwen_vl_interface': 8.123568145630641e-06, 'learning_rate/action_model': 8.12356814563064e-05, 'epoch': 0.96}) +[2026-06-17 20:08:19,089][starVLA.training.train_starvla][INFO] - Step 2124, Loss: {'train/loss': 0.433353453874588, 'train/grad_norm_pre_clip': 12.429662704467773, 'timing/data': 0.029632763005793095, 'timing/model': 1.182210996747017, 'learning_rate/qwen_vl_interface': 8.121845795596282e-06, 'learning_rate/action_model': 8.121845795596282e-05, 'epoch': 0.96}) +[2026-06-17 20:08:23,606][starVLA.training.train_starvla][INFO] - Step 2125, Loss: {'train/loss': 0.32982543110847473, 'train/grad_norm_pre_clip': 8.772759437561035, 'timing/data': 0.03430333361029625, 'timing/model': 1.1378205679357052, 'learning_rate/qwen_vl_interface': 8.1201228502252e-06, 'learning_rate/action_model': 8.120122850225199e-05, 'epoch': 0.96}) +[2026-06-17 20:08:27,340][starVLA.training.train_starvla][INFO] - Step 2126, Loss: {'train/loss': 0.4291921854019165, 'train/grad_norm_pre_clip': 23.899648666381836, 'timing/data': 0.03392791096121073, 'timing/model': 0.9388559721410275, 'learning_rate/qwen_vl_interface': 8.118399309874563e-06, 'learning_rate/action_model': 8.118399309874563e-05, 'epoch': 0.96}) +[2026-06-17 20:08:31,229][starVLA.training.train_starvla][INFO] - Step 2127, Loss: {'train/loss': 0.5855684280395508, 'train/grad_norm_pre_clip': 15.571789741516113, 'timing/data': 0.034341804683208466, 'timing/model': 0.9381647240370512, 'learning_rate/qwen_vl_interface': 8.116675174901663e-06, 'learning_rate/action_model': 8.116675174901662e-05, 'epoch': 0.96}) +[2026-06-17 20:08:35,200][starVLA.training.train_starvla][INFO] - Step 2128, Loss: {'train/loss': 0.5221052169799805, 'train/grad_norm_pre_clip': 12.023866653442383, 'timing/data': 0.03094188030809164, 'timing/model': 0.9589098999276757, 'learning_rate/qwen_vl_interface': 8.114950445663915e-06, 'learning_rate/action_model': 8.114950445663913e-05, 'epoch': 0.97}) +[2026-06-17 20:08:39,213][starVLA.training.train_starvla][INFO] - Step 2129, Loss: {'train/loss': 0.27045902609825134, 'train/grad_norm_pre_clip': 10.148076057434082, 'timing/data': 0.03907052334398031, 'timing/model': 0.9998460542410612, 'learning_rate/qwen_vl_interface': 8.113225122518854e-06, 'learning_rate/action_model': 8.113225122518854e-05, 'epoch': 0.97}) +[2026-06-17 20:08:43,791][starVLA.training.train_starvla][INFO] - Step 2130, Loss: {'train/loss': 0.29374921321868896, 'train/grad_norm_pre_clip': 10.134286880493164, 'timing/data': 0.030605987645685673, 'timing/model': 0.9267284534871578, 'learning_rate/qwen_vl_interface': 8.111499205824148e-06, 'learning_rate/action_model': 8.111499205824147e-05, 'epoch': 0.97}) +[2026-06-17 20:08:47,633][starVLA.training.train_starvla][INFO] - Step 2131, Loss: {'train/loss': 0.4406017065048218, 'train/grad_norm_pre_clip': 13.392496109008789, 'timing/data': 0.03532417584210634, 'timing/model': 1.1683611832559109, 'learning_rate/qwen_vl_interface': 8.109772695937575e-06, 'learning_rate/action_model': 8.109772695937574e-05, 'epoch': 0.97}) +[2026-06-17 20:08:51,391][starVLA.training.train_starvla][INFO] - Step 2132, Loss: {'train/loss': 0.49960705637931824, 'train/grad_norm_pre_clip': 13.336756706237793, 'timing/data': 0.0329681895673275, 'timing/model': 0.9520443473011255, 'learning_rate/qwen_vl_interface': 8.108045593217045e-06, 'learning_rate/action_model': 8.108045593217045e-05, 'epoch': 0.97}) +[2026-06-17 20:08:55,553][starVLA.training.train_starvla][INFO] - Step 2133, Loss: {'train/loss': 0.5473213195800781, 'train/grad_norm_pre_clip': 11.83365249633789, 'timing/data': 0.04289344884455204, 'timing/model': 0.9461201224476099, 'learning_rate/qwen_vl_interface': 8.106317898020589e-06, 'learning_rate/action_model': 8.106317898020589e-05, 'epoch': 0.97}) +[2026-06-17 20:08:59,648][starVLA.training.train_starvla][INFO] - Step 2134, Loss: {'train/loss': 0.5594422221183777, 'train/grad_norm_pre_clip': 10.602970123291016, 'timing/data': 0.033199507743120193, 'timing/model': 0.9488587398082018, 'learning_rate/qwen_vl_interface': 8.104589610706359e-06, 'learning_rate/action_model': 8.104589610706359e-05, 'epoch': 0.97}) +[2026-06-17 20:09:03,629][starVLA.training.train_starvla][INFO] - Step 2135, Loss: {'train/loss': 0.4676760137081146, 'train/grad_norm_pre_clip': 13.595244407653809, 'timing/data': 0.037456694059073925, 'timing/model': 0.9563541458919644, 'learning_rate/qwen_vl_interface': 8.102860731632632e-06, 'learning_rate/action_model': 8.102860731632631e-05, 'epoch': 0.97}) +[2026-06-17 20:09:07,701][starVLA.training.train_starvla][INFO] - Step 2136, Loss: {'train/loss': 0.38661810755729675, 'train/grad_norm_pre_clip': 9.627015113830566, 'timing/data': 0.03624431695789099, 'timing/model': 0.9737601336091757, 'learning_rate/qwen_vl_interface': 8.101131261157802e-06, 'learning_rate/action_model': 8.101131261157802e-05, 'epoch': 0.97}) +[2026-06-17 20:09:11,667][starVLA.training.train_starvla][INFO] - Step 2137, Loss: {'train/loss': 0.9826101660728455, 'train/grad_norm_pre_clip': 25.72684097290039, 'timing/data': 0.04661550186574459, 'timing/model': 0.9648038148880005, 'learning_rate/qwen_vl_interface': 8.099401199640397e-06, 'learning_rate/action_model': 8.099401199640397e-05, 'epoch': 0.97}) +[2026-06-17 20:09:16,353][starVLA.training.train_starvla][INFO] - Step 2138, Loss: {'train/loss': 0.4056508541107178, 'train/grad_norm_pre_clip': 11.972405433654785, 'timing/data': 0.037649817764759064, 'timing/model': 1.8167006419971585, 'learning_rate/qwen_vl_interface': 8.097670547439057e-06, 'learning_rate/action_model': 8.097670547439058e-05, 'epoch': 0.97}) +[2026-06-17 20:09:20,115][starVLA.training.train_starvla][INFO] - Step 2139, Loss: {'train/loss': 0.4663754105567932, 'train/grad_norm_pre_clip': 13.206084251403809, 'timing/data': 0.03526865504682064, 'timing/model': 0.9547097440809011, 'learning_rate/qwen_vl_interface': 8.095939304912549e-06, 'learning_rate/action_model': 8.095939304912548e-05, 'epoch': 0.97}) +[2026-06-17 20:09:24,115][starVLA.training.train_starvla][INFO] - Step 2140, Loss: {'train/loss': 0.7161005139350891, 'train/grad_norm_pre_clip': 26.507936477661133, 'timing/data': 0.02880383562296629, 'timing/model': 0.9773348290473223, 'learning_rate/qwen_vl_interface': 8.094207472419759e-06, 'learning_rate/action_model': 8.094207472419759e-05, 'epoch': 0.97}) +[2026-06-17 20:09:28,089][starVLA.training.train_starvla][INFO] - Step 2141, Loss: {'train/loss': 0.46926334500312805, 'train/grad_norm_pre_clip': 12.032214164733887, 'timing/data': 0.0357380248606205, 'timing/model': 0.9532077284529805, 'learning_rate/qwen_vl_interface': 8.092475050319699e-06, 'learning_rate/action_model': 8.092475050319698e-05, 'epoch': 0.97}) +[2026-06-17 20:09:32,138][starVLA.training.train_starvla][INFO] - Step 2142, Loss: {'train/loss': 0.5260144472122192, 'train/grad_norm_pre_clip': 19.570457458496094, 'timing/data': 0.037275020964443684, 'timing/model': 0.961689468473196, 'learning_rate/qwen_vl_interface': 8.090742038971503e-06, 'learning_rate/action_model': 8.090742038971503e-05, 'epoch': 0.97}) +[2026-06-17 20:09:36,108][starVLA.training.train_starvla][INFO] - Step 2143, Loss: {'train/loss': 0.5924323797225952, 'train/grad_norm_pre_clip': 14.531770706176758, 'timing/data': 0.04021223355084658, 'timing/model': 0.963312329724431, 'learning_rate/qwen_vl_interface': 8.089008438734427e-06, 'learning_rate/action_model': 8.089008438734427e-05, 'epoch': 0.97}) +[2026-06-17 20:09:40,137][starVLA.training.train_starvla][INFO] - Step 2144, Loss: {'train/loss': 0.5421704649925232, 'train/grad_norm_pre_clip': 12.394641876220703, 'timing/data': 0.05485953018069267, 'timing/model': 0.9586653988808393, 'learning_rate/qwen_vl_interface': 8.087274249967844e-06, 'learning_rate/action_model': 8.087274249967845e-05, 'epoch': 0.97}) +[2026-06-17 20:09:44,263][starVLA.training.train_starvla][INFO] - Step 2145, Loss: {'train/loss': 0.5289888978004456, 'train/grad_norm_pre_clip': 17.8746337890625, 'timing/data': 0.03590292204171419, 'timing/model': 1.3964380659162998, 'learning_rate/qwen_vl_interface': 8.08553947303126e-06, 'learning_rate/action_model': 8.085539473031259e-05, 'epoch': 0.97}) +[2026-06-17 20:09:48,043][starVLA.training.train_starvla][INFO] - Step 2146, Loss: {'train/loss': 0.4250180721282959, 'train/grad_norm_pre_clip': 17.422260284423828, 'timing/data': 0.033323279581964016, 'timing/model': 0.9619159903377295, 'learning_rate/qwen_vl_interface': 8.083804108284288e-06, 'learning_rate/action_model': 8.083804108284287e-05, 'epoch': 0.97}) +[2026-06-17 20:09:52,001][starVLA.training.train_starvla][INFO] - Step 2147, Loss: {'train/loss': 0.6596292853355408, 'train/grad_norm_pre_clip': 14.775264739990234, 'timing/data': 0.0341610312461853, 'timing/model': 0.9661041526123881, 'learning_rate/qwen_vl_interface': 8.082068156086675e-06, 'learning_rate/action_model': 8.082068156086675e-05, 'epoch': 0.97}) +[2026-06-17 20:09:55,959][starVLA.training.train_starvla][INFO] - Step 2148, Loss: {'train/loss': 0.4910530149936676, 'train/grad_norm_pre_clip': 8.95455265045166, 'timing/data': 0.038405694998800755, 'timing/model': 0.9554009027779102, 'learning_rate/qwen_vl_interface': 8.080331616798289e-06, 'learning_rate/action_model': 8.080331616798287e-05, 'epoch': 0.97}) +[2026-06-17 20:09:59,840][starVLA.training.train_starvla][INFO] - Step 2149, Loss: {'train/loss': 0.4420883357524872, 'train/grad_norm_pre_clip': 13.682483673095703, 'timing/data': 0.0452968729659915, 'timing/model': 0.9616716476157308, 'learning_rate/qwen_vl_interface': 8.078594490779112e-06, 'learning_rate/action_model': 8.078594490779111e-05, 'epoch': 0.97}) +[2026-06-17 20:10:03,728][starVLA.training.train_starvla][INFO] - Step 2150, Loss: {'train/loss': 0.231000155210495, 'train/grad_norm_pre_clip': 14.7710599899292, 'timing/data': 0.0349276177585125, 'timing/model': 0.9308843174949288, 'learning_rate/qwen_vl_interface': 8.076856778389253e-06, 'learning_rate/action_model': 8.076856778389252e-05, 'epoch': 0.98}) +[2026-06-17 20:10:07,585][starVLA.training.train_starvla][INFO] - Step 2151, Loss: {'train/loss': 0.54364013671875, 'train/grad_norm_pre_clip': 13.715845108032227, 'timing/data': 0.03977297432720661, 'timing/model': 1.146293549798429, 'learning_rate/qwen_vl_interface': 8.075118479988942e-06, 'learning_rate/action_model': 8.075118479988941e-05, 'epoch': 0.98}) +[2026-06-17 20:10:11,868][starVLA.training.train_starvla][INFO] - Step 2152, Loss: {'train/loss': 0.2697941064834595, 'train/grad_norm_pre_clip': 7.360048294067383, 'timing/data': 0.03739475831389427, 'timing/model': 1.4757523043081164, 'learning_rate/qwen_vl_interface': 8.07337959593853e-06, 'learning_rate/action_model': 8.073379595938529e-05, 'epoch': 0.98}) +[2026-06-17 20:10:16,737][starVLA.training.train_starvla][INFO] - Step 2153, Loss: {'train/loss': 0.4829545319080353, 'train/grad_norm_pre_clip': 10.531083106994629, 'timing/data': 0.0340125048533082, 'timing/model': 0.9535429310053587, 'learning_rate/qwen_vl_interface': 8.071640126598488e-06, 'learning_rate/action_model': 8.071640126598488e-05, 'epoch': 0.98}) +[2026-06-17 20:10:20,736][starVLA.training.train_starvla][INFO] - Step 2154, Loss: {'train/loss': 0.5690874457359314, 'train/grad_norm_pre_clip': 11.334203720092773, 'timing/data': 0.042077318765223026, 'timing/model': 0.9398106774315238, 'learning_rate/qwen_vl_interface': 8.069900072329414e-06, 'learning_rate/action_model': 8.069900072329415e-05, 'epoch': 0.98}) +[2026-06-17 20:10:24,623][starVLA.training.train_starvla][INFO] - Step 2155, Loss: {'train/loss': 0.4583609700202942, 'train/grad_norm_pre_clip': 13.285430908203125, 'timing/data': 0.04137733392417431, 'timing/model': 1.000416903756559, 'learning_rate/qwen_vl_interface': 8.068159433492021e-06, 'learning_rate/action_model': 8.068159433492021e-05, 'epoch': 0.98}) +[2026-06-17 20:10:28,462][starVLA.training.train_starvla][INFO] - Step 2156, Loss: {'train/loss': 0.654525876045227, 'train/grad_norm_pre_clip': 12.205374717712402, 'timing/data': 0.049106284976005554, 'timing/model': 0.9391817422583699, 'learning_rate/qwen_vl_interface': 8.066418210447144e-06, 'learning_rate/action_model': 8.066418210447143e-05, 'epoch': 0.98}) +[2026-06-17 20:10:32,446][starVLA.training.train_starvla][INFO] - Step 2157, Loss: {'train/loss': 0.49221426248550415, 'train/grad_norm_pre_clip': 14.993636131286621, 'timing/data': 0.03478236868977547, 'timing/model': 0.973102992400527, 'learning_rate/qwen_vl_interface': 8.064676403555741e-06, 'learning_rate/action_model': 8.064676403555741e-05, 'epoch': 0.98}) +[2026-06-17 20:10:36,369][starVLA.training.train_starvla][INFO] - Step 2158, Loss: {'train/loss': 0.8956692814826965, 'train/grad_norm_pre_clip': 14.781730651855469, 'timing/data': 0.03732937015593052, 'timing/model': 1.2000611694529653, 'learning_rate/qwen_vl_interface': 8.06293401317889e-06, 'learning_rate/action_model': 8.062934013178891e-05, 'epoch': 0.98}) +[2026-06-17 20:10:40,511][starVLA.training.train_starvla][INFO] - Step 2159, Loss: {'train/loss': 0.48068562150001526, 'train/grad_norm_pre_clip': 10.566494941711426, 'timing/data': 0.037503138184547424, 'timing/model': 1.317394376732409, 'learning_rate/qwen_vl_interface': 8.061191039677793e-06, 'learning_rate/action_model': 8.061191039677793e-05, 'epoch': 0.98}) +[2026-06-17 20:10:44,266][starVLA.training.train_starvla][INFO] - Step 2160, Loss: {'train/loss': 0.3869757652282715, 'train/grad_norm_pre_clip': 10.5015230178833, 'timing/data': 0.04045569896697998, 'timing/model': 1.0415536211803555, 'learning_rate/qwen_vl_interface': 8.059447483413768e-06, 'learning_rate/action_model': 8.059447483413767e-05, 'epoch': 0.98}) +[2026-06-17 20:10:48,247][starVLA.training.train_starvla][INFO] - Step 2161, Loss: {'train/loss': 0.5046629309654236, 'train/grad_norm_pre_clip': 11.416593551635742, 'timing/data': 0.05868864431977272, 'timing/model': 0.9501213012263179, 'learning_rate/qwen_vl_interface': 8.057703344748258e-06, 'learning_rate/action_model': 8.057703344748257e-05, 'epoch': 0.98}) +[2026-06-17 20:10:52,210][starVLA.training.train_starvla][INFO] - Step 2162, Loss: {'train/loss': 0.388873428106308, 'train/grad_norm_pre_clip': 9.982077598571777, 'timing/data': 0.028067896142601967, 'timing/model': 0.9994026757776737, 'learning_rate/qwen_vl_interface': 8.05595862404282e-06, 'learning_rate/action_model': 8.05595862404282e-05, 'epoch': 0.98}) +[2026-06-17 20:10:56,262][starVLA.training.train_starvla][INFO] - Step 2163, Loss: {'train/loss': 0.516111433506012, 'train/grad_norm_pre_clip': 13.560968399047852, 'timing/data': 0.03819232061505318, 'timing/model': 1.0743722757324576, 'learning_rate/qwen_vl_interface': 8.054213321659143e-06, 'learning_rate/action_model': 8.054213321659142e-05, 'epoch': 0.98}) +[2026-06-17 20:11:00,330][starVLA.training.train_starvla][INFO] - Step 2164, Loss: {'train/loss': 0.4588805139064789, 'train/grad_norm_pre_clip': 12.645410537719727, 'timing/data': 0.03117549419403076, 'timing/model': 1.0636498415842652, 'learning_rate/qwen_vl_interface': 8.052467437959025e-06, 'learning_rate/action_model': 8.052467437959025e-05, 'epoch': 0.98}) +[2026-06-17 20:11:04,316][starVLA.training.train_starvla][INFO] - Step 2165, Loss: {'train/loss': 0.5965826511383057, 'train/grad_norm_pre_clip': 13.330306053161621, 'timing/data': 0.04188964609056711, 'timing/model': 1.2456806041300297, 'learning_rate/qwen_vl_interface': 8.050720973304393e-06, 'learning_rate/action_model': 8.050720973304391e-05, 'epoch': 0.98}) +[2026-06-17 20:11:08,027][starVLA.training.train_starvla][INFO] - Step 2166, Loss: {'train/loss': 0.49770525097846985, 'train/grad_norm_pre_clip': 11.74274730682373, 'timing/data': 0.04532265476882458, 'timing/model': 0.992096628062427, 'learning_rate/qwen_vl_interface': 8.048973928057288e-06, 'learning_rate/action_model': 8.048973928057288e-05, 'epoch': 0.98}) +[2026-06-17 20:11:12,024][starVLA.training.train_starvla][INFO] - Step 2167, Loss: {'train/loss': 0.38992226123809814, 'train/grad_norm_pre_clip': 12.151383399963379, 'timing/data': 0.03230581432580948, 'timing/model': 0.988658020272851, 'learning_rate/qwen_vl_interface': 8.047226302579877e-06, 'learning_rate/action_model': 8.047226302579877e-05, 'epoch': 0.98}) +[2026-06-17 20:11:16,004][starVLA.training.train_starvla][INFO] - Step 2168, Loss: {'train/loss': 0.4221571683883667, 'train/grad_norm_pre_clip': 13.17205810546875, 'timing/data': 0.03921774309128523, 'timing/model': 1.0253531513735652, 'learning_rate/qwen_vl_interface': 8.045478097234443e-06, 'learning_rate/action_model': 8.045478097234444e-05, 'epoch': 0.98}) +[2026-06-17 20:11:19,873][starVLA.training.train_starvla][INFO] - Step 2169, Loss: {'train/loss': 0.38445279002189636, 'train/grad_norm_pre_clip': 12.099935531616211, 'timing/data': 0.03197078872472048, 'timing/model': 0.9456059243530035, 'learning_rate/qwen_vl_interface': 8.043729312383393e-06, 'learning_rate/action_model': 8.043729312383392e-05, 'epoch': 0.98}) +[2026-06-17 20:11:23,855][starVLA.training.train_starvla][INFO] - Step 2170, Loss: {'train/loss': 0.42440107464790344, 'train/grad_norm_pre_clip': 12.341249465942383, 'timing/data': 0.029873820021748543, 'timing/model': 0.9484950881451368, 'learning_rate/qwen_vl_interface': 8.041979948389248e-06, 'learning_rate/action_model': 8.041979948389248e-05, 'epoch': 0.98}) +[2026-06-17 20:11:27,900][starVLA.training.train_starvla][INFO] - Step 2171, Loss: {'train/loss': 0.4341734051704407, 'train/grad_norm_pre_clip': 13.324052810668945, 'timing/data': 0.030853229574859142, 'timing/model': 0.9508271655067801, 'learning_rate/qwen_vl_interface': 8.040230005614656e-06, 'learning_rate/action_model': 8.040230005614656e-05, 'epoch': 0.98}) +[2026-06-17 20:11:31,797][starVLA.training.train_starvla][INFO] - Step 2172, Loss: {'train/loss': 0.3356861472129822, 'train/grad_norm_pre_clip': 7.21486759185791, 'timing/data': 0.039032441563904285, 'timing/model': 1.1472794590517879, 'learning_rate/qwen_vl_interface': 8.038479484422383e-06, 'learning_rate/action_model': 8.038479484422383e-05, 'epoch': 0.99}) +[2026-06-17 20:11:35,535][starVLA.training.train_starvla][INFO] - Step 2173, Loss: {'train/loss': 0.5866531729698181, 'train/grad_norm_pre_clip': 15.267285346984863, 'timing/data': 0.03804030828177929, 'timing/model': 0.9451140128076077, 'learning_rate/qwen_vl_interface': 8.03672838517531e-06, 'learning_rate/action_model': 8.036728385175311e-05, 'epoch': 0.99}) +[2026-06-17 20:11:39,393][starVLA.training.train_starvla][INFO] - Step 2174, Loss: {'train/loss': 0.7699754238128662, 'train/grad_norm_pre_clip': 15.549325942993164, 'timing/data': 0.04618288483470678, 'timing/model': 0.9546381710097194, 'learning_rate/qwen_vl_interface': 8.034976708236446e-06, 'learning_rate/action_model': 8.034976708236445e-05, 'epoch': 0.99}) +[2026-06-17 20:11:43,627][starVLA.training.train_starvla][INFO] - Step 2175, Loss: {'train/loss': 0.5636348724365234, 'train/grad_norm_pre_clip': 12.49765396118164, 'timing/data': 0.03636390995234251, 'timing/model': 1.0534006292000413, 'learning_rate/qwen_vl_interface': 8.033224453968913e-06, 'learning_rate/action_model': 8.033224453968913e-05, 'epoch': 0.99}) +[2026-06-17 20:11:47,732][starVLA.training.train_starvla][INFO] - Step 2176, Loss: {'train/loss': 0.35857492685317993, 'train/grad_norm_pre_clip': 10.20386791229248, 'timing/data': 0.0381761658936739, 'timing/model': 0.9536277493461967, 'learning_rate/qwen_vl_interface': 8.031471622735956e-06, 'learning_rate/action_model': 8.031471622735956e-05, 'epoch': 0.99}) +[2026-06-17 20:11:51,589][starVLA.training.train_starvla][INFO] - Step 2177, Loss: {'train/loss': 0.6384340524673462, 'train/grad_norm_pre_clip': 15.179840087890625, 'timing/data': 0.03752718213945627, 'timing/model': 0.9512913385406137, 'learning_rate/qwen_vl_interface': 8.029718214900937e-06, 'learning_rate/action_model': 8.029718214900937e-05, 'epoch': 0.99}) +[2026-06-17 20:11:55,424][starVLA.training.train_starvla][INFO] - Step 2178, Loss: {'train/loss': 0.6063007712364197, 'train/grad_norm_pre_clip': 14.38281536102295, 'timing/data': 0.029866419732570648, 'timing/model': 0.9506619852036238, 'learning_rate/qwen_vl_interface': 8.027964230827342e-06, 'learning_rate/action_model': 8.027964230827342e-05, 'epoch': 0.99}) +[2026-06-17 20:11:59,250][starVLA.training.train_starvla][INFO] - Step 2179, Loss: {'train/loss': 0.549515426158905, 'train/grad_norm_pre_clip': 12.004607200622559, 'timing/data': 0.03566018957644701, 'timing/model': 1.113852396607399, 'learning_rate/qwen_vl_interface': 8.026209670878776e-06, 'learning_rate/action_model': 8.026209670878775e-05, 'epoch': 0.99}) +[2026-06-17 20:12:02,929][starVLA.training.train_starvla][INFO] - Step 2180, Loss: {'train/loss': 0.3642827272415161, 'train/grad_norm_pre_clip': 11.760638236999512, 'timing/data': 0.038951785303652287, 'timing/model': 0.953473879955709, 'learning_rate/qwen_vl_interface': 8.024454535418952e-06, 'learning_rate/action_model': 8.024454535418951e-05, 'epoch': 0.99}) +[2026-06-17 20:12:06,747][starVLA.training.train_starvla][INFO] - Step 2181, Loss: {'train/loss': 0.7692185640335083, 'train/grad_norm_pre_clip': 18.256431579589844, 'timing/data': 0.03219457156956196, 'timing/model': 0.9410955999046564, 'learning_rate/qwen_vl_interface': 8.02269882481172e-06, 'learning_rate/action_model': 8.022698824811721e-05, 'epoch': 0.99}) +[2026-06-17 20:12:10,630][starVLA.training.train_starvla][INFO] - Step 2182, Loss: {'train/loss': 0.48492103815078735, 'train/grad_norm_pre_clip': 12.68044662475586, 'timing/data': 0.04218465182930231, 'timing/model': 0.9520646324381232, 'learning_rate/qwen_vl_interface': 8.02094253942104e-06, 'learning_rate/action_model': 8.02094253942104e-05, 'epoch': 0.99}) +[2026-06-17 20:12:14,534][starVLA.training.train_starvla][INFO] - Step 2183, Loss: {'train/loss': 0.4806586503982544, 'train/grad_norm_pre_clip': 17.416208267211914, 'timing/data': 0.03664183523505926, 'timing/model': 0.9565433654934168, 'learning_rate/qwen_vl_interface': 8.019185679610988e-06, 'learning_rate/action_model': 8.019185679610988e-05, 'epoch': 0.99}) +[2026-06-17 20:12:18,337][starVLA.training.train_starvla][INFO] - Step 2184, Loss: {'train/loss': 0.6460536122322083, 'train/grad_norm_pre_clip': 16.34712791442871, 'timing/data': 0.03469425439834595, 'timing/model': 0.9329252978786826, 'learning_rate/qwen_vl_interface': 8.017428245745765e-06, 'learning_rate/action_model': 8.017428245745763e-05, 'epoch': 0.99}) +[2026-06-17 20:12:22,162][starVLA.training.train_starvla][INFO] - Step 2185, Loss: {'train/loss': 0.38697350025177, 'train/grad_norm_pre_clip': 10.97642993927002, 'timing/data': 0.0313856303691864, 'timing/model': 1.1479031359776855, 'learning_rate/qwen_vl_interface': 8.015670238189688e-06, 'learning_rate/action_model': 8.015670238189688e-05, 'epoch': 0.99}) +[2026-06-17 20:12:26,035][starVLA.training.train_starvla][INFO] - Step 2186, Loss: {'train/loss': 0.30719172954559326, 'train/grad_norm_pre_clip': 7.967296600341797, 'timing/data': 0.03240535594522953, 'timing/model': 1.1410493087023497, 'learning_rate/qwen_vl_interface': 8.013911657307196e-06, 'learning_rate/action_model': 8.013911657307196e-05, 'epoch': 0.99}) +[2026-06-17 20:12:29,697][starVLA.training.train_starvla][INFO] - Step 2187, Loss: {'train/loss': 0.6121892929077148, 'train/grad_norm_pre_clip': 17.549049377441406, 'timing/data': 0.04333543498069048, 'timing/model': 0.948149542324245, 'learning_rate/qwen_vl_interface': 8.012152503462843e-06, 'learning_rate/action_model': 8.012152503462842e-05, 'epoch': 0.99}) +[2026-06-17 20:12:33,570][starVLA.training.train_starvla][INFO] - Step 2188, Loss: {'train/loss': 0.5969744920730591, 'train/grad_norm_pre_clip': 11.399980545043945, 'timing/data': 0.03464497905224562, 'timing/model': 0.9536091703921556, 'learning_rate/qwen_vl_interface': 8.010392777021304e-06, 'learning_rate/action_model': 8.010392777021303e-05, 'epoch': 0.99}) +[2026-06-17 20:12:37,405][starVLA.training.train_starvla][INFO] - Step 2189, Loss: {'train/loss': 0.6928184628486633, 'train/grad_norm_pre_clip': 13.245302200317383, 'timing/data': 0.04121156223118305, 'timing/model': 0.9525628983974457, 'learning_rate/qwen_vl_interface': 8.008632478347372e-06, 'learning_rate/action_model': 8.008632478347371e-05, 'epoch': 0.99}) +[2026-06-17 20:12:41,348][starVLA.training.train_starvla][INFO] - Step 2190, Loss: {'train/loss': 0.4932200610637665, 'train/grad_norm_pre_clip': 13.073737144470215, 'timing/data': 0.03260401263833046, 'timing/model': 0.9454001951962709, 'learning_rate/qwen_vl_interface': 8.006871607805957e-06, 'learning_rate/action_model': 8.006871607805957e-05, 'epoch': 0.99}) +[2026-06-17 20:12:45,356][starVLA.training.train_starvla][INFO] - Step 2191, Loss: {'train/loss': 0.3701871931552887, 'train/grad_norm_pre_clip': 9.611470222473145, 'timing/data': 0.03384026698768139, 'timing/model': 0.9658602392300963, 'learning_rate/qwen_vl_interface': 8.005110165762093e-06, 'learning_rate/action_model': 8.005110165762093e-05, 'epoch': 0.99}) +[2026-06-17 20:12:49,354][starVLA.training.train_starvla][INFO] - Step 2192, Loss: {'train/loss': 0.4229937195777893, 'train/grad_norm_pre_clip': 10.660392761230469, 'timing/data': 0.03723096288740635, 'timing/model': 1.2124968320131302, 'learning_rate/qwen_vl_interface': 8.003348152580927e-06, 'learning_rate/action_model': 8.003348152580927e-05, 'epoch': 0.99}) +[2026-06-17 20:12:53,304][starVLA.training.train_starvla][INFO] - Step 2193, Loss: {'train/loss': 0.418686181306839, 'train/grad_norm_pre_clip': 10.958828926086426, 'timing/data': 0.03297022357583046, 'timing/model': 1.1698487186804414, 'learning_rate/qwen_vl_interface': 8.001585568627724e-06, 'learning_rate/action_model': 8.001585568627725e-05, 'epoch': 0.99}) +[2026-06-17 20:13:01,522][starVLA.training.train_starvla][INFO] - Step 2194, Loss: {'train/loss': 0.29305267333984375, 'train/grad_norm_pre_clip': 6.987218379974365, 'timing/data': 0.041229802183806896, 'timing/model': 1.0387299368157983, 'learning_rate/qwen_vl_interface': 7.999822414267874e-06, 'learning_rate/action_model': 7.999822414267874e-05, 'epoch': 1.0}) +[2026-06-17 20:13:05,505][starVLA.training.train_starvla][INFO] - Step 2195, Loss: {'train/loss': 0.43692734837532043, 'train/grad_norm_pre_clip': 11.428327560424805, 'timing/data': 0.033486816100776196, 'timing/model': 0.9678330617025495, 'learning_rate/qwen_vl_interface': 7.998058689866879e-06, 'learning_rate/action_model': 7.998058689866879e-05, 'epoch': 1.0}) +[2026-06-17 20:13:09,481][starVLA.training.train_starvla][INFO] - Step 2196, Loss: {'train/loss': 0.7589839696884155, 'train/grad_norm_pre_clip': 13.281668663024902, 'timing/data': 0.030711648054420948, 'timing/model': 0.9963155090808868, 'learning_rate/qwen_vl_interface': 7.99629439579036e-06, 'learning_rate/action_model': 7.99629439579036e-05, 'epoch': 1.0}) +[2026-06-17 20:13:14,218][starVLA.training.train_starvla][INFO] - Step 2197, Loss: {'train/loss': 0.3337288200855255, 'train/grad_norm_pre_clip': 10.497801780700684, 'timing/data': 0.03750651981681585, 'timing/model': 1.7639611205086112, 'learning_rate/qwen_vl_interface': 7.994529532404059e-06, 'learning_rate/action_model': 7.994529532404058e-05, 'epoch': 1.0}) +[2026-06-17 20:13:21,029][starVLA.training.train_starvla][INFO] - Step 2198, Loss: {'train/loss': 0.23978212475776672, 'train/grad_norm_pre_clip': 9.132196426391602, 'timing/data': 0.0371190095320344, 'timing/model': 3.6463408740237355, 'learning_rate/qwen_vl_interface': 7.99276410007383e-06, 'learning_rate/action_model': 7.99276410007383e-05, 'epoch': 1.0}) +[2026-06-17 20:13:27,621][starVLA.training.train_starvla][INFO] - Step 2199, Loss: {'train/loss': 0.28678643703460693, 'train/grad_norm_pre_clip': 8.365948677062988, 'timing/data': 0.04033403471112251, 'timing/model': 2.836248906329274, 'learning_rate/qwen_vl_interface': 7.990998099165652e-06, 'learning_rate/action_model': 7.990998099165652e-05, 'epoch': 1.0}) +[2026-06-17 20:16:11,515][starVLA.training.train_starvla][INFO] - Step 2200, Loss: {'train/loss': 0.5150988101959229, 'train/grad_norm_pre_clip': 12.19600772857666, 'eval/loss': 0.45184844732284546, 'eval/latency_8/loss': 0.45184844732284546, 'eval/demon_fire_precision': 0.9061351427763642, 'eval/demon_fire_recall': 0.9956508232370301, 'eval/demon_fire_f1': 0.9487862640615748, 'eval/demon_fire_rate_model': 0.9841402337228714, 'eval/demon_fire_rate_teacher': 0.8956594323873122, 'eval/demon_left_precision': 0.8795620437956204, 'eval/demon_left_recall': 0.5160599571734475, 'eval/demon_left_f1': 0.650472334682861, 'eval/demon_left_rate_model': 0.07623817473567056, 'eval/demon_left_rate_teacher': 0.12993878686700056, 'eval/demon_right_precision': 0.67125, 'eval/demon_right_recall': 0.8136363636363636, 'eval/demon_right_f1': 0.7356164383561643, 'eval/demon_right_rate_model': 0.22259321090706732, 'eval/demon_right_rate_teacher': 0.18363939899833054, 'eval/demon_move_f1': 0.6930443865195126, 'eval/demon_cc_f1': 0.8542208710062215, 'eval/latency_8/demon_fire_precision': 0.9061351427763642, 'eval/latency_8/demon_fire_recall': 0.9956508232370301, 'eval/latency_8/demon_fire_f1': 0.9487862640615748, 'eval/latency_8/demon_fire_rate_model': 0.9841402337228714, 'eval/latency_8/demon_fire_rate_teacher': 0.8956594323873122, 'eval/latency_8/demon_left_precision': 0.8795620437956204, 'eval/latency_8/demon_left_recall': 0.5160599571734475, 'eval/latency_8/demon_left_f1': 0.650472334682861, 'eval/latency_8/demon_left_rate_model': 0.07623817473567056, 'eval/latency_8/demon_left_rate_teacher': 0.12993878686700056, 'eval/latency_8/demon_right_precision': 0.67125, 'eval/latency_8/demon_right_recall': 0.8136363636363636, 'eval/latency_8/demon_right_f1': 0.7356164383561643, 'eval/latency_8/demon_right_rate_model': 0.22259321090706732, 'eval/latency_8/demon_right_rate_teacher': 0.18363939899833054, 'eval/latency_8/demon_move_f1': 0.6930443865195126, 'eval/latency_8/demon_cc_f1': 0.8542208710062215, 'timing/data': 0.02972791250795126, 'timing/model': 0.9628256037831306, 'learning_rate/qwen_vl_interface': 7.98923153004562e-06, 'learning_rate/action_model': 7.98923153004562e-05, 'epoch': 1.0}) +[2026-06-17 20:16:11,518][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_2200_state +[2026-06-17 20:16:11,527][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 20:16:18,483][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_2200_state/model.safetensors +[2026-06-17 20:16:34,140][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_2200_state/optimizer.bin +[2026-06-17 20:16:34,143][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_2200_state/sampler.bin +[2026-06-17 20:16:34,143][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_2200_state/sampler_1.bin +[2026-06-17 20:16:34,147][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_2200_state/random_states_0.pkl +[2026-06-17 20:16:34,152][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2200 +[2026-06-17 20:16:34,152][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 20:16:34,164][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 20:16:38,513][starVLA.training.train_starvla][INFO] - Step 2201, Loss: {'train/loss': 0.5177972316741943, 'train/grad_norm_pre_clip': 12.195659637451172, 'timing/data': 0.04416742641478777, 'timing/model': 1.3769920971244574, 'learning_rate/qwen_vl_interface': 7.987464393079943e-06, 'learning_rate/action_model': 7.987464393079943e-05, 'epoch': 1.0}) +[2026-06-17 20:16:42,779][starVLA.training.train_starvla][INFO] - Step 2202, Loss: {'train/loss': 0.3996599614620209, 'train/grad_norm_pre_clip': 10.102629661560059, 'timing/data': 0.03473459091037512, 'timing/model': 1.5000270903110504, 'learning_rate/qwen_vl_interface': 7.985696688634953e-06, 'learning_rate/action_model': 7.985696688634952e-05, 'epoch': 1.0}) +[2026-06-17 20:16:46,500][starVLA.training.train_starvla][INFO] - Step 2203, Loss: {'train/loss': 0.5869538187980652, 'train/grad_norm_pre_clip': 12.758806228637695, 'timing/data': 0.03758208453655243, 'timing/model': 0.9436843981966376, 'learning_rate/qwen_vl_interface': 7.983928417077095e-06, 'learning_rate/action_model': 7.983928417077095e-05, 'epoch': 1.0}) +[2026-06-17 20:16:50,890][starVLA.training.train_starvla][INFO] - Step 2204, Loss: {'train/loss': 0.33064574003219604, 'train/grad_norm_pre_clip': 12.769143104553223, 'timing/data': 0.010914703831076622, 'timing/model': 1.0802155155688524, 'learning_rate/qwen_vl_interface': 7.982159578772935e-06, 'learning_rate/action_model': 7.982159578772934e-05, 'epoch': 1.0}) +[2026-06-17 20:16:55,262][starVLA.training.train_starvla][INFO] - Step 2205, Loss: {'train/loss': 0.38353100419044495, 'train/grad_norm_pre_clip': 13.670807838439941, 'timing/data': 0.0005774460732936859, 'timing/model': 0.9661703538149595, 'learning_rate/qwen_vl_interface': 7.980390174089154e-06, 'learning_rate/action_model': 7.980390174089153e-05, 'epoch': 1.0}) +[2026-06-17 20:16:59,261][starVLA.training.train_starvla][INFO] - Step 2206, Loss: {'train/loss': 0.7028171420097351, 'train/grad_norm_pre_clip': 15.366034507751465, 'timing/data': 0.0010276669636368752, 'timing/model': 0.9710728358477354, 'learning_rate/qwen_vl_interface': 7.978620203392551e-06, 'learning_rate/action_model': 7.97862020339255e-05, 'epoch': 1.0}) +[2026-06-17 20:17:03,316][starVLA.training.train_starvla][INFO] - Step 2207, Loss: {'train/loss': 0.8829862475395203, 'train/grad_norm_pre_clip': 19.09072494506836, 'timing/data': 0.0007570711895823479, 'timing/model': 0.9546417072415352, 'learning_rate/qwen_vl_interface': 7.976849667050044e-06, 'learning_rate/action_model': 7.976849667050044e-05, 'epoch': 1.0}) +[2026-06-17 20:17:07,535][starVLA.training.train_starvla][INFO] - Step 2208, Loss: {'train/loss': 0.7222192287445068, 'train/grad_norm_pre_clip': 13.370753288269043, 'timing/data': 0.000867750495672226, 'timing/model': 1.3646739134564996, 'learning_rate/qwen_vl_interface': 7.975078565428667e-06, 'learning_rate/action_model': 7.975078565428668e-05, 'epoch': 1.0}) +[2026-06-17 20:17:11,513][starVLA.training.train_starvla][INFO] - Step 2209, Loss: {'train/loss': 0.33410924673080444, 'train/grad_norm_pre_clip': 11.69913101196289, 'timing/data': 0.0008415775373578072, 'timing/model': 1.212074656970799, 'learning_rate/qwen_vl_interface': 7.97330689889557e-06, 'learning_rate/action_model': 7.97330689889557e-05, 'epoch': 1.0}) +[2026-06-17 20:17:15,331][starVLA.training.train_starvla][INFO] - Step 2210, Loss: {'train/loss': 0.3865439295768738, 'train/grad_norm_pre_clip': 27.152246475219727, 'timing/data': 0.0007560830563306808, 'timing/model': 1.0315896598622203, 'learning_rate/qwen_vl_interface': 7.971534667818026e-06, 'learning_rate/action_model': 7.971534667818025e-05, 'epoch': 1.0}) +[2026-06-17 20:17:19,365][starVLA.training.train_starvla][INFO] - Step 2211, Loss: {'train/loss': 0.513225793838501, 'train/grad_norm_pre_clip': 13.399514198303223, 'timing/data': 0.0010896772146224976, 'timing/model': 0.9554003700613976, 'learning_rate/qwen_vl_interface': 7.969761872563413e-06, 'learning_rate/action_model': 7.969761872563413e-05, 'epoch': 1.0}) +[2026-06-17 20:17:23,370][starVLA.training.train_starvla][INFO] - Step 2212, Loss: {'train/loss': 0.816652774810791, 'train/grad_norm_pre_clip': 16.820037841796875, 'timing/data': 0.0005794772878289223, 'timing/model': 0.9854119485244155, 'learning_rate/qwen_vl_interface': 7.96798851349924e-06, 'learning_rate/action_model': 7.96798851349924e-05, 'epoch': 1.0}) +[2026-06-17 20:17:27,325][starVLA.training.train_starvla][INFO] - Step 2213, Loss: {'train/loss': 0.2898991107940674, 'train/grad_norm_pre_clip': 7.954626560211182, 'timing/data': 0.0005032578483223915, 'timing/model': 0.9728394830599427, 'learning_rate/qwen_vl_interface': 7.966214590993122e-06, 'learning_rate/action_model': 7.966214590993122e-05, 'epoch': 1.0}) +[2026-06-17 20:17:31,321][starVLA.training.train_starvla][INFO] - Step 2214, Loss: {'train/loss': 0.542393684387207, 'train/grad_norm_pre_clip': 12.20105266571045, 'timing/data': 0.0006310287863016129, 'timing/model': 0.9669718174263835, 'learning_rate/qwen_vl_interface': 7.964440105412797e-06, 'learning_rate/action_model': 7.964440105412796e-05, 'epoch': 1.0}) +[2026-06-17 20:17:35,377][starVLA.training.train_starvla][INFO] - Step 2215, Loss: {'train/loss': 0.588346540927887, 'train/grad_norm_pre_clip': 16.35738182067871, 'timing/data': 0.03795270621776581, 'timing/model': 1.1902424367144704, 'learning_rate/qwen_vl_interface': 7.962665057126116e-06, 'learning_rate/action_model': 7.962665057126117e-05, 'epoch': 1.0}) +[2026-06-17 20:17:39,298][starVLA.training.train_starvla][INFO] - Step 2216, Loss: {'train/loss': 0.7054333090782166, 'train/grad_norm_pre_clip': 14.79379940032959, 'timing/data': 0.001147627830505371, 'timing/model': 1.134370910935104, 'learning_rate/qwen_vl_interface': 7.960889446501052e-06, 'learning_rate/action_model': 7.960889446501052e-05, 'epoch': 1.01}) +[2026-06-17 20:17:43,023][starVLA.training.train_starvla][INFO] - Step 2217, Loss: {'train/loss': 0.7961389422416687, 'train/grad_norm_pre_clip': 16.86250877380371, 'timing/data': 0.0010387981310486794, 'timing/model': 0.9569424269720912, 'learning_rate/qwen_vl_interface': 7.959113273905688e-06, 'learning_rate/action_model': 7.959113273905688e-05, 'epoch': 1.01}) +[2026-06-17 20:17:47,037][starVLA.training.train_starvla][INFO] - Step 2218, Loss: {'train/loss': 0.6642935276031494, 'train/grad_norm_pre_clip': 13.183043479919434, 'timing/data': 0.0006022816523909569, 'timing/model': 0.9721353426575661, 'learning_rate/qwen_vl_interface': 7.957336539708226e-06, 'learning_rate/action_model': 7.957336539708225e-05, 'epoch': 1.01}) +[2026-06-17 20:17:51,541][starVLA.training.train_starvla][INFO] - Step 2219, Loss: {'train/loss': 0.35867637395858765, 'train/grad_norm_pre_clip': 11.151095390319824, 'timing/data': 0.0009752120822668076, 'timing/model': 0.9685222944244742, 'learning_rate/qwen_vl_interface': 7.955559244276987e-06, 'learning_rate/action_model': 7.955559244276987e-05, 'epoch': 1.01}) +[2026-06-17 20:17:56,658][starVLA.training.train_starvla][INFO] - Step 2220, Loss: {'train/loss': 0.6615292429924011, 'train/grad_norm_pre_clip': 17.494842529296875, 'timing/data': 0.0008071670308709145, 'timing/model': 1.0164286438375711, 'learning_rate/qwen_vl_interface': 7.953781387980404e-06, 'learning_rate/action_model': 7.953781387980404e-05, 'epoch': 1.01}) +[2026-06-17 20:18:00,528][starVLA.training.train_starvla][INFO] - Step 2221, Loss: {'train/loss': 0.3700610101222992, 'train/grad_norm_pre_clip': 9.859292030334473, 'timing/data': 0.0009039584547281265, 'timing/model': 1.1285971961915493, 'learning_rate/qwen_vl_interface': 7.952002971187032e-06, 'learning_rate/action_model': 7.952002971187032e-05, 'epoch': 1.01}) +[2026-06-17 20:18:04,456][starVLA.training.train_starvla][INFO] - Step 2222, Loss: {'train/loss': 0.5254344940185547, 'train/grad_norm_pre_clip': 12.964130401611328, 'timing/data': 0.0009574070572853088, 'timing/model': 1.1507561095058918, 'learning_rate/qwen_vl_interface': 7.950223994265534e-06, 'learning_rate/action_model': 7.950223994265533e-05, 'epoch': 1.01}) +[2026-06-17 20:18:08,145][starVLA.training.train_starvla][INFO] - Step 2223, Loss: {'train/loss': 0.5797502994537354, 'train/grad_norm_pre_clip': 14.986278533935547, 'timing/data': 0.0009178109467029572, 'timing/model': 0.9392571216449142, 'learning_rate/qwen_vl_interface': 7.9484444575847e-06, 'learning_rate/action_model': 7.948444457584699e-05, 'epoch': 1.01}) +[2026-06-17 20:18:12,044][starVLA.training.train_starvla][INFO] - Step 2224, Loss: {'train/loss': 0.7248326539993286, 'train/grad_norm_pre_clip': 17.59456443786621, 'timing/data': 0.0005504954606294632, 'timing/model': 0.9590586898848414, 'learning_rate/qwen_vl_interface': 7.946664361513423e-06, 'learning_rate/action_model': 7.946664361513421e-05, 'epoch': 1.01}) +[2026-06-17 20:18:16,641][starVLA.training.train_starvla][INFO] - Step 2225, Loss: {'train/loss': 0.4203486144542694, 'train/grad_norm_pre_clip': 11.134857177734375, 'timing/data': 0.0006697028875350952, 'timing/model': 1.096320703625679, 'learning_rate/qwen_vl_interface': 7.944883706420724e-06, 'learning_rate/action_model': 7.944883706420724e-05, 'epoch': 1.01}) +[2026-06-17 20:18:21,638][starVLA.training.train_starvla][INFO] - Step 2226, Loss: {'train/loss': 0.529590904712677, 'train/grad_norm_pre_clip': 15.925119400024414, 'timing/data': 0.0005788281559944153, 'timing/model': 2.0519483871757984, 'learning_rate/qwen_vl_interface': 7.943102492675732e-06, 'learning_rate/action_model': 7.943102492675732e-05, 'epoch': 1.01}) +[2026-06-17 20:18:27,898][starVLA.training.train_starvla][INFO] - Step 2227, Loss: {'train/loss': 0.46620291471481323, 'train/grad_norm_pre_clip': 16.27733612060547, 'timing/data': 0.0005283774808049202, 'timing/model': 0.9612697251141071, 'learning_rate/qwen_vl_interface': 7.941320720647694e-06, 'learning_rate/action_model': 7.941320720647695e-05, 'epoch': 1.01}) +[2026-06-17 20:18:31,767][starVLA.training.train_starvla][INFO] - Step 2228, Loss: {'train/loss': 0.3771626353263855, 'train/grad_norm_pre_clip': 11.040213584899902, 'timing/data': 0.0005012201145291328, 'timing/model': 0.9362199148163199, 'learning_rate/qwen_vl_interface': 7.939538390705976e-06, 'learning_rate/action_model': 7.939538390705976e-05, 'epoch': 1.01}) +[2026-06-17 20:18:35,703][starVLA.training.train_starvla][INFO] - Step 2229, Loss: {'train/loss': 0.534249484539032, 'train/grad_norm_pre_clip': 15.898957252502441, 'timing/data': 0.0004836302250623703, 'timing/model': 1.1356491008773446, 'learning_rate/qwen_vl_interface': 7.937755503220057e-06, 'learning_rate/action_model': 7.937755503220056e-05, 'epoch': 1.01}) +[2026-06-17 20:18:39,333][starVLA.training.train_starvla][INFO] - Step 2230, Loss: {'train/loss': 0.8434813618659973, 'train/grad_norm_pre_clip': 14.968400955200195, 'timing/data': 0.000855298712849617, 'timing/model': 0.9375641737133265, 'learning_rate/qwen_vl_interface': 7.93597205855953e-06, 'learning_rate/action_model': 7.935972058559529e-05, 'epoch': 1.01}) +[2026-06-17 20:18:43,398][starVLA.training.train_starvla][INFO] - Step 2231, Loss: {'train/loss': 0.42433568835258484, 'train/grad_norm_pre_clip': 11.926064491271973, 'timing/data': 0.0005447762086987495, 'timing/model': 0.9479046110063791, 'learning_rate/qwen_vl_interface': 7.934188057094103e-06, 'learning_rate/action_model': 7.934188057094103e-05, 'epoch': 1.01}) +[2026-06-17 20:18:47,231][starVLA.training.train_starvla][INFO] - Step 2232, Loss: {'train/loss': 0.45760706067085266, 'train/grad_norm_pre_clip': 13.051234245300293, 'timing/data': 0.0012030964717268944, 'timing/model': 0.9366741562262177, 'learning_rate/qwen_vl_interface': 7.932403499193604e-06, 'learning_rate/action_model': 7.932403499193605e-05, 'epoch': 1.01}) +[2026-06-17 20:18:51,103][starVLA.training.train_starvla][INFO] - Step 2233, Loss: {'train/loss': 0.3861152231693268, 'train/grad_norm_pre_clip': 12.077497482299805, 'timing/data': 0.0004872828722000122, 'timing/model': 0.9669764945283532, 'learning_rate/qwen_vl_interface': 7.930618385227975e-06, 'learning_rate/action_model': 7.930618385227973e-05, 'epoch': 1.01}) +[2026-06-17 20:18:54,937][starVLA.training.train_starvla][INFO] - Step 2234, Loss: {'train/loss': 0.3056447505950928, 'train/grad_norm_pre_clip': 10.70788288116455, 'timing/data': 0.0010484373196959496, 'timing/model': 0.9675567075610161, 'learning_rate/qwen_vl_interface': 7.928832715567268e-06, 'learning_rate/action_model': 7.928832715567267e-05, 'epoch': 1.01}) +[2026-06-17 20:18:58,901][starVLA.training.train_starvla][INFO] - Step 2235, Loss: {'train/loss': 0.28842899203300476, 'train/grad_norm_pre_clip': 10.011622428894043, 'timing/data': 0.0007526939734816551, 'timing/model': 1.1316608767956495, 'learning_rate/qwen_vl_interface': 7.927046490581655e-06, 'learning_rate/action_model': 7.927046490581655e-05, 'epoch': 1.01}) +[2026-06-17 20:19:02,746][starVLA.training.train_starvla][INFO] - Step 2236, Loss: {'train/loss': 0.6313334703445435, 'train/grad_norm_pre_clip': 19.544857025146484, 'timing/data': 0.0008845077827572823, 'timing/model': 1.157374607399106, 'learning_rate/qwen_vl_interface': 7.925259710641428e-06, 'learning_rate/action_model': 7.925259710641427e-05, 'epoch': 1.01}) +[2026-06-17 20:19:06,442][starVLA.training.train_starvla][INFO] - Step 2237, Loss: {'train/loss': 0.34212207794189453, 'train/grad_norm_pre_clip': 12.692490577697754, 'timing/data': 0.0008537322282791138, 'timing/model': 0.9419731581583619, 'learning_rate/qwen_vl_interface': 7.923472376116982e-06, 'learning_rate/action_model': 7.923472376116982e-05, 'epoch': 1.01}) +[2026-06-17 20:19:10,312][starVLA.training.train_starvla][INFO] - Step 2238, Loss: {'train/loss': 0.4312768578529358, 'train/grad_norm_pre_clip': 14.264471054077148, 'timing/data': 0.0005008717998862267, 'timing/model': 0.9493874851614237, 'learning_rate/qwen_vl_interface': 7.921684487378836e-06, 'learning_rate/action_model': 7.921684487378837e-05, 'epoch': 1.02}) +[2026-06-17 20:19:14,189][starVLA.training.train_starvla][INFO] - Step 2239, Loss: {'train/loss': 0.4936806261539459, 'train/grad_norm_pre_clip': 11.231244087219238, 'timing/data': 0.0005097724497318268, 'timing/model': 0.950490552932024, 'learning_rate/qwen_vl_interface': 7.919896044797622e-06, 'learning_rate/action_model': 7.91989604479762e-05, 'epoch': 1.02}) +[2026-06-17 20:19:17,935][starVLA.training.train_starvla][INFO] - Step 2240, Loss: {'train/loss': 0.42947494983673096, 'train/grad_norm_pre_clip': 13.77319049835205, 'timing/data': 0.0008055102080106735, 'timing/model': 0.9645463265478611, 'learning_rate/qwen_vl_interface': 7.918107048744083e-06, 'learning_rate/action_model': 7.918107048744083e-05, 'epoch': 1.02}) +[2026-06-17 20:19:21,942][starVLA.training.train_starvla][INFO] - Step 2241, Loss: {'train/loss': 0.22768236696720123, 'train/grad_norm_pre_clip': 9.071054458618164, 'timing/data': 0.0005462151020765305, 'timing/model': 0.9432221343740821, 'learning_rate/qwen_vl_interface': 7.916317499589083e-06, 'learning_rate/action_model': 7.916317499589083e-05, 'epoch': 1.02}) +[2026-06-17 20:19:25,822][starVLA.training.train_starvla][INFO] - Step 2242, Loss: {'train/loss': 0.5888251066207886, 'train/grad_norm_pre_clip': 18.630306243896484, 'timing/data': 0.0016934964805841446, 'timing/model': 1.126455712132156, 'learning_rate/qwen_vl_interface': 7.914527397703596e-06, 'learning_rate/action_model': 7.914527397703595e-05, 'epoch': 1.02}) +[2026-06-17 20:19:29,428][starVLA.training.train_starvla][INFO] - Step 2243, Loss: {'train/loss': 0.37560319900512695, 'train/grad_norm_pre_clip': 12.358043670654297, 'timing/data': 0.0012266766279935837, 'timing/model': 0.9311750559136271, 'learning_rate/qwen_vl_interface': 7.912736743458712e-06, 'learning_rate/action_model': 7.912736743458711e-05, 'epoch': 1.02}) +[2026-06-17 20:19:33,309][starVLA.training.train_starvla][INFO] - Step 2244, Loss: {'train/loss': 0.3496447801589966, 'train/grad_norm_pre_clip': 14.033485412597656, 'timing/data': 0.0007595960050821304, 'timing/model': 0.9682648358866572, 'learning_rate/qwen_vl_interface': 7.910945537225638e-06, 'learning_rate/action_model': 7.910945537225637e-05, 'epoch': 1.02}) +[2026-06-17 20:19:37,154][starVLA.training.train_starvla][INFO] - Step 2245, Loss: {'train/loss': 0.7039337158203125, 'train/grad_norm_pre_clip': 16.506086349487305, 'timing/data': 0.0005233399569988251, 'timing/model': 0.9299898631870747, 'learning_rate/qwen_vl_interface': 7.909153779375688e-06, 'learning_rate/action_model': 7.909153779375688e-05, 'epoch': 1.02}) +[2026-06-17 20:19:40,943][starVLA.training.train_starvla][INFO] - Step 2246, Loss: {'train/loss': 0.7516448497772217, 'train/grad_norm_pre_clip': 16.208354949951172, 'timing/data': 0.0008918233215808868, 'timing/model': 0.943769290111959, 'learning_rate/qwen_vl_interface': 7.907361470280299e-06, 'learning_rate/action_model': 7.907361470280299e-05, 'epoch': 1.02}) +[2026-06-17 20:19:44,798][starVLA.training.train_starvla][INFO] - Step 2247, Loss: {'train/loss': 0.3752014636993408, 'train/grad_norm_pre_clip': 12.082337379455566, 'timing/data': 0.0005224458873271942, 'timing/model': 0.9399842685088515, 'learning_rate/qwen_vl_interface': 7.905568610311018e-06, 'learning_rate/action_model': 7.905568610311017e-05, 'epoch': 1.02}) +[2026-06-17 20:19:48,626][starVLA.training.train_starvla][INFO] - Step 2248, Loss: {'train/loss': 0.41363146901130676, 'train/grad_norm_pre_clip': 14.13459300994873, 'timing/data': 0.0006320048123598099, 'timing/model': 1.1404108041897416, 'learning_rate/qwen_vl_interface': 7.903775199839506e-06, 'learning_rate/action_model': 7.903775199839505e-05, 'epoch': 1.02}) +[2026-06-17 20:19:52,450][starVLA.training.train_starvla][INFO] - Step 2249, Loss: {'train/loss': 0.5668007731437683, 'train/grad_norm_pre_clip': 14.852195739746094, 'timing/data': 0.0011139670386910439, 'timing/model': 1.1068934425711632, 'learning_rate/qwen_vl_interface': 7.901981239237539e-06, 'learning_rate/action_model': 7.901981239237539e-05, 'epoch': 1.02}) +[2026-06-17 20:19:56,215][starVLA.training.train_starvla][INFO] - Step 2250, Loss: {'train/loss': 0.3673795759677887, 'train/grad_norm_pre_clip': 9.73439884185791, 'timing/data': 0.0005758823826909065, 'timing/model': 1.0315897725522518, 'learning_rate/qwen_vl_interface': 7.900186728877009e-06, 'learning_rate/action_model': 7.90018672887701e-05, 'epoch': 1.02}) +[2026-06-17 20:20:00,139][starVLA.training.train_starvla][INFO] - Step 2251, Loss: {'train/loss': 0.4916447401046753, 'train/grad_norm_pre_clip': 14.79312515258789, 'timing/data': 0.0010301508009433746, 'timing/model': 0.9440286410972476, 'learning_rate/qwen_vl_interface': 7.898391669129919e-06, 'learning_rate/action_model': 7.898391669129918e-05, 'epoch': 1.02}) +[2026-06-17 20:20:03,990][starVLA.training.train_starvla][INFO] - Step 2252, Loss: {'train/loss': 0.3974670469760895, 'train/grad_norm_pre_clip': 19.57744598388672, 'timing/data': 0.0005753468722105026, 'timing/model': 0.9520263010635972, 'learning_rate/qwen_vl_interface': 7.896596060368384e-06, 'learning_rate/action_model': 7.896596060368384e-05, 'epoch': 1.02}) +[2026-06-17 20:20:07,917][starVLA.training.train_starvla][INFO] - Step 2253, Loss: {'train/loss': 0.7657111287117004, 'train/grad_norm_pre_clip': 16.751270294189453, 'timing/data': 0.0008750064298510551, 'timing/model': 0.9614131553098559, 'learning_rate/qwen_vl_interface': 7.89479990296464e-06, 'learning_rate/action_model': 7.894799902964639e-05, 'epoch': 1.02}) +[2026-06-17 20:20:11,857][starVLA.training.train_starvla][INFO] - Step 2254, Loss: {'train/loss': 0.3451536297798157, 'train/grad_norm_pre_clip': 43.054813385009766, 'timing/data': 0.0008607832714915276, 'timing/model': 0.9637303184717894, 'learning_rate/qwen_vl_interface': 7.893003197291028e-06, 'learning_rate/action_model': 7.893003197291027e-05, 'epoch': 1.02}) +[2026-06-17 20:20:15,759][starVLA.training.train_starvla][INFO] - Step 2255, Loss: {'train/loss': 0.6807569861412048, 'train/grad_norm_pre_clip': 49.39039611816406, 'timing/data': 0.0008303355425596237, 'timing/model': 1.1597723783925176, 'learning_rate/qwen_vl_interface': 7.89120594372001e-06, 'learning_rate/action_model': 7.89120594372001e-05, 'epoch': 1.02}) +[2026-06-17 20:20:19,752][starVLA.training.train_starvla][INFO] - Step 2256, Loss: {'train/loss': 0.3949664533138275, 'train/grad_norm_pre_clip': 18.341739654541016, 'timing/data': 0.0010243281722068787, 'timing/model': 1.2087260829284787, 'learning_rate/qwen_vl_interface': 7.88940814262416e-06, 'learning_rate/action_model': 7.889408142624158e-05, 'epoch': 1.02}) +[2026-06-17 20:20:23,668][starVLA.training.train_starvla][INFO] - Step 2257, Loss: {'train/loss': 0.6164372563362122, 'train/grad_norm_pre_clip': 18.262584686279297, 'timing/data': 0.0008974410593509674, 'timing/model': 0.963078748434782, 'learning_rate/qwen_vl_interface': 7.887609794376158e-06, 'learning_rate/action_model': 7.887609794376158e-05, 'epoch': 1.02}) +[2026-06-17 20:20:27,657][starVLA.training.train_starvla][INFO] - Step 2258, Loss: {'train/loss': 0.5348460078239441, 'train/grad_norm_pre_clip': 15.491192817687988, 'timing/data': 0.00101399514824152, 'timing/model': 0.9530446380376816, 'learning_rate/qwen_vl_interface': 7.88581089934881e-06, 'learning_rate/action_model': 7.88581089934881e-05, 'epoch': 1.02}) +[2026-06-17 20:20:31,509][starVLA.training.train_starvla][INFO] - Step 2259, Loss: {'train/loss': 0.42071518301963806, 'train/grad_norm_pre_clip': 16.055156707763672, 'timing/data': 0.0005830880254507065, 'timing/model': 0.9525859551504254, 'learning_rate/qwen_vl_interface': 7.884011457915025e-06, 'learning_rate/action_model': 7.884011457915024e-05, 'epoch': 1.02}) +[2026-06-17 20:20:35,389][starVLA.training.train_starvla][INFO] - Step 2260, Loss: {'train/loss': 0.47357261180877686, 'train/grad_norm_pre_clip': 16.21888542175293, 'timing/data': 0.0014672297984361649, 'timing/model': 0.9475638838484883, 'learning_rate/qwen_vl_interface': 7.882211470447832e-06, 'learning_rate/action_model': 7.882211470447831e-05, 'epoch': 1.03}) +[2026-06-17 20:20:39,211][starVLA.training.train_starvla][INFO] - Step 2261, Loss: {'train/loss': 0.6044195890426636, 'train/grad_norm_pre_clip': 15.256905555725098, 'timing/data': 0.0007167225703597069, 'timing/model': 0.9466788815334439, 'learning_rate/qwen_vl_interface': 7.880410937320367e-06, 'learning_rate/action_model': 7.880410937320368e-05, 'epoch': 1.03}) +[2026-06-17 20:20:43,137][starVLA.training.train_starvla][INFO] - Step 2262, Loss: {'train/loss': 0.3182407021522522, 'train/grad_norm_pre_clip': 20.27520751953125, 'timing/data': 0.0005349740386009216, 'timing/model': 1.1557794800028205, 'learning_rate/qwen_vl_interface': 7.878609858905885e-06, 'learning_rate/action_model': 7.878609858905886e-05, 'epoch': 1.03}) +[2026-06-17 20:20:46,821][starVLA.training.train_starvla][INFO] - Step 2263, Loss: {'train/loss': 0.2700377404689789, 'train/grad_norm_pre_clip': 7.218409538269043, 'timing/data': 0.0005229339003562927, 'timing/model': 0.9416023995727301, 'learning_rate/qwen_vl_interface': 7.876808235577751e-06, 'learning_rate/action_model': 7.87680823557775e-05, 'epoch': 1.03}) +[2026-06-17 20:20:52,284][starVLA.training.train_starvla][INFO] - Step 2264, Loss: {'train/loss': 0.39522790908813477, 'train/grad_norm_pre_clip': 13.701383590698242, 'timing/data': 0.0008600009605288506, 'timing/model': 0.9661586275324225, 'learning_rate/qwen_vl_interface': 7.875006067709444e-06, 'learning_rate/action_model': 7.875006067709443e-05, 'epoch': 1.03}) +[2026-06-17 20:20:56,137][starVLA.training.train_starvla][INFO] - Step 2265, Loss: {'train/loss': 0.2515926957130432, 'train/grad_norm_pre_clip': 7.380555629730225, 'timing/data': 0.0005214409902691841, 'timing/model': 0.9440492689609528, 'learning_rate/qwen_vl_interface': 7.873203355674553e-06, 'learning_rate/action_model': 7.873203355674552e-05, 'epoch': 1.03}) +[2026-06-17 20:20:59,966][starVLA.training.train_starvla][INFO] - Step 2266, Loss: {'train/loss': 0.37161555886268616, 'train/grad_norm_pre_clip': 12.810235977172852, 'timing/data': 0.0008468125015497208, 'timing/model': 0.9389259954914451, 'learning_rate/qwen_vl_interface': 7.871400099846784e-06, 'learning_rate/action_model': 7.871400099846783e-05, 'epoch': 1.03}) +[2026-06-17 20:21:03,936][starVLA.training.train_starvla][INFO] - Step 2267, Loss: {'train/loss': 0.5697312951087952, 'train/grad_norm_pre_clip': 15.400882720947266, 'timing/data': 0.0007444471120834351, 'timing/model': 0.9679330606013536, 'learning_rate/qwen_vl_interface': 7.869596300599954e-06, 'learning_rate/action_model': 7.869596300599953e-05, 'epoch': 1.03}) +[2026-06-17 20:21:07,988][starVLA.training.train_starvla][INFO] - Step 2268, Loss: {'train/loss': 0.6034035682678223, 'train/grad_norm_pre_clip': 15.143411636352539, 'timing/data': 0.0008353777229785919, 'timing/model': 1.0125478561967611, 'learning_rate/qwen_vl_interface': 7.867791958307992e-06, 'learning_rate/action_model': 7.867791958307992e-05, 'epoch': 1.03}) +[2026-06-17 20:21:12,046][starVLA.training.train_starvla][INFO] - Step 2269, Loss: {'train/loss': 0.4068300426006317, 'train/grad_norm_pre_clip': 13.796900749206543, 'timing/data': 0.0007263058796525002, 'timing/model': 1.204964687116444, 'learning_rate/qwen_vl_interface': 7.865987073344942e-06, 'learning_rate/action_model': 7.865987073344942e-05, 'epoch': 1.03}) +[2026-06-17 20:21:15,762][starVLA.training.train_starvla][INFO] - Step 2270, Loss: {'train/loss': 0.35383450984954834, 'train/grad_norm_pre_clip': 11.103281021118164, 'timing/data': 0.0006789891049265862, 'timing/model': 0.9810065217316151, 'learning_rate/qwen_vl_interface': 7.864181646084953e-06, 'learning_rate/action_model': 7.864181646084953e-05, 'epoch': 1.03}) +[2026-06-17 20:21:19,770][starVLA.training.train_starvla][INFO] - Step 2271, Loss: {'train/loss': 0.3595764935016632, 'train/grad_norm_pre_clip': 12.953848838806152, 'timing/data': 0.0006189560517668724, 'timing/model': 0.95433568675071, 'learning_rate/qwen_vl_interface': 7.862375676902298e-06, 'learning_rate/action_model': 7.862375676902297e-05, 'epoch': 1.03}) +[2026-06-17 20:21:24,390][starVLA.training.train_starvla][INFO] - Step 2272, Loss: {'train/loss': 0.530415952205658, 'train/grad_norm_pre_clip': 14.836660385131836, 'timing/data': 0.0005652094259858131, 'timing/model': 0.9600511230528355, 'learning_rate/qwen_vl_interface': 7.860569166171353e-06, 'learning_rate/action_model': 7.860569166171354e-05, 'epoch': 1.03}) +[2026-06-17 20:21:28,498][starVLA.training.train_starvla][INFO] - Step 2273, Loss: {'train/loss': 0.33345094323158264, 'train/grad_norm_pre_clip': 12.211844444274902, 'timing/data': 0.0010731043294072151, 'timing/model': 1.054349864833057, 'learning_rate/qwen_vl_interface': 7.858762114266613e-06, 'learning_rate/action_model': 7.858762114266612e-05, 'epoch': 1.03}) +[2026-06-17 20:21:32,460][starVLA.training.train_starvla][INFO] - Step 2274, Loss: {'train/loss': 0.25822314620018005, 'train/grad_norm_pre_clip': 11.1019868850708, 'timing/data': 0.001009361818432808, 'timing/model': 0.9521101396530867, 'learning_rate/qwen_vl_interface': 7.856954521562677e-06, 'learning_rate/action_model': 7.856954521562677e-05, 'epoch': 1.03}) +[2026-06-17 20:21:36,574][starVLA.training.train_starvla][INFO] - Step 2275, Loss: {'train/loss': 0.7304595112800598, 'train/grad_norm_pre_clip': 21.76358413696289, 'timing/data': 0.0011755432933568954, 'timing/model': 1.1891664303839207, 'learning_rate/qwen_vl_interface': 7.855146388434264e-06, 'learning_rate/action_model': 7.855146388434264e-05, 'epoch': 1.03}) +[2026-06-17 20:21:40,765][starVLA.training.train_starvla][INFO] - Step 2276, Loss: {'train/loss': 0.5855827331542969, 'train/grad_norm_pre_clip': 23.452896118164062, 'timing/data': 0.0008687656372785568, 'timing/model': 1.2118402477353811, 'learning_rate/qwen_vl_interface': 7.853337715256202e-06, 'learning_rate/action_model': 7.853337715256203e-05, 'epoch': 1.03}) +[2026-06-17 20:21:44,696][starVLA.training.train_starvla][INFO] - Step 2277, Loss: {'train/loss': 0.5013564229011536, 'train/grad_norm_pre_clip': 16.217287063598633, 'timing/data': 0.0008445894345641136, 'timing/model': 0.9492348209023476, 'learning_rate/qwen_vl_interface': 7.851528502403431e-06, 'learning_rate/action_model': 7.851528502403431e-05, 'epoch': 1.03}) +[2026-06-17 20:21:48,574][starVLA.training.train_starvla][INFO] - Step 2278, Loss: {'train/loss': 0.523888885974884, 'train/grad_norm_pre_clip': 11.318347930908203, 'timing/data': 0.0006494810804724693, 'timing/model': 0.9394336864352226, 'learning_rate/qwen_vl_interface': 7.849718750251001e-06, 'learning_rate/action_model': 7.849718750251002e-05, 'epoch': 1.03}) +[2026-06-17 20:21:52,780][starVLA.training.train_starvla][INFO] - Step 2279, Loss: {'train/loss': 0.3352334499359131, 'train/grad_norm_pre_clip': 10.576921463012695, 'timing/data': 0.0007294556125998497, 'timing/model': 0.9439196391031146, 'learning_rate/qwen_vl_interface': 7.847908459174077e-06, 'learning_rate/action_model': 7.847908459174078e-05, 'epoch': 1.03}) +[2026-06-17 20:21:56,618][starVLA.training.train_starvla][INFO] - Step 2280, Loss: {'train/loss': 0.28684574365615845, 'train/grad_norm_pre_clip': 10.343060493469238, 'timing/data': 0.0005263825878500938, 'timing/model': 0.9414752935990691, 'learning_rate/qwen_vl_interface': 7.846097629547934e-06, 'learning_rate/action_model': 7.846097629547933e-05, 'epoch': 1.03}) +[2026-06-17 20:22:00,415][starVLA.training.train_starvla][INFO] - Step 2281, Loss: {'train/loss': 0.5709332823753357, 'train/grad_norm_pre_clip': 13.673745155334473, 'timing/data': 0.0008844928815960884, 'timing/model': 0.9584753829985857, 'learning_rate/qwen_vl_interface': 7.844286261747959e-06, 'learning_rate/action_model': 7.844286261747958e-05, 'epoch': 1.03}) +[2026-06-17 20:22:04,287][starVLA.training.train_starvla][INFO] - Step 2282, Loss: {'train/loss': 0.4530198574066162, 'train/grad_norm_pre_clip': 13.845074653625488, 'timing/data': 0.00139668770134449, 'timing/model': 1.1485096225515008, 'learning_rate/qwen_vl_interface': 7.84247435614965e-06, 'learning_rate/action_model': 7.84247435614965e-05, 'epoch': 1.04}) +[2026-06-17 20:22:08,418][starVLA.training.train_starvla][INFO] - Step 2283, Loss: {'train/loss': 0.5836381316184998, 'train/grad_norm_pre_clip': 12.507685661315918, 'timing/data': 0.0005519874393939972, 'timing/model': 1.3381823915988207, 'learning_rate/qwen_vl_interface': 7.840661913128618e-06, 'learning_rate/action_model': 7.840661913128619e-05, 'epoch': 1.04}) +[2026-06-17 20:22:12,517][starVLA.training.train_starvla][INFO] - Step 2284, Loss: {'train/loss': 0.3290565609931946, 'train/grad_norm_pre_clip': 11.525738716125488, 'timing/data': 0.000625384971499443, 'timing/model': 0.948707896284759, 'learning_rate/qwen_vl_interface': 7.838848933060583e-06, 'learning_rate/action_model': 7.838848933060583e-05, 'epoch': 1.04}) +[2026-06-17 20:22:16,503][starVLA.training.train_starvla][INFO] - Step 2285, Loss: {'train/loss': 0.43093207478523254, 'train/grad_norm_pre_clip': 11.690272331237793, 'timing/data': 0.0008351663127541542, 'timing/model': 0.9252756796777248, 'learning_rate/qwen_vl_interface': 7.83703541632138e-06, 'learning_rate/action_model': 7.83703541632138e-05, 'epoch': 1.04}) +[2026-06-17 20:22:20,515][starVLA.training.train_starvla][INFO] - Step 2286, Loss: {'train/loss': 0.3381463885307312, 'train/grad_norm_pre_clip': 9.479482650756836, 'timing/data': 0.0010339664295315742, 'timing/model': 0.949498139321804, 'learning_rate/qwen_vl_interface': 7.835221363286949e-06, 'learning_rate/action_model': 7.835221363286949e-05, 'epoch': 1.04}) +[2026-06-17 20:22:24,589][starVLA.training.train_starvla][INFO] - Step 2287, Loss: {'train/loss': 0.5015164017677307, 'train/grad_norm_pre_clip': 30.883317947387695, 'timing/data': 0.0005273157730698586, 'timing/model': 1.0567825473845005, 'learning_rate/qwen_vl_interface': 7.83340677433335e-06, 'learning_rate/action_model': 7.83340677433335e-05, 'epoch': 1.04}) +[2026-06-17 20:22:28,451][starVLA.training.train_starvla][INFO] - Step 2288, Loss: {'train/loss': 0.4878382980823517, 'train/grad_norm_pre_clip': 13.418094635009766, 'timing/data': 0.000529976561665535, 'timing/model': 0.9523519100621343, 'learning_rate/qwen_vl_interface': 7.831591649836745e-06, 'learning_rate/action_model': 7.831591649836744e-05, 'epoch': 1.04}) +[2026-06-17 20:22:32,360][starVLA.training.train_starvla][INFO] - Step 2289, Loss: {'train/loss': 0.330201119184494, 'train/grad_norm_pre_clip': 10.469797134399414, 'timing/data': 0.0012384401634335518, 'timing/model': 1.191214439459145, 'learning_rate/qwen_vl_interface': 7.829775990173413e-06, 'learning_rate/action_model': 7.829775990173413e-05, 'epoch': 1.04}) +[2026-06-17 20:22:36,106][starVLA.training.train_starvla][INFO] - Step 2290, Loss: {'train/loss': 0.6805821061134338, 'train/grad_norm_pre_clip': 15.968993186950684, 'timing/data': 0.0005686096847057343, 'timing/model': 0.937199261970818, 'learning_rate/qwen_vl_interface': 7.827959795719744e-06, 'learning_rate/action_model': 7.827959795719744e-05, 'epoch': 1.04}) +[2026-06-17 20:22:40,117][starVLA.training.train_starvla][INFO] - Step 2291, Loss: {'train/loss': 0.33249908685684204, 'train/grad_norm_pre_clip': 13.086090087890625, 'timing/data': 0.0007777130231261253, 'timing/model': 1.0445046788081527, 'learning_rate/qwen_vl_interface': 7.826143066852233e-06, 'learning_rate/action_model': 7.826143066852233e-05, 'epoch': 1.04}) +[2026-06-17 20:22:44,699][starVLA.training.train_starvla][INFO] - Step 2292, Loss: {'train/loss': 0.4274362027645111, 'train/grad_norm_pre_clip': 12.713098526000977, 'timing/data': 0.000578979030251503, 'timing/model': 0.9360471554100513, 'learning_rate/qwen_vl_interface': 7.824325803947494e-06, 'learning_rate/action_model': 7.824325803947495e-05, 'epoch': 1.04}) +[2026-06-17 20:22:48,577][starVLA.training.train_starvla][INFO] - Step 2293, Loss: {'train/loss': 0.5020738244056702, 'train/grad_norm_pre_clip': 15.140630722045898, 'timing/data': 0.000905454158782959, 'timing/model': 0.9489872604608536, 'learning_rate/qwen_vl_interface': 7.822508007382246e-06, 'learning_rate/action_model': 7.822508007382245e-05, 'epoch': 1.04}) +[2026-06-17 20:22:52,511][starVLA.training.train_starvla][INFO] - Step 2294, Loss: {'train/loss': 0.40428104996681213, 'train/grad_norm_pre_clip': 15.8935546875, 'timing/data': 0.0005838898941874504, 'timing/model': 0.9577714167535305, 'learning_rate/qwen_vl_interface': 7.820689677533319e-06, 'learning_rate/action_model': 7.82068967753332e-05, 'epoch': 1.04}) +[2026-06-17 20:22:56,497][starVLA.training.train_starvla][INFO] - Step 2295, Loss: {'train/loss': 0.44015392661094666, 'train/grad_norm_pre_clip': 12.772862434387207, 'timing/data': 0.0011050058528780937, 'timing/model': 1.206890569999814, 'learning_rate/qwen_vl_interface': 7.818870814777655e-06, 'learning_rate/action_model': 7.818870814777655e-05, 'epoch': 1.04}) +[2026-06-17 20:23:00,468][starVLA.training.train_starvla][INFO] - Step 2296, Loss: {'train/loss': 0.398221880197525, 'train/grad_norm_pre_clip': 12.083719253540039, 'timing/data': 0.0012725172564387321, 'timing/model': 1.2392308646813035, 'learning_rate/qwen_vl_interface': 7.81705141949231e-06, 'learning_rate/action_model': 7.817051419492308e-05, 'epoch': 1.04}) +[2026-06-17 20:23:04,257][starVLA.training.train_starvla][INFO] - Step 2297, Loss: {'train/loss': 0.5785155892372131, 'train/grad_norm_pre_clip': 15.025179862976074, 'timing/data': 0.0006459038704633713, 'timing/model': 1.008210949599743, 'learning_rate/qwen_vl_interface': 7.81523149205444e-06, 'learning_rate/action_model': 7.81523149205444e-05, 'epoch': 1.04}) +[2026-06-17 20:23:08,350][starVLA.training.train_starvla][INFO] - Step 2298, Loss: {'train/loss': 0.5568565726280212, 'train/grad_norm_pre_clip': 13.724217414855957, 'timing/data': 0.000970412977039814, 'timing/model': 1.0127977104857564, 'learning_rate/qwen_vl_interface': 7.813411032841323e-06, 'learning_rate/action_model': 7.813411032841323e-05, 'epoch': 1.04}) +[2026-06-17 20:23:12,354][starVLA.training.train_starvla][INFO] - Step 2299, Loss: {'train/loss': 0.5105207562446594, 'train/grad_norm_pre_clip': 36.53931427001953, 'timing/data': 0.0006464878097176552, 'timing/model': 0.9608150515705347, 'learning_rate/qwen_vl_interface': 7.81159004223034e-06, 'learning_rate/action_model': 7.81159004223034e-05, 'epoch': 1.04}) +[2026-06-17 20:25:31,035][starVLA.training.train_starvla][INFO] - Step 2300, Loss: {'train/loss': 0.2918035089969635, 'train/grad_norm_pre_clip': 11.51010513305664, 'eval/loss': 0.4453102946281433, 'eval/latency_8/loss': 0.4453102946281433, 'eval/demon_fire_precision': 0.9493945979509469, 'eval/demon_fire_recall': 0.9499844672258465, 'eval/demon_fire_f1': 0.9496894409937888, 'eval/demon_fire_rate_model': 0.8962159154145799, 'eval/demon_fire_rate_teacher': 0.8956594323873122, 'eval/demon_left_precision': 0.7414529914529915, 'eval/demon_left_recall': 0.7430406852248393, 'eval/demon_left_f1': 0.7422459893048129, 'eval/demon_left_rate_model': 0.1302170283806344, 'eval/demon_left_rate_teacher': 0.12993878686700056, 'eval/demon_right_precision': 0.7906976744186046, 'eval/demon_right_recall': 0.7212121212121212, 'eval/demon_right_f1': 0.7543581616481774, 'eval/demon_right_rate_model': 0.16750139120756816, 'eval/demon_right_rate_teacher': 0.18363939899833054, 'eval/demon_move_f1': 0.7483020754764951, 'eval/demon_cc_f1': 0.878741822485254, 'eval/latency_8/demon_fire_precision': 0.9493945979509469, 'eval/latency_8/demon_fire_recall': 0.9499844672258465, 'eval/latency_8/demon_fire_f1': 0.9496894409937888, 'eval/latency_8/demon_fire_rate_model': 0.8962159154145799, 'eval/latency_8/demon_fire_rate_teacher': 0.8956594323873122, 'eval/latency_8/demon_left_precision': 0.7414529914529915, 'eval/latency_8/demon_left_recall': 0.7430406852248393, 'eval/latency_8/demon_left_f1': 0.7422459893048129, 'eval/latency_8/demon_left_rate_model': 0.1302170283806344, 'eval/latency_8/demon_left_rate_teacher': 0.12993878686700056, 'eval/latency_8/demon_right_precision': 0.7906976744186046, 'eval/latency_8/demon_right_recall': 0.7212121212121212, 'eval/latency_8/demon_right_f1': 0.7543581616481774, 'eval/latency_8/demon_right_rate_model': 0.16750139120756816, 'eval/latency_8/demon_right_rate_teacher': 0.18363939899833054, 'eval/latency_8/demon_move_f1': 0.7483020754764951, 'eval/latency_8/demon_cc_f1': 0.878741822485254, 'timing/data': 0.0007815444841980934, 'timing/model': 0.9862493108958006, 'learning_rate/qwen_vl_interface': 7.809768520598986e-06, 'learning_rate/action_model': 7.809768520598984e-05, 'epoch': 1.04}) +[2026-06-17 20:25:31,037][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_2300_state +[2026-06-17 20:25:31,041][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 20:25:38,115][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_2300_state/model.safetensors +[2026-06-17 20:25:54,469][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_2300_state/optimizer.bin +[2026-06-17 20:25:54,470][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_2300_state/sampler.bin +[2026-06-17 20:25:54,470][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_2300_state/sampler_1.bin +[2026-06-17 20:25:54,475][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_2300_state/random_states_0.pkl +[2026-06-17 20:25:54,482][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2300 +[2026-06-17 20:25:54,482][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 20:25:54,495][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 20:25:58,340][starVLA.training.train_starvla][INFO] - Step 2301, Loss: {'train/loss': 0.5202246308326721, 'train/grad_norm_pre_clip': 12.672072410583496, 'timing/data': 0.0007081916555762291, 'timing/model': 0.9384558666497469, 'learning_rate/qwen_vl_interface': 7.80794646832486e-06, 'learning_rate/action_model': 7.80794646832486e-05, 'epoch': 1.04}) +[2026-06-17 20:26:02,173][starVLA.training.train_starvla][INFO] - Step 2302, Loss: {'train/loss': 0.29601752758026123, 'train/grad_norm_pre_clip': 10.647835731506348, 'timing/data': 0.0005504777655005455, 'timing/model': 0.9416906973347068, 'learning_rate/qwen_vl_interface': 7.80612388578568e-06, 'learning_rate/action_model': 7.80612388578568e-05, 'epoch': 1.04}) +[2026-06-17 20:26:06,034][starVLA.training.train_starvla][INFO] - Step 2303, Loss: {'train/loss': 0.6741721630096436, 'train/grad_norm_pre_clip': 14.82612133026123, 'timing/data': 0.0005752081051468849, 'timing/model': 0.9662489211186767, 'learning_rate/qwen_vl_interface': 7.80430077335927e-06, 'learning_rate/action_model': 7.80430077335927e-05, 'epoch': 1.04}) +[2026-06-17 20:26:09,821][starVLA.training.train_starvla][INFO] - Step 2304, Loss: {'train/loss': 0.569553554058075, 'train/grad_norm_pre_clip': 16.168380737304688, 'timing/data': 0.0005078203976154327, 'timing/model': 1.1142737390473485, 'learning_rate/qwen_vl_interface': 7.802477131423561e-06, 'learning_rate/action_model': 7.80247713142356e-05, 'epoch': 1.05}) +[2026-06-17 20:26:13,611][starVLA.training.train_starvla][INFO] - Step 2305, Loss: {'train/loss': 0.20243437588214874, 'train/grad_norm_pre_clip': 7.984750747680664, 'timing/data': 0.000979917123913765, 'timing/model': 1.1065240940079093, 'learning_rate/qwen_vl_interface': 7.800652960356594e-06, 'learning_rate/action_model': 7.800652960356594e-05, 'epoch': 1.05}) +[2026-06-17 20:26:17,320][starVLA.training.train_starvla][INFO] - Step 2306, Loss: {'train/loss': 0.18970313668251038, 'train/grad_norm_pre_clip': 7.865107536315918, 'timing/data': 0.0005217595025897026, 'timing/model': 0.944066590629518, 'learning_rate/qwen_vl_interface': 7.798828260536524e-06, 'learning_rate/action_model': 7.798828260536524e-05, 'epoch': 1.05}) +[2026-06-17 20:26:21,245][starVLA.training.train_starvla][INFO] - Step 2307, Loss: {'train/loss': 0.3709186613559723, 'train/grad_norm_pre_clip': 9.574262619018555, 'timing/data': 0.0011435039341449738, 'timing/model': 0.9726459961384535, 'learning_rate/qwen_vl_interface': 7.797003032341615e-06, 'learning_rate/action_model': 7.797003032341614e-05, 'epoch': 1.05}) +[2026-06-17 20:26:25,123][starVLA.training.train_starvla][INFO] - Step 2308, Loss: {'train/loss': 0.39748692512512207, 'train/grad_norm_pre_clip': 11.035333633422852, 'timing/data': 0.0008958587422966957, 'timing/model': 0.9876112630590796, 'learning_rate/qwen_vl_interface': 7.795177276150234e-06, 'learning_rate/action_model': 7.795177276150234e-05, 'epoch': 1.05}) +[2026-06-17 20:26:30,131][starVLA.training.train_starvla][INFO] - Step 2309, Loss: {'train/loss': 1.064484715461731, 'train/grad_norm_pre_clip': 24.4315185546875, 'timing/data': 0.0008091935887932777, 'timing/model': 1.9726776173338294, 'learning_rate/qwen_vl_interface': 7.793350992340867e-06, 'learning_rate/action_model': 7.793350992340867e-05, 'epoch': 1.05}) +[2026-06-17 20:26:34,039][starVLA.training.train_starvla][INFO] - Step 2310, Loss: {'train/loss': 0.4122997224330902, 'train/grad_norm_pre_clip': 12.114670753479004, 'timing/data': 0.0008957730606198311, 'timing/model': 0.9338116943836212, 'learning_rate/qwen_vl_interface': 7.791524181292102e-06, 'learning_rate/action_model': 7.791524181292101e-05, 'epoch': 1.05}) +[2026-06-17 20:26:37,861][starVLA.training.train_starvla][INFO] - Step 2311, Loss: {'train/loss': 0.5143656730651855, 'train/grad_norm_pre_clip': 11.334586143493652, 'timing/data': 0.0005557183176279068, 'timing/model': 1.0975591000169516, 'learning_rate/qwen_vl_interface': 7.789696843382638e-06, 'learning_rate/action_model': 7.789696843382639e-05, 'epoch': 1.05}) +[2026-06-17 20:26:41,529][starVLA.training.train_starvla][INFO] - Step 2312, Loss: {'train/loss': 0.3793562948703766, 'train/grad_norm_pre_clip': 13.974774360656738, 'timing/data': 0.0005010496824979782, 'timing/model': 0.9605909055098891, 'learning_rate/qwen_vl_interface': 7.787868978991288e-06, 'learning_rate/action_model': 7.787868978991287e-05, 'epoch': 1.05}) +[2026-06-17 20:26:45,382][starVLA.training.train_starvla][INFO] - Step 2313, Loss: {'train/loss': 0.460165798664093, 'train/grad_norm_pre_clip': 21.601905822753906, 'timing/data': 0.0005275839939713478, 'timing/model': 0.9620304238051176, 'learning_rate/qwen_vl_interface': 7.786040588496966e-06, 'learning_rate/action_model': 7.786040588496965e-05, 'epoch': 1.05}) +[2026-06-17 20:26:49,285][starVLA.training.train_starvla][INFO] - Step 2314, Loss: {'train/loss': 0.42383262515068054, 'train/grad_norm_pre_clip': 11.526745796203613, 'timing/data': 0.0006712330505251884, 'timing/model': 0.9521767012774944, 'learning_rate/qwen_vl_interface': 7.784211672278703e-06, 'learning_rate/action_model': 7.784211672278703e-05, 'epoch': 1.05}) +[2026-06-17 20:26:53,138][starVLA.training.train_starvla][INFO] - Step 2315, Loss: {'train/loss': 0.37882864475250244, 'train/grad_norm_pre_clip': 12.865527153015137, 'timing/data': 0.0007133232429623604, 'timing/model': 0.9524762639775872, 'learning_rate/qwen_vl_interface': 7.78238223071563e-06, 'learning_rate/action_model': 7.78238223071563e-05, 'epoch': 1.05}) +[2026-06-17 20:26:57,017][starVLA.training.train_starvla][INFO] - Step 2316, Loss: {'train/loss': 0.5877005457878113, 'train/grad_norm_pre_clip': 15.54934024810791, 'timing/data': 0.0015844125300645828, 'timing/model': 0.9466034332290292, 'learning_rate/qwen_vl_interface': 7.780552264187e-06, 'learning_rate/action_model': 7.780552264186999e-05, 'epoch': 1.05}) +[2026-06-17 20:27:00,808][starVLA.training.train_starvla][INFO] - Step 2317, Loss: {'train/loss': 0.6071749329566956, 'train/grad_norm_pre_clip': 17.05265235900879, 'timing/data': 0.0009135864675045013, 'timing/model': 0.9462767969816923, 'learning_rate/qwen_vl_interface': 7.77872177307216e-06, 'learning_rate/action_model': 7.778721773072159e-05, 'epoch': 1.05}) +[2026-06-17 20:27:04,795][starVLA.training.train_starvla][INFO] - Step 2318, Loss: {'train/loss': 0.3834853172302246, 'train/grad_norm_pre_clip': 19.183637619018555, 'timing/data': 0.000581003725528717, 'timing/model': 1.123126775957644, 'learning_rate/qwen_vl_interface': 7.776890757750575e-06, 'learning_rate/action_model': 7.776890757750575e-05, 'epoch': 1.05}) +[2026-06-17 20:27:08,444][starVLA.training.train_starvla][INFO] - Step 2319, Loss: {'train/loss': 0.3461460769176483, 'train/grad_norm_pre_clip': 11.424848556518555, 'timing/data': 0.0012476909905672073, 'timing/model': 0.9557141866534948, 'learning_rate/qwen_vl_interface': 7.775059218601816e-06, 'learning_rate/action_model': 7.775059218601816e-05, 'epoch': 1.05}) +[2026-06-17 20:27:12,282][starVLA.training.train_starvla][INFO] - Step 2320, Loss: {'train/loss': 0.2626487910747528, 'train/grad_norm_pre_clip': 11.222582817077637, 'timing/data': 0.0005263378843665123, 'timing/model': 0.9391464544460177, 'learning_rate/qwen_vl_interface': 7.773227156005567e-06, 'learning_rate/action_model': 7.773227156005567e-05, 'epoch': 1.05}) +[2026-06-17 20:27:16,149][starVLA.training.train_starvla][INFO] - Step 2321, Loss: {'train/loss': 0.44393691420555115, 'train/grad_norm_pre_clip': 14.406306266784668, 'timing/data': 0.0010776864364743233, 'timing/model': 0.9597054459154606, 'learning_rate/qwen_vl_interface': 7.771394570341612e-06, 'learning_rate/action_model': 7.771394570341612e-05, 'epoch': 1.05}) +[2026-06-17 20:27:19,993][starVLA.training.train_starvla][INFO] - Step 2322, Loss: {'train/loss': 0.7023468017578125, 'train/grad_norm_pre_clip': 15.806120872497559, 'timing/data': 0.0005221124738454819, 'timing/model': 0.9502229001373053, 'learning_rate/qwen_vl_interface': 7.769561461989848e-06, 'learning_rate/action_model': 7.769561461989848e-05, 'epoch': 1.05}) +[2026-06-17 20:27:24,185][starVLA.training.train_starvla][INFO] - Step 2323, Loss: {'train/loss': 0.39791804552078247, 'train/grad_norm_pre_clip': 12.731282234191895, 'timing/data': 0.0009259805083274841, 'timing/model': 1.2942674960941076, 'learning_rate/qwen_vl_interface': 7.767727831330285e-06, 'learning_rate/action_model': 7.767727831330285e-05, 'epoch': 1.05}) +[2026-06-17 20:27:28,196][starVLA.training.train_starvla][INFO] - Step 2324, Loss: {'train/loss': 0.38590073585510254, 'train/grad_norm_pre_clip': 12.007562637329102, 'timing/data': 0.0006660223007202148, 'timing/model': 1.2024791026487947, 'learning_rate/qwen_vl_interface': 7.765893678743029e-06, 'learning_rate/action_model': 7.76589367874303e-05, 'epoch': 1.05}) +[2026-06-17 20:27:31,868][starVLA.training.train_starvla][INFO] - Step 2325, Loss: {'train/loss': 0.43032020330429077, 'train/grad_norm_pre_clip': 13.568370819091797, 'timing/data': 0.0005095247179269791, 'timing/model': 0.9455371992662549, 'learning_rate/qwen_vl_interface': 7.76405900460831e-06, 'learning_rate/action_model': 7.76405900460831e-05, 'epoch': 1.05}) +[2026-06-17 20:27:35,756][starVLA.training.train_starvla][INFO] - Step 2326, Loss: {'train/loss': 0.41505318880081177, 'train/grad_norm_pre_clip': 13.422711372375488, 'timing/data': 0.0009965449571609497, 'timing/model': 0.9600932141765952, 'learning_rate/qwen_vl_interface': 7.762223809306453e-06, 'learning_rate/action_model': 7.762223809306452e-05, 'epoch': 1.06}) +[2026-06-17 20:27:39,598][starVLA.training.train_starvla][INFO] - Step 2327, Loss: {'train/loss': 0.5082342624664307, 'train/grad_norm_pre_clip': 14.544971466064453, 'timing/data': 0.0005103060975670815, 'timing/model': 0.9454788183793426, 'learning_rate/qwen_vl_interface': 7.760388093217897e-06, 'learning_rate/action_model': 7.760388093217898e-05, 'epoch': 1.06}) +[2026-06-17 20:27:44,472][starVLA.training.train_starvla][INFO] - Step 2328, Loss: {'train/loss': 0.47363319993019104, 'train/grad_norm_pre_clip': 13.00738525390625, 'timing/data': 0.0005145939067006111, 'timing/model': 1.6915122009813786, 'learning_rate/qwen_vl_interface': 7.758551856723187e-06, 'learning_rate/action_model': 7.758551856723188e-05, 'epoch': 1.06}) +[2026-06-17 20:27:48,932][starVLA.training.train_starvla][INFO] - Step 2329, Loss: {'train/loss': 0.4046143889427185, 'train/grad_norm_pre_clip': 16.910776138305664, 'timing/data': 0.0008371509611606598, 'timing/model': 1.1676121577620506, 'learning_rate/qwen_vl_interface': 7.756715100202979e-06, 'learning_rate/action_model': 7.75671510020298e-05, 'epoch': 1.06}) +[2026-06-17 20:27:52,727][starVLA.training.train_starvla][INFO] - Step 2330, Loss: {'train/loss': 0.29632458090782166, 'train/grad_norm_pre_clip': 11.556458473205566, 'timing/data': 0.0005090385675430298, 'timing/model': 1.1077511152252555, 'learning_rate/qwen_vl_interface': 7.754877824038035e-06, 'learning_rate/action_model': 7.754877824038035e-05, 'epoch': 1.06}) +[2026-06-17 20:27:56,509][starVLA.training.train_starvla][INFO] - Step 2331, Loss: {'train/loss': 0.38003313541412354, 'train/grad_norm_pre_clip': 38.5745964050293, 'timing/data': 0.0006823921576142311, 'timing/model': 1.102106286212802, 'learning_rate/qwen_vl_interface': 7.75304002860922e-06, 'learning_rate/action_model': 7.75304002860922e-05, 'epoch': 1.06}) +[2026-06-17 20:28:00,153][starVLA.training.train_starvla][INFO] - Step 2332, Loss: {'train/loss': 0.38910672068595886, 'train/grad_norm_pre_clip': 60.41568374633789, 'timing/data': 0.0004944391548633575, 'timing/model': 0.9778857668861747, 'learning_rate/qwen_vl_interface': 7.751201714297513e-06, 'learning_rate/action_model': 7.751201714297512e-05, 'epoch': 1.06}) +[2026-06-17 20:28:04,045][starVLA.training.train_starvla][INFO] - Step 2333, Loss: {'train/loss': 0.4248993992805481, 'train/grad_norm_pre_clip': 12.465109825134277, 'timing/data': 0.000501084141433239, 'timing/model': 0.9780629631131887, 'learning_rate/qwen_vl_interface': 7.749362881484002e-06, 'learning_rate/action_model': 7.749362881484001e-05, 'epoch': 1.06}) +[2026-06-17 20:28:07,882][starVLA.training.train_starvla][INFO] - Step 2334, Loss: {'train/loss': 0.669607937335968, 'train/grad_norm_pre_clip': 14.917034149169922, 'timing/data': 0.0004857005551457405, 'timing/model': 0.9416464073583484, 'learning_rate/qwen_vl_interface': 7.747523530549874e-06, 'learning_rate/action_model': 7.747523530549875e-05, 'epoch': 1.06}) +[2026-06-17 20:28:11,742][starVLA.training.train_starvla][INFO] - Step 2335, Loss: {'train/loss': 0.3317616581916809, 'train/grad_norm_pre_clip': 14.63786792755127, 'timing/data': 0.0006709704175591469, 'timing/model': 0.9500790536403656, 'learning_rate/qwen_vl_interface': 7.745683661876432e-06, 'learning_rate/action_model': 7.745683661876432e-05, 'epoch': 1.06}) +[2026-06-17 20:28:15,573][starVLA.training.train_starvla][INFO] - Step 2336, Loss: {'train/loss': 0.4760127067565918, 'train/grad_norm_pre_clip': 14.553960800170898, 'timing/data': 0.0005838442593812943, 'timing/model': 0.9476152705028653, 'learning_rate/qwen_vl_interface': 7.743843275845081e-06, 'learning_rate/action_model': 7.74384327584508e-05, 'epoch': 1.06}) +[2026-06-17 20:28:19,454][starVLA.training.train_starvla][INFO] - Step 2337, Loss: {'train/loss': 0.5061586499214172, 'train/grad_norm_pre_clip': 24.60893440246582, 'timing/data': 0.0006081247702240944, 'timing/model': 1.1617686282843351, 'learning_rate/qwen_vl_interface': 7.742002372837336e-06, 'learning_rate/action_model': 7.742002372837335e-05, 'epoch': 1.06}) +[2026-06-17 20:28:23,205][starVLA.training.train_starvla][INFO] - Step 2338, Loss: {'train/loss': 0.2968659996986389, 'train/grad_norm_pre_clip': 10.698155403137207, 'timing/data': 0.000866960734128952, 'timing/model': 1.0124665340408683, 'learning_rate/qwen_vl_interface': 7.740160953234815e-06, 'learning_rate/action_model': 7.740160953234815e-05, 'epoch': 1.06}) +[2026-06-17 20:28:27,062][starVLA.training.train_starvla][INFO] - Step 2339, Loss: {'train/loss': 0.5436383485794067, 'train/grad_norm_pre_clip': 13.771125793457031, 'timing/data': 0.0007578190416097641, 'timing/model': 0.9586385991424322, 'learning_rate/qwen_vl_interface': 7.738319017419252e-06, 'learning_rate/action_model': 7.738319017419252e-05, 'epoch': 1.06}) +[2026-06-17 20:28:31,326][starVLA.training.train_starvla][INFO] - Step 2340, Loss: {'train/loss': 0.5894468426704407, 'train/grad_norm_pre_clip': 22.201562881469727, 'timing/data': 0.0007625753059983253, 'timing/model': 1.1481034522876143, 'learning_rate/qwen_vl_interface': 7.736476565772478e-06, 'learning_rate/action_model': 7.736476565772478e-05, 'epoch': 1.06}) +[2026-06-17 20:28:35,237][starVLA.training.train_starvla][INFO] - Step 2341, Loss: {'train/loss': 0.3838784694671631, 'train/grad_norm_pre_clip': 12.086197853088379, 'timing/data': 0.0008837198838591576, 'timing/model': 0.9450557567179203, 'learning_rate/qwen_vl_interface': 7.734633598676435e-06, 'learning_rate/action_model': 7.734633598676434e-05, 'epoch': 1.06}) +[2026-06-17 20:28:39,411][starVLA.training.train_starvla][INFO] - Step 2342, Loss: {'train/loss': 0.8437212109565735, 'train/grad_norm_pre_clip': 17.100088119506836, 'timing/data': 0.0012285839766263962, 'timing/model': 1.2467340715229511, 'learning_rate/qwen_vl_interface': 7.732790116513176e-06, 'learning_rate/action_model': 7.732790116513176e-05, 'epoch': 1.06}) +[2026-06-17 20:28:45,258][starVLA.training.train_starvla][INFO] - Step 2343, Loss: {'train/loss': 0.45207086205482483, 'train/grad_norm_pre_clip': 14.520574569702148, 'timing/data': 0.0005205702036619186, 'timing/model': 1.203622955828905, 'learning_rate/qwen_vl_interface': 7.730946119664852e-06, 'learning_rate/action_model': 7.730946119664853e-05, 'epoch': 1.06}) +[2026-06-17 20:28:49,096][starVLA.training.train_starvla][INFO] - Step 2344, Loss: {'train/loss': 0.41504940390586853, 'train/grad_norm_pre_clip': 12.906717300415039, 'timing/data': 0.0005144365131855011, 'timing/model': 1.0744303679093719, 'learning_rate/qwen_vl_interface': 7.729101608513729e-06, 'learning_rate/action_model': 7.72910160851373e-05, 'epoch': 1.06}) +[2026-06-17 20:28:53,014][starVLA.training.train_starvla][INFO] - Step 2345, Loss: {'train/loss': 0.41744813323020935, 'train/grad_norm_pre_clip': 12.899221420288086, 'timing/data': 0.0005136085674166679, 'timing/model': 1.0569515246897936, 'learning_rate/qwen_vl_interface': 7.727256583442175e-06, 'learning_rate/action_model': 7.727256583442175e-05, 'epoch': 1.06}) +[2026-06-17 20:28:56,955][starVLA.training.train_starvla][INFO] - Step 2346, Loss: {'train/loss': 0.39768868684768677, 'train/grad_norm_pre_clip': 14.471405029296875, 'timing/data': 0.0004726368933916092, 'timing/model': 1.0031990483403206, 'learning_rate/qwen_vl_interface': 7.725411044832665e-06, 'learning_rate/action_model': 7.725411044832665e-05, 'epoch': 1.06}) +[2026-06-17 20:29:00,986][starVLA.training.train_starvla][INFO] - Step 2347, Loss: {'train/loss': 0.47819384932518005, 'train/grad_norm_pre_clip': 11.231820106506348, 'timing/data': 0.0008878922089934349, 'timing/model': 0.996482364833355, 'learning_rate/qwen_vl_interface': 7.723564993067783e-06, 'learning_rate/action_model': 7.723564993067783e-05, 'epoch': 1.06}) +[2026-06-17 20:29:04,914][starVLA.training.train_starvla][INFO] - Step 2348, Loss: {'train/loss': 0.4700343608856201, 'train/grad_norm_pre_clip': 11.348320960998535, 'timing/data': 0.0008194399997591972, 'timing/model': 0.9543948117643595, 'learning_rate/qwen_vl_interface': 7.721718428530214e-06, 'learning_rate/action_model': 7.721718428530214e-05, 'epoch': 1.07}) +[2026-06-17 20:29:08,731][starVLA.training.train_starvla][INFO] - Step 2349, Loss: {'train/loss': 0.4678211808204651, 'train/grad_norm_pre_clip': 14.574094772338867, 'timing/data': 0.000821869820356369, 'timing/model': 0.9324640762060881, 'learning_rate/qwen_vl_interface': 7.719871351602757e-06, 'learning_rate/action_model': 7.719871351602757e-05, 'epoch': 1.07}) +[2026-06-17 20:29:12,619][starVLA.training.train_starvla][INFO] - Step 2350, Loss: {'train/loss': 0.28739407658576965, 'train/grad_norm_pre_clip': 8.433221817016602, 'timing/data': 0.0005368897691369057, 'timing/model': 1.13664925750345, 'learning_rate/qwen_vl_interface': 7.718023762668309e-06, 'learning_rate/action_model': 7.718023762668309e-05, 'epoch': 1.07}) +[2026-06-17 20:29:16,298][starVLA.training.train_starvla][INFO] - Step 2351, Loss: {'train/loss': 0.3897588551044464, 'train/grad_norm_pre_clip': 11.218515396118164, 'timing/data': 0.0010412530973553658, 'timing/model': 1.0103623699396849, 'learning_rate/qwen_vl_interface': 7.716175662109882e-06, 'learning_rate/action_model': 7.716175662109881e-05, 'epoch': 1.07}) +[2026-06-17 20:29:20,083][starVLA.training.train_starvla][INFO] - Step 2352, Loss: {'train/loss': 0.2979142963886261, 'train/grad_norm_pre_clip': 9.453559875488281, 'timing/data': 0.0004797438159584999, 'timing/model': 0.9321653144434094, 'learning_rate/qwen_vl_interface': 7.714327050310586e-06, 'learning_rate/action_model': 7.714327050310585e-05, 'epoch': 1.07}) +[2026-06-17 20:29:23,926][starVLA.training.train_starvla][INFO] - Step 2353, Loss: {'train/loss': 0.7122642993927002, 'train/grad_norm_pre_clip': 17.029991149902344, 'timing/data': 0.0005176384001970291, 'timing/model': 0.952953614294529, 'learning_rate/qwen_vl_interface': 7.71247792765364e-06, 'learning_rate/action_model': 7.712477927653638e-05, 'epoch': 1.07}) +[2026-06-17 20:29:27,764][starVLA.training.train_starvla][INFO] - Step 2354, Loss: {'train/loss': 0.5428532958030701, 'train/grad_norm_pre_clip': 13.064797401428223, 'timing/data': 0.0005370210856199265, 'timing/model': 0.9372747419402003, 'learning_rate/qwen_vl_interface': 7.71062829452237e-06, 'learning_rate/action_model': 7.710628294522369e-05, 'epoch': 1.07}) +[2026-06-17 20:29:31,614][starVLA.training.train_starvla][INFO] - Step 2355, Loss: {'train/loss': 0.6976228356361389, 'train/grad_norm_pre_clip': 17.003515243530273, 'timing/data': 0.0005148416385054588, 'timing/model': 0.9480521362274885, 'learning_rate/qwen_vl_interface': 7.708778151300206e-06, 'learning_rate/action_model': 7.708778151300206e-05, 'epoch': 1.07}) +[2026-06-17 20:29:35,434][starVLA.training.train_starvla][INFO] - Step 2356, Loss: {'train/loss': 0.5740746259689331, 'train/grad_norm_pre_clip': 13.672264099121094, 'timing/data': 0.0007042884826660156, 'timing/model': 0.9538524886593223, 'learning_rate/qwen_vl_interface': 7.706927498370687e-06, 'learning_rate/action_model': 7.706927498370686e-05, 'epoch': 1.07}) +[2026-06-17 20:29:39,566][starVLA.training.train_starvla][INFO] - Step 2357, Loss: {'train/loss': 0.2975538372993469, 'train/grad_norm_pre_clip': 8.609091758728027, 'timing/data': 0.0008461037650704384, 'timing/model': 1.1741130352020264, 'learning_rate/qwen_vl_interface': 7.705076336117452e-06, 'learning_rate/action_model': 7.705076336117451e-05, 'epoch': 1.07}) +[2026-06-17 20:29:43,703][starVLA.training.train_starvla][INFO] - Step 2358, Loss: {'train/loss': 0.8698392510414124, 'train/grad_norm_pre_clip': 20.883634567260742, 'timing/data': 0.0004935311153531075, 'timing/model': 0.9594432720914483, 'learning_rate/qwen_vl_interface': 7.703224664924252e-06, 'learning_rate/action_model': 7.703224664924251e-05, 'epoch': 1.07}) +[2026-06-17 20:29:47,542][starVLA.training.train_starvla][INFO] - Step 2359, Loss: {'train/loss': 0.3793066740036011, 'train/grad_norm_pre_clip': 11.361291885375977, 'timing/data': 0.0006962958723306656, 'timing/model': 0.9385953946039081, 'learning_rate/qwen_vl_interface': 7.70137248517494e-06, 'learning_rate/action_model': 7.701372485174939e-05, 'epoch': 1.07}) +[2026-06-17 20:29:51,437][starVLA.training.train_starvla][INFO] - Step 2360, Loss: {'train/loss': 0.4597921073436737, 'train/grad_norm_pre_clip': 13.507085800170898, 'timing/data': 0.0010232292115688324, 'timing/model': 1.0125232059508562, 'learning_rate/qwen_vl_interface': 7.699519797253471e-06, 'learning_rate/action_model': 7.699519797253472e-05, 'epoch': 1.07}) +[2026-06-17 20:29:55,267][starVLA.training.train_starvla][INFO] - Step 2361, Loss: {'train/loss': 0.25719887018203735, 'train/grad_norm_pre_clip': 10.620686531066895, 'timing/data': 0.0007596360519528389, 'timing/model': 0.938386526890099, 'learning_rate/qwen_vl_interface': 7.697666601543915e-06, 'learning_rate/action_model': 7.697666601543915e-05, 'epoch': 1.07}) +[2026-06-17 20:29:59,131][starVLA.training.train_starvla][INFO] - Step 2362, Loss: {'train/loss': 0.4442175626754761, 'train/grad_norm_pre_clip': 13.133721351623535, 'timing/data': 0.0006500948220491409, 'timing/model': 0.9506146386265755, 'learning_rate/qwen_vl_interface': 7.695812898430438e-06, 'learning_rate/action_model': 7.695812898430437e-05, 'epoch': 1.07}) +[2026-06-17 20:30:02,981][starVLA.training.train_starvla][INFO] - Step 2363, Loss: {'train/loss': 0.4976135790348053, 'train/grad_norm_pre_clip': 11.504074096679688, 'timing/data': 0.000510229729115963, 'timing/model': 1.1136132264509797, 'learning_rate/qwen_vl_interface': 7.693958688297316e-06, 'learning_rate/action_model': 7.693958688297314e-05, 'epoch': 1.07}) +[2026-06-17 20:30:06,797][starVLA.training.train_starvla][INFO] - Step 2364, Loss: {'train/loss': 0.38659995794296265, 'train/grad_norm_pre_clip': 12.98953914642334, 'timing/data': 0.0014770515263080597, 'timing/model': 1.1188313690945506, 'learning_rate/qwen_vl_interface': 7.692103971528925e-06, 'learning_rate/action_model': 7.692103971528926e-05, 'epoch': 1.07}) +[2026-06-17 20:30:10,463][starVLA.training.train_starvla][INFO] - Step 2365, Loss: {'train/loss': 0.5692240595817566, 'train/grad_norm_pre_clip': 12.328841209411621, 'timing/data': 0.0005496013909578323, 'timing/model': 0.9465459659695625, 'learning_rate/qwen_vl_interface': 7.690248748509757e-06, 'learning_rate/action_model': 7.690248748509757e-05, 'epoch': 1.07}) +[2026-06-17 20:30:14,371][starVLA.training.train_starvla][INFO] - Step 2366, Loss: {'train/loss': 0.3878708481788635, 'train/grad_norm_pre_clip': 12.286430358886719, 'timing/data': 0.00048512034118175507, 'timing/model': 0.9350104471668601, 'learning_rate/qwen_vl_interface': 7.688393019624396e-06, 'learning_rate/action_model': 7.688393019624395e-05, 'epoch': 1.07}) +[2026-06-17 20:30:18,216][starVLA.training.train_starvla][INFO] - Step 2367, Loss: {'train/loss': 0.45632293820381165, 'train/grad_norm_pre_clip': 14.584397315979004, 'timing/data': 0.0005030399188399315, 'timing/model': 0.9446722371503711, 'learning_rate/qwen_vl_interface': 7.686536785257539e-06, 'learning_rate/action_model': 7.686536785257538e-05, 'epoch': 1.07}) +[2026-06-17 20:30:22,074][starVLA.training.train_starvla][INFO] - Step 2368, Loss: {'train/loss': 0.510735273361206, 'train/grad_norm_pre_clip': 14.460530281066895, 'timing/data': 0.0005485741421580315, 'timing/model': 0.9275081763044, 'learning_rate/qwen_vl_interface': 7.684680045793982e-06, 'learning_rate/action_model': 7.684680045793983e-05, 'epoch': 1.07}) +[2026-06-17 20:30:25,866][starVLA.training.train_starvla][INFO] - Step 2369, Loss: {'train/loss': 0.47219356894493103, 'train/grad_norm_pre_clip': 10.396833419799805, 'timing/data': 0.0004423782229423523, 'timing/model': 0.9300508005544543, 'learning_rate/qwen_vl_interface': 7.682822801618634e-06, 'learning_rate/action_model': 7.682822801618633e-05, 'epoch': 1.07}) +[2026-06-17 20:30:29,875][starVLA.training.train_starvla][INFO] - Step 2370, Loss: {'train/loss': 0.6135812997817993, 'train/grad_norm_pre_clip': 17.7657413482666, 'timing/data': 0.0007925024256110191, 'timing/model': 1.1823103921487927, 'learning_rate/qwen_vl_interface': 7.680965053116498e-06, 'learning_rate/action_model': 7.680965053116499e-05, 'epoch': 1.08}) +[2026-06-17 20:30:33,561][starVLA.training.train_starvla][INFO] - Step 2371, Loss: {'train/loss': 0.32796013355255127, 'train/grad_norm_pre_clip': 9.63411808013916, 'timing/data': 0.0005069458857178688, 'timing/model': 0.9251909889280796, 'learning_rate/qwen_vl_interface': 7.679106800672695e-06, 'learning_rate/action_model': 7.679106800672695e-05, 'epoch': 1.08}) +[2026-06-17 20:30:37,348][starVLA.training.train_starvla][INFO] - Step 2372, Loss: {'train/loss': 0.3467661738395691, 'train/grad_norm_pre_clip': 12.949210166931152, 'timing/data': 0.0009964238852262497, 'timing/model': 0.9424696704372764, 'learning_rate/qwen_vl_interface': 7.677248044672438e-06, 'learning_rate/action_model': 7.677248044672437e-05, 'epoch': 1.08}) +[2026-06-17 20:30:41,120][starVLA.training.train_starvla][INFO] - Step 2373, Loss: {'train/loss': 0.34066668152809143, 'train/grad_norm_pre_clip': 12.851435661315918, 'timing/data': 0.0005522631108760834, 'timing/model': 0.9328616065904498, 'learning_rate/qwen_vl_interface': 7.675388785501048e-06, 'learning_rate/action_model': 7.675388785501047e-05, 'epoch': 1.08}) +[2026-06-17 20:30:45,038][starVLA.training.train_starvla][INFO] - Step 2374, Loss: {'train/loss': 0.2978058159351349, 'train/grad_norm_pre_clip': 9.93416976928711, 'timing/data': 0.0008963868021965027, 'timing/model': 0.9477693708613515, 'learning_rate/qwen_vl_interface': 7.673529023543951e-06, 'learning_rate/action_model': 7.673529023543951e-05, 'epoch': 1.08}) +[2026-06-17 20:30:49,507][starVLA.training.train_starvla][INFO] - Step 2375, Loss: {'train/loss': 0.31083357334136963, 'train/grad_norm_pre_clip': 10.969466209411621, 'timing/data': 0.0012745363637804985, 'timing/model': 0.9530373746529222, 'learning_rate/qwen_vl_interface': 7.671668759186683e-06, 'learning_rate/action_model': 7.671668759186683e-05, 'epoch': 1.08}) +[2026-06-17 20:30:53,310][starVLA.training.train_starvla][INFO] - Step 2376, Loss: {'train/loss': 0.5418122410774231, 'train/grad_norm_pre_clip': 13.491387367248535, 'timing/data': 0.0009608957916498184, 'timing/model': 1.1029186183586717, 'learning_rate/qwen_vl_interface': 7.669807992814875e-06, 'learning_rate/action_model': 7.669807992814875e-05, 'epoch': 1.08}) +[2026-06-17 20:30:57,115][starVLA.training.train_starvla][INFO] - Step 2377, Loss: {'train/loss': 0.42783355712890625, 'train/grad_norm_pre_clip': 11.446526527404785, 'timing/data': 0.0006290245801210403, 'timing/model': 1.118192440830171, 'learning_rate/qwen_vl_interface': 7.667946724814264e-06, 'learning_rate/action_model': 7.667946724814265e-05, 'epoch': 1.08}) +[2026-06-17 20:31:00,734][starVLA.training.train_starvla][INFO] - Step 2378, Loss: {'train/loss': 0.3193848431110382, 'train/grad_norm_pre_clip': 9.968757629394531, 'timing/data': 0.0008218567818403244, 'timing/model': 0.9375605676323175, 'learning_rate/qwen_vl_interface': 7.666084955570698e-06, 'learning_rate/action_model': 7.666084955570698e-05, 'epoch': 1.08}) +[2026-06-17 20:31:04,530][starVLA.training.train_starvla][INFO] - Step 2379, Loss: {'train/loss': 0.4073698818683624, 'train/grad_norm_pre_clip': 12.116082191467285, 'timing/data': 0.0007857140153646469, 'timing/model': 0.9458169722929597, 'learning_rate/qwen_vl_interface': 7.664222685470119e-06, 'learning_rate/action_model': 7.664222685470118e-05, 'epoch': 1.08}) +[2026-06-17 20:31:08,347][starVLA.training.train_starvla][INFO] - Step 2380, Loss: {'train/loss': 0.27422693371772766, 'train/grad_norm_pre_clip': 9.711002349853516, 'timing/data': 0.0012264205142855644, 'timing/model': 0.9321237979456782, 'learning_rate/qwen_vl_interface': 7.66235991489858e-06, 'learning_rate/action_model': 7.66235991489858e-05, 'epoch': 1.08}) +[2026-06-17 20:31:12,176][starVLA.training.train_starvla][INFO] - Step 2381, Loss: {'train/loss': 0.4505669176578522, 'train/grad_norm_pre_clip': 13.87105655670166, 'timing/data': 0.0010349676012992859, 'timing/model': 0.9385946681722999, 'learning_rate/qwen_vl_interface': 7.660496644242233e-06, 'learning_rate/action_model': 7.660496644242233e-05, 'epoch': 1.08}) +[2026-06-17 20:31:16,024][starVLA.training.train_starvla][INFO] - Step 2382, Loss: {'train/loss': 0.4985220730304718, 'train/grad_norm_pre_clip': 12.512679100036621, 'timing/data': 0.0005377763882279396, 'timing/model': 0.9386490695178509, 'learning_rate/qwen_vl_interface': 7.658632873887342e-06, 'learning_rate/action_model': 7.658632873887342e-05, 'epoch': 1.08}) +[2026-06-17 20:31:20,138][starVLA.training.train_starvla][INFO] - Step 2383, Loss: {'train/loss': 0.47316160798072815, 'train/grad_norm_pre_clip': 12.283123016357422, 'timing/data': 0.0008961586281657219, 'timing/model': 1.4191426476463675, 'learning_rate/qwen_vl_interface': 7.656768604220262e-06, 'learning_rate/action_model': 7.656768604220261e-05, 'epoch': 1.08}) +[2026-06-17 20:31:24,124][starVLA.training.train_starvla][INFO] - Step 2384, Loss: {'train/loss': 0.5588129758834839, 'train/grad_norm_pre_clip': 18.791866302490234, 'timing/data': 0.0005899639800190926, 'timing/model': 1.1729707466438413, 'learning_rate/qwen_vl_interface': 7.654903835627459e-06, 'learning_rate/action_model': 7.654903835627458e-05, 'epoch': 1.08}) +[2026-06-17 20:31:27,796][starVLA.training.train_starvla][INFO] - Step 2385, Loss: {'train/loss': 0.3109270930290222, 'train/grad_norm_pre_clip': 10.751753807067871, 'timing/data': 0.0004992373287677765, 'timing/model': 0.9437878224998713, 'learning_rate/qwen_vl_interface': 7.653038568495505e-06, 'learning_rate/action_model': 7.653038568495506e-05, 'epoch': 1.08}) +[2026-06-17 20:31:31,656][starVLA.training.train_starvla][INFO] - Step 2386, Loss: {'train/loss': 0.19665610790252686, 'train/grad_norm_pre_clip': 9.089364051818848, 'timing/data': 0.000496232882142067, 'timing/model': 0.9395610634237528, 'learning_rate/qwen_vl_interface': 7.651172803211068e-06, 'learning_rate/action_model': 7.651172803211068e-05, 'epoch': 1.08}) +[2026-06-17 20:31:35,516][starVLA.training.train_starvla][INFO] - Step 2387, Loss: {'train/loss': 0.4640781581401825, 'train/grad_norm_pre_clip': 11.040254592895508, 'timing/data': 0.0008485214784741402, 'timing/model': 0.9508673176169395, 'learning_rate/qwen_vl_interface': 7.649306540160928e-06, 'learning_rate/action_model': 7.649306540160927e-05, 'epoch': 1.08}) +[2026-06-17 20:31:40,007][starVLA.training.train_starvla][INFO] - Step 2388, Loss: {'train/loss': 0.27620452642440796, 'train/grad_norm_pre_clip': 11.933170318603516, 'timing/data': 0.0006431955844163895, 'timing/model': 0.9539078017696738, 'learning_rate/qwen_vl_interface': 7.647439779731959e-06, 'learning_rate/action_model': 7.647439779731958e-05, 'epoch': 1.08}) +[2026-06-17 20:31:43,882][starVLA.training.train_starvla][INFO] - Step 2389, Loss: {'train/loss': 0.47221532464027405, 'train/grad_norm_pre_clip': 14.81876277923584, 'timing/data': 0.0005331914871931076, 'timing/model': 0.9350929390639067, 'learning_rate/qwen_vl_interface': 7.645572522311144e-06, 'learning_rate/action_model': 7.645572522311144e-05, 'epoch': 1.08}) +[2026-06-17 20:31:47,791][starVLA.training.train_starvla][INFO] - Step 2390, Loss: {'train/loss': 0.4446706175804138, 'train/grad_norm_pre_clip': 15.06656551361084, 'timing/data': 0.000834951177239418, 'timing/model': 1.1235036039724946, 'learning_rate/qwen_vl_interface': 7.643704768285568e-06, 'learning_rate/action_model': 7.643704768285568e-05, 'epoch': 1.08}) +[2026-06-17 20:31:51,474][starVLA.training.train_starvla][INFO] - Step 2391, Loss: {'train/loss': 0.39049962162971497, 'train/grad_norm_pre_clip': 11.568358421325684, 'timing/data': 0.0008327765390276909, 'timing/model': 0.9859311217442155, 'learning_rate/qwen_vl_interface': 7.641836518042417e-06, 'learning_rate/action_model': 7.641836518042416e-05, 'epoch': 1.08}) +[2026-06-17 20:31:55,337][starVLA.training.train_starvla][INFO] - Step 2392, Loss: {'train/loss': 0.2649889588356018, 'train/grad_norm_pre_clip': 10.293784141540527, 'timing/data': 0.0008862307295203209, 'timing/model': 0.9455105336382985, 'learning_rate/qwen_vl_interface': 7.639967771968983e-06, 'learning_rate/action_model': 7.639967771968983e-05, 'epoch': 1.09}) +[2026-06-17 20:31:59,155][starVLA.training.train_starvla][INFO] - Step 2393, Loss: {'train/loss': 0.3716652989387512, 'train/grad_norm_pre_clip': 19.714866638183594, 'timing/data': 0.0008535357192158699, 'timing/model': 0.9689870290458202, 'learning_rate/qwen_vl_interface': 7.638098530452657e-06, 'learning_rate/action_model': 7.638098530452658e-05, 'epoch': 1.09}) +[2026-06-17 20:32:02,982][starVLA.training.train_starvla][INFO] - Step 2394, Loss: {'train/loss': 0.26728230714797974, 'train/grad_norm_pre_clip': 10.703530311584473, 'timing/data': 0.0007085930556058884, 'timing/model': 0.9505309658125043, 'learning_rate/qwen_vl_interface': 7.636228793880935e-06, 'learning_rate/action_model': 7.636228793880935e-05, 'epoch': 1.09}) +[2026-06-17 20:32:06,900][starVLA.training.train_starvla][INFO] - Step 2395, Loss: {'train/loss': 0.5418453812599182, 'train/grad_norm_pre_clip': 47.04216384887695, 'timing/data': 0.0008068298920989037, 'timing/model': 1.0253272522240877, 'learning_rate/qwen_vl_interface': 7.63435856264142e-06, 'learning_rate/action_model': 7.63435856264142e-05, 'epoch': 1.09}) +[2026-06-17 20:32:10,801][starVLA.training.train_starvla][INFO] - Step 2396, Loss: {'train/loss': 0.2989482581615448, 'train/grad_norm_pre_clip': 14.431708335876465, 'timing/data': 0.0007029445841908455, 'timing/model': 1.1721027018502355, 'learning_rate/qwen_vl_interface': 7.632487837121806e-06, 'learning_rate/action_model': 7.632487837121806e-05, 'epoch': 1.09}) +[2026-06-17 20:32:14,602][starVLA.training.train_starvla][INFO] - Step 2397, Loss: {'train/loss': 0.5276523232460022, 'train/grad_norm_pre_clip': 13.482951164245605, 'timing/data': 0.0005960604175925255, 'timing/model': 1.092697098851204, 'learning_rate/qwen_vl_interface': 7.630616617709902e-06, 'learning_rate/action_model': 7.630616617709902e-05, 'epoch': 1.09}) +[2026-06-17 20:32:18,243][starVLA.training.train_starvla][INFO] - Step 2398, Loss: {'train/loss': 0.2837650179862976, 'train/grad_norm_pre_clip': 11.342513084411621, 'timing/data': 0.0009612059220671654, 'timing/model': 0.9470678539946675, 'learning_rate/qwen_vl_interface': 7.6287449047936095e-06, 'learning_rate/action_model': 7.628744904793609e-05, 'epoch': 1.09}) +[2026-06-17 20:32:22,043][starVLA.training.train_starvla][INFO] - Step 2399, Loss: {'train/loss': 0.5841007232666016, 'train/grad_norm_pre_clip': 15.199706077575684, 'timing/data': 0.0008948575705289841, 'timing/model': 0.9342360896989703, 'learning_rate/qwen_vl_interface': 7.6268726987609384e-06, 'learning_rate/action_model': 7.626872698760939e-05, 'epoch': 1.09}) +[2026-06-17 20:34:28,662][starVLA.training.train_starvla][INFO] - Step 2400, Loss: {'train/loss': 0.38219091296195984, 'train/grad_norm_pre_clip': 13.642955780029297, 'eval/loss': 0.4106820225715637, 'eval/latency_8/loss': 0.4106820225715637, 'eval/demon_fire_precision': 0.9392115558230515, 'eval/demon_fire_recall': 0.969555762659211, 'eval/demon_fire_f1': 0.9541424640782636, 'eval/demon_fire_rate_model': 0.924596549805231, 'eval/demon_fire_rate_teacher': 0.8956594323873122, 'eval/demon_left_precision': 0.7407407407407407, 'eval/demon_left_recall': 0.6852248394004282, 'eval/demon_left_f1': 0.7119021134593994, 'eval/demon_left_rate_model': 0.12020033388981637, 'eval/demon_left_rate_teacher': 0.12993878686700056, 'eval/demon_right_precision': 0.7728, 'eval/demon_right_recall': 0.7318181818181818, 'eval/demon_right_f1': 0.7517509727626459, 'eval/demon_right_rate_model': 0.17390094602114636, 'eval/demon_right_rate_teacher': 0.18363939899833054, 'eval/demon_move_f1': 0.7318265431110227, 'eval/demon_cc_f1': 0.8744497662823252, 'eval/latency_8/demon_fire_precision': 0.9392115558230515, 'eval/latency_8/demon_fire_recall': 0.969555762659211, 'eval/latency_8/demon_fire_f1': 0.9541424640782636, 'eval/latency_8/demon_fire_rate_model': 0.924596549805231, 'eval/latency_8/demon_fire_rate_teacher': 0.8956594323873122, 'eval/latency_8/demon_left_precision': 0.7407407407407407, 'eval/latency_8/demon_left_recall': 0.6852248394004282, 'eval/latency_8/demon_left_f1': 0.7119021134593994, 'eval/latency_8/demon_left_rate_model': 0.12020033388981637, 'eval/latency_8/demon_left_rate_teacher': 0.12993878686700056, 'eval/latency_8/demon_right_precision': 0.7728, 'eval/latency_8/demon_right_recall': 0.7318181818181818, 'eval/latency_8/demon_right_f1': 0.7517509727626459, 'eval/latency_8/demon_right_rate_model': 0.17390094602114636, 'eval/latency_8/demon_right_rate_teacher': 0.18363939899833054, 'eval/latency_8/demon_move_f1': 0.7318265431110227, 'eval/latency_8/demon_cc_f1': 0.8744497662823252, 'timing/data': 0.001745765097439289, 'timing/model': 0.9353650873526931, 'learning_rate/qwen_vl_interface': 7.625e-06, 'learning_rate/action_model': 7.625e-05, 'epoch': 1.09}) +[2026-06-17 20:34:28,665][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_2400_state +[2026-06-17 20:34:28,670][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 20:34:35,416][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_2400_state/model.safetensors +[2026-06-17 20:34:51,498][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_2400_state/optimizer.bin +[2026-06-17 20:34:51,499][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_2400_state/sampler.bin +[2026-06-17 20:34:51,499][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_2400_state/sampler_1.bin +[2026-06-17 20:34:51,503][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_2400_state/random_states_0.pkl +[2026-06-17 20:34:51,508][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2400 +[2026-06-17 20:34:51,510][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 20:34:51,521][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 20:34:55,434][starVLA.training.train_starvla][INFO] - Step 2401, Loss: {'train/loss': 0.36844950914382935, 'train/grad_norm_pre_clip': 11.696499824523926, 'timing/data': 0.0006676120683550835, 'timing/model': 0.952899201773107, 'learning_rate/qwen_vl_interface': 7.6231268088990075e-06, 'learning_rate/action_model': 7.623126808899008e-05, 'epoch': 1.09}) +[2026-06-17 20:34:59,270][starVLA.training.train_starvla][INFO] - Step 2402, Loss: {'train/loss': 0.4160645604133606, 'train/grad_norm_pre_clip': 11.641776084899902, 'timing/data': 0.0005522333085536957, 'timing/model': 0.9330520881339908, 'learning_rate/qwen_vl_interface': 7.621253125846273e-06, 'learning_rate/action_model': 7.621253125846273e-05, 'epoch': 1.09}) +[2026-06-17 20:35:03,109][starVLA.training.train_starvla][INFO] - Step 2403, Loss: {'train/loss': 0.4885656535625458, 'train/grad_norm_pre_clip': 17.5037899017334, 'timing/data': 0.0005333013832569122, 'timing/model': 0.9476811569184065, 'learning_rate/qwen_vl_interface': 7.619378951230213e-06, 'learning_rate/action_model': 7.619378951230213e-05, 'epoch': 1.09}) +[2026-06-17 20:35:06,905][starVLA.training.train_starvla][INFO] - Step 2404, Loss: {'train/loss': 0.5921487808227539, 'train/grad_norm_pre_clip': 15.322441101074219, 'timing/data': 0.0007137646898627281, 'timing/model': 0.9453384885564446, 'learning_rate/qwen_vl_interface': 7.617504285439346e-06, 'learning_rate/action_model': 7.617504285439346e-05, 'epoch': 1.09}) +[2026-06-17 20:35:10,745][starVLA.training.train_starvla][INFO] - Step 2405, Loss: {'train/loss': 0.3317998945713043, 'train/grad_norm_pre_clip': 13.23872184753418, 'timing/data': 0.0005748560652136803, 'timing/model': 0.967514518648386, 'learning_rate/qwen_vl_interface': 7.615629128862295e-06, 'learning_rate/action_model': 7.615629128862295e-05, 'epoch': 1.09}) +[2026-06-17 20:35:14,577][starVLA.training.train_starvla][INFO] - Step 2406, Loss: {'train/loss': 0.60758376121521, 'train/grad_norm_pre_clip': 17.595558166503906, 'timing/data': 0.0004955986514687538, 'timing/model': 1.1048757815733552, 'learning_rate/qwen_vl_interface': 7.61375348188778e-06, 'learning_rate/action_model': 7.61375348188778e-05, 'epoch': 1.09}) +[2026-06-17 20:35:18,348][starVLA.training.train_starvla][INFO] - Step 2407, Loss: {'train/loss': 0.38845348358154297, 'train/grad_norm_pre_clip': 14.007159233093262, 'timing/data': 0.00048484280705451965, 'timing/model': 0.9357523880898952, 'learning_rate/qwen_vl_interface': 7.611877344904624e-06, 'learning_rate/action_model': 7.611877344904624e-05, 'epoch': 1.09}) +[2026-06-17 20:35:22,163][starVLA.training.train_starvla][INFO] - Step 2408, Loss: {'train/loss': 0.49370458722114563, 'train/grad_norm_pre_clip': 16.973039627075195, 'timing/data': 0.0013153031468391418, 'timing/model': 0.976146743632853, 'learning_rate/qwen_vl_interface': 7.610000718301753e-06, 'learning_rate/action_model': 7.610000718301753e-05, 'epoch': 1.09}) +[2026-06-17 20:35:25,953][starVLA.training.train_starvla][INFO] - Step 2409, Loss: {'train/loss': 0.5176064968109131, 'train/grad_norm_pre_clip': 15.689006805419922, 'timing/data': 0.0006024027243256569, 'timing/model': 0.9656112343072891, 'learning_rate/qwen_vl_interface': 7.608123602468195e-06, 'learning_rate/action_model': 7.608123602468194e-05, 'epoch': 1.09}) +[2026-06-17 20:35:30,020][starVLA.training.train_starvla][INFO] - Step 2410, Loss: {'train/loss': 0.5088638663291931, 'train/grad_norm_pre_clip': 18.639835357666016, 'timing/data': 0.000918341800570488, 'timing/model': 0.9796174997463822, 'learning_rate/qwen_vl_interface': 7.606245997793074e-06, 'learning_rate/action_model': 7.606245997793075e-05, 'epoch': 1.09}) +[2026-06-17 20:35:33,844][starVLA.training.train_starvla][INFO] - Step 2411, Loss: {'train/loss': 0.5310910940170288, 'train/grad_norm_pre_clip': 13.53986644744873, 'timing/data': 0.0008916091173887253, 'timing/model': 0.935647938400507, 'learning_rate/qwen_vl_interface': 7.604367904665625e-06, 'learning_rate/action_model': 7.604367904665624e-05, 'epoch': 1.09}) +[2026-06-17 20:35:37,834][starVLA.training.train_starvla][INFO] - Step 2412, Loss: {'train/loss': 0.4607195258140564, 'train/grad_norm_pre_clip': 20.264755249023438, 'timing/data': 0.0005421917885541916, 'timing/model': 1.127950606867671, 'learning_rate/qwen_vl_interface': 7.602489323475174e-06, 'learning_rate/action_model': 7.602489323475174e-05, 'epoch': 1.09}) +[2026-06-17 20:35:41,838][starVLA.training.train_starvla][INFO] - Step 2413, Loss: {'train/loss': 0.42497074604034424, 'train/grad_norm_pre_clip': 16.993772506713867, 'timing/data': 0.0005196500569581985, 'timing/model': 1.1037608180195093, 'learning_rate/qwen_vl_interface': 7.600610254611155e-06, 'learning_rate/action_model': 7.600610254611155e-05, 'epoch': 1.09}) +[2026-06-17 20:35:45,606][starVLA.training.train_starvla][INFO] - Step 2414, Loss: {'train/loss': 0.5767402052879333, 'train/grad_norm_pre_clip': 16.4826602935791, 'timing/data': 0.0005139308050274849, 'timing/model': 0.9666824657469988, 'learning_rate/qwen_vl_interface': 7.5987306984631025e-06, 'learning_rate/action_model': 7.598730698463102e-05, 'epoch': 1.1}) +[2026-06-17 20:35:49,625][starVLA.training.train_starvla][INFO] - Step 2415, Loss: {'train/loss': 0.38395246863365173, 'train/grad_norm_pre_clip': 13.832996368408203, 'timing/data': 0.0005766265094280243, 'timing/model': 1.0850477600470185, 'learning_rate/qwen_vl_interface': 7.596850655420647e-06, 'learning_rate/action_model': 7.596850655420646e-05, 'epoch': 1.1}) +[2026-06-17 20:35:54,105][starVLA.training.train_starvla][INFO] - Step 2416, Loss: {'train/loss': 0.31856465339660645, 'train/grad_norm_pre_clip': 10.789921760559082, 'timing/data': 0.0010660719126462936, 'timing/model': 1.5168288461863995, 'learning_rate/qwen_vl_interface': 7.5949701258735265e-06, 'learning_rate/action_model': 7.594970125873526e-05, 'epoch': 1.1}) +[2026-06-17 20:35:57,935][starVLA.training.train_starvla][INFO] - Step 2417, Loss: {'train/loss': 0.405070960521698, 'train/grad_norm_pre_clip': 14.007759094238281, 'timing/data': 0.0009622750803828239, 'timing/model': 0.9442985123023391, 'learning_rate/qwen_vl_interface': 7.5930891102115735e-06, 'learning_rate/action_model': 7.593089110211573e-05, 'epoch': 1.1}) +[2026-06-17 20:36:01,745][starVLA.training.train_starvla][INFO] - Step 2418, Loss: {'train/loss': 0.3636796474456787, 'train/grad_norm_pre_clip': 13.233283042907715, 'timing/data': 0.0006903205066919327, 'timing/model': 0.9444048330187798, 'learning_rate/qwen_vl_interface': 7.59120760882473e-06, 'learning_rate/action_model': 7.59120760882473e-05, 'epoch': 1.1}) +[2026-06-17 20:36:05,663][starVLA.training.train_starvla][INFO] - Step 2419, Loss: {'train/loss': 0.6210758686065674, 'train/grad_norm_pre_clip': 18.412282943725586, 'timing/data': 0.0009179906919598579, 'timing/model': 1.1419816194102168, 'learning_rate/qwen_vl_interface': 7.5893256221030275e-06, 'learning_rate/action_model': 7.589325622103027e-05, 'epoch': 1.1}) +[2026-06-17 20:36:09,565][starVLA.training.train_starvla][INFO] - Step 2420, Loss: {'train/loss': 0.4545076787471771, 'train/grad_norm_pre_clip': 18.38399314880371, 'timing/data': 0.000859774649143219, 'timing/model': 1.1707767909392715, 'learning_rate/qwen_vl_interface': 7.587443150436608e-06, 'learning_rate/action_model': 7.587443150436608e-05, 'epoch': 1.1}) +[2026-06-17 20:36:13,278][starVLA.training.train_starvla][INFO] - Step 2421, Loss: {'train/loss': 0.5869956016540527, 'train/grad_norm_pre_clip': 16.423294067382812, 'timing/data': 0.0005076108500361443, 'timing/model': 0.9468462448567152, 'learning_rate/qwen_vl_interface': 7.585560194215706e-06, 'learning_rate/action_model': 7.585560194215706e-05, 'epoch': 1.1}) +[2026-06-17 20:36:17,120][starVLA.training.train_starvla][INFO] - Step 2422, Loss: {'train/loss': 0.362408310174942, 'train/grad_norm_pre_clip': 10.567556381225586, 'timing/data': 0.0007027136161923409, 'timing/model': 0.9395017018541694, 'learning_rate/qwen_vl_interface': 7.583676753830665e-06, 'learning_rate/action_model': 7.583676753830664e-05, 'epoch': 1.1}) +[2026-06-17 20:36:20,941][starVLA.training.train_starvla][INFO] - Step 2423, Loss: {'train/loss': 0.8479802012443542, 'train/grad_norm_pre_clip': 21.229022979736328, 'timing/data': 0.0009252429008483887, 'timing/model': 0.9393812958151102, 'learning_rate/qwen_vl_interface': 7.58179282967192e-06, 'learning_rate/action_model': 7.58179282967192e-05, 'epoch': 1.1}) +[2026-06-17 20:36:24,738][starVLA.training.train_starvla][INFO] - Step 2424, Loss: {'train/loss': 0.7346493005752563, 'train/grad_norm_pre_clip': 19.909048080444336, 'timing/data': 0.000848076306283474, 'timing/model': 0.9282269077375531, 'learning_rate/qwen_vl_interface': 7.579908422130013e-06, 'learning_rate/action_model': 7.579908422130014e-05, 'epoch': 1.1}) +[2026-06-17 20:36:28,539][starVLA.training.train_starvla][INFO] - Step 2425, Loss: {'train/loss': 0.5798515677452087, 'train/grad_norm_pre_clip': 14.416481971740723, 'timing/data': 0.0013151783496141434, 'timing/model': 0.9313626969233155, 'learning_rate/qwen_vl_interface': 7.578023531595583e-06, 'learning_rate/action_model': 7.578023531595583e-05, 'epoch': 1.1}) +[2026-06-17 20:36:32,325][starVLA.training.train_starvla][INFO] - Step 2426, Loss: {'train/loss': 0.5424919724464417, 'train/grad_norm_pre_clip': 18.119155883789062, 'timing/data': 0.0005396567285060883, 'timing/model': 1.09238117095083, 'learning_rate/qwen_vl_interface': 7.576138158459369e-06, 'learning_rate/action_model': 7.576138158459368e-05, 'epoch': 1.1}) +[2026-06-17 20:36:36,016][starVLA.training.train_starvla][INFO] - Step 2427, Loss: {'train/loss': 0.6071339249610901, 'train/grad_norm_pre_clip': 18.135494232177734, 'timing/data': 0.0008685989305377007, 'timing/model': 1.000328803434968, 'learning_rate/qwen_vl_interface': 7.574252303112213e-06, 'learning_rate/action_model': 7.574252303112212e-05, 'epoch': 1.1}) +[2026-06-17 20:36:40,071][starVLA.training.train_starvla][INFO] - Step 2428, Loss: {'train/loss': 0.5672111511230469, 'train/grad_norm_pre_clip': 18.258432388305664, 'timing/data': 0.0005364082753658295, 'timing/model': 1.1355189122259617, 'learning_rate/qwen_vl_interface': 7.572365965945053e-06, 'learning_rate/action_model': 7.572365965945052e-05, 'epoch': 1.1}) +[2026-06-17 20:36:44,144][starVLA.training.train_starvla][INFO] - Step 2429, Loss: {'train/loss': 0.3425939083099365, 'train/grad_norm_pre_clip': 12.261396408081055, 'timing/data': 0.0005663912743330002, 'timing/model': 0.9434906672686338, 'learning_rate/qwen_vl_interface': 7.570479147348929e-06, 'learning_rate/action_model': 7.570479147348929e-05, 'epoch': 1.1}) +[2026-06-17 20:36:48,874][starVLA.training.train_starvla][INFO] - Step 2430, Loss: {'train/loss': 0.46053197979927063, 'train/grad_norm_pre_clip': 16.850210189819336, 'timing/data': 0.00047612283378839493, 'timing/model': 0.9323558611795306, 'learning_rate/qwen_vl_interface': 7.568591847714983e-06, 'learning_rate/action_model': 7.568591847714982e-05, 'epoch': 1.1}) +[2026-06-17 20:36:53,667][starVLA.training.train_starvla][INFO] - Step 2431, Loss: {'train/loss': 0.28696340322494507, 'train/grad_norm_pre_clip': 13.068934440612793, 'timing/data': 0.0006490349769592285, 'timing/model': 0.9387624468654394, 'learning_rate/qwen_vl_interface': 7.5667040674344495e-06, 'learning_rate/action_model': 7.566704067434449e-05, 'epoch': 1.1}) +[2026-06-17 20:36:57,675][starVLA.training.train_starvla][INFO] - Step 2432, Loss: {'train/loss': 0.2691831588745117, 'train/grad_norm_pre_clip': 11.781353950500488, 'timing/data': 0.0005861548706889153, 'timing/model': 1.2227192586287856, 'learning_rate/qwen_vl_interface': 7.564815806898669e-06, 'learning_rate/action_model': 7.564815806898668e-05, 'epoch': 1.1}) +[2026-06-17 20:37:01,619][starVLA.training.train_starvla][INFO] - Step 2433, Loss: {'train/loss': 0.31825488805770874, 'train/grad_norm_pre_clip': 12.186370849609375, 'timing/data': 0.0005437033250927925, 'timing/model': 1.176668232306838, 'learning_rate/qwen_vl_interface': 7.562927066499084e-06, 'learning_rate/action_model': 7.562927066499083e-05, 'epoch': 1.1}) +[2026-06-17 20:37:05,268][starVLA.training.train_starvla][INFO] - Step 2434, Loss: {'train/loss': 0.3509637713432312, 'train/grad_norm_pre_clip': 11.493006706237793, 'timing/data': 0.0005188649520277977, 'timing/model': 0.9322436330839992, 'learning_rate/qwen_vl_interface': 7.5610378466272274e-06, 'learning_rate/action_model': 7.561037846627227e-05, 'epoch': 1.1}) +[2026-06-17 20:37:09,331][starVLA.training.train_starvla][INFO] - Step 2435, Loss: {'train/loss': 0.9774880409240723, 'train/grad_norm_pre_clip': 21.30556297302246, 'timing/data': 0.0008993195369839668, 'timing/model': 1.0512542203068733, 'learning_rate/qwen_vl_interface': 7.5591481476747384e-06, 'learning_rate/action_model': 7.559148147674739e-05, 'epoch': 1.1}) +[2026-06-17 20:37:13,195][starVLA.training.train_starvla][INFO] - Step 2436, Loss: {'train/loss': 0.2299349457025528, 'train/grad_norm_pre_clip': 9.80163860321045, 'timing/data': 0.0005364855751395226, 'timing/model': 0.9505673376843333, 'learning_rate/qwen_vl_interface': 7.557257970033355e-06, 'learning_rate/action_model': 7.557257970033354e-05, 'epoch': 1.11}) +[2026-06-17 20:37:17,038][starVLA.training.train_starvla][INFO] - Step 2437, Loss: {'train/loss': 0.6229321360588074, 'train/grad_norm_pre_clip': 13.955917358398438, 'timing/data': 0.0005570519715547562, 'timing/model': 0.9518096158280969, 'learning_rate/qwen_vl_interface': 7.55536731409491e-06, 'learning_rate/action_model': 7.55536731409491e-05, 'epoch': 1.11}) +[2026-06-17 20:37:20,943][starVLA.training.train_starvla][INFO] - Step 2438, Loss: {'train/loss': 0.3671034276485443, 'train/grad_norm_pre_clip': 10.900961875915527, 'timing/data': 0.0009965309873223305, 'timing/model': 0.9389238990843296, 'learning_rate/qwen_vl_interface': 7.553476180251341e-06, 'learning_rate/action_model': 7.55347618025134e-05, 'epoch': 1.11}) +[2026-06-17 20:37:24,961][starVLA.training.train_starvla][INFO] - Step 2439, Loss: {'train/loss': 0.30402401089668274, 'train/grad_norm_pre_clip': 11.182510375976562, 'timing/data': 0.0008141752332448959, 'timing/model': 1.1562842326238751, 'learning_rate/qwen_vl_interface': 7.551584568894681e-06, 'learning_rate/action_model': 7.551584568894681e-05, 'epoch': 1.11}) +[2026-06-17 20:37:28,814][starVLA.training.train_starvla][INFO] - Step 2440, Loss: {'train/loss': 0.3283235430717468, 'train/grad_norm_pre_clip': 14.523250579833984, 'timing/data': 0.0005386443808674812, 'timing/model': 1.129901167936623, 'learning_rate/qwen_vl_interface': 7.5496924804170635e-06, 'learning_rate/action_model': 7.549692480417063e-05, 'epoch': 1.11}) +[2026-06-17 20:37:32,507][starVLA.training.train_starvla][INFO] - Step 2441, Loss: {'train/loss': 0.5165645480155945, 'train/grad_norm_pre_clip': 13.091814994812012, 'timing/data': 0.000524972565472126, 'timing/model': 0.9598411461338401, 'learning_rate/qwen_vl_interface': 7.5477999152107195e-06, 'learning_rate/action_model': 7.54779991521072e-05, 'epoch': 1.11}) +[2026-06-17 20:37:36,348][starVLA.training.train_starvla][INFO] - Step 2442, Loss: {'train/loss': 0.5241848230361938, 'train/grad_norm_pre_clip': 19.440763473510742, 'timing/data': 0.0005521737039089203, 'timing/model': 0.9391308501362801, 'learning_rate/qwen_vl_interface': 7.545906873667982e-06, 'learning_rate/action_model': 7.545906873667981e-05, 'epoch': 1.11}) +[2026-06-17 20:37:40,280][starVLA.training.train_starvla][INFO] - Step 2443, Loss: {'train/loss': 0.3309686481952667, 'train/grad_norm_pre_clip': 10.943065643310547, 'timing/data': 0.0004928791895508766, 'timing/model': 0.9488350711762905, 'learning_rate/qwen_vl_interface': 7.544013356181277e-06, 'learning_rate/action_model': 7.544013356181277e-05, 'epoch': 1.11}) +[2026-06-17 20:37:45,548][starVLA.training.train_starvla][INFO] - Step 2444, Loss: {'train/loss': 0.466676265001297, 'train/grad_norm_pre_clip': 15.136763572692871, 'timing/data': 0.001359371468424797, 'timing/model': 0.9453127598389983, 'learning_rate/qwen_vl_interface': 7.542119363143135e-06, 'learning_rate/action_model': 7.542119363143136e-05, 'epoch': 1.11}) +[2026-06-17 20:37:51,426][starVLA.training.train_starvla][INFO] - Step 2445, Loss: {'train/loss': 0.45584970712661743, 'train/grad_norm_pre_clip': 14.895730972290039, 'timing/data': 0.0006439704447984695, 'timing/model': 0.9690189938992262, 'learning_rate/qwen_vl_interface': 7.540224894946184e-06, 'learning_rate/action_model': 7.540224894946184e-05, 'epoch': 1.11}) +[2026-06-17 20:37:55,309][starVLA.training.train_starvla][INFO] - Step 2446, Loss: {'train/loss': 0.4009994566440582, 'train/grad_norm_pre_clip': 28.867216110229492, 'timing/data': 0.00048437435179948807, 'timing/model': 1.1281803408637643, 'learning_rate/qwen_vl_interface': 7.538329951983148e-06, 'learning_rate/action_model': 7.538329951983147e-05, 'epoch': 1.11}) +[2026-06-17 20:37:59,532][starVLA.training.train_starvla][INFO] - Step 2447, Loss: {'train/loss': 0.33996322751045227, 'train/grad_norm_pre_clip': 12.132166862487793, 'timing/data': 0.0008391598239541054, 'timing/model': 1.5241035697981715, 'learning_rate/qwen_vl_interface': 7.536434534646851e-06, 'learning_rate/action_model': 7.53643453464685e-05, 'epoch': 1.11}) +[2026-06-17 20:38:03,402][starVLA.training.train_starvla][INFO] - Step 2448, Loss: {'train/loss': 0.6215811371803284, 'train/grad_norm_pre_clip': 21.60831069946289, 'timing/data': 0.0010219672694802284, 'timing/model': 0.9577197805047035, 'learning_rate/qwen_vl_interface': 7.534538643330213e-06, 'learning_rate/action_model': 7.534538643330214e-05, 'epoch': 1.11}) +[2026-06-17 20:38:07,266][starVLA.training.train_starvla][INFO] - Step 2449, Loss: {'train/loss': 0.4362499415874481, 'train/grad_norm_pre_clip': 17.655208587646484, 'timing/data': 0.0005357973277568817, 'timing/model': 0.9450528677552938, 'learning_rate/qwen_vl_interface': 7.5326422784262566e-06, 'learning_rate/action_model': 7.532642278426256e-05, 'epoch': 1.11}) +[2026-06-17 20:38:11,142][starVLA.training.train_starvla][INFO] - Step 2450, Loss: {'train/loss': 0.34264007210731506, 'train/grad_norm_pre_clip': 12.424002647399902, 'timing/data': 0.0018449053168296814, 'timing/model': 0.9545739330351353, 'learning_rate/qwen_vl_interface': 7.5307454403281e-06, 'learning_rate/action_model': 7.5307454403281e-05, 'epoch': 1.11}) +[2026-06-17 20:38:15,052][starVLA.training.train_starvla][INFO] - Step 2451, Loss: {'train/loss': 0.3885306715965271, 'train/grad_norm_pre_clip': 12.768654823303223, 'timing/data': 0.0005012881010770798, 'timing/model': 0.9238800415769219, 'learning_rate/qwen_vl_interface': 7.528848129428958e-06, 'learning_rate/action_model': 7.528848129428958e-05, 'epoch': 1.11}) +[2026-06-17 20:38:18,905][starVLA.training.train_starvla][INFO] - Step 2452, Loss: {'train/loss': 0.5870432257652283, 'train/grad_norm_pre_clip': 14.032419204711914, 'timing/data': 0.0006174910813570023, 'timing/model': 1.1251807250082493, 'learning_rate/qwen_vl_interface': 7.526950346122149e-06, 'learning_rate/action_model': 7.526950346122148e-05, 'epoch': 1.11}) +[2026-06-17 20:38:24,028][starVLA.training.train_starvla][INFO] - Step 2453, Loss: {'train/loss': 0.5017073750495911, 'train/grad_norm_pre_clip': 15.062843322753906, 'timing/data': 0.0005945218726992607, 'timing/model': 1.0931475572288036, 'learning_rate/qwen_vl_interface': 7.525052090801083e-06, 'learning_rate/action_model': 7.525052090801083e-05, 'epoch': 1.11}) +[2026-06-17 20:38:27,797][starVLA.training.train_starvla][INFO] - Step 2454, Loss: {'train/loss': 0.6039831638336182, 'train/grad_norm_pre_clip': 14.671672821044922, 'timing/data': 0.000898679718375206, 'timing/model': 0.9438954517245293, 'learning_rate/qwen_vl_interface': 7.523153363859269e-06, 'learning_rate/action_model': 7.523153363859269e-05, 'epoch': 1.11}) +[2026-06-17 20:38:31,713][starVLA.training.train_starvla][INFO] - Step 2455, Loss: {'train/loss': 0.16415464878082275, 'train/grad_norm_pre_clip': 7.211867332458496, 'timing/data': 0.0005443021655082703, 'timing/model': 1.0032312562689185, 'learning_rate/qwen_vl_interface': 7.521254165690317e-06, 'learning_rate/action_model': 7.521254165690317e-05, 'epoch': 1.11}) +[2026-06-17 20:38:35,791][starVLA.training.train_starvla][INFO] - Step 2456, Loss: {'train/loss': 0.5758532881736755, 'train/grad_norm_pre_clip': 15.504631996154785, 'timing/data': 0.0013792291283607483, 'timing/model': 0.9973744470626116, 'learning_rate/qwen_vl_interface': 7.519354496687933e-06, 'learning_rate/action_model': 7.519354496687933e-05, 'epoch': 1.11}) +[2026-06-17 20:38:39,811][starVLA.training.train_starvla][INFO] - Step 2457, Loss: {'train/loss': 0.3492676615715027, 'train/grad_norm_pre_clip': 11.145950317382812, 'timing/data': 0.0005090376362204552, 'timing/model': 0.9793553333729506, 'learning_rate/qwen_vl_interface': 7.517454357245918e-06, 'learning_rate/action_model': 7.517454357245917e-05, 'epoch': 1.11}) +[2026-06-17 20:38:43,692][starVLA.training.train_starvla][INFO] - Step 2458, Loss: {'train/loss': 0.42346593737602234, 'train/grad_norm_pre_clip': 15.455450057983398, 'timing/data': 0.0006573917344212532, 'timing/model': 0.9430219270288944, 'learning_rate/qwen_vl_interface': 7.515553747758176e-06, 'learning_rate/action_model': 7.515553747758176e-05, 'epoch': 1.12}) +[2026-06-17 20:38:48,560][starVLA.training.train_starvla][INFO] - Step 2459, Loss: {'train/loss': 0.406633198261261, 'train/grad_norm_pre_clip': 10.990345001220703, 'timing/data': 0.0008618906140327454, 'timing/model': 1.1958365077152848, 'learning_rate/qwen_vl_interface': 7.513652668618705e-06, 'learning_rate/action_model': 7.513652668618705e-05, 'epoch': 1.12}) +[2026-06-17 20:38:52,382][starVLA.training.train_starvla][INFO] - Step 2460, Loss: {'train/loss': 0.51474928855896, 'train/grad_norm_pre_clip': 14.36564826965332, 'timing/data': 0.0006051827222108841, 'timing/model': 1.0738160945475101, 'learning_rate/qwen_vl_interface': 7.5117511202215994e-06, 'learning_rate/action_model': 7.5117511202216e-05, 'epoch': 1.12}) +[2026-06-17 20:38:56,004][starVLA.training.train_starvla][INFO] - Step 2461, Loss: {'train/loss': 0.49945658445358276, 'train/grad_norm_pre_clip': 22.60791778564453, 'timing/data': 0.0005284007638692856, 'timing/model': 0.9495694544166327, 'learning_rate/qwen_vl_interface': 7.509849102961053e-06, 'learning_rate/action_model': 7.509849102961053e-05, 'epoch': 1.12}) +[2026-06-17 20:38:59,898][starVLA.training.train_starvla][INFO] - Step 2462, Loss: {'train/loss': 0.4183969497680664, 'train/grad_norm_pre_clip': 12.367288589477539, 'timing/data': 0.0006128037348389626, 'timing/model': 0.952147901058197, 'learning_rate/qwen_vl_interface': 7.507946617231357e-06, 'learning_rate/action_model': 7.507946617231356e-05, 'epoch': 1.12}) +[2026-06-17 20:39:03,888][starVLA.training.train_starvla][INFO] - Step 2463, Loss: {'train/loss': 0.3598165810108185, 'train/grad_norm_pre_clip': 10.542962074279785, 'timing/data': 0.0005207322537899017, 'timing/model': 0.9371948195621371, 'learning_rate/qwen_vl_interface': 7.5060436634268965e-06, 'learning_rate/action_model': 7.506043663426896e-05, 'epoch': 1.12}) +[2026-06-17 20:39:07,694][starVLA.training.train_starvla][INFO] - Step 2464, Loss: {'train/loss': 0.13196362555027008, 'train/grad_norm_pre_clip': 6.789918899536133, 'timing/data': 0.0005305437371134758, 'timing/model': 1.0088697308674455, 'learning_rate/qwen_vl_interface': 7.504140241942157e-06, 'learning_rate/action_model': 7.504140241942157e-05, 'epoch': 1.12}) +[2026-06-17 20:39:11,638][starVLA.training.train_starvla][INFO] - Step 2465, Loss: {'train/loss': 0.413979709148407, 'train/grad_norm_pre_clip': 16.64556312561035, 'timing/data': 0.01063349936157465, 'timing/model': 0.9429760230705142, 'learning_rate/qwen_vl_interface': 7.502236353171718e-06, 'learning_rate/action_model': 7.502236353171718e-05, 'epoch': 1.12}) +[2026-06-17 20:39:15,461][starVLA.training.train_starvla][INFO] - Step 2466, Loss: {'train/loss': 0.3665950894355774, 'train/grad_norm_pre_clip': 13.67611026763916, 'timing/data': 0.011405284516513348, 'timing/model': 1.0955426823347807, 'learning_rate/qwen_vl_interface': 7.500331997510262e-06, 'learning_rate/action_model': 7.500331997510262e-05, 'epoch': 1.12}) +[2026-06-17 20:39:19,182][starVLA.training.train_starvla][INFO] - Step 2467, Loss: {'train/loss': 0.5805344581604004, 'train/grad_norm_pre_clip': 15.292606353759766, 'timing/data': 0.010046782903373241, 'timing/model': 0.9480184400454164, 'learning_rate/qwen_vl_interface': 7.498427175352561e-06, 'learning_rate/action_model': 7.49842717535256e-05, 'epoch': 1.12}) +[2026-06-17 20:39:23,096][starVLA.training.train_starvla][INFO] - Step 2468, Loss: {'train/loss': 0.3687814474105835, 'train/grad_norm_pre_clip': 16.506587982177734, 'timing/data': 0.015856098383665085, 'timing/model': 1.0057315044105053, 'learning_rate/qwen_vl_interface': 7.496521887093487e-06, 'learning_rate/action_model': 7.496521887093486e-05, 'epoch': 1.12}) +[2026-06-17 20:39:26,836][starVLA.training.train_starvla][INFO] - Step 2469, Loss: {'train/loss': 0.505864679813385, 'train/grad_norm_pre_clip': 15.994156837463379, 'timing/data': 0.010704748332500458, 'timing/model': 0.9434707527980208, 'learning_rate/qwen_vl_interface': 7.494616133128008e-06, 'learning_rate/action_model': 7.494616133128008e-05, 'epoch': 1.12}) +[2026-06-17 20:39:32,100][starVLA.training.train_starvla][INFO] - Step 2470, Loss: {'train/loss': 0.6926020383834839, 'train/grad_norm_pre_clip': 52.66463851928711, 'timing/data': 0.012595361098647118, 'timing/model': 0.9458406679332256, 'learning_rate/qwen_vl_interface': 7.4927099138511895e-06, 'learning_rate/action_model': 7.492709913851189e-05, 'epoch': 1.12}) +[2026-06-17 20:39:37,193][starVLA.training.train_starvla][INFO] - Step 2471, Loss: {'train/loss': 0.5943603515625, 'train/grad_norm_pre_clip': 17.856468200683594, 'timing/data': 0.01198124885559082, 'timing/model': 0.9562640637159348, 'learning_rate/qwen_vl_interface': 7.490803229658192e-06, 'learning_rate/action_model': 7.490803229658192e-05, 'epoch': 1.12}) +[2026-06-17 20:39:41,291][starVLA.training.train_starvla][INFO] - Step 2472, Loss: {'train/loss': 0.244268000125885, 'train/grad_norm_pre_clip': 15.410222053527832, 'timing/data': 0.013384047895669937, 'timing/model': 1.4004546394571662, 'learning_rate/qwen_vl_interface': 7.488896080944274e-06, 'learning_rate/action_model': 7.488896080944273e-05, 'epoch': 1.12}) +[2026-06-17 20:39:45,485][starVLA.training.train_starvla][INFO] - Step 2473, Loss: {'train/loss': 0.5923616886138916, 'train/grad_norm_pre_clip': 16.56674575805664, 'timing/data': 0.016071699559688568, 'timing/model': 1.4178120354190469, 'learning_rate/qwen_vl_interface': 7.4869884681047896e-06, 'learning_rate/action_model': 7.486988468104789e-05, 'epoch': 1.12}) +[2026-06-17 20:39:50,882][starVLA.training.train_starvla][INFO] - Step 2474, Loss: {'train/loss': 0.5654970407485962, 'train/grad_norm_pre_clip': 16.967145919799805, 'timing/data': 0.009179597720503807, 'timing/model': 2.6866108179092407, 'learning_rate/qwen_vl_interface': 7.485080391535189e-06, 'learning_rate/action_model': 7.485080391535188e-05, 'epoch': 1.12}) +[2026-06-17 20:39:54,854][starVLA.training.train_starvla][INFO] - Step 2475, Loss: {'train/loss': 0.5512813329696655, 'train/grad_norm_pre_clip': 15.038763046264648, 'timing/data': 0.010799253359436989, 'timing/model': 0.945919775404036, 'learning_rate/qwen_vl_interface': 7.483171851631018e-06, 'learning_rate/action_model': 7.483171851631017e-05, 'epoch': 1.12}) +[2026-06-17 20:39:58,735][starVLA.training.train_starvla][INFO] - Step 2476, Loss: {'train/loss': 0.488788366317749, 'train/grad_norm_pre_clip': 16.497900009155273, 'timing/data': 0.00978928990662098, 'timing/model': 0.9360020579770207, 'learning_rate/qwen_vl_interface': 7.481262848787919e-06, 'learning_rate/action_model': 7.481262848787919e-05, 'epoch': 1.12}) +[2026-06-17 20:40:03,047][starVLA.training.train_starvla][INFO] - Step 2477, Loss: {'train/loss': 0.276690274477005, 'train/grad_norm_pre_clip': 14.474212646484375, 'timing/data': 0.007860040292143822, 'timing/model': 0.9491297518834472, 'learning_rate/qwen_vl_interface': 7.479353383401631e-06, 'learning_rate/action_model': 7.47935338340163e-05, 'epoch': 1.12}) +[2026-06-17 20:40:07,069][starVLA.training.train_starvla][INFO] - Step 2478, Loss: {'train/loss': 0.2967689335346222, 'train/grad_norm_pre_clip': 12.029500007629395, 'timing/data': 0.007702270522713661, 'timing/model': 0.9587175976485014, 'learning_rate/qwen_vl_interface': 7.477443455867987e-06, 'learning_rate/action_model': 7.477443455867988e-05, 'epoch': 1.12}) +[2026-06-17 20:40:10,963][starVLA.training.train_starvla][INFO] - Step 2479, Loss: {'train/loss': 0.4765018820762634, 'train/grad_norm_pre_clip': 16.35422134399414, 'timing/data': 0.010455767624080181, 'timing/model': 1.1440406581386924, 'learning_rate/qwen_vl_interface': 7.475533066582917e-06, 'learning_rate/action_model': 7.475533066582918e-05, 'epoch': 1.12}) +[2026-06-17 20:40:15,496][starVLA.training.train_starvla][INFO] - Step 2480, Loss: {'train/loss': 0.5104604363441467, 'train/grad_norm_pre_clip': 14.105213165283203, 'timing/data': 0.008092840202152729, 'timing/model': 1.7600963823497295, 'learning_rate/qwen_vl_interface': 7.47362221594245e-06, 'learning_rate/action_model': 7.47362221594245e-05, 'epoch': 1.13}) +[2026-06-17 20:40:19,223][starVLA.training.train_starvla][INFO] - Step 2481, Loss: {'train/loss': 0.46236592531204224, 'train/grad_norm_pre_clip': 14.73755931854248, 'timing/data': 0.01126146037131548, 'timing/model': 0.9986236775293946, 'learning_rate/qwen_vl_interface': 7.4717109043427025e-06, 'learning_rate/action_model': 7.471710904342702e-05, 'epoch': 1.13}) +[2026-06-17 20:40:23,216][starVLA.training.train_starvla][INFO] - Step 2482, Loss: {'train/loss': 0.16623321175575256, 'train/grad_norm_pre_clip': 9.209656715393066, 'timing/data': 0.01587000396102667, 'timing/model': 0.958903044462204, 'learning_rate/qwen_vl_interface': 7.469799132179894e-06, 'learning_rate/action_model': 7.469799132179895e-05, 'epoch': 1.13}) +[2026-06-17 20:40:28,565][starVLA.training.train_starvla][INFO] - Step 2483, Loss: {'train/loss': 0.9539651870727539, 'train/grad_norm_pre_clip': 16.895235061645508, 'timing/data': 0.01123411301523447, 'timing/model': 0.9455743385478854, 'learning_rate/qwen_vl_interface': 7.467886899850338e-06, 'learning_rate/action_model': 7.467886899850338e-05, 'epoch': 1.13}) +[2026-06-17 20:40:32,383][starVLA.training.train_starvla][INFO] - Step 2484, Loss: {'train/loss': 0.1704893559217453, 'train/grad_norm_pre_clip': 8.1088285446167, 'timing/data': 0.009197073988616467, 'timing/model': 0.9542208127677441, 'learning_rate/qwen_vl_interface': 7.465974207750442e-06, 'learning_rate/action_model': 7.465974207750442e-05, 'epoch': 1.13}) +[2026-06-17 20:40:37,257][starVLA.training.train_starvla][INFO] - Step 2485, Loss: {'train/loss': 0.19164742529392242, 'train/grad_norm_pre_clip': 10.657512664794922, 'timing/data': 0.009627114050090313, 'timing/model': 1.3895938200876117, 'learning_rate/qwen_vl_interface': 7.464061056276707e-06, 'learning_rate/action_model': 7.464061056276707e-05, 'epoch': 1.13}) +[2026-06-17 20:40:41,816][starVLA.training.train_starvla][INFO] - Step 2486, Loss: {'train/loss': 0.5122068524360657, 'train/grad_norm_pre_clip': 15.174729347229004, 'timing/data': 0.01477768924087286, 'timing/model': 1.1506096795201302, 'learning_rate/qwen_vl_interface': 7.462147445825733e-06, 'learning_rate/action_model': 7.462147445825733e-05, 'epoch': 1.13}) +[2026-06-17 20:40:46,906][starVLA.training.train_starvla][INFO] - Step 2487, Loss: {'train/loss': 0.36639511585235596, 'train/grad_norm_pre_clip': 12.394977569580078, 'timing/data': 0.016240907832980156, 'timing/model': 0.9572254782542586, 'learning_rate/qwen_vl_interface': 7.460233376794211e-06, 'learning_rate/action_model': 7.46023337679421e-05, 'epoch': 1.13}) +[2026-06-17 20:40:50,961][starVLA.training.train_starvla][INFO] - Step 2488, Loss: {'train/loss': 0.8432286381721497, 'train/grad_norm_pre_clip': 22.53571319580078, 'timing/data': 0.011879665777087212, 'timing/model': 0.9957679687067866, 'learning_rate/qwen_vl_interface': 7.458318849578934e-06, 'learning_rate/action_model': 7.458318849578933e-05, 'epoch': 1.13}) +[2026-06-17 20:40:54,903][starVLA.training.train_starvla][INFO] - Step 2489, Loss: {'train/loss': 0.5884489417076111, 'train/grad_norm_pre_clip': 14.90475845336914, 'timing/data': 0.00916311889886856, 'timing/model': 1.023369519971311, 'learning_rate/qwen_vl_interface': 7.456403864576782e-06, 'learning_rate/action_model': 7.456403864576782e-05, 'epoch': 1.13}) +[2026-06-17 20:40:58,813][starVLA.training.train_starvla][INFO] - Step 2490, Loss: {'train/loss': 0.5892996191978455, 'train/grad_norm_pre_clip': 17.141090393066406, 'timing/data': 0.009307161904871464, 'timing/model': 0.9802338117733598, 'learning_rate/qwen_vl_interface': 7.4544884221847354e-06, 'learning_rate/action_model': 7.454488422184735e-05, 'epoch': 1.13}) +[2026-06-17 20:41:02,670][starVLA.training.train_starvla][INFO] - Step 2491, Loss: {'train/loss': 0.4203023612499237, 'train/grad_norm_pre_clip': 13.560483932495117, 'timing/data': 0.011938857845962048, 'timing/model': 0.9436527481302619, 'learning_rate/qwen_vl_interface': 7.452572522799867e-06, 'learning_rate/action_model': 7.452572522799866e-05, 'epoch': 1.13}) +[2026-06-17 20:41:06,541][starVLA.training.train_starvla][INFO] - Step 2492, Loss: {'train/loss': 0.2805175483226776, 'train/grad_norm_pre_clip': 10.528621673583984, 'timing/data': 0.01156708225607872, 'timing/model': 0.967905392870307, 'learning_rate/qwen_vl_interface': 7.450656166819344e-06, 'learning_rate/action_model': 7.450656166819344e-05, 'epoch': 1.13}) +[2026-06-17 20:41:10,461][starVLA.training.train_starvla][INFO] - Step 2493, Loss: {'train/loss': 0.46090665459632874, 'train/grad_norm_pre_clip': 17.167661666870117, 'timing/data': 0.00943800900131464, 'timing/model': 1.186490273103118, 'learning_rate/qwen_vl_interface': 7.4487393546404295e-06, 'learning_rate/action_model': 7.44873935464043e-05, 'epoch': 1.13}) +[2026-06-17 20:41:14,249][starVLA.training.train_starvla][INFO] - Step 2494, Loss: {'train/loss': 0.465830534696579, 'train/grad_norm_pre_clip': 16.038148880004883, 'timing/data': 0.009431235492229462, 'timing/model': 0.9607553612440825, 'learning_rate/qwen_vl_interface': 7.446822086660481e-06, 'learning_rate/action_model': 7.446822086660481e-05, 'epoch': 1.13}) +[2026-06-17 20:41:18,257][starVLA.training.train_starvla][INFO] - Step 2495, Loss: {'train/loss': 0.17827819287776947, 'train/grad_norm_pre_clip': 9.726991653442383, 'timing/data': 0.00910617969930172, 'timing/model': 1.0624984567984939, 'learning_rate/qwen_vl_interface': 7.444904363276951e-06, 'learning_rate/action_model': 7.444904363276951e-05, 'epoch': 1.13}) +[2026-06-17 20:41:22,136][starVLA.training.train_starvla][INFO] - Step 2496, Loss: {'train/loss': 0.2746131122112274, 'train/grad_norm_pre_clip': 12.889750480651855, 'timing/data': 0.009445480071008205, 'timing/model': 0.947555479593575, 'learning_rate/qwen_vl_interface': 7.442986184887383e-06, 'learning_rate/action_model': 7.442986184887383e-05, 'epoch': 1.13}) +[2026-06-17 20:41:25,947][starVLA.training.train_starvla][INFO] - Step 2497, Loss: {'train/loss': 0.4385719299316406, 'train/grad_norm_pre_clip': 23.331832885742188, 'timing/data': 0.011133942753076553, 'timing/model': 0.9386807270348072, 'learning_rate/qwen_vl_interface': 7.4410675518894205e-06, 'learning_rate/action_model': 7.44106755188942e-05, 'epoch': 1.13}) +[2026-06-17 20:41:29,742][starVLA.training.train_starvla][INFO] - Step 2498, Loss: {'train/loss': 0.6373937129974365, 'train/grad_norm_pre_clip': 14.52017879486084, 'timing/data': 0.009215723723173141, 'timing/model': 0.9406650681048632, 'learning_rate/qwen_vl_interface': 7.439148464680795e-06, 'learning_rate/action_model': 7.439148464680796e-05, 'epoch': 1.13}) +[2026-06-17 20:41:33,661][starVLA.training.train_starvla][INFO] - Step 2499, Loss: {'train/loss': 0.4082415699958801, 'train/grad_norm_pre_clip': 17.311357498168945, 'timing/data': 0.014230797998607159, 'timing/model': 1.155935738235712, 'learning_rate/qwen_vl_interface': 7.4372289236593395e-06, 'learning_rate/action_model': 7.43722892365934e-05, 'epoch': 1.13}) +[2026-06-17 20:43:53,933][starVLA.training.train_starvla][INFO] - Step 2500, Loss: {'train/loss': 0.6165683269500732, 'train/grad_norm_pre_clip': 14.756574630737305, 'eval/loss': 0.40635964274406433, 'eval/latency_8/loss': 0.40635964274406433, 'eval/demon_fire_precision': 0.9191134139320668, 'eval/demon_fire_recall': 0.9919229574401989, 'eval/demon_fire_f1': 0.9541311818317646, 'eval/demon_fire_rate_model': 0.9666110183639399, 'eval/demon_fire_rate_teacher': 0.8956594323873122, 'eval/demon_left_precision': 0.8675496688741722, 'eval/demon_left_recall': 0.5610278372591007, 'eval/demon_left_f1': 0.6814044213263979, 'eval/demon_left_rate_model': 0.08402893711741792, 'eval/demon_left_rate_teacher': 0.12993878686700056, 'eval/demon_right_precision': 0.751453488372093, 'eval/demon_right_recall': 0.7833333333333333, 'eval/demon_right_f1': 0.7670623145400594, 'eval/demon_right_rate_model': 0.1914301613800779, 'eval/demon_right_rate_teacher': 0.18363939899833054, 'eval/demon_move_f1': 0.7242333679332287, 'eval/demon_cc_f1': 0.8711690324863435, 'eval/latency_8/demon_fire_precision': 0.9191134139320668, 'eval/latency_8/demon_fire_recall': 0.9919229574401989, 'eval/latency_8/demon_fire_f1': 0.9541311818317646, 'eval/latency_8/demon_fire_rate_model': 0.9666110183639399, 'eval/latency_8/demon_fire_rate_teacher': 0.8956594323873122, 'eval/latency_8/demon_left_precision': 0.8675496688741722, 'eval/latency_8/demon_left_recall': 0.5610278372591007, 'eval/latency_8/demon_left_f1': 0.6814044213263979, 'eval/latency_8/demon_left_rate_model': 0.08402893711741792, 'eval/latency_8/demon_left_rate_teacher': 0.12993878686700056, 'eval/latency_8/demon_right_precision': 0.751453488372093, 'eval/latency_8/demon_right_recall': 0.7833333333333333, 'eval/latency_8/demon_right_f1': 0.7670623145400594, 'eval/latency_8/demon_right_rate_model': 0.1914301613800779, 'eval/latency_8/demon_right_rate_teacher': 0.18363939899833054, 'eval/latency_8/demon_move_f1': 0.7242333679332287, 'eval/latency_8/demon_cc_f1': 0.8711690324863435, 'timing/data': 0.00808202289044857, 'timing/model': 1.125391609966755, 'learning_rate/qwen_vl_interface': 7.435308929222973e-06, 'learning_rate/action_model': 7.435308929222972e-05, 'epoch': 1.13}) +[2026-06-17 20:43:53,951][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_2500_state +[2026-06-17 20:43:53,956][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 20:44:00,574][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_2500_state/model.safetensors +[2026-06-17 20:44:16,373][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_2500_state/optimizer.bin +[2026-06-17 20:44:16,374][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_2500_state/sampler.bin +[2026-06-17 20:44:16,374][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_2500_state/sampler_1.bin +[2026-06-17 20:44:16,377][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_2500_state/random_states_0.pkl +[2026-06-17 20:44:16,382][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2500 +[2026-06-17 20:44:16,382][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 20:44:16,396][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 20:44:20,304][starVLA.training.train_starvla][INFO] - Step 2501, Loss: {'train/loss': 0.3533133566379547, 'train/grad_norm_pre_clip': 12.903119087219238, 'timing/data': 0.009516874328255653, 'timing/model': 1.1540137715637684, 'learning_rate/qwen_vl_interface': 7.433388481769716e-06, 'learning_rate/action_model': 7.433388481769715e-05, 'epoch': 1.13}) +[2026-06-17 20:44:24,180][starVLA.training.train_starvla][INFO] - Step 2502, Loss: {'train/loss': 0.537521243095398, 'train/grad_norm_pre_clip': 18.732452392578125, 'timing/data': 0.009508442133665085, 'timing/model': 1.1302159233018756, 'learning_rate/qwen_vl_interface': 7.431467581697675e-06, 'learning_rate/action_model': 7.431467581697674e-05, 'epoch': 1.14}) +[2026-06-17 20:44:27,875][starVLA.training.train_starvla][INFO] - Step 2503, Loss: {'train/loss': 0.26081132888793945, 'train/grad_norm_pre_clip': 12.73261547088623, 'timing/data': 0.007853783667087555, 'timing/model': 0.9310698574408889, 'learning_rate/qwen_vl_interface': 7.429546229405057e-06, 'learning_rate/action_model': 7.429546229405057e-05, 'epoch': 1.14}) +[2026-06-17 20:44:33,574][starVLA.training.train_starvla][INFO] - Step 2504, Loss: {'train/loss': 0.2478512078523636, 'train/grad_norm_pre_clip': 10.926804542541504, 'timing/data': 0.00909747090190649, 'timing/model': 0.9348967615514994, 'learning_rate/qwen_vl_interface': 7.427624425290162e-06, 'learning_rate/action_model': 7.427624425290161e-05, 'epoch': 1.14}) +[2026-06-17 20:44:37,490][starVLA.training.train_starvla][INFO] - Step 2505, Loss: {'train/loss': 0.4149060547351837, 'train/grad_norm_pre_clip': 16.3151912689209, 'timing/data': 0.008512815460562706, 'timing/model': 0.9462484726682305, 'learning_rate/qwen_vl_interface': 7.425702169751379e-06, 'learning_rate/action_model': 7.425702169751378e-05, 'epoch': 1.14}) +[2026-06-17 20:44:42,776][starVLA.training.train_starvla][INFO] - Step 2506, Loss: {'train/loss': 0.5416938662528992, 'train/grad_norm_pre_clip': 16.97120475769043, 'timing/data': 0.009450331330299377, 'timing/model': 0.9339071158319712, 'learning_rate/qwen_vl_interface': 7.4237794631871965e-06, 'learning_rate/action_model': 7.423779463187196e-05, 'epoch': 1.14}) +[2026-06-17 20:44:46,566][starVLA.training.train_starvla][INFO] - Step 2507, Loss: {'train/loss': 0.39825567603111267, 'train/grad_norm_pre_clip': 12.991085052490234, 'timing/data': 0.012289744801819324, 'timing/model': 0.9362260485067964, 'learning_rate/qwen_vl_interface': 7.42185630599619e-06, 'learning_rate/action_model': 7.42185630599619e-05, 'epoch': 1.14}) +[2026-06-17 20:44:50,529][starVLA.training.train_starvla][INFO] - Step 2508, Loss: {'train/loss': 0.40495213866233826, 'train/grad_norm_pre_clip': 12.2139253616333, 'timing/data': 0.017101156525313854, 'timing/model': 1.1247989470139146, 'learning_rate/qwen_vl_interface': 7.419932698577036e-06, 'learning_rate/action_model': 7.419932698577035e-05, 'epoch': 1.14}) +[2026-06-17 20:44:54,341][starVLA.training.train_starvla][INFO] - Step 2509, Loss: {'train/loss': 0.1978687345981598, 'train/grad_norm_pre_clip': 9.530743598937988, 'timing/data': 0.010688718408346176, 'timing/model': 0.9493505330756307, 'learning_rate/qwen_vl_interface': 7.418008641328498e-06, 'learning_rate/action_model': 7.418008641328497e-05, 'epoch': 1.14}) +[2026-06-17 20:44:58,209][starVLA.training.train_starvla][INFO] - Step 2510, Loss: {'train/loss': 0.5217657685279846, 'train/grad_norm_pre_clip': 16.454124450683594, 'timing/data': 0.007275951094925404, 'timing/model': 0.9851314490661025, 'learning_rate/qwen_vl_interface': 7.416084134649435e-06, 'learning_rate/action_model': 7.416084134649435e-05, 'epoch': 1.14}) +[2026-06-17 20:45:02,155][starVLA.training.train_starvla][INFO] - Step 2511, Loss: {'train/loss': 0.2648431360721588, 'train/grad_norm_pre_clip': 12.328254699707031, 'timing/data': 0.011660589836537838, 'timing/model': 1.011010967195034, 'learning_rate/qwen_vl_interface': 7.414159178938801e-06, 'learning_rate/action_model': 7.414159178938801e-05, 'epoch': 1.14}) +[2026-06-17 20:45:06,201][starVLA.training.train_starvla][INFO] - Step 2512, Loss: {'train/loss': 0.4226474463939667, 'train/grad_norm_pre_clip': 16.749107360839844, 'timing/data': 0.011688729748129845, 'timing/model': 1.0014739613980055, 'learning_rate/qwen_vl_interface': 7.412233774595639e-06, 'learning_rate/action_model': 7.41223377459564e-05, 'epoch': 1.14}) +[2026-06-17 20:45:11,732][starVLA.training.train_starvla][INFO] - Step 2513, Loss: {'train/loss': 0.40003594756126404, 'train/grad_norm_pre_clip': 14.879217147827148, 'timing/data': 0.015411030501127243, 'timing/model': 2.1295416792854667, 'learning_rate/qwen_vl_interface': 7.410307922019091e-06, 'learning_rate/action_model': 7.41030792201909e-05, 'epoch': 1.14}) +[2026-06-17 20:45:16,699][starVLA.training.train_starvla][INFO] - Step 2514, Loss: {'train/loss': 0.22789189219474792, 'train/grad_norm_pre_clip': 8.67659854888916, 'timing/data': 0.013694310560822487, 'timing/model': 1.8230044599622488, 'learning_rate/qwen_vl_interface': 7.408381621608387e-06, 'learning_rate/action_model': 7.408381621608387e-05, 'epoch': 1.14}) +[2026-06-17 20:45:20,723][starVLA.training.train_starvla][INFO] - Step 2515, Loss: {'train/loss': 0.44018879532814026, 'train/grad_norm_pre_clip': 14.031380653381348, 'timing/data': 0.01367832999676466, 'timing/model': 1.3057959852740169, 'learning_rate/qwen_vl_interface': 7.406454873762848e-06, 'learning_rate/action_model': 7.406454873762848e-05, 'epoch': 1.14}) +[2026-06-17 20:45:24,569][starVLA.training.train_starvla][INFO] - Step 2516, Loss: {'train/loss': 0.4485425651073456, 'train/grad_norm_pre_clip': 12.744357109069824, 'timing/data': 0.00878832582384348, 'timing/model': 1.1195046678185463, 'learning_rate/qwen_vl_interface': 7.404527678881894e-06, 'learning_rate/action_model': 7.404527678881894e-05, 'epoch': 1.14}) +[2026-06-17 20:45:28,451][starVLA.training.train_starvla][INFO] - Step 2517, Loss: {'train/loss': 0.6393772959709167, 'train/grad_norm_pre_clip': 14.355219841003418, 'timing/data': 0.010075714439153671, 'timing/model': 0.9586615832522511, 'learning_rate/qwen_vl_interface': 7.402600037365034e-06, 'learning_rate/action_model': 7.402600037365034e-05, 'epoch': 1.14}) +[2026-06-17 20:45:32,343][starVLA.training.train_starvla][INFO] - Step 2518, Loss: {'train/loss': 0.5801749229431152, 'train/grad_norm_pre_clip': 17.206315994262695, 'timing/data': 0.009926244616508484, 'timing/model': 0.9427227480337024, 'learning_rate/qwen_vl_interface': 7.400671949611869e-06, 'learning_rate/action_model': 7.40067194961187e-05, 'epoch': 1.14}) +[2026-06-17 20:45:36,133][starVLA.training.train_starvla][INFO] - Step 2519, Loss: {'train/loss': 0.5003592371940613, 'train/grad_norm_pre_clip': 18.69684410095215, 'timing/data': 0.01375868171453476, 'timing/model': 0.9401258770376444, 'learning_rate/qwen_vl_interface': 7.398743416022095e-06, 'learning_rate/action_model': 7.398743416022095e-05, 'epoch': 1.14}) +[2026-06-17 20:45:39,954][starVLA.training.train_starvla][INFO] - Step 2520, Loss: {'train/loss': 0.4197371304035187, 'train/grad_norm_pre_clip': 15.490596771240234, 'timing/data': 0.010958798229694366, 'timing/model': 0.9413587395101786, 'learning_rate/qwen_vl_interface': 7.3968144369954995e-06, 'learning_rate/action_model': 7.396814436995499e-05, 'epoch': 1.14}) +[2026-06-17 20:45:43,865][starVLA.training.train_starvla][INFO] - Step 2521, Loss: {'train/loss': 0.6781740784645081, 'train/grad_norm_pre_clip': 20.709978103637695, 'timing/data': 0.018552672117948532, 'timing/model': 1.1814891146495938, 'learning_rate/qwen_vl_interface': 7.39488501293196e-06, 'learning_rate/action_model': 7.39488501293196e-05, 'epoch': 1.14}) +[2026-06-17 20:45:47,668][starVLA.training.train_starvla][INFO] - Step 2522, Loss: {'train/loss': 0.6405459046363831, 'train/grad_norm_pre_clip': 15.936378479003906, 'timing/data': 0.014038783498108387, 'timing/model': 1.0709290327504277, 'learning_rate/qwen_vl_interface': 7.392955144231452e-06, 'learning_rate/action_model': 7.392955144231452e-05, 'epoch': 1.14}) +[2026-06-17 20:45:51,355][starVLA.training.train_starvla][INFO] - Step 2523, Loss: {'train/loss': 0.40895265340805054, 'train/grad_norm_pre_clip': 28.69108772277832, 'timing/data': 0.007864220067858696, 'timing/model': 0.9474392822012305, 'learning_rate/qwen_vl_interface': 7.391024831294034e-06, 'learning_rate/action_model': 7.391024831294033e-05, 'epoch': 1.14}) +[2026-06-17 20:45:55,145][starVLA.training.train_starvla][INFO] - Step 2524, Loss: {'train/loss': 0.31955960392951965, 'train/grad_norm_pre_clip': 9.822000503540039, 'timing/data': 0.009206462651491165, 'timing/model': 0.9351148633286357, 'learning_rate/qwen_vl_interface': 7.389094074519866e-06, 'learning_rate/action_model': 7.389094074519866e-05, 'epoch': 1.15}) +[2026-06-17 20:45:59,077][starVLA.training.train_starvla][INFO] - Step 2525, Loss: {'train/loss': 0.2161962389945984, 'train/grad_norm_pre_clip': 9.51877212524414, 'timing/data': 0.00761662982404232, 'timing/model': 0.9334127316251397, 'learning_rate/qwen_vl_interface': 7.387162874309194e-06, 'learning_rate/action_model': 7.387162874309193e-05, 'epoch': 1.15}) +[2026-06-17 20:46:02,847][starVLA.training.train_starvla][INFO] - Step 2526, Loss: {'train/loss': 0.47009292244911194, 'train/grad_norm_pre_clip': 14.644882202148438, 'timing/data': 0.01008590403944254, 'timing/model': 0.9426472093909979, 'learning_rate/qwen_vl_interface': 7.385231231062359e-06, 'learning_rate/action_model': 7.385231231062358e-05, 'epoch': 1.15}) +[2026-06-17 20:46:06,682][starVLA.training.train_starvla][INFO] - Step 2527, Loss: {'train/loss': 0.3204224407672882, 'train/grad_norm_pre_clip': 15.19143009185791, 'timing/data': 0.00836962554603815, 'timing/model': 0.9467973709106445, 'learning_rate/qwen_vl_interface': 7.383299145179789e-06, 'learning_rate/action_model': 7.383299145179789e-05, 'epoch': 1.15}) +[2026-06-17 20:46:10,557][starVLA.training.train_starvla][INFO] - Step 2528, Loss: {'train/loss': 0.618599534034729, 'train/grad_norm_pre_clip': 14.580530166625977, 'timing/data': 0.008812110871076584, 'timing/model': 1.1730573372915387, 'learning_rate/qwen_vl_interface': 7.381366617062014e-06, 'learning_rate/action_model': 7.381366617062014e-05, 'epoch': 1.15}) +[2026-06-17 20:46:14,240][starVLA.training.train_starvla][INFO] - Step 2529, Loss: {'train/loss': 0.27793440222740173, 'train/grad_norm_pre_clip': 11.825645446777344, 'timing/data': 0.010083673521876335, 'timing/model': 0.9465165939182043, 'learning_rate/qwen_vl_interface': 7.379433647109645e-06, 'learning_rate/action_model': 7.379433647109645e-05, 'epoch': 1.15}) +[2026-06-17 20:46:18,044][starVLA.training.train_starvla][INFO] - Step 2530, Loss: {'train/loss': 0.3416246771812439, 'train/grad_norm_pre_clip': 11.86194133758545, 'timing/data': 0.008023195900022984, 'timing/model': 0.9417142374441028, 'learning_rate/qwen_vl_interface': 7.37750023572339e-06, 'learning_rate/action_model': 7.37750023572339e-05, 'epoch': 1.15}) +[2026-06-17 20:46:21,837][starVLA.training.train_starvla][INFO] - Step 2531, Loss: {'train/loss': 0.5682961344718933, 'train/grad_norm_pre_clip': 17.636730194091797, 'timing/data': 0.015520472079515457, 'timing/model': 0.9425777308642864, 'learning_rate/qwen_vl_interface': 7.375566383304047e-06, 'learning_rate/action_model': 7.375566383304047e-05, 'epoch': 1.15}) +[2026-06-17 20:46:25,739][starVLA.training.train_starvla][INFO] - Step 2532, Loss: {'train/loss': 0.5618619918823242, 'train/grad_norm_pre_clip': 16.552013397216797, 'timing/data': 0.008551371283829212, 'timing/model': 0.94651098549366, 'learning_rate/qwen_vl_interface': 7.373632090252505e-06, 'learning_rate/action_model': 7.373632090252505e-05, 'epoch': 1.15}) +[2026-06-17 20:46:29,944][starVLA.training.train_starvla][INFO] - Step 2533, Loss: {'train/loss': 0.23700207471847534, 'train/grad_norm_pre_clip': 11.147128105163574, 'timing/data': 0.009629080072045326, 'timing/model': 0.9395834859460592, 'learning_rate/qwen_vl_interface': 7.371697356969746e-06, 'learning_rate/action_model': 7.371697356969746e-05, 'epoch': 1.15}) +[2026-06-17 20:46:33,798][starVLA.training.train_starvla][INFO] - Step 2534, Loss: {'train/loss': 0.5278835892677307, 'train/grad_norm_pre_clip': 15.161499977111816, 'timing/data': 0.0135740851983428, 'timing/model': 1.1352833975106478, 'learning_rate/qwen_vl_interface': 7.36976218385684e-06, 'learning_rate/action_model': 7.36976218385684e-05, 'epoch': 1.15}) +[2026-06-17 20:46:38,378][starVLA.training.train_starvla][INFO] - Step 2535, Loss: {'train/loss': 0.42015111446380615, 'train/grad_norm_pre_clip': 17.34140396118164, 'timing/data': 0.019882814027369022, 'timing/model': 1.87338597048074, 'learning_rate/qwen_vl_interface': 7.367826571314955e-06, 'learning_rate/action_model': 7.367826571314954e-05, 'epoch': 1.15}) +[2026-06-17 20:46:42,061][starVLA.training.train_starvla][INFO] - Step 2536, Loss: {'train/loss': 0.301815927028656, 'train/grad_norm_pre_clip': 15.816361427307129, 'timing/data': 0.008236352354288101, 'timing/model': 0.967197279445827, 'learning_rate/qwen_vl_interface': 7.36589051974534e-06, 'learning_rate/action_model': 7.36589051974534e-05, 'epoch': 1.15}) +[2026-06-17 20:46:46,038][starVLA.training.train_starvla][INFO] - Step 2537, Loss: {'train/loss': 0.4120454788208008, 'train/grad_norm_pre_clip': 13.114423751831055, 'timing/data': 0.013562936335802078, 'timing/model': 0.9954166328534484, 'learning_rate/qwen_vl_interface': 7.363954029549348e-06, 'learning_rate/action_model': 7.363954029549348e-05, 'epoch': 1.15}) +[2026-06-17 20:46:49,996][starVLA.training.train_starvla][INFO] - Step 2538, Loss: {'train/loss': 0.2613854706287384, 'train/grad_norm_pre_clip': 9.716080665588379, 'timing/data': 0.012784721329808235, 'timing/model': 0.9541004449129105, 'learning_rate/qwen_vl_interface': 7.362017101128406e-06, 'learning_rate/action_model': 7.362017101128406e-05, 'epoch': 1.15}) +[2026-06-17 20:46:53,922][starVLA.training.train_starvla][INFO] - Step 2539, Loss: {'train/loss': 0.7037624716758728, 'train/grad_norm_pre_clip': 15.255321502685547, 'timing/data': 0.00974816083908081, 'timing/model': 0.9535670494660735, 'learning_rate/qwen_vl_interface': 7.360079734884049e-06, 'learning_rate/action_model': 7.360079734884049e-05, 'epoch': 1.15}) +[2026-06-17 20:46:57,831][starVLA.training.train_starvla][INFO] - Step 2540, Loss: {'train/loss': 0.3068748414516449, 'train/grad_norm_pre_clip': 10.891366958618164, 'timing/data': 0.011990075930953026, 'timing/model': 0.9595430670306087, 'learning_rate/qwen_vl_interface': 7.358141931217893e-06, 'learning_rate/action_model': 7.358141931217892e-05, 'epoch': 1.15}) +[2026-06-17 20:47:01,834][starVLA.training.train_starvla][INFO] - Step 2541, Loss: {'train/loss': 0.44296807050704956, 'train/grad_norm_pre_clip': 14.961494445800781, 'timing/data': 0.014643734320998192, 'timing/model': 1.2649744022637606, 'learning_rate/qwen_vl_interface': 7.356203690531643e-06, 'learning_rate/action_model': 7.356203690531643e-05, 'epoch': 1.15}) +[2026-06-17 20:47:05,810][starVLA.training.train_starvla][INFO] - Step 2542, Loss: {'train/loss': 0.2107827514410019, 'train/grad_norm_pre_clip': 13.68372917175293, 'timing/data': 0.008696344681084156, 'timing/model': 1.2157005788758397, 'learning_rate/qwen_vl_interface': 7.354265013227104e-06, 'learning_rate/action_model': 7.354265013227104e-05, 'epoch': 1.15}) +[2026-06-17 20:47:09,502][starVLA.training.train_starvla][INFO] - Step 2543, Loss: {'train/loss': 0.361550509929657, 'train/grad_norm_pre_clip': 13.511883735656738, 'timing/data': 0.010881947353482246, 'timing/model': 0.9461609926074743, 'learning_rate/qwen_vl_interface': 7.352325899706163e-06, 'learning_rate/action_model': 7.352325899706162e-05, 'epoch': 1.15}) +[2026-06-17 20:47:13,402][starVLA.training.train_starvla][INFO] - Step 2544, Loss: {'train/loss': 0.7415915727615356, 'train/grad_norm_pre_clip': 17.689815521240234, 'timing/data': 0.010513624176383018, 'timing/model': 0.9502100730314851, 'learning_rate/qwen_vl_interface': 7.350386350370802e-06, 'learning_rate/action_model': 7.350386350370801e-05, 'epoch': 1.15}) +[2026-06-17 20:47:17,486][starVLA.training.train_starvla][INFO] - Step 2545, Loss: {'train/loss': 0.4533115327358246, 'train/grad_norm_pre_clip': 13.851198196411133, 'timing/data': 0.00796846766024828, 'timing/model': 1.0184744857251644, 'learning_rate/qwen_vl_interface': 7.348446365623089e-06, 'learning_rate/action_model': 7.348446365623088e-05, 'epoch': 1.15}) +[2026-06-17 20:47:21,257][starVLA.training.train_starvla][INFO] - Step 2546, Loss: {'train/loss': 0.33447936177253723, 'train/grad_norm_pre_clip': 14.56357479095459, 'timing/data': 0.008682387880980968, 'timing/model': 0.9359967028722167, 'learning_rate/qwen_vl_interface': 7.346505945865186e-06, 'learning_rate/action_model': 7.346505945865186e-05, 'epoch': 1.16}) +[2026-06-17 20:47:25,300][starVLA.training.train_starvla][INFO] - Step 2547, Loss: {'train/loss': 0.33130621910095215, 'train/grad_norm_pre_clip': 13.210494995117188, 'timing/data': 0.008075999096035957, 'timing/model': 0.9527127631008625, 'learning_rate/qwen_vl_interface': 7.344565091499345e-06, 'learning_rate/action_model': 7.344565091499344e-05, 'epoch': 1.16}) +[2026-06-17 20:47:29,118][starVLA.training.train_starvla][INFO] - Step 2548, Loss: {'train/loss': 0.48258981108665466, 'train/grad_norm_pre_clip': 13.927529335021973, 'timing/data': 0.007164842449128628, 'timing/model': 1.126724517904222, 'learning_rate/qwen_vl_interface': 7.342623802927907e-06, 'learning_rate/action_model': 7.342623802927907e-05, 'epoch': 1.16}) +[2026-06-17 20:47:32,739][starVLA.training.train_starvla][INFO] - Step 2549, Loss: {'train/loss': 0.5326881408691406, 'train/grad_norm_pre_clip': 18.864912033081055, 'timing/data': 0.010838591493666172, 'timing/model': 0.9461273541674018, 'learning_rate/qwen_vl_interface': 7.340682080553303e-06, 'learning_rate/action_model': 7.340682080553303e-05, 'epoch': 1.16}) +[2026-06-17 20:47:36,655][starVLA.training.train_starvla][INFO] - Step 2550, Loss: {'train/loss': 0.2974049150943756, 'train/grad_norm_pre_clip': 13.483414649963379, 'timing/data': 0.008160177618265152, 'timing/model': 0.9487696941941977, 'learning_rate/qwen_vl_interface': 7.338739924778053e-06, 'learning_rate/action_model': 7.338739924778053e-05, 'epoch': 1.16}) +[2026-06-17 20:47:40,541][starVLA.training.train_starvla][INFO] - Step 2551, Loss: {'train/loss': 0.5546836853027344, 'train/grad_norm_pre_clip': 17.350997924804688, 'timing/data': 0.0073391953483223915, 'timing/model': 0.9281944511458278, 'learning_rate/qwen_vl_interface': 7.336797336004771e-06, 'learning_rate/action_model': 7.33679733600477e-05, 'epoch': 1.16}) +[2026-06-17 20:47:44,364][starVLA.training.train_starvla][INFO] - Step 2552, Loss: {'train/loss': 0.6476879715919495, 'train/grad_norm_pre_clip': 20.833532333374023, 'timing/data': 0.019394777715206146, 'timing/model': 0.9395763147622347, 'learning_rate/qwen_vl_interface': 7.334854314636153e-06, 'learning_rate/action_model': 7.334854314636152e-05, 'epoch': 1.16}) +[2026-06-17 20:47:48,324][starVLA.training.train_starvla][INFO] - Step 2553, Loss: {'train/loss': 0.4288121461868286, 'train/grad_norm_pre_clip': 11.5023832321167, 'timing/data': 0.015194160863757133, 'timing/model': 0.9552244283258915, 'learning_rate/qwen_vl_interface': 7.3329108610749955e-06, 'learning_rate/action_model': 7.332910861074995e-05, 'epoch': 1.16}) +[2026-06-17 20:47:52,227][starVLA.training.train_starvla][INFO] - Step 2554, Loss: {'train/loss': 0.6227725148200989, 'train/grad_norm_pre_clip': 23.730802536010742, 'timing/data': 0.010765604674816132, 'timing/model': 0.9811842860653996, 'learning_rate/qwen_vl_interface': 7.330966975724173e-06, 'learning_rate/action_model': 7.330966975724173e-05, 'epoch': 1.16}) +[2026-06-17 20:47:56,128][starVLA.training.train_starvla][INFO] - Step 2555, Loss: {'train/loss': 0.4329104721546173, 'train/grad_norm_pre_clip': 23.848848342895508, 'timing/data': 0.009825637564063072, 'timing/model': 1.1644161846488714, 'learning_rate/qwen_vl_interface': 7.329022658986657e-06, 'learning_rate/action_model': 7.329022658986658e-05, 'epoch': 1.16}) +[2026-06-17 20:47:59,729][starVLA.training.train_starvla][INFO] - Step 2556, Loss: {'train/loss': 0.5805011987686157, 'train/grad_norm_pre_clip': 15.120101928710938, 'timing/data': 0.008238933980464935, 'timing/model': 0.9374171309173107, 'learning_rate/qwen_vl_interface': 7.327077911265508e-06, 'learning_rate/action_model': 7.327077911265507e-05, 'epoch': 1.16}) +[2026-06-17 20:48:03,634][starVLA.training.train_starvla][INFO] - Step 2557, Loss: {'train/loss': 0.521490752696991, 'train/grad_norm_pre_clip': 18.1884708404541, 'timing/data': 0.009137626737356186, 'timing/model': 0.9702426819130778, 'learning_rate/qwen_vl_interface': 7.325132732963872e-06, 'learning_rate/action_model': 7.325132732963872e-05, 'epoch': 1.16}) +[2026-06-17 20:48:07,459][starVLA.training.train_starvla][INFO] - Step 2558, Loss: {'train/loss': 0.32116496562957764, 'train/grad_norm_pre_clip': 21.423189163208008, 'timing/data': 0.012396621517837048, 'timing/model': 0.937309174798429, 'learning_rate/qwen_vl_interface': 7.323187124484988e-06, 'learning_rate/action_model': 7.323187124484988e-05, 'epoch': 1.16}) +[2026-06-17 20:48:11,395][starVLA.training.train_starvla][INFO] - Step 2559, Loss: {'train/loss': 0.3662213087081909, 'train/grad_norm_pre_clip': 12.339062690734863, 'timing/data': 0.011679495684802532, 'timing/model': 1.0195171171799302, 'learning_rate/qwen_vl_interface': 7.3212410862321815e-06, 'learning_rate/action_model': 7.321241086232182e-05, 'epoch': 1.16}) +[2026-06-17 20:48:15,231][starVLA.training.train_starvla][INFO] - Step 2560, Loss: {'train/loss': 0.22678761184215546, 'train/grad_norm_pre_clip': 12.573423385620117, 'timing/data': 0.009468494914472103, 'timing/model': 0.9609039779752493, 'learning_rate/qwen_vl_interface': 7.319294618608869e-06, 'learning_rate/action_model': 7.319294618608868e-05, 'epoch': 1.16}) +[2026-06-17 20:48:19,275][starVLA.training.train_starvla][INFO] - Step 2561, Loss: {'train/loss': 0.38740697503089905, 'train/grad_norm_pre_clip': 14.553738594055176, 'timing/data': 0.01176092866808176, 'timing/model': 1.2922584628686309, 'learning_rate/qwen_vl_interface': 7.317347722018557e-06, 'learning_rate/action_model': 7.317347722018557e-05, 'epoch': 1.16}) +[2026-06-17 20:48:23,268][starVLA.training.train_starvla][INFO] - Step 2562, Loss: {'train/loss': 0.5782699584960938, 'train/grad_norm_pre_clip': 17.741744995117188, 'timing/data': 0.010864156298339367, 'timing/model': 1.145614730194211, 'learning_rate/qwen_vl_interface': 7.3154003968648355e-06, 'learning_rate/action_model': 7.315400396864835e-05, 'epoch': 1.16}) +[2026-06-17 20:48:28,948][starVLA.training.train_starvla][INFO] - Step 2563, Loss: {'train/loss': 0.43140849471092224, 'train/grad_norm_pre_clip': 13.508929252624512, 'timing/data': 0.009741249494254589, 'timing/model': 0.9386548101902008, 'learning_rate/qwen_vl_interface': 7.313452643551392e-06, 'learning_rate/action_model': 7.313452643551392e-05, 'epoch': 1.16}) +[2026-06-17 20:48:34,818][starVLA.training.train_starvla][INFO] - Step 2564, Loss: {'train/loss': 0.3966064453125, 'train/grad_norm_pre_clip': 14.273356437683105, 'timing/data': 0.0075698914006352425, 'timing/model': 0.9812516057863832, 'learning_rate/qwen_vl_interface': 7.311504462481993e-06, 'learning_rate/action_model': 7.311504462481993e-05, 'epoch': 1.16}) +[2026-06-17 20:48:38,738][starVLA.training.train_starvla][INFO] - Step 2565, Loss: {'train/loss': 0.37252745032310486, 'train/grad_norm_pre_clip': 13.632577896118164, 'timing/data': 0.009891767054796219, 'timing/model': 0.927548311650753, 'learning_rate/qwen_vl_interface': 7.3095558540605e-06, 'learning_rate/action_model': 7.309555854060499e-05, 'epoch': 1.16}) +[2026-06-17 20:48:42,822][starVLA.training.train_starvla][INFO] - Step 2566, Loss: {'train/loss': 0.30964896082878113, 'train/grad_norm_pre_clip': 16.347850799560547, 'timing/data': 0.01806824281811714, 'timing/model': 0.9542667223140597, 'learning_rate/qwen_vl_interface': 7.307606818690865e-06, 'learning_rate/action_model': 7.307606818690864e-05, 'epoch': 1.16}) +[2026-06-17 20:48:46,638][starVLA.training.train_starvla][INFO] - Step 2567, Loss: {'train/loss': 0.2304692268371582, 'train/grad_norm_pre_clip': 9.6593017578125, 'timing/data': 0.008179794065654278, 'timing/model': 0.9390547648072243, 'learning_rate/qwen_vl_interface': 7.30565735677712e-06, 'learning_rate/action_model': 7.305657356777119e-05, 'epoch': 1.16}) +[2026-06-17 20:48:50,512][starVLA.training.train_starvla][INFO] - Step 2568, Loss: {'train/loss': 0.3202958405017853, 'train/grad_norm_pre_clip': 14.801575660705566, 'timing/data': 0.012161005288362503, 'timing/model': 1.150204548612237, 'learning_rate/qwen_vl_interface': 7.303707468723395e-06, 'learning_rate/action_model': 7.303707468723395e-05, 'epoch': 1.17}) +[2026-06-17 20:48:54,358][starVLA.training.train_starvla][INFO] - Step 2569, Loss: {'train/loss': 0.31212592124938965, 'train/grad_norm_pre_clip': 13.588665008544922, 'timing/data': 0.008583830669522285, 'timing/model': 1.150044403038919, 'learning_rate/qwen_vl_interface': 7.301757154933899e-06, 'learning_rate/action_model': 7.3017571549339e-05, 'epoch': 1.17}) +[2026-06-17 20:48:58,154][starVLA.training.train_starvla][INFO] - Step 2570, Loss: {'train/loss': 0.3874534070491791, 'train/grad_norm_pre_clip': 14.01288890838623, 'timing/data': 0.01408337987959385, 'timing/model': 0.9617024306207895, 'learning_rate/qwen_vl_interface': 7.29980641581294e-06, 'learning_rate/action_model': 7.299806415812939e-05, 'epoch': 1.17}) +[2026-06-17 20:49:02,034][starVLA.training.train_starvla][INFO] - Step 2571, Loss: {'train/loss': 0.5644020438194275, 'train/grad_norm_pre_clip': 15.575037956237793, 'timing/data': 0.01290106400847435, 'timing/model': 0.9436461636796594, 'learning_rate/qwen_vl_interface': 7.297855251764903e-06, 'learning_rate/action_model': 7.297855251764903e-05, 'epoch': 1.17}) +[2026-06-17 20:49:05,859][starVLA.training.train_starvla][INFO] - Step 2572, Loss: {'train/loss': 0.3573093116283417, 'train/grad_norm_pre_clip': 19.04542350769043, 'timing/data': 0.008480901829898357, 'timing/model': 0.9569629495963454, 'learning_rate/qwen_vl_interface': 7.295903663194269e-06, 'learning_rate/action_model': 7.29590366319427e-05, 'epoch': 1.17}) +[2026-06-17 20:49:09,701][starVLA.training.train_starvla][INFO] - Step 2573, Loss: {'train/loss': 0.31765317916870117, 'train/grad_norm_pre_clip': 12.855342864990234, 'timing/data': 0.00918327085673809, 'timing/model': 0.9496648777276278, 'learning_rate/qwen_vl_interface': 7.293951650505606e-06, 'learning_rate/action_model': 7.293951650505607e-05, 'epoch': 1.17}) +[2026-06-17 20:49:14,679][starVLA.training.train_starvla][INFO] - Step 2574, Loss: {'train/loss': 0.6749927997589111, 'train/grad_norm_pre_clip': 24.185068130493164, 'timing/data': 0.01299092173576355, 'timing/model': 0.9683461217209697, 'learning_rate/qwen_vl_interface': 7.291999214103565e-06, 'learning_rate/action_model': 7.291999214103564e-05, 'epoch': 1.17}) +[2026-06-17 20:49:18,875][starVLA.training.train_starvla][INFO] - Step 2575, Loss: {'train/loss': 0.6443478465080261, 'train/grad_norm_pre_clip': 17.421489715576172, 'timing/data': 0.012576623819768429, 'timing/model': 1.379451404325664, 'learning_rate/qwen_vl_interface': 7.29004635439289e-06, 'learning_rate/action_model': 7.29004635439289e-05, 'epoch': 1.17}) +[2026-06-17 20:49:22,550][starVLA.training.train_starvla][INFO] - Step 2576, Loss: {'train/loss': 0.3311353027820587, 'train/grad_norm_pre_clip': 15.765789031982422, 'timing/data': 0.008177673444151878, 'timing/model': 0.945946172811091, 'learning_rate/qwen_vl_interface': 7.28809307177841e-06, 'learning_rate/action_model': 7.28809307177841e-05, 'epoch': 1.17}) +[2026-06-17 20:49:26,555][starVLA.training.train_starvla][INFO] - Step 2577, Loss: {'train/loss': 0.3999854624271393, 'train/grad_norm_pre_clip': 16.265350341796875, 'timing/data': 0.011492935940623283, 'timing/model': 0.9990891488268971, 'learning_rate/qwen_vl_interface': 7.286139366665044e-06, 'learning_rate/action_model': 7.286139366665043e-05, 'epoch': 1.17}) +[2026-06-17 20:49:30,347][starVLA.training.train_starvla][INFO] - Step 2578, Loss: {'train/loss': 0.517509400844574, 'train/grad_norm_pre_clip': 17.49065399169922, 'timing/data': 0.008626957423985004, 'timing/model': 0.9357254067435861, 'learning_rate/qwen_vl_interface': 7.284185239457796e-06, 'learning_rate/action_model': 7.284185239457795e-05, 'epoch': 1.17}) +[2026-06-17 20:49:34,221][starVLA.training.train_starvla][INFO] - Step 2579, Loss: {'train/loss': 0.771827757358551, 'train/grad_norm_pre_clip': 24.674753189086914, 'timing/data': 0.009986772201955318, 'timing/model': 0.9374587768688798, 'learning_rate/qwen_vl_interface': 7.282230690561759e-06, 'learning_rate/action_model': 7.28223069056176e-05, 'epoch': 1.17}) +[2026-06-17 20:49:38,068][starVLA.training.train_starvla][INFO] - Step 2580, Loss: {'train/loss': 0.31684088706970215, 'train/grad_norm_pre_clip': 13.918383598327637, 'timing/data': 0.007843793369829655, 'timing/model': 0.9373728148639202, 'learning_rate/qwen_vl_interface': 7.280275720382113e-06, 'learning_rate/action_model': 7.280275720382113e-05, 'epoch': 1.17}) +[2026-06-17 20:49:41,939][starVLA.training.train_starvla][INFO] - Step 2581, Loss: {'train/loss': 0.7468472719192505, 'train/grad_norm_pre_clip': 19.896333694458008, 'timing/data': 0.007924859412014484, 'timing/model': 1.1320709381252527, 'learning_rate/qwen_vl_interface': 7.2783203293241246e-06, 'learning_rate/action_model': 7.278320329324125e-05, 'epoch': 1.17}) +[2026-06-17 20:49:46,018][starVLA.training.train_starvla][INFO] - Step 2582, Loss: {'train/loss': 0.5127719640731812, 'train/grad_norm_pre_clip': 17.76857566833496, 'timing/data': 0.008930249139666557, 'timing/model': 1.379302498884499, 'learning_rate/qwen_vl_interface': 7.276364517793152e-06, 'learning_rate/action_model': 7.276364517793151e-05, 'epoch': 1.17}) +[2026-06-17 20:49:49,658][starVLA.training.train_starvla][INFO] - Step 2583, Loss: {'train/loss': 0.34166860580444336, 'train/grad_norm_pre_clip': 13.2316255569458, 'timing/data': 0.010705466382205486, 'timing/model': 0.9385185269638896, 'learning_rate/qwen_vl_interface': 7.274408286194631e-06, 'learning_rate/action_model': 7.27440828619463e-05, 'epoch': 1.17}) +[2026-06-17 20:49:53,581][starVLA.training.train_starvla][INFO] - Step 2584, Loss: {'train/loss': 0.2536633610725403, 'train/grad_norm_pre_clip': 10.062143325805664, 'timing/data': 0.00930403359234333, 'timing/model': 0.9603287372738123, 'learning_rate/qwen_vl_interface': 7.272451634934096e-06, 'learning_rate/action_model': 7.272451634934096e-05, 'epoch': 1.17}) +[2026-06-17 20:49:57,611][starVLA.training.train_starvla][INFO] - Step 2585, Loss: {'train/loss': 0.3702196180820465, 'train/grad_norm_pre_clip': 16.770442962646484, 'timing/data': 0.008787905797362328, 'timing/model': 0.9476896738633513, 'learning_rate/qwen_vl_interface': 7.270494564417161e-06, 'learning_rate/action_model': 7.270494564417161e-05, 'epoch': 1.17}) +[2026-06-17 20:50:01,552][starVLA.training.train_starvla][INFO] - Step 2586, Loss: {'train/loss': 0.3490942716598511, 'train/grad_norm_pre_clip': 11.142555236816406, 'timing/data': 0.009402569383382797, 'timing/model': 0.9937457293272018, 'learning_rate/qwen_vl_interface': 7.268537075049527e-06, 'learning_rate/action_model': 7.268537075049526e-05, 'epoch': 1.17}) +[2026-06-17 20:50:05,646][starVLA.training.train_starvla][INFO] - Step 2587, Loss: {'train/loss': 0.2498353123664856, 'train/grad_norm_pre_clip': 10.817687034606934, 'timing/data': 0.009311782196164131, 'timing/model': 0.9460316272452474, 'learning_rate/qwen_vl_interface': 7.2665791672369875e-06, 'learning_rate/action_model': 7.266579167236988e-05, 'epoch': 1.17}) +[2026-06-17 20:50:09,602][starVLA.training.train_starvla][INFO] - Step 2588, Loss: {'train/loss': 0.2329678237438202, 'train/grad_norm_pre_clip': 9.130581855773926, 'timing/data': 0.013850326649844646, 'timing/model': 1.2172149326652288, 'learning_rate/qwen_vl_interface': 7.264620841385415e-06, 'learning_rate/action_model': 7.264620841385415e-05, 'epoch': 1.17}) +[2026-06-17 20:50:13,523][starVLA.training.train_starvla][INFO] - Step 2589, Loss: {'train/loss': 0.31715768575668335, 'train/grad_norm_pre_clip': 12.066630363464355, 'timing/data': 0.010442242957651615, 'timing/model': 1.143018082715571, 'learning_rate/qwen_vl_interface': 7.262662097900775e-06, 'learning_rate/action_model': 7.262662097900774e-05, 'epoch': 1.17}) +[2026-06-17 20:50:17,153][starVLA.training.train_starvla][INFO] - Step 2590, Loss: {'train/loss': 0.3128471374511719, 'train/grad_norm_pre_clip': 21.277698516845703, 'timing/data': 0.010299231857061386, 'timing/model': 0.933099739253521, 'learning_rate/qwen_vl_interface': 7.260702937189116e-06, 'learning_rate/action_model': 7.260702937189115e-05, 'epoch': 1.18}) +[2026-06-17 20:50:21,016][starVLA.training.train_starvla][INFO] - Step 2591, Loss: {'train/loss': 0.2743949592113495, 'train/grad_norm_pre_clip': 12.880898475646973, 'timing/data': 0.013078357093036175, 'timing/model': 0.9354315884411335, 'learning_rate/qwen_vl_interface': 7.258743359656574e-06, 'learning_rate/action_model': 7.258743359656573e-05, 'epoch': 1.18}) +[2026-06-17 20:50:24,882][starVLA.training.train_starvla][INFO] - Step 2592, Loss: {'train/loss': 0.4196065068244934, 'train/grad_norm_pre_clip': 15.217171669006348, 'timing/data': 0.014873835258185863, 'timing/model': 0.9322871686890721, 'learning_rate/qwen_vl_interface': 7.2567833657093725e-06, 'learning_rate/action_model': 7.256783365709371e-05, 'epoch': 1.18}) +[2026-06-17 20:50:28,714][starVLA.training.train_starvla][INFO] - Step 2593, Loss: {'train/loss': 0.25320208072662354, 'train/grad_norm_pre_clip': 11.759306907653809, 'timing/data': 0.010012084618210793, 'timing/model': 0.9511062307283282, 'learning_rate/qwen_vl_interface': 7.254822955753821e-06, 'learning_rate/action_model': 7.25482295575382e-05, 'epoch': 1.18}) +[2026-06-17 20:50:32,991][starVLA.training.train_starvla][INFO] - Step 2594, Loss: {'train/loss': 0.4111156165599823, 'train/grad_norm_pre_clip': 24.477102279663086, 'timing/data': 0.010306118056178093, 'timing/model': 0.9600009694695473, 'learning_rate/qwen_vl_interface': 7.2528621301963135e-06, 'learning_rate/action_model': 7.252862130196313e-05, 'epoch': 1.18}) +[2026-06-17 20:50:37,133][starVLA.training.train_starvla][INFO] - Step 2595, Loss: {'train/loss': 0.6236850023269653, 'train/grad_norm_pre_clip': 14.112405776977539, 'timing/data': 0.009640797041356564, 'timing/model': 1.4457949241623282, 'learning_rate/qwen_vl_interface': 7.25090088944333e-06, 'learning_rate/action_model': 7.250900889443329e-05, 'epoch': 1.18}) +[2026-06-17 20:50:40,804][starVLA.training.train_starvla][INFO] - Step 2596, Loss: {'train/loss': 0.44646522402763367, 'train/grad_norm_pre_clip': 15.20776081085205, 'timing/data': 0.009681598283350468, 'timing/model': 0.9496314264833927, 'learning_rate/qwen_vl_interface': 7.248939233901438e-06, 'learning_rate/action_model': 7.248939233901439e-05, 'epoch': 1.18}) +[2026-06-17 20:50:44,642][starVLA.training.train_starvla][INFO] - Step 2597, Loss: {'train/loss': 0.6156591176986694, 'train/grad_norm_pre_clip': 23.382192611694336, 'timing/data': 0.017024733126163483, 'timing/model': 0.9370707925409079, 'learning_rate/qwen_vl_interface': 7.2469771639772944e-06, 'learning_rate/action_model': 7.246977163977294e-05, 'epoch': 1.18}) +[2026-06-17 20:50:48,441][starVLA.training.train_starvla][INFO] - Step 2598, Loss: {'train/loss': 0.4860893189907074, 'train/grad_norm_pre_clip': 20.321624755859375, 'timing/data': 0.01372453011572361, 'timing/model': 0.9490376738831401, 'learning_rate/qwen_vl_interface': 7.245014680077633e-06, 'learning_rate/action_model': 7.245014680077633e-05, 'epoch': 1.18}) +[2026-06-17 20:50:52,366][starVLA.training.train_starvla][INFO] - Step 2599, Loss: {'train/loss': 0.15185114741325378, 'train/grad_norm_pre_clip': 9.531331062316895, 'timing/data': 0.00822445098310709, 'timing/model': 0.9719857890158892, 'learning_rate/qwen_vl_interface': 7.243051782609285e-06, 'learning_rate/action_model': 7.243051782609285e-05, 'epoch': 1.18}) +[2026-06-17 20:53:07,613][starVLA.training.train_starvla][INFO] - Step 2600, Loss: {'train/loss': 0.4881960153579712, 'train/grad_norm_pre_clip': 22.02237319946289, 'eval/loss': 0.39800652861595154, 'eval/latency_8/loss': 0.39800652861595154, 'eval/demon_fire_precision': 0.9468277945619336, 'eval/demon_fire_recall': 0.9735942839391115, 'eval/demon_fire_f1': 0.9600245060499311, 'eval/demon_fire_rate_model': 0.9209794101279911, 'eval/demon_fire_rate_teacher': 0.8956594323873122, 'eval/demon_left_precision': 0.8174157303370787, 'eval/demon_left_recall': 0.6231263383297645, 'eval/demon_left_f1': 0.707168894289186, 'eval/demon_left_rate_model': 0.09905397885364496, 'eval/demon_left_rate_teacher': 0.12993878686700056, 'eval/demon_right_precision': 0.8375451263537906, 'eval/demon_right_recall': 0.703030303030303, 'eval/demon_right_f1': 0.7644151565074135, 'eval/demon_right_rate_model': 0.15414579855314414, 'eval/demon_right_rate_teacher': 0.18363939899833054, 'eval/demon_move_f1': 0.7357920253982997, 'eval/demon_cc_f1': 0.8796065290167988, 'eval/latency_8/demon_fire_precision': 0.9468277945619336, 'eval/latency_8/demon_fire_recall': 0.9735942839391115, 'eval/latency_8/demon_fire_f1': 0.9600245060499311, 'eval/latency_8/demon_fire_rate_model': 0.9209794101279911, 'eval/latency_8/demon_fire_rate_teacher': 0.8956594323873122, 'eval/latency_8/demon_left_precision': 0.8174157303370787, 'eval/latency_8/demon_left_recall': 0.6231263383297645, 'eval/latency_8/demon_left_f1': 0.707168894289186, 'eval/latency_8/demon_left_rate_model': 0.09905397885364496, 'eval/latency_8/demon_left_rate_teacher': 0.12993878686700056, 'eval/latency_8/demon_right_precision': 0.8375451263537906, 'eval/latency_8/demon_right_recall': 0.703030303030303, 'eval/latency_8/demon_right_f1': 0.7644151565074135, 'eval/latency_8/demon_right_rate_model': 0.15414579855314414, 'eval/latency_8/demon_right_rate_teacher': 0.18363939899833054, 'eval/latency_8/demon_move_f1': 0.7357920253982997, 'eval/latency_8/demon_cc_f1': 0.8796065290167988, 'timing/data': 0.01378275640308857, 'timing/model': 0.9721745047718287, 'learning_rate/qwen_vl_interface': 7.241088471979155e-06, 'learning_rate/action_model': 7.241088471979155e-05, 'epoch': 1.18}) +[2026-06-17 20:53:07,616][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_2600_state +[2026-06-17 20:53:07,621][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 20:53:14,300][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_2600_state/model.safetensors +[2026-06-17 20:53:29,648][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_2600_state/optimizer.bin +[2026-06-17 20:53:29,649][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_2600_state/sampler.bin +[2026-06-17 20:53:29,650][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_2600_state/sampler_1.bin +[2026-06-17 20:53:29,652][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_2600_state/random_states_0.pkl +[2026-06-17 20:53:29,659][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2600 +[2026-06-17 20:53:29,659][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 20:53:29,670][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 20:53:33,605][starVLA.training.train_starvla][INFO] - Step 2601, Loss: {'train/loss': 0.29396891593933105, 'train/grad_norm_pre_clip': 15.832700729370117, 'timing/data': 0.007925847545266151, 'timing/model': 0.9331092201173306, 'learning_rate/qwen_vl_interface': 7.239124748594244e-06, 'learning_rate/action_model': 7.239124748594244e-05, 'epoch': 1.18}) +[2026-06-17 20:53:37,600][starVLA.training.train_starvla][INFO] - Step 2602, Loss: {'train/loss': 0.7623154520988464, 'train/grad_norm_pre_clip': 21.31593132019043, 'timing/data': 0.012217936106026173, 'timing/model': 1.087131884880364, 'learning_rate/qwen_vl_interface': 7.237160612861632e-06, 'learning_rate/action_model': 7.237160612861632e-05, 'epoch': 1.18}) +[2026-06-17 20:53:41,425][starVLA.training.train_starvla][INFO] - Step 2603, Loss: {'train/loss': 0.5611968636512756, 'train/grad_norm_pre_clip': 31.20338249206543, 'timing/data': 0.00994555838406086, 'timing/model': 0.945642801001668, 'learning_rate/qwen_vl_interface': 7.235196065188485e-06, 'learning_rate/action_model': 7.235196065188485e-05, 'epoch': 1.18}) +[2026-06-17 20:53:45,208][starVLA.training.train_starvla][INFO] - Step 2604, Loss: {'train/loss': 0.5644246935844421, 'train/grad_norm_pre_clip': 21.009716033935547, 'timing/data': 0.008283690549433231, 'timing/model': 1.108264091424644, 'learning_rate/qwen_vl_interface': 7.23323110598206e-06, 'learning_rate/action_model': 7.23323110598206e-05, 'epoch': 1.18}) +[2026-06-17 20:53:48,871][starVLA.training.train_starvla][INFO] - Step 2605, Loss: {'train/loss': 0.41997504234313965, 'train/grad_norm_pre_clip': 17.979053497314453, 'timing/data': 0.013909226283431053, 'timing/model': 0.9384678974747658, 'learning_rate/qwen_vl_interface': 7.231265735649691e-06, 'learning_rate/action_model': 7.231265735649691e-05, 'epoch': 1.18}) +[2026-06-17 20:53:52,743][starVLA.training.train_starvla][INFO] - Step 2606, Loss: {'train/loss': 0.19967693090438843, 'train/grad_norm_pre_clip': 12.001269340515137, 'timing/data': 0.008964348584413528, 'timing/model': 0.9801586084067822, 'learning_rate/qwen_vl_interface': 7.229299954598805e-06, 'learning_rate/action_model': 7.229299954598804e-05, 'epoch': 1.18}) +[2026-06-17 20:53:56,550][starVLA.training.train_starvla][INFO] - Step 2607, Loss: {'train/loss': 0.252286434173584, 'train/grad_norm_pre_clip': 11.80523681640625, 'timing/data': 0.008011063560843468, 'timing/model': 0.9491534791886806, 'learning_rate/qwen_vl_interface': 7.227333763236907e-06, 'learning_rate/action_model': 7.227333763236906e-05, 'epoch': 1.18}) +[2026-06-17 20:54:00,333][starVLA.training.train_starvla][INFO] - Step 2608, Loss: {'train/loss': 0.37895670533180237, 'train/grad_norm_pre_clip': 13.715948104858398, 'timing/data': 0.007990545593202114, 'timing/model': 0.9341947436332703, 'learning_rate/qwen_vl_interface': 7.225367161971592e-06, 'learning_rate/action_model': 7.225367161971591e-05, 'epoch': 1.18}) +[2026-06-17 20:54:04,241][starVLA.training.train_starvla][INFO] - Step 2609, Loss: {'train/loss': 0.5513231158256531, 'train/grad_norm_pre_clip': 26.700634002685547, 'timing/data': 0.009798189625144005, 'timing/model': 0.9386118641123176, 'learning_rate/qwen_vl_interface': 7.223400151210539e-06, 'learning_rate/action_model': 7.223400151210538e-05, 'epoch': 1.18}) +[2026-06-17 20:54:08,217][starVLA.training.train_starvla][INFO] - Step 2610, Loss: {'train/loss': 0.2379262000322342, 'train/grad_norm_pre_clip': 16.049087524414062, 'timing/data': 0.01018297579139471, 'timing/model': 1.2578121479600668, 'learning_rate/qwen_vl_interface': 7.2214327313615105e-06, 'learning_rate/action_model': 7.22143273136151e-05, 'epoch': 1.18}) +[2026-06-17 20:54:12,013][starVLA.training.train_starvla][INFO] - Step 2611, Loss: {'train/loss': 0.4865492582321167, 'train/grad_norm_pre_clip': 17.999591827392578, 'timing/data': 0.008171855472028255, 'timing/model': 1.1109705297276378, 'learning_rate/qwen_vl_interface': 7.219464902832357e-06, 'learning_rate/action_model': 7.219464902832356e-05, 'epoch': 1.18}) +[2026-06-17 20:54:15,634][starVLA.training.train_starvla][INFO] - Step 2612, Loss: {'train/loss': 0.6308655738830566, 'train/grad_norm_pre_clip': 21.601898193359375, 'timing/data': 0.009384168311953545, 'timing/model': 0.9457605043426156, 'learning_rate/qwen_vl_interface': 7.2174966660310085e-06, 'learning_rate/action_model': 7.217496666031008e-05, 'epoch': 1.18}) +[2026-06-17 20:54:19,437][starVLA.training.train_starvla][INFO] - Step 2613, Loss: {'train/loss': 0.38002657890319824, 'train/grad_norm_pre_clip': 16.86375617980957, 'timing/data': 0.011963658034801483, 'timing/model': 0.9493592446669936, 'learning_rate/qwen_vl_interface': 7.215528021365487e-06, 'learning_rate/action_model': 7.215528021365487e-05, 'epoch': 1.19}) +[2026-06-17 20:54:23,290][starVLA.training.train_starvla][INFO] - Step 2614, Loss: {'train/loss': 0.6134620904922485, 'train/grad_norm_pre_clip': 19.753646850585938, 'timing/data': 0.010263563133776188, 'timing/model': 0.9513774644583464, 'learning_rate/qwen_vl_interface': 7.213558969243888e-06, 'learning_rate/action_model': 7.213558969243888e-05, 'epoch': 1.19}) +[2026-06-17 20:54:27,174][starVLA.training.train_starvla][INFO] - Step 2615, Loss: {'train/loss': 0.2958911955356598, 'train/grad_norm_pre_clip': 14.358506202697754, 'timing/data': 0.01380093116313219, 'timing/model': 0.9407423259690404, 'learning_rate/qwen_vl_interface': 7.211589510074404e-06, 'learning_rate/action_model': 7.211589510074404e-05, 'epoch': 1.19}) +[2026-06-17 20:54:31,002][starVLA.training.train_starvla][INFO] - Step 2616, Loss: {'train/loss': 0.21947868168354034, 'train/grad_norm_pre_clip': 10.791391372680664, 'timing/data': 0.020273416303098202, 'timing/model': 0.9428512481972575, 'learning_rate/qwen_vl_interface': 7.209619644265302e-06, 'learning_rate/action_model': 7.209619644265302e-05, 'epoch': 1.19}) +[2026-06-17 20:54:34,812][starVLA.training.train_starvla][INFO] - Step 2617, Loss: {'train/loss': 0.3907173275947571, 'train/grad_norm_pre_clip': 15.284072875976562, 'timing/data': 0.00964177493005991, 'timing/model': 1.1169808087870479, 'learning_rate/qwen_vl_interface': 7.20764937222494e-06, 'learning_rate/action_model': 7.20764937222494e-05, 'epoch': 1.19}) +[2026-06-17 20:54:38,476][starVLA.training.train_starvla][INFO] - Step 2618, Loss: {'train/loss': 0.29840755462646484, 'train/grad_norm_pre_clip': 11.85009765625, 'timing/data': 0.00943840853869915, 'timing/model': 0.9425139864906669, 'learning_rate/qwen_vl_interface': 7.205678694361757e-06, 'learning_rate/action_model': 7.205678694361757e-05, 'epoch': 1.19}) +[2026-06-17 20:54:42,385][starVLA.training.train_starvla][INFO] - Step 2619, Loss: {'train/loss': 0.23837676644325256, 'train/grad_norm_pre_clip': 12.218350410461426, 'timing/data': 0.00938450824469328, 'timing/model': 0.9418306536972523, 'learning_rate/qwen_vl_interface': 7.203707611084276e-06, 'learning_rate/action_model': 7.203707611084275e-05, 'epoch': 1.19}) +[2026-06-17 20:54:46,304][starVLA.training.train_starvla][INFO] - Step 2620, Loss: {'train/loss': 0.41470691561698914, 'train/grad_norm_pre_clip': 17.62226676940918, 'timing/data': 0.009310876950621605, 'timing/model': 0.9449629792943597, 'learning_rate/qwen_vl_interface': 7.2017361228011055e-06, 'learning_rate/action_model': 7.201736122801104e-05, 'epoch': 1.19}) +[2026-06-17 20:54:50,139][starVLA.training.train_starvla][INFO] - Step 2621, Loss: {'train/loss': 0.4052158296108246, 'train/grad_norm_pre_clip': 14.884828567504883, 'timing/data': 0.009307553991675377, 'timing/model': 0.9515467416495085, 'learning_rate/qwen_vl_interface': 7.199764229920935e-06, 'learning_rate/action_model': 7.199764229920934e-05, 'epoch': 1.19}) +[2026-06-17 20:54:53,959][starVLA.training.train_starvla][INFO] - Step 2622, Loss: {'train/loss': 0.6290794014930725, 'train/grad_norm_pre_clip': 19.380138397216797, 'timing/data': 0.013034666888415813, 'timing/model': 0.9414740083739161, 'learning_rate/qwen_vl_interface': 7.197791932852543e-06, 'learning_rate/action_model': 7.197791932852542e-05, 'epoch': 1.19}) +[2026-06-17 20:54:58,033][starVLA.training.train_starvla][INFO] - Step 2623, Loss: {'train/loss': 0.20679298043251038, 'train/grad_norm_pre_clip': 9.454602241516113, 'timing/data': 0.010887700133025646, 'timing/model': 1.3427843637764454, 'learning_rate/qwen_vl_interface': 7.195819232004787e-06, 'learning_rate/action_model': 7.195819232004787e-05, 'epoch': 1.19}) +[2026-06-17 20:55:01,895][starVLA.training.train_starvla][INFO] - Step 2624, Loss: {'train/loss': 0.524782657623291, 'train/grad_norm_pre_clip': 15.84988784790039, 'timing/data': 0.016084574162960052, 'timing/model': 1.158351176418364, 'learning_rate/qwen_vl_interface': 7.1938461277866105e-06, 'learning_rate/action_model': 7.19384612778661e-05, 'epoch': 1.19}) +[2026-06-17 20:55:05,541][starVLA.training.train_starvla][INFO] - Step 2625, Loss: {'train/loss': 0.4373971223831177, 'train/grad_norm_pre_clip': 18.60487937927246, 'timing/data': 0.011172176338732243, 'timing/model': 0.9503930872306228, 'learning_rate/qwen_vl_interface': 7.19187262060704e-06, 'learning_rate/action_model': 7.19187262060704e-05, 'epoch': 1.19}) +[2026-06-17 20:55:09,543][starVLA.training.train_starvla][INFO] - Step 2626, Loss: {'train/loss': 0.5106409788131714, 'train/grad_norm_pre_clip': 18.465675354003906, 'timing/data': 0.010491832159459591, 'timing/model': 0.9497080752626061, 'learning_rate/qwen_vl_interface': 7.189898710875185e-06, 'learning_rate/action_model': 7.189898710875184e-05, 'epoch': 1.19}) +[2026-06-17 20:55:13,503][starVLA.training.train_starvla][INFO] - Step 2627, Loss: {'train/loss': 0.2941179871559143, 'train/grad_norm_pre_clip': 11.479667663574219, 'timing/data': 0.00993666797876358, 'timing/model': 1.0240673748776317, 'learning_rate/qwen_vl_interface': 7.187924399000241e-06, 'learning_rate/action_model': 7.18792439900024e-05, 'epoch': 1.19}) +[2026-06-17 20:55:17,312][starVLA.training.train_starvla][INFO] - Step 2628, Loss: {'train/loss': 0.30714184045791626, 'train/grad_norm_pre_clip': 12.045929908752441, 'timing/data': 0.008617331273853779, 'timing/model': 0.929809377528727, 'learning_rate/qwen_vl_interface': 7.185949685391483e-06, 'learning_rate/action_model': 7.185949685391483e-05, 'epoch': 1.19}) +[2026-06-17 20:55:21,144][starVLA.training.train_starvla][INFO] - Step 2629, Loss: {'train/loss': 0.37014201283454895, 'train/grad_norm_pre_clip': 14.17267894744873, 'timing/data': 0.011697419919073582, 'timing/model': 0.9372459342703223, 'learning_rate/qwen_vl_interface': 7.183974570458272e-06, 'learning_rate/action_model': 7.183974570458272e-05, 'epoch': 1.19}) +[2026-06-17 20:55:24,992][starVLA.training.train_starvla][INFO] - Step 2630, Loss: {'train/loss': 0.5937412977218628, 'train/grad_norm_pre_clip': 20.615779876708984, 'timing/data': 0.010956201702356339, 'timing/model': 1.1311765052378178, 'learning_rate/qwen_vl_interface': 7.181999054610053e-06, 'learning_rate/action_model': 7.181999054610052e-05, 'epoch': 1.19}) +[2026-06-17 20:55:28,825][starVLA.training.train_starvla][INFO] - Step 2631, Loss: {'train/loss': 0.5204837918281555, 'train/grad_norm_pre_clip': 19.908655166625977, 'timing/data': 0.009961985051631927, 'timing/model': 1.1085308687761426, 'learning_rate/qwen_vl_interface': 7.180023138256349e-06, 'learning_rate/action_model': 7.180023138256348e-05, 'epoch': 1.19}) +[2026-06-17 20:55:32,493][starVLA.training.train_starvla][INFO] - Step 2632, Loss: {'train/loss': 0.553165853023529, 'train/grad_norm_pre_clip': 20.80644989013672, 'timing/data': 0.011308261193335056, 'timing/model': 0.9464616058394313, 'learning_rate/qwen_vl_interface': 7.178046821806771e-06, 'learning_rate/action_model': 7.17804682180677e-05, 'epoch': 1.19}) +[2026-06-17 20:55:36,439][starVLA.training.train_starvla][INFO] - Step 2633, Loss: {'train/loss': 0.7127493023872375, 'train/grad_norm_pre_clip': 20.0247745513916, 'timing/data': 0.008262262679636478, 'timing/model': 0.9327081888914108, 'learning_rate/qwen_vl_interface': 7.176070105671016e-06, 'learning_rate/action_model': 7.176070105671015e-05, 'epoch': 1.19}) +[2026-06-17 20:55:40,428][starVLA.training.train_starvla][INFO] - Step 2634, Loss: {'train/loss': 0.23223835229873657, 'train/grad_norm_pre_clip': 8.88884449005127, 'timing/data': 0.011356023140251637, 'timing/model': 0.9285532310605049, 'learning_rate/qwen_vl_interface': 7.174092990258852e-06, 'learning_rate/action_model': 7.174092990258851e-05, 'epoch': 1.19}) +[2026-06-17 20:55:44,165][starVLA.training.train_starvla][INFO] - Step 2635, Loss: {'train/loss': 0.6792044043540955, 'train/grad_norm_pre_clip': 17.571800231933594, 'timing/data': 0.010671048425137997, 'timing/model': 0.9233152875676751, 'learning_rate/qwen_vl_interface': 7.172115475980142e-06, 'learning_rate/action_model': 7.172115475980142e-05, 'epoch': 1.2}) +[2026-06-17 20:55:48,229][starVLA.training.train_starvla][INFO] - Step 2636, Loss: {'train/loss': 0.24568872153759003, 'train/grad_norm_pre_clip': 10.290848731994629, 'timing/data': 0.008043062873184681, 'timing/model': 0.9520048955455422, 'learning_rate/qwen_vl_interface': 7.170137563244825e-06, 'learning_rate/action_model': 7.170137563244824e-05, 'epoch': 1.2}) +[2026-06-17 20:55:52,061][starVLA.training.train_starvla][INFO] - Step 2637, Loss: {'train/loss': 0.3875611424446106, 'train/grad_norm_pre_clip': 15.172957420349121, 'timing/data': 0.007566937245428562, 'timing/model': 1.1233359221369028, 'learning_rate/qwen_vl_interface': 7.168159252462927e-06, 'learning_rate/action_model': 7.168159252462926e-05, 'epoch': 1.2}) +[2026-06-17 20:55:55,743][starVLA.training.train_starvla][INFO] - Step 2638, Loss: {'train/loss': 0.26403358578681946, 'train/grad_norm_pre_clip': 10.332856178283691, 'timing/data': 0.008397042751312256, 'timing/model': 0.9380912706255913, 'learning_rate/qwen_vl_interface': 7.166180544044549e-06, 'learning_rate/action_model': 7.166180544044549e-05, 'epoch': 1.2}) +[2026-06-17 20:55:59,616][starVLA.training.train_starvla][INFO] - Step 2639, Loss: {'train/loss': 0.7732899785041809, 'train/grad_norm_pre_clip': 18.72027587890625, 'timing/data': 0.011777560226619244, 'timing/model': 0.9580056127160788, 'learning_rate/qwen_vl_interface': 7.164201438399884e-06, 'learning_rate/action_model': 7.164201438399884e-05, 'epoch': 1.2}) +[2026-06-17 20:56:03,446][starVLA.training.train_starvla][INFO] - Step 2640, Loss: {'train/loss': 0.42382487654685974, 'train/grad_norm_pre_clip': 14.46104907989502, 'timing/data': 0.009550000540912151, 'timing/model': 0.9662344455718994, 'learning_rate/qwen_vl_interface': 7.1622219359392e-06, 'learning_rate/action_model': 7.162221935939199e-05, 'epoch': 1.2}) +[2026-06-17 20:56:07,366][starVLA.training.train_starvla][INFO] - Step 2641, Loss: {'train/loss': 0.3518582284450531, 'train/grad_norm_pre_clip': 15.794332504272461, 'timing/data': 0.007575824856758118, 'timing/model': 0.9656416298821568, 'learning_rate/qwen_vl_interface': 7.160242037072852e-06, 'learning_rate/action_model': 7.160242037072851e-05, 'epoch': 1.2}) +[2026-06-17 20:56:11,314][starVLA.training.train_starvla][INFO] - Step 2642, Loss: {'train/loss': 0.15144386887550354, 'train/grad_norm_pre_clip': 8.702126502990723, 'timing/data': 0.016329679638147354, 'timing/model': 0.9511053152382374, 'learning_rate/qwen_vl_interface': 7.158261742211275e-06, 'learning_rate/action_model': 7.158261742211274e-05, 'epoch': 1.2}) +[2026-06-17 20:56:15,145][starVLA.training.train_starvla][INFO] - Step 2643, Loss: {'train/loss': 0.46963876485824585, 'train/grad_norm_pre_clip': 14.667752265930176, 'timing/data': 0.007267482578754425, 'timing/model': 1.0966730462387204, 'learning_rate/qwen_vl_interface': 7.156281051764983e-06, 'learning_rate/action_model': 7.156281051764983e-05, 'epoch': 1.2}) +[2026-06-17 20:56:18,966][starVLA.training.train_starvla][INFO] - Step 2644, Loss: {'train/loss': 0.3927454948425293, 'train/grad_norm_pre_clip': 25.154464721679688, 'timing/data': 0.016108611598610878, 'timing/model': 1.1065730229020119, 'learning_rate/qwen_vl_interface': 7.15429996614458e-06, 'learning_rate/action_model': 7.15429996614458e-05, 'epoch': 1.2}) +[2026-06-17 20:56:22,734][starVLA.training.train_starvla][INFO] - Step 2645, Loss: {'train/loss': 0.19168734550476074, 'train/grad_norm_pre_clip': 11.624953269958496, 'timing/data': 0.007746962830424309, 'timing/model': 0.9518139269202948, 'learning_rate/qwen_vl_interface': 7.152318485760743e-06, 'learning_rate/action_model': 7.152318485760742e-05, 'epoch': 1.2}) +[2026-06-17 20:56:26,540][starVLA.training.train_starvla][INFO] - Step 2646, Loss: {'train/loss': 0.4205515682697296, 'train/grad_norm_pre_clip': 17.3750057220459, 'timing/data': 0.0074251629412174225, 'timing/model': 0.9301724378019571, 'learning_rate/qwen_vl_interface': 7.150336611024239e-06, 'learning_rate/action_model': 7.150336611024239e-05, 'epoch': 1.2}) +[2026-06-17 20:56:30,635][starVLA.training.train_starvla][INFO] - Step 2647, Loss: {'train/loss': 0.7290677428245544, 'train/grad_norm_pre_clip': 22.159881591796875, 'timing/data': 0.017868418246507645, 'timing/model': 0.9508991679176688, 'learning_rate/qwen_vl_interface': 7.1483543423459086e-06, 'learning_rate/action_model': 7.148354342345908e-05, 'epoch': 1.2}) +[2026-06-17 20:56:34,439][starVLA.training.train_starvla][INFO] - Step 2648, Loss: {'train/loss': 0.4344717860221863, 'train/grad_norm_pre_clip': 15.612293243408203, 'timing/data': 0.009201212786138058, 'timing/model': 0.9418955296278, 'learning_rate/qwen_vl_interface': 7.146371680136682e-06, 'learning_rate/action_model': 7.146371680136681e-05, 'epoch': 1.2}) +[2026-06-17 20:56:38,320][starVLA.training.train_starvla][INFO] - Step 2649, Loss: {'train/loss': 0.2276919037103653, 'train/grad_norm_pre_clip': 11.399773597717285, 'timing/data': 0.01569301076233387, 'timing/model': 0.9505548262968659, 'learning_rate/qwen_vl_interface': 7.144388624807565e-06, 'learning_rate/action_model': 7.144388624807566e-05, 'epoch': 1.2}) +[2026-06-17 20:56:42,114][starVLA.training.train_starvla][INFO] - Step 2650, Loss: {'train/loss': 0.21279238164424896, 'train/grad_norm_pre_clip': 12.711929321289062, 'timing/data': 0.009281759150326252, 'timing/model': 1.1220850134268403, 'learning_rate/qwen_vl_interface': 7.142405176769648e-06, 'learning_rate/action_model': 7.142405176769648e-05, 'epoch': 1.2}) +[2026-06-17 20:56:45,905][starVLA.training.train_starvla][INFO] - Step 2651, Loss: {'train/loss': 0.5274242162704468, 'train/grad_norm_pre_clip': 18.523014068603516, 'timing/data': 0.007536198943853378, 'timing/model': 1.1087737819179893, 'learning_rate/qwen_vl_interface': 7.140421336434102e-06, 'learning_rate/action_model': 7.140421336434102e-05, 'epoch': 1.2}) +[2026-06-17 20:56:49,579][starVLA.training.train_starvla][INFO] - Step 2652, Loss: {'train/loss': 0.6403723955154419, 'train/grad_norm_pre_clip': 22.460458755493164, 'timing/data': 0.009124241769313812, 'timing/model': 0.9961702385917306, 'learning_rate/qwen_vl_interface': 7.138437104212176e-06, 'learning_rate/action_model': 7.138437104212176e-05, 'epoch': 1.2}) +[2026-06-17 20:56:53,415][starVLA.training.train_starvla][INFO] - Step 2653, Loss: {'train/loss': 0.3502233028411865, 'train/grad_norm_pre_clip': 16.341325759887695, 'timing/data': 0.007750318385660648, 'timing/model': 0.9485495332628489, 'learning_rate/qwen_vl_interface': 7.136452480515208e-06, 'learning_rate/action_model': 7.136452480515208e-05, 'epoch': 1.2}) +[2026-06-17 20:56:57,249][starVLA.training.train_starvla][INFO] - Step 2654, Loss: {'train/loss': 0.6163628101348877, 'train/grad_norm_pre_clip': 19.5893497467041, 'timing/data': 0.011615577153861523, 'timing/model': 0.9389619482681155, 'learning_rate/qwen_vl_interface': 7.134467465754612e-06, 'learning_rate/action_model': 7.134467465754612e-05, 'epoch': 1.2}) +[2026-06-17 20:57:01,006][starVLA.training.train_starvla][INFO] - Step 2655, Loss: {'train/loss': 0.7306153178215027, 'train/grad_norm_pre_clip': 26.645069122314453, 'timing/data': 0.007700455375015736, 'timing/model': 0.9468654552474618, 'learning_rate/qwen_vl_interface': 7.132482060341881e-06, 'learning_rate/action_model': 7.13248206034188e-05, 'epoch': 1.2}) +[2026-06-17 20:57:04,909][starVLA.training.train_starvla][INFO] - Step 2656, Loss: {'train/loss': 0.5518364310264587, 'train/grad_norm_pre_clip': 20.0512752532959, 'timing/data': 0.008017662912607193, 'timing/model': 1.0407022014260292, 'learning_rate/qwen_vl_interface': 7.1304962646885935e-06, 'learning_rate/action_model': 7.130496264688593e-05, 'epoch': 1.2}) +[2026-06-17 20:57:08,714][starVLA.training.train_starvla][INFO] - Step 2657, Loss: {'train/loss': 0.4569632411003113, 'train/grad_norm_pre_clip': 17.83800506591797, 'timing/data': 0.00967705063521862, 'timing/model': 1.124205886386335, 'learning_rate/qwen_vl_interface': 7.128510079206408e-06, 'learning_rate/action_model': 7.128510079206407e-05, 'epoch': 1.21}) +[2026-06-17 20:57:12,326][starVLA.training.train_starvla][INFO] - Step 2658, Loss: {'train/loss': 0.30864277482032776, 'train/grad_norm_pre_clip': 12.930276870727539, 'timing/data': 0.009185869246721268, 'timing/model': 0.9364319713786244, 'learning_rate/qwen_vl_interface': 7.126523504307061e-06, 'learning_rate/action_model': 7.126523504307061e-05, 'epoch': 1.21}) +[2026-06-17 20:57:16,138][starVLA.training.train_starvla][INFO] - Step 2659, Loss: {'train/loss': 0.3248855173587799, 'train/grad_norm_pre_clip': 12.839272499084473, 'timing/data': 0.008459554053843021, 'timing/model': 0.935169629752636, 'learning_rate/qwen_vl_interface': 7.124536540402371e-06, 'learning_rate/action_model': 7.124536540402372e-05, 'epoch': 1.21}) +[2026-06-17 20:57:20,033][starVLA.training.train_starvla][INFO] - Step 2660, Loss: {'train/loss': 0.3657497763633728, 'train/grad_norm_pre_clip': 15.268141746520996, 'timing/data': 0.010149035602807999, 'timing/model': 0.9391631167382002, 'learning_rate/qwen_vl_interface': 7.1225491879042405e-06, 'learning_rate/action_model': 7.12254918790424e-05, 'epoch': 1.21}) +[2026-06-17 20:57:23,854][starVLA.training.train_starvla][INFO] - Step 2661, Loss: {'train/loss': 0.5543188452720642, 'train/grad_norm_pre_clip': 30.1595516204834, 'timing/data': 0.00782719999551773, 'timing/model': 0.9392721932381392, 'learning_rate/qwen_vl_interface': 7.120561447224649e-06, 'learning_rate/action_model': 7.120561447224649e-05, 'epoch': 1.21}) +[2026-06-17 20:57:27,661][starVLA.training.train_starvla][INFO] - Step 2662, Loss: {'train/loss': 0.4366910457611084, 'train/grad_norm_pre_clip': 16.568830490112305, 'timing/data': 0.01567354891449213, 'timing/model': 0.9447713037952781, 'learning_rate/qwen_vl_interface': 7.118573318775656e-06, 'learning_rate/action_model': 7.118573318775655e-05, 'epoch': 1.21}) +[2026-06-17 20:57:31,510][starVLA.training.train_starvla][INFO] - Step 2663, Loss: {'train/loss': 0.5284622311592102, 'train/grad_norm_pre_clip': 25.88312530517578, 'timing/data': 0.018216640688478947, 'timing/model': 0.9467206122353673, 'learning_rate/qwen_vl_interface': 7.116584802969403e-06, 'learning_rate/action_model': 7.116584802969402e-05, 'epoch': 1.21}) +[2026-06-17 20:57:35,320][starVLA.training.train_starvla][INFO] - Step 2664, Loss: {'train/loss': 0.9281043410301208, 'train/grad_norm_pre_clip': 24.58334732055664, 'timing/data': 0.008401377126574516, 'timing/model': 1.11662223841995, 'learning_rate/qwen_vl_interface': 7.114595900218112e-06, 'learning_rate/action_model': 7.114595900218112e-05, 'epoch': 1.21}) +[2026-06-17 20:57:38,930][starVLA.training.train_starvla][INFO] - Step 2665, Loss: {'train/loss': 0.5974401235580444, 'train/grad_norm_pre_clip': 19.072254180908203, 'timing/data': 0.008448682725429535, 'timing/model': 0.9271940561011434, 'learning_rate/qwen_vl_interface': 7.112606610934084e-06, 'learning_rate/action_model': 7.112606610934084e-05, 'epoch': 1.21}) +[2026-06-17 20:57:42,738][starVLA.training.train_starvla][INFO] - Step 2666, Loss: {'train/loss': 0.3215699791908264, 'train/grad_norm_pre_clip': 13.890424728393555, 'timing/data': 0.009093426167964935, 'timing/model': 0.9325148975476623, 'learning_rate/qwen_vl_interface': 7.110616935529702e-06, 'learning_rate/action_model': 7.110616935529702e-05, 'epoch': 1.21}) +[2026-06-17 20:57:46,635][starVLA.training.train_starvla][INFO] - Step 2667, Loss: {'train/loss': 0.4767163395881653, 'train/grad_norm_pre_clip': 14.86527156829834, 'timing/data': 0.00799933448433876, 'timing/model': 0.9331933064386249, 'learning_rate/qwen_vl_interface': 7.108626874417426e-06, 'learning_rate/action_model': 7.108626874417426e-05, 'epoch': 1.21}) +[2026-06-17 20:57:50,441][starVLA.training.train_starvla][INFO] - Step 2668, Loss: {'train/loss': 0.7989611625671387, 'train/grad_norm_pre_clip': 21.647926330566406, 'timing/data': 0.007536293938755989, 'timing/model': 0.934495166875422, 'learning_rate/qwen_vl_interface': 7.106636428009801e-06, 'learning_rate/action_model': 7.1066364280098e-05, 'epoch': 1.21}) +[2026-06-17 20:57:54,307][starVLA.training.train_starvla][INFO] - Step 2669, Loss: {'train/loss': 0.270415335893631, 'train/grad_norm_pre_clip': 13.808145523071289, 'timing/data': 0.008024017326533794, 'timing/model': 0.9472594708204269, 'learning_rate/qwen_vl_interface': 7.104645596719445e-06, 'learning_rate/action_model': 7.104645596719445e-05, 'epoch': 1.21}) +[2026-06-17 20:57:58,113][starVLA.training.train_starvla][INFO] - Step 2670, Loss: {'train/loss': 0.28788599371910095, 'train/grad_norm_pre_clip': 13.954930305480957, 'timing/data': 0.0086293900385499, 'timing/model': 1.115822745487094, 'learning_rate/qwen_vl_interface': 7.102654380959063e-06, 'learning_rate/action_model': 7.102654380959063e-05, 'epoch': 1.21}) +[2026-06-17 20:58:01,935][starVLA.training.train_starvla][INFO] - Step 2671, Loss: {'train/loss': 0.40404337644577026, 'train/grad_norm_pre_clip': 17.465518951416016, 'timing/data': 0.008480559103190899, 'timing/model': 1.134328469634056, 'learning_rate/qwen_vl_interface': 7.100662781141432e-06, 'learning_rate/action_model': 7.100662781141432e-05, 'epoch': 1.21}) +[2026-06-17 20:58:05,629][starVLA.training.train_starvla][INFO] - Step 2672, Loss: {'train/loss': 0.3366069197654724, 'train/grad_norm_pre_clip': 18.979087829589844, 'timing/data': 0.010220280848443508, 'timing/model': 0.9662729371339083, 'learning_rate/qwen_vl_interface': 7.0986707976794155e-06, 'learning_rate/action_model': 7.098670797679415e-05, 'epoch': 1.21}) +[2026-06-17 20:58:09,474][starVLA.training.train_starvla][INFO] - Step 2673, Loss: {'train/loss': 0.2541121244430542, 'train/grad_norm_pre_clip': 11.828213691711426, 'timing/data': 0.010468036867678165, 'timing/model': 0.9678686689585447, 'learning_rate/qwen_vl_interface': 7.096678430985955e-06, 'learning_rate/action_model': 7.096678430985954e-05, 'epoch': 1.21}) +[2026-06-17 20:58:13,453][starVLA.training.train_starvla][INFO] - Step 2674, Loss: {'train/loss': 0.22134119272232056, 'train/grad_norm_pre_clip': 20.380970001220703, 'timing/data': 0.0215066010132432, 'timing/model': 0.9833810674026608, 'learning_rate/qwen_vl_interface': 7.094685681474066e-06, 'learning_rate/action_model': 7.094685681474065e-05, 'epoch': 1.21}) +[2026-06-17 20:58:17,332][starVLA.training.train_starvla][INFO] - Step 2675, Loss: {'train/loss': 0.33214661478996277, 'train/grad_norm_pre_clip': 16.163179397583008, 'timing/data': 0.009172572754323483, 'timing/model': 0.972657224163413, 'learning_rate/qwen_vl_interface': 7.092692549556853e-06, 'learning_rate/action_model': 7.092692549556853e-05, 'epoch': 1.21}) +[2026-06-17 20:58:21,166][starVLA.training.train_starvla][INFO] - Step 2676, Loss: {'train/loss': 0.5619080662727356, 'train/grad_norm_pre_clip': 17.327383041381836, 'timing/data': 0.011577242985367775, 'timing/model': 0.9395424276590347, 'learning_rate/qwen_vl_interface': 7.0906990356474895e-06, 'learning_rate/action_model': 7.090699035647489e-05, 'epoch': 1.21}) +[2026-06-17 20:58:24,965][starVLA.training.train_starvla][INFO] - Step 2677, Loss: {'train/loss': 0.5572570562362671, 'train/grad_norm_pre_clip': 18.568702697753906, 'timing/data': 0.010768312960863113, 'timing/model': 1.0914351642131805, 'learning_rate/qwen_vl_interface': 7.088705140159238e-06, 'learning_rate/action_model': 7.088705140159237e-05, 'epoch': 1.21}) +[2026-06-17 20:58:28,760][starVLA.training.train_starvla][INFO] - Step 2678, Loss: {'train/loss': 0.4182480275630951, 'train/grad_norm_pre_clip': 16.810901641845703, 'timing/data': 0.014034940861165524, 'timing/model': 1.0622996045276523, 'learning_rate/qwen_vl_interface': 7.086710863505428e-06, 'learning_rate/action_model': 7.086710863505428e-05, 'epoch': 1.21}) +[2026-06-17 20:58:32,574][starVLA.training.train_starvla][INFO] - Step 2679, Loss: {'train/loss': 0.30766090750694275, 'train/grad_norm_pre_clip': 18.11609649658203, 'timing/data': 0.00936275627464056, 'timing/model': 0.932257148437202, 'learning_rate/qwen_vl_interface': 7.084716206099485e-06, 'learning_rate/action_model': 7.084716206099484e-05, 'epoch': 1.22}) +[2026-06-17 20:58:36,528][starVLA.training.train_starvla][INFO] - Step 2680, Loss: {'train/loss': 0.4240405261516571, 'train/grad_norm_pre_clip': 16.088403701782227, 'timing/data': 0.016330277547240257, 'timing/model': 0.9483841201290488, 'learning_rate/qwen_vl_interface': 7.082721168354894e-06, 'learning_rate/action_model': 7.082721168354894e-05, 'epoch': 1.22}) +[2026-06-17 20:58:40,336][starVLA.training.train_starvla][INFO] - Step 2681, Loss: {'train/loss': 0.4778735637664795, 'train/grad_norm_pre_clip': 15.618232727050781, 'timing/data': 0.010408584959805012, 'timing/model': 0.9491214584559202, 'learning_rate/qwen_vl_interface': 7.080725750685235e-06, 'learning_rate/action_model': 7.080725750685234e-05, 'epoch': 1.22}) +[2026-06-17 20:58:44,044][starVLA.training.train_starvla][INFO] - Step 2682, Loss: {'train/loss': 0.6898378133773804, 'train/grad_norm_pre_clip': 22.235017776489258, 'timing/data': 0.013981141149997711, 'timing/model': 0.9369854535907507, 'learning_rate/qwen_vl_interface': 7.078729953504157e-06, 'learning_rate/action_model': 7.078729953504156e-05, 'epoch': 1.22}) +[2026-06-17 20:58:47,835][starVLA.training.train_starvla][INFO] - Step 2683, Loss: {'train/loss': 0.23099972307682037, 'train/grad_norm_pre_clip': 11.771924018859863, 'timing/data': 0.01249589491635561, 'timing/model': 0.9523511538282037, 'learning_rate/qwen_vl_interface': 7.076733777225391e-06, 'learning_rate/action_model': 7.07673377722539e-05, 'epoch': 1.22}) +[2026-06-17 20:58:51,676][starVLA.training.train_starvla][INFO] - Step 2684, Loss: {'train/loss': 0.45253080129623413, 'train/grad_norm_pre_clip': 15.316298484802246, 'timing/data': 0.00885334238409996, 'timing/model': 1.1773562850430608, 'learning_rate/qwen_vl_interface': 7.074737222262749e-06, 'learning_rate/action_model': 7.074737222262748e-05, 'epoch': 1.22}) +[2026-06-17 20:58:55,428][starVLA.training.train_starvla][INFO] - Step 2685, Loss: {'train/loss': 0.4007570445537567, 'train/grad_norm_pre_clip': 14.180170059204102, 'timing/data': 0.011984393000602722, 'timing/model': 0.9378700070083141, 'learning_rate/qwen_vl_interface': 7.072740289030115e-06, 'learning_rate/action_model': 7.072740289030115e-05, 'epoch': 1.22}) +[2026-06-17 20:58:59,451][starVLA.training.train_starvla][INFO] - Step 2686, Loss: {'train/loss': 0.3836731016635895, 'train/grad_norm_pre_clip': 18.99369239807129, 'timing/data': 0.012393471784889698, 'timing/model': 0.9304032800719142, 'learning_rate/qwen_vl_interface': 7.07074297794146e-06, 'learning_rate/action_model': 7.07074297794146e-05, 'epoch': 1.22}) +[2026-06-17 20:59:03,270][starVLA.training.train_starvla][INFO] - Step 2687, Loss: {'train/loss': 0.3988531231880188, 'train/grad_norm_pre_clip': 16.392183303833008, 'timing/data': 0.013403542339801788, 'timing/model': 0.9451461248099804, 'learning_rate/qwen_vl_interface': 7.068745289410826e-06, 'learning_rate/action_model': 7.068745289410827e-05, 'epoch': 1.22}) +[2026-06-17 20:59:07,136][starVLA.training.train_starvla][INFO] - Step 2688, Loss: {'train/loss': 0.5908668041229248, 'train/grad_norm_pre_clip': 14.821222305297852, 'timing/data': 0.009487788192927837, 'timing/model': 0.9527818327769637, 'learning_rate/qwen_vl_interface': 7.066747223852337e-06, 'learning_rate/action_model': 7.066747223852337e-05, 'epoch': 1.22}) +[2026-06-17 20:59:10,941][starVLA.training.train_starvla][INFO] - Step 2689, Loss: {'train/loss': 0.36781013011932373, 'train/grad_norm_pre_clip': 14.142745971679688, 'timing/data': 0.008028421550989151, 'timing/model': 0.937745489180088, 'learning_rate/qwen_vl_interface': 7.064748781680192e-06, 'learning_rate/action_model': 7.064748781680192e-05, 'epoch': 1.22}) +[2026-06-17 20:59:14,750][starVLA.training.train_starvla][INFO] - Step 2690, Loss: {'train/loss': 0.38844093680381775, 'train/grad_norm_pre_clip': 15.473999977111816, 'timing/data': 0.008732817135751247, 'timing/model': 1.1026089545339346, 'learning_rate/qwen_vl_interface': 7.062749963308673e-06, 'learning_rate/action_model': 7.062749963308672e-05, 'epoch': 1.22}) +[2026-06-17 20:59:18,600][starVLA.training.train_starvla][INFO] - Step 2691, Loss: {'train/loss': 0.4239688515663147, 'train/grad_norm_pre_clip': 20.3828125, 'timing/data': 0.008299950510263443, 'timing/model': 1.1373611437156796, 'learning_rate/qwen_vl_interface': 7.060750769152135e-06, 'learning_rate/action_model': 7.060750769152135e-05, 'epoch': 1.22}) +[2026-06-17 20:59:22,233][starVLA.training.train_starvla][INFO] - Step 2692, Loss: {'train/loss': 0.43375372886657715, 'train/grad_norm_pre_clip': 12.753994941711426, 'timing/data': 0.007938997820019722, 'timing/model': 0.9355998104438186, 'learning_rate/qwen_vl_interface': 7.058751199625014e-06, 'learning_rate/action_model': 7.058751199625015e-05, 'epoch': 1.22}) +[2026-06-17 20:59:26,100][starVLA.training.train_starvla][INFO] - Step 2693, Loss: {'train/loss': 0.5651751160621643, 'train/grad_norm_pre_clip': 14.414962768554688, 'timing/data': 0.012675213627517223, 'timing/model': 0.9532115561887622, 'learning_rate/qwen_vl_interface': 7.056751255141824e-06, 'learning_rate/action_model': 7.056751255141824e-05, 'epoch': 1.22}) +[2026-06-17 20:59:29,946][starVLA.training.train_starvla][INFO] - Step 2694, Loss: {'train/loss': 0.22586292028427124, 'train/grad_norm_pre_clip': 12.203015327453613, 'timing/data': 0.00756246130913496, 'timing/model': 0.9313608733937144, 'learning_rate/qwen_vl_interface': 7.054750936117155e-06, 'learning_rate/action_model': 7.054750936117156e-05, 'epoch': 1.22}) +[2026-06-17 20:59:33,820][starVLA.training.train_starvla][INFO] - Step 2695, Loss: {'train/loss': 0.15236185491085052, 'train/grad_norm_pre_clip': 8.126136779785156, 'timing/data': 0.008879275992512703, 'timing/model': 0.972423255443573, 'learning_rate/qwen_vl_interface': 7.0527502429656745e-06, 'learning_rate/action_model': 7.052750242965674e-05, 'epoch': 1.22}) +[2026-06-17 20:59:37,607][starVLA.training.train_starvla][INFO] - Step 2696, Loss: {'train/loss': 0.2702554166316986, 'train/grad_norm_pre_clip': 11.653961181640625, 'timing/data': 0.010013051331043243, 'timing/model': 0.9374076295644045, 'learning_rate/qwen_vl_interface': 7.050749176102128e-06, 'learning_rate/action_model': 7.050749176102129e-05, 'epoch': 1.22}) +[2026-06-17 20:59:41,631][starVLA.training.train_starvla][INFO] - Step 2697, Loss: {'train/loss': 0.2090148776769638, 'train/grad_norm_pre_clip': 10.915060997009277, 'timing/data': 0.012195943854749203, 'timing/model': 1.1219001589342952, 'learning_rate/qwen_vl_interface': 7.048747735941342e-06, 'learning_rate/action_model': 7.048747735941341e-05, 'epoch': 1.22}) +[2026-06-17 20:59:45,416][starVLA.training.train_starvla][INFO] - Step 2698, Loss: {'train/loss': 0.5694043040275574, 'train/grad_norm_pre_clip': 18.12668800354004, 'timing/data': 0.009872909635305405, 'timing/model': 1.0994442645460367, 'learning_rate/qwen_vl_interface': 7.046745922898212e-06, 'learning_rate/action_model': 7.046745922898211e-05, 'epoch': 1.22}) +[2026-06-17 20:59:49,049][starVLA.training.train_starvla][INFO] - Step 2699, Loss: {'train/loss': 0.2572896480560303, 'train/grad_norm_pre_clip': 13.297773361206055, 'timing/data': 0.007938262075185776, 'timing/model': 0.9423594921827316, 'learning_rate/qwen_vl_interface': 7.04474373738772e-06, 'learning_rate/action_model': 7.04474373738772e-05, 'epoch': 1.22}) +[2026-06-17 21:02:03,470][starVLA.training.train_starvla][INFO] - Step 2700, Loss: {'train/loss': 0.25582972168922424, 'train/grad_norm_pre_clip': 13.328193664550781, 'eval/loss': 0.41503241658210754, 'eval/latency_8/loss': 0.41503241658210754, 'eval/demon_fire_precision': 0.9305515027721039, 'eval/demon_fire_recall': 0.9906803355079217, 'eval/demon_fire_f1': 0.9596749924766776, 'eval/demon_fire_rate_model': 0.9535336672231497, 'eval/demon_fire_rate_teacher': 0.8956594323873122, 'eval/demon_left_precision': 0.7419354838709677, 'eval/demon_left_recall': 0.7387580299785867, 'eval/demon_left_f1': 0.7403433476394851, 'eval/demon_left_rate_model': 0.1293823038397329, 'eval/demon_left_rate_teacher': 0.12993878686700056, 'eval/demon_right_precision': 0.8577319587628865, 'eval/demon_right_recall': 0.6303030303030303, 'eval/demon_right_f1': 0.7266375545851528, 'eval/demon_right_rate_model': 0.13494713411240958, 'eval/demon_right_rate_teacher': 0.18363939899833054, 'eval/demon_move_f1': 0.7334904511123189, 'eval/demon_cc_f1': 0.8784128291124248, 'eval/latency_8/demon_fire_precision': 0.9305515027721039, 'eval/latency_8/demon_fire_recall': 0.9906803355079217, 'eval/latency_8/demon_fire_f1': 0.9596749924766776, 'eval/latency_8/demon_fire_rate_model': 0.9535336672231497, 'eval/latency_8/demon_fire_rate_teacher': 0.8956594323873122, 'eval/latency_8/demon_left_precision': 0.7419354838709677, 'eval/latency_8/demon_left_recall': 0.7387580299785867, 'eval/latency_8/demon_left_f1': 0.7403433476394851, 'eval/latency_8/demon_left_rate_model': 0.1293823038397329, 'eval/latency_8/demon_left_rate_teacher': 0.12993878686700056, 'eval/latency_8/demon_right_precision': 0.8577319587628865, 'eval/latency_8/demon_right_recall': 0.6303030303030303, 'eval/latency_8/demon_right_f1': 0.7266375545851528, 'eval/latency_8/demon_right_rate_model': 0.13494713411240958, 'eval/latency_8/demon_right_rate_teacher': 0.18363939899833054, 'eval/latency_8/demon_move_f1': 0.7334904511123189, 'eval/latency_8/demon_cc_f1': 0.8784128291124248, 'timing/data': 0.00869786273688078, 'timing/model': 0.945325868204236, 'learning_rate/qwen_vl_interface': 7.042741179824919e-06, 'learning_rate/action_model': 7.042741179824918e-05, 'epoch': 1.22}) +[2026-06-17 21:02:03,479][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_2700_state +[2026-06-17 21:02:03,484][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 21:02:10,678][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_2700_state/model.safetensors +[2026-06-17 21:02:26,405][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_2700_state/optimizer.bin +[2026-06-17 21:02:26,406][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_2700_state/sampler.bin +[2026-06-17 21:02:26,406][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_2700_state/sampler_1.bin +[2026-06-17 21:02:26,408][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_2700_state/random_states_0.pkl +[2026-06-17 21:02:26,413][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2700 +[2026-06-17 21:02:26,413][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 21:02:26,423][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 21:02:30,115][starVLA.training.train_starvla][INFO] - Step 2701, Loss: {'train/loss': 0.4538571834564209, 'train/grad_norm_pre_clip': 16.720794677734375, 'timing/data': 0.007832346484065056, 'timing/model': 0.9465911714360118, 'learning_rate/qwen_vl_interface': 7.040738250624944e-06, 'learning_rate/action_model': 7.040738250624944e-05, 'epoch': 1.23}) +[2026-06-17 21:02:34,163][starVLA.training.train_starvla][INFO] - Step 2702, Loss: {'train/loss': 0.37112876772880554, 'train/grad_norm_pre_clip': 13.858757019042969, 'timing/data': 0.009619656018912792, 'timing/model': 1.1629607221111655, 'learning_rate/qwen_vl_interface': 7.038734950203001e-06, 'learning_rate/action_model': 7.038734950203001e-05, 'epoch': 1.23}) +[2026-06-17 21:02:39,595][starVLA.training.train_starvla][INFO] - Step 2703, Loss: {'train/loss': 0.46726229786872864, 'train/grad_norm_pre_clip': 15.122020721435547, 'timing/data': 0.012561127543449402, 'timing/model': 1.1224913857877254, 'learning_rate/qwen_vl_interface': 7.036731278974379e-06, 'learning_rate/action_model': 7.036731278974378e-05, 'epoch': 1.23}) +[2026-06-17 21:02:43,522][starVLA.training.train_starvla][INFO] - Step 2704, Loss: {'train/loss': 0.16145606338977814, 'train/grad_norm_pre_clip': 10.323264122009277, 'timing/data': 0.01441252138465643, 'timing/model': 1.042082592844963, 'learning_rate/qwen_vl_interface': 7.034727237354442e-06, 'learning_rate/action_model': 7.034727237354441e-05, 'epoch': 1.23}) +[2026-06-17 21:02:47,819][starVLA.training.train_starvla][INFO] - Step 2705, Loss: {'train/loss': 0.3383033871650696, 'train/grad_norm_pre_clip': 18.61598014831543, 'timing/data': 0.011200414970517159, 'timing/model': 1.4156639380380511, 'learning_rate/qwen_vl_interface': 7.0327228257586244e-06, 'learning_rate/action_model': 7.032722825758625e-05, 'epoch': 1.23}) +[2026-06-17 21:02:51,641][starVLA.training.train_starvla][INFO] - Step 2706, Loss: {'train/loss': 0.7031025886535645, 'train/grad_norm_pre_clip': 16.339937210083008, 'timing/data': 0.009428351186215878, 'timing/model': 1.1402575764805079, 'learning_rate/qwen_vl_interface': 7.030718044602451e-06, 'learning_rate/action_model': 7.03071804460245e-05, 'epoch': 1.23}) +[2026-06-17 21:02:55,617][starVLA.training.train_starvla][INFO] - Step 2707, Loss: {'train/loss': 0.4052499532699585, 'train/grad_norm_pre_clip': 15.67322063446045, 'timing/data': 0.014742684550583363, 'timing/model': 1.1324343606829643, 'learning_rate/qwen_vl_interface': 7.028712894301507e-06, 'learning_rate/action_model': 7.028712894301506e-05, 'epoch': 1.23}) +[2026-06-17 21:02:59,243][starVLA.training.train_starvla][INFO] - Step 2708, Loss: {'train/loss': 0.3280501961708069, 'train/grad_norm_pre_clip': 14.956055641174316, 'timing/data': 0.009735393337905407, 'timing/model': 0.9253361793234944, 'learning_rate/qwen_vl_interface': 7.026707375271469e-06, 'learning_rate/action_model': 7.026707375271468e-05, 'epoch': 1.23}) +[2026-06-17 21:03:03,065][starVLA.training.train_starvla][INFO] - Step 2709, Loss: {'train/loss': 0.48277410864830017, 'train/grad_norm_pre_clip': 24.748214721679688, 'timing/data': 0.013095579110085964, 'timing/model': 0.9399544624611735, 'learning_rate/qwen_vl_interface': 7.024701487928079e-06, 'learning_rate/action_model': 7.024701487928078e-05, 'epoch': 1.23}) +[2026-06-17 21:03:06,937][starVLA.training.train_starvla][INFO] - Step 2710, Loss: {'train/loss': 0.34728527069091797, 'train/grad_norm_pre_clip': 15.567012786865234, 'timing/data': 0.0079743517562747, 'timing/model': 0.9384512696415186, 'learning_rate/qwen_vl_interface': 7.02269523268716e-06, 'learning_rate/action_model': 7.022695232687159e-05, 'epoch': 1.23}) +[2026-06-17 21:03:10,730][starVLA.training.train_starvla][INFO] - Step 2711, Loss: {'train/loss': 0.45379409193992615, 'train/grad_norm_pre_clip': 21.70222282409668, 'timing/data': 0.012152091600000858, 'timing/model': 0.9395048059523106, 'learning_rate/qwen_vl_interface': 7.020688609964612e-06, 'learning_rate/action_model': 7.020688609964612e-05, 'epoch': 1.23}) +[2026-06-17 21:03:14,585][starVLA.training.train_starvla][INFO] - Step 2712, Loss: {'train/loss': 0.5264573693275452, 'train/grad_norm_pre_clip': 17.96492576599121, 'timing/data': 0.009439688175916672, 'timing/model': 0.9641065262258053, 'learning_rate/qwen_vl_interface': 7.018681620176412e-06, 'learning_rate/action_model': 7.018681620176412e-05, 'epoch': 1.23}) +[2026-06-17 21:03:18,427][starVLA.training.train_starvla][INFO] - Step 2713, Loss: {'train/loss': 0.4141550362110138, 'train/grad_norm_pre_clip': 21.594005584716797, 'timing/data': 0.008910514414310455, 'timing/model': 1.1333942580968142, 'learning_rate/qwen_vl_interface': 7.016674263738606e-06, 'learning_rate/action_model': 7.016674263738607e-05, 'epoch': 1.23}) +[2026-06-17 21:03:22,054][starVLA.training.train_starvla][INFO] - Step 2714, Loss: {'train/loss': 0.32736921310424805, 'train/grad_norm_pre_clip': 15.1780366897583, 'timing/data': 0.009093469940125942, 'timing/model': 0.9355533653870225, 'learning_rate/qwen_vl_interface': 7.014666541067325e-06, 'learning_rate/action_model': 7.014666541067325e-05, 'epoch': 1.23}) +[2026-06-17 21:03:25,896][starVLA.training.train_starvla][INFO] - Step 2715, Loss: {'train/loss': 0.17261025309562683, 'train/grad_norm_pre_clip': 10.246150016784668, 'timing/data': 0.009444799274206161, 'timing/model': 0.9398718234151602, 'learning_rate/qwen_vl_interface': 7.012658452578774e-06, 'learning_rate/action_model': 7.012658452578773e-05, 'epoch': 1.23}) +[2026-06-17 21:03:29,680][starVLA.training.train_starvla][INFO] - Step 2716, Loss: {'train/loss': 0.4104016423225403, 'train/grad_norm_pre_clip': 15.02055549621582, 'timing/data': 0.010076018050312996, 'timing/model': 0.9422346530482173, 'learning_rate/qwen_vl_interface': 7.010649998689225e-06, 'learning_rate/action_model': 7.010649998689224e-05, 'epoch': 1.23}) +[2026-06-17 21:03:33,625][starVLA.training.train_starvla][INFO] - Step 2717, Loss: {'train/loss': 0.4238978624343872, 'train/grad_norm_pre_clip': 21.507389068603516, 'timing/data': 0.00983242318034172, 'timing/model': 0.9497543796896935, 'learning_rate/qwen_vl_interface': 7.0086411798150385e-06, 'learning_rate/action_model': 7.008641179815038e-05, 'epoch': 1.23}) +[2026-06-17 21:03:37,457][starVLA.training.train_starvla][INFO] - Step 2718, Loss: {'train/loss': 0.20363959670066833, 'train/grad_norm_pre_clip': 10.705227851867676, 'timing/data': 0.01207461766898632, 'timing/model': 0.9505397295579314, 'learning_rate/qwen_vl_interface': 7.006631996372642e-06, 'learning_rate/action_model': 7.006631996372643e-05, 'epoch': 1.23}) +[2026-06-17 21:03:41,239][starVLA.training.train_starvla][INFO] - Step 2719, Loss: {'train/loss': 0.17963631451129913, 'train/grad_norm_pre_clip': 14.817888259887695, 'timing/data': 0.007433961145579815, 'timing/model': 0.9328959006816149, 'learning_rate/qwen_vl_interface': 7.004622448778545e-06, 'learning_rate/action_model': 7.004622448778544e-05, 'epoch': 1.23}) +[2026-06-17 21:03:45,112][starVLA.training.train_starvla][INFO] - Step 2720, Loss: {'train/loss': 0.7126381993293762, 'train/grad_norm_pre_clip': 20.807106018066406, 'timing/data': 0.00657684076577425, 'timing/model': 1.1344542978331447, 'learning_rate/qwen_vl_interface': 7.0026125374493244e-06, 'learning_rate/action_model': 7.002612537449324e-05, 'epoch': 1.23}) +[2026-06-17 21:03:48,915][starVLA.training.train_starvla][INFO] - Step 2721, Loss: {'train/loss': 0.5691549181938171, 'train/grad_norm_pre_clip': 15.647459983825684, 'timing/data': 0.015426906757056713, 'timing/model': 1.0385346915572882, 'learning_rate/qwen_vl_interface': 7.000602262801639e-06, 'learning_rate/action_model': 7.000602262801638e-05, 'epoch': 1.23}) +[2026-06-17 21:03:52,827][starVLA.training.train_starvla][INFO] - Step 2722, Loss: {'train/loss': 0.2075108289718628, 'train/grad_norm_pre_clip': 10.965662002563477, 'timing/data': 0.007043309509754181, 'timing/model': 1.0437673199921846, 'learning_rate/qwen_vl_interface': 6.9985916252522225e-06, 'learning_rate/action_model': 6.998591625252222e-05, 'epoch': 1.23}) +[2026-06-17 21:03:56,636][starVLA.training.train_starvla][INFO] - Step 2723, Loss: {'train/loss': 0.20338864624500275, 'train/grad_norm_pre_clip': 10.653383255004883, 'timing/data': 0.007621078751981258, 'timing/model': 0.9292616592720151, 'learning_rate/qwen_vl_interface': 6.9965806252178785e-06, 'learning_rate/action_model': 6.996580625217877e-05, 'epoch': 1.24}) +[2026-06-17 21:04:00,435][starVLA.training.train_starvla][INFO] - Step 2724, Loss: {'train/loss': 0.6008468866348267, 'train/grad_norm_pre_clip': 35.396732330322266, 'timing/data': 0.00790525320917368, 'timing/model': 0.9514036048203707, 'learning_rate/qwen_vl_interface': 6.994569263115491e-06, 'learning_rate/action_model': 6.994569263115491e-05, 'epoch': 1.24}) +[2026-06-17 21:04:04,240][starVLA.training.train_starvla][INFO] - Step 2725, Loss: {'train/loss': 0.49282726645469666, 'train/grad_norm_pre_clip': 17.087444305419922, 'timing/data': 0.007955613546073437, 'timing/model': 0.9368285350501537, 'learning_rate/qwen_vl_interface': 6.992557539362019e-06, 'learning_rate/action_model': 6.99255753936202e-05, 'epoch': 1.24}) +[2026-06-17 21:04:08,109][starVLA.training.train_starvla][INFO] - Step 2726, Loss: {'train/loss': 0.5987008810043335, 'train/grad_norm_pre_clip': 15.027565956115723, 'timing/data': 0.013016321696341038, 'timing/model': 1.149629677645862, 'learning_rate/qwen_vl_interface': 6.990545454374495e-06, 'learning_rate/action_model': 6.990545454374494e-05, 'epoch': 1.24}) +[2026-06-17 21:04:12,025][starVLA.training.train_starvla][INFO] - Step 2727, Loss: {'train/loss': 0.31192803382873535, 'train/grad_norm_pre_clip': 15.428630828857422, 'timing/data': 0.009745384566485882, 'timing/model': 1.1824372690171003, 'learning_rate/qwen_vl_interface': 6.988533008570024e-06, 'learning_rate/action_model': 6.988533008570024e-05, 'epoch': 1.24}) +[2026-06-17 21:04:15,693][starVLA.training.train_starvla][INFO] - Step 2728, Loss: {'train/loss': 0.8005260825157166, 'train/grad_norm_pre_clip': 20.084014892578125, 'timing/data': 0.0099489139392972, 'timing/model': 0.9494295716285706, 'learning_rate/qwen_vl_interface': 6.98652020236579e-06, 'learning_rate/action_model': 6.986520202365789e-05, 'epoch': 1.24}) +[2026-06-17 21:04:20,397][starVLA.training.train_starvla][INFO] - Step 2729, Loss: {'train/loss': 0.4239632487297058, 'train/grad_norm_pre_clip': 13.805331230163574, 'timing/data': 0.009670140221714973, 'timing/model': 0.9357801079750061, 'learning_rate/qwen_vl_interface': 6.98450703617905e-06, 'learning_rate/action_model': 6.98450703617905e-05, 'epoch': 1.24}) +[2026-06-17 21:04:24,244][starVLA.training.train_starvla][INFO] - Step 2730, Loss: {'train/loss': 0.2872347831726074, 'train/grad_norm_pre_clip': 14.447612762451172, 'timing/data': 0.008101869374513626, 'timing/model': 0.9374892292544246, 'learning_rate/qwen_vl_interface': 6.982493510427134e-06, 'learning_rate/action_model': 6.982493510427133e-05, 'epoch': 1.24}) +[2026-06-17 21:04:28,228][starVLA.training.train_starvla][INFO] - Step 2731, Loss: {'train/loss': 0.3061986267566681, 'train/grad_norm_pre_clip': 13.00175952911377, 'timing/data': 0.012104076333343983, 'timing/model': 0.9538953136652708, 'learning_rate/qwen_vl_interface': 6.980479625527449e-06, 'learning_rate/action_model': 6.980479625527448e-05, 'epoch': 1.24}) +[2026-06-17 21:04:32,054][starVLA.training.train_starvla][INFO] - Step 2732, Loss: {'train/loss': 0.4432518780231476, 'train/grad_norm_pre_clip': 16.318513870239258, 'timing/data': 0.007816461846232414, 'timing/model': 0.948614526540041, 'learning_rate/qwen_vl_interface': 6.978465381897477e-06, 'learning_rate/action_model': 6.978465381897477e-05, 'epoch': 1.24}) +[2026-06-17 21:04:36,272][starVLA.training.train_starvla][INFO] - Step 2733, Loss: {'train/loss': 0.25553905963897705, 'train/grad_norm_pre_clip': 12.696157455444336, 'timing/data': 0.009344243444502354, 'timing/model': 1.4181597456336021, 'learning_rate/qwen_vl_interface': 6.97645077995477e-06, 'learning_rate/action_model': 6.97645077995477e-05, 'epoch': 1.24}) +[2026-06-17 21:04:40,095][starVLA.training.train_starvla][INFO] - Step 2734, Loss: {'train/loss': 0.5118246078491211, 'train/grad_norm_pre_clip': 18.32809066772461, 'timing/data': 0.0072336746379733086, 'timing/model': 1.1056062644347548, 'learning_rate/qwen_vl_interface': 6.97443582011696e-06, 'learning_rate/action_model': 6.974435820116959e-05, 'epoch': 1.24}) +[2026-06-17 21:04:44,476][starVLA.training.train_starvla][INFO] - Step 2735, Loss: {'train/loss': 0.23444920778274536, 'train/grad_norm_pre_clip': 12.088091850280762, 'timing/data': 0.009260986931622028, 'timing/model': 0.9521641526371241, 'learning_rate/qwen_vl_interface': 6.972420502801748e-06, 'learning_rate/action_model': 6.972420502801747e-05, 'epoch': 1.24}) +[2026-06-17 21:04:49,630][starVLA.training.train_starvla][INFO] - Step 2736, Loss: {'train/loss': 0.1477443128824234, 'train/grad_norm_pre_clip': 8.62501049041748, 'timing/data': 0.012810750864446163, 'timing/model': 0.9370141820982099, 'learning_rate/qwen_vl_interface': 6.970404828426911e-06, 'learning_rate/action_model': 6.97040482842691e-05, 'epoch': 1.24}) +[2026-06-17 21:04:54,325][starVLA.training.train_starvla][INFO] - Step 2737, Loss: {'train/loss': 0.45058658719062805, 'train/grad_norm_pre_clip': 21.762264251708984, 'timing/data': 0.011737382039427757, 'timing/model': 1.0479129310697317, 'learning_rate/qwen_vl_interface': 6.9683887974103044e-06, 'learning_rate/action_model': 6.968388797410305e-05, 'epoch': 1.24}) +[2026-06-17 21:04:58,222][starVLA.training.train_starvla][INFO] - Step 2738, Loss: {'train/loss': 0.3812861740589142, 'train/grad_norm_pre_clip': 13.374969482421875, 'timing/data': 0.008465702645480633, 'timing/model': 0.950274245813489, 'learning_rate/qwen_vl_interface': 6.966372410169848e-06, 'learning_rate/action_model': 6.966372410169848e-05, 'epoch': 1.24}) +[2026-06-17 21:05:02,122][starVLA.training.train_starvla][INFO] - Step 2739, Loss: {'train/loss': 0.36207789182662964, 'train/grad_norm_pre_clip': 10.558082580566406, 'timing/data': 0.008338631130754948, 'timing/model': 0.9719112226739526, 'learning_rate/qwen_vl_interface': 6.964355667123548e-06, 'learning_rate/action_model': 6.964355667123547e-05, 'epoch': 1.24}) +[2026-06-17 21:05:06,341][starVLA.training.train_starvla][INFO] - Step 2740, Loss: {'train/loss': 0.358124703168869, 'train/grad_norm_pre_clip': 20.963369369506836, 'timing/data': 0.009160668589174747, 'timing/model': 1.265920851379633, 'learning_rate/qwen_vl_interface': 6.96233856868947e-06, 'learning_rate/action_model': 6.962338568689469e-05, 'epoch': 1.24}) +[2026-06-17 21:05:10,161][starVLA.training.train_starvla][INFO] - Step 2741, Loss: {'train/loss': 0.40018197894096375, 'train/grad_norm_pre_clip': 23.549297332763672, 'timing/data': 0.0094819450750947, 'timing/model': 0.966592276468873, 'learning_rate/qwen_vl_interface': 6.960321115285767e-06, 'learning_rate/action_model': 6.960321115285767e-05, 'epoch': 1.24}) +[2026-06-17 21:05:14,645][starVLA.training.train_starvla][INFO] - Step 2742, Loss: {'train/loss': 0.7300770878791809, 'train/grad_norm_pre_clip': 26.361608505249023, 'timing/data': 0.012329470366239548, 'timing/model': 0.9544087173417211, 'learning_rate/qwen_vl_interface': 6.958303307330655e-06, 'learning_rate/action_model': 6.958303307330655e-05, 'epoch': 1.24}) +[2026-06-17 21:05:18,437][starVLA.training.train_starvla][INFO] - Step 2743, Loss: {'train/loss': 0.12732040882110596, 'train/grad_norm_pre_clip': 8.245646476745605, 'timing/data': 0.010206521488726139, 'timing/model': 0.949404738843441, 'learning_rate/qwen_vl_interface': 6.956285145242429e-06, 'learning_rate/action_model': 6.956285145242428e-05, 'epoch': 1.24}) +[2026-06-17 21:05:22,243][starVLA.training.train_starvla][INFO] - Step 2744, Loss: {'train/loss': 0.5017869472503662, 'train/grad_norm_pre_clip': 35.70578384399414, 'timing/data': 0.009104464203119278, 'timing/model': 0.9390125386416912, 'learning_rate/qwen_vl_interface': 6.954266629439456e-06, 'learning_rate/action_model': 6.954266629439457e-05, 'epoch': 1.24}) +[2026-06-17 21:05:26,233][starVLA.training.train_starvla][INFO] - Step 2745, Loss: {'train/loss': 0.16911910474300385, 'train/grad_norm_pre_clip': 9.038775444030762, 'timing/data': 0.008569661527872086, 'timing/model': 1.0263215089216828, 'learning_rate/qwen_vl_interface': 6.952247760340176e-06, 'learning_rate/action_model': 6.952247760340176e-05, 'epoch': 1.25}) +[2026-06-17 21:05:30,132][starVLA.training.train_starvla][INFO] - Step 2746, Loss: {'train/loss': 0.622674822807312, 'train/grad_norm_pre_clip': 16.64797592163086, 'timing/data': 0.010589232668280602, 'timing/model': 1.0979860378429294, 'learning_rate/qwen_vl_interface': 6.950228538363108e-06, 'learning_rate/action_model': 6.950228538363107e-05, 'epoch': 1.25}) +[2026-06-17 21:05:34,211][starVLA.training.train_starvla][INFO] - Step 2747, Loss: {'train/loss': 0.647441565990448, 'train/grad_norm_pre_clip': 22.637216567993164, 'timing/data': 0.014226243831217289, 'timing/model': 1.3337527327239513, 'learning_rate/qwen_vl_interface': 6.948208963926829e-06, 'learning_rate/action_model': 6.94820896392683e-05, 'epoch': 1.25}) +[2026-06-17 21:05:37,931][starVLA.training.train_starvla][INFO] - Step 2748, Loss: {'train/loss': 0.49128445982933044, 'train/grad_norm_pre_clip': 20.01995086669922, 'timing/data': 0.00987412128597498, 'timing/model': 0.9534479221329093, 'learning_rate/qwen_vl_interface': 6.946189037450007e-06, 'learning_rate/action_model': 6.946189037450007e-05, 'epoch': 1.25}) +[2026-06-17 21:05:41,823][starVLA.training.train_starvla][INFO] - Step 2749, Loss: {'train/loss': 0.7032023072242737, 'train/grad_norm_pre_clip': 22.995786666870117, 'timing/data': 0.008749107830226421, 'timing/model': 0.9413638636469841, 'learning_rate/qwen_vl_interface': 6.9441687593513715e-06, 'learning_rate/action_model': 6.944168759351372e-05, 'epoch': 1.25}) +[2026-06-17 21:05:46,060][starVLA.training.train_starvla][INFO] - Step 2750, Loss: {'train/loss': 0.4818628132343292, 'train/grad_norm_pre_clip': 24.051795959472656, 'timing/data': 0.01244864147156477, 'timing/model': 0.9772150022909045, 'learning_rate/qwen_vl_interface': 6.942148130049732e-06, 'learning_rate/action_model': 6.942148130049731e-05, 'epoch': 1.25}) +[2026-06-17 21:05:49,963][starVLA.training.train_starvla][INFO] - Step 2751, Loss: {'train/loss': 0.5045343041419983, 'train/grad_norm_pre_clip': 17.27193832397461, 'timing/data': 0.008171215653419495, 'timing/model': 0.9883936047554016, 'learning_rate/qwen_vl_interface': 6.940127149963961e-06, 'learning_rate/action_model': 6.94012714996396e-05, 'epoch': 1.25}) +[2026-06-17 21:05:53,849][starVLA.training.train_starvla][INFO] - Step 2752, Loss: {'train/loss': 0.16046009957790375, 'train/grad_norm_pre_clip': 10.931282043457031, 'timing/data': 0.010171692818403244, 'timing/model': 0.9447597349062562, 'learning_rate/qwen_vl_interface': 6.9381058195130146e-06, 'learning_rate/action_model': 6.938105819513014e-05, 'epoch': 1.25}) +[2026-06-17 21:05:57,733][starVLA.training.train_starvla][INFO] - Step 2753, Loss: {'train/loss': 0.19858230650424957, 'train/grad_norm_pre_clip': 11.811334609985352, 'timing/data': 0.00973994005471468, 'timing/model': 1.1259221844375134, 'learning_rate/qwen_vl_interface': 6.936084139115915e-06, 'learning_rate/action_model': 6.936084139115915e-05, 'epoch': 1.25}) +[2026-06-17 21:06:01,598][starVLA.training.train_starvla][INFO] - Step 2754, Loss: {'train/loss': 0.4445275068283081, 'train/grad_norm_pre_clip': 16.969438552856445, 'timing/data': 0.010759755969047546, 'timing/model': 1.1108001321554184, 'learning_rate/qwen_vl_interface': 6.934062109191759e-06, 'learning_rate/action_model': 6.934062109191759e-05, 'epoch': 1.25}) +[2026-06-17 21:06:05,253][starVLA.training.train_starvla][INFO] - Step 2755, Loss: {'train/loss': 0.49857088923454285, 'train/grad_norm_pre_clip': 15.796965599060059, 'timing/data': 0.008335443213582039, 'timing/model': 0.9560330919921398, 'learning_rate/qwen_vl_interface': 6.9320397301597155e-06, 'learning_rate/action_model': 6.932039730159715e-05, 'epoch': 1.25}) +[2026-06-17 21:06:09,052][starVLA.training.train_starvla][INFO] - Step 2756, Loss: {'train/loss': 0.4358977973461151, 'train/grad_norm_pre_clip': 18.5513858795166, 'timing/data': 0.007863990031182766, 'timing/model': 0.9366352474316955, 'learning_rate/qwen_vl_interface': 6.930017002439027e-06, 'learning_rate/action_model': 6.930017002439027e-05, 'epoch': 1.25}) +[2026-06-17 21:06:12,934][starVLA.training.train_starvla][INFO] - Step 2757, Loss: {'train/loss': 0.3065545856952667, 'train/grad_norm_pre_clip': 18.419570922851562, 'timing/data': 0.009207875467836857, 'timing/model': 0.9680805439129472, 'learning_rate/qwen_vl_interface': 6.927993926449005e-06, 'learning_rate/action_model': 6.927993926449005e-05, 'epoch': 1.25}) +[2026-06-17 21:06:16,742][starVLA.training.train_starvla][INFO] - Step 2758, Loss: {'train/loss': 0.7206605672836304, 'train/grad_norm_pre_clip': 19.0106258392334, 'timing/data': 0.013363351114094257, 'timing/model': 0.9263829970732331, 'learning_rate/qwen_vl_interface': 6.925970502609036e-06, 'learning_rate/action_model': 6.925970502609036e-05, 'epoch': 1.25}) +[2026-06-17 21:06:20,595][starVLA.training.train_starvla][INFO] - Step 2759, Loss: {'train/loss': 0.21422246098518372, 'train/grad_norm_pre_clip': 10.658705711364746, 'timing/data': 0.008270924910902977, 'timing/model': 0.9481241134926677, 'learning_rate/qwen_vl_interface': 6.92394673133858e-06, 'learning_rate/action_model': 6.92394673133858e-05, 'epoch': 1.25}) +[2026-06-17 21:06:24,437][starVLA.training.train_starvla][INFO] - Step 2760, Loss: {'train/loss': 0.5071715712547302, 'train/grad_norm_pre_clip': 17.473085403442383, 'timing/data': 0.014890567399561405, 'timing/model': 1.1449634497985244, 'learning_rate/qwen_vl_interface': 6.921922613057165e-06, 'learning_rate/action_model': 6.921922613057164e-05, 'epoch': 1.25}) +[2026-06-17 21:06:28,175][starVLA.training.train_starvla][INFO] - Step 2761, Loss: {'train/loss': 0.21557918190956116, 'train/grad_norm_pre_clip': 12.359721183776855, 'timing/data': 0.009331759065389633, 'timing/model': 0.9820880629122257, 'learning_rate/qwen_vl_interface': 6.9198981481843915e-06, 'learning_rate/action_model': 6.919898148184392e-05, 'epoch': 1.25}) +[2026-06-17 21:06:32,201][starVLA.training.train_starvla][INFO] - Step 2762, Loss: {'train/loss': 0.36284658312797546, 'train/grad_norm_pre_clip': 17.688980102539062, 'timing/data': 0.012649898417294025, 'timing/model': 0.9501161416992545, 'learning_rate/qwen_vl_interface': 6.917873337139936e-06, 'learning_rate/action_model': 6.917873337139937e-05, 'epoch': 1.25}) +[2026-06-17 21:06:36,008][starVLA.training.train_starvla][INFO] - Step 2763, Loss: {'train/loss': 0.5496010780334473, 'train/grad_norm_pre_clip': 20.482494354248047, 'timing/data': 0.006723265163600445, 'timing/model': 0.9463499737903476, 'learning_rate/qwen_vl_interface': 6.915848180343545e-06, 'learning_rate/action_model': 6.915848180343545e-05, 'epoch': 1.25}) +[2026-06-17 21:06:40,004][starVLA.training.train_starvla][INFO] - Step 2764, Loss: {'train/loss': 0.41404247283935547, 'train/grad_norm_pre_clip': 16.95015525817871, 'timing/data': 0.014716601930558681, 'timing/model': 0.9436669265851378, 'learning_rate/qwen_vl_interface': 6.913822678215031e-06, 'learning_rate/action_model': 6.913822678215031e-05, 'epoch': 1.25}) +[2026-06-17 21:06:43,832][starVLA.training.train_starvla][INFO] - Step 2765, Loss: {'train/loss': 0.4078788161277771, 'train/grad_norm_pre_clip': 14.166293144226074, 'timing/data': 0.008151902817189693, 'timing/model': 0.9339063735678792, 'learning_rate/qwen_vl_interface': 6.911796831174287e-06, 'learning_rate/action_model': 6.911796831174286e-05, 'epoch': 1.25}) +[2026-06-17 21:06:47,994][starVLA.training.train_starvla][INFO] - Step 2766, Loss: {'train/loss': 0.253976047039032, 'train/grad_norm_pre_clip': 14.452964782714844, 'timing/data': 0.008358737453818321, 'timing/model': 0.9511372884735465, 'learning_rate/qwen_vl_interface': 6.909770639641271e-06, 'learning_rate/action_model': 6.909770639641271e-05, 'epoch': 1.25}) +[2026-06-17 21:06:51,911][starVLA.training.train_starvla][INFO] - Step 2767, Loss: {'train/loss': 0.27503830194473267, 'train/grad_norm_pre_clip': 10.320405960083008, 'timing/data': 0.008687891066074371, 'timing/model': 1.2194214900955558, 'learning_rate/qwen_vl_interface': 6.907744104036015e-06, 'learning_rate/action_model': 6.907744104036015e-05, 'epoch': 1.26}) +[2026-06-17 21:06:55,563][starVLA.training.train_starvla][INFO] - Step 2768, Loss: {'train/loss': 0.5177661180496216, 'train/grad_norm_pre_clip': 18.7840576171875, 'timing/data': 0.008563002571463585, 'timing/model': 0.9696806725114584, 'learning_rate/qwen_vl_interface': 6.905717224778623e-06, 'learning_rate/action_model': 6.905717224778622e-05, 'epoch': 1.26}) +[2026-06-17 21:06:59,418][starVLA.training.train_starvla][INFO] - Step 2769, Loss: {'train/loss': 0.3441489636898041, 'train/grad_norm_pre_clip': 17.59630584716797, 'timing/data': 0.014843973331153393, 'timing/model': 0.9516241913661361, 'learning_rate/qwen_vl_interface': 6.903690002289268e-06, 'learning_rate/action_model': 6.903690002289267e-05, 'epoch': 1.26}) +[2026-06-17 21:07:03,448][starVLA.training.train_starvla][INFO] - Step 2770, Loss: {'train/loss': 0.421934574842453, 'train/grad_norm_pre_clip': 19.39775276184082, 'timing/data': 0.008027812466025352, 'timing/model': 0.9399032285436988, 'learning_rate/qwen_vl_interface': 6.901662436988199e-06, 'learning_rate/action_model': 6.901662436988199e-05, 'epoch': 1.26}) +[2026-06-17 21:07:08,628][starVLA.training.train_starvla][INFO] - Step 2771, Loss: {'train/loss': 0.2480292171239853, 'train/grad_norm_pre_clip': 10.13060188293457, 'timing/data': 0.01577700302004814, 'timing/model': 2.2328605335205793, 'learning_rate/qwen_vl_interface': 6.899634529295726e-06, 'learning_rate/action_model': 6.899634529295726e-05, 'epoch': 1.26}) +[2026-06-17 21:07:13,426][starVLA.training.train_starvla][INFO] - Step 2772, Loss: {'train/loss': 0.49405938386917114, 'train/grad_norm_pre_clip': 22.454538345336914, 'timing/data': 0.010446647182106972, 'timing/model': 1.3560845842584968, 'learning_rate/qwen_vl_interface': 6.897606279632241e-06, 'learning_rate/action_model': 6.897606279632241e-05, 'epoch': 1.26}) +[2026-06-17 21:07:17,561][starVLA.training.train_starvla][INFO] - Step 2773, Loss: {'train/loss': 0.270347535610199, 'train/grad_norm_pre_clip': 15.330190658569336, 'timing/data': 0.008398822508752346, 'timing/model': 1.417134934104979, 'learning_rate/qwen_vl_interface': 6.895577688418202e-06, 'learning_rate/action_model': 6.895577688418202e-05, 'epoch': 1.26}) +[2026-06-17 21:07:21,494][starVLA.training.train_starvla][INFO] - Step 2774, Loss: {'train/loss': 0.5681242346763611, 'train/grad_norm_pre_clip': 22.644649505615234, 'timing/data': 0.015684853307902813, 'timing/model': 1.157824182882905, 'learning_rate/qwen_vl_interface': 6.893548756074136e-06, 'learning_rate/action_model': 6.893548756074136e-05, 'epoch': 1.26}) +[2026-06-17 21:07:25,141][starVLA.training.train_starvla][INFO] - Step 2775, Loss: {'train/loss': 0.7300020456314087, 'train/grad_norm_pre_clip': 19.904325485229492, 'timing/data': 0.00859732087701559, 'timing/model': 0.9385401364415884, 'learning_rate/qwen_vl_interface': 6.891519483020647e-06, 'learning_rate/action_model': 6.891519483020647e-05, 'epoch': 1.26}) +[2026-06-17 21:07:29,358][starVLA.training.train_starvla][INFO] - Step 2776, Loss: {'train/loss': 0.1792498081922531, 'train/grad_norm_pre_clip': 11.79388427734375, 'timing/data': 0.010843954049050808, 'timing/model': 0.9428643556311727, 'learning_rate/qwen_vl_interface': 6.8894898696784e-06, 'learning_rate/action_model': 6.889489869678399e-05, 'epoch': 1.26}) +[2026-06-17 21:07:33,254][starVLA.training.train_starvla][INFO] - Step 2777, Loss: {'train/loss': 0.2579863667488098, 'train/grad_norm_pre_clip': 12.780747413635254, 'timing/data': 0.009129010140895844, 'timing/model': 0.9605150585994124, 'learning_rate/qwen_vl_interface': 6.887459916468141e-06, 'learning_rate/action_model': 6.88745991646814e-05, 'epoch': 1.26}) +[2026-06-17 21:07:37,145][starVLA.training.train_starvla][INFO] - Step 2778, Loss: {'train/loss': 0.6460657119750977, 'train/grad_norm_pre_clip': 23.114805221557617, 'timing/data': 0.009056183509528637, 'timing/model': 0.9422193365171552, 'learning_rate/qwen_vl_interface': 6.885429623810678e-06, 'learning_rate/action_model': 6.885429623810677e-05, 'epoch': 1.26}) +[2026-06-17 21:07:41,198][starVLA.training.train_starvla][INFO] - Step 2779, Loss: {'train/loss': 0.25575029850006104, 'train/grad_norm_pre_clip': 12.026495933532715, 'timing/data': 0.008002929389476776, 'timing/model': 0.9517910722643137, 'learning_rate/qwen_vl_interface': 6.883398992126894e-06, 'learning_rate/action_model': 6.883398992126894e-05, 'epoch': 1.26}) +[2026-06-17 21:07:45,051][starVLA.training.train_starvla][INFO] - Step 2780, Loss: {'train/loss': 0.3211841285228729, 'train/grad_norm_pre_clip': 11.848576545715332, 'timing/data': 0.020694445818662643, 'timing/model': 1.1205723090097308, 'learning_rate/qwen_vl_interface': 6.8813680218377425e-06, 'learning_rate/action_model': 6.881368021837741e-05, 'epoch': 1.26}) +[2026-06-17 21:07:48,870][starVLA.training.train_starvla][INFO] - Step 2781, Loss: {'train/loss': 0.3020867109298706, 'train/grad_norm_pre_clip': 13.704794883728027, 'timing/data': 0.0092499740421772, 'timing/model': 1.088109988719225, 'learning_rate/qwen_vl_interface': 6.8793367133642445e-06, 'learning_rate/action_model': 6.879336713364243e-05, 'epoch': 1.26}) +[2026-06-17 21:07:52,595][starVLA.training.train_starvla][INFO] - Step 2782, Loss: {'train/loss': 0.4131709039211273, 'train/grad_norm_pre_clip': 19.116962432861328, 'timing/data': 0.007825462147593498, 'timing/model': 1.0123316198587418, 'learning_rate/qwen_vl_interface': 6.877305067127492e-06, 'learning_rate/action_model': 6.877305067127492e-05, 'epoch': 1.26}) +[2026-06-17 21:07:56,437][starVLA.training.train_starvla][INFO] - Step 2783, Loss: {'train/loss': 0.6690549850463867, 'train/grad_norm_pre_clip': 21.048564910888672, 'timing/data': 0.008313365280628204, 'timing/model': 0.9348862189799547, 'learning_rate/qwen_vl_interface': 6.875273083548648e-06, 'learning_rate/action_model': 6.875273083548648e-05, 'epoch': 1.26}) +[2026-06-17 21:08:00,238][starVLA.training.train_starvla][INFO] - Step 2784, Loss: {'train/loss': 0.422465980052948, 'train/grad_norm_pre_clip': 27.467260360717773, 'timing/data': 0.008220166899263859, 'timing/model': 0.9428375931456685, 'learning_rate/qwen_vl_interface': 6.873240763048945e-06, 'learning_rate/action_model': 6.873240763048945e-05, 'epoch': 1.26}) +[2026-06-17 21:08:03,946][starVLA.training.train_starvla][INFO] - Step 2785, Loss: {'train/loss': 0.34275493025779724, 'train/grad_norm_pre_clip': 16.62221908569336, 'timing/data': 0.008810261264443398, 'timing/model': 0.9348170971497893, 'learning_rate/qwen_vl_interface': 6.871208106049685e-06, 'learning_rate/action_model': 6.871208106049685e-05, 'epoch': 1.26}) +[2026-06-17 21:08:07,810][starVLA.training.train_starvla][INFO] - Step 2786, Loss: {'train/loss': 0.40424779057502747, 'train/grad_norm_pre_clip': 15.819746017456055, 'timing/data': 0.00904780812561512, 'timing/model': 0.9588342886418104, 'learning_rate/qwen_vl_interface': 6.869175112972239e-06, 'learning_rate/action_model': 6.86917511297224e-05, 'epoch': 1.26}) +[2026-06-17 21:08:11,721][starVLA.training.train_starvla][INFO] - Step 2787, Loss: {'train/loss': 0.15501244366168976, 'train/grad_norm_pre_clip': 11.630967140197754, 'timing/data': 0.00994423683732748, 'timing/model': 1.1329042138531804, 'learning_rate/qwen_vl_interface': 6.867141784238052e-06, 'learning_rate/action_model': 6.867141784238051e-05, 'epoch': 1.26}) +[2026-06-17 21:08:15,583][starVLA.training.train_starvla][INFO] - Step 2788, Loss: {'train/loss': 0.503213107585907, 'train/grad_norm_pre_clip': 21.238420486450195, 'timing/data': 0.007119867950677872, 'timing/model': 0.9513209508731961, 'learning_rate/qwen_vl_interface': 6.865108120268632e-06, 'learning_rate/action_model': 6.865108120268632e-05, 'epoch': 1.26}) +[2026-06-17 21:08:19,443][starVLA.training.train_starvla][INFO] - Step 2789, Loss: {'train/loss': 0.2413690686225891, 'train/grad_norm_pre_clip': 10.92878246307373, 'timing/data': 0.009094160050153732, 'timing/model': 0.9411640651524067, 'learning_rate/qwen_vl_interface': 6.863074121485561e-06, 'learning_rate/action_model': 6.86307412148556e-05, 'epoch': 1.27}) +[2026-06-17 21:08:23,256][starVLA.training.train_starvla][INFO] - Step 2790, Loss: {'train/loss': 0.22645904123783112, 'train/grad_norm_pre_clip': 10.927647590637207, 'timing/data': 0.008471885696053505, 'timing/model': 0.9501548707485199, 'learning_rate/qwen_vl_interface': 6.861039788310489e-06, 'learning_rate/action_model': 6.861039788310489e-05, 'epoch': 1.27}) +[2026-06-17 21:08:27,200][starVLA.training.train_starvla][INFO] - Step 2791, Loss: {'train/loss': 0.08867427706718445, 'train/grad_norm_pre_clip': 6.681674957275391, 'timing/data': 0.010039483197033405, 'timing/model': 0.9373935917392373, 'learning_rate/qwen_vl_interface': 6.859005121165135e-06, 'learning_rate/action_model': 6.859005121165135e-05, 'epoch': 1.27}) +[2026-06-17 21:08:32,037][starVLA.training.train_starvla][INFO] - Step 2792, Loss: {'train/loss': 0.1821969449520111, 'train/grad_norm_pre_clip': 12.642040252685547, 'timing/data': 0.014074220322072506, 'timing/model': 0.9503699578344822, 'learning_rate/qwen_vl_interface': 6.856970120471289e-06, 'learning_rate/action_model': 6.856970120471289e-05, 'epoch': 1.27}) +[2026-06-17 21:08:36,003][starVLA.training.train_starvla][INFO] - Step 2793, Loss: {'train/loss': 0.5130712985992432, 'train/grad_norm_pre_clip': 18.871252059936523, 'timing/data': 0.019758261740207672, 'timing/model': 1.2332478165626526, 'learning_rate/qwen_vl_interface': 6.854934786650805e-06, 'learning_rate/action_model': 6.854934786650806e-05, 'epoch': 1.27}) +[2026-06-17 21:08:41,401][starVLA.training.train_starvla][INFO] - Step 2794, Loss: {'train/loss': 0.6261273622512817, 'train/grad_norm_pre_clip': 21.443038940429688, 'timing/data': 0.010211378335952759, 'timing/model': 1.149816570803523, 'learning_rate/qwen_vl_interface': 6.852899120125615e-06, 'learning_rate/action_model': 6.852899120125616e-05, 'epoch': 1.27}) +[2026-06-17 21:08:45,069][starVLA.training.train_starvla][INFO] - Step 2795, Loss: {'train/loss': 0.6434154510498047, 'train/grad_norm_pre_clip': 39.53977966308594, 'timing/data': 0.011899464763700962, 'timing/model': 0.9496557908132672, 'learning_rate/qwen_vl_interface': 6.850863121317712e-06, 'learning_rate/action_model': 6.850863121317712e-05, 'epoch': 1.27}) +[2026-06-17 21:08:49,588][starVLA.training.train_starvla][INFO] - Step 2796, Loss: {'train/loss': 0.4771682620048523, 'train/grad_norm_pre_clip': 19.352142333984375, 'timing/data': 0.008376211859285831, 'timing/model': 0.9584851088002324, 'learning_rate/qwen_vl_interface': 6.848826790649162e-06, 'learning_rate/action_model': 6.848826790649163e-05, 'epoch': 1.27}) +[2026-06-17 21:08:54,123][starVLA.training.train_starvla][INFO] - Step 2797, Loss: {'train/loss': 0.21838390827178955, 'train/grad_norm_pre_clip': 12.187714576721191, 'timing/data': 0.009942452423274517, 'timing/model': 0.9517222587019205, 'learning_rate/qwen_vl_interface': 6.846790128542098e-06, 'learning_rate/action_model': 6.846790128542098e-05, 'epoch': 1.27}) +[2026-06-17 21:08:57,929][starVLA.training.train_starvla][INFO] - Step 2798, Loss: {'train/loss': 0.24787510931491852, 'train/grad_norm_pre_clip': 13.652703285217285, 'timing/data': 0.012120857834815979, 'timing/model': 0.9643631707876921, 'learning_rate/qwen_vl_interface': 6.8447531354187214e-06, 'learning_rate/action_model': 6.844753135418722e-05, 'epoch': 1.27}) +[2026-06-17 21:09:01,920][starVLA.training.train_starvla][INFO] - Step 2799, Loss: {'train/loss': 0.31153252720832825, 'train/grad_norm_pre_clip': 12.507745742797852, 'timing/data': 0.009715153835713863, 'timing/model': 0.9863274963572621, 'learning_rate/qwen_vl_interface': 6.842715811701305e-06, 'learning_rate/action_model': 6.842715811701304e-05, 'epoch': 1.27}) +[2026-06-17 21:11:18,359][starVLA.training.train_starvla][INFO] - Step 2800, Loss: {'train/loss': 0.8455542922019958, 'train/grad_norm_pre_clip': 25.617345809936523, 'eval/loss': 0.3898636996746063, 'eval/latency_8/loss': 0.3898636996746063, 'eval/demon_fire_precision': 0.9424698795180723, 'eval/demon_fire_recall': 0.9720410065237651, 'eval/demon_fire_f1': 0.9570270683590762, 'eval/demon_fire_rate_model': 0.9237618252643295, 'eval/demon_fire_rate_teacher': 0.8956594323873122, 'eval/demon_left_precision': 0.7749419953596288, 'eval/demon_left_recall': 0.715203426124197, 'eval/demon_left_f1': 0.7438752783964366, 'eval/demon_left_rate_model': 0.11992209237618252, 'eval/demon_left_rate_teacher': 0.12993878686700056, 'eval/demon_right_precision': 0.7417974322396577, 'eval/demon_right_recall': 0.7878787878787878, 'eval/demon_right_f1': 0.7641440117560617, 'eval/demon_right_rate_model': 0.19504730105731774, 'eval/demon_right_rate_teacher': 0.18363939899833054, 'eval/demon_move_f1': 0.7540096450762492, 'eval/demon_cc_f1': 0.8855006392467565, 'eval/latency_8/demon_fire_precision': 0.9424698795180723, 'eval/latency_8/demon_fire_recall': 0.9720410065237651, 'eval/latency_8/demon_fire_f1': 0.9570270683590762, 'eval/latency_8/demon_fire_rate_model': 0.9237618252643295, 'eval/latency_8/demon_fire_rate_teacher': 0.8956594323873122, 'eval/latency_8/demon_left_precision': 0.7749419953596288, 'eval/latency_8/demon_left_recall': 0.715203426124197, 'eval/latency_8/demon_left_f1': 0.7438752783964366, 'eval/latency_8/demon_left_rate_model': 0.11992209237618252, 'eval/latency_8/demon_left_rate_teacher': 0.12993878686700056, 'eval/latency_8/demon_right_precision': 0.7417974322396577, 'eval/latency_8/demon_right_recall': 0.7878787878787878, 'eval/latency_8/demon_right_f1': 0.7641440117560617, 'eval/latency_8/demon_right_rate_model': 0.19504730105731774, 'eval/latency_8/demon_right_rate_teacher': 0.18363939899833054, 'eval/latency_8/demon_move_f1': 0.7540096450762492, 'eval/latency_8/demon_cc_f1': 0.8855006392467565, 'timing/data': 0.00807640329003334, 'timing/model': 1.122043733485043, 'learning_rate/qwen_vl_interface': 6.840678157812184e-06, 'learning_rate/action_model': 6.840678157812184e-05, 'epoch': 1.27}) +[2026-06-17 21:11:18,363][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_2800_state +[2026-06-17 21:11:18,367][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 21:11:25,356][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_2800_state/model.safetensors +[2026-06-17 21:11:41,353][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_2800_state/optimizer.bin +[2026-06-17 21:11:41,354][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_2800_state/sampler.bin +[2026-06-17 21:11:41,354][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_2800_state/sampler_1.bin +[2026-06-17 21:11:41,357][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_2800_state/random_states_0.pkl +[2026-06-17 21:11:41,360][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2800 +[2026-06-17 21:11:41,360][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 21:11:41,371][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 21:11:45,290][starVLA.training.train_starvla][INFO] - Step 2801, Loss: {'train/loss': 0.33135974407196045, 'train/grad_norm_pre_clip': 12.46490478515625, 'timing/data': 0.013159743510186672, 'timing/model': 0.938574256375432, 'learning_rate/qwen_vl_interface': 6.8386401741737725e-06, 'learning_rate/action_model': 6.838640174173773e-05, 'epoch': 1.27}) +[2026-06-17 21:11:49,223][starVLA.training.train_starvla][INFO] - Step 2802, Loss: {'train/loss': 0.28051698207855225, 'train/grad_norm_pre_clip': 14.78930377960205, 'timing/data': 0.01236910093575716, 'timing/model': 1.2293547037988901, 'learning_rate/qwen_vl_interface': 6.836601861208542e-06, 'learning_rate/action_model': 6.836601861208541e-05, 'epoch': 1.27}) +[2026-06-17 21:11:53,148][starVLA.training.train_starvla][INFO] - Step 2803, Loss: {'train/loss': 0.2858567535877228, 'train/grad_norm_pre_clip': 16.757671356201172, 'timing/data': 0.010176380164921284, 'timing/model': 1.1486678048968315, 'learning_rate/qwen_vl_interface': 6.834563219339039e-06, 'learning_rate/action_model': 6.834563219339038e-05, 'epoch': 1.27}) +[2026-06-17 21:11:56,843][starVLA.training.train_starvla][INFO] - Step 2804, Loss: {'train/loss': 0.3146604001522064, 'train/grad_norm_pre_clip': 14.205718994140625, 'timing/data': 0.01438398938626051, 'timing/model': 0.9328708900138736, 'learning_rate/qwen_vl_interface': 6.832524248987874e-06, 'learning_rate/action_model': 6.832524248987874e-05, 'epoch': 1.27}) +[2026-06-17 21:12:00,636][starVLA.training.train_starvla][INFO] - Step 2805, Loss: {'train/loss': 0.36984387040138245, 'train/grad_norm_pre_clip': 16.191701889038086, 'timing/data': 0.01126839593052864, 'timing/model': 0.9407399734482169, 'learning_rate/qwen_vl_interface': 6.8304849505777314e-06, 'learning_rate/action_model': 6.830484950577732e-05, 'epoch': 1.27}) +[2026-06-17 21:12:04,485][starVLA.training.train_starvla][INFO] - Step 2806, Loss: {'train/loss': 0.6229298710823059, 'train/grad_norm_pre_clip': 22.011445999145508, 'timing/data': 0.012327441945672035, 'timing/model': 0.9453151347115636, 'learning_rate/qwen_vl_interface': 6.828445324531357e-06, 'learning_rate/action_model': 6.828445324531357e-05, 'epoch': 1.27}) +[2026-06-17 21:12:08,333][starVLA.training.train_starvla][INFO] - Step 2807, Loss: {'train/loss': 0.3421134650707245, 'train/grad_norm_pre_clip': 19.27932357788086, 'timing/data': 0.016281126998364925, 'timing/model': 0.9608073169365525, 'learning_rate/qwen_vl_interface': 6.826405371271567e-06, 'learning_rate/action_model': 6.826405371271567e-05, 'epoch': 1.27}) +[2026-06-17 21:12:12,193][starVLA.training.train_starvla][INFO] - Step 2808, Loss: {'train/loss': 0.18582668900489807, 'train/grad_norm_pre_clip': 10.61572551727295, 'timing/data': 0.007352408021688461, 'timing/model': 0.9479972971603274, 'learning_rate/qwen_vl_interface': 6.82436509122125e-06, 'learning_rate/action_model': 6.82436509122125e-05, 'epoch': 1.27}) +[2026-06-17 21:12:16,091][starVLA.training.train_starvla][INFO] - Step 2809, Loss: {'train/loss': 0.507728099822998, 'train/grad_norm_pre_clip': 32.75663757324219, 'timing/data': 0.00898712407797575, 'timing/model': 1.1361423563212156, 'learning_rate/qwen_vl_interface': 6.822324484803353e-06, 'learning_rate/action_model': 6.822324484803353e-05, 'epoch': 1.27}) +[2026-06-17 21:12:19,750][starVLA.training.train_starvla][INFO] - Step 2810, Loss: {'train/loss': 0.26386287808418274, 'train/grad_norm_pre_clip': 20.066415786743164, 'timing/data': 0.0074990540742874146, 'timing/model': 0.9382423209026456, 'learning_rate/qwen_vl_interface': 6.8202835524409e-06, 'learning_rate/action_model': 6.8202835524409e-05, 'epoch': 1.27}) +[2026-06-17 21:12:23,654][starVLA.training.train_starvla][INFO] - Step 2811, Loss: {'train/loss': 0.26343345642089844, 'train/grad_norm_pre_clip': 19.43131446838379, 'timing/data': 0.008327343501150608, 'timing/model': 0.9563173055648804, 'learning_rate/qwen_vl_interface': 6.818242294556977e-06, 'learning_rate/action_model': 6.818242294556977e-05, 'epoch': 1.28}) +[2026-06-17 21:12:28,458][starVLA.training.train_starvla][INFO] - Step 2812, Loss: {'train/loss': 0.3610892593860626, 'train/grad_norm_pre_clip': 19.647817611694336, 'timing/data': 0.009364290162920952, 'timing/model': 0.9399349857121706, 'learning_rate/qwen_vl_interface': 6.816200711574738e-06, 'learning_rate/action_model': 6.816200711574738e-05, 'epoch': 1.28}) +[2026-06-17 21:12:34,202][starVLA.training.train_starvla][INFO] - Step 2813, Loss: {'train/loss': 0.2472926527261734, 'train/grad_norm_pre_clip': 15.245096206665039, 'timing/data': 0.010172534734010696, 'timing/model': 0.9501362647861242, 'learning_rate/qwen_vl_interface': 6.814158803917406e-06, 'learning_rate/action_model': 6.814158803917406e-05, 'epoch': 1.28}) +[2026-06-17 21:12:39,047][starVLA.training.train_starvla][INFO] - Step 2814, Loss: {'train/loss': 0.310639351606369, 'train/grad_norm_pre_clip': 13.445556640625, 'timing/data': 0.009594659321010113, 'timing/model': 1.397296529263258, 'learning_rate/qwen_vl_interface': 6.812116572008271e-06, 'learning_rate/action_model': 6.812116572008271e-05, 'epoch': 1.28}) +[2026-06-17 21:12:42,944][starVLA.training.train_starvla][INFO] - Step 2815, Loss: {'train/loss': 0.5319316983222961, 'train/grad_norm_pre_clip': 21.75752830505371, 'timing/data': 0.009038305841386318, 'timing/model': 1.1442960975691676, 'learning_rate/qwen_vl_interface': 6.8100740162706906e-06, 'learning_rate/action_model': 6.81007401627069e-05, 'epoch': 1.28}) +[2026-06-17 21:12:46,819][starVLA.training.train_starvla][INFO] - Step 2816, Loss: {'train/loss': 0.14315535128116608, 'train/grad_norm_pre_clip': 11.063981056213379, 'timing/data': 0.009977824985980988, 'timing/model': 1.125627065077424, 'learning_rate/qwen_vl_interface': 6.808031137128087e-06, 'learning_rate/action_model': 6.808031137128086e-05, 'epoch': 1.28}) +[2026-06-17 21:12:50,956][starVLA.training.train_starvla][INFO] - Step 2817, Loss: {'train/loss': 0.25197187066078186, 'train/grad_norm_pre_clip': 13.795027732849121, 'timing/data': 0.008165108039975166, 'timing/model': 0.9365157810971141, 'learning_rate/qwen_vl_interface': 6.805987935003951e-06, 'learning_rate/action_model': 6.805987935003951e-05, 'epoch': 1.28}) +[2026-06-17 21:12:54,852][starVLA.training.train_starvla][INFO] - Step 2818, Loss: {'train/loss': 0.19953270256519318, 'train/grad_norm_pre_clip': 10.336109161376953, 'timing/data': 0.007604918442666531, 'timing/model': 0.942836937494576, 'learning_rate/qwen_vl_interface': 6.803944410321843e-06, 'learning_rate/action_model': 6.803944410321844e-05, 'epoch': 1.28}) +[2026-06-17 21:12:58,739][starVLA.training.train_starvla][INFO] - Step 2819, Loss: {'train/loss': 0.345416784286499, 'train/grad_norm_pre_clip': 19.823501586914062, 'timing/data': 0.009297630749642849, 'timing/model': 0.9391993619501591, 'learning_rate/qwen_vl_interface': 6.801900563505388e-06, 'learning_rate/action_model': 6.801900563505388e-05, 'epoch': 1.28}) +[2026-06-17 21:13:02,556][starVLA.training.train_starvla][INFO] - Step 2820, Loss: {'train/loss': 0.5056421160697937, 'train/grad_norm_pre_clip': 21.240320205688477, 'timing/data': 0.014212778769433498, 'timing/model': 0.9342294931411743, 'learning_rate/qwen_vl_interface': 6.7998563949782755e-06, 'learning_rate/action_model': 6.799856394978276e-05, 'epoch': 1.28}) +[2026-06-17 21:13:06,452][starVLA.training.train_starvla][INFO] - Step 2821, Loss: {'train/loss': 0.4048221707344055, 'train/grad_norm_pre_clip': 17.033279418945312, 'timing/data': 0.01294303871691227, 'timing/model': 0.9472780907526612, 'learning_rate/qwen_vl_interface': 6.797811905164263e-06, 'learning_rate/action_model': 6.797811905164263e-05, 'epoch': 1.28}) +[2026-06-17 21:13:10,432][starVLA.training.train_starvla][INFO] - Step 2822, Loss: {'train/loss': 0.8399181365966797, 'train/grad_norm_pre_clip': 23.892580032348633, 'timing/data': 0.008061757311224937, 'timing/model': 1.1365970931947231, 'learning_rate/qwen_vl_interface': 6.795767094487179e-06, 'learning_rate/action_model': 6.795767094487179e-05, 'epoch': 1.28}) +[2026-06-17 21:13:14,154][starVLA.training.train_starvla][INFO] - Step 2823, Loss: {'train/loss': 0.1686801314353943, 'train/grad_norm_pre_clip': 11.433355331420898, 'timing/data': 0.010868092998862267, 'timing/model': 0.941898662596941, 'learning_rate/qwen_vl_interface': 6.793721963370913e-06, 'learning_rate/action_model': 6.793721963370913e-05, 'epoch': 1.28}) +[2026-06-17 21:13:18,069][starVLA.training.train_starvla][INFO] - Step 2824, Loss: {'train/loss': 0.20680806040763855, 'train/grad_norm_pre_clip': 13.789969444274902, 'timing/data': 0.006804721429944038, 'timing/model': 0.9435025211423635, 'learning_rate/qwen_vl_interface': 6.7916765122394215e-06, 'learning_rate/action_model': 6.791676512239421e-05, 'epoch': 1.28}) +[2026-06-17 21:13:21,962][starVLA.training.train_starvla][INFO] - Step 2825, Loss: {'train/loss': 0.44333237409591675, 'train/grad_norm_pre_clip': 17.866043090820312, 'timing/data': 0.00965743325650692, 'timing/model': 0.9441101383417845, 'learning_rate/qwen_vl_interface': 6.7896307415167305e-06, 'learning_rate/action_model': 6.78963074151673e-05, 'epoch': 1.28}) +[2026-06-17 21:13:27,087][starVLA.training.train_starvla][INFO] - Step 2826, Loss: {'train/loss': 0.3005842864513397, 'train/grad_norm_pre_clip': 21.250682830810547, 'timing/data': 0.008344534784555435, 'timing/model': 0.9910396449267864, 'learning_rate/qwen_vl_interface': 6.787584651626929e-06, 'learning_rate/action_model': 6.78758465162693e-05, 'epoch': 1.28}) +[2026-06-17 21:13:31,024][starVLA.training.train_starvla][INFO] - Step 2827, Loss: {'train/loss': 0.5851831436157227, 'train/grad_norm_pre_clip': 17.422012329101562, 'timing/data': 0.014739214442670345, 'timing/model': 0.9718894278630614, 'learning_rate/qwen_vl_interface': 6.785538242994177e-06, 'learning_rate/action_model': 6.785538242994177e-05, 'epoch': 1.28}) +[2026-06-17 21:13:34,948][starVLA.training.train_starvla][INFO] - Step 2828, Loss: {'train/loss': 0.3278558850288391, 'train/grad_norm_pre_clip': 14.699625968933105, 'timing/data': 0.01072040107101202, 'timing/model': 0.9484193455427885, 'learning_rate/qwen_vl_interface': 6.783491516042694e-06, 'learning_rate/action_model': 6.783491516042693e-05, 'epoch': 1.28}) +[2026-06-17 21:13:39,441][starVLA.training.train_starvla][INFO] - Step 2829, Loss: {'train/loss': 0.17989946901798248, 'train/grad_norm_pre_clip': 16.7259578704834, 'timing/data': 0.00928458571434021, 'timing/model': 1.7032362930476665, 'learning_rate/qwen_vl_interface': 6.781444471196768e-06, 'learning_rate/action_model': 6.781444471196769e-05, 'epoch': 1.28}) +[2026-06-17 21:13:43,368][starVLA.training.train_starvla][INFO] - Step 2830, Loss: {'train/loss': 0.27344319224357605, 'train/grad_norm_pre_clip': 11.9006929397583, 'timing/data': 0.009649229235947132, 'timing/model': 0.9366091107949615, 'learning_rate/qwen_vl_interface': 6.7793971088807575e-06, 'learning_rate/action_model': 6.779397108880757e-05, 'epoch': 1.28}) +[2026-06-17 21:13:47,232][starVLA.training.train_starvla][INFO] - Step 2831, Loss: {'train/loss': 0.3850567042827606, 'train/grad_norm_pre_clip': 20.35279655456543, 'timing/data': 0.013068889267742634, 'timing/model': 0.9546849643811584, 'learning_rate/qwen_vl_interface': 6.77734942951908e-06, 'learning_rate/action_model': 6.77734942951908e-05, 'epoch': 1.28}) +[2026-06-17 21:13:51,047][starVLA.training.train_starvla][INFO] - Step 2832, Loss: {'train/loss': 0.5637616515159607, 'train/grad_norm_pre_clip': 29.228515625, 'timing/data': 0.008896825835108757, 'timing/model': 0.939035783521831, 'learning_rate/qwen_vl_interface': 6.7753014335362224e-06, 'learning_rate/action_model': 6.775301433536222e-05, 'epoch': 1.28}) +[2026-06-17 21:13:54,927][starVLA.training.train_starvla][INFO] - Step 2833, Loss: {'train/loss': 0.21158504486083984, 'train/grad_norm_pre_clip': 16.886075973510742, 'timing/data': 0.012791854329407215, 'timing/model': 0.9494888484477997, 'learning_rate/qwen_vl_interface': 6.773253121356735e-06, 'learning_rate/action_model': 6.773253121356735e-05, 'epoch': 1.29}) +[2026-06-17 21:13:58,781][starVLA.training.train_starvla][INFO] - Step 2834, Loss: {'train/loss': 0.18401354551315308, 'train/grad_norm_pre_clip': 9.136256217956543, 'timing/data': 0.01612717565149069, 'timing/model': 0.946893360465765, 'learning_rate/qwen_vl_interface': 6.771204493405239e-06, 'learning_rate/action_model': 6.771204493405239e-05, 'epoch': 1.29}) +[2026-06-17 21:14:02,650][starVLA.training.train_starvla][INFO] - Step 2835, Loss: {'train/loss': 0.46943214535713196, 'train/grad_norm_pre_clip': 20.797561645507812, 'timing/data': 0.00815643835812807, 'timing/model': 1.1619148710742593, 'learning_rate/qwen_vl_interface': 6.769155550106413e-06, 'learning_rate/action_model': 6.769155550106413e-05, 'epoch': 1.29}) +[2026-06-17 21:14:06,527][starVLA.training.train_starvla][INFO] - Step 2836, Loss: {'train/loss': 0.3307011127471924, 'train/grad_norm_pre_clip': 11.978911399841309, 'timing/data': 0.008488057181239128, 'timing/model': 1.123744947835803, 'learning_rate/qwen_vl_interface': 6.767106291885006e-06, 'learning_rate/action_model': 6.767106291885006e-05, 'epoch': 1.29}) +[2026-06-17 21:14:10,298][starVLA.training.train_starvla][INFO] - Step 2837, Loss: {'train/loss': 0.36411920189857483, 'train/grad_norm_pre_clip': 15.666892051696777, 'timing/data': 0.009916989132761955, 'timing/model': 1.035709804855287, 'learning_rate/qwen_vl_interface': 6.765056719165833e-06, 'learning_rate/action_model': 6.765056719165834e-05, 'epoch': 1.29}) +[2026-06-17 21:14:14,240][starVLA.training.train_starvla][INFO] - Step 2838, Loss: {'train/loss': 0.23889844119548798, 'train/grad_norm_pre_clip': 13.875798225402832, 'timing/data': 0.01620271150022745, 'timing/model': 0.9654953088611364, 'learning_rate/qwen_vl_interface': 6.763006832373772e-06, 'learning_rate/action_model': 6.763006832373771e-05, 'epoch': 1.29}) +[2026-06-17 21:14:18,041][starVLA.training.train_starvla][INFO] - Step 2839, Loss: {'train/loss': 0.5531703233718872, 'train/grad_norm_pre_clip': 14.687253952026367, 'timing/data': 0.013198875822126865, 'timing/model': 0.9445035718381405, 'learning_rate/qwen_vl_interface': 6.760956631933766e-06, 'learning_rate/action_model': 6.760956631933765e-05, 'epoch': 1.29}) +[2026-06-17 21:14:21,842][starVLA.training.train_starvla][INFO] - Step 2840, Loss: {'train/loss': 0.4935973584651947, 'train/grad_norm_pre_clip': 21.786096572875977, 'timing/data': 0.009151908569037914, 'timing/model': 0.9403387028723955, 'learning_rate/qwen_vl_interface': 6.758906118270826e-06, 'learning_rate/action_model': 6.758906118270825e-05, 'epoch': 1.29}) +[2026-06-17 21:14:25,827][starVLA.training.train_starvla][INFO] - Step 2841, Loss: {'train/loss': 0.25607830286026, 'train/grad_norm_pre_clip': 11.244611740112305, 'timing/data': 0.010707051493227482, 'timing/model': 0.9601695733144879, 'learning_rate/qwen_vl_interface': 6.756855291810024e-06, 'learning_rate/action_model': 6.756855291810024e-05, 'epoch': 1.29}) +[2026-06-17 21:14:29,711][starVLA.training.train_starvla][INFO] - Step 2842, Loss: {'train/loss': 0.3582603633403778, 'train/grad_norm_pre_clip': 16.320358276367188, 'timing/data': 0.013728633522987366, 'timing/model': 1.137407454662025, 'learning_rate/qwen_vl_interface': 6.754804152976495e-06, 'learning_rate/action_model': 6.754804152976495e-05, 'epoch': 1.29}) +[2026-06-17 21:14:33,567][starVLA.training.train_starvla][INFO] - Step 2843, Loss: {'train/loss': 0.5658631920814514, 'train/grad_norm_pre_clip': 36.113948822021484, 'timing/data': 0.011108620092272758, 'timing/model': 1.1485930541530252, 'learning_rate/qwen_vl_interface': 6.752752702195448e-06, 'learning_rate/action_model': 6.752752702195448e-05, 'epoch': 1.29}) +[2026-06-17 21:14:38,691][starVLA.training.train_starvla][INFO] - Step 2844, Loss: {'train/loss': 0.5698527097702026, 'train/grad_norm_pre_clip': 18.55857276916504, 'timing/data': 0.012913553975522518, 'timing/model': 0.9648992186412215, 'learning_rate/qwen_vl_interface': 6.750700939892151e-06, 'learning_rate/action_model': 6.75070093989215e-05, 'epoch': 1.29}) +[2026-06-17 21:14:43,344][starVLA.training.train_starvla][INFO] - Step 2845, Loss: {'train/loss': 0.7559092044830322, 'train/grad_norm_pre_clip': 22.572484970092773, 'timing/data': 0.01664184220135212, 'timing/model': 0.9465717179700732, 'learning_rate/qwen_vl_interface': 6.748648866491931e-06, 'learning_rate/action_model': 6.748648866491932e-05, 'epoch': 1.29}) +[2026-06-17 21:14:47,997][starVLA.training.train_starvla][INFO] - Step 2846, Loss: {'train/loss': 0.5414098501205444, 'train/grad_norm_pre_clip': 19.284015655517578, 'timing/data': 0.010541324503719807, 'timing/model': 1.0305809983983636, 'learning_rate/qwen_vl_interface': 6.7465964824201915e-06, 'learning_rate/action_model': 6.746596482420192e-05, 'epoch': 1.29}) +[2026-06-17 21:14:51,814][starVLA.training.train_starvla][INFO] - Step 2847, Loss: {'train/loss': 0.32996514439582825, 'train/grad_norm_pre_clip': 14.319007873535156, 'timing/data': 0.009803971275687218, 'timing/model': 0.9576428011059761, 'learning_rate/qwen_vl_interface': 6.744543788102388e-06, 'learning_rate/action_model': 6.744543788102388e-05, 'epoch': 1.29}) +[2026-06-17 21:14:55,834][starVLA.training.train_starvla][INFO] - Step 2848, Loss: {'train/loss': 0.3299486041069031, 'train/grad_norm_pre_clip': 13.832608222961426, 'timing/data': 0.010272938758134842, 'timing/model': 0.962965409271419, 'learning_rate/qwen_vl_interface': 6.742490783964051e-06, 'learning_rate/action_model': 6.742490783964051e-05, 'epoch': 1.29}) +[2026-06-17 21:14:59,847][starVLA.training.train_starvla][INFO] - Step 2849, Loss: {'train/loss': 0.48604974150657654, 'train/grad_norm_pre_clip': 18.602458953857422, 'timing/data': 0.01328002568334341, 'timing/model': 1.1512789763510227, 'learning_rate/qwen_vl_interface': 6.740437470430766e-06, 'learning_rate/action_model': 6.740437470430766e-05, 'epoch': 1.29}) +[2026-06-17 21:15:04,736][starVLA.training.train_starvla][INFO] - Step 2850, Loss: {'train/loss': 0.4285288453102112, 'train/grad_norm_pre_clip': 15.958891868591309, 'timing/data': 0.00857024546712637, 'timing/model': 0.9528881246224046, 'learning_rate/qwen_vl_interface': 6.7383838479281915e-06, 'learning_rate/action_model': 6.738383847928191e-05, 'epoch': 1.29}) +[2026-06-17 21:15:09,004][starVLA.training.train_starvla][INFO] - Step 2851, Loss: {'train/loss': 0.5080475211143494, 'train/grad_norm_pre_clip': 35.783138275146484, 'timing/data': 0.008894400671124458, 'timing/model': 0.9527752120047808, 'learning_rate/qwen_vl_interface': 6.736329916882044e-06, 'learning_rate/action_model': 6.736329916882044e-05, 'epoch': 1.29}) +[2026-06-17 21:15:12,949][starVLA.training.train_starvla][INFO] - Step 2852, Loss: {'train/loss': 0.7832629084587097, 'train/grad_norm_pre_clip': 21.129865646362305, 'timing/data': 0.013272631913423538, 'timing/model': 0.9503903081640601, 'learning_rate/qwen_vl_interface': 6.734275677718105e-06, 'learning_rate/action_model': 6.734275677718105e-05, 'epoch': 1.29}) +[2026-06-17 21:15:17,603][starVLA.training.train_starvla][INFO] - Step 2853, Loss: {'train/loss': 0.2514978051185608, 'train/grad_norm_pre_clip': 14.342342376708984, 'timing/data': 0.009289653971791267, 'timing/model': 1.4615507200360298, 'learning_rate/qwen_vl_interface': 6.73222113086222e-06, 'learning_rate/action_model': 6.73222113086222e-05, 'epoch': 1.29}) +[2026-06-17 21:15:22,711][starVLA.training.train_starvla][INFO] - Step 2854, Loss: {'train/loss': 0.2926616668701172, 'train/grad_norm_pre_clip': 11.555917739868164, 'timing/data': 0.00671393983066082, 'timing/model': 2.2043580496683717, 'learning_rate/qwen_vl_interface': 6.7301662767403e-06, 'learning_rate/action_model': 6.7301662767403e-05, 'epoch': 1.29}) +[2026-06-17 21:15:26,600][starVLA.training.train_starvla][INFO] - Step 2855, Loss: {'train/loss': 0.27013084292411804, 'train/grad_norm_pre_clip': 13.929807662963867, 'timing/data': 0.009975009597837925, 'timing/model': 1.1899677300825715, 'learning_rate/qwen_vl_interface': 6.728111115778319e-06, 'learning_rate/action_model': 6.728111115778318e-05, 'epoch': 1.3}) +[2026-06-17 21:15:30,491][starVLA.training.train_starvla][INFO] - Step 2856, Loss: {'train/loss': 0.25286540389060974, 'train/grad_norm_pre_clip': 10.480259895324707, 'timing/data': 0.009190818294882774, 'timing/model': 1.1869806069880724, 'learning_rate/qwen_vl_interface': 6.7260556484023135e-06, 'learning_rate/action_model': 6.726055648402313e-05, 'epoch': 1.3}) +[2026-06-17 21:15:34,335][starVLA.training.train_starvla][INFO] - Step 2857, Loss: {'train/loss': 0.4445386230945587, 'train/grad_norm_pre_clip': 19.00804901123047, 'timing/data': 0.013448773883283138, 'timing/model': 1.0716605531051755, 'learning_rate/qwen_vl_interface': 6.723999875038381e-06, 'learning_rate/action_model': 6.72399987503838e-05, 'epoch': 1.3}) +[2026-06-17 21:15:38,314][starVLA.training.train_starvla][INFO] - Step 2858, Loss: {'train/loss': 0.4489205479621887, 'train/grad_norm_pre_clip': 29.06484603881836, 'timing/data': 0.011253444477915764, 'timing/model': 0.9574260357767344, 'learning_rate/qwen_vl_interface': 6.721943796112691e-06, 'learning_rate/action_model': 6.721943796112691e-05, 'epoch': 1.3}) +[2026-06-17 21:15:42,445][starVLA.training.train_starvla][INFO] - Step 2859, Loss: {'train/loss': 0.4392548203468323, 'train/grad_norm_pre_clip': 15.980453491210938, 'timing/data': 0.01082749292254448, 'timing/model': 1.0138689763844013, 'learning_rate/qwen_vl_interface': 6.719887412051468e-06, 'learning_rate/action_model': 6.719887412051468e-05, 'epoch': 1.3}) +[2026-06-17 21:15:46,354][starVLA.training.train_starvla][INFO] - Step 2860, Loss: {'train/loss': 0.2824759781360626, 'train/grad_norm_pre_clip': 12.76445198059082, 'timing/data': 0.0083073191344738, 'timing/model': 0.9796426882967353, 'learning_rate/qwen_vl_interface': 6.7178307232810015e-06, 'learning_rate/action_model': 6.717830723281e-05, 'epoch': 1.3}) +[2026-06-17 21:15:50,306][starVLA.training.train_starvla][INFO] - Step 2861, Loss: {'train/loss': 0.3250962495803833, 'train/grad_norm_pre_clip': 15.325506210327148, 'timing/data': 0.008958653546869755, 'timing/model': 0.9849889259785414, 'learning_rate/qwen_vl_interface': 6.715773730227646e-06, 'learning_rate/action_model': 6.715773730227646e-05, 'epoch': 1.3}) +[2026-06-17 21:15:54,283][starVLA.training.train_starvla][INFO] - Step 2862, Loss: {'train/loss': 0.4521162211894989, 'train/grad_norm_pre_clip': 19.57303810119629, 'timing/data': 0.007786327041685581, 'timing/model': 1.1519903587177396, 'learning_rate/qwen_vl_interface': 6.7137164333178216e-06, 'learning_rate/action_model': 6.713716433317821e-05, 'epoch': 1.3}) +[2026-06-17 21:15:58,262][starVLA.training.train_starvla][INFO] - Step 2863, Loss: {'train/loss': 0.36057737469673157, 'train/grad_norm_pre_clip': 17.136112213134766, 'timing/data': 0.010070358403027058, 'timing/model': 1.1824553161859512, 'learning_rate/qwen_vl_interface': 6.711658832978004e-06, 'learning_rate/action_model': 6.711658832978003e-05, 'epoch': 1.3}) +[2026-06-17 21:16:01,988][starVLA.training.train_starvla][INFO] - Step 2864, Loss: {'train/loss': 0.3950592279434204, 'train/grad_norm_pre_clip': 20.664146423339844, 'timing/data': 0.01040752325206995, 'timing/model': 0.949296304024756, 'learning_rate/qwen_vl_interface': 6.709600929634737e-06, 'learning_rate/action_model': 6.709600929634738e-05, 'epoch': 1.3}) +[2026-06-17 21:16:05,865][starVLA.training.train_starvla][INFO] - Step 2865, Loss: {'train/loss': 0.2514372766017914, 'train/grad_norm_pre_clip': 12.307464599609375, 'timing/data': 0.00827492494136095, 'timing/model': 0.950001479126513, 'learning_rate/qwen_vl_interface': 6.707542723714629e-06, 'learning_rate/action_model': 6.707542723714629e-05, 'epoch': 1.3}) +[2026-06-17 21:16:09,768][starVLA.training.train_starvla][INFO] - Step 2866, Loss: {'train/loss': 0.35546520352363586, 'train/grad_norm_pre_clip': 12.951038360595703, 'timing/data': 0.009448236785829067, 'timing/model': 0.9564289283007383, 'learning_rate/qwen_vl_interface': 6.7054842156443435e-06, 'learning_rate/action_model': 6.705484215644343e-05, 'epoch': 1.3}) +[2026-06-17 21:16:13,689][starVLA.training.train_starvla][INFO] - Step 2867, Loss: {'train/loss': 0.4294353127479553, 'train/grad_norm_pre_clip': 18.90677833557129, 'timing/data': 0.008970513939857483, 'timing/model': 0.9561807448044419, 'learning_rate/qwen_vl_interface': 6.703425405850614e-06, 'learning_rate/action_model': 6.703425405850613e-05, 'epoch': 1.3}) +[2026-06-17 21:16:17,545][starVLA.training.train_starvla][INFO] - Step 2868, Loss: {'train/loss': 0.28364014625549316, 'train/grad_norm_pre_clip': 13.455394744873047, 'timing/data': 0.010897262953221798, 'timing/model': 0.974204896017909, 'learning_rate/qwen_vl_interface': 6.701366294760236e-06, 'learning_rate/action_model': 6.701366294760235e-05, 'epoch': 1.3}) +[2026-06-17 21:16:21,440][starVLA.training.train_starvla][INFO] - Step 2869, Loss: {'train/loss': 0.39728090167045593, 'train/grad_norm_pre_clip': 17.55461311340332, 'timing/data': 0.009876045398414135, 'timing/model': 1.0990581531077623, 'learning_rate/qwen_vl_interface': 6.6993068828000605e-06, 'learning_rate/action_model': 6.69930688280006e-05, 'epoch': 1.3}) +[2026-06-17 21:16:25,359][starVLA.training.train_starvla][INFO] - Step 2870, Loss: {'train/loss': 0.7196397185325623, 'train/grad_norm_pre_clip': 14.950178146362305, 'timing/data': 0.010409092530608177, 'timing/model': 1.0678823255002499, 'learning_rate/qwen_vl_interface': 6.697247170397011e-06, 'learning_rate/action_model': 6.697247170397011e-05, 'epoch': 1.3}) +[2026-06-17 21:16:29,204][starVLA.training.train_starvla][INFO] - Step 2871, Loss: {'train/loss': 0.2635246217250824, 'train/grad_norm_pre_clip': 13.037569999694824, 'timing/data': 0.00781856756657362, 'timing/model': 0.9449712289497256, 'learning_rate/qwen_vl_interface': 6.6951871579780645e-06, 'learning_rate/action_model': 6.695187157978064e-05, 'epoch': 1.3}) +[2026-06-17 21:16:33,120][starVLA.training.train_starvla][INFO] - Step 2872, Loss: {'train/loss': 0.4371703863143921, 'train/grad_norm_pre_clip': 14.155692100524902, 'timing/data': 0.009493633173406124, 'timing/model': 0.9560479400679469, 'learning_rate/qwen_vl_interface': 6.693126845970267e-06, 'learning_rate/action_model': 6.693126845970267e-05, 'epoch': 1.3}) +[2026-06-17 21:16:37,119][starVLA.training.train_starvla][INFO] - Step 2873, Loss: {'train/loss': 0.3062455654144287, 'train/grad_norm_pre_clip': 15.381258964538574, 'timing/data': 0.009258881211280823, 'timing/model': 1.0348843494430184, 'learning_rate/qwen_vl_interface': 6.691066234800719e-06, 'learning_rate/action_model': 6.691066234800719e-05, 'epoch': 1.3}) +[2026-06-17 21:16:41,056][starVLA.training.train_starvla][INFO] - Step 2874, Loss: {'train/loss': 0.1498744636774063, 'train/grad_norm_pre_clip': 10.745757102966309, 'timing/data': 0.009928853251039982, 'timing/model': 0.9740572571754456, 'learning_rate/qwen_vl_interface': 6.689005324896591e-06, 'learning_rate/action_model': 6.689005324896591e-05, 'epoch': 1.3}) +[2026-06-17 21:16:44,952][starVLA.training.train_starvla][INFO] - Step 2875, Loss: {'train/loss': 0.5297141075134277, 'train/grad_norm_pre_clip': 18.127174377441406, 'timing/data': 0.00770918745547533, 'timing/model': 0.9597119484096766, 'learning_rate/qwen_vl_interface': 6.686944116685111e-06, 'learning_rate/action_model': 6.686944116685111e-05, 'epoch': 1.3}) +[2026-06-17 21:16:48,990][starVLA.training.train_starvla][INFO] - Step 2876, Loss: {'train/loss': 0.39165088534355164, 'train/grad_norm_pre_clip': 21.128982543945312, 'timing/data': 0.0075735729187726974, 'timing/model': 1.182274715974927, 'learning_rate/qwen_vl_interface': 6.684882610593568e-06, 'learning_rate/action_model': 6.684882610593567e-05, 'epoch': 1.3}) +[2026-06-17 21:16:52,704][starVLA.training.train_starvla][INFO] - Step 2877, Loss: {'train/loss': 0.4254499673843384, 'train/grad_norm_pre_clip': 28.618715286254883, 'timing/data': 0.015945148654282093, 'timing/model': 0.9717139787971973, 'learning_rate/qwen_vl_interface': 6.682820807049317e-06, 'learning_rate/action_model': 6.682820807049317e-05, 'epoch': 1.31}) +[2026-06-17 21:16:56,708][starVLA.training.train_starvla][INFO] - Step 2878, Loss: {'train/loss': 0.4274176359176636, 'train/grad_norm_pre_clip': 16.6603946685791, 'timing/data': 0.011142302304506302, 'timing/model': 0.9653979679569602, 'learning_rate/qwen_vl_interface': 6.680758706479772e-06, 'learning_rate/action_model': 6.680758706479772e-05, 'epoch': 1.31}) +[2026-06-17 21:17:03,310][starVLA.training.train_starvla][INFO] - Step 2879, Loss: {'train/loss': 0.7029077410697937, 'train/grad_norm_pre_clip': 17.896154403686523, 'timing/data': 0.018658630549907684, 'timing/model': 0.9504352789372206, 'learning_rate/qwen_vl_interface': 6.678696309312405e-06, 'learning_rate/action_model': 6.678696309312406e-05, 'epoch': 1.31}) +[2026-06-17 21:17:08,580][starVLA.training.train_starvla][INFO] - Step 2880, Loss: {'train/loss': 0.15040460228919983, 'train/grad_norm_pre_clip': 12.080744743347168, 'timing/data': 0.011417594738304615, 'timing/model': 1.0097588589414954, 'learning_rate/qwen_vl_interface': 6.676633615974758e-06, 'learning_rate/action_model': 6.676633615974757e-05, 'epoch': 1.31}) +[2026-06-17 21:17:12,521][starVLA.training.train_starvla][INFO] - Step 2881, Loss: {'train/loss': 0.7030673027038574, 'train/grad_norm_pre_clip': 21.29324722290039, 'timing/data': 0.00808775331825018, 'timing/model': 0.9582703076303005, 'learning_rate/qwen_vl_interface': 6.6745706268944265e-06, 'learning_rate/action_model': 6.674570626894426e-05, 'epoch': 1.31}) +[2026-06-17 21:17:16,438][starVLA.training.train_starvla][INFO] - Step 2882, Loss: {'train/loss': 0.4928251802921295, 'train/grad_norm_pre_clip': 10.743515968322754, 'timing/data': 0.009498863480985165, 'timing/model': 1.1239772103726864, 'learning_rate/qwen_vl_interface': 6.672507342499071e-06, 'learning_rate/action_model': 6.67250734249907e-05, 'epoch': 1.31}) +[2026-06-17 21:17:21,366][starVLA.training.train_starvla][INFO] - Step 2883, Loss: {'train/loss': 0.20029135048389435, 'train/grad_norm_pre_clip': 12.830780029296875, 'timing/data': 0.01785274315625429, 'timing/model': 1.2259349497035146, 'learning_rate/qwen_vl_interface': 6.670443763216412e-06, 'learning_rate/action_model': 6.670443763216411e-05, 'epoch': 1.31}) +[2026-06-17 21:17:25,157][starVLA.training.train_starvla][INFO] - Step 2884, Loss: {'train/loss': 0.5364236831665039, 'train/grad_norm_pre_clip': 24.00432777404785, 'timing/data': 0.009939662180840969, 'timing/model': 0.9530058251693845, 'learning_rate/qwen_vl_interface': 6.668379889474233e-06, 'learning_rate/action_model': 6.668379889474234e-05, 'epoch': 1.31}) +[2026-06-17 21:17:29,031][starVLA.training.train_starvla][INFO] - Step 2885, Loss: {'train/loss': 0.5040779113769531, 'train/grad_norm_pre_clip': 16.539329528808594, 'timing/data': 0.011553790420293808, 'timing/model': 0.9489446533843875, 'learning_rate/qwen_vl_interface': 6.666315721700378e-06, 'learning_rate/action_model': 6.666315721700377e-05, 'epoch': 1.31}) +[2026-06-17 21:17:32,867][starVLA.training.train_starvla][INFO] - Step 2886, Loss: {'train/loss': 0.3726801574230194, 'train/grad_norm_pre_clip': 16.888973236083984, 'timing/data': 0.019605856388807297, 'timing/model': 0.949057949706912, 'learning_rate/qwen_vl_interface': 6.664251260322748e-06, 'learning_rate/action_model': 6.664251260322749e-05, 'epoch': 1.31}) +[2026-06-17 21:17:36,731][starVLA.training.train_starvla][INFO] - Step 2887, Loss: {'train/loss': 0.46763789653778076, 'train/grad_norm_pre_clip': 23.621307373046875, 'timing/data': 0.016136640682816505, 'timing/model': 0.954202476888895, 'learning_rate/qwen_vl_interface': 6.66218650576931e-06, 'learning_rate/action_model': 6.66218650576931e-05, 'epoch': 1.31}) +[2026-06-17 21:17:40,658][starVLA.training.train_starvla][INFO] - Step 2888, Loss: {'train/loss': 0.5810668468475342, 'train/grad_norm_pre_clip': 21.206295013427734, 'timing/data': 0.013669433072209358, 'timing/model': 0.9617061913013458, 'learning_rate/qwen_vl_interface': 6.66012145846809e-06, 'learning_rate/action_model': 6.660121458468091e-05, 'epoch': 1.31}) +[2026-06-17 21:17:44,522][starVLA.training.train_starvla][INFO] - Step 2889, Loss: {'train/loss': 0.5028459429740906, 'train/grad_norm_pre_clip': 15.084938049316406, 'timing/data': 0.00859703030437231, 'timing/model': 1.1186102321371436, 'learning_rate/qwen_vl_interface': 6.658056118847176e-06, 'learning_rate/action_model': 6.658056118847175e-05, 'epoch': 1.31}) +[2026-06-17 21:17:48,404][starVLA.training.train_starvla][INFO] - Step 2890, Loss: {'train/loss': 0.3286040127277374, 'train/grad_norm_pre_clip': 13.893732070922852, 'timing/data': 0.007952881045639515, 'timing/model': 1.0886366050690413, 'learning_rate/qwen_vl_interface': 6.655990487334713e-06, 'learning_rate/action_model': 6.655990487334712e-05, 'epoch': 1.31}) +[2026-06-17 21:17:52,128][starVLA.training.train_starvla][INFO] - Step 2891, Loss: {'train/loss': 0.35601383447647095, 'train/grad_norm_pre_clip': 12.313316345214844, 'timing/data': 0.008464553393423557, 'timing/model': 1.0215600105002522, 'learning_rate/qwen_vl_interface': 6.653924564358907e-06, 'learning_rate/action_model': 6.653924564358907e-05, 'epoch': 1.31}) +[2026-06-17 21:17:56,628][starVLA.training.train_starvla][INFO] - Step 2892, Loss: {'train/loss': 0.3825986087322235, 'train/grad_norm_pre_clip': 17.57088851928711, 'timing/data': 0.008975266478955746, 'timing/model': 0.9528490537777543, 'learning_rate/qwen_vl_interface': 6.6518583503480294e-06, 'learning_rate/action_model': 6.651858350348029e-05, 'epoch': 1.31}) +[2026-06-17 21:18:00,754][starVLA.training.train_starvla][INFO] - Step 2893, Loss: {'train/loss': 0.29202762246131897, 'train/grad_norm_pre_clip': 22.780593872070312, 'timing/data': 0.013074533082544804, 'timing/model': 0.9446086417883635, 'learning_rate/qwen_vl_interface': 6.649791845730408e-06, 'learning_rate/action_model': 6.649791845730408e-05, 'epoch': 1.31}) +[2026-06-17 21:18:04,536][starVLA.training.train_starvla][INFO] - Step 2894, Loss: {'train/loss': 0.3497019112110138, 'train/grad_norm_pre_clip': 14.56586742401123, 'timing/data': 0.008898907341063023, 'timing/model': 0.9414570806547999, 'learning_rate/qwen_vl_interface': 6.64772505093443e-06, 'learning_rate/action_model': 6.647725050934429e-05, 'epoch': 1.31}) +[2026-06-17 21:18:08,346][starVLA.training.train_starvla][INFO] - Step 2895, Loss: {'train/loss': 0.4546254575252533, 'train/grad_norm_pre_clip': 24.021499633789062, 'timing/data': 0.007944994606077671, 'timing/model': 0.9379308512434363, 'learning_rate/qwen_vl_interface': 6.645657966388543e-06, 'learning_rate/action_model': 6.645657966388543e-05, 'epoch': 1.31}) +[2026-06-17 21:18:12,221][starVLA.training.train_starvla][INFO] - Step 2896, Loss: {'train/loss': 0.5386995077133179, 'train/grad_norm_pre_clip': 15.170050621032715, 'timing/data': 0.007919766940176487, 'timing/model': 1.1181665575131774, 'learning_rate/qwen_vl_interface': 6.643590592521261e-06, 'learning_rate/action_model': 6.643590592521261e-05, 'epoch': 1.31}) +[2026-06-17 21:18:16,029][starVLA.training.train_starvla][INFO] - Step 2897, Loss: {'train/loss': 0.3468901813030243, 'train/grad_norm_pre_clip': 16.149002075195312, 'timing/data': 0.008867394179105759, 'timing/model': 0.960590461269021, 'learning_rate/qwen_vl_interface': 6.641522929761148e-06, 'learning_rate/action_model': 6.641522929761148e-05, 'epoch': 1.31}) +[2026-06-17 21:18:19,904][starVLA.training.train_starvla][INFO] - Step 2898, Loss: {'train/loss': 0.12897300720214844, 'train/grad_norm_pre_clip': 8.803959846496582, 'timing/data': 0.008624549023807049, 'timing/model': 0.9413420325145125, 'learning_rate/qwen_vl_interface': 6.639454978536835e-06, 'learning_rate/action_model': 6.639454978536834e-05, 'epoch': 1.31}) +[2026-06-17 21:18:24,017][starVLA.training.train_starvla][INFO] - Step 2899, Loss: {'train/loss': 0.20208176970481873, 'train/grad_norm_pre_clip': 22.757205963134766, 'timing/data': 0.008667855523526669, 'timing/model': 0.9426980298012495, 'learning_rate/qwen_vl_interface': 6.63738673927701e-06, 'learning_rate/action_model': 6.63738673927701e-05, 'epoch': 1.32}) +[2026-06-17 21:20:39,078][starVLA.training.train_starvla][INFO] - Step 2900, Loss: {'train/loss': 0.34988167881965637, 'train/grad_norm_pre_clip': 19.230182647705078, 'eval/loss': 0.3735124170780182, 'eval/latency_8/loss': 0.3735124170780182, 'eval/demon_fire_precision': 0.9395017793594306, 'eval/demon_fire_recall': 0.9841565703634669, 'eval/demon_fire_f1': 0.9613108784706419, 'eval/demon_fire_rate_model': 0.9382303839732888, 'eval/demon_fire_rate_teacher': 0.8956594323873122, 'eval/demon_left_precision': 0.7057613168724279, 'eval/demon_left_recall': 0.734475374732334, 'eval/demon_left_f1': 0.719832109129066, 'eval/demon_left_rate_model': 0.1352253756260434, 'eval/demon_left_rate_teacher': 0.12993878686700056, 'eval/demon_right_precision': 0.7974683544303798, 'eval/demon_right_recall': 0.7636363636363637, 'eval/demon_right_f1': 0.780185758513932, 'eval/demon_right_rate_model': 0.1758486366165832, 'eval/demon_right_rate_teacher': 0.18363939899833054, 'eval/demon_move_f1': 0.750008933821499, 'eval/demon_cc_f1': 0.8863931337213955, 'eval/latency_8/demon_fire_precision': 0.9395017793594306, 'eval/latency_8/demon_fire_recall': 0.9841565703634669, 'eval/latency_8/demon_fire_f1': 0.9613108784706419, 'eval/latency_8/demon_fire_rate_model': 0.9382303839732888, 'eval/latency_8/demon_fire_rate_teacher': 0.8956594323873122, 'eval/latency_8/demon_left_precision': 0.7057613168724279, 'eval/latency_8/demon_left_recall': 0.734475374732334, 'eval/latency_8/demon_left_f1': 0.719832109129066, 'eval/latency_8/demon_left_rate_model': 0.1352253756260434, 'eval/latency_8/demon_left_rate_teacher': 0.12993878686700056, 'eval/latency_8/demon_right_precision': 0.7974683544303798, 'eval/latency_8/demon_right_recall': 0.7636363636363637, 'eval/latency_8/demon_right_f1': 0.780185758513932, 'eval/latency_8/demon_right_rate_model': 0.1758486366165832, 'eval/latency_8/demon_right_rate_teacher': 0.18363939899833054, 'eval/latency_8/demon_move_f1': 0.750008933821499, 'eval/latency_8/demon_cc_f1': 0.8863931337213955, 'timing/data': 0.00817202404141426, 'timing/model': 0.9836546424776316, 'learning_rate/qwen_vl_interface': 6.63531821241042e-06, 'learning_rate/action_model': 6.63531821241042e-05, 'epoch': 1.32}) +[2026-06-17 21:20:39,091][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_2900_state +[2026-06-17 21:20:39,097][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 21:20:45,743][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_2900_state/model.safetensors +[2026-06-17 21:21:01,666][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_2900_state/optimizer.bin +[2026-06-17 21:21:01,667][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_2900_state/sampler.bin +[2026-06-17 21:21:01,667][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_2900_state/sampler_1.bin +[2026-06-17 21:21:01,669][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_2900_state/random_states_0.pkl +[2026-06-17 21:21:01,674][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2900 +[2026-06-17 21:21:01,674][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 21:21:01,685][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 21:21:05,571][starVLA.training.train_starvla][INFO] - Step 2901, Loss: {'train/loss': 0.29567471146583557, 'train/grad_norm_pre_clip': 12.879932403564453, 'timing/data': 0.009007059969007969, 'timing/model': 0.9332555243745446, 'learning_rate/qwen_vl_interface': 6.633249398365874e-06, 'learning_rate/action_model': 6.633249398365873e-05, 'epoch': 1.32}) +[2026-06-17 21:21:09,400][starVLA.training.train_starvla][INFO] - Step 2902, Loss: {'train/loss': 0.3926764130592346, 'train/grad_norm_pre_clip': 15.196206092834473, 'timing/data': 0.010897992178797722, 'timing/model': 0.9464971059933305, 'learning_rate/qwen_vl_interface': 6.6311802975722395e-06, 'learning_rate/action_model': 6.63118029757224e-05, 'epoch': 1.32}) +[2026-06-17 21:21:13,346][starVLA.training.train_starvla][INFO] - Step 2903, Loss: {'train/loss': 0.7327929139137268, 'train/grad_norm_pre_clip': 25.711833953857422, 'timing/data': 0.007438438944518566, 'timing/model': 0.9641041532158852, 'learning_rate/qwen_vl_interface': 6.629110910458444e-06, 'learning_rate/action_model': 6.629110910458444e-05, 'epoch': 1.32}) +[2026-06-17 21:21:17,307][starVLA.training.train_starvla][INFO] - Step 2904, Loss: {'train/loss': 0.408633828163147, 'train/grad_norm_pre_clip': 14.689303398132324, 'timing/data': 0.008764965459704399, 'timing/model': 1.1638006512075663, 'learning_rate/qwen_vl_interface': 6.627041237453471e-06, 'learning_rate/action_model': 6.627041237453471e-05, 'epoch': 1.32}) +[2026-06-17 21:21:21,069][starVLA.training.train_starvla][INFO] - Step 2905, Loss: {'train/loss': 0.4549405574798584, 'train/grad_norm_pre_clip': 19.70006561279297, 'timing/data': 0.013807779178023338, 'timing/model': 1.087024975568056, 'learning_rate/qwen_vl_interface': 6.624971278986369e-06, 'learning_rate/action_model': 6.62497127898637e-05, 'epoch': 1.32}) +[2026-06-17 21:21:24,831][starVLA.training.train_starvla][INFO] - Step 2906, Loss: {'train/loss': 0.372310072183609, 'train/grad_norm_pre_clip': 16.83456039428711, 'timing/data': 0.01570943370461464, 'timing/model': 0.9489385979250073, 'learning_rate/qwen_vl_interface': 6.622901035486241e-06, 'learning_rate/action_model': 6.622901035486241e-05, 'epoch': 1.32}) +[2026-06-17 21:21:28,726][starVLA.training.train_starvla][INFO] - Step 2907, Loss: {'train/loss': 0.21642829477787018, 'train/grad_norm_pre_clip': 13.650933265686035, 'timing/data': 0.009039021097123623, 'timing/model': 0.9491986008360982, 'learning_rate/qwen_vl_interface': 6.620830507382251e-06, 'learning_rate/action_model': 6.620830507382251e-05, 'epoch': 1.32}) +[2026-06-17 21:21:32,694][starVLA.training.train_starvla][INFO] - Step 2908, Loss: {'train/loss': 0.4840683043003082, 'train/grad_norm_pre_clip': 16.60289764404297, 'timing/data': 0.012576749548316002, 'timing/model': 0.9386437414214015, 'learning_rate/qwen_vl_interface': 6.618759695103623e-06, 'learning_rate/action_model': 6.618759695103622e-05, 'epoch': 1.32}) +[2026-06-17 21:21:36,457][starVLA.training.train_starvla][INFO] - Step 2909, Loss: {'train/loss': 0.38667094707489014, 'train/grad_norm_pre_clip': 16.386919021606445, 'timing/data': 0.009065359830856323, 'timing/model': 0.957319526001811, 'learning_rate/qwen_vl_interface': 6.616688599079635e-06, 'learning_rate/action_model': 6.616688599079634e-05, 'epoch': 1.32}) +[2026-06-17 21:21:40,424][starVLA.training.train_starvla][INFO] - Step 2910, Loss: {'train/loss': 0.2986696660518646, 'train/grad_norm_pre_clip': 20.04216194152832, 'timing/data': 0.009175638668239117, 'timing/model': 0.9764055712148547, 'learning_rate/qwen_vl_interface': 6.614617219739632e-06, 'learning_rate/action_model': 6.614617219739632e-05, 'epoch': 1.32}) +[2026-06-17 21:21:44,444][starVLA.training.train_starvla][INFO] - Step 2911, Loss: {'train/loss': 0.29733994603157043, 'train/grad_norm_pre_clip': 12.559049606323242, 'timing/data': 0.0129856513813138, 'timing/model': 1.3144960599020123, 'learning_rate/qwen_vl_interface': 6.612545557513008e-06, 'learning_rate/action_model': 6.612545557513008e-05, 'epoch': 1.32}) +[2026-06-17 21:21:48,036][starVLA.training.train_starvla][INFO] - Step 2912, Loss: {'train/loss': 0.5424313545227051, 'train/grad_norm_pre_clip': 19.002960205078125, 'timing/data': 0.010128164663910866, 'timing/model': 0.9370596213266253, 'learning_rate/qwen_vl_interface': 6.610473612829227e-06, 'learning_rate/action_model': 6.610473612829226e-05, 'epoch': 1.32}) +[2026-06-17 21:21:51,828][starVLA.training.train_starvla][INFO] - Step 2913, Loss: {'train/loss': 0.3085117042064667, 'train/grad_norm_pre_clip': 20.968093872070312, 'timing/data': 0.00789421983063221, 'timing/model': 0.9415800385177135, 'learning_rate/qwen_vl_interface': 6.608401386117801e-06, 'learning_rate/action_model': 6.6084013861178e-05, 'epoch': 1.32}) +[2026-06-17 21:21:55,653][starVLA.training.train_starvla][INFO] - Step 2914, Loss: {'train/loss': 0.14512743055820465, 'train/grad_norm_pre_clip': 12.283536911010742, 'timing/data': 0.00971787329763174, 'timing/model': 0.9740155339241028, 'learning_rate/qwen_vl_interface': 6.606328877808306e-06, 'learning_rate/action_model': 6.606328877808305e-05, 'epoch': 1.32}) +[2026-06-17 21:21:59,431][starVLA.training.train_starvla][INFO] - Step 2915, Loss: {'train/loss': 0.3243727385997772, 'train/grad_norm_pre_clip': 13.944013595581055, 'timing/data': 0.015496168285608292, 'timing/model': 0.92921057716012, 'learning_rate/qwen_vl_interface': 6.604256088330374e-06, 'learning_rate/action_model': 6.604256088330374e-05, 'epoch': 1.32}) +[2026-06-17 21:22:03,242][starVLA.training.train_starvla][INFO] - Step 2916, Loss: {'train/loss': 0.43387654423713684, 'train/grad_norm_pre_clip': 19.455753326416016, 'timing/data': 0.00998311210423708, 'timing/model': 0.940759539604187, 'learning_rate/qwen_vl_interface': 6.602183018113698e-06, 'learning_rate/action_model': 6.602183018113698e-05, 'epoch': 1.32}) +[2026-06-17 21:22:07,227][starVLA.training.train_starvla][INFO] - Step 2917, Loss: {'train/loss': 0.3306083083152771, 'train/grad_norm_pre_clip': 14.605406761169434, 'timing/data': 0.009736825712025166, 'timing/model': 1.2726960452273488, 'learning_rate/qwen_vl_interface': 6.600109667588029e-06, 'learning_rate/action_model': 6.60010966758803e-05, 'epoch': 1.32}) +[2026-06-17 21:22:11,055][starVLA.training.train_starvla][INFO] - Step 2918, Loss: {'train/loss': 0.4505995810031891, 'train/grad_norm_pre_clip': 25.628334045410156, 'timing/data': 0.011771097779273987, 'timing/model': 1.1109527107328176, 'learning_rate/qwen_vl_interface': 6.598036037183172e-06, 'learning_rate/action_model': 6.598036037183172e-05, 'epoch': 1.32}) +[2026-06-17 21:22:14,827][starVLA.training.train_starvla][INFO] - Step 2919, Loss: {'train/loss': 0.3413349986076355, 'train/grad_norm_pre_clip': 24.74530601501465, 'timing/data': 0.008087323047220707, 'timing/model': 0.9481875337660313, 'learning_rate/qwen_vl_interface': 6.595962127328993e-06, 'learning_rate/action_model': 6.595962127328993e-05, 'epoch': 1.32}) +[2026-06-17 21:22:18,892][starVLA.training.train_starvla][INFO] - Step 2920, Loss: {'train/loss': 0.4167509377002716, 'train/grad_norm_pre_clip': 14.809345245361328, 'timing/data': 0.009408706799149513, 'timing/model': 1.1860379753634334, 'learning_rate/qwen_vl_interface': 6.593887938455418e-06, 'learning_rate/action_model': 6.593887938455418e-05, 'epoch': 1.32}) +[2026-06-17 21:22:22,689][starVLA.training.train_starvla][INFO] - Step 2921, Loss: {'train/loss': 0.20647631585597992, 'train/grad_norm_pre_clip': 14.337158203125, 'timing/data': 0.013252661563456059, 'timing/model': 0.9423825535923243, 'learning_rate/qwen_vl_interface': 6.591813470992428e-06, 'learning_rate/action_model': 6.591813470992427e-05, 'epoch': 1.33}) +[2026-06-17 21:22:26,554][starVLA.training.train_starvla][INFO] - Step 2922, Loss: {'train/loss': 0.25680944323539734, 'train/grad_norm_pre_clip': 14.018677711486816, 'timing/data': 0.008719772100448608, 'timing/model': 0.9480904927477241, 'learning_rate/qwen_vl_interface': 6.58973872537006e-06, 'learning_rate/action_model': 6.589738725370059e-05, 'epoch': 1.33}) +[2026-06-17 21:22:30,430][starVLA.training.train_starvla][INFO] - Step 2923, Loss: {'train/loss': 0.2824650704860687, 'train/grad_norm_pre_clip': 13.874825477600098, 'timing/data': 0.006806200370192528, 'timing/model': 0.9566036369651556, 'learning_rate/qwen_vl_interface': 6.587663702018413e-06, 'learning_rate/action_model': 6.587663702018412e-05, 'epoch': 1.33}) +[2026-06-17 21:22:34,268][starVLA.training.train_starvla][INFO] - Step 2924, Loss: {'train/loss': 0.2529847025871277, 'train/grad_norm_pre_clip': 15.270881652832031, 'timing/data': 0.007034142501652241, 'timing/model': 1.1142044635489583, 'learning_rate/qwen_vl_interface': 6.585588401367641e-06, 'learning_rate/action_model': 6.58558840136764e-05, 'epoch': 1.33}) +[2026-06-17 21:22:38,324][starVLA.training.train_starvla][INFO] - Step 2925, Loss: {'train/loss': 0.21200913190841675, 'train/grad_norm_pre_clip': 16.4440975189209, 'timing/data': 0.008145643398165703, 'timing/model': 1.2032959563657641, 'learning_rate/qwen_vl_interface': 6.583512823847957e-06, 'learning_rate/action_model': 6.583512823847958e-05, 'epoch': 1.33}) +[2026-06-17 21:22:42,136][starVLA.training.train_starvla][INFO] - Step 2926, Loss: {'train/loss': 0.31230196356773376, 'train/grad_norm_pre_clip': 15.94510555267334, 'timing/data': 0.010209312662482262, 'timing/model': 0.9448109846562147, 'learning_rate/qwen_vl_interface': 6.5814369698896285e-06, 'learning_rate/action_model': 6.581436969889628e-05, 'epoch': 1.33}) +[2026-06-17 21:22:45,969][starVLA.training.train_starvla][INFO] - Step 2927, Loss: {'train/loss': 0.27639472484588623, 'train/grad_norm_pre_clip': 13.865644454956055, 'timing/data': 0.007983580231666565, 'timing/model': 0.9372524982318282, 'learning_rate/qwen_vl_interface': 6.5793608399229855e-06, 'learning_rate/action_model': 6.579360839922985e-05, 'epoch': 1.33}) +[2026-06-17 21:22:49,808][starVLA.training.train_starvla][INFO] - Step 2928, Loss: {'train/loss': 0.4058457612991333, 'train/grad_norm_pre_clip': 22.79786491394043, 'timing/data': 0.00937892496585846, 'timing/model': 0.9374643433839083, 'learning_rate/qwen_vl_interface': 6.577284434378408e-06, 'learning_rate/action_model': 6.577284434378408e-05, 'epoch': 1.33}) +[2026-06-17 21:22:53,646][starVLA.training.train_starvla][INFO] - Step 2929, Loss: {'train/loss': 0.20351259410381317, 'train/grad_norm_pre_clip': 15.117318153381348, 'timing/data': 0.015241811983287334, 'timing/model': 0.968958780169487, 'learning_rate/qwen_vl_interface': 6.5752077536863404e-06, 'learning_rate/action_model': 6.57520775368634e-05, 'epoch': 1.33}) +[2026-06-17 21:22:57,560][starVLA.training.train_starvla][INFO] - Step 2930, Loss: {'train/loss': 0.2987067401409149, 'train/grad_norm_pre_clip': 20.953460693359375, 'timing/data': 0.009563379921019077, 'timing/model': 0.9948260970413685, 'learning_rate/qwen_vl_interface': 6.573130798277277e-06, 'learning_rate/action_model': 6.573130798277277e-05, 'epoch': 1.33}) +[2026-06-17 21:23:01,510][starVLA.training.train_starvla][INFO] - Step 2931, Loss: {'train/loss': 0.1762365847826004, 'train/grad_norm_pre_clip': 18.980297088623047, 'timing/data': 0.011125044897198677, 'timing/model': 1.1321047926321626, 'learning_rate/qwen_vl_interface': 6.571053568581778e-06, 'learning_rate/action_model': 6.571053568581777e-05, 'epoch': 1.33}) +[2026-06-17 21:23:05,221][starVLA.training.train_starvla][INFO] - Step 2932, Loss: {'train/loss': 0.5014398694038391, 'train/grad_norm_pre_clip': 19.553974151611328, 'timing/data': 0.009391797706484795, 'timing/model': 0.9694782560691237, 'learning_rate/qwen_vl_interface': 6.568976065030453e-06, 'learning_rate/action_model': 6.568976065030453e-05, 'epoch': 1.33}) +[2026-06-17 21:23:09,216][starVLA.training.train_starvla][INFO] - Step 2933, Loss: {'train/loss': 0.41654470562934875, 'train/grad_norm_pre_clip': 23.50142478942871, 'timing/data': 0.007138006389141083, 'timing/model': 0.9545191759243608, 'learning_rate/qwen_vl_interface': 6.566898288053968e-06, 'learning_rate/action_model': 6.566898288053968e-05, 'epoch': 1.33}) +[2026-06-17 21:23:13,097][starVLA.training.train_starvla][INFO] - Step 2934, Loss: {'train/loss': 0.43537384271621704, 'train/grad_norm_pre_clip': 15.17504596710205, 'timing/data': 0.006605485454201698, 'timing/model': 0.9520726026967168, 'learning_rate/qwen_vl_interface': 6.564820238083053e-06, 'learning_rate/action_model': 6.564820238083052e-05, 'epoch': 1.33}) +[2026-06-17 21:23:16,933][starVLA.training.train_starvla][INFO] - Step 2935, Loss: {'train/loss': 0.1679219752550125, 'train/grad_norm_pre_clip': 10.84361457824707, 'timing/data': 0.008188546635210514, 'timing/model': 0.9456854416057467, 'learning_rate/qwen_vl_interface': 6.562741915548486e-06, 'learning_rate/action_model': 6.562741915548486e-05, 'epoch': 1.33}) +[2026-06-17 21:23:20,768][starVLA.training.train_starvla][INFO] - Step 2936, Loss: {'train/loss': 0.17469002306461334, 'train/grad_norm_pre_clip': 10.577401161193848, 'timing/data': 0.019768626429140568, 'timing/model': 0.9399938499554992, 'learning_rate/qwen_vl_interface': 6.5606633208811096e-06, 'learning_rate/action_model': 6.560663320881109e-05, 'epoch': 1.33}) +[2026-06-17 21:23:24,737][starVLA.training.train_starvla][INFO] - Step 2937, Loss: {'train/loss': 0.5073802471160889, 'train/grad_norm_pre_clip': 18.900611877441406, 'timing/data': 0.009704346768558025, 'timing/model': 0.9463091660290956, 'learning_rate/qwen_vl_interface': 6.558584454511814e-06, 'learning_rate/action_model': 6.558584454511814e-05, 'epoch': 1.33}) +[2026-06-17 21:23:28,595][starVLA.training.train_starvla][INFO] - Step 2938, Loss: {'train/loss': 0.28878384828567505, 'train/grad_norm_pre_clip': 19.824323654174805, 'timing/data': 0.016711684875190258, 'timing/model': 1.1329829767346382, 'learning_rate/qwen_vl_interface': 6.556505316871555e-06, 'learning_rate/action_model': 6.556505316871554e-05, 'epoch': 1.33}) +[2026-06-17 21:23:32,234][starVLA.training.train_starvla][INFO] - Step 2939, Loss: {'train/loss': 0.5853523015975952, 'train/grad_norm_pre_clip': 19.458528518676758, 'timing/data': 0.00941486470401287, 'timing/model': 0.9426378766074777, 'learning_rate/qwen_vl_interface': 6.554425908391337e-06, 'learning_rate/action_model': 6.554425908391336e-05, 'epoch': 1.33}) +[2026-06-17 21:23:36,083][starVLA.training.train_starvla][INFO] - Step 2940, Loss: {'train/loss': 0.20360791683197021, 'train/grad_norm_pre_clip': 15.409947395324707, 'timing/data': 0.006832106038928032, 'timing/model': 1.0002240892499685, 'learning_rate/qwen_vl_interface': 6.552346229502224e-06, 'learning_rate/action_model': 6.552346229502224e-05, 'epoch': 1.33}) +[2026-06-17 21:23:40,031][starVLA.training.train_starvla][INFO] - Step 2941, Loss: {'train/loss': 0.6381149291992188, 'train/grad_norm_pre_clip': 30.14044952392578, 'timing/data': 0.00938887894153595, 'timing/model': 0.9642749633640051, 'learning_rate/qwen_vl_interface': 6.550266280635337e-06, 'learning_rate/action_model': 6.550266280635336e-05, 'epoch': 1.33}) +[2026-06-17 21:23:43,926][starVLA.training.train_starvla][INFO] - Step 2942, Loss: {'train/loss': 0.6779389977455139, 'train/grad_norm_pre_clip': 23.09035301208496, 'timing/data': 0.007740196771919727, 'timing/model': 0.9484652187675238, 'learning_rate/qwen_vl_interface': 6.54818606222185e-06, 'learning_rate/action_model': 6.548186062221849e-05, 'epoch': 1.33}) +[2026-06-17 21:23:47,733][starVLA.training.train_starvla][INFO] - Step 2943, Loss: {'train/loss': 0.19580203294754028, 'train/grad_norm_pre_clip': 11.381291389465332, 'timing/data': 0.009518005885183811, 'timing/model': 0.9462435496971011, 'learning_rate/qwen_vl_interface': 6.546105574692997e-06, 'learning_rate/action_model': 6.546105574692997e-05, 'epoch': 1.34}) +[2026-06-17 21:23:51,716][starVLA.training.train_starvla][INFO] - Step 2944, Loss: {'train/loss': 0.45421063899993896, 'train/grad_norm_pre_clip': 22.428726196289062, 'timing/data': 0.013539766892790794, 'timing/model': 1.1293132985010743, 'learning_rate/qwen_vl_interface': 6.544024818480061e-06, 'learning_rate/action_model': 6.544024818480062e-05, 'epoch': 1.34}) +[2026-06-17 21:23:55,579][starVLA.training.train_starvla][INFO] - Step 2945, Loss: {'train/loss': 0.49983689188957214, 'train/grad_norm_pre_clip': 21.57099151611328, 'timing/data': 0.008937803097069263, 'timing/model': 1.1847657784819603, 'learning_rate/qwen_vl_interface': 6.5419437940143905e-06, 'learning_rate/action_model': 6.54194379401439e-05, 'epoch': 1.34}) +[2026-06-17 21:23:59,278][starVLA.training.train_starvla][INFO] - Step 2946, Loss: {'train/loss': 0.24773231148719788, 'train/grad_norm_pre_clip': 14.383651733398438, 'timing/data': 0.009458094835281372, 'timing/model': 0.9718152107670903, 'learning_rate/qwen_vl_interface': 6.5398625017273795e-06, 'learning_rate/action_model': 6.539862501727379e-05, 'epoch': 1.34}) +[2026-06-17 21:24:03,197][starVLA.training.train_starvla][INFO] - Step 2947, Loss: {'train/loss': 0.20856241881847382, 'train/grad_norm_pre_clip': 14.964042663574219, 'timing/data': 0.009339042007923126, 'timing/model': 0.9501249985769391, 'learning_rate/qwen_vl_interface': 6.537780942050486e-06, 'learning_rate/action_model': 6.537780942050486e-05, 'epoch': 1.34}) +[2026-06-17 21:24:07,071][starVLA.training.train_starvla][INFO] - Step 2948, Loss: {'train/loss': 0.44003620743751526, 'train/grad_norm_pre_clip': 18.553342819213867, 'timing/data': 0.009904930368065834, 'timing/model': 0.9549802849069238, 'learning_rate/qwen_vl_interface': 6.535699115415216e-06, 'learning_rate/action_model': 6.535699115415216e-05, 'epoch': 1.34}) +[2026-06-17 21:24:11,304][starVLA.training.train_starvla][INFO] - Step 2949, Loss: {'train/loss': 0.37889111042022705, 'train/grad_norm_pre_clip': 15.350210189819336, 'timing/data': 0.008505208417773247, 'timing/model': 0.9572534058243036, 'learning_rate/qwen_vl_interface': 6.533617022253137e-06, 'learning_rate/action_model': 6.533617022253137e-05, 'epoch': 1.34}) +[2026-06-17 21:24:15,494][starVLA.training.train_starvla][INFO] - Step 2950, Loss: {'train/loss': 0.8219473361968994, 'train/grad_norm_pre_clip': 32.46955490112305, 'timing/data': 0.01010153815150261, 'timing/model': 1.2353204544633627, 'learning_rate/qwen_vl_interface': 6.531534662995866e-06, 'learning_rate/action_model': 6.531534662995866e-05, 'epoch': 1.34}) +[2026-06-17 21:24:19,366][starVLA.training.train_starvla][INFO] - Step 2951, Loss: {'train/loss': 0.6262741088867188, 'train/grad_norm_pre_clip': 24.74233627319336, 'timing/data': 0.0189021872356534, 'timing/model': 1.1447580447420478, 'learning_rate/qwen_vl_interface': 6.529452038075082e-06, 'learning_rate/action_model': 6.529452038075081e-05, 'epoch': 1.34}) +[2026-06-17 21:24:23,221][starVLA.training.train_starvla][INFO] - Step 2952, Loss: {'train/loss': 0.42666295170783997, 'train/grad_norm_pre_clip': 20.78791046142578, 'timing/data': 0.013124903663992882, 'timing/model': 1.0920475395396352, 'learning_rate/qwen_vl_interface': 6.5273691479225124e-06, 'learning_rate/action_model': 6.527369147922511e-05, 'epoch': 1.34}) +[2026-06-17 21:24:26,904][starVLA.training.train_starvla][INFO] - Step 2953, Loss: {'train/loss': 0.7073471546173096, 'train/grad_norm_pre_clip': 23.283567428588867, 'timing/data': 0.010200338438153267, 'timing/model': 0.948950226418674, 'learning_rate/qwen_vl_interface': 6.525285992969945e-06, 'learning_rate/action_model': 6.525285992969945e-05, 'epoch': 1.34}) +[2026-06-17 21:24:30,730][starVLA.training.train_starvla][INFO] - Step 2954, Loss: {'train/loss': 0.3715454041957855, 'train/grad_norm_pre_clip': 23.578113555908203, 'timing/data': 0.009181411005556583, 'timing/model': 0.9263220187276602, 'learning_rate/qwen_vl_interface': 6.523202573649219e-06, 'learning_rate/action_model': 6.523202573649219e-05, 'epoch': 1.34}) +[2026-06-17 21:24:34,622][starVLA.training.train_starvla][INFO] - Step 2955, Loss: {'train/loss': 0.5845996141433716, 'train/grad_norm_pre_clip': 20.872093200683594, 'timing/data': 0.018628683872520924, 'timing/model': 0.9701607506722212, 'learning_rate/qwen_vl_interface': 6.5211188903922294e-06, 'learning_rate/action_model': 6.52111889039223e-05, 'epoch': 1.34}) +[2026-06-17 21:24:38,458][starVLA.training.train_starvla][INFO] - Step 2956, Loss: {'train/loss': 0.26962921023368835, 'train/grad_norm_pre_clip': 12.609023094177246, 'timing/data': 0.010511571541428566, 'timing/model': 0.938251655548811, 'learning_rate/qwen_vl_interface': 6.519034943630926e-06, 'learning_rate/action_model': 6.519034943630927e-05, 'epoch': 1.34}) +[2026-06-17 21:24:42,621][starVLA.training.train_starvla][INFO] - Step 2957, Loss: {'train/loss': 0.21390654146671295, 'train/grad_norm_pre_clip': 23.566747665405273, 'timing/data': 0.008869119919836521, 'timing/model': 0.9740728419274092, 'learning_rate/qwen_vl_interface': 6.5169507337973125e-06, 'learning_rate/action_model': 6.516950733797312e-05, 'epoch': 1.34}) +[2026-06-17 21:24:46,804][starVLA.training.train_starvla][INFO] - Step 2958, Loss: {'train/loss': 0.4846982955932617, 'train/grad_norm_pre_clip': 13.953104019165039, 'timing/data': 0.023000498302280903, 'timing/model': 1.1747178165242076, 'learning_rate/qwen_vl_interface': 6.51486626132345e-06, 'learning_rate/action_model': 6.51486626132345e-05, 'epoch': 1.34}) +[2026-06-17 21:24:50,443][starVLA.training.train_starvla][INFO] - Step 2959, Loss: {'train/loss': 0.2671201229095459, 'train/grad_norm_pre_clip': 17.389509201049805, 'timing/data': 0.007756161503493786, 'timing/model': 0.9446210218593478, 'learning_rate/qwen_vl_interface': 6.512781526641449e-06, 'learning_rate/action_model': 6.512781526641449e-05, 'epoch': 1.34}) +[2026-06-17 21:24:57,370][starVLA.training.train_starvla][INFO] - Step 2960, Loss: {'train/loss': 0.6508485674858093, 'train/grad_norm_pre_clip': 27.838802337646484, 'timing/data': 0.010985772125422955, 'timing/model': 2.322327592410147, 'learning_rate/qwen_vl_interface': 6.51069653018348e-06, 'learning_rate/action_model': 6.51069653018348e-05, 'epoch': 1.34}) +[2026-06-17 21:25:01,636][starVLA.training.train_starvla][INFO] - Step 2961, Loss: {'train/loss': 0.34906336665153503, 'train/grad_norm_pre_clip': 20.837032318115234, 'timing/data': 0.00832200888544321, 'timing/model': 0.9433208061382174, 'learning_rate/qwen_vl_interface': 6.508611272381763e-06, 'learning_rate/action_model': 6.508611272381763e-05, 'epoch': 1.34}) +[2026-06-17 21:25:05,716][starVLA.training.train_starvla][INFO] - Step 2962, Loss: {'train/loss': 0.32177573442459106, 'train/grad_norm_pre_clip': 14.791864395141602, 'timing/data': 0.009073611348867416, 'timing/model': 1.027830021455884, 'learning_rate/qwen_vl_interface': 6.506525753668576e-06, 'learning_rate/action_model': 6.506525753668576e-05, 'epoch': 1.34}) +[2026-06-17 21:25:09,553][starVLA.training.train_starvla][INFO] - Step 2963, Loss: {'train/loss': 0.6131139993667603, 'train/grad_norm_pre_clip': 20.915264129638672, 'timing/data': 0.008922656066715717, 'timing/model': 0.9537772350013256, 'learning_rate/qwen_vl_interface': 6.504439974476246e-06, 'learning_rate/action_model': 6.504439974476246e-05, 'epoch': 1.34}) +[2026-06-17 21:25:13,445][starVLA.training.train_starvla][INFO] - Step 2964, Loss: {'train/loss': 0.48914650082588196, 'train/grad_norm_pre_clip': 20.48876953125, 'timing/data': 0.009151003323495388, 'timing/model': 1.1167761329561472, 'learning_rate/qwen_vl_interface': 6.50235393523716e-06, 'learning_rate/action_model': 6.502353935237159e-05, 'epoch': 1.34}) +[2026-06-17 21:25:17,255][starVLA.training.train_starvla][INFO] - Step 2965, Loss: {'train/loss': 0.34670132398605347, 'train/grad_norm_pre_clip': 13.144143104553223, 'timing/data': 0.009839307516813278, 'timing/model': 1.1405050083994865, 'learning_rate/qwen_vl_interface': 6.500267636383758e-06, 'learning_rate/action_model': 6.500267636383758e-05, 'epoch': 1.35}) +[2026-06-17 21:25:20,893][starVLA.training.train_starvla][INFO] - Step 2966, Loss: {'train/loss': 0.38959458470344543, 'train/grad_norm_pre_clip': 17.183176040649414, 'timing/data': 0.014990243129432201, 'timing/model': 0.9317621206864715, 'learning_rate/qwen_vl_interface': 6.498181078348527e-06, 'learning_rate/action_model': 6.498181078348527e-05, 'epoch': 1.35}) +[2026-06-17 21:25:24,735][starVLA.training.train_starvla][INFO] - Step 2967, Loss: {'train/loss': 0.5001604557037354, 'train/grad_norm_pre_clip': 17.151212692260742, 'timing/data': 0.0101783974096179, 'timing/model': 0.9514286313205957, 'learning_rate/qwen_vl_interface': 6.496094261564018e-06, 'learning_rate/action_model': 6.496094261564018e-05, 'epoch': 1.35}) +[2026-06-17 21:25:28,585][starVLA.training.train_starvla][INFO] - Step 2968, Loss: {'train/loss': 0.4487426280975342, 'train/grad_norm_pre_clip': 16.76704216003418, 'timing/data': 0.012482765130698681, 'timing/model': 0.9233561558648944, 'learning_rate/qwen_vl_interface': 6.494007186462824e-06, 'learning_rate/action_model': 6.494007186462823e-05, 'epoch': 1.35}) +[2026-06-17 21:25:32,382][starVLA.training.train_starvla][INFO] - Step 2969, Loss: {'train/loss': 0.2797076404094696, 'train/grad_norm_pre_clip': 13.153767585754395, 'timing/data': 0.010310729034245014, 'timing/model': 0.9377245232462883, 'learning_rate/qwen_vl_interface': 6.491919853477604e-06, 'learning_rate/action_model': 6.491919853477604e-05, 'epoch': 1.35}) +[2026-06-17 21:25:36,226][starVLA.training.train_starvla][INFO] - Step 2970, Loss: {'train/loss': 0.2398889809846878, 'train/grad_norm_pre_clip': 12.22664737701416, 'timing/data': 0.011574527248740196, 'timing/model': 0.95452387817204, 'learning_rate/qwen_vl_interface': 6.4898322630410615e-06, 'learning_rate/action_model': 6.489832263041062e-05, 'epoch': 1.35}) +[2026-06-17 21:25:40,056][starVLA.training.train_starvla][INFO] - Step 2971, Loss: {'train/loss': 0.4074482023715973, 'train/grad_norm_pre_clip': 19.306163787841797, 'timing/data': 0.019360709935426712, 'timing/model': 1.119647934101522, 'learning_rate/qwen_vl_interface': 6.487744415585958e-06, 'learning_rate/action_model': 6.487744415585958e-05, 'epoch': 1.35}) +[2026-06-17 21:25:43,988][starVLA.training.train_starvla][INFO] - Step 2972, Loss: {'train/loss': 0.1262008398771286, 'train/grad_norm_pre_clip': 8.159456253051758, 'timing/data': 0.01933875121176243, 'timing/model': 1.148874256759882, 'learning_rate/qwen_vl_interface': 6.485656311545103e-06, 'learning_rate/action_model': 6.485656311545103e-05, 'epoch': 1.35}) +[2026-06-17 21:25:47,664][starVLA.training.train_starvla][INFO] - Step 2973, Loss: {'train/loss': 0.20312157273292542, 'train/grad_norm_pre_clip': 11.947943687438965, 'timing/data': 0.010755482129752636, 'timing/model': 0.9367837235331535, 'learning_rate/qwen_vl_interface': 6.483567951351366e-06, 'learning_rate/action_model': 6.483567951351365e-05, 'epoch': 1.35}) +[2026-06-17 21:25:51,545][starVLA.training.train_starvla][INFO] - Step 2974, Loss: {'train/loss': 0.34474605321884155, 'train/grad_norm_pre_clip': 13.826501846313477, 'timing/data': 0.008236703462898731, 'timing/model': 0.9395297691226006, 'learning_rate/qwen_vl_interface': 6.481479335437667e-06, 'learning_rate/action_model': 6.481479335437667e-05, 'epoch': 1.35}) +[2026-06-17 21:25:55,344][starVLA.training.train_starvla][INFO] - Step 2975, Loss: {'train/loss': 0.2159145474433899, 'train/grad_norm_pre_clip': 10.310161590576172, 'timing/data': 0.00961467158049345, 'timing/model': 0.9378840373829007, 'learning_rate/qwen_vl_interface': 6.479390464236974e-06, 'learning_rate/action_model': 6.479390464236974e-05, 'epoch': 1.35}) +[2026-06-17 21:25:59,238][starVLA.training.train_starvla][INFO] - Step 2976, Loss: {'train/loss': 0.5651825070381165, 'train/grad_norm_pre_clip': 21.142236709594727, 'timing/data': 0.016400881111621857, 'timing/model': 0.9695359552279115, 'learning_rate/qwen_vl_interface': 6.477301338182317e-06, 'learning_rate/action_model': 6.477301338182317e-05, 'epoch': 1.35}) +[2026-06-17 21:26:03,296][starVLA.training.train_starvla][INFO] - Step 2977, Loss: {'train/loss': 0.621040940284729, 'train/grad_norm_pre_clip': 20.393306732177734, 'timing/data': 0.00870032049715519, 'timing/model': 1.0210547670722008, 'learning_rate/qwen_vl_interface': 6.47521195770677e-06, 'learning_rate/action_model': 6.47521195770677e-05, 'epoch': 1.35}) +[2026-06-17 21:26:11,440][starVLA.training.train_starvla][INFO] - Step 2978, Loss: {'train/loss': 0.5950296521186829, 'train/grad_norm_pre_clip': 20.27867889404297, 'timing/data': 0.010804601944983006, 'timing/model': 2.6731936670839787, 'learning_rate/qwen_vl_interface': 6.473122323243467e-06, 'learning_rate/action_model': 6.473122323243467e-05, 'epoch': 1.35}) +[2026-06-17 21:26:15,266][starVLA.training.train_starvla][INFO] - Step 2979, Loss: {'train/loss': 0.5491877794265747, 'train/grad_norm_pre_clip': 20.833984375, 'timing/data': 0.010020372457802296, 'timing/model': 0.9512661630287766, 'learning_rate/qwen_vl_interface': 6.471032435225591e-06, 'learning_rate/action_model': 6.47103243522559e-05, 'epoch': 1.35}) +[2026-06-17 21:26:19,239][starVLA.training.train_starvla][INFO] - Step 2980, Loss: {'train/loss': 0.3936362862586975, 'train/grad_norm_pre_clip': 22.257219314575195, 'timing/data': 0.009719532914459705, 'timing/model': 0.9511013654991984, 'learning_rate/qwen_vl_interface': 6.468942294086376e-06, 'learning_rate/action_model': 6.468942294086376e-05, 'epoch': 1.35}) +[2026-06-17 21:26:23,093][starVLA.training.train_starvla][INFO] - Step 2981, Loss: {'train/loss': 0.41563594341278076, 'train/grad_norm_pre_clip': 16.08724021911621, 'timing/data': 0.009551191702485085, 'timing/model': 0.9846809785813093, 'learning_rate/qwen_vl_interface': 6.466851900259113e-06, 'learning_rate/action_model': 6.466851900259112e-05, 'epoch': 1.35}) +[2026-06-17 21:26:26,958][starVLA.training.train_starvla][INFO] - Step 2982, Loss: {'train/loss': 0.3706490099430084, 'train/grad_norm_pre_clip': 20.40328025817871, 'timing/data': 0.007952660322189331, 'timing/model': 0.9474394088611007, 'learning_rate/qwen_vl_interface': 6.464761254177139e-06, 'learning_rate/action_model': 6.464761254177138e-05, 'epoch': 1.35}) +[2026-06-17 21:26:31,077][starVLA.training.train_starvla][INFO] - Step 2983, Loss: {'train/loss': 1.3105993270874023, 'train/grad_norm_pre_clip': 29.834270477294922, 'timing/data': 0.008433190174400806, 'timing/model': 0.9543926753103733, 'learning_rate/qwen_vl_interface': 6.462670356273852e-06, 'learning_rate/action_model': 6.462670356273852e-05, 'epoch': 1.35}) +[2026-06-17 21:26:36,526][starVLA.training.train_starvla][INFO] - Step 2984, Loss: {'train/loss': 0.4894133508205414, 'train/grad_norm_pre_clip': 23.603153228759766, 'timing/data': 0.01713031902909279, 'timing/model': 0.9663241170346737, 'learning_rate/qwen_vl_interface': 6.460579206982695e-06, 'learning_rate/action_model': 6.460579206982696e-05, 'epoch': 1.35}) +[2026-06-17 21:26:40,587][starVLA.training.train_starvla][INFO] - Step 2985, Loss: {'train/loss': 0.24168285727500916, 'train/grad_norm_pre_clip': 11.853034019470215, 'timing/data': 0.009874601848423481, 'timing/model': 1.2805735701695085, 'learning_rate/qwen_vl_interface': 6.458487806737166e-06, 'learning_rate/action_model': 6.458487806737165e-05, 'epoch': 1.35}) +[2026-06-17 21:26:44,240][starVLA.training.train_starvla][INFO] - Step 2986, Loss: {'train/loss': 0.4683379530906677, 'train/grad_norm_pre_clip': 20.646240234375, 'timing/data': 0.007615760900080204, 'timing/model': 0.941426151432097, 'learning_rate/qwen_vl_interface': 6.456396155970814e-06, 'learning_rate/action_model': 6.456396155970814e-05, 'epoch': 1.35}) +[2026-06-17 21:26:48,243][starVLA.training.train_starvla][INFO] - Step 2987, Loss: {'train/loss': 0.2507682144641876, 'train/grad_norm_pre_clip': 15.21391773223877, 'timing/data': 0.007978091016411781, 'timing/model': 0.9521909290924668, 'learning_rate/qwen_vl_interface': 6.4543042551172404e-06, 'learning_rate/action_model': 6.45430425511724e-05, 'epoch': 1.36}) +[2026-06-17 21:26:52,137][starVLA.training.train_starvla][INFO] - Step 2988, Loss: {'train/loss': 0.2535528838634491, 'train/grad_norm_pre_clip': 12.650062561035156, 'timing/data': 0.009267650544643402, 'timing/model': 0.9559100130572915, 'learning_rate/qwen_vl_interface': 6.4522121046101e-06, 'learning_rate/action_model': 6.452212104610099e-05, 'epoch': 1.36}) +[2026-06-17 21:26:55,992][starVLA.training.train_starvla][INFO] - Step 2989, Loss: {'train/loss': 0.468683660030365, 'train/grad_norm_pre_clip': 18.183250427246094, 'timing/data': 0.02122076228260994, 'timing/model': 0.9355821032077074, 'learning_rate/qwen_vl_interface': 6.450119704883095e-06, 'learning_rate/action_model': 6.450119704883096e-05, 'epoch': 1.36}) +[2026-06-17 21:27:00,715][starVLA.training.train_starvla][INFO] - Step 2990, Loss: {'train/loss': 0.23352965712547302, 'train/grad_norm_pre_clip': 16.83440589904785, 'timing/data': 0.009477771818637848, 'timing/model': 1.8032599491998553, 'learning_rate/qwen_vl_interface': 6.448027056369983e-06, 'learning_rate/action_model': 6.448027056369983e-05, 'epoch': 1.36}) +[2026-06-17 21:27:04,685][starVLA.training.train_starvla][INFO] - Step 2991, Loss: {'train/loss': 0.29043737053871155, 'train/grad_norm_pre_clip': 15.015900611877441, 'timing/data': 0.011232415214180946, 'timing/model': 1.246137160807848, 'learning_rate/qwen_vl_interface': 6.445934159504576e-06, 'learning_rate/action_model': 6.445934159504576e-05, 'epoch': 1.36}) +[2026-06-17 21:27:08,608][starVLA.training.train_starvla][INFO] - Step 2992, Loss: {'train/loss': 0.3827158212661743, 'train/grad_norm_pre_clip': 15.7998046875, 'timing/data': 0.008360481821000576, 'timing/model': 1.1408542254939675, 'learning_rate/qwen_vl_interface': 6.44384101472073e-06, 'learning_rate/action_model': 6.443841014720729e-05, 'epoch': 1.36}) +[2026-06-17 21:27:12,231][starVLA.training.train_starvla][INFO] - Step 2993, Loss: {'train/loss': 0.3578979969024658, 'train/grad_norm_pre_clip': 17.741111755371094, 'timing/data': 0.008168607018887997, 'timing/model': 0.9374637613072991, 'learning_rate/qwen_vl_interface': 6.441747622452357e-06, 'learning_rate/action_model': 6.441747622452356e-05, 'epoch': 1.36}) +[2026-06-17 21:27:16,234][starVLA.training.train_starvla][INFO] - Step 2994, Loss: {'train/loss': 0.23277808725833893, 'train/grad_norm_pre_clip': 19.670644760131836, 'timing/data': 0.010107151232659817, 'timing/model': 0.9426865428686142, 'learning_rate/qwen_vl_interface': 6.4396539831334195e-06, 'learning_rate/action_model': 6.439653983133419e-05, 'epoch': 1.36}) +[2026-06-17 21:27:20,082][starVLA.training.train_starvla][INFO] - Step 2995, Loss: {'train/loss': 0.1522216945886612, 'train/grad_norm_pre_clip': 10.720894813537598, 'timing/data': 0.008959030732512474, 'timing/model': 0.9426493737846613, 'learning_rate/qwen_vl_interface': 6.437560097197928e-06, 'learning_rate/action_model': 6.437560097197927e-05, 'epoch': 1.36}) +[2026-06-17 21:27:23,934][starVLA.training.train_starvla][INFO] - Step 2996, Loss: {'train/loss': 0.41142886877059937, 'train/grad_norm_pre_clip': 17.139633178710938, 'timing/data': 0.010110422968864441, 'timing/model': 0.9475043276324868, 'learning_rate/qwen_vl_interface': 6.4354659650799546e-06, 'learning_rate/action_model': 6.435465965079955e-05, 'epoch': 1.36}) +[2026-06-17 21:27:27,754][starVLA.training.train_starvla][INFO] - Step 2997, Loss: {'train/loss': 0.3270932734012604, 'train/grad_norm_pre_clip': 17.759681701660156, 'timing/data': 0.013277706690132618, 'timing/model': 0.9338823081925511, 'learning_rate/qwen_vl_interface': 6.433371587213609e-06, 'learning_rate/action_model': 6.433371587213607e-05, 'epoch': 1.36}) +[2026-06-17 21:27:31,630][starVLA.training.train_starvla][INFO] - Step 2998, Loss: {'train/loss': 0.3201076090335846, 'train/grad_norm_pre_clip': 18.51475715637207, 'timing/data': 0.01317060086876154, 'timing/model': 1.1441694451496005, 'learning_rate/qwen_vl_interface': 6.431276964033062e-06, 'learning_rate/action_model': 6.431276964033062e-05, 'epoch': 1.36}) +[2026-06-17 21:27:35,563][starVLA.training.train_starvla][INFO] - Step 2999, Loss: {'train/loss': 0.15052413940429688, 'train/grad_norm_pre_clip': 12.416182518005371, 'timing/data': 0.009075753390789032, 'timing/model': 1.2159549910575151, 'learning_rate/qwen_vl_interface': 6.429182095972526e-06, 'learning_rate/action_model': 6.429182095972526e-05, 'epoch': 1.36}) +[2026-06-17 21:29:55,162][starVLA.training.train_starvla][INFO] - Step 3000, Loss: {'train/loss': 0.282379150390625, 'train/grad_norm_pre_clip': 14.973878860473633, 'eval/loss': 0.42012539505958557, 'eval/latency_8/loss': 0.42012539505958557, 'eval/demon_fire_precision': 0.9175729822552948, 'eval/demon_fire_recall': 0.9959614787200994, 'eval/demon_fire_f1': 0.9551616266944735, 'eval/demon_fire_rate_model': 0.9721758486366165, 'eval/demon_fire_rate_teacher': 0.8956594323873122, 'eval/demon_left_precision': 0.7729357798165137, 'eval/demon_left_recall': 0.721627408993576, 'eval/demon_left_f1': 0.7464008859357697, 'eval/demon_left_rate_model': 0.1213132999443517, 'eval/demon_left_rate_teacher': 0.12993878686700056, 'eval/demon_right_precision': 0.8455598455598455, 'eval/demon_right_recall': 0.6636363636363637, 'eval/demon_right_f1': 0.7436332767402377, 'eval/demon_right_rate_model': 0.1441291040623261, 'eval/demon_right_rate_teacher': 0.18363939899833054, 'eval/demon_move_f1': 0.7450170813380037, 'eval/demon_cc_f1': 0.8806417041913994, 'eval/latency_8/demon_fire_precision': 0.9175729822552948, 'eval/latency_8/demon_fire_recall': 0.9959614787200994, 'eval/latency_8/demon_fire_f1': 0.9551616266944735, 'eval/latency_8/demon_fire_rate_model': 0.9721758486366165, 'eval/latency_8/demon_fire_rate_teacher': 0.8956594323873122, 'eval/latency_8/demon_left_precision': 0.7729357798165137, 'eval/latency_8/demon_left_recall': 0.721627408993576, 'eval/latency_8/demon_left_f1': 0.7464008859357697, 'eval/latency_8/demon_left_rate_model': 0.1213132999443517, 'eval/latency_8/demon_left_rate_teacher': 0.12993878686700056, 'eval/latency_8/demon_right_precision': 0.8455598455598455, 'eval/latency_8/demon_right_recall': 0.6636363636363637, 'eval/latency_8/demon_right_f1': 0.7436332767402377, 'eval/latency_8/demon_right_rate_model': 0.1441291040623261, 'eval/latency_8/demon_right_rate_teacher': 0.18363939899833054, 'eval/latency_8/demon_move_f1': 0.7450170813380037, 'eval/latency_8/demon_cc_f1': 0.8806417041913994, 'timing/data': 0.014501686207950115, 'timing/model': 1.0181688135489821, 'learning_rate/qwen_vl_interface': 6.427086983466275e-06, 'learning_rate/action_model': 6.427086983466275e-05, 'epoch': 1.36}) +[2026-06-17 21:29:55,163][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_3000_state +[2026-06-17 21:29:55,184][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 21:30:01,760][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_3000_state/model.safetensors +[2026-06-17 21:30:17,820][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_3000_state/optimizer.bin +[2026-06-17 21:30:17,821][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_3000_state/sampler.bin +[2026-06-17 21:30:17,821][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_3000_state/sampler_1.bin +[2026-06-17 21:30:17,824][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_3000_state/random_states_0.pkl +[2026-06-17 21:30:17,829][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3000 +[2026-06-17 21:30:17,829][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 21:30:17,841][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 21:30:21,840][starVLA.training.train_starvla][INFO] - Step 3001, Loss: {'train/loss': 0.3641837239265442, 'train/grad_norm_pre_clip': 17.477741241455078, 'timing/data': 0.009946286678314209, 'timing/model': 1.213712370954454, 'learning_rate/qwen_vl_interface': 6.4249916269486255e-06, 'learning_rate/action_model': 6.424991626948626e-05, 'epoch': 1.36}) +[2026-06-17 21:30:25,641][starVLA.training.train_starvla][INFO] - Step 3002, Loss: {'train/loss': 0.8208589553833008, 'train/grad_norm_pre_clip': 22.047903060913086, 'timing/data': 0.010458811186254025, 'timing/model': 0.9420640263706446, 'learning_rate/qwen_vl_interface': 6.422896026853948e-06, 'learning_rate/action_model': 6.422896026853947e-05, 'epoch': 1.36}) +[2026-06-17 21:30:30,866][starVLA.training.train_starvla][INFO] - Step 3003, Loss: {'train/loss': 0.6363849639892578, 'train/grad_norm_pre_clip': 18.497852325439453, 'timing/data': 0.008951687254011631, 'timing/model': 1.0072306832298636, 'learning_rate/qwen_vl_interface': 6.420800183616661e-06, 'learning_rate/action_model': 6.420800183616661e-05, 'epoch': 1.36}) +[2026-06-17 21:30:35,622][starVLA.training.train_starvla][INFO] - Step 3004, Loss: {'train/loss': 0.7273926734924316, 'train/grad_norm_pre_clip': 24.375207901000977, 'timing/data': 0.010327501222491264, 'timing/model': 1.0017976090312004, 'learning_rate/qwen_vl_interface': 6.4187040976712355e-06, 'learning_rate/action_model': 6.418704097671235e-05, 'epoch': 1.36}) +[2026-06-17 21:30:39,442][starVLA.training.train_starvla][INFO] - Step 3005, Loss: {'train/loss': 0.17655058205127716, 'train/grad_norm_pre_clip': 11.867588996887207, 'timing/data': 0.00858725793659687, 'timing/model': 0.9348526056855917, 'learning_rate/qwen_vl_interface': 6.416607769452195e-06, 'learning_rate/action_model': 6.416607769452194e-05, 'epoch': 1.36}) +[2026-06-17 21:30:43,311][starVLA.training.train_starvla][INFO] - Step 3006, Loss: {'train/loss': 0.39909639954566956, 'train/grad_norm_pre_clip': 20.969022750854492, 'timing/data': 0.012501941993832588, 'timing/model': 0.9440711410716176, 'learning_rate/qwen_vl_interface': 6.414511199394106e-06, 'learning_rate/action_model': 6.414511199394105e-05, 'epoch': 1.36}) +[2026-06-17 21:30:47,546][starVLA.training.train_starvla][INFO] - Step 3007, Loss: {'train/loss': 0.6592251062393188, 'train/grad_norm_pre_clip': 25.17105484008789, 'timing/data': 0.01620140764862299, 'timing/model': 1.2913690069690347, 'learning_rate/qwen_vl_interface': 6.4124143879315935e-06, 'learning_rate/action_model': 6.412414387931593e-05, 'epoch': 1.36}) +[2026-06-17 21:30:51,256][starVLA.training.train_starvla][INFO] - Step 3008, Loss: {'train/loss': 0.29647505283355713, 'train/grad_norm_pre_clip': 16.73004150390625, 'timing/data': 0.009419973939657211, 'timing/model': 0.9521806798875332, 'learning_rate/qwen_vl_interface': 6.410317335499328e-06, 'learning_rate/action_model': 6.410317335499328e-05, 'epoch': 1.36}) +[2026-06-17 21:30:55,992][starVLA.training.train_starvla][INFO] - Step 3009, Loss: {'train/loss': 0.20562289655208588, 'train/grad_norm_pre_clip': 14.763821601867676, 'timing/data': 0.008959931321442127, 'timing/model': 0.9927710862830281, 'learning_rate/qwen_vl_interface': 6.40822004253203e-06, 'learning_rate/action_model': 6.40822004253203e-05, 'epoch': 1.37}) +[2026-06-17 21:30:59,858][starVLA.training.train_starvla][INFO] - Step 3010, Loss: {'train/loss': 0.2986622154712677, 'train/grad_norm_pre_clip': 13.766218185424805, 'timing/data': 0.00993392150849104, 'timing/model': 0.9306844649836421, 'learning_rate/qwen_vl_interface': 6.406122509464471e-06, 'learning_rate/action_model': 6.40612250946447e-05, 'epoch': 1.37}) +[2026-06-17 21:31:03,695][starVLA.training.train_starvla][INFO] - Step 3011, Loss: {'train/loss': 0.24249441921710968, 'train/grad_norm_pre_clip': 19.771934509277344, 'timing/data': 0.009959802962839603, 'timing/model': 0.9424954438582063, 'learning_rate/qwen_vl_interface': 6.404024736731472e-06, 'learning_rate/action_model': 6.404024736731472e-05, 'epoch': 1.37}) +[2026-06-17 21:31:07,608][starVLA.training.train_starvla][INFO] - Step 3012, Loss: {'train/loss': 0.26509514451026917, 'train/grad_norm_pre_clip': 14.82925033569336, 'timing/data': 0.008959321305155754, 'timing/model': 0.9431443680077791, 'learning_rate/qwen_vl_interface': 6.401926724767903e-06, 'learning_rate/action_model': 6.401926724767902e-05, 'epoch': 1.37}) +[2026-06-17 21:31:11,544][starVLA.training.train_starvla][INFO] - Step 3013, Loss: {'train/loss': 0.5272916555404663, 'train/grad_norm_pre_clip': 30.355512619018555, 'timing/data': 0.009155327454209328, 'timing/model': 0.9412784026935697, 'learning_rate/qwen_vl_interface': 6.399828474008683e-06, 'learning_rate/action_model': 6.399828474008682e-05, 'epoch': 1.37}) +[2026-06-17 21:31:15,455][starVLA.training.train_starvla][INFO] - Step 3014, Loss: {'train/loss': 0.5016072988510132, 'train/grad_norm_pre_clip': 17.942960739135742, 'timing/data': 0.008336391299962997, 'timing/model': 1.1325538381934166, 'learning_rate/qwen_vl_interface': 6.397729984888785e-06, 'learning_rate/action_model': 6.397729984888785e-05, 'epoch': 1.37}) +[2026-06-17 21:31:19,245][starVLA.training.train_starvla][INFO] - Step 3015, Loss: {'train/loss': 0.4154193103313446, 'train/grad_norm_pre_clip': 15.123160362243652, 'timing/data': 0.008206355385482311, 'timing/model': 0.9455564990639687, 'learning_rate/qwen_vl_interface': 6.395631257843224e-06, 'learning_rate/action_model': 6.395631257843224e-05, 'epoch': 1.37}) +[2026-06-17 21:31:23,394][starVLA.training.train_starvla][INFO] - Step 3016, Loss: {'train/loss': 0.3753589689731598, 'train/grad_norm_pre_clip': 16.74308967590332, 'timing/data': 0.00816630944609642, 'timing/model': 0.9500803621485829, 'learning_rate/qwen_vl_interface': 6.393532293307073e-06, 'learning_rate/action_model': 6.393532293307072e-05, 'epoch': 1.37}) +[2026-06-17 21:31:27,358][starVLA.training.train_starvla][INFO] - Step 3017, Loss: {'train/loss': 0.45965448021888733, 'train/grad_norm_pre_clip': 19.611433029174805, 'timing/data': 0.008211968466639519, 'timing/model': 0.9467967776581645, 'learning_rate/qwen_vl_interface': 6.391433091715447e-06, 'learning_rate/action_model': 6.391433091715446e-05, 'epoch': 1.37}) +[2026-06-17 21:31:31,298][starVLA.training.train_starvla][INFO] - Step 3018, Loss: {'train/loss': 0.20475922524929047, 'train/grad_norm_pre_clip': 12.262630462646484, 'timing/data': 0.0074070533737540245, 'timing/model': 0.9470797842368484, 'learning_rate/qwen_vl_interface': 6.389333653503511e-06, 'learning_rate/action_model': 6.38933365350351e-05, 'epoch': 1.37}) +[2026-06-17 21:31:35,349][starVLA.training.train_starvla][INFO] - Step 3019, Loss: {'train/loss': 0.3945273458957672, 'train/grad_norm_pre_clip': 23.890140533447266, 'timing/data': 0.007133805193006992, 'timing/model': 0.950926935300231, 'learning_rate/qwen_vl_interface': 6.387233979106484e-06, 'learning_rate/action_model': 6.387233979106483e-05, 'epoch': 1.37}) +[2026-06-17 21:31:40,716][starVLA.training.train_starvla][INFO] - Step 3020, Loss: {'train/loss': 0.5279678106307983, 'train/grad_norm_pre_clip': 21.53611946105957, 'timing/data': 0.008563702926039696, 'timing/model': 1.1483909040689468, 'learning_rate/qwen_vl_interface': 6.385134068959629e-06, 'learning_rate/action_model': 6.385134068959628e-05, 'epoch': 1.37}) +[2026-06-17 21:31:45,745][starVLA.training.train_starvla][INFO] - Step 3021, Loss: {'train/loss': 0.23728309571743011, 'train/grad_norm_pre_clip': 10.631848335266113, 'timing/data': 0.012128650210797787, 'timing/model': 1.189398406073451, 'learning_rate/qwen_vl_interface': 6.3830339234982596e-06, 'learning_rate/action_model': 6.38303392349826e-05, 'epoch': 1.37}) +[2026-06-17 21:31:50,683][starVLA.training.train_starvla][INFO] - Step 3022, Loss: {'train/loss': 0.2872238755226135, 'train/grad_norm_pre_clip': 14.098567008972168, 'timing/data': 0.010080705396831036, 'timing/model': 0.9566453704610467, 'learning_rate/qwen_vl_interface': 6.38093354315774e-06, 'learning_rate/action_model': 6.38093354315774e-05, 'epoch': 1.37}) +[2026-06-17 21:31:55,352][starVLA.training.train_starvla][INFO] - Step 3023, Loss: {'train/loss': 0.14867834746837616, 'train/grad_norm_pre_clip': 9.053766250610352, 'timing/data': 0.009235592558979988, 'timing/model': 0.9490049751475453, 'learning_rate/qwen_vl_interface': 6.378832928373479e-06, 'learning_rate/action_model': 6.378832928373478e-05, 'epoch': 1.37}) +[2026-06-17 21:31:59,237][starVLA.training.train_starvla][INFO] - Step 3024, Loss: {'train/loss': 0.39726677536964417, 'train/grad_norm_pre_clip': 17.356321334838867, 'timing/data': 0.007412474602460861, 'timing/model': 0.9446777775883675, 'learning_rate/qwen_vl_interface': 6.376732079580936e-06, 'learning_rate/action_model': 6.376732079580936e-05, 'epoch': 1.37}) +[2026-06-17 21:32:03,112][starVLA.training.train_starvla][INFO] - Step 3025, Loss: {'train/loss': 0.4276875853538513, 'train/grad_norm_pre_clip': 20.1660099029541, 'timing/data': 0.008539771661162376, 'timing/model': 0.9525739587843418, 'learning_rate/qwen_vl_interface': 6.374630997215623e-06, 'learning_rate/action_model': 6.374630997215623e-05, 'epoch': 1.37}) +[2026-06-17 21:32:07,076][starVLA.training.train_starvla][INFO] - Step 3026, Loss: {'train/loss': 0.34372779726982117, 'train/grad_norm_pre_clip': 17.761594772338867, 'timing/data': 0.008830086328089237, 'timing/model': 0.956780381500721, 'learning_rate/qwen_vl_interface': 6.372529681713093e-06, 'learning_rate/action_model': 6.372529681713094e-05, 'epoch': 1.37}) +[2026-06-17 21:32:11,357][starVLA.training.train_starvla][INFO] - Step 3027, Loss: {'train/loss': 0.1817673146724701, 'train/grad_norm_pre_clip': 7.97730827331543, 'timing/data': 0.010646420530974865, 'timing/model': 1.1650861138477921, 'learning_rate/qwen_vl_interface': 6.370428133508953e-06, 'learning_rate/action_model': 6.370428133508953e-05, 'epoch': 1.37}) +[2026-06-17 21:32:15,062][starVLA.training.train_starvla][INFO] - Step 3028, Loss: {'train/loss': 0.31940579414367676, 'train/grad_norm_pre_clip': 16.73123550415039, 'timing/data': 0.012899346649646759, 'timing/model': 0.952697416767478, 'learning_rate/qwen_vl_interface': 6.368326353038852e-06, 'learning_rate/action_model': 6.368326353038853e-05, 'epoch': 1.37}) +[2026-06-17 21:32:18,952][starVLA.training.train_starvla][INFO] - Step 3029, Loss: {'train/loss': 0.3107988238334656, 'train/grad_norm_pre_clip': 21.957719802856445, 'timing/data': 0.008062448352575302, 'timing/model': 0.9397747525945306, 'learning_rate/qwen_vl_interface': 6.366224340738499e-06, 'learning_rate/action_model': 6.366224340738499e-05, 'epoch': 1.37}) +[2026-06-17 21:32:22,768][starVLA.training.train_starvla][INFO] - Step 3030, Loss: {'train/loss': 0.3946569859981537, 'train/grad_norm_pre_clip': 19.11182403564453, 'timing/data': 0.010271307080984116, 'timing/model': 0.9583266843110323, 'learning_rate/qwen_vl_interface': 6.3641220970436355e-06, 'learning_rate/action_model': 6.364122097043635e-05, 'epoch': 1.37}) +[2026-06-17 21:32:26,929][starVLA.training.train_starvla][INFO] - Step 3031, Loss: {'train/loss': 0.6000752449035645, 'train/grad_norm_pre_clip': 25.974979400634766, 'timing/data': 0.012852097861468792, 'timing/model': 0.9607665725052357, 'learning_rate/qwen_vl_interface': 6.362019622390065e-06, 'learning_rate/action_model': 6.362019622390066e-05, 'epoch': 1.38}) +[2026-06-17 21:32:31,050][starVLA.training.train_starvla][INFO] - Step 3032, Loss: {'train/loss': 0.5180870890617371, 'train/grad_norm_pre_clip': 26.417264938354492, 'timing/data': 0.009022311307489872, 'timing/model': 0.959079073742032, 'learning_rate/qwen_vl_interface': 6.359916917213629e-06, 'learning_rate/action_model': 6.359916917213629e-05, 'epoch': 1.38}) +[2026-06-17 21:32:34,998][starVLA.training.train_starvla][INFO] - Step 3033, Loss: {'train/loss': 0.42570558190345764, 'train/grad_norm_pre_clip': 14.493769645690918, 'timing/data': 0.01157423947006464, 'timing/model': 1.1411409555003047, 'learning_rate/qwen_vl_interface': 6.357813981950223e-06, 'learning_rate/action_model': 6.357813981950223e-05, 'epoch': 1.38}) +[2026-06-17 21:32:38,915][starVLA.training.train_starvla][INFO] - Step 3034, Loss: {'train/loss': 0.5179840326309204, 'train/grad_norm_pre_clip': 22.113346099853516, 'timing/data': 0.008547506295144558, 'timing/model': 1.164122418500483, 'learning_rate/qwen_vl_interface': 6.355710817035784e-06, 'learning_rate/action_model': 6.355710817035783e-05, 'epoch': 1.38}) +[2026-06-17 21:32:43,227][starVLA.training.train_starvla][INFO] - Step 3035, Loss: {'train/loss': 0.2860044836997986, 'train/grad_norm_pre_clip': 13.296014785766602, 'timing/data': 0.014412714168429375, 'timing/model': 0.9656322421506047, 'learning_rate/qwen_vl_interface': 6.353607422906304e-06, 'learning_rate/action_model': 6.353607422906304e-05, 'epoch': 1.38}) +[2026-06-17 21:32:49,975][starVLA.training.train_starvla][INFO] - Step 3036, Loss: {'train/loss': 0.39767688512802124, 'train/grad_norm_pre_clip': 14.102470397949219, 'timing/data': 0.011543860659003258, 'timing/model': 0.9522777963429689, 'learning_rate/qwen_vl_interface': 6.351503799997819e-06, 'learning_rate/action_model': 6.351503799997818e-05, 'epoch': 1.38}) +[2026-06-17 21:32:56,574][starVLA.training.train_starvla][INFO] - Step 3037, Loss: {'train/loss': 0.46032577753067017, 'train/grad_norm_pre_clip': 17.093948364257812, 'timing/data': 0.01056066807359457, 'timing/model': 2.0064339069649577, 'learning_rate/qwen_vl_interface': 6.349399948746409e-06, 'learning_rate/action_model': 6.349399948746409e-05, 'epoch': 1.38}) +[2026-06-17 21:33:02,930][starVLA.training.train_starvla][INFO] - Step 3038, Loss: {'train/loss': 0.3961065709590912, 'train/grad_norm_pre_clip': 20.419937133789062, 'timing/data': 0.019426051527261734, 'timing/model': 2.456290134228766, 'learning_rate/qwen_vl_interface': 6.347295869588208e-06, 'learning_rate/action_model': 6.347295869588208e-05, 'epoch': 1.38}) +[2026-06-17 21:33:06,903][starVLA.training.train_starvla][INFO] - Step 3039, Loss: {'train/loss': 0.507857620716095, 'train/grad_norm_pre_clip': 20.02661895751953, 'timing/data': 0.009123009629547596, 'timing/model': 0.9671213980764151, 'learning_rate/qwen_vl_interface': 6.345191562959392e-06, 'learning_rate/action_model': 6.345191562959391e-05, 'epoch': 1.38}) +[2026-06-17 21:33:10,850][starVLA.training.train_starvla][INFO] - Step 3040, Loss: {'train/loss': 0.5452530384063721, 'train/grad_norm_pre_clip': 25.37664794921875, 'timing/data': 0.011820358224213123, 'timing/model': 1.179825745522976, 'learning_rate/qwen_vl_interface': 6.343087029296187e-06, 'learning_rate/action_model': 6.343087029296186e-05, 'epoch': 1.38}) +[2026-06-17 21:33:14,922][starVLA.training.train_starvla][INFO] - Step 3041, Loss: {'train/loss': 0.22786062955856323, 'train/grad_norm_pre_clip': 12.769097328186035, 'timing/data': 0.007589825429022312, 'timing/model': 1.2386710634455085, 'learning_rate/qwen_vl_interface': 6.340982269034863e-06, 'learning_rate/action_model': 6.340982269034863e-05, 'epoch': 1.38}) +[2026-06-17 21:33:18,739][starVLA.training.train_starvla][INFO] - Step 3042, Loss: {'train/loss': 0.6223850846290588, 'train/grad_norm_pre_clip': 22.18174171447754, 'timing/data': 0.008424101397395134, 'timing/model': 0.9432942280545831, 'learning_rate/qwen_vl_interface': 6.3388772826117415e-06, 'learning_rate/action_model': 6.338877282611742e-05, 'epoch': 1.38}) +[2026-06-17 21:33:22,643][starVLA.training.train_starvla][INFO] - Step 3043, Loss: {'train/loss': 0.4554869532585144, 'train/grad_norm_pre_clip': 21.2406063079834, 'timing/data': 0.009362703189253807, 'timing/model': 0.9532673545181751, 'learning_rate/qwen_vl_interface': 6.3367720704631895e-06, 'learning_rate/action_model': 6.33677207046319e-05, 'epoch': 1.38}) +[2026-06-17 21:33:26,473][starVLA.training.train_starvla][INFO] - Step 3044, Loss: {'train/loss': 0.41752344369888306, 'train/grad_norm_pre_clip': 17.48045539855957, 'timing/data': 0.009104819968342781, 'timing/model': 0.9456402882933617, 'learning_rate/qwen_vl_interface': 6.334666633025616e-06, 'learning_rate/action_model': 6.334666633025615e-05, 'epoch': 1.38}) +[2026-06-17 21:33:30,344][starVLA.training.train_starvla][INFO] - Step 3045, Loss: {'train/loss': 0.5388497710227966, 'train/grad_norm_pre_clip': 24.663179397583008, 'timing/data': 0.009827397763729095, 'timing/model': 0.940328711643815, 'learning_rate/qwen_vl_interface': 6.332560970735482e-06, 'learning_rate/action_model': 6.332560970735482e-05, 'epoch': 1.38}) +[2026-06-17 21:33:34,253][starVLA.training.train_starvla][INFO] - Step 3046, Loss: {'train/loss': 0.5451290607452393, 'train/grad_norm_pre_clip': 20.64742088317871, 'timing/data': 0.010180257260799408, 'timing/model': 0.9527431912720203, 'learning_rate/qwen_vl_interface': 6.3304550840292945e-06, 'learning_rate/action_model': 6.330455084029294e-05, 'epoch': 1.38}) +[2026-06-17 21:33:38,127][starVLA.training.train_starvla][INFO] - Step 3047, Loss: {'train/loss': 0.352059006690979, 'train/grad_norm_pre_clip': 22.52486801147461, 'timing/data': 0.009679055772721767, 'timing/model': 1.1373874424025416, 'learning_rate/qwen_vl_interface': 6.328348973343605e-06, 'learning_rate/action_model': 6.328348973343605e-05, 'epoch': 1.38}) +[2026-06-17 21:33:41,792][starVLA.training.train_starvla][INFO] - Step 3048, Loss: {'train/loss': 0.5296458601951599, 'train/grad_norm_pre_clip': 24.334487915039062, 'timing/data': 0.010820586234331131, 'timing/model': 0.9497303934767842, 'learning_rate/qwen_vl_interface': 6.326242639115014e-06, 'learning_rate/action_model': 6.326242639115015e-05, 'epoch': 1.38}) +[2026-06-17 21:33:45,634][starVLA.training.train_starvla][INFO] - Step 3049, Loss: {'train/loss': 0.23500671982765198, 'train/grad_norm_pre_clip': 19.476760864257812, 'timing/data': 0.010634436272084713, 'timing/model': 0.9458195278421044, 'learning_rate/qwen_vl_interface': 6.324136081780166e-06, 'learning_rate/action_model': 6.324136081780165e-05, 'epoch': 1.38}) +[2026-06-17 21:33:49,476][starVLA.training.train_starvla][INFO] - Step 3050, Loss: {'train/loss': 0.4824064075946808, 'train/grad_norm_pre_clip': 19.99752426147461, 'timing/data': 0.009660901501774788, 'timing/model': 0.9660839168354869, 'learning_rate/qwen_vl_interface': 6.322029301775751e-06, 'learning_rate/action_model': 6.32202930177575e-05, 'epoch': 1.38}) +[2026-06-17 21:33:53,347][starVLA.training.train_starvla][INFO] - Step 3051, Loss: {'train/loss': 0.4878444969654083, 'train/grad_norm_pre_clip': 22.826663970947266, 'timing/data': 0.010147899389266968, 'timing/model': 0.9398961011320353, 'learning_rate/qwen_vl_interface': 6.3199222995385105e-06, 'learning_rate/action_model': 6.31992229953851e-05, 'epoch': 1.38}) +[2026-06-17 21:33:57,501][starVLA.training.train_starvla][INFO] - Step 3052, Loss: {'train/loss': 1.0747555494308472, 'train/grad_norm_pre_clip': 25.089435577392578, 'timing/data': 0.010049195028841496, 'timing/model': 1.0177292125299573, 'learning_rate/qwen_vl_interface': 6.317815075505223e-06, 'learning_rate/action_model': 6.317815075505222e-05, 'epoch': 1.38}) +[2026-06-17 21:34:01,338][starVLA.training.train_starvla][INFO] - Step 3053, Loss: {'train/loss': 0.13778510689735413, 'train/grad_norm_pre_clip': 8.403541564941406, 'timing/data': 0.015283472836017609, 'timing/model': 1.144645377062261, 'learning_rate/qwen_vl_interface': 6.315707630112724e-06, 'learning_rate/action_model': 6.315707630112723e-05, 'epoch': 1.39}) +[2026-06-17 21:34:05,413][starVLA.training.train_starvla][INFO] - Step 3054, Loss: {'train/loss': 0.6088685393333435, 'train/grad_norm_pre_clip': 47.76517868041992, 'timing/data': 0.011601652950048447, 'timing/model': 1.1910518584772944, 'learning_rate/qwen_vl_interface': 6.313599963797884e-06, 'learning_rate/action_model': 6.313599963797884e-05, 'epoch': 1.39}) +[2026-06-17 21:34:09,123][starVLA.training.train_starvla][INFO] - Step 3055, Loss: {'train/loss': 0.41847196221351624, 'train/grad_norm_pre_clip': 16.212886810302734, 'timing/data': 0.009653045795857906, 'timing/model': 0.9639460369944572, 'learning_rate/qwen_vl_interface': 6.311492076997629e-06, 'learning_rate/action_model': 6.311492076997629e-05, 'epoch': 1.39}) +[2026-06-17 21:34:13,017][starVLA.training.train_starvla][INFO] - Step 3056, Loss: {'train/loss': 0.21056026220321655, 'train/grad_norm_pre_clip': 12.425131797790527, 'timing/data': 0.008246063254773617, 'timing/model': 0.9536474347114563, 'learning_rate/qwen_vl_interface': 6.309383970148922e-06, 'learning_rate/action_model': 6.309383970148922e-05, 'epoch': 1.39}) +[2026-06-17 21:34:16,890][starVLA.training.train_starvla][INFO] - Step 3057, Loss: {'train/loss': 0.14378122985363007, 'train/grad_norm_pre_clip': 13.768125534057617, 'timing/data': 0.011911693960428238, 'timing/model': 0.9486054982990026, 'learning_rate/qwen_vl_interface': 6.3072756436887805e-06, 'learning_rate/action_model': 6.30727564368878e-05, 'epoch': 1.39}) +[2026-06-17 21:34:20,739][starVLA.training.train_starvla][INFO] - Step 3058, Loss: {'train/loss': 0.1756950467824936, 'train/grad_norm_pre_clip': 11.12911319732666, 'timing/data': 0.009731614030897617, 'timing/model': 0.9331471547484398, 'learning_rate/qwen_vl_interface': 6.30516709805426e-06, 'learning_rate/action_model': 6.30516709805426e-05, 'epoch': 1.39}) +[2026-06-17 21:34:24,680][starVLA.training.train_starvla][INFO] - Step 3059, Loss: {'train/loss': 0.4146154820919037, 'train/grad_norm_pre_clip': 19.578367233276367, 'timing/data': 0.008641121909022331, 'timing/model': 0.9731640517711639, 'learning_rate/qwen_vl_interface': 6.303058333682464e-06, 'learning_rate/action_model': 6.303058333682463e-05, 'epoch': 1.39}) +[2026-06-17 21:34:28,690][starVLA.training.train_starvla][INFO] - Step 3060, Loss: {'train/loss': 0.2251662015914917, 'train/grad_norm_pre_clip': 15.972925186157227, 'timing/data': 0.014767858199775219, 'timing/model': 1.1574002783745527, 'learning_rate/qwen_vl_interface': 6.300949351010542e-06, 'learning_rate/action_model': 6.300949351010543e-05, 'epoch': 1.39}) +[2026-06-17 21:34:32,573][starVLA.training.train_starvla][INFO] - Step 3061, Loss: {'train/loss': 0.10433291643857956, 'train/grad_norm_pre_clip': 7.661121845245361, 'timing/data': 0.015582979656755924, 'timing/model': 1.1440238635987043, 'learning_rate/qwen_vl_interface': 6.298840150475689e-06, 'learning_rate/action_model': 6.298840150475689e-05, 'epoch': 1.39}) +[2026-06-17 21:34:36,700][starVLA.training.train_starvla][INFO] - Step 3062, Loss: {'train/loss': 0.49560630321502686, 'train/grad_norm_pre_clip': 24.17889404296875, 'timing/data': 0.016142881475389004, 'timing/model': 1.2777298763394356, 'learning_rate/qwen_vl_interface': 6.296730732515145e-06, 'learning_rate/action_model': 6.296730732515145e-05, 'epoch': 1.39}) +[2026-06-17 21:34:40,706][starVLA.training.train_starvla][INFO] - Step 3063, Loss: {'train/loss': 0.5792681574821472, 'train/grad_norm_pre_clip': 22.823755264282227, 'timing/data': 0.01115186046808958, 'timing/model': 0.9653404979035258, 'learning_rate/qwen_vl_interface': 6.294621097566193e-06, 'learning_rate/action_model': 6.294621097566193e-05, 'epoch': 1.39}) +[2026-06-17 21:34:44,537][starVLA.training.train_starvla][INFO] - Step 3064, Loss: {'train/loss': 0.19397415220737457, 'train/grad_norm_pre_clip': 12.914299964904785, 'timing/data': 0.009359501302242279, 'timing/model': 0.9392896154895425, 'learning_rate/qwen_vl_interface': 6.292511246066163e-06, 'learning_rate/action_model': 6.292511246066163e-05, 'epoch': 1.39}) +[2026-06-17 21:34:48,344][starVLA.training.train_starvla][INFO] - Step 3065, Loss: {'train/loss': 0.43118634819984436, 'train/grad_norm_pre_clip': 18.642791748046875, 'timing/data': 0.007757571525871754, 'timing/model': 0.9419061271473765, 'learning_rate/qwen_vl_interface': 6.290401178452431e-06, 'learning_rate/action_model': 6.29040117845243e-05, 'epoch': 1.39}) +[2026-06-17 21:34:52,278][starVLA.training.train_starvla][INFO] - Step 3066, Loss: {'train/loss': 0.7528460025787354, 'train/grad_norm_pre_clip': 138.03001403808594, 'timing/data': 0.013497516512870789, 'timing/model': 1.0095808384940028, 'learning_rate/qwen_vl_interface': 6.288290895162414e-06, 'learning_rate/action_model': 6.288290895162413e-05, 'epoch': 1.39}) +[2026-06-17 21:34:58,315][starVLA.training.train_starvla][INFO] - Step 3067, Loss: {'train/loss': 0.32509633898735046, 'train/grad_norm_pre_clip': 18.09379005432129, 'timing/data': 0.008253749459981918, 'timing/model': 3.2833262225613, 'learning_rate/qwen_vl_interface': 6.286180396633579e-06, 'learning_rate/action_model': 6.286180396633579e-05, 'epoch': 1.39}) +[2026-06-17 21:35:02,214][starVLA.training.train_starvla][INFO] - Step 3068, Loss: {'train/loss': 0.5724765658378601, 'train/grad_norm_pre_clip': 17.7019100189209, 'timing/data': 0.011730806902050972, 'timing/model': 1.0364408688619733, 'learning_rate/qwen_vl_interface': 6.28406968330343e-06, 'learning_rate/action_model': 6.28406968330343e-05, 'epoch': 1.39}) +[2026-06-17 21:35:06,335][starVLA.training.train_starvla][INFO] - Step 3069, Loss: {'train/loss': 0.26189008355140686, 'train/grad_norm_pre_clip': 15.790380477905273, 'timing/data': 0.014706242829561234, 'timing/model': 1.05829949490726, 'learning_rate/qwen_vl_interface': 6.281958755609526e-06, 'learning_rate/action_model': 6.281958755609526e-05, 'epoch': 1.39}) +[2026-06-17 21:35:10,141][starVLA.training.train_starvla][INFO] - Step 3070, Loss: {'train/loss': 0.3320273756980896, 'train/grad_norm_pre_clip': 22.977373123168945, 'timing/data': 0.013729195110499859, 'timing/model': 0.9540354572236538, 'learning_rate/qwen_vl_interface': 6.279847613989459e-06, 'learning_rate/action_model': 6.279847613989459e-05, 'epoch': 1.39}) +[2026-06-17 21:35:14,128][starVLA.training.train_starvla][INFO] - Step 3071, Loss: {'train/loss': 0.33586856722831726, 'train/grad_norm_pre_clip': 18.590534210205078, 'timing/data': 0.010995582677423954, 'timing/model': 0.9425679901614785, 'learning_rate/qwen_vl_interface': 6.2777362588808745e-06, 'learning_rate/action_model': 6.277736258880874e-05, 'epoch': 1.39}) +[2026-06-17 21:35:17,992][starVLA.training.train_starvla][INFO] - Step 3072, Loss: {'train/loss': 0.27409714460372925, 'train/grad_norm_pre_clip': 20.79109764099121, 'timing/data': 0.009420610032975674, 'timing/model': 0.9574785195291042, 'learning_rate/qwen_vl_interface': 6.275624690721458e-06, 'learning_rate/action_model': 6.275624690721458e-05, 'epoch': 1.39}) +[2026-06-17 21:35:21,872][starVLA.training.train_starvla][INFO] - Step 3073, Loss: {'train/loss': 0.3875109851360321, 'train/grad_norm_pre_clip': 21.797088623046875, 'timing/data': 0.008343221619725227, 'timing/model': 1.1688625738024712, 'learning_rate/qwen_vl_interface': 6.27351290994894e-06, 'learning_rate/action_model': 6.27351290994894e-05, 'epoch': 1.39}) +[2026-06-17 21:35:26,066][starVLA.training.train_starvla][INFO] - Step 3074, Loss: {'train/loss': 0.3447704315185547, 'train/grad_norm_pre_clip': 12.250890731811523, 'timing/data': 0.008940347470343113, 'timing/model': 1.1184857916086912, 'learning_rate/qwen_vl_interface': 6.271400917001092e-06, 'learning_rate/action_model': 6.271400917001092e-05, 'epoch': 1.39}) +[2026-06-17 21:35:29,732][starVLA.training.train_starvla][INFO] - Step 3075, Loss: {'train/loss': 0.23055274784564972, 'train/grad_norm_pre_clip': 18.154626846313477, 'timing/data': 0.009846598841249943, 'timing/model': 0.9419593578204513, 'learning_rate/qwen_vl_interface': 6.269288712315737e-06, 'learning_rate/action_model': 6.269288712315737e-05, 'epoch': 1.4}) +[2026-06-17 21:35:33,614][starVLA.training.train_starvla][INFO] - Step 3076, Loss: {'train/loss': 0.3575347661972046, 'train/grad_norm_pre_clip': 21.96720314025879, 'timing/data': 0.010000319220125675, 'timing/model': 0.9522325117141008, 'learning_rate/qwen_vl_interface': 6.267176296330734e-06, 'learning_rate/action_model': 6.267176296330733e-05, 'epoch': 1.4}) +[2026-06-17 21:35:37,554][starVLA.training.train_starvla][INFO] - Step 3077, Loss: {'train/loss': 0.3216387927532196, 'train/grad_norm_pre_clip': 12.104433059692383, 'timing/data': 0.015096934512257576, 'timing/model': 0.9487512679770589, 'learning_rate/qwen_vl_interface': 6.265063669483992e-06, 'learning_rate/action_model': 6.265063669483991e-05, 'epoch': 1.4}) +[2026-06-17 21:35:41,396][starVLA.training.train_starvla][INFO] - Step 3078, Loss: {'train/loss': 0.22090990841388702, 'train/grad_norm_pre_clip': 13.443644523620605, 'timing/data': 0.01599958259612322, 'timing/model': 0.9568561678752303, 'learning_rate/qwen_vl_interface': 6.262950832213457e-06, 'learning_rate/action_model': 6.262950832213456e-05, 'epoch': 1.4}) +[2026-06-17 21:35:45,712][starVLA.training.train_starvla][INFO] - Step 3079, Loss: {'train/loss': 0.3927290737628937, 'train/grad_norm_pre_clip': 19.240571975708008, 'timing/data': 0.009686683304607868, 'timing/model': 1.0358556853607297, 'learning_rate/qwen_vl_interface': 6.260837784957125e-06, 'learning_rate/action_model': 6.260837784957126e-05, 'epoch': 1.4}) +[2026-06-17 21:35:49,700][starVLA.training.train_starvla][INFO] - Step 3080, Loss: {'train/loss': 0.6062870621681213, 'train/grad_norm_pre_clip': 22.8662166595459, 'timing/data': 0.008422243408858776, 'timing/model': 1.1397296842187643, 'learning_rate/qwen_vl_interface': 6.258724528153034e-06, 'learning_rate/action_model': 6.258724528153033e-05, 'epoch': 1.4}) +[2026-06-17 21:35:53,548][starVLA.training.train_starvla][INFO] - Step 3081, Loss: {'train/loss': 0.3897040784358978, 'train/grad_norm_pre_clip': 21.44049835205078, 'timing/data': 0.011418468318879604, 'timing/model': 1.1307038087397814, 'learning_rate/qwen_vl_interface': 6.256611062239261e-06, 'learning_rate/action_model': 6.25661106223926e-05, 'epoch': 1.4}) +[2026-06-17 21:35:57,219][starVLA.training.train_starvla][INFO] - Step 3082, Loss: {'train/loss': 0.3951781988143921, 'train/grad_norm_pre_clip': 18.128408432006836, 'timing/data': 0.013539723120629787, 'timing/model': 0.9436985300853848, 'learning_rate/qwen_vl_interface': 6.254497387653933e-06, 'learning_rate/action_model': 6.254497387653932e-05, 'epoch': 1.4}) +[2026-06-17 21:36:01,077][starVLA.training.train_starvla][INFO] - Step 3083, Loss: {'train/loss': 0.4472788870334625, 'train/grad_norm_pre_clip': 18.424062728881836, 'timing/data': 0.00929549615830183, 'timing/model': 0.9378524469211698, 'learning_rate/qwen_vl_interface': 6.252383504835213e-06, 'learning_rate/action_model': 6.252383504835213e-05, 'epoch': 1.4}) +[2026-06-17 21:36:05,575][starVLA.training.train_starvla][INFO] - Step 3084, Loss: {'train/loss': 0.1863154172897339, 'train/grad_norm_pre_clip': 11.462141990661621, 'timing/data': 0.011365686543285847, 'timing/model': 0.9507257286459208, 'learning_rate/qwen_vl_interface': 6.250269414221317e-06, 'learning_rate/action_model': 6.250269414221317e-05, 'epoch': 1.4}) +[2026-06-17 21:36:12,505][starVLA.training.train_starvla][INFO] - Step 3085, Loss: {'train/loss': 0.43824514746665955, 'train/grad_norm_pre_clip': 19.12996482849121, 'timing/data': 0.009479142725467682, 'timing/model': 1.0241743195801973, 'learning_rate/qwen_vl_interface': 6.248155116250494e-06, 'learning_rate/action_model': 6.248155116250493e-05, 'epoch': 1.4}) +[2026-06-17 21:36:16,543][starVLA.training.train_starvla][INFO] - Step 3086, Loss: {'train/loss': 0.3848596215248108, 'train/grad_norm_pre_clip': 22.109521865844727, 'timing/data': 0.011745821684598923, 'timing/model': 0.9517176588997245, 'learning_rate/qwen_vl_interface': 6.246040611361042e-06, 'learning_rate/action_model': 6.246040611361041e-05, 'epoch': 1.4}) +[2026-06-17 21:36:20,552][starVLA.training.train_starvla][INFO] - Step 3087, Loss: {'train/loss': 0.20212422311306, 'train/grad_norm_pre_clip': 17.41162109375, 'timing/data': 0.020307214930653572, 'timing/model': 1.1865503005683422, 'learning_rate/qwen_vl_interface': 6.2439258999912976e-06, 'learning_rate/action_model': 6.243925899991298e-05, 'epoch': 1.4}) +[2026-06-17 21:36:24,510][starVLA.training.train_starvla][INFO] - Step 3088, Loss: {'train/loss': 0.37984699010849, 'train/grad_norm_pre_clip': 20.605405807495117, 'timing/data': 0.0097458241507411, 'timing/model': 0.9442403148859739, 'learning_rate/qwen_vl_interface': 6.241810982579646e-06, 'learning_rate/action_model': 6.241810982579646e-05, 'epoch': 1.4}) +[2026-06-17 21:36:28,403][starVLA.training.train_starvla][INFO] - Step 3089, Loss: {'train/loss': 0.2312113493680954, 'train/grad_norm_pre_clip': 15.435068130493164, 'timing/data': 0.016485110856592655, 'timing/model': 0.934473211877048, 'learning_rate/qwen_vl_interface': 6.239695859564511e-06, 'learning_rate/action_model': 6.23969585956451e-05, 'epoch': 1.4}) +[2026-06-17 21:36:32,304][starVLA.training.train_starvla][INFO] - Step 3090, Loss: {'train/loss': 0.6396040916442871, 'train/grad_norm_pre_clip': 21.161245346069336, 'timing/data': 0.0140497712418437, 'timing/model': 0.9521373203024268, 'learning_rate/qwen_vl_interface': 6.2375805313843564e-06, 'learning_rate/action_model': 6.237580531384356e-05, 'epoch': 1.4}) +[2026-06-17 21:36:36,465][starVLA.training.train_starvla][INFO] - Step 3091, Loss: {'train/loss': 0.217702716588974, 'train/grad_norm_pre_clip': 13.399263381958008, 'timing/data': 0.008821958675980568, 'timing/model': 0.9582947352901101, 'learning_rate/qwen_vl_interface': 6.2354649984777e-06, 'learning_rate/action_model': 6.235464998477699e-05, 'epoch': 1.4}) +[2026-06-17 21:36:40,377][starVLA.training.train_starvla][INFO] - Step 3092, Loss: {'train/loss': 0.2658185362815857, 'train/grad_norm_pre_clip': 14.437493324279785, 'timing/data': 0.008232847787439823, 'timing/model': 0.9879887579008937, 'learning_rate/qwen_vl_interface': 6.2333492612830835e-06, 'learning_rate/action_model': 6.233349261283083e-05, 'epoch': 1.4}) +[2026-06-17 21:36:44,257][starVLA.training.train_starvla][INFO] - Step 3093, Loss: {'train/loss': 0.28237488865852356, 'train/grad_norm_pre_clip': 13.550865173339844, 'timing/data': 0.007758624851703644, 'timing/model': 0.9667997108772397, 'learning_rate/qwen_vl_interface': 6.231233320239112e-06, 'learning_rate/action_model': 6.231233320239112e-05, 'epoch': 1.4}) +[2026-06-17 21:36:48,270][starVLA.training.train_starvla][INFO] - Step 3094, Loss: {'train/loss': 0.26796284317970276, 'train/grad_norm_pre_clip': 13.304579734802246, 'timing/data': 0.00894503016024828, 'timing/model': 1.2303801914677024, 'learning_rate/qwen_vl_interface': 6.229117175784414e-06, 'learning_rate/action_model': 6.229117175784413e-05, 'epoch': 1.4}) +[2026-06-17 21:36:51,936][starVLA.training.train_starvla][INFO] - Step 3095, Loss: {'train/loss': 0.5503485798835754, 'train/grad_norm_pre_clip': 25.453369140625, 'timing/data': 0.016806461848318577, 'timing/model': 0.9359644092619419, 'learning_rate/qwen_vl_interface': 6.227000828357672e-06, 'learning_rate/action_model': 6.227000828357672e-05, 'epoch': 1.4}) +[2026-06-17 21:36:55,841][starVLA.training.train_starvla][INFO] - Step 3096, Loss: {'train/loss': 0.25257670879364014, 'train/grad_norm_pre_clip': 25.47258186340332, 'timing/data': 0.014064020477235317, 'timing/model': 0.9747359789907932, 'learning_rate/qwen_vl_interface': 6.2248842783976085e-06, 'learning_rate/action_model': 6.224884278397608e-05, 'epoch': 1.4}) +[2026-06-17 21:37:01,742][starVLA.training.train_starvla][INFO] - Step 3097, Loss: {'train/loss': 0.3911325931549072, 'train/grad_norm_pre_clip': 39.221900939941406, 'timing/data': 0.012280628085136414, 'timing/model': 0.9660362899303436, 'learning_rate/qwen_vl_interface': 6.222767526342983e-06, 'learning_rate/action_model': 6.222767526342983e-05, 'epoch': 1.41}) +[2026-06-17 21:37:07,090][starVLA.training.train_starvla][INFO] - Step 3098, Loss: {'train/loss': 0.40970084071159363, 'train/grad_norm_pre_clip': 22.85703468322754, 'timing/data': 0.010584056377410889, 'timing/model': 2.3662258982658386, 'learning_rate/qwen_vl_interface': 6.220650572632605e-06, 'learning_rate/action_model': 6.220650572632604e-05, 'epoch': 1.41}) +[2026-06-17 21:37:12,176][starVLA.training.train_starvla][INFO] - Step 3099, Loss: {'train/loss': 0.44495806097984314, 'train/grad_norm_pre_clip': 76.2358627319336, 'timing/data': 0.009189355187118053, 'timing/model': 1.0073397811502218, 'learning_rate/qwen_vl_interface': 6.218533417705318e-06, 'learning_rate/action_model': 6.218533417705318e-05, 'epoch': 1.41}) +[2026-06-17 21:39:14,884][starVLA.training.train_starvla][INFO] - Step 3100, Loss: {'train/loss': 0.66650390625, 'train/grad_norm_pre_clip': 22.791194915771484, 'eval/loss': 0.3945449888706207, 'eval/latency_8/loss': 0.3945449888706207, 'eval/demon_fire_precision': 0.9621520175164217, 'eval/demon_fire_recall': 0.9555762659210935, 'eval/demon_fire_f1': 0.9588528678304239, 'eval/demon_fire_rate_model': 0.8895381190873678, 'eval/demon_fire_rate_teacher': 0.8956594323873122, 'eval/demon_left_precision': 0.8128205128205128, 'eval/demon_left_recall': 0.6788008565310493, 'eval/demon_left_f1': 0.7397899649941657, 'eval/demon_left_rate_model': 0.10851419031719532, 'eval/demon_left_rate_teacher': 0.12993878686700056, 'eval/demon_right_precision': 0.8233151183970856, 'eval/demon_right_recall': 0.6848484848484848, 'eval/demon_right_f1': 0.7477253928866832, 'eval/demon_right_rate_model': 0.15275459098497496, 'eval/demon_right_rate_teacher': 0.18363939899833054, 'eval/demon_move_f1': 0.7437576789404244, 'eval/demon_cc_f1': 0.8823040056929743, 'eval/latency_8/demon_fire_precision': 0.9621520175164217, 'eval/latency_8/demon_fire_recall': 0.9555762659210935, 'eval/latency_8/demon_fire_f1': 0.9588528678304239, 'eval/latency_8/demon_fire_rate_model': 0.8895381190873678, 'eval/latency_8/demon_fire_rate_teacher': 0.8956594323873122, 'eval/latency_8/demon_left_precision': 0.8128205128205128, 'eval/latency_8/demon_left_recall': 0.6788008565310493, 'eval/latency_8/demon_left_f1': 0.7397899649941657, 'eval/latency_8/demon_left_rate_model': 0.10851419031719532, 'eval/latency_8/demon_left_rate_teacher': 0.12993878686700056, 'eval/latency_8/demon_right_precision': 0.8233151183970856, 'eval/latency_8/demon_right_recall': 0.6848484848484848, 'eval/latency_8/demon_right_f1': 0.7477253928866832, 'eval/latency_8/demon_right_rate_model': 0.15275459098497496, 'eval/latency_8/demon_right_rate_teacher': 0.18363939899833054, 'eval/latency_8/demon_move_f1': 0.7437576789404244, 'eval/latency_8/demon_cc_f1': 0.8823040056929743, 'timing/data': 0.010916575789451599, 'timing/model': 1.2522329641506076, 'learning_rate/qwen_vl_interface': 6.216416062000012e-06, 'learning_rate/action_model': 6.216416062000013e-05, 'epoch': 1.41}) +[2026-06-17 21:39:14,885][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_3100_state +[2026-06-17 21:39:14,890][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 21:39:21,416][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_3100_state/model.safetensors +[2026-06-17 21:39:37,450][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_3100_state/optimizer.bin +[2026-06-17 21:39:37,451][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_3100_state/sampler.bin +[2026-06-17 21:39:37,451][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_3100_state/sampler_1.bin +[2026-06-17 21:39:37,455][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_3100_state/random_states_0.pkl +[2026-06-17 21:39:37,458][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3100 +[2026-06-17 21:39:37,458][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 21:39:37,469][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 21:39:41,416][starVLA.training.train_starvla][INFO] - Step 3101, Loss: {'train/loss': 0.4579044282436371, 'train/grad_norm_pre_clip': 25.17777442932129, 'timing/data': 0.009052402339875698, 'timing/model': 0.9402377353981137, 'learning_rate/qwen_vl_interface': 6.214298505955615e-06, 'learning_rate/action_model': 6.214298505955614e-05, 'epoch': 1.41}) +[2026-06-17 21:39:45,204][starVLA.training.train_starvla][INFO] - Step 3102, Loss: {'train/loss': 0.24102452397346497, 'train/grad_norm_pre_clip': 15.473306655883789, 'timing/data': 0.012674241326749325, 'timing/model': 0.938123319298029, 'learning_rate/qwen_vl_interface': 6.2121807500111e-06, 'learning_rate/action_model': 6.212180750011099e-05, 'epoch': 1.41}) +[2026-06-17 21:39:49,104][starVLA.training.train_starvla][INFO] - Step 3103, Loss: {'train/loss': 0.42526718974113464, 'train/grad_norm_pre_clip': 19.675634384155273, 'timing/data': 0.01139095425605774, 'timing/model': 1.207329099997878, 'learning_rate/qwen_vl_interface': 6.210062794605482e-06, 'learning_rate/action_model': 6.21006279460548e-05, 'epoch': 1.41}) +[2026-06-17 21:39:52,803][starVLA.training.train_starvla][INFO] - Step 3104, Loss: {'train/loss': 0.2632846236228943, 'train/grad_norm_pre_clip': 13.122282028198242, 'timing/data': 0.007651816122233868, 'timing/model': 0.9409572724252939, 'learning_rate/qwen_vl_interface': 6.207944640177812e-06, 'learning_rate/action_model': 6.207944640177812e-05, 'epoch': 1.41}) +[2026-06-17 21:39:56,845][starVLA.training.train_starvla][INFO] - Step 3105, Loss: {'train/loss': 0.5796000957489014, 'train/grad_norm_pre_clip': 20.000316619873047, 'timing/data': 0.012688584625720978, 'timing/model': 0.9459750996902585, 'learning_rate/qwen_vl_interface': 6.205826287167188e-06, 'learning_rate/action_model': 6.205826287167188e-05, 'epoch': 1.41}) +[2026-06-17 21:40:00,846][starVLA.training.train_starvla][INFO] - Step 3106, Loss: {'train/loss': 0.3474184572696686, 'train/grad_norm_pre_clip': 16.026453018188477, 'timing/data': 0.0101618068292737, 'timing/model': 1.0170796578750014, 'learning_rate/qwen_vl_interface': 6.203707736012745e-06, 'learning_rate/action_model': 6.203707736012745e-05, 'epoch': 1.41}) +[2026-06-17 21:40:06,720][starVLA.training.train_starvla][INFO] - Step 3107, Loss: {'train/loss': 0.3878505229949951, 'train/grad_norm_pre_clip': 19.072900772094727, 'timing/data': 0.07362415175884962, 'timing/model': 2.8232663720846176, 'learning_rate/qwen_vl_interface': 6.201588987153665e-06, 'learning_rate/action_model': 6.201588987153664e-05, 'epoch': 1.41}) +[2026-06-17 21:40:11,526][starVLA.training.train_starvla][INFO] - Step 3108, Loss: {'train/loss': 0.24738644063472748, 'train/grad_norm_pre_clip': 13.525249481201172, 'timing/data': 0.0093776760622859, 'timing/model': 0.9675921238958836, 'learning_rate/qwen_vl_interface': 6.19947004102916e-06, 'learning_rate/action_model': 6.199470041029159e-05, 'epoch': 1.41}) +[2026-06-17 21:40:17,134][starVLA.training.train_starvla][INFO] - Step 3109, Loss: {'train/loss': 0.5080111026763916, 'train/grad_norm_pre_clip': 22.174184799194336, 'timing/data': 0.011016465723514557, 'timing/model': 1.2045659814029932, 'learning_rate/qwen_vl_interface': 6.197350898078496e-06, 'learning_rate/action_model': 6.197350898078495e-05, 'epoch': 1.41}) +[2026-06-17 21:40:20,982][starVLA.training.train_starvla][INFO] - Step 3110, Loss: {'train/loss': 0.29575955867767334, 'train/grad_norm_pre_clip': 15.583256721496582, 'timing/data': 0.012368714436888695, 'timing/model': 1.1286398125812411, 'learning_rate/qwen_vl_interface': 6.195231558740972e-06, 'learning_rate/action_model': 6.195231558740971e-05, 'epoch': 1.41}) +[2026-06-17 21:40:24,789][starVLA.training.train_starvla][INFO] - Step 3111, Loss: {'train/loss': 0.3064596652984619, 'train/grad_norm_pre_clip': 59.299110412597656, 'timing/data': 0.00912014301866293, 'timing/model': 0.9258068352937698, 'learning_rate/qwen_vl_interface': 6.1931120234559275e-06, 'learning_rate/action_model': 6.193112023455927e-05, 'epoch': 1.41}) +[2026-06-17 21:40:28,648][starVLA.training.train_starvla][INFO] - Step 3112, Loss: {'train/loss': 0.38339295983314514, 'train/grad_norm_pre_clip': 16.08497428894043, 'timing/data': 0.00789723452180624, 'timing/model': 0.9408561326563358, 'learning_rate/qwen_vl_interface': 6.190992292662747e-06, 'learning_rate/action_model': 6.190992292662746e-05, 'epoch': 1.41}) +[2026-06-17 21:40:32,542][starVLA.training.train_starvla][INFO] - Step 3113, Loss: {'train/loss': 0.41617435216903687, 'train/grad_norm_pre_clip': 18.375324249267578, 'timing/data': 0.009022030048072338, 'timing/model': 0.9336078343912959, 'learning_rate/qwen_vl_interface': 6.18887236680085e-06, 'learning_rate/action_model': 6.18887236680085e-05, 'epoch': 1.41}) +[2026-06-17 21:40:36,799][starVLA.training.train_starvla][INFO] - Step 3114, Loss: {'train/loss': 0.48124533891677856, 'train/grad_norm_pre_clip': 19.941566467285156, 'timing/data': 0.008982416242361069, 'timing/model': 1.362092082388699, 'learning_rate/qwen_vl_interface': 6.186752246309702e-06, 'learning_rate/action_model': 6.186752246309702e-05, 'epoch': 1.41}) +[2026-06-17 21:40:40,629][starVLA.training.train_starvla][INFO] - Step 3115, Loss: {'train/loss': 0.3842860758304596, 'train/grad_norm_pre_clip': 16.309001922607422, 'timing/data': 0.012540846131742, 'timing/model': 1.1150618763640523, 'learning_rate/qwen_vl_interface': 6.184631931628806e-06, 'learning_rate/action_model': 6.184631931628806e-05, 'epoch': 1.41}) +[2026-06-17 21:40:44,576][starVLA.training.train_starvla][INFO] - Step 3116, Loss: {'train/loss': 0.2830714285373688, 'train/grad_norm_pre_clip': 22.585094451904297, 'timing/data': 0.009080247022211552, 'timing/model': 1.0485313441604376, 'learning_rate/qwen_vl_interface': 6.1825114231977036e-06, 'learning_rate/action_model': 6.182511423197704e-05, 'epoch': 1.41}) +[2026-06-17 21:40:48,219][starVLA.training.train_starvla][INFO] - Step 3117, Loss: {'train/loss': 0.2812742590904236, 'train/grad_norm_pre_clip': 12.675564765930176, 'timing/data': 0.010547335259616375, 'timing/model': 0.9382983585819602, 'learning_rate/qwen_vl_interface': 6.18039072145598e-06, 'learning_rate/action_model': 6.18039072145598e-05, 'epoch': 1.41}) +[2026-06-17 21:40:51,971][starVLA.training.train_starvla][INFO] - Step 3118, Loss: {'train/loss': 0.29106244444847107, 'train/grad_norm_pre_clip': 13.396571159362793, 'timing/data': 0.014129942283034325, 'timing/model': 0.9344152370467782, 'learning_rate/qwen_vl_interface': 6.17826982684326e-06, 'learning_rate/action_model': 6.178269826843259e-05, 'epoch': 1.41}) +[2026-06-17 21:40:55,841][starVLA.training.train_starvla][INFO] - Step 3119, Loss: {'train/loss': 0.49321556091308594, 'train/grad_norm_pre_clip': 21.674522399902344, 'timing/data': 0.010843313299119473, 'timing/model': 0.9342451095581055, 'learning_rate/qwen_vl_interface': 6.176148739799205e-06, 'learning_rate/action_model': 6.176148739799205e-05, 'epoch': 1.42}) +[2026-06-17 21:40:59,650][starVLA.training.train_starvla][INFO] - Step 3120, Loss: {'train/loss': 0.1227603554725647, 'train/grad_norm_pre_clip': 8.902252197265625, 'timing/data': 0.010498138144612312, 'timing/model': 0.9286591801792383, 'learning_rate/qwen_vl_interface': 6.1740274607635195e-06, 'learning_rate/action_model': 6.174027460763519e-05, 'epoch': 1.42}) +[2026-06-17 21:41:03,436][starVLA.training.train_starvla][INFO] - Step 3121, Loss: {'train/loss': 0.3832922875881195, 'train/grad_norm_pre_clip': 22.386552810668945, 'timing/data': 0.008001655340194702, 'timing/model': 0.9417325500398874, 'learning_rate/qwen_vl_interface': 6.1719059901759474e-06, 'learning_rate/action_model': 6.171905990175947e-05, 'epoch': 1.42}) +[2026-06-17 21:41:07,295][starVLA.training.train_starvla][INFO] - Step 3122, Loss: {'train/loss': 0.3396921157836914, 'train/grad_norm_pre_clip': 14.781649589538574, 'timing/data': 0.009471459314227104, 'timing/model': 1.176832684315741, 'learning_rate/qwen_vl_interface': 6.169784328476272e-06, 'learning_rate/action_model': 6.169784328476271e-05, 'epoch': 1.42}) +[2026-06-17 21:41:10,946][starVLA.training.train_starvla][INFO] - Step 3123, Loss: {'train/loss': 0.29969772696495056, 'train/grad_norm_pre_clip': 15.937431335449219, 'timing/data': 0.012263675220310688, 'timing/model': 0.9401820059865713, 'learning_rate/qwen_vl_interface': 6.167662476104314e-06, 'learning_rate/action_model': 6.167662476104314e-05, 'epoch': 1.42}) +[2026-06-17 21:41:14,843][starVLA.training.train_starvla][INFO] - Step 3124, Loss: {'train/loss': 0.3876382112503052, 'train/grad_norm_pre_clip': 22.2012882232666, 'timing/data': 0.011045399121940136, 'timing/model': 0.9523824788630009, 'learning_rate/qwen_vl_interface': 6.165540433499939e-06, 'learning_rate/action_model': 6.165540433499939e-05, 'epoch': 1.42}) +[2026-06-17 21:41:18,809][starVLA.training.train_starvla][INFO] - Step 3125, Loss: {'train/loss': 0.28753599524497986, 'train/grad_norm_pre_clip': 50.61397171020508, 'timing/data': 0.006950081326067448, 'timing/model': 0.933531554415822, 'learning_rate/qwen_vl_interface': 6.163418201103045e-06, 'learning_rate/action_model': 6.163418201103045e-05, 'epoch': 1.42}) +[2026-06-17 21:41:22,566][starVLA.training.train_starvla][INFO] - Step 3126, Loss: {'train/loss': 0.5176896452903748, 'train/grad_norm_pre_clip': 19.745729446411133, 'timing/data': 0.008801106363534927, 'timing/model': 0.9378777351230383, 'learning_rate/qwen_vl_interface': 6.161295779353578e-06, 'learning_rate/action_model': 6.161295779353577e-05, 'epoch': 1.42}) +[2026-06-17 21:41:26,323][starVLA.training.train_starvla][INFO] - Step 3127, Loss: {'train/loss': 0.450219988822937, 'train/grad_norm_pre_clip': 22.294204711914062, 'timing/data': 0.011416633613407612, 'timing/model': 0.9287598067894578, 'learning_rate/qwen_vl_interface': 6.159173168691512e-06, 'learning_rate/action_model': 6.159173168691512e-05, 'epoch': 1.42}) +[2026-06-17 21:41:30,107][starVLA.training.train_starvla][INFO] - Step 3128, Loss: {'train/loss': 0.549162745475769, 'train/grad_norm_pre_clip': 23.23665428161621, 'timing/data': 0.00804520957171917, 'timing/model': 1.1215612208470702, 'learning_rate/qwen_vl_interface': 6.157050369556871e-06, 'learning_rate/action_model': 6.157050369556872e-05, 'epoch': 1.42}) +[2026-06-17 21:41:33,898][starVLA.training.train_starvla][INFO] - Step 3129, Loss: {'train/loss': 0.7925808429718018, 'train/grad_norm_pre_clip': 24.187665939331055, 'timing/data': 0.009133178740739822, 'timing/model': 1.1052501499652863, 'learning_rate/qwen_vl_interface': 6.154927382389714e-06, 'learning_rate/action_model': 6.154927382389714e-05, 'epoch': 1.42}) +[2026-06-17 21:41:37,619][starVLA.training.train_starvla][INFO] - Step 3130, Loss: {'train/loss': 0.3825174570083618, 'train/grad_norm_pre_clip': 18.473819732666016, 'timing/data': 0.00851458590477705, 'timing/model': 0.9639742756262422, 'learning_rate/qwen_vl_interface': 6.152804207630135e-06, 'learning_rate/action_model': 6.152804207630136e-05, 'epoch': 1.42}) +[2026-06-17 21:41:41,469][starVLA.training.train_starvla][INFO] - Step 3131, Loss: {'train/loss': 0.3391905426979065, 'train/grad_norm_pre_clip': 17.737512588500977, 'timing/data': 0.01695914287120104, 'timing/model': 0.9429444139823318, 'learning_rate/qwen_vl_interface': 6.150680845718274e-06, 'learning_rate/action_model': 6.150680845718274e-05, 'epoch': 1.42}) +[2026-06-17 21:41:45,334][starVLA.training.train_starvla][INFO] - Step 3132, Loss: {'train/loss': 0.20578224956989288, 'train/grad_norm_pre_clip': 10.295002937316895, 'timing/data': 0.01071932166814804, 'timing/model': 0.9418987892568111, 'learning_rate/qwen_vl_interface': 6.148557297094303e-06, 'learning_rate/action_model': 6.148557297094303e-05, 'epoch': 1.42}) +[2026-06-17 21:41:49,309][starVLA.training.train_starvla][INFO] - Step 3133, Loss: {'train/loss': 0.4897690415382385, 'train/grad_norm_pre_clip': 17.634511947631836, 'timing/data': 0.008719704113900661, 'timing/model': 1.0292397532612085, 'learning_rate/qwen_vl_interface': 6.1464335621984375e-06, 'learning_rate/action_model': 6.146433562198437e-05, 'epoch': 1.42}) +[2026-06-17 21:41:53,215][starVLA.training.train_starvla][INFO] - Step 3134, Loss: {'train/loss': 0.3006228506565094, 'train/grad_norm_pre_clip': 16.224088668823242, 'timing/data': 0.011651157401502132, 'timing/model': 1.023496893234551, 'learning_rate/qwen_vl_interface': 6.14430964147093e-06, 'learning_rate/action_model': 6.14430964147093e-05, 'epoch': 1.42}) +[2026-06-17 21:41:57,276][starVLA.training.train_starvla][INFO] - Step 3135, Loss: {'train/loss': 0.24532555043697357, 'train/grad_norm_pre_clip': 15.243430137634277, 'timing/data': 0.007100026123225689, 'timing/model': 1.1305170459672809, 'learning_rate/qwen_vl_interface': 6.1421855353520735e-06, 'learning_rate/action_model': 6.142185535352073e-05, 'epoch': 1.42}) +[2026-06-17 21:42:01,085][starVLA.training.train_starvla][INFO] - Step 3136, Loss: {'train/loss': 0.4663670063018799, 'train/grad_norm_pre_clip': 15.729592323303223, 'timing/data': 0.011565526016056538, 'timing/model': 1.099866557866335, 'learning_rate/qwen_vl_interface': 6.140061244282194e-06, 'learning_rate/action_model': 6.140061244282194e-05, 'epoch': 1.42}) +[2026-06-17 21:42:06,316][starVLA.training.train_starvla][INFO] - Step 3137, Loss: {'train/loss': 0.4882015287876129, 'train/grad_norm_pre_clip': 24.399877548217773, 'timing/data': 0.008241118863224983, 'timing/model': 1.0609053736552596, 'learning_rate/qwen_vl_interface': 6.137936768701659e-06, 'learning_rate/action_model': 6.137936768701659e-05, 'epoch': 1.42}) +[2026-06-17 21:42:10,241][starVLA.training.train_starvla][INFO] - Step 3138, Loss: {'train/loss': 0.23429927229881287, 'train/grad_norm_pre_clip': 12.050826072692871, 'timing/data': 0.00796614121645689, 'timing/model': 0.9422190273180604, 'learning_rate/qwen_vl_interface': 6.135812109050879e-06, 'learning_rate/action_model': 6.135812109050879e-05, 'epoch': 1.42}) +[2026-06-17 21:42:14,153][starVLA.training.train_starvla][INFO] - Step 3139, Loss: {'train/loss': 0.19271741807460785, 'train/grad_norm_pre_clip': 12.016884803771973, 'timing/data': 0.009627674706280231, 'timing/model': 0.9418468503281474, 'learning_rate/qwen_vl_interface': 6.1336872657702936e-06, 'learning_rate/action_model': 6.133687265770293e-05, 'epoch': 1.42}) +[2026-06-17 21:42:18,031][starVLA.training.train_starvla][INFO] - Step 3140, Loss: {'train/loss': 0.2671422064304352, 'train/grad_norm_pre_clip': 17.297876358032227, 'timing/data': 0.021106346510350704, 'timing/model': 0.939909509383142, 'learning_rate/qwen_vl_interface': 6.131562239300389e-06, 'learning_rate/action_model': 6.13156223930039e-05, 'epoch': 1.42}) +[2026-06-17 21:42:21,832][starVLA.training.train_starvla][INFO] - Step 3141, Loss: {'train/loss': 0.608216404914856, 'train/grad_norm_pre_clip': 69.4746322631836, 'timing/data': 0.011670966632664204, 'timing/model': 0.9294080091640353, 'learning_rate/qwen_vl_interface': 6.129437030081683e-06, 'learning_rate/action_model': 6.129437030081682e-05, 'epoch': 1.42}) +[2026-06-17 21:42:25,661][starVLA.training.train_starvla][INFO] - Step 3142, Loss: {'train/loss': 0.3301660120487213, 'train/grad_norm_pre_clip': 19.334896087646484, 'timing/data': 0.012767629697918892, 'timing/model': 1.1213656468316913, 'learning_rate/qwen_vl_interface': 6.1273116385547336e-06, 'learning_rate/action_model': 6.127311638554734e-05, 'epoch': 1.43}) +[2026-06-17 21:42:29,330][starVLA.training.train_starvla][INFO] - Step 3143, Loss: {'train/loss': 0.420584112405777, 'train/grad_norm_pre_clip': 23.73443603515625, 'timing/data': 0.009415414184331894, 'timing/model': 0.9433919284492731, 'learning_rate/qwen_vl_interface': 6.125186065160138e-06, 'learning_rate/action_model': 6.125186065160138e-05, 'epoch': 1.43}) +[2026-06-17 21:42:33,231][starVLA.training.train_starvla][INFO] - Step 3144, Loss: {'train/loss': 0.17300942540168762, 'train/grad_norm_pre_clip': 11.753555297851562, 'timing/data': 0.011228921823203564, 'timing/model': 0.9370448747649789, 'learning_rate/qwen_vl_interface': 6.123060310338529e-06, 'learning_rate/action_model': 6.123060310338529e-05, 'epoch': 1.43}) +[2026-06-17 21:42:37,029][starVLA.training.train_starvla][INFO] - Step 3145, Loss: {'train/loss': 0.49280059337615967, 'train/grad_norm_pre_clip': 16.376256942749023, 'timing/data': 0.012073440477252007, 'timing/model': 0.9307160368189216, 'learning_rate/qwen_vl_interface': 6.120934374530579e-06, 'learning_rate/action_model': 6.12093437453058e-05, 'epoch': 1.43}) +[2026-06-17 21:42:40,993][starVLA.training.train_starvla][INFO] - Step 3146, Loss: {'train/loss': 0.4014548659324646, 'train/grad_norm_pre_clip': 15.72849178314209, 'timing/data': 0.006904100999236107, 'timing/model': 1.03186387848109, 'learning_rate/qwen_vl_interface': 6.118808258176996e-06, 'learning_rate/action_model': 6.118808258176995e-05, 'epoch': 1.43}) +[2026-06-17 21:42:44,839][starVLA.training.train_starvla][INFO] - Step 3147, Loss: {'train/loss': 0.272378146648407, 'train/grad_norm_pre_clip': 19.689050674438477, 'timing/data': 0.010408448986709118, 'timing/model': 0.9377338048070669, 'learning_rate/qwen_vl_interface': 6.116681961718525e-06, 'learning_rate/action_model': 6.116681961718524e-05, 'epoch': 1.43}) +[2026-06-17 21:42:49,035][starVLA.training.train_starvla][INFO] - Step 3148, Loss: {'train/loss': 0.43211716413497925, 'train/grad_norm_pre_clip': 21.769058227539062, 'timing/data': 0.00955782551318407, 'timing/model': 1.154012924991548, 'learning_rate/qwen_vl_interface': 6.114555485595952e-06, 'learning_rate/action_model': 6.114555485595952e-05, 'epoch': 1.43}) +[2026-06-17 21:42:52,947][starVLA.training.train_starvla][INFO] - Step 3149, Loss: {'train/loss': 0.46883583068847656, 'train/grad_norm_pre_clip': 27.264150619506836, 'timing/data': 0.01118918415158987, 'timing/model': 1.1515909805893898, 'learning_rate/qwen_vl_interface': 6.112428830250096e-06, 'learning_rate/action_model': 6.112428830250095e-05, 'epoch': 1.43}) +[2026-06-17 21:42:56,655][starVLA.training.train_starvla][INFO] - Step 3150, Loss: {'train/loss': 0.3899388015270233, 'train/grad_norm_pre_clip': 16.83568000793457, 'timing/data': 0.01100164745002985, 'timing/model': 0.9357256628572941, 'learning_rate/qwen_vl_interface': 6.110301996121816e-06, 'learning_rate/action_model': 6.110301996121816e-05, 'epoch': 1.43}) +[2026-06-17 21:43:00,662][starVLA.training.train_starvla][INFO] - Step 3151, Loss: {'train/loss': 0.4540535509586334, 'train/grad_norm_pre_clip': 20.893875122070312, 'timing/data': 0.009481635875999928, 'timing/model': 0.9825760591775179, 'learning_rate/qwen_vl_interface': 6.108174983652007e-06, 'learning_rate/action_model': 6.108174983652006e-05, 'epoch': 1.43}) +[2026-06-17 21:43:04,701][starVLA.training.train_starvla][INFO] - Step 3152, Loss: {'train/loss': 0.17580223083496094, 'train/grad_norm_pre_clip': 13.923306465148926, 'timing/data': 0.01196521520614624, 'timing/model': 0.9506612569093704, 'learning_rate/qwen_vl_interface': 6.106047793281598e-06, 'learning_rate/action_model': 6.106047793281598e-05, 'epoch': 1.43}) +[2026-06-17 21:43:08,667][starVLA.training.train_starvla][INFO] - Step 3153, Loss: {'train/loss': 0.5101135969161987, 'train/grad_norm_pre_clip': 16.938936233520508, 'timing/data': 0.009359369985759258, 'timing/model': 0.9363875761628151, 'learning_rate/qwen_vl_interface': 6.1039204254515626e-06, 'learning_rate/action_model': 6.103920425451563e-05, 'epoch': 1.43}) +[2026-06-17 21:43:12,598][starVLA.training.train_starvla][INFO] - Step 3154, Loss: {'train/loss': 0.210958331823349, 'train/grad_norm_pre_clip': 16.521879196166992, 'timing/data': 0.009809750132262707, 'timing/model': 0.9597530281171203, 'learning_rate/qwen_vl_interface': 6.101792880602901e-06, 'learning_rate/action_model': 6.101792880602901e-05, 'epoch': 1.43}) +[2026-06-17 21:43:16,611][starVLA.training.train_starvla][INFO] - Step 3155, Loss: {'train/loss': 0.19470590353012085, 'train/grad_norm_pre_clip': 11.800724029541016, 'timing/data': 0.009429228492081165, 'timing/model': 1.215766198001802, 'learning_rate/qwen_vl_interface': 6.099665159176663e-06, 'learning_rate/action_model': 6.099665159176663e-05, 'epoch': 1.43}) +[2026-06-17 21:43:20,412][starVLA.training.train_starvla][INFO] - Step 3156, Loss: {'train/loss': 0.5066767930984497, 'train/grad_norm_pre_clip': 21.382532119750977, 'timing/data': 0.01032085157930851, 'timing/model': 1.1162445228546858, 'learning_rate/qwen_vl_interface': 6.09753726161392e-06, 'learning_rate/action_model': 6.09753726161392e-05, 'epoch': 1.43}) +[2026-06-17 21:43:24,118][starVLA.training.train_starvla][INFO] - Step 3157, Loss: {'train/loss': 0.445497065782547, 'train/grad_norm_pre_clip': 19.303054809570312, 'timing/data': 0.009552612900733948, 'timing/model': 0.9603914078325033, 'learning_rate/qwen_vl_interface': 6.095409188355793e-06, 'learning_rate/action_model': 6.095409188355793e-05, 'epoch': 1.43}) +[2026-06-17 21:43:27,932][starVLA.training.train_starvla][INFO] - Step 3158, Loss: {'train/loss': 0.20924566686153412, 'train/grad_norm_pre_clip': 12.05412769317627, 'timing/data': 0.008874074555933475, 'timing/model': 0.9379914049059153, 'learning_rate/qwen_vl_interface': 6.0932809398434315e-06, 'learning_rate/action_model': 6.093280939843431e-05, 'epoch': 1.43}) +[2026-06-17 21:43:31,731][starVLA.training.train_starvla][INFO] - Step 3159, Loss: {'train/loss': 0.2610381841659546, 'train/grad_norm_pre_clip': 16.994495391845703, 'timing/data': 0.009982986375689507, 'timing/model': 0.9342185901477933, 'learning_rate/qwen_vl_interface': 6.091152516518022e-06, 'learning_rate/action_model': 6.0911525165180216e-05, 'epoch': 1.43}) +[2026-06-17 21:43:35,604][starVLA.training.train_starvla][INFO] - Step 3160, Loss: {'train/loss': 0.20648126304149628, 'train/grad_norm_pre_clip': 11.467508316040039, 'timing/data': 0.009054253809154034, 'timing/model': 0.9483723491430283, 'learning_rate/qwen_vl_interface': 6.089023918820792e-06, 'learning_rate/action_model': 6.089023918820792e-05, 'epoch': 1.43}) +[2026-06-17 21:43:39,409][starVLA.training.train_starvla][INFO] - Step 3161, Loss: {'train/loss': 0.6012300848960876, 'train/grad_norm_pre_clip': 19.137985229492188, 'timing/data': 0.007879160344600677, 'timing/model': 0.9569008098915219, 'learning_rate/qwen_vl_interface': 6.0868951471929986e-06, 'learning_rate/action_model': 6.086895147192998e-05, 'epoch': 1.43}) +[2026-06-17 21:43:43,305][starVLA.training.train_starvla][INFO] - Step 3162, Loss: {'train/loss': 0.25384870171546936, 'train/grad_norm_pre_clip': 17.204139709472656, 'timing/data': 0.007127782329916954, 'timing/model': 1.1652765106409788, 'learning_rate/qwen_vl_interface': 6.084766202075944e-06, 'learning_rate/action_model': 6.0847662020759435e-05, 'epoch': 1.43}) +[2026-06-17 21:43:46,934][starVLA.training.train_starvla][INFO] - Step 3163, Loss: {'train/loss': 0.19371941685676575, 'train/grad_norm_pre_clip': 12.586136817932129, 'timing/data': 0.0077045150101184845, 'timing/model': 0.9352129874750972, 'learning_rate/qwen_vl_interface': 6.082637083910955e-06, 'learning_rate/action_model': 6.0826370839109546e-05, 'epoch': 1.43}) +[2026-06-17 21:43:50,742][starVLA.training.train_starvla][INFO] - Step 3164, Loss: {'train/loss': 0.5663213729858398, 'train/grad_norm_pre_clip': 23.674976348876953, 'timing/data': 0.014830839820206165, 'timing/model': 0.9414869593456388, 'learning_rate/qwen_vl_interface': 6.080507793139403e-06, 'learning_rate/action_model': 6.080507793139403e-05, 'epoch': 1.44}) +[2026-06-17 21:43:54,532][starVLA.training.train_starvla][INFO] - Step 3165, Loss: {'train/loss': 0.31426531076431274, 'train/grad_norm_pre_clip': 14.030631065368652, 'timing/data': 0.008181956596672535, 'timing/model': 0.9390632789582014, 'learning_rate/qwen_vl_interface': 6.0783783302026925e-06, 'learning_rate/action_model': 6.078378330202692e-05, 'epoch': 1.44}) +[2026-06-17 21:43:58,433][starVLA.training.train_starvla][INFO] - Step 3166, Loss: {'train/loss': 0.29939016699790955, 'train/grad_norm_pre_clip': 15.459346771240234, 'timing/data': 0.009366637095808983, 'timing/model': 0.9400618365034461, 'learning_rate/qwen_vl_interface': 6.076248695542262e-06, 'learning_rate/action_model': 6.076248695542261e-05, 'epoch': 1.44}) +[2026-06-17 21:44:02,328][starVLA.training.train_starvla][INFO] - Step 3167, Loss: {'train/loss': 0.4340657591819763, 'train/grad_norm_pre_clip': 21.858861923217773, 'timing/data': 0.009019349701702595, 'timing/model': 0.9318241998553276, 'learning_rate/qwen_vl_interface': 6.074118889599589e-06, 'learning_rate/action_model': 6.0741188895995884e-05, 'epoch': 1.44}) +[2026-06-17 21:44:06,202][starVLA.training.train_starvla][INFO] - Step 3168, Loss: {'train/loss': 0.06645001471042633, 'train/grad_norm_pre_clip': 7.125612258911133, 'timing/data': 0.007478510960936546, 'timing/model': 0.951330091804266, 'learning_rate/qwen_vl_interface': 6.07198891281618e-06, 'learning_rate/action_model': 6.07198891281618e-05, 'epoch': 1.44}) +[2026-06-17 21:44:10,114][starVLA.training.train_starvla][INFO] - Step 3169, Loss: {'train/loss': 0.6713588833808899, 'train/grad_norm_pre_clip': 27.953039169311523, 'timing/data': 0.014097628183662891, 'timing/model': 1.1524034002795815, 'learning_rate/qwen_vl_interface': 6.0698587656335895e-06, 'learning_rate/action_model': 6.069858765633589e-05, 'epoch': 1.44}) +[2026-06-17 21:44:13,785][starVLA.training.train_starvla][INFO] - Step 3170, Loss: {'train/loss': 0.49388211965560913, 'train/grad_norm_pre_clip': 26.64495086669922, 'timing/data': 0.00978285912424326, 'timing/model': 0.9353599734604359, 'learning_rate/qwen_vl_interface': 6.067728448493392e-06, 'learning_rate/action_model': 6.067728448493392e-05, 'epoch': 1.44}) +[2026-06-17 21:44:19,423][starVLA.training.train_starvla][INFO] - Step 3171, Loss: {'train/loss': 0.34438446164131165, 'train/grad_norm_pre_clip': 20.39401626586914, 'timing/data': 0.01129993051290512, 'timing/model': 0.9563726875931025, 'learning_rate/qwen_vl_interface': 6.065597961837209e-06, 'learning_rate/action_model': 6.0655979618372086e-05, 'epoch': 1.44}) +[2026-06-17 21:44:23,443][starVLA.training.train_starvla][INFO] - Step 3172, Loss: {'train/loss': 0.5590048432350159, 'train/grad_norm_pre_clip': 26.288305282592773, 'timing/data': 0.00726429745554924, 'timing/model': 1.012311402708292, 'learning_rate/qwen_vl_interface': 6.063467306106688e-06, 'learning_rate/action_model': 6.0634673061066885e-05, 'epoch': 1.44}) +[2026-06-17 21:44:28,490][starVLA.training.train_starvla][INFO] - Step 3173, Loss: {'train/loss': 0.6411734223365784, 'train/grad_norm_pre_clip': 29.499483108520508, 'timing/data': 0.012054098770022392, 'timing/model': 0.9599798107519746, 'learning_rate/qwen_vl_interface': 6.061336481743522e-06, 'learning_rate/action_model': 6.061336481743522e-05, 'epoch': 1.44}) +[2026-06-17 21:44:32,428][starVLA.training.train_starvla][INFO] - Step 3174, Loss: {'train/loss': 0.2829018831253052, 'train/grad_norm_pre_clip': 13.620881080627441, 'timing/data': 0.016232753172516823, 'timing/model': 0.9308605007827282, 'learning_rate/qwen_vl_interface': 6.059205489189429e-06, 'learning_rate/action_model': 6.059205489189429e-05, 'epoch': 1.44}) +[2026-06-17 21:44:36,264][starVLA.training.train_starvla][INFO] - Step 3175, Loss: {'train/loss': 0.5562028288841248, 'train/grad_norm_pre_clip': 22.852954864501953, 'timing/data': 0.010917018167674541, 'timing/model': 1.1417378447949886, 'learning_rate/qwen_vl_interface': 6.057074328886168e-06, 'learning_rate/action_model': 6.057074328886168e-05, 'epoch': 1.44}) +[2026-06-17 21:44:40,097][starVLA.training.train_starvla][INFO] - Step 3176, Loss: {'train/loss': 0.7343141436576843, 'train/grad_norm_pre_clip': 23.593475341796875, 'timing/data': 0.008784748613834381, 'timing/model': 1.1081471359357238, 'learning_rate/qwen_vl_interface': 6.0549430012755295e-06, 'learning_rate/action_model': 6.0549430012755295e-05, 'epoch': 1.44}) +[2026-06-17 21:44:43,732][starVLA.training.train_starvla][INFO] - Step 3177, Loss: {'train/loss': 0.13220728933811188, 'train/grad_norm_pre_clip': 10.571155548095703, 'timing/data': 0.00996348075568676, 'timing/model': 0.9422278134152293, 'learning_rate/qwen_vl_interface': 6.052811506799341e-06, 'learning_rate/action_model': 6.052811506799341e-05, 'epoch': 1.44}) +[2026-06-17 21:44:47,596][starVLA.training.train_starvla][INFO] - Step 3178, Loss: {'train/loss': 0.4263728857040405, 'train/grad_norm_pre_clip': 18.77811050415039, 'timing/data': 0.010412493720650673, 'timing/model': 0.9342139158397913, 'learning_rate/qwen_vl_interface': 6.050679845899462e-06, 'learning_rate/action_model': 6.050679845899462e-05, 'epoch': 1.44}) +[2026-06-17 21:44:51,502][starVLA.training.train_starvla][INFO] - Step 3179, Loss: {'train/loss': 0.4691894054412842, 'train/grad_norm_pre_clip': 19.654315948486328, 'timing/data': 0.014590460807085037, 'timing/model': 1.023959282785654, 'learning_rate/qwen_vl_interface': 6.048548019017791e-06, 'learning_rate/action_model': 6.04854801901779e-05, 'epoch': 1.44}) +[2026-06-17 21:44:55,329][starVLA.training.train_starvla][INFO] - Step 3180, Loss: {'train/loss': 0.5896975994110107, 'train/grad_norm_pre_clip': 27.157489776611328, 'timing/data': 0.007739582099020481, 'timing/model': 0.935367570258677, 'learning_rate/qwen_vl_interface': 6.046416026596256e-06, 'learning_rate/action_model': 6.046416026596255e-05, 'epoch': 1.44}) +[2026-06-17 21:44:59,181][starVLA.training.train_starvla][INFO] - Step 3181, Loss: {'train/loss': 0.5522249341011047, 'train/grad_norm_pre_clip': 22.634841918945312, 'timing/data': 0.00987847801297903, 'timing/model': 0.9474848881363869, 'learning_rate/qwen_vl_interface': 6.044283869076822e-06, 'learning_rate/action_model': 6.0442838690768224e-05, 'epoch': 1.44}) +[2026-06-17 21:45:03,077][starVLA.training.train_starvla][INFO] - Step 3182, Loss: {'train/loss': 0.2930584251880646, 'train/grad_norm_pre_clip': 12.036050796508789, 'timing/data': 0.0106933219358325, 'timing/model': 1.1291968002915382, 'learning_rate/qwen_vl_interface': 6.042151546901487e-06, 'learning_rate/action_model': 6.042151546901487e-05, 'epoch': 1.44}) +[2026-06-17 21:45:06,922][starVLA.training.train_starvla][INFO] - Step 3183, Loss: {'train/loss': 0.1715126484632492, 'train/grad_norm_pre_clip': 11.048850059509277, 'timing/data': 0.009475535713136196, 'timing/model': 1.1271430868655443, 'learning_rate/qwen_vl_interface': 6.0400190605122846e-06, 'learning_rate/action_model': 6.040019060512284e-05, 'epoch': 1.44}) +[2026-06-17 21:45:10,625][starVLA.training.train_starvla][INFO] - Step 3184, Loss: {'train/loss': 0.2752380669116974, 'train/grad_norm_pre_clip': 14.80377197265625, 'timing/data': 0.010281486436724663, 'timing/model': 0.9560523051768541, 'learning_rate/qwen_vl_interface': 6.0378864103512805e-06, 'learning_rate/action_model': 6.03788641035128e-05, 'epoch': 1.44}) +[2026-06-17 21:45:14,431][starVLA.training.train_starvla][INFO] - Step 3185, Loss: {'train/loss': 0.1470303237438202, 'train/grad_norm_pre_clip': 12.800139427185059, 'timing/data': 0.008901398628950119, 'timing/model': 0.9360253596678376, 'learning_rate/qwen_vl_interface': 6.035753596860573e-06, 'learning_rate/action_model': 6.0357535968605725e-05, 'epoch': 1.44}) +[2026-06-17 21:45:18,301][starVLA.training.train_starvla][INFO] - Step 3186, Loss: {'train/loss': 0.3564719259738922, 'train/grad_norm_pre_clip': 24.270723342895508, 'timing/data': 0.008136621676385403, 'timing/model': 0.9421025365591049, 'learning_rate/qwen_vl_interface': 6.033620620482303e-06, 'learning_rate/action_model': 6.033620620482303e-05, 'epoch': 1.45}) +[2026-06-17 21:45:22,047][starVLA.training.train_starvla][INFO] - Step 3187, Loss: {'train/loss': 0.3327881991863251, 'train/grad_norm_pre_clip': 16.6093807220459, 'timing/data': 0.014502021484076977, 'timing/model': 0.9307989273220301, 'learning_rate/qwen_vl_interface': 6.031487481658633e-06, 'learning_rate/action_model': 6.031487481658633e-05, 'epoch': 1.45}) +[2026-06-17 21:45:25,881][starVLA.training.train_starvla][INFO] - Step 3188, Loss: {'train/loss': 0.47498220205307007, 'train/grad_norm_pre_clip': 18.106109619140625, 'timing/data': 0.00895853340625763, 'timing/model': 0.9520098371431231, 'learning_rate/qwen_vl_interface': 6.029354180831769e-06, 'learning_rate/action_model': 6.029354180831769e-05, 'epoch': 1.45}) +[2026-06-17 21:45:29,708][starVLA.training.train_starvla][INFO] - Step 3189, Loss: {'train/loss': 0.3710057735443115, 'train/grad_norm_pre_clip': 17.962703704833984, 'timing/data': 0.009502124041318893, 'timing/model': 1.1229342222213745, 'learning_rate/qwen_vl_interface': 6.027220718443943e-06, 'learning_rate/action_model': 6.027220718443943e-05, 'epoch': 1.45}) +[2026-06-17 21:45:33,332][starVLA.training.train_starvla][INFO] - Step 3190, Loss: {'train/loss': 0.25714319944381714, 'train/grad_norm_pre_clip': 16.35117530822754, 'timing/data': 0.008704199455678463, 'timing/model': 0.9378167139366269, 'learning_rate/qwen_vl_interface': 6.025087094937426e-06, 'learning_rate/action_model': 6.025087094937426e-05, 'epoch': 1.45}) +[2026-06-17 21:45:37,195][starVLA.training.train_starvla][INFO] - Step 3191, Loss: {'train/loss': 0.3617626428604126, 'train/grad_norm_pre_clip': 13.679258346557617, 'timing/data': 0.009040665812790394, 'timing/model': 0.9942351384088397, 'learning_rate/qwen_vl_interface': 6.022953310754521e-06, 'learning_rate/action_model': 6.022953310754521e-05, 'epoch': 1.45}) +[2026-06-17 21:45:40,936][starVLA.training.train_starvla][INFO] - Step 3192, Loss: {'train/loss': 0.48279961943626404, 'train/grad_norm_pre_clip': 22.823436737060547, 'timing/data': 0.011050505563616753, 'timing/model': 0.9257593676447868, 'learning_rate/qwen_vl_interface': 6.020819366337562e-06, 'learning_rate/action_model': 6.020819366337562e-05, 'epoch': 1.45}) +[2026-06-17 21:45:44,941][starVLA.training.train_starvla][INFO] - Step 3193, Loss: {'train/loss': 0.4313045144081116, 'train/grad_norm_pre_clip': 17.160829544067383, 'timing/data': 0.007366843521595001, 'timing/model': 0.989903537556529, 'learning_rate/qwen_vl_interface': 6.018685262128921e-06, 'learning_rate/action_model': 6.01868526212892e-05, 'epoch': 1.45}) +[2026-06-17 21:45:49,361][starVLA.training.train_starvla][INFO] - Step 3194, Loss: {'train/loss': 0.5156069397926331, 'train/grad_norm_pre_clip': 15.448104858398438, 'timing/data': 0.01026058942079544, 'timing/model': 0.9474744098260999, 'learning_rate/qwen_vl_interface': 6.016550998570995e-06, 'learning_rate/action_model': 6.0165509985709944e-05, 'epoch': 1.45}) +[2026-06-17 21:45:53,924][starVLA.training.train_starvla][INFO] - Step 3195, Loss: {'train/loss': 0.10354301333427429, 'train/grad_norm_pre_clip': 7.018373966217041, 'timing/data': 0.013370215892791748, 'timing/model': 1.774369371123612, 'learning_rate/qwen_vl_interface': 6.014416576106224e-06, 'learning_rate/action_model': 6.0144165761062244e-05, 'epoch': 1.45}) +[2026-06-17 21:45:57,813][starVLA.training.train_starvla][INFO] - Step 3196, Loss: {'train/loss': 0.3790827989578247, 'train/grad_norm_pre_clip': 12.946883201599121, 'timing/data': 0.010018724016845226, 'timing/model': 1.1522350842133164, 'learning_rate/qwen_vl_interface': 6.012281995177074e-06, 'learning_rate/action_model': 6.012281995177074e-05, 'epoch': 1.45}) +[2026-06-17 21:46:01,482][starVLA.training.train_starvla][INFO] - Step 3197, Loss: {'train/loss': 0.25811752676963806, 'train/grad_norm_pre_clip': 11.875149726867676, 'timing/data': 0.008952462114393711, 'timing/model': 0.9422412486746907, 'learning_rate/qwen_vl_interface': 6.010147256226046e-06, 'learning_rate/action_model': 6.0101472562260454e-05, 'epoch': 1.45}) +[2026-06-17 21:46:06,636][starVLA.training.train_starvla][INFO] - Step 3198, Loss: {'train/loss': 0.5894408822059631, 'train/grad_norm_pre_clip': 21.499170303344727, 'timing/data': 0.008673966862261295, 'timing/model': 0.9766845051199198, 'learning_rate/qwen_vl_interface': 6.008012359695673e-06, 'learning_rate/action_model': 6.0080123596956726e-05, 'epoch': 1.45}) +[2026-06-17 21:46:10,728][starVLA.training.train_starvla][INFO] - Step 3199, Loss: {'train/loss': 0.43907809257507324, 'train/grad_norm_pre_clip': 18.086040496826172, 'timing/data': 0.013191488571465015, 'timing/model': 0.9936715150251985, 'learning_rate/qwen_vl_interface': 6.005877306028522e-06, 'learning_rate/action_model': 6.005877306028522e-05, 'epoch': 1.45}) +[2026-06-17 21:48:17,097][starVLA.training.train_starvla][INFO] - Step 3200, Loss: {'train/loss': 0.394106388092041, 'train/grad_norm_pre_clip': 15.317242622375488, 'eval/loss': 0.3818633258342743, 'eval/latency_8/loss': 0.3818633258342743, 'eval/demon_fire_precision': 0.9396449704142011, 'eval/demon_fire_recall': 0.9866418142280211, 'eval/demon_fire_f1': 0.9625700863767237, 'eval/demon_fire_rate_model': 0.9404563160823595, 'eval/demon_fire_rate_teacher': 0.8956594323873122, 'eval/demon_left_precision': 0.8186528497409327, 'eval/demon_left_recall': 0.6766595289079229, 'eval/demon_left_f1': 0.7409144196951934, 'eval/demon_left_rate_model': 0.10740122426266, 'eval/demon_left_rate_teacher': 0.12993878686700056, 'eval/demon_right_precision': 0.8380414312617702, 'eval/demon_right_recall': 0.6742424242424242, 'eval/demon_right_f1': 0.7472712006717044, 'eval/demon_right_rate_model': 0.14774624373956594, 'eval/demon_right_rate_teacher': 0.18363939899833054, 'eval/demon_move_f1': 0.744092810183449, 'eval/demon_cc_f1': 0.8846462937992549, 'eval/latency_8/demon_fire_precision': 0.9396449704142011, 'eval/latency_8/demon_fire_recall': 0.9866418142280211, 'eval/latency_8/demon_fire_f1': 0.9625700863767237, 'eval/latency_8/demon_fire_rate_model': 0.9404563160823595, 'eval/latency_8/demon_fire_rate_teacher': 0.8956594323873122, 'eval/latency_8/demon_left_precision': 0.8186528497409327, 'eval/latency_8/demon_left_recall': 0.6766595289079229, 'eval/latency_8/demon_left_f1': 0.7409144196951934, 'eval/latency_8/demon_left_rate_model': 0.10740122426266, 'eval/latency_8/demon_left_rate_teacher': 0.12993878686700056, 'eval/latency_8/demon_right_precision': 0.8380414312617702, 'eval/latency_8/demon_right_recall': 0.6742424242424242, 'eval/latency_8/demon_right_f1': 0.7472712006717044, 'eval/latency_8/demon_right_rate_model': 0.14774624373956594, 'eval/latency_8/demon_right_rate_teacher': 0.18363939899833054, 'eval/latency_8/demon_move_f1': 0.744092810183449, 'eval/latency_8/demon_cc_f1': 0.8846462937992549, 'timing/data': 0.008430032059550285, 'timing/model': 0.9968022042885423, 'learning_rate/qwen_vl_interface': 6.003742095667194e-06, 'learning_rate/action_model': 6.003742095667194e-05, 'epoch': 1.45}) +[2026-06-17 21:48:17,100][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_3200_state +[2026-06-17 21:48:17,105][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 21:48:23,689][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_3200_state/model.safetensors +[2026-06-17 21:48:39,281][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_3200_state/optimizer.bin +[2026-06-17 21:48:39,282][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_3200_state/sampler.bin +[2026-06-17 21:48:39,282][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_3200_state/sampler_1.bin +[2026-06-17 21:48:39,283][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_3200_state/random_states_0.pkl +[2026-06-17 21:48:39,286][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3200 +[2026-06-17 21:48:39,286][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 21:48:39,301][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 21:48:43,093][starVLA.training.train_starvla][INFO] - Step 3201, Loss: {'train/loss': 0.6564207077026367, 'train/grad_norm_pre_clip': 25.75064468383789, 'timing/data': 0.011396832764148712, 'timing/model': 0.9371682750061154, 'learning_rate/qwen_vl_interface': 6.001606729054318e-06, 'learning_rate/action_model': 6.0016067290543176e-05, 'epoch': 1.45}) +[2026-06-17 21:48:47,041][starVLA.training.train_starvla][INFO] - Step 3202, Loss: {'train/loss': 0.30243510007858276, 'train/grad_norm_pre_clip': 50.977928161621094, 'timing/data': 0.008094443008303642, 'timing/model': 0.9492092011496425, 'learning_rate/qwen_vl_interface': 5.999471206632559e-06, 'learning_rate/action_model': 5.999471206632559e-05, 'epoch': 1.45}) +[2026-06-17 21:48:51,020][starVLA.training.train_starvla][INFO] - Step 3203, Loss: {'train/loss': 0.6316738128662109, 'train/grad_norm_pre_clip': 21.74134063720703, 'timing/data': 0.011852411553263664, 'timing/model': 0.9796473905444145, 'learning_rate/qwen_vl_interface': 5.99733552884461e-06, 'learning_rate/action_model': 5.99733552884461e-05, 'epoch': 1.45}) +[2026-06-17 21:48:54,964][starVLA.training.train_starvla][INFO] - Step 3204, Loss: {'train/loss': 0.613206684589386, 'train/grad_norm_pre_clip': 22.503982543945312, 'timing/data': 0.007774785161018372, 'timing/model': 1.1485010040923953, 'learning_rate/qwen_vl_interface': 5.995199696133204e-06, 'learning_rate/action_model': 5.9951996961332034e-05, 'epoch': 1.45}) +[2026-06-17 21:48:58,759][starVLA.training.train_starvla][INFO] - Step 3205, Loss: {'train/loss': 0.22853082418441772, 'train/grad_norm_pre_clip': 13.20676326751709, 'timing/data': 0.011037612333893776, 'timing/model': 1.0663389153778553, 'learning_rate/qwen_vl_interface': 5.993063708941098e-06, 'learning_rate/action_model': 5.993063708941098e-05, 'epoch': 1.45}) +[2026-06-17 21:49:03,456][starVLA.training.train_starvla][INFO] - Step 3206, Loss: {'train/loss': 0.2289479374885559, 'train/grad_norm_pre_clip': 10.83875560760498, 'timing/data': 0.010326537303626537, 'timing/model': 0.9507036525756121, 'learning_rate/qwen_vl_interface': 5.990927567711085e-06, 'learning_rate/action_model': 5.990927567711085e-05, 'epoch': 1.45}) +[2026-06-17 21:49:07,834][starVLA.training.train_starvla][INFO] - Step 3207, Loss: {'train/loss': 0.28905779123306274, 'train/grad_norm_pre_clip': 18.18225860595703, 'timing/data': 0.009496570564806461, 'timing/model': 0.9663105392828584, 'learning_rate/qwen_vl_interface': 5.988791272885991e-06, 'learning_rate/action_model': 5.988791272885991e-05, 'epoch': 1.45}) +[2026-06-17 21:49:11,829][starVLA.training.train_starvla][INFO] - Step 3208, Loss: {'train/loss': 0.34139344096183777, 'train/grad_norm_pre_clip': 17.207836151123047, 'timing/data': 0.008799020200967789, 'timing/model': 0.9702442353591323, 'learning_rate/qwen_vl_interface': 5.986654824908671e-06, 'learning_rate/action_model': 5.9866548249086706e-05, 'epoch': 1.46}) +[2026-06-17 21:49:15,803][starVLA.training.train_starvla][INFO] - Step 3209, Loss: {'train/loss': 0.3023386001586914, 'train/grad_norm_pre_clip': 16.47818374633789, 'timing/data': 0.009168833494186401, 'timing/model': 1.007435985840857, 'learning_rate/qwen_vl_interface': 5.9845182242220125e-06, 'learning_rate/action_model': 5.984518224222012e-05, 'epoch': 1.46}) +[2026-06-17 21:49:19,645][starVLA.training.train_starvla][INFO] - Step 3210, Loss: {'train/loss': 0.27268344163894653, 'train/grad_norm_pre_clip': 203.09349060058594, 'timing/data': 0.009511293843388557, 'timing/model': 0.9522305820137262, 'learning_rate/qwen_vl_interface': 5.982381471268935e-06, 'learning_rate/action_model': 5.9823814712689354e-05, 'epoch': 1.46}) +[2026-06-17 21:49:23,512][starVLA.training.train_starvla][INFO] - Step 3211, Loss: {'train/loss': 0.29920336604118347, 'train/grad_norm_pre_clip': 12.008967399597168, 'timing/data': 0.00852093193680048, 'timing/model': 1.1162514798343182, 'learning_rate/qwen_vl_interface': 5.980244566492392e-06, 'learning_rate/action_model': 5.9802445664923925e-05, 'epoch': 1.46}) +[2026-06-17 21:49:27,221][starVLA.training.train_starvla][INFO] - Step 3212, Loss: {'train/loss': 0.13230957090854645, 'train/grad_norm_pre_clip': 10.220683097839355, 'timing/data': 0.012063764035701752, 'timing/model': 0.963139814324677, 'learning_rate/qwen_vl_interface': 5.978107510335365e-06, 'learning_rate/action_model': 5.9781075103353647e-05, 'epoch': 1.46}) +[2026-06-17 21:49:31,076][starVLA.training.train_starvla][INFO] - Step 3213, Loss: {'train/loss': 0.43055474758148193, 'train/grad_norm_pre_clip': 17.791786193847656, 'timing/data': 0.02049637772142887, 'timing/model': 0.9419641736894846, 'learning_rate/qwen_vl_interface': 5.975970303240867e-06, 'learning_rate/action_model': 5.9759703032408664e-05, 'epoch': 1.46}) +[2026-06-17 21:49:34,939][starVLA.training.train_starvla][INFO] - Step 3214, Loss: {'train/loss': 0.316397100687027, 'train/grad_norm_pre_clip': 12.083740234375, 'timing/data': 0.01067375298589468, 'timing/model': 0.9448077119886875, 'learning_rate/qwen_vl_interface': 5.973832945651944e-06, 'learning_rate/action_model': 5.9738329456519435e-05, 'epoch': 1.46}) +[2026-06-17 21:49:38,765][starVLA.training.train_starvla][INFO] - Step 3215, Loss: {'train/loss': 0.315197616815567, 'train/grad_norm_pre_clip': 15.766602516174316, 'timing/data': 0.010963546112179756, 'timing/model': 0.9440169120207429, 'learning_rate/qwen_vl_interface': 5.971695438011672e-06, 'learning_rate/action_model': 5.971695438011672e-05, 'epoch': 1.46}) +[2026-06-17 21:49:42,857][starVLA.training.train_starvla][INFO] - Step 3216, Loss: {'train/loss': 0.4291514456272125, 'train/grad_norm_pre_clip': 20.924560546875, 'timing/data': 0.007825275883078575, 'timing/model': 0.9511006847023964, 'learning_rate/qwen_vl_interface': 5.969557780763159e-06, 'learning_rate/action_model': 5.9695577807631585e-05, 'epoch': 1.46}) +[2026-06-17 21:49:48,374][starVLA.training.train_starvla][INFO] - Step 3217, Loss: {'train/loss': 0.5187930464744568, 'train/grad_norm_pre_clip': 21.15732765197754, 'timing/data': 0.008522897027432919, 'timing/model': 1.1110280342400074, 'learning_rate/qwen_vl_interface': 5.9674199743495456e-06, 'learning_rate/action_model': 5.967419974349545e-05, 'epoch': 1.46}) +[2026-06-17 21:49:52,190][starVLA.training.train_starvla][INFO] - Step 3218, Loss: {'train/loss': 0.2718111574649811, 'train/grad_norm_pre_clip': 14.114950180053711, 'timing/data': 0.008248542435467243, 'timing/model': 1.1160317612811923, 'learning_rate/qwen_vl_interface': 5.965282019213996e-06, 'learning_rate/action_model': 5.965282019213996e-05, 'epoch': 1.46}) +[2026-06-17 21:49:55,872][starVLA.training.train_starvla][INFO] - Step 3219, Loss: {'train/loss': 0.45236140489578247, 'train/grad_norm_pre_clip': 17.244678497314453, 'timing/data': 0.009043264202773571, 'timing/model': 0.9421142227947712, 'learning_rate/qwen_vl_interface': 5.963143915799718e-06, 'learning_rate/action_model': 5.9631439157997184e-05, 'epoch': 1.46}) +[2026-06-17 21:49:59,752][starVLA.training.train_starvla][INFO] - Step 3220, Loss: {'train/loss': 0.2516646087169647, 'train/grad_norm_pre_clip': 14.19039535522461, 'timing/data': 0.012637434527277946, 'timing/model': 0.9653016068041325, 'learning_rate/qwen_vl_interface': 5.961005664549936e-06, 'learning_rate/action_model': 5.9610056645499354e-05, 'epoch': 1.46}) +[2026-06-17 21:50:03,684][starVLA.training.train_starvla][INFO] - Step 3221, Loss: {'train/loss': 0.2840871810913086, 'train/grad_norm_pre_clip': 14.13385009765625, 'timing/data': 0.00958817359060049, 'timing/model': 0.9450258892029524, 'learning_rate/qwen_vl_interface': 5.958867265907916e-06, 'learning_rate/action_model': 5.958867265907916e-05, 'epoch': 1.46}) +[2026-06-17 21:50:07,500][starVLA.training.train_starvla][INFO] - Step 3222, Loss: {'train/loss': 0.14383776485919952, 'train/grad_norm_pre_clip': 9.860185623168945, 'timing/data': 0.012369041331112385, 'timing/model': 0.948534476570785, 'learning_rate/qwen_vl_interface': 5.956728720316949e-06, 'learning_rate/action_model': 5.9567287203169485e-05, 'epoch': 1.46}) +[2026-06-17 21:50:11,370][starVLA.training.train_starvla][INFO] - Step 3223, Loss: {'train/loss': 0.32888588309288025, 'train/grad_norm_pre_clip': 14.31444263458252, 'timing/data': 0.007898302748799324, 'timing/model': 0.930770218372345, 'learning_rate/qwen_vl_interface': 5.954590028220357e-06, 'learning_rate/action_model': 5.954590028220357e-05, 'epoch': 1.46}) +[2026-06-17 21:50:15,207][starVLA.training.train_starvla][INFO] - Step 3224, Loss: {'train/loss': 0.3872443437576294, 'train/grad_norm_pre_clip': 15.38440990447998, 'timing/data': 0.009670417755842209, 'timing/model': 1.1147097572684288, 'learning_rate/qwen_vl_interface': 5.952451190061496e-06, 'learning_rate/action_model': 5.9524511900614956e-05, 'epoch': 1.46}) +[2026-06-17 21:50:18,870][starVLA.training.train_starvla][INFO] - Step 3225, Loss: {'train/loss': 0.5546001195907593, 'train/grad_norm_pre_clip': 23.979589462280273, 'timing/data': 0.01407970767468214, 'timing/model': 0.9316672096028924, 'learning_rate/qwen_vl_interface': 5.950312206283743e-06, 'learning_rate/action_model': 5.9503122062837434e-05, 'epoch': 1.46}) +[2026-06-17 21:50:22,833][starVLA.training.train_starvla][INFO] - Step 3226, Loss: {'train/loss': 0.7237421870231628, 'train/grad_norm_pre_clip': 27.396522521972656, 'timing/data': 0.008846559561789036, 'timing/model': 0.9608413781970739, 'learning_rate/qwen_vl_interface': 5.948173077330522e-06, 'learning_rate/action_model': 5.948173077330522e-05, 'epoch': 1.46}) +[2026-06-17 21:50:26,634][starVLA.training.train_starvla][INFO] - Step 3227, Loss: {'train/loss': 0.14545810222625732, 'train/grad_norm_pre_clip': 10.051223754882812, 'timing/data': 0.01053006574511528, 'timing/model': 0.9366352949291468, 'learning_rate/qwen_vl_interface': 5.9460338036452665e-06, 'learning_rate/action_model': 5.946033803645267e-05, 'epoch': 1.46}) +[2026-06-17 21:50:30,438][starVLA.training.train_starvla][INFO] - Step 3228, Loss: {'train/loss': 0.42190858721733093, 'train/grad_norm_pre_clip': 16.62795639038086, 'timing/data': 0.01019158773124218, 'timing/model': 0.9462355310097337, 'learning_rate/qwen_vl_interface': 5.943894385671456e-06, 'learning_rate/action_model': 5.9438943856714556e-05, 'epoch': 1.46}) +[2026-06-17 21:50:34,423][starVLA.training.train_starvla][INFO] - Step 3229, Loss: {'train/loss': 0.38629060983657837, 'train/grad_norm_pre_clip': 15.078855514526367, 'timing/data': 0.010426744818687439, 'timing/model': 0.9520126013085246, 'learning_rate/qwen_vl_interface': 5.941754823852593e-06, 'learning_rate/action_model': 5.9417548238525924e-05, 'epoch': 1.46}) +[2026-06-17 21:50:38,328][starVLA.training.train_starvla][INFO] - Step 3230, Loss: {'train/loss': 0.18403500318527222, 'train/grad_norm_pre_clip': 13.764532089233398, 'timing/data': 0.0071580670773983, 'timing/model': 0.9592640604823828, 'learning_rate/qwen_vl_interface': 5.939615118632208e-06, 'learning_rate/action_model': 5.939615118632208e-05, 'epoch': 1.47}) +[2026-06-17 21:50:42,129][starVLA.training.train_starvla][INFO] - Step 3231, Loss: {'train/loss': 0.4476173520088196, 'train/grad_norm_pre_clip': 14.066537857055664, 'timing/data': 0.008126704953610897, 'timing/model': 1.1147103272378445, 'learning_rate/qwen_vl_interface': 5.937475270453872e-06, 'learning_rate/action_model': 5.937475270453872e-05, 'epoch': 1.47}) +[2026-06-17 21:50:45,914][starVLA.training.train_starvla][INFO] - Step 3232, Loss: {'train/loss': 0.239755317568779, 'train/grad_norm_pre_clip': 16.477331161499023, 'timing/data': 0.011774105951189995, 'timing/model': 0.9515534481033683, 'learning_rate/qwen_vl_interface': 5.935335279761169e-06, 'learning_rate/action_model': 5.935335279761168e-05, 'epoch': 1.47}) +[2026-06-17 21:50:49,745][starVLA.training.train_starvla][INFO] - Step 3233, Loss: {'train/loss': 0.4450269341468811, 'train/grad_norm_pre_clip': 21.554407119750977, 'timing/data': 0.00819743424654007, 'timing/model': 0.9463726906105876, 'learning_rate/qwen_vl_interface': 5.933195146997725e-06, 'learning_rate/action_model': 5.9331951469977254e-05, 'epoch': 1.47}) +[2026-06-17 21:50:53,589][starVLA.training.train_starvla][INFO] - Step 3234, Loss: {'train/loss': 0.34613680839538574, 'train/grad_norm_pre_clip': 20.275861740112305, 'timing/data': 0.01011475920677185, 'timing/model': 0.9463674416765571, 'learning_rate/qwen_vl_interface': 5.931054872607193e-06, 'learning_rate/action_model': 5.9310548726071925e-05, 'epoch': 1.47}) +[2026-06-17 21:50:57,423][starVLA.training.train_starvla][INFO] - Step 3235, Loss: {'train/loss': 0.22994926571846008, 'train/grad_norm_pre_clip': 16.593721389770508, 'timing/data': 0.008310091681778431, 'timing/model': 0.9362105010077357, 'learning_rate/qwen_vl_interface': 5.928914457033252e-06, 'learning_rate/action_model': 5.928914457033252e-05, 'epoch': 1.47}) +[2026-06-17 21:51:01,192][starVLA.training.train_starvla][INFO] - Step 3236, Loss: {'train/loss': 0.5582430362701416, 'train/grad_norm_pre_clip': 21.000537872314453, 'timing/data': 0.0074262069538235664, 'timing/model': 0.9336821073666215, 'learning_rate/qwen_vl_interface': 5.926773900719616e-06, 'learning_rate/action_model': 5.9267739007196157e-05, 'epoch': 1.47}) +[2026-06-17 21:51:05,083][starVLA.training.train_starvla][INFO] - Step 3237, Loss: {'train/loss': 0.205290749669075, 'train/grad_norm_pre_clip': 14.088025093078613, 'timing/data': 0.0088195176795125, 'timing/model': 1.1679296251386404, 'learning_rate/qwen_vl_interface': 5.92463320411002e-06, 'learning_rate/action_model': 5.92463320411002e-05, 'epoch': 1.47}) +[2026-06-17 21:51:08,936][starVLA.training.train_starvla][INFO] - Step 3238, Loss: {'train/loss': 0.140769362449646, 'train/grad_norm_pre_clip': 14.556107521057129, 'timing/data': 0.008144775405526161, 'timing/model': 1.1406788881868124, 'learning_rate/qwen_vl_interface': 5.922492367648236e-06, 'learning_rate/action_model': 5.922492367648236e-05, 'epoch': 1.47}) +[2026-06-17 21:51:12,546][starVLA.training.train_starvla][INFO] - Step 3239, Loss: {'train/loss': 0.39025676250457764, 'train/grad_norm_pre_clip': 25.46837043762207, 'timing/data': 0.008178231306374073, 'timing/model': 0.9471339397132397, 'learning_rate/qwen_vl_interface': 5.920351391778059e-06, 'learning_rate/action_model': 5.920351391778059e-05, 'epoch': 1.47}) +[2026-06-17 21:51:16,313][starVLA.training.train_starvla][INFO] - Step 3240, Loss: {'train/loss': 0.3358144164085388, 'train/grad_norm_pre_clip': 19.48633575439453, 'timing/data': 0.017675643786787987, 'timing/model': 0.9169407989829779, 'learning_rate/qwen_vl_interface': 5.918210276943318e-06, 'learning_rate/action_model': 5.918210276943318e-05, 'epoch': 1.47}) +[2026-06-17 21:51:20,149][starVLA.training.train_starvla][INFO] - Step 3241, Loss: {'train/loss': 0.3987047076225281, 'train/grad_norm_pre_clip': 20.97567367553711, 'timing/data': 0.010631716810166836, 'timing/model': 0.941104156896472, 'learning_rate/qwen_vl_interface': 5.916069023587867e-06, 'learning_rate/action_model': 5.916069023587867e-05, 'epoch': 1.47}) +[2026-06-17 21:51:23,935][starVLA.training.train_starvla][INFO] - Step 3242, Loss: {'train/loss': 0.38883352279663086, 'train/grad_norm_pre_clip': 16.88030242919922, 'timing/data': 0.009987341240048409, 'timing/model': 0.9509890610352159, 'learning_rate/qwen_vl_interface': 5.913927632155592e-06, 'learning_rate/action_model': 5.913927632155591e-05, 'epoch': 1.47}) +[2026-06-17 21:51:28,105][starVLA.training.train_starvla][INFO] - Step 3243, Loss: {'train/loss': 0.17480051517486572, 'train/grad_norm_pre_clip': 10.420056343078613, 'timing/data': 0.012290309183299541, 'timing/model': 1.2398011293262243, 'learning_rate/qwen_vl_interface': 5.911786103090402e-06, 'learning_rate/action_model': 5.911786103090402e-05, 'epoch': 1.47}) +[2026-06-17 21:51:32,909][starVLA.training.train_starvla][INFO] - Step 3244, Loss: {'train/loss': 0.23234567046165466, 'train/grad_norm_pre_clip': 14.251843452453613, 'timing/data': 0.00805218517780304, 'timing/model': 1.2322031883522868, 'learning_rate/qwen_vl_interface': 5.909644436836243e-06, 'learning_rate/action_model': 5.909644436836244e-05, 'epoch': 1.47}) +[2026-06-17 21:51:36,718][starVLA.training.train_starvla][INFO] - Step 3245, Loss: {'train/loss': 0.1779496669769287, 'train/grad_norm_pre_clip': 13.18783950805664, 'timing/data': 0.007378395646810532, 'timing/model': 1.1150299198925495, 'learning_rate/qwen_vl_interface': 5.907502633837082e-06, 'learning_rate/action_model': 5.907502633837082e-05, 'epoch': 1.47}) +[2026-06-17 21:51:40,347][starVLA.training.train_starvla][INFO] - Step 3246, Loss: {'train/loss': 0.2986064851284027, 'train/grad_norm_pre_clip': 15.312186241149902, 'timing/data': 0.008066866546869278, 'timing/model': 0.9345774566754699, 'learning_rate/qwen_vl_interface': 5.905360694536918e-06, 'learning_rate/action_model': 5.905360694536918e-05, 'epoch': 1.47}) +[2026-06-17 21:51:44,307][starVLA.training.train_starvla][INFO] - Step 3247, Loss: {'train/loss': 0.7983866333961487, 'train/grad_norm_pre_clip': 24.677223205566406, 'timing/data': 0.018950002267956734, 'timing/model': 0.9477525474503636, 'learning_rate/qwen_vl_interface': 5.903218619379776e-06, 'learning_rate/action_model': 5.903218619379776e-05, 'epoch': 1.47}) +[2026-06-17 21:51:48,269][starVLA.training.train_starvla][INFO] - Step 3248, Loss: {'train/loss': 0.2631634473800659, 'train/grad_norm_pre_clip': 14.253229141235352, 'timing/data': 0.008662219159305096, 'timing/model': 1.0079473806545138, 'learning_rate/qwen_vl_interface': 5.901076408809712e-06, 'learning_rate/action_model': 5.9010764088097114e-05, 'epoch': 1.47}) +[2026-06-17 21:51:52,107][starVLA.training.train_starvla][INFO] - Step 3249, Loss: {'train/loss': 0.3137892186641693, 'train/grad_norm_pre_clip': 19.288036346435547, 'timing/data': 0.012244113720953465, 'timing/model': 0.9575470983982086, 'learning_rate/qwen_vl_interface': 5.898934063270806e-06, 'learning_rate/action_model': 5.8989340632708056e-05, 'epoch': 1.47}) +[2026-06-17 21:51:56,163][starVLA.training.train_starvla][INFO] - Step 3250, Loss: {'train/loss': 0.3309788405895233, 'train/grad_norm_pre_clip': 25.881898880004883, 'timing/data': 0.009713435545563698, 'timing/model': 1.1488281777128577, 'learning_rate/qwen_vl_interface': 5.896791583207173e-06, 'learning_rate/action_model': 5.896791583207173e-05, 'epoch': 1.47}) +[2026-06-17 21:52:00,040][starVLA.training.train_starvla][INFO] - Step 3251, Loss: {'train/loss': 0.5892313122749329, 'train/grad_norm_pre_clip': 29.471954345703125, 'timing/data': 0.009459222666919231, 'timing/model': 1.1305567594245076, 'learning_rate/qwen_vl_interface': 5.894648969062947e-06, 'learning_rate/action_model': 5.894648969062947e-05, 'epoch': 1.47}) +[2026-06-17 21:52:03,729][starVLA.training.train_starvla][INFO] - Step 3252, Loss: {'train/loss': 0.14283110201358795, 'train/grad_norm_pre_clip': 10.946000099182129, 'timing/data': 0.008424017578363419, 'timing/model': 0.9576732739806175, 'learning_rate/qwen_vl_interface': 5.892506221282297e-06, 'learning_rate/action_model': 5.892506221282297e-05, 'epoch': 1.48}) +[2026-06-17 21:52:07,656][starVLA.training.train_starvla][INFO] - Step 3253, Loss: {'train/loss': 0.5493659973144531, 'train/grad_norm_pre_clip': 17.928436279296875, 'timing/data': 0.010645122267305851, 'timing/model': 0.9449578309431672, 'learning_rate/qwen_vl_interface': 5.890363340309415e-06, 'learning_rate/action_model': 5.890363340309415e-05, 'epoch': 1.48}) +[2026-06-17 21:52:14,948][starVLA.training.train_starvla][INFO] - Step 3254, Loss: {'train/loss': 0.3350819945335388, 'train/grad_norm_pre_clip': 20.266300201416016, 'timing/data': 0.007736559957265854, 'timing/model': 2.140929396264255, 'learning_rate/qwen_vl_interface': 5.888220326588523e-06, 'learning_rate/action_model': 5.8882203265885226e-05, 'epoch': 1.48}) +[2026-06-17 21:52:18,847][starVLA.training.train_starvla][INFO] - Step 3255, Loss: {'train/loss': 0.24395738542079926, 'train/grad_norm_pre_clip': 17.447959899902344, 'timing/data': 0.010613820515573025, 'timing/model': 0.9518776498734951, 'learning_rate/qwen_vl_interface': 5.8860771805638725e-06, 'learning_rate/action_model': 5.886077180563872e-05, 'epoch': 1.48}) +[2026-06-17 21:52:22,799][starVLA.training.train_starvla][INFO] - Step 3256, Loss: {'train/loss': 0.34773021936416626, 'train/grad_norm_pre_clip': 16.559526443481445, 'timing/data': 0.01749745849519968, 'timing/model': 1.0524534899741411, 'learning_rate/qwen_vl_interface': 5.883933902679736e-06, 'learning_rate/action_model': 5.8839339026797355e-05, 'epoch': 1.48}) +[2026-06-17 21:52:26,728][starVLA.training.train_starvla][INFO] - Step 3257, Loss: {'train/loss': 0.560064971446991, 'train/grad_norm_pre_clip': 36.193275451660156, 'timing/data': 0.009942022152245045, 'timing/model': 1.144032122567296, 'learning_rate/qwen_vl_interface': 5.8817904933804214e-06, 'learning_rate/action_model': 5.881790493380421e-05, 'epoch': 1.48}) +[2026-06-17 21:52:31,509][starVLA.training.train_starvla][INFO] - Step 3258, Loss: {'train/loss': 0.3058127164840698, 'train/grad_norm_pre_clip': 21.817358016967773, 'timing/data': 0.016273624263703823, 'timing/model': 1.1411998514086008, 'learning_rate/qwen_vl_interface': 5.8796469531102565e-06, 'learning_rate/action_model': 5.879646953110256e-05, 'epoch': 1.48}) +[2026-06-17 21:52:35,447][starVLA.training.train_starvla][INFO] - Step 3259, Loss: {'train/loss': 0.4054499566555023, 'train/grad_norm_pre_clip': 19.079710006713867, 'timing/data': 0.015975586138665676, 'timing/model': 0.9498049197718501, 'learning_rate/qwen_vl_interface': 5.877503282313601e-06, 'learning_rate/action_model': 5.877503282313601e-05, 'epoch': 1.48}) +[2026-06-17 21:52:39,356][starVLA.training.train_starvla][INFO] - Step 3260, Loss: {'train/loss': 0.48040780425071716, 'train/grad_norm_pre_clip': 21.89034080505371, 'timing/data': 0.012956840917468071, 'timing/model': 0.9669859036803246, 'learning_rate/qwen_vl_interface': 5.875359481434839e-06, 'learning_rate/action_model': 5.875359481434839e-05, 'epoch': 1.48}) +[2026-06-17 21:52:43,336][starVLA.training.train_starvla][INFO] - Step 3261, Loss: {'train/loss': 0.15276826918125153, 'train/grad_norm_pre_clip': 11.212496757507324, 'timing/data': 0.0160342026501894, 'timing/model': 0.9452920025214553, 'learning_rate/qwen_vl_interface': 5.873215550918384e-06, 'learning_rate/action_model': 5.873215550918384e-05, 'epoch': 1.48}) +[2026-06-17 21:52:47,161][starVLA.training.train_starvla][INFO] - Step 3262, Loss: {'train/loss': 0.4463594853878021, 'train/grad_norm_pre_clip': 18.193422317504883, 'timing/data': 0.007966889999806881, 'timing/model': 0.9653220735490322, 'learning_rate/qwen_vl_interface': 5.871071491208676e-06, 'learning_rate/action_model': 5.871071491208676e-05, 'epoch': 1.48}) +[2026-06-17 21:52:51,080][starVLA.training.train_starvla][INFO] - Step 3263, Loss: {'train/loss': 0.39905181527137756, 'train/grad_norm_pre_clip': 17.473691940307617, 'timing/data': 0.01008645910769701, 'timing/model': 0.9876865604892373, 'learning_rate/qwen_vl_interface': 5.868927302750178e-06, 'learning_rate/action_model': 5.868927302750178e-05, 'epoch': 1.48}) +[2026-06-17 21:52:55,451][starVLA.training.train_starvla][INFO] - Step 3264, Loss: {'train/loss': 0.2583157420158386, 'train/grad_norm_pre_clip': 15.998058319091797, 'timing/data': 0.007358118891716003, 'timing/model': 1.6441000867635012, 'learning_rate/qwen_vl_interface': 5.866782985987386e-06, 'learning_rate/action_model': 5.866782985987386e-05, 'epoch': 1.48}) +[2026-06-17 21:52:59,363][starVLA.training.train_starvla][INFO] - Step 3265, Loss: {'train/loss': 0.35623273253440857, 'train/grad_norm_pre_clip': 18.019424438476562, 'timing/data': 0.008177832700312138, 'timing/model': 1.164018944837153, 'learning_rate/qwen_vl_interface': 5.864638541364815e-06, 'learning_rate/action_model': 5.8646385413648154e-05, 'epoch': 1.48}) +[2026-06-17 21:53:03,047][starVLA.training.train_starvla][INFO] - Step 3266, Loss: {'train/loss': 0.4116021990776062, 'train/grad_norm_pre_clip': 22.64571762084961, 'timing/data': 0.012179271318018436, 'timing/model': 0.9434961890801787, 'learning_rate/qwen_vl_interface': 5.862493969327015e-06, 'learning_rate/action_model': 5.862493969327014e-05, 'epoch': 1.48}) +[2026-06-17 21:53:06,950][starVLA.training.train_starvla][INFO] - Step 3267, Loss: {'train/loss': 0.07774028182029724, 'train/grad_norm_pre_clip': 6.221593856811523, 'timing/data': 0.008884006179869175, 'timing/model': 0.9815845498815179, 'learning_rate/qwen_vl_interface': 5.860349270318555e-06, 'learning_rate/action_model': 5.860349270318555e-05, 'epoch': 1.48}) +[2026-06-17 21:53:10,891][starVLA.training.train_starvla][INFO] - Step 3268, Loss: {'train/loss': 0.20397202670574188, 'train/grad_norm_pre_clip': 13.015457153320312, 'timing/data': 0.010819495655596256, 'timing/model': 0.9481261037290096, 'learning_rate/qwen_vl_interface': 5.8582044447840346e-06, 'learning_rate/action_model': 5.858204444784034e-05, 'epoch': 1.48}) +[2026-06-17 21:53:14,680][starVLA.training.train_starvla][INFO] - Step 3269, Loss: {'train/loss': 0.3845374882221222, 'train/grad_norm_pre_clip': 20.435054779052734, 'timing/data': 0.008470576256513596, 'timing/model': 0.9402250004932284, 'learning_rate/qwen_vl_interface': 5.8560594931680785e-06, 'learning_rate/action_model': 5.856059493168078e-05, 'epoch': 1.48}) +[2026-06-17 21:53:18,558][starVLA.training.train_starvla][INFO] - Step 3270, Loss: {'train/loss': 0.3742082417011261, 'train/grad_norm_pre_clip': 18.221965789794922, 'timing/data': 0.007799485698342323, 'timing/model': 0.9575092839077115, 'learning_rate/qwen_vl_interface': 5.853914415915337e-06, 'learning_rate/action_model': 5.853914415915337e-05, 'epoch': 1.48}) +[2026-06-17 21:53:22,412][starVLA.training.train_starvla][INFO] - Step 3271, Loss: {'train/loss': 0.4489530920982361, 'train/grad_norm_pre_clip': 20.11472511291504, 'timing/data': 0.008714433759450912, 'timing/model': 1.1288342159241438, 'learning_rate/qwen_vl_interface': 5.8517692134704875e-06, 'learning_rate/action_model': 5.851769213470487e-05, 'epoch': 1.48}) +[2026-06-17 21:53:26,043][starVLA.training.train_starvla][INFO] - Step 3272, Loss: {'train/loss': 0.502159833908081, 'train/grad_norm_pre_clip': 24.190128326416016, 'timing/data': 0.016940372064709663, 'timing/model': 0.9498458411544561, 'learning_rate/qwen_vl_interface': 5.849623886278232e-06, 'learning_rate/action_model': 5.849623886278231e-05, 'epoch': 1.48}) +[2026-06-17 21:53:29,898][starVLA.training.train_starvla][INFO] - Step 3273, Loss: {'train/loss': 0.14247967302799225, 'train/grad_norm_pre_clip': 13.546968460083008, 'timing/data': 0.007457521744072437, 'timing/model': 0.9424543185159564, 'learning_rate/qwen_vl_interface': 5.8474784347833e-06, 'learning_rate/action_model': 5.8474784347833e-05, 'epoch': 1.48}) +[2026-06-17 21:53:33,790][starVLA.training.train_starvla][INFO] - Step 3274, Loss: {'train/loss': 0.27826836705207825, 'train/grad_norm_pre_clip': 20.03481674194336, 'timing/data': 0.010203910991549492, 'timing/model': 0.9980861758813262, 'learning_rate/qwen_vl_interface': 5.845332859430446e-06, 'learning_rate/action_model': 5.845332859430446e-05, 'epoch': 1.49}) +[2026-06-17 21:53:37,664][starVLA.training.train_starvla][INFO] - Step 3275, Loss: {'train/loss': 0.4483318626880646, 'train/grad_norm_pre_clip': 24.785011291503906, 'timing/data': 0.008725721389055252, 'timing/model': 0.9321664338931441, 'learning_rate/qwen_vl_interface': 5.843187160664449e-06, 'learning_rate/action_model': 5.8431871606644495e-05, 'epoch': 1.49}) +[2026-06-17 21:53:41,556][starVLA.training.train_starvla][INFO] - Step 3276, Loss: {'train/loss': 0.759696364402771, 'train/grad_norm_pre_clip': 23.674602508544922, 'timing/data': 0.008793652057647705, 'timing/model': 0.9369634464383125, 'learning_rate/qwen_vl_interface': 5.841041338930115e-06, 'learning_rate/action_model': 5.8410413389301155e-05, 'epoch': 1.49}) +[2026-06-17 21:53:45,361][starVLA.training.train_starvla][INFO] - Step 3277, Loss: {'train/loss': 0.5003364682197571, 'train/grad_norm_pre_clip': 18.155475616455078, 'timing/data': 0.00799550861120224, 'timing/model': 0.9344927063211799, 'learning_rate/qwen_vl_interface': 5.8388953946722775e-06, 'learning_rate/action_model': 5.8388953946722766e-05, 'epoch': 1.49}) +[2026-06-17 21:53:49,237][starVLA.training.train_starvla][INFO] - Step 3278, Loss: {'train/loss': 0.7471534013748169, 'train/grad_norm_pre_clip': 23.42294692993164, 'timing/data': 0.009753414429724216, 'timing/model': 1.1253236522898078, 'learning_rate/qwen_vl_interface': 5.836749328335789e-06, 'learning_rate/action_model': 5.8367493283357885e-05, 'epoch': 1.49}) +[2026-06-17 21:53:52,900][starVLA.training.train_starvla][INFO] - Step 3279, Loss: {'train/loss': 0.4508717656135559, 'train/grad_norm_pre_clip': 20.439476013183594, 'timing/data': 0.015347084030508995, 'timing/model': 0.9392702411860228, 'learning_rate/qwen_vl_interface': 5.8346031403655345e-06, 'learning_rate/action_model': 5.8346031403655345e-05, 'epoch': 1.49}) +[2026-06-17 21:53:56,740][starVLA.training.train_starvla][INFO] - Step 3280, Loss: {'train/loss': 0.16494148969650269, 'train/grad_norm_pre_clip': 39.157859802246094, 'timing/data': 0.009951367974281311, 'timing/model': 0.9482420375570655, 'learning_rate/qwen_vl_interface': 5.832456831206418e-06, 'learning_rate/action_model': 5.8324568312064176e-05, 'epoch': 1.49}) +[2026-06-17 21:54:00,831][starVLA.training.train_starvla][INFO] - Step 3281, Loss: {'train/loss': 0.3136615455150604, 'train/grad_norm_pre_clip': 15.881585121154785, 'timing/data': 0.009261361323297024, 'timing/model': 0.9388453373685479, 'learning_rate/qwen_vl_interface': 5.830310401303377e-06, 'learning_rate/action_model': 5.830310401303377e-05, 'epoch': 1.49}) +[2026-06-17 21:54:04,695][starVLA.training.train_starvla][INFO] - Step 3282, Loss: {'train/loss': 0.3861633539199829, 'train/grad_norm_pre_clip': 21.26931381225586, 'timing/data': 0.009884853847324848, 'timing/model': 0.9994603265076876, 'learning_rate/qwen_vl_interface': 5.8281638511013625e-06, 'learning_rate/action_model': 5.828163851101363e-05, 'epoch': 1.49}) +[2026-06-17 21:54:08,638][starVLA.training.train_starvla][INFO] - Step 3283, Loss: {'train/loss': 0.3021753430366516, 'train/grad_norm_pre_clip': 18.259490966796875, 'timing/data': 0.009957496076822281, 'timing/model': 0.932547252625227, 'learning_rate/qwen_vl_interface': 5.8260171810453605e-06, 'learning_rate/action_model': 5.82601718104536e-05, 'epoch': 1.49}) +[2026-06-17 21:54:12,464][starVLA.training.train_starvla][INFO] - Step 3284, Loss: {'train/loss': 0.22928543388843536, 'train/grad_norm_pre_clip': 16.38951873779297, 'timing/data': 0.008857027627527714, 'timing/model': 1.1091961180791259, 'learning_rate/qwen_vl_interface': 5.823870391580376e-06, 'learning_rate/action_model': 5.823870391580376e-05, 'epoch': 1.49}) +[2026-06-17 21:54:16,413][starVLA.training.train_starvla][INFO] - Step 3285, Loss: {'train/loss': 0.1616775095462799, 'train/grad_norm_pre_clip': 11.677117347717285, 'timing/data': 0.009107544086873531, 'timing/model': 1.1183441393077374, 'learning_rate/qwen_vl_interface': 5.821723483151444e-06, 'learning_rate/action_model': 5.821723483151443e-05, 'epoch': 1.49}) +[2026-06-17 21:54:20,056][starVLA.training.train_starvla][INFO] - Step 3286, Loss: {'train/loss': 0.5431374907493591, 'train/grad_norm_pre_clip': 20.26789093017578, 'timing/data': 0.013873795978724957, 'timing/model': 0.9308695755898952, 'learning_rate/qwen_vl_interface': 5.819576456203617e-06, 'learning_rate/action_model': 5.8195764562036174e-05, 'epoch': 1.49}) +[2026-06-17 21:54:24,139][starVLA.training.train_starvla][INFO] - Step 3287, Loss: {'train/loss': 0.25012877583503723, 'train/grad_norm_pre_clip': 15.657448768615723, 'timing/data': 0.009636590257287025, 'timing/model': 0.9383556852117181, 'learning_rate/qwen_vl_interface': 5.817429311181977e-06, 'learning_rate/action_model': 5.8174293111819764e-05, 'epoch': 1.49}) +[2026-06-17 21:54:28,031][starVLA.training.train_starvla][INFO] - Step 3288, Loss: {'train/loss': 0.7360954284667969, 'train/grad_norm_pre_clip': 23.601215362548828, 'timing/data': 0.010748963803052902, 'timing/model': 0.9659433076158166, 'learning_rate/qwen_vl_interface': 5.815282048531633e-06, 'learning_rate/action_model': 5.815282048531633e-05, 'epoch': 1.49}) +[2026-06-17 21:54:32,136][starVLA.training.train_starvla][INFO] - Step 3289, Loss: {'train/loss': 0.27373960614204407, 'train/grad_norm_pre_clip': 16.707962036132812, 'timing/data': 0.01016466785222292, 'timing/model': 0.9439996015280485, 'learning_rate/qwen_vl_interface': 5.81313466869771e-06, 'learning_rate/action_model': 5.8131346686977096e-05, 'epoch': 1.49}) +[2026-06-17 21:54:36,078][starVLA.training.train_starvla][INFO] - Step 3290, Loss: {'train/loss': 0.4305361807346344, 'train/grad_norm_pre_clip': 22.101818084716797, 'timing/data': 0.006626513786613941, 'timing/model': 0.9422434428706765, 'learning_rate/qwen_vl_interface': 5.810987172125364e-06, 'learning_rate/action_model': 5.810987172125364e-05, 'epoch': 1.49}) +[2026-06-17 21:54:41,043][starVLA.training.train_starvla][INFO] - Step 3291, Loss: {'train/loss': 0.24524807929992676, 'train/grad_norm_pre_clip': 14.416910171508789, 'timing/data': 0.015034090727567673, 'timing/model': 1.8207955649122596, 'learning_rate/qwen_vl_interface': 5.808839559259774e-06, 'learning_rate/action_model': 5.8088395592597746e-05, 'epoch': 1.49}) +[2026-06-17 21:54:44,963][starVLA.training.train_starvla][INFO] - Step 3292, Loss: {'train/loss': 0.2943148910999298, 'train/grad_norm_pre_clip': 15.43323802947998, 'timing/data': 0.006577197462320328, 'timing/model': 1.1021404340863228, 'learning_rate/qwen_vl_interface': 5.806691830546143e-06, 'learning_rate/action_model': 5.8066918305461427e-05, 'epoch': 1.49}) +[2026-06-17 21:54:48,647][starVLA.training.train_starvla][INFO] - Step 3293, Loss: {'train/loss': 0.35603660345077515, 'train/grad_norm_pre_clip': 33.40275192260742, 'timing/data': 0.008297284133732319, 'timing/model': 0.9747538724914193, 'learning_rate/qwen_vl_interface': 5.804543986429696e-06, 'learning_rate/action_model': 5.804543986429696e-05, 'epoch': 1.49}) +[2026-06-17 21:54:52,642][starVLA.training.train_starvla][INFO] - Step 3294, Loss: {'train/loss': 0.301868736743927, 'train/grad_norm_pre_clip': 14.957135200500488, 'timing/data': 0.009000398218631744, 'timing/model': 0.9540265388786793, 'learning_rate/qwen_vl_interface': 5.802396027355683e-06, 'learning_rate/action_model': 5.8023960273556835e-05, 'epoch': 1.49}) +[2026-06-17 21:54:56,856][starVLA.training.train_starvla][INFO] - Step 3295, Loss: {'train/loss': 0.17583967745304108, 'train/grad_norm_pre_clip': 14.967348098754883, 'timing/data': 0.014591077342629433, 'timing/model': 0.9426515605300665, 'learning_rate/qwen_vl_interface': 5.800247953769381e-06, 'learning_rate/action_model': 5.800247953769381e-05, 'epoch': 1.49}) +[2026-06-17 21:55:00,985][starVLA.training.train_starvla][INFO] - Step 3296, Loss: {'train/loss': 0.34888994693756104, 'train/grad_norm_pre_clip': 17.479278564453125, 'timing/data': 0.007416036911308765, 'timing/model': 0.9702156987041235, 'learning_rate/qwen_vl_interface': 5.798099766116084e-06, 'learning_rate/action_model': 5.798099766116084e-05, 'epoch': 1.5}) +[2026-06-17 21:55:05,567][starVLA.training.train_starvla][INFO] - Step 3297, Loss: {'train/loss': 0.5317166447639465, 'train/grad_norm_pre_clip': 21.068689346313477, 'timing/data': 0.013108327984809875, 'timing/model': 0.9914551489055157, 'learning_rate/qwen_vl_interface': 5.7959514648411175e-06, 'learning_rate/action_model': 5.795951464841117e-05, 'epoch': 1.5}) +[2026-06-17 21:55:09,465][starVLA.training.train_starvla][INFO] - Step 3298, Loss: {'train/loss': 0.37321189045906067, 'train/grad_norm_pre_clip': 21.29033851623535, 'timing/data': 0.007629883475601673, 'timing/model': 1.1382857970893383, 'learning_rate/qwen_vl_interface': 5.793803050389824e-06, 'learning_rate/action_model': 5.793803050389824e-05, 'epoch': 1.5}) +[2026-06-17 21:55:13,281][starVLA.training.train_starvla][INFO] - Step 3299, Loss: {'train/loss': 0.5509675145149231, 'train/grad_norm_pre_clip': 17.185440063476562, 'timing/data': 0.009878180921077728, 'timing/model': 0.963518213480711, 'learning_rate/qwen_vl_interface': 5.7916545232075725e-06, 'learning_rate/action_model': 5.791654523207573e-05, 'epoch': 1.5}) +[2026-06-17 21:57:17,266][starVLA.training.train_starvla][INFO] - Step 3300, Loss: {'train/loss': 0.34133657813072205, 'train/grad_norm_pre_clip': 14.830479621887207, 'eval/loss': 0.34942731261253357, 'eval/latency_8/loss': 0.34942731261253357, 'eval/demon_fire_precision': 0.9438669438669439, 'eval/demon_fire_recall': 0.9872631251941597, 'eval/demon_fire_f1': 0.9650774369875494, 'eval/demon_fire_rate_model': 0.9368391764051196, 'eval/demon_fire_rate_teacher': 0.8956594323873122, 'eval/demon_left_precision': 0.8342105263157895, 'eval/demon_left_recall': 0.6788008565310493, 'eval/demon_left_f1': 0.7485242030696576, 'eval/demon_left_rate_model': 0.10573177518085698, 'eval/demon_left_rate_teacher': 0.12993878686700056, 'eval/demon_right_precision': 0.7949852507374632, 'eval/demon_right_recall': 0.8166666666666667, 'eval/demon_right_f1': 0.8056801195814649, 'eval/demon_right_rate_model': 0.18864774624373956, 'eval/demon_right_rate_teacher': 0.18363939899833054, 'eval/demon_move_f1': 0.7771021613255613, 'eval/demon_cc_f1': 0.8997822331892488, 'eval/latency_8/demon_fire_precision': 0.9438669438669439, 'eval/latency_8/demon_fire_recall': 0.9872631251941597, 'eval/latency_8/demon_fire_f1': 0.9650774369875494, 'eval/latency_8/demon_fire_rate_model': 0.9368391764051196, 'eval/latency_8/demon_fire_rate_teacher': 0.8956594323873122, 'eval/latency_8/demon_left_precision': 0.8342105263157895, 'eval/latency_8/demon_left_recall': 0.6788008565310493, 'eval/latency_8/demon_left_f1': 0.7485242030696576, 'eval/latency_8/demon_left_rate_model': 0.10573177518085698, 'eval/latency_8/demon_left_rate_teacher': 0.12993878686700056, 'eval/latency_8/demon_right_precision': 0.7949852507374632, 'eval/latency_8/demon_right_recall': 0.8166666666666667, 'eval/latency_8/demon_right_f1': 0.8056801195814649, 'eval/latency_8/demon_right_rate_model': 0.18864774624373956, 'eval/latency_8/demon_right_rate_teacher': 0.18363939899833054, 'eval/latency_8/demon_move_f1': 0.7771021613255613, 'eval/latency_8/demon_cc_f1': 0.8997822331892488, 'timing/data': 0.008167698979377747, 'timing/model': 2.0065491879358888, 'learning_rate/qwen_vl_interface': 5.789505883739756e-06, 'learning_rate/action_model': 5.7895058837397554e-05, 'epoch': 1.5}) +[2026-06-17 21:57:17,269][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_3300_state +[2026-06-17 21:57:17,274][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 21:57:23,915][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_3300_state/model.safetensors +[2026-06-17 21:57:39,362][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_3300_state/optimizer.bin +[2026-06-17 21:57:39,363][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_3300_state/sampler.bin +[2026-06-17 21:57:39,364][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_3300_state/sampler_1.bin +[2026-06-17 21:57:39,366][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_3300_state/random_states_0.pkl +[2026-06-17 21:57:39,370][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3300 +[2026-06-17 21:57:39,370][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 21:57:39,381][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 21:57:43,364][starVLA.training.train_starvla][INFO] - Step 3301, Loss: {'train/loss': 0.41423478722572327, 'train/grad_norm_pre_clip': 19.6772403717041, 'timing/data': 0.011294418945908546, 'timing/model': 1.128390566445887, 'learning_rate/qwen_vl_interface': 5.78735713243179e-06, 'learning_rate/action_model': 5.78735713243179e-05, 'epoch': 1.5}) +[2026-06-17 21:57:47,169][starVLA.training.train_starvla][INFO] - Step 3302, Loss: {'train/loss': 0.27834179997444153, 'train/grad_norm_pre_clip': 15.20294189453125, 'timing/data': 0.009008144959807396, 'timing/model': 1.1351435938850045, 'learning_rate/qwen_vl_interface': 5.785208269729111e-06, 'learning_rate/action_model': 5.785208269729111e-05, 'epoch': 1.5}) +[2026-06-17 21:57:50,984][starVLA.training.train_starvla][INFO] - Step 3303, Loss: {'train/loss': 0.26999732851982117, 'train/grad_norm_pre_clip': 14.358428001403809, 'timing/data': 0.012194803915917873, 'timing/model': 0.9403076665475965, 'learning_rate/qwen_vl_interface': 5.783059296077183e-06, 'learning_rate/action_model': 5.7830592960771826e-05, 'epoch': 1.5}) +[2026-06-17 21:57:54,801][starVLA.training.train_starvla][INFO] - Step 3304, Loss: {'train/loss': 0.12348906695842743, 'train/grad_norm_pre_clip': 13.431856155395508, 'timing/data': 0.01039770245552063, 'timing/model': 0.9294104492291808, 'learning_rate/qwen_vl_interface': 5.780910211921488e-06, 'learning_rate/action_model': 5.7809102119214874e-05, 'epoch': 1.5}) +[2026-06-17 21:57:58,708][starVLA.training.train_starvla][INFO] - Step 3305, Loss: {'train/loss': 0.4374702572822571, 'train/grad_norm_pre_clip': 22.676122665405273, 'timing/data': 0.006845241412520409, 'timing/model': 1.1305055599659681, 'learning_rate/qwen_vl_interface': 5.778761017707536e-06, 'learning_rate/action_model': 5.778761017707536e-05, 'epoch': 1.5}) +[2026-06-17 21:58:02,548][starVLA.training.train_starvla][INFO] - Step 3306, Loss: {'train/loss': 0.35179656744003296, 'train/grad_norm_pre_clip': 16.909812927246094, 'timing/data': 0.01025393046438694, 'timing/model': 0.9208526946604252, 'learning_rate/qwen_vl_interface': 5.7766117138808545e-06, 'learning_rate/action_model': 5.776611713880854e-05, 'epoch': 1.5}) +[2026-06-17 21:58:06,337][starVLA.training.train_starvla][INFO] - Step 3307, Loss: {'train/loss': 0.24610602855682373, 'train/grad_norm_pre_clip': 23.06523323059082, 'timing/data': 0.00856741052120924, 'timing/model': 1.1215011132881045, 'learning_rate/qwen_vl_interface': 5.774462300886999e-06, 'learning_rate/action_model': 5.774462300886999e-05, 'epoch': 1.5}) +[2026-06-17 21:58:09,966][starVLA.training.train_starvla][INFO] - Step 3308, Loss: {'train/loss': 0.21307721734046936, 'train/grad_norm_pre_clip': 12.214582443237305, 'timing/data': 0.008310365490615368, 'timing/model': 0.9339422900229692, 'learning_rate/qwen_vl_interface': 5.772312779171545e-06, 'learning_rate/action_model': 5.7723127791715445e-05, 'epoch': 1.5}) +[2026-06-17 21:58:14,062][starVLA.training.train_starvla][INFO] - Step 3309, Loss: {'train/loss': 0.24404674768447876, 'train/grad_norm_pre_clip': 13.178582191467285, 'timing/data': 0.010087571106851101, 'timing/model': 0.9701465675607324, 'learning_rate/qwen_vl_interface': 5.770163149180089e-06, 'learning_rate/action_model': 5.770163149180089e-05, 'epoch': 1.5}) +[2026-06-17 21:58:17,915][starVLA.training.train_starvla][INFO] - Step 3310, Loss: {'train/loss': 0.36242279410362244, 'train/grad_norm_pre_clip': 13.925859451293945, 'timing/data': 0.010525747202336788, 'timing/model': 0.9263103250414133, 'learning_rate/qwen_vl_interface': 5.768013411358252e-06, 'learning_rate/action_model': 5.768013411358252e-05, 'epoch': 1.5}) +[2026-06-17 21:58:21,717][starVLA.training.train_starvla][INFO] - Step 3311, Loss: {'train/loss': 0.3991629481315613, 'train/grad_norm_pre_clip': 17.689027786254883, 'timing/data': 0.010582957416772842, 'timing/model': 0.9332433827221394, 'learning_rate/qwen_vl_interface': 5.765863566151677e-06, 'learning_rate/action_model': 5.7658635661516767e-05, 'epoch': 1.5}) +[2026-06-17 21:58:26,288][starVLA.training.train_starvla][INFO] - Step 3312, Loss: {'train/loss': 0.3478074073791504, 'train/grad_norm_pre_clip': 17.82089614868164, 'timing/data': 0.010250844992697239, 'timing/model': 1.7281539645045996, 'learning_rate/qwen_vl_interface': 5.763713614006033e-06, 'learning_rate/action_model': 5.763713614006032e-05, 'epoch': 1.5}) +[2026-06-17 21:58:30,143][starVLA.training.train_starvla][INFO] - Step 3313, Loss: {'train/loss': 0.2439449578523636, 'train/grad_norm_pre_clip': 13.121691703796387, 'timing/data': 0.01119093969464302, 'timing/model': 1.1924517899751663, 'learning_rate/qwen_vl_interface': 5.761563555367e-06, 'learning_rate/action_model': 5.761563555367e-05, 'epoch': 1.5}) +[2026-06-17 21:58:33,753][starVLA.training.train_starvla][INFO] - Step 3314, Loss: {'train/loss': 0.3543337285518646, 'train/grad_norm_pre_clip': 21.31768226623535, 'timing/data': 0.009147428907454014, 'timing/model': 0.9259543744847178, 'learning_rate/qwen_vl_interface': 5.759413390680293e-06, 'learning_rate/action_model': 5.759413390680293e-05, 'epoch': 1.5}) +[2026-06-17 21:58:37,545][starVLA.training.train_starvla][INFO] - Step 3315, Loss: {'train/loss': 0.41370490193367004, 'train/grad_norm_pre_clip': 21.300846099853516, 'timing/data': 0.01386344339698553, 'timing/model': 0.9281920483335853, 'learning_rate/qwen_vl_interface': 5.757263120391643e-06, 'learning_rate/action_model': 5.757263120391643e-05, 'epoch': 1.5}) +[2026-06-17 21:58:41,648][starVLA.training.train_starvla][INFO] - Step 3316, Loss: {'train/loss': 0.49064818024635315, 'train/grad_norm_pre_clip': 17.005512237548828, 'timing/data': 0.008549848571419716, 'timing/model': 0.9466761685907841, 'learning_rate/qwen_vl_interface': 5.7551127449468005e-06, 'learning_rate/action_model': 5.7551127449468004e-05, 'epoch': 1.5}) +[2026-06-17 21:58:45,555][starVLA.training.train_starvla][INFO] - Step 3317, Loss: {'train/loss': 0.4928576946258545, 'train/grad_norm_pre_clip': 26.314849853515625, 'timing/data': 0.010539949871599674, 'timing/model': 0.933032657019794, 'learning_rate/qwen_vl_interface': 5.752962264791543e-06, 'learning_rate/action_model': 5.752962264791543e-05, 'epoch': 1.5}) +[2026-06-17 21:58:50,074][starVLA.training.train_starvla][INFO] - Step 3318, Loss: {'train/loss': 0.3113117516040802, 'train/grad_norm_pre_clip': 16.572511672973633, 'timing/data': 0.00923634972423315, 'timing/model': 0.9383812490850687, 'learning_rate/qwen_vl_interface': 5.750811680371666e-06, 'learning_rate/action_model': 5.750811680371666e-05, 'epoch': 1.51}) +[2026-06-17 21:58:54,245][starVLA.training.train_starvla][INFO] - Step 3319, Loss: {'train/loss': 0.1360706239938736, 'train/grad_norm_pre_clip': 10.622323036193848, 'timing/data': 0.008687615394592285, 'timing/model': 0.9376562461256981, 'learning_rate/qwen_vl_interface': 5.7486609921329914e-06, 'learning_rate/action_model': 5.748660992132992e-05, 'epoch': 1.51}) +[2026-06-17 21:58:58,161][starVLA.training.train_starvla][INFO] - Step 3320, Loss: {'train/loss': 0.5495041608810425, 'train/grad_norm_pre_clip': 20.937854766845703, 'timing/data': 0.008275190368294716, 'timing/model': 1.128597043454647, 'learning_rate/qwen_vl_interface': 5.746510200521354e-06, 'learning_rate/action_model': 5.746510200521355e-05, 'epoch': 1.51}) +[2026-06-17 21:59:01,810][starVLA.training.train_starvla][INFO] - Step 3321, Loss: {'train/loss': 0.20180495083332062, 'train/grad_norm_pre_clip': 17.40549659729004, 'timing/data': 0.009878436103463173, 'timing/model': 0.9656899906694889, 'learning_rate/qwen_vl_interface': 5.7443593059826206e-06, 'learning_rate/action_model': 5.744359305982621e-05, 'epoch': 1.51}) +[2026-06-17 21:59:05,607][starVLA.training.train_starvla][INFO] - Step 3322, Loss: {'train/loss': 0.3692234456539154, 'train/grad_norm_pre_clip': 15.850069999694824, 'timing/data': 0.010617446154356003, 'timing/model': 0.9404519572854042, 'learning_rate/qwen_vl_interface': 5.742208308962669e-06, 'learning_rate/action_model': 5.742208308962669e-05, 'epoch': 1.51}) +[2026-06-17 21:59:09,439][starVLA.training.train_starvla][INFO] - Step 3323, Loss: {'train/loss': 0.15125896036624908, 'train/grad_norm_pre_clip': 10.604259490966797, 'timing/data': 0.009616535156965256, 'timing/model': 0.9442274589091539, 'learning_rate/qwen_vl_interface': 5.740057209907405e-06, 'learning_rate/action_model': 5.7400572099074046e-05, 'epoch': 1.51}) +[2026-06-17 21:59:13,389][starVLA.training.train_starvla][INFO] - Step 3324, Loss: {'train/loss': 0.14334137737751007, 'train/grad_norm_pre_clip': 12.044113159179688, 'timing/data': 0.00788280088454485, 'timing/model': 1.0061711138114333, 'learning_rate/qwen_vl_interface': 5.7379060092627545e-06, 'learning_rate/action_model': 5.7379060092627544e-05, 'epoch': 1.51}) +[2026-06-17 21:59:17,183][starVLA.training.train_starvla][INFO] - Step 3325, Loss: {'train/loss': 0.26605382561683655, 'train/grad_norm_pre_clip': 13.905217170715332, 'timing/data': 0.01277914084494114, 'timing/model': 0.9319745767861605, 'learning_rate/qwen_vl_interface': 5.735754707474663e-06, 'learning_rate/action_model': 5.735754707474663e-05, 'epoch': 1.51}) +[2026-06-17 21:59:21,001][starVLA.training.train_starvla][INFO] - Step 3326, Loss: {'train/loss': 0.17681007087230682, 'train/grad_norm_pre_clip': 12.960041046142578, 'timing/data': 0.008718322031199932, 'timing/model': 1.1116880681365728, 'learning_rate/qwen_vl_interface': 5.7336033049891e-06, 'learning_rate/action_model': 5.733603304989099e-05, 'epoch': 1.51}) +[2026-06-17 21:59:24,878][starVLA.training.train_starvla][INFO] - Step 3327, Loss: {'train/loss': 0.3358052670955658, 'train/grad_norm_pre_clip': 16.781360626220703, 'timing/data': 0.008980576880276203, 'timing/model': 1.1616700934246182, 'learning_rate/qwen_vl_interface': 5.731451802252049e-06, 'learning_rate/action_model': 5.731451802252048e-05, 'epoch': 1.51}) +[2026-06-17 21:59:28,726][starVLA.training.train_starvla][INFO] - Step 3328, Loss: {'train/loss': 0.3329756557941437, 'train/grad_norm_pre_clip': 14.388566970825195, 'timing/data': 0.015590555034577847, 'timing/model': 0.9699370237067342, 'learning_rate/qwen_vl_interface': 5.729300199709522e-06, 'learning_rate/action_model': 5.729300199709522e-05, 'epoch': 1.51}) +[2026-06-17 21:59:32,588][starVLA.training.train_starvla][INFO] - Step 3329, Loss: {'train/loss': 0.13708920776844025, 'train/grad_norm_pre_clip': 11.726049423217773, 'timing/data': 0.017343124374747276, 'timing/model': 0.9360787644982338, 'learning_rate/qwen_vl_interface': 5.7271484978075475e-06, 'learning_rate/action_model': 5.7271484978075475e-05, 'epoch': 1.51}) +[2026-06-17 21:59:36,371][starVLA.training.train_starvla][INFO] - Step 3330, Loss: {'train/loss': 0.2830316126346588, 'train/grad_norm_pre_clip': 19.839712142944336, 'timing/data': 0.014365469105541706, 'timing/model': 0.9327207980677485, 'learning_rate/qwen_vl_interface': 5.724996696992176e-06, 'learning_rate/action_model': 5.7249966969921764e-05, 'epoch': 1.51}) +[2026-06-17 21:59:40,203][starVLA.training.train_starvla][INFO] - Step 3331, Loss: {'train/loss': 0.2948170304298401, 'train/grad_norm_pre_clip': 21.31633949279785, 'timing/data': 0.013649721629917622, 'timing/model': 0.9276218581944704, 'learning_rate/qwen_vl_interface': 5.72284479770948e-06, 'learning_rate/action_model': 5.72284479770948e-05, 'epoch': 1.51}) +[2026-06-17 21:59:44,228][starVLA.training.train_starvla][INFO] - Step 3332, Loss: {'train/loss': 0.5326941013336182, 'train/grad_norm_pre_clip': 19.777320861816406, 'timing/data': 0.011039636097848415, 'timing/model': 0.928128344938159, 'learning_rate/qwen_vl_interface': 5.720692800405547e-06, 'learning_rate/action_model': 5.720692800405547e-05, 'epoch': 1.51}) +[2026-06-17 21:59:48,002][starVLA.training.train_starvla][INFO] - Step 3333, Loss: {'train/loss': 0.27616962790489197, 'train/grad_norm_pre_clip': 17.428808212280273, 'timing/data': 0.00952011439949274, 'timing/model': 1.0945647144690156, 'learning_rate/qwen_vl_interface': 5.71854070552649e-06, 'learning_rate/action_model': 5.7185407055264904e-05, 'epoch': 1.51}) +[2026-06-17 21:59:51,645][starVLA.training.train_starvla][INFO] - Step 3334, Loss: {'train/loss': 0.2370537668466568, 'train/grad_norm_pre_clip': 15.921817779541016, 'timing/data': 0.007748235948383808, 'timing/model': 0.939675877802074, 'learning_rate/qwen_vl_interface': 5.71638851351844e-06, 'learning_rate/action_model': 5.71638851351844e-05, 'epoch': 1.51}) +[2026-06-17 21:59:55,993][starVLA.training.train_starvla][INFO] - Step 3335, Loss: {'train/loss': 0.5254367589950562, 'train/grad_norm_pre_clip': 25.96601104736328, 'timing/data': 0.009326617233455181, 'timing/model': 0.9470335263758898, 'learning_rate/qwen_vl_interface': 5.714236224827551e-06, 'learning_rate/action_model': 5.714236224827551e-05, 'epoch': 1.51}) +[2026-06-17 21:59:59,847][starVLA.training.train_starvla][INFO] - Step 3336, Loss: {'train/loss': 0.36934879422187805, 'train/grad_norm_pre_clip': 21.17959976196289, 'timing/data': 0.009591598063707352, 'timing/model': 0.9322620686143637, 'learning_rate/qwen_vl_interface': 5.712083839899991e-06, 'learning_rate/action_model': 5.712083839899991e-05, 'epoch': 1.51}) +[2026-06-17 22:00:03,698][starVLA.training.train_starvla][INFO] - Step 3337, Loss: {'train/loss': 0.15447279810905457, 'train/grad_norm_pre_clip': 14.523677825927734, 'timing/data': 0.00975792109966278, 'timing/model': 0.9299950776621699, 'learning_rate/qwen_vl_interface': 5.709931359181956e-06, 'learning_rate/action_model': 5.709931359181956e-05, 'epoch': 1.51}) +[2026-06-17 22:00:07,613][starVLA.training.train_starvla][INFO] - Step 3338, Loss: {'train/loss': 0.2436189353466034, 'train/grad_norm_pre_clip': 14.046231269836426, 'timing/data': 0.007990170270204544, 'timing/model': 0.9449348347261548, 'learning_rate/qwen_vl_interface': 5.707778783119654e-06, 'learning_rate/action_model': 5.707778783119654e-05, 'epoch': 1.51}) +[2026-06-17 22:00:11,441][starVLA.training.train_starvla][INFO] - Step 3339, Loss: {'train/loss': 0.2700934410095215, 'train/grad_norm_pre_clip': 19.347122192382812, 'timing/data': 0.008054920472204685, 'timing/model': 1.1401453716680408, 'learning_rate/qwen_vl_interface': 5.705626112159318e-06, 'learning_rate/action_model': 5.7056261121593177e-05, 'epoch': 1.51}) +[2026-06-17 22:00:15,218][starVLA.training.train_starvla][INFO] - Step 3340, Loss: {'train/loss': 0.194008931517601, 'train/grad_norm_pre_clip': 15.197266578674316, 'timing/data': 0.009483899921178818, 'timing/model': 1.097911387681961, 'learning_rate/qwen_vl_interface': 5.703473346747199e-06, 'learning_rate/action_model': 5.703473346747199e-05, 'epoch': 1.52}) +[2026-06-17 22:00:18,815][starVLA.training.train_starvla][INFO] - Step 3341, Loss: {'train/loss': 0.2156848907470703, 'train/grad_norm_pre_clip': 16.124988555908203, 'timing/data': 0.00771101750433445, 'timing/model': 0.9280765512958169, 'learning_rate/qwen_vl_interface': 5.701320487329567e-06, 'learning_rate/action_model': 5.701320487329567e-05, 'epoch': 1.52}) +[2026-06-17 22:00:22,609][starVLA.training.train_starvla][INFO] - Step 3342, Loss: {'train/loss': 0.7231948971748352, 'train/grad_norm_pre_clip': 21.320335388183594, 'timing/data': 0.008539593778550625, 'timing/model': 0.9292256841436028, 'learning_rate/qwen_vl_interface': 5.699167534352712e-06, 'learning_rate/action_model': 5.699167534352712e-05, 'epoch': 1.52}) +[2026-06-17 22:00:26,338][starVLA.training.train_starvla][INFO] - Step 3343, Loss: {'train/loss': 0.45217105746269226, 'train/grad_norm_pre_clip': 26.77001190185547, 'timing/data': 0.012590655125677586, 'timing/model': 0.9240135364234447, 'learning_rate/qwen_vl_interface': 5.697014488262945e-06, 'learning_rate/action_model': 5.697014488262945e-05, 'epoch': 1.52}) +[2026-06-17 22:00:30,098][starVLA.training.train_starvla][INFO] - Step 3344, Loss: {'train/loss': 0.3491594195365906, 'train/grad_norm_pre_clip': 18.940610885620117, 'timing/data': 0.00992223247885704, 'timing/model': 0.9304195437580347, 'learning_rate/qwen_vl_interface': 5.694861349506591e-06, 'learning_rate/action_model': 5.6948613495065903e-05, 'epoch': 1.52}) +[2026-06-17 22:00:34,021][starVLA.training.train_starvla][INFO] - Step 3345, Loss: {'train/loss': 0.28973138332366943, 'train/grad_norm_pre_clip': 45.758506774902344, 'timing/data': 0.011634111404418945, 'timing/model': 0.9994664341211319, 'learning_rate/qwen_vl_interface': 5.692708118530003e-06, 'learning_rate/action_model': 5.6927081185300025e-05, 'epoch': 1.52}) +[2026-06-17 22:00:37,990][starVLA.training.train_starvla][INFO] - Step 3346, Loss: {'train/loss': 0.21741032600402832, 'train/grad_norm_pre_clip': 15.266443252563477, 'timing/data': 0.0076100947335362434, 'timing/model': 1.2932913042604923, 'learning_rate/qwen_vl_interface': 5.6905547957795446e-06, 'learning_rate/action_model': 5.690554795779545e-05, 'epoch': 1.52}) +[2026-06-17 22:00:41,841][starVLA.training.train_starvla][INFO] - Step 3347, Loss: {'train/loss': 0.4788128137588501, 'train/grad_norm_pre_clip': 27.369314193725586, 'timing/data': 0.0110649848356843, 'timing/model': 0.9347832286730409, 'learning_rate/qwen_vl_interface': 5.688401381701604e-06, 'learning_rate/action_model': 5.688401381701603e-05, 'epoch': 1.52}) +[2026-06-17 22:00:45,711][starVLA.training.train_starvla][INFO] - Step 3348, Loss: {'train/loss': 0.640751838684082, 'train/grad_norm_pre_clip': 28.35832977294922, 'timing/data': 0.00969746895134449, 'timing/model': 0.932790644466877, 'learning_rate/qwen_vl_interface': 5.686247876742583e-06, 'learning_rate/action_model': 5.686247876742583e-05, 'epoch': 1.52}) +[2026-06-17 22:00:49,853][starVLA.training.train_starvla][INFO] - Step 3349, Loss: {'train/loss': 0.5121109485626221, 'train/grad_norm_pre_clip': 18.27410125732422, 'timing/data': 0.015758488327264786, 'timing/model': 1.0037416769191623, 'learning_rate/qwen_vl_interface': 5.684094281348908e-06, 'learning_rate/action_model': 5.684094281348908e-05, 'epoch': 1.52}) +[2026-06-17 22:00:53,673][starVLA.training.train_starvla][INFO] - Step 3350, Loss: {'train/loss': 0.46092545986175537, 'train/grad_norm_pre_clip': 21.118438720703125, 'timing/data': 0.008110720664262772, 'timing/model': 0.9258506782352924, 'learning_rate/qwen_vl_interface': 5.681940595967023e-06, 'learning_rate/action_model': 5.681940595967024e-05, 'epoch': 1.52}) +[2026-06-17 22:00:57,681][starVLA.training.train_starvla][INFO] - Step 3351, Loss: {'train/loss': 0.5328153967857361, 'train/grad_norm_pre_clip': 35.21403121948242, 'timing/data': 0.0068480633199214935, 'timing/model': 1.1109015280380845, 'learning_rate/qwen_vl_interface': 5.679786821043388e-06, 'learning_rate/action_model': 5.679786821043388e-05, 'epoch': 1.52}) +[2026-06-17 22:01:01,482][starVLA.training.train_starvla][INFO] - Step 3352, Loss: {'train/loss': 0.2130906730890274, 'train/grad_norm_pre_clip': 21.96612548828125, 'timing/data': 0.00885777361690998, 'timing/model': 0.9265438234433532, 'learning_rate/qwen_vl_interface': 5.677632957024482e-06, 'learning_rate/action_model': 5.6776329570244826e-05, 'epoch': 1.52}) +[2026-06-17 22:01:05,284][starVLA.training.train_starvla][INFO] - Step 3353, Loss: {'train/loss': 0.15521340072155, 'train/grad_norm_pre_clip': 12.037531852722168, 'timing/data': 0.008474163711071014, 'timing/model': 1.1169755766168237, 'learning_rate/qwen_vl_interface': 5.675479004356803e-06, 'learning_rate/action_model': 5.675479004356803e-05, 'epoch': 1.52}) +[2026-06-17 22:01:08,898][starVLA.training.train_starvla][INFO] - Step 3354, Loss: {'train/loss': 0.5180948376655579, 'train/grad_norm_pre_clip': 21.95619010925293, 'timing/data': 0.009143614210188389, 'timing/model': 0.9278174145147204, 'learning_rate/qwen_vl_interface': 5.673324963486871e-06, 'learning_rate/action_model': 5.67332496348687e-05, 'epoch': 1.52}) +[2026-06-17 22:01:12,980][starVLA.training.train_starvla][INFO] - Step 3355, Loss: {'train/loss': 0.48957163095474243, 'train/grad_norm_pre_clip': 17.906543731689453, 'timing/data': 0.00873530376702547, 'timing/model': 0.9284812239930034, 'learning_rate/qwen_vl_interface': 5.671170834861215e-06, 'learning_rate/action_model': 5.671170834861215e-05, 'epoch': 1.52}) +[2026-06-17 22:01:16,829][starVLA.training.train_starvla][INFO] - Step 3356, Loss: {'train/loss': 0.4739694595336914, 'train/grad_norm_pre_clip': 27.88495635986328, 'timing/data': 0.011522592045366764, 'timing/model': 0.9387598941102624, 'learning_rate/qwen_vl_interface': 5.6690166189263925e-06, 'learning_rate/action_model': 5.669016618926393e-05, 'epoch': 1.52}) +[2026-06-17 22:01:22,129][starVLA.training.train_starvla][INFO] - Step 3357, Loss: {'train/loss': 0.5447702407836914, 'train/grad_norm_pre_clip': 28.113388061523438, 'timing/data': 0.014055823907256126, 'timing/model': 0.9589442582800984, 'learning_rate/qwen_vl_interface': 5.666862316128977e-06, 'learning_rate/action_model': 5.666862316128977e-05, 'epoch': 1.52}) +[2026-06-17 22:01:26,022][starVLA.training.train_starvla][INFO] - Step 3358, Loss: {'train/loss': 0.6514608860015869, 'train/grad_norm_pre_clip': 16.56976318359375, 'timing/data': 0.009482881054282188, 'timing/model': 0.9442473212257028, 'learning_rate/qwen_vl_interface': 5.664707926915551e-06, 'learning_rate/action_model': 5.66470792691555e-05, 'epoch': 1.52}) +[2026-06-17 22:01:29,828][starVLA.training.train_starvla][INFO] - Step 3359, Loss: {'train/loss': 0.39845365285873413, 'train/grad_norm_pre_clip': 17.99233627319336, 'timing/data': 0.011265870183706284, 'timing/model': 1.1251891804859042, 'learning_rate/qwen_vl_interface': 5.662553451732729e-06, 'learning_rate/action_model': 5.6625534517327285e-05, 'epoch': 1.52}) +[2026-06-17 22:01:33,609][starVLA.training.train_starvla][INFO] - Step 3360, Loss: {'train/loss': 0.10043345391750336, 'train/grad_norm_pre_clip': 10.083181381225586, 'timing/data': 0.00986386463046074, 'timing/model': 1.1018115133047104, 'learning_rate/qwen_vl_interface': 5.660398891027129e-06, 'learning_rate/action_model': 5.660398891027129e-05, 'epoch': 1.52}) +[2026-06-17 22:01:37,196][starVLA.training.train_starvla][INFO] - Step 3361, Loss: {'train/loss': 0.6200584173202515, 'train/grad_norm_pre_clip': 34.05319595336914, 'timing/data': 0.013310412876307964, 'timing/model': 0.9258573455736041, 'learning_rate/qwen_vl_interface': 5.658244245245399e-06, 'learning_rate/action_model': 5.658244245245399e-05, 'epoch': 1.52}) +[2026-06-17 22:01:41,010][starVLA.training.train_starvla][INFO] - Step 3362, Loss: {'train/loss': 0.15251903235912323, 'train/grad_norm_pre_clip': 12.122955322265625, 'timing/data': 0.008732701651751995, 'timing/model': 0.9267140356823802, 'learning_rate/qwen_vl_interface': 5.6560895148341985e-06, 'learning_rate/action_model': 5.656089514834199e-05, 'epoch': 1.53}) +[2026-06-17 22:01:44,866][starVLA.training.train_starvla][INFO] - Step 3363, Loss: {'train/loss': 0.7354592084884644, 'train/grad_norm_pre_clip': 17.35626220703125, 'timing/data': 0.014354974031448364, 'timing/model': 0.9344350937753916, 'learning_rate/qwen_vl_interface': 5.653934700240204e-06, 'learning_rate/action_model': 5.6539347002402034e-05, 'epoch': 1.53}) +[2026-06-17 22:01:48,626][starVLA.training.train_starvla][INFO] - Step 3364, Loss: {'train/loss': 0.23199868202209473, 'train/grad_norm_pre_clip': 23.666074752807617, 'timing/data': 0.01600172370672226, 'timing/model': 0.9212459204718471, 'learning_rate/qwen_vl_interface': 5.651779801910111e-06, 'learning_rate/action_model': 5.651779801910111e-05, 'epoch': 1.53}) +[2026-06-17 22:01:52,409][starVLA.training.train_starvla][INFO] - Step 3365, Loss: {'train/loss': 0.26176905632019043, 'train/grad_norm_pre_clip': 14.042840957641602, 'timing/data': 0.008122829720377922, 'timing/model': 0.9346761833876371, 'learning_rate/qwen_vl_interface': 5.649624820290631e-06, 'learning_rate/action_model': 5.649624820290631e-05, 'epoch': 1.53}) +[2026-06-17 22:01:56,147][starVLA.training.train_starvla][INFO] - Step 3366, Loss: {'train/loss': 0.26913610100746155, 'train/grad_norm_pre_clip': 16.62857437133789, 'timing/data': 0.014743341132998466, 'timing/model': 1.0943457521498203, 'learning_rate/qwen_vl_interface': 5.647469755828501e-06, 'learning_rate/action_model': 5.647469755828501e-05, 'epoch': 1.53}) +[2026-06-17 22:01:59,862][starVLA.training.train_starvla][INFO] - Step 3367, Loss: {'train/loss': 0.41382184624671936, 'train/grad_norm_pre_clip': 16.523731231689453, 'timing/data': 0.010978683829307556, 'timing/model': 1.0553968753665686, 'learning_rate/qwen_vl_interface': 5.645314608970457e-06, 'learning_rate/action_model': 5.6453146089704566e-05, 'epoch': 1.53}) +[2026-06-17 22:02:03,435][starVLA.training.train_starvla][INFO] - Step 3368, Loss: {'train/loss': 0.4821985363960266, 'train/grad_norm_pre_clip': 20.78961944580078, 'timing/data': 0.008974392898380756, 'timing/model': 0.9246252300217748, 'learning_rate/qwen_vl_interface': 5.643159380163273e-06, 'learning_rate/action_model': 5.643159380163273e-05, 'epoch': 1.53}) +[2026-06-17 22:02:07,192][starVLA.training.train_starvla][INFO] - Step 3369, Loss: {'train/loss': 0.255677193403244, 'train/grad_norm_pre_clip': 14.539240837097168, 'timing/data': 0.009934460744261742, 'timing/model': 0.924299294129014, 'learning_rate/qwen_vl_interface': 5.641004069853725e-06, 'learning_rate/action_model': 5.641004069853725e-05, 'epoch': 1.53}) +[2026-06-17 22:02:10,949][starVLA.training.train_starvla][INFO] - Step 3370, Loss: {'train/loss': 0.4564342498779297, 'train/grad_norm_pre_clip': 21.06542205810547, 'timing/data': 0.010604809038341045, 'timing/model': 0.9260422298684716, 'learning_rate/qwen_vl_interface': 5.638848678488613e-06, 'learning_rate/action_model': 5.638848678488613e-05, 'epoch': 1.53}) +[2026-06-17 22:02:14,643][starVLA.training.train_starvla][INFO] - Step 3371, Loss: {'train/loss': 0.27293211221694946, 'train/grad_norm_pre_clip': 19.937097549438477, 'timing/data': 0.00995468720793724, 'timing/model': 0.9244750719517469, 'learning_rate/qwen_vl_interface': 5.636693206514752e-06, 'learning_rate/action_model': 5.636693206514752e-05, 'epoch': 1.53}) +[2026-06-17 22:02:18,386][starVLA.training.train_starvla][INFO] - Step 3372, Loss: {'train/loss': 0.37856850028038025, 'train/grad_norm_pre_clip': 14.019624710083008, 'timing/data': 0.008328383788466454, 'timing/model': 0.9273484526202083, 'learning_rate/qwen_vl_interface': 5.634537654378973e-06, 'learning_rate/action_model': 5.634537654378973e-05, 'epoch': 1.53}) +[2026-06-17 22:02:22,118][starVLA.training.train_starvla][INFO] - Step 3373, Loss: {'train/loss': 0.26641377806663513, 'train/grad_norm_pre_clip': 12.914216041564941, 'timing/data': 0.01356678456068039, 'timing/model': 1.0876781987026334, 'learning_rate/qwen_vl_interface': 5.632382022528123e-06, 'learning_rate/action_model': 5.632382022528123e-05, 'epoch': 1.53}) +[2026-06-17 22:02:25,687][starVLA.training.train_starvla][INFO] - Step 3374, Loss: {'train/loss': 0.2528318762779236, 'train/grad_norm_pre_clip': 11.618949890136719, 'timing/data': 0.00977249164134264, 'timing/model': 0.921407638117671, 'learning_rate/qwen_vl_interface': 5.6302263114090705e-06, 'learning_rate/action_model': 5.63022631140907e-05, 'epoch': 1.53}) +[2026-06-17 22:02:29,444][starVLA.training.train_starvla][INFO] - Step 3375, Loss: {'train/loss': 0.2391146570444107, 'train/grad_norm_pre_clip': 17.72926139831543, 'timing/data': 0.008009974844753742, 'timing/model': 0.9250365318730474, 'learning_rate/qwen_vl_interface': 5.628070521468693e-06, 'learning_rate/action_model': 5.628070521468692e-05, 'epoch': 1.53}) +[2026-06-17 22:02:33,130][starVLA.training.train_starvla][INFO] - Step 3376, Loss: {'train/loss': 0.38881993293762207, 'train/grad_norm_pre_clip': 17.11092758178711, 'timing/data': 0.011169984936714172, 'timing/model': 0.923123306594789, 'learning_rate/qwen_vl_interface': 5.625914653153889e-06, 'learning_rate/action_model': 5.625914653153889e-05, 'epoch': 1.53}) +[2026-06-17 22:02:36,930][starVLA.training.train_starvla][INFO] - Step 3377, Loss: {'train/loss': 0.4150922894477844, 'train/grad_norm_pre_clip': 18.3956356048584, 'timing/data': 0.00911637768149376, 'timing/model': 0.9269912056624889, 'learning_rate/qwen_vl_interface': 5.623758706911574e-06, 'learning_rate/action_model': 5.6237587069115737e-05, 'epoch': 1.53}) +[2026-06-17 22:02:40,763][starVLA.training.train_starvla][INFO] - Step 3378, Loss: {'train/loss': 0.522083044052124, 'train/grad_norm_pre_clip': 19.984020233154297, 'timing/data': 0.011818823404610157, 'timing/model': 0.9388998411595821, 'learning_rate/qwen_vl_interface': 5.621602683188677e-06, 'learning_rate/action_model': 5.621602683188677e-05, 'epoch': 1.53}) +[2026-06-17 22:02:44,509][starVLA.training.train_starvla][INFO] - Step 3379, Loss: {'train/loss': 0.1867194026708603, 'train/grad_norm_pre_clip': 12.021934509277344, 'timing/data': 0.010413167998194695, 'timing/model': 1.0868721306324005, 'learning_rate/qwen_vl_interface': 5.619446582432143e-06, 'learning_rate/action_model': 5.619446582432143e-05, 'epoch': 1.53}) +[2026-06-17 22:02:48,300][starVLA.training.train_starvla][INFO] - Step 3380, Loss: {'train/loss': 0.3488650321960449, 'train/grad_norm_pre_clip': 15.334505081176758, 'timing/data': 0.012623464688658714, 'timing/model': 1.1138111762702465, 'learning_rate/qwen_vl_interface': 5.617290405088936e-06, 'learning_rate/action_model': 5.617290405088936e-05, 'epoch': 1.53}) +[2026-06-17 22:02:51,957][starVLA.training.train_starvla][INFO] - Step 3381, Loss: {'train/loss': 0.26110538840293884, 'train/grad_norm_pre_clip': 14.360306739807129, 'timing/data': 0.009365330450236797, 'timing/model': 0.927490571513772, 'learning_rate/qwen_vl_interface': 5.615134151606033e-06, 'learning_rate/action_model': 5.615134151606033e-05, 'epoch': 1.53}) +[2026-06-17 22:02:55,763][starVLA.training.train_starvla][INFO] - Step 3382, Loss: {'train/loss': 0.6210935711860657, 'train/grad_norm_pre_clip': 21.618444442749023, 'timing/data': 0.011700707487761974, 'timing/model': 0.928426013328135, 'learning_rate/qwen_vl_interface': 5.612977822430425e-06, 'learning_rate/action_model': 5.612977822430425e-05, 'epoch': 1.53}) +[2026-06-17 22:02:59,506][starVLA.training.train_starvla][INFO] - Step 3383, Loss: {'train/loss': 0.3973962366580963, 'train/grad_norm_pre_clip': 23.78399085998535, 'timing/data': 0.007794732227921486, 'timing/model': 0.9272488411515951, 'learning_rate/qwen_vl_interface': 5.610821418009127e-06, 'learning_rate/action_model': 5.610821418009127e-05, 'epoch': 1.53}) +[2026-06-17 22:03:03,245][starVLA.training.train_starvla][INFO] - Step 3384, Loss: {'train/loss': 0.3158891201019287, 'train/grad_norm_pre_clip': 14.951045989990234, 'timing/data': 0.009755689650774002, 'timing/model': 0.9280292401090264, 'learning_rate/qwen_vl_interface': 5.608664938789159e-06, 'learning_rate/action_model': 5.608664938789159e-05, 'epoch': 1.54}) +[2026-06-17 22:03:07,087][starVLA.training.train_starvla][INFO] - Step 3385, Loss: {'train/loss': 0.34228020906448364, 'train/grad_norm_pre_clip': 16.252132415771484, 'timing/data': 0.0093504274263978, 'timing/model': 0.9703574273735285, 'learning_rate/qwen_vl_interface': 5.606508385217564e-06, 'learning_rate/action_model': 5.6065083852175636e-05, 'epoch': 1.54}) +[2026-06-17 22:03:10,869][starVLA.training.train_starvla][INFO] - Step 3386, Loss: {'train/loss': 0.8737280368804932, 'train/grad_norm_pre_clip': 24.835960388183594, 'timing/data': 0.010220775380730629, 'timing/model': 1.1097597358748317, 'learning_rate/qwen_vl_interface': 5.604351757741396e-06, 'learning_rate/action_model': 5.6043517577413964e-05, 'epoch': 1.54}) +[2026-06-17 22:03:14,738][starVLA.training.train_starvla][INFO] - Step 3387, Loss: {'train/loss': 0.305502712726593, 'train/grad_norm_pre_clip': 19.401277542114258, 'timing/data': 0.009464625269174576, 'timing/model': 1.1555132856592536, 'learning_rate/qwen_vl_interface': 5.602195056807727e-06, 'learning_rate/action_model': 5.602195056807727e-05, 'epoch': 1.54}) +[2026-06-17 22:03:18,330][starVLA.training.train_starvla][INFO] - Step 3388, Loss: {'train/loss': 0.6355239152908325, 'train/grad_norm_pre_clip': 25.202560424804688, 'timing/data': 0.01449795812368393, 'timing/model': 0.9261200157925487, 'learning_rate/qwen_vl_interface': 5.600038282863647e-06, 'learning_rate/action_model': 5.6000382828636466e-05, 'epoch': 1.54}) +[2026-06-17 22:03:22,106][starVLA.training.train_starvla][INFO] - Step 3389, Loss: {'train/loss': 0.5009310841560364, 'train/grad_norm_pre_clip': 98.4647445678711, 'timing/data': 0.012252365238964558, 'timing/model': 0.9282923052087426, 'learning_rate/qwen_vl_interface': 5.597881436356252e-06, 'learning_rate/action_model': 5.597881436356252e-05, 'epoch': 1.54}) +[2026-06-17 22:03:25,895][starVLA.training.train_starvla][INFO] - Step 3390, Loss: {'train/loss': 0.3492238521575928, 'train/grad_norm_pre_clip': 24.30419158935547, 'timing/data': 0.007517979480326176, 'timing/model': 0.9438068252056837, 'learning_rate/qwen_vl_interface': 5.595724517732661e-06, 'learning_rate/action_model': 5.595724517732661e-05, 'epoch': 1.54}) +[2026-06-17 22:03:29,826][starVLA.training.train_starvla][INFO] - Step 3391, Loss: {'train/loss': 0.4987104535102844, 'train/grad_norm_pre_clip': 95.44398498535156, 'timing/data': 0.008169244043529034, 'timing/model': 0.9545066673308611, 'learning_rate/qwen_vl_interface': 5.5935675274400046e-06, 'learning_rate/action_model': 5.593567527440005e-05, 'epoch': 1.54}) +[2026-06-17 22:03:33,647][starVLA.training.train_starvla][INFO] - Step 3392, Loss: {'train/loss': 0.30070486664772034, 'train/grad_norm_pre_clip': 15.73725700378418, 'timing/data': 0.009202105924487114, 'timing/model': 0.920785010792315, 'learning_rate/qwen_vl_interface': 5.591410465925431e-06, 'learning_rate/action_model': 5.5914104659254306e-05, 'epoch': 1.54}) +[2026-06-17 22:03:37,548][starVLA.training.train_starvla][INFO] - Step 3393, Loss: {'train/loss': 0.22575125098228455, 'train/grad_norm_pre_clip': 16.73099136352539, 'timing/data': 0.008302773348987103, 'timing/model': 1.100714928470552, 'learning_rate/qwen_vl_interface': 5.589253333636101e-06, 'learning_rate/action_model': 5.589253333636101e-05, 'epoch': 1.54}) +[2026-06-17 22:03:41,351][starVLA.training.train_starvla][INFO] - Step 3394, Loss: {'train/loss': 0.3808490037918091, 'train/grad_norm_pre_clip': 16.77955436706543, 'timing/data': 0.008376294746994972, 'timing/model': 0.9300898276269436, 'learning_rate/qwen_vl_interface': 5.587096131019189e-06, 'learning_rate/action_model': 5.5870961310191896e-05, 'epoch': 1.54}) +[2026-06-17 22:03:45,163][starVLA.training.train_starvla][INFO] - Step 3395, Loss: {'train/loss': 0.20352426171302795, 'train/grad_norm_pre_clip': 15.16230583190918, 'timing/data': 0.009548826143145561, 'timing/model': 0.9296617256477475, 'learning_rate/qwen_vl_interface': 5.5849388585218865e-06, 'learning_rate/action_model': 5.5849388585218866e-05, 'epoch': 1.54}) +[2026-06-17 22:03:48,913][starVLA.training.train_starvla][INFO] - Step 3396, Loss: {'train/loss': 0.1958666443824768, 'train/grad_norm_pre_clip': 13.17229175567627, 'timing/data': 0.008800718002021313, 'timing/model': 0.9219390396028757, 'learning_rate/qwen_vl_interface': 5.582781516591398e-06, 'learning_rate/action_model': 5.5827815165913977e-05, 'epoch': 1.54}) +[2026-06-17 22:03:52,682][starVLA.training.train_starvla][INFO] - Step 3397, Loss: {'train/loss': 0.2421945482492447, 'train/grad_norm_pre_clip': 13.644503593444824, 'timing/data': 0.009160825982689857, 'timing/model': 0.9292360125109553, 'learning_rate/qwen_vl_interface': 5.580624105674945e-06, 'learning_rate/action_model': 5.580624105674944e-05, 'epoch': 1.54}) +[2026-06-17 22:03:56,503][starVLA.training.train_starvla][INFO] - Step 3398, Loss: {'train/loss': 0.38867831230163574, 'train/grad_norm_pre_clip': 24.033952713012695, 'timing/data': 0.00993906520307064, 'timing/model': 0.9635757775977254, 'learning_rate/qwen_vl_interface': 5.578466626219756e-06, 'learning_rate/action_model': 5.578466626219756e-05, 'epoch': 1.54}) +[2026-06-17 22:04:00,287][starVLA.training.train_starvla][INFO] - Step 3399, Loss: {'train/loss': 0.3904409408569336, 'train/grad_norm_pre_clip': 21.714799880981445, 'timing/data': 0.007519479840993881, 'timing/model': 0.9380566421896219, 'learning_rate/qwen_vl_interface': 5.576309078673083e-06, 'learning_rate/action_model': 5.576309078673083e-05, 'epoch': 1.54}) +[2026-06-17 22:06:13,075][starVLA.training.train_starvla][INFO] - Step 3400, Loss: {'train/loss': 0.31392452120780945, 'train/grad_norm_pre_clip': 20.64044952392578, 'eval/loss': 0.3504974842071533, 'eval/latency_8/loss': 0.3504974842071533, 'eval/demon_fire_precision': 0.9596799015081564, 'eval/demon_fire_recall': 0.968623796210003, 'eval/demon_fire_f1': 0.9641311069882499, 'eval/demon_fire_rate_model': 0.9040066777963273, 'eval/demon_fire_rate_teacher': 0.8956594323873122, 'eval/demon_left_precision': 0.785234899328859, 'eval/demon_left_recall': 0.7516059957173448, 'eval/demon_left_f1': 0.7680525164113786, 'eval/demon_left_rate_model': 0.12437395659432388, 'eval/demon_left_rate_teacher': 0.12993878686700056, 'eval/demon_right_precision': 0.7776162790697675, 'eval/demon_right_recall': 0.8106060606060606, 'eval/demon_right_f1': 0.7937685459940653, 'eval/demon_right_rate_model': 0.1914301613800779, 'eval/demon_right_rate_teacher': 0.18363939899833054, 'eval/demon_move_f1': 0.780910531202722, 'eval/demon_cc_f1': 0.9007310903149244, 'eval/latency_8/demon_fire_precision': 0.9596799015081564, 'eval/latency_8/demon_fire_recall': 0.968623796210003, 'eval/latency_8/demon_fire_f1': 0.9641311069882499, 'eval/latency_8/demon_fire_rate_model': 0.9040066777963273, 'eval/latency_8/demon_fire_rate_teacher': 0.8956594323873122, 'eval/latency_8/demon_left_precision': 0.785234899328859, 'eval/latency_8/demon_left_recall': 0.7516059957173448, 'eval/latency_8/demon_left_f1': 0.7680525164113786, 'eval/latency_8/demon_left_rate_model': 0.12437395659432388, 'eval/latency_8/demon_left_rate_teacher': 0.12993878686700056, 'eval/latency_8/demon_right_precision': 0.7776162790697675, 'eval/latency_8/demon_right_recall': 0.8106060606060606, 'eval/latency_8/demon_right_f1': 0.7937685459940653, 'eval/latency_8/demon_right_rate_model': 0.1914301613800779, 'eval/latency_8/demon_right_rate_teacher': 0.18363939899833054, 'eval/latency_8/demon_move_f1': 0.780910531202722, 'eval/latency_8/demon_cc_f1': 0.9007310903149244, 'timing/data': 0.008942272514104843, 'timing/model': 1.2359908698126674, 'learning_rate/qwen_vl_interface': 5.5741514634821885e-06, 'learning_rate/action_model': 5.574151463482188e-05, 'epoch': 1.54}) +[2026-06-17 22:06:13,078][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_3400_state +[2026-06-17 22:06:13,083][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 22:06:19,702][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_3400_state/model.safetensors +[2026-06-17 22:06:35,294][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_3400_state/optimizer.bin +[2026-06-17 22:06:35,295][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_3400_state/sampler.bin +[2026-06-17 22:06:35,295][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_3400_state/sampler_1.bin +[2026-06-17 22:06:35,297][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_3400_state/random_states_0.pkl +[2026-06-17 22:06:35,300][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3400 +[2026-06-17 22:06:35,300][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 22:06:35,312][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 22:06:39,139][starVLA.training.train_starvla][INFO] - Step 3401, Loss: {'train/loss': 0.3597593307495117, 'train/grad_norm_pre_clip': 25.36095428466797, 'timing/data': 0.009853449650108814, 'timing/model': 0.9257620852440596, 'learning_rate/qwen_vl_interface': 5.571993781094345e-06, 'learning_rate/action_model': 5.5719937810943444e-05, 'epoch': 1.54}) +[2026-06-17 22:06:42,949][starVLA.training.train_starvla][INFO] - Step 3402, Loss: {'train/loss': 0.25246816873550415, 'train/grad_norm_pre_clip': 15.268746376037598, 'timing/data': 0.01791535783559084, 'timing/model': 1.098222404718399, 'learning_rate/qwen_vl_interface': 5.569836031956845e-06, 'learning_rate/action_model': 5.5698360319568454e-05, 'epoch': 1.54}) +[2026-06-17 22:06:46,623][starVLA.training.train_starvla][INFO] - Step 3403, Loss: {'train/loss': 0.29638224840164185, 'train/grad_norm_pre_clip': 23.45789909362793, 'timing/data': 0.02786330319941044, 'timing/model': 0.9307608185335994, 'learning_rate/qwen_vl_interface': 5.5676782165169915e-06, 'learning_rate/action_model': 5.567678216516991e-05, 'epoch': 1.54}) +[2026-06-17 22:06:50,535][starVLA.training.train_starvla][INFO] - Step 3404, Loss: {'train/loss': 0.49708646535873413, 'train/grad_norm_pre_clip': 21.01568603515625, 'timing/data': 0.028261818923056126, 'timing/model': 0.9779821811243892, 'learning_rate/qwen_vl_interface': 5.565520335222101e-06, 'learning_rate/action_model': 5.565520335222101e-05, 'epoch': 1.54}) +[2026-06-17 22:06:54,424][starVLA.training.train_starvla][INFO] - Step 3405, Loss: {'train/loss': 0.33251726627349854, 'train/grad_norm_pre_clip': 20.709280014038086, 'timing/data': 0.0342820156365633, 'timing/model': 0.954592571593821, 'learning_rate/qwen_vl_interface': 5.563362388519507e-06, 'learning_rate/action_model': 5.5633623885195066e-05, 'epoch': 1.54}) +[2026-06-17 22:06:58,273][starVLA.training.train_starvla][INFO] - Step 3406, Loss: {'train/loss': 0.3535347580909729, 'train/grad_norm_pre_clip': 17.660551071166992, 'timing/data': 0.016458779573440552, 'timing/model': 0.9741429202258587, 'learning_rate/qwen_vl_interface': 5.56120437685655e-06, 'learning_rate/action_model': 5.56120437685655e-05, 'epoch': 1.55}) +[2026-06-17 22:07:02,224][starVLA.training.train_starvla][INFO] - Step 3407, Loss: {'train/loss': 0.4029162526130676, 'train/grad_norm_pre_clip': 17.782258987426758, 'timing/data': 0.016714217141270638, 'timing/model': 0.9563300292938948, 'learning_rate/qwen_vl_interface': 5.559046300680593e-06, 'learning_rate/action_model': 5.5590463006805925e-05, 'epoch': 1.55}) +[2026-06-17 22:07:06,130][starVLA.training.train_starvla][INFO] - Step 3408, Loss: {'train/loss': 0.36968034505844116, 'train/grad_norm_pre_clip': 20.572101593017578, 'timing/data': 0.015228296630084515, 'timing/model': 1.1456501390784979, 'learning_rate/qwen_vl_interface': 5.556888160439006e-06, 'learning_rate/action_model': 5.556888160439006e-05, 'epoch': 1.55}) +[2026-06-17 22:07:09,978][starVLA.training.train_starvla][INFO] - Step 3409, Loss: {'train/loss': 0.37030163407325745, 'train/grad_norm_pre_clip': 18.644262313842773, 'timing/data': 0.025349967181682587, 'timing/model': 1.0948216207325459, 'learning_rate/qwen_vl_interface': 5.5547299565791725e-06, 'learning_rate/action_model': 5.5547299565791723e-05, 'epoch': 1.55}) +[2026-06-17 22:07:13,642][starVLA.training.train_starvla][INFO] - Step 3410, Loss: {'train/loss': 0.18223708868026733, 'train/grad_norm_pre_clip': 10.725305557250977, 'timing/data': 0.017482220195233822, 'timing/model': 0.9492034241557121, 'learning_rate/qwen_vl_interface': 5.55257168954849e-06, 'learning_rate/action_model': 5.5525716895484905e-05, 'epoch': 1.55}) +[2026-06-17 22:07:17,458][starVLA.training.train_starvla][INFO] - Step 3411, Loss: {'train/loss': 0.1302664577960968, 'train/grad_norm_pre_clip': 9.110773086547852, 'timing/data': 0.029881438240408897, 'timing/model': 0.945581910200417, 'learning_rate/qwen_vl_interface': 5.550413359794371e-06, 'learning_rate/action_model': 5.550413359794371e-05, 'epoch': 1.55}) +[2026-06-17 22:07:21,295][starVLA.training.train_starvla][INFO] - Step 3412, Loss: {'train/loss': 0.563989520072937, 'train/grad_norm_pre_clip': 23.7524471282959, 'timing/data': 0.01985839754343033, 'timing/model': 0.9372120024636388, 'learning_rate/qwen_vl_interface': 5.548254967764239e-06, 'learning_rate/action_model': 5.5482549677642394e-05, 'epoch': 1.55}) +[2026-06-17 22:07:25,380][starVLA.training.train_starvla][INFO] - Step 3413, Loss: {'train/loss': 0.3254387676715851, 'train/grad_norm_pre_clip': 17.839719772338867, 'timing/data': 0.016228006221354008, 'timing/model': 0.9378527142107487, 'learning_rate/qwen_vl_interface': 5.546096513905532e-06, 'learning_rate/action_model': 5.546096513905532e-05, 'epoch': 1.55}) +[2026-06-17 22:07:30,492][starVLA.training.train_starvla][INFO] - Step 3414, Loss: {'train/loss': 0.37504008412361145, 'train/grad_norm_pre_clip': 17.988401412963867, 'timing/data': 0.016211475245654583, 'timing/model': 0.9527754634618759, 'learning_rate/qwen_vl_interface': 5.5439379986657e-06, 'learning_rate/action_model': 5.543937998665699e-05, 'epoch': 1.55}) +[2026-06-17 22:07:35,330][starVLA.training.train_starvla][INFO] - Step 3415, Loss: {'train/loss': 0.0684717446565628, 'train/grad_norm_pre_clip': 5.931771278381348, 'timing/data': 0.024750842712819576, 'timing/model': 1.1980320308357477, 'learning_rate/qwen_vl_interface': 5.5417794224922025e-06, 'learning_rate/action_model': 5.5417794224922023e-05, 'epoch': 1.55}) +[2026-06-17 22:07:39,302][starVLA.training.train_starvla][INFO] - Step 3416, Loss: {'train/loss': 0.3794592022895813, 'train/grad_norm_pre_clip': 23.16716957092285, 'timing/data': 0.00047553982585668564, 'timing/model': 0.9520179573446512, 'learning_rate/qwen_vl_interface': 5.539620785832517e-06, 'learning_rate/action_model': 5.5396207858325166e-05, 'epoch': 1.55}) +[2026-06-17 22:07:43,130][starVLA.training.train_starvla][INFO] - Step 3417, Loss: {'train/loss': 0.3879730701446533, 'train/grad_norm_pre_clip': 18.000072479248047, 'timing/data': 0.0010146256536245346, 'timing/model': 0.9292267095297575, 'learning_rate/qwen_vl_interface': 5.537462089134133e-06, 'learning_rate/action_model': 5.537462089134132e-05, 'epoch': 1.55}) +[2026-06-17 22:07:47,892][starVLA.training.train_starvla][INFO] - Step 3418, Loss: {'train/loss': 0.3644401431083679, 'train/grad_norm_pre_clip': 17.90081787109375, 'timing/data': 0.0006154151633381844, 'timing/model': 0.9508653599768877, 'learning_rate/qwen_vl_interface': 5.535303332844546e-06, 'learning_rate/action_model': 5.5353033328445455e-05, 'epoch': 1.55}) +[2026-06-17 22:07:52,295][starVLA.training.train_starvla][INFO] - Step 3419, Loss: {'train/loss': 0.6317351460456848, 'train/grad_norm_pre_clip': 27.333646774291992, 'timing/data': 0.000498797744512558, 'timing/model': 0.9563311086967587, 'learning_rate/qwen_vl_interface': 5.533144517411273e-06, 'learning_rate/action_model': 5.533144517411273e-05, 'epoch': 1.55}) +[2026-06-17 22:07:56,152][starVLA.training.train_starvla][INFO] - Step 3420, Loss: {'train/loss': 0.588857114315033, 'train/grad_norm_pre_clip': 24.331329345703125, 'timing/data': 0.0009068707004189491, 'timing/model': 0.9526988351717591, 'learning_rate/qwen_vl_interface': 5.530985643281834e-06, 'learning_rate/action_model': 5.530985643281834e-05, 'epoch': 1.55}) +[2026-06-17 22:08:00,432][starVLA.training.train_starvla][INFO] - Step 3421, Loss: {'train/loss': 0.39267030358314514, 'train/grad_norm_pre_clip': 18.4968318939209, 'timing/data': 0.0017814934253692627, 'timing/model': 1.2017835536971688, 'learning_rate/qwen_vl_interface': 5.528826710903774e-06, 'learning_rate/action_model': 5.528826710903774e-05, 'epoch': 1.55}) +[2026-06-17 22:08:04,311][starVLA.training.train_starvla][INFO] - Step 3422, Loss: {'train/loss': 0.25089192390441895, 'train/grad_norm_pre_clip': 15.8892822265625, 'timing/data': 0.0005233278498053551, 'timing/model': 1.1336626689881086, 'learning_rate/qwen_vl_interface': 5.526667720724632e-06, 'learning_rate/action_model': 5.526667720724632e-05, 'epoch': 1.55}) +[2026-06-17 22:08:07,971][starVLA.training.train_starvla][INFO] - Step 3423, Loss: {'train/loss': 0.3701035678386688, 'train/grad_norm_pre_clip': 18.242815017700195, 'timing/data': 0.0005628792569041252, 'timing/model': 0.9496654253453016, 'learning_rate/qwen_vl_interface': 5.5245086731919764e-06, 'learning_rate/action_model': 5.524508673191976e-05, 'epoch': 1.55}) +[2026-06-17 22:08:11,997][starVLA.training.train_starvla][INFO] - Step 3424, Loss: {'train/loss': 0.2939131557941437, 'train/grad_norm_pre_clip': 17.00301170349121, 'timing/data': 0.007954449392855167, 'timing/model': 0.9762561665847898, 'learning_rate/qwen_vl_interface': 5.522349568753376e-06, 'learning_rate/action_model': 5.522349568753377e-05, 'epoch': 1.55}) +[2026-06-17 22:08:15,744][starVLA.training.train_starvla][INFO] - Step 3425, Loss: {'train/loss': 0.31522420048713684, 'train/grad_norm_pre_clip': 12.244109153747559, 'timing/data': 0.007391162216663361, 'timing/model': 0.9416238311678171, 'learning_rate/qwen_vl_interface': 5.520190407856419e-06, 'learning_rate/action_model': 5.5201904078564185e-05, 'epoch': 1.55}) +[2026-06-17 22:08:19,591][starVLA.training.train_starvla][INFO] - Step 3426, Loss: {'train/loss': 0.6233699321746826, 'train/grad_norm_pre_clip': 23.246706008911133, 'timing/data': 0.009999451227486134, 'timing/model': 0.9489010209217668, 'learning_rate/qwen_vl_interface': 5.5180311909486974e-06, 'learning_rate/action_model': 5.518031190948697e-05, 'epoch': 1.55}) +[2026-06-17 22:08:23,528][starVLA.training.train_starvla][INFO] - Step 3427, Loss: {'train/loss': 0.18067368865013123, 'train/grad_norm_pre_clip': 12.01773738861084, 'timing/data': 0.009424182586371899, 'timing/model': 1.0404452364891768, 'learning_rate/qwen_vl_interface': 5.515871918477821e-06, 'learning_rate/action_model': 5.515871918477821e-05, 'epoch': 1.55}) +[2026-06-17 22:08:27,408][starVLA.training.train_starvla][INFO] - Step 3428, Loss: {'train/loss': 0.3190346956253052, 'train/grad_norm_pre_clip': 18.373868942260742, 'timing/data': 0.008024097420275211, 'timing/model': 1.1203611977398396, 'learning_rate/qwen_vl_interface': 5.513712590891414e-06, 'learning_rate/action_model': 5.513712590891413e-05, 'epoch': 1.56}) +[2026-06-17 22:08:31,249][starVLA.training.train_starvla][INFO] - Step 3429, Loss: {'train/loss': 0.35951367020606995, 'train/grad_norm_pre_clip': 22.284387588500977, 'timing/data': 0.008944693021476269, 'timing/model': 1.15909006819129, 'learning_rate/qwen_vl_interface': 5.511553208637098e-06, 'learning_rate/action_model': 5.511553208637098e-05, 'epoch': 1.56}) +[2026-06-17 22:08:34,853][starVLA.training.train_starvla][INFO] - Step 3430, Loss: {'train/loss': 0.521258533000946, 'train/grad_norm_pre_clip': 22.658479690551758, 'timing/data': 0.009477831423282623, 'timing/model': 0.9342649914324284, 'learning_rate/qwen_vl_interface': 5.509393772162524e-06, 'learning_rate/action_model': 5.509393772162523e-05, 'epoch': 1.56}) +[2026-06-17 22:08:38,740][starVLA.training.train_starvla][INFO] - Step 3431, Loss: {'train/loss': 0.3407425880432129, 'train/grad_norm_pre_clip': 18.200008392333984, 'timing/data': 0.009097271598875523, 'timing/model': 0.9419860979542136, 'learning_rate/qwen_vl_interface': 5.50723428191534e-06, 'learning_rate/action_model': 5.50723428191534e-05, 'epoch': 1.56}) +[2026-06-17 22:08:42,542][starVLA.training.train_starvla][INFO] - Step 3432, Loss: {'train/loss': 0.6737605333328247, 'train/grad_norm_pre_clip': 22.2425479888916, 'timing/data': 0.010809027589857578, 'timing/model': 0.9477221965789795, 'learning_rate/qwen_vl_interface': 5.505074738343212e-06, 'learning_rate/action_model': 5.505074738343212e-05, 'epoch': 1.56}) +[2026-06-17 22:08:46,334][starVLA.training.train_starvla][INFO] - Step 3433, Loss: {'train/loss': 0.32848259806632996, 'train/grad_norm_pre_clip': 21.159286499023438, 'timing/data': 0.008317464962601662, 'timing/model': 0.9333682777360082, 'learning_rate/qwen_vl_interface': 5.502915141893819e-06, 'learning_rate/action_model': 5.5029151418938184e-05, 'epoch': 1.56}) +[2026-06-17 22:08:50,124][starVLA.training.train_starvla][INFO] - Step 3434, Loss: {'train/loss': 0.24929414689540863, 'train/grad_norm_pre_clip': 13.964735984802246, 'timing/data': 0.01104093436151743, 'timing/model': 0.9349248558282852, 'learning_rate/qwen_vl_interface': 5.5007554930148436e-06, 'learning_rate/action_model': 5.5007554930148434e-05, 'epoch': 1.56}) +[2026-06-17 22:08:53,927][starVLA.training.train_starvla][INFO] - Step 3435, Loss: {'train/loss': 0.4087850749492645, 'train/grad_norm_pre_clip': 20.856237411499023, 'timing/data': 0.007835897617042065, 'timing/model': 1.119575240649283, 'learning_rate/qwen_vl_interface': 5.498595792153985e-06, 'learning_rate/action_model': 5.498595792153984e-05, 'epoch': 1.56}) +[2026-06-17 22:08:57,744][starVLA.training.train_starvla][INFO] - Step 3436, Loss: {'train/loss': 0.26567304134368896, 'train/grad_norm_pre_clip': 16.48905372619629, 'timing/data': 0.00859692320227623, 'timing/model': 1.1457698307931423, 'learning_rate/qwen_vl_interface': 5.49643603975895e-06, 'learning_rate/action_model': 5.4964360397589495e-05, 'epoch': 1.56}) +[2026-06-17 22:09:01,424][starVLA.training.train_starvla][INFO] - Step 3437, Loss: {'train/loss': 0.29673728346824646, 'train/grad_norm_pre_clip': 14.650805473327637, 'timing/data': 0.010036975145339966, 'timing/model': 0.9649054137989879, 'learning_rate/qwen_vl_interface': 5.494276236277458e-06, 'learning_rate/action_model': 5.4942762362774584e-05, 'epoch': 1.56}) +[2026-06-17 22:09:05,431][starVLA.training.train_starvla][INFO] - Step 3438, Loss: {'train/loss': 0.6206226944923401, 'train/grad_norm_pre_clip': 21.951263427734375, 'timing/data': 0.007502276450395584, 'timing/model': 0.9527523284777999, 'learning_rate/qwen_vl_interface': 5.492116382157242e-06, 'learning_rate/action_model': 5.492116382157242e-05, 'epoch': 1.56}) +[2026-06-17 22:09:09,300][starVLA.training.train_starvla][INFO] - Step 3439, Loss: {'train/loss': 0.16019074618816376, 'train/grad_norm_pre_clip': 14.080419540405273, 'timing/data': 0.007556392811238766, 'timing/model': 0.9507511146366596, 'learning_rate/qwen_vl_interface': 5.489956477846038e-06, 'learning_rate/action_model': 5.489956477846038e-05, 'epoch': 1.56}) +[2026-06-17 22:09:13,241][starVLA.training.train_starvla][INFO] - Step 3440, Loss: {'train/loss': 0.0703476220369339, 'train/grad_norm_pre_clip': 5.855223655700684, 'timing/data': 0.009086629375815392, 'timing/model': 0.9343991028144956, 'learning_rate/qwen_vl_interface': 5.4877965237915984e-06, 'learning_rate/action_model': 5.4877965237915986e-05, 'epoch': 1.56}) +[2026-06-17 22:09:17,114][starVLA.training.train_starvla][INFO] - Step 3441, Loss: {'train/loss': 0.20689134299755096, 'train/grad_norm_pre_clip': 16.008405685424805, 'timing/data': 0.006877649575471878, 'timing/model': 0.9420604296028614, 'learning_rate/qwen_vl_interface': 5.485636520441684e-06, 'learning_rate/action_model': 5.4856365204416836e-05, 'epoch': 1.56}) +[2026-06-17 22:09:21,052][starVLA.training.train_starvla][INFO] - Step 3442, Loss: {'train/loss': 0.19989486038684845, 'train/grad_norm_pre_clip': 15.190535545349121, 'timing/data': 0.01337690930813551, 'timing/model': 1.1494253929704428, 'learning_rate/qwen_vl_interface': 5.483476468244063e-06, 'learning_rate/action_model': 5.483476468244063e-05, 'epoch': 1.56}) +[2026-06-17 22:09:24,903][starVLA.training.train_starvla][INFO] - Step 3443, Loss: {'train/loss': 0.3192785680294037, 'train/grad_norm_pre_clip': 15.897520065307617, 'timing/data': 0.011762506328523159, 'timing/model': 1.1297775795683265, 'learning_rate/qwen_vl_interface': 5.481316367646522e-06, 'learning_rate/action_model': 5.4813163676465216e-05, 'epoch': 1.56}) +[2026-06-17 22:09:28,538][starVLA.training.train_starvla][INFO] - Step 3444, Loss: {'train/loss': 0.34550002217292786, 'train/grad_norm_pre_clip': 17.774656295776367, 'timing/data': 0.00952371396124363, 'timing/model': 0.9475921532139182, 'learning_rate/qwen_vl_interface': 5.479156219096847e-06, 'learning_rate/action_model': 5.479156219096847e-05, 'epoch': 1.56}) +[2026-06-17 22:09:32,402][starVLA.training.train_starvla][INFO] - Step 3445, Loss: {'train/loss': 0.2509671151638031, 'train/grad_norm_pre_clip': 16.85984230041504, 'timing/data': 0.008415629155933857, 'timing/model': 0.950568008236587, 'learning_rate/qwen_vl_interface': 5.476996023042843e-06, 'learning_rate/action_model': 5.476996023042843e-05, 'epoch': 1.56}) +[2026-06-17 22:09:36,338][starVLA.training.train_starvla][INFO] - Step 3446, Loss: {'train/loss': 0.2719070017337799, 'train/grad_norm_pre_clip': 16.590593338012695, 'timing/data': 0.009572768583893776, 'timing/model': 0.9425884559750557, 'learning_rate/qwen_vl_interface': 5.474835779932317e-06, 'learning_rate/action_model': 5.474835779932317e-05, 'epoch': 1.56}) +[2026-06-17 22:09:40,233][starVLA.training.train_starvla][INFO] - Step 3447, Loss: {'train/loss': 0.6979886293411255, 'train/grad_norm_pre_clip': 22.502910614013672, 'timing/data': 0.013569329865276814, 'timing/model': 0.9496559426188469, 'learning_rate/qwen_vl_interface': 5.472675490213093e-06, 'learning_rate/action_model': 5.472675490213093e-05, 'epoch': 1.56}) +[2026-06-17 22:09:44,047][starVLA.training.train_starvla][INFO] - Step 3448, Loss: {'train/loss': 0.4442427456378937, 'train/grad_norm_pre_clip': 20.130613327026367, 'timing/data': 0.009480344131588936, 'timing/model': 0.9317910429090261, 'learning_rate/qwen_vl_interface': 5.470515154333e-06, 'learning_rate/action_model': 5.470515154333e-05, 'epoch': 1.56}) +[2026-06-17 22:09:47,936][starVLA.training.train_starvla][INFO] - Step 3449, Loss: {'train/loss': 0.44846656918525696, 'train/grad_norm_pre_clip': 21.364795684814453, 'timing/data': 0.0159213924780488, 'timing/model': 1.1569907059893012, 'learning_rate/qwen_vl_interface': 5.468354772739879e-06, 'learning_rate/action_model': 5.468354772739879e-05, 'epoch': 1.56}) +[2026-06-17 22:09:51,537][starVLA.training.train_starvla][INFO] - Step 3450, Loss: {'train/loss': 0.2972412407398224, 'train/grad_norm_pre_clip': 13.448437690734863, 'timing/data': 0.007004059851169586, 'timing/model': 0.9393427884206176, 'learning_rate/qwen_vl_interface': 5.466194345881578e-06, 'learning_rate/action_model': 5.4661943458815776e-05, 'epoch': 1.57}) +[2026-06-17 22:09:55,535][starVLA.training.train_starvla][INFO] - Step 3451, Loss: {'train/loss': 0.1616150438785553, 'train/grad_norm_pre_clip': 12.814098358154297, 'timing/data': 0.01356253121048212, 'timing/model': 0.9618457155302167, 'learning_rate/qwen_vl_interface': 5.4640338742059564e-06, 'learning_rate/action_model': 5.464033874205956e-05, 'epoch': 1.57}) +[2026-06-17 22:09:59,358][starVLA.training.train_starvla][INFO] - Step 3452, Loss: {'train/loss': 0.33507513999938965, 'train/grad_norm_pre_clip': 16.25869369506836, 'timing/data': 0.008773518726229668, 'timing/model': 0.963083777576685, 'learning_rate/qwen_vl_interface': 5.461873358160883e-06, 'learning_rate/action_model': 5.461873358160883e-05, 'epoch': 1.57}) +[2026-06-17 22:10:03,219][starVLA.training.train_starvla][INFO] - Step 3453, Loss: {'train/loss': 0.3479481041431427, 'train/grad_norm_pre_clip': 19.810422897338867, 'timing/data': 0.009155678562819958, 'timing/model': 0.945141926407814, 'learning_rate/qwen_vl_interface': 5.459712798194234e-06, 'learning_rate/action_model': 5.459712798194233e-05, 'epoch': 1.57}) +[2026-06-17 22:10:07,239][starVLA.training.train_starvla][INFO] - Step 3454, Loss: {'train/loss': 0.26895245909690857, 'train/grad_norm_pre_clip': 16.55919075012207, 'timing/data': 0.012621914967894554, 'timing/model': 0.9436589498072863, 'learning_rate/qwen_vl_interface': 5.457552194753896e-06, 'learning_rate/action_model': 5.457552194753895e-05, 'epoch': 1.57}) +[2026-06-17 22:10:11,170][starVLA.training.train_starvla][INFO] - Step 3455, Loss: {'train/loss': 0.26656192541122437, 'train/grad_norm_pre_clip': 13.751382827758789, 'timing/data': 0.008320174179971218, 'timing/model': 1.15885691344738, 'learning_rate/qwen_vl_interface': 5.455391548287764e-06, 'learning_rate/action_model': 5.455391548287764e-05, 'epoch': 1.57}) +[2026-06-17 22:10:14,842][starVLA.training.train_starvla][INFO] - Step 3456, Loss: {'train/loss': 0.395256906747818, 'train/grad_norm_pre_clip': 18.567527770996094, 'timing/data': 0.008396419696509838, 'timing/model': 0.9293134417384863, 'learning_rate/qwen_vl_interface': 5.453230859243745e-06, 'learning_rate/action_model': 5.4532308592437444e-05, 'epoch': 1.57}) +[2026-06-17 22:10:18,750][starVLA.training.train_starvla][INFO] - Step 3457, Loss: {'train/loss': 0.3655296564102173, 'train/grad_norm_pre_clip': 18.89244270324707, 'timing/data': 0.008928059600293636, 'timing/model': 0.956546668894589, 'learning_rate/qwen_vl_interface': 5.451070128069748e-06, 'learning_rate/action_model': 5.4510701280697485e-05, 'epoch': 1.57}) +[2026-06-17 22:10:22,647][starVLA.training.train_starvla][INFO] - Step 3458, Loss: {'train/loss': 0.35712578892707825, 'train/grad_norm_pre_clip': 17.095556259155273, 'timing/data': 0.008216993883252144, 'timing/model': 0.9347344068810344, 'learning_rate/qwen_vl_interface': 5.4489093552136986e-06, 'learning_rate/action_model': 5.4489093552136984e-05, 'epoch': 1.57}) +[2026-06-17 22:10:26,666][starVLA.training.train_starvla][INFO] - Step 3459, Loss: {'train/loss': 0.3141375184059143, 'train/grad_norm_pre_clip': 22.51565170288086, 'timing/data': 0.008830136619508266, 'timing/model': 0.9320886414498091, 'learning_rate/qwen_vl_interface': 5.446748541123526e-06, 'learning_rate/action_model': 5.4467485411235254e-05, 'epoch': 1.57}) +[2026-06-17 22:10:30,626][starVLA.training.train_starvla][INFO] - Step 3460, Loss: {'train/loss': 0.2194170355796814, 'train/grad_norm_pre_clip': 11.990655899047852, 'timing/data': 0.010172809474170208, 'timing/model': 0.9719555778428912, 'learning_rate/qwen_vl_interface': 5.444587686247168e-06, 'learning_rate/action_model': 5.4445876862471675e-05, 'epoch': 1.57}) +[2026-06-17 22:10:34,407][starVLA.training.train_starvla][INFO] - Step 3461, Loss: {'train/loss': 0.22305141389369965, 'train/grad_norm_pre_clip': 16.00429344177246, 'timing/data': 0.014499008655548096, 'timing/model': 0.9459616998210549, 'learning_rate/qwen_vl_interface': 5.442426791032574e-06, 'learning_rate/action_model': 5.4424267910325746e-05, 'epoch': 1.57}) +[2026-06-17 22:10:38,317][starVLA.training.train_starvla][INFO] - Step 3462, Loss: {'train/loss': 0.131414532661438, 'train/grad_norm_pre_clip': 10.071439743041992, 'timing/data': 0.008439681492745876, 'timing/model': 1.2157189771533012, 'learning_rate/qwen_vl_interface': 5.4402658559277e-06, 'learning_rate/action_model': 5.4402658559276995e-05, 'epoch': 1.57}) +[2026-06-17 22:10:41,931][starVLA.training.train_starvla][INFO] - Step 3463, Loss: {'train/loss': 0.15062645077705383, 'train/grad_norm_pre_clip': 15.563812255859375, 'timing/data': 0.017318077385425568, 'timing/model': 0.935335929505527, 'learning_rate/qwen_vl_interface': 5.438104881380507e-06, 'learning_rate/action_model': 5.438104881380507e-05, 'epoch': 1.57}) +[2026-06-17 22:10:45,800][starVLA.training.train_starvla][INFO] - Step 3464, Loss: {'train/loss': 0.428817480802536, 'train/grad_norm_pre_clip': 21.068111419677734, 'timing/data': 0.01210804469883442, 'timing/model': 0.9752925634384155, 'learning_rate/qwen_vl_interface': 5.435943867838971e-06, 'learning_rate/action_model': 5.435943867838971e-05, 'epoch': 1.57}) +[2026-06-17 22:10:49,592][starVLA.training.train_starvla][INFO] - Step 3465, Loss: {'train/loss': 0.712136447429657, 'train/grad_norm_pre_clip': 24.491668701171875, 'timing/data': 0.008484568446874619, 'timing/model': 0.9388250475749373, 'learning_rate/qwen_vl_interface': 5.433782815751072e-06, 'learning_rate/action_model': 5.433782815751072e-05, 'epoch': 1.57}) +[2026-06-17 22:10:53,390][starVLA.training.train_starvla][INFO] - Step 3466, Loss: {'train/loss': 0.3431532382965088, 'train/grad_norm_pre_clip': 20.134300231933594, 'timing/data': 0.017642000690102577, 'timing/model': 0.9427876705303788, 'learning_rate/qwen_vl_interface': 5.431621725564797e-06, 'learning_rate/action_model': 5.4316217255647974e-05, 'epoch': 1.57}) +[2026-06-17 22:10:57,262][starVLA.training.train_starvla][INFO] - Step 3467, Loss: {'train/loss': 0.2622559368610382, 'train/grad_norm_pre_clip': 23.777280807495117, 'timing/data': 0.01124516874551773, 'timing/model': 0.9355627102777362, 'learning_rate/qwen_vl_interface': 5.429460597728142e-06, 'learning_rate/action_model': 5.429460597728142e-05, 'epoch': 1.57}) +[2026-06-17 22:11:01,103][starVLA.training.train_starvla][INFO] - Step 3468, Loss: {'train/loss': 0.20624659955501556, 'train/grad_norm_pre_clip': 21.672138214111328, 'timing/data': 0.007884040474891663, 'timing/model': 1.0980699127539992, 'learning_rate/qwen_vl_interface': 5.4272994326891114e-06, 'learning_rate/action_model': 5.427299432689111e-05, 'epoch': 1.57}) +[2026-06-17 22:11:05,228][starVLA.training.train_starvla][INFO] - Step 3469, Loss: {'train/loss': 0.32577601075172424, 'train/grad_norm_pre_clip': 24.79046058654785, 'timing/data': 0.0167532442137599, 'timing/model': 1.2465027179569006, 'learning_rate/qwen_vl_interface': 5.425138230895719e-06, 'learning_rate/action_model': 5.425138230895719e-05, 'epoch': 1.57}) +[2026-06-17 22:11:08,940][starVLA.training.train_starvla][INFO] - Step 3470, Loss: {'train/loss': 0.2892068326473236, 'train/grad_norm_pre_clip': 21.42431640625, 'timing/data': 0.013358759693801403, 'timing/model': 0.9491048669442534, 'learning_rate/qwen_vl_interface': 5.422976992795984e-06, 'learning_rate/action_model': 5.422976992795984e-05, 'epoch': 1.57}) +[2026-06-17 22:11:14,048][starVLA.training.train_starvla][INFO] - Step 3471, Loss: {'train/loss': 0.5289576053619385, 'train/grad_norm_pre_clip': 26.96099281311035, 'timing/data': 0.010132129304111004, 'timing/model': 0.9426314644515514, 'learning_rate/qwen_vl_interface': 5.420815718837929e-06, 'learning_rate/action_model': 5.42081571883793e-05, 'epoch': 1.57}) +[2026-06-17 22:11:17,898][starVLA.training.train_starvla][INFO] - Step 3472, Loss: {'train/loss': 0.8039032220840454, 'train/grad_norm_pre_clip': 35.618831634521484, 'timing/data': 0.009419706650078297, 'timing/model': 0.9360675616189837, 'learning_rate/qwen_vl_interface': 5.418654409469595e-06, 'learning_rate/action_model': 5.4186544094695946e-05, 'epoch': 1.58}) +[2026-06-17 22:11:21,805][starVLA.training.train_starvla][INFO] - Step 3473, Loss: {'train/loss': 0.20954857766628265, 'train/grad_norm_pre_clip': 15.711337089538574, 'timing/data': 0.006212804466485977, 'timing/model': 0.9535514106974006, 'learning_rate/qwen_vl_interface': 5.416493065139019e-06, 'learning_rate/action_model': 5.4164930651390186e-05, 'epoch': 1.58}) +[2026-06-17 22:11:26,027][starVLA.training.train_starvla][INFO] - Step 3474, Loss: {'train/loss': 0.21131674945354462, 'train/grad_norm_pre_clip': 17.554513931274414, 'timing/data': 0.007152276113629341, 'timing/model': 1.0349906841292977, 'learning_rate/qwen_vl_interface': 5.4143316862942516e-06, 'learning_rate/action_model': 5.4143316862942514e-05, 'epoch': 1.58}) +[2026-06-17 22:11:29,941][starVLA.training.train_starvla][INFO] - Step 3475, Loss: {'train/loss': 0.3238318860530853, 'train/grad_norm_pre_clip': 21.198062896728516, 'timing/data': 0.010793136432766914, 'timing/model': 1.1874430254101753, 'learning_rate/qwen_vl_interface': 5.412170273383348e-06, 'learning_rate/action_model': 5.412170273383348e-05, 'epoch': 1.58}) +[2026-06-17 22:11:34,025][starVLA.training.train_starvla][INFO] - Step 3476, Loss: {'train/loss': 0.23757530748844147, 'train/grad_norm_pre_clip': 19.840028762817383, 'timing/data': 0.019029248505830765, 'timing/model': 1.0613608807325363, 'learning_rate/qwen_vl_interface': 5.410008826854374e-06, 'learning_rate/action_model': 5.410008826854373e-05, 'epoch': 1.58}) +[2026-06-17 22:11:37,991][starVLA.training.train_starvla][INFO] - Step 3477, Loss: {'train/loss': 0.0763811469078064, 'train/grad_norm_pre_clip': 7.412847995758057, 'timing/data': 0.011188971810042858, 'timing/model': 0.9385190354660153, 'learning_rate/qwen_vl_interface': 5.407847347155395e-06, 'learning_rate/action_model': 5.407847347155395e-05, 'epoch': 1.58}) +[2026-06-17 22:11:41,798][starVLA.training.train_starvla][INFO] - Step 3478, Loss: {'train/loss': 0.44170159101486206, 'train/grad_norm_pre_clip': 21.670106887817383, 'timing/data': 0.01083006989210844, 'timing/model': 0.9363137353211641, 'learning_rate/qwen_vl_interface': 5.4056858347344934e-06, 'learning_rate/action_model': 5.405685834734493e-05, 'epoch': 1.58}) +[2026-06-17 22:11:45,654][starVLA.training.train_starvla][INFO] - Step 3479, Loss: {'train/loss': 0.2387763261795044, 'train/grad_norm_pre_clip': 18.99003791809082, 'timing/data': 0.010130518116056919, 'timing/model': 0.9348204210400581, 'learning_rate/qwen_vl_interface': 5.403524290039751e-06, 'learning_rate/action_model': 5.403524290039751e-05, 'epoch': 1.58}) +[2026-06-17 22:11:49,465][starVLA.training.train_starvla][INFO] - Step 3480, Loss: {'train/loss': 0.5214020609855652, 'train/grad_norm_pre_clip': 22.011629104614258, 'timing/data': 0.014779392629861832, 'timing/model': 0.9267668733373284, 'learning_rate/qwen_vl_interface': 5.401362713519257e-06, 'learning_rate/action_model': 5.4013627135192565e-05, 'epoch': 1.58}) +[2026-06-17 22:11:53,417][starVLA.training.train_starvla][INFO] - Step 3481, Loss: {'train/loss': 0.9141685962677002, 'train/grad_norm_pre_clip': 29.103368759155273, 'timing/data': 0.008683925494551659, 'timing/model': 1.2517694383859634, 'learning_rate/qwen_vl_interface': 5.399201105621111e-06, 'learning_rate/action_model': 5.39920110562111e-05, 'epoch': 1.58}) +[2026-06-17 22:11:57,164][starVLA.training.train_starvla][INFO] - Step 3482, Loss: {'train/loss': 0.6597943902015686, 'train/grad_norm_pre_clip': 46.60979461669922, 'timing/data': 0.008609933778643608, 'timing/model': 1.073719535022974, 'learning_rate/qwen_vl_interface': 5.397039466793412e-06, 'learning_rate/action_model': 5.3970394667934123e-05, 'epoch': 1.58}) +[2026-06-17 22:12:00,851][starVLA.training.train_starvla][INFO] - Step 3483, Loss: {'train/loss': 0.29628339409828186, 'train/grad_norm_pre_clip': 20.06740951538086, 'timing/data': 0.00892951712012291, 'timing/model': 0.9634428629651666, 'learning_rate/qwen_vl_interface': 5.394877797484278e-06, 'learning_rate/action_model': 5.3948777974842777e-05, 'epoch': 1.58}) +[2026-06-17 22:12:04,641][starVLA.training.train_starvla][INFO] - Step 3484, Loss: {'train/loss': 0.2819528579711914, 'train/grad_norm_pre_clip': 20.17751693725586, 'timing/data': 0.0075433943420648575, 'timing/model': 0.9335544044151902, 'learning_rate/qwen_vl_interface': 5.392716098141817e-06, 'learning_rate/action_model': 5.3927160981418165e-05, 'epoch': 1.58}) +[2026-06-17 22:12:08,473][starVLA.training.train_starvla][INFO] - Step 3485, Loss: {'train/loss': 0.2081763595342636, 'train/grad_norm_pre_clip': 13.220544815063477, 'timing/data': 0.01025329902768135, 'timing/model': 0.9385468577966094, 'learning_rate/qwen_vl_interface': 5.390554369214157e-06, 'learning_rate/action_model': 5.390554369214157e-05, 'epoch': 1.58}) +[2026-06-17 22:12:12,349][starVLA.training.train_starvla][INFO] - Step 3486, Loss: {'train/loss': 0.3447783589363098, 'train/grad_norm_pre_clip': 16.730989456176758, 'timing/data': 0.009787865914404392, 'timing/model': 0.9404280912131071, 'learning_rate/qwen_vl_interface': 5.388392611149425e-06, 'learning_rate/action_model': 5.388392611149424e-05, 'epoch': 1.58}) +[2026-06-17 22:12:16,251][starVLA.training.train_starvla][INFO] - Step 3487, Loss: {'train/loss': 0.33625343441963196, 'train/grad_norm_pre_clip': 19.7271728515625, 'timing/data': 0.013307090848684311, 'timing/model': 0.9780021226033568, 'learning_rate/qwen_vl_interface': 5.386230824395754e-06, 'learning_rate/action_model': 5.386230824395754e-05, 'epoch': 1.58}) +[2026-06-17 22:12:20,109][starVLA.training.train_starvla][INFO] - Step 3488, Loss: {'train/loss': 0.6681228280067444, 'train/grad_norm_pre_clip': 28.590639114379883, 'timing/data': 0.012327069416642189, 'timing/model': 1.1398675348609686, 'learning_rate/qwen_vl_interface': 5.384069009401288e-06, 'learning_rate/action_model': 5.384069009401288e-05, 'epoch': 1.58}) +[2026-06-17 22:12:23,932][starVLA.training.train_starvla][INFO] - Step 3489, Loss: {'train/loss': 0.1647389531135559, 'train/grad_norm_pre_clip': 13.852225303649902, 'timing/data': 0.007404659874737263, 'timing/model': 1.068743398413062, 'learning_rate/qwen_vl_interface': 5.381907166614171e-06, 'learning_rate/action_model': 5.3819071666141704e-05, 'epoch': 1.58}) +[2026-06-17 22:12:27,657][starVLA.training.train_starvla][INFO] - Step 3490, Loss: {'train/loss': 0.24361498653888702, 'train/grad_norm_pre_clip': 18.212310791015625, 'timing/data': 0.012905807234346867, 'timing/model': 0.9436417976394296, 'learning_rate/qwen_vl_interface': 5.379745296482559e-06, 'learning_rate/action_model': 5.379745296482559e-05, 'epoch': 1.58}) +[2026-06-17 22:12:31,550][starVLA.training.train_starvla][INFO] - Step 3491, Loss: {'train/loss': 0.2942666709423065, 'train/grad_norm_pre_clip': 17.259267807006836, 'timing/data': 0.007417435757815838, 'timing/model': 0.9533674931153655, 'learning_rate/qwen_vl_interface': 5.377583399454605e-06, 'learning_rate/action_model': 5.3775833994546044e-05, 'epoch': 1.58}) +[2026-06-17 22:12:35,452][starVLA.training.train_starvla][INFO] - Step 3492, Loss: {'train/loss': 0.22469694912433624, 'train/grad_norm_pre_clip': 13.838268280029297, 'timing/data': 0.008588462136685848, 'timing/model': 0.94989222753793, 'learning_rate/qwen_vl_interface': 5.375421475978476e-06, 'learning_rate/action_model': 5.375421475978476e-05, 'epoch': 1.58}) +[2026-06-17 22:12:39,358][starVLA.training.train_starvla][INFO] - Step 3493, Loss: {'train/loss': 0.29754772782325745, 'train/grad_norm_pre_clip': 18.220762252807617, 'timing/data': 0.008964618667960167, 'timing/model': 0.9337192680686712, 'learning_rate/qwen_vl_interface': 5.37325952650234e-06, 'learning_rate/action_model': 5.37325952650234e-05, 'epoch': 1.58}) +[2026-06-17 22:12:43,306][starVLA.training.train_starvla][INFO] - Step 3494, Loss: {'train/loss': 0.5003787279129028, 'train/grad_norm_pre_clip': 23.50596046447754, 'timing/data': 0.01215269323438406, 'timing/model': 0.9451853791251779, 'learning_rate/qwen_vl_interface': 5.371097551474371e-06, 'learning_rate/action_model': 5.371097551474371e-05, 'epoch': 1.59}) +[2026-06-17 22:12:48,454][starVLA.training.train_starvla][INFO] - Step 3495, Loss: {'train/loss': 0.14577177166938782, 'train/grad_norm_pre_clip': 13.890769004821777, 'timing/data': 0.012313133105635643, 'timing/model': 1.208730787038803, 'learning_rate/qwen_vl_interface': 5.36893555134275e-06, 'learning_rate/action_model': 5.3689355513427494e-05, 'epoch': 1.59}) +[2026-06-17 22:12:52,111][starVLA.training.train_starvla][INFO] - Step 3496, Loss: {'train/loss': 0.21712298691272736, 'train/grad_norm_pre_clip': 15.06553840637207, 'timing/data': 0.013304571621119976, 'timing/model': 0.9268993176519871, 'learning_rate/qwen_vl_interface': 5.366773526555661e-06, 'learning_rate/action_model': 5.366773526555661e-05, 'epoch': 1.59}) +[2026-06-17 22:12:55,954][starVLA.training.train_starvla][INFO] - Step 3497, Loss: {'train/loss': 0.28696247935295105, 'train/grad_norm_pre_clip': 20.707326889038086, 'timing/data': 0.014235888607800007, 'timing/model': 0.9367027096450329, 'learning_rate/qwen_vl_interface': 5.364611477561294e-06, 'learning_rate/action_model': 5.3646114775612945e-05, 'epoch': 1.59}) +[2026-06-17 22:12:59,784][starVLA.training.train_starvla][INFO] - Step 3498, Loss: {'train/loss': 0.38547664880752563, 'train/grad_norm_pre_clip': 24.275495529174805, 'timing/data': 0.01270736288279295, 'timing/model': 0.9298709919676185, 'learning_rate/qwen_vl_interface': 5.362449404807845e-06, 'learning_rate/action_model': 5.362449404807844e-05, 'epoch': 1.59}) +[2026-06-17 22:13:03,585][starVLA.training.train_starvla][INFO] - Step 3499, Loss: {'train/loss': 0.2893122136592865, 'train/grad_norm_pre_clip': 22.70732879638672, 'timing/data': 0.010625911876559258, 'timing/model': 0.9361915169283748, 'learning_rate/qwen_vl_interface': 5.360287308743512e-06, 'learning_rate/action_model': 5.3602873087435126e-05, 'epoch': 1.59}) +[2026-06-17 22:15:35,285][starVLA.training.train_starvla][INFO] - Step 3500, Loss: {'train/loss': 0.4587576389312744, 'train/grad_norm_pre_clip': 28.530582427978516, 'eval/loss': 0.34005340933799744, 'eval/latency_8/loss': 0.34005340933799744, 'eval/demon_fire_precision': 0.9541868932038835, 'eval/demon_fire_recall': 0.9770114942528736, 'eval/demon_fire_f1': 0.9654643131235611, 'eval/demon_fire_rate_model': 0.9170840289371174, 'eval/demon_fire_rate_teacher': 0.8956594323873122, 'eval/demon_left_precision': 0.8898809523809523, 'eval/demon_left_recall': 0.6402569593147751, 'eval/demon_left_f1': 0.7447073474470735, 'eval/demon_left_rate_model': 0.09348914858096828, 'eval/demon_left_rate_teacher': 0.12993878686700056, 'eval/demon_right_precision': 0.7840909090909091, 'eval/demon_right_recall': 0.8363636363636363, 'eval/demon_right_f1': 0.8093841642228738, 'eval/demon_right_rate_model': 0.19588202559821927, 'eval/demon_right_rate_teacher': 0.18363939899833054, 'eval/demon_move_f1': 0.7770457558349737, 'eval/demon_cc_f1': 0.8999949088980314, 'eval/latency_8/demon_fire_precision': 0.9541868932038835, 'eval/latency_8/demon_fire_recall': 0.9770114942528736, 'eval/latency_8/demon_fire_f1': 0.9654643131235611, 'eval/latency_8/demon_fire_rate_model': 0.9170840289371174, 'eval/latency_8/demon_fire_rate_teacher': 0.8956594323873122, 'eval/latency_8/demon_left_precision': 0.8898809523809523, 'eval/latency_8/demon_left_recall': 0.6402569593147751, 'eval/latency_8/demon_left_f1': 0.7447073474470735, 'eval/latency_8/demon_left_rate_model': 0.09348914858096828, 'eval/latency_8/demon_left_rate_teacher': 0.12993878686700056, 'eval/latency_8/demon_right_precision': 0.7840909090909091, 'eval/latency_8/demon_right_recall': 0.8363636363636363, 'eval/latency_8/demon_right_f1': 0.8093841642228738, 'eval/latency_8/demon_right_rate_model': 0.19588202559821927, 'eval/latency_8/demon_right_rate_teacher': 0.18363939899833054, 'eval/latency_8/demon_move_f1': 0.7770457558349737, 'eval/latency_8/demon_cc_f1': 0.8999949088980314, 'timing/data': 0.008727067150175571, 'timing/model': 0.9446978205814958, 'learning_rate/qwen_vl_interface': 5.358125189816502e-06, 'learning_rate/action_model': 5.358125189816502e-05, 'epoch': 1.59}) +[2026-06-17 22:15:35,288][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_3500_state +[2026-06-17 22:15:35,294][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 22:15:41,820][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_3500_state/model.safetensors +[2026-06-17 22:15:57,179][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_3500_state/optimizer.bin +[2026-06-17 22:15:57,180][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_3500_state/sampler.bin +[2026-06-17 22:15:57,180][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_3500_state/sampler_1.bin +[2026-06-17 22:15:57,182][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_3500_state/random_states_0.pkl +[2026-06-17 22:15:57,186][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3500 +[2026-06-17 22:15:57,186][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 22:15:57,197][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 22:16:01,034][starVLA.training.train_starvla][INFO] - Step 3501, Loss: {'train/loss': 0.6023451089859009, 'train/grad_norm_pre_clip': 23.77626609802246, 'timing/data': 0.011272886767983437, 'timing/model': 0.9361360827460885, 'learning_rate/qwen_vl_interface': 5.355963048475024e-06, 'learning_rate/action_model': 5.3559630484750244e-05, 'epoch': 1.59}) +[2026-06-17 22:16:04,794][starVLA.training.train_starvla][INFO] - Step 3502, Loss: {'train/loss': 0.4434801936149597, 'train/grad_norm_pre_clip': 25.968791961669922, 'timing/data': 0.009305805899202824, 'timing/model': 0.9309202581644058, 'learning_rate/qwen_vl_interface': 5.3538008851672916e-06, 'learning_rate/action_model': 5.3538008851672916e-05, 'epoch': 1.59}) +[2026-06-17 22:16:08,547][starVLA.training.train_starvla][INFO] - Step 3503, Loss: {'train/loss': 0.2384372055530548, 'train/grad_norm_pre_clip': 18.035011291503906, 'timing/data': 0.0074493736028671265, 'timing/model': 0.9451936911791563, 'learning_rate/qwen_vl_interface': 5.351638700341525e-06, 'learning_rate/action_model': 5.351638700341525e-05, 'epoch': 1.59}) +[2026-06-17 22:16:13,562][starVLA.training.train_starvla][INFO] - Step 3504, Loss: {'train/loss': 0.23033490777015686, 'train/grad_norm_pre_clip': 25.962331771850586, 'timing/data': 0.0123915309086442, 'timing/model': 2.135786642320454, 'learning_rate/qwen_vl_interface': 5.349476494445947e-06, 'learning_rate/action_model': 5.3494764944459466e-05, 'epoch': 1.59}) +[2026-06-17 22:16:17,227][starVLA.training.train_starvla][INFO] - Step 3505, Loss: {'train/loss': 0.647956132888794, 'train/grad_norm_pre_clip': 25.648422241210938, 'timing/data': 0.011657928116619587, 'timing/model': 0.9969544112682343, 'learning_rate/qwen_vl_interface': 5.3473142679287824e-06, 'learning_rate/action_model': 5.3473142679287826e-05, 'epoch': 1.59}) +[2026-06-17 22:16:21,041][starVLA.training.train_starvla][INFO] - Step 3506, Loss: {'train/loss': 0.2548537254333496, 'train/grad_norm_pre_clip': 16.129751205444336, 'timing/data': 0.014885025098919868, 'timing/model': 0.9275369560346007, 'learning_rate/qwen_vl_interface': 5.345152021238268e-06, 'learning_rate/action_model': 5.345152021238268e-05, 'epoch': 1.59}) +[2026-06-17 22:16:25,676][starVLA.training.train_starvla][INFO] - Step 3507, Loss: {'train/loss': 0.5143944025039673, 'train/grad_norm_pre_clip': 19.59326934814453, 'timing/data': 0.009609168395400047, 'timing/model': 0.9343587337061763, 'learning_rate/qwen_vl_interface': 5.342989754822635e-06, 'learning_rate/action_model': 5.342989754822635e-05, 'epoch': 1.59}) +[2026-06-17 22:16:29,553][starVLA.training.train_starvla][INFO] - Step 3508, Loss: {'train/loss': 0.12200190871953964, 'train/grad_norm_pre_clip': 17.118696212768555, 'timing/data': 0.009443055838346481, 'timing/model': 0.923362085595727, 'learning_rate/qwen_vl_interface': 5.340827469130126e-06, 'learning_rate/action_model': 5.3408274691301255e-05, 'epoch': 1.59}) +[2026-06-17 22:16:33,388][starVLA.training.train_starvla][INFO] - Step 3509, Loss: {'train/loss': 0.38323649764060974, 'train/grad_norm_pre_clip': 15.019915580749512, 'timing/data': 0.008494589477777481, 'timing/model': 0.9284612126648426, 'learning_rate/qwen_vl_interface': 5.338665164608985e-06, 'learning_rate/action_model': 5.3386651646089845e-05, 'epoch': 1.59}) +[2026-06-17 22:16:39,158][starVLA.training.train_starvla][INFO] - Step 3510, Loss: {'train/loss': 0.17255805432796478, 'train/grad_norm_pre_clip': 13.840868949890137, 'timing/data': 0.012139621190726757, 'timing/model': 3.0704140039160848, 'learning_rate/qwen_vl_interface': 5.336502841707461e-06, 'learning_rate/action_model': 5.33650284170746e-05, 'epoch': 1.59}) +[2026-06-17 22:16:43,091][starVLA.training.train_starvla][INFO] - Step 3511, Loss: {'train/loss': 0.2295544296503067, 'train/grad_norm_pre_clip': 38.638919830322266, 'timing/data': 0.010382554493844509, 'timing/model': 0.9316843682900071, 'learning_rate/qwen_vl_interface': 5.334340500873803e-06, 'learning_rate/action_model': 5.3343405008738034e-05, 'epoch': 1.59}) +[2026-06-17 22:16:47,041][starVLA.training.train_starvla][INFO] - Step 3512, Loss: {'train/loss': 0.14330348372459412, 'train/grad_norm_pre_clip': 13.150323867797852, 'timing/data': 0.011181008070707321, 'timing/model': 1.0712253274396062, 'learning_rate/qwen_vl_interface': 5.33217814255627e-06, 'learning_rate/action_model': 5.33217814255627e-05, 'epoch': 1.59}) +[2026-06-17 22:16:50,861][starVLA.training.train_starvla][INFO] - Step 3513, Loss: {'train/loss': 0.3787483870983124, 'train/grad_norm_pre_clip': 22.155902862548828, 'timing/data': 0.008662736043334007, 'timing/model': 0.9270579060539603, 'learning_rate/qwen_vl_interface': 5.330015767203118e-06, 'learning_rate/action_model': 5.330015767203118e-05, 'epoch': 1.59}) +[2026-06-17 22:16:54,784][starVLA.training.train_starvla][INFO] - Step 3514, Loss: {'train/loss': 0.09986013174057007, 'train/grad_norm_pre_clip': 8.688258171081543, 'timing/data': 0.010255697183310986, 'timing/model': 0.924726590514183, 'learning_rate/qwen_vl_interface': 5.3278533752626135e-06, 'learning_rate/action_model': 5.3278533752626136e-05, 'epoch': 1.59}) +[2026-06-17 22:16:58,558][starVLA.training.train_starvla][INFO] - Step 3515, Loss: {'train/loss': 0.2801954448223114, 'train/grad_norm_pre_clip': 20.359647750854492, 'timing/data': 0.011030596680939198, 'timing/model': 0.9268514402210712, 'learning_rate/qwen_vl_interface': 5.325690967183019e-06, 'learning_rate/action_model': 5.325690967183019e-05, 'epoch': 1.59}) +[2026-06-17 22:17:02,310][starVLA.training.train_starvla][INFO] - Step 3516, Loss: {'train/loss': 0.27483516931533813, 'train/grad_norm_pre_clip': 16.734800338745117, 'timing/data': 0.007680350914597511, 'timing/model': 1.087232076562941, 'learning_rate/qwen_vl_interface': 5.323528543412607e-06, 'learning_rate/action_model': 5.323528543412607e-05, 'epoch': 1.6}) +[2026-06-17 22:17:06,077][starVLA.training.train_starvla][INFO] - Step 3517, Loss: {'train/loss': 0.9687514305114746, 'train/grad_norm_pre_clip': 29.261188507080078, 'timing/data': 0.00847968552261591, 'timing/model': 1.081606786698103, 'learning_rate/qwen_vl_interface': 5.321366104399648e-06, 'learning_rate/action_model': 5.321366104399648e-05, 'epoch': 1.6}) +[2026-06-17 22:17:10,258][starVLA.training.train_starvla][INFO] - Step 3518, Loss: {'train/loss': 0.2790371775627136, 'train/grad_norm_pre_clip': 16.919530868530273, 'timing/data': 0.008132404647767544, 'timing/model': 0.9574722275137901, 'learning_rate/qwen_vl_interface': 5.319203650592419e-06, 'learning_rate/action_model': 5.3192036505924194e-05, 'epoch': 1.6}) +[2026-06-17 22:17:15,071][starVLA.training.train_starvla][INFO] - Step 3519, Loss: {'train/loss': 0.36503493785858154, 'train/grad_norm_pre_clip': 24.225069046020508, 'timing/data': 0.010694928467273712, 'timing/model': 1.0038210591301322, 'learning_rate/qwen_vl_interface': 5.317041182439199e-06, 'learning_rate/action_model': 5.3170411824391985e-05, 'epoch': 1.6}) +[2026-06-17 22:17:19,082][starVLA.training.train_starvla][INFO] - Step 3520, Loss: {'train/loss': 0.25946176052093506, 'train/grad_norm_pre_clip': 18.503948211669922, 'timing/data': 0.00821535475552082, 'timing/model': 0.9863695912063122, 'learning_rate/qwen_vl_interface': 5.314878700388267e-06, 'learning_rate/action_model': 5.314878700388267e-05, 'epoch': 1.6}) +[2026-06-17 22:17:22,899][starVLA.training.train_starvla][INFO] - Step 3521, Loss: {'train/loss': 0.10049880295991898, 'train/grad_norm_pre_clip': 9.27839469909668, 'timing/data': 0.009797838516533375, 'timing/model': 0.9394519776105881, 'learning_rate/qwen_vl_interface': 5.312716204887914e-06, 'learning_rate/action_model': 5.3127162048879145e-05, 'epoch': 1.6}) +[2026-06-17 22:17:26,768][starVLA.training.train_starvla][INFO] - Step 3522, Loss: {'train/loss': 0.6003267168998718, 'train/grad_norm_pre_clip': 24.525867462158203, 'timing/data': 0.008541734889149666, 'timing/model': 0.9937490886077285, 'learning_rate/qwen_vl_interface': 5.3105536963864226e-06, 'learning_rate/action_model': 5.310553696386422e-05, 'epoch': 1.6}) +[2026-06-17 22:17:30,523][starVLA.training.train_starvla][INFO] - Step 3523, Loss: {'train/loss': 0.19396011531352997, 'train/grad_norm_pre_clip': 17.40708351135254, 'timing/data': 0.008617223240435123, 'timing/model': 1.0968971299007535, 'learning_rate/qwen_vl_interface': 5.3083911753320875e-06, 'learning_rate/action_model': 5.308391175332087e-05, 'epoch': 1.6}) +[2026-06-17 22:17:34,318][starVLA.training.train_starvla][INFO] - Step 3524, Loss: {'train/loss': 0.22158898413181305, 'train/grad_norm_pre_clip': 17.54315948486328, 'timing/data': 0.01980031095445156, 'timing/model': 1.096479319036007, 'learning_rate/qwen_vl_interface': 5.306228642173194e-06, 'learning_rate/action_model': 5.306228642173194e-05, 'epoch': 1.6}) +[2026-06-17 22:17:37,927][starVLA.training.train_starvla][INFO] - Step 3525, Loss: {'train/loss': 0.3384607434272766, 'train/grad_norm_pre_clip': 19.701913833618164, 'timing/data': 0.006677723489701748, 'timing/model': 0.93321184348315, 'learning_rate/qwen_vl_interface': 5.304066097358046e-06, 'learning_rate/action_model': 5.304066097358046e-05, 'epoch': 1.6}) +[2026-06-17 22:17:41,753][starVLA.training.train_starvla][INFO] - Step 3526, Loss: {'train/loss': 0.37117230892181396, 'train/grad_norm_pre_clip': 25.185142517089844, 'timing/data': 0.007455989718437195, 'timing/model': 0.9666868150234222, 'learning_rate/qwen_vl_interface': 5.301903541334937e-06, 'learning_rate/action_model': 5.301903541334937e-05, 'epoch': 1.6}) +[2026-06-17 22:17:45,666][starVLA.training.train_starvla][INFO] - Step 3527, Loss: {'train/loss': 0.22218923270702362, 'train/grad_norm_pre_clip': 19.850210189819336, 'timing/data': 0.009525652974843979, 'timing/model': 1.0001030983403325, 'learning_rate/qwen_vl_interface': 5.299740974552168e-06, 'learning_rate/action_model': 5.299740974552167e-05, 'epoch': 1.6}) +[2026-06-17 22:17:49,688][starVLA.training.train_starvla][INFO] - Step 3528, Loss: {'train/loss': 0.42101556062698364, 'train/grad_norm_pre_clip': 27.374513626098633, 'timing/data': 0.008366066962480545, 'timing/model': 0.9501646533608437, 'learning_rate/qwen_vl_interface': 5.297578397458042e-06, 'learning_rate/action_model': 5.297578397458042e-05, 'epoch': 1.6}) +[2026-06-17 22:17:53,914][starVLA.training.train_starvla][INFO] - Step 3529, Loss: {'train/loss': 0.2073736935853958, 'train/grad_norm_pre_clip': 15.252650260925293, 'timing/data': 0.009898717515170574, 'timing/model': 0.947141395881772, 'learning_rate/qwen_vl_interface': 5.295415810500862e-06, 'learning_rate/action_model': 5.295415810500862e-05, 'epoch': 1.6}) +[2026-06-17 22:17:57,709][starVLA.training.train_starvla][INFO] - Step 3530, Loss: {'train/loss': 0.3752328157424927, 'train/grad_norm_pre_clip': 12.519685745239258, 'timing/data': 0.010562569834291935, 'timing/model': 1.1109835021197796, 'learning_rate/qwen_vl_interface': 5.293253214128938e-06, 'learning_rate/action_model': 5.293253214128937e-05, 'epoch': 1.6}) +[2026-06-17 22:18:01,330][starVLA.training.train_starvla][INFO] - Step 3531, Loss: {'train/loss': 0.1998664140701294, 'train/grad_norm_pre_clip': 15.659255981445312, 'timing/data': 0.009618453681468964, 'timing/model': 0.930815557949245, 'learning_rate/qwen_vl_interface': 5.291090608790572e-06, 'learning_rate/action_model': 5.291090608790572e-05, 'epoch': 1.6}) +[2026-06-17 22:18:07,294][starVLA.training.train_starvla][INFO] - Step 3532, Loss: {'train/loss': 0.2702014148235321, 'train/grad_norm_pre_clip': 15.887683868408203, 'timing/data': 0.011375542730093002, 'timing/model': 2.9907498508691788, 'learning_rate/qwen_vl_interface': 5.28892799493408e-06, 'learning_rate/action_model': 5.2889279949340796e-05, 'epoch': 1.6}) +[2026-06-17 22:18:11,137][starVLA.training.train_starvla][INFO] - Step 3533, Loss: {'train/loss': 0.22563083469867706, 'train/grad_norm_pre_clip': 14.165605545043945, 'timing/data': 0.009032927453517914, 'timing/model': 0.9449252746999264, 'learning_rate/qwen_vl_interface': 5.286765373007774e-06, 'learning_rate/action_model': 5.2867653730077735e-05, 'epoch': 1.6}) +[2026-06-17 22:18:15,299][starVLA.training.train_starvla][INFO] - Step 3534, Loss: {'train/loss': 0.34715157747268677, 'train/grad_norm_pre_clip': 16.664005279541016, 'timing/data': 0.014989441260695457, 'timing/model': 0.9447431256994605, 'learning_rate/qwen_vl_interface': 5.284602743459968e-06, 'learning_rate/action_model': 5.284602743459968e-05, 'epoch': 1.6}) +[2026-06-17 22:18:19,137][starVLA.training.train_starvla][INFO] - Step 3535, Loss: {'train/loss': 0.3265337347984314, 'train/grad_norm_pre_clip': 24.48270034790039, 'timing/data': 0.010199243202805519, 'timing/model': 0.9368524933233857, 'learning_rate/qwen_vl_interface': 5.282440106738973e-06, 'learning_rate/action_model': 5.282440106738973e-05, 'epoch': 1.6}) +[2026-06-17 22:18:22,946][starVLA.training.train_starvla][INFO] - Step 3536, Loss: {'train/loss': 0.2035878747701645, 'train/grad_norm_pre_clip': 14.885531425476074, 'timing/data': 0.008034483529627323, 'timing/model': 1.1310949670150876, 'learning_rate/qwen_vl_interface': 5.280277463293111e-06, 'learning_rate/action_model': 5.280277463293111e-05, 'epoch': 1.6}) +[2026-06-17 22:18:26,763][starVLA.training.train_starvla][INFO] - Step 3537, Loss: {'train/loss': 0.15478993952274323, 'train/grad_norm_pre_clip': 14.47674560546875, 'timing/data': 0.009335589595139027, 'timing/model': 1.1210208982229233, 'learning_rate/qwen_vl_interface': 5.2781148135706975e-06, 'learning_rate/action_model': 5.278114813570697e-05, 'epoch': 1.6}) +[2026-06-17 22:18:30,355][starVLA.training.train_starvla][INFO] - Step 3538, Loss: {'train/loss': 0.18497157096862793, 'train/grad_norm_pre_clip': 20.590423583984375, 'timing/data': 0.008512464351952076, 'timing/model': 0.9246592149138451, 'learning_rate/qwen_vl_interface': 5.275952158020054e-06, 'learning_rate/action_model': 5.275952158020053e-05, 'epoch': 1.61}) +[2026-06-17 22:18:34,145][starVLA.training.train_starvla][INFO] - Step 3539, Loss: {'train/loss': 0.11246191710233688, 'train/grad_norm_pre_clip': 9.606557846069336, 'timing/data': 0.007924527861177921, 'timing/model': 0.9247984485700727, 'learning_rate/qwen_vl_interface': 5.273789497089497e-06, 'learning_rate/action_model': 5.2737894970894974e-05, 'epoch': 1.61}) +[2026-06-17 22:18:37,906][starVLA.training.train_starvla][INFO] - Step 3540, Loss: {'train/loss': 0.390085905790329, 'train/grad_norm_pre_clip': 19.649856567382812, 'timing/data': 0.007851239293813705, 'timing/model': 0.923070820979774, 'learning_rate/qwen_vl_interface': 5.271626831227355e-06, 'learning_rate/action_model': 5.2716268312273544e-05, 'epoch': 1.61}) +[2026-06-17 22:18:41,667][starVLA.training.train_starvla][INFO] - Step 3541, Loss: {'train/loss': 0.5617212057113647, 'train/grad_norm_pre_clip': 25.43678092956543, 'timing/data': 0.006380702368915081, 'timing/model': 0.9241030560806394, 'learning_rate/qwen_vl_interface': 5.2694641608819465e-06, 'learning_rate/action_model': 5.269464160881946e-05, 'epoch': 1.61}) +[2026-06-17 22:18:45,473][starVLA.training.train_starvla][INFO] - Step 3542, Loss: {'train/loss': 0.253470778465271, 'train/grad_norm_pre_clip': 23.520973205566406, 'timing/data': 0.008840210735797882, 'timing/model': 0.924949019216001, 'learning_rate/qwen_vl_interface': 5.267301486501595e-06, 'learning_rate/action_model': 5.2673014865015945e-05, 'epoch': 1.61}) +[2026-06-17 22:18:49,221][starVLA.training.train_starvla][INFO] - Step 3543, Loss: {'train/loss': 0.37938472628593445, 'train/grad_norm_pre_clip': 23.54158592224121, 'timing/data': 0.009135519154369831, 'timing/model': 1.1053170440718532, 'learning_rate/qwen_vl_interface': 5.265138808534628e-06, 'learning_rate/action_model': 5.265138808534628e-05, 'epoch': 1.61}) +[2026-06-17 22:18:52,980][starVLA.training.train_starvla][INFO] - Step 3544, Loss: {'train/loss': 0.2367456555366516, 'train/grad_norm_pre_clip': 13.39539623260498, 'timing/data': 0.007676452398300171, 'timing/model': 1.1006203824654222, 'learning_rate/qwen_vl_interface': 5.26297612742937e-06, 'learning_rate/action_model': 5.26297612742937e-05, 'epoch': 1.61}) +[2026-06-17 22:18:56,579][starVLA.training.train_starvla][INFO] - Step 3545, Loss: {'train/loss': 0.36049550771713257, 'train/grad_norm_pre_clip': 25.289255142211914, 'timing/data': 0.0072678327560424805, 'timing/model': 0.9319344898685813, 'learning_rate/qwen_vl_interface': 5.2608134436341476e-06, 'learning_rate/action_model': 5.260813443634147e-05, 'epoch': 1.61}) +[2026-06-17 22:19:00,362][starVLA.training.train_starvla][INFO] - Step 3546, Loss: {'train/loss': 0.2619099020957947, 'train/grad_norm_pre_clip': 14.708686828613281, 'timing/data': 0.008415346965193748, 'timing/model': 0.9277779515832663, 'learning_rate/qwen_vl_interface': 5.258650757597285e-06, 'learning_rate/action_model': 5.258650757597286e-05, 'epoch': 1.61}) +[2026-06-17 22:19:04,138][starVLA.training.train_starvla][INFO] - Step 3547, Loss: {'train/loss': 0.41493478417396545, 'train/grad_norm_pre_clip': 22.708660125732422, 'timing/data': 0.01017109677195549, 'timing/model': 0.925694732926786, 'learning_rate/qwen_vl_interface': 5.256488069767113e-06, 'learning_rate/action_model': 5.256488069767113e-05, 'epoch': 1.61}) +[2026-06-17 22:19:07,916][starVLA.training.train_starvla][INFO] - Step 3548, Loss: {'train/loss': 0.38664788007736206, 'train/grad_norm_pre_clip': 18.158939361572266, 'timing/data': 0.01210560742765665, 'timing/model': 0.926695385016501, 'learning_rate/qwen_vl_interface': 5.254325380591955e-06, 'learning_rate/action_model': 5.254325380591955e-05, 'epoch': 1.61}) +[2026-06-17 22:19:11,663][starVLA.training.train_starvla][INFO] - Step 3549, Loss: {'train/loss': 0.30166348814964294, 'train/grad_norm_pre_clip': 21.924724578857422, 'timing/data': 0.00815134309232235, 'timing/model': 0.925081692636013, 'learning_rate/qwen_vl_interface': 5.252162690520142e-06, 'learning_rate/action_model': 5.252162690520142e-05, 'epoch': 1.61}) +[2026-06-17 22:19:15,470][starVLA.training.train_starvla][INFO] - Step 3550, Loss: {'train/loss': 0.3682767152786255, 'train/grad_norm_pre_clip': 26.3040828704834, 'timing/data': 0.008043629117310047, 'timing/model': 1.1076398147270083, 'learning_rate/qwen_vl_interface': 5.2500000000000006e-06, 'learning_rate/action_model': 5.25e-05, 'epoch': 1.61}) +[2026-06-17 22:19:19,049][starVLA.training.train_starvla][INFO] - Step 3551, Loss: {'train/loss': 0.22940678894519806, 'train/grad_norm_pre_clip': 19.577022552490234, 'timing/data': 0.00946976151317358, 'timing/model': 0.9250931013375521, 'learning_rate/qwen_vl_interface': 5.24783730947986e-06, 'learning_rate/action_model': 5.2478373094798594e-05, 'epoch': 1.61}) +[2026-06-17 22:19:22,802][starVLA.training.train_starvla][INFO] - Step 3552, Loss: {'train/loss': 0.2046070396900177, 'train/grad_norm_pre_clip': 18.93319320678711, 'timing/data': 0.008998367935419083, 'timing/model': 0.9235176211223006, 'learning_rate/qwen_vl_interface': 5.245674619408045e-06, 'learning_rate/action_model': 5.245674619408045e-05, 'epoch': 1.61}) +[2026-06-17 22:19:26,487][starVLA.training.train_starvla][INFO] - Step 3553, Loss: {'train/loss': 0.33922117948532104, 'train/grad_norm_pre_clip': 23.58732032775879, 'timing/data': 0.008309823460876942, 'timing/model': 0.9203286608681083, 'learning_rate/qwen_vl_interface': 5.243511930232889e-06, 'learning_rate/action_model': 5.243511930232888e-05, 'epoch': 1.61}) +[2026-06-17 22:19:30,224][starVLA.training.train_starvla][INFO] - Step 3554, Loss: {'train/loss': 0.1363300383090973, 'train/grad_norm_pre_clip': 17.670438766479492, 'timing/data': 0.00920010358095169, 'timing/model': 0.923235896974802, 'learning_rate/qwen_vl_interface': 5.2413492424027165e-06, 'learning_rate/action_model': 5.241349242402717e-05, 'epoch': 1.61}) +[2026-06-17 22:19:34,015][starVLA.training.train_starvla][INFO] - Step 3555, Loss: {'train/loss': 0.5290384292602539, 'train/grad_norm_pre_clip': 23.51468849182129, 'timing/data': 0.008976366370916367, 'timing/model': 0.9500722466036677, 'learning_rate/qwen_vl_interface': 5.239186556365855e-06, 'learning_rate/action_model': 5.239186556365855e-05, 'epoch': 1.61}) +[2026-06-17 22:19:37,772][starVLA.training.train_starvla][INFO] - Step 3556, Loss: {'train/loss': 0.283816933631897, 'train/grad_norm_pre_clip': 25.597368240356445, 'timing/data': 0.009256797842681408, 'timing/model': 1.0890233842656016, 'learning_rate/qwen_vl_interface': 5.23702387257063e-06, 'learning_rate/action_model': 5.23702387257063e-05, 'epoch': 1.61}) +[2026-06-17 22:19:41,518][starVLA.training.train_starvla][INFO] - Step 3557, Loss: {'train/loss': 0.27721959352493286, 'train/grad_norm_pre_clip': 17.924928665161133, 'timing/data': 0.01361426617950201, 'timing/model': 1.0723190726712346, 'learning_rate/qwen_vl_interface': 5.234861191465373e-06, 'learning_rate/action_model': 5.234861191465372e-05, 'epoch': 1.61}) +[2026-06-17 22:19:45,202][starVLA.training.train_starvla][INFO] - Step 3558, Loss: {'train/loss': 0.3222155272960663, 'train/grad_norm_pre_clip': 23.633251190185547, 'timing/data': 0.01231407094746828, 'timing/model': 0.9378747036680579, 'learning_rate/qwen_vl_interface': 5.232698513498407e-06, 'learning_rate/action_model': 5.2326985134984066e-05, 'epoch': 1.61}) +[2026-06-17 22:19:49,809][starVLA.training.train_starvla][INFO] - Step 3559, Loss: {'train/loss': 0.07947090268135071, 'train/grad_norm_pre_clip': 8.097723960876465, 'timing/data': 0.00889355968683958, 'timing/model': 0.9438514178618789, 'learning_rate/qwen_vl_interface': 5.230535839118057e-06, 'learning_rate/action_model': 5.230535839118057e-05, 'epoch': 1.61}) +[2026-06-17 22:19:53,636][starVLA.training.train_starvla][INFO] - Step 3560, Loss: {'train/loss': 0.40580272674560547, 'train/grad_norm_pre_clip': 25.23420524597168, 'timing/data': 0.008749488741159439, 'timing/model': 0.9490877334028482, 'learning_rate/qwen_vl_interface': 5.2283731687726455e-06, 'learning_rate/action_model': 5.2283731687726454e-05, 'epoch': 1.62}) +[2026-06-17 22:19:57,538][starVLA.training.train_starvla][INFO] - Step 3561, Loss: {'train/loss': 0.22932900488376617, 'train/grad_norm_pre_clip': 31.914623260498047, 'timing/data': 0.022602918557822704, 'timing/model': 0.9347339514642954, 'learning_rate/qwen_vl_interface': 5.226210502910504e-06, 'learning_rate/action_model': 5.226210502910504e-05, 'epoch': 1.62}) +[2026-06-17 22:20:01,335][starVLA.training.train_starvla][INFO] - Step 3562, Loss: {'train/loss': 0.14301934838294983, 'train/grad_norm_pre_clip': 13.755513191223145, 'timing/data': 0.008830972947180271, 'timing/model': 0.9482739353552461, 'learning_rate/qwen_vl_interface': 5.224047841979949e-06, 'learning_rate/action_model': 5.2240478419799485e-05, 'epoch': 1.62}) +[2026-06-17 22:20:05,199][starVLA.training.train_starvla][INFO] - Step 3563, Loss: {'train/loss': 0.22244389355182648, 'train/grad_norm_pre_clip': 24.99622917175293, 'timing/data': 0.012273126281797886, 'timing/model': 1.1523873656988144, 'learning_rate/qwen_vl_interface': 5.2218851864293044e-06, 'learning_rate/action_model': 5.221885186429304e-05, 'epoch': 1.62}) +[2026-06-17 22:20:09,070][starVLA.training.train_starvla][INFO] - Step 3564, Loss: {'train/loss': 0.45988309383392334, 'train/grad_norm_pre_clip': 25.77550506591797, 'timing/data': 0.0074403248727321625, 'timing/model': 1.1613765638321638, 'learning_rate/qwen_vl_interface': 5.21972253670689e-06, 'learning_rate/action_model': 5.21972253670689e-05, 'epoch': 1.62}) +[2026-06-17 22:20:12,755][starVLA.training.train_starvla][INFO] - Step 3565, Loss: {'train/loss': 0.18687690794467926, 'train/grad_norm_pre_clip': 23.068683624267578, 'timing/data': 0.008957636542618275, 'timing/model': 0.9407460475340486, 'learning_rate/qwen_vl_interface': 5.217559893261028e-06, 'learning_rate/action_model': 5.217559893261028e-05, 'epoch': 1.62}) +[2026-06-17 22:20:16,646][starVLA.training.train_starvla][INFO] - Step 3566, Loss: {'train/loss': 0.6802995204925537, 'train/grad_norm_pre_clip': 29.609663009643555, 'timing/data': 0.016442355699837208, 'timing/model': 0.9469684539362788, 'learning_rate/qwen_vl_interface': 5.2153972565400355e-06, 'learning_rate/action_model': 5.215397256540035e-05, 'epoch': 1.62}) +[2026-06-17 22:20:20,469][starVLA.training.train_starvla][INFO] - Step 3567, Loss: {'train/loss': 0.3283703327178955, 'train/grad_norm_pre_clip': 26.720291137695312, 'timing/data': 0.008945195004343987, 'timing/model': 0.9369878359138966, 'learning_rate/qwen_vl_interface': 5.213234626992227e-06, 'learning_rate/action_model': 5.213234626992226e-05, 'epoch': 1.62}) +[2026-06-17 22:20:24,340][starVLA.training.train_starvla][INFO] - Step 3568, Loss: {'train/loss': 0.3312791585922241, 'train/grad_norm_pre_clip': 23.95735740661621, 'timing/data': 0.00858161598443985, 'timing/model': 0.9459097608923912, 'learning_rate/qwen_vl_interface': 5.21107200506592e-06, 'learning_rate/action_model': 5.21107200506592e-05, 'epoch': 1.62}) +[2026-06-17 22:20:28,192][starVLA.training.train_starvla][INFO] - Step 3569, Loss: {'train/loss': 0.4157850444316864, 'train/grad_norm_pre_clip': 20.19036293029785, 'timing/data': 0.008911083452403545, 'timing/model': 0.9727519759908319, 'learning_rate/qwen_vl_interface': 5.208909391209429e-06, 'learning_rate/action_model': 5.208909391209429e-05, 'epoch': 1.62}) +[2026-06-17 22:20:32,095][starVLA.training.train_starvla][INFO] - Step 3570, Loss: {'train/loss': 0.5712520480155945, 'train/grad_norm_pre_clip': 24.749370574951172, 'timing/data': 0.007823997177183628, 'timing/model': 1.1870676493272185, 'learning_rate/qwen_vl_interface': 5.206746785871066e-06, 'learning_rate/action_model': 5.2067467858710664e-05, 'epoch': 1.62}) +[2026-06-17 22:20:35,789][starVLA.training.train_starvla][INFO] - Step 3571, Loss: {'train/loss': 0.2700786292552948, 'train/grad_norm_pre_clip': 18.66376304626465, 'timing/data': 0.012099540792405605, 'timing/model': 0.9776297556236386, 'learning_rate/qwen_vl_interface': 5.204584189499139e-06, 'learning_rate/action_model': 5.2045841894991385e-05, 'epoch': 1.62}) +[2026-06-17 22:20:39,691][starVLA.training.train_starvla][INFO] - Step 3572, Loss: {'train/loss': 0.3645406663417816, 'train/grad_norm_pre_clip': 20.69806480407715, 'timing/data': 0.016601120121777058, 'timing/model': 0.953731432557106, 'learning_rate/qwen_vl_interface': 5.202421602541959e-06, 'learning_rate/action_model': 5.2024216025419594e-05, 'epoch': 1.62}) +[2026-06-17 22:20:43,525][starVLA.training.train_starvla][INFO] - Step 3573, Loss: {'train/loss': 0.445338636636734, 'train/grad_norm_pre_clip': 26.617860794067383, 'timing/data': 0.014597669243812561, 'timing/model': 0.9557458944618702, 'learning_rate/qwen_vl_interface': 5.200259025447835e-06, 'learning_rate/action_model': 5.2002590254478345e-05, 'epoch': 1.62}) +[2026-06-17 22:20:47,439][starVLA.training.train_starvla][INFO] - Step 3574, Loss: {'train/loss': 0.3652741014957428, 'train/grad_norm_pre_clip': 19.62149429321289, 'timing/data': 0.015705274417996407, 'timing/model': 0.9606068292632699, 'learning_rate/qwen_vl_interface': 5.198096458665064e-06, 'learning_rate/action_model': 5.198096458665064e-05, 'epoch': 1.62}) +[2026-06-17 22:20:51,304][starVLA.training.train_starvla][INFO] - Step 3575, Loss: {'train/loss': 0.43905356526374817, 'train/grad_norm_pre_clip': 26.457300186157227, 'timing/data': 0.00912762526422739, 'timing/model': 0.944278122857213, 'learning_rate/qwen_vl_interface': 5.195933902641955e-06, 'learning_rate/action_model': 5.195933902641955e-05, 'epoch': 1.62}) +[2026-06-17 22:20:55,346][starVLA.training.train_starvla][INFO] - Step 3576, Loss: {'train/loss': 0.3228774666786194, 'train/grad_norm_pre_clip': 20.83235740661621, 'timing/data': 0.009175071492791176, 'timing/model': 1.2820992143824697, 'learning_rate/qwen_vl_interface': 5.193771357826807e-06, 'learning_rate/action_model': 5.193771357826806e-05, 'epoch': 1.62}) +[2026-06-17 22:20:59,238][starVLA.training.train_starvla][INFO] - Step 3577, Loss: {'train/loss': 0.48028919100761414, 'train/grad_norm_pre_clip': 22.1892032623291, 'timing/data': 0.009342453442513943, 'timing/model': 1.0863428628072143, 'learning_rate/qwen_vl_interface': 5.191608824667916e-06, 'learning_rate/action_model': 5.1916088246679165e-05, 'epoch': 1.62}) +[2026-06-17 22:21:02,938][starVLA.training.train_starvla][INFO] - Step 3578, Loss: {'train/loss': 0.3415198028087616, 'train/grad_norm_pre_clip': 21.505786895751953, 'timing/data': 0.00982745923101902, 'timing/model': 0.9502048194408417, 'learning_rate/qwen_vl_interface': 5.189446303613578e-06, 'learning_rate/action_model': 5.1894463036135775e-05, 'epoch': 1.62}) +[2026-06-17 22:21:06,884][starVLA.training.train_starvla][INFO] - Step 3579, Loss: {'train/loss': 0.2172178030014038, 'train/grad_norm_pre_clip': 15.09153938293457, 'timing/data': 0.008764448575675488, 'timing/model': 0.9412530343979597, 'learning_rate/qwen_vl_interface': 5.187283795112087e-06, 'learning_rate/action_model': 5.1872837951120866e-05, 'epoch': 1.62}) +[2026-06-17 22:21:10,724][starVLA.training.train_starvla][INFO] - Step 3580, Loss: {'train/loss': 0.39339736104011536, 'train/grad_norm_pre_clip': 17.51658821105957, 'timing/data': 0.010119521990418434, 'timing/model': 0.9459655163809657, 'learning_rate/qwen_vl_interface': 5.185121299611734e-06, 'learning_rate/action_model': 5.1851212996117335e-05, 'epoch': 1.62}) +[2026-06-17 22:21:14,530][starVLA.training.train_starvla][INFO] - Step 3581, Loss: {'train/loss': 0.1741049885749817, 'train/grad_norm_pre_clip': 13.644373893737793, 'timing/data': 0.008883065544068813, 'timing/model': 0.9254594966769218, 'learning_rate/qwen_vl_interface': 5.182958817560803e-06, 'learning_rate/action_model': 5.1829588175608026e-05, 'epoch': 1.62}) +[2026-06-17 22:21:18,262][starVLA.training.train_starvla][INFO] - Step 3582, Loss: {'train/loss': 0.3242349624633789, 'train/grad_norm_pre_clip': 25.607351303100586, 'timing/data': 0.0098633524030447, 'timing/model': 0.920110865496099, 'learning_rate/qwen_vl_interface': 5.180796349407583e-06, 'learning_rate/action_model': 5.180796349407583e-05, 'epoch': 1.63}) +[2026-06-17 22:21:22,051][starVLA.training.train_starvla][INFO] - Step 3583, Loss: {'train/loss': 0.21574518084526062, 'train/grad_norm_pre_clip': 18.49285125732422, 'timing/data': 0.009051862172782421, 'timing/model': 1.1276873294264078, 'learning_rate/qwen_vl_interface': 5.178633895600354e-06, 'learning_rate/action_model': 5.178633895600353e-05, 'epoch': 1.63}) +[2026-06-17 22:21:25,843][starVLA.training.train_starvla][INFO] - Step 3584, Loss: {'train/loss': 0.7354255318641663, 'train/grad_norm_pre_clip': 23.616085052490234, 'timing/data': 0.00867822952568531, 'timing/model': 1.1484747305512428, 'learning_rate/qwen_vl_interface': 5.176471456587394e-06, 'learning_rate/action_model': 5.176471456587394e-05, 'epoch': 1.63}) +[2026-06-17 22:21:29,409][starVLA.training.train_starvla][INFO] - Step 3585, Loss: {'train/loss': 0.15721319615840912, 'train/grad_norm_pre_clip': 13.305779457092285, 'timing/data': 0.012683185748755932, 'timing/model': 0.9181712241843343, 'learning_rate/qwen_vl_interface': 5.1743090328169805e-06, 'learning_rate/action_model': 5.1743090328169805e-05, 'epoch': 1.63}) +[2026-06-17 22:21:33,164][starVLA.training.train_starvla][INFO] - Step 3586, Loss: {'train/loss': 0.46391627192497253, 'train/grad_norm_pre_clip': 19.633005142211914, 'timing/data': 0.010809935629367828, 'timing/model': 0.9301005778834224, 'learning_rate/qwen_vl_interface': 5.172146624737388e-06, 'learning_rate/action_model': 5.1721466247373875e-05, 'epoch': 1.63}) +[2026-06-17 22:21:37,009][starVLA.training.train_starvla][INFO] - Step 3587, Loss: {'train/loss': 0.3111191391944885, 'train/grad_norm_pre_clip': 20.660194396972656, 'timing/data': 0.00878739170730114, 'timing/model': 0.9555901540443301, 'learning_rate/qwen_vl_interface': 5.169984232796883e-06, 'learning_rate/action_model': 5.169984232796883e-05, 'epoch': 1.63}) +[2026-06-17 22:21:40,843][starVLA.training.train_starvla][INFO] - Step 3588, Loss: {'train/loss': 0.17673934996128082, 'train/grad_norm_pre_clip': 14.214279174804688, 'timing/data': 0.010210664942860603, 'timing/model': 0.9237681226804852, 'learning_rate/qwen_vl_interface': 5.1678218574437325e-06, 'learning_rate/action_model': 5.167821857443732e-05, 'epoch': 1.63}) +[2026-06-17 22:21:44,661][starVLA.training.train_starvla][INFO] - Step 3589, Loss: {'train/loss': 0.3568248748779297, 'train/grad_norm_pre_clip': 19.38100814819336, 'timing/data': 0.018978754989802837, 'timing/model': 0.926322765648365, 'learning_rate/qwen_vl_interface': 5.165659499126198e-06, 'learning_rate/action_model': 5.165659499126198e-05, 'epoch': 1.63}) +[2026-06-17 22:21:49,082][starVLA.training.train_starvla][INFO] - Step 3590, Loss: {'train/loss': 0.29940879344940186, 'train/grad_norm_pre_clip': 23.409408569335938, 'timing/data': 0.011681555770337582, 'timing/model': 1.1454395931214094, 'learning_rate/qwen_vl_interface': 5.1634971582925405e-06, 'learning_rate/action_model': 5.163497158292541e-05, 'epoch': 1.63}) +[2026-06-17 22:21:53,013][starVLA.training.train_starvla][INFO] - Step 3591, Loss: {'train/loss': 0.22108334302902222, 'train/grad_norm_pre_clip': 14.126352310180664, 'timing/data': 0.02144666574895382, 'timing/model': 0.9427452320232987, 'learning_rate/qwen_vl_interface': 5.161334835391016e-06, 'learning_rate/action_model': 5.161334835391016e-05, 'epoch': 1.63}) +[2026-06-17 22:21:56,967][starVLA.training.train_starvla][INFO] - Step 3592, Loss: {'train/loss': 0.48635029792785645, 'train/grad_norm_pre_clip': 20.46468162536621, 'timing/data': 0.010367896407842636, 'timing/model': 1.0530367605388165, 'learning_rate/qwen_vl_interface': 5.159172530869875e-06, 'learning_rate/action_model': 5.159172530869874e-05, 'epoch': 1.63}) +[2026-06-17 22:22:00,780][starVLA.training.train_starvla][INFO] - Step 3593, Loss: {'train/loss': 0.5325191617012024, 'train/grad_norm_pre_clip': 21.819602966308594, 'timing/data': 0.008725449442863464, 'timing/model': 0.9345119409263134, 'learning_rate/qwen_vl_interface': 5.157010245177366e-06, 'learning_rate/action_model': 5.1570102451773664e-05, 'epoch': 1.63}) +[2026-06-17 22:22:04,986][starVLA.training.train_starvla][INFO] - Step 3594, Loss: {'train/loss': 0.37523841857910156, 'train/grad_norm_pre_clip': 19.526473999023438, 'timing/data': 0.009806474670767784, 'timing/model': 0.9385124975815415, 'learning_rate/qwen_vl_interface': 5.1548479787617346e-06, 'learning_rate/action_model': 5.1548479787617346e-05, 'epoch': 1.63}) +[2026-06-17 22:22:09,033][starVLA.training.train_starvla][INFO] - Step 3595, Loss: {'train/loss': 0.26564687490463257, 'train/grad_norm_pre_clip': 18.326086044311523, 'timing/data': 0.011124144308269024, 'timing/model': 0.9387544896453619, 'learning_rate/qwen_vl_interface': 5.152685732071219e-06, 'learning_rate/action_model': 5.1526857320712185e-05, 'epoch': 1.63}) +[2026-06-17 22:22:12,892][starVLA.training.train_starvla][INFO] - Step 3596, Loss: {'train/loss': 0.44087761640548706, 'train/grad_norm_pre_clip': 21.89644432067871, 'timing/data': 0.013983342796564102, 'timing/model': 1.1647360054776073, 'learning_rate/qwen_vl_interface': 5.1505235055540545e-06, 'learning_rate/action_model': 5.1505235055540545e-05, 'epoch': 1.63}) +[2026-06-17 22:22:16,635][starVLA.training.train_starvla][INFO] - Step 3597, Loss: {'train/loss': 0.35678958892822266, 'train/grad_norm_pre_clip': 14.977983474731445, 'timing/data': 0.010436595417559147, 'timing/model': 1.071462158113718, 'learning_rate/qwen_vl_interface': 5.148361299658477e-06, 'learning_rate/action_model': 5.148361299658476e-05, 'epoch': 1.63}) +[2026-06-17 22:22:20,265][starVLA.training.train_starvla][INFO] - Step 3598, Loss: {'train/loss': 0.2192777693271637, 'train/grad_norm_pre_clip': 16.408809661865234, 'timing/data': 0.012080677784979343, 'timing/model': 0.9302835119888186, 'learning_rate/qwen_vl_interface': 5.1461991148327095e-06, 'learning_rate/action_model': 5.1461991148327095e-05, 'epoch': 1.63}) +[2026-06-17 22:22:24,033][starVLA.training.train_starvla][INFO] - Step 3599, Loss: {'train/loss': 0.3667755126953125, 'train/grad_norm_pre_clip': 18.175867080688477, 'timing/data': 0.00966602098196745, 'timing/model': 0.9230738105252385, 'learning_rate/qwen_vl_interface': 5.144036951524977e-06, 'learning_rate/action_model': 5.144036951524977e-05, 'epoch': 1.63}) +[2026-06-17 22:24:26,110][starVLA.training.train_starvla][INFO] - Step 3600, Loss: {'train/loss': 0.4117913246154785, 'train/grad_norm_pre_clip': 21.978479385375977, 'eval/loss': 0.32791274785995483, 'eval/latency_8/loss': 0.32791274785995483, 'eval/demon_fire_precision': 0.9420461265523359, 'eval/demon_fire_recall': 0.9897483690587139, 'eval/demon_fire_f1': 0.9653082866232389, 'eval/demon_fire_rate_model': 0.9410127991096272, 'eval/demon_fire_rate_teacher': 0.8956594323873122, 'eval/demon_left_precision': 0.907051282051282, 'eval/demon_left_recall': 0.6059957173447538, 'eval/demon_left_f1': 0.7265725288831836, 'eval/demon_left_rate_model': 0.08681135225375626, 'eval/demon_left_rate_teacher': 0.12993878686700056, 'eval/demon_right_precision': 0.7513089005235603, 'eval/demon_right_recall': 0.8696969696969697, 'eval/demon_right_f1': 0.8061797752808988, 'eval/demon_right_rate_model': 0.2125765164162493, 'eval/demon_right_rate_teacher': 0.18363939899833054, 'eval/demon_move_f1': 0.7663761520820411, 'eval/demon_cc_f1': 0.8955681408468775, 'eval/latency_8/demon_fire_precision': 0.9420461265523359, 'eval/latency_8/demon_fire_recall': 0.9897483690587139, 'eval/latency_8/demon_fire_f1': 0.9653082866232389, 'eval/latency_8/demon_fire_rate_model': 0.9410127991096272, 'eval/latency_8/demon_fire_rate_teacher': 0.8956594323873122, 'eval/latency_8/demon_left_precision': 0.907051282051282, 'eval/latency_8/demon_left_recall': 0.6059957173447538, 'eval/latency_8/demon_left_f1': 0.7265725288831836, 'eval/latency_8/demon_left_rate_model': 0.08681135225375626, 'eval/latency_8/demon_left_rate_teacher': 0.12993878686700056, 'eval/latency_8/demon_right_precision': 0.7513089005235603, 'eval/latency_8/demon_right_recall': 0.8696969696969697, 'eval/latency_8/demon_right_f1': 0.8061797752808988, 'eval/latency_8/demon_right_rate_model': 0.2125765164162493, 'eval/latency_8/demon_right_rate_teacher': 0.18363939899833054, 'eval/latency_8/demon_move_f1': 0.7663761520820411, 'eval/latency_8/demon_cc_f1': 0.8955681408468775, 'timing/data': 0.007952737621963024, 'timing/model': 0.9874421507120132, 'learning_rate/qwen_vl_interface': 5.141874810183498e-06, 'learning_rate/action_model': 5.141874810183498e-05, 'epoch': 1.63}) +[2026-06-17 22:24:26,116][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_3600_state +[2026-06-17 22:24:26,121][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 22:24:32,813][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_3600_state/model.safetensors +[2026-06-17 22:24:48,084][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_3600_state/optimizer.bin +[2026-06-17 22:24:48,088][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_3600_state/sampler.bin +[2026-06-17 22:24:48,088][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_3600_state/sampler_1.bin +[2026-06-17 22:24:48,090][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_3600_state/random_states_0.pkl +[2026-06-17 22:24:48,094][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3600 +[2026-06-17 22:24:48,094][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 22:24:48,105][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 22:24:51,987][starVLA.training.train_starvla][INFO] - Step 3601, Loss: {'train/loss': 0.3756406903266907, 'train/grad_norm_pre_clip': 19.152538299560547, 'timing/data': 0.00827697478234768, 'timing/model': 0.9274017997086048, 'learning_rate/qwen_vl_interface': 5.139712691256489e-06, 'learning_rate/action_model': 5.1397126912564885e-05, 'epoch': 1.63}) +[2026-06-17 22:24:55,788][starVLA.training.train_starvla][INFO] - Step 3602, Loss: {'train/loss': 0.32108205556869507, 'train/grad_norm_pre_clip': 28.589866638183594, 'timing/data': 0.00830897781997919, 'timing/model': 0.9328161990270019, 'learning_rate/qwen_vl_interface': 5.137550595192156e-06, 'learning_rate/action_model': 5.137550595192156e-05, 'epoch': 1.63}) +[2026-06-17 22:24:59,622][starVLA.training.train_starvla][INFO] - Step 3603, Loss: {'train/loss': 0.3880901038646698, 'train/grad_norm_pre_clip': 19.182626724243164, 'timing/data': 0.008787930011749268, 'timing/model': 0.9302844852209091, 'learning_rate/qwen_vl_interface': 5.135388522438707e-06, 'learning_rate/action_model': 5.1353885224387066e-05, 'epoch': 1.63}) +[2026-06-17 22:25:03,516][starVLA.training.train_starvla][INFO] - Step 3604, Loss: {'train/loss': 0.13284075260162354, 'train/grad_norm_pre_clip': 11.423677444458008, 'timing/data': 0.009211408905684948, 'timing/model': 0.9329879926517606, 'learning_rate/qwen_vl_interface': 5.133226473444341e-06, 'learning_rate/action_model': 5.13322647344434e-05, 'epoch': 1.64}) +[2026-06-17 22:25:07,796][starVLA.training.train_starvla][INFO] - Step 3605, Loss: {'train/loss': 0.742498517036438, 'train/grad_norm_pre_clip': 22.711103439331055, 'timing/data': 0.021428844891488552, 'timing/model': 0.9424001798033714, 'learning_rate/qwen_vl_interface': 5.1310644486572515e-06, 'learning_rate/action_model': 5.131064448657251e-05, 'epoch': 1.64}) +[2026-06-17 22:25:11,624][starVLA.training.train_starvla][INFO] - Step 3606, Loss: {'train/loss': 0.1860821545124054, 'train/grad_norm_pre_clip': 15.208518981933594, 'timing/data': 0.010051732882857323, 'timing/model': 1.1336065726354718, 'learning_rate/qwen_vl_interface': 5.12890244852563e-06, 'learning_rate/action_model': 5.12890244852563e-05, 'epoch': 1.64}) +[2026-06-17 22:25:15,424][starVLA.training.train_starvla][INFO] - Step 3607, Loss: {'train/loss': 0.2611575424671173, 'train/grad_norm_pre_clip': 58.25888442993164, 'timing/data': 0.008873425424098969, 'timing/model': 1.1054473500698805, 'learning_rate/qwen_vl_interface': 5.126740473497661e-06, 'learning_rate/action_model': 5.126740473497661e-05, 'epoch': 1.64}) +[2026-06-17 22:25:19,376][starVLA.training.train_starvla][INFO] - Step 3608, Loss: {'train/loss': 0.4129103720188141, 'train/grad_norm_pre_clip': 19.784704208374023, 'timing/data': 0.01147764828056097, 'timing/model': 0.9536828212440014, 'learning_rate/qwen_vl_interface': 5.124578524021525e-06, 'learning_rate/action_model': 5.124578524021525e-05, 'epoch': 1.64}) +[2026-06-17 22:25:23,172][starVLA.training.train_starvla][INFO] - Step 3609, Loss: {'train/loss': 0.44134461879730225, 'train/grad_norm_pre_clip': 18.58249855041504, 'timing/data': 0.008436840958893299, 'timing/model': 0.9563018884509802, 'learning_rate/qwen_vl_interface': 5.122416600545396e-06, 'learning_rate/action_model': 5.122416600545397e-05, 'epoch': 1.64}) +[2026-06-17 22:25:27,089][starVLA.training.train_starvla][INFO] - Step 3610, Loss: {'train/loss': 0.13392147421836853, 'train/grad_norm_pre_clip': 11.923391342163086, 'timing/data': 0.013509027659893036, 'timing/model': 0.9509856505319476, 'learning_rate/qwen_vl_interface': 5.1202547035174436e-06, 'learning_rate/action_model': 5.120254703517443e-05, 'epoch': 1.64}) +[2026-06-17 22:25:31,053][starVLA.training.train_starvla][INFO] - Step 3611, Loss: {'train/loss': 0.4327053725719452, 'train/grad_norm_pre_clip': 19.207862854003906, 'timing/data': 0.009595373645424843, 'timing/model': 0.9305521976202726, 'learning_rate/qwen_vl_interface': 5.118092833385829e-06, 'learning_rate/action_model': 5.1180928333858293e-05, 'epoch': 1.64}) +[2026-06-17 22:25:35,058][starVLA.training.train_starvla][INFO] - Step 3612, Loss: {'train/loss': 0.23627306520938873, 'train/grad_norm_pre_clip': 21.434492111206055, 'timing/data': 0.0165734039619565, 'timing/model': 1.117484682239592, 'learning_rate/qwen_vl_interface': 5.115930990598713e-06, 'learning_rate/action_model': 5.115930990598713e-05, 'epoch': 1.64}) +[2026-06-17 22:25:38,838][starVLA.training.train_starvla][INFO] - Step 3613, Loss: {'train/loss': 0.12682583928108215, 'train/grad_norm_pre_clip': 10.934531211853027, 'timing/data': 0.009879241697490215, 'timing/model': 1.1032331949099898, 'learning_rate/qwen_vl_interface': 5.113769175604248e-06, 'learning_rate/action_model': 5.113769175604247e-05, 'epoch': 1.64}) +[2026-06-17 22:25:42,792][starVLA.training.train_starvla][INFO] - Step 3614, Loss: {'train/loss': 0.22233882546424866, 'train/grad_norm_pre_clip': 23.071964263916016, 'timing/data': 0.009556438773870468, 'timing/model': 0.9364485070109367, 'learning_rate/qwen_vl_interface': 5.111607388850576e-06, 'learning_rate/action_model': 5.111607388850576e-05, 'epoch': 1.64}) +[2026-06-17 22:25:46,694][starVLA.training.train_starvla][INFO] - Step 3615, Loss: {'train/loss': 0.15511232614517212, 'train/grad_norm_pre_clip': 12.853921890258789, 'timing/data': 0.017286483198404312, 'timing/model': 0.9282967066392303, 'learning_rate/qwen_vl_interface': 5.1094456307858444e-06, 'learning_rate/action_model': 5.109445630785844e-05, 'epoch': 1.64}) +[2026-06-17 22:25:50,561][starVLA.training.train_starvla][INFO] - Step 3616, Loss: {'train/loss': 0.3679313659667969, 'train/grad_norm_pre_clip': 20.96219825744629, 'timing/data': 0.014964784495532513, 'timing/model': 0.9554172810167074, 'learning_rate/qwen_vl_interface': 5.107283901858184e-06, 'learning_rate/action_model': 5.107283901858184e-05, 'epoch': 1.64}) +[2026-06-17 22:25:54,361][starVLA.training.train_starvla][INFO] - Step 3617, Loss: {'train/loss': 0.20029228925704956, 'train/grad_norm_pre_clip': 17.690364837646484, 'timing/data': 0.008907683193683624, 'timing/model': 0.9419245915487409, 'learning_rate/qwen_vl_interface': 5.105122202515726e-06, 'learning_rate/action_model': 5.1051222025157255e-05, 'epoch': 1.64}) +[2026-06-17 22:25:58,172][starVLA.training.train_starvla][INFO] - Step 3618, Loss: {'train/loss': 0.3945464491844177, 'train/grad_norm_pre_clip': 27.65805435180664, 'timing/data': 0.007199037820100784, 'timing/model': 0.924943283200264, 'learning_rate/qwen_vl_interface': 5.102960533206588e-06, 'learning_rate/action_model': 5.1029605332065874e-05, 'epoch': 1.64}) +[2026-06-17 22:26:02,477][starVLA.training.train_starvla][INFO] - Step 3619, Loss: {'train/loss': 0.2593073546886444, 'train/grad_norm_pre_clip': 16.49652862548828, 'timing/data': 0.01575793046504259, 'timing/model': 1.131055174395442, 'learning_rate/qwen_vl_interface': 5.100798894378891e-06, 'learning_rate/action_model': 5.100798894378891e-05, 'epoch': 1.64}) +[2026-06-17 22:26:06,273][starVLA.training.train_starvla][INFO] - Step 3620, Loss: {'train/loss': 0.21255898475646973, 'train/grad_norm_pre_clip': 15.369274139404297, 'timing/data': 0.009751109406352043, 'timing/model': 1.128134286031127, 'learning_rate/qwen_vl_interface': 5.098637286480744e-06, 'learning_rate/action_model': 5.098637286480744e-05, 'epoch': 1.64}) +[2026-06-17 22:26:10,180][starVLA.training.train_starvla][INFO] - Step 3621, Loss: {'train/loss': 0.35454845428466797, 'train/grad_norm_pre_clip': 20.475128173828125, 'timing/data': 0.00922208558768034, 'timing/model': 0.9375612661242485, 'learning_rate/qwen_vl_interface': 5.096475709960252e-06, 'learning_rate/action_model': 5.096475709960251e-05, 'epoch': 1.64}) +[2026-06-17 22:26:14,463][starVLA.training.train_starvla][INFO] - Step 3622, Loss: {'train/loss': 0.3878702223300934, 'train/grad_norm_pre_clip': 17.128311157226562, 'timing/data': 0.010976005345582962, 'timing/model': 0.9657092439010739, 'learning_rate/qwen_vl_interface': 5.094314165265506e-06, 'learning_rate/action_model': 5.0943141652655056e-05, 'epoch': 1.64}) +[2026-06-17 22:26:18,445][starVLA.training.train_starvla][INFO] - Step 3623, Loss: {'train/loss': 0.31360581517219543, 'train/grad_norm_pre_clip': 19.025978088378906, 'timing/data': 0.009553024545311928, 'timing/model': 1.1033206721767783, 'learning_rate/qwen_vl_interface': 5.092152652844606e-06, 'learning_rate/action_model': 5.092152652844606e-05, 'epoch': 1.64}) +[2026-06-17 22:26:22,264][starVLA.training.train_starvla][INFO] - Step 3624, Loss: {'train/loss': 0.16232261061668396, 'train/grad_norm_pre_clip': 11.303281784057617, 'timing/data': 0.014319582842290401, 'timing/model': 0.9299678960815072, 'learning_rate/qwen_vl_interface': 5.089991173145628e-06, 'learning_rate/action_model': 5.0899911731456285e-05, 'epoch': 1.64}) +[2026-06-17 22:26:26,305][starVLA.training.train_starvla][INFO] - Step 3625, Loss: {'train/loss': 0.2227131724357605, 'train/grad_norm_pre_clip': 16.49237060546875, 'timing/data': 0.009529301896691322, 'timing/model': 1.1688718581572175, 'learning_rate/qwen_vl_interface': 5.087829726616653e-06, 'learning_rate/action_model': 5.087829726616653e-05, 'epoch': 1.64}) +[2026-06-17 22:26:30,157][starVLA.training.train_starvla][INFO] - Step 3626, Loss: {'train/loss': 0.31613394618034363, 'train/grad_norm_pre_clip': 18.356491088867188, 'timing/data': 0.008196480572223663, 'timing/model': 1.1258213426917791, 'learning_rate/qwen_vl_interface': 5.08566831370575e-06, 'learning_rate/action_model': 5.0856683137057504e-05, 'epoch': 1.65}) +[2026-06-17 22:26:33,766][starVLA.training.train_starvla][INFO] - Step 3627, Loss: {'train/loss': 0.30837365984916687, 'train/grad_norm_pre_clip': 21.36563491821289, 'timing/data': 0.015025490894913673, 'timing/model': 0.9335376918315887, 'learning_rate/qwen_vl_interface': 5.083506934860982e-06, 'learning_rate/action_model': 5.083506934860982e-05, 'epoch': 1.65}) +[2026-06-17 22:26:37,970][starVLA.training.train_starvla][INFO] - Step 3628, Loss: {'train/loss': 0.1073237732052803, 'train/grad_norm_pre_clip': 9.88547134399414, 'timing/data': 0.008478567935526371, 'timing/model': 1.0504718087613583, 'learning_rate/qwen_vl_interface': 5.081345590530407e-06, 'learning_rate/action_model': 5.081345590530407e-05, 'epoch': 1.65}) +[2026-06-17 22:26:41,837][starVLA.training.train_starvla][INFO] - Step 3629, Loss: {'train/loss': 0.16604548692703247, 'train/grad_norm_pre_clip': 13.072808265686035, 'timing/data': 0.008128086104989052, 'timing/model': 1.0011107986792922, 'learning_rate/qwen_vl_interface': 5.079184281162069e-06, 'learning_rate/action_model': 5.0791842811620694e-05, 'epoch': 1.65}) +[2026-06-17 22:26:45,886][starVLA.training.train_starvla][INFO] - Step 3630, Loss: {'train/loss': 0.23558330535888672, 'train/grad_norm_pre_clip': 18.800764083862305, 'timing/data': 0.00807165913283825, 'timing/model': 1.080483946017921, 'learning_rate/qwen_vl_interface': 5.077023007204018e-06, 'learning_rate/action_model': 5.077023007204018e-05, 'epoch': 1.65}) +[2026-06-17 22:26:49,723][starVLA.training.train_starvla][INFO] - Step 3631, Loss: {'train/loss': 0.5510400533676147, 'train/grad_norm_pre_clip': 26.35887336730957, 'timing/data': 0.008655120618641376, 'timing/model': 0.9313417337834835, 'learning_rate/qwen_vl_interface': 5.074861769104281e-06, 'learning_rate/action_model': 5.074861769104281e-05, 'epoch': 1.65}) +[2026-06-17 22:26:53,597][starVLA.training.train_starvla][INFO] - Step 3632, Loss: {'train/loss': 0.2806819677352905, 'train/grad_norm_pre_clip': 17.99755096435547, 'timing/data': 0.006865976378321648, 'timing/model': 1.156620896421373, 'learning_rate/qwen_vl_interface': 5.07270056731089e-06, 'learning_rate/action_model': 5.07270056731089e-05, 'epoch': 1.65}) +[2026-06-17 22:26:57,555][starVLA.training.train_starvla][INFO] - Step 3633, Loss: {'train/loss': 0.2506338655948639, 'train/grad_norm_pre_clip': 14.459896087646484, 'timing/data': 0.00892410334199667, 'timing/model': 1.2404102887958288, 'learning_rate/qwen_vl_interface': 5.070539402271859e-06, 'learning_rate/action_model': 5.070539402271859e-05, 'epoch': 1.65}) +[2026-06-17 22:27:01,184][starVLA.training.train_starvla][INFO] - Step 3634, Loss: {'train/loss': 0.8371372818946838, 'train/grad_norm_pre_clip': 30.75914764404297, 'timing/data': 0.010668298229575157, 'timing/model': 0.9404116049408913, 'learning_rate/qwen_vl_interface': 5.068378274435205e-06, 'learning_rate/action_model': 5.068378274435205e-05, 'epoch': 1.65}) +[2026-06-17 22:27:05,094][starVLA.training.train_starvla][INFO] - Step 3635, Loss: {'train/loss': 0.454471617937088, 'train/grad_norm_pre_clip': 32.11377716064453, 'timing/data': 0.009726082906126976, 'timing/model': 0.9438453251495957, 'learning_rate/qwen_vl_interface': 5.066217184248931e-06, 'learning_rate/action_model': 5.0662171842489305e-05, 'epoch': 1.65}) +[2026-06-17 22:27:09,271][starVLA.training.train_starvla][INFO] - Step 3636, Loss: {'train/loss': 0.37640446424484253, 'train/grad_norm_pre_clip': 23.882549285888672, 'timing/data': 0.01085684821009636, 'timing/model': 0.9390940675511956, 'learning_rate/qwen_vl_interface': 5.0640561321610295e-06, 'learning_rate/action_model': 5.064056132161029e-05, 'epoch': 1.65}) +[2026-06-17 22:27:13,296][starVLA.training.train_starvla][INFO] - Step 3637, Loss: {'train/loss': 0.12219303101301193, 'train/grad_norm_pre_clip': 13.81950569152832, 'timing/data': 0.009940612129867077, 'timing/model': 0.9339805906638503, 'learning_rate/qwen_vl_interface': 5.061895118619493e-06, 'learning_rate/action_model': 5.061895118619493e-05, 'epoch': 1.65}) +[2026-06-17 22:27:17,156][starVLA.training.train_starvla][INFO] - Step 3638, Loss: {'train/loss': 0.48981136083602905, 'train/grad_norm_pre_clip': 24.397260665893555, 'timing/data': 0.012407868169248104, 'timing/model': 0.9264776147902012, 'learning_rate/qwen_vl_interface': 5.059734144072302e-06, 'learning_rate/action_model': 5.059734144072302e-05, 'epoch': 1.65}) +[2026-06-17 22:27:21,062][starVLA.training.train_starvla][INFO] - Step 3639, Loss: {'train/loss': 0.1935882866382599, 'train/grad_norm_pre_clip': 20.133188247680664, 'timing/data': 0.014618574641644955, 'timing/model': 1.107768221758306, 'learning_rate/qwen_vl_interface': 5.0575732089674275e-06, 'learning_rate/action_model': 5.057573208967428e-05, 'epoch': 1.65}) +[2026-06-17 22:27:24,993][starVLA.training.train_starvla][INFO] - Step 3640, Loss: {'train/loss': 0.2082710862159729, 'train/grad_norm_pre_clip': 17.181812286376953, 'timing/data': 0.02137887291610241, 'timing/model': 1.1281643342226744, 'learning_rate/qwen_vl_interface': 5.055412313752832e-06, 'learning_rate/action_model': 5.055412313752832e-05, 'epoch': 1.65}) +[2026-06-17 22:27:28,672][starVLA.training.train_starvla][INFO] - Step 3641, Loss: {'train/loss': 0.31230276823043823, 'train/grad_norm_pre_clip': 18.460493087768555, 'timing/data': 0.009545997716486454, 'timing/model': 0.9293502513319254, 'learning_rate/qwen_vl_interface': 5.053251458876474e-06, 'learning_rate/action_model': 5.0532514588764744e-05, 'epoch': 1.65}) +[2026-06-17 22:27:32,492][starVLA.training.train_starvla][INFO] - Step 3642, Loss: {'train/loss': 0.09912844747304916, 'train/grad_norm_pre_clip': 15.284878730773926, 'timing/data': 0.009964493103325367, 'timing/model': 0.9322699159383774, 'learning_rate/qwen_vl_interface': 5.0510906447863025e-06, 'learning_rate/action_model': 5.051090644786303e-05, 'epoch': 1.65}) +[2026-06-17 22:27:36,299][starVLA.training.train_starvla][INFO] - Step 3643, Loss: {'train/loss': 0.1596105396747589, 'train/grad_norm_pre_clip': 16.68250846862793, 'timing/data': 0.008945398032665253, 'timing/model': 0.9317291108891368, 'learning_rate/qwen_vl_interface': 5.048929871930252e-06, 'learning_rate/action_model': 5.048929871930251e-05, 'epoch': 1.65}) +[2026-06-17 22:27:40,172][starVLA.training.train_starvla][INFO] - Step 3644, Loss: {'train/loss': 0.45471590757369995, 'train/grad_norm_pre_clip': 23.19770050048828, 'timing/data': 0.014654749073088169, 'timing/model': 0.9290560893714428, 'learning_rate/qwen_vl_interface': 5.046769140756257e-06, 'learning_rate/action_model': 5.046769140756257e-05, 'epoch': 1.65}) +[2026-06-17 22:27:43,988][starVLA.training.train_starvla][INFO] - Step 3645, Loss: {'train/loss': 0.3626430928707123, 'train/grad_norm_pre_clip': 22.767553329467773, 'timing/data': 0.01916566863656044, 'timing/model': 0.9477025968953967, 'learning_rate/qwen_vl_interface': 5.044608451712237e-06, 'learning_rate/action_model': 5.044608451712237e-05, 'epoch': 1.65}) +[2026-06-17 22:27:47,751][starVLA.training.train_starvla][INFO] - Step 3646, Loss: {'train/loss': 0.25639545917510986, 'train/grad_norm_pre_clip': 18.855762481689453, 'timing/data': 0.012687986716628075, 'timing/model': 1.0842537051066756, 'learning_rate/qwen_vl_interface': 5.042447805246105e-06, 'learning_rate/action_model': 5.042447805246106e-05, 'epoch': 1.65}) +[2026-06-17 22:27:51,359][starVLA.training.train_starvla][INFO] - Step 3647, Loss: {'train/loss': 0.5293630361557007, 'train/grad_norm_pre_clip': 29.60313606262207, 'timing/data': 0.018536816351115704, 'timing/model': 0.9311342788860202, 'learning_rate/qwen_vl_interface': 5.040287201805767e-06, 'learning_rate/action_model': 5.040287201805767e-05, 'epoch': 1.65}) +[2026-06-17 22:27:55,205][starVLA.training.train_starvla][INFO] - Step 3648, Loss: {'train/loss': 0.2919008433818817, 'train/grad_norm_pre_clip': 19.268871307373047, 'timing/data': 0.016423351131379604, 'timing/model': 0.9244620278477669, 'learning_rate/qwen_vl_interface': 5.038126641839119e-06, 'learning_rate/action_model': 5.0381266418391184e-05, 'epoch': 1.65}) +[2026-06-17 22:27:58,902][starVLA.training.train_starvla][INFO] - Step 3649, Loss: {'train/loss': 0.3047417998313904, 'train/grad_norm_pre_clip': 18.983808517456055, 'timing/data': 0.017287220805883408, 'timing/model': 0.9179883655160666, 'learning_rate/qwen_vl_interface': 5.0359661257940455e-06, 'learning_rate/action_model': 5.035966125794045e-05, 'epoch': 1.66}) +[2026-06-17 22:28:02,704][starVLA.training.train_starvla][INFO] - Step 3650, Loss: {'train/loss': 0.07866715639829636, 'train/grad_norm_pre_clip': 9.508163452148438, 'timing/data': 0.010507296770811081, 'timing/model': 0.9265663754194975, 'learning_rate/qwen_vl_interface': 5.033805654118424e-06, 'learning_rate/action_model': 5.033805654118424e-05, 'epoch': 1.66}) +[2026-06-17 22:28:06,478][starVLA.training.train_starvla][INFO] - Step 3651, Loss: {'train/loss': 0.430458664894104, 'train/grad_norm_pre_clip': 21.543941497802734, 'timing/data': 0.019984515383839607, 'timing/model': 0.9263975257053971, 'learning_rate/qwen_vl_interface': 5.031645227260122e-06, 'learning_rate/action_model': 5.0316452272601224e-05, 'epoch': 1.66}) +[2026-06-17 22:28:10,272][starVLA.training.train_starvla][INFO] - Step 3652, Loss: {'train/loss': 0.4749254882335663, 'train/grad_norm_pre_clip': 25.968456268310547, 'timing/data': 0.009260422550141811, 'timing/model': 1.1186591209843755, 'learning_rate/qwen_vl_interface': 5.029484845667001e-06, 'learning_rate/action_model': 5.029484845667001e-05, 'epoch': 1.66}) +[2026-06-17 22:28:13,996][starVLA.training.train_starvla][INFO] - Step 3653, Loss: {'train/loss': 0.19667014479637146, 'train/grad_norm_pre_clip': 16.63617515563965, 'timing/data': 0.009666305966675282, 'timing/model': 1.0735115492716432, 'learning_rate/qwen_vl_interface': 5.027324509786908e-06, 'learning_rate/action_model': 5.027324509786908e-05, 'epoch': 1.66}) +[2026-06-17 22:28:17,576][starVLA.training.train_starvla][INFO] - Step 3654, Loss: {'train/loss': 0.42003417015075684, 'train/grad_norm_pre_clip': 23.927518844604492, 'timing/data': 0.009901574812829494, 'timing/model': 0.9286575261503458, 'learning_rate/qwen_vl_interface': 5.025164220067683e-06, 'learning_rate/action_model': 5.025164220067683e-05, 'epoch': 1.66}) +[2026-06-17 22:28:21,360][starVLA.training.train_starvla][INFO] - Step 3655, Loss: {'train/loss': 0.6545437574386597, 'train/grad_norm_pre_clip': 26.34756851196289, 'timing/data': 0.008912624791264534, 'timing/model': 0.9276654617860913, 'learning_rate/qwen_vl_interface': 5.023003976957159e-06, 'learning_rate/action_model': 5.023003976957159e-05, 'epoch': 1.66}) +[2026-06-17 22:28:25,113][starVLA.training.train_starvla][INFO] - Step 3656, Loss: {'train/loss': 0.4106281101703644, 'train/grad_norm_pre_clip': 24.105239868164062, 'timing/data': 0.016579880379140377, 'timing/model': 0.9232850410044193, 'learning_rate/qwen_vl_interface': 5.020843780903155e-06, 'learning_rate/action_model': 5.020843780903155e-05, 'epoch': 1.66}) +[2026-06-17 22:28:28,867][starVLA.training.train_starvla][INFO] - Step 3657, Loss: {'train/loss': 0.5095666646957397, 'train/grad_norm_pre_clip': 27.24610710144043, 'timing/data': 0.009139836765825748, 'timing/model': 0.9208504557609558, 'learning_rate/qwen_vl_interface': 5.01868363235348e-06, 'learning_rate/action_model': 5.01868363235348e-05, 'epoch': 1.66}) +[2026-06-17 22:28:32,623][starVLA.training.train_starvla][INFO] - Step 3658, Loss: {'train/loss': 0.10254672914743423, 'train/grad_norm_pre_clip': 9.431578636169434, 'timing/data': 0.010919297114014626, 'timing/model': 0.9253869280219078, 'learning_rate/qwen_vl_interface': 5.016523531755937e-06, 'learning_rate/action_model': 5.016523531755937e-05, 'epoch': 1.66}) +[2026-06-17 22:28:36,357][starVLA.training.train_starvla][INFO] - Step 3659, Loss: {'train/loss': 0.40681159496307373, 'train/grad_norm_pre_clip': 20.98427963256836, 'timing/data': 0.009141006506979465, 'timing/model': 1.0945008900016546, 'learning_rate/qwen_vl_interface': 5.014363479558318e-06, 'learning_rate/action_model': 5.014363479558318e-05, 'epoch': 1.66}) +[2026-06-17 22:28:40,103][starVLA.training.train_starvla][INFO] - Step 3660, Loss: {'train/loss': 0.2289319634437561, 'train/grad_norm_pre_clip': 20.183000564575195, 'timing/data': 0.007679091766476631, 'timing/model': 1.0967661645263433, 'learning_rate/qwen_vl_interface': 5.012203476208403e-06, 'learning_rate/action_model': 5.0122034762084025e-05, 'epoch': 1.66}) +[2026-06-17 22:28:43,695][starVLA.training.train_starvla][INFO] - Step 3661, Loss: {'train/loss': 0.5387031435966492, 'train/grad_norm_pre_clip': 31.267230987548828, 'timing/data': 0.006879194639623165, 'timing/model': 0.9222307661548257, 'learning_rate/qwen_vl_interface': 5.010043522153965e-06, 'learning_rate/action_model': 5.0100435221539646e-05, 'epoch': 1.66}) +[2026-06-17 22:28:47,451][starVLA.training.train_starvla][INFO] - Step 3662, Loss: {'train/loss': 0.4762962758541107, 'train/grad_norm_pre_clip': 31.099674224853516, 'timing/data': 0.007638430222868919, 'timing/model': 0.9242525845766068, 'learning_rate/qwen_vl_interface': 5.007883617842759e-06, 'learning_rate/action_model': 5.007883617842759e-05, 'epoch': 1.66}) +[2026-06-17 22:28:51,238][starVLA.training.train_starvla][INFO] - Step 3663, Loss: {'train/loss': 0.2342115044593811, 'train/grad_norm_pre_clip': 19.163461685180664, 'timing/data': 0.010730426758527756, 'timing/model': 0.9256132235750556, 'learning_rate/qwen_vl_interface': 5.005723763722543e-06, 'learning_rate/action_model': 5.005723763722543e-05, 'epoch': 1.66}) +[2026-06-17 22:28:54,982][starVLA.training.train_starvla][INFO] - Step 3664, Loss: {'train/loss': 0.3357260227203369, 'train/grad_norm_pre_clip': 19.522769927978516, 'timing/data': 0.008054197765886784, 'timing/model': 0.9274815106764436, 'learning_rate/qwen_vl_interface': 5.003563960241052e-06, 'learning_rate/action_model': 5.003563960241052e-05, 'epoch': 1.66}) +[2026-06-17 22:28:58,815][starVLA.training.train_starvla][INFO] - Step 3665, Loss: {'train/loss': 0.2720009684562683, 'train/grad_norm_pre_clip': 35.16765594482422, 'timing/data': 0.007893567904829979, 'timing/model': 1.006997355259955, 'learning_rate/qwen_vl_interface': 5.001404207846017e-06, 'learning_rate/action_model': 5.001404207846017e-05, 'epoch': 1.66}) +[2026-06-17 22:29:02,623][starVLA.training.train_starvla][INFO] - Step 3666, Loss: {'train/loss': 0.31498634815216064, 'train/grad_norm_pre_clip': 22.72057342529297, 'timing/data': 0.011787815019488335, 'timing/model': 1.1488250605762005, 'learning_rate/qwen_vl_interface': 4.999244506985158e-06, 'learning_rate/action_model': 4.999244506985158e-05, 'epoch': 1.66}) +[2026-06-17 22:29:06,257][starVLA.training.train_starvla][INFO] - Step 3667, Loss: {'train/loss': 0.6547360420227051, 'train/grad_norm_pre_clip': 25.627769470214844, 'timing/data': 0.010991326533257961, 'timing/model': 0.9348568497225642, 'learning_rate/qwen_vl_interface': 4.997084858106182e-06, 'learning_rate/action_model': 4.997084858106183e-05, 'epoch': 1.66}) +[2026-06-17 22:29:10,023][starVLA.training.train_starvla][INFO] - Step 3668, Loss: {'train/loss': 0.22965900599956512, 'train/grad_norm_pre_clip': 12.183987617492676, 'timing/data': 0.006827850826084614, 'timing/model': 0.9238599240779877, 'learning_rate/qwen_vl_interface': 4.994925261656789e-06, 'learning_rate/action_model': 4.994925261656789e-05, 'epoch': 1.66}) +[2026-06-17 22:29:13,775][starVLA.training.train_starvla][INFO] - Step 3669, Loss: {'train/loss': 0.25719618797302246, 'train/grad_norm_pre_clip': 16.67531967163086, 'timing/data': 0.0067978594452142715, 'timing/model': 0.9343462772667408, 'learning_rate/qwen_vl_interface': 4.99276571808466e-06, 'learning_rate/action_model': 4.9927657180846605e-05, 'epoch': 1.66}) +[2026-06-17 22:29:17,545][starVLA.training.train_starvla][INFO] - Step 3670, Loss: {'train/loss': 0.5397465825080872, 'train/grad_norm_pre_clip': 31.72494125366211, 'timing/data': 0.0067601315677165985, 'timing/model': 0.9209069991484284, 'learning_rate/qwen_vl_interface': 4.9906062278374775e-06, 'learning_rate/action_model': 4.990606227837478e-05, 'epoch': 1.66}) +[2026-06-17 22:29:21,285][starVLA.training.train_starvla][INFO] - Step 3671, Loss: {'train/loss': 0.2615787386894226, 'train/grad_norm_pre_clip': 19.304018020629883, 'timing/data': 0.009024050086736679, 'timing/model': 0.9229250615462661, 'learning_rate/qwen_vl_interface': 4.988446791362903e-06, 'learning_rate/action_model': 4.988446791362903e-05, 'epoch': 1.67}) +[2026-06-17 22:29:25,050][starVLA.training.train_starvla][INFO] - Step 3672, Loss: {'train/loss': 0.264831006526947, 'train/grad_norm_pre_clip': 20.645463943481445, 'timing/data': 0.01273883506655693, 'timing/model': 1.1087935892865062, 'learning_rate/qwen_vl_interface': 4.986287409108589e-06, 'learning_rate/action_model': 4.9862874091085893e-05, 'epoch': 1.67}) +[2026-06-17 22:29:28,788][starVLA.training.train_starvla][INFO] - Step 3673, Loss: {'train/loss': 0.1882772296667099, 'train/grad_norm_pre_clip': 14.20443344116211, 'timing/data': 0.015916756354272366, 'timing/model': 1.075674102641642, 'learning_rate/qwen_vl_interface': 4.984128081522179e-06, 'learning_rate/action_model': 4.984128081522178e-05, 'epoch': 1.67}) +[2026-06-17 22:29:32,370][starVLA.training.train_starvla][INFO] - Step 3674, Loss: {'train/loss': 0.4563449025154114, 'train/grad_norm_pre_clip': 23.79526138305664, 'timing/data': 0.012869705446064472, 'timing/model': 0.9173927903175354, 'learning_rate/qwen_vl_interface': 4.981968809051304e-06, 'learning_rate/action_model': 4.981968809051304e-05, 'epoch': 1.67}) +[2026-06-17 22:29:36,094][starVLA.training.train_starvla][INFO] - Step 3675, Loss: {'train/loss': 0.2603128254413605, 'train/grad_norm_pre_clip': 22.012819290161133, 'timing/data': 0.00875340960919857, 'timing/model': 0.9232575753703713, 'learning_rate/qwen_vl_interface': 4.979809592143583e-06, 'learning_rate/action_model': 4.979809592143583e-05, 'epoch': 1.67}) +[2026-06-17 22:29:39,844][starVLA.training.train_starvla][INFO] - Step 3676, Loss: {'train/loss': 0.25250399112701416, 'train/grad_norm_pre_clip': 17.19222640991211, 'timing/data': 0.00972922332584858, 'timing/model': 0.9300867598503828, 'learning_rate/qwen_vl_interface': 4.977650431246624e-06, 'learning_rate/action_model': 4.977650431246624e-05, 'epoch': 1.67}) +[2026-06-17 22:29:43,713][starVLA.training.train_starvla][INFO] - Step 3677, Loss: {'train/loss': 0.26837629079818726, 'train/grad_norm_pre_clip': 23.18560791015625, 'timing/data': 0.009678728878498077, 'timing/model': 1.026529461145401, 'learning_rate/qwen_vl_interface': 4.9754913268080255e-06, 'learning_rate/action_model': 4.975491326808025e-05, 'epoch': 1.67}) +[2026-06-17 22:29:47,461][starVLA.training.train_starvla][INFO] - Step 3678, Loss: {'train/loss': 0.37203192710876465, 'train/grad_norm_pre_clip': 21.57872200012207, 'timing/data': 0.007572031579911709, 'timing/model': 0.925464422442019, 'learning_rate/qwen_vl_interface': 4.973332279275369e-06, 'learning_rate/action_model': 4.973332279275369e-05, 'epoch': 1.67}) +[2026-06-17 22:29:51,224][starVLA.training.train_starvla][INFO] - Step 3679, Loss: {'train/loss': 0.20562484860420227, 'train/grad_norm_pre_clip': 16.642486572265625, 'timing/data': 0.008816909976303577, 'timing/model': 1.1054389281198382, 'learning_rate/qwen_vl_interface': 4.971173289096229e-06, 'learning_rate/action_model': 4.9711732890962286e-05, 'epoch': 1.67}) +[2026-06-17 22:29:55,034][starVLA.training.train_starvla][INFO] - Step 3680, Loss: {'train/loss': 0.38996583223342896, 'train/grad_norm_pre_clip': 21.056900024414062, 'timing/data': 0.014814147725701332, 'timing/model': 1.1326903644949198, 'learning_rate/qwen_vl_interface': 4.969014356718165e-06, 'learning_rate/action_model': 4.969014356718165e-05, 'epoch': 1.67}) +[2026-06-17 22:29:58,645][starVLA.training.train_starvla][INFO] - Step 3681, Loss: {'train/loss': 0.3406209647655487, 'train/grad_norm_pre_clip': 16.445205688476562, 'timing/data': 0.013831542804837227, 'timing/model': 0.9316329834982753, 'learning_rate/qwen_vl_interface': 4.9668554825887285e-06, 'learning_rate/action_model': 4.966855482588728e-05, 'epoch': 1.67}) +[2026-06-17 22:30:02,461][starVLA.training.train_starvla][INFO] - Step 3682, Loss: {'train/loss': 0.1678110957145691, 'train/grad_norm_pre_clip': 16.613544464111328, 'timing/data': 0.009226655587553978, 'timing/model': 0.931747023947537, 'learning_rate/qwen_vl_interface': 4.964696667155455e-06, 'learning_rate/action_model': 4.964696667155454e-05, 'epoch': 1.67}) +[2026-06-17 22:30:06,354][starVLA.training.train_starvla][INFO] - Step 3683, Loss: {'train/loss': 0.3554389178752899, 'train/grad_norm_pre_clip': 20.75832176208496, 'timing/data': 0.009799377992749214, 'timing/model': 0.9642797140404582, 'learning_rate/qwen_vl_interface': 4.96253791086587e-06, 'learning_rate/action_model': 4.96253791086587e-05, 'epoch': 1.67}) +[2026-06-17 22:30:10,108][starVLA.training.train_starvla][INFO] - Step 3684, Loss: {'train/loss': 0.31367531418800354, 'train/grad_norm_pre_clip': 20.652023315429688, 'timing/data': 0.010351911187171936, 'timing/model': 0.9265790702775121, 'learning_rate/qwen_vl_interface': 4.960379214167483e-06, 'learning_rate/action_model': 4.960379214167483e-05, 'epoch': 1.67}) +[2026-06-17 22:30:13,860][starVLA.training.train_starvla][INFO] - Step 3685, Loss: {'train/loss': 0.10239386558532715, 'train/grad_norm_pre_clip': 10.894303321838379, 'timing/data': 0.007400980219244957, 'timing/model': 0.926130747422576, 'learning_rate/qwen_vl_interface': 4.9582205775077994e-06, 'learning_rate/action_model': 4.958220577507799e-05, 'epoch': 1.67}) +[2026-06-17 22:30:17,618][starVLA.training.train_starvla][INFO] - Step 3686, Loss: {'train/loss': 0.18539203703403473, 'train/grad_norm_pre_clip': 13.520125389099121, 'timing/data': 0.009653464891016483, 'timing/model': 1.0925052957609296, 'learning_rate/qwen_vl_interface': 4.956062001334302e-06, 'learning_rate/action_model': 4.956062001334302e-05, 'epoch': 1.67}) +[2026-06-17 22:30:21,201][starVLA.training.train_starvla][INFO] - Step 3687, Loss: {'train/loss': 0.47997984290122986, 'train/grad_norm_pre_clip': 18.900272369384766, 'timing/data': 0.01373132411390543, 'timing/model': 0.9203202975913882, 'learning_rate/qwen_vl_interface': 4.953903486094469e-06, 'learning_rate/action_model': 4.953903486094469e-05, 'epoch': 1.67}) +[2026-06-17 22:30:24,928][starVLA.training.train_starvla][INFO] - Step 3688, Loss: {'train/loss': 0.314215749502182, 'train/grad_norm_pre_clip': 17.731355667114258, 'timing/data': 0.014512565918266773, 'timing/model': 0.9191220793873072, 'learning_rate/qwen_vl_interface': 4.951745032235762e-06, 'learning_rate/action_model': 4.951745032235762e-05, 'epoch': 1.67}) +[2026-06-17 22:30:28,717][starVLA.training.train_starvla][INFO] - Step 3689, Loss: {'train/loss': 0.43324291706085205, 'train/grad_norm_pre_clip': 28.906932830810547, 'timing/data': 0.009898154065012932, 'timing/model': 0.9192049875855446, 'learning_rate/qwen_vl_interface': 4.949586640205631e-06, 'learning_rate/action_model': 4.94958664020563e-05, 'epoch': 1.67}) +[2026-06-17 22:30:32,436][starVLA.training.train_starvla][INFO] - Step 3690, Loss: {'train/loss': 0.15792535245418549, 'train/grad_norm_pre_clip': 13.415691375732422, 'timing/data': 0.008016013540327549, 'timing/model': 0.9211659822613001, 'learning_rate/qwen_vl_interface': 4.9474283104515124e-06, 'learning_rate/action_model': 4.9474283104515126e-05, 'epoch': 1.67}) +[2026-06-17 22:30:36,165][starVLA.training.train_starvla][INFO] - Step 3691, Loss: {'train/loss': 0.30260494351387024, 'train/grad_norm_pre_clip': 15.40179443359375, 'timing/data': 0.009932971559464931, 'timing/model': 0.9234917834401131, 'learning_rate/qwen_vl_interface': 4.945270043420829e-06, 'learning_rate/action_model': 4.945270043420828e-05, 'epoch': 1.67}) +[2026-06-17 22:30:39,890][starVLA.training.train_starvla][INFO] - Step 3692, Loss: {'train/loss': 0.28610238432884216, 'train/grad_norm_pre_clip': 18.10616683959961, 'timing/data': 0.007815955206751823, 'timing/model': 1.0916093569248915, 'learning_rate/qwen_vl_interface': 4.9431118395609954e-06, 'learning_rate/action_model': 4.9431118395609954e-05, 'epoch': 1.67}) +[2026-06-17 22:30:43,586][starVLA.training.train_starvla][INFO] - Step 3693, Loss: {'train/loss': 0.2978112995624542, 'train/grad_norm_pre_clip': 19.734403610229492, 'timing/data': 0.008656262420117855, 'timing/model': 1.0599949397146702, 'learning_rate/qwen_vl_interface': 4.940953699319407e-06, 'learning_rate/action_model': 4.940953699319407e-05, 'epoch': 1.68}) +[2026-06-17 22:30:47,138][starVLA.training.train_starvla][INFO] - Step 3694, Loss: {'train/loss': 0.3752240836620331, 'train/grad_norm_pre_clip': 23.321630477905273, 'timing/data': 0.007644595578312874, 'timing/model': 0.9237061832100153, 'learning_rate/qwen_vl_interface': 4.93879562314345e-06, 'learning_rate/action_model': 4.93879562314345e-05, 'epoch': 1.68}) +[2026-06-17 22:30:50,902][starVLA.training.train_starvla][INFO] - Step 3695, Loss: {'train/loss': 0.2876921594142914, 'train/grad_norm_pre_clip': 13.978986740112305, 'timing/data': 0.013023332692682743, 'timing/model': 0.933282857760787, 'learning_rate/qwen_vl_interface': 4.936637611480495e-06, 'learning_rate/action_model': 4.9366376114804945e-05, 'epoch': 1.68}) +[2026-06-17 22:30:54,645][starVLA.training.train_starvla][INFO] - Step 3696, Loss: {'train/loss': 0.25509127974510193, 'train/grad_norm_pre_clip': 11.386343955993652, 'timing/data': 0.007673605345189571, 'timing/model': 0.9257528651505709, 'learning_rate/qwen_vl_interface': 4.9344796647779e-06, 'learning_rate/action_model': 4.9344796647778995e-05, 'epoch': 1.68}) +[2026-06-17 22:30:58,382][starVLA.training.train_starvla][INFO] - Step 3697, Loss: {'train/loss': 0.12828221917152405, 'train/grad_norm_pre_clip': 16.31905174255371, 'timing/data': 0.013245235197246075, 'timing/model': 0.9203039295971394, 'learning_rate/qwen_vl_interface': 4.9323217834830105e-06, 'learning_rate/action_model': 4.9323217834830105e-05, 'epoch': 1.68}) +[2026-06-17 22:31:02,106][starVLA.training.train_starvla][INFO] - Step 3698, Loss: {'train/loss': 0.3113583028316498, 'train/grad_norm_pre_clip': 18.09096908569336, 'timing/data': 0.0077627068385481834, 'timing/model': 0.9250521454960108, 'learning_rate/qwen_vl_interface': 4.930163968043155e-06, 'learning_rate/action_model': 4.930163968043155e-05, 'epoch': 1.68}) +[2026-06-17 22:31:05,870][starVLA.training.train_starvla][INFO] - Step 3699, Loss: {'train/loss': 0.34543710947036743, 'train/grad_norm_pre_clip': 18.96765899658203, 'timing/data': 0.015855700708925724, 'timing/model': 1.1075857235118747, 'learning_rate/qwen_vl_interface': 4.928006218905656e-06, 'learning_rate/action_model': 4.928006218905656e-05, 'epoch': 1.68}) +[2026-06-17 22:33:04,957][starVLA.training.train_starvla][INFO] - Step 3700, Loss: {'train/loss': 0.5094696283340454, 'train/grad_norm_pre_clip': 25.50249671936035, 'eval/loss': 0.3265551030635834, 'eval/latency_8/loss': 0.3265551030635834, 'eval/demon_fire_precision': 0.9584859584859585, 'eval/demon_fire_recall': 0.9754582168375272, 'eval/demon_fire_f1': 0.9668976135488838, 'eval/demon_fire_rate_model': 0.9115191986644408, 'eval/demon_fire_rate_teacher': 0.8956594323873122, 'eval/demon_left_precision': 0.7923250564334086, 'eval/demon_left_recall': 0.7516059957173448, 'eval/demon_left_f1': 0.7714285714285715, 'eval/demon_left_rate_model': 0.12326099053978853, 'eval/demon_left_rate_teacher': 0.12993878686700056, 'eval/demon_right_precision': 0.8404802744425386, 'eval/demon_right_recall': 0.7424242424242424, 'eval/demon_right_f1': 0.7884151246983105, 'eval/demon_right_rate_model': 0.16221480244852532, 'eval/demon_right_rate_teacher': 0.18363939899833054, 'eval/demon_move_f1': 0.779921848063441, 'eval/demon_cc_f1': 0.9020232347067978, 'eval/latency_8/demon_fire_precision': 0.9584859584859585, 'eval/latency_8/demon_fire_recall': 0.9754582168375272, 'eval/latency_8/demon_fire_f1': 0.9668976135488838, 'eval/latency_8/demon_fire_rate_model': 0.9115191986644408, 'eval/latency_8/demon_fire_rate_teacher': 0.8956594323873122, 'eval/latency_8/demon_left_precision': 0.7923250564334086, 'eval/latency_8/demon_left_recall': 0.7516059957173448, 'eval/latency_8/demon_left_f1': 0.7714285714285715, 'eval/latency_8/demon_left_rate_model': 0.12326099053978853, 'eval/latency_8/demon_left_rate_teacher': 0.12993878686700056, 'eval/latency_8/demon_right_precision': 0.8404802744425386, 'eval/latency_8/demon_right_recall': 0.7424242424242424, 'eval/latency_8/demon_right_f1': 0.7884151246983105, 'eval/latency_8/demon_right_rate_model': 0.16221480244852532, 'eval/latency_8/demon_right_rate_teacher': 0.18363939899833054, 'eval/latency_8/demon_move_f1': 0.779921848063441, 'eval/latency_8/demon_cc_f1': 0.9020232347067978, 'timing/data': 0.010310404933989048, 'timing/model': 1.1157983280718327, 'learning_rate/qwen_vl_interface': 4.9258485365178135e-06, 'learning_rate/action_model': 4.925848536517814e-05, 'epoch': 1.68}) +[2026-06-17 22:33:04,959][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_3700_state +[2026-06-17 22:33:04,964][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 22:33:11,581][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_3700_state/model.safetensors +[2026-06-17 22:33:27,174][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_3700_state/optimizer.bin +[2026-06-17 22:33:27,181][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_3700_state/sampler.bin +[2026-06-17 22:33:27,181][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_3700_state/sampler_1.bin +[2026-06-17 22:33:27,183][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_3700_state/random_states_0.pkl +[2026-06-17 22:33:27,186][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3700 +[2026-06-17 22:33:27,186][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 22:33:27,196][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 22:33:31,069][starVLA.training.train_starvla][INFO] - Step 3701, Loss: {'train/loss': 0.2489173859357834, 'train/grad_norm_pre_clip': 18.518293380737305, 'timing/data': 0.009367210790514946, 'timing/model': 1.1834363406524062, 'learning_rate/qwen_vl_interface': 4.923690921326918e-06, 'learning_rate/action_model': 4.9236909213269175e-05, 'epoch': 1.68}) +[2026-06-17 22:33:34,804][starVLA.training.train_starvla][INFO] - Step 3702, Loss: {'train/loss': 0.15932799875736237, 'train/grad_norm_pre_clip': 12.30373764038086, 'timing/data': 0.011731141246855259, 'timing/model': 1.0661439942196012, 'learning_rate/qwen_vl_interface': 4.921533373780244e-06, 'learning_rate/action_model': 4.921533373780244e-05, 'epoch': 1.68}) +[2026-06-17 22:33:38,379][starVLA.training.train_starvla][INFO] - Step 3703, Loss: {'train/loss': 0.47789084911346436, 'train/grad_norm_pre_clip': 84.25164794921875, 'timing/data': 0.009140245616436005, 'timing/model': 0.9205359071493149, 'learning_rate/qwen_vl_interface': 4.919375894325057e-06, 'learning_rate/action_model': 4.9193758943250566e-05, 'epoch': 1.68}) +[2026-06-17 22:33:42,141][starVLA.training.train_starvla][INFO] - Step 3704, Loss: {'train/loss': 0.5931743383407593, 'train/grad_norm_pre_clip': 29.344837188720703, 'timing/data': 0.011052815243601799, 'timing/model': 0.9301807340234518, 'learning_rate/qwen_vl_interface': 4.917218483408603e-06, 'learning_rate/action_model': 4.9172184834086035e-05, 'epoch': 1.68}) +[2026-06-17 22:33:46,128][starVLA.training.train_starvla][INFO] - Step 3705, Loss: {'train/loss': 0.19979678094387054, 'train/grad_norm_pre_clip': 14.243185043334961, 'timing/data': 0.009982195682823658, 'timing/model': 0.933824835345149, 'learning_rate/qwen_vl_interface': 4.915061141478114e-06, 'learning_rate/action_model': 4.915061141478113e-05, 'epoch': 1.68}) +[2026-06-17 22:33:49,951][starVLA.training.train_starvla][INFO] - Step 3706, Loss: {'train/loss': 0.26801320910453796, 'train/grad_norm_pre_clip': 20.820180892944336, 'timing/data': 0.010328051634132862, 'timing/model': 0.9377104667946696, 'learning_rate/qwen_vl_interface': 4.912903868980812e-06, 'learning_rate/action_model': 4.9129038689808115e-05, 'epoch': 1.68}) +[2026-06-17 22:33:53,716][starVLA.training.train_starvla][INFO] - Step 3707, Loss: {'train/loss': 0.5846230387687683, 'train/grad_norm_pre_clip': 24.295812606811523, 'timing/data': 0.00927919801324606, 'timing/model': 0.9274675231426954, 'learning_rate/qwen_vl_interface': 4.9107466663639005e-06, 'learning_rate/action_model': 4.910746666363901e-05, 'epoch': 1.68}) +[2026-06-17 22:33:57,555][starVLA.training.train_starvla][INFO] - Step 3708, Loss: {'train/loss': 0.4682859778404236, 'train/grad_norm_pre_clip': 24.415231704711914, 'timing/data': 0.009088383987545967, 'timing/model': 1.1069376412779093, 'learning_rate/qwen_vl_interface': 4.9085895340745705e-06, 'learning_rate/action_model': 4.9085895340745705e-05, 'epoch': 1.68}) +[2026-06-17 22:34:01,313][starVLA.training.train_starvla][INFO] - Step 3709, Loss: {'train/loss': 0.3438158929347992, 'train/grad_norm_pre_clip': 21.117504119873047, 'timing/data': 0.015699801966547966, 'timing/model': 1.051411553286016, 'learning_rate/qwen_vl_interface': 4.906432472559996e-06, 'learning_rate/action_model': 4.9064324725599955e-05, 'epoch': 1.68}) +[2026-06-17 22:34:05,071][starVLA.training.train_starvla][INFO] - Step 3710, Loss: {'train/loss': 0.31532055139541626, 'train/grad_norm_pre_clip': 19.42225456237793, 'timing/data': 0.007800453342497349, 'timing/model': 1.0795577000826597, 'learning_rate/qwen_vl_interface': 4.90427548226734e-06, 'learning_rate/action_model': 4.9042754822673396e-05, 'epoch': 1.68}) +[2026-06-17 22:34:08,853][starVLA.training.train_starvla][INFO] - Step 3711, Loss: {'train/loss': 0.22045490145683289, 'train/grad_norm_pre_clip': 17.086488723754883, 'timing/data': 0.009345044381916523, 'timing/model': 0.9222383666783571, 'learning_rate/qwen_vl_interface': 4.902118563643751e-06, 'learning_rate/action_model': 4.902118563643751e-05, 'epoch': 1.68}) +[2026-06-17 22:34:12,945][starVLA.training.train_starvla][INFO] - Step 3712, Loss: {'train/loss': 0.18037420511245728, 'train/grad_norm_pre_clip': 12.87398910522461, 'timing/data': 0.010924186557531357, 'timing/model': 0.9329806268215179, 'learning_rate/qwen_vl_interface': 4.899961717136356e-06, 'learning_rate/action_model': 4.899961717136355e-05, 'epoch': 1.68}) +[2026-06-17 22:34:16,737][starVLA.training.train_starvla][INFO] - Step 3713, Loss: {'train/loss': 0.31210121512413025, 'train/grad_norm_pre_clip': 20.210397720336914, 'timing/data': 0.008014955557882786, 'timing/model': 0.9266287460923195, 'learning_rate/qwen_vl_interface': 4.897804943192273e-06, 'learning_rate/action_model': 4.897804943192273e-05, 'epoch': 1.68}) +[2026-06-17 22:34:20,949][starVLA.training.train_starvla][INFO] - Step 3714, Loss: {'train/loss': 0.3176926076412201, 'train/grad_norm_pre_clip': 22.557817459106445, 'timing/data': 0.01820277702063322, 'timing/model': 0.928675951436162, 'learning_rate/qwen_vl_interface': 4.895648242258605e-06, 'learning_rate/action_model': 4.895648242258605e-05, 'epoch': 1.68}) +[2026-06-17 22:34:24,699][starVLA.training.train_starvla][INFO] - Step 3715, Loss: {'train/loss': 0.06435737013816833, 'train/grad_norm_pre_clip': 5.515146732330322, 'timing/data': 0.010155357420444489, 'timing/model': 1.097141365520656, 'learning_rate/qwen_vl_interface': 4.893491614782438e-06, 'learning_rate/action_model': 4.8934916147824375e-05, 'epoch': 1.69}) +[2026-06-17 22:34:28,277][starVLA.training.train_starvla][INFO] - Step 3716, Loss: {'train/loss': 0.2511042654514313, 'train/grad_norm_pre_clip': 22.921308517456055, 'timing/data': 0.009619532153010368, 'timing/model': 0.9251089794561267, 'learning_rate/qwen_vl_interface': 4.891335061210841e-06, 'learning_rate/action_model': 4.891335061210841e-05, 'epoch': 1.69}) +[2026-06-17 22:34:32,249][starVLA.training.train_starvla][INFO] - Step 3717, Loss: {'train/loss': 0.5273864269256592, 'train/grad_norm_pre_clip': 24.195405960083008, 'timing/data': 0.011504151858389378, 'timing/model': 0.9232500521466136, 'learning_rate/qwen_vl_interface': 4.889178581990874e-06, 'learning_rate/action_model': 4.8891785819908734e-05, 'epoch': 1.69}) +[2026-06-17 22:34:36,181][starVLA.training.train_starvla][INFO] - Step 3718, Loss: {'train/loss': 0.22945280373096466, 'train/grad_norm_pre_clip': 19.047876358032227, 'timing/data': 0.007970570586621761, 'timing/model': 1.1312959361821413, 'learning_rate/qwen_vl_interface': 4.887022177569576e-06, 'learning_rate/action_model': 4.8870221775695754e-05, 'epoch': 1.69}) +[2026-06-17 22:34:39,946][starVLA.training.train_starvla][INFO] - Step 3719, Loss: {'train/loss': 0.3303734064102173, 'train/grad_norm_pre_clip': 18.316558837890625, 'timing/data': 0.008093904703855515, 'timing/model': 0.9347889218479395, 'learning_rate/qwen_vl_interface': 4.88486584839397e-06, 'learning_rate/action_model': 4.8848658483939695e-05, 'epoch': 1.69}) +[2026-06-17 22:34:43,717][starVLA.training.train_starvla][INFO] - Step 3720, Loss: {'train/loss': 0.16063010692596436, 'train/grad_norm_pre_clip': 15.027905464172363, 'timing/data': 0.010037132538855076, 'timing/model': 0.9372472940012813, 'learning_rate/qwen_vl_interface': 4.882709594911065e-06, 'learning_rate/action_model': 4.8827095949110655e-05, 'epoch': 1.69}) +[2026-06-17 22:34:47,495][starVLA.training.train_starvla][INFO] - Step 3721, Loss: {'train/loss': 0.31654173135757446, 'train/grad_norm_pre_clip': 15.855732917785645, 'timing/data': 0.01015471015125513, 'timing/model': 1.104058414697647, 'learning_rate/qwen_vl_interface': 4.880553417567858e-06, 'learning_rate/action_model': 4.880553417567858e-05, 'epoch': 1.69}) +[2026-06-17 22:34:51,189][starVLA.training.train_starvla][INFO] - Step 3722, Loss: {'train/loss': 0.4017719030380249, 'train/grad_norm_pre_clip': 42.18406295776367, 'timing/data': 0.011407879181206226, 'timing/model': 1.0467110136523843, 'learning_rate/qwen_vl_interface': 4.878397316811324e-06, 'learning_rate/action_model': 4.878397316811324e-05, 'epoch': 1.69}) +[2026-06-17 22:34:54,766][starVLA.training.train_starvla][INFO] - Step 3723, Loss: {'train/loss': 0.21378280222415924, 'train/grad_norm_pre_clip': 16.354604721069336, 'timing/data': 0.013675066642463207, 'timing/model': 0.9163352567702532, 'learning_rate/qwen_vl_interface': 4.876241293088428e-06, 'learning_rate/action_model': 4.876241293088428e-05, 'epoch': 1.69}) +[2026-06-17 22:34:58,695][starVLA.training.train_starvla][INFO] - Step 3724, Loss: {'train/loss': 0.6684834957122803, 'train/grad_norm_pre_clip': 30.98908042907715, 'timing/data': 0.009910309687256813, 'timing/model': 0.9355164477601647, 'learning_rate/qwen_vl_interface': 4.874085346846111e-06, 'learning_rate/action_model': 4.87408534684611e-05, 'epoch': 1.69}) +[2026-06-17 22:35:02,492][starVLA.training.train_starvla][INFO] - Step 3725, Loss: {'train/loss': 0.4770532548427582, 'train/grad_norm_pre_clip': 25.56026840209961, 'timing/data': 0.01017834898084402, 'timing/model': 0.9406301937997341, 'learning_rate/qwen_vl_interface': 4.871929478531309e-06, 'learning_rate/action_model': 4.871929478531309e-05, 'epoch': 1.69}) +[2026-06-17 22:35:06,294][starVLA.training.train_starvla][INFO] - Step 3726, Loss: {'train/loss': 0.3879682421684265, 'train/grad_norm_pre_clip': 34.25325393676758, 'timing/data': 0.00911876279860735, 'timing/model': 0.930085132829845, 'learning_rate/qwen_vl_interface': 4.869773688590932e-06, 'learning_rate/action_model': 4.8697736885909325e-05, 'epoch': 1.69}) +[2026-06-17 22:35:10,411][starVLA.training.train_starvla][INFO] - Step 3727, Loss: {'train/loss': 0.3294331729412079, 'train/grad_norm_pre_clip': 23.9531307220459, 'timing/data': 0.009374705143272877, 'timing/model': 1.1598670855164528, 'learning_rate/qwen_vl_interface': 4.867617977471878e-06, 'learning_rate/action_model': 4.867617977471878e-05, 'epoch': 1.69}) +[2026-06-17 22:35:14,174][starVLA.training.train_starvla][INFO] - Step 3728, Loss: {'train/loss': 0.39538097381591797, 'train/grad_norm_pre_clip': 16.247690200805664, 'timing/data': 0.006647111847996712, 'timing/model': 1.0841314727440476, 'learning_rate/qwen_vl_interface': 4.865462345621029e-06, 'learning_rate/action_model': 4.865462345621029e-05, 'epoch': 1.69}) +[2026-06-17 22:35:17,796][starVLA.training.train_starvla][INFO] - Step 3729, Loss: {'train/loss': 0.1601092368364334, 'train/grad_norm_pre_clip': 13.821738243103027, 'timing/data': 0.010321250185370445, 'timing/model': 0.9491167338564992, 'learning_rate/qwen_vl_interface': 4.86330679348525e-06, 'learning_rate/action_model': 4.863306793485249e-05, 'epoch': 1.69}) +[2026-06-17 22:35:21,563][starVLA.training.train_starvla][INFO] - Step 3730, Loss: {'train/loss': 0.36424681544303894, 'train/grad_norm_pre_clip': 22.893274307250977, 'timing/data': 0.010853654704988003, 'timing/model': 0.940757105126977, 'learning_rate/qwen_vl_interface': 4.8611513215113894e-06, 'learning_rate/action_model': 4.861151321511389e-05, 'epoch': 1.69}) +[2026-06-17 22:35:25,332][starVLA.training.train_starvla][INFO] - Step 3731, Loss: {'train/loss': 0.23882007598876953, 'train/grad_norm_pre_clip': 17.371444702148438, 'timing/data': 0.009229928255081177, 'timing/model': 0.931014446541667, 'learning_rate/qwen_vl_interface': 4.858995930146276e-06, 'learning_rate/action_model': 4.858995930146275e-05, 'epoch': 1.69}) +[2026-06-17 22:35:29,064][starVLA.training.train_starvla][INFO] - Step 3732, Loss: {'train/loss': 0.5142517685890198, 'train/grad_norm_pre_clip': 28.5754337310791, 'timing/data': 0.009779703803360462, 'timing/model': 0.9293407266959548, 'learning_rate/qwen_vl_interface': 4.856840619836729e-06, 'learning_rate/action_model': 4.856840619836728e-05, 'epoch': 1.69}) +[2026-06-17 22:35:32,787][starVLA.training.train_starvla][INFO] - Step 3733, Loss: {'train/loss': 0.17661841213703156, 'train/grad_norm_pre_clip': 35.50719451904297, 'timing/data': 0.007779202423989773, 'timing/model': 0.9412440257146955, 'learning_rate/qwen_vl_interface': 4.854685391029543e-06, 'learning_rate/action_model': 4.854685391029543e-05, 'epoch': 1.69}) +[2026-06-17 22:35:36,540][starVLA.training.train_starvla][INFO] - Step 3734, Loss: {'train/loss': 0.4711996912956238, 'train/grad_norm_pre_clip': 25.728649139404297, 'timing/data': 0.01019225176423788, 'timing/model': 1.114269908517599, 'learning_rate/qwen_vl_interface': 4.852530244171502e-06, 'learning_rate/action_model': 4.852530244171502e-05, 'epoch': 1.69}) +[2026-06-17 22:35:40,126][starVLA.training.train_starvla][INFO] - Step 3735, Loss: {'train/loss': 0.1643814742565155, 'train/grad_norm_pre_clip': 16.382022857666016, 'timing/data': 0.007804992608726025, 'timing/model': 0.9310256345197558, 'learning_rate/qwen_vl_interface': 4.850375179709368e-06, 'learning_rate/action_model': 4.8503751797093674e-05, 'epoch': 1.69}) +[2026-06-17 22:35:43,881][starVLA.training.train_starvla][INFO] - Step 3736, Loss: {'train/loss': 0.28322169184684753, 'train/grad_norm_pre_clip': 19.695337295532227, 'timing/data': 0.010446379892528057, 'timing/model': 0.928001381456852, 'learning_rate/qwen_vl_interface': 4.84822019808989e-06, 'learning_rate/action_model': 4.84822019808989e-05, 'epoch': 1.69}) +[2026-06-17 22:35:47,685][starVLA.training.train_starvla][INFO] - Step 3737, Loss: {'train/loss': 0.3650800287723541, 'train/grad_norm_pre_clip': 23.70835304260254, 'timing/data': 0.009443786926567554, 'timing/model': 0.9346502972766757, 'learning_rate/qwen_vl_interface': 4.846065299759799e-06, 'learning_rate/action_model': 4.846065299759799e-05, 'epoch': 1.7}) +[2026-06-17 22:35:51,451][starVLA.training.train_starvla][INFO] - Step 3738, Loss: {'train/loss': 0.37937265634536743, 'train/grad_norm_pre_clip': 19.331514358520508, 'timing/data': 0.011167232878506184, 'timing/model': 0.9330302830785513, 'learning_rate/qwen_vl_interface': 4.843910485165802e-06, 'learning_rate/action_model': 4.843910485165802e-05, 'epoch': 1.7}) +[2026-06-17 22:35:55,225][starVLA.training.train_starvla][INFO] - Step 3739, Loss: {'train/loss': 0.19832506775856018, 'train/grad_norm_pre_clip': 18.796222686767578, 'timing/data': 0.008094054646790028, 'timing/model': 0.9332031225785613, 'learning_rate/qwen_vl_interface': 4.8417557547546015e-06, 'learning_rate/action_model': 4.8417557547546015e-05, 'epoch': 1.7}) +[2026-06-17 22:35:58,953][starVLA.training.train_starvla][INFO] - Step 3740, Loss: {'train/loss': 0.12432819604873657, 'train/grad_norm_pre_clip': 12.913742065429688, 'timing/data': 0.009432910941541195, 'timing/model': 0.9300769446417689, 'learning_rate/qwen_vl_interface': 4.839601108972872e-06, 'learning_rate/action_model': 4.839601108972871e-05, 'epoch': 1.7}) +[2026-06-17 22:36:02,692][starVLA.training.train_starvla][INFO] - Step 3741, Loss: {'train/loss': 0.19812610745429993, 'train/grad_norm_pre_clip': 19.571975708007812, 'timing/data': 0.015071673318743706, 'timing/model': 1.0925561049953103, 'learning_rate/qwen_vl_interface': 4.837446548267275e-06, 'learning_rate/action_model': 4.837446548267274e-05, 'epoch': 1.7}) +[2026-06-17 22:36:06,275][starVLA.training.train_starvla][INFO] - Step 3742, Loss: {'train/loss': 0.45883074402809143, 'train/grad_norm_pre_clip': 21.552095413208008, 'timing/data': 0.0075804246589541435, 'timing/model': 0.9360714638605714, 'learning_rate/qwen_vl_interface': 4.835292073084449e-06, 'learning_rate/action_model': 4.8352920730844486e-05, 'epoch': 1.7}) +[2026-06-17 22:36:10,118][starVLA.training.train_starvla][INFO] - Step 3743, Loss: {'train/loss': 0.3369427025318146, 'train/grad_norm_pre_clip': 20.624752044677734, 'timing/data': 0.009486491791903973, 'timing/model': 0.9351729080080986, 'learning_rate/qwen_vl_interface': 4.833137683871025e-06, 'learning_rate/action_model': 4.833137683871025e-05, 'epoch': 1.7}) +[2026-06-17 22:36:13,838][starVLA.training.train_starvla][INFO] - Step 3744, Loss: {'train/loss': 0.13747526705265045, 'train/grad_norm_pre_clip': 10.898624420166016, 'timing/data': 0.008221677504479885, 'timing/model': 0.9329527532681823, 'learning_rate/qwen_vl_interface': 4.830983381073609e-06, 'learning_rate/action_model': 4.830983381073608e-05, 'epoch': 1.7}) +[2026-06-17 22:36:17,622][starVLA.training.train_starvla][INFO] - Step 3745, Loss: {'train/loss': 0.34648528695106506, 'train/grad_norm_pre_clip': 23.187421798706055, 'timing/data': 0.014078284613788128, 'timing/model': 0.94143230561167, 'learning_rate/qwen_vl_interface': 4.828829165138787e-06, 'learning_rate/action_model': 4.828829165138787e-05, 'epoch': 1.7}) +[2026-06-17 22:36:21,345][starVLA.training.train_starvla][INFO] - Step 3746, Loss: {'train/loss': 0.2574811279773712, 'train/grad_norm_pre_clip': 18.805076599121094, 'timing/data': 0.009368312545120716, 'timing/model': 0.9311261717230082, 'learning_rate/qwen_vl_interface': 4.826675036513132e-06, 'learning_rate/action_model': 4.826675036513132e-05, 'epoch': 1.7}) +[2026-06-17 22:36:25,115][starVLA.training.train_starvla][INFO] - Step 3747, Loss: {'train/loss': 0.3640753924846649, 'train/grad_norm_pre_clip': 21.4054012298584, 'timing/data': 0.010255265049636364, 'timing/model': 1.108612259849906, 'learning_rate/qwen_vl_interface': 4.824520995643198e-06, 'learning_rate/action_model': 4.8245209956431984e-05, 'epoch': 1.7}) +[2026-06-17 22:36:28,789][starVLA.training.train_starvla][INFO] - Step 3748, Loss: {'train/loss': 0.20749132335186005, 'train/grad_norm_pre_clip': 13.163893699645996, 'timing/data': 0.013236838392913342, 'timing/model': 1.0259075118228793, 'learning_rate/qwen_vl_interface': 4.82236704297552e-06, 'learning_rate/action_model': 4.82236704297552e-05, 'epoch': 1.7}) +[2026-06-17 22:36:32,359][starVLA.training.train_starvla][INFO] - Step 3749, Loss: {'train/loss': 0.37637901306152344, 'train/grad_norm_pre_clip': 17.69468879699707, 'timing/data': 0.008296906016767025, 'timing/model': 0.9320831280201674, 'learning_rate/qwen_vl_interface': 4.820213178956613e-06, 'learning_rate/action_model': 4.820213178956612e-05, 'epoch': 1.7}) +[2026-06-17 22:36:36,101][starVLA.training.train_starvla][INFO] - Step 3750, Loss: {'train/loss': 0.3318578898906708, 'train/grad_norm_pre_clip': 17.66461753845215, 'timing/data': 0.008846921846270561, 'timing/model': 0.9345247466117144, 'learning_rate/qwen_vl_interface': 4.818059404032977e-06, 'learning_rate/action_model': 4.818059404032977e-05, 'epoch': 1.7}) +[2026-06-17 22:36:39,847][starVLA.training.train_starvla][INFO] - Step 3751, Loss: {'train/loss': 0.17634768784046173, 'train/grad_norm_pre_clip': 20.52159881591797, 'timing/data': 0.009341733530163765, 'timing/model': 0.93262062035501, 'learning_rate/qwen_vl_interface': 4.815905718651092e-06, 'learning_rate/action_model': 4.815905718651092e-05, 'epoch': 1.7}) +[2026-06-17 22:36:43,586][starVLA.training.train_starvla][INFO] - Step 3752, Loss: {'train/loss': 0.31565061211586, 'train/grad_norm_pre_clip': 16.828754425048828, 'timing/data': 0.011215699836611748, 'timing/model': 0.9238985693082213, 'learning_rate/qwen_vl_interface': 4.813752123257419e-06, 'learning_rate/action_model': 4.813752123257418e-05, 'epoch': 1.7}) +[2026-06-17 22:36:47,309][starVLA.training.train_starvla][INFO] - Step 3753, Loss: {'train/loss': 0.4157741069793701, 'train/grad_norm_pre_clip': 15.625288009643555, 'timing/data': 0.009132626466453075, 'timing/model': 0.9221995109692216, 'learning_rate/qwen_vl_interface': 4.811598618298398e-06, 'learning_rate/action_model': 4.811598618298398e-05, 'epoch': 1.7}) +[2026-06-17 22:36:51,050][starVLA.training.train_starvla][INFO] - Step 3754, Loss: {'train/loss': 0.3405699133872986, 'train/grad_norm_pre_clip': 17.32632064819336, 'timing/data': 0.013383789919316769, 'timing/model': 1.0897298753261566, 'learning_rate/qwen_vl_interface': 4.809445204220457e-06, 'learning_rate/action_model': 4.809445204220457e-05, 'epoch': 1.7}) +[2026-06-17 22:36:54,616][starVLA.training.train_starvla][INFO] - Step 3755, Loss: {'train/loss': 0.4340091049671173, 'train/grad_norm_pre_clip': 21.201051712036133, 'timing/data': 0.00974541436880827, 'timing/model': 0.9208080265671015, 'learning_rate/qwen_vl_interface': 4.8072918814699985e-06, 'learning_rate/action_model': 4.8072918814699986e-05, 'epoch': 1.7}) +[2026-06-17 22:36:58,352][starVLA.training.train_starvla][INFO] - Step 3756, Loss: {'train/loss': 0.2830005884170532, 'train/grad_norm_pre_clip': 15.820869445800781, 'timing/data': 0.007708406075835228, 'timing/model': 0.9218902802094817, 'learning_rate/qwen_vl_interface': 4.80513865049341e-06, 'learning_rate/action_model': 4.80513865049341e-05, 'epoch': 1.7}) +[2026-06-17 22:37:02,102][starVLA.training.train_starvla][INFO] - Step 3757, Loss: {'train/loss': 0.05815267562866211, 'train/grad_norm_pre_clip': 10.015954971313477, 'timing/data': 0.010880980640649796, 'timing/model': 0.9278675932437181, 'learning_rate/qwen_vl_interface': 4.802985511737057e-06, 'learning_rate/action_model': 4.802985511737057e-05, 'epoch': 1.7}) +[2026-06-17 22:37:05,849][starVLA.training.train_starvla][INFO] - Step 3758, Loss: {'train/loss': 0.2501130700111389, 'train/grad_norm_pre_clip': 15.41013240814209, 'timing/data': 0.010799234732985497, 'timing/model': 0.9231550898402929, 'learning_rate/qwen_vl_interface': 4.80083246564729e-06, 'learning_rate/action_model': 4.8008324656472894e-05, 'epoch': 1.7}) +[2026-06-17 22:37:09,595][starVLA.training.train_starvla][INFO] - Step 3759, Loss: {'train/loss': 0.4498388469219208, 'train/grad_norm_pre_clip': 19.037841796875, 'timing/data': 0.010765227489173412, 'timing/model': 0.9254225995391607, 'learning_rate/qwen_vl_interface': 4.798679512670436e-06, 'learning_rate/action_model': 4.798679512670435e-05, 'epoch': 1.71}) +[2026-06-17 22:37:13,352][starVLA.training.train_starvla][INFO] - Step 3760, Loss: {'train/loss': 0.3460330367088318, 'train/grad_norm_pre_clip': 18.18067169189453, 'timing/data': 0.00917817186564207, 'timing/model': 0.9268185803666711, 'learning_rate/qwen_vl_interface': 4.796526653252802e-06, 'learning_rate/action_model': 4.796526653252802e-05, 'epoch': 1.71}) +[2026-06-17 22:37:17,089][starVLA.training.train_starvla][INFO] - Step 3761, Loss: {'train/loss': 0.3246203362941742, 'train/grad_norm_pre_clip': 19.566957473754883, 'timing/data': 0.008468573912978172, 'timing/model': 1.089962461963296, 'learning_rate/qwen_vl_interface': 4.794373887840683e-06, 'learning_rate/action_model': 4.794373887840682e-05, 'epoch': 1.71}) +[2026-06-17 22:37:20,705][starVLA.training.train_starvla][INFO] - Step 3762, Loss: {'train/loss': 0.21679112315177917, 'train/grad_norm_pre_clip': 9.185970306396484, 'timing/data': 0.016907173208892345, 'timing/model': 0.9202102795243263, 'learning_rate/qwen_vl_interface': 4.792221216880347e-06, 'learning_rate/action_model': 4.792221216880347e-05, 'epoch': 1.71}) +[2026-06-17 22:37:24,551][starVLA.training.train_starvla][INFO] - Step 3763, Loss: {'train/loss': 0.27344271540641785, 'train/grad_norm_pre_clip': 17.57465362548828, 'timing/data': 0.009702462702989578, 'timing/model': 0.9218720970675349, 'learning_rate/qwen_vl_interface': 4.790068640818046e-06, 'learning_rate/action_model': 4.790068640818045e-05, 'epoch': 1.71}) +[2026-06-17 22:37:28,253][starVLA.training.train_starvla][INFO] - Step 3764, Loss: {'train/loss': 0.1238030344247818, 'train/grad_norm_pre_clip': 10.870423316955566, 'timing/data': 0.008727306500077248, 'timing/model': 0.9233936481177807, 'learning_rate/qwen_vl_interface': 4.787916160100008e-06, 'learning_rate/action_model': 4.787916160100008e-05, 'epoch': 1.71}) +[2026-06-17 22:37:31,971][starVLA.training.train_starvla][INFO] - Step 3765, Loss: {'train/loss': 0.08354970812797546, 'train/grad_norm_pre_clip': 8.66215705871582, 'timing/data': 0.013509925454854965, 'timing/model': 0.9214308448135853, 'learning_rate/qwen_vl_interface': 4.78576377517245e-06, 'learning_rate/action_model': 4.7857637751724504e-05, 'epoch': 1.71}) +[2026-06-17 22:37:35,717][starVLA.training.train_starvla][INFO] - Step 3766, Loss: {'train/loss': 0.46647515892982483, 'train/grad_norm_pre_clip': 20.507129669189453, 'timing/data': 0.012420300394296646, 'timing/model': 0.9265501974150538, 'learning_rate/qwen_vl_interface': 4.783611486481561e-06, 'learning_rate/action_model': 4.7836114864815604e-05, 'epoch': 1.71}) +[2026-06-17 22:37:39,445][starVLA.training.train_starvla][INFO] - Step 3767, Loss: {'train/loss': 0.2162240743637085, 'train/grad_norm_pre_clip': 19.706880569458008, 'timing/data': 0.010946466587483883, 'timing/model': 1.0828103553503752, 'learning_rate/qwen_vl_interface': 4.781459294473511e-06, 'learning_rate/action_model': 4.781459294473511e-05, 'epoch': 1.71}) +[2026-06-17 22:37:43,114][starVLA.training.train_starvla][INFO] - Step 3768, Loss: {'train/loss': 0.6428328156471252, 'train/grad_norm_pre_clip': 22.345104217529297, 'timing/data': 0.010231654159724712, 'timing/model': 1.0297646569088101, 'learning_rate/qwen_vl_interface': 4.779307199594454e-06, 'learning_rate/action_model': 4.779307199594454e-05, 'epoch': 1.71}) +[2026-06-17 22:37:46,694][starVLA.training.train_starvla][INFO] - Step 3769, Loss: {'train/loss': 0.41861802339553833, 'train/grad_norm_pre_clip': 29.930072784423828, 'timing/data': 0.008589192293584347, 'timing/model': 0.9320901120081544, 'learning_rate/qwen_vl_interface': 4.777155202290522e-06, 'learning_rate/action_model': 4.777155202290521e-05, 'epoch': 1.71}) +[2026-06-17 22:37:50,484][starVLA.training.train_starvla][INFO] - Step 3770, Loss: {'train/loss': 0.163085475564003, 'train/grad_norm_pre_clip': 18.107723236083984, 'timing/data': 0.010333037935197353, 'timing/model': 0.9361481629312038, 'learning_rate/qwen_vl_interface': 4.775003303007824e-06, 'learning_rate/action_model': 4.775003303007824e-05, 'epoch': 1.71}) +[2026-06-17 22:37:54,265][starVLA.training.train_starvla][INFO] - Step 3771, Loss: {'train/loss': 0.2036135047674179, 'train/grad_norm_pre_clip': 21.42778778076172, 'timing/data': 0.013329025357961655, 'timing/model': 0.9143971605226398, 'learning_rate/qwen_vl_interface': 4.772851502192453e-06, 'learning_rate/action_model': 4.772851502192452e-05, 'epoch': 1.71}) +[2026-06-17 22:37:58,003][starVLA.training.train_starvla][INFO] - Step 3772, Loss: {'train/loss': 0.36785688996315, 'train/grad_norm_pre_clip': 25.096359252929688, 'timing/data': 0.008277487941086292, 'timing/model': 0.9203963670879602, 'learning_rate/qwen_vl_interface': 4.770699800290478e-06, 'learning_rate/action_model': 4.770699800290478e-05, 'epoch': 1.71}) +[2026-06-17 22:38:01,756][starVLA.training.train_starvla][INFO] - Step 3773, Loss: {'train/loss': 0.13866950571537018, 'train/grad_norm_pre_clip': 12.069790840148926, 'timing/data': 0.01561452355235815, 'timing/model': 0.9229430947452784, 'learning_rate/qwen_vl_interface': 4.768548197747953e-06, 'learning_rate/action_model': 4.768548197747953e-05, 'epoch': 1.71}) +[2026-06-17 22:38:05,503][starVLA.training.train_starvla][INFO] - Step 3774, Loss: {'train/loss': 0.08234991878271103, 'train/grad_norm_pre_clip': 8.746781349182129, 'timing/data': 0.009755745530128479, 'timing/model': 1.0961811700835824, 'learning_rate/qwen_vl_interface': 4.766396695010903e-06, 'learning_rate/action_model': 4.766396695010903e-05, 'epoch': 1.71}) +[2026-06-17 22:38:09,090][starVLA.training.train_starvla][INFO] - Step 3775, Loss: {'train/loss': 0.3656173348426819, 'train/grad_norm_pre_clip': 26.654151916503906, 'timing/data': 0.008627180010080338, 'timing/model': 0.92385099735111, 'learning_rate/qwen_vl_interface': 4.764245292525337e-06, 'learning_rate/action_model': 4.7642452925253376e-05, 'epoch': 1.71}) +[2026-06-17 22:38:12,853][starVLA.training.train_starvla][INFO] - Step 3776, Loss: {'train/loss': 0.3755822479724884, 'train/grad_norm_pre_clip': 47.68674850463867, 'timing/data': 0.011173073202371597, 'timing/model': 0.9257826218381524, 'learning_rate/qwen_vl_interface': 4.762093990737247e-06, 'learning_rate/action_model': 4.762093990737246e-05, 'epoch': 1.71}) +[2026-06-17 22:38:16,603][starVLA.training.train_starvla][INFO] - Step 3777, Loss: {'train/loss': 0.374336302280426, 'train/grad_norm_pre_clip': 20.361942291259766, 'timing/data': 0.01010901853442192, 'timing/model': 0.9212664347141981, 'learning_rate/qwen_vl_interface': 4.759942790092596e-06, 'learning_rate/action_model': 4.759942790092596e-05, 'epoch': 1.71}) +[2026-06-17 22:38:20,344][starVLA.training.train_starvla][INFO] - Step 3778, Loss: {'train/loss': 0.1653135120868683, 'train/grad_norm_pre_clip': 16.09370994567871, 'timing/data': 0.010008377954363823, 'timing/model': 0.9264507154002786, 'learning_rate/qwen_vl_interface': 4.757791691037332e-06, 'learning_rate/action_model': 4.757791691037332e-05, 'epoch': 1.71}) +[2026-06-17 22:38:24,088][starVLA.training.train_starvla][INFO] - Step 3779, Loss: {'train/loss': 0.4650292694568634, 'train/grad_norm_pre_clip': 21.911758422851562, 'timing/data': 0.01150533463805914, 'timing/model': 0.9221936948597431, 'learning_rate/qwen_vl_interface': 4.755640694017381e-06, 'learning_rate/action_model': 4.755640694017381e-05, 'epoch': 1.71}) +[2026-06-17 22:38:27,821][starVLA.training.train_starvla][INFO] - Step 3780, Loss: {'train/loss': 0.062144193798303604, 'train/grad_norm_pre_clip': 7.230190753936768, 'timing/data': 0.008817797526717186, 'timing/model': 0.922188994474709, 'learning_rate/qwen_vl_interface': 4.753489799478647e-06, 'learning_rate/action_model': 4.7534897994786464e-05, 'epoch': 1.71}) +[2026-06-17 22:38:31,544][starVLA.training.train_starvla][INFO] - Step 3781, Loss: {'train/loss': 0.6305420994758606, 'train/grad_norm_pre_clip': 19.351011276245117, 'timing/data': 0.008992117829620838, 'timing/model': 1.0849431287497282, 'learning_rate/qwen_vl_interface': 4.7513390078670105e-06, 'learning_rate/action_model': 4.75133900786701e-05, 'epoch': 1.72}) +[2026-06-17 22:38:35,111][starVLA.training.train_starvla][INFO] - Step 3782, Loss: {'train/loss': 0.23516745865345, 'train/grad_norm_pre_clip': 15.277485847473145, 'timing/data': 0.010103014297783375, 'timing/model': 0.9231114778667688, 'learning_rate/qwen_vl_interface': 4.749188319628333e-06, 'learning_rate/action_model': 4.749188319628333e-05, 'epoch': 1.72}) +[2026-06-17 22:38:38,847][starVLA.training.train_starvla][INFO] - Step 3783, Loss: {'train/loss': 0.48611247539520264, 'train/grad_norm_pre_clip': 29.019132614135742, 'timing/data': 0.00805059913545847, 'timing/model': 0.920099719427526, 'learning_rate/qwen_vl_interface': 4.747037735208457e-06, 'learning_rate/action_model': 4.7470377352084575e-05, 'epoch': 1.72}) +[2026-06-17 22:38:42,531][starVLA.training.train_starvla][INFO] - Step 3784, Loss: {'train/loss': 0.20854994654655457, 'train/grad_norm_pre_clip': 18.366655349731445, 'timing/data': 0.012313350103795528, 'timing/model': 0.9274209914728999, 'learning_rate/qwen_vl_interface': 4.7448872550532006e-06, 'learning_rate/action_model': 4.7448872550532e-05, 'epoch': 1.72}) +[2026-06-17 22:38:46,249][starVLA.training.train_starvla][INFO] - Step 3785, Loss: {'train/loss': 0.6880213618278503, 'train/grad_norm_pre_clip': 27.968109130859375, 'timing/data': 0.009410201571881771, 'timing/model': 0.9226943906396627, 'learning_rate/qwen_vl_interface': 4.74273687960836e-06, 'learning_rate/action_model': 4.74273687960836e-05, 'epoch': 1.72}) +[2026-06-17 22:38:49,989][starVLA.training.train_starvla][INFO] - Step 3786, Loss: {'train/loss': 0.18136700987815857, 'train/grad_norm_pre_clip': 22.352954864501953, 'timing/data': 0.009292928501963615, 'timing/model': 0.9238996300846338, 'learning_rate/qwen_vl_interface': 4.7405866093197075e-06, 'learning_rate/action_model': 4.740586609319707e-05, 'epoch': 1.72}) +[2026-06-17 22:38:53,724][starVLA.training.train_starvla][INFO] - Step 3787, Loss: {'train/loss': 0.4450589418411255, 'train/grad_norm_pre_clip': 21.49473762512207, 'timing/data': 0.010130099020898342, 'timing/model': 1.0798660228028893, 'learning_rate/qwen_vl_interface': 4.738436444633002e-06, 'learning_rate/action_model': 4.738436444633001e-05, 'epoch': 1.72}) +[2026-06-17 22:38:57,398][starVLA.training.train_starvla][INFO] - Step 3788, Loss: {'train/loss': 0.36905941367149353, 'train/grad_norm_pre_clip': 25.127731323242188, 'timing/data': 0.009312473237514496, 'timing/model': 1.0283077508211136, 'learning_rate/qwen_vl_interface': 4.73628638599397e-06, 'learning_rate/action_model': 4.73628638599397e-05, 'epoch': 1.72}) +[2026-06-17 22:39:00,978][starVLA.training.train_starvla][INFO] - Step 3789, Loss: {'train/loss': 0.24769145250320435, 'train/grad_norm_pre_clip': 23.24126625061035, 'timing/data': 0.011491318233311176, 'timing/model': 0.9238007264211774, 'learning_rate/qwen_vl_interface': 4.734136433848323e-06, 'learning_rate/action_model': 4.7341364338483224e-05, 'epoch': 1.72}) +[2026-06-17 22:39:04,710][starVLA.training.train_starvla][INFO] - Step 3790, Loss: {'train/loss': 0.31646937131881714, 'train/grad_norm_pre_clip': 19.334930419921875, 'timing/data': 0.010887366719543934, 'timing/model': 0.9261828297749162, 'learning_rate/qwen_vl_interface': 4.731986588641748e-06, 'learning_rate/action_model': 4.7319865886417484e-05, 'epoch': 1.72}) +[2026-06-17 22:39:08,455][starVLA.training.train_starvla][INFO] - Step 3791, Loss: {'train/loss': 0.21870797872543335, 'train/grad_norm_pre_clip': 14.093422889709473, 'timing/data': 0.008585712872445583, 'timing/model': 0.9250226346775889, 'learning_rate/qwen_vl_interface': 4.729836850819912e-06, 'learning_rate/action_model': 4.729836850819912e-05, 'epoch': 1.72}) +[2026-06-17 22:39:12,201][starVLA.training.train_starvla][INFO] - Step 3792, Loss: {'train/loss': 0.27845343947410583, 'train/grad_norm_pre_clip': 22.88398551940918, 'timing/data': 0.007891588844358921, 'timing/model': 0.9259214587509632, 'learning_rate/qwen_vl_interface': 4.727687220828458e-06, 'learning_rate/action_model': 4.727687220828458e-05, 'epoch': 1.72}) +[2026-06-17 22:39:15,935][starVLA.training.train_starvla][INFO] - Step 3793, Loss: {'train/loss': 0.5139700770378113, 'train/grad_norm_pre_clip': 23.564151763916016, 'timing/data': 0.009089524857699871, 'timing/model': 0.9240646129474044, 'learning_rate/qwen_vl_interface': 4.7255376991130005e-06, 'learning_rate/action_model': 4.7255376991130005e-05, 'epoch': 1.72}) +[2026-06-17 22:39:19,660][starVLA.training.train_starvla][INFO] - Step 3794, Loss: {'train/loss': 0.5892546772956848, 'train/grad_norm_pre_clip': 27.510683059692383, 'timing/data': 0.012609953992068768, 'timing/model': 1.0801841337233782, 'learning_rate/qwen_vl_interface': 4.723388286119147e-06, 'learning_rate/action_model': 4.723388286119146e-05, 'epoch': 1.72}) +[2026-06-17 22:39:23,239][starVLA.training.train_starvla][INFO] - Step 3795, Loss: {'train/loss': 0.5127375721931458, 'train/grad_norm_pre_clip': 33.9027099609375, 'timing/data': 0.00869381707161665, 'timing/model': 0.9296993520110846, 'learning_rate/qwen_vl_interface': 4.721238982292467e-06, 'learning_rate/action_model': 4.721238982292466e-05, 'epoch': 1.72}) +[2026-06-17 22:39:26,970][starVLA.training.train_starvla][INFO] - Step 3796, Loss: {'train/loss': 0.3008665144443512, 'train/grad_norm_pre_clip': 19.225950241088867, 'timing/data': 0.008943343535065651, 'timing/model': 0.9180662026628852, 'learning_rate/qwen_vl_interface': 4.719089788078514e-06, 'learning_rate/action_model': 4.719089788078514e-05, 'epoch': 1.72}) +[2026-06-17 22:39:30,741][starVLA.training.train_starvla][INFO] - Step 3797, Loss: {'train/loss': 0.3842657804489136, 'train/grad_norm_pre_clip': 19.24190902709961, 'timing/data': 0.008934753946959972, 'timing/model': 0.9267890052869916, 'learning_rate/qwen_vl_interface': 4.716940703922818e-06, 'learning_rate/action_model': 4.716940703922818e-05, 'epoch': 1.72}) +[2026-06-17 22:39:34,480][starVLA.training.train_starvla][INFO] - Step 3798, Loss: {'train/loss': 0.12003428488969803, 'train/grad_norm_pre_clip': 10.011191368103027, 'timing/data': 0.00848960131406784, 'timing/model': 0.9257797775790095, 'learning_rate/qwen_vl_interface': 4.714791730270889e-06, 'learning_rate/action_model': 4.71479173027089e-05, 'epoch': 1.72}) +[2026-06-17 22:39:38,326][starVLA.training.train_starvla][INFO] - Step 3799, Loss: {'train/loss': 0.4280043840408325, 'train/grad_norm_pre_clip': 18.07441520690918, 'timing/data': 0.00947433803230524, 'timing/model': 0.9553510546684265, 'learning_rate/qwen_vl_interface': 4.712642867568213e-06, 'learning_rate/action_model': 4.712642867568212e-05, 'epoch': 1.72}) +[2026-06-17 22:41:38,293][starVLA.training.train_starvla][INFO] - Step 3800, Loss: {'train/loss': 0.14814594388008118, 'train/grad_norm_pre_clip': 13.784464836120605, 'eval/loss': 0.3114781379699707, 'eval/latency_8/loss': 0.3114781379699707, 'eval/demon_fire_precision': 0.9514242878560719, 'eval/demon_fire_recall': 0.9857098477788133, 'eval/demon_fire_f1': 0.968263655782728, 'eval/demon_fire_rate_model': 0.927935447968837, 'eval/demon_fire_rate_teacher': 0.8956594323873122, 'eval/demon_left_precision': 0.8452380952380952, 'eval/demon_left_recall': 0.7601713062098501, 'eval/demon_left_f1': 0.8004509582863584, 'eval/demon_left_rate_model': 0.11686143572621036, 'eval/demon_left_rate_teacher': 0.12993878686700056, 'eval/demon_right_precision': 0.7908961593172119, 'eval/demon_right_recall': 0.8424242424242424, 'eval/demon_right_f1': 0.8158473954512107, 'eval/demon_right_rate_model': 0.19560378408458542, 'eval/demon_right_rate_teacher': 0.18363939899833054, 'eval/demon_move_f1': 0.8081491768687845, 'eval/demon_cc_f1': 0.913941296039582, 'eval/latency_8/demon_fire_precision': 0.9514242878560719, 'eval/latency_8/demon_fire_recall': 0.9857098477788133, 'eval/latency_8/demon_fire_f1': 0.968263655782728, 'eval/latency_8/demon_fire_rate_model': 0.927935447968837, 'eval/latency_8/demon_fire_rate_teacher': 0.8956594323873122, 'eval/latency_8/demon_left_precision': 0.8452380952380952, 'eval/latency_8/demon_left_recall': 0.7601713062098501, 'eval/latency_8/demon_left_f1': 0.8004509582863584, 'eval/latency_8/demon_left_rate_model': 0.11686143572621036, 'eval/latency_8/demon_left_rate_teacher': 0.12993878686700056, 'eval/latency_8/demon_right_precision': 0.7908961593172119, 'eval/latency_8/demon_right_recall': 0.8424242424242424, 'eval/latency_8/demon_right_f1': 0.8158473954512107, 'eval/latency_8/demon_right_rate_model': 0.19560378408458542, 'eval/latency_8/demon_right_rate_teacher': 0.18363939899833054, 'eval/latency_8/demon_move_f1': 0.8081491768687845, 'eval/latency_8/demon_cc_f1': 0.913941296039582, 'timing/data': 0.007907401770353317, 'timing/model': 1.0328570157289505, 'learning_rate/qwen_vl_interface': 4.7104941162602436e-06, 'learning_rate/action_model': 4.710494116260244e-05, 'epoch': 1.72}) +[2026-06-17 22:41:38,295][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_3800_state +[2026-06-17 22:41:38,299][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 22:41:44,958][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_3800_state/model.safetensors +[2026-06-17 22:42:00,281][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_3800_state/optimizer.bin +[2026-06-17 22:42:00,284][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_3800_state/sampler.bin +[2026-06-17 22:42:00,284][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_3800_state/sampler_1.bin +[2026-06-17 22:42:00,286][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_3800_state/random_states_0.pkl +[2026-06-17 22:42:00,289][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3800 +[2026-06-17 22:42:00,289][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 22:42:00,301][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 22:42:04,205][starVLA.training.train_starvla][INFO] - Step 3801, Loss: {'train/loss': 0.4583096504211426, 'train/grad_norm_pre_clip': 23.637523651123047, 'timing/data': 0.007636993192136288, 'timing/model': 0.9222940253093839, 'learning_rate/qwen_vl_interface': 4.708345476792428e-06, 'learning_rate/action_model': 4.7083454767924276e-05, 'epoch': 1.72}) +[2026-06-17 22:42:08,004][starVLA.training.train_starvla][INFO] - Step 3802, Loss: {'train/loss': 0.1699160933494568, 'train/grad_norm_pre_clip': 13.532194137573242, 'timing/data': 0.010211480781435966, 'timing/model': 0.9264414515346289, 'learning_rate/qwen_vl_interface': 4.706196949610178e-06, 'learning_rate/action_model': 4.7061969496101776e-05, 'epoch': 1.72}) +[2026-06-17 22:42:11,796][starVLA.training.train_starvla][INFO] - Step 3803, Loss: {'train/loss': 0.5007616877555847, 'train/grad_norm_pre_clip': 23.780851364135742, 'timing/data': 0.009230056777596474, 'timing/model': 1.0968529311940074, 'learning_rate/qwen_vl_interface': 4.7040485351588845e-06, 'learning_rate/action_model': 4.7040485351588845e-05, 'epoch': 1.73}) +[2026-06-17 22:42:15,489][starVLA.training.train_starvla][INFO] - Step 3804, Loss: {'train/loss': 0.2923097610473633, 'train/grad_norm_pre_clip': 14.53921127319336, 'timing/data': 0.009026318788528442, 'timing/model': 1.0464704260230064, 'learning_rate/qwen_vl_interface': 4.701900233883916e-06, 'learning_rate/action_model': 4.7019002338839156e-05, 'epoch': 1.73}) +[2026-06-17 22:42:19,078][starVLA.training.train_starvla][INFO] - Step 3805, Loss: {'train/loss': 0.19255781173706055, 'train/grad_norm_pre_clip': 16.4564151763916, 'timing/data': 0.009775836020708084, 'timing/model': 0.9162324741482735, 'learning_rate/qwen_vl_interface': 4.69975204623062e-06, 'learning_rate/action_model': 4.69975204623062e-05, 'epoch': 1.73}) +[2026-06-17 22:42:22,804][starVLA.training.train_starvla][INFO] - Step 3806, Loss: {'train/loss': 0.1970348358154297, 'train/grad_norm_pre_clip': 13.608370780944824, 'timing/data': 0.009529809467494488, 'timing/model': 0.9262415803968906, 'learning_rate/qwen_vl_interface': 4.697603972644319e-06, 'learning_rate/action_model': 4.697603972644318e-05, 'epoch': 1.73}) +[2026-06-17 22:42:26,555][starVLA.training.train_starvla][INFO] - Step 3807, Loss: {'train/loss': 0.24507664144039154, 'train/grad_norm_pre_clip': 14.537896156311035, 'timing/data': 0.010261335410177708, 'timing/model': 0.9226228157058358, 'learning_rate/qwen_vl_interface': 4.695456013570306e-06, 'learning_rate/action_model': 4.695456013570306e-05, 'epoch': 1.73}) +[2026-06-17 22:42:30,349][starVLA.training.train_starvla][INFO] - Step 3808, Loss: {'train/loss': 0.2729906737804413, 'train/grad_norm_pre_clip': 22.184877395629883, 'timing/data': 0.008791468106210232, 'timing/model': 0.9238631138578057, 'learning_rate/qwen_vl_interface': 4.693308169453858e-06, 'learning_rate/action_model': 4.693308169453858e-05, 'epoch': 1.73}) +[2026-06-17 22:42:34,119][starVLA.training.train_starvla][INFO] - Step 3809, Loss: {'train/loss': 0.28750908374786377, 'train/grad_norm_pre_clip': 13.118148803710938, 'timing/data': 0.008733918890357018, 'timing/model': 0.9305429290980101, 'learning_rate/qwen_vl_interface': 4.691160440740226e-06, 'learning_rate/action_model': 4.691160440740226e-05, 'epoch': 1.73}) +[2026-06-17 22:42:37,909][starVLA.training.train_starvla][INFO] - Step 3810, Loss: {'train/loss': 0.3304947316646576, 'train/grad_norm_pre_clip': 24.241952896118164, 'timing/data': 0.01064863707870245, 'timing/model': 1.1035654237493873, 'learning_rate/qwen_vl_interface': 4.689012827874637e-06, 'learning_rate/action_model': 4.689012827874637e-05, 'epoch': 1.73}) +[2026-06-17 22:42:41,495][starVLA.training.train_starvla][INFO] - Step 3811, Loss: {'train/loss': 0.13908037543296814, 'train/grad_norm_pre_clip': 17.153745651245117, 'timing/data': 0.009270314127206802, 'timing/model': 0.9307092735543847, 'learning_rate/qwen_vl_interface': 4.6868653313022905e-06, 'learning_rate/action_model': 4.68686533130229e-05, 'epoch': 1.73}) +[2026-06-17 22:42:45,311][starVLA.training.train_starvla][INFO] - Step 3812, Loss: {'train/loss': 0.13421547412872314, 'train/grad_norm_pre_clip': 10.928229331970215, 'timing/data': 0.009328325279057026, 'timing/model': 0.9341859240084887, 'learning_rate/qwen_vl_interface': 4.684717951468368e-06, 'learning_rate/action_model': 4.6847179514683684e-05, 'epoch': 1.73}) +[2026-06-17 22:42:49,103][starVLA.training.train_starvla][INFO] - Step 3813, Loss: {'train/loss': 0.5641016960144043, 'train/grad_norm_pre_clip': 25.214204788208008, 'timing/data': 0.008584484457969666, 'timing/model': 0.9232723535969853, 'learning_rate/qwen_vl_interface': 4.682570688818024e-06, 'learning_rate/action_model': 4.682570688818024e-05, 'epoch': 1.73}) +[2026-06-17 22:42:52,881][starVLA.training.train_starvla][INFO] - Step 3814, Loss: {'train/loss': 0.3886713683605194, 'train/grad_norm_pre_clip': 16.219730377197266, 'timing/data': 0.013103088364005089, 'timing/model': 0.9224962713196874, 'learning_rate/qwen_vl_interface': 4.680423543796385e-06, 'learning_rate/action_model': 4.680423543796385e-05, 'epoch': 1.73}) +[2026-06-17 22:42:56,860][starVLA.training.train_starvla][INFO] - Step 3815, Loss: {'train/loss': 0.12686851620674133, 'train/grad_norm_pre_clip': 10.938885688781738, 'timing/data': 0.009223494678735733, 'timing/model': 0.9316392876207829, 'learning_rate/qwen_vl_interface': 4.678276516848558e-06, 'learning_rate/action_model': 4.678276516848557e-05, 'epoch': 1.73}) +[2026-06-17 22:43:00,873][starVLA.training.train_starvla][INFO] - Step 3816, Loss: {'train/loss': 0.4382023215293884, 'train/grad_norm_pre_clip': 22.138065338134766, 'timing/data': 0.00720601249486208, 'timing/model': 1.352385688573122, 'learning_rate/qwen_vl_interface': 4.676129608419625e-06, 'learning_rate/action_model': 4.676129608419625e-05, 'epoch': 1.73}) +[2026-06-17 22:43:04,475][starVLA.training.train_starvla][INFO] - Step 3817, Loss: {'train/loss': 0.42732423543930054, 'train/grad_norm_pre_clip': 24.989269256591797, 'timing/data': 0.009912324137985706, 'timing/model': 0.9282185081392527, 'learning_rate/qwen_vl_interface': 4.6739828189546415e-06, 'learning_rate/action_model': 4.6739828189546417e-05, 'epoch': 1.73}) +[2026-06-17 22:43:08,248][starVLA.training.train_starvla][INFO] - Step 3818, Loss: {'train/loss': 0.1068919375538826, 'train/grad_norm_pre_clip': 10.735560417175293, 'timing/data': 0.012440699152648449, 'timing/model': 0.9277680721133947, 'learning_rate/qwen_vl_interface': 4.671836148898641e-06, 'learning_rate/action_model': 4.6718361488986404e-05, 'epoch': 1.73}) +[2026-06-17 22:43:12,020][starVLA.training.train_starvla][INFO] - Step 3819, Loss: {'train/loss': 0.23282910883426666, 'train/grad_norm_pre_clip': 13.181888580322266, 'timing/data': 0.009775137528777122, 'timing/model': 0.9332315362989902, 'learning_rate/qwen_vl_interface': 4.6696895986966254e-06, 'learning_rate/action_model': 4.669689598696625e-05, 'epoch': 1.73}) +[2026-06-17 22:43:15,804][starVLA.training.train_starvla][INFO] - Step 3820, Loss: {'train/loss': 0.3447452783584595, 'train/grad_norm_pre_clip': 31.185022354125977, 'timing/data': 0.008215000852942467, 'timing/model': 0.9288834417238832, 'learning_rate/qwen_vl_interface': 4.667543168793583e-06, 'learning_rate/action_model': 4.667543168793583e-05, 'epoch': 1.73}) +[2026-06-17 22:43:19,567][starVLA.training.train_starvla][INFO] - Step 3821, Loss: {'train/loss': 0.386769562959671, 'train/grad_norm_pre_clip': 25.81707000732422, 'timing/data': 0.009642533957958221, 'timing/model': 0.9199908310547471, 'learning_rate/qwen_vl_interface': 4.665396859634468e-06, 'learning_rate/action_model': 4.665396859634468e-05, 'epoch': 1.73}) +[2026-06-17 22:43:23,288][starVLA.training.train_starvla][INFO] - Step 3822, Loss: {'train/loss': 0.1255306750535965, 'train/grad_norm_pre_clip': 17.282794952392578, 'timing/data': 0.007696538232266903, 'timing/model': 0.937672239728272, 'learning_rate/qwen_vl_interface': 4.663250671664212e-06, 'learning_rate/action_model': 4.663250671664211e-05, 'epoch': 1.73}) +[2026-06-17 22:43:26,999][starVLA.training.train_starvla][INFO] - Step 3823, Loss: {'train/loss': 0.3110828995704651, 'train/grad_norm_pre_clip': 19.731433868408203, 'timing/data': 0.011798880062997341, 'timing/model': 1.0585013693198562, 'learning_rate/qwen_vl_interface': 4.6611046053277245e-06, 'learning_rate/action_model': 4.661104605327724e-05, 'epoch': 1.73}) +[2026-06-17 22:43:30,581][starVLA.training.train_starvla][INFO] - Step 3824, Loss: {'train/loss': 0.5942156910896301, 'train/grad_norm_pre_clip': 27.925331115722656, 'timing/data': 0.011586163192987442, 'timing/model': 0.9159151278436184, 'learning_rate/qwen_vl_interface': 4.658958661069886e-06, 'learning_rate/action_model': 4.6589586610698856e-05, 'epoch': 1.73}) +[2026-06-17 22:43:34,606][starVLA.training.train_starvla][INFO] - Step 3825, Loss: {'train/loss': 0.27661454677581787, 'train/grad_norm_pre_clip': 15.970293998718262, 'timing/data': 0.010967491194605827, 'timing/model': 0.9272658936679363, 'learning_rate/qwen_vl_interface': 4.6568128393355535e-06, 'learning_rate/action_model': 4.656812839335553e-05, 'epoch': 1.74}) +[2026-06-17 22:43:38,446][starVLA.training.train_starvla][INFO] - Step 3826, Loss: {'train/loss': 0.42773061990737915, 'train/grad_norm_pre_clip': 21.521549224853516, 'timing/data': 0.019425205886363983, 'timing/model': 0.9663199642673135, 'learning_rate/qwen_vl_interface': 4.654667140569555e-06, 'learning_rate/action_model': 4.6546671405695544e-05, 'epoch': 1.74}) +[2026-06-17 22:43:42,289][starVLA.training.train_starvla][INFO] - Step 3827, Loss: {'train/loss': 0.2366943508386612, 'train/grad_norm_pre_clip': 18.191192626953125, 'timing/data': 0.010451363399624825, 'timing/model': 0.985422819852829, 'learning_rate/qwen_vl_interface': 4.652521565216701e-06, 'learning_rate/action_model': 4.652521565216701e-05, 'epoch': 1.74}) +[2026-06-17 22:43:46,106][starVLA.training.train_starvla][INFO] - Step 3828, Loss: {'train/loss': 0.262139230966568, 'train/grad_norm_pre_clip': 19.985143661499023, 'timing/data': 0.008160353638231754, 'timing/model': 0.9355859439820051, 'learning_rate/qwen_vl_interface': 4.65037611372177e-06, 'learning_rate/action_model': 4.650376113721769e-05, 'epoch': 1.74}) +[2026-06-17 22:43:49,881][starVLA.training.train_starvla][INFO] - Step 3829, Loss: {'train/loss': 0.25041642785072327, 'train/grad_norm_pre_clip': 19.963781356811523, 'timing/data': 0.009295282885432243, 'timing/model': 1.1156526347622275, 'learning_rate/qwen_vl_interface': 4.648230786529514e-06, 'learning_rate/action_model': 4.648230786529513e-05, 'epoch': 1.74}) +[2026-06-17 22:43:53,471][starVLA.training.train_starvla][INFO] - Step 3830, Loss: {'train/loss': 0.412036657333374, 'train/grad_norm_pre_clip': 25.31814193725586, 'timing/data': 0.013189089484512806, 'timing/model': 0.9255636623129249, 'learning_rate/qwen_vl_interface': 4.646085584084664e-06, 'learning_rate/action_model': 4.6460855840846637e-05, 'epoch': 1.74}) +[2026-06-17 22:43:57,394][starVLA.training.train_starvla][INFO] - Step 3831, Loss: {'train/loss': 0.3930725157260895, 'train/grad_norm_pre_clip': 25.75665283203125, 'timing/data': 0.008677156642079353, 'timing/model': 0.9219828164204955, 'learning_rate/qwen_vl_interface': 4.643940506831923e-06, 'learning_rate/action_model': 4.6439405068319226e-05, 'epoch': 1.74}) +[2026-06-17 22:44:01,200][starVLA.training.train_starvla][INFO] - Step 3832, Loss: {'train/loss': 0.32654714584350586, 'train/grad_norm_pre_clip': 19.151140213012695, 'timing/data': 0.007047275081276894, 'timing/model': 0.982461697421968, 'learning_rate/qwen_vl_interface': 4.641795555215967e-06, 'learning_rate/action_model': 4.6417955552159676e-05, 'epoch': 1.74}) +[2026-06-17 22:44:05,179][starVLA.training.train_starvla][INFO] - Step 3833, Loss: {'train/loss': 0.4986152946949005, 'train/grad_norm_pre_clip': 21.239503860473633, 'timing/data': 0.008300933055579662, 'timing/model': 0.9394772201776505, 'learning_rate/qwen_vl_interface': 4.639650729681445e-06, 'learning_rate/action_model': 4.6396507296814454e-05, 'epoch': 1.74}) +[2026-06-17 22:44:08,970][starVLA.training.train_starvla][INFO] - Step 3834, Loss: {'train/loss': 0.532752275466919, 'train/grad_norm_pre_clip': 31.67677879333496, 'timing/data': 0.00808460358530283, 'timing/model': 0.930936748161912, 'learning_rate/qwen_vl_interface': 4.637506030672987e-06, 'learning_rate/action_model': 4.637506030672987e-05, 'epoch': 1.74}) +[2026-06-17 22:44:13,360][starVLA.training.train_starvla][INFO] - Step 3835, Loss: {'train/loss': 0.041673533618450165, 'train/grad_norm_pre_clip': 6.313006401062012, 'timing/data': 0.010578995570540428, 'timing/model': 1.1256875488907099, 'learning_rate/qwen_vl_interface': 4.635361458635186e-06, 'learning_rate/action_model': 4.6353614586351864e-05, 'epoch': 1.74}) +[2026-06-17 22:44:17,556][starVLA.training.train_starvla][INFO] - Step 3836, Loss: {'train/loss': 0.31607604026794434, 'train/grad_norm_pre_clip': 22.37143898010254, 'timing/data': 0.010482718236744404, 'timing/model': 1.1094367131590843, 'learning_rate/qwen_vl_interface': 4.633217014012616e-06, 'learning_rate/action_model': 4.633217014012616e-05, 'epoch': 1.74}) +[2026-06-17 22:44:21,157][starVLA.training.train_starvla][INFO] - Step 3837, Loss: {'train/loss': 0.4463787078857422, 'train/grad_norm_pre_clip': 30.15446662902832, 'timing/data': 0.010149963200092316, 'timing/model': 0.9232617523521185, 'learning_rate/qwen_vl_interface': 4.631072697249822e-06, 'learning_rate/action_model': 4.631072697249822e-05, 'epoch': 1.74}) +[2026-06-17 22:44:24,935][starVLA.training.train_starvla][INFO] - Step 3838, Loss: {'train/loss': 0.3624575436115265, 'train/grad_norm_pre_clip': 25.144123077392578, 'timing/data': 0.00984916090965271, 'timing/model': 0.9278161134570837, 'learning_rate/qwen_vl_interface': 4.628928508791325e-06, 'learning_rate/action_model': 4.628928508791324e-05, 'epoch': 1.74}) +[2026-06-17 22:44:28,706][starVLA.training.train_starvla][INFO] - Step 3839, Loss: {'train/loss': 0.18179550766944885, 'train/grad_norm_pre_clip': 14.940956115722656, 'timing/data': 0.009640160948038101, 'timing/model': 0.9275466781109571, 'learning_rate/qwen_vl_interface': 4.626784449081617e-06, 'learning_rate/action_model': 4.626784449081617e-05, 'epoch': 1.74}) +[2026-06-17 22:44:32,904][starVLA.training.train_starvla][INFO] - Step 3840, Loss: {'train/loss': 0.1596015989780426, 'train/grad_norm_pre_clip': 15.585723876953125, 'timing/data': 0.014037352986633778, 'timing/model': 0.9342353697866201, 'learning_rate/qwen_vl_interface': 4.62464051856516e-06, 'learning_rate/action_model': 4.6246405185651596e-05, 'epoch': 1.74}) +[2026-06-17 22:44:36,695][starVLA.training.train_starvla][INFO] - Step 3841, Loss: {'train/loss': 0.3792019486427307, 'train/grad_norm_pre_clip': 25.038301467895508, 'timing/data': 0.010229872539639473, 'timing/model': 0.9334874460473657, 'learning_rate/qwen_vl_interface': 4.622496717686401e-06, 'learning_rate/action_model': 4.6224967176864005e-05, 'epoch': 1.74}) +[2026-06-17 22:44:40,517][starVLA.training.train_starvla][INFO] - Step 3842, Loss: {'train/loss': 0.38056477904319763, 'train/grad_norm_pre_clip': 19.163366317749023, 'timing/data': 0.010142761282622814, 'timing/model': 1.1129974145442247, 'learning_rate/qwen_vl_interface': 4.620353046889745e-06, 'learning_rate/action_model': 4.620353046889745e-05, 'epoch': 1.74}) +[2026-06-17 22:44:44,675][starVLA.training.train_starvla][INFO] - Step 3843, Loss: {'train/loss': 0.4278518259525299, 'train/grad_norm_pre_clip': 27.446855545043945, 'timing/data': 0.008465329185128212, 'timing/model': 1.4391032764688134, 'learning_rate/qwen_vl_interface': 4.6182095066195805e-06, 'learning_rate/action_model': 4.61820950661958e-05, 'epoch': 1.74}) +[2026-06-17 22:44:48,284][starVLA.training.train_starvla][INFO] - Step 3844, Loss: {'train/loss': 0.2773607075214386, 'train/grad_norm_pre_clip': 27.33160400390625, 'timing/data': 0.007455693557858467, 'timing/model': 0.9235538514330983, 'learning_rate/qwen_vl_interface': 4.616066097320264e-06, 'learning_rate/action_model': 4.6160660973202636e-05, 'epoch': 1.74}) +[2026-06-17 22:44:52,142][starVLA.training.train_starvla][INFO] - Step 3845, Loss: {'train/loss': 0.42127200961112976, 'train/grad_norm_pre_clip': 28.695362091064453, 'timing/data': 0.02185662928968668, 'timing/model': 0.9287763023748994, 'learning_rate/qwen_vl_interface': 4.613922819436129e-06, 'learning_rate/action_model': 4.613922819436128e-05, 'epoch': 1.74}) +[2026-06-17 22:44:55,900][starVLA.training.train_starvla][INFO] - Step 3846, Loss: {'train/loss': 0.45357462763786316, 'train/grad_norm_pre_clip': 20.044767379760742, 'timing/data': 0.007379321381449699, 'timing/model': 0.9223802890628576, 'learning_rate/qwen_vl_interface': 4.611779673411478e-06, 'learning_rate/action_model': 4.6117796734114785e-05, 'epoch': 1.74}) +[2026-06-17 22:44:59,769][starVLA.training.train_starvla][INFO] - Step 3847, Loss: {'train/loss': 0.12397225946187973, 'train/grad_norm_pre_clip': 11.347142219543457, 'timing/data': 0.007124530151486397, 'timing/model': 0.9211212573572993, 'learning_rate/qwen_vl_interface': 4.609636659690588e-06, 'learning_rate/action_model': 4.609636659690587e-05, 'epoch': 1.75}) +[2026-06-17 22:45:03,530][starVLA.training.train_starvla][INFO] - Step 3848, Loss: {'train/loss': 0.3201786279678345, 'train/grad_norm_pre_clip': 24.99612045288086, 'timing/data': 0.007445916533470154, 'timing/model': 0.9272469785064459, 'learning_rate/qwen_vl_interface': 4.607493778717705e-06, 'learning_rate/action_model': 4.6074937787177055e-05, 'epoch': 1.75}) +[2026-06-17 22:45:07,466][starVLA.training.train_starvla][INFO] - Step 3849, Loss: {'train/loss': 0.35354650020599365, 'train/grad_norm_pre_clip': 28.066614151000977, 'timing/data': 0.006978233344852924, 'timing/model': 1.2701580012217164, 'learning_rate/qwen_vl_interface': 4.605351030937054e-06, 'learning_rate/action_model': 4.605351030937055e-05, 'epoch': 1.75}) +[2026-06-17 22:45:11,080][starVLA.training.train_starvla][INFO] - Step 3850, Loss: {'train/loss': 0.7021939754486084, 'train/grad_norm_pre_clip': 30.997709274291992, 'timing/data': 0.010996418073773384, 'timing/model': 0.9382509393617511, 'learning_rate/qwen_vl_interface': 4.60320841679283e-06, 'learning_rate/action_model': 4.60320841679283e-05, 'epoch': 1.75}) +[2026-06-17 22:45:14,834][starVLA.training.train_starvla][INFO] - Step 3851, Loss: {'train/loss': 0.19406633079051971, 'train/grad_norm_pre_clip': 20.125152587890625, 'timing/data': 0.011783276684582233, 'timing/model': 0.9276677239686251, 'learning_rate/qwen_vl_interface': 4.6010659367291945e-06, 'learning_rate/action_model': 4.601065936729194e-05, 'epoch': 1.75}) +[2026-06-17 22:45:18,566][starVLA.training.train_starvla][INFO] - Step 3852, Loss: {'train/loss': 0.3240203261375427, 'train/grad_norm_pre_clip': 23.47237205505371, 'timing/data': 0.006335237063467503, 'timing/model': 0.9294521491974592, 'learning_rate/qwen_vl_interface': 4.598923591190289e-06, 'learning_rate/action_model': 4.598923591190289e-05, 'epoch': 1.75}) +[2026-06-17 22:45:22,330][starVLA.training.train_starvla][INFO] - Step 3853, Loss: {'train/loss': 0.3831709921360016, 'train/grad_norm_pre_clip': 25.44171142578125, 'timing/data': 0.006244176998734474, 'timing/model': 0.927075382322073, 'learning_rate/qwen_vl_interface': 4.596781380620225e-06, 'learning_rate/action_model': 4.5967813806202245e-05, 'epoch': 1.75}) +[2026-06-17 22:45:26,357][starVLA.training.train_starvla][INFO] - Step 3854, Loss: {'train/loss': 0.49672433733940125, 'train/grad_norm_pre_clip': 27.37397575378418, 'timing/data': 0.006525992415845394, 'timing/model': 0.9339500572532415, 'learning_rate/qwen_vl_interface': 4.5946393054630845e-06, 'learning_rate/action_model': 4.5946393054630845e-05, 'epoch': 1.75}) +[2026-06-17 22:45:30,214][starVLA.training.train_starvla][INFO] - Step 3855, Loss: {'train/loss': 0.23632432520389557, 'train/grad_norm_pre_clip': 19.193742752075195, 'timing/data': 0.010904448106884956, 'timing/model': 1.17597237508744, 'learning_rate/qwen_vl_interface': 4.592497366162919e-06, 'learning_rate/action_model': 4.592497366162918e-05, 'epoch': 1.75}) +[2026-06-17 22:45:33,949][starVLA.training.train_starvla][INFO] - Step 3856, Loss: {'train/loss': 0.10371553152799606, 'train/grad_norm_pre_clip': 9.3705472946167, 'timing/data': 0.00697188638150692, 'timing/model': 1.0699119456112385, 'learning_rate/qwen_vl_interface': 4.590355563163758e-06, 'learning_rate/action_model': 4.5903555631637574e-05, 'epoch': 1.75}) +[2026-06-17 22:45:37,535][starVLA.training.train_starvla][INFO] - Step 3857, Loss: {'train/loss': 0.1995183527469635, 'train/grad_norm_pre_clip': 15.028121948242188, 'timing/data': 0.008675131015479565, 'timing/model': 0.9246297338977456, 'learning_rate/qwen_vl_interface': 4.588213896909598e-06, 'learning_rate/action_model': 4.588213896909598e-05, 'epoch': 1.75}) +[2026-06-17 22:45:41,393][starVLA.training.train_starvla][INFO] - Step 3858, Loss: {'train/loss': 0.44939711689949036, 'train/grad_norm_pre_clip': 22.454044342041016, 'timing/data': 0.008506054058670998, 'timing/model': 0.9266191991046071, 'learning_rate/qwen_vl_interface': 4.58607236784441e-06, 'learning_rate/action_model': 4.58607236784441e-05, 'epoch': 1.75}) +[2026-06-17 22:45:45,160][starVLA.training.train_starvla][INFO] - Step 3859, Loss: {'train/loss': 0.49129360914230347, 'train/grad_norm_pre_clip': 16.561744689941406, 'timing/data': 0.011502818204462528, 'timing/model': 0.9234881503507495, 'learning_rate/qwen_vl_interface': 4.583930976412133e-06, 'learning_rate/action_model': 4.583930976412133e-05, 'epoch': 1.75}) +[2026-06-17 22:45:48,927][starVLA.training.train_starvla][INFO] - Step 3860, Loss: {'train/loss': 0.26902464032173157, 'train/grad_norm_pre_clip': 20.46599006652832, 'timing/data': 0.010572107508778572, 'timing/model': 0.9224397530779243, 'learning_rate/qwen_vl_interface': 4.5817897230566834e-06, 'learning_rate/action_model': 4.581789723056683e-05, 'epoch': 1.75}) +[2026-06-17 22:45:52,703][starVLA.training.train_starvla][INFO] - Step 3861, Loss: {'train/loss': 0.23814129829406738, 'train/grad_norm_pre_clip': 16.738136291503906, 'timing/data': 0.009751904755830765, 'timing/model': 0.9230968821793795, 'learning_rate/qwen_vl_interface': 4.579648608221943e-06, 'learning_rate/action_model': 4.579648608221943e-05, 'epoch': 1.75}) +[2026-06-17 22:45:56,470][starVLA.training.train_starvla][INFO] - Step 3862, Loss: {'train/loss': 0.3144274353981018, 'train/grad_norm_pre_clip': 26.483808517456055, 'timing/data': 0.009851708076894283, 'timing/model': 1.0843359539285302, 'learning_rate/qwen_vl_interface': 4.577507632351766e-06, 'learning_rate/action_model': 4.5775076323517664e-05, 'epoch': 1.75}) +[2026-06-17 22:46:00,243][starVLA.training.train_starvla][INFO] - Step 3863, Loss: {'train/loss': 0.3831612467765808, 'train/grad_norm_pre_clip': 26.807331085205078, 'timing/data': 0.008954251185059547, 'timing/model': 1.0923320138826966, 'learning_rate/qwen_vl_interface': 4.575366795889981e-06, 'learning_rate/action_model': 4.575366795889981e-05, 'epoch': 1.75}) +[2026-06-17 22:46:03,818][starVLA.training.train_starvla][INFO] - Step 3864, Loss: {'train/loss': 0.3741552233695984, 'train/grad_norm_pre_clip': 24.596397399902344, 'timing/data': 0.008624313399195671, 'timing/model': 0.93026805203408, 'learning_rate/qwen_vl_interface': 4.573226099280386e-06, 'learning_rate/action_model': 4.573226099280386e-05, 'epoch': 1.75}) +[2026-06-17 22:46:07,572][starVLA.training.train_starvla][INFO] - Step 3865, Loss: {'train/loss': 0.09898779541254044, 'train/grad_norm_pre_clip': 9.903159141540527, 'timing/data': 0.007907056249678135, 'timing/model': 0.9260079069063067, 'learning_rate/qwen_vl_interface': 4.571085542966748e-06, 'learning_rate/action_model': 4.5710855429667485e-05, 'epoch': 1.75}) +[2026-06-17 22:46:11,319][starVLA.training.train_starvla][INFO] - Step 3866, Loss: {'train/loss': 0.2695872187614441, 'train/grad_norm_pre_clip': 23.289684295654297, 'timing/data': 0.008415301330387592, 'timing/model': 0.9243399202823639, 'learning_rate/qwen_vl_interface': 4.568945127392807e-06, 'learning_rate/action_model': 4.5689451273928066e-05, 'epoch': 1.75}) +[2026-06-17 22:46:15,072][starVLA.training.train_starvla][INFO] - Step 3867, Loss: {'train/loss': 0.3716433346271515, 'train/grad_norm_pre_clip': 19.07024383544922, 'timing/data': 0.009601685218513012, 'timing/model': 0.9253165265545249, 'learning_rate/qwen_vl_interface': 4.566804853002275e-06, 'learning_rate/action_model': 4.566804853002275e-05, 'epoch': 1.75}) +[2026-06-17 22:46:18,823][starVLA.training.train_starvla][INFO] - Step 3868, Loss: {'train/loss': 0.40143823623657227, 'train/grad_norm_pre_clip': 25.159076690673828, 'timing/data': 0.008957095444202423, 'timing/model': 0.9229776682332158, 'learning_rate/qwen_vl_interface': 4.564664720238833e-06, 'learning_rate/action_model': 4.564664720238833e-05, 'epoch': 1.75}) +[2026-06-17 22:46:22,556][starVLA.training.train_starvla][INFO] - Step 3869, Loss: {'train/loss': 0.11007476598024368, 'train/grad_norm_pre_clip': 11.25478458404541, 'timing/data': 0.009369878098368645, 'timing/model': 1.0851452481001616, 'learning_rate/qwen_vl_interface': 4.5625247295461314e-06, 'learning_rate/action_model': 4.562524729546131e-05, 'epoch': 1.76}) +[2026-06-17 22:46:26,146][starVLA.training.train_starvla][INFO] - Step 3870, Loss: {'train/loss': 0.31723424792289734, 'train/grad_norm_pre_clip': 18.339740753173828, 'timing/data': 0.011762282811105251, 'timing/model': 0.9266164563596249, 'learning_rate/qwen_vl_interface': 4.560384881367791e-06, 'learning_rate/action_model': 4.5603848813677904e-05, 'epoch': 1.76}) +[2026-06-17 22:46:29,894][starVLA.training.train_starvla][INFO] - Step 3871, Loss: {'train/loss': 0.30526039004325867, 'train/grad_norm_pre_clip': 19.608802795410156, 'timing/data': 0.008850792422890663, 'timing/model': 0.9281632108613849, 'learning_rate/qwen_vl_interface': 4.5582451761474085e-06, 'learning_rate/action_model': 4.558245176147408e-05, 'epoch': 1.76}) +[2026-06-17 22:46:33,596][starVLA.training.train_starvla][INFO] - Step 3872, Loss: {'train/loss': 0.20988351106643677, 'train/grad_norm_pre_clip': 19.747482299804688, 'timing/data': 0.006286519579589367, 'timing/model': 0.9243745924904943, 'learning_rate/qwen_vl_interface': 4.556105614328545e-06, 'learning_rate/action_model': 4.5561056143285455e-05, 'epoch': 1.76}) +[2026-06-17 22:46:37,361][starVLA.training.train_starvla][INFO] - Step 3873, Loss: {'train/loss': 0.3350648581981659, 'train/grad_norm_pre_clip': 24.972061157226562, 'timing/data': 0.010773577727377415, 'timing/model': 0.9558775136247277, 'learning_rate/qwen_vl_interface': 4.553966196354734e-06, 'learning_rate/action_model': 4.553966196354734e-05, 'epoch': 1.76}) +[2026-06-17 22:46:41,101][starVLA.training.train_starvla][INFO] - Step 3874, Loss: {'train/loss': 0.211495503783226, 'train/grad_norm_pre_clip': 12.350956916809082, 'timing/data': 0.00896762777119875, 'timing/model': 0.926894586533308, 'learning_rate/qwen_vl_interface': 4.55182692266948e-06, 'learning_rate/action_model': 4.551826922669479e-05, 'epoch': 1.76}) +[2026-06-17 22:46:44,839][starVLA.training.train_starvla][INFO] - Step 3875, Loss: {'train/loss': 0.38554373383522034, 'train/grad_norm_pre_clip': 35.37550735473633, 'timing/data': 0.010851073078811169, 'timing/model': 1.0794163951650262, 'learning_rate/qwen_vl_interface': 4.549687793716257e-06, 'learning_rate/action_model': 4.549687793716257e-05, 'epoch': 1.76}) +[2026-06-17 22:46:48,593][starVLA.training.train_starvla][INFO] - Step 3876, Loss: {'train/loss': 0.19244809448719025, 'train/grad_norm_pre_clip': 18.994327545166016, 'timing/data': 0.009866045787930489, 'timing/model': 1.0793532766401768, 'learning_rate/qwen_vl_interface': 4.547548809938507e-06, 'learning_rate/action_model': 4.5475488099385075e-05, 'epoch': 1.76}) +[2026-06-17 22:46:52,339][starVLA.training.train_starvla][INFO] - Step 3877, Loss: {'train/loss': 0.3745283782482147, 'train/grad_norm_pre_clip': 27.707796096801758, 'timing/data': 0.013897147960960865, 'timing/model': 1.0544669162482023, 'learning_rate/qwen_vl_interface': 4.545409971779644e-06, 'learning_rate/action_model': 4.545409971779644e-05, 'epoch': 1.76}) +[2026-06-17 22:46:56,088][starVLA.training.train_starvla][INFO] - Step 3878, Loss: {'train/loss': 0.3101138174533844, 'train/grad_norm_pre_clip': 23.528928756713867, 'timing/data': 0.01099483948200941, 'timing/model': 0.9271273380145431, 'learning_rate/qwen_vl_interface': 4.543271279683052e-06, 'learning_rate/action_model': 4.543271279683053e-05, 'epoch': 1.76}) +[2026-06-17 22:46:59,816][starVLA.training.train_starvla][INFO] - Step 3879, Loss: {'train/loss': 0.1983286738395691, 'train/grad_norm_pre_clip': 16.985111236572266, 'timing/data': 0.007712340913712978, 'timing/model': 0.9206324080005288, 'learning_rate/qwen_vl_interface': 4.541132734092086e-06, 'learning_rate/action_model': 4.541132734092086e-05, 'epoch': 1.76}) +[2026-06-17 22:47:03,546][starVLA.training.train_starvla][INFO] - Step 3880, Loss: {'train/loss': 0.33437642455101013, 'train/grad_norm_pre_clip': 27.52489471435547, 'timing/data': 0.011171027086675167, 'timing/model': 0.921029107645154, 'learning_rate/qwen_vl_interface': 4.538994335450066e-06, 'learning_rate/action_model': 4.5389943354500664e-05, 'epoch': 1.76}) +[2026-06-17 22:47:07,294][starVLA.training.train_starvla][INFO] - Step 3881, Loss: {'train/loss': 0.41067856550216675, 'train/grad_norm_pre_clip': 47.470455169677734, 'timing/data': 0.009038074873387814, 'timing/model': 0.9237396018579602, 'learning_rate/qwen_vl_interface': 4.5368560842002836e-06, 'learning_rate/action_model': 4.5368560842002834e-05, 'epoch': 1.76}) +[2026-06-17 22:47:11,029][starVLA.training.train_starvla][INFO] - Step 3882, Loss: {'train/loss': 0.23133619129657745, 'train/grad_norm_pre_clip': 21.949737548828125, 'timing/data': 0.011172187514603138, 'timing/model': 1.0851785354316235, 'learning_rate/qwen_vl_interface': 4.534717980786005e-06, 'learning_rate/action_model': 4.5347179807860046e-05, 'epoch': 1.76}) +[2026-06-17 22:47:14,761][starVLA.training.train_starvla][INFO] - Step 3883, Loss: {'train/loss': 0.19779489934444427, 'train/grad_norm_pre_clip': 16.037431716918945, 'timing/data': 0.008569128811359406, 'timing/model': 1.0828392757102847, 'learning_rate/qwen_vl_interface': 4.532580025650457e-06, 'learning_rate/action_model': 4.5325800256504574e-05, 'epoch': 1.76}) +[2026-06-17 22:47:18,320][starVLA.training.train_starvla][INFO] - Step 3884, Loss: {'train/loss': 0.14898385107517242, 'train/grad_norm_pre_clip': 18.758413314819336, 'timing/data': 0.01060305256396532, 'timing/model': 0.9180974755436182, 'learning_rate/qwen_vl_interface': 4.530442219236842e-06, 'learning_rate/action_model': 4.530442219236842e-05, 'epoch': 1.76}) +[2026-06-17 22:47:22,045][starVLA.training.train_starvla][INFO] - Step 3885, Loss: {'train/loss': 0.2898932695388794, 'train/grad_norm_pre_clip': 23.61268424987793, 'timing/data': 0.006517674773931503, 'timing/model': 0.922131952829659, 'learning_rate/qwen_vl_interface': 4.5283045619883285e-06, 'learning_rate/action_model': 4.5283045619883285e-05, 'epoch': 1.76}) +[2026-06-17 22:47:25,823][starVLA.training.train_starvla][INFO] - Step 3886, Loss: {'train/loss': 0.48795634508132935, 'train/grad_norm_pre_clip': 33.6595458984375, 'timing/data': 0.006532078608870506, 'timing/model': 0.9380325619131327, 'learning_rate/qwen_vl_interface': 4.526167054348057e-06, 'learning_rate/action_model': 4.526167054348057e-05, 'epoch': 1.76}) +[2026-06-17 22:47:29,561][starVLA.training.train_starvla][INFO] - Step 3887, Loss: {'train/loss': 0.12189821898937225, 'train/grad_norm_pre_clip': 11.206512451171875, 'timing/data': 0.006674066185951233, 'timing/model': 0.922087469138205, 'learning_rate/qwen_vl_interface': 4.524029696759135e-06, 'learning_rate/action_model': 4.5240296967591354e-05, 'epoch': 1.76}) +[2026-06-17 22:47:33,309][starVLA.training.train_starvla][INFO] - Step 3888, Loss: {'train/loss': 0.36201900243759155, 'train/grad_norm_pre_clip': 27.695032119750977, 'timing/data': 0.009456438943743706, 'timing/model': 0.9183446047827601, 'learning_rate/qwen_vl_interface': 4.5218924896646354e-06, 'learning_rate/action_model': 4.521892489664635e-05, 'epoch': 1.76}) +[2026-06-17 22:47:37,073][starVLA.training.train_starvla][INFO] - Step 3889, Loss: {'train/loss': 0.3403104245662689, 'train/grad_norm_pre_clip': 21.4686222076416, 'timing/data': 0.010170497000217438, 'timing/model': 1.1138255093246698, 'learning_rate/qwen_vl_interface': 4.519755433507609e-06, 'learning_rate/action_model': 4.5197554335076086e-05, 'epoch': 1.76}) +[2026-06-17 22:47:40,654][starVLA.training.train_starvla][INFO] - Step 3890, Loss: {'train/loss': 0.3775997459888458, 'train/grad_norm_pre_clip': 16.795589447021484, 'timing/data': 0.012517373077571392, 'timing/model': 0.920106953009963, 'learning_rate/qwen_vl_interface': 4.517618528731066e-06, 'learning_rate/action_model': 4.517618528731065e-05, 'epoch': 1.76}) +[2026-06-17 22:47:44,401][starVLA.training.train_starvla][INFO] - Step 3891, Loss: {'train/loss': 0.3340291380882263, 'train/grad_norm_pre_clip': 19.347932815551758, 'timing/data': 0.008855977095663548, 'timing/model': 0.9220232693478465, 'learning_rate/qwen_vl_interface': 4.515481775777989e-06, 'learning_rate/action_model': 4.515481775777988e-05, 'epoch': 1.77}) +[2026-06-17 22:47:48,084][starVLA.training.train_starvla][INFO] - Step 3892, Loss: {'train/loss': 0.32571008801460266, 'train/grad_norm_pre_clip': 16.831531524658203, 'timing/data': 0.0076639242470264435, 'timing/model': 0.9216427616775036, 'learning_rate/qwen_vl_interface': 4.51334517509133e-06, 'learning_rate/action_model': 4.5133451750913305e-05, 'epoch': 1.77}) +[2026-06-17 22:47:51,842][starVLA.training.train_starvla][INFO] - Step 3893, Loss: {'train/loss': 0.11618619412183762, 'train/grad_norm_pre_clip': 20.7266845703125, 'timing/data': 0.012408332899212837, 'timing/model': 0.919404024258256, 'learning_rate/qwen_vl_interface': 4.511208727114011e-06, 'learning_rate/action_model': 4.51120872711401e-05, 'epoch': 1.77}) +[2026-06-17 22:47:55,571][starVLA.training.train_starvla][INFO] - Step 3894, Loss: {'train/loss': 0.5404388308525085, 'train/grad_norm_pre_clip': 23.31566047668457, 'timing/data': 0.011506190523505211, 'timing/model': 0.9250580621883273, 'learning_rate/qwen_vl_interface': 4.5090724322889175e-06, 'learning_rate/action_model': 4.5090724322889174e-05, 'epoch': 1.77}) +[2026-06-17 22:47:59,345][starVLA.training.train_starvla][INFO] - Step 3895, Loss: {'train/loss': 0.3047625422477722, 'train/grad_norm_pre_clip': 20.19275665283203, 'timing/data': 0.009698702953755856, 'timing/model': 1.091049786657095, 'learning_rate/qwen_vl_interface': 4.506936291058902e-06, 'learning_rate/action_model': 4.506936291058902e-05, 'epoch': 1.77}) +[2026-06-17 22:48:03,066][starVLA.training.train_starvla][INFO] - Step 3896, Loss: {'train/loss': 0.3127751648426056, 'train/grad_norm_pre_clip': 23.2078914642334, 'timing/data': 0.018722712993621826, 'timing/model': 1.0655000768601894, 'learning_rate/qwen_vl_interface': 4.504800303866798e-06, 'learning_rate/action_model': 4.504800303866798e-05, 'epoch': 1.77}) +[2026-06-17 22:48:06,741][starVLA.training.train_starvla][INFO] - Step 3897, Loss: {'train/loss': 0.2689777612686157, 'train/grad_norm_pre_clip': 16.68942642211914, 'timing/data': 0.011406774632632732, 'timing/model': 1.002041781321168, 'learning_rate/qwen_vl_interface': 4.50266447115539e-06, 'learning_rate/action_model': 4.5026644711553904e-05, 'epoch': 1.77}) +[2026-06-17 22:48:10,539][starVLA.training.train_starvla][INFO] - Step 3898, Loss: {'train/loss': 0.40094247460365295, 'train/grad_norm_pre_clip': 28.923940658569336, 'timing/data': 0.00943929236382246, 'timing/model': 0.9245610618963838, 'learning_rate/qwen_vl_interface': 4.500528793367444e-06, 'learning_rate/action_model': 4.5005287933674435e-05, 'epoch': 1.77}) +[2026-06-17 22:48:14,358][starVLA.training.train_starvla][INFO] - Step 3899, Loss: {'train/loss': 0.2717256247997284, 'train/grad_norm_pre_clip': 25.93048858642578, 'timing/data': 0.00843745470046997, 'timing/model': 0.9382949648424983, 'learning_rate/qwen_vl_interface': 4.498393270945683e-06, 'learning_rate/action_model': 4.498393270945683e-05, 'epoch': 1.77}) +[2026-06-17 22:50:11,473][starVLA.training.train_starvla][INFO] - Step 3900, Loss: {'train/loss': 0.47599920630455017, 'train/grad_norm_pre_clip': 30.107248306274414, 'eval/loss': 0.31643420457839966, 'eval/latency_8/loss': 0.31643420457839966, 'eval/demon_fire_precision': 0.9553005134400483, 'eval/demon_fire_recall': 0.9826032929481205, 'eval/demon_fire_f1': 0.9687595712098009, 'eval/demon_fire_rate_model': 0.9212576516416249, 'eval/demon_fire_rate_teacher': 0.8956594323873122, 'eval/demon_left_precision': 0.8277404921700223, 'eval/demon_left_recall': 0.7922912205567452, 'eval/demon_left_f1': 0.8096280087527353, 'eval/demon_left_rate_model': 0.12437395659432388, 'eval/demon_left_rate_teacher': 0.12993878686700056, 'eval/demon_right_precision': 0.7997054491899853, 'eval/demon_right_recall': 0.8227272727272728, 'eval/demon_right_f1': 0.8110530246452576, 'eval/demon_right_rate_model': 0.1889259877573734, 'eval/demon_right_rate_teacher': 0.18363939899833054, 'eval/demon_move_f1': 0.8103405166989965, 'eval/demon_cc_f1': 0.9150903733536345, 'eval/latency_8/demon_fire_precision': 0.9553005134400483, 'eval/latency_8/demon_fire_recall': 0.9826032929481205, 'eval/latency_8/demon_fire_f1': 0.9687595712098009, 'eval/latency_8/demon_fire_rate_model': 0.9212576516416249, 'eval/latency_8/demon_fire_rate_teacher': 0.8956594323873122, 'eval/latency_8/demon_left_precision': 0.8277404921700223, 'eval/latency_8/demon_left_recall': 0.7922912205567452, 'eval/latency_8/demon_left_f1': 0.8096280087527353, 'eval/latency_8/demon_left_rate_model': 0.12437395659432388, 'eval/latency_8/demon_left_rate_teacher': 0.12993878686700056, 'eval/latency_8/demon_right_precision': 0.7997054491899853, 'eval/latency_8/demon_right_recall': 0.8227272727272728, 'eval/latency_8/demon_right_f1': 0.8110530246452576, 'eval/latency_8/demon_right_rate_model': 0.1889259877573734, 'eval/latency_8/demon_right_rate_teacher': 0.18363939899833054, 'eval/latency_8/demon_move_f1': 0.8103405166989965, 'eval/latency_8/demon_cc_f1': 0.9150903733536345, 'timing/data': 0.008921025320887566, 'timing/model': 0.9313310272991657, 'learning_rate/qwen_vl_interface': 4.496257904332806e-06, 'learning_rate/action_model': 4.4962579043328064e-05, 'epoch': 1.77}) +[2026-06-17 22:50:11,474][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_3900_state +[2026-06-17 22:50:11,479][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 22:50:18,107][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_3900_state/model.safetensors +[2026-06-17 22:50:33,407][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_3900_state/optimizer.bin +[2026-06-17 22:50:33,407][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_3900_state/sampler.bin +[2026-06-17 22:50:33,408][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_3900_state/sampler_1.bin +[2026-06-17 22:50:33,411][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_3900_state/random_states_0.pkl +[2026-06-17 22:50:33,415][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3900 +[2026-06-17 22:50:33,415][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 22:50:33,425][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 22:50:37,244][starVLA.training.train_starvla][INFO] - Step 3901, Loss: {'train/loss': 0.3421543836593628, 'train/grad_norm_pre_clip': 19.74652862548828, 'timing/data': 0.009575209580361843, 'timing/model': 0.9292702320963144, 'learning_rate/qwen_vl_interface': 4.494122693971478e-06, 'learning_rate/action_model': 4.4941226939714785e-05, 'epoch': 1.77}) +[2026-06-17 22:50:41,028][starVLA.training.train_starvla][INFO] - Step 3902, Loss: {'train/loss': 0.22261618077754974, 'train/grad_norm_pre_clip': 18.71198844909668, 'timing/data': 0.016666839830577374, 'timing/model': 0.9351045060902834, 'learning_rate/qwen_vl_interface': 4.491987640304327e-06, 'learning_rate/action_model': 4.4919876403043265e-05, 'epoch': 1.77}) +[2026-06-17 22:50:44,958][starVLA.training.train_starvla][INFO] - Step 3903, Loss: {'train/loss': 0.7353923916816711, 'train/grad_norm_pre_clip': 36.8815803527832, 'timing/data': 0.008532073348760605, 'timing/model': 0.9639407955110073, 'learning_rate/qwen_vl_interface': 4.489852743773955e-06, 'learning_rate/action_model': 4.489852743773955e-05, 'epoch': 1.77}) +[2026-06-17 22:50:48,826][starVLA.training.train_starvla][INFO] - Step 3904, Loss: {'train/loss': 0.1974073350429535, 'train/grad_norm_pre_clip': 16.412662506103516, 'timing/data': 0.009489428251981735, 'timing/model': 0.9628686904907227, 'learning_rate/qwen_vl_interface': 4.487718004822928e-06, 'learning_rate/action_model': 4.487718004822928e-05, 'epoch': 1.77}) +[2026-06-17 22:50:52,725][starVLA.training.train_starvla][INFO] - Step 3905, Loss: {'train/loss': 0.37111493945121765, 'train/grad_norm_pre_clip': 17.5993709564209, 'timing/data': 0.00878993421792984, 'timing/model': 1.131747705861926, 'learning_rate/qwen_vl_interface': 4.485583423893778e-06, 'learning_rate/action_model': 4.4855834238937774e-05, 'epoch': 1.77}) +[2026-06-17 22:50:56,367][starVLA.training.train_starvla][INFO] - Step 3906, Loss: {'train/loss': 0.20393893122673035, 'train/grad_norm_pre_clip': 16.640905380249023, 'timing/data': 0.009130884893238544, 'timing/model': 0.9245365494862199, 'learning_rate/qwen_vl_interface': 4.483449001429005e-06, 'learning_rate/action_model': 4.483449001429005e-05, 'epoch': 1.77}) +[2026-06-17 22:51:00,298][starVLA.training.train_starvla][INFO] - Step 3907, Loss: {'train/loss': 0.27695196866989136, 'train/grad_norm_pre_clip': 18.23622703552246, 'timing/data': 0.008556847460567951, 'timing/model': 0.9343155790120363, 'learning_rate/qwen_vl_interface': 4.481314737871081e-06, 'learning_rate/action_model': 4.4813147378710804e-05, 'epoch': 1.77}) +[2026-06-17 22:51:04,100][starVLA.training.train_starvla][INFO] - Step 3908, Loss: {'train/loss': 0.08784394711256027, 'train/grad_norm_pre_clip': 12.079202651977539, 'timing/data': 0.013239489868283272, 'timing/model': 0.9335689842700958, 'learning_rate/qwen_vl_interface': 4.47918063366244e-06, 'learning_rate/action_model': 4.4791806336624395e-05, 'epoch': 1.77}) +[2026-06-17 22:51:08,186][starVLA.training.train_starvla][INFO] - Step 3909, Loss: {'train/loss': 0.48385170102119446, 'train/grad_norm_pre_clip': 21.801023483276367, 'timing/data': 0.010188235901296139, 'timing/model': 1.0194058194756508, 'learning_rate/qwen_vl_interface': 4.477046689245481e-06, 'learning_rate/action_model': 4.477046689245481e-05, 'epoch': 1.77}) +[2026-06-17 22:51:12,214][starVLA.training.train_starvla][INFO] - Step 3910, Loss: {'train/loss': 0.21727029979228973, 'train/grad_norm_pre_clip': 19.607606887817383, 'timing/data': 0.014537121169269085, 'timing/model': 0.9377337945625186, 'learning_rate/qwen_vl_interface': 4.474912905062575e-06, 'learning_rate/action_model': 4.474912905062575e-05, 'epoch': 1.77}) +[2026-06-17 22:51:15,987][starVLA.training.train_starvla][INFO] - Step 3911, Loss: {'train/loss': 0.30202516913414, 'train/grad_norm_pre_clip': 16.072174072265625, 'timing/data': 0.012043667025864124, 'timing/model': 1.0894416030496359, 'learning_rate/qwen_vl_interface': 4.472779281556059e-06, 'learning_rate/action_model': 4.4727792815560585e-05, 'epoch': 1.77}) +[2026-06-17 22:51:19,685][starVLA.training.train_starvla][INFO] - Step 3912, Loss: {'train/loss': 0.5805655717849731, 'train/grad_norm_pre_clip': 28.959609985351562, 'timing/data': 0.01954597979784012, 'timing/model': 1.0251378240063787, 'learning_rate/qwen_vl_interface': 4.470645819168233e-06, 'learning_rate/action_model': 4.4706458191682324e-05, 'epoch': 1.77}) +[2026-06-17 22:51:23,278][starVLA.training.train_starvla][INFO] - Step 3913, Loss: {'train/loss': 0.3074423670768738, 'train/grad_norm_pre_clip': 30.418428421020508, 'timing/data': 0.015556060709059238, 'timing/model': 0.921471199952066, 'learning_rate/qwen_vl_interface': 4.468512518341367e-06, 'learning_rate/action_model': 4.4685125183413674e-05, 'epoch': 1.78}) +[2026-06-17 22:51:27,027][starVLA.training.train_starvla][INFO] - Step 3914, Loss: {'train/loss': 0.48550841212272644, 'train/grad_norm_pre_clip': 27.94936180114746, 'timing/data': 0.008336836472153664, 'timing/model': 0.9278336223214865, 'learning_rate/qwen_vl_interface': 4.466379379517699e-06, 'learning_rate/action_model': 4.4663793795176985e-05, 'epoch': 1.78}) +[2026-06-17 22:51:30,794][starVLA.training.train_starvla][INFO] - Step 3915, Loss: {'train/loss': 0.44072505831718445, 'train/grad_norm_pre_clip': 23.543148040771484, 'timing/data': 0.008993028663098812, 'timing/model': 0.9222666146233678, 'learning_rate/qwen_vl_interface': 4.464246403139429e-06, 'learning_rate/action_model': 4.464246403139428e-05, 'epoch': 1.78}) +[2026-06-17 22:51:34,571][starVLA.training.train_starvla][INFO] - Step 3916, Loss: {'train/loss': 0.39004120230674744, 'train/grad_norm_pre_clip': 30.065523147583008, 'timing/data': 0.014408794231712818, 'timing/model': 0.9224468106403947, 'learning_rate/qwen_vl_interface': 4.462113589648723e-06, 'learning_rate/action_model': 4.462113589648723e-05, 'epoch': 1.78}) +[2026-06-17 22:51:38,301][starVLA.training.train_starvla][INFO] - Step 3917, Loss: {'train/loss': 0.22830520570278168, 'train/grad_norm_pre_clip': 24.13764190673828, 'timing/data': 0.00807966385036707, 'timing/model': 0.9298671949654818, 'learning_rate/qwen_vl_interface': 4.459980939487718e-06, 'learning_rate/action_model': 4.4599809394877176e-05, 'epoch': 1.78}) +[2026-06-17 22:51:42,125][starVLA.training.train_starvla][INFO] - Step 3918, Loss: {'train/loss': 0.2782840132713318, 'train/grad_norm_pre_clip': 21.22124671936035, 'timing/data': 0.008108692243695259, 'timing/model': 1.1252907579764724, 'learning_rate/qwen_vl_interface': 4.457848453098514e-06, 'learning_rate/action_model': 4.4578484530985144e-05, 'epoch': 1.78}) +[2026-06-17 22:51:45,825][starVLA.training.train_starvla][INFO] - Step 3919, Loss: {'train/loss': 0.4279753267765045, 'train/grad_norm_pre_clip': 28.720115661621094, 'timing/data': 0.010292110964655876, 'timing/model': 0.9470179919153452, 'learning_rate/qwen_vl_interface': 4.45571613092318e-06, 'learning_rate/action_model': 4.4557161309231794e-05, 'epoch': 1.78}) +[2026-06-17 22:51:49,647][starVLA.training.train_starvla][INFO] - Step 3920, Loss: {'train/loss': 0.3288589417934418, 'train/grad_norm_pre_clip': 23.459720611572266, 'timing/data': 0.007178829051554203, 'timing/model': 0.9333651755005121, 'learning_rate/qwen_vl_interface': 4.453583973403746e-06, 'learning_rate/action_model': 4.453583973403746e-05, 'epoch': 1.78}) +[2026-06-17 22:51:53,441][starVLA.training.train_starvla][INFO] - Step 3921, Loss: {'train/loss': 0.26347121596336365, 'train/grad_norm_pre_clip': 17.684940338134766, 'timing/data': 0.007643876597285271, 'timing/model': 0.9296416817232966, 'learning_rate/qwen_vl_interface': 4.451451980982209e-06, 'learning_rate/action_model': 4.4514519809822086e-05, 'epoch': 1.78}) +[2026-06-17 22:51:57,188][starVLA.training.train_starvla][INFO] - Step 3922, Loss: {'train/loss': 0.37499234080314636, 'train/grad_norm_pre_clip': 21.043556213378906, 'timing/data': 0.008120622485876083, 'timing/model': 0.9296969762071967, 'learning_rate/qwen_vl_interface': 4.4493201541005385e-06, 'learning_rate/action_model': 4.4493201541005384e-05, 'epoch': 1.78}) +[2026-06-17 22:52:00,932][starVLA.training.train_starvla][INFO] - Step 3923, Loss: {'train/loss': 0.17655692994594574, 'train/grad_norm_pre_clip': 15.776178359985352, 'timing/data': 0.016899594105780125, 'timing/model': 0.9252800447866321, 'learning_rate/qwen_vl_interface': 4.447188493200661e-06, 'learning_rate/action_model': 4.447188493200661e-05, 'epoch': 1.78}) +[2026-06-17 22:52:04,691][starVLA.training.train_starvla][INFO] - Step 3924, Loss: {'train/loss': 0.2621331214904785, 'train/grad_norm_pre_clip': 22.631492614746094, 'timing/data': 0.00881240051239729, 'timing/model': 1.0928816758096218, 'learning_rate/qwen_vl_interface': 4.445056998724472e-06, 'learning_rate/action_model': 4.4450569987244716e-05, 'epoch': 1.78}) +[2026-06-17 22:52:08,309][starVLA.training.train_starvla][INFO] - Step 3925, Loss: {'train/loss': 0.15087398886680603, 'train/grad_norm_pre_clip': 13.67111873626709, 'timing/data': 0.011072715744376183, 'timing/model': 0.9300020858645439, 'learning_rate/qwen_vl_interface': 4.4429256711138335e-06, 'learning_rate/action_model': 4.442925671113833e-05, 'epoch': 1.78}) +[2026-06-17 22:52:12,066][starVLA.training.train_starvla][INFO] - Step 3926, Loss: {'train/loss': 0.4107125699520111, 'train/grad_norm_pre_clip': 27.071510314941406, 'timing/data': 0.009172643534839153, 'timing/model': 0.9248225232586265, 'learning_rate/qwen_vl_interface': 4.440794510810572e-06, 'learning_rate/action_model': 4.440794510810572e-05, 'epoch': 1.78}) +[2026-06-17 22:52:15,826][starVLA.training.train_starvla][INFO] - Step 3927, Loss: {'train/loss': 0.2567521035671234, 'train/grad_norm_pre_clip': 20.16120147705078, 'timing/data': 0.007872048765420914, 'timing/model': 0.9273430267348886, 'learning_rate/qwen_vl_interface': 4.438663518256481e-06, 'learning_rate/action_model': 4.438663518256481e-05, 'epoch': 1.78}) +[2026-06-17 22:52:19,567][starVLA.training.train_starvla][INFO] - Step 3928, Loss: {'train/loss': 0.16471652686595917, 'train/grad_norm_pre_clip': 8.918960571289062, 'timing/data': 0.007813677191734314, 'timing/model': 0.9282763814553618, 'learning_rate/qwen_vl_interface': 4.436532693893311e-06, 'learning_rate/action_model': 4.436532693893311e-05, 'epoch': 1.78}) +[2026-06-17 22:52:23,342][starVLA.training.train_starvla][INFO] - Step 3929, Loss: {'train/loss': 0.280505895614624, 'train/grad_norm_pre_clip': 22.40940284729004, 'timing/data': 0.009568266570568085, 'timing/model': 0.9300929652526975, 'learning_rate/qwen_vl_interface': 4.434402038162793e-06, 'learning_rate/action_model': 4.434402038162793e-05, 'epoch': 1.78}) +[2026-06-17 22:52:27,154][starVLA.training.train_starvla][INFO] - Step 3930, Loss: {'train/loss': 0.24846605956554413, 'train/grad_norm_pre_clip': 15.464106559753418, 'timing/data': 0.014295312575995922, 'timing/model': 1.1394283547997475, 'learning_rate/qwen_vl_interface': 4.432271551506609e-06, 'learning_rate/action_model': 4.4322715515066094e-05, 'epoch': 1.78}) +[2026-06-17 22:52:30,843][starVLA.training.train_starvla][INFO] - Step 3931, Loss: {'train/loss': 0.14652206003665924, 'train/grad_norm_pre_clip': 14.62572956085205, 'timing/data': 0.012880438938736916, 'timing/model': 1.036190644837916, 'learning_rate/qwen_vl_interface': 4.4301412343664125e-06, 'learning_rate/action_model': 4.4301412343664125e-05, 'epoch': 1.78}) +[2026-06-17 22:52:34,426][starVLA.training.train_starvla][INFO] - Step 3932, Loss: {'train/loss': 0.134432852268219, 'train/grad_norm_pre_clip': 10.746195793151855, 'timing/data': 0.01466628909111023, 'timing/model': 0.921661825850606, 'learning_rate/qwen_vl_interface': 4.428011087183819e-06, 'learning_rate/action_model': 4.428011087183819e-05, 'epoch': 1.78}) +[2026-06-17 22:52:38,184][starVLA.training.train_starvla][INFO] - Step 3933, Loss: {'train/loss': 0.17265884578227997, 'train/grad_norm_pre_clip': 14.959680557250977, 'timing/data': 0.013775232248008251, 'timing/model': 0.9302954021841288, 'learning_rate/qwen_vl_interface': 4.4258811104004125e-06, 'learning_rate/action_model': 4.425881110400413e-05, 'epoch': 1.78}) +[2026-06-17 22:52:41,919][starVLA.training.train_starvla][INFO] - Step 3934, Loss: {'train/loss': 0.4117092490196228, 'train/grad_norm_pre_clip': 21.146533966064453, 'timing/data': 0.009123300202190876, 'timing/model': 0.9301582835614681, 'learning_rate/qwen_vl_interface': 4.4237513044577394e-06, 'learning_rate/action_model': 4.4237513044577386e-05, 'epoch': 1.78}) +[2026-06-17 22:52:45,661][starVLA.training.train_starvla][INFO] - Step 3935, Loss: {'train/loss': 0.24009226262569427, 'train/grad_norm_pre_clip': 16.152725219726562, 'timing/data': 0.008091865107417107, 'timing/model': 0.9268110608682036, 'learning_rate/qwen_vl_interface': 4.421621669797309e-06, 'learning_rate/action_model': 4.421621669797308e-05, 'epoch': 1.79}) +[2026-06-17 22:52:49,417][starVLA.training.train_starvla][INFO] - Step 3936, Loss: {'train/loss': 0.296671599149704, 'train/grad_norm_pre_clip': 17.28120231628418, 'timing/data': 0.007312443107366562, 'timing/model': 0.9302570754662156, 'learning_rate/qwen_vl_interface': 4.419492206860599e-06, 'learning_rate/action_model': 4.419492206860598e-05, 'epoch': 1.79}) +[2026-06-17 22:52:53,163][starVLA.training.train_starvla][INFO] - Step 3937, Loss: {'train/loss': 0.18518365919589996, 'train/grad_norm_pre_clip': 15.667155265808105, 'timing/data': 0.008735496550798416, 'timing/model': 1.0886315098032355, 'learning_rate/qwen_vl_interface': 4.417362916089048e-06, 'learning_rate/action_model': 4.417362916089047e-05, 'epoch': 1.79}) +[2026-06-17 22:52:56,846][starVLA.training.train_starvla][INFO] - Step 3938, Loss: {'train/loss': 0.273154079914093, 'train/grad_norm_pre_clip': 21.005916595458984, 'timing/data': 0.008858591318130493, 'timing/model': 1.0292571019381285, 'learning_rate/qwen_vl_interface': 4.415233797924059e-06, 'learning_rate/action_model': 4.415233797924059e-05, 'epoch': 1.79}) +[2026-06-17 22:53:00,445][starVLA.training.train_starvla][INFO] - Step 3939, Loss: {'train/loss': 0.14243291318416595, 'train/grad_norm_pre_clip': 15.39123821258545, 'timing/data': 0.013739202171564102, 'timing/model': 0.9223913969472051, 'learning_rate/qwen_vl_interface': 4.413104852807002e-06, 'learning_rate/action_model': 4.413104852807001e-05, 'epoch': 1.79}) +[2026-06-17 22:53:04,197][starVLA.training.train_starvla][INFO] - Step 3940, Loss: {'train/loss': 0.12860873341560364, 'train/grad_norm_pre_clip': 13.694912910461426, 'timing/data': 0.007844904437661171, 'timing/model': 0.9282535305246711, 'learning_rate/qwen_vl_interface': 4.41097608117921e-06, 'learning_rate/action_model': 4.41097608117921e-05, 'epoch': 1.79}) +[2026-06-17 22:53:07,969][starVLA.training.train_starvla][INFO] - Step 3941, Loss: {'train/loss': 0.21504858136177063, 'train/grad_norm_pre_clip': 22.1386775970459, 'timing/data': 0.008733097463846207, 'timing/model': 0.9294939571991563, 'learning_rate/qwen_vl_interface': 4.40884748348198e-06, 'learning_rate/action_model': 4.40884748348198e-05, 'epoch': 1.79}) +[2026-06-17 22:53:11,726][starVLA.training.train_starvla][INFO] - Step 3942, Loss: {'train/loss': 0.23412105441093445, 'train/grad_norm_pre_clip': 24.512996673583984, 'timing/data': 0.008492084220051765, 'timing/model': 0.92783756647259, 'learning_rate/qwen_vl_interface': 4.406719060156572e-06, 'learning_rate/action_model': 4.4067190601565716e-05, 'epoch': 1.79}) +[2026-06-17 22:53:15,500][starVLA.training.train_starvla][INFO] - Step 3943, Loss: {'train/loss': 0.21441325545310974, 'train/grad_norm_pre_clip': 16.333751678466797, 'timing/data': 0.0072930846363306046, 'timing/model': 0.9252860695123672, 'learning_rate/qwen_vl_interface': 4.404590811644209e-06, 'learning_rate/action_model': 4.4045908116442087e-05, 'epoch': 1.79}) +[2026-06-17 22:53:19,269][starVLA.training.train_starvla][INFO] - Step 3944, Loss: {'train/loss': 0.4165957570075989, 'train/grad_norm_pre_clip': 34.42571258544922, 'timing/data': 0.01013138610869646, 'timing/model': 1.0911583816632628, 'learning_rate/qwen_vl_interface': 4.40246273838608e-06, 'learning_rate/action_model': 4.40246273838608e-05, 'epoch': 1.79}) +[2026-06-17 22:53:22,967][starVLA.training.train_starvla][INFO] - Step 3945, Loss: {'train/loss': 0.08811604976654053, 'train/grad_norm_pre_clip': 11.473824501037598, 'timing/data': 0.011279474012553692, 'timing/model': 1.0105018522590399, 'learning_rate/qwen_vl_interface': 4.400334840823339e-06, 'learning_rate/action_model': 4.400334840823339e-05, 'epoch': 1.79}) +[2026-06-17 22:53:26,747][starVLA.training.train_starvla][INFO] - Step 3946, Loss: {'train/loss': 0.3166889548301697, 'train/grad_norm_pre_clip': 21.774751663208008, 'timing/data': 0.00862831249833107, 'timing/model': 0.9281215323135257, 'learning_rate/qwen_vl_interface': 4.398207119397098e-06, 'learning_rate/action_model': 4.398207119397098e-05, 'epoch': 1.79}) +[2026-06-17 22:53:30,501][starVLA.training.train_starvla][INFO] - Step 3947, Loss: {'train/loss': 0.2810770869255066, 'train/grad_norm_pre_clip': 50.297813415527344, 'timing/data': 0.010129226371645927, 'timing/model': 0.9266516948118806, 'learning_rate/qwen_vl_interface': 4.396079574548439e-06, 'learning_rate/action_model': 4.396079574548439e-05, 'epoch': 1.79}) +[2026-06-17 22:53:34,236][starVLA.training.train_starvla][INFO] - Step 3948, Loss: {'train/loss': 0.2543033957481384, 'train/grad_norm_pre_clip': 22.18080711364746, 'timing/data': 0.007919137366116047, 'timing/model': 0.9325053840875626, 'learning_rate/qwen_vl_interface': 4.393952206718403e-06, 'learning_rate/action_model': 4.3939522067184026e-05, 'epoch': 1.79}) +[2026-06-17 22:53:38,010][starVLA.training.train_starvla][INFO] - Step 3949, Loss: {'train/loss': 0.2757904827594757, 'train/grad_norm_pre_clip': 23.534175872802734, 'timing/data': 0.00803151074796915, 'timing/model': 0.9273471925407648, 'learning_rate/qwen_vl_interface': 4.391825016347996e-06, 'learning_rate/action_model': 4.391825016347996e-05, 'epoch': 1.79}) +[2026-06-17 22:53:41,752][starVLA.training.train_starvla][INFO] - Step 3950, Loss: {'train/loss': 0.2167448103427887, 'train/grad_norm_pre_clip': 14.543148040771484, 'timing/data': 0.009210791438817978, 'timing/model': 1.0826287101954222, 'learning_rate/qwen_vl_interface': 4.3896980038781844e-06, 'learning_rate/action_model': 4.3896980038781846e-05, 'epoch': 1.79}) +[2026-06-17 22:53:45,446][starVLA.training.train_starvla][INFO] - Step 3951, Loss: {'train/loss': 0.18261107802391052, 'train/grad_norm_pre_clip': 32.13511276245117, 'timing/data': 0.008417638018727303, 'timing/model': 1.035625302232802, 'learning_rate/qwen_vl_interface': 4.3875711697499054e-06, 'learning_rate/action_model': 4.387571169749905e-05, 'epoch': 1.79}) +[2026-06-17 22:53:49,025][starVLA.training.train_starvla][INFO] - Step 3952, Loss: {'train/loss': 0.5105304718017578, 'train/grad_norm_pre_clip': 34.76091766357422, 'timing/data': 0.00818604789674282, 'timing/model': 0.9263353254646063, 'learning_rate/qwen_vl_interface': 4.38544451440405e-06, 'learning_rate/action_model': 4.3854445144040495e-05, 'epoch': 1.79}) +[2026-06-17 22:53:52,778][starVLA.training.train_starvla][INFO] - Step 3953, Loss: {'train/loss': 0.20796169340610504, 'train/grad_norm_pre_clip': 21.79243278503418, 'timing/data': 0.016175976023077965, 'timing/model': 0.9222070686519146, 'learning_rate/qwen_vl_interface': 4.383318038281475e-06, 'learning_rate/action_model': 4.383318038281475e-05, 'epoch': 1.79}) +[2026-06-17 22:53:56,527][starVLA.training.train_starvla][INFO] - Step 3954, Loss: {'train/loss': 0.48176273703575134, 'train/grad_norm_pre_clip': 32.1236457824707, 'timing/data': 0.006543139927089214, 'timing/model': 0.9269821541383862, 'learning_rate/qwen_vl_interface': 4.381191741823005e-06, 'learning_rate/action_model': 4.3811917418230046e-05, 'epoch': 1.79}) +[2026-06-17 22:54:00,264][starVLA.training.train_starvla][INFO] - Step 3955, Loss: {'train/loss': 0.21079935133457184, 'train/grad_norm_pre_clip': 27.43117332458496, 'timing/data': 0.007182233966886997, 'timing/model': 0.9201335357502103, 'learning_rate/qwen_vl_interface': 4.379065625469422e-06, 'learning_rate/action_model': 4.3790656254694215e-05, 'epoch': 1.79}) +[2026-06-17 22:54:04,012][starVLA.training.train_starvla][INFO] - Step 3956, Loss: {'train/loss': 0.2693691849708557, 'train/grad_norm_pre_clip': 26.27364158630371, 'timing/data': 0.007741314359009266, 'timing/model': 0.9261776031926274, 'learning_rate/qwen_vl_interface': 4.3769396896614724e-06, 'learning_rate/action_model': 4.376939689661472e-05, 'epoch': 1.79}) +[2026-06-17 22:54:07,749][starVLA.training.train_starvla][INFO] - Step 3957, Loss: {'train/loss': 0.5034294128417969, 'train/grad_norm_pre_clip': 28.42747688293457, 'timing/data': 0.007947242818772793, 'timing/model': 1.0869676563888788, 'learning_rate/qwen_vl_interface': 4.374813934839862e-06, 'learning_rate/action_model': 4.374813934839862e-05, 'epoch': 1.8}) +[2026-06-17 22:54:11,445][starVLA.training.train_starvla][INFO] - Step 3958, Loss: {'train/loss': 0.20500987768173218, 'train/grad_norm_pre_clip': 15.757819175720215, 'timing/data': 0.0073373038321733475, 'timing/model': 1.0343373799696565, 'learning_rate/qwen_vl_interface': 4.372688361445267e-06, 'learning_rate/action_model': 4.3726883614452665e-05, 'epoch': 1.8}) +[2026-06-17 22:54:15,064][starVLA.training.train_starvla][INFO] - Step 3959, Loss: {'train/loss': 0.12705716490745544, 'train/grad_norm_pre_clip': 20.150104522705078, 'timing/data': 0.0116198705509305, 'timing/model': 0.9261389262974262, 'learning_rate/qwen_vl_interface': 4.370562969918319e-06, 'learning_rate/action_model': 4.370562969918319e-05, 'epoch': 1.8}) +[2026-06-17 22:54:18,824][starVLA.training.train_starvla][INFO] - Step 3960, Loss: {'train/loss': 0.16802886128425598, 'train/grad_norm_pre_clip': 14.43608283996582, 'timing/data': 0.007910075597465038, 'timing/model': 0.9216246996074915, 'learning_rate/qwen_vl_interface': 4.368437760699613e-06, 'learning_rate/action_model': 4.3684377606996124e-05, 'epoch': 1.8}) +[2026-06-17 22:54:22,586][starVLA.training.train_starvla][INFO] - Step 3961, Loss: {'train/loss': 0.14843465387821198, 'train/grad_norm_pre_clip': 16.79551887512207, 'timing/data': 0.009212581440806389, 'timing/model': 0.9287401987239718, 'learning_rate/qwen_vl_interface': 4.366312734229706e-06, 'learning_rate/action_model': 4.366312734229706e-05, 'epoch': 1.8}) +[2026-06-17 22:54:26,323][starVLA.training.train_starvla][INFO] - Step 3962, Loss: {'train/loss': 0.4113742709159851, 'train/grad_norm_pre_clip': 30.939226150512695, 'timing/data': 0.013427188619971275, 'timing/model': 0.9255896657705307, 'learning_rate/qwen_vl_interface': 4.364187890949122e-06, 'learning_rate/action_model': 4.364187890949122e-05, 'epoch': 1.8}) +[2026-06-17 22:54:30,056][starVLA.training.train_starvla][INFO] - Step 3963, Loss: {'train/loss': 0.5481643080711365, 'train/grad_norm_pre_clip': 37.23508071899414, 'timing/data': 0.00822839792817831, 'timing/model': 0.9188125049695373, 'learning_rate/qwen_vl_interface': 4.362063231298342e-06, 'learning_rate/action_model': 4.3620632312983425e-05, 'epoch': 1.8}) +[2026-06-17 22:54:33,793][starVLA.training.train_starvla][INFO] - Step 3964, Loss: {'train/loss': 0.05595717579126358, 'train/grad_norm_pre_clip': 10.045047760009766, 'timing/data': 0.01186427567154169, 'timing/model': 1.0751875014975667, 'learning_rate/qwen_vl_interface': 4.359938755717807e-06, 'learning_rate/action_model': 4.359938755717807e-05, 'epoch': 1.8}) +[2026-06-17 22:54:37,364][starVLA.training.train_starvla][INFO] - Step 3965, Loss: {'train/loss': 0.3299049139022827, 'train/grad_norm_pre_clip': 22.029415130615234, 'timing/data': 0.006472093053162098, 'timing/model': 0.9257356645539403, 'learning_rate/qwen_vl_interface': 4.357814464647929e-06, 'learning_rate/action_model': 4.357814464647929e-05, 'epoch': 1.8}) +[2026-06-17 22:54:41,110][starVLA.training.train_starvla][INFO] - Step 3966, Loss: {'train/loss': 0.17632491886615753, 'train/grad_norm_pre_clip': 22.47019386291504, 'timing/data': 0.015400994569063187, 'timing/model': 0.9222812196239829, 'learning_rate/qwen_vl_interface': 4.355690358529071e-06, 'learning_rate/action_model': 4.3556903585290705e-05, 'epoch': 1.8}) +[2026-06-17 22:54:44,867][starVLA.training.train_starvla][INFO] - Step 3967, Loss: {'train/loss': 0.28576892614364624, 'train/grad_norm_pre_clip': 25.23063850402832, 'timing/data': 0.016474575735628605, 'timing/model': 0.9264203272759914, 'learning_rate/qwen_vl_interface': 4.353566437801564e-06, 'learning_rate/action_model': 4.353566437801563e-05, 'epoch': 1.8}) +[2026-06-17 22:54:48,610][starVLA.training.train_starvla][INFO] - Step 3968, Loss: {'train/loss': 0.18037158250808716, 'train/grad_norm_pre_clip': 20.243501663208008, 'timing/data': 0.007364717312157154, 'timing/model': 0.9339058259502053, 'learning_rate/qwen_vl_interface': 4.351442702905698e-06, 'learning_rate/action_model': 4.3514427029056976e-05, 'epoch': 1.8}) +[2026-06-17 22:54:52,365][starVLA.training.train_starvla][INFO] - Step 3969, Loss: {'train/loss': 0.15184640884399414, 'train/grad_norm_pre_clip': 13.65555191040039, 'timing/data': 0.006294434890151024, 'timing/model': 0.923361430875957, 'learning_rate/qwen_vl_interface': 4.349319154281727e-06, 'learning_rate/action_model': 4.349319154281728e-05, 'epoch': 1.8}) +[2026-06-17 22:54:56,121][starVLA.training.train_starvla][INFO] - Step 3970, Loss: {'train/loss': 0.40515175461769104, 'train/grad_norm_pre_clip': 29.984773635864258, 'timing/data': 0.007318117655813694, 'timing/model': 1.1208969540894032, 'learning_rate/qwen_vl_interface': 4.347195792369867e-06, 'learning_rate/action_model': 4.347195792369867e-05, 'epoch': 1.8}) +[2026-06-17 22:54:59,832][starVLA.training.train_starvla][INFO] - Step 3971, Loss: {'train/loss': 0.1109081506729126, 'train/grad_norm_pre_clip': 16.939504623413086, 'timing/data': 0.007359879091382027, 'timing/model': 1.0584402224048972, 'learning_rate/qwen_vl_interface': 4.345072617610288e-06, 'learning_rate/action_model': 4.345072617610288e-05, 'epoch': 1.8}) +[2026-06-17 22:55:03,417][starVLA.training.train_starvla][INFO] - Step 3972, Loss: {'train/loss': 0.815208911895752, 'train/grad_norm_pre_clip': 31.355594635009766, 'timing/data': 0.012786182574927807, 'timing/model': 0.9275246486067772, 'learning_rate/qwen_vl_interface': 4.342949630443129e-06, 'learning_rate/action_model': 4.342949630443129e-05, 'epoch': 1.8}) +[2026-06-17 22:55:07,286][starVLA.training.train_starvla][INFO] - Step 3973, Loss: {'train/loss': 0.2401183396577835, 'train/grad_norm_pre_clip': 21.281558990478516, 'timing/data': 0.008036325685679913, 'timing/model': 0.9556742301210761, 'learning_rate/qwen_vl_interface': 4.340826831308488e-06, 'learning_rate/action_model': 4.340826831308488e-05, 'epoch': 1.8}) +[2026-06-17 22:55:11,127][starVLA.training.train_starvla][INFO] - Step 3974, Loss: {'train/loss': 0.1678076535463333, 'train/grad_norm_pre_clip': 19.844348907470703, 'timing/data': 0.016241399571299553, 'timing/model': 0.9462985582649708, 'learning_rate/qwen_vl_interface': 4.338704220646424e-06, 'learning_rate/action_model': 4.338704220646424e-05, 'epoch': 1.8}) +[2026-06-17 22:55:14,935][starVLA.training.train_starvla][INFO] - Step 3975, Loss: {'train/loss': 0.2465554177761078, 'train/grad_norm_pre_clip': 15.940549850463867, 'timing/data': 0.009773731231689453, 'timing/model': 0.9390130843967199, 'learning_rate/qwen_vl_interface': 4.336581798896955e-06, 'learning_rate/action_model': 4.336581798896955e-05, 'epoch': 1.8}) +[2026-06-17 22:55:18,777][starVLA.training.train_starvla][INFO] - Step 3976, Loss: {'train/loss': 0.45709890127182007, 'train/grad_norm_pre_clip': 28.645753860473633, 'timing/data': 0.011762761510908604, 'timing/model': 0.9843517001718283, 'learning_rate/qwen_vl_interface': 4.334459566500062e-06, 'learning_rate/action_model': 4.334459566500062e-05, 'epoch': 1.8}) +[2026-06-17 22:55:22,623][starVLA.training.train_starvla][INFO] - Step 3977, Loss: {'train/loss': 0.2708374857902527, 'train/grad_norm_pre_clip': 21.253250122070312, 'timing/data': 0.008709701709449291, 'timing/model': 1.113139315508306, 'learning_rate/qwen_vl_interface': 4.332337523895688e-06, 'learning_rate/action_model': 4.3323375238956876e-05, 'epoch': 1.8}) +[2026-06-17 22:55:26,776][starVLA.training.train_starvla][INFO] - Step 3978, Loss: {'train/loss': 0.23843829333782196, 'train/grad_norm_pre_clip': 16.72804069519043, 'timing/data': 0.0074348123744130135, 'timing/model': 1.0454554446041584, 'learning_rate/qwen_vl_interface': 4.330215671523731e-06, 'learning_rate/action_model': 4.330215671523731e-05, 'epoch': 1.8}) +[2026-06-17 22:55:30,385][starVLA.training.train_starvla][INFO] - Step 3979, Loss: {'train/loss': 0.10473266243934631, 'train/grad_norm_pre_clip': 11.635004043579102, 'timing/data': 0.013525422662496567, 'timing/model': 0.9325423529371619, 'learning_rate/qwen_vl_interface': 4.328094009824054e-06, 'learning_rate/action_model': 4.328094009824054e-05, 'epoch': 1.81}) +[2026-06-17 22:55:34,146][starVLA.training.train_starvla][INFO] - Step 3980, Loss: {'train/loss': 0.27186840772628784, 'train/grad_norm_pre_clip': 22.199386596679688, 'timing/data': 0.00905852671712637, 'timing/model': 0.921609622426331, 'learning_rate/qwen_vl_interface': 4.3259725392364825e-06, 'learning_rate/action_model': 4.325972539236482e-05, 'epoch': 1.81}) +[2026-06-17 22:55:37,957][starVLA.training.train_starvla][INFO] - Step 3981, Loss: {'train/loss': 0.14425049722194672, 'train/grad_norm_pre_clip': 18.654069900512695, 'timing/data': 0.010899389162659645, 'timing/model': 0.9338952014222741, 'learning_rate/qwen_vl_interface': 4.323851260200797e-06, 'learning_rate/action_model': 4.323851260200796e-05, 'epoch': 1.81}) +[2026-06-17 22:55:41,812][starVLA.training.train_starvla][INFO] - Step 3982, Loss: {'train/loss': 0.16866201162338257, 'train/grad_norm_pre_clip': 20.805660247802734, 'timing/data': 0.011584565974771976, 'timing/model': 0.9298296943306923, 'learning_rate/qwen_vl_interface': 4.321730173156743e-06, 'learning_rate/action_model': 4.321730173156743e-05, 'epoch': 1.81}) +[2026-06-17 22:55:45,737][starVLA.training.train_starvla][INFO] - Step 3983, Loss: {'train/loss': 0.2111043632030487, 'train/grad_norm_pre_clip': 15.533358573913574, 'timing/data': 0.012787654995918274, 'timing/model': 0.9311229614540935, 'learning_rate/qwen_vl_interface': 4.31960927854402e-06, 'learning_rate/action_model': 4.31960927854402e-05, 'epoch': 1.81}) +[2026-06-17 22:55:49,511][starVLA.training.train_starvla][INFO] - Step 3984, Loss: {'train/loss': 0.6678263545036316, 'train/grad_norm_pre_clip': 34.12751388549805, 'timing/data': 0.007615634240210056, 'timing/model': 1.0987158827483654, 'learning_rate/qwen_vl_interface': 4.317488576802298e-06, 'learning_rate/action_model': 4.3174885768022986e-05, 'epoch': 1.81}) +[2026-06-17 22:55:53,123][starVLA.training.train_starvla][INFO] - Step 3985, Loss: {'train/loss': 0.13159243762493134, 'train/grad_norm_pre_clip': 15.571233749389648, 'timing/data': 0.008036233484745026, 'timing/model': 0.9383856868371367, 'learning_rate/qwen_vl_interface': 4.315368068371197e-06, 'learning_rate/action_model': 4.315368068371197e-05, 'epoch': 1.81}) +[2026-06-17 22:55:56,911][starVLA.training.train_starvla][INFO] - Step 3986, Loss: {'train/loss': 0.23967903852462769, 'train/grad_norm_pre_clip': 26.308822631835938, 'timing/data': 0.008709284476935863, 'timing/model': 0.924680975265801, 'learning_rate/qwen_vl_interface': 4.313247753690298e-06, 'learning_rate/action_model': 4.3132477536902985e-05, 'epoch': 1.81}) +[2026-06-17 22:56:00,656][starVLA.training.train_starvla][INFO] - Step 3987, Loss: {'train/loss': 0.2130310982465744, 'train/grad_norm_pre_clip': 18.452707290649414, 'timing/data': 0.009262028150260448, 'timing/model': 0.9254196165129542, 'learning_rate/qwen_vl_interface': 4.311127633199151e-06, 'learning_rate/action_model': 4.3111276331991504e-05, 'epoch': 1.81}) +[2026-06-17 22:56:04,410][starVLA.training.train_starvla][INFO] - Step 3988, Loss: {'train/loss': 0.5260831713676453, 'train/grad_norm_pre_clip': 27.35315704345703, 'timing/data': 0.013965590856969357, 'timing/model': 0.9307341584935784, 'learning_rate/qwen_vl_interface': 4.309007707337255e-06, 'learning_rate/action_model': 4.309007707337255e-05, 'epoch': 1.81}) +[2026-06-17 22:56:08,184][starVLA.training.train_starvla][INFO] - Step 3989, Loss: {'train/loss': 0.3782062232494354, 'train/grad_norm_pre_clip': 23.129087448120117, 'timing/data': 0.009445559233427048, 'timing/model': 0.9271360449492931, 'learning_rate/qwen_vl_interface': 4.3068879765440745e-06, 'learning_rate/action_model': 4.306887976544074e-05, 'epoch': 1.81}) +[2026-06-17 22:56:11,938][starVLA.training.train_starvla][INFO] - Step 3990, Loss: {'train/loss': 0.28284314274787903, 'train/grad_norm_pre_clip': 23.753124237060547, 'timing/data': 0.009223254397511482, 'timing/model': 1.09229436237365, 'learning_rate/qwen_vl_interface': 4.304768441259029e-06, 'learning_rate/action_model': 4.3047684412590285e-05, 'epoch': 1.81}) +[2026-06-17 22:56:15,685][starVLA.training.train_starvla][INFO] - Step 3991, Loss: {'train/loss': 0.20511102676391602, 'train/grad_norm_pre_clip': 23.660503387451172, 'timing/data': 0.0078074391931295395, 'timing/model': 1.045853086747229, 'learning_rate/qwen_vl_interface': 4.302649101921505e-06, 'learning_rate/action_model': 4.302649101921505e-05, 'epoch': 1.81}) +[2026-06-17 22:56:19,264][starVLA.training.train_starvla][INFO] - Step 3992, Loss: {'train/loss': 0.6242260932922363, 'train/grad_norm_pre_clip': 40.81690979003906, 'timing/data': 0.01026955246925354, 'timing/model': 0.9271378451958299, 'learning_rate/qwen_vl_interface': 4.300529958970841e-06, 'learning_rate/action_model': 4.300529958970841e-05, 'epoch': 1.81}) +[2026-06-17 22:56:23,006][starVLA.training.train_starvla][INFO] - Step 3993, Loss: {'train/loss': 0.31561797857284546, 'train/grad_norm_pre_clip': 23.395774841308594, 'timing/data': 0.01141422800719738, 'timing/model': 0.9276180742308497, 'learning_rate/qwen_vl_interface': 4.298411012846338e-06, 'learning_rate/action_model': 4.2984110128463376e-05, 'epoch': 1.81}) +[2026-06-17 22:56:26,754][starVLA.training.train_starvla][INFO] - Step 3994, Loss: {'train/loss': 0.11617725342512131, 'train/grad_norm_pre_clip': 12.336038589477539, 'timing/data': 0.008363095112144947, 'timing/model': 0.9254268957301974, 'learning_rate/qwen_vl_interface': 4.296292263987255e-06, 'learning_rate/action_model': 4.2962922639872556e-05, 'epoch': 1.81}) +[2026-06-17 22:56:30,484][starVLA.training.train_starvla][INFO] - Step 3995, Loss: {'train/loss': 0.16020189225673676, 'train/grad_norm_pre_clip': 15.700109481811523, 'timing/data': 0.012573719024658203, 'timing/model': 0.9245324293151498, 'learning_rate/qwen_vl_interface': 4.294173712832813e-06, 'learning_rate/action_model': 4.294173712832813e-05, 'epoch': 1.81}) +[2026-06-17 22:56:34,228][starVLA.training.train_starvla][INFO] - Step 3996, Loss: {'train/loss': 0.3017779588699341, 'train/grad_norm_pre_clip': 22.613656997680664, 'timing/data': 0.012827405706048012, 'timing/model': 0.9273129450157285, 'learning_rate/qwen_vl_interface': 4.292055359822191e-06, 'learning_rate/action_model': 4.29205535982219e-05, 'epoch': 1.81}) +[2026-06-17 22:56:37,960][starVLA.training.train_starvla][INFO] - Step 3997, Loss: {'train/loss': 0.3417394757270813, 'train/grad_norm_pre_clip': 36.21857452392578, 'timing/data': 0.008885429240763187, 'timing/model': 1.0866872435435653, 'learning_rate/qwen_vl_interface': 4.289937205394519e-06, 'learning_rate/action_model': 4.2899372053945185e-05, 'epoch': 1.81}) +[2026-06-17 22:56:41,671][starVLA.training.train_starvla][INFO] - Step 3998, Loss: {'train/loss': 0.32797202467918396, 'train/grad_norm_pre_clip': 29.124313354492188, 'timing/data': 0.007433343678712845, 'timing/model': 1.043921179138124, 'learning_rate/qwen_vl_interface': 4.287819249988901e-06, 'learning_rate/action_model': 4.2878192499889007e-05, 'epoch': 1.81}) +[2026-06-17 22:56:45,819][starVLA.training.train_starvla][INFO] - Step 3999, Loss: {'train/loss': 0.10005746781826019, 'train/grad_norm_pre_clip': 9.822306632995605, 'timing/data': 0.008316993713378906, 'timing/model': 0.9519926067441702, 'learning_rate/qwen_vl_interface': 4.285701494044387e-06, 'learning_rate/action_model': 4.285701494044387e-05, 'epoch': 1.81}) +[2026-06-17 22:58:39,172][starVLA.training.train_starvla][INFO] - Step 4000, Loss: {'train/loss': 0.06419527530670166, 'train/grad_norm_pre_clip': 9.38299560546875, 'eval/loss': 0.3143514394760132, 'eval/latency_8/loss': 0.3143514394760132, 'eval/demon_fire_precision': 0.9657513113236655, 'eval/demon_fire_recall': 0.9723516620068344, 'eval/demon_fire_f1': 0.9690402476780186, 'eval/demon_fire_rate_model': 0.9017807456872565, 'eval/demon_fire_rate_teacher': 0.8956594323873122, 'eval/demon_left_precision': 0.9020172910662824, 'eval/demon_left_recall': 0.6702355460385439, 'eval/demon_left_f1': 0.7690417690417691, 'eval/demon_left_rate_model': 0.09654980523094045, 'eval/demon_left_rate_teacher': 0.12993878686700056, 'eval/demon_right_precision': 0.8231797919762258, 'eval/demon_right_recall': 0.8393939393939394, 'eval/demon_right_f1': 0.8312078019504875, 'eval/demon_right_rate_model': 0.1872565386755704, 'eval/demon_right_rate_teacher': 0.18363939899833054, 'eval/demon_move_f1': 0.8001247854961283, 'eval/demon_cc_f1': 0.911323043130266, 'eval/latency_8/demon_fire_precision': 0.9657513113236655, 'eval/latency_8/demon_fire_recall': 0.9723516620068344, 'eval/latency_8/demon_fire_f1': 0.9690402476780186, 'eval/latency_8/demon_fire_rate_model': 0.9017807456872565, 'eval/latency_8/demon_fire_rate_teacher': 0.8956594323873122, 'eval/latency_8/demon_left_precision': 0.9020172910662824, 'eval/latency_8/demon_left_recall': 0.6702355460385439, 'eval/latency_8/demon_left_f1': 0.7690417690417691, 'eval/latency_8/demon_left_rate_model': 0.09654980523094045, 'eval/latency_8/demon_left_rate_teacher': 0.12993878686700056, 'eval/latency_8/demon_right_precision': 0.8231797919762258, 'eval/latency_8/demon_right_recall': 0.8393939393939394, 'eval/latency_8/demon_right_f1': 0.8312078019504875, 'eval/latency_8/demon_right_rate_model': 0.1872565386755704, 'eval/latency_8/demon_right_rate_teacher': 0.18363939899833054, 'eval/latency_8/demon_move_f1': 0.8001247854961283, 'eval/latency_8/demon_cc_f1': 0.911323043130266, 'timing/data': 0.0076475245878100395, 'timing/model': 0.9443940361961722, 'learning_rate/qwen_vl_interface': 4.283583937999991e-06, 'learning_rate/action_model': 4.2835839379999906e-05, 'epoch': 1.81}) +[2026-06-17 22:58:39,173][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_4000_state +[2026-06-17 22:58:39,178][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 22:58:45,794][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_4000_state/model.safetensors +[2026-06-17 22:59:01,319][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_4000_state/optimizer.bin +[2026-06-17 22:59:01,320][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_4000_state/sampler.bin +[2026-06-17 22:59:01,321][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_4000_state/sampler_1.bin +[2026-06-17 22:59:01,322][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_4000_state/random_states_0.pkl +[2026-06-17 22:59:01,327][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4000 +[2026-06-17 22:59:01,327][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 22:59:01,338][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 22:59:05,245][starVLA.training.train_starvla][INFO] - Step 4001, Loss: {'train/loss': 0.27683714032173157, 'train/grad_norm_pre_clip': 30.4389705657959, 'timing/data': 0.010935096070170403, 'timing/model': 0.9471525009721518, 'learning_rate/qwen_vl_interface': 4.2814665822946825e-06, 'learning_rate/action_model': 4.281466582294682e-05, 'epoch': 1.82}) +[2026-06-17 22:59:09,049][starVLA.training.train_starvla][INFO] - Step 4002, Loss: {'train/loss': 0.32561495900154114, 'train/grad_norm_pre_clip': 19.946035385131836, 'timing/data': 0.01906414143741131, 'timing/model': 0.9250001320615411, 'learning_rate/qwen_vl_interface': 4.279349427367395e-06, 'learning_rate/action_model': 4.279349427367395e-05, 'epoch': 1.82}) +[2026-06-17 22:59:12,901][starVLA.training.train_starvla][INFO] - Step 4003, Loss: {'train/loss': 0.18422776460647583, 'train/grad_norm_pre_clip': 18.113353729248047, 'timing/data': 0.020521347410976887, 'timing/model': 0.9405591497197747, 'learning_rate/qwen_vl_interface': 4.277232473657018e-06, 'learning_rate/action_model': 4.277232473657018e-05, 'epoch': 1.82}) +[2026-06-17 22:59:16,664][starVLA.training.train_starvla][INFO] - Step 4004, Loss: {'train/loss': 0.20242439210414886, 'train/grad_norm_pre_clip': 19.548559188842773, 'timing/data': 0.017874490469694138, 'timing/model': 0.9255471937358379, 'learning_rate/qwen_vl_interface': 4.2751157216023934e-06, 'learning_rate/action_model': 4.2751157216023936e-05, 'epoch': 1.82}) +[2026-06-17 22:59:20,426][starVLA.training.train_starvla][INFO] - Step 4005, Loss: {'train/loss': 0.28459590673446655, 'train/grad_norm_pre_clip': 21.350698471069336, 'timing/data': 0.01809992454946041, 'timing/model': 0.9241989180445671, 'learning_rate/qwen_vl_interface': 4.272999171642329e-06, 'learning_rate/action_model': 4.272999171642329e-05, 'epoch': 1.82}) +[2026-06-17 22:59:24,194][starVLA.training.train_starvla][INFO] - Step 4006, Loss: {'train/loss': 0.31269392371177673, 'train/grad_norm_pre_clip': 21.867277145385742, 'timing/data': 0.016681752167642117, 'timing/model': 1.0899071479216218, 'learning_rate/qwen_vl_interface': 4.270882824215588e-06, 'learning_rate/action_model': 4.270882824215588e-05, 'epoch': 1.82}) +[2026-06-17 22:59:27,817][starVLA.training.train_starvla][INFO] - Step 4007, Loss: {'train/loss': 0.115683414041996, 'train/grad_norm_pre_clip': 16.311325073242188, 'timing/data': 0.020462962798774242, 'timing/model': 0.9279412338510156, 'learning_rate/qwen_vl_interface': 4.268766679760891e-06, 'learning_rate/action_model': 4.268766679760891e-05, 'epoch': 1.82}) +[2026-06-17 22:59:31,574][starVLA.training.train_starvla][INFO] - Step 4008, Loss: {'train/loss': 0.5588948130607605, 'train/grad_norm_pre_clip': 27.195276260375977, 'timing/data': 0.022080859169363976, 'timing/model': 0.9244838915765285, 'learning_rate/qwen_vl_interface': 4.266650738716916e-06, 'learning_rate/action_model': 4.266650738716916e-05, 'epoch': 1.82}) +[2026-06-17 22:59:35,293][starVLA.training.train_starvla][INFO] - Step 4009, Loss: {'train/loss': 0.2624358534812927, 'train/grad_norm_pre_clip': 24.227052688598633, 'timing/data': 0.01693783886730671, 'timing/model': 0.9351711133494973, 'learning_rate/qwen_vl_interface': 4.264535001522302e-06, 'learning_rate/action_model': 4.2645350015223025e-05, 'epoch': 1.82}) +[2026-06-17 22:59:39,043][starVLA.training.train_starvla][INFO] - Step 4010, Loss: {'train/loss': 0.3692159354686737, 'train/grad_norm_pre_clip': 24.25498390197754, 'timing/data': 0.018718644976615906, 'timing/model': 0.927114449441433, 'learning_rate/qwen_vl_interface': 4.262419468615644e-06, 'learning_rate/action_model': 4.2624194686156435e-05, 'epoch': 1.82}) +[2026-06-17 22:59:42,814][starVLA.training.train_starvla][INFO] - Step 4011, Loss: {'train/loss': 0.2265869677066803, 'train/grad_norm_pre_clip': 18.1900577545166, 'timing/data': 0.01805006619542837, 'timing/model': 0.932186302728951, 'learning_rate/qwen_vl_interface': 4.2603041404354925e-06, 'learning_rate/action_model': 4.260304140435492e-05, 'epoch': 1.82}) +[2026-06-17 22:59:46,581][starVLA.training.train_starvla][INFO] - Step 4012, Loss: {'train/loss': 0.2137300819158554, 'train/grad_norm_pre_clip': 15.434157371520996, 'timing/data': 0.03191390447318554, 'timing/model': 1.0923840086907148, 'learning_rate/qwen_vl_interface': 4.258189017420356e-06, 'learning_rate/action_model': 4.258189017420355e-05, 'epoch': 1.82}) +[2026-06-17 22:59:50,181][starVLA.training.train_starvla][INFO] - Step 4013, Loss: {'train/loss': 0.0878843367099762, 'train/grad_norm_pre_clip': 13.262042045593262, 'timing/data': 0.02258945908397436, 'timing/model': 0.9282304849475622, 'learning_rate/qwen_vl_interface': 4.2560741000087035e-06, 'learning_rate/action_model': 4.2560741000087034e-05, 'epoch': 1.82}) +[2026-06-17 22:59:53,937][starVLA.training.train_starvla][INFO] - Step 4014, Loss: {'train/loss': 0.13331091403961182, 'train/grad_norm_pre_clip': 17.00878143310547, 'timing/data': 0.02100240346044302, 'timing/model': 0.9321100935339928, 'learning_rate/qwen_vl_interface': 4.253959388638961e-06, 'learning_rate/action_model': 4.2539593886389604e-05, 'epoch': 1.82}) +[2026-06-17 22:59:57,683][starVLA.training.train_starvla][INFO] - Step 4015, Loss: {'train/loss': 0.11549966037273407, 'train/grad_norm_pre_clip': 16.11229133605957, 'timing/data': 0.01929517276585102, 'timing/model': 0.927891775034368, 'learning_rate/qwen_vl_interface': 4.251844883749507e-06, 'learning_rate/action_model': 4.2518448837495066e-05, 'epoch': 1.82}) +[2026-06-17 23:00:01,434][starVLA.training.train_starvla][INFO] - Step 4016, Loss: {'train/loss': 0.26132458448410034, 'train/grad_norm_pre_clip': 16.909624099731445, 'timing/data': 0.01917354203760624, 'timing/model': 0.9293001256883144, 'learning_rate/qwen_vl_interface': 4.249730585778684e-06, 'learning_rate/action_model': 4.2497305857786834e-05, 'epoch': 1.82}) +[2026-06-17 23:00:05,194][starVLA.training.train_starvla][INFO] - Step 4017, Loss: {'train/loss': 0.30845433473587036, 'train/grad_norm_pre_clip': 25.329252243041992, 'timing/data': 0.01791474875062704, 'timing/model': 0.9305278705433011, 'learning_rate/qwen_vl_interface': 4.247616495164788e-06, 'learning_rate/action_model': 4.247616495164787e-05, 'epoch': 1.82}) +[2026-06-17 23:00:08,954][starVLA.training.train_starvla][INFO] - Step 4018, Loss: {'train/loss': 0.2786012887954712, 'train/grad_norm_pre_clip': 27.729061126708984, 'timing/data': 0.019597887992858887, 'timing/model': 1.0921682678163052, 'learning_rate/qwen_vl_interface': 4.24550261234607e-06, 'learning_rate/action_model': 4.2455026123460694e-05, 'epoch': 1.82}) +[2026-06-17 23:00:12,718][starVLA.training.train_starvla][INFO] - Step 4019, Loss: {'train/loss': 0.25617557764053345, 'train/grad_norm_pre_clip': 19.785818099975586, 'timing/data': 0.016620670445263386, 'timing/model': 1.0924162482842803, 'learning_rate/qwen_vl_interface': 4.2433889377607405e-06, 'learning_rate/action_model': 4.24338893776074e-05, 'epoch': 1.82}) +[2026-06-17 23:00:16,300][starVLA.training.train_starvla][INFO] - Step 4020, Loss: {'train/loss': 0.49461573362350464, 'train/grad_norm_pre_clip': 23.925966262817383, 'timing/data': 0.015934843569993973, 'timing/model': 0.925644570030272, 'learning_rate/qwen_vl_interface': 4.241275471846968e-06, 'learning_rate/action_model': 4.241275471846968e-05, 'epoch': 1.82}) +[2026-06-17 23:00:20,056][starVLA.training.train_starvla][INFO] - Step 4021, Loss: {'train/loss': 0.27993327379226685, 'train/grad_norm_pre_clip': 27.22645378112793, 'timing/data': 0.01697499118745327, 'timing/model': 0.9258250119164586, 'learning_rate/qwen_vl_interface': 4.239162215042876e-06, 'learning_rate/action_model': 4.239162215042876e-05, 'epoch': 1.82}) +[2026-06-17 23:00:23,800][starVLA.training.train_starvla][INFO] - Step 4022, Loss: {'train/loss': 0.5192641019821167, 'train/grad_norm_pre_clip': 34.50122833251953, 'timing/data': 0.01945190131664276, 'timing/model': 0.9296757644042373, 'learning_rate/qwen_vl_interface': 4.237049167786544e-06, 'learning_rate/action_model': 4.2370491677865444e-05, 'epoch': 1.82}) +[2026-06-17 23:00:27,686][starVLA.training.train_starvla][INFO] - Step 4023, Loss: {'train/loss': 0.08755933493375778, 'train/grad_norm_pre_clip': 10.131734848022461, 'timing/data': 0.026111196726560593, 'timing/model': 1.049122360534966, 'learning_rate/qwen_vl_interface': 4.234936330516009e-06, 'learning_rate/action_model': 4.234936330516009e-05, 'epoch': 1.83}) +[2026-06-17 23:00:31,469][starVLA.training.train_starvla][INFO] - Step 4024, Loss: {'train/loss': 0.18235807120800018, 'train/grad_norm_pre_clip': 19.260282516479492, 'timing/data': 0.01808664482086897, 'timing/model': 0.9261647555977106, 'learning_rate/qwen_vl_interface': 4.2328237036692675e-06, 'learning_rate/action_model': 4.232823703669267e-05, 'epoch': 1.83}) +[2026-06-17 23:00:35,214][starVLA.training.train_starvla][INFO] - Step 4025, Loss: {'train/loss': 0.1202613040804863, 'train/grad_norm_pre_clip': 15.344489097595215, 'timing/data': 0.027156323194503784, 'timing/model': 1.0808202428743243, 'learning_rate/qwen_vl_interface': 4.2307112876842645e-06, 'learning_rate/action_model': 4.2307112876842645e-05, 'epoch': 1.83}) +[2026-06-17 23:00:38,803][starVLA.training.train_starvla][INFO] - Step 4026, Loss: {'train/loss': 0.09370116144418716, 'train/grad_norm_pre_clip': 11.178559303283691, 'timing/data': 0.02886039763689041, 'timing/model': 0.9238770762458444, 'learning_rate/qwen_vl_interface': 4.228599082998908e-06, 'learning_rate/action_model': 4.228599082998908e-05, 'epoch': 1.83}) +[2026-06-17 23:00:42,541][starVLA.training.train_starvla][INFO] - Step 4027, Loss: {'train/loss': 0.3262191414833069, 'train/grad_norm_pre_clip': 33.07562255859375, 'timing/data': 0.020688232965767384, 'timing/model': 0.9236711664125323, 'learning_rate/qwen_vl_interface': 4.226487090051062e-06, 'learning_rate/action_model': 4.2264870900510624e-05, 'epoch': 1.83}) +[2026-06-17 23:00:46,221][starVLA.training.train_starvla][INFO] - Step 4028, Loss: {'train/loss': 0.08357063680887222, 'train/grad_norm_pre_clip': 10.901731491088867, 'timing/data': 0.01915866043418646, 'timing/model': 0.9264493612572551, 'learning_rate/qwen_vl_interface': 4.224375309278543e-06, 'learning_rate/action_model': 4.2243753092785434e-05, 'epoch': 1.83}) +[2026-06-17 23:00:49,964][starVLA.training.train_starvla][INFO] - Step 4029, Loss: {'train/loss': 0.2591332793235779, 'train/grad_norm_pre_clip': 23.433269500732422, 'timing/data': 0.02152366191148758, 'timing/model': 0.9260013801977038, 'learning_rate/qwen_vl_interface': 4.222263741119127e-06, 'learning_rate/action_model': 4.222263741119126e-05, 'epoch': 1.83}) +[2026-06-17 23:00:53,720][starVLA.training.train_starvla][INFO] - Step 4030, Loss: {'train/loss': 0.34939876198768616, 'train/grad_norm_pre_clip': 31.05730628967285, 'timing/data': 0.022180688567459583, 'timing/model': 0.9265336077660322, 'learning_rate/qwen_vl_interface': 4.220152386010541e-06, 'learning_rate/action_model': 4.2201523860105405e-05, 'epoch': 1.83}) +[2026-06-17 23:00:57,449][starVLA.training.train_starvla][INFO] - Step 4031, Loss: {'train/loss': 0.200688436627388, 'train/grad_norm_pre_clip': 17.75370216369629, 'timing/data': 0.017074554227292538, 'timing/model': 1.0821188893169165, 'learning_rate/qwen_vl_interface': 4.218041244390476e-06, 'learning_rate/action_model': 4.218041244390475e-05, 'epoch': 1.83}) +[2026-06-17 23:01:01,178][starVLA.training.train_starvla][INFO] - Step 4032, Loss: {'train/loss': 0.4358941912651062, 'train/grad_norm_pre_clip': 33.86339569091797, 'timing/data': 0.015427768230438232, 'timing/model': 1.0793373370543122, 'learning_rate/qwen_vl_interface': 4.215930316696571e-06, 'learning_rate/action_model': 4.215930316696571e-05, 'epoch': 1.83}) +[2026-06-17 23:01:04,765][starVLA.training.train_starvla][INFO] - Step 4033, Loss: {'train/loss': 0.09581979364156723, 'train/grad_norm_pre_clip': 11.256389617919922, 'timing/data': 0.021394806914031506, 'timing/model': 0.9260682379826903, 'learning_rate/qwen_vl_interface': 4.213819603366424e-06, 'learning_rate/action_model': 4.213819603366424e-05, 'epoch': 1.83}) +[2026-06-17 23:01:08,526][starVLA.training.train_starvla][INFO] - Step 4034, Loss: {'train/loss': 0.432761013507843, 'train/grad_norm_pre_clip': 22.008039474487305, 'timing/data': 0.018172578886151314, 'timing/model': 0.9265548484399915, 'learning_rate/qwen_vl_interface': 4.211709104837587e-06, 'learning_rate/action_model': 4.211709104837587e-05, 'epoch': 1.83}) +[2026-06-17 23:01:12,298][starVLA.training.train_starvla][INFO] - Step 4035, Loss: {'train/loss': 0.17247694730758667, 'train/grad_norm_pre_clip': 18.378890991210938, 'timing/data': 0.021756169386208057, 'timing/model': 0.9273910596966743, 'learning_rate/qwen_vl_interface': 4.2095988215475705e-06, 'learning_rate/action_model': 4.20959882154757e-05, 'epoch': 1.83}) +[2026-06-17 23:01:16,003][starVLA.training.train_starvla][INFO] - Step 4036, Loss: {'train/loss': 0.41883498430252075, 'train/grad_norm_pre_clip': 24.003828048706055, 'timing/data': 0.02278313133865595, 'timing/model': 0.9252233551815152, 'learning_rate/qwen_vl_interface': 4.207488753933838e-06, 'learning_rate/action_model': 4.207488753933838e-05, 'epoch': 1.83}) +[2026-06-17 23:01:19,737][starVLA.training.train_starvla][INFO] - Step 4037, Loss: {'train/loss': 0.17217546701431274, 'train/grad_norm_pre_clip': 22.384300231933594, 'timing/data': 0.016608036123216152, 'timing/model': 0.9258808074519038, 'learning_rate/qwen_vl_interface': 4.205378902433807e-06, 'learning_rate/action_model': 4.205378902433807e-05, 'epoch': 1.83}) +[2026-06-17 23:01:23,469][starVLA.training.train_starvla][INFO] - Step 4038, Loss: {'train/loss': 0.36259958148002625, 'train/grad_norm_pre_clip': 25.059860229492188, 'timing/data': 0.015637408010661602, 'timing/model': 1.0830855295062065, 'learning_rate/qwen_vl_interface': 4.203269267484856e-06, 'learning_rate/action_model': 4.203269267484856e-05, 'epoch': 1.83}) +[2026-06-17 23:01:27,040][starVLA.training.train_starvla][INFO] - Step 4039, Loss: {'train/loss': 0.29049819707870483, 'train/grad_norm_pre_clip': 29.465412139892578, 'timing/data': 0.020562604069709778, 'timing/model': 0.923039480112493, 'learning_rate/qwen_vl_interface': 4.201159849524313e-06, 'learning_rate/action_model': 4.201159849524312e-05, 'epoch': 1.83}) +[2026-06-17 23:01:30,793][starVLA.training.train_starvla][INFO] - Step 4040, Loss: {'train/loss': 0.6709024906158447, 'train/grad_norm_pre_clip': 31.078081130981445, 'timing/data': 0.02630100306123495, 'timing/model': 0.9292276101186872, 'learning_rate/qwen_vl_interface': 4.19905064898946e-06, 'learning_rate/action_model': 4.199050648989459e-05, 'epoch': 1.83}) +[2026-06-17 23:01:34,554][starVLA.training.train_starvla][INFO] - Step 4041, Loss: {'train/loss': 0.1793573647737503, 'train/grad_norm_pre_clip': 24.267221450805664, 'timing/data': 0.024953984655439854, 'timing/model': 0.9402015246450901, 'learning_rate/qwen_vl_interface': 4.196941666317537e-06, 'learning_rate/action_model': 4.196941666317537e-05, 'epoch': 1.83}) +[2026-06-17 23:01:38,319][starVLA.training.train_starvla][INFO] - Step 4042, Loss: {'train/loss': 0.4665178954601288, 'train/grad_norm_pre_clip': 24.665573120117188, 'timing/data': 0.02144843526184559, 'timing/model': 0.9232615474611521, 'learning_rate/qwen_vl_interface': 4.194832901945741e-06, 'learning_rate/action_model': 4.194832901945741e-05, 'epoch': 1.83}) +[2026-06-17 23:01:42,039][starVLA.training.train_starvla][INFO] - Step 4043, Loss: {'train/loss': 0.2641373872756958, 'train/grad_norm_pre_clip': 21.058040618896484, 'timing/data': 0.018552652560174465, 'timing/model': 0.92186866607517, 'learning_rate/qwen_vl_interface': 4.192724356311221e-06, 'learning_rate/action_model': 4.19272435631122e-05, 'epoch': 1.83}) +[2026-06-17 23:01:45,837][starVLA.training.train_starvla][INFO] - Step 4044, Loss: {'train/loss': 0.3330029249191284, 'train/grad_norm_pre_clip': 20.134645462036133, 'timing/data': 0.015652460977435112, 'timing/model': 0.9269103417173028, 'learning_rate/qwen_vl_interface': 4.190616029851079e-06, 'learning_rate/action_model': 4.190616029851079e-05, 'epoch': 1.83}) +[2026-06-17 23:01:49,578][starVLA.training.train_starvla][INFO] - Step 4045, Loss: {'train/loss': 0.5292132496833801, 'train/grad_norm_pre_clip': 63.42908477783203, 'timing/data': 0.015810374170541763, 'timing/model': 1.0802615210413933, 'learning_rate/qwen_vl_interface': 4.1885079230023715e-06, 'learning_rate/action_model': 4.1885079230023715e-05, 'epoch': 1.84}) +[2026-06-17 23:01:53,172][starVLA.training.train_starvla][INFO] - Step 4046, Loss: {'train/loss': 0.46720048785209656, 'train/grad_norm_pre_clip': 29.491697311401367, 'timing/data': 0.019296742044389248, 'timing/model': 0.9297015154734254, 'learning_rate/qwen_vl_interface': 4.186400036202117e-06, 'learning_rate/action_model': 4.1864000362021175e-05, 'epoch': 1.84}) +[2026-06-17 23:01:56,921][starVLA.training.train_starvla][INFO] - Step 4047, Loss: {'train/loss': 0.4686751067638397, 'train/grad_norm_pre_clip': 23.209522247314453, 'timing/data': 0.015893670730292797, 'timing/model': 0.9261860819533467, 'learning_rate/qwen_vl_interface': 4.184292369887279e-06, 'learning_rate/action_model': 4.1842923698872785e-05, 'epoch': 1.84}) +[2026-06-17 23:02:00,670][starVLA.training.train_starvla][INFO] - Step 4048, Loss: {'train/loss': 0.2197900116443634, 'train/grad_norm_pre_clip': 22.812053680419922, 'timing/data': 0.016629493795335293, 'timing/model': 0.9267336018383503, 'learning_rate/qwen_vl_interface': 4.182184924494778e-06, 'learning_rate/action_model': 4.1821849244947773e-05, 'epoch': 1.84}) +[2026-06-17 23:02:04,413][starVLA.training.train_starvla][INFO] - Step 4049, Loss: {'train/loss': 0.33508509397506714, 'train/grad_norm_pre_clip': 25.943923950195312, 'timing/data': 0.024765539914369583, 'timing/model': 0.9286992894485593, 'learning_rate/qwen_vl_interface': 4.1800777004614915e-06, 'learning_rate/action_model': 4.180077700461492e-05, 'epoch': 1.84}) +[2026-06-17 23:02:08,195][starVLA.training.train_starvla][INFO] - Step 4050, Loss: {'train/loss': 0.6683397889137268, 'train/grad_norm_pre_clip': 36.08775329589844, 'timing/data': 0.015003837645053864, 'timing/model': 0.932257454842329, 'learning_rate/qwen_vl_interface': 4.177970698224249e-06, 'learning_rate/action_model': 4.17797069822425e-05, 'epoch': 1.84}) +[2026-06-17 23:02:11,955][starVLA.training.train_starvla][INFO] - Step 4051, Loss: {'train/loss': 0.20687122642993927, 'train/grad_norm_pre_clip': 16.918365478515625, 'timing/data': 0.019443356432020664, 'timing/model': 1.091206157580018, 'learning_rate/qwen_vl_interface': 4.175863918219836e-06, 'learning_rate/action_model': 4.175863918219836e-05, 'epoch': 1.84}) +[2026-06-17 23:02:15,701][starVLA.training.train_starvla][INFO] - Step 4052, Loss: {'train/loss': 0.5461449027061462, 'train/grad_norm_pre_clip': 31.724990844726562, 'timing/data': 0.014040179550647736, 'timing/model': 1.0874171052128077, 'learning_rate/qwen_vl_interface': 4.173757360884986e-06, 'learning_rate/action_model': 4.173757360884986e-05, 'epoch': 1.84}) +[2026-06-17 23:02:19,458][starVLA.training.train_starvla][INFO] - Step 4053, Loss: {'train/loss': 0.17029497027397156, 'train/grad_norm_pre_clip': 17.043474197387695, 'timing/data': 0.01474742777645588, 'timing/model': 0.9352966910228133, 'learning_rate/qwen_vl_interface': 4.171651026656396e-06, 'learning_rate/action_model': 4.171651026656395e-05, 'epoch': 1.84}) +[2026-06-17 23:02:23,409][starVLA.training.train_starvla][INFO] - Step 4054, Loss: {'train/loss': 0.5200703144073486, 'train/grad_norm_pre_clip': 24.34296989440918, 'timing/data': 0.018733923323452473, 'timing/model': 0.9410918820649385, 'learning_rate/qwen_vl_interface': 4.1695449159707066e-06, 'learning_rate/action_model': 4.169544915970707e-05, 'epoch': 1.84}) +[2026-06-17 23:02:27,463][starVLA.training.train_starvla][INFO] - Step 4055, Loss: {'train/loss': 0.19120724499225616, 'train/grad_norm_pre_clip': 17.621055603027344, 'timing/data': 0.01621063519269228, 'timing/model': 0.9906687410548329, 'learning_rate/qwen_vl_interface': 4.167439029264519e-06, 'learning_rate/action_model': 4.167439029264519e-05, 'epoch': 1.84}) +[2026-06-17 23:02:31,373][starVLA.training.train_starvla][INFO] - Step 4056, Loss: {'train/loss': 0.3083847165107727, 'train/grad_norm_pre_clip': 17.31412124633789, 'timing/data': 0.015154982917010784, 'timing/model': 1.0353868817910552, 'learning_rate/qwen_vl_interface': 4.165333366974385e-06, 'learning_rate/action_model': 4.165333366974385e-05, 'epoch': 1.84}) +[2026-06-17 23:02:35,206][starVLA.training.train_starvla][INFO] - Step 4057, Loss: {'train/loss': 0.306318074464798, 'train/grad_norm_pre_clip': 25.305843353271484, 'timing/data': 0.020228632725775242, 'timing/model': 0.9373698327690363, 'learning_rate/qwen_vl_interface': 4.163227929536812e-06, 'learning_rate/action_model': 4.163227929536812e-05, 'epoch': 1.84}) +[2026-06-17 23:02:39,055][starVLA.training.train_starvla][INFO] - Step 4058, Loss: {'train/loss': 0.5636496543884277, 'train/grad_norm_pre_clip': 33.24620819091797, 'timing/data': 0.020690109580755234, 'timing/model': 1.1602122345939279, 'learning_rate/qwen_vl_interface': 4.1611227173882596e-06, 'learning_rate/action_model': 4.1611227173882596e-05, 'epoch': 1.84}) +[2026-06-17 23:02:42,837][starVLA.training.train_starvla][INFO] - Step 4059, Loss: {'train/loss': 0.34466424584388733, 'train/grad_norm_pre_clip': 21.00715446472168, 'timing/data': 0.01530471257865429, 'timing/model': 1.0776983806863427, 'learning_rate/qwen_vl_interface': 4.1590177309651365e-06, 'learning_rate/action_model': 4.159017730965136e-05, 'epoch': 1.84}) +[2026-06-17 23:02:46,482][starVLA.training.train_starvla][INFO] - Step 4060, Loss: {'train/loss': 0.0934024304151535, 'train/grad_norm_pre_clip': 9.474924087524414, 'timing/data': 0.015933189541101456, 'timing/model': 0.9415897252038121, 'learning_rate/qwen_vl_interface': 4.156912970703815e-06, 'learning_rate/action_model': 4.1569129707038155e-05, 'epoch': 1.84}) +[2026-06-17 23:02:50,294][starVLA.training.train_starvla][INFO] - Step 4061, Loss: {'train/loss': 0.2671080231666565, 'train/grad_norm_pre_clip': 24.108116149902344, 'timing/data': 0.018170827068388462, 'timing/model': 0.9365325160324574, 'learning_rate/qwen_vl_interface': 4.15480843704061e-06, 'learning_rate/action_model': 4.154808437040609e-05, 'epoch': 1.84}) +[2026-06-17 23:02:54,111][starVLA.training.train_starvla][INFO] - Step 4062, Loss: {'train/loss': 0.14638154208660126, 'train/grad_norm_pre_clip': 18.885623931884766, 'timing/data': 0.017642629332840443, 'timing/model': 0.9272196143865585, 'learning_rate/qwen_vl_interface': 4.152704130411794e-06, 'learning_rate/action_model': 4.152704130411793e-05, 'epoch': 1.84}) +[2026-06-17 23:02:57,828][starVLA.training.train_starvla][INFO] - Step 4063, Loss: {'train/loss': 0.22039030492305756, 'train/grad_norm_pre_clip': 16.701406478881836, 'timing/data': 0.017201277427375317, 'timing/model': 0.9392856601625681, 'learning_rate/qwen_vl_interface': 4.150600051253591e-06, 'learning_rate/action_model': 4.150600051253591e-05, 'epoch': 1.84}) +[2026-06-17 23:03:01,582][starVLA.training.train_starvla][INFO] - Step 4064, Loss: {'train/loss': 0.3896675407886505, 'train/grad_norm_pre_clip': 16.726858139038086, 'timing/data': 0.017218207009136677, 'timing/model': 0.9364518206566572, 'learning_rate/qwen_vl_interface': 4.148496200002183e-06, 'learning_rate/action_model': 4.148496200002183e-05, 'epoch': 1.84}) +[2026-06-17 23:03:05,465][starVLA.training.train_starvla][INFO] - Step 4065, Loss: {'train/loss': 0.3693767488002777, 'train/grad_norm_pre_clip': 26.292495727539062, 'timing/data': 0.018085090443491936, 'timing/model': 1.161370306275785, 'learning_rate/qwen_vl_interface': 4.146392577093697e-06, 'learning_rate/action_model': 4.1463925770936974e-05, 'epoch': 1.84}) +[2026-06-17 23:03:09,051][starVLA.training.train_starvla][INFO] - Step 4066, Loss: {'train/loss': 0.11757159978151321, 'train/grad_norm_pre_clip': 9.353447914123535, 'timing/data': 0.014998373575508595, 'timing/model': 0.9254597527906299, 'learning_rate/qwen_vl_interface': 4.144289182964218e-06, 'learning_rate/action_model': 4.144289182964218e-05, 'epoch': 1.84}) +[2026-06-17 23:03:12,810][starVLA.training.train_starvla][INFO] - Step 4067, Loss: {'train/loss': 0.34766778349876404, 'train/grad_norm_pre_clip': 22.39272689819336, 'timing/data': 0.016748445108532906, 'timing/model': 0.92731371242553, 'learning_rate/qwen_vl_interface': 4.14218601804978e-06, 'learning_rate/action_model': 4.1421860180497796e-05, 'epoch': 1.85}) +[2026-06-17 23:03:16,540][starVLA.training.train_starvla][INFO] - Step 4068, Loss: {'train/loss': 0.20812110602855682, 'train/grad_norm_pre_clip': 20.42268943786621, 'timing/data': 0.014860793948173523, 'timing/model': 0.9291719757020473, 'learning_rate/qwen_vl_interface': 4.1400830827863725e-06, 'learning_rate/action_model': 4.1400830827863725e-05, 'epoch': 1.85}) +[2026-06-17 23:03:20,337][starVLA.training.train_starvla][INFO] - Step 4069, Loss: {'train/loss': 0.3569907248020172, 'train/grad_norm_pre_clip': 21.244808197021484, 'timing/data': 0.014663568697869778, 'timing/model': 0.9284925172105432, 'learning_rate/qwen_vl_interface': 4.137980377609937e-06, 'learning_rate/action_model': 4.137980377609937e-05, 'epoch': 1.85}) +[2026-06-17 23:03:24,111][starVLA.training.train_starvla][INFO] - Step 4070, Loss: {'train/loss': 0.19664521515369415, 'train/grad_norm_pre_clip': 15.863611221313477, 'timing/data': 0.01927085965871811, 'timing/model': 0.9272558949887753, 'learning_rate/qwen_vl_interface': 4.135877902956364e-06, 'learning_rate/action_model': 4.1358779029563636e-05, 'epoch': 1.85}) +[2026-06-17 23:03:27,880][starVLA.training.train_starvla][INFO] - Step 4071, Loss: {'train/loss': 0.2895067036151886, 'train/grad_norm_pre_clip': 26.784181594848633, 'timing/data': 0.01641869079321623, 'timing/model': 1.1199363907799125, 'learning_rate/qwen_vl_interface': 4.133775659261503e-06, 'learning_rate/action_model': 4.133775659261502e-05, 'epoch': 1.85}) +[2026-06-17 23:03:31,648][starVLA.training.train_starvla][INFO] - Step 4072, Loss: {'train/loss': 0.25147363543510437, 'train/grad_norm_pre_clip': 20.347400665283203, 'timing/data': 0.026033077389001846, 'timing/model': 1.0659904852509499, 'learning_rate/qwen_vl_interface': 4.131673646961149e-06, 'learning_rate/action_model': 4.131673646961149e-05, 'epoch': 1.85}) +[2026-06-17 23:03:35,250][starVLA.training.train_starvla][INFO] - Step 4073, Loss: {'train/loss': 0.26586589217185974, 'train/grad_norm_pre_clip': 22.08540153503418, 'timing/data': 0.02589874155819416, 'timing/model': 0.9275508187711239, 'learning_rate/qwen_vl_interface': 4.129571866491051e-06, 'learning_rate/action_model': 4.1295718664910504e-05, 'epoch': 1.85}) +[2026-06-17 23:03:39,089][starVLA.training.train_starvla][INFO] - Step 4074, Loss: {'train/loss': 0.3423755168914795, 'train/grad_norm_pre_clip': 23.743967056274414, 'timing/data': 0.013340509496629238, 'timing/model': 1.0120503567159176, 'learning_rate/qwen_vl_interface': 4.127470318286908e-06, 'learning_rate/action_model': 4.127470318286908e-05, 'epoch': 1.85}) +[2026-06-17 23:03:42,864][starVLA.training.train_starvla][INFO] - Step 4075, Loss: {'train/loss': 0.2725105881690979, 'train/grad_norm_pre_clip': 20.76202392578125, 'timing/data': 0.01490241289138794, 'timing/model': 0.9296438926830888, 'learning_rate/qwen_vl_interface': 4.125369002784378e-06, 'learning_rate/action_model': 4.1253690027843776e-05, 'epoch': 1.85}) +[2026-06-17 23:03:46,713][starVLA.training.train_starvla][INFO] - Step 4076, Loss: {'train/loss': 0.48706939816474915, 'train/grad_norm_pre_clip': 20.306224822998047, 'timing/data': 0.020278718322515488, 'timing/model': 0.9580453094094992, 'learning_rate/qwen_vl_interface': 4.123267920419064e-06, 'learning_rate/action_model': 4.123267920419064e-05, 'epoch': 1.85}) +[2026-06-17 23:03:50,563][starVLA.training.train_starvla][INFO] - Step 4077, Loss: {'train/loss': 0.33309513330459595, 'train/grad_norm_pre_clip': 23.207439422607422, 'timing/data': 0.0173740703612566, 'timing/model': 0.96351732686162, 'learning_rate/qwen_vl_interface': 4.121167071626522e-06, 'learning_rate/action_model': 4.121167071626522e-05, 'epoch': 1.85}) +[2026-06-17 23:03:54,505][starVLA.training.train_starvla][INFO] - Step 4078, Loss: {'train/loss': 0.24872717261314392, 'train/grad_norm_pre_clip': 15.925444602966309, 'timing/data': 0.027428774163126945, 'timing/model': 1.1318778665736318, 'learning_rate/qwen_vl_interface': 4.119066456842261e-06, 'learning_rate/action_model': 4.119066456842261e-05, 'epoch': 1.85}) +[2026-06-17 23:03:58,972][starVLA.training.train_starvla][INFO] - Step 4079, Loss: {'train/loss': 0.2738184630870819, 'train/grad_norm_pre_clip': 21.254987716674805, 'timing/data': 0.03107957635074854, 'timing/model': 1.2042981060221791, 'learning_rate/qwen_vl_interface': 4.1169660765017416e-06, 'learning_rate/action_model': 4.1169660765017414e-05, 'epoch': 1.85}) +[2026-06-17 23:04:02,680][starVLA.training.train_starvla][INFO] - Step 4080, Loss: {'train/loss': 0.40914684534072876, 'train/grad_norm_pre_clip': 20.542593002319336, 'timing/data': 0.017225426621735096, 'timing/model': 0.93632887955755, 'learning_rate/qwen_vl_interface': 4.1148659310403725e-06, 'learning_rate/action_model': 4.1148659310403723e-05, 'epoch': 1.85}) +[2026-06-17 23:04:06,509][starVLA.training.train_starvla][INFO] - Step 4081, Loss: {'train/loss': 0.4741537570953369, 'train/grad_norm_pre_clip': 24.332714080810547, 'timing/data': 0.014024579897522926, 'timing/model': 0.939237006008625, 'learning_rate/qwen_vl_interface': 4.112766020893516e-06, 'learning_rate/action_model': 4.1127660208935164e-05, 'epoch': 1.85}) +[2026-06-17 23:04:10,409][starVLA.training.train_starvla][INFO] - Step 4082, Loss: {'train/loss': 0.22639016807079315, 'train/grad_norm_pre_clip': 17.404949188232422, 'timing/data': 0.02659923303872347, 'timing/model': 0.9439397156238556, 'learning_rate/qwen_vl_interface': 4.11066634649649e-06, 'learning_rate/action_model': 4.1106663464964895e-05, 'epoch': 1.85}) +[2026-06-17 23:04:14,223][starVLA.training.train_starvla][INFO] - Step 4083, Loss: {'train/loss': 0.3917746841907501, 'train/grad_norm_pre_clip': 29.900304794311523, 'timing/data': 0.015387460589408875, 'timing/model': 0.9408776117488742, 'learning_rate/qwen_vl_interface': 4.1085669082845545e-06, 'learning_rate/action_model': 4.108566908284555e-05, 'epoch': 1.85}) +[2026-06-17 23:04:18,178][starVLA.training.train_starvla][INFO] - Step 4084, Loss: {'train/loss': 0.14245247840881348, 'train/grad_norm_pre_clip': 14.469629287719727, 'timing/data': 0.017765354365110397, 'timing/model': 0.9309710906818509, 'learning_rate/qwen_vl_interface': 4.10646770669293e-06, 'learning_rate/action_model': 4.106467706692929e-05, 'epoch': 1.85}) +[2026-06-17 23:04:22,059][starVLA.training.train_starvla][INFO] - Step 4085, Loss: {'train/loss': 0.2414509803056717, 'train/grad_norm_pre_clip': 19.703739166259766, 'timing/data': 0.021899809129536152, 'timing/model': 1.1229467699304223, 'learning_rate/qwen_vl_interface': 4.104368742156774e-06, 'learning_rate/action_model': 4.104368742156774e-05, 'epoch': 1.85}) +[2026-06-17 23:04:25,887][starVLA.training.train_starvla][INFO] - Step 4086, Loss: {'train/loss': 0.2621382176876068, 'train/grad_norm_pre_clip': 18.86027717590332, 'timing/data': 0.019586306996643543, 'timing/model': 1.048964463174343, 'learning_rate/qwen_vl_interface': 4.102270015111217e-06, 'learning_rate/action_model': 4.102270015111217e-05, 'epoch': 1.85}) +[2026-06-17 23:04:29,559][starVLA.training.train_starvla][INFO] - Step 4087, Loss: {'train/loss': 0.25617608428001404, 'train/grad_norm_pre_clip': 14.467824935913086, 'timing/data': 0.01599728036671877, 'timing/model': 0.9286894546821713, 'learning_rate/qwen_vl_interface': 4.100171525991318e-06, 'learning_rate/action_model': 4.100171525991318e-05, 'epoch': 1.85}) +[2026-06-17 23:04:33,338][starVLA.training.train_starvla][INFO] - Step 4088, Loss: {'train/loss': 0.36561310291290283, 'train/grad_norm_pre_clip': 23.337717056274414, 'timing/data': 0.01791771501302719, 'timing/model': 0.9258534470573068, 'learning_rate/qwen_vl_interface': 4.098073275232098e-06, 'learning_rate/action_model': 4.098073275232098e-05, 'epoch': 1.85}) +[2026-06-17 23:04:37,275][starVLA.training.train_starvla][INFO] - Step 4089, Loss: {'train/loss': 0.20721247792243958, 'train/grad_norm_pre_clip': 12.325984001159668, 'timing/data': 0.016995057463645935, 'timing/model': 0.929058950394392, 'learning_rate/qwen_vl_interface': 4.09597526326853e-06, 'learning_rate/action_model': 4.095975263268529e-05, 'epoch': 1.86}) +[2026-06-17 23:04:41,407][starVLA.training.train_starvla][INFO] - Step 4090, Loss: {'train/loss': 0.1809382438659668, 'train/grad_norm_pre_clip': 20.097614288330078, 'timing/data': 0.0185387022793293, 'timing/model': 1.2530331537127495, 'learning_rate/qwen_vl_interface': 4.09387749053553e-06, 'learning_rate/action_model': 4.0938774905355296e-05, 'epoch': 1.86}) +[2026-06-17 23:04:45,229][starVLA.training.train_starvla][INFO] - Step 4091, Loss: {'train/loss': 0.25430646538734436, 'train/grad_norm_pre_clip': 28.406862258911133, 'timing/data': 0.019663244485855103, 'timing/model': 1.119817621074617, 'learning_rate/qwen_vl_interface': 4.0917799574679725e-06, 'learning_rate/action_model': 4.091779957467973e-05, 'epoch': 1.86}) +[2026-06-17 23:04:48,997][starVLA.training.train_starvla][INFO] - Step 4092, Loss: {'train/loss': 0.1618257462978363, 'train/grad_norm_pre_clip': 20.622112274169922, 'timing/data': 0.017829984426498413, 'timing/model': 1.0704485634341836, 'learning_rate/qwen_vl_interface': 4.0896826645006725e-06, 'learning_rate/action_model': 4.089682664500673e-05, 'epoch': 1.86}) +[2026-06-17 23:04:52,921][starVLA.training.train_starvla][INFO] - Step 4093, Loss: {'train/loss': 0.19941812753677368, 'train/grad_norm_pre_clip': 21.794858932495117, 'timing/data': 0.017119616270065308, 'timing/model': 0.9345950046554208, 'learning_rate/qwen_vl_interface': 4.087585612068408e-06, 'learning_rate/action_model': 4.0875856120684074e-05, 'epoch': 1.86}) +[2026-06-17 23:04:57,033][starVLA.training.train_starvla][INFO] - Step 4094, Loss: {'train/loss': 0.1323496252298355, 'train/grad_norm_pre_clip': 12.472121238708496, 'timing/data': 0.017349490895867348, 'timing/model': 0.9405693989247084, 'learning_rate/qwen_vl_interface': 4.0854888006058955e-06, 'learning_rate/action_model': 4.085488800605896e-05, 'epoch': 1.86}) +[2026-06-17 23:05:00,855][starVLA.training.train_starvla][INFO] - Step 4095, Loss: {'train/loss': 0.247674360871315, 'train/grad_norm_pre_clip': 19.802982330322266, 'timing/data': 0.01693871896713972, 'timing/model': 0.9324932536110282, 'learning_rate/qwen_vl_interface': 4.083392230547808e-06, 'learning_rate/action_model': 4.083392230547808e-05, 'epoch': 1.86}) +[2026-06-17 23:05:04,668][starVLA.training.train_starvla][INFO] - Step 4096, Loss: {'train/loss': 0.5859590172767639, 'train/grad_norm_pre_clip': 22.857824325561523, 'timing/data': 0.017779434099793434, 'timing/model': 0.9364074282348156, 'learning_rate/qwen_vl_interface': 4.081295902328765e-06, 'learning_rate/action_model': 4.081295902328765e-05, 'epoch': 1.86}) +[2026-06-17 23:05:08,406][starVLA.training.train_starvla][INFO] - Step 4097, Loss: {'train/loss': 0.5067346096038818, 'train/grad_norm_pre_clip': 27.821701049804688, 'timing/data': 0.01708371937274933, 'timing/model': 0.9430482788011432, 'learning_rate/qwen_vl_interface': 4.079199816383341e-06, 'learning_rate/action_model': 4.0791998163833404e-05, 'epoch': 1.86}) +[2026-06-17 23:05:12,268][starVLA.training.train_starvla][INFO] - Step 4098, Loss: {'train/loss': 0.09621655941009521, 'train/grad_norm_pre_clip': 15.38917350769043, 'timing/data': 0.019764727912843227, 'timing/model': 1.1704671252518892, 'learning_rate/qwen_vl_interface': 4.077103973146054e-06, 'learning_rate/action_model': 4.077103973146054e-05, 'epoch': 1.86}) +[2026-06-17 23:05:15,992][starVLA.training.train_starvla][INFO] - Step 4099, Loss: {'train/loss': 0.20949168503284454, 'train/grad_norm_pre_clip': 15.272490501403809, 'timing/data': 0.015996171161532402, 'timing/model': 0.9804695015773177, 'learning_rate/qwen_vl_interface': 4.075008373051375e-06, 'learning_rate/action_model': 4.075008373051374e-05, 'epoch': 1.86}) +[2026-06-17 23:07:09,694][starVLA.training.train_starvla][INFO] - Step 4100, Loss: {'train/loss': 0.2217094898223877, 'train/grad_norm_pre_clip': 23.16082000732422, 'eval/loss': 0.29949286580085754, 'eval/latency_8/loss': 0.29949286580085754, 'eval/demon_fire_precision': 0.9591280653950953, 'eval/demon_fire_recall': 0.9841565703634669, 'eval/demon_fire_f1': 0.9714811407543699, 'eval/demon_fire_rate_model': 0.9190317195325542, 'eval/demon_fire_rate_teacher': 0.8956594323873122, 'eval/demon_left_precision': 0.8454332552693209, 'eval/demon_left_recall': 0.7730192719486081, 'eval/demon_left_f1': 0.807606263982103, 'eval/demon_left_rate_model': 0.1188091263216472, 'eval/demon_left_rate_teacher': 0.12993878686700056, 'eval/demon_right_precision': 0.8573770491803279, 'eval/demon_right_recall': 0.7924242424242425, 'eval/demon_right_f1': 0.8236220472440946, 'eval/demon_right_rate_model': 0.16972732331663884, 'eval/demon_right_rate_teacher': 0.18363939899833054, 'eval/demon_move_f1': 0.8156141556130988, 'eval/demon_cc_f1': 0.9188049959153632, 'eval/latency_8/demon_fire_precision': 0.9591280653950953, 'eval/latency_8/demon_fire_recall': 0.9841565703634669, 'eval/latency_8/demon_fire_f1': 0.9714811407543699, 'eval/latency_8/demon_fire_rate_model': 0.9190317195325542, 'eval/latency_8/demon_fire_rate_teacher': 0.8956594323873122, 'eval/latency_8/demon_left_precision': 0.8454332552693209, 'eval/latency_8/demon_left_recall': 0.7730192719486081, 'eval/latency_8/demon_left_f1': 0.807606263982103, 'eval/latency_8/demon_left_rate_model': 0.1188091263216472, 'eval/latency_8/demon_left_rate_teacher': 0.12993878686700056, 'eval/latency_8/demon_right_precision': 0.8573770491803279, 'eval/latency_8/demon_right_recall': 0.7924242424242425, 'eval/latency_8/demon_right_f1': 0.8236220472440946, 'eval/latency_8/demon_right_rate_model': 0.16972732331663884, 'eval/latency_8/demon_right_rate_teacher': 0.18363939899833054, 'eval/latency_8/demon_move_f1': 0.8156141556130988, 'eval/latency_8/demon_cc_f1': 0.9188049959153632, 'timing/data': 0.020586755126714706, 'timing/model': 0.9420943967998028, 'learning_rate/qwen_vl_interface': 4.072913016533726e-06, 'learning_rate/action_model': 4.072913016533726e-05, 'epoch': 1.86}) +[2026-06-17 23:07:09,697][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_4100_state +[2026-06-17 23:07:09,702][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 23:07:16,186][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_4100_state/model.safetensors +[2026-06-17 23:07:31,986][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_4100_state/optimizer.bin +[2026-06-17 23:07:31,986][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_4100_state/sampler.bin +[2026-06-17 23:07:31,986][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_4100_state/sampler_1.bin +[2026-06-17 23:07:31,988][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_4100_state/random_states_0.pkl +[2026-06-17 23:07:31,993][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4100 +[2026-06-17 23:07:31,993][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 23:07:32,003][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 23:07:35,609][starVLA.training.train_starvla][INFO] - Step 4101, Loss: {'train/loss': 0.3310648500919342, 'train/grad_norm_pre_clip': 21.48297119140625, 'timing/data': 0.01908854115754366, 'timing/model': 0.9289856143295765, 'learning_rate/qwen_vl_interface': 4.070817904027474e-06, 'learning_rate/action_model': 4.070817904027474e-05, 'epoch': 1.86}) +[2026-06-17 23:07:39,446][starVLA.training.train_starvla][INFO] - Step 4102, Loss: {'train/loss': 0.2633315622806549, 'train/grad_norm_pre_clip': 27.653059005737305, 'timing/data': 0.018592659384012222, 'timing/model': 0.9343053679913282, 'learning_rate/qwen_vl_interface': 4.06872303596694e-06, 'learning_rate/action_model': 4.06872303596694e-05, 'epoch': 1.86}) +[2026-06-17 23:07:43,199][starVLA.training.train_starvla][INFO] - Step 4103, Loss: {'train/loss': 0.4104219675064087, 'train/grad_norm_pre_clip': 25.746994018554688, 'timing/data': 0.020117186941206455, 'timing/model': 0.9251241171732545, 'learning_rate/qwen_vl_interface': 4.066628412786391e-06, 'learning_rate/action_model': 4.066628412786391e-05, 'epoch': 1.86}) +[2026-06-17 23:07:47,182][starVLA.training.train_starvla][INFO] - Step 4104, Loss: {'train/loss': 0.3649747371673584, 'train/grad_norm_pre_clip': 29.146059036254883, 'timing/data': 0.01643195003271103, 'timing/model': 0.9273193823173642, 'learning_rate/qwen_vl_interface': 4.0645340349200466e-06, 'learning_rate/action_model': 4.0645340349200464e-05, 'epoch': 1.86}) +[2026-06-17 23:07:50,924][starVLA.training.train_starvla][INFO] - Step 4105, Loss: {'train/loss': 0.11993161588907242, 'train/grad_norm_pre_clip': 11.442355155944824, 'timing/data': 0.022375669330358505, 'timing/model': 0.9262450113892555, 'learning_rate/qwen_vl_interface': 4.062439902802073e-06, 'learning_rate/action_model': 4.0624399028020725e-05, 'epoch': 1.86}) +[2026-06-17 23:07:54,689][starVLA.training.train_starvla][INFO] - Step 4106, Loss: {'train/loss': 0.19840694963932037, 'train/grad_norm_pre_clip': 16.179706573486328, 'timing/data': 0.02377466671168804, 'timing/model': 1.1000458505004644, 'learning_rate/qwen_vl_interface': 4.060346016866584e-06, 'learning_rate/action_model': 4.060346016866584e-05, 'epoch': 1.86}) +[2026-06-17 23:07:58,589][starVLA.training.train_starvla][INFO] - Step 4107, Loss: {'train/loss': 0.2047579139471054, 'train/grad_norm_pre_clip': 19.143253326416016, 'timing/data': 0.0224771061912179, 'timing/model': 1.0598022108897567, 'learning_rate/qwen_vl_interface': 4.058252377547645e-06, 'learning_rate/action_model': 4.058252377547645e-05, 'epoch': 1.86}) +[2026-06-17 23:08:02,237][starVLA.training.train_starvla][INFO] - Step 4108, Loss: {'train/loss': 0.5452102422714233, 'train/grad_norm_pre_clip': 54.509674072265625, 'timing/data': 0.019257755018770695, 'timing/model': 0.9337419467046857, 'learning_rate/qwen_vl_interface': 4.056158985279272e-06, 'learning_rate/action_model': 4.0561589852792715e-05, 'epoch': 1.86}) +[2026-06-17 23:08:06,054][starVLA.training.train_starvla][INFO] - Step 4109, Loss: {'train/loss': 0.6684363484382629, 'train/grad_norm_pre_clip': 40.55443572998047, 'timing/data': 0.016595272347331047, 'timing/model': 0.9619475081562996, 'learning_rate/qwen_vl_interface': 4.054065840495425e-06, 'learning_rate/action_model': 4.054065840495425e-05, 'epoch': 1.86}) +[2026-06-17 23:08:10,139][starVLA.training.train_starvla][INFO] - Step 4110, Loss: {'train/loss': 0.09689515084028244, 'train/grad_norm_pre_clip': 12.936483383178711, 'timing/data': 0.01941419579088688, 'timing/model': 0.9389518974348903, 'learning_rate/qwen_vl_interface': 4.0519729436300165e-06, 'learning_rate/action_model': 4.051972943630016e-05, 'epoch': 1.86}) +[2026-06-17 23:08:14,097][starVLA.training.train_starvla][INFO] - Step 4111, Loss: {'train/loss': 0.5451772212982178, 'train/grad_norm_pre_clip': 23.78729248046875, 'timing/data': 0.01899871416389942, 'timing/model': 1.0134197594597936, 'learning_rate/qwen_vl_interface': 4.049880295116906e-06, 'learning_rate/action_model': 4.049880295116906e-05, 'epoch': 1.87}) +[2026-06-17 23:08:18,705][starVLA.training.train_starvla][INFO] - Step 4112, Loss: {'train/loss': 0.29885128140449524, 'train/grad_norm_pre_clip': 22.319459915161133, 'timing/data': 0.023270539939403534, 'timing/model': 1.4711242048069835, 'learning_rate/qwen_vl_interface': 4.047787895389902e-06, 'learning_rate/action_model': 4.0477878953899016e-05, 'epoch': 1.87}) +[2026-06-17 23:08:22,600][starVLA.training.train_starvla][INFO] - Step 4113, Loss: {'train/loss': 0.24757082760334015, 'train/grad_norm_pre_clip': 21.79022216796875, 'timing/data': 0.02147130947560072, 'timing/model': 1.1308401776477695, 'learning_rate/qwen_vl_interface': 4.0456957448827615e-06, 'learning_rate/action_model': 4.045695744882761e-05, 'epoch': 1.87}) +[2026-06-17 23:08:26,248][starVLA.training.train_starvla][INFO] - Step 4114, Loss: {'train/loss': 0.3335014879703522, 'train/grad_norm_pre_clip': 24.89368438720703, 'timing/data': 0.018272943794727325, 'timing/model': 0.9598954617977142, 'learning_rate/qwen_vl_interface': 4.043603844029188e-06, 'learning_rate/action_model': 4.0436038440291874e-05, 'epoch': 1.87}) +[2026-06-17 23:08:30,117][starVLA.training.train_starvla][INFO] - Step 4115, Loss: {'train/loss': 0.26327112317085266, 'train/grad_norm_pre_clip': 25.408193588256836, 'timing/data': 0.02087877131998539, 'timing/model': 0.9845186332240701, 'learning_rate/qwen_vl_interface': 4.041512193262836e-06, 'learning_rate/action_model': 4.041512193262836e-05, 'epoch': 1.87}) +[2026-06-17 23:08:34,032][starVLA.training.train_starvla][INFO] - Step 4116, Loss: {'train/loss': 0.11131247878074646, 'train/grad_norm_pre_clip': 12.200394630432129, 'timing/data': 0.02173604443669319, 'timing/model': 1.0438291616737843, 'learning_rate/qwen_vl_interface': 4.039420793017306e-06, 'learning_rate/action_model': 4.039420793017306e-05, 'epoch': 1.87}) +[2026-06-17 23:08:37,805][starVLA.training.train_starvla][INFO] - Step 4117, Loss: {'train/loss': 0.2622845470905304, 'train/grad_norm_pre_clip': 24.038719177246094, 'timing/data': 0.022841338999569416, 'timing/model': 0.9325086493045092, 'learning_rate/qwen_vl_interface': 4.03732964372615e-06, 'learning_rate/action_model': 4.037329643726149e-05, 'epoch': 1.87}) +[2026-06-17 23:08:41,609][starVLA.training.train_starvla][INFO] - Step 4118, Loss: {'train/loss': 0.09052503108978271, 'train/grad_norm_pre_clip': 10.78715991973877, 'timing/data': 0.022120407782495022, 'timing/model': 0.9251410514116287, 'learning_rate/qwen_vl_interface': 4.035238745822861e-06, 'learning_rate/action_model': 4.035238745822861e-05, 'epoch': 1.87}) +[2026-06-17 23:08:45,397][starVLA.training.train_starvla][INFO] - Step 4119, Loss: {'train/loss': 0.2665785253047943, 'train/grad_norm_pre_clip': 23.262451171875, 'timing/data': 0.020282614044845104, 'timing/model': 1.1267677024006844, 'learning_rate/qwen_vl_interface': 4.033148099740889e-06, 'learning_rate/action_model': 4.0331480997408885e-05, 'epoch': 1.87}) +[2026-06-17 23:08:49,252][starVLA.training.train_starvla][INFO] - Step 4120, Loss: {'train/loss': 0.26296308636665344, 'train/grad_norm_pre_clip': 23.291515350341797, 'timing/data': 0.015335028991103172, 'timing/model': 1.0994518268853426, 'learning_rate/qwen_vl_interface': 4.031057705913626e-06, 'learning_rate/action_model': 4.031057705913626e-05, 'epoch': 1.87}) +[2026-06-17 23:08:52,877][starVLA.training.train_starvla][INFO] - Step 4121, Loss: {'train/loss': 0.18470393121242523, 'train/grad_norm_pre_clip': 21.43006706237793, 'timing/data': 0.016931315883994102, 'timing/model': 0.940285200253129, 'learning_rate/qwen_vl_interface': 4.028967564774411e-06, 'learning_rate/action_model': 4.0289675647744106e-05, 'epoch': 1.87}) +[2026-06-17 23:08:56,759][starVLA.training.train_starvla][INFO] - Step 4122, Loss: {'train/loss': 0.26505976915359497, 'train/grad_norm_pre_clip': 19.78169822692871, 'timing/data': 0.015989921055734158, 'timing/model': 0.9265351993963122, 'learning_rate/qwen_vl_interface': 4.026877676756534e-06, 'learning_rate/action_model': 4.026877676756534e-05, 'epoch': 1.87}) +[2026-06-17 23:09:00,514][starVLA.training.train_starvla][INFO] - Step 4123, Loss: {'train/loss': 0.40850692987442017, 'train/grad_norm_pre_clip': 26.60913848876953, 'timing/data': 0.01720060221850872, 'timing/model': 0.922545094974339, 'learning_rate/qwen_vl_interface': 4.0247880422932305e-06, 'learning_rate/action_model': 4.02478804229323e-05, 'epoch': 1.87}) +[2026-06-17 23:09:04,352][starVLA.training.train_starvla][INFO] - Step 4124, Loss: {'train/loss': 0.19727477431297302, 'train/grad_norm_pre_clip': 25.10338592529297, 'timing/data': 0.016801530495285988, 'timing/model': 0.993320663459599, 'learning_rate/qwen_vl_interface': 4.0226986618176855e-06, 'learning_rate/action_model': 4.022698661817686e-05, 'epoch': 1.87}) +[2026-06-17 23:09:08,413][starVLA.training.train_starvla][INFO] - Step 4125, Loss: {'train/loss': 0.2966115474700928, 'train/grad_norm_pre_clip': 21.973005294799805, 'timing/data': 0.017096216790378094, 'timing/model': 0.9868685808032751, 'learning_rate/qwen_vl_interface': 4.0206095357630256e-06, 'learning_rate/action_model': 4.020609535763026e-05, 'epoch': 1.87}) +[2026-06-17 23:09:12,270][starVLA.training.train_starvla][INFO] - Step 4126, Loss: {'train/loss': 0.34154289960861206, 'train/grad_norm_pre_clip': 22.268835067749023, 'timing/data': 0.019395802170038223, 'timing/model': 1.1244710879400373, 'learning_rate/qwen_vl_interface': 4.018520664562334e-06, 'learning_rate/action_model': 4.018520664562334e-05, 'epoch': 1.87}) +[2026-06-17 23:09:16,189][starVLA.training.train_starvla][INFO] - Step 4127, Loss: {'train/loss': 0.3287392258644104, 'train/grad_norm_pre_clip': 20.72658920288086, 'timing/data': 0.021797984838485718, 'timing/model': 1.1238422859460115, 'learning_rate/qwen_vl_interface': 4.016432048648635e-06, 'learning_rate/action_model': 4.016432048648635e-05, 'epoch': 1.87}) +[2026-06-17 23:09:19,971][starVLA.training.train_starvla][INFO] - Step 4128, Loss: {'train/loss': 0.3664194345474243, 'train/grad_norm_pre_clip': 19.671497344970703, 'timing/data': 0.02227355819195509, 'timing/model': 0.9342821901664138, 'learning_rate/qwen_vl_interface': 4.014343688454898e-06, 'learning_rate/action_model': 4.0143436884548976e-05, 'epoch': 1.87}) +[2026-06-17 23:09:23,755][starVLA.training.train_starvla][INFO] - Step 4129, Loss: {'train/loss': 0.3527306914329529, 'train/grad_norm_pre_clip': 23.7784366607666, 'timing/data': 0.014941019006073475, 'timing/model': 0.9403978260233998, 'learning_rate/qwen_vl_interface': 4.012255584414044e-06, 'learning_rate/action_model': 4.0122555844140436e-05, 'epoch': 1.87}) +[2026-06-17 23:09:27,998][starVLA.training.train_starvla][INFO] - Step 4130, Loss: {'train/loss': 0.3333264887332916, 'train/grad_norm_pre_clip': 21.03680992126465, 'timing/data': 0.019097950309515, 'timing/model': 1.315598277375102, 'learning_rate/qwen_vl_interface': 4.010167736958939e-06, 'learning_rate/action_model': 4.010167736958938e-05, 'epoch': 1.87}) +[2026-06-17 23:09:31,857][starVLA.training.train_starvla][INFO] - Step 4131, Loss: {'train/loss': 0.16333550214767456, 'train/grad_norm_pre_clip': 33.01214599609375, 'timing/data': 0.02832952979952097, 'timing/model': 0.9505460513755679, 'learning_rate/qwen_vl_interface': 4.008080146522397e-06, 'learning_rate/action_model': 4.008080146522396e-05, 'epoch': 1.87}) +[2026-06-17 23:09:35,645][starVLA.training.train_starvla][INFO] - Step 4132, Loss: {'train/loss': 0.0638611912727356, 'train/grad_norm_pre_clip': 8.451623916625977, 'timing/data': 0.017214606516063213, 'timing/model': 0.936511468142271, 'learning_rate/qwen_vl_interface': 4.005992813537176e-06, 'learning_rate/action_model': 4.005992813537176e-05, 'epoch': 1.87}) +[2026-06-17 23:09:39,435][starVLA.training.train_starvla][INFO] - Step 4133, Loss: {'train/loss': 0.39420050382614136, 'train/grad_norm_pre_clip': 18.650854110717773, 'timing/data': 0.013780487701296806, 'timing/model': 1.096730233170092, 'learning_rate/qwen_vl_interface': 4.003905738435984e-06, 'learning_rate/action_model': 4.003905738435984e-05, 'epoch': 1.88}) +[2026-06-17 23:09:43,030][starVLA.training.train_starvla][INFO] - Step 4134, Loss: {'train/loss': 0.4852979779243469, 'train/grad_norm_pre_clip': 27.14523696899414, 'timing/data': 0.016932101920247078, 'timing/model': 0.9395702220499516, 'learning_rate/qwen_vl_interface': 4.001818921651475e-06, 'learning_rate/action_model': 4.0018189216514746e-05, 'epoch': 1.88}) +[2026-06-17 23:09:46,823][starVLA.training.train_starvla][INFO] - Step 4135, Loss: {'train/loss': 0.17926639318466187, 'train/grad_norm_pre_clip': 15.077069282531738, 'timing/data': 0.020430748350918293, 'timing/model': 0.9290367420762777, 'learning_rate/qwen_vl_interface': 3.999732363616244e-06, 'learning_rate/action_model': 3.999732363616244e-05, 'epoch': 1.88}) +[2026-06-17 23:09:50,632][starVLA.training.train_starvla][INFO] - Step 4136, Loss: {'train/loss': 0.30285218358039856, 'train/grad_norm_pre_clip': 16.742984771728516, 'timing/data': 0.02341552823781967, 'timing/model': 0.9269423540681601, 'learning_rate/qwen_vl_interface': 3.9976460647628394e-06, 'learning_rate/action_model': 3.997646064762839e-05, 'epoch': 1.88}) +[2026-06-17 23:09:54,473][starVLA.training.train_starvla][INFO] - Step 4137, Loss: {'train/loss': 0.3904682993888855, 'train/grad_norm_pre_clip': 25.525789260864258, 'timing/data': 0.018342910334467888, 'timing/model': 0.9612945578992367, 'learning_rate/qwen_vl_interface': 3.995560025523755e-06, 'learning_rate/action_model': 3.995560025523754e-05, 'epoch': 1.88}) +[2026-06-17 23:09:58,323][starVLA.training.train_starvla][INFO] - Step 4138, Loss: {'train/loss': 0.20274020731449127, 'train/grad_norm_pre_clip': 17.853485107421875, 'timing/data': 0.01906877849251032, 'timing/model': 0.94038016628474, 'learning_rate/qwen_vl_interface': 3.993474246331426e-06, 'learning_rate/action_model': 3.993474246331426e-05, 'epoch': 1.88}) +[2026-06-17 23:10:02,184][starVLA.training.train_starvla][INFO] - Step 4139, Loss: {'train/loss': 0.15892621874809265, 'train/grad_norm_pre_clip': 18.55354881286621, 'timing/data': 0.020390414632856846, 'timing/model': 1.159197692759335, 'learning_rate/qwen_vl_interface': 3.991388727618237e-06, 'learning_rate/action_model': 3.991388727618237e-05, 'epoch': 1.88}) +[2026-06-17 23:10:06,127][starVLA.training.train_starvla][INFO] - Step 4140, Loss: {'train/loss': 0.33320656418800354, 'train/grad_norm_pre_clip': 25.27643394470215, 'timing/data': 0.01838257536292076, 'timing/model': 1.1518428949639201, 'learning_rate/qwen_vl_interface': 3.989303469816521e-06, 'learning_rate/action_model': 3.9893034698165205e-05, 'epoch': 1.88}) +[2026-06-17 23:10:09,812][starVLA.training.train_starvla][INFO] - Step 4141, Loss: {'train/loss': 0.24727967381477356, 'train/grad_norm_pre_clip': 16.011709213256836, 'timing/data': 0.017791638150811195, 'timing/model': 0.9463168196380138, 'learning_rate/qwen_vl_interface': 3.9872184733585526e-06, 'learning_rate/action_model': 3.987218473358552e-05, 'epoch': 1.88}) +[2026-06-17 23:10:13,675][starVLA.training.train_starvla][INFO] - Step 4142, Loss: {'train/loss': 0.31779536604881287, 'train/grad_norm_pre_clip': 19.01966667175293, 'timing/data': 0.016770072281360626, 'timing/model': 0.9473070343956351, 'learning_rate/qwen_vl_interface': 3.985133738676552e-06, 'learning_rate/action_model': 3.985133738676552e-05, 'epoch': 1.88}) +[2026-06-17 23:10:17,693][starVLA.training.train_starvla][INFO] - Step 4143, Loss: {'train/loss': 0.5448101758956909, 'train/grad_norm_pre_clip': 34.441505432128906, 'timing/data': 0.02105572074651718, 'timing/model': 0.9306393507868052, 'learning_rate/qwen_vl_interface': 3.983049266202688e-06, 'learning_rate/action_model': 3.983049266202687e-05, 'epoch': 1.88}) +[2026-06-17 23:10:21,606][starVLA.training.train_starvla][INFO] - Step 4144, Loss: {'train/loss': 0.2103586047887802, 'train/grad_norm_pre_clip': 15.295215606689453, 'timing/data': 0.019520843401551247, 'timing/model': 0.9353758990764618, 'learning_rate/qwen_vl_interface': 3.980965056369075e-06, 'learning_rate/action_model': 3.980965056369075e-05, 'epoch': 1.88}) +[2026-06-17 23:10:25,412][starVLA.training.train_starvla][INFO] - Step 4145, Loss: {'train/loss': 0.05402234569191933, 'train/grad_norm_pre_clip': 7.054142475128174, 'timing/data': 0.019916852936148643, 'timing/model': 0.9303568750619888, 'learning_rate/qwen_vl_interface': 3.978881109607772e-06, 'learning_rate/action_model': 3.978881109607772e-05, 'epoch': 1.88}) +[2026-06-17 23:10:29,208][starVLA.training.train_starvla][INFO] - Step 4146, Loss: {'train/loss': 0.25608375668525696, 'train/grad_norm_pre_clip': 23.032333374023438, 'timing/data': 0.016888907179236412, 'timing/model': 1.1177718602120876, 'learning_rate/qwen_vl_interface': 3.976797426350783e-06, 'learning_rate/action_model': 3.976797426350783e-05, 'epoch': 1.88}) +[2026-06-17 23:10:33,192][starVLA.training.train_starvla][INFO] - Step 4147, Loss: {'train/loss': 0.4554109275341034, 'train/grad_norm_pre_clip': 35.444766998291016, 'timing/data': 0.015474976040422916, 'timing/model': 1.0613611629232764, 'learning_rate/qwen_vl_interface': 3.974714007030055e-06, 'learning_rate/action_model': 3.974714007030056e-05, 'epoch': 1.88}) +[2026-06-17 23:10:36,798][starVLA.training.train_starvla][INFO] - Step 4148, Loss: {'train/loss': 0.2964860200881958, 'train/grad_norm_pre_clip': 29.27957534790039, 'timing/data': 0.014522338286042213, 'timing/model': 0.9475804837420583, 'learning_rate/qwen_vl_interface': 3.972630852077488e-06, 'learning_rate/action_model': 3.972630852077488e-05, 'epoch': 1.88}) +[2026-06-17 23:10:40,605][starVLA.training.train_starvla][INFO] - Step 4149, Loss: {'train/loss': 0.36434054374694824, 'train/grad_norm_pre_clip': 24.40552520751953, 'timing/data': 0.01812727376818657, 'timing/model': 0.9356275266036391, 'learning_rate/qwen_vl_interface': 3.9705479619249195e-06, 'learning_rate/action_model': 3.9705479619249197e-05, 'epoch': 1.88}) +[2026-06-17 23:10:44,358][starVLA.training.train_starvla][INFO] - Step 4150, Loss: {'train/loss': 0.2566080689430237, 'train/grad_norm_pre_clip': 16.21588706970215, 'timing/data': 0.015230025164783001, 'timing/model': 0.9358334317803383, 'learning_rate/qwen_vl_interface': 3.9684653370041345e-06, 'learning_rate/action_model': 3.968465337004134e-05, 'epoch': 1.88}) +[2026-06-17 23:10:48,419][starVLA.training.train_starvla][INFO] - Step 4151, Loss: {'train/loss': 0.32662972807884216, 'train/grad_norm_pre_clip': 23.2842960357666, 'timing/data': 0.015172665938735008, 'timing/model': 1.1608718689531088, 'learning_rate/qwen_vl_interface': 3.966382977746866e-06, 'learning_rate/action_model': 3.9663829777468656e-05, 'epoch': 1.88}) +[2026-06-17 23:10:52,159][starVLA.training.train_starvla][INFO] - Step 4152, Loss: {'train/loss': 0.2886699438095093, 'train/grad_norm_pre_clip': 16.487573623657227, 'timing/data': 0.023685737513005733, 'timing/model': 0.9282520487904549, 'learning_rate/qwen_vl_interface': 3.9643008845847845e-06, 'learning_rate/action_model': 3.964300884584785e-05, 'epoch': 1.88}) +[2026-06-17 23:10:56,189][starVLA.training.train_starvla][INFO] - Step 4153, Loss: {'train/loss': 0.319796621799469, 'train/grad_norm_pre_clip': 27.725337982177734, 'timing/data': 0.019021845422685146, 'timing/model': 1.2180304415524006, 'learning_rate/qwen_vl_interface': 3.962219057949517e-06, 'learning_rate/action_model': 3.9622190579495166e-05, 'epoch': 1.88}) +[2026-06-17 23:10:59,972][starVLA.training.train_starvla][INFO] - Step 4154, Loss: {'train/loss': 0.29348692297935486, 'train/grad_norm_pre_clip': 18.655933380126953, 'timing/data': 0.017519216984510422, 'timing/model': 0.9268814930692315, 'learning_rate/qwen_vl_interface': 3.960137498272619e-06, 'learning_rate/action_model': 3.9601374982726194e-05, 'epoch': 1.88}) +[2026-06-17 23:11:03,746][starVLA.training.train_starvla][INFO] - Step 4155, Loss: {'train/loss': 0.0711815282702446, 'train/grad_norm_pre_clip': 9.61336898803711, 'timing/data': 0.014453394338488579, 'timing/model': 0.9284681268036366, 'learning_rate/qwen_vl_interface': 3.958056205985611e-06, 'learning_rate/action_model': 3.95805620598561e-05, 'epoch': 1.89}) +[2026-06-17 23:11:07,482][starVLA.training.train_starvla][INFO] - Step 4156, Loss: {'train/loss': 0.2702402174472809, 'train/grad_norm_pre_clip': 35.12362289428711, 'timing/data': 0.015815703198313713, 'timing/model': 0.9405438788235188, 'learning_rate/qwen_vl_interface': 3.95597518151994e-06, 'learning_rate/action_model': 3.955975181519939e-05, 'epoch': 1.89}) +[2026-06-17 23:11:11,449][starVLA.training.train_starvla][INFO] - Step 4157, Loss: {'train/loss': 0.16943271458148956, 'train/grad_norm_pre_clip': 17.358104705810547, 'timing/data': 0.019304754212498665, 'timing/model': 0.9792420091107488, 'learning_rate/qwen_vl_interface': 3.953894425307005e-06, 'learning_rate/action_model': 3.9538944253070055e-05, 'epoch': 1.89}) +[2026-06-17 23:11:15,404][starVLA.training.train_starvla][INFO] - Step 4158, Loss: {'train/loss': 0.3732137680053711, 'train/grad_norm_pre_clip': 22.271190643310547, 'timing/data': 0.01601427234709263, 'timing/model': 1.004823287948966, 'learning_rate/qwen_vl_interface': 3.951813937778151e-06, 'learning_rate/action_model': 3.95181393777815e-05, 'epoch': 1.89}) +[2026-06-17 23:11:19,380][starVLA.training.train_starvla][INFO] - Step 4159, Loss: {'train/loss': 0.06102992966771126, 'train/grad_norm_pre_clip': 27.88760757446289, 'timing/data': 0.01625768654048443, 'timing/model': 1.1616899259388447, 'learning_rate/qwen_vl_interface': 3.949733719364664e-06, 'learning_rate/action_model': 3.949733719364664e-05, 'epoch': 1.89}) +[2026-06-17 23:11:24,799][starVLA.training.train_starvla][INFO] - Step 4160, Loss: {'train/loss': 0.45565104484558105, 'train/grad_norm_pre_clip': 22.471168518066406, 'timing/data': 0.01844778284430504, 'timing/model': 1.17366695497185, 'learning_rate/qwen_vl_interface': 3.947653770497778e-06, 'learning_rate/action_model': 3.947653770497778e-05, 'epoch': 1.89}) +[2026-06-17 23:11:29,698][starVLA.training.train_starvla][INFO] - Step 4161, Loss: {'train/loss': 0.049915753304958344, 'train/grad_norm_pre_clip': 7.653646469116211, 'timing/data': 0.017637290991842747, 'timing/model': 1.0186237804591656, 'learning_rate/qwen_vl_interface': 3.945574091608664e-06, 'learning_rate/action_model': 3.9455740916086636e-05, 'epoch': 1.89}) +[2026-06-17 23:11:34,866][starVLA.training.train_starvla][INFO] - Step 4162, Loss: {'train/loss': 0.08996632695198059, 'train/grad_norm_pre_clip': 13.022175788879395, 'timing/data': 0.014557612128555775, 'timing/model': 0.9363317862153053, 'learning_rate/qwen_vl_interface': 3.9434946831284465e-06, 'learning_rate/action_model': 3.943494683128447e-05, 'epoch': 1.89}) +[2026-06-17 23:11:38,820][starVLA.training.train_starvla][INFO] - Step 4163, Loss: {'train/loss': 0.39822274446487427, 'train/grad_norm_pre_clip': 23.440624237060547, 'timing/data': 0.017536578699946404, 'timing/model': 0.9446795517578721, 'learning_rate/qwen_vl_interface': 3.941415545488187e-06, 'learning_rate/action_model': 3.941415545488187e-05, 'epoch': 1.89}) +[2026-06-17 23:11:43,220][starVLA.training.train_starvla][INFO] - Step 4164, Loss: {'train/loss': 0.3883327543735504, 'train/grad_norm_pre_clip': 25.2023868560791, 'timing/data': 0.017324996180832386, 'timing/model': 1.5154499430209398, 'learning_rate/qwen_vl_interface': 3.939336679118892e-06, 'learning_rate/action_model': 3.939336679118892e-05, 'epoch': 1.89}) +[2026-06-17 23:11:47,152][starVLA.training.train_starvla][INFO] - Step 4165, Loss: {'train/loss': 0.1901279240846634, 'train/grad_norm_pre_clip': 20.751873016357422, 'timing/data': 0.017434573732316494, 'timing/model': 0.9459838839247823, 'learning_rate/qwen_vl_interface': 3.937258084451514e-06, 'learning_rate/action_model': 3.937258084451514e-05, 'epoch': 1.89}) +[2026-06-17 23:11:51,034][starVLA.training.train_starvla][INFO] - Step 4166, Loss: {'train/loss': 0.2344566285610199, 'train/grad_norm_pre_clip': 23.28903579711914, 'timing/data': 0.017837208695709705, 'timing/model': 1.1372894085943699, 'learning_rate/qwen_vl_interface': 3.9351797619169485e-06, 'learning_rate/action_model': 3.935179761916949e-05, 'epoch': 1.89}) +[2026-06-17 23:11:54,815][starVLA.training.train_starvla][INFO] - Step 4167, Loss: {'train/loss': 0.38321471214294434, 'train/grad_norm_pre_clip': 25.772754669189453, 'timing/data': 0.024692779406905174, 'timing/model': 1.0982150258496404, 'learning_rate/qwen_vl_interface': 3.933101711946034e-06, 'learning_rate/action_model': 3.9331017119460336e-05, 'epoch': 1.89}) +[2026-06-17 23:11:58,476][starVLA.training.train_starvla][INFO] - Step 4168, Loss: {'train/loss': 0.07950406521558762, 'train/grad_norm_pre_clip': 9.54988956451416, 'timing/data': 0.025680577382445335, 'timing/model': 0.9303100276738405, 'learning_rate/qwen_vl_interface': 3.93102393496955e-06, 'learning_rate/action_model': 3.93102393496955e-05, 'epoch': 1.89}) +[2026-06-17 23:12:02,338][starVLA.training.train_starvla][INFO] - Step 4169, Loss: {'train/loss': 0.3542449176311493, 'train/grad_norm_pre_clip': 23.270517349243164, 'timing/data': 0.017962872982025146, 'timing/model': 0.944941065274179, 'learning_rate/qwen_vl_interface': 3.928946431418223e-06, 'learning_rate/action_model': 3.9289464314182225e-05, 'epoch': 1.89}) +[2026-06-17 23:12:06,141][starVLA.training.train_starvla][INFO] - Step 4170, Loss: {'train/loss': 0.17866231501102448, 'train/grad_norm_pre_clip': 19.95380973815918, 'timing/data': 0.01741250418126583, 'timing/model': 0.9470688793808222, 'learning_rate/qwen_vl_interface': 3.926869201722723e-06, 'learning_rate/action_model': 3.926869201722723e-05, 'epoch': 1.89}) +[2026-06-17 23:12:10,102][starVLA.training.train_starvla][INFO] - Step 4171, Loss: {'train/loss': 0.1550183892250061, 'train/grad_norm_pre_clip': 18.64798927307129, 'timing/data': 0.01501847431063652, 'timing/model': 0.939225316978991, 'learning_rate/qwen_vl_interface': 3.9247922463136615e-06, 'learning_rate/action_model': 3.924792246313662e-05, 'epoch': 1.89}) +[2026-06-17 23:12:13,945][starVLA.training.train_starvla][INFO] - Step 4172, Loss: {'train/loss': 0.478049635887146, 'train/grad_norm_pre_clip': 23.391027450561523, 'timing/data': 0.016887448728084564, 'timing/model': 0.9407579740509391, 'learning_rate/qwen_vl_interface': 3.922715565621593e-06, 'learning_rate/action_model': 3.9227155656215926e-05, 'epoch': 1.89}) +[2026-06-17 23:12:17,766][starVLA.training.train_starvla][INFO] - Step 4173, Loss: {'train/loss': 0.24701884388923645, 'train/grad_norm_pre_clip': 20.22821617126465, 'timing/data': 0.02002850826829672, 'timing/model': 1.1238692263141274, 'learning_rate/qwen_vl_interface': 3.9206391600770164e-06, 'learning_rate/action_model': 3.920639160077016e-05, 'epoch': 1.89}) +[2026-06-17 23:12:21,853][starVLA.training.train_starvla][INFO] - Step 4174, Loss: {'train/loss': 0.3754720687866211, 'train/grad_norm_pre_clip': 33.292903900146484, 'timing/data': 0.025679592974483967, 'timing/model': 1.2738828416913748, 'learning_rate/qwen_vl_interface': 3.918563030110373e-06, 'learning_rate/action_model': 3.918563030110373e-05, 'epoch': 1.89}) +[2026-06-17 23:12:25,708][starVLA.training.train_starvla][INFO] - Step 4175, Loss: {'train/loss': 0.6483370661735535, 'train/grad_norm_pre_clip': 35.02394104003906, 'timing/data': 0.015983711928129196, 'timing/model': 0.9458714425563812, 'learning_rate/qwen_vl_interface': 3.916487176152045e-06, 'learning_rate/action_model': 3.916487176152045e-05, 'epoch': 1.89}) +[2026-06-17 23:12:29,485][starVLA.training.train_starvla][INFO] - Step 4176, Loss: {'train/loss': 0.24190370738506317, 'train/grad_norm_pre_clip': 19.63973617553711, 'timing/data': 0.026245808228850365, 'timing/model': 0.940675999969244, 'learning_rate/qwen_vl_interface': 3.914411598632359e-06, 'learning_rate/action_model': 3.914411598632359e-05, 'epoch': 1.89}) +[2026-06-17 23:12:33,321][starVLA.training.train_starvla][INFO] - Step 4177, Loss: {'train/loss': 0.20385880768299103, 'train/grad_norm_pre_clip': 18.256715774536133, 'timing/data': 0.016422527842223644, 'timing/model': 0.9604575326666236, 'learning_rate/qwen_vl_interface': 3.912336297981588e-06, 'learning_rate/action_model': 3.912336297981588e-05, 'epoch': 1.89}) +[2026-06-17 23:12:37,401][starVLA.training.train_starvla][INFO] - Step 4178, Loss: {'train/loss': 0.806764543056488, 'train/grad_norm_pre_clip': 30.58823013305664, 'timing/data': 0.017558397725224495, 'timing/model': 0.9854161469265819, 'learning_rate/qwen_vl_interface': 3.910261274629941e-06, 'learning_rate/action_model': 3.910261274629941e-05, 'epoch': 1.9}) +[2026-06-17 23:12:41,290][starVLA.training.train_starvla][INFO] - Step 4179, Loss: {'train/loss': 0.17210637032985687, 'train/grad_norm_pre_clip': 23.371442794799805, 'timing/data': 0.016001670621335506, 'timing/model': 0.9821138177067041, 'learning_rate/qwen_vl_interface': 3.908186529007576e-06, 'learning_rate/action_model': 3.908186529007575e-05, 'epoch': 1.9}) +[2026-06-17 23:12:45,158][starVLA.training.train_starvla][INFO] - Step 4180, Loss: {'train/loss': 0.14518766105175018, 'train/grad_norm_pre_clip': 14.523202896118164, 'timing/data': 0.01737319864332676, 'timing/model': 1.1129080187529325, 'learning_rate/qwen_vl_interface': 3.906112061544582e-06, 'learning_rate/action_model': 3.906112061544582e-05, 'epoch': 1.9}) +[2026-06-17 23:12:48,891][starVLA.training.train_starvla][INFO] - Step 4181, Loss: {'train/loss': 0.1527969390153885, 'train/grad_norm_pre_clip': 20.49275016784668, 'timing/data': 0.013951234519481659, 'timing/model': 0.9478352498263121, 'learning_rate/qwen_vl_interface': 3.904037872671008e-06, 'learning_rate/action_model': 3.904037872671008e-05, 'epoch': 1.9}) +[2026-06-17 23:12:52,742][starVLA.training.train_starvla][INFO] - Step 4182, Loss: {'train/loss': 0.40364086627960205, 'train/grad_norm_pre_clip': 32.84576416015625, 'timing/data': 0.018604349344968796, 'timing/model': 0.9370216894894838, 'learning_rate/qwen_vl_interface': 3.90196396281683e-06, 'learning_rate/action_model': 3.9019639628168295e-05, 'epoch': 1.9}) +[2026-06-17 23:12:56,604][starVLA.training.train_starvla][INFO] - Step 4183, Loss: {'train/loss': 0.20570281147956848, 'train/grad_norm_pre_clip': 19.846099853515625, 'timing/data': 0.01591285690665245, 'timing/model': 0.944945527240634, 'learning_rate/qwen_vl_interface': 3.899890332411972e-06, 'learning_rate/action_model': 3.8998903324119724e-05, 'epoch': 1.9}) +[2026-06-17 23:13:00,566][starVLA.training.train_starvla][INFO] - Step 4184, Loss: {'train/loss': 0.3637090027332306, 'train/grad_norm_pre_clip': 24.21755599975586, 'timing/data': 0.01603239867836237, 'timing/model': 1.0003716880455613, 'learning_rate/qwen_vl_interface': 3.897816981886302e-06, 'learning_rate/action_model': 3.897816981886302e-05, 'epoch': 1.9}) +[2026-06-17 23:13:04,470][starVLA.training.train_starvla][INFO] - Step 4185, Loss: {'train/loss': 0.5136865377426147, 'train/grad_norm_pre_clip': 22.2799072265625, 'timing/data': 0.01657764334231615, 'timing/model': 0.9538349648937583, 'learning_rate/qwen_vl_interface': 3.895743911669627e-06, 'learning_rate/action_model': 3.895743911669627e-05, 'epoch': 1.9}) +[2026-06-17 23:13:08,391][starVLA.training.train_starvla][INFO] - Step 4186, Loss: {'train/loss': 0.4572584927082062, 'train/grad_norm_pre_clip': 29.26221466064453, 'timing/data': 0.02223942056298256, 'timing/model': 1.1536838561296463, 'learning_rate/qwen_vl_interface': 3.893671122191698e-06, 'learning_rate/action_model': 3.8936711221916974e-05, 'epoch': 1.9}) +[2026-06-17 23:13:12,306][starVLA.training.train_starvla][INFO] - Step 4187, Loss: {'train/loss': 0.1902884989976883, 'train/grad_norm_pre_clip': 18.03229522705078, 'timing/data': 0.020095820538699627, 'timing/model': 1.1559446146711707, 'learning_rate/qwen_vl_interface': 3.8915986138822e-06, 'learning_rate/action_model': 3.8915986138822e-05, 'epoch': 1.9}) +[2026-06-17 23:13:15,961][starVLA.training.train_starvla][INFO] - Step 4188, Loss: {'train/loss': 0.20297771692276, 'train/grad_norm_pre_clip': 16.938308715820312, 'timing/data': 0.01848526019603014, 'timing/model': 0.9625484235584736, 'learning_rate/qwen_vl_interface': 3.889526387170775e-06, 'learning_rate/action_model': 3.889526387170775e-05, 'epoch': 1.9}) +[2026-06-17 23:13:19,893][starVLA.training.train_starvla][INFO] - Step 4189, Loss: {'train/loss': 0.5133256912231445, 'train/grad_norm_pre_clip': 27.154878616333008, 'timing/data': 0.017291334457695484, 'timing/model': 0.9862336926162243, 'learning_rate/qwen_vl_interface': 3.8874544424869924e-06, 'learning_rate/action_model': 3.8874544424869926e-05, 'epoch': 1.9}) +[2026-06-17 23:13:23,780][starVLA.training.train_starvla][INFO] - Step 4190, Loss: {'train/loss': 0.2738238573074341, 'train/grad_norm_pre_clip': 24.241741180419922, 'timing/data': 0.015772482380270958, 'timing/model': 0.9334532357752323, 'learning_rate/qwen_vl_interface': 3.885382780260371e-06, 'learning_rate/action_model': 3.885382780260371e-05, 'epoch': 1.9}) +[2026-06-17 23:13:27,628][starVLA.training.train_starvla][INFO] - Step 4191, Loss: {'train/loss': 0.11289005726575851, 'train/grad_norm_pre_clip': 12.197086334228516, 'timing/data': 0.019229071214795113, 'timing/model': 0.924579051323235, 'learning_rate/qwen_vl_interface': 3.883311400920366e-06, 'learning_rate/action_model': 3.883311400920366e-05, 'epoch': 1.9}) +[2026-06-17 23:13:31,441][starVLA.training.train_starvla][INFO] - Step 4192, Loss: {'train/loss': 0.24758405983448029, 'train/grad_norm_pre_clip': 21.768583297729492, 'timing/data': 0.014750218950212002, 'timing/model': 0.9355681883171201, 'learning_rate/qwen_vl_interface': 3.88124030489638e-06, 'learning_rate/action_model': 3.88124030489638e-05, 'epoch': 1.9}) +[2026-06-17 23:13:35,212][starVLA.training.train_starvla][INFO] - Step 4193, Loss: {'train/loss': 0.17709219455718994, 'train/grad_norm_pre_clip': 15.513779640197754, 'timing/data': 0.017705241218209267, 'timing/model': 1.1018448136746883, 'learning_rate/qwen_vl_interface': 3.8791694926177495e-06, 'learning_rate/action_model': 3.87916949261775e-05, 'epoch': 1.9}) +[2026-06-17 23:13:39,047][starVLA.training.train_starvla][INFO] - Step 4194, Loss: {'train/loss': 0.2430877983570099, 'train/grad_norm_pre_clip': 22.558612823486328, 'timing/data': 0.017442782409489155, 'timing/model': 1.0822932105511427, 'learning_rate/qwen_vl_interface': 3.877098964513759e-06, 'learning_rate/action_model': 3.877098964513759e-05, 'epoch': 1.9}) +[2026-06-17 23:13:42,639][starVLA.training.train_starvla][INFO] - Step 4195, Loss: {'train/loss': 0.11881381273269653, 'train/grad_norm_pre_clip': 19.256940841674805, 'timing/data': 0.01666625589132309, 'timing/model': 0.927380339242518, 'learning_rate/qwen_vl_interface': 3.875028721013633e-06, 'learning_rate/action_model': 3.875028721013632e-05, 'epoch': 1.9}) +[2026-06-17 23:13:46,427][starVLA.training.train_starvla][INFO] - Step 4196, Loss: {'train/loss': 0.21553830802440643, 'train/grad_norm_pre_clip': 14.673243522644043, 'timing/data': 0.017577863298356533, 'timing/model': 0.9437528168782592, 'learning_rate/qwen_vl_interface': 3.87295876254653e-06, 'learning_rate/action_model': 3.8729587625465304e-05, 'epoch': 1.9}) +[2026-06-17 23:13:50,237][starVLA.training.train_starvla][INFO] - Step 4197, Loss: {'train/loss': 0.16717670857906342, 'train/grad_norm_pre_clip': 19.679088592529297, 'timing/data': 0.01820049900561571, 'timing/model': 0.9422789514064789, 'learning_rate/qwen_vl_interface': 3.870889089541558e-06, 'learning_rate/action_model': 3.870889089541558e-05, 'epoch': 1.9}) +[2026-06-17 23:13:54,103][starVLA.training.train_starvla][INFO] - Step 4198, Loss: {'train/loss': 0.15331590175628662, 'train/grad_norm_pre_clip': 14.776965141296387, 'timing/data': 0.01579525787383318, 'timing/model': 1.0085865203291178, 'learning_rate/qwen_vl_interface': 3.868819702427762e-06, 'learning_rate/action_model': 3.8688197024277616e-05, 'epoch': 1.9}) +[2026-06-17 23:13:58,989][starVLA.training.train_starvla][INFO] - Step 4199, Loss: {'train/loss': 0.19315128028392792, 'train/grad_norm_pre_clip': 23.186141967773438, 'timing/data': 0.017228228971362114, 'timing/model': 0.9567215433344245, 'learning_rate/qwen_vl_interface': 3.866750601634128e-06, 'learning_rate/action_model': 3.8667506016341277e-05, 'epoch': 1.9}) +[2026-06-17 23:16:10,256][starVLA.training.train_starvla][INFO] - Step 4200, Loss: {'train/loss': 0.40533721446990967, 'train/grad_norm_pre_clip': 32.56580352783203, 'eval/loss': 0.308839350938797, 'eval/latency_8/loss': 0.308839350938797, 'eval/demon_fire_precision': 0.94907683144729, 'eval/demon_fire_recall': 0.9900590245417832, 'eval/demon_fire_f1': 0.9691348639197203, 'eval/demon_fire_rate_model': 0.9343350027824151, 'eval/demon_fire_rate_teacher': 0.8956594323873122, 'eval/demon_left_precision': 0.8826530612244898, 'eval/demon_left_recall': 0.7408993576017131, 'eval/demon_left_f1': 0.8055878928987195, 'eval/demon_left_rate_model': 0.10907067334446299, 'eval/demon_left_rate_teacher': 0.12993878686700056, 'eval/demon_right_precision': 0.8623548922056384, 'eval/demon_right_recall': 0.7878787878787878, 'eval/demon_right_f1': 0.8234362628661915, 'eval/demon_right_rate_model': 0.167779632721202, 'eval/demon_right_rate_teacher': 0.18363939899833054, 'eval/demon_move_f1': 0.8145120778824555, 'eval/demon_cc_f1': 0.916916073657044, 'eval/latency_8/demon_fire_precision': 0.94907683144729, 'eval/latency_8/demon_fire_recall': 0.9900590245417832, 'eval/latency_8/demon_fire_f1': 0.9691348639197203, 'eval/latency_8/demon_fire_rate_model': 0.9343350027824151, 'eval/latency_8/demon_fire_rate_teacher': 0.8956594323873122, 'eval/latency_8/demon_left_precision': 0.8826530612244898, 'eval/latency_8/demon_left_recall': 0.7408993576017131, 'eval/latency_8/demon_left_f1': 0.8055878928987195, 'eval/latency_8/demon_left_rate_model': 0.10907067334446299, 'eval/latency_8/demon_left_rate_teacher': 0.12993878686700056, 'eval/latency_8/demon_right_precision': 0.8623548922056384, 'eval/latency_8/demon_right_recall': 0.7878787878787878, 'eval/latency_8/demon_right_f1': 0.8234362628661915, 'eval/latency_8/demon_right_rate_model': 0.167779632721202, 'eval/latency_8/demon_right_rate_teacher': 0.18363939899833054, 'eval/latency_8/demon_move_f1': 0.8145120778824555, 'eval/latency_8/demon_cc_f1': 0.916916073657044, 'timing/data': 0.016927579417824745, 'timing/model': 1.2672975054010749, 'learning_rate/qwen_vl_interface': 3.864681787589581e-06, 'learning_rate/action_model': 3.8646817875895816e-05, 'epoch': 1.91}) +[2026-06-17 23:16:10,293][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_4200_state +[2026-06-17 23:16:10,298][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 23:16:16,744][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_4200_state/model.safetensors +[2026-06-17 23:16:31,651][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_4200_state/optimizer.bin +[2026-06-17 23:16:31,651][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_4200_state/sampler.bin +[2026-06-17 23:16:31,652][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_4200_state/sampler_1.bin +[2026-06-17 23:16:31,653][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_4200_state/random_states_0.pkl +[2026-06-17 23:16:31,656][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4200 +[2026-06-17 23:16:31,656][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 23:16:31,666][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 23:16:36,177][starVLA.training.train_starvla][INFO] - Step 4201, Loss: {'train/loss': 0.22589920461177826, 'train/grad_norm_pre_clip': 26.097299575805664, 'timing/data': 0.01993489172309637, 'timing/model': 1.1471299780532718, 'learning_rate/qwen_vl_interface': 3.862613260722991e-06, 'learning_rate/action_model': 3.862613260722992e-05, 'epoch': 1.91}) +[2026-06-17 23:16:40,265][starVLA.training.train_starvla][INFO] - Step 4202, Loss: {'train/loss': 0.2655303180217743, 'train/grad_norm_pre_clip': 30.34784507751465, 'timing/data': 0.016254443675279617, 'timing/model': 1.2300360137596726, 'learning_rate/qwen_vl_interface': 3.860545021463166e-06, 'learning_rate/action_model': 3.860545021463167e-05, 'epoch': 1.91}) +[2026-06-17 23:16:46,781][starVLA.training.train_starvla][INFO] - Step 4203, Loss: {'train/loss': 0.1971823275089264, 'train/grad_norm_pre_clip': 16.17017936706543, 'timing/data': 0.016348207369446754, 'timing/model': 1.1471939105540514, 'learning_rate/qwen_vl_interface': 3.858477070238854e-06, 'learning_rate/action_model': 3.8584770702388534e-05, 'epoch': 1.91}) +[2026-06-17 23:16:50,501][starVLA.training.train_starvla][INFO] - Step 4204, Loss: {'train/loss': 0.2926870882511139, 'train/grad_norm_pre_clip': 27.552684783935547, 'timing/data': 0.021301177330315113, 'timing/model': 0.9455074444413185, 'learning_rate/qwen_vl_interface': 3.856409407478741e-06, 'learning_rate/action_model': 3.856409407478741e-05, 'epoch': 1.91}) +[2026-06-17 23:16:54,429][starVLA.training.train_starvla][INFO] - Step 4205, Loss: {'train/loss': 0.2595866620540619, 'train/grad_norm_pre_clip': 29.030000686645508, 'timing/data': 0.016701715067029, 'timing/model': 0.9457517471164465, 'learning_rate/qwen_vl_interface': 3.854342033611456e-06, 'learning_rate/action_model': 3.854342033611456e-05, 'epoch': 1.91}) +[2026-06-17 23:16:58,781][starVLA.training.train_starvla][INFO] - Step 4206, Loss: {'train/loss': 0.20903640985488892, 'train/grad_norm_pre_clip': 20.787538528442383, 'timing/data': 0.01301173772662878, 'timing/model': 1.4634835151955485, 'learning_rate/qwen_vl_interface': 3.852274949065571e-06, 'learning_rate/action_model': 3.852274949065571e-05, 'epoch': 1.91}) +[2026-06-17 23:17:02,564][starVLA.training.train_starvla][INFO] - Step 4207, Loss: {'train/loss': 0.1850602626800537, 'train/grad_norm_pre_clip': 20.569515228271484, 'timing/data': 0.01417173258960247, 'timing/model': 0.9455195842310786, 'learning_rate/qwen_vl_interface': 3.850208154269593e-06, 'learning_rate/action_model': 3.8502081542695934e-05, 'epoch': 1.91}) +[2026-06-17 23:17:08,267][starVLA.training.train_starvla][INFO] - Step 4208, Loss: {'train/loss': 0.10342855751514435, 'train/grad_norm_pre_clip': 17.451263427734375, 'timing/data': 0.018186544068157673, 'timing/model': 0.9444508999586105, 'learning_rate/qwen_vl_interface': 3.8481416496519725e-06, 'learning_rate/action_model': 3.848141649651972e-05, 'epoch': 1.91}) +[2026-06-17 23:17:14,561][starVLA.training.train_starvla][INFO] - Step 4209, Loss: {'train/loss': 0.2292141169309616, 'train/grad_norm_pre_clip': 29.005489349365234, 'timing/data': 0.014214560389518738, 'timing/model': 1.1650012005120516, 'learning_rate/qwen_vl_interface': 3.846075435641093e-06, 'learning_rate/action_model': 3.846075435641093e-05, 'epoch': 1.91}) +[2026-06-17 23:17:18,271][starVLA.training.train_starvla][INFO] - Step 4210, Loss: {'train/loss': 0.27522042393684387, 'train/grad_norm_pre_clip': 24.899612426757812, 'timing/data': 0.017699165269732475, 'timing/model': 0.9395599365234375, 'learning_rate/qwen_vl_interface': 3.844009512665289e-06, 'learning_rate/action_model': 3.844009512665289e-05, 'epoch': 1.91}) +[2026-06-17 23:17:22,956][starVLA.training.train_starvla][INFO] - Step 4211, Loss: {'train/loss': 0.26165327429771423, 'train/grad_norm_pre_clip': 22.733640670776367, 'timing/data': 0.013832191936671734, 'timing/model': 0.9322368707507849, 'learning_rate/qwen_vl_interface': 3.841943881152826e-06, 'learning_rate/action_model': 3.841943881152826e-05, 'epoch': 1.91}) +[2026-06-17 23:17:26,717][starVLA.training.train_starvla][INFO] - Step 4212, Loss: {'train/loss': 0.3440726697444916, 'train/grad_norm_pre_clip': 31.89559555053711, 'timing/data': 0.015552845783531666, 'timing/model': 0.9455431224778295, 'learning_rate/qwen_vl_interface': 3.83987854153191e-06, 'learning_rate/action_model': 3.83987854153191e-05, 'epoch': 1.91}) +[2026-06-17 23:17:31,586][starVLA.training.train_starvla][INFO] - Step 4213, Loss: {'train/loss': 0.13422825932502747, 'train/grad_norm_pre_clip': 15.853474617004395, 'timing/data': 0.01869258377701044, 'timing/model': 1.836504734121263, 'learning_rate/qwen_vl_interface': 3.83781349423069e-06, 'learning_rate/action_model': 3.83781349423069e-05, 'epoch': 1.91}) +[2026-06-17 23:17:35,459][starVLA.training.train_starvla][INFO] - Step 4214, Loss: {'train/loss': 0.26442810893058777, 'train/grad_norm_pre_clip': 20.26174545288086, 'timing/data': 0.015892612747848034, 'timing/model': 0.9377678940072656, 'learning_rate/qwen_vl_interface': 3.835748739677253e-06, 'learning_rate/action_model': 3.835748739677253e-05, 'epoch': 1.91}) +[2026-06-17 23:17:39,566][starVLA.training.train_starvla][INFO] - Step 4215, Loss: {'train/loss': 0.052283938974142075, 'train/grad_norm_pre_clip': 5.698178291320801, 'timing/data': 0.015060164034366608, 'timing/model': 1.131306572817266, 'learning_rate/qwen_vl_interface': 3.8336842782996245e-06, 'learning_rate/action_model': 3.8336842782996244e-05, 'epoch': 1.91}) +[2026-06-17 23:17:43,997][starVLA.training.train_starvla][INFO] - Step 4216, Loss: {'train/loss': 0.11559345573186874, 'train/grad_norm_pre_clip': 9.229331016540527, 'timing/data': 0.018056362867355347, 'timing/model': 1.1375623904168606, 'learning_rate/qwen_vl_interface': 3.831620110525766e-06, 'learning_rate/action_model': 3.831620110525767e-05, 'epoch': 1.91}) +[2026-06-17 23:17:48,343][starVLA.training.train_starvla][INFO] - Step 4217, Loss: {'train/loss': 0.29329752922058105, 'train/grad_norm_pre_clip': 26.17510986328125, 'timing/data': 0.013880538754165173, 'timing/model': 0.94536252040416, 'learning_rate/qwen_vl_interface': 3.829556236783588e-06, 'learning_rate/action_model': 3.8295562367835884e-05, 'epoch': 1.91}) +[2026-06-17 23:17:52,132][starVLA.training.train_starvla][INFO] - Step 4218, Loss: {'train/loss': 0.29006096720695496, 'train/grad_norm_pre_clip': 16.853235244750977, 'timing/data': 0.014621593989431858, 'timing/model': 0.9304979098960757, 'learning_rate/qwen_vl_interface': 3.827492657500931e-06, 'learning_rate/action_model': 3.8274926575009305e-05, 'epoch': 1.91}) +[2026-06-17 23:17:55,904][starVLA.training.train_starvla][INFO] - Step 4219, Loss: {'train/loss': 0.4107678234577179, 'train/grad_norm_pre_clip': 28.248844146728516, 'timing/data': 0.0186216551810503, 'timing/model': 0.932007790543139, 'learning_rate/qwen_vl_interface': 3.8254293731055754e-06, 'learning_rate/action_model': 3.825429373105576e-05, 'epoch': 1.91}) +[2026-06-17 23:17:59,659][starVLA.training.train_starvla][INFO] - Step 4220, Loss: {'train/loss': 0.13854967057704926, 'train/grad_norm_pre_clip': 17.253551483154297, 'timing/data': 0.02040868904441595, 'timing/model': 0.924775755032897, 'learning_rate/qwen_vl_interface': 3.823366384025243e-06, 'learning_rate/action_model': 3.8233663840252426e-05, 'epoch': 1.91}) +[2026-06-17 23:18:03,433][starVLA.training.train_starvla][INFO] - Step 4221, Loss: {'train/loss': 0.39495861530303955, 'train/grad_norm_pre_clip': 25.186779022216797, 'timing/data': 0.01606629602611065, 'timing/model': 0.9373050583526492, 'learning_rate/qwen_vl_interface': 3.821303690687595e-06, 'learning_rate/action_model': 3.821303690687595e-05, 'epoch': 1.91}) +[2026-06-17 23:18:07,443][starVLA.training.train_starvla][INFO] - Step 4222, Loss: {'train/loss': 0.11673424392938614, 'train/grad_norm_pre_clip': 15.424193382263184, 'timing/data': 0.06716351769864559, 'timing/model': 1.225729949772358, 'learning_rate/qwen_vl_interface': 3.81924129352023e-06, 'learning_rate/action_model': 3.8192412935202304e-05, 'epoch': 1.92}) +[2026-06-17 23:18:11,282][starVLA.training.train_starvla][INFO] - Step 4223, Loss: {'train/loss': 0.45796817541122437, 'train/grad_norm_pre_clip': 29.501802444458008, 'timing/data': 0.016739018261432648, 'timing/model': 1.1702871955931187, 'learning_rate/qwen_vl_interface': 3.817179192950682e-06, 'learning_rate/action_model': 3.8171791929506826e-05, 'epoch': 1.92}) +[2026-06-17 23:18:15,029][starVLA.training.train_starvla][INFO] - Step 4224, Loss: {'train/loss': 0.4054913818836212, 'train/grad_norm_pre_clip': 29.899930953979492, 'timing/data': 0.01679995283484459, 'timing/model': 0.9300952591001987, 'learning_rate/qwen_vl_interface': 3.815117389406432e-06, 'learning_rate/action_model': 3.8151173894064324e-05, 'epoch': 1.92}) +[2026-06-17 23:18:18,882][starVLA.training.train_starvla][INFO] - Step 4225, Loss: {'train/loss': 0.37406155467033386, 'train/grad_norm_pre_clip': 22.000186920166016, 'timing/data': 0.019029499031603336, 'timing/model': 0.9342394461855292, 'learning_rate/qwen_vl_interface': 3.8130558833148906e-06, 'learning_rate/action_model': 3.813055883314891e-05, 'epoch': 1.92}) +[2026-06-17 23:18:25,295][starVLA.training.train_starvla][INFO] - Step 4226, Loss: {'train/loss': 0.41482678055763245, 'train/grad_norm_pre_clip': 31.843660354614258, 'timing/data': 0.021457928232848644, 'timing/model': 3.45587988011539, 'learning_rate/qwen_vl_interface': 3.81099467510341e-06, 'learning_rate/action_model': 3.81099467510341e-05, 'epoch': 1.92}) +[2026-06-17 23:18:29,851][starVLA.training.train_starvla][INFO] - Step 4227, Loss: {'train/loss': 0.45761388540267944, 'train/grad_norm_pre_clip': 26.483943939208984, 'timing/data': 0.016356169246137142, 'timing/model': 0.929426521062851, 'learning_rate/qwen_vl_interface': 3.8089337651992815e-06, 'learning_rate/action_model': 3.808933765199281e-05, 'epoch': 1.92}) +[2026-06-17 23:18:33,707][starVLA.training.train_starvla][INFO] - Step 4228, Loss: {'train/loss': 0.10265114158391953, 'train/grad_norm_pre_clip': 12.173185348510742, 'timing/data': 0.0214419886469841, 'timing/model': 0.9385841051116586, 'learning_rate/qwen_vl_interface': 3.8068731540297344e-06, 'learning_rate/action_model': 3.806873154029734e-05, 'epoch': 1.92}) +[2026-06-17 23:18:37,545][starVLA.training.train_starvla][INFO] - Step 4229, Loss: {'train/loss': 0.5396503210067749, 'train/grad_norm_pre_clip': 33.81694030761719, 'timing/data': 0.021326429210603237, 'timing/model': 1.14981300663203, 'learning_rate/qwen_vl_interface': 3.8048128420219366e-06, 'learning_rate/action_model': 3.804812842021936e-05, 'epoch': 1.92}) +[2026-06-17 23:18:41,145][starVLA.training.train_starvla][INFO] - Step 4230, Loss: {'train/loss': 0.2789389193058014, 'train/grad_norm_pre_clip': 21.61943817138672, 'timing/data': 0.021083415485918522, 'timing/model': 0.9324062541127205, 'learning_rate/qwen_vl_interface': 3.80275282960299e-06, 'learning_rate/action_model': 3.80275282960299e-05, 'epoch': 1.92}) +[2026-06-17 23:18:44,953][starVLA.training.train_starvla][INFO] - Step 4231, Loss: {'train/loss': 0.529802143573761, 'train/grad_norm_pre_clip': 21.44204330444336, 'timing/data': 0.022330869920551777, 'timing/model': 0.9249111050739884, 'learning_rate/qwen_vl_interface': 3.8006931171999394e-06, 'learning_rate/action_model': 3.800693117199939e-05, 'epoch': 1.92}) +[2026-06-17 23:18:48,755][starVLA.training.train_starvla][INFO] - Step 4232, Loss: {'train/loss': 0.1386820673942566, 'train/grad_norm_pre_clip': 15.29651165008545, 'timing/data': 0.014232257381081581, 'timing/model': 0.9327402617782354, 'learning_rate/qwen_vl_interface': 3.798633705239766e-06, 'learning_rate/action_model': 3.7986337052397655e-05, 'epoch': 1.92}) +[2026-06-17 23:18:52,788][starVLA.training.train_starvla][INFO] - Step 4233, Loss: {'train/loss': 0.27059316635131836, 'train/grad_norm_pre_clip': 25.09018898010254, 'timing/data': 0.017892307601869106, 'timing/model': 0.9376571364700794, 'learning_rate/qwen_vl_interface': 3.796574594149387e-06, 'learning_rate/action_model': 3.796574594149387e-05, 'epoch': 1.92}) +[2026-06-17 23:18:56,638][starVLA.training.train_starvla][INFO] - Step 4234, Loss: {'train/loss': 0.5482633113861084, 'train/grad_norm_pre_clip': 28.475473403930664, 'timing/data': 0.014887701719999313, 'timing/model': 0.9314667861908674, 'learning_rate/qwen_vl_interface': 3.794515784355657e-06, 'learning_rate/action_model': 3.7945157843556575e-05, 'epoch': 1.92}) +[2026-06-17 23:19:00,384][starVLA.training.train_starvla][INFO] - Step 4235, Loss: {'train/loss': 0.30038896203041077, 'train/grad_norm_pre_clip': 40.36341094970703, 'timing/data': 0.016466915607452393, 'timing/model': 0.9310300527140498, 'learning_rate/qwen_vl_interface': 3.7924572762853727e-06, 'learning_rate/action_model': 3.7924572762853725e-05, 'epoch': 1.92}) +[2026-06-17 23:19:04,226][starVLA.training.train_starvla][INFO] - Step 4236, Loss: {'train/loss': 0.1793939769268036, 'train/grad_norm_pre_clip': 13.14164924621582, 'timing/data': 0.01626473944634199, 'timing/model': 1.1351453326642513, 'learning_rate/qwen_vl_interface': 3.7903990703652643e-06, 'learning_rate/action_model': 3.790399070365264e-05, 'epoch': 1.92}) +[2026-06-17 23:19:07,899][starVLA.training.train_starvla][INFO] - Step 4237, Loss: {'train/loss': 0.3536679148674011, 'train/grad_norm_pre_clip': 23.37212562561035, 'timing/data': 0.02904821466654539, 'timing/model': 0.9510661382228136, 'learning_rate/qwen_vl_interface': 3.7883411670219976e-06, 'learning_rate/action_model': 3.788341167021998e-05, 'epoch': 1.92}) +[2026-06-17 23:19:11,747][starVLA.training.train_starvla][INFO] - Step 4238, Loss: {'train/loss': 0.236257404088974, 'train/grad_norm_pre_clip': 25.72063636779785, 'timing/data': 0.015482932329177856, 'timing/model': 0.9363862294703722, 'learning_rate/qwen_vl_interface': 3.7862835666821795e-06, 'learning_rate/action_model': 3.78628356668218e-05, 'epoch': 1.92}) +[2026-06-17 23:19:15,582][starVLA.training.train_starvla][INFO] - Step 4239, Loss: {'train/loss': 0.20618225634098053, 'train/grad_norm_pre_clip': 20.899206161499023, 'timing/data': 0.026206964626908302, 'timing/model': 0.9440690549090505, 'learning_rate/qwen_vl_interface': 3.7842262697723543e-06, 'learning_rate/action_model': 3.784226269772354e-05, 'epoch': 1.92}) +[2026-06-17 23:19:19,494][starVLA.training.train_starvla][INFO] - Step 4240, Loss: {'train/loss': 0.5476793646812439, 'train/grad_norm_pre_clip': 25.3297061920166, 'timing/data': 0.016616337932646275, 'timing/model': 1.015564164146781, 'learning_rate/qwen_vl_interface': 3.782169276719e-06, 'learning_rate/action_model': 3.782169276719e-05, 'epoch': 1.92}) +[2026-06-17 23:19:23,339][starVLA.training.train_starvla][INFO] - Step 4241, Loss: {'train/loss': 0.45400017499923706, 'train/grad_norm_pre_clip': 37.1303596496582, 'timing/data': 0.01545232068747282, 'timing/model': 0.9506849907338619, 'learning_rate/qwen_vl_interface': 3.7801125879485355e-06, 'learning_rate/action_model': 3.7801125879485356e-05, 'epoch': 1.92}) +[2026-06-17 23:19:27,207][starVLA.training.train_starvla][INFO] - Step 4242, Loss: {'train/loss': 0.46510764956474304, 'train/grad_norm_pre_clip': 35.291507720947266, 'timing/data': 0.01496074441820383, 'timing/model': 1.1279594581574202, 'learning_rate/qwen_vl_interface': 3.77805620388731e-06, 'learning_rate/action_model': 3.7780562038873105e-05, 'epoch': 1.92}) +[2026-06-17 23:19:31,043][starVLA.training.train_starvla][INFO] - Step 4243, Loss: {'train/loss': 0.07449080795049667, 'train/grad_norm_pre_clip': 7.998510360717773, 'timing/data': 0.0257082125172019, 'timing/model': 1.1326091503724456, 'learning_rate/qwen_vl_interface': 3.7760001249616204e-06, 'learning_rate/action_model': 3.77600012496162e-05, 'epoch': 1.92}) +[2026-06-17 23:19:34,709][starVLA.training.train_starvla][INFO] - Step 4244, Loss: {'train/loss': 0.35702604055404663, 'train/grad_norm_pre_clip': 23.151832580566406, 'timing/data': 0.02697180863469839, 'timing/model': 0.950686501339078, 'learning_rate/qwen_vl_interface': 3.7739443515976898e-06, 'learning_rate/action_model': 3.7739443515976894e-05, 'epoch': 1.93}) +[2026-06-17 23:19:38,529][starVLA.training.train_starvla][INFO] - Step 4245, Loss: {'train/loss': 0.14458319544792175, 'train/grad_norm_pre_clip': 12.802834510803223, 'timing/data': 0.02071959897875786, 'timing/model': 0.930491590872407, 'learning_rate/qwen_vl_interface': 3.7718888842216822e-06, 'learning_rate/action_model': 3.771888884221682e-05, 'epoch': 1.93}) +[2026-06-17 23:19:42,361][starVLA.training.train_starvla][INFO] - Step 4246, Loss: {'train/loss': 0.34329137206077576, 'train/grad_norm_pre_clip': 25.745683670043945, 'timing/data': 0.015430054627358913, 'timing/model': 0.9476127699017525, 'learning_rate/qwen_vl_interface': 3.7698337232597002e-06, 'learning_rate/action_model': 3.7698337232597e-05, 'epoch': 1.93}) +[2026-06-17 23:19:46,256][starVLA.training.train_starvla][INFO] - Step 4247, Loss: {'train/loss': 0.35780996084213257, 'train/grad_norm_pre_clip': 31.187297821044922, 'timing/data': 0.01643595565110445, 'timing/model': 0.9496619030833244, 'learning_rate/qwen_vl_interface': 3.7677788691377814e-06, 'learning_rate/action_model': 3.7677788691377817e-05, 'epoch': 1.93}) +[2026-06-17 23:19:50,285][starVLA.training.train_starvla][INFO] - Step 4248, Loss: {'train/loss': 0.27721869945526123, 'train/grad_norm_pre_clip': 28.82412338256836, 'timing/data': 0.014371991157531738, 'timing/model': 0.9773479932919145, 'learning_rate/qwen_vl_interface': 3.7657243222818978e-06, 'learning_rate/action_model': 3.7657243222818976e-05, 'epoch': 1.93}) +[2026-06-17 23:19:54,154][starVLA.training.train_starvla][INFO] - Step 4249, Loss: {'train/loss': 0.2809548079967499, 'train/grad_norm_pre_clip': 27.25296974182129, 'timing/data': 0.01666670199483633, 'timing/model': 1.1403634529560804, 'learning_rate/qwen_vl_interface': 3.7636700831179575e-06, 'learning_rate/action_model': 3.763670083117957e-05, 'epoch': 1.93}) +[2026-06-17 23:19:58,100][starVLA.training.train_starvla][INFO] - Step 4250, Loss: {'train/loss': 0.17878839373588562, 'train/grad_norm_pre_clip': 14.273178100585938, 'timing/data': 0.016566705889999866, 'timing/model': 1.2032483061775565, 'learning_rate/qwen_vl_interface': 3.7616161520718096e-06, 'learning_rate/action_model': 3.76161615207181e-05, 'epoch': 1.93}) +[2026-06-17 23:20:01,863][starVLA.training.train_starvla][INFO] - Step 4251, Loss: {'train/loss': 0.32973119616508484, 'train/grad_norm_pre_clip': 25.42282485961914, 'timing/data': 0.017117872834205627, 'timing/model': 0.9697973756119609, 'learning_rate/qwen_vl_interface': 3.7595625295692352e-06, 'learning_rate/action_model': 3.759562529569235e-05, 'epoch': 1.93}) +[2026-06-17 23:20:05,757][starVLA.training.train_starvla][INFO] - Step 4252, Loss: {'train/loss': 0.05162894353270531, 'train/grad_norm_pre_clip': 7.133416175842285, 'timing/data': 0.0139016043394804, 'timing/model': 0.9523074999451637, 'learning_rate/qwen_vl_interface': 3.7575092160359523e-06, 'learning_rate/action_model': 3.757509216035952e-05, 'epoch': 1.93}) +[2026-06-17 23:20:09,671][starVLA.training.train_starvla][INFO] - Step 4253, Loss: {'train/loss': 0.16969870030879974, 'train/grad_norm_pre_clip': 19.38755226135254, 'timing/data': 0.015895644202828407, 'timing/model': 0.9530411269515753, 'learning_rate/qwen_vl_interface': 3.7554562118976135e-06, 'learning_rate/action_model': 3.7554562118976136e-05, 'epoch': 1.93}) +[2026-06-17 23:20:13,629][starVLA.training.train_starvla][INFO] - Step 4254, Loss: {'train/loss': 0.3304394781589508, 'train/grad_norm_pre_clip': 26.45540428161621, 'timing/data': 0.017917931079864502, 'timing/model': 1.0014548879116774, 'learning_rate/qwen_vl_interface': 3.7534035175798113e-06, 'learning_rate/action_model': 3.753403517579811e-05, 'epoch': 1.93}) +[2026-06-17 23:20:17,567][starVLA.training.train_starvla][INFO] - Step 4255, Loss: {'train/loss': 0.2910114824771881, 'train/grad_norm_pre_clip': 23.737966537475586, 'timing/data': 0.015861605294048786, 'timing/model': 0.9491952788084745, 'learning_rate/qwen_vl_interface': 3.7513511335080706e-06, 'learning_rate/action_model': 3.75135113350807e-05, 'epoch': 1.93}) +[2026-06-17 23:20:21,623][starVLA.training.train_starvla][INFO] - Step 4256, Loss: {'train/loss': 0.2184504121541977, 'train/grad_norm_pre_clip': 20.03525161743164, 'timing/data': 0.015951081179082394, 'timing/model': 1.210739355534315, 'learning_rate/qwen_vl_interface': 3.74929906010785e-06, 'learning_rate/action_model': 3.74929906010785e-05, 'epoch': 1.93}) +[2026-06-17 23:20:25,397][starVLA.training.train_starvla][INFO] - Step 4257, Loss: {'train/loss': 0.5866492986679077, 'train/grad_norm_pre_clip': 24.679275512695312, 'timing/data': 0.01860534120351076, 'timing/model': 1.029548717662692, 'learning_rate/qwen_vl_interface': 3.747247297804553e-06, 'learning_rate/action_model': 3.7472472978045526e-05, 'epoch': 1.93}) +[2026-06-17 23:20:29,326][starVLA.training.train_starvla][INFO] - Step 4258, Loss: {'train/loss': 0.2126416712999344, 'train/grad_norm_pre_clip': 18.30119514465332, 'timing/data': 0.024154520593583584, 'timing/model': 0.9894183622673154, 'learning_rate/qwen_vl_interface': 3.745195847023506e-06, 'learning_rate/action_model': 3.745195847023506e-05, 'epoch': 1.93}) +[2026-06-17 23:20:33,192][starVLA.training.train_starvla][INFO] - Step 4259, Loss: {'train/loss': 0.2980433404445648, 'train/grad_norm_pre_clip': 26.056503295898438, 'timing/data': 0.026967324316501617, 'timing/model': 0.9802316380664706, 'learning_rate/qwen_vl_interface': 3.7431447081899795e-06, 'learning_rate/action_model': 3.7431447081899796e-05, 'epoch': 1.93}) +[2026-06-17 23:20:37,218][starVLA.training.train_starvla][INFO] - Step 4260, Loss: {'train/loss': 0.21615710854530334, 'train/grad_norm_pre_clip': 18.433774948120117, 'timing/data': 0.02127217873930931, 'timing/model': 0.9529085764661431, 'learning_rate/qwen_vl_interface': 3.7410938817291753e-06, 'learning_rate/action_model': 3.741093881729175e-05, 'epoch': 1.93}) +[2026-06-17 23:20:41,130][starVLA.training.train_starvla][INFO] - Step 4261, Loss: {'train/loss': 0.46996697783470154, 'train/grad_norm_pre_clip': 30.761871337890625, 'timing/data': 0.015408260747790337, 'timing/model': 0.9856323450803757, 'learning_rate/qwen_vl_interface': 3.739043368066234e-06, 'learning_rate/action_model': 3.739043368066234e-05, 'epoch': 1.93}) +[2026-06-17 23:20:45,343][starVLA.training.train_starvla][INFO] - Step 4262, Loss: {'train/loss': 0.3406175971031189, 'train/grad_norm_pre_clip': 34.4301872253418, 'timing/data': 0.01601708773523569, 'timing/model': 1.4321969151496887, 'learning_rate/qwen_vl_interface': 3.7369931676262304e-06, 'learning_rate/action_model': 3.7369931676262304e-05, 'epoch': 1.93}) +[2026-06-17 23:20:49,162][starVLA.training.train_starvla][INFO] - Step 4263, Loss: {'train/loss': 0.11842797696590424, 'train/grad_norm_pre_clip': 16.971038818359375, 'timing/data': 0.030251773074269295, 'timing/model': 1.1244199387729168, 'learning_rate/qwen_vl_interface': 3.7349432808341673e-06, 'learning_rate/action_model': 3.7349432808341674e-05, 'epoch': 1.93}) +[2026-06-17 23:20:52,900][starVLA.training.train_starvla][INFO] - Step 4264, Loss: {'train/loss': 0.2550313472747803, 'train/grad_norm_pre_clip': 18.64052391052246, 'timing/data': 0.02159480843693018, 'timing/model': 0.9596586525440216, 'learning_rate/qwen_vl_interface': 3.7328937081149952e-06, 'learning_rate/action_model': 3.732893708114995e-05, 'epoch': 1.93}) +[2026-06-17 23:20:56,743][starVLA.training.train_starvla][INFO] - Step 4265, Loss: {'train/loss': 0.23712776601314545, 'train/grad_norm_pre_clip': 18.645862579345703, 'timing/data': 0.01945251878350973, 'timing/model': 0.9447019463405013, 'learning_rate/qwen_vl_interface': 3.7308444498935894e-06, 'learning_rate/action_model': 3.730844449893589e-05, 'epoch': 1.93}) +[2026-06-17 23:21:00,746][starVLA.training.train_starvla][INFO] - Step 4266, Loss: {'train/loss': 0.232143372297287, 'train/grad_norm_pre_clip': 18.85837745666504, 'timing/data': 0.018679648637771606, 'timing/model': 0.9432062404230237, 'learning_rate/qwen_vl_interface': 3.7287955065947634e-06, 'learning_rate/action_model': 3.728795506594763e-05, 'epoch': 1.94}) +[2026-06-17 23:21:04,580][starVLA.training.train_starvla][INFO] - Step 4267, Loss: {'train/loss': 0.08358725905418396, 'train/grad_norm_pre_clip': 10.81954288482666, 'timing/data': 0.016804642044007778, 'timing/model': 0.9537749951705337, 'learning_rate/qwen_vl_interface': 3.7267468786432655e-06, 'learning_rate/action_model': 3.7267468786432654e-05, 'epoch': 1.94}) +[2026-06-17 23:21:08,440][starVLA.training.train_starvla][INFO] - Step 4268, Loss: {'train/loss': 0.4115839898586273, 'train/grad_norm_pre_clip': 24.580841064453125, 'timing/data': 0.017185906879603863, 'timing/model': 0.9605295201763511, 'learning_rate/qwen_vl_interface': 3.7246985664637787e-06, 'learning_rate/action_model': 3.724698566463778e-05, 'epoch': 1.94}) +[2026-06-17 23:21:12,398][starVLA.training.train_starvla][INFO] - Step 4269, Loss: {'train/loss': 0.4628680348396301, 'train/grad_norm_pre_clip': 24.911535263061523, 'timing/data': 0.01644018292427063, 'timing/model': 1.1898667570203543, 'learning_rate/qwen_vl_interface': 3.722650570480922e-06, 'learning_rate/action_model': 3.722650570480922e-05, 'epoch': 1.94}) +[2026-06-17 23:21:16,233][starVLA.training.train_starvla][INFO] - Step 4270, Loss: {'train/loss': 0.44311195611953735, 'train/grad_norm_pre_clip': 24.608196258544922, 'timing/data': 0.016072258353233337, 'timing/model': 1.138155434280634, 'learning_rate/qwen_vl_interface': 3.7206028911192445e-06, 'learning_rate/action_model': 3.720602891119244e-05, 'epoch': 1.94}) +[2026-06-17 23:21:19,853][starVLA.training.train_starvla][INFO] - Step 4271, Loss: {'train/loss': 0.40628960728645325, 'train/grad_norm_pre_clip': 28.434526443481445, 'timing/data': 0.024644787423312664, 'timing/model': 0.9344003600999713, 'learning_rate/qwen_vl_interface': 3.7185555288032315e-06, 'learning_rate/action_model': 3.7185555288032316e-05, 'epoch': 1.94}) +[2026-06-17 23:21:24,293][starVLA.training.train_starvla][INFO] - Step 4272, Loss: {'train/loss': 0.09634794294834137, 'train/grad_norm_pre_clip': 10.135534286499023, 'timing/data': 0.01841845829039812, 'timing/model': 1.0145740350708365, 'learning_rate/qwen_vl_interface': 3.716508483957308e-06, 'learning_rate/action_model': 3.716508483957308e-05, 'epoch': 1.94}) +[2026-06-17 23:21:29,799][starVLA.training.train_starvla][INFO] - Step 4273, Loss: {'train/loss': 0.23023252189159393, 'train/grad_norm_pre_clip': 21.813217163085938, 'timing/data': 0.020082369446754456, 'timing/model': 0.9826944684609771, 'learning_rate/qwen_vl_interface': 3.7144617570058245e-06, 'learning_rate/action_model': 3.7144617570058245e-05, 'epoch': 1.94}) +[2026-06-17 23:21:33,902][starVLA.training.train_starvla][INFO] - Step 4274, Loss: {'train/loss': 0.3239572048187256, 'train/grad_norm_pre_clip': 22.19086265563965, 'timing/data': 0.015474937856197357, 'timing/model': 0.9613418150693178, 'learning_rate/qwen_vl_interface': 3.7124153483730706e-06, 'learning_rate/action_model': 3.712415348373071e-05, 'epoch': 1.94}) +[2026-06-17 23:21:38,931][starVLA.training.train_starvla][INFO] - Step 4275, Loss: {'train/loss': 0.215005561709404, 'train/grad_norm_pre_clip': 18.569015502929688, 'timing/data': 0.020783857442438602, 'timing/model': 2.0464270086959004, 'learning_rate/qwen_vl_interface': 3.71036925848327e-06, 'learning_rate/action_model': 3.71036925848327e-05, 'epoch': 1.94}) +[2026-06-17 23:21:43,685][starVLA.training.train_starvla][INFO] - Step 4276, Loss: {'train/loss': 0.1351122260093689, 'train/grad_norm_pre_clip': 14.841373443603516, 'timing/data': 0.018457677215337753, 'timing/model': 1.2076281541958451, 'learning_rate/qwen_vl_interface': 3.7083234877605804e-06, 'learning_rate/action_model': 3.7083234877605806e-05, 'epoch': 1.94}) +[2026-06-17 23:21:47,366][starVLA.training.train_starvla][INFO] - Step 4277, Loss: {'train/loss': 0.26961663365364075, 'train/grad_norm_pre_clip': 31.52123260498047, 'timing/data': 0.01988869532942772, 'timing/model': 0.9376569101586938, 'learning_rate/qwen_vl_interface': 3.70627803662909e-06, 'learning_rate/action_model': 3.70627803662909e-05, 'epoch': 1.94}) +[2026-06-17 23:21:51,324][starVLA.training.train_starvla][INFO] - Step 4278, Loss: {'train/loss': 0.1458052694797516, 'train/grad_norm_pre_clip': 16.45653533935547, 'timing/data': 0.021898708306252956, 'timing/model': 0.9740683436393738, 'learning_rate/qwen_vl_interface': 3.7042329055128216e-06, 'learning_rate/action_model': 3.704232905512821e-05, 'epoch': 1.94}) +[2026-06-17 23:21:55,170][starVLA.training.train_starvla][INFO] - Step 4279, Loss: {'train/loss': 0.18762654066085815, 'train/grad_norm_pre_clip': 17.223472595214844, 'timing/data': 0.029579059220850468, 'timing/model': 0.9368796646595001, 'learning_rate/qwen_vl_interface': 3.7021880948357375e-06, 'learning_rate/action_model': 3.702188094835737e-05, 'epoch': 1.94}) +[2026-06-17 23:21:59,044][starVLA.training.train_starvla][INFO] - Step 4280, Loss: {'train/loss': 0.05188094452023506, 'train/grad_norm_pre_clip': 7.543667316436768, 'timing/data': 0.016026186756789684, 'timing/model': 0.9370412938296795, 'learning_rate/qwen_vl_interface': 3.700143605021726e-06, 'learning_rate/action_model': 3.7001436050217256e-05, 'epoch': 1.94}) +[2026-06-17 23:22:02,948][starVLA.training.train_starvla][INFO] - Step 4281, Loss: {'train/loss': 0.20936746895313263, 'train/grad_norm_pre_clip': 16.024946212768555, 'timing/data': 0.01626934390515089, 'timing/model': 0.9739913968369365, 'learning_rate/qwen_vl_interface': 3.6980994364946133e-06, 'learning_rate/action_model': 3.698099436494613e-05, 'epoch': 1.94}) +[2026-06-17 23:22:06,857][starVLA.training.train_starvla][INFO] - Step 4282, Loss: {'train/loss': 0.036674272269010544, 'train/grad_norm_pre_clip': 4.004546165466309, 'timing/data': 0.021358614787459373, 'timing/model': 0.9518470522016287, 'learning_rate/qwen_vl_interface': 3.696055589678156e-06, 'learning_rate/action_model': 3.6960555896781554e-05, 'epoch': 1.94}) +[2026-06-17 23:22:10,943][starVLA.training.train_starvla][INFO] - Step 4283, Loss: {'train/loss': 0.2700496017932892, 'train/grad_norm_pre_clip': 19.74445152282715, 'timing/data': 0.022131619043648243, 'timing/model': 1.2512117605656385, 'learning_rate/qwen_vl_interface': 3.694012064996049e-06, 'learning_rate/action_model': 3.694012064996049e-05, 'epoch': 1.94}) +[2026-06-17 23:22:14,630][starVLA.training.train_starvla][INFO] - Step 4284, Loss: {'train/loss': 0.2950369119644165, 'train/grad_norm_pre_clip': 27.75455665588379, 'timing/data': 0.03192470036447048, 'timing/model': 0.9393876260146499, 'learning_rate/qwen_vl_interface': 3.6919688628719153e-06, 'learning_rate/action_model': 3.691968862871915e-05, 'epoch': 1.94}) +[2026-06-17 23:22:18,442][starVLA.training.train_starvla][INFO] - Step 4285, Loss: {'train/loss': 0.27502530813217163, 'train/grad_norm_pre_clip': 27.032493591308594, 'timing/data': 0.025028306990861893, 'timing/model': 0.9374648705124855, 'learning_rate/qwen_vl_interface': 3.689925983729311e-06, 'learning_rate/action_model': 3.689925983729311e-05, 'epoch': 1.94}) +[2026-06-17 23:22:22,253][starVLA.training.train_starvla][INFO] - Step 4286, Loss: {'train/loss': 0.2182217836380005, 'train/grad_norm_pre_clip': 23.705686569213867, 'timing/data': 0.02052915282547474, 'timing/model': 0.936784834600985, 'learning_rate/qwen_vl_interface': 3.68788342799173e-06, 'learning_rate/action_model': 3.6878834279917294e-05, 'epoch': 1.94}) +[2026-06-17 23:22:26,123][starVLA.training.train_starvla][INFO] - Step 4287, Loss: {'train/loss': 0.10614990442991257, 'train/grad_norm_pre_clip': 14.233802795410156, 'timing/data': 0.021205231547355652, 'timing/model': 0.9551549162715673, 'learning_rate/qwen_vl_interface': 3.6858411960825953e-06, 'learning_rate/action_model': 3.685841196082595e-05, 'epoch': 1.94}) +[2026-06-17 23:22:30,114][starVLA.training.train_starvla][INFO] - Step 4288, Loss: {'train/loss': 0.1831628680229187, 'train/grad_norm_pre_clip': 23.118751525878906, 'timing/data': 0.018010572530329227, 'timing/model': 0.9656492248177528, 'learning_rate/qwen_vl_interface': 3.6837992884252636e-06, 'learning_rate/action_model': 3.683799288425264e-05, 'epoch': 1.95}) +[2026-06-17 23:22:34,066][starVLA.training.train_starvla][INFO] - Step 4289, Loss: {'train/loss': 0.20963622629642487, 'train/grad_norm_pre_clip': 18.657655715942383, 'timing/data': 0.02956337295472622, 'timing/model': 1.1799568440765142, 'learning_rate/qwen_vl_interface': 3.6817577054430243e-06, 'learning_rate/action_model': 3.6817577054430244e-05, 'epoch': 1.95}) +[2026-06-17 23:22:38,111][starVLA.training.train_starvla][INFO] - Step 4290, Loss: {'train/loss': 0.20775866508483887, 'train/grad_norm_pre_clip': 18.160587310791016, 'timing/data': 0.019080132246017456, 'timing/model': 1.1999110635370016, 'learning_rate/qwen_vl_interface': 3.6797164475591003e-06, 'learning_rate/action_model': 3.6797164475591e-05, 'epoch': 1.95}) +[2026-06-17 23:22:42,062][starVLA.training.train_starvla][INFO] - Step 4291, Loss: {'train/loss': 0.0532766655087471, 'train/grad_norm_pre_clip': 6.917531490325928, 'timing/data': 0.021449985913932323, 'timing/model': 0.9473645593971014, 'learning_rate/qwen_vl_interface': 3.677675515196648e-06, 'learning_rate/action_model': 3.677675515196648e-05, 'epoch': 1.95}) +[2026-06-17 23:22:46,041][starVLA.training.train_starvla][INFO] - Step 4292, Loss: {'train/loss': 0.48774558305740356, 'train/grad_norm_pre_clip': 22.87543296813965, 'timing/data': 0.020450534299016, 'timing/model': 0.9737759893760085, 'learning_rate/qwen_vl_interface': 3.6756349087787524e-06, 'learning_rate/action_model': 3.6756349087787524e-05, 'epoch': 1.95}) +[2026-06-17 23:22:49,964][starVLA.training.train_starvla][INFO] - Step 4293, Loss: {'train/loss': 0.3573880195617676, 'train/grad_norm_pre_clip': 23.962881088256836, 'timing/data': 0.029986483044922352, 'timing/model': 0.9781061913818121, 'learning_rate/qwen_vl_interface': 3.6735946287284325e-06, 'learning_rate/action_model': 3.6735946287284324e-05, 'epoch': 1.95}) +[2026-06-17 23:22:53,927][starVLA.training.train_starvla][INFO] - Step 4294, Loss: {'train/loss': 0.5473222136497498, 'train/grad_norm_pre_clip': 29.573640823364258, 'timing/data': 0.01739179901778698, 'timing/model': 1.0488254176452756, 'learning_rate/qwen_vl_interface': 3.6715546754686444e-06, 'learning_rate/action_model': 3.6715546754686444e-05, 'epoch': 1.95}) +[2026-06-17 23:22:57,755][starVLA.training.train_starvla][INFO] - Step 4295, Loss: {'train/loss': 0.4016304016113281, 'train/grad_norm_pre_clip': 29.6728515625, 'timing/data': 0.01552252285182476, 'timing/model': 0.9414152801036835, 'learning_rate/qwen_vl_interface': 3.6695150494222697e-06, 'learning_rate/action_model': 3.6695150494222695e-05, 'epoch': 1.95}) +[2026-06-17 23:23:01,621][starVLA.training.train_starvla][INFO] - Step 4296, Loss: {'train/loss': 0.3140513002872467, 'train/grad_norm_pre_clip': 33.010032653808594, 'timing/data': 0.020969241857528687, 'timing/model': 1.1497969720512629, 'learning_rate/qwen_vl_interface': 3.6674757510121255e-06, 'learning_rate/action_model': 3.6674757510121255e-05, 'epoch': 1.95}) +[2026-06-17 23:23:05,443][starVLA.training.train_starvla][INFO] - Step 4297, Loss: {'train/loss': 0.2586309313774109, 'train/grad_norm_pre_clip': 21.012353897094727, 'timing/data': 0.014267275109887123, 'timing/model': 1.1382529735565186, 'learning_rate/qwen_vl_interface': 3.6654367806609615e-06, 'learning_rate/action_model': 3.665436780660961e-05, 'epoch': 1.95}) +[2026-06-17 23:23:09,140][starVLA.training.train_starvla][INFO] - Step 4298, Loss: {'train/loss': 0.33470821380615234, 'train/grad_norm_pre_clip': 28.251882553100586, 'timing/data': 0.01736801117658615, 'timing/model': 0.9466891614720225, 'learning_rate/qwen_vl_interface': 3.6633981387914597e-06, 'learning_rate/action_model': 3.66339813879146e-05, 'epoch': 1.95}) +[2026-06-17 23:23:14,763][starVLA.training.train_starvla][INFO] - Step 4299, Loss: {'train/loss': 0.186363086104393, 'train/grad_norm_pre_clip': 23.928483963012695, 'timing/data': 0.016151984222233295, 'timing/model': 0.9467726303264499, 'learning_rate/qwen_vl_interface': 3.6613598258262295e-06, 'learning_rate/action_model': 3.6613598258262296e-05, 'epoch': 1.95}) +[2026-06-17 23:25:23,798][starVLA.training.train_starvla][INFO] - Step 4300, Loss: {'train/loss': 0.36812812089920044, 'train/grad_norm_pre_clip': 31.10795783996582, 'eval/loss': 0.2952694892883301, 'eval/latency_8/loss': 0.2952694892883301, 'eval/demon_fire_precision': 0.9598418010343779, 'eval/demon_fire_recall': 0.9801180490835664, 'eval/demon_fire_f1': 0.9698739624961574, 'eval/demon_fire_rate_model': 0.9145798553144129, 'eval/demon_fire_rate_teacher': 0.8956594323873122, 'eval/demon_left_precision': 0.8805970149253731, 'eval/demon_left_recall': 0.7580299785867237, 'eval/demon_left_f1': 0.8147295742232451, 'eval/demon_left_rate_model': 0.11185308848080133, 'eval/demon_left_rate_teacher': 0.12993878686700056, 'eval/demon_right_precision': 0.8415384615384616, 'eval/demon_right_recall': 0.8287878787878787, 'eval/demon_right_f1': 0.835114503816794, 'eval/demon_right_rate_model': 0.1808569838619922, 'eval/demon_right_rate_teacher': 0.18363939899833054, 'eval/demon_move_f1': 0.8249220390200196, 'eval/demon_cc_f1': 0.9213074504828409, 'eval/latency_8/demon_fire_precision': 0.9598418010343779, 'eval/latency_8/demon_fire_recall': 0.9801180490835664, 'eval/latency_8/demon_fire_f1': 0.9698739624961574, 'eval/latency_8/demon_fire_rate_model': 0.9145798553144129, 'eval/latency_8/demon_fire_rate_teacher': 0.8956594323873122, 'eval/latency_8/demon_left_precision': 0.8805970149253731, 'eval/latency_8/demon_left_recall': 0.7580299785867237, 'eval/latency_8/demon_left_f1': 0.8147295742232451, 'eval/latency_8/demon_left_rate_model': 0.11185308848080133, 'eval/latency_8/demon_left_rate_teacher': 0.12993878686700056, 'eval/latency_8/demon_right_precision': 0.8415384615384616, 'eval/latency_8/demon_right_recall': 0.8287878787878787, 'eval/latency_8/demon_right_f1': 0.835114503816794, 'eval/latency_8/demon_right_rate_model': 0.1808569838619922, 'eval/latency_8/demon_right_rate_teacher': 0.18363939899833054, 'eval/latency_8/demon_move_f1': 0.8249220390200196, 'eval/latency_8/demon_cc_f1': 0.9213074504828409, 'timing/data': 0.018674355931580067, 'timing/model': 1.7548184897750616, 'learning_rate/qwen_vl_interface': 3.6593218421878157e-06, 'learning_rate/action_model': 3.659321842187816e-05, 'epoch': 1.95}) +[2026-06-17 23:25:23,801][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_4300_state +[2026-06-17 23:25:23,806][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 23:25:30,648][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_4300_state/model.safetensors +[2026-06-17 23:25:46,130][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_4300_state/optimizer.bin +[2026-06-17 23:25:46,130][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_4300_state/sampler.bin +[2026-06-17 23:25:46,131][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_4300_state/sampler_1.bin +[2026-06-17 23:25:46,133][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_4300_state/random_states_0.pkl +[2026-06-17 23:25:46,138][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4300 +[2026-06-17 23:25:46,139][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 23:25:46,149][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 23:25:50,091][starVLA.training.train_starvla][INFO] - Step 4301, Loss: {'train/loss': 0.41076964139938354, 'train/grad_norm_pre_clip': 29.829341888427734, 'timing/data': 0.022204574197530746, 'timing/model': 0.9527986440807581, 'learning_rate/qwen_vl_interface': 3.657284188298697e-06, 'learning_rate/action_model': 3.657284188298697e-05, 'epoch': 1.95}) +[2026-06-17 23:25:53,959][starVLA.training.train_starvla][INFO] - Step 4302, Loss: {'train/loss': 0.08490431308746338, 'train/grad_norm_pre_clip': 12.659452438354492, 'timing/data': 0.022473405115306377, 'timing/model': 0.9324665032327175, 'learning_rate/qwen_vl_interface': 3.65524686458128e-06, 'learning_rate/action_model': 3.65524686458128e-05, 'epoch': 1.95}) +[2026-06-17 23:25:57,797][starVLA.training.train_starvla][INFO] - Step 4303, Loss: {'train/loss': 0.27037379145622253, 'train/grad_norm_pre_clip': 16.476730346679688, 'timing/data': 0.020632685162127018, 'timing/model': 0.9503644052892923, 'learning_rate/qwen_vl_interface': 3.653209871457904e-06, 'learning_rate/action_model': 3.653209871457904e-05, 'epoch': 1.95}) +[2026-06-17 23:26:01,674][starVLA.training.train_starvla][INFO] - Step 4304, Loss: {'train/loss': 0.2605487108230591, 'train/grad_norm_pre_clip': 21.457971572875977, 'timing/data': 0.02237823698669672, 'timing/model': 0.9561848435550928, 'learning_rate/qwen_vl_interface': 3.6511732093508376e-06, 'learning_rate/action_model': 3.651173209350838e-05, 'epoch': 1.95}) +[2026-06-17 23:26:05,548][starVLA.training.train_starvla][INFO] - Step 4305, Loss: {'train/loss': 0.19609415531158447, 'train/grad_norm_pre_clip': 21.013385772705078, 'timing/data': 0.018107349053025246, 'timing/model': 1.1299303323030472, 'learning_rate/qwen_vl_interface': 3.6491368786822885e-06, 'learning_rate/action_model': 3.6491368786822884e-05, 'epoch': 1.95}) +[2026-06-17 23:26:09,250][starVLA.training.train_starvla][INFO] - Step 4306, Loss: {'train/loss': 0.062450163066387177, 'train/grad_norm_pre_clip': 10.45799732208252, 'timing/data': 0.028918717056512833, 'timing/model': 0.9604589892551303, 'learning_rate/qwen_vl_interface': 3.647100879874386e-06, 'learning_rate/action_model': 3.6471008798743855e-05, 'epoch': 1.95}) +[2026-06-17 23:26:13,142][starVLA.training.train_starvla][INFO] - Step 4307, Loss: {'train/loss': 0.655467689037323, 'train/grad_norm_pre_clip': 40.64419174194336, 'timing/data': 0.015041852369904518, 'timing/model': 0.9391640825197101, 'learning_rate/qwen_vl_interface': 3.645065213349195e-06, 'learning_rate/action_model': 3.645065213349195e-05, 'epoch': 1.95}) +[2026-06-17 23:26:17,056][starVLA.training.train_starvla][INFO] - Step 4308, Loss: {'train/loss': 0.2908302843570709, 'train/grad_norm_pre_clip': 29.734952926635742, 'timing/data': 0.024194631725549698, 'timing/model': 0.9754977868869901, 'learning_rate/qwen_vl_interface': 3.6430298795287135e-06, 'learning_rate/action_model': 3.643029879528713e-05, 'epoch': 1.95}) +[2026-06-17 23:26:20,932][starVLA.training.train_starvla][INFO] - Step 4309, Loss: {'train/loss': 0.3222559988498688, 'train/grad_norm_pre_clip': 28.415102005004883, 'timing/data': 0.017447980120778084, 'timing/model': 0.9518566643819213, 'learning_rate/qwen_vl_interface': 3.640994878834866e-06, 'learning_rate/action_model': 3.640994878834866e-05, 'epoch': 1.95}) +[2026-06-17 23:26:24,748][starVLA.training.train_starvla][INFO] - Step 4310, Loss: {'train/loss': 0.1038355678319931, 'train/grad_norm_pre_clip': 17.24931526184082, 'timing/data': 0.01709267683327198, 'timing/model': 0.9635085975751281, 'learning_rate/qwen_vl_interface': 3.6389602116895136e-06, 'learning_rate/action_model': 3.6389602116895134e-05, 'epoch': 1.96}) +[2026-06-17 23:26:28,635][starVLA.training.train_starvla][INFO] - Step 4311, Loss: {'train/loss': 0.2604829668998718, 'train/grad_norm_pre_clip': 24.397245407104492, 'timing/data': 0.01903969794511795, 'timing/model': 0.9848343739286065, 'learning_rate/qwen_vl_interface': 3.6369258785144395e-06, 'learning_rate/action_model': 3.636925878514439e-05, 'epoch': 1.96}) +[2026-06-17 23:26:33,323][starVLA.training.train_starvla][INFO] - Step 4312, Loss: {'train/loss': 0.10977774113416672, 'train/grad_norm_pre_clip': 15.38431167602539, 'timing/data': 0.016848250292241573, 'timing/model': 2.014043712988496, 'learning_rate/qwen_vl_interface': 3.6348918797313693e-06, 'learning_rate/action_model': 3.634891879731369e-05, 'epoch': 1.96}) +[2026-06-17 23:26:36,954][starVLA.training.train_starvla][INFO] - Step 4313, Loss: {'train/loss': 0.32595768570899963, 'train/grad_norm_pre_clip': 21.820241928100586, 'timing/data': 0.01793458592146635, 'timing/model': 0.9343407927080989, 'learning_rate/qwen_vl_interface': 3.6328582157619495e-06, 'learning_rate/action_model': 3.6328582157619496e-05, 'epoch': 1.96}) +[2026-06-17 23:26:40,753][starVLA.training.train_starvla][INFO] - Step 4314, Loss: {'train/loss': 0.25838035345077515, 'train/grad_norm_pre_clip': 25.097627639770508, 'timing/data': 0.019210261292755604, 'timing/model': 0.9346978338435292, 'learning_rate/qwen_vl_interface': 3.6308248870277613e-06, 'learning_rate/action_model': 3.630824887027761e-05, 'epoch': 1.96}) +[2026-06-17 23:26:44,630][starVLA.training.train_starvla][INFO] - Step 4315, Loss: {'train/loss': 0.1468796730041504, 'train/grad_norm_pre_clip': 16.294076919555664, 'timing/data': 0.015936847776174545, 'timing/model': 0.9661312773823738, 'learning_rate/qwen_vl_interface': 3.628791893950316e-06, 'learning_rate/action_model': 3.628791893950316e-05, 'epoch': 1.96}) +[2026-06-17 23:26:48,476][starVLA.training.train_starvla][INFO] - Step 4316, Loss: {'train/loss': 0.22120638191699982, 'train/grad_norm_pre_clip': 23.711984634399414, 'timing/data': 0.018873218446969986, 'timing/model': 0.9389252681285143, 'learning_rate/qwen_vl_interface': 3.6267592369510563e-06, 'learning_rate/action_model': 3.6267592369510564e-05, 'epoch': 1.96}) +[2026-06-17 23:26:52,331][starVLA.training.train_starvla][INFO] - Step 4317, Loss: {'train/loss': 0.10753793269395828, 'train/grad_norm_pre_clip': 13.382194519042969, 'timing/data': 0.01636734325438738, 'timing/model': 0.9297893019393086, 'learning_rate/qwen_vl_interface': 3.624726916451355e-06, 'learning_rate/action_model': 3.624726916451355e-05, 'epoch': 1.96}) +[2026-06-17 23:26:56,129][starVLA.training.train_starvla][INFO] - Step 4318, Loss: {'train/loss': 0.34839510917663574, 'train/grad_norm_pre_clip': 26.732986450195312, 'timing/data': 0.015117130242288113, 'timing/model': 1.1226774277165532, 'learning_rate/qwen_vl_interface': 3.622694932872509e-06, 'learning_rate/action_model': 3.622694932872509e-05, 'epoch': 1.96}) +[2026-06-17 23:26:59,951][starVLA.training.train_starvla][INFO] - Step 4319, Loss: {'train/loss': 0.2882659137248993, 'train/grad_norm_pre_clip': 25.187286376953125, 'timing/data': 0.03743718005716801, 'timing/model': 1.132414424791932, 'learning_rate/qwen_vl_interface': 3.620663286635758e-06, 'learning_rate/action_model': 3.620663286635758e-05, 'epoch': 1.96}) +[2026-06-17 23:27:03,621][starVLA.training.train_starvla][INFO] - Step 4320, Loss: {'train/loss': 0.21034343540668488, 'train/grad_norm_pre_clip': 19.47432518005371, 'timing/data': 0.02067219465970993, 'timing/model': 0.9895535316318274, 'learning_rate/qwen_vl_interface': 3.618631978162259e-06, 'learning_rate/action_model': 3.618631978162259e-05, 'epoch': 1.96}) +[2026-06-17 23:27:07,502][starVLA.training.train_starvla][INFO] - Step 4321, Loss: {'train/loss': 0.10951302200555801, 'train/grad_norm_pre_clip': 13.320923805236816, 'timing/data': 0.015909901820123196, 'timing/model': 0.9754377407953143, 'learning_rate/qwen_vl_interface': 3.6166010078731073e-06, 'learning_rate/action_model': 3.6166010078731075e-05, 'epoch': 1.96}) +[2026-06-17 23:27:11,415][starVLA.training.train_starvla][INFO] - Step 4322, Loss: {'train/loss': 0.019774843007326126, 'train/grad_norm_pre_clip': 2.1414718627929688, 'timing/data': 0.01625611912459135, 'timing/model': 0.9635649621486664, 'learning_rate/qwen_vl_interface': 3.614570376189323e-06, 'learning_rate/action_model': 3.614570376189323e-05, 'epoch': 1.96}) +[2026-06-17 23:27:15,395][starVLA.training.train_starvla][INFO] - Step 4323, Loss: {'train/loss': 0.12264691293239594, 'train/grad_norm_pre_clip': 18.795671463012695, 'timing/data': 0.019076500087976456, 'timing/model': 0.992615750990808, 'learning_rate/qwen_vl_interface': 3.6125400835318604e-06, 'learning_rate/action_model': 3.61254008353186e-05, 'epoch': 1.96}) +[2026-06-17 23:27:19,461][starVLA.training.train_starvla][INFO] - Step 4324, Loss: {'train/loss': 0.3186604976654053, 'train/grad_norm_pre_clip': 17.993925094604492, 'timing/data': 0.015757613815367222, 'timing/model': 0.9384460272267461, 'learning_rate/qwen_vl_interface': 3.6105101303216016e-06, 'learning_rate/action_model': 3.610510130321601e-05, 'epoch': 1.96}) +[2026-06-17 23:27:23,407][starVLA.training.train_starvla][INFO] - Step 4325, Loss: {'train/loss': 0.053077250719070435, 'train/grad_norm_pre_clip': 11.097772598266602, 'timing/data': 0.015349062159657478, 'timing/model': 1.2797261448577046, 'learning_rate/qwen_vl_interface': 3.608480516979354e-06, 'learning_rate/action_model': 3.608480516979354e-05, 'epoch': 1.96}) +[2026-06-17 23:27:28,828][starVLA.training.train_starvla][INFO] - Step 4326, Loss: {'train/loss': 0.29605504870414734, 'train/grad_norm_pre_clip': 15.351348876953125, 'timing/data': 0.018975337035954, 'timing/model': 2.5607832400128245, 'learning_rate/qwen_vl_interface': 3.6064512439258646e-06, 'learning_rate/action_model': 3.606451243925865e-05, 'epoch': 1.96}) +[2026-06-17 23:27:36,314][starVLA.training.train_starvla][INFO] - Step 4327, Loss: {'train/loss': 0.24627292156219482, 'train/grad_norm_pre_clip': 17.713409423828125, 'timing/data': 0.015218431130051613, 'timing/model': 3.0220339922234416, 'learning_rate/qwen_vl_interface': 3.6044223115818e-06, 'learning_rate/action_model': 3.6044223115818e-05, 'epoch': 1.96}) +[2026-06-17 23:27:41,542][starVLA.training.train_starvla][INFO] - Step 4328, Loss: {'train/loss': 0.11905449628829956, 'train/grad_norm_pre_clip': 14.11971378326416, 'timing/data': 0.01512610912322998, 'timing/model': 0.9309538472443819, 'learning_rate/qwen_vl_interface': 3.6023937203677606e-06, 'learning_rate/action_model': 3.6023937203677604e-05, 'epoch': 1.96}) +[2026-06-17 23:27:45,337][starVLA.training.train_starvla][INFO] - Step 4329, Loss: {'train/loss': 0.1310872584581375, 'train/grad_norm_pre_clip': 15.856535911560059, 'timing/data': 0.020763125270605087, 'timing/model': 0.9449465917423368, 'learning_rate/qwen_vl_interface': 3.6003654707042746e-06, 'learning_rate/action_model': 3.6003654707042745e-05, 'epoch': 1.96}) +[2026-06-17 23:27:49,235][starVLA.training.train_starvla][INFO] - Step 4330, Loss: {'train/loss': 0.409549742937088, 'train/grad_norm_pre_clip': 25.580522537231445, 'timing/data': 0.019722284749150276, 'timing/model': 0.9793359441682696, 'learning_rate/qwen_vl_interface': 3.5983375630118033e-06, 'learning_rate/action_model': 3.598337563011803e-05, 'epoch': 1.96}) +[2026-06-17 23:27:55,421][starVLA.training.train_starvla][INFO] - Step 4331, Loss: {'train/loss': 0.019152812659740448, 'train/grad_norm_pre_clip': 3.9694266319274902, 'timing/data': 0.01875413116067648, 'timing/model': 1.2272098576650023, 'learning_rate/qwen_vl_interface': 3.596309997710733e-06, 'learning_rate/action_model': 3.596309997710733e-05, 'epoch': 1.96}) +[2026-06-17 23:27:59,471][starVLA.training.train_starvla][INFO] - Step 4332, Loss: {'train/loss': 0.5746273398399353, 'train/grad_norm_pre_clip': 23.721839904785156, 'timing/data': 0.01783787738531828, 'timing/model': 1.2045665681362152, 'learning_rate/qwen_vl_interface': 3.5942827752213788e-06, 'learning_rate/action_model': 3.594282775221379e-05, 'epoch': 1.97}) +[2026-06-17 23:28:07,460][starVLA.training.train_starvla][INFO] - Step 4333, Loss: {'train/loss': 0.1524791419506073, 'train/grad_norm_pre_clip': 18.82522964477539, 'timing/data': 0.017849686555564404, 'timing/model': 0.9610023098066449, 'learning_rate/qwen_vl_interface': 3.592255895963986e-06, 'learning_rate/action_model': 3.592255895963986e-05, 'epoch': 1.97}) +[2026-06-17 23:28:11,451][starVLA.training.train_starvla][INFO] - Step 4334, Loss: {'train/loss': 0.16944342851638794, 'train/grad_norm_pre_clip': 26.621204376220703, 'timing/data': 0.017315425910055637, 'timing/model': 0.9675392555072904, 'learning_rate/qwen_vl_interface': 3.590229360358731e-06, 'learning_rate/action_model': 3.590229360358731e-05, 'epoch': 1.97}) +[2026-06-17 23:28:15,353][starVLA.training.train_starvla][INFO] - Step 4335, Loss: {'train/loss': 0.3878576457500458, 'train/grad_norm_pre_clip': 190.89358520507812, 'timing/data': 0.019766654819250107, 'timing/model': 0.9563406175002456, 'learning_rate/qwen_vl_interface': 3.5882031688257147e-06, 'learning_rate/action_model': 3.588203168825715e-05, 'epoch': 1.97}) +[2026-06-17 23:28:19,800][starVLA.training.train_starvla][INFO] - Step 4336, Loss: {'train/loss': 0.38490810990333557, 'train/grad_norm_pre_clip': 31.898435592651367, 'timing/data': 0.017998725175857544, 'timing/model': 0.9694666862487793, 'learning_rate/qwen_vl_interface': 3.5861773217849694e-06, 'learning_rate/action_model': 3.5861773217849693e-05, 'epoch': 1.97}) +[2026-06-17 23:28:23,657][starVLA.training.train_starvla][INFO] - Step 4337, Loss: {'train/loss': 0.19323094189167023, 'train/grad_norm_pre_clip': 28.44171714782715, 'timing/data': 0.01899829600006342, 'timing/model': 0.9426839593797922, 'learning_rate/qwen_vl_interface': 3.5841518196564566e-06, 'learning_rate/action_model': 3.584151819656457e-05, 'epoch': 1.97}) +[2026-06-17 23:28:27,543][starVLA.training.train_starvla][INFO] - Step 4338, Loss: {'train/loss': 0.08409811556339264, 'train/grad_norm_pre_clip': 15.814229965209961, 'timing/data': 0.022258958779275417, 'timing/model': 1.1684340797364712, 'learning_rate/qwen_vl_interface': 3.582126662860065e-06, 'learning_rate/action_model': 3.582126662860065e-05, 'epoch': 1.97}) +[2026-06-17 23:28:31,917][starVLA.training.train_starvla][INFO] - Step 4339, Loss: {'train/loss': 0.20366401970386505, 'train/grad_norm_pre_clip': 24.839826583862305, 'timing/data': 0.02129021566361189, 'timing/model': 1.7000743318349123, 'learning_rate/qwen_vl_interface': 3.5801018518156096e-06, 'learning_rate/action_model': 3.5801018518156096e-05, 'epoch': 1.97}) +[2026-06-17 23:28:36,790][starVLA.training.train_starvla][INFO] - Step 4340, Loss: {'train/loss': 0.2276919037103653, 'train/grad_norm_pre_clip': 47.02529525756836, 'timing/data': 0.019061836414039135, 'timing/model': 0.9431219231337309, 'learning_rate/qwen_vl_interface': 3.5780773869428366e-06, 'learning_rate/action_model': 3.5780773869428364e-05, 'epoch': 1.97}) +[2026-06-17 23:28:40,672][starVLA.training.train_starvla][INFO] - Step 4341, Loss: {'train/loss': 0.14472997188568115, 'train/grad_norm_pre_clip': 20.38626480102539, 'timing/data': 0.01886565238237381, 'timing/model': 0.9412133283913136, 'learning_rate/qwen_vl_interface': 3.576053268661422e-06, 'learning_rate/action_model': 3.5760532686614216e-05, 'epoch': 1.97}) +[2026-06-17 23:28:44,645][starVLA.training.train_starvla][INFO] - Step 4342, Loss: {'train/loss': 0.4321226179599762, 'train/grad_norm_pre_clip': 26.082338333129883, 'timing/data': 0.025980334728956223, 'timing/model': 0.9552577026188374, 'learning_rate/qwen_vl_interface': 3.5740294973909643e-06, 'learning_rate/action_model': 3.5740294973909644e-05, 'epoch': 1.97}) +[2026-06-17 23:28:48,804][starVLA.training.train_starvla][INFO] - Step 4343, Loss: {'train/loss': 0.21649618446826935, 'train/grad_norm_pre_clip': 30.387388229370117, 'timing/data': 0.017924923449754715, 'timing/model': 1.231331559829414, 'learning_rate/qwen_vl_interface': 3.5720060735509976e-06, 'learning_rate/action_model': 3.5720060735509975e-05, 'epoch': 1.97}) +[2026-06-17 23:28:53,382][starVLA.training.train_starvla][INFO] - Step 4344, Loss: {'train/loss': 0.33109670877456665, 'train/grad_norm_pre_clip': 22.141477584838867, 'timing/data': 0.024616685695946217, 'timing/model': 1.6920561641454697, 'learning_rate/qwen_vl_interface': 3.569982997560974e-06, 'learning_rate/action_model': 3.5699829975609736e-05, 'epoch': 1.97}) +[2026-06-17 23:28:57,346][starVLA.training.train_starvla][INFO] - Step 4345, Loss: {'train/loss': 0.22139592468738556, 'train/grad_norm_pre_clip': 37.068302154541016, 'timing/data': 0.018651621416211128, 'timing/model': 1.1929518394172192, 'learning_rate/qwen_vl_interface': 3.567960269840286e-06, 'learning_rate/action_model': 3.567960269840286e-05, 'epoch': 1.97}) +[2026-06-17 23:29:01,050][starVLA.training.train_starvla][INFO] - Step 4346, Loss: {'train/loss': 0.2459101527929306, 'train/grad_norm_pre_clip': 17.589244842529297, 'timing/data': 0.019724528305232525, 'timing/model': 0.9569442290812731, 'learning_rate/qwen_vl_interface': 3.5659378908082427e-06, 'learning_rate/action_model': 3.5659378908082426e-05, 'epoch': 1.97}) +[2026-06-17 23:29:04,970][starVLA.training.train_starvla][INFO] - Step 4347, Loss: {'train/loss': 0.12563030421733856, 'train/grad_norm_pre_clip': 20.28307342529297, 'timing/data': 0.01855885051190853, 'timing/model': 0.9583210786804557, 'learning_rate/qwen_vl_interface': 3.5639158608840862e-06, 'learning_rate/action_model': 3.563915860884086e-05, 'epoch': 1.97}) +[2026-06-17 23:29:08,861][starVLA.training.train_starvla][INFO] - Step 4348, Loss: {'train/loss': 0.3851545751094818, 'train/grad_norm_pre_clip': 23.814857482910156, 'timing/data': 0.01741529256105423, 'timing/model': 0.9526022328063846, 'learning_rate/qwen_vl_interface': 3.5618941804869866e-06, 'learning_rate/action_model': 3.561894180486987e-05, 'epoch': 1.97}) +[2026-06-17 23:29:12,884][starVLA.training.train_starvla][INFO] - Step 4349, Loss: {'train/loss': 0.21738296747207642, 'train/grad_norm_pre_clip': 22.91712760925293, 'timing/data': 0.01655527390539646, 'timing/model': 0.9494881881400943, 'learning_rate/qwen_vl_interface': 3.5598728500360395e-06, 'learning_rate/action_model': 3.55987285003604e-05, 'epoch': 1.97}) +[2026-06-17 23:29:16,740][starVLA.training.train_starvla][INFO] - Step 4350, Loss: {'train/loss': 0.40607544779777527, 'train/grad_norm_pre_clip': 22.341594696044922, 'timing/data': 0.01654668990522623, 'timing/model': 0.9444856094196439, 'learning_rate/qwen_vl_interface': 3.557851869950272e-06, 'learning_rate/action_model': 3.557851869950272e-05, 'epoch': 1.97}) +[2026-06-17 23:29:20,614][starVLA.training.train_starvla][INFO] - Step 4351, Loss: {'train/loss': 0.34474343061447144, 'train/grad_norm_pre_clip': 26.695627212524414, 'timing/data': 0.018518432043492794, 'timing/model': 1.167043887078762, 'learning_rate/qwen_vl_interface': 3.555831240648628e-06, 'learning_rate/action_model': 3.5558312406486274e-05, 'epoch': 1.97}) +[2026-06-17 23:29:24,706][starVLA.training.train_starvla][INFO] - Step 4352, Loss: {'train/loss': 0.29149264097213745, 'train/grad_norm_pre_clip': 22.915557861328125, 'timing/data': 0.016827167011797428, 'timing/model': 1.388846293091774, 'learning_rate/qwen_vl_interface': 3.553810962549994e-06, 'learning_rate/action_model': 3.553810962549994e-05, 'epoch': 1.97}) +[2026-06-17 23:29:28,344][starVLA.training.train_starvla][INFO] - Step 4353, Loss: {'train/loss': 0.08180306106805801, 'train/grad_norm_pre_clip': 12.055059432983398, 'timing/data': 0.01637046691030264, 'timing/model': 0.9367862697690725, 'learning_rate/qwen_vl_interface': 3.5517910360731723e-06, 'learning_rate/action_model': 3.551791036073172e-05, 'epoch': 1.97}) +[2026-06-17 23:29:32,141][starVLA.training.train_starvla][INFO] - Step 4354, Loss: {'train/loss': 0.17742793262004852, 'train/grad_norm_pre_clip': 26.634305953979492, 'timing/data': 0.0172354718670249, 'timing/model': 0.9422953510656953, 'learning_rate/qwen_vl_interface': 3.549771461636896e-06, 'learning_rate/action_model': 3.549771461636896e-05, 'epoch': 1.98}) +[2026-06-17 23:29:35,979][starVLA.training.train_starvla][INFO] - Step 4355, Loss: {'train/loss': 0.25273987650871277, 'train/grad_norm_pre_clip': 30.81739616394043, 'timing/data': 0.018932802602648735, 'timing/model': 0.9495606748387218, 'learning_rate/qwen_vl_interface': 3.5477522396598236e-06, 'learning_rate/action_model': 3.5477522396598235e-05, 'epoch': 1.98}) +[2026-06-17 23:29:39,959][starVLA.training.train_starvla][INFO] - Step 4356, Loss: {'train/loss': 0.3271322548389435, 'train/grad_norm_pre_clip': 27.107330322265625, 'timing/data': 0.018035144545137882, 'timing/model': 0.9559674253687263, 'learning_rate/qwen_vl_interface': 3.5457333705605456e-06, 'learning_rate/action_model': 3.5457333705605456e-05, 'epoch': 1.98}) +[2026-06-17 23:29:44,527][starVLA.training.train_starvla][INFO] - Step 4357, Loss: {'train/loss': 0.18056830763816833, 'train/grad_norm_pre_clip': 26.961585998535156, 'timing/data': 0.02180854231119156, 'timing/model': 1.5963939083740115, 'learning_rate/qwen_vl_interface': 3.543714854757574e-06, 'learning_rate/action_model': 3.5437148547575736e-05, 'epoch': 1.98}) +[2026-06-17 23:29:48,533][starVLA.training.train_starvla][INFO] - Step 4358, Loss: {'train/loss': 0.18017855286598206, 'train/grad_norm_pre_clip': 21.69998550415039, 'timing/data': 0.017409631051123142, 'timing/model': 1.2179239746183157, 'learning_rate/qwen_vl_interface': 3.5416966926693464e-06, 'learning_rate/action_model': 3.5416966926693465e-05, 'epoch': 1.98}) +[2026-06-17 23:29:52,335][starVLA.training.train_starvla][INFO] - Step 4359, Loss: {'train/loss': 0.35420963168144226, 'train/grad_norm_pre_clip': 22.69628143310547, 'timing/data': 0.020557851530611515, 'timing/model': 1.103268014267087, 'learning_rate/qwen_vl_interface': 3.539678884714235e-06, 'learning_rate/action_model': 3.539678884714235e-05, 'epoch': 1.98}) +[2026-06-17 23:29:56,342][starVLA.training.train_starvla][INFO] - Step 4360, Loss: {'train/loss': 0.3350995182991028, 'train/grad_norm_pre_clip': 25.7425594329834, 'timing/data': 0.025137807242572308, 'timing/model': 0.9600283661857247, 'learning_rate/qwen_vl_interface': 3.5376614313105307e-06, 'learning_rate/action_model': 3.53766143131053e-05, 'epoch': 1.98}) +[2026-06-17 23:30:00,145][starVLA.training.train_starvla][INFO] - Step 4361, Loss: {'train/loss': 0.15414126217365265, 'train/grad_norm_pre_clip': 44.493900299072266, 'timing/data': 0.022856581024825573, 'timing/model': 0.9464438362047076, 'learning_rate/qwen_vl_interface': 3.5356443328764546e-06, 'learning_rate/action_model': 3.5356443328764546e-05, 'epoch': 1.98}) +[2026-06-17 23:30:04,049][starVLA.training.train_starvla][INFO] - Step 4362, Loss: {'train/loss': 0.3355092406272888, 'train/grad_norm_pre_clip': 25.36354637145996, 'timing/data': 0.017188172787427902, 'timing/model': 0.9461459368467331, 'learning_rate/qwen_vl_interface': 3.5336275898301515e-06, 'learning_rate/action_model': 3.5336275898301514e-05, 'epoch': 1.98}) +[2026-06-17 23:30:07,852][starVLA.training.train_starvla][INFO] - Step 4363, Loss: {'train/loss': 0.31688636541366577, 'train/grad_norm_pre_clip': 30.343061447143555, 'timing/data': 0.01783180981874466, 'timing/model': 0.9332540091127157, 'learning_rate/qwen_vl_interface': 3.531611202589697e-06, 'learning_rate/action_model': 3.531611202589697e-05, 'epoch': 1.98}) +[2026-06-17 23:30:11,804][starVLA.training.train_starvla][INFO] - Step 4364, Loss: {'train/loss': 0.5659310817718506, 'train/grad_norm_pre_clip': 31.059656143188477, 'timing/data': 0.01900332886725664, 'timing/model': 0.94486996345222, 'learning_rate/qwen_vl_interface': 3.5295951715730902e-06, 'learning_rate/action_model': 3.52959517157309e-05, 'epoch': 1.98}) +[2026-06-17 23:30:15,821][starVLA.training.train_starvla][INFO] - Step 4365, Loss: {'train/loss': 0.13587510585784912, 'train/grad_norm_pre_clip': 24.339435577392578, 'timing/data': 0.027371114119887352, 'timing/model': 1.2419846234843135, 'learning_rate/qwen_vl_interface': 3.527579497198255e-06, 'learning_rate/action_model': 3.527579497198255e-05, 'epoch': 1.98}) +[2026-06-17 23:30:19,562][starVLA.training.train_starvla][INFO] - Step 4366, Loss: {'train/loss': 0.51121586561203, 'train/grad_norm_pre_clip': 27.730297088623047, 'timing/data': 0.02832919266074896, 'timing/model': 0.9854465145617723, 'learning_rate/qwen_vl_interface': 3.525564179883042e-06, 'learning_rate/action_model': 3.525564179883042e-05, 'epoch': 1.98}) +[2026-06-17 23:30:23,580][starVLA.training.train_starvla][INFO] - Step 4367, Loss: {'train/loss': 0.14773891866207123, 'train/grad_norm_pre_clip': 10.504622459411621, 'timing/data': 0.02243656199425459, 'timing/model': 0.955968682654202, 'learning_rate/qwen_vl_interface': 3.5235492200452308e-06, 'learning_rate/action_model': 3.523549220045231e-05, 'epoch': 1.98}) +[2026-06-17 23:30:27,542][starVLA.training.train_starvla][INFO] - Step 4368, Loss: {'train/loss': 0.08383098244667053, 'train/grad_norm_pre_clip': 16.9438419342041, 'timing/data': 0.014330345205962658, 'timing/model': 0.9727460481226444, 'learning_rate/qwen_vl_interface': 3.521534618102525e-06, 'learning_rate/action_model': 3.5215346181025246e-05, 'epoch': 1.98}) +[2026-06-17 23:30:31,502][starVLA.training.train_starvla][INFO] - Step 4369, Loss: {'train/loss': 0.5836532711982727, 'train/grad_norm_pre_clip': 30.89308738708496, 'timing/data': 0.02041560970246792, 'timing/model': 0.9626428410410881, 'learning_rate/qwen_vl_interface': 3.5195203744725516e-06, 'learning_rate/action_model': 3.5195203744725516e-05, 'epoch': 1.98}) +[2026-06-17 23:30:35,929][starVLA.training.train_starvla][INFO] - Step 4370, Loss: {'train/loss': 0.12708748877048492, 'train/grad_norm_pre_clip': 18.66089630126953, 'timing/data': 0.015311338007450104, 'timing/model': 0.956574210897088, 'learning_rate/qwen_vl_interface': 3.5175064895728665e-06, 'learning_rate/action_model': 3.5175064895728664e-05, 'epoch': 1.98}) +[2026-06-17 23:30:39,802][starVLA.training.train_starvla][INFO] - Step 4371, Loss: {'train/loss': 0.44445228576660156, 'train/grad_norm_pre_clip': 27.756500244140625, 'timing/data': 0.009251024574041367, 'timing/model': 1.1153731448575854, 'learning_rate/qwen_vl_interface': 3.515492963820951e-06, 'learning_rate/action_model': 3.515492963820951e-05, 'epoch': 1.98}) +[2026-06-17 23:30:43,615][starVLA.training.train_starvla][INFO] - Step 4372, Loss: {'train/loss': 0.40383049845695496, 'train/grad_norm_pre_clip': 32.68964385986328, 'timing/data': 0.009815046563744545, 'timing/model': 1.117809240706265, 'learning_rate/qwen_vl_interface': 3.5134797976342123e-06, 'learning_rate/action_model': 3.513479797634212e-05, 'epoch': 1.98}) +[2026-06-17 23:30:47,327][starVLA.training.train_starvla][INFO] - Step 4373, Loss: {'train/loss': 0.10232317447662354, 'train/grad_norm_pre_clip': 13.505664825439453, 'timing/data': 0.01020803488790989, 'timing/model': 0.9720504097640514, 'learning_rate/qwen_vl_interface': 3.5114669914299776e-06, 'learning_rate/action_model': 3.511466991429978e-05, 'epoch': 1.98}) +[2026-06-17 23:30:51,510][starVLA.training.train_starvla][INFO] - Step 4374, Loss: {'train/loss': 0.3757491409778595, 'train/grad_norm_pre_clip': 28.76609992980957, 'timing/data': 0.018148580566048622, 'timing/model': 1.0341968862339854, 'learning_rate/qwen_vl_interface': 3.5094545456255065e-06, 'learning_rate/action_model': 3.509454545625506e-05, 'epoch': 1.98}) +[2026-06-17 23:30:55,534][starVLA.training.train_starvla][INFO] - Step 4375, Loss: {'train/loss': 0.1961967796087265, 'train/grad_norm_pre_clip': 18.76470375061035, 'timing/data': 0.013402247801423073, 'timing/model': 0.9473273223266006, 'learning_rate/qwen_vl_interface': 3.5074424606379826e-06, 'learning_rate/action_model': 3.5074424606379824e-05, 'epoch': 1.98}) +[2026-06-17 23:30:59,471][starVLA.training.train_starvla][INFO] - Step 4376, Loss: {'train/loss': 0.29933032393455505, 'train/grad_norm_pre_clip': 30.49391746520996, 'timing/data': 0.015355242416262627, 'timing/model': 0.9508008668199182, 'learning_rate/qwen_vl_interface': 3.505430736884511e-06, 'learning_rate/action_model': 3.5054307368845105e-05, 'epoch': 1.99}) +[2026-06-17 23:31:03,372][starVLA.training.train_starvla][INFO] - Step 4377, Loss: {'train/loss': 0.3162234425544739, 'train/grad_norm_pre_clip': 16.532094955444336, 'timing/data': 0.012166534550487995, 'timing/model': 0.986174238845706, 'learning_rate/qwen_vl_interface': 3.503419374782123e-06, 'learning_rate/action_model': 3.503419374782123e-05, 'epoch': 1.99}) +[2026-06-17 23:31:07,423][starVLA.training.train_starvla][INFO] - Step 4378, Loss: {'train/loss': 0.23051650822162628, 'train/grad_norm_pre_clip': 27.348398208618164, 'timing/data': 0.008573571220040321, 'timing/model': 1.2424368252977729, 'learning_rate/qwen_vl_interface': 3.50140837474778e-06, 'learning_rate/action_model': 3.50140837474778e-05, 'epoch': 1.99}) +[2026-06-17 23:31:11,241][starVLA.training.train_starvla][INFO] - Step 4379, Loss: {'train/loss': 0.20006927847862244, 'train/grad_norm_pre_clip': 20.313373565673828, 'timing/data': 0.008275588974356651, 'timing/model': 1.1130662430077791, 'learning_rate/qwen_vl_interface': 3.4993977371983625e-06, 'learning_rate/action_model': 3.499397737198362e-05, 'epoch': 1.99}) +[2026-06-17 23:31:14,924][starVLA.training.train_starvla][INFO] - Step 4380, Loss: {'train/loss': 0.3751581013202667, 'train/grad_norm_pre_clip': 47.41130065917969, 'timing/data': 0.01667890790849924, 'timing/model': 0.9614595286548138, 'learning_rate/qwen_vl_interface': 3.497387462550676e-06, 'learning_rate/action_model': 3.4973874625506757e-05, 'epoch': 1.99}) +[2026-06-17 23:31:18,776][starVLA.training.train_starvla][INFO] - Step 4381, Loss: {'train/loss': 0.2584487199783325, 'train/grad_norm_pre_clip': 24.220998764038086, 'timing/data': 0.014184211380779743, 'timing/model': 0.9887368101626635, 'learning_rate/qwen_vl_interface': 3.4953775512214568e-06, 'learning_rate/action_model': 3.495377551221457e-05, 'epoch': 1.99}) +[2026-06-17 23:31:22,703][starVLA.training.train_starvla][INFO] - Step 4382, Loss: {'train/loss': 0.42396122217178345, 'train/grad_norm_pre_clip': 27.5085506439209, 'timing/data': 0.008350448682904243, 'timing/model': 1.0313603011891246, 'learning_rate/qwen_vl_interface': 3.4933680036273583e-06, 'learning_rate/action_model': 3.493368003627358e-05, 'epoch': 1.99}) +[2026-06-17 23:31:26,534][starVLA.training.train_starvla][INFO] - Step 4383, Loss: {'train/loss': 0.18704615533351898, 'train/grad_norm_pre_clip': 20.029335021972656, 'timing/data': 0.00973530299961567, 'timing/model': 0.9445520034059882, 'learning_rate/qwen_vl_interface': 3.491358820184963e-06, 'learning_rate/action_model': 3.491358820184963e-05, 'epoch': 1.99}) +[2026-06-17 23:31:30,383][starVLA.training.train_starvla][INFO] - Step 4384, Loss: {'train/loss': 0.3639909029006958, 'train/grad_norm_pre_clip': 25.597721099853516, 'timing/data': 0.00835993979126215, 'timing/model': 0.9438788220286369, 'learning_rate/qwen_vl_interface': 3.4893500013107758e-06, 'learning_rate/action_model': 3.4893500013107754e-05, 'epoch': 1.99}) +[2026-06-17 23:31:34,223][starVLA.training.train_starvla][INFO] - Step 4385, Loss: {'train/loss': 0.06664537638425827, 'train/grad_norm_pre_clip': 11.166147232055664, 'timing/data': 0.009643771685659885, 'timing/model': 1.1344789871945977, 'learning_rate/qwen_vl_interface': 3.4873415474212287e-06, 'learning_rate/action_model': 3.487341547421228e-05, 'epoch': 1.99}) +[2026-06-17 23:31:37,912][starVLA.training.train_starvla][INFO] - Step 4386, Loss: {'train/loss': 0.38432472944259644, 'train/grad_norm_pre_clip': 23.896549224853516, 'timing/data': 0.009595226496458054, 'timing/model': 0.9506457364186645, 'learning_rate/qwen_vl_interface': 3.4853334589326764e-06, 'learning_rate/action_model': 3.4853334589326765e-05, 'epoch': 1.99}) +[2026-06-17 23:31:41,774][starVLA.training.train_starvla][INFO] - Step 4387, Loss: {'train/loss': 0.3441120982170105, 'train/grad_norm_pre_clip': 28.29671287536621, 'timing/data': 0.009691624902188778, 'timing/model': 0.9821569602936506, 'learning_rate/qwen_vl_interface': 3.483325736261394e-06, 'learning_rate/action_model': 3.483325736261394e-05, 'epoch': 1.99}) +[2026-06-17 23:31:45,660][starVLA.training.train_starvla][INFO] - Step 4388, Loss: {'train/loss': 0.3194991946220398, 'train/grad_norm_pre_clip': 19.004972457885742, 'timing/data': 0.013274215161800385, 'timing/model': 0.9661991968750954, 'learning_rate/qwen_vl_interface': 3.48131837982359e-06, 'learning_rate/action_model': 3.48131837982359e-05, 'epoch': 1.99}) +[2026-06-17 23:31:49,735][starVLA.training.train_starvla][INFO] - Step 4389, Loss: {'train/loss': 0.2297995239496231, 'train/grad_norm_pre_clip': 25.14854621887207, 'timing/data': 0.011859199032187462, 'timing/model': 0.9649958880618215, 'learning_rate/qwen_vl_interface': 3.4793113900353885e-06, 'learning_rate/action_model': 3.479311390035389e-05, 'epoch': 1.99}) +[2026-06-17 23:31:53,608][starVLA.training.train_starvla][INFO] - Step 4390, Loss: {'train/loss': 0.15984943509101868, 'train/grad_norm_pre_clip': 18.664026260375977, 'timing/data': 0.009304801933467388, 'timing/model': 0.943289521150291, 'learning_rate/qwen_vl_interface': 3.4773047673128412e-06, 'learning_rate/action_model': 3.477304767312841e-05, 'epoch': 1.99}) +[2026-06-17 23:31:57,507][starVLA.training.train_starvla][INFO] - Step 4391, Loss: {'train/loss': 0.21838968992233276, 'train/grad_norm_pre_clip': 20.780439376831055, 'timing/data': 0.007071373984217644, 'timing/model': 0.9934601644054055, 'learning_rate/qwen_vl_interface': 3.4752985120719225e-06, 'learning_rate/action_model': 3.475298512071922e-05, 'epoch': 1.99}) +[2026-06-17 23:32:01,458][starVLA.training.train_starvla][INFO] - Step 4392, Loss: {'train/loss': 0.49388331174850464, 'train/grad_norm_pre_clip': 48.55086135864258, 'timing/data': 0.009936179034411907, 'timing/model': 1.1907546892762184, 'learning_rate/qwen_vl_interface': 3.4732926247285324e-06, 'learning_rate/action_model': 3.4732926247285325e-05, 'epoch': 1.99}) +[2026-06-17 23:32:05,141][starVLA.training.train_starvla][INFO] - Step 4393, Loss: {'train/loss': 0.31871509552001953, 'train/grad_norm_pre_clip': 20.00262451171875, 'timing/data': 0.008179305121302605, 'timing/model': 0.947440036572516, 'learning_rate/qwen_vl_interface': 3.471287105698494e-06, 'learning_rate/action_model': 3.471287105698494e-05, 'epoch': 1.99}) +[2026-06-17 23:32:09,037][starVLA.training.train_starvla][INFO] - Step 4394, Loss: {'train/loss': 0.12844258546829224, 'train/grad_norm_pre_clip': 14.35400104522705, 'timing/data': 0.010861788876354694, 'timing/model': 0.9487183000892401, 'learning_rate/qwen_vl_interface': 3.4692819553975514e-06, 'learning_rate/action_model': 3.469281955397551e-05, 'epoch': 1.99}) +[2026-06-17 23:32:12,907][starVLA.training.train_starvla][INFO] - Step 4395, Loss: {'train/loss': 0.2576161324977875, 'train/grad_norm_pre_clip': 26.470447540283203, 'timing/data': 0.014978798106312752, 'timing/model': 0.9586143456399441, 'learning_rate/qwen_vl_interface': 3.4672771742413754e-06, 'learning_rate/action_model': 3.467277174241375e-05, 'epoch': 1.99}) +[2026-06-17 23:32:16,733][starVLA.training.train_starvla][INFO] - Step 4396, Loss: {'train/loss': 0.32491591572761536, 'train/grad_norm_pre_clip': 47.66020965576172, 'timing/data': 0.00834018737077713, 'timing/model': 0.9379143035039306, 'learning_rate/qwen_vl_interface': 3.46527276264556e-06, 'learning_rate/action_model': 3.46527276264556e-05, 'epoch': 1.99}) +[2026-06-17 23:32:20,564][starVLA.training.train_starvla][INFO] - Step 4397, Loss: {'train/loss': 0.3427407741546631, 'train/grad_norm_pre_clip': 26.609407424926758, 'timing/data': 0.008923543617129326, 'timing/model': 0.9316803244873881, 'learning_rate/qwen_vl_interface': 3.4632687210256218e-06, 'learning_rate/action_model': 3.463268721025621e-05, 'epoch': 1.99}) +[2026-06-17 23:32:24,699][starVLA.training.train_starvla][INFO] - Step 4398, Loss: {'train/loss': 0.11865972727537155, 'train/grad_norm_pre_clip': 20.459836959838867, 'timing/data': 0.010283246636390686, 'timing/model': 1.1941663166508079, 'learning_rate/qwen_vl_interface': 3.461265049796999e-06, 'learning_rate/action_model': 3.461265049796999e-05, 'epoch': 2.0}) +[2026-06-17 23:32:28,978][starVLA.training.train_starvla][INFO] - Step 4399, Loss: {'train/loss': 0.09313114732503891, 'train/grad_norm_pre_clip': 12.535874366760254, 'timing/data': 0.013621555641293526, 'timing/model': 1.2847713930532336, 'learning_rate/qwen_vl_interface': 3.459261749375057e-06, 'learning_rate/action_model': 3.4592617493750567e-05, 'epoch': 2.0}) +[2026-06-17 23:34:44,067][starVLA.training.train_starvla][INFO] - Step 4400, Loss: {'train/loss': 0.36438268423080444, 'train/grad_norm_pre_clip': 14.346116065979004, 'eval/loss': 0.288151353597641, 'eval/latency_8/loss': 0.288151353597641, 'eval/demon_fire_precision': 0.968440594059406, 'eval/demon_fire_recall': 0.9723516620068344, 'eval/demon_fire_f1': 0.9703921872577895, 'eval/demon_fire_rate_model': 0.899276572064552, 'eval/demon_fire_rate_teacher': 0.8956594323873122, 'eval/demon_left_precision': 0.8899755501222494, 'eval/demon_left_recall': 0.7794432548179872, 'eval/demon_left_f1': 0.8310502283105022, 'eval/demon_left_rate_model': 0.11380077907623817, 'eval/demon_left_rate_teacher': 0.12993878686700056, 'eval/demon_right_precision': 0.848531684698609, 'eval/demon_right_recall': 0.8318181818181818, 'eval/demon_right_f1': 0.8400918133129303, 'eval/demon_right_rate_model': 0.1800222593210907, 'eval/demon_right_rate_teacher': 0.18363939899833054, 'eval/demon_move_f1': 0.8355710208117162, 'eval/demon_cc_f1': 0.9255884843980249, 'eval/latency_8/demon_fire_precision': 0.968440594059406, 'eval/latency_8/demon_fire_recall': 0.9723516620068344, 'eval/latency_8/demon_fire_f1': 0.9703921872577895, 'eval/latency_8/demon_fire_rate_model': 0.899276572064552, 'eval/latency_8/demon_fire_rate_teacher': 0.8956594323873122, 'eval/latency_8/demon_left_precision': 0.8899755501222494, 'eval/latency_8/demon_left_recall': 0.7794432548179872, 'eval/latency_8/demon_left_f1': 0.8310502283105022, 'eval/latency_8/demon_left_rate_model': 0.11380077907623817, 'eval/latency_8/demon_left_rate_teacher': 0.12993878686700056, 'eval/latency_8/demon_right_precision': 0.848531684698609, 'eval/latency_8/demon_right_recall': 0.8318181818181818, 'eval/latency_8/demon_right_f1': 0.8400918133129303, 'eval/latency_8/demon_right_rate_model': 0.1800222593210907, 'eval/latency_8/demon_right_rate_teacher': 0.18363939899833054, 'eval/latency_8/demon_move_f1': 0.8355710208117162, 'eval/latency_8/demon_cc_f1': 0.9255884843980249, 'timing/data': 0.009031827561557293, 'timing/model': 0.9495140081271529, 'learning_rate/qwen_vl_interface': 3.457258820175082e-06, 'learning_rate/action_model': 3.457258820175082e-05, 'epoch': 2.0}) +[2026-06-17 23:34:44,071][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_4400_state +[2026-06-17 23:34:44,093][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 23:34:51,027][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_4400_state/model.safetensors +[2026-06-17 23:35:06,878][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_4400_state/optimizer.bin +[2026-06-17 23:35:06,878][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_4400_state/sampler.bin +[2026-06-17 23:35:06,879][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_4400_state/sampler_1.bin +[2026-06-17 23:35:06,881][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_4400_state/random_states_0.pkl +[2026-06-17 23:35:06,885][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4400 +[2026-06-17 23:35:06,886][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 23:35:06,899][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 23:35:11,019][starVLA.training.train_starvla][INFO] - Step 4401, Loss: {'train/loss': 0.3524044156074524, 'train/grad_norm_pre_clip': 25.287349700927734, 'timing/data': 0.010671794414520264, 'timing/model': 1.2816808270290494, 'learning_rate/qwen_vl_interface': 3.4552562626122817e-06, 'learning_rate/action_model': 3.455256262612281e-05, 'epoch': 2.0}) +[2026-06-17 23:35:14,930][starVLA.training.train_starvla][INFO] - Step 4402, Loss: {'train/loss': 0.20789574086666107, 'train/grad_norm_pre_clip': 28.417448043823242, 'timing/data': 0.01737904269248247, 'timing/model': 1.0143603906035423, 'learning_rate/qwen_vl_interface': 3.453254077101789e-06, 'learning_rate/action_model': 3.453254077101789e-05, 'epoch': 2.0}) +[2026-06-17 23:35:18,946][starVLA.training.train_starvla][INFO] - Step 4403, Loss: {'train/loss': 0.3844989240169525, 'train/grad_norm_pre_clip': 25.84613037109375, 'timing/data': 0.011293299496173859, 'timing/model': 0.9579341867938638, 'learning_rate/qwen_vl_interface': 3.45125226405866e-06, 'learning_rate/action_model': 3.4512522640586595e-05, 'epoch': 2.0}) +[2026-06-17 23:35:22,852][starVLA.training.train_starvla][INFO] - Step 4404, Loss: {'train/loss': 0.1216232106089592, 'train/grad_norm_pre_clip': 18.09677505493164, 'timing/data': 0.011635414324700832, 'timing/model': 0.957696108147502, 'learning_rate/qwen_vl_interface': 3.449250823897872e-06, 'learning_rate/action_model': 3.449250823897872e-05, 'epoch': 2.0}) +[2026-06-17 23:35:26,735][starVLA.training.train_starvla][INFO] - Step 4405, Loss: {'train/loss': 0.10509154200553894, 'train/grad_norm_pre_clip': 15.09067153930664, 'timing/data': 0.019350042566657066, 'timing/model': 0.9428769238293171, 'learning_rate/qwen_vl_interface': 3.447249757034328e-06, 'learning_rate/action_model': 3.4472497570343276e-05, 'epoch': 2.0}) +[2026-06-17 23:35:30,632][starVLA.training.train_starvla][INFO] - Step 4406, Loss: {'train/loss': 0.18664559721946716, 'train/grad_norm_pre_clip': 18.726760864257812, 'timing/data': 0.009984959848225117, 'timing/model': 0.9645263701677322, 'learning_rate/qwen_vl_interface': 3.445249063882845e-06, 'learning_rate/action_model': 3.445249063882845e-05, 'epoch': 2.0}) +[2026-06-17 23:35:34,605][starVLA.training.train_starvla][INFO] - Step 4407, Loss: {'train/loss': 0.21067304909229279, 'train/grad_norm_pre_clip': 21.189376831054688, 'timing/data': 0.010238468647003174, 'timing/model': 1.155743115581572, 'learning_rate/qwen_vl_interface': 3.4432487448581762e-06, 'learning_rate/action_model': 3.443248744858176e-05, 'epoch': 2.0}) +[2026-06-17 23:35:38,511][starVLA.training.train_starvla][INFO] - Step 4408, Loss: {'train/loss': 0.1873895823955536, 'train/grad_norm_pre_clip': 14.726428985595703, 'timing/data': 0.009623268619179726, 'timing/model': 1.1701452173292637, 'learning_rate/qwen_vl_interface': 3.4412488003749867e-06, 'learning_rate/action_model': 3.4412488003749866e-05, 'epoch': 2.0}) +[2026-06-17 23:35:42,522][starVLA.training.train_starvla][INFO] - Step 4409, Loss: {'train/loss': 0.3930489420890808, 'train/grad_norm_pre_clip': 27.416542053222656, 'timing/data': 0.011343595571815968, 'timing/model': 0.986110526137054, 'learning_rate/qwen_vl_interface': 3.4392492308478657e-06, 'learning_rate/action_model': 3.439249230847866e-05, 'epoch': 2.0}) +[2026-06-17 23:35:46,466][starVLA.training.train_starvla][INFO] - Step 4410, Loss: {'train/loss': 0.33930277824401855, 'train/grad_norm_pre_clip': 29.297426223754883, 'timing/data': 0.009491353295743465, 'timing/model': 0.9524497631937265, 'learning_rate/qwen_vl_interface': 3.437250036691328e-06, 'learning_rate/action_model': 3.437250036691328e-05, 'epoch': 2.0}) +[2026-06-17 23:35:50,485][starVLA.training.train_starvla][INFO] - Step 4411, Loss: {'train/loss': 0.16841425001621246, 'train/grad_norm_pre_clip': 15.28742790222168, 'timing/data': 0.011895380914211273, 'timing/model': 1.0260457582771778, 'learning_rate/qwen_vl_interface': 3.4352512183198094e-06, 'learning_rate/action_model': 3.4352512183198095e-05, 'epoch': 2.0}) +[2026-06-17 23:35:54,460][starVLA.training.train_starvla][INFO] - Step 4412, Loss: {'train/loss': 0.5321052670478821, 'train/grad_norm_pre_clip': 30.740808486938477, 'timing/data': 0.010943681932985783, 'timing/model': 0.9509734241291881, 'learning_rate/qwen_vl_interface': 3.4332527761476662e-06, 'learning_rate/action_model': 3.433252776147666e-05, 'epoch': 2.0}) +[2026-06-17 23:35:58,528][starVLA.training.train_starvla][INFO] - Step 4413, Loss: {'train/loss': 0.1348736584186554, 'train/grad_norm_pre_clip': 24.1707763671875, 'timing/data': 0.01612362079322338, 'timing/model': 0.9560085851699114, 'learning_rate/qwen_vl_interface': 3.431254710589174e-06, 'learning_rate/action_model': 3.431254710589174e-05, 'epoch': 2.0}) +[2026-06-17 23:36:02,403][starVLA.training.train_starvla][INFO] - Step 4414, Loss: {'train/loss': 0.46253082156181335, 'train/grad_norm_pre_clip': 30.36270523071289, 'timing/data': 0.009139204397797585, 'timing/model': 1.1442415807396173, 'learning_rate/qwen_vl_interface': 3.4292570220585407e-06, 'learning_rate/action_model': 3.4292570220585407e-05, 'epoch': 2.0}) +[2026-06-17 23:36:06,107][starVLA.training.train_starvla][INFO] - Step 4415, Loss: {'train/loss': 0.2737046480178833, 'train/grad_norm_pre_clip': 18.67610740661621, 'timing/data': 0.013194495812058449, 'timing/model': 0.9484371906146407, 'learning_rate/qwen_vl_interface': 3.4272597109698857e-06, 'learning_rate/action_model': 3.427259710969886e-05, 'epoch': 2.0}) +[2026-06-17 23:36:09,998][starVLA.training.train_starvla][INFO] - Step 4416, Loss: {'train/loss': 0.15214771032333374, 'train/grad_norm_pre_clip': 23.96556282043457, 'timing/data': 0.009054833091795444, 'timing/model': 0.9555620672181249, 'learning_rate/qwen_vl_interface': 3.4252627777372533e-06, 'learning_rate/action_model': 3.425262777737253e-05, 'epoch': 2.0}) +[2026-06-17 23:36:13,857][starVLA.training.train_starvla][INFO] - Step 4417, Loss: {'train/loss': 0.11364977806806564, 'train/grad_norm_pre_clip': 16.346006393432617, 'timing/data': 0.010633563622832298, 'timing/model': 0.953911847434938, 'learning_rate/qwen_vl_interface': 3.4232662227746095e-06, 'learning_rate/action_model': 3.423266222774609e-05, 'epoch': 2.0}) +[2026-06-17 23:36:17,793][starVLA.training.train_starvla][INFO] - Step 4418, Loss: {'train/loss': 0.4002574384212494, 'train/grad_norm_pre_clip': 37.900245666503906, 'timing/data': 0.008151250891387463, 'timing/model': 0.9580546831712127, 'learning_rate/qwen_vl_interface': 3.4212700464958448e-06, 'learning_rate/action_model': 3.421270046495845e-05, 'epoch': 2.0}) +[2026-06-17 23:36:21,663][starVLA.training.train_starvla][INFO] - Step 4419, Loss: {'train/loss': 0.19027656316757202, 'train/grad_norm_pre_clip': 27.887197494506836, 'timing/data': 0.007835146971046925, 'timing/model': 0.950559196062386, 'learning_rate/qwen_vl_interface': 3.4192742493147686e-06, 'learning_rate/action_model': 3.419274249314768e-05, 'epoch': 2.0}) +[2026-06-17 23:36:25,598][starVLA.training.train_starvla][INFO] - Step 4420, Loss: {'train/loss': 0.17452386021614075, 'train/grad_norm_pre_clip': 21.53037452697754, 'timing/data': 0.0167981693521142, 'timing/model': 0.9442212982103229, 'learning_rate/qwen_vl_interface': 3.4172788316451065e-06, 'learning_rate/action_model': 3.417278831645107e-05, 'epoch': 2.01}) +[2026-06-17 23:36:29,431][starVLA.training.train_starvla][INFO] - Step 4421, Loss: {'train/loss': 0.4694761037826538, 'train/grad_norm_pre_clip': 30.003311157226562, 'timing/data': 0.009920874610543251, 'timing/model': 1.1380576370283961, 'learning_rate/qwen_vl_interface': 3.4152837939005177e-06, 'learning_rate/action_model': 3.415283793900518e-05, 'epoch': 2.01}) +[2026-06-17 23:36:33,267][starVLA.training.train_starvla][INFO] - Step 4422, Loss: {'train/loss': 0.36201024055480957, 'train/grad_norm_pre_clip': 30.956884384155273, 'timing/data': 0.0120115140452981, 'timing/model': 0.9911574553698301, 'learning_rate/qwen_vl_interface': 3.413289136494572e-06, 'learning_rate/action_model': 3.413289136494572e-05, 'epoch': 2.01}) +[2026-06-17 23:36:37,150][starVLA.training.train_starvla][INFO] - Step 4423, Loss: {'train/loss': 0.3182162642478943, 'train/grad_norm_pre_clip': 23.702295303344727, 'timing/data': 0.007980754598975182, 'timing/model': 0.9412880344316363, 'learning_rate/qwen_vl_interface': 3.411294859840765e-06, 'learning_rate/action_model': 3.411294859840764e-05, 'epoch': 2.01}) +[2026-06-17 23:36:41,122][starVLA.training.train_starvla][INFO] - Step 4424, Loss: {'train/loss': 0.2651207447052002, 'train/grad_norm_pre_clip': 21.991008758544922, 'timing/data': 0.00901443138718605, 'timing/model': 0.9540081033483148, 'learning_rate/qwen_vl_interface': 3.409300964352511e-06, 'learning_rate/action_model': 3.409300964352511e-05, 'epoch': 2.01}) +[2026-06-17 23:36:46,078][starVLA.training.train_starvla][INFO] - Step 4425, Loss: {'train/loss': 0.4210294187068939, 'train/grad_norm_pre_clip': 36.622432708740234, 'timing/data': 0.010754657909274101, 'timing/model': 0.9546545036137104, 'learning_rate/qwen_vl_interface': 3.407307450443148e-06, 'learning_rate/action_model': 3.407307450443148e-05, 'epoch': 2.01}) +[2026-06-17 23:36:51,428][starVLA.training.train_starvla][INFO] - Step 4426, Loss: {'train/loss': 0.33612656593322754, 'train/grad_norm_pre_clip': 21.266586303710938, 'timing/data': 0.013071656227111816, 'timing/model': 2.325018501840532, 'learning_rate/qwen_vl_interface': 3.4053143185259353e-06, 'learning_rate/action_model': 3.405314318525935e-05, 'epoch': 2.01}) +[2026-06-17 23:36:55,619][starVLA.training.train_starvla][INFO] - Step 4427, Loss: {'train/loss': 0.2954779863357544, 'train/grad_norm_pre_clip': 22.772653579711914, 'timing/data': 0.012408052571117878, 'timing/model': 1.3922594375908375, 'learning_rate/qwen_vl_interface': 3.4033215690140475e-06, 'learning_rate/action_model': 3.4033215690140476e-05, 'epoch': 2.01}) +[2026-06-17 23:36:59,453][starVLA.training.train_starvla][INFO] - Step 4428, Loss: {'train/loss': 0.21292197704315186, 'train/grad_norm_pre_clip': 18.842748641967773, 'timing/data': 0.007218477316200733, 'timing/model': 1.0993525190278888, 'learning_rate/qwen_vl_interface': 3.4013292023205852e-06, 'learning_rate/action_model': 3.401329202320585e-05, 'epoch': 2.01}) +[2026-06-17 23:37:04,317][starVLA.training.train_starvla][INFO] - Step 4429, Loss: {'train/loss': 0.19669750332832336, 'train/grad_norm_pre_clip': 21.180856704711914, 'timing/data': 0.011494908481836319, 'timing/model': 1.6566520696505904, 'learning_rate/qwen_vl_interface': 3.3993372188585694e-06, 'learning_rate/action_model': 3.3993372188585694e-05, 'epoch': 2.01}) +[2026-06-17 23:37:08,371][starVLA.training.train_starvla][INFO] - Step 4430, Loss: {'train/loss': 0.26426923274993896, 'train/grad_norm_pre_clip': 28.30182647705078, 'timing/data': 0.0090863686054945, 'timing/model': 0.9983310429379344, 'learning_rate/qwen_vl_interface': 3.3973456190409397e-06, 'learning_rate/action_model': 3.39734561904094e-05, 'epoch': 2.01}) +[2026-06-17 23:37:12,255][starVLA.training.train_starvla][INFO] - Step 4431, Loss: {'train/loss': 0.21347180008888245, 'train/grad_norm_pre_clip': 18.528955459594727, 'timing/data': 0.00968908704817295, 'timing/model': 0.9503560308367014, 'learning_rate/qwen_vl_interface': 3.395354403280555e-06, 'learning_rate/action_model': 3.395354403280555e-05, 'epoch': 2.01}) +[2026-06-17 23:37:16,148][starVLA.training.train_starvla][INFO] - Step 4432, Loss: {'train/loss': 0.08138613402843475, 'train/grad_norm_pre_clip': 11.421172142028809, 'timing/data': 0.007991275750100613, 'timing/model': 0.9314484912902117, 'learning_rate/qwen_vl_interface': 3.3933635719902e-06, 'learning_rate/action_model': 3.3933635719901995e-05, 'epoch': 2.01}) +[2026-06-17 23:37:19,949][starVLA.training.train_starvla][INFO] - Step 4433, Loss: {'train/loss': 0.08775344491004944, 'train/grad_norm_pre_clip': 17.0474796295166, 'timing/data': 0.007803804241120815, 'timing/model': 0.9563317904248834, 'learning_rate/qwen_vl_interface': 3.3913731255825754e-06, 'learning_rate/action_model': 3.3913731255825754e-05, 'epoch': 2.01}) +[2026-06-17 23:37:23,819][starVLA.training.train_starvla][INFO] - Step 4434, Loss: {'train/loss': 0.09483375400304794, 'train/grad_norm_pre_clip': 12.004955291748047, 'timing/data': 0.009269513189792633, 'timing/model': 1.1839821934700012, 'learning_rate/qwen_vl_interface': 3.3893830644703003e-06, 'learning_rate/action_model': 3.3893830644703e-05, 'epoch': 2.01}) +[2026-06-17 23:37:27,599][starVLA.training.train_starvla][INFO] - Step 4435, Loss: {'train/loss': 0.24716611206531525, 'train/grad_norm_pre_clip': 21.299280166625977, 'timing/data': 0.016071478836238384, 'timing/model': 1.1062237527221441, 'learning_rate/qwen_vl_interface': 3.3873933890659174e-06, 'learning_rate/action_model': 3.3873933890659174e-05, 'epoch': 2.01}) +[2026-06-17 23:37:31,476][starVLA.training.train_starvla][INFO] - Step 4436, Loss: {'train/loss': 0.29299554228782654, 'train/grad_norm_pre_clip': 20.16555404663086, 'timing/data': 0.006647263653576374, 'timing/model': 0.9464092534035444, 'learning_rate/qwen_vl_interface': 3.3854040997818897e-06, 'learning_rate/action_model': 3.38540409978189e-05, 'epoch': 2.01}) +[2026-06-17 23:37:35,438][starVLA.training.train_starvla][INFO] - Step 4437, Loss: {'train/loss': 0.19375768303871155, 'train/grad_norm_pre_clip': 17.858768463134766, 'timing/data': 0.013428469188511372, 'timing/model': 0.9869989668950438, 'learning_rate/qwen_vl_interface': 3.383415197030599e-06, 'learning_rate/action_model': 3.383415197030599e-05, 'epoch': 2.01}) +[2026-06-17 23:37:40,405][starVLA.training.train_starvla][INFO] - Step 4438, Loss: {'train/loss': 0.21273009479045868, 'train/grad_norm_pre_clip': 23.214643478393555, 'timing/data': 0.01012688223272562, 'timing/model': 2.0186903895810246, 'learning_rate/qwen_vl_interface': 3.3814266812243473e-06, 'learning_rate/action_model': 3.3814266812243474e-05, 'epoch': 2.01}) +[2026-06-17 23:37:44,348][starVLA.training.train_starvla][INFO] - Step 4439, Loss: {'train/loss': 0.15202803909778595, 'train/grad_norm_pre_clip': 18.869312286376953, 'timing/data': 0.012079084292054176, 'timing/model': 0.9340912839397788, 'learning_rate/qwen_vl_interface': 3.3794385527753522e-06, 'learning_rate/action_model': 3.379438552775352e-05, 'epoch': 2.01}) +[2026-06-17 23:37:48,271][starVLA.training.train_starvla][INFO] - Step 4440, Loss: {'train/loss': 0.49145421385765076, 'train/grad_norm_pre_clip': 36.922760009765625, 'timing/data': 0.012367334216833115, 'timing/model': 0.9431620072573423, 'learning_rate/qwen_vl_interface': 3.3774508120957606e-06, 'learning_rate/action_model': 3.37745081209576e-05, 'epoch': 2.01}) +[2026-06-17 23:37:52,123][starVLA.training.train_starvla][INFO] - Step 4441, Loss: {'train/loss': 0.17459887266159058, 'train/grad_norm_pre_clip': 19.30167007446289, 'timing/data': 0.009773223660886288, 'timing/model': 1.1007536444813013, 'learning_rate/qwen_vl_interface': 3.37546345959763e-06, 'learning_rate/action_model': 3.3754634595976295e-05, 'epoch': 2.01}) +[2026-06-17 23:37:55,811][starVLA.training.train_starvla][INFO] - Step 4442, Loss: {'train/loss': 0.18721206486225128, 'train/grad_norm_pre_clip': 23.737703323364258, 'timing/data': 0.007890472188591957, 'timing/model': 0.9641329059377313, 'learning_rate/qwen_vl_interface': 3.3734764956929407e-06, 'learning_rate/action_model': 3.37347649569294e-05, 'epoch': 2.02}) +[2026-06-17 23:37:59,629][starVLA.training.train_starvla][INFO] - Step 4443, Loss: {'train/loss': 0.26463356614112854, 'train/grad_norm_pre_clip': 17.961557388305664, 'timing/data': 0.01555623672902584, 'timing/model': 0.9345641918480396, 'learning_rate/qwen_vl_interface': 3.3714899207935937e-06, 'learning_rate/action_model': 3.371489920793594e-05, 'epoch': 2.02}) +[2026-06-17 23:38:03,456][starVLA.training.train_starvla][INFO] - Step 4444, Loss: {'train/loss': 0.4468677341938019, 'train/grad_norm_pre_clip': 30.446359634399414, 'timing/data': 0.014339566230773926, 'timing/model': 0.9442738406360149, 'learning_rate/qwen_vl_interface': 3.369503735311407e-06, 'learning_rate/action_model': 3.369503735311407e-05, 'epoch': 2.02}) +[2026-06-17 23:38:07,408][starVLA.training.train_starvla][INFO] - Step 4445, Loss: {'train/loss': 0.23820489645004272, 'train/grad_norm_pre_clip': 198.21046447753906, 'timing/data': 0.020999683067202568, 'timing/model': 0.9557294575497508, 'learning_rate/qwen_vl_interface': 3.3675179396581208e-06, 'learning_rate/action_model': 3.367517939658121e-05, 'epoch': 2.02}) +[2026-06-17 23:38:11,252][starVLA.training.train_starvla][INFO] - Step 4446, Loss: {'train/loss': 0.37227922677993774, 'train/grad_norm_pre_clip': 25.193931579589844, 'timing/data': 0.008234268054366112, 'timing/model': 0.9445769051089883, 'learning_rate/qwen_vl_interface': 3.365532534245389e-06, 'learning_rate/action_model': 3.3655325342453894e-05, 'epoch': 2.02}) +[2026-06-17 23:38:15,139][starVLA.training.train_starvla][INFO] - Step 4447, Loss: {'train/loss': 0.11471269279718399, 'train/grad_norm_pre_clip': 22.813600540161133, 'timing/data': 0.009223638102412224, 'timing/model': 1.1649225633591413, 'learning_rate/qwen_vl_interface': 3.3635475194847927e-06, 'learning_rate/action_model': 3.363547519484793e-05, 'epoch': 2.02}) +[2026-06-17 23:38:18,915][starVLA.training.train_starvla][INFO] - Step 4448, Loss: {'train/loss': 0.07923717796802521, 'train/grad_norm_pre_clip': 18.271150588989258, 'timing/data': 0.012760993093252182, 'timing/model': 1.0536995083093643, 'learning_rate/qwen_vl_interface': 3.361562895787825e-06, 'learning_rate/action_model': 3.361562895787825e-05, 'epoch': 2.02}) +[2026-06-17 23:38:22,538][starVLA.training.train_starvla][INFO] - Step 4449, Loss: {'train/loss': 0.25049588084220886, 'train/grad_norm_pre_clip': 42.1375617980957, 'timing/data': 0.011852990835905075, 'timing/model': 0.9402172453701496, 'learning_rate/qwen_vl_interface': 3.3595786635659e-06, 'learning_rate/action_model': 3.3595786635659e-05, 'epoch': 2.02}) +[2026-06-17 23:38:26,353][starVLA.training.train_starvla][INFO] - Step 4450, Loss: {'train/loss': 0.23674717545509338, 'train/grad_norm_pre_clip': 19.79336166381836, 'timing/data': 0.008015993982553482, 'timing/model': 0.9469921849668026, 'learning_rate/qwen_vl_interface': 3.357594823230353e-06, 'learning_rate/action_model': 3.3575948232303525e-05, 'epoch': 2.02}) +[2026-06-17 23:38:30,312][starVLA.training.train_starvla][INFO] - Step 4451, Loss: {'train/loss': 0.43434157967567444, 'train/grad_norm_pre_clip': 30.75082015991211, 'timing/data': 0.007996444590389729, 'timing/model': 1.0325898565351963, 'learning_rate/qwen_vl_interface': 3.3556113751924357e-06, 'learning_rate/action_model': 3.355611375192436e-05, 'epoch': 2.02}) +[2026-06-17 23:38:34,138][starVLA.training.train_starvla][INFO] - Step 4452, Loss: {'train/loss': 0.1496042013168335, 'train/grad_norm_pre_clip': 13.662850379943848, 'timing/data': 0.0085976617410779, 'timing/model': 0.9510040376335382, 'learning_rate/qwen_vl_interface': 3.3536283198633183e-06, 'learning_rate/action_model': 3.353628319863318e-05, 'epoch': 2.02}) +[2026-06-17 23:38:37,958][starVLA.training.train_starvla][INFO] - Step 4453, Loss: {'train/loss': 0.18913604319095612, 'train/grad_norm_pre_clip': 14.755960464477539, 'timing/data': 0.01106112077832222, 'timing/model': 0.9458196284249425, 'learning_rate/qwen_vl_interface': 3.3516456576540913e-06, 'learning_rate/action_model': 3.3516456576540914e-05, 'epoch': 2.02}) +[2026-06-17 23:38:41,862][starVLA.training.train_starvla][INFO] - Step 4454, Loss: {'train/loss': 0.18365254998207092, 'train/grad_norm_pre_clip': 28.2613468170166, 'timing/data': 0.008861919865012169, 'timing/model': 1.195204964838922, 'learning_rate/qwen_vl_interface': 3.349663388975763e-06, 'learning_rate/action_model': 3.3496633889757625e-05, 'epoch': 2.02}) +[2026-06-17 23:38:45,611][starVLA.training.train_starvla][INFO] - Step 4455, Loss: {'train/loss': 0.19164502620697021, 'train/grad_norm_pre_clip': 30.857107162475586, 'timing/data': 0.010497266426682472, 'timing/model': 0.947049886919558, 'learning_rate/qwen_vl_interface': 3.347681514239258e-06, 'learning_rate/action_model': 3.347681514239257e-05, 'epoch': 2.02}) +[2026-06-17 23:38:49,449][starVLA.training.train_starvla][INFO] - Step 4456, Loss: {'train/loss': 0.19528716802597046, 'train/grad_norm_pre_clip': 30.395122528076172, 'timing/data': 0.011272215284407139, 'timing/model': 0.944551020860672, 'learning_rate/qwen_vl_interface': 3.345700033855421e-06, 'learning_rate/action_model': 3.345700033855421e-05, 'epoch': 2.02}) +[2026-06-17 23:38:53,345][starVLA.training.train_starvla][INFO] - Step 4457, Loss: {'train/loss': 0.2768978774547577, 'train/grad_norm_pre_clip': 27.870019912719727, 'timing/data': 0.011390183120965958, 'timing/model': 0.9584282543510199, 'learning_rate/qwen_vl_interface': 3.3437189482350173e-06, 'learning_rate/action_model': 3.343718948235017e-05, 'epoch': 2.02}) +[2026-06-17 23:38:57,144][starVLA.training.train_starvla][INFO] - Step 4458, Loss: {'train/loss': 0.18993708491325378, 'train/grad_norm_pre_clip': 12.596914291381836, 'timing/data': 0.007840003818273544, 'timing/model': 0.9382960367947817, 'learning_rate/qwen_vl_interface': 3.341738257788727e-06, 'learning_rate/action_model': 3.3417382577887266e-05, 'epoch': 2.02}) +[2026-06-17 23:39:01,001][starVLA.training.train_starvla][INFO] - Step 4459, Loss: {'train/loss': 0.4260770380496979, 'train/grad_norm_pre_clip': 18.89691925048828, 'timing/data': 0.011120864190161228, 'timing/model': 0.9567506369203329, 'learning_rate/qwen_vl_interface': 3.3397579629271487e-06, 'learning_rate/action_model': 3.3397579629271486e-05, 'epoch': 2.02}) +[2026-06-17 23:39:04,902][starVLA.training.train_starvla][INFO] - Step 4460, Loss: {'train/loss': 0.16887272894382477, 'train/grad_norm_pre_clip': 20.770784378051758, 'timing/data': 0.01241418719291687, 'timing/model': 0.9679800923913717, 'learning_rate/qwen_vl_interface': 3.3377780640607997e-06, 'learning_rate/action_model': 3.3377780640608e-05, 'epoch': 2.02}) +[2026-06-17 23:39:08,729][starVLA.training.train_starvla][INFO] - Step 4461, Loss: {'train/loss': 0.376483291387558, 'train/grad_norm_pre_clip': 160.5216522216797, 'timing/data': 0.0005565760657191277, 'timing/model': 1.112279417924583, 'learning_rate/qwen_vl_interface': 3.3357985616001166e-06, 'learning_rate/action_model': 3.335798561600116e-05, 'epoch': 2.02}) +[2026-06-17 23:39:12,433][starVLA.training.train_starvla][INFO] - Step 4462, Loss: {'train/loss': 0.619473934173584, 'train/grad_norm_pre_clip': 33.52372360229492, 'timing/data': 0.0006519900634884834, 'timing/model': 0.9771035760641098, 'learning_rate/qwen_vl_interface': 3.3338194559554524e-06, 'learning_rate/action_model': 3.333819455955453e-05, 'epoch': 2.02}) +[2026-06-17 23:39:16,402][starVLA.training.train_starvla][INFO] - Step 4463, Loss: {'train/loss': 0.5256674289703369, 'train/grad_norm_pre_clip': 30.90160369873047, 'timing/data': 0.0004945425316691399, 'timing/model': 1.0426947241649032, 'learning_rate/qwen_vl_interface': 3.3318407475370758e-06, 'learning_rate/action_model': 3.3318407475370755e-05, 'epoch': 2.02}) +[2026-06-17 23:39:20,242][starVLA.training.train_starvla][INFO] - Step 4464, Loss: {'train/loss': 0.2900208830833435, 'train/grad_norm_pre_clip': 36.043174743652344, 'timing/data': 0.0012463973835110664, 'timing/model': 0.9354712311178446, 'learning_rate/qwen_vl_interface': 3.329862436755176e-06, 'learning_rate/action_model': 3.329862436755176e-05, 'epoch': 2.03}) +[2026-06-17 23:39:24,409][starVLA.training.train_starvla][INFO] - Step 4465, Loss: {'train/loss': 0.1867491602897644, 'train/grad_norm_pre_clip': 17.17068862915039, 'timing/data': 0.0009044259786605835, 'timing/model': 1.2174842869862914, 'learning_rate/qwen_vl_interface': 3.3278845240198592e-06, 'learning_rate/action_model': 3.3278845240198595e-05, 'epoch': 2.03}) +[2026-06-17 23:39:28,626][starVLA.training.train_starvla][INFO] - Step 4466, Loss: {'train/loss': 0.1811305433511734, 'train/grad_norm_pre_clip': 19.840473175048828, 'timing/data': 0.0009159054607152939, 'timing/model': 0.9424117114394903, 'learning_rate/qwen_vl_interface': 3.325907009741149e-06, 'learning_rate/action_model': 3.3259070097411485e-05, 'epoch': 2.03}) +[2026-06-17 23:39:32,646][starVLA.training.train_starvla][INFO] - Step 4467, Loss: {'train/loss': 0.3055938482284546, 'train/grad_norm_pre_clip': 24.761056900024414, 'timing/data': 0.0008033774793148041, 'timing/model': 1.1478322204202414, 'learning_rate/qwen_vl_interface': 3.323929894328987e-06, 'learning_rate/action_model': 3.323929894328987e-05, 'epoch': 2.03}) +[2026-06-17 23:39:39,564][starVLA.training.train_starvla][INFO] - Step 4468, Loss: {'train/loss': 0.2253144532442093, 'train/grad_norm_pre_clip': 23.20004653930664, 'timing/data': 0.0006014043465256691, 'timing/model': 1.1333254054188728, 'learning_rate/qwen_vl_interface': 3.321953178193228e-06, 'learning_rate/action_model': 3.321953178193228e-05, 'epoch': 2.03}) +[2026-06-17 23:39:43,541][starVLA.training.train_starvla][INFO] - Step 4469, Loss: {'train/loss': 0.09726713597774506, 'train/grad_norm_pre_clip': 10.70016098022461, 'timing/data': 0.0008553480729460716, 'timing/model': 0.942259319126606, 'learning_rate/qwen_vl_interface': 3.319976861743653e-06, 'learning_rate/action_model': 3.319976861743653e-05, 'epoch': 2.03}) +[2026-06-17 23:39:47,456][starVLA.training.train_starvla][INFO] - Step 4470, Loss: {'train/loss': 0.6016010642051697, 'train/grad_norm_pre_clip': 22.332347869873047, 'timing/data': 0.001202218234539032, 'timing/model': 0.9519302062690258, 'learning_rate/qwen_vl_interface': 3.3180009453899496e-06, 'learning_rate/action_model': 3.318000945389949e-05, 'epoch': 2.03}) +[2026-06-17 23:39:51,361][starVLA.training.train_starvla][INFO] - Step 4471, Loss: {'train/loss': 0.22788342833518982, 'train/grad_norm_pre_clip': 24.29013442993164, 'timing/data': 0.0008652377873659134, 'timing/model': 0.9647693382576108, 'learning_rate/qwen_vl_interface': 3.3160254295417287e-06, 'learning_rate/action_model': 3.3160254295417285e-05, 'epoch': 2.03}) +[2026-06-17 23:39:55,247][starVLA.training.train_starvla][INFO] - Step 4472, Loss: {'train/loss': 0.16027124226093292, 'train/grad_norm_pre_clip': 18.792003631591797, 'timing/data': 0.0004961108788847923, 'timing/model': 0.9749998673796654, 'learning_rate/qwen_vl_interface': 3.314050314608518e-06, 'learning_rate/action_model': 3.314050314608518e-05, 'epoch': 2.03}) +[2026-06-17 23:39:59,192][starVLA.training.train_starvla][INFO] - Step 4473, Loss: {'train/loss': 0.18259279429912567, 'train/grad_norm_pre_clip': 20.97780990600586, 'timing/data': 0.0005135927349328995, 'timing/model': 0.9583557648584247, 'learning_rate/qwen_vl_interface': 3.3120756009997606e-06, 'learning_rate/action_model': 3.312075600999761e-05, 'epoch': 2.03}) +[2026-06-17 23:40:03,071][starVLA.training.train_starvla][INFO] - Step 4474, Loss: {'train/loss': 0.3106445372104645, 'train/grad_norm_pre_clip': 28.207561492919922, 'timing/data': 0.0008252756670117378, 'timing/model': 1.123570160008967, 'learning_rate/qwen_vl_interface': 3.310101289124817e-06, 'learning_rate/action_model': 3.310101289124817e-05, 'epoch': 2.03}) +[2026-06-17 23:40:06,948][starVLA.training.train_starvla][INFO] - Step 4475, Loss: {'train/loss': 0.1832803636789322, 'train/grad_norm_pre_clip': 19.42119026184082, 'timing/data': 0.0004991311579942703, 'timing/model': 1.1261148815974593, 'learning_rate/qwen_vl_interface': 3.3081273793929615e-06, 'learning_rate/action_model': 3.308127379392961e-05, 'epoch': 2.03}) +[2026-06-17 23:40:10,805][starVLA.training.train_starvla][INFO] - Step 4476, Loss: {'train/loss': 0.24127286672592163, 'train/grad_norm_pre_clip': 17.26279067993164, 'timing/data': 0.000573059543967247, 'timing/model': 1.0471586575731635, 'learning_rate/qwen_vl_interface': 3.306153872213392e-06, 'learning_rate/action_model': 3.306153872213392e-05, 'epoch': 2.03}) +[2026-06-17 23:40:14,796][starVLA.training.train_starvla][INFO] - Step 4477, Loss: {'train/loss': 0.1124868243932724, 'train/grad_norm_pre_clip': 12.586282730102539, 'timing/data': 0.0009336834773421288, 'timing/model': 0.9903657538816333, 'learning_rate/qwen_vl_interface': 3.304180767995215e-06, 'learning_rate/action_model': 3.3041807679952146e-05, 'epoch': 2.03}) +[2026-06-17 23:40:18,725][starVLA.training.train_starvla][INFO] - Step 4478, Loss: {'train/loss': 0.24020929634571075, 'train/grad_norm_pre_clip': 23.260005950927734, 'timing/data': 0.0004891632124781609, 'timing/model': 0.9564707465469837, 'learning_rate/qwen_vl_interface': 3.3022080671474588e-06, 'learning_rate/action_model': 3.302208067147459e-05, 'epoch': 2.03}) +[2026-06-17 23:40:24,079][starVLA.training.train_starvla][INFO] - Step 4479, Loss: {'train/loss': 0.12880568206310272, 'train/grad_norm_pre_clip': 15.625879287719727, 'timing/data': 0.000527714379131794, 'timing/model': 2.5239976961165667, 'learning_rate/qwen_vl_interface': 3.300235770079066e-06, 'learning_rate/action_model': 3.300235770079066e-05, 'epoch': 2.03}) +[2026-06-17 23:40:28,015][starVLA.training.train_starvla][INFO] - Step 4480, Loss: {'train/loss': 0.3004896640777588, 'train/grad_norm_pre_clip': 29.842897415161133, 'timing/data': 0.0005232365801930428, 'timing/model': 0.9681228557601571, 'learning_rate/qwen_vl_interface': 3.298263877198896e-06, 'learning_rate/action_model': 3.298263877198896e-05, 'epoch': 2.03}) +[2026-06-17 23:40:31,835][starVLA.training.train_starvla][INFO] - Step 4481, Loss: {'train/loss': 0.31582003831863403, 'train/grad_norm_pre_clip': 29.58449935913086, 'timing/data': 0.0005274070426821709, 'timing/model': 1.1218990702182055, 'learning_rate/qwen_vl_interface': 3.296292388915726e-06, 'learning_rate/action_model': 3.296292388915726e-05, 'epoch': 2.03}) +[2026-06-17 23:40:35,527][starVLA.training.train_starvla][INFO] - Step 4482, Loss: {'train/loss': 0.22240181267261505, 'train/grad_norm_pre_clip': 20.03976821899414, 'timing/data': 0.0010587992146611214, 'timing/model': 0.9620864288881421, 'learning_rate/qwen_vl_interface': 3.294321305638243e-06, 'learning_rate/action_model': 3.294321305638243e-05, 'epoch': 2.03}) +[2026-06-17 23:40:39,534][starVLA.training.train_starvla][INFO] - Step 4483, Loss: {'train/loss': 0.2383834272623062, 'train/grad_norm_pre_clip': 24.95765495300293, 'timing/data': 0.0006303777918219566, 'timing/model': 0.9478242890909314, 'learning_rate/qwen_vl_interface': 3.292350627775061e-06, 'learning_rate/action_model': 3.292350627775061e-05, 'epoch': 2.03}) +[2026-06-17 23:40:43,402][starVLA.training.train_starvla][INFO] - Step 4484, Loss: {'train/loss': 0.13328249752521515, 'train/grad_norm_pre_clip': 17.698156356811523, 'timing/data': 0.0005794595927000046, 'timing/model': 0.9527536677196622, 'learning_rate/qwen_vl_interface': 3.2903803557346997e-06, 'learning_rate/action_model': 3.2903803557347e-05, 'epoch': 2.03}) +[2026-06-17 23:40:47,326][starVLA.training.train_starvla][INFO] - Step 4485, Loss: {'train/loss': 0.07502869516611099, 'train/grad_norm_pre_clip': 10.77034854888916, 'timing/data': 0.001345532014966011, 'timing/model': 0.9428387992084026, 'learning_rate/qwen_vl_interface': 3.2884104899255986e-06, 'learning_rate/action_model': 3.288410489925599e-05, 'epoch': 2.03}) +[2026-06-17 23:40:51,539][starVLA.training.train_starvla][INFO] - Step 4486, Loss: {'train/loss': 0.09881503880023956, 'train/grad_norm_pre_clip': 10.196122169494629, 'timing/data': 0.00048817042261362076, 'timing/model': 1.1984854508191347, 'learning_rate/qwen_vl_interface': 3.2864410307561124e-06, 'learning_rate/action_model': 3.286441030756112e-05, 'epoch': 2.04}) +[2026-06-17 23:40:55,387][starVLA.training.train_starvla][INFO] - Step 4487, Loss: {'train/loss': 0.26939281821250916, 'train/grad_norm_pre_clip': 27.0234317779541, 'timing/data': 0.0004953565075993538, 'timing/model': 0.9341438990086317, 'learning_rate/qwen_vl_interface': 3.284471978634515e-06, 'learning_rate/action_model': 3.284471978634515e-05, 'epoch': 2.04}) +[2026-06-17 23:40:59,213][starVLA.training.train_starvla][INFO] - Step 4488, Loss: {'train/loss': 0.5054274797439575, 'train/grad_norm_pre_clip': 35.4602165222168, 'timing/data': 0.0009933654218912125, 'timing/model': 1.116380832158029, 'learning_rate/qwen_vl_interface': 3.2825033339689926e-06, 'learning_rate/action_model': 3.282503333968993e-05, 'epoch': 2.04}) +[2026-06-17 23:41:02,919][starVLA.training.train_starvla][INFO] - Step 4489, Loss: {'train/loss': 0.47036656737327576, 'train/grad_norm_pre_clip': 36.08477783203125, 'timing/data': 0.0005149571225047112, 'timing/model': 0.9530519209802151, 'learning_rate/qwen_vl_interface': 3.2805350971676447e-06, 'learning_rate/action_model': 3.280535097167645e-05, 'epoch': 2.04}) +[2026-06-17 23:41:06,785][starVLA.training.train_starvla][INFO] - Step 4490, Loss: {'train/loss': 0.29520508646965027, 'train/grad_norm_pre_clip': 21.772476196289062, 'timing/data': 0.0007887780666351318, 'timing/model': 0.9904195005074143, 'learning_rate/qwen_vl_interface': 3.2785672686384893e-06, 'learning_rate/action_model': 3.278567268638489e-05, 'epoch': 2.04}) +[2026-06-17 23:41:10,843][starVLA.training.train_starvla][INFO] - Step 4491, Loss: {'train/loss': 0.12019513547420502, 'train/grad_norm_pre_clip': 13.14928913116455, 'timing/data': 0.0014584101736545563, 'timing/model': 0.9468021132051945, 'learning_rate/qwen_vl_interface': 3.2765998487894622e-06, 'learning_rate/action_model': 3.276599848789462e-05, 'epoch': 2.04}) +[2026-06-17 23:41:14,906][starVLA.training.train_starvla][INFO] - Step 4492, Loss: {'train/loss': 0.15978458523750305, 'train/grad_norm_pre_clip': 21.09806251525879, 'timing/data': 0.0004912465810775757, 'timing/model': 0.971437687985599, 'learning_rate/qwen_vl_interface': 3.2746328380284098e-06, 'learning_rate/action_model': 3.27463283802841e-05, 'epoch': 2.04}) +[2026-06-17 23:41:19,525][starVLA.training.train_starvla][INFO] - Step 4493, Loss: {'train/loss': 0.22260963916778564, 'train/grad_norm_pre_clip': 25.272314071655273, 'timing/data': 0.0007088249549269676, 'timing/model': 0.9590176427736878, 'learning_rate/qwen_vl_interface': 3.2726662367630953e-06, 'learning_rate/action_model': 3.272666236763095e-05, 'epoch': 2.04}) +[2026-06-17 23:41:23,517][starVLA.training.train_starvla][INFO] - Step 4494, Loss: {'train/loss': 0.3864091634750366, 'train/grad_norm_pre_clip': 33.37868118286133, 'timing/data': 0.001103319227695465, 'timing/model': 1.1737966611981392, 'learning_rate/qwen_vl_interface': 3.270700045401197e-06, 'learning_rate/action_model': 3.2707000454011974e-05, 'epoch': 2.04}) +[2026-06-17 23:41:27,680][starVLA.training.train_starvla][INFO] - Step 4495, Loss: {'train/loss': 0.13626734912395477, 'train/grad_norm_pre_clip': 20.86489486694336, 'timing/data': 0.0009329430758953094, 'timing/model': 1.1351383468136191, 'learning_rate/qwen_vl_interface': 3.26873426435031e-06, 'learning_rate/action_model': 3.2687342643503096e-05, 'epoch': 2.04}) +[2026-06-17 23:41:31,326][starVLA.training.train_starvla][INFO] - Step 4496, Loss: {'train/loss': 0.3906475901603699, 'train/grad_norm_pre_clip': 33.393272399902344, 'timing/data': 0.0004702955484390259, 'timing/model': 0.9504000516608357, 'learning_rate/qwen_vl_interface': 3.266768894017942e-06, 'learning_rate/action_model': 3.266768894017942e-05, 'epoch': 2.04}) +[2026-06-17 23:41:35,252][starVLA.training.train_starvla][INFO] - Step 4497, Loss: {'train/loss': 0.2899446487426758, 'train/grad_norm_pre_clip': 21.035600662231445, 'timing/data': 0.0008139479905366898, 'timing/model': 0.9734500553458929, 'learning_rate/qwen_vl_interface': 3.264803934811514e-06, 'learning_rate/action_model': 3.264803934811514e-05, 'epoch': 2.04}) +[2026-06-17 23:41:39,619][starVLA.training.train_starvla][INFO] - Step 4498, Loss: {'train/loss': 0.12042005360126495, 'train/grad_norm_pre_clip': 17.817930221557617, 'timing/data': 0.0005182698369026184, 'timing/model': 0.9715138766914606, 'learning_rate/qwen_vl_interface': 3.2628393871383706e-06, 'learning_rate/action_model': 3.26283938713837e-05, 'epoch': 2.04}) +[2026-06-17 23:41:43,605][starVLA.training.train_starvla][INFO] - Step 4499, Loss: {'train/loss': 0.24358908832073212, 'train/grad_norm_pre_clip': 29.782859802246094, 'timing/data': 0.0006727902218699455, 'timing/model': 0.985708431340754, 'learning_rate/qwen_vl_interface': 3.2608752514057585e-06, 'learning_rate/action_model': 3.260875251405758e-05, 'epoch': 2.04}) +[2026-06-17 23:43:56,701][starVLA.training.train_starvla][INFO] - Step 4500, Loss: {'train/loss': 0.09295659512281418, 'train/grad_norm_pre_clip': 17.420251846313477, 'eval/loss': 0.29860174655914307, 'eval/latency_8/loss': 0.29860174655914307, 'eval/demon_fire_precision': 0.9648533007334963, 'eval/demon_fire_recall': 0.9807393600497049, 'eval/demon_fire_f1': 0.9727314743490987, 'eval/demon_fire_rate_model': 0.9104062326099054, 'eval/demon_fire_rate_teacher': 0.8956594323873122, 'eval/demon_left_precision': 0.8463302752293578, 'eval/demon_left_recall': 0.7901498929336188, 'eval/demon_left_f1': 0.8172757475083057, 'eval/demon_left_rate_model': 0.1213132999443517, 'eval/demon_left_rate_teacher': 0.12993878686700056, 'eval/demon_right_precision': 0.836890243902439, 'eval/demon_right_recall': 0.8318181818181818, 'eval/demon_right_f1': 0.8343465045592705, 'eval/demon_right_rate_model': 0.18252643294379522, 'eval/demon_right_rate_teacher': 0.18363939899833054, 'eval/demon_move_f1': 0.8258111260337881, 'eval/demon_cc_f1': 0.9234443798799132, 'eval/latency_8/demon_fire_precision': 0.9648533007334963, 'eval/latency_8/demon_fire_recall': 0.9807393600497049, 'eval/latency_8/demon_fire_f1': 0.9727314743490987, 'eval/latency_8/demon_fire_rate_model': 0.9104062326099054, 'eval/latency_8/demon_fire_rate_teacher': 0.8956594323873122, 'eval/latency_8/demon_left_precision': 0.8463302752293578, 'eval/latency_8/demon_left_recall': 0.7901498929336188, 'eval/latency_8/demon_left_f1': 0.8172757475083057, 'eval/latency_8/demon_left_rate_model': 0.1213132999443517, 'eval/latency_8/demon_left_rate_teacher': 0.12993878686700056, 'eval/latency_8/demon_right_precision': 0.836890243902439, 'eval/latency_8/demon_right_recall': 0.8318181818181818, 'eval/latency_8/demon_right_f1': 0.8343465045592705, 'eval/latency_8/demon_right_rate_model': 0.18252643294379522, 'eval/latency_8/demon_right_rate_teacher': 0.18363939899833054, 'eval/latency_8/demon_move_f1': 0.8258111260337881, 'eval/latency_8/demon_cc_f1': 0.9234443798799132, 'timing/data': 0.0006526652723550797, 'timing/model': 1.0057920645922422, 'learning_rate/qwen_vl_interface': 3.2589115280208473e-06, 'learning_rate/action_model': 3.258911528020847e-05, 'epoch': 2.04}) +[2026-06-17 23:43:56,704][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_4500_state +[2026-06-17 23:43:56,710][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 23:44:03,347][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_4500_state/model.safetensors +[2026-06-17 23:44:19,174][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_4500_state/optimizer.bin +[2026-06-17 23:44:19,175][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_4500_state/sampler.bin +[2026-06-17 23:44:19,176][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_4500_state/sampler_1.bin +[2026-06-17 23:44:19,178][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_4500_state/random_states_0.pkl +[2026-06-17 23:44:19,186][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4500 +[2026-06-17 23:44:19,186][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 23:44:19,203][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 23:44:23,148][starVLA.training.train_starvla][INFO] - Step 4501, Loss: {'train/loss': 0.5452500581741333, 'train/grad_norm_pre_clip': 28.59071159362793, 'timing/data': 0.0007279710844159126, 'timing/model': 0.9541902774944901, 'learning_rate/qwen_vl_interface': 3.2569482173907164e-06, 'learning_rate/action_model': 3.2569482173907165e-05, 'epoch': 2.04}) +[2026-06-17 23:44:27,044][starVLA.training.train_starvla][INFO] - Step 4502, Loss: {'train/loss': 0.30581042170524597, 'train/grad_norm_pre_clip': 29.374996185302734, 'timing/data': 0.0009110458195209503, 'timing/model': 0.9571182979270816, 'learning_rate/qwen_vl_interface': 3.2549853199223668e-06, 'learning_rate/action_model': 3.254985319922367e-05, 'epoch': 2.04}) +[2026-06-17 23:44:31,060][starVLA.training.train_starvla][INFO] - Step 4503, Loss: {'train/loss': 0.33043956756591797, 'train/grad_norm_pre_clip': 29.002017974853516, 'timing/data': 0.0005558440461754799, 'timing/model': 1.176449884660542, 'learning_rate/qwen_vl_interface': 3.253022836022708e-06, 'learning_rate/action_model': 3.253022836022708e-05, 'epoch': 2.04}) +[2026-06-17 23:44:34,744][starVLA.training.train_starvla][INFO] - Step 4504, Loss: {'train/loss': 0.2425990253686905, 'train/grad_norm_pre_clip': 24.47861671447754, 'timing/data': 0.0008091004565358162, 'timing/model': 0.9555218545719981, 'learning_rate/qwen_vl_interface': 3.2510607660985627e-06, 'learning_rate/action_model': 3.251060766098563e-05, 'epoch': 2.04}) +[2026-06-17 23:44:38,650][starVLA.training.train_starvla][INFO] - Step 4505, Loss: {'train/loss': 0.09927503764629364, 'train/grad_norm_pre_clip': 15.095880508422852, 'timing/data': 0.0009302329272031784, 'timing/model': 0.9585810294374824, 'learning_rate/qwen_vl_interface': 3.249099110556672e-06, 'learning_rate/action_model': 3.249099110556672e-05, 'epoch': 2.04}) +[2026-06-17 23:44:42,801][starVLA.training.train_starvla][INFO] - Step 4506, Loss: {'train/loss': 0.3329417109489441, 'train/grad_norm_pre_clip': 31.093875885009766, 'timing/data': 0.0005334764719009399, 'timing/model': 0.9592849742621183, 'learning_rate/qwen_vl_interface': 3.247137869803689e-06, 'learning_rate/action_model': 3.2471378698036886e-05, 'epoch': 2.04}) +[2026-06-17 23:44:46,826][starVLA.training.train_starvla][INFO] - Step 4507, Loss: {'train/loss': 0.36405685544013977, 'train/grad_norm_pre_clip': 33.92802047729492, 'timing/data': 0.016177835874259472, 'timing/model': 1.0078880731016397, 'learning_rate/qwen_vl_interface': 3.2451770442461813e-06, 'learning_rate/action_model': 3.245177044246181e-05, 'epoch': 2.04}) +[2026-06-17 23:44:50,836][starVLA.training.train_starvla][INFO] - Step 4508, Loss: {'train/loss': 0.10845573991537094, 'train/grad_norm_pre_clip': 13.496281623840332, 'timing/data': 0.014269632287323475, 'timing/model': 0.9637077059596777, 'learning_rate/qwen_vl_interface': 3.2432166342906286e-06, 'learning_rate/action_model': 3.2432166342906284e-05, 'epoch': 2.05}) +[2026-06-17 23:44:54,973][starVLA.training.train_starvla][INFO] - Step 4509, Loss: {'train/loss': 0.15135914087295532, 'train/grad_norm_pre_clip': 15.177589416503906, 'timing/data': 0.011214775033295155, 'timing/model': 0.9898914508521557, 'learning_rate/qwen_vl_interface': 3.241256640343427e-06, 'learning_rate/action_model': 3.241256640343427e-05, 'epoch': 2.05}) +[2026-06-17 23:44:58,942][starVLA.training.train_starvla][INFO] - Step 4510, Loss: {'train/loss': 0.4054306745529175, 'train/grad_norm_pre_clip': 34.7947883605957, 'timing/data': 0.010740074329078197, 'timing/model': 1.2443259581923485, 'learning_rate/qwen_vl_interface': 3.239297062810886e-06, 'learning_rate/action_model': 3.239297062810886e-05, 'epoch': 2.05}) +[2026-06-17 23:45:02,857][starVLA.training.train_starvla][INFO] - Step 4511, Loss: {'train/loss': 0.23124507069587708, 'train/grad_norm_pre_clip': 23.58516502380371, 'timing/data': 0.0100162448361516, 'timing/model': 1.137061427347362, 'learning_rate/qwen_vl_interface': 3.2373379020992252e-06, 'learning_rate/action_model': 3.237337902099225e-05, 'epoch': 2.05}) +[2026-06-17 23:45:06,685][starVLA.training.train_starvla][INFO] - Step 4512, Loss: {'train/loss': 0.5294187664985657, 'train/grad_norm_pre_clip': 33.30946350097656, 'timing/data': 0.010403148829936981, 'timing/model': 0.9554553376510739, 'learning_rate/qwen_vl_interface': 3.2353791586145855e-06, 'learning_rate/action_model': 3.2353791586145855e-05, 'epoch': 2.05}) +[2026-06-17 23:45:10,758][starVLA.training.train_starvla][INFO] - Step 4513, Loss: {'train/loss': 0.18389184772968292, 'train/grad_norm_pre_clip': 36.61473083496094, 'timing/data': 0.013438553549349308, 'timing/model': 0.9718540087342262, 'learning_rate/qwen_vl_interface': 3.2334208327630158e-06, 'learning_rate/action_model': 3.233420832763016e-05, 'epoch': 2.05}) +[2026-06-17 23:45:14,746][starVLA.training.train_starvla][INFO] - Step 4514, Loss: {'train/loss': 0.3788169026374817, 'train/grad_norm_pre_clip': 37.21075439453125, 'timing/data': 0.011254146695137024, 'timing/model': 0.9461357248947024, 'learning_rate/qwen_vl_interface': 3.231462924950475e-06, 'learning_rate/action_model': 3.2314629249504746e-05, 'epoch': 2.05}) +[2026-06-17 23:45:18,641][starVLA.training.train_starvla][INFO] - Step 4515, Loss: {'train/loss': 0.263298898935318, 'train/grad_norm_pre_clip': 25.421348571777344, 'timing/data': 0.008558039553463459, 'timing/model': 0.973777012899518, 'learning_rate/qwen_vl_interface': 3.2295054355828396e-06, 'learning_rate/action_model': 3.22950543558284e-05, 'epoch': 2.05}) +[2026-06-17 23:45:22,568][starVLA.training.train_starvla][INFO] - Step 4516, Loss: {'train/loss': 0.26891860365867615, 'train/grad_norm_pre_clip': 32.2734375, 'timing/data': 0.010733005590736866, 'timing/model': 0.9742496134713292, 'learning_rate/qwen_vl_interface': 3.227548365065905e-06, 'learning_rate/action_model': 3.227548365065905e-05, 'epoch': 2.05}) +[2026-06-17 23:45:26,515][starVLA.training.train_starvla][INFO] - Step 4517, Loss: {'train/loss': 0.30789703130722046, 'train/grad_norm_pre_clip': 30.71344566345215, 'timing/data': 0.010606679134070873, 'timing/model': 1.159694661386311, 'learning_rate/qwen_vl_interface': 3.2255917138053695e-06, 'learning_rate/action_model': 3.2255917138053696e-05, 'epoch': 2.05}) +[2026-06-17 23:45:30,336][starVLA.training.train_starvla][INFO] - Step 4518, Loss: {'train/loss': 0.5159745812416077, 'train/grad_norm_pre_clip': 28.258623123168945, 'timing/data': 0.011374833062291145, 'timing/model': 0.9799705175682902, 'learning_rate/qwen_vl_interface': 3.2236354822068515e-06, 'learning_rate/action_model': 3.223635482206852e-05, 'epoch': 2.05}) +[2026-06-17 23:45:34,253][starVLA.training.train_starvla][INFO] - Step 4519, Loss: {'train/loss': 0.46444806456565857, 'train/grad_norm_pre_clip': 23.348037719726562, 'timing/data': 0.013121099211275578, 'timing/model': 0.9452808573842049, 'learning_rate/qwen_vl_interface': 3.2216796706758757e-06, 'learning_rate/action_model': 3.221679670675876e-05, 'epoch': 2.05}) +[2026-06-17 23:45:38,226][starVLA.training.train_starvla][INFO] - Step 4520, Loss: {'train/loss': 0.2193441241979599, 'train/grad_norm_pre_clip': 31.865345001220703, 'timing/data': 0.009695236571133137, 'timing/model': 1.0616324814036489, 'learning_rate/qwen_vl_interface': 3.219724279617888e-06, 'learning_rate/action_model': 3.219724279617888e-05, 'epoch': 2.05}) +[2026-06-17 23:45:42,215][starVLA.training.train_starvla][INFO] - Step 4521, Loss: {'train/loss': 0.30541864037513733, 'train/grad_norm_pre_clip': 20.409481048583984, 'timing/data': 0.016480837017297745, 'timing/model': 0.9594602026045322, 'learning_rate/qwen_vl_interface': 3.2177693094382424e-06, 'learning_rate/action_model': 3.2177693094382424e-05, 'epoch': 2.05}) +[2026-06-17 23:45:47,625][starVLA.training.train_starvla][INFO] - Step 4522, Loss: {'train/loss': 0.23396068811416626, 'train/grad_norm_pre_clip': 21.111595153808594, 'timing/data': 0.008775115944445133, 'timing/model': 0.9509064005687833, 'learning_rate/qwen_vl_interface': 3.2158147605422052e-06, 'learning_rate/action_model': 3.215814760542205e-05, 'epoch': 2.05}) +[2026-06-17 23:45:51,499][starVLA.training.train_starvla][INFO] - Step 4523, Loss: {'train/loss': 0.20662246644496918, 'train/grad_norm_pre_clip': 15.667129516601562, 'timing/data': 0.011992909014225006, 'timing/model': 1.1566510470584035, 'learning_rate/qwen_vl_interface': 3.2138606333349576e-06, 'learning_rate/action_model': 3.213860633334957e-05, 'epoch': 2.05}) +[2026-06-17 23:45:55,752][starVLA.training.train_starvla][INFO] - Step 4524, Loss: {'train/loss': 0.13551107048988342, 'train/grad_norm_pre_clip': 15.468461990356445, 'timing/data': 0.019785975106060505, 'timing/model': 1.4098423328250647, 'learning_rate/qwen_vl_interface': 3.2119069282215913e-06, 'learning_rate/action_model': 3.211906928221591e-05, 'epoch': 2.05}) +[2026-06-17 23:45:59,402][starVLA.training.train_starvla][INFO] - Step 4525, Loss: {'train/loss': 0.46445348858833313, 'train/grad_norm_pre_clip': 42.042259216308594, 'timing/data': 0.00954368058592081, 'timing/model': 0.9343457976356149, 'learning_rate/qwen_vl_interface': 3.2099536456071118e-06, 'learning_rate/action_model': 3.209953645607112e-05, 'epoch': 2.05}) +[2026-06-17 23:46:03,291][starVLA.training.train_starvla][INFO] - Step 4526, Loss: {'train/loss': 0.2542439103126526, 'train/grad_norm_pre_clip': 25.680063247680664, 'timing/data': 0.011121156625449657, 'timing/model': 0.9559226520359516, 'learning_rate/qwen_vl_interface': 3.2080007858964347e-06, 'learning_rate/action_model': 3.2080007858964345e-05, 'epoch': 2.05}) +[2026-06-17 23:46:07,158][starVLA.training.train_starvla][INFO] - Step 4527, Loss: {'train/loss': 0.1037837490439415, 'train/grad_norm_pre_clip': 20.290067672729492, 'timing/data': 0.010720326565206051, 'timing/model': 0.954560867510736, 'learning_rate/qwen_vl_interface': 3.206048349494396e-06, 'learning_rate/action_model': 3.206048349494396e-05, 'epoch': 2.05}) +[2026-06-17 23:46:11,128][starVLA.training.train_starvla][INFO] - Step 4528, Loss: {'train/loss': 0.206363707780838, 'train/grad_norm_pre_clip': 18.854644775390625, 'timing/data': 0.008163888938724995, 'timing/model': 0.960628061555326, 'learning_rate/qwen_vl_interface': 3.2040963368057323e-06, 'learning_rate/action_model': 3.204096336805732e-05, 'epoch': 2.05}) +[2026-06-17 23:46:15,067][starVLA.training.train_starvla][INFO] - Step 4529, Loss: {'train/loss': 0.348650187253952, 'train/grad_norm_pre_clip': 61.165687561035156, 'timing/data': 0.009750897996127605, 'timing/model': 0.9477700218558311, 'learning_rate/qwen_vl_interface': 3.2021447482350994e-06, 'learning_rate/action_model': 3.202144748235099e-05, 'epoch': 2.05}) +[2026-06-17 23:46:18,996][starVLA.training.train_starvla][INFO] - Step 4530, Loss: {'train/loss': 0.4641471803188324, 'train/grad_norm_pre_clip': 32.3796272277832, 'timing/data': 0.009620997123420238, 'timing/model': 1.1486467588692904, 'learning_rate/qwen_vl_interface': 3.200193584187061e-06, 'learning_rate/action_model': 3.200193584187061e-05, 'epoch': 2.06}) +[2026-06-17 23:46:22,974][starVLA.training.train_starvla][INFO] - Step 4531, Loss: {'train/loss': 0.2354028970003128, 'train/grad_norm_pre_clip': 20.4438533782959, 'timing/data': 0.015642491169273853, 'timing/model': 1.143618573434651, 'learning_rate/qwen_vl_interface': 3.198242845066101e-06, 'learning_rate/action_model': 3.198242845066101e-05, 'epoch': 2.06}) +[2026-06-17 23:46:26,788][starVLA.training.train_starvla][INFO] - Step 4532, Loss: {'train/loss': 0.33091896772384644, 'train/grad_norm_pre_clip': 25.943138122558594, 'timing/data': 0.009381878189742565, 'timing/model': 1.0923026502132416, 'learning_rate/qwen_vl_interface': 3.1962925312766087e-06, 'learning_rate/action_model': 3.196292531276608e-05, 'epoch': 2.06}) +[2026-06-17 23:46:30,615][starVLA.training.train_starvla][INFO] - Step 4533, Loss: {'train/loss': 0.12462910264730453, 'train/grad_norm_pre_clip': 22.457883834838867, 'timing/data': 0.009099839255213737, 'timing/model': 0.9329364104196429, 'learning_rate/qwen_vl_interface': 3.19434264322288e-06, 'learning_rate/action_model': 3.19434264322288e-05, 'epoch': 2.06}) +[2026-06-17 23:46:36,050][starVLA.training.train_starvla][INFO] - Step 4534, Loss: {'train/loss': 0.45979633927345276, 'train/grad_norm_pre_clip': 27.7373104095459, 'timing/data': 0.026932465843856335, 'timing/model': 2.0966387381777167, 'learning_rate/qwen_vl_interface': 3.1923931813091364e-06, 'learning_rate/action_model': 3.192393181309136e-05, 'epoch': 2.06}) +[2026-06-17 23:46:40,045][starVLA.training.train_starvla][INFO] - Step 4535, Loss: {'train/loss': 0.3716789484024048, 'train/grad_norm_pre_clip': 43.27857208251953, 'timing/data': 0.016008751466870308, 'timing/model': 0.9631618214771152, 'learning_rate/qwen_vl_interface': 3.1904441459395e-06, 'learning_rate/action_model': 3.1904441459395e-05, 'epoch': 2.06}) +[2026-06-17 23:46:44,202][starVLA.training.train_starvla][INFO] - Step 4536, Loss: {'train/loss': 0.26150694489479065, 'train/grad_norm_pre_clip': 24.480945587158203, 'timing/data': 0.019970187917351723, 'timing/model': 1.0467424597591162, 'learning_rate/qwen_vl_interface': 3.1884955375180088e-06, 'learning_rate/action_model': 3.1884955375180085e-05, 'epoch': 2.06}) +[2026-06-17 23:46:49,214][starVLA.training.train_starvla][INFO] - Step 4537, Loss: {'train/loss': 0.32856735587120056, 'train/grad_norm_pre_clip': 30.571407318115234, 'timing/data': 0.022522962652146816, 'timing/model': 2.253399448469281, 'learning_rate/qwen_vl_interface': 3.18654735644861e-06, 'learning_rate/action_model': 3.1865473564486095e-05, 'epoch': 2.06}) +[2026-06-17 23:46:52,929][starVLA.training.train_starvla][INFO] - Step 4538, Loss: {'train/loss': 0.12363354116678238, 'train/grad_norm_pre_clip': 14.144601821899414, 'timing/data': 0.019523894414305687, 'timing/model': 0.9728449117392302, 'learning_rate/qwen_vl_interface': 3.184599603135165e-06, 'learning_rate/action_model': 3.1845996031351646e-05, 'epoch': 2.06}) +[2026-06-17 23:46:57,038][starVLA.training.train_starvla][INFO] - Step 4539, Loss: {'train/loss': 0.3693304657936096, 'train/grad_norm_pre_clip': 25.69264030456543, 'timing/data': 0.03202309925109148, 'timing/model': 0.9610612327232957, 'learning_rate/qwen_vl_interface': 3.1826522779814453e-06, 'learning_rate/action_model': 3.182652277981445e-05, 'epoch': 2.06}) +[2026-06-17 23:47:00,931][starVLA.training.train_starvla][INFO] - Step 4540, Loss: {'train/loss': 0.15522553026676178, 'train/grad_norm_pre_clip': 19.738706588745117, 'timing/data': 0.018871246837079525, 'timing/model': 0.9614571761339903, 'learning_rate/qwen_vl_interface': 3.1807053813911317e-06, 'learning_rate/action_model': 3.180705381391132e-05, 'epoch': 2.06}) +[2026-06-17 23:47:04,891][starVLA.training.train_starvla][INFO] - Step 4541, Loss: {'train/loss': 0.24092116951942444, 'train/grad_norm_pre_clip': 18.890172958374023, 'timing/data': 0.023191210813820362, 'timing/model': 0.9826133623719215, 'learning_rate/qwen_vl_interface': 3.17875891376782e-06, 'learning_rate/action_model': 3.1787589137678196e-05, 'epoch': 2.06}) +[2026-06-17 23:47:11,078][starVLA.training.train_starvla][INFO] - Step 4542, Loss: {'train/loss': 0.19380314648151398, 'train/grad_norm_pre_clip': 20.4599609375, 'timing/data': 0.017237809486687183, 'timing/model': 0.9991388078778982, 'learning_rate/qwen_vl_interface': 3.176812875515014e-06, 'learning_rate/action_model': 3.176812875515014e-05, 'epoch': 2.06}) +[2026-06-17 23:47:16,400][starVLA.training.train_starvla][INFO] - Step 4543, Loss: {'train/loss': 0.2731019854545593, 'train/grad_norm_pre_clip': 20.15638542175293, 'timing/data': 0.01889631524682045, 'timing/model': 2.3628311855718493, 'learning_rate/qwen_vl_interface': 3.1748672670361304e-06, 'learning_rate/action_model': 3.17486726703613e-05, 'epoch': 2.06}) +[2026-06-17 23:47:20,416][starVLA.training.train_starvla][INFO] - Step 4544, Loss: {'train/loss': 0.32569408416748047, 'train/grad_norm_pre_clip': 20.99640655517578, 'timing/data': 0.023073554039001465, 'timing/model': 1.1704809721559286, 'learning_rate/qwen_vl_interface': 3.172922088734493e-06, 'learning_rate/action_model': 3.172922088734493e-05, 'epoch': 2.06}) +[2026-06-17 23:47:24,191][starVLA.training.train_starvla][INFO] - Step 4545, Loss: {'train/loss': 0.2533283531665802, 'train/grad_norm_pre_clip': 22.821889877319336, 'timing/data': 0.025310573168098927, 'timing/model': 0.9704237831756473, 'learning_rate/qwen_vl_interface': 3.1709773410133434e-06, 'learning_rate/action_model': 3.170977341013343e-05, 'epoch': 2.06}) +[2026-06-17 23:47:28,261][starVLA.training.train_starvla][INFO] - Step 4546, Loss: {'train/loss': 0.2957925498485565, 'train/grad_norm_pre_clip': 31.200437545776367, 'timing/data': 0.017898159101605415, 'timing/model': 0.9448066763579845, 'learning_rate/qwen_vl_interface': 3.16903302427583e-06, 'learning_rate/action_model': 3.16903302427583e-05, 'epoch': 2.06}) +[2026-06-17 23:47:32,185][starVLA.training.train_starvla][INFO] - Step 4547, Loss: {'train/loss': 0.4429544508457184, 'train/grad_norm_pre_clip': 23.18252182006836, 'timing/data': 0.023670942522585392, 'timing/model': 0.9645705195143819, 'learning_rate/qwen_vl_interface': 3.1670891389250073e-06, 'learning_rate/action_model': 3.167089138925007e-05, 'epoch': 2.06}) +[2026-06-17 23:47:36,162][starVLA.training.train_starvla][INFO] - Step 4548, Loss: {'train/loss': 0.14792224764823914, 'train/grad_norm_pre_clip': 15.772923469543457, 'timing/data': 0.018398351967334747, 'timing/model': 0.9702083496376872, 'learning_rate/qwen_vl_interface': 3.1651456853638464e-06, 'learning_rate/action_model': 3.165145685363846e-05, 'epoch': 2.06}) +[2026-06-17 23:47:40,119][starVLA.training.train_starvla][INFO] - Step 4549, Loss: {'train/loss': 0.21225886046886444, 'train/grad_norm_pre_clip': 18.66901969909668, 'timing/data': 0.017195110209286213, 'timing/model': 0.9547769064083695, 'learning_rate/qwen_vl_interface': 3.1632026639952303e-06, 'learning_rate/action_model': 3.16320266399523e-05, 'epoch': 2.06}) +[2026-06-17 23:47:44,024][starVLA.training.train_starvla][INFO] - Step 4550, Loss: {'train/loss': 0.31413012742996216, 'train/grad_norm_pre_clip': 19.866544723510742, 'timing/data': 0.017636152915656567, 'timing/model': 1.1200851807370782, 'learning_rate/qwen_vl_interface': 3.161260075221947e-06, 'learning_rate/action_model': 3.161260075221947e-05, 'epoch': 2.06}) +[2026-06-17 23:47:47,993][starVLA.training.train_starvla][INFO] - Step 4551, Loss: {'train/loss': 0.43043747544288635, 'train/grad_norm_pre_clip': 30.449193954467773, 'timing/data': 0.015513960272073746, 'timing/model': 1.237486643716693, 'learning_rate/qwen_vl_interface': 3.1593179194467e-06, 'learning_rate/action_model': 3.1593179194467e-05, 'epoch': 2.06}) +[2026-06-17 23:47:51,683][starVLA.training.train_starvla][INFO] - Step 4552, Loss: {'train/loss': 0.1650143265724182, 'train/grad_norm_pre_clip': 24.477649688720703, 'timing/data': 0.016730330884456635, 'timing/model': 0.9656080864369869, 'learning_rate/qwen_vl_interface': 3.157376197072094e-06, 'learning_rate/action_model': 3.157376197072094e-05, 'epoch': 2.07}) +[2026-06-17 23:47:55,981][starVLA.training.train_starvla][INFO] - Step 4553, Loss: {'train/loss': 0.1976696252822876, 'train/grad_norm_pre_clip': 21.907367706298828, 'timing/data': 0.02638186141848564, 'timing/model': 0.9909845208749175, 'learning_rate/qwen_vl_interface': 3.1554349085006557e-06, 'learning_rate/action_model': 3.155434908500656e-05, 'epoch': 2.07}) +[2026-06-17 23:47:59,871][starVLA.training.train_starvla][INFO] - Step 4554, Loss: {'train/loss': 0.15186423063278198, 'train/grad_norm_pre_clip': 17.314603805541992, 'timing/data': 0.021132181398570538, 'timing/model': 0.9487914592027664, 'learning_rate/qwen_vl_interface': 3.1534940541348152e-06, 'learning_rate/action_model': 3.153494054134815e-05, 'epoch': 2.07}) +[2026-06-17 23:48:03,872][starVLA.training.train_starvla][INFO] - Step 4555, Loss: {'train/loss': 0.18953941762447357, 'train/grad_norm_pre_clip': 17.240970611572266, 'timing/data': 0.0202512564137578, 'timing/model': 0.9637902304530144, 'learning_rate/qwen_vl_interface': 3.1515536343769133e-06, 'learning_rate/action_model': 3.151553634376913e-05, 'epoch': 2.07}) +[2026-06-17 23:48:07,857][starVLA.training.train_starvla][INFO] - Step 4556, Loss: {'train/loss': 0.2328486293554306, 'train/grad_norm_pre_clip': 21.9031982421875, 'timing/data': 0.02287864312529564, 'timing/model': 0.9878409830853343, 'learning_rate/qwen_vl_interface': 3.1496136496292e-06, 'learning_rate/action_model': 3.1496136496292e-05, 'epoch': 2.07}) +[2026-06-17 23:48:12,493][starVLA.training.train_starvla][INFO] - Step 4557, Loss: {'train/loss': 0.3714316487312317, 'train/grad_norm_pre_clip': 32.760337829589844, 'timing/data': 0.020433999598026276, 'timing/model': 1.1574889365583658, 'learning_rate/qwen_vl_interface': 3.1476741002938383e-06, 'learning_rate/action_model': 3.147674100293838e-05, 'epoch': 2.07}) +[2026-06-17 23:48:16,645][starVLA.training.train_starvla][INFO] - Step 4558, Loss: {'train/loss': 0.1973765790462494, 'train/grad_norm_pre_clip': 20.018383026123047, 'timing/data': 0.015944601967930794, 'timing/model': 1.3816365227103233, 'learning_rate/qwen_vl_interface': 3.145734986772897e-06, 'learning_rate/action_model': 3.1457349867728966e-05, 'epoch': 2.07}) +[2026-06-17 23:48:20,342][starVLA.training.train_starvla][INFO] - Step 4559, Loss: {'train/loss': 0.26793715357780457, 'train/grad_norm_pre_clip': 19.363658905029297, 'timing/data': 0.026538945734500885, 'timing/model': 0.9450527103617787, 'learning_rate/qwen_vl_interface': 3.1437963094683566e-06, 'learning_rate/action_model': 3.1437963094683564e-05, 'epoch': 2.07}) +[2026-06-17 23:48:24,320][starVLA.training.train_starvla][INFO] - Step 4560, Loss: {'train/loss': 0.09060429036617279, 'train/grad_norm_pre_clip': 19.019073486328125, 'timing/data': 0.025996445678174496, 'timing/model': 0.9607375245541334, 'learning_rate/qwen_vl_interface': 3.1418580687821103e-06, 'learning_rate/action_model': 3.14185806878211e-05, 'epoch': 2.07}) +[2026-06-17 23:48:28,322][starVLA.training.train_starvla][INFO] - Step 4561, Loss: {'train/loss': 0.4408656358718872, 'train/grad_norm_pre_clip': 33.5413818359375, 'timing/data': 0.019839276559650898, 'timing/model': 0.9663854325190187, 'learning_rate/qwen_vl_interface': 3.1399202651159532e-06, 'learning_rate/action_model': 3.1399202651159534e-05, 'epoch': 2.07}) +[2026-06-17 23:48:32,300][starVLA.training.train_starvla][INFO] - Step 4562, Loss: {'train/loss': 0.2815626561641693, 'train/grad_norm_pre_clip': 18.46245574951172, 'timing/data': 0.018034356646239758, 'timing/model': 0.947011997923255, 'learning_rate/qwen_vl_interface': 3.1379828988715955e-06, 'learning_rate/action_model': 3.1379828988715956e-05, 'epoch': 2.07}) +[2026-06-17 23:48:36,265][starVLA.training.train_starvla][INFO] - Step 4563, Loss: {'train/loss': 0.19170616567134857, 'train/grad_norm_pre_clip': 25.716358184814453, 'timing/data': 0.01855235267430544, 'timing/model': 0.9514985699206591, 'learning_rate/qwen_vl_interface': 3.136045970450654e-06, 'learning_rate/action_model': 3.136045970450654e-05, 'epoch': 2.07}) +[2026-06-17 23:48:40,310][starVLA.training.train_starvla][INFO] - Step 4564, Loss: {'train/loss': 0.24698680639266968, 'train/grad_norm_pre_clip': 24.06540870666504, 'timing/data': 0.02539294771850109, 'timing/model': 1.1438452368602157, 'learning_rate/qwen_vl_interface': 3.1341094802546595e-06, 'learning_rate/action_model': 3.1341094802546595e-05, 'epoch': 2.07}) +[2026-06-17 23:48:43,979][starVLA.training.train_starvla][INFO] - Step 4565, Loss: {'train/loss': 0.1895410418510437, 'train/grad_norm_pre_clip': 16.273216247558594, 'timing/data': 0.017532261088490486, 'timing/model': 0.9528245907276869, 'learning_rate/qwen_vl_interface': 3.132173428685048e-06, 'learning_rate/action_model': 3.132173428685048e-05, 'epoch': 2.07}) +[2026-06-17 23:48:47,819][starVLA.training.train_starvla][INFO] - Step 4566, Loss: {'train/loss': 0.1781589537858963, 'train/grad_norm_pre_clip': 16.65346336364746, 'timing/data': 0.021680930629372597, 'timing/model': 0.9398420481011271, 'learning_rate/qwen_vl_interface': 3.13023781614316e-06, 'learning_rate/action_model': 3.13023781614316e-05, 'epoch': 2.07}) +[2026-06-17 23:48:51,577][starVLA.training.train_starvla][INFO] - Step 4567, Loss: {'train/loss': 0.22931678593158722, 'train/grad_norm_pre_clip': 14.201700210571289, 'timing/data': 0.023429890163242817, 'timing/model': 0.9371315706521273, 'learning_rate/qwen_vl_interface': 3.1283026430302553e-06, 'learning_rate/action_model': 3.128302643030255e-05, 'epoch': 2.07}) +[2026-06-17 23:48:55,430][starVLA.training.train_starvla][INFO] - Step 4568, Loss: {'train/loss': 0.30929017066955566, 'train/grad_norm_pre_clip': 22.089759826660156, 'timing/data': 0.02850531041622162, 'timing/model': 0.9475468946620822, 'learning_rate/qwen_vl_interface': 3.1263679097474958e-06, 'learning_rate/action_model': 3.126367909747496e-05, 'epoch': 2.07}) +[2026-06-17 23:48:59,225][starVLA.training.train_starvla][INFO] - Step 4569, Loss: {'train/loss': 0.1998545527458191, 'train/grad_norm_pre_clip': 18.452592849731445, 'timing/data': 0.016559856943786144, 'timing/model': 0.945498501881957, 'learning_rate/qwen_vl_interface': 3.124433616695955e-06, 'learning_rate/action_model': 3.1244336166959545e-05, 'epoch': 2.07}) +[2026-06-17 23:49:03,256][starVLA.training.train_starvla][INFO] - Step 4570, Loss: {'train/loss': 0.4610975384712219, 'train/grad_norm_pre_clip': 30.994787216186523, 'timing/data': 0.0088709881529212, 'timing/model': 1.156352112069726, 'learning_rate/qwen_vl_interface': 3.1224997642766115e-06, 'learning_rate/action_model': 3.122499764276611e-05, 'epoch': 2.07}) +[2026-06-17 23:49:07,330][starVLA.training.train_starvla][INFO] - Step 4571, Loss: {'train/loss': 0.19304554164409637, 'train/grad_norm_pre_clip': 26.344274520874023, 'timing/data': 0.009644724428653717, 'timing/model': 1.1433291248977184, 'learning_rate/qwen_vl_interface': 3.1205663528903563e-06, 'learning_rate/action_model': 3.1205663528903565e-05, 'epoch': 2.07}) +[2026-06-17 23:49:11,143][starVLA.training.train_starvla][INFO] - Step 4572, Loss: {'train/loss': 0.3504088222980499, 'train/grad_norm_pre_clip': 25.644655227661133, 'timing/data': 0.011129521764814854, 'timing/model': 0.9638156341388822, 'learning_rate/qwen_vl_interface': 3.118633382937987e-06, 'learning_rate/action_model': 3.118633382937987e-05, 'epoch': 2.07}) +[2026-06-17 23:49:15,041][starVLA.training.train_starvla][INFO] - Step 4573, Loss: {'train/loss': 0.38625767827033997, 'train/grad_norm_pre_clip': 21.72886085510254, 'timing/data': 0.013622848317027092, 'timing/model': 0.9383174479007721, 'learning_rate/qwen_vl_interface': 3.116700854820209e-06, 'learning_rate/action_model': 3.116700854820209e-05, 'epoch': 2.07}) +[2026-06-17 23:49:18,962][starVLA.training.train_starvla][INFO] - Step 4574, Loss: {'train/loss': 0.12340458482503891, 'train/grad_norm_pre_clip': 16.23758888244629, 'timing/data': 0.014785612933337688, 'timing/model': 0.965926113538444, 'learning_rate/qwen_vl_interface': 3.1147687689376443e-06, 'learning_rate/action_model': 3.1147687689376444e-05, 'epoch': 2.08}) +[2026-06-17 23:49:22,914][starVLA.training.train_starvla][INFO] - Step 4575, Loss: {'train/loss': 0.46849361062049866, 'train/grad_norm_pre_clip': 35.80012512207031, 'timing/data': 0.008928393013775349, 'timing/model': 0.9606258533895016, 'learning_rate/qwen_vl_interface': 3.1128371256908087e-06, 'learning_rate/action_model': 3.112837125690809e-05, 'epoch': 2.08}) +[2026-06-17 23:49:26,810][starVLA.training.train_starvla][INFO] - Step 4576, Loss: {'train/loss': 0.1682378649711609, 'train/grad_norm_pre_clip': 17.211414337158203, 'timing/data': 0.010192854329943657, 'timing/model': 0.9831881634891033, 'learning_rate/qwen_vl_interface': 3.1109059254801368e-06, 'learning_rate/action_model': 3.1109059254801366e-05, 'epoch': 2.08}) +[2026-06-17 23:49:30,677][starVLA.training.train_starvla][INFO] - Step 4577, Loss: {'train/loss': 0.2228112816810608, 'train/grad_norm_pre_clip': 33.10048294067383, 'timing/data': 0.009455010294914246, 'timing/model': 1.1525274803861976, 'learning_rate/qwen_vl_interface': 3.108975168705966e-06, 'learning_rate/action_model': 3.1089751687059656e-05, 'epoch': 2.08}) +[2026-06-17 23:49:34,506][starVLA.training.train_starvla][INFO] - Step 4578, Loss: {'train/loss': 0.18021877110004425, 'train/grad_norm_pre_clip': 16.484634399414062, 'timing/data': 0.008572679944336414, 'timing/model': 1.0981671251356602, 'learning_rate/qwen_vl_interface': 3.107044855768549e-06, 'learning_rate/action_model': 3.107044855768549e-05, 'epoch': 2.08}) +[2026-06-17 23:49:38,323][starVLA.training.train_starvla][INFO] - Step 4579, Loss: {'train/loss': 0.3420943021774292, 'train/grad_norm_pre_clip': 32.241127014160156, 'timing/data': 0.0112704336643219, 'timing/model': 0.9691963847726583, 'learning_rate/qwen_vl_interface': 3.1051149870680413e-06, 'learning_rate/action_model': 3.105114987068041e-05, 'epoch': 2.08}) +[2026-06-17 23:49:42,182][starVLA.training.train_starvla][INFO] - Step 4580, Loss: {'train/loss': 0.38609960675239563, 'train/grad_norm_pre_clip': 25.660612106323242, 'timing/data': 0.012055772356688976, 'timing/model': 0.9481392614543438, 'learning_rate/qwen_vl_interface': 3.103185563004503e-06, 'learning_rate/action_model': 3.103185563004503e-05, 'epoch': 2.08}) +[2026-06-17 23:49:46,100][starVLA.training.train_starvla][INFO] - Step 4581, Loss: {'train/loss': 0.2841636538505554, 'train/grad_norm_pre_clip': 26.408172607421875, 'timing/data': 0.009039566852152348, 'timing/model': 0.9527649348601699, 'learning_rate/qwen_vl_interface': 3.101256583977905e-06, 'learning_rate/action_model': 3.101256583977905e-05, 'epoch': 2.08}) +[2026-06-17 23:49:49,951][starVLA.training.train_starvla][INFO] - Step 4582, Loss: {'train/loss': 0.05927680805325508, 'train/grad_norm_pre_clip': 10.380899429321289, 'timing/data': 0.007807228714227676, 'timing/model': 0.9420726653188467, 'learning_rate/qwen_vl_interface': 3.099328050388131e-06, 'learning_rate/action_model': 3.0993280503881305e-05, 'epoch': 2.08}) +[2026-06-17 23:49:53,901][starVLA.training.train_starvla][INFO] - Step 4583, Loss: {'train/loss': 0.49597153067588806, 'train/grad_norm_pre_clip': 32.02859878540039, 'timing/data': 0.008792340755462646, 'timing/model': 0.9568710308521986, 'learning_rate/qwen_vl_interface': 3.0973999626349667e-06, 'learning_rate/action_model': 3.097399962634966e-05, 'epoch': 2.08}) +[2026-06-17 23:49:57,924][starVLA.training.train_starvla][INFO] - Step 4584, Loss: {'train/loss': 0.0583980418741703, 'train/grad_norm_pre_clip': 7.894381523132324, 'timing/data': 0.009108892641961575, 'timing/model': 1.1784450765699148, 'learning_rate/qwen_vl_interface': 3.095472321118107e-06, 'learning_rate/action_model': 3.095472321118107e-05, 'epoch': 2.08}) +[2026-06-17 23:50:01,906][starVLA.training.train_starvla][INFO] - Step 4585, Loss: {'train/loss': 0.15685701370239258, 'train/grad_norm_pre_clip': 21.90180206298828, 'timing/data': 0.01249656267464161, 'timing/model': 1.2602676004171371, 'learning_rate/qwen_vl_interface': 3.093545126237153e-06, 'learning_rate/action_model': 3.0935451262371527e-05, 'epoch': 2.08}) +[2026-06-17 23:50:06,064][starVLA.training.train_starvla][INFO] - Step 4586, Loss: {'train/loss': 0.092656210064888, 'train/grad_norm_pre_clip': 15.153395652770996, 'timing/data': 0.009695746935904026, 'timing/model': 0.9610415371134877, 'learning_rate/qwen_vl_interface': 3.091618378391615e-06, 'learning_rate/action_model': 3.0916183783916146e-05, 'epoch': 2.08}) +[2026-06-17 23:50:10,021][starVLA.training.train_starvla][INFO] - Step 4587, Loss: {'train/loss': 0.05691460892558098, 'train/grad_norm_pre_clip': 7.300962448120117, 'timing/data': 0.010894197039306164, 'timing/model': 0.9630350256338716, 'learning_rate/qwen_vl_interface': 3.0896920779809097e-06, 'learning_rate/action_model': 3.08969207798091e-05, 'epoch': 2.08}) +[2026-06-17 23:50:14,027][starVLA.training.train_starvla][INFO] - Step 4588, Loss: {'train/loss': 0.13844723999500275, 'train/grad_norm_pre_clip': 16.58260726928711, 'timing/data': 0.009113756939768791, 'timing/model': 0.968313685618341, 'learning_rate/qwen_vl_interface': 3.0877662254043605e-06, 'learning_rate/action_model': 3.0877662254043606e-05, 'epoch': 2.08}) +[2026-06-17 23:50:17,868][starVLA.training.train_starvla][INFO] - Step 4589, Loss: {'train/loss': 0.3974626660346985, 'train/grad_norm_pre_clip': 27.641315460205078, 'timing/data': 0.0068482402712106705, 'timing/model': 0.9427727805450559, 'learning_rate/qwen_vl_interface': 3.0858408210612e-06, 'learning_rate/action_model': 3.0858408210612e-05, 'epoch': 2.08}) +[2026-06-17 23:50:21,726][starVLA.training.train_starvla][INFO] - Step 4590, Loss: {'train/loss': 0.3291302025318146, 'train/grad_norm_pre_clip': 41.886131286621094, 'timing/data': 0.011886416003108025, 'timing/model': 0.9517740570008755, 'learning_rate/qwen_vl_interface': 3.0839158653505663e-06, 'learning_rate/action_model': 3.0839158653505664e-05, 'epoch': 2.08}) +[2026-06-17 23:50:25,556][starVLA.training.train_starvla][INFO] - Step 4591, Loss: {'train/loss': 0.05804279446601868, 'train/grad_norm_pre_clip': 15.764199256896973, 'timing/data': 0.00686352513730526, 'timing/model': 1.115308117121458, 'learning_rate/qwen_vl_interface': 3.081991358671504e-06, 'learning_rate/action_model': 3.081991358671504e-05, 'epoch': 2.08}) +[2026-06-17 23:50:29,206][starVLA.training.train_starvla][INFO] - Step 4592, Loss: {'train/loss': 0.6168422102928162, 'train/grad_norm_pre_clip': 35.417240142822266, 'timing/data': 0.010160905309021473, 'timing/model': 0.958809033036232, 'learning_rate/qwen_vl_interface': 3.080067301422965e-06, 'learning_rate/action_model': 3.080067301422965e-05, 'epoch': 2.08}) +[2026-06-17 23:50:33,003][starVLA.training.train_starvla][INFO] - Step 4593, Loss: {'train/loss': 0.18228265643119812, 'train/grad_norm_pre_clip': 24.126232147216797, 'timing/data': 0.007894081994891167, 'timing/model': 0.9391282815486193, 'learning_rate/qwen_vl_interface': 3.07814369400381e-06, 'learning_rate/action_model': 3.07814369400381e-05, 'epoch': 2.08}) +[2026-06-17 23:50:36,848][starVLA.training.train_starvla][INFO] - Step 4594, Loss: {'train/loss': 0.41997551918029785, 'train/grad_norm_pre_clip': 39.68946075439453, 'timing/data': 0.014854883775115013, 'timing/model': 0.9775335108861327, 'learning_rate/qwen_vl_interface': 3.076220536812807e-06, 'learning_rate/action_model': 3.076220536812807e-05, 'epoch': 2.08}) +[2026-06-17 23:50:40,962][starVLA.training.train_starvla][INFO] - Step 4595, Loss: {'train/loss': 0.11174941062927246, 'train/grad_norm_pre_clip': 15.345760345458984, 'timing/data': 0.015498743392527103, 'timing/model': 0.9935524007305503, 'learning_rate/qwen_vl_interface': 3.0742978302486214e-06, 'learning_rate/action_model': 3.074297830248621e-05, 'epoch': 2.08}) +[2026-06-17 23:50:49,626][starVLA.training.train_starvla][INFO] - Step 4596, Loss: {'train/loss': 0.19450680911540985, 'train/grad_norm_pre_clip': 12.331683158874512, 'timing/data': 0.014787053689360619, 'timing/model': 4.135022853501141, 'learning_rate/qwen_vl_interface': 3.0723755747098387e-06, 'learning_rate/action_model': 3.072375574709839e-05, 'epoch': 2.09}) +[2026-06-17 23:50:53,713][starVLA.training.train_starvla][INFO] - Step 4597, Loss: {'train/loss': 0.5067481398582458, 'train/grad_norm_pre_clip': 36.59172821044922, 'timing/data': 0.00989613588899374, 'timing/model': 1.25390759203583, 'learning_rate/qwen_vl_interface': 3.070453770594943e-06, 'learning_rate/action_model': 3.070453770594943e-05, 'epoch': 2.09}) +[2026-06-17 23:50:57,704][starVLA.training.train_starvla][INFO] - Step 4598, Loss: {'train/loss': 0.39499178528785706, 'train/grad_norm_pre_clip': 29.576326370239258, 'timing/data': 0.0092768594622612, 'timing/model': 1.1437428267672658, 'learning_rate/qwen_vl_interface': 3.0685324183023264e-06, 'learning_rate/action_model': 3.068532418302326e-05, 'epoch': 2.09}) +[2026-06-17 23:51:02,496][starVLA.training.train_starvla][INFO] - Step 4599, Loss: {'train/loss': 0.13275901973247528, 'train/grad_norm_pre_clip': 15.534775733947754, 'timing/data': 0.012182313948869705, 'timing/model': 0.94613391533494, 'learning_rate/qwen_vl_interface': 3.0666115182302863e-06, 'learning_rate/action_model': 3.0666115182302865e-05, 'epoch': 2.09}) +[2026-06-17 23:53:06,144][starVLA.training.train_starvla][INFO] - Step 4600, Loss: {'train/loss': 0.42135006189346313, 'train/grad_norm_pre_clip': 21.005521774291992, 'eval/loss': 0.28547972440719604, 'eval/latency_8/loss': 0.28547972440719604, 'eval/demon_fire_precision': 0.9678407350689127, 'eval/demon_fire_recall': 0.9816713264989128, 'eval/demon_fire_f1': 0.9747069710055521, 'eval/demon_fire_rate_model': 0.9084585420144685, 'eval/demon_fire_rate_teacher': 0.8956594323873122, 'eval/demon_left_precision': 0.8855721393034826, 'eval/demon_left_recall': 0.7623126338329764, 'eval/demon_left_f1': 0.8193325661680092, 'eval/demon_left_rate_model': 0.11185308848080133, 'eval/demon_left_rate_teacher': 0.12993878686700056, 'eval/demon_right_precision': 0.8600953895071543, 'eval/demon_right_recall': 0.8196969696969697, 'eval/demon_right_f1': 0.8394103956555469, 'eval/demon_right_rate_model': 0.1750139120756817, 'eval/demon_right_rate_teacher': 0.18363939899833054, 'eval/demon_move_f1': 0.829371480911778, 'eval/demon_cc_f1': 0.9260251709359877, 'eval/latency_8/demon_fire_precision': 0.9678407350689127, 'eval/latency_8/demon_fire_recall': 0.9816713264989128, 'eval/latency_8/demon_fire_f1': 0.9747069710055521, 'eval/latency_8/demon_fire_rate_model': 0.9084585420144685, 'eval/latency_8/demon_fire_rate_teacher': 0.8956594323873122, 'eval/latency_8/demon_left_precision': 0.8855721393034826, 'eval/latency_8/demon_left_recall': 0.7623126338329764, 'eval/latency_8/demon_left_f1': 0.8193325661680092, 'eval/latency_8/demon_left_rate_model': 0.11185308848080133, 'eval/latency_8/demon_left_rate_teacher': 0.12993878686700056, 'eval/latency_8/demon_right_precision': 0.8600953895071543, 'eval/latency_8/demon_right_recall': 0.8196969696969697, 'eval/latency_8/demon_right_f1': 0.8394103956555469, 'eval/latency_8/demon_right_rate_model': 0.1750139120756817, 'eval/latency_8/demon_right_rate_teacher': 0.18363939899833054, 'eval/latency_8/demon_move_f1': 0.829371480911778, 'eval/latency_8/demon_cc_f1': 0.9260251709359877, 'timing/data': 0.008210527710616589, 'timing/model': 2.277166100218892, 'learning_rate/qwen_vl_interface': 3.064691070777027e-06, 'learning_rate/action_model': 3.064691070777027e-05, 'epoch': 2.09}) +[2026-06-17 23:53:06,148][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_4600_state +[2026-06-17 23:53:06,153][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 23:53:12,945][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_4600_state/model.safetensors +[2026-06-17 23:53:28,163][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_4600_state/optimizer.bin +[2026-06-17 23:53:28,164][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_4600_state/sampler.bin +[2026-06-17 23:53:28,164][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_4600_state/sampler_1.bin +[2026-06-17 23:53:28,166][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_4600_state/random_states_0.pkl +[2026-06-17 23:53:28,169][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4600 +[2026-06-17 23:53:28,170][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 23:53:28,181][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 23:53:32,008][starVLA.training.train_starvla][INFO] - Step 4601, Loss: {'train/loss': 0.41490814089775085, 'train/grad_norm_pre_clip': 33.76096725463867, 'timing/data': 0.017201798036694527, 'timing/model': 0.9600338703021407, 'learning_rate/qwen_vl_interface': 3.062771076340662e-06, 'learning_rate/action_model': 3.0627710763406617e-05, 'epoch': 2.09}) +[2026-06-17 23:53:35,980][starVLA.training.train_starvla][INFO] - Step 4602, Loss: {'train/loss': 0.28700587153434753, 'train/grad_norm_pre_clip': 23.99867057800293, 'timing/data': 0.014936631545424461, 'timing/model': 0.9685440426692367, 'learning_rate/qwen_vl_interface': 3.0608515353192046e-06, 'learning_rate/action_model': 3.060851535319205e-05, 'epoch': 2.09}) +[2026-06-17 23:53:39,937][starVLA.training.train_starvla][INFO] - Step 4603, Loss: {'train/loss': 0.1711050271987915, 'train/grad_norm_pre_clip': 16.727516174316406, 'timing/data': 0.010805884376168251, 'timing/model': 0.9802919588983059, 'learning_rate/qwen_vl_interface': 3.0589324481105814e-06, 'learning_rate/action_model': 3.058932448110581e-05, 'epoch': 2.09}) +[2026-06-17 23:53:43,984][starVLA.training.train_starvla][INFO] - Step 4604, Loss: {'train/loss': 0.3395780324935913, 'train/grad_norm_pre_clip': 29.51945686340332, 'timing/data': 0.009160024113953114, 'timing/model': 0.9892856106162071, 'learning_rate/qwen_vl_interface': 3.0570138151126185e-06, 'learning_rate/action_model': 3.057013815112618e-05, 'epoch': 2.09}) +[2026-06-17 23:53:48,709][starVLA.training.train_starvla][INFO] - Step 4605, Loss: {'train/loss': 0.2513740360736847, 'train/grad_norm_pre_clip': 26.35675048828125, 'timing/data': 0.0007928693667054176, 'timing/model': 1.0762317040935159, 'learning_rate/qwen_vl_interface': 3.0550956367230506e-06, 'learning_rate/action_model': 3.05509563672305e-05, 'epoch': 2.09}) +[2026-06-17 23:53:52,783][starVLA.training.train_starvla][INFO] - Step 4606, Loss: {'train/loss': 0.21435967087745667, 'train/grad_norm_pre_clip': 29.84510040283203, 'timing/data': 0.0005920166149735451, 'timing/model': 1.2342910170555115, 'learning_rate/qwen_vl_interface': 3.0531779133395184e-06, 'learning_rate/action_model': 3.053177913339518e-05, 'epoch': 2.09}) +[2026-06-17 23:53:56,930][starVLA.training.train_starvla][INFO] - Step 4607, Loss: {'train/loss': 0.3911114037036896, 'train/grad_norm_pre_clip': 42.64923095703125, 'timing/data': 0.0005579832941293716, 'timing/model': 1.1900734612718225, 'learning_rate/qwen_vl_interface': 3.0512606453595708e-06, 'learning_rate/action_model': 3.0512606453595704e-05, 'epoch': 2.09}) +[2026-06-17 23:54:00,640][starVLA.training.train_starvla][INFO] - Step 4608, Loss: {'train/loss': 0.22793613374233246, 'train/grad_norm_pre_clip': 13.494490623474121, 'timing/data': 0.0005880966782569885, 'timing/model': 0.9674501046538353, 'learning_rate/qwen_vl_interface': 3.049343833180658e-06, 'learning_rate/action_model': 3.049343833180658e-05, 'epoch': 2.09}) +[2026-06-17 23:54:04,553][starVLA.training.train_starvla][INFO] - Step 4609, Loss: {'train/loss': 0.15251292288303375, 'train/grad_norm_pre_clip': 17.00456428527832, 'timing/data': 0.0009008655324578285, 'timing/model': 0.9519309150055051, 'learning_rate/qwen_vl_interface': 3.0474274772001355e-06, 'learning_rate/action_model': 3.0474274772001355e-05, 'epoch': 2.09}) +[2026-06-17 23:54:08,458][starVLA.training.train_starvla][INFO] - Step 4610, Loss: {'train/loss': 0.1827995777130127, 'train/grad_norm_pre_clip': 21.704309463500977, 'timing/data': 0.000986817292869091, 'timing/model': 0.9583196602761745, 'learning_rate/qwen_vl_interface': 3.045511577815265e-06, 'learning_rate/action_model': 3.045511577815265e-05, 'epoch': 2.09}) +[2026-06-17 23:54:12,400][starVLA.training.train_starvla][INFO] - Step 4611, Loss: {'train/loss': 0.5080304145812988, 'train/grad_norm_pre_clip': 32.697628021240234, 'timing/data': 0.0006399219855666161, 'timing/model': 0.9581930637359619, 'learning_rate/qwen_vl_interface': 3.043596135423218e-06, 'learning_rate/action_model': 3.043596135423218e-05, 'epoch': 2.09}) +[2026-06-17 23:54:16,270][starVLA.training.train_starvla][INFO] - Step 4612, Loss: {'train/loss': 0.4059840142726898, 'train/grad_norm_pre_clip': 27.348661422729492, 'timing/data': 0.0005619414150714874, 'timing/model': 0.9946331866085529, 'learning_rate/qwen_vl_interface': 3.041681150421067e-06, 'learning_rate/action_model': 3.041681150421067e-05, 'epoch': 2.09}) +[2026-06-17 23:54:20,157][starVLA.training.train_starvla][INFO] - Step 4613, Loss: {'train/loss': 0.10731904953718185, 'train/grad_norm_pre_clip': 19.85016441345215, 'timing/data': 0.0005331886932253838, 'timing/model': 1.1425238503143191, 'learning_rate/qwen_vl_interface': 3.039766623205791e-06, 'learning_rate/action_model': 3.039766623205791e-05, 'epoch': 2.09}) +[2026-06-17 23:54:24,008][starVLA.training.train_starvla][INFO] - Step 4614, Loss: {'train/loss': 0.12006568908691406, 'train/grad_norm_pre_clip': 24.21584701538086, 'timing/data': 0.0012557972222566605, 'timing/model': 0.9491038881242275, 'learning_rate/qwen_vl_interface': 3.037852554174269e-06, 'learning_rate/action_model': 3.0378525541742692e-05, 'epoch': 2.09}) +[2026-06-17 23:54:27,999][starVLA.training.train_starvla][INFO] - Step 4615, Loss: {'train/loss': 0.19953308999538422, 'train/grad_norm_pre_clip': 29.411178588867188, 'timing/data': 0.0005809934809803963, 'timing/model': 0.9564181286841631, 'learning_rate/qwen_vl_interface': 3.035938943723295e-06, 'learning_rate/action_model': 3.035938943723295e-05, 'epoch': 2.09}) +[2026-06-17 23:54:31,936][starVLA.training.train_starvla][INFO] - Step 4616, Loss: {'train/loss': 0.12702715396881104, 'train/grad_norm_pre_clip': 16.615901947021484, 'timing/data': 0.0004995530471205711, 'timing/model': 0.9849621830508113, 'learning_rate/qwen_vl_interface': 3.03402579224956e-06, 'learning_rate/action_model': 3.03402579224956e-05, 'epoch': 2.09}) +[2026-06-17 23:54:35,904][starVLA.training.train_starvla][INFO] - Step 4617, Loss: {'train/loss': 0.2727930545806885, 'train/grad_norm_pre_clip': 33.20204162597656, 'timing/data': 0.0005588168278336525, 'timing/model': 0.957516017369926, 'learning_rate/qwen_vl_interface': 3.032113100149663e-06, 'learning_rate/action_model': 3.0321131001496628e-05, 'epoch': 2.09}) +[2026-06-17 23:54:39,780][starVLA.training.train_starvla][INFO] - Step 4618, Loss: {'train/loss': 0.26349198818206787, 'train/grad_norm_pre_clip': 23.001657485961914, 'timing/data': 0.0006003528833389282, 'timing/model': 0.9672372797504067, 'learning_rate/qwen_vl_interface': 3.0302008678201067e-06, 'learning_rate/action_model': 3.0302008678201067e-05, 'epoch': 2.1}) +[2026-06-17 23:54:43,733][starVLA.training.train_starvla][INFO] - Step 4619, Loss: {'train/loss': 0.3534146547317505, 'train/grad_norm_pre_clip': 20.915056228637695, 'timing/data': 0.0006168056279420853, 'timing/model': 0.9501977981999516, 'learning_rate/qwen_vl_interface': 3.028289095657299e-06, 'learning_rate/action_model': 3.028289095657299e-05, 'epoch': 2.1}) +[2026-06-17 23:54:49,722][starVLA.training.train_starvla][INFO] - Step 4620, Loss: {'train/loss': 0.26557210087776184, 'train/grad_norm_pre_clip': 22.076929092407227, 'timing/data': 0.0010236669331789017, 'timing/model': 3.220283660106361, 'learning_rate/qwen_vl_interface': 3.026377784057552e-06, 'learning_rate/action_model': 3.026377784057552e-05, 'epoch': 2.1}) +[2026-06-17 23:54:53,509][starVLA.training.train_starvla][INFO] - Step 4621, Loss: {'train/loss': 0.2752336859703064, 'train/grad_norm_pre_clip': 30.95618438720703, 'timing/data': 0.0005786418914794922, 'timing/model': 1.0342944981530309, 'learning_rate/qwen_vl_interface': 3.0244669334170822e-06, 'learning_rate/action_model': 3.0244669334170822e-05, 'epoch': 2.1}) +[2026-06-17 23:54:57,406][starVLA.training.train_starvla][INFO] - Step 4622, Loss: {'train/loss': 0.2072584480047226, 'train/grad_norm_pre_clip': 22.94346046447754, 'timing/data': 0.000586547888815403, 'timing/model': 0.9718906357884407, 'learning_rate/qwen_vl_interface': 3.0225565441320156e-06, 'learning_rate/action_model': 3.0225565441320152e-05, 'epoch': 2.1}) +[2026-06-17 23:55:01,319][starVLA.training.train_starvla][INFO] - Step 4623, Loss: {'train/loss': 0.28962716460227966, 'train/grad_norm_pre_clip': 24.53546142578125, 'timing/data': 0.0005839327350258827, 'timing/model': 0.9658882757648826, 'learning_rate/qwen_vl_interface': 3.0206466165983716e-06, 'learning_rate/action_model': 3.0206466165983716e-05, 'epoch': 2.1}) +[2026-06-17 23:55:05,212][starVLA.training.train_starvla][INFO] - Step 4624, Loss: {'train/loss': 0.33140507340431213, 'train/grad_norm_pre_clip': 28.38568687438965, 'timing/data': 0.0005575064569711685, 'timing/model': 0.9567657029256225, 'learning_rate/qwen_vl_interface': 3.0187371512120833e-06, 'learning_rate/action_model': 3.0187371512120832e-05, 'epoch': 2.1}) +[2026-06-17 23:55:09,047][starVLA.training.train_starvla][INFO] - Step 4625, Loss: {'train/loss': 0.49179255962371826, 'train/grad_norm_pre_clip': 37.58924865722656, 'timing/data': 0.0008827093988656998, 'timing/model': 0.9437491958960891, 'learning_rate/qwen_vl_interface': 3.0168281483689825e-06, 'learning_rate/action_model': 3.0168281483689825e-05, 'epoch': 2.1}) +[2026-06-17 23:55:12,900][starVLA.training.train_starvla][INFO] - Step 4626, Loss: {'train/loss': 0.47618600726127625, 'train/grad_norm_pre_clip': 35.694522857666016, 'timing/data': 0.0004533473402261734, 'timing/model': 0.9651763178408146, 'learning_rate/qwen_vl_interface': 3.014919608464812e-06, 'learning_rate/action_model': 3.0149196084648117e-05, 'epoch': 2.1}) +[2026-06-17 23:55:16,771][starVLA.training.train_starvla][INFO] - Step 4627, Loss: {'train/loss': 0.23370632529258728, 'train/grad_norm_pre_clip': 33.20920944213867, 'timing/data': 0.0007809894159436226, 'timing/model': 1.1380835240706801, 'learning_rate/qwen_vl_interface': 3.0130115318952124e-06, 'learning_rate/action_model': 3.0130115318952122e-05, 'epoch': 2.1}) +[2026-06-17 23:55:20,449][starVLA.training.train_starvla][INFO] - Step 4628, Loss: {'train/loss': 0.1942070573568344, 'train/grad_norm_pre_clip': 18.72455596923828, 'timing/data': 0.0005196807906031609, 'timing/model': 0.9420602694153786, 'learning_rate/qwen_vl_interface': 3.011103919055726e-06, 'learning_rate/action_model': 3.011103919055726e-05, 'epoch': 2.1}) +[2026-06-17 23:55:24,282][starVLA.training.train_starvla][INFO] - Step 4629, Loss: {'train/loss': 0.19338402152061462, 'train/grad_norm_pre_clip': 26.955839157104492, 'timing/data': 0.0005934294313192368, 'timing/model': 0.9434344423934817, 'learning_rate/qwen_vl_interface': 3.009196770341809e-06, 'learning_rate/action_model': 3.009196770341809e-05, 'epoch': 2.1}) +[2026-06-17 23:55:28,098][starVLA.training.train_starvla][INFO] - Step 4630, Loss: {'train/loss': 0.0603400394320488, 'train/grad_norm_pre_clip': 8.920947074890137, 'timing/data': 0.0006039245054125786, 'timing/model': 0.960038079880178, 'learning_rate/qwen_vl_interface': 3.0072900861488125e-06, 'learning_rate/action_model': 3.0072900861488125e-05, 'epoch': 2.1}) +[2026-06-17 23:55:31,940][starVLA.training.train_starvla][INFO] - Step 4631, Loss: {'train/loss': 0.23902477324008942, 'train/grad_norm_pre_clip': 14.526806831359863, 'timing/data': 0.0009159278124570847, 'timing/model': 0.942068318836391, 'learning_rate/qwen_vl_interface': 3.0053838668719944e-06, 'learning_rate/action_model': 3.0053838668719943e-05, 'epoch': 2.1}) +[2026-06-17 23:55:35,792][starVLA.training.train_starvla][INFO] - Step 4632, Loss: {'train/loss': 0.33380717039108276, 'train/grad_norm_pre_clip': 31.87687873840332, 'timing/data': 0.0005334373563528061, 'timing/model': 0.9507792238146067, 'learning_rate/qwen_vl_interface': 3.003478112906515e-06, 'learning_rate/action_model': 3.0034781129065148e-05, 'epoch': 2.1}) +[2026-06-17 23:55:39,977][starVLA.training.train_starvla][INFO] - Step 4633, Loss: {'train/loss': 0.14198829233646393, 'train/grad_norm_pre_clip': 14.540928840637207, 'timing/data': 0.0009379750117659569, 'timing/model': 1.4163410076871514, 'learning_rate/qwen_vl_interface': 3.0015728246474405e-06, 'learning_rate/action_model': 3.0015728246474405e-05, 'epoch': 2.1}) +[2026-06-17 23:55:44,266][starVLA.training.train_starvla][INFO] - Step 4634, Loss: {'train/loss': 0.32589104771614075, 'train/grad_norm_pre_clip': 40.08460235595703, 'timing/data': 0.0012587206438183784, 'timing/model': 1.1689421590417624, 'learning_rate/qwen_vl_interface': 2.99966800248974e-06, 'learning_rate/action_model': 2.99966800248974e-05, 'epoch': 2.1}) +[2026-06-17 23:55:48,106][starVLA.training.train_starvla][INFO] - Step 4635, Loss: {'train/loss': 0.2822684645652771, 'train/grad_norm_pre_clip': 26.306997299194336, 'timing/data': 0.0005293730646371841, 'timing/model': 0.9672053037211299, 'learning_rate/qwen_vl_interface': 2.9977636468282817e-06, 'learning_rate/action_model': 2.9977636468282816e-05, 'epoch': 2.1}) +[2026-06-17 23:55:53,397][starVLA.training.train_starvla][INFO] - Step 4636, Loss: {'train/loss': 0.28808096051216125, 'train/grad_norm_pre_clip': 24.85223388671875, 'timing/data': 0.001191270537674427, 'timing/model': 0.9573844466358423, 'learning_rate/qwen_vl_interface': 2.995859758057846e-06, 'learning_rate/action_model': 2.9958597580578456e-05, 'epoch': 2.1}) +[2026-06-17 23:55:57,345][starVLA.training.train_starvla][INFO] - Step 4637, Loss: {'train/loss': 0.24533571302890778, 'train/grad_norm_pre_clip': 15.145934104919434, 'timing/data': 0.0012430883944034576, 'timing/model': 0.9577252008020878, 'learning_rate/qwen_vl_interface': 2.993956336573106e-06, 'learning_rate/action_model': 2.9939563365731062e-05, 'epoch': 2.1}) +[2026-06-17 23:56:01,256][starVLA.training.train_starvla][INFO] - Step 4638, Loss: {'train/loss': 0.0413627102971077, 'train/grad_norm_pre_clip': 8.371429443359375, 'timing/data': 0.0005237702280282974, 'timing/model': 0.9388155061751604, 'learning_rate/qwen_vl_interface': 2.9920533827686453e-06, 'learning_rate/action_model': 2.9920533827686453e-05, 'epoch': 2.1}) +[2026-06-17 23:56:05,262][starVLA.training.train_starvla][INFO] - Step 4639, Loss: {'train/loss': 0.16456519067287445, 'train/grad_norm_pre_clip': 20.72837257385254, 'timing/data': 0.0011451467871665955, 'timing/model': 0.9569300729781389, 'learning_rate/qwen_vl_interface': 2.9901508970389477e-06, 'learning_rate/action_model': 2.9901508970389476e-05, 'epoch': 2.1}) +[2026-06-17 23:56:09,168][starVLA.training.train_starvla][INFO] - Step 4640, Loss: {'train/loss': 0.1222570389509201, 'train/grad_norm_pre_clip': 21.434242248535156, 'timing/data': 0.0009772386401891708, 'timing/model': 1.1597254322841763, 'learning_rate/qwen_vl_interface': 2.988248879778401e-06, 'learning_rate/action_model': 2.988248879778401e-05, 'epoch': 2.11}) +[2026-06-17 23:56:13,255][starVLA.training.train_starvla][INFO] - Step 4641, Loss: {'train/loss': 0.373331218957901, 'train/grad_norm_pre_clip': 18.929645538330078, 'timing/data': 0.0013702455908060074, 'timing/model': 1.2516415975987911, 'learning_rate/qwen_vl_interface': 2.986347331381297e-06, 'learning_rate/action_model': 2.9863473313812967e-05, 'epoch': 2.11}) +[2026-06-17 23:56:16,979][starVLA.training.train_starvla][INFO] - Step 4642, Loss: {'train/loss': 0.3706732988357544, 'train/grad_norm_pre_clip': 40.770206451416016, 'timing/data': 0.000533808022737503, 'timing/model': 0.9445001976564527, 'learning_rate/qwen_vl_interface': 2.984446252241826e-06, 'learning_rate/action_model': 2.984446252241826e-05, 'epoch': 2.11}) +[2026-06-17 23:56:21,417][starVLA.training.train_starvla][INFO] - Step 4643, Loss: {'train/loss': 0.1522424966096878, 'train/grad_norm_pre_clip': 11.140861511230469, 'timing/data': 0.0005673728883266449, 'timing/model': 1.0474627185612917, 'learning_rate/qwen_vl_interface': 2.982545642754083e-06, 'learning_rate/action_model': 2.9825456427540826e-05, 'epoch': 2.11}) +[2026-06-17 23:56:25,807][starVLA.training.train_starvla][INFO] - Step 4644, Loss: {'train/loss': 0.6104716658592224, 'train/grad_norm_pre_clip': 39.27216720581055, 'timing/data': 0.0011035017669200897, 'timing/model': 1.028607907705009, 'learning_rate/qwen_vl_interface': 2.9806455033120697e-06, 'learning_rate/action_model': 2.9806455033120696e-05, 'epoch': 2.11}) +[2026-06-17 23:56:29,714][starVLA.training.train_starvla][INFO] - Step 4645, Loss: {'train/loss': 0.4318757951259613, 'train/grad_norm_pre_clip': 26.547271728515625, 'timing/data': 0.0005757417529821396, 'timing/model': 0.9675296749919653, 'learning_rate/qwen_vl_interface': 2.978745834309685e-06, 'learning_rate/action_model': 2.9787458343096846e-05, 'epoch': 2.11}) +[2026-06-17 23:56:33,664][starVLA.training.train_starvla][INFO] - Step 4646, Loss: {'train/loss': 0.35470497608184814, 'train/grad_norm_pre_clip': 34.61435317993164, 'timing/data': 0.0008313283324241638, 'timing/model': 0.9919305965304375, 'learning_rate/qwen_vl_interface': 2.9768466361407335e-06, 'learning_rate/action_model': 2.9768466361407334e-05, 'epoch': 2.11}) +[2026-06-17 23:56:37,707][starVLA.training.train_starvla][INFO] - Step 4647, Loss: {'train/loss': 0.41601815819740295, 'train/grad_norm_pre_clip': 20.667192459106445, 'timing/data': 0.0005917707458138466, 'timing/model': 1.2279690001159906, 'learning_rate/qwen_vl_interface': 2.97494790919892e-06, 'learning_rate/action_model': 2.9749479091989197e-05, 'epoch': 2.11}) +[2026-06-17 23:56:44,546][starVLA.training.train_starvla][INFO] - Step 4648, Loss: {'train/loss': 0.12104041874408722, 'train/grad_norm_pre_clip': 14.00218391418457, 'timing/data': 0.0005058916285634041, 'timing/model': 2.782564213499427, 'learning_rate/qwen_vl_interface': 2.9730496538778524e-06, 'learning_rate/action_model': 2.9730496538778523e-05, 'epoch': 2.11}) +[2026-06-17 23:56:48,249][starVLA.training.train_starvla][INFO] - Step 4649, Loss: {'train/loss': 0.23742952942848206, 'train/grad_norm_pre_clip': 24.507017135620117, 'timing/data': 0.0009890664368867874, 'timing/model': 0.952629760839045, 'learning_rate/qwen_vl_interface': 2.9711518705710422e-06, 'learning_rate/action_model': 2.9711518705710424e-05, 'epoch': 2.11}) +[2026-06-17 23:56:52,231][starVLA.training.train_starvla][INFO] - Step 4650, Loss: {'train/loss': 0.25985902547836304, 'train/grad_norm_pre_clip': 21.62436294555664, 'timing/data': 0.0006029000505805016, 'timing/model': 0.9676365777850151, 'learning_rate/qwen_vl_interface': 2.9692545596719014e-06, 'learning_rate/action_model': 2.969254559671901e-05, 'epoch': 2.11}) +[2026-06-17 23:56:56,219][starVLA.training.train_starvla][INFO] - Step 4651, Loss: {'train/loss': 0.07185246795415878, 'train/grad_norm_pre_clip': 18.008203506469727, 'timing/data': 0.0007720114663243294, 'timing/model': 0.9798302762210369, 'learning_rate/qwen_vl_interface': 2.9673577215737454e-06, 'learning_rate/action_model': 2.9673577215737452e-05, 'epoch': 2.11}) +[2026-06-17 23:57:00,193][starVLA.training.train_starvla][INFO] - Step 4652, Loss: {'train/loss': 0.220366433262825, 'train/grad_norm_pre_clip': 20.809951782226562, 'timing/data': 0.0005749789997935295, 'timing/model': 0.951565875671804, 'learning_rate/qwen_vl_interface': 2.9654613566697886e-06, 'learning_rate/action_model': 2.9654613566697887e-05, 'epoch': 2.11}) +[2026-06-17 23:57:04,391][starVLA.training.train_starvla][INFO] - Step 4653, Loss: {'train/loss': 0.35516494512557983, 'train/grad_norm_pre_clip': 21.964956283569336, 'timing/data': 0.0009224386885762215, 'timing/model': 0.9710849123075604, 'learning_rate/qwen_vl_interface': 2.9635654653531516e-06, 'learning_rate/action_model': 2.9635654653531515e-05, 'epoch': 2.11}) +[2026-06-17 23:57:12,250][starVLA.training.train_starvla][INFO] - Step 4654, Loss: {'train/loss': 0.08624471724033356, 'train/grad_norm_pre_clip': 15.302276611328125, 'timing/data': 0.0008199475705623627, 'timing/model': 3.592881018295884, 'learning_rate/qwen_vl_interface': 2.961670048016852e-06, 'learning_rate/action_model': 2.9616700480168518e-05, 'epoch': 2.11}) +[2026-06-17 23:57:16,296][starVLA.training.train_starvla][INFO] - Step 4655, Loss: {'train/loss': 0.15229520201683044, 'train/grad_norm_pre_clip': 19.461912155151367, 'timing/data': 0.051887194626033306, 'timing/model': 1.0098099177703261, 'learning_rate/qwen_vl_interface': 2.9597751050538177e-06, 'learning_rate/action_model': 2.9597751050538176e-05, 'epoch': 2.11}) +[2026-06-17 23:57:20,562][starVLA.training.train_starvla][INFO] - Step 4656, Loss: {'train/loss': 0.25357335805892944, 'train/grad_norm_pre_clip': 84.44589233398438, 'timing/data': 0.0008260942995548248, 'timing/model': 0.962872689589858, 'learning_rate/qwen_vl_interface': 2.957880636856866e-06, 'learning_rate/action_model': 2.9578806368568655e-05, 'epoch': 2.11}) +[2026-06-17 23:57:25,415][starVLA.training.train_starvla][INFO] - Step 4657, Loss: {'train/loss': 0.33153387904167175, 'train/grad_norm_pre_clip': 21.3268985748291, 'timing/data': 0.0005865953862667084, 'timing/model': 1.0639730375260115, 'learning_rate/qwen_vl_interface': 2.955986643818723e-06, 'learning_rate/action_model': 2.9559866438187228e-05, 'epoch': 2.11}) +[2026-06-17 23:57:29,462][starVLA.training.train_starvla][INFO] - Step 4658, Loss: {'train/loss': 0.19622208178043365, 'train/grad_norm_pre_clip': 15.227259635925293, 'timing/data': 0.0005718190222978592, 'timing/model': 0.9799308069050312, 'learning_rate/qwen_vl_interface': 2.954093126332019e-06, 'learning_rate/action_model': 2.9540931263320194e-05, 'epoch': 2.11}) +[2026-06-17 23:57:33,484][starVLA.training.train_starvla][INFO] - Step 4659, Loss: {'train/loss': 0.10015787184238434, 'train/grad_norm_pre_clip': 17.65245246887207, 'timing/data': 0.0005898075178265572, 'timing/model': 0.9558299537748098, 'learning_rate/qwen_vl_interface': 2.9522000847892807e-06, 'learning_rate/action_model': 2.9522000847892807e-05, 'epoch': 2.11}) +[2026-06-17 23:57:39,405][starVLA.training.train_starvla][INFO] - Step 4660, Loss: {'train/loss': 0.07623028755187988, 'train/grad_norm_pre_clip': 10.484601974487305, 'timing/data': 0.0014522839337587357, 'timing/model': 1.241557976230979, 'learning_rate/qwen_vl_interface': 2.9503075195829393e-06, 'learning_rate/action_model': 2.9503075195829393e-05, 'epoch': 2.11}) +[2026-06-17 23:57:43,398][starVLA.training.train_starvla][INFO] - Step 4661, Loss: {'train/loss': 0.436540812253952, 'train/grad_norm_pre_clip': 34.423484802246094, 'timing/data': 0.00137270987033844, 'timing/model': 1.1749679958447814, 'learning_rate/qwen_vl_interface': 2.9484154311053203e-06, 'learning_rate/action_model': 2.94841543110532e-05, 'epoch': 2.11}) +[2026-06-17 23:57:47,069][starVLA.training.train_starvla][INFO] - Step 4662, Loss: {'train/loss': 0.4519155025482178, 'train/grad_norm_pre_clip': 25.69502067565918, 'timing/data': 0.0005967812612652779, 'timing/model': 0.942454987205565, 'learning_rate/qwen_vl_interface': 2.94652381974866e-06, 'learning_rate/action_model': 2.94652381974866e-05, 'epoch': 2.12}) +[2026-06-17 23:57:50,925][starVLA.training.train_starvla][INFO] - Step 4663, Loss: {'train/loss': 0.2172115445137024, 'train/grad_norm_pre_clip': 20.190765380859375, 'timing/data': 0.0005638115108013153, 'timing/model': 0.9665920929983258, 'learning_rate/qwen_vl_interface': 2.9446326859050916e-06, 'learning_rate/action_model': 2.9446326859050916e-05, 'epoch': 2.12}) +[2026-06-17 23:57:54,858][starVLA.training.train_starvla][INFO] - Step 4664, Loss: {'train/loss': 0.23225441575050354, 'train/grad_norm_pre_clip': 20.3662166595459, 'timing/data': 0.0005927411839365959, 'timing/model': 0.9468543203547597, 'learning_rate/qwen_vl_interface': 2.9427420299666472e-06, 'learning_rate/action_model': 2.942742029966647e-05, 'epoch': 2.12}) +[2026-06-17 23:57:58,761][starVLA.training.train_starvla][INFO] - Step 4665, Loss: {'train/loss': 0.22701680660247803, 'train/grad_norm_pre_clip': 18.25162124633789, 'timing/data': 0.0010435562580823898, 'timing/model': 0.9584225174039602, 'learning_rate/qwen_vl_interface': 2.940851852325263e-06, 'learning_rate/action_model': 2.9408518523252632e-05, 'epoch': 2.12}) +[2026-06-17 23:58:02,709][starVLA.training.train_starvla][INFO] - Step 4666, Loss: {'train/loss': 0.3051181733608246, 'train/grad_norm_pre_clip': 24.928783416748047, 'timing/data': 0.0011044768616557121, 'timing/model': 0.9568218505010009, 'learning_rate/qwen_vl_interface': 2.9389621533727745e-06, 'learning_rate/action_model': 2.9389621533727747e-05, 'epoch': 2.12}) +[2026-06-17 23:58:06,592][starVLA.training.train_starvla][INFO] - Step 4667, Loss: {'train/loss': 0.3116896152496338, 'train/grad_norm_pre_clip': 23.927291870117188, 'timing/data': 0.0005768639966845512, 'timing/model': 1.1567526748403907, 'learning_rate/qwen_vl_interface': 2.9370729335009185e-06, 'learning_rate/action_model': 2.9370729335009184e-05, 'epoch': 2.12}) +[2026-06-17 23:58:10,320][starVLA.training.train_starvla][INFO] - Step 4668, Loss: {'train/loss': 0.31973928213119507, 'train/grad_norm_pre_clip': 20.209274291992188, 'timing/data': 0.0007523940876126289, 'timing/model': 0.955770212225616, 'learning_rate/qwen_vl_interface': 2.93518419310133e-06, 'learning_rate/action_model': 2.93518419310133e-05, 'epoch': 2.12}) +[2026-06-17 23:58:14,408][starVLA.training.train_starvla][INFO] - Step 4669, Loss: {'train/loss': 0.23358313739299774, 'train/grad_norm_pre_clip': 24.206735610961914, 'timing/data': 0.0005848081782460213, 'timing/model': 0.9555999236181378, 'learning_rate/qwen_vl_interface': 2.9332959325655537e-06, 'learning_rate/action_model': 2.9332959325655535e-05, 'epoch': 2.12}) +[2026-06-17 23:58:20,101][starVLA.training.train_starvla][INFO] - Step 4670, Loss: {'train/loss': 0.37266308069229126, 'train/grad_norm_pre_clip': 31.80025291442871, 'timing/data': 0.001017221249639988, 'timing/model': 1.1415071180090308, 'learning_rate/qwen_vl_interface': 2.9314081522850206e-06, 'learning_rate/action_model': 2.9314081522850205e-05, 'epoch': 2.12}) +[2026-06-17 23:58:24,210][starVLA.training.train_starvla][INFO] - Step 4671, Loss: {'train/loss': 0.12039308249950409, 'train/grad_norm_pre_clip': 19.90553092956543, 'timing/data': 0.0010472601279616356, 'timing/model': 0.9653372066095471, 'learning_rate/qwen_vl_interface': 2.9295208526510726e-06, 'learning_rate/action_model': 2.9295208526510725e-05, 'epoch': 2.12}) +[2026-06-17 23:58:29,197][starVLA.training.train_starvla][INFO] - Step 4672, Loss: {'train/loss': 0.14526742696762085, 'train/grad_norm_pre_clip': 21.85296630859375, 'timing/data': 0.0007805079221725464, 'timing/model': 1.025094530545175, 'learning_rate/qwen_vl_interface': 2.9276340340549476e-06, 'learning_rate/action_model': 2.9276340340549473e-05, 'epoch': 2.12}) +[2026-06-17 23:58:33,112][starVLA.training.train_starvla][INFO] - Step 4673, Loss: {'train/loss': 0.2301243543624878, 'train/grad_norm_pre_clip': 20.7112979888916, 'timing/data': 0.0009428849443793297, 'timing/model': 1.1398054454475641, 'learning_rate/qwen_vl_interface': 2.9257476968877875e-06, 'learning_rate/action_model': 2.9257476968877874e-05, 'epoch': 2.12}) +[2026-06-17 23:58:36,970][starVLA.training.train_starvla][INFO] - Step 4674, Loss: {'train/loss': 0.4505279064178467, 'train/grad_norm_pre_clip': 29.018278121948242, 'timing/data': 0.0005732662975788116, 'timing/model': 1.1306331465020776, 'learning_rate/qwen_vl_interface': 2.923861841540631e-06, 'learning_rate/action_model': 2.923861841540631e-05, 'epoch': 2.12}) +[2026-06-17 23:58:40,681][starVLA.training.train_starvla][INFO] - Step 4675, Loss: {'train/loss': 0.14661626517772675, 'train/grad_norm_pre_clip': 19.242910385131836, 'timing/data': 0.0005483012646436691, 'timing/model': 0.9553473116829991, 'learning_rate/qwen_vl_interface': 2.9219764684044196e-06, 'learning_rate/action_model': 2.9219764684044193e-05, 'epoch': 2.12}) +[2026-06-17 23:58:44,740][starVLA.training.train_starvla][INFO] - Step 4676, Loss: {'train/loss': 0.04344988241791725, 'train/grad_norm_pre_clip': 8.647672653198242, 'timing/data': 0.001069018617272377, 'timing/model': 0.9456689395010471, 'learning_rate/qwen_vl_interface': 2.920091577869988e-06, 'learning_rate/action_model': 2.9200915778699877e-05, 'epoch': 2.12}) +[2026-06-17 23:58:48,595][starVLA.training.train_starvla][INFO] - Step 4677, Loss: {'train/loss': 0.2715966999530792, 'train/grad_norm_pre_clip': 30.605018615722656, 'timing/data': 0.0006386861205101013, 'timing/model': 0.9542447235435247, 'learning_rate/qwen_vl_interface': 2.918207170328081e-06, 'learning_rate/action_model': 2.918207170328081e-05, 'epoch': 2.12}) +[2026-06-17 23:58:52,454][starVLA.training.train_starvla][INFO] - Step 4678, Loss: {'train/loss': 0.16684015095233917, 'train/grad_norm_pre_clip': 18.208160400390625, 'timing/data': 0.0009899865835905075, 'timing/model': 0.9594814879819751, 'learning_rate/qwen_vl_interface': 2.9163232461693373e-06, 'learning_rate/action_model': 2.9163232461693373e-05, 'epoch': 2.12}) +[2026-06-17 23:58:56,419][starVLA.training.train_starvla][INFO] - Step 4679, Loss: {'train/loss': 0.26402533054351807, 'train/grad_norm_pre_clip': 24.881546020507812, 'timing/data': 0.0005667479708790779, 'timing/model': 0.9961955081671476, 'learning_rate/qwen_vl_interface': 2.9144398057842948e-06, 'learning_rate/action_model': 2.9144398057842948e-05, 'epoch': 2.12}) +[2026-06-17 23:59:00,543][starVLA.training.train_starvla][INFO] - Step 4680, Loss: {'train/loss': 0.3960910439491272, 'train/grad_norm_pre_clip': 35.697566986083984, 'timing/data': 0.0009646648541092873, 'timing/model': 1.2519505200907588, 'learning_rate/qwen_vl_interface': 2.912556849563394e-06, 'learning_rate/action_model': 2.9125568495633943e-05, 'epoch': 2.12}) +[2026-06-17 23:59:04,647][starVLA.training.train_starvla][INFO] - Step 4681, Loss: {'train/loss': 0.39584866166114807, 'train/grad_norm_pre_clip': 38.88455581665039, 'timing/data': 0.0005826912820339203, 'timing/model': 1.1882641855627298, 'learning_rate/qwen_vl_interface': 2.910674377896974e-06, 'learning_rate/action_model': 2.910674377896974e-05, 'epoch': 2.12}) +[2026-06-17 23:59:09,093][starVLA.training.train_starvla][INFO] - Step 4682, Loss: {'train/loss': 0.12248857319355011, 'train/grad_norm_pre_clip': 19.159645080566406, 'timing/data': 0.0008123284205794334, 'timing/model': 0.9795615607872605, 'learning_rate/qwen_vl_interface': 2.9087923911752723e-06, 'learning_rate/action_model': 2.9087923911752725e-05, 'epoch': 2.12}) +[2026-06-17 23:59:13,103][starVLA.training.train_starvla][INFO] - Step 4683, Loss: {'train/loss': 0.06319435685873032, 'train/grad_norm_pre_clip': 9.683890342712402, 'timing/data': 0.0007142294198274612, 'timing/model': 1.004925636574626, 'learning_rate/qwen_vl_interface': 2.906910889788425e-06, 'learning_rate/action_model': 2.906910889788425e-05, 'epoch': 2.12}) +[2026-06-17 23:59:17,106][starVLA.training.train_starvla][INFO] - Step 4684, Loss: {'train/loss': 0.1665620654821396, 'train/grad_norm_pre_clip': 11.976967811584473, 'timing/data': 0.0005808193236589432, 'timing/model': 1.0485228765755892, 'learning_rate/qwen_vl_interface': 2.905029874126476e-06, 'learning_rate/action_model': 2.905029874126476e-05, 'epoch': 2.13}) +[2026-06-17 23:59:21,212][starVLA.training.train_starvla][INFO] - Step 4685, Loss: {'train/loss': 0.19595031440258026, 'train/grad_norm_pre_clip': 19.304899215698242, 'timing/data': 0.0010072235018014908, 'timing/model': 1.0441046049818397, 'learning_rate/qwen_vl_interface': 2.9031493445793558e-06, 'learning_rate/action_model': 2.9031493445793558e-05, 'epoch': 2.13}) +[2026-06-17 23:59:25,400][starVLA.training.train_starvla][INFO] - Step 4686, Loss: {'train/loss': 0.10119666159152985, 'train/grad_norm_pre_clip': 21.539962768554688, 'timing/data': 0.000541912391781807, 'timing/model': 1.2200666358694434, 'learning_rate/qwen_vl_interface': 2.9012693015369007e-06, 'learning_rate/action_model': 2.9012693015369008e-05, 'epoch': 2.13}) +[2026-06-17 23:59:29,255][starVLA.training.train_starvla][INFO] - Step 4687, Loss: {'train/loss': 0.25374114513397217, 'train/grad_norm_pre_clip': 21.47637176513672, 'timing/data': 0.0005243364721536636, 'timing/model': 1.1173272151499987, 'learning_rate/qwen_vl_interface': 2.8993897453888454e-06, 'learning_rate/action_model': 2.899389745388845e-05, 'epoch': 2.13}) +[2026-06-17 23:59:33,338][starVLA.training.train_starvla][INFO] - Step 4688, Loss: {'train/loss': 0.18084226548671722, 'train/grad_norm_pre_clip': 17.56999397277832, 'timing/data': 0.0006648832932114601, 'timing/model': 1.2847672998905182, 'learning_rate/qwen_vl_interface': 2.8975106765248267e-06, 'learning_rate/action_model': 2.8975106765248267e-05, 'epoch': 2.13}) +[2026-06-17 23:59:39,121][starVLA.training.train_starvla][INFO] - Step 4689, Loss: {'train/loss': 0.34998685121536255, 'train/grad_norm_pre_clip': 26.64803695678711, 'timing/data': 0.0005717622116208076, 'timing/model': 1.852709055878222, 'learning_rate/qwen_vl_interface': 2.895632095334378e-06, 'learning_rate/action_model': 2.895632095334378e-05, 'epoch': 2.13}) +[2026-06-17 23:59:43,187][starVLA.training.train_starvla][INFO] - Step 4690, Loss: {'train/loss': 0.16016817092895508, 'train/grad_norm_pre_clip': 22.626667022705078, 'timing/data': 0.0011540642008185387, 'timing/model': 1.0046033626422286, 'learning_rate/qwen_vl_interface': 2.8937540022069267e-06, 'learning_rate/action_model': 2.8937540022069265e-05, 'epoch': 2.13}) +[2026-06-17 23:59:47,165][starVLA.training.train_starvla][INFO] - Step 4691, Loss: {'train/loss': 0.12398112565279007, 'train/grad_norm_pre_clip': 11.384718894958496, 'timing/data': 0.0011971639469265938, 'timing/model': 0.9474766962230206, 'learning_rate/qwen_vl_interface': 2.8918763975318076e-06, 'learning_rate/action_model': 2.8918763975318076e-05, 'epoch': 2.13}) +[2026-06-17 23:59:50,968][starVLA.training.train_starvla][INFO] - Step 4692, Loss: {'train/loss': 0.140672966837883, 'train/grad_norm_pre_clip': 16.035581588745117, 'timing/data': 0.0012394348159432411, 'timing/model': 0.9595555467531085, 'learning_rate/qwen_vl_interface': 2.889999281698248e-06, 'learning_rate/action_model': 2.889999281698248e-05, 'epoch': 2.13}) +[2026-06-17 23:59:55,047][starVLA.training.train_starvla][INFO] - Step 4693, Loss: {'train/loss': 0.3214704394340515, 'train/grad_norm_pre_clip': 25.57600975036621, 'timing/data': 0.0010966071859002113, 'timing/model': 0.9959696745499969, 'learning_rate/qwen_vl_interface': 2.8881226550953772e-06, 'learning_rate/action_model': 2.888122655095377e-05, 'epoch': 2.13}) +[2026-06-17 23:59:59,130][starVLA.training.train_starvla][INFO] - Step 4694, Loss: {'train/loss': 0.13353870809078217, 'train/grad_norm_pre_clip': 18.385574340820312, 'timing/data': 0.0008916202932596207, 'timing/model': 1.2607687199488282, 'learning_rate/qwen_vl_interface': 2.8862465181122212e-06, 'learning_rate/action_model': 2.886246518112221e-05, 'epoch': 2.13}) +[2026-06-18 00:00:02,964][starVLA.training.train_starvla][INFO] - Step 4695, Loss: {'train/loss': 0.2137201428413391, 'train/grad_norm_pre_clip': 21.532148361206055, 'timing/data': 0.0007399506866931915, 'timing/model': 0.9703089790418744, 'learning_rate/qwen_vl_interface': 2.884370871137706e-06, 'learning_rate/action_model': 2.884370871137706e-05, 'epoch': 2.13}) +[2026-06-18 00:00:07,122][starVLA.training.train_starvla][INFO] - Step 4696, Loss: {'train/loss': 0.48940926790237427, 'train/grad_norm_pre_clip': 30.846529006958008, 'timing/data': 0.000753382220864296, 'timing/model': 1.026098190806806, 'learning_rate/qwen_vl_interface': 2.8824957145606548e-06, 'learning_rate/action_model': 2.8824957145606548e-05, 'epoch': 2.13}) +[2026-06-18 00:00:11,233][starVLA.training.train_starvla][INFO] - Step 4697, Loss: {'train/loss': 0.4095405042171478, 'train/grad_norm_pre_clip': 30.647167205810547, 'timing/data': 0.0005262866616249084, 'timing/model': 0.9535496979951859, 'learning_rate/qwen_vl_interface': 2.8806210487697866e-06, 'learning_rate/action_model': 2.8806210487697867e-05, 'epoch': 2.13}) +[2026-06-18 00:00:16,107][starVLA.training.train_starvla][INFO] - Step 4698, Loss: {'train/loss': 0.06073066592216492, 'train/grad_norm_pre_clip': 9.145949363708496, 'timing/data': 0.000831902027130127, 'timing/model': 1.2416911302134395, 'learning_rate/qwen_vl_interface': 2.87874687415373e-06, 'learning_rate/action_model': 2.87874687415373e-05, 'epoch': 2.13}) +[2026-06-18 00:00:20,146][starVLA.training.train_starvla][INFO] - Step 4699, Loss: {'train/loss': 0.14524000883102417, 'train/grad_norm_pre_clip': 23.985469818115234, 'timing/data': 0.0006354795768857002, 'timing/model': 0.9935565460473299, 'learning_rate/qwen_vl_interface': 2.876873191100995e-06, 'learning_rate/action_model': 2.8768731911009948e-05, 'epoch': 2.13}) +[2026-06-18 00:02:31,777][starVLA.training.train_starvla][INFO] - Step 4700, Loss: {'train/loss': 0.2414267510175705, 'train/grad_norm_pre_clip': 32.83827209472656, 'eval/loss': 0.2809803783893585, 'eval/latency_8/loss': 0.2809803783893585, 'eval/demon_fire_precision': 0.9663608562691132, 'eval/demon_fire_recall': 0.9816713264989128, 'eval/demon_fire_f1': 0.9739559254122361, 'eval/demon_fire_rate_model': 0.9098497495826378, 'eval/demon_fire_rate_teacher': 0.8956594323873122, 'eval/demon_left_precision': 0.9329446064139941, 'eval/demon_left_recall': 0.6852248394004282, 'eval/demon_left_f1': 0.7901234567901235, 'eval/demon_left_rate_model': 0.09543683917640512, 'eval/demon_left_rate_teacher': 0.12993878686700056, 'eval/demon_right_precision': 0.8019662921348315, 'eval/demon_right_recall': 0.8651515151515151, 'eval/demon_right_f1': 0.8323615160349854, 'eval/demon_right_rate_model': 0.19810795770728992, 'eval/demon_right_rate_teacher': 0.18363939899833054, 'eval/demon_move_f1': 0.8112424864125545, 'eval/demon_cc_f1': 0.9186773310220404, 'eval/latency_8/demon_fire_precision': 0.9663608562691132, 'eval/latency_8/demon_fire_recall': 0.9816713264989128, 'eval/latency_8/demon_fire_f1': 0.9739559254122361, 'eval/latency_8/demon_fire_rate_model': 0.9098497495826378, 'eval/latency_8/demon_fire_rate_teacher': 0.8956594323873122, 'eval/latency_8/demon_left_precision': 0.9329446064139941, 'eval/latency_8/demon_left_recall': 0.6852248394004282, 'eval/latency_8/demon_left_f1': 0.7901234567901235, 'eval/latency_8/demon_left_rate_model': 0.09543683917640512, 'eval/latency_8/demon_left_rate_teacher': 0.12993878686700056, 'eval/latency_8/demon_right_precision': 0.8019662921348315, 'eval/latency_8/demon_right_recall': 0.8651515151515151, 'eval/latency_8/demon_right_f1': 0.8323615160349854, 'eval/latency_8/demon_right_rate_model': 0.19810795770728992, 'eval/latency_8/demon_right_rate_teacher': 0.18363939899833054, 'eval/latency_8/demon_move_f1': 0.8112424864125545, 'eval/latency_8/demon_cc_f1': 0.9186773310220404, 'timing/data': 0.0013659261167049408, 'timing/model': 1.2809672886505723, 'learning_rate/qwen_vl_interface': 2.8750000000000013e-06, 'learning_rate/action_model': 2.875000000000001e-05, 'epoch': 2.13}) +[2026-06-18 00:02:31,780][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_4700_state +[2026-06-18 00:02:31,786][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 00:02:38,002][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_4700_state/model.safetensors +[2026-06-18 00:02:53,179][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_4700_state/optimizer.bin +[2026-06-18 00:02:53,180][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_4700_state/sampler.bin +[2026-06-18 00:02:53,180][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_4700_state/sampler_1.bin +[2026-06-18 00:02:53,181][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_4700_state/random_states_0.pkl +[2026-06-18 00:02:53,185][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4700 +[2026-06-18 00:02:53,185][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 00:02:53,201][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 00:02:57,029][starVLA.training.train_starvla][INFO] - Step 4701, Loss: {'train/loss': 0.15453100204467773, 'train/grad_norm_pre_clip': 20.015195846557617, 'timing/data': 0.0006942320615053177, 'timing/model': 0.9409317942336202, 'learning_rate/qwen_vl_interface': 2.8731273012390614e-06, 'learning_rate/action_model': 2.8731273012390615e-05, 'epoch': 2.13}) +[2026-06-18 00:03:00,878][starVLA.training.train_starvla][INFO] - Step 4702, Loss: {'train/loss': 0.48151111602783203, 'train/grad_norm_pre_clip': 39.275978088378906, 'timing/data': 0.0010453136637806892, 'timing/model': 1.188781863078475, 'learning_rate/qwen_vl_interface': 2.8712550952063912e-06, 'learning_rate/action_model': 2.871255095206391e-05, 'epoch': 2.13}) +[2026-06-18 00:03:04,802][starVLA.training.train_starvla][INFO] - Step 4703, Loss: {'train/loss': 0.34202271699905396, 'train/grad_norm_pre_clip': 28.765703201293945, 'timing/data': 0.0017155110836029053, 'timing/model': 1.2151429457589984, 'learning_rate/qwen_vl_interface': 2.8693833822901012e-06, 'learning_rate/action_model': 2.8693833822901012e-05, 'epoch': 2.13}) +[2026-06-18 00:03:08,714][starVLA.training.train_starvla][INFO] - Step 4704, Loss: {'train/loss': 0.17962101101875305, 'train/grad_norm_pre_clip': 19.91706657409668, 'timing/data': 0.00047397147864103317, 'timing/model': 1.234398065134883, 'learning_rate/qwen_vl_interface': 2.867512162878196e-06, 'learning_rate/action_model': 2.867512162878196e-05, 'epoch': 2.13}) +[2026-06-18 00:03:12,545][starVLA.training.train_starvla][INFO] - Step 4705, Loss: {'train/loss': 0.21017630398273468, 'train/grad_norm_pre_clip': 21.214065551757812, 'timing/data': 0.0005159471184015274, 'timing/model': 0.9627330442890525, 'learning_rate/qwen_vl_interface': 2.8656414373585812e-06, 'learning_rate/action_model': 2.865641437358581e-05, 'epoch': 2.13}) +[2026-06-18 00:03:16,404][starVLA.training.train_starvla][INFO] - Step 4706, Loss: {'train/loss': 0.34126877784729004, 'train/grad_norm_pre_clip': 29.63188934326172, 'timing/data': 0.000914938747882843, 'timing/model': 0.9147303327918053, 'learning_rate/qwen_vl_interface': 2.8637712061190644e-06, 'learning_rate/action_model': 2.863771206119064e-05, 'epoch': 2.13}) +[2026-06-18 00:03:20,125][starVLA.training.train_starvla][INFO] - Step 4707, Loss: {'train/loss': 0.6300243735313416, 'train/grad_norm_pre_clip': 28.95001983642578, 'timing/data': 0.0010254289954900742, 'timing/model': 0.9199124528095126, 'learning_rate/qwen_vl_interface': 2.861901469547344e-06, 'learning_rate/action_model': 2.861901469547344e-05, 'epoch': 2.14}) +[2026-06-18 00:03:24,147][starVLA.training.train_starvla][INFO] - Step 4708, Loss: {'train/loss': 0.14503613114356995, 'train/grad_norm_pre_clip': 13.108515739440918, 'timing/data': 0.0005226284265518188, 'timing/model': 0.9543855907395482, 'learning_rate/qwen_vl_interface': 2.8600322280310185e-06, 'learning_rate/action_model': 2.8600322280310183e-05, 'epoch': 2.14}) +[2026-06-18 00:03:28,669][starVLA.training.train_starvla][INFO] - Step 4709, Loss: {'train/loss': 0.4859796166419983, 'train/grad_norm_pre_clip': 34.882606506347656, 'timing/data': 0.0011332407593727112, 'timing/model': 1.7127616805955768, 'learning_rate/qwen_vl_interface': 2.8581634819575843e-06, 'learning_rate/action_model': 2.8581634819575843e-05, 'epoch': 2.14}) +[2026-06-18 00:03:32,519][starVLA.training.train_starvla][INFO] - Step 4710, Loss: {'train/loss': 0.09701014310121536, 'train/grad_norm_pre_clip': 13.94487190246582, 'timing/data': 0.0004964061081409454, 'timing/model': 1.1663728794083, 'learning_rate/qwen_vl_interface': 2.856295231714434e-06, 'learning_rate/action_model': 2.8562952317144337e-05, 'epoch': 2.14}) +[2026-06-18 00:03:36,141][starVLA.training.train_starvla][INFO] - Step 4711, Loss: {'train/loss': 0.468698650598526, 'train/grad_norm_pre_clip': 22.659263610839844, 'timing/data': 0.0005151955410838127, 'timing/model': 0.9316890789195895, 'learning_rate/qwen_vl_interface': 2.854427477688857e-06, 'learning_rate/action_model': 2.854427477688857e-05, 'epoch': 2.14}) +[2026-06-18 00:03:40,322][starVLA.training.train_starvla][INFO] - Step 4712, Loss: {'train/loss': 0.250467985868454, 'train/grad_norm_pre_clip': 24.750152587890625, 'timing/data': 0.0005324510857462883, 'timing/model': 1.1560430163517594, 'learning_rate/qwen_vl_interface': 2.8525602202680424e-06, 'learning_rate/action_model': 2.8525602202680424e-05, 'epoch': 2.14}) +[2026-06-18 00:03:44,132][starVLA.training.train_starvla][INFO] - Step 4713, Loss: {'train/loss': 0.06206611543893814, 'train/grad_norm_pre_clip': 9.863316535949707, 'timing/data': 0.0008362755179405212, 'timing/model': 0.9231048179790378, 'learning_rate/qwen_vl_interface': 2.8506934598390736e-06, 'learning_rate/action_model': 2.8506934598390735e-05, 'epoch': 2.14}) +[2026-06-18 00:03:47,887][starVLA.training.train_starvla][INFO] - Step 4714, Loss: {'train/loss': 0.46033143997192383, 'train/grad_norm_pre_clip': 20.028553009033203, 'timing/data': 0.00048326142132282257, 'timing/model': 0.9304059529677033, 'learning_rate/qwen_vl_interface': 2.848827196788932e-06, 'learning_rate/action_model': 2.848827196788932e-05, 'epoch': 2.14}) +[2026-06-18 00:03:51,800][starVLA.training.train_starvla][INFO] - Step 4715, Loss: {'train/loss': 0.32136282324790955, 'train/grad_norm_pre_clip': 24.3447322845459, 'timing/data': 0.0005003642290830612, 'timing/model': 0.9537090286612511, 'learning_rate/qwen_vl_interface': 2.8469614315044975e-06, 'learning_rate/action_model': 2.8469614315044975e-05, 'epoch': 2.14}) +[2026-06-18 00:03:55,610][starVLA.training.train_starvla][INFO] - Step 4716, Loss: {'train/loss': 0.5353696942329407, 'train/grad_norm_pre_clip': 32.793880462646484, 'timing/data': 0.00047384947538375854, 'timing/model': 1.129395597614348, 'learning_rate/qwen_vl_interface': 2.8450961643725405e-06, 'learning_rate/action_model': 2.8450961643725405e-05, 'epoch': 2.14}) +[2026-06-18 00:03:59,462][starVLA.training.train_starvla][INFO] - Step 4717, Loss: {'train/loss': 0.2576224207878113, 'train/grad_norm_pre_clip': 27.73849105834961, 'timing/data': 0.000859895721077919, 'timing/model': 0.9597440594807267, 'learning_rate/qwen_vl_interface': 2.8432313957797415e-06, 'learning_rate/action_model': 2.8432313957797414e-05, 'epoch': 2.14}) +[2026-06-18 00:04:03,294][starVLA.training.train_starvla][INFO] - Step 4718, Loss: {'train/loss': 0.06503164768218994, 'train/grad_norm_pre_clip': 9.50888729095459, 'timing/data': 0.0005512675270438194, 'timing/model': 0.9410845963284373, 'learning_rate/qwen_vl_interface': 2.8413671261126613e-06, 'learning_rate/action_model': 2.8413671261126613e-05, 'epoch': 2.14}) +[2026-06-18 00:04:08,028][starVLA.training.train_starvla][INFO] - Step 4719, Loss: {'train/loss': 0.20387187600135803, 'train/grad_norm_pre_clip': 20.66666603088379, 'timing/data': 0.000637163408100605, 'timing/model': 1.7516892571002245, 'learning_rate/qwen_vl_interface': 2.8395033557577668e-06, 'learning_rate/action_model': 2.8395033557577668e-05, 'epoch': 2.14}) +[2026-06-18 00:04:12,562][starVLA.training.train_starvla][INFO] - Step 4720, Loss: {'train/loss': 0.13344265520572662, 'train/grad_norm_pre_clip': 26.192485809326172, 'timing/data': 0.0008278768509626389, 'timing/model': 0.9433811521157622, 'learning_rate/qwen_vl_interface': 2.8376400851014214e-06, 'learning_rate/action_model': 2.8376400851014214e-05, 'epoch': 2.14}) +[2026-06-18 00:04:16,767][starVLA.training.train_starvla][INFO] - Step 4721, Loss: {'train/loss': 0.42313411831855774, 'train/grad_norm_pre_clip': 40.92177963256836, 'timing/data': 0.0006481669843196869, 'timing/model': 0.9465519851073623, 'learning_rate/qwen_vl_interface': 2.8357773145298823e-06, 'learning_rate/action_model': 2.8357773145298823e-05, 'epoch': 2.14}) +[2026-06-18 00:04:20,660][starVLA.training.train_starvla][INFO] - Step 4722, Loss: {'train/loss': 0.15735793113708496, 'train/grad_norm_pre_clip': 17.78858184814453, 'timing/data': 0.0004370352253317833, 'timing/model': 1.1436657551676035, 'learning_rate/qwen_vl_interface': 2.8339150444293055e-06, 'learning_rate/action_model': 2.8339150444293055e-05, 'epoch': 2.14}) +[2026-06-18 00:04:24,414][starVLA.training.train_starvla][INFO] - Step 4723, Loss: {'train/loss': 0.06115134060382843, 'train/grad_norm_pre_clip': 12.648880004882812, 'timing/data': 0.0004363739863038063, 'timing/model': 1.091262562200427, 'learning_rate/qwen_vl_interface': 2.8320532751857365e-06, 'learning_rate/action_model': 2.8320532751857366e-05, 'epoch': 2.14}) +[2026-06-18 00:04:28,021][starVLA.training.train_starvla][INFO] - Step 4724, Loss: {'train/loss': 0.11417810618877411, 'train/grad_norm_pre_clip': 16.243938446044922, 'timing/data': 0.0004350859671831131, 'timing/model': 0.9510102765634656, 'learning_rate/qwen_vl_interface': 2.830192007185127e-06, 'learning_rate/action_model': 2.830192007185127e-05, 'epoch': 2.14}) +[2026-06-18 00:04:31,798][starVLA.training.train_starvla][INFO] - Step 4725, Loss: {'train/loss': 0.12701557576656342, 'train/grad_norm_pre_clip': 25.563146591186523, 'timing/data': 0.0007229652255773544, 'timing/model': 0.9394840616732836, 'learning_rate/qwen_vl_interface': 2.828331240813318e-06, 'learning_rate/action_model': 2.8283312408133176e-05, 'epoch': 2.14}) +[2026-06-18 00:04:35,549][starVLA.training.train_starvla][INFO] - Step 4726, Loss: {'train/loss': 0.17659209668636322, 'train/grad_norm_pre_clip': 18.480592727661133, 'timing/data': 0.0005134819075465202, 'timing/model': 0.9236552529036999, 'learning_rate/qwen_vl_interface': 2.826470976456049e-06, 'learning_rate/action_model': 2.826470976456049e-05, 'epoch': 2.14}) +[2026-06-18 00:04:39,292][starVLA.training.train_starvla][INFO] - Step 4727, Loss: {'train/loss': 0.28436896204948425, 'train/grad_norm_pre_clip': 25.840490341186523, 'timing/data': 0.0008660499006509781, 'timing/model': 0.9263189295306802, 'learning_rate/qwen_vl_interface': 2.8246112144989548e-06, 'learning_rate/action_model': 2.8246112144989543e-05, 'epoch': 2.14}) +[2026-06-18 00:04:43,130][starVLA.training.train_starvla][INFO] - Step 4728, Loss: {'train/loss': 0.08708200603723526, 'train/grad_norm_pre_clip': 12.617378234863281, 'timing/data': 0.0006875656545162201, 'timing/model': 0.9795669447630644, 'learning_rate/qwen_vl_interface': 2.8227519553275644e-06, 'learning_rate/action_model': 2.8227519553275645e-05, 'epoch': 2.14}) +[2026-06-18 00:04:47,057][starVLA.training.train_starvla][INFO] - Step 4729, Loss: {'train/loss': 0.1828349381685257, 'train/grad_norm_pre_clip': 25.85659408569336, 'timing/data': 0.0004650093615055084, 'timing/model': 1.1908108405768871, 'learning_rate/qwen_vl_interface': 2.8208931993273066e-06, 'learning_rate/action_model': 2.8208931993273062e-05, 'epoch': 2.15}) +[2026-06-18 00:04:50,735][starVLA.training.train_starvla][INFO] - Step 4730, Loss: {'train/loss': 0.2534475326538086, 'train/grad_norm_pre_clip': 17.65947151184082, 'timing/data': 0.0012455778196454048, 'timing/model': 0.9330692179501057, 'learning_rate/qwen_vl_interface': 2.8190349468835006e-06, 'learning_rate/action_model': 2.8190349468835004e-05, 'epoch': 2.15}) +[2026-06-18 00:04:54,523][starVLA.training.train_starvla][INFO] - Step 4731, Loss: {'train/loss': 0.20969516038894653, 'train/grad_norm_pre_clip': 29.574993133544922, 'timing/data': 0.0009122695773839951, 'timing/model': 0.9310644390061498, 'learning_rate/qwen_vl_interface': 2.817177198381369e-06, 'learning_rate/action_model': 2.817177198381369e-05, 'epoch': 2.15}) +[2026-06-18 00:04:58,346][starVLA.training.train_starvla][INFO] - Step 4732, Loss: {'train/loss': 0.13463492691516876, 'train/grad_norm_pre_clip': 22.078615188598633, 'timing/data': 0.0007393639534711838, 'timing/model': 0.9530412433668971, 'learning_rate/qwen_vl_interface': 2.8153199542060206e-06, 'learning_rate/action_model': 2.8153199542060206e-05, 'epoch': 2.15}) +[2026-06-18 00:05:02,122][starVLA.training.train_starvla][INFO] - Step 4733, Loss: {'train/loss': 0.06518076360225677, 'train/grad_norm_pre_clip': 10.116450309753418, 'timing/data': 0.0006428789347410202, 'timing/model': 0.95522366091609, 'learning_rate/qwen_vl_interface': 2.813463214742465e-06, 'learning_rate/action_model': 2.813463214742465e-05, 'epoch': 2.15}) +[2026-06-18 00:05:05,935][starVLA.training.train_starvla][INFO] - Step 4734, Loss: {'train/loss': 0.7014437317848206, 'train/grad_norm_pre_clip': 41.581626892089844, 'timing/data': 0.0009197527542710304, 'timing/model': 0.9472369207069278, 'learning_rate/qwen_vl_interface': 2.8116069803756047e-06, 'learning_rate/action_model': 2.8116069803756047e-05, 'epoch': 2.15}) +[2026-06-18 00:05:09,733][starVLA.training.train_starvla][INFO] - Step 4735, Loss: {'train/loss': 0.06903840601444244, 'train/grad_norm_pre_clip': 13.729927062988281, 'timing/data': 0.00046726781874895096, 'timing/model': 1.1038347529247403, 'learning_rate/qwen_vl_interface': 2.809751251490244e-06, 'learning_rate/action_model': 2.809751251490244e-05, 'epoch': 2.15}) +[2026-06-18 00:05:13,459][starVLA.training.train_starvla][INFO] - Step 4736, Loss: {'train/loss': 0.5474786758422852, 'train/grad_norm_pre_clip': 28.202754974365234, 'timing/data': 0.00045977160334587097, 'timing/model': 1.0449754633009434, 'learning_rate/qwen_vl_interface': 2.807896028471076e-06, 'learning_rate/action_model': 2.807896028471076e-05, 'epoch': 2.15}) +[2026-06-18 00:05:17,159][starVLA.training.train_starvla][INFO] - Step 4737, Loss: {'train/loss': 0.15455205738544464, 'train/grad_norm_pre_clip': 16.670076370239258, 'timing/data': 0.0004901336506009102, 'timing/model': 0.9483872810378671, 'learning_rate/qwen_vl_interface': 2.806041311702688e-06, 'learning_rate/action_model': 2.806041311702688e-05, 'epoch': 2.15}) +[2026-06-18 00:05:20,948][starVLA.training.train_starvla][INFO] - Step 4738, Loss: {'train/loss': 0.4647022485733032, 'train/grad_norm_pre_clip': 32.25732421875, 'timing/data': 0.0004525454714894295, 'timing/model': 0.9632339393720031, 'learning_rate/qwen_vl_interface': 2.8041871015695626e-06, 'learning_rate/action_model': 2.8041871015695626e-05, 'epoch': 2.15}) +[2026-06-18 00:05:24,820][starVLA.training.train_starvla][INFO] - Step 4739, Loss: {'train/loss': 0.34209883213043213, 'train/grad_norm_pre_clip': 23.619230270385742, 'timing/data': 0.0004974342882633209, 'timing/model': 0.9472213294357061, 'learning_rate/qwen_vl_interface': 2.802333398456086e-06, 'learning_rate/action_model': 2.8023333984560856e-05, 'epoch': 2.15}) +[2026-06-18 00:05:28,727][starVLA.training.train_starvla][INFO] - Step 4740, Loss: {'train/loss': 0.14250527322292328, 'train/grad_norm_pre_clip': 25.461673736572266, 'timing/data': 0.0007724538445472717, 'timing/model': 0.9693000260740519, 'learning_rate/qwen_vl_interface': 2.8004802027465296e-06, 'learning_rate/action_model': 2.8004802027465293e-05, 'epoch': 2.15}) +[2026-06-18 00:05:32,425][starVLA.training.train_starvla][INFO] - Step 4741, Loss: {'train/loss': 0.2129908800125122, 'train/grad_norm_pre_clip': 26.991535186767578, 'timing/data': 0.000876476988196373, 'timing/model': 0.9332172255963087, 'learning_rate/qwen_vl_interface': 2.7986275148250624e-06, 'learning_rate/action_model': 2.7986275148250622e-05, 'epoch': 2.15}) +[2026-06-18 00:05:36,219][starVLA.training.train_starvla][INFO] - Step 4742, Loss: {'train/loss': 0.12834535539150238, 'train/grad_norm_pre_clip': 15.31096076965332, 'timing/data': 0.0005550812929868698, 'timing/model': 1.1469704266637564, 'learning_rate/qwen_vl_interface': 2.7967753350757498e-06, 'learning_rate/action_model': 2.7967753350757497e-05, 'epoch': 2.15}) +[2026-06-18 00:05:39,881][starVLA.training.train_starvla][INFO] - Step 4743, Loss: {'train/loss': 0.42561817169189453, 'train/grad_norm_pre_clip': 24.23138999938965, 'timing/data': 0.0009123915806412697, 'timing/model': 0.931504569016397, 'learning_rate/qwen_vl_interface': 2.7949236638825496e-06, 'learning_rate/action_model': 2.7949236638825495e-05, 'epoch': 2.15}) +[2026-06-18 00:05:44,155][starVLA.training.train_starvla][INFO] - Step 4744, Loss: {'train/loss': 0.1305040568113327, 'train/grad_norm_pre_clip': 21.080007553100586, 'timing/data': 0.0006539551541209221, 'timing/model': 0.9428842552006245, 'learning_rate/qwen_vl_interface': 2.7930725016293155e-06, 'learning_rate/action_model': 2.793072501629315e-05, 'epoch': 2.15}) +[2026-06-18 00:05:47,939][starVLA.training.train_starvla][INFO] - Step 4745, Loss: {'train/loss': 0.3551141321659088, 'train/grad_norm_pre_clip': 34.84550094604492, 'timing/data': 0.0004692990332841873, 'timing/model': 0.9262531874701381, 'learning_rate/qwen_vl_interface': 2.7912218486997954e-06, 'learning_rate/action_model': 2.7912218486997954e-05, 'epoch': 2.15}) +[2026-06-18 00:05:51,701][starVLA.training.train_starvla][INFO] - Step 4746, Loss: {'train/loss': 0.29016560316085815, 'train/grad_norm_pre_clip': 21.08564567565918, 'timing/data': 0.0008534947410225868, 'timing/model': 0.9529795860871673, 'learning_rate/qwen_vl_interface': 2.7893717054776327e-06, 'learning_rate/action_model': 2.7893717054776327e-05, 'epoch': 2.15}) +[2026-06-18 00:05:55,426][starVLA.training.train_starvla][INFO] - Step 4747, Loss: {'train/loss': 0.22615686058998108, 'train/grad_norm_pre_clip': 26.627798080444336, 'timing/data': 0.0005514314398169518, 'timing/model': 0.9248123187571764, 'learning_rate/qwen_vl_interface': 2.7875220723463624e-06, 'learning_rate/action_model': 2.7875220723463625e-05, 'epoch': 2.15}) +[2026-06-18 00:05:59,257][starVLA.training.train_starvla][INFO] - Step 4748, Loss: {'train/loss': 0.1111060157418251, 'train/grad_norm_pre_clip': 18.568845748901367, 'timing/data': 0.0004799133166670799, 'timing/model': 1.1742941923439503, 'learning_rate/qwen_vl_interface': 2.7856729496894165e-06, 'learning_rate/action_model': 2.7856729496894162e-05, 'epoch': 2.15}) +[2026-06-18 00:06:02,878][starVLA.training.train_starvla][INFO] - Step 4749, Loss: {'train/loss': 0.36732298135757446, 'train/grad_norm_pre_clip': 25.681943893432617, 'timing/data': 0.0005627050995826721, 'timing/model': 0.9374717026948929, 'learning_rate/qwen_vl_interface': 2.783824337890118e-06, 'learning_rate/action_model': 2.783824337890118e-05, 'epoch': 2.15}) +[2026-06-18 00:06:06,640][starVLA.training.train_starvla][INFO] - Step 4750, Loss: {'train/loss': 0.18308454751968384, 'train/grad_norm_pre_clip': 29.36708641052246, 'timing/data': 0.0009545488283038139, 'timing/model': 0.9283791529014707, 'learning_rate/qwen_vl_interface': 2.7819762373316916e-06, 'learning_rate/action_model': 2.7819762373316917e-05, 'epoch': 2.15}) +[2026-06-18 00:06:11,610][starVLA.training.train_starvla][INFO] - Step 4751, Loss: {'train/loss': 0.2762983739376068, 'train/grad_norm_pre_clip': 39.54450607299805, 'timing/data': 0.0004863422363996506, 'timing/model': 0.9521654052659869, 'learning_rate/qwen_vl_interface': 2.7801286483972458e-06, 'learning_rate/action_model': 2.7801286483972457e-05, 'epoch': 2.16}) +[2026-06-18 00:06:15,426][starVLA.training.train_starvla][INFO] - Step 4752, Loss: {'train/loss': 0.1332981139421463, 'train/grad_norm_pre_clip': 18.08865737915039, 'timing/data': 0.0008374759927392006, 'timing/model': 0.935186673887074, 'learning_rate/qwen_vl_interface': 2.778281571469786e-06, 'learning_rate/action_model': 2.7782815714697858e-05, 'epoch': 2.16}) +[2026-06-18 00:06:19,180][starVLA.training.train_starvla][INFO] - Step 4753, Loss: {'train/loss': 0.10126878321170807, 'train/grad_norm_pre_clip': 14.367722511291504, 'timing/data': 0.001172681339085102, 'timing/model': 0.9240571036934853, 'learning_rate/qwen_vl_interface': 2.776435006932219e-06, 'learning_rate/action_model': 2.7764350069322187e-05, 'epoch': 2.16}) +[2026-06-18 00:06:23,197][starVLA.training.train_starvla][INFO] - Step 4754, Loss: {'train/loss': 0.21365255117416382, 'train/grad_norm_pre_clip': 20.946788787841797, 'timing/data': 0.0012899292632937431, 'timing/model': 0.9449636116623878, 'learning_rate/qwen_vl_interface': 2.7745889551673355e-06, 'learning_rate/action_model': 2.7745889551673353e-05, 'epoch': 2.16}) +[2026-06-18 00:06:27,005][starVLA.training.train_starvla][INFO] - Step 4755, Loss: {'train/loss': 0.4203035235404968, 'train/grad_norm_pre_clip': 27.218311309814453, 'timing/data': 0.0005028154700994492, 'timing/model': 1.1311371438205242, 'learning_rate/qwen_vl_interface': 2.772743416557826e-06, 'learning_rate/action_model': 2.772743416557826e-05, 'epoch': 2.16}) +[2026-06-18 00:06:30,561][starVLA.training.train_starvla][INFO] - Step 4756, Loss: {'train/loss': 0.2154405564069748, 'train/grad_norm_pre_clip': 26.15964126586914, 'timing/data': 0.0005406532436609268, 'timing/model': 0.9202946778386831, 'learning_rate/qwen_vl_interface': 2.770898391486272e-06, 'learning_rate/action_model': 2.7708983914862718e-05, 'epoch': 2.16}) +[2026-06-18 00:06:34,313][starVLA.training.train_starvla][INFO] - Step 4757, Loss: {'train/loss': 0.26498666405677795, 'train/grad_norm_pre_clip': 27.69048500061035, 'timing/data': 0.0005601923912763596, 'timing/model': 0.943665330298245, 'learning_rate/qwen_vl_interface': 2.7690538803351484e-06, 'learning_rate/action_model': 2.7690538803351484e-05, 'epoch': 2.16}) +[2026-06-18 00:06:38,321][starVLA.training.train_starvla][INFO] - Step 4758, Loss: {'train/loss': 0.3583984375, 'train/grad_norm_pre_clip': 24.568326950073242, 'timing/data': 0.00048235151916742325, 'timing/model': 0.9373516365885735, 'learning_rate/qwen_vl_interface': 2.7672098834868253e-06, 'learning_rate/action_model': 2.7672098834868253e-05, 'epoch': 2.16}) +[2026-06-18 00:06:42,057][starVLA.training.train_starvla][INFO] - Step 4759, Loss: {'train/loss': 0.2840189039707184, 'train/grad_norm_pre_clip': 28.204971313476562, 'timing/data': 0.0005510104820132256, 'timing/model': 0.9250167198479176, 'learning_rate/qwen_vl_interface': 2.7653664013235655e-06, 'learning_rate/action_model': 2.7653664013235653e-05, 'epoch': 2.16}) +[2026-06-18 00:06:45,902][starVLA.training.train_starvla][INFO] - Step 4760, Loss: {'train/loss': 0.2589039206504822, 'train/grad_norm_pre_clip': 33.58364486694336, 'timing/data': 0.00047664251178503036, 'timing/model': 0.9477015705779195, 'learning_rate/qwen_vl_interface': 2.763523434227524e-06, 'learning_rate/action_model': 2.763523434227524e-05, 'epoch': 2.16}) +[2026-06-18 00:06:49,776][starVLA.training.train_starvla][INFO] - Step 4761, Loss: {'train/loss': 0.5066924691200256, 'train/grad_norm_pre_clip': 28.843080520629883, 'timing/data': 0.00044622551649808884, 'timing/model': 0.9281414551660419, 'learning_rate/qwen_vl_interface': 2.76168098258075e-06, 'learning_rate/action_model': 2.76168098258075e-05, 'epoch': 2.16}) +[2026-06-18 00:06:53,655][starVLA.training.train_starvla][INFO] - Step 4762, Loss: {'train/loss': 0.2456497848033905, 'train/grad_norm_pre_clip': 26.976116180419922, 'timing/data': 0.0007155193015933037, 'timing/model': 1.113925433717668, 'learning_rate/qwen_vl_interface': 2.7598390467651865e-06, 'learning_rate/action_model': 2.7598390467651863e-05, 'epoch': 2.16}) +[2026-06-18 00:06:57,375][starVLA.training.train_starvla][INFO] - Step 4763, Loss: {'train/loss': 0.18777604401111603, 'train/grad_norm_pre_clip': 13.119646072387695, 'timing/data': 0.0004860805347561836, 'timing/model': 0.9814283614978194, 'learning_rate/qwen_vl_interface': 2.757997627162665e-06, 'learning_rate/action_model': 2.7579976271626652e-05, 'epoch': 2.16}) +[2026-06-18 00:07:01,187][starVLA.training.train_starvla][INFO] - Step 4764, Loss: {'train/loss': 0.15843357145786285, 'train/grad_norm_pre_clip': 16.86331558227539, 'timing/data': 0.0005344673991203308, 'timing/model': 0.9341958565637469, 'learning_rate/qwen_vl_interface': 2.7561567241549203e-06, 'learning_rate/action_model': 2.7561567241549202e-05, 'epoch': 2.16}) +[2026-06-18 00:07:05,025][starVLA.training.train_starvla][INFO] - Step 4765, Loss: {'train/loss': 0.164455845952034, 'train/grad_norm_pre_clip': 17.986207962036133, 'timing/data': 0.0005169864743947983, 'timing/model': 0.9917300678789616, 'learning_rate/qwen_vl_interface': 2.7543163381235705e-06, 'learning_rate/action_model': 2.7543163381235702e-05, 'epoch': 2.16}) +[2026-06-18 00:07:08,893][starVLA.training.train_starvla][INFO] - Step 4766, Loss: {'train/loss': 0.19856247305870056, 'train/grad_norm_pre_clip': 24.30853271484375, 'timing/data': 0.0004915716126561165, 'timing/model': 1.0203598644584417, 'learning_rate/qwen_vl_interface': 2.752476469450128e-06, 'learning_rate/action_model': 2.7524764694501283e-05, 'epoch': 2.16}) +[2026-06-18 00:07:12,732][starVLA.training.train_starvla][INFO] - Step 4767, Loss: {'train/loss': 0.230185404419899, 'train/grad_norm_pre_clip': 19.16001319885254, 'timing/data': 0.0006593763828277588, 'timing/model': 0.9708485016599298, 'learning_rate/qwen_vl_interface': 2.750637118515999e-06, 'learning_rate/action_model': 2.750637118515999e-05, 'epoch': 2.16}) +[2026-06-18 00:07:16,674][starVLA.training.train_starvla][INFO] - Step 4768, Loss: {'train/loss': 0.1697486788034439, 'train/grad_norm_pre_clip': 15.58619499206543, 'timing/data': 0.00047806277871131897, 'timing/model': 1.0827001985162497, 'learning_rate/qwen_vl_interface': 2.7487982857024875e-06, 'learning_rate/action_model': 2.7487982857024876e-05, 'epoch': 2.16}) +[2026-06-18 00:07:20,642][starVLA.training.train_starvla][INFO] - Step 4769, Loss: {'train/loss': 0.3992173373699188, 'train/grad_norm_pre_clip': 27.650434494018555, 'timing/data': 0.0005324613302946091, 'timing/model': 1.1983385952189565, 'learning_rate/qwen_vl_interface': 2.746959971390782e-06, 'learning_rate/action_model': 2.7469599713907817e-05, 'epoch': 2.16}) +[2026-06-18 00:07:24,487][starVLA.training.train_starvla][INFO] - Step 4770, Loss: {'train/loss': 0.32767876982688904, 'train/grad_norm_pre_clip': 23.7023983001709, 'timing/data': 0.0004921825602650642, 'timing/model': 0.939395060762763, 'learning_rate/qwen_vl_interface': 2.745122175961967e-06, 'learning_rate/action_model': 2.7451221759619672e-05, 'epoch': 2.16}) +[2026-06-18 00:07:28,265][starVLA.training.train_starvla][INFO] - Step 4771, Loss: {'train/loss': 0.22553320229053497, 'train/grad_norm_pre_clip': 17.887035369873047, 'timing/data': 0.0005895290523767471, 'timing/model': 0.9392130337655544, 'learning_rate/qwen_vl_interface': 2.743284899797022e-06, 'learning_rate/action_model': 2.7432848997970218e-05, 'epoch': 2.16}) +[2026-06-18 00:07:32,038][starVLA.training.train_starvla][INFO] - Step 4772, Loss: {'train/loss': 0.09347102046012878, 'train/grad_norm_pre_clip': 12.831687927246094, 'timing/data': 0.0004491908475756645, 'timing/model': 0.9520252039656043, 'learning_rate/qwen_vl_interface': 2.741448143276813e-06, 'learning_rate/action_model': 2.741448143276813e-05, 'epoch': 2.16}) +[2026-06-18 00:07:35,998][starVLA.training.train_starvla][INFO] - Step 4773, Loss: {'train/loss': 0.1777953803539276, 'train/grad_norm_pre_clip': 20.217023849487305, 'timing/data': 0.0005425885319709778, 'timing/model': 0.9331974452361465, 'learning_rate/qwen_vl_interface': 2.7396119067821043e-06, 'learning_rate/action_model': 2.7396119067821043e-05, 'epoch': 2.17}) +[2026-06-18 00:07:39,844][starVLA.training.train_starvla][INFO] - Step 4774, Loss: {'train/loss': 0.08460415154695511, 'train/grad_norm_pre_clip': 22.025638580322266, 'timing/data': 0.00050318893045187, 'timing/model': 0.919374430552125, 'learning_rate/qwen_vl_interface': 2.737776190693548e-06, 'learning_rate/action_model': 2.7377761906935477e-05, 'epoch': 2.17}) +[2026-06-18 00:07:43,584][starVLA.training.train_starvla][INFO] - Step 4775, Loss: {'train/loss': 0.5031927824020386, 'train/grad_norm_pre_clip': 41.812896728515625, 'timing/data': 0.000433359295129776, 'timing/model': 0.9287098851054907, 'learning_rate/qwen_vl_interface': 2.7359409953916914e-06, 'learning_rate/action_model': 2.7359409953916913e-05, 'epoch': 2.17}) +[2026-06-18 00:07:47,363][starVLA.training.train_starvla][INFO] - Step 4776, Loss: {'train/loss': 0.2640901207923889, 'train/grad_norm_pre_clip': 18.870508193969727, 'timing/data': 0.0004910556599497795, 'timing/model': 1.1204726221039891, 'learning_rate/qwen_vl_interface': 2.734106321256971e-06, 'learning_rate/action_model': 2.7341063212569705e-05, 'epoch': 2.17}) +[2026-06-18 00:07:51,023][starVLA.training.train_starvla][INFO] - Step 4777, Loss: {'train/loss': 0.12005125731229782, 'train/grad_norm_pre_clip': 16.489709854125977, 'timing/data': 0.0004744855687022209, 'timing/model': 0.9273246750235558, 'learning_rate/qwen_vl_interface': 2.732272168669718e-06, 'learning_rate/action_model': 2.732272168669718e-05, 'epoch': 2.17}) +[2026-06-18 00:07:55,366][starVLA.training.train_starvla][INFO] - Step 4778, Loss: {'train/loss': 0.2615611255168915, 'train/grad_norm_pre_clip': 26.10257339477539, 'timing/data': 0.0005073696374893188, 'timing/model': 1.0076025314629078, 'learning_rate/qwen_vl_interface': 2.7304385380101512e-06, 'learning_rate/action_model': 2.730438538010151e-05, 'epoch': 2.17}) +[2026-06-18 00:08:00,438][starVLA.training.train_starvla][INFO] - Step 4779, Loss: {'train/loss': 0.2934403121471405, 'train/grad_norm_pre_clip': 40.376155853271484, 'timing/data': 0.0012530144304037094, 'timing/model': 0.9222601894289255, 'learning_rate/qwen_vl_interface': 2.7286054296583904e-06, 'learning_rate/action_model': 2.72860542965839e-05, 'epoch': 2.17}) +[2026-06-18 00:08:04,256][starVLA.training.train_starvla][INFO] - Step 4780, Loss: {'train/loss': 0.04956800490617752, 'train/grad_norm_pre_clip': 7.747035503387451, 'timing/data': 0.0005010124295949936, 'timing/model': 0.9209079807624221, 'learning_rate/qwen_vl_interface': 2.726772843994435e-06, 'learning_rate/action_model': 2.7267728439944352e-05, 'epoch': 2.17}) +[2026-06-18 00:08:08,044][starVLA.training.train_starvla][INFO] - Step 4781, Loss: {'train/loss': 0.17539899051189423, 'train/grad_norm_pre_clip': 20.33022117614746, 'timing/data': 0.00048342812806367874, 'timing/model': 0.9327893797308207, 'learning_rate/qwen_vl_interface': 2.7249407813981834e-06, 'learning_rate/action_model': 2.724940781398183e-05, 'epoch': 2.17}) +[2026-06-18 00:08:11,829][starVLA.training.train_starvla][INFO] - Step 4782, Loss: {'train/loss': 0.07417075335979462, 'train/grad_norm_pre_clip': 11.372307777404785, 'timing/data': 0.0004353681579232216, 'timing/model': 0.9622530909255147, 'learning_rate/qwen_vl_interface': 2.7231092422494255e-06, 'learning_rate/action_model': 2.7231092422494253e-05, 'epoch': 2.17}) +[2026-06-18 00:08:15,692][starVLA.training.train_starvla][INFO] - Step 4783, Loss: {'train/loss': 0.07217857241630554, 'train/grad_norm_pre_clip': 12.38785457611084, 'timing/data': 0.0004292512312531471, 'timing/model': 1.1169216400012374, 'learning_rate/qwen_vl_interface': 2.7212782269278417e-06, 'learning_rate/action_model': 2.7212782269278413e-05, 'epoch': 2.17}) +[2026-06-18 00:08:19,352][starVLA.training.train_starvla][INFO] - Step 4784, Loss: {'train/loss': 0.07618486881256104, 'train/grad_norm_pre_clip': 13.650428771972656, 'timing/data': 0.0009918129071593285, 'timing/model': 0.9511905955150723, 'learning_rate/qwen_vl_interface': 2.719447735813004e-06, 'learning_rate/action_model': 2.719447735813004e-05, 'epoch': 2.17}) +[2026-06-18 00:08:23,138][starVLA.training.train_starvla][INFO] - Step 4785, Loss: {'train/loss': 0.17068064212799072, 'train/grad_norm_pre_clip': 19.756103515625, 'timing/data': 0.0008753808215260506, 'timing/model': 0.9387109242379665, 'learning_rate/qwen_vl_interface': 2.7176177692843697e-06, 'learning_rate/action_model': 2.7176177692843697e-05, 'epoch': 2.17}) +[2026-06-18 00:08:26,928][starVLA.training.train_starvla][INFO] - Step 4786, Loss: {'train/loss': 0.11436053365468979, 'train/grad_norm_pre_clip': 21.282848358154297, 'timing/data': 0.0008537359535694122, 'timing/model': 0.9408593736588955, 'learning_rate/qwen_vl_interface': 2.7157883277212985e-06, 'learning_rate/action_model': 2.7157883277212985e-05, 'epoch': 2.17}) +[2026-06-18 00:08:30,709][starVLA.training.train_starvla][INFO] - Step 4787, Loss: {'train/loss': 0.15981736779212952, 'train/grad_norm_pre_clip': 22.356369018554688, 'timing/data': 0.0004993192851543427, 'timing/model': 0.9401535103097558, 'learning_rate/qwen_vl_interface': 2.7139594115030347e-06, 'learning_rate/action_model': 2.7139594115030348e-05, 'epoch': 2.17}) +[2026-06-18 00:08:34,504][starVLA.training.train_starvla][INFO] - Step 4788, Loss: {'train/loss': 0.3231070041656494, 'train/grad_norm_pre_clip': 21.75028419494629, 'timing/data': 0.00044280849397182465, 'timing/model': 0.9610768435522914, 'learning_rate/qwen_vl_interface': 2.712131021008714e-06, 'learning_rate/action_model': 2.7121310210087137e-05, 'epoch': 2.17}) +[2026-06-18 00:08:38,312][starVLA.training.train_starvla][INFO] - Step 4789, Loss: {'train/loss': 0.27678948640823364, 'train/grad_norm_pre_clip': 22.121822357177734, 'timing/data': 0.0005036722868680954, 'timing/model': 0.953474260866642, 'learning_rate/qwen_vl_interface': 2.7103031566173622e-06, 'learning_rate/action_model': 2.710303156617362e-05, 'epoch': 2.17}) +[2026-06-18 00:08:42,262][starVLA.training.train_starvla][INFO] - Step 4790, Loss: {'train/loss': 0.07928682863712311, 'train/grad_norm_pre_clip': 9.934433937072754, 'timing/data': 0.0008769780397415161, 'timing/model': 1.3282032813876867, 'learning_rate/qwen_vl_interface': 2.7084758187078995e-06, 'learning_rate/action_model': 2.7084758187078997e-05, 'epoch': 2.17}) +[2026-06-18 00:08:45,983][starVLA.training.train_starvla][INFO] - Step 4791, Loss: {'train/loss': 0.17984376847743988, 'train/grad_norm_pre_clip': 20.513797760009766, 'timing/data': 0.0012791343033313751, 'timing/model': 0.926192132756114, 'learning_rate/qwen_vl_interface': 2.7066490076591344e-06, 'learning_rate/action_model': 2.7066490076591343e-05, 'epoch': 2.17}) +[2026-06-18 00:08:51,315][starVLA.training.train_starvla][INFO] - Step 4792, Loss: {'train/loss': 0.2011852115392685, 'train/grad_norm_pre_clip': 26.29635238647461, 'timing/data': 0.0004907790571451187, 'timing/model': 0.9604382179677486, 'learning_rate/qwen_vl_interface': 2.704822723849765e-06, 'learning_rate/action_model': 2.7048227238497648e-05, 'epoch': 2.17}) +[2026-06-18 00:08:55,358][starVLA.training.train_starvla][INFO] - Step 4793, Loss: {'train/loss': 0.2235858142375946, 'train/grad_norm_pre_clip': 21.113527297973633, 'timing/data': 0.0007546134293079376, 'timing/model': 0.9419738696888089, 'learning_rate/qwen_vl_interface': 2.702996967658388e-06, 'learning_rate/action_model': 2.7029969676583876e-05, 'epoch': 2.17}) +[2026-06-18 00:09:01,754][starVLA.training.train_starvla][INFO] - Step 4794, Loss: {'train/loss': 0.12044277787208557, 'train/grad_norm_pre_clip': 19.343307495117188, 'timing/data': 0.00047903694212436676, 'timing/model': 1.8168537989258766, 'learning_rate/qwen_vl_interface': 2.7011717394634767e-06, 'learning_rate/action_model': 2.701171739463477e-05, 'epoch': 2.17}) +[2026-06-18 00:09:07,793][starVLA.training.train_starvla][INFO] - Step 4795, Loss: {'train/loss': 0.47963330149650574, 'train/grad_norm_pre_clip': 27.439393997192383, 'timing/data': 0.0005208803340792656, 'timing/model': 1.8207623129710555, 'learning_rate/qwen_vl_interface': 2.699347039643408e-06, 'learning_rate/action_model': 2.699347039643408e-05, 'epoch': 2.18}) +[2026-06-18 00:09:11,643][starVLA.training.train_starvla][INFO] - Step 4796, Loss: {'train/loss': 0.12845729291439056, 'train/grad_norm_pre_clip': 32.480960845947266, 'timing/data': 0.0008576065301895142, 'timing/model': 1.1377141289412975, 'learning_rate/qwen_vl_interface': 2.6975228685764398e-06, 'learning_rate/action_model': 2.69752286857644e-05, 'epoch': 2.18}) +[2026-06-18 00:09:15,657][starVLA.training.train_starvla][INFO] - Step 4797, Loss: {'train/loss': 0.13112132251262665, 'train/grad_norm_pre_clip': 22.24087905883789, 'timing/data': 0.0009171077981591225, 'timing/model': 1.2709218468517065, 'learning_rate/qwen_vl_interface': 2.6956992266407307e-06, 'learning_rate/action_model': 2.6956992266407306e-05, 'epoch': 2.18}) +[2026-06-18 00:09:19,681][starVLA.training.train_starvla][INFO] - Step 4798, Loss: {'train/loss': 0.37806111574172974, 'train/grad_norm_pre_clip': 24.147497177124023, 'timing/data': 0.00046298839151859283, 'timing/model': 0.9428102010861039, 'learning_rate/qwen_vl_interface': 2.693876114214321e-06, 'learning_rate/action_model': 2.6938761142143208e-05, 'epoch': 2.18}) +[2026-06-18 00:09:24,173][starVLA.training.train_starvla][INFO] - Step 4799, Loss: {'train/loss': 0.10671661794185638, 'train/grad_norm_pre_clip': 17.92930793762207, 'timing/data': 0.0006563430652022362, 'timing/model': 0.9333390025421977, 'learning_rate/qwen_vl_interface': 2.6920535316751417e-06, 'learning_rate/action_model': 2.6920535316751416e-05, 'epoch': 2.18}) +[2026-06-18 00:11:29,496][starVLA.training.train_starvla][INFO] - Step 4800, Loss: {'train/loss': 0.08430483937263489, 'train/grad_norm_pre_clip': 16.42705535888672, 'eval/loss': 0.2775426506996155, 'eval/latency_8/loss': 0.2775426506996155, 'eval/demon_fire_precision': 0.9657806293919952, 'eval/demon_fire_recall': 0.9819819819819819, 'eval/demon_fire_f1': 0.9738139248305606, 'eval/demon_fire_rate_model': 0.9106844741235393, 'eval/demon_fire_rate_teacher': 0.8956594323873122, 'eval/demon_left_precision': 0.9030612244897959, 'eval/demon_left_recall': 0.7580299785867237, 'eval/demon_left_f1': 0.8242142025611177, 'eval/demon_left_rate_model': 0.10907067334446299, 'eval/demon_left_rate_teacher': 0.12993878686700056, 'eval/demon_right_precision': 0.8323529411764706, 'eval/demon_right_recall': 0.8575757575757575, 'eval/demon_right_f1': 0.8447761194029851, 'eval/demon_right_rate_model': 0.18920422927100725, 'eval/demon_right_rate_teacher': 0.18363939899833054, 'eval/demon_move_f1': 0.8344951609820515, 'eval/demon_cc_f1': 0.9273668251624096, 'eval/latency_8/demon_fire_precision': 0.9657806293919952, 'eval/latency_8/demon_fire_recall': 0.9819819819819819, 'eval/latency_8/demon_fire_f1': 0.9738139248305606, 'eval/latency_8/demon_fire_rate_model': 0.9106844741235393, 'eval/latency_8/demon_fire_rate_teacher': 0.8956594323873122, 'eval/latency_8/demon_left_precision': 0.9030612244897959, 'eval/latency_8/demon_left_recall': 0.7580299785867237, 'eval/latency_8/demon_left_f1': 0.8242142025611177, 'eval/latency_8/demon_left_rate_model': 0.10907067334446299, 'eval/latency_8/demon_left_rate_teacher': 0.12993878686700056, 'eval/latency_8/demon_right_precision': 0.8323529411764706, 'eval/latency_8/demon_right_recall': 0.8575757575757575, 'eval/latency_8/demon_right_f1': 0.8447761194029851, 'eval/latency_8/demon_right_rate_model': 0.18920422927100725, 'eval/latency_8/demon_right_rate_teacher': 0.18363939899833054, 'eval/latency_8/demon_move_f1': 0.8344951609820515, 'eval/latency_8/demon_cc_f1': 0.9273668251624096, 'timing/data': 0.0007429150864481926, 'timing/model': 0.9143667984753847, 'learning_rate/qwen_vl_interface': 2.6902314794010164e-06, 'learning_rate/action_model': 2.6902314794010165e-05, 'epoch': 2.18}) +[2026-06-18 00:11:29,498][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_4800_state +[2026-06-18 00:11:29,503][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 00:11:36,263][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_4800_state/model.safetensors +[2026-06-18 00:11:51,926][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_4800_state/optimizer.bin +[2026-06-18 00:11:51,928][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_4800_state/sampler.bin +[2026-06-18 00:11:51,928][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_4800_state/sampler_1.bin +[2026-06-18 00:11:51,930][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_4800_state/random_states_0.pkl +[2026-06-18 00:11:51,933][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4800 +[2026-06-18 00:11:51,933][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 00:11:51,943][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 00:11:55,887][starVLA.training.train_starvla][INFO] - Step 4801, Loss: {'train/loss': 0.13423678278923035, 'train/grad_norm_pre_clip': 15.242818832397461, 'timing/data': 0.0008311811834573746, 'timing/model': 0.9539157217368484, 'learning_rate/qwen_vl_interface': 2.6884099577696615e-06, 'learning_rate/action_model': 2.6884099577696615e-05, 'epoch': 2.18}) +[2026-06-18 00:11:59,717][starVLA.training.train_starvla][INFO] - Step 4802, Loss: {'train/loss': 0.8415499329566956, 'train/grad_norm_pre_clip': 53.884273529052734, 'timing/data': 0.0005386341363191605, 'timing/model': 0.9445012174546719, 'learning_rate/qwen_vl_interface': 2.686588967158679e-06, 'learning_rate/action_model': 2.6865889671586792e-05, 'epoch': 2.18}) +[2026-06-18 00:12:03,478][starVLA.training.train_starvla][INFO] - Step 4803, Loss: {'train/loss': 0.13474784791469574, 'train/grad_norm_pre_clip': 32.50579833984375, 'timing/data': 0.0005111005157232285, 'timing/model': 0.9243211010470986, 'learning_rate/qwen_vl_interface': 2.6847685079455614e-06, 'learning_rate/action_model': 2.6847685079455613e-05, 'epoch': 2.18}) +[2026-06-18 00:12:07,253][starVLA.training.train_starvla][INFO] - Step 4804, Loss: {'train/loss': 0.22167310118675232, 'train/grad_norm_pre_clip': 27.200267791748047, 'timing/data': 0.00046823639422655106, 'timing/model': 0.9351852759718895, 'learning_rate/qwen_vl_interface': 2.6829485805076927e-06, 'learning_rate/action_model': 2.6829485805076925e-05, 'epoch': 2.18}) +[2026-06-18 00:12:11,037][starVLA.training.train_starvla][INFO] - Step 4805, Loss: {'train/loss': 0.2727539837360382, 'train/grad_norm_pre_clip': 27.93590545654297, 'timing/data': 0.0011461582034826279, 'timing/model': 1.1610215697437525, 'learning_rate/qwen_vl_interface': 2.6811291852223452e-06, 'learning_rate/action_model': 2.6811291852223454e-05, 'epoch': 2.18}) +[2026-06-18 00:12:14,581][starVLA.training.train_starvla][INFO] - Step 4806, Loss: {'train/loss': 0.41812363266944885, 'train/grad_norm_pre_clip': 34.87681579589844, 'timing/data': 0.0004719765856862068, 'timing/model': 0.9220029041171074, 'learning_rate/qwen_vl_interface': 2.679310322466683e-06, 'learning_rate/action_model': 2.6793103224666827e-05, 'epoch': 2.18}) +[2026-06-18 00:12:18,370][starVLA.training.train_starvla][INFO] - Step 4807, Loss: {'train/loss': 0.20095457136631012, 'train/grad_norm_pre_clip': 24.56034278869629, 'timing/data': 0.00051169004291296, 'timing/model': 0.9458045195788145, 'learning_rate/qwen_vl_interface': 2.6774919926177563e-06, 'learning_rate/action_model': 2.677491992617756e-05, 'epoch': 2.18}) +[2026-06-18 00:12:22,799][starVLA.training.train_starvla][INFO] - Step 4808, Loss: {'train/loss': 0.15809860825538635, 'train/grad_norm_pre_clip': 22.895170211791992, 'timing/data': 0.0005882643163204193, 'timing/model': 0.9563992861658335, 'learning_rate/qwen_vl_interface': 2.675674196052507e-06, 'learning_rate/action_model': 2.6756741960525072e-05, 'epoch': 2.18}) +[2026-06-18 00:12:26,755][starVLA.training.train_starvla][INFO] - Step 4809, Loss: {'train/loss': 0.2715611159801483, 'train/grad_norm_pre_clip': 18.672033309936523, 'timing/data': 0.0005049342289566994, 'timing/model': 0.9776624254882336, 'learning_rate/qwen_vl_interface': 2.6738569331477677e-06, 'learning_rate/action_model': 2.6738569331477674e-05, 'epoch': 2.18}) +[2026-06-18 00:12:30,556][starVLA.training.train_starvla][INFO] - Step 4810, Loss: {'train/loss': 0.07427893579006195, 'train/grad_norm_pre_clip': 9.685919761657715, 'timing/data': 0.0004617534577846527, 'timing/model': 0.9932602494955063, 'learning_rate/qwen_vl_interface': 2.672040204280258e-06, 'learning_rate/action_model': 2.672040204280258e-05, 'epoch': 2.18}) +[2026-06-18 00:12:34,380][starVLA.training.train_starvla][INFO] - Step 4811, Loss: {'train/loss': 0.3150872588157654, 'train/grad_norm_pre_clip': 25.98714828491211, 'timing/data': 0.0006204554811120033, 'timing/model': 0.9682228118181229, 'learning_rate/qwen_vl_interface': 2.6702240098265864e-06, 'learning_rate/action_model': 2.6702240098265864e-05, 'epoch': 2.18}) +[2026-06-18 00:12:38,631][starVLA.training.train_starvla][INFO] - Step 4812, Loss: {'train/loss': 0.13808393478393555, 'train/grad_norm_pre_clip': 19.242324829101562, 'timing/data': 0.0006593512371182442, 'timing/model': 1.1334690107032657, 'learning_rate/qwen_vl_interface': 2.6684083501632573e-06, 'learning_rate/action_model': 2.6684083501632574e-05, 'epoch': 2.18}) +[2026-06-18 00:12:42,287][starVLA.training.train_starvla][INFO] - Step 4813, Loss: {'train/loss': 0.25973615050315857, 'train/grad_norm_pre_clip': 28.711042404174805, 'timing/data': 0.0005466211587190628, 'timing/model': 0.9676878033205867, 'learning_rate/qwen_vl_interface': 2.6665932256666525e-06, 'learning_rate/action_model': 2.6665932256666527e-05, 'epoch': 2.18}) +[2026-06-18 00:12:47,532][starVLA.training.train_starvla][INFO] - Step 4814, Loss: {'train/loss': 0.1561780571937561, 'train/grad_norm_pre_clip': 19.338289260864258, 'timing/data': 0.0009990902617573738, 'timing/model': 1.4310233099386096, 'learning_rate/qwen_vl_interface': 2.6647786367130513e-06, 'learning_rate/action_model': 2.664778636713051e-05, 'epoch': 2.18}) +[2026-06-18 00:12:51,495][starVLA.training.train_starvla][INFO] - Step 4815, Loss: {'train/loss': 0.45712387561798096, 'train/grad_norm_pre_clip': 49.548973083496094, 'timing/data': 0.0005034506320953369, 'timing/model': 0.971592060290277, 'learning_rate/qwen_vl_interface': 2.6629645836786217e-06, 'learning_rate/action_model': 2.6629645836786217e-05, 'epoch': 2.18}) +[2026-06-18 00:12:56,411][starVLA.training.train_starvla][INFO] - Step 4816, Loss: {'train/loss': 0.09917180240154266, 'train/grad_norm_pre_clip': 15.68514633178711, 'timing/data': 0.000768173485994339, 'timing/model': 0.9881972568109632, 'learning_rate/qwen_vl_interface': 2.661151066939417e-06, 'learning_rate/action_model': 2.6611510669394168e-05, 'epoch': 2.18}) +[2026-06-18 00:13:00,330][starVLA.training.train_starvla][INFO] - Step 4817, Loss: {'train/loss': 0.17913514375686646, 'train/grad_norm_pre_clip': 17.57512855529785, 'timing/data': 0.0005690557882189751, 'timing/model': 0.943466984666884, 'learning_rate/qwen_vl_interface': 2.6593380868713843e-06, 'learning_rate/action_model': 2.659338086871384e-05, 'epoch': 2.19}) +[2026-06-18 00:13:04,296][starVLA.training.train_starvla][INFO] - Step 4818, Loss: {'train/loss': 0.2249225527048111, 'train/grad_norm_pre_clip': 21.536664962768555, 'timing/data': 0.0010861046612262726, 'timing/model': 1.1016662092879415, 'learning_rate/qwen_vl_interface': 2.6575256438503507e-06, 'learning_rate/action_model': 2.6575256438503504e-05, 'epoch': 2.19}) +[2026-06-18 00:13:08,116][starVLA.training.train_starvla][INFO] - Step 4819, Loss: {'train/loss': 0.1040886864066124, 'train/grad_norm_pre_clip': 18.42900276184082, 'timing/data': 0.0006188107654452324, 'timing/model': 1.1317034708335996, 'learning_rate/qwen_vl_interface': 2.6557137382520413e-06, 'learning_rate/action_model': 2.6557137382520412e-05, 'epoch': 2.19}) +[2026-06-18 00:13:11,723][starVLA.training.train_starvla][INFO] - Step 4820, Loss: {'train/loss': 0.320012629032135, 'train/grad_norm_pre_clip': 29.29386329650879, 'timing/data': 0.0005070166662335396, 'timing/model': 0.9199014967307448, 'learning_rate/qwen_vl_interface': 2.6539023704520666e-06, 'learning_rate/action_model': 2.6539023704520664e-05, 'epoch': 2.19}) +[2026-06-18 00:13:15,584][starVLA.training.train_starvla][INFO] - Step 4821, Loss: {'train/loss': 0.1533556580543518, 'train/grad_norm_pre_clip': 22.633071899414062, 'timing/data': 0.0004635229706764221, 'timing/model': 0.9295450570061803, 'learning_rate/qwen_vl_interface': 2.6520915408259234e-06, 'learning_rate/action_model': 2.6520915408259235e-05, 'epoch': 2.19}) +[2026-06-18 00:13:19,486][starVLA.training.train_starvla][INFO] - Step 4822, Loss: {'train/loss': 0.24449950456619263, 'train/grad_norm_pre_clip': 21.575868606567383, 'timing/data': 0.0004900358617305756, 'timing/model': 0.9284663582220674, 'learning_rate/qwen_vl_interface': 2.650281249749e-06, 'learning_rate/action_model': 2.6502812497489998e-05, 'epoch': 2.19}) +[2026-06-18 00:13:24,190][starVLA.training.train_starvla][INFO] - Step 4823, Loss: {'train/loss': 0.09090347588062286, 'train/grad_norm_pre_clip': 21.727643966674805, 'timing/data': 0.0005325926467776299, 'timing/model': 0.9606091948226094, 'learning_rate/qwen_vl_interface': 2.6484714975965704e-06, 'learning_rate/action_model': 2.6484714975965703e-05, 'epoch': 2.19}) +[2026-06-18 00:13:28,185][starVLA.training.train_starvla][INFO] - Step 4824, Loss: {'train/loss': 0.04181631654500961, 'train/grad_norm_pre_clip': 10.291580200195312, 'timing/data': 0.0006813723593950272, 'timing/model': 1.0127312457188964, 'learning_rate/qwen_vl_interface': 2.646662284743799e-06, 'learning_rate/action_model': 2.646662284743799e-05, 'epoch': 2.19}) +[2026-06-18 00:13:33,434][starVLA.training.train_starvla][INFO] - Step 4825, Loss: {'train/loss': 0.36860334873199463, 'train/grad_norm_pre_clip': 32.857666015625, 'timing/data': 0.00048709288239479065, 'timing/model': 2.554576175287366, 'learning_rate/qwen_vl_interface': 2.6448536115657353e-06, 'learning_rate/action_model': 2.6448536115657353e-05, 'epoch': 2.19}) +[2026-06-18 00:13:37,414][starVLA.training.train_starvla][INFO] - Step 4826, Loss: {'train/loss': 0.3702244162559509, 'train/grad_norm_pre_clip': 27.279094696044922, 'timing/data': 0.0005220025777816772, 'timing/model': 1.249263278208673, 'learning_rate/qwen_vl_interface': 2.6430454784373248e-06, 'learning_rate/action_model': 2.643045478437325e-05, 'epoch': 2.19}) +[2026-06-18 00:13:41,079][starVLA.training.train_starvla][INFO] - Step 4827, Loss: {'train/loss': 0.1917383074760437, 'train/grad_norm_pre_clip': 29.248403549194336, 'timing/data': 0.0004802914336323738, 'timing/model': 0.9248116780072451, 'learning_rate/qwen_vl_interface': 2.6412378857333903e-06, 'learning_rate/action_model': 2.6412378857333903e-05, 'epoch': 2.19}) +[2026-06-18 00:13:47,425][starVLA.training.train_starvla][INFO] - Step 4828, Loss: {'train/loss': 0.2981083393096924, 'train/grad_norm_pre_clip': 27.115530014038086, 'timing/data': 0.0005305353552103043, 'timing/model': 0.9489908665418625, 'learning_rate/qwen_vl_interface': 2.639430833828649e-06, 'learning_rate/action_model': 2.6394308338286488e-05, 'epoch': 2.19}) +[2026-06-18 00:13:51,145][starVLA.training.train_starvla][INFO] - Step 4829, Loss: {'train/loss': 0.09202861040830612, 'train/grad_norm_pre_clip': 15.446426391601562, 'timing/data': 0.0004913685843348503, 'timing/model': 0.9302129298448563, 'learning_rate/qwen_vl_interface': 2.637624323097703e-06, 'learning_rate/action_model': 2.6376243230977026e-05, 'epoch': 2.19}) +[2026-06-18 00:13:54,951][starVLA.training.train_starvla][INFO] - Step 4830, Loss: {'train/loss': 0.3480168581008911, 'train/grad_norm_pre_clip': 29.238706588745117, 'timing/data': 0.0010102596133947372, 'timing/model': 0.9506778409704566, 'learning_rate/qwen_vl_interface': 2.635818353915048e-06, 'learning_rate/action_model': 2.635818353915048e-05, 'epoch': 2.19}) +[2026-06-18 00:13:59,650][starVLA.training.train_starvla][INFO] - Step 4831, Loss: {'train/loss': 0.3452998399734497, 'train/grad_norm_pre_clip': 33.222084045410156, 'timing/data': 0.0005581201985478401, 'timing/model': 0.9506748486310244, 'learning_rate/qwen_vl_interface': 2.634012926655063e-06, 'learning_rate/action_model': 2.6340129266550628e-05, 'epoch': 2.19}) +[2026-06-18 00:14:03,774][starVLA.training.train_starvla][INFO] - Step 4832, Loss: {'train/loss': 0.05231994390487671, 'train/grad_norm_pre_clip': 8.734227180480957, 'timing/data': 0.0011746613308787346, 'timing/model': 1.1346618989482522, 'learning_rate/qwen_vl_interface': 2.632208041692009e-06, 'learning_rate/action_model': 2.6322080416920088e-05, 'epoch': 2.19}) +[2026-06-18 00:14:07,546][starVLA.training.train_starvla][INFO] - Step 4833, Loss: {'train/loss': 0.28777796030044556, 'train/grad_norm_pre_clip': 33.32963180541992, 'timing/data': 0.0005743950605392456, 'timing/model': 1.1103912284597754, 'learning_rate/qwen_vl_interface': 2.630403699400047e-06, 'learning_rate/action_model': 2.6304036994000468e-05, 'epoch': 2.19}) +[2026-06-18 00:14:11,095][starVLA.training.train_starvla][INFO] - Step 4834, Loss: {'train/loss': 0.34804296493530273, 'train/grad_norm_pre_clip': 49.75871658325195, 'timing/data': 0.00048715248703956604, 'timing/model': 0.9198694890365005, 'learning_rate/qwen_vl_interface': 2.628599900153217e-06, 'learning_rate/action_model': 2.628599900153217e-05, 'epoch': 2.19}) +[2026-06-18 00:14:14,921][starVLA.training.train_starvla][INFO] - Step 4835, Loss: {'train/loss': 0.17611268162727356, 'train/grad_norm_pre_clip': 24.587705612182617, 'timing/data': 0.0013717804104089737, 'timing/model': 0.9469098607078195, 'learning_rate/qwen_vl_interface': 2.6267966443254484e-06, 'learning_rate/action_model': 2.6267966443254483e-05, 'epoch': 2.19}) +[2026-06-18 00:14:18,758][starVLA.training.train_starvla][INFO] - Step 4836, Loss: {'train/loss': 0.17981508374214172, 'train/grad_norm_pre_clip': 25.318456649780273, 'timing/data': 0.00045730452984571457, 'timing/model': 0.9750429261475801, 'learning_rate/qwen_vl_interface': 2.6249939322905576e-06, 'learning_rate/action_model': 2.6249939322905576e-05, 'epoch': 2.19}) +[2026-06-18 00:14:22,484][starVLA.training.train_starvla][INFO] - Step 4837, Loss: {'train/loss': 0.24676279723644257, 'train/grad_norm_pre_clip': 19.36298942565918, 'timing/data': 0.0004665059968829155, 'timing/model': 0.9564504092559218, 'learning_rate/qwen_vl_interface': 2.6231917644222505e-06, 'learning_rate/action_model': 2.6231917644222504e-05, 'epoch': 2.19}) +[2026-06-18 00:14:26,467][starVLA.training.train_starvla][INFO] - Step 4838, Loss: {'train/loss': 0.2920758128166199, 'train/grad_norm_pre_clip': 29.163320541381836, 'timing/data': 0.0005356092005968094, 'timing/model': 0.9882928412407637, 'learning_rate/qwen_vl_interface': 2.621390141094116e-06, 'learning_rate/action_model': 2.621390141094116e-05, 'epoch': 2.19}) +[2026-06-18 00:14:30,327][starVLA.training.train_starvla][INFO] - Step 4839, Loss: {'train/loss': 0.14125318825244904, 'train/grad_norm_pre_clip': 11.700118064880371, 'timing/data': 0.0005028983578085899, 'timing/model': 1.0940767657011747, 'learning_rate/qwen_vl_interface': 2.6195890626796345e-06, 'learning_rate/action_model': 2.6195890626796345e-05, 'epoch': 2.2}) +[2026-06-18 00:14:34,856][starVLA.training.train_starvla][INFO] - Step 4840, Loss: {'train/loss': 0.45536819100379944, 'train/grad_norm_pre_clip': 44.888668060302734, 'timing/data': 0.00046465080231428146, 'timing/model': 1.8844769978895783, 'learning_rate/qwen_vl_interface': 2.6177885295521706e-06, 'learning_rate/action_model': 2.6177885295521702e-05, 'epoch': 2.2}) +[2026-06-18 00:14:39,340][starVLA.training.train_starvla][INFO] - Step 4841, Loss: {'train/loss': 0.2766716778278351, 'train/grad_norm_pre_clip': 24.56887435913086, 'timing/data': 0.0005516102537512779, 'timing/model': 0.9273630697280169, 'learning_rate/qwen_vl_interface': 2.6159885420849766e-06, 'learning_rate/action_model': 2.6159885420849766e-05, 'epoch': 2.2}) +[2026-06-18 00:14:43,036][starVLA.training.train_starvla][INFO] - Step 4842, Loss: {'train/loss': 0.3751644194126129, 'train/grad_norm_pre_clip': 21.354637145996094, 'timing/data': 0.0004788422957062721, 'timing/model': 0.9260400524362922, 'learning_rate/qwen_vl_interface': 2.6141891006511927e-06, 'learning_rate/action_model': 2.6141891006511925e-05, 'epoch': 2.2}) +[2026-06-18 00:14:46,823][starVLA.training.train_starvla][INFO] - Step 4843, Loss: {'train/loss': 0.6697171330451965, 'train/grad_norm_pre_clip': 36.66227722167969, 'timing/data': 0.0004756757989525795, 'timing/model': 0.947762044146657, 'learning_rate/qwen_vl_interface': 2.612390205623842e-06, 'learning_rate/action_model': 2.612390205623842e-05, 'epoch': 2.2}) +[2026-06-18 00:14:50,569][starVLA.training.train_starvla][INFO] - Step 4844, Loss: {'train/loss': 0.042626895010471344, 'train/grad_norm_pre_clip': 10.456128120422363, 'timing/data': 0.001188710331916809, 'timing/model': 0.9276836980134249, 'learning_rate/qwen_vl_interface': 2.6105918573758425e-06, 'learning_rate/action_model': 2.6105918573758427e-05, 'epoch': 2.2}) +[2026-06-18 00:14:55,109][starVLA.training.train_starvla][INFO] - Step 4845, Loss: {'train/loss': 0.14947275817394257, 'train/grad_norm_pre_clip': 14.15329360961914, 'timing/data': 0.0005006026476621628, 'timing/model': 1.828938815742731, 'learning_rate/qwen_vl_interface': 2.608794056279991e-06, 'learning_rate/action_model': 2.6087940562799906e-05, 'epoch': 2.2}) +[2026-06-18 00:14:58,877][starVLA.training.train_starvla][INFO] - Step 4846, Loss: {'train/loss': 0.5180199146270752, 'train/grad_norm_pre_clip': 45.66769027709961, 'timing/data': 0.0007351357489824295, 'timing/model': 1.0287143802270293, 'learning_rate/qwen_vl_interface': 2.6069968027089744e-06, 'learning_rate/action_model': 2.6069968027089742e-05, 'epoch': 2.2}) +[2026-06-18 00:15:02,588][starVLA.training.train_starvla][INFO] - Step 4847, Loss: {'train/loss': 0.2141016274690628, 'train/grad_norm_pre_clip': 21.509105682373047, 'timing/data': 0.0005564698949456215, 'timing/model': 0.9662195555865765, 'learning_rate/qwen_vl_interface': 2.605200097035362e-06, 'learning_rate/action_model': 2.605200097035362e-05, 'epoch': 2.2}) +[2026-06-18 00:15:06,550][starVLA.training.train_starvla][INFO] - Step 4848, Loss: {'train/loss': 0.3082420229911804, 'train/grad_norm_pre_clip': 27.06574821472168, 'timing/data': 0.000984223559498787, 'timing/model': 0.9503269772976637, 'learning_rate/qwen_vl_interface': 2.6034039396316167e-06, 'learning_rate/action_model': 2.6034039396316167e-05, 'epoch': 2.2}) +[2026-06-18 00:15:10,572][starVLA.training.train_starvla][INFO] - Step 4849, Loss: {'train/loss': 0.3058944046497345, 'train/grad_norm_pre_clip': 34.241905212402344, 'timing/data': 0.00047559943050146103, 'timing/model': 1.1089303391054273, 'learning_rate/qwen_vl_interface': 2.601608330870083e-06, 'learning_rate/action_model': 2.601608330870083e-05, 'epoch': 2.2}) +[2026-06-18 00:15:14,328][starVLA.training.train_starvla][INFO] - Step 4850, Loss: {'train/loss': 0.4077816307544708, 'train/grad_norm_pre_clip': 31.70283317565918, 'timing/data': 0.0005178377032279968, 'timing/model': 0.9280918017029762, 'learning_rate/qwen_vl_interface': 2.5998132711229918e-06, 'learning_rate/action_model': 2.5998132711229915e-05, 'epoch': 2.2}) +[2026-06-18 00:15:18,161][starVLA.training.train_starvla][INFO] - Step 4851, Loss: {'train/loss': 0.4079481065273285, 'train/grad_norm_pre_clip': 32.03044891357422, 'timing/data': 0.0009698206558823586, 'timing/model': 0.9285590434446931, 'learning_rate/qwen_vl_interface': 2.598018760762461e-06, 'learning_rate/action_model': 2.598018760762461e-05, 'epoch': 2.2}) +[2026-06-18 00:15:21,920][starVLA.training.train_starvla][INFO] - Step 4852, Loss: {'train/loss': 0.3039049506187439, 'train/grad_norm_pre_clip': 26.185482025146484, 'timing/data': 0.0004896270111203194, 'timing/model': 1.1087839379906654, 'learning_rate/qwen_vl_interface': 2.5962248001604958e-06, 'learning_rate/action_model': 2.5962248001604956e-05, 'epoch': 2.2}) +[2026-06-18 00:15:25,531][starVLA.training.train_starvla][INFO] - Step 4853, Loss: {'train/loss': 0.03036898747086525, 'train/grad_norm_pre_clip': 5.288050174713135, 'timing/data': 0.0009418502449989319, 'timing/model': 0.9414652474224567, 'learning_rate/qwen_vl_interface': 2.5944313896889837e-06, 'learning_rate/action_model': 2.594431389688984e-05, 'epoch': 2.2}) +[2026-06-18 00:15:29,320][starVLA.training.train_starvla][INFO] - Step 4854, Loss: {'train/loss': 0.1255006343126297, 'train/grad_norm_pre_clip': 17.471614837646484, 'timing/data': 0.000466909259557724, 'timing/model': 0.954049302265048, 'learning_rate/qwen_vl_interface': 2.592638529719703e-06, 'learning_rate/action_model': 2.5926385297197033e-05, 'epoch': 2.2}) +[2026-06-18 00:15:33,313][starVLA.training.train_starvla][INFO] - Step 4855, Loss: {'train/loss': 0.13084763288497925, 'train/grad_norm_pre_clip': 19.863893508911133, 'timing/data': 0.0005854954943060875, 'timing/model': 0.9478983711451292, 'learning_rate/qwen_vl_interface': 2.590846220624314e-06, 'learning_rate/action_model': 2.5908462206243138e-05, 'epoch': 2.2}) +[2026-06-18 00:15:37,200][starVLA.training.train_starvla][INFO] - Step 4856, Loss: {'train/loss': 0.10657019913196564, 'train/grad_norm_pre_clip': 12.489738464355469, 'timing/data': 0.0006839465349912643, 'timing/model': 1.0083930576220155, 'learning_rate/qwen_vl_interface': 2.5890544627743652e-06, 'learning_rate/action_model': 2.589054462774365e-05, 'epoch': 2.2}) +[2026-06-18 00:15:41,026][starVLA.training.train_starvla][INFO] - Step 4857, Loss: {'train/loss': 0.17158754169940948, 'train/grad_norm_pre_clip': 18.218929290771484, 'timing/data': 0.0005252258852124214, 'timing/model': 0.9485756875947118, 'learning_rate/qwen_vl_interface': 2.58726325654129e-06, 'learning_rate/action_model': 2.5872632565412898e-05, 'epoch': 2.2}) +[2026-06-18 00:15:44,812][starVLA.training.train_starvla][INFO] - Step 4858, Loss: {'train/loss': 0.036458466202020645, 'train/grad_norm_pre_clip': 7.684119701385498, 'timing/data': 0.0011158939450979233, 'timing/model': 1.1022513546049595, 'learning_rate/qwen_vl_interface': 2.5854726022964043e-06, 'learning_rate/action_model': 2.5854726022964042e-05, 'epoch': 2.2}) +[2026-06-18 00:15:48,620][starVLA.training.train_starvla][INFO] - Step 4859, Loss: {'train/loss': 0.046227097511291504, 'train/grad_norm_pre_clip': 6.920654773712158, 'timing/data': 0.0004540523514151573, 'timing/model': 1.1000156737864017, 'learning_rate/qwen_vl_interface': 2.5836825004109173e-06, 'learning_rate/action_model': 2.5836825004109172e-05, 'epoch': 2.2}) +[2026-06-18 00:15:52,198][starVLA.training.train_starvla][INFO] - Step 4860, Loss: {'train/loss': 0.5439620614051819, 'train/grad_norm_pre_clip': 44.09484100341797, 'timing/data': 0.0004829205572605133, 'timing/model': 0.9188658408820629, 'learning_rate/qwen_vl_interface': 2.5818929512559193e-06, 'learning_rate/action_model': 2.5818929512559193e-05, 'epoch': 2.2}) +[2026-06-18 00:15:55,929][starVLA.training.train_starvla][INFO] - Step 4861, Loss: {'train/loss': 0.168501615524292, 'train/grad_norm_pre_clip': 14.250862121582031, 'timing/data': 0.0012347390875220299, 'timing/model': 0.9178765276446939, 'learning_rate/qwen_vl_interface': 2.580103955202381e-06, 'learning_rate/action_model': 2.580103955202381e-05, 'epoch': 2.21}) +[2026-06-18 00:15:59,617][starVLA.training.train_starvla][INFO] - Step 4862, Loss: {'train/loss': 0.15457770228385925, 'train/grad_norm_pre_clip': 27.08127212524414, 'timing/data': 0.0008820444345474243, 'timing/model': 0.9173378953710198, 'learning_rate/qwen_vl_interface': 2.5783155126211647e-06, 'learning_rate/action_model': 2.5783155126211644e-05, 'epoch': 2.21}) +[2026-06-18 00:16:03,354][starVLA.training.train_starvla][INFO] - Step 4863, Loss: {'train/loss': 0.2316645383834839, 'train/grad_norm_pre_clip': 17.636943817138672, 'timing/data': 0.0006918171420693398, 'timing/model': 0.9294985551387072, 'learning_rate/qwen_vl_interface': 2.576527623883019e-06, 'learning_rate/action_model': 2.576527623883019e-05, 'epoch': 2.21}) +[2026-06-18 00:16:07,078][starVLA.training.train_starvla][INFO] - Step 4864, Loss: {'train/loss': 0.39470160007476807, 'train/grad_norm_pre_clip': 28.416912078857422, 'timing/data': 0.0005189329385757446, 'timing/model': 0.9223064510151744, 'learning_rate/qwen_vl_interface': 2.5747402893585732e-06, 'learning_rate/action_model': 2.574740289358573e-05, 'epoch': 2.21}) +[2026-06-18 00:16:10,940][starVLA.training.train_starvla][INFO] - Step 4865, Loss: {'train/loss': 0.33224186301231384, 'train/grad_norm_pre_clip': 42.992000579833984, 'timing/data': 0.0007114540785551071, 'timing/model': 1.146640382707119, 'learning_rate/qwen_vl_interface': 2.5729535094183445e-06, 'learning_rate/action_model': 2.5729535094183444e-05, 'epoch': 2.21}) +[2026-06-18 00:16:14,892][starVLA.training.train_starvla][INFO] - Step 4866, Loss: {'train/loss': 0.45879825949668884, 'train/grad_norm_pre_clip': 28.856834411621094, 'timing/data': 0.0006479052826762199, 'timing/model': 1.1528006736189127, 'learning_rate/qwen_vl_interface': 2.571167284432734e-06, 'learning_rate/action_model': 2.5711672844327338e-05, 'epoch': 2.21}) +[2026-06-18 00:16:18,655][starVLA.training.train_starvla][INFO] - Step 4867, Loss: {'train/loss': 0.17309033870697021, 'train/grad_norm_pre_clip': 28.788734436035156, 'timing/data': 0.00048084184527397156, 'timing/model': 1.0145951732993126, 'learning_rate/qwen_vl_interface': 2.569381614772028e-06, 'learning_rate/action_model': 2.5693816147720278e-05, 'epoch': 2.21}) +[2026-06-18 00:16:22,440][starVLA.training.train_starvla][INFO] - Step 4868, Loss: {'train/loss': 0.12396963685750961, 'train/grad_norm_pre_clip': 20.118473052978516, 'timing/data': 0.0007985746487975121, 'timing/model': 0.9432849157601595, 'learning_rate/qwen_vl_interface': 2.5675965008063976e-06, 'learning_rate/action_model': 2.5675965008063974e-05, 'epoch': 2.21}) +[2026-06-18 00:16:26,225][starVLA.training.train_starvla][INFO] - Step 4869, Loss: {'train/loss': 0.28009232878685, 'train/grad_norm_pre_clip': 31.652761459350586, 'timing/data': 0.0006701620295643806, 'timing/model': 0.9337442247197032, 'learning_rate/qwen_vl_interface': 2.565811942905899e-06, 'learning_rate/action_model': 2.5658119429058993e-05, 'epoch': 2.21}) +[2026-06-18 00:16:30,063][starVLA.training.train_starvla][INFO] - Step 4870, Loss: {'train/loss': 0.335595965385437, 'train/grad_norm_pre_clip': 25.000404357910156, 'timing/data': 0.00046671926975250244, 'timing/model': 0.9318427471444011, 'learning_rate/qwen_vl_interface': 2.5640279414404723e-06, 'learning_rate/action_model': 2.5640279414404722e-05, 'epoch': 2.21}) +[2026-06-18 00:16:34,128][starVLA.training.train_starvla][INFO] - Step 4871, Loss: {'train/loss': 0.1495332568883896, 'train/grad_norm_pre_clip': 23.856584548950195, 'timing/data': 0.0005532754585146904, 'timing/model': 1.2222842629998922, 'learning_rate/qwen_vl_interface': 2.5622444967799444e-06, 'learning_rate/action_model': 2.5622444967799443e-05, 'epoch': 2.21}) +[2026-06-18 00:16:37,894][starVLA.training.train_starvla][INFO] - Step 4872, Loss: {'train/loss': 0.1499263346195221, 'train/grad_norm_pre_clip': 28.634925842285156, 'timing/data': 0.0005644327029585838, 'timing/model': 1.1194850960746408, 'learning_rate/qwen_vl_interface': 2.5604616092940245e-06, 'learning_rate/action_model': 2.5604616092940242e-05, 'epoch': 2.21}) +[2026-06-18 00:16:41,703][starVLA.training.train_starvla][INFO] - Step 4873, Loss: {'train/loss': 0.2317255586385727, 'train/grad_norm_pre_clip': 26.01340103149414, 'timing/data': 0.000683310441672802, 'timing/model': 1.0948092164471745, 'learning_rate/qwen_vl_interface': 2.5586792793523052e-06, 'learning_rate/action_model': 2.5586792793523052e-05, 'epoch': 2.21}) +[2026-06-18 00:16:45,500][starVLA.training.train_starvla][INFO] - Step 4874, Loss: {'train/loss': 0.336424320936203, 'train/grad_norm_pre_clip': 36.618682861328125, 'timing/data': 0.0004871031269431114, 'timing/model': 1.0245443154126406, 'learning_rate/qwen_vl_interface': 2.556897507324271e-06, 'learning_rate/action_model': 2.5568975073242705e-05, 'epoch': 2.21}) +[2026-06-18 00:16:49,307][starVLA.training.train_starvla][INFO] - Step 4875, Loss: {'train/loss': 0.18652072548866272, 'train/grad_norm_pre_clip': 29.255218505859375, 'timing/data': 0.0006843060255050659, 'timing/model': 0.938313527032733, 'learning_rate/qwen_vl_interface': 2.5551162935792782e-06, 'learning_rate/action_model': 2.5551162935792784e-05, 'epoch': 2.21}) +[2026-06-18 00:16:53,218][starVLA.training.train_starvla][INFO] - Step 4876, Loss: {'train/loss': 0.17997249960899353, 'train/grad_norm_pre_clip': 32.765411376953125, 'timing/data': 0.000563478097319603, 'timing/model': 0.932724948041141, 'learning_rate/qwen_vl_interface': 2.5533356384865777e-06, 'learning_rate/action_model': 2.5533356384865776e-05, 'epoch': 2.21}) +[2026-06-18 00:16:56,923][starVLA.training.train_starvla][INFO] - Step 4877, Loss: {'train/loss': 0.12388340383768082, 'train/grad_norm_pre_clip': 19.172168731689453, 'timing/data': 0.0005325069651007652, 'timing/model': 0.9286708403378725, 'learning_rate/qwen_vl_interface': 2.5515555424153017e-06, 'learning_rate/action_model': 2.5515555424153016e-05, 'epoch': 2.21}) +[2026-06-18 00:17:00,611][starVLA.training.train_starvla][INFO] - Step 4878, Loss: {'train/loss': 0.41915374994277954, 'train/grad_norm_pre_clip': 31.038129806518555, 'timing/data': 0.0004918919876217842, 'timing/model': 0.9288853462785482, 'learning_rate/qwen_vl_interface': 2.5497760057344663e-06, 'learning_rate/action_model': 2.549776005734466e-05, 'epoch': 2.21}) +[2026-06-18 00:17:04,417][starVLA.training.train_starvla][INFO] - Step 4879, Loss: {'train/loss': 0.11606206744909286, 'train/grad_norm_pre_clip': 13.389450073242188, 'timing/data': 0.0005264170467853546, 'timing/model': 1.102942582219839, 'learning_rate/qwen_vl_interface': 2.5479970288129714e-06, 'learning_rate/action_model': 2.5479970288129714e-05, 'epoch': 2.21}) +[2026-06-18 00:17:07,986][starVLA.training.train_starvla][INFO] - Step 4880, Loss: {'train/loss': 0.45953765511512756, 'train/grad_norm_pre_clip': 63.7119026184082, 'timing/data': 0.0005335798487067223, 'timing/model': 0.9267049198970199, 'learning_rate/qwen_vl_interface': 2.546218612019597e-06, 'learning_rate/action_model': 2.5462186120195967e-05, 'epoch': 2.21}) +[2026-06-18 00:17:11,926][starVLA.training.train_starvla][INFO] - Step 4881, Loss: {'train/loss': 0.3300102949142456, 'train/grad_norm_pre_clip': 29.95234489440918, 'timing/data': 0.0008723726496100426, 'timing/model': 0.9377429112792015, 'learning_rate/qwen_vl_interface': 2.544440755723015e-06, 'learning_rate/action_model': 2.5444407557230147e-05, 'epoch': 2.21}) +[2026-06-18 00:17:15,677][starVLA.training.train_starvla][INFO] - Step 4882, Loss: {'train/loss': 0.2613540291786194, 'train/grad_norm_pre_clip': 20.362537384033203, 'timing/data': 0.00047122687101364136, 'timing/model': 0.9188196901232004, 'learning_rate/qwen_vl_interface': 2.5426634602917756e-06, 'learning_rate/action_model': 2.5426634602917755e-05, 'epoch': 2.21}) +[2026-06-18 00:17:19,389][starVLA.training.train_starvla][INFO] - Step 4883, Loss: {'train/loss': 0.22703281044960022, 'train/grad_norm_pre_clip': 27.392940521240234, 'timing/data': 0.000482446514070034, 'timing/model': 0.917498741298914, 'learning_rate/qwen_vl_interface': 2.540886726094315e-06, 'learning_rate/action_model': 2.540886726094315e-05, 'epoch': 2.22}) +[2026-06-18 00:17:23,114][starVLA.training.train_starvla][INFO] - Step 4884, Loss: {'train/loss': 0.09830676019191742, 'train/grad_norm_pre_clip': 13.18932819366455, 'timing/data': 0.000534839928150177, 'timing/model': 0.9244244275614619, 'learning_rate/qwen_vl_interface': 2.5391105534989503e-06, 'learning_rate/action_model': 2.53911055349895e-05, 'epoch': 2.22}) +[2026-06-18 00:17:26,816][starVLA.training.train_starvla][INFO] - Step 4885, Loss: {'train/loss': 0.13979589939117432, 'train/grad_norm_pre_clip': 37.60799026489258, 'timing/data': 0.0004778476431965828, 'timing/model': 1.0775619205087423, 'learning_rate/qwen_vl_interface': 2.537334942873885e-06, 'learning_rate/action_model': 2.5373349428738846e-05, 'epoch': 2.22}) +[2026-06-18 00:17:30,457][starVLA.training.train_starvla][INFO] - Step 4886, Loss: {'train/loss': 0.14843778312206268, 'train/grad_norm_pre_clip': 10.602174758911133, 'timing/data': 0.00046234577894210815, 'timing/model': 1.0186644410714507, 'learning_rate/qwen_vl_interface': 2.535559894587205e-06, 'learning_rate/action_model': 2.535559894587205e-05, 'epoch': 2.22}) +[2026-06-18 00:17:34,038][starVLA.training.train_starvla][INFO] - Step 4887, Loss: {'train/loss': 0.09204676747322083, 'train/grad_norm_pre_clip': 14.58486557006836, 'timing/data': 0.0009000683203339577, 'timing/model': 0.9458781629800797, 'learning_rate/qwen_vl_interface': 2.533785409006878e-06, 'learning_rate/action_model': 2.533785409006878e-05, 'epoch': 2.22}) +[2026-06-18 00:17:37,759][starVLA.training.train_starvla][INFO] - Step 4888, Loss: {'train/loss': 0.22384969890117645, 'train/grad_norm_pre_clip': 21.755159378051758, 'timing/data': 0.0004966557025909424, 'timing/model': 0.9240542072802782, 'learning_rate/qwen_vl_interface': 2.532011486500762e-06, 'learning_rate/action_model': 2.532011486500762e-05, 'epoch': 2.22}) +[2026-06-18 00:17:41,747][starVLA.training.train_starvla][INFO] - Step 4889, Loss: {'train/loss': 0.8102607131004333, 'train/grad_norm_pre_clip': 44.850868225097656, 'timing/data': 0.0005918750539422035, 'timing/model': 0.9489725567400455, 'learning_rate/qwen_vl_interface': 2.530238127436588e-06, 'learning_rate/action_model': 2.5302381274365882e-05, 'epoch': 2.22}) +[2026-06-18 00:17:45,569][starVLA.training.train_starvla][INFO] - Step 4890, Loss: {'train/loss': 0.1602390557527542, 'train/grad_norm_pre_clip': 22.249183654785156, 'timing/data': 0.000501411035656929, 'timing/model': 0.9277829471975565, 'learning_rate/qwen_vl_interface': 2.528465332181977e-06, 'learning_rate/action_model': 2.5284653321819767e-05, 'epoch': 2.22}) +[2026-06-18 00:17:49,317][starVLA.training.train_starvla][INFO] - Step 4891, Loss: {'train/loss': 0.21906346082687378, 'train/grad_norm_pre_clip': 22.376964569091797, 'timing/data': 0.00048553477972745895, 'timing/model': 0.9239808907732368, 'learning_rate/qwen_vl_interface': 2.526693101104429e-06, 'learning_rate/action_model': 2.526693101104429e-05, 'epoch': 2.22}) +[2026-06-18 00:17:53,276][starVLA.training.train_starvla][INFO] - Step 4892, Loss: {'train/loss': 0.1159808412194252, 'train/grad_norm_pre_clip': 23.70794677734375, 'timing/data': 0.0010717902332544327, 'timing/model': 1.204268272034824, 'learning_rate/qwen_vl_interface': 2.524921434571333e-06, 'learning_rate/action_model': 2.524921434571333e-05, 'epoch': 2.22}) +[2026-06-18 00:17:57,111][starVLA.training.train_starvla][INFO] - Step 4893, Loss: {'train/loss': 0.25676384568214417, 'train/grad_norm_pre_clip': 29.837007522583008, 'timing/data': 0.0005051679909229279, 'timing/model': 0.9796766992658377, 'learning_rate/qwen_vl_interface': 2.5231503329499585e-06, 'learning_rate/action_model': 2.5231503329499583e-05, 'epoch': 2.22}) +[2026-06-18 00:18:01,378][starVLA.training.train_starvla][INFO] - Step 4894, Loss: {'train/loss': 0.12205924838781357, 'train/grad_norm_pre_clip': 22.5849552154541, 'timing/data': 0.0006551891565322876, 'timing/model': 0.9293943224474788, 'learning_rate/qwen_vl_interface': 2.5213797966074506e-06, 'learning_rate/action_model': 2.5213797966074505e-05, 'epoch': 2.22}) +[2026-06-18 00:18:05,509][starVLA.training.train_starvla][INFO] - Step 4895, Loss: {'train/loss': 0.324512243270874, 'train/grad_norm_pre_clip': 28.537160873413086, 'timing/data': 0.0006216336041688919, 'timing/model': 0.987630732357502, 'learning_rate/qwen_vl_interface': 2.519609825910848e-06, 'learning_rate/action_model': 2.5196098259108482e-05, 'epoch': 2.22}) +[2026-06-18 00:18:09,809][starVLA.training.train_starvla][INFO] - Step 4896, Loss: {'train/loss': 0.08487074077129364, 'train/grad_norm_pre_clip': 12.760285377502441, 'timing/data': 0.0006751818582415581, 'timing/model': 0.9322730377316475, 'learning_rate/qwen_vl_interface': 2.517840421227067e-06, 'learning_rate/action_model': 2.517840421227067e-05, 'epoch': 2.22}) +[2026-06-18 00:18:13,930][starVLA.training.train_starvla][INFO] - Step 4897, Loss: {'train/loss': 0.09411603957414627, 'train/grad_norm_pre_clip': 15.215333938598633, 'timing/data': 0.0006949584931135178, 'timing/model': 1.1174733871594071, 'learning_rate/qwen_vl_interface': 2.5160715829229067e-06, 'learning_rate/action_model': 2.5160715829229066e-05, 'epoch': 2.22}) +[2026-06-18 00:18:19,112][starVLA.training.train_starvla][INFO] - Step 4898, Loss: {'train/loss': 0.03673563897609711, 'train/grad_norm_pre_clip': 6.335018157958984, 'timing/data': 0.00048614107072353363, 'timing/model': 1.1038341596722603, 'learning_rate/qwen_vl_interface': 2.5143033113650484e-06, 'learning_rate/action_model': 2.5143033113650483e-05, 'epoch': 2.22}) +[2026-06-18 00:18:22,889][starVLA.training.train_starvla][INFO] - Step 4899, Loss: {'train/loss': 0.10435608774423599, 'train/grad_norm_pre_clip': 23.655052185058594, 'timing/data': 0.0007518874481320381, 'timing/model': 1.096318269148469, 'learning_rate/qwen_vl_interface': 2.5125356069200572e-06, 'learning_rate/action_model': 2.5125356069200574e-05, 'epoch': 2.22}) +[2026-06-18 00:20:18,541][starVLA.training.train_starvla][INFO] - Step 4900, Loss: {'train/loss': 0.2474091798067093, 'train/grad_norm_pre_clip': 34.50523376464844, 'eval/loss': 0.2761273682117462, 'eval/latency_8/loss': 0.2761273682117462, 'eval/demon_fire_precision': 0.9615384615384616, 'eval/demon_fire_recall': 0.9863311587449518, 'eval/demon_fire_f1': 0.9737770280631806, 'eval/demon_fire_rate_model': 0.9187534780189204, 'eval/demon_fire_rate_teacher': 0.8956594323873122, 'eval/demon_left_precision': 0.9095477386934674, 'eval/demon_left_recall': 0.7751605995717344, 'eval/demon_left_f1': 0.8369942196531793, 'eval/demon_left_rate_model': 0.110740122426266, 'eval/demon_left_rate_teacher': 0.12993878686700056, 'eval/demon_right_precision': 0.8675078864353313, 'eval/demon_right_recall': 0.8333333333333334, 'eval/demon_right_f1': 0.8500772797527049, 'eval/demon_right_rate_model': 0.17640511964385086, 'eval/demon_right_rate_teacher': 0.18363939899833054, 'eval/demon_move_f1': 0.8435357497029421, 'eval/demon_cc_f1': 0.9306686676706548, 'eval/latency_8/demon_fire_precision': 0.9615384615384616, 'eval/latency_8/demon_fire_recall': 0.9863311587449518, 'eval/latency_8/demon_fire_f1': 0.9737770280631806, 'eval/latency_8/demon_fire_rate_model': 0.9187534780189204, 'eval/latency_8/demon_fire_rate_teacher': 0.8956594323873122, 'eval/latency_8/demon_left_precision': 0.9095477386934674, 'eval/latency_8/demon_left_recall': 0.7751605995717344, 'eval/latency_8/demon_left_f1': 0.8369942196531793, 'eval/latency_8/demon_left_rate_model': 0.110740122426266, 'eval/latency_8/demon_left_rate_teacher': 0.12993878686700056, 'eval/latency_8/demon_right_precision': 0.8675078864353313, 'eval/latency_8/demon_right_recall': 0.8333333333333334, 'eval/latency_8/demon_right_f1': 0.8500772797527049, 'eval/latency_8/demon_right_rate_model': 0.17640511964385086, 'eval/latency_8/demon_right_rate_teacher': 0.18363939899833054, 'eval/latency_8/demon_move_f1': 0.8435357497029421, 'eval/latency_8/demon_cc_f1': 0.9306686676706548, 'timing/data': 0.000529884360730648, 'timing/model': 1.0056751910597086, 'learning_rate/qwen_vl_interface': 2.5107684699543817e-06, 'learning_rate/action_model': 2.5107684699543816e-05, 'epoch': 2.22}) +[2026-06-18 00:20:18,542][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_4900_state +[2026-06-18 00:20:18,547][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 00:20:24,683][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_4900_state/model.safetensors +[2026-06-18 00:20:39,953][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_4900_state/optimizer.bin +[2026-06-18 00:20:39,954][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_4900_state/sampler.bin +[2026-06-18 00:20:39,956][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_4900_state/sampler_1.bin +[2026-06-18 00:20:39,959][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_4900_state/random_states_0.pkl +[2026-06-18 00:20:39,962][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4900 +[2026-06-18 00:20:39,962][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 00:20:39,972][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 00:20:43,707][starVLA.training.train_starvla][INFO] - Step 4901, Loss: {'train/loss': 0.13009896874427795, 'train/grad_norm_pre_clip': 13.052412033081055, 'timing/data': 0.0005311965942382812, 'timing/model': 1.0881229816004634, 'learning_rate/qwen_vl_interface': 2.509001900834349e-06, 'learning_rate/action_model': 2.5090019008343492e-05, 'epoch': 2.22}) +[2026-06-18 00:20:47,492][starVLA.training.train_starvla][INFO] - Step 4902, Loss: {'train/loss': 0.14709129929542542, 'train/grad_norm_pre_clip': 24.32646942138672, 'timing/data': 0.0005282089114189148, 'timing/model': 1.030256506986916, 'learning_rate/qwen_vl_interface': 2.507235899926172e-06, 'learning_rate/action_model': 2.5072358999261718e-05, 'epoch': 2.22}) +[2026-06-18 00:20:51,220][starVLA.training.train_starvla][INFO] - Step 4903, Loss: {'train/loss': 0.09710296988487244, 'train/grad_norm_pre_clip': 16.95561981201172, 'timing/data': 0.0005077272653579712, 'timing/model': 0.9834019914269447, 'learning_rate/qwen_vl_interface': 2.5054704675959445e-06, 'learning_rate/action_model': 2.5054704675959445e-05, 'epoch': 2.22}) +[2026-06-18 00:20:55,026][starVLA.training.train_starvla][INFO] - Step 4904, Loss: {'train/loss': 0.12304027378559113, 'train/grad_norm_pre_clip': 17.594675064086914, 'timing/data': 0.0005062296986579895, 'timing/model': 0.9320695335045457, 'learning_rate/qwen_vl_interface': 2.5037056042096418e-06, 'learning_rate/action_model': 2.5037056042096417e-05, 'epoch': 2.22}) +[2026-06-18 00:20:59,024][starVLA.training.train_starvla][INFO] - Step 4905, Loss: {'train/loss': 0.2584441304206848, 'train/grad_norm_pre_clip': 33.2545280456543, 'timing/data': 0.0008025234565138817, 'timing/model': 0.9365249210968614, 'learning_rate/qwen_vl_interface': 2.5019413101331212e-06, 'learning_rate/action_model': 2.5019413101331214e-05, 'epoch': 2.23}) +[2026-06-18 00:21:02,787][starVLA.training.train_starvla][INFO] - Step 4906, Loss: {'train/loss': 0.021829703822731972, 'train/grad_norm_pre_clip': 3.727451801300049, 'timing/data': 0.0005046017467975616, 'timing/model': 0.9376336801797152, 'learning_rate/qwen_vl_interface': 2.5001775857321254e-06, 'learning_rate/action_model': 2.5001775857321253e-05, 'epoch': 2.23}) +[2026-06-18 00:21:06,835][starVLA.training.train_starvla][INFO] - Step 4907, Loss: {'train/loss': 0.41735538840293884, 'train/grad_norm_pre_clip': 27.931529998779297, 'timing/data': 0.000579209066927433, 'timing/model': 0.948679412715137, 'learning_rate/qwen_vl_interface': 2.4984144313722765e-06, 'learning_rate/action_model': 2.4984144313722766e-05, 'epoch': 2.23}) +[2026-06-18 00:21:10,649][starVLA.training.train_starvla][INFO] - Step 4908, Loss: {'train/loss': 0.3049532473087311, 'train/grad_norm_pre_clip': 23.731290817260742, 'timing/data': 0.0010243309661746025, 'timing/model': 1.1446867799386382, 'learning_rate/qwen_vl_interface': 2.496651847419076e-06, 'learning_rate/action_model': 2.496651847419076e-05, 'epoch': 2.23}) +[2026-06-18 00:21:14,238][starVLA.training.train_starvla][INFO] - Step 4909, Loss: {'train/loss': 0.292140930891037, 'train/grad_norm_pre_clip': 34.94725036621094, 'timing/data': 0.000554598867893219, 'timing/model': 0.9287043316289783, 'learning_rate/qwen_vl_interface': 2.4948898342379076e-06, 'learning_rate/action_model': 2.4948898342379075e-05, 'epoch': 2.23}) +[2026-06-18 00:21:17,977][starVLA.training.train_starvla][INFO] - Step 4910, Loss: {'train/loss': 0.32069048285484314, 'train/grad_norm_pre_clip': 31.927818298339844, 'timing/data': 0.0005608163774013519, 'timing/model': 0.9234695853665471, 'learning_rate/qwen_vl_interface': 2.4931283921940434e-06, 'learning_rate/action_model': 2.4931283921940434e-05, 'epoch': 2.23}) +[2026-06-18 00:21:21,934][starVLA.training.train_starvla][INFO] - Step 4911, Loss: {'train/loss': 0.16170664131641388, 'train/grad_norm_pre_clip': 23.861541748046875, 'timing/data': 0.0014113802462816238, 'timing/model': 0.9191597923636436, 'learning_rate/qwen_vl_interface': 2.4913675216526302e-06, 'learning_rate/action_model': 2.4913675216526304e-05, 'epoch': 2.23}) +[2026-06-18 00:21:25,715][starVLA.training.train_starvla][INFO] - Step 4912, Loss: {'train/loss': 0.19202925264835358, 'train/grad_norm_pre_clip': 17.10027503967285, 'timing/data': 0.000489761121571064, 'timing/model': 0.9195169601589441, 'learning_rate/qwen_vl_interface': 2.4896072229786996e-06, 'learning_rate/action_model': 2.4896072229786995e-05, 'epoch': 2.23}) +[2026-06-18 00:21:29,400][starVLA.training.train_starvla][INFO] - Step 4913, Loss: {'train/loss': 0.3437694013118744, 'train/grad_norm_pre_clip': 29.23013687133789, 'timing/data': 0.0008724704384803772, 'timing/model': 0.9158422462642193, 'learning_rate/qwen_vl_interface': 2.4878474965371587e-06, 'learning_rate/action_model': 2.4878474965371586e-05, 'epoch': 2.23}) +[2026-06-18 00:21:33,130][starVLA.training.train_starvla][INFO] - Step 4914, Loss: {'train/loss': 0.3136897385120392, 'train/grad_norm_pre_clip': 28.772371292114258, 'timing/data': 0.0005031861364841461, 'timing/model': 0.915164552628994, 'learning_rate/qwen_vl_interface': 2.4860883426928053e-06, 'learning_rate/action_model': 2.4860883426928053e-05, 'epoch': 2.23}) +[2026-06-18 00:21:37,062][starVLA.training.train_starvla][INFO] - Step 4915, Loss: {'train/loss': 0.34536343812942505, 'train/grad_norm_pre_clip': 31.05986976623535, 'timing/data': 0.0005477480590343475, 'timing/model': 1.110145197249949, 'learning_rate/qwen_vl_interface': 2.4843297618103127e-06, 'learning_rate/action_model': 2.4843297618103128e-05, 'epoch': 2.23}) +[2026-06-18 00:21:40,847][starVLA.training.train_starvla][INFO] - Step 4916, Loss: {'train/loss': 0.27310413122177124, 'train/grad_norm_pre_clip': 31.048452377319336, 'timing/data': 0.0005319565534591675, 'timing/model': 0.9172636773437262, 'learning_rate/qwen_vl_interface': 2.482571754254237e-06, 'learning_rate/action_model': 2.4825717542542372e-05, 'epoch': 2.23}) +[2026-06-18 00:21:44,588][starVLA.training.train_starvla][INFO] - Step 4917, Loss: {'train/loss': 0.4485040307044983, 'train/grad_norm_pre_clip': 48.32809066772461, 'timing/data': 0.0005000801756978035, 'timing/model': 0.9147785771638155, 'learning_rate/qwen_vl_interface': 2.4808143203890137e-06, 'learning_rate/action_model': 2.480814320389014e-05, 'epoch': 2.23}) +[2026-06-18 00:21:48,302][starVLA.training.train_starvla][INFO] - Step 4918, Loss: {'train/loss': 0.26775017380714417, 'train/grad_norm_pre_clip': 31.142656326293945, 'timing/data': 0.0004993090406060219, 'timing/model': 0.9126600725576282, 'learning_rate/qwen_vl_interface': 2.4790574605789627e-06, 'learning_rate/action_model': 2.4790574605789624e-05, 'epoch': 2.23}) +[2026-06-18 00:21:52,187][starVLA.training.train_starvla][INFO] - Step 4919, Loss: {'train/loss': 0.07784151285886765, 'train/grad_norm_pre_clip': 11.963335990905762, 'timing/data': 0.0005099698901176453, 'timing/model': 1.0619645873084664, 'learning_rate/qwen_vl_interface': 2.4773011751882807e-06, 'learning_rate/action_model': 2.4773011751882804e-05, 'epoch': 2.23}) +[2026-06-18 00:21:55,939][starVLA.training.train_starvla][INFO] - Step 4920, Loss: {'train/loss': 0.13279768824577332, 'train/grad_norm_pre_clip': 17.998647689819336, 'timing/data': 0.000851907767355442, 'timing/model': 0.9418447278439999, 'learning_rate/qwen_vl_interface': 2.4755454645810473e-06, 'learning_rate/action_model': 2.4755454645810473e-05, 'epoch': 2.23}) +[2026-06-18 00:21:59,704][starVLA.training.train_starvla][INFO] - Step 4921, Loss: {'train/loss': 0.02894686721265316, 'train/grad_norm_pre_clip': 5.834237098693848, 'timing/data': 0.0004680883139371872, 'timing/model': 0.9212053334340453, 'learning_rate/qwen_vl_interface': 2.473790329121228e-06, 'learning_rate/action_model': 2.473790329121228e-05, 'epoch': 2.23}) +[2026-06-18 00:22:03,451][starVLA.training.train_starvla][INFO] - Step 4922, Loss: {'train/loss': 0.2638583183288574, 'train/grad_norm_pre_clip': 30.725473403930664, 'timing/data': 0.0004924489185214043, 'timing/model': 1.1079909447580576, 'learning_rate/qwen_vl_interface': 2.4720357691726595e-06, 'learning_rate/action_model': 2.4720357691726594e-05, 'epoch': 2.23}) +[2026-06-18 00:22:07,067][starVLA.training.train_starvla][INFO] - Step 4923, Loss: {'train/loss': 0.12291041016578674, 'train/grad_norm_pre_clip': 19.249792098999023, 'timing/data': 0.0004896298050880432, 'timing/model': 0.9310484323650599, 'learning_rate/qwen_vl_interface': 2.4702817850990645e-06, 'learning_rate/action_model': 2.470281785099064e-05, 'epoch': 2.23}) +[2026-06-18 00:22:10,975][starVLA.training.train_starvla][INFO] - Step 4924, Loss: {'train/loss': 0.3747597932815552, 'train/grad_norm_pre_clip': 27.49423599243164, 'timing/data': 0.0004979921504855156, 'timing/model': 1.045251744799316, 'learning_rate/qwen_vl_interface': 2.4685283772640452e-06, 'learning_rate/action_model': 2.468528377264045e-05, 'epoch': 2.23}) +[2026-06-18 00:22:14,700][starVLA.training.train_starvla][INFO] - Step 4925, Loss: {'train/loss': 0.496271014213562, 'train/grad_norm_pre_clip': 38.72625732421875, 'timing/data': 0.00047926977276802063, 'timing/model': 0.9214024543762207, 'learning_rate/qwen_vl_interface': 2.466775546031088e-06, 'learning_rate/action_model': 2.4667755460310878e-05, 'epoch': 2.23}) +[2026-06-18 00:22:18,554][starVLA.training.train_starvla][INFO] - Step 4926, Loss: {'train/loss': 0.374537855386734, 'train/grad_norm_pre_clip': 35.51707077026367, 'timing/data': 0.0004980200901627541, 'timing/model': 0.9156183321028948, 'learning_rate/qwen_vl_interface': 2.465023291763555e-06, 'learning_rate/action_model': 2.465023291763555e-05, 'epoch': 2.23}) +[2026-06-18 00:22:22,460][starVLA.training.train_starvla][INFO] - Step 4927, Loss: {'train/loss': 0.23500674962997437, 'train/grad_norm_pre_clip': 27.36060905456543, 'timing/data': 0.0008140411227941513, 'timing/model': 1.0325368354097009, 'learning_rate/qwen_vl_interface': 2.4632716148246903e-06, 'learning_rate/action_model': 2.4632716148246903e-05, 'epoch': 2.24}) +[2026-06-18 00:22:26,198][starVLA.training.train_starvla][INFO] - Step 4928, Loss: {'train/loss': 0.12273076921701431, 'train/grad_norm_pre_clip': 21.679794311523438, 'timing/data': 0.0005449457094073296, 'timing/model': 1.0853130873292685, 'learning_rate/qwen_vl_interface': 2.4615205155776185e-06, 'learning_rate/action_model': 2.4615205155776188e-05, 'epoch': 2.24}) +[2026-06-18 00:22:30,049][starVLA.training.train_starvla][INFO] - Step 4929, Loss: {'train/loss': 0.13728606700897217, 'train/grad_norm_pre_clip': 27.266923904418945, 'timing/data': 0.000568099319934845, 'timing/model': 1.0987914381548762, 'learning_rate/qwen_vl_interface': 2.4597699943853444e-06, 'learning_rate/action_model': 2.4597699943853442e-05, 'epoch': 2.24}) +[2026-06-18 00:22:33,663][starVLA.training.train_starvla][INFO] - Step 4930, Loss: {'train/loss': 0.2223268449306488, 'train/grad_norm_pre_clip': 57.459251403808594, 'timing/data': 0.0005000550299882889, 'timing/model': 0.9239732977002859, 'learning_rate/qwen_vl_interface': 2.458020051610753e-06, 'learning_rate/action_model': 2.4580200516107528e-05, 'epoch': 2.24}) +[2026-06-18 00:22:37,672][starVLA.training.train_starvla][INFO] - Step 4931, Loss: {'train/loss': 0.17129643261432648, 'train/grad_norm_pre_clip': 25.52669334411621, 'timing/data': 0.0008604181930422783, 'timing/model': 1.162138681858778, 'learning_rate/qwen_vl_interface': 2.456270687616609e-06, 'learning_rate/action_model': 2.4562706876166087e-05, 'epoch': 2.24}) +[2026-06-18 00:22:41,454][starVLA.training.train_starvla][INFO] - Step 4932, Loss: {'train/loss': 0.16749241948127747, 'train/grad_norm_pre_clip': 25.114093780517578, 'timing/data': 0.0006449185311794281, 'timing/model': 0.9380782246589661, 'learning_rate/qwen_vl_interface': 2.4545219027655577e-06, 'learning_rate/action_model': 2.4545219027655574e-05, 'epoch': 2.24}) +[2026-06-18 00:22:45,230][starVLA.training.train_starvla][INFO] - Step 4933, Loss: {'train/loss': 0.02558816783130169, 'train/grad_norm_pre_clip': 6.478996753692627, 'timing/data': 0.00046709273010492325, 'timing/model': 0.9502363884821534, 'learning_rate/qwen_vl_interface': 2.452773697420124e-06, 'learning_rate/action_model': 2.452773697420124e-05, 'epoch': 2.24}) +[2026-06-18 00:22:48,951][starVLA.training.train_starvla][INFO] - Step 4934, Loss: {'train/loss': 0.17180891335010529, 'train/grad_norm_pre_clip': 27.504589080810547, 'timing/data': 0.0005103955045342445, 'timing/model': 0.9146242383867502, 'learning_rate/qwen_vl_interface': 2.4510260719427127e-06, 'learning_rate/action_model': 2.4510260719427127e-05, 'epoch': 2.24}) +[2026-06-18 00:22:52,830][starVLA.training.train_starvla][INFO] - Step 4935, Loss: {'train/loss': 0.20745833218097687, 'train/grad_norm_pre_clip': 25.15427589416504, 'timing/data': 0.0005308398976922035, 'timing/model': 1.1135210879147053, 'learning_rate/qwen_vl_interface': 2.4492790266956092e-06, 'learning_rate/action_model': 2.449279026695609e-05, 'epoch': 2.24}) +[2026-06-18 00:22:56,893][starVLA.training.train_starvla][INFO] - Step 4936, Loss: {'train/loss': 0.3595998287200928, 'train/grad_norm_pre_clip': 32.53257369995117, 'timing/data': 0.0006495509296655655, 'timing/model': 1.2714251084253192, 'learning_rate/qwen_vl_interface': 2.4475325620409765e-06, 'learning_rate/action_model': 2.4475325620409764e-05, 'epoch': 2.24}) +[2026-06-18 00:23:00,518][starVLA.training.train_starvla][INFO] - Step 4937, Loss: {'train/loss': 0.2668699622154236, 'train/grad_norm_pre_clip': 24.640167236328125, 'timing/data': 0.0005090357735753059, 'timing/model': 0.9450730755925179, 'learning_rate/qwen_vl_interface': 2.445786678340859e-06, 'learning_rate/action_model': 2.4457866783408588e-05, 'epoch': 2.24}) +[2026-06-18 00:23:04,309][starVLA.training.train_starvla][INFO] - Step 4938, Loss: {'train/loss': 0.182464599609375, 'train/grad_norm_pre_clip': 16.955543518066406, 'timing/data': 0.0008402168750762939, 'timing/model': 0.920918533578515, 'learning_rate/qwen_vl_interface': 2.444041375957179e-06, 'learning_rate/action_model': 2.444041375957179e-05, 'epoch': 2.24}) +[2026-06-18 00:23:08,144][starVLA.training.train_starvla][INFO] - Step 4939, Loss: {'train/loss': 0.08101718872785568, 'train/grad_norm_pre_clip': 15.870223999023438, 'timing/data': 0.0004926631227135658, 'timing/model': 1.0363702028989792, 'learning_rate/qwen_vl_interface': 2.4422966552517434e-06, 'learning_rate/action_model': 2.4422966552517433e-05, 'epoch': 2.24}) +[2026-06-18 00:23:11,834][starVLA.training.train_starvla][INFO] - Step 4940, Loss: {'train/loss': 0.5673174262046814, 'train/grad_norm_pre_clip': 35.96207046508789, 'timing/data': 0.00047819502651691437, 'timing/model': 0.9308223947882652, 'learning_rate/qwen_vl_interface': 2.440552516586232e-06, 'learning_rate/action_model': 2.440552516586232e-05, 'epoch': 2.24}) +[2026-06-18 00:23:15,677][starVLA.training.train_starvla][INFO] - Step 4941, Loss: {'train/loss': 0.12858182191848755, 'train/grad_norm_pre_clip': 30.1165771484375, 'timing/data': 0.00047633610665798187, 'timing/model': 0.9191559981554747, 'learning_rate/qwen_vl_interface': 2.4388089603222085e-06, 'learning_rate/action_model': 2.4388089603222087e-05, 'epoch': 2.24}) +[2026-06-18 00:23:19,411][starVLA.training.train_starvla][INFO] - Step 4942, Loss: {'train/loss': 0.3212812542915344, 'train/grad_norm_pre_clip': 13.992947578430176, 'timing/data': 0.0004855962470173836, 'timing/model': 1.0999626824632287, 'learning_rate/qwen_vl_interface': 2.4370659868211095e-06, 'learning_rate/action_model': 2.4370659868211097e-05, 'epoch': 2.24}) +[2026-06-18 00:23:22,972][starVLA.training.train_starvla][INFO] - Step 4943, Loss: {'train/loss': 0.49329882860183716, 'train/grad_norm_pre_clip': 43.982025146484375, 'timing/data': 0.0005550142377614975, 'timing/model': 0.9230045434087515, 'learning_rate/qwen_vl_interface': 2.4353235964442597e-06, 'learning_rate/action_model': 2.4353235964442598e-05, 'epoch': 2.24}) +[2026-06-18 00:23:26,707][starVLA.training.train_starvla][INFO] - Step 4944, Loss: {'train/loss': 0.24270881712436676, 'train/grad_norm_pre_clip': 30.506887435913086, 'timing/data': 0.0005127135664224625, 'timing/model': 0.9140234896913171, 'learning_rate/qwen_vl_interface': 2.433581789552857e-06, 'learning_rate/action_model': 2.433581789552857e-05, 'epoch': 2.24}) +[2026-06-18 00:23:30,385][starVLA.training.train_starvla][INFO] - Step 4945, Loss: {'train/loss': 0.6387937664985657, 'train/grad_norm_pre_clip': 44.63092041015625, 'timing/data': 0.0005370862782001495, 'timing/model': 0.9357994366437197, 'learning_rate/qwen_vl_interface': 2.4318405665079806e-06, 'learning_rate/action_model': 2.4318405665079805e-05, 'epoch': 2.24}) +[2026-06-18 00:23:34,396][starVLA.training.train_starvla][INFO] - Step 4946, Loss: {'train/loss': 0.15375679731369019, 'train/grad_norm_pre_clip': 40.83612060546875, 'timing/data': 0.0005359724164009094, 'timing/model': 1.0581440664827824, 'learning_rate/qwen_vl_interface': 2.4300999276705868e-06, 'learning_rate/action_model': 2.4300999276705868e-05, 'epoch': 2.24}) +[2026-06-18 00:23:38,474][starVLA.training.train_starvla][INFO] - Step 4947, Loss: {'train/loss': 0.27123895287513733, 'train/grad_norm_pre_clip': 30.924198150634766, 'timing/data': 0.0004822826012969017, 'timing/model': 1.0033065536990762, 'learning_rate/qwen_vl_interface': 2.428359873401512e-06, 'learning_rate/action_model': 2.4283598734015124e-05, 'epoch': 2.24}) +[2026-06-18 00:23:42,318][starVLA.training.train_starvla][INFO] - Step 4948, Loss: {'train/loss': 0.3379664421081543, 'train/grad_norm_pre_clip': 48.54373550415039, 'timing/data': 0.0005500307306647301, 'timing/model': 1.0972488056868315, 'learning_rate/qwen_vl_interface': 2.4266204040614722e-06, 'learning_rate/action_model': 2.4266204040614724e-05, 'epoch': 2.24}) +[2026-06-18 00:23:46,089][starVLA.training.train_starvla][INFO] - Step 4949, Loss: {'train/loss': 0.3262031376361847, 'train/grad_norm_pre_clip': 43.70747756958008, 'timing/data': 0.0011671008542180061, 'timing/model': 1.0935083916410804, 'learning_rate/qwen_vl_interface': 2.424881520011059e-06, 'learning_rate/action_model': 2.424881520011059e-05, 'epoch': 2.25}) +[2026-06-18 00:23:49,712][starVLA.training.train_starvla][INFO] - Step 4950, Loss: {'train/loss': 0.13876566290855408, 'train/grad_norm_pre_clip': 13.019110679626465, 'timing/data': 0.0006708307191729546, 'timing/model': 0.9386748392134905, 'learning_rate/qwen_vl_interface': 2.4231432216107494e-06, 'learning_rate/action_model': 2.4231432216107494e-05, 'epoch': 2.25}) +[2026-06-18 00:23:53,836][starVLA.training.train_starvla][INFO] - Step 4951, Loss: {'train/loss': 0.06327994912862778, 'train/grad_norm_pre_clip': 13.846400260925293, 'timing/data': 0.0013884138315916061, 'timing/model': 1.1850308002904058, 'learning_rate/qwen_vl_interface': 2.4214055092208907e-06, 'learning_rate/action_model': 2.4214055092208907e-05, 'epoch': 2.25}) +[2026-06-18 00:23:57,708][starVLA.training.train_starvla][INFO] - Step 4952, Loss: {'train/loss': 0.22783510386943817, 'train/grad_norm_pre_clip': 28.047334671020508, 'timing/data': 0.0011203214526176453, 'timing/model': 0.9647496975958347, 'learning_rate/qwen_vl_interface': 2.4196683832017134e-06, 'learning_rate/action_model': 2.4196683832017135e-05, 'epoch': 2.25}) +[2026-06-18 00:24:01,537][starVLA.training.train_starvla][INFO] - Step 4953, Loss: {'train/loss': 0.08495748788118362, 'train/grad_norm_pre_clip': 12.290482521057129, 'timing/data': 0.001197543926537037, 'timing/model': 0.9372896999120712, 'learning_rate/qwen_vl_interface': 2.4179318439133245e-06, 'learning_rate/action_model': 2.417931843913324e-05, 'epoch': 2.25}) +[2026-06-18 00:24:05,613][starVLA.training.train_starvla][INFO] - Step 4954, Loss: {'train/loss': 0.11817777901887894, 'train/grad_norm_pre_clip': 16.418615341186523, 'timing/data': 0.000659245066344738, 'timing/model': 1.0304778022691607, 'learning_rate/qwen_vl_interface': 2.4161958917157126e-06, 'learning_rate/action_model': 2.4161958917157126e-05, 'epoch': 2.25}) +[2026-06-18 00:24:09,433][starVLA.training.train_starvla][INFO] - Step 4955, Loss: {'train/loss': 0.32300591468811035, 'train/grad_norm_pre_clip': 24.73797607421875, 'timing/data': 0.0005729673430323601, 'timing/model': 1.1036397842690349, 'learning_rate/qwen_vl_interface': 2.4144605269687445e-06, 'learning_rate/action_model': 2.4144605269687446e-05, 'epoch': 2.25}) +[2026-06-18 00:24:13,066][starVLA.training.train_starvla][INFO] - Step 4956, Loss: {'train/loss': 0.43699371814727783, 'train/grad_norm_pre_clip': 32.36508560180664, 'timing/data': 0.0005019363015890121, 'timing/model': 0.9191149557009339, 'learning_rate/qwen_vl_interface': 2.412725750032156e-06, 'learning_rate/action_model': 2.4127257500321563e-05, 'epoch': 2.25}) +[2026-06-18 00:24:16,995][starVLA.training.train_starvla][INFO] - Step 4957, Loss: {'train/loss': 0.1913873553276062, 'train/grad_norm_pre_clip': 22.22528076171875, 'timing/data': 0.0004942147061228752, 'timing/model': 1.059217625297606, 'learning_rate/qwen_vl_interface': 2.4109915612655745e-06, 'learning_rate/action_model': 2.4109915612655745e-05, 'epoch': 2.25}) +[2026-06-18 00:24:20,681][starVLA.training.train_starvla][INFO] - Step 4958, Loss: {'train/loss': 0.1415058970451355, 'train/grad_norm_pre_clip': 13.563239097595215, 'timing/data': 0.0004902798682451248, 'timing/model': 0.924373546615243, 'learning_rate/qwen_vl_interface': 2.409257961028498e-06, 'learning_rate/action_model': 2.4092579610284978e-05, 'epoch': 2.25}) +[2026-06-18 00:24:24,548][starVLA.training.train_starvla][INFO] - Step 4959, Loss: {'train/loss': 0.09347040206193924, 'train/grad_norm_pre_clip': 19.705974578857422, 'timing/data': 0.000698186457157135, 'timing/model': 0.9277700688689947, 'learning_rate/qwen_vl_interface': 2.4075249496803026e-06, 'learning_rate/action_model': 2.4075249496803025e-05, 'epoch': 2.25}) +[2026-06-18 00:24:28,293][starVLA.training.train_starvla][INFO] - Step 4960, Loss: {'train/loss': 0.4036013185977936, 'train/grad_norm_pre_clip': 30.78888702392578, 'timing/data': 0.000505899079144001, 'timing/model': 0.9207515539601445, 'learning_rate/qwen_vl_interface': 2.405792527580243e-06, 'learning_rate/action_model': 2.405792527580243e-05, 'epoch': 2.25}) +[2026-06-18 00:24:32,092][starVLA.training.train_starvla][INFO] - Step 4961, Loss: {'train/loss': 0.25326740741729736, 'train/grad_norm_pre_clip': 27.59805679321289, 'timing/data': 0.000725393183529377, 'timing/model': 1.1370000476017594, 'learning_rate/qwen_vl_interface': 2.404060695087454e-06, 'learning_rate/action_model': 2.4040606950874536e-05, 'epoch': 2.25}) +[2026-06-18 00:24:35,829][starVLA.training.train_starvla][INFO] - Step 4962, Loss: {'train/loss': 0.028392642736434937, 'train/grad_norm_pre_clip': 14.884323120117188, 'timing/data': 0.0005309954285621643, 'timing/model': 1.091856419108808, 'learning_rate/qwen_vl_interface': 2.4023294525609437e-06, 'learning_rate/action_model': 2.4023294525609437e-05, 'epoch': 2.25}) +[2026-06-18 00:24:39,445][starVLA.training.train_starvla][INFO] - Step 4963, Loss: {'train/loss': 0.26799124479293823, 'train/grad_norm_pre_clip': 30.279203414916992, 'timing/data': 0.0009092660620808601, 'timing/model': 0.931511789560318, 'learning_rate/qwen_vl_interface': 2.400598800359604e-06, 'learning_rate/action_model': 2.4005988003596035e-05, 'epoch': 2.25}) +[2026-06-18 00:24:43,219][starVLA.training.train_starvla][INFO] - Step 4964, Loss: {'train/loss': 0.32597944140434265, 'train/grad_norm_pre_clip': 21.473331451416016, 'timing/data': 0.000633598305284977, 'timing/model': 0.9286913620308042, 'learning_rate/qwen_vl_interface': 2.398868738842198e-06, 'learning_rate/action_model': 2.3988687388421978e-05, 'epoch': 2.25}) +[2026-06-18 00:24:47,469][starVLA.training.train_starvla][INFO] - Step 4965, Loss: {'train/loss': 0.02388831600546837, 'train/grad_norm_pre_clip': 4.786489963531494, 'timing/data': 0.0005840957164764404, 'timing/model': 1.1837619533762336, 'learning_rate/qwen_vl_interface': 2.3971392683673705e-06, 'learning_rate/action_model': 2.3971392683673706e-05, 'epoch': 2.25}) +[2026-06-18 00:24:52,057][starVLA.training.train_starvla][INFO] - Step 4966, Loss: {'train/loss': 0.18451736867427826, 'train/grad_norm_pre_clip': 28.439281463623047, 'timing/data': 0.0009418129920959473, 'timing/model': 0.9378036698326468, 'learning_rate/qwen_vl_interface': 2.3954103892936425e-06, 'learning_rate/action_model': 2.3954103892936424e-05, 'epoch': 2.25}) +[2026-06-18 00:24:55,811][starVLA.training.train_starvla][INFO] - Step 4967, Loss: {'train/loss': 0.32139676809310913, 'train/grad_norm_pre_clip': 29.98169708251953, 'timing/data': 0.0008221091702580452, 'timing/model': 0.9522921470925212, 'learning_rate/qwen_vl_interface': 2.393682101979411e-06, 'learning_rate/action_model': 2.3936821019794105e-05, 'epoch': 2.25}) +[2026-06-18 00:24:59,618][starVLA.training.train_starvla][INFO] - Step 4968, Loss: {'train/loss': 0.40114742517471313, 'train/grad_norm_pre_clip': 31.652307510375977, 'timing/data': 0.0008768904954195023, 'timing/model': 1.1133621521294117, 'learning_rate/qwen_vl_interface': 2.3919544067829546e-06, 'learning_rate/action_model': 2.3919544067829545e-05, 'epoch': 2.25}) +[2026-06-18 00:25:03,442][starVLA.training.train_starvla][INFO] - Step 4969, Loss: {'train/loss': 0.16484445333480835, 'train/grad_norm_pre_clip': 27.127042770385742, 'timing/data': 0.0004929434508085251, 'timing/model': 1.166608914732933, 'learning_rate/qwen_vl_interface': 2.3902273040624267e-06, 'learning_rate/action_model': 2.390227304062427e-05, 'epoch': 2.25}) +[2026-06-18 00:25:07,638][starVLA.training.train_starvla][INFO] - Step 4970, Loss: {'train/loss': 0.09165103733539581, 'train/grad_norm_pre_clip': 26.20341682434082, 'timing/data': 0.0005144523456692696, 'timing/model': 1.5072664702311158, 'learning_rate/qwen_vl_interface': 2.388500794175855e-06, 'learning_rate/action_model': 2.3885007941758548e-05, 'epoch': 2.25}) +[2026-06-18 00:25:11,462][starVLA.training.train_starvla][INFO] - Step 4971, Loss: {'train/loss': 0.4110105335712433, 'train/grad_norm_pre_clip': 30.6086483001709, 'timing/data': 0.00047617778182029724, 'timing/model': 0.9302571257576346, 'learning_rate/qwen_vl_interface': 2.386774877481145e-06, 'learning_rate/action_model': 2.3867748774811452e-05, 'epoch': 2.26}) +[2026-06-18 00:25:15,645][starVLA.training.train_starvla][INFO] - Step 4972, Loss: {'train/loss': 0.25209683179855347, 'train/grad_norm_pre_clip': 37.958431243896484, 'timing/data': 0.00048119667917490005, 'timing/model': 0.9389426186680794, 'learning_rate/qwen_vl_interface': 2.385049554336087e-06, 'learning_rate/action_model': 2.385049554336087e-05, 'epoch': 2.26}) +[2026-06-18 00:25:19,428][starVLA.training.train_starvla][INFO] - Step 4973, Loss: {'train/loss': 0.43470120429992676, 'train/grad_norm_pre_clip': 30.65774917602539, 'timing/data': 0.0005127079784870148, 'timing/model': 0.9374947948381305, 'learning_rate/qwen_vl_interface': 2.3833248250983383e-06, 'learning_rate/action_model': 2.3833248250983383e-05, 'epoch': 2.26}) +[2026-06-18 00:25:23,232][starVLA.training.train_starvla][INFO] - Step 4974, Loss: {'train/loss': 0.18625903129577637, 'train/grad_norm_pre_clip': 27.82891273498535, 'timing/data': 0.0010798387229442596, 'timing/model': 0.9294648701325059, 'learning_rate/qwen_vl_interface': 2.3816006901254396e-06, 'learning_rate/action_model': 2.3816006901254396e-05, 'epoch': 2.26}) +[2026-06-18 00:25:27,009][starVLA.training.train_starvla][INFO] - Step 4975, Loss: {'train/loss': 0.33971288800239563, 'train/grad_norm_pre_clip': 25.198762893676758, 'timing/data': 0.0008005369454622269, 'timing/model': 1.1208972502499819, 'learning_rate/qwen_vl_interface': 2.3798771497748012e-06, 'learning_rate/action_model': 2.379877149774801e-05, 'epoch': 2.26}) +[2026-06-18 00:25:30,987][starVLA.training.train_starvla][INFO] - Step 4976, Loss: {'train/loss': 0.09587827324867249, 'train/grad_norm_pre_clip': 22.913999557495117, 'timing/data': 0.0004758276045322418, 'timing/model': 1.1649455009028316, 'learning_rate/qwen_vl_interface': 2.3781542044037195e-06, 'learning_rate/action_model': 2.3781542044037193e-05, 'epoch': 2.26}) +[2026-06-18 00:25:34,693][starVLA.training.train_starvla][INFO] - Step 4977, Loss: {'train/loss': 0.130327969789505, 'train/grad_norm_pre_clip': 28.78824234008789, 'timing/data': 0.00047460664063692093, 'timing/model': 0.9540313808247447, 'learning_rate/qwen_vl_interface': 2.376431854369362e-06, 'learning_rate/action_model': 2.3764318543693618e-05, 'epoch': 2.26}) +[2026-06-18 00:25:38,532][starVLA.training.train_starvla][INFO] - Step 4978, Loss: {'train/loss': 0.21801374852657318, 'train/grad_norm_pre_clip': 21.180091857910156, 'timing/data': 0.0012993160635232925, 'timing/model': 0.9354996914044023, 'learning_rate/qwen_vl_interface': 2.374710100028773e-06, 'learning_rate/action_model': 2.3747101000287728e-05, 'epoch': 2.26}) +[2026-06-18 00:25:42,644][starVLA.training.train_starvla][INFO] - Step 4979, Loss: {'train/loss': 0.1730629950761795, 'train/grad_norm_pre_clip': 17.672719955444336, 'timing/data': 0.0004837792366743088, 'timing/model': 0.9363874485716224, 'learning_rate/qwen_vl_interface': 2.3729889417388736e-06, 'learning_rate/action_model': 2.3729889417388737e-05, 'epoch': 2.26}) +[2026-06-18 00:25:46,431][starVLA.training.train_starvla][INFO] - Step 4980, Loss: {'train/loss': 0.2094118744134903, 'train/grad_norm_pre_clip': 34.84784698486328, 'timing/data': 0.0005038250237703323, 'timing/model': 0.9485318819060922, 'learning_rate/qwen_vl_interface': 2.3712683798564628e-06, 'learning_rate/action_model': 2.3712683798564625e-05, 'epoch': 2.26}) +[2026-06-18 00:25:50,201][starVLA.training.train_starvla][INFO] - Step 4981, Loss: {'train/loss': 0.3234088122844696, 'train/grad_norm_pre_clip': 28.61559295654297, 'timing/data': 0.0007599443197250366, 'timing/model': 0.939513967372477, 'learning_rate/qwen_vl_interface': 2.3695484147382146e-06, 'learning_rate/action_model': 2.3695484147382147e-05, 'epoch': 2.26}) +[2026-06-18 00:25:54,013][starVLA.training.train_starvla][INFO] - Step 4982, Loss: {'train/loss': 0.09832371026277542, 'train/grad_norm_pre_clip': 14.370237350463867, 'timing/data': 0.0013530626893043518, 'timing/model': 1.1154824690893292, 'learning_rate/qwen_vl_interface': 2.367829046740678e-06, 'learning_rate/action_model': 2.3678290467406776e-05, 'epoch': 2.26}) +[2026-06-18 00:25:57,595][starVLA.training.train_starvla][INFO] - Step 4983, Loss: {'train/loss': 0.23765933513641357, 'train/grad_norm_pre_clip': 32.03675079345703, 'timing/data': 0.0009104786440730095, 'timing/model': 0.9369102828204632, 'learning_rate/qwen_vl_interface': 2.3661102762202846e-06, 'learning_rate/action_model': 2.3661102762202843e-05, 'epoch': 2.26}) +[2026-06-18 00:26:01,357][starVLA.training.train_starvla][INFO] - Step 4984, Loss: {'train/loss': 0.10104714334011078, 'train/grad_norm_pre_clip': 22.621877670288086, 'timing/data': 0.0005017202347517014, 'timing/model': 0.9282030891627073, 'learning_rate/qwen_vl_interface': 2.364392103533332e-06, 'learning_rate/action_model': 2.364392103533332e-05, 'epoch': 2.26}) +[2026-06-18 00:26:05,118][starVLA.training.train_starvla][INFO] - Step 4985, Loss: {'train/loss': 0.11121257394552231, 'train/grad_norm_pre_clip': 16.509899139404297, 'timing/data': 0.0004974314942955971, 'timing/model': 0.9261151952669024, 'learning_rate/qwen_vl_interface': 2.362674529036002e-06, 'learning_rate/action_model': 2.362674529036002e-05, 'epoch': 2.26}) +[2026-06-18 00:26:08,917][starVLA.training.train_starvla][INFO] - Step 4986, Loss: {'train/loss': 0.06080833449959755, 'train/grad_norm_pre_clip': 10.791227340698242, 'timing/data': 0.0005071209743618965, 'timing/model': 0.9313176646828651, 'learning_rate/qwen_vl_interface': 2.360957553084347e-06, 'learning_rate/action_model': 2.360957553084347e-05, 'epoch': 2.26}) +[2026-06-18 00:26:13,000][starVLA.training.train_starvla][INFO] - Step 4987, Loss: {'train/loss': 0.1529417634010315, 'train/grad_norm_pre_clip': 20.401718139648438, 'timing/data': 0.0005040308460593224, 'timing/model': 0.9503071373328567, 'learning_rate/qwen_vl_interface': 2.359241176034302e-06, 'learning_rate/action_model': 2.359241176034302e-05, 'epoch': 2.26}) +[2026-06-18 00:26:16,783][starVLA.training.train_starvla][INFO] - Step 4988, Loss: {'train/loss': 0.10650808364152908, 'train/grad_norm_pre_clip': 20.4915828704834, 'timing/data': 0.000537114217877388, 'timing/model': 1.1266586063429713, 'learning_rate/qwen_vl_interface': 2.3575253982416732e-06, 'learning_rate/action_model': 2.357525398241673e-05, 'epoch': 2.26}) +[2026-06-18 00:26:20,829][starVLA.training.train_starvla][INFO] - Step 4989, Loss: {'train/loss': 0.11389109492301941, 'train/grad_norm_pre_clip': 22.174467086791992, 'timing/data': 0.0005092760547995567, 'timing/model': 1.1252294536679983, 'learning_rate/qwen_vl_interface': 2.3558102200621385e-06, 'learning_rate/action_model': 2.3558102200621386e-05, 'epoch': 2.26}) +[2026-06-18 00:26:24,414][starVLA.training.train_starvla][INFO] - Step 4990, Loss: {'train/loss': 0.13204993307590485, 'train/grad_norm_pre_clip': 29.101608276367188, 'timing/data': 0.00048153288662433624, 'timing/model': 0.9297963930293918, 'learning_rate/qwen_vl_interface': 2.3540956418512607e-06, 'learning_rate/action_model': 2.354095641851261e-05, 'epoch': 2.26}) +[2026-06-18 00:26:28,505][starVLA.training.train_starvla][INFO] - Step 4991, Loss: {'train/loss': 0.1735910326242447, 'train/grad_norm_pre_clip': 14.48189640045166, 'timing/data': 0.0015356158837676048, 'timing/model': 0.9308082824572921, 'learning_rate/qwen_vl_interface': 2.352381663964473e-06, 'learning_rate/action_model': 2.352381663964473e-05, 'epoch': 2.26}) +[2026-06-18 00:26:32,349][starVLA.training.train_starvla][INFO] - Step 4992, Loss: {'train/loss': 0.2650177776813507, 'train/grad_norm_pre_clip': 56.921905517578125, 'timing/data': 0.0004915241152048111, 'timing/model': 1.0394149795174599, 'learning_rate/qwen_vl_interface': 2.3506682867570846e-06, 'learning_rate/action_model': 2.3506682867570844e-05, 'epoch': 2.26}) +[2026-06-18 00:26:36,101][starVLA.training.train_starvla][INFO] - Step 4993, Loss: {'train/loss': 0.3826826214790344, 'train/grad_norm_pre_clip': 41.426483154296875, 'timing/data': 0.0005169548094272614, 'timing/model': 0.9321950525045395, 'learning_rate/qwen_vl_interface': 2.3489555105842796e-06, 'learning_rate/action_model': 2.3489555105842794e-05, 'epoch': 2.27}) +[2026-06-18 00:26:40,274][starVLA.training.train_starvla][INFO] - Step 4994, Loss: {'train/loss': 0.04609914496541023, 'train/grad_norm_pre_clip': 9.607142448425293, 'timing/data': 0.0007021324709057808, 'timing/model': 0.9861538764089346, 'learning_rate/qwen_vl_interface': 2.347243335801119e-06, 'learning_rate/action_model': 2.347243335801119e-05, 'epoch': 2.27}) +[2026-06-18 00:26:44,073][starVLA.training.train_starvla][INFO] - Step 4995, Loss: {'train/loss': 0.08636404573917389, 'train/grad_norm_pre_clip': 16.305265426635742, 'timing/data': 0.0009621903300285339, 'timing/model': 1.138687158934772, 'learning_rate/qwen_vl_interface': 2.3455317627625385e-06, 'learning_rate/action_model': 2.3455317627625385e-05, 'epoch': 2.27}) +[2026-06-18 00:26:48,005][starVLA.training.train_starvla][INFO] - Step 4996, Loss: {'train/loss': 0.31064605712890625, 'train/grad_norm_pre_clip': 27.15074920654297, 'timing/data': 0.000607670284807682, 'timing/model': 1.251576004549861, 'learning_rate/qwen_vl_interface': 2.3438207918233495e-06, 'learning_rate/action_model': 2.3438207918233494e-05, 'epoch': 2.27}) +[2026-06-18 00:26:51,627][starVLA.training.train_starvla][INFO] - Step 4997, Loss: {'train/loss': 0.39419326186180115, 'train/grad_norm_pre_clip': 43.121131896972656, 'timing/data': 0.0013105794787406921, 'timing/model': 0.929629173129797, 'learning_rate/qwen_vl_interface': 2.3421104233382366e-06, 'learning_rate/action_model': 2.3421104233382364e-05, 'epoch': 2.27}) +[2026-06-18 00:26:55,785][starVLA.training.train_starvla][INFO] - Step 4998, Loss: {'train/loss': 0.21788504719734192, 'train/grad_norm_pre_clip': 48.1601676940918, 'timing/data': 0.00051919836550951, 'timing/model': 0.9234365290030837, 'learning_rate/qwen_vl_interface': 2.340400657661762e-06, 'learning_rate/action_model': 2.340400657661762e-05, 'epoch': 2.27}) +[2026-06-18 00:26:59,550][starVLA.training.train_starvla][INFO] - Step 4999, Loss: {'train/loss': 0.20731081068515778, 'train/grad_norm_pre_clip': 24.053983688354492, 'timing/data': 0.000513601116836071, 'timing/model': 0.9190488159656525, 'learning_rate/qwen_vl_interface': 2.338691495148363e-06, 'learning_rate/action_model': 2.3386914951483626e-05, 'epoch': 2.27}) +[2026-06-18 00:28:54,612][starVLA.training.train_starvla][INFO] - Step 5000, Loss: {'train/loss': 0.1431736946105957, 'train/grad_norm_pre_clip': 23.86264419555664, 'eval/loss': 0.27498874068260193, 'eval/latency_8/loss': 0.27498874068260193, 'eval/demon_fire_precision': 0.9709517923362175, 'eval/demon_fire_recall': 0.9760795278036657, 'eval/demon_fire_f1': 0.9735089078233927, 'eval/demon_fire_rate_model': 0.9003895381190874, 'eval/demon_fire_rate_teacher': 0.8956594323873122, 'eval/demon_left_precision': 0.8347280334728033, 'eval/demon_left_recall': 0.854389721627409, 'eval/demon_left_f1': 0.8444444444444446, 'eval/demon_left_rate_model': 0.13299944351697274, 'eval/demon_left_rate_teacher': 0.12993878686700056, 'eval/demon_right_precision': 0.9065743944636678, 'eval/demon_right_recall': 0.793939393939394, 'eval/demon_right_f1': 0.8465266558966075, 'eval/demon_right_rate_model': 0.16082359488035614, 'eval/demon_right_rate_teacher': 0.18363939899833054, 'eval/demon_move_f1': 0.8454855501705261, 'eval/demon_cc_f1': 0.9312078799331567, 'eval/latency_8/demon_fire_precision': 0.9709517923362175, 'eval/latency_8/demon_fire_recall': 0.9760795278036657, 'eval/latency_8/demon_fire_f1': 0.9735089078233927, 'eval/latency_8/demon_fire_rate_model': 0.9003895381190874, 'eval/latency_8/demon_fire_rate_teacher': 0.8956594323873122, 'eval/latency_8/demon_left_precision': 0.8347280334728033, 'eval/latency_8/demon_left_recall': 0.854389721627409, 'eval/latency_8/demon_left_f1': 0.8444444444444446, 'eval/latency_8/demon_left_rate_model': 0.13299944351697274, 'eval/latency_8/demon_left_rate_teacher': 0.12993878686700056, 'eval/latency_8/demon_right_precision': 0.9065743944636678, 'eval/latency_8/demon_right_recall': 0.793939393939394, 'eval/latency_8/demon_right_f1': 0.8465266558966075, 'eval/latency_8/demon_right_rate_model': 0.16082359488035614, 'eval/latency_8/demon_right_rate_teacher': 0.18363939899833054, 'eval/latency_8/demon_move_f1': 0.8454855501705261, 'eval/latency_8/demon_cc_f1': 0.9312078799331567, 'timing/data': 0.0005473913624882698, 'timing/model': 0.9686016328632832, 'learning_rate/qwen_vl_interface': 2.336982936152347e-06, 'learning_rate/action_model': 2.3369829361523468e-05, 'epoch': 2.27}) +[2026-06-18 00:28:54,615][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_5000_state +[2026-06-18 00:28:54,620][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 00:29:01,070][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_5000_state/model.safetensors +[2026-06-18 00:29:16,304][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_5000_state/optimizer.bin +[2026-06-18 00:29:16,305][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_5000_state/sampler.bin +[2026-06-18 00:29:16,305][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_5000_state/sampler_1.bin +[2026-06-18 00:29:16,371][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_5000_state/random_states_0.pkl +[2026-06-18 00:29:16,375][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 5000 +[2026-06-18 00:29:16,375][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 00:29:16,387][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 00:29:20,281][starVLA.training.train_starvla][INFO] - Step 5001, Loss: {'train/loss': 0.3274919390678406, 'train/grad_norm_pre_clip': 41.58316421508789, 'timing/data': 0.0008310237899422646, 'timing/model': 0.929080163128674, 'learning_rate/qwen_vl_interface': 2.3352749810279047e-06, 'learning_rate/action_model': 2.3352749810279047e-05, 'epoch': 2.27}) +[2026-06-18 00:29:24,126][starVLA.training.train_starvla][INFO] - Step 5002, Loss: {'train/loss': 0.07519393414258957, 'train/grad_norm_pre_clip': 23.576204299926758, 'timing/data': 0.0007453653961420059, 'timing/model': 0.9289141399785876, 'learning_rate/qwen_vl_interface': 2.3335676301290972e-06, 'learning_rate/action_model': 2.3335676301290973e-05, 'epoch': 2.27}) +[2026-06-18 00:29:27,906][starVLA.training.train_starvla][INFO] - Step 5003, Loss: {'train/loss': 0.05486537888646126, 'train/grad_norm_pre_clip': 11.190903663635254, 'timing/data': 0.0007544020190834999, 'timing/model': 0.9436195595189929, 'learning_rate/qwen_vl_interface': 2.331860883809857e-06, 'learning_rate/action_model': 2.3318608838098565e-05, 'epoch': 2.27}) +[2026-06-18 00:29:31,731][starVLA.training.train_starvla][INFO] - Step 5004, Loss: {'train/loss': 0.2807435095310211, 'train/grad_norm_pre_clip': 32.977787017822266, 'timing/data': 0.0005213329568505287, 'timing/model': 1.1178518664091825, 'learning_rate/qwen_vl_interface': 2.330154742423993e-06, 'learning_rate/action_model': 2.330154742423993e-05, 'epoch': 2.27}) +[2026-06-18 00:29:35,548][starVLA.training.train_starvla][INFO] - Step 5005, Loss: {'train/loss': 0.3924882709980011, 'train/grad_norm_pre_clip': 44.50672149658203, 'timing/data': 0.0005000568926334381, 'timing/model': 1.1234209416434169, 'learning_rate/qwen_vl_interface': 2.328449206325195e-06, 'learning_rate/action_model': 2.3284492063251947e-05, 'epoch': 2.27}) +[2026-06-18 00:29:41,712][starVLA.training.train_starvla][INFO] - Step 5006, Loss: {'train/loss': 0.3399502635002136, 'train/grad_norm_pre_clip': 35.29599380493164, 'timing/data': 0.0008692629635334015, 'timing/model': 2.27774066850543, 'learning_rate/qwen_vl_interface': 2.32674427586702e-06, 'learning_rate/action_model': 2.3267442758670198e-05, 'epoch': 2.27}) +[2026-06-18 00:29:46,752][starVLA.training.train_starvla][INFO] - Step 5007, Loss: {'train/loss': 0.2465188354253769, 'train/grad_norm_pre_clip': 27.059520721435547, 'timing/data': 0.0009670164436101913, 'timing/model': 0.9324542535468936, 'learning_rate/qwen_vl_interface': 2.3250399514029018e-06, 'learning_rate/action_model': 2.3250399514029016e-05, 'epoch': 2.27}) +[2026-06-18 00:29:50,619][starVLA.training.train_starvla][INFO] - Step 5008, Loss: {'train/loss': 0.09029653668403625, 'train/grad_norm_pre_clip': 18.413482666015625, 'timing/data': 0.0004664352163672447, 'timing/model': 0.9445467069745064, 'learning_rate/qwen_vl_interface': 2.32333623328615e-06, 'learning_rate/action_model': 2.3233362332861498e-05, 'epoch': 2.27}) +[2026-06-18 00:29:55,121][starVLA.training.train_starvla][INFO] - Step 5009, Loss: {'train/loss': 0.11768391728401184, 'train/grad_norm_pre_clip': 35.32958984375, 'timing/data': 0.0005096485838294029, 'timing/model': 1.6485232328996062, 'learning_rate/qwen_vl_interface': 2.3216331218699457e-06, 'learning_rate/action_model': 2.3216331218699456e-05, 'epoch': 2.27}) +[2026-06-18 00:29:58,905][starVLA.training.train_starvla][INFO] - Step 5010, Loss: {'train/loss': 0.5161091685295105, 'train/grad_norm_pre_clip': 42.656585693359375, 'timing/data': 0.00044059567153453827, 'timing/model': 0.9454584317281842, 'learning_rate/qwen_vl_interface': 2.3199306175073465e-06, 'learning_rate/action_model': 2.3199306175073467e-05, 'epoch': 2.27}) +[2026-06-18 00:30:03,226][starVLA.training.train_starvla][INFO] - Step 5011, Loss: {'train/loss': 0.10407435894012451, 'train/grad_norm_pre_clip': 14.245816230773926, 'timing/data': 0.000718437135219574, 'timing/model': 1.1564499055966735, 'learning_rate/qwen_vl_interface': 2.318228720551283e-06, 'learning_rate/action_model': 2.318228720551283e-05, 'epoch': 2.27}) +[2026-06-18 00:30:06,911][starVLA.training.train_starvla][INFO] - Step 5012, Loss: {'train/loss': 0.12031083554029465, 'train/grad_norm_pre_clip': 30.367551803588867, 'timing/data': 0.0004925969988107681, 'timing/model': 0.9392556548118591, 'learning_rate/qwen_vl_interface': 2.3165274313545606e-06, 'learning_rate/action_model': 2.3165274313545606e-05, 'epoch': 2.27}) +[2026-06-18 00:30:10,764][starVLA.training.train_starvla][INFO] - Step 5013, Loss: {'train/loss': 0.10524693131446838, 'train/grad_norm_pre_clip': 19.235034942626953, 'timing/data': 0.0006940178573131561, 'timing/model': 0.9356958884745836, 'learning_rate/qwen_vl_interface': 2.3148267502698584e-06, 'learning_rate/action_model': 2.3148267502698584e-05, 'epoch': 2.27}) +[2026-06-18 00:30:14,999][starVLA.training.train_starvla][INFO] - Step 5014, Loss: {'train/loss': 0.33968591690063477, 'train/grad_norm_pre_clip': 22.321271896362305, 'timing/data': 0.0005118744447827339, 'timing/model': 1.0174937564879656, 'learning_rate/qwen_vl_interface': 2.313126677649729e-06, 'learning_rate/action_model': 2.313126677649729e-05, 'epoch': 2.27}) +[2026-06-18 00:30:18,837][starVLA.training.train_starvla][INFO] - Step 5015, Loss: {'train/loss': 0.1582956165075302, 'train/grad_norm_pre_clip': 20.515682220458984, 'timing/data': 0.0009090239182114601, 'timing/model': 0.9494364066049457, 'learning_rate/qwen_vl_interface': 2.3114272138465975e-06, 'learning_rate/action_model': 2.3114272138465977e-05, 'epoch': 2.28}) +[2026-06-18 00:30:22,643][starVLA.training.train_starvla][INFO] - Step 5016, Loss: {'train/loss': 0.30123576521873474, 'train/grad_norm_pre_clip': 28.511672973632812, 'timing/data': 0.0006125578656792641, 'timing/model': 0.9433829793706536, 'learning_rate/qwen_vl_interface': 2.3097283592127706e-06, 'learning_rate/action_model': 2.3097283592127706e-05, 'epoch': 2.28}) +[2026-06-18 00:30:26,551][starVLA.training.train_starvla][INFO] - Step 5017, Loss: {'train/loss': 0.323818176984787, 'train/grad_norm_pre_clip': 17.413619995117188, 'timing/data': 0.0007678978145122528, 'timing/model': 1.0218978356570005, 'learning_rate/qwen_vl_interface': 2.3080301141004165e-06, 'learning_rate/action_model': 2.3080301141004165e-05, 'epoch': 2.28}) +[2026-06-18 00:30:30,512][starVLA.training.train_starvla][INFO] - Step 5018, Loss: {'train/loss': 0.2474716752767563, 'train/grad_norm_pre_clip': 40.90311050415039, 'timing/data': 0.0007544877007603645, 'timing/model': 1.1969755413010716, 'learning_rate/qwen_vl_interface': 2.3063324788615843e-06, 'learning_rate/action_model': 2.306332478861584e-05, 'epoch': 2.28}) +[2026-06-18 00:30:34,127][starVLA.training.train_starvla][INFO] - Step 5019, Loss: {'train/loss': 0.12949497997760773, 'train/grad_norm_pre_clip': 14.495686531066895, 'timing/data': 0.0007150545716285706, 'timing/model': 0.9400412235409021, 'learning_rate/qwen_vl_interface': 2.304635453848198e-06, 'learning_rate/action_model': 2.3046354538481983e-05, 'epoch': 2.28}) +[2026-06-18 00:30:38,030][starVLA.training.train_starvla][INFO] - Step 5020, Loss: {'train/loss': 0.1917516589164734, 'train/grad_norm_pre_clip': 23.966642379760742, 'timing/data': 0.000927322544157505, 'timing/model': 0.9468064643442631, 'learning_rate/qwen_vl_interface': 2.3029390394120517e-06, 'learning_rate/action_model': 2.3029390394120513e-05, 'epoch': 2.28}) +[2026-06-18 00:30:42,077][starVLA.training.train_starvla][INFO] - Step 5021, Loss: {'train/loss': 0.06842847913503647, 'train/grad_norm_pre_clip': 14.415229797363281, 'timing/data': 0.0013733245432376862, 'timing/model': 1.0101300738751888, 'learning_rate/qwen_vl_interface': 2.301243235904813e-06, 'learning_rate/action_model': 2.3012432359048133e-05, 'epoch': 2.28}) +[2026-06-18 00:30:45,936][starVLA.training.train_starvla][INFO] - Step 5022, Loss: {'train/loss': 0.24347612261772156, 'train/grad_norm_pre_clip': 34.156898498535156, 'timing/data': 0.0005505159497261047, 'timing/model': 0.9418014427646995, 'learning_rate/qwen_vl_interface': 2.299548043678026e-06, 'learning_rate/action_model': 2.299548043678026e-05, 'epoch': 2.28}) +[2026-06-18 00:30:49,721][starVLA.training.train_starvla][INFO] - Step 5023, Loss: {'train/loss': 0.14030857384204865, 'train/grad_norm_pre_clip': 17.609575271606445, 'timing/data': 0.0008482402190566063, 'timing/model': 0.9319264097139239, 'learning_rate/qwen_vl_interface': 2.297853463083103e-06, 'learning_rate/action_model': 2.297853463083103e-05, 'epoch': 2.28}) +[2026-06-18 00:30:53,535][starVLA.training.train_starvla][INFO] - Step 5024, Loss: {'train/loss': 0.09266103804111481, 'train/grad_norm_pre_clip': 17.177324295043945, 'timing/data': 0.0005559716373682022, 'timing/model': 1.1575693264603615, 'learning_rate/qwen_vl_interface': 2.296159494471335e-06, 'learning_rate/action_model': 2.2961594944713347e-05, 'epoch': 2.28}) +[2026-06-18 00:30:57,099][starVLA.training.train_starvla][INFO] - Step 5025, Loss: {'train/loss': 0.2075466513633728, 'train/grad_norm_pre_clip': 27.350500106811523, 'timing/data': 0.0004947120323777199, 'timing/model': 0.9219780564308167, 'learning_rate/qwen_vl_interface': 2.294466138193881e-06, 'learning_rate/action_model': 2.2944661381938812e-05, 'epoch': 2.28}) +[2026-06-18 00:31:01,025][starVLA.training.train_starvla][INFO] - Step 5026, Loss: {'train/loss': 0.02811429649591446, 'train/grad_norm_pre_clip': 7.392724514007568, 'timing/data': 0.0005457866936922073, 'timing/model': 0.9293293673545122, 'learning_rate/qwen_vl_interface': 2.2927733946017777e-06, 'learning_rate/action_model': 2.2927733946017778e-05, 'epoch': 2.28}) +[2026-06-18 00:31:04,871][starVLA.training.train_starvla][INFO] - Step 5027, Loss: {'train/loss': 0.2249683141708374, 'train/grad_norm_pre_clip': 40.376441955566406, 'timing/data': 0.0009634159505367279, 'timing/model': 0.9264059578999877, 'learning_rate/qwen_vl_interface': 2.2910812640459314e-06, 'learning_rate/action_model': 2.2910812640459315e-05, 'epoch': 2.28}) +[2026-06-18 00:31:08,790][starVLA.training.train_starvla][INFO] - Step 5028, Loss: {'train/loss': 0.13651292026042938, 'train/grad_norm_pre_clip': 28.518035888671875, 'timing/data': 0.0014385655522346497, 'timing/model': 0.9320001266896725, 'learning_rate/qwen_vl_interface': 2.289389746877123e-06, 'learning_rate/action_model': 2.2893897468771228e-05, 'epoch': 2.28}) +[2026-06-18 00:31:13,175][starVLA.training.train_starvla][INFO] - Step 5029, Loss: {'train/loss': 0.04709584265947342, 'train/grad_norm_pre_clip': 9.757759094238281, 'timing/data': 0.0012854412198066711, 'timing/model': 0.9494878696277738, 'learning_rate/qwen_vl_interface': 2.2876988434460046e-06, 'learning_rate/action_model': 2.2876988434460046e-05, 'epoch': 2.28}) +[2026-06-18 00:31:16,956][starVLA.training.train_starvla][INFO] - Step 5030, Loss: {'train/loss': 0.12672974169254303, 'train/grad_norm_pre_clip': 13.074124336242676, 'timing/data': 0.0011112205684185028, 'timing/model': 0.9319692403078079, 'learning_rate/qwen_vl_interface': 2.2860085541031048e-06, 'learning_rate/action_model': 2.2860085541031047e-05, 'epoch': 2.28}) +[2026-06-18 00:31:20,764][starVLA.training.train_starvla][INFO] - Step 5031, Loss: {'train/loss': 0.14626671373844147, 'train/grad_norm_pre_clip': 25.65296173095703, 'timing/data': 0.0004918575286865234, 'timing/model': 1.1364338574931026, 'learning_rate/qwen_vl_interface': 2.284318879198823e-06, 'learning_rate/action_model': 2.284318879198823e-05, 'epoch': 2.28}) +[2026-06-18 00:31:24,360][starVLA.training.train_starvla][INFO] - Step 5032, Loss: {'train/loss': 0.2018151730298996, 'train/grad_norm_pre_clip': 26.10054588317871, 'timing/data': 0.0005108918994665146, 'timing/model': 0.9216071562841535, 'learning_rate/qwen_vl_interface': 2.2826298190834266e-06, 'learning_rate/action_model': 2.2826298190834266e-05, 'epoch': 2.28}) +[2026-06-18 00:31:28,560][starVLA.training.train_starvla][INFO] - Step 5033, Loss: {'train/loss': 0.6069827079772949, 'train/grad_norm_pre_clip': 53.29685592651367, 'timing/data': 0.0005332138389348984, 'timing/model': 0.9432925628498197, 'learning_rate/qwen_vl_interface': 2.2809413741070603e-06, 'learning_rate/action_model': 2.28094137410706e-05, 'epoch': 2.28}) +[2026-06-18 00:31:32,318][starVLA.training.train_starvla][INFO] - Step 5034, Loss: {'train/loss': 0.026048870757222176, 'train/grad_norm_pre_clip': 6.800732612609863, 'timing/data': 0.0004989001899957657, 'timing/model': 0.9198092864826322, 'learning_rate/qwen_vl_interface': 2.2792535446197435e-06, 'learning_rate/action_model': 2.2792535446197436e-05, 'epoch': 2.28}) +[2026-06-18 00:31:36,044][starVLA.training.train_starvla][INFO] - Step 5035, Loss: {'train/loss': 0.12753571569919586, 'train/grad_norm_pre_clip': 22.906949996948242, 'timing/data': 0.0004804292693734169, 'timing/model': 0.9203295726329088, 'learning_rate/qwen_vl_interface': 2.2775663309713643e-06, 'learning_rate/action_model': 2.2775663309713645e-05, 'epoch': 2.28}) +[2026-06-18 00:31:39,810][starVLA.training.train_starvla][INFO] - Step 5036, Loss: {'train/loss': 0.16597767174243927, 'train/grad_norm_pre_clip': 17.34261131286621, 'timing/data': 0.0005253693088889122, 'timing/model': 0.9255279367789626, 'learning_rate/qwen_vl_interface': 2.2758797335116843e-06, 'learning_rate/action_model': 2.2758797335116844e-05, 'epoch': 2.28}) +[2026-06-18 00:31:43,564][starVLA.training.train_starvla][INFO] - Step 5037, Loss: {'train/loss': 0.468107134103775, 'train/grad_norm_pre_clip': 43.762245178222656, 'timing/data': 0.0005269581452012062, 'timing/model': 1.081105719320476, 'learning_rate/qwen_vl_interface': 2.2741937525903324e-06, 'learning_rate/action_model': 2.2741937525903324e-05, 'epoch': 2.29}) +[2026-06-18 00:31:47,236][starVLA.training.train_starvla][INFO] - Step 5038, Loss: {'train/loss': 0.15224170684814453, 'train/grad_norm_pre_clip': 24.67986488342285, 'timing/data': 0.001173701137304306, 'timing/model': 1.0379541143774986, 'learning_rate/qwen_vl_interface': 2.2725083885568198e-06, 'learning_rate/action_model': 2.2725083885568197e-05, 'epoch': 2.29}) +[2026-06-18 00:31:50,801][starVLA.training.train_starvla][INFO] - Step 5039, Loss: {'train/loss': 0.2721318304538727, 'train/grad_norm_pre_clip': 33.97735595703125, 'timing/data': 0.00050310418009758, 'timing/model': 0.9270546501502395, 'learning_rate/qwen_vl_interface': 2.2708236417605217e-06, 'learning_rate/action_model': 2.2708236417605216e-05, 'epoch': 2.29}) +[2026-06-18 00:31:54,521][starVLA.training.train_starvla][INFO] - Step 5040, Loss: {'train/loss': 0.3855077028274536, 'train/grad_norm_pre_clip': 48.08179473876953, 'timing/data': 0.0005082916468381882, 'timing/model': 0.9325159667059779, 'learning_rate/qwen_vl_interface': 2.2691395125506885e-06, 'learning_rate/action_model': 2.2691395125506885e-05, 'epoch': 2.29}) +[2026-06-18 00:31:58,246][starVLA.training.train_starvla][INFO] - Step 5041, Loss: {'train/loss': 0.14223934710025787, 'train/grad_norm_pre_clip': 27.607622146606445, 'timing/data': 0.00044170767068862915, 'timing/model': 0.9122121939435601, 'learning_rate/qwen_vl_interface': 2.2674560012764426e-06, 'learning_rate/action_model': 2.2674560012764423e-05, 'epoch': 2.29}) +[2026-06-18 00:32:02,301][starVLA.training.train_starvla][INFO] - Step 5042, Loss: {'train/loss': 0.296549528837204, 'train/grad_norm_pre_clip': 69.69952392578125, 'timing/data': 0.0004398571327328682, 'timing/model': 1.2273475714027882, 'learning_rate/qwen_vl_interface': 2.2657731082867764e-06, 'learning_rate/action_model': 2.265773108286776e-05, 'epoch': 2.29}) +[2026-06-18 00:32:06,403][starVLA.training.train_starvla][INFO] - Step 5043, Loss: {'train/loss': 0.06803172826766968, 'train/grad_norm_pre_clip': 16.063671112060547, 'timing/data': 0.0005078110843896866, 'timing/model': 1.009287092834711, 'learning_rate/qwen_vl_interface': 2.264090833930555e-06, 'learning_rate/action_model': 2.2640908339305553e-05, 'epoch': 2.29}) +[2026-06-18 00:32:10,581][starVLA.training.train_starvla][INFO] - Step 5044, Loss: {'train/loss': 0.7414246797561646, 'train/grad_norm_pre_clip': 64.9811782836914, 'timing/data': 0.0005167024210095406, 'timing/model': 1.1059306198731065, 'learning_rate/qwen_vl_interface': 2.2624091785565165e-06, 'learning_rate/action_model': 2.2624091785565167e-05, 'epoch': 2.29}) +[2026-06-18 00:32:14,260][starVLA.training.train_starvla][INFO] - Step 5045, Loss: {'train/loss': 0.06636271625757217, 'train/grad_norm_pre_clip': 14.81281566619873, 'timing/data': 0.0004912577569484711, 'timing/model': 0.9245804166421294, 'learning_rate/qwen_vl_interface': 2.2607281425132728e-06, 'learning_rate/action_model': 2.260728142513273e-05, 'epoch': 2.29}) +[2026-06-18 00:32:18,181][starVLA.training.train_starvla][INFO] - Step 5046, Loss: {'train/loss': 0.1508287638425827, 'train/grad_norm_pre_clip': 16.099422454833984, 'timing/data': 0.000506112352013588, 'timing/model': 0.9169488837942481, 'learning_rate/qwen_vl_interface': 2.2590477261493e-06, 'learning_rate/action_model': 2.2590477261493e-05, 'epoch': 2.29}) +[2026-06-18 00:32:21,905][starVLA.training.train_starvla][INFO] - Step 5047, Loss: {'train/loss': 0.1948845535516739, 'train/grad_norm_pre_clip': 20.235280990600586, 'timing/data': 0.0004972992464900017, 'timing/model': 0.918468083254993, 'learning_rate/qwen_vl_interface': 2.257367929812953e-06, 'learning_rate/action_model': 2.2573679298129525e-05, 'epoch': 2.29}) +[2026-06-18 00:32:25,630][starVLA.training.train_starvla][INFO] - Step 5048, Loss: {'train/loss': 0.48015493154525757, 'train/grad_norm_pre_clip': 31.14975357055664, 'timing/data': 0.001578853465616703, 'timing/model': 0.9196219053119421, 'learning_rate/qwen_vl_interface': 2.2556887538524514e-06, 'learning_rate/action_model': 2.2556887538524512e-05, 'epoch': 2.29}) +[2026-06-18 00:32:29,305][starVLA.training.train_starvla][INFO] - Step 5049, Loss: {'train/loss': 0.3595799505710602, 'train/grad_norm_pre_clip': 25.566024780273438, 'timing/data': 0.0004541892558336258, 'timing/model': 0.9180034175515175, 'learning_rate/qwen_vl_interface': 2.2540101986158953e-06, 'learning_rate/action_model': 2.2540101986158953e-05, 'epoch': 2.29}) +[2026-06-18 00:32:33,275][starVLA.training.train_starvla][INFO] - Step 5050, Loss: {'train/loss': 0.08493004739284515, 'train/grad_norm_pre_clip': 26.015810012817383, 'timing/data': 0.0009670741856098175, 'timing/model': 1.2196936178952456, 'learning_rate/qwen_vl_interface': 2.2523322644512506e-06, 'learning_rate/action_model': 2.2523322644512504e-05, 'epoch': 2.29}) +[2026-06-18 00:32:36,843][starVLA.training.train_starvla][INFO] - Step 5051, Loss: {'train/loss': 0.46479785442352295, 'train/grad_norm_pre_clip': 40.6257438659668, 'timing/data': 0.0009777061641216278, 'timing/model': 0.918922713957727, 'learning_rate/qwen_vl_interface': 2.2506549517063497e-06, 'learning_rate/action_model': 2.2506549517063496e-05, 'epoch': 2.29}) +[2026-06-18 00:32:40,575][starVLA.training.train_starvla][INFO] - Step 5052, Loss: {'train/loss': 0.6991934180259705, 'train/grad_norm_pre_clip': 53.90612030029297, 'timing/data': 0.0004652021452784538, 'timing/model': 0.9252612693235278, 'learning_rate/qwen_vl_interface': 2.248978260728907e-06, 'learning_rate/action_model': 2.2489782607289067e-05, 'epoch': 2.29}) +[2026-06-18 00:32:44,311][starVLA.training.train_starvla][INFO] - Step 5053, Loss: {'train/loss': 0.19176775217056274, 'train/grad_norm_pre_clip': 22.818927764892578, 'timing/data': 0.00047560594975948334, 'timing/model': 0.9166631391271949, 'learning_rate/qwen_vl_interface': 2.2473021918664987e-06, 'learning_rate/action_model': 2.247302191866499e-05, 'epoch': 2.29}) +[2026-06-18 00:32:48,017][starVLA.training.train_starvla][INFO] - Step 5054, Loss: {'train/loss': 0.1904093325138092, 'train/grad_norm_pre_clip': 25.657440185546875, 'timing/data': 0.00044029857963323593, 'timing/model': 0.9144564494490623, 'learning_rate/qwen_vl_interface': 2.245626745466578e-06, 'learning_rate/action_model': 2.2456267454665782e-05, 'epoch': 2.29}) +[2026-06-18 00:32:51,827][starVLA.training.train_starvla][INFO] - Step 5055, Loss: {'train/loss': 0.34497585892677307, 'train/grad_norm_pre_clip': 33.109554290771484, 'timing/data': 0.0012031393125653267, 'timing/model': 0.916988667100668, 'learning_rate/qwen_vl_interface': 2.2439519218764657e-06, 'learning_rate/action_model': 2.2439519218764656e-05, 'epoch': 2.29}) +[2026-06-18 00:32:55,567][starVLA.training.train_starvla][INFO] - Step 5056, Loss: {'train/loss': 0.186228409409523, 'train/grad_norm_pre_clip': 26.815980911254883, 'timing/data': 0.0004481431096792221, 'timing/model': 0.920351973734796, 'learning_rate/qwen_vl_interface': 2.242277721443355e-06, 'learning_rate/action_model': 2.2422777214433545e-05, 'epoch': 2.29}) +[2026-06-18 00:32:59,311][starVLA.training.train_starvla][INFO] - Step 5057, Loss: {'train/loss': 0.08903481066226959, 'train/grad_norm_pre_clip': 14.466194152832031, 'timing/data': 0.000440206378698349, 'timing/model': 1.089821550063789, 'learning_rate/qwen_vl_interface': 2.2406041445143086e-06, 'learning_rate/action_model': 2.2406041445143086e-05, 'epoch': 2.29}) +[2026-06-18 00:33:02,859][starVLA.training.train_starvla][INFO] - Step 5058, Loss: {'train/loss': 0.5611711144447327, 'train/grad_norm_pre_clip': 32.69188690185547, 'timing/data': 0.0004949979484081268, 'timing/model': 0.9168485132977366, 'learning_rate/qwen_vl_interface': 2.2389311914362614e-06, 'learning_rate/action_model': 2.2389311914362616e-05, 'epoch': 2.29}) +[2026-06-18 00:33:06,577][starVLA.training.train_starvla][INFO] - Step 5059, Loss: {'train/loss': 0.09471846371889114, 'train/grad_norm_pre_clip': 27.45641326904297, 'timing/data': 0.000509326346218586, 'timing/model': 0.9192207381129265, 'learning_rate/qwen_vl_interface': 2.2372588625560185e-06, 'learning_rate/action_model': 2.2372588625560185e-05, 'epoch': 2.3}) +[2026-06-18 00:33:10,381][starVLA.training.train_starvla][INFO] - Step 5060, Loss: {'train/loss': 0.11319959163665771, 'train/grad_norm_pre_clip': 23.477388381958008, 'timing/data': 0.0004981942474842072, 'timing/model': 0.995839306153357, 'learning_rate/qwen_vl_interface': 2.2355871582202544e-06, 'learning_rate/action_model': 2.2355871582202544e-05, 'epoch': 2.3}) +[2026-06-18 00:33:14,111][starVLA.training.train_starvla][INFO] - Step 5061, Loss: {'train/loss': 0.3925681412220001, 'train/grad_norm_pre_clip': 33.053958892822266, 'timing/data': 0.0013844352215528488, 'timing/model': 0.9204838862642646, 'learning_rate/qwen_vl_interface': 2.2339160787755175e-06, 'learning_rate/action_model': 2.2339160787755174e-05, 'epoch': 2.3}) +[2026-06-18 00:33:17,836][starVLA.training.train_starvla][INFO] - Step 5062, Loss: {'train/loss': 0.46272698044776917, 'train/grad_norm_pre_clip': 27.810195922851562, 'timing/data': 0.0004822695627808571, 'timing/model': 0.9147015940397978, 'learning_rate/qwen_vl_interface': 2.23224562456822e-06, 'learning_rate/action_model': 2.23224562456822e-05, 'epoch': 2.3}) +[2026-06-18 00:33:21,629][starVLA.training.train_starvla][INFO] - Step 5063, Loss: {'train/loss': 0.1672334522008896, 'train/grad_norm_pre_clip': 24.32489585876465, 'timing/data': 0.00043739378452301025, 'timing/model': 0.9177881320938468, 'learning_rate/qwen_vl_interface': 2.2305757959446538e-06, 'learning_rate/action_model': 2.2305757959446537e-05, 'epoch': 2.3}) +[2026-06-18 00:33:25,814][starVLA.training.train_starvla][INFO] - Step 5064, Loss: {'train/loss': 0.08673327416181564, 'train/grad_norm_pre_clip': 15.815120697021484, 'timing/data': 0.0005219802260398865, 'timing/model': 1.3092694906517863, 'learning_rate/qwen_vl_interface': 2.228906593250976e-06, 'learning_rate/action_model': 2.228906593250976e-05, 'epoch': 2.3}) +[2026-06-18 00:33:29,803][starVLA.training.train_starvla][INFO] - Step 5065, Loss: {'train/loss': 0.3767860233783722, 'train/grad_norm_pre_clip': 61.91710662841797, 'timing/data': 0.00044438522309064865, 'timing/model': 1.3238521348685026, 'learning_rate/qwen_vl_interface': 2.2272380168332105e-06, 'learning_rate/action_model': 2.2272380168332103e-05, 'epoch': 2.3}) +[2026-06-18 00:33:33,721][starVLA.training.train_starvla][INFO] - Step 5066, Loss: {'train/loss': 0.12628710269927979, 'train/grad_norm_pre_clip': 18.806514739990234, 'timing/data': 0.0010748105123639107, 'timing/model': 0.9323583180084825, 'learning_rate/qwen_vl_interface': 2.225570067037255e-06, 'learning_rate/action_model': 2.2255700670372548e-05, 'epoch': 2.3}) +[2026-06-18 00:33:37,669][starVLA.training.train_starvla][INFO] - Step 5067, Loss: {'train/loss': 0.03599580004811287, 'train/grad_norm_pre_clip': 9.76717758178711, 'timing/data': 0.0005178758874535561, 'timing/model': 0.9268481032922864, 'learning_rate/qwen_vl_interface': 2.223902744208881e-06, 'learning_rate/action_model': 2.2239027442088808e-05, 'epoch': 2.3}) +[2026-06-18 00:33:41,406][starVLA.training.train_starvla][INFO] - Step 5068, Loss: {'train/loss': 0.2302725613117218, 'train/grad_norm_pre_clip': 39.7940559387207, 'timing/data': 0.0004393989220261574, 'timing/model': 0.9214241523295641, 'learning_rate/qwen_vl_interface': 2.2222360486937238e-06, 'learning_rate/action_model': 2.2222360486937235e-05, 'epoch': 2.3}) +[2026-06-18 00:33:45,176][starVLA.training.train_starvla][INFO] - Step 5069, Loss: {'train/loss': 0.11481418460607529, 'train/grad_norm_pre_clip': 24.215356826782227, 'timing/data': 0.00047469791024923325, 'timing/model': 0.933593574911356, 'learning_rate/qwen_vl_interface': 2.220569980837292e-06, 'learning_rate/action_model': 2.220569980837292e-05, 'epoch': 2.3}) +[2026-06-18 00:33:49,071][starVLA.training.train_starvla][INFO] - Step 5070, Loss: {'train/loss': 0.1930135041475296, 'train/grad_norm_pre_clip': 23.881122589111328, 'timing/data': 0.00048020854592323303, 'timing/model': 0.9250458888709545, 'learning_rate/qwen_vl_interface': 2.2189045409849625e-06, 'learning_rate/action_model': 2.2189045409849628e-05, 'epoch': 2.3}) +[2026-06-18 00:33:53,125][starVLA.training.train_starvla][INFO] - Step 5071, Loss: {'train/loss': 0.1917409598827362, 'train/grad_norm_pre_clip': 26.282249450683594, 'timing/data': 0.0005008243024349213, 'timing/model': 1.2807393642142415, 'learning_rate/qwen_vl_interface': 2.2172397294819843e-06, 'learning_rate/action_model': 2.2172397294819843e-05, 'epoch': 2.3}) +[2026-06-18 00:33:56,686][starVLA.training.train_starvla][INFO] - Step 5072, Loss: {'train/loss': 0.029154280200600624, 'train/grad_norm_pre_clip': 5.940386772155762, 'timing/data': 0.0008974559605121613, 'timing/model': 0.920854888856411, 'learning_rate/qwen_vl_interface': 2.215575546673473e-06, 'learning_rate/action_model': 2.215575546673473e-05, 'epoch': 2.3}) +[2026-06-18 00:34:00,597][starVLA.training.train_starvla][INFO] - Step 5073, Loss: {'train/loss': 0.1790497750043869, 'train/grad_norm_pre_clip': 24.988393783569336, 'timing/data': 0.0006805034354329109, 'timing/model': 0.9225571798160672, 'learning_rate/qwen_vl_interface': 2.213911992904416e-06, 'learning_rate/action_model': 2.213911992904416e-05, 'epoch': 2.3}) +[2026-06-18 00:34:04,364][starVLA.training.train_starvla][INFO] - Step 5074, Loss: {'train/loss': 0.060092706233263016, 'train/grad_norm_pre_clip': 12.556095123291016, 'timing/data': 0.0007611289620399475, 'timing/model': 0.9246221352368593, 'learning_rate/qwen_vl_interface': 2.2122490685196703e-06, 'learning_rate/action_model': 2.2122490685196704e-05, 'epoch': 2.3}) +[2026-06-18 00:34:08,122][starVLA.training.train_starvla][INFO] - Step 5075, Loss: {'train/loss': 0.11191806942224503, 'train/grad_norm_pre_clip': 18.1772518157959, 'timing/data': 0.0009087417274713516, 'timing/model': 0.9374820599332452, 'learning_rate/qwen_vl_interface': 2.210586773863961e-06, 'learning_rate/action_model': 2.210586773863961e-05, 'epoch': 2.3}) +[2026-06-18 00:34:11,940][starVLA.training.train_starvla][INFO] - Step 5076, Loss: {'train/loss': 0.5158669352531433, 'train/grad_norm_pre_clip': 48.9721794128418, 'timing/data': 0.0008575664833188057, 'timing/model': 0.9228706434369087, 'learning_rate/qwen_vl_interface': 2.208925109281884e-06, 'learning_rate/action_model': 2.208925109281884e-05, 'epoch': 2.3}) +[2026-06-18 00:34:15,716][starVLA.training.train_starvla][INFO] - Step 5077, Loss: {'train/loss': 0.1487007886171341, 'train/grad_norm_pre_clip': 19.992755889892578, 'timing/data': 0.0004825908690690994, 'timing/model': 0.9212184585630894, 'learning_rate/qwen_vl_interface': 2.2072640751179026e-06, 'learning_rate/action_model': 2.2072640751179025e-05, 'epoch': 2.3}) +[2026-06-18 00:34:19,489][starVLA.training.train_starvla][INFO] - Step 5078, Loss: {'train/loss': 0.07066570222377777, 'train/grad_norm_pre_clip': 16.388193130493164, 'timing/data': 0.0015064049512147903, 'timing/model': 1.142226668074727, 'learning_rate/qwen_vl_interface': 2.205603671716356e-06, 'learning_rate/action_model': 2.2056036717163557e-05, 'epoch': 2.3}) +[2026-06-18 00:34:23,481][starVLA.training.train_starvla][INFO] - Step 5079, Loss: {'train/loss': 0.22508040070533752, 'train/grad_norm_pre_clip': 24.44001007080078, 'timing/data': 0.0005035819485783577, 'timing/model': 0.9262323994189501, 'learning_rate/qwen_vl_interface': 2.2039438994214424e-06, 'learning_rate/action_model': 2.2039438994214424e-05, 'epoch': 2.3}) +[2026-06-18 00:34:27,472][starVLA.training.train_starvla][INFO] - Step 5080, Loss: {'train/loss': 0.1957847625017166, 'train/grad_norm_pre_clip': 18.16238021850586, 'timing/data': 0.0005988040938973427, 'timing/model': 0.9340307237580419, 'learning_rate/qwen_vl_interface': 2.202284758577235e-06, 'learning_rate/action_model': 2.2022847585772347e-05, 'epoch': 2.3}) +[2026-06-18 00:34:31,311][starVLA.training.train_starvla][INFO] - Step 5081, Loss: {'train/loss': 0.1832953691482544, 'train/grad_norm_pre_clip': 22.60800552368164, 'timing/data': 0.0004998026415705681, 'timing/model': 0.9269694052636623, 'learning_rate/qwen_vl_interface': 2.2006262495276774e-06, 'learning_rate/action_model': 2.2006262495276775e-05, 'epoch': 2.31}) +[2026-06-18 00:34:35,070][starVLA.training.train_starvla][INFO] - Step 5082, Loss: {'train/loss': 0.29224878549575806, 'train/grad_norm_pre_clip': 24.038095474243164, 'timing/data': 0.0006727594882249832, 'timing/model': 0.926334310323, 'learning_rate/qwen_vl_interface': 2.1989683726165795e-06, 'learning_rate/action_model': 2.1989683726165795e-05, 'epoch': 2.31}) +[2026-06-18 00:34:39,033][starVLA.training.train_starvla][INFO] - Step 5083, Loss: {'train/loss': 0.28200769424438477, 'train/grad_norm_pre_clip': 31.832279205322266, 'timing/data': 0.0004779435694217682, 'timing/model': 1.1700234012678266, 'learning_rate/qwen_vl_interface': 2.197311128187623e-06, 'learning_rate/action_model': 2.1973111281876232e-05, 'epoch': 2.31}) +[2026-06-18 00:34:42,821][starVLA.training.train_starvla][INFO] - Step 5084, Loss: {'train/loss': 0.04575430601835251, 'train/grad_norm_pre_clip': 8.7366943359375, 'timing/data': 0.00047602225095033646, 'timing/model': 0.9591925796121359, 'learning_rate/qwen_vl_interface': 2.1956545165843517e-06, 'learning_rate/action_model': 2.1956545165843517e-05, 'epoch': 2.31}) +[2026-06-18 00:34:46,668][starVLA.training.train_starvla][INFO] - Step 5085, Loss: {'train/loss': 0.15178847312927246, 'train/grad_norm_pre_clip': 21.01715087890625, 'timing/data': 0.0004922486841678619, 'timing/model': 1.142399220727384, 'learning_rate/qwen_vl_interface': 2.1939985381501866e-06, 'learning_rate/action_model': 2.1939985381501866e-05, 'epoch': 2.31}) +[2026-06-18 00:34:50,243][starVLA.training.train_starvla][INFO] - Step 5086, Loss: {'train/loss': 0.20237086713314056, 'train/grad_norm_pre_clip': 23.78521156311035, 'timing/data': 0.0004984671249985695, 'timing/model': 0.9200311983004212, 'learning_rate/qwen_vl_interface': 2.1923431932284135e-06, 'learning_rate/action_model': 2.1923431932284136e-05, 'epoch': 2.31}) +[2026-06-18 00:34:54,008][starVLA.training.train_starvla][INFO] - Step 5087, Loss: {'train/loss': 0.0990833044052124, 'train/grad_norm_pre_clip': 15.000144004821777, 'timing/data': 0.00048022717237472534, 'timing/model': 0.9293154273182154, 'learning_rate/qwen_vl_interface': 2.190688482162187e-06, 'learning_rate/action_model': 2.190688482162187e-05, 'epoch': 2.31}) +[2026-06-18 00:34:58,087][starVLA.training.train_starvla][INFO] - Step 5088, Loss: {'train/loss': 0.08103062957525253, 'train/grad_norm_pre_clip': 18.464580535888672, 'timing/data': 0.0004960857331752777, 'timing/model': 1.0112863341346383, 'learning_rate/qwen_vl_interface': 2.189034405294529e-06, 'learning_rate/action_model': 2.189034405294529e-05, 'epoch': 2.31}) +[2026-06-18 00:35:01,882][starVLA.training.train_starvla][INFO] - Step 5089, Loss: {'train/loss': 0.1174626275897026, 'train/grad_norm_pre_clip': 18.985403060913086, 'timing/data': 0.0005277041345834732, 'timing/model': 0.923374124802649, 'learning_rate/qwen_vl_interface': 2.1873809629683333e-06, 'learning_rate/action_model': 2.1873809629683333e-05, 'epoch': 2.31}) +[2026-06-18 00:35:05,668][starVLA.training.train_starvla][INFO] - Step 5090, Loss: {'train/loss': 0.19179955124855042, 'train/grad_norm_pre_clip': 19.36419105529785, 'timing/data': 0.00048054009675979614, 'timing/model': 0.9386694552376866, 'learning_rate/qwen_vl_interface': 2.185728155526359e-06, 'learning_rate/action_model': 2.1857281555263588e-05, 'epoch': 2.31}) +[2026-06-18 00:35:09,434][starVLA.training.train_starvla][INFO] - Step 5091, Loss: {'train/loss': 0.21227850019931793, 'train/grad_norm_pre_clip': 18.15016746520996, 'timing/data': 0.0007875310257077217, 'timing/model': 0.9407801469787955, 'learning_rate/qwen_vl_interface': 2.184075983311234e-06, 'learning_rate/action_model': 2.184075983311234e-05, 'epoch': 2.31}) +[2026-06-18 00:35:13,465][starVLA.training.train_starvla][INFO] - Step 5092, Loss: {'train/loss': 0.1123238205909729, 'train/grad_norm_pre_clip': 20.29810905456543, 'timing/data': 0.0005426909774541855, 'timing/model': 1.1323899030685425, 'learning_rate/qwen_vl_interface': 2.18242444666546e-06, 'learning_rate/action_model': 2.18242444666546e-05, 'epoch': 2.31}) +[2026-06-18 00:35:17,202][starVLA.training.train_starvla][INFO] - Step 5093, Loss: {'train/loss': 0.15134143829345703, 'train/grad_norm_pre_clip': 44.87363052368164, 'timing/data': 0.0004914989694952965, 'timing/model': 0.9345404226332903, 'learning_rate/qwen_vl_interface': 2.1807735459313974e-06, 'learning_rate/action_model': 2.180773545931397e-05, 'epoch': 2.31}) +[2026-06-18 00:35:21,087][starVLA.training.train_starvla][INFO] - Step 5094, Loss: {'train/loss': 0.3483223617076874, 'train/grad_norm_pre_clip': 34.75521469116211, 'timing/data': 0.0006806384772062302, 'timing/model': 0.999161034822464, 'learning_rate/qwen_vl_interface': 2.1791232814512815e-06, 'learning_rate/action_model': 2.1791232814512815e-05, 'epoch': 2.31}) +[2026-06-18 00:35:24,907][starVLA.training.train_starvla][INFO] - Step 5095, Loss: {'train/loss': 0.2842188775539398, 'train/grad_norm_pre_clip': 39.083839416503906, 'timing/data': 0.0004949094727635384, 'timing/model': 0.9676354564726353, 'learning_rate/qwen_vl_interface': 2.1774736535672117e-06, 'learning_rate/action_model': 2.1774736535672115e-05, 'epoch': 2.31}) +[2026-06-18 00:35:29,006][starVLA.training.train_starvla][INFO] - Step 5096, Loss: {'train/loss': 0.23577220737934113, 'train/grad_norm_pre_clip': 30.25244140625, 'timing/data': 0.0005222391337156296, 'timing/model': 0.9707681452855468, 'learning_rate/qwen_vl_interface': 2.1758246626211617e-06, 'learning_rate/action_model': 2.1758246626211617e-05, 'epoch': 2.31}) +[2026-06-18 00:35:33,250][starVLA.training.train_starvla][INFO] - Step 5097, Loss: {'train/loss': 0.27585461735725403, 'train/grad_norm_pre_clip': 26.897876739501953, 'timing/data': 0.00048399902880191803, 'timing/model': 0.9739104202017188, 'learning_rate/qwen_vl_interface': 2.174176308954969e-06, 'learning_rate/action_model': 2.174176308954969e-05, 'epoch': 2.31}) +[2026-06-18 00:35:37,085][starVLA.training.train_starvla][INFO] - Step 5098, Loss: {'train/loss': 0.2233562171459198, 'train/grad_norm_pre_clip': 45.907997131347656, 'timing/data': 0.00047191325575113297, 'timing/model': 0.9443588927388191, 'learning_rate/qwen_vl_interface': 2.172528592910335e-06, 'learning_rate/action_model': 2.1725285929103348e-05, 'epoch': 2.31}) +[2026-06-18 00:35:40,965][starVLA.training.train_starvla][INFO] - Step 5099, Loss: {'train/loss': 0.23267270624637604, 'train/grad_norm_pre_clip': 23.028812408447266, 'timing/data': 0.0012417864054441452, 'timing/model': 1.1414857544004917, 'learning_rate/qwen_vl_interface': 2.170881514828834e-06, 'learning_rate/action_model': 2.170881514828834e-05, 'epoch': 2.31}) +[2026-06-18 00:37:39,398][starVLA.training.train_starvla][INFO] - Step 5100, Loss: {'train/loss': 0.1339293271303177, 'train/grad_norm_pre_clip': 39.00149154663086, 'eval/loss': 0.2731367349624634, 'eval/latency_8/loss': 0.2731367349624634, 'eval/demon_fire_precision': 0.968721251149954, 'eval/demon_fire_recall': 0.9813606710158435, 'eval/demon_fire_f1': 0.975, 'eval/demon_fire_rate_model': 0.9073455759599333, 'eval/demon_fire_rate_teacher': 0.8956594323873122, 'eval/demon_left_precision': 0.926829268292683, 'eval/demon_left_recall': 0.7323340471092077, 'eval/demon_left_f1': 0.8181818181818182, 'eval/demon_left_rate_model': 0.10267111853088481, 'eval/demon_left_rate_teacher': 0.12993878686700056, 'eval/demon_right_precision': 0.8569254185692542, 'eval/demon_right_recall': 0.853030303030303, 'eval/demon_right_f1': 0.8549734244495063, 'eval/demon_right_rate_model': 0.18280467445742904, 'eval/demon_right_rate_teacher': 0.18363939899833054, 'eval/demon_move_f1': 0.8365776213156623, 'eval/demon_cc_f1': 0.9288909624346484, 'eval/latency_8/demon_fire_precision': 0.968721251149954, 'eval/latency_8/demon_fire_recall': 0.9813606710158435, 'eval/latency_8/demon_fire_f1': 0.975, 'eval/latency_8/demon_fire_rate_model': 0.9073455759599333, 'eval/latency_8/demon_fire_rate_teacher': 0.8956594323873122, 'eval/latency_8/demon_left_precision': 0.926829268292683, 'eval/latency_8/demon_left_recall': 0.7323340471092077, 'eval/latency_8/demon_left_f1': 0.8181818181818182, 'eval/latency_8/demon_left_rate_model': 0.10267111853088481, 'eval/latency_8/demon_left_rate_teacher': 0.12993878686700056, 'eval/latency_8/demon_right_precision': 0.8569254185692542, 'eval/latency_8/demon_right_recall': 0.853030303030303, 'eval/latency_8/demon_right_f1': 0.8549734244495063, 'eval/latency_8/demon_right_rate_model': 0.18280467445742904, 'eval/latency_8/demon_right_rate_teacher': 0.18363939899833054, 'eval/latency_8/demon_move_f1': 0.8365776213156623, 'eval/latency_8/demon_cc_f1': 0.9288909624346484, 'timing/data': 0.0005847830325365067, 'timing/model': 0.9319557659327984, 'learning_rate/qwen_vl_interface': 2.1692350750519093e-06, 'learning_rate/action_model': 2.1692350750519092e-05, 'epoch': 2.31}) +[2026-06-18 00:37:39,399][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_5100_state +[2026-06-18 00:37:39,404][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 00:37:45,578][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_5100_state/model.safetensors +[2026-06-18 00:38:00,550][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_5100_state/optimizer.bin +[2026-06-18 00:38:00,551][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_5100_state/sampler.bin +[2026-06-18 00:38:00,551][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_5100_state/sampler_1.bin +[2026-06-18 00:38:00,554][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_5100_state/random_states_0.pkl +[2026-06-18 00:38:00,557][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 5100 +[2026-06-18 00:38:00,559][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 00:38:00,569][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 00:38:04,489][starVLA.training.train_starvla][INFO] - Step 5101, Loss: {'train/loss': 0.23212000727653503, 'train/grad_norm_pre_clip': 16.843387603759766, 'timing/data': 0.0008498867973685265, 'timing/model': 1.1580615909770131, 'learning_rate/qwen_vl_interface': 2.167589273920869e-06, 'learning_rate/action_model': 2.167589273920869e-05, 'epoch': 2.31}) +[2026-06-18 00:38:08,357][starVLA.training.train_starvla][INFO] - Step 5102, Loss: {'train/loss': 0.063251793384552, 'train/grad_norm_pre_clip': 20.419057846069336, 'timing/data': 0.00048389844596385956, 'timing/model': 0.9257335523143411, 'learning_rate/qwen_vl_interface': 2.1659441117768888e-06, 'learning_rate/action_model': 2.1659441117768886e-05, 'epoch': 2.31}) +[2026-06-18 00:38:12,133][starVLA.training.train_starvla][INFO] - Step 5103, Loss: {'train/loss': 0.11576920747756958, 'train/grad_norm_pre_clip': 18.045713424682617, 'timing/data': 0.0005821436643600464, 'timing/model': 0.9369913935661316, 'learning_rate/qwen_vl_interface': 2.164299588961013e-06, 'learning_rate/action_model': 2.1642995889610128e-05, 'epoch': 2.32}) +[2026-06-18 00:38:15,982][starVLA.training.train_starvla][INFO] - Step 5104, Loss: {'train/loss': 0.42559027671813965, 'train/grad_norm_pre_clip': 48.46693801879883, 'timing/data': 0.0004893559962511063, 'timing/model': 0.9710659012198448, 'learning_rate/qwen_vl_interface': 2.162655705814153e-06, 'learning_rate/action_model': 2.1626557058141527e-05, 'epoch': 2.32}) +[2026-06-18 00:38:19,885][starVLA.training.train_starvla][INFO] - Step 5105, Loss: {'train/loss': 0.5062668919563293, 'train/grad_norm_pre_clip': 26.898761749267578, 'timing/data': 0.0008461968973278999, 'timing/model': 0.9202368576079607, 'learning_rate/qwen_vl_interface': 2.161012462677087e-06, 'learning_rate/action_model': 2.161012462677087e-05, 'epoch': 2.32}) +[2026-06-18 00:38:23,636][starVLA.training.train_starvla][INFO] - Step 5106, Loss: {'train/loss': 0.25426068902015686, 'train/grad_norm_pre_clip': 20.57292366027832, 'timing/data': 0.0006725871935486794, 'timing/model': 0.9257328659296036, 'learning_rate/qwen_vl_interface': 2.1593698598904612e-06, 'learning_rate/action_model': 2.1593698598904614e-05, 'epoch': 2.32}) +[2026-06-18 00:38:27,506][starVLA.training.train_starvla][INFO] - Step 5107, Loss: {'train/loss': 0.07172143459320068, 'train/grad_norm_pre_clip': 15.035042762756348, 'timing/data': 0.000477413646876812, 'timing/model': 1.0158959338441491, 'learning_rate/qwen_vl_interface': 2.1577278977947896e-06, 'learning_rate/action_model': 2.1577278977947897e-05, 'epoch': 2.32}) +[2026-06-18 00:38:31,294][starVLA.training.train_starvla][INFO] - Step 5108, Loss: {'train/loss': 0.25627633929252625, 'train/grad_norm_pre_clip': 20.266796112060547, 'timing/data': 0.0005265753716230392, 'timing/model': 1.149024836719036, 'learning_rate/qwen_vl_interface': 2.1560865767304524e-06, 'learning_rate/action_model': 2.156086576730452e-05, 'epoch': 2.32}) +[2026-06-18 00:38:34,837][starVLA.training.train_starvla][INFO] - Step 5109, Loss: {'train/loss': 0.26110249757766724, 'train/grad_norm_pre_clip': 32.86330032348633, 'timing/data': 0.0004943693056702614, 'timing/model': 0.9160430980846286, 'learning_rate/qwen_vl_interface': 2.154445897037696e-06, 'learning_rate/action_model': 2.154445897037696e-05, 'epoch': 2.32}) +[2026-06-18 00:38:38,587][starVLA.training.train_starvla][INFO] - Step 5110, Loss: {'train/loss': 0.02430967055261135, 'train/grad_norm_pre_clip': 5.682138919830322, 'timing/data': 0.0005285739898681641, 'timing/model': 0.9194999663159251, 'learning_rate/qwen_vl_interface': 2.1528058590566353e-06, 'learning_rate/action_model': 2.1528058590566353e-05, 'epoch': 2.32}) +[2026-06-18 00:38:42,460][starVLA.training.train_starvla][INFO] - Step 5111, Loss: {'train/loss': 0.3565236032009125, 'train/grad_norm_pre_clip': 30.87369155883789, 'timing/data': 0.000507851131260395, 'timing/model': 1.0085332533344626, 'learning_rate/qwen_vl_interface': 2.1511664631272543e-06, 'learning_rate/action_model': 2.1511664631272545e-05, 'epoch': 2.32}) +[2026-06-18 00:38:46,373][starVLA.training.train_starvla][INFO] - Step 5112, Loss: {'train/loss': 0.22493651509284973, 'train/grad_norm_pre_clip': 22.2056884765625, 'timing/data': 0.000493132509291172, 'timing/model': 0.9242088682949543, 'learning_rate/qwen_vl_interface': 2.1495277095894015e-06, 'learning_rate/action_model': 2.1495277095894014e-05, 'epoch': 2.32}) +[2026-06-18 00:38:50,594][starVLA.training.train_starvla][INFO] - Step 5113, Loss: {'train/loss': 0.2557324767112732, 'train/grad_norm_pre_clip': 28.32395362854004, 'timing/data': 0.0010288171470165253, 'timing/model': 0.9948381865397096, 'learning_rate/qwen_vl_interface': 2.147889598782787e-06, 'learning_rate/action_model': 2.147889598782787e-05, 'epoch': 2.32}) +[2026-06-18 00:38:54,323][starVLA.training.train_starvla][INFO] - Step 5114, Loss: {'train/loss': 0.1828746497631073, 'train/grad_norm_pre_clip': 20.12809181213379, 'timing/data': 0.0004888130351901054, 'timing/model': 0.926882503554225, 'learning_rate/qwen_vl_interface': 2.1462521310469987e-06, 'learning_rate/action_model': 2.1462521310469987e-05, 'epoch': 2.32}) +[2026-06-18 00:38:58,322][starVLA.training.train_starvla][INFO] - Step 5115, Loss: {'train/loss': 0.16228385269641876, 'train/grad_norm_pre_clip': 26.561016082763672, 'timing/data': 0.0005204938352108002, 'timing/model': 1.3296449482440948, 'learning_rate/qwen_vl_interface': 2.144615306721484e-06, 'learning_rate/action_model': 2.144615306721484e-05, 'epoch': 2.32}) +[2026-06-18 00:39:01,877][starVLA.training.train_starvla][INFO] - Step 5116, Loss: {'train/loss': 0.14837607741355896, 'train/grad_norm_pre_clip': 25.03725814819336, 'timing/data': 0.00048394128680229187, 'timing/model': 0.925671779550612, 'learning_rate/qwen_vl_interface': 2.1429791261455575e-06, 'learning_rate/action_model': 2.1429791261455572e-05, 'epoch': 2.32}) +[2026-06-18 00:39:05,648][starVLA.training.train_starvla][INFO] - Step 5117, Loss: {'train/loss': 0.286494642496109, 'train/grad_norm_pre_clip': 48.26110076904297, 'timing/data': 0.0004843408241868019, 'timing/model': 0.9247919488698244, 'learning_rate/qwen_vl_interface': 2.1413435896584006e-06, 'learning_rate/action_model': 2.1413435896584005e-05, 'epoch': 2.32}) +[2026-06-18 00:39:09,387][starVLA.training.train_starvla][INFO] - Step 5118, Loss: {'train/loss': 0.2662346363067627, 'train/grad_norm_pre_clip': 28.305221557617188, 'timing/data': 0.0005118502303957939, 'timing/model': 0.9325279425829649, 'learning_rate/qwen_vl_interface': 2.1397086975990633e-06, 'learning_rate/action_model': 2.139708697599063e-05, 'epoch': 2.32}) +[2026-06-18 00:39:13,132][starVLA.training.train_starvla][INFO] - Step 5119, Loss: {'train/loss': 0.3358251750469208, 'train/grad_norm_pre_clip': 30.37412452697754, 'timing/data': 0.001316978596150875, 'timing/model': 0.9215435106307268, 'learning_rate/qwen_vl_interface': 2.1380744503064597e-06, 'learning_rate/action_model': 2.1380744503064598e-05, 'epoch': 2.32}) +[2026-06-18 00:39:17,004][starVLA.training.train_starvla][INFO] - Step 5120, Loss: {'train/loss': 0.049001485109329224, 'train/grad_norm_pre_clip': 12.425552368164062, 'timing/data': 0.0004933187738060951, 'timing/model': 1.0087271612137556, 'learning_rate/qwen_vl_interface': 2.1364408481193714e-06, 'learning_rate/action_model': 2.1364408481193712e-05, 'epoch': 2.32}) +[2026-06-18 00:39:20,918][starVLA.training.train_starvla][INFO] - Step 5121, Loss: {'train/loss': 0.19551776349544525, 'train/grad_norm_pre_clip': 24.295166015625, 'timing/data': 0.0004457533359527588, 'timing/model': 0.9240758735686541, 'learning_rate/qwen_vl_interface': 2.1348078913764454e-06, 'learning_rate/action_model': 2.1348078913764452e-05, 'epoch': 2.32}) +[2026-06-18 00:39:24,680][starVLA.training.train_starvla][INFO] - Step 5122, Loss: {'train/loss': 0.16665343940258026, 'train/grad_norm_pre_clip': 22.379989624023438, 'timing/data': 0.0004994114860892296, 'timing/model': 1.1090205339714885, 'learning_rate/qwen_vl_interface': 2.1331755804161957e-06, 'learning_rate/action_model': 2.133175580416196e-05, 'epoch': 2.32}) +[2026-06-18 00:39:28,277][starVLA.training.train_starvla][INFO] - Step 5123, Loss: {'train/loss': 0.3806813955307007, 'train/grad_norm_pre_clip': 27.70846939086914, 'timing/data': 0.0008856579661369324, 'timing/model': 0.9240377601236105, 'learning_rate/qwen_vl_interface': 2.131543915577002e-06, 'learning_rate/action_model': 2.131543915577002e-05, 'epoch': 2.32}) +[2026-06-18 00:39:32,025][starVLA.training.train_starvla][INFO] - Step 5124, Loss: {'train/loss': 0.4297622740268707, 'train/grad_norm_pre_clip': 29.69141387939453, 'timing/data': 0.00046151597052812576, 'timing/model': 0.9222783436998725, 'learning_rate/qwen_vl_interface': 2.129912897197108e-06, 'learning_rate/action_model': 2.1299128971971084e-05, 'epoch': 2.32}) +[2026-06-18 00:39:35,817][starVLA.training.train_starvla][INFO] - Step 5125, Loss: {'train/loss': 0.23807302117347717, 'train/grad_norm_pre_clip': 40.70587921142578, 'timing/data': 0.000740797258913517, 'timing/model': 0.9486474087461829, 'learning_rate/qwen_vl_interface': 2.1282825256146305e-06, 'learning_rate/action_model': 2.12828252561463e-05, 'epoch': 2.33}) +[2026-06-18 00:39:39,595][starVLA.training.train_starvla][INFO] - Step 5126, Loss: {'train/loss': 0.10200914740562439, 'train/grad_norm_pre_clip': 17.937639236450195, 'timing/data': 0.0005106814205646515, 'timing/model': 0.919625106267631, 'learning_rate/qwen_vl_interface': 2.1266528011675453e-06, 'learning_rate/action_model': 2.126652801167545e-05, 'epoch': 2.33}) +[2026-06-18 00:39:43,374][starVLA.training.train_starvla][INFO] - Step 5127, Loss: {'train/loss': 0.04318796098232269, 'train/grad_norm_pre_clip': 9.70138168334961, 'timing/data': 0.0004756692796945572, 'timing/model': 0.9715495742857456, 'learning_rate/qwen_vl_interface': 2.125023724193694e-06, 'learning_rate/action_model': 2.1250237241936936e-05, 'epoch': 2.33}) +[2026-06-18 00:39:47,242][starVLA.training.train_starvla][INFO] - Step 5128, Loss: {'train/loss': 0.3610791862010956, 'train/grad_norm_pre_clip': 39.18613815307617, 'timing/data': 0.00044462457299232483, 'timing/model': 0.9165463773533702, 'learning_rate/qwen_vl_interface': 2.1233952950307857e-06, 'learning_rate/action_model': 2.1233952950307855e-05, 'epoch': 2.33}) +[2026-06-18 00:39:51,034][starVLA.training.train_starvla][INFO] - Step 5129, Loss: {'train/loss': 0.17006415128707886, 'train/grad_norm_pre_clip': 18.08657455444336, 'timing/data': 0.0009719831869006157, 'timing/model': 1.122958680614829, 'learning_rate/qwen_vl_interface': 2.1217675140163984e-06, 'learning_rate/action_model': 2.1217675140163983e-05, 'epoch': 2.33}) +[2026-06-18 00:39:54,623][starVLA.training.train_starvla][INFO] - Step 5130, Loss: {'train/loss': 0.47691601514816284, 'train/grad_norm_pre_clip': 51.843807220458984, 'timing/data': 0.0004980545490980148, 'timing/model': 0.9270566627383232, 'learning_rate/qwen_vl_interface': 2.1201403814879715e-06, 'learning_rate/action_model': 2.1201403814879717e-05, 'epoch': 2.33}) +[2026-06-18 00:39:58,348][starVLA.training.train_starvla][INFO] - Step 5131, Loss: {'train/loss': 0.09138335287570953, 'train/grad_norm_pre_clip': 19.963502883911133, 'timing/data': 0.00046079978346824646, 'timing/model': 0.9179613459855318, 'learning_rate/qwen_vl_interface': 2.1185138977828122e-06, 'learning_rate/action_model': 2.118513897782812e-05, 'epoch': 2.33}) +[2026-06-18 00:40:02,097][starVLA.training.train_starvla][INFO] - Step 5132, Loss: {'train/loss': 0.1989133358001709, 'train/grad_norm_pre_clip': 29.450380325317383, 'timing/data': 0.0005133692175149918, 'timing/model': 0.9415080277249217, 'learning_rate/qwen_vl_interface': 2.116888063238091e-06, 'learning_rate/action_model': 2.116888063238091e-05, 'epoch': 2.33}) +[2026-06-18 00:40:05,984][starVLA.training.train_starvla][INFO] - Step 5133, Loss: {'train/loss': 0.24267785251140594, 'train/grad_norm_pre_clip': 14.150274276733398, 'timing/data': 0.0013715773820877075, 'timing/model': 0.9279532888904214, 'learning_rate/qwen_vl_interface': 2.1152628781908455e-06, 'learning_rate/action_model': 2.1152628781908455e-05, 'epoch': 2.33}) +[2026-06-18 00:40:09,710][starVLA.training.train_starvla][INFO] - Step 5134, Loss: {'train/loss': 0.19772274792194366, 'train/grad_norm_pre_clip': 32.02131271362305, 'timing/data': 0.0004896214231848717, 'timing/model': 0.9262413866817951, 'learning_rate/qwen_vl_interface': 2.1136383429779787e-06, 'learning_rate/action_model': 2.1136383429779788e-05, 'epoch': 2.33}) +[2026-06-18 00:40:13,627][starVLA.training.train_starvla][INFO] - Step 5135, Loss: {'train/loss': 0.3001756966114044, 'train/grad_norm_pre_clip': 31.39127540588379, 'timing/data': 0.0005345586687326431, 'timing/model': 0.9187804274260998, 'learning_rate/qwen_vl_interface': 2.1120144579362582e-06, 'learning_rate/action_model': 2.1120144579362582e-05, 'epoch': 2.33}) +[2026-06-18 00:40:17,366][starVLA.training.train_starvla][INFO] - Step 5136, Loss: {'train/loss': 0.0887134000658989, 'train/grad_norm_pre_clip': 10.256939888000488, 'timing/data': 0.0005235923454165459, 'timing/model': 1.0960737997666001, 'learning_rate/qwen_vl_interface': 2.1103912234023175e-06, 'learning_rate/action_model': 2.1103912234023175e-05, 'epoch': 2.33}) +[2026-06-18 00:40:21,075][starVLA.training.train_starvla][INFO] - Step 5137, Loss: {'train/loss': 0.45813584327697754, 'train/grad_norm_pre_clip': 25.814786911010742, 'timing/data': 0.0007814168930053711, 'timing/model': 0.9227750170975924, 'learning_rate/qwen_vl_interface': 2.1087686397126544e-06, 'learning_rate/action_model': 2.1087686397126545e-05, 'epoch': 2.33}) +[2026-06-18 00:40:24,881][starVLA.training.train_starvla][INFO] - Step 5138, Loss: {'train/loss': 0.19483894109725952, 'train/grad_norm_pre_clip': 31.308475494384766, 'timing/data': 0.0004891399294137955, 'timing/model': 0.9311573812738061, 'learning_rate/qwen_vl_interface': 2.107146707203632e-06, 'learning_rate/action_model': 2.1071467072036323e-05, 'epoch': 2.33}) +[2026-06-18 00:40:28,964][starVLA.training.train_starvla][INFO] - Step 5139, Loss: {'train/loss': 0.029586143791675568, 'train/grad_norm_pre_clip': 5.109161376953125, 'timing/data': 0.000739571638405323, 'timing/model': 1.149641209281981, 'learning_rate/qwen_vl_interface': 2.105525426211478e-06, 'learning_rate/action_model': 2.1055254262114778e-05, 'epoch': 2.33}) +[2026-06-18 00:40:32,876][starVLA.training.train_starvla][INFO] - Step 5140, Loss: {'train/loss': 0.24151042103767395, 'train/grad_norm_pre_clip': 19.49581527709961, 'timing/data': 0.000484309159219265, 'timing/model': 0.9632761543616652, 'learning_rate/qwen_vl_interface': 2.1039047970722894e-06, 'learning_rate/action_model': 2.1039047970722893e-05, 'epoch': 2.33}) +[2026-06-18 00:40:36,637][starVLA.training.train_starvla][INFO] - Step 5141, Loss: {'train/loss': 0.24533861875534058, 'train/grad_norm_pre_clip': 20.636255264282227, 'timing/data': 0.0004924489185214043, 'timing/model': 0.9256347259506583, 'learning_rate/qwen_vl_interface': 2.1022848201220204e-06, 'learning_rate/action_model': 2.1022848201220204e-05, 'epoch': 2.33}) +[2026-06-18 00:40:40,611][starVLA.training.train_starvla][INFO] - Step 5142, Loss: {'train/loss': 0.13080668449401855, 'train/grad_norm_pre_clip': 31.806079864501953, 'timing/data': 0.0004344005137681961, 'timing/model': 1.144155533052981, 'learning_rate/qwen_vl_interface': 2.1006654956964935e-06, 'learning_rate/action_model': 2.1006654956964936e-05, 'epoch': 2.33}) +[2026-06-18 00:40:44,369][starVLA.training.train_starvla][INFO] - Step 5143, Loss: {'train/loss': 0.3148545026779175, 'train/grad_norm_pre_clip': 31.712297439575195, 'timing/data': 0.0014011217281222343, 'timing/model': 1.0867626024410129, 'learning_rate/qwen_vl_interface': 2.0990468241314004e-06, 'learning_rate/action_model': 2.0990468241314002e-05, 'epoch': 2.33}) +[2026-06-18 00:40:48,365][starVLA.training.train_starvla][INFO] - Step 5144, Loss: {'train/loss': 0.21279190480709076, 'train/grad_norm_pre_clip': 23.30956268310547, 'timing/data': 0.0004959963262081146, 'timing/model': 0.921044965274632, 'learning_rate/qwen_vl_interface': 2.0974288057622905e-06, 'learning_rate/action_model': 2.0974288057622903e-05, 'epoch': 2.33}) +[2026-06-18 00:40:52,557][starVLA.training.train_starvla][INFO] - Step 5145, Loss: {'train/loss': 0.0267656147480011, 'train/grad_norm_pre_clip': 4.9322710037231445, 'timing/data': 0.0004929983988404274, 'timing/model': 0.9289196999743581, 'learning_rate/qwen_vl_interface': 2.095811440924584e-06, 'learning_rate/action_model': 2.0958114409245835e-05, 'epoch': 2.33}) +[2026-06-18 00:40:56,532][starVLA.training.train_starvla][INFO] - Step 5146, Loss: {'train/loss': 0.23025569319725037, 'train/grad_norm_pre_clip': 34.383750915527344, 'timing/data': 0.0011866902932524681, 'timing/model': 0.9383690161630511, 'learning_rate/qwen_vl_interface': 2.0941947299535554e-06, 'learning_rate/action_model': 2.0941947299535552e-05, 'epoch': 2.33}) +[2026-06-18 00:41:00,324][starVLA.training.train_starvla][INFO] - Step 5147, Loss: {'train/loss': 0.16524499654769897, 'train/grad_norm_pre_clip': 22.856470108032227, 'timing/data': 0.00048136990517377853, 'timing/model': 0.9337291596457362, 'learning_rate/qwen_vl_interface': 2.0925786731843573e-06, 'learning_rate/action_model': 2.0925786731843572e-05, 'epoch': 2.34}) +[2026-06-18 00:41:04,166][starVLA.training.train_starvla][INFO] - Step 5148, Loss: {'train/loss': 0.18794816732406616, 'train/grad_norm_pre_clip': 25.411514282226562, 'timing/data': 0.000483599491417408, 'timing/model': 0.9463342409580946, 'learning_rate/qwen_vl_interface': 2.0909632709519976e-06, 'learning_rate/action_model': 2.090963270951998e-05, 'epoch': 2.34}) +[2026-06-18 00:41:07,935][starVLA.training.train_starvla][INFO] - Step 5149, Loss: {'train/loss': 0.1856871098279953, 'train/grad_norm_pre_clip': 24.92286491394043, 'timing/data': 0.0004494544118642807, 'timing/model': 0.9331572353839874, 'learning_rate/qwen_vl_interface': 2.0893485235913514e-06, 'learning_rate/action_model': 2.0893485235913514e-05, 'epoch': 2.34}) +[2026-06-18 00:41:12,648][starVLA.training.train_starvla][INFO] - Step 5150, Loss: {'train/loss': 0.16464462876319885, 'train/grad_norm_pre_clip': 32.43399429321289, 'timing/data': 0.0005678534507751465, 'timing/model': 1.1030830442905426, 'learning_rate/qwen_vl_interface': 2.0877344314371575e-06, 'learning_rate/action_model': 2.0877344314371573e-05, 'epoch': 2.34}) +[2026-06-18 00:41:16,470][starVLA.training.train_starvla][INFO] - Step 5151, Loss: {'train/loss': 0.21763819456100464, 'train/grad_norm_pre_clip': 24.39617919921875, 'timing/data': 0.0004836646839976311, 'timing/model': 0.9597366573289037, 'learning_rate/qwen_vl_interface': 2.086120994824018e-06, 'learning_rate/action_model': 2.0861209948240178e-05, 'epoch': 2.34}) +[2026-06-18 00:41:20,351][starVLA.training.train_starvla][INFO] - Step 5152, Loss: {'train/loss': 0.08444821834564209, 'train/grad_norm_pre_clip': 15.252493858337402, 'timing/data': 0.0005320906639099121, 'timing/model': 0.9325649747624993, 'learning_rate/qwen_vl_interface': 2.0845082140864012e-06, 'learning_rate/action_model': 2.084508214086401e-05, 'epoch': 2.34}) +[2026-06-18 00:41:24,161][starVLA.training.train_starvla][INFO] - Step 5153, Loss: {'train/loss': 0.37738949060440063, 'train/grad_norm_pre_clip': 37.417030334472656, 'timing/data': 0.000499628484249115, 'timing/model': 0.930174863897264, 'learning_rate/qwen_vl_interface': 2.0828960895586365e-06, 'learning_rate/action_model': 2.0828960895586364e-05, 'epoch': 2.34}) +[2026-06-18 00:41:27,992][starVLA.training.train_starvla][INFO] - Step 5154, Loss: {'train/loss': 0.2187592089176178, 'train/grad_norm_pre_clip': 27.18841552734375, 'timing/data': 0.0005165310576558113, 'timing/model': 0.9519911119714379, 'learning_rate/qwen_vl_interface': 2.0812846215749227e-06, 'learning_rate/action_model': 2.0812846215749227e-05, 'epoch': 2.34}) +[2026-06-18 00:41:31,778][starVLA.training.train_starvla][INFO] - Step 5155, Loss: {'train/loss': 0.26799827814102173, 'train/grad_norm_pre_clip': 31.461000442504883, 'timing/data': 0.0005576862022280693, 'timing/model': 0.9309454085305333, 'learning_rate/qwen_vl_interface': 2.079673810469316e-06, 'learning_rate/action_model': 2.0796738104693158e-05, 'epoch': 2.34}) +[2026-06-18 00:41:35,767][starVLA.training.train_starvla][INFO] - Step 5156, Loss: {'train/loss': 0.07455620169639587, 'train/grad_norm_pre_clip': 15.286898612976074, 'timing/data': 0.0006467932835221291, 'timing/model': 1.012021129950881, 'learning_rate/qwen_vl_interface': 2.0780636565757392e-06, 'learning_rate/action_model': 2.0780636565757392e-05, 'epoch': 2.34}) +[2026-06-18 00:41:39,508][starVLA.training.train_starvla][INFO] - Step 5157, Loss: {'train/loss': 0.15069378912448883, 'train/grad_norm_pre_clip': 26.971067428588867, 'timing/data': 0.0004974035546183586, 'timing/model': 1.0978191765025258, 'learning_rate/qwen_vl_interface': 2.0764541602279777e-06, 'learning_rate/action_model': 2.0764541602279778e-05, 'epoch': 2.34}) +[2026-06-18 00:41:43,074][starVLA.training.train_starvla][INFO] - Step 5158, Loss: {'train/loss': 0.07685936242341995, 'train/grad_norm_pre_clip': 22.747207641601562, 'timing/data': 0.0006924718618392944, 'timing/model': 0.9258603854104877, 'learning_rate/qwen_vl_interface': 2.0748453217596856e-06, 'learning_rate/action_model': 2.0748453217596856e-05, 'epoch': 2.34}) +[2026-06-18 00:41:46,814][starVLA.training.train_starvla][INFO] - Step 5159, Loss: {'train/loss': 0.26035842299461365, 'train/grad_norm_pre_clip': 30.229890823364258, 'timing/data': 0.0006726039573550224, 'timing/model': 0.9264709018170834, 'learning_rate/qwen_vl_interface': 2.0732371415043766e-06, 'learning_rate/action_model': 2.0732371415043765e-05, 'epoch': 2.34}) +[2026-06-18 00:41:50,831][starVLA.training.train_starvla][INFO] - Step 5160, Loss: {'train/loss': 0.3439705967903137, 'train/grad_norm_pre_clip': 36.70635223388672, 'timing/data': 0.0013923076912760735, 'timing/model': 0.9257460795342922, 'learning_rate/qwen_vl_interface': 2.071629619795425e-06, 'learning_rate/action_model': 2.071629619795425e-05, 'epoch': 2.34}) +[2026-06-18 00:41:54,606][starVLA.training.train_starvla][INFO] - Step 5161, Loss: {'train/loss': 0.12206912785768509, 'train/grad_norm_pre_clip': 21.863840103149414, 'timing/data': 0.00048489682376384735, 'timing/model': 0.9586504977196455, 'learning_rate/qwen_vl_interface': 2.070022756966071e-06, 'learning_rate/action_model': 2.0700227569660712e-05, 'epoch': 2.34}) +[2026-06-18 00:41:58,415][starVLA.training.train_starvla][INFO] - Step 5162, Loss: {'train/loss': 0.0942237600684166, 'train/grad_norm_pre_clip': 14.3948335647583, 'timing/data': 0.0009960541501641273, 'timing/model': 0.9353915611281991, 'learning_rate/qwen_vl_interface': 2.068416553349424e-06, 'learning_rate/action_model': 2.068416553349424e-05, 'epoch': 2.34}) +[2026-06-18 00:42:02,132][starVLA.training.train_starvla][INFO] - Step 5163, Loss: {'train/loss': 0.07686261832714081, 'train/grad_norm_pre_clip': 20.216350555419922, 'timing/data': 0.0004933308809995651, 'timing/model': 0.9232059502974153, 'learning_rate/qwen_vl_interface': 2.0668110092784484e-06, 'learning_rate/action_model': 2.0668110092784484e-05, 'epoch': 2.34}) +[2026-06-18 00:42:06,096][starVLA.training.train_starvla][INFO] - Step 5164, Loss: {'train/loss': 0.10359162092208862, 'train/grad_norm_pre_clip': 17.828149795532227, 'timing/data': 0.0004778364673256874, 'timing/model': 1.149853155016899, 'learning_rate/qwen_vl_interface': 2.0652061250859758e-06, 'learning_rate/action_model': 2.0652061250859756e-05, 'epoch': 2.34}) +[2026-06-18 00:42:09,759][starVLA.training.train_starvla][INFO] - Step 5165, Loss: {'train/loss': 0.2870847284793854, 'train/grad_norm_pre_clip': 38.99840545654297, 'timing/data': 0.0004840698093175888, 'timing/model': 0.9361423021182418, 'learning_rate/qwen_vl_interface': 2.0636019011047005e-06, 'learning_rate/action_model': 2.0636019011047003e-05, 'epoch': 2.34}) +[2026-06-18 00:42:13,519][starVLA.training.train_starvla][INFO] - Step 5166, Loss: {'train/loss': 0.3583309054374695, 'train/grad_norm_pre_clip': 29.608261108398438, 'timing/data': 0.0008345535025000572, 'timing/model': 0.9196323743090034, 'learning_rate/qwen_vl_interface': 2.0619983376671796e-06, 'learning_rate/action_model': 2.0619983376671793e-05, 'epoch': 2.34}) +[2026-06-18 00:42:17,242][starVLA.training.train_starvla][INFO] - Step 5167, Loss: {'train/loss': 0.04110855609178543, 'train/grad_norm_pre_clip': 7.40352201461792, 'timing/data': 0.0004912903532385826, 'timing/model': 0.917865083552897, 'learning_rate/qwen_vl_interface': 2.060395435105834e-06, 'learning_rate/action_model': 2.060395435105834e-05, 'epoch': 2.34}) +[2026-06-18 00:42:20,982][starVLA.training.train_starvla][INFO] - Step 5168, Loss: {'train/loss': 0.11471184343099594, 'train/grad_norm_pre_clip': 23.05009651184082, 'timing/data': 0.0005178889259696007, 'timing/model': 0.9242023443803191, 'learning_rate/qwen_vl_interface': 2.0587931937529465e-06, 'learning_rate/action_model': 2.0587931937529464e-05, 'epoch': 2.34}) +[2026-06-18 00:42:24,721][starVLA.training.train_starvla][INFO] - Step 5169, Loss: {'train/loss': 0.1744183748960495, 'train/grad_norm_pre_clip': 116.83055877685547, 'timing/data': 0.0005317078903317451, 'timing/model': 0.9256912302225828, 'learning_rate/qwen_vl_interface': 2.0571916139406648e-06, 'learning_rate/action_model': 2.0571916139406647e-05, 'epoch': 2.35}) +[2026-06-18 00:42:28,472][starVLA.training.train_starvla][INFO] - Step 5170, Loss: {'train/loss': 0.14769792556762695, 'train/grad_norm_pre_clip': 17.958160400390625, 'timing/data': 0.0007977522909641266, 'timing/model': 0.9295949134975672, 'learning_rate/qwen_vl_interface': 2.055590696000996e-06, 'learning_rate/action_model': 2.0555906960009962e-05, 'epoch': 2.35}) +[2026-06-18 00:42:32,230][starVLA.training.train_starvla][INFO] - Step 5171, Loss: {'train/loss': 0.15092317759990692, 'train/grad_norm_pre_clip': 17.849485397338867, 'timing/data': 0.00047002919018268585, 'timing/model': 1.1234506163746119, 'learning_rate/qwen_vl_interface': 2.0539904402658144e-06, 'learning_rate/action_model': 2.0539904402658142e-05, 'epoch': 2.35}) +[2026-06-18 00:42:35,864][starVLA.training.train_starvla][INFO] - Step 5172, Loss: {'train/loss': 0.5509500503540039, 'train/grad_norm_pre_clip': 37.20287322998047, 'timing/data': 0.0004878733307123184, 'timing/model': 0.9374524988234043, 'learning_rate/qwen_vl_interface': 2.052390847066851e-06, 'learning_rate/action_model': 2.052390847066851e-05, 'epoch': 2.35}) +[2026-06-18 00:42:39,783][starVLA.training.train_starvla][INFO] - Step 5173, Loss: {'train/loss': 0.190424382686615, 'train/grad_norm_pre_clip': 17.606693267822266, 'timing/data': 0.0004994766786694527, 'timing/model': 0.9295060988515615, 'learning_rate/qwen_vl_interface': 2.0507919167357093e-06, 'learning_rate/action_model': 2.050791916735709e-05, 'epoch': 2.35}) +[2026-06-18 00:42:43,548][starVLA.training.train_starvla][INFO] - Step 5174, Loss: {'train/loss': 0.17349423468112946, 'train/grad_norm_pre_clip': 20.731077194213867, 'timing/data': 0.00048540253192186356, 'timing/model': 0.930742003954947, 'learning_rate/qwen_vl_interface': 2.049193649603844e-06, 'learning_rate/action_model': 2.049193649603844e-05, 'epoch': 2.35}) +[2026-06-18 00:42:47,325][starVLA.training.train_starvla][INFO] - Step 5175, Loss: {'train/loss': 0.1804410070180893, 'train/grad_norm_pre_clip': 28.43660545349121, 'timing/data': 0.0005080439150333405, 'timing/model': 0.9548284495249391, 'learning_rate/qwen_vl_interface': 2.047596046002578e-06, 'learning_rate/action_model': 2.0475960460025778e-05, 'epoch': 2.35}) +[2026-06-18 00:42:51,062][starVLA.training.train_starvla][INFO] - Step 5176, Loss: {'train/loss': 0.3247041404247284, 'train/grad_norm_pre_clip': 38.81120300292969, 'timing/data': 0.0005074292421340942, 'timing/model': 0.9317128071561456, 'learning_rate/qwen_vl_interface': 2.0459991062630994e-06, 'learning_rate/action_model': 2.045999106263099e-05, 'epoch': 2.35}) +[2026-06-18 00:42:55,180][starVLA.training.train_starvla][INFO] - Step 5177, Loss: {'train/loss': 0.0815042182803154, 'train/grad_norm_pre_clip': 16.399677276611328, 'timing/data': 0.00046445056796073914, 'timing/model': 1.0538960741832852, 'learning_rate/qwen_vl_interface': 2.0444028307164532e-06, 'learning_rate/action_model': 2.0444028307164534e-05, 'epoch': 2.35}) +[2026-06-18 00:42:59,041][starVLA.training.train_starvla][INFO] - Step 5178, Loss: {'train/loss': 0.008452455513179302, 'train/grad_norm_pre_clip': 1.2534515857696533, 'timing/data': 0.0005146423354744911, 'timing/model': 1.0800783187150955, 'learning_rate/qwen_vl_interface': 2.0428072196935516e-06, 'learning_rate/action_model': 2.0428072196935514e-05, 'epoch': 2.35}) +[2026-06-18 00:43:02,595][starVLA.training.train_starvla][INFO] - Step 5179, Loss: {'train/loss': 0.3088303208351135, 'train/grad_norm_pre_clip': 27.06856346130371, 'timing/data': 0.0008288361132144928, 'timing/model': 0.9280871702358127, 'learning_rate/qwen_vl_interface': 2.041212273525161e-06, 'learning_rate/action_model': 2.041212273525161e-05, 'epoch': 2.35}) +[2026-06-18 00:43:06,354][starVLA.training.train_starvla][INFO] - Step 5180, Loss: {'train/loss': 0.12409201264381409, 'train/grad_norm_pre_clip': 31.96213150024414, 'timing/data': 0.0009120004251599312, 'timing/model': 0.927728989161551, 'learning_rate/qwen_vl_interface': 2.03961799254192e-06, 'learning_rate/action_model': 2.03961799254192e-05, 'epoch': 2.35}) +[2026-06-18 00:43:10,107][starVLA.training.train_starvla][INFO] - Step 5181, Loss: {'train/loss': 0.04134822636842728, 'train/grad_norm_pre_clip': 7.523351192474365, 'timing/data': 0.0004944819957017899, 'timing/model': 0.9251067023724318, 'learning_rate/qwen_vl_interface': 2.0380243770743246e-06, 'learning_rate/action_model': 2.0380243770743247e-05, 'epoch': 2.35}) +[2026-06-18 00:43:13,964][starVLA.training.train_starvla][INFO] - Step 5182, Loss: {'train/loss': 0.2764489948749542, 'train/grad_norm_pre_clip': 31.54545783996582, 'timing/data': 0.0005008857697248459, 'timing/model': 0.9324924815446138, 'learning_rate/qwen_vl_interface': 2.0364314274527318e-06, 'learning_rate/action_model': 2.0364314274527314e-05, 'epoch': 2.35}) +[2026-06-18 00:43:17,930][starVLA.training.train_starvla][INFO] - Step 5183, Loss: {'train/loss': 0.22668811678886414, 'train/grad_norm_pre_clip': 23.47604751586914, 'timing/data': 0.0004989113658666611, 'timing/model': 0.9268806139007211, 'learning_rate/qwen_vl_interface': 2.034839144007362e-06, 'learning_rate/action_model': 2.034839144007362e-05, 'epoch': 2.35}) +[2026-06-18 00:43:21,762][starVLA.training.train_starvla][INFO] - Step 5184, Loss: {'train/loss': 0.0728381872177124, 'train/grad_norm_pre_clip': 12.337523460388184, 'timing/data': 0.0005551772192120552, 'timing/model': 0.9279450923204422, 'learning_rate/qwen_vl_interface': 2.033247527068297e-06, 'learning_rate/action_model': 2.0332475270682972e-05, 'epoch': 2.35}) +[2026-06-18 00:43:25,553][starVLA.training.train_starvla][INFO] - Step 5185, Loss: {'train/loss': 0.47706326842308044, 'train/grad_norm_pre_clip': 31.179218292236328, 'timing/data': 0.0004912428557872772, 'timing/model': 1.1131853079423308, 'learning_rate/qwen_vl_interface': 2.0316565769654816e-06, 'learning_rate/action_model': 2.0316565769654816e-05, 'epoch': 2.35}) +[2026-06-18 00:43:29,124][starVLA.training.train_starvla][INFO] - Step 5186, Loss: {'train/loss': 0.1892145872116089, 'train/grad_norm_pre_clip': 23.388294219970703, 'timing/data': 0.0008260607719421387, 'timing/model': 0.9210111889988184, 'learning_rate/qwen_vl_interface': 2.0300662940287193e-06, 'learning_rate/action_model': 2.0300662940287192e-05, 'epoch': 2.35}) +[2026-06-18 00:43:32,942][starVLA.training.train_starvla][INFO] - Step 5187, Loss: {'train/loss': 0.3178425431251526, 'train/grad_norm_pre_clip': 19.6805362701416, 'timing/data': 0.000498088076710701, 'timing/model': 0.9276770725846291, 'learning_rate/qwen_vl_interface': 2.0284766785876812e-06, 'learning_rate/action_model': 2.0284766785876812e-05, 'epoch': 2.35}) +[2026-06-18 00:43:36,712][starVLA.training.train_starvla][INFO] - Step 5188, Loss: {'train/loss': 0.03915611654520035, 'train/grad_norm_pre_clip': 12.275972366333008, 'timing/data': 0.0006735026836395264, 'timing/model': 0.9348495816811919, 'learning_rate/qwen_vl_interface': 2.026887730971894e-06, 'learning_rate/action_model': 2.0268877309718938e-05, 'epoch': 2.35}) +[2026-06-18 00:43:40,653][starVLA.training.train_starvla][INFO] - Step 5189, Loss: {'train/loss': 0.13167321681976318, 'train/grad_norm_pre_clip': 16.80491065979004, 'timing/data': 0.000527295283973217, 'timing/model': 0.9193757381290197, 'learning_rate/qwen_vl_interface': 2.0252994515107475e-06, 'learning_rate/action_model': 2.0252994515107472e-05, 'epoch': 2.35}) +[2026-06-18 00:43:44,992][starVLA.training.train_starvla][INFO] - Step 5190, Loss: {'train/loss': 0.07335089892148972, 'train/grad_norm_pre_clip': 21.761301040649414, 'timing/data': 0.00047070253640413284, 'timing/model': 0.9302481599152088, 'learning_rate/qwen_vl_interface': 2.023711840533493e-06, 'learning_rate/action_model': 2.023711840533493e-05, 'epoch': 2.35}) +[2026-06-18 00:43:49,044][starVLA.training.train_starvla][INFO] - Step 5191, Loss: {'train/loss': 0.45079153776168823, 'train/grad_norm_pre_clip': 33.94160079956055, 'timing/data': 0.00047864578664302826, 'timing/model': 0.9373566256836057, 'learning_rate/qwen_vl_interface': 2.0221248983692475e-06, 'learning_rate/action_model': 2.0221248983692474e-05, 'epoch': 2.36}) +[2026-06-18 00:43:52,869][starVLA.training.train_starvla][INFO] - Step 5192, Loss: {'train/loss': 0.19443762302398682, 'train/grad_norm_pre_clip': 25.60299301147461, 'timing/data': 0.0004972442984580994, 'timing/model': 1.1301211100071669, 'learning_rate/qwen_vl_interface': 2.020538625346984e-06, 'learning_rate/action_model': 2.020538625346984e-05, 'epoch': 2.36}) +[2026-06-18 00:43:56,863][starVLA.training.train_starvla][INFO] - Step 5193, Loss: {'train/loss': 0.21702753007411957, 'train/grad_norm_pre_clip': 24.415485382080078, 'timing/data': 0.0005431687459349632, 'timing/model': 0.9255812158808112, 'learning_rate/qwen_vl_interface': 2.0189530217955382e-06, 'learning_rate/action_model': 2.018953021795538e-05, 'epoch': 2.36}) +[2026-06-18 00:44:01,069][starVLA.training.train_starvla][INFO] - Step 5194, Loss: {'train/loss': 0.8456305265426636, 'train/grad_norm_pre_clip': 48.63704299926758, 'timing/data': 0.0012412453070282936, 'timing/model': 1.0132140815258026, 'learning_rate/qwen_vl_interface': 2.0173680880436076e-06, 'learning_rate/action_model': 2.0173680880436076e-05, 'epoch': 2.36}) +[2026-06-18 00:44:05,039][starVLA.training.train_starvla][INFO] - Step 5195, Loss: {'train/loss': 0.06000111252069473, 'train/grad_norm_pre_clip': 12.112083435058594, 'timing/data': 0.0005383072420954704, 'timing/model': 0.9261699765920639, 'learning_rate/qwen_vl_interface': 2.015783824419751e-06, 'learning_rate/action_model': 2.0157838244197508e-05, 'epoch': 2.36}) +[2026-06-18 00:44:08,870][starVLA.training.train_starvla][INFO] - Step 5196, Loss: {'train/loss': 0.2390012890100479, 'train/grad_norm_pre_clip': 20.66999626159668, 'timing/data': 0.0004925038665533066, 'timing/model': 0.9392998153343797, 'learning_rate/qwen_vl_interface': 2.0142002312523877e-06, 'learning_rate/action_model': 2.014200231252388e-05, 'epoch': 2.36}) +[2026-06-18 00:44:12,886][starVLA.training.train_starvla][INFO] - Step 5197, Loss: {'train/loss': 0.1825595647096634, 'train/grad_norm_pre_clip': 21.724233627319336, 'timing/data': 0.0006030350923538208, 'timing/model': 0.9492217181250453, 'learning_rate/qwen_vl_interface': 2.0126173088697988e-06, 'learning_rate/action_model': 2.0126173088697987e-05, 'epoch': 2.36}) +[2026-06-18 00:44:16,645][starVLA.training.train_starvla][INFO] - Step 5198, Loss: {'train/loss': 0.1639813631772995, 'train/grad_norm_pre_clip': 110.92081451416016, 'timing/data': 0.00046220235526561737, 'timing/model': 0.9254734516143799, 'learning_rate/qwen_vl_interface': 2.0110350576001244e-06, 'learning_rate/action_model': 2.0110350576001245e-05, 'epoch': 2.36}) +[2026-06-18 00:44:20,939][starVLA.training.train_starvla][INFO] - Step 5199, Loss: {'train/loss': 0.08331486582756042, 'train/grad_norm_pre_clip': 22.810791015625, 'timing/data': 0.0005312711000442505, 'timing/model': 1.0746985264122486, 'learning_rate/qwen_vl_interface': 2.0094534777713684e-06, 'learning_rate/action_model': 2.0094534777713684e-05, 'epoch': 2.36}) +[2026-06-18 00:46:13,319][starVLA.training.train_starvla][INFO] - Step 5200, Loss: {'train/loss': 0.07969518005847931, 'train/grad_norm_pre_clip': 14.995600700378418, 'eval/loss': 0.267953485250473, 'eval/latency_8/loss': 0.267953485250473, 'eval/demon_fire_precision': 0.9633333333333334, 'eval/demon_fire_recall': 0.987573780677229, 'eval/demon_fire_f1': 0.9753029605767757, 'eval/demon_fire_rate_model': 0.9181969949916527, 'eval/demon_fire_rate_teacher': 0.8956594323873122, 'eval/demon_left_precision': 0.8990610328638498, 'eval/demon_left_recall': 0.8201284796573876, 'eval/demon_left_f1': 0.8577827547592385, 'eval/demon_left_rate_model': 0.11853088480801335, 'eval/demon_left_rate_teacher': 0.12993878686700056, 'eval/demon_right_precision': 0.8430493273542601, 'eval/demon_right_recall': 0.8545454545454545, 'eval/demon_right_f1': 0.8487584650112867, 'eval/demon_right_rate_model': 0.18614357262103506, 'eval/demon_right_rate_teacher': 0.18363939899833054, 'eval/demon_move_f1': 0.8532706098852626, 'eval/demon_cc_f1': 0.9351789649460233, 'eval/latency_8/demon_fire_precision': 0.9633333333333334, 'eval/latency_8/demon_fire_recall': 0.987573780677229, 'eval/latency_8/demon_fire_f1': 0.9753029605767757, 'eval/latency_8/demon_fire_rate_model': 0.9181969949916527, 'eval/latency_8/demon_fire_rate_teacher': 0.8956594323873122, 'eval/latency_8/demon_left_precision': 0.8990610328638498, 'eval/latency_8/demon_left_recall': 0.8201284796573876, 'eval/latency_8/demon_left_f1': 0.8577827547592385, 'eval/latency_8/demon_left_rate_model': 0.11853088480801335, 'eval/latency_8/demon_left_rate_teacher': 0.12993878686700056, 'eval/latency_8/demon_right_precision': 0.8430493273542601, 'eval/latency_8/demon_right_recall': 0.8545454545454545, 'eval/latency_8/demon_right_f1': 0.8487584650112867, 'eval/latency_8/demon_right_rate_model': 0.18614357262103506, 'eval/latency_8/demon_right_rate_teacher': 0.18363939899833054, 'eval/latency_8/demon_move_f1': 0.8532706098852626, 'eval/latency_8/demon_cc_f1': 0.9351789649460233, 'timing/data': 0.0005160942673683167, 'timing/model': 0.9216053560376167, 'learning_rate/qwen_vl_interface': 2.007872569711394e-06, 'learning_rate/action_model': 2.007872569711394e-05, 'epoch': 2.36}) +[2026-06-18 00:46:13,322][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_5200_state +[2026-06-18 00:46:13,326][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 00:46:19,677][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_5200_state/model.safetensors +[2026-06-18 00:46:35,033][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_5200_state/optimizer.bin +[2026-06-18 00:46:35,033][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_5200_state/sampler.bin +[2026-06-18 00:46:35,034][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_5200_state/sampler_1.bin +[2026-06-18 00:46:35,036][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_5200_state/random_states_0.pkl +[2026-06-18 00:46:35,038][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 5200 +[2026-06-18 00:46:35,039][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 00:46:35,050][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 00:46:38,812][starVLA.training.train_starvla][INFO] - Step 5201, Loss: {'train/loss': 0.3312484920024872, 'train/grad_norm_pre_clip': 27.35041046142578, 'timing/data': 0.00048340950161218643, 'timing/model': 1.117632569745183, 'learning_rate/qwen_vl_interface': 2.0062923337479223e-06, 'learning_rate/action_model': 2.006292333747922e-05, 'epoch': 2.36}) +[2026-06-18 00:46:42,384][starVLA.training.train_starvla][INFO] - Step 5202, Loss: {'train/loss': 0.12102639675140381, 'train/grad_norm_pre_clip': 15.799269676208496, 'timing/data': 0.0004939865320920944, 'timing/model': 0.9242684645578265, 'learning_rate/qwen_vl_interface': 2.0047127702085437e-06, 'learning_rate/action_model': 2.0047127702085437e-05, 'epoch': 2.36}) +[2026-06-18 00:46:46,117][starVLA.training.train_starvla][INFO] - Step 5203, Loss: {'train/loss': 0.20841704308986664, 'train/grad_norm_pre_clip': 21.03388786315918, 'timing/data': 0.0005319202318787575, 'timing/model': 0.9237747602164745, 'learning_rate/qwen_vl_interface': 2.003133879420698e-06, 'learning_rate/action_model': 2.003133879420698e-05, 'epoch': 2.36}) +[2026-06-18 00:46:49,858][starVLA.training.train_starvla][INFO] - Step 5204, Loss: {'train/loss': 0.2379741072654724, 'train/grad_norm_pre_clip': 25.3183650970459, 'timing/data': 0.001307436265051365, 'timing/model': 0.9383934671059251, 'learning_rate/qwen_vl_interface': 2.001555661711691e-06, 'learning_rate/action_model': 2.001555661711691e-05, 'epoch': 2.36}) +[2026-06-18 00:46:53,726][starVLA.training.train_starvla][INFO] - Step 5205, Loss: {'train/loss': 0.26843875646591187, 'train/grad_norm_pre_clip': 23.13434410095215, 'timing/data': 0.0005019446834921837, 'timing/model': 0.9779414525255561, 'learning_rate/qwen_vl_interface': 1.999978117408693e-06, 'learning_rate/action_model': 1.999978117408693e-05, 'epoch': 2.36}) +[2026-06-18 00:46:57,884][starVLA.training.train_starvla][INFO] - Step 5206, Loss: {'train/loss': 0.1718922108411789, 'train/grad_norm_pre_clip': 20.707181930541992, 'timing/data': 0.0009072273969650269, 'timing/model': 0.9278626069426537, 'learning_rate/qwen_vl_interface': 1.9984012468387285e-06, 'learning_rate/action_model': 1.9984012468387285e-05, 'epoch': 2.36}) +[2026-06-18 00:47:01,756][starVLA.training.train_starvla][INFO] - Step 5207, Loss: {'train/loss': 0.13054825365543365, 'train/grad_norm_pre_clip': 26.10310935974121, 'timing/data': 0.00044265110045671463, 'timing/model': 0.9225637307390571, 'learning_rate/qwen_vl_interface': 1.9968250503286857e-06, 'learning_rate/action_model': 1.996825050328686e-05, 'epoch': 2.36}) +[2026-06-18 00:47:05,554][starVLA.training.train_starvla][INFO] - Step 5208, Loss: {'train/loss': 0.14029498398303986, 'train/grad_norm_pre_clip': 19.430051803588867, 'timing/data': 0.0004955437034368515, 'timing/model': 1.1104850331321359, 'learning_rate/qwen_vl_interface': 1.9952495282053073e-06, 'learning_rate/action_model': 1.9952495282053073e-05, 'epoch': 2.36}) +[2026-06-18 00:47:09,380][starVLA.training.train_starvla][INFO] - Step 5209, Loss: {'train/loss': 0.09788186848163605, 'train/grad_norm_pre_clip': 17.142030715942383, 'timing/data': 0.0010887552052736282, 'timing/model': 1.1699435887858272, 'learning_rate/qwen_vl_interface': 1.993674680795206e-06, 'learning_rate/action_model': 1.9936746807952058e-05, 'epoch': 2.36}) +[2026-06-18 00:47:13,189][starVLA.training.train_starvla][INFO] - Step 5210, Loss: {'train/loss': 0.18795299530029297, 'train/grad_norm_pre_clip': 46.27382278442383, 'timing/data': 0.000527932308614254, 'timing/model': 0.9574108151718974, 'learning_rate/qwen_vl_interface': 1.992100508424847e-06, 'learning_rate/action_model': 1.9921005084248467e-05, 'epoch': 2.36}) +[2026-06-18 00:47:16,929][starVLA.training.train_starvla][INFO] - Step 5211, Loss: {'train/loss': 0.39018526673316956, 'train/grad_norm_pre_clip': 79.7153091430664, 'timing/data': 0.000495200976729393, 'timing/model': 0.9266135208308697, 'learning_rate/qwen_vl_interface': 1.990527011420558e-06, 'learning_rate/action_model': 1.990527011420558e-05, 'epoch': 2.36}) +[2026-06-18 00:47:20,708][starVLA.training.train_starvla][INFO] - Step 5212, Loss: {'train/loss': 0.2409307211637497, 'train/grad_norm_pre_clip': 42.01496505737305, 'timing/data': 0.0004824493080377579, 'timing/model': 0.931223850697279, 'learning_rate/qwen_vl_interface': 1.9889541901085274e-06, 'learning_rate/action_model': 1.9889541901085273e-05, 'epoch': 2.36}) +[2026-06-18 00:47:24,501][starVLA.training.train_starvla][INFO] - Step 5213, Loss: {'train/loss': 0.504184901714325, 'train/grad_norm_pre_clip': 49.8763313293457, 'timing/data': 0.0008333157747983932, 'timing/model': 0.9298701072111726, 'learning_rate/qwen_vl_interface': 1.987382044814802e-06, 'learning_rate/action_model': 1.987382044814802e-05, 'epoch': 2.36}) +[2026-06-18 00:47:28,674][starVLA.training.train_starvla][INFO] - Step 5214, Loss: {'train/loss': 0.049616847187280655, 'train/grad_norm_pre_clip': 15.744881629943848, 'timing/data': 0.00044353585690259933, 'timing/model': 1.1642930563539267, 'learning_rate/qwen_vl_interface': 1.9858105758652895e-06, 'learning_rate/action_model': 1.9858105758652892e-05, 'epoch': 2.37}) +[2026-06-18 00:47:32,956][starVLA.training.train_starvla][INFO] - Step 5215, Loss: {'train/loss': 0.3123282492160797, 'train/grad_norm_pre_clip': 34.086891174316406, 'timing/data': 0.0007487926632165909, 'timing/model': 1.5118289310485125, 'learning_rate/qwen_vl_interface': 1.9842397835857555e-06, 'learning_rate/action_model': 1.9842397835857554e-05, 'epoch': 2.37}) +[2026-06-18 00:47:36,628][starVLA.training.train_starvla][INFO] - Step 5216, Loss: {'train/loss': 0.10011032223701477, 'train/grad_norm_pre_clip': 21.243305206298828, 'timing/data': 0.0010767364874482155, 'timing/model': 0.9272257117554545, 'learning_rate/qwen_vl_interface': 1.982669668301831e-06, 'learning_rate/action_model': 1.982669668301831e-05, 'epoch': 2.37}) +[2026-06-18 00:47:40,389][starVLA.training.train_starvla][INFO] - Step 5217, Loss: {'train/loss': 0.4562526047229767, 'train/grad_norm_pre_clip': 36.69404220581055, 'timing/data': 0.00047956034541130066, 'timing/model': 0.9265259336680174, 'learning_rate/qwen_vl_interface': 1.9811002303389993e-06, 'learning_rate/action_model': 1.9811002303389993e-05, 'epoch': 2.37}) +[2026-06-18 00:47:44,178][starVLA.training.train_starvla][INFO] - Step 5218, Loss: {'train/loss': 0.06112099811434746, 'train/grad_norm_pre_clip': 14.480364799499512, 'timing/data': 0.000573732890188694, 'timing/model': 0.92044607270509, 'learning_rate/qwen_vl_interface': 1.9795314700226068e-06, 'learning_rate/action_model': 1.9795314700226066e-05, 'epoch': 2.37}) +[2026-06-18 00:47:48,592][starVLA.training.train_starvla][INFO] - Step 5219, Loss: {'train/loss': 0.17327196896076202, 'train/grad_norm_pre_clip': 31.294818878173828, 'timing/data': 0.00044494494795799255, 'timing/model': 1.2616124507039785, 'learning_rate/qwen_vl_interface': 1.9779633876778587e-06, 'learning_rate/action_model': 1.9779633876778586e-05, 'epoch': 2.37}) +[2026-06-18 00:47:53,561][starVLA.training.train_starvla][INFO] - Step 5220, Loss: {'train/loss': 0.09592283517122269, 'train/grad_norm_pre_clip': 28.834121704101562, 'timing/data': 0.0005053533241152763, 'timing/model': 1.7010026630014181, 'learning_rate/qwen_vl_interface': 1.9763959836298225e-06, 'learning_rate/action_model': 1.9763959836298225e-05, 'epoch': 2.37}) +[2026-06-18 00:47:57,499][starVLA.training.train_starvla][INFO] - Step 5221, Loss: {'train/loss': 0.11874152719974518, 'train/grad_norm_pre_clip': 20.61497688293457, 'timing/data': 0.0005903290584683418, 'timing/model': 1.1336426390334964, 'learning_rate/qwen_vl_interface': 1.9748292582034228e-06, 'learning_rate/action_model': 1.9748292582034228e-05, 'epoch': 2.37}) +[2026-06-18 00:48:01,356][starVLA.training.train_starvla][INFO] - Step 5222, Loss: {'train/loss': 0.0775160863995552, 'train/grad_norm_pre_clip': 16.213796615600586, 'timing/data': 0.0004794122651219368, 'timing/model': 1.1307088062167168, 'learning_rate/qwen_vl_interface': 1.9732632117234405e-06, 'learning_rate/action_model': 1.9732632117234404e-05, 'epoch': 2.37}) +[2026-06-18 00:48:04,944][starVLA.training.train_starvla][INFO] - Step 5223, Loss: {'train/loss': 0.1882585734128952, 'train/grad_norm_pre_clip': 18.02127456665039, 'timing/data': 0.0005295919254422188, 'timing/model': 0.9212190508842468, 'learning_rate/qwen_vl_interface': 1.9716978445145195e-06, 'learning_rate/action_model': 1.9716978445145193e-05, 'epoch': 2.37}) +[2026-06-18 00:48:08,741][starVLA.training.train_starvla][INFO] - Step 5224, Loss: {'train/loss': 0.04736462980508804, 'train/grad_norm_pre_clip': 10.391268730163574, 'timing/data': 0.0005306461825966835, 'timing/model': 0.9285725429654121, 'learning_rate/qwen_vl_interface': 1.9701331569011634e-06, 'learning_rate/action_model': 1.9701331569011633e-05, 'epoch': 2.37}) +[2026-06-18 00:48:12,475][starVLA.training.train_starvla][INFO] - Step 5225, Loss: {'train/loss': 0.29282131791114807, 'train/grad_norm_pre_clip': 27.89081382751465, 'timing/data': 0.0006933445110917091, 'timing/model': 0.925847863778472, 'learning_rate/qwen_vl_interface': 1.968569149207734e-06, 'learning_rate/action_model': 1.968569149207734e-05, 'epoch': 2.37}) +[2026-06-18 00:48:16,485][starVLA.training.train_starvla][INFO] - Step 5226, Loss: {'train/loss': 0.06421821564435959, 'train/grad_norm_pre_clip': 15.441797256469727, 'timing/data': 0.00048188772052526474, 'timing/model': 0.9550401400774717, 'learning_rate/qwen_vl_interface': 1.9670058217584505e-06, 'learning_rate/action_model': 1.9670058217584504e-05, 'epoch': 2.37}) +[2026-06-18 00:48:20,401][starVLA.training.train_starvla][INFO] - Step 5227, Loss: {'train/loss': 0.24519126117229462, 'train/grad_norm_pre_clip': 26.170753479003906, 'timing/data': 0.0005291858687996864, 'timing/model': 1.083512886427343, 'learning_rate/qwen_vl_interface': 1.965443174877393e-06, 'learning_rate/action_model': 1.965443174877393e-05, 'epoch': 2.37}) +[2026-06-18 00:48:24,322][starVLA.training.train_starvla][INFO] - Step 5228, Loss: {'train/loss': 0.027131127193570137, 'train/grad_norm_pre_clip': 7.681804656982422, 'timing/data': 0.0011969953775405884, 'timing/model': 0.9282357012853026, 'learning_rate/qwen_vl_interface': 1.9638812088885003e-06, 'learning_rate/action_model': 1.9638812088885e-05, 'epoch': 2.37}) +[2026-06-18 00:48:28,057][starVLA.training.train_starvla][INFO] - Step 5229, Loss: {'train/loss': 0.16003185510635376, 'train/grad_norm_pre_clip': 20.09899139404297, 'timing/data': 0.00044030603021383286, 'timing/model': 1.093489051796496, 'learning_rate/qwen_vl_interface': 1.9623199241155684e-06, 'learning_rate/action_model': 1.9623199241155683e-05, 'epoch': 2.37}) +[2026-06-18 00:48:31,680][starVLA.training.train_starvla][INFO] - Step 5230, Loss: {'train/loss': 0.3477751314640045, 'train/grad_norm_pre_clip': 31.036758422851562, 'timing/data': 0.0005083093419671059, 'timing/model': 0.9550901614129543, 'learning_rate/qwen_vl_interface': 1.9607593208822536e-06, 'learning_rate/action_model': 1.9607593208822535e-05, 'epoch': 2.37}) +[2026-06-18 00:48:35,640][starVLA.training.train_starvla][INFO] - Step 5231, Loss: {'train/loss': 0.06804663687944412, 'train/grad_norm_pre_clip': 13.90048599243164, 'timing/data': 0.00045976322144269943, 'timing/model': 0.9243512889370322, 'learning_rate/qwen_vl_interface': 1.9591993995120715e-06, 'learning_rate/action_model': 1.9591993995120712e-05, 'epoch': 2.37}) +[2026-06-18 00:48:39,359][starVLA.training.train_starvla][INFO] - Step 5232, Loss: {'train/loss': 0.17915427684783936, 'train/grad_norm_pre_clip': 25.36701011657715, 'timing/data': 0.0005191322416067123, 'timing/model': 0.921438897959888, 'learning_rate/qwen_vl_interface': 1.9576401603283944e-06, 'learning_rate/action_model': 1.9576401603283944e-05, 'epoch': 2.37}) +[2026-06-18 00:48:43,091][starVLA.training.train_starvla][INFO] - Step 5233, Loss: {'train/loss': 0.34299519658088684, 'train/grad_norm_pre_clip': 29.939313888549805, 'timing/data': 0.00046963803470134735, 'timing/model': 0.9206305220723152, 'learning_rate/qwen_vl_interface': 1.956081603654455e-06, 'learning_rate/action_model': 1.956081603654455e-05, 'epoch': 2.37}) +[2026-06-18 00:48:46,875][starVLA.training.train_starvla][INFO] - Step 5234, Loss: {'train/loss': 0.32230499386787415, 'train/grad_norm_pre_clip': 26.31539535522461, 'timing/data': 0.000482780858874321, 'timing/model': 0.9280795073136687, 'learning_rate/qwen_vl_interface': 1.9545237298133423e-06, 'learning_rate/action_model': 1.9545237298133424e-05, 'epoch': 2.37}) +[2026-06-18 00:48:50,691][starVLA.training.train_starvla][INFO] - Step 5235, Loss: {'train/loss': 0.41592124104499817, 'train/grad_norm_pre_clip': 31.325212478637695, 'timing/data': 0.0010355403646826744, 'timing/model': 0.9487279141321778, 'learning_rate/qwen_vl_interface': 1.9529665391280096e-06, 'learning_rate/action_model': 1.95296653912801e-05, 'epoch': 2.37}) +[2026-06-18 00:48:54,472][starVLA.training.train_starvla][INFO] - Step 5236, Loss: {'train/loss': 0.061877597123384476, 'train/grad_norm_pre_clip': 15.16475772857666, 'timing/data': 0.0004926500841975212, 'timing/model': 1.1027230899780989, 'learning_rate/qwen_vl_interface': 1.9514100319212594e-06, 'learning_rate/action_model': 1.9514100319212593e-05, 'epoch': 2.38}) +[2026-06-18 00:48:58,033][starVLA.training.train_starvla][INFO] - Step 5237, Loss: {'train/loss': 0.4411664605140686, 'train/grad_norm_pre_clip': 38.78858947753906, 'timing/data': 0.000814458355307579, 'timing/model': 0.9290622230619192, 'learning_rate/qwen_vl_interface': 1.949854208515757e-06, 'learning_rate/action_model': 1.9498542085157573e-05, 'epoch': 2.38}) +[2026-06-18 00:49:01,931][starVLA.training.train_starvla][INFO] - Step 5238, Loss: {'train/loss': 0.05185805261135101, 'train/grad_norm_pre_clip': 11.269783973693848, 'timing/data': 0.0004587080329656601, 'timing/model': 1.1076508425176144, 'learning_rate/qwen_vl_interface': 1.9482990692340305e-06, 'learning_rate/action_model': 1.9482990692340305e-05, 'epoch': 2.38}) +[2026-06-18 00:49:05,726][starVLA.training.train_starvla][INFO] - Step 5239, Loss: {'train/loss': 0.19852393865585327, 'train/grad_norm_pre_clip': 21.421016693115234, 'timing/data': 0.0010888176038861275, 'timing/model': 0.9361155144870281, 'learning_rate/qwen_vl_interface': 1.94674461439846e-06, 'learning_rate/action_model': 1.9467446143984598e-05, 'epoch': 2.38}) +[2026-06-18 00:49:09,465][starVLA.training.train_starvla][INFO] - Step 5240, Loss: {'train/loss': 0.2735411822795868, 'train/grad_norm_pre_clip': 28.732257843017578, 'timing/data': 0.0004921900108456612, 'timing/model': 0.921254837885499, 'learning_rate/qwen_vl_interface': 1.9451908443312863e-06, 'learning_rate/action_model': 1.9451908443312865e-05, 'epoch': 2.38}) +[2026-06-18 00:49:13,185][starVLA.training.train_starvla][INFO] - Step 5241, Loss: {'train/loss': 0.36189982295036316, 'train/grad_norm_pre_clip': 37.728824615478516, 'timing/data': 0.0004929192364215851, 'timing/model': 0.9214133387431502, 'learning_rate/qwen_vl_interface': 1.943637759354604e-06, 'learning_rate/action_model': 1.9436377593546042e-05, 'epoch': 2.38}) +[2026-06-18 00:49:16,920][starVLA.training.train_starvla][INFO] - Step 5242, Loss: {'train/loss': 0.2319464087486267, 'train/grad_norm_pre_clip': 29.747167587280273, 'timing/data': 0.0004350375384092331, 'timing/model': 0.9237416498363018, 'learning_rate/qwen_vl_interface': 1.942085359790373e-06, 'learning_rate/action_model': 1.942085359790373e-05, 'epoch': 2.38}) +[2026-06-18 00:49:20,650][starVLA.training.train_starvla][INFO] - Step 5243, Loss: {'train/loss': 0.0917808935046196, 'train/grad_norm_pre_clip': 16.726699829101562, 'timing/data': 0.000502399168908596, 'timing/model': 1.0881418818607926, 'learning_rate/qwen_vl_interface': 1.940533645960407e-06, 'learning_rate/action_model': 1.940533645960407e-05, 'epoch': 2.38}) +[2026-06-18 00:49:24,265][starVLA.training.train_starvla][INFO] - Step 5244, Loss: {'train/loss': 0.22045208513736725, 'train/grad_norm_pre_clip': 33.52308654785156, 'timing/data': 0.0004628729075193405, 'timing/model': 0.9568723533302546, 'learning_rate/qwen_vl_interface': 1.9389826181863767e-06, 'learning_rate/action_model': 1.9389826181863767e-05, 'epoch': 2.38}) +[2026-06-18 00:49:28,083][starVLA.training.train_starvla][INFO] - Step 5245, Loss: {'train/loss': 0.08878198266029358, 'train/grad_norm_pre_clip': 14.190114974975586, 'timing/data': 0.0008649853989481926, 'timing/model': 0.9277170253917575, 'learning_rate/qwen_vl_interface': 1.937432276789813e-06, 'learning_rate/action_model': 1.9374322767898127e-05, 'epoch': 2.38}) +[2026-06-18 00:49:31,906][starVLA.training.train_starvla][INFO] - Step 5246, Loss: {'train/loss': 0.17172536253929138, 'train/grad_norm_pre_clip': 25.370304107666016, 'timing/data': 0.00048497971147298813, 'timing/model': 0.9305344596505165, 'learning_rate/qwen_vl_interface': 1.935882622092103e-06, 'learning_rate/action_model': 1.935882622092103e-05, 'epoch': 2.38}) +[2026-06-18 00:49:35,757][starVLA.training.train_starvla][INFO] - Step 5247, Loss: {'train/loss': 0.32076314091682434, 'train/grad_norm_pre_clip': 37.959861755371094, 'timing/data': 0.00048820860683918, 'timing/model': 0.919243267737329, 'learning_rate/qwen_vl_interface': 1.9343336544144907e-06, 'learning_rate/action_model': 1.9343336544144906e-05, 'epoch': 2.38}) +[2026-06-18 00:49:39,498][starVLA.training.train_starvla][INFO] - Step 5248, Loss: {'train/loss': 0.37059545516967773, 'train/grad_norm_pre_clip': 32.13614273071289, 'timing/data': 0.00047070160508155823, 'timing/model': 0.9290033420547843, 'learning_rate/qwen_vl_interface': 1.9327853740780787e-06, 'learning_rate/action_model': 1.932785374078079e-05, 'epoch': 2.38}) +[2026-06-18 00:49:43,226][starVLA.training.train_starvla][INFO] - Step 5249, Loss: {'train/loss': 0.29893621802330017, 'train/grad_norm_pre_clip': 21.719894409179688, 'timing/data': 0.00041431281715631485, 'timing/model': 0.9161248486489058, 'learning_rate/qwen_vl_interface': 1.9312377814038314e-06, 'learning_rate/action_model': 1.9312377814038312e-05, 'epoch': 2.38}) +[2026-06-18 00:49:46,923][starVLA.training.train_starvla][INFO] - Step 5250, Loss: {'train/loss': 0.16696703433990479, 'train/grad_norm_pre_clip': 47.797874450683594, 'timing/data': 0.0004260353744029999, 'timing/model': 1.0875948630273342, 'learning_rate/qwen_vl_interface': 1.9296908767125612e-06, 'learning_rate/action_model': 1.929690876712561e-05, 'epoch': 2.38}) +[2026-06-18 00:49:50,489][starVLA.training.train_starvla][INFO] - Step 5251, Loss: {'train/loss': 0.3214089274406433, 'train/grad_norm_pre_clip': 38.700157165527344, 'timing/data': 0.0007155165076255798, 'timing/model': 0.9383259816095233, 'learning_rate/qwen_vl_interface': 1.928144660324945e-06, 'learning_rate/action_model': 1.928144660324945e-05, 'epoch': 2.38}) +[2026-06-18 00:49:55,147][starVLA.training.train_starvla][INFO] - Step 5252, Loss: {'train/loss': 0.1547725796699524, 'train/grad_norm_pre_clip': 17.287372589111328, 'timing/data': 0.001033923588693142, 'timing/model': 0.9354195324704051, 'learning_rate/qwen_vl_interface': 1.9265991325615135e-06, 'learning_rate/action_model': 1.9265991325615135e-05, 'epoch': 2.38}) +[2026-06-18 00:49:59,091][starVLA.training.train_starvla][INFO] - Step 5253, Loss: {'train/loss': 0.2925567030906677, 'train/grad_norm_pre_clip': 35.46874237060547, 'timing/data': 0.0004858197644352913, 'timing/model': 0.9240108281373978, 'learning_rate/qwen_vl_interface': 1.9250542937426604e-06, 'learning_rate/action_model': 1.9250542937426602e-05, 'epoch': 2.38}) +[2026-06-18 00:50:02,861][starVLA.training.train_starvla][INFO] - Step 5254, Loss: {'train/loss': 0.3311372995376587, 'train/grad_norm_pre_clip': 42.59485626220703, 'timing/data': 0.0011092321947216988, 'timing/model': 0.9246983779594302, 'learning_rate/qwen_vl_interface': 1.9235101441886307e-06, 'learning_rate/action_model': 1.923510144188631e-05, 'epoch': 2.38}) +[2026-06-18 00:50:06,606][starVLA.training.train_starvla][INFO] - Step 5255, Loss: {'train/loss': 0.14509867131710052, 'train/grad_norm_pre_clip': 20.543888092041016, 'timing/data': 0.0005370276048779488, 'timing/model': 0.9273044317960739, 'learning_rate/qwen_vl_interface': 1.921966684219525e-06, 'learning_rate/action_model': 1.921966684219525e-05, 'epoch': 2.38}) +[2026-06-18 00:50:10,353][starVLA.training.train_starvla][INFO] - Step 5256, Loss: {'train/loss': 0.1956549882888794, 'train/grad_norm_pre_clip': 21.93208885192871, 'timing/data': 0.00047438032925128937, 'timing/model': 0.9230618523433805, 'learning_rate/qwen_vl_interface': 1.920423914155308e-06, 'learning_rate/action_model': 1.920423914155308e-05, 'epoch': 2.38}) +[2026-06-18 00:50:14,088][starVLA.training.train_starvla][INFO] - Step 5257, Loss: {'train/loss': 0.1413291096687317, 'train/grad_norm_pre_clip': 31.693668365478516, 'timing/data': 0.0006900755688548088, 'timing/model': 1.0911234095692635, 'learning_rate/qwen_vl_interface': 1.9188818343157962e-06, 'learning_rate/action_model': 1.9188818343157964e-05, 'epoch': 2.38}) +[2026-06-18 00:50:17,668][starVLA.training.train_starvla][INFO] - Step 5258, Loss: {'train/loss': 0.19899514317512512, 'train/grad_norm_pre_clip': 19.08766746520996, 'timing/data': 0.0004664324223995209, 'timing/model': 0.9214798826724291, 'learning_rate/qwen_vl_interface': 1.9173404450206654e-06, 'learning_rate/action_model': 1.9173404450206654e-05, 'epoch': 2.39}) +[2026-06-18 00:50:21,497][starVLA.training.train_starvla][INFO] - Step 5259, Loss: {'train/loss': 0.12998934090137482, 'train/grad_norm_pre_clip': 24.90549087524414, 'timing/data': 0.0005204211920499802, 'timing/model': 0.9311385583132505, 'learning_rate/qwen_vl_interface': 1.9157997465894452e-06, 'learning_rate/action_model': 1.9157997465894454e-05, 'epoch': 2.39}) +[2026-06-18 00:50:25,359][starVLA.training.train_starvla][INFO] - Step 5260, Loss: {'train/loss': 0.22875909507274628, 'train/grad_norm_pre_clip': 28.540395736694336, 'timing/data': 0.0007518446072936058, 'timing/model': 0.9635624531656504, 'learning_rate/qwen_vl_interface': 1.9142597393415264e-06, 'learning_rate/action_model': 1.914259739341526e-05, 'epoch': 2.39}) +[2026-06-18 00:50:29,188][starVLA.training.train_starvla][INFO] - Step 5261, Loss: {'train/loss': 0.22680805623531342, 'train/grad_norm_pre_clip': 33.911556243896484, 'timing/data': 0.0004863934591412544, 'timing/model': 0.9243981288745999, 'learning_rate/qwen_vl_interface': 1.9127204235961524e-06, 'learning_rate/action_model': 1.9127204235961525e-05, 'epoch': 2.39}) +[2026-06-18 00:50:33,145][starVLA.training.train_starvla][INFO] - Step 5262, Loss: {'train/loss': 0.07448559999465942, 'train/grad_norm_pre_clip': 20.00068473815918, 'timing/data': 0.0004961490631103516, 'timing/model': 0.9261792786419392, 'learning_rate/qwen_vl_interface': 1.911181799672426e-06, 'learning_rate/action_model': 1.9111817996724264e-05, 'epoch': 2.39}) +[2026-06-18 00:50:37,471][starVLA.training.train_starvla][INFO] - Step 5263, Loss: {'train/loss': 0.3349114954471588, 'train/grad_norm_pre_clip': 34.89006423950195, 'timing/data': 0.0008341614156961441, 'timing/model': 1.5035281600430608, 'learning_rate/qwen_vl_interface': 1.909643867889305e-06, 'learning_rate/action_model': 1.909643867889305e-05, 'epoch': 2.39}) +[2026-06-18 00:50:41,222][starVLA.training.train_starvla][INFO] - Step 5264, Loss: {'train/loss': 0.1130697950720787, 'train/grad_norm_pre_clip': 16.24854278564453, 'timing/data': 0.0005020340904593468, 'timing/model': 1.0923068607226014, 'learning_rate/qwen_vl_interface': 1.908106628565605e-06, 'learning_rate/action_model': 1.9081066285656047e-05, 'epoch': 2.39}) +[2026-06-18 00:50:44,820][starVLA.training.train_starvla][INFO] - Step 5265, Loss: {'train/loss': 0.17418646812438965, 'train/grad_norm_pre_clip': 24.138914108276367, 'timing/data': 0.0004982519894838333, 'timing/model': 0.9261376075446606, 'learning_rate/qwen_vl_interface': 1.9065700820199965e-06, 'learning_rate/action_model': 1.9065700820199963e-05, 'epoch': 2.39}) +[2026-06-18 00:50:48,562][starVLA.training.train_starvla][INFO] - Step 5266, Loss: {'train/loss': 0.05302630364894867, 'train/grad_norm_pre_clip': 16.592567443847656, 'timing/data': 0.000499228946864605, 'timing/model': 0.9297704137861729, 'learning_rate/qwen_vl_interface': 1.9050342285710063e-06, 'learning_rate/action_model': 1.9050342285710062e-05, 'epoch': 2.39}) +[2026-06-18 00:50:52,409][starVLA.training.train_starvla][INFO] - Step 5267, Loss: {'train/loss': 0.17679527401924133, 'train/grad_norm_pre_clip': 44.01826095581055, 'timing/data': 0.0005145799368619919, 'timing/model': 0.9529545279219747, 'learning_rate/qwen_vl_interface': 1.9034990685370217e-06, 'learning_rate/action_model': 1.9034990685370217e-05, 'epoch': 2.39}) +[2026-06-18 00:50:56,309][starVLA.training.train_starvla][INFO] - Step 5268, Loss: {'train/loss': 0.08608589321374893, 'train/grad_norm_pre_clip': 19.333683013916016, 'timing/data': 0.0004945481196045876, 'timing/model': 0.9858098505064845, 'learning_rate/qwen_vl_interface': 1.901964602236282e-06, 'learning_rate/action_model': 1.901964602236282e-05, 'epoch': 2.39}) +[2026-06-18 00:51:00,504][starVLA.training.train_starvla][INFO] - Step 5269, Loss: {'train/loss': 0.3261792063713074, 'train/grad_norm_pre_clip': 30.295473098754883, 'timing/data': 0.0006351694464683533, 'timing/model': 0.927913149818778, 'learning_rate/qwen_vl_interface': 1.900430829986882e-06, 'learning_rate/action_model': 1.900430829986882e-05, 'epoch': 2.39}) +[2026-06-18 00:51:04,770][starVLA.training.train_starvla][INFO] - Step 5270, Loss: {'train/loss': 0.0641118660569191, 'train/grad_norm_pre_clip': 10.788521766662598, 'timing/data': 0.0011139987036585808, 'timing/model': 1.1426446279510856, 'learning_rate/qwen_vl_interface': 1.898897752106773e-06, 'learning_rate/action_model': 1.898897752106773e-05, 'epoch': 2.39}) +[2026-06-18 00:51:08,609][starVLA.training.train_starvla][INFO] - Step 5271, Loss: {'train/loss': 0.28971409797668457, 'train/grad_norm_pre_clip': 27.876558303833008, 'timing/data': 0.0010354118421673775, 'timing/model': 1.1058435421437025, 'learning_rate/qwen_vl_interface': 1.8973653689137683e-06, 'learning_rate/action_model': 1.8973653689137684e-05, 'epoch': 2.39}) +[2026-06-18 00:51:12,353][starVLA.training.train_starvla][INFO] - Step 5272, Loss: {'train/loss': 0.40614211559295654, 'train/grad_norm_pre_clip': 29.86311149597168, 'timing/data': 0.0010759588330984116, 'timing/model': 0.9262764360755682, 'learning_rate/qwen_vl_interface': 1.8958336807255298e-06, 'learning_rate/action_model': 1.8958336807255297e-05, 'epoch': 2.39}) +[2026-06-18 00:51:16,115][starVLA.training.train_starvla][INFO] - Step 5273, Loss: {'train/loss': 0.13731707632541656, 'train/grad_norm_pre_clip': 21.068466186523438, 'timing/data': 0.001412898302078247, 'timing/model': 0.9360107472166419, 'learning_rate/qwen_vl_interface': 1.8943026878595785e-06, 'learning_rate/action_model': 1.8943026878595786e-05, 'epoch': 2.39}) +[2026-06-18 00:51:19,858][starVLA.training.train_starvla][INFO] - Step 5274, Loss: {'train/loss': 0.2767905592918396, 'train/grad_norm_pre_clip': 28.59921646118164, 'timing/data': 0.0005936017259955406, 'timing/model': 0.9219273794442415, 'learning_rate/qwen_vl_interface': 1.8927723906332913e-06, 'learning_rate/action_model': 1.8927723906332912e-05, 'epoch': 2.39}) +[2026-06-18 00:51:23,734][starVLA.training.train_starvla][INFO] - Step 5275, Loss: {'train/loss': 0.10813901573419571, 'train/grad_norm_pre_clip': 23.904531478881836, 'timing/data': 0.0010601300746202469, 'timing/model': 0.942422685213387, 'learning_rate/qwen_vl_interface': 1.8912427893638996e-06, 'learning_rate/action_model': 1.8912427893638996e-05, 'epoch': 2.39}) +[2026-06-18 00:51:27,524][starVLA.training.train_starvla][INFO] - Step 5276, Loss: {'train/loss': 0.08727292716503143, 'train/grad_norm_pre_clip': 13.836485862731934, 'timing/data': 0.0010901745408773422, 'timing/model': 0.9255973109975457, 'learning_rate/qwen_vl_interface': 1.8897138843684918e-06, 'learning_rate/action_model': 1.8897138843684917e-05, 'epoch': 2.39}) +[2026-06-18 00:51:31,254][starVLA.training.train_starvla][INFO] - Step 5277, Loss: {'train/loss': 0.03276970982551575, 'train/grad_norm_pre_clip': 9.238615036010742, 'timing/data': 0.00048171449452638626, 'timing/model': 0.9260805211961269, 'learning_rate/qwen_vl_interface': 1.8881856759640122e-06, 'learning_rate/action_model': 1.8881856759640122e-05, 'epoch': 2.39}) +[2026-06-18 00:51:35,278][starVLA.training.train_starvla][INFO] - Step 5278, Loss: {'train/loss': 0.04018283635377884, 'train/grad_norm_pre_clip': 23.6920108795166, 'timing/data': 0.0005116425454616547, 'timing/model': 1.1263576848432422, 'learning_rate/qwen_vl_interface': 1.8866581644672589e-06, 'learning_rate/action_model': 1.8866581644672586e-05, 'epoch': 2.39}) +[2026-06-18 00:51:39,138][starVLA.training.train_starvla][INFO] - Step 5279, Loss: {'train/loss': 0.05233707278966904, 'train/grad_norm_pre_clip': 11.51445484161377, 'timing/data': 0.0008705584332346916, 'timing/model': 0.9384489823132753, 'learning_rate/qwen_vl_interface': 1.8851313501948872e-06, 'learning_rate/action_model': 1.885131350194887e-05, 'epoch': 2.39}) +[2026-06-18 00:51:43,172][starVLA.training.train_starvla][INFO] - Step 5280, Loss: {'train/loss': 0.13591986894607544, 'train/grad_norm_pre_clip': 20.70888328552246, 'timing/data': 0.00091518834233284, 'timing/model': 0.9448763690888882, 'learning_rate/qwen_vl_interface': 1.8836052334634072e-06, 'learning_rate/action_model': 1.883605233463407e-05, 'epoch': 2.4}) +[2026-06-18 00:51:47,154][starVLA.training.train_starvla][INFO] - Step 5281, Loss: {'train/loss': 0.342146635055542, 'train/grad_norm_pre_clip': 29.814395904541016, 'timing/data': 0.0005007823929190636, 'timing/model': 1.1464022807776928, 'learning_rate/qwen_vl_interface': 1.8820798145891834e-06, 'learning_rate/action_model': 1.8820798145891832e-05, 'epoch': 2.4}) +[2026-06-18 00:51:51,108][starVLA.training.train_starvla][INFO] - Step 5282, Loss: {'train/loss': 0.6012547612190247, 'train/grad_norm_pre_clip': 36.023895263671875, 'timing/data': 0.0004937425255775452, 'timing/model': 0.9205462532117963, 'learning_rate/qwen_vl_interface': 1.88055509388844e-06, 'learning_rate/action_model': 1.88055509388844e-05, 'epoch': 2.4}) +[2026-06-18 00:51:55,033][starVLA.training.train_starvla][INFO] - Step 5283, Loss: {'train/loss': 0.17264911532402039, 'train/grad_norm_pre_clip': 24.488933563232422, 'timing/data': 0.0005426900461316109, 'timing/model': 0.92224161606282, 'learning_rate/qwen_vl_interface': 1.8790310716772525e-06, 'learning_rate/action_model': 1.8790310716772524e-05, 'epoch': 2.4}) +[2026-06-18 00:51:58,766][starVLA.training.train_starvla][INFO] - Step 5284, Loss: {'train/loss': 0.11058034747838974, 'train/grad_norm_pre_clip': 14.910563468933105, 'timing/data': 0.0006444267928600311, 'timing/model': 0.9287500558421016, 'learning_rate/qwen_vl_interface': 1.8775077482715496e-06, 'learning_rate/action_model': 1.8775077482715494e-05, 'epoch': 2.4}) +[2026-06-18 00:52:02,513][starVLA.training.train_starvla][INFO] - Step 5285, Loss: {'train/loss': 0.030841564759612083, 'train/grad_norm_pre_clip': 11.077117919921875, 'timing/data': 0.0009108874946832657, 'timing/model': 1.090052803978324, 'learning_rate/qwen_vl_interface': 1.8759851239871176e-06, 'learning_rate/action_model': 1.8759851239871175e-05, 'epoch': 2.4}) +[2026-06-18 00:52:06,334][starVLA.training.train_starvla][INFO] - Step 5286, Loss: {'train/loss': 0.1918633133172989, 'train/grad_norm_pre_clip': 19.549678802490234, 'timing/data': 0.0005273502320051193, 'timing/model': 0.9271619468927383, 'learning_rate/qwen_vl_interface': 1.8744631991396023e-06, 'learning_rate/action_model': 1.8744631991396023e-05, 'epoch': 2.4}) +[2026-06-18 00:52:10,760][starVLA.training.train_starvla][INFO] - Step 5287, Loss: {'train/loss': 0.09358181804418564, 'train/grad_norm_pre_clip': 22.32398796081543, 'timing/data': 0.0005112644284963608, 'timing/model': 1.3461430603638291, 'learning_rate/qwen_vl_interface': 1.8729419740444977e-06, 'learning_rate/action_model': 1.8729419740444977e-05, 'epoch': 2.4}) +[2026-06-18 00:52:14,970][starVLA.training.train_starvla][INFO] - Step 5288, Loss: {'train/loss': 0.07476341724395752, 'train/grad_norm_pre_clip': 16.79102325439453, 'timing/data': 0.000579136423766613, 'timing/model': 1.0023135896772146, 'learning_rate/qwen_vl_interface': 1.8714214490171554e-06, 'learning_rate/action_model': 1.8714214490171552e-05, 'epoch': 2.4}) +[2026-06-18 00:52:18,786][starVLA.training.train_starvla][INFO] - Step 5289, Loss: {'train/loss': 0.4326496720314026, 'train/grad_norm_pre_clip': 60.91533279418945, 'timing/data': 0.0005054324865341187, 'timing/model': 0.922983642667532, 'learning_rate/qwen_vl_interface': 1.8699016243727826e-06, 'learning_rate/action_model': 1.8699016243727827e-05, 'epoch': 2.4}) +[2026-06-18 00:52:22,595][starVLA.training.train_starvla][INFO] - Step 5290, Loss: {'train/loss': 0.33275383710861206, 'train/grad_norm_pre_clip': 30.24561309814453, 'timing/data': 0.001189609058201313, 'timing/model': 0.9958889558911324, 'learning_rate/qwen_vl_interface': 1.8683825004264402e-06, 'learning_rate/action_model': 1.86838250042644e-05, 'epoch': 2.4}) +[2026-06-18 00:52:26,337][starVLA.training.train_starvla][INFO] - Step 5291, Loss: {'train/loss': 0.06606275588274002, 'train/grad_norm_pre_clip': 12.279054641723633, 'timing/data': 0.00043005868792533875, 'timing/model': 0.9250558661296964, 'learning_rate/qwen_vl_interface': 1.8668640774930444e-06, 'learning_rate/action_model': 1.8668640774930442e-05, 'epoch': 2.4}) +[2026-06-18 00:52:30,427][starVLA.training.train_starvla][INFO] - Step 5292, Loss: {'train/loss': 0.3201315701007843, 'train/grad_norm_pre_clip': 42.02970504760742, 'timing/data': 0.00044647976756095886, 'timing/model': 1.085558919236064, 'learning_rate/qwen_vl_interface': 1.8653463558873651e-06, 'learning_rate/action_model': 1.8653463558873652e-05, 'epoch': 2.4}) +[2026-06-18 00:52:34,588][starVLA.training.train_starvla][INFO] - Step 5293, Loss: {'train/loss': 0.0741298720240593, 'train/grad_norm_pre_clip': 22.320341110229492, 'timing/data': 0.000890740193426609, 'timing/model': 0.9643300417810678, 'learning_rate/qwen_vl_interface': 1.863829335924029e-06, 'learning_rate/action_model': 1.8638293359240288e-05, 'epoch': 2.4}) +[2026-06-18 00:52:38,370][starVLA.training.train_starvla][INFO] - Step 5294, Loss: {'train/loss': 0.18538916110992432, 'train/grad_norm_pre_clip': 23.840940475463867, 'timing/data': 0.000493338331580162, 'timing/model': 0.9264168245717883, 'learning_rate/qwen_vl_interface': 1.8623130179175148e-06, 'learning_rate/action_model': 1.8623130179175146e-05, 'epoch': 2.4}) +[2026-06-18 00:52:42,128][starVLA.training.train_starvla][INFO] - Step 5295, Loss: {'train/loss': 0.2127343863248825, 'train/grad_norm_pre_clip': 20.534961700439453, 'timing/data': 0.0004882039502263069, 'timing/model': 0.92601287830621, 'learning_rate/qwen_vl_interface': 1.8607974021821573e-06, 'learning_rate/action_model': 1.8607974021821572e-05, 'epoch': 2.4}) +[2026-06-18 00:52:46,166][starVLA.training.train_starvla][INFO] - Step 5296, Loss: {'train/loss': 0.33737701177597046, 'train/grad_norm_pre_clip': 32.59876251220703, 'timing/data': 0.0005265725776553154, 'timing/model': 0.9288877630606294, 'learning_rate/qwen_vl_interface': 1.8592824890321443e-06, 'learning_rate/action_model': 1.8592824890321442e-05, 'epoch': 2.4}) +[2026-06-18 00:52:50,236][starVLA.training.train_starvla][INFO] - Step 5297, Loss: {'train/loss': 0.125261589884758, 'train/grad_norm_pre_clip': 26.568145751953125, 'timing/data': 0.0012502055615186691, 'timing/model': 0.930610965937376, 'learning_rate/qwen_vl_interface': 1.8577682787815223e-06, 'learning_rate/action_model': 1.8577682787815225e-05, 'epoch': 2.4}) +[2026-06-18 00:52:54,195][starVLA.training.train_starvla][INFO] - Step 5298, Loss: {'train/loss': 0.23499687016010284, 'train/grad_norm_pre_clip': 39.64263153076172, 'timing/data': 0.0005372064188122749, 'timing/model': 1.0870470087975264, 'learning_rate/qwen_vl_interface': 1.856254771744185e-06, 'learning_rate/action_model': 1.856254771744185e-05, 'epoch': 2.4}) +[2026-06-18 00:52:58,044][starVLA.training.train_starvla][INFO] - Step 5299, Loss: {'train/loss': 0.09570927917957306, 'train/grad_norm_pre_clip': 20.6937198638916, 'timing/data': 0.0008619623258709908, 'timing/model': 1.1074904026463628, 'learning_rate/qwen_vl_interface': 1.8547419682338838e-06, 'learning_rate/action_model': 1.8547419682338836e-05, 'epoch': 2.4}) +[2026-06-18 00:54:49,884][starVLA.training.train_starvla][INFO] - Step 5300, Loss: {'train/loss': 0.13480105996131897, 'train/grad_norm_pre_clip': 22.38409423828125, 'eval/loss': 0.2684350907802582, 'eval/latency_8/loss': 0.2684350907802582, 'eval/demon_fire_precision': 0.9733581164807931, 'eval/demon_fire_recall': 0.9760795278036657, 'eval/demon_fire_f1': 0.9747169225996588, 'eval/demon_fire_rate_model': 0.8981636060100167, 'eval/demon_fire_rate_teacher': 0.8956594323873122, 'eval/demon_left_precision': 0.8943661971830986, 'eval/demon_left_recall': 0.815845824411135, 'eval/demon_left_f1': 0.8533034714445689, 'eval/demon_left_rate_model': 0.11853088480801335, 'eval/demon_left_rate_teacher': 0.12993878686700056, 'eval/demon_right_precision': 0.8837579617834395, 'eval/demon_right_recall': 0.8409090909090909, 'eval/demon_right_f1': 0.8618012422360247, 'eval/demon_right_rate_model': 0.17473567056204786, 'eval/demon_right_rate_teacher': 0.18363939899833054, 'eval/demon_move_f1': 0.8575523568402967, 'eval/demon_cc_f1': 0.9363382744210352, 'eval/latency_8/demon_fire_precision': 0.9733581164807931, 'eval/latency_8/demon_fire_recall': 0.9760795278036657, 'eval/latency_8/demon_fire_f1': 0.9747169225996588, 'eval/latency_8/demon_fire_rate_model': 0.8981636060100167, 'eval/latency_8/demon_fire_rate_teacher': 0.8956594323873122, 'eval/latency_8/demon_left_precision': 0.8943661971830986, 'eval/latency_8/demon_left_recall': 0.815845824411135, 'eval/latency_8/demon_left_f1': 0.8533034714445689, 'eval/latency_8/demon_left_rate_model': 0.11853088480801335, 'eval/latency_8/demon_left_rate_teacher': 0.12993878686700056, 'eval/latency_8/demon_right_precision': 0.8837579617834395, 'eval/latency_8/demon_right_recall': 0.8409090909090909, 'eval/latency_8/demon_right_f1': 0.8618012422360247, 'eval/latency_8/demon_right_rate_model': 0.17473567056204786, 'eval/latency_8/demon_right_rate_teacher': 0.18363939899833054, 'eval/latency_8/demon_move_f1': 0.8575523568402967, 'eval/latency_8/demon_cc_f1': 0.9363382744210352, 'timing/data': 0.000467187725007534, 'timing/model': 0.920559193007648, 'learning_rate/qwen_vl_interface': 1.8532298685642275e-06, 'learning_rate/action_model': 1.8532298685642274e-05, 'epoch': 2.4}) +[2026-06-18 00:54:49,888][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_5300_state +[2026-06-18 00:54:49,894][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 00:54:56,168][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_5300_state/model.safetensors +[2026-06-18 00:55:11,394][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_5300_state/optimizer.bin +[2026-06-18 00:55:11,394][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_5300_state/sampler.bin +[2026-06-18 00:55:11,395][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_5300_state/sampler_1.bin +[2026-06-18 00:55:11,396][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_5300_state/random_states_0.pkl +[2026-06-18 00:55:11,399][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 5300 +[2026-06-18 00:55:11,399][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 00:55:11,410][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 00:55:15,444][starVLA.training.train_starvla][INFO] - Step 5301, Loss: {'train/loss': 0.39474907517433167, 'train/grad_norm_pre_clip': 32.74366760253906, 'timing/data': 0.0009172782301902771, 'timing/model': 1.3303124429658055, 'learning_rate/qwen_vl_interface': 1.8517184730486745e-06, 'learning_rate/action_model': 1.8517184730486746e-05, 'epoch': 2.4}) +[2026-06-18 00:55:19,019][starVLA.training.train_starvla][INFO] - Step 5302, Loss: {'train/loss': 0.24858741462230682, 'train/grad_norm_pre_clip': 37.284812927246094, 'timing/data': 0.0006401268765330315, 'timing/model': 0.9514226811006665, 'learning_rate/qwen_vl_interface': 1.8502077820005408e-06, 'learning_rate/action_model': 1.8502077820005407e-05, 'epoch': 2.41}) +[2026-06-18 00:55:23,002][starVLA.training.train_starvla][INFO] - Step 5303, Loss: {'train/loss': 0.028858860954642296, 'train/grad_norm_pre_clip': 6.241490364074707, 'timing/data': 0.00069419015198946, 'timing/model': 0.9172517359256744, 'learning_rate/qwen_vl_interface': 1.8486977957329888e-06, 'learning_rate/action_model': 1.8486977957329886e-05, 'epoch': 2.41}) +[2026-06-18 00:55:26,882][starVLA.training.train_starvla][INFO] - Step 5304, Loss: {'train/loss': 0.1435157209634781, 'train/grad_norm_pre_clip': 17.645694732666016, 'timing/data': 0.0005290573462843895, 'timing/model': 1.0365959275513887, 'learning_rate/qwen_vl_interface': 1.847188514559045e-06, 'learning_rate/action_model': 1.847188514559045e-05, 'epoch': 2.41}) +[2026-06-18 00:55:30,671][starVLA.training.train_starvla][INFO] - Step 5305, Loss: {'train/loss': 0.2483893185853958, 'train/grad_norm_pre_clip': 41.71230697631836, 'timing/data': 0.000416802242398262, 'timing/model': 0.9114303831011057, 'learning_rate/qwen_vl_interface': 1.8456799387915839e-06, 'learning_rate/action_model': 1.8456799387915838e-05, 'epoch': 2.41}) +[2026-06-18 00:55:34,442][starVLA.training.train_starvla][INFO] - Step 5306, Loss: {'train/loss': 0.048298947513103485, 'train/grad_norm_pre_clip': 18.722993850708008, 'timing/data': 0.0005325833335518837, 'timing/model': 0.9203916098922491, 'learning_rate/qwen_vl_interface': 1.8441720687433345e-06, 'learning_rate/action_model': 1.8441720687433344e-05, 'epoch': 2.41}) +[2026-06-18 00:55:38,178][starVLA.training.train_starvla][INFO] - Step 5307, Loss: {'train/loss': 0.3868264853954315, 'train/grad_norm_pre_clip': 31.93214988708496, 'timing/data': 0.0005074869841337204, 'timing/model': 1.0952766155824065, 'learning_rate/qwen_vl_interface': 1.8426649047268808e-06, 'learning_rate/action_model': 1.8426649047268807e-05, 'epoch': 2.41}) +[2026-06-18 00:55:41,743][starVLA.training.train_starvla][INFO] - Step 5308, Loss: {'train/loss': 0.42940831184387207, 'train/grad_norm_pre_clip': 42.36391067504883, 'timing/data': 0.000496402382850647, 'timing/model': 0.9251621123403311, 'learning_rate/qwen_vl_interface': 1.8411584470546582e-06, 'learning_rate/action_model': 1.841158447054658e-05, 'epoch': 2.41}) +[2026-06-18 00:55:45,479][starVLA.training.train_starvla][INFO] - Step 5309, Loss: {'train/loss': 0.15608856081962585, 'train/grad_norm_pre_clip': 30.631929397583008, 'timing/data': 0.0007912842556834221, 'timing/model': 0.914556666277349, 'learning_rate/qwen_vl_interface': 1.8396526960389575e-06, 'learning_rate/action_model': 1.8396526960389575e-05, 'epoch': 2.41}) +[2026-06-18 00:55:49,312][starVLA.training.train_starvla][INFO] - Step 5310, Loss: {'train/loss': 0.26941588521003723, 'train/grad_norm_pre_clip': 33.55607986450195, 'timing/data': 0.0004930319264531136, 'timing/model': 0.9222631799057126, 'learning_rate/qwen_vl_interface': 1.8381476519919215e-06, 'learning_rate/action_model': 1.8381476519919212e-05, 'epoch': 2.41}) +[2026-06-18 00:55:53,012][starVLA.training.train_starvla][INFO] - Step 5311, Loss: {'train/loss': 0.3660791516304016, 'train/grad_norm_pre_clip': 31.580272674560547, 'timing/data': 0.0005203625187277794, 'timing/model': 0.9144575111567974, 'learning_rate/qwen_vl_interface': 1.8366433152255511e-06, 'learning_rate/action_model': 1.8366433152255512e-05, 'epoch': 2.41}) +[2026-06-18 00:55:56,733][starVLA.training.train_starvla][INFO] - Step 5312, Loss: {'train/loss': 0.3158653974533081, 'train/grad_norm_pre_clip': 36.73723220825195, 'timing/data': 0.0005063703283667564, 'timing/model': 0.9139268985018134, 'learning_rate/qwen_vl_interface': 1.8351396860516931e-06, 'learning_rate/action_model': 1.835139686051693e-05, 'epoch': 2.41}) +[2026-06-18 00:56:00,629][starVLA.training.train_starvla][INFO] - Step 5313, Loss: {'train/loss': 0.18351873755455017, 'train/grad_norm_pre_clip': 29.895931243896484, 'timing/data': 0.00046502891927957535, 'timing/model': 0.9217945886775851, 'learning_rate/qwen_vl_interface': 1.8336367647820527e-06, 'learning_rate/action_model': 1.8336367647820528e-05, 'epoch': 2.41}) +[2026-06-18 00:56:04,324][starVLA.training.train_starvla][INFO] - Step 5314, Loss: {'train/loss': 0.11675620079040527, 'train/grad_norm_pre_clip': 28.273216247558594, 'timing/data': 0.0010127006098628044, 'timing/model': 1.051409131847322, 'learning_rate/qwen_vl_interface': 1.832134551728186e-06, 'learning_rate/action_model': 1.832134551728186e-05, 'epoch': 2.41}) +[2026-06-18 00:56:07,875][starVLA.training.train_starvla][INFO] - Step 5315, Loss: {'train/loss': 0.24627289175987244, 'train/grad_norm_pre_clip': 38.14111328125, 'timing/data': 0.0005710776895284653, 'timing/model': 0.9202229138463736, 'learning_rate/qwen_vl_interface': 1.8306330472015051e-06, 'learning_rate/action_model': 1.8306330472015053e-05, 'epoch': 2.41}) +[2026-06-18 00:56:11,631][starVLA.training.train_starvla][INFO] - Step 5316, Loss: {'train/loss': 0.4563128352165222, 'train/grad_norm_pre_clip': 42.775054931640625, 'timing/data': 0.0010646851733326912, 'timing/model': 0.9382621347904205, 'learning_rate/qwen_vl_interface': 1.8291322515132757e-06, 'learning_rate/action_model': 1.8291322515132756e-05, 'epoch': 2.41}) +[2026-06-18 00:56:15,491][starVLA.training.train_starvla][INFO] - Step 5317, Loss: {'train/loss': 0.33020827174186707, 'train/grad_norm_pre_clip': 47.49747085571289, 'timing/data': 0.0007247580215334892, 'timing/model': 0.9379073111340404, 'learning_rate/qwen_vl_interface': 1.8276321649746076e-06, 'learning_rate/action_model': 1.8276321649746076e-05, 'epoch': 2.41}) +[2026-06-18 00:56:19,432][starVLA.training.train_starvla][INFO] - Step 5318, Loss: {'train/loss': 0.22406093776226044, 'train/grad_norm_pre_clip': 15.505236625671387, 'timing/data': 0.000718412920832634, 'timing/model': 0.9197516348212957, 'learning_rate/qwen_vl_interface': 1.826132787896476e-06, 'learning_rate/action_model': 1.826132787896476e-05, 'epoch': 2.41}) +[2026-06-18 00:56:23,244][starVLA.training.train_starvla][INFO] - Step 5319, Loss: {'train/loss': 0.19593380391597748, 'train/grad_norm_pre_clip': 23.633127212524414, 'timing/data': 0.0005570435896515846, 'timing/model': 0.9319464089348912, 'learning_rate/qwen_vl_interface': 1.8246341205897018e-06, 'learning_rate/action_model': 1.8246341205897017e-05, 'epoch': 2.41}) +[2026-06-18 00:56:27,165][starVLA.training.train_starvla][INFO] - Step 5320, Loss: {'train/loss': 0.1438441276550293, 'train/grad_norm_pre_clip': 25.79060173034668, 'timing/data': 0.0005543222650885582, 'timing/model': 1.125192134641111, 'learning_rate/qwen_vl_interface': 1.8231361633649606e-06, 'learning_rate/action_model': 1.8231361633649606e-05, 'epoch': 2.41}) +[2026-06-18 00:56:30,803][starVLA.training.train_starvla][INFO] - Step 5321, Loss: {'train/loss': 0.08083757013082504, 'train/grad_norm_pre_clip': 14.915576934814453, 'timing/data': 0.00066382996737957, 'timing/model': 0.9447395568713546, 'learning_rate/qwen_vl_interface': 1.8216389165327797e-06, 'learning_rate/action_model': 1.8216389165327798e-05, 'epoch': 2.41}) +[2026-06-18 00:56:34,593][starVLA.training.train_starvla][INFO] - Step 5322, Loss: {'train/loss': 0.09918314218521118, 'train/grad_norm_pre_clip': 21.98094940185547, 'timing/data': 0.0005000252276659012, 'timing/model': 0.924464974552393, 'learning_rate/qwen_vl_interface': 1.8201423804035406e-06, 'learning_rate/action_model': 1.8201423804035405e-05, 'epoch': 2.41}) +[2026-06-18 00:56:38,357][starVLA.training.train_starvla][INFO] - Step 5323, Loss: {'train/loss': 0.04346809908747673, 'train/grad_norm_pre_clip': 37.87473678588867, 'timing/data': 0.0004736287519335747, 'timing/model': 0.9218919854611158, 'learning_rate/qwen_vl_interface': 1.8186465552874769e-06, 'learning_rate/action_model': 1.8186465552874766e-05, 'epoch': 2.41}) +[2026-06-18 00:56:42,155][starVLA.training.train_starvla][INFO] - Step 5324, Loss: {'train/loss': 0.11693574488162994, 'train/grad_norm_pre_clip': 15.701966285705566, 'timing/data': 0.0004507647827267647, 'timing/model': 0.9741859408095479, 'learning_rate/qwen_vl_interface': 1.8171514414946745e-06, 'learning_rate/action_model': 1.8171514414946745e-05, 'epoch': 2.42}) +[2026-06-18 00:56:45,961][starVLA.training.train_starvla][INFO] - Step 5325, Loss: {'train/loss': 0.2786043882369995, 'train/grad_norm_pre_clip': 28.763124465942383, 'timing/data': 0.0005415035411715508, 'timing/model': 0.9241523947566748, 'learning_rate/qwen_vl_interface': 1.815657039335072e-06, 'learning_rate/action_model': 1.815657039335072e-05, 'epoch': 2.42}) +[2026-06-18 00:56:49,722][starVLA.training.train_starvla][INFO] - Step 5326, Loss: {'train/loss': 0.09313713014125824, 'train/grad_norm_pre_clip': 24.313541412353516, 'timing/data': 0.0004929797723889351, 'timing/model': 1.1155391903594136, 'learning_rate/qwen_vl_interface': 1.8141633491184612e-06, 'learning_rate/action_model': 1.814163349118461e-05, 'epoch': 2.42}) +[2026-06-18 00:56:53,490][starVLA.training.train_starvla][INFO] - Step 5327, Loss: {'train/loss': 0.044046998023986816, 'train/grad_norm_pre_clip': 12.72826099395752, 'timing/data': 0.001539750024676323, 'timing/model': 1.0674013569951057, 'learning_rate/qwen_vl_interface': 1.8126703711544848e-06, 'learning_rate/action_model': 1.812670371154485e-05, 'epoch': 2.42}) +[2026-06-18 00:56:57,035][starVLA.training.train_starvla][INFO] - Step 5328, Loss: {'train/loss': 0.31147223711013794, 'train/grad_norm_pre_clip': 22.28459358215332, 'timing/data': 0.0004705432802438736, 'timing/model': 0.9162534149363637, 'learning_rate/qwen_vl_interface': 1.8111781057526381e-06, 'learning_rate/action_model': 1.811178105752638e-05, 'epoch': 2.42}) +[2026-06-18 00:57:00,750][starVLA.training.train_starvla][INFO] - Step 5329, Loss: {'train/loss': 0.13846652209758759, 'train/grad_norm_pre_clip': 23.285585403442383, 'timing/data': 0.0004958193749189377, 'timing/model': 0.9187108529731631, 'learning_rate/qwen_vl_interface': 1.8096865532222714e-06, 'learning_rate/action_model': 1.8096865532222714e-05, 'epoch': 2.42}) +[2026-06-18 00:57:04,499][starVLA.training.train_starvla][INFO] - Step 5330, Loss: {'train/loss': 0.17838338017463684, 'train/grad_norm_pre_clip': 24.068037033081055, 'timing/data': 0.000552925281226635, 'timing/model': 0.9256302146241069, 'learning_rate/qwen_vl_interface': 1.8081957138725861e-06, 'learning_rate/action_model': 1.808195713872586e-05, 'epoch': 2.42}) +[2026-06-18 00:57:08,452][starVLA.training.train_starvla][INFO] - Step 5331, Loss: {'train/loss': 0.07360303401947021, 'train/grad_norm_pre_clip': 10.873625755310059, 'timing/data': 0.0006840676069259644, 'timing/model': 0.9343497259542346, 'learning_rate/qwen_vl_interface': 1.8067055880126311e-06, 'learning_rate/action_model': 1.806705588012631e-05, 'epoch': 2.42}) +[2026-06-18 00:57:12,213][starVLA.training.train_starvla][INFO] - Step 5332, Loss: {'train/loss': 0.2382855862379074, 'train/grad_norm_pre_clip': 24.548095703125, 'timing/data': 0.0013012532144784927, 'timing/model': 0.9300908530130982, 'learning_rate/qwen_vl_interface': 1.8052161759513114e-06, 'learning_rate/action_model': 1.8052161759513115e-05, 'epoch': 2.42}) +[2026-06-18 00:57:15,945][starVLA.training.train_starvla][INFO] - Step 5333, Loss: {'train/loss': 0.24304907023906708, 'train/grad_norm_pre_clip': 29.68269920349121, 'timing/data': 0.0005670851096510887, 'timing/model': 1.0816184291616082, 'learning_rate/qwen_vl_interface': 1.8037274779973868e-06, 'learning_rate/action_model': 1.8037274779973867e-05, 'epoch': 2.42}) +[2026-06-18 00:57:19,666][starVLA.training.train_starvla][INFO] - Step 5334, Loss: {'train/loss': 0.5068778991699219, 'train/grad_norm_pre_clip': 36.87255859375, 'timing/data': 0.0005971109494566917, 'timing/model': 1.0872506694868207, 'learning_rate/qwen_vl_interface': 1.8022394944594643e-06, 'learning_rate/action_model': 1.802239494459464e-05, 'epoch': 2.42}) +[2026-06-18 00:57:23,228][starVLA.training.train_starvla][INFO] - Step 5335, Loss: {'train/loss': 0.455101877450943, 'train/grad_norm_pre_clip': 48.282958984375, 'timing/data': 0.0006791297346353531, 'timing/model': 0.9224536567926407, 'learning_rate/qwen_vl_interface': 1.8007522256460062e-06, 'learning_rate/action_model': 1.8007522256460062e-05, 'epoch': 2.42}) +[2026-06-18 00:57:26,934][starVLA.training.train_starvla][INFO] - Step 5336, Loss: {'train/loss': 0.28369882702827454, 'train/grad_norm_pre_clip': 28.099071502685547, 'timing/data': 0.0005956180393695831, 'timing/model': 0.9160124622285366, 'learning_rate/qwen_vl_interface': 1.7992656718653201e-06, 'learning_rate/action_model': 1.79926567186532e-05, 'epoch': 2.42}) +[2026-06-18 00:57:30,690][starVLA.training.train_starvla][INFO] - Step 5337, Loss: {'train/loss': 0.455517053604126, 'train/grad_norm_pre_clip': 35.65219497680664, 'timing/data': 0.000691533088684082, 'timing/model': 0.984053609892726, 'learning_rate/qwen_vl_interface': 1.797779833425575e-06, 'learning_rate/action_model': 1.797779833425575e-05, 'epoch': 2.42}) +[2026-06-18 00:57:34,413][starVLA.training.train_starvla][INFO] - Step 5338, Loss: {'train/loss': 0.14100542664527893, 'train/grad_norm_pre_clip': 22.5687313079834, 'timing/data': 0.0006102491170167923, 'timing/model': 0.9173258692026138, 'learning_rate/qwen_vl_interface': 1.7962947106347846e-06, 'learning_rate/action_model': 1.7962947106347845e-05, 'epoch': 2.42}) +[2026-06-18 00:57:38,163][starVLA.training.train_starvla][INFO] - Step 5339, Loss: {'train/loss': 0.138778418302536, 'train/grad_norm_pre_clip': 21.019309997558594, 'timing/data': 0.0011724773794412613, 'timing/model': 0.9339189855381846, 'learning_rate/qwen_vl_interface': 1.794810303800818e-06, 'learning_rate/action_model': 1.794810303800818e-05, 'epoch': 2.42}) +[2026-06-18 00:57:41,916][starVLA.training.train_starvla][INFO] - Step 5340, Loss: {'train/loss': 0.4017528295516968, 'train/grad_norm_pre_clip': 27.727855682373047, 'timing/data': 0.0006623547524213791, 'timing/model': 1.1052192086353898, 'learning_rate/qwen_vl_interface': 1.7933266132313924e-06, 'learning_rate/action_model': 1.7933266132313923e-05, 'epoch': 2.42}) +[2026-06-18 00:57:45,534][starVLA.training.train_starvla][INFO] - Step 5341, Loss: {'train/loss': 0.026201095432043076, 'train/grad_norm_pre_clip': 7.02433967590332, 'timing/data': 0.000586184673011303, 'timing/model': 0.9236382488161325, 'learning_rate/qwen_vl_interface': 1.7918436392340803e-06, 'learning_rate/action_model': 1.7918436392340802e-05, 'epoch': 2.42}) +[2026-06-18 00:57:49,260][starVLA.training.train_starvla][INFO] - Step 5342, Loss: {'train/loss': 0.1315770149230957, 'train/grad_norm_pre_clip': 20.533477783203125, 'timing/data': 0.0006811069324612617, 'timing/model': 0.9236419750377536, 'learning_rate/qwen_vl_interface': 1.7903613821163022e-06, 'learning_rate/action_model': 1.790361382116302e-05, 'epoch': 2.42}) +[2026-06-18 00:57:53,027][starVLA.training.train_starvla][INFO] - Step 5343, Loss: {'train/loss': 0.17897555232048035, 'train/grad_norm_pre_clip': 24.568775177001953, 'timing/data': 0.0006850650534033775, 'timing/model': 0.9256968125700951, 'learning_rate/qwen_vl_interface': 1.7888798421853313e-06, 'learning_rate/action_model': 1.7888798421853313e-05, 'epoch': 2.42}) +[2026-06-18 00:57:57,010][starVLA.training.train_starvla][INFO] - Step 5344, Loss: {'train/loss': 0.2633214294910431, 'train/grad_norm_pre_clip': 32.50225067138672, 'timing/data': 0.0006633857265114784, 'timing/model': 0.9201972456648946, 'learning_rate/qwen_vl_interface': 1.7873990197482962e-06, 'learning_rate/action_model': 1.787399019748296e-05, 'epoch': 2.42}) +[2026-06-18 00:58:00,728][starVLA.training.train_starvla][INFO] - Step 5345, Loss: {'train/loss': 0.1666439324617386, 'train/grad_norm_pre_clip': 15.950799942016602, 'timing/data': 0.0009414199739694595, 'timing/model': 0.9243495734408498, 'learning_rate/qwen_vl_interface': 1.785918915112169e-06, 'learning_rate/action_model': 1.785918915112169e-05, 'epoch': 2.42}) +[2026-06-18 00:58:04,443][starVLA.training.train_starvla][INFO] - Step 5346, Loss: {'train/loss': 0.1890680342912674, 'train/grad_norm_pre_clip': 17.118196487426758, 'timing/data': 0.0005845148116350174, 'timing/model': 0.9236692124977708, 'learning_rate/qwen_vl_interface': 1.7844395285837773e-06, 'learning_rate/action_model': 1.7844395285837774e-05, 'epoch': 2.43}) +[2026-06-18 00:58:08,168][starVLA.training.train_starvla][INFO] - Step 5347, Loss: {'train/loss': 0.4098314344882965, 'train/grad_norm_pre_clip': 32.06144332885742, 'timing/data': 0.0006303489208221436, 'timing/model': 1.0981167061254382, 'learning_rate/qwen_vl_interface': 1.7829608604698003e-06, 'learning_rate/action_model': 1.7829608604698e-05, 'epoch': 2.43}) +[2026-06-18 00:58:11,792][starVLA.training.train_starvla][INFO] - Step 5348, Loss: {'train/loss': 0.37521329522132874, 'train/grad_norm_pre_clip': 37.81882858276367, 'timing/data': 0.0008042706176638603, 'timing/model': 0.9231481524184346, 'learning_rate/qwen_vl_interface': 1.7814829110767685e-06, 'learning_rate/action_model': 1.7814829110767685e-05, 'epoch': 2.43}) +[2026-06-18 00:58:15,517][starVLA.training.train_starvla][INFO] - Step 5349, Loss: {'train/loss': 0.12757608294487, 'train/grad_norm_pre_clip': 16.269519805908203, 'timing/data': 0.0006479956209659576, 'timing/model': 0.9127861429005861, 'learning_rate/qwen_vl_interface': 1.7800056807110627e-06, 'learning_rate/action_model': 1.7800056807110626e-05, 'epoch': 2.43}) +[2026-06-18 00:58:19,193][starVLA.training.train_starvla][INFO] - Step 5350, Loss: {'train/loss': 0.33361199498176575, 'train/grad_norm_pre_clip': 38.18937301635742, 'timing/data': 0.0005039982497692108, 'timing/model': 0.9176423456519842, 'learning_rate/qwen_vl_interface': 1.7785291696789103e-06, 'learning_rate/action_model': 1.7785291696789102e-05, 'epoch': 2.43}) +[2026-06-18 00:58:22,979][starVLA.training.train_starvla][INFO] - Step 5351, Loss: {'train/loss': 0.21343986690044403, 'train/grad_norm_pre_clip': 30.927507400512695, 'timing/data': 0.0007624495774507523, 'timing/model': 0.9217095440253615, 'learning_rate/qwen_vl_interface': 1.7770533782863983e-06, 'learning_rate/action_model': 1.7770533782863983e-05, 'epoch': 2.43}) +[2026-06-18 00:58:26,749][starVLA.training.train_starvla][INFO] - Step 5352, Loss: {'train/loss': 0.05851761996746063, 'train/grad_norm_pre_clip': 15.279494285583496, 'timing/data': 0.0005363048985600471, 'timing/model': 0.9226783057674766, 'learning_rate/qwen_vl_interface': 1.7755783068394578e-06, 'learning_rate/action_model': 1.7755783068394577e-05, 'epoch': 2.43}) +[2026-06-18 00:58:30,457][starVLA.training.train_starvla][INFO] - Step 5353, Loss: {'train/loss': 0.31766581535339355, 'train/grad_norm_pre_clip': 24.205284118652344, 'timing/data': 0.0008769892156124115, 'timing/model': 1.0813094032928348, 'learning_rate/qwen_vl_interface': 1.7741039556438733e-06, 'learning_rate/action_model': 1.7741039556438733e-05, 'epoch': 2.43}) +[2026-06-18 00:58:34,164][starVLA.training.train_starvla][INFO] - Step 5354, Loss: {'train/loss': 0.23174713551998138, 'train/grad_norm_pre_clip': 21.020824432373047, 'timing/data': 0.0004864279180765152, 'timing/model': 1.0855507235974073, 'learning_rate/qwen_vl_interface': 1.7726303250052785e-06, 'learning_rate/action_model': 1.7726303250052784e-05, 'epoch': 2.43}) +[2026-06-18 00:58:37,729][starVLA.training.train_starvla][INFO] - Step 5355, Loss: {'train/loss': 0.29491204023361206, 'train/grad_norm_pre_clip': 25.26202964782715, 'timing/data': 0.0007919296622276306, 'timing/model': 0.9158409340307117, 'learning_rate/qwen_vl_interface': 1.7711574152291591e-06, 'learning_rate/action_model': 1.771157415229159e-05, 'epoch': 2.43}) +[2026-06-18 00:58:41,786][starVLA.training.train_starvla][INFO] - Step 5356, Loss: {'train/loss': 0.1362527310848236, 'train/grad_norm_pre_clip': 25.12004852294922, 'timing/data': 0.00046000536531209946, 'timing/model': 0.9627329185605049, 'learning_rate/qwen_vl_interface': 1.7696852266208512e-06, 'learning_rate/action_model': 1.7696852266208512e-05, 'epoch': 2.43}) +[2026-06-18 00:58:46,779][starVLA.training.train_starvla][INFO] - Step 5357, Loss: {'train/loss': 0.10657459497451782, 'train/grad_norm_pre_clip': 16.456045150756836, 'timing/data': 0.0005083763971924782, 'timing/model': 1.931676403619349, 'learning_rate/qwen_vl_interface': 1.7682137594855408e-06, 'learning_rate/action_model': 1.768213759485541e-05, 'epoch': 2.43}) +[2026-06-18 00:58:50,893][starVLA.training.train_starvla][INFO] - Step 5358, Loss: {'train/loss': 0.3980081379413605, 'train/grad_norm_pre_clip': 32.02999496459961, 'timing/data': 0.0005194200202822685, 'timing/model': 0.9238349497318268, 'learning_rate/qwen_vl_interface': 1.7667430141282649e-06, 'learning_rate/action_model': 1.7667430141282648e-05, 'epoch': 2.43}) +[2026-06-18 00:58:54,672][starVLA.training.train_starvla][INFO] - Step 5359, Loss: {'train/loss': 0.23695945739746094, 'train/grad_norm_pre_clip': 31.28352165222168, 'timing/data': 0.0012290235608816147, 'timing/model': 0.9302025567740202, 'learning_rate/qwen_vl_interface': 1.7652729908539101e-06, 'learning_rate/action_model': 1.76527299085391e-05, 'epoch': 2.43}) +[2026-06-18 00:58:58,448][starVLA.training.train_starvla][INFO] - Step 5360, Loss: {'train/loss': 0.15115918219089508, 'train/grad_norm_pre_clip': 19.722843170166016, 'timing/data': 0.0004947995766997337, 'timing/model': 1.1281925793737173, 'learning_rate/qwen_vl_interface': 1.7638036899672137e-06, 'learning_rate/action_model': 1.7638036899672136e-05, 'epoch': 2.43}) +[2026-06-18 00:59:01,999][starVLA.training.train_starvla][INFO] - Step 5361, Loss: {'train/loss': 0.20316825807094574, 'train/grad_norm_pre_clip': 35.13460159301758, 'timing/data': 0.00047165900468826294, 'timing/model': 0.9149957802146673, 'learning_rate/qwen_vl_interface': 1.7623351117727628e-06, 'learning_rate/action_model': 1.7623351117727626e-05, 'epoch': 2.43}) +[2026-06-18 00:59:05,737][starVLA.training.train_starvla][INFO] - Step 5362, Loss: {'train/loss': 0.16617931425571442, 'train/grad_norm_pre_clip': 25.494308471679688, 'timing/data': 0.0004859957844018936, 'timing/model': 0.9234510278329253, 'learning_rate/qwen_vl_interface': 1.7608672565749973e-06, 'learning_rate/action_model': 1.760867256574997e-05, 'epoch': 2.43}) +[2026-06-18 00:59:09,449][starVLA.training.train_starvla][INFO] - Step 5363, Loss: {'train/loss': 0.6241130232810974, 'train/grad_norm_pre_clip': 40.338768005371094, 'timing/data': 0.0005040401592850685, 'timing/model': 0.9247418874874711, 'learning_rate/qwen_vl_interface': 1.7594001246782039e-06, 'learning_rate/action_model': 1.7594001246782037e-05, 'epoch': 2.43}) +[2026-06-18 00:59:13,209][starVLA.training.train_starvla][INFO] - Step 5364, Loss: {'train/loss': 0.06672710180282593, 'train/grad_norm_pre_clip': 17.835874557495117, 'timing/data': 0.0004968941211700439, 'timing/model': 0.9235724620521069, 'learning_rate/qwen_vl_interface': 1.7579337163865221e-06, 'learning_rate/action_model': 1.757933716386522e-05, 'epoch': 2.43}) +[2026-06-18 00:59:17,110][starVLA.training.train_starvla][INFO] - Step 5365, Loss: {'train/loss': 0.2883518636226654, 'train/grad_norm_pre_clip': 21.204626083374023, 'timing/data': 0.0007371474057435989, 'timing/model': 1.0022627478465438, 'learning_rate/qwen_vl_interface': 1.7564680320039356e-06, 'learning_rate/action_model': 1.7564680320039358e-05, 'epoch': 2.43}) +[2026-06-18 00:59:20,831][starVLA.training.train_starvla][INFO] - Step 5366, Loss: {'train/loss': 0.0390266515314579, 'train/grad_norm_pre_clip': 7.257471084594727, 'timing/data': 0.0005046380683779716, 'timing/model': 1.0819623321294785, 'learning_rate/qwen_vl_interface': 1.7550030718342867e-06, 'learning_rate/action_model': 1.7550030718342867e-05, 'epoch': 2.43}) +[2026-06-18 00:59:24,685][starVLA.training.train_starvla][INFO] - Step 5367, Loss: {'train/loss': 0.18085643649101257, 'train/grad_norm_pre_clip': 45.83486557006836, 'timing/data': 0.000567547045648098, 'timing/model': 1.0403414638713002, 'learning_rate/qwen_vl_interface': 1.753538836181262e-06, 'learning_rate/action_model': 1.753538836181262e-05, 'epoch': 2.43}) +[2026-06-18 00:59:28,246][starVLA.training.train_starvla][INFO] - Step 5368, Loss: {'train/loss': 0.29095086455345154, 'train/grad_norm_pre_clip': 30.583585739135742, 'timing/data': 0.0005435990169644356, 'timing/model': 0.9205219158902764, 'learning_rate/qwen_vl_interface': 1.7520753253483987e-06, 'learning_rate/action_model': 1.7520753253483986e-05, 'epoch': 2.44}) +[2026-06-18 00:59:31,953][starVLA.training.train_starvla][INFO] - Step 5369, Loss: {'train/loss': 0.06051180511713028, 'train/grad_norm_pre_clip': 10.911526679992676, 'timing/data': 0.0006938306614756584, 'timing/model': 0.9251101473346353, 'learning_rate/qwen_vl_interface': 1.7506125396390847e-06, 'learning_rate/action_model': 1.7506125396390847e-05, 'epoch': 2.44}) +[2026-06-18 00:59:35,665][starVLA.training.train_starvla][INFO] - Step 5370, Loss: {'train/loss': 0.233555868268013, 'train/grad_norm_pre_clip': 22.879783630371094, 'timing/data': 0.0004833787679672241, 'timing/model': 0.9178312085568905, 'learning_rate/qwen_vl_interface': 1.7491504793565564e-06, 'learning_rate/action_model': 1.7491504793565562e-05, 'epoch': 2.44}) +[2026-06-18 00:59:39,393][starVLA.training.train_starvla][INFO] - Step 5371, Loss: {'train/loss': 0.0617382675409317, 'train/grad_norm_pre_clip': 16.84516143798828, 'timing/data': 0.0005297716706991196, 'timing/model': 0.9266374688595533, 'learning_rate/qwen_vl_interface': 1.7476891448039012e-06, 'learning_rate/action_model': 1.7476891448039013e-05, 'epoch': 2.44}) +[2026-06-18 00:59:43,146][starVLA.training.train_starvla][INFO] - Step 5372, Loss: {'train/loss': 0.1094742864370346, 'train/grad_norm_pre_clip': 21.205663681030273, 'timing/data': 0.0014617014676332474, 'timing/model': 0.9204986784607172, 'learning_rate/qwen_vl_interface': 1.7462285362840534e-06, 'learning_rate/action_model': 1.746228536284053e-05, 'epoch': 2.44}) +[2026-06-18 00:59:46,871][starVLA.training.train_starvla][INFO] - Step 5373, Loss: {'train/loss': 0.03865407034754753, 'train/grad_norm_pre_clip': 11.129362106323242, 'timing/data': 0.0004886602982878685, 'timing/model': 1.0795474387705326, 'learning_rate/qwen_vl_interface': 1.7447686540998027e-06, 'learning_rate/action_model': 1.7447686540998026e-05, 'epoch': 2.44}) +[2026-06-18 00:59:50,415][starVLA.training.train_starvla][INFO] - Step 5374, Loss: {'train/loss': 0.377239853143692, 'train/grad_norm_pre_clip': 27.656808853149414, 'timing/data': 0.0005006641149520874, 'timing/model': 0.9220067616552114, 'learning_rate/qwen_vl_interface': 1.743309498553781e-06, 'learning_rate/action_model': 1.743309498553781e-05, 'epoch': 2.44}) +[2026-06-18 00:59:54,188][starVLA.training.train_starvla][INFO] - Step 5375, Loss: {'train/loss': 0.079762302339077, 'train/grad_norm_pre_clip': 18.615585327148438, 'timing/data': 0.0005353903397917747, 'timing/model': 0.9351939465850592, 'learning_rate/qwen_vl_interface': 1.7418510699484728e-06, 'learning_rate/action_model': 1.7418510699484727e-05, 'epoch': 2.44}) +[2026-06-18 00:59:57,866][starVLA.training.train_starvla][INFO] - Step 5376, Loss: {'train/loss': 0.07000737637281418, 'train/grad_norm_pre_clip': 11.131107330322266, 'timing/data': 0.00048435479402542114, 'timing/model': 0.9195371149107814, 'learning_rate/qwen_vl_interface': 1.7403933685862118e-06, 'learning_rate/action_model': 1.7403933685862118e-05, 'epoch': 2.44}) +[2026-06-18 01:00:01,788][starVLA.training.train_starvla][INFO] - Step 5377, Loss: {'train/loss': 0.08781344443559647, 'train/grad_norm_pre_clip': 20.209901809692383, 'timing/data': 0.0008637700229883194, 'timing/model': 0.9284507799893618, 'learning_rate/qwen_vl_interface': 1.7389363947691837e-06, 'learning_rate/action_model': 1.7389363947691838e-05, 'epoch': 2.44}) +[2026-06-18 01:00:05,805][starVLA.training.train_starvla][INFO] - Step 5378, Loss: {'train/loss': 0.28669115900993347, 'train/grad_norm_pre_clip': 53.09508514404297, 'timing/data': 0.0014801155775785446, 'timing/model': 0.9249487183988094, 'learning_rate/qwen_vl_interface': 1.7374801487994198e-06, 'learning_rate/action_model': 1.7374801487994198e-05, 'epoch': 2.44}) +[2026-06-18 01:00:09,746][starVLA.training.train_starvla][INFO] - Step 5379, Loss: {'train/loss': 0.06773820519447327, 'train/grad_norm_pre_clip': 10.626093864440918, 'timing/data': 0.0007279962301254272, 'timing/model': 1.2892564553767443, 'learning_rate/qwen_vl_interface': 1.7360246309787982e-06, 'learning_rate/action_model': 1.736024630978798e-05, 'epoch': 2.44}) +[2026-06-18 01:00:13,575][starVLA.training.train_starvla][INFO] - Step 5380, Loss: {'train/loss': 0.19124369323253632, 'train/grad_norm_pre_clip': 25.869070053100586, 'timing/data': 0.0005214782431721687, 'timing/model': 1.144112702459097, 'learning_rate/qwen_vl_interface': 1.7345698416090534e-06, 'learning_rate/action_model': 1.7345698416090533e-05, 'epoch': 2.44}) +[2026-06-18 01:00:17,153][starVLA.training.train_starvla][INFO] - Step 5381, Loss: {'train/loss': 0.21818387508392334, 'train/grad_norm_pre_clip': 28.776708602905273, 'timing/data': 0.0005210861563682556, 'timing/model': 0.9202072061598301, 'learning_rate/qwen_vl_interface': 1.7331157809917632e-06, 'learning_rate/action_model': 1.733115780991763e-05, 'epoch': 2.44}) +[2026-06-18 01:00:20,917][starVLA.training.train_starvla][INFO] - Step 5382, Loss: {'train/loss': 0.1494034081697464, 'train/grad_norm_pre_clip': 21.181747436523438, 'timing/data': 0.0004991507157683372, 'timing/model': 0.9235370401293039, 'learning_rate/qwen_vl_interface': 1.7316624494283557e-06, 'learning_rate/action_model': 1.7316624494283557e-05, 'epoch': 2.44}) +[2026-06-18 01:00:24,653][starVLA.training.train_starvla][INFO] - Step 5383, Loss: {'train/loss': 0.11483372002840042, 'train/grad_norm_pre_clip': 18.42278480529785, 'timing/data': 0.0005161333829164505, 'timing/model': 0.9243851844221354, 'learning_rate/qwen_vl_interface': 1.7302098472201085e-06, 'learning_rate/action_model': 1.7302098472201084e-05, 'epoch': 2.44}) +[2026-06-18 01:00:28,473][starVLA.training.train_starvla][INFO] - Step 5384, Loss: {'train/loss': 0.10040034353733063, 'train/grad_norm_pre_clip': 15.82456111907959, 'timing/data': 0.0005184877663850784, 'timing/model': 0.9190030600875616, 'learning_rate/qwen_vl_interface': 1.728757974668147e-06, 'learning_rate/action_model': 1.728757974668147e-05, 'epoch': 2.44}) +[2026-06-18 01:00:32,206][starVLA.training.train_starvla][INFO] - Step 5385, Loss: {'train/loss': 0.7287943363189697, 'train/grad_norm_pre_clip': 116.88522338867188, 'timing/data': 0.00048580486327409744, 'timing/model': 0.9183599436655641, 'learning_rate/qwen_vl_interface': 1.727306832073447e-06, 'learning_rate/action_model': 1.7273068320734472e-05, 'epoch': 2.44}) +[2026-06-18 01:00:35,934][starVLA.training.train_starvla][INFO] - Step 5386, Loss: {'train/loss': 0.2590031027793884, 'train/grad_norm_pre_clip': 27.704574584960938, 'timing/data': 0.0008174562826752663, 'timing/model': 1.0874539855867624, 'learning_rate/qwen_vl_interface': 1.7258564197368308e-06, 'learning_rate/action_model': 1.725856419736831e-05, 'epoch': 2.44}) +[2026-06-18 01:00:39,666][starVLA.training.train_starvla][INFO] - Step 5387, Loss: {'train/loss': 0.19285646080970764, 'train/grad_norm_pre_clip': 26.3537654876709, 'timing/data': 0.0004875706508755684, 'timing/model': 1.0885234558954835, 'learning_rate/qwen_vl_interface': 1.7244067379589708e-06, 'learning_rate/action_model': 1.7244067379589707e-05, 'epoch': 2.44}) +[2026-06-18 01:00:43,214][starVLA.training.train_starvla][INFO] - Step 5388, Loss: {'train/loss': 0.040886882692575455, 'train/grad_norm_pre_clip': 8.345138549804688, 'timing/data': 0.0005071777850389481, 'timing/model': 0.9208822008222342, 'learning_rate/qwen_vl_interface': 1.7229577870403886e-06, 'learning_rate/action_model': 1.7229577870403886e-05, 'epoch': 2.44}) +[2026-06-18 01:00:46,970][starVLA.training.train_starvla][INFO] - Step 5389, Loss: {'train/loss': 0.142253115773201, 'train/grad_norm_pre_clip': 36.81510543823242, 'timing/data': 0.0005322480574250221, 'timing/model': 0.936150997877121, 'learning_rate/qwen_vl_interface': 1.7215095672814517e-06, 'learning_rate/action_model': 1.7215095672814516e-05, 'epoch': 2.44}) +[2026-06-18 01:00:50,752][starVLA.training.train_starvla][INFO] - Step 5390, Loss: {'train/loss': 0.35562196373939514, 'train/grad_norm_pre_clip': 28.986228942871094, 'timing/data': 0.0009899288415908813, 'timing/model': 1.0291005000472069, 'learning_rate/qwen_vl_interface': 1.7200620789823769e-06, 'learning_rate/action_model': 1.7200620789823767e-05, 'epoch': 2.45}) +[2026-06-18 01:00:54,486][starVLA.training.train_starvla][INFO] - Step 5391, Loss: {'train/loss': 0.06955987215042114, 'train/grad_norm_pre_clip': 17.32213592529297, 'timing/data': 0.00049605593085289, 'timing/model': 0.9399314867332578, 'learning_rate/qwen_vl_interface': 1.7186153224432331e-06, 'learning_rate/action_model': 1.718615322443233e-05, 'epoch': 2.45}) +[2026-06-18 01:00:58,196][starVLA.training.train_starvla][INFO] - Step 5392, Loss: {'train/loss': 0.13779382407665253, 'train/grad_norm_pre_clip': 27.585472106933594, 'timing/data': 0.0009915679693222046, 'timing/model': 0.9180730832740664, 'learning_rate/qwen_vl_interface': 1.7171692979639342e-06, 'learning_rate/action_model': 1.717169297963934e-05, 'epoch': 2.45}) +[2026-06-18 01:01:01,947][starVLA.training.train_starvla][INFO] - Step 5393, Loss: {'train/loss': 0.1449759155511856, 'train/grad_norm_pre_clip': 16.787521362304688, 'timing/data': 0.0005001630634069443, 'timing/model': 1.112659621052444, 'learning_rate/qwen_vl_interface': 1.7157240058442397e-06, 'learning_rate/action_model': 1.71572400584424e-05, 'epoch': 2.45}) +[2026-06-18 01:01:05,922][starVLA.training.train_starvla][INFO] - Step 5394, Loss: {'train/loss': 0.1371689885854721, 'train/grad_norm_pre_clip': 27.75925636291504, 'timing/data': 0.0004869457334280014, 'timing/model': 1.0873804315924644, 'learning_rate/qwen_vl_interface': 1.7142794463837608e-06, 'learning_rate/action_model': 1.714279446383761e-05, 'epoch': 2.45}) +[2026-06-18 01:01:09,558][starVLA.training.train_starvla][INFO] - Step 5395, Loss: {'train/loss': 0.21398474276065826, 'train/grad_norm_pre_clip': 23.623416900634766, 'timing/data': 0.0004912866279482841, 'timing/model': 0.9790489925071597, 'learning_rate/qwen_vl_interface': 1.712835619881959e-06, 'learning_rate/action_model': 1.7128356198819588e-05, 'epoch': 2.45}) +[2026-06-18 01:01:13,351][starVLA.training.train_starvla][INFO] - Step 5396, Loss: {'train/loss': 0.0776205062866211, 'train/grad_norm_pre_clip': 28.688329696655273, 'timing/data': 0.00048434827476739883, 'timing/model': 0.9196903984993696, 'learning_rate/qwen_vl_interface': 1.7113925266381383e-06, 'learning_rate/action_model': 1.7113925266381383e-05, 'epoch': 2.45}) +[2026-06-18 01:01:17,168][starVLA.training.train_starvla][INFO] - Step 5397, Loss: {'train/loss': 0.08069328218698502, 'train/grad_norm_pre_clip': 22.652036666870117, 'timing/data': 0.0011435700580477715, 'timing/model': 0.9160649972036481, 'learning_rate/qwen_vl_interface': 1.7099501669514569e-06, 'learning_rate/action_model': 1.7099501669514568e-05, 'epoch': 2.45}) +[2026-06-18 01:01:20,948][starVLA.training.train_starvla][INFO] - Step 5398, Loss: {'train/loss': 0.268749475479126, 'train/grad_norm_pre_clip': 38.46670150756836, 'timing/data': 0.0007205633446574211, 'timing/model': 0.9951501907780766, 'learning_rate/qwen_vl_interface': 1.7085085411209118e-06, 'learning_rate/action_model': 1.7085085411209118e-05, 'epoch': 2.45}) +[2026-06-18 01:01:24,721][starVLA.training.train_starvla][INFO] - Step 5399, Loss: {'train/loss': 0.25935712456703186, 'train/grad_norm_pre_clip': 29.83963394165039, 'timing/data': 0.0004861457273364067, 'timing/model': 0.9309959262609482, 'learning_rate/qwen_vl_interface': 1.7070676494453587e-06, 'learning_rate/action_model': 1.7070676494453586e-05, 'epoch': 2.45}) +[2026-06-18 01:03:15,007][starVLA.training.train_starvla][INFO] - Step 5400, Loss: {'train/loss': 0.07447182387113571, 'train/grad_norm_pre_clip': 18.65127944946289, 'eval/loss': 0.27185627818107605, 'eval/latency_8/loss': 0.27185627818107605, 'eval/demon_fire_precision': 0.9659367396593674, 'eval/demon_fire_recall': 0.9866418142280211, 'eval/demon_fire_f1': 0.9761794990010758, 'eval/demon_fire_rate_model': 0.9148580968280468, 'eval/demon_fire_rate_teacher': 0.8956594323873122, 'eval/demon_left_precision': 0.905, 'eval/demon_left_recall': 0.7751605995717344, 'eval/demon_left_f1': 0.8350634371395618, 'eval/demon_left_rate_model': 0.11129660545353366, 'eval/demon_left_rate_teacher': 0.12993878686700056, 'eval/demon_right_precision': 0.8839427662957074, 'eval/demon_right_recall': 0.8424242424242424, 'eval/demon_right_f1': 0.8626842513576416, 'eval/demon_right_rate_model': 0.1750139120756817, 'eval/demon_right_rate_teacher': 0.18363939899833054, 'eval/demon_move_f1': 0.8488738442486017, 'eval/demon_cc_f1': 0.9341510925605094, 'eval/latency_8/demon_fire_precision': 0.9659367396593674, 'eval/latency_8/demon_fire_recall': 0.9866418142280211, 'eval/latency_8/demon_fire_f1': 0.9761794990010758, 'eval/latency_8/demon_fire_rate_model': 0.9148580968280468, 'eval/latency_8/demon_fire_rate_teacher': 0.8956594323873122, 'eval/latency_8/demon_left_precision': 0.905, 'eval/latency_8/demon_left_recall': 0.7751605995717344, 'eval/latency_8/demon_left_f1': 0.8350634371395618, 'eval/latency_8/demon_left_rate_model': 0.11129660545353366, 'eval/latency_8/demon_left_rate_teacher': 0.12993878686700056, 'eval/latency_8/demon_right_precision': 0.8839427662957074, 'eval/latency_8/demon_right_recall': 0.8424242424242424, 'eval/latency_8/demon_right_f1': 0.8626842513576416, 'eval/latency_8/demon_right_rate_model': 0.1750139120756817, 'eval/latency_8/demon_right_rate_teacher': 0.18363939899833054, 'eval/latency_8/demon_move_f1': 0.8488738442486017, 'eval/latency_8/demon_cc_f1': 0.9341510925605094, 'timing/data': 0.0005095675587654114, 'timing/model': 1.0951048405840993, 'learning_rate/qwen_vl_interface': 1.705627492223494e-06, 'learning_rate/action_model': 1.705627492223494e-05, 'epoch': 2.45}) +[2026-06-18 01:03:15,020][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_5400_state +[2026-06-18 01:03:15,025][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 01:03:21,602][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_5400_state/model.safetensors +[2026-06-18 01:03:37,771][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_5400_state/optimizer.bin +[2026-06-18 01:03:37,772][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_5400_state/sampler.bin +[2026-06-18 01:03:37,772][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_5400_state/sampler_1.bin +[2026-06-18 01:03:37,794][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_5400_state/random_states_0.pkl +[2026-06-18 01:03:37,798][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 5400 +[2026-06-18 01:03:37,799][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 01:03:37,811][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 01:03:42,114][starVLA.training.train_starvla][INFO] - Step 5401, Loss: {'train/loss': 0.085041344165802, 'train/grad_norm_pre_clip': 25.784549713134766, 'timing/data': 0.0008561974391341209, 'timing/model': 0.9214526163414121, 'learning_rate/qwen_vl_interface': 1.7041880697538644e-06, 'learning_rate/action_model': 1.704188069753864e-05, 'epoch': 2.45}) +[2026-06-18 01:03:45,873][starVLA.training.train_starvla][INFO] - Step 5402, Loss: {'train/loss': 0.39694392681121826, 'train/grad_norm_pre_clip': 34.71554183959961, 'timing/data': 0.0005556186661124229, 'timing/model': 1.1349879167973995, 'learning_rate/qwen_vl_interface': 1.702749382334863e-06, 'learning_rate/action_model': 1.7027493823348632e-05, 'epoch': 2.45}) +[2026-06-18 01:03:49,497][starVLA.training.train_starvla][INFO] - Step 5403, Loss: {'train/loss': 0.39211416244506836, 'train/grad_norm_pre_clip': 37.07783508300781, 'timing/data': 0.0004826616495847702, 'timing/model': 0.9180587269365788, 'learning_rate/qwen_vl_interface': 1.7013114302647318e-06, 'learning_rate/action_model': 1.7013114302647318e-05, 'epoch': 2.45}) +[2026-06-18 01:03:53,303][starVLA.training.train_starvla][INFO] - Step 5404, Loss: {'train/loss': 0.5704151391983032, 'train/grad_norm_pre_clip': 41.16889953613281, 'timing/data': 0.0004957150667905807, 'timing/model': 0.9216511305421591, 'learning_rate/qwen_vl_interface': 1.69987421384156e-06, 'learning_rate/action_model': 1.69987421384156e-05, 'epoch': 2.45}) +[2026-06-18 01:03:57,024][starVLA.training.train_starvla][INFO] - Step 5405, Loss: {'train/loss': 0.4702899754047394, 'train/grad_norm_pre_clip': 52.405029296875, 'timing/data': 0.0004989141598343849, 'timing/model': 0.9212070209905505, 'learning_rate/qwen_vl_interface': 1.6984377333632815e-06, 'learning_rate/action_model': 1.6984377333632813e-05, 'epoch': 2.45}) +[2026-06-18 01:04:00,748][starVLA.training.train_starvla][INFO] - Step 5406, Loss: {'train/loss': 0.1178143098950386, 'train/grad_norm_pre_clip': 19.203140258789062, 'timing/data': 0.000509507954120636, 'timing/model': 0.9183100992813706, 'learning_rate/qwen_vl_interface': 1.697001989127686e-06, 'learning_rate/action_model': 1.697001989127686e-05, 'epoch': 2.45}) +[2026-06-18 01:04:04,476][starVLA.training.train_starvla][INFO] - Step 5407, Loss: {'train/loss': 0.11186999082565308, 'train/grad_norm_pre_clip': 118.90750122070312, 'timing/data': 0.0004941895604133606, 'timing/model': 0.925106106325984, 'learning_rate/qwen_vl_interface': 1.6955669814324003e-06, 'learning_rate/action_model': 1.6955669814324e-05, 'epoch': 2.45}) +[2026-06-18 01:04:08,215][starVLA.training.train_starvla][INFO] - Step 5408, Loss: {'train/loss': 0.05170521140098572, 'train/grad_norm_pre_clip': 10.521032333374023, 'timing/data': 0.000781363807618618, 'timing/model': 1.0906935511156917, 'learning_rate/qwen_vl_interface': 1.6941327105749034e-06, 'learning_rate/action_model': 1.694132710574903e-05, 'epoch': 2.45}) +[2026-06-18 01:04:11,901][starVLA.training.train_starvla][INFO] - Step 5409, Loss: {'train/loss': 0.15635086596012115, 'train/grad_norm_pre_clip': 42.79851150512695, 'timing/data': 0.00048086978495121, 'timing/model': 1.034446799196303, 'learning_rate/qwen_vl_interface': 1.6926991768525205e-06, 'learning_rate/action_model': 1.6926991768525205e-05, 'epoch': 2.45}) +[2026-06-18 01:04:15,471][starVLA.training.train_starvla][INFO] - Step 5410, Loss: {'train/loss': 0.2177669107913971, 'train/grad_norm_pre_clip': 29.82047462463379, 'timing/data': 0.0004868796095252037, 'timing/model': 0.9157352466136217, 'learning_rate/qwen_vl_interface': 1.6912663805624278e-06, 'learning_rate/action_model': 1.6912663805624275e-05, 'epoch': 2.45}) +[2026-06-18 01:04:19,665][starVLA.training.train_starvla][INFO] - Step 5411, Loss: {'train/loss': 0.1659490317106247, 'train/grad_norm_pre_clip': 25.026216506958008, 'timing/data': 0.00047705601900815964, 'timing/model': 1.225488705560565, 'learning_rate/qwen_vl_interface': 1.6898343220016446e-06, 'learning_rate/action_model': 1.6898343220016447e-05, 'epoch': 2.45}) +[2026-06-18 01:04:23,682][starVLA.training.train_starvla][INFO] - Step 5412, Loss: {'train/loss': 0.12633967399597168, 'train/grad_norm_pre_clip': 16.14610481262207, 'timing/data': 0.0004794420674443245, 'timing/model': 0.9397974470630288, 'learning_rate/qwen_vl_interface': 1.6884030014670353e-06, 'learning_rate/action_model': 1.688403001467035e-05, 'epoch': 2.46}) +[2026-06-18 01:04:27,399][starVLA.training.train_starvla][INFO] - Step 5413, Loss: {'train/loss': 0.6300186514854431, 'train/grad_norm_pre_clip': 32.91868591308594, 'timing/data': 0.0004355683922767639, 'timing/model': 0.9234058372676373, 'learning_rate/qwen_vl_interface': 1.6869724192553168e-06, 'learning_rate/action_model': 1.686972419255317e-05, 'epoch': 2.46}) +[2026-06-18 01:04:31,145][starVLA.training.train_starvla][INFO] - Step 5414, Loss: {'train/loss': 0.2851725220680237, 'train/grad_norm_pre_clip': 25.988224029541016, 'timing/data': 0.00047951750457286835, 'timing/model': 0.9315000688657165, 'learning_rate/qwen_vl_interface': 1.6855425756630506e-06, 'learning_rate/action_model': 1.6855425756630506e-05, 'epoch': 2.46}) +[2026-06-18 01:04:34,977][starVLA.training.train_starvla][INFO] - Step 5415, Loss: {'train/loss': 0.06903956830501556, 'train/grad_norm_pre_clip': 14.197932243347168, 'timing/data': 0.000478135421872139, 'timing/model': 1.1135682025924325, 'learning_rate/qwen_vl_interface': 1.684113470986644e-06, 'learning_rate/action_model': 1.6841134709866438e-05, 'epoch': 2.46}) +[2026-06-18 01:04:38,547][starVLA.training.train_starvla][INFO] - Step 5416, Loss: {'train/loss': 0.5729576349258423, 'train/grad_norm_pre_clip': 51.35447311401367, 'timing/data': 0.0006558224558830261, 'timing/model': 0.9180732937529683, 'learning_rate/qwen_vl_interface': 1.682685105522351e-06, 'learning_rate/action_model': 1.682685105522351e-05, 'epoch': 2.46}) +[2026-06-18 01:04:42,289][starVLA.training.train_starvla][INFO] - Step 5417, Loss: {'train/loss': 0.24723246693611145, 'train/grad_norm_pre_clip': 35.14087677001953, 'timing/data': 0.0004938794299960136, 'timing/model': 0.9181184535846114, 'learning_rate/qwen_vl_interface': 1.6812574795662751e-06, 'learning_rate/action_model': 1.681257479566275e-05, 'epoch': 2.46}) +[2026-06-18 01:04:46,103][starVLA.training.train_starvla][INFO] - Step 5418, Loss: {'train/loss': 0.14365170896053314, 'train/grad_norm_pre_clip': 20.704837799072266, 'timing/data': 0.0004680277779698372, 'timing/model': 0.951468656770885, 'learning_rate/qwen_vl_interface': 1.6798305934143635e-06, 'learning_rate/action_model': 1.6798305934143633e-05, 'epoch': 2.46}) +[2026-06-18 01:04:49,937][starVLA.training.train_starvla][INFO] - Step 5419, Loss: {'train/loss': 0.11201982945203781, 'train/grad_norm_pre_clip': 19.084468841552734, 'timing/data': 0.001146465539932251, 'timing/model': 0.9213877040892839, 'learning_rate/qwen_vl_interface': 1.6784044473624118e-06, 'learning_rate/action_model': 1.6784044473624118e-05, 'epoch': 2.46}) +[2026-06-18 01:04:53,675][starVLA.training.train_starvla][INFO] - Step 5420, Loss: {'train/loss': 0.2239447385072708, 'train/grad_norm_pre_clip': 31.667795181274414, 'timing/data': 0.0008458858355879784, 'timing/model': 0.9184863902628422, 'learning_rate/qwen_vl_interface': 1.6769790417060611e-06, 'learning_rate/action_model': 1.676979041706061e-05, 'epoch': 2.46}) +[2026-06-18 01:04:57,402][starVLA.training.train_starvla][INFO] - Step 5421, Loss: {'train/loss': 0.29834309220314026, 'train/grad_norm_pre_clip': 33.77755355834961, 'timing/data': 0.000875343568623066, 'timing/model': 0.9180648820474744, 'learning_rate/qwen_vl_interface': 1.6755543767407998e-06, 'learning_rate/action_model': 1.6755543767407998e-05, 'epoch': 2.46}) +[2026-06-18 01:05:01,110][starVLA.training.train_starvla][INFO] - Step 5422, Loss: {'train/loss': 0.5298110842704773, 'train/grad_norm_pre_clip': 32.16709899902344, 'timing/data': 0.0004984373226761818, 'timing/model': 1.0855736266821623, 'learning_rate/qwen_vl_interface': 1.6741304527619624e-06, 'learning_rate/action_model': 1.674130452761962e-05, 'epoch': 2.46}) +[2026-06-18 01:05:04,692][starVLA.training.train_starvla][INFO] - Step 5423, Loss: {'train/loss': 0.21296323835849762, 'train/grad_norm_pre_clip': 33.488670349121094, 'timing/data': 0.0004875427111983299, 'timing/model': 0.9314649533480406, 'learning_rate/qwen_vl_interface': 1.672707270064729e-06, 'learning_rate/action_model': 1.6727072700647288e-05, 'epoch': 2.46}) +[2026-06-18 01:05:08,544][starVLA.training.train_starvla][INFO] - Step 5424, Loss: {'train/loss': 0.11535272002220154, 'train/grad_norm_pre_clip': 23.58670425415039, 'timing/data': 0.0004903329536318779, 'timing/model': 0.9245409443974495, 'learning_rate/qwen_vl_interface': 1.6712848289441288e-06, 'learning_rate/action_model': 1.6712848289441287e-05, 'epoch': 2.46}) +[2026-06-18 01:05:12,610][starVLA.training.train_starvla][INFO] - Step 5425, Loss: {'train/loss': 0.12098202109336853, 'train/grad_norm_pre_clip': 22.782678604125977, 'timing/data': 0.0005028285086154938, 'timing/model': 0.9219528008252382, 'learning_rate/qwen_vl_interface': 1.669863129695036e-06, 'learning_rate/action_model': 1.669863129695036e-05, 'epoch': 2.46}) +[2026-06-18 01:05:16,381][starVLA.training.train_starvla][INFO] - Step 5426, Loss: {'train/loss': 0.05367013439536095, 'train/grad_norm_pre_clip': 14.135875701904297, 'timing/data': 0.0009012464433908463, 'timing/model': 0.9519101027399302, 'learning_rate/qwen_vl_interface': 1.6684421726121675e-06, 'learning_rate/action_model': 1.6684421726121674e-05, 'epoch': 2.46}) +[2026-06-18 01:05:20,103][starVLA.training.train_starvla][INFO] - Step 5427, Loss: {'train/loss': 0.3061923086643219, 'train/grad_norm_pre_clip': 28.708223342895508, 'timing/data': 0.0005363449454307556, 'timing/model': 0.9288111748173833, 'learning_rate/qwen_vl_interface': 1.6670219579900888e-06, 'learning_rate/action_model': 1.6670219579900887e-05, 'epoch': 2.46}) +[2026-06-18 01:05:24,002][starVLA.training.train_starvla][INFO] - Step 5428, Loss: {'train/loss': 0.5310440063476562, 'train/grad_norm_pre_clip': 40.02519989013672, 'timing/data': 0.0005103806033730507, 'timing/model': 1.236111811362207, 'learning_rate/qwen_vl_interface': 1.665602486123215e-06, 'learning_rate/action_model': 1.665602486123215e-05, 'epoch': 2.46}) +[2026-06-18 01:05:27,724][starVLA.training.train_starvla][INFO] - Step 5429, Loss: {'train/loss': 0.14988675713539124, 'train/grad_norm_pre_clip': 33.40593338012695, 'timing/data': 0.0004870779812335968, 'timing/model': 1.069546576589346, 'learning_rate/qwen_vl_interface': 1.6641837573058034e-06, 'learning_rate/action_model': 1.6641837573058035e-05, 'epoch': 2.46}) +[2026-06-18 01:05:31,344][starVLA.training.train_starvla][INFO] - Step 5430, Loss: {'train/loss': 0.25152817368507385, 'train/grad_norm_pre_clip': 27.384307861328125, 'timing/data': 0.0005604419857263565, 'timing/model': 0.926482998766005, 'learning_rate/qwen_vl_interface': 1.6627657718319568e-06, 'learning_rate/action_model': 1.6627657718319568e-05, 'epoch': 2.46}) +[2026-06-18 01:05:35,112][starVLA.training.train_starvla][INFO] - Step 5431, Loss: {'train/loss': 0.38082408905029297, 'train/grad_norm_pre_clip': 27.407901763916016, 'timing/data': 0.0005081184208393097, 'timing/model': 0.9153408091515303, 'learning_rate/qwen_vl_interface': 1.6613485299956265e-06, 'learning_rate/action_model': 1.6613485299956262e-05, 'epoch': 2.46}) +[2026-06-18 01:05:38,791][starVLA.training.train_starvla][INFO] - Step 5432, Loss: {'train/loss': 0.09938742220401764, 'train/grad_norm_pre_clip': 23.903331756591797, 'timing/data': 0.000809001736342907, 'timing/model': 0.9212266383692622, 'learning_rate/qwen_vl_interface': 1.6599320320906064e-06, 'learning_rate/action_model': 1.6599320320906066e-05, 'epoch': 2.46}) +[2026-06-18 01:05:42,504][starVLA.training.train_starvla][INFO] - Step 5433, Loss: {'train/loss': 0.507915198802948, 'train/grad_norm_pre_clip': 36.36415100097656, 'timing/data': 0.0004910510033369064, 'timing/model': 0.9352558152750134, 'learning_rate/qwen_vl_interface': 1.65851627841054e-06, 'learning_rate/action_model': 1.65851627841054e-05, 'epoch': 2.46}) +[2026-06-18 01:05:46,312][starVLA.training.train_starvla][INFO] - Step 5434, Loss: {'train/loss': 0.1446378082036972, 'train/grad_norm_pre_clip': 24.68893814086914, 'timing/data': 0.0005134753882884979, 'timing/model': 0.9612974235787988, 'learning_rate/qwen_vl_interface': 1.657101269248913e-06, 'learning_rate/action_model': 1.657101269248913e-05, 'epoch': 2.47}) +[2026-06-18 01:05:50,120][starVLA.training.train_starvla][INFO] - Step 5435, Loss: {'train/loss': 0.4456390142440796, 'train/grad_norm_pre_clip': 38.744049072265625, 'timing/data': 0.0012061195448040962, 'timing/model': 1.1196150612086058, 'learning_rate/qwen_vl_interface': 1.65568700489906e-06, 'learning_rate/action_model': 1.65568700489906e-05, 'epoch': 2.47}) +[2026-06-18 01:05:54,148][starVLA.training.train_starvla][INFO] - Step 5436, Loss: {'train/loss': 0.5866925716400146, 'train/grad_norm_pre_clip': 44.44321823120117, 'timing/data': 0.0009831152856349945, 'timing/model': 1.1113134175539017, 'learning_rate/qwen_vl_interface': 1.6542734856541577e-06, 'learning_rate/action_model': 1.6542734856541576e-05, 'epoch': 2.47}) +[2026-06-18 01:05:57,746][starVLA.training.train_starvla][INFO] - Step 5437, Loss: {'train/loss': 0.12716498970985413, 'train/grad_norm_pre_clip': 26.518924713134766, 'timing/data': 0.0010992754250764847, 'timing/model': 0.9215714884921908, 'learning_rate/qwen_vl_interface': 1.652860711807231e-06, 'learning_rate/action_model': 1.652860711807231e-05, 'epoch': 2.47}) +[2026-06-18 01:06:01,468][starVLA.training.train_starvla][INFO] - Step 5438, Loss: {'train/loss': 0.3494514226913452, 'train/grad_norm_pre_clip': 37.06951141357422, 'timing/data': 0.0011487770825624466, 'timing/model': 0.9261369798332453, 'learning_rate/qwen_vl_interface': 1.6514486836511484e-06, 'learning_rate/action_model': 1.6514486836511484e-05, 'epoch': 2.47}) +[2026-06-18 01:06:05,245][starVLA.training.train_starvla][INFO] - Step 5439, Loss: {'train/loss': 0.30053481459617615, 'train/grad_norm_pre_clip': 30.12409019470215, 'timing/data': 0.0005078744143247604, 'timing/model': 0.9394946321845055, 'learning_rate/qwen_vl_interface': 1.6500374014786288e-06, 'learning_rate/action_model': 1.650037401478629e-05, 'epoch': 2.47}) +[2026-06-18 01:06:09,110][starVLA.training.train_starvla][INFO] - Step 5440, Loss: {'train/loss': 0.20066791772842407, 'train/grad_norm_pre_clip': 49.30868911743164, 'timing/data': 0.000519048422574997, 'timing/model': 0.9441190529614687, 'learning_rate/qwen_vl_interface': 1.6486268655822277e-06, 'learning_rate/action_model': 1.6486268655822277e-05, 'epoch': 2.47}) +[2026-06-18 01:06:13,605][starVLA.training.train_starvla][INFO] - Step 5441, Loss: {'train/loss': 0.41251325607299805, 'train/grad_norm_pre_clip': 40.06085205078125, 'timing/data': 0.0008684666827321053, 'timing/model': 1.5971667431294918, 'learning_rate/qwen_vl_interface': 1.6472170762543515e-06, 'learning_rate/action_model': 1.6472170762543514e-05, 'epoch': 2.47}) +[2026-06-18 01:06:18,436][starVLA.training.train_starvla][INFO] - Step 5442, Loss: {'train/loss': 0.0650860071182251, 'train/grad_norm_pre_clip': 12.20915412902832, 'timing/data': 0.0004696398973464966, 'timing/model': 1.1088702250272036, 'learning_rate/qwen_vl_interface': 1.6458080337872534e-06, 'learning_rate/action_model': 1.6458080337872535e-05, 'epoch': 2.47}) +[2026-06-18 01:06:23,437][starVLA.training.train_starvla][INFO] - Step 5443, Loss: {'train/loss': 0.16880744695663452, 'train/grad_norm_pre_clip': 25.32257080078125, 'timing/data': 0.0008158823475241661, 'timing/model': 0.9386429991573095, 'learning_rate/qwen_vl_interface': 1.644399738473029e-06, 'learning_rate/action_model': 1.644399738473029e-05, 'epoch': 2.47}) +[2026-06-18 01:06:27,262][starVLA.training.train_starvla][INFO] - Step 5444, Loss: {'train/loss': 0.17971214652061462, 'train/grad_norm_pre_clip': 18.845905303955078, 'timing/data': 0.0008306661620736122, 'timing/model': 0.9672611011192203, 'learning_rate/qwen_vl_interface': 1.6429921906036175e-06, 'learning_rate/action_model': 1.6429921906036176e-05, 'epoch': 2.47}) +[2026-06-18 01:06:31,761][starVLA.training.train_starvla][INFO] - Step 5445, Loss: {'train/loss': 0.3787098824977875, 'train/grad_norm_pre_clip': 48.11774444580078, 'timing/data': 0.00047580525279045105, 'timing/model': 0.9255694979801774, 'learning_rate/qwen_vl_interface': 1.6415853904708068e-06, 'learning_rate/action_model': 1.6415853904708066e-05, 'epoch': 2.47}) +[2026-06-18 01:06:35,541][starVLA.training.train_starvla][INFO] - Step 5446, Loss: {'train/loss': 0.26302820444107056, 'train/grad_norm_pre_clip': 31.842851638793945, 'timing/data': 0.0004924014210700989, 'timing/model': 0.9455657787621021, 'learning_rate/qwen_vl_interface': 1.6401793383662274e-06, 'learning_rate/action_model': 1.6401793383662273e-05, 'epoch': 2.47}) +[2026-06-18 01:06:39,281][starVLA.training.train_starvla][INFO] - Step 5447, Loss: {'train/loss': 0.3071056604385376, 'train/grad_norm_pre_clip': 30.285520553588867, 'timing/data': 0.0004714084789156914, 'timing/model': 0.9176000850275159, 'learning_rate/qwen_vl_interface': 1.6387740345813553e-06, 'learning_rate/action_model': 1.6387740345813552e-05, 'epoch': 2.47}) +[2026-06-18 01:06:43,054][starVLA.training.train_starvla][INFO] - Step 5448, Loss: {'train/loss': 0.16620707511901855, 'train/grad_norm_pre_clip': 21.95697021484375, 'timing/data': 0.00048394594341516495, 'timing/model': 1.090855911374092, 'learning_rate/qwen_vl_interface': 1.6373694794075115e-06, 'learning_rate/action_model': 1.6373694794075115e-05, 'epoch': 2.47}) +[2026-06-18 01:06:46,735][starVLA.training.train_starvla][INFO] - Step 5449, Loss: {'train/loss': 0.18910910189151764, 'train/grad_norm_pre_clip': 33.38378143310547, 'timing/data': 0.0006237942725419998, 'timing/model': 1.0227680560201406, 'learning_rate/qwen_vl_interface': 1.6359656731358626e-06, 'learning_rate/action_model': 1.6359656731358626e-05, 'epoch': 2.47}) +[2026-06-18 01:06:50,345][starVLA.training.train_starvla][INFO] - Step 5450, Loss: {'train/loss': 0.14773088693618774, 'train/grad_norm_pre_clip': 18.2408390045166, 'timing/data': 0.0010582879185676575, 'timing/model': 0.9396244660019875, 'learning_rate/qwen_vl_interface': 1.634562616057418e-06, 'learning_rate/action_model': 1.634562616057418e-05, 'epoch': 2.47}) +[2026-06-18 01:06:54,102][starVLA.training.train_starvla][INFO] - Step 5451, Loss: {'train/loss': 0.3101043999195099, 'train/grad_norm_pre_clip': 32.893856048583984, 'timing/data': 0.0004895394667983055, 'timing/model': 0.9310239069163799, 'learning_rate/qwen_vl_interface': 1.6331603084630338e-06, 'learning_rate/action_model': 1.6331603084630338e-05, 'epoch': 2.47}) +[2026-06-18 01:06:57,843][starVLA.training.train_starvla][INFO] - Step 5452, Loss: {'train/loss': 0.058326080441474915, 'train/grad_norm_pre_clip': 11.090371131896973, 'timing/data': 0.0009932760149240494, 'timing/model': 0.9245881605893373, 'learning_rate/qwen_vl_interface': 1.6317587506434076e-06, 'learning_rate/action_model': 1.6317587506434076e-05, 'epoch': 2.47}) +[2026-06-18 01:07:01,623][starVLA.training.train_starvla][INFO] - Step 5453, Loss: {'train/loss': 0.4391076862812042, 'train/grad_norm_pre_clip': 48.338966369628906, 'timing/data': 0.001238376833498478, 'timing/model': 0.9134848853573203, 'learning_rate/qwen_vl_interface': 1.6303579428890881e-06, 'learning_rate/action_model': 1.630357942889088e-05, 'epoch': 2.47}) +[2026-06-18 01:07:05,474][starVLA.training.train_starvla][INFO] - Step 5454, Loss: {'train/loss': 0.10518015921115875, 'train/grad_norm_pre_clip': 15.743062019348145, 'timing/data': 0.0004743989557027817, 'timing/model': 0.9355038469657302, 'learning_rate/qwen_vl_interface': 1.6289578854904626e-06, 'learning_rate/action_model': 1.6289578854904625e-05, 'epoch': 2.47}) +[2026-06-18 01:07:09,449][starVLA.training.train_starvla][INFO] - Step 5455, Loss: {'train/loss': 0.07989640533924103, 'train/grad_norm_pre_clip': 13.953812599182129, 'timing/data': 0.0005488432943820953, 'timing/model': 1.0885297115892172, 'learning_rate/qwen_vl_interface': 1.6275585787377627e-06, 'learning_rate/action_model': 1.6275585787377626e-05, 'epoch': 2.47}) +[2026-06-18 01:07:13,018][starVLA.training.train_starvla][INFO] - Step 5456, Loss: {'train/loss': 0.10877104103565216, 'train/grad_norm_pre_clip': 15.42606258392334, 'timing/data': 0.0005489001050591469, 'timing/model': 0.9218564424663782, 'learning_rate/qwen_vl_interface': 1.6261600229210662e-06, 'learning_rate/action_model': 1.626160022921066e-05, 'epoch': 2.48}) +[2026-06-18 01:07:16,749][starVLA.training.train_starvla][INFO] - Step 5457, Loss: {'train/loss': 0.17690123617649078, 'train/grad_norm_pre_clip': 23.586048126220703, 'timing/data': 0.0005487985908985138, 'timing/model': 0.9177944101393223, 'learning_rate/qwen_vl_interface': 1.6247622183302972e-06, 'learning_rate/action_model': 1.624762218330297e-05, 'epoch': 2.48}) +[2026-06-18 01:07:20,568][starVLA.training.train_starvla][INFO] - Step 5458, Loss: {'train/loss': 0.1790200173854828, 'train/grad_norm_pre_clip': 22.739011764526367, 'timing/data': 0.00048650987446308136, 'timing/model': 0.9265410751104355, 'learning_rate/qwen_vl_interface': 1.6233651652552218e-06, 'learning_rate/action_model': 1.6233651652552217e-05, 'epoch': 2.48}) +[2026-06-18 01:07:24,561][starVLA.training.train_starvla][INFO] - Step 5459, Loss: {'train/loss': 0.1443396657705307, 'train/grad_norm_pre_clip': 22.505687713623047, 'timing/data': 0.0010383734479546547, 'timing/model': 0.9208550229668617, 'learning_rate/qwen_vl_interface': 1.6219688639854507e-06, 'learning_rate/action_model': 1.6219688639854506e-05, 'epoch': 2.48}) +[2026-06-18 01:07:28,375][starVLA.training.train_starvla][INFO] - Step 5460, Loss: {'train/loss': 0.31702134013175964, 'train/grad_norm_pre_clip': 17.3668155670166, 'timing/data': 0.0008916445076465607, 'timing/model': 0.9614394856616855, 'learning_rate/qwen_vl_interface': 1.620573314810436e-06, 'learning_rate/action_model': 1.620573314810436e-05, 'epoch': 2.48}) +[2026-06-18 01:07:32,479][starVLA.training.train_starvla][INFO] - Step 5461, Loss: {'train/loss': 0.2378978431224823, 'train/grad_norm_pre_clip': 18.576963424682617, 'timing/data': 0.0005616089329123497, 'timing/model': 1.0911863325163722, 'learning_rate/qwen_vl_interface': 1.6191785180194797e-06, 'learning_rate/action_model': 1.6191785180194796e-05, 'epoch': 2.48}) +[2026-06-18 01:07:36,232][starVLA.training.train_starvla][INFO] - Step 5462, Loss: {'train/loss': 0.09097079187631607, 'train/grad_norm_pre_clip': 21.559402465820312, 'timing/data': 0.00047963298857212067, 'timing/model': 1.0811661714687943, 'learning_rate/qwen_vl_interface': 1.6177844739017236e-06, 'learning_rate/action_model': 1.6177844739017236e-05, 'epoch': 2.48}) +[2026-06-18 01:07:39,810][starVLA.training.train_starvla][INFO] - Step 5463, Loss: {'train/loss': 0.0794299766421318, 'train/grad_norm_pre_clip': 16.306922912597656, 'timing/data': 0.0009209560230374336, 'timing/model': 0.9216578416526318, 'learning_rate/qwen_vl_interface': 1.6163911827461546e-06, 'learning_rate/action_model': 1.6163911827461546e-05, 'epoch': 2.48}) +[2026-06-18 01:07:43,601][starVLA.training.train_starvla][INFO] - Step 5464, Loss: {'train/loss': 0.12838441133499146, 'train/grad_norm_pre_clip': 20.34891700744629, 'timing/data': 0.0010253461077809334, 'timing/model': 0.9354279674589634, 'learning_rate/qwen_vl_interface': 1.6149986448416032e-06, 'learning_rate/action_model': 1.614998644841603e-05, 'epoch': 2.48}) +[2026-06-18 01:07:47,361][starVLA.training.train_starvla][INFO] - Step 5465, Loss: {'train/loss': 0.21611981093883514, 'train/grad_norm_pre_clip': 30.676618576049805, 'timing/data': 0.0013223318383097649, 'timing/model': 0.9177253637462854, 'learning_rate/qwen_vl_interface': 1.6136068604767444e-06, 'learning_rate/action_model': 1.6136068604767442e-05, 'epoch': 2.48}) +[2026-06-18 01:07:51,601][starVLA.training.train_starvla][INFO] - Step 5466, Loss: {'train/loss': 0.060840338468551636, 'train/grad_norm_pre_clip': 12.344012260437012, 'timing/data': 0.0011491170153021812, 'timing/model': 1.1568039851263165, 'learning_rate/qwen_vl_interface': 1.6122158299400962e-06, 'learning_rate/action_model': 1.612215829940096e-05, 'epoch': 2.48}) +[2026-06-18 01:07:55,340][starVLA.training.train_starvla][INFO] - Step 5467, Loss: {'train/loss': 0.4341046214103699, 'train/grad_norm_pre_clip': 48.196285247802734, 'timing/data': 0.0005018021911382675, 'timing/model': 0.9287717901170254, 'learning_rate/qwen_vl_interface': 1.6108255535200193e-06, 'learning_rate/action_model': 1.6108255535200193e-05, 'epoch': 2.48}) +[2026-06-18 01:07:59,042][starVLA.training.train_starvla][INFO] - Step 5468, Loss: {'train/loss': 0.1411241739988327, 'train/grad_norm_pre_clip': 19.98223876953125, 'timing/data': 0.0005126530304551125, 'timing/model': 1.0699512278661132, 'learning_rate/qwen_vl_interface': 1.6094360315047236e-06, 'learning_rate/action_model': 1.6094360315047237e-05, 'epoch': 2.48}) +[2026-06-18 01:08:02,810][starVLA.training.train_starvla][INFO] - Step 5469, Loss: {'train/loss': 0.3055810034275055, 'train/grad_norm_pre_clip': 30.422569274902344, 'timing/data': 0.0005968613550066948, 'timing/model': 1.1322246380150318, 'learning_rate/qwen_vl_interface': 1.608047264182255e-06, 'learning_rate/action_model': 1.608047264182255e-05, 'epoch': 2.48}) +[2026-06-18 01:08:06,403][starVLA.training.train_starvla][INFO] - Step 5470, Loss: {'train/loss': 0.13781440258026123, 'train/grad_norm_pre_clip': 13.886237144470215, 'timing/data': 0.0004983628168702126, 'timing/model': 0.9284719247370958, 'learning_rate/qwen_vl_interface': 1.6066592518405077e-06, 'learning_rate/action_model': 1.6066592518405076e-05, 'epoch': 2.48}) +[2026-06-18 01:08:10,177][starVLA.training.train_starvla][INFO] - Step 5471, Loss: {'train/loss': 0.142767533659935, 'train/grad_norm_pre_clip': 21.790616989135742, 'timing/data': 0.0004924889653921127, 'timing/model': 0.9259136226028204, 'learning_rate/qwen_vl_interface': 1.605271994767216e-06, 'learning_rate/action_model': 1.605271994767216e-05, 'epoch': 2.48}) +[2026-06-18 01:08:13,880][starVLA.training.train_starvla][INFO] - Step 5472, Loss: {'train/loss': 0.34459027647972107, 'train/grad_norm_pre_clip': 24.53020477294922, 'timing/data': 0.0005971742793917656, 'timing/model': 0.9277420714497566, 'learning_rate/qwen_vl_interface': 1.6038854932499635e-06, 'learning_rate/action_model': 1.6038854932499633e-05, 'epoch': 2.48}) +[2026-06-18 01:08:17,607][starVLA.training.train_starvla][INFO] - Step 5473, Loss: {'train/loss': 0.1287619024515152, 'train/grad_norm_pre_clip': 17.31080436706543, 'timing/data': 0.0006279507651925087, 'timing/model': 0.9223250709474087, 'learning_rate/qwen_vl_interface': 1.6024997475761727e-06, 'learning_rate/action_model': 1.602499747576173e-05, 'epoch': 2.48}) +[2026-06-18 01:08:21,347][starVLA.training.train_starvla][INFO] - Step 5474, Loss: {'train/loss': 0.03423341363668442, 'train/grad_norm_pre_clip': 7.399284362792969, 'timing/data': 0.0004930300638079643, 'timing/model': 0.920680277980864, 'learning_rate/qwen_vl_interface': 1.6011147580331065e-06, 'learning_rate/action_model': 1.6011147580331066e-05, 'epoch': 2.48}) +[2026-06-18 01:08:25,798][starVLA.training.train_starvla][INFO] - Step 5475, Loss: {'train/loss': 0.08453318476676941, 'train/grad_norm_pre_clip': 14.386114120483398, 'timing/data': 0.0005013057962059975, 'timing/model': 1.6564984684810042, 'learning_rate/qwen_vl_interface': 1.5997305249078788e-06, 'learning_rate/action_model': 1.5997305249078787e-05, 'epoch': 2.48}) +[2026-06-18 01:08:29,545][starVLA.training.train_starvla][INFO] - Step 5476, Loss: {'train/loss': 0.2784186005592346, 'train/grad_norm_pre_clip': 33.35527038574219, 'timing/data': 0.0005745477974414825, 'timing/model': 1.0890581496059895, 'learning_rate/qwen_vl_interface': 1.5983470484874412e-06, 'learning_rate/action_model': 1.598347048487441e-05, 'epoch': 2.48}) +[2026-06-18 01:08:33,102][starVLA.training.train_starvla][INFO] - Step 5477, Loss: {'train/loss': 0.06442231684923172, 'train/grad_norm_pre_clip': 17.52850914001465, 'timing/data': 0.0004936391487717628, 'timing/model': 0.9230455998331308, 'learning_rate/qwen_vl_interface': 1.5969643290585893e-06, 'learning_rate/action_model': 1.5969643290585892e-05, 'epoch': 2.48}) +[2026-06-18 01:08:36,837][starVLA.training.train_starvla][INFO] - Step 5478, Loss: {'train/loss': 0.4449390470981598, 'train/grad_norm_pre_clip': 38.67399215698242, 'timing/data': 0.0004661697894334793, 'timing/model': 0.929255548864603, 'learning_rate/qwen_vl_interface': 1.5955823669079632e-06, 'learning_rate/action_model': 1.595582366907963e-05, 'epoch': 2.49}) +[2026-06-18 01:08:41,475][starVLA.training.train_starvla][INFO] - Step 5479, Loss: {'train/loss': 0.15568549931049347, 'train/grad_norm_pre_clip': 54.543609619140625, 'timing/data': 0.0007463889196515083, 'timing/model': 0.9665009053424001, 'learning_rate/qwen_vl_interface': 1.594201162322045e-06, 'learning_rate/action_model': 1.594201162322045e-05, 'epoch': 2.49}) +[2026-06-18 01:08:45,230][starVLA.training.train_starvla][INFO] - Step 5480, Loss: {'train/loss': 0.20986559987068176, 'train/grad_norm_pre_clip': 21.51314926147461, 'timing/data': 0.0005897004157304764, 'timing/model': 0.9387840507552028, 'learning_rate/qwen_vl_interface': 1.592820715587159e-06, 'learning_rate/action_model': 1.5928207155871593e-05, 'epoch': 2.49}) +[2026-06-18 01:08:48,983][starVLA.training.train_starvla][INFO] - Step 5481, Loss: {'train/loss': 0.748677670955658, 'train/grad_norm_pre_clip': 45.84559631347656, 'timing/data': 0.000548679381608963, 'timing/model': 0.9177740467712283, 'learning_rate/qwen_vl_interface': 1.5914410269894743e-06, 'learning_rate/action_model': 1.591441026989474e-05, 'epoch': 2.49}) +[2026-06-18 01:08:52,684][starVLA.training.train_starvla][INFO] - Step 5482, Loss: {'train/loss': 0.14364300668239594, 'train/grad_norm_pre_clip': 20.692127227783203, 'timing/data': 0.0005120011046528816, 'timing/model': 1.086782188154757, 'learning_rate/qwen_vl_interface': 1.5900620968150015e-06, 'learning_rate/action_model': 1.5900620968150015e-05, 'epoch': 2.49}) +[2026-06-18 01:08:56,250][starVLA.training.train_starvla][INFO] - Step 5483, Loss: {'train/loss': 0.19929809868335724, 'train/grad_norm_pre_clip': 23.57993507385254, 'timing/data': 0.0005724458023905754, 'timing/model': 0.9283439191058278, 'learning_rate/qwen_vl_interface': 1.5886839253495944e-06, 'learning_rate/action_model': 1.5886839253495943e-05, 'epoch': 2.49}) +[2026-06-18 01:09:00,014][starVLA.training.train_starvla][INFO] - Step 5484, Loss: {'train/loss': 0.19143161177635193, 'train/grad_norm_pre_clip': 29.02454376220703, 'timing/data': 0.0008188094943761826, 'timing/model': 0.9334050295874476, 'learning_rate/qwen_vl_interface': 1.5873065128789495e-06, 'learning_rate/action_model': 1.5873065128789493e-05, 'epoch': 2.49}) +[2026-06-18 01:09:03,718][starVLA.training.train_starvla][INFO] - Step 5485, Loss: {'train/loss': 0.03166304901242256, 'train/grad_norm_pre_clip': 9.500557899475098, 'timing/data': 0.0004892293363809586, 'timing/model': 0.9229244468733668, 'learning_rate/qwen_vl_interface': 1.5859298596886038e-06, 'learning_rate/action_model': 1.585929859688604e-05, 'epoch': 2.49}) +[2026-06-18 01:09:07,757][starVLA.training.train_starvla][INFO] - Step 5486, Loss: {'train/loss': 0.1823969930410385, 'train/grad_norm_pre_clip': 25.502262115478516, 'timing/data': 0.0004810132086277008, 'timing/model': 0.9219917254522443, 'learning_rate/qwen_vl_interface': 1.5845539660639429e-06, 'learning_rate/action_model': 1.5845539660639426e-05, 'epoch': 2.49}) +[2026-06-18 01:09:11,482][starVLA.training.train_starvla][INFO] - Step 5487, Loss: {'train/loss': 0.1613241285085678, 'train/grad_norm_pre_clip': 32.18556594848633, 'timing/data': 0.0006162067875266075, 'timing/model': 0.9292501788586378, 'learning_rate/qwen_vl_interface': 1.5831788322901901e-06, 'learning_rate/action_model': 1.58317883229019e-05, 'epoch': 2.49}) +[2026-06-18 01:09:15,227][starVLA.training.train_starvla][INFO] - Step 5488, Loss: {'train/loss': 0.5335002541542053, 'train/grad_norm_pre_clip': 38.56809616088867, 'timing/data': 0.001022944226861, 'timing/model': 1.0973844463005662, 'learning_rate/qwen_vl_interface': 1.5818044586524092e-06, 'learning_rate/action_model': 1.5818044586524092e-05, 'epoch': 2.49}) +[2026-06-18 01:09:19,050][starVLA.training.train_starvla][INFO] - Step 5489, Loss: {'train/loss': 0.04848260059952736, 'train/grad_norm_pre_clip': 8.26887321472168, 'timing/data': 0.0008806567639112473, 'timing/model': 1.0072173699736595, 'learning_rate/qwen_vl_interface': 1.5804308454355102e-06, 'learning_rate/action_model': 1.5804308454355102e-05, 'epoch': 2.49}) +[2026-06-18 01:09:22,624][starVLA.training.train_starvla][INFO] - Step 5490, Loss: {'train/loss': 0.17205283045768738, 'train/grad_norm_pre_clip': 19.743240356445312, 'timing/data': 0.0006917808204889297, 'timing/model': 0.9311355939134955, 'learning_rate/qwen_vl_interface': 1.5790579929242475e-06, 'learning_rate/action_model': 1.5790579929242475e-05, 'epoch': 2.49}) +[2026-06-18 01:09:26,348][starVLA.training.train_starvla][INFO] - Step 5491, Loss: {'train/loss': 0.12300962954759598, 'train/grad_norm_pre_clip': 24.07221221923828, 'timing/data': 0.0009737871587276459, 'timing/model': 0.9283334333449602, 'learning_rate/qwen_vl_interface': 1.577685901403212e-06, 'learning_rate/action_model': 1.577685901403212e-05, 'epoch': 2.49}) +[2026-06-18 01:09:31,244][starVLA.training.train_starvla][INFO] - Step 5492, Loss: {'train/loss': 0.10455682873725891, 'train/grad_norm_pre_clip': 17.456764221191406, 'timing/data': 0.0010595452040433884, 'timing/model': 1.7386938510462642, 'learning_rate/qwen_vl_interface': 1.5763145711568414e-06, 'learning_rate/action_model': 1.5763145711568413e-05, 'epoch': 2.49}) +[2026-06-18 01:09:35,234][starVLA.training.train_starvla][INFO] - Step 5493, Loss: {'train/loss': 0.18462492525577545, 'train/grad_norm_pre_clip': 33.32294464111328, 'timing/data': 0.0005896855145692825, 'timing/model': 1.1001452449709177, 'learning_rate/qwen_vl_interface': 1.5749440024694129e-06, 'learning_rate/action_model': 1.574944002469413e-05, 'epoch': 2.49}) +[2026-06-18 01:09:39,238][starVLA.training.train_starvla][INFO] - Step 5494, Loss: {'train/loss': 0.1719098836183548, 'train/grad_norm_pre_clip': 34.2889518737793, 'timing/data': 0.0008020438253879547, 'timing/model': 0.9298928314819932, 'learning_rate/qwen_vl_interface': 1.5735741956250476e-06, 'learning_rate/action_model': 1.5735741956250475e-05, 'epoch': 2.49}) +[2026-06-18 01:09:43,087][starVLA.training.train_starvla][INFO] - Step 5495, Loss: {'train/loss': 0.24845559895038605, 'train/grad_norm_pre_clip': 21.314645767211914, 'timing/data': 0.001216791570186615, 'timing/model': 1.1883509354665875, 'learning_rate/qwen_vl_interface': 1.5722051509077068e-06, 'learning_rate/action_model': 1.572205150907707e-05, 'epoch': 2.49}) +[2026-06-18 01:09:46,960][starVLA.training.train_starvla][INFO] - Step 5496, Loss: {'train/loss': 0.35596975684165955, 'train/grad_norm_pre_clip': 27.727142333984375, 'timing/data': 0.00048300810158252716, 'timing/model': 0.9253249745815992, 'learning_rate/qwen_vl_interface': 1.570836868601196e-06, 'learning_rate/action_model': 1.570836868601196e-05, 'epoch': 2.49}) +[2026-06-18 01:09:50,837][starVLA.training.train_starvla][INFO] - Step 5497, Loss: {'train/loss': 0.18363751471042633, 'train/grad_norm_pre_clip': 21.916351318359375, 'timing/data': 0.0008403519168496132, 'timing/model': 1.0215086340904236, 'learning_rate/qwen_vl_interface': 1.5694693489891616e-06, 'learning_rate/action_model': 1.5694693489891617e-05, 'epoch': 2.49}) +[2026-06-18 01:09:54,558][starVLA.training.train_starvla][INFO] - Step 5498, Loss: {'train/loss': 0.2367038130760193, 'train/grad_norm_pre_clip': 33.67975616455078, 'timing/data': 0.0010230839252471924, 'timing/model': 0.9330769218504429, 'learning_rate/qwen_vl_interface': 1.5681025923550914e-06, 'learning_rate/action_model': 1.568102592355091e-05, 'epoch': 2.49}) +[2026-06-18 01:09:59,685][starVLA.training.train_starvla][INFO] - Step 5499, Loss: {'train/loss': 0.3335275948047638, 'train/grad_norm_pre_clip': 28.672780990600586, 'timing/data': 0.0005553308874368668, 'timing/model': 1.0078033534809947, 'learning_rate/qwen_vl_interface': 1.5667365989823161e-06, 'learning_rate/action_model': 1.566736598982316e-05, 'epoch': 2.49}) +[2026-06-18 01:11:56,494][starVLA.training.train_starvla][INFO] - Step 5500, Loss: {'train/loss': 0.025301769375801086, 'train/grad_norm_pre_clip': 6.084932804107666, 'eval/loss': 0.25843727588653564, 'eval/latency_8/loss': 0.25843727588653564, 'eval/demon_fire_precision': 0.9733663672963766, 'eval/demon_fire_recall': 0.976390183286735, 'eval/demon_fire_f1': 0.9748759305210919, 'eval/demon_fire_rate_model': 0.8984418475236505, 'eval/demon_fire_rate_teacher': 0.8956594323873122, 'eval/demon_left_precision': 0.8865248226950354, 'eval/demon_left_recall': 0.8029978586723768, 'eval/demon_left_f1': 0.8426966292134831, 'eval/demon_left_rate_model': 0.11769616026711185, 'eval/demon_left_rate_teacher': 0.12993878686700056, 'eval/demon_right_precision': 0.8835725677830941, 'eval/demon_right_recall': 0.8393939393939394, 'eval/demon_right_f1': 0.8609168609168609, 'eval/demon_right_rate_model': 0.17445742904841402, 'eval/demon_right_rate_teacher': 0.18363939899833054, 'eval/demon_move_f1': 0.8518067450651721, 'eval/demon_cc_f1': 0.9343762768268886, 'eval/latency_8/demon_fire_precision': 0.9733663672963766, 'eval/latency_8/demon_fire_recall': 0.976390183286735, 'eval/latency_8/demon_fire_f1': 0.9748759305210919, 'eval/latency_8/demon_fire_rate_model': 0.8984418475236505, 'eval/latency_8/demon_fire_rate_teacher': 0.8956594323873122, 'eval/latency_8/demon_left_precision': 0.8865248226950354, 'eval/latency_8/demon_left_recall': 0.8029978586723768, 'eval/latency_8/demon_left_f1': 0.8426966292134831, 'eval/latency_8/demon_left_rate_model': 0.11769616026711185, 'eval/latency_8/demon_left_rate_teacher': 0.12993878686700056, 'eval/latency_8/demon_right_precision': 0.8835725677830941, 'eval/latency_8/demon_right_recall': 0.8393939393939394, 'eval/latency_8/demon_right_f1': 0.8609168609168609, 'eval/latency_8/demon_right_rate_model': 0.17445742904841402, 'eval/latency_8/demon_right_rate_teacher': 0.18363939899833054, 'eval/latency_8/demon_move_f1': 0.8518067450651721, 'eval/latency_8/demon_cc_f1': 0.9343762768268886, 'timing/data': 0.0007011434063315392, 'timing/model': 0.9498138437047601, 'learning_rate/qwen_vl_interface': 1.5653713691540053e-06, 'learning_rate/action_model': 1.5653713691540054e-05, 'epoch': 2.5}) +[2026-06-18 01:11:56,498][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_5500_state +[2026-06-18 01:11:56,502][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 01:12:02,874][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_5500_state/model.safetensors +[2026-06-18 01:12:18,146][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_5500_state/optimizer.bin +[2026-06-18 01:12:18,147][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_5500_state/sampler.bin +[2026-06-18 01:12:18,147][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_5500_state/sampler_1.bin +[2026-06-18 01:12:18,150][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_5500_state/random_states_0.pkl +[2026-06-18 01:12:18,153][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 5500 +[2026-06-18 01:12:18,153][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 01:12:18,163][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 01:12:21,869][starVLA.training.train_starvla][INFO] - Step 5501, Loss: {'train/loss': 0.49148857593536377, 'train/grad_norm_pre_clip': 27.621280670166016, 'timing/data': 0.0005019446834921837, 'timing/model': 0.9304041918367147, 'learning_rate/qwen_vl_interface': 1.5640069031531774e-06, 'learning_rate/action_model': 1.5640069031531772e-05, 'epoch': 2.5}) +[2026-06-18 01:12:25,746][starVLA.training.train_starvla][INFO] - Step 5502, Loss: {'train/loss': 0.14590340852737427, 'train/grad_norm_pre_clip': 17.15972328186035, 'timing/data': 0.0005116499960422516, 'timing/model': 0.9427578067407012, 'learning_rate/qwen_vl_interface': 1.5626432012626835e-06, 'learning_rate/action_model': 1.5626432012626836e-05, 'epoch': 2.5}) +[2026-06-18 01:12:29,467][starVLA.training.train_starvla][INFO] - Step 5503, Loss: {'train/loss': 0.09491035342216492, 'train/grad_norm_pre_clip': 18.70244598388672, 'timing/data': 0.0010031433776021004, 'timing/model': 1.0762216662988067, 'learning_rate/qwen_vl_interface': 1.5612802637652202e-06, 'learning_rate/action_model': 1.5612802637652203e-05, 'epoch': 2.5}) +[2026-06-18 01:12:33,180][starVLA.training.train_starvla][INFO] - Step 5504, Loss: {'train/loss': 0.2584649920463562, 'train/grad_norm_pre_clip': 22.67719078063965, 'timing/data': 0.0008703656494617462, 'timing/model': 1.1025371560826898, 'learning_rate/qwen_vl_interface': 1.5599180909433292e-06, 'learning_rate/action_model': 1.5599180909433292e-05, 'epoch': 2.5}) +[2026-06-18 01:12:36,756][starVLA.training.train_starvla][INFO] - Step 5505, Loss: {'train/loss': 0.11500455439090729, 'train/grad_norm_pre_clip': 24.40721893310547, 'timing/data': 0.0006816303357481956, 'timing/model': 0.9289922285825014, 'learning_rate/qwen_vl_interface': 1.5585566830793883e-06, 'learning_rate/action_model': 1.558556683079388e-05, 'epoch': 2.5}) +[2026-06-18 01:12:40,577][starVLA.training.train_starvla][INFO] - Step 5506, Loss: {'train/loss': 0.18102914094924927, 'train/grad_norm_pre_clip': 29.943113327026367, 'timing/data': 0.0009822864085435867, 'timing/model': 0.9512326968833804, 'learning_rate/qwen_vl_interface': 1.5571960404556204e-06, 'learning_rate/action_model': 1.5571960404556203e-05, 'epoch': 2.5}) +[2026-06-18 01:12:44,330][starVLA.training.train_starvla][INFO] - Step 5507, Loss: {'train/loss': 0.04862082749605179, 'train/grad_norm_pre_clip': 10.694164276123047, 'timing/data': 0.0006954846903681755, 'timing/model': 0.917165944352746, 'learning_rate/qwen_vl_interface': 1.5558361633540848e-06, 'learning_rate/action_model': 1.5558361633540848e-05, 'epoch': 2.5}) +[2026-06-18 01:12:48,016][starVLA.training.train_starvla][INFO] - Step 5508, Loss: {'train/loss': 0.08805086463689804, 'train/grad_norm_pre_clip': 13.048334121704102, 'timing/data': 0.00045332498848438263, 'timing/model': 0.9452923331409693, 'learning_rate/qwen_vl_interface': 1.5544770520566884e-06, 'learning_rate/action_model': 1.5544770520566885e-05, 'epoch': 2.5}) +[2026-06-18 01:12:53,026][starVLA.training.train_starvla][INFO] - Step 5509, Loss: {'train/loss': 0.12307193875312805, 'train/grad_norm_pre_clip': 23.008161544799805, 'timing/data': 0.0005005002021789551, 'timing/model': 1.059665978886187, 'learning_rate/qwen_vl_interface': 1.553118706845176e-06, 'learning_rate/action_model': 1.5531187068451757e-05, 'epoch': 2.5}) +[2026-06-18 01:12:56,818][starVLA.training.train_starvla][INFO] - Step 5510, Loss: {'train/loss': 0.24191033840179443, 'train/grad_norm_pre_clip': 27.114540100097656, 'timing/data': 0.00047293491661548615, 'timing/model': 1.0973719982430339, 'learning_rate/qwen_vl_interface': 1.5517611280011333e-06, 'learning_rate/action_model': 1.551761128001133e-05, 'epoch': 2.5}) +[2026-06-18 01:13:00,660][starVLA.training.train_starvla][INFO] - Step 5511, Loss: {'train/loss': 0.32259848713874817, 'train/grad_norm_pre_clip': 37.83845138549805, 'timing/data': 0.0004749307408928871, 'timing/model': 0.9173185573890805, 'learning_rate/qwen_vl_interface': 1.5504043158059886e-06, 'learning_rate/action_model': 1.5504043158059885e-05, 'epoch': 2.5}) +[2026-06-18 01:13:04,424][starVLA.training.train_starvla][INFO] - Step 5512, Loss: {'train/loss': 0.16276545822620392, 'train/grad_norm_pre_clip': 22.532913208007812, 'timing/data': 0.0005408944562077522, 'timing/model': 0.935503238812089, 'learning_rate/qwen_vl_interface': 1.54904827054101e-06, 'learning_rate/action_model': 1.54904827054101e-05, 'epoch': 2.5}) +[2026-06-18 01:13:08,270][starVLA.training.train_starvla][INFO] - Step 5513, Loss: {'train/loss': 0.3699336349964142, 'train/grad_norm_pre_clip': 34.75733947753906, 'timing/data': 0.0010765744373202324, 'timing/model': 0.9941862476989627, 'learning_rate/qwen_vl_interface': 1.5476929924873074e-06, 'learning_rate/action_model': 1.5476929924873073e-05, 'epoch': 2.5}) +[2026-06-18 01:13:12,130][starVLA.training.train_starvla][INFO] - Step 5514, Loss: {'train/loss': 0.11485128104686737, 'train/grad_norm_pre_clip': 32.74607849121094, 'timing/data': 0.0006526065990328789, 'timing/model': 0.92896927613765, 'learning_rate/qwen_vl_interface': 1.5463384819258302e-06, 'learning_rate/action_model': 1.54633848192583e-05, 'epoch': 2.5}) +[2026-06-18 01:13:15,891][starVLA.training.train_starvla][INFO] - Step 5515, Loss: {'train/loss': 0.5243843197822571, 'train/grad_norm_pre_clip': 36.938655853271484, 'timing/data': 0.000486922450363636, 'timing/model': 0.9276371775195003, 'learning_rate/qwen_vl_interface': 1.544984739137373e-06, 'learning_rate/action_model': 1.544984739137373e-05, 'epoch': 2.5}) +[2026-06-18 01:13:20,015][starVLA.training.train_starvla][INFO] - Step 5516, Loss: {'train/loss': 0.19350296258926392, 'train/grad_norm_pre_clip': 22.335208892822266, 'timing/data': 0.0008461475372314453, 'timing/model': 1.1300265416502953, 'learning_rate/qwen_vl_interface': 1.543631764402566e-06, 'learning_rate/action_model': 1.5436317644025658e-05, 'epoch': 2.5}) +[2026-06-18 01:13:23,724][starVLA.training.train_starvla][INFO] - Step 5517, Loss: {'train/loss': 0.19461148977279663, 'train/grad_norm_pre_clip': 28.0345516204834, 'timing/data': 0.0006967224180698395, 'timing/model': 1.033797087147832, 'learning_rate/qwen_vl_interface': 1.5422795580018813e-06, 'learning_rate/action_model': 1.542279558001881e-05, 'epoch': 2.5}) +[2026-06-18 01:13:27,400][starVLA.training.train_starvla][INFO] - Step 5518, Loss: {'train/loss': 0.0149559760466218, 'train/grad_norm_pre_clip': 2.928642988204956, 'timing/data': 0.0005069421604275703, 'timing/model': 0.9387505762279034, 'learning_rate/qwen_vl_interface': 1.5409281202156329e-06, 'learning_rate/action_model': 1.540928120215633e-05, 'epoch': 2.5}) +[2026-06-18 01:13:31,238][starVLA.training.train_starvla][INFO] - Step 5519, Loss: {'train/loss': 0.5448346734046936, 'train/grad_norm_pre_clip': 36.78175735473633, 'timing/data': 0.0004636356607079506, 'timing/model': 0.9238526821136475, 'learning_rate/qwen_vl_interface': 1.5395774513239777e-06, 'learning_rate/action_model': 1.539577451323978e-05, 'epoch': 2.5}) +[2026-06-18 01:13:35,013][starVLA.training.train_starvla][INFO] - Step 5520, Loss: {'train/loss': 0.08107220381498337, 'train/grad_norm_pre_clip': 16.89297103881836, 'timing/data': 0.0014111604541540146, 'timing/model': 0.9426029492169619, 'learning_rate/qwen_vl_interface': 1.5382275516069112e-06, 'learning_rate/action_model': 1.5382275516069113e-05, 'epoch': 2.5}) +[2026-06-18 01:13:38,814][starVLA.training.train_starvla][INFO] - Step 5521, Loss: {'train/loss': 0.2777774930000305, 'train/grad_norm_pre_clip': 33.19541931152344, 'timing/data': 0.0004986263811588287, 'timing/model': 0.9429068630561233, 'learning_rate/qwen_vl_interface': 1.5368784213442653e-06, 'learning_rate/action_model': 1.5368784213442654e-05, 'epoch': 2.5}) +[2026-06-18 01:13:42,584][starVLA.training.train_starvla][INFO] - Step 5522, Loss: {'train/loss': 0.15311214327812195, 'train/grad_norm_pre_clip': 31.811859130859375, 'timing/data': 0.0008581178262829781, 'timing/model': 0.928941142745316, 'learning_rate/qwen_vl_interface': 1.5355300608157172e-06, 'learning_rate/action_model': 1.5355300608157173e-05, 'epoch': 2.51}) +[2026-06-18 01:13:46,597][starVLA.training.train_starvla][INFO] - Step 5523, Loss: {'train/loss': 0.13288475573062897, 'train/grad_norm_pre_clip': 22.239187240600586, 'timing/data': 0.0013903835788369179, 'timing/model': 1.1152098160237074, 'learning_rate/qwen_vl_interface': 1.5341824703007854e-06, 'learning_rate/action_model': 1.5341824703007854e-05, 'epoch': 2.51}) +[2026-06-18 01:13:50,222][starVLA.training.train_starvla][INFO] - Step 5524, Loss: {'train/loss': 0.34202906489372253, 'train/grad_norm_pre_clip': 30.812482833862305, 'timing/data': 0.0004833843559026718, 'timing/model': 0.9527354249730706, 'learning_rate/qwen_vl_interface': 1.5328356500788263e-06, 'learning_rate/action_model': 1.5328356500788264e-05, 'epoch': 2.51}) +[2026-06-18 01:13:54,040][starVLA.training.train_starvla][INFO] - Step 5525, Loss: {'train/loss': 0.42070022225379944, 'train/grad_norm_pre_clip': 33.36295700073242, 'timing/data': 0.0004945332184433937, 'timing/model': 0.9299240009859204, 'learning_rate/qwen_vl_interface': 1.5314896004290366e-06, 'learning_rate/action_model': 1.5314896004290365e-05, 'epoch': 2.51}) +[2026-06-18 01:13:57,826][starVLA.training.train_starvla][INFO] - Step 5526, Loss: {'train/loss': 0.18897514045238495, 'train/grad_norm_pre_clip': 27.66002655029297, 'timing/data': 0.0005947472527623177, 'timing/model': 0.9373924238607287, 'learning_rate/qwen_vl_interface': 1.530144321630454e-06, 'learning_rate/action_model': 1.530144321630454e-05, 'epoch': 2.51}) +[2026-06-18 01:14:01,631][starVLA.training.train_starvla][INFO] - Step 5527, Loss: {'train/loss': 0.34231603145599365, 'train/grad_norm_pre_clip': 24.352598190307617, 'timing/data': 0.0005677668377757072, 'timing/model': 0.9296246860176325, 'learning_rate/qwen_vl_interface': 1.5287998139619564e-06, 'learning_rate/action_model': 1.5287998139619562e-05, 'epoch': 2.51}) +[2026-06-18 01:14:05,456][starVLA.training.train_starvla][INFO] - Step 5528, Loss: {'train/loss': 0.279728502035141, 'train/grad_norm_pre_clip': 28.24607276916504, 'timing/data': 0.0004953481256961823, 'timing/model': 0.9690565653145313, 'learning_rate/qwen_vl_interface': 1.5274560777022613e-06, 'learning_rate/action_model': 1.5274560777022613e-05, 'epoch': 2.51}) +[2026-06-18 01:14:09,386][starVLA.training.train_starvla][INFO] - Step 5529, Loss: {'train/loss': 0.0997086837887764, 'train/grad_norm_pre_clip': 80.72345733642578, 'timing/data': 0.0006700046360492706, 'timing/model': 0.9301707232370973, 'learning_rate/qwen_vl_interface': 1.5261131131299274e-06, 'learning_rate/action_model': 1.5261131131299272e-05, 'epoch': 2.51}) +[2026-06-18 01:14:13,257][starVLA.training.train_starvla][INFO] - Step 5530, Loss: {'train/loss': 0.4257422387599945, 'train/grad_norm_pre_clip': 25.87185287475586, 'timing/data': 0.00064825639128685, 'timing/model': 1.166793281212449, 'learning_rate/qwen_vl_interface': 1.5247709205233518e-06, 'learning_rate/action_model': 1.5247709205233518e-05, 'epoch': 2.51}) +[2026-06-18 01:14:16,874][starVLA.training.train_starvla][INFO] - Step 5531, Loss: {'train/loss': 0.2514883279800415, 'train/grad_norm_pre_clip': 21.267372131347656, 'timing/data': 0.0011560441926121712, 'timing/model': 0.9403936360031366, 'learning_rate/qwen_vl_interface': 1.5234295001607732e-06, 'learning_rate/action_model': 1.5234295001607732e-05, 'epoch': 2.51}) +[2026-06-18 01:14:20,692][starVLA.training.train_starvla][INFO] - Step 5532, Loss: {'train/loss': 0.10038439929485321, 'train/grad_norm_pre_clip': 21.452165603637695, 'timing/data': 0.0004964172840118408, 'timing/model': 0.9677329370751977, 'learning_rate/qwen_vl_interface': 1.5220888523202687e-06, 'learning_rate/action_model': 1.5220888523202686e-05, 'epoch': 2.51}) +[2026-06-18 01:14:24,637][starVLA.training.train_starvla][INFO] - Step 5533, Loss: {'train/loss': 0.15550175309181213, 'train/grad_norm_pre_clip': 18.8963565826416, 'timing/data': 0.0005327481776475906, 'timing/model': 0.9299119785428047, 'learning_rate/qwen_vl_interface': 1.5207489772797557e-06, 'learning_rate/action_model': 1.5207489772797556e-05, 'epoch': 2.51}) +[2026-06-18 01:14:28,417][starVLA.training.train_starvla][INFO] - Step 5534, Loss: {'train/loss': 0.016166362911462784, 'train/grad_norm_pre_clip': 2.537226915359497, 'timing/data': 0.0005907975137233734, 'timing/model': 0.9298042561858892, 'learning_rate/qwen_vl_interface': 1.5194098753169933e-06, 'learning_rate/action_model': 1.5194098753169933e-05, 'epoch': 2.51}) +[2026-06-18 01:14:32,384][starVLA.training.train_starvla][INFO] - Step 5535, Loss: {'train/loss': 0.21528658270835876, 'train/grad_norm_pre_clip': 21.830669403076172, 'timing/data': 0.001196182332932949, 'timing/model': 0.9252031026408076, 'learning_rate/qwen_vl_interface': 1.5180715467095799e-06, 'learning_rate/action_model': 1.5180715467095798e-05, 'epoch': 2.51}) +[2026-06-18 01:14:36,127][starVLA.training.train_starvla][INFO] - Step 5536, Loss: {'train/loss': 0.2020084112882614, 'train/grad_norm_pre_clip': 17.044300079345703, 'timing/data': 0.0005806656554341316, 'timing/model': 0.9219136368483305, 'learning_rate/qwen_vl_interface': 1.5167339917349482e-06, 'learning_rate/action_model': 1.5167339917349482e-05, 'epoch': 2.51}) +[2026-06-18 01:14:39,886][starVLA.training.train_starvla][INFO] - Step 5537, Loss: {'train/loss': 0.08421534299850464, 'train/grad_norm_pre_clip': 22.90172576904297, 'timing/data': 0.0013150395825505257, 'timing/model': 1.0985079612582922, 'learning_rate/qwen_vl_interface': 1.5153972106703778e-06, 'learning_rate/action_model': 1.5153972106703779e-05, 'epoch': 2.51}) +[2026-06-18 01:14:43,543][starVLA.training.train_starvla][INFO] - Step 5538, Loss: {'train/loss': 0.2673085331916809, 'train/grad_norm_pre_clip': 23.750532150268555, 'timing/data': 0.0005351901054382324, 'timing/model': 0.9285409487783909, 'learning_rate/qwen_vl_interface': 1.5140612037929844e-06, 'learning_rate/action_model': 1.5140612037929844e-05, 'epoch': 2.51}) +[2026-06-18 01:14:47,597][starVLA.training.train_starvla][INFO] - Step 5539, Loss: {'train/loss': 0.3923547863960266, 'train/grad_norm_pre_clip': 62.4322624206543, 'timing/data': 0.0009597865864634514, 'timing/model': 0.9785864995792508, 'learning_rate/qwen_vl_interface': 1.512725971379724e-06, 'learning_rate/action_model': 1.5127259713797239e-05, 'epoch': 2.51}) +[2026-06-18 01:14:51,536][starVLA.training.train_starvla][INFO] - Step 5540, Loss: {'train/loss': 0.14684560894966125, 'train/grad_norm_pre_clip': 35.31332778930664, 'timing/data': 0.0007997555658221245, 'timing/model': 0.9570225868374109, 'learning_rate/qwen_vl_interface': 1.5113915137073905e-06, 'learning_rate/action_model': 1.5113915137073906e-05, 'epoch': 2.51}) +[2026-06-18 01:14:55,540][starVLA.training.train_starvla][INFO] - Step 5541, Loss: {'train/loss': 0.3077060580253601, 'train/grad_norm_pre_clip': 23.62989044189453, 'timing/data': 0.0004992960020899773, 'timing/model': 0.9305227594450116, 'learning_rate/qwen_vl_interface': 1.5100578310526198e-06, 'learning_rate/action_model': 1.5100578310526197e-05, 'epoch': 2.51}) +[2026-06-18 01:14:59,336][starVLA.training.train_starvla][INFO] - Step 5542, Loss: {'train/loss': 0.0743866041302681, 'train/grad_norm_pre_clip': 14.203296661376953, 'timing/data': 0.0006893463432788849, 'timing/model': 0.937228717841208, 'learning_rate/qwen_vl_interface': 1.508724923691885e-06, 'learning_rate/action_model': 1.5087249236918848e-05, 'epoch': 2.51}) +[2026-06-18 01:15:03,500][starVLA.training.train_starvla][INFO] - Step 5543, Loss: {'train/loss': 0.10039958357810974, 'train/grad_norm_pre_clip': 16.254058837890625, 'timing/data': 0.002760731615126133, 'timing/model': 1.4158316254615784, 'learning_rate/qwen_vl_interface': 1.5073927919014986e-06, 'learning_rate/action_model': 1.5073927919014985e-05, 'epoch': 2.51}) +[2026-06-18 01:15:07,401][starVLA.training.train_starvla][INFO] - Step 5544, Loss: {'train/loss': 0.31243473291397095, 'train/grad_norm_pre_clip': 57.90003204345703, 'timing/data': 0.0010304413735866547, 'timing/model': 1.17587885633111, 'learning_rate/qwen_vl_interface': 1.506061435957614e-06, 'learning_rate/action_model': 1.506061435957614e-05, 'epoch': 2.52}) +[2026-06-18 01:15:11,029][starVLA.training.train_starvla][INFO] - Step 5545, Loss: {'train/loss': 0.13186432421207428, 'train/grad_norm_pre_clip': 32.9659309387207, 'timing/data': 0.0009713442996144295, 'timing/model': 0.9350727656856179, 'learning_rate/qwen_vl_interface': 1.5047308561362225e-06, 'learning_rate/action_model': 1.5047308561362225e-05, 'epoch': 2.52}) +[2026-06-18 01:15:14,899][starVLA.training.train_starvla][INFO] - Step 5546, Loss: {'train/loss': 0.09199017286300659, 'train/grad_norm_pre_clip': 19.15695571899414, 'timing/data': 0.000719151459634304, 'timing/model': 0.9457157319411635, 'learning_rate/qwen_vl_interface': 1.5034010527131548e-06, 'learning_rate/action_model': 1.5034010527131549e-05, 'epoch': 2.52}) +[2026-06-18 01:15:18,885][starVLA.training.train_starvla][INFO] - Step 5547, Loss: {'train/loss': 0.06484542042016983, 'train/grad_norm_pre_clip': 35.514190673828125, 'timing/data': 0.0008705491200089455, 'timing/model': 1.0481637185439467, 'learning_rate/qwen_vl_interface': 1.5020720259640792e-06, 'learning_rate/action_model': 1.5020720259640792e-05, 'epoch': 2.52}) +[2026-06-18 01:15:23,252][starVLA.training.train_starvla][INFO] - Step 5548, Loss: {'train/loss': 0.4306684136390686, 'train/grad_norm_pre_clip': 42.078338623046875, 'timing/data': 0.0009367074817419052, 'timing/model': 0.9358051922172308, 'learning_rate/qwen_vl_interface': 1.5007437761645064e-06, 'learning_rate/action_model': 1.5007437761645063e-05, 'epoch': 2.52}) +[2026-06-18 01:15:27,147][starVLA.training.train_starvla][INFO] - Step 5549, Loss: {'train/loss': 0.1837112009525299, 'train/grad_norm_pre_clip': 28.720558166503906, 'timing/data': 0.0006665773689746857, 'timing/model': 0.9486037837341428, 'learning_rate/qwen_vl_interface': 1.4994163035897845e-06, 'learning_rate/action_model': 1.4994163035897846e-05, 'epoch': 2.52}) +[2026-06-18 01:15:30,986][starVLA.training.train_starvla][INFO] - Step 5550, Loss: {'train/loss': 0.2063273787498474, 'train/grad_norm_pre_clip': 34.55842971801758, 'timing/data': 0.0005176616832613945, 'timing/model': 1.1219996036961675, 'learning_rate/qwen_vl_interface': 1.4980896085150976e-06, 'learning_rate/action_model': 1.4980896085150975e-05, 'epoch': 2.52}) +[2026-06-18 01:15:35,545][starVLA.training.train_starvla][INFO] - Step 5551, Loss: {'train/loss': 0.3847467303276062, 'train/grad_norm_pre_clip': 43.858768463134766, 'timing/data': 0.00047897733747959137, 'timing/model': 1.1034723240882158, 'learning_rate/qwen_vl_interface': 1.4967636912154707e-06, 'learning_rate/action_model': 1.4967636912154706e-05, 'epoch': 2.52}) +[2026-06-18 01:15:39,574][starVLA.training.train_starvla][INFO] - Step 5552, Loss: {'train/loss': 0.4019920527935028, 'train/grad_norm_pre_clip': 33.907772064208984, 'timing/data': 0.058408104814589024, 'timing/model': 1.1888119084760547, 'learning_rate/qwen_vl_interface': 1.4954385519657701e-06, 'learning_rate/action_model': 1.49543855196577e-05, 'epoch': 2.52}) +[2026-06-18 01:15:43,741][starVLA.training.train_starvla][INFO] - Step 5553, Loss: {'train/loss': 0.24887435138225555, 'train/grad_norm_pre_clip': 29.584259033203125, 'timing/data': 0.0004921285435557365, 'timing/model': 0.9452446885406971, 'learning_rate/qwen_vl_interface': 1.4941141910406984e-06, 'learning_rate/action_model': 1.4941141910406984e-05, 'epoch': 2.52}) +[2026-06-18 01:15:47,565][starVLA.training.train_starvla][INFO] - Step 5554, Loss: {'train/loss': 0.10119087249040604, 'train/grad_norm_pre_clip': 25.629220962524414, 'timing/data': 0.0006714314222335815, 'timing/model': 0.9356919154524803, 'learning_rate/qwen_vl_interface': 1.4927906087147966e-06, 'learning_rate/action_model': 1.4927906087147967e-05, 'epoch': 2.52}) +[2026-06-18 01:15:51,414][starVLA.training.train_starvla][INFO] - Step 5555, Loss: {'train/loss': 0.14532561600208282, 'train/grad_norm_pre_clip': 20.163867950439453, 'timing/data': 0.0004893457517027855, 'timing/model': 0.9387512858957052, 'learning_rate/qwen_vl_interface': 1.491467805262444e-06, 'learning_rate/action_model': 1.4914678052624439e-05, 'epoch': 2.52}) +[2026-06-18 01:15:55,212][starVLA.training.train_starvla][INFO] - Step 5556, Loss: {'train/loss': 0.10417235642671585, 'train/grad_norm_pre_clip': 27.023536682128906, 'timing/data': 0.0006134966388344765, 'timing/model': 0.9668414043262601, 'learning_rate/qwen_vl_interface': 1.4901457809578598e-06, 'learning_rate/action_model': 1.4901457809578597e-05, 'epoch': 2.52}) +[2026-06-18 01:15:59,206][starVLA.training.train_starvla][INFO] - Step 5557, Loss: {'train/loss': 0.12059849500656128, 'train/grad_norm_pre_clip': 43.69123077392578, 'timing/data': 0.0005573146045207977, 'timing/model': 1.3193589765578508, 'learning_rate/qwen_vl_interface': 1.4888245360751014e-06, 'learning_rate/action_model': 1.4888245360751013e-05, 'epoch': 2.52}) +[2026-06-18 01:16:02,927][starVLA.training.train_starvla][INFO] - Step 5558, Loss: {'train/loss': 0.048987530171871185, 'train/grad_norm_pre_clip': 11.579183578491211, 'timing/data': 0.0006621219217777252, 'timing/model': 0.9385284287855029, 'learning_rate/qwen_vl_interface': 1.4875040708880631e-06, 'learning_rate/action_model': 1.487504070888063e-05, 'epoch': 2.52}) +[2026-06-18 01:16:06,757][starVLA.training.train_starvla][INFO] - Step 5559, Loss: {'train/loss': 0.23924322426319122, 'train/grad_norm_pre_clip': 22.85471534729004, 'timing/data': 0.0005252938717603683, 'timing/model': 0.9832357624545693, 'learning_rate/qwen_vl_interface': 1.4861843856704802e-06, 'learning_rate/action_model': 1.4861843856704801e-05, 'epoch': 2.52}) +[2026-06-18 01:16:10,671][starVLA.training.train_starvla][INFO] - Step 5560, Loss: {'train/loss': 0.20293334126472473, 'train/grad_norm_pre_clip': 31.7273006439209, 'timing/data': 0.0005040466785430908, 'timing/model': 0.9233734617009759, 'learning_rate/qwen_vl_interface': 1.4848654806959243e-06, 'learning_rate/action_model': 1.4848654806959242e-05, 'epoch': 2.52}) +[2026-06-18 01:16:14,474][starVLA.training.train_starvla][INFO] - Step 5561, Loss: {'train/loss': 0.43201887607574463, 'train/grad_norm_pre_clip': 28.684900283813477, 'timing/data': 0.0005195131525397301, 'timing/model': 0.9221824733540416, 'learning_rate/qwen_vl_interface': 1.483547356237805e-06, 'learning_rate/action_model': 1.4835473562378049e-05, 'epoch': 2.52}) +[2026-06-18 01:16:18,341][starVLA.training.train_starvla][INFO] - Step 5562, Loss: {'train/loss': 0.2426089197397232, 'train/grad_norm_pre_clip': 26.960418701171875, 'timing/data': 0.00121281947940588, 'timing/model': 0.9776641679927707, 'learning_rate/qwen_vl_interface': 1.4822300125693709e-06, 'learning_rate/action_model': 1.4822300125693708e-05, 'epoch': 2.52}) +[2026-06-18 01:16:22,124][starVLA.training.train_starvla][INFO] - Step 5563, Loss: {'train/loss': 0.04952654615044594, 'train/grad_norm_pre_clip': 11.906702041625977, 'timing/data': 0.000528440810739994, 'timing/model': 1.1029608910903335, 'learning_rate/qwen_vl_interface': 1.480913449963712e-06, 'learning_rate/action_model': 1.480913449963712e-05, 'epoch': 2.52}) +[2026-06-18 01:16:26,223][starVLA.training.train_starvla][INFO] - Step 5564, Loss: {'train/loss': 0.06454352289438248, 'train/grad_norm_pre_clip': 16.580322265625, 'timing/data': 0.0004955017939209938, 'timing/model': 1.148123317398131, 'learning_rate/qwen_vl_interface': 1.4795976686937488e-06, 'learning_rate/action_model': 1.4795976686937487e-05, 'epoch': 2.52}) +[2026-06-18 01:16:29,820][starVLA.training.train_starvla][INFO] - Step 5565, Loss: {'train/loss': 0.04557270556688309, 'train/grad_norm_pre_clip': 9.051623344421387, 'timing/data': 0.0004918072372674942, 'timing/model': 0.9258146220818162, 'learning_rate/qwen_vl_interface': 1.4782826690322453e-06, 'learning_rate/action_model': 1.4782826690322451e-05, 'epoch': 2.52}) +[2026-06-18 01:16:33,648][starVLA.training.train_starvla][INFO] - Step 5566, Loss: {'train/loss': 0.13406914472579956, 'train/grad_norm_pre_clip': 23.328998565673828, 'timing/data': 0.0004824623465538025, 'timing/model': 0.936776296235621, 'learning_rate/qwen_vl_interface': 1.4769684512518045e-06, 'learning_rate/action_model': 1.4769684512518045e-05, 'epoch': 2.53}) +[2026-06-18 01:16:37,440][starVLA.training.train_starvla][INFO] - Step 5567, Loss: {'train/loss': 0.1569146364927292, 'train/grad_norm_pre_clip': 28.09286117553711, 'timing/data': 0.0004992466419935226, 'timing/model': 0.928724599070847, 'learning_rate/qwen_vl_interface': 1.4756550156248633e-06, 'learning_rate/action_model': 1.4756550156248632e-05, 'epoch': 2.53}) +[2026-06-18 01:16:41,273][starVLA.training.train_starvla][INFO] - Step 5568, Loss: {'train/loss': 0.20393428206443787, 'train/grad_norm_pre_clip': 23.065616607666016, 'timing/data': 0.0008323462679982185, 'timing/model': 0.980805354192853, 'learning_rate/qwen_vl_interface': 1.4743423624237e-06, 'learning_rate/action_model': 1.4743423624236998e-05, 'epoch': 2.53}) +[2026-06-18 01:16:45,108][starVLA.training.train_starvla][INFO] - Step 5569, Loss: {'train/loss': 0.16535796225070953, 'train/grad_norm_pre_clip': 38.81013488769531, 'timing/data': 0.001184859313070774, 'timing/model': 0.94544424302876, 'learning_rate/qwen_vl_interface': 1.4730304919204253e-06, 'learning_rate/action_model': 1.4730304919204252e-05, 'epoch': 2.53}) +[2026-06-18 01:16:48,886][starVLA.training.train_starvla][INFO] - Step 5570, Loss: {'train/loss': 0.17259205877780914, 'train/grad_norm_pre_clip': 24.97956657409668, 'timing/data': 0.0006472617387771606, 'timing/model': 1.0834044003859162, 'learning_rate/qwen_vl_interface': 1.4717194043869953e-06, 'learning_rate/action_model': 1.4717194043869953e-05, 'epoch': 2.53}) +[2026-06-18 01:16:52,523][starVLA.training.train_starvla][INFO] - Step 5571, Loss: {'train/loss': 0.14631111919879913, 'train/grad_norm_pre_clip': 18.82345199584961, 'timing/data': 0.001193137839436531, 'timing/model': 0.9240859290584922, 'learning_rate/qwen_vl_interface': 1.4704091000951982e-06, 'learning_rate/action_model': 1.4704091000951981e-05, 'epoch': 2.53}) +[2026-06-18 01:16:56,357][starVLA.training.train_starvla][INFO] - Step 5572, Loss: {'train/loss': 0.10152167081832886, 'train/grad_norm_pre_clip': 23.08605194091797, 'timing/data': 0.0004927376285195351, 'timing/model': 0.9348767809569836, 'learning_rate/qwen_vl_interface': 1.4690995793166615e-06, 'learning_rate/action_model': 1.4690995793166616e-05, 'epoch': 2.53}) +[2026-06-18 01:17:00,167][starVLA.training.train_starvla][INFO] - Step 5573, Loss: {'train/loss': 0.10779732465744019, 'train/grad_norm_pre_clip': 20.853910446166992, 'timing/data': 0.0004984168335795403, 'timing/model': 0.988017225638032, 'learning_rate/qwen_vl_interface': 1.4677908423228506e-06, 'learning_rate/action_model': 1.4677908423228507e-05, 'epoch': 2.53}) +[2026-06-18 01:17:04,018][starVLA.training.train_starvla][INFO] - Step 5574, Loss: {'train/loss': 0.3546333909034729, 'train/grad_norm_pre_clip': 32.64208221435547, 'timing/data': 0.0004919040948152542, 'timing/model': 0.9389751022681594, 'learning_rate/qwen_vl_interface': 1.4664828893850677e-06, 'learning_rate/action_model': 1.4664828893850676e-05, 'epoch': 2.53}) +[2026-06-18 01:17:07,907][starVLA.training.train_starvla][INFO] - Step 5575, Loss: {'train/loss': 0.13215242326259613, 'train/grad_norm_pre_clip': 16.9510555267334, 'timing/data': 0.0006871521472930908, 'timing/model': 0.9541356824338436, 'learning_rate/qwen_vl_interface': 1.4651757207744539e-06, 'learning_rate/action_model': 1.4651757207744537e-05, 'epoch': 2.53}) +[2026-06-18 01:17:11,889][starVLA.training.train_starvla][INFO] - Step 5576, Loss: {'train/loss': 0.22223268449306488, 'train/grad_norm_pre_clip': 29.23015785217285, 'timing/data': 0.0005476363003253937, 'timing/model': 1.2612982681021094, 'learning_rate/qwen_vl_interface': 1.4638693367619837e-06, 'learning_rate/action_model': 1.4638693367619834e-05, 'epoch': 2.53}) +[2026-06-18 01:17:17,722][starVLA.training.train_starvla][INFO] - Step 5577, Loss: {'train/loss': 0.2109038233757019, 'train/grad_norm_pre_clip': 28.396739959716797, 'timing/data': 0.0004884917289018631, 'timing/model': 2.6438951157033443, 'learning_rate/qwen_vl_interface': 1.4625637376184768e-06, 'learning_rate/action_model': 1.4625637376184768e-05, 'epoch': 2.53}) +[2026-06-18 01:17:24,582][starVLA.training.train_starvla][INFO] - Step 5578, Loss: {'train/loss': 0.22770559787750244, 'train/grad_norm_pre_clip': 48.07779312133789, 'timing/data': 0.0011163046583533287, 'timing/model': 2.444962282665074, 'learning_rate/qwen_vl_interface': 1.4612589236145811e-06, 'learning_rate/action_model': 1.461258923614581e-05, 'epoch': 2.53}) +[2026-06-18 01:17:29,922][starVLA.training.train_starvla][INFO] - Step 5579, Loss: {'train/loss': 0.2524329125881195, 'train/grad_norm_pre_clip': 27.344486236572266, 'timing/data': 0.0005294084548950195, 'timing/model': 0.9320048773661256, 'learning_rate/qwen_vl_interface': 1.4599548950207878e-06, 'learning_rate/action_model': 1.4599548950207878e-05, 'epoch': 2.53}) +[2026-06-18 01:17:34,271][starVLA.training.train_starvla][INFO] - Step 5580, Loss: {'train/loss': 0.20447063446044922, 'train/grad_norm_pre_clip': 25.567459106445312, 'timing/data': 0.0006792806088924408, 'timing/model': 1.371254114434123, 'learning_rate/qwen_vl_interface': 1.4586516521074214e-06, 'learning_rate/action_model': 1.4586516521074214e-05, 'epoch': 2.53}) +[2026-06-18 01:17:38,089][starVLA.training.train_starvla][INFO] - Step 5581, Loss: {'train/loss': 0.09877196699380875, 'train/grad_norm_pre_clip': 15.268608093261719, 'timing/data': 0.0005489010363817215, 'timing/model': 0.9432068401947618, 'learning_rate/qwen_vl_interface': 1.457349195144649e-06, 'learning_rate/action_model': 1.457349195144649e-05, 'epoch': 2.53}) +[2026-06-18 01:17:41,973][starVLA.training.train_starvla][INFO] - Step 5582, Loss: {'train/loss': 0.10345017910003662, 'train/grad_norm_pre_clip': 14.089033126831055, 'timing/data': 0.0005331002175807953, 'timing/model': 0.992038082331419, 'learning_rate/qwen_vl_interface': 1.4560475244024702e-06, 'learning_rate/action_model': 1.4560475244024702e-05, 'epoch': 2.53}) +[2026-06-18 01:17:45,878][starVLA.training.train_starvla][INFO] - Step 5583, Loss: {'train/loss': 0.09730187058448792, 'train/grad_norm_pre_clip': 18.00235366821289, 'timing/data': 0.0006912890821695328, 'timing/model': 1.1388081209734082, 'learning_rate/qwen_vl_interface': 1.4547466401507213e-06, 'learning_rate/action_model': 1.4547466401507211e-05, 'epoch': 2.53}) +[2026-06-18 01:17:49,743][starVLA.training.train_starvla][INFO] - Step 5584, Loss: {'train/loss': 0.14190395176410675, 'train/grad_norm_pre_clip': 16.172260284423828, 'timing/data': 0.0004890067502856255, 'timing/model': 1.1051090899854898, 'learning_rate/qwen_vl_interface': 1.4534465426590763e-06, 'learning_rate/action_model': 1.4534465426590763e-05, 'epoch': 2.53}) +[2026-06-18 01:17:53,448][starVLA.training.train_starvla][INFO] - Step 5585, Loss: {'train/loss': 0.10516076534986496, 'train/grad_norm_pre_clip': 19.012420654296875, 'timing/data': 0.0010212091729044914, 'timing/model': 0.9625229239463806, 'learning_rate/qwen_vl_interface': 1.4521472321970496e-06, 'learning_rate/action_model': 1.4521472321970494e-05, 'epoch': 2.53}) +[2026-06-18 01:17:57,401][starVLA.training.train_starvla][INFO] - Step 5586, Loss: {'train/loss': 0.2786438763141632, 'train/grad_norm_pre_clip': 56.25346374511719, 'timing/data': 0.0008302638307213783, 'timing/model': 0.9562575938180089, 'learning_rate/qwen_vl_interface': 1.450848709033988e-06, 'learning_rate/action_model': 1.4508487090339878e-05, 'epoch': 2.53}) +[2026-06-18 01:18:01,254][starVLA.training.train_starvla][INFO] - Step 5587, Loss: {'train/loss': 0.3295849561691284, 'train/grad_norm_pre_clip': 57.45473098754883, 'timing/data': 0.0015937145799398422, 'timing/model': 0.9496904360130429, 'learning_rate/qwen_vl_interface': 1.4495509734390769e-06, 'learning_rate/action_model': 1.4495509734390769e-05, 'epoch': 2.53}) +[2026-06-18 01:18:05,033][starVLA.training.train_starvla][INFO] - Step 5588, Loss: {'train/loss': 0.11977303773164749, 'train/grad_norm_pre_clip': 17.2747745513916, 'timing/data': 0.001116734929382801, 'timing/model': 0.9331272030249238, 'learning_rate/qwen_vl_interface': 1.448254025681339e-06, 'learning_rate/action_model': 1.448254025681339e-05, 'epoch': 2.54}) +[2026-06-18 01:18:08,949][starVLA.training.train_starvla][INFO] - Step 5589, Loss: {'train/loss': 0.2765408456325531, 'train/grad_norm_pre_clip': 22.12213134765625, 'timing/data': 0.0011341609060764313, 'timing/model': 0.9843646856024861, 'learning_rate/qwen_vl_interface': 1.4469578660296327e-06, 'learning_rate/action_model': 1.4469578660296326e-05, 'epoch': 2.54}) +[2026-06-18 01:18:13,643][starVLA.training.train_starvla][INFO] - Step 5590, Loss: {'train/loss': 0.31925615668296814, 'train/grad_norm_pre_clip': 41.24190139770508, 'timing/data': 0.0008400985971093178, 'timing/model': 1.6446679271757603, 'learning_rate/qwen_vl_interface': 1.4456624947526524e-06, 'learning_rate/action_model': 1.4456624947526524e-05, 'epoch': 2.54}) +[2026-06-18 01:18:18,504][starVLA.training.train_starvla][INFO] - Step 5591, Loss: {'train/loss': 0.15689617395401, 'train/grad_norm_pre_clip': 23.354768753051758, 'timing/data': 0.0009914366528391838, 'timing/model': 1.6428957302123308, 'learning_rate/qwen_vl_interface': 1.4443679121189306e-06, 'learning_rate/action_model': 1.4443679121189305e-05, 'epoch': 2.54}) +[2026-06-18 01:18:22,167][starVLA.training.train_starvla][INFO] - Step 5592, Loss: {'train/loss': 0.301252156496048, 'train/grad_norm_pre_clip': 35.946842193603516, 'timing/data': 0.0004855189472436905, 'timing/model': 0.9370950413867831, 'learning_rate/qwen_vl_interface': 1.4430741183968361e-06, 'learning_rate/action_model': 1.443074118396836e-05, 'epoch': 2.54}) +[2026-06-18 01:18:26,019][starVLA.training.train_starvla][INFO] - Step 5593, Loss: {'train/loss': 0.22358423471450806, 'train/grad_norm_pre_clip': 28.079662322998047, 'timing/data': 0.000507003627717495, 'timing/model': 0.9574159691110253, 'learning_rate/qwen_vl_interface': 1.441781113854573e-06, 'learning_rate/action_model': 1.4417811138545728e-05, 'epoch': 2.54}) +[2026-06-18 01:18:30,363][starVLA.training.train_starvla][INFO] - Step 5594, Loss: {'train/loss': 0.1386062651872635, 'train/grad_norm_pre_clip': 18.834814071655273, 'timing/data': 0.0004993239417672157, 'timing/model': 1.4909373484551907, 'learning_rate/qwen_vl_interface': 1.4404888987601828e-06, 'learning_rate/action_model': 1.4404888987601828e-05, 'epoch': 2.54}) +[2026-06-18 01:18:36,147][starVLA.training.train_starvla][INFO] - Step 5595, Loss: {'train/loss': 0.5148876309394836, 'train/grad_norm_pre_clip': 36.25958251953125, 'timing/data': 0.0010206401348114014, 'timing/model': 0.9965734239667654, 'learning_rate/qwen_vl_interface': 1.4391974733815417e-06, 'learning_rate/action_model': 1.4391974733815416e-05, 'epoch': 2.54}) +[2026-06-18 01:18:40,419][starVLA.training.train_starvla][INFO] - Step 5596, Loss: {'train/loss': 0.1740587055683136, 'train/grad_norm_pre_clip': 36.89518356323242, 'timing/data': 0.0005564233288168907, 'timing/model': 0.9668763689696789, 'learning_rate/qwen_vl_interface': 1.437906837986368e-06, 'learning_rate/action_model': 1.4379068379863677e-05, 'epoch': 2.54}) +[2026-06-18 01:18:44,405][starVLA.training.train_starvla][INFO] - Step 5597, Loss: {'train/loss': 0.17747023701667786, 'train/grad_norm_pre_clip': 31.26165771484375, 'timing/data': 0.0004854481667280197, 'timing/model': 1.1787148611620069, 'learning_rate/qwen_vl_interface': 1.436616992842207e-06, 'learning_rate/action_model': 1.4366169928422068e-05, 'epoch': 2.54}) +[2026-06-18 01:18:49,511][starVLA.training.train_starvla][INFO] - Step 5598, Loss: {'train/loss': 0.10261794179677963, 'train/grad_norm_pre_clip': 24.17565155029297, 'timing/data': 0.0009129121899604797, 'timing/model': 1.1125199934467673, 'learning_rate/qwen_vl_interface': 1.4353279382164459e-06, 'learning_rate/action_model': 1.4353279382164458e-05, 'epoch': 2.54}) +[2026-06-18 01:18:54,747][starVLA.training.train_starvla][INFO] - Step 5599, Loss: {'train/loss': 0.11768443882465363, 'train/grad_norm_pre_clip': 19.43688201904297, 'timing/data': 0.0005053021013736725, 'timing/model': 2.2887618113309145, 'learning_rate/qwen_vl_interface': 1.4340396743763084e-06, 'learning_rate/action_model': 1.4340396743763085e-05, 'epoch': 2.54}) +[2026-06-18 01:21:12,542][starVLA.training.train_starvla][INFO] - Step 5600, Loss: {'train/loss': 0.2825862169265747, 'train/grad_norm_pre_clip': 30.544662475585938, 'eval/loss': 0.2570497989654541, 'eval/latency_8/loss': 0.2570497989654541, 'eval/demon_fire_precision': 0.9687882496940025, 'eval/demon_fire_recall': 0.9835352593973283, 'eval/demon_fire_f1': 0.976106058270387, 'eval/demon_fire_rate_model': 0.90929326655537, 'eval/demon_fire_rate_teacher': 0.8956594323873122, 'eval/demon_left_precision': 0.8784403669724771, 'eval/demon_left_recall': 0.8201284796573876, 'eval/demon_left_f1': 0.84828349944629, 'eval/demon_left_rate_model': 0.1213132999443517, 'eval/demon_left_rate_teacher': 0.12993878686700056, 'eval/demon_right_precision': 0.8934426229508197, 'eval/demon_right_recall': 0.8257575757575758, 'eval/demon_right_f1': 0.8582677165354332, 'eval/demon_right_rate_model': 0.16972732331663884, 'eval/demon_right_rate_teacher': 0.18363939899833054, 'eval/demon_move_f1': 0.8532756079908617, 'eval/demon_cc_f1': 0.9356974927550277, 'eval/latency_8/demon_fire_precision': 0.9687882496940025, 'eval/latency_8/demon_fire_recall': 0.9835352593973283, 'eval/latency_8/demon_fire_f1': 0.976106058270387, 'eval/latency_8/demon_fire_rate_model': 0.90929326655537, 'eval/latency_8/demon_fire_rate_teacher': 0.8956594323873122, 'eval/latency_8/demon_left_precision': 0.8784403669724771, 'eval/latency_8/demon_left_recall': 0.8201284796573876, 'eval/latency_8/demon_left_f1': 0.84828349944629, 'eval/latency_8/demon_left_rate_model': 0.1213132999443517, 'eval/latency_8/demon_left_rate_teacher': 0.12993878686700056, 'eval/latency_8/demon_right_precision': 0.8934426229508197, 'eval/latency_8/demon_right_recall': 0.8257575757575758, 'eval/latency_8/demon_right_f1': 0.8582677165354332, 'eval/latency_8/demon_right_rate_model': 0.16972732331663884, 'eval/latency_8/demon_right_rate_teacher': 0.18363939899833054, 'eval/latency_8/demon_move_f1': 0.8532756079908617, 'eval/latency_8/demon_cc_f1': 0.9356974927550277, 'timing/data': 0.001061098650097847, 'timing/model': 0.9785184860229492, 'learning_rate/qwen_vl_interface': 1.4327522015888535e-06, 'learning_rate/action_model': 1.4327522015888534e-05, 'epoch': 2.54}) +[2026-06-18 01:21:12,545][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_5600_state +[2026-06-18 01:21:12,550][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 01:21:18,700][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_5600_state/model.safetensors +[2026-06-18 01:21:34,067][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_5600_state/optimizer.bin +[2026-06-18 01:21:34,068][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_5600_state/sampler.bin +[2026-06-18 01:21:34,068][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_5600_state/sampler_1.bin +[2026-06-18 01:21:34,070][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_5600_state/random_states_0.pkl +[2026-06-18 01:21:34,077][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 5600 +[2026-06-18 01:21:34,077][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 01:21:34,090][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 01:21:37,949][starVLA.training.train_starvla][INFO] - Step 5601, Loss: {'train/loss': 0.1987999975681305, 'train/grad_norm_pre_clip': 27.00794219970703, 'timing/data': 0.0008740704506635666, 'timing/model': 0.9450847897678614, 'learning_rate/qwen_vl_interface': 1.4314655201209756e-06, 'learning_rate/action_model': 1.4314655201209757e-05, 'epoch': 2.54}) +[2026-06-18 01:21:41,940][starVLA.training.train_starvla][INFO] - Step 5602, Loss: {'train/loss': 0.1678980141878128, 'train/grad_norm_pre_clip': 31.365957260131836, 'timing/data': 0.00047268159687519073, 'timing/model': 0.9217921923846006, 'learning_rate/qwen_vl_interface': 1.4301796302394019e-06, 'learning_rate/action_model': 1.4301796302394018e-05, 'epoch': 2.54}) +[2026-06-18 01:21:45,798][starVLA.training.train_starvla][INFO] - Step 5603, Loss: {'train/loss': 0.2834404706954956, 'train/grad_norm_pre_clip': 41.133296966552734, 'timing/data': 0.0005690623074769974, 'timing/model': 0.9855106826871634, 'learning_rate/qwen_vl_interface': 1.4288945322107025e-06, 'learning_rate/action_model': 1.4288945322107026e-05, 'epoch': 2.54}) +[2026-06-18 01:21:49,629][starVLA.training.train_starvla][INFO] - Step 5604, Loss: {'train/loss': 0.22399486601352692, 'train/grad_norm_pre_clip': 20.5799560546875, 'timing/data': 0.000569147989153862, 'timing/model': 0.9485955182462931, 'learning_rate/qwen_vl_interface': 1.4276102263012781e-06, 'learning_rate/action_model': 1.427610226301278e-05, 'epoch': 2.54}) +[2026-06-18 01:21:53,438][starVLA.training.train_starvla][INFO] - Step 5605, Loss: {'train/loss': 0.5629342198371887, 'train/grad_norm_pre_clip': 43.60744857788086, 'timing/data': 0.0006131930276751518, 'timing/model': 0.9331147996708751, 'learning_rate/qwen_vl_interface': 1.426326712777367e-06, 'learning_rate/action_model': 1.4263267127773669e-05, 'epoch': 2.54}) +[2026-06-18 01:21:57,263][starVLA.training.train_starvla][INFO] - Step 5606, Loss: {'train/loss': 0.16920512914657593, 'train/grad_norm_pre_clip': 21.831995010375977, 'timing/data': 0.0005107996985316277, 'timing/model': 1.1239237310364842, 'learning_rate/qwen_vl_interface': 1.425043991905043e-06, 'learning_rate/action_model': 1.425043991905043e-05, 'epoch': 2.54}) +[2026-06-18 01:22:00,952][starVLA.training.train_starvla][INFO] - Step 5607, Loss: {'train/loss': 0.24934686720371246, 'train/grad_norm_pre_clip': 18.976852416992188, 'timing/data': 0.0006456151604652405, 'timing/model': 0.9611973902210593, 'learning_rate/qwen_vl_interface': 1.4237620639502149e-06, 'learning_rate/action_model': 1.4237620639502147e-05, 'epoch': 2.54}) +[2026-06-18 01:22:04,796][starVLA.training.train_starvla][INFO] - Step 5608, Loss: {'train/loss': 0.6026440858840942, 'train/grad_norm_pre_clip': 62.51591873168945, 'timing/data': 0.000596298836171627, 'timing/model': 0.9521629130467772, 'learning_rate/qwen_vl_interface': 1.4224809291786281e-06, 'learning_rate/action_model': 1.4224809291786282e-05, 'epoch': 2.54}) +[2026-06-18 01:22:08,645][starVLA.training.train_starvla][INFO] - Step 5609, Loss: {'train/loss': 0.4515914022922516, 'train/grad_norm_pre_clip': 43.23206329345703, 'timing/data': 0.0007554516196250916, 'timing/model': 0.9568772474303842, 'learning_rate/qwen_vl_interface': 1.4212005878558622e-06, 'learning_rate/action_model': 1.4212005878558622e-05, 'epoch': 2.54}) +[2026-06-18 01:22:12,436][starVLA.training.train_starvla][INFO] - Step 5610, Loss: {'train/loss': 0.06150307133793831, 'train/grad_norm_pre_clip': 13.312417030334473, 'timing/data': 0.0009425980970263481, 'timing/model': 0.9431554712355137, 'learning_rate/qwen_vl_interface': 1.4199210402473368e-06, 'learning_rate/action_model': 1.4199210402473367e-05, 'epoch': 2.55}) +[2026-06-18 01:22:16,223][starVLA.training.train_starvla][INFO] - Step 5611, Loss: {'train/loss': 0.07684523612260818, 'train/grad_norm_pre_clip': 15.337723731994629, 'timing/data': 0.0006321584805846214, 'timing/model': 0.9267000919207931, 'learning_rate/qwen_vl_interface': 1.4186422866183e-06, 'learning_rate/action_model': 1.4186422866182999e-05, 'epoch': 2.55}) +[2026-06-18 01:22:20,007][starVLA.training.train_starvla][INFO] - Step 5612, Loss: {'train/loss': 0.40359821915626526, 'train/grad_norm_pre_clip': 33.26053237915039, 'timing/data': 0.0005339663475751877, 'timing/model': 1.1240801271051168, 'learning_rate/qwen_vl_interface': 1.41736432723384e-06, 'learning_rate/action_model': 1.41736432723384e-05, 'epoch': 2.55}) +[2026-06-18 01:22:23,733][starVLA.training.train_starvla][INFO] - Step 5613, Loss: {'train/loss': 0.16646656394004822, 'train/grad_norm_pre_clip': 22.01688003540039, 'timing/data': 0.0004996899515390396, 'timing/model': 1.0542574590072036, 'learning_rate/qwen_vl_interface': 1.4160871623588774e-06, 'learning_rate/action_model': 1.4160871623588772e-05, 'epoch': 2.55}) +[2026-06-18 01:22:27,419][starVLA.training.train_starvla][INFO] - Step 5614, Loss: {'train/loss': 0.3543626070022583, 'train/grad_norm_pre_clip': 23.73817253112793, 'timing/data': 0.000988820567727089, 'timing/model': 0.9380202377215028, 'learning_rate/qwen_vl_interface': 1.414810792258173e-06, 'learning_rate/action_model': 1.414810792258173e-05, 'epoch': 2.55}) +[2026-06-18 01:22:33,341][starVLA.training.train_starvla][INFO] - Step 5615, Loss: {'train/loss': 0.34275802969932556, 'train/grad_norm_pre_clip': 39.98693084716797, 'timing/data': 0.0010516857728362083, 'timing/model': 0.9309215769171715, 'learning_rate/qwen_vl_interface': 1.4135352171963186e-06, 'learning_rate/action_model': 1.4135352171963185e-05, 'epoch': 2.55}) +[2026-06-18 01:22:37,143][starVLA.training.train_starvla][INFO] - Step 5616, Loss: {'train/loss': 0.1320653110742569, 'train/grad_norm_pre_clip': 19.236183166503906, 'timing/data': 0.0004978002980351448, 'timing/model': 0.9267426198348403, 'learning_rate/qwen_vl_interface': 1.412260437437741e-06, 'learning_rate/action_model': 1.412260437437741e-05, 'epoch': 2.55}) +[2026-06-18 01:22:40,951][starVLA.training.train_starvla][INFO] - Step 5617, Loss: {'train/loss': 0.04853714257478714, 'train/grad_norm_pre_clip': 9.00466251373291, 'timing/data': 0.0005130534991621971, 'timing/model': 0.9270371962338686, 'learning_rate/qwen_vl_interface': 1.4109864532467051e-06, 'learning_rate/action_model': 1.4109864532467052e-05, 'epoch': 2.55}) +[2026-06-18 01:22:44,801][starVLA.training.train_starvla][INFO] - Step 5618, Loss: {'train/loss': 0.1778675764799118, 'train/grad_norm_pre_clip': 36.57207107543945, 'timing/data': 0.0006675785407423973, 'timing/model': 0.9443795634433627, 'learning_rate/qwen_vl_interface': 1.4097132648873082e-06, 'learning_rate/action_model': 1.4097132648873083e-05, 'epoch': 2.55}) +[2026-06-18 01:22:48,648][starVLA.training.train_starvla][INFO] - Step 5619, Loss: {'train/loss': 0.08001568913459778, 'train/grad_norm_pre_clip': 21.743085861206055, 'timing/data': 0.0004977332428097725, 'timing/model': 1.109620344825089, 'learning_rate/qwen_vl_interface': 1.4084408726234838e-06, 'learning_rate/action_model': 1.4084408726234838e-05, 'epoch': 2.55}) +[2026-06-18 01:22:52,332][starVLA.training.train_starvla][INFO] - Step 5620, Loss: {'train/loss': 0.03282076492905617, 'train/grad_norm_pre_clip': 5.005053997039795, 'timing/data': 0.0009525669738650322, 'timing/model': 0.958225934766233, 'learning_rate/qwen_vl_interface': 1.407169276719e-06, 'learning_rate/action_model': 1.4071692767190001e-05, 'epoch': 2.55}) +[2026-06-18 01:22:56,113][starVLA.training.train_starvla][INFO] - Step 5621, Loss: {'train/loss': 0.5729537606239319, 'train/grad_norm_pre_clip': 61.725521087646484, 'timing/data': 0.0005159620195627213, 'timing/model': 0.9401655001565814, 'learning_rate/qwen_vl_interface': 1.4058984774374606e-06, 'learning_rate/action_model': 1.4058984774374607e-05, 'epoch': 2.55}) +[2026-06-18 01:23:00,130][starVLA.training.train_starvla][INFO] - Step 5622, Loss: {'train/loss': 0.22082991898059845, 'train/grad_norm_pre_clip': 29.894508361816406, 'timing/data': 0.0005154367536306381, 'timing/model': 0.9566012201830745, 'learning_rate/qwen_vl_interface': 1.404628475042303e-06, 'learning_rate/action_model': 1.404628475042303e-05, 'epoch': 2.55}) +[2026-06-18 01:23:04,172][starVLA.training.train_starvla][INFO] - Step 5623, Loss: {'train/loss': 0.15860813856124878, 'train/grad_norm_pre_clip': 28.156476974487305, 'timing/data': 0.0005093347281217575, 'timing/model': 1.193264159373939, 'learning_rate/qwen_vl_interface': 1.4033592697967998e-06, 'learning_rate/action_model': 1.4033592697967998e-05, 'epoch': 2.55}) +[2026-06-18 01:23:07,981][starVLA.training.train_starvla][INFO] - Step 5624, Loss: {'train/loss': 0.10487685352563858, 'train/grad_norm_pre_clip': 25.475658416748047, 'timing/data': 0.0006860010325908661, 'timing/model': 0.956382124684751, 'learning_rate/qwen_vl_interface': 1.4020908619640602e-06, 'learning_rate/action_model': 1.4020908619640602e-05, 'epoch': 2.55}) +[2026-06-18 01:23:11,886][starVLA.training.train_starvla][INFO] - Step 5625, Loss: {'train/loss': 0.16930682957172394, 'train/grad_norm_pre_clip': 20.32782554626465, 'timing/data': 0.000767257995903492, 'timing/model': 1.2036053696647286, 'learning_rate/qwen_vl_interface': 1.4008232518070248e-06, 'learning_rate/action_model': 1.4008232518070247e-05, 'epoch': 2.55}) +[2026-06-18 01:23:15,518][starVLA.training.train_starvla][INFO] - Step 5626, Loss: {'train/loss': 0.49640318751335144, 'train/grad_norm_pre_clip': 32.90658950805664, 'timing/data': 0.0005296459421515465, 'timing/model': 0.9400539891794324, 'learning_rate/qwen_vl_interface': 1.3995564395884712e-06, 'learning_rate/action_model': 1.399556439588471e-05, 'epoch': 2.55}) +[2026-06-18 01:23:19,379][starVLA.training.train_starvla][INFO] - Step 5627, Loss: {'train/loss': 0.07331375032663345, 'train/grad_norm_pre_clip': 10.644621849060059, 'timing/data': 0.0009722700342535973, 'timing/model': 0.9947730591520667, 'learning_rate/qwen_vl_interface': 1.3982904255710094e-06, 'learning_rate/action_model': 1.3982904255710094e-05, 'epoch': 2.55}) +[2026-06-18 01:23:23,310][starVLA.training.train_starvla][INFO] - Step 5628, Loss: {'train/loss': 0.2829355001449585, 'train/grad_norm_pre_clip': 26.503747940063477, 'timing/data': 0.0005391109734773636, 'timing/model': 1.0365388384088874, 'learning_rate/qwen_vl_interface': 1.3970252100170892e-06, 'learning_rate/action_model': 1.3970252100170892e-05, 'epoch': 2.55}) +[2026-06-18 01:23:27,167][starVLA.training.train_starvla][INFO] - Step 5629, Loss: {'train/loss': 0.27589377760887146, 'train/grad_norm_pre_clip': 62.00212478637695, 'timing/data': 0.0011244481429457664, 'timing/model': 0.9877668432891369, 'learning_rate/qwen_vl_interface': 1.3957607931889885e-06, 'learning_rate/action_model': 1.3957607931889884e-05, 'epoch': 2.55}) +[2026-06-18 01:23:31,026][starVLA.training.train_starvla][INFO] - Step 5630, Loss: {'train/loss': 0.11451563239097595, 'train/grad_norm_pre_clip': 28.100866317749023, 'timing/data': 0.0004451395943760872, 'timing/model': 0.9611372807994485, 'learning_rate/qwen_vl_interface': 1.3944971753488238e-06, 'learning_rate/action_model': 1.3944971753488236e-05, 'epoch': 2.55}) +[2026-06-18 01:23:34,889][starVLA.training.train_starvla][INFO] - Step 5631, Loss: {'train/loss': 0.12991836667060852, 'train/grad_norm_pre_clip': 26.635637283325195, 'timing/data': 0.000550772063434124, 'timing/model': 1.0031033530831337, 'learning_rate/qwen_vl_interface': 1.3932343567585408e-06, 'learning_rate/action_model': 1.3932343567585407e-05, 'epoch': 2.55}) +[2026-06-18 01:23:38,878][starVLA.training.train_starvla][INFO] - Step 5632, Loss: {'train/loss': 0.07275953143835068, 'train/grad_norm_pre_clip': 19.151086807250977, 'timing/data': 0.0011785011738538742, 'timing/model': 1.2906744796782732, 'learning_rate/qwen_vl_interface': 1.3919723376799268e-06, 'learning_rate/action_model': 1.3919723376799269e-05, 'epoch': 2.56}) +[2026-06-18 01:23:42,767][starVLA.training.train_starvla][INFO] - Step 5633, Loss: {'train/loss': 0.3350418210029602, 'train/grad_norm_pre_clip': 27.232107162475586, 'timing/data': 0.0004779445007443428, 'timing/model': 0.9319831049069762, 'learning_rate/qwen_vl_interface': 1.3907111183745992e-06, 'learning_rate/action_model': 1.3907111183745991e-05, 'epoch': 2.56}) +[2026-06-18 01:23:46,927][starVLA.training.train_starvla][INFO] - Step 5634, Loss: {'train/loss': 0.05392979830503464, 'train/grad_norm_pre_clip': 16.78595733642578, 'timing/data': 0.000470050610601902, 'timing/model': 0.9539975114166737, 'learning_rate/qwen_vl_interface': 1.3894506991040089e-06, 'learning_rate/action_model': 1.3894506991040087e-05, 'epoch': 2.56}) +[2026-06-18 01:23:50,703][starVLA.training.train_starvla][INFO] - Step 5635, Loss: {'train/loss': 0.23623143136501312, 'train/grad_norm_pre_clip': 26.91985321044922, 'timing/data': 0.000464094802737236, 'timing/model': 0.9357570158317685, 'learning_rate/qwen_vl_interface': 1.3881910801294433e-06, 'learning_rate/action_model': 1.3881910801294431e-05, 'epoch': 2.56}) +[2026-06-18 01:23:56,125][starVLA.training.train_starvla][INFO] - Step 5636, Loss: {'train/loss': 0.078676238656044, 'train/grad_norm_pre_clip': 18.011730194091797, 'timing/data': 0.0009626448154449463, 'timing/model': 0.9282778967171907, 'learning_rate/qwen_vl_interface': 1.3869322617120223e-06, 'learning_rate/action_model': 1.3869322617120222e-05, 'epoch': 2.56}) +[2026-06-18 01:24:01,506][starVLA.training.train_starvla][INFO] - Step 5637, Loss: {'train/loss': 0.14471936225891113, 'train/grad_norm_pre_clip': 18.82188606262207, 'timing/data': 0.0004906188696622849, 'timing/model': 2.435425723902881, 'learning_rate/qwen_vl_interface': 1.3856742441127e-06, 'learning_rate/action_model': 1.3856742441126999e-05, 'epoch': 2.56}) +[2026-06-18 01:24:06,820][starVLA.training.train_starvla][INFO] - Step 5638, Loss: {'train/loss': 0.39791348576545715, 'train/grad_norm_pre_clip': 39.23548126220703, 'timing/data': 0.0004704520106315613, 'timing/model': 2.3996285209432244, 'learning_rate/qwen_vl_interface': 1.384417027592264e-06, 'learning_rate/action_model': 1.3844170275922638e-05, 'epoch': 2.56}) +[2026-06-18 01:24:10,691][starVLA.training.train_starvla][INFO] - Step 5639, Loss: {'train/loss': 0.0986396074295044, 'train/grad_norm_pre_clip': 13.803359031677246, 'timing/data': 0.0014117872342467308, 'timing/model': 1.2130346009507775, 'learning_rate/qwen_vl_interface': 1.3831606124113397e-06, 'learning_rate/action_model': 1.3831606124113398e-05, 'epoch': 2.56}) +[2026-06-18 01:24:14,443][starVLA.training.train_starvla][INFO] - Step 5640, Loss: {'train/loss': 0.1726827174425125, 'train/grad_norm_pre_clip': 24.93395233154297, 'timing/data': 0.00044330209493637085, 'timing/model': 0.9727861080318689, 'learning_rate/qwen_vl_interface': 1.3819049988303809e-06, 'learning_rate/action_model': 1.381904998830381e-05, 'epoch': 2.56}) +[2026-06-18 01:24:18,226][starVLA.training.train_starvla][INFO] - Step 5641, Loss: {'train/loss': 0.27043062448501587, 'train/grad_norm_pre_clip': 22.87153434753418, 'timing/data': 0.0009948574006557465, 'timing/model': 0.9240739587694407, 'learning_rate/qwen_vl_interface': 1.3806501871096774e-06, 'learning_rate/action_model': 1.3806501871096774e-05, 'epoch': 2.56}) +[2026-06-18 01:24:22,360][starVLA.training.train_starvla][INFO] - Step 5642, Loss: {'train/loss': 0.07633432000875473, 'train/grad_norm_pre_clip': 11.92487907409668, 'timing/data': 0.0004754820838570595, 'timing/model': 1.244936479255557, 'learning_rate/qwen_vl_interface': 1.3793961775093528e-06, 'learning_rate/action_model': 1.3793961775093528e-05, 'epoch': 2.56}) +[2026-06-18 01:24:26,257][starVLA.training.train_starvla][INFO] - Step 5643, Loss: {'train/loss': 0.26550623774528503, 'train/grad_norm_pre_clip': 42.94987487792969, 'timing/data': 0.00044280756264925003, 'timing/model': 0.9734124038368464, 'learning_rate/qwen_vl_interface': 1.378142970289367e-06, 'learning_rate/action_model': 1.378142970289367e-05, 'epoch': 2.56}) +[2026-06-18 01:24:30,311][starVLA.training.train_starvla][INFO] - Step 5644, Loss: {'train/loss': 0.26976439356803894, 'train/grad_norm_pre_clip': 22.136804580688477, 'timing/data': 0.0008525727316737175, 'timing/model': 1.0005381936207414, 'learning_rate/qwen_vl_interface': 1.3768905657095109e-06, 'learning_rate/action_model': 1.3768905657095107e-05, 'epoch': 2.56}) +[2026-06-18 01:24:34,208][starVLA.training.train_starvla][INFO] - Step 5645, Loss: {'train/loss': 0.2794873118400574, 'train/grad_norm_pre_clip': 26.275917053222656, 'timing/data': 0.0004577953368425369, 'timing/model': 1.0487102530896664, 'learning_rate/qwen_vl_interface': 1.3756389640294066e-06, 'learning_rate/action_model': 1.3756389640294064e-05, 'epoch': 2.56}) +[2026-06-18 01:24:38,012][starVLA.training.train_starvla][INFO] - Step 5646, Loss: {'train/loss': 0.35705217719078064, 'train/grad_norm_pre_clip': 30.355175018310547, 'timing/data': 0.00045319460332393646, 'timing/model': 1.133996701799333, 'learning_rate/qwen_vl_interface': 1.3743881655085131e-06, 'learning_rate/action_model': 1.3743881655085131e-05, 'epoch': 2.56}) +[2026-06-18 01:24:43,011][starVLA.training.train_starvla][INFO] - Step 5647, Loss: {'train/loss': 0.06030440703034401, 'train/grad_norm_pre_clip': 9.899301528930664, 'timing/data': 0.0008156159892678261, 'timing/model': 1.0349575160071254, 'learning_rate/qwen_vl_interface': 1.3731381704061248e-06, 'learning_rate/action_model': 1.3731381704061246e-05, 'epoch': 2.56}) +[2026-06-18 01:24:48,070][starVLA.training.train_starvla][INFO] - Step 5648, Loss: {'train/loss': 0.040431953966617584, 'train/grad_norm_pre_clip': 9.703140258789062, 'timing/data': 0.00048247165977954865, 'timing/model': 1.000720102339983, 'learning_rate/qwen_vl_interface': 1.3718889789813647e-06, 'learning_rate/action_model': 1.3718889789813646e-05, 'epoch': 2.56}) +[2026-06-18 01:24:52,001][starVLA.training.train_starvla][INFO] - Step 5649, Loss: {'train/loss': 0.1423058658838272, 'train/grad_norm_pre_clip': 24.010820388793945, 'timing/data': 0.0011469535529613495, 'timing/model': 0.9630072992295027, 'learning_rate/qwen_vl_interface': 1.370640591493193e-06, 'learning_rate/action_model': 1.370640591493193e-05, 'epoch': 2.56}) +[2026-06-18 01:24:55,875][starVLA.training.train_starvla][INFO] - Step 5650, Loss: {'train/loss': 0.22769320011138916, 'train/grad_norm_pre_clip': 14.463265419006348, 'timing/data': 0.0007012328132987022, 'timing/model': 0.9827848980203271, 'learning_rate/qwen_vl_interface': 1.3693930082004006e-06, 'learning_rate/action_model': 1.3693930082004006e-05, 'epoch': 2.56}) +[2026-06-18 01:24:59,734][starVLA.training.train_starvla][INFO] - Step 5651, Loss: {'train/loss': 0.11733291298151016, 'train/grad_norm_pre_clip': 26.898860931396484, 'timing/data': 0.0007409080862998962, 'timing/model': 0.9530088324099779, 'learning_rate/qwen_vl_interface': 1.3681462293616131e-06, 'learning_rate/action_model': 1.3681462293616131e-05, 'epoch': 2.56}) +[2026-06-18 01:25:03,729][starVLA.training.train_starvla][INFO] - Step 5652, Loss: {'train/loss': 0.025411875918507576, 'train/grad_norm_pre_clip': 3.647134304046631, 'timing/data': 0.0007596081122756004, 'timing/model': 0.9350517522543669, 'learning_rate/qwen_vl_interface': 1.3669002552352898e-06, 'learning_rate/action_model': 1.3669002552352897e-05, 'epoch': 2.56}) +[2026-06-18 01:25:07,633][starVLA.training.train_starvla][INFO] - Step 5653, Loss: {'train/loss': 0.04879848659038544, 'train/grad_norm_pre_clip': 45.55698013305664, 'timing/data': 0.0005071377381682396, 'timing/model': 1.1552818613126874, 'learning_rate/qwen_vl_interface': 1.3656550860797223e-06, 'learning_rate/action_model': 1.3656550860797223e-05, 'epoch': 2.56}) +[2026-06-18 01:25:11,291][starVLA.training.train_starvla][INFO] - Step 5654, Loss: {'train/loss': 0.24916577339172363, 'train/grad_norm_pre_clip': 33.70913314819336, 'timing/data': 0.0007988112047314644, 'timing/model': 0.9506398588418961, 'learning_rate/qwen_vl_interface': 1.3644107221530344e-06, 'learning_rate/action_model': 1.3644107221530345e-05, 'epoch': 2.57}) +[2026-06-18 01:25:15,134][starVLA.training.train_starvla][INFO] - Step 5655, Loss: {'train/loss': 0.0566570945084095, 'train/grad_norm_pre_clip': 10.4956636428833, 'timing/data': 0.0008631311357021332, 'timing/model': 0.9365009572356939, 'learning_rate/qwen_vl_interface': 1.3631671637131856e-06, 'learning_rate/action_model': 1.3631671637131855e-05, 'epoch': 2.57}) +[2026-06-18 01:25:18,981][starVLA.training.train_starvla][INFO] - Step 5656, Loss: {'train/loss': 0.2163778692483902, 'train/grad_norm_pre_clip': 22.955846786499023, 'timing/data': 0.00042245350778102875, 'timing/model': 1.0068496027961373, 'learning_rate/qwen_vl_interface': 1.3619244110179655e-06, 'learning_rate/action_model': 1.3619244110179654e-05, 'epoch': 2.57}) +[2026-06-18 01:25:22,986][starVLA.training.train_starvla][INFO] - Step 5657, Loss: {'train/loss': 0.11749153584241867, 'train/grad_norm_pre_clip': 25.063682556152344, 'timing/data': 0.0006912453100085258, 'timing/model': 0.9459781935438514, 'learning_rate/qwen_vl_interface': 1.360682464324998e-06, 'learning_rate/action_model': 1.360682464324998e-05, 'epoch': 2.57}) +[2026-06-18 01:25:26,881][starVLA.training.train_starvla][INFO] - Step 5658, Loss: {'train/loss': 0.13059823215007782, 'train/grad_norm_pre_clip': 19.876279830932617, 'timing/data': 0.0004981989040970802, 'timing/model': 0.9341716226190329, 'learning_rate/qwen_vl_interface': 1.3594413238917425e-06, 'learning_rate/action_model': 1.3594413238917425e-05, 'epoch': 2.57}) +[2026-06-18 01:25:30,664][starVLA.training.train_starvla][INFO] - Step 5659, Loss: {'train/loss': 0.19840934872627258, 'train/grad_norm_pre_clip': 21.964008331298828, 'timing/data': 0.00043365173041820526, 'timing/model': 0.922052564099431, 'learning_rate/qwen_vl_interface': 1.3582009899754864e-06, 'learning_rate/action_model': 1.3582009899754864e-05, 'epoch': 2.57}) +[2026-06-18 01:25:34,555][starVLA.training.train_starvla][INFO] - Step 5660, Loss: {'train/loss': 0.43029066920280457, 'train/grad_norm_pre_clip': 46.82194519042969, 'timing/data': 0.0004450995475053787, 'timing/model': 1.1765251830220222, 'learning_rate/qwen_vl_interface': 1.3569614628333515e-06, 'learning_rate/action_model': 1.3569614628333516e-05, 'epoch': 2.57}) +[2026-06-18 01:25:38,156][starVLA.training.train_starvla][INFO] - Step 5661, Loss: {'train/loss': 0.4381108283996582, 'train/grad_norm_pre_clip': 38.94988250732422, 'timing/data': 0.0006176037713885307, 'timing/model': 0.92716602422297, 'learning_rate/qwen_vl_interface': 1.3557227427222954e-06, 'learning_rate/action_model': 1.3557227427222954e-05, 'epoch': 2.57}) +[2026-06-18 01:25:41,946][starVLA.training.train_starvla][INFO] - Step 5662, Loss: {'train/loss': 0.23674415051937103, 'train/grad_norm_pre_clip': 32.148162841796875, 'timing/data': 0.0004431558772921562, 'timing/model': 0.9716607900336385, 'learning_rate/qwen_vl_interface': 1.3544848298991056e-06, 'learning_rate/action_model': 1.3544848298991056e-05, 'epoch': 2.57}) +[2026-06-18 01:25:45,709][starVLA.training.train_starvla][INFO] - Step 5663, Loss: {'train/loss': 0.19951415061950684, 'train/grad_norm_pre_clip': 15.378510475158691, 'timing/data': 0.00042768754065036774, 'timing/model': 0.9275073641911149, 'learning_rate/qwen_vl_interface': 1.3532477246204037e-06, 'learning_rate/action_model': 1.3532477246204036e-05, 'epoch': 2.57}) +[2026-06-18 01:25:49,628][starVLA.training.train_starvla][INFO] - Step 5664, Loss: {'train/loss': 0.22725696861743927, 'train/grad_norm_pre_clip': 29.27096176147461, 'timing/data': 0.0008171079680323601, 'timing/model': 0.9261707570403814, 'learning_rate/qwen_vl_interface': 1.3520114271426382e-06, 'learning_rate/action_model': 1.3520114271426381e-05, 'epoch': 2.57}) +[2026-06-18 01:25:53,426][starVLA.training.train_starvla][INFO] - Step 5665, Loss: {'train/loss': 0.22032372653484344, 'train/grad_norm_pre_clip': 42.98252868652344, 'timing/data': 0.0004933737218379974, 'timing/model': 0.945404976606369, 'learning_rate/qwen_vl_interface': 1.3507759377221001e-06, 'learning_rate/action_model': 1.3507759377221003e-05, 'epoch': 2.57}) +[2026-06-18 01:25:57,358][starVLA.training.train_starvla][INFO] - Step 5666, Loss: {'train/loss': 0.02340845949947834, 'train/grad_norm_pre_clip': 3.749840021133423, 'timing/data': 0.0004529273137450218, 'timing/model': 1.0284874271601439, 'learning_rate/qwen_vl_interface': 1.3495412566149058e-06, 'learning_rate/action_model': 1.3495412566149057e-05, 'epoch': 2.57}) +[2026-06-18 01:26:01,504][starVLA.training.train_starvla][INFO] - Step 5667, Loss: {'train/loss': 0.08128298074007034, 'train/grad_norm_pre_clip': 13.114907264709473, 'timing/data': 0.0006069773808121681, 'timing/model': 1.3473978955298662, 'learning_rate/qwen_vl_interface': 1.3483073840770067e-06, 'learning_rate/action_model': 1.3483073840770066e-05, 'epoch': 2.57}) +[2026-06-18 01:26:05,114][starVLA.training.train_starvla][INFO] - Step 5668, Loss: {'train/loss': 0.14276041090488434, 'train/grad_norm_pre_clip': 19.978134155273438, 'timing/data': 0.0006367452442646027, 'timing/model': 0.9372594822198153, 'learning_rate/qwen_vl_interface': 1.3470743203641846e-06, 'learning_rate/action_model': 1.3470743203641844e-05, 'epoch': 2.57}) +[2026-06-18 01:26:08,945][starVLA.training.train_starvla][INFO] - Step 5669, Loss: {'train/loss': 0.031022388488054276, 'train/grad_norm_pre_clip': 14.533690452575684, 'timing/data': 0.0008332710713148117, 'timing/model': 0.9336311323568225, 'learning_rate/qwen_vl_interface': 1.3458420657320555e-06, 'learning_rate/action_model': 1.3458420657320555e-05, 'epoch': 2.57}) +[2026-06-18 01:26:12,735][starVLA.training.train_starvla][INFO] - Step 5670, Loss: {'train/loss': 0.16358113288879395, 'train/grad_norm_pre_clip': 21.356060028076172, 'timing/data': 0.0004887823015451431, 'timing/model': 0.952321550808847, 'learning_rate/qwen_vl_interface': 1.344610620436067e-06, 'learning_rate/action_model': 1.344610620436067e-05, 'epoch': 2.57}) +[2026-06-18 01:26:16,834][starVLA.training.train_starvla][INFO] - Step 5671, Loss: {'train/loss': 0.08813431859016418, 'train/grad_norm_pre_clip': 19.980228424072266, 'timing/data': 0.00048432033509016037, 'timing/model': 1.1705832835286856, 'learning_rate/qwen_vl_interface': 1.3433799847314989e-06, 'learning_rate/action_model': 1.3433799847314988e-05, 'epoch': 2.57}) +[2026-06-18 01:26:20,648][starVLA.training.train_starvla][INFO] - Step 5672, Loss: {'train/loss': 0.21959426999092102, 'train/grad_norm_pre_clip': 38.21464538574219, 'timing/data': 0.0008087828755378723, 'timing/model': 0.9504532730206847, 'learning_rate/qwen_vl_interface': 1.3421501588734655e-06, 'learning_rate/action_model': 1.3421501588734656e-05, 'epoch': 2.57}) +[2026-06-18 01:26:24,487][starVLA.training.train_starvla][INFO] - Step 5673, Loss: {'train/loss': 0.3848271369934082, 'train/grad_norm_pre_clip': 30.481033325195312, 'timing/data': 0.00044571608304977417, 'timing/model': 0.9479943849146366, 'learning_rate/qwen_vl_interface': 1.3409211431169084e-06, 'learning_rate/action_model': 1.3409211431169083e-05, 'epoch': 2.57}) +[2026-06-18 01:26:28,492][starVLA.training.train_starvla][INFO] - Step 5674, Loss: {'train/loss': 0.1491844356060028, 'train/grad_norm_pre_clip': 20.818044662475586, 'timing/data': 0.0005125533789396286, 'timing/model': 1.314583020284772, 'learning_rate/qwen_vl_interface': 1.3396929377166042e-06, 'learning_rate/action_model': 1.3396929377166043e-05, 'epoch': 2.57}) +[2026-06-18 01:26:32,143][starVLA.training.train_starvla][INFO] - Step 5675, Loss: {'train/loss': 0.06695470958948135, 'train/grad_norm_pre_clip': 14.779410362243652, 'timing/data': 0.0007799174636602402, 'timing/model': 0.930428515188396, 'learning_rate/qwen_vl_interface': 1.338465542927161e-06, 'learning_rate/action_model': 1.3384655429271608e-05, 'epoch': 2.57}) +[2026-06-18 01:26:36,006][starVLA.training.train_starvla][INFO] - Step 5676, Loss: {'train/loss': 0.31413835287094116, 'train/grad_norm_pre_clip': 48.101810455322266, 'timing/data': 0.0004428848624229431, 'timing/model': 0.9399367459118366, 'learning_rate/qwen_vl_interface': 1.3372389590030204e-06, 'learning_rate/action_model': 1.3372389590030204e-05, 'epoch': 2.58}) +[2026-06-18 01:26:39,819][starVLA.training.train_starvla][INFO] - Step 5677, Loss: {'train/loss': 0.22742106020450592, 'train/grad_norm_pre_clip': 24.034822463989258, 'timing/data': 0.00045892875641584396, 'timing/model': 0.9414339549839497, 'learning_rate/qwen_vl_interface': 1.3360131861984551e-06, 'learning_rate/action_model': 1.3360131861984552e-05, 'epoch': 2.58}) +[2026-06-18 01:26:43,716][starVLA.training.train_starvla][INFO] - Step 5678, Loss: {'train/loss': 0.506475031375885, 'train/grad_norm_pre_clip': 38.020267486572266, 'timing/data': 0.000458364374935627, 'timing/model': 0.9835896929726005, 'learning_rate/qwen_vl_interface': 1.3347882247675662e-06, 'learning_rate/action_model': 1.3347882247675661e-05, 'epoch': 2.58}) +[2026-06-18 01:26:47,621][starVLA.training.train_starvla][INFO] - Step 5679, Loss: {'train/loss': 0.03630002588033676, 'train/grad_norm_pre_clip': 10.52220630645752, 'timing/data': 0.0004906309768557549, 'timing/model': 0.9466433981433511, 'learning_rate/qwen_vl_interface': 1.3335640749642923e-06, 'learning_rate/action_model': 1.3335640749642923e-05, 'epoch': 2.58}) +[2026-06-18 01:26:51,485][starVLA.training.train_starvla][INFO] - Step 5680, Loss: {'train/loss': 0.2831732928752899, 'train/grad_norm_pre_clip': 27.297910690307617, 'timing/data': 0.0005275066941976547, 'timing/model': 0.9359002970159054, 'learning_rate/qwen_vl_interface': 1.3323407370423993e-06, 'learning_rate/action_model': 1.3323407370423994e-05, 'epoch': 2.58}) +[2026-06-18 01:26:55,410][starVLA.training.train_starvla][INFO] - Step 5681, Loss: {'train/loss': 0.05573564022779465, 'train/grad_norm_pre_clip': 13.588133811950684, 'timing/data': 0.0005446961149573326, 'timing/model': 1.1048241965472698, 'learning_rate/qwen_vl_interface': 1.3311182112554877e-06, 'learning_rate/action_model': 1.3311182112554876e-05, 'epoch': 2.58}) +[2026-06-18 01:27:00,018][starVLA.training.train_starvla][INFO] - Step 5682, Loss: {'train/loss': 0.06941673159599304, 'train/grad_norm_pre_clip': 22.512937545776367, 'timing/data': 0.0012222351506352425, 'timing/model': 1.1245622457936406, 'learning_rate/qwen_vl_interface': 1.3298964978569873e-06, 'learning_rate/action_model': 1.3298964978569872e-05, 'epoch': 2.58}) +[2026-06-18 01:27:03,877][starVLA.training.train_starvla][INFO] - Step 5683, Loss: {'train/loss': 0.09020502865314484, 'train/grad_norm_pre_clip': 27.98015785217285, 'timing/data': 0.0007552178576588631, 'timing/model': 0.9488118290901184, 'learning_rate/qwen_vl_interface': 1.3286755971001618e-06, 'learning_rate/action_model': 1.3286755971001618e-05, 'epoch': 2.58}) +[2026-06-18 01:27:08,503][starVLA.training.train_starvla][INFO] - Step 5684, Loss: {'train/loss': 0.43840429186820984, 'train/grad_norm_pre_clip': 50.170989990234375, 'timing/data': 0.0004660375416278839, 'timing/model': 0.9538453258574009, 'learning_rate/qwen_vl_interface': 1.327455509238104e-06, 'learning_rate/action_model': 1.3274555092381039e-05, 'epoch': 2.58}) +[2026-06-18 01:27:12,390][starVLA.training.train_starvla][INFO] - Step 5685, Loss: {'train/loss': 0.020383115857839584, 'train/grad_norm_pre_clip': 4.028287410736084, 'timing/data': 0.0005120532587170601, 'timing/model': 0.9366547306999564, 'learning_rate/qwen_vl_interface': 1.32623623452374e-06, 'learning_rate/action_model': 1.32623623452374e-05, 'epoch': 2.58}) +[2026-06-18 01:27:16,428][starVLA.training.train_starvla][INFO] - Step 5686, Loss: {'train/loss': 0.17029090225696564, 'train/grad_norm_pre_clip': 23.304418563842773, 'timing/data': 0.001061093993484974, 'timing/model': 0.9372053183615208, 'learning_rate/qwen_vl_interface': 1.325017773209827e-06, 'learning_rate/action_model': 1.325017773209827e-05, 'epoch': 2.58}) +[2026-06-18 01:27:20,213][starVLA.training.train_starvla][INFO] - Step 5687, Loss: {'train/loss': 0.4825981557369232, 'train/grad_norm_pre_clip': 26.131196975708008, 'timing/data': 0.0005010729655623436, 'timing/model': 0.9424814451485872, 'learning_rate/qwen_vl_interface': 1.323800125548953e-06, 'learning_rate/action_model': 1.323800125548953e-05, 'epoch': 2.58}) +[2026-06-18 01:27:24,389][starVLA.training.train_starvla][INFO] - Step 5688, Loss: {'train/loss': 0.32835742831230164, 'train/grad_norm_pre_clip': 39.342708587646484, 'timing/data': 0.0006605293601751328, 'timing/model': 1.1288137445226312, 'learning_rate/qwen_vl_interface': 1.3225832917935376e-06, 'learning_rate/action_model': 1.3225832917935377e-05, 'epoch': 2.58}) +[2026-06-18 01:27:32,643][starVLA.training.train_starvla][INFO] - Step 5689, Loss: {'train/loss': 0.25739309191703796, 'train/grad_norm_pre_clip': 35.43885803222656, 'timing/data': 0.0006526652723550797, 'timing/model': 0.9350680913776159, 'learning_rate/qwen_vl_interface': 1.3213672721958309e-06, 'learning_rate/action_model': 1.321367272195831e-05, 'epoch': 2.58}) +[2026-06-18 01:27:36,453][starVLA.training.train_starvla][INFO] - Step 5690, Loss: {'train/loss': 0.25600603222846985, 'train/grad_norm_pre_clip': 37.515586853027344, 'timing/data': 0.0007954537868499756, 'timing/model': 0.9598013004288077, 'learning_rate/qwen_vl_interface': 1.3201520670079172e-06, 'learning_rate/action_model': 1.3201520670079171e-05, 'epoch': 2.58}) +[2026-06-18 01:27:40,235][starVLA.training.train_starvla][INFO] - Step 5691, Loss: {'train/loss': 0.2905897796154022, 'train/grad_norm_pre_clip': 37.11112976074219, 'timing/data': 0.000452578067779541, 'timing/model': 0.9188377661630511, 'learning_rate/qwen_vl_interface': 1.31893767648171e-06, 'learning_rate/action_model': 1.3189376764817099e-05, 'epoch': 2.58}) +[2026-06-18 01:27:44,025][starVLA.training.train_starvla][INFO] - Step 5692, Loss: {'train/loss': 0.053534798324108124, 'train/grad_norm_pre_clip': 10.145845413208008, 'timing/data': 0.0004883222281932831, 'timing/model': 0.9395826691761613, 'learning_rate/qwen_vl_interface': 1.3177241008689513e-06, 'learning_rate/action_model': 1.3177241008689511e-05, 'epoch': 2.58}) +[2026-06-18 01:27:47,845][starVLA.training.train_starvla][INFO] - Step 5693, Loss: {'train/loss': 0.1182507798075676, 'train/grad_norm_pre_clip': 21.122644424438477, 'timing/data': 0.000507790595293045, 'timing/model': 0.9242941988632083, 'learning_rate/qwen_vl_interface': 1.316511340421216e-06, 'learning_rate/action_model': 1.316511340421216e-05, 'epoch': 2.58}) +[2026-06-18 01:27:51,597][starVLA.training.train_starvla][INFO] - Step 5694, Loss: {'train/loss': 0.23566290736198425, 'train/grad_norm_pre_clip': 16.890443801879883, 'timing/data': 0.0007622521370649338, 'timing/model': 0.9421289088204503, 'learning_rate/qwen_vl_interface': 1.315299395389914e-06, 'learning_rate/action_model': 1.315299395389914e-05, 'epoch': 2.58}) +[2026-06-18 01:27:55,585][starVLA.training.train_starvla][INFO] - Step 5695, Loss: {'train/loss': 0.08450207114219666, 'train/grad_norm_pre_clip': 34.14179611206055, 'timing/data': 0.0008174600079655647, 'timing/model': 1.1884007649496198, 'learning_rate/qwen_vl_interface': 1.3140882660262818e-06, 'learning_rate/action_model': 1.3140882660262818e-05, 'epoch': 2.58}) +[2026-06-18 01:27:59,209][starVLA.training.train_starvla][INFO] - Step 5696, Loss: {'train/loss': 0.38644909858703613, 'train/grad_norm_pre_clip': 30.805513381958008, 'timing/data': 0.0004893476143479347, 'timing/model': 0.938555620610714, 'learning_rate/qwen_vl_interface': 1.3128779525813865e-06, 'learning_rate/action_model': 1.3128779525813864e-05, 'epoch': 2.58}) +[2026-06-18 01:28:03,101][starVLA.training.train_starvla][INFO] - Step 5697, Loss: {'train/loss': 0.08102001249790192, 'train/grad_norm_pre_clip': 13.754525184631348, 'timing/data': 0.0004830872640013695, 'timing/model': 1.0254473118111491, 'learning_rate/qwen_vl_interface': 1.3116684553061288e-06, 'learning_rate/action_model': 1.3116684553061287e-05, 'epoch': 2.58}) +[2026-06-18 01:28:09,285][starVLA.training.train_starvla][INFO] - Step 5698, Loss: {'train/loss': 0.20353461802005768, 'train/grad_norm_pre_clip': 35.4149284362793, 'timing/data': 0.00048389192670583725, 'timing/model': 0.9375067669898272, 'learning_rate/qwen_vl_interface': 1.310459774451238e-06, 'learning_rate/action_model': 1.3104597744512379e-05, 'epoch': 2.59}) +[2026-06-18 01:28:13,361][starVLA.training.train_starvla][INFO] - Step 5699, Loss: {'train/loss': 0.41342273354530334, 'train/grad_norm_pre_clip': 34.03193283081055, 'timing/data': 0.0007395688444375992, 'timing/model': 0.9715006984770298, 'learning_rate/qwen_vl_interface': 1.3092519102672744e-06, 'learning_rate/action_model': 1.3092519102672742e-05, 'epoch': 2.59}) +[2026-06-18 01:30:30,774][starVLA.training.train_starvla][INFO] - Step 5700, Loss: {'train/loss': 0.103763647377491, 'train/grad_norm_pre_clip': 27.69197654724121, 'eval/loss': 0.2625134587287903, 'eval/latency_8/loss': 0.2625134587287903, 'eval/demon_fire_precision': 0.9690374003678724, 'eval/demon_fire_recall': 0.9819819819819819, 'eval/demon_fire_f1': 0.9754667489584941, 'eval/demon_fire_rate_model': 0.907623817473567, 'eval/demon_fire_rate_teacher': 0.8956594323873122, 'eval/demon_left_precision': 0.914004914004914, 'eval/demon_left_recall': 0.7965738758029979, 'eval/demon_left_f1': 0.851258581235698, 'eval/demon_left_rate_model': 0.1132442960489705, 'eval/demon_left_rate_teacher': 0.12993878686700056, 'eval/demon_right_precision': 0.8786936236391913, 'eval/demon_right_recall': 0.8560606060606061, 'eval/demon_right_f1': 0.8672294704528012, 'eval/demon_right_rate_model': 0.17890929326655536, 'eval/demon_right_rate_teacher': 0.18363939899833054, 'eval/demon_move_f1': 0.8592440258442495, 'eval/demon_cc_f1': 0.937427370691127, 'eval/latency_8/demon_fire_precision': 0.9690374003678724, 'eval/latency_8/demon_fire_recall': 0.9819819819819819, 'eval/latency_8/demon_fire_f1': 0.9754667489584941, 'eval/latency_8/demon_fire_rate_model': 0.907623817473567, 'eval/latency_8/demon_fire_rate_teacher': 0.8956594323873122, 'eval/latency_8/demon_left_precision': 0.914004914004914, 'eval/latency_8/demon_left_recall': 0.7965738758029979, 'eval/latency_8/demon_left_f1': 0.851258581235698, 'eval/latency_8/demon_left_rate_model': 0.1132442960489705, 'eval/latency_8/demon_left_rate_teacher': 0.12993878686700056, 'eval/latency_8/demon_right_precision': 0.8786936236391913, 'eval/latency_8/demon_right_recall': 0.8560606060606061, 'eval/latency_8/demon_right_f1': 0.8672294704528012, 'eval/latency_8/demon_right_rate_model': 0.17890929326655536, 'eval/latency_8/demon_right_rate_teacher': 0.18363939899833054, 'eval/latency_8/demon_move_f1': 0.8592440258442495, 'eval/latency_8/demon_cc_f1': 0.937427370691127, 'timing/data': 0.0004883278161287308, 'timing/model': 1.7909821141511202, 'learning_rate/qwen_vl_interface': 1.3080448630046312e-06, 'learning_rate/action_model': 1.308044863004631e-05, 'epoch': 2.59}) +[2026-06-18 01:30:30,776][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_5700_state +[2026-06-18 01:30:30,781][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 01:30:37,509][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_5700_state/model.safetensors +[2026-06-18 01:30:51,652][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_5700_state/optimizer.bin +[2026-06-18 01:30:51,653][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_5700_state/sampler.bin +[2026-06-18 01:30:51,653][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_5700_state/sampler_1.bin +[2026-06-18 01:30:51,654][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_5700_state/random_states_0.pkl +[2026-06-18 01:30:51,658][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 5700 +[2026-06-18 01:30:51,658][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 01:30:51,669][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 01:30:55,528][starVLA.training.train_starvla][INFO] - Step 5701, Loss: {'train/loss': 0.11229334026575089, 'train/grad_norm_pre_clip': 21.36223030090332, 'timing/data': 0.0011441241949796677, 'timing/model': 0.955936124548316, 'learning_rate/qwen_vl_interface': 1.3068386329135287e-06, 'learning_rate/action_model': 1.3068386329135287e-05, 'epoch': 2.59}) +[2026-06-18 01:30:59,264][starVLA.training.train_starvla][INFO] - Step 5702, Loss: {'train/loss': 0.2148052304983139, 'train/grad_norm_pre_clip': 40.649932861328125, 'timing/data': 0.0010966062545776367, 'timing/model': 0.925633211620152, 'learning_rate/qwen_vl_interface': 1.3056332202440211e-06, 'learning_rate/action_model': 1.305633220244021e-05, 'epoch': 2.59}) +[2026-06-18 01:31:03,005][starVLA.training.train_starvla][INFO] - Step 5703, Loss: {'train/loss': 0.09287670254707336, 'train/grad_norm_pre_clip': 18.071151733398438, 'timing/data': 0.0009041819721460342, 'timing/model': 0.9287348380312324, 'learning_rate/qwen_vl_interface': 1.3044286252459914e-06, 'learning_rate/action_model': 1.3044286252459915e-05, 'epoch': 2.59}) +[2026-06-18 01:31:06,814][starVLA.training.train_starvla][INFO] - Step 5704, Loss: {'train/loss': 0.15820303559303284, 'train/grad_norm_pre_clip': 28.273958206176758, 'timing/data': 0.00084720179438591, 'timing/model': 1.1149488119408488, 'learning_rate/qwen_vl_interface': 1.3032248481691523e-06, 'learning_rate/action_model': 1.3032248481691523e-05, 'epoch': 2.59}) +[2026-06-18 01:31:10,433][starVLA.training.train_starvla][INFO] - Step 5705, Loss: {'train/loss': 0.04326556995511055, 'train/grad_norm_pre_clip': 9.541494369506836, 'timing/data': 0.0005119424313306808, 'timing/model': 0.957520168274641, 'learning_rate/qwen_vl_interface': 1.302021889263051e-06, 'learning_rate/action_model': 1.3020218892630508e-05, 'epoch': 2.59}) +[2026-06-18 01:31:14,252][starVLA.training.train_starvla][INFO] - Step 5706, Loss: {'train/loss': 0.35375213623046875, 'train/grad_norm_pre_clip': 29.21575927734375, 'timing/data': 0.0008912123739719391, 'timing/model': 0.9361472772434354, 'learning_rate/qwen_vl_interface': 1.3008197487770593e-06, 'learning_rate/action_model': 1.3008197487770593e-05, 'epoch': 2.59}) +[2026-06-18 01:31:18,992][starVLA.training.train_starvla][INFO] - Step 5707, Loss: {'train/loss': 0.21171346306800842, 'train/grad_norm_pre_clip': 36.22182846069336, 'timing/data': 0.00047305505722761154, 'timing/model': 1.9576222095638514, 'learning_rate/qwen_vl_interface': 1.2996184269603835e-06, 'learning_rate/action_model': 1.2996184269603833e-05, 'epoch': 2.59}) +[2026-06-18 01:31:23,101][starVLA.training.train_starvla][INFO] - Step 5708, Loss: {'train/loss': 0.05400031805038452, 'train/grad_norm_pre_clip': 12.925955772399902, 'timing/data': 0.0005192598327994347, 'timing/model': 0.9697081698104739, 'learning_rate/qwen_vl_interface': 1.2984179240620567e-06, 'learning_rate/action_model': 1.2984179240620567e-05, 'epoch': 2.59}) +[2026-06-18 01:31:27,104][starVLA.training.train_starvla][INFO] - Step 5709, Loss: {'train/loss': 0.2175586074590683, 'train/grad_norm_pre_clip': 16.646860122680664, 'timing/data': 0.00044903531670570374, 'timing/model': 0.941820053383708, 'learning_rate/qwen_vl_interface': 1.2972182403309474e-06, 'learning_rate/action_model': 1.2972182403309474e-05, 'epoch': 2.59}) +[2026-06-18 01:31:31,105][starVLA.training.train_starvla][INFO] - Step 5710, Loss: {'train/loss': 0.39459002017974854, 'train/grad_norm_pre_clip': 32.38755416870117, 'timing/data': 0.0004953835159540176, 'timing/model': 1.145303855650127, 'learning_rate/qwen_vl_interface': 1.2960193760157507e-06, 'learning_rate/action_model': 1.2960193760157507e-05, 'epoch': 2.59}) +[2026-06-18 01:31:34,708][starVLA.training.train_starvla][INFO] - Step 5711, Loss: {'train/loss': 0.05889969319105148, 'train/grad_norm_pre_clip': 13.371000289916992, 'timing/data': 0.0004988806322216988, 'timing/model': 0.9362116064876318, 'learning_rate/qwen_vl_interface': 1.294821331364992e-06, 'learning_rate/action_model': 1.294821331364992e-05, 'epoch': 2.59}) +[2026-06-18 01:31:38,625][starVLA.training.train_starvla][INFO] - Step 5712, Loss: {'train/loss': 0.15929654240608215, 'train/grad_norm_pre_clip': 29.41705894470215, 'timing/data': 0.0009224154055118561, 'timing/model': 0.9345344072207808, 'learning_rate/qwen_vl_interface': 1.2936241066270278e-06, 'learning_rate/action_model': 1.2936241066270278e-05, 'epoch': 2.59}) +[2026-06-18 01:31:42,506][starVLA.training.train_starvla][INFO] - Step 5713, Loss: {'train/loss': 0.209927037358284, 'train/grad_norm_pre_clip': 23.504165649414062, 'timing/data': 0.000496499240398407, 'timing/model': 0.9412593701854348, 'learning_rate/qwen_vl_interface': 1.2924277020500427e-06, 'learning_rate/action_model': 1.2924277020500428e-05, 'epoch': 2.59}) +[2026-06-18 01:31:46,289][starVLA.training.train_starvla][INFO] - Step 5714, Loss: {'train/loss': 0.713498592376709, 'train/grad_norm_pre_clip': 31.55320930480957, 'timing/data': 0.0004922924563288689, 'timing/model': 0.9450566852465272, 'learning_rate/qwen_vl_interface': 1.291232117882054e-06, 'learning_rate/action_model': 1.2912321178820538e-05, 'epoch': 2.59}) +[2026-06-18 01:31:50,083][starVLA.training.train_starvla][INFO] - Step 5715, Loss: {'train/loss': 0.08918548375368118, 'train/grad_norm_pre_clip': 23.460416793823242, 'timing/data': 0.0004573613405227661, 'timing/model': 0.9377668211236596, 'learning_rate/qwen_vl_interface': 1.2900373543709065e-06, 'learning_rate/action_model': 1.2900373543709065e-05, 'epoch': 2.59}) +[2026-06-18 01:31:53,965][starVLA.training.train_starvla][INFO] - Step 5716, Loss: {'train/loss': 0.016996143385767937, 'train/grad_norm_pre_clip': 8.33353328704834, 'timing/data': 0.0006679091602563858, 'timing/model': 1.160625689662993, 'learning_rate/qwen_vl_interface': 1.288843411764277e-06, 'learning_rate/action_model': 1.288843411764277e-05, 'epoch': 2.59}) +[2026-06-18 01:31:57,768][starVLA.training.train_starvla][INFO] - Step 5717, Loss: {'train/loss': 0.072136290371418, 'train/grad_norm_pre_clip': 17.995344161987305, 'timing/data': 0.0005375109612941742, 'timing/model': 1.1019884068518877, 'learning_rate/qwen_vl_interface': 1.2876502903096709e-06, 'learning_rate/action_model': 1.2876502903096708e-05, 'epoch': 2.59}) +[2026-06-18 01:32:01,443][starVLA.training.train_starvla][INFO] - Step 5718, Loss: {'train/loss': 0.2180936187505722, 'train/grad_norm_pre_clip': 21.527973175048828, 'timing/data': 0.000453827902674675, 'timing/model': 0.9591000899672508, 'learning_rate/qwen_vl_interface': 1.286457990254423e-06, 'learning_rate/action_model': 1.286457990254423e-05, 'epoch': 2.59}) +[2026-06-18 01:32:05,449][starVLA.training.train_starvla][INFO] - Step 5719, Loss: {'train/loss': 0.15025177597999573, 'train/grad_norm_pre_clip': 27.260347366333008, 'timing/data': 0.0004829550161957741, 'timing/model': 0.9873384283855557, 'learning_rate/qwen_vl_interface': 1.2852665118456971e-06, 'learning_rate/action_model': 1.2852665118456972e-05, 'epoch': 2.59}) +[2026-06-18 01:32:09,263][starVLA.training.train_starvla][INFO] - Step 5720, Loss: {'train/loss': 0.2350921779870987, 'train/grad_norm_pre_clip': 39.675899505615234, 'timing/data': 0.0004646517336368561, 'timing/model': 0.9365208456292748, 'learning_rate/qwen_vl_interface': 1.2840758553304927e-06, 'learning_rate/action_model': 1.2840758553304927e-05, 'epoch': 2.6}) +[2026-06-18 01:32:13,169][starVLA.training.train_starvla][INFO] - Step 5721, Loss: {'train/loss': 0.19026756286621094, 'train/grad_norm_pre_clip': 35.852230072021484, 'timing/data': 0.0004747360944747925, 'timing/model': 0.9366470389068127, 'learning_rate/qwen_vl_interface': 1.2828860209556293e-06, 'learning_rate/action_model': 1.2828860209556292e-05, 'epoch': 2.6}) +[2026-06-18 01:32:17,001][starVLA.training.train_starvla][INFO] - Step 5722, Loss: {'train/loss': 0.3116178810596466, 'train/grad_norm_pre_clip': 39.290950775146484, 'timing/data': 0.0007126759737730026, 'timing/model': 0.9517458071932197, 'learning_rate/qwen_vl_interface': 1.2816970089677617e-06, 'learning_rate/action_model': 1.2816970089677615e-05, 'epoch': 2.6}) +[2026-06-18 01:32:20,838][starVLA.training.train_starvla][INFO] - Step 5723, Loss: {'train/loss': 0.25324738025665283, 'train/grad_norm_pre_clip': 25.140172958374023, 'timing/data': 0.0005132546648383141, 'timing/model': 1.116768615320325, 'learning_rate/qwen_vl_interface': 1.280508819613375e-06, 'learning_rate/action_model': 1.280508819613375e-05, 'epoch': 2.6}) +[2026-06-18 01:32:24,725][starVLA.training.train_starvla][INFO] - Step 5724, Loss: {'train/loss': 0.29157233238220215, 'train/grad_norm_pre_clip': 22.29022789001465, 'timing/data': 0.0005173487588763237, 'timing/model': 1.1813336899504066, 'learning_rate/qwen_vl_interface': 1.2793214531387812e-06, 'learning_rate/action_model': 1.2793214531387812e-05, 'epoch': 2.6}) +[2026-06-18 01:32:28,314][starVLA.training.train_starvla][INFO] - Step 5725, Loss: {'train/loss': 0.21537598967552185, 'train/grad_norm_pre_clip': 68.8281478881836, 'timing/data': 0.000523625873029232, 'timing/model': 0.9391414439305663, 'learning_rate/qwen_vl_interface': 1.2781349097901246e-06, 'learning_rate/action_model': 1.2781349097901247e-05, 'epoch': 2.6}) +[2026-06-18 01:32:32,113][starVLA.training.train_starvla][INFO] - Step 5726, Loss: {'train/loss': 0.15413609147071838, 'train/grad_norm_pre_clip': 37.742103576660156, 'timing/data': 0.0004997663199901581, 'timing/model': 0.9467411367222667, 'learning_rate/qwen_vl_interface': 1.2769491898133723e-06, 'learning_rate/action_model': 1.2769491898133722e-05, 'epoch': 2.6}) +[2026-06-18 01:32:35,924][starVLA.training.train_starvla][INFO] - Step 5727, Loss: {'train/loss': 0.07142319530248642, 'train/grad_norm_pre_clip': 15.198690414428711, 'timing/data': 0.000558151863515377, 'timing/model': 0.9426738126203418, 'learning_rate/qwen_vl_interface': 1.2757642934543293e-06, 'learning_rate/action_model': 1.2757642934543293e-05, 'epoch': 2.6}) +[2026-06-18 01:32:39,684][starVLA.training.train_starvla][INFO] - Step 5728, Loss: {'train/loss': 0.17265157401561737, 'train/grad_norm_pre_clip': 24.891868591308594, 'timing/data': 0.0004631243646144867, 'timing/model': 0.9349415516480803, 'learning_rate/qwen_vl_interface': 1.274580220958625e-06, 'learning_rate/action_model': 1.274580220958625e-05, 'epoch': 2.6}) +[2026-06-18 01:32:43,583][starVLA.training.train_starvla][INFO] - Step 5729, Loss: {'train/loss': 0.3315849304199219, 'train/grad_norm_pre_clip': 26.073766708374023, 'timing/data': 0.0004871618002653122, 'timing/model': 0.9391702637076378, 'learning_rate/qwen_vl_interface': 1.2733969725717184e-06, 'learning_rate/action_model': 1.2733969725717184e-05, 'epoch': 2.6}) +[2026-06-18 01:32:47,383][starVLA.training.train_starvla][INFO] - Step 5730, Loss: {'train/loss': 0.011366624385118484, 'train/grad_norm_pre_clip': 2.5637402534484863, 'timing/data': 0.00048276595771312714, 'timing/model': 1.1123876655474305, 'learning_rate/qwen_vl_interface': 1.272214548538899e-06, 'learning_rate/action_model': 1.2722145485388989e-05, 'epoch': 2.6}) +[2026-06-18 01:32:51,018][starVLA.training.train_starvla][INFO] - Step 5731, Loss: {'train/loss': 0.3166063725948334, 'train/grad_norm_pre_clip': 36.45722579956055, 'timing/data': 0.0004817824810743332, 'timing/model': 0.937990247271955, 'learning_rate/qwen_vl_interface': 1.2710329491052834e-06, 'learning_rate/action_model': 1.2710329491052833e-05, 'epoch': 2.6}) +[2026-06-18 01:32:54,922][starVLA.training.train_starvla][INFO] - Step 5732, Loss: {'train/loss': 0.0845852866768837, 'train/grad_norm_pre_clip': 23.44341278076172, 'timing/data': 0.001390143297612667, 'timing/model': 0.932680381461978, 'learning_rate/qwen_vl_interface': 1.2698521745158188e-06, 'learning_rate/action_model': 1.2698521745158187e-05, 'epoch': 2.6}) +[2026-06-18 01:32:58,724][starVLA.training.train_starvla][INFO] - Step 5733, Loss: {'train/loss': 0.09371288865804672, 'train/grad_norm_pre_clip': 16.73291015625, 'timing/data': 0.0012649400159716606, 'timing/model': 0.9212433379143476, 'learning_rate/qwen_vl_interface': 1.2686722250152804e-06, 'learning_rate/action_model': 1.2686722250152806e-05, 'epoch': 2.6}) +[2026-06-18 01:33:02,698][starVLA.training.train_starvla][INFO] - Step 5734, Loss: {'train/loss': 0.16384361684322357, 'train/grad_norm_pre_clip': 25.51272201538086, 'timing/data': 0.0004731602966785431, 'timing/model': 0.9434374840930104, 'learning_rate/qwen_vl_interface': 1.2674931008482763e-06, 'learning_rate/action_model': 1.2674931008482763e-05, 'epoch': 2.6}) +[2026-06-18 01:33:07,724][starVLA.training.train_starvla][INFO] - Step 5735, Loss: {'train/loss': 0.1927640289068222, 'train/grad_norm_pre_clip': 43.84165954589844, 'timing/data': 0.0005156686529517174, 'timing/model': 0.9661792796105146, 'learning_rate/qwen_vl_interface': 1.2663148022592365e-06, 'learning_rate/action_model': 1.2663148022592363e-05, 'epoch': 2.6}) +[2026-06-18 01:33:11,596][starVLA.training.train_starvla][INFO] - Step 5736, Loss: {'train/loss': 0.1993694007396698, 'train/grad_norm_pre_clip': 34.577392578125, 'timing/data': 0.001012871041893959, 'timing/model': 1.1549210995435715, 'learning_rate/qwen_vl_interface': 1.2651373294924249e-06, 'learning_rate/action_model': 1.2651373294924248e-05, 'epoch': 2.6}) +[2026-06-18 01:33:16,805][starVLA.training.train_starvla][INFO] - Step 5737, Loss: {'train/loss': 0.07826138287782669, 'train/grad_norm_pre_clip': 17.378644943237305, 'timing/data': 0.0008880160748958588, 'timing/model': 1.105202387087047, 'learning_rate/qwen_vl_interface': 1.2639606827919313e-06, 'learning_rate/action_model': 1.2639606827919311e-05, 'epoch': 2.6}) +[2026-06-18 01:33:21,305][starVLA.training.train_starvla][INFO] - Step 5738, Loss: {'train/loss': 0.10798653960227966, 'train/grad_norm_pre_clip': 21.881263732910156, 'timing/data': 0.00042020995169878006, 'timing/model': 0.9506075382232666, 'learning_rate/qwen_vl_interface': 1.2627848624016788e-06, 'learning_rate/action_model': 1.2627848624016788e-05, 'epoch': 2.6}) +[2026-06-18 01:33:27,019][starVLA.training.train_starvla][INFO] - Step 5739, Loss: {'train/loss': 0.0709153562784195, 'train/grad_norm_pre_clip': 20.876977920532227, 'timing/data': 0.0004514157772064209, 'timing/model': 0.942663456313312, 'learning_rate/qwen_vl_interface': 1.261609868565416e-06, 'learning_rate/action_model': 1.2616098685654159e-05, 'epoch': 2.6}) +[2026-06-18 01:33:30,822][starVLA.training.train_starvla][INFO] - Step 5740, Loss: {'train/loss': 0.42913421988487244, 'train/grad_norm_pre_clip': 40.406700134277344, 'timing/data': 0.0007099593058228493, 'timing/model': 0.9515972528606653, 'learning_rate/qwen_vl_interface': 1.2604357015267165e-06, 'learning_rate/action_model': 1.2604357015267166e-05, 'epoch': 2.6}) +[2026-06-18 01:33:34,710][starVLA.training.train_starvla][INFO] - Step 5741, Loss: {'train/loss': 0.5905506610870361, 'train/grad_norm_pre_clip': 38.81455993652344, 'timing/data': 0.00045847147703170776, 'timing/model': 0.9442876596003771, 'learning_rate/qwen_vl_interface': 1.2592623615289906e-06, 'learning_rate/action_model': 1.2592623615289906e-05, 'epoch': 2.6}) +[2026-06-18 01:33:38,485][starVLA.training.train_starvla][INFO] - Step 5742, Loss: {'train/loss': 0.17037692666053772, 'train/grad_norm_pre_clip': 30.407978057861328, 'timing/data': 0.0004341639578342438, 'timing/model': 0.9342998927459121, 'learning_rate/qwen_vl_interface': 1.258089848815471e-06, 'learning_rate/action_model': 1.2580898488154708e-05, 'epoch': 2.6}) +[2026-06-18 01:33:42,283][starVLA.training.train_starvla][INFO] - Step 5743, Loss: {'train/loss': 0.16195715963840485, 'train/grad_norm_pre_clip': 20.278324127197266, 'timing/data': 0.0004248563200235367, 'timing/model': 1.1208931216970086, 'learning_rate/qwen_vl_interface': 1.256918163629221e-06, 'learning_rate/action_model': 1.256918163629221e-05, 'epoch': 2.61}) +[2026-06-18 01:33:46,125][starVLA.training.train_starvla][INFO] - Step 5744, Loss: {'train/loss': 0.09910636395215988, 'train/grad_norm_pre_clip': 13.426029205322266, 'timing/data': 0.00046390388160943985, 'timing/model': 1.103320344351232, 'learning_rate/qwen_vl_interface': 1.2557473062131332e-06, 'learning_rate/action_model': 1.2557473062131331e-05, 'epoch': 2.61}) +[2026-06-18 01:33:49,770][starVLA.training.train_starvla][INFO] - Step 5745, Loss: {'train/loss': 0.08564144372940063, 'train/grad_norm_pre_clip': 141.67311096191406, 'timing/data': 0.0004908842965960503, 'timing/model': 0.9319231370463967, 'learning_rate/qwen_vl_interface': 1.2545772768099263e-06, 'learning_rate/action_model': 1.2545772768099263e-05, 'epoch': 2.61}) +[2026-06-18 01:33:53,563][starVLA.training.train_starvla][INFO] - Step 5746, Loss: {'train/loss': 0.027076423168182373, 'train/grad_norm_pre_clip': 5.817987442016602, 'timing/data': 0.00043793488293886185, 'timing/model': 0.930389310233295, 'learning_rate/qwen_vl_interface': 1.2534080756621488e-06, 'learning_rate/action_model': 1.2534080756621488e-05, 'epoch': 2.61}) +[2026-06-18 01:33:57,427][starVLA.training.train_starvla][INFO] - Step 5747, Loss: {'train/loss': 0.05615108832716942, 'train/grad_norm_pre_clip': 15.07815933227539, 'timing/data': 0.0005498910322785378, 'timing/model': 0.947520000860095, 'learning_rate/qwen_vl_interface': 1.2522397030121782e-06, 'learning_rate/action_model': 1.2522397030121783e-05, 'epoch': 2.61}) +[2026-06-18 01:34:01,390][starVLA.training.train_starvla][INFO] - Step 5748, Loss: {'train/loss': 0.10506796836853027, 'train/grad_norm_pre_clip': 24.13722038269043, 'timing/data': 0.0004722690209746361, 'timing/model': 1.0188037790358067, 'learning_rate/qwen_vl_interface': 1.2510721591022194e-06, 'learning_rate/action_model': 1.2510721591022192e-05, 'epoch': 2.61}) +[2026-06-18 01:34:05,183][starVLA.training.train_starvla][INFO] - Step 5749, Loss: {'train/loss': 0.41701096296310425, 'train/grad_norm_pre_clip': 43.44219207763672, 'timing/data': 0.000430121086537838, 'timing/model': 0.932702025398612, 'learning_rate/qwen_vl_interface': 1.2499054441743047e-06, 'learning_rate/action_model': 1.2499054441743047e-05, 'epoch': 2.61}) +[2026-06-18 01:34:09,038][starVLA.training.train_starvla][INFO] - Step 5750, Loss: {'train/loss': 0.3241814970970154, 'train/grad_norm_pre_clip': 43.250003814697266, 'timing/data': 0.000712759792804718, 'timing/model': 1.1039562374353409, 'learning_rate/qwen_vl_interface': 1.2487395584702955e-06, 'learning_rate/action_model': 1.2487395584702954e-05, 'epoch': 2.61}) +[2026-06-18 01:34:12,701][starVLA.training.train_starvla][INFO] - Step 5751, Loss: {'train/loss': 0.49694252014160156, 'train/grad_norm_pre_clip': 38.37181091308594, 'timing/data': 0.0004967972636222839, 'timing/model': 0.9414547542110085, 'learning_rate/qwen_vl_interface': 1.2475745022318803e-06, 'learning_rate/action_model': 1.2475745022318803e-05, 'epoch': 2.61}) +[2026-06-18 01:34:16,533][starVLA.training.train_starvla][INFO] - Step 5752, Loss: {'train/loss': 0.4564913213253021, 'train/grad_norm_pre_clip': 24.851146697998047, 'timing/data': 0.000413261353969574, 'timing/model': 0.962630876339972, 'learning_rate/qwen_vl_interface': 1.2464102757005788e-06, 'learning_rate/action_model': 1.2464102757005788e-05, 'epoch': 2.61}) +[2026-06-18 01:34:20,274][starVLA.training.train_starvla][INFO] - Step 5753, Loss: {'train/loss': 0.12624746561050415, 'train/grad_norm_pre_clip': 27.632707595825195, 'timing/data': 0.0004351213574409485, 'timing/model': 0.929198089055717, 'learning_rate/qwen_vl_interface': 1.2452468791177357e-06, 'learning_rate/action_model': 1.2452468791177357e-05, 'epoch': 2.61}) +[2026-06-18 01:34:24,244][starVLA.training.train_starvla][INFO] - Step 5754, Loss: {'train/loss': 0.13751652836799622, 'train/grad_norm_pre_clip': 36.554725646972656, 'timing/data': 0.0006769169121980667, 'timing/model': 0.9638514192774892, 'learning_rate/qwen_vl_interface': 1.2440843127245235e-06, 'learning_rate/action_model': 1.2440843127245235e-05, 'epoch': 2.61}) +[2026-06-18 01:34:28,205][starVLA.training.train_starvla][INFO] - Step 5755, Loss: {'train/loss': 0.19541683793067932, 'train/grad_norm_pre_clip': 29.302812576293945, 'timing/data': 0.012818158604204655, 'timing/model': 0.9494673972949386, 'learning_rate/qwen_vl_interface': 1.2429225767619427e-06, 'learning_rate/action_model': 1.2429225767619427e-05, 'epoch': 2.61}) +[2026-06-18 01:34:32,188][starVLA.training.train_starvla][INFO] - Step 5756, Loss: {'train/loss': 0.2000141739845276, 'train/grad_norm_pre_clip': 17.70355796813965, 'timing/data': 0.012017696164548397, 'timing/model': 1.2273300914093852, 'learning_rate/qwen_vl_interface': 1.2417616714708246e-06, 'learning_rate/action_model': 1.2417616714708247e-05, 'epoch': 2.61}) +[2026-06-18 01:34:36,120][starVLA.training.train_starvla][INFO] - Step 5757, Loss: {'train/loss': 0.05927303433418274, 'train/grad_norm_pre_clip': 36.55538558959961, 'timing/data': 0.011869333684444427, 'timing/model': 1.204375684261322, 'learning_rate/qwen_vl_interface': 1.2406015970918247e-06, 'learning_rate/action_model': 1.2406015970918247e-05, 'epoch': 2.61}) +[2026-06-18 01:34:39,774][starVLA.training.train_starvla][INFO] - Step 5758, Loss: {'train/loss': 0.05547234043478966, 'train/grad_norm_pre_clip': 12.31534481048584, 'timing/data': 0.010893628932535648, 'timing/model': 0.9455770934000611, 'learning_rate/qwen_vl_interface': 1.2394423538654296e-06, 'learning_rate/action_model': 1.2394423538654296e-05, 'epoch': 2.61}) +[2026-06-18 01:34:43,634][starVLA.training.train_starvla][INFO] - Step 5759, Loss: {'train/loss': 0.27026674151420593, 'train/grad_norm_pre_clip': 37.75923538208008, 'timing/data': 0.009691267274320126, 'timing/model': 0.9419447984546423, 'learning_rate/qwen_vl_interface': 1.2382839420319478e-06, 'learning_rate/action_model': 1.2382839420319478e-05, 'epoch': 2.61}) +[2026-06-18 01:34:47,663][starVLA.training.train_starvla][INFO] - Step 5760, Loss: {'train/loss': 0.2712450921535492, 'train/grad_norm_pre_clip': 29.277973175048828, 'timing/data': 0.008772858418524265, 'timing/model': 0.938294586725533, 'learning_rate/qwen_vl_interface': 1.2371263618315225e-06, 'learning_rate/action_model': 1.2371263618315224e-05, 'epoch': 2.61}) +[2026-06-18 01:34:51,552][starVLA.training.train_starvla][INFO] - Step 5761, Loss: {'train/loss': 0.13869218528270721, 'train/grad_norm_pre_clip': 16.83814239501953, 'timing/data': 0.010795902460813522, 'timing/model': 0.9630481470376253, 'learning_rate/qwen_vl_interface': 1.2359696135041202e-06, 'learning_rate/action_model': 1.2359696135041201e-05, 'epoch': 2.61}) +[2026-06-18 01:34:56,213][starVLA.training.train_starvla][INFO] - Step 5762, Loss: {'train/loss': 0.46201834082603455, 'train/grad_norm_pre_clip': 43.127601623535156, 'timing/data': 0.017043082043528557, 'timing/model': 1.672610123641789, 'learning_rate/qwen_vl_interface': 1.2348136972895369e-06, 'learning_rate/action_model': 1.2348136972895368e-05, 'epoch': 2.61}) +[2026-06-18 01:35:00,161][starVLA.training.train_starvla][INFO] - Step 5763, Loss: {'train/loss': 0.23946797847747803, 'train/grad_norm_pre_clip': 35.85860824584961, 'timing/data': 0.015303879044950008, 'timing/model': 1.1674503460526466, 'learning_rate/qwen_vl_interface': 1.2336586134273945e-06, 'learning_rate/action_model': 1.2336586134273945e-05, 'epoch': 2.61}) +[2026-06-18 01:35:03,992][starVLA.training.train_starvla][INFO] - Step 5764, Loss: {'train/loss': 0.0690101757645607, 'train/grad_norm_pre_clip': 13.612273216247559, 'timing/data': 0.011487167328596115, 'timing/model': 1.121260041370988, 'learning_rate/qwen_vl_interface': 1.2325043621571428e-06, 'learning_rate/action_model': 1.2325043621571427e-05, 'epoch': 2.61}) +[2026-06-18 01:35:08,894][starVLA.training.train_starvla][INFO] - Step 5765, Loss: {'train/loss': 0.32304647564888, 'train/grad_norm_pre_clip': 34.265159606933594, 'timing/data': 0.018612324260175228, 'timing/model': 0.9390516076236963, 'learning_rate/qwen_vl_interface': 1.2313509437180603e-06, 'learning_rate/action_model': 1.2313509437180604e-05, 'epoch': 2.62}) +[2026-06-18 01:35:12,834][starVLA.training.train_starvla][INFO] - Step 5766, Loss: {'train/loss': 0.026480654254555702, 'train/grad_norm_pre_clip': 8.046361923217773, 'timing/data': 0.009810376912355423, 'timing/model': 0.9297876255586743, 'learning_rate/qwen_vl_interface': 1.2301983583492502e-06, 'learning_rate/action_model': 1.2301983583492502e-05, 'epoch': 2.62}) +[2026-06-18 01:35:16,623][starVLA.training.train_starvla][INFO] - Step 5767, Loss: {'train/loss': 0.21933384239673615, 'train/grad_norm_pre_clip': 30.224468231201172, 'timing/data': 0.009823002852499485, 'timing/model': 0.935625459998846, 'learning_rate/qwen_vl_interface': 1.229046606289648e-06, 'learning_rate/action_model': 1.229046606289648e-05, 'epoch': 2.62}) +[2026-06-18 01:35:20,411][starVLA.training.train_starvla][INFO] - Step 5768, Loss: {'train/loss': 0.2634522616863251, 'train/grad_norm_pre_clip': 25.242311477661133, 'timing/data': 0.008739055134356022, 'timing/model': 0.9554959405213594, 'learning_rate/qwen_vl_interface': 1.2278956877780102e-06, 'learning_rate/action_model': 1.2278956877780101e-05, 'epoch': 2.62}) +[2026-06-18 01:35:24,228][starVLA.training.train_starvla][INFO] - Step 5769, Loss: {'train/loss': 0.18908056616783142, 'train/grad_norm_pre_clip': 24.420576095581055, 'timing/data': 0.012341748923063278, 'timing/model': 0.9574473835527897, 'learning_rate/qwen_vl_interface': 1.2267456030529245e-06, 'learning_rate/action_model': 1.2267456030529245e-05, 'epoch': 2.62}) +[2026-06-18 01:35:28,024][starVLA.training.train_starvla][INFO] - Step 5770, Loss: {'train/loss': 0.43481382727622986, 'train/grad_norm_pre_clip': 33.22280502319336, 'timing/data': 0.008063032291829586, 'timing/model': 1.1053260816261172, 'learning_rate/qwen_vl_interface': 1.225596352352803e-06, 'learning_rate/action_model': 1.225596352352803e-05, 'epoch': 2.62}) +[2026-06-18 01:35:31,627][starVLA.training.train_starvla][INFO] - Step 5771, Loss: {'train/loss': 0.03572919964790344, 'train/grad_norm_pre_clip': 7.956958770751953, 'timing/data': 0.010745089501142502, 'timing/model': 0.9322574278339744, 'learning_rate/qwen_vl_interface': 1.2244479359158895e-06, 'learning_rate/action_model': 1.2244479359158895e-05, 'epoch': 2.62}) +[2026-06-18 01:35:35,699][starVLA.training.train_starvla][INFO] - Step 5772, Loss: {'train/loss': 0.21560686826705933, 'train/grad_norm_pre_clip': 26.290727615356445, 'timing/data': 0.007366151548922062, 'timing/model': 1.1934736566618085, 'learning_rate/qwen_vl_interface': 1.2233003539802512e-06, 'learning_rate/action_model': 1.2233003539802513e-05, 'epoch': 2.62}) +[2026-06-18 01:35:39,462][starVLA.training.train_starvla][INFO] - Step 5773, Loss: {'train/loss': 0.2292037159204483, 'train/grad_norm_pre_clip': 35.462074279785156, 'timing/data': 0.012191841378808022, 'timing/model': 0.9676418732851744, 'learning_rate/qwen_vl_interface': 1.2221536067837807e-06, 'learning_rate/action_model': 1.2221536067837808e-05, 'epoch': 2.62}) +[2026-06-18 01:35:43,330][starVLA.training.train_starvla][INFO] - Step 5774, Loss: {'train/loss': 0.20742267370224, 'train/grad_norm_pre_clip': 33.2122802734375, 'timing/data': 0.009647130034863949, 'timing/model': 0.9465670818462968, 'learning_rate/qwen_vl_interface': 1.2210076945642025e-06, 'learning_rate/action_model': 1.2210076945642024e-05, 'epoch': 2.62}) +[2026-06-18 01:35:47,126][starVLA.training.train_starvla][INFO] - Step 5775, Loss: {'train/loss': 0.051196835935115814, 'train/grad_norm_pre_clip': 11.951616287231445, 'timing/data': 0.010920645669102669, 'timing/model': 0.9447776637971401, 'learning_rate/qwen_vl_interface': 1.219862617559065e-06, 'learning_rate/action_model': 1.219862617559065e-05, 'epoch': 2.62}) +[2026-06-18 01:35:50,895][starVLA.training.train_starvla][INFO] - Step 5776, Loss: {'train/loss': 0.09874669462442398, 'train/grad_norm_pre_clip': 18.302831649780273, 'timing/data': 0.017799620516598225, 'timing/model': 1.1059399293735623, 'learning_rate/qwen_vl_interface': 1.2187183760057431e-06, 'learning_rate/action_model': 1.218718376005743e-05, 'epoch': 2.62}) +[2026-06-18 01:35:54,702][starVLA.training.train_starvla][INFO] - Step 5777, Loss: {'train/loss': 0.2891329824924469, 'train/grad_norm_pre_clip': 33.04602813720703, 'timing/data': 0.012040314264595509, 'timing/model': 1.106539061293006, 'learning_rate/qwen_vl_interface': 1.2175749701414393e-06, 'learning_rate/action_model': 1.2175749701414394e-05, 'epoch': 2.62}) +[2026-06-18 01:35:58,319][starVLA.training.train_starvla][INFO] - Step 5778, Loss: {'train/loss': 0.06839895993471146, 'train/grad_norm_pre_clip': 20.540573120117188, 'timing/data': 0.011391563341021538, 'timing/model': 0.935768611729145, 'learning_rate/qwen_vl_interface': 1.216432400203184e-06, 'learning_rate/action_model': 1.2164324002031841e-05, 'epoch': 2.62}) +[2026-06-18 01:36:02,206][starVLA.training.train_starvla][INFO] - Step 5779, Loss: {'train/loss': 0.19934172928333282, 'train/grad_norm_pre_clip': 19.87494659423828, 'timing/data': 0.007000983692705631, 'timing/model': 0.9411320928484201, 'learning_rate/qwen_vl_interface': 1.215290666427832e-06, 'learning_rate/action_model': 1.2152906664278318e-05, 'epoch': 2.62}) +[2026-06-18 01:36:06,114][starVLA.training.train_starvla][INFO] - Step 5780, Loss: {'train/loss': 0.2433360069990158, 'train/grad_norm_pre_clip': 19.93034553527832, 'timing/data': 0.008133742958307266, 'timing/model': 1.0575546650215983, 'learning_rate/qwen_vl_interface': 1.214149769052067e-06, 'learning_rate/action_model': 1.214149769052067e-05, 'epoch': 2.62}) +[2026-06-18 01:36:09,930][starVLA.training.train_starvla][INFO] - Step 5781, Loss: {'train/loss': 0.6131598353385925, 'train/grad_norm_pre_clip': 32.944271087646484, 'timing/data': 0.010713104158639908, 'timing/model': 0.9481189101934433, 'learning_rate/qwen_vl_interface': 1.2130097083123971e-06, 'learning_rate/action_model': 1.213009708312397e-05, 'epoch': 2.62}) +[2026-06-18 01:36:13,690][starVLA.training.train_starvla][INFO] - Step 5782, Loss: {'train/loss': 0.031930580735206604, 'train/grad_norm_pre_clip': 7.965605735778809, 'timing/data': 0.008977416902780533, 'timing/model': 0.936075845733285, 'learning_rate/qwen_vl_interface': 1.2118704844451587e-06, 'learning_rate/action_model': 1.2118704844451586e-05, 'epoch': 2.62}) +[2026-06-18 01:36:17,528][starVLA.training.train_starvla][INFO] - Step 5783, Loss: {'train/loss': 0.1832198202610016, 'train/grad_norm_pre_clip': 68.99931335449219, 'timing/data': 0.011801722459495068, 'timing/model': 1.15789641905576, 'learning_rate/qwen_vl_interface': 1.2107320976865138e-06, 'learning_rate/action_model': 1.2107320976865139e-05, 'epoch': 2.62}) +[2026-06-18 01:36:21,312][starVLA.training.train_starvla][INFO] - Step 5784, Loss: {'train/loss': 0.1349247843027115, 'train/grad_norm_pre_clip': 19.833667755126953, 'timing/data': 0.012454673647880554, 'timing/model': 1.129698771983385, 'learning_rate/qwen_vl_interface': 1.209594548272451e-06, 'learning_rate/action_model': 1.209594548272451e-05, 'epoch': 2.62}) +[2026-06-18 01:36:24,998][starVLA.training.train_starvla][INFO] - Step 5785, Loss: {'train/loss': 0.22021278738975525, 'train/grad_norm_pre_clip': 31.013282775878906, 'timing/data': 0.010032493621110916, 'timing/model': 0.9453876158222556, 'learning_rate/qwen_vl_interface': 1.2084578364387869e-06, 'learning_rate/action_model': 1.208457836438787e-05, 'epoch': 2.62}) +[2026-06-18 01:36:28,876][starVLA.training.train_starvla][INFO] - Step 5786, Loss: {'train/loss': 0.28664037585258484, 'train/grad_norm_pre_clip': 32.27863311767578, 'timing/data': 0.010215037502348423, 'timing/model': 1.007752819918096, 'learning_rate/qwen_vl_interface': 1.2073219624211635e-06, 'learning_rate/action_model': 1.2073219624211635e-05, 'epoch': 2.62}) +[2026-06-18 01:36:32,740][starVLA.training.train_starvla][INFO] - Step 5787, Loss: {'train/loss': 0.25410139560699463, 'train/grad_norm_pre_clip': 28.360607147216797, 'timing/data': 0.010293246246874332, 'timing/model': 0.9286626484245062, 'learning_rate/qwen_vl_interface': 1.2061869264550458e-06, 'learning_rate/action_model': 1.2061869264550457e-05, 'epoch': 2.63}) +[2026-06-18 01:36:36,517][starVLA.training.train_starvla][INFO] - Step 5788, Loss: {'train/loss': 0.08279109746217728, 'train/grad_norm_pre_clip': 27.14735984802246, 'timing/data': 0.007847408764064312, 'timing/model': 0.9436897709965706, 'learning_rate/qwen_vl_interface': 1.2050527287757288e-06, 'learning_rate/action_model': 1.2050527287757289e-05, 'epoch': 2.63}) +[2026-06-18 01:36:40,422][starVLA.training.train_starvla][INFO] - Step 5789, Loss: {'train/loss': 0.21534158289432526, 'train/grad_norm_pre_clip': 67.51519012451172, 'timing/data': 0.017181670293211937, 'timing/model': 0.9402444642037153, 'learning_rate/qwen_vl_interface': 1.203919369618335e-06, 'learning_rate/action_model': 1.2039193696183348e-05, 'epoch': 2.63}) +[2026-06-18 01:36:44,373][starVLA.training.train_starvla][INFO] - Step 5790, Loss: {'train/loss': 0.029297545552253723, 'train/grad_norm_pre_clip': 8.743192672729492, 'timing/data': 0.008376405574381351, 'timing/model': 1.2890681847929955, 'learning_rate/qwen_vl_interface': 1.202786849217809e-06, 'learning_rate/action_model': 1.202786849217809e-05, 'epoch': 2.63}) +[2026-06-18 01:36:48,724][starVLA.training.train_starvla][INFO] - Step 5791, Loss: {'train/loss': 0.15335965156555176, 'train/grad_norm_pre_clip': 23.422157287597656, 'timing/data': 0.015467524528503418, 'timing/model': 1.623045440763235, 'learning_rate/qwen_vl_interface': 1.2016551678089247e-06, 'learning_rate/action_model': 1.2016551678089246e-05, 'epoch': 2.63}) +[2026-06-18 01:36:52,632][starVLA.training.train_starvla][INFO] - Step 5792, Loss: {'train/loss': 0.6005213260650635, 'train/grad_norm_pre_clip': 52.91617202758789, 'timing/data': 0.016207345761358738, 'timing/model': 0.9394765812903643, 'learning_rate/qwen_vl_interface': 1.2005243256262803e-06, 'learning_rate/action_model': 1.2005243256262802e-05, 'epoch': 2.63}) +[2026-06-18 01:36:56,464][starVLA.training.train_starvla][INFO] - Step 5793, Loss: {'train/loss': 0.09566459059715271, 'train/grad_norm_pre_clip': 21.870338439941406, 'timing/data': 0.013507924973964691, 'timing/model': 0.9265665020793676, 'learning_rate/qwen_vl_interface': 1.1993943229043003e-06, 'learning_rate/action_model': 1.1993943229043003e-05, 'epoch': 2.63}) +[2026-06-18 01:37:00,355][starVLA.training.train_starvla][INFO] - Step 5794, Loss: {'train/loss': 0.031502146273851395, 'train/grad_norm_pre_clip': 8.837995529174805, 'timing/data': 0.00784257985651493, 'timing/model': 0.9273604825139046, 'learning_rate/qwen_vl_interface': 1.1982651598772358e-06, 'learning_rate/action_model': 1.1982651598772356e-05, 'epoch': 2.63}) +[2026-06-18 01:37:04,229][starVLA.training.train_starvla][INFO] - Step 5795, Loss: {'train/loss': 0.13202613592147827, 'train/grad_norm_pre_clip': 31.797603607177734, 'timing/data': 0.015738346613943577, 'timing/model': 0.960958513431251, 'learning_rate/qwen_vl_interface': 1.197136836779164e-06, 'learning_rate/action_model': 1.1971368367791639e-05, 'epoch': 2.63}) +[2026-06-18 01:37:08,053][starVLA.training.train_starvla][INFO] - Step 5796, Loss: {'train/loss': 0.36657366156578064, 'train/grad_norm_pre_clip': 47.6760368347168, 'timing/data': 0.008344576694071293, 'timing/model': 1.1582694509997964, 'learning_rate/qwen_vl_interface': 1.1960093538439866e-06, 'learning_rate/action_model': 1.1960093538439864e-05, 'epoch': 2.63}) +[2026-06-18 01:37:13,352][starVLA.training.train_starvla][INFO] - Step 5797, Loss: {'train/loss': 0.059084512293338776, 'train/grad_norm_pre_clip': 16.259090423583984, 'timing/data': 0.00865020789206028, 'timing/model': 2.596349968574941, 'learning_rate/qwen_vl_interface': 1.1948827113054323e-06, 'learning_rate/action_model': 1.1948827113054323e-05, 'epoch': 2.63}) +[2026-06-18 01:37:18,145][starVLA.training.train_starvla][INFO] - Step 5798, Loss: {'train/loss': 0.3278975486755371, 'train/grad_norm_pre_clip': 33.66575241088867, 'timing/data': 0.013584003783762455, 'timing/model': 1.9642841424793005, 'learning_rate/qwen_vl_interface': 1.1937569093970558e-06, 'learning_rate/action_model': 1.1937569093970558e-05, 'epoch': 2.63}) +[2026-06-18 01:37:23,702][starVLA.training.train_starvla][INFO] - Step 5799, Loss: {'train/loss': 0.2577487826347351, 'train/grad_norm_pre_clip': 28.943408966064453, 'timing/data': 0.009781179949641228, 'timing/model': 1.0230593709275126, 'learning_rate/qwen_vl_interface': 1.1926319483522353e-06, 'learning_rate/action_model': 1.1926319483522353e-05, 'epoch': 2.63}) +[2026-06-18 01:39:30,997][starVLA.training.train_starvla][INFO] - Step 5800, Loss: {'train/loss': 0.07027220726013184, 'train/grad_norm_pre_clip': 15.09267807006836, 'eval/loss': 0.2647063434123993, 'eval/latency_8/loss': 0.2647063434123993, 'eval/demon_fire_precision': 0.9679193400549955, 'eval/demon_fire_recall': 0.9841565703634669, 'eval/demon_fire_f1': 0.9759704251386322, 'eval/demon_fire_rate_model': 0.9106844741235393, 'eval/demon_fire_rate_teacher': 0.8956594323873122, 'eval/demon_left_precision': 0.9051094890510949, 'eval/demon_left_recall': 0.7965738758029979, 'eval/demon_left_f1': 0.847380410022779, 'eval/demon_left_rate_model': 0.11435726210350584, 'eval/demon_left_rate_teacher': 0.12993878686700056, 'eval/demon_right_precision': 0.8840125391849529, 'eval/demon_right_recall': 0.8545454545454545, 'eval/demon_right_f1': 0.8690292758089369, 'eval/demon_right_rate_model': 0.1775180856983862, 'eval/demon_right_rate_teacher': 0.18363939899833054, 'eval/demon_move_f1': 0.858204842915858, 'eval/demon_cc_f1': 0.9373812974372429, 'eval/latency_8/demon_fire_precision': 0.9679193400549955, 'eval/latency_8/demon_fire_recall': 0.9841565703634669, 'eval/latency_8/demon_fire_f1': 0.9759704251386322, 'eval/latency_8/demon_fire_rate_model': 0.9106844741235393, 'eval/latency_8/demon_fire_rate_teacher': 0.8956594323873122, 'eval/latency_8/demon_left_precision': 0.9051094890510949, 'eval/latency_8/demon_left_recall': 0.7965738758029979, 'eval/latency_8/demon_left_f1': 0.847380410022779, 'eval/latency_8/demon_left_rate_model': 0.11435726210350584, 'eval/latency_8/demon_left_rate_teacher': 0.12993878686700056, 'eval/latency_8/demon_right_precision': 0.8840125391849529, 'eval/latency_8/demon_right_recall': 0.8545454545454545, 'eval/latency_8/demon_right_f1': 0.8690292758089369, 'eval/latency_8/demon_right_rate_model': 0.1775180856983862, 'eval/latency_8/demon_right_rate_teacher': 0.18363939899833054, 'eval/latency_8/demon_move_f1': 0.858204842915858, 'eval/latency_8/demon_cc_f1': 0.9373812974372429, 'timing/data': 0.009774306789040565, 'timing/model': 1.1762650283053517, 'learning_rate/qwen_vl_interface': 1.1915078284041793e-06, 'learning_rate/action_model': 1.1915078284041793e-05, 'epoch': 2.63}) +[2026-06-18 01:39:31,005][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_5800_state +[2026-06-18 01:39:31,009][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 01:39:37,448][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_5800_state/model.safetensors +[2026-06-18 01:39:52,663][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_5800_state/optimizer.bin +[2026-06-18 01:39:52,664][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_5800_state/sampler.bin +[2026-06-18 01:39:52,664][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_5800_state/sampler_1.bin +[2026-06-18 01:39:52,666][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_5800_state/random_states_0.pkl +[2026-06-18 01:39:52,679][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 5800 +[2026-06-18 01:39:52,679][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 01:39:52,692][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 01:39:56,742][starVLA.training.train_starvla][INFO] - Step 5801, Loss: {'train/loss': 0.32982298731803894, 'train/grad_norm_pre_clip': 60.338233947753906, 'timing/data': 0.013564114458858967, 'timing/model': 0.9524327348917723, 'learning_rate/qwen_vl_interface': 1.1903845497859185e-06, 'learning_rate/action_model': 1.1903845497859184e-05, 'epoch': 2.63}) +[2026-06-18 01:40:00,627][starVLA.training.train_starvla][INFO] - Step 5802, Loss: {'train/loss': 0.3340170383453369, 'train/grad_norm_pre_clip': 32.74571990966797, 'timing/data': 0.03059396706521511, 'timing/model': 0.9570313403382897, 'learning_rate/qwen_vl_interface': 1.1892621127303067e-06, 'learning_rate/action_model': 1.1892621127303066e-05, 'epoch': 2.63}) +[2026-06-18 01:40:04,730][starVLA.training.train_starvla][INFO] - Step 5803, Loss: {'train/loss': 0.13925954699516296, 'train/grad_norm_pre_clip': 17.353431701660156, 'timing/data': 0.022952459752559662, 'timing/model': 0.9595942627638578, 'learning_rate/qwen_vl_interface': 1.1881405174700292e-06, 'learning_rate/action_model': 1.1881405174700291e-05, 'epoch': 2.63}) +[2026-06-18 01:40:08,881][starVLA.training.train_starvla][INFO] - Step 5804, Loss: {'train/loss': 0.16937299072742462, 'train/grad_norm_pre_clip': 78.13591766357422, 'timing/data': 0.014669246971607208, 'timing/model': 1.1713528018444777, 'learning_rate/qwen_vl_interface': 1.1870197642375935e-06, 'learning_rate/action_model': 1.1870197642375935e-05, 'epoch': 2.63}) +[2026-06-18 01:40:12,837][starVLA.training.train_starvla][INFO] - Step 5805, Loss: {'train/loss': 0.1396612524986267, 'train/grad_norm_pre_clip': 30.34473419189453, 'timing/data': 0.01600303966552019, 'timing/model': 0.9242009278386831, 'learning_rate/qwen_vl_interface': 1.1858998532653326e-06, 'learning_rate/action_model': 1.1858998532653325e-05, 'epoch': 2.63}) +[2026-06-18 01:40:16,726][starVLA.training.train_starvla][INFO] - Step 5806, Loss: {'train/loss': 0.09422128647565842, 'train/grad_norm_pre_clip': 16.49680519104004, 'timing/data': 0.01748052518814802, 'timing/model': 1.169158497825265, 'learning_rate/qwen_vl_interface': 1.1847807847854054e-06, 'learning_rate/action_model': 1.1847807847854054e-05, 'epoch': 2.63}) +[2026-06-18 01:40:20,610][starVLA.training.train_starvla][INFO] - Step 5807, Loss: {'train/loss': 0.44320809841156006, 'train/grad_norm_pre_clip': 33.300323486328125, 'timing/data': 0.020840484648942947, 'timing/model': 1.1043308833613992, 'learning_rate/qwen_vl_interface': 1.1836625590297959e-06, 'learning_rate/action_model': 1.1836625590297958e-05, 'epoch': 2.63}) +[2026-06-18 01:40:24,288][starVLA.training.train_starvla][INFO] - Step 5808, Loss: {'train/loss': 0.22660571336746216, 'train/grad_norm_pre_clip': 41.7338752746582, 'timing/data': 0.0272674597799778, 'timing/model': 0.9523012796416879, 'learning_rate/qwen_vl_interface': 1.1825451762303134e-06, 'learning_rate/action_model': 1.1825451762303133e-05, 'epoch': 2.63}) +[2026-06-18 01:40:28,122][starVLA.training.train_starvla][INFO] - Step 5809, Loss: {'train/loss': 0.05509661138057709, 'train/grad_norm_pre_clip': 13.21497917175293, 'timing/data': 0.01540272869169712, 'timing/model': 0.9169123936444521, 'learning_rate/qwen_vl_interface': 1.1814286366185923e-06, 'learning_rate/action_model': 1.1814286366185922e-05, 'epoch': 2.64}) +[2026-06-18 01:40:32,012][starVLA.training.train_starvla][INFO] - Step 5810, Loss: {'train/loss': 0.20408639311790466, 'train/grad_norm_pre_clip': 17.040193557739258, 'timing/data': 0.018409240059554577, 'timing/model': 0.9340654900297523, 'learning_rate/qwen_vl_interface': 1.1803129404260932e-06, 'learning_rate/action_model': 1.1803129404260931e-05, 'epoch': 2.64}) +[2026-06-18 01:40:35,914][starVLA.training.train_starvla][INFO] - Step 5811, Loss: {'train/loss': 0.05840257182717323, 'train/grad_norm_pre_clip': 16.738998413085938, 'timing/data': 0.02424048725515604, 'timing/model': 0.9570888495072722, 'learning_rate/qwen_vl_interface': 1.1791980878841008e-06, 'learning_rate/action_model': 1.1791980878841008e-05, 'epoch': 2.64}) +[2026-06-18 01:40:39,802][starVLA.training.train_starvla][INFO] - Step 5812, Loss: {'train/loss': 0.155385360121727, 'train/grad_norm_pre_clip': 33.62489700317383, 'timing/data': 0.016051700338721275, 'timing/model': 0.9789842749014497, 'learning_rate/qwen_vl_interface': 1.1780840792237246e-06, 'learning_rate/action_model': 1.1780840792237246e-05, 'epoch': 2.64}) +[2026-06-18 01:40:43,620][starVLA.training.train_starvla][INFO] - Step 5813, Loss: {'train/loss': 0.3007608950138092, 'train/grad_norm_pre_clip': 30.9216251373291, 'timing/data': 0.024908307939767838, 'timing/model': 1.12554691452533, 'learning_rate/qwen_vl_interface': 1.1769709146759002e-06, 'learning_rate/action_model': 1.1769709146759e-05, 'epoch': 2.64}) +[2026-06-18 01:40:47,248][starVLA.training.train_starvla][INFO] - Step 5814, Loss: {'train/loss': 0.09474606066942215, 'train/grad_norm_pre_clip': 23.56458854675293, 'timing/data': 0.020852395333349705, 'timing/model': 0.9330298565328121, 'learning_rate/qwen_vl_interface': 1.175858594471388e-06, 'learning_rate/action_model': 1.175858594471388e-05, 'epoch': 2.64}) +[2026-06-18 01:40:51,017][starVLA.training.train_starvla][INFO] - Step 5815, Loss: {'train/loss': 0.11927486211061478, 'train/grad_norm_pre_clip': 19.858386993408203, 'timing/data': 0.02231360785663128, 'timing/model': 0.922494881786406, 'learning_rate/qwen_vl_interface': 1.174747118840775e-06, 'learning_rate/action_model': 1.1747471188407749e-05, 'epoch': 2.64}) +[2026-06-18 01:40:54,728][starVLA.training.train_starvla][INFO] - Step 5816, Loss: {'train/loss': 0.29002177715301514, 'train/grad_norm_pre_clip': 38.832271575927734, 'timing/data': 0.017361478880047798, 'timing/model': 0.9442130429670215, 'learning_rate/qwen_vl_interface': 1.173636488014468e-06, 'learning_rate/action_model': 1.173636488014468e-05, 'epoch': 2.64}) +[2026-06-18 01:40:58,518][starVLA.training.train_starvla][INFO] - Step 5817, Loss: {'train/loss': 0.06791552901268005, 'train/grad_norm_pre_clip': 13.55369758605957, 'timing/data': 0.019565449096262455, 'timing/model': 0.9637581678107381, 'learning_rate/qwen_vl_interface': 1.1725267022227024e-06, 'learning_rate/action_model': 1.1725267022227025e-05, 'epoch': 2.64}) +[2026-06-18 01:41:02,431][starVLA.training.train_starvla][INFO] - Step 5818, Loss: {'train/loss': 0.17188259959220886, 'train/grad_norm_pre_clip': 31.616952896118164, 'timing/data': 0.021287795156240463, 'timing/model': 0.9277441948652267, 'learning_rate/qwen_vl_interface': 1.17141776169554e-06, 'learning_rate/action_model': 1.17141776169554e-05, 'epoch': 2.64}) +[2026-06-18 01:41:06,261][starVLA.training.train_starvla][INFO] - Step 5819, Loss: {'train/loss': 0.17146585881710052, 'train/grad_norm_pre_clip': 19.771150588989258, 'timing/data': 0.022128982469439507, 'timing/model': 1.1664369190111756, 'learning_rate/qwen_vl_interface': 1.170309666662864e-06, 'learning_rate/action_model': 1.170309666662864e-05, 'epoch': 2.64}) +[2026-06-18 01:41:10,213][starVLA.training.train_starvla][INFO] - Step 5820, Loss: {'train/loss': 0.2842392027378082, 'train/grad_norm_pre_clip': 21.510576248168945, 'timing/data': 0.01840130425989628, 'timing/model': 1.1335277948528528, 'learning_rate/qwen_vl_interface': 1.1692024173543858e-06, 'learning_rate/action_model': 1.1692024173543859e-05, 'epoch': 2.64}) +[2026-06-18 01:41:13,973][starVLA.training.train_starvla][INFO] - Step 5821, Loss: {'train/loss': 0.16937255859375, 'train/grad_norm_pre_clip': 26.57288360595703, 'timing/data': 0.015120853669941425, 'timing/model': 0.9498446397483349, 'learning_rate/qwen_vl_interface': 1.168096013999635e-06, 'learning_rate/action_model': 1.168096013999635e-05, 'epoch': 2.64}) +[2026-06-18 01:41:17,925][starVLA.training.train_starvla][INFO] - Step 5822, Loss: {'train/loss': 0.13421760499477386, 'train/grad_norm_pre_clip': 31.639314651489258, 'timing/data': 0.017222920432686806, 'timing/model': 0.9371119365096092, 'learning_rate/qwen_vl_interface': 1.1669904568279742e-06, 'learning_rate/action_model': 1.1669904568279741e-05, 'epoch': 2.64}) +[2026-06-18 01:41:21,779][starVLA.training.train_starvla][INFO] - Step 5823, Loss: {'train/loss': 0.028383243829011917, 'train/grad_norm_pre_clip': 5.328353404998779, 'timing/data': 0.024784574285149574, 'timing/model': 0.9771394357085228, 'learning_rate/qwen_vl_interface': 1.165885746068585e-06, 'learning_rate/action_model': 1.1658857460685848e-05, 'epoch': 2.64}) +[2026-06-18 01:41:25,601][starVLA.training.train_starvla][INFO] - Step 5824, Loss: {'train/loss': 0.04985596239566803, 'train/grad_norm_pre_clip': 14.314126968383789, 'timing/data': 0.02613420318812132, 'timing/model': 0.9361163303256035, 'learning_rate/qwen_vl_interface': 1.1647818819504754e-06, 'learning_rate/action_model': 1.1647818819504756e-05, 'epoch': 2.64}) +[2026-06-18 01:41:29,644][starVLA.training.train_starvla][INFO] - Step 5825, Loss: {'train/loss': 0.4037698805332184, 'train/grad_norm_pre_clip': 30.822359085083008, 'timing/data': 0.015920368023216724, 'timing/model': 0.9371781395748258, 'learning_rate/qwen_vl_interface': 1.1636788647024776e-06, 'learning_rate/action_model': 1.1636788647024776e-05, 'epoch': 2.64}) +[2026-06-18 01:41:33,546][starVLA.training.train_starvla][INFO] - Step 5826, Loss: {'train/loss': 0.2689904272556305, 'train/grad_norm_pre_clip': 33.188316345214844, 'timing/data': 0.016292636282742023, 'timing/model': 1.13458286318928, 'learning_rate/qwen_vl_interface': 1.1625766945532485e-06, 'learning_rate/action_model': 1.1625766945532484e-05, 'epoch': 2.64}) +[2026-06-18 01:41:37,308][starVLA.training.train_starvla][INFO] - Step 5827, Loss: {'train/loss': 0.06988692283630371, 'train/grad_norm_pre_clip': 17.04132843017578, 'timing/data': 0.018658260814845562, 'timing/model': 1.1158097302541137, 'learning_rate/qwen_vl_interface': 1.1614753717312683e-06, 'learning_rate/action_model': 1.1614753717312681e-05, 'epoch': 2.64}) +[2026-06-18 01:41:40,931][starVLA.training.train_starvla][INFO] - Step 5828, Loss: {'train/loss': 0.13194972276687622, 'train/grad_norm_pre_clip': 22.204355239868164, 'timing/data': 0.017782443203032017, 'timing/model': 0.9313225504010916, 'learning_rate/qwen_vl_interface': 1.1603748964648423e-06, 'learning_rate/action_model': 1.1603748964648422e-05, 'epoch': 2.64}) +[2026-06-18 01:41:44,728][starVLA.training.train_starvla][INFO] - Step 5829, Loss: {'train/loss': 0.2534881830215454, 'train/grad_norm_pre_clip': 35.665802001953125, 'timing/data': 0.015950024127960205, 'timing/model': 0.9359865728765726, 'learning_rate/qwen_vl_interface': 1.1592752689821024e-06, 'learning_rate/action_model': 1.1592752689821024e-05, 'epoch': 2.64}) +[2026-06-18 01:41:48,682][starVLA.training.train_starvla][INFO] - Step 5830, Loss: {'train/loss': 0.3378615379333496, 'train/grad_norm_pre_clip': 25.188270568847656, 'timing/data': 0.0004867464303970337, 'timing/model': 0.9248422821983695, 'learning_rate/qwen_vl_interface': 1.1581764895110007e-06, 'learning_rate/action_model': 1.1581764895110007e-05, 'epoch': 2.64}) +[2026-06-18 01:41:52,568][starVLA.training.train_starvla][INFO] - Step 5831, Loss: {'train/loss': 0.2468166947364807, 'train/grad_norm_pre_clip': 25.69997787475586, 'timing/data': 0.0010339263826608658, 'timing/model': 0.9294648710638285, 'learning_rate/qwen_vl_interface': 1.1570785582793153e-06, 'learning_rate/action_model': 1.1570785582793154e-05, 'epoch': 2.65}) +[2026-06-18 01:41:56,546][starVLA.training.train_starvla][INFO] - Step 5832, Loss: {'train/loss': 0.054159507155418396, 'train/grad_norm_pre_clip': 14.24064826965332, 'timing/data': 0.00046578142791986465, 'timing/model': 0.9287563366815448, 'learning_rate/qwen_vl_interface': 1.1559814755146484e-06, 'learning_rate/action_model': 1.1559814755146483e-05, 'epoch': 2.65}) +[2026-06-18 01:42:00,513][starVLA.training.train_starvla][INFO] - Step 5833, Loss: {'train/loss': 0.20107293128967285, 'train/grad_norm_pre_clip': 47.69062423706055, 'timing/data': 0.0009878762066364288, 'timing/model': 1.2512388145551085, 'learning_rate/qwen_vl_interface': 1.154885241444428e-06, 'learning_rate/action_model': 1.1548852414444277e-05, 'epoch': 2.65}) +[2026-06-18 01:42:04,228][starVLA.training.train_starvla][INFO] - Step 5834, Loss: {'train/loss': 0.26869216561317444, 'train/grad_norm_pre_clip': 45.58481216430664, 'timing/data': 0.001054292544722557, 'timing/model': 0.9766766782850027, 'learning_rate/qwen_vl_interface': 1.1537898562959039e-06, 'learning_rate/action_model': 1.1537898562959037e-05, 'epoch': 2.65}) +[2026-06-18 01:42:08,409][starVLA.training.train_starvla][INFO] - Step 5835, Loss: {'train/loss': 0.10301799327135086, 'train/grad_norm_pre_clip': 22.460737228393555, 'timing/data': 0.0008430508896708488, 'timing/model': 1.2164465868845582, 'learning_rate/qwen_vl_interface': 1.152695320296149e-06, 'learning_rate/action_model': 1.1526953202961488e-05, 'epoch': 2.65}) +[2026-06-18 01:42:14,090][starVLA.training.train_starvla][INFO] - Step 5836, Loss: {'train/loss': 0.4451817572116852, 'train/grad_norm_pre_clip': 30.867340087890625, 'timing/data': 0.004041684791445732, 'timing/model': 1.8994313543662429, 'learning_rate/qwen_vl_interface': 1.1516016336720635e-06, 'learning_rate/action_model': 1.1516016336720633e-05, 'epoch': 2.65}) +[2026-06-18 01:42:18,108][starVLA.training.train_starvla][INFO] - Step 5837, Loss: {'train/loss': 0.10771556198596954, 'train/grad_norm_pre_clip': 25.63340950012207, 'timing/data': 0.0004487382248044014, 'timing/model': 0.9542572945356369, 'learning_rate/qwen_vl_interface': 1.1505087966503697e-06, 'learning_rate/action_model': 1.1505087966503697e-05, 'epoch': 2.65}) +[2026-06-18 01:42:21,994][starVLA.training.train_starvla][INFO] - Step 5838, Loss: {'train/loss': 0.25094449520111084, 'train/grad_norm_pre_clip': 28.773666381835938, 'timing/data': 0.0010752715170383453, 'timing/model': 1.041553931310773, 'learning_rate/qwen_vl_interface': 1.1494168094576138e-06, 'learning_rate/action_model': 1.1494168094576138e-05, 'epoch': 2.65}) +[2026-06-18 01:42:25,932][starVLA.training.train_starvla][INFO] - Step 5839, Loss: {'train/loss': 0.11079303175210953, 'train/grad_norm_pre_clip': 28.015823364257812, 'timing/data': 0.000516875647008419, 'timing/model': 1.0900833737105131, 'learning_rate/qwen_vl_interface': 1.1483256723201656e-06, 'learning_rate/action_model': 1.1483256723201655e-05, 'epoch': 2.65}) +[2026-06-18 01:42:29,814][starVLA.training.train_starvla][INFO] - Step 5840, Loss: {'train/loss': 0.03559362515807152, 'train/grad_norm_pre_clip': 8.734435081481934, 'timing/data': 0.000521412119269371, 'timing/model': 1.1398954400792718, 'learning_rate/qwen_vl_interface': 1.1472353854642187e-06, 'learning_rate/action_model': 1.1472353854642187e-05, 'epoch': 2.65}) +[2026-06-18 01:42:34,196][starVLA.training.train_starvla][INFO] - Step 5841, Loss: {'train/loss': 0.04843900352716446, 'train/grad_norm_pre_clip': 10.087257385253906, 'timing/data': 0.00042286328971385956, 'timing/model': 1.7239154744893312, 'learning_rate/qwen_vl_interface': 1.1461459491157914e-06, 'learning_rate/action_model': 1.1461459491157914e-05, 'epoch': 2.65}) +[2026-06-18 01:42:41,454][starVLA.training.train_starvla][INFO] - Step 5842, Loss: {'train/loss': 0.3670330345630646, 'train/grad_norm_pre_clip': 34.447757720947266, 'timing/data': 0.0008583385497331619, 'timing/model': 0.9510218212381005, 'learning_rate/qwen_vl_interface': 1.145057363500725e-06, 'learning_rate/action_model': 1.1450573635007251e-05, 'epoch': 2.65}) +[2026-06-18 01:42:45,362][starVLA.training.train_starvla][INFO] - Step 5843, Loss: {'train/loss': 0.03395687788724899, 'train/grad_norm_pre_clip': 11.907337188720703, 'timing/data': 0.0005900748074054718, 'timing/model': 0.9715269394218922, 'learning_rate/qwen_vl_interface': 1.1439696288446845e-06, 'learning_rate/action_model': 1.1439696288446844e-05, 'epoch': 2.65}) +[2026-06-18 01:42:49,440][starVLA.training.train_starvla][INFO] - Step 5844, Loss: {'train/loss': 0.09108912944793701, 'train/grad_norm_pre_clip': 26.358318328857422, 'timing/data': 0.00045389123260974884, 'timing/model': 1.1638570465147495, 'learning_rate/qwen_vl_interface': 1.1428827453731575e-06, 'learning_rate/action_model': 1.1428827453731574e-05, 'epoch': 2.65}) +[2026-06-18 01:42:53,334][starVLA.training.train_starvla][INFO] - Step 5845, Loss: {'train/loss': 0.11367862671613693, 'train/grad_norm_pre_clip': 20.57400131225586, 'timing/data': 0.00043859053403139114, 'timing/model': 0.9470774158835411, 'learning_rate/qwen_vl_interface': 1.1417967133114575e-06, 'learning_rate/action_model': 1.1417967133114573e-05, 'epoch': 2.65}) +[2026-06-18 01:42:57,131][starVLA.training.train_starvla][INFO] - Step 5846, Loss: {'train/loss': 0.34446051716804504, 'train/grad_norm_pre_clip': 36.53718185424805, 'timing/data': 0.0005531283095479012, 'timing/model': 1.0992149915546179, 'learning_rate/qwen_vl_interface': 1.140711532884718e-06, 'learning_rate/action_model': 1.140711532884718e-05, 'epoch': 2.65}) +[2026-06-18 01:43:00,724][starVLA.training.train_starvla][INFO] - Step 5847, Loss: {'train/loss': 0.17313256859779358, 'train/grad_norm_pre_clip': 121.85086059570312, 'timing/data': 0.00043635908514261246, 'timing/model': 0.942222498357296, 'learning_rate/qwen_vl_interface': 1.1396272043179004e-06, 'learning_rate/action_model': 1.1396272043179005e-05, 'epoch': 2.65}) +[2026-06-18 01:43:04,614][starVLA.training.train_starvla][INFO] - Step 5848, Loss: {'train/loss': 0.10088598728179932, 'train/grad_norm_pre_clip': 25.13146209716797, 'timing/data': 0.00044639967381954193, 'timing/model': 0.9380014883354306, 'learning_rate/qwen_vl_interface': 1.138543727835788e-06, 'learning_rate/action_model': 1.138543727835788e-05, 'epoch': 2.65}) +[2026-06-18 01:43:08,434][starVLA.training.train_starvla][INFO] - Step 5849, Loss: {'train/loss': 0.11655136197805405, 'train/grad_norm_pre_clip': 8.085841178894043, 'timing/data': 0.0005069868639111519, 'timing/model': 0.9244184121489525, 'learning_rate/qwen_vl_interface': 1.1374611036629832e-06, 'learning_rate/action_model': 1.1374611036629831e-05, 'epoch': 2.65}) +[2026-06-18 01:43:12,213][starVLA.training.train_starvla][INFO] - Step 5850, Loss: {'train/loss': 0.024248458445072174, 'train/grad_norm_pre_clip': 6.6374664306640625, 'timing/data': 0.0004915352910757065, 'timing/model': 0.9371351087465882, 'learning_rate/qwen_vl_interface': 1.136379332023916e-06, 'learning_rate/action_model': 1.1363793320239161e-05, 'epoch': 2.65}) +[2026-06-18 01:43:16,119][starVLA.training.train_starvla][INFO] - Step 5851, Loss: {'train/loss': 0.10687124729156494, 'train/grad_norm_pre_clip': 25.865074157714844, 'timing/data': 0.0004652291536331177, 'timing/model': 0.9723109323531389, 'learning_rate/qwen_vl_interface': 1.1352984131428413e-06, 'learning_rate/action_model': 1.1352984131428413e-05, 'epoch': 2.65}) +[2026-06-18 01:43:19,861][starVLA.training.train_starvla][INFO] - Step 5852, Loss: {'train/loss': 0.314578652381897, 'train/grad_norm_pre_clip': 32.43219757080078, 'timing/data': 0.0008232938125729561, 'timing/model': 1.0866508726030588, 'learning_rate/qwen_vl_interface': 1.1342183472438329e-06, 'learning_rate/action_model': 1.1342183472438328e-05, 'epoch': 2.65}) +[2026-06-18 01:43:23,911][starVLA.training.train_starvla][INFO] - Step 5853, Loss: {'train/loss': 0.4367882311344147, 'train/grad_norm_pre_clip': 40.15690994262695, 'timing/data': 0.0005064290016889572, 'timing/model': 1.3777660597115755, 'learning_rate/qwen_vl_interface': 1.1331391345507905e-06, 'learning_rate/action_model': 1.1331391345507905e-05, 'epoch': 2.66}) +[2026-06-18 01:43:27,537][starVLA.training.train_starvla][INFO] - Step 5854, Loss: {'train/loss': 0.06443197280168533, 'train/grad_norm_pre_clip': 17.244340896606445, 'timing/data': 0.0004461100324988365, 'timing/model': 0.9229949116706848, 'learning_rate/qwen_vl_interface': 1.1320607752874341e-06, 'learning_rate/action_model': 1.1320607752874341e-05, 'epoch': 2.66}) +[2026-06-18 01:43:31,377][starVLA.training.train_starvla][INFO] - Step 5855, Loss: {'train/loss': 0.082760289311409, 'train/grad_norm_pre_clip': 13.834382057189941, 'timing/data': 0.0004719328135251999, 'timing/model': 0.928255838342011, 'learning_rate/qwen_vl_interface': 1.1309832696773104e-06, 'learning_rate/action_model': 1.1309832696773104e-05, 'epoch': 2.66}) +[2026-06-18 01:43:35,224][starVLA.training.train_starvla][INFO] - Step 5856, Loss: {'train/loss': 0.19758346676826477, 'train/grad_norm_pre_clip': 24.480722427368164, 'timing/data': 0.0007073776796460152, 'timing/model': 0.9409248614683747, 'learning_rate/qwen_vl_interface': 1.1299066179437876e-06, 'learning_rate/action_model': 1.1299066179437876e-05, 'epoch': 2.66}) +[2026-06-18 01:43:38,974][starVLA.training.train_starvla][INFO] - Step 5857, Loss: {'train/loss': 0.12211719155311584, 'train/grad_norm_pre_clip': 23.451231002807617, 'timing/data': 0.00043730903416872025, 'timing/model': 0.9906895412132144, 'learning_rate/qwen_vl_interface': 1.1288308203100565e-06, 'learning_rate/action_model': 1.1288308203100564e-05, 'epoch': 2.66}) +[2026-06-18 01:43:42,757][starVLA.training.train_starvla][INFO] - Step 5858, Loss: {'train/loss': 0.13529105484485626, 'train/grad_norm_pre_clip': 26.940996170043945, 'timing/data': 0.0004190690815448761, 'timing/model': 0.9219216890633106, 'learning_rate/qwen_vl_interface': 1.12775587699913e-06, 'learning_rate/action_model': 1.12775587699913e-05, 'epoch': 2.66}) +[2026-06-18 01:43:46,549][starVLA.training.train_starvla][INFO] - Step 5859, Loss: {'train/loss': 0.07156016677618027, 'train/grad_norm_pre_clip': 16.776233673095703, 'timing/data': 0.0005213115364313126, 'timing/model': 1.1073928559198976, 'learning_rate/qwen_vl_interface': 1.126681788233847e-06, 'learning_rate/action_model': 1.1266817882338468e-05, 'epoch': 2.66}) +[2026-06-18 01:43:50,121][starVLA.training.train_starvla][INFO] - Step 5860, Loss: {'train/loss': 0.04701778292655945, 'train/grad_norm_pre_clip': 13.576990127563477, 'timing/data': 0.0004918863996863365, 'timing/model': 0.9149255827069283, 'learning_rate/qwen_vl_interface': 1.1256085542368655e-06, 'learning_rate/action_model': 1.1256085542368655e-05, 'epoch': 2.66}) +[2026-06-18 01:43:53,962][starVLA.training.train_starvla][INFO] - Step 5861, Loss: {'train/loss': 0.20338183641433716, 'train/grad_norm_pre_clip': 33.70709228515625, 'timing/data': 0.0007717525586485863, 'timing/model': 0.9765864070504904, 'learning_rate/qwen_vl_interface': 1.1245361752306676e-06, 'learning_rate/action_model': 1.1245361752306677e-05, 'epoch': 2.66}) +[2026-06-18 01:43:57,753][starVLA.training.train_starvla][INFO] - Step 5862, Loss: {'train/loss': 0.310086190700531, 'train/grad_norm_pre_clip': 33.29673767089844, 'timing/data': 0.0008070077747106552, 'timing/model': 0.960862978361547, 'learning_rate/qwen_vl_interface': 1.1234646514375625e-06, 'learning_rate/action_model': 1.1234646514375624e-05, 'epoch': 2.66}) +[2026-06-18 01:44:01,626][starVLA.training.train_starvla][INFO] - Step 5863, Loss: {'train/loss': 0.22276054322719574, 'train/grad_norm_pre_clip': 30.972412109375, 'timing/data': 0.0005054017528891563, 'timing/model': 0.927277086302638, 'learning_rate/qwen_vl_interface': 1.122393983079674e-06, 'learning_rate/action_model': 1.122393983079674e-05, 'epoch': 2.66}) +[2026-06-18 01:44:05,517][starVLA.training.train_starvla][INFO] - Step 5864, Loss: {'train/loss': 0.07296334952116013, 'train/grad_norm_pre_clip': 19.466413497924805, 'timing/data': 0.0009117322042584419, 'timing/model': 1.0413486994802952, 'learning_rate/qwen_vl_interface': 1.1213241703789534e-06, 'learning_rate/action_model': 1.1213241703789534e-05, 'epoch': 2.66}) +[2026-06-18 01:44:11,579][starVLA.training.train_starvla][INFO] - Step 5865, Loss: {'train/loss': 0.11651057749986649, 'train/grad_norm_pre_clip': 19.479780197143555, 'timing/data': 0.0004689199849963188, 'timing/model': 0.9396991515532136, 'learning_rate/qwen_vl_interface': 1.1202552135571759e-06, 'learning_rate/action_model': 1.1202552135571759e-05, 'epoch': 2.66}) +[2026-06-18 01:44:16,015][starVLA.training.train_starvla][INFO] - Step 5866, Loss: {'train/loss': 0.08372151851654053, 'train/grad_norm_pre_clip': 19.232437133789062, 'timing/data': 0.0004325034096837044, 'timing/model': 1.0820776503533125, 'learning_rate/qwen_vl_interface': 1.1191871128359367e-06, 'learning_rate/action_model': 1.1191871128359365e-05, 'epoch': 2.66}) +[2026-06-18 01:44:19,662][starVLA.training.train_starvla][INFO] - Step 5867, Loss: {'train/loss': 0.030858639627695084, 'train/grad_norm_pre_clip': 6.754158020019531, 'timing/data': 0.0004957308992743492, 'timing/model': 0.9222971526905894, 'learning_rate/qwen_vl_interface': 1.1181198684366548e-06, 'learning_rate/action_model': 1.1181198684366548e-05, 'epoch': 2.66}) +[2026-06-18 01:44:23,427][starVLA.training.train_starvla][INFO] - Step 5868, Loss: {'train/loss': 0.047615714371204376, 'train/grad_norm_pre_clip': 11.437450408935547, 'timing/data': 0.000839604064822197, 'timing/model': 0.9267354011535645, 'learning_rate/qwen_vl_interface': 1.117053480580568e-06, 'learning_rate/action_model': 1.1170534805805679e-05, 'epoch': 2.66}) +[2026-06-18 01:44:27,309][starVLA.training.train_starvla][INFO] - Step 5869, Loss: {'train/loss': 0.048581812530756, 'train/grad_norm_pre_clip': 11.647114753723145, 'timing/data': 0.0010745776817202568, 'timing/model': 0.9590568887069821, 'learning_rate/qwen_vl_interface': 1.1159879494887429e-06, 'learning_rate/action_model': 1.115987949488743e-05, 'epoch': 2.66}) +[2026-06-18 01:44:31,140][starVLA.training.train_starvla][INFO] - Step 5870, Loss: {'train/loss': 0.5739916563034058, 'train/grad_norm_pre_clip': 55.22062301635742, 'timing/data': 0.00045173242688179016, 'timing/model': 0.9188759056851268, 'learning_rate/qwen_vl_interface': 1.1149232753820645e-06, 'learning_rate/action_model': 1.1149232753820647e-05, 'epoch': 2.66}) +[2026-06-18 01:44:34,896][starVLA.training.train_starvla][INFO] - Step 5871, Loss: {'train/loss': 0.2013045698404312, 'train/grad_norm_pre_clip': 15.95356559753418, 'timing/data': 0.0004990100860595703, 'timing/model': 0.9410905307158828, 'learning_rate/qwen_vl_interface': 1.1138594584812406e-06, 'learning_rate/action_model': 1.1138594584812406e-05, 'epoch': 2.66}) +[2026-06-18 01:44:38,667][starVLA.training.train_starvla][INFO] - Step 5872, Loss: {'train/loss': 0.1835145205259323, 'train/grad_norm_pre_clip': 25.520755767822266, 'timing/data': 0.00043851323425769806, 'timing/model': 1.094591828994453, 'learning_rate/qwen_vl_interface': 1.1127964990068017e-06, 'learning_rate/action_model': 1.1127964990068017e-05, 'epoch': 2.66}) +[2026-06-18 01:44:42,462][starVLA.training.train_starvla][INFO] - Step 5873, Loss: {'train/loss': 0.27112877368927, 'train/grad_norm_pre_clip': 40.3483772277832, 'timing/data': 0.0005321139469742775, 'timing/model': 0.9307450344786048, 'learning_rate/qwen_vl_interface': 1.1117343971791005e-06, 'learning_rate/action_model': 1.1117343971791006e-05, 'epoch': 2.66}) +[2026-06-18 01:44:46,235][starVLA.training.train_starvla][INFO] - Step 5874, Loss: {'train/loss': 0.08467746526002884, 'train/grad_norm_pre_clip': 32.50884246826172, 'timing/data': 0.0011790022253990173, 'timing/model': 0.9287359174340963, 'learning_rate/qwen_vl_interface': 1.1106731532183114e-06, 'learning_rate/action_model': 1.1106731532183114e-05, 'epoch': 2.66}) +[2026-06-18 01:44:50,022][starVLA.training.train_starvla][INFO] - Step 5875, Loss: {'train/loss': 0.03998372703790665, 'train/grad_norm_pre_clip': 11.783504486083984, 'timing/data': 0.0004179682582616806, 'timing/model': 0.9443808253854513, 'learning_rate/qwen_vl_interface': 1.1096127673444309e-06, 'learning_rate/action_model': 1.109612767344431e-05, 'epoch': 2.67}) +[2026-06-18 01:44:53,820][starVLA.training.train_starvla][INFO] - Step 5876, Loss: {'train/loss': 0.28649669885635376, 'train/grad_norm_pre_clip': 31.387237548828125, 'timing/data': 0.0005731349810957909, 'timing/model': 0.947385223582387, 'learning_rate/qwen_vl_interface': 1.108553239777281e-06, 'learning_rate/action_model': 1.108553239777281e-05, 'epoch': 2.67}) +[2026-06-18 01:44:57,653][starVLA.training.train_starvla][INFO] - Step 5877, Loss: {'train/loss': 0.15751409530639648, 'train/grad_norm_pre_clip': 35.238162994384766, 'timing/data': 0.0004996061325073242, 'timing/model': 0.9648736948147416, 'learning_rate/qwen_vl_interface': 1.1074945707365004e-06, 'learning_rate/action_model': 1.1074945707365003e-05, 'epoch': 2.67}) +[2026-06-18 01:45:01,487][starVLA.training.train_starvla][INFO] - Step 5878, Loss: {'train/loss': 0.08406943827867508, 'train/grad_norm_pre_clip': 16.21870231628418, 'timing/data': 0.0004412597045302391, 'timing/model': 0.9373907297849655, 'learning_rate/qwen_vl_interface': 1.1064367604415522e-06, 'learning_rate/action_model': 1.1064367604415521e-05, 'epoch': 2.67}) +[2026-06-18 01:45:05,358][starVLA.training.train_starvla][INFO] - Step 5879, Loss: {'train/loss': 0.23176997900009155, 'train/grad_norm_pre_clip': 26.45209312438965, 'timing/data': 0.0005167489871382713, 'timing/model': 1.1172201614826918, 'learning_rate/qwen_vl_interface': 1.105379809111721e-06, 'learning_rate/action_model': 1.105379809111721e-05, 'epoch': 2.67}) +[2026-06-18 01:45:09,139][starVLA.training.train_starvla][INFO] - Step 5880, Loss: {'train/loss': 0.510604977607727, 'train/grad_norm_pre_clip': 38.037227630615234, 'timing/data': 0.0005119489505887032, 'timing/model': 0.9677894962951541, 'learning_rate/qwen_vl_interface': 1.1043237169661167e-06, 'learning_rate/action_model': 1.1043237169661166e-05, 'epoch': 2.67}) +[2026-06-18 01:45:13,085][starVLA.training.train_starvla][INFO] - Step 5881, Loss: {'train/loss': 0.5095754265785217, 'train/grad_norm_pre_clip': 42.41976547241211, 'timing/data': 0.00043519120663404465, 'timing/model': 0.932292933575809, 'learning_rate/qwen_vl_interface': 1.103268484223667e-06, 'learning_rate/action_model': 1.1032684842236669e-05, 'epoch': 2.67}) +[2026-06-18 01:45:16,853][starVLA.training.train_starvla][INFO] - Step 5882, Loss: {'train/loss': 0.09009604156017303, 'train/grad_norm_pre_clip': 26.282180786132812, 'timing/data': 0.0004518544301390648, 'timing/model': 0.9267563773319125, 'learning_rate/qwen_vl_interface': 1.102214111103123e-06, 'learning_rate/action_model': 1.102214111103123e-05, 'epoch': 2.67}) +[2026-06-18 01:45:20,633][starVLA.training.train_starvla][INFO] - Step 5883, Loss: {'train/loss': 0.15469306707382202, 'train/grad_norm_pre_clip': 22.941577911376953, 'timing/data': 0.0004913369193673134, 'timing/model': 0.9590428089722991, 'learning_rate/qwen_vl_interface': 1.1011605978230552e-06, 'learning_rate/action_model': 1.1011605978230551e-05, 'epoch': 2.67}) +[2026-06-18 01:45:24,533][starVLA.training.train_starvla][INFO] - Step 5884, Loss: {'train/loss': 0.38741952180862427, 'train/grad_norm_pre_clip': 44.10284423828125, 'timing/data': 0.0006948541849851608, 'timing/model': 0.9331274479627609, 'learning_rate/qwen_vl_interface': 1.100107944601861e-06, 'learning_rate/action_model': 1.100107944601861e-05, 'epoch': 2.67}) +[2026-06-18 01:45:28,450][starVLA.training.train_starvla][INFO] - Step 5885, Loss: {'train/loss': 0.12112614512443542, 'train/grad_norm_pre_clip': 20.565977096557617, 'timing/data': 0.0005192188546061516, 'timing/model': 1.1434359094128013, 'learning_rate/qwen_vl_interface': 1.0990561516577552e-06, 'learning_rate/action_model': 1.0990561516577552e-05, 'epoch': 2.67}) +[2026-06-18 01:45:32,572][starVLA.training.train_starvla][INFO] - Step 5886, Loss: {'train/loss': 0.47328251600265503, 'train/grad_norm_pre_clip': 42.54762268066406, 'timing/data': 0.0008590631186962128, 'timing/model': 1.1603033328428864, 'learning_rate/qwen_vl_interface': 1.0980052192087753e-06, 'learning_rate/action_model': 1.0980052192087754e-05, 'epoch': 2.67}) +[2026-06-18 01:45:36,223][starVLA.training.train_starvla][INFO] - Step 5887, Loss: {'train/loss': 0.04420320689678192, 'train/grad_norm_pre_clip': 22.346616744995117, 'timing/data': 0.00047411397099494934, 'timing/model': 0.9501634882763028, 'learning_rate/qwen_vl_interface': 1.096955147472782e-06, 'learning_rate/action_model': 1.0969551474727821e-05, 'epoch': 2.67}) +[2026-06-18 01:45:39,998][starVLA.training.train_starvla][INFO] - Step 5888, Loss: {'train/loss': 0.24561867117881775, 'train/grad_norm_pre_clip': 33.03032302856445, 'timing/data': 0.0005448786541819572, 'timing/model': 0.937396015971899, 'learning_rate/qwen_vl_interface': 1.095905936667455e-06, 'learning_rate/action_model': 1.0959059366674552e-05, 'epoch': 2.67}) +[2026-06-18 01:45:43,819][starVLA.training.train_starvla][INFO] - Step 5889, Loss: {'train/loss': 0.430683970451355, 'train/grad_norm_pre_clip': 43.822265625, 'timing/data': 0.0004884200170636177, 'timing/model': 0.9440033063292503, 'learning_rate/qwen_vl_interface': 1.0948575870102977e-06, 'learning_rate/action_model': 1.0948575870102978e-05, 'epoch': 2.67}) +[2026-06-18 01:45:47,778][starVLA.training.train_starvla][INFO] - Step 5890, Loss: {'train/loss': 0.0816466435790062, 'train/grad_norm_pre_clip': 29.239307403564453, 'timing/data': 0.00043253786861896515, 'timing/model': 0.9252997720614076, 'learning_rate/qwen_vl_interface': 1.0938100987186325e-06, 'learning_rate/action_model': 1.0938100987186323e-05, 'epoch': 2.67}) +[2026-06-18 01:45:51,656][starVLA.training.train_starvla][INFO] - Step 5891, Loss: {'train/loss': 0.23977699875831604, 'train/grad_norm_pre_clip': 26.181917190551758, 'timing/data': 0.00045928917825222015, 'timing/model': 0.9259195756167173, 'learning_rate/qwen_vl_interface': 1.092763472009608e-06, 'learning_rate/action_model': 1.092763472009608e-05, 'epoch': 2.67}) +[2026-06-18 01:45:55,535][starVLA.training.train_starvla][INFO] - Step 5892, Loss: {'train/loss': 0.34016698598861694, 'train/grad_norm_pre_clip': 25.063350677490234, 'timing/data': 0.0006948532536625862, 'timing/model': 1.141128709539771, 'learning_rate/qwen_vl_interface': 1.0917177071001884e-06, 'learning_rate/action_model': 1.0917177071001884e-05, 'epoch': 2.67}) +[2026-06-18 01:45:59,297][starVLA.training.train_starvla][INFO] - Step 5893, Loss: {'train/loss': 0.23571659624576569, 'train/grad_norm_pre_clip': 36.55033493041992, 'timing/data': 0.0007364805787801743, 'timing/model': 1.1099570896476507, 'learning_rate/qwen_vl_interface': 1.0906728042071627e-06, 'learning_rate/action_model': 1.0906728042071627e-05, 'epoch': 2.67}) +[2026-06-18 01:46:02,973][starVLA.training.train_starvla][INFO] - Step 5894, Loss: {'train/loss': 0.13328060507774353, 'train/grad_norm_pre_clip': 13.414684295654297, 'timing/data': 0.00048592593520879745, 'timing/model': 1.0027707535773516, 'learning_rate/qwen_vl_interface': 1.089628763547139e-06, 'learning_rate/action_model': 1.0896287635471389e-05, 'epoch': 2.67}) +[2026-06-18 01:46:06,821][starVLA.training.train_starvla][INFO] - Step 5895, Loss: {'train/loss': 0.08291377127170563, 'train/grad_norm_pre_clip': 18.065088272094727, 'timing/data': 0.0007871994748711586, 'timing/model': 0.940619133412838, 'learning_rate/qwen_vl_interface': 1.0885855853365505e-06, 'learning_rate/action_model': 1.0885855853365503e-05, 'epoch': 2.67}) +[2026-06-18 01:46:10,569][starVLA.training.train_starvla][INFO] - Step 5896, Loss: {'train/loss': 0.20742996037006378, 'train/grad_norm_pre_clip': 30.598552703857422, 'timing/data': 0.0005115820094943047, 'timing/model': 0.9406620664522052, 'learning_rate/qwen_vl_interface': 1.0875432697916492e-06, 'learning_rate/action_model': 1.087543269791649e-05, 'epoch': 2.67}) +[2026-06-18 01:46:14,416][starVLA.training.train_starvla][INFO] - Step 5897, Loss: {'train/loss': 0.12088824808597565, 'train/grad_norm_pre_clip': 31.48779296875, 'timing/data': 0.0005400329828262329, 'timing/model': 0.9576951432973146, 'learning_rate/qwen_vl_interface': 1.0865018171285047e-06, 'learning_rate/action_model': 1.0865018171285047e-05, 'epoch': 2.68}) +[2026-06-18 01:46:19,372][starVLA.training.train_starvla][INFO] - Step 5898, Loss: {'train/loss': 0.3338082432746887, 'train/grad_norm_pre_clip': 30.828113555908203, 'timing/data': 0.0005002496764063835, 'timing/model': 1.0012508127838373, 'learning_rate/qwen_vl_interface': 1.0854612275630146e-06, 'learning_rate/action_model': 1.0854612275630146e-05, 'epoch': 2.68}) +[2026-06-18 01:46:24,104][starVLA.training.train_starvla][INFO] - Step 5899, Loss: {'train/loss': 0.05722388997673988, 'train/grad_norm_pre_clip': 15.383245468139648, 'timing/data': 0.0008156299591064453, 'timing/model': 1.1994986729696393, 'learning_rate/qwen_vl_interface': 1.0844215013108936e-06, 'learning_rate/action_model': 1.0844215013108936e-05, 'epoch': 2.68}) +[2026-06-18 01:48:35,424][starVLA.training.train_starvla][INFO] - Step 5900, Loss: {'train/loss': 0.048976872116327286, 'train/grad_norm_pre_clip': 16.850343704223633, 'eval/loss': 0.2619936466217041, 'eval/latency_8/loss': 0.2619936466217041, 'eval/demon_fire_precision': 0.9711479435236341, 'eval/demon_fire_recall': 0.9829139484311898, 'eval/demon_fire_f1': 0.9769955226184962, 'eval/demon_fire_rate_model': 0.9065108514190318, 'eval/demon_fire_rate_teacher': 0.8956594323873122, 'eval/demon_left_precision': 0.9278074866310161, 'eval/demon_left_recall': 0.7430406852248393, 'eval/demon_left_f1': 0.8252080856123661, 'eval/demon_left_rate_model': 0.10406232609905398, 'eval/demon_left_rate_teacher': 0.12993878686700056, 'eval/demon_right_precision': 0.8432510885341074, 'eval/demon_right_recall': 0.8803030303030303, 'eval/demon_right_f1': 0.8613787991104522, 'eval/demon_right_rate_model': 0.19170840289371174, 'eval/demon_right_rate_teacher': 0.18363939899833054, 'eval/demon_move_f1': 0.8432934423614091, 'eval/demon_cc_f1': 0.9326354142359609, 'eval/latency_8/demon_fire_precision': 0.9711479435236341, 'eval/latency_8/demon_fire_recall': 0.9829139484311898, 'eval/latency_8/demon_fire_f1': 0.9769955226184962, 'eval/latency_8/demon_fire_rate_model': 0.9065108514190318, 'eval/latency_8/demon_fire_rate_teacher': 0.8956594323873122, 'eval/latency_8/demon_left_precision': 0.9278074866310161, 'eval/latency_8/demon_left_recall': 0.7430406852248393, 'eval/latency_8/demon_left_f1': 0.8252080856123661, 'eval/latency_8/demon_left_rate_model': 0.10406232609905398, 'eval/latency_8/demon_left_rate_teacher': 0.12993878686700056, 'eval/latency_8/demon_right_precision': 0.8432510885341074, 'eval/latency_8/demon_right_recall': 0.8803030303030303, 'eval/latency_8/demon_right_f1': 0.8613787991104522, 'eval/latency_8/demon_right_rate_model': 0.19170840289371174, 'eval/latency_8/demon_right_rate_teacher': 0.18363939899833054, 'eval/latency_8/demon_move_f1': 0.8432934423614091, 'eval/latency_8/demon_cc_f1': 0.9326354142359609, 'timing/data': 0.0011556372046470642, 'timing/model': 1.1869560154154897, 'learning_rate/qwen_vl_interface': 1.0833826385876772e-06, 'learning_rate/action_model': 1.0833826385876771e-05, 'epoch': 2.68}) +[2026-06-18 01:48:35,429][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_5900_state +[2026-06-18 01:48:35,433][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 01:48:41,949][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_5900_state/model.safetensors +[2026-06-18 01:48:57,176][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_5900_state/optimizer.bin +[2026-06-18 01:48:57,176][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_5900_state/sampler.bin +[2026-06-18 01:48:57,177][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_5900_state/sampler_1.bin +[2026-06-18 01:48:57,178][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_5900_state/random_states_0.pkl +[2026-06-18 01:48:57,182][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 5900 +[2026-06-18 01:48:57,182][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 01:48:57,199][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 01:49:01,137][starVLA.training.train_starvla][INFO] - Step 5901, Loss: {'train/loss': 0.24214814603328705, 'train/grad_norm_pre_clip': 19.830482482910156, 'timing/data': 0.0005567120388150215, 'timing/model': 1.150400129146874, 'learning_rate/qwen_vl_interface': 1.0823446396087239e-06, 'learning_rate/action_model': 1.0823446396087239e-05, 'epoch': 2.68}) +[2026-06-18 01:49:04,910][starVLA.training.train_starvla][INFO] - Step 5902, Loss: {'train/loss': 0.04830768704414368, 'train/grad_norm_pre_clip': 10.74120044708252, 'timing/data': 0.0004593105986714363, 'timing/model': 1.0997683117166162, 'learning_rate/qwen_vl_interface': 1.0813075045892116e-06, 'learning_rate/action_model': 1.0813075045892116e-05, 'epoch': 2.68}) +[2026-06-18 01:49:08,640][starVLA.training.train_starvla][INFO] - Step 5903, Loss: {'train/loss': 0.19988127052783966, 'train/grad_norm_pre_clip': 19.89805030822754, 'timing/data': 0.018401626497507095, 'timing/model': 0.9875166164711118, 'learning_rate/qwen_vl_interface': 1.0802712337441383e-06, 'learning_rate/action_model': 1.0802712337441384e-05, 'epoch': 2.68}) +[2026-06-18 01:49:12,461][starVLA.training.train_starvla][INFO] - Step 5904, Loss: {'train/loss': 0.24847795069217682, 'train/grad_norm_pre_clip': 27.248489379882812, 'timing/data': 0.010363707318902016, 'timing/model': 0.978816956281662, 'learning_rate/qwen_vl_interface': 1.0792358272883262e-06, 'learning_rate/action_model': 1.0792358272883262e-05, 'epoch': 2.68}) +[2026-06-18 01:49:16,278][starVLA.training.train_starvla][INFO] - Step 5905, Loss: {'train/loss': 0.0786222442984581, 'train/grad_norm_pre_clip': 23.145172119140625, 'timing/data': 0.008722406812012196, 'timing/model': 0.9528614459559321, 'learning_rate/qwen_vl_interface': 1.0782012854364145e-06, 'learning_rate/action_model': 1.0782012854364145e-05, 'epoch': 2.68}) +[2026-06-18 01:49:20,289][starVLA.training.train_starvla][INFO] - Step 5906, Loss: {'train/loss': 0.10929518193006516, 'train/grad_norm_pre_clip': 18.66623306274414, 'timing/data': 0.00971510075032711, 'timing/model': 0.9547353917732835, 'learning_rate/qwen_vl_interface': 1.0771676084028652e-06, 'learning_rate/action_model': 1.0771676084028652e-05, 'epoch': 2.68}) +[2026-06-18 01:49:24,295][starVLA.training.train_starvla][INFO] - Step 5907, Loss: {'train/loss': 0.24995414912700653, 'train/grad_norm_pre_clip': 19.810871124267578, 'timing/data': 0.009645952843129635, 'timing/model': 0.9981582015752792, 'learning_rate/qwen_vl_interface': 1.0761347964019611e-06, 'learning_rate/action_model': 1.076134796401961e-05, 'epoch': 2.68}) +[2026-06-18 01:49:28,176][starVLA.training.train_starvla][INFO] - Step 5908, Loss: {'train/loss': 0.22450602054595947, 'train/grad_norm_pre_clip': 37.8796272277832, 'timing/data': 0.0070289792492985725, 'timing/model': 1.1175531018525362, 'learning_rate/qwen_vl_interface': 1.075102849647804e-06, 'learning_rate/action_model': 1.0751028496478038e-05, 'epoch': 2.68}) +[2026-06-18 01:49:31,844][starVLA.training.train_starvla][INFO] - Step 5909, Loss: {'train/loss': 0.30558204650878906, 'train/grad_norm_pre_clip': 41.66373825073242, 'timing/data': 0.009133116342127323, 'timing/model': 0.932223454117775, 'learning_rate/qwen_vl_interface': 1.0740717683543193e-06, 'learning_rate/action_model': 1.0740717683543191e-05, 'epoch': 2.68}) +[2026-06-18 01:49:35,694][starVLA.training.train_starvla][INFO] - Step 5910, Loss: {'train/loss': 0.07087656110525131, 'train/grad_norm_pre_clip': 24.214187622070312, 'timing/data': 0.02086404711008072, 'timing/model': 0.9491639779880643, 'learning_rate/qwen_vl_interface': 1.0730415527352518e-06, 'learning_rate/action_model': 1.0730415527352517e-05, 'epoch': 2.68}) +[2026-06-18 01:49:39,434][starVLA.training.train_starvla][INFO] - Step 5911, Loss: {'train/loss': 0.33559346199035645, 'train/grad_norm_pre_clip': 42.32905197143555, 'timing/data': 0.007972953841090202, 'timing/model': 0.944547669030726, 'learning_rate/qwen_vl_interface': 1.0720122030041638e-06, 'learning_rate/action_model': 1.0720122030041637e-05, 'epoch': 2.68}) +[2026-06-18 01:49:43,228][starVLA.training.train_starvla][INFO] - Step 5912, Loss: {'train/loss': 0.19538559019565582, 'train/grad_norm_pre_clip': 36.34052276611328, 'timing/data': 0.00968138687312603, 'timing/model': 0.9423301266506314, 'learning_rate/qwen_vl_interface': 1.0709837193744412e-06, 'learning_rate/action_model': 1.070983719374441e-05, 'epoch': 2.68}) +[2026-06-18 01:49:47,351][starVLA.training.train_starvla][INFO] - Step 5913, Loss: {'train/loss': 0.18629196286201477, 'train/grad_norm_pre_clip': 28.078022003173828, 'timing/data': 0.008113793097436428, 'timing/model': 0.9491670085117221, 'learning_rate/qwen_vl_interface': 1.069956102059291e-06, 'learning_rate/action_model': 1.069956102059291e-05, 'epoch': 2.68}) +[2026-06-18 01:49:53,915][starVLA.training.train_starvla][INFO] - Step 5914, Loss: {'train/loss': 0.2928982377052307, 'train/grad_norm_pre_clip': 24.174964904785156, 'timing/data': 0.008657282218337059, 'timing/model': 1.1372300824150443, 'learning_rate/qwen_vl_interface': 1.0689293512717387e-06, 'learning_rate/action_model': 1.0689293512717385e-05, 'epoch': 2.68}) +[2026-06-18 01:49:57,559][starVLA.training.train_starvla][INFO] - Step 5915, Loss: {'train/loss': 0.07850293070077896, 'train/grad_norm_pre_clip': 22.359893798828125, 'timing/data': 0.011071908287703991, 'timing/model': 0.9272030778229237, 'learning_rate/qwen_vl_interface': 1.0679034672246322e-06, 'learning_rate/action_model': 1.0679034672246322e-05, 'epoch': 2.68}) +[2026-06-18 01:50:01,320][starVLA.training.train_starvla][INFO] - Step 5916, Loss: {'train/loss': 0.10548419505357742, 'train/grad_norm_pre_clip': 17.373741149902344, 'timing/data': 0.009085345081984997, 'timing/model': 0.9415124775841832, 'learning_rate/qwen_vl_interface': 1.0668784501306348e-06, 'learning_rate/action_model': 1.0668784501306348e-05, 'epoch': 2.68}) +[2026-06-18 01:50:05,140][starVLA.training.train_starvla][INFO] - Step 5917, Loss: {'train/loss': 0.4644184112548828, 'train/grad_norm_pre_clip': 43.9558219909668, 'timing/data': 0.008729450404644012, 'timing/model': 0.927697223611176, 'learning_rate/qwen_vl_interface': 1.0658543002022377e-06, 'learning_rate/action_model': 1.0658543002022375e-05, 'epoch': 2.68}) +[2026-06-18 01:50:08,950][starVLA.training.train_starvla][INFO] - Step 5918, Loss: {'train/loss': 0.23154813051223755, 'train/grad_norm_pre_clip': 27.948427200317383, 'timing/data': 0.013194316998124123, 'timing/model': 0.9264526218175888, 'learning_rate/qwen_vl_interface': 1.0648310176517454e-06, 'learning_rate/action_model': 1.0648310176517453e-05, 'epoch': 2.68}) +[2026-06-18 01:50:12,733][starVLA.training.train_starvla][INFO] - Step 5919, Loss: {'train/loss': 0.29053109884262085, 'train/grad_norm_pre_clip': 32.105628967285156, 'timing/data': 0.008679820224642754, 'timing/model': 0.9258642653003335, 'learning_rate/qwen_vl_interface': 1.0638086026912873e-06, 'learning_rate/action_model': 1.0638086026912872e-05, 'epoch': 2.69}) +[2026-06-18 01:50:16,529][starVLA.training.train_starvla][INFO] - Step 5920, Loss: {'train/loss': 0.10184313356876373, 'train/grad_norm_pre_clip': 26.1185302734375, 'timing/data': 0.009420840069651604, 'timing/model': 1.102825816720724, 'learning_rate/qwen_vl_interface': 1.0627870555328102e-06, 'learning_rate/action_model': 1.06278705553281e-05, 'epoch': 2.69}) +[2026-06-18 01:50:20,288][starVLA.training.train_starvla][INFO] - Step 5921, Loss: {'train/loss': 0.15255795419216156, 'train/grad_norm_pre_clip': 18.7042179107666, 'timing/data': 0.00911663006991148, 'timing/model': 1.0812123995274305, 'learning_rate/qwen_vl_interface': 1.061766376388082e-06, 'learning_rate/action_model': 1.0617663763880821e-05, 'epoch': 2.69}) +[2026-06-18 01:50:23,958][starVLA.training.train_starvla][INFO] - Step 5922, Loss: {'train/loss': 0.31447434425354004, 'train/grad_norm_pre_clip': 32.993709564208984, 'timing/data': 0.01057721022516489, 'timing/model': 0.9549484141170979, 'learning_rate/qwen_vl_interface': 1.0607465654686908e-06, 'learning_rate/action_model': 1.0607465654686907e-05, 'epoch': 2.69}) +[2026-06-18 01:50:27,830][starVLA.training.train_starvla][INFO] - Step 5923, Loss: {'train/loss': 0.22076597809791565, 'train/grad_norm_pre_clip': 16.570741653442383, 'timing/data': 0.0071801068261265755, 'timing/model': 0.9482130212709308, 'learning_rate/qwen_vl_interface': 1.0597276229860434e-06, 'learning_rate/action_model': 1.0597276229860434e-05, 'epoch': 2.69}) +[2026-06-18 01:50:31,724][starVLA.training.train_starvla][INFO] - Step 5924, Loss: {'train/loss': 0.3752950131893158, 'train/grad_norm_pre_clip': 50.443546295166016, 'timing/data': 0.011757134459912777, 'timing/model': 0.9371663173660636, 'learning_rate/qwen_vl_interface': 1.0587095491513709e-06, 'learning_rate/action_model': 1.0587095491513709e-05, 'epoch': 2.69}) +[2026-06-18 01:50:35,500][starVLA.training.train_starvla][INFO] - Step 5925, Loss: {'train/loss': 0.0732397735118866, 'train/grad_norm_pre_clip': 15.334490776062012, 'timing/data': 0.01145391259342432, 'timing/model': 0.9222465138882399, 'learning_rate/qwen_vl_interface': 1.0576923441757174e-06, 'learning_rate/action_model': 1.0576923441757176e-05, 'epoch': 2.69}) +[2026-06-18 01:50:39,312][starVLA.training.train_starvla][INFO] - Step 5926, Loss: {'train/loss': 0.3207881450653076, 'train/grad_norm_pre_clip': 34.727508544921875, 'timing/data': 0.012638775631785393, 'timing/model': 0.9417248768731952, 'learning_rate/qwen_vl_interface': 1.0566760082699515e-06, 'learning_rate/action_model': 1.0566760082699514e-05, 'epoch': 2.69}) +[2026-06-18 01:50:43,116][starVLA.training.train_starvla][INFO] - Step 5927, Loss: {'train/loss': 0.14738525450229645, 'train/grad_norm_pre_clip': 33.21937942504883, 'timing/data': 0.009942930191755295, 'timing/model': 1.1274306112900376, 'learning_rate/qwen_vl_interface': 1.0556605416447626e-06, 'learning_rate/action_model': 1.0556605416447625e-05, 'epoch': 2.69}) +[2026-06-18 01:50:46,916][starVLA.training.train_starvla][INFO] - Step 5928, Loss: {'train/loss': 0.1892690658569336, 'train/grad_norm_pre_clip': 25.215747833251953, 'timing/data': 0.008908355608582497, 'timing/model': 1.1099289925768971, 'learning_rate/qwen_vl_interface': 1.054645944510656e-06, 'learning_rate/action_model': 1.054645944510656e-05, 'epoch': 2.69}) +[2026-06-18 01:50:50,527][starVLA.training.train_starvla][INFO] - Step 5929, Loss: {'train/loss': 0.11513182520866394, 'train/grad_norm_pre_clip': 24.856616973876953, 'timing/data': 0.018835192546248436, 'timing/model': 0.9388226205483079, 'learning_rate/qwen_vl_interface': 1.0536322170779605e-06, 'learning_rate/action_model': 1.0536322170779606e-05, 'epoch': 2.69}) +[2026-06-18 01:50:54,352][starVLA.training.train_starvla][INFO] - Step 5930, Loss: {'train/loss': 0.6813694834709167, 'train/grad_norm_pre_clip': 40.296993255615234, 'timing/data': 0.009022891521453857, 'timing/model': 0.9354682601988316, 'learning_rate/qwen_vl_interface': 1.0526193595568203e-06, 'learning_rate/action_model': 1.0526193595568203e-05, 'epoch': 2.69}) +[2026-06-18 01:50:58,131][starVLA.training.train_starvla][INFO] - Step 5931, Loss: {'train/loss': 0.17460311949253082, 'train/grad_norm_pre_clip': 23.274534225463867, 'timing/data': 0.008173679932951927, 'timing/model': 0.9436478279531002, 'learning_rate/qwen_vl_interface': 1.0516073721572043e-06, 'learning_rate/action_model': 1.0516073721572044e-05, 'epoch': 2.69}) +[2026-06-18 01:51:02,006][starVLA.training.train_starvla][INFO] - Step 5932, Loss: {'train/loss': 0.06082133576273918, 'train/grad_norm_pre_clip': 17.21516227722168, 'timing/data': 0.015150316059589386, 'timing/model': 0.963799862191081, 'learning_rate/qwen_vl_interface': 1.0505962550888982e-06, 'learning_rate/action_model': 1.0505962550888982e-05, 'epoch': 2.69}) +[2026-06-18 01:51:05,941][starVLA.training.train_starvla][INFO] - Step 5933, Loss: {'train/loss': 0.5019173622131348, 'train/grad_norm_pre_clip': 40.141151428222656, 'timing/data': 0.010375704616308212, 'timing/model': 0.9354514461010695, 'learning_rate/qwen_vl_interface': 1.0495860085615065e-06, 'learning_rate/action_model': 1.0495860085615064e-05, 'epoch': 2.69}) +[2026-06-18 01:51:10,128][starVLA.training.train_starvla][INFO] - Step 5934, Loss: {'train/loss': 0.6489251852035522, 'train/grad_norm_pre_clip': 54.128021240234375, 'timing/data': 0.0089896684512496, 'timing/model': 1.170581636019051, 'learning_rate/qwen_vl_interface': 1.0485766327844558e-06, 'learning_rate/action_model': 1.0485766327844557e-05, 'epoch': 2.69}) +[2026-06-18 01:51:14,813][starVLA.training.train_starvla][INFO] - Step 5935, Loss: {'train/loss': 0.28107646107673645, 'train/grad_norm_pre_clip': 17.022695541381836, 'timing/data': 0.009424337185919285, 'timing/model': 1.8214000258594751, 'learning_rate/qwen_vl_interface': 1.0475681279669892e-06, 'learning_rate/action_model': 1.0475681279669891e-05, 'epoch': 2.69}) +[2026-06-18 01:51:18,735][starVLA.training.train_starvla][INFO] - Step 5936, Loss: {'train/loss': 0.10379081219434738, 'train/grad_norm_pre_clip': 21.50455665588379, 'timing/data': 0.011027952656149864, 'timing/model': 0.9595855316147208, 'learning_rate/qwen_vl_interface': 1.0465604943181717e-06, 'learning_rate/action_model': 1.0465604943181717e-05, 'epoch': 2.69}) +[2026-06-18 01:51:22,703][starVLA.training.train_starvla][INFO] - Step 5937, Loss: {'train/loss': 0.09020498394966125, 'train/grad_norm_pre_clip': 26.473552703857422, 'timing/data': 0.012517490424215794, 'timing/model': 1.042248741723597, 'learning_rate/qwen_vl_interface': 1.0455537320468861e-06, 'learning_rate/action_model': 1.045553732046886e-05, 'epoch': 2.69}) +[2026-06-18 01:51:26,532][starVLA.training.train_starvla][INFO] - Step 5938, Loss: {'train/loss': 0.21975266933441162, 'train/grad_norm_pre_clip': 28.714096069335938, 'timing/data': 0.006943703629076481, 'timing/model': 0.9570563547313213, 'learning_rate/qwen_vl_interface': 1.0445478413618382e-06, 'learning_rate/action_model': 1.0445478413618382e-05, 'epoch': 2.69}) +[2026-06-18 01:51:30,401][starVLA.training.train_starvla][INFO] - Step 5939, Loss: {'train/loss': 0.05055199936032295, 'train/grad_norm_pre_clip': 17.60270118713379, 'timing/data': 0.01952600944787264, 'timing/model': 1.016712094657123, 'learning_rate/qwen_vl_interface': 1.0435428224715472e-06, 'learning_rate/action_model': 1.0435428224715471e-05, 'epoch': 2.69}) +[2026-06-18 01:51:34,196][starVLA.training.train_starvla][INFO] - Step 5940, Loss: {'train/loss': 0.023919977247714996, 'train/grad_norm_pre_clip': 6.87147855758667, 'timing/data': 0.00990764144808054, 'timing/model': 1.0934809893369675, 'learning_rate/qwen_vl_interface': 1.0425386755843547e-06, 'learning_rate/action_model': 1.0425386755843547e-05, 'epoch': 2.69}) +[2026-06-18 01:51:38,031][starVLA.training.train_starvla][INFO] - Step 5941, Loss: {'train/loss': 0.2591574192047119, 'train/grad_norm_pre_clip': 33.50202560424805, 'timing/data': 0.009109406732022762, 'timing/model': 1.098242332227528, 'learning_rate/qwen_vl_interface': 1.0415354009084219e-06, 'learning_rate/action_model': 1.041535400908422e-05, 'epoch': 2.7}) +[2026-06-18 01:51:41,637][starVLA.training.train_starvla][INFO] - Step 5942, Loss: {'train/loss': 0.3233133554458618, 'train/grad_norm_pre_clip': 45.868385314941406, 'timing/data': 0.009874642826616764, 'timing/model': 0.9291505422443151, 'learning_rate/qwen_vl_interface': 1.0405329986517301e-06, 'learning_rate/action_model': 1.0405329986517301e-05, 'epoch': 2.7}) +[2026-06-18 01:51:45,421][starVLA.training.train_starvla][INFO] - Step 5943, Loss: {'train/loss': 0.40185466408729553, 'train/grad_norm_pre_clip': 32.01940155029297, 'timing/data': 0.011585735715925694, 'timing/model': 0.9301218204200268, 'learning_rate/qwen_vl_interface': 1.0395314690220776e-06, 'learning_rate/action_model': 1.0395314690220776e-05, 'epoch': 2.7}) +[2026-06-18 01:51:49,426][starVLA.training.train_starvla][INFO] - Step 5944, Loss: {'train/loss': 0.2066245973110199, 'train/grad_norm_pre_clip': 35.1577033996582, 'timing/data': 0.010539780370891094, 'timing/model': 1.1341730765998363, 'learning_rate/qwen_vl_interface': 1.0385308122270815e-06, 'learning_rate/action_model': 1.0385308122270814e-05, 'epoch': 2.7}) +[2026-06-18 01:51:53,222][starVLA.training.train_starvla][INFO] - Step 5945, Loss: {'train/loss': 0.4489539563655853, 'train/grad_norm_pre_clip': 27.392179489135742, 'timing/data': 0.010560154914855957, 'timing/model': 0.9395449897274375, 'learning_rate/qwen_vl_interface': 1.0375310284741787e-06, 'learning_rate/action_model': 1.0375310284741788e-05, 'epoch': 2.7}) +[2026-06-18 01:51:57,231][starVLA.training.train_starvla][INFO] - Step 5946, Loss: {'train/loss': 0.23535434901714325, 'train/grad_norm_pre_clip': 25.250574111938477, 'timing/data': 0.009435232728719711, 'timing/model': 0.9304182920604944, 'learning_rate/qwen_vl_interface': 1.0365321179706276e-06, 'learning_rate/action_model': 1.0365321179706276e-05, 'epoch': 2.7}) +[2026-06-18 01:52:01,028][starVLA.training.train_starvla][INFO] - Step 5947, Loss: {'train/loss': 0.09933726489543915, 'train/grad_norm_pre_clip': 23.387514114379883, 'timing/data': 0.009134805761277676, 'timing/model': 1.1143172848969698, 'learning_rate/qwen_vl_interface': 1.035534080923503e-06, 'learning_rate/action_model': 1.0355340809235029e-05, 'epoch': 2.7}) +[2026-06-18 01:52:04,821][starVLA.training.train_starvla][INFO] - Step 5948, Loss: {'train/loss': 0.07722518593072891, 'train/grad_norm_pre_clip': 12.963573455810547, 'timing/data': 0.010565697215497494, 'timing/model': 1.1150356428697705, 'learning_rate/qwen_vl_interface': 1.0345369175396972e-06, 'learning_rate/action_model': 1.0345369175396972e-05, 'epoch': 2.7}) +[2026-06-18 01:52:08,430][starVLA.training.train_starvla][INFO] - Step 5949, Loss: {'train/loss': 0.21317006647586823, 'train/grad_norm_pre_clip': 21.138078689575195, 'timing/data': 0.009725748561322689, 'timing/model': 0.9345354260876775, 'learning_rate/qwen_vl_interface': 1.0335406280259255e-06, 'learning_rate/action_model': 1.0335406280259254e-05, 'epoch': 2.7}) +[2026-06-18 01:52:12,231][starVLA.training.train_starvla][INFO] - Step 5950, Loss: {'train/loss': 0.14226366579532623, 'train/grad_norm_pre_clip': 13.61208438873291, 'timing/data': 0.010872011072933674, 'timing/model': 0.9334950353950262, 'learning_rate/qwen_vl_interface': 1.0325452125887177e-06, 'learning_rate/action_model': 1.0325452125887178e-05, 'epoch': 2.7}) +[2026-06-18 01:52:16,036][starVLA.training.train_starvla][INFO] - Step 5951, Loss: {'train/loss': 0.2347874939441681, 'train/grad_norm_pre_clip': 43.528602600097656, 'timing/data': 0.012272261083126068, 'timing/model': 0.945113418623805, 'learning_rate/qwen_vl_interface': 1.031550671434426e-06, 'learning_rate/action_model': 1.031550671434426e-05, 'epoch': 2.7}) +[2026-06-18 01:52:19,882][starVLA.training.train_starvla][INFO] - Step 5952, Loss: {'train/loss': 0.07156068831682205, 'train/grad_norm_pre_clip': 18.534183502197266, 'timing/data': 0.009575978852808475, 'timing/model': 0.921758190728724, 'learning_rate/qwen_vl_interface': 1.0305570047692197e-06, 'learning_rate/action_model': 1.0305570047692195e-05, 'epoch': 2.7}) +[2026-06-18 01:52:23,662][starVLA.training.train_starvla][INFO] - Step 5953, Loss: {'train/loss': 0.21650534868240356, 'train/grad_norm_pre_clip': 39.72309875488281, 'timing/data': 0.010723749175667763, 'timing/model': 0.939636810682714, 'learning_rate/qwen_vl_interface': 1.0295642127990863e-06, 'learning_rate/action_model': 1.0295642127990863e-05, 'epoch': 2.7}) +[2026-06-18 01:52:27,523][starVLA.training.train_starvla][INFO] - Step 5954, Loss: {'train/loss': 0.05493118241429329, 'train/grad_norm_pre_clip': 12.516692161560059, 'timing/data': 0.009085128083825111, 'timing/model': 1.1354682436212897, 'learning_rate/qwen_vl_interface': 1.0285722957298326e-06, 'learning_rate/action_model': 1.0285722957298326e-05, 'epoch': 2.7}) +[2026-06-18 01:52:31,427][starVLA.training.train_starvla][INFO] - Step 5955, Loss: {'train/loss': 0.11038284003734589, 'train/grad_norm_pre_clip': 23.472524642944336, 'timing/data': 0.013085894286632538, 'timing/model': 0.9352842811495066, 'learning_rate/qwen_vl_interface': 1.027581253767085e-06, 'learning_rate/action_model': 1.027581253767085e-05, 'epoch': 2.7}) +[2026-06-18 01:52:35,331][starVLA.training.train_starvla][INFO] - Step 5956, Loss: {'train/loss': 0.020427783951163292, 'train/grad_norm_pre_clip': 3.7133164405822754, 'timing/data': 0.01378918532282114, 'timing/model': 0.9483440602198243, 'learning_rate/qwen_vl_interface': 1.0265910871162858e-06, 'learning_rate/action_model': 1.0265910871162858e-05, 'epoch': 2.7}) +[2026-06-18 01:52:39,088][starVLA.training.train_starvla][INFO] - Step 5957, Loss: {'train/loss': 0.024602359160780907, 'train/grad_norm_pre_clip': 7.401587963104248, 'timing/data': 0.012643035501241684, 'timing/model': 0.9385256301611662, 'learning_rate/qwen_vl_interface': 1.0256017959827008e-06, 'learning_rate/action_model': 1.0256017959827008e-05, 'epoch': 2.7}) +[2026-06-18 01:52:42,922][starVLA.training.train_starvla][INFO] - Step 5958, Loss: {'train/loss': 0.09220082312822342, 'train/grad_norm_pre_clip': 23.248674392700195, 'timing/data': 0.008436525240540504, 'timing/model': 0.9300741031765938, 'learning_rate/qwen_vl_interface': 1.0246133805714084e-06, 'learning_rate/action_model': 1.0246133805714084e-05, 'epoch': 2.7}) +[2026-06-18 01:52:48,820][starVLA.training.train_starvla][INFO] - Step 5959, Loss: {'train/loss': 0.05790761485695839, 'train/grad_norm_pre_clip': 14.594786643981934, 'timing/data': 0.009302820079028606, 'timing/model': 1.1164038218557835, 'learning_rate/qwen_vl_interface': 1.0236258410873076e-06, 'learning_rate/action_model': 1.0236258410873076e-05, 'epoch': 2.7}) +[2026-06-18 01:52:52,862][starVLA.training.train_starvla][INFO] - Step 5960, Loss: {'train/loss': 0.024618076160550117, 'train/grad_norm_pre_clip': 5.982157230377197, 'timing/data': 0.009869285859167576, 'timing/model': 1.1561490427702665, 'learning_rate/qwen_vl_interface': 1.0226391777351188e-06, 'learning_rate/action_model': 1.0226391777351189e-05, 'epoch': 2.7}) +[2026-06-18 01:52:56,767][starVLA.training.train_starvla][INFO] - Step 5961, Loss: {'train/loss': 0.4015065133571625, 'train/grad_norm_pre_clip': 28.876277923583984, 'timing/data': 0.008247416466474533, 'timing/model': 1.1612212704494596, 'learning_rate/qwen_vl_interface': 1.0216533907193777e-06, 'learning_rate/action_model': 1.0216533907193776e-05, 'epoch': 2.7}) +[2026-06-18 01:53:00,516][starVLA.training.train_starvla][INFO] - Step 5962, Loss: {'train/loss': 0.0539659708738327, 'train/grad_norm_pre_clip': 14.729580879211426, 'timing/data': 0.008358369581401348, 'timing/model': 0.9505037823691964, 'learning_rate/qwen_vl_interface': 1.020668480244438e-06, 'learning_rate/action_model': 1.020668480244438e-05, 'epoch': 2.7}) +[2026-06-18 01:53:06,197][starVLA.training.train_starvla][INFO] - Step 5963, Loss: {'train/loss': 0.4203440845012665, 'train/grad_norm_pre_clip': 26.11496925354004, 'timing/data': 0.008661878295242786, 'timing/model': 0.99785899464041, 'learning_rate/qwen_vl_interface': 1.0196844465144735e-06, 'learning_rate/action_model': 1.0196844465144734e-05, 'epoch': 2.71}) +[2026-06-18 01:53:11,603][starVLA.training.train_starvla][INFO] - Step 5964, Loss: {'train/loss': 0.059948742389678955, 'train/grad_norm_pre_clip': 18.65090560913086, 'timing/data': 0.007070389576256275, 'timing/model': 1.0140836359933019, 'learning_rate/qwen_vl_interface': 1.0187012897334755e-06, 'learning_rate/action_model': 1.0187012897334755e-05, 'epoch': 2.71}) +[2026-06-18 01:53:15,513][starVLA.training.train_starvla][INFO] - Step 5965, Loss: {'train/loss': 0.032955896109342575, 'train/grad_norm_pre_clip': 14.226595878601074, 'timing/data': 0.030770556069910526, 'timing/model': 0.9750879239290953, 'learning_rate/qwen_vl_interface': 1.017719010105253e-06, 'learning_rate/action_model': 1.017719010105253e-05, 'epoch': 2.71}) +[2026-06-18 01:53:19,417][starVLA.training.train_starvla][INFO] - Step 5966, Loss: {'train/loss': 0.017915302887558937, 'train/grad_norm_pre_clip': 3.866264820098877, 'timing/data': 0.009625528939068317, 'timing/model': 0.9499172177165747, 'learning_rate/qwen_vl_interface': 1.0167376078334337e-06, 'learning_rate/action_model': 1.0167376078334337e-05, 'epoch': 2.71}) +[2026-06-18 01:53:23,204][starVLA.training.train_starvla][INFO] - Step 5967, Loss: {'train/loss': 0.30073538422584534, 'train/grad_norm_pre_clip': 20.312326431274414, 'timing/data': 0.010323751717805862, 'timing/model': 1.1310902424156666, 'learning_rate/qwen_vl_interface': 1.0157570831214634e-06, 'learning_rate/action_model': 1.0157570831214634e-05, 'epoch': 2.71}) +[2026-06-18 01:53:26,980][starVLA.training.train_starvla][INFO] - Step 5968, Loss: {'train/loss': 0.09622229635715485, 'train/grad_norm_pre_clip': 21.065502166748047, 'timing/data': 0.010357682593166828, 'timing/model': 1.1166011337190866, 'learning_rate/qwen_vl_interface': 1.0147774361726066e-06, 'learning_rate/action_model': 1.0147774361726064e-05, 'epoch': 2.71}) +[2026-06-18 01:53:31,177][starVLA.training.train_starvla][INFO] - Step 5969, Loss: {'train/loss': 0.2163645178079605, 'train/grad_norm_pre_clip': 33.49632263183594, 'timing/data': 0.010070372372865677, 'timing/model': 1.1066916584968567, 'learning_rate/qwen_vl_interface': 1.013798667189944e-06, 'learning_rate/action_model': 1.013798667189944e-05, 'epoch': 2.71}) +[2026-06-18 01:53:35,129][starVLA.training.train_starvla][INFO] - Step 5970, Loss: {'train/loss': 0.046594444662332535, 'train/grad_norm_pre_clip': 8.872299194335938, 'timing/data': 0.009895674884319305, 'timing/model': 0.9328336343169212, 'learning_rate/qwen_vl_interface': 1.0128207763763755e-06, 'learning_rate/action_model': 1.0128207763763755e-05, 'epoch': 2.71}) +[2026-06-18 01:53:38,960][starVLA.training.train_starvla][INFO] - Step 5971, Loss: {'train/loss': 0.08932039886713028, 'train/grad_norm_pre_clip': 15.908917427062988, 'timing/data': 0.014495204202830791, 'timing/model': 0.9778839768841863, 'learning_rate/qwen_vl_interface': 1.0118437639346204e-06, 'learning_rate/action_model': 1.0118437639346204e-05, 'epoch': 2.71}) +[2026-06-18 01:53:42,765][starVLA.training.train_starvla][INFO] - Step 5972, Loss: {'train/loss': 0.023379845544695854, 'train/grad_norm_pre_clip': 5.257406711578369, 'timing/data': 0.012041636742651463, 'timing/model': 0.9354984983801842, 'learning_rate/qwen_vl_interface': 1.0108676300672144e-06, 'learning_rate/action_model': 1.0108676300672144e-05, 'epoch': 2.71}) +[2026-06-18 01:53:46,585][starVLA.training.train_starvla][INFO] - Step 5973, Loss: {'train/loss': 0.18457120656967163, 'train/grad_norm_pre_clip': 24.438547134399414, 'timing/data': 0.012045095674693584, 'timing/model': 0.9256076700985432, 'learning_rate/qwen_vl_interface': 1.0098923749765086e-06, 'learning_rate/action_model': 1.0098923749765086e-05, 'epoch': 2.71}) +[2026-06-18 01:53:50,649][starVLA.training.train_starvla][INFO] - Step 5974, Loss: {'train/loss': 0.12385103106498718, 'train/grad_norm_pre_clip': 35.83887481689453, 'timing/data': 0.008730362169444561, 'timing/model': 1.1286911247298121, 'learning_rate/qwen_vl_interface': 1.008917998864676e-06, 'learning_rate/action_model': 1.008917998864676e-05, 'epoch': 2.71}) +[2026-06-18 01:53:54,367][starVLA.training.train_starvla][INFO] - Step 5975, Loss: {'train/loss': 0.35480767488479614, 'train/grad_norm_pre_clip': 26.3154296875, 'timing/data': 0.009618646465241909, 'timing/model': 0.9473769012838602, 'learning_rate/qwen_vl_interface': 1.0079445019337063e-06, 'learning_rate/action_model': 1.0079445019337063e-05, 'epoch': 2.71}) +[2026-06-18 01:53:58,301][starVLA.training.train_starvla][INFO] - Step 5976, Loss: {'train/loss': 0.4502721428871155, 'train/grad_norm_pre_clip': 24.478015899658203, 'timing/data': 0.010467861779034138, 'timing/model': 0.9431435018777847, 'learning_rate/qwen_vl_interface': 1.0069718843854058e-06, 'learning_rate/action_model': 1.0069718843854058e-05, 'epoch': 2.71}) +[2026-06-18 01:54:02,058][starVLA.training.train_starvla][INFO] - Step 5977, Loss: {'train/loss': 0.12627406418323517, 'train/grad_norm_pre_clip': 26.000898361206055, 'timing/data': 0.008687696419656277, 'timing/model': 0.9486940111964941, 'learning_rate/qwen_vl_interface': 1.0060001464214007e-06, 'learning_rate/action_model': 1.0060001464214006e-05, 'epoch': 2.71}) +[2026-06-18 01:54:05,946][starVLA.training.train_starvla][INFO] - Step 5978, Loss: {'train/loss': 0.2404499650001526, 'train/grad_norm_pre_clip': 38.15095901489258, 'timing/data': 0.009652048349380493, 'timing/model': 0.9629033068194985, 'learning_rate/qwen_vl_interface': 1.00502928824313e-06, 'learning_rate/action_model': 1.00502928824313e-05, 'epoch': 2.71}) +[2026-06-18 01:54:09,884][starVLA.training.train_starvla][INFO] - Step 5979, Loss: {'train/loss': 0.10952174663543701, 'train/grad_norm_pre_clip': 24.290605545043945, 'timing/data': 0.007890413515269756, 'timing/model': 0.940153188072145, 'learning_rate/qwen_vl_interface': 1.0040593100518567e-06, 'learning_rate/action_model': 1.0040593100518566e-05, 'epoch': 2.71}) +[2026-06-18 01:54:14,142][starVLA.training.train_starvla][INFO] - Step 5980, Loss: {'train/loss': 0.15750855207443237, 'train/grad_norm_pre_clip': 15.827850341796875, 'timing/data': 0.007997904904186726, 'timing/model': 1.0999851487576962, 'learning_rate/qwen_vl_interface': 1.003090212048657e-06, 'learning_rate/action_model': 1.003090212048657e-05, 'epoch': 2.71}) +[2026-06-18 01:54:18,231][starVLA.training.train_starvla][INFO] - Step 5981, Loss: {'train/loss': 0.029706446453928947, 'train/grad_norm_pre_clip': 15.859905242919922, 'timing/data': 0.015123086981475353, 'timing/model': 1.1452533453702927, 'learning_rate/qwen_vl_interface': 1.0021219944344263e-06, 'learning_rate/action_model': 1.0021219944344262e-05, 'epoch': 2.71}) +[2026-06-18 01:54:23,331][starVLA.training.train_starvla][INFO] - Step 5982, Loss: {'train/loss': 0.4828985035419464, 'train/grad_norm_pre_clip': 55.84092330932617, 'timing/data': 0.01275398675352335, 'timing/model': 2.360326264053583, 'learning_rate/qwen_vl_interface': 1.0011546574098773e-06, 'learning_rate/action_model': 1.0011546574098772e-05, 'epoch': 2.71}) +[2026-06-18 01:54:27,417][starVLA.training.train_starvla][INFO] - Step 5983, Loss: {'train/loss': 0.13848769664764404, 'train/grad_norm_pre_clip': 26.7827205657959, 'timing/data': 0.00935471523553133, 'timing/model': 0.9567184317857027, 'learning_rate/qwen_vl_interface': 1.0001882011755396e-06, 'learning_rate/action_model': 1.0001882011755395e-05, 'epoch': 2.71}) +[2026-06-18 01:54:31,207][starVLA.training.train_starvla][INFO] - Step 5984, Loss: {'train/loss': 0.07010199129581451, 'train/grad_norm_pre_clip': 14.328913688659668, 'timing/data': 0.008878897875547409, 'timing/model': 0.9309596782550216, 'learning_rate/qwen_vl_interface': 9.992226259317605e-07, 'learning_rate/action_model': 9.992226259317605e-06, 'epoch': 2.71}) +[2026-06-18 01:54:36,653][starVLA.training.train_starvla][INFO] - Step 5985, Loss: {'train/loss': 0.1979612410068512, 'train/grad_norm_pre_clip': 28.778532028198242, 'timing/data': 0.0072322990745306015, 'timing/model': 0.9238925129175186, 'learning_rate/qwen_vl_interface': 9.982579318787044e-07, 'learning_rate/action_model': 9.982579318787044e-06, 'epoch': 2.72}) +[2026-06-18 01:54:40,433][starVLA.training.train_starvla][INFO] - Step 5986, Loss: {'train/loss': 0.1469414234161377, 'train/grad_norm_pre_clip': 19.62275505065918, 'timing/data': 0.007970675826072693, 'timing/model': 0.9356118813157082, 'learning_rate/qwen_vl_interface': 9.97294119216356e-07, 'learning_rate/action_model': 9.97294119216356e-06, 'epoch': 2.72}) +[2026-06-18 01:54:44,262][starVLA.training.train_starvla][INFO] - Step 5987, Loss: {'train/loss': 0.22939437627792358, 'train/grad_norm_pre_clip': 51.408897399902344, 'timing/data': 0.010034721344709396, 'timing/model': 1.1351260943338275, 'learning_rate/qwen_vl_interface': 9.96331188144512e-07, 'learning_rate/action_model': 9.96331188144512e-06, 'epoch': 2.72}) +[2026-06-18 01:54:48,128][starVLA.training.train_starvla][INFO] - Step 5988, Loss: {'train/loss': 0.09968171268701553, 'train/grad_norm_pre_clip': 27.203317642211914, 'timing/data': 0.010037546046078205, 'timing/model': 1.1193463755771518, 'learning_rate/qwen_vl_interface': 9.953691388627887e-07, 'learning_rate/action_model': 9.953691388627886e-06, 'epoch': 2.72}) +[2026-06-18 01:54:51,796][starVLA.training.train_starvla][INFO] - Step 5989, Loss: {'train/loss': 0.17277459800243378, 'train/grad_norm_pre_clip': 23.93410873413086, 'timing/data': 0.009693746455013752, 'timing/model': 0.9483604151755571, 'learning_rate/qwen_vl_interface': 9.94407971570622e-07, 'learning_rate/action_model': 9.94407971570622e-06, 'epoch': 2.72}) +[2026-06-18 01:54:55,589][starVLA.training.train_starvla][INFO] - Step 5990, Loss: {'train/loss': 0.2415241003036499, 'train/grad_norm_pre_clip': 29.51499366760254, 'timing/data': 0.01030490268021822, 'timing/model': 0.9381627375259995, 'learning_rate/qwen_vl_interface': 9.934476864672618e-07, 'learning_rate/action_model': 9.934476864672617e-06, 'epoch': 2.72}) +[2026-06-18 01:54:59,909][starVLA.training.train_starvla][INFO] - Step 5991, Loss: {'train/loss': 0.10267356038093567, 'train/grad_norm_pre_clip': 32.40931701660156, 'timing/data': 0.010012706741690636, 'timing/model': 0.9170685196295381, 'learning_rate/qwen_vl_interface': 9.924882837517781e-07, 'learning_rate/action_model': 9.924882837517782e-06, 'epoch': 2.72}) +[2026-06-18 01:55:03,733][starVLA.training.train_starvla][INFO] - Step 5992, Loss: {'train/loss': 0.31303831934928894, 'train/grad_norm_pre_clip': 30.788179397583008, 'timing/data': 0.009789420291781425, 'timing/model': 0.9393434394150972, 'learning_rate/qwen_vl_interface': 9.91529763623052e-07, 'learning_rate/action_model': 9.91529763623052e-06, 'epoch': 2.72}) +[2026-06-18 01:55:07,731][starVLA.training.train_starvla][INFO] - Step 5993, Loss: {'train/loss': 0.23202794790267944, 'train/grad_norm_pre_clip': 22.281017303466797, 'timing/data': 0.010752311907708645, 'timing/model': 1.0057450979948044, 'learning_rate/qwen_vl_interface': 9.905721262797894e-07, 'learning_rate/action_model': 9.905721262797892e-06, 'epoch': 2.72}) +[2026-06-18 01:55:11,587][starVLA.training.train_starvla][INFO] - Step 5994, Loss: {'train/loss': 0.04772551357746124, 'train/grad_norm_pre_clip': 10.019391059875488, 'timing/data': 0.009165479801595211, 'timing/model': 1.1426794724538922, 'learning_rate/qwen_vl_interface': 9.896153719205082e-07, 'learning_rate/action_model': 9.89615371920508e-06, 'epoch': 2.72}) +[2026-06-18 01:55:15,300][starVLA.training.train_starvla][INFO] - Step 5995, Loss: {'train/loss': 0.1681143343448639, 'train/grad_norm_pre_clip': 14.934078216552734, 'timing/data': 0.010048016905784607, 'timing/model': 0.9409699104726315, 'learning_rate/qwen_vl_interface': 9.886595007435448e-07, 'learning_rate/action_model': 9.886595007435447e-06, 'epoch': 2.72}) +[2026-06-18 01:55:19,052][starVLA.training.train_starvla][INFO] - Step 5996, Loss: {'train/loss': 0.08542574197053909, 'train/grad_norm_pre_clip': 18.73609733581543, 'timing/data': 0.013333829119801521, 'timing/model': 0.9207818657159805, 'learning_rate/qwen_vl_interface': 9.877045129470524e-07, 'learning_rate/action_model': 9.877045129470524e-06, 'epoch': 2.72}) +[2026-06-18 01:55:23,263][starVLA.training.train_starvla][INFO] - Step 5997, Loss: {'train/loss': 0.3547401428222656, 'train/grad_norm_pre_clip': 29.8613338470459, 'timing/data': 0.009906643070280552, 'timing/model': 0.9363407017663121, 'learning_rate/qwen_vl_interface': 9.86750408729001e-07, 'learning_rate/action_model': 9.867504087290009e-06, 'epoch': 2.72}) +[2026-06-18 01:55:27,139][starVLA.training.train_starvla][INFO] - Step 5998, Loss: {'train/loss': 0.13700230419635773, 'train/grad_norm_pre_clip': 45.763038635253906, 'timing/data': 0.013026091270148754, 'timing/model': 0.9448979208245873, 'learning_rate/qwen_vl_interface': 9.857971882871773e-07, 'learning_rate/action_model': 9.857971882871772e-06, 'epoch': 2.72}) +[2026-06-18 01:55:31,698][starVLA.training.train_starvla][INFO] - Step 5999, Loss: {'train/loss': 0.14245422184467316, 'train/grad_norm_pre_clip': 32.452945709228516, 'timing/data': 0.009321397170424461, 'timing/model': 1.0216669077053666, 'learning_rate/qwen_vl_interface': 9.848448518191838e-07, 'learning_rate/action_model': 9.848448518191836e-06, 'epoch': 2.72}) +[2026-06-18 01:57:36,215][starVLA.training.train_starvla][INFO] - Step 6000, Loss: {'train/loss': 0.07953370362520218, 'train/grad_norm_pre_clip': 19.067529678344727, 'eval/loss': 0.2623364329338074, 'eval/latency_8/loss': 0.2623364329338074, 'eval/demon_fire_precision': 0.9679975617189881, 'eval/demon_fire_recall': 0.9866418142280211, 'eval/demon_fire_f1': 0.9772307692307691, 'eval/demon_fire_rate_model': 0.9129104062326099, 'eval/demon_fire_rate_teacher': 0.8956594323873122, 'eval/demon_left_precision': 0.9088669950738916, 'eval/demon_left_recall': 0.7901498929336188, 'eval/demon_left_f1': 0.8453608247422679, 'eval/demon_left_rate_model': 0.11296605453533667, 'eval/demon_left_rate_teacher': 0.12993878686700056, 'eval/demon_right_precision': 0.8694992412746586, 'eval/demon_right_recall': 0.8681818181818182, 'eval/demon_right_f1': 0.8688400303260045, 'eval/demon_right_rate_model': 0.18336115748469672, 'eval/demon_right_rate_teacher': 0.18363939899833054, 'eval/demon_move_f1': 0.8571004275341363, 'eval/demon_cc_f1': 0.9377984983676196, 'eval/latency_8/demon_fire_precision': 0.9679975617189881, 'eval/latency_8/demon_fire_recall': 0.9866418142280211, 'eval/latency_8/demon_fire_f1': 0.9772307692307691, 'eval/latency_8/demon_fire_rate_model': 0.9129104062326099, 'eval/latency_8/demon_fire_rate_teacher': 0.8956594323873122, 'eval/latency_8/demon_left_precision': 0.9088669950738916, 'eval/latency_8/demon_left_recall': 0.7901498929336188, 'eval/latency_8/demon_left_f1': 0.8453608247422679, 'eval/latency_8/demon_left_rate_model': 0.11296605453533667, 'eval/latency_8/demon_left_rate_teacher': 0.12993878686700056, 'eval/latency_8/demon_right_precision': 0.8694992412746586, 'eval/latency_8/demon_right_recall': 0.8681818181818182, 'eval/latency_8/demon_right_f1': 0.8688400303260045, 'eval/latency_8/demon_right_rate_model': 0.18336115748469672, 'eval/latency_8/demon_right_rate_teacher': 0.18363939899833054, 'eval/latency_8/demon_move_f1': 0.8571004275341363, 'eval/latency_8/demon_cc_f1': 0.9377984983676196, 'timing/data': 0.010047566145658493, 'timing/model': 1.137773085385561, 'learning_rate/qwen_vl_interface': 9.838933995224439e-07, 'learning_rate/action_model': 9.83893399522444e-06, 'epoch': 2.72}) +[2026-06-18 01:57:36,216][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_6000_state +[2026-06-18 01:57:36,223][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 01:57:42,629][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_6000_state/model.safetensors +[2026-06-18 01:57:57,788][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_6000_state/optimizer.bin +[2026-06-18 01:57:57,789][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_6000_state/sampler.bin +[2026-06-18 01:57:57,789][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_6000_state/sampler_1.bin +[2026-06-18 01:57:57,791][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_6000_state/random_states_0.pkl +[2026-06-18 01:57:57,794][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 6000 +[2026-06-18 01:57:57,794][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 01:57:57,806][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 01:58:01,623][starVLA.training.train_starvla][INFO] - Step 6001, Loss: {'train/loss': 0.1647903323173523, 'train/grad_norm_pre_clip': 36.259521484375, 'timing/data': 0.01643703319132328, 'timing/model': 0.930408425629139, 'learning_rate/qwen_vl_interface': 9.829428315941921e-07, 'learning_rate/action_model': 9.829428315941922e-06, 'epoch': 2.72}) +[2026-06-18 01:58:05,338][starVLA.training.train_starvla][INFO] - Step 6002, Loss: {'train/loss': 0.028194071725010872, 'train/grad_norm_pre_clip': 8.93626880645752, 'timing/data': 0.008879470638930798, 'timing/model': 0.9246487962082028, 'learning_rate/qwen_vl_interface': 9.819931482314823e-07, 'learning_rate/action_model': 9.819931482314823e-06, 'epoch': 2.72}) +[2026-06-18 01:58:10,963][starVLA.training.train_starvla][INFO] - Step 6003, Loss: {'train/loss': 0.09646511822938919, 'train/grad_norm_pre_clip': 30.250898361206055, 'timing/data': 0.06573969684541225, 'timing/model': 2.8738233111798763, 'learning_rate/qwen_vl_interface': 9.81044349631185e-07, 'learning_rate/action_model': 9.81044349631185e-06, 'epoch': 2.72}) +[2026-06-18 01:58:15,543][starVLA.training.train_starvla][INFO] - Step 6004, Loss: {'train/loss': 0.16244706511497498, 'train/grad_norm_pre_clip': 29.631366729736328, 'timing/data': 0.013366990722715855, 'timing/model': 1.8619956690818071, 'learning_rate/qwen_vl_interface': 9.800964359899886e-07, 'learning_rate/action_model': 9.800964359899886e-06, 'epoch': 2.72}) +[2026-06-18 01:58:19,601][starVLA.training.train_starvla][INFO] - Step 6005, Loss: {'train/loss': 0.3183738887310028, 'train/grad_norm_pre_clip': 34.65798568725586, 'timing/data': 0.012951882556080818, 'timing/model': 0.9435884999111295, 'learning_rate/qwen_vl_interface': 9.791494075043963e-07, 'learning_rate/action_model': 9.791494075043963e-06, 'epoch': 2.72}) +[2026-06-18 01:58:23,430][starVLA.training.train_starvla][INFO] - Step 6006, Loss: {'train/loss': 0.1375548243522644, 'train/grad_norm_pre_clip': 20.705425262451172, 'timing/data': 0.01012520119547844, 'timing/model': 0.941853241994977, 'learning_rate/qwen_vl_interface': 9.78203264370727e-07, 'learning_rate/action_model': 9.78203264370727e-06, 'epoch': 2.72}) +[2026-06-18 01:58:27,412][starVLA.training.train_starvla][INFO] - Step 6007, Loss: {'train/loss': 0.15059347450733185, 'train/grad_norm_pre_clip': 19.43039321899414, 'timing/data': 0.008726225234568119, 'timing/model': 1.032265997491777, 'learning_rate/qwen_vl_interface': 9.772580067851162e-07, 'learning_rate/action_model': 9.772580067851161e-06, 'epoch': 2.73}) +[2026-06-18 01:58:32,436][starVLA.training.train_starvla][INFO] - Step 6008, Loss: {'train/loss': 0.3130405843257904, 'train/grad_norm_pre_clip': 30.845741271972656, 'timing/data': 0.01237707119435072, 'timing/model': 0.9626862294971943, 'learning_rate/qwen_vl_interface': 9.763136349435195e-07, 'learning_rate/action_model': 9.763136349435194e-06, 'epoch': 2.73}) +[2026-06-18 01:58:36,278][starVLA.training.train_starvla][INFO] - Step 6009, Loss: {'train/loss': 0.2023446261882782, 'train/grad_norm_pre_clip': 52.79304504394531, 'timing/data': 0.015451088547706604, 'timing/model': 0.9331253282725811, 'learning_rate/qwen_vl_interface': 9.753701490417048e-07, 'learning_rate/action_model': 9.753701490417047e-06, 'epoch': 2.73}) +[2026-06-18 01:58:40,143][starVLA.training.train_starvla][INFO] - Step 6010, Loss: {'train/loss': 0.04187484085559845, 'train/grad_norm_pre_clip': 13.375822067260742, 'timing/data': 0.016082962974905968, 'timing/model': 1.1000767229124904, 'learning_rate/qwen_vl_interface': 9.744275492752584e-07, 'learning_rate/action_model': 9.744275492752583e-06, 'epoch': 2.73}) +[2026-06-18 01:58:43,738][starVLA.training.train_starvla][INFO] - Step 6011, Loss: {'train/loss': 0.14751984179019928, 'train/grad_norm_pre_clip': 27.16978645324707, 'timing/data': 0.00871602538973093, 'timing/model': 0.9202776541933417, 'learning_rate/qwen_vl_interface': 9.734858358395808e-07, 'learning_rate/action_model': 9.734858358395807e-06, 'epoch': 2.73}) +[2026-06-18 01:58:47,673][starVLA.training.train_starvla][INFO] - Step 6012, Loss: {'train/loss': 0.19463969767093658, 'train/grad_norm_pre_clip': 27.3035945892334, 'timing/data': 0.01019563339650631, 'timing/model': 1.102548472583294, 'learning_rate/qwen_vl_interface': 9.72545008929892e-07, 'learning_rate/action_model': 9.725450089298921e-06, 'epoch': 2.73}) +[2026-06-18 01:58:51,535][starVLA.training.train_starvla][INFO] - Step 6013, Loss: {'train/loss': 0.2555293142795563, 'train/grad_norm_pre_clip': 27.708120346069336, 'timing/data': 0.019338756799697876, 'timing/model': 0.9563620612025261, 'learning_rate/qwen_vl_interface': 9.71605068741225e-07, 'learning_rate/action_model': 9.716050687412251e-06, 'epoch': 2.73}) +[2026-06-18 01:58:55,324][starVLA.training.train_starvla][INFO] - Step 6014, Loss: {'train/loss': 0.1020471528172493, 'train/grad_norm_pre_clip': 19.845304489135742, 'timing/data': 0.008436462841928005, 'timing/model': 0.9246462732553482, 'learning_rate/qwen_vl_interface': 9.706660154684312e-07, 'learning_rate/action_model': 9.706660154684312e-06, 'epoch': 2.73}) +[2026-06-18 01:58:59,177][starVLA.training.train_starvla][INFO] - Step 6015, Loss: {'train/loss': 0.21879208087921143, 'train/grad_norm_pre_clip': 24.175413131713867, 'timing/data': 0.008162431418895721, 'timing/model': 0.931799853220582, 'learning_rate/qwen_vl_interface': 9.697278493061773e-07, 'learning_rate/action_model': 9.697278493061773e-06, 'epoch': 2.73}) +[2026-06-18 01:59:03,020][starVLA.training.train_starvla][INFO] - Step 6016, Loss: {'train/loss': 0.08946885168552399, 'train/grad_norm_pre_clip': 18.749303817749023, 'timing/data': 0.011064690537750721, 'timing/model': 1.0929315462708473, 'learning_rate/qwen_vl_interface': 9.687905704489461e-07, 'learning_rate/action_model': 9.687905704489462e-06, 'epoch': 2.73}) +[2026-06-18 01:59:06,624][starVLA.training.train_starvla][INFO] - Step 6017, Loss: {'train/loss': 0.2007923126220703, 'train/grad_norm_pre_clip': 29.451183319091797, 'timing/data': 0.008174390532076359, 'timing/model': 0.9256414184346795, 'learning_rate/qwen_vl_interface': 9.678541790910361e-07, 'learning_rate/action_model': 9.678541790910361e-06, 'epoch': 2.73}) +[2026-06-18 01:59:10,726][starVLA.training.train_starvla][INFO] - Step 6018, Loss: {'train/loss': 0.15006288886070251, 'train/grad_norm_pre_clip': 34.349693298339844, 'timing/data': 0.014123926870524883, 'timing/model': 1.1780155608430505, 'learning_rate/qwen_vl_interface': 9.669186754265619e-07, 'learning_rate/action_model': 9.669186754265618e-06, 'epoch': 2.73}) +[2026-06-18 01:59:14,524][starVLA.training.train_starvla][INFO] - Step 6019, Loss: {'train/loss': 0.2623860239982605, 'train/grad_norm_pre_clip': 30.656282424926758, 'timing/data': 0.017468156293034554, 'timing/model': 0.939635256305337, 'learning_rate/qwen_vl_interface': 9.65984059649457e-07, 'learning_rate/action_model': 9.65984059649457e-06, 'epoch': 2.73}) +[2026-06-18 01:59:18,530][starVLA.training.train_starvla][INFO] - Step 6020, Loss: {'train/loss': 0.18964296579360962, 'train/grad_norm_pre_clip': 31.319019317626953, 'timing/data': 0.009707946330308914, 'timing/model': 0.9229858107864857, 'learning_rate/qwen_vl_interface': 9.650503319534652e-07, 'learning_rate/action_model': 9.650503319534651e-06, 'epoch': 2.73}) +[2026-06-18 01:59:22,314][starVLA.training.train_starvla][INFO] - Step 6021, Loss: {'train/loss': 0.22238045930862427, 'train/grad_norm_pre_clip': 39.644962310791016, 'timing/data': 0.009636069647967815, 'timing/model': 0.9320873292163014, 'learning_rate/qwen_vl_interface': 9.641174925321494e-07, 'learning_rate/action_model': 9.641174925321494e-06, 'epoch': 2.73}) +[2026-06-18 01:59:26,334][starVLA.training.train_starvla][INFO] - Step 6022, Loss: {'train/loss': 0.10679115355014801, 'train/grad_norm_pre_clip': 28.964006423950195, 'timing/data': 0.009064735844731331, 'timing/model': 1.0205060224980116, 'learning_rate/qwen_vl_interface': 9.631855415788904e-07, 'learning_rate/action_model': 9.631855415788904e-06, 'epoch': 2.73}) +[2026-06-18 01:59:30,204][starVLA.training.train_starvla][INFO] - Step 6023, Loss: {'train/loss': 0.3548388183116913, 'train/grad_norm_pre_clip': 28.746492385864258, 'timing/data': 0.016332062892615795, 'timing/model': 1.1165973469614983, 'learning_rate/qwen_vl_interface': 9.622544792868812e-07, 'learning_rate/action_model': 9.622544792868812e-06, 'epoch': 2.73}) +[2026-06-18 01:59:33,888][starVLA.training.train_starvla][INFO] - Step 6024, Loss: {'train/loss': 0.32879284024238586, 'train/grad_norm_pre_clip': 98.5449447631836, 'timing/data': 0.01262713223695755, 'timing/model': 0.9364147922024131, 'learning_rate/qwen_vl_interface': 9.613243058491334e-07, 'learning_rate/action_model': 9.613243058491334e-06, 'epoch': 2.73}) +[2026-06-18 01:59:37,728][starVLA.training.train_starvla][INFO] - Step 6025, Loss: {'train/loss': 0.14749476313591003, 'train/grad_norm_pre_clip': 28.43196678161621, 'timing/data': 0.008434451185166836, 'timing/model': 0.9288352988660336, 'learning_rate/qwen_vl_interface': 9.603950214584695e-07, 'learning_rate/action_model': 9.603950214584696e-06, 'epoch': 2.73}) +[2026-06-18 01:59:41,635][starVLA.training.train_starvla][INFO] - Step 6026, Loss: {'train/loss': 0.29986169934272766, 'train/grad_norm_pre_clip': 24.192106246948242, 'timing/data': 0.010122429579496384, 'timing/model': 1.0523640485480428, 'learning_rate/qwen_vl_interface': 9.59466626307535e-07, 'learning_rate/action_model': 9.594666263075349e-06, 'epoch': 2.73}) +[2026-06-18 01:59:45,592][starVLA.training.train_starvla][INFO] - Step 6027, Loss: {'train/loss': 0.11157462745904922, 'train/grad_norm_pre_clip': 16.301443099975586, 'timing/data': 0.009003539569675922, 'timing/model': 0.9371237559244037, 'learning_rate/qwen_vl_interface': 9.585391205887851e-07, 'learning_rate/action_model': 9.585391205887851e-06, 'epoch': 2.73}) +[2026-06-18 01:59:49,406][starVLA.training.train_starvla][INFO] - Step 6028, Loss: {'train/loss': 0.042328156530857086, 'train/grad_norm_pre_clip': 7.979053974151611, 'timing/data': 0.00888882763683796, 'timing/model': 0.9431528253480792, 'learning_rate/qwen_vl_interface': 9.57612504494494e-07, 'learning_rate/action_model': 9.57612504494494e-06, 'epoch': 2.73}) +[2026-06-18 01:59:53,237][starVLA.training.train_starvla][INFO] - Step 6029, Loss: {'train/loss': 0.16216681897640228, 'train/grad_norm_pre_clip': 27.22406578063965, 'timing/data': 0.012380794622004032, 'timing/model': 1.1518016252666712, 'learning_rate/qwen_vl_interface': 9.566867782167494e-07, 'learning_rate/action_model': 9.566867782167494e-06, 'epoch': 2.74}) +[2026-06-18 01:59:56,923][starVLA.training.train_starvla][INFO] - Step 6030, Loss: {'train/loss': 0.07647936791181564, 'train/grad_norm_pre_clip': 16.222593307495117, 'timing/data': 0.01484429370611906, 'timing/model': 1.0326886950060725, 'learning_rate/qwen_vl_interface': 9.55761941947455e-07, 'learning_rate/action_model': 9.557619419474551e-06, 'epoch': 2.74}) +[2026-06-18 02:00:00,602][starVLA.training.train_starvla][INFO] - Step 6031, Loss: {'train/loss': 0.05893027037382126, 'train/grad_norm_pre_clip': 12.76893424987793, 'timing/data': 0.009409870021045208, 'timing/model': 0.9829964973032475, 'learning_rate/qwen_vl_interface': 9.548379958783308e-07, 'learning_rate/action_model': 9.548379958783308e-06, 'epoch': 2.74}) +[2026-06-18 02:00:04,671][starVLA.training.train_starvla][INFO] - Step 6032, Loss: {'train/loss': 0.1527816355228424, 'train/grad_norm_pre_clip': 39.28242874145508, 'timing/data': 0.011394224129617214, 'timing/model': 0.9243367994204164, 'learning_rate/qwen_vl_interface': 9.539149402009114e-07, 'learning_rate/action_model': 9.539149402009115e-06, 'epoch': 2.74}) +[2026-06-18 02:00:08,954][starVLA.training.train_starvla][INFO] - Step 6033, Loss: {'train/loss': 0.22200429439544678, 'train/grad_norm_pre_clip': 42.456974029541016, 'timing/data': 0.013651255518198013, 'timing/model': 0.9319465449079871, 'learning_rate/qwen_vl_interface': 9.529927751065493e-07, 'learning_rate/action_model': 9.529927751065492e-06, 'epoch': 2.74}) +[2026-06-18 02:00:12,813][starVLA.training.train_starvla][INFO] - Step 6034, Loss: {'train/loss': 0.08930167555809021, 'train/grad_norm_pre_clip': 26.183183670043945, 'timing/data': 0.015067294239997864, 'timing/model': 0.9437781702727079, 'learning_rate/qwen_vl_interface': 9.520715007864082e-07, 'learning_rate/action_model': 9.52071500786408e-06, 'epoch': 2.74}) +[2026-06-18 02:00:16,556][starVLA.training.train_starvla][INFO] - Step 6035, Loss: {'train/loss': 0.10578934848308563, 'train/grad_norm_pre_clip': 17.095970153808594, 'timing/data': 0.016511418856680393, 'timing/model': 0.935095633380115, 'learning_rate/qwen_vl_interface': 9.511511174314689e-07, 'learning_rate/action_model': 9.511511174314688e-06, 'epoch': 2.74}) +[2026-06-18 02:00:20,331][starVLA.training.train_starvla][INFO] - Step 6036, Loss: {'train/loss': 0.1196659728884697, 'train/grad_norm_pre_clip': 15.42768669128418, 'timing/data': 0.014011327177286148, 'timing/model': 1.0981456311419606, 'learning_rate/qwen_vl_interface': 9.502316252325283e-07, 'learning_rate/action_model': 9.502316252325283e-06, 'epoch': 2.74}) +[2026-06-18 02:00:24,043][starVLA.training.train_starvla][INFO] - Step 6037, Loss: {'train/loss': 0.10271468013525009, 'train/grad_norm_pre_clip': 18.496383666992188, 'timing/data': 0.008392860181629658, 'timing/model': 0.9836570573970675, 'learning_rate/qwen_vl_interface': 9.493130243801993e-07, 'learning_rate/action_model': 9.493130243801992e-06, 'epoch': 2.74}) +[2026-06-18 02:00:27,822][starVLA.training.train_starvla][INFO] - Step 6038, Loss: {'train/loss': 0.3175370991230011, 'train/grad_norm_pre_clip': 36.26099395751953, 'timing/data': 0.01114278007298708, 'timing/model': 0.9283659858629107, 'learning_rate/qwen_vl_interface': 9.483953150649091e-07, 'learning_rate/action_model': 9.483953150649091e-06, 'epoch': 2.74}) +[2026-06-18 02:00:32,813][starVLA.training.train_starvla][INFO] - Step 6039, Loss: {'train/loss': 0.16712574660778046, 'train/grad_norm_pre_clip': 22.798131942749023, 'timing/data': 0.007796269841492176, 'timing/model': 0.9616223173215985, 'learning_rate/qwen_vl_interface': 9.474784974768982e-07, 'learning_rate/action_model': 9.474784974768982e-06, 'epoch': 2.74}) +[2026-06-18 02:00:36,774][starVLA.training.train_starvla][INFO] - Step 6040, Loss: {'train/loss': 0.06306682527065277, 'train/grad_norm_pre_clip': 258.5664367675781, 'timing/data': 0.007310943678021431, 'timing/model': 0.9645590931177139, 'learning_rate/qwen_vl_interface': 9.465625718062236e-07, 'learning_rate/action_model': 9.465625718062235e-06, 'epoch': 2.74}) +[2026-06-18 02:00:40,725][starVLA.training.train_starvla][INFO] - Step 6041, Loss: {'train/loss': 0.33083003759384155, 'train/grad_norm_pre_clip': 38.47543716430664, 'timing/data': 0.009279112331569195, 'timing/model': 0.9547951621934772, 'learning_rate/qwen_vl_interface': 9.456475382427591e-07, 'learning_rate/action_model': 9.45647538242759e-06, 'epoch': 2.74}) +[2026-06-18 02:00:44,480][starVLA.training.train_starvla][INFO] - Step 6042, Loss: {'train/loss': 0.289145827293396, 'train/grad_norm_pre_clip': 32.6733512878418, 'timing/data': 0.009730761870741844, 'timing/model': 1.1056157955899835, 'learning_rate/qwen_vl_interface': 9.447333969761915e-07, 'learning_rate/action_model': 9.447333969761915e-06, 'epoch': 2.74}) +[2026-06-18 02:00:48,430][starVLA.training.train_starvla][INFO] - Step 6043, Loss: {'train/loss': 0.27908921241760254, 'train/grad_norm_pre_clip': 132.99366760253906, 'timing/data': 0.008828078396618366, 'timing/model': 0.927561204880476, 'learning_rate/qwen_vl_interface': 9.438201481960235e-07, 'learning_rate/action_model': 9.438201481960234e-06, 'epoch': 2.74}) +[2026-06-18 02:00:52,290][starVLA.training.train_starvla][INFO] - Step 6044, Loss: {'train/loss': 0.3492174744606018, 'train/grad_norm_pre_clip': 29.134410858154297, 'timing/data': 0.008393463678658009, 'timing/model': 0.9951423918828368, 'learning_rate/qwen_vl_interface': 9.429077920915724e-07, 'learning_rate/action_model': 9.429077920915724e-06, 'epoch': 2.74}) +[2026-06-18 02:00:56,183][starVLA.training.train_starvla][INFO] - Step 6045, Loss: {'train/loss': 0.17901799082756042, 'train/grad_norm_pre_clip': 28.25763702392578, 'timing/data': 0.008442949503660202, 'timing/model': 0.9474390111863613, 'learning_rate/qwen_vl_interface': 9.419963288519711e-07, 'learning_rate/action_model': 9.41996328851971e-06, 'epoch': 2.74}) +[2026-06-18 02:01:00,129][starVLA.training.train_starvla][INFO] - Step 6046, Loss: {'train/loss': 0.055910658091306686, 'train/grad_norm_pre_clip': 14.799736976623535, 'timing/data': 0.009043127298355103, 'timing/model': 0.9847632348537445, 'learning_rate/qwen_vl_interface': 9.410857586661659e-07, 'learning_rate/action_model': 9.410857586661659e-06, 'epoch': 2.74}) +[2026-06-18 02:01:05,921][starVLA.training.train_starvla][INFO] - Step 6047, Loss: {'train/loss': 0.31340551376342773, 'train/grad_norm_pre_clip': 28.498668670654297, 'timing/data': 0.009169408120214939, 'timing/model': 2.745936546474695, 'learning_rate/qwen_vl_interface': 9.401760817229197e-07, 'learning_rate/action_model': 9.401760817229196e-06, 'epoch': 2.74}) +[2026-06-18 02:01:09,900][starVLA.training.train_starvla][INFO] - Step 6048, Loss: {'train/loss': 0.22566550970077515, 'train/grad_norm_pre_clip': 22.649398803710938, 'timing/data': 0.008879820816218853, 'timing/model': 1.142197814770043, 'learning_rate/qwen_vl_interface': 9.392672982108091e-07, 'learning_rate/action_model': 9.392672982108091e-06, 'epoch': 2.74}) +[2026-06-18 02:01:15,425][starVLA.training.train_starvla][INFO] - Step 6049, Loss: {'train/loss': 0.11146571487188339, 'train/grad_norm_pre_clip': 21.646198272705078, 'timing/data': 0.012782236561179161, 'timing/model': 0.9592853393405676, 'learning_rate/qwen_vl_interface': 9.383594083182265e-07, 'learning_rate/action_model': 9.383594083182265e-06, 'epoch': 2.74}) +[2026-06-18 02:01:19,502][starVLA.training.train_starvla][INFO] - Step 6050, Loss: {'train/loss': 0.2556183338165283, 'train/grad_norm_pre_clip': 39.18698501586914, 'timing/data': 0.00801145564764738, 'timing/model': 0.9379308968782425, 'learning_rate/qwen_vl_interface': 9.374524122333769e-07, 'learning_rate/action_model': 9.374524122333769e-06, 'epoch': 2.74}) +[2026-06-18 02:01:23,328][starVLA.training.train_starvla][INFO] - Step 6051, Loss: {'train/loss': 0.16760818660259247, 'train/grad_norm_pre_clip': 30.275325775146484, 'timing/data': 0.00798182561993599, 'timing/model': 0.9620945826172829, 'learning_rate/qwen_vl_interface': 9.365463101442842e-07, 'learning_rate/action_model': 9.365463101442842e-06, 'epoch': 2.75}) +[2026-06-18 02:01:27,235][starVLA.training.train_starvla][INFO] - Step 6052, Loss: {'train/loss': 0.024370601400732994, 'train/grad_norm_pre_clip': 5.270589828491211, 'timing/data': 0.009036889299750328, 'timing/model': 0.9459597980603576, 'learning_rate/qwen_vl_interface': 9.356411022387823e-07, 'learning_rate/action_model': 9.356411022387824e-06, 'epoch': 2.75}) +[2026-06-18 02:01:31,019][starVLA.training.train_starvla][INFO] - Step 6053, Loss: {'train/loss': 0.1663566380739212, 'train/grad_norm_pre_clip': 37.23966598510742, 'timing/data': 0.008675425313413143, 'timing/model': 0.9277135310694575, 'learning_rate/qwen_vl_interface': 9.347367887045236e-07, 'learning_rate/action_model': 9.347367887045236e-06, 'epoch': 2.75}) +[2026-06-18 02:01:34,788][starVLA.training.train_starvla][INFO] - Step 6054, Loss: {'train/loss': 0.04061185196042061, 'train/grad_norm_pre_clip': 9.072158813476562, 'timing/data': 0.018427731469273567, 'timing/model': 0.9835278587415814, 'learning_rate/qwen_vl_interface': 9.338333697289697e-07, 'learning_rate/action_model': 9.338333697289696e-06, 'epoch': 2.75}) +[2026-06-18 02:01:38,796][starVLA.training.train_starvla][INFO] - Step 6055, Loss: {'train/loss': 0.4146743416786194, 'train/grad_norm_pre_clip': 37.04209518432617, 'timing/data': 0.008100445382297039, 'timing/model': 1.0613365843892097, 'learning_rate/qwen_vl_interface': 9.329308454994044e-07, 'learning_rate/action_model': 9.329308454994043e-06, 'epoch': 2.75}) +[2026-06-18 02:01:42,563][starVLA.training.train_starvla][INFO] - Step 6056, Loss: {'train/loss': 0.2658270299434662, 'train/grad_norm_pre_clip': 27.27271270751953, 'timing/data': 0.013054149225354195, 'timing/model': 0.9219133518636227, 'learning_rate/qwen_vl_interface': 9.320292162029192e-07, 'learning_rate/action_model': 9.320292162029191e-06, 'epoch': 2.75}) +[2026-06-18 02:01:46,336][starVLA.training.train_starvla][INFO] - Step 6057, Loss: {'train/loss': 0.47185438871383667, 'train/grad_norm_pre_clip': 32.04575729370117, 'timing/data': 0.007082315161824226, 'timing/model': 0.9393861507996917, 'learning_rate/qwen_vl_interface': 9.311284820264245e-07, 'learning_rate/action_model': 9.311284820264244e-06, 'epoch': 2.75}) +[2026-06-18 02:01:50,303][starVLA.training.train_starvla][INFO] - Step 6058, Loss: {'train/loss': 0.25906869769096375, 'train/grad_norm_pre_clip': 50.96139144897461, 'timing/data': 0.008648809045553207, 'timing/model': 0.9323316840454936, 'learning_rate/qwen_vl_interface': 9.302286431566428e-07, 'learning_rate/action_model': 9.302286431566428e-06, 'epoch': 2.75}) +[2026-06-18 02:01:54,117][starVLA.training.train_starvla][INFO] - Step 6059, Loss: {'train/loss': 0.2416776865720749, 'train/grad_norm_pre_clip': 36.192100524902344, 'timing/data': 0.0089187640696764, 'timing/model': 0.9529808070510626, 'learning_rate/qwen_vl_interface': 9.293296997801113e-07, 'learning_rate/action_model': 9.293296997801112e-06, 'epoch': 2.75}) +[2026-06-18 02:01:58,035][starVLA.training.train_starvla][INFO] - Step 6060, Loss: {'train/loss': 0.13371968269348145, 'train/grad_norm_pre_clip': 20.90365219116211, 'timing/data': 0.0077179791405797005, 'timing/model': 0.9274016134440899, 'learning_rate/qwen_vl_interface': 9.284316520831825e-07, 'learning_rate/action_model': 9.284316520831825e-06, 'epoch': 2.75}) +[2026-06-18 02:02:01,819][starVLA.training.train_starvla][INFO] - Step 6061, Loss: {'train/loss': 0.06970375031232834, 'train/grad_norm_pre_clip': 21.682228088378906, 'timing/data': 0.008083835244178772, 'timing/model': 1.1164376977831125, 'learning_rate/qwen_vl_interface': 9.275345002520225e-07, 'learning_rate/action_model': 9.275345002520225e-06, 'epoch': 2.75}) +[2026-06-18 02:02:05,429][starVLA.training.train_starvla][INFO] - Step 6062, Loss: {'train/loss': 0.05288861691951752, 'train/grad_norm_pre_clip': 15.808521270751953, 'timing/data': 0.01101190410554409, 'timing/model': 0.9414281072095037, 'learning_rate/qwen_vl_interface': 9.266382444726117e-07, 'learning_rate/action_model': 9.266382444726117e-06, 'epoch': 2.75}) +[2026-06-18 02:02:09,219][starVLA.training.train_starvla][INFO] - Step 6063, Loss: {'train/loss': 0.04488431289792061, 'train/grad_norm_pre_clip': 14.764681816101074, 'timing/data': 0.010696991346776485, 'timing/model': 0.9308948870748281, 'learning_rate/qwen_vl_interface': 9.257428849307446e-07, 'learning_rate/action_model': 9.257428849307445e-06, 'epoch': 2.75}) +[2026-06-18 02:02:13,137][starVLA.training.train_starvla][INFO] - Step 6064, Loss: {'train/loss': 0.0806979089975357, 'train/grad_norm_pre_clip': 19.923181533813477, 'timing/data': 0.01232176274061203, 'timing/model': 1.030789758078754, 'learning_rate/qwen_vl_interface': 9.248484218120307e-07, 'learning_rate/action_model': 9.248484218120307e-06, 'epoch': 2.75}) +[2026-06-18 02:02:16,968][starVLA.training.train_starvla][INFO] - Step 6065, Loss: {'train/loss': 0.26876047253608704, 'train/grad_norm_pre_clip': 23.12993621826172, 'timing/data': 0.012006642296910286, 'timing/model': 0.9334252271801233, 'learning_rate/qwen_vl_interface': 9.239548553018925e-07, 'learning_rate/action_model': 9.239548553018924e-06, 'epoch': 2.75}) +[2026-06-18 02:02:20,851][starVLA.training.train_starvla][INFO] - Step 6066, Loss: {'train/loss': 0.04986398667097092, 'train/grad_norm_pre_clip': 13.19123363494873, 'timing/data': 0.01877561677247286, 'timing/model': 0.9330240385606885, 'learning_rate/qwen_vl_interface': 9.230621855855687e-07, 'learning_rate/action_model': 9.230621855855687e-06, 'epoch': 2.75}) +[2026-06-18 02:02:24,810][starVLA.training.train_starvla][INFO] - Step 6067, Loss: {'train/loss': 0.23146885633468628, 'train/grad_norm_pre_clip': 25.610578536987305, 'timing/data': 0.010607372038066387, 'timing/model': 1.151143291965127, 'learning_rate/qwen_vl_interface': 9.221704128481107e-07, 'learning_rate/action_model': 9.221704128481106e-06, 'epoch': 2.75}) +[2026-06-18 02:02:28,728][starVLA.training.train_starvla][INFO] - Step 6068, Loss: {'train/loss': 0.33696597814559937, 'train/grad_norm_pre_clip': 27.2220401763916, 'timing/data': 0.01115914061665535, 'timing/model': 1.180970192886889, 'learning_rate/qwen_vl_interface': 9.212795372743815e-07, 'learning_rate/action_model': 9.212795372743813e-06, 'epoch': 2.75}) +[2026-06-18 02:02:32,344][starVLA.training.train_starvla][INFO] - Step 6069, Loss: {'train/loss': 0.1005171462893486, 'train/grad_norm_pre_clip': 25.487619400024414, 'timing/data': 0.007639368996024132, 'timing/model': 0.9569871835410595, 'learning_rate/qwen_vl_interface': 9.203895590490617e-07, 'learning_rate/action_model': 9.203895590490616e-06, 'epoch': 2.75}) +[2026-06-18 02:02:36,383][starVLA.training.train_starvla][INFO] - Step 6070, Loss: {'train/loss': 0.10377923399209976, 'train/grad_norm_pre_clip': 23.318418502807617, 'timing/data': 0.00916147232055664, 'timing/model': 0.9428942296653986, 'learning_rate/qwen_vl_interface': 9.195004783566464e-07, 'learning_rate/action_model': 9.195004783566463e-06, 'epoch': 2.75}) +[2026-06-18 02:02:40,233][starVLA.training.train_starvla][INFO] - Step 6071, Loss: {'train/loss': 0.10200966894626617, 'train/grad_norm_pre_clip': 31.561168670654297, 'timing/data': 0.010694355703890324, 'timing/model': 0.9374203765764832, 'learning_rate/qwen_vl_interface': 9.186122953814414e-07, 'learning_rate/action_model': 9.186122953814413e-06, 'epoch': 2.75}) +[2026-06-18 02:02:44,428][starVLA.training.train_starvla][INFO] - Step 6072, Loss: {'train/loss': 0.31363019347190857, 'train/grad_norm_pre_clip': 18.814390182495117, 'timing/data': 0.00741917546838522, 'timing/model': 1.1535213980823755, 'learning_rate/qwen_vl_interface': 9.177250103075686e-07, 'learning_rate/action_model': 9.177250103075686e-06, 'epoch': 2.75}) +[2026-06-18 02:02:48,426][starVLA.training.train_starvla][INFO] - Step 6073, Loss: {'train/loss': 0.10957720130681992, 'train/grad_norm_pre_clip': 27.49687385559082, 'timing/data': 0.008600475266575813, 'timing/model': 0.9545145193114877, 'learning_rate/qwen_vl_interface': 9.168386233189627e-07, 'learning_rate/action_model': 9.168386233189627e-06, 'epoch': 2.76}) +[2026-06-18 02:02:54,212][starVLA.training.train_starvla][INFO] - Step 6074, Loss: {'train/loss': 0.14878727495670319, 'train/grad_norm_pre_clip': 30.654438018798828, 'timing/data': 0.018390418961644173, 'timing/model': 1.151054142974317, 'learning_rate/qwen_vl_interface': 9.159531345993726e-07, 'learning_rate/action_model': 9.159531345993725e-06, 'epoch': 2.76}) +[2026-06-18 02:02:58,230][starVLA.training.train_starvla][INFO] - Step 6075, Loss: {'train/loss': 0.0912834107875824, 'train/grad_norm_pre_clip': 19.335384368896484, 'timing/data': 0.0171424662694335, 'timing/model': 1.1939225234091282, 'learning_rate/qwen_vl_interface': 9.150685443323619e-07, 'learning_rate/action_model': 9.150685443323618e-06, 'epoch': 2.76}) +[2026-06-18 02:03:01,997][starVLA.training.train_starvla][INFO] - Step 6076, Loss: {'train/loss': 0.06357418745756149, 'train/grad_norm_pre_clip': 12.764261245727539, 'timing/data': 0.008924287743866444, 'timing/model': 0.9494896307587624, 'learning_rate/qwen_vl_interface': 9.141848527013059e-07, 'learning_rate/action_model': 9.141848527013059e-06, 'epoch': 2.76}) +[2026-06-18 02:03:06,037][starVLA.training.train_starvla][INFO] - Step 6077, Loss: {'train/loss': 0.24574105441570282, 'train/grad_norm_pre_clip': 29.773958206176758, 'timing/data': 0.009358835406601429, 'timing/model': 0.9700456662103534, 'learning_rate/qwen_vl_interface': 9.13302059889396e-07, 'learning_rate/action_model': 9.13302059889396e-06, 'epoch': 2.76}) +[2026-06-18 02:03:10,004][starVLA.training.train_starvla][INFO] - Step 6078, Loss: {'train/loss': 0.2759803235530853, 'train/grad_norm_pre_clip': 29.222387313842773, 'timing/data': 0.00809636153280735, 'timing/model': 1.0075612887740135, 'learning_rate/qwen_vl_interface': 9.12420166079635e-07, 'learning_rate/action_model': 9.12420166079635e-06, 'epoch': 2.76}) +[2026-06-18 02:03:13,834][starVLA.training.train_starvla][INFO] - Step 6079, Loss: {'train/loss': 0.3996680676937103, 'train/grad_norm_pre_clip': 31.160303115844727, 'timing/data': 0.00938478298485279, 'timing/model': 0.9453083127737045, 'learning_rate/qwen_vl_interface': 9.115391714548418e-07, 'learning_rate/action_model': 9.115391714548418e-06, 'epoch': 2.76}) +[2026-06-18 02:03:17,727][starVLA.training.train_starvla][INFO] - Step 6080, Loss: {'train/loss': 0.1271120309829712, 'train/grad_norm_pre_clip': 14.634178161621094, 'timing/data': 0.008969128131866455, 'timing/model': 1.0476478040218353, 'learning_rate/qwen_vl_interface': 9.106590761976454e-07, 'learning_rate/action_model': 9.106590761976455e-06, 'epoch': 2.76}) +[2026-06-18 02:03:21,636][starVLA.training.train_starvla][INFO] - Step 6081, Loss: {'train/loss': 0.03577528893947601, 'train/grad_norm_pre_clip': 8.58005428314209, 'timing/data': 0.008871063590049744, 'timing/model': 1.1493587903678417, 'learning_rate/qwen_vl_interface': 9.097798804904934e-07, 'learning_rate/action_model': 9.097798804904933e-06, 'epoch': 2.76}) +[2026-06-18 02:03:25,292][starVLA.training.train_starvla][INFO] - Step 6082, Loss: {'train/loss': 0.2912323772907257, 'train/grad_norm_pre_clip': 33.24380874633789, 'timing/data': 0.008612660691142082, 'timing/model': 0.9409938985481858, 'learning_rate/qwen_vl_interface': 9.089015845156422e-07, 'learning_rate/action_model': 9.089015845156422e-06, 'epoch': 2.76}) +[2026-06-18 02:03:29,400][starVLA.training.train_starvla][INFO] - Step 6083, Loss: {'train/loss': 0.2915969491004944, 'train/grad_norm_pre_clip': 45.21598815917969, 'timing/data': 0.008612748235464096, 'timing/model': 0.9529099818319082, 'learning_rate/qwen_vl_interface': 9.080241884551627e-07, 'learning_rate/action_model': 9.080241884551627e-06, 'epoch': 2.76}) +[2026-06-18 02:03:33,107][starVLA.training.train_starvla][INFO] - Step 6084, Loss: {'train/loss': 0.17866799235343933, 'train/grad_norm_pre_clip': 30.361387252807617, 'timing/data': 0.014051136560738087, 'timing/model': 0.9225219022482634, 'learning_rate/qwen_vl_interface': 9.071476924909417e-07, 'learning_rate/action_model': 9.071476924909416e-06, 'epoch': 2.76}) +[2026-06-18 02:03:37,007][starVLA.training.train_starvla][INFO] - Step 6085, Loss: {'train/loss': 0.04848852381110191, 'train/grad_norm_pre_clip': 16.262617111206055, 'timing/data': 0.01159616932272911, 'timing/model': 0.9489137511700392, 'learning_rate/qwen_vl_interface': 9.062720968046773e-07, 'learning_rate/action_model': 9.062720968046773e-06, 'epoch': 2.76}) +[2026-06-18 02:03:40,985][starVLA.training.train_starvla][INFO] - Step 6086, Loss: {'train/loss': 0.4687225818634033, 'train/grad_norm_pre_clip': 25.364770889282227, 'timing/data': 0.008369413204491138, 'timing/model': 0.9759178925305605, 'learning_rate/qwen_vl_interface': 9.053974015778817e-07, 'learning_rate/action_model': 9.053974015778817e-06, 'epoch': 2.76}) +[2026-06-18 02:03:44,791][starVLA.training.train_starvla][INFO] - Step 6087, Loss: {'train/loss': 0.1189422756433487, 'train/grad_norm_pre_clip': 35.43191146850586, 'timing/data': 0.009726676158607006, 'timing/model': 1.1114900009706616, 'learning_rate/qwen_vl_interface': 9.045236069918788e-07, 'learning_rate/action_model': 9.045236069918788e-06, 'epoch': 2.76}) +[2026-06-18 02:03:48,729][starVLA.training.train_starvla][INFO] - Step 6088, Loss: {'train/loss': 0.27562758326530457, 'train/grad_norm_pre_clip': 26.248443603515625, 'timing/data': 0.006935896351933479, 'timing/model': 1.2418670356273651, 'learning_rate/qwen_vl_interface': 9.036507132278085e-07, 'learning_rate/action_model': 9.036507132278086e-06, 'epoch': 2.76}) +[2026-06-18 02:03:52,516][starVLA.training.train_starvla][INFO] - Step 6089, Loss: {'train/loss': 0.20463363826274872, 'train/grad_norm_pre_clip': 13.973275184631348, 'timing/data': 0.00858138594776392, 'timing/model': 1.0377532942220569, 'learning_rate/qwen_vl_interface': 9.02778720466622e-07, 'learning_rate/action_model': 9.02778720466622e-06, 'epoch': 2.76}) +[2026-06-18 02:03:56,501][starVLA.training.train_starvla][INFO] - Step 6090, Loss: {'train/loss': 0.18240030109882355, 'train/grad_norm_pre_clip': 38.691524505615234, 'timing/data': 0.018376905471086502, 'timing/model': 1.0358917089179158, 'learning_rate/qwen_vl_interface': 9.019076288890852e-07, 'learning_rate/action_model': 9.019076288890852e-06, 'epoch': 2.76}) +[2026-06-18 02:04:02,404][starVLA.training.train_starvla][INFO] - Step 6091, Loss: {'train/loss': 0.24206335842609406, 'train/grad_norm_pre_clip': 38.74472427368164, 'timing/data': 0.009749067015945911, 'timing/model': 0.9441875955089927, 'learning_rate/qwen_vl_interface': 9.010374386757755e-07, 'learning_rate/action_model': 9.010374386757754e-06, 'epoch': 2.76}) +[2026-06-18 02:04:06,241][starVLA.training.train_starvla][INFO] - Step 6092, Loss: {'train/loss': 0.0355108305811882, 'train/grad_norm_pre_clip': 11.556973457336426, 'timing/data': 0.011779987253248692, 'timing/model': 0.9298966247588396, 'learning_rate/qwen_vl_interface': 9.001681500070842e-07, 'learning_rate/action_model': 9.001681500070842e-06, 'epoch': 2.76}) +[2026-06-18 02:04:10,734][starVLA.training.train_starvla][INFO] - Step 6093, Loss: {'train/loss': 0.07648532837629318, 'train/grad_norm_pre_clip': 21.597492218017578, 'timing/data': 0.00937497615814209, 'timing/model': 1.6004920024424791, 'learning_rate/qwen_vl_interface': 8.992997630632161e-07, 'learning_rate/action_model': 8.99299763063216e-06, 'epoch': 2.76}) +[2026-06-18 02:04:14,648][starVLA.training.train_starvla][INFO] - Step 6094, Loss: {'train/loss': 0.102815181016922, 'train/grad_norm_pre_clip': 18.24323272705078, 'timing/data': 0.010301263071596622, 'timing/model': 1.176398642361164, 'learning_rate/qwen_vl_interface': 8.984322780241879e-07, 'learning_rate/action_model': 8.984322780241878e-06, 'epoch': 2.76}) +[2026-06-18 02:04:18,520][starVLA.training.train_starvla][INFO] - Step 6095, Loss: {'train/loss': 0.022546928375959396, 'train/grad_norm_pre_clip': 5.630387306213379, 'timing/data': 0.009381632320582867, 'timing/model': 1.1461278619244695, 'learning_rate/qwen_vl_interface': 8.975656950698325e-07, 'learning_rate/action_model': 8.975656950698324e-06, 'epoch': 2.77}) +[2026-06-18 02:04:22,132][starVLA.training.train_starvla][INFO] - Step 6096, Loss: {'train/loss': 0.12557092308998108, 'train/grad_norm_pre_clip': 24.825809478759766, 'timing/data': 0.007641781121492386, 'timing/model': 0.9394620815292001, 'learning_rate/qwen_vl_interface': 8.967000143797913e-07, 'learning_rate/action_model': 8.967000143797913e-06, 'epoch': 2.77}) +[2026-06-18 02:04:25,934][starVLA.training.train_starvla][INFO] - Step 6097, Loss: {'train/loss': 0.23061509430408478, 'train/grad_norm_pre_clip': 46.19504928588867, 'timing/data': 0.008366039954125881, 'timing/model': 0.9389119977131486, 'learning_rate/qwen_vl_interface': 8.958352361335214e-07, 'learning_rate/action_model': 8.958352361335213e-06, 'epoch': 2.77}) +[2026-06-18 02:04:29,777][starVLA.training.train_starvla][INFO] - Step 6098, Loss: {'train/loss': 0.11281365901231766, 'train/grad_norm_pre_clip': 32.222938537597656, 'timing/data': 0.020826062187552452, 'timing/model': 0.9308766620233655, 'learning_rate/qwen_vl_interface': 8.949713605102917e-07, 'learning_rate/action_model': 8.949713605102916e-06, 'epoch': 2.77}) +[2026-06-18 02:04:33,694][starVLA.training.train_starvla][INFO] - Step 6099, Loss: {'train/loss': 0.09322728216648102, 'train/grad_norm_pre_clip': 37.094120025634766, 'timing/data': 0.00048544909805059433, 'timing/model': 0.9349293792620301, 'learning_rate/qwen_vl_interface': 8.941083876891857e-07, 'learning_rate/action_model': 8.941083876891856e-06, 'epoch': 2.77}) +[2026-06-18 02:06:47,174][starVLA.training.train_starvla][INFO] - Step 6100, Loss: {'train/loss': 0.24444063007831573, 'train/grad_norm_pre_clip': 34.77846908569336, 'eval/loss': 0.25648003816604614, 'eval/latency_8/loss': 0.25648003816604614, 'eval/demon_fire_precision': 0.9720430107526882, 'eval/demon_fire_recall': 0.9829139484311898, 'eval/demon_fire_f1': 0.9774482545566884, 'eval/demon_fire_rate_model': 0.9056761268781303, 'eval/demon_fire_rate_teacher': 0.8956594323873122, 'eval/demon_left_precision': 0.9108910891089109, 'eval/demon_left_recall': 0.7880085653104925, 'eval/demon_left_f1': 0.8450057405281285, 'eval/demon_left_rate_model': 0.112409571508069, 'eval/demon_left_rate_teacher': 0.12993878686700056, 'eval/demon_right_precision': 0.88125, 'eval/demon_right_recall': 0.8545454545454545, 'eval/demon_right_f1': 0.8676923076923077, 'eval/demon_right_rate_model': 0.17807456872565386, 'eval/demon_right_rate_teacher': 0.18363939899833054, 'eval/demon_move_f1': 0.8563490241102181, 'eval/demon_cc_f1': 0.9376685827450993, 'eval/latency_8/demon_fire_precision': 0.9720430107526882, 'eval/latency_8/demon_fire_recall': 0.9829139484311898, 'eval/latency_8/demon_fire_f1': 0.9774482545566884, 'eval/latency_8/demon_fire_rate_model': 0.9056761268781303, 'eval/latency_8/demon_fire_rate_teacher': 0.8956594323873122, 'eval/latency_8/demon_left_precision': 0.9108910891089109, 'eval/latency_8/demon_left_recall': 0.7880085653104925, 'eval/latency_8/demon_left_f1': 0.8450057405281285, 'eval/latency_8/demon_left_rate_model': 0.112409571508069, 'eval/latency_8/demon_left_rate_teacher': 0.12993878686700056, 'eval/latency_8/demon_right_precision': 0.88125, 'eval/latency_8/demon_right_recall': 0.8545454545454545, 'eval/latency_8/demon_right_f1': 0.8676923076923077, 'eval/latency_8/demon_right_rate_model': 0.17807456872565386, 'eval/latency_8/demon_right_rate_teacher': 0.18363939899833054, 'eval/latency_8/demon_move_f1': 0.8563490241102181, 'eval/latency_8/demon_cc_f1': 0.9376685827450993, 'timing/data': 0.0012160837650299072, 'timing/model': 0.9885604232549667, 'learning_rate/qwen_vl_interface': 8.932463178490991e-07, 'learning_rate/action_model': 8.93246317849099e-06, 'epoch': 2.77}) +[2026-06-18 02:06:47,175][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_6100_state +[2026-06-18 02:06:47,181][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 02:06:53,717][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_6100_state/model.safetensors +[2026-06-18 02:07:08,986][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_6100_state/optimizer.bin +[2026-06-18 02:07:08,987][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_6100_state/sampler.bin +[2026-06-18 02:07:08,987][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_6100_state/sampler_1.bin +[2026-06-18 02:07:08,990][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_6100_state/random_states_0.pkl +[2026-06-18 02:07:08,993][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 6100 +[2026-06-18 02:07:08,993][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 02:07:09,004][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 02:07:12,766][starVLA.training.train_starvla][INFO] - Step 6101, Loss: {'train/loss': 0.2767009437084198, 'train/grad_norm_pre_clip': 25.171663284301758, 'timing/data': 0.0007229587063193321, 'timing/model': 0.9173004189506173, 'learning_rate/qwen_vl_interface': 8.923851511687379e-07, 'learning_rate/action_model': 8.923851511687378e-06, 'epoch': 2.77}) +[2026-06-18 02:07:16,501][starVLA.training.train_starvla][INFO] - Step 6102, Loss: {'train/loss': 0.13476164638996124, 'train/grad_norm_pre_clip': 37.53266143798828, 'timing/data': 0.001129993237555027, 'timing/model': 0.9191537545993924, 'learning_rate/qwen_vl_interface': 8.915248878266228e-07, 'learning_rate/action_model': 8.915248878266227e-06, 'epoch': 2.77}) +[2026-06-18 02:07:20,202][starVLA.training.train_starvla][INFO] - Step 6103, Loss: {'train/loss': 0.05926615372300148, 'train/grad_norm_pre_clip': 13.583542823791504, 'timing/data': 0.0005546500906348228, 'timing/model': 1.076397666707635, 'learning_rate/qwen_vl_interface': 8.906655280010895e-07, 'learning_rate/action_model': 8.906655280010896e-06, 'epoch': 2.77}) +[2026-06-18 02:07:23,953][starVLA.training.train_starvla][INFO] - Step 6104, Loss: {'train/loss': 0.09945353120565414, 'train/grad_norm_pre_clip': 15.051993370056152, 'timing/data': 0.0006799129769206047, 'timing/model': 1.104379347525537, 'learning_rate/qwen_vl_interface': 8.898070718702831e-07, 'learning_rate/action_model': 8.898070718702831e-06, 'epoch': 2.77}) +[2026-06-18 02:07:27,730][starVLA.training.train_starvla][INFO] - Step 6105, Loss: {'train/loss': 0.08614613860845566, 'train/grad_norm_pre_clip': 18.190990447998047, 'timing/data': 0.0011067576706409454, 'timing/model': 1.144007888622582, 'learning_rate/qwen_vl_interface': 8.889495196121628e-07, 'learning_rate/action_model': 8.889495196121627e-06, 'epoch': 2.77}) +[2026-06-18 02:07:31,465][starVLA.training.train_starvla][INFO] - Step 6106, Loss: {'train/loss': 0.08068209141492844, 'train/grad_norm_pre_clip': 14.947141647338867, 'timing/data': 0.00048738159239292145, 'timing/model': 0.922499873675406, 'learning_rate/qwen_vl_interface': 8.880928714044994e-07, 'learning_rate/action_model': 8.880928714044993e-06, 'epoch': 2.77}) +[2026-06-18 02:07:35,266][starVLA.training.train_starvla][INFO] - Step 6107, Loss: {'train/loss': 0.1362036019563675, 'train/grad_norm_pre_clip': 16.105146408081055, 'timing/data': 0.000985647551715374, 'timing/model': 1.0004773596301675, 'learning_rate/qwen_vl_interface': 8.872371274248773e-07, 'learning_rate/action_model': 8.872371274248772e-06, 'epoch': 2.77}) +[2026-06-18 02:07:38,976][starVLA.training.train_starvla][INFO] - Step 6108, Loss: {'train/loss': 0.014399169944226742, 'train/grad_norm_pre_clip': 7.952975273132324, 'timing/data': 0.0005633393302559853, 'timing/model': 0.9333516331389546, 'learning_rate/qwen_vl_interface': 8.863822878506931e-07, 'learning_rate/action_model': 8.86382287850693e-06, 'epoch': 2.77}) +[2026-06-18 02:07:42,754][starVLA.training.train_starvla][INFO] - Step 6109, Loss: {'train/loss': 0.1640356183052063, 'train/grad_norm_pre_clip': 20.889284133911133, 'timing/data': 0.0005012145265936852, 'timing/model': 0.9182236604392529, 'learning_rate/qwen_vl_interface': 8.855283528591566e-07, 'learning_rate/action_model': 8.855283528591565e-06, 'epoch': 2.77}) +[2026-06-18 02:07:47,051][starVLA.training.train_starvla][INFO] - Step 6110, Loss: {'train/loss': 0.13659760355949402, 'train/grad_norm_pre_clip': 29.689041137695312, 'timing/data': 0.0006935512647032738, 'timing/model': 1.0992730800062418, 'learning_rate/qwen_vl_interface': 8.846753226272884e-07, 'learning_rate/action_model': 8.846753226272883e-06, 'epoch': 2.77}) +[2026-06-18 02:07:50,608][starVLA.training.train_starvla][INFO] - Step 6111, Loss: {'train/loss': 0.11140355467796326, 'train/grad_norm_pre_clip': 23.627817153930664, 'timing/data': 0.0005328981205821037, 'timing/model': 0.9184908708557487, 'learning_rate/qwen_vl_interface': 8.838231973319235e-07, 'learning_rate/action_model': 8.838231973319235e-06, 'epoch': 2.77}) +[2026-06-18 02:07:54,381][starVLA.training.train_starvla][INFO] - Step 6112, Loss: {'train/loss': 0.09094254672527313, 'train/grad_norm_pre_clip': 25.47788429260254, 'timing/data': 0.0009730877354741096, 'timing/model': 0.9196596760302782, 'learning_rate/qwen_vl_interface': 8.829719771497072e-07, 'learning_rate/action_model': 8.829719771497071e-06, 'epoch': 2.77}) +[2026-06-18 02:07:58,434][starVLA.training.train_starvla][INFO] - Step 6113, Loss: {'train/loss': 0.20759141445159912, 'train/grad_norm_pre_clip': 21.664199829101562, 'timing/data': 0.0004933606833219528, 'timing/model': 0.9213320445269346, 'learning_rate/qwen_vl_interface': 8.821216622570999e-07, 'learning_rate/action_model': 8.821216622570998e-06, 'epoch': 2.77}) +[2026-06-18 02:08:02,155][starVLA.training.train_starvla][INFO] - Step 6114, Loss: {'train/loss': 0.2233479619026184, 'train/grad_norm_pre_clip': 24.726795196533203, 'timing/data': 0.0005187075585126877, 'timing/model': 0.9337716409936547, 'learning_rate/qwen_vl_interface': 8.812722528303727e-07, 'learning_rate/action_model': 8.812722528303727e-06, 'epoch': 2.77}) +[2026-06-18 02:08:05,889][starVLA.training.train_starvla][INFO] - Step 6115, Loss: {'train/loss': 0.18689602613449097, 'train/grad_norm_pre_clip': 26.78025245666504, 'timing/data': 0.0005452446639537811, 'timing/model': 0.9282958572730422, 'learning_rate/qwen_vl_interface': 8.804237490456076e-07, 'learning_rate/action_model': 8.804237490456075e-06, 'epoch': 2.77}) +[2026-06-18 02:08:09,660][starVLA.training.train_starvla][INFO] - Step 6116, Loss: {'train/loss': 0.10177537053823471, 'train/grad_norm_pre_clip': 18.02050018310547, 'timing/data': 0.0004900144413113594, 'timing/model': 1.1234417660161853, 'learning_rate/qwen_vl_interface': 8.795761510787007e-07, 'learning_rate/action_model': 8.795761510787006e-06, 'epoch': 2.77}) +[2026-06-18 02:08:13,564][starVLA.training.train_starvla][INFO] - Step 6117, Loss: {'train/loss': 0.08676119148731232, 'train/grad_norm_pre_clip': 15.629148483276367, 'timing/data': 0.0009001987054944038, 'timing/model': 1.2514750938862562, 'learning_rate/qwen_vl_interface': 8.78729459105361e-07, 'learning_rate/action_model': 8.78729459105361e-06, 'epoch': 2.78}) +[2026-06-18 02:08:17,139][starVLA.training.train_starvla][INFO] - Step 6118, Loss: {'train/loss': 0.20612800121307373, 'train/grad_norm_pre_clip': 27.233848571777344, 'timing/data': 0.0010630898177623749, 'timing/model': 0.9192060837522149, 'learning_rate/qwen_vl_interface': 8.778836733011079e-07, 'learning_rate/action_model': 8.778836733011079e-06, 'epoch': 2.78}) +[2026-06-18 02:08:21,058][starVLA.training.train_starvla][INFO] - Step 6119, Loss: {'train/loss': 0.39629438519477844, 'train/grad_norm_pre_clip': 39.80220031738281, 'timing/data': 0.0004910901188850403, 'timing/model': 0.9201841354370117, 'learning_rate/qwen_vl_interface': 8.770387938412753e-07, 'learning_rate/action_model': 8.770387938412753e-06, 'epoch': 2.78}) +[2026-06-18 02:08:24,795][starVLA.training.train_starvla][INFO] - Step 6120, Loss: {'train/loss': 0.17251980304718018, 'train/grad_norm_pre_clip': 32.72291564941406, 'timing/data': 0.00050392746925354, 'timing/model': 0.9389526704326272, 'learning_rate/qwen_vl_interface': 8.76194820901004e-07, 'learning_rate/action_model': 8.76194820901004e-06, 'epoch': 2.78}) +[2026-06-18 02:08:28,803][starVLA.training.train_starvla][INFO] - Step 6121, Loss: {'train/loss': 0.3447522819042206, 'train/grad_norm_pre_clip': 42.27912902832031, 'timing/data': 0.000543360598385334, 'timing/model': 0.9242742517963052, 'learning_rate/qwen_vl_interface': 8.753517546552545e-07, 'learning_rate/action_model': 8.753517546552544e-06, 'epoch': 2.78}) +[2026-06-18 02:08:32,607][starVLA.training.train_starvla][INFO] - Step 6122, Loss: {'train/loss': 0.05379701033234596, 'train/grad_norm_pre_clip': 16.205047607421875, 'timing/data': 0.0004576807841658592, 'timing/model': 0.9306357046589255, 'learning_rate/qwen_vl_interface': 8.74509595278793e-07, 'learning_rate/action_model': 8.74509595278793e-06, 'epoch': 2.78}) +[2026-06-18 02:08:36,331][starVLA.training.train_starvla][INFO] - Step 6123, Loss: {'train/loss': 0.5291755199432373, 'train/grad_norm_pre_clip': 25.0185546875, 'timing/data': 0.0008042454719543457, 'timing/model': 1.0883240913972259, 'learning_rate/qwen_vl_interface': 8.736683429462003e-07, 'learning_rate/action_model': 8.736683429462004e-06, 'epoch': 2.78}) +[2026-06-18 02:08:39,884][starVLA.training.train_starvla][INFO] - Step 6124, Loss: {'train/loss': 0.2236504703760147, 'train/grad_norm_pre_clip': 35.22398376464844, 'timing/data': 0.00047990400344133377, 'timing/model': 0.9098206348717213, 'learning_rate/qwen_vl_interface': 8.728279978318696e-07, 'learning_rate/action_model': 8.728279978318695e-06, 'epoch': 2.78}) +[2026-06-18 02:08:43,590][starVLA.training.train_starvla][INFO] - Step 6125, Loss: {'train/loss': 0.3367210328578949, 'train/grad_norm_pre_clip': 28.339651107788086, 'timing/data': 0.0007286854088306427, 'timing/model': 0.9114311756566167, 'learning_rate/qwen_vl_interface': 8.719885601100052e-07, 'learning_rate/action_model': 8.719885601100051e-06, 'epoch': 2.78}) +[2026-06-18 02:08:47,347][starVLA.training.train_starvla][INFO] - Step 6126, Loss: {'train/loss': 0.03805459290742874, 'train/grad_norm_pre_clip': 10.25921630859375, 'timing/data': 0.0008977875113487244, 'timing/model': 0.9208109248429537, 'learning_rate/qwen_vl_interface': 8.71150029954623e-07, 'learning_rate/action_model': 8.71150029954623e-06, 'epoch': 2.78}) +[2026-06-18 02:08:51,064][starVLA.training.train_starvla][INFO] - Step 6127, Loss: {'train/loss': 0.03046482801437378, 'train/grad_norm_pre_clip': 7.604512691497803, 'timing/data': 0.0009622946381568909, 'timing/model': 0.9106770474463701, 'learning_rate/qwen_vl_interface': 8.703124075395504e-07, 'learning_rate/action_model': 8.703124075395504e-06, 'epoch': 2.78}) +[2026-06-18 02:08:54,772][starVLA.training.train_starvla][INFO] - Step 6128, Loss: {'train/loss': 0.23289994895458221, 'train/grad_norm_pre_clip': 27.137401580810547, 'timing/data': 0.0011238064616918564, 'timing/model': 0.9150595832616091, 'learning_rate/qwen_vl_interface': 8.694756930384305e-07, 'learning_rate/action_model': 8.694756930384306e-06, 'epoch': 2.78}) +[2026-06-18 02:08:58,470][starVLA.training.train_starvla][INFO] - Step 6129, Loss: {'train/loss': 0.2951030135154724, 'train/grad_norm_pre_clip': 38.287376403808594, 'timing/data': 0.0006883740425109863, 'timing/model': 1.0723752789199352, 'learning_rate/qwen_vl_interface': 8.686398866247121e-07, 'learning_rate/action_model': 8.68639886624712e-06, 'epoch': 2.78}) +[2026-06-18 02:09:02,366][starVLA.training.train_starvla][INFO] - Step 6130, Loss: {'train/loss': 0.16917628049850464, 'train/grad_norm_pre_clip': 22.28139877319336, 'timing/data': 0.000523286871612072, 'timing/model': 1.1143962405622005, 'learning_rate/qwen_vl_interface': 8.6780498847166e-07, 'learning_rate/action_model': 8.678049884716599e-06, 'epoch': 2.78}) +[2026-06-18 02:09:05,969][starVLA.training.train_starvla][INFO] - Step 6131, Loss: {'train/loss': 0.07386169582605362, 'train/grad_norm_pre_clip': 27.14447784423828, 'timing/data': 0.0009388141334056854, 'timing/model': 0.9264629250392318, 'learning_rate/qwen_vl_interface': 8.669709987523479e-07, 'learning_rate/action_model': 8.669709987523477e-06, 'epoch': 2.78}) +[2026-06-18 02:09:09,742][starVLA.training.train_starvla][INFO] - Step 6132, Loss: {'train/loss': 0.10699649900197983, 'train/grad_norm_pre_clip': 26.21440887451172, 'timing/data': 0.0004954561591148376, 'timing/model': 0.9213788751512766, 'learning_rate/qwen_vl_interface': 8.661379176396651e-07, 'learning_rate/action_model': 8.66137917639665e-06, 'epoch': 2.78}) +[2026-06-18 02:09:13,497][starVLA.training.train_starvla][INFO] - Step 6133, Loss: {'train/loss': 0.03603607043623924, 'train/grad_norm_pre_clip': 10.5377779006958, 'timing/data': 0.0005055274814367294, 'timing/model': 0.9248823309317231, 'learning_rate/qwen_vl_interface': 8.653057453063092e-07, 'learning_rate/action_model': 8.653057453063091e-06, 'epoch': 2.78}) +[2026-06-18 02:09:17,246][starVLA.training.train_starvla][INFO] - Step 6134, Loss: {'train/loss': 0.36651143431663513, 'train/grad_norm_pre_clip': 29.280010223388672, 'timing/data': 0.000496196560561657, 'timing/model': 0.9183981940150261, 'learning_rate/qwen_vl_interface': 8.644744819247908e-07, 'learning_rate/action_model': 8.644744819247908e-06, 'epoch': 2.78}) +[2026-06-18 02:09:20,963][starVLA.training.train_starvla][INFO] - Step 6135, Loss: {'train/loss': 0.030507436022162437, 'train/grad_norm_pre_clip': 9.231806755065918, 'timing/data': 0.001459469087421894, 'timing/model': 0.9215562157332897, 'learning_rate/qwen_vl_interface': 8.636441276674308e-07, 'learning_rate/action_model': 8.636441276674308e-06, 'epoch': 2.78}) +[2026-06-18 02:09:24,703][starVLA.training.train_starvla][INFO] - Step 6136, Loss: {'train/loss': 0.20920580625534058, 'train/grad_norm_pre_clip': 31.494524002075195, 'timing/data': 0.0004842756316065788, 'timing/model': 1.0917016817256808, 'learning_rate/qwen_vl_interface': 8.628146827063632e-07, 'learning_rate/action_model': 8.628146827063633e-06, 'epoch': 2.78}) +[2026-06-18 02:09:28,458][starVLA.training.train_starvla][INFO] - Step 6137, Loss: {'train/loss': 0.13586577773094177, 'train/grad_norm_pre_clip': 22.46726417541504, 'timing/data': 0.0005144663155078888, 'timing/model': 1.1254405388608575, 'learning_rate/qwen_vl_interface': 8.619861472135332e-07, 'learning_rate/action_model': 8.619861472135332e-06, 'epoch': 2.78}) +[2026-06-18 02:09:32,121][starVLA.training.train_starvla][INFO] - Step 6138, Loss: {'train/loss': 0.1381656974554062, 'train/grad_norm_pre_clip': 25.37919044494629, 'timing/data': 0.0005195271223783493, 'timing/model': 0.9234030414372683, 'learning_rate/qwen_vl_interface': 8.611585213606964e-07, 'learning_rate/action_model': 8.611585213606964e-06, 'epoch': 2.78}) +[2026-06-18 02:09:35,987][starVLA.training.train_starvla][INFO] - Step 6139, Loss: {'train/loss': 0.5884754657745361, 'train/grad_norm_pre_clip': 69.71334075927734, 'timing/data': 0.000522940419614315, 'timing/model': 0.9362865351140499, 'learning_rate/qwen_vl_interface': 8.603318053194206e-07, 'learning_rate/action_model': 8.603318053194205e-06, 'epoch': 2.79}) +[2026-06-18 02:09:39,675][starVLA.training.train_starvla][INFO] - Step 6140, Loss: {'train/loss': 0.14006605744361877, 'train/grad_norm_pre_clip': 15.334087371826172, 'timing/data': 0.0004888996481895447, 'timing/model': 0.921345959417522, 'learning_rate/qwen_vl_interface': 8.595059992610846e-07, 'learning_rate/action_model': 8.595059992610846e-06, 'epoch': 2.79}) +[2026-06-18 02:09:43,368][starVLA.training.train_starvla][INFO] - Step 6141, Loss: {'train/loss': 0.5116445422172546, 'train/grad_norm_pre_clip': 43.30113220214844, 'timing/data': 0.0004792753607034683, 'timing/model': 0.914581710472703, 'learning_rate/qwen_vl_interface': 8.586811033568796e-07, 'learning_rate/action_model': 8.586811033568797e-06, 'epoch': 2.79}) +[2026-06-18 02:09:47,095][starVLA.training.train_starvla][INFO] - Step 6142, Loss: {'train/loss': 0.30230286717414856, 'train/grad_norm_pre_clip': 44.211273193359375, 'timing/data': 0.001133485697209835, 'timing/model': 0.9185744747519493, 'learning_rate/qwen_vl_interface': 8.578571177778076e-07, 'learning_rate/action_model': 8.578571177778075e-06, 'epoch': 2.79}) +[2026-06-18 02:09:50,841][starVLA.training.train_starvla][INFO] - Step 6143, Loss: {'train/loss': 0.20707006752490997, 'train/grad_norm_pre_clip': 30.44686508178711, 'timing/data': 0.0004901643842458725, 'timing/model': 1.1028572348877788, 'learning_rate/qwen_vl_interface': 8.570340426946805e-07, 'learning_rate/action_model': 8.570340426946804e-06, 'epoch': 2.79}) +[2026-06-18 02:09:54,556][starVLA.training.train_starvla][INFO] - Step 6144, Loss: {'train/loss': 0.3337973356246948, 'train/grad_norm_pre_clip': 24.7420597076416, 'timing/data': 0.0004978450015187263, 'timing/model': 1.0815196298062801, 'learning_rate/qwen_vl_interface': 8.562118782781241e-07, 'learning_rate/action_model': 8.56211878278124e-06, 'epoch': 2.79}) +[2026-06-18 02:09:58,374][starVLA.training.train_starvla][INFO] - Step 6145, Loss: {'train/loss': 0.041232965886592865, 'train/grad_norm_pre_clip': 9.407835960388184, 'timing/data': 0.0005097612738609314, 'timing/model': 0.9178385799750686, 'learning_rate/qwen_vl_interface': 8.553906246985722e-07, 'learning_rate/action_model': 8.553906246985721e-06, 'epoch': 2.79}) +[2026-06-18 02:10:02,120][starVLA.training.train_starvla][INFO] - Step 6146, Loss: {'train/loss': 0.3118968605995178, 'train/grad_norm_pre_clip': 41.55756378173828, 'timing/data': 0.0005000326782464981, 'timing/model': 0.9466079147532582, 'learning_rate/qwen_vl_interface': 8.545702821262733e-07, 'learning_rate/action_model': 8.545702821262733e-06, 'epoch': 2.79}) +[2026-06-18 02:10:05,932][starVLA.training.train_starvla][INFO] - Step 6147, Loss: {'train/loss': 0.1787135899066925, 'train/grad_norm_pre_clip': 21.96809196472168, 'timing/data': 0.0004772748798131943, 'timing/model': 0.932143421843648, 'learning_rate/qwen_vl_interface': 8.537508507312844e-07, 'learning_rate/action_model': 8.537508507312844e-06, 'epoch': 2.79}) +[2026-06-18 02:10:09,605][starVLA.training.train_starvla][INFO] - Step 6148, Loss: {'train/loss': 0.3998488783836365, 'train/grad_norm_pre_clip': 29.018081665039062, 'timing/data': 0.00048755668103694916, 'timing/model': 0.9177645398303866, 'learning_rate/qwen_vl_interface': 8.52932330683476e-07, 'learning_rate/action_model': 8.529323306834759e-06, 'epoch': 2.79}) +[2026-06-18 02:10:13,319][starVLA.training.train_starvla][INFO] - Step 6149, Loss: {'train/loss': 0.13913792371749878, 'train/grad_norm_pre_clip': 22.077505111694336, 'timing/data': 0.000492253340780735, 'timing/model': 0.9219012232497334, 'learning_rate/qwen_vl_interface': 8.52114722152525e-07, 'learning_rate/action_model': 8.521147221525249e-06, 'epoch': 2.79}) +[2026-06-18 02:10:17,144][starVLA.training.train_starvla][INFO] - Step 6150, Loss: {'train/loss': 0.13312819600105286, 'train/grad_norm_pre_clip': 32.38688278198242, 'timing/data': 0.0008966168388724327, 'timing/model': 1.1353001240640879, 'learning_rate/qwen_vl_interface': 8.51298025307925e-07, 'learning_rate/action_model': 8.51298025307925e-06, 'epoch': 2.79}) +[2026-06-18 02:10:20,870][starVLA.training.train_starvla][INFO] - Step 6151, Loss: {'train/loss': 0.33801785111427307, 'train/grad_norm_pre_clip': 51.338497161865234, 'timing/data': 0.00043459702283143997, 'timing/model': 1.0921024680137634, 'learning_rate/qwen_vl_interface': 8.504822403189783e-07, 'learning_rate/action_model': 8.504822403189782e-06, 'epoch': 2.79}) +[2026-06-18 02:10:24,789][starVLA.training.train_starvla][INFO] - Step 6152, Loss: {'train/loss': 0.020664682611823082, 'train/grad_norm_pre_clip': 5.82256555557251, 'timing/data': 0.0004705730825662613, 'timing/model': 1.1079437425360084, 'learning_rate/qwen_vl_interface': 8.496673673547967e-07, 'learning_rate/action_model': 8.496673673547968e-06, 'epoch': 2.79}) +[2026-06-18 02:10:28,691][starVLA.training.train_starvla][INFO] - Step 6153, Loss: {'train/loss': 0.15991821885108948, 'train/grad_norm_pre_clip': 15.572248458862305, 'timing/data': 0.0005346294492483139, 'timing/model': 0.9140796093270183, 'learning_rate/qwen_vl_interface': 8.488534065843046e-07, 'learning_rate/action_model': 8.488534065843045e-06, 'epoch': 2.79}) +[2026-06-18 02:10:32,399][starVLA.training.train_starvla][INFO] - Step 6154, Loss: {'train/loss': 0.17047569155693054, 'train/grad_norm_pre_clip': 39.82986068725586, 'timing/data': 0.0004738643765449524, 'timing/model': 0.9163556937128305, 'learning_rate/qwen_vl_interface': 8.480403581762378e-07, 'learning_rate/action_model': 8.480403581762378e-06, 'epoch': 2.79}) +[2026-06-18 02:10:36,123][starVLA.training.train_starvla][INFO] - Step 6155, Loss: {'train/loss': 0.301227331161499, 'train/grad_norm_pre_clip': 38.210243225097656, 'timing/data': 0.0005035791546106339, 'timing/model': 0.9195839799940586, 'learning_rate/qwen_vl_interface': 8.472282222991411e-07, 'learning_rate/action_model': 8.47228222299141e-06, 'epoch': 2.79}) +[2026-06-18 02:10:39,887][starVLA.training.train_starvla][INFO] - Step 6156, Loss: {'train/loss': 0.22701294720172882, 'train/grad_norm_pre_clip': 32.36093521118164, 'timing/data': 0.0008151708170771599, 'timing/model': 1.0856867637485266, 'learning_rate/qwen_vl_interface': 8.464169991213705e-07, 'learning_rate/action_model': 8.464169991213706e-06, 'epoch': 2.79}) +[2026-06-18 02:10:43,651][starVLA.training.train_starvla][INFO] - Step 6157, Loss: {'train/loss': 0.07177220284938812, 'train/grad_norm_pre_clip': 15.030634880065918, 'timing/data': 0.0004938011988997459, 'timing/model': 1.0319714862853289, 'learning_rate/qwen_vl_interface': 8.456066888110964e-07, 'learning_rate/action_model': 8.456066888110964e-06, 'epoch': 2.79}) +[2026-06-18 02:10:47,465][starVLA.training.train_starvla][INFO] - Step 6158, Loss: {'train/loss': 0.01985178515315056, 'train/grad_norm_pre_clip': 3.9126625061035156, 'timing/data': 0.0004936773329973221, 'timing/model': 0.9193688929080963, 'learning_rate/qwen_vl_interface': 8.447972915362941e-07, 'learning_rate/action_model': 8.44797291536294e-06, 'epoch': 2.79}) +[2026-06-18 02:10:51,203][starVLA.training.train_starvla][INFO] - Step 6159, Loss: {'train/loss': 0.1178102120757103, 'train/grad_norm_pre_clip': 17.98894500732422, 'timing/data': 0.0004821252077817917, 'timing/model': 0.9482510285452008, 'learning_rate/qwen_vl_interface': 8.439888074647536e-07, 'learning_rate/action_model': 8.439888074647536e-06, 'epoch': 2.79}) +[2026-06-18 02:10:55,056][starVLA.training.train_starvla][INFO] - Step 6160, Loss: {'train/loss': 0.11293443292379379, 'train/grad_norm_pre_clip': 30.75547981262207, 'timing/data': 0.0005200495943427086, 'timing/model': 0.9213652918115258, 'learning_rate/qwen_vl_interface': 8.431812367640732e-07, 'learning_rate/action_model': 8.431812367640731e-06, 'epoch': 2.79}) +[2026-06-18 02:10:58,847][starVLA.training.train_starvla][INFO] - Step 6161, Loss: {'train/loss': 0.15788929164409637, 'train/grad_norm_pre_clip': 19.442296981811523, 'timing/data': 0.00048457272350788116, 'timing/model': 0.9239434590563178, 'learning_rate/qwen_vl_interface': 8.423745796016659e-07, 'learning_rate/action_model': 8.42374579601666e-06, 'epoch': 2.8}) +[2026-06-18 02:11:02,564][starVLA.training.train_starvla][INFO] - Step 6162, Loss: {'train/loss': 0.028293170034885406, 'train/grad_norm_pre_clip': 7.30438756942749, 'timing/data': 0.0008094487711787224, 'timing/model': 0.922676695510745, 'learning_rate/qwen_vl_interface': 8.41568836144751e-07, 'learning_rate/action_model': 8.41568836144751e-06, 'epoch': 2.8}) +[2026-06-18 02:11:06,369][starVLA.training.train_starvla][INFO] - Step 6163, Loss: {'train/loss': 0.5321105718612671, 'train/grad_norm_pre_clip': 44.7362060546875, 'timing/data': 0.000714106485247612, 'timing/model': 1.1424610363319516, 'learning_rate/qwen_vl_interface': 8.407640065603601e-07, 'learning_rate/action_model': 8.4076400656036e-06, 'epoch': 2.8}) +[2026-06-18 02:11:09,981][starVLA.training.train_starvla][INFO] - Step 6164, Loss: {'train/loss': 0.1978272795677185, 'train/grad_norm_pre_clip': 41.13080978393555, 'timing/data': 0.0006882771849632263, 'timing/model': 0.933219812810421, 'learning_rate/qwen_vl_interface': 8.399600910153336e-07, 'learning_rate/action_model': 8.399600910153336e-06, 'epoch': 2.8}) +[2026-06-18 02:11:13,718][starVLA.training.train_starvla][INFO] - Step 6165, Loss: {'train/loss': 0.1345251053571701, 'train/grad_norm_pre_clip': 31.881593704223633, 'timing/data': 0.000484393909573555, 'timing/model': 0.9230435164645314, 'learning_rate/qwen_vl_interface': 8.391570896763266e-07, 'learning_rate/action_model': 8.391570896763267e-06, 'epoch': 2.8}) +[2026-06-18 02:11:17,423][starVLA.training.train_starvla][INFO] - Step 6166, Loss: {'train/loss': 0.3058723509311676, 'train/grad_norm_pre_clip': 31.400009155273438, 'timing/data': 0.00047691911458969116, 'timing/model': 0.9168219109997153, 'learning_rate/qwen_vl_interface': 8.38355002709802e-07, 'learning_rate/action_model': 8.38355002709802e-06, 'epoch': 2.8}) +[2026-06-18 02:11:21,221][starVLA.training.train_starvla][INFO] - Step 6167, Loss: {'train/loss': 0.02917000651359558, 'train/grad_norm_pre_clip': 6.296507358551025, 'timing/data': 0.0005103172734379768, 'timing/model': 0.9223941368982196, 'learning_rate/qwen_vl_interface': 8.375538302820314e-07, 'learning_rate/action_model': 8.375538302820315e-06, 'epoch': 2.8}) +[2026-06-18 02:11:24,982][starVLA.training.train_starvla][INFO] - Step 6168, Loss: {'train/loss': 0.11489474773406982, 'train/grad_norm_pre_clip': 18.19746971130371, 'timing/data': 0.0004725838080048561, 'timing/model': 0.9269511019811034, 'learning_rate/qwen_vl_interface': 8.367535725591003e-07, 'learning_rate/action_model': 8.367535725591003e-06, 'epoch': 2.8}) +[2026-06-18 02:11:28,695][starVLA.training.train_starvla][INFO] - Step 6169, Loss: {'train/loss': 0.15297313034534454, 'train/grad_norm_pre_clip': 21.723392486572266, 'timing/data': 0.0011070901528000832, 'timing/model': 1.0797886718064547, 'learning_rate/qwen_vl_interface': 8.359542297069024e-07, 'learning_rate/action_model': 8.359542297069024e-06, 'epoch': 2.8}) +[2026-06-18 02:11:32,492][starVLA.training.train_starvla][INFO] - Step 6170, Loss: {'train/loss': 0.15215080976486206, 'train/grad_norm_pre_clip': 27.20393943786621, 'timing/data': 0.0005393326282501221, 'timing/model': 1.0705135110765696, 'learning_rate/qwen_vl_interface': 8.351558018911423e-07, 'learning_rate/action_model': 8.351558018911423e-06, 'epoch': 2.8}) +[2026-06-18 02:11:36,095][starVLA.training.train_starvla][INFO] - Step 6171, Loss: {'train/loss': 0.3036065101623535, 'train/grad_norm_pre_clip': 40.967777252197266, 'timing/data': 0.0008767712861299515, 'timing/model': 0.9807921508327127, 'learning_rate/qwen_vl_interface': 8.343582892773352e-07, 'learning_rate/action_model': 8.343582892773351e-06, 'epoch': 2.8}) +[2026-06-18 02:11:39,795][starVLA.training.train_starvla][INFO] - Step 6172, Loss: {'train/loss': 0.40488025546073914, 'train/grad_norm_pre_clip': 31.228851318359375, 'timing/data': 0.0005521206185221672, 'timing/model': 0.9145445423200727, 'learning_rate/qwen_vl_interface': 8.335616920308061e-07, 'learning_rate/action_model': 8.33561692030806e-06, 'epoch': 2.8}) +[2026-06-18 02:11:43,524][starVLA.training.train_starvla][INFO] - Step 6173, Loss: {'train/loss': 0.20265595614910126, 'train/grad_norm_pre_clip': 24.898937225341797, 'timing/data': 0.0004904372617602348, 'timing/model': 0.9150901287794113, 'learning_rate/qwen_vl_interface': 8.327660103166907e-07, 'learning_rate/action_model': 8.327660103166906e-06, 'epoch': 2.8}) +[2026-06-18 02:11:47,247][starVLA.training.train_starvla][INFO] - Step 6174, Loss: {'train/loss': 0.3177134394645691, 'train/grad_norm_pre_clip': 39.86586380004883, 'timing/data': 0.0010269861668348312, 'timing/model': 0.9123131614178419, 'learning_rate/qwen_vl_interface': 8.319712442999335e-07, 'learning_rate/action_model': 8.319712442999335e-06, 'epoch': 2.8}) +[2026-06-18 02:11:51,016][starVLA.training.train_starvla][INFO] - Step 6175, Loss: {'train/loss': 0.09406992793083191, 'train/grad_norm_pre_clip': 19.094013214111328, 'timing/data': 0.0006937617436051369, 'timing/model': 0.9191339621320367, 'learning_rate/qwen_vl_interface': 8.311773941452924e-07, 'learning_rate/action_model': 8.311773941452925e-06, 'epoch': 2.8}) +[2026-06-18 02:11:54,802][starVLA.training.train_starvla][INFO] - Step 6176, Loss: {'train/loss': 0.28304213285446167, 'train/grad_norm_pre_clip': 30.981170654296875, 'timing/data': 0.001094253733754158, 'timing/model': 1.1259932927787304, 'learning_rate/qwen_vl_interface': 8.303844600173326e-07, 'learning_rate/action_model': 8.303844600173325e-06, 'epoch': 2.8}) +[2026-06-18 02:11:58,565][starVLA.training.train_starvla][INFO] - Step 6177, Loss: {'train/loss': 0.3465898633003235, 'train/grad_norm_pre_clip': 36.330406188964844, 'timing/data': 0.00048809219151735306, 'timing/model': 1.090405348688364, 'learning_rate/qwen_vl_interface': 8.295924420804297e-07, 'learning_rate/action_model': 8.295924420804296e-06, 'epoch': 2.8}) +[2026-06-18 02:12:02,108][starVLA.training.train_starvla][INFO] - Step 6178, Loss: {'train/loss': 0.16303128004074097, 'train/grad_norm_pre_clip': 49.684654235839844, 'timing/data': 0.0008715298026800156, 'timing/model': 0.916068309918046, 'learning_rate/qwen_vl_interface': 8.288013404987691e-07, 'learning_rate/action_model': 8.28801340498769e-06, 'epoch': 2.8}) +[2026-06-18 02:12:05,838][starVLA.training.train_starvla][INFO] - Step 6179, Loss: {'train/loss': 0.05371435731649399, 'train/grad_norm_pre_clip': 9.596700668334961, 'timing/data': 0.0005054119974374771, 'timing/model': 0.9189798207953572, 'learning_rate/qwen_vl_interface': 8.28011155436349e-07, 'learning_rate/action_model': 8.28011155436349e-06, 'epoch': 2.8}) +[2026-06-18 02:12:09,840][starVLA.training.train_starvla][INFO] - Step 6180, Loss: {'train/loss': 0.04559702426195145, 'train/grad_norm_pre_clip': 18.27702522277832, 'timing/data': 0.0011828858405351639, 'timing/model': 1.0248739887028933, 'learning_rate/qwen_vl_interface': 8.272218870569748e-07, 'learning_rate/action_model': 8.272218870569748e-06, 'epoch': 2.8}) +[2026-06-18 02:12:13,677][starVLA.training.train_starvla][INFO] - Step 6181, Loss: {'train/loss': 0.11320474743843079, 'train/grad_norm_pre_clip': 30.847267150878906, 'timing/data': 0.0004931949079036713, 'timing/model': 0.9436454605311155, 'learning_rate/qwen_vl_interface': 8.264335355242625e-07, 'learning_rate/action_model': 8.264335355242625e-06, 'epoch': 2.8}) +[2026-06-18 02:12:17,444][starVLA.training.train_starvla][INFO] - Step 6182, Loss: {'train/loss': 0.331547349691391, 'train/grad_norm_pre_clip': 30.003192901611328, 'timing/data': 0.00048480182886123657, 'timing/model': 0.9282371243461967, 'learning_rate/qwen_vl_interface': 8.256461010016376e-07, 'learning_rate/action_model': 8.256461010016376e-06, 'epoch': 2.8}) +[2026-06-18 02:12:21,159][starVLA.training.train_starvla][INFO] - Step 6183, Loss: {'train/loss': 0.03228411078453064, 'train/grad_norm_pre_clip': 11.29637622833252, 'timing/data': 0.000535392202436924, 'timing/model': 1.080329985357821, 'learning_rate/qwen_vl_interface': 8.248595836523368e-07, 'learning_rate/action_model': 8.248595836523368e-06, 'epoch': 2.81}) +[2026-06-18 02:12:24,901][starVLA.training.train_starvla][INFO] - Step 6184, Loss: {'train/loss': 0.2260873168706894, 'train/grad_norm_pre_clip': 33.0362548828125, 'timing/data': 0.0004892954602837563, 'timing/model': 1.1054052067920566, 'learning_rate/qwen_vl_interface': 8.240739836394068e-07, 'learning_rate/action_model': 8.240739836394068e-06, 'epoch': 2.81}) +[2026-06-18 02:12:28,537][starVLA.training.train_starvla][INFO] - Step 6185, Loss: {'train/loss': 0.24799570441246033, 'train/grad_norm_pre_clip': 37.67242431640625, 'timing/data': 0.00046838633716106415, 'timing/model': 0.9275605482980609, 'learning_rate/qwen_vl_interface': 8.232893011257019e-07, 'learning_rate/action_model': 8.232893011257018e-06, 'epoch': 2.81}) +[2026-06-18 02:12:32,272][starVLA.training.train_starvla][INFO] - Step 6186, Loss: {'train/loss': 0.07847722619771957, 'train/grad_norm_pre_clip': 19.015560150146484, 'timing/data': 0.00048586074262857437, 'timing/model': 0.9212766038253903, 'learning_rate/qwen_vl_interface': 8.225055362738886e-07, 'learning_rate/action_model': 8.225055362738885e-06, 'epoch': 2.81}) +[2026-06-18 02:12:36,481][starVLA.training.train_starvla][INFO] - Step 6187, Loss: {'train/loss': 0.17108504474163055, 'train/grad_norm_pre_clip': 26.1975040435791, 'timing/data': 0.000509927049279213, 'timing/model': 0.9202405326068401, 'learning_rate/qwen_vl_interface': 8.217226892464411e-07, 'learning_rate/action_model': 8.21722689246441e-06, 'epoch': 2.81}) +[2026-06-18 02:12:40,242][starVLA.training.train_starvla][INFO] - Step 6188, Loss: {'train/loss': 0.06151974946260452, 'train/grad_norm_pre_clip': 22.127124786376953, 'timing/data': 0.0010153036564588547, 'timing/model': 0.9687285991385579, 'learning_rate/qwen_vl_interface': 8.209407602056457e-07, 'learning_rate/action_model': 8.209407602056457e-06, 'epoch': 2.81}) +[2026-06-18 02:12:44,162][starVLA.training.train_starvla][INFO] - Step 6189, Loss: {'train/loss': 0.21830712258815765, 'train/grad_norm_pre_clip': 43.27888107299805, 'timing/data': 0.0004524495452642441, 'timing/model': 1.0607049111276865, 'learning_rate/qwen_vl_interface': 8.201597493135962e-07, 'learning_rate/action_model': 8.201597493135961e-06, 'epoch': 2.81}) +[2026-06-18 02:12:48,537][starVLA.training.train_starvla][INFO] - Step 6190, Loss: {'train/loss': 0.25279805064201355, 'train/grad_norm_pre_clip': 54.411094665527344, 'timing/data': 0.0004676477983593941, 'timing/model': 1.0839389096945524, 'learning_rate/qwen_vl_interface': 8.193796567321982e-07, 'learning_rate/action_model': 8.193796567321982e-06, 'epoch': 2.81}) +[2026-06-18 02:12:52,096][starVLA.training.train_starvla][INFO] - Step 6191, Loss: {'train/loss': 0.24086783826351166, 'train/grad_norm_pre_clip': 21.882919311523438, 'timing/data': 0.0004886258393526077, 'timing/model': 0.9205737402662635, 'learning_rate/qwen_vl_interface': 8.186004826231644e-07, 'learning_rate/action_model': 8.186004826231644e-06, 'epoch': 2.81}) +[2026-06-18 02:12:55,801][starVLA.training.train_starvla][INFO] - Step 6192, Loss: {'train/loss': 0.04134032875299454, 'train/grad_norm_pre_clip': 10.490434646606445, 'timing/data': 0.00046663451939821243, 'timing/model': 0.9138175053521991, 'learning_rate/qwen_vl_interface': 8.178222271480192e-07, 'learning_rate/action_model': 8.178222271480191e-06, 'epoch': 2.81}) +[2026-06-18 02:12:59,583][starVLA.training.train_starvla][INFO] - Step 6193, Loss: {'train/loss': 0.1833740472793579, 'train/grad_norm_pre_clip': 25.055500030517578, 'timing/data': 0.0014891717582941055, 'timing/model': 0.999162052758038, 'learning_rate/qwen_vl_interface': 8.170448904680943e-07, 'learning_rate/action_model': 8.170448904680943e-06, 'epoch': 2.81}) +[2026-06-18 02:13:03,541][starVLA.training.train_starvla][INFO] - Step 6194, Loss: {'train/loss': 0.28090399503707886, 'train/grad_norm_pre_clip': 40.808902740478516, 'timing/data': 0.0004926230758428574, 'timing/model': 0.9700524089857936, 'learning_rate/qwen_vl_interface': 8.162684727445347e-07, 'learning_rate/action_model': 8.162684727445346e-06, 'epoch': 2.81}) +[2026-06-18 02:13:07,492][starVLA.training.train_starvla][INFO] - Step 6195, Loss: {'train/loss': 0.2705014944076538, 'train/grad_norm_pre_clip': 27.147781372070312, 'timing/data': 0.0006898576393723488, 'timing/model': 1.0048275729641318, 'learning_rate/qwen_vl_interface': 8.154929741382922e-07, 'learning_rate/action_model': 8.154929741382922e-06, 'epoch': 2.81}) +[2026-06-18 02:13:11,526][starVLA.training.train_starvla][INFO] - Step 6196, Loss: {'train/loss': 0.13304445147514343, 'train/grad_norm_pre_clip': 32.016109466552734, 'timing/data': 0.0008632456883788109, 'timing/model': 1.1152012096717954, 'learning_rate/qwen_vl_interface': 8.147183948101266e-07, 'learning_rate/action_model': 8.147183948101265e-06, 'epoch': 2.81}) +[2026-06-18 02:13:15,207][starVLA.training.train_starvla][INFO] - Step 6197, Loss: {'train/loss': 0.13940057158470154, 'train/grad_norm_pre_clip': 29.45990753173828, 'timing/data': 0.0006593912839889526, 'timing/model': 1.028014617972076, 'learning_rate/qwen_vl_interface': 8.139447349206111e-07, 'learning_rate/action_model': 8.139447349206111e-06, 'epoch': 2.81}) +[2026-06-18 02:13:18,837][starVLA.training.train_starvla][INFO] - Step 6198, Loss: {'train/loss': 0.11872110515832901, 'train/grad_norm_pre_clip': 20.748271942138672, 'timing/data': 0.0006851833313703537, 'timing/model': 0.944598670117557, 'learning_rate/qwen_vl_interface': 8.131719946301255e-07, 'learning_rate/action_model': 8.131719946301255e-06, 'epoch': 2.81}) +[2026-06-18 02:13:22,539][starVLA.training.train_starvla][INFO] - Step 6199, Loss: {'train/loss': 0.2397506833076477, 'train/grad_norm_pre_clip': 39.70068359375, 'timing/data': 0.0004819948226213455, 'timing/model': 0.9191833809018135, 'learning_rate/qwen_vl_interface': 8.124001740988595e-07, 'learning_rate/action_model': 8.124001740988595e-06, 'epoch': 2.81}) +[2026-06-18 02:15:35,188][starVLA.training.train_starvla][INFO] - Step 6200, Loss: {'train/loss': 0.3893163204193115, 'train/grad_norm_pre_clip': 37.36257553100586, 'eval/loss': 0.25742942094802856, 'eval/latency_8/loss': 0.25742942094802856, 'eval/demon_fire_precision': 0.9688359303391384, 'eval/demon_fire_recall': 0.9850885368126747, 'eval/demon_fire_f1': 0.9768946395563771, 'eval/demon_fire_rate_model': 0.9106844741235393, 'eval/demon_fire_rate_teacher': 0.8956594323873122, 'eval/demon_left_precision': 0.9193954659949622, 'eval/demon_left_recall': 0.7815845824411135, 'eval/demon_left_f1': 0.8449074074074074, 'eval/demon_left_rate_model': 0.11046188091263216, 'eval/demon_left_rate_teacher': 0.12993878686700056, 'eval/demon_right_precision': 0.8631578947368421, 'eval/demon_right_recall': 0.8696969696969697, 'eval/demon_right_f1': 0.8664150943396228, 'eval/demon_right_rate_model': 0.18503060656649972, 'eval/demon_right_rate_teacher': 0.18363939899833054, 'eval/demon_move_f1': 0.8556612508735151, 'eval/demon_cc_f1': 0.9370645623413717, 'eval/latency_8/demon_fire_precision': 0.9688359303391384, 'eval/latency_8/demon_fire_recall': 0.9850885368126747, 'eval/latency_8/demon_fire_f1': 0.9768946395563771, 'eval/latency_8/demon_fire_rate_model': 0.9106844741235393, 'eval/latency_8/demon_fire_rate_teacher': 0.8956594323873122, 'eval/latency_8/demon_left_precision': 0.9193954659949622, 'eval/latency_8/demon_left_recall': 0.7815845824411135, 'eval/latency_8/demon_left_f1': 0.8449074074074074, 'eval/latency_8/demon_left_rate_model': 0.11046188091263216, 'eval/latency_8/demon_left_rate_teacher': 0.12993878686700056, 'eval/latency_8/demon_right_precision': 0.8631578947368421, 'eval/latency_8/demon_right_recall': 0.8696969696969697, 'eval/latency_8/demon_right_f1': 0.8664150943396228, 'eval/latency_8/demon_right_rate_model': 0.18503060656649972, 'eval/latency_8/demon_right_rate_teacher': 0.18363939899833054, 'eval/latency_8/demon_move_f1': 0.8556612508735151, 'eval/latency_8/demon_cc_f1': 0.9370645623413717, 'timing/data': 0.000917467288672924, 'timing/model': 1.0312838293612003, 'learning_rate/qwen_vl_interface': 8.116292734868126e-07, 'learning_rate/action_model': 8.116292734868125e-06, 'epoch': 2.81}) +[2026-06-18 02:15:35,206][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_6200_state +[2026-06-18 02:15:35,211][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 02:15:41,773][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_6200_state/model.safetensors +[2026-06-18 02:15:56,934][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_6200_state/optimizer.bin +[2026-06-18 02:15:56,935][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_6200_state/sampler.bin +[2026-06-18 02:15:56,937][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_6200_state/sampler_1.bin +[2026-06-18 02:15:56,940][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_6200_state/random_states_0.pkl +[2026-06-18 02:15:56,943][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 6200 +[2026-06-18 02:15:56,943][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 02:15:56,954][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 02:16:00,762][starVLA.training.train_starvla][INFO] - Step 6201, Loss: {'train/loss': 0.14560255408287048, 'train/grad_norm_pre_clip': 28.0533504486084, 'timing/data': 0.000544404610991478, 'timing/model': 0.9385716766119003, 'learning_rate/qwen_vl_interface': 8.108592929537928e-07, 'learning_rate/action_model': 8.108592929537928e-06, 'epoch': 2.81}) +[2026-06-18 02:16:04,633][starVLA.training.train_starvla][INFO] - Step 6202, Loss: {'train/loss': 0.32223138213157654, 'train/grad_norm_pre_clip': 39.81798553466797, 'timing/data': 0.0008226046338677406, 'timing/model': 0.9375492883846164, 'learning_rate/qwen_vl_interface': 8.100902326594186e-07, 'learning_rate/action_model': 8.100902326594186e-06, 'epoch': 2.81}) +[2026-06-18 02:16:08,433][starVLA.training.train_starvla][INFO] - Step 6203, Loss: {'train/loss': 0.044768210500478745, 'train/grad_norm_pre_clip': 11.167008399963379, 'timing/data': 0.0008406993001699448, 'timing/model': 0.9308381890878081, 'learning_rate/qwen_vl_interface': 8.093220927631171e-07, 'learning_rate/action_model': 8.09322092763117e-06, 'epoch': 2.81}) +[2026-06-18 02:16:12,215][starVLA.training.train_starvla][INFO] - Step 6204, Loss: {'train/loss': 0.32940563559532166, 'train/grad_norm_pre_clip': 31.367748260498047, 'timing/data': 0.0007361834868788719, 'timing/model': 1.122485714033246, 'learning_rate/qwen_vl_interface': 8.085548734241233e-07, 'learning_rate/action_model': 8.085548734241232e-06, 'epoch': 2.81}) +[2026-06-18 02:16:15,773][starVLA.training.train_starvla][INFO] - Step 6205, Loss: {'train/loss': 0.020361697301268578, 'train/grad_norm_pre_clip': 4.30704927444458, 'timing/data': 0.0005145641043782234, 'timing/model': 0.9142268169671297, 'learning_rate/qwen_vl_interface': 8.077885748014843e-07, 'learning_rate/action_model': 8.077885748014841e-06, 'epoch': 2.82}) +[2026-06-18 02:16:19,499][starVLA.training.train_starvla][INFO] - Step 6206, Loss: {'train/loss': 0.05390232801437378, 'train/grad_norm_pre_clip': 14.152600288391113, 'timing/data': 0.0005794819444417953, 'timing/model': 0.9297731993719935, 'learning_rate/qwen_vl_interface': 8.070231970540532e-07, 'learning_rate/action_model': 8.070231970540532e-06, 'epoch': 2.82}) +[2026-06-18 02:16:23,316][starVLA.training.train_starvla][INFO] - Step 6207, Loss: {'train/loss': 0.5109109878540039, 'train/grad_norm_pre_clip': 41.16361999511719, 'timing/data': 0.0007300497964024544, 'timing/model': 1.0309170559048653, 'learning_rate/qwen_vl_interface': 8.062587403404934e-07, 'learning_rate/action_model': 8.062587403404934e-06, 'epoch': 2.82}) +[2026-06-18 02:16:27,096][starVLA.training.train_starvla][INFO] - Step 6208, Loss: {'train/loss': 0.10706979036331177, 'train/grad_norm_pre_clip': 22.39352035522461, 'timing/data': 0.0005617709830403328, 'timing/model': 0.9300763048231602, 'learning_rate/qwen_vl_interface': 8.054952048192793e-07, 'learning_rate/action_model': 8.054952048192794e-06, 'epoch': 2.82}) +[2026-06-18 02:16:30,880][starVLA.training.train_starvla][INFO] - Step 6209, Loss: {'train/loss': 0.20599596202373505, 'train/grad_norm_pre_clip': 26.851890563964844, 'timing/data': 0.0006897160783410072, 'timing/model': 0.9275319641456008, 'learning_rate/qwen_vl_interface': 8.047325906486926e-07, 'learning_rate/action_model': 8.047325906486925e-06, 'epoch': 2.82}) +[2026-06-18 02:16:34,898][starVLA.training.train_starvla][INFO] - Step 6210, Loss: {'train/loss': 0.38353005051612854, 'train/grad_norm_pre_clip': 33.931678771972656, 'timing/data': 0.00048035476356744766, 'timing/model': 1.3181398278102279, 'learning_rate/qwen_vl_interface': 8.039708979868229e-07, 'learning_rate/action_model': 8.039708979868227e-06, 'epoch': 2.82}) +[2026-06-18 02:16:38,668][starVLA.training.train_starvla][INFO] - Step 6211, Loss: {'train/loss': 0.16549190878868103, 'train/grad_norm_pre_clip': 18.9249267578125, 'timing/data': 0.0004842737689614296, 'timing/model': 1.0866652997210622, 'learning_rate/qwen_vl_interface': 8.032101269915687e-07, 'learning_rate/action_model': 8.032101269915687e-06, 'epoch': 2.82}) +[2026-06-18 02:16:42,603][starVLA.training.train_starvla][INFO] - Step 6212, Loss: {'train/loss': 0.11677925288677216, 'train/grad_norm_pre_clip': 25.3791561126709, 'timing/data': 0.0008329041302204132, 'timing/model': 0.9824860636144876, 'learning_rate/qwen_vl_interface': 8.024502778206411e-07, 'learning_rate/action_model': 8.02450277820641e-06, 'epoch': 2.82}) +[2026-06-18 02:16:46,401][starVLA.training.train_starvla][INFO] - Step 6213, Loss: {'train/loss': 0.019339343532919884, 'train/grad_norm_pre_clip': 4.400068759918213, 'timing/data': 0.00047075189650058746, 'timing/model': 0.9259733129292727, 'learning_rate/qwen_vl_interface': 8.016913506315566e-07, 'learning_rate/action_model': 8.016913506315565e-06, 'epoch': 2.82}) +[2026-06-18 02:16:50,477][starVLA.training.train_starvla][INFO] - Step 6214, Loss: {'train/loss': 0.049398571252822876, 'train/grad_norm_pre_clip': 24.18497085571289, 'timing/data': 0.00048327352851629257, 'timing/model': 1.1652972120791674, 'learning_rate/qwen_vl_interface': 8.009333455816416e-07, 'learning_rate/action_model': 8.009333455816417e-06, 'epoch': 2.82}) +[2026-06-18 02:16:54,588][starVLA.training.train_starvla][INFO] - Step 6215, Loss: {'train/loss': 0.1625925451517105, 'train/grad_norm_pre_clip': 35.158973693847656, 'timing/data': 0.00049615278840065, 'timing/model': 0.9378959536552429, 'learning_rate/qwen_vl_interface': 8.001762628280318e-07, 'learning_rate/action_model': 8.001762628280319e-06, 'epoch': 2.82}) +[2026-06-18 02:16:58,885][starVLA.training.train_starvla][INFO] - Step 6216, Loss: {'train/loss': 0.15016600489616394, 'train/grad_norm_pre_clip': 18.509647369384766, 'timing/data': 0.0004815841093659401, 'timing/model': 0.9303251029923558, 'learning_rate/qwen_vl_interface': 7.994201025276708e-07, 'learning_rate/action_model': 7.994201025276708e-06, 'epoch': 2.82}) +[2026-06-18 02:17:02,770][starVLA.training.train_starvla][INFO] - Step 6217, Loss: {'train/loss': 0.22086933255195618, 'train/grad_norm_pre_clip': 33.348228454589844, 'timing/data': 0.000875718891620636, 'timing/model': 1.1779367038980126, 'learning_rate/qwen_vl_interface': 7.986648648373115e-07, 'learning_rate/action_model': 7.986648648373115e-06, 'epoch': 2.82}) +[2026-06-18 02:17:06,805][starVLA.training.train_starvla][INFO] - Step 6218, Loss: {'train/loss': 0.21156539022922516, 'train/grad_norm_pre_clip': 34.7479248046875, 'timing/data': 0.0004882980138063431, 'timing/model': 1.2221629237756133, 'learning_rate/qwen_vl_interface': 7.979105499135157e-07, 'learning_rate/action_model': 7.979105499135155e-06, 'epoch': 2.82}) +[2026-06-18 02:17:10,453][starVLA.training.train_starvla][INFO] - Step 6219, Loss: {'train/loss': 0.0649130716919899, 'train/grad_norm_pre_clip': 9.643582344055176, 'timing/data': 0.0005424246191978455, 'timing/model': 0.9343846226111054, 'learning_rate/qwen_vl_interface': 7.971571579126533e-07, 'learning_rate/action_model': 7.971571579126534e-06, 'epoch': 2.82}) +[2026-06-18 02:17:14,323][starVLA.training.train_starvla][INFO] - Step 6220, Loss: {'train/loss': 0.07115588337182999, 'train/grad_norm_pre_clip': 19.884164810180664, 'timing/data': 0.0005018655210733414, 'timing/model': 0.9488112442195415, 'learning_rate/qwen_vl_interface': 7.964046889909038e-07, 'learning_rate/action_model': 7.964046889909038e-06, 'epoch': 2.82}) +[2026-06-18 02:17:18,109][starVLA.training.train_starvla][INFO] - Step 6221, Loss: {'train/loss': 0.2572192847728729, 'train/grad_norm_pre_clip': 38.88996887207031, 'timing/data': 0.0010297438129782677, 'timing/model': 0.9335576854646206, 'learning_rate/qwen_vl_interface': 7.956531433042543e-07, 'learning_rate/action_model': 7.956531433042542e-06, 'epoch': 2.82}) +[2026-06-18 02:17:21,915][starVLA.training.train_starvla][INFO] - Step 6222, Loss: {'train/loss': 0.1556122750043869, 'train/grad_norm_pre_clip': 37.991722106933594, 'timing/data': 0.0004955660551786423, 'timing/model': 0.9576489944010973, 'learning_rate/qwen_vl_interface': 7.94902521008501e-07, 'learning_rate/action_model': 7.94902521008501e-06, 'epoch': 2.82}) +[2026-06-18 02:17:25,851][starVLA.training.train_starvla][INFO] - Step 6223, Loss: {'train/loss': 0.7364950776100159, 'train/grad_norm_pre_clip': 70.18103790283203, 'timing/data': 0.000510471872985363, 'timing/model': 0.9512912062928081, 'learning_rate/qwen_vl_interface': 7.941528222592501e-07, 'learning_rate/action_model': 7.941528222592501e-06, 'epoch': 2.82}) +[2026-06-18 02:17:29,740][starVLA.training.train_starvla][INFO] - Step 6224, Loss: {'train/loss': 0.1201242133975029, 'train/grad_norm_pre_clip': 15.491814613342285, 'timing/data': 0.0007157661020755768, 'timing/model': 1.1332707051187754, 'learning_rate/qwen_vl_interface': 7.934040472119132e-07, 'learning_rate/action_model': 7.934040472119133e-06, 'epoch': 2.82}) +[2026-06-18 02:17:33,428][starVLA.training.train_starvla][INFO] - Step 6225, Loss: {'train/loss': 0.036849211901426315, 'train/grad_norm_pre_clip': 9.657788276672363, 'timing/data': 0.0005999468266963959, 'timing/model': 1.0034751696512103, 'learning_rate/qwen_vl_interface': 7.926561960217127e-07, 'learning_rate/action_model': 7.926561960217126e-06, 'epoch': 2.82}) +[2026-06-18 02:17:37,274][starVLA.training.train_starvla][INFO] - Step 6226, Loss: {'train/loss': 0.24183061718940735, 'train/grad_norm_pre_clip': 30.184803009033203, 'timing/data': 0.0004857173189520836, 'timing/model': 0.9496645797044039, 'learning_rate/qwen_vl_interface': 7.919092688436783e-07, 'learning_rate/action_model': 7.919092688436783e-06, 'epoch': 2.82}) +[2026-06-18 02:17:41,125][starVLA.training.train_starvla][INFO] - Step 6227, Loss: {'train/loss': 0.05927353352308273, 'train/grad_norm_pre_clip': 20.868083953857422, 'timing/data': 0.0005775792524218559, 'timing/model': 0.9250319711863995, 'learning_rate/qwen_vl_interface': 7.911632658326505e-07, 'learning_rate/action_model': 7.911632658326505e-06, 'epoch': 2.83}) +[2026-06-18 02:17:44,907][starVLA.training.train_starvla][INFO] - Step 6228, Loss: {'train/loss': 0.08530452847480774, 'train/grad_norm_pre_clip': 22.780071258544922, 'timing/data': 0.0004858691245317459, 'timing/model': 0.934702611528337, 'learning_rate/qwen_vl_interface': 7.904181871432757e-07, 'learning_rate/action_model': 7.904181871432757e-06, 'epoch': 2.83}) +[2026-06-18 02:17:48,735][starVLA.training.train_starvla][INFO] - Step 6229, Loss: {'train/loss': 0.14847137033939362, 'train/grad_norm_pre_clip': 27.409305572509766, 'timing/data': 0.0009696604683995247, 'timing/model': 0.9615799132734537, 'learning_rate/qwen_vl_interface': 7.896740329300091e-07, 'learning_rate/action_model': 7.89674032930009e-06, 'epoch': 2.83}) +[2026-06-18 02:17:52,520][starVLA.training.train_starvla][INFO] - Step 6230, Loss: {'train/loss': 0.33750632405281067, 'train/grad_norm_pre_clip': 65.36526489257812, 'timing/data': 0.0008061351254582405, 'timing/model': 1.1214655740186572, 'learning_rate/qwen_vl_interface': 7.889308033471156e-07, 'learning_rate/action_model': 7.889308033471155e-06, 'epoch': 2.83}) +[2026-06-18 02:17:56,292][starVLA.training.train_starvla][INFO] - Step 6231, Loss: {'train/loss': 0.02656482718884945, 'train/grad_norm_pre_clip': 10.731972694396973, 'timing/data': 0.0004978002980351448, 'timing/model': 1.0746376849710941, 'learning_rate/qwen_vl_interface': 7.881884985486662e-07, 'learning_rate/action_model': 7.881884985486662e-06, 'epoch': 2.83}) +[2026-06-18 02:17:59,935][starVLA.training.train_starvla][INFO] - Step 6232, Loss: {'train/loss': 0.07152538001537323, 'train/grad_norm_pre_clip': 17.031620025634766, 'timing/data': 0.0004925420507788658, 'timing/model': 0.9642477910965681, 'learning_rate/qwen_vl_interface': 7.874471186885427e-07, 'learning_rate/action_model': 7.874471186885427e-06, 'epoch': 2.83}) +[2026-06-18 02:18:03,727][starVLA.training.train_starvla][INFO] - Step 6233, Loss: {'train/loss': 0.3343683183193207, 'train/grad_norm_pre_clip': 51.483787536621094, 'timing/data': 0.00047435425221920013, 'timing/model': 0.9442945923656225, 'learning_rate/qwen_vl_interface': 7.867066639204326e-07, 'learning_rate/action_model': 7.867066639204327e-06, 'epoch': 2.83}) +[2026-06-18 02:18:07,538][starVLA.training.train_starvla][INFO] - Step 6234, Loss: {'train/loss': 0.12181830406188965, 'train/grad_norm_pre_clip': 32.55192565917969, 'timing/data': 0.0006441818550229073, 'timing/model': 0.9294121703132987, 'learning_rate/qwen_vl_interface': 7.859671343978348e-07, 'learning_rate/action_model': 7.859671343978348e-06, 'epoch': 2.83}) +[2026-06-18 02:18:11,426][starVLA.training.train_starvla][INFO] - Step 6235, Loss: {'train/loss': 0.25897878408432007, 'train/grad_norm_pre_clip': 33.895381927490234, 'timing/data': 0.000646510161459446, 'timing/model': 0.9371753595769405, 'learning_rate/qwen_vl_interface': 7.852285302740528e-07, 'learning_rate/action_model': 7.852285302740527e-06, 'epoch': 2.83}) +[2026-06-18 02:18:15,385][starVLA.training.train_starvla][INFO] - Step 6236, Loss: {'train/loss': 0.2222720980644226, 'train/grad_norm_pre_clip': 30.8793888092041, 'timing/data': 0.0004808371886610985, 'timing/model': 0.9345647552981973, 'learning_rate/qwen_vl_interface': 7.844908517022005e-07, 'learning_rate/action_model': 7.844908517022005e-06, 'epoch': 2.83}) +[2026-06-18 02:18:19,415][starVLA.training.train_starvla][INFO] - Step 6237, Loss: {'train/loss': 0.11040773242712021, 'train/grad_norm_pre_clip': 27.202621459960938, 'timing/data': 0.0007255570963025093, 'timing/model': 1.1094725094735622, 'learning_rate/qwen_vl_interface': 7.837540988352001e-07, 'learning_rate/action_model': 7.837540988352e-06, 'epoch': 2.83}) +[2026-06-18 02:18:24,331][starVLA.training.train_starvla][INFO] - Step 6238, Loss: {'train/loss': 0.1879023313522339, 'train/grad_norm_pre_clip': 38.72386932373047, 'timing/data': 0.0006513064727187157, 'timing/model': 1.1017923802137375, 'learning_rate/qwen_vl_interface': 7.83018271825782e-07, 'learning_rate/action_model': 7.83018271825782e-06, 'epoch': 2.83}) +[2026-06-18 02:18:28,002][starVLA.training.train_starvla][INFO] - Step 6239, Loss: {'train/loss': 0.11022445559501648, 'train/grad_norm_pre_clip': 17.597097396850586, 'timing/data': 0.0005039162933826447, 'timing/model': 0.9359692195430398, 'learning_rate/qwen_vl_interface': 7.822833708264821e-07, 'learning_rate/action_model': 7.822833708264821e-06, 'epoch': 2.83}) +[2026-06-18 02:18:31,957][starVLA.training.train_starvla][INFO] - Step 6240, Loss: {'train/loss': 0.15029455721378326, 'train/grad_norm_pre_clip': 12.229991912841797, 'timing/data': 0.0004956750199198723, 'timing/model': 1.1303372578695416, 'learning_rate/qwen_vl_interface': 7.81549395989646e-07, 'learning_rate/action_model': 7.81549395989646e-06, 'epoch': 2.83}) +[2026-06-18 02:18:35,937][starVLA.training.train_starvla][INFO] - Step 6241, Loss: {'train/loss': 0.4048982858657837, 'train/grad_norm_pre_clip': 56.89779281616211, 'timing/data': 0.0004934109747409821, 'timing/model': 0.9602401750162244, 'learning_rate/qwen_vl_interface': 7.808163474674291e-07, 'learning_rate/action_model': 7.80816347467429e-06, 'epoch': 2.83}) +[2026-06-18 02:18:39,775][starVLA.training.train_starvla][INFO] - Step 6242, Loss: {'train/loss': 0.2449609935283661, 'train/grad_norm_pre_clip': 66.32020568847656, 'timing/data': 0.00048394594341516495, 'timing/model': 0.9312276672571898, 'learning_rate/qwen_vl_interface': 7.800842254117922e-07, 'learning_rate/action_model': 7.80084225411792e-06, 'epoch': 2.83}) +[2026-06-18 02:18:43,576][starVLA.training.train_starvla][INFO] - Step 6243, Loss: {'train/loss': 0.058521684259176254, 'train/grad_norm_pre_clip': 13.175108909606934, 'timing/data': 0.00088494922965765, 'timing/model': 0.9235979290679097, 'learning_rate/qwen_vl_interface': 7.793530299745061e-07, 'learning_rate/action_model': 7.793530299745061e-06, 'epoch': 2.83}) +[2026-06-18 02:18:47,407][starVLA.training.train_starvla][INFO] - Step 6244, Loss: {'train/loss': 0.5411206483840942, 'train/grad_norm_pre_clip': 35.63367462158203, 'timing/data': 0.0006970269605517387, 'timing/model': 1.1065224828198552, 'learning_rate/qwen_vl_interface': 7.786227613071464e-07, 'learning_rate/action_model': 7.786227613071464e-06, 'epoch': 2.83}) +[2026-06-18 02:18:51,088][starVLA.training.train_starvla][INFO] - Step 6245, Loss: {'train/loss': 0.02026568353176117, 'train/grad_norm_pre_clip': 4.588949680328369, 'timing/data': 0.0005000894889235497, 'timing/model': 1.0052575068548322, 'learning_rate/qwen_vl_interface': 7.778934195610998e-07, 'learning_rate/action_model': 7.778934195610998e-06, 'epoch': 2.83}) +[2026-06-18 02:18:54,901][starVLA.training.train_starvla][INFO] - Step 6246, Loss: {'train/loss': 0.32083365321159363, 'train/grad_norm_pre_clip': 36.33356857299805, 'timing/data': 0.0009572328999638557, 'timing/model': 0.9486318929120898, 'learning_rate/qwen_vl_interface': 7.771650048875598e-07, 'learning_rate/action_model': 7.771650048875597e-06, 'epoch': 2.83}) +[2026-06-18 02:18:59,905][starVLA.training.train_starvla][INFO] - Step 6247, Loss: {'train/loss': 0.0988488718867302, 'train/grad_norm_pre_clip': 23.574413299560547, 'timing/data': 0.0004905741661787033, 'timing/model': 0.9481172524392605, 'learning_rate/qwen_vl_interface': 7.764375174375276e-07, 'learning_rate/action_model': 7.764375174375276e-06, 'epoch': 2.83}) +[2026-06-18 02:19:03,714][starVLA.training.train_starvla][INFO] - Step 6248, Loss: {'train/loss': 0.6986960172653198, 'train/grad_norm_pre_clip': 52.11170196533203, 'timing/data': 0.0005564186722040176, 'timing/model': 0.9462492167949677, 'learning_rate/qwen_vl_interface': 7.757109573618116e-07, 'learning_rate/action_model': 7.757109573618115e-06, 'epoch': 2.83}) +[2026-06-18 02:19:07,626][starVLA.training.train_starvla][INFO] - Step 6249, Loss: {'train/loss': 0.18336056172847748, 'train/grad_norm_pre_clip': 24.126752853393555, 'timing/data': 0.0005151545628905296, 'timing/model': 1.0389223005622625, 'learning_rate/qwen_vl_interface': 7.749853248110287e-07, 'learning_rate/action_model': 7.749853248110287e-06, 'epoch': 2.83}) +[2026-06-18 02:19:11,439][starVLA.training.train_starvla][INFO] - Step 6250, Loss: {'train/loss': 0.05022348463535309, 'train/grad_norm_pre_clip': 10.25890064239502, 'timing/data': 0.0005399323999881744, 'timing/model': 1.14831165317446, 'learning_rate/qwen_vl_interface': 7.742606199356032e-07, 'learning_rate/action_model': 7.74260619935603e-06, 'epoch': 2.84}) +[2026-06-18 02:19:15,203][starVLA.training.train_starvla][INFO] - Step 6251, Loss: {'train/loss': 0.26518169045448303, 'train/grad_norm_pre_clip': 45.91474914550781, 'timing/data': 0.0008414750918745995, 'timing/model': 1.0730486754328012, 'learning_rate/qwen_vl_interface': 7.735368428857665e-07, 'learning_rate/action_model': 7.735368428857665e-06, 'epoch': 2.84}) +[2026-06-18 02:19:18,836][starVLA.training.train_starvla][INFO] - Step 6252, Loss: {'train/loss': 0.2189590483903885, 'train/grad_norm_pre_clip': 16.268184661865234, 'timing/data': 0.0009804386645555496, 'timing/model': 0.9483834207057953, 'learning_rate/qwen_vl_interface': 7.728139938115605e-07, 'learning_rate/action_model': 7.728139938115604e-06, 'epoch': 2.84}) +[2026-06-18 02:19:22,770][starVLA.training.train_starvla][INFO] - Step 6253, Loss: {'train/loss': 0.18358147144317627, 'train/grad_norm_pre_clip': 26.186227798461914, 'timing/data': 0.0009607644751667976, 'timing/model': 0.9664498902857304, 'learning_rate/qwen_vl_interface': 7.720920728628306e-07, 'learning_rate/action_model': 7.720920728628306e-06, 'epoch': 2.84}) +[2026-06-18 02:19:26,609][starVLA.training.train_starvla][INFO] - Step 6254, Loss: {'train/loss': 0.19720004498958588, 'train/grad_norm_pre_clip': 28.410888671875, 'timing/data': 0.000532771460711956, 'timing/model': 0.937807715497911, 'learning_rate/qwen_vl_interface': 7.713710801892327e-07, 'learning_rate/action_model': 7.713710801892326e-06, 'epoch': 2.84}) +[2026-06-18 02:19:30,431][starVLA.training.train_starvla][INFO] - Step 6255, Loss: {'train/loss': 0.1696963906288147, 'train/grad_norm_pre_clip': 14.370591163635254, 'timing/data': 0.0007788734510540962, 'timing/model': 0.9220834588631988, 'learning_rate/qwen_vl_interface': 7.70651015940228e-07, 'learning_rate/action_model': 7.70651015940228e-06, 'epoch': 2.84}) +[2026-06-18 02:19:34,198][starVLA.training.train_starvla][INFO] - Step 6256, Loss: {'train/loss': 0.6440185308456421, 'train/grad_norm_pre_clip': 33.94816589355469, 'timing/data': 0.0005004741251468658, 'timing/model': 0.92818424385041, 'learning_rate/qwen_vl_interface': 7.699318802650891e-07, 'learning_rate/action_model': 7.69931880265089e-06, 'epoch': 2.84}) +[2026-06-18 02:19:37,993][starVLA.training.train_starvla][INFO] - Step 6257, Loss: {'train/loss': 0.01934996247291565, 'train/grad_norm_pre_clip': 4.786098480224609, 'timing/data': 0.0005112439393997192, 'timing/model': 1.1013997569680214, 'learning_rate/qwen_vl_interface': 7.692136733128918e-07, 'learning_rate/action_model': 7.692136733128917e-06, 'epoch': 2.84}) +[2026-06-18 02:19:41,792][starVLA.training.train_starvla][INFO] - Step 6258, Loss: {'train/loss': 0.25609090924263, 'train/grad_norm_pre_clip': 39.42850112915039, 'timing/data': 0.0005061402916908264, 'timing/model': 1.1161459926515818, 'learning_rate/qwen_vl_interface': 7.684963952325207e-07, 'learning_rate/action_model': 7.684963952325206e-06, 'epoch': 2.84}) +[2026-06-18 02:19:45,397][starVLA.training.train_starvla][INFO] - Step 6259, Loss: {'train/loss': 0.1781964749097824, 'train/grad_norm_pre_clip': 26.20172119140625, 'timing/data': 0.0004631122574210167, 'timing/model': 0.9389549363404512, 'learning_rate/qwen_vl_interface': 7.6778004617267e-07, 'learning_rate/action_model': 7.6778004617267e-06, 'epoch': 2.84}) +[2026-06-18 02:19:49,273][starVLA.training.train_starvla][INFO] - Step 6260, Loss: {'train/loss': 0.3232639729976654, 'train/grad_norm_pre_clip': 93.6912612915039, 'timing/data': 0.0004827473312616348, 'timing/model': 0.9382775686681271, 'learning_rate/qwen_vl_interface': 7.670646262818387e-07, 'learning_rate/action_model': 7.670646262818387e-06, 'epoch': 2.84}) +[2026-06-18 02:19:53,551][starVLA.training.train_starvla][INFO] - Step 6261, Loss: {'train/loss': 0.25114017724990845, 'train/grad_norm_pre_clip': 22.685121536254883, 'timing/data': 0.0005125459283590317, 'timing/model': 0.9295810395851731, 'learning_rate/qwen_vl_interface': 7.663501357083342e-07, 'learning_rate/action_model': 7.66350135708334e-06, 'epoch': 2.84}) +[2026-06-18 02:19:57,407][starVLA.training.train_starvla][INFO] - Step 6262, Loss: {'train/loss': 0.3004642128944397, 'train/grad_norm_pre_clip': 28.240758895874023, 'timing/data': 0.0006589610129594803, 'timing/model': 0.9430337222293019, 'learning_rate/qwen_vl_interface': 7.656365746002715e-07, 'learning_rate/action_model': 7.656365746002716e-06, 'epoch': 2.84}) +[2026-06-18 02:20:01,212][starVLA.training.train_starvla][INFO] - Step 6263, Loss: {'train/loss': 0.2718643248081207, 'train/grad_norm_pre_clip': 33.91621017456055, 'timing/data': 0.0009490838274359703, 'timing/model': 0.9584143748506904, 'learning_rate/qwen_vl_interface': 7.649239431055718e-07, 'learning_rate/action_model': 7.649239431055717e-06, 'epoch': 2.84}) +[2026-06-18 02:20:04,987][starVLA.training.train_starvla][INFO] - Step 6264, Loss: {'train/loss': 0.21347828209400177, 'train/grad_norm_pre_clip': 23.463329315185547, 'timing/data': 0.00048386864364147186, 'timing/model': 1.1083151400089264, 'learning_rate/qwen_vl_interface': 7.642122413719652e-07, 'learning_rate/action_model': 7.642122413719651e-06, 'epoch': 2.84}) +[2026-06-18 02:20:08,713][starVLA.training.train_starvla][INFO] - Step 6265, Loss: {'train/loss': 0.21321120858192444, 'train/grad_norm_pre_clip': 19.48995590209961, 'timing/data': 0.0005069496110081673, 'timing/model': 0.9332266673445702, 'learning_rate/qwen_vl_interface': 7.635014695469877e-07, 'learning_rate/action_model': 7.635014695469878e-06, 'epoch': 2.84}) +[2026-06-18 02:20:12,537][starVLA.training.train_starvla][INFO] - Step 6266, Loss: {'train/loss': 0.45754724740982056, 'train/grad_norm_pre_clip': 35.931522369384766, 'timing/data': 0.0006611933931708336, 'timing/model': 0.9343310641124845, 'learning_rate/qwen_vl_interface': 7.627916277779832e-07, 'learning_rate/action_model': 7.627916277779831e-06, 'epoch': 2.84}) +[2026-06-18 02:20:16,360][starVLA.training.train_starvla][INFO] - Step 6267, Loss: {'train/loss': 0.058783192187547684, 'train/grad_norm_pre_clip': 24.426677703857422, 'timing/data': 0.000501963309943676, 'timing/model': 0.9956640331074595, 'learning_rate/qwen_vl_interface': 7.62082716212103e-07, 'learning_rate/action_model': 7.62082716212103e-06, 'epoch': 2.84}) +[2026-06-18 02:20:20,129][starVLA.training.train_starvla][INFO] - Step 6268, Loss: {'train/loss': 0.12397740036249161, 'train/grad_norm_pre_clip': 31.526607513427734, 'timing/data': 0.0007797032594680786, 'timing/model': 0.9257402438670397, 'learning_rate/qwen_vl_interface': 7.613747349963048e-07, 'learning_rate/action_model': 7.613747349963048e-06, 'epoch': 2.84}) +[2026-06-18 02:20:23,916][starVLA.training.train_starvla][INFO] - Step 6269, Loss: {'train/loss': 0.11041872948408127, 'train/grad_norm_pre_clip': 15.253801345825195, 'timing/data': 0.0007313713431358337, 'timing/model': 0.931880715303123, 'learning_rate/qwen_vl_interface': 7.606676842773531e-07, 'learning_rate/action_model': 7.60667684277353e-06, 'epoch': 2.84}) +[2026-06-18 02:20:27,711][starVLA.training.train_starvla][INFO] - Step 6270, Loss: {'train/loss': 0.41121238470077515, 'train/grad_norm_pre_clip': 33.44819259643555, 'timing/data': 0.0007712636142969131, 'timing/model': 1.1104629635810852, 'learning_rate/qwen_vl_interface': 7.599615642018218e-07, 'learning_rate/action_model': 7.599615642018218e-06, 'epoch': 2.84}) +[2026-06-18 02:20:31,474][starVLA.training.train_starvla][INFO] - Step 6271, Loss: {'train/loss': 0.1842956691980362, 'train/grad_norm_pre_clip': 20.2819766998291, 'timing/data': 0.0006109261885285378, 'timing/model': 1.0755377104505897, 'learning_rate/qwen_vl_interface': 7.592563749160908e-07, 'learning_rate/action_model': 7.592563749160908e-06, 'epoch': 2.84}) +[2026-06-18 02:20:35,193][starVLA.training.train_starvla][INFO] - Step 6272, Loss: {'train/loss': 0.2509707808494568, 'train/grad_norm_pre_clip': 36.01504898071289, 'timing/data': 0.0004947483539581299, 'timing/model': 1.0089721838012338, 'learning_rate/qwen_vl_interface': 7.585521165663445e-07, 'learning_rate/action_model': 7.585521165663445e-06, 'epoch': 2.85}) +[2026-06-18 02:20:39,220][starVLA.training.train_starvla][INFO] - Step 6273, Loss: {'train/loss': 0.08394545316696167, 'train/grad_norm_pre_clip': 28.488557815551758, 'timing/data': 0.0005945507436990738, 'timing/model': 1.031966426409781, 'learning_rate/qwen_vl_interface': 7.578487892985769e-07, 'learning_rate/action_model': 7.578487892985769e-06, 'epoch': 2.85}) +[2026-06-18 02:20:43,064][starVLA.training.train_starvla][INFO] - Step 6274, Loss: {'train/loss': 0.350689172744751, 'train/grad_norm_pre_clip': 22.106698989868164, 'timing/data': 0.0006816843524575233, 'timing/model': 0.9792135711759329, 'learning_rate/qwen_vl_interface': 7.571463932585901e-07, 'learning_rate/action_model': 7.571463932585901e-06, 'epoch': 2.85}) +[2026-06-18 02:20:47,221][starVLA.training.train_starvla][INFO] - Step 6275, Loss: {'train/loss': 0.060530368238687515, 'train/grad_norm_pre_clip': 15.535506248474121, 'timing/data': 0.0004961565136909485, 'timing/model': 0.9373372355476022, 'learning_rate/qwen_vl_interface': 7.564449285919897e-07, 'learning_rate/action_model': 7.564449285919897e-06, 'epoch': 2.85}) +[2026-06-18 02:20:51,071][starVLA.training.train_starvla][INFO] - Step 6276, Loss: {'train/loss': 0.18281735479831696, 'train/grad_norm_pre_clip': 42.99617385864258, 'timing/data': 0.000511527992784977, 'timing/model': 1.0041770031675696, 'learning_rate/qwen_vl_interface': 7.557443954441924e-07, 'learning_rate/action_model': 7.557443954441923e-06, 'epoch': 2.85}) +[2026-06-18 02:20:55,008][starVLA.training.train_starvla][INFO] - Step 6277, Loss: {'train/loss': 0.21780885756015778, 'train/grad_norm_pre_clip': 23.49283218383789, 'timing/data': 0.0008615581318736076, 'timing/model': 1.1987958420068026, 'learning_rate/qwen_vl_interface': 7.550447939604161e-07, 'learning_rate/action_model': 7.550447939604161e-06, 'epoch': 2.85}) +[2026-06-18 02:20:58,999][starVLA.training.train_starvla][INFO] - Step 6278, Loss: {'train/loss': 0.41752228140830994, 'train/grad_norm_pre_clip': 47.15390396118164, 'timing/data': 0.0005732467398047447, 'timing/model': 1.1449789116159081, 'learning_rate/qwen_vl_interface': 7.543461242856916e-07, 'learning_rate/action_model': 7.543461242856916e-06, 'epoch': 2.85}) +[2026-06-18 02:21:02,782][starVLA.training.train_starvla][INFO] - Step 6279, Loss: {'train/loss': 0.07151991128921509, 'train/grad_norm_pre_clip': 17.105031967163086, 'timing/data': 0.0010049082338809967, 'timing/model': 0.9866787735372782, 'learning_rate/qwen_vl_interface': 7.536483865648531e-07, 'learning_rate/action_model': 7.536483865648531e-06, 'epoch': 2.85}) +[2026-06-18 02:21:08,394][starVLA.training.train_starvla][INFO] - Step 6280, Loss: {'train/loss': 0.35377228260040283, 'train/grad_norm_pre_clip': 22.099040985107422, 'timing/data': 0.000490131787955761, 'timing/model': 1.0236770007759333, 'learning_rate/qwen_vl_interface': 7.52951580942542e-07, 'learning_rate/action_model': 7.5295158094254195e-06, 'epoch': 2.85}) +[2026-06-18 02:21:12,259][starVLA.training.train_starvla][INFO] - Step 6281, Loss: {'train/loss': 0.2321656495332718, 'train/grad_norm_pre_clip': 34.63286209106445, 'timing/data': 0.000525384210050106, 'timing/model': 0.9493432240560651, 'learning_rate/qwen_vl_interface': 7.522557075632078e-07, 'learning_rate/action_model': 7.522557075632078e-06, 'epoch': 2.85}) +[2026-06-18 02:21:16,057][starVLA.training.train_starvla][INFO] - Step 6282, Loss: {'train/loss': 0.2196410447359085, 'train/grad_norm_pre_clip': 25.461402893066406, 'timing/data': 0.0004916833713650703, 'timing/model': 0.934736767783761, 'learning_rate/qwen_vl_interface': 7.515607665711051e-07, 'learning_rate/action_model': 7.515607665711051e-06, 'epoch': 2.85}) +[2026-06-18 02:21:19,822][starVLA.training.train_starvla][INFO] - Step 6283, Loss: {'train/loss': 0.14657793939113617, 'train/grad_norm_pre_clip': 33.510189056396484, 'timing/data': 0.000555010512471199, 'timing/model': 0.9275974649935961, 'learning_rate/qwen_vl_interface': 7.508667581102959e-07, 'learning_rate/action_model': 7.508667581102958e-06, 'epoch': 2.85}) +[2026-06-18 02:21:23,792][starVLA.training.train_starvla][INFO] - Step 6284, Loss: {'train/loss': 0.026833947747945786, 'train/grad_norm_pre_clip': 8.314535140991211, 'timing/data': 0.0005169594660401344, 'timing/model': 1.1009797537699342, 'learning_rate/qwen_vl_interface': 7.50173682324648e-07, 'learning_rate/action_model': 7.501736823246481e-06, 'epoch': 2.85}) +[2026-06-18 02:21:27,618][starVLA.training.train_starvla][INFO] - Step 6285, Loss: {'train/loss': 0.02569897659122944, 'train/grad_norm_pre_clip': 8.329925537109375, 'timing/data': 0.0009194687008857727, 'timing/model': 1.1420398345217109, 'learning_rate/qwen_vl_interface': 7.494815393578394e-07, 'learning_rate/action_model': 7.494815393578393e-06, 'epoch': 2.85}) +[2026-06-18 02:21:31,225][starVLA.training.train_starvla][INFO] - Step 6286, Loss: {'train/loss': 0.09572559595108032, 'train/grad_norm_pre_clip': 30.125017166137695, 'timing/data': 0.0007375581189990044, 'timing/model': 0.9351944178342819, 'learning_rate/qwen_vl_interface': 7.487903293533501e-07, 'learning_rate/action_model': 7.4879032935335e-06, 'epoch': 2.85}) +[2026-06-18 02:21:35,525][starVLA.training.train_starvla][INFO] - Step 6287, Loss: {'train/loss': 0.18263167142868042, 'train/grad_norm_pre_clip': 22.157197952270508, 'timing/data': 0.0014032190665602684, 'timing/model': 0.9223406622186303, 'learning_rate/qwen_vl_interface': 7.481000524544686e-07, 'learning_rate/action_model': 7.481000524544686e-06, 'epoch': 2.85}) +[2026-06-18 02:21:39,303][starVLA.training.train_starvla][INFO] - Step 6288, Loss: {'train/loss': 0.5052634477615356, 'train/grad_norm_pre_clip': 48.649574279785156, 'timing/data': 0.0006239358335733414, 'timing/model': 0.9364680638536811, 'learning_rate/qwen_vl_interface': 7.474107088042898e-07, 'learning_rate/action_model': 7.474107088042898e-06, 'epoch': 2.85}) +[2026-06-18 02:21:43,119][starVLA.training.train_starvla][INFO] - Step 6289, Loss: {'train/loss': 0.03152196854352951, 'train/grad_norm_pre_clip': 10.6754789352417, 'timing/data': 0.0004943152889609337, 'timing/model': 0.9509479338303208, 'learning_rate/qwen_vl_interface': 7.467222985457167e-07, 'learning_rate/action_model': 7.4672229854571675e-06, 'epoch': 2.85}) +[2026-06-18 02:21:46,927][starVLA.training.train_starvla][INFO] - Step 6290, Loss: {'train/loss': 0.3060321509838104, 'train/grad_norm_pre_clip': 31.8487491607666, 'timing/data': 0.0015099141746759415, 'timing/model': 0.9317545248195529, 'learning_rate/qwen_vl_interface': 7.46034821821458e-07, 'learning_rate/action_model': 7.46034821821458e-06, 'epoch': 2.85}) +[2026-06-18 02:21:50,706][starVLA.training.train_starvla][INFO] - Step 6291, Loss: {'train/loss': 0.07790671288967133, 'train/grad_norm_pre_clip': 20.135408401489258, 'timing/data': 0.00048796553164720535, 'timing/model': 1.1186620350927114, 'learning_rate/qwen_vl_interface': 7.453482787740255e-07, 'learning_rate/action_model': 7.4534827877402545e-06, 'epoch': 2.85}) +[2026-06-18 02:21:54,721][starVLA.training.train_starvla][INFO] - Step 6292, Loss: {'train/loss': 0.25416746735572815, 'train/grad_norm_pre_clip': 22.80702018737793, 'timing/data': 0.0012410357594490051, 'timing/model': 1.1530561354011297, 'learning_rate/qwen_vl_interface': 7.446626695457427e-07, 'learning_rate/action_model': 7.446626695457427e-06, 'epoch': 2.85}) +[2026-06-18 02:21:58,582][starVLA.training.train_starvla][INFO] - Step 6293, Loss: {'train/loss': 0.4892638027667999, 'train/grad_norm_pre_clip': 29.13564109802246, 'timing/data': 0.00046016275882720947, 'timing/model': 0.9307935731485486, 'learning_rate/qwen_vl_interface': 7.439779942787367e-07, 'learning_rate/action_model': 7.439779942787367e-06, 'epoch': 2.85}) +[2026-06-18 02:22:02,338][starVLA.training.train_starvla][INFO] - Step 6294, Loss: {'train/loss': 0.09074555337429047, 'train/grad_norm_pre_clip': 21.239421844482422, 'timing/data': 0.0013453979045152664, 'timing/model': 0.9347682353109121, 'learning_rate/qwen_vl_interface': 7.432942531149409e-07, 'learning_rate/action_model': 7.432942531149409e-06, 'epoch': 2.86}) +[2026-06-18 02:22:06,112][starVLA.training.train_starvla][INFO] - Step 6295, Loss: {'train/loss': 0.028268013149499893, 'train/grad_norm_pre_clip': 7.93780517578125, 'timing/data': 0.0013781478628516197, 'timing/model': 0.936935686506331, 'learning_rate/qwen_vl_interface': 7.42611446196096e-07, 'learning_rate/action_model': 7.426114461960959e-06, 'epoch': 2.86}) +[2026-06-18 02:22:09,954][starVLA.training.train_starvla][INFO] - Step 6296, Loss: {'train/loss': 0.06883065402507782, 'train/grad_norm_pre_clip': 23.868221282958984, 'timing/data': 0.0006851786747574806, 'timing/model': 0.9611216001212597, 'learning_rate/qwen_vl_interface': 7.419295736637483e-07, 'learning_rate/action_model': 7.419295736637483e-06, 'epoch': 2.86}) +[2026-06-18 02:22:13,738][starVLA.training.train_starvla][INFO] - Step 6297, Loss: {'train/loss': 0.18703724443912506, 'train/grad_norm_pre_clip': 38.672847747802734, 'timing/data': 0.0004944121465086937, 'timing/model': 1.118653504177928, 'learning_rate/qwen_vl_interface': 7.412486356592512e-07, 'learning_rate/action_model': 7.412486356592511e-06, 'epoch': 2.86}) +[2026-06-18 02:22:17,543][starVLA.training.train_starvla][INFO] - Step 6298, Loss: {'train/loss': 0.043967727571725845, 'train/grad_norm_pre_clip': 10.449198722839355, 'timing/data': 0.0005460148677229881, 'timing/model': 1.1003306154161692, 'learning_rate/qwen_vl_interface': 7.405686323237623e-07, 'learning_rate/action_model': 7.405686323237623e-06, 'epoch': 2.86}) +[2026-06-18 02:22:21,200][starVLA.training.train_starvla][INFO] - Step 6299, Loss: {'train/loss': 0.020330986008048058, 'train/grad_norm_pre_clip': 11.500624656677246, 'timing/data': 0.0005522510036826134, 'timing/model': 0.939876570366323, 'learning_rate/qwen_vl_interface': 7.398895637982504e-07, 'learning_rate/action_model': 7.398895637982503e-06, 'epoch': 2.86}) +[2026-06-18 02:24:30,883][starVLA.training.train_starvla][INFO] - Step 6300, Loss: {'train/loss': 0.2404252588748932, 'train/grad_norm_pre_clip': 29.645931243896484, 'eval/loss': 0.26119670271873474, 'eval/latency_8/loss': 0.26119670271873474, 'eval/demon_fire_precision': 0.971156796563363, 'eval/demon_fire_recall': 0.983224603914259, 'eval/demon_fire_f1': 0.9771534424205002, 'eval/demon_fire_rate_model': 0.9067890929326655, 'eval/demon_fire_rate_teacher': 0.8956594323873122, 'eval/demon_left_precision': 0.927860696517413, 'eval/demon_left_recall': 0.7987152034261242, 'eval/demon_left_f1': 0.8584579976985041, 'eval/demon_left_rate_model': 0.11185308848080133, 'eval/demon_left_rate_teacher': 0.12993878686700056, 'eval/demon_right_precision': 0.8725038402457758, 'eval/demon_right_recall': 0.8606060606060606, 'eval/demon_right_f1': 0.8665141113653698, 'eval/demon_right_rate_model': 0.18113522537562604, 'eval/demon_right_rate_teacher': 0.18363939899833054, 'eval/demon_move_f1': 0.8624860545319369, 'eval/demon_cc_f1': 0.9396745658524289, 'eval/latency_8/demon_fire_precision': 0.971156796563363, 'eval/latency_8/demon_fire_recall': 0.983224603914259, 'eval/latency_8/demon_fire_f1': 0.9771534424205002, 'eval/latency_8/demon_fire_rate_model': 0.9067890929326655, 'eval/latency_8/demon_fire_rate_teacher': 0.8956594323873122, 'eval/latency_8/demon_left_precision': 0.927860696517413, 'eval/latency_8/demon_left_recall': 0.7987152034261242, 'eval/latency_8/demon_left_f1': 0.8584579976985041, 'eval/latency_8/demon_left_rate_model': 0.11185308848080133, 'eval/latency_8/demon_left_rate_teacher': 0.12993878686700056, 'eval/latency_8/demon_right_precision': 0.8725038402457758, 'eval/latency_8/demon_right_recall': 0.8606060606060606, 'eval/latency_8/demon_right_f1': 0.8665141113653698, 'eval/latency_8/demon_right_rate_model': 0.18113522537562604, 'eval/latency_8/demon_right_rate_teacher': 0.18363939899833054, 'eval/latency_8/demon_move_f1': 0.8624860545319369, 'eval/latency_8/demon_cc_f1': 0.9396745658524289, 'timing/data': 0.0004913406446576118, 'timing/model': 0.9570347415283322, 'learning_rate/qwen_vl_interface': 7.392114302234838e-07, 'learning_rate/action_model': 7.3921143022348385e-06, 'epoch': 2.86}) +[2026-06-18 02:24:30,889][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_6300_state +[2026-06-18 02:24:30,895][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 02:24:37,538][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_6300_state/model.safetensors +[2026-06-18 02:24:51,869][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_6300_state/optimizer.bin +[2026-06-18 02:24:51,870][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_6300_state/sampler.bin +[2026-06-18 02:24:51,870][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_6300_state/sampler_1.bin +[2026-06-18 02:24:51,874][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_6300_state/random_states_0.pkl +[2026-06-18 02:24:51,877][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 6300 +[2026-06-18 02:24:51,877][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 02:24:51,890][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 02:24:55,624][starVLA.training.train_starvla][INFO] - Step 6301, Loss: {'train/loss': 0.11806657910346985, 'train/grad_norm_pre_clip': 19.62353515625, 'timing/data': 0.0009816419333219528, 'timing/model': 0.9294726569205523, 'learning_rate/qwen_vl_interface': 7.385342317400418e-07, 'learning_rate/action_model': 7.385342317400417e-06, 'epoch': 2.86}) +[2026-06-18 02:25:01,219][starVLA.training.train_starvla][INFO] - Step 6302, Loss: {'train/loss': 0.13524065911769867, 'train/grad_norm_pre_clip': 28.373525619506836, 'timing/data': 0.000545150600373745, 'timing/model': 2.7471042284741998, 'learning_rate/qwen_vl_interface': 7.378579684883069e-07, 'learning_rate/action_model': 7.378579684883069e-06, 'epoch': 2.86}) +[2026-06-18 02:25:08,399][starVLA.training.train_starvla][INFO] - Step 6303, Loss: {'train/loss': 0.26584509015083313, 'train/grad_norm_pre_clip': 43.21946716308594, 'timing/data': 0.0005971826612949371, 'timing/model': 2.8506427695974708, 'learning_rate/qwen_vl_interface': 7.371826406084714e-07, 'learning_rate/action_model': 7.371826406084714e-06, 'epoch': 2.86}) +[2026-06-18 02:25:12,268][starVLA.training.train_starvla][INFO] - Step 6304, Loss: {'train/loss': 0.2371472269296646, 'train/grad_norm_pre_clip': 24.17222023010254, 'timing/data': 0.0009139981120824814, 'timing/model': 0.9273203397169709, 'learning_rate/qwen_vl_interface': 7.365082482405315e-07, 'learning_rate/action_model': 7.365082482405315e-06, 'epoch': 2.86}) +[2026-06-18 02:25:16,125][starVLA.training.train_starvla][INFO] - Step 6305, Loss: {'train/loss': 0.13380253314971924, 'train/grad_norm_pre_clip': 32.86995315551758, 'timing/data': 0.0005663968622684479, 'timing/model': 0.9490162059664726, 'learning_rate/qwen_vl_interface': 7.358347915242877e-07, 'learning_rate/action_model': 7.358347915242876e-06, 'epoch': 2.86}) +[2026-06-18 02:25:19,898][starVLA.training.train_starvla][INFO] - Step 6306, Loss: {'train/loss': 0.13880838453769684, 'train/grad_norm_pre_clip': 31.241273880004883, 'timing/data': 0.0005856938660144806, 'timing/model': 0.9345623049885035, 'learning_rate/qwen_vl_interface': 7.351622705993481e-07, 'learning_rate/action_model': 7.351622705993481e-06, 'epoch': 2.86}) +[2026-06-18 02:25:23,673][starVLA.training.train_starvla][INFO] - Step 6307, Loss: {'train/loss': 0.14992278814315796, 'train/grad_norm_pre_clip': 28.099040985107422, 'timing/data': 0.0005009043961763382, 'timing/model': 1.1047317627817392, 'learning_rate/qwen_vl_interface': 7.344906856051288e-07, 'learning_rate/action_model': 7.344906856051288e-06, 'epoch': 2.86}) +[2026-06-18 02:25:27,232][starVLA.training.train_starvla][INFO] - Step 6308, Loss: {'train/loss': 0.3338073492050171, 'train/grad_norm_pre_clip': 53.733009338378906, 'timing/data': 0.0005103778094053268, 'timing/model': 0.9197914823889732, 'learning_rate/qwen_vl_interface': 7.338200366808493e-07, 'learning_rate/action_model': 7.338200366808492e-06, 'epoch': 2.86}) +[2026-06-18 02:25:30,982][starVLA.training.train_starvla][INFO] - Step 6309, Loss: {'train/loss': 0.36086684465408325, 'train/grad_norm_pre_clip': 30.772579193115234, 'timing/data': 0.0004883389919996262, 'timing/model': 0.9230636795982718, 'learning_rate/qwen_vl_interface': 7.331503239655361e-07, 'learning_rate/action_model': 7.33150323965536e-06, 'epoch': 2.86}) +[2026-06-18 02:25:34,717][starVLA.training.train_starvla][INFO] - Step 6310, Loss: {'train/loss': 0.342736154794693, 'train/grad_norm_pre_clip': 29.117401123046875, 'timing/data': 0.000501479022204876, 'timing/model': 0.9465383300557733, 'learning_rate/qwen_vl_interface': 7.324815475980207e-07, 'learning_rate/action_model': 7.324815475980207e-06, 'epoch': 2.86}) +[2026-06-18 02:25:38,523][starVLA.training.train_starvla][INFO] - Step 6311, Loss: {'train/loss': 0.15405917167663574, 'train/grad_norm_pre_clip': 23.289077758789062, 'timing/data': 0.0005556419491767883, 'timing/model': 0.9550279453396797, 'learning_rate/qwen_vl_interface': 7.318137077169416e-07, 'learning_rate/action_model': 7.318137077169415e-06, 'epoch': 2.86}) +[2026-06-18 02:25:42,375][starVLA.training.train_starvla][INFO] - Step 6312, Loss: {'train/loss': 0.1019434928894043, 'train/grad_norm_pre_clip': 18.071109771728516, 'timing/data': 0.0005552340298891068, 'timing/model': 1.0216659866273403, 'learning_rate/qwen_vl_interface': 7.311468044607425e-07, 'learning_rate/action_model': 7.311468044607425e-06, 'epoch': 2.86}) +[2026-06-18 02:25:46,241][starVLA.training.train_starvla][INFO] - Step 6313, Loss: {'train/loss': 0.015836017206311226, 'train/grad_norm_pre_clip': 3.1831178665161133, 'timing/data': 0.00047540199011564255, 'timing/model': 1.0092703253030777, 'learning_rate/qwen_vl_interface': 7.304808379676739e-07, 'learning_rate/action_model': 7.3048083796767394e-06, 'epoch': 2.86}) +[2026-06-18 02:25:50,236][starVLA.training.train_starvla][INFO] - Step 6314, Loss: {'train/loss': 0.3310661315917969, 'train/grad_norm_pre_clip': 34.435211181640625, 'timing/data': 0.0012743854895234108, 'timing/model': 1.0623017270117998, 'learning_rate/qwen_vl_interface': 7.298158083757903e-07, 'learning_rate/action_model': 7.298158083757903e-06, 'epoch': 2.86}) +[2026-06-18 02:25:53,982][starVLA.training.train_starvla][INFO] - Step 6315, Loss: {'train/loss': 0.4240153729915619, 'train/grad_norm_pre_clip': 41.804115295410156, 'timing/data': 0.0004932116717100143, 'timing/model': 0.9211112940683961, 'learning_rate/qwen_vl_interface': 7.291517158229536e-07, 'learning_rate/action_model': 7.291517158229535e-06, 'epoch': 2.86}) +[2026-06-18 02:25:57,834][starVLA.training.train_starvla][INFO] - Step 6316, Loss: {'train/loss': 0.10745227336883545, 'train/grad_norm_pre_clip': 28.873308181762695, 'timing/data': 0.0005908133462071419, 'timing/model': 0.9441479202359915, 'learning_rate/qwen_vl_interface': 7.284885604468312e-07, 'learning_rate/action_model': 7.284885604468311e-06, 'epoch': 2.87}) +[2026-06-18 02:26:01,632][starVLA.training.train_starvla][INFO] - Step 6317, Loss: {'train/loss': 0.36987370252609253, 'train/grad_norm_pre_clip': 28.27171516418457, 'timing/data': 0.0007487237453460693, 'timing/model': 0.9442755337804556, 'learning_rate/qwen_vl_interface': 7.278263423848948e-07, 'learning_rate/action_model': 7.278263423848948e-06, 'epoch': 2.87}) +[2026-06-18 02:26:05,837][starVLA.training.train_starvla][INFO] - Step 6318, Loss: {'train/loss': 0.47457966208457947, 'train/grad_norm_pre_clip': 75.71126556396484, 'timing/data': 0.0005449159070849419, 'timing/model': 1.1591127282008529, 'learning_rate/qwen_vl_interface': 7.271650617744239e-07, 'learning_rate/action_model': 7.271650617744238e-06, 'epoch': 2.87}) +[2026-06-18 02:26:09,664][starVLA.training.train_starvla][INFO] - Step 6319, Loss: {'train/loss': 0.17171774804592133, 'train/grad_norm_pre_clip': 23.483182907104492, 'timing/data': 0.00048863235861063, 'timing/model': 0.938110738992691, 'learning_rate/qwen_vl_interface': 7.265047187525033e-07, 'learning_rate/action_model': 7.265047187525033e-06, 'epoch': 2.87}) +[2026-06-18 02:26:13,562][starVLA.training.train_starvla][INFO] - Step 6320, Loss: {'train/loss': 0.013645854778587818, 'train/grad_norm_pre_clip': 2.9308438301086426, 'timing/data': 0.0005137091502547264, 'timing/model': 1.1281416630372405, 'learning_rate/qwen_vl_interface': 7.258453134560212e-07, 'learning_rate/action_model': 7.258453134560211e-06, 'epoch': 2.87}) +[2026-06-18 02:26:17,264][starVLA.training.train_starvla][INFO] - Step 6321, Loss: {'train/loss': 0.08435449004173279, 'train/grad_norm_pre_clip': 20.578393936157227, 'timing/data': 0.0005590561777353287, 'timing/model': 1.0550592551007867, 'learning_rate/qwen_vl_interface': 7.251868460216743e-07, 'learning_rate/action_model': 7.251868460216743e-06, 'epoch': 2.87}) +[2026-06-18 02:26:20,849][starVLA.training.train_starvla][INFO] - Step 6322, Loss: {'train/loss': 0.16598933935165405, 'train/grad_norm_pre_clip': 30.676557540893555, 'timing/data': 0.0008141221478581429, 'timing/model': 0.9264709819108248, 'learning_rate/qwen_vl_interface': 7.245293165859629e-07, 'learning_rate/action_model': 7.245293165859629e-06, 'epoch': 2.87}) +[2026-06-18 02:26:24,658][starVLA.training.train_starvla][INFO] - Step 6323, Loss: {'train/loss': 0.18000292778015137, 'train/grad_norm_pre_clip': 23.933032989501953, 'timing/data': 0.0009055910632014275, 'timing/model': 0.929658267647028, 'learning_rate/qwen_vl_interface': 7.238727252851935e-07, 'learning_rate/action_model': 7.238727252851935e-06, 'epoch': 2.87}) +[2026-06-18 02:26:28,462][starVLA.training.train_starvla][INFO] - Step 6324, Loss: {'train/loss': 0.3083498179912567, 'train/grad_norm_pre_clip': 58.33384323120117, 'timing/data': 0.0005870135501027107, 'timing/model': 0.930898672901094, 'learning_rate/qwen_vl_interface': 7.23217072255479e-07, 'learning_rate/action_model': 7.23217072255479e-06, 'epoch': 2.87}) +[2026-06-18 02:26:32,158][starVLA.training.train_starvla][INFO] - Step 6325, Loss: {'train/loss': 0.2723146677017212, 'train/grad_norm_pre_clip': 26.54205322265625, 'timing/data': 0.0005082497373223305, 'timing/model': 0.9252927452325821, 'learning_rate/qwen_vl_interface': 7.225623576327361e-07, 'learning_rate/action_model': 7.22562357632736e-06, 'epoch': 2.87}) +[2026-06-18 02:26:35,908][starVLA.training.train_starvla][INFO] - Step 6326, Loss: {'train/loss': 0.20999735593795776, 'train/grad_norm_pre_clip': 84.85369873046875, 'timing/data': 0.00048539135605096817, 'timing/model': 0.9330773456022143, 'learning_rate/qwen_vl_interface': 7.219085815526887e-07, 'learning_rate/action_model': 7.219085815526887e-06, 'epoch': 2.87}) +[2026-06-18 02:26:39,792][starVLA.training.train_starvla][INFO] - Step 6327, Loss: {'train/loss': 0.2009146362543106, 'train/grad_norm_pre_clip': 28.714387893676758, 'timing/data': 0.0005404846742749214, 'timing/model': 1.169815151952207, 'learning_rate/qwen_vl_interface': 7.212557441508641e-07, 'learning_rate/action_model': 7.212557441508641e-06, 'epoch': 2.87}) +[2026-06-18 02:26:43,438][starVLA.training.train_starvla][INFO] - Step 6328, Loss: {'train/loss': 0.019904781132936478, 'train/grad_norm_pre_clip': 7.000481605529785, 'timing/data': 0.0007506683468818665, 'timing/model': 0.9589370107278228, 'learning_rate/qwen_vl_interface': 7.206038455625972e-07, 'learning_rate/action_model': 7.206038455625971e-06, 'epoch': 2.87}) +[2026-06-18 02:26:47,246][starVLA.training.train_starvla][INFO] - Step 6329, Loss: {'train/loss': 0.1689198911190033, 'train/grad_norm_pre_clip': 38.26572799682617, 'timing/data': 0.0005029104650020599, 'timing/model': 0.9316246621310711, 'learning_rate/qwen_vl_interface': 7.199528859230271e-07, 'learning_rate/action_model': 7.199528859230272e-06, 'epoch': 2.87}) +[2026-06-18 02:26:50,955][starVLA.training.train_starvla][INFO] - Step 6330, Loss: {'train/loss': 0.18100935220718384, 'train/grad_norm_pre_clip': 28.063535690307617, 'timing/data': 0.0005334680899977684, 'timing/model': 0.9270177921280265, 'learning_rate/qwen_vl_interface': 7.19302865367098e-07, 'learning_rate/action_model': 7.19302865367098e-06, 'epoch': 2.87}) +[2026-06-18 02:26:54,714][starVLA.training.train_starvla][INFO] - Step 6331, Loss: {'train/loss': 0.171888530254364, 'train/grad_norm_pre_clip': 28.816333770751953, 'timing/data': 0.000863262452185154, 'timing/model': 0.9210813986137509, 'learning_rate/qwen_vl_interface': 7.186537840295599e-07, 'learning_rate/action_model': 7.186537840295598e-06, 'epoch': 2.87}) +[2026-06-18 02:26:58,529][starVLA.training.train_starvla][INFO] - Step 6332, Loss: {'train/loss': 0.49291089177131653, 'train/grad_norm_pre_clip': 47.742149353027344, 'timing/data': 0.0005423100665211678, 'timing/model': 0.9383101314306259, 'learning_rate/qwen_vl_interface': 7.18005642044969e-07, 'learning_rate/action_model': 7.180056420449689e-06, 'epoch': 2.87}) +[2026-06-18 02:27:02,410][starVLA.training.train_starvla][INFO] - Step 6333, Loss: {'train/loss': 0.04187818244099617, 'train/grad_norm_pre_clip': 11.215426445007324, 'timing/data': 0.0007044840604066849, 'timing/model': 1.1580848824232817, 'learning_rate/qwen_vl_interface': 7.173584395476852e-07, 'learning_rate/action_model': 7.173584395476851e-06, 'epoch': 2.87}) +[2026-06-18 02:27:06,300][starVLA.training.train_starvla][INFO] - Step 6334, Loss: {'train/loss': 0.2043064534664154, 'train/grad_norm_pre_clip': 27.003969192504883, 'timing/data': 0.0008113076910376549, 'timing/model': 1.1758232908323407, 'learning_rate/qwen_vl_interface': 7.16712176671874e-07, 'learning_rate/action_model': 7.16712176671874e-06, 'epoch': 2.87}) +[2026-06-18 02:27:10,112][starVLA.training.train_starvla][INFO] - Step 6335, Loss: {'train/loss': 0.23975658416748047, 'train/grad_norm_pre_clip': 30.19744300842285, 'timing/data': 0.0011681197211146355, 'timing/model': 0.9263783469796181, 'learning_rate/qwen_vl_interface': 7.160668535515059e-07, 'learning_rate/action_model': 7.1606685355150585e-06, 'epoch': 2.87}) +[2026-06-18 02:27:13,977][starVLA.training.train_starvla][INFO] - Step 6336, Loss: {'train/loss': 0.16667377948760986, 'train/grad_norm_pre_clip': 32.39094924926758, 'timing/data': 0.0005079815164208412, 'timing/model': 0.9548388365656137, 'learning_rate/qwen_vl_interface': 7.154224703203584e-07, 'learning_rate/action_model': 7.154224703203584e-06, 'epoch': 2.87}) +[2026-06-18 02:27:17,799][starVLA.training.train_starvla][INFO] - Step 6337, Loss: {'train/loss': 0.2054622620344162, 'train/grad_norm_pre_clip': 23.59893035888672, 'timing/data': 0.0004885876551270485, 'timing/model': 0.9409607229754329, 'learning_rate/qwen_vl_interface': 7.147790271120122e-07, 'learning_rate/action_model': 7.147790271120122e-06, 'epoch': 2.87}) +[2026-06-18 02:27:21,586][starVLA.training.train_starvla][INFO] - Step 6338, Loss: {'train/loss': 0.38803982734680176, 'train/grad_norm_pre_clip': 28.873979568481445, 'timing/data': 0.0006098542362451553, 'timing/model': 0.9213490327820182, 'learning_rate/qwen_vl_interface': 7.141365240598545e-07, 'learning_rate/action_model': 7.141365240598545e-06, 'epoch': 2.88}) +[2026-06-18 02:27:25,519][starVLA.training.train_starvla][INFO] - Step 6339, Loss: {'train/loss': 0.08366265147924423, 'train/grad_norm_pre_clip': 17.625146865844727, 'timing/data': 0.0005880445241928101, 'timing/model': 0.99697155226022, 'learning_rate/qwen_vl_interface': 7.134949612970749e-07, 'learning_rate/action_model': 7.134949612970748e-06, 'epoch': 2.88}) +[2026-06-18 02:27:29,481][starVLA.training.train_starvla][INFO] - Step 6340, Loss: {'train/loss': 0.20108595490455627, 'train/grad_norm_pre_clip': 20.573001861572266, 'timing/data': 0.0006983466446399689, 'timing/model': 1.272516862489283, 'learning_rate/qwen_vl_interface': 7.128543389566718e-07, 'learning_rate/action_model': 7.128543389566718e-06, 'epoch': 2.88}) +[2026-06-18 02:27:33,294][starVLA.training.train_starvla][INFO] - Step 6341, Loss: {'train/loss': 0.6201779842376709, 'train/grad_norm_pre_clip': 46.912261962890625, 'timing/data': 0.0005142837762832642, 'timing/model': 1.1652775472030044, 'learning_rate/qwen_vl_interface': 7.122146571714464e-07, 'learning_rate/action_model': 7.122146571714463e-06, 'epoch': 2.88}) +[2026-06-18 02:27:36,858][starVLA.training.train_starvla][INFO] - Step 6342, Loss: {'train/loss': 0.06419798731803894, 'train/grad_norm_pre_clip': 14.062822341918945, 'timing/data': 0.00048150401562452316, 'timing/model': 0.9229267956689, 'learning_rate/qwen_vl_interface': 7.115759160740058e-07, 'learning_rate/action_model': 7.115759160740057e-06, 'epoch': 2.88}) +[2026-06-18 02:27:40,911][starVLA.training.train_starvla][INFO] - Step 6343, Loss: {'train/loss': 0.1656334400177002, 'train/grad_norm_pre_clip': 23.634607315063477, 'timing/data': 0.0009558815509080887, 'timing/model': 1.2035082895308733, 'learning_rate/qwen_vl_interface': 7.109381157967607e-07, 'learning_rate/action_model': 7.109381157967607e-06, 'epoch': 2.88}) +[2026-06-18 02:27:44,678][starVLA.training.train_starvla][INFO] - Step 6344, Loss: {'train/loss': 0.054150182753801346, 'train/grad_norm_pre_clip': 18.536136627197266, 'timing/data': 0.0004960205405950546, 'timing/model': 0.9229660117998719, 'learning_rate/qwen_vl_interface': 7.103012564719289e-07, 'learning_rate/action_model': 7.103012564719288e-06, 'epoch': 2.88}) +[2026-06-18 02:27:48,411][starVLA.training.train_starvla][INFO] - Step 6345, Loss: {'train/loss': 0.15174086391925812, 'train/grad_norm_pre_clip': 50.42109298706055, 'timing/data': 0.001137775368988514, 'timing/model': 0.9271413553506136, 'learning_rate/qwen_vl_interface': 7.096653382315315e-07, 'learning_rate/action_model': 7.096653382315315e-06, 'epoch': 2.88}) +[2026-06-18 02:27:52,247][starVLA.training.train_starvla][INFO] - Step 6346, Loss: {'train/loss': 0.2885478436946869, 'train/grad_norm_pre_clip': 158.82382202148438, 'timing/data': 0.0005202749744057655, 'timing/model': 0.9247724637389183, 'learning_rate/qwen_vl_interface': 7.090303612073946e-07, 'learning_rate/action_model': 7.090303612073946e-06, 'epoch': 2.88}) +[2026-06-18 02:27:56,019][starVLA.training.train_starvla][INFO] - Step 6347, Loss: {'train/loss': 0.299667626619339, 'train/grad_norm_pre_clip': 39.53629684448242, 'timing/data': 0.0004895692691206932, 'timing/model': 1.119583873078227, 'learning_rate/qwen_vl_interface': 7.083963255311508e-07, 'learning_rate/action_model': 7.083963255311509e-06, 'epoch': 2.88}) +[2026-06-18 02:27:59,599][starVLA.training.train_starvla][INFO] - Step 6348, Loss: {'train/loss': 0.12857186794281006, 'train/grad_norm_pre_clip': 23.389863967895508, 'timing/data': 0.0005065705627202988, 'timing/model': 0.9323552660644054, 'learning_rate/qwen_vl_interface': 7.07763231334236e-07, 'learning_rate/action_model': 7.077632313342359e-06, 'epoch': 2.88}) +[2026-06-18 02:28:03,611][starVLA.training.train_starvla][INFO] - Step 6349, Loss: {'train/loss': 0.2823329269886017, 'train/grad_norm_pre_clip': 38.52109146118164, 'timing/data': 0.0005344776436686516, 'timing/model': 0.9308877494186163, 'learning_rate/qwen_vl_interface': 7.071310787478902e-07, 'learning_rate/action_model': 7.071310787478902e-06, 'epoch': 2.88}) +[2026-06-18 02:28:07,352][starVLA.training.train_starvla][INFO] - Step 6350, Loss: {'train/loss': 0.1325356513261795, 'train/grad_norm_pre_clip': 24.372392654418945, 'timing/data': 0.0005107130855321884, 'timing/model': 0.9230320984497666, 'learning_rate/qwen_vl_interface': 7.0649986790316e-07, 'learning_rate/action_model': 7.0649986790315995e-06, 'epoch': 2.88}) +[2026-06-18 02:28:11,121][starVLA.training.train_starvla][INFO] - Step 6351, Loss: {'train/loss': 0.3174523711204529, 'train/grad_norm_pre_clip': 50.15728759765625, 'timing/data': 0.0004907529801130295, 'timing/model': 0.9333767602220178, 'learning_rate/qwen_vl_interface': 7.058695989308965e-07, 'learning_rate/action_model': 7.058695989308965e-06, 'epoch': 2.88}) +[2026-06-18 02:28:14,897][starVLA.training.train_starvla][INFO] - Step 6352, Loss: {'train/loss': 0.19880472123622894, 'train/grad_norm_pre_clip': 39.06153106689453, 'timing/data': 0.0005233753472566605, 'timing/model': 0.920798422768712, 'learning_rate/qwen_vl_interface': 7.052402719617554e-07, 'learning_rate/action_model': 7.052402719617554e-06, 'epoch': 2.88}) +[2026-06-18 02:28:18,676][starVLA.training.train_starvla][INFO] - Step 6353, Loss: {'train/loss': 0.21387295424938202, 'train/grad_norm_pre_clip': 34.8759880065918, 'timing/data': 0.00048515573143959045, 'timing/model': 0.9212352354079485, 'learning_rate/qwen_vl_interface': 7.046118871261952e-07, 'learning_rate/action_model': 7.0461188712619514e-06, 'epoch': 2.88}) +[2026-06-18 02:28:22,442][starVLA.training.train_starvla][INFO] - Step 6354, Loss: {'train/loss': 0.04028068482875824, 'train/grad_norm_pre_clip': 16.580310821533203, 'timing/data': 0.0010780822485685349, 'timing/model': 1.1058807587251067, 'learning_rate/qwen_vl_interface': 7.039844445544828e-07, 'learning_rate/action_model': 7.039844445544827e-06, 'epoch': 2.88}) +[2026-06-18 02:28:26,056][starVLA.training.train_starvla][INFO] - Step 6355, Loss: {'train/loss': 0.07702922821044922, 'train/grad_norm_pre_clip': 16.22728157043457, 'timing/data': 0.000496329739689827, 'timing/model': 0.9209850449115038, 'learning_rate/qwen_vl_interface': 7.033579443766864e-07, 'learning_rate/action_model': 7.033579443766864e-06, 'epoch': 2.88}) +[2026-06-18 02:28:29,821][starVLA.training.train_starvla][INFO] - Step 6356, Loss: {'train/loss': 0.14997261762619019, 'train/grad_norm_pre_clip': 40.935821533203125, 'timing/data': 0.0004947343841195107, 'timing/model': 0.917509357444942, 'learning_rate/qwen_vl_interface': 7.027323867226807e-07, 'learning_rate/action_model': 7.0273238672268065e-06, 'epoch': 2.88}) +[2026-06-18 02:28:33,761][starVLA.training.train_starvla][INFO] - Step 6357, Loss: {'train/loss': 0.18445876240730286, 'train/grad_norm_pre_clip': 31.536165237426758, 'timing/data': 0.0005516456440091133, 'timing/model': 0.93705203384161, 'learning_rate/qwen_vl_interface': 7.021077717221441e-07, 'learning_rate/action_model': 7.0210777172214415e-06, 'epoch': 2.88}) +[2026-06-18 02:28:37,588][starVLA.training.train_starvla][INFO] - Step 6358, Loss: {'train/loss': 0.14316542446613312, 'train/grad_norm_pre_clip': 22.182598114013672, 'timing/data': 0.0007643653079867363, 'timing/model': 0.9586012568324804, 'learning_rate/qwen_vl_interface': 7.014840995045602e-07, 'learning_rate/action_model': 7.014840995045602e-06, 'epoch': 2.88}) +[2026-06-18 02:28:41,403][starVLA.training.train_starvla][INFO] - Step 6359, Loss: {'train/loss': 0.21343164145946503, 'train/grad_norm_pre_clip': 25.173839569091797, 'timing/data': 0.00047833751887083054, 'timing/model': 0.9950537364929914, 'learning_rate/qwen_vl_interface': 7.008613701992172e-07, 'learning_rate/action_model': 7.008613701992171e-06, 'epoch': 2.88}) +[2026-06-18 02:28:45,236][starVLA.training.train_starvla][INFO] - Step 6360, Loss: {'train/loss': 0.1362815648317337, 'train/grad_norm_pre_clip': 23.242677688598633, 'timing/data': 0.000500447116792202, 'timing/model': 1.1134726824238896, 'learning_rate/qwen_vl_interface': 7.002395839352063e-07, 'learning_rate/action_model': 7.002395839352063e-06, 'epoch': 2.89}) +[2026-06-18 02:28:49,064][starVLA.training.train_starvla][INFO] - Step 6361, Loss: {'train/loss': 0.18719840049743652, 'train/grad_norm_pre_clip': 36.14310073852539, 'timing/data': 0.0005352441221475601, 'timing/model': 1.1042886897921562, 'learning_rate/qwen_vl_interface': 6.996187408414268e-07, 'learning_rate/action_model': 6.996187408414268e-06, 'epoch': 2.89}) +[2026-06-18 02:28:52,674][starVLA.training.train_starvla][INFO] - Step 6362, Loss: {'train/loss': 0.14586661756038666, 'train/grad_norm_pre_clip': 21.506563186645508, 'timing/data': 0.0004844600334763527, 'timing/model': 0.9260505605489016, 'learning_rate/qwen_vl_interface': 6.98998841046578e-07, 'learning_rate/action_model': 6.989988410465779e-06, 'epoch': 2.89}) +[2026-06-18 02:28:56,438][starVLA.training.train_starvla][INFO] - Step 6363, Loss: {'train/loss': 0.335679829120636, 'train/grad_norm_pre_clip': 46.74583435058594, 'timing/data': 0.0004889657720923424, 'timing/model': 0.9227279210463166, 'learning_rate/qwen_vl_interface': 6.983798846791667e-07, 'learning_rate/action_model': 6.983798846791667e-06, 'epoch': 2.89}) +[2026-06-18 02:29:00,371][starVLA.training.train_starvla][INFO] - Step 6364, Loss: {'train/loss': 0.09828179329633713, 'train/grad_norm_pre_clip': 23.695777893066406, 'timing/data': 0.0009339936077594757, 'timing/model': 0.9205652512609959, 'learning_rate/qwen_vl_interface': 6.977618718675023e-07, 'learning_rate/action_model': 6.977618718675024e-06, 'epoch': 2.89}) +[2026-06-18 02:29:04,196][starVLA.training.train_starvla][INFO] - Step 6365, Loss: {'train/loss': 0.03215105086565018, 'train/grad_norm_pre_clip': 10.685893058776855, 'timing/data': 0.0009466772899031639, 'timing/model': 0.9596008686348796, 'learning_rate/qwen_vl_interface': 6.971448027397007e-07, 'learning_rate/action_model': 6.971448027397007e-06, 'epoch': 2.89}) +[2026-06-18 02:29:08,032][starVLA.training.train_starvla][INFO] - Step 6366, Loss: {'train/loss': 0.06852017343044281, 'train/grad_norm_pre_clip': 16.51285743713379, 'timing/data': 0.0005208225920796394, 'timing/model': 0.9243166064843535, 'learning_rate/qwen_vl_interface': 6.965286774236814e-07, 'learning_rate/action_model': 6.965286774236813e-06, 'epoch': 2.89}) +[2026-06-18 02:29:11,871][starVLA.training.train_starvla][INFO] - Step 6367, Loss: {'train/loss': 0.14759671688079834, 'train/grad_norm_pre_clip': 22.101675033569336, 'timing/data': 0.0005235429853200912, 'timing/model': 1.193687742576003, 'learning_rate/qwen_vl_interface': 6.959134960471664e-07, 'learning_rate/action_model': 6.959134960471664e-06, 'epoch': 2.89}) +[2026-06-18 02:29:15,625][starVLA.training.train_starvla][INFO] - Step 6368, Loss: {'train/loss': 0.07125457376241684, 'train/grad_norm_pre_clip': 21.662067413330078, 'timing/data': 0.0005388474091887474, 'timing/model': 1.1029254691675305, 'learning_rate/qwen_vl_interface': 6.952992587376836e-07, 'learning_rate/action_model': 6.952992587376836e-06, 'epoch': 2.89}) +[2026-06-18 02:29:19,372][starVLA.training.train_starvla][INFO] - Step 6369, Loss: {'train/loss': 0.08346105366945267, 'train/grad_norm_pre_clip': 20.430944442749023, 'timing/data': 0.0005616694688796997, 'timing/model': 0.9200651803985238, 'learning_rate/qwen_vl_interface': 6.946859656225662e-07, 'learning_rate/action_model': 6.946859656225661e-06, 'epoch': 2.89}) +[2026-06-18 02:29:23,697][starVLA.training.train_starvla][INFO] - Step 6370, Loss: {'train/loss': 0.2140120267868042, 'train/grad_norm_pre_clip': 42.62514877319336, 'timing/data': 0.0013927649706602097, 'timing/model': 0.9226806219667196, 'learning_rate/qwen_vl_interface': 6.940736168289497e-07, 'learning_rate/action_model': 6.940736168289497e-06, 'epoch': 2.89}) +[2026-06-18 02:29:28,483][starVLA.training.train_starvla][INFO] - Step 6371, Loss: {'train/loss': 0.3539934456348419, 'train/grad_norm_pre_clip': 51.53495788574219, 'timing/data': 0.0004797978326678276, 'timing/model': 0.9170716926455498, 'learning_rate/qwen_vl_interface': 6.934622124837751e-07, 'learning_rate/action_model': 6.93462212483775e-06, 'epoch': 2.89}) +[2026-06-18 02:29:32,201][starVLA.training.train_starvla][INFO] - Step 6372, Loss: {'train/loss': 0.07905415445566177, 'train/grad_norm_pre_clip': 14.633606910705566, 'timing/data': 0.0009474493563175201, 'timing/model': 0.9374726936221123, 'learning_rate/qwen_vl_interface': 6.928517527137864e-07, 'learning_rate/action_model': 6.9285175271378635e-06, 'epoch': 2.89}) +[2026-06-18 02:29:36,217][starVLA.training.train_starvla][INFO] - Step 6373, Loss: {'train/loss': 0.34033212065696716, 'train/grad_norm_pre_clip': 40.35891342163086, 'timing/data': 0.0004829559475183487, 'timing/model': 0.9576244130730629, 'learning_rate/qwen_vl_interface': 6.922422376455333e-07, 'learning_rate/action_model': 6.9224223764553326e-06, 'epoch': 2.89}) +[2026-06-18 02:29:40,344][starVLA.training.train_starvla][INFO] - Step 6374, Loss: {'train/loss': 0.18315406143665314, 'train/grad_norm_pre_clip': 18.0935115814209, 'timing/data': 0.0007653245702385902, 'timing/model': 1.4293880509212613, 'learning_rate/qwen_vl_interface': 6.916336674053688e-07, 'learning_rate/action_model': 6.9163366740536875e-06, 'epoch': 2.89}) +[2026-06-18 02:29:44,000][starVLA.training.train_starvla][INFO] - Step 6375, Loss: {'train/loss': 0.3212103843688965, 'train/grad_norm_pre_clip': 116.88870239257812, 'timing/data': 0.0009011747315526009, 'timing/model': 0.9388389997184277, 'learning_rate/qwen_vl_interface': 6.9102604211945e-07, 'learning_rate/action_model': 6.910260421194501e-06, 'epoch': 2.89}) +[2026-06-18 02:29:48,180][starVLA.training.train_starvla][INFO] - Step 6376, Loss: {'train/loss': 0.06907975673675537, 'train/grad_norm_pre_clip': 15.172978401184082, 'timing/data': 0.0005072662606835365, 'timing/model': 1.1005211435258389, 'learning_rate/qwen_vl_interface': 6.904193619137381e-07, 'learning_rate/action_model': 6.904193619137381e-06, 'epoch': 2.89}) +[2026-06-18 02:29:52,142][starVLA.training.train_starvla][INFO] - Step 6377, Loss: {'train/loss': 0.05698196962475777, 'train/grad_norm_pre_clip': 22.606809616088867, 'timing/data': 0.0005159396678209305, 'timing/model': 0.9225474260747433, 'learning_rate/qwen_vl_interface': 6.898136269139989e-07, 'learning_rate/action_model': 6.8981362691399885e-06, 'epoch': 2.89}) +[2026-06-18 02:29:55,969][starVLA.training.train_starvla][INFO] - Step 6378, Loss: {'train/loss': 0.19887664914131165, 'train/grad_norm_pre_clip': 21.277551651000977, 'timing/data': 0.0009117927402257919, 'timing/model': 0.9344137283042073, 'learning_rate/qwen_vl_interface': 6.89208837245802e-07, 'learning_rate/action_model': 6.892088372458019e-06, 'epoch': 2.89}) +[2026-06-18 02:29:59,869][starVLA.training.train_starvla][INFO] - Step 6379, Loss: {'train/loss': 0.2105780690908432, 'train/grad_norm_pre_clip': 23.726184844970703, 'timing/data': 0.0010099811479449272, 'timing/model': 0.9786998135969043, 'learning_rate/qwen_vl_interface': 6.886049930345204e-07, 'learning_rate/action_model': 6.8860499303452035e-06, 'epoch': 2.89}) +[2026-06-18 02:30:03,809][starVLA.training.train_starvla][INFO] - Step 6380, Loss: {'train/loss': 0.22299422323703766, 'train/grad_norm_pre_clip': 22.397174835205078, 'timing/data': 0.0006640050560235977, 'timing/model': 1.1250092843547463, 'learning_rate/qwen_vl_interface': 6.880020944053326e-07, 'learning_rate/action_model': 6.880020944053325e-06, 'epoch': 2.89}) +[2026-06-18 02:30:07,628][starVLA.training.train_starvla][INFO] - Step 6381, Loss: {'train/loss': 0.12589724361896515, 'train/grad_norm_pre_clip': 23.06512451171875, 'timing/data': 0.0006150463595986366, 'timing/model': 1.1382898278534412, 'learning_rate/qwen_vl_interface': 6.87400141483219e-07, 'learning_rate/action_model': 6.87400141483219e-06, 'epoch': 2.89}) +[2026-06-18 02:30:11,379][starVLA.training.train_starvla][INFO] - Step 6382, Loss: {'train/loss': 0.016076451167464256, 'train/grad_norm_pre_clip': 3.5862503051757812, 'timing/data': 0.0005151554942131042, 'timing/model': 1.0011497652158141, 'learning_rate/qwen_vl_interface': 6.867991343929655e-07, 'learning_rate/action_model': 6.867991343929656e-06, 'epoch': 2.9}) +[2026-06-18 02:30:15,307][starVLA.training.train_starvla][INFO] - Step 6383, Loss: {'train/loss': 0.1324405074119568, 'train/grad_norm_pre_clip': 21.396026611328125, 'timing/data': 0.00047487858682870865, 'timing/model': 0.9446639064699411, 'learning_rate/qwen_vl_interface': 6.861990732591622e-07, 'learning_rate/action_model': 6.861990732591622e-06, 'epoch': 2.9}) +[2026-06-18 02:30:19,085][starVLA.training.train_starvla][INFO] - Step 6384, Loss: {'train/loss': 0.05122428014874458, 'train/grad_norm_pre_clip': 12.16421890258789, 'timing/data': 0.0004724031314253807, 'timing/model': 0.9255474731326103, 'learning_rate/qwen_vl_interface': 6.855999582062017e-07, 'learning_rate/action_model': 6.855999582062017e-06, 'epoch': 2.9}) +[2026-06-18 02:30:23,114][starVLA.training.train_starvla][INFO] - Step 6385, Loss: {'train/loss': 0.6939824223518372, 'train/grad_norm_pre_clip': 41.83298873901367, 'timing/data': 0.0005010487511754036, 'timing/model': 0.9247007593512535, 'learning_rate/qwen_vl_interface': 6.850017893582818e-07, 'learning_rate/action_model': 6.850017893582818e-06, 'epoch': 2.9}) +[2026-06-18 02:30:26,850][starVLA.training.train_starvla][INFO] - Step 6386, Loss: {'train/loss': 0.12084097415208817, 'train/grad_norm_pre_clip': 27.867626190185547, 'timing/data': 0.00044984370470046997, 'timing/model': 0.9146020645275712, 'learning_rate/qwen_vl_interface': 6.844045668394024e-07, 'learning_rate/action_model': 6.844045668394024e-06, 'epoch': 2.9}) +[2026-06-18 02:30:30,558][starVLA.training.train_starvla][INFO] - Step 6387, Loss: {'train/loss': 0.2628624737262726, 'train/grad_norm_pre_clip': 27.069847106933594, 'timing/data': 0.0004559271037578583, 'timing/model': 1.0715064005926251, 'learning_rate/qwen_vl_interface': 6.838082907733694e-07, 'learning_rate/action_model': 6.838082907733694e-06, 'epoch': 2.9}) +[2026-06-18 02:30:34,300][starVLA.training.train_starvla][INFO] - Step 6388, Loss: {'train/loss': 0.491280734539032, 'train/grad_norm_pre_clip': 50.89372253417969, 'timing/data': 0.0006038863211870193, 'timing/model': 1.0982079971581697, 'learning_rate/qwen_vl_interface': 6.832129612837912e-07, 'learning_rate/action_model': 6.832129612837912e-06, 'epoch': 2.9}) +[2026-06-18 02:30:38,148][starVLA.training.train_starvla][INFO] - Step 6389, Loss: {'train/loss': 0.154927060008049, 'train/grad_norm_pre_clip': 31.412620544433594, 'timing/data': 0.00048781558871269226, 'timing/model': 0.9771413151174784, 'learning_rate/qwen_vl_interface': 6.826185784940801e-07, 'learning_rate/action_model': 6.826185784940801e-06, 'epoch': 2.9}) +[2026-06-18 02:30:41,945][starVLA.training.train_starvla][INFO] - Step 6390, Loss: {'train/loss': 0.3817182779312134, 'train/grad_norm_pre_clip': 35.65686798095703, 'timing/data': 0.0005343416705727577, 'timing/model': 0.9360584039241076, 'learning_rate/qwen_vl_interface': 6.820251425274526e-07, 'learning_rate/action_model': 6.820251425274527e-06, 'epoch': 2.9}) +[2026-06-18 02:30:45,739][starVLA.training.train_starvla][INFO] - Step 6391, Loss: {'train/loss': 0.07901017367839813, 'train/grad_norm_pre_clip': 31.98048973083496, 'timing/data': 0.0005279732868075371, 'timing/model': 0.9659616453573108, 'learning_rate/qwen_vl_interface': 6.81432653506928e-07, 'learning_rate/action_model': 6.814326535069279e-06, 'epoch': 2.9}) +[2026-06-18 02:30:49,525][starVLA.training.train_starvla][INFO] - Step 6392, Loss: {'train/loss': 0.2900189161300659, 'train/grad_norm_pre_clip': 31.41871452331543, 'timing/data': 0.0004789326339960098, 'timing/model': 0.9428900312632322, 'learning_rate/qwen_vl_interface': 6.808411115553305e-07, 'learning_rate/action_model': 6.808411115553305e-06, 'epoch': 2.9}) +[2026-06-18 02:30:53,339][starVLA.training.train_starvla][INFO] - Step 6393, Loss: {'train/loss': 0.15876032412052155, 'train/grad_norm_pre_clip': 24.798343658447266, 'timing/data': 0.0004915026947855949, 'timing/model': 0.9353347076103091, 'learning_rate/qwen_vl_interface': 6.802505167952871e-07, 'learning_rate/action_model': 6.80250516795287e-06, 'epoch': 2.9}) +[2026-06-18 02:30:57,120][starVLA.training.train_starvla][INFO] - Step 6394, Loss: {'train/loss': 0.0305798277258873, 'train/grad_norm_pre_clip': 8.641000747680664, 'timing/data': 0.0005359500646591187, 'timing/model': 1.1093091974034905, 'learning_rate/qwen_vl_interface': 6.796608693492293e-07, 'learning_rate/action_model': 6.7966086934922925e-06, 'epoch': 2.9}) +[2026-06-18 02:31:00,735][starVLA.training.train_starvla][INFO] - Step 6395, Loss: {'train/loss': 0.10143549740314484, 'train/grad_norm_pre_clip': 18.679576873779297, 'timing/data': 0.0007813619449734688, 'timing/model': 0.938792816363275, 'learning_rate/qwen_vl_interface': 6.790721693393912e-07, 'learning_rate/action_model': 6.790721693393912e-06, 'epoch': 2.9}) +[2026-06-18 02:31:04,501][starVLA.training.train_starvla][INFO] - Step 6396, Loss: {'train/loss': 0.1399773806333542, 'train/grad_norm_pre_clip': 33.70701217651367, 'timing/data': 0.00048377178609371185, 'timing/model': 0.9254158642143011, 'learning_rate/qwen_vl_interface': 6.784844168878108e-07, 'learning_rate/action_model': 6.784844168878107e-06, 'epoch': 2.9}) +[2026-06-18 02:31:08,250][starVLA.training.train_starvla][INFO] - Step 6397, Loss: {'train/loss': 0.12921133637428284, 'train/grad_norm_pre_clip': 26.67707633972168, 'timing/data': 0.0004915026947855949, 'timing/model': 0.9257839592173696, 'learning_rate/qwen_vl_interface': 6.778976121163289e-07, 'learning_rate/action_model': 6.778976121163288e-06, 'epoch': 2.9}) +[2026-06-18 02:31:12,088][starVLA.training.train_starvla][INFO] - Step 6398, Loss: {'train/loss': 0.1585550308227539, 'train/grad_norm_pre_clip': 43.604881286621094, 'timing/data': 0.0005596242845058441, 'timing/model': 0.9346136385574937, 'learning_rate/qwen_vl_interface': 6.773117551465927e-07, 'learning_rate/action_model': 6.773117551465927e-06, 'epoch': 2.9}) +[2026-06-18 02:31:15,868][starVLA.training.train_starvla][INFO] - Step 6399, Loss: {'train/loss': 0.2110472023487091, 'train/grad_norm_pre_clip': 29.552743911743164, 'timing/data': 0.0005019642412662506, 'timing/model': 0.9260996114462614, 'learning_rate/qwen_vl_interface': 6.767268461000498e-07, 'learning_rate/action_model': 6.7672684610004975e-06, 'epoch': 2.9}) +[2026-06-18 02:33:17,445][starVLA.training.train_starvla][INFO] - Step 6400, Loss: {'train/loss': 0.10500757396221161, 'train/grad_norm_pre_clip': 28.87102508544922, 'eval/loss': 0.25418195128440857, 'eval/latency_8/loss': 0.25418195128440857, 'eval/demon_fire_precision': 0.9755114693118413, 'eval/demon_fire_recall': 0.9776328052190121, 'eval/demon_fire_f1': 0.9765709852598914, 'eval/demon_fire_rate_model': 0.897607122982749, 'eval/demon_fire_rate_teacher': 0.8956594323873122, 'eval/demon_left_precision': 0.8969555035128806, 'eval/demon_left_recall': 0.8201284796573876, 'eval/demon_left_f1': 0.8568232662192394, 'eval/demon_left_rate_model': 0.1188091263216472, 'eval/demon_left_rate_teacher': 0.12993878686700056, 'eval/demon_right_precision': 0.8648648648648649, 'eval/demon_right_recall': 0.8727272727272727, 'eval/demon_right_f1': 0.8687782805429864, 'eval/demon_right_rate_model': 0.18530884808013356, 'eval/demon_right_rate_teacher': 0.18363939899833054, 'eval/demon_move_f1': 0.8628007733811129, 'eval/demon_cc_f1': 0.9394094028526445, 'eval/latency_8/demon_fire_precision': 0.9755114693118413, 'eval/latency_8/demon_fire_recall': 0.9776328052190121, 'eval/latency_8/demon_fire_f1': 0.9765709852598914, 'eval/latency_8/demon_fire_rate_model': 0.897607122982749, 'eval/latency_8/demon_fire_rate_teacher': 0.8956594323873122, 'eval/latency_8/demon_left_precision': 0.8969555035128806, 'eval/latency_8/demon_left_recall': 0.8201284796573876, 'eval/latency_8/demon_left_f1': 0.8568232662192394, 'eval/latency_8/demon_left_rate_model': 0.1188091263216472, 'eval/latency_8/demon_left_rate_teacher': 0.12993878686700056, 'eval/latency_8/demon_right_precision': 0.8648648648648649, 'eval/latency_8/demon_right_recall': 0.8727272727272727, 'eval/latency_8/demon_right_f1': 0.8687782805429864, 'eval/latency_8/demon_right_rate_model': 0.18530884808013356, 'eval/latency_8/demon_right_rate_teacher': 0.18363939899833054, 'eval/latency_8/demon_move_f1': 0.8628007733811129, 'eval/latency_8/demon_cc_f1': 0.9394094028526445, 'timing/data': 0.0004917113110423088, 'timing/model': 1.1692065745592117, 'learning_rate/qwen_vl_interface': 6.761428850979542e-07, 'learning_rate/action_model': 6.7614288509795415e-06, 'epoch': 2.9}) +[2026-06-18 02:33:17,448][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_6400_state +[2026-06-18 02:33:17,452][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 02:33:23,890][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_6400_state/model.safetensors +[2026-06-18 02:33:37,820][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_6400_state/optimizer.bin +[2026-06-18 02:33:37,823][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_6400_state/sampler.bin +[2026-06-18 02:33:37,823][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_6400_state/sampler_1.bin +[2026-06-18 02:33:37,826][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_6400_state/random_states_0.pkl +[2026-06-18 02:33:37,829][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 6400 +[2026-06-18 02:33:37,829][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 02:33:37,840][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 02:33:41,622][starVLA.training.train_starvla][INFO] - Step 6401, Loss: {'train/loss': 0.16134272515773773, 'train/grad_norm_pre_clip': 25.935123443603516, 'timing/data': 0.0010487502440810204, 'timing/model': 0.9265848882496357, 'learning_rate/qwen_vl_interface': 6.755598722613579e-07, 'learning_rate/action_model': 6.755598722613579e-06, 'epoch': 2.9}) +[2026-06-18 02:33:45,477][starVLA.training.train_starvla][INFO] - Step 6402, Loss: {'train/loss': 0.21280966699123383, 'train/grad_norm_pre_clip': 24.630083084106445, 'timing/data': 0.00048187654465436935, 'timing/model': 1.0009159883484244, 'learning_rate/qwen_vl_interface': 6.749778077111241e-07, 'learning_rate/action_model': 6.749778077111241e-06, 'epoch': 2.9}) +[2026-06-18 02:33:49,381][starVLA.training.train_starvla][INFO] - Step 6403, Loss: {'train/loss': 0.09158625453710556, 'train/grad_norm_pre_clip': 22.94921112060547, 'timing/data': 0.000504777766764164, 'timing/model': 1.0909684812650084, 'learning_rate/qwen_vl_interface': 6.743966915679128e-07, 'learning_rate/action_model': 6.743966915679129e-06, 'epoch': 2.9}) +[2026-06-18 02:33:53,038][starVLA.training.train_starvla][INFO] - Step 6404, Loss: {'train/loss': 0.06032910197973251, 'train/grad_norm_pre_clip': 28.223609924316406, 'timing/data': 0.0004832586273550987, 'timing/model': 0.9230867512524128, 'learning_rate/qwen_vl_interface': 6.738165239521909e-07, 'learning_rate/action_model': 6.7381652395219094e-06, 'epoch': 2.91}) +[2026-06-18 02:33:56,831][starVLA.training.train_starvla][INFO] - Step 6405, Loss: {'train/loss': 0.12085391581058502, 'train/grad_norm_pre_clip': 24.930892944335938, 'timing/data': 0.0008470788598060608, 'timing/model': 0.9314596112817526, 'learning_rate/qwen_vl_interface': 6.732373049842284e-07, 'learning_rate/action_model': 6.7323730498422846e-06, 'epoch': 2.91}) +[2026-06-18 02:34:00,758][starVLA.training.train_starvla][INFO] - Step 6406, Loss: {'train/loss': 0.14785321056842804, 'train/grad_norm_pre_clip': 29.11441421508789, 'timing/data': 0.00044832751154899597, 'timing/model': 0.9256522776558995, 'learning_rate/qwen_vl_interface': 6.726590347840972e-07, 'learning_rate/action_model': 6.726590347840972e-06, 'epoch': 2.91}) +[2026-06-18 02:34:04,553][starVLA.training.train_starvla][INFO] - Step 6407, Loss: {'train/loss': 0.3474176526069641, 'train/grad_norm_pre_clip': 21.758577346801758, 'timing/data': 0.0010363301262259483, 'timing/model': 0.9121026657521725, 'learning_rate/qwen_vl_interface': 6.720817134716736e-07, 'learning_rate/action_model': 6.720817134716736e-06, 'epoch': 2.91}) +[2026-06-18 02:34:08,653][starVLA.training.train_starvla][INFO] - Step 6408, Loss: {'train/loss': 0.26414552330970764, 'train/grad_norm_pre_clip': 31.60382652282715, 'timing/data': 0.0005006389692425728, 'timing/model': 1.2482219748198986, 'learning_rate/qwen_vl_interface': 6.71505341166636e-07, 'learning_rate/action_model': 6.71505341166636e-06, 'epoch': 2.91}) +[2026-06-18 02:34:12,485][starVLA.training.train_starvla][INFO] - Step 6409, Loss: {'train/loss': 0.155190572142601, 'train/grad_norm_pre_clip': 30.11331558227539, 'timing/data': 0.0006973976269364357, 'timing/model': 1.1398036191239953, 'learning_rate/qwen_vl_interface': 6.709299179884689e-07, 'learning_rate/action_model': 6.709299179884689e-06, 'epoch': 2.91}) +[2026-06-18 02:34:16,400][starVLA.training.train_starvla][INFO] - Step 6410, Loss: {'train/loss': 0.11538062244653702, 'train/grad_norm_pre_clip': 175.2180633544922, 'timing/data': 0.002191353589296341, 'timing/model': 1.041393954306841, 'learning_rate/qwen_vl_interface': 6.703554440564569e-07, 'learning_rate/action_model': 6.703554440564569e-06, 'epoch': 2.91}) +[2026-06-18 02:34:19,965][starVLA.training.train_starvla][INFO] - Step 6411, Loss: {'train/loss': 0.12045355886220932, 'train/grad_norm_pre_clip': 80.10875701904297, 'timing/data': 0.0005552181974053383, 'timing/model': 0.9281422337517142, 'learning_rate/qwen_vl_interface': 6.697819194896889e-07, 'learning_rate/action_model': 6.697819194896889e-06, 'epoch': 2.91}) +[2026-06-18 02:34:23,732][starVLA.training.train_starvla][INFO] - Step 6412, Loss: {'train/loss': 0.4023608863353729, 'train/grad_norm_pre_clip': 39.466182708740234, 'timing/data': 0.0005506612360477448, 'timing/model': 0.9188945321366191, 'learning_rate/qwen_vl_interface': 6.692093444070571e-07, 'learning_rate/action_model': 6.692093444070572e-06, 'epoch': 2.91}) +[2026-06-18 02:34:27,475][starVLA.training.train_starvla][INFO] - Step 6413, Loss: {'train/loss': 0.32189083099365234, 'train/grad_norm_pre_clip': 32.867431640625, 'timing/data': 0.0004982287064194679, 'timing/model': 0.9148633116856217, 'learning_rate/qwen_vl_interface': 6.686377189272584e-07, 'learning_rate/action_model': 6.686377189272583e-06, 'epoch': 2.91}) +[2026-06-18 02:34:31,342][starVLA.training.train_starvla][INFO] - Step 6414, Loss: {'train/loss': 0.12253943830728531, 'train/grad_norm_pre_clip': 26.989437103271484, 'timing/data': 0.0005274992436170578, 'timing/model': 0.9551322991028428, 'learning_rate/qwen_vl_interface': 6.680670431687898e-07, 'learning_rate/action_model': 6.680670431687898e-06, 'epoch': 2.91}) +[2026-06-18 02:34:35,115][starVLA.training.train_starvla][INFO] - Step 6415, Loss: {'train/loss': 0.3401773273944855, 'train/grad_norm_pre_clip': 29.183128356933594, 'timing/data': 0.0007367106154561043, 'timing/model': 0.9319411208853126, 'learning_rate/qwen_vl_interface': 6.674973172499536e-07, 'learning_rate/action_model': 6.674973172499535e-06, 'epoch': 2.91}) +[2026-06-18 02:34:38,935][starVLA.training.train_starvla][INFO] - Step 6416, Loss: {'train/loss': 0.45865026116371155, 'train/grad_norm_pre_clip': 34.0557861328125, 'timing/data': 0.00042961351573467255, 'timing/model': 1.1774426810443401, 'learning_rate/qwen_vl_interface': 6.669285412888544e-07, 'learning_rate/action_model': 6.669285412888544e-06, 'epoch': 2.91}) +[2026-06-18 02:34:42,848][starVLA.training.train_starvla][INFO] - Step 6417, Loss: {'train/loss': 0.36492496728897095, 'train/grad_norm_pre_clip': 35.60000228881836, 'timing/data': 0.0006382698193192482, 'timing/model': 0.9272590437904, 'learning_rate/qwen_vl_interface': 6.663607154034003e-07, 'learning_rate/action_model': 6.663607154034003e-06, 'epoch': 2.91}) +[2026-06-18 02:34:46,615][starVLA.training.train_starvla][INFO] - Step 6418, Loss: {'train/loss': 0.254215270280838, 'train/grad_norm_pre_clip': 30.090879440307617, 'timing/data': 0.00047733820974826813, 'timing/model': 0.915388286113739, 'learning_rate/qwen_vl_interface': 6.657938397113019e-07, 'learning_rate/action_model': 6.657938397113018e-06, 'epoch': 2.91}) +[2026-06-18 02:34:50,666][starVLA.training.train_starvla][INFO] - Step 6419, Loss: {'train/loss': 0.26496827602386475, 'train/grad_norm_pre_clip': 27.96451759338379, 'timing/data': 0.0005006063729524612, 'timing/model': 1.1963641690090299, 'learning_rate/qwen_vl_interface': 6.652279143300734e-07, 'learning_rate/action_model': 6.652279143300734e-06, 'epoch': 2.91}) +[2026-06-18 02:34:54,461][starVLA.training.train_starvla][INFO] - Step 6420, Loss: {'train/loss': 0.31468695402145386, 'train/grad_norm_pre_clip': 34.23820114135742, 'timing/data': 0.0004813326522707939, 'timing/model': 0.9231161205098033, 'learning_rate/qwen_vl_interface': 6.64662939377032e-07, 'learning_rate/action_model': 6.646629393770321e-06, 'epoch': 2.91}) +[2026-06-18 02:34:58,165][starVLA.training.train_starvla][INFO] - Step 6421, Loss: {'train/loss': 0.2710053026676178, 'train/grad_norm_pre_clip': 26.21116065979004, 'timing/data': 0.00047504156827926636, 'timing/model': 0.9198744269087911, 'learning_rate/qwen_vl_interface': 6.64098914969297e-07, 'learning_rate/action_model': 6.6409891496929695e-06, 'epoch': 2.91}) +[2026-06-18 02:35:01,904][starVLA.training.train_starvla][INFO] - Step 6422, Loss: {'train/loss': 0.1697688102722168, 'train/grad_norm_pre_clip': 26.27788734436035, 'timing/data': 0.0009996239095926285, 'timing/model': 1.091399623081088, 'learning_rate/qwen_vl_interface': 6.635358412237912e-07, 'learning_rate/action_model': 6.635358412237912e-06, 'epoch': 2.91}) +[2026-06-18 02:35:05,834][starVLA.training.train_starvla][INFO] - Step 6423, Loss: {'train/loss': 0.3309789299964905, 'train/grad_norm_pre_clip': 46.839691162109375, 'timing/data': 0.0009101377800107002, 'timing/model': 1.1632341537624598, 'learning_rate/qwen_vl_interface': 6.629737182572418e-07, 'learning_rate/action_model': 6.629737182572418e-06, 'epoch': 2.91}) +[2026-06-18 02:35:09,716][starVLA.training.train_starvla][INFO] - Step 6424, Loss: {'train/loss': 0.11486048996448517, 'train/grad_norm_pre_clip': 19.75061798095703, 'timing/data': 0.0009348951280117035, 'timing/model': 1.139552197419107, 'learning_rate/qwen_vl_interface': 6.62412546186176e-07, 'learning_rate/action_model': 6.62412546186176e-06, 'epoch': 2.91}) +[2026-06-18 02:35:13,702][starVLA.training.train_starvla][INFO] - Step 6425, Loss: {'train/loss': 0.3049624264240265, 'train/grad_norm_pre_clip': 41.017173767089844, 'timing/data': 0.0005827387794852257, 'timing/model': 1.0385572975501418, 'learning_rate/qwen_vl_interface': 6.61852325126926e-07, 'learning_rate/action_model': 6.61852325126926e-06, 'epoch': 2.91}) +[2026-06-18 02:35:17,571][starVLA.training.train_starvla][INFO] - Step 6426, Loss: {'train/loss': 0.3337258994579315, 'train/grad_norm_pre_clip': 35.72698974609375, 'timing/data': 0.0007965508848428726, 'timing/model': 0.9680592585355043, 'learning_rate/qwen_vl_interface': 6.61293055195625e-07, 'learning_rate/action_model': 6.6129305519562504e-06, 'epoch': 2.92}) +[2026-06-18 02:35:21,453][starVLA.training.train_starvla][INFO] - Step 6427, Loss: {'train/loss': 0.09648507833480835, 'train/grad_norm_pre_clip': 23.59782600402832, 'timing/data': 0.00046815257519483566, 'timing/model': 0.929805213585496, 'learning_rate/qwen_vl_interface': 6.607347365082127e-07, 'learning_rate/action_model': 6.607347365082127e-06, 'epoch': 2.92}) +[2026-06-18 02:35:25,979][starVLA.training.train_starvla][INFO] - Step 6428, Loss: {'train/loss': 0.40585780143737793, 'train/grad_norm_pre_clip': 49.4389533996582, 'timing/data': 0.0004844348877668381, 'timing/model': 0.9309746054932475, 'learning_rate/qwen_vl_interface': 6.601773691804273e-07, 'learning_rate/action_model': 6.6017736918042735e-06, 'epoch': 2.92}) +[2026-06-18 02:35:29,869][starVLA.training.train_starvla][INFO] - Step 6429, Loss: {'train/loss': 0.19327206909656525, 'train/grad_norm_pre_clip': 32.85342025756836, 'timing/data': 0.0012270165607333183, 'timing/model': 1.1663615433499217, 'learning_rate/qwen_vl_interface': 6.596209533278122e-07, 'learning_rate/action_model': 6.596209533278122e-06, 'epoch': 2.92}) +[2026-06-18 02:35:33,678][starVLA.training.train_starvla][INFO] - Step 6430, Loss: {'train/loss': 0.28604891896247864, 'train/grad_norm_pre_clip': 29.551300048828125, 'timing/data': 0.0007281312718987465, 'timing/model': 1.0570113537833095, 'learning_rate/qwen_vl_interface': 6.590654890657125e-07, 'learning_rate/action_model': 6.590654890657125e-06, 'epoch': 2.92}) +[2026-06-18 02:35:37,550][starVLA.training.train_starvla][INFO] - Step 6431, Loss: {'train/loss': 0.31626778841018677, 'train/grad_norm_pre_clip': 49.15925979614258, 'timing/data': 0.001111331395804882, 'timing/model': 0.9325826829299331, 'learning_rate/qwen_vl_interface': 6.585109765092772e-07, 'learning_rate/action_model': 6.585109765092772e-06, 'epoch': 2.92}) +[2026-06-18 02:35:41,429][starVLA.training.train_starvla][INFO] - Step 6432, Loss: {'train/loss': 0.17391209304332733, 'train/grad_norm_pre_clip': 36.25740432739258, 'timing/data': 0.00048395898193120956, 'timing/model': 0.9430878898128867, 'learning_rate/qwen_vl_interface': 6.579574157734575e-07, 'learning_rate/action_model': 6.579574157734575e-06, 'epoch': 2.92}) +[2026-06-18 02:35:45,915][starVLA.training.train_starvla][INFO] - Step 6433, Loss: {'train/loss': 0.5254502296447754, 'train/grad_norm_pre_clip': 34.132957458496094, 'timing/data': 0.0006473483517765999, 'timing/model': 1.5425249617546797, 'learning_rate/qwen_vl_interface': 6.574048069730062e-07, 'learning_rate/action_model': 6.574048069730063e-06, 'epoch': 2.92}) +[2026-06-18 02:35:49,681][starVLA.training.train_starvla][INFO] - Step 6434, Loss: {'train/loss': 0.08299271017313004, 'train/grad_norm_pre_clip': 25.71120834350586, 'timing/data': 0.0008505191653966904, 'timing/model': 0.9940807335078716, 'learning_rate/qwen_vl_interface': 6.568531502224808e-07, 'learning_rate/action_model': 6.568531502224808e-06, 'epoch': 2.92}) +[2026-06-18 02:35:53,542][starVLA.training.train_starvla][INFO] - Step 6435, Loss: {'train/loss': 0.07683224231004715, 'train/grad_norm_pre_clip': 13.380995750427246, 'timing/data': 0.0007794369012117386, 'timing/model': 0.94587894808501, 'learning_rate/qwen_vl_interface': 6.563024456362395e-07, 'learning_rate/action_model': 6.563024456362395e-06, 'epoch': 2.92}) +[2026-06-18 02:35:57,322][starVLA.training.train_starvla][INFO] - Step 6436, Loss: {'train/loss': 0.16369843482971191, 'train/grad_norm_pre_clip': 32.6810188293457, 'timing/data': 0.0007906388491392136, 'timing/model': 1.1215994739905, 'learning_rate/qwen_vl_interface': 6.557526933284443e-07, 'learning_rate/action_model': 6.557526933284443e-06, 'epoch': 2.92}) +[2026-06-18 02:36:00,920][starVLA.training.train_starvla][INFO] - Step 6437, Loss: {'train/loss': 0.1305765062570572, 'train/grad_norm_pre_clip': 21.17401123046875, 'timing/data': 0.0005010329186916351, 'timing/model': 0.946465983055532, 'learning_rate/qwen_vl_interface': 6.552038934130595e-07, 'learning_rate/action_model': 6.552038934130594e-06, 'epoch': 2.92}) +[2026-06-18 02:36:04,673][starVLA.training.train_starvla][INFO] - Step 6438, Loss: {'train/loss': 0.26813921332359314, 'train/grad_norm_pre_clip': 26.35309410095215, 'timing/data': 0.0004694061353802681, 'timing/model': 0.9232060806825757, 'learning_rate/qwen_vl_interface': 6.546560460038514e-07, 'learning_rate/action_model': 6.546560460038514e-06, 'epoch': 2.92}) +[2026-06-18 02:36:08,366][starVLA.training.train_starvla][INFO] - Step 6439, Loss: {'train/loss': 0.29366782307624817, 'train/grad_norm_pre_clip': 22.97231101989746, 'timing/data': 0.00048497505486011505, 'timing/model': 0.9236699678003788, 'learning_rate/qwen_vl_interface': 6.5410915121439e-07, 'learning_rate/action_model': 6.5410915121439e-06, 'epoch': 2.92}) +[2026-06-18 02:36:12,337][starVLA.training.train_starvla][INFO] - Step 6440, Loss: {'train/loss': 0.4230627417564392, 'train/grad_norm_pre_clip': 31.142311096191406, 'timing/data': 0.0005686338990926743, 'timing/model': 1.172921184450388, 'learning_rate/qwen_vl_interface': 6.535632091580474e-07, 'learning_rate/action_model': 6.535632091580474e-06, 'epoch': 2.92}) +[2026-06-18 02:36:16,130][starVLA.training.train_starvla][INFO] - Step 6441, Loss: {'train/loss': 0.29841744899749756, 'train/grad_norm_pre_clip': 40.81279754638672, 'timing/data': 0.0014547882601618767, 'timing/model': 0.9295036680996418, 'learning_rate/qwen_vl_interface': 6.530182199479965e-07, 'learning_rate/action_model': 6.5301821994799645e-06, 'epoch': 2.92}) +[2026-06-18 02:36:19,971][starVLA.training.train_starvla][INFO] - Step 6442, Loss: {'train/loss': 0.11620534956455231, 'train/grad_norm_pre_clip': 31.376903533935547, 'timing/data': 0.0005017900839447975, 'timing/model': 1.1238018842414021, 'learning_rate/qwen_vl_interface': 6.524741836972162e-07, 'learning_rate/action_model': 6.524741836972162e-06, 'epoch': 2.92}) +[2026-06-18 02:36:23,736][starVLA.training.train_starvla][INFO] - Step 6443, Loss: {'train/loss': 0.16379986703395844, 'train/grad_norm_pre_clip': 25.29682159423828, 'timing/data': 0.0005035819485783577, 'timing/model': 1.0538860596716404, 'learning_rate/qwen_vl_interface': 6.519311005184848e-07, 'learning_rate/action_model': 6.5193110051848476e-06, 'epoch': 2.92}) +[2026-06-18 02:36:27,430][starVLA.training.train_starvla][INFO] - Step 6444, Loss: {'train/loss': 0.17977255582809448, 'train/grad_norm_pre_clip': 26.819551467895508, 'timing/data': 0.0008942242711782455, 'timing/model': 1.056592920795083, 'learning_rate/qwen_vl_interface': 6.51388970524383e-07, 'learning_rate/action_model': 6.513889705243829e-06, 'epoch': 2.92}) +[2026-06-18 02:36:31,628][starVLA.training.train_starvla][INFO] - Step 6445, Loss: {'train/loss': 0.08109111338853836, 'train/grad_norm_pre_clip': 16.61128807067871, 'timing/data': 0.0005715666338801384, 'timing/model': 0.927623437717557, 'learning_rate/qwen_vl_interface': 6.508477938272963e-07, 'learning_rate/action_model': 6.508477938272962e-06, 'epoch': 2.92}) +[2026-06-18 02:36:35,460][starVLA.training.train_starvla][INFO] - Step 6446, Loss: {'train/loss': 0.020688217133283615, 'train/grad_norm_pre_clip': 7.608153820037842, 'timing/data': 0.0004916060715913773, 'timing/model': 0.9282894413918257, 'learning_rate/qwen_vl_interface': 6.503075705394113e-07, 'learning_rate/action_model': 6.5030757053941125e-06, 'epoch': 2.92}) +[2026-06-18 02:36:39,213][starVLA.training.train_starvla][INFO] - Step 6447, Loss: {'train/loss': 0.1494726687669754, 'train/grad_norm_pre_clip': 20.550804138183594, 'timing/data': 0.0006627365946769714, 'timing/model': 0.9322761436924338, 'learning_rate/qwen_vl_interface': 6.497683007727166e-07, 'learning_rate/action_model': 6.497683007727166e-06, 'epoch': 2.92}) +[2026-06-18 02:36:42,994][starVLA.training.train_starvla][INFO] - Step 6448, Loss: {'train/loss': 0.3034542500972748, 'train/grad_norm_pre_clip': 32.187313079833984, 'timing/data': 0.0009278403595089912, 'timing/model': 0.9211064595729113, 'learning_rate/qwen_vl_interface': 6.492299846390024e-07, 'learning_rate/action_model': 6.492299846390023e-06, 'epoch': 2.93}) +[2026-06-18 02:36:46,757][starVLA.training.train_starvla][INFO] - Step 6449, Loss: {'train/loss': 0.11747024953365326, 'train/grad_norm_pre_clip': 19.340978622436523, 'timing/data': 0.0004979511722922325, 'timing/model': 1.1077897241339087, 'learning_rate/qwen_vl_interface': 6.486926222498633e-07, 'learning_rate/action_model': 6.486926222498633e-06, 'epoch': 2.93}) +[2026-06-18 02:36:50,537][starVLA.training.train_starvla][INFO] - Step 6450, Loss: {'train/loss': 0.058271344751119614, 'train/grad_norm_pre_clip': 19.87750816345215, 'timing/data': 0.00046626757830381393, 'timing/model': 1.095441804267466, 'learning_rate/qwen_vl_interface': 6.481562137166952e-07, 'learning_rate/action_model': 6.481562137166952e-06, 'epoch': 2.93}) +[2026-06-18 02:36:54,231][starVLA.training.train_starvla][INFO] - Step 6451, Loss: {'train/loss': 0.0829639658331871, 'train/grad_norm_pre_clip': 17.869211196899414, 'timing/data': 0.0005204780027270317, 'timing/model': 0.9280561907216907, 'learning_rate/qwen_vl_interface': 6.476207591506954e-07, 'learning_rate/action_model': 6.476207591506954e-06, 'epoch': 2.93}) +[2026-06-18 02:36:59,092][starVLA.training.train_starvla][INFO] - Step 6452, Loss: {'train/loss': 0.1749500036239624, 'train/grad_norm_pre_clip': 15.168370246887207, 'timing/data': 0.0005421899259090424, 'timing/model': 0.922504304908216, 'learning_rate/qwen_vl_interface': 6.470862586628645e-07, 'learning_rate/action_model': 6.470862586628646e-06, 'epoch': 2.93}) +[2026-06-18 02:37:03,257][starVLA.training.train_starvla][INFO] - Step 6453, Loss: {'train/loss': 0.15082168579101562, 'train/grad_norm_pre_clip': 29.57391357421875, 'timing/data': 0.0013304995372891426, 'timing/model': 0.9597168229520321, 'learning_rate/qwen_vl_interface': 6.465527123640056e-07, 'learning_rate/action_model': 6.465527123640056e-06, 'epoch': 2.93}) +[2026-06-18 02:37:07,333][starVLA.training.train_starvla][INFO] - Step 6454, Loss: {'train/loss': 0.12610571086406708, 'train/grad_norm_pre_clip': 25.812101364135742, 'timing/data': 0.0011138692498207092, 'timing/model': 0.9993736073374748, 'learning_rate/qwen_vl_interface': 6.460201203647229e-07, 'learning_rate/action_model': 6.460201203647229e-06, 'epoch': 2.93}) +[2026-06-18 02:37:11,296][starVLA.training.train_starvla][INFO] - Step 6455, Loss: {'train/loss': 0.11621280014514923, 'train/grad_norm_pre_clip': 24.75142478942871, 'timing/data': 0.00048756133764982224, 'timing/model': 1.0815178407356143, 'learning_rate/qwen_vl_interface': 6.454884827754224e-07, 'learning_rate/action_model': 6.4548848277542235e-06, 'epoch': 2.93}) +[2026-06-18 02:37:15,842][starVLA.training.train_starvla][INFO] - Step 6456, Loss: {'train/loss': 0.22810570895671844, 'train/grad_norm_pre_clip': 29.618532180786133, 'timing/data': 0.0006845081225037575, 'timing/model': 1.144644184038043, 'learning_rate/qwen_vl_interface': 6.449577997063156e-07, 'learning_rate/action_model': 6.449577997063155e-06, 'epoch': 2.93}) +[2026-06-18 02:37:19,733][starVLA.training.train_starvla][INFO] - Step 6457, Loss: {'train/loss': 0.097586490213871, 'train/grad_norm_pre_clip': 22.649757385253906, 'timing/data': 0.0008324114605784416, 'timing/model': 1.1424649758264422, 'learning_rate/qwen_vl_interface': 6.444280712674112e-07, 'learning_rate/action_model': 6.444280712674111e-06, 'epoch': 2.93}) +[2026-06-18 02:37:23,397][starVLA.training.train_starvla][INFO] - Step 6458, Loss: {'train/loss': 0.21425117552280426, 'train/grad_norm_pre_clip': 31.44951820373535, 'timing/data': 0.0004921304062008858, 'timing/model': 0.9607753297314048, 'learning_rate/qwen_vl_interface': 6.438992975685237e-07, 'learning_rate/action_model': 6.438992975685237e-06, 'epoch': 2.93}) +[2026-06-18 02:37:27,623][starVLA.training.train_starvla][INFO] - Step 6459, Loss: {'train/loss': 0.0877167209982872, 'train/grad_norm_pre_clip': 15.914546966552734, 'timing/data': 0.0008160676807165146, 'timing/model': 1.3550941068679094, 'learning_rate/qwen_vl_interface': 6.433714787192677e-07, 'learning_rate/action_model': 6.4337147871926766e-06, 'epoch': 2.93}) +[2026-06-18 02:37:31,392][starVLA.training.train_starvla][INFO] - Step 6460, Loss: {'train/loss': 0.0669310986995697, 'train/grad_norm_pre_clip': 17.561216354370117, 'timing/data': 0.00046394579112529755, 'timing/model': 0.911404300481081, 'learning_rate/qwen_vl_interface': 6.428446148290615e-07, 'learning_rate/action_model': 6.428446148290615e-06, 'epoch': 2.93}) +[2026-06-18 02:37:35,285][starVLA.training.train_starvla][INFO] - Step 6461, Loss: {'train/loss': 0.05217386782169342, 'train/grad_norm_pre_clip': 11.633211135864258, 'timing/data': 0.0012327199801802635, 'timing/model': 0.9230411741882563, 'learning_rate/qwen_vl_interface': 6.423187060071248e-07, 'learning_rate/action_model': 6.423187060071248e-06, 'epoch': 2.93}) +[2026-06-18 02:37:39,115][starVLA.training.train_starvla][INFO] - Step 6462, Loss: {'train/loss': 0.24163345992565155, 'train/grad_norm_pre_clip': 21.648040771484375, 'timing/data': 0.0007986510172486305, 'timing/model': 1.1310531962662935, 'learning_rate/qwen_vl_interface': 6.417937523624778e-07, 'learning_rate/action_model': 6.417937523624778e-06, 'epoch': 2.93}) +[2026-06-18 02:37:42,829][starVLA.training.train_starvla][INFO] - Step 6463, Loss: {'train/loss': 0.15810638666152954, 'train/grad_norm_pre_clip': 33.497962951660156, 'timing/data': 0.0009710267186164856, 'timing/model': 1.0425574323162436, 'learning_rate/qwen_vl_interface': 6.412697540039442e-07, 'learning_rate/action_model': 6.412697540039442e-06, 'epoch': 2.93}) +[2026-06-18 02:37:46,799][starVLA.training.train_starvla][INFO] - Step 6464, Loss: {'train/loss': 0.11221799999475479, 'train/grad_norm_pre_clip': 25.398221969604492, 'timing/data': 0.0009942930191755295, 'timing/model': 1.235998003743589, 'learning_rate/qwen_vl_interface': 6.407467110401505e-07, 'learning_rate/action_model': 6.407467110401505e-06, 'epoch': 2.93}) +[2026-06-18 02:37:50,756][starVLA.training.train_starvla][INFO] - Step 6465, Loss: {'train/loss': 0.0971643254160881, 'train/grad_norm_pre_clip': 22.3900089263916, 'timing/data': 0.0009281225502490997, 'timing/model': 0.9652574947103858, 'learning_rate/qwen_vl_interface': 6.402246235795228e-07, 'learning_rate/action_model': 6.402246235795227e-06, 'epoch': 2.93}) +[2026-06-18 02:37:54,615][starVLA.training.train_starvla][INFO] - Step 6466, Loss: {'train/loss': 0.07219002395868301, 'train/grad_norm_pre_clip': 18.056026458740234, 'timing/data': 0.0005138721317052841, 'timing/model': 0.9336931277066469, 'learning_rate/qwen_vl_interface': 6.397034917302913e-07, 'learning_rate/action_model': 6.397034917302913e-06, 'epoch': 2.93}) +[2026-06-18 02:37:58,407][starVLA.training.train_starvla][INFO] - Step 6467, Loss: {'train/loss': 0.5821738839149475, 'train/grad_norm_pre_clip': 47.797977447509766, 'timing/data': 0.000732719898223877, 'timing/model': 0.950911395251751, 'learning_rate/qwen_vl_interface': 6.391833156004871e-07, 'learning_rate/action_model': 6.39183315600487e-06, 'epoch': 2.93}) +[2026-06-18 02:38:02,150][starVLA.training.train_starvla][INFO] - Step 6468, Loss: {'train/loss': 0.1583118736743927, 'train/grad_norm_pre_clip': 30.12224578857422, 'timing/data': 0.0011964663863182068, 'timing/model': 0.934645569883287, 'learning_rate/qwen_vl_interface': 6.386640952979427e-07, 'learning_rate/action_model': 6.386640952979426e-06, 'epoch': 2.93}) +[2026-06-18 02:38:05,862][starVLA.training.train_starvla][INFO] - Step 6469, Loss: {'train/loss': 0.06166777387261391, 'train/grad_norm_pre_clip': 12.627309799194336, 'timing/data': 0.0005883807316422462, 'timing/model': 1.0745431259274483, 'learning_rate/qwen_vl_interface': 6.381458309302937e-07, 'learning_rate/action_model': 6.381458309302937e-06, 'epoch': 2.93}) +[2026-06-18 02:38:09,446][starVLA.training.train_starvla][INFO] - Step 6470, Loss: {'train/loss': 0.48394471406936646, 'train/grad_norm_pre_clip': 40.17372512817383, 'timing/data': 0.000508197583258152, 'timing/model': 0.9356212038546801, 'learning_rate/qwen_vl_interface': 6.376285226049762e-07, 'learning_rate/action_model': 6.376285226049762e-06, 'epoch': 2.94}) +[2026-06-18 02:38:13,187][starVLA.training.train_starvla][INFO] - Step 6471, Loss: {'train/loss': 0.12189195305109024, 'train/grad_norm_pre_clip': 16.346214294433594, 'timing/data': 0.001241006888449192, 'timing/model': 0.9249241696670651, 'learning_rate/qwen_vl_interface': 6.371121704292292e-07, 'learning_rate/action_model': 6.371121704292292e-06, 'epoch': 2.94}) +[2026-06-18 02:38:16,935][starVLA.training.train_starvla][INFO] - Step 6472, Loss: {'train/loss': 0.039649032056331635, 'train/grad_norm_pre_clip': 8.551453590393066, 'timing/data': 0.00047550443559885025, 'timing/model': 0.9287294475361705, 'learning_rate/qwen_vl_interface': 6.365967745100926e-07, 'learning_rate/action_model': 6.3659677451009256e-06, 'epoch': 2.94}) +[2026-06-18 02:38:20,671][starVLA.training.train_starvla][INFO] - Step 6473, Loss: {'train/loss': 0.2769653797149658, 'train/grad_norm_pre_clip': 20.7626953125, 'timing/data': 0.0004961462691426277, 'timing/model': 0.9246322652325034, 'learning_rate/qwen_vl_interface': 6.360823349544092e-07, 'learning_rate/action_model': 6.360823349544092e-06, 'epoch': 2.94}) +[2026-06-18 02:38:24,815][starVLA.training.train_starvla][INFO] - Step 6474, Loss: {'train/loss': 0.0492851585149765, 'train/grad_norm_pre_clip': 13.019002914428711, 'timing/data': 0.0005102409049868584, 'timing/model': 0.9226246103644371, 'learning_rate/qwen_vl_interface': 6.355688518688222e-07, 'learning_rate/action_model': 6.355688518688221e-06, 'epoch': 2.94}) +[2026-06-18 02:38:28,572][starVLA.training.train_starvla][INFO] - Step 6475, Loss: {'train/loss': 0.1792038083076477, 'train/grad_norm_pre_clip': 41.3286018371582, 'timing/data': 0.0009207213297486305, 'timing/model': 1.113480368629098, 'learning_rate/qwen_vl_interface': 6.350563253597781e-07, 'learning_rate/action_model': 6.350563253597781e-06, 'epoch': 2.94}) +[2026-06-18 02:38:32,141][starVLA.training.train_starvla][INFO] - Step 6476, Loss: {'train/loss': 0.19325099885463715, 'train/grad_norm_pre_clip': 30.001907348632812, 'timing/data': 0.0004797130823135376, 'timing/model': 0.9257897138595581, 'learning_rate/qwen_vl_interface': 6.345447555335236e-07, 'learning_rate/action_model': 6.3454475553352345e-06, 'epoch': 2.94}) +[2026-06-18 02:38:35,873][starVLA.training.train_starvla][INFO] - Step 6477, Loss: {'train/loss': 0.424237459897995, 'train/grad_norm_pre_clip': 49.715736389160156, 'timing/data': 0.0004938086494803429, 'timing/model': 0.9167434293776751, 'learning_rate/qwen_vl_interface': 6.340341424961072e-07, 'learning_rate/action_model': 6.340341424961071e-06, 'epoch': 2.94}) +[2026-06-18 02:38:39,584][starVLA.training.train_starvla][INFO] - Step 6478, Loss: {'train/loss': 0.4338605999946594, 'train/grad_norm_pre_clip': 35.06383514404297, 'timing/data': 0.0005960259586572647, 'timing/model': 0.9202224919572473, 'learning_rate/qwen_vl_interface': 6.335244863533802e-07, 'learning_rate/action_model': 6.3352448635338016e-06, 'epoch': 2.94}) +[2026-06-18 02:38:43,334][starVLA.training.train_starvla][INFO] - Step 6479, Loss: {'train/loss': 0.280428409576416, 'train/grad_norm_pre_clip': 31.760339736938477, 'timing/data': 0.0005148928612470627, 'timing/model': 0.9212733609601855, 'learning_rate/qwen_vl_interface': 6.330157872109953e-07, 'learning_rate/action_model': 6.330157872109953e-06, 'epoch': 2.94}) +[2026-06-18 02:38:47,053][starVLA.training.train_starvla][INFO] - Step 6480, Loss: {'train/loss': 0.25759220123291016, 'train/grad_norm_pre_clip': 14.691232681274414, 'timing/data': 0.0004805438220500946, 'timing/model': 0.912719214335084, 'learning_rate/qwen_vl_interface': 6.325080451744053e-07, 'learning_rate/action_model': 6.325080451744053e-06, 'epoch': 2.94}) +[2026-06-18 02:38:50,803][starVLA.training.train_starvla][INFO] - Step 6481, Loss: {'train/loss': 0.10365265607833862, 'train/grad_norm_pre_clip': 30.308883666992188, 'timing/data': 0.00048622768372297287, 'timing/model': 0.917607001028955, 'learning_rate/qwen_vl_interface': 6.320012603488663e-07, 'learning_rate/action_model': 6.320012603488662e-06, 'epoch': 2.94}) +[2026-06-18 02:38:54,597][starVLA.training.train_starvla][INFO] - Step 6482, Loss: {'train/loss': 0.18481341004371643, 'train/grad_norm_pre_clip': 32.16184997558594, 'timing/data': 0.00047834962606430054, 'timing/model': 1.1271371627226472, 'learning_rate/qwen_vl_interface': 6.314954328394351e-07, 'learning_rate/action_model': 6.3149543283943515e-06, 'epoch': 2.94}) +[2026-06-18 02:38:58,211][starVLA.training.train_starvla][INFO] - Step 6483, Loss: {'train/loss': 0.3042945861816406, 'train/grad_norm_pre_clip': 52.957401275634766, 'timing/data': 0.0013387426733970642, 'timing/model': 0.9366111820563674, 'learning_rate/qwen_vl_interface': 6.309905627509707e-07, 'learning_rate/action_model': 6.309905627509706e-06, 'epoch': 2.94}) +[2026-06-18 02:39:01,954][starVLA.training.train_starvla][INFO] - Step 6484, Loss: {'train/loss': 0.34022143483161926, 'train/grad_norm_pre_clip': 41.33216857910156, 'timing/data': 0.0005003102123737335, 'timing/model': 0.9242437798529863, 'learning_rate/qwen_vl_interface': 6.304866501881327e-07, 'learning_rate/action_model': 6.304866501881327e-06, 'epoch': 2.94}) +[2026-06-18 02:39:07,695][starVLA.training.train_starvla][INFO] - Step 6485, Loss: {'train/loss': 0.44454583525657654, 'train/grad_norm_pre_clip': 37.78867721557617, 'timing/data': 0.0007534902542829514, 'timing/model': 0.9527906430885196, 'learning_rate/qwen_vl_interface': 6.299836952553831e-07, 'learning_rate/action_model': 6.299836952553831e-06, 'epoch': 2.94}) +[2026-06-18 02:39:12,007][starVLA.training.train_starvla][INFO] - Step 6486, Loss: {'train/loss': 0.21597033739089966, 'train/grad_norm_pre_clip': 46.91450500488281, 'timing/data': 0.0011169658973813057, 'timing/model': 1.056995514780283, 'learning_rate/qwen_vl_interface': 6.294816980569845e-07, 'learning_rate/action_model': 6.294816980569845e-06, 'epoch': 2.94}) +[2026-06-18 02:39:15,963][starVLA.training.train_starvla][INFO] - Step 6487, Loss: {'train/loss': 0.3580937087535858, 'train/grad_norm_pre_clip': 23.221790313720703, 'timing/data': 0.0006323475390672684, 'timing/model': 0.95363097358495, 'learning_rate/qwen_vl_interface': 6.289806586970021e-07, 'learning_rate/action_model': 6.289806586970021e-06, 'epoch': 2.94}) +[2026-06-18 02:39:19,803][starVLA.training.train_starvla][INFO] - Step 6488, Loss: {'train/loss': 0.029337914660573006, 'train/grad_norm_pre_clip': 7.310654163360596, 'timing/data': 0.0004995297640562057, 'timing/model': 1.119784009642899, 'learning_rate/qwen_vl_interface': 6.284805772793011e-07, 'learning_rate/action_model': 6.284805772793011e-06, 'epoch': 2.94}) +[2026-06-18 02:39:23,702][starVLA.training.train_starvla][INFO] - Step 6489, Loss: {'train/loss': 0.4197643995285034, 'train/grad_norm_pre_clip': 33.043540954589844, 'timing/data': 0.0005232812836766243, 'timing/model': 1.1206345465034246, 'learning_rate/qwen_vl_interface': 6.279814539075499e-07, 'learning_rate/action_model': 6.2798145390754985e-06, 'epoch': 2.94}) +[2026-06-18 02:39:27,424][starVLA.training.train_starvla][INFO] - Step 6490, Loss: {'train/loss': 0.14316141605377197, 'train/grad_norm_pre_clip': 22.511178970336914, 'timing/data': 0.0006203306838870049, 'timing/model': 0.9370524873957038, 'learning_rate/qwen_vl_interface': 6.274832886852172e-07, 'learning_rate/action_model': 6.2748328868521724e-06, 'epoch': 2.94}) +[2026-06-18 02:39:31,230][starVLA.training.train_starvla][INFO] - Step 6491, Loss: {'train/loss': 0.11498268693685532, 'train/grad_norm_pre_clip': 22.03310203552246, 'timing/data': 0.0005324585363268852, 'timing/model': 0.9458373952656984, 'learning_rate/qwen_vl_interface': 6.269860817155721e-07, 'learning_rate/action_model': 6.269860817155721e-06, 'epoch': 2.94}) +[2026-06-18 02:39:35,263][starVLA.training.train_starvla][INFO] - Step 6492, Loss: {'train/loss': 0.12082548439502716, 'train/grad_norm_pre_clip': 25.532310485839844, 'timing/data': 0.0007890118286013603, 'timing/model': 0.9270618250593543, 'learning_rate/qwen_vl_interface': 6.264898331016865e-07, 'learning_rate/action_model': 6.264898331016865e-06, 'epoch': 2.95}) +[2026-06-18 02:39:39,020][starVLA.training.train_starvla][INFO] - Step 6493, Loss: {'train/loss': 0.1569523960351944, 'train/grad_norm_pre_clip': 27.80207061767578, 'timing/data': 0.0008316356688737869, 'timing/model': 0.9508010288700461, 'learning_rate/qwen_vl_interface': 6.259945429464343e-07, 'learning_rate/action_model': 6.259945429464342e-06, 'epoch': 2.95}) +[2026-06-18 02:39:42,904][starVLA.training.train_starvla][INFO] - Step 6494, Loss: {'train/loss': 0.1492508351802826, 'train/grad_norm_pre_clip': 16.71839714050293, 'timing/data': 0.0004877867177128792, 'timing/model': 0.9258712409064174, 'learning_rate/qwen_vl_interface': 6.255002113524885e-07, 'learning_rate/action_model': 6.255002113524886e-06, 'epoch': 2.95}) +[2026-06-18 02:39:46,698][starVLA.training.train_starvla][INFO] - Step 6495, Loss: {'train/loss': 0.2779805362224579, 'train/grad_norm_pre_clip': 23.86944580078125, 'timing/data': 0.0004955586045980453, 'timing/model': 1.1148489164188504, 'learning_rate/qwen_vl_interface': 6.250068384223255e-07, 'learning_rate/action_model': 6.250068384223255e-06, 'epoch': 2.95}) +[2026-06-18 02:39:50,460][starVLA.training.train_starvla][INFO] - Step 6496, Loss: {'train/loss': 0.18925093114376068, 'train/grad_norm_pre_clip': 21.387046813964844, 'timing/data': 0.0013369163498282433, 'timing/model': 1.0639698691666126, 'learning_rate/qwen_vl_interface': 6.245144242582214e-07, 'learning_rate/action_model': 6.245144242582214e-06, 'epoch': 2.95}) +[2026-06-18 02:39:54,091][starVLA.training.train_starvla][INFO] - Step 6497, Loss: {'train/loss': 0.07436773926019669, 'train/grad_norm_pre_clip': 16.40045738220215, 'timing/data': 0.0009219339117407799, 'timing/model': 0.9729581233114004, 'learning_rate/qwen_vl_interface': 6.240229689622546e-07, 'learning_rate/action_model': 6.240229689622545e-06, 'epoch': 2.95}) +[2026-06-18 02:39:57,994][starVLA.training.train_starvla][INFO] - Step 6498, Loss: {'train/loss': 0.10380243510007858, 'train/grad_norm_pre_clip': 42.70162582397461, 'timing/data': 0.000490841455757618, 'timing/model': 0.9573935819789767, 'learning_rate/qwen_vl_interface': 6.235324726363042e-07, 'learning_rate/action_model': 6.235324726363041e-06, 'epoch': 2.95}) +[2026-06-18 02:40:04,146][starVLA.training.train_starvla][INFO] - Step 6499, Loss: {'train/loss': 0.1372462958097458, 'train/grad_norm_pre_clip': 22.942062377929688, 'timing/data': 0.001127549447119236, 'timing/model': 0.9518513847142458, 'learning_rate/qwen_vl_interface': 6.230429353820503e-07, 'learning_rate/action_model': 6.230429353820503e-06, 'epoch': 2.95}) +[2026-06-18 02:42:12,083][starVLA.training.train_starvla][INFO] - Step 6500, Loss: {'train/loss': 0.20789377391338348, 'train/grad_norm_pre_clip': 30.078920364379883, 'eval/loss': 0.2495536208152771, 'eval/latency_8/loss': 0.2495536208152771, 'eval/demon_fire_precision': 0.9731812577065352, 'eval/demon_fire_recall': 0.9807393600497049, 'eval/demon_fire_f1': 0.9769456908556398, 'eval/demon_fire_rate_model': 0.902615470228158, 'eval/demon_fire_rate_teacher': 0.8956594323873122, 'eval/demon_left_precision': 0.9106280193236715, 'eval/demon_left_recall': 0.8072805139186295, 'eval/demon_left_f1': 0.8558456299659478, 'eval/demon_left_rate_model': 0.11519198664440734, 'eval/demon_left_rate_teacher': 0.12993878686700056, 'eval/demon_right_precision': 0.8602941176470589, 'eval/demon_right_recall': 0.8863636363636364, 'eval/demon_right_f1': 0.873134328358209, 'eval/demon_right_rate_model': 0.18920422927100725, 'eval/demon_right_rate_teacher': 0.18363939899833054, 'eval/demon_move_f1': 0.8644899791620784, 'eval/demon_cc_f1': 0.9402523514633733, 'eval/latency_8/demon_fire_precision': 0.9731812577065352, 'eval/latency_8/demon_fire_recall': 0.9807393600497049, 'eval/latency_8/demon_fire_f1': 0.9769456908556398, 'eval/latency_8/demon_fire_rate_model': 0.902615470228158, 'eval/latency_8/demon_fire_rate_teacher': 0.8956594323873122, 'eval/latency_8/demon_left_precision': 0.9106280193236715, 'eval/latency_8/demon_left_recall': 0.8072805139186295, 'eval/latency_8/demon_left_f1': 0.8558456299659478, 'eval/latency_8/demon_left_rate_model': 0.11519198664440734, 'eval/latency_8/demon_left_rate_teacher': 0.12993878686700056, 'eval/latency_8/demon_right_precision': 0.8602941176470589, 'eval/latency_8/demon_right_recall': 0.8863636363636364, 'eval/latency_8/demon_right_f1': 0.873134328358209, 'eval/latency_8/demon_right_rate_model': 0.18920422927100725, 'eval/latency_8/demon_right_rate_teacher': 0.18363939899833054, 'eval/latency_8/demon_move_f1': 0.8644899791620784, 'eval/latency_8/demon_cc_f1': 0.9402523514633733, 'timing/data': 0.0007073506712913513, 'timing/model': 0.9327803608030081, 'learning_rate/qwen_vl_interface': 6.225543573009747e-07, 'learning_rate/action_model': 6.225543573009748e-06, 'epoch': 2.95}) +[2026-06-18 02:42:12,091][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_6500_state +[2026-06-18 02:42:12,096][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 02:42:18,448][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_6500_state/model.safetensors +[2026-06-18 02:42:33,786][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_6500_state/optimizer.bin +[2026-06-18 02:42:33,788][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_6500_state/sampler.bin +[2026-06-18 02:42:33,788][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_6500_state/sampler_1.bin +[2026-06-18 02:42:33,791][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_6500_state/random_states_0.pkl +[2026-06-18 02:42:33,794][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 6500 +[2026-06-18 02:42:33,795][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 02:42:33,806][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 02:42:37,596][starVLA.training.train_starvla][INFO] - Step 6501, Loss: {'train/loss': 0.025137796998023987, 'train/grad_norm_pre_clip': 7.8700270652771, 'timing/data': 0.0004949085414409637, 'timing/model': 0.9231430292129517, 'learning_rate/qwen_vl_interface': 6.2206673849436e-07, 'learning_rate/action_model': 6.2206673849436e-06, 'epoch': 2.95}) +[2026-06-18 02:42:41,291][starVLA.training.train_starvla][INFO] - Step 6502, Loss: {'train/loss': 0.5603075623512268, 'train/grad_norm_pre_clip': 58.62620162963867, 'timing/data': 0.00046901870518922806, 'timing/model': 0.9112794864922762, 'learning_rate/qwen_vl_interface': 6.215800790632905e-07, 'learning_rate/action_model': 6.215800790632904e-06, 'epoch': 2.95}) +[2026-06-18 02:42:45,282][starVLA.training.train_starvla][INFO] - Step 6503, Loss: {'train/loss': 0.2456607222557068, 'train/grad_norm_pre_clip': 28.389225006103516, 'timing/data': 0.0009469659999012947, 'timing/model': 0.9203327922150493, 'learning_rate/qwen_vl_interface': 6.210943791086506e-07, 'learning_rate/action_model': 6.210943791086506e-06, 'epoch': 2.95}) +[2026-06-18 02:42:49,263][starVLA.training.train_starvla][INFO] - Step 6504, Loss: {'train/loss': 0.11781655251979828, 'train/grad_norm_pre_clip': 25.343849182128906, 'timing/data': 0.0007061930373311043, 'timing/model': 1.251744986511767, 'learning_rate/qwen_vl_interface': 6.206096387311278e-07, 'learning_rate/action_model': 6.206096387311277e-06, 'epoch': 2.95}) +[2026-06-18 02:42:52,957][starVLA.training.train_starvla][INFO] - Step 6505, Loss: {'train/loss': 0.3377925753593445, 'train/grad_norm_pre_clip': 57.00529479980469, 'timing/data': 0.0004897722974419594, 'timing/model': 0.9250678978860378, 'learning_rate/qwen_vl_interface': 6.201258580312078e-07, 'learning_rate/action_model': 6.201258580312077e-06, 'epoch': 2.95}) +[2026-06-18 02:42:56,957][starVLA.training.train_starvla][INFO] - Step 6506, Loss: {'train/loss': 0.03863593190908432, 'train/grad_norm_pre_clip': 6.437044143676758, 'timing/data': 0.0005155354738235474, 'timing/model': 0.9310778984799981, 'learning_rate/qwen_vl_interface': 6.196430371091792e-07, 'learning_rate/action_model': 6.196430371091792e-06, 'epoch': 2.95}) +[2026-06-18 02:43:00,836][starVLA.training.train_starvla][INFO] - Step 6507, Loss: {'train/loss': 0.23642149567604065, 'train/grad_norm_pre_clip': 43.72136688232422, 'timing/data': 0.0005013216286897659, 'timing/model': 0.9367155758664012, 'learning_rate/qwen_vl_interface': 6.191611760651312e-07, 'learning_rate/action_model': 6.191611760651312e-06, 'epoch': 2.95}) +[2026-06-18 02:43:04,635][starVLA.training.train_starvla][INFO] - Step 6508, Loss: {'train/loss': 0.42121461033821106, 'train/grad_norm_pre_clip': 50.31100845336914, 'timing/data': 0.0007797451689839363, 'timing/model': 0.9445170238614082, 'learning_rate/qwen_vl_interface': 6.18680274998955e-07, 'learning_rate/action_model': 6.186802749989549e-06, 'epoch': 2.95}) +[2026-06-18 02:43:08,538][starVLA.training.train_starvla][INFO] - Step 6509, Loss: {'train/loss': 0.4052567780017853, 'train/grad_norm_pre_clip': 33.03094482421875, 'timing/data': 0.0005249064415693283, 'timing/model': 0.9555778056383133, 'learning_rate/qwen_vl_interface': 6.182003340103413e-07, 'learning_rate/action_model': 6.1820033401034135e-06, 'epoch': 2.95}) +[2026-06-18 02:43:12,523][starVLA.training.train_starvla][INFO] - Step 6510, Loss: {'train/loss': 0.056518521159887314, 'train/grad_norm_pre_clip': 14.664589881896973, 'timing/data': 0.0004901271313428879, 'timing/model': 1.168168567121029, 'learning_rate/qwen_vl_interface': 6.177213531987819e-07, 'learning_rate/action_model': 6.177213531987819e-06, 'epoch': 2.95}) +[2026-06-18 02:43:16,127][starVLA.training.train_starvla][INFO] - Step 6511, Loss: {'train/loss': 0.13878895342350006, 'train/grad_norm_pre_clip': 35.031681060791016, 'timing/data': 0.00048162415623664856, 'timing/model': 0.9139886647462845, 'learning_rate/qwen_vl_interface': 6.172433326635714e-07, 'learning_rate/action_model': 6.172433326635713e-06, 'epoch': 2.95}) +[2026-06-18 02:43:19,881][starVLA.training.train_starvla][INFO] - Step 6512, Loss: {'train/loss': 0.1953427493572235, 'train/grad_norm_pre_clip': 21.611316680908203, 'timing/data': 0.0006048362702131271, 'timing/model': 0.9311601389199495, 'learning_rate/qwen_vl_interface': 6.167662725038025e-07, 'learning_rate/action_model': 6.1676627250380255e-06, 'epoch': 2.95}) +[2026-06-18 02:43:23,990][starVLA.training.train_starvla][INFO] - Step 6513, Loss: {'train/loss': 0.14238670468330383, 'train/grad_norm_pre_clip': 39.06808853149414, 'timing/data': 0.0004851343110203743, 'timing/model': 1.1549517512321472, 'learning_rate/qwen_vl_interface': 6.16290172818371e-07, 'learning_rate/action_model': 6.16290172818371e-06, 'epoch': 2.95}) +[2026-06-18 02:43:27,804][starVLA.training.train_starvla][INFO] - Step 6514, Loss: {'train/loss': 0.18235228955745697, 'train/grad_norm_pre_clip': 17.461606979370117, 'timing/data': 0.0004717698320746422, 'timing/model': 0.9506381377577782, 'learning_rate/qwen_vl_interface': 6.158150337059725e-07, 'learning_rate/action_model': 6.158150337059725e-06, 'epoch': 2.96}) +[2026-06-18 02:43:33,247][starVLA.training.train_starvla][INFO] - Step 6515, Loss: {'train/loss': 0.2878080904483795, 'train/grad_norm_pre_clip': 35.825050354003906, 'timing/data': 0.0005134707316756248, 'timing/model': 0.9366250475868583, 'learning_rate/qwen_vl_interface': 6.153408552651045e-07, 'learning_rate/action_model': 6.153408552651045e-06, 'epoch': 2.96}) +[2026-06-18 02:43:37,061][starVLA.training.train_starvla][INFO] - Step 6516, Loss: {'train/loss': 0.4740126430988312, 'train/grad_norm_pre_clip': 34.258575439453125, 'timing/data': 0.00044044945389032364, 'timing/model': 0.9270476195961237, 'learning_rate/qwen_vl_interface': 6.14867637594064e-07, 'learning_rate/action_model': 6.14867637594064e-06, 'epoch': 2.96}) +[2026-06-18 02:43:42,083][starVLA.training.train_starvla][INFO] - Step 6517, Loss: {'train/loss': 0.02481316588819027, 'train/grad_norm_pre_clip': 5.622711658477783, 'timing/data': 0.0005090627819299698, 'timing/model': 2.392412167042494, 'learning_rate/qwen_vl_interface': 6.143953807909497e-07, 'learning_rate/action_model': 6.143953807909496e-06, 'epoch': 2.96}) +[2026-06-18 02:43:45,907][starVLA.training.train_starvla][INFO] - Step 6518, Loss: {'train/loss': 0.06030905619263649, 'train/grad_norm_pre_clip': 14.309131622314453, 'timing/data': 0.000492805615067482, 'timing/model': 0.9537292774766684, 'learning_rate/qwen_vl_interface': 6.139240849536616e-07, 'learning_rate/action_model': 6.1392408495366156e-06, 'epoch': 2.96}) +[2026-06-18 02:43:49,742][starVLA.training.train_starvla][INFO] - Step 6519, Loss: {'train/loss': 0.1322583109140396, 'train/grad_norm_pre_clip': 15.424535751342773, 'timing/data': 0.0010092342272400856, 'timing/model': 0.9192285453900695, 'learning_rate/qwen_vl_interface': 6.134537501798993e-07, 'learning_rate/action_model': 6.1345375017989925e-06, 'epoch': 2.96}) +[2026-06-18 02:43:53,522][starVLA.training.train_starvla][INFO] - Step 6520, Loss: {'train/loss': 0.05601271986961365, 'train/grad_norm_pre_clip': 15.541192054748535, 'timing/data': 0.0015776818618178368, 'timing/model': 0.926030857488513, 'learning_rate/qwen_vl_interface': 6.129843765671634e-07, 'learning_rate/action_model': 6.129843765671633e-06, 'epoch': 2.96}) +[2026-06-18 02:43:57,449][starVLA.training.train_starvla][INFO] - Step 6521, Loss: {'train/loss': 0.1655299961566925, 'train/grad_norm_pre_clip': 26.677688598632812, 'timing/data': 0.0011503798887133598, 'timing/model': 0.9245641995221376, 'learning_rate/qwen_vl_interface': 6.125159642127559e-07, 'learning_rate/action_model': 6.125159642127558e-06, 'epoch': 2.96}) +[2026-06-18 02:44:01,294][starVLA.training.train_starvla][INFO] - Step 6522, Loss: {'train/loss': 0.12086119502782822, 'train/grad_norm_pre_clip': 33.12516403198242, 'timing/data': 0.0006536915898323059, 'timing/model': 0.9322434859350324, 'learning_rate/qwen_vl_interface': 6.120485132137793e-07, 'learning_rate/action_model': 6.120485132137794e-06, 'epoch': 2.96}) +[2026-06-18 02:44:05,431][starVLA.training.train_starvla][INFO] - Step 6523, Loss: {'train/loss': 0.01750156469643116, 'train/grad_norm_pre_clip': 3.2117116451263428, 'timing/data': 0.000489051453769207, 'timing/model': 1.164231994189322, 'learning_rate/qwen_vl_interface': 6.115820236671369e-07, 'learning_rate/action_model': 6.115820236671369e-06, 'epoch': 2.96}) +[2026-06-18 02:44:09,249][starVLA.training.train_starvla][INFO] - Step 6524, Loss: {'train/loss': 0.392304927110672, 'train/grad_norm_pre_clip': 64.89600372314453, 'timing/data': 0.0006117643788456917, 'timing/model': 1.125199617817998, 'learning_rate/qwen_vl_interface': 6.111164956695316e-07, 'learning_rate/action_model': 6.111164956695316e-06, 'epoch': 2.96}) +[2026-06-18 02:44:12,959][starVLA.training.train_starvla][INFO] - Step 6525, Loss: {'train/loss': 0.054684434086084366, 'train/grad_norm_pre_clip': 10.359344482421875, 'timing/data': 0.0005225222557783127, 'timing/model': 0.9784564720466733, 'learning_rate/qwen_vl_interface': 6.106519293174684e-07, 'learning_rate/action_model': 6.106519293174684e-06, 'epoch': 2.96}) +[2026-06-18 02:44:16,913][starVLA.training.train_starvla][INFO] - Step 6526, Loss: {'train/loss': 0.1241743192076683, 'train/grad_norm_pre_clip': 21.94989013671875, 'timing/data': 0.00046970415860414505, 'timing/model': 1.039323851466179, 'learning_rate/qwen_vl_interface': 6.101883247072525e-07, 'learning_rate/action_model': 6.1018832470725245e-06, 'epoch': 2.96}) +[2026-06-18 02:44:20,781][starVLA.training.train_starvla][INFO] - Step 6527, Loss: {'train/loss': 0.15578585863113403, 'train/grad_norm_pre_clip': 26.494779586791992, 'timing/data': 0.0004964759573340416, 'timing/model': 0.9232256840914488, 'learning_rate/qwen_vl_interface': 6.097256819349899e-07, 'learning_rate/action_model': 6.0972568193498985e-06, 'epoch': 2.96}) +[2026-06-18 02:44:25,946][starVLA.training.train_starvla][INFO] - Step 6528, Loss: {'train/loss': 0.0331992506980896, 'train/grad_norm_pre_clip': 14.05036735534668, 'timing/data': 0.0010624006390571594, 'timing/model': 2.2614690586924553, 'learning_rate/qwen_vl_interface': 6.092640010965864e-07, 'learning_rate/action_model': 6.092640010965864e-06, 'epoch': 2.96}) +[2026-06-18 02:44:33,810][starVLA.training.train_starvla][INFO] - Step 6529, Loss: {'train/loss': 0.08850017935037613, 'train/grad_norm_pre_clip': 26.412174224853516, 'timing/data': 0.0004883846268057823, 'timing/model': 0.9307988500222564, 'learning_rate/qwen_vl_interface': 6.088032822877488e-07, 'learning_rate/action_model': 6.088032822877488e-06, 'epoch': 2.96}) +[2026-06-18 02:44:39,260][starVLA.training.train_starvla][INFO] - Step 6530, Loss: {'train/loss': 0.22738058865070343, 'train/grad_norm_pre_clip': 33.094451904296875, 'timing/data': 0.0005342531949281693, 'timing/model': 1.187682406976819, 'learning_rate/qwen_vl_interface': 6.083435256039854e-07, 'learning_rate/action_model': 6.083435256039854e-06, 'epoch': 2.96}) +[2026-06-18 02:44:42,980][starVLA.training.train_starvla][INFO] - Step 6531, Loss: {'train/loss': 0.026889268308877945, 'train/grad_norm_pre_clip': 6.170071601867676, 'timing/data': 0.0005307001993060112, 'timing/model': 1.0705333286896348, 'learning_rate/qwen_vl_interface': 6.078847311406033e-07, 'learning_rate/action_model': 6.078847311406032e-06, 'epoch': 2.96}) +[2026-06-18 02:44:46,573][starVLA.training.train_starvla][INFO] - Step 6532, Loss: {'train/loss': 0.05182885006070137, 'train/grad_norm_pre_clip': 11.52718448638916, 'timing/data': 0.0004856772720813751, 'timing/model': 0.9223014926537871, 'learning_rate/qwen_vl_interface': 6.07426898992712e-07, 'learning_rate/action_model': 6.074268989927119e-06, 'epoch': 2.96}) +[2026-06-18 02:44:50,542][starVLA.training.train_starvla][INFO] - Step 6533, Loss: {'train/loss': 0.0906921923160553, 'train/grad_norm_pre_clip': 30.71463394165039, 'timing/data': 0.0008174227550625801, 'timing/model': 0.918251265771687, 'learning_rate/qwen_vl_interface': 6.069700292552202e-07, 'learning_rate/action_model': 6.069700292552202e-06, 'epoch': 2.96}) +[2026-06-18 02:44:54,323][starVLA.training.train_starvla][INFO] - Step 6534, Loss: {'train/loss': 0.7087558507919312, 'train/grad_norm_pre_clip': 37.91632843017578, 'timing/data': 0.000506572425365448, 'timing/model': 0.9406558545306325, 'learning_rate/qwen_vl_interface': 6.065141220228372e-07, 'learning_rate/action_model': 6.065141220228372e-06, 'epoch': 2.96}) +[2026-06-18 02:44:58,058][starVLA.training.train_starvla][INFO] - Step 6535, Loss: {'train/loss': 0.4087759256362915, 'train/grad_norm_pre_clip': 31.73150062561035, 'timing/data': 0.00046915002167224884, 'timing/model': 0.9271924151107669, 'learning_rate/qwen_vl_interface': 6.060591773900732e-07, 'learning_rate/action_model': 6.060591773900733e-06, 'epoch': 2.96}) +[2026-06-18 02:45:02,013][starVLA.training.train_starvla][INFO] - Step 6536, Loss: {'train/loss': 0.02260042540729046, 'train/grad_norm_pre_clip': 5.133965492248535, 'timing/data': 0.0005193054676055908, 'timing/model': 1.0369669403880835, 'learning_rate/qwen_vl_interface': 6.056051954512388e-07, 'learning_rate/action_model': 6.0560519545123875e-06, 'epoch': 2.97}) +[2026-06-18 02:45:05,883][starVLA.training.train_starvla][INFO] - Step 6537, Loss: {'train/loss': 0.2667447030544281, 'train/grad_norm_pre_clip': 24.701526641845703, 'timing/data': 0.0006577270105481148, 'timing/model': 1.134629663079977, 'learning_rate/qwen_vl_interface': 6.051521763004455e-07, 'learning_rate/action_model': 6.051521763004456e-06, 'epoch': 2.97}) +[2026-06-18 02:45:09,483][starVLA.training.train_starvla][INFO] - Step 6538, Loss: {'train/loss': 0.15720918774604797, 'train/grad_norm_pre_clip': 17.66136360168457, 'timing/data': 0.0004679309204220772, 'timing/model': 0.9199864063411951, 'learning_rate/qwen_vl_interface': 6.047001200316038e-07, 'learning_rate/action_model': 6.047001200316037e-06, 'epoch': 2.97}) +[2026-06-18 02:45:13,526][starVLA.training.train_starvla][INFO] - Step 6539, Loss: {'train/loss': 0.3940292000770569, 'train/grad_norm_pre_clip': 59.80788040161133, 'timing/data': 0.00047383736819028854, 'timing/model': 0.9591069053858519, 'learning_rate/qwen_vl_interface': 6.042490267384255e-07, 'learning_rate/action_model': 6.0424902673842545e-06, 'epoch': 2.97}) +[2026-06-18 02:45:17,357][starVLA.training.train_starvla][INFO] - Step 6540, Loss: {'train/loss': 0.1781468689441681, 'train/grad_norm_pre_clip': 30.74122428894043, 'timing/data': 0.0004904149100184441, 'timing/model': 0.9732474135234952, 'learning_rate/qwen_vl_interface': 6.03798896514423e-07, 'learning_rate/action_model': 6.0379889651442295e-06, 'epoch': 2.97}) +[2026-06-18 02:45:21,497][starVLA.training.train_starvla][INFO] - Step 6541, Loss: {'train/loss': 0.09585299342870712, 'train/grad_norm_pre_clip': 20.476530075073242, 'timing/data': 0.0005197701975703239, 'timing/model': 1.2207181435078382, 'learning_rate/qwen_vl_interface': 6.033497294529093e-07, 'learning_rate/action_model': 6.033497294529093e-06, 'epoch': 2.97}) +[2026-06-18 02:45:25,281][starVLA.training.train_starvla][INFO] - Step 6542, Loss: {'train/loss': 0.13499732315540314, 'train/grad_norm_pre_clip': 23.540489196777344, 'timing/data': 0.0004923595115542412, 'timing/model': 0.9261206639930606, 'learning_rate/qwen_vl_interface': 6.029015256469967e-07, 'learning_rate/action_model': 6.0290152564699666e-06, 'epoch': 2.97}) +[2026-06-18 02:45:29,086][starVLA.training.train_starvla][INFO] - Step 6543, Loss: {'train/loss': 0.08185526728630066, 'train/grad_norm_pre_clip': 20.51644515991211, 'timing/data': 0.000707150436937809, 'timing/model': 0.9150145621970296, 'learning_rate/qwen_vl_interface': 6.024542851895976e-07, 'learning_rate/action_model': 6.0245428518959756e-06, 'epoch': 2.97}) +[2026-06-18 02:45:32,832][starVLA.training.train_starvla][INFO] - Step 6544, Loss: {'train/loss': 0.26158997416496277, 'train/grad_norm_pre_clip': 59.289794921875, 'timing/data': 0.0004717539995908737, 'timing/model': 1.1269018156453967, 'learning_rate/qwen_vl_interface': 6.020080081734267e-07, 'learning_rate/action_model': 6.020080081734266e-06, 'epoch': 2.97}) +[2026-06-18 02:45:36,404][starVLA.training.train_starvla][INFO] - Step 6545, Loss: {'train/loss': 0.1494373083114624, 'train/grad_norm_pre_clip': 30.767589569091797, 'timing/data': 0.0005383724346756935, 'timing/model': 0.9133986365050077, 'learning_rate/qwen_vl_interface': 6.015626946909974e-07, 'learning_rate/action_model': 6.015626946909973e-06, 'epoch': 2.97}) +[2026-06-18 02:45:40,213][starVLA.training.train_starvla][INFO] - Step 6546, Loss: {'train/loss': 0.2134406566619873, 'train/grad_norm_pre_clip': 27.13030242919922, 'timing/data': 0.0004822099581360817, 'timing/model': 0.9477813458070159, 'learning_rate/qwen_vl_interface': 6.011183448346233e-07, 'learning_rate/action_model': 6.011183448346233e-06, 'epoch': 2.97}) +[2026-06-18 02:45:44,024][starVLA.training.train_starvla][INFO] - Step 6547, Loss: {'train/loss': 0.0686296746134758, 'train/grad_norm_pre_clip': 20.439027786254883, 'timing/data': 0.0004614153876900673, 'timing/model': 0.9457718338817358, 'learning_rate/qwen_vl_interface': 6.006749586964183e-07, 'learning_rate/action_model': 6.006749586964183e-06, 'epoch': 2.97}) +[2026-06-18 02:45:47,842][starVLA.training.train_starvla][INFO] - Step 6548, Loss: {'train/loss': 0.19847610592842102, 'train/grad_norm_pre_clip': 23.331621170043945, 'timing/data': 0.0007551712915301323, 'timing/model': 0.9393491093069315, 'learning_rate/qwen_vl_interface': 6.002325363682974e-07, 'learning_rate/action_model': 6.0023253636829736e-06, 'epoch': 2.97}) +[2026-06-18 02:45:51,688][starVLA.training.train_starvla][INFO] - Step 6549, Loss: {'train/loss': 0.11668732762336731, 'train/grad_norm_pre_clip': 20.635221481323242, 'timing/data': 0.0005835527554154396, 'timing/model': 0.9313476514071226, 'learning_rate/qwen_vl_interface': 5.997910779419756e-07, 'learning_rate/action_model': 5.9979107794197565e-06, 'epoch': 2.97}) +[2026-06-18 02:45:55,872][starVLA.training.train_starvla][INFO] - Step 6550, Loss: {'train/loss': 0.05042020231485367, 'train/grad_norm_pre_clip': 14.979218482971191, 'timing/data': 0.00048733968287706375, 'timing/model': 1.4918499365448952, 'learning_rate/qwen_vl_interface': 5.993505835089661e-07, 'learning_rate/action_model': 5.993505835089661e-06, 'epoch': 2.97}) +[2026-06-18 02:45:59,706][starVLA.training.train_starvla][INFO] - Step 6551, Loss: {'train/loss': 0.43332457542419434, 'train/grad_norm_pre_clip': 62.55706787109375, 'timing/data': 0.0004779566079378128, 'timing/model': 1.0832294318825006, 'learning_rate/qwen_vl_interface': 5.98911053160586e-07, 'learning_rate/action_model': 5.989110531605859e-06, 'epoch': 2.97}) +[2026-06-18 02:46:03,394][starVLA.training.train_starvla][INFO] - Step 6552, Loss: {'train/loss': 0.16476623713970184, 'train/grad_norm_pre_clip': 18.972820281982422, 'timing/data': 0.0009028483182191849, 'timing/model': 0.964817170985043, 'learning_rate/qwen_vl_interface': 5.984724869879494e-07, 'learning_rate/action_model': 5.984724869879493e-06, 'epoch': 2.97}) +[2026-06-18 02:46:07,410][starVLA.training.train_starvla][INFO] - Step 6553, Loss: {'train/loss': 0.3681199252605438, 'train/grad_norm_pre_clip': 21.811899185180664, 'timing/data': 0.0005369698628783226, 'timing/model': 0.940275733359158, 'learning_rate/qwen_vl_interface': 5.980348850819711e-07, 'learning_rate/action_model': 5.98034885081971e-06, 'epoch': 2.97}) +[2026-06-18 02:46:12,787][starVLA.training.train_starvla][INFO] - Step 6554, Loss: {'train/loss': 0.06881040334701538, 'train/grad_norm_pre_clip': 30.794815063476562, 'timing/data': 0.0006050067022442818, 'timing/model': 0.9272201545536518, 'learning_rate/qwen_vl_interface': 5.975982475333664e-07, 'learning_rate/action_model': 5.975982475333663e-06, 'epoch': 2.97}) +[2026-06-18 02:46:17,065][starVLA.training.train_starvla][INFO] - Step 6555, Loss: {'train/loss': 0.037315890192985535, 'train/grad_norm_pre_clip': 9.816662788391113, 'timing/data': 0.000571434386074543, 'timing/model': 1.3245993936434388, 'learning_rate/qwen_vl_interface': 5.971625744326516e-07, 'learning_rate/action_model': 5.971625744326515e-06, 'epoch': 2.97}) +[2026-06-18 02:46:20,849][starVLA.training.train_starvla][INFO] - Step 6556, Loss: {'train/loss': 0.10557177662849426, 'train/grad_norm_pre_clip': 25.22342300415039, 'timing/data': 0.0004929304122924805, 'timing/model': 0.9188325712457299, 'learning_rate/qwen_vl_interface': 5.967278658701421e-07, 'learning_rate/action_model': 5.967278658701421e-06, 'epoch': 2.97}) +[2026-06-18 02:46:24,595][starVLA.training.train_starvla][INFO] - Step 6557, Loss: {'train/loss': 0.2598491907119751, 'train/grad_norm_pre_clip': 27.042844772338867, 'timing/data': 0.0005392273887991905, 'timing/model': 1.1011151634156704, 'learning_rate/qwen_vl_interface': 5.962941219359527e-07, 'learning_rate/action_model': 5.962941219359527e-06, 'epoch': 2.97}) +[2026-06-18 02:46:28,323][starVLA.training.train_starvla][INFO] - Step 6558, Loss: {'train/loss': 0.3017028570175171, 'train/grad_norm_pre_clip': 45.4657096862793, 'timing/data': 0.0004659360274672508, 'timing/model': 1.0438185995444655, 'learning_rate/qwen_vl_interface': 5.958613427199998e-07, 'learning_rate/action_model': 5.958613427199997e-06, 'epoch': 2.98}) +[2026-06-18 02:46:31,950][starVLA.training.train_starvla][INFO] - Step 6559, Loss: {'train/loss': 0.09071548283100128, 'train/grad_norm_pre_clip': 31.58388328552246, 'timing/data': 0.000980611890554428, 'timing/model': 0.9324183827266097, 'learning_rate/qwen_vl_interface': 5.954295283119985e-07, 'learning_rate/action_model': 5.954295283119985e-06, 'epoch': 2.98}) +[2026-06-18 02:46:36,491][starVLA.training.train_starvla][INFO] - Step 6560, Loss: {'train/loss': 0.17485977709293365, 'train/grad_norm_pre_clip': 42.72146224975586, 'timing/data': 0.0006869779899716377, 'timing/model': 1.0011780206114054, 'learning_rate/qwen_vl_interface': 5.949986788014643e-07, 'learning_rate/action_model': 5.949986788014643e-06, 'epoch': 2.98}) +[2026-06-18 02:46:41,031][starVLA.training.train_starvla][INFO] - Step 6561, Loss: {'train/loss': 0.01861034706234932, 'train/grad_norm_pre_clip': 4.772688865661621, 'timing/data': 0.0008154697716236115, 'timing/model': 1.3904127171263099, 'learning_rate/qwen_vl_interface': 5.94568794277713e-07, 'learning_rate/action_model': 5.94568794277713e-06, 'epoch': 2.98}) +[2026-06-18 02:46:44,948][starVLA.training.train_starvla][INFO] - Step 6562, Loss: {'train/loss': 0.031397078186273575, 'train/grad_norm_pre_clip': 13.955950736999512, 'timing/data': 0.0004990268498659134, 'timing/model': 0.9424723535776138, 'learning_rate/qwen_vl_interface': 5.941398748298603e-07, 'learning_rate/action_model': 5.941398748298603e-06, 'epoch': 2.98}) +[2026-06-18 02:46:48,742][starVLA.training.train_starvla][INFO] - Step 6563, Loss: {'train/loss': 0.09790575504302979, 'train/grad_norm_pre_clip': 25.806100845336914, 'timing/data': 0.0004987139254808426, 'timing/model': 0.9180418644100428, 'learning_rate/qwen_vl_interface': 5.937119205468217e-07, 'learning_rate/action_model': 5.9371192054682165e-06, 'epoch': 2.98}) +[2026-06-18 02:46:52,554][starVLA.training.train_starvla][INFO] - Step 6564, Loss: {'train/loss': 0.1642463356256485, 'train/grad_norm_pre_clip': 32.401939392089844, 'timing/data': 0.00048313383013010025, 'timing/model': 1.1436513122171164, 'learning_rate/qwen_vl_interface': 5.93284931517312e-07, 'learning_rate/action_model': 5.93284931517312e-06, 'epoch': 2.98}) +[2026-06-18 02:46:56,187][starVLA.training.train_starvla][INFO] - Step 6565, Loss: {'train/loss': 0.020740415900945663, 'train/grad_norm_pre_clip': 7.077453136444092, 'timing/data': 0.0004929937422275543, 'timing/model': 0.9366217693313956, 'learning_rate/qwen_vl_interface': 5.928589078298468e-07, 'learning_rate/action_model': 5.928589078298468e-06, 'epoch': 2.98}) +[2026-06-18 02:47:00,154][starVLA.training.train_starvla][INFO] - Step 6566, Loss: {'train/loss': 0.09115520119667053, 'train/grad_norm_pre_clip': 27.193735122680664, 'timing/data': 0.0005041938275098801, 'timing/model': 0.9236700171604753, 'learning_rate/qwen_vl_interface': 5.924338495727416e-07, 'learning_rate/action_model': 5.9243384957274155e-06, 'epoch': 2.98}) +[2026-06-18 02:47:03,929][starVLA.training.train_starvla][INFO] - Step 6567, Loss: {'train/loss': 0.22277261316776276, 'train/grad_norm_pre_clip': 30.39252471923828, 'timing/data': 0.001003064215183258, 'timing/model': 0.933017042465508, 'learning_rate/qwen_vl_interface': 5.92009756834111e-07, 'learning_rate/action_model': 5.9200975683411094e-06, 'epoch': 2.98}) +[2026-06-18 02:47:07,702][starVLA.training.train_starvla][INFO] - Step 6568, Loss: {'train/loss': 0.20033925771713257, 'train/grad_norm_pre_clip': 19.033140182495117, 'timing/data': 0.0007247868925333023, 'timing/model': 0.925127868540585, 'learning_rate/qwen_vl_interface': 5.915866297018697e-07, 'learning_rate/action_model': 5.915866297018696e-06, 'epoch': 2.98}) +[2026-06-18 02:47:11,441][starVLA.training.train_starvla][INFO] - Step 6569, Loss: {'train/loss': 0.0538501963019371, 'train/grad_norm_pre_clip': 15.815190315246582, 'timing/data': 0.0007649073377251625, 'timing/model': 0.9299188377335668, 'learning_rate/qwen_vl_interface': 5.911644682637329e-07, 'learning_rate/action_model': 5.911644682637329e-06, 'epoch': 2.98}) +[2026-06-18 02:47:15,192][starVLA.training.train_starvla][INFO] - Step 6570, Loss: {'train/loss': 0.08647888153791428, 'train/grad_norm_pre_clip': 17.80362892150879, 'timing/data': 0.0008737621828913689, 'timing/model': 1.0964695643633604, 'learning_rate/qwen_vl_interface': 5.907432726072149e-07, 'learning_rate/action_model': 5.9074327260721485e-06, 'epoch': 2.98}) +[2026-06-18 02:47:19,111][starVLA.training.train_starvla][INFO] - Step 6571, Loss: {'train/loss': 0.2055249959230423, 'train/grad_norm_pre_clip': 25.269176483154297, 'timing/data': 0.0008796434849500656, 'timing/model': 1.1727001806721091, 'learning_rate/qwen_vl_interface': 5.903230428196305e-07, 'learning_rate/action_model': 5.9032304281963045e-06, 'epoch': 2.98}) +[2026-06-18 02:47:22,805][starVLA.training.train_starvla][INFO] - Step 6572, Loss: {'train/loss': 0.20290027558803558, 'train/grad_norm_pre_clip': 19.63822364807129, 'timing/data': 0.0005651460960507393, 'timing/model': 0.955885617993772, 'learning_rate/qwen_vl_interface': 5.899037789880929e-07, 'learning_rate/action_model': 5.899037789880929e-06, 'epoch': 2.98}) +[2026-06-18 02:47:26,770][starVLA.training.train_starvla][INFO] - Step 6573, Loss: {'train/loss': 0.14156967401504517, 'train/grad_norm_pre_clip': 11.312894821166992, 'timing/data': 0.0005030892789363861, 'timing/model': 0.9404543824493885, 'learning_rate/qwen_vl_interface': 5.894854811995166e-07, 'learning_rate/action_model': 5.894854811995165e-06, 'epoch': 2.98}) +[2026-06-18 02:47:30,664][starVLA.training.train_starvla][INFO] - Step 6574, Loss: {'train/loss': 0.13958841562271118, 'train/grad_norm_pre_clip': 25.103839874267578, 'timing/data': 0.0007151979953050613, 'timing/model': 1.0323069663718343, 'learning_rate/qwen_vl_interface': 5.890681495406152e-07, 'learning_rate/action_model': 5.890681495406151e-06, 'epoch': 2.98}) +[2026-06-18 02:47:34,617][starVLA.training.train_starvla][INFO] - Step 6575, Loss: {'train/loss': 0.1353389173746109, 'train/grad_norm_pre_clip': 32.78341293334961, 'timing/data': 0.0010137343779206276, 'timing/model': 0.9769804263487458, 'learning_rate/qwen_vl_interface': 5.886517840979015e-07, 'learning_rate/action_model': 5.886517840979016e-06, 'epoch': 2.98}) +[2026-06-18 02:47:39,855][starVLA.training.train_starvla][INFO] - Step 6576, Loss: {'train/loss': 0.2113146334886551, 'train/grad_norm_pre_clip': 26.16851234436035, 'timing/data': 0.00046620890498161316, 'timing/model': 2.3807439180091023, 'learning_rate/qwen_vl_interface': 5.882363849576889e-07, 'learning_rate/action_model': 5.882363849576889e-06, 'epoch': 2.98}) +[2026-06-18 02:47:43,707][starVLA.training.train_starvla][INFO] - Step 6577, Loss: {'train/loss': 0.15441064536571503, 'train/grad_norm_pre_clip': 24.635841369628906, 'timing/data': 0.0008186064660549164, 'timing/model': 1.1321036741137505, 'learning_rate/qwen_vl_interface': 5.8782195220609e-07, 'learning_rate/action_model': 5.878219522060899e-06, 'epoch': 2.98}) +[2026-06-18 02:47:47,562][starVLA.training.train_starvla][INFO] - Step 6578, Loss: {'train/loss': 0.20889230072498322, 'train/grad_norm_pre_clip': 27.567476272583008, 'timing/data': 0.0004790816456079483, 'timing/model': 1.194497874006629, 'learning_rate/qwen_vl_interface': 5.874084859290172e-07, 'learning_rate/action_model': 5.874084859290171e-06, 'epoch': 2.98}) +[2026-06-18 02:47:51,318][starVLA.training.train_starvla][INFO] - Step 6579, Loss: {'train/loss': 0.12899960577487946, 'train/grad_norm_pre_clip': 18.420936584472656, 'timing/data': 0.0004765437915921211, 'timing/model': 0.9503558492287993, 'learning_rate/qwen_vl_interface': 5.869959862121822e-07, 'learning_rate/action_model': 5.869959862121822e-06, 'epoch': 2.98}) +[2026-06-18 02:47:55,158][starVLA.training.train_starvla][INFO] - Step 6580, Loss: {'train/loss': 0.18246299028396606, 'train/grad_norm_pre_clip': 37.738651275634766, 'timing/data': 0.000763697549700737, 'timing/model': 0.9832649426534772, 'learning_rate/qwen_vl_interface': 5.86584453141097e-07, 'learning_rate/action_model': 5.86584453141097e-06, 'epoch': 2.99}) +[2026-06-18 02:47:59,271][starVLA.training.train_starvla][INFO] - Step 6581, Loss: {'train/loss': 0.1501702517271042, 'train/grad_norm_pre_clip': 35.95287322998047, 'timing/data': 0.0005115615203976631, 'timing/model': 1.3139324337244034, 'learning_rate/qwen_vl_interface': 5.861738868010723e-07, 'learning_rate/action_model': 5.861738868010723e-06, 'epoch': 2.99}) +[2026-06-18 02:48:03,087][starVLA.training.train_starvla][INFO] - Step 6582, Loss: {'train/loss': 0.31462717056274414, 'train/grad_norm_pre_clip': 38.60884094238281, 'timing/data': 0.0005294596776366234, 'timing/model': 0.9202631544321775, 'learning_rate/qwen_vl_interface': 5.857642872772198e-07, 'learning_rate/action_model': 5.8576428727721986e-06, 'epoch': 2.99}) +[2026-06-18 02:48:06,940][starVLA.training.train_starvla][INFO] - Step 6583, Loss: {'train/loss': 0.297782301902771, 'train/grad_norm_pre_clip': 30.901514053344727, 'timing/data': 0.0008336631581187248, 'timing/model': 0.9553157771006227, 'learning_rate/qwen_vl_interface': 5.853556546544491e-07, 'learning_rate/action_model': 5.8535565465444904e-06, 'epoch': 2.99}) +[2026-06-18 02:48:10,842][starVLA.training.train_starvla][INFO] - Step 6584, Loss: {'train/loss': 0.2627929151058197, 'train/grad_norm_pre_clip': 28.15166473388672, 'timing/data': 0.0007127858698368073, 'timing/model': 1.173830984160304, 'learning_rate/qwen_vl_interface': 5.849479890174708e-07, 'learning_rate/action_model': 5.8494798901747075e-06, 'epoch': 2.99}) +[2026-06-18 02:48:14,562][starVLA.training.train_starvla][INFO] - Step 6585, Loss: {'train/loss': 0.12286083400249481, 'train/grad_norm_pre_clip': 24.523099899291992, 'timing/data': 0.0005000028759241104, 'timing/model': 0.9948158105835319, 'learning_rate/qwen_vl_interface': 5.845412904507948e-07, 'learning_rate/action_model': 5.845412904507948e-06, 'epoch': 2.99}) +[2026-06-18 02:48:18,426][starVLA.training.train_starvla][INFO] - Step 6586, Loss: {'train/loss': 0.10623791813850403, 'train/grad_norm_pre_clip': 31.62189292907715, 'timing/data': 0.000527448020875454, 'timing/model': 0.9248935710638762, 'learning_rate/qwen_vl_interface': 5.841355590387291e-07, 'learning_rate/action_model': 5.841355590387291e-06, 'epoch': 2.99}) +[2026-06-18 02:48:22,222][starVLA.training.train_starvla][INFO] - Step 6587, Loss: {'train/loss': 0.2908633351325989, 'train/grad_norm_pre_clip': 56.761077880859375, 'timing/data': 0.0008202921599149704, 'timing/model': 0.9297167779877782, 'learning_rate/qwen_vl_interface': 5.837307948653821e-07, 'learning_rate/action_model': 5.837307948653821e-06, 'epoch': 2.99}) +[2026-06-18 02:48:26,027][starVLA.training.train_starvla][INFO] - Step 6588, Loss: {'train/loss': 0.08038588613271713, 'train/grad_norm_pre_clip': 18.540870666503906, 'timing/data': 0.0004796469584107399, 'timing/model': 0.9392792880535126, 'learning_rate/qwen_vl_interface': 5.833269980146631e-07, 'learning_rate/action_model': 5.83326998014663e-06, 'epoch': 2.99}) +[2026-06-18 02:48:29,800][starVLA.training.train_starvla][INFO] - Step 6589, Loss: {'train/loss': 0.1014019250869751, 'train/grad_norm_pre_clip': 23.497581481933594, 'timing/data': 0.0013760579749941826, 'timing/model': 0.9404590548947453, 'learning_rate/qwen_vl_interface': 5.829241685702791e-07, 'learning_rate/action_model': 5.82924168570279e-06, 'epoch': 2.99}) +[2026-06-18 02:48:33,556][starVLA.training.train_starvla][INFO] - Step 6590, Loss: {'train/loss': 0.299227774143219, 'train/grad_norm_pre_clip': 41.16285705566406, 'timing/data': 0.00047140195965766907, 'timing/model': 0.9216649867594242, 'learning_rate/qwen_vl_interface': 5.825223066157365e-07, 'learning_rate/action_model': 5.8252230661573645e-06, 'epoch': 2.99}) +[2026-06-18 02:48:37,412][starVLA.training.train_starvla][INFO] - Step 6591, Loss: {'train/loss': 0.2540487051010132, 'train/grad_norm_pre_clip': 43.14348220825195, 'timing/data': 0.000583539716899395, 'timing/model': 1.1503666285425425, 'learning_rate/qwen_vl_interface': 5.821214122343423e-07, 'learning_rate/action_model': 5.821214122343422e-06, 'epoch': 2.99}) +[2026-06-18 02:48:41,000][starVLA.training.train_starvla][INFO] - Step 6592, Loss: {'train/loss': 0.17502443492412567, 'train/grad_norm_pre_clip': 20.751432418823242, 'timing/data': 0.0005106022581458092, 'timing/model': 0.9293569400906563, 'learning_rate/qwen_vl_interface': 5.817214855092024e-07, 'learning_rate/action_model': 5.817214855092024e-06, 'epoch': 2.99}) +[2026-06-18 02:48:44,770][starVLA.training.train_starvla][INFO] - Step 6593, Loss: {'train/loss': 0.411240816116333, 'train/grad_norm_pre_clip': 24.308290481567383, 'timing/data': 0.0005925679579377174, 'timing/model': 0.9297396363690495, 'learning_rate/qwen_vl_interface': 5.813225265232216e-07, 'learning_rate/action_model': 5.813225265232215e-06, 'epoch': 2.99}) +[2026-06-18 02:48:48,585][starVLA.training.train_starvla][INFO] - Step 6594, Loss: {'train/loss': 0.14769209921360016, 'train/grad_norm_pre_clip': 22.081199645996094, 'timing/data': 0.0010607652366161346, 'timing/model': 0.9633463053032756, 'learning_rate/qwen_vl_interface': 5.809245353591048e-07, 'learning_rate/action_model': 5.8092453535910485e-06, 'epoch': 2.99}) +[2026-06-18 02:48:52,329][starVLA.training.train_starvla][INFO] - Step 6595, Loss: {'train/loss': 0.11444665491580963, 'train/grad_norm_pre_clip': 26.5352725982666, 'timing/data': 0.0005450770258903503, 'timing/model': 0.9220453556627035, 'learning_rate/qwen_vl_interface': 5.805275120993561e-07, 'learning_rate/action_model': 5.805275120993561e-06, 'epoch': 2.99}) +[2026-06-18 02:48:56,104][starVLA.training.train_starvla][INFO] - Step 6596, Loss: {'train/loss': 0.032188549637794495, 'train/grad_norm_pre_clip': 7.466246604919434, 'timing/data': 0.00048203207552433014, 'timing/model': 0.9190146559849381, 'learning_rate/qwen_vl_interface': 5.801314568262787e-07, 'learning_rate/action_model': 5.801314568262787e-06, 'epoch': 2.99}) +[2026-06-18 02:48:59,818][starVLA.training.train_starvla][INFO] - Step 6597, Loss: {'train/loss': 0.4169325530529022, 'train/grad_norm_pre_clip': 41.732398986816406, 'timing/data': 0.0004886416718363762, 'timing/model': 1.0835255971178412, 'learning_rate/qwen_vl_interface': 5.797363696219751e-07, 'learning_rate/action_model': 5.797363696219751e-06, 'epoch': 2.99}) +[2026-06-18 02:49:03,568][starVLA.training.train_starvla][INFO] - Step 6598, Loss: {'train/loss': 0.3577946126461029, 'train/grad_norm_pre_clip': 46.80303955078125, 'timing/data': 0.0005027595907449722, 'timing/model': 1.1045344537124038, 'learning_rate/qwen_vl_interface': 5.793422505683474e-07, 'learning_rate/action_model': 5.793422505683474e-06, 'epoch': 2.99}) +[2026-06-18 02:49:07,118][starVLA.training.train_starvla][INFO] - Step 6599, Loss: {'train/loss': 0.27673086524009705, 'train/grad_norm_pre_clip': 25.287710189819336, 'timing/data': 0.0004894640296697617, 'timing/model': 0.9174156952649355, 'learning_rate/qwen_vl_interface': 5.789490997470978e-07, 'learning_rate/action_model': 5.789490997470978e-06, 'epoch': 2.99}) +[2026-06-18 02:51:02,849][starVLA.training.train_starvla][INFO] - Step 6600, Loss: {'train/loss': 0.24264152348041534, 'train/grad_norm_pre_clip': 27.123310089111328, 'eval/loss': 0.2545144855976105, 'eval/latency_8/loss': 0.2545144855976105, 'eval/demon_fire_precision': 0.9720344191763983, 'eval/demon_fire_recall': 0.9826032929481205, 'eval/demon_fire_f1': 0.977290282712807, 'eval/demon_fire_rate_model': 0.9053978853644964, 'eval/demon_fire_rate_teacher': 0.8956594323873122, 'eval/demon_left_precision': 0.9213759213759214, 'eval/demon_left_recall': 0.8029978586723768, 'eval/demon_left_f1': 0.8581235697940504, 'eval/demon_left_rate_model': 0.1132442960489705, 'eval/demon_left_rate_teacher': 0.12993878686700056, 'eval/demon_right_precision': 0.8622222222222222, 'eval/demon_right_recall': 0.8818181818181818, 'eval/demon_right_f1': 0.8719101123595505, 'eval/demon_right_rate_model': 0.18781302170283806, 'eval/demon_right_rate_teacher': 0.18363939899833054, 'eval/demon_move_f1': 0.8650168410768004, 'eval/demon_cc_f1': 0.9406627492603556, 'eval/latency_8/demon_fire_precision': 0.9720344191763983, 'eval/latency_8/demon_fire_recall': 0.9826032929481205, 'eval/latency_8/demon_fire_f1': 0.977290282712807, 'eval/latency_8/demon_fire_rate_model': 0.9053978853644964, 'eval/latency_8/demon_fire_rate_teacher': 0.8956594323873122, 'eval/latency_8/demon_left_precision': 0.9213759213759214, 'eval/latency_8/demon_left_recall': 0.8029978586723768, 'eval/latency_8/demon_left_f1': 0.8581235697940504, 'eval/latency_8/demon_left_rate_model': 0.1132442960489705, 'eval/latency_8/demon_left_rate_teacher': 0.12993878686700056, 'eval/latency_8/demon_right_precision': 0.8622222222222222, 'eval/latency_8/demon_right_recall': 0.8818181818181818, 'eval/latency_8/demon_right_f1': 0.8719101123595505, 'eval/latency_8/demon_right_rate_model': 0.18781302170283806, 'eval/latency_8/demon_right_rate_teacher': 0.18363939899833054, 'eval/latency_8/demon_move_f1': 0.8650168410768004, 'eval/latency_8/demon_cc_f1': 0.9406627492603556, 'timing/data': 0.00048691127449274063, 'timing/model': 1.1663949992507696, 'learning_rate/qwen_vl_interface': 5.785569172397254e-07, 'learning_rate/action_model': 5.785569172397254e-06, 'epoch': 2.99}) +[2026-06-18 02:51:02,864][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_6600_state +[2026-06-18 02:51:02,869][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 02:51:09,214][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_6600_state/model.safetensors +[2026-06-18 02:51:24,681][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_6600_state/optimizer.bin +[2026-06-18 02:51:24,682][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_6600_state/sampler.bin +[2026-06-18 02:51:24,682][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_6600_state/sampler_1.bin +[2026-06-18 02:51:24,684][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_6600_state/random_states_0.pkl +[2026-06-18 02:51:24,686][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 6600 +[2026-06-18 02:51:24,687][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 02:51:24,700][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 02:51:28,332][starVLA.training.train_starvla][INFO] - Step 6601, Loss: {'train/loss': 0.20924493670463562, 'train/grad_norm_pre_clip': 18.04124641418457, 'timing/data': 0.0004946626722812653, 'timing/model': 0.9316794108599424, 'learning_rate/qwen_vl_interface': 5.781657031275306e-07, 'learning_rate/action_model': 5.781657031275306e-06, 'epoch': 2.99}) +[2026-06-18 02:51:32,132][starVLA.training.train_starvla][INFO] - Step 6602, Loss: {'train/loss': 0.035193245857954025, 'train/grad_norm_pre_clip': 9.991592407226562, 'timing/data': 0.00045781396329402924, 'timing/model': 0.9157439451664686, 'learning_rate/qwen_vl_interface': 5.777754574916129e-07, 'learning_rate/action_model': 5.7777545749161285e-06, 'epoch': 3.0}) +[2026-06-18 02:51:35,853][starVLA.training.train_starvla][INFO] - Step 6603, Loss: {'train/loss': 0.029103662818670273, 'train/grad_norm_pre_clip': 6.821150302886963, 'timing/data': 0.0004945332184433937, 'timing/model': 0.9185758829116821, 'learning_rate/qwen_vl_interface': 5.773861804128701e-07, 'learning_rate/action_model': 5.773861804128701e-06, 'epoch': 3.0}) +[2026-06-18 02:51:39,605][starVLA.training.train_starvla][INFO] - Step 6604, Loss: {'train/loss': 0.034560564905405045, 'train/grad_norm_pre_clip': 7.576910972595215, 'timing/data': 0.0004823952913284302, 'timing/model': 0.919131226837635, 'learning_rate/qwen_vl_interface': 5.769978719720011e-07, 'learning_rate/action_model': 5.76997871972001e-06, 'epoch': 3.0}) +[2026-06-18 02:51:43,404][starVLA.training.train_starvla][INFO] - Step 6605, Loss: {'train/loss': 0.11480952054262161, 'train/grad_norm_pre_clip': 20.078317642211914, 'timing/data': 0.0004973206669092178, 'timing/model': 0.9234944600611925, 'learning_rate/qwen_vl_interface': 5.766105322495002e-07, 'learning_rate/action_model': 5.766105322495002e-06, 'epoch': 3.0}) +[2026-06-18 02:51:47,197][starVLA.training.train_starvla][INFO] - Step 6606, Loss: {'train/loss': 0.3129339814186096, 'train/grad_norm_pre_clip': 39.13043975830078, 'timing/data': 0.0007201405242085457, 'timing/model': 1.103748707100749, 'learning_rate/qwen_vl_interface': 5.762241613256654e-07, 'learning_rate/action_model': 5.762241613256654e-06, 'epoch': 3.0}) +[2026-06-18 02:51:50,950][starVLA.training.train_starvla][INFO] - Step 6607, Loss: {'train/loss': 0.2172888070344925, 'train/grad_norm_pre_clip': 34.808719635009766, 'timing/data': 0.00048237573355436325, 'timing/model': 1.091157391667366, 'learning_rate/qwen_vl_interface': 5.758387592805916e-07, 'learning_rate/action_model': 5.758387592805915e-06, 'epoch': 3.0}) +[2026-06-18 02:51:54,541][starVLA.training.train_starvla][INFO] - Step 6608, Loss: {'train/loss': 0.07574418187141418, 'train/grad_norm_pre_clip': 21.296850204467773, 'timing/data': 0.0008398424834012985, 'timing/model': 0.9234826136380434, 'learning_rate/qwen_vl_interface': 5.754543261941719e-07, 'learning_rate/action_model': 5.754543261941718e-06, 'epoch': 3.0}) +[2026-06-18 02:51:58,277][starVLA.training.train_starvla][INFO] - Step 6609, Loss: {'train/loss': 0.2799888849258423, 'train/grad_norm_pre_clip': 38.14313888549805, 'timing/data': 0.00046580657362937927, 'timing/model': 0.9213564228266478, 'learning_rate/qwen_vl_interface': 5.750708621461011e-07, 'learning_rate/action_model': 5.75070862146101e-06, 'epoch': 3.0}) +[2026-06-18 02:52:02,006][starVLA.training.train_starvla][INFO] - Step 6610, Loss: {'train/loss': 0.06818952411413193, 'train/grad_norm_pre_clip': 15.468788146972656, 'timing/data': 0.0006813574582338333, 'timing/model': 0.9253311036154628, 'learning_rate/qwen_vl_interface': 5.746883672158705e-07, 'learning_rate/action_model': 5.746883672158705e-06, 'epoch': 3.0}) +[2026-06-18 02:52:05,807][starVLA.training.train_starvla][INFO] - Step 6611, Loss: {'train/loss': 0.16423936188220978, 'train/grad_norm_pre_clip': 27.656885147094727, 'timing/data': 0.0004990147426724434, 'timing/model': 0.9369728891178966, 'learning_rate/qwen_vl_interface': 5.743068414827726e-07, 'learning_rate/action_model': 5.743068414827726e-06, 'epoch': 3.0}) +[2026-06-18 02:52:09,717][starVLA.training.train_starvla][INFO] - Step 6612, Loss: {'train/loss': 0.27335286140441895, 'train/grad_norm_pre_clip': 30.919214248657227, 'timing/data': 0.0006632860749959946, 'timing/model': 0.9236162276938558, 'learning_rate/qwen_vl_interface': 5.739262850258968e-07, 'learning_rate/action_model': 5.7392628502589684e-06, 'epoch': 3.0}) +[2026-06-18 02:52:13,721][starVLA.training.train_starvla][INFO] - Step 6613, Loss: {'train/loss': 0.04211368411779404, 'train/grad_norm_pre_clip': 25.540756225585938, 'timing/data': 0.3063909690827131, 'timing/model': 1.1163205290213227, 'learning_rate/qwen_vl_interface': 5.735466979241352e-07, 'learning_rate/action_model': 5.735466979241352e-06, 'epoch': 3.0}) +[2026-06-18 02:52:17,434][starVLA.training.train_starvla][INFO] - Step 6614, Loss: {'train/loss': 0.10302752256393433, 'train/grad_norm_pre_clip': 27.352418899536133, 'timing/data': 0.0005087461322546005, 'timing/model': 0.9316681316122413, 'learning_rate/qwen_vl_interface': 5.731680802561744e-07, 'learning_rate/action_model': 5.731680802561745e-06, 'epoch': 3.0}) +[2026-06-18 02:52:21,182][starVLA.training.train_starvla][INFO] - Step 6615, Loss: {'train/loss': 0.15431389212608337, 'train/grad_norm_pre_clip': 22.444517135620117, 'timing/data': 0.0009000347927212715, 'timing/model': 0.9284260412678123, 'learning_rate/qwen_vl_interface': 5.727904321005031e-07, 'learning_rate/action_model': 5.727904321005031e-06, 'epoch': 3.0}) +[2026-06-18 02:52:24,981][starVLA.training.train_starvla][INFO] - Step 6616, Loss: {'train/loss': 0.20530185103416443, 'train/grad_norm_pre_clip': 25.21136474609375, 'timing/data': 0.0004887357354164124, 'timing/model': 0.9248060313984752, 'learning_rate/qwen_vl_interface': 5.724137535354077e-07, 'learning_rate/action_model': 5.724137535354077e-06, 'epoch': 3.0}) +[2026-06-18 02:52:28,720][starVLA.training.train_starvla][INFO] - Step 6617, Loss: {'train/loss': 0.1147279217839241, 'train/grad_norm_pre_clip': 24.880481719970703, 'timing/data': 0.00047091487795114517, 'timing/model': 0.9349582381546497, 'learning_rate/qwen_vl_interface': 5.720380446389748e-07, 'learning_rate/action_model': 5.720380446389748e-06, 'epoch': 3.0}) +[2026-06-18 02:52:32,500][starVLA.training.train_starvla][INFO] - Step 6618, Loss: {'train/loss': 0.1100061908364296, 'train/grad_norm_pre_clip': 26.3168888092041, 'timing/data': 0.0005170013755559921, 'timing/model': 0.9228298859670758, 'learning_rate/qwen_vl_interface': 5.716633054890897e-07, 'learning_rate/action_model': 5.716633054890896e-06, 'epoch': 3.0}) +[2026-06-18 02:52:36,235][starVLA.training.train_starvla][INFO] - Step 6619, Loss: {'train/loss': 0.3070855736732483, 'train/grad_norm_pre_clip': 41.661006927490234, 'timing/data': 0.00047038309276103973, 'timing/model': 1.0998211288824677, 'learning_rate/qwen_vl_interface': 5.712895361634345e-07, 'learning_rate/action_model': 5.712895361634345e-06, 'epoch': 3.0}) +[2026-06-18 02:52:40,196][starVLA.training.train_starvla][INFO] - Step 6620, Loss: {'train/loss': 0.12538480758666992, 'train/grad_norm_pre_clip': 29.960342407226562, 'timing/data': 0.0005429256707429886, 'timing/model': 1.0904442854225636, 'learning_rate/qwen_vl_interface': 5.709167367394936e-07, 'learning_rate/action_model': 5.709167367394936e-06, 'epoch': 3.0}) +[2026-06-18 02:52:43,778][starVLA.training.train_starvla][INFO] - Step 6621, Loss: {'train/loss': 0.21580740809440613, 'train/grad_norm_pre_clip': 34.465354919433594, 'timing/data': 0.0005226870998740196, 'timing/model': 0.9259863402694464, 'learning_rate/qwen_vl_interface': 5.705449072945479e-07, 'learning_rate/action_model': 5.705449072945479e-06, 'epoch': 3.0}) +[2026-06-18 02:52:47,507][starVLA.training.train_starvla][INFO] - Step 6622, Loss: {'train/loss': 0.06237740069627762, 'train/grad_norm_pre_clip': 13.210437774658203, 'timing/data': 0.000796332024037838, 'timing/model': 0.9171530865132809, 'learning_rate/qwen_vl_interface': 5.70174047905678e-07, 'learning_rate/action_model': 5.701740479056779e-06, 'epoch': 3.0}) +[2026-06-18 02:52:51,241][starVLA.training.train_starvla][INFO] - Step 6623, Loss: {'train/loss': 0.13252323865890503, 'train/grad_norm_pre_clip': 21.58514976501465, 'timing/data': 0.00048715341836214066, 'timing/model': 0.9216891955584288, 'learning_rate/qwen_vl_interface': 5.69804158649764e-07, 'learning_rate/action_model': 5.6980415864976405e-06, 'epoch': 3.0}) +[2026-06-18 02:52:55,014][starVLA.training.train_starvla][INFO] - Step 6624, Loss: {'train/loss': 0.18134291470050812, 'train/grad_norm_pre_clip': 29.536970138549805, 'timing/data': 0.0005030203610658646, 'timing/model': 0.9212304577231407, 'learning_rate/qwen_vl_interface': 5.69435239603484e-07, 'learning_rate/action_model': 5.69435239603484e-06, 'epoch': 3.01}) +[2026-06-18 02:52:58,795][starVLA.training.train_starvla][INFO] - Step 6625, Loss: {'train/loss': 0.10960675776004791, 'train/grad_norm_pre_clip': 17.319469451904297, 'timing/data': 0.0004991739988327026, 'timing/model': 0.9267904181033373, 'learning_rate/qwen_vl_interface': 5.690672908433149e-07, 'learning_rate/action_model': 5.690672908433149e-06, 'epoch': 3.01}) +[2026-06-18 02:53:02,687][starVLA.training.train_starvla][INFO] - Step 6626, Loss: {'train/loss': 0.15798623859882355, 'train/grad_norm_pre_clip': 26.44709587097168, 'timing/data': 0.0005286745727062225, 'timing/model': 1.255365821532905, 'learning_rate/qwen_vl_interface': 5.68700312445534e-07, 'learning_rate/action_model': 5.687003124455339e-06, 'epoch': 3.01}) +[2026-06-18 02:53:06,263][starVLA.training.train_starvla][INFO] - Step 6627, Loss: {'train/loss': 0.11880875378847122, 'train/grad_norm_pre_clip': 38.507747650146484, 'timing/data': 0.0004941355437040329, 'timing/model': 0.9326018895953894, 'learning_rate/qwen_vl_interface': 5.683343044862151e-07, 'learning_rate/action_model': 5.68334304486215e-06, 'epoch': 3.01}) +[2026-06-18 02:53:10,233][starVLA.training.train_starvla][INFO] - Step 6628, Loss: {'train/loss': 0.38288190960884094, 'train/grad_norm_pre_clip': 28.809158325195312, 'timing/data': 0.0005158856511116028, 'timing/model': 0.9361339630559087, 'learning_rate/qwen_vl_interface': 5.679692670412324e-07, 'learning_rate/action_model': 5.679692670412324e-06, 'epoch': 3.01}) +[2026-06-18 02:53:14,017][starVLA.training.train_starvla][INFO] - Step 6629, Loss: {'train/loss': 0.43439069390296936, 'train/grad_norm_pre_clip': 38.300132751464844, 'timing/data': 0.00048811454325914383, 'timing/model': 0.9216348538175225, 'learning_rate/qwen_vl_interface': 5.67605200186259e-07, 'learning_rate/action_model': 5.67605200186259e-06, 'epoch': 3.01}) +[2026-06-18 02:53:17,782][starVLA.training.train_starvla][INFO] - Step 6630, Loss: {'train/loss': 0.22508542239665985, 'train/grad_norm_pre_clip': 27.77269744873047, 'timing/data': 0.0005202721804380417, 'timing/model': 0.9334897743538022, 'learning_rate/qwen_vl_interface': 5.67242103996766e-07, 'learning_rate/action_model': 5.67242103996766e-06, 'epoch': 3.01}) +[2026-06-18 02:53:21,753][starVLA.training.train_starvla][INFO] - Step 6631, Loss: {'train/loss': 0.4544145464897156, 'train/grad_norm_pre_clip': 35.0751953125, 'timing/data': 0.0005035987123847008, 'timing/model': 0.9211346888914704, 'learning_rate/qwen_vl_interface': 5.668799785480242e-07, 'learning_rate/action_model': 5.668799785480242e-06, 'epoch': 3.01}) +[2026-06-18 02:53:25,511][starVLA.training.train_starvla][INFO] - Step 6632, Loss: {'train/loss': 0.3166826069355011, 'train/grad_norm_pre_clip': 36.22602081298828, 'timing/data': 0.0004908135160803795, 'timing/model': 1.0934021202847362, 'learning_rate/qwen_vl_interface': 5.665188239151019e-07, 'learning_rate/action_model': 5.665188239151018e-06, 'epoch': 3.01}) +[2026-06-18 02:53:29,294][starVLA.training.train_starvla][INFO] - Step 6633, Loss: {'train/loss': 0.054121848195791245, 'train/grad_norm_pre_clip': 15.092694282531738, 'timing/data': 0.00045484956353902817, 'timing/model': 1.1104480354115367, 'learning_rate/qwen_vl_interface': 5.661586401728671e-07, 'learning_rate/action_model': 5.6615864017286705e-06, 'epoch': 3.01}) +[2026-06-18 02:53:32,935][starVLA.training.train_starvla][INFO] - Step 6634, Loss: {'train/loss': 0.045318353921175, 'train/grad_norm_pre_clip': 11.198019981384277, 'timing/data': 0.0008626943454146385, 'timing/model': 0.9489176766946912, 'learning_rate/qwen_vl_interface': 5.657994273959857e-07, 'learning_rate/action_model': 5.657994273959856e-06, 'epoch': 3.01}) +[2026-06-18 02:53:36,682][starVLA.training.train_starvla][INFO] - Step 6635, Loss: {'train/loss': 0.3499867022037506, 'train/grad_norm_pre_clip': 27.95879554748535, 'timing/data': 0.0004895925521850586, 'timing/model': 0.9228452863171697, 'learning_rate/qwen_vl_interface': 5.654411856589233e-07, 'learning_rate/action_model': 5.654411856589233e-06, 'epoch': 3.01}) +[2026-06-18 02:53:40,421][starVLA.training.train_starvla][INFO] - Step 6636, Loss: {'train/loss': 0.2713162302970886, 'train/grad_norm_pre_clip': 34.87712860107422, 'timing/data': 0.0008324366062879562, 'timing/model': 0.9240132821723819, 'learning_rate/qwen_vl_interface': 5.650839150359444e-07, 'learning_rate/action_model': 5.650839150359444e-06, 'epoch': 3.01}) +[2026-06-18 02:53:44,191][starVLA.training.train_starvla][INFO] - Step 6637, Loss: {'train/loss': 0.04294373095035553, 'train/grad_norm_pre_clip': 16.95964241027832, 'timing/data': 0.0005173105746507645, 'timing/model': 0.9292104244232178, 'learning_rate/qwen_vl_interface': 5.647276156011114e-07, 'learning_rate/action_model': 5.6472761560111135e-06, 'epoch': 3.01}) +[2026-06-18 02:53:47,943][starVLA.training.train_starvla][INFO] - Step 6638, Loss: {'train/loss': 0.28215742111206055, 'train/grad_norm_pre_clip': 30.66899871826172, 'timing/data': 0.00047737546265125275, 'timing/model': 0.9274104367941618, 'learning_rate/qwen_vl_interface': 5.643722874282846e-07, 'learning_rate/action_model': 5.643722874282846e-06, 'epoch': 3.01}) +[2026-06-18 02:53:51,687][starVLA.training.train_starvla][INFO] - Step 6639, Loss: {'train/loss': 0.3332024812698364, 'train/grad_norm_pre_clip': 48.31485366821289, 'timing/data': 0.0008174153044819832, 'timing/model': 1.0898059383034706, 'learning_rate/qwen_vl_interface': 5.640179305911253e-07, 'learning_rate/action_model': 5.640179305911253e-06, 'epoch': 3.01}) +[2026-06-18 02:53:55,275][starVLA.training.train_starvla][INFO] - Step 6640, Loss: {'train/loss': 0.21828605234622955, 'train/grad_norm_pre_clip': 45.820648193359375, 'timing/data': 0.0010503251105546951, 'timing/model': 0.9194939285516739, 'learning_rate/qwen_vl_interface': 5.636645451630908e-07, 'learning_rate/action_model': 5.6366454516309084e-06, 'epoch': 3.01}) +[2026-06-18 02:53:59,073][starVLA.training.train_starvla][INFO] - Step 6641, Loss: {'train/loss': 0.1903475821018219, 'train/grad_norm_pre_clip': 22.255403518676758, 'timing/data': 0.00046780239790678024, 'timing/model': 0.9189829202368855, 'learning_rate/qwen_vl_interface': 5.63312131217439e-07, 'learning_rate/action_model': 5.633121312174389e-06, 'epoch': 3.01}) +[2026-06-18 02:54:02,850][starVLA.training.train_starvla][INFO] - Step 6642, Loss: {'train/loss': 0.1354871392250061, 'train/grad_norm_pre_clip': 14.210371017456055, 'timing/data': 0.000516924075782299, 'timing/model': 0.9279794162139297, 'learning_rate/qwen_vl_interface': 5.629606888272253e-07, 'learning_rate/action_model': 5.629606888272252e-06, 'epoch': 3.01}) +[2026-06-18 02:54:06,619][starVLA.training.train_starvla][INFO] - Step 6643, Loss: {'train/loss': 0.19112809002399445, 'train/grad_norm_pre_clip': 27.034605026245117, 'timing/data': 0.0005030697211623192, 'timing/model': 0.9492022702470422, 'learning_rate/qwen_vl_interface': 5.626102180653046e-07, 'learning_rate/action_model': 5.626102180653045e-06, 'epoch': 3.01}) +[2026-06-18 02:54:10,479][starVLA.training.train_starvla][INFO] - Step 6644, Loss: {'train/loss': 0.18964140117168427, 'train/grad_norm_pre_clip': 27.128768920898438, 'timing/data': 0.00047990307211875916, 'timing/model': 0.9391802502796054, 'learning_rate/qwen_vl_interface': 5.62260719004329e-07, 'learning_rate/action_model': 5.6226071900432895e-06, 'epoch': 3.01}) +[2026-06-18 02:54:14,523][starVLA.training.train_starvla][INFO] - Step 6645, Loss: {'train/loss': 0.26546257734298706, 'train/grad_norm_pre_clip': 32.59812545776367, 'timing/data': 0.0004936251789331436, 'timing/model': 0.9269990585744381, 'learning_rate/qwen_vl_interface': 5.619121917167504e-07, 'learning_rate/action_model': 5.619121917167505e-06, 'epoch': 3.01}) +[2026-06-18 02:54:18,310][starVLA.training.train_starvla][INFO] - Step 6646, Loss: {'train/loss': 0.07234359532594681, 'train/grad_norm_pre_clip': 20.745256423950195, 'timing/data': 0.0008784746751189232, 'timing/model': 1.1082990244030952, 'learning_rate/qwen_vl_interface': 5.615646362748199e-07, 'learning_rate/action_model': 5.6156463627482e-06, 'epoch': 3.02}) +[2026-06-18 02:54:21,946][starVLA.training.train_starvla][INFO] - Step 6647, Loss: {'train/loss': 0.06378637254238129, 'train/grad_norm_pre_clip': 18.153249740600586, 'timing/data': 0.0007736273109912872, 'timing/model': 0.9435306005179882, 'learning_rate/qwen_vl_interface': 5.612180527505854e-07, 'learning_rate/action_model': 5.612180527505854e-06, 'epoch': 3.02}) +[2026-06-18 02:54:25,741][starVLA.training.train_starvla][INFO] - Step 6648, Loss: {'train/loss': 0.13711601495742798, 'train/grad_norm_pre_clip': 22.917070388793945, 'timing/data': 0.0004836106672883034, 'timing/model': 0.9346093647181988, 'learning_rate/qwen_vl_interface': 5.608724412158931e-07, 'learning_rate/action_model': 5.608724412158931e-06, 'epoch': 3.02}) +[2026-06-18 02:54:29,508][starVLA.training.train_starvla][INFO] - Step 6649, Loss: {'train/loss': 0.26420384645462036, 'train/grad_norm_pre_clip': 28.974891662597656, 'timing/data': 0.0005098460242152214, 'timing/model': 0.9236559504643083, 'learning_rate/qwen_vl_interface': 5.605278017423897e-07, 'learning_rate/action_model': 5.605278017423897e-06, 'epoch': 3.02}) +[2026-06-18 02:54:33,369][starVLA.training.train_starvla][INFO] - Step 6650, Loss: {'train/loss': 0.16832183301448822, 'train/grad_norm_pre_clip': 24.399658203125, 'timing/data': 0.0006218506023287773, 'timing/model': 0.9914328446611762, 'learning_rate/qwen_vl_interface': 5.601841344015193e-07, 'learning_rate/action_model': 5.601841344015193e-06, 'epoch': 3.02}) +[2026-06-18 02:54:38,280][starVLA.training.train_starvla][INFO] - Step 6651, Loss: {'train/loss': 0.07401321083307266, 'train/grad_norm_pre_clip': 16.92547035217285, 'timing/data': 0.0004821429029107094, 'timing/model': 0.9344858564436436, 'learning_rate/qwen_vl_interface': 5.598414392645239e-07, 'learning_rate/action_model': 5.598414392645239e-06, 'epoch': 3.02}) +[2026-06-18 02:54:42,057][starVLA.training.train_starvla][INFO] - Step 6652, Loss: {'train/loss': 0.05493893474340439, 'train/grad_norm_pre_clip': 13.613655090332031, 'timing/data': 0.00047486182302236557, 'timing/model': 1.1363131357356906, 'learning_rate/qwen_vl_interface': 5.594997164024454e-07, 'learning_rate/action_model': 5.594997164024454e-06, 'epoch': 3.02}) +[2026-06-18 02:54:45,789][starVLA.training.train_starvla][INFO] - Step 6653, Loss: {'train/loss': 0.28516149520874023, 'train/grad_norm_pre_clip': 35.045631408691406, 'timing/data': 0.0005024122074246407, 'timing/model': 1.082942795008421, 'learning_rate/qwen_vl_interface': 5.591589658861228e-07, 'learning_rate/action_model': 5.591589658861228e-06, 'epoch': 3.02}) +[2026-06-18 02:54:50,101][starVLA.training.train_starvla][INFO] - Step 6654, Loss: {'train/loss': 0.02327311784029007, 'train/grad_norm_pre_clip': 8.148775100708008, 'timing/data': 0.0005428558215498924, 'timing/model': 1.1840341771021485, 'learning_rate/qwen_vl_interface': 5.588191877861945e-07, 'learning_rate/action_model': 5.588191877861944e-06, 'epoch': 3.02}) +[2026-06-18 02:54:53,848][starVLA.training.train_starvla][INFO] - Step 6655, Loss: {'train/loss': 0.10960771888494492, 'train/grad_norm_pre_clip': 30.11127281188965, 'timing/data': 0.0004845401272177696, 'timing/model': 0.9315405515953898, 'learning_rate/qwen_vl_interface': 5.584803821730961e-07, 'learning_rate/action_model': 5.58480382173096e-06, 'epoch': 3.02}) +[2026-06-18 02:54:57,918][starVLA.training.train_starvla][INFO] - Step 6656, Loss: {'train/loss': 0.24302737414836884, 'train/grad_norm_pre_clip': 26.096620559692383, 'timing/data': 0.0007895287126302719, 'timing/model': 1.0818977104499936, 'learning_rate/qwen_vl_interface': 5.581425491170627e-07, 'learning_rate/action_model': 5.581425491170626e-06, 'epoch': 3.02}) +[2026-06-18 02:55:01,703][starVLA.training.train_starvla][INFO] - Step 6657, Loss: {'train/loss': 0.32399770617485046, 'train/grad_norm_pre_clip': 35.562320709228516, 'timing/data': 0.0006906269118189812, 'timing/model': 0.9384265029802918, 'learning_rate/qwen_vl_interface': 5.578056886881274e-07, 'learning_rate/action_model': 5.578056886881274e-06, 'epoch': 3.02}) +[2026-06-18 02:55:05,502][starVLA.training.train_starvla][INFO] - Step 6658, Loss: {'train/loss': 0.03133680298924446, 'train/grad_norm_pre_clip': 6.890794277191162, 'timing/data': 0.0004914216697216034, 'timing/model': 0.947395283728838, 'learning_rate/qwen_vl_interface': 5.574698009561223e-07, 'learning_rate/action_model': 5.574698009561223e-06, 'epoch': 3.02}) +[2026-06-18 02:55:09,442][starVLA.training.train_starvla][INFO] - Step 6659, Loss: {'train/loss': 0.05628994107246399, 'train/grad_norm_pre_clip': 18.226219177246094, 'timing/data': 0.0005063330754637718, 'timing/model': 1.1995666651055217, 'learning_rate/qwen_vl_interface': 5.571348859906763e-07, 'learning_rate/action_model': 5.571348859906764e-06, 'epoch': 3.02}) +[2026-06-18 02:55:13,372][starVLA.training.train_starvla][INFO] - Step 6660, Loss: {'train/loss': 0.19056057929992676, 'train/grad_norm_pre_clip': 52.422916412353516, 'timing/data': 0.0005070241168141365, 'timing/model': 1.1206294111907482, 'learning_rate/qwen_vl_interface': 5.568009438612186e-07, 'learning_rate/action_model': 5.568009438612186e-06, 'epoch': 3.02}) +[2026-06-18 02:55:17,057][starVLA.training.train_starvla][INFO] - Step 6661, Loss: {'train/loss': 0.1974114179611206, 'train/grad_norm_pre_clip': 25.289125442504883, 'timing/data': 0.0004801526665687561, 'timing/model': 0.9824287099763751, 'learning_rate/qwen_vl_interface': 5.564679746369755e-07, 'learning_rate/action_model': 5.5646797463697556e-06, 'epoch': 3.02}) +[2026-06-18 02:55:20,957][starVLA.training.train_starvla][INFO] - Step 6662, Loss: {'train/loss': 0.25508660078048706, 'train/grad_norm_pre_clip': 22.73637580871582, 'timing/data': 0.0005328208208084106, 'timing/model': 0.96913231164217, 'learning_rate/qwen_vl_interface': 5.561359783869712e-07, 'learning_rate/action_model': 5.5613597838697116e-06, 'epoch': 3.02}) +[2026-06-18 02:55:24,824][starVLA.training.train_starvla][INFO] - Step 6663, Loss: {'train/loss': 0.11686241626739502, 'train/grad_norm_pre_clip': 14.45985221862793, 'timing/data': 0.0005294084548950195, 'timing/model': 0.939438053406775, 'learning_rate/qwen_vl_interface': 5.558049551800291e-07, 'learning_rate/action_model': 5.558049551800292e-06, 'epoch': 3.02}) +[2026-06-18 02:55:28,566][starVLA.training.train_starvla][INFO] - Step 6664, Loss: {'train/loss': 0.18889214098453522, 'train/grad_norm_pre_clip': 21.533681869506836, 'timing/data': 0.0004897834733128548, 'timing/model': 0.9291431186720729, 'learning_rate/qwen_vl_interface': 5.554749050847718e-07, 'learning_rate/action_model': 5.5547490508477176e-06, 'epoch': 3.02}) +[2026-06-18 02:55:32,383][starVLA.training.train_starvla][INFO] - Step 6665, Loss: {'train/loss': 0.05640706792473793, 'train/grad_norm_pre_clip': 19.460081100463867, 'timing/data': 0.0004853270947933197, 'timing/model': 0.9412634326145053, 'learning_rate/qwen_vl_interface': 5.55145828169618e-07, 'learning_rate/action_model': 5.5514582816961805e-06, 'epoch': 3.02}) +[2026-06-18 02:55:36,209][starVLA.training.train_starvla][INFO] - Step 6666, Loss: {'train/loss': 0.2796039283275604, 'train/grad_norm_pre_clip': 28.618671417236328, 'timing/data': 0.0004906682297587395, 'timing/model': 1.1494174404069781, 'learning_rate/qwen_vl_interface': 5.548177245027859e-07, 'learning_rate/action_model': 5.548177245027859e-06, 'epoch': 3.02}) +[2026-06-18 02:55:39,823][starVLA.training.train_starvla][INFO] - Step 6667, Loss: {'train/loss': 0.2026491016149521, 'train/grad_norm_pre_clip': 32.317962646484375, 'timing/data': 0.0005000298842787743, 'timing/model': 0.9397060908377171, 'learning_rate/qwen_vl_interface': 5.544905941522913e-07, 'learning_rate/action_model': 5.544905941522913e-06, 'epoch': 3.02}) +[2026-06-18 02:55:43,601][starVLA.training.train_starvla][INFO] - Step 6668, Loss: {'train/loss': 0.17999282479286194, 'train/grad_norm_pre_clip': 39.1031608581543, 'timing/data': 0.00048657041043043137, 'timing/model': 0.9254700355231762, 'learning_rate/qwen_vl_interface': 5.541644371859498e-07, 'learning_rate/action_model': 5.541644371859498e-06, 'epoch': 3.03}) +[2026-06-18 02:55:47,341][starVLA.training.train_starvla][INFO] - Step 6669, Loss: {'train/loss': 0.23752282559871674, 'train/grad_norm_pre_clip': 25.552947998046875, 'timing/data': 0.000781266950070858, 'timing/model': 0.9295465406030416, 'learning_rate/qwen_vl_interface': 5.53839253671373e-07, 'learning_rate/action_model': 5.5383925367137305e-06, 'epoch': 3.03}) +[2026-06-18 02:55:51,140][starVLA.training.train_starvla][INFO] - Step 6670, Loss: {'train/loss': 0.03394746780395508, 'train/grad_norm_pre_clip': 22.957107543945312, 'timing/data': 0.0005114041268825531, 'timing/model': 0.9275725465267897, 'learning_rate/qwen_vl_interface': 5.535150436759725e-07, 'learning_rate/action_model': 5.535150436759725e-06, 'epoch': 3.03}) +[2026-06-18 02:55:54,930][starVLA.training.train_starvla][INFO] - Step 6671, Loss: {'train/loss': 0.055566661059856415, 'train/grad_norm_pre_clip': 13.54665470123291, 'timing/data': 0.000493139959871769, 'timing/model': 0.9287212649360299, 'learning_rate/qwen_vl_interface': 5.531918072669566e-07, 'learning_rate/action_model': 5.531918072669565e-06, 'epoch': 3.03}) +[2026-06-18 02:55:58,722][starVLA.training.train_starvla][INFO] - Step 6672, Loss: {'train/loss': 0.2141350507736206, 'train/grad_norm_pre_clip': 17.626792907714844, 'timing/data': 0.00047613028436899185, 'timing/model': 1.104302642866969, 'learning_rate/qwen_vl_interface': 5.528695445113336e-07, 'learning_rate/action_model': 5.528695445113335e-06, 'epoch': 3.03}) +[2026-06-18 02:56:02,514][starVLA.training.train_starvla][INFO] - Step 6673, Loss: {'train/loss': 0.6238458156585693, 'train/grad_norm_pre_clip': 47.795780181884766, 'timing/data': 0.0005072653293609619, 'timing/model': 1.137793249450624, 'learning_rate/qwen_vl_interface': 5.525482554759079e-07, 'learning_rate/action_model': 5.5254825547590795e-06, 'epoch': 3.03}) +[2026-06-18 02:56:06,281][starVLA.training.train_starvla][INFO] - Step 6674, Loss: {'train/loss': 0.2518819570541382, 'train/grad_norm_pre_clip': 37.160499572753906, 'timing/data': 0.0004851054400205612, 'timing/model': 1.0103100799024105, 'learning_rate/qwen_vl_interface': 5.522279402272834e-07, 'learning_rate/action_model': 5.522279402272834e-06, 'epoch': 3.03}) +[2026-06-18 02:56:10,142][starVLA.training.train_starvla][INFO] - Step 6675, Loss: {'train/loss': 0.3916996419429779, 'train/grad_norm_pre_clip': 33.71251678466797, 'timing/data': 0.0005301283672451973, 'timing/model': 0.9356760010123253, 'learning_rate/qwen_vl_interface': 5.519085988318627e-07, 'learning_rate/action_model': 5.5190859883186265e-06, 'epoch': 3.03}) +[2026-06-18 02:56:14,075][starVLA.training.train_starvla][INFO] - Step 6676, Loss: {'train/loss': 0.052249643951654434, 'train/grad_norm_pre_clip': 17.241209030151367, 'timing/data': 0.0006621181964874268, 'timing/model': 0.9431972643360496, 'learning_rate/qwen_vl_interface': 5.515902313558443e-07, 'learning_rate/action_model': 5.515902313558444e-06, 'epoch': 3.03}) +[2026-06-18 02:56:18,043][starVLA.training.train_starvla][INFO] - Step 6677, Loss: {'train/loss': 0.02896704711019993, 'train/grad_norm_pre_clip': 12.642457008361816, 'timing/data': 0.00047665834426879883, 'timing/model': 0.9442909611389041, 'learning_rate/qwen_vl_interface': 5.512728378652266e-07, 'learning_rate/action_model': 5.512728378652266e-06, 'epoch': 3.03}) +[2026-06-18 02:56:22,125][starVLA.training.train_starvla][INFO] - Step 6678, Loss: {'train/loss': 0.09049613773822784, 'train/grad_norm_pre_clip': 22.309825897216797, 'timing/data': 0.0005015106871724129, 'timing/model': 1.1970359291881323, 'learning_rate/qwen_vl_interface': 5.509564184258055e-07, 'learning_rate/action_model': 5.509564184258055e-06, 'epoch': 3.03}) +[2026-06-18 02:56:25,978][starVLA.training.train_starvla][INFO] - Step 6679, Loss: {'train/loss': 0.15191194415092468, 'train/grad_norm_pre_clip': 15.775012016296387, 'timing/data': 0.0008497759699821472, 'timing/model': 1.1598243862390518, 'learning_rate/qwen_vl_interface': 5.506409731031754e-07, 'learning_rate/action_model': 5.506409731031755e-06, 'epoch': 3.03}) +[2026-06-18 02:56:29,886][starVLA.training.train_starvla][INFO] - Step 6680, Loss: {'train/loss': 0.28401923179626465, 'train/grad_norm_pre_clip': 32.502052307128906, 'timing/data': 0.0007846001535654068, 'timing/model': 1.1731019914150238, 'learning_rate/qwen_vl_interface': 5.503265019627289e-07, 'learning_rate/action_model': 5.503265019627289e-06, 'epoch': 3.03}) +[2026-06-18 02:56:33,516][starVLA.training.train_starvla][INFO] - Step 6681, Loss: {'train/loss': 0.3311034142971039, 'train/grad_norm_pre_clip': 31.403745651245117, 'timing/data': 0.000507512129843235, 'timing/model': 0.9331259494647384, 'learning_rate/qwen_vl_interface': 5.500130050696548e-07, 'learning_rate/action_model': 5.500130050696548e-06, 'epoch': 3.03}) +[2026-06-18 02:56:37,414][starVLA.training.train_starvla][INFO] - Step 6682, Loss: {'train/loss': 0.19975046813488007, 'train/grad_norm_pre_clip': 36.711055755615234, 'timing/data': 0.0007020030170679092, 'timing/model': 1.0213832948356867, 'learning_rate/qwen_vl_interface': 5.497004824889424e-07, 'learning_rate/action_model': 5.497004824889423e-06, 'epoch': 3.03}) +[2026-06-18 02:56:41,281][starVLA.training.train_starvla][INFO] - Step 6683, Loss: {'train/loss': 0.1558283269405365, 'train/grad_norm_pre_clip': 76.59113311767578, 'timing/data': 0.0011630412191152573, 'timing/model': 1.0013685366138816, 'learning_rate/qwen_vl_interface': 5.493889342853779e-07, 'learning_rate/action_model': 5.493889342853778e-06, 'epoch': 3.03}) +[2026-06-18 02:56:45,216][starVLA.training.train_starvla][INFO] - Step 6684, Loss: {'train/loss': 0.16124914586544037, 'train/grad_norm_pre_clip': 25.509654998779297, 'timing/data': 0.0014202818274497986, 'timing/model': 1.0289144832640886, 'learning_rate/qwen_vl_interface': 5.490783605235451e-07, 'learning_rate/action_model': 5.49078360523545e-06, 'epoch': 3.03}) +[2026-06-18 02:56:49,036][starVLA.training.train_starvla][INFO] - Step 6685, Loss: {'train/loss': 0.04602846875786781, 'train/grad_norm_pre_clip': 13.79183292388916, 'timing/data': 0.001415221020579338, 'timing/model': 0.9391414597630501, 'learning_rate/qwen_vl_interface': 5.487687612678265e-07, 'learning_rate/action_model': 5.4876876126782645e-06, 'epoch': 3.03}) +[2026-06-18 02:56:52,889][starVLA.training.train_starvla][INFO] - Step 6686, Loss: {'train/loss': 0.42106470465660095, 'train/grad_norm_pre_clip': 29.90086555480957, 'timing/data': 0.000626005232334137, 'timing/model': 1.159416296519339, 'learning_rate/qwen_vl_interface': 5.484601365824032e-07, 'learning_rate/action_model': 5.4846013658240315e-06, 'epoch': 3.03}) +[2026-06-18 02:56:56,556][starVLA.training.train_starvla][INFO] - Step 6687, Loss: {'train/loss': 0.11131671071052551, 'train/grad_norm_pre_clip': 22.55231475830078, 'timing/data': 0.0008827392011880875, 'timing/model': 0.9440164361149073, 'learning_rate/qwen_vl_interface': 5.481524865312518e-07, 'learning_rate/action_model': 5.481524865312518e-06, 'epoch': 3.03}) +[2026-06-18 02:57:00,399][starVLA.training.train_starvla][INFO] - Step 6688, Loss: {'train/loss': 0.2625564932823181, 'train/grad_norm_pre_clip': 30.89905548095703, 'timing/data': 0.0005535660311579704, 'timing/model': 1.017674165777862, 'learning_rate/qwen_vl_interface': 5.478458111781496e-07, 'learning_rate/action_model': 5.478458111781496e-06, 'epoch': 3.03}) +[2026-06-18 02:57:04,166][starVLA.training.train_starvla][INFO] - Step 6689, Loss: {'train/loss': 0.2360168844461441, 'train/grad_norm_pre_clip': 34.42887878417969, 'timing/data': 0.0005318503826856613, 'timing/model': 0.9244216121733189, 'learning_rate/qwen_vl_interface': 5.475401105866712e-07, 'learning_rate/action_model': 5.4754011058667114e-06, 'epoch': 3.03}) +[2026-06-18 02:57:07,929][starVLA.training.train_starvla][INFO] - Step 6690, Loss: {'train/loss': 0.14886537194252014, 'train/grad_norm_pre_clip': 22.119770050048828, 'timing/data': 0.0004978897050023079, 'timing/model': 0.9330180436372757, 'learning_rate/qwen_vl_interface': 5.47235384820187e-07, 'learning_rate/action_model': 5.47235384820187e-06, 'epoch': 3.04}) +[2026-06-18 02:57:11,697][starVLA.training.train_starvla][INFO] - Step 6691, Loss: {'train/loss': 0.06242411956191063, 'train/grad_norm_pre_clip': 13.574305534362793, 'timing/data': 0.0004977034404873848, 'timing/model': 0.932823808863759, 'learning_rate/qwen_vl_interface': 5.469316339418681e-07, 'learning_rate/action_model': 5.469316339418681e-06, 'epoch': 3.04}) +[2026-06-18 02:57:15,447][starVLA.training.train_starvla][INFO] - Step 6692, Loss: {'train/loss': 0.2930944263935089, 'train/grad_norm_pre_clip': 22.836679458618164, 'timing/data': 0.0004893569275736809, 'timing/model': 0.9277652883902192, 'learning_rate/qwen_vl_interface': 5.466288580146825e-07, 'learning_rate/action_model': 5.466288580146825e-06, 'epoch': 3.04}) +[2026-06-18 02:57:19,221][starVLA.training.train_starvla][INFO] - Step 6693, Loss: {'train/loss': 0.490679532289505, 'train/grad_norm_pre_clip': 28.769346237182617, 'timing/data': 0.000488518737256527, 'timing/model': 1.1132544158026576, 'learning_rate/qwen_vl_interface': 5.463270571013954e-07, 'learning_rate/action_model': 5.463270571013954e-06, 'epoch': 3.04}) +[2026-06-18 02:57:22,838][starVLA.training.train_starvla][INFO] - Step 6694, Loss: {'train/loss': 0.6635448336601257, 'train/grad_norm_pre_clip': 71.11457061767578, 'timing/data': 0.0005926042795181274, 'timing/model': 0.9573563979938626, 'learning_rate/qwen_vl_interface': 5.460262312645709e-07, 'learning_rate/action_model': 5.460262312645709e-06, 'epoch': 3.04}) +[2026-06-18 02:57:26,778][starVLA.training.train_starvla][INFO] - Step 6695, Loss: {'train/loss': 0.3648854196071625, 'train/grad_norm_pre_clip': 62.04178237915039, 'timing/data': 0.0004954366013407707, 'timing/model': 0.9285362726077437, 'learning_rate/qwen_vl_interface': 5.457263805665707e-07, 'learning_rate/action_model': 5.457263805665707e-06, 'epoch': 3.04}) +[2026-06-18 02:57:30,519][starVLA.training.train_starvla][INFO] - Step 6696, Loss: {'train/loss': 0.2006513774394989, 'train/grad_norm_pre_clip': 22.70136070251465, 'timing/data': 0.0004892973229289055, 'timing/model': 0.932792779058218, 'learning_rate/qwen_vl_interface': 5.454275050695529e-07, 'learning_rate/action_model': 5.454275050695529e-06, 'epoch': 3.04}) +[2026-06-18 02:57:34,306][starVLA.training.train_starvla][INFO] - Step 6697, Loss: {'train/loss': 0.21267098188400269, 'train/grad_norm_pre_clip': 23.323421478271484, 'timing/data': 0.0013387799263000488, 'timing/model': 0.939202630892396, 'learning_rate/qwen_vl_interface': 5.451296048354764e-07, 'learning_rate/action_model': 5.4512960483547635e-06, 'epoch': 3.04}) +[2026-06-18 02:57:38,162][starVLA.training.train_starvla][INFO] - Step 6698, Loss: {'train/loss': 0.10218866914510727, 'train/grad_norm_pre_clip': 44.819984436035156, 'timing/data': 0.0005257716402411461, 'timing/model': 0.9295385116711259, 'learning_rate/qwen_vl_interface': 5.448326799260951e-07, 'learning_rate/action_model': 5.448326799260952e-06, 'epoch': 3.04}) +[2026-06-18 02:57:42,167][starVLA.training.train_starvla][INFO] - Step 6699, Loss: {'train/loss': 0.13603202998638153, 'train/grad_norm_pre_clip': 28.187379837036133, 'timing/data': 0.0005406979471445084, 'timing/model': 1.3333717556670308, 'learning_rate/qwen_vl_interface': 5.445367304029627e-07, 'learning_rate/action_model': 5.445367304029626e-06, 'epoch': 3.04}) +[2026-06-18 02:59:38,052][starVLA.training.train_starvla][INFO] - Step 6700, Loss: {'train/loss': 0.14094392955303192, 'train/grad_norm_pre_clip': 23.391551971435547, 'eval/loss': 0.24997328221797943, 'eval/latency_8/loss': 0.24997328221797943, 'eval/demon_fire_precision': 0.9708767627222563, 'eval/demon_fire_recall': 0.9838459148803976, 'eval/demon_fire_f1': 0.9773183150748342, 'eval/demon_fire_rate_model': 0.907623817473567, 'eval/demon_fire_rate_teacher': 0.8956594323873122, 'eval/demon_left_precision': 0.9225, 'eval/demon_left_recall': 0.7901498929336188, 'eval/demon_left_f1': 0.8512110726643598, 'eval/demon_left_rate_model': 0.11129660545353366, 'eval/demon_left_rate_teacher': 0.12993878686700056, 'eval/demon_right_precision': 0.8656716417910447, 'eval/demon_right_recall': 0.8787878787878788, 'eval/demon_right_f1': 0.8721804511278195, 'eval/demon_right_rate_model': 0.1864218141346689, 'eval/demon_right_rate_teacher': 0.18363939899833054, 'eval/demon_move_f1': 0.8616957618960897, 'eval/demon_cc_f1': 0.9394996265083454, 'eval/latency_8/demon_fire_precision': 0.9708767627222563, 'eval/latency_8/demon_fire_recall': 0.9838459148803976, 'eval/latency_8/demon_fire_f1': 0.9773183150748342, 'eval/latency_8/demon_fire_rate_model': 0.907623817473567, 'eval/latency_8/demon_fire_rate_teacher': 0.8956594323873122, 'eval/latency_8/demon_left_precision': 0.9225, 'eval/latency_8/demon_left_recall': 0.7901498929336188, 'eval/latency_8/demon_left_f1': 0.8512110726643598, 'eval/latency_8/demon_left_rate_model': 0.11129660545353366, 'eval/latency_8/demon_left_rate_teacher': 0.12993878686700056, 'eval/latency_8/demon_right_precision': 0.8656716417910447, 'eval/latency_8/demon_right_recall': 0.8787878787878788, 'eval/latency_8/demon_right_f1': 0.8721804511278195, 'eval/latency_8/demon_right_rate_model': 0.1864218141346689, 'eval/latency_8/demon_right_rate_teacher': 0.18363939899833054, 'eval/latency_8/demon_move_f1': 0.8616957618960897, 'eval/latency_8/demon_cc_f1': 0.9394996265083454, 'timing/data': 0.0004856623709201813, 'timing/model': 1.1414531664922833, 'learning_rate/qwen_vl_interface': 5.442417563274289e-07, 'learning_rate/action_model': 5.442417563274289e-06, 'epoch': 3.04}) +[2026-06-18 02:59:38,102][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_6700_state +[2026-06-18 02:59:38,108][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 02:59:44,401][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_6700_state/model.safetensors +[2026-06-18 02:59:59,817][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_6700_state/optimizer.bin +[2026-06-18 02:59:59,820][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_6700_state/sampler.bin +[2026-06-18 02:59:59,821][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_6700_state/sampler_1.bin +[2026-06-18 02:59:59,824][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_6700_state/random_states_0.pkl +[2026-06-18 02:59:59,827][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 6700 +[2026-06-18 02:59:59,827][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 02:59:59,837][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 03:00:03,592][starVLA.training.train_starvla][INFO] - Step 6701, Loss: {'train/loss': 0.020963594317436218, 'train/grad_norm_pre_clip': 5.547630310058594, 'timing/data': 0.0005485527217388153, 'timing/model': 1.0945955188944936, 'learning_rate/qwen_vl_interface': 5.439477577606429e-07, 'learning_rate/action_model': 5.4394775776064295e-06, 'epoch': 3.04}) +[2026-06-18 03:00:07,387][starVLA.training.train_starvla][INFO] - Step 6702, Loss: {'train/loss': 0.056693438440561295, 'train/grad_norm_pre_clip': 14.78531551361084, 'timing/data': 0.0005105007439851761, 'timing/model': 1.1358061339706182, 'learning_rate/qwen_vl_interface': 5.436547347635506e-07, 'learning_rate/action_model': 5.436547347635506e-06, 'epoch': 3.04}) +[2026-06-18 03:00:12,078][starVLA.training.train_starvla][INFO] - Step 6703, Loss: {'train/loss': 0.07117899507284164, 'train/grad_norm_pre_clip': 9.928475379943848, 'timing/data': 0.0007688440382480621, 'timing/model': 1.9870234318077564, 'learning_rate/qwen_vl_interface': 5.433626873968962e-07, 'learning_rate/action_model': 5.433626873968962e-06, 'epoch': 3.04}) +[2026-06-18 03:00:16,049][starVLA.training.train_starvla][INFO] - Step 6704, Loss: {'train/loss': 0.038980890065431595, 'train/grad_norm_pre_clip': 7.1815338134765625, 'timing/data': 0.0005025211721658707, 'timing/model': 0.9487194828689098, 'learning_rate/qwen_vl_interface': 5.43071615721221e-07, 'learning_rate/action_model': 5.43071615721221e-06, 'epoch': 3.04}) +[2026-06-18 03:00:20,169][starVLA.training.train_starvla][INFO] - Step 6705, Loss: {'train/loss': 0.06408166140317917, 'train/grad_norm_pre_clip': 12.033306121826172, 'timing/data': 0.0004954533651471138, 'timing/model': 0.9292885838076472, 'learning_rate/qwen_vl_interface': 5.427815197968647e-07, 'learning_rate/action_model': 5.4278151979686465e-06, 'epoch': 3.04}) +[2026-06-18 03:00:23,951][starVLA.training.train_starvla][INFO] - Step 6706, Loss: {'train/loss': 0.33190280199050903, 'train/grad_norm_pre_clip': 34.71791458129883, 'timing/data': 0.0004929853603243828, 'timing/model': 0.9233548017218709, 'learning_rate/qwen_vl_interface': 5.424923996839647e-07, 'learning_rate/action_model': 5.4249239968396466e-06, 'epoch': 3.04}) +[2026-06-18 03:00:27,726][starVLA.training.train_starvla][INFO] - Step 6707, Loss: {'train/loss': 0.23641523718833923, 'train/grad_norm_pre_clip': 31.653146743774414, 'timing/data': 0.0004939939826726913, 'timing/model': 0.9273386057466269, 'learning_rate/qwen_vl_interface': 5.422042554424557e-07, 'learning_rate/action_model': 5.4220425544245565e-06, 'epoch': 3.04}) +[2026-06-18 03:00:31,569][starVLA.training.train_starvla][INFO] - Step 6708, Loss: {'train/loss': 0.010400699451565742, 'train/grad_norm_pre_clip': 3.8374269008636475, 'timing/data': 0.0005620056763291359, 'timing/model': 1.1765478234738111, 'learning_rate/qwen_vl_interface': 5.419170871320705e-07, 'learning_rate/action_model': 5.419170871320706e-06, 'epoch': 3.04}) +[2026-06-18 03:00:35,508][starVLA.training.train_starvla][INFO] - Step 6709, Loss: {'train/loss': 0.44196373224258423, 'train/grad_norm_pre_clip': 36.864105224609375, 'timing/data': 0.000501180998980999, 'timing/model': 1.1760717742145061, 'learning_rate/qwen_vl_interface': 5.41630894812339e-07, 'learning_rate/action_model': 5.41630894812339e-06, 'epoch': 3.04}) +[2026-06-18 03:00:39,098][starVLA.training.train_starvla][INFO] - Step 6710, Loss: {'train/loss': 0.333096444606781, 'train/grad_norm_pre_clip': 36.170841217041016, 'timing/data': 0.0004981150850653648, 'timing/model': 0.9275964321568608, 'learning_rate/qwen_vl_interface': 5.413456785425896e-07, 'learning_rate/action_model': 5.413456785425896e-06, 'epoch': 3.04}) +[2026-06-18 03:00:42,837][starVLA.training.train_starvla][INFO] - Step 6711, Loss: {'train/loss': 0.34314218163490295, 'train/grad_norm_pre_clip': 24.269712448120117, 'timing/data': 0.0006006648764014244, 'timing/model': 0.9344557579606771, 'learning_rate/qwen_vl_interface': 5.410614383819474e-07, 'learning_rate/action_model': 5.4106143838194736e-06, 'epoch': 3.04}) +[2026-06-18 03:00:46,633][starVLA.training.train_starvla][INFO] - Step 6712, Loss: {'train/loss': 0.08524037897586823, 'train/grad_norm_pre_clip': 35.64956283569336, 'timing/data': 0.0008809920400381088, 'timing/model': 0.9327065134420991, 'learning_rate/qwen_vl_interface': 5.407781743893364e-07, 'learning_rate/action_model': 5.407781743893364e-06, 'epoch': 3.05}) +[2026-06-18 03:00:50,408][starVLA.training.train_starvla][INFO] - Step 6713, Loss: {'train/loss': 0.10885582864284515, 'train/grad_norm_pre_clip': 29.493257522583008, 'timing/data': 0.0011924989521503448, 'timing/model': 0.9431018820032477, 'learning_rate/qwen_vl_interface': 5.404958866234772e-07, 'learning_rate/action_model': 5.404958866234772e-06, 'epoch': 3.05}) +[2026-06-18 03:00:54,201][starVLA.training.train_starvla][INFO] - Step 6714, Loss: {'train/loss': 0.08058957755565643, 'train/grad_norm_pre_clip': 18.001148223876953, 'timing/data': 0.0008251480758190155, 'timing/model': 1.117097801528871, 'learning_rate/qwen_vl_interface': 5.402145751428884e-07, 'learning_rate/action_model': 5.402145751428884e-06, 'epoch': 3.05}) +[2026-06-18 03:00:58,240][starVLA.training.train_starvla][INFO] - Step 6715, Loss: {'train/loss': 0.03669741377234459, 'train/grad_norm_pre_clip': 10.556414604187012, 'timing/data': 0.0008688876405358315, 'timing/model': 1.0515911923721433, 'learning_rate/qwen_vl_interface': 5.399342400058856e-07, 'learning_rate/action_model': 5.399342400058856e-06, 'epoch': 3.05}) +[2026-06-18 03:01:03,721][starVLA.training.train_starvla][INFO] - Step 6716, Loss: {'train/loss': 0.25084826350212097, 'train/grad_norm_pre_clip': 36.032135009765625, 'timing/data': 0.0004578772932291031, 'timing/model': 2.6969784516841173, 'learning_rate/qwen_vl_interface': 5.396548812705838e-07, 'learning_rate/action_model': 5.396548812705839e-06, 'epoch': 3.05}) +[2026-06-18 03:01:09,509][starVLA.training.train_starvla][INFO] - Step 6717, Loss: {'train/loss': 0.02242535911500454, 'train/grad_norm_pre_clip': 4.963892936706543, 'timing/data': 0.0005007591098546982, 'timing/model': 1.0529080545529723, 'learning_rate/qwen_vl_interface': 5.393764989948933e-07, 'learning_rate/action_model': 5.393764989948933e-06, 'epoch': 3.05}) +[2026-06-18 03:01:13,570][starVLA.training.train_starvla][INFO] - Step 6718, Loss: {'train/loss': 0.12743748724460602, 'train/grad_norm_pre_clip': 23.751758575439453, 'timing/data': 0.0005550626665353775, 'timing/model': 0.9610463967546821, 'learning_rate/qwen_vl_interface': 5.390990932365234e-07, 'learning_rate/action_model': 5.390990932365233e-06, 'epoch': 3.05}) +[2026-06-18 03:01:18,269][starVLA.training.train_starvla][INFO] - Step 6719, Loss: {'train/loss': 0.131860613822937, 'train/grad_norm_pre_clip': 36.78837585449219, 'timing/data': 0.0007256567478179932, 'timing/model': 1.1808781987056136, 'learning_rate/qwen_vl_interface': 5.388226640529808e-07, 'learning_rate/action_model': 5.388226640529808e-06, 'epoch': 3.05}) +[2026-06-18 03:01:22,105][starVLA.training.train_starvla][INFO] - Step 6720, Loss: {'train/loss': 0.12535801529884338, 'train/grad_norm_pre_clip': 14.692404747009277, 'timing/data': 0.0014033755287528038, 'timing/model': 0.9546107798814774, 'learning_rate/qwen_vl_interface': 5.385472115015697e-07, 'learning_rate/action_model': 5.385472115015696e-06, 'epoch': 3.05}) +[2026-06-18 03:01:25,922][starVLA.training.train_starvla][INFO] - Step 6721, Loss: {'train/loss': 0.031202811747789383, 'train/grad_norm_pre_clip': 13.629173278808594, 'timing/data': 0.0014254050329327583, 'timing/model': 1.1302281422540545, 'learning_rate/qwen_vl_interface': 5.382727356393909e-07, 'learning_rate/action_model': 5.382727356393909e-06, 'epoch': 3.05}) +[2026-06-18 03:01:29,587][starVLA.training.train_starvla][INFO] - Step 6722, Loss: {'train/loss': 0.45385223627090454, 'train/grad_norm_pre_clip': 52.27024841308594, 'timing/data': 0.00048511289060115814, 'timing/model': 0.9311210755258799, 'learning_rate/qwen_vl_interface': 5.379992365233442e-07, 'learning_rate/action_model': 5.379992365233442e-06, 'epoch': 3.05}) +[2026-06-18 03:01:33,509][starVLA.training.train_starvla][INFO] - Step 6723, Loss: {'train/loss': 0.20358461141586304, 'train/grad_norm_pre_clip': 44.323856353759766, 'timing/data': 0.0012343721464276314, 'timing/model': 0.9514667969197035, 'learning_rate/qwen_vl_interface': 5.377267142101261e-07, 'learning_rate/action_model': 5.37726714210126e-06, 'epoch': 3.05}) +[2026-06-18 03:01:37,297][starVLA.training.train_starvla][INFO] - Step 6724, Loss: {'train/loss': 0.34458598494529724, 'train/grad_norm_pre_clip': 23.979162216186523, 'timing/data': 0.0006284955888986588, 'timing/model': 0.9317515920847654, 'learning_rate/qwen_vl_interface': 5.374551687562306e-07, 'learning_rate/action_model': 5.374551687562306e-06, 'epoch': 3.05}) +[2026-06-18 03:01:41,251][starVLA.training.train_starvla][INFO] - Step 6725, Loss: {'train/loss': 0.02705475129187107, 'train/grad_norm_pre_clip': 7.073597431182861, 'timing/data': 0.0005014380440115929, 'timing/model': 0.9330306928604841, 'learning_rate/qwen_vl_interface': 5.371846002179494e-07, 'learning_rate/action_model': 5.371846002179493e-06, 'epoch': 3.05}) +[2026-06-18 03:01:45,055][starVLA.training.train_starvla][INFO] - Step 6726, Loss: {'train/loss': 0.15765002369880676, 'train/grad_norm_pre_clip': 16.134645462036133, 'timing/data': 0.0004497496411204338, 'timing/model': 0.936580658890307, 'learning_rate/qwen_vl_interface': 5.36915008651372e-07, 'learning_rate/action_model': 5.369150086513719e-06, 'epoch': 3.05}) +[2026-06-18 03:01:48,935][starVLA.training.train_starvla][INFO] - Step 6727, Loss: {'train/loss': 0.06010596081614494, 'train/grad_norm_pre_clip': 17.426620483398438, 'timing/data': 0.0004734927788376808, 'timing/model': 1.1207831846550107, 'learning_rate/qwen_vl_interface': 5.366463941123853e-07, 'learning_rate/action_model': 5.366463941123853e-06, 'epoch': 3.05}) +[2026-06-18 03:01:52,768][starVLA.training.train_starvla][INFO] - Step 6728, Loss: {'train/loss': 0.1316310465335846, 'train/grad_norm_pre_clip': 21.451637268066406, 'timing/data': 0.0005163112655282021, 'timing/model': 1.1263695089146495, 'learning_rate/qwen_vl_interface': 5.36378756656672e-07, 'learning_rate/action_model': 5.36378756656672e-06, 'epoch': 3.05}) +[2026-06-18 03:01:56,569][starVLA.training.train_starvla][INFO] - Step 6729, Loss: {'train/loss': 0.08883039653301239, 'train/grad_norm_pre_clip': 24.159164428710938, 'timing/data': 0.000856972299516201, 'timing/model': 1.0185318477451801, 'learning_rate/qwen_vl_interface': 5.361120963397151e-07, 'learning_rate/action_model': 5.361120963397152e-06, 'epoch': 3.05}) +[2026-06-18 03:02:00,448][starVLA.training.train_starvla][INFO] - Step 6730, Loss: {'train/loss': 0.15271052718162537, 'train/grad_norm_pre_clip': 19.97819709777832, 'timing/data': 0.0011324156075716019, 'timing/model': 0.9518134659156203, 'learning_rate/qwen_vl_interface': 5.358464132167927e-07, 'learning_rate/action_model': 5.358464132167927e-06, 'epoch': 3.05}) +[2026-06-18 03:02:04,308][starVLA.training.train_starvla][INFO] - Step 6731, Loss: {'train/loss': 0.07756348699331284, 'train/grad_norm_pre_clip': 22.081159591674805, 'timing/data': 0.0005032923072576523, 'timing/model': 0.9640423124656081, 'learning_rate/qwen_vl_interface': 5.355817073429818e-07, 'learning_rate/action_model': 5.355817073429818e-06, 'epoch': 3.05}) +[2026-06-18 03:02:08,332][starVLA.training.train_starvla][INFO] - Step 6732, Loss: {'train/loss': 0.16790200769901276, 'train/grad_norm_pre_clip': 29.417566299438477, 'timing/data': 0.0004969574511051178, 'timing/model': 0.9501699646934867, 'learning_rate/qwen_vl_interface': 5.353179787731557e-07, 'learning_rate/action_model': 5.353179787731557e-06, 'epoch': 3.05}) +[2026-06-18 03:02:12,165][starVLA.training.train_starvla][INFO] - Step 6733, Loss: {'train/loss': 0.1506487876176834, 'train/grad_norm_pre_clip': 41.742427825927734, 'timing/data': 0.0005187308415770531, 'timing/model': 0.9377224249765277, 'learning_rate/qwen_vl_interface': 5.350552275619857e-07, 'learning_rate/action_model': 5.350552275619857e-06, 'epoch': 3.05}) +[2026-06-18 03:02:16,168][starVLA.training.train_starvla][INFO] - Step 6734, Loss: {'train/loss': 0.08362024277448654, 'train/grad_norm_pre_clip': 24.33376121520996, 'timing/data': 0.00048606470227241516, 'timing/model': 1.2048791470006108, 'learning_rate/qwen_vl_interface': 5.347934537639405e-07, 'learning_rate/action_model': 5.347934537639405e-06, 'epoch': 3.06}) +[2026-06-18 03:02:20,022][starVLA.training.train_starvla][INFO] - Step 6735, Loss: {'train/loss': 0.07798884809017181, 'train/grad_norm_pre_clip': 21.21169662475586, 'timing/data': 0.0005106786265969276, 'timing/model': 1.0724768741056323, 'learning_rate/qwen_vl_interface': 5.345326574332865e-07, 'learning_rate/action_model': 5.345326574332865e-06, 'epoch': 3.06}) +[2026-06-18 03:02:23,677][starVLA.training.train_starvla][INFO] - Step 6736, Loss: {'train/loss': 0.20815089344978333, 'train/grad_norm_pre_clip': 28.09388542175293, 'timing/data': 0.0006358902901411057, 'timing/model': 0.9327771430835128, 'learning_rate/qwen_vl_interface': 5.342728386240867e-07, 'learning_rate/action_model': 5.3427283862408665e-06, 'epoch': 3.06}) +[2026-06-18 03:02:27,545][starVLA.training.train_starvla][INFO] - Step 6737, Loss: {'train/loss': 0.3177136778831482, 'train/grad_norm_pre_clip': 42.24947738647461, 'timing/data': 0.000497954897582531, 'timing/model': 0.9356301398947835, 'learning_rate/qwen_vl_interface': 5.340139973902013e-07, 'learning_rate/action_model': 5.340139973902013e-06, 'epoch': 3.06}) +[2026-06-18 03:02:31,438][starVLA.training.train_starvla][INFO] - Step 6738, Loss: {'train/loss': 0.14896991848945618, 'train/grad_norm_pre_clip': 29.177288055419922, 'timing/data': 0.0005287742242217064, 'timing/model': 0.960783981718123, 'learning_rate/qwen_vl_interface': 5.337561337852892e-07, 'learning_rate/action_model': 5.337561337852892e-06, 'epoch': 3.06}) +[2026-06-18 03:02:35,174][starVLA.training.train_starvla][INFO] - Step 6739, Loss: {'train/loss': 0.20640170574188232, 'train/grad_norm_pre_clip': 22.46399688720703, 'timing/data': 0.0005322601646184921, 'timing/model': 0.9337664525955915, 'learning_rate/qwen_vl_interface': 5.334992478628054e-07, 'learning_rate/action_model': 5.334992478628054e-06, 'epoch': 3.06}) +[2026-06-18 03:02:38,958][starVLA.training.train_starvla][INFO] - Step 6740, Loss: {'train/loss': 0.43937021493911743, 'train/grad_norm_pre_clip': 37.17523956298828, 'timing/data': 0.0004709949716925621, 'timing/model': 0.9399655004963279, 'learning_rate/qwen_vl_interface': 5.332433396760027e-07, 'learning_rate/action_model': 5.332433396760028e-06, 'epoch': 3.06}) +[2026-06-18 03:02:42,817][starVLA.training.train_starvla][INFO] - Step 6741, Loss: {'train/loss': 0.3036922216415405, 'train/grad_norm_pre_clip': 47.11151123046875, 'timing/data': 0.0007780147716403008, 'timing/model': 1.124794657342136, 'learning_rate/qwen_vl_interface': 5.329884092779312e-07, 'learning_rate/action_model': 5.329884092779312e-06, 'epoch': 3.06}) +[2026-06-18 03:02:46,552][starVLA.training.train_starvla][INFO] - Step 6742, Loss: {'train/loss': 0.014668687246739864, 'train/grad_norm_pre_clip': 4.100611686706543, 'timing/data': 0.0004883958026766777, 'timing/model': 1.015527380630374, 'learning_rate/qwen_vl_interface': 5.327344567214386e-07, 'learning_rate/action_model': 5.327344567214385e-06, 'epoch': 3.06}) +[2026-06-18 03:02:50,869][starVLA.training.train_starvla][INFO] - Step 6743, Loss: {'train/loss': 0.021249381825327873, 'train/grad_norm_pre_clip': 7.399304389953613, 'timing/data': 0.0006359517574310303, 'timing/model': 1.2036105394363403, 'learning_rate/qwen_vl_interface': 5.324814820591683e-07, 'learning_rate/action_model': 5.324814820591682e-06, 'epoch': 3.06}) +[2026-06-18 03:02:54,760][starVLA.training.train_starvla][INFO] - Step 6744, Loss: {'train/loss': 0.2146281749010086, 'train/grad_norm_pre_clip': 16.06775665283203, 'timing/data': 0.0008326917886734009, 'timing/model': 0.9547137152403593, 'learning_rate/qwen_vl_interface': 5.322294853435631e-07, 'learning_rate/action_model': 5.322294853435631e-06, 'epoch': 3.06}) +[2026-06-18 03:02:58,701][starVLA.training.train_starvla][INFO] - Step 6745, Loss: {'train/loss': 0.20421102643013, 'train/grad_norm_pre_clip': 30.285171508789062, 'timing/data': 0.0005062632262706757, 'timing/model': 0.9495669808238745, 'learning_rate/qwen_vl_interface': 5.319784666268626e-07, 'learning_rate/action_model': 5.319784666268626e-06, 'epoch': 3.06}) +[2026-06-18 03:03:02,505][starVLA.training.train_starvla][INFO] - Step 6746, Loss: {'train/loss': 0.12227137386798859, 'train/grad_norm_pre_clip': 27.923091888427734, 'timing/data': 0.0004917187616229057, 'timing/model': 0.9518654169514775, 'learning_rate/qwen_vl_interface': 5.317284259611024e-07, 'learning_rate/action_model': 5.317284259611024e-06, 'epoch': 3.06}) +[2026-06-18 03:03:06,333][starVLA.training.train_starvla][INFO] - Step 6747, Loss: {'train/loss': 0.4889020025730133, 'train/grad_norm_pre_clip': 34.03879928588867, 'timing/data': 0.0008049001917243004, 'timing/model': 1.1497064400464296, 'learning_rate/qwen_vl_interface': 5.314793633981169e-07, 'learning_rate/action_model': 5.314793633981169e-06, 'epoch': 3.06}) +[2026-06-18 03:03:10,257][starVLA.training.train_starvla][INFO] - Step 6748, Loss: {'train/loss': 0.19166728854179382, 'train/grad_norm_pre_clip': 30.534637451171875, 'timing/data': 0.000497741624712944, 'timing/model': 1.2019101502373815, 'learning_rate/qwen_vl_interface': 5.312312789895369e-07, 'learning_rate/action_model': 5.3123127898953685e-06, 'epoch': 3.06}) +[2026-06-18 03:03:14,042][starVLA.training.train_starvla][INFO] - Step 6749, Loss: {'train/loss': 0.050326768308877945, 'train/grad_norm_pre_clip': 21.77652359008789, 'timing/data': 0.0006798114627599716, 'timing/model': 0.9342694003134966, 'learning_rate/qwen_vl_interface': 5.3098417278679e-07, 'learning_rate/action_model': 5.3098417278679e-06, 'epoch': 3.06}) +[2026-06-18 03:03:18,021][starVLA.training.train_starvla][INFO] - Step 6750, Loss: {'train/loss': 0.09729543328285217, 'train/grad_norm_pre_clip': 16.847341537475586, 'timing/data': 0.00042275991290807724, 'timing/model': 0.9536245940253139, 'learning_rate/qwen_vl_interface': 5.307380448411022e-07, 'learning_rate/action_model': 5.307380448411022e-06, 'epoch': 3.06}) +[2026-06-18 03:03:21,867][starVLA.training.train_starvla][INFO] - Step 6751, Loss: {'train/loss': 0.4342808723449707, 'train/grad_norm_pre_clip': 36.4622802734375, 'timing/data': 0.0004926193505525589, 'timing/model': 0.9825393818318844, 'learning_rate/qwen_vl_interface': 5.304928952034955e-07, 'learning_rate/action_model': 5.304928952034956e-06, 'epoch': 3.06}) +[2026-06-18 03:03:25,886][starVLA.training.train_starvla][INFO] - Step 6752, Loss: {'train/loss': 0.18598896265029907, 'train/grad_norm_pre_clip': 30.383764266967773, 'timing/data': 0.0007659867405891418, 'timing/model': 0.9795355880632997, 'learning_rate/qwen_vl_interface': 5.302487239247905e-07, 'learning_rate/action_model': 5.302487239247904e-06, 'epoch': 3.06}) +[2026-06-18 03:03:29,745][starVLA.training.train_starvla][INFO] - Step 6753, Loss: {'train/loss': 0.09447110444307327, 'train/grad_norm_pre_clip': 14.990657806396484, 'timing/data': 0.0005814032629132271, 'timing/model': 0.9364043790847063, 'learning_rate/qwen_vl_interface': 5.300055310556034e-07, 'learning_rate/action_model': 5.300055310556034e-06, 'epoch': 3.06}) +[2026-06-18 03:03:33,563][starVLA.training.train_starvla][INFO] - Step 6754, Loss: {'train/loss': 0.08184657245874405, 'train/grad_norm_pre_clip': 29.350215911865234, 'timing/data': 0.00045608170330524445, 'timing/model': 1.0979464733973145, 'learning_rate/qwen_vl_interface': 5.297633166463482e-07, 'learning_rate/action_model': 5.297633166463481e-06, 'epoch': 3.06}) +[2026-06-18 03:03:37,486][starVLA.training.train_starvla][INFO] - Step 6755, Loss: {'train/loss': 0.15196557343006134, 'train/grad_norm_pre_clip': 33.53521728515625, 'timing/data': 0.0005036639049649239, 'timing/model': 1.205059110186994, 'learning_rate/qwen_vl_interface': 5.295220807472373e-07, 'learning_rate/action_model': 5.295220807472374e-06, 'epoch': 3.06}) +[2026-06-18 03:03:41,287][starVLA.training.train_starvla][INFO] - Step 6756, Loss: {'train/loss': 0.14726416766643524, 'train/grad_norm_pre_clip': 14.887232780456543, 'timing/data': 0.0006736675277352333, 'timing/model': 0.9573034876957536, 'learning_rate/qwen_vl_interface': 5.292818234082786e-07, 'learning_rate/action_model': 5.292818234082787e-06, 'epoch': 3.07}) +[2026-06-18 03:03:45,198][starVLA.training.train_starvla][INFO] - Step 6757, Loss: {'train/loss': 0.016899580135941505, 'train/grad_norm_pre_clip': 5.400786876678467, 'timing/data': 0.0006324369460344315, 'timing/model': 0.9731258638203144, 'learning_rate/qwen_vl_interface': 5.290425446792772e-07, 'learning_rate/action_model': 5.2904254467927716e-06, 'epoch': 3.07}) +[2026-06-18 03:03:49,135][starVLA.training.train_starvla][INFO] - Step 6758, Loss: {'train/loss': 0.4302145540714264, 'train/grad_norm_pre_clip': 50.32695388793945, 'timing/data': 0.0004808194935321808, 'timing/model': 1.0684256665408611, 'learning_rate/qwen_vl_interface': 5.288042446098364e-07, 'learning_rate/action_model': 5.2880424460983634e-06, 'epoch': 3.07}) +[2026-06-18 03:03:53,022][starVLA.training.train_starvla][INFO] - Step 6759, Loss: {'train/loss': 0.1737077385187149, 'train/grad_norm_pre_clip': 31.241539001464844, 'timing/data': 0.0005181431770324707, 'timing/model': 0.9335063993930817, 'learning_rate/qwen_vl_interface': 5.28566923249356e-07, 'learning_rate/action_model': 5.28566923249356e-06, 'epoch': 3.07}) +[2026-06-18 03:03:56,855][starVLA.training.train_starvla][INFO] - Step 6760, Loss: {'train/loss': 0.20263418555259705, 'train/grad_norm_pre_clip': 33.086856842041016, 'timing/data': 0.0004975665360689163, 'timing/model': 0.9308504443615675, 'learning_rate/qwen_vl_interface': 5.283305806470329e-07, 'learning_rate/action_model': 5.2833058064703286e-06, 'epoch': 3.07}) +[2026-06-18 03:04:00,684][starVLA.training.train_starvla][INFO] - Step 6761, Loss: {'train/loss': 0.2842363715171814, 'train/grad_norm_pre_clip': 24.603498458862305, 'timing/data': 0.0004854472354054451, 'timing/model': 1.1484667900949717, 'learning_rate/qwen_vl_interface': 5.280952168518612e-07, 'learning_rate/action_model': 5.280952168518612e-06, 'epoch': 3.07}) +[2026-06-18 03:04:04,300][starVLA.training.train_starvla][INFO] - Step 6762, Loss: {'train/loss': 0.19879983365535736, 'train/grad_norm_pre_clip': 26.812265396118164, 'timing/data': 0.0005179941654205322, 'timing/model': 0.934265567921102, 'learning_rate/qwen_vl_interface': 5.278608319126318e-07, 'learning_rate/action_model': 5.278608319126318e-06, 'epoch': 3.07}) +[2026-06-18 03:04:08,116][starVLA.training.train_starvla][INFO] - Step 6763, Loss: {'train/loss': 0.040511492639780045, 'train/grad_norm_pre_clip': 10.632744789123535, 'timing/data': 0.0005872556939721107, 'timing/model': 0.9307734230533242, 'learning_rate/qwen_vl_interface': 5.276274258779332e-07, 'learning_rate/action_model': 5.2762742587793315e-06, 'epoch': 3.07}) +[2026-06-18 03:04:11,860][starVLA.training.train_starvla][INFO] - Step 6764, Loss: {'train/loss': 0.09053418040275574, 'train/grad_norm_pre_clip': 14.501187324523926, 'timing/data': 0.0008690021932125092, 'timing/model': 0.9342901390045881, 'learning_rate/qwen_vl_interface': 5.273949987961508e-07, 'learning_rate/action_model': 5.273949987961508e-06, 'epoch': 3.07}) +[2026-06-18 03:04:15,934][starVLA.training.train_starvla][INFO] - Step 6765, Loss: {'train/loss': 0.29554587602615356, 'train/grad_norm_pre_clip': 54.563926696777344, 'timing/data': 0.0009513013064861298, 'timing/model': 0.9384500840678811, 'learning_rate/qwen_vl_interface': 5.271635507154671e-07, 'learning_rate/action_model': 5.2716355071546706e-06, 'epoch': 3.07}) +[2026-06-18 03:04:19,936][starVLA.training.train_starvla][INFO] - Step 6766, Loss: {'train/loss': 0.10602355003356934, 'train/grad_norm_pre_clip': 20.232177734375, 'timing/data': 0.0007467698305845261, 'timing/model': 0.9332632701843977, 'learning_rate/qwen_vl_interface': 5.269330816838614e-07, 'learning_rate/action_model': 5.269330816838613e-06, 'epoch': 3.07}) +[2026-06-18 03:04:23,740][starVLA.training.train_starvla][INFO] - Step 6767, Loss: {'train/loss': 0.10504346340894699, 'train/grad_norm_pre_clip': 18.136980056762695, 'timing/data': 0.0008377861231565475, 'timing/model': 0.9373743608593941, 'learning_rate/qwen_vl_interface': 5.267035917491095e-07, 'learning_rate/action_model': 5.267035917491094e-06, 'epoch': 3.07}) +[2026-06-18 03:04:27,961][starVLA.training.train_starvla][INFO] - Step 6768, Loss: {'train/loss': 0.25782468914985657, 'train/grad_norm_pre_clip': 26.429195404052734, 'timing/data': 0.0004923418164253235, 'timing/model': 1.3806710569187999, 'learning_rate/qwen_vl_interface': 5.264750809587862e-07, 'learning_rate/action_model': 5.264750809587863e-06, 'epoch': 3.07}) +[2026-06-18 03:04:31,688][starVLA.training.train_starvla][INFO] - Step 6769, Loss: {'train/loss': 0.16016893088817596, 'train/grad_norm_pre_clip': 20.645832061767578, 'timing/data': 0.0004936661571264267, 'timing/model': 0.9631862677633762, 'learning_rate/qwen_vl_interface': 5.262475493602609e-07, 'learning_rate/action_model': 5.26247549360261e-06, 'epoch': 3.07}) +[2026-06-18 03:04:36,141][starVLA.training.train_starvla][INFO] - Step 6770, Loss: {'train/loss': 0.10003051906824112, 'train/grad_norm_pre_clip': 22.428022384643555, 'timing/data': 0.0007362598553299904, 'timing/model': 1.2410263773053885, 'learning_rate/qwen_vl_interface': 5.260209970007018e-07, 'learning_rate/action_model': 5.260209970007018e-06, 'epoch': 3.07}) +[2026-06-18 03:04:40,605][starVLA.training.train_starvla][INFO] - Step 6771, Loss: {'train/loss': 0.21654528379440308, 'train/grad_norm_pre_clip': 41.324806213378906, 'timing/data': 0.0005992241203784943, 'timing/model': 1.1745530432090163, 'learning_rate/qwen_vl_interface': 5.257954239270732e-07, 'learning_rate/action_model': 5.257954239270732e-06, 'epoch': 3.07}) +[2026-06-18 03:04:44,681][starVLA.training.train_starvla][INFO] - Step 6772, Loss: {'train/loss': 0.042829837650060654, 'train/grad_norm_pre_clip': 13.677419662475586, 'timing/data': 0.0008136788383126259, 'timing/model': 0.9517331793904305, 'learning_rate/qwen_vl_interface': 5.255708301861367e-07, 'learning_rate/action_model': 5.255708301861367e-06, 'epoch': 3.07}) +[2026-06-18 03:04:48,860][starVLA.training.train_starvla][INFO] - Step 6773, Loss: {'train/loss': 0.056572895497083664, 'train/grad_norm_pre_clip': 18.187381744384766, 'timing/data': 0.0011891517788171768, 'timing/model': 0.9394320836290717, 'learning_rate/qwen_vl_interface': 5.253472158244506e-07, 'learning_rate/action_model': 5.253472158244505e-06, 'epoch': 3.07}) +[2026-06-18 03:04:52,865][starVLA.training.train_starvla][INFO] - Step 6774, Loss: {'train/loss': 0.11364727467298508, 'train/grad_norm_pre_clip': 23.11824607849121, 'timing/data': 0.0005525127053260803, 'timing/model': 1.194136245176196, 'learning_rate/qwen_vl_interface': 5.251245808883707e-07, 'learning_rate/action_model': 5.251245808883707e-06, 'epoch': 3.07}) +[2026-06-18 03:04:56,813][starVLA.training.train_starvla][INFO] - Step 6775, Loss: {'train/loss': 0.28392642736434937, 'train/grad_norm_pre_clip': 25.826047897338867, 'timing/data': 0.0008710157126188278, 'timing/model': 1.2177911922335625, 'learning_rate/qwen_vl_interface': 5.249029254240495e-07, 'learning_rate/action_model': 5.249029254240495e-06, 'epoch': 3.07}) +[2026-06-18 03:05:00,437][starVLA.training.train_starvla][INFO] - Step 6776, Loss: {'train/loss': 0.11415521055459976, 'train/grad_norm_pre_clip': 26.717863082885742, 'timing/data': 0.0005146749317646027, 'timing/model': 0.9517078408971429, 'learning_rate/qwen_vl_interface': 5.246822494774364e-07, 'learning_rate/action_model': 5.246822494774363e-06, 'epoch': 3.07}) +[2026-06-18 03:05:04,328][starVLA.training.train_starvla][INFO] - Step 6777, Loss: {'train/loss': 0.1540321707725525, 'train/grad_norm_pre_clip': 31.607280731201172, 'timing/data': 0.0006945403292775154, 'timing/model': 0.9375513410195708, 'learning_rate/qwen_vl_interface': 5.244625530942774e-07, 'learning_rate/action_model': 5.244625530942774e-06, 'epoch': 3.07}) +[2026-06-18 03:05:08,205][starVLA.training.train_starvla][INFO] - Step 6778, Loss: {'train/loss': 0.07803113013505936, 'train/grad_norm_pre_clip': 20.9339656829834, 'timing/data': 0.0004840157926082611, 'timing/model': 0.9378688354045153, 'learning_rate/qwen_vl_interface': 5.242438363201161e-07, 'learning_rate/action_model': 5.2424383632011605e-06, 'epoch': 3.07}) +[2026-06-18 03:05:12,040][starVLA.training.train_starvla][INFO] - Step 6779, Loss: {'train/loss': 0.3064994215965271, 'train/grad_norm_pre_clip': 39.46403884887695, 'timing/data': 0.0005988730117678642, 'timing/model': 0.9323594495654106, 'learning_rate/qwen_vl_interface': 5.240260992002928e-07, 'learning_rate/action_model': 5.240260992002928e-06, 'epoch': 3.08}) +[2026-06-18 03:05:15,961][starVLA.training.train_starvla][INFO] - Step 6780, Loss: {'train/loss': 0.03912771865725517, 'train/grad_norm_pre_clip': 10.666722297668457, 'timing/data': 0.0004823068156838417, 'timing/model': 0.9494866756722331, 'learning_rate/qwen_vl_interface': 5.238093417799446e-07, 'learning_rate/action_model': 5.238093417799446e-06, 'epoch': 3.08}) +[2026-06-18 03:05:19,889][starVLA.training.train_starvla][INFO] - Step 6781, Loss: {'train/loss': 0.01802200637757778, 'train/grad_norm_pre_clip': 4.711869716644287, 'timing/data': 0.0007657948881387711, 'timing/model': 1.187866222113371, 'learning_rate/qwen_vl_interface': 5.235935641040048e-07, 'learning_rate/action_model': 5.235935641040048e-06, 'epoch': 3.08}) +[2026-06-18 03:05:23,734][starVLA.training.train_starvla][INFO] - Step 6782, Loss: {'train/loss': 0.21180163323879242, 'train/grad_norm_pre_clip': 22.29521942138672, 'timing/data': 0.00047521665692329407, 'timing/model': 1.1593263009563088, 'learning_rate/qwen_vl_interface': 5.233787662172059e-07, 'learning_rate/action_model': 5.23378766217206e-06, 'epoch': 3.08}) +[2026-06-18 03:05:27,535][starVLA.training.train_starvla][INFO] - Step 6783, Loss: {'train/loss': 0.06486500054597855, 'train/grad_norm_pre_clip': 32.872283935546875, 'timing/data': 0.0012176921591162682, 'timing/model': 0.9954393496736884, 'learning_rate/qwen_vl_interface': 5.231649481640741e-07, 'learning_rate/action_model': 5.231649481640741e-06, 'epoch': 3.08}) +[2026-06-18 03:05:31,537][starVLA.training.train_starvla][INFO] - Step 6784, Loss: {'train/loss': 0.05433173105120659, 'train/grad_norm_pre_clip': 13.128453254699707, 'timing/data': 0.0009871972724795341, 'timing/model': 0.9437721595168114, 'learning_rate/qwen_vl_interface': 5.229521099889351e-07, 'learning_rate/action_model': 5.229521099889352e-06, 'epoch': 3.08}) +[2026-06-18 03:05:35,570][starVLA.training.train_starvla][INFO] - Step 6785, Loss: {'train/loss': 0.04345010966062546, 'train/grad_norm_pre_clip': 11.21873664855957, 'timing/data': 0.0006768554449081421, 'timing/model': 0.9585207663476467, 'learning_rate/qwen_vl_interface': 5.227402517359105e-07, 'learning_rate/action_model': 5.2274025173591045e-06, 'epoch': 3.08}) +[2026-06-18 03:05:39,629][starVLA.training.train_starvla][INFO] - Step 6786, Loss: {'train/loss': 0.22288121283054352, 'train/grad_norm_pre_clip': 24.523096084594727, 'timing/data': 0.0005287313833832741, 'timing/model': 0.9567891443148255, 'learning_rate/qwen_vl_interface': 5.225293734489182e-07, 'learning_rate/action_model': 5.225293734489181e-06, 'epoch': 3.08}) +[2026-06-18 03:05:43,560][starVLA.training.train_starvla][INFO] - Step 6787, Loss: {'train/loss': 0.1984390914440155, 'train/grad_norm_pre_clip': 34.88840866088867, 'timing/data': 0.00048599857836961746, 'timing/model': 0.9531926782801747, 'learning_rate/qwen_vl_interface': 5.223194751716733e-07, 'learning_rate/action_model': 5.223194751716733e-06, 'epoch': 3.08}) +[2026-06-18 03:05:47,599][starVLA.training.train_starvla][INFO] - Step 6788, Loss: {'train/loss': 0.0834459513425827, 'train/grad_norm_pre_clip': 24.59333610534668, 'timing/data': 0.0008779196068644524, 'timing/model': 1.1939677372574806, 'learning_rate/qwen_vl_interface': 5.221105569476893e-07, 'learning_rate/action_model': 5.221105569476892e-06, 'epoch': 3.08}) +[2026-06-18 03:05:51,354][starVLA.training.train_starvla][INFO] - Step 6789, Loss: {'train/loss': 0.6312658190727234, 'train/grad_norm_pre_clip': 38.220359802246094, 'timing/data': 0.0004916554316878319, 'timing/model': 0.9585801539942622, 'learning_rate/qwen_vl_interface': 5.21902618820274e-07, 'learning_rate/action_model': 5.21902618820274e-06, 'epoch': 3.08}) +[2026-06-18 03:05:55,343][starVLA.training.train_starvla][INFO] - Step 6790, Loss: {'train/loss': 0.15782992541790009, 'train/grad_norm_pre_clip': 23.356643676757812, 'timing/data': 0.0006244713440537453, 'timing/model': 0.9617385501042008, 'learning_rate/qwen_vl_interface': 5.216956608325331e-07, 'learning_rate/action_model': 5.2169566083253304e-06, 'epoch': 3.08}) +[2026-06-18 03:05:59,166][starVLA.training.train_starvla][INFO] - Step 6791, Loss: {'train/loss': 0.08509169518947601, 'train/grad_norm_pre_clip': 20.22994613647461, 'timing/data': 0.0005366271361708641, 'timing/model': 0.9471821952611208, 'learning_rate/qwen_vl_interface': 5.2148968302737e-07, 'learning_rate/action_model': 5.2148968302737005e-06, 'epoch': 3.08}) +[2026-06-18 03:06:03,057][starVLA.training.train_starvla][INFO] - Step 6792, Loss: {'train/loss': 0.2035597711801529, 'train/grad_norm_pre_clip': 24.61697006225586, 'timing/data': 0.001221061684191227, 'timing/model': 0.952243429608643, 'learning_rate/qwen_vl_interface': 5.212846854474834e-07, 'learning_rate/action_model': 5.212846854474833e-06, 'epoch': 3.08}) +[2026-06-18 03:06:07,114][starVLA.training.train_starvla][INFO] - Step 6793, Loss: {'train/loss': 0.1390710473060608, 'train/grad_norm_pre_clip': 14.682451248168945, 'timing/data': 0.000468703918159008, 'timing/model': 0.9661751929670572, 'learning_rate/qwen_vl_interface': 5.2108066813537e-07, 'learning_rate/action_model': 5.2108066813537e-06, 'epoch': 3.08}) +[2026-06-18 03:06:11,068][starVLA.training.train_starvla][INFO] - Step 6794, Loss: {'train/loss': 0.44492262601852417, 'train/grad_norm_pre_clip': 34.04039764404297, 'timing/data': 0.00048772338777780533, 'timing/model': 1.1167223425582051, 'learning_rate/qwen_vl_interface': 5.208776311333223e-07, 'learning_rate/action_model': 5.208776311333223e-06, 'epoch': 3.08}) +[2026-06-18 03:06:15,888][starVLA.training.train_starvla][INFO] - Step 6795, Loss: {'train/loss': 0.1795651614665985, 'train/grad_norm_pre_clip': 27.901290893554688, 'timing/data': 0.0007272055372595787, 'timing/model': 1.127673259936273, 'learning_rate/qwen_vl_interface': 5.206755744834312e-07, 'learning_rate/action_model': 5.2067557448343125e-06, 'epoch': 3.08}) +[2026-06-18 03:06:19,838][starVLA.training.train_starvla][INFO] - Step 6796, Loss: {'train/loss': 0.12818044424057007, 'train/grad_norm_pre_clip': 26.921356201171875, 'timing/data': 0.0008895862847566605, 'timing/model': 0.9469527062028646, 'learning_rate/qwen_vl_interface': 5.20474498227582e-07, 'learning_rate/action_model': 5.2047449822758196e-06, 'epoch': 3.08}) +[2026-06-18 03:06:23,626][starVLA.training.train_starvla][INFO] - Step 6797, Loss: {'train/loss': 0.2587810456752777, 'train/grad_norm_pre_clip': 28.432708740234375, 'timing/data': 0.0008218465372920036, 'timing/model': 0.9445595340803266, 'learning_rate/qwen_vl_interface': 5.202744024074583e-07, 'learning_rate/action_model': 5.202744024074584e-06, 'epoch': 3.08}) +[2026-06-18 03:06:27,428][starVLA.training.train_starvla][INFO] - Step 6798, Loss: {'train/loss': 0.2892395555973053, 'train/grad_norm_pre_clip': 26.052236557006836, 'timing/data': 0.0005773575976490974, 'timing/model': 0.9293778482824564, 'learning_rate/qwen_vl_interface': 5.200752870645407e-07, 'learning_rate/action_model': 5.200752870645406e-06, 'epoch': 3.08}) +[2026-06-18 03:06:31,184][starVLA.training.train_starvla][INFO] - Step 6799, Loss: {'train/loss': 0.0819995105266571, 'train/grad_norm_pre_clip': 18.96651840209961, 'timing/data': 0.0004948480054736137, 'timing/model': 0.9505814332515001, 'learning_rate/qwen_vl_interface': 5.198771522401054e-07, 'learning_rate/action_model': 5.198771522401054e-06, 'epoch': 3.08}) +[2026-06-18 03:08:42,764][starVLA.training.train_starvla][INFO] - Step 6800, Loss: {'train/loss': 0.3018222749233246, 'train/grad_norm_pre_clip': 43.21131134033203, 'eval/loss': 0.2567683458328247, 'eval/latency_8/loss': 0.2567683458328247, 'eval/demon_fire_precision': 0.9699846860643185, 'eval/demon_fire_recall': 0.9838459148803976, 'eval/demon_fire_f1': 0.9768661320172732, 'eval/demon_fire_rate_model': 0.9084585420144685, 'eval/demon_fire_rate_teacher': 0.8956594323873122, 'eval/demon_left_precision': 0.9154589371980676, 'eval/demon_left_recall': 0.8115631691648822, 'eval/demon_left_f1': 0.8603859250851306, 'eval/demon_left_rate_model': 0.11519198664440734, 'eval/demon_left_rate_teacher': 0.12993878686700056, 'eval/demon_right_precision': 0.8689759036144579, 'eval/demon_right_recall': 0.8742424242424243, 'eval/demon_right_f1': 0.8716012084592145, 'eval/demon_right_rate_model': 0.18475236505286588, 'eval/demon_right_rate_teacher': 0.18363939899833054, 'eval/demon_move_f1': 0.8659935667721725, 'eval/demon_cc_f1': 0.9407336845642695, 'eval/latency_8/demon_fire_precision': 0.9699846860643185, 'eval/latency_8/demon_fire_recall': 0.9838459148803976, 'eval/latency_8/demon_fire_f1': 0.9768661320172732, 'eval/latency_8/demon_fire_rate_model': 0.9084585420144685, 'eval/latency_8/demon_fire_rate_teacher': 0.8956594323873122, 'eval/latency_8/demon_left_precision': 0.9154589371980676, 'eval/latency_8/demon_left_recall': 0.8115631691648822, 'eval/latency_8/demon_left_f1': 0.8603859250851306, 'eval/latency_8/demon_left_rate_model': 0.11519198664440734, 'eval/latency_8/demon_left_rate_teacher': 0.12993878686700056, 'eval/latency_8/demon_right_precision': 0.8689759036144579, 'eval/latency_8/demon_right_recall': 0.8742424242424243, 'eval/latency_8/demon_right_f1': 0.8716012084592145, 'eval/latency_8/demon_right_rate_model': 0.18475236505286588, 'eval/latency_8/demon_right_rate_teacher': 0.18363939899833054, 'eval/latency_8/demon_move_f1': 0.8659935667721725, 'eval/latency_8/demon_cc_f1': 0.9407336845642695, 'timing/data': 0.0004954393953084946, 'timing/model': 0.9448268199339509, 'learning_rate/qwen_vl_interface': 5.196799979752265e-07, 'learning_rate/action_model': 5.196799979752264e-06, 'epoch': 3.08}) +[2026-06-18 03:08:42,774][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_6800_state +[2026-06-18 03:08:42,778][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 03:08:48,942][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_6800_state/model.safetensors +[2026-06-18 03:09:03,447][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_6800_state/optimizer.bin +[2026-06-18 03:09:03,448][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_6800_state/sampler.bin +[2026-06-18 03:09:03,448][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_6800_state/sampler_1.bin +[2026-06-18 03:09:03,450][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_6800_state/random_states_0.pkl +[2026-06-18 03:09:03,453][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 6800 +[2026-06-18 03:09:03,453][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 03:09:03,463][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 03:09:07,223][starVLA.training.train_starvla][INFO] - Step 6801, Loss: {'train/loss': 0.5289273858070374, 'train/grad_norm_pre_clip': 42.01029586791992, 'timing/data': 0.0004958556964993477, 'timing/model': 0.9232795983552933, 'learning_rate/qwen_vl_interface': 5.194838243107742e-07, 'learning_rate/action_model': 5.194838243107742e-06, 'epoch': 3.09}) +[2026-06-18 03:09:10,975][starVLA.training.train_starvla][INFO] - Step 6802, Loss: {'train/loss': 0.31405118107795715, 'train/grad_norm_pre_clip': 23.948129653930664, 'timing/data': 0.00048516783863306046, 'timing/model': 0.9244782328605652, 'learning_rate/qwen_vl_interface': 5.192886312874148e-07, 'learning_rate/action_model': 5.1928863128741475e-06, 'epoch': 3.09}) +[2026-06-18 03:09:14,726][starVLA.training.train_starvla][INFO] - Step 6803, Loss: {'train/loss': 0.296129047870636, 'train/grad_norm_pre_clip': 29.423133850097656, 'timing/data': 0.00045825447887182236, 'timing/model': 0.9238576106727123, 'learning_rate/qwen_vl_interface': 5.190944189456133e-07, 'learning_rate/action_model': 5.190944189456132e-06, 'epoch': 3.09}) +[2026-06-18 03:09:18,464][starVLA.training.train_starvla][INFO] - Step 6804, Loss: {'train/loss': 0.578253448009491, 'train/grad_norm_pre_clip': 48.72529602050781, 'timing/data': 0.0004654228687286377, 'timing/model': 1.1014316668733954, 'learning_rate/qwen_vl_interface': 5.189011873256295e-07, 'learning_rate/action_model': 5.189011873256295e-06, 'epoch': 3.09}) +[2026-06-18 03:09:22,022][starVLA.training.train_starvla][INFO] - Step 6805, Loss: {'train/loss': 0.025883832946419716, 'train/grad_norm_pre_clip': 14.74626636505127, 'timing/data': 0.0004489654675126076, 'timing/model': 0.926034870557487, 'learning_rate/qwen_vl_interface': 5.187089364675195e-07, 'learning_rate/action_model': 5.187089364675194e-06, 'epoch': 3.09}) +[2026-06-18 03:09:25,793][starVLA.training.train_starvla][INFO] - Step 6806, Loss: {'train/loss': 0.11632324010133743, 'train/grad_norm_pre_clip': 24.867069244384766, 'timing/data': 0.0004910575225949287, 'timing/model': 0.9260434219613671, 'learning_rate/qwen_vl_interface': 5.18517666411139e-07, 'learning_rate/action_model': 5.18517666411139e-06, 'epoch': 3.09}) +[2026-06-18 03:09:29,581][starVLA.training.train_starvla][INFO] - Step 6807, Loss: {'train/loss': 0.0855393260717392, 'train/grad_norm_pre_clip': 33.52787399291992, 'timing/data': 0.0007963972166180611, 'timing/model': 0.9454317912459373, 'learning_rate/qwen_vl_interface': 5.18327377196137e-07, 'learning_rate/action_model': 5.18327377196137e-06, 'epoch': 3.09}) +[2026-06-18 03:09:33,337][starVLA.training.train_starvla][INFO] - Step 6808, Loss: {'train/loss': 0.07904676347970963, 'train/grad_norm_pre_clip': 20.26276206970215, 'timing/data': 0.001516931690275669, 'timing/model': 0.9202415505424142, 'learning_rate/qwen_vl_interface': 5.181380688619617e-07, 'learning_rate/action_model': 5.181380688619617e-06, 'epoch': 3.09}) +[2026-06-18 03:09:37,060][starVLA.training.train_starvla][INFO] - Step 6809, Loss: {'train/loss': 0.41333162784576416, 'train/grad_norm_pre_clip': 37.11181640625, 'timing/data': 0.0011296626180410385, 'timing/model': 0.9189735893160105, 'learning_rate/qwen_vl_interface': 5.179497414478561e-07, 'learning_rate/action_model': 5.179497414478562e-06, 'epoch': 3.09}) +[2026-06-18 03:09:40,781][starVLA.training.train_starvla][INFO] - Step 6810, Loss: {'train/loss': 0.5350748300552368, 'train/grad_norm_pre_clip': 39.57170486450195, 'timing/data': 0.0004380829632282257, 'timing/model': 0.9229874834418297, 'learning_rate/qwen_vl_interface': 5.177623949928608e-07, 'learning_rate/action_model': 5.177623949928608e-06, 'epoch': 3.09}) +[2026-06-18 03:09:44,515][starVLA.training.train_starvla][INFO] - Step 6811, Loss: {'train/loss': 0.49979597330093384, 'train/grad_norm_pre_clip': 53.169410705566406, 'timing/data': 0.00044291187077760696, 'timing/model': 1.1051449105143547, 'learning_rate/qwen_vl_interface': 5.175760295358137e-07, 'learning_rate/action_model': 5.175760295358136e-06, 'epoch': 3.09}) +[2026-06-18 03:09:48,073][starVLA.training.train_starvla][INFO] - Step 6812, Loss: {'train/loss': 0.3267945945262909, 'train/grad_norm_pre_clip': 33.441078186035156, 'timing/data': 0.0004959981888532639, 'timing/model': 0.9180913837626576, 'learning_rate/qwen_vl_interface': 5.173906451153477e-07, 'learning_rate/action_model': 5.173906451153477e-06, 'epoch': 3.09}) +[2026-06-18 03:09:51,805][starVLA.training.train_starvla][INFO] - Step 6813, Loss: {'train/loss': 0.19775596261024475, 'train/grad_norm_pre_clip': 23.920652389526367, 'timing/data': 0.00047255121171474457, 'timing/model': 0.9184712721034884, 'learning_rate/qwen_vl_interface': 5.172062417698931e-07, 'learning_rate/action_model': 5.1720624176989304e-06, 'epoch': 3.09}) +[2026-06-18 03:09:55,492][starVLA.training.train_starvla][INFO] - Step 6814, Loss: {'train/loss': 0.2585980296134949, 'train/grad_norm_pre_clip': 31.441694259643555, 'timing/data': 0.0005082376301288605, 'timing/model': 0.9183424450457096, 'learning_rate/qwen_vl_interface': 5.170228195376778e-07, 'learning_rate/action_model': 5.170228195376778e-06, 'epoch': 3.09}) +[2026-06-18 03:09:59,195][starVLA.training.train_starvla][INFO] - Step 6815, Loss: {'train/loss': 0.1774231493473053, 'train/grad_norm_pre_clip': 38.06871795654297, 'timing/data': 0.0008271234109997749, 'timing/model': 0.9240466626361012, 'learning_rate/qwen_vl_interface': 5.168403784567247e-07, 'learning_rate/action_model': 5.1684037845672475e-06, 'epoch': 3.09}) +[2026-06-18 03:10:03,094][starVLA.training.train_starvla][INFO] - Step 6816, Loss: {'train/loss': 0.17510005831718445, 'train/grad_norm_pre_clip': 16.9039306640625, 'timing/data': 0.0005429796874523163, 'timing/model': 0.9217312764376402, 'learning_rate/qwen_vl_interface': 5.166589185648543e-07, 'learning_rate/action_model': 5.1665891856485425e-06, 'epoch': 3.09}) +[2026-06-18 03:10:06,800][starVLA.training.train_starvla][INFO] - Step 6817, Loss: {'train/loss': 0.2567204236984253, 'train/grad_norm_pre_clip': 32.45410919189453, 'timing/data': 0.0004998911172151566, 'timing/model': 0.9459367711097002, 'learning_rate/qwen_vl_interface': 5.164784398996836e-07, 'learning_rate/action_model': 5.164784398996836e-06, 'epoch': 3.09}) +[2026-06-18 03:10:10,672][starVLA.training.train_starvla][INFO] - Step 6818, Loss: {'train/loss': 0.05781674385070801, 'train/grad_norm_pre_clip': 16.978357315063477, 'timing/data': 0.00047339126467704773, 'timing/model': 1.0844730269163847, 'learning_rate/qwen_vl_interface': 5.162989424986253e-07, 'learning_rate/action_model': 5.162989424986253e-06, 'epoch': 3.09}) +[2026-06-18 03:10:14,307][starVLA.training.train_starvla][INFO] - Step 6819, Loss: {'train/loss': 0.08551832288503647, 'train/grad_norm_pre_clip': 21.315542221069336, 'timing/data': 0.0005088476464152336, 'timing/model': 0.9238197281956673, 'learning_rate/qwen_vl_interface': 5.161204263988904e-07, 'learning_rate/action_model': 5.161204263988904e-06, 'epoch': 3.09}) +[2026-06-18 03:10:18,037][starVLA.training.train_starvla][INFO] - Step 6820, Loss: {'train/loss': 0.06141038239002228, 'train/grad_norm_pre_clip': 11.588666915893555, 'timing/data': 0.0004904251545667648, 'timing/model': 0.9175936914980412, 'learning_rate/qwen_vl_interface': 5.159428916374845e-07, 'learning_rate/action_model': 5.159428916374845e-06, 'epoch': 3.09}) +[2026-06-18 03:10:22,046][starVLA.training.train_starvla][INFO] - Step 6821, Loss: {'train/loss': 0.05566229671239853, 'train/grad_norm_pre_clip': 15.835651397705078, 'timing/data': 0.0005201809108257294, 'timing/model': 0.9178475206717849, 'learning_rate/qwen_vl_interface': 5.157663382512118e-07, 'learning_rate/action_model': 5.157663382512119e-06, 'epoch': 3.09}) +[2026-06-18 03:10:25,740][starVLA.training.train_starvla][INFO] - Step 6822, Loss: {'train/loss': 0.029627889394760132, 'train/grad_norm_pre_clip': 6.110497951507568, 'timing/data': 0.0004911664873361588, 'timing/model': 0.9212287561967969, 'learning_rate/qwen_vl_interface': 5.155907662766711e-07, 'learning_rate/action_model': 5.155907662766711e-06, 'epoch': 3.09}) +[2026-06-18 03:10:29,612][starVLA.training.train_starvla][INFO] - Step 6823, Loss: {'train/loss': 0.0910501480102539, 'train/grad_norm_pre_clip': 27.375972747802734, 'timing/data': 0.000739268958568573, 'timing/model': 0.9192772647365928, 'learning_rate/qwen_vl_interface': 5.15416175750259e-07, 'learning_rate/action_model': 5.154161757502589e-06, 'epoch': 3.1}) +[2026-06-18 03:10:33,761][starVLA.training.train_starvla][INFO] - Step 6824, Loss: {'train/loss': 0.24171346426010132, 'train/grad_norm_pre_clip': 47.88925552368164, 'timing/data': 0.0004817293956875801, 'timing/model': 1.1123251812532544, 'learning_rate/qwen_vl_interface': 5.15242566708168e-07, 'learning_rate/action_model': 5.15242566708168e-06, 'epoch': 3.1}) +[2026-06-18 03:10:37,507][starVLA.training.train_starvla][INFO] - Step 6825, Loss: {'train/loss': 0.3889376223087311, 'train/grad_norm_pre_clip': 32.33852005004883, 'timing/data': 0.0005009640008211136, 'timing/model': 1.0500028086826205, 'learning_rate/qwen_vl_interface': 5.150699391863884e-07, 'learning_rate/action_model': 5.150699391863884e-06, 'epoch': 3.1}) +[2026-06-18 03:10:41,058][starVLA.training.train_starvla][INFO] - Step 6826, Loss: {'train/loss': 0.2658030092716217, 'train/grad_norm_pre_clip': 39.984840393066406, 'timing/data': 0.0004922449588775635, 'timing/model': 0.9137252727523446, 'learning_rate/qwen_vl_interface': 5.148982932207052e-07, 'learning_rate/action_model': 5.148982932207052e-06, 'epoch': 3.1}) +[2026-06-18 03:10:44,775][starVLA.training.train_starvla][INFO] - Step 6827, Loss: {'train/loss': 0.23319074511528015, 'train/grad_norm_pre_clip': 40.84898376464844, 'timing/data': 0.0004907753318548203, 'timing/model': 0.9208748051896691, 'learning_rate/qwen_vl_interface': 5.147276288467008e-07, 'learning_rate/action_model': 5.147276288467008e-06, 'epoch': 3.1}) +[2026-06-18 03:10:48,493][starVLA.training.train_starvla][INFO] - Step 6828, Loss: {'train/loss': 0.16034412384033203, 'train/grad_norm_pre_clip': 24.542129516601562, 'timing/data': 0.00047491397708654404, 'timing/model': 0.9173723189160228, 'learning_rate/qwen_vl_interface': 5.145579460997544e-07, 'learning_rate/action_model': 5.145579460997544e-06, 'epoch': 3.1}) +[2026-06-18 03:10:52,124][starVLA.training.train_starvla][INFO] - Step 6829, Loss: {'train/loss': 0.4288886785507202, 'train/grad_norm_pre_clip': 49.03382110595703, 'timing/data': 0.0004742033779621124, 'timing/model': 0.9207671601325274, 'learning_rate/qwen_vl_interface': 5.143892450150411e-07, 'learning_rate/action_model': 5.143892450150411e-06, 'epoch': 3.1}) +[2026-06-18 03:10:55,835][starVLA.training.train_starvla][INFO] - Step 6830, Loss: {'train/loss': 0.2081361562013626, 'train/grad_norm_pre_clip': 25.556550979614258, 'timing/data': 0.0005441335961222649, 'timing/model': 0.9235990140587091, 'learning_rate/qwen_vl_interface': 5.142215256275338e-07, 'learning_rate/action_model': 5.142215256275338e-06, 'epoch': 3.1}) +[2026-06-18 03:10:59,788][starVLA.training.train_starvla][INFO] - Step 6831, Loss: {'train/loss': 0.23572978377342224, 'train/grad_norm_pre_clip': 36.37590026855469, 'timing/data': 0.00047861039638519287, 'timing/model': 1.09021607786417, 'learning_rate/qwen_vl_interface': 5.140547879719993e-07, 'learning_rate/action_model': 5.140547879719993e-06, 'epoch': 3.1}) +[2026-06-18 03:11:03,350][starVLA.training.train_starvla][INFO] - Step 6832, Loss: {'train/loss': 0.24830447137355804, 'train/grad_norm_pre_clip': 29.422136306762695, 'timing/data': 0.0008266950026154518, 'timing/model': 0.9157120659947395, 'learning_rate/qwen_vl_interface': 5.138890320830038e-07, 'learning_rate/action_model': 5.138890320830038e-06, 'epoch': 3.1}) +[2026-06-18 03:11:07,061][starVLA.training.train_starvla][INFO] - Step 6833, Loss: {'train/loss': 0.01778494007885456, 'train/grad_norm_pre_clip': 3.5722830295562744, 'timing/data': 0.000541437417268753, 'timing/model': 0.9196294499561191, 'learning_rate/qwen_vl_interface': 5.137242579949081e-07, 'learning_rate/action_model': 5.137242579949081e-06, 'epoch': 3.1}) +[2026-06-18 03:11:10,855][starVLA.training.train_starvla][INFO] - Step 6834, Loss: {'train/loss': 0.05904914438724518, 'train/grad_norm_pre_clip': 10.599403381347656, 'timing/data': 0.0004862891510128975, 'timing/model': 0.9842478632926941, 'learning_rate/qwen_vl_interface': 5.135604657418702e-07, 'learning_rate/action_model': 5.135604657418702e-06, 'epoch': 3.1}) +[2026-06-18 03:11:14,566][starVLA.training.train_starvla][INFO] - Step 6835, Loss: {'train/loss': 0.37636491656303406, 'train/grad_norm_pre_clip': 26.677440643310547, 'timing/data': 0.0005280738696455956, 'timing/model': 0.925873039290309, 'learning_rate/qwen_vl_interface': 5.133976553578443e-07, 'learning_rate/action_model': 5.1339765535784425e-06, 'epoch': 3.1}) +[2026-06-18 03:11:18,297][starVLA.training.train_starvla][INFO] - Step 6836, Loss: {'train/loss': 0.2474907487630844, 'train/grad_norm_pre_clip': 26.685691833496094, 'timing/data': 0.0008815815672278404, 'timing/model': 0.9195566149428487, 'learning_rate/qwen_vl_interface': 5.132358268765813e-07, 'learning_rate/action_model': 5.132358268765813e-06, 'epoch': 3.1}) +[2026-06-18 03:11:22,060][starVLA.training.train_starvla][INFO] - Step 6837, Loss: {'train/loss': 0.2468542605638504, 'train/grad_norm_pre_clip': 38.12578582763672, 'timing/data': 0.0006917314603924751, 'timing/model': 1.1075233779847622, 'learning_rate/qwen_vl_interface': 5.130749803316284e-07, 'learning_rate/action_model': 5.130749803316284e-06, 'epoch': 3.1}) +[2026-06-18 03:11:25,820][starVLA.training.train_starvla][INFO] - Step 6838, Loss: {'train/loss': 0.019665423780679703, 'train/grad_norm_pre_clip': 4.577331066131592, 'timing/data': 0.00045629125088453293, 'timing/model': 1.0963470209389925, 'learning_rate/qwen_vl_interface': 5.129151157563294e-07, 'learning_rate/action_model': 5.129151157563293e-06, 'epoch': 3.1}) +[2026-06-18 03:11:29,794][starVLA.training.train_starvla][INFO] - Step 6839, Loss: {'train/loss': 0.2599710524082184, 'train/grad_norm_pre_clip': 42.59001922607422, 'timing/data': 0.0004711747169494629, 'timing/model': 1.3539726370945573, 'learning_rate/qwen_vl_interface': 5.127562331838241e-07, 'learning_rate/action_model': 5.12756233183824e-06, 'epoch': 3.1}) +[2026-06-18 03:11:33,564][starVLA.training.train_starvla][INFO] - Step 6840, Loss: {'train/loss': 0.34057188034057617, 'train/grad_norm_pre_clip': 55.604087829589844, 'timing/data': 0.0004909122362732887, 'timing/model': 0.914057987742126, 'learning_rate/qwen_vl_interface': 5.125983326470492e-07, 'learning_rate/action_model': 5.125983326470492e-06, 'epoch': 3.1}) +[2026-06-18 03:11:37,582][starVLA.training.train_starvla][INFO] - Step 6841, Loss: {'train/loss': 0.05643417686223984, 'train/grad_norm_pre_clip': 9.585057258605957, 'timing/data': 0.001121465116739273, 'timing/model': 0.9172995826229453, 'learning_rate/qwen_vl_interface': 5.124414141787372e-07, 'learning_rate/action_model': 5.124414141787372e-06, 'epoch': 3.1}) +[2026-06-18 03:11:41,361][starVLA.training.train_starvla][INFO] - Step 6842, Loss: {'train/loss': 0.2221611589193344, 'train/grad_norm_pre_clip': 27.798370361328125, 'timing/data': 0.0004987642168998718, 'timing/model': 0.9281424069777131, 'learning_rate/qwen_vl_interface': 5.122854778114185e-07, 'learning_rate/action_model': 5.122854778114185e-06, 'epoch': 3.1}) +[2026-06-18 03:11:45,217][starVLA.training.train_starvla][INFO] - Step 6843, Loss: {'train/loss': 0.6070556640625, 'train/grad_norm_pre_clip': 37.391117095947266, 'timing/data': 0.0004871934652328491, 'timing/model': 0.9186337683349848, 'learning_rate/qwen_vl_interface': 5.12130523577418e-07, 'learning_rate/action_model': 5.12130523577418e-06, 'epoch': 3.1}) +[2026-06-18 03:11:49,055][starVLA.training.train_starvla][INFO] - Step 6844, Loss: {'train/loss': 0.19213183224201202, 'train/grad_norm_pre_clip': 18.611440658569336, 'timing/data': 0.0005632974207401276, 'timing/model': 1.1156270047649741, 'learning_rate/qwen_vl_interface': 5.119765515088579e-07, 'learning_rate/action_model': 5.1197655150885785e-06, 'epoch': 3.1}) +[2026-06-18 03:11:52,991][starVLA.training.train_starvla][INFO] - Step 6845, Loss: {'train/loss': 0.0954928919672966, 'train/grad_norm_pre_clip': 15.941476821899414, 'timing/data': 0.0009519131854176521, 'timing/model': 1.1864979211241007, 'learning_rate/qwen_vl_interface': 5.11823561637657e-07, 'learning_rate/action_model': 5.1182356163765705e-06, 'epoch': 3.11}) +[2026-06-18 03:11:56,613][starVLA.training.train_starvla][INFO] - Step 6846, Loss: {'train/loss': 0.25293049216270447, 'train/grad_norm_pre_clip': 20.65227699279785, 'timing/data': 0.0005072448402643204, 'timing/model': 0.929347924888134, 'learning_rate/qwen_vl_interface': 5.116715539955306e-07, 'learning_rate/action_model': 5.116715539955305e-06, 'epoch': 3.11}) +[2026-06-18 03:12:00,597][starVLA.training.train_starvla][INFO] - Step 6847, Loss: {'train/loss': 0.035815782845020294, 'train/grad_norm_pre_clip': 7.066670894622803, 'timing/data': 0.0004927497357130051, 'timing/model': 0.9191752281039953, 'learning_rate/qwen_vl_interface': 5.115205286139895e-07, 'learning_rate/action_model': 5.115205286139895e-06, 'epoch': 3.11}) +[2026-06-18 03:12:04,368][starVLA.training.train_starvla][INFO] - Step 6848, Loss: {'train/loss': 0.1700620949268341, 'train/grad_norm_pre_clip': 21.75425910949707, 'timing/data': 0.0005003660917282104, 'timing/model': 0.9607766848057508, 'learning_rate/qwen_vl_interface': 5.11370485524342e-07, 'learning_rate/action_model': 5.1137048552434205e-06, 'epoch': 3.11}) +[2026-06-18 03:12:08,095][starVLA.training.train_starvla][INFO] - Step 6849, Loss: {'train/loss': 0.2096087634563446, 'train/grad_norm_pre_clip': 12.604022979736328, 'timing/data': 0.0006083743646740913, 'timing/model': 0.9213224733248353, 'learning_rate/qwen_vl_interface': 5.112214247576915e-07, 'learning_rate/action_model': 5.1122142475769155e-06, 'epoch': 3.11}) +[2026-06-18 03:12:11,875][starVLA.training.train_starvla][INFO] - Step 6850, Loss: {'train/loss': 0.3272976875305176, 'train/grad_norm_pre_clip': 34.72770690917969, 'timing/data': 0.00048456620424985886, 'timing/model': 0.9243852980434895, 'learning_rate/qwen_vl_interface': 5.110733463449384e-07, 'learning_rate/action_model': 5.110733463449385e-06, 'epoch': 3.11}) +[2026-06-18 03:12:15,588][starVLA.training.train_starvla][INFO] - Step 6851, Loss: {'train/loss': 0.11894573271274567, 'train/grad_norm_pre_clip': 20.529794692993164, 'timing/data': 0.0006488114595413208, 'timing/model': 1.0795572437345982, 'learning_rate/qwen_vl_interface': 5.109262503167808e-07, 'learning_rate/action_model': 5.109262503167808e-06, 'epoch': 3.11}) +[2026-06-18 03:12:19,155][starVLA.training.train_starvla][INFO] - Step 6852, Loss: {'train/loss': 0.1515779048204422, 'train/grad_norm_pre_clip': 34.515995025634766, 'timing/data': 0.0009734118357300758, 'timing/model': 0.913031586445868, 'learning_rate/qwen_vl_interface': 5.107801367037106e-07, 'learning_rate/action_model': 5.107801367037107e-06, 'epoch': 3.11}) +[2026-06-18 03:12:22,876][starVLA.training.train_starvla][INFO] - Step 6853, Loss: {'train/loss': 0.22141559422016144, 'train/grad_norm_pre_clip': 31.54824447631836, 'timing/data': 0.0005056178197264671, 'timing/model': 0.9207793334499002, 'learning_rate/qwen_vl_interface': 5.106350055360173e-07, 'learning_rate/action_model': 5.106350055360173e-06, 'epoch': 3.11}) +[2026-06-18 03:12:26,571][starVLA.training.train_starvla][INFO] - Step 6854, Loss: {'train/loss': 0.20919814705848694, 'train/grad_norm_pre_clip': 30.015140533447266, 'timing/data': 0.0005440516397356987, 'timing/model': 0.9142315620556474, 'learning_rate/qwen_vl_interface': 5.104908568437875e-07, 'learning_rate/action_model': 5.104908568437875e-06, 'epoch': 3.11}) +[2026-06-18 03:12:30,292][starVLA.training.train_starvla][INFO] - Step 6855, Loss: {'train/loss': 0.1414373815059662, 'train/grad_norm_pre_clip': 33.0472526550293, 'timing/data': 0.0005495809018611908, 'timing/model': 0.9195992350578308, 'learning_rate/qwen_vl_interface': 5.103476906569034e-07, 'learning_rate/action_model': 5.103476906569033e-06, 'epoch': 3.11}) +[2026-06-18 03:12:33,984][starVLA.training.train_starvla][INFO] - Step 6856, Loss: {'train/loss': 0.08301975578069687, 'train/grad_norm_pre_clip': 18.512285232543945, 'timing/data': 0.00047285109758377075, 'timing/model': 0.9155054753646255, 'learning_rate/qwen_vl_interface': 5.102055070050428e-07, 'learning_rate/action_model': 5.102055070050427e-06, 'epoch': 3.11}) +[2026-06-18 03:12:37,784][starVLA.training.train_starvla][INFO] - Step 6857, Loss: {'train/loss': 0.298592746257782, 'train/grad_norm_pre_clip': 21.621871948242188, 'timing/data': 0.0004995819181203842, 'timing/model': 0.9753965688869357, 'learning_rate/qwen_vl_interface': 5.100643059176803e-07, 'learning_rate/action_model': 5.100643059176802e-06, 'epoch': 3.11}) +[2026-06-18 03:12:41,546][starVLA.training.train_starvla][INFO] - Step 6858, Loss: {'train/loss': 0.05613434687256813, 'train/grad_norm_pre_clip': 11.385675430297852, 'timing/data': 0.00045842770487070084, 'timing/model': 1.128802009858191, 'learning_rate/qwen_vl_interface': 5.099240874240883e-07, 'learning_rate/action_model': 5.099240874240883e-06, 'epoch': 3.11}) +[2026-06-18 03:12:45,090][starVLA.training.train_starvla][INFO] - Step 6859, Loss: {'train/loss': 0.24249036610126495, 'train/grad_norm_pre_clip': 22.866989135742188, 'timing/data': 0.0004869820550084114, 'timing/model': 0.9189909109845757, 'learning_rate/qwen_vl_interface': 5.097848515533331e-07, 'learning_rate/action_model': 5.0978485155333305e-06, 'epoch': 3.11}) +[2026-06-18 03:12:49,102][starVLA.training.train_starvla][INFO] - Step 6860, Loss: {'train/loss': 0.26629501581192017, 'train/grad_norm_pre_clip': 26.46308135986328, 'timing/data': 0.00043657608330249786, 'timing/model': 1.1897672535851598, 'learning_rate/qwen_vl_interface': 5.096465983342791e-07, 'learning_rate/action_model': 5.096465983342791e-06, 'epoch': 3.11}) +[2026-06-18 03:12:53,547][starVLA.training.train_starvla][INFO] - Step 6861, Loss: {'train/loss': 0.13953185081481934, 'train/grad_norm_pre_clip': 27.096147537231445, 'timing/data': 0.00048822537064552307, 'timing/model': 0.936611833050847, 'learning_rate/qwen_vl_interface': 5.095093277955859e-07, 'learning_rate/action_model': 5.095093277955859e-06, 'epoch': 3.11}) +[2026-06-18 03:12:57,563][starVLA.training.train_starvla][INFO] - Step 6862, Loss: {'train/loss': 0.4065243601799011, 'train/grad_norm_pre_clip': 40.937904357910156, 'timing/data': 0.0005245581269264221, 'timing/model': 0.9421474486589432, 'learning_rate/qwen_vl_interface': 5.093730399657101e-07, 'learning_rate/action_model': 5.0937303996571e-06, 'epoch': 3.11}) +[2026-06-18 03:13:01,376][starVLA.training.train_starvla][INFO] - Step 6863, Loss: {'train/loss': 0.16131415963172913, 'train/grad_norm_pre_clip': 17.668331146240234, 'timing/data': 0.0005251970142126083, 'timing/model': 0.9413054417818785, 'learning_rate/qwen_vl_interface': 5.092377348729043e-07, 'learning_rate/action_model': 5.092377348729043e-06, 'epoch': 3.11}) +[2026-06-18 03:13:05,213][starVLA.training.train_starvla][INFO] - Step 6864, Loss: {'train/loss': 0.127811998128891, 'train/grad_norm_pre_clip': 33.01756286621094, 'timing/data': 0.0005818363279104233, 'timing/model': 1.116429090499878, 'learning_rate/qwen_vl_interface': 5.091034125452171e-07, 'learning_rate/action_model': 5.091034125452171e-06, 'epoch': 3.11}) +[2026-06-18 03:13:08,970][starVLA.training.train_starvla][INFO] - Step 6865, Loss: {'train/loss': 0.17380951344966888, 'train/grad_norm_pre_clip': 34.7301025390625, 'timing/data': 0.0004834812134504318, 'timing/model': 1.0882911020889878, 'learning_rate/qwen_vl_interface': 5.089700730104944e-07, 'learning_rate/action_model': 5.089700730104944e-06, 'epoch': 3.11}) +[2026-06-18 03:13:12,577][starVLA.training.train_starvla][INFO] - Step 6866, Loss: {'train/loss': 0.3410432040691376, 'train/grad_norm_pre_clip': 42.76957321166992, 'timing/data': 0.0018181977793574333, 'timing/model': 0.9213888887315989, 'learning_rate/qwen_vl_interface': 5.08837716296377e-07, 'learning_rate/action_model': 5.0883771629637696e-06, 'epoch': 3.11}) +[2026-06-18 03:13:16,406][starVLA.training.train_starvla][INFO] - Step 6867, Loss: {'train/loss': 0.23063914477825165, 'train/grad_norm_pre_clip': 36.33112716674805, 'timing/data': 0.0007236618548631668, 'timing/model': 0.9654580503702164, 'learning_rate/qwen_vl_interface': 5.087063424303026e-07, 'learning_rate/action_model': 5.087063424303026e-06, 'epoch': 3.12}) +[2026-06-18 03:13:20,152][starVLA.training.train_starvla][INFO] - Step 6868, Loss: {'train/loss': 0.2528946101665497, 'train/grad_norm_pre_clip': 35.164588928222656, 'timing/data': 0.0010139578953385353, 'timing/model': 0.9319941690191627, 'learning_rate/qwen_vl_interface': 5.085759514395052e-07, 'learning_rate/action_model': 5.085759514395051e-06, 'epoch': 3.12}) +[2026-06-18 03:13:24,148][starVLA.training.train_starvla][INFO] - Step 6869, Loss: {'train/loss': 0.13567590713500977, 'train/grad_norm_pre_clip': 23.834430694580078, 'timing/data': 0.0005636336281895638, 'timing/model': 0.9344924613833427, 'learning_rate/qwen_vl_interface': 5.084465433510151e-07, 'learning_rate/action_model': 5.0844654335101515e-06, 'epoch': 3.12}) +[2026-06-18 03:13:27,973][starVLA.training.train_starvla][INFO] - Step 6870, Loss: {'train/loss': 0.11233673989772797, 'train/grad_norm_pre_clip': 24.992616653442383, 'timing/data': 0.000500841997563839, 'timing/model': 0.9192813793197274, 'learning_rate/qwen_vl_interface': 5.083181181916588e-07, 'learning_rate/action_model': 5.083181181916588e-06, 'epoch': 3.12}) +[2026-06-18 03:13:31,749][starVLA.training.train_starvla][INFO] - Step 6871, Loss: {'train/loss': 0.28590214252471924, 'train/grad_norm_pre_clip': 29.458723068237305, 'timing/data': 0.000614408403635025, 'timing/model': 1.1432974198833108, 'learning_rate/qwen_vl_interface': 5.081906759880585e-07, 'learning_rate/action_model': 5.081906759880585e-06, 'epoch': 3.12}) +[2026-06-18 03:13:35,446][starVLA.training.train_starvla][INFO] - Step 6872, Loss: {'train/loss': 0.03899785876274109, 'train/grad_norm_pre_clip': 16.75880241394043, 'timing/data': 0.0006575863808393478, 'timing/model': 1.0431759022176266, 'learning_rate/qwen_vl_interface': 5.080642167666338e-07, 'learning_rate/action_model': 5.080642167666337e-06, 'epoch': 3.12}) +[2026-06-18 03:13:39,020][starVLA.training.train_starvla][INFO] - Step 6873, Loss: {'train/loss': 0.2164941281080246, 'train/grad_norm_pre_clip': 31.909934997558594, 'timing/data': 0.001200912520289421, 'timing/model': 0.9206482144072652, 'learning_rate/qwen_vl_interface': 5.079387405535995e-07, 'learning_rate/action_model': 5.079387405535995e-06, 'epoch': 3.12}) +[2026-06-18 03:13:42,749][starVLA.training.train_starvla][INFO] - Step 6874, Loss: {'train/loss': 0.19505634903907776, 'train/grad_norm_pre_clip': 43.89966583251953, 'timing/data': 0.0005263332277536392, 'timing/model': 0.9155601439997554, 'learning_rate/qwen_vl_interface': 5.07814247374967e-07, 'learning_rate/action_model': 5.078142473749671e-06, 'epoch': 3.12}) +[2026-06-18 03:13:46,490][starVLA.training.train_starvla][INFO] - Step 6875, Loss: {'train/loss': 0.25871723890304565, 'train/grad_norm_pre_clip': 40.519447326660156, 'timing/data': 0.0005316678434610367, 'timing/model': 0.9208854306489229, 'learning_rate/qwen_vl_interface': 5.07690737256544e-07, 'learning_rate/action_model': 5.07690737256544e-06, 'epoch': 3.12}) +[2026-06-18 03:13:50,179][starVLA.training.train_starvla][INFO] - Step 6876, Loss: {'train/loss': 0.02991589531302452, 'train/grad_norm_pre_clip': 8.30702018737793, 'timing/data': 0.0007541058585047722, 'timing/model': 0.9302906095981598, 'learning_rate/qwen_vl_interface': 5.075682102239336e-07, 'learning_rate/action_model': 5.075682102239335e-06, 'epoch': 3.12}) +[2026-06-18 03:13:53,969][starVLA.training.train_starvla][INFO] - Step 6877, Loss: {'train/loss': 0.19036683440208435, 'train/grad_norm_pre_clip': 18.710350036621094, 'timing/data': 0.0005296468734741211, 'timing/model': 0.9201070349663496, 'learning_rate/qwen_vl_interface': 5.074466663025369e-07, 'learning_rate/action_model': 5.074466663025369e-06, 'epoch': 3.12}) +[2026-06-18 03:13:57,682][starVLA.training.train_starvla][INFO] - Step 6878, Loss: {'train/loss': 0.3651711344718933, 'train/grad_norm_pre_clip': 31.16108512878418, 'timing/data': 0.00048460252583026886, 'timing/model': 1.0961640616878867, 'learning_rate/qwen_vl_interface': 5.07326105517549e-07, 'learning_rate/action_model': 5.07326105517549e-06, 'epoch': 3.12}) +[2026-06-18 03:14:01,242][starVLA.training.train_starvla][INFO] - Step 6879, Loss: {'train/loss': 0.19567735493183136, 'train/grad_norm_pre_clip': 28.666471481323242, 'timing/data': 0.00046549271792173386, 'timing/model': 0.9299463778734207, 'learning_rate/qwen_vl_interface': 5.072065278939628e-07, 'learning_rate/action_model': 5.072065278939627e-06, 'epoch': 3.12}) +[2026-06-18 03:14:05,083][starVLA.training.train_starvla][INFO] - Step 6880, Loss: {'train/loss': 0.22218000888824463, 'train/grad_norm_pre_clip': 25.764598846435547, 'timing/data': 0.006302353926002979, 'timing/model': 0.9841656321659684, 'learning_rate/qwen_vl_interface': 5.070879334565671e-07, 'learning_rate/action_model': 5.070879334565671e-06, 'epoch': 3.12}) +[2026-06-18 03:14:09,106][starVLA.training.train_starvla][INFO] - Step 6881, Loss: {'train/loss': 0.462812602519989, 'train/grad_norm_pre_clip': 42.04585647583008, 'timing/data': 0.0005697049200534821, 'timing/model': 1.1780433487147093, 'learning_rate/qwen_vl_interface': 5.069703222299464e-07, 'learning_rate/action_model': 5.069703222299463e-06, 'epoch': 3.12}) +[2026-06-18 03:14:12,854][starVLA.training.train_starvla][INFO] - Step 6882, Loss: {'train/loss': 0.1521606296300888, 'train/grad_norm_pre_clip': 23.354509353637695, 'timing/data': 0.0005164537578821182, 'timing/model': 0.9294452015310526, 'learning_rate/qwen_vl_interface': 5.068536942384813e-07, 'learning_rate/action_model': 5.068536942384812e-06, 'epoch': 3.12}) +[2026-06-18 03:14:16,606][starVLA.training.train_starvla][INFO] - Step 6883, Loss: {'train/loss': 0.27610260248184204, 'train/grad_norm_pre_clip': 32.76250457763672, 'timing/data': 0.0005498481914401054, 'timing/model': 0.9370454354211688, 'learning_rate/qwen_vl_interface': 5.067380495063494e-07, 'learning_rate/action_model': 5.067380495063494e-06, 'epoch': 3.12}) +[2026-06-18 03:14:20,329][starVLA.training.train_starvla][INFO] - Step 6884, Loss: {'train/loss': 0.08449951559305191, 'train/grad_norm_pre_clip': 19.49568748474121, 'timing/data': 0.000492660328745842, 'timing/model': 1.0880594942718744, 'learning_rate/qwen_vl_interface': 5.066233880575239e-07, 'learning_rate/action_model': 5.0662338805752395e-06, 'epoch': 3.12}) +[2026-06-18 03:14:24,051][starVLA.training.train_starvla][INFO] - Step 6885, Loss: {'train/loss': 0.2826712727546692, 'train/grad_norm_pre_clip': 40.16659164428711, 'timing/data': 0.0006590811535716057, 'timing/model': 1.0819857502356172, 'learning_rate/qwen_vl_interface': 5.065097099157743e-07, 'learning_rate/action_model': 5.065097099157743e-06, 'epoch': 3.12}) +[2026-06-18 03:14:27,846][starVLA.training.train_starvla][INFO] - Step 6886, Loss: {'train/loss': 0.04299816116690636, 'train/grad_norm_pre_clip': 11.566768646240234, 'timing/data': 0.0004831179976463318, 'timing/model': 0.9148296359926462, 'learning_rate/qwen_vl_interface': 5.063970151046654e-07, 'learning_rate/action_model': 5.0639701510466545e-06, 'epoch': 3.12}) +[2026-06-18 03:14:31,563][starVLA.training.train_starvla][INFO] - Step 6887, Loss: {'train/loss': 0.14595210552215576, 'train/grad_norm_pre_clip': 37.38372802734375, 'timing/data': 0.0004510246217250824, 'timing/model': 0.9207497555762529, 'learning_rate/qwen_vl_interface': 5.062853036475604e-07, 'learning_rate/action_model': 5.062853036475603e-06, 'epoch': 3.12}) +[2026-06-18 03:14:35,322][starVLA.training.train_starvla][INFO] - Step 6888, Loss: {'train/loss': 0.296741247177124, 'train/grad_norm_pre_clip': 27.050724029541016, 'timing/data': 0.000504596158862114, 'timing/model': 0.9171656165271997, 'learning_rate/qwen_vl_interface': 5.061745755676161e-07, 'learning_rate/action_model': 5.061745755676161e-06, 'epoch': 3.12}) +[2026-06-18 03:14:39,029][starVLA.training.train_starvla][INFO] - Step 6889, Loss: {'train/loss': 0.6993182301521301, 'train/grad_norm_pre_clip': 30.816457748413086, 'timing/data': 0.0005193967372179031, 'timing/model': 0.9243303136900067, 'learning_rate/qwen_vl_interface': 5.060648308877866e-07, 'learning_rate/action_model': 5.0606483088778656e-06, 'epoch': 3.13}) +[2026-06-18 03:14:42,800][starVLA.training.train_starvla][INFO] - Step 6890, Loss: {'train/loss': 0.5188360810279846, 'train/grad_norm_pre_clip': 42.13398742675781, 'timing/data': 0.0004894677549600601, 'timing/model': 0.9305527489632368, 'learning_rate/qwen_vl_interface': 5.059560696308229e-07, 'learning_rate/action_model': 5.059560696308228e-06, 'epoch': 3.13}) +[2026-06-18 03:14:48,472][starVLA.training.train_starvla][INFO] - Step 6891, Loss: {'train/loss': 0.22982943058013916, 'train/grad_norm_pre_clip': 19.657655715942383, 'timing/data': 0.00112238060683012, 'timing/model': 1.577165343798697, 'learning_rate/qwen_vl_interface': 5.0584829181927e-07, 'learning_rate/action_model': 5.058482918192699e-06, 'epoch': 3.13}) +[2026-06-18 03:14:52,901][starVLA.training.train_starvla][INFO] - Step 6892, Loss: {'train/loss': 0.09942381829023361, 'train/grad_norm_pre_clip': 27.636926651000977, 'timing/data': 0.0004741717129945755, 'timing/model': 1.5465108267962933, 'learning_rate/qwen_vl_interface': 5.05741497475472e-07, 'learning_rate/action_model': 5.05741497475472e-06, 'epoch': 3.13}) +[2026-06-18 03:14:56,492][starVLA.training.train_starvla][INFO] - Step 6893, Loss: {'train/loss': 0.25311779975891113, 'train/grad_norm_pre_clip': 32.253318786621094, 'timing/data': 0.001185886561870575, 'timing/model': 0.9473766982555389, 'learning_rate/qwen_vl_interface': 5.056356866215663e-07, 'learning_rate/action_model': 5.056356866215662e-06, 'epoch': 3.13}) +[2026-06-18 03:15:00,311][starVLA.training.train_starvla][INFO] - Step 6894, Loss: {'train/loss': 0.1285141259431839, 'train/grad_norm_pre_clip': 21.93162727355957, 'timing/data': 0.0006814468652009964, 'timing/model': 0.9105155980214477, 'learning_rate/qwen_vl_interface': 5.055308592794876e-07, 'learning_rate/action_model': 5.0553085927948756e-06, 'epoch': 3.13}) +[2026-06-18 03:15:04,036][starVLA.training.train_starvla][INFO] - Step 6895, Loss: {'train/loss': 0.5072516202926636, 'train/grad_norm_pre_clip': 50.087074279785156, 'timing/data': 0.0008101174607872963, 'timing/model': 0.9252933664247394, 'learning_rate/qwen_vl_interface': 5.054270154709674e-07, 'learning_rate/action_model': 5.054270154709673e-06, 'epoch': 3.13}) +[2026-06-18 03:15:07,810][starVLA.training.train_starvla][INFO] - Step 6896, Loss: {'train/loss': 0.20186451077461243, 'train/grad_norm_pre_clip': 29.310508728027344, 'timing/data': 0.0005020108073949814, 'timing/model': 0.9218186363577843, 'learning_rate/qwen_vl_interface': 5.053241552175322e-07, 'learning_rate/action_model': 5.053241552175321e-06, 'epoch': 3.13}) +[2026-06-18 03:15:11,874][starVLA.training.train_starvla][INFO] - Step 6897, Loss: {'train/loss': 0.5728015899658203, 'train/grad_norm_pre_clip': 71.4516372680664, 'timing/data': 0.0006441418081521988, 'timing/model': 1.151362169533968, 'learning_rate/qwen_vl_interface': 5.05222278540505e-07, 'learning_rate/action_model': 5.05222278540505e-06, 'epoch': 3.13}) +[2026-06-18 03:15:16,060][starVLA.training.train_starvla][INFO] - Step 6898, Loss: {'train/loss': 0.16098998486995697, 'train/grad_norm_pre_clip': 27.05887794494629, 'timing/data': 0.0005081044510006905, 'timing/model': 1.1750653572380543, 'learning_rate/qwen_vl_interface': 5.051213854610051e-07, 'learning_rate/action_model': 5.0512138546100514e-06, 'epoch': 3.13}) +[2026-06-18 03:15:19,949][starVLA.training.train_starvla][INFO] - Step 6899, Loss: {'train/loss': 0.1530812382698059, 'train/grad_norm_pre_clip': 27.281002044677734, 'timing/data': 0.0007818304002285004, 'timing/model': 0.9338782923296094, 'learning_rate/qwen_vl_interface': 5.050214759999478e-07, 'learning_rate/action_model': 5.050214759999478e-06, 'epoch': 3.13}) +[2026-06-18 03:17:18,183][starVLA.training.train_starvla][INFO] - Step 6900, Loss: {'train/loss': 0.40942466259002686, 'train/grad_norm_pre_clip': 57.04109573364258, 'eval/loss': 0.2521917223930359, 'eval/latency_8/loss': 0.2521917223930359, 'eval/demon_fire_precision': 0.9740340030911901, 'eval/demon_fire_recall': 0.9788754271512892, 'eval/demon_fire_f1': 0.9764487139758289, 'eval/demon_fire_rate_model': 0.9001112966054535, 'eval/demon_fire_rate_teacher': 0.8956594323873122, 'eval/demon_left_precision': 0.9152542372881356, 'eval/demon_left_recall': 0.8094218415417559, 'eval/demon_left_f1': 0.859090909090909, 'eval/demon_left_rate_model': 0.11491374513077351, 'eval/demon_left_rate_teacher': 0.12993878686700056, 'eval/demon_right_precision': 0.8798751950078003, 'eval/demon_right_recall': 0.8545454545454545, 'eval/demon_right_f1': 0.8670253651037663, 'eval/demon_right_rate_model': 0.1783528102392877, 'eval/demon_right_rate_teacher': 0.18363939899833054, 'eval/demon_move_f1': 0.8630581370973376, 'eval/demon_cc_f1': 0.9394216945999005, 'eval/latency_8/demon_fire_precision': 0.9740340030911901, 'eval/latency_8/demon_fire_recall': 0.9788754271512892, 'eval/latency_8/demon_fire_f1': 0.9764487139758289, 'eval/latency_8/demon_fire_rate_model': 0.9001112966054535, 'eval/latency_8/demon_fire_rate_teacher': 0.8956594323873122, 'eval/latency_8/demon_left_precision': 0.9152542372881356, 'eval/latency_8/demon_left_recall': 0.8094218415417559, 'eval/latency_8/demon_left_f1': 0.859090909090909, 'eval/latency_8/demon_left_rate_model': 0.11491374513077351, 'eval/latency_8/demon_left_rate_teacher': 0.12993878686700056, 'eval/latency_8/demon_right_precision': 0.8798751950078003, 'eval/latency_8/demon_right_recall': 0.8545454545454545, 'eval/latency_8/demon_right_f1': 0.8670253651037663, 'eval/latency_8/demon_right_rate_model': 0.1783528102392877, 'eval/latency_8/demon_right_rate_teacher': 0.18363939899833054, 'eval/latency_8/demon_move_f1': 0.8630581370973376, 'eval/latency_8/demon_cc_f1': 0.9394216945999005, 'timing/data': 0.00046744104474782944, 'timing/model': 0.9198498530313373, 'learning_rate/qwen_vl_interface': 5.049225501780441e-07, 'learning_rate/action_model': 5.049225501780441e-06, 'epoch': 3.13}) +[2026-06-18 03:17:18,204][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_6900_state +[2026-06-18 03:17:18,224][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 03:17:24,503][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_6900_state/model.safetensors +[2026-06-18 03:17:38,476][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_6900_state/optimizer.bin +[2026-06-18 03:17:38,477][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_6900_state/sampler.bin +[2026-06-18 03:17:38,478][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_6900_state/sampler_1.bin +[2026-06-18 03:17:38,479][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_6900_state/random_states_0.pkl +[2026-06-18 03:17:38,483][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 6900 +[2026-06-18 03:17:38,483][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 03:17:38,494][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 03:17:42,378][starVLA.training.train_starvla][INFO] - Step 6901, Loss: {'train/loss': 0.0533517561852932, 'train/grad_norm_pre_clip': 10.257792472839355, 'timing/data': 0.0005764877423644066, 'timing/model': 1.2077408395707607, 'learning_rate/qwen_vl_interface': 5.048246080158016e-07, 'learning_rate/action_model': 5.048246080158016e-06, 'epoch': 3.13}) +[2026-06-18 03:17:45,962][starVLA.training.train_starvla][INFO] - Step 6902, Loss: {'train/loss': 0.5178934335708618, 'train/grad_norm_pre_clip': 61.661651611328125, 'timing/data': 0.000927346758544445, 'timing/model': 0.930400681681931, 'learning_rate/qwen_vl_interface': 5.047276495335243e-07, 'learning_rate/action_model': 5.047276495335242e-06, 'epoch': 3.13}) +[2026-06-18 03:17:49,760][starVLA.training.train_starvla][INFO] - Step 6903, Loss: {'train/loss': 0.1987503170967102, 'train/grad_norm_pre_clip': 33.61823272705078, 'timing/data': 0.0006722193211317062, 'timing/model': 0.9152452852576971, 'learning_rate/qwen_vl_interface': 5.046316747513115e-07, 'learning_rate/action_model': 5.046316747513115e-06, 'epoch': 3.13}) +[2026-06-18 03:17:53,869][starVLA.training.train_starvla][INFO] - Step 6904, Loss: {'train/loss': 0.42377129197120667, 'train/grad_norm_pre_clip': 36.71344757080078, 'timing/data': 0.000503217801451683, 'timing/model': 1.2106386963278055, 'learning_rate/qwen_vl_interface': 5.045366836890585e-07, 'learning_rate/action_model': 5.045366836890584e-06, 'epoch': 3.13}) +[2026-06-18 03:17:57,761][starVLA.training.train_starvla][INFO] - Step 6905, Loss: {'train/loss': 0.1839780956506729, 'train/grad_norm_pre_clip': 34.07586669921875, 'timing/data': 0.0005193594843149185, 'timing/model': 0.9236488519236445, 'learning_rate/qwen_vl_interface': 5.044426763664572e-07, 'learning_rate/action_model': 5.044426763664571e-06, 'epoch': 3.13}) +[2026-06-18 03:18:01,558][starVLA.training.train_starvla][INFO] - Step 6906, Loss: {'train/loss': 0.09008122235536575, 'train/grad_norm_pre_clip': 23.325593948364258, 'timing/data': 0.0005238242447376251, 'timing/model': 0.9248264068737626, 'learning_rate/qwen_vl_interface': 5.043496528029954e-07, 'learning_rate/action_model': 5.043496528029954e-06, 'epoch': 3.13}) +[2026-06-18 03:18:05,318][starVLA.training.train_starvla][INFO] - Step 6907, Loss: {'train/loss': 0.16021180152893066, 'train/grad_norm_pre_clip': 47.9417610168457, 'timing/data': 0.00044485460966825485, 'timing/model': 0.9384494451805949, 'learning_rate/qwen_vl_interface': 5.042576130179573e-07, 'learning_rate/action_model': 5.042576130179572e-06, 'epoch': 3.13}) +[2026-06-18 03:18:09,051][starVLA.training.train_starvla][INFO] - Step 6908, Loss: {'train/loss': 0.08808165043592453, 'train/grad_norm_pre_clip': 17.840412139892578, 'timing/data': 0.0008752793073654175, 'timing/model': 1.0961855631321669, 'learning_rate/qwen_vl_interface': 5.041665570304227e-07, 'learning_rate/action_model': 5.041665570304226e-06, 'epoch': 3.13}) +[2026-06-18 03:18:12,620][starVLA.training.train_starvla][INFO] - Step 6909, Loss: {'train/loss': 0.3644880950450897, 'train/grad_norm_pre_clip': 26.922780990600586, 'timing/data': 0.0007900875061750412, 'timing/model': 0.9236763063818216, 'learning_rate/qwen_vl_interface': 5.040764848592675e-07, 'learning_rate/action_model': 5.040764848592675e-06, 'epoch': 3.13}) +[2026-06-18 03:18:17,177][starVLA.training.train_starvla][INFO] - Step 6910, Loss: {'train/loss': 0.04423241317272186, 'train/grad_norm_pre_clip': 11.838598251342773, 'timing/data': 0.0014286842197179794, 'timing/model': 0.9213322401046753, 'learning_rate/qwen_vl_interface': 5.039873965231644e-07, 'learning_rate/action_model': 5.039873965231643e-06, 'epoch': 3.13}) +[2026-06-18 03:18:21,157][starVLA.training.train_starvla][INFO] - Step 6911, Loss: {'train/loss': 0.5316553711891174, 'train/grad_norm_pre_clip': 43.17437744140625, 'timing/data': 0.0008728094398975372, 'timing/model': 0.9239934822544456, 'learning_rate/qwen_vl_interface': 5.0389929204058e-07, 'learning_rate/action_model': 5.0389929204058e-06, 'epoch': 3.14}) +[2026-06-18 03:18:25,381][starVLA.training.train_starvla][INFO] - Step 6912, Loss: {'train/loss': 0.17411130666732788, 'train/grad_norm_pre_clip': 18.604066848754883, 'timing/data': 0.0005079833790659904, 'timing/model': 0.9172314358875155, 'learning_rate/qwen_vl_interface': 5.038121714297802e-07, 'learning_rate/action_model': 5.038121714297802e-06, 'epoch': 3.14}) +[2026-06-18 03:18:29,126][starVLA.training.train_starvla][INFO] - Step 6913, Loss: {'train/loss': 0.17367792129516602, 'train/grad_norm_pre_clip': 47.8760986328125, 'timing/data': 0.0007382389158010483, 'timing/model': 0.9389398489147425, 'learning_rate/qwen_vl_interface': 5.037260347088236e-07, 'learning_rate/action_model': 5.037260347088236e-06, 'epoch': 3.14}) +[2026-06-18 03:18:32,905][starVLA.training.train_starvla][INFO] - Step 6914, Loss: {'train/loss': 0.10846307128667831, 'train/grad_norm_pre_clip': 32.15657043457031, 'timing/data': 0.0005542626604437828, 'timing/model': 0.9671052368357778, 'learning_rate/qwen_vl_interface': 5.03640881895568e-07, 'learning_rate/action_model': 5.03640881895568e-06, 'epoch': 3.14}) +[2026-06-18 03:18:36,614][starVLA.training.train_starvla][INFO] - Step 6915, Loss: {'train/loss': 0.14035801589488983, 'train/grad_norm_pre_clip': 26.80175018310547, 'timing/data': 0.0005287695676088333, 'timing/model': 1.0668566189706326, 'learning_rate/qwen_vl_interface': 5.035567130076642e-07, 'learning_rate/action_model': 5.035567130076642e-06, 'epoch': 3.14}) +[2026-06-18 03:18:40,245][starVLA.training.train_starvla][INFO] - Step 6916, Loss: {'train/loss': 0.1064508855342865, 'train/grad_norm_pre_clip': 26.745250701904297, 'timing/data': 0.0006165327504277229, 'timing/model': 0.9460202939808369, 'learning_rate/qwen_vl_interface': 5.034735280625617e-07, 'learning_rate/action_model': 5.034735280625616e-06, 'epoch': 3.14}) +[2026-06-18 03:18:43,996][starVLA.training.train_starvla][INFO] - Step 6917, Loss: {'train/loss': 0.18515750765800476, 'train/grad_norm_pre_clip': 23.029329299926758, 'timing/data': 0.0006887409836053848, 'timing/model': 0.9206148171797395, 'learning_rate/qwen_vl_interface': 5.033913270775039e-07, 'learning_rate/action_model': 5.033913270775039e-06, 'epoch': 3.14}) +[2026-06-18 03:18:47,778][starVLA.training.train_starvla][INFO] - Step 6918, Loss: {'train/loss': 0.332990825176239, 'train/grad_norm_pre_clip': 72.7976303100586, 'timing/data': 0.00045572593808174133, 'timing/model': 0.9547098344191909, 'learning_rate/qwen_vl_interface': 5.03310110069532e-07, 'learning_rate/action_model': 5.03310110069532e-06, 'epoch': 3.14}) +[2026-06-18 03:18:51,556][starVLA.training.train_starvla][INFO] - Step 6919, Loss: {'train/loss': 0.1931016892194748, 'train/grad_norm_pre_clip': 31.178939819335938, 'timing/data': 0.0007325615733861923, 'timing/model': 0.93609176017344, 'learning_rate/qwen_vl_interface': 5.032298770554818e-07, 'learning_rate/action_model': 5.032298770554818e-06, 'epoch': 3.14}) +[2026-06-18 03:18:55,565][starVLA.training.train_starvla][INFO] - Step 6920, Loss: {'train/loss': 0.2843768000602722, 'train/grad_norm_pre_clip': 32.69841003417969, 'timing/data': 0.0005050869658589363, 'timing/model': 1.146834995597601, 'learning_rate/qwen_vl_interface': 5.031506280519855e-07, 'learning_rate/action_model': 5.031506280519855e-06, 'epoch': 3.14}) +[2026-06-18 03:18:59,391][starVLA.training.train_starvla][INFO] - Step 6921, Loss: {'train/loss': 0.1368110328912735, 'train/grad_norm_pre_clip': 30.205041885375977, 'timing/data': 0.0005218842998147011, 'timing/model': 1.1734385685995221, 'learning_rate/qwen_vl_interface': 5.030723630754723e-07, 'learning_rate/action_model': 5.030723630754723e-06, 'epoch': 3.14}) +[2026-06-18 03:19:03,396][starVLA.training.train_starvla][INFO] - Step 6922, Loss: {'train/loss': 0.21583400666713715, 'train/grad_norm_pre_clip': 55.84014892578125, 'timing/data': 0.0007844781503081322, 'timing/model': 1.0575914653018117, 'learning_rate/qwen_vl_interface': 5.02995082142166e-07, 'learning_rate/action_model': 5.029950821421659e-06, 'epoch': 3.14}) +[2026-06-18 03:19:07,014][starVLA.training.train_starvla][INFO] - Step 6923, Loss: {'train/loss': 0.2851325571537018, 'train/grad_norm_pre_clip': 24.62308692932129, 'timing/data': 0.0004922272637486458, 'timing/model': 0.937891723588109, 'learning_rate/qwen_vl_interface': 5.029187852680869e-07, 'learning_rate/action_model': 5.029187852680868e-06, 'epoch': 3.14}) +[2026-06-18 03:19:10,824][starVLA.training.train_starvla][INFO] - Step 6924, Loss: {'train/loss': 0.3255644142627716, 'train/grad_norm_pre_clip': 40.539363861083984, 'timing/data': 0.0005498593673110008, 'timing/model': 0.9346763137727976, 'learning_rate/qwen_vl_interface': 5.028434724690521e-07, 'learning_rate/action_model': 5.028434724690521e-06, 'epoch': 3.14}) +[2026-06-18 03:19:14,639][starVLA.training.train_starvla][INFO] - Step 6925, Loss: {'train/loss': 0.5456339716911316, 'train/grad_norm_pre_clip': 51.41569900512695, 'timing/data': 0.00048625096678733826, 'timing/model': 0.9593937247991562, 'learning_rate/qwen_vl_interface': 5.027691437606734e-07, 'learning_rate/action_model': 5.027691437606733e-06, 'epoch': 3.14}) +[2026-06-18 03:19:18,422][starVLA.training.train_starvla][INFO] - Step 6926, Loss: {'train/loss': 0.0415072962641716, 'train/grad_norm_pre_clip': 11.983762741088867, 'timing/data': 0.0005075996741652489, 'timing/model': 1.0012222472578287, 'learning_rate/qwen_vl_interface': 5.026957991583597e-07, 'learning_rate/action_model': 5.0269579915835965e-06, 'epoch': 3.14}) +[2026-06-18 03:19:22,191][starVLA.training.train_starvla][INFO] - Step 6927, Loss: {'train/loss': 0.17554862797260284, 'train/grad_norm_pre_clip': 20.023923873901367, 'timing/data': 0.000547739677131176, 'timing/model': 0.929873613640666, 'learning_rate/qwen_vl_interface': 5.026234386773145e-07, 'learning_rate/action_model': 5.026234386773145e-06, 'epoch': 3.14}) +[2026-06-18 03:19:26,228][starVLA.training.train_starvla][INFO] - Step 6928, Loss: {'train/loss': 0.21488262712955475, 'train/grad_norm_pre_clip': 24.686351776123047, 'timing/data': 0.0011281250044703484, 'timing/model': 1.3614201257005334, 'learning_rate/qwen_vl_interface': 5.025520623325398e-07, 'learning_rate/action_model': 5.025520623325397e-06, 'epoch': 3.14}) +[2026-06-18 03:19:30,029][starVLA.training.train_starvla][INFO] - Step 6929, Loss: {'train/loss': 0.12343379110097885, 'train/grad_norm_pre_clip': 16.697933197021484, 'timing/data': 0.0004951711744070053, 'timing/model': 0.9170250287279487, 'learning_rate/qwen_vl_interface': 5.024816701388302e-07, 'learning_rate/action_model': 5.024816701388302e-06, 'epoch': 3.14}) +[2026-06-18 03:19:33,758][starVLA.training.train_starvla][INFO] - Step 6930, Loss: {'train/loss': 0.024914974346756935, 'train/grad_norm_pre_clip': 7.434966087341309, 'timing/data': 0.0004558516666293144, 'timing/model': 0.9200448328629136, 'learning_rate/qwen_vl_interface': 5.024122621107795e-07, 'learning_rate/action_model': 5.024122621107795e-06, 'epoch': 3.14}) +[2026-06-18 03:19:37,425][starVLA.training.train_starvla][INFO] - Step 6931, Loss: {'train/loss': 0.10686583817005157, 'train/grad_norm_pre_clip': 12.349413871765137, 'timing/data': 0.0006235456094145775, 'timing/model': 0.9191707652062178, 'learning_rate/qwen_vl_interface': 5.023438382627749e-07, 'learning_rate/action_model': 5.0234383826277486e-06, 'epoch': 3.14}) +[2026-06-18 03:19:41,173][starVLA.training.train_starvla][INFO] - Step 6932, Loss: {'train/loss': 0.35804203152656555, 'train/grad_norm_pre_clip': 34.707664489746094, 'timing/data': 0.0005091801285743713, 'timing/model': 0.9416973851621151, 'learning_rate/qwen_vl_interface': 5.022763986090018e-07, 'learning_rate/action_model': 5.022763986090018e-06, 'epoch': 3.14}) +[2026-06-18 03:19:45,420][starVLA.training.train_starvla][INFO] - Step 6933, Loss: {'train/loss': 0.115373894572258, 'train/grad_norm_pre_clip': 17.448348999023438, 'timing/data': 0.0009481115266680717, 'timing/model': 0.958079781383276, 'learning_rate/qwen_vl_interface': 5.022099431634398e-07, 'learning_rate/action_model': 5.022099431634398e-06, 'epoch': 3.15}) +[2026-06-18 03:19:49,421][starVLA.training.train_starvla][INFO] - Step 6934, Loss: {'train/loss': 0.2623976767063141, 'train/grad_norm_pre_clip': 18.61060905456543, 'timing/data': 0.0015809210017323494, 'timing/model': 1.1939057996496558, 'learning_rate/qwen_vl_interface': 5.021444719398656e-07, 'learning_rate/action_model': 5.021444719398656e-06, 'epoch': 3.15}) +[2026-06-18 03:19:53,276][starVLA.training.train_starvla][INFO] - Step 6935, Loss: {'train/loss': 0.10291141271591187, 'train/grad_norm_pre_clip': 21.929710388183594, 'timing/data': 0.0005365302786231041, 'timing/model': 1.11639847420156, 'learning_rate/qwen_vl_interface': 5.020799849518513e-07, 'learning_rate/action_model': 5.020799849518513e-06, 'epoch': 3.15}) +[2026-06-18 03:19:56,900][starVLA.training.train_starvla][INFO] - Step 6936, Loss: {'train/loss': 0.11436636745929718, 'train/grad_norm_pre_clip': 9.798590660095215, 'timing/data': 0.000493406318128109, 'timing/model': 0.929703782312572, 'learning_rate/qwen_vl_interface': 5.020164822127652e-07, 'learning_rate/action_model': 5.020164822127652e-06, 'epoch': 3.15}) +[2026-06-18 03:20:00,669][starVLA.training.train_starvla][INFO] - Step 6937, Loss: {'train/loss': 0.09440509974956512, 'train/grad_norm_pre_clip': 20.7490291595459, 'timing/data': 0.0004480471834540367, 'timing/model': 0.9344330057501793, 'learning_rate/qwen_vl_interface': 5.019539637357715e-07, 'learning_rate/action_model': 5.019539637357714e-06, 'epoch': 3.15}) +[2026-06-18 03:20:04,473][starVLA.training.train_starvla][INFO] - Step 6938, Loss: {'train/loss': 0.2656362056732178, 'train/grad_norm_pre_clip': 49.21751403808594, 'timing/data': 0.0004987828433513641, 'timing/model': 0.9345899419859052, 'learning_rate/qwen_vl_interface': 5.0189242953383e-07, 'learning_rate/action_model': 5.018924295338299e-06, 'epoch': 3.15}) +[2026-06-18 03:20:08,658][starVLA.training.train_starvla][INFO] - Step 6939, Loss: {'train/loss': 0.10490650683641434, 'train/grad_norm_pre_clip': 26.254180908203125, 'timing/data': 0.0005164574831724167, 'timing/model': 1.2309937113896012, 'learning_rate/qwen_vl_interface': 5.018318796196967e-07, 'learning_rate/action_model': 5.018318796196967e-06, 'epoch': 3.15}) +[2026-06-18 03:20:12,594][starVLA.training.train_starvla][INFO] - Step 6940, Loss: {'train/loss': 0.14698685705661774, 'train/grad_norm_pre_clip': 16.544403076171875, 'timing/data': 0.0004509389400482178, 'timing/model': 0.9450952252373099, 'learning_rate/qwen_vl_interface': 5.017723140059243e-07, 'learning_rate/action_model': 5.017723140059242e-06, 'epoch': 3.15}) +[2026-06-18 03:20:21,046][starVLA.training.train_starvla][INFO] - Step 6941, Loss: {'train/loss': 0.13046729564666748, 'train/grad_norm_pre_clip': 16.001516342163086, 'timing/data': 0.00044274888932704926, 'timing/model': 1.0921311574056745, 'learning_rate/qwen_vl_interface': 5.017137327048606e-07, 'learning_rate/action_model': 5.017137327048606e-06, 'epoch': 3.15}) +[2026-06-18 03:20:25,414][starVLA.training.train_starvla][INFO] - Step 6942, Loss: {'train/loss': 0.15549218654632568, 'train/grad_norm_pre_clip': 23.54275131225586, 'timing/data': 0.0008266828954219818, 'timing/model': 1.1749688982963562, 'learning_rate/qwen_vl_interface': 5.016561357286498e-07, 'learning_rate/action_model': 5.016561357286498e-06, 'epoch': 3.15}) +[2026-06-18 03:20:29,105][starVLA.training.train_starvla][INFO] - Step 6943, Loss: {'train/loss': 0.29077425599098206, 'train/grad_norm_pre_clip': 27.159259796142578, 'timing/data': 0.0008698431774973869, 'timing/model': 0.9566906364634633, 'learning_rate/qwen_vl_interface': 5.015995230892312e-07, 'learning_rate/action_model': 5.015995230892311e-06, 'epoch': 3.15}) +[2026-06-18 03:20:32,909][starVLA.training.train_starvla][INFO] - Step 6944, Loss: {'train/loss': 0.28652042150497437, 'train/grad_norm_pre_clip': 33.5451774597168, 'timing/data': 0.0004565846174955368, 'timing/model': 0.9383944030851126, 'learning_rate/qwen_vl_interface': 5.015438947983409e-07, 'learning_rate/action_model': 5.015438947983409e-06, 'epoch': 3.15}) +[2026-06-18 03:20:37,653][starVLA.training.train_starvla][INFO] - Step 6945, Loss: {'train/loss': 0.1523905098438263, 'train/grad_norm_pre_clip': 24.146970748901367, 'timing/data': 0.0004921620711684227, 'timing/model': 1.0448091374710202, 'learning_rate/qwen_vl_interface': 5.014892508675108e-07, 'learning_rate/action_model': 5.014892508675107e-06, 'epoch': 3.15}) +[2026-06-18 03:20:41,422][starVLA.training.train_starvla][INFO] - Step 6946, Loss: {'train/loss': 0.02497084252536297, 'train/grad_norm_pre_clip': 5.416917324066162, 'timing/data': 0.0009166505187749863, 'timing/model': 0.9262200128287077, 'learning_rate/qwen_vl_interface': 5.014355913080683e-07, 'learning_rate/action_model': 5.014355913080683e-06, 'epoch': 3.15}) +[2026-06-18 03:20:45,224][starVLA.training.train_starvla][INFO] - Step 6947, Loss: {'train/loss': 0.04849883168935776, 'train/grad_norm_pre_clip': 15.70799732208252, 'timing/data': 0.0008457405492663383, 'timing/model': 0.9330783681944013, 'learning_rate/qwen_vl_interface': 5.013829161311377e-07, 'learning_rate/action_model': 5.013829161311377e-06, 'epoch': 3.15}) +[2026-06-18 03:20:48,996][starVLA.training.train_starvla][INFO] - Step 6948, Loss: {'train/loss': 0.3919597566127777, 'train/grad_norm_pre_clip': 33.24500274658203, 'timing/data': 0.0009376909583806992, 'timing/model': 1.1002326384186745, 'learning_rate/qwen_vl_interface': 5.013312253476383e-07, 'learning_rate/action_model': 5.0133122534763826e-06, 'epoch': 3.15}) +[2026-06-18 03:20:52,633][starVLA.training.train_starvla][INFO] - Step 6949, Loss: {'train/loss': 0.11553733795881271, 'train/grad_norm_pre_clip': 23.56972885131836, 'timing/data': 0.0007879380136728287, 'timing/model': 0.9565137615427375, 'learning_rate/qwen_vl_interface': 5.012805189682861e-07, 'learning_rate/action_model': 5.01280518968286e-06, 'epoch': 3.15}) +[2026-06-18 03:20:56,694][starVLA.training.train_starvla][INFO] - Step 6950, Loss: {'train/loss': 0.04854367673397064, 'train/grad_norm_pre_clip': 11.400684356689453, 'timing/data': 0.0004926975816488266, 'timing/model': 1.14812854770571, 'learning_rate/qwen_vl_interface': 5.012307970035915e-07, 'learning_rate/action_model': 5.012307970035915e-06, 'epoch': 3.15}) +[2026-06-18 03:21:00,629][starVLA.training.train_starvla][INFO] - Step 6951, Loss: {'train/loss': 0.2877773344516754, 'train/grad_norm_pre_clip': 36.41520309448242, 'timing/data': 0.0008247671648859978, 'timing/model': 0.9396728835999966, 'learning_rate/qwen_vl_interface': 5.011820594638626e-07, 'learning_rate/action_model': 5.0118205946386256e-06, 'epoch': 3.15}) +[2026-06-18 03:21:04,425][starVLA.training.train_starvla][INFO] - Step 6952, Loss: {'train/loss': 0.06586717814207077, 'train/grad_norm_pre_clip': 15.730060577392578, 'timing/data': 0.00047207437455654144, 'timing/model': 0.9404887016862631, 'learning_rate/qwen_vl_interface': 5.011343063592032e-07, 'learning_rate/action_model': 5.011343063592032e-06, 'epoch': 3.15}) +[2026-06-18 03:21:08,229][starVLA.training.train_starvla][INFO] - Step 6953, Loss: {'train/loss': 0.0884726494550705, 'train/grad_norm_pre_clip': 22.212060928344727, 'timing/data': 0.0007971888408064842, 'timing/model': 0.9373264070600271, 'learning_rate/qwen_vl_interface': 5.010875376995121e-07, 'learning_rate/action_model': 5.01087537699512e-06, 'epoch': 3.15}) +[2026-06-18 03:21:12,097][starVLA.training.train_starvla][INFO] - Step 6954, Loss: {'train/loss': 0.1284276843070984, 'train/grad_norm_pre_clip': 19.40803337097168, 'timing/data': 0.0009008888155221939, 'timing/model': 0.9586819047108293, 'learning_rate/qwen_vl_interface': 5.010417534944839e-07, 'learning_rate/action_model': 5.010417534944838e-06, 'epoch': 3.15}) +[2026-06-18 03:21:16,045][starVLA.training.train_starvla][INFO] - Step 6955, Loss: {'train/loss': 0.19084718823432922, 'train/grad_norm_pre_clip': 28.733177185058594, 'timing/data': 0.0005005691200494766, 'timing/model': 1.2660292824730277, 'learning_rate/qwen_vl_interface': 5.009969537536106e-07, 'learning_rate/action_model': 5.009969537536106e-06, 'epoch': 3.16}) +[2026-06-18 03:21:19,824][starVLA.training.train_starvla][INFO] - Step 6956, Loss: {'train/loss': 0.1132415235042572, 'train/grad_norm_pre_clip': 29.45988655090332, 'timing/data': 0.0006992164999246597, 'timing/model': 0.9634432308375835, 'learning_rate/qwen_vl_interface': 5.00953138486179e-07, 'learning_rate/action_model': 5.00953138486179e-06, 'epoch': 3.16}) +[2026-06-18 03:21:23,818][starVLA.training.train_starvla][INFO] - Step 6957, Loss: {'train/loss': 0.12283031642436981, 'train/grad_norm_pre_clip': 19.445192337036133, 'timing/data': 0.0009338846430182457, 'timing/model': 1.0762426014989614, 'learning_rate/qwen_vl_interface': 5.009103077012723e-07, 'learning_rate/action_model': 5.009103077012723e-06, 'epoch': 3.16}) +[2026-06-18 03:21:31,154][starVLA.training.train_starvla][INFO] - Step 6958, Loss: {'train/loss': 0.2288169413805008, 'train/grad_norm_pre_clip': 24.40667724609375, 'timing/data': 0.0007288726046681404, 'timing/model': 2.2858388628810644, 'learning_rate/qwen_vl_interface': 5.008684614077684e-07, 'learning_rate/action_model': 5.008684614077684e-06, 'epoch': 3.16}) +[2026-06-18 03:21:35,125][starVLA.training.train_starvla][INFO] - Step 6959, Loss: {'train/loss': 0.22258757054805756, 'train/grad_norm_pre_clip': 20.513092041015625, 'timing/data': 0.0006224466487765312, 'timing/model': 0.9279424799606204, 'learning_rate/qwen_vl_interface': 5.008275996143429e-07, 'learning_rate/action_model': 5.0082759961434295e-06, 'epoch': 3.16}) +[2026-06-18 03:21:38,980][starVLA.training.train_starvla][INFO] - Step 6960, Loss: {'train/loss': 0.24801239371299744, 'train/grad_norm_pre_clip': 28.580242156982422, 'timing/data': 0.0005557611584663391, 'timing/model': 0.9625784130766988, 'learning_rate/qwen_vl_interface': 5.007877223294668e-07, 'learning_rate/action_model': 5.0078772232946675e-06, 'epoch': 3.16}) +[2026-06-18 03:21:42,911][starVLA.training.train_starvla][INFO] - Step 6961, Loss: {'train/loss': 0.2195957452058792, 'train/grad_norm_pre_clip': 26.552213668823242, 'timing/data': 0.000708606094121933, 'timing/model': 1.1829267414286733, 'learning_rate/qwen_vl_interface': 5.007488295614058e-07, 'learning_rate/action_model': 5.007488295614058e-06, 'epoch': 3.16}) +[2026-06-18 03:21:46,846][starVLA.training.train_starvla][INFO] - Step 6962, Loss: {'train/loss': 0.32553964853286743, 'train/grad_norm_pre_clip': 33.00594711303711, 'timing/data': 0.0004316139966249466, 'timing/model': 1.107068171724677, 'learning_rate/qwen_vl_interface': 5.007109213182231e-07, 'learning_rate/action_model': 5.007109213182231e-06, 'epoch': 3.16}) +[2026-06-18 03:21:50,474][starVLA.training.train_starvla][INFO] - Step 6963, Loss: {'train/loss': 0.04354345053434372, 'train/grad_norm_pre_clip': 13.422571182250977, 'timing/data': 0.0006807409226894379, 'timing/model': 0.933907886967063, 'learning_rate/qwen_vl_interface': 5.006739976077769e-07, 'learning_rate/action_model': 5.0067399760777695e-06, 'epoch': 3.16}) +[2026-06-18 03:21:54,582][starVLA.training.train_starvla][INFO] - Step 6964, Loss: {'train/loss': 0.025146793574094772, 'train/grad_norm_pre_clip': 4.7987446784973145, 'timing/data': 0.0005037738010287285, 'timing/model': 0.9466628236696124, 'learning_rate/qwen_vl_interface': 5.006380584377213e-07, 'learning_rate/action_model': 5.006380584377213e-06, 'epoch': 3.16}) +[2026-06-18 03:21:58,624][starVLA.training.train_starvla][INFO] - Step 6965, Loss: {'train/loss': 0.14367017149925232, 'train/grad_norm_pre_clip': 16.819316864013672, 'timing/data': 0.0010952316224575043, 'timing/model': 0.9252701597288251, 'learning_rate/qwen_vl_interface': 5.006031038155068e-07, 'learning_rate/action_model': 5.006031038155067e-06, 'epoch': 3.16}) +[2026-06-18 03:22:02,431][starVLA.training.train_starvla][INFO] - Step 6966, Loss: {'train/loss': 0.2143690586090088, 'train/grad_norm_pre_clip': 31.512332916259766, 'timing/data': 0.0006736433133482933, 'timing/model': 0.9495817432180047, 'learning_rate/qwen_vl_interface': 5.00569133748379e-07, 'learning_rate/action_model': 5.00569133748379e-06, 'epoch': 3.16}) +[2026-06-18 03:22:06,615][starVLA.training.train_starvla][INFO] - Step 6967, Loss: {'train/loss': 0.0987858697772026, 'train/grad_norm_pre_clip': 17.375080108642578, 'timing/data': 0.0005030343309044838, 'timing/model': 1.034993619658053, 'learning_rate/qwen_vl_interface': 5.005361482433811e-07, 'learning_rate/action_model': 5.005361482433811e-06, 'epoch': 3.16}) +[2026-06-18 03:22:10,565][starVLA.training.train_starvla][INFO] - Step 6968, Loss: {'train/loss': 0.4930253326892853, 'train/grad_norm_pre_clip': 33.0404167175293, 'timing/data': 0.0005449289456009865, 'timing/model': 1.1716704843565822, 'learning_rate/qwen_vl_interface': 5.005041473073501e-07, 'learning_rate/action_model': 5.005041473073501e-06, 'epoch': 3.16}) +[2026-06-18 03:22:14,629][starVLA.training.train_starvla][INFO] - Step 6969, Loss: {'train/loss': 0.10074129700660706, 'train/grad_norm_pre_clip': 15.647157669067383, 'timing/data': 0.0005819946527481079, 'timing/model': 1.0759764667600393, 'learning_rate/qwen_vl_interface': 5.004731309469201e-07, 'learning_rate/action_model': 5.004731309469201e-06, 'epoch': 3.16}) +[2026-06-18 03:22:18,224][starVLA.training.train_starvla][INFO] - Step 6970, Loss: {'train/loss': 0.13132795691490173, 'train/grad_norm_pre_clip': 24.782533645629883, 'timing/data': 0.000765778124332428, 'timing/model': 0.9308740925043821, 'learning_rate/qwen_vl_interface': 5.004430991685201e-07, 'learning_rate/action_model': 5.0044309916852016e-06, 'epoch': 3.16}) +[2026-06-18 03:22:22,108][starVLA.training.train_starvla][INFO] - Step 6971, Loss: {'train/loss': 0.37024280428886414, 'train/grad_norm_pre_clip': 22.95440673828125, 'timing/data': 0.000500035472214222, 'timing/model': 0.9514760281890631, 'learning_rate/qwen_vl_interface': 5.004140519783773e-07, 'learning_rate/action_model': 5.0041405197837725e-06, 'epoch': 3.16}) +[2026-06-18 03:22:26,005][starVLA.training.train_starvla][INFO] - Step 6972, Loss: {'train/loss': 0.07941273599863052, 'train/grad_norm_pre_clip': 23.623245239257812, 'timing/data': 0.0004759775474667549, 'timing/model': 1.0250103063881397, 'learning_rate/qwen_vl_interface': 5.003859893825117e-07, 'learning_rate/action_model': 5.003859893825117e-06, 'epoch': 3.16}) +[2026-06-18 03:22:29,847][starVLA.training.train_starvla][INFO] - Step 6973, Loss: {'train/loss': 0.05514562129974365, 'train/grad_norm_pre_clip': 14.251428604125977, 'timing/data': 0.0004366496577858925, 'timing/model': 0.9697834867984056, 'learning_rate/qwen_vl_interface': 5.003589113867419e-07, 'learning_rate/action_model': 5.003589113867419e-06, 'epoch': 3.16}) +[2026-06-18 03:22:33,629][starVLA.training.train_starvla][INFO] - Step 6974, Loss: {'train/loss': 0.16010203957557678, 'train/grad_norm_pre_clip': 28.324607849121094, 'timing/data': 0.0007926896214485168, 'timing/model': 0.9244313454255462, 'learning_rate/qwen_vl_interface': 5.003328179966799e-07, 'learning_rate/action_model': 5.003328179966798e-06, 'epoch': 3.16}) +[2026-06-18 03:22:37,425][starVLA.training.train_starvla][INFO] - Step 6975, Loss: {'train/loss': 0.28918829560279846, 'train/grad_norm_pre_clip': 31.12940788269043, 'timing/data': 0.0007233694195747375, 'timing/model': 1.1219922685995698, 'learning_rate/qwen_vl_interface': 5.003077092177363e-07, 'learning_rate/action_model': 5.003077092177363e-06, 'epoch': 3.16}) +[2026-06-18 03:22:41,084][starVLA.training.train_starvla][INFO] - Step 6976, Loss: {'train/loss': 0.5053467750549316, 'train/grad_norm_pre_clip': 38.965763092041016, 'timing/data': 0.0004794243723154068, 'timing/model': 0.9166541239246726, 'learning_rate/qwen_vl_interface': 5.002835850551146e-07, 'learning_rate/action_model': 5.002835850551146e-06, 'epoch': 3.16}) +[2026-06-18 03:22:44,867][starVLA.training.train_starvla][INFO] - Step 6977, Loss: {'train/loss': 0.12061149626970291, 'train/grad_norm_pre_clip': 18.176191329956055, 'timing/data': 0.000490877777338028, 'timing/model': 0.9187729684635997, 'learning_rate/qwen_vl_interface': 5.002604455138174e-07, 'learning_rate/action_model': 5.002604455138173e-06, 'epoch': 3.17}) +[2026-06-18 03:22:48,567][starVLA.training.train_starvla][INFO] - Step 6978, Loss: {'train/loss': 0.07145863026380539, 'train/grad_norm_pre_clip': 13.351162910461426, 'timing/data': 0.0004666326567530632, 'timing/model': 0.9309089332818985, 'learning_rate/qwen_vl_interface': 5.002382905986407e-07, 'learning_rate/action_model': 5.002382905986407e-06, 'epoch': 3.17}) +[2026-06-18 03:22:52,551][starVLA.training.train_starvla][INFO] - Step 6979, Loss: {'train/loss': 0.44191426038742065, 'train/grad_norm_pre_clip': 42.91062545776367, 'timing/data': 0.00046399421989917755, 'timing/model': 0.9392927549779415, 'learning_rate/qwen_vl_interface': 5.002171203141771e-07, 'learning_rate/action_model': 5.0021712031417704e-06, 'epoch': 3.17}) +[2026-06-18 03:22:56,506][starVLA.training.train_starvla][INFO] - Step 6980, Loss: {'train/loss': 0.1760750710964203, 'train/grad_norm_pre_clip': 29.01515769958496, 'timing/data': 0.0005405722185969353, 'timing/model': 0.9452750338241458, 'learning_rate/qwen_vl_interface': 5.001969346648154e-07, 'learning_rate/action_model': 5.001969346648154e-06, 'epoch': 3.17}) +[2026-06-18 03:23:00,343][starVLA.training.train_starvla][INFO] - Step 6981, Loss: {'train/loss': 0.18589621782302856, 'train/grad_norm_pre_clip': 16.813480377197266, 'timing/data': 0.0010738493874669075, 'timing/model': 1.167299522086978, 'learning_rate/qwen_vl_interface': 5.0017773365474e-07, 'learning_rate/action_model': 5.001777336547399e-06, 'epoch': 3.17}) +[2026-06-18 03:23:04,165][starVLA.training.train_starvla][INFO] - Step 6982, Loss: {'train/loss': 0.13315673172473907, 'train/grad_norm_pre_clip': 46.493446350097656, 'timing/data': 0.0005662767216563225, 'timing/model': 1.1192055335268378, 'learning_rate/qwen_vl_interface': 5.001595172879317e-07, 'learning_rate/action_model': 5.001595172879316e-06, 'epoch': 3.17}) +[2026-06-18 03:23:07,837][starVLA.training.train_starvla][INFO] - Step 6983, Loss: {'train/loss': 0.14674289524555206, 'train/grad_norm_pre_clip': 26.055137634277344, 'timing/data': 0.0008876416832208633, 'timing/model': 0.935698002576828, 'learning_rate/qwen_vl_interface': 5.001422855681669e-07, 'learning_rate/action_model': 5.001422855681669e-06, 'epoch': 3.17}) +[2026-06-18 03:23:11,657][starVLA.training.train_starvla][INFO] - Step 6984, Loss: {'train/loss': 0.12144249677658081, 'train/grad_norm_pre_clip': 17.22562599182129, 'timing/data': 0.0007629022002220154, 'timing/model': 0.95697021484375, 'learning_rate/qwen_vl_interface': 5.001260384990169e-07, 'learning_rate/action_model': 5.001260384990169e-06, 'epoch': 3.17}) +[2026-06-18 03:23:15,449][starVLA.training.train_starvla][INFO] - Step 6985, Loss: {'train/loss': 0.12038077414035797, 'train/grad_norm_pre_clip': 17.478288650512695, 'timing/data': 0.0004437696188688278, 'timing/model': 0.925680723041296, 'learning_rate/qwen_vl_interface': 5.001107760838504e-07, 'learning_rate/action_model': 5.001107760838504e-06, 'epoch': 3.17}) +[2026-06-18 03:23:19,233][starVLA.training.train_starvla][INFO] - Step 6986, Loss: {'train/loss': 0.41349202394485474, 'train/grad_norm_pre_clip': 41.05442810058594, 'timing/data': 0.0009849602356553078, 'timing/model': 0.9429582031443715, 'learning_rate/qwen_vl_interface': 5.000964983258306e-07, 'learning_rate/action_model': 5.000964983258306e-06, 'epoch': 3.17}) +[2026-06-18 03:23:23,441][starVLA.training.train_starvla][INFO] - Step 6987, Loss: {'train/loss': 0.09419269859790802, 'train/grad_norm_pre_clip': 37.31244659423828, 'timing/data': 0.0006637563928961754, 'timing/model': 1.0097946031019092, 'learning_rate/qwen_vl_interface': 5.000832052279185e-07, 'learning_rate/action_model': 5.000832052279185e-06, 'epoch': 3.17}) +[2026-06-18 03:23:27,353][starVLA.training.train_starvla][INFO] - Step 6988, Loss: {'train/loss': 0.02628183551132679, 'train/grad_norm_pre_clip': 8.51785659790039, 'timing/data': 0.00048164743930101395, 'timing/model': 1.1883262731134892, 'learning_rate/qwen_vl_interface': 5.000708967928686e-07, 'learning_rate/action_model': 5.000708967928685e-06, 'epoch': 3.17}) +[2026-06-18 03:23:31,216][starVLA.training.train_starvla][INFO] - Step 6989, Loss: {'train/loss': 0.09673623740673065, 'train/grad_norm_pre_clip': 21.649120330810547, 'timing/data': 0.0004522949457168579, 'timing/model': 1.1421038135886192, 'learning_rate/qwen_vl_interface': 5.00059573023233e-07, 'learning_rate/action_model': 5.0005957302323305e-06, 'epoch': 3.17}) +[2026-06-18 03:23:35,000][starVLA.training.train_starvla][INFO] - Step 6990, Loss: {'train/loss': 0.25453391671180725, 'train/grad_norm_pre_clip': 110.3025894165039, 'timing/data': 0.0005428371950984001, 'timing/model': 0.9376164972782135, 'learning_rate/qwen_vl_interface': 5.000492339213594e-07, 'learning_rate/action_model': 5.000492339213594e-06, 'epoch': 3.17}) +[2026-06-18 03:23:39,353][starVLA.training.train_starvla][INFO] - Step 6991, Loss: {'train/loss': 0.3063591718673706, 'train/grad_norm_pre_clip': 39.446495056152344, 'timing/data': 0.0007153330370783806, 'timing/model': 1.2822751058265567, 'learning_rate/qwen_vl_interface': 5.000398794893907e-07, 'learning_rate/action_model': 5.000398794893906e-06, 'epoch': 3.17}) +[2026-06-18 03:23:43,402][starVLA.training.train_starvla][INFO] - Step 6992, Loss: {'train/loss': 0.24158288538455963, 'train/grad_norm_pre_clip': 65.2166748046875, 'timing/data': 0.0004606219008564949, 'timing/model': 0.9746087063103914, 'learning_rate/qwen_vl_interface': 5.000315097292659e-07, 'learning_rate/action_model': 5.000315097292659e-06, 'epoch': 3.17}) +[2026-06-18 03:23:47,317][starVLA.training.train_starvla][INFO] - Step 6993, Loss: {'train/loss': 0.1360921710729599, 'train/grad_norm_pre_clip': 32.770111083984375, 'timing/data': 0.0006416207179427147, 'timing/model': 0.9380793990567327, 'learning_rate/qwen_vl_interface': 5.000241246427207e-07, 'learning_rate/action_model': 5.000241246427207e-06, 'epoch': 3.17}) +[2026-06-18 03:23:51,372][starVLA.training.train_starvla][INFO] - Step 6994, Loss: {'train/loss': 0.1679932028055191, 'train/grad_norm_pre_clip': 17.959497451782227, 'timing/data': 0.0005310205742716789, 'timing/model': 0.9387289378792048, 'learning_rate/qwen_vl_interface': 5.000177242312855e-07, 'learning_rate/action_model': 5.000177242312855e-06, 'epoch': 3.17}) +[2026-06-18 03:23:55,404][starVLA.training.train_starvla][INFO] - Step 6995, Loss: {'train/loss': 0.22022008895874023, 'train/grad_norm_pre_clip': 26.567577362060547, 'timing/data': 0.0005951281636953354, 'timing/model': 1.2631246792152524, 'learning_rate/qwen_vl_interface': 5.00012308496287e-07, 'learning_rate/action_model': 5.000123084962871e-06, 'epoch': 3.17}) +[2026-06-18 03:23:59,108][starVLA.training.train_starvla][INFO] - Step 6996, Loss: {'train/loss': 0.02413790300488472, 'train/grad_norm_pre_clip': 4.824780464172363, 'timing/data': 0.0005128439515829086, 'timing/model': 1.0137470597401261, 'learning_rate/qwen_vl_interface': 5.000078774388487e-07, 'learning_rate/action_model': 5.000078774388487e-06, 'epoch': 3.17}) +[2026-06-18 03:24:02,930][starVLA.training.train_starvla][INFO] - Step 6997, Loss: {'train/loss': 0.15758444368839264, 'train/grad_norm_pre_clip': 26.31064796447754, 'timing/data': 0.0008120415732264519, 'timing/model': 0.9341634558513761, 'learning_rate/qwen_vl_interface': 5.000044310598879e-07, 'learning_rate/action_model': 5.00004431059888e-06, 'epoch': 3.17}) +[2026-06-18 03:24:06,748][starVLA.training.train_starvla][INFO] - Step 6998, Loss: {'train/loss': 0.16847243905067444, 'train/grad_norm_pre_clip': 19.006587982177734, 'timing/data': 0.000537944957613945, 'timing/model': 0.9666877929121256, 'learning_rate/qwen_vl_interface': 5.000019693601204e-07, 'learning_rate/action_model': 5.000019693601204e-06, 'epoch': 3.17}) +[2026-06-18 03:24:10,679][starVLA.training.train_starvla][INFO] - Step 6999, Loss: {'train/loss': 0.11201319843530655, 'train/grad_norm_pre_clip': 27.7429141998291, 'timing/data': 0.0004685400053858757, 'timing/model': 0.9642353719100356, 'learning_rate/qwen_vl_interface': 5.000004923400556e-07, 'learning_rate/action_model': 5.0000049234005565e-06, 'epoch': 3.18}) +[2026-06-18 03:26:26,928][starVLA.training.train_starvla][INFO] - Step 7000, Loss: {'train/loss': 0.2264287769794464, 'train/grad_norm_pre_clip': 47.99951934814453, 'eval/loss': 0.2504505217075348, 'eval/latency_8/loss': 0.2504505217075348, 'eval/demon_fire_precision': 0.9749613601236476, 'eval/demon_fire_recall': 0.9798073936004971, 'eval/demon_fire_f1': 0.9773783700030989, 'eval/demon_fire_rate_model': 0.9001112966054535, 'eval/demon_fire_rate_teacher': 0.8956594323873122, 'eval/demon_left_precision': 0.916267942583732, 'eval/demon_left_recall': 0.8201284796573876, 'eval/demon_left_f1': 0.8655367231638418, 'eval/demon_left_rate_model': 0.11630495269894268, 'eval/demon_left_rate_teacher': 0.12993878686700056, 'eval/demon_right_precision': 0.8796296296296297, 'eval/demon_right_recall': 0.8636363636363636, 'eval/demon_right_f1': 0.8715596330275229, 'eval/demon_right_rate_model': 0.18030050083472454, 'eval/demon_right_rate_teacher': 0.18363939899833054, 'eval/demon_move_f1': 0.8685481780956823, 'eval/demon_cc_f1': 0.9419693813720712, 'eval/latency_8/demon_fire_precision': 0.9749613601236476, 'eval/latency_8/demon_fire_recall': 0.9798073936004971, 'eval/latency_8/demon_fire_f1': 0.9773783700030989, 'eval/latency_8/demon_fire_rate_model': 0.9001112966054535, 'eval/latency_8/demon_fire_rate_teacher': 0.8956594323873122, 'eval/latency_8/demon_left_precision': 0.916267942583732, 'eval/latency_8/demon_left_recall': 0.8201284796573876, 'eval/latency_8/demon_left_f1': 0.8655367231638418, 'eval/latency_8/demon_left_rate_model': 0.11630495269894268, 'eval/latency_8/demon_left_rate_teacher': 0.12993878686700056, 'eval/latency_8/demon_right_precision': 0.8796296296296297, 'eval/latency_8/demon_right_recall': 0.8636363636363636, 'eval/latency_8/demon_right_f1': 0.8715596330275229, 'eval/latency_8/demon_right_rate_model': 0.18030050083472454, 'eval/latency_8/demon_right_rate_teacher': 0.18363939899833054, 'eval/latency_8/demon_move_f1': 0.8685481780956823, 'eval/latency_8/demon_cc_f1': 0.9419693813720712, 'timing/data': 0.0009375326335430145, 'timing/model': 0.9375580791383982, 'learning_rate/qwen_vl_interface': 5e-07, 'learning_rate/action_model': 4.9999999999999996e-06, 'epoch': 3.18}) +[2026-06-18 03:26:26,939][accelerate.accelerator][INFO] - Saving current state to /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_7000_state +[2026-06-18 03:26:26,963][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-18 03:26:33,087][accelerate.checkpointing][INFO] - Model weights saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_7000_state/model.safetensors +[2026-06-18 03:26:48,666][accelerate.checkpointing][INFO] - Optimizer state saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_7000_state/optimizer.bin +[2026-06-18 03:26:48,666][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_7000_state/sampler.bin +[2026-06-18 03:26:48,667][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_7000_state/sampler_1.bin +[2026-06-18 03:26:48,668][accelerate.checkpointing][INFO] - Random states saved in /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints/steps_7000_state/random_states_0.pkl +[2026-06-18 03:26:48,671][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 7000 +[2026-06-18 03:26:48,671][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-18 03:26:48,682][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-18 03:26:48,683][starVLA.training.train_starvla][INFO] - Training complete. Final training state is saved under /inspire/hdd/global_user/liumingyu-253208120284/lzj/starvla/results/Checkpoints/demon_attack_fix_latency_8_200ep/checkpoints +[2026-06-18 03:26:48,698][starVLA.training.train_starvla][INFO] - ... and that's all, folks! diff --git a/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_8_200ep/summary.jsonl b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_8_200ep/summary.jsonl new file mode 100644 index 0000000000000000000000000000000000000000..67f4020ac3871f475553f86f0c342379cb905502 --- /dev/null +++ b/latency-aware/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_8_200ep/summary.jsonl @@ -0,0 +1,70 @@ +{"steps": 100} +{"steps": 200} +{"steps": 300} +{"steps": 400} +{"steps": 500} +{"steps": 600} +{"steps": 700} +{"steps": 800} +{"steps": 900} +{"steps": 1000} +{"steps": 1100} +{"steps": 1200} +{"steps": 1300} +{"steps": 1400} +{"steps": 1500} +{"steps": 1600} +{"steps": 1700} +{"steps": 1800} +{"steps": 1900} +{"steps": 2000} +{"steps": 2100} +{"steps": 2200} +{"steps": 2300} +{"steps": 2400} +{"steps": 2500} +{"steps": 2600} +{"steps": 2700} +{"steps": 2800} +{"steps": 2900} +{"steps": 3000} +{"steps": 3100} +{"steps": 3200} +{"steps": 3300} +{"steps": 3400} +{"steps": 3500} +{"steps": 3600} +{"steps": 3700} +{"steps": 3800} +{"steps": 3900} +{"steps": 4000} +{"steps": 4100} +{"steps": 4200} +{"steps": 4300} +{"steps": 4400} +{"steps": 4500} +{"steps": 4600} +{"steps": 4700} +{"steps": 4800} +{"steps": 4900} +{"steps": 5000} +{"steps": 5100} +{"steps": 5200} +{"steps": 5300} +{"steps": 5400} +{"steps": 5500} +{"steps": 5600} +{"steps": 5700} +{"steps": 5800} +{"steps": 5900} +{"steps": 6000} +{"steps": 6100} +{"steps": 6200} +{"steps": 6300} +{"steps": 6400} +{"steps": 6500} +{"steps": 6600} +{"steps": 6700} +{"steps": 6800} +{"steps": 6900} +{"steps": 7000} diff --git a/latency-aware/demon-attack/vla/starvla-wanoft-h8/wan_oft_demon_attack_fix_latency_6_context5_standard_sft_5000_effbs128_224_currentce/checkpoints/steps_5000_pytorch_model.pt b/latency-aware/demon-attack/vla/starvla-wanoft-h8/wan_oft_demon_attack_fix_latency_6_context5_standard_sft_5000_effbs128_224_currentce/checkpoints/steps_5000_pytorch_model.pt new file mode 100644 index 0000000000000000000000000000000000000000..518eff0577508bd50060211b67b606f2b1ec6355 --- /dev/null +++ b/latency-aware/demon-attack/vla/starvla-wanoft-h8/wan_oft_demon_attack_fix_latency_6_context5_standard_sft_5000_effbs128_224_currentce/checkpoints/steps_5000_pytorch_model.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:791c66ee0313063e3eeec01a3eeb788e642f8e22952e232caa6562998d0754dd +size 25211531190 diff --git a/latency-aware/demon-attack/vla/starvla-wanoft-h8/wan_oft_demon_attack_fix_latency_6_context5_standard_sft_5000_effbs128_224_currentce/config.full.yaml b/latency-aware/demon-attack/vla/starvla-wanoft-h8/wan_oft_demon_attack_fix_latency_6_context5_standard_sft_5000_effbs128_224_currentce/config.full.yaml new file mode 100644 index 0000000000000000000000000000000000000000..ba476199dd7541bf36003090c6e7ad1148206d23 --- /dev/null +++ b/latency-aware/demon-attack/vla/starvla-wanoft-h8/wan_oft_demon_attack_fix_latency_6_context5_standard_sft_5000_effbs128_224_currentce/config.full.yaml @@ -0,0 +1,280 @@ +framework: + name: WanOFT + world_model: + base_wm: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/wanoft_bench/third_party/starVLA/playground/Pretrained_models/Wan-AI/Wan2.2-TI2V-5B-Diffusers + extract_layers: + - -1 + num_frames: 5 + qwenvl: + base_vlm: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/wanoft_bench/third_party/starVLA/playground/Pretrained_models/Wan-AI/Wan2.2-TI2V-5B-Diffusers + attn_implementation: sdpa + flex_backend: triton + enable_gradient_checkpointing: true + action_model: + action_model_type: MLP + action_dim: 7 + action_hidden_dim: 3072 + action_horizon: 8 + future_action_window_size: 7 + past_action_window_size: 0 + loss_type: current_discrete_ce + class_weights: null + future_loss_weight: null + action_query_source: mean + action_query_num_heads: 24 + state_dim: 7 + action_env_dim: 6 + kv_memory: + enabled: false + window: 4 + rollout_len: 8 + packed_train: false + rebased_sink: true +datasets: + vla_data: + dataset_py: lerobot_datasets + include_state: true + data_root_dir: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/wanoft_bench/third_party/starVLA/data/demon_attack_fix_latency_6_200ep_context5 + data_mix: demon_attack_train__bridge + eval_data_mix: null + custom_mixtures_path: null + action_type: discrete + sequential_step_sampling: false + eval_sequential_step_sampling: null + num_workers: 8 + eval_num_workers: 8 + prefetch_factor: 4 + persistent_workers: true + pin_memory: true + shuffle: true + action_balance: + enabled: false + strategy: balanced_epoch + action_key: action_id + target_flap_fraction: 0.3 + noop_id: 0 + flap_id: 1 + latency_curriculum: + enabled: false + strategy: exclusive + latencies: null + phase_steps: null + phase_distributions: null + new_latency_passes: 1.0 + replay_passes: 0.25 + target_total_passes: 2.0 + final_equalization: true + step_budget_mode: auto + eval_at_phase_end: false + save_at_phase_end: false + computed_plan: null + per_device_batch_size: 1 + load_all_data_for_training: true + num_obs_frames: 1 + image_mode: single + prompt_mode: raw + stitch_grid: + - 2 + - 2 + obs_image_size: + - 224 + - 224 + video_backend: torchvision_av + observation_indices: + - -4 + - -3 + - -2 + - -1 + - 0 + language_indices: + - 0 + state_indices: + - 0 + action_indices: + - 0 + - 1 + - 2 + - 3 + - 4 + - 5 + - 6 + - 7 + pack_image_sequence: true + image_sequence_length: 5 + context_images_column: observation.context_images +dataset: + source_hf: latency-sensitive-bench/memory-data + config_name: demon_attack_fixed_latency_6_200ep_7k2steps + source_subdir: null + converted_name: demon_attack_train__bridge + single_source_hf: '' + mixed_source_hf: '' + single_converted_name: demon_attack_train__bridge + mixed_converted_name: demon_attack_mixed_latency_train__bridge + single_latency_filter: null + mixed_latency_filter: null + force_download: false + setup_force: false + skip_verification: false + target_latency_unit: observation_steps + verify_rows: 200 + max_episodes: 200 + episodes_per_latency: null + latency_filter: null + debug_subset: + enabled: false + max_episodes: 5 + suffix: debug +base_model: + repo_id: Wan-AI/Wan2.2-TI2V-5B-Diffusers +initialization: + checkpoint_local_dir: playground/Pretrained_models/WM4A-Wan2d2-OFT-LIBERO-4in1 + checkpoint_hf_repo_id: StarVLA/WM4A-Wan2d2-OFT-LIBERO-4in1 + checkpoint_filename: checkpoints/steps_60000_pytorch_model.pt +trainer: + max_train_steps: 5000 + num_warmup_steps: 0 + save_interval: 500 + eval_interval: 100 + eval_num_batches: 100 + per_latency_eval_num_batches: null + eval_action_classification: false + eval_action_classification_interval: null + cc_f1_tolerance: 1 + learning_rate: + base: 2.5e-05 + qwen_vl_interface: 1.0e-05 + action_model: 0.0001 + action_query_proj: 0.0001 + lr_scheduler_type: cosine_with_min_lr + scheduler_specific_kwargs: + min_lr: 1.0e-06 + freeze_modules: '' + freeze_vit: false + freeze_tied_embedding: false + freeze_llm_layers: [] + loss_scale: + vla: 1.0 + vlm: 0.1 + max_grad_norm: 1.0 + weight_decay: 0.0 + logging_frequency: 1 + profile_timing: + enabled: false + log_interval: 10 + gradient_clipping: 1.0 + gradient_accumulation_steps: 16 + distributed_backend: deepspeed + is_resume: true + pretrained_checkpoint: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/wanoft_bench/third_party/starVLA/results/Checkpoints/wan_oft_demon_attack_fix_latency_6_context5_standard_sft_5000_effbs128_224_currentce/checkpoints/steps_4000_state + resume_step: 4000 + reload_modules: null + optimizer: + name: AdamW + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + weight_decay: 1.0e-08 + fused: true + save_format: pt +workspace_dir: WORKSPACE_DIR +run_root_dir: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/wanoft_bench/third_party/starVLA/results/Checkpoints +seed: 42 +wandb_entity: zihanwang-ai-northwestern-university +wandb_project: starVLA_rl_games +auth: + env_file: null + hf_token_env: HF_TOKEN + wandb_api_key_env: WANDB_API_KEY +paths: + run_root_dir: results/Checkpoints + dataset_local_dir: data/demon_attack_fix_latency_6_200ep_context5 + dataset_cache_dir: null + base_model_dir: playground/Pretrained_models/Wan-AI/Wan2.2-TI2V-5B-Diffusers + accelerate_config: starVLA/config/deepseeds/deepspeed_zero2.yaml +launch: + use_accelerate: true + gpus: null + num_processes: 1 + dry_run: false +conda: + enabled: true + env_name: null +rl_games: + model_alias: wan_oft + env_eval: + image_size: 224 + frameskip: 4 + image_transform: raw_rgb + prompt_mode: raw + ghost_trail: + history_frames: 5 + gamma: 1.3 + min_alpha: 35 + scroll_px_per_step: 4.0 + ground_fraction: 0.22 + seed: 42 + fixed_episode_seeds: true + latency_seed_stride: 0 + task_seed_stride: 0 + task_description: 'You are playing Demon Attack from a single game image. Choose + exactly one action from: NOOP, FIRE, RIGHT, LEFT, RIGHTFIRE, LEFTFIRE.' + eval_parallel_envs: 5 + action_chunk_execution: + enabled: false + chunk_size: null + demon_attack: + noop_max: 30 + enabled: true + eval_backend: eval_core + distributed_mode: rank_sharded + vectorized: + enabled: false + batch_size: 1 + latency: + prompt_map_path: data/demon_attack_fix_latency_6_200ep_context5/demon_attack_train__bridge/latency_prompt_map.json + mode: single + values: + - 0 + mid_train: + enabled: false + interval_steps: 100 + latencies: + - 0 + num_episodes: 5 + max_steps_per_episode: 3600 + post_train: + enabled: false + latencies: + - 6 + num_episodes: 20 + max_steps_per_episode: 3600 + task: demon_attack + initialization_mode: bridge + action_carrier: bridge +model: wan_oft +env: demon_attack +init: wan_oft_libero +mode: single +checkpoint: + load: auto + hf_repo_id: null + save_best_model: false + save_final_model: true + save_pt_file: true + save_training_state: true + save_safetensors_file: false + local: + keep_last_n: 1 + sync: + enabled: false + repo_id: null + keep_last_n: 0 + sync_every_n_checkpoints: 1 + resume_policy: local_latest +run_id: wan_oft_demon_attack_fix_latency_6_context5_standard_sft_5000_effbs128_224_currentce +output_dir: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/wanoft_bench/third_party/starVLA/results/Checkpoints/wan_oft_demon_attack_fix_latency_6_context5_standard_sft_5000_effbs128_224_currentce +config_yaml: null +is_debug: false +version_id: '0.21' diff --git a/latency-aware/demon-attack/vla/starvla-wanoft-h8/wan_oft_demon_attack_fix_latency_6_context5_standard_sft_5000_effbs128_224_currentce/config.yaml b/latency-aware/demon-attack/vla/starvla-wanoft-h8/wan_oft_demon_attack_fix_latency_6_context5_standard_sft_5000_effbs128_224_currentce/config.yaml new file mode 100644 index 0000000000000000000000000000000000000000..9df6eda43084bef403a3b010cd8b61c67ce29b0d --- /dev/null +++ b/latency-aware/demon-attack/vla/starvla-wanoft-h8/wan_oft_demon_attack_fix_latency_6_context5_standard_sft_5000_effbs128_224_currentce/config.yaml @@ -0,0 +1,89 @@ +checkpoint: + local: + keep_last_n: 1 + save_best_model: false + save_final_model: true + save_pt_file: true + save_safetensors_file: false + save_training_state: true + sync: + enabled: false + keep_last_n: 0 + repo_id: null +datasets: + vla_data: + data_mix: demon_attack_train__bridge + dataset_py: lerobot_datasets + eval_data_mix: null + latency_curriculum: + enabled: false + obs_image_size: + - 224 + - 224 + per_device_batch_size: 1 +framework: + action_model: + action_dim: 7 + action_env_dim: 6 + action_hidden_dim: 3072 + action_horizon: 8 + action_model_type: MLP + action_query_source: mean + class_weights: null + loss_type: current_discrete_ce + name: WanOFT + qwenvl: + base_vlm: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/wanoft_bench/third_party/starVLA/playground/Pretrained_models/Wan-AI/Wan2.2-TI2V-5B-Diffusers + world_model: + base_wm: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/wanoft_bench/third_party/starVLA/playground/Pretrained_models/Wan-AI/Wan2.2-TI2V-5B-Diffusers + extract_layers: + - -1 + num_frames: 5 +output_dir: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/wanoft_bench/third_party/starVLA/results/Checkpoints/wan_oft_demon_attack_fix_latency_6_context5_standard_sft_5000_effbs128_224_currentce +rl_games: + env_eval: + enabled: true + eval_backend: eval_core + mid_train: + enabled: false + interval_steps: 100 + prompt_mode: raw +run_id: wan_oft_demon_attack_fix_latency_6_context5_standard_sft_5000_effbs128_224_currentce +run_root_dir: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/wanoft_bench/third_party/starVLA/results/Checkpoints +seed: 42 +trainer: + distributed_backend: deepspeed + eval_action_classification: false + eval_action_classification_interval: null + eval_interval: 100 + freeze_llm_layers: [] + freeze_modules: '' + freeze_tied_embedding: false + freeze_vit: false + gradient_accumulation_steps: 16 + is_resume: true + learning_rate: + action_model: 0.0001 + action_query_proj: 0.0001 + base: 2.5e-05 + qwen_vl_interface: 1.0e-05 + logging_frequency: 1 + lr_scheduler_type: cosine_with_min_lr + max_train_steps: 5000 + num_warmup_steps: 0 + optimizer: + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + fused: true + weight_decay: 1.0e-08 + pretrained_checkpoint: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/wanoft_bench/third_party/starVLA/results/Checkpoints/wan_oft_demon_attack_fix_latency_6_context5_standard_sft_5000_effbs128_224_currentce/checkpoints/steps_4000_state + profile_timing: + enabled: false + resume_step: 4000 + save_interval: 500 + scheduler_specific_kwargs: + min_lr: 1.0e-06 +wandb_entity: zihanwang-ai-northwestern-university +wandb_project: starVLA_rl_games diff --git a/latency-aware/demon-attack/vla/starvla-wanoft-h8/wan_oft_demon_attack_fix_latency_6_context5_standard_sft_5000_effbs128_224_currentce/dataset_statistics.json b/latency-aware/demon-attack/vla/starvla-wanoft-h8/wan_oft_demon_attack_fix_latency_6_context5_standard_sft_5000_effbs128_224_currentce/dataset_statistics.json new file mode 100644 index 0000000000000000000000000000000000000000..f9ec5ee85e5d33fe9b8a3fb41a763ba4c21394c2 --- /dev/null +++ b/latency-aware/demon-attack/vla/starvla-wanoft-h8/wan_oft_demon_attack_fix_latency_6_context5_standard_sft_5000_effbs128_224_currentce/dataset_statistics.json @@ -0,0 +1,127 @@ +{ + "new_embodiment": { + "action": { + "mean": [ + 0.05213047191500664, + 0.25790393352508545, + 0.17633378505706787, + 0.08325807005167007, + 0.18504540622234344, + 0.24532835185527802, + 0.0 + ], + "std": [ + 0.22280143201351166, + 0.4365144968032837, + 0.3812168538570404, + 0.2774599492549896, + 0.3870716392993927, + 0.4308818578720093, + 0.0 + ], + "max": [ + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + 0.0 + ], + "mask": [ + true, + true, + true, + true, + true, + true, + true + ] + }, + "state": { + "mean": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ] + }, + "num_transitions": 1055141, + "num_trajectories": 198 + } +} \ No newline at end of file diff --git a/latency-aware/demon-attack/vla/starvla-wanoft-h8/wan_oft_demon_attack_fix_latency_6_context5_standard_sft_5000_effbs128_224_currentce/hydra/.hydra/config.yaml b/latency-aware/demon-attack/vla/starvla-wanoft-h8/wan_oft_demon_attack_fix_latency_6_context5_standard_sft_5000_effbs128_224_currentce/hydra/.hydra/config.yaml new file mode 100644 index 0000000000000000000000000000000000000000..ec3780f2060484350f479ffa83ae858188485501 --- /dev/null +++ b/latency-aware/demon-attack/vla/starvla-wanoft-h8/wan_oft_demon_attack_fix_latency_6_context5_standard_sft_5000_effbs128_224_currentce/hydra/.hydra/config.yaml @@ -0,0 +1,277 @@ +framework: + qwenvl: + base_vlm: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/wanoft_bench/third_party/starVLA/playground/Pretrained_models/Wan-AI/Wan2.2-TI2V-5B-Diffusers + attn_implementation: sdpa + flex_backend: triton + enable_gradient_checkpointing: true + action_model: + state_dim: 7 + loss_type: current_discrete_ce + action_horizon: 8 + future_action_window_size: 7 + past_action_window_size: 0 + action_model_type: MLP + action_dim: 7 + action_env_dim: 6 + action_query_source: mean + action_query_num_heads: 24 + kv_memory: + enabled: false + window: 4 + rollout_len: 8 + packed_train: false + rebased_sink: true + name: WanOFT + world_model: + base_wm: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/wanoft_bench/third_party/starVLA/playground/Pretrained_models/Wan-AI/Wan2.2-TI2V-5B-Diffusers + extract_layers: + - -1 + num_frames: 5 +datasets: + vla_data: + dataset_py: lerobot_datasets + include_state: true + data_root_dir: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/wanoft_bench/third_party/starVLA/data/demon_attack_fix_latency_6_200ep_context5 + data_mix: demon_attack_train__bridge + eval_data_mix: null + custom_mixtures_path: null + action_type: discrete + sequential_step_sampling: false + eval_sequential_step_sampling: null + num_workers: 8 + eval_num_workers: 8 + prefetch_factor: 4 + persistent_workers: true + pin_memory: true + shuffle: true + action_balance: + enabled: false + strategy: balanced_epoch + action_key: action_id + target_flap_fraction: 0.3 + noop_id: 0 + flap_id: 1 + latency_curriculum: + enabled: false + strategy: exclusive + latencies: null + phase_steps: null + phase_distributions: null + new_latency_passes: 1.0 + replay_passes: 0.25 + target_total_passes: 2.0 + final_equalization: true + step_budget_mode: auto + eval_at_phase_end: false + save_at_phase_end: false + computed_plan: null + per_device_batch_size: 1 + load_all_data_for_training: true + num_obs_frames: 1 + image_mode: single + prompt_mode: raw + stitch_grid: + - 2 + - 2 + obs_image_size: + - 224 + - 224 + video_backend: torchvision_av + observation_indices: + - -4 + - -3 + - -2 + - -1 + - 0 + language_indices: + - 0 + state_indices: + - 0 + action_indices: + - 0 + - 1 + - 2 + - 3 + - 4 + - 5 + - 6 + - 7 + pack_image_sequence: true + image_sequence_length: 5 + context_images_column: observation.context_images +dataset: + source_hf: latency-sensitive-bench/memory-rollouts + config_name: demon_attack_fixed_latency_6_200ep_7k2steps + source_subdir: null + converted_name: demon_attack_train__bridge + single_source_hf: '' + mixed_source_hf: '' + single_converted_name: demon_attack_train__bridge + mixed_converted_name: demon_attack_mixed_latency_train__bridge + single_latency_filter: null + mixed_latency_filter: null + force_download: false + setup_force: false + skip_verification: false + target_latency_unit: observation_steps + verify_rows: 200 + max_episodes: 200 + episodes_per_latency: null + latency_filter: null + debug_subset: + enabled: false + max_episodes: 5 + suffix: debug +base_model: + repo_id: Wan-AI/Wan2.2-TI2V-5B-Diffusers +initialization: + checkpoint_local_dir: playground/Pretrained_models/WM4A-Wan2d2-OFT-LIBERO-4in1 + checkpoint_hf_repo_id: StarVLA/WM4A-Wan2d2-OFT-LIBERO-4in1 + checkpoint_filename: checkpoints/steps_60000_pytorch_model.pt +trainer: + max_train_steps: 5000 + num_warmup_steps: 0 + save_interval: 500 + eval_interval: 100 + eval_num_batches: 100 + per_latency_eval_num_batches: null + eval_action_classification: false + eval_action_classification_interval: null + cc_f1_tolerance: 1 + learning_rate: + base: 2.5e-05 + qwen_vl_interface: 1.0e-05 + action_model: 0.0001 + action_query_proj: 0.0001 + lr_scheduler_type: cosine_with_min_lr + scheduler_specific_kwargs: + min_lr: 1.0e-06 + freeze_modules: '' + freeze_vit: false + freeze_tied_embedding: false + freeze_llm_layers: [] + loss_scale: + vla: 1.0 + vlm: 0.1 + max_grad_norm: 1.0 + weight_decay: 0.0 + logging_frequency: 1 + profile_timing: + enabled: false + log_interval: 10 + gradient_clipping: 1.0 + gradient_accumulation_steps: 16 + distributed_backend: deepspeed + is_resume: true + pretrained_checkpoint: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/wanoft_bench/third_party/starVLA/results/Checkpoints/wan_oft_demon_attack_fix_latency_6_context5_standard_sft_5000_effbs128_224_currentce/checkpoints/steps_4000_state + resume_step: 4000 + reload_modules: null + optimizer: + name: AdamW + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + weight_decay: 1.0e-08 + fused: true + save_format: pt +workspace_dir: WORKSPACE_DIR +run_root_dir: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/wanoft_bench/third_party/starVLA/results/Checkpoints +seed: 42 +wandb_entity: ${oc.env:WANDB_ENTITY} +wandb_project: ${oc.env:WANDB_PROJECT,starVLA_rl_games} +auth: + env_file: null + hf_token_env: HF_TOKEN + wandb_api_key_env: WANDB_API_KEY +paths: + run_root_dir: results/Checkpoints + dataset_local_dir: data/demon_attack_fix_latency_6_200ep_context5 + dataset_cache_dir: null + base_model_dir: playground/Pretrained_models/Wan-AI/Wan2.2-TI2V-5B-Diffusers + accelerate_config: starVLA/config/deepseeds/deepspeed_zero2.yaml +launch: + use_accelerate: true + gpus: null + num_processes: 1 + dry_run: false +conda: + enabled: true + env_name: null +rl_games: + model_alias: wan_oft + env_eval: + image_size: 224 + frameskip: 4 + image_transform: raw_rgb + prompt_mode: raw + ghost_trail: + history_frames: 5 + gamma: 1.3 + min_alpha: 35 + scroll_px_per_step: 4.0 + ground_fraction: 0.22 + seed: 42 + fixed_episode_seeds: true + latency_seed_stride: 0 + task_seed_stride: 0 + task_description: 'You are playing Demon Attack from a single game image. Choose + exactly one action from: NOOP, FIRE, RIGHT, LEFT, RIGHTFIRE, LEFTFIRE.' + eval_parallel_envs: 5 + action_chunk_execution: + enabled: false + chunk_size: null + demon_attack: + noop_max: 30 + enabled: true + eval_backend: eval_core + distributed_mode: rank_sharded + vectorized: + enabled: false + batch_size: 1 + latency: + prompt_map_path: data/demon_attack_fix_latency_6_200ep_context5/demon_attack_train__bridge/latency_prompt_map.json + mode: single + values: + - 0 + mid_train: + enabled: false + interval_steps: 100 + latencies: + - 0 + num_episodes: 5 + max_steps_per_episode: 3600 + post_train: + enabled: false + latencies: + - 6 + num_episodes: 20 + max_steps_per_episode: 3600 + task: demon_attack + initialization_mode: bridge + action_carrier: bridge +model: wan_oft +env: demon_attack +init: wan_oft_libero +mode: single +checkpoint: + load: auto + hf_repo_id: null + save_best_model: false + save_final_model: true + save_pt_file: true + save_training_state: true + save_safetensors_file: false + local: + keep_last_n: 1 + sync: + enabled: false + repo_id: null + keep_last_n: 0 + sync_every_n_checkpoints: 1 + resume_policy: local_latest +run_id: wan_oft_demon_attack_fix_latency_6_context5_standard_sft_5000_effbs128_224_currentce +output_dir: null +config_yaml: null +is_debug: false +version_id: 0.21 diff --git a/latency-aware/flappy/small-policy/memory_rollout_teacher:flappy:teacher:fixed_l3:fs4:obs30:stride1:seed0/train/memory_rollout_teacher:flappy:teacher:fixed_l3:fs4:obs30:stride1:seed0/checkpoint_p0/checkpoint_000048840_25034752.pth b/latency-aware/flappy/small-policy/memory_rollout_teacher:flappy:teacher:fixed_l3:fs4:obs30:stride1:seed0/train/memory_rollout_teacher:flappy:teacher:fixed_l3:fs4:obs30:stride1:seed0/checkpoint_p0/checkpoint_000048840_25034752.pth new file mode 100644 index 0000000000000000000000000000000000000000..463cefb462bc9ed87aa3a344a8b5a2f3c715ad76 --- /dev/null +++ b/latency-aware/flappy/small-policy/memory_rollout_teacher:flappy:teacher:fixed_l3:fs4:obs30:stride1:seed0/train/memory_rollout_teacher:flappy:teacher:fixed_l3:fs4:obs30:stride1:seed0/checkpoint_p0/checkpoint_000048840_25034752.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b6142e6465fd2a56b7ac2fad4b171f5ce67e892bd148da05033cf5fbba1ff126 +size 8171325 diff --git a/latency-aware/flappy/small-policy/memory_rollout_teacher:flappy:teacher:fixed_l3:fs4:obs30:stride1:seed0/train/memory_rollout_teacher:flappy:teacher:fixed_l3:fs4:obs30:stride1:seed0/config.json b/latency-aware/flappy/small-policy/memory_rollout_teacher:flappy:teacher:fixed_l3:fs4:obs30:stride1:seed0/train/memory_rollout_teacher:flappy:teacher:fixed_l3:fs4:obs30:stride1:seed0/config.json new file mode 100644 index 0000000000000000000000000000000000000000..aefe4d2a0d4368b749c478a30f9d82c404614907 --- /dev/null +++ b/latency-aware/flappy/small-policy/memory_rollout_teacher:flappy:teacher:fixed_l3:fs4:obs30:stride1:seed0/train/memory_rollout_teacher:flappy:teacher:fixed_l3:fs4:obs30:stride1:seed0/config.json @@ -0,0 +1,266 @@ +{ + "help": false, + "algo": "APPO", + "env": "latency_flappy", + "experiment": "flappy_teacher_fixed_l3_obs30_seed0", + "train_dir": "results/checkpoints/memory/rollout_teacher", + "restart_behavior": "resume", + "device": "gpu", + "seed": 0, + "num_policies": 1, + "async_rl": true, + "serial_mode": false, + "batched_sampling": true, + "num_batches_to_accumulate": 2, + "worker_num_splits": 1, + "policy_workers_per_policy": 1, + "max_policy_lag": 400, + "num_workers": 2, + "num_envs_per_worker": 1, + "batch_size": 4096, + "num_batches_per_epoch": 8, + "num_epochs": 8, + "rollout": 128, + "recurrence": 1, + "shuffle_minibatches": false, + "gamma": 0.99, + "reward_scale": 1.0, + "reward_clip": 1000.0, + "value_bootstrap": false, + "normalize_returns": true, + "exploration_loss_coeff": 0.003, + "value_loss_coeff": 0.5, + "kl_loss_coeff": 0.0, + "exploration_loss": "entropy", + "gae_lambda": 0.95, + "ppo_clip_ratio": 0.1, + "ppo_clip_value": 0.2, + "with_vtrace": false, + "vtrace_rho": 1.0, + "vtrace_c": 1.0, + "optimizer": "adam", + "adam_eps": 1e-05, + "adam_beta1": 0.9, + "adam_beta2": 0.999, + "max_grad_norm": 0.5, + "learning_rate": 0.00025, + "lr_schedule": "linear_decay", + "lr_schedule_kl_threshold": 0.008, + "lr_adaptive_min": 1e-06, + "lr_adaptive_max": 0.01, + "obs_subtract_mean": 0.0, + "obs_scale": 255.0, + "normalize_input": true, + "normalize_input_keys": null, + "decorrelate_experience_max_seconds": 0, + "decorrelate_envs_on_one_worker": true, + "actor_worker_gpus": [ + 0 + ], + "set_workers_cpu_affinity": true, + "force_envs_single_thread": false, + "default_niceness": 0, + "log_to_file": true, + "experiment_summaries_interval": 1, + "flush_summaries_interval": 30, + "stats_avg": 100, + "summaries_use_frameskip": true, + "heartbeat_interval": 20, + "heartbeat_reporting_interval": 180, + "train_for_env_steps": 25000000, + "train_for_seconds": 10000000000, + "save_every_sec": 600, + "keep_checkpoints": 5, + "load_checkpoint_kind": "latest", + "save_milestones_sec": -1, + "save_best_every_sec": 5, + "save_best_metric": "reward", + "save_best_after": 100000, + "benchmark": false, + "encoder_mlp_layers": [ + 512, + 512 + ], + "encoder_conv_architecture": "convnet_atari", + "encoder_conv_mlp_layers": [ + 512 + ], + "use_rnn": false, + "rnn_size": 512, + "rnn_type": "gru", + "rnn_num_layers": 1, + "decoder_mlp_layers": [], + "nonlinearity": "elu", + "policy_initialization": "orthogonal", + "policy_init_gain": 1.0, + "actor_critic_share_weights": true, + "adaptive_stddev": true, + "continuous_tanh_scale": 0.0, + "initial_stddev": 1.0, + "use_env_info_cache": false, + "env_gpu_actions": true, + "env_gpu_observations": true, + "env_frameskip": 1, + "env_framestack": 1, + "pixel_format": "CHW", + "use_record_episode_statistics": false, + "with_wandb": true, + "wandb_user": null, + "wandb_project": "latency-sensitive-bench", + "wandb_group": "flappy-obs30-fixed_l3", + "wandb_job_type": "sample_factory", + "wandb_tags": [ + "memory_rollout_teacher", + "flappy", + "teacher", + "fixed", + "fixed_l3", + "obs30", + "seed0" + ], + "with_pbt": false, + "pbt_mix_policies_in_one_env": true, + "pbt_period_env_steps": 5000000, + "pbt_start_mutation": 20000000, + "pbt_replace_fraction": 0.3, + "pbt_mutation_rate": 0.15, + "pbt_replace_reward_gap": 0.1, + "pbt_replace_reward_gap_absolute": 1e-06, + "pbt_optimize_gamma": false, + "pbt_target_objective": "true_objective", + "pbt_perturb_min": 1.1, + "pbt_perturb_max": 1.5, + "gym_id": "FlappyBird-v0", + "env_fps": 30.0, + "obs_fps": 30.0, + "use_lidar": false, + "normalize_obs": true, + "audio_on": false, + "screen_size": "", + "obs_resize": "84,84", + "use_gpu_render": true, + "simulator": "gpu", + "gpu_render_device": "auto", + "gpu_render_batch_size": 128, + "gpu_render_profile": false, + "gpu_render_profile_interval": 200, + "pipe_gap": 100, + "bird_color": "yellow", + "pipe_color": "green", + "background": "day", + "score_limit": -1, + "frame_stack": 4, + "debug": false, + "debug_timelimit_diagnostics": false, + "max_episode_steps": 0, + "mode": "train", + "latency_type": "fixed", + "fixed_latency_ms": 100.0, + "mean_latency_ms": null, + "std_latency_ms": null, + "min_latency_ms": null, + "max_latency_ms": null, + "latency_seed": null, + "add_latency_info": false, + "max_pending_actions": null, + "hold_policy": "one_frame_then_noop", + "ordering_policy": "latest_ready", + "eval_episodes": 100, + "eval_parallel_envs": 50, + "eval_latency_raw_frame_values": "0,1,2,3,4,5", + "eval_max_steps": 7200, + "eval_deterministic": true, + "eval_raw_reward": false, + "episode_metrics_path": "results/checkpoints/memory/rollout_teacher/flappy_teacher_fixed_l3_obs30_seed0/episode_metrics.jsonl", + "command_line": "--mode train --algo APPO --env latency_flappy --experiment flappy_teacher_fixed_l3_obs30_seed0 --train_dir results/checkpoints/memory/rollout_teacher --restart_behavior resume --device gpu --actor_worker_gpus 0 --env_gpu_observations True --env_gpu_actions True --gpu-render-batch-size 128 --seed 0 --episode_metrics_path results/checkpoints/memory/rollout_teacher/flappy_teacher_fixed_l3_obs30_seed0/episode_metrics.jsonl --train_for_env_steps 25000000 --num_workers 2 --num_envs_per_worker 1 --num_policies 1 --batch_size 4096 --rollout 128 --recurrence 1 --num_epochs 8 --num_batches_per_epoch 8 --worker_num_splits 1 --max_policy_lag 400 --learning_rate 0.00025 --gamma 0.99 --gae_lambda 0.95 --ppo_clip_ratio 0.1 --ppo_clip_value 0.2 --value_loss_coeff 0.5 --max_grad_norm 0.5 --save_every_sec 600 --keep_checkpoints 5 --stats_avg 100 --experiment_summaries_interval 1 --batched_sampling True --async_rl True --use_rnn False --normalize_returns True --normalize_input True --latency-type fixed --fixed-latency-ms 100.0 --add-latency-info False --eval-episodes 100 --eval-parallel-envs 50 --eval-max-steps 7200 --eval-deterministic True --with_wandb True --wandb_project latency-sensitive-bench --wandb_group flappy-obs30-fixed_l3 --wandb_job_type sample_factory --wandb_tags memory_rollout_teacher flappy teacher fixed fixed_l3 obs30 seed0 --gym_id FlappyBird-v0 --env-fps 30 --obs-fps 30.0 --use_lidar False --normalize_obs True --audio_on False --obs_resize 84,84 --use-gpu-render True --simulator gpu --gpu-render-device auto --gpu-render-profile False --gpu-render-profile-interval 200 --pipe_gap 100 --bird_color yellow --pipe_color green --background day --frame_stack 4 --debug False --debug-timelimit-diagnostics False --hold-policy one_frame_then_noop --ordering-policy latest_ready", + "cli_args": { + "algo": "APPO", + "env": "latency_flappy", + "experiment": "flappy_teacher_fixed_l3_obs30_seed0", + "train_dir": "results/checkpoints/memory/rollout_teacher", + "restart_behavior": "resume", + "device": "gpu", + "seed": 0, + "num_policies": 1, + "async_rl": true, + "batched_sampling": true, + "worker_num_splits": 1, + "max_policy_lag": 400, + "num_workers": 2, + "num_envs_per_worker": 1, + "batch_size": 4096, + "num_batches_per_epoch": 8, + "num_epochs": 8, + "rollout": 128, + "recurrence": 1, + "gamma": 0.99, + "normalize_returns": true, + "value_loss_coeff": 0.5, + "gae_lambda": 0.95, + "ppo_clip_ratio": 0.1, + "ppo_clip_value": 0.2, + "max_grad_norm": 0.5, + "learning_rate": 0.00025, + "normalize_input": true, + "actor_worker_gpus": [ + 0 + ], + "experiment_summaries_interval": 1, + "stats_avg": 100, + "train_for_env_steps": 25000000, + "save_every_sec": 600, + "keep_checkpoints": 5, + "use_rnn": false, + "env_gpu_actions": true, + "env_gpu_observations": true, + "with_wandb": true, + "wandb_project": "latency-sensitive-bench", + "wandb_group": "flappy-obs30-fixed_l3", + "wandb_job_type": "sample_factory", + "wandb_tags": [ + "memory_rollout_teacher", + "flappy", + "teacher", + "fixed", + "fixed_l3", + "obs30", + "seed0" + ], + "gym_id": "FlappyBird-v0", + "env_fps": 30.0, + "obs_fps": 30.0, + "use_lidar": false, + "normalize_obs": true, + "audio_on": false, + "obs_resize": "84,84", + "use_gpu_render": true, + "simulator": "gpu", + "gpu_render_device": "auto", + "gpu_render_batch_size": 128, + "gpu_render_profile": false, + "gpu_render_profile_interval": 200, + "pipe_gap": 100, + "bird_color": "yellow", + "pipe_color": "green", + "background": "day", + "frame_stack": 4, + "debug": false, + "debug_timelimit_diagnostics": false, + "mode": "train", + "latency_type": "fixed", + "fixed_latency_ms": 100.0, + "add_latency_info": false, + "hold_policy": "one_frame_then_noop", + "ordering_policy": "latest_ready", + "eval_episodes": 100, + "eval_parallel_envs": 50, + "eval_max_steps": 7200, + "eval_deterministic": true, + "episode_metrics_path": "results/checkpoints/memory/rollout_teacher/flappy_teacher_fixed_l3_obs30_seed0/episode_metrics.jsonl" + }, + "git_hash": "bedb62f3a97309f6dacdb6a2953e3312d9776765", + "git_repo_name": "git@github.com:ZihanWang314/latency-sensitive-bench.git", + "eval_env_frameskip": 1, + "output_dir": "outputs/memory/rollout_teacher/flappy/train/teacher/fixed_l3/obs30/seed_0", + "wandb_unique_id": "flappy-obs30-fixed_l3-s0" +} \ No newline at end of file diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_5000_model.safetensors b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_5000_model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..d9e597b48a3f948cd940b0de899aa3602ca7ec80 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_5000_model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ad4903982cf1ec6f29c8ae4edea25e10b42353d56e0755b5ed123f5df3351754 +size 9138230516 diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_curriculum_cumulative_40ep_per_latency/config.full.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_curriculum_cumulative_40ep_per_latency/config.full.yaml new file mode 100644 index 0000000000000000000000000000000000000000..0af8d16fcb4e2ed1b05d95ea7cdcc2ec62d45ffc --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_curriculum_cumulative_40ep_per_latency/config.full.yaml @@ -0,0 +1,226 @@ +framework: + name: QwenOFT + qwenvl: + base_vlm: /workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + attn_implementation: flash_attention_2 + enable_gradient_checkpointing: true + action_model: + action_model_type: MLP + action_dim: 7 + action_hidden_dim: 2560 + future_action_window_size: 0 + past_action_window_size: 0 + loss_type: discrete_ce + state_dim: 7 + action_horizon: 1 + action_env_dim: 2 +datasets: + vla_data: + dataset_py: lerobot_datasets + include_state: true + data_root_dir: /workspace/starVLA/data/flappy_mixed_latency_40ep_per_lat + data_mix: flappy_mixed_latency_train__bridge + eval_data_mix: flappy_mixed_latency_train__bridge__val + custom_mixtures_path: null + action_type: discrete + sequential_step_sampling: true + eval_sequential_step_sampling: false + num_workers: 4 + eval_num_workers: 0 + shuffle: true + action_balance: + enabled: false + strategy: balanced_epoch + action_key: action_id + target_flap_fraction: 0.3 + noop_id: 0 + flap_id: 1 + latency_curriculum: + enabled: true + strategy: cumulative + latencies: + - 0 + - 1 + - 2 + - 3 + - 4 + - 5 + phase_steps: null + per_device_batch_size: 32 + load_all_data_for_training: true + obs_image_size: null + video_backend: torchvision_av +dataset: + source_hf: '' + config_name: null + source_subdir: null + converted_name: flappy_mixed_latency_train + single_source_hf: '' + mixed_source_hf: '' + single_converted_name: flappy_train + mixed_converted_name: flappy_mixed_latency_train + single_latency_filter: null + mixed_latency_filter: null + force_download: false + setup_force: false + skip_verification: false + verify_rows: 200 + max_episodes: null + episodes_per_latency: null + latency_filter: null + debug_subset: + enabled: false + max_episodes: 5 + suffix: debug +base_model: + repo_id: Qwen/Qwen3-VL-4B-Instruct +initialization: + checkpoint_local_dir: playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_hf_repo_id: StarVLA/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_filename: checkpoints/steps_5000_pytorch_model.pt +trainer: + max_train_steps: 5000 + num_warmup_steps: 100 + save_interval: 100 + eval_interval: 100 + eval_num_batches: 100 + per_latency_eval_num_batches: null + eval_action_classification: true + eval_action_classification_interval: null + cc_f1_tolerance: 1 + learning_rate: + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + action_model: 0.0001 + lr_scheduler_type: cosine_with_min_lr + scheduler_specific_kwargs: + min_lr: 1.0e-06 + freeze_modules: '' + loss_scale: + vla: 1.0 + vlm: 0.1 + max_grad_norm: 1.0 + weight_decay: 0.0 + logging_frequency: 1 + gradient_clipping: 1.0 + gradient_accumulation_steps: 4 + distributed_backend: none + is_resume: false + pretrained_checkpoint: null + resume_step: 0 + reload_modules: null + optimizer: + name: AdamW + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + weight_decay: 1.0e-08 + fused: false + save_format: pt +workspace_dir: WORKSPACE_DIR +run_root_dir: /workspace/starVLA/results/Checkpoints +seed: 42 +wandb_entity: dongqianyu99-zhejiang-university +wandb_project: starVLA_rl_games +auth: + env_file: null + hf_token_env: HF_TOKEN + wandb_api_key_env: WANDB_API_KEY +paths: + run_root_dir: results/Checkpoints + dataset_local_dir: playground/Datasets/rl_games + dataset_cache_dir: null + base_model_dir: playground/Pretrained_models/Qwen3-VL-4B-Instruct + accelerate_config: starVLA/config/deepseeds/deepspeed_zero2.yaml +launch: + use_accelerate: true + gpus: null + num_processes: 1 + dry_run: false +conda: + enabled: true + env_name: null +rl_games: + model_alias: openvla + env_eval: + image_size: 224 + frameskip: 1 + seed: 42 + fixed_episode_seeds: true + latency_seed_stride: 0 + task_seed_stride: 0 + task_description: '' + enabled: true + distributed_mode: none + vectorized: + enabled: false + batch_size: 1 + latency: + prompt_map_path: /workspace/starVLA/data/flappy_mixed_latency_40ep_per_lat/flappy_mixed_latency_train__bridge/latency_prompt_map.json + mode: mixed + values: + - 0 + - 1 + - 2 + - 3 + - 4 + - 5 + mid_train: + enabled: false + interval_steps: 100 + latencies: + - 0 + - 1 + - 2 + - 3 + - 4 + - 5 + num_episodes: 5 + max_steps_per_episode: 3600 + post_train: + enabled: false + latencies: + - 0 + - 1 + - 2 + - 3 + - 4 + num_episodes: 5 + max_steps_per_episode: 3600 + task: flappy + initialization_mode: bridge + action_carrier: bridge +model: openvla +env: flappy +init: bridge +bridge_base_model: + repo_id: + openvla: Qwen/Qwen3-VL-4B-Instruct + pi0: StarVLA/Qwen2.5-VL-3B-Instruct-Action + pi05: Qwen/Qwen3-VL-4B-Instruct + gr00t: Qwen/Qwen3-VL-4B-Instruct + local_dir: + openvla: playground/Pretrained_models/Qwen3-VL-4B-Instruct + pi0: playground/Pretrained_models/Qwen2.5-VL-3B-Instruct-Action + pi05: playground/Pretrained_models/Qwen3-VL-4B-Instruct + gr00t: playground/Pretrained_models/Qwen3-VL-4B-Instruct +mode: curriculum_cumulative +checkpoint: + load: auto + hf_repo_id: null + save_best_model: false + save_pt_file: false + local: + keep_last_n: 1 + sync: + enabled: false + repo_id: null + keep_last_n: 0 + sync_every_n_checkpoints: 1 + resume_policy: local_latest +run_id: flappy_curriculum_cumulative_40ep_per_latency +output_dir: /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency +config_yaml: null +is_debug: false +version_id: '0.21' diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_curriculum_cumulative_40ep_per_latency/config.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_curriculum_cumulative_40ep_per_latency/config.yaml new file mode 100644 index 0000000000000000000000000000000000000000..ef2f858ec9c4709ab613d5061cd3b166741ad329 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_curriculum_cumulative_40ep_per_latency/config.yaml @@ -0,0 +1,84 @@ +checkpoint: + local: + keep_last_n: 1 + save_best_model: false + save_pt_file: false + sync: + enabled: false + keep_last_n: 0 + repo_id: null +datasets: + vla_data: + data_mix: flappy_mixed_latency_train__bridge + dataset_py: lerobot_datasets + eval_data_mix: flappy_mixed_latency_train__bridge__val + latency_curriculum: + enabled: true + latencies: + - 0 + - 1 + - 2 + - 3 + - 4 + - 5 + phase_steps: null + strategy: cumulative + obs_image_size: null + per_device_batch_size: 32 +framework: + action_model: + action_dim: 7 + action_env_dim: 2 + action_hidden_dim: 2560 + action_horizon: 1 + action_model_type: MLP + loss_type: discrete_ce + name: QwenOFT + qwenvl: + attn_implementation: flash_attention_2 + base_vlm: /workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + enable_gradient_checkpointing: true +output_dir: /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency +rl_games: + env_eval: + enabled: true + mid_train: + enabled: false + interval_steps: 100 + task: flappy +run_id: flappy_curriculum_cumulative_40ep_per_latency +run_root_dir: /workspace/starVLA/results/Checkpoints +seed: 42 +trainer: + distributed_backend: none + eval_action_classification: true + eval_action_classification_interval: null + eval_interval: 100 + eval_num_batches: 100 + freeze_modules: '' + gradient_accumulation_steps: 4 + gradient_clipping: 1.0 + is_resume: false + learning_rate: + action_model: 0.0001 + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + logging_frequency: 1 + lr_scheduler_type: cosine_with_min_lr + max_train_steps: 5000 + num_warmup_steps: 100 + optimizer: + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + fused: false + weight_decay: 1.0e-08 + per_latency_eval_num_batches: null + pretrained_checkpoint: null + reload_modules: null + save_interval: 100 + scheduler_specific_kwargs: + min_lr: 1.0e-06 +wandb_entity: dongqianyu99-zhejiang-university +wandb_project: starVLA_rl_games diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_curriculum_cumulative_40ep_per_latency/dataset_statistics.json b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_curriculum_cumulative_40ep_per_latency/dataset_statistics.json new file mode 100644 index 0000000000000000000000000000000000000000..00c5bd8bf289b8861d281fa80c419f1969ba9d24 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_curriculum_cumulative_40ep_per_latency/dataset_statistics.json @@ -0,0 +1,127 @@ +{ + "new_embodiment": { + "action": { + "mean": [ + 0.6633877158164978, + 0.3366122841835022, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.4729187786579132, + 0.4729187786579132, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "mask": [ + true, + true, + true, + true, + true, + true, + true + ] + }, + "state": { + "mean": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ] + }, + "num_transitions": 341369, + "num_trajectories": 200 + } +} \ No newline at end of file diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_curriculum_cumulative_40ep_per_latency/dataset_statistics_eval.json b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_curriculum_cumulative_40ep_per_latency/dataset_statistics_eval.json new file mode 100644 index 0000000000000000000000000000000000000000..83b9bfe2a5a1d63a6bf5004f3b884056fb467ea8 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_curriculum_cumulative_40ep_per_latency/dataset_statistics_eval.json @@ -0,0 +1,127 @@ +{ + "new_embodiment": { + "action": { + "mean": [ + 0.8398444652557373, + 0.1601555496454239, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.3672837018966675, + 0.3672837018966675, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "mask": [ + true, + true, + true, + true, + true, + true, + true + ] + }, + "state": { + "mean": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ] + }, + "num_transitions": 360000, + "num_trajectories": 100 + } +} \ No newline at end of file diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_curriculum_cumulative_40ep_per_latency/hydra/.hydra/config.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_curriculum_cumulative_40ep_per_latency/hydra/.hydra/config.yaml new file mode 100644 index 0000000000000000000000000000000000000000..18df0364019d83de4e9f4c1a052b3e8cc218cc6f --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_curriculum_cumulative_40ep_per_latency/hydra/.hydra/config.yaml @@ -0,0 +1,224 @@ +framework: + qwenvl: + base_vlm: /workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + attn_implementation: flash_attention_2 + enable_gradient_checkpointing: true + action_model: + state_dim: 7 + loss_type: discrete_ce + action_horizon: 1 + future_action_window_size: 0 + past_action_window_size: 0 + action_dim: 7 + action_env_dim: 2 + name: QwenOFT +datasets: + vla_data: + dataset_py: lerobot_datasets + include_state: true + data_root_dir: /workspace/starVLA/data/flappy_mixed_latency_40ep_per_lat + data_mix: flappy_mixed_latency_train__bridge + eval_data_mix: flappy_mixed_latency_train__bridge__val + custom_mixtures_path: null + action_type: discrete + sequential_step_sampling: true + eval_sequential_step_sampling: false + num_workers: 4 + eval_num_workers: 0 + shuffle: true + action_balance: + enabled: false + strategy: balanced_epoch + action_key: action_id + target_flap_fraction: 0.3 + noop_id: 0 + flap_id: 1 + latency_curriculum: + enabled: true + strategy: cumulative + latencies: + - 0 + - 1 + - 2 + - 3 + - 4 + - 5 + phase_steps: null + per_device_batch_size: 32 + load_all_data_for_training: true + obs_image_size: null + video_backend: torchvision_av +dataset: + source_hf: ${dataset.mixed_source_hf} + config_name: null + source_subdir: null + converted_name: ${dataset.mixed_converted_name} + single_source_hf: '' + mixed_source_hf: '' + single_converted_name: flappy_train + mixed_converted_name: flappy_mixed_latency_train + single_latency_filter: null + mixed_latency_filter: null + force_download: false + setup_force: false + skip_verification: false + verify_rows: 200 + max_episodes: null + episodes_per_latency: null + latency_filter: ${dataset.mixed_latency_filter} + debug_subset: + enabled: false + max_episodes: 5 + suffix: debug +base_model: + repo_id: ${bridge_base_model.repo_id.${model}} +initialization: + checkpoint_local_dir: playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_hf_repo_id: StarVLA/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_filename: checkpoints/steps_5000_pytorch_model.pt +trainer: + max_train_steps: 5000 + num_warmup_steps: 100 + save_interval: 100 + eval_interval: 100 + eval_num_batches: 100 + per_latency_eval_num_batches: null + eval_action_classification: true + eval_action_classification_interval: null + cc_f1_tolerance: 1 + learning_rate: + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + action_model: 0.0001 + lr_scheduler_type: cosine_with_min_lr + scheduler_specific_kwargs: + min_lr: 1.0e-06 + freeze_modules: '' + loss_scale: + vla: 1.0 + vlm: 0.1 + max_grad_norm: 1.0 + weight_decay: 0.0 + logging_frequency: 1 + gradient_clipping: 1.0 + gradient_accumulation_steps: 4 + distributed_backend: none + is_resume: false + pretrained_checkpoint: /workspace/starVLA/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt + resume_step: 0 + reload_modules: null + optimizer: + name: AdamW + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + weight_decay: 1.0e-08 + fused: false + save_format: pt +workspace_dir: WORKSPACE_DIR +run_root_dir: /workspace/starVLA/results/Checkpoints +seed: 42 +wandb_entity: dongqianyu99-zhejiang-university +wandb_project: starVLA_rl_games +auth: + env_file: null + hf_token_env: HF_TOKEN + wandb_api_key_env: WANDB_API_KEY +paths: + run_root_dir: results/Checkpoints + dataset_local_dir: playground/Datasets/rl_games + dataset_cache_dir: null + base_model_dir: ${bridge_base_model.local_dir.${model}} + accelerate_config: starVLA/config/deepseeds/deepspeed_zero2.yaml +launch: + use_accelerate: true + gpus: null + num_processes: 1 + dry_run: false +conda: + enabled: true + env_name: null +rl_games: + model_alias: openvla + env_eval: + image_size: 224 + frameskip: 1 + seed: 42 + fixed_episode_seeds: true + latency_seed_stride: 0 + task_seed_stride: 0 + task_description: '' + enabled: true + distributed_mode: none + vectorized: + enabled: false + batch_size: 1 + latency: + prompt_map_path: /workspace/starVLA/data/flappy_mixed_latency_40ep_per_lat/flappy_mixed_latency_train__bridge/latency_prompt_map.json + mode: mixed + values: + - 0 + - 1 + - 2 + - 3 + - 4 + - 5 + mid_train: + enabled: false + interval_steps: 100 + latencies: + - 0 + - 1 + - 2 + - 3 + - 4 + - 5 + num_episodes: 5 + max_steps_per_episode: 3600 + post_train: + enabled: false + latencies: + - 0 + - 1 + - 2 + - 3 + - 4 + num_episodes: 5 + max_steps_per_episode: 3600 + task: flappy + initialization_mode: bridge + action_carrier: bridge +model: openvla +env: flappy +init: bridge +bridge_base_model: + repo_id: + openvla: Qwen/Qwen3-VL-4B-Instruct + pi0: StarVLA/Qwen2.5-VL-3B-Instruct-Action + pi05: Qwen/Qwen3-VL-4B-Instruct + gr00t: Qwen/Qwen3-VL-4B-Instruct + local_dir: + openvla: playground/Pretrained_models/Qwen3-VL-4B-Instruct + pi0: playground/Pretrained_models/Qwen2.5-VL-3B-Instruct-Action + pi05: playground/Pretrained_models/Qwen3-VL-4B-Instruct + gr00t: playground/Pretrained_models/Qwen3-VL-4B-Instruct +mode: curriculum_cumulative +checkpoint: + load: auto + hf_repo_id: null + save_best_model: false + save_pt_file: false + local: + keep_last_n: 1 + sync: + enabled: false + repo_id: null + keep_last_n: 0 + sync_every_n_checkpoints: 1 + resume_policy: local_latest +run_id: flappy_curriculum_cumulative_40ep_per_latency +output_dir: null +config_yaml: null +is_debug: false +version_id: '0.21' diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_curriculum_cumulative_40ep_per_latency/hydra/.hydra/hydra.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_curriculum_cumulative_40ep_per_latency/hydra/.hydra/hydra.yaml new file mode 100644 index 0000000000000000000000000000000000000000..ea8c2bdc62d2ce600267bd688751270040afa21a --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_curriculum_cumulative_40ep_per_latency/hydra/.hydra/hydra.yaml @@ -0,0 +1,255 @@ +hydra: + run: + dir: ${run_root_dir}/${run_id}/hydra + sweep: + dir: multirun/${now:%Y-%m-%d}/${now:%H-%M-%S} + subdir: ${hydra.job.num} + launcher: + _target_: hydra._internal.core_plugins.basic_launcher.BasicLauncher + sweeper: + _target_: hydra._internal.core_plugins.basic_sweeper.BasicSweeper + max_batch_size: null + params: null + help: + app_name: ${hydra.job.name} + header: '${hydra.help.app_name} is powered by Hydra. + + ' + footer: 'Powered by Hydra (https://hydra.cc) + + Use --hydra-help to view Hydra specific help + + ' + template: '${hydra.help.header} + + == Configuration groups == + + Compose your configuration from those groups (group=option) + + + $APP_CONFIG_GROUPS + + + == Config == + + Override anything in the config (foo.bar=value) + + + $CONFIG + + + ${hydra.help.footer} + + ' + hydra_help: + template: 'Hydra (${hydra.runtime.version}) + + See https://hydra.cc for more info. + + + == Flags == + + $FLAGS_HELP + + + == Configuration groups == + + Compose your configuration from those groups (For example, append hydra/job_logging=disabled + to command line) + + + $HYDRA_CONFIG_GROUPS + + + Use ''--cfg hydra'' to Show the Hydra config. + + ' + hydra_help: ??? + hydra_logging: + version: 1 + formatters: + simple: + format: '[%(asctime)s][HYDRA] %(message)s' + handlers: + console: + class: logging.StreamHandler + formatter: simple + stream: ext://sys.stdout + root: + level: INFO + handlers: + - console + loggers: + logging_example: + level: DEBUG + disable_existing_loggers: false + job_logging: + version: 1 + formatters: + simple: + format: '[%(asctime)s][%(name)s][%(levelname)s] - %(message)s' + handlers: + console: + class: logging.StreamHandler + formatter: simple + stream: ext://sys.stdout + file: + class: logging.FileHandler + formatter: simple + filename: ${hydra.runtime.output_dir}/${hydra.job.name}.log + root: + level: INFO + handlers: + - console + - file + disable_existing_loggers: false + env: {} + mode: RUN + searchpath: [] + callbacks: {} + output_subdir: .hydra + overrides: + hydra: + - hydra.mode=RUN + task: + - model=openvla + - env=flappy + - init=bridge + - mode=curriculum_cumulative + - run_id=flappy_curriculum_cumulative_40ep_per_latency + - run_root_dir=/workspace/starVLA/results/Checkpoints + - seed=42 + - wandb_entity=dongqianyu99-zhejiang-university + - wandb_project=starVLA_rl_games + - rl_games.env_eval.enabled=true + - checkpoint.sync.enabled=false + - checkpoint.sync.keep_last_n=0 + - checkpoint.local.keep_last_n=1 + - checkpoint.save_best_model=false + - checkpoint.save_pt_file=false + - trainer.is_resume=false + - trainer.pretrained_checkpoint=/workspace/starVLA/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt + - trainer.resume_step=0 + - trainer.max_train_steps=5000 + - trainer.num_warmup_steps=100 + - trainer.save_interval=100 + - trainer.eval_interval=100 + - trainer.eval_num_batches=100 + - trainer.logging_frequency=1 + - trainer.gradient_accumulation_steps=4 + - trainer.distributed_backend=none + - trainer.learning_rate.base=2e-05 + - trainer.learning_rate.qwen_vl_interface=1e-05 + - trainer.learning_rate.action_model=0.0001 + - trainer.lr_scheduler_type=cosine_with_min_lr + - trainer.scheduler_specific_kwargs.min_lr=1e-06 + - trainer.loss_scale.vla=1.0 + - trainer.loss_scale.vlm=0.1 + - trainer.max_grad_norm=1.0 + - trainer.weight_decay=0.0 + - trainer.gradient_clipping=1.0 + - trainer.optimizer.name=AdamW + - trainer.optimizer.betas=[0.9,0.95] + - trainer.optimizer.eps=1e-08 + - trainer.optimizer.weight_decay=1e-08 + - trainer.optimizer.fused=false + - trainer.save_format=pt + - framework.name=QwenOFT + - framework.qwenvl.attn_implementation=flash_attention_2 + - framework.qwenvl.enable_gradient_checkpointing=true + - framework.action_model.action_dim=7 + - framework.action_model.action_env_dim=2 + - framework.action_model.state_dim=7 + - framework.action_model.loss_type=discrete_ce + - framework.action_model.action_horizon=1 + - framework.action_model.future_action_window_size=0 + - framework.action_model.past_action_window_size=0 + - datasets.vla_data.include_state=true + - datasets.vla_data.action_type=discrete + - datasets.vla_data.sequential_step_sampling=true + - datasets.vla_data.shuffle=true + - datasets.vla_data.action_balance.enabled=false + - datasets.vla_data.action_balance.strategy=balanced_epoch + - datasets.vla_data.action_balance.action_key=action_id + - datasets.vla_data.action_balance.target_flap_fraction=0.3 + - datasets.vla_data.action_balance.noop_id=0 + - datasets.vla_data.action_balance.flap_id=1 + - datasets.vla_data.latency_curriculum.enabled=true + - datasets.vla_data.latency_curriculum.strategy=cumulative + - datasets.vla_data.latency_curriculum.latencies=[0,1,2,3,4,5] + - datasets.vla_data.per_device_batch_size=32 + - datasets.vla_data.load_all_data_for_training=true + - datasets.vla_data.video_backend=torchvision_av + - datasets.vla_data.data_root_dir=/workspace/starVLA/data/flappy_mixed_latency_40ep_per_lat + - datasets.vla_data.data_mix=flappy_mixed_latency_train__bridge + - datasets.vla_data.eval_data_mix=flappy_mixed_latency_train__bridge__val + - framework.qwenvl.base_vlm=/workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + - rl_games.task=flappy + - rl_games.model_alias=openvla + - rl_games.initialization_mode=bridge + - rl_games.action_carrier=bridge + - rl_games.env_eval.distributed_mode=none + - rl_games.env_eval.latency.mode=mixed + - rl_games.env_eval.frameskip=1 + - rl_games.env_eval.image_size=224 + - rl_games.env_eval.seed=42 + - rl_games.env_eval.fixed_episode_seeds=true + - rl_games.env_eval.latency_seed_stride=0 + - rl_games.env_eval.task_seed_stride=0 + - rl_games.env_eval.latency.values=[0,1,2,3,4,5] + - rl_games.env_eval.latency.prompt_map_path=/workspace/starVLA/data/flappy_mixed_latency_40ep_per_lat/flappy_mixed_latency_train__bridge/latency_prompt_map.json + - rl_games.env_eval.mid_train.enabled=false + - rl_games.env_eval.mid_train.interval_steps=100 + - rl_games.env_eval.mid_train.num_episodes=5 + - rl_games.env_eval.mid_train.max_steps_per_episode=3600 + - rl_games.env_eval.mid_train.latencies=[0,1,2,3,4,5] + - rl_games.env_eval.post_train.enabled=false + - rl_games.env_eval.post_train.num_episodes=5 + - rl_games.env_eval.post_train.max_steps_per_episode=3600 + - rl_games.env_eval.post_train.latencies=[0,1,2,3,4] + job: + name: train_starvla_hydra + chdir: false + override_dirname: checkpoint.local.keep_last_n=1,checkpoint.save_best_model=false,checkpoint.save_pt_file=false,checkpoint.sync.enabled=false,checkpoint.sync.keep_last_n=0,datasets.vla_data.action_balance.action_key=action_id,datasets.vla_data.action_balance.enabled=false,datasets.vla_data.action_balance.flap_id=1,datasets.vla_data.action_balance.noop_id=0,datasets.vla_data.action_balance.strategy=balanced_epoch,datasets.vla_data.action_balance.target_flap_fraction=0.3,datasets.vla_data.action_type=discrete,datasets.vla_data.data_mix=flappy_mixed_latency_train__bridge,datasets.vla_data.data_root_dir=/workspace/starVLA/data/flappy_mixed_latency_40ep_per_lat,datasets.vla_data.eval_data_mix=flappy_mixed_latency_train__bridge__val,datasets.vla_data.include_state=true,datasets.vla_data.latency_curriculum.enabled=true,datasets.vla_data.latency_curriculum.latencies=[0,1,2,3,4,5],datasets.vla_data.latency_curriculum.strategy=cumulative,datasets.vla_data.load_all_data_for_training=true,datasets.vla_data.per_device_batch_size=32,datasets.vla_data.sequential_step_sampling=true,datasets.vla_data.shuffle=true,datasets.vla_data.video_backend=torchvision_av,env=flappy,framework.action_model.action_dim=7,framework.action_model.action_env_dim=2,framework.action_model.action_horizon=1,framework.action_model.future_action_window_size=0,framework.action_model.loss_type=discrete_ce,framework.action_model.past_action_window_size=0,framework.action_model.state_dim=7,framework.name=QwenOFT,framework.qwenvl.attn_implementation=flash_attention_2,framework.qwenvl.base_vlm=/workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct,framework.qwenvl.enable_gradient_checkpointing=true,init=bridge,mode=curriculum_cumulative,model=openvla,rl_games.action_carrier=bridge,rl_games.env_eval.distributed_mode=none,rl_games.env_eval.enabled=true,rl_games.env_eval.fixed_episode_seeds=true,rl_games.env_eval.frameskip=1,rl_games.env_eval.image_size=224,rl_games.env_eval.latency.mode=mixed,rl_games.env_eval.latency.prompt_map_path=/workspace/starVLA/data/flappy_mixed_latency_40ep_per_lat/flappy_mixed_latency_train__bridge/latency_prompt_map.json,rl_games.env_eval.latency.values=[0,1,2,3,4,5],rl_games.env_eval.latency_seed_stride=0,rl_games.env_eval.mid_train.enabled=false,rl_games.env_eval.mid_train.interval_steps=100,rl_games.env_eval.mid_train.latencies=[0,1,2,3,4,5],rl_games.env_eval.mid_train.max_steps_per_episode=3600,rl_games.env_eval.mid_train.num_episodes=5,rl_games.env_eval.post_train.enabled=false,rl_games.env_eval.post_train.latencies=[0,1,2,3,4],rl_games.env_eval.post_train.max_steps_per_episode=3600,rl_games.env_eval.post_train.num_episodes=5,rl_games.env_eval.seed=42,rl_games.env_eval.task_seed_stride=0,rl_games.initialization_mode=bridge,rl_games.model_alias=openvla,rl_games.task=flappy,run_id=flappy_curriculum_cumulative_40ep_per_latency,run_root_dir=/workspace/starVLA/results/Checkpoints,seed=42,trainer.distributed_backend=none,trainer.eval_interval=100,trainer.eval_num_batches=100,trainer.gradient_accumulation_steps=4,trainer.gradient_clipping=1.0,trainer.is_resume=false,trainer.learning_rate.action_model=0.0001,trainer.learning_rate.base=2e-05,trainer.learning_rate.qwen_vl_interface=1e-05,trainer.logging_frequency=1,trainer.loss_scale.vla=1.0,trainer.loss_scale.vlm=0.1,trainer.lr_scheduler_type=cosine_with_min_lr,trainer.max_grad_norm=1.0,trainer.max_train_steps=5000,trainer.num_warmup_steps=100,trainer.optimizer.betas=[0.9,0.95],trainer.optimizer.eps=1e-08,trainer.optimizer.fused=false,trainer.optimizer.name=AdamW,trainer.optimizer.weight_decay=1e-08,trainer.pretrained_checkpoint=/workspace/starVLA/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt,trainer.resume_step=0,trainer.save_format=pt,trainer.save_interval=100,trainer.scheduler_specific_kwargs.min_lr=1e-06,trainer.weight_decay=0.0,wandb_entity=dongqianyu99-zhejiang-university,wandb_project=starVLA_rl_games + id: ??? + num: ??? + config_name: train + env_set: {} + env_copy: [] + config: + override_dirname: + kv_sep: '=' + item_sep: ',' + exclude_keys: [] + runtime: + version: 1.3.3 + version_base: '1.1' + cwd: /workspace/starVLA + config_sources: + - path: hydra.conf + schema: pkg + provider: hydra + - path: /workspace/starVLA/examples/rl_games/config + schema: file + provider: main + - path: '' + schema: structured + provider: schema + output_dir: /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/hydra + choices: + cross_task_setup: null + checkpoint: default + mode: curriculum_cumulative + init: bridge + env: flappy + model: openvla + hydra/env: default + hydra/callbacks: null + hydra/job_logging: default + hydra/hydra_logging: default + hydra/hydra_help: default + hydra/help: default + hydra/sweeper: basic + hydra/launcher: basic + hydra/output: default + verbose: false diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_curriculum_cumulative_40ep_per_latency/hydra/.hydra/overrides.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_curriculum_cumulative_40ep_per_latency/hydra/.hydra/overrides.yaml new file mode 100644 index 0000000000000000000000000000000000000000..0f99d541ff0fbc34012ae6060c6b2966cdd928b9 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_curriculum_cumulative_40ep_per_latency/hydra/.hydra/overrides.yaml @@ -0,0 +1,95 @@ +- model=openvla +- env=flappy +- init=bridge +- mode=curriculum_cumulative +- run_id=flappy_curriculum_cumulative_40ep_per_latency +- run_root_dir=/workspace/starVLA/results/Checkpoints +- seed=42 +- wandb_entity=dongqianyu99-zhejiang-university +- wandb_project=starVLA_rl_games +- rl_games.env_eval.enabled=true +- checkpoint.sync.enabled=false +- checkpoint.sync.keep_last_n=0 +- checkpoint.local.keep_last_n=1 +- checkpoint.save_best_model=false +- checkpoint.save_pt_file=false +- trainer.is_resume=false +- trainer.pretrained_checkpoint=/workspace/starVLA/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt +- trainer.resume_step=0 +- trainer.max_train_steps=5000 +- trainer.num_warmup_steps=100 +- trainer.save_interval=100 +- trainer.eval_interval=100 +- trainer.eval_num_batches=100 +- trainer.logging_frequency=1 +- trainer.gradient_accumulation_steps=4 +- trainer.distributed_backend=none +- trainer.learning_rate.base=2e-05 +- trainer.learning_rate.qwen_vl_interface=1e-05 +- trainer.learning_rate.action_model=0.0001 +- trainer.lr_scheduler_type=cosine_with_min_lr +- trainer.scheduler_specific_kwargs.min_lr=1e-06 +- trainer.loss_scale.vla=1.0 +- trainer.loss_scale.vlm=0.1 +- trainer.max_grad_norm=1.0 +- trainer.weight_decay=0.0 +- trainer.gradient_clipping=1.0 +- trainer.optimizer.name=AdamW +- trainer.optimizer.betas=[0.9,0.95] +- trainer.optimizer.eps=1e-08 +- trainer.optimizer.weight_decay=1e-08 +- trainer.optimizer.fused=false +- trainer.save_format=pt +- framework.name=QwenOFT +- framework.qwenvl.attn_implementation=flash_attention_2 +- framework.qwenvl.enable_gradient_checkpointing=true +- framework.action_model.action_dim=7 +- framework.action_model.action_env_dim=2 +- framework.action_model.state_dim=7 +- framework.action_model.loss_type=discrete_ce +- framework.action_model.action_horizon=1 +- framework.action_model.future_action_window_size=0 +- framework.action_model.past_action_window_size=0 +- datasets.vla_data.include_state=true +- datasets.vla_data.action_type=discrete +- datasets.vla_data.sequential_step_sampling=true +- datasets.vla_data.shuffle=true +- datasets.vla_data.action_balance.enabled=false +- datasets.vla_data.action_balance.strategy=balanced_epoch +- datasets.vla_data.action_balance.action_key=action_id +- datasets.vla_data.action_balance.target_flap_fraction=0.3 +- datasets.vla_data.action_balance.noop_id=0 +- datasets.vla_data.action_balance.flap_id=1 +- datasets.vla_data.latency_curriculum.enabled=true +- datasets.vla_data.latency_curriculum.strategy=cumulative +- datasets.vla_data.latency_curriculum.latencies=[0,1,2,3,4,5] +- datasets.vla_data.per_device_batch_size=32 +- datasets.vla_data.load_all_data_for_training=true +- datasets.vla_data.video_backend=torchvision_av +- datasets.vla_data.data_root_dir=/workspace/starVLA/data/flappy_mixed_latency_40ep_per_lat +- datasets.vla_data.data_mix=flappy_mixed_latency_train__bridge +- datasets.vla_data.eval_data_mix=flappy_mixed_latency_train__bridge__val +- framework.qwenvl.base_vlm=/workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct +- rl_games.task=flappy +- rl_games.model_alias=openvla +- rl_games.initialization_mode=bridge +- rl_games.action_carrier=bridge +- rl_games.env_eval.distributed_mode=none +- rl_games.env_eval.latency.mode=mixed +- rl_games.env_eval.frameskip=1 +- rl_games.env_eval.image_size=224 +- rl_games.env_eval.seed=42 +- rl_games.env_eval.fixed_episode_seeds=true +- rl_games.env_eval.latency_seed_stride=0 +- rl_games.env_eval.task_seed_stride=0 +- rl_games.env_eval.latency.values=[0,1,2,3,4,5] +- rl_games.env_eval.latency.prompt_map_path=/workspace/starVLA/data/flappy_mixed_latency_40ep_per_lat/flappy_mixed_latency_train__bridge/latency_prompt_map.json +- rl_games.env_eval.mid_train.enabled=false +- rl_games.env_eval.mid_train.interval_steps=100 +- rl_games.env_eval.mid_train.num_episodes=5 +- rl_games.env_eval.mid_train.max_steps_per_episode=3600 +- rl_games.env_eval.mid_train.latencies=[0,1,2,3,4,5] +- rl_games.env_eval.post_train.enabled=false +- rl_games.env_eval.post_train.num_episodes=5 +- rl_games.env_eval.post_train.max_steps_per_episode=3600 +- rl_games.env_eval.post_train.latencies=[0,1,2,3,4] diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_curriculum_cumulative_40ep_per_latency/hydra/train_starvla_hydra.log b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_curriculum_cumulative_40ep_per_latency/hydra/train_starvla_hydra.log new file mode 100644 index 0000000000000000000000000000000000000000..5836f7fad845853f83d1555cb0c4aa31371a266d --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_curriculum_cumulative_40ep_per_latency/hydra/train_starvla_hydra.log @@ -0,0 +1,5758 @@ +[2026-06-19 09:48:54,970][starVLA.training.train_starvla][INFO] - VLA Training :: Warming Up +[2026-06-19 09:48:54,980][starVLA.training.train_starvla][INFO] - ✅ Configuration wrapped for access tracking +[2026-06-19 09:49:00,371][starVLA.training.train_starvla][INFO] - Preloaded model checkpoint before Accelerator/DeepSpeed initialization: /workspace/starVLA/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt +[2026-06-19 09:49:00,397][starVLA.training.train_starvla][INFO] - Creating VLA Dataset with Mixture `flappy_mixed_latency_train__bridge` +[2026-06-19 09:49:00,930][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Behavior (data_config): ['BEHAVIOR_challenge'] +[2026-06-19 09:49:00,930][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from DOMINO: ['robotwin'] +[2026-06-19 09:49:00,930][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from DOMINO (data_config): [] +[2026-06-19 09:49:00,930][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from DOMINO (data_config): ['domino', 'domino_clean', 'domino_random', 'domino_cotrain'] +[2026-06-19 09:49:00,930][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Franka: ['custom_robot_config', 'demo_sim_franka_delta_joints', 'SO101'] +[2026-06-19 09:49:00,930][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Franka (data_config): [] +[2026-06-19 09:49:00,930][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Franka (data_config): ['custom_dataset', 'custom_dataset_2', 'demo_sim_pick_place', 'SO101_pick'] +[2026-06-19 09:49:00,930][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from LIBERO: ['libero_franka'] +[2026-06-19 09:49:00,930][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from LIBERO (data_config): [] +[2026-06-19 09:49:00,930][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from LIBERO (data_config): ['libero_all', 'libero_goal', 'multi_robot'] +[2026-06-19 09:49:00,930][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from RoboChallenge_table30v2: ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-19 09:49:00,930][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from RoboChallenge_table30v2 (data_config): ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-19 09:49:00,931][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from RoboChallenge_table30v2 (data_config): ['robochallenge_table30v2_shred_paper', 'robochallenge_table30v2_ur5_all', 'robochallenge_table30v2_arx5_all', 'robochallenge_table30v2_dosw1_all'] +[2026-06-19 09:49:00,931][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_365: ['panda_omron_robocasa365'] +[2026-06-19 09:49:00,931][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_365 (data_config): [] +[2026-06-19 09:49:00,931][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_365 (data_config): ['robocasa365_open_drawer_target_human', 'robocasa365_atomic_target_human_all', 'robocasa365_composite_target_human_all', 'robocasa365_target_human_all'] +[2026-06-19 09:49:00,931][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_tabletop: ['fourier_gr1_arms_waist'] +[2026-06-19 09:49:00,931][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_tabletop (data_config): [] +[2026-06-19 09:49:00,931][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_tabletop (data_config): ['fourier_gr1_unified_1000'] +[2026-06-19 09:49:00,931][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robotwin: ['robotwin', 'robotwin50', 'arx_x5'] +[2026-06-19 09:49:00,931][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robotwin (data_config): [] +[2026-06-19 09:49:00,931][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robotwin (data_config): ['robotwin_all', 'robotwin_all_50', 'robotwin', 'robotwin_task1', 'robotwin_task2', 'arx_x5'] +[2026-06-19 09:49:00,931][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from SimplerEnv: ['oxe_droid', 'oxe_bridge', 'oxe_rt1'] +[2026-06-19 09:49:00,931][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from SimplerEnv (data_config): [] +[2026-06-19 09:49:00,931][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from SimplerEnv (data_config): ['bridge', 'bridge_rt_1'] +[2026-06-19 09:49:00,931][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from VLA-Arena: ['vla_arena_franka'] +[2026-06-19 09:49:00,931][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from VLA-Arena (data_config): [] +[2026-06-19 09:49:00,931][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from VLA-Arena (data_config): ['vla_arena_L0_S', 'vla_arena_L0_M', 'vla_arena_L0_L'] +[2026-06-19 09:49:00,931][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from rl_games: ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-19 09:49:00,931][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from rl_games (data_config): ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-19 09:49:00,931][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from rl_games (data_config): ['flappy_train', 'flappy_train__bridge', 'flappy_mixed_latency_train', 'flappy_mixed_latency_train__bridge', 'demon_attack_train', 'demon_attack_train__bridge', 'demon_attack_mixed_latency_train', 'demon_attack_mixed_latency_train__bridge', 'deadly_corridor_train', 'deadly_corridor_train__bridge', 'deadly_corridor_mixed_latency_train', 'deadly_corridor_mixed_latency_train__bridge'] +[2026-06-19 09:49:00,982][starVLA.training.train_starvla][INFO] - Creating VLA Eval Dataset with Mixture `flappy_mixed_latency_train__bridge__val` +[2026-06-19 09:49:01,029][starVLA.training.train_starvla][INFO] - 📝 Full config saved at /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/config.full.yaml +[2026-06-19 09:49:01,035][starVLA.training.train_starvla][INFO] - 📊 Accessed config snapshot saved at /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/config.yaml +[2026-06-19 09:49:01,035][starVLA.training.train_starvla][INFO] - No pretrained checkpoint provided. Starting training from scratch. +[2026-06-19 09:49:03,622][starVLA.training.train_starvla][INFO] - ***** Training Configuration ***** +[2026-06-19 09:49:03,622][starVLA.training.train_starvla][INFO] - Total optimization steps = 5000 +[2026-06-19 09:49:03,622][starVLA.training.train_starvla][INFO] - Per device batch size = 32 +[2026-06-19 09:49:03,622][starVLA.training.train_starvla][INFO] - Gradient accumulation steps = 4 +[2026-06-19 09:49:03,622][starVLA.training.train_starvla][INFO] - Total batch size = 128 +[2026-06-19 09:49:03,818][starVLA.training.train_starvla][INFO] - Latency curriculum phase 0 at step 0: active_latencies=[0], dataset_size=62322 +[2026-06-19 09:49:19,763][starVLA.training.train_starvla][INFO] - Step 1, Loss: {'train/loss': 0.6944701075553894, 'train/grad_norm_pre_clip': 7.275081634521484, 'timing/data': 0.005204409011639655, 'timing/model': 0.903549375012517, 'learning_rate/qwen_vl_interface': 1.0000000000000001e-07, 'learning_rate/action_model': 1.0000000000000002e-06, 'epoch': 0.0}) +[2026-06-19 09:49:23,185][starVLA.training.train_starvla][INFO] - Step 2, Loss: {'train/loss': 0.695019543170929, 'train/grad_norm_pre_clip': 8.896716117858887, 'timing/data': 0.006866144016385078, 'timing/model': 0.8874795190058649, 'learning_rate/qwen_vl_interface': 2.0000000000000002e-07, 'learning_rate/action_model': 2.0000000000000003e-06, 'epoch': 0.0}) +[2026-06-19 09:49:26,594][starVLA.training.train_starvla][INFO] - Step 3, Loss: {'train/loss': 0.6884669661521912, 'train/grad_norm_pre_clip': 5.439120292663574, 'timing/data': 0.00507265399210155, 'timing/model': 0.8872074970277026, 'learning_rate/qwen_vl_interface': 3.0000000000000004e-07, 'learning_rate/action_model': 3e-06, 'epoch': 0.01}) +[2026-06-19 09:49:30,008][starVLA.training.train_starvla][INFO] - Step 4, Loss: {'train/loss': 0.6848503947257996, 'train/grad_norm_pre_clip': 2.711014747619629, 'timing/data': 0.005985297961160541, 'timing/model': 0.8923496690113097, 'learning_rate/qwen_vl_interface': 4.0000000000000003e-07, 'learning_rate/action_model': 4.000000000000001e-06, 'epoch': 0.01}) +[2026-06-19 09:49:33,425][starVLA.training.train_starvla][INFO] - Step 5, Loss: {'train/loss': 0.6310445070266724, 'train/grad_norm_pre_clip': 9.425329208374023, 'timing/data': 0.005402190960012376, 'timing/model': 0.8961127999937162, 'learning_rate/qwen_vl_interface': 5.000000000000001e-07, 'learning_rate/action_model': 5e-06, 'epoch': 0.01}) +[2026-06-19 09:49:36,839][starVLA.training.train_starvla][INFO] - Step 6, Loss: {'train/loss': 0.6356132626533508, 'train/grad_norm_pre_clip': 4.688987731933594, 'timing/data': 0.006110499030910432, 'timing/model': 0.8963408219860867, 'learning_rate/qwen_vl_interface': 6.000000000000001e-07, 'learning_rate/action_model': 6e-06, 'epoch': 0.01}) +[2026-06-19 09:49:40,255][starVLA.training.train_starvla][INFO] - Step 7, Loss: {'train/loss': 0.6128319501876831, 'train/grad_norm_pre_clip': 3.8045051097869873, 'timing/data': 0.00520487199537456, 'timing/model': 0.8914563929429278, 'learning_rate/qwen_vl_interface': 7.000000000000001e-07, 'learning_rate/action_model': 7.000000000000001e-06, 'epoch': 0.01}) +[2026-06-19 09:49:43,794][starVLA.training.train_starvla][INFO] - Step 8, Loss: {'train/loss': 0.4996437430381775, 'train/grad_norm_pre_clip': 6.843039512634277, 'timing/data': 0.005441900924779475, 'timing/model': 1.0192958959378302, 'learning_rate/qwen_vl_interface': 8.000000000000001e-07, 'learning_rate/action_model': 8.000000000000001e-06, 'epoch': 0.02}) +[2026-06-19 09:49:47,209][starVLA.training.train_starvla][INFO] - Step 9, Loss: {'train/loss': 0.47738903760910034, 'train/grad_norm_pre_clip': 4.9253315925598145, 'timing/data': 0.005502865999005735, 'timing/model': 0.8957376120379195, 'learning_rate/qwen_vl_interface': 9.000000000000001e-07, 'learning_rate/action_model': 9e-06, 'epoch': 0.02}) +[2026-06-19 09:49:50,633][starVLA.training.train_starvla][INFO] - Step 10, Loss: {'train/loss': 0.3782957196235657, 'train/grad_norm_pre_clip': 5.369585037231445, 'timing/data': 0.00537220505066216, 'timing/model': 0.895377159002237, 'learning_rate/qwen_vl_interface': 1.0000000000000002e-06, 'learning_rate/action_model': 1e-05, 'epoch': 0.02}) +[2026-06-19 09:49:54,060][starVLA.training.train_starvla][INFO] - Step 11, Loss: {'train/loss': 0.6051784753799438, 'train/grad_norm_pre_clip': 1.474735975265503, 'timing/data': 0.0057997399708256125, 'timing/model': 0.8975600550184026, 'learning_rate/qwen_vl_interface': 1.1e-06, 'learning_rate/action_model': 1.1000000000000001e-05, 'epoch': 0.02}) +[2026-06-19 09:49:57,507][starVLA.training.train_starvla][INFO] - Step 12, Loss: {'train/loss': 0.5745930075645447, 'train/grad_norm_pre_clip': 1.4881997108459473, 'timing/data': 0.005015961010940373, 'timing/model': 0.9067769770044833, 'learning_rate/qwen_vl_interface': 1.2000000000000002e-06, 'learning_rate/action_model': 1.2e-05, 'epoch': 0.02}) +[2026-06-19 09:50:00,951][starVLA.training.train_starvla][INFO] - Step 13, Loss: {'train/loss': 0.5816644430160522, 'train/grad_norm_pre_clip': 2.0270116329193115, 'timing/data': 0.005381409078836441, 'timing/model': 0.9049094730289653, 'learning_rate/qwen_vl_interface': 1.3e-06, 'learning_rate/action_model': 1.3000000000000001e-05, 'epoch': 0.03}) +[2026-06-19 09:50:04,406][starVLA.training.train_starvla][INFO] - Step 14, Loss: {'train/loss': 0.43412595987319946, 'train/grad_norm_pre_clip': 0.21987493336200714, 'timing/data': 0.005323635064996779, 'timing/model': 0.9089396289782599, 'learning_rate/qwen_vl_interface': 1.4000000000000001e-06, 'learning_rate/action_model': 1.4000000000000001e-05, 'epoch': 0.03}) +[2026-06-19 09:50:07,841][starVLA.training.train_starvla][INFO] - Step 15, Loss: {'train/loss': 0.43517011404037476, 'train/grad_norm_pre_clip': 0.2503543198108673, 'timing/data': 0.005434992024675012, 'timing/model': 0.899528287933208, 'learning_rate/qwen_vl_interface': 1.5e-06, 'learning_rate/action_model': 1.5e-05, 'epoch': 0.03}) +[2026-06-19 09:50:11,280][starVLA.training.train_starvla][INFO] - Step 16, Loss: {'train/loss': 0.5478962659835815, 'train/grad_norm_pre_clip': 2.057711601257324, 'timing/data': 0.005633449996821582, 'timing/model': 0.8994418510701507, 'learning_rate/qwen_vl_interface': 1.6000000000000001e-06, 'learning_rate/action_model': 1.6000000000000003e-05, 'epoch': 0.03}) +[2026-06-19 09:50:14,724][starVLA.training.train_starvla][INFO] - Step 17, Loss: {'train/loss': 0.5497311353683472, 'train/grad_norm_pre_clip': 2.094496965408325, 'timing/data': 0.005571813904680312, 'timing/model': 0.8983348889742047, 'learning_rate/qwen_vl_interface': 1.7000000000000002e-06, 'learning_rate/action_model': 1.7000000000000003e-05, 'epoch': 0.03}) +[2026-06-19 09:50:18,178][starVLA.training.train_starvla][INFO] - Step 18, Loss: {'train/loss': 0.43365904688835144, 'train/grad_norm_pre_clip': 0.38009175658226013, 'timing/data': 0.005407441989518702, 'timing/model': 0.9055233090184629, 'learning_rate/qwen_vl_interface': 1.8000000000000001e-06, 'learning_rate/action_model': 1.8e-05, 'epoch': 0.04}) +[2026-06-19 09:50:21,638][starVLA.training.train_starvla][INFO] - Step 19, Loss: {'train/loss': 0.37777814269065857, 'train/grad_norm_pre_clip': 0.8787947297096252, 'timing/data': 0.005611800006590784, 'timing/model': 0.9052291070111096, 'learning_rate/qwen_vl_interface': 1.9000000000000002e-06, 'learning_rate/action_model': 1.9e-05, 'epoch': 0.04}) +[2026-06-19 09:50:25,099][starVLA.training.train_starvla][INFO] - Step 20, Loss: {'train/loss': 0.4327293932437897, 'train/grad_norm_pre_clip': 0.19310811161994934, 'timing/data': 0.005389319965615869, 'timing/model': 0.9008790210355073, 'learning_rate/qwen_vl_interface': 2.0000000000000003e-06, 'learning_rate/action_model': 2e-05, 'epoch': 0.04}) +[2026-06-19 09:50:28,555][starVLA.training.train_starvla][INFO] - Step 21, Loss: {'train/loss': 0.43239298462867737, 'train/grad_norm_pre_clip': 0.13166004419326782, 'timing/data': 0.004900477011688054, 'timing/model': 0.9070339449681342, 'learning_rate/qwen_vl_interface': 2.1000000000000002e-06, 'learning_rate/action_model': 2.1e-05, 'epoch': 0.04}) +[2026-06-19 09:50:32,018][starVLA.training.train_starvla][INFO] - Step 22, Loss: {'train/loss': 0.5894824266433716, 'train/grad_norm_pre_clip': 2.366154909133911, 'timing/data': 0.0063303010538220406, 'timing/model': 0.9017830129014328, 'learning_rate/qwen_vl_interface': 2.2e-06, 'learning_rate/action_model': 2.2000000000000003e-05, 'epoch': 0.05}) +[2026-06-19 09:50:35,475][starVLA.training.train_starvla][INFO] - Step 23, Loss: {'train/loss': 0.33648818731307983, 'train/grad_norm_pre_clip': 2.101897954940796, 'timing/data': 0.0056555019691586494, 'timing/model': 0.9070828559342772, 'learning_rate/qwen_vl_interface': 2.3000000000000004e-06, 'learning_rate/action_model': 2.3000000000000003e-05, 'epoch': 0.05}) +[2026-06-19 09:50:38,942][starVLA.training.train_starvla][INFO] - Step 24, Loss: {'train/loss': 0.33679476380348206, 'train/grad_norm_pre_clip': 2.1430232524871826, 'timing/data': 0.005827545071952045, 'timing/model': 0.9052050169557333, 'learning_rate/qwen_vl_interface': 2.4000000000000003e-06, 'learning_rate/action_model': 2.4e-05, 'epoch': 0.05}) +[2026-06-19 09:50:42,400][starVLA.training.train_starvla][INFO] - Step 25, Loss: {'train/loss': 0.5819238424301147, 'train/grad_norm_pre_clip': 2.2656660079956055, 'timing/data': 0.00559443305246532, 'timing/model': 0.9048151939641684, 'learning_rate/qwen_vl_interface': 2.5e-06, 'learning_rate/action_model': 2.5e-05, 'epoch': 0.05}) +[2026-06-19 09:50:45,864][starVLA.training.train_starvla][INFO] - Step 26, Loss: {'train/loss': 0.431635320186615, 'train/grad_norm_pre_clip': 1.117140769958496, 'timing/data': 0.005569037050008774, 'timing/model': 0.9041582880308852, 'learning_rate/qwen_vl_interface': 2.6e-06, 'learning_rate/action_model': 2.6000000000000002e-05, 'epoch': 0.05}) +[2026-06-19 09:50:49,325][starVLA.training.train_starvla][INFO] - Step 27, Loss: {'train/loss': 0.5314164161682129, 'train/grad_norm_pre_clip': 1.408348798751831, 'timing/data': 0.005530990078113973, 'timing/model': 0.9042735520051792, 'learning_rate/qwen_vl_interface': 2.7000000000000004e-06, 'learning_rate/action_model': 2.7000000000000002e-05, 'epoch': 0.06}) +[2026-06-19 09:50:52,799][starVLA.training.train_starvla][INFO] - Step 28, Loss: {'train/loss': 0.48058509826660156, 'train/grad_norm_pre_clip': 0.42126548290252686, 'timing/data': 0.006349284085445106, 'timing/model': 0.9133947109803557, 'learning_rate/qwen_vl_interface': 2.8000000000000003e-06, 'learning_rate/action_model': 2.8000000000000003e-05, 'epoch': 0.06}) +[2026-06-19 09:50:56,262][starVLA.training.train_starvla][INFO] - Step 29, Loss: {'train/loss': 0.3463231921195984, 'train/grad_norm_pre_clip': 2.707091808319092, 'timing/data': 0.005705837975256145, 'timing/model': 0.9058105100411922, 'learning_rate/qwen_vl_interface': 2.9e-06, 'learning_rate/action_model': 2.9e-05, 'epoch': 0.06}) +[2026-06-19 09:50:59,725][starVLA.training.train_starvla][INFO] - Step 30, Loss: {'train/loss': 0.3455757796764374, 'train/grad_norm_pre_clip': 2.7131965160369873, 'timing/data': 0.005340202944353223, 'timing/model': 0.9042364740744233, 'learning_rate/qwen_vl_interface': 3e-06, 'learning_rate/action_model': 3e-05, 'epoch': 0.06}) +[2026-06-19 09:51:03,202][starVLA.training.train_starvla][INFO] - Step 31, Loss: {'train/loss': 0.5732350945472717, 'train/grad_norm_pre_clip': 2.107424259185791, 'timing/data': 0.0057289969408884645, 'timing/model': 0.9068767509888858, 'learning_rate/qwen_vl_interface': 3.1000000000000004e-06, 'learning_rate/action_model': 3.1e-05, 'epoch': 0.06}) +[2026-06-19 09:51:06,671][starVLA.training.train_starvla][INFO] - Step 32, Loss: {'train/loss': 0.6726718544960022, 'train/grad_norm_pre_clip': 3.8455448150634766, 'timing/data': 0.005934752058237791, 'timing/model': 0.9059047499904409, 'learning_rate/qwen_vl_interface': 3.2000000000000003e-06, 'learning_rate/action_model': 3.2000000000000005e-05, 'epoch': 0.07}) +[2026-06-19 09:51:10,142][starVLA.training.train_starvla][INFO] - Step 33, Loss: {'train/loss': 0.42980092763900757, 'train/grad_norm_pre_clip': 1.1272363662719727, 'timing/data': 0.005795857985503972, 'timing/model': 0.9079766970826313, 'learning_rate/qwen_vl_interface': 3.3000000000000006e-06, 'learning_rate/action_model': 3.3e-05, 'epoch': 0.07}) +[2026-06-19 09:51:13,620][starVLA.training.train_starvla][INFO] - Step 34, Loss: {'train/loss': 0.4793846011161804, 'train/grad_norm_pre_clip': 0.692075788974762, 'timing/data': 0.005599495954811573, 'timing/model': 0.9083579659927636, 'learning_rate/qwen_vl_interface': 3.4000000000000005e-06, 'learning_rate/action_model': 3.4000000000000007e-05, 'epoch': 0.07}) +[2026-06-19 09:51:17,221][starVLA.training.train_starvla][INFO] - Step 35, Loss: {'train/loss': 0.5600757002830505, 'train/grad_norm_pre_clip': 1.8725627660751343, 'timing/data': 0.0053635480580851436, 'timing/model': 0.9060283950529993, 'learning_rate/qwen_vl_interface': 3.5e-06, 'learning_rate/action_model': 3.5e-05, 'epoch': 0.07}) +[2026-06-19 09:51:20,702][starVLA.training.train_starvla][INFO] - Step 36, Loss: {'train/loss': 0.4765932261943817, 'train/grad_norm_pre_clip': 0.682374119758606, 'timing/data': 0.005536591983400285, 'timing/model': 0.9139190170681104, 'learning_rate/qwen_vl_interface': 3.6000000000000003e-06, 'learning_rate/action_model': 3.6e-05, 'epoch': 0.07}) +[2026-06-19 09:51:24,164][starVLA.training.train_starvla][INFO] - Step 37, Loss: {'train/loss': 0.5132882595062256, 'train/grad_norm_pre_clip': 0.8849804997444153, 'timing/data': 0.0052815469680354, 'timing/model': 0.9055283019552007, 'learning_rate/qwen_vl_interface': 3.7e-06, 'learning_rate/action_model': 3.7e-05, 'epoch': 0.08}) +[2026-06-19 09:51:27,644][starVLA.training.train_starvla][INFO] - Step 38, Loss: {'train/loss': 0.5435429215431213, 'train/grad_norm_pre_clip': 2.838461399078369, 'timing/data': 0.005822046077810228, 'timing/model': 0.9062130820238963, 'learning_rate/qwen_vl_interface': 3.8000000000000005e-06, 'learning_rate/action_model': 3.8e-05, 'epoch': 0.08}) +[2026-06-19 09:51:31,118][starVLA.training.train_starvla][INFO] - Step 39, Loss: {'train/loss': 0.45238983631134033, 'train/grad_norm_pre_clip': 3.8687305450439453, 'timing/data': 0.005776761914603412, 'timing/model': 0.9091687350301072, 'learning_rate/qwen_vl_interface': 3.900000000000001e-06, 'learning_rate/action_model': 3.9000000000000006e-05, 'epoch': 0.08}) +[2026-06-19 09:51:34,591][starVLA.training.train_starvla][INFO] - Step 40, Loss: {'train/loss': 0.47320374846458435, 'train/grad_norm_pre_clip': 2.4110612869262695, 'timing/data': 0.005973858991637826, 'timing/model': 0.908278348040767, 'learning_rate/qwen_vl_interface': 4.000000000000001e-06, 'learning_rate/action_model': 4e-05, 'epoch': 0.08}) +[2026-06-19 09:51:38,061][starVLA.training.train_starvla][INFO] - Step 41, Loss: {'train/loss': 0.5639286637306213, 'train/grad_norm_pre_clip': 6.9409565925598145, 'timing/data': 0.005928111029788852, 'timing/model': 0.9079993750201538, 'learning_rate/qwen_vl_interface': 4.1e-06, 'learning_rate/action_model': 4.1e-05, 'epoch': 0.08}) +[2026-06-19 09:51:41,552][starVLA.training.train_starvla][INFO] - Step 42, Loss: {'train/loss': 0.434324711561203, 'train/grad_norm_pre_clip': 2.1418089866638184, 'timing/data': 0.006048155017197132, 'timing/model': 0.9067326829535887, 'learning_rate/qwen_vl_interface': 4.2000000000000004e-06, 'learning_rate/action_model': 4.2e-05, 'epoch': 0.09}) +[2026-06-19 09:51:45,032][starVLA.training.train_starvla][INFO] - Step 43, Loss: {'train/loss': 0.35008135437965393, 'train/grad_norm_pre_clip': 3.6593704223632812, 'timing/data': 0.006083045969717205, 'timing/model': 0.9050936329877004, 'learning_rate/qwen_vl_interface': 4.3e-06, 'learning_rate/action_model': 4.3e-05, 'epoch': 0.09}) +[2026-06-19 09:51:48,497][starVLA.training.train_starvla][INFO] - Step 44, Loss: {'train/loss': 0.5492233633995056, 'train/grad_norm_pre_clip': 7.5922722816467285, 'timing/data': 0.006614184007048607, 'timing/model': 0.9057958379853517, 'learning_rate/qwen_vl_interface': 4.4e-06, 'learning_rate/action_model': 4.4000000000000006e-05, 'epoch': 0.09}) +[2026-06-19 09:51:51,979][starVLA.training.train_starvla][INFO] - Step 45, Loss: {'train/loss': 0.31377995014190674, 'train/grad_norm_pre_clip': 8.027508735656738, 'timing/data': 0.00634593004360795, 'timing/model': 0.9087105669314042, 'learning_rate/qwen_vl_interface': 4.5e-06, 'learning_rate/action_model': 4.5e-05, 'epoch': 0.09}) +[2026-06-19 09:51:55,457][starVLA.training.train_starvla][INFO] - Step 46, Loss: {'train/loss': 0.5285704135894775, 'train/grad_norm_pre_clip': 6.039198398590088, 'timing/data': 0.006416552932932973, 'timing/model': 0.9081102530471981, 'learning_rate/qwen_vl_interface': 4.600000000000001e-06, 'learning_rate/action_model': 4.600000000000001e-05, 'epoch': 0.09}) +[2026-06-19 09:51:58,939][starVLA.training.train_starvla][INFO] - Step 47, Loss: {'train/loss': 0.34669506549835205, 'train/grad_norm_pre_clip': 4.484374046325684, 'timing/data': 0.006047318922355771, 'timing/model': 0.9129797350615263, 'learning_rate/qwen_vl_interface': 4.7e-06, 'learning_rate/action_model': 4.7e-05, 'epoch': 0.1}) +[2026-06-19 09:52:02,428][starVLA.training.train_starvla][INFO] - Step 48, Loss: {'train/loss': 0.23427347838878632, 'train/grad_norm_pre_clip': 6.951675891876221, 'timing/data': 0.005915151908993721, 'timing/model': 0.9059335110941902, 'learning_rate/qwen_vl_interface': 4.800000000000001e-06, 'learning_rate/action_model': 4.8e-05, 'epoch': 0.1}) +[2026-06-19 09:52:05,905][starVLA.training.train_starvla][INFO] - Step 49, Loss: {'train/loss': 0.5077255368232727, 'train/grad_norm_pre_clip': 17.38299560546875, 'timing/data': 0.00646149693056941, 'timing/model': 0.9096201140200719, 'learning_rate/qwen_vl_interface': 4.9000000000000005e-06, 'learning_rate/action_model': 4.9e-05, 'epoch': 0.1}) +[2026-06-19 09:52:09,390][starVLA.training.train_starvla][INFO] - Step 50, Loss: {'train/loss': 0.30428221821784973, 'train/grad_norm_pre_clip': 7.344770908355713, 'timing/data': 0.0059462880017235875, 'timing/model': 0.9076536670327187, 'learning_rate/qwen_vl_interface': 5e-06, 'learning_rate/action_model': 5e-05, 'epoch': 0.1}) +[2026-06-19 09:52:12,867][starVLA.training.train_starvla][INFO] - Step 51, Loss: {'train/loss': 0.5802416801452637, 'train/grad_norm_pre_clip': 17.46518898010254, 'timing/data': 0.0060947490856051445, 'timing/model': 0.9081030130619183, 'learning_rate/qwen_vl_interface': 5.1e-06, 'learning_rate/action_model': 5.1000000000000006e-05, 'epoch': 0.1}) +[2026-06-19 09:52:16,377][starVLA.training.train_starvla][INFO] - Step 52, Loss: {'train/loss': 0.5002238750457764, 'train/grad_norm_pre_clip': 8.939926147460938, 'timing/data': 0.008535998989827931, 'timing/model': 0.9174048929708079, 'learning_rate/qwen_vl_interface': 5.2e-06, 'learning_rate/action_model': 5.2000000000000004e-05, 'epoch': 0.11}) +[2026-06-19 09:52:19,858][starVLA.training.train_starvla][INFO] - Step 53, Loss: {'train/loss': 0.4522019922733307, 'train/grad_norm_pre_clip': 5.1611433029174805, 'timing/data': 0.006264619063585997, 'timing/model': 0.9061252609826624, 'learning_rate/qwen_vl_interface': 5.300000000000001e-06, 'learning_rate/action_model': 5.300000000000001e-05, 'epoch': 0.11}) +[2026-06-19 09:52:23,341][starVLA.training.train_starvla][INFO] - Step 54, Loss: {'train/loss': 0.4385750889778137, 'train/grad_norm_pre_clip': 16.33895492553711, 'timing/data': 0.006112748058512807, 'timing/model': 0.9120809659361839, 'learning_rate/qwen_vl_interface': 5.400000000000001e-06, 'learning_rate/action_model': 5.4000000000000005e-05, 'epoch': 0.11}) +[2026-06-19 09:52:26,837][starVLA.training.train_starvla][INFO] - Step 55, Loss: {'train/loss': 0.4816053807735443, 'train/grad_norm_pre_clip': 19.83543586730957, 'timing/data': 0.006074949051253498, 'timing/model': 0.9170047020306811, 'learning_rate/qwen_vl_interface': 5.500000000000001e-06, 'learning_rate/action_model': 5.500000000000001e-05, 'epoch': 0.11}) +[2026-06-19 09:52:30,319][starVLA.training.train_starvla][INFO] - Step 56, Loss: {'train/loss': 0.5297531485557556, 'train/grad_norm_pre_clip': 11.84608268737793, 'timing/data': 0.00602355704177171, 'timing/model': 0.913077863981016, 'learning_rate/qwen_vl_interface': 5.600000000000001e-06, 'learning_rate/action_model': 5.6000000000000006e-05, 'epoch': 0.12}) +[2026-06-19 09:52:33,804][starVLA.training.train_starvla][INFO] - Step 57, Loss: {'train/loss': 0.4025862514972687, 'train/grad_norm_pre_clip': 11.633943557739258, 'timing/data': 0.005839603021740913, 'timing/model': 0.9131836900487542, 'learning_rate/qwen_vl_interface': 5.7e-06, 'learning_rate/action_model': 5.6999999999999996e-05, 'epoch': 0.12}) +[2026-06-19 09:52:37,314][starVLA.training.train_starvla][INFO] - Step 58, Loss: {'train/loss': 0.6136500835418701, 'train/grad_norm_pre_clip': 14.336127281188965, 'timing/data': 0.005617366987280548, 'timing/model': 0.9104294220451266, 'learning_rate/qwen_vl_interface': 5.8e-06, 'learning_rate/action_model': 5.8e-05, 'epoch': 0.12}) +[2026-06-19 09:52:40,794][starVLA.training.train_starvla][INFO] - Step 59, Loss: {'train/loss': 0.4631809890270233, 'train/grad_norm_pre_clip': 17.81282615661621, 'timing/data': 0.006056429003365338, 'timing/model': 0.9079859369667247, 'learning_rate/qwen_vl_interface': 5.9e-06, 'learning_rate/action_model': 5.9e-05, 'epoch': 0.12}) +[2026-06-19 09:52:44,275][starVLA.training.train_starvla][INFO] - Step 60, Loss: {'train/loss': 0.43035024404525757, 'train/grad_norm_pre_clip': 13.91558837890625, 'timing/data': 0.005903616081923246, 'timing/model': 0.9115793829550967, 'learning_rate/qwen_vl_interface': 6e-06, 'learning_rate/action_model': 6e-05, 'epoch': 0.12}) +[2026-06-19 09:52:47,887][starVLA.training.train_starvla][INFO] - Step 61, Loss: {'train/loss': 0.31111806631088257, 'train/grad_norm_pre_clip': 8.361395835876465, 'timing/data': 0.005630946950986981, 'timing/model': 1.0380574020091444, 'learning_rate/qwen_vl_interface': 6.1e-06, 'learning_rate/action_model': 6.1e-05, 'epoch': 0.13}) +[2026-06-19 09:52:51,360][starVLA.training.train_starvla][INFO] - Step 62, Loss: {'train/loss': 0.5119650959968567, 'train/grad_norm_pre_clip': 17.0977783203125, 'timing/data': 0.005853337934240699, 'timing/model': 0.9075285230064765, 'learning_rate/qwen_vl_interface': 6.200000000000001e-06, 'learning_rate/action_model': 6.2e-05, 'epoch': 0.13}) +[2026-06-19 09:52:54,839][starVLA.training.train_starvla][INFO] - Step 63, Loss: {'train/loss': 0.31259942054748535, 'train/grad_norm_pre_clip': 15.067095756530762, 'timing/data': 0.005869148997589946, 'timing/model': 0.9072771890787408, 'learning_rate/qwen_vl_interface': 6.300000000000001e-06, 'learning_rate/action_model': 6.3e-05, 'epoch': 0.13}) +[2026-06-19 09:52:58,332][starVLA.training.train_starvla][INFO] - Step 64, Loss: {'train/loss': 0.3934931755065918, 'train/grad_norm_pre_clip': 11.584129333496094, 'timing/data': 0.0063273339765146375, 'timing/model': 0.9130472989054397, 'learning_rate/qwen_vl_interface': 6.4000000000000006e-06, 'learning_rate/action_model': 6.400000000000001e-05, 'epoch': 0.13}) +[2026-06-19 09:53:01,812][starVLA.training.train_starvla][INFO] - Step 65, Loss: {'train/loss': 0.37198328971862793, 'train/grad_norm_pre_clip': 7.708217620849609, 'timing/data': 0.005883643054403365, 'timing/model': 0.9116934579797089, 'learning_rate/qwen_vl_interface': 6.5000000000000004e-06, 'learning_rate/action_model': 6.500000000000001e-05, 'epoch': 0.13}) +[2026-06-19 09:53:05,290][starVLA.training.train_starvla][INFO] - Step 66, Loss: {'train/loss': 0.3784519135951996, 'train/grad_norm_pre_clip': 4.4706645011901855, 'timing/data': 0.0062933380249887705, 'timing/model': 0.9093046099878848, 'learning_rate/qwen_vl_interface': 6.600000000000001e-06, 'learning_rate/action_model': 6.6e-05, 'epoch': 0.14}) +[2026-06-19 09:53:08,790][starVLA.training.train_starvla][INFO] - Step 67, Loss: {'train/loss': 0.2453058511018753, 'train/grad_norm_pre_clip': 5.591751575469971, 'timing/data': 0.006078789010643959, 'timing/model': 0.9180507709970698, 'learning_rate/qwen_vl_interface': 6.700000000000001e-06, 'learning_rate/action_model': 6.7e-05, 'epoch': 0.14}) +[2026-06-19 09:53:12,273][starVLA.training.train_starvla][INFO] - Step 68, Loss: {'train/loss': 0.48006293177604675, 'train/grad_norm_pre_clip': 22.918970108032227, 'timing/data': 0.005887001054361463, 'timing/model': 0.9095148720080033, 'learning_rate/qwen_vl_interface': 6.800000000000001e-06, 'learning_rate/action_model': 6.800000000000001e-05, 'epoch': 0.14}) +[2026-06-19 09:53:15,760][starVLA.training.train_starvla][INFO] - Step 69, Loss: {'train/loss': 0.33527034521102905, 'train/grad_norm_pre_clip': 26.45132064819336, 'timing/data': 0.0060510660987347364, 'timing/model': 0.9105166449444368, 'learning_rate/qwen_vl_interface': 6.9e-06, 'learning_rate/action_model': 6.9e-05, 'epoch': 0.14}) +[2026-06-19 09:53:19,241][starVLA.training.train_starvla][INFO] - Step 70, Loss: {'train/loss': 0.18162964284420013, 'train/grad_norm_pre_clip': 10.974984169006348, 'timing/data': 0.006063741049729288, 'timing/model': 0.9095172859961167, 'learning_rate/qwen_vl_interface': 7e-06, 'learning_rate/action_model': 7e-05, 'epoch': 0.14}) +[2026-06-19 09:53:22,749][starVLA.training.train_starvla][INFO] - Step 71, Loss: {'train/loss': 0.2141963541507721, 'train/grad_norm_pre_clip': 11.090849876403809, 'timing/data': 0.005519666010513902, 'timing/model': 0.9209214219590649, 'learning_rate/qwen_vl_interface': 7.100000000000001e-06, 'learning_rate/action_model': 7.1e-05, 'epoch': 0.15}) +[2026-06-19 09:53:26,224][starVLA.training.train_starvla][INFO] - Step 72, Loss: {'train/loss': 0.37151411175727844, 'train/grad_norm_pre_clip': 19.856361389160156, 'timing/data': 0.00577408995013684, 'timing/model': 0.9073968939483166, 'learning_rate/qwen_vl_interface': 7.2000000000000005e-06, 'learning_rate/action_model': 7.2e-05, 'epoch': 0.15}) +[2026-06-19 09:53:29,718][starVLA.training.train_starvla][INFO] - Step 73, Loss: {'train/loss': 0.5091533660888672, 'train/grad_norm_pre_clip': 47.7156867980957, 'timing/data': 0.005816344055347145, 'timing/model': 0.9095485489815474, 'learning_rate/qwen_vl_interface': 7.3e-06, 'learning_rate/action_model': 7.3e-05, 'epoch': 0.15}) +[2026-06-19 09:53:33,217][starVLA.training.train_starvla][INFO] - Step 74, Loss: {'train/loss': 0.4466147720813751, 'train/grad_norm_pre_clip': 20.55555534362793, 'timing/data': 0.005458009894937277, 'timing/model': 0.9117864560103044, 'learning_rate/qwen_vl_interface': 7.4e-06, 'learning_rate/action_model': 7.4e-05, 'epoch': 0.15}) +[2026-06-19 09:53:36,712][starVLA.training.train_starvla][INFO] - Step 75, Loss: {'train/loss': 0.37733781337738037, 'train/grad_norm_pre_clip': 12.496191024780273, 'timing/data': 0.005931664956733584, 'timing/model': 0.9129390181042254, 'learning_rate/qwen_vl_interface': 7.500000000000001e-06, 'learning_rate/action_model': 7.500000000000001e-05, 'epoch': 0.15}) +[2026-06-19 09:53:40,194][starVLA.training.train_starvla][INFO] - Step 76, Loss: {'train/loss': 0.3692062795162201, 'train/grad_norm_pre_clip': 11.986197471618652, 'timing/data': 0.006384527077898383, 'timing/model': 0.9108379350509495, 'learning_rate/qwen_vl_interface': 7.600000000000001e-06, 'learning_rate/action_model': 7.6e-05, 'epoch': 0.16}) +[2026-06-19 09:53:43,688][starVLA.training.train_starvla][INFO] - Step 77, Loss: {'train/loss': 0.33958008885383606, 'train/grad_norm_pre_clip': 7.560705184936523, 'timing/data': 0.006076775025576353, 'timing/model': 0.9075430419761688, 'learning_rate/qwen_vl_interface': 7.7e-06, 'learning_rate/action_model': 7.7e-05, 'epoch': 0.16}) +[2026-06-19 09:53:47,183][starVLA.training.train_starvla][INFO] - Step 78, Loss: {'train/loss': 0.5014641880989075, 'train/grad_norm_pre_clip': 12.680343627929688, 'timing/data': 0.005732543999329209, 'timing/model': 0.9094324810430408, 'learning_rate/qwen_vl_interface': 7.800000000000002e-06, 'learning_rate/action_model': 7.800000000000001e-05, 'epoch': 0.16}) +[2026-06-19 09:53:50,665][starVLA.training.train_starvla][INFO] - Step 79, Loss: {'train/loss': 0.32460224628448486, 'train/grad_norm_pre_clip': 10.122261047363281, 'timing/data': 0.00570679095108062, 'timing/model': 0.9095189109211788, 'learning_rate/qwen_vl_interface': 7.9e-06, 'learning_rate/action_model': 7.900000000000001e-05, 'epoch': 0.16}) +[2026-06-19 09:53:54,145][starVLA.training.train_starvla][INFO] - Step 80, Loss: {'train/loss': 0.26560044288635254, 'train/grad_norm_pre_clip': 4.142768859863281, 'timing/data': 0.005613469053059816, 'timing/model': 0.9099658940685913, 'learning_rate/qwen_vl_interface': 8.000000000000001e-06, 'learning_rate/action_model': 8e-05, 'epoch': 0.16}) +[2026-06-19 09:53:57,629][starVLA.training.train_starvla][INFO] - Step 81, Loss: {'train/loss': 0.3651237487792969, 'train/grad_norm_pre_clip': 11.747939109802246, 'timing/data': 0.0056278209667652845, 'timing/model': 0.9100476440507919, 'learning_rate/qwen_vl_interface': 8.1e-06, 'learning_rate/action_model': 8.1e-05, 'epoch': 0.17}) +[2026-06-19 09:54:01,123][starVLA.training.train_starvla][INFO] - Step 82, Loss: {'train/loss': 0.31050118803977966, 'train/grad_norm_pre_clip': 11.991750717163086, 'timing/data': 0.006005432922393084, 'timing/model': 0.9131998090306297, 'learning_rate/qwen_vl_interface': 8.2e-06, 'learning_rate/action_model': 8.2e-05, 'epoch': 0.17}) +[2026-06-19 09:54:04,618][starVLA.training.train_starvla][INFO] - Step 83, Loss: {'train/loss': 0.4662417471408844, 'train/grad_norm_pre_clip': 26.588945388793945, 'timing/data': 0.006375727942213416, 'timing/model': 0.9111769739538431, 'learning_rate/qwen_vl_interface': 8.3e-06, 'learning_rate/action_model': 8.3e-05, 'epoch': 0.17}) +[2026-06-19 09:54:08,113][starVLA.training.train_starvla][INFO] - Step 84, Loss: {'train/loss': 0.5097554922103882, 'train/grad_norm_pre_clip': 30.164031982421875, 'timing/data': 0.00572779297363013, 'timing/model': 0.9196295869769529, 'learning_rate/qwen_vl_interface': 8.400000000000001e-06, 'learning_rate/action_model': 8.4e-05, 'epoch': 0.17}) +[2026-06-19 09:54:11,616][starVLA.training.train_starvla][INFO] - Step 85, Loss: {'train/loss': 0.30956146121025085, 'train/grad_norm_pre_clip': 20.865427017211914, 'timing/data': 0.0064745630370453, 'timing/model': 0.9160268750274554, 'learning_rate/qwen_vl_interface': 8.5e-06, 'learning_rate/action_model': 8.5e-05, 'epoch': 0.17}) +[2026-06-19 09:54:15,111][starVLA.training.train_starvla][INFO] - Step 86, Loss: {'train/loss': 0.7105323076248169, 'train/grad_norm_pre_clip': 49.0867805480957, 'timing/data': 0.006108941975980997, 'timing/model': 0.9109813070390373, 'learning_rate/qwen_vl_interface': 8.6e-06, 'learning_rate/action_model': 8.6e-05, 'epoch': 0.18}) +[2026-06-19 09:54:18,606][starVLA.training.train_starvla][INFO] - Step 87, Loss: {'train/loss': 0.46161210536956787, 'train/grad_norm_pre_clip': 31.771806716918945, 'timing/data': 0.005610188934952021, 'timing/model': 0.9112637790385634, 'learning_rate/qwen_vl_interface': 8.700000000000001e-06, 'learning_rate/action_model': 8.7e-05, 'epoch': 0.18}) +[2026-06-19 09:54:22,085][starVLA.training.train_starvla][INFO] - Step 88, Loss: {'train/loss': 0.5772420763969421, 'train/grad_norm_pre_clip': 47.357810974121094, 'timing/data': 0.005866685067303479, 'timing/model': 0.9072707860032097, 'learning_rate/qwen_vl_interface': 8.8e-06, 'learning_rate/action_model': 8.800000000000001e-05, 'epoch': 0.18}) +[2026-06-19 09:54:25,652][starVLA.training.train_starvla][INFO] - Step 89, Loss: {'train/loss': 0.40321266651153564, 'train/grad_norm_pre_clip': 20.655080795288086, 'timing/data': 0.0058524589985609055, 'timing/model': 0.9108877569669858, 'learning_rate/qwen_vl_interface': 8.900000000000001e-06, 'learning_rate/action_model': 8.900000000000001e-05, 'epoch': 0.18}) +[2026-06-19 09:54:29,141][starVLA.training.train_starvla][INFO] - Step 90, Loss: {'train/loss': 0.3986331820487976, 'train/grad_norm_pre_clip': 21.24953269958496, 'timing/data': 0.005517591955140233, 'timing/model': 0.9139523609774187, 'learning_rate/qwen_vl_interface': 9e-06, 'learning_rate/action_model': 9e-05, 'epoch': 0.18}) +[2026-06-19 09:54:32,624][starVLA.training.train_starvla][INFO] - Step 91, Loss: {'train/loss': 0.4371972978115082, 'train/grad_norm_pre_clip': 18.223251342773438, 'timing/data': 0.006023655994795263, 'timing/model': 0.9125477579655126, 'learning_rate/qwen_vl_interface': 9.100000000000001e-06, 'learning_rate/action_model': 9.1e-05, 'epoch': 0.19}) +[2026-06-19 09:54:36,113][starVLA.training.train_starvla][INFO] - Step 92, Loss: {'train/loss': 0.400955468416214, 'train/grad_norm_pre_clip': 21.08047103881836, 'timing/data': 0.006358593003824353, 'timing/model': 0.9135222779586911, 'learning_rate/qwen_vl_interface': 9.200000000000002e-06, 'learning_rate/action_model': 9.200000000000001e-05, 'epoch': 0.19}) +[2026-06-19 09:54:39,592][starVLA.training.train_starvla][INFO] - Step 93, Loss: {'train/loss': 0.38104429841041565, 'train/grad_norm_pre_clip': 8.500959396362305, 'timing/data': 0.005391231970861554, 'timing/model': 0.9090852970257401, 'learning_rate/qwen_vl_interface': 9.3e-06, 'learning_rate/action_model': 9.300000000000001e-05, 'epoch': 0.19}) +[2026-06-19 09:54:43,077][starVLA.training.train_starvla][INFO] - Step 94, Loss: {'train/loss': 0.2337939292192459, 'train/grad_norm_pre_clip': 11.739623069763184, 'timing/data': 0.005492941942065954, 'timing/model': 0.9137330900412053, 'learning_rate/qwen_vl_interface': 9.4e-06, 'learning_rate/action_model': 9.4e-05, 'epoch': 0.19}) +[2026-06-19 09:54:46,568][starVLA.training.train_starvla][INFO] - Step 95, Loss: {'train/loss': 0.11158688366413116, 'train/grad_norm_pre_clip': 2.360137701034546, 'timing/data': 0.00585439708083868, 'timing/model': 0.9096570260589942, 'learning_rate/qwen_vl_interface': 9.5e-06, 'learning_rate/action_model': 9.5e-05, 'epoch': 0.2}) +[2026-06-19 09:54:50,059][starVLA.training.train_starvla][INFO] - Step 96, Loss: {'train/loss': 0.44860702753067017, 'train/grad_norm_pre_clip': 15.593944549560547, 'timing/data': 0.0055748759768903255, 'timing/model': 0.9130436560371891, 'learning_rate/qwen_vl_interface': 9.600000000000001e-06, 'learning_rate/action_model': 9.6e-05, 'epoch': 0.2}) +[2026-06-19 09:54:53,552][starVLA.training.train_starvla][INFO] - Step 97, Loss: {'train/loss': 0.6346007585525513, 'train/grad_norm_pre_clip': 17.346057891845703, 'timing/data': 0.00523484090808779, 'timing/model': 0.9115602979436517, 'learning_rate/qwen_vl_interface': 9.7e-06, 'learning_rate/action_model': 9.7e-05, 'epoch': 0.2}) +[2026-06-19 09:54:57,044][starVLA.training.train_starvla][INFO] - Step 98, Loss: {'train/loss': 0.4736871123313904, 'train/grad_norm_pre_clip': 22.157411575317383, 'timing/data': 0.006283832946792245, 'timing/model': 0.9151120720198378, 'learning_rate/qwen_vl_interface': 9.800000000000001e-06, 'learning_rate/action_model': 9.8e-05, 'epoch': 0.2}) +[2026-06-19 09:55:00,535][starVLA.training.train_starvla][INFO] - Step 99, Loss: {'train/loss': 0.3247431218624115, 'train/grad_norm_pre_clip': 12.717652320861816, 'timing/data': 0.0055549280950799584, 'timing/model': 0.9126502809813246, 'learning_rate/qwen_vl_interface': 9.9e-06, 'learning_rate/action_model': 9.900000000000001e-05, 'epoch': 0.2}) +[2026-06-19 09:55:04,035][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 100 +[2026-06-19 09:57:56,853][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 100 +[2026-06-19 09:57:56,854][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 100 +[2026-06-19 09:57:57,140][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_openvla/include -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_openvla/include -fPIC -c /tmp/tmp91htnufh/test.c -o /tmp/tmp91htnufh/test.o +[2026-06-19 09:57:57,158][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat /tmp/tmp91htnufh/test.o -laio -o /tmp/tmp91htnufh/a.out +[2026-06-19 09:57:57,393][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_openvla/include -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_openvla/include -fPIC -c /tmp/tmp6l3wo6jl/test.c -o /tmp/tmp6l3wo6jl/test.o +[2026-06-19 09:57:57,405][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat /tmp/tmp6l3wo6jl/test.o -L/usr/local/cuda -L/usr/local/cuda/lib64 -lcufile -o /tmp/tmp6l3wo6jl/a.out +[2026-06-19 09:57:57,430][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_openvla/include -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_openvla/include -fPIC -c /tmp/tmpqsfw1xlu/test.c -o /tmp/tmpqsfw1xlu/test.o +[2026-06-19 09:57:57,439][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat /tmp/tmpqsfw1xlu/test.o -laio -o /tmp/tmpqsfw1xlu/a.out +[2026-06-19 10:01:06,403][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 100 +[2026-06-19 10:01:06,404][starVLA.training.train_starvla][INFO] - Step 100, Loss: {'train/loss': 0.5317091345787048, 'train/grad_norm_pre_clip': 18.20232391357422, 'eval/loss': 0.36543670296669006, 'eval/latency_0/loss': 0.26884081959724426, 'eval/latency_1/loss': 0.300489217042923, 'eval/latency_2/loss': 0.4084911048412323, 'eval/latency_3/loss': 0.40051034092903137, 'eval/latency_4/loss': 0.4437788128852844, 'eval/flappy/loss': 0.36543670296669006, 'eval/flappy/latency_0/loss': 0.26884081959724426, 'eval/flappy/latency_1/loss': 0.300489217042923, 'eval/flappy/latency_2/loss': 0.4084911048412323, 'eval/flappy/latency_3/loss': 0.40051034092903137, 'eval/flappy/latency_4/loss': 0.4437788128852844, 'eval/flap_precision': 0.0, 'eval/flap_recall': 0.0, 'eval/flap_f1': 0.0, 'eval/flap_rate_model': 0.0, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.0, 'eval/latency_0/flap_precision': 0.0, 'eval/latency_0/flap_recall': 0.0, 'eval/latency_0/flap_f1': 0.0, 'eval/latency_0/flap_rate_model': 0.0, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.0, 'timing/data': 0.005951486993581057, 'timing/model': 0.913118113996461, 'learning_rate/qwen_vl_interface': 1e-05, 'learning_rate/action_model': 0.0001, 'epoch': 0.21}) +[2026-06-19 10:01:06,404][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_100_state +[2026-06-19 10:01:06,407][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-19 10:01:09,824][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_100_state/model.safetensors +[2026-06-19 10:01:17,024][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_100_state/optimizer.bin +[2026-06-19 10:01:17,025][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_100_state/sampler.bin +[2026-06-19 10:01:17,025][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_100_state/sampler_1.bin +[2026-06-19 10:01:17,026][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_100_state/random_states_0.pkl +[2026-06-19 10:01:17,027][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 100 +[2026-06-19 10:01:17,027][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-19 10:01:17,058][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-19 10:01:21,450][starVLA.training.train_starvla][INFO] - Step 101, Loss: {'train/loss': 0.4661291837692261, 'train/grad_norm_pre_clip': 4.528980255126953, 'timing/data': 0.004796260967850685, 'timing/model': 1.2526817499892786, 'learning_rate/qwen_vl_interface': 9.999999023727212e-06, 'learning_rate/action_model': 9.99999902372721e-05, 'epoch': 0.21}) +[2026-06-19 10:01:25,827][starVLA.training.train_starvla][INFO] - Step 102, Loss: {'train/loss': 0.2894204258918762, 'train/grad_norm_pre_clip': 5.321714401245117, 'timing/data': 0.00491689401678741, 'timing/model': 1.2550095500191674, 'learning_rate/qwen_vl_interface': 9.999996094909247e-06, 'learning_rate/action_model': 9.999996094909248e-05, 'epoch': 0.21}) +[2026-06-19 10:01:30,084][starVLA.training.train_starvla][INFO] - Step 103, Loss: {'train/loss': 0.5214000344276428, 'train/grad_norm_pre_clip': 8.196742057800293, 'timing/data': 0.004902548040263355, 'timing/model': 1.1002728760940954, 'learning_rate/qwen_vl_interface': 9.999991213547314e-06, 'learning_rate/action_model': 9.999991213547313e-05, 'epoch': 0.21}) +[2026-06-19 10:01:34,504][starVLA.training.train_starvla][INFO] - Step 104, Loss: {'train/loss': 0.3794873356819153, 'train/grad_norm_pre_clip': 11.621459007263184, 'timing/data': 0.004978389944881201, 'timing/model': 1.103063976042904, 'learning_rate/qwen_vl_interface': 9.999984379643413e-06, 'learning_rate/action_model': 9.999984379643412e-05, 'epoch': 0.21}) +[2026-06-19 10:01:38,916][starVLA.training.train_starvla][INFO] - Step 105, Loss: {'train/loss': 0.22643764317035675, 'train/grad_norm_pre_clip': 6.345125198364258, 'timing/data': 0.004894098034128547, 'timing/model': 1.1014752640621737, 'learning_rate/qwen_vl_interface': 9.999975593200357e-06, 'learning_rate/action_model': 9.999975593200357e-05, 'epoch': 0.22}) +[2026-06-19 10:01:43,349][starVLA.training.train_starvla][INFO] - Step 106, Loss: {'train/loss': 0.20024840533733368, 'train/grad_norm_pre_clip': 7.158163070678711, 'timing/data': 0.004881148925051093, 'timing/model': 1.1042664159322158, 'learning_rate/qwen_vl_interface': 9.999964854221756e-06, 'learning_rate/action_model': 9.999964854221754e-05, 'epoch': 0.22}) +[2026-06-19 10:01:47,730][starVLA.training.train_starvla][INFO] - Step 107, Loss: {'train/loss': 0.2755604386329651, 'train/grad_norm_pre_clip': 4.9149651527404785, 'timing/data': 0.004654988995753229, 'timing/model': 1.1026697070337832, 'learning_rate/qwen_vl_interface': 9.999952162712027e-06, 'learning_rate/action_model': 9.999952162712026e-05, 'epoch': 0.22}) +[2026-06-19 10:01:52,148][starVLA.training.train_starvla][INFO] - Step 108, Loss: {'train/loss': 0.22579698264598846, 'train/grad_norm_pre_clip': 10.307814598083496, 'timing/data': 0.00485966494306922, 'timing/model': 1.2711620409972966, 'learning_rate/qwen_vl_interface': 9.999937518676382e-06, 'learning_rate/action_model': 9.999937518676382e-05, 'epoch': 0.22}) +[2026-06-19 10:01:56,417][starVLA.training.train_starvla][INFO] - Step 109, Loss: {'train/loss': 0.20724153518676758, 'train/grad_norm_pre_clip': 7.166304588317871, 'timing/data': 0.004566704970784485, 'timing/model': 1.1071986800525337, 'learning_rate/qwen_vl_interface': 9.999920922120846e-06, 'learning_rate/action_model': 9.999920922120846e-05, 'epoch': 0.22}) +[2026-06-19 10:02:00,878][starVLA.training.train_starvla][INFO] - Step 110, Loss: {'train/loss': 0.5883675813674927, 'train/grad_norm_pre_clip': 22.2379150390625, 'timing/data': 0.004453934961929917, 'timing/model': 1.113672923995182, 'learning_rate/qwen_vl_interface': 9.999902373052239e-06, 'learning_rate/action_model': 9.999902373052239e-05, 'epoch': 0.23}) +[2026-06-19 10:02:05,323][starVLA.training.train_starvla][INFO] - Step 111, Loss: {'train/loss': 0.48836949467658997, 'train/grad_norm_pre_clip': 14.061561584472656, 'timing/data': 0.00497693195939064, 'timing/model': 1.107589719002135, 'learning_rate/qwen_vl_interface': 9.999881871478188e-06, 'learning_rate/action_model': 9.999881871478187e-05, 'epoch': 0.23}) +[2026-06-19 10:02:09,723][starVLA.training.train_starvla][INFO] - Step 112, Loss: {'train/loss': 0.41547152400016785, 'train/grad_norm_pre_clip': 18.29488182067871, 'timing/data': 0.005041592055931687, 'timing/model': 1.1112572599668056, 'learning_rate/qwen_vl_interface': 9.999859417407115e-06, 'learning_rate/action_model': 9.999859417407115e-05, 'epoch': 0.23}) +[2026-06-19 10:02:14,153][starVLA.training.train_starvla][INFO] - Step 113, Loss: {'train/loss': 0.44983208179473877, 'train/grad_norm_pre_clip': 6.305392742156982, 'timing/data': 0.00533187307883054, 'timing/model': 1.1071012249449268, 'learning_rate/qwen_vl_interface': 9.999835010848254e-06, 'learning_rate/action_model': 9.999835010848254e-05, 'epoch': 0.23}) +[2026-06-19 10:02:18,589][starVLA.training.train_starvla][INFO] - Step 114, Loss: {'train/loss': 0.8404441475868225, 'train/grad_norm_pre_clip': 29.176883697509766, 'timing/data': 0.00477611692622304, 'timing/model': 1.2689285710221156, 'learning_rate/qwen_vl_interface': 9.999808651811639e-06, 'learning_rate/action_model': 9.999808651811639e-05, 'epoch': 0.23}) +[2026-06-19 10:02:22,861][starVLA.training.train_starvla][INFO] - Step 115, Loss: {'train/loss': 0.9336323142051697, 'train/grad_norm_pre_clip': 38.10781478881836, 'timing/data': 0.004574346938170493, 'timing/model': 1.1078938520513475, 'learning_rate/qwen_vl_interface': 9.999780340308101e-06, 'learning_rate/action_model': 9.999780340308101e-05, 'epoch': 0.24}) +[2026-06-19 10:02:27,301][starVLA.training.train_starvla][INFO] - Step 116, Loss: {'train/loss': 0.6745496988296509, 'train/grad_norm_pre_clip': 26.373760223388672, 'timing/data': 0.005062271025963128, 'timing/model': 1.1123406459810212, 'learning_rate/qwen_vl_interface': 9.999750076349282e-06, 'learning_rate/action_model': 9.999750076349282e-05, 'epoch': 0.24}) +[2026-06-19 10:02:31,761][starVLA.training.train_starvla][INFO] - Step 117, Loss: {'train/loss': 0.4947783350944519, 'train/grad_norm_pre_clip': 13.36811351776123, 'timing/data': 0.004872447112575173, 'timing/model': 1.1165212619816884, 'learning_rate/qwen_vl_interface': 9.999717859947618e-06, 'learning_rate/action_model': 9.999717859947618e-05, 'epoch': 0.24}) +[2026-06-19 10:02:36,213][starVLA.training.train_starvla][INFO] - Step 118, Loss: {'train/loss': 0.3887178301811218, 'train/grad_norm_pre_clip': 4.863559722900391, 'timing/data': 0.004959831014275551, 'timing/model': 1.1168902839999646, 'learning_rate/qwen_vl_interface': 9.999683691116356e-06, 'learning_rate/action_model': 9.999683691116355e-05, 'epoch': 0.24}) +[2026-06-19 10:02:40,640][starVLA.training.train_starvla][INFO] - Step 119, Loss: {'train/loss': 0.39185526967048645, 'train/grad_norm_pre_clip': 13.483393669128418, 'timing/data': 0.004965557949617505, 'timing/model': 1.114944247994572, 'learning_rate/qwen_vl_interface': 9.999647569869539e-06, 'learning_rate/action_model': 9.99964756986954e-05, 'epoch': 0.24}) +[2026-06-19 10:02:45,099][starVLA.training.train_starvla][INFO] - Step 120, Loss: {'train/loss': 0.4141940176486969, 'train/grad_norm_pre_clip': 11.526262283325195, 'timing/data': 0.0050487000262364745, 'timing/model': 1.277454617084004, 'learning_rate/qwen_vl_interface': 9.999609496222017e-06, 'learning_rate/action_model': 9.999609496222017e-05, 'epoch': 0.25}) +[2026-06-19 10:02:49,410][starVLA.training.train_starvla][INFO] - Step 121, Loss: {'train/loss': 0.18819628655910492, 'train/grad_norm_pre_clip': 5.0765557289123535, 'timing/data': 0.0050152069889009, 'timing/model': 1.1155401959549636, 'learning_rate/qwen_vl_interface': 9.99956947018944e-06, 'learning_rate/action_model': 9.999569470189439e-05, 'epoch': 0.25}) +[2026-06-19 10:02:53,856][starVLA.training.train_starvla][INFO] - Step 122, Loss: {'train/loss': 0.2499796450138092, 'train/grad_norm_pre_clip': 3.4993772506713867, 'timing/data': 0.00478006899356842, 'timing/model': 1.1172527029411867, 'learning_rate/qwen_vl_interface': 9.99952749178826e-06, 'learning_rate/action_model': 9.999527491788259e-05, 'epoch': 0.25}) +[2026-06-19 10:02:58,312][starVLA.training.train_starvla][INFO] - Step 123, Loss: {'train/loss': 0.23173806071281433, 'train/grad_norm_pre_clip': 6.981877326965332, 'timing/data': 0.004727901075966656, 'timing/model': 1.1185802270192653, 'learning_rate/qwen_vl_interface': 9.999483561035732e-06, 'learning_rate/action_model': 9.999483561035732e-05, 'epoch': 0.25}) +[2026-06-19 10:03:02,737][starVLA.training.train_starvla][INFO] - Step 124, Loss: {'train/loss': 0.32618623971939087, 'train/grad_norm_pre_clip': 15.30481243133545, 'timing/data': 0.004917364101856947, 'timing/model': 1.1150939429644495, 'learning_rate/qwen_vl_interface': 9.999437677949917e-06, 'learning_rate/action_model': 9.999437677949917e-05, 'epoch': 0.25}) +[2026-06-19 10:03:07,191][starVLA.training.train_starvla][INFO] - Step 125, Loss: {'train/loss': 0.43584030866622925, 'train/grad_norm_pre_clip': 21.456939697265625, 'timing/data': 0.004665990942157805, 'timing/model': 1.1133251040009782, 'learning_rate/qwen_vl_interface': 9.999389842549674e-06, 'learning_rate/action_model': 9.999389842549673e-05, 'epoch': 0.26}) +[2026-06-19 10:03:11,672][starVLA.training.train_starvla][INFO] - Step 126, Loss: {'train/loss': 0.42655128240585327, 'train/grad_norm_pre_clip': 13.480406761169434, 'timing/data': 0.005099068977870047, 'timing/model': 1.2916447700699791, 'learning_rate/qwen_vl_interface': 9.999340054854669e-06, 'learning_rate/action_model': 9.999340054854667e-05, 'epoch': 0.26}) +[2026-06-19 10:03:15,993][starVLA.training.train_starvla][INFO] - Step 127, Loss: {'train/loss': 0.5577496886253357, 'train/grad_norm_pre_clip': 25.807044982910156, 'timing/data': 0.004469288978725672, 'timing/model': 1.120012179017067, 'learning_rate/qwen_vl_interface': 9.999288314885362e-06, 'learning_rate/action_model': 9.999288314885362e-05, 'epoch': 0.26}) +[2026-06-19 10:03:20,469][starVLA.training.train_starvla][INFO] - Step 128, Loss: {'train/loss': 0.3259488046169281, 'train/grad_norm_pre_clip': 12.012213706970215, 'timing/data': 0.004877623054198921, 'timing/model': 1.118740836973302, 'learning_rate/qwen_vl_interface': 9.999234622663027e-06, 'learning_rate/action_model': 9.999234622663028e-05, 'epoch': 0.26}) +[2026-06-19 10:03:24,872][starVLA.training.train_starvla][INFO] - Step 129, Loss: {'train/loss': 0.3162407875061035, 'train/grad_norm_pre_clip': 23.38697052001953, 'timing/data': 0.004958753008395433, 'timing/model': 1.1141911499435082, 'learning_rate/qwen_vl_interface': 9.999178978209732e-06, 'learning_rate/action_model': 9.999178978209732e-05, 'epoch': 0.26}) +[2026-06-19 10:03:29,290][starVLA.training.train_starvla][INFO] - Step 130, Loss: {'train/loss': 0.5168586373329163, 'train/grad_norm_pre_clip': 10.618642807006836, 'timing/data': 0.004795752000063658, 'timing/model': 1.1134572949958965, 'learning_rate/qwen_vl_interface': 9.999121381548354e-06, 'learning_rate/action_model': 9.999121381548353e-05, 'epoch': 0.27}) +[2026-06-19 10:03:33,764][starVLA.training.train_starvla][INFO] - Step 131, Loss: {'train/loss': 0.5422360897064209, 'train/grad_norm_pre_clip': 21.33623695373535, 'timing/data': 0.004809436039067805, 'timing/model': 1.1202216999372467, 'learning_rate/qwen_vl_interface': 9.999061832702563e-06, 'learning_rate/action_model': 9.999061832702563e-05, 'epoch': 0.27}) +[2026-06-19 10:03:38,245][starVLA.training.train_starvla][INFO] - Step 132, Loss: {'train/loss': 0.4214218556880951, 'train/grad_norm_pre_clip': 23.675092697143555, 'timing/data': 0.005047912942245603, 'timing/model': 1.2892511959653348, 'learning_rate/qwen_vl_interface': 9.999000331696842e-06, 'learning_rate/action_model': 9.999000331696841e-05, 'epoch': 0.27}) +[2026-06-19 10:03:42,552][starVLA.training.train_starvla][INFO] - Step 133, Loss: {'train/loss': 0.47464966773986816, 'train/grad_norm_pre_clip': 22.051902770996094, 'timing/data': 0.005092154024168849, 'timing/model': 1.1159598090453073, 'learning_rate/qwen_vl_interface': 9.99893687855647e-06, 'learning_rate/action_model': 9.99893687855647e-05, 'epoch': 0.27}) +[2026-06-19 10:03:47,074][starVLA.training.train_starvla][INFO] - Step 134, Loss: {'train/loss': 0.5049344301223755, 'train/grad_norm_pre_clip': 12.204620361328125, 'timing/data': 0.005079459981061518, 'timing/model': 1.1176676519680768, 'learning_rate/qwen_vl_interface': 9.998871473307529e-06, 'learning_rate/action_model': 9.99887147330753e-05, 'epoch': 0.28}) +[2026-06-19 10:03:51,494][starVLA.training.train_starvla][INFO] - Step 135, Loss: {'train/loss': 0.38705554604530334, 'train/grad_norm_pre_clip': 8.564762115478516, 'timing/data': 0.004942853935062885, 'timing/model': 1.1158810750348493, 'learning_rate/qwen_vl_interface': 9.998804115976907e-06, 'learning_rate/action_model': 9.998804115976908e-05, 'epoch': 0.28}) +[2026-06-19 10:03:55,936][starVLA.training.train_starvla][INFO] - Step 136, Loss: {'train/loss': 0.28519710898399353, 'train/grad_norm_pre_clip': 3.384675979614258, 'timing/data': 0.005052530090324581, 'timing/model': 1.1179445579182357, 'learning_rate/qwen_vl_interface': 9.99873480659229e-06, 'learning_rate/action_model': 9.99873480659229e-05, 'epoch': 0.28}) +[2026-06-19 10:04:00,416][starVLA.training.train_starvla][INFO] - Step 137, Loss: {'train/loss': 0.3179330825805664, 'train/grad_norm_pre_clip': 3.747114658355713, 'timing/data': 0.005012789973989129, 'timing/model': 1.1301300559425727, 'learning_rate/qwen_vl_interface': 9.99866354518217e-06, 'learning_rate/action_model': 9.99866354518217e-05, 'epoch': 0.28}) +[2026-06-19 10:04:04,879][starVLA.training.train_starvla][INFO] - Step 138, Loss: {'train/loss': 0.2879076600074768, 'train/grad_norm_pre_clip': 2.281262159347534, 'timing/data': 0.005396003951318562, 'timing/model': 1.2805669320514426, 'learning_rate/qwen_vl_interface': 9.99859033177584e-06, 'learning_rate/action_model': 9.99859033177584e-05, 'epoch': 0.28}) +[2026-06-19 10:04:09,193][starVLA.training.train_starvla][INFO] - Step 139, Loss: {'train/loss': 0.5581611394882202, 'train/grad_norm_pre_clip': 9.49895191192627, 'timing/data': 0.007358236005529761, 'timing/model': 1.1229615280171856, 'learning_rate/qwen_vl_interface': 9.998515166403394e-06, 'learning_rate/action_model': 9.998515166403394e-05, 'epoch': 0.29}) +[2026-06-19 10:04:13,658][starVLA.training.train_starvla][INFO] - Step 140, Loss: {'train/loss': 0.4733981490135193, 'train/grad_norm_pre_clip': 6.551326751708984, 'timing/data': 0.005192909040488303, 'timing/model': 1.118404372013174, 'learning_rate/qwen_vl_interface': 9.99843804909573e-06, 'learning_rate/action_model': 9.99843804909573e-05, 'epoch': 0.29}) +[2026-06-19 10:04:18,089][starVLA.training.train_starvla][INFO] - Step 141, Loss: {'train/loss': 0.2152072638273239, 'train/grad_norm_pre_clip': 6.126567363739014, 'timing/data': 0.005231703049503267, 'timing/model': 1.1193040229845792, 'learning_rate/qwen_vl_interface': 9.998358979884548e-06, 'learning_rate/action_model': 9.998358979884547e-05, 'epoch': 0.29}) +[2026-06-19 10:04:22,576][starVLA.training.train_starvla][INFO] - Step 142, Loss: {'train/loss': 0.3639909625053406, 'train/grad_norm_pre_clip': 4.139636516571045, 'timing/data': 0.004878610954619944, 'timing/model': 1.1175260549644008, 'learning_rate/qwen_vl_interface': 9.998277958802352e-06, 'learning_rate/action_model': 9.998277958802351e-05, 'epoch': 0.29}) +[2026-06-19 10:04:27,072][starVLA.training.train_starvla][INFO] - Step 143, Loss: {'train/loss': 0.22652143239974976, 'train/grad_norm_pre_clip': 8.678366661071777, 'timing/data': 0.004685858963057399, 'timing/model': 1.1165272030048072, 'learning_rate/qwen_vl_interface': 9.998194985882442e-06, 'learning_rate/action_model': 9.998194985882442e-05, 'epoch': 0.29}) +[2026-06-19 10:04:31,572][starVLA.training.train_starvla][INFO] - Step 144, Loss: {'train/loss': 0.29775360226631165, 'train/grad_norm_pre_clip': 5.379061698913574, 'timing/data': 0.00478112103883177, 'timing/model': 1.2833599579753354, 'learning_rate/qwen_vl_interface': 9.99811006115893e-06, 'learning_rate/action_model': 9.998110061158931e-05, 'epoch': 0.3}) +[2026-06-19 10:04:35,991][starVLA.training.train_starvla][INFO] - Step 145, Loss: {'train/loss': 0.6008157730102539, 'train/grad_norm_pre_clip': 14.804922103881836, 'timing/data': 0.004841443034820259, 'timing/model': 1.2169509490486234, 'learning_rate/qwen_vl_interface': 9.998023184666726e-06, 'learning_rate/action_model': 9.998023184666726e-05, 'epoch': 0.3}) +[2026-06-19 10:04:40,301][starVLA.training.train_starvla][INFO] - Step 146, Loss: {'train/loss': 0.5638160705566406, 'train/grad_norm_pre_clip': 18.091571807861328, 'timing/data': 0.0049108139937743545, 'timing/model': 1.1142573269316927, 'learning_rate/qwen_vl_interface': 9.997934356441538e-06, 'learning_rate/action_model': 9.997934356441537e-05, 'epoch': 0.3}) +[2026-06-19 10:04:44,735][starVLA.training.train_starvla][INFO] - Step 147, Loss: {'train/loss': 0.5450429320335388, 'train/grad_norm_pre_clip': 22.349435806274414, 'timing/data': 0.004923791973851621, 'timing/model': 1.1234346280107275, 'learning_rate/qwen_vl_interface': 9.997843576519879e-06, 'learning_rate/action_model': 9.997843576519878e-05, 'epoch': 0.3}) +[2026-06-19 10:04:49,214][starVLA.training.train_starvla][INFO] - Step 148, Loss: {'train/loss': 0.2814144194126129, 'train/grad_norm_pre_clip': 3.1239125728607178, 'timing/data': 0.00476421695202589, 'timing/model': 1.1172468019649386, 'learning_rate/qwen_vl_interface': 9.99775084493907e-06, 'learning_rate/action_model': 9.997750844939069e-05, 'epoch': 0.3}) +[2026-06-19 10:04:53,686][starVLA.training.train_starvla][INFO] - Step 149, Loss: {'train/loss': 0.3361172378063202, 'train/grad_norm_pre_clip': 4.901087284088135, 'timing/data': 0.005076308036223054, 'timing/model': 1.121257861959748, 'learning_rate/qwen_vl_interface': 9.997656161737227e-06, 'learning_rate/action_model': 9.997656161737226e-05, 'epoch': 0.31}) +[2026-06-19 10:04:58,166][starVLA.training.train_starvla][INFO] - Step 150, Loss: {'train/loss': 0.2716732323169708, 'train/grad_norm_pre_clip': 8.864017486572266, 'timing/data': 0.004546741023659706, 'timing/model': 1.287207096000202, 'learning_rate/qwen_vl_interface': 9.99755952695327e-06, 'learning_rate/action_model': 9.997559526953269e-05, 'epoch': 0.31}) +[2026-06-19 10:05:02,605][starVLA.training.train_starvla][INFO] - Step 151, Loss: {'train/loss': 0.34798675775527954, 'train/grad_norm_pre_clip': 9.075765609741211, 'timing/data': 0.004900239058770239, 'timing/model': 1.2432593010598794, 'learning_rate/qwen_vl_interface': 9.99746094062692e-06, 'learning_rate/action_model': 9.99746094062692e-05, 'epoch': 0.31}) +[2026-06-19 10:05:06,921][starVLA.training.train_starvla][INFO] - Step 152, Loss: {'train/loss': 0.3256785273551941, 'train/grad_norm_pre_clip': 4.271297454833984, 'timing/data': 0.004872852936387062, 'timing/model': 1.1159325250191614, 'learning_rate/qwen_vl_interface': 9.997360402798706e-06, 'learning_rate/action_model': 9.997360402798705e-05, 'epoch': 0.31}) +[2026-06-19 10:05:11,369][starVLA.training.train_starvla][INFO] - Step 153, Loss: {'train/loss': 0.45778968930244446, 'train/grad_norm_pre_clip': 7.0720906257629395, 'timing/data': 0.004912509000860155, 'timing/model': 1.1174024129286408, 'learning_rate/qwen_vl_interface': 9.997257913509955e-06, 'learning_rate/action_model': 9.997257913509955e-05, 'epoch': 0.31}) +[2026-06-19 10:05:15,835][starVLA.training.train_starvla][INFO] - Step 154, Loss: {'train/loss': 0.4202347993850708, 'train/grad_norm_pre_clip': 12.41564655303955, 'timing/data': 0.005285336053930223, 'timing/model': 1.116365818073973, 'learning_rate/qwen_vl_interface': 9.997153472802794e-06, 'learning_rate/action_model': 9.997153472802795e-05, 'epoch': 0.32}) +[2026-06-19 10:05:20,292][starVLA.training.train_starvla][INFO] - Step 155, Loss: {'train/loss': 0.32104915380477905, 'train/grad_norm_pre_clip': 13.043535232543945, 'timing/data': 0.004965409985743463, 'timing/model': 1.1183890480315313, 'learning_rate/qwen_vl_interface': 9.997047080720157e-06, 'learning_rate/action_model': 9.997047080720158e-05, 'epoch': 0.32}) +[2026-06-19 10:05:24,766][starVLA.training.train_starvla][INFO] - Step 156, Loss: {'train/loss': 0.22770202159881592, 'train/grad_norm_pre_clip': 4.000877380371094, 'timing/data': 0.0047956620110198855, 'timing/model': 1.2855325270211324, 'learning_rate/qwen_vl_interface': 9.996938737305777e-06, 'learning_rate/action_model': 9.996938737305776e-05, 'epoch': 0.32}) +[2026-06-19 10:05:29,083][starVLA.training.train_starvla][INFO] - Step 157, Loss: {'train/loss': 0.1984463930130005, 'train/grad_norm_pre_clip': 10.387568473815918, 'timing/data': 0.004669471993111074, 'timing/model': 1.1196602089330554, 'learning_rate/qwen_vl_interface': 9.996828442604189e-06, 'learning_rate/action_model': 9.99682844260419e-05, 'epoch': 0.32}) +[2026-06-19 10:05:33,541][starVLA.training.train_starvla][INFO] - Step 158, Loss: {'train/loss': 0.21643415093421936, 'train/grad_norm_pre_clip': 7.862414360046387, 'timing/data': 0.004796015098690987, 'timing/model': 1.1130258420016617, 'learning_rate/qwen_vl_interface': 9.996716196660733e-06, 'learning_rate/action_model': 9.996716196660731e-05, 'epoch': 0.32}) +[2026-06-19 10:05:38,026][starVLA.training.train_starvla][INFO] - Step 159, Loss: {'train/loss': 0.28205516934394836, 'train/grad_norm_pre_clip': 5.270236968994141, 'timing/data': 0.00508415000513196, 'timing/model': 1.1167649449780583, 'learning_rate/qwen_vl_interface': 9.996601999521546e-06, 'learning_rate/action_model': 9.996601999521546e-05, 'epoch': 0.33}) +[2026-06-19 10:05:42,513][starVLA.training.train_starvla][INFO] - Step 160, Loss: {'train/loss': 0.4065133035182953, 'train/grad_norm_pre_clip': 16.750810623168945, 'timing/data': 0.004746056976728141, 'timing/model': 1.1156110729789361, 'learning_rate/qwen_vl_interface': 9.996485851233572e-06, 'learning_rate/action_model': 9.996485851233572e-05, 'epoch': 0.33}) +[2026-06-19 10:05:46,940][starVLA.training.train_starvla][INFO] - Step 161, Loss: {'train/loss': 0.27535727620124817, 'train/grad_norm_pre_clip': 17.174959182739258, 'timing/data': 0.005190194002352655, 'timing/model': 1.12126698193606, 'learning_rate/qwen_vl_interface': 9.996367751844556e-06, 'learning_rate/action_model': 9.996367751844555e-05, 'epoch': 0.33}) +[2026-06-19 10:05:51,364][starVLA.training.train_starvla][INFO] - Step 162, Loss: {'train/loss': 0.3304619789123535, 'train/grad_norm_pre_clip': 16.085838317871094, 'timing/data': 0.0051406529964879155, 'timing/model': 1.2416544220177457, 'learning_rate/qwen_vl_interface': 9.996247701403043e-06, 'learning_rate/action_model': 9.996247701403042e-05, 'epoch': 0.33}) +[2026-06-19 10:05:55,674][starVLA.training.train_starvla][INFO] - Step 163, Loss: {'train/loss': 0.4340633749961853, 'train/grad_norm_pre_clip': 7.720938682556152, 'timing/data': 0.0046321500558406115, 'timing/model': 1.122510546934791, 'learning_rate/qwen_vl_interface': 9.99612569995838e-06, 'learning_rate/action_model': 9.99612569995838e-05, 'epoch': 0.33}) +[2026-06-19 10:06:00,156][starVLA.training.train_starvla][INFO] - Step 164, Loss: {'train/loss': 0.9919240474700928, 'train/grad_norm_pre_clip': 29.15822410583496, 'timing/data': 0.0045380680821835995, 'timing/model': 1.1139129899675027, 'learning_rate/qwen_vl_interface': 9.99600174756072e-06, 'learning_rate/action_model': 9.99600174756072e-05, 'epoch': 0.34}) +[2026-06-19 10:06:04,618][starVLA.training.train_starvla][INFO] - Step 165, Loss: {'train/loss': 0.6405217051506042, 'train/grad_norm_pre_clip': 23.29446792602539, 'timing/data': 0.005015695001929998, 'timing/model': 1.1194626649376005, 'learning_rate/qwen_vl_interface': 9.995875844261012e-06, 'learning_rate/action_model': 9.995875844261013e-05, 'epoch': 0.34}) +[2026-06-19 10:06:09,083][starVLA.training.train_starvla][INFO] - Step 166, Loss: {'train/loss': 0.6516820192337036, 'train/grad_norm_pre_clip': 20.65144157409668, 'timing/data': 0.004869429976679385, 'timing/model': 1.1179561379831284, 'learning_rate/qwen_vl_interface': 9.995747990111015e-06, 'learning_rate/action_model': 9.995747990111013e-05, 'epoch': 0.34}) +[2026-06-19 10:06:13,513][starVLA.training.train_starvla][INFO] - Step 167, Loss: {'train/loss': 0.46295633912086487, 'train/grad_norm_pre_clip': 7.4322638511657715, 'timing/data': 0.004722543992102146, 'timing/model': 1.121212902944535, 'learning_rate/qwen_vl_interface': 9.99561818516328e-06, 'learning_rate/action_model': 9.99561818516328e-05, 'epoch': 0.34}) +[2026-06-19 10:06:17,962][starVLA.training.train_starvla][INFO] - Step 168, Loss: {'train/loss': 0.2228948473930359, 'train/grad_norm_pre_clip': 4.564469337463379, 'timing/data': 0.004899455001577735, 'timing/model': 1.2546885670162737, 'learning_rate/qwen_vl_interface': 9.995486429471166e-06, 'learning_rate/action_model': 9.995486429471166e-05, 'epoch': 0.35}) +[2026-06-19 10:06:22,280][starVLA.training.train_starvla][INFO] - Step 169, Loss: {'train/loss': 0.6316932439804077, 'train/grad_norm_pre_clip': 16.85113525390625, 'timing/data': 0.004635122022591531, 'timing/model': 1.133056538994424, 'learning_rate/qwen_vl_interface': 9.995352723088833e-06, 'learning_rate/action_model': 9.995352723088834e-05, 'epoch': 0.35}) +[2026-06-19 10:06:26,759][starVLA.training.train_starvla][INFO] - Step 170, Loss: {'train/loss': 0.4177529513835907, 'train/grad_norm_pre_clip': 7.950399875640869, 'timing/data': 0.005151559016667306, 'timing/model': 1.1208901079371572, 'learning_rate/qwen_vl_interface': 9.995217066071245e-06, 'learning_rate/action_model': 9.995217066071246e-05, 'epoch': 0.35}) +[2026-06-19 10:06:31,265][starVLA.training.train_starvla][INFO] - Step 171, Loss: {'train/loss': 0.5062970519065857, 'train/grad_norm_pre_clip': 12.73657512664795, 'timing/data': 0.0049972470151260495, 'timing/model': 1.1250393210211769, 'learning_rate/qwen_vl_interface': 9.995079458474163e-06, 'learning_rate/action_model': 9.995079458474164e-05, 'epoch': 0.35}) +[2026-06-19 10:06:35,740][starVLA.training.train_starvla][INFO] - Step 172, Loss: {'train/loss': 0.40778079628944397, 'train/grad_norm_pre_clip': 11.748313903808594, 'timing/data': 0.004327566013671458, 'timing/model': 1.11992859095335, 'learning_rate/qwen_vl_interface': 9.994939900354154e-06, 'learning_rate/action_model': 9.994939900354154e-05, 'epoch': 0.35}) +[2026-06-19 10:06:40,198][starVLA.training.train_starvla][INFO] - Step 173, Loss: {'train/loss': 0.40626606345176697, 'train/grad_norm_pre_clip': 10.5462007522583, 'timing/data': 0.004704023944213986, 'timing/model': 1.2697668430628255, 'learning_rate/qwen_vl_interface': 9.994798391768582e-06, 'learning_rate/action_model': 9.994798391768582e-05, 'epoch': 0.36}) +[2026-06-19 10:06:44,513][starVLA.training.train_starvla][INFO] - Step 174, Loss: {'train/loss': 0.4729000926017761, 'train/grad_norm_pre_clip': 4.596169948577881, 'timing/data': 0.0048022790579125285, 'timing/model': 1.1219464800087735, 'learning_rate/qwen_vl_interface': 9.99465493277562e-06, 'learning_rate/action_model': 9.994654932775618e-05, 'epoch': 0.36}) +[2026-06-19 10:06:48,997][starVLA.training.train_starvla][INFO] - Step 175, Loss: {'train/loss': 0.4484832286834717, 'train/grad_norm_pre_clip': 4.4744768142700195, 'timing/data': 0.004887346061877906, 'timing/model': 1.1174641890684143, 'learning_rate/qwen_vl_interface': 9.994509523434235e-06, 'learning_rate/action_model': 9.994509523434235e-05, 'epoch': 0.36}) +[2026-06-19 10:06:53,375][starVLA.training.train_starvla][INFO] - Step 176, Loss: {'train/loss': 0.4951256811618805, 'train/grad_norm_pre_clip': 13.151954650878906, 'timing/data': 0.005007453029975295, 'timing/model': 1.1160685670329258, 'learning_rate/qwen_vl_interface': 9.994362163804202e-06, 'learning_rate/action_model': 9.9943621638042e-05, 'epoch': 0.36}) +[2026-06-19 10:06:57,788][starVLA.training.train_starvla][INFO] - Step 177, Loss: {'train/loss': 0.46281737089157104, 'train/grad_norm_pre_clip': 12.875020980834961, 'timing/data': 0.0052296549547463655, 'timing/model': 1.114925883943215, 'learning_rate/qwen_vl_interface': 9.994212853946092e-06, 'learning_rate/action_model': 9.994212853946092e-05, 'epoch': 0.36}) +[2026-06-19 10:07:02,211][starVLA.training.train_starvla][INFO] - Step 178, Loss: {'train/loss': 0.46344488859176636, 'train/grad_norm_pre_clip': 8.455229759216309, 'timing/data': 0.005017800023779273, 'timing/model': 1.1225453439401463, 'learning_rate/qwen_vl_interface': 9.994061593921283e-06, 'learning_rate/action_model': 9.994061593921282e-05, 'epoch': 0.37}) +[2026-06-19 10:07:06,680][starVLA.training.train_starvla][INFO] - Step 179, Loss: {'train/loss': 0.4420040249824524, 'train/grad_norm_pre_clip': 7.336363792419434, 'timing/data': 0.004915676079690456, 'timing/model': 1.285515393014066, 'learning_rate/qwen_vl_interface': 9.993908383791951e-06, 'learning_rate/action_model': 9.993908383791951e-05, 'epoch': 0.37}) +[2026-06-19 10:07:10,992][starVLA.training.train_starvla][INFO] - Step 180, Loss: {'train/loss': 0.33984196186065674, 'train/grad_norm_pre_clip': 1.8176970481872559, 'timing/data': 0.00509550201240927, 'timing/model': 1.1159971179440618, 'learning_rate/qwen_vl_interface': 9.993753223621076e-06, 'learning_rate/action_model': 9.993753223621077e-05, 'epoch': 0.37}) +[2026-06-19 10:07:15,498][starVLA.training.train_starvla][INFO] - Step 181, Loss: {'train/loss': 0.1595594882965088, 'train/grad_norm_pre_clip': 1.998254656791687, 'timing/data': 0.0048688509268686175, 'timing/model': 1.124381031957455, 'learning_rate/qwen_vl_interface': 9.99359611347244e-06, 'learning_rate/action_model': 9.993596113472439e-05, 'epoch': 0.37}) +[2026-06-19 10:07:19,970][starVLA.training.train_starvla][INFO] - Step 182, Loss: {'train/loss': 0.7931908965110779, 'train/grad_norm_pre_clip': 24.2257080078125, 'timing/data': 0.0052538360469043255, 'timing/model': 1.1161474119871855, 'learning_rate/qwen_vl_interface': 9.99343705341062e-06, 'learning_rate/action_model': 9.99343705341062e-05, 'epoch': 0.37}) +[2026-06-19 10:07:24,400][starVLA.training.train_starvla][INFO] - Step 183, Loss: {'train/loss': 0.791357159614563, 'train/grad_norm_pre_clip': 25.927492141723633, 'timing/data': 0.005297339987009764, 'timing/model': 1.1265495569678023, 'learning_rate/qwen_vl_interface': 9.993276043501004e-06, 'learning_rate/action_model': 9.993276043501004e-05, 'epoch': 0.38}) +[2026-06-19 10:07:28,872][starVLA.training.train_starvla][INFO] - Step 184, Loss: {'train/loss': 0.5685124397277832, 'train/grad_norm_pre_clip': 9.921564102172852, 'timing/data': 0.005228682071901858, 'timing/model': 1.1255573179805651, 'learning_rate/qwen_vl_interface': 9.993113083809776e-06, 'learning_rate/action_model': 9.993113083809775e-05, 'epoch': 0.38}) +[2026-06-19 10:07:33,364][starVLA.training.train_starvla][INFO] - Step 185, Loss: {'train/loss': 0.8516623973846436, 'train/grad_norm_pre_clip': 14.176165580749512, 'timing/data': 0.005013010930269957, 'timing/model': 1.2998035970376804, 'learning_rate/qwen_vl_interface': 9.992948174403921e-06, 'learning_rate/action_model': 9.992948174403922e-05, 'epoch': 0.38}) +[2026-06-19 10:07:37,672][starVLA.training.train_starvla][INFO] - Step 186, Loss: {'train/loss': 0.3384264409542084, 'train/grad_norm_pre_clip': 6.954573631286621, 'timing/data': 0.0047944479156285524, 'timing/model': 1.1173175649018958, 'learning_rate/qwen_vl_interface': 9.99278131535123e-06, 'learning_rate/action_model': 9.992781315351229e-05, 'epoch': 0.38}) +[2026-06-19 10:07:42,179][starVLA.training.train_starvla][INFO] - Step 187, Loss: {'train/loss': 0.4237060546875, 'train/grad_norm_pre_clip': 6.406613826751709, 'timing/data': 0.004781105904839933, 'timing/model': 1.1252019910607487, 'learning_rate/qwen_vl_interface': 9.992612506720289e-06, 'learning_rate/action_model': 9.99261250672029e-05, 'epoch': 0.38}) +[2026-06-19 10:07:46,676][starVLA.training.train_starvla][INFO] - Step 188, Loss: {'train/loss': 0.24022382497787476, 'train/grad_norm_pre_clip': 4.277698040008545, 'timing/data': 0.005203361972235143, 'timing/model': 1.136172377038747, 'learning_rate/qwen_vl_interface': 9.992441748580492e-06, 'learning_rate/action_model': 9.99244174858049e-05, 'epoch': 0.39}) +[2026-06-19 10:07:51,137][starVLA.training.train_starvla][INFO] - Step 189, Loss: {'train/loss': 0.34041547775268555, 'train/grad_norm_pre_clip': 5.651200771331787, 'timing/data': 0.004731849068775773, 'timing/model': 1.1168038700707257, 'learning_rate/qwen_vl_interface': 9.992269041002028e-06, 'learning_rate/action_model': 9.992269041002027e-05, 'epoch': 0.39}) +[2026-06-19 10:07:55,587][starVLA.training.train_starvla][INFO] - Step 190, Loss: {'train/loss': 0.41576409339904785, 'train/grad_norm_pre_clip': 3.5695698261260986, 'timing/data': 0.005137462052516639, 'timing/model': 1.1220078819897026, 'learning_rate/qwen_vl_interface': 9.992094384055894e-06, 'learning_rate/action_model': 9.992094384055895e-05, 'epoch': 0.39}) +[2026-06-19 10:08:00,080][starVLA.training.train_starvla][INFO] - Step 191, Loss: {'train/loss': 0.30604809522628784, 'train/grad_norm_pre_clip': 4.014668941497803, 'timing/data': 0.005640254006721079, 'timing/model': 1.3064166429685429, 'learning_rate/qwen_vl_interface': 9.991917777813884e-06, 'learning_rate/action_model': 9.991917777813884e-05, 'epoch': 0.39}) +[2026-06-19 10:08:04,387][starVLA.training.train_starvla][INFO] - Step 192, Loss: {'train/loss': 0.3868506848812103, 'train/grad_norm_pre_clip': 4.072452545166016, 'timing/data': 0.004909783019684255, 'timing/model': 1.1191366199636832, 'learning_rate/qwen_vl_interface': 9.991739222348593e-06, 'learning_rate/action_model': 9.991739222348592e-05, 'epoch': 0.39}) +[2026-06-19 10:08:08,852][starVLA.training.train_starvla][INFO] - Step 193, Loss: {'train/loss': 0.20177039504051208, 'train/grad_norm_pre_clip': 3.3222293853759766, 'timing/data': 0.0048781499499455094, 'timing/model': 1.1193701869342476, 'learning_rate/qwen_vl_interface': 9.991558717733418e-06, 'learning_rate/action_model': 9.991558717733417e-05, 'epoch': 0.4}) +[2026-06-19 10:08:13,328][starVLA.training.train_starvla][INFO] - Step 194, Loss: {'train/loss': 0.20564411580562592, 'train/grad_norm_pre_clip': 2.8040599822998047, 'timing/data': 0.005132513004355133, 'timing/model': 1.1167324620764703, 'learning_rate/qwen_vl_interface': 9.99137626404256e-06, 'learning_rate/action_model': 9.991376264042559e-05, 'epoch': 0.4}) +[2026-06-19 10:08:17,776][starVLA.training.train_starvla][INFO] - Step 195, Loss: {'train/loss': 0.7930744290351868, 'train/grad_norm_pre_clip': 14.003883361816406, 'timing/data': 0.004876127000898123, 'timing/model': 1.128437620936893, 'learning_rate/qwen_vl_interface': 9.991191861351015e-06, 'learning_rate/action_model': 9.991191861351014e-05, 'epoch': 0.4}) +[2026-06-19 10:08:22,266][starVLA.training.train_starvla][INFO] - Step 196, Loss: {'train/loss': 0.2273266464471817, 'train/grad_norm_pre_clip': 5.7981672286987305, 'timing/data': 0.0048755950992926955, 'timing/model': 1.1241517410380766, 'learning_rate/qwen_vl_interface': 9.991005509734587e-06, 'learning_rate/action_model': 9.991005509734587e-05, 'epoch': 0.4}) +[2026-06-19 10:08:26,737][starVLA.training.train_starvla][INFO] - Step 197, Loss: {'train/loss': 0.4021521210670471, 'train/grad_norm_pre_clip': 9.812507629394531, 'timing/data': 0.004738025018014014, 'timing/model': 1.2747030858881772, 'learning_rate/qwen_vl_interface': 9.990817209269878e-06, 'learning_rate/action_model': 9.990817209269878e-05, 'epoch': 0.4}) +[2026-06-19 10:08:31,060][starVLA.training.train_starvla][INFO] - Step 198, Loss: {'train/loss': 0.16309300065040588, 'train/grad_norm_pre_clip': 8.773386001586914, 'timing/data': 0.005021809949539602, 'timing/model': 1.1227418830385432, 'learning_rate/qwen_vl_interface': 9.990626960034291e-06, 'learning_rate/action_model': 9.990626960034291e-05, 'epoch': 0.41}) +[2026-06-19 10:08:35,548][starVLA.training.train_starvla][INFO] - Step 199, Loss: {'train/loss': 0.2576808035373688, 'train/grad_norm_pre_clip': 5.932347774505615, 'timing/data': 0.0051643409533426166, 'timing/model': 1.1206333910813555, 'learning_rate/qwen_vl_interface': 9.990434762106029e-06, 'learning_rate/action_model': 9.990434762106028e-05, 'epoch': 0.41}) +[2026-06-19 10:08:39,986][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 200 +[2026-06-19 10:11:33,232][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 200 +[2026-06-19 10:11:33,233][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 200 +[2026-06-19 10:14:37,080][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 200 +[2026-06-19 10:14:37,081][starVLA.training.train_starvla][INFO] - Step 200, Loss: {'train/loss': 0.2950425446033478, 'train/grad_norm_pre_clip': 9.000149726867676, 'eval/loss': 0.34354308247566223, 'eval/latency_0/loss': 0.17599539458751678, 'eval/latency_1/loss': 0.2519521415233612, 'eval/latency_2/loss': 0.4138123691082001, 'eval/latency_3/loss': 0.3831031620502472, 'eval/latency_4/loss': 0.4821121096611023, 'eval/flappy/loss': 0.34354308247566223, 'eval/flappy/latency_0/loss': 0.17599539458751678, 'eval/flappy/latency_1/loss': 0.2519521415233612, 'eval/flappy/latency_2/loss': 0.4138123691082001, 'eval/flappy/latency_3/loss': 0.3831031620502472, 'eval/flappy/latency_4/loss': 0.4821121096611023, 'eval/flap_precision': 0.5303030303030303, 'eval/flap_recall': 0.6363636363636364, 'eval/flap_f1': 0.578512396694215, 'eval/flap_rate_model': 0.09166666666666666, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.578512396694215, 'eval/latency_0/flap_precision': 0.5303030303030303, 'eval/latency_0/flap_recall': 0.6363636363636364, 'eval/latency_0/flap_f1': 0.578512396694215, 'eval/latency_0/flap_rate_model': 0.09166666666666666, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.578512396694215, 'timing/data': 0.005133206956088543, 'timing/model': 1.118274859036319, 'learning_rate/qwen_vl_interface': 9.990240615564099e-06, 'learning_rate/action_model': 9.990240615564099e-05, 'epoch': 0.41}) +[2026-06-19 10:14:37,081][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_200_state +[2026-06-19 10:14:37,084][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-19 10:14:40,822][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_200_state/model.safetensors +[2026-06-19 10:14:48,547][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_200_state/optimizer.bin +[2026-06-19 10:14:48,548][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_200_state/sampler.bin +[2026-06-19 10:14:48,548][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_200_state/sampler_1.bin +[2026-06-19 10:14:48,549][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_200_state/random_states_0.pkl +[2026-06-19 10:14:48,573][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 200 +[2026-06-19 10:14:48,573][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-19 10:14:48,628][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-19 10:14:53,056][starVLA.training.train_starvla][INFO] - Step 201, Loss: {'train/loss': 0.47910037636756897, 'train/grad_norm_pre_clip': 10.980539321899414, 'timing/data': 0.0051824579713866115, 'timing/model': 1.0986403001006693, 'learning_rate/qwen_vl_interface': 9.990044520488305e-06, 'learning_rate/action_model': 9.990044520488305e-05, 'epoch': 0.41}) +[2026-06-19 10:14:57,464][starVLA.training.train_starvla][INFO] - Step 202, Loss: {'train/loss': 0.6369997262954712, 'train/grad_norm_pre_clip': 11.057939529418945, 'timing/data': 0.0051392229506745934, 'timing/model': 1.2721980329370126, 'learning_rate/qwen_vl_interface': 9.989846476959258e-06, 'learning_rate/action_model': 9.989846476959257e-05, 'epoch': 0.41}) +[2026-06-19 10:15:01,735][starVLA.training.train_starvla][INFO] - Step 203, Loss: {'train/loss': 0.3333664536476135, 'train/grad_norm_pre_clip': 4.5565185546875, 'timing/data': 0.004885291098617017, 'timing/model': 1.1103541280608624, 'learning_rate/qwen_vl_interface': 9.989646485058361e-06, 'learning_rate/action_model': 9.989646485058361e-05, 'epoch': 0.42}) +[2026-06-19 10:15:06,162][starVLA.training.train_starvla][INFO] - Step 204, Loss: {'train/loss': 0.4841929078102112, 'train/grad_norm_pre_clip': 10.240880012512207, 'timing/data': 0.004999132012017071, 'timing/model': 1.1121815830701962, 'learning_rate/qwen_vl_interface': 9.98944454486783e-06, 'learning_rate/action_model': 9.989444544867829e-05, 'epoch': 0.42}) +[2026-06-19 10:15:10,483][starVLA.training.train_starvla][INFO] - Step 205, Loss: {'train/loss': 0.3443784713745117, 'train/grad_norm_pre_clip': 4.726736068725586, 'timing/data': 0.0052283980185166, 'timing/model': 1.1022922219708562, 'learning_rate/qwen_vl_interface': 9.989240656470667e-06, 'learning_rate/action_model': 9.989240656470666e-05, 'epoch': 0.42}) +[2026-06-19 10:15:14,849][starVLA.training.train_starvla][INFO] - Step 206, Loss: {'train/loss': 0.3648601472377777, 'train/grad_norm_pre_clip': 10.041862487792969, 'timing/data': 0.005268196924589574, 'timing/model': 1.1083241669693962, 'learning_rate/qwen_vl_interface': 9.98903481995069e-06, 'learning_rate/action_model': 9.989034819950689e-05, 'epoch': 0.42}) +[2026-06-19 10:15:19,250][starVLA.training.train_starvla][INFO] - Step 207, Loss: {'train/loss': 0.47672227025032043, 'train/grad_norm_pre_clip': 12.179482460021973, 'timing/data': 0.0052692589815706015, 'timing/model': 1.1037831150460988, 'learning_rate/qwen_vl_interface': 9.988827035392506e-06, 'learning_rate/action_model': 9.988827035392507e-05, 'epoch': 0.43}) +[2026-06-19 10:15:23,688][starVLA.training.train_starvla][INFO] - Step 208, Loss: {'train/loss': 0.4016803503036499, 'train/grad_norm_pre_clip': 9.22350025177002, 'timing/data': 0.005011365981772542, 'timing/model': 1.281612416030839, 'learning_rate/qwen_vl_interface': 9.988617302881529e-06, 'learning_rate/action_model': 9.988617302881529e-05, 'epoch': 0.43}) +[2026-06-19 10:15:27,960][starVLA.training.train_starvla][INFO] - Step 209, Loss: {'train/loss': 0.2506307363510132, 'train/grad_norm_pre_clip': 4.344839572906494, 'timing/data': 0.004616292077116668, 'timing/model': 1.1074461630778387, 'learning_rate/qwen_vl_interface': 9.988405622503972e-06, 'learning_rate/action_model': 9.988405622503971e-05, 'epoch': 0.43}) +[2026-06-19 10:15:32,392][starVLA.training.train_starvla][INFO] - Step 210, Loss: {'train/loss': 0.19241267442703247, 'train/grad_norm_pre_clip': 5.3041486740112305, 'timing/data': 0.004926261026412249, 'timing/model': 1.107759059057571, 'learning_rate/qwen_vl_interface': 9.988191994346848e-06, 'learning_rate/action_model': 9.988191994346848e-05, 'epoch': 0.43}) +[2026-06-19 10:15:36,838][starVLA.training.train_starvla][INFO] - Step 211, Loss: {'train/loss': 0.4797971546649933, 'train/grad_norm_pre_clip': 12.134422302246094, 'timing/data': 0.004938571946695447, 'timing/model': 1.1051614750176668, 'learning_rate/qwen_vl_interface': 9.987976418497974e-06, 'learning_rate/action_model': 9.987976418497972e-05, 'epoch': 0.43}) +[2026-06-19 10:15:41,253][starVLA.training.train_starvla][INFO] - Step 212, Loss: {'train/loss': 0.21555101871490479, 'train/grad_norm_pre_clip': 4.127810478210449, 'timing/data': 0.005012867972254753, 'timing/model': 1.114417522912845, 'learning_rate/qwen_vl_interface': 9.987758895045964e-06, 'learning_rate/action_model': 9.987758895045963e-05, 'epoch': 0.44}) +[2026-06-19 10:15:45,650][starVLA.training.train_starvla][INFO] - Step 213, Loss: {'train/loss': 0.38771170377731323, 'train/grad_norm_pre_clip': 9.18257999420166, 'timing/data': 0.004786583944223821, 'timing/model': 1.1123295639408752, 'learning_rate/qwen_vl_interface': 9.98753942408023e-06, 'learning_rate/action_model': 9.98753942408023e-05, 'epoch': 0.44}) +[2026-06-19 10:15:50,129][starVLA.training.train_starvla][INFO] - Step 214, Loss: {'train/loss': 0.33038806915283203, 'train/grad_norm_pre_clip': 8.766448974609375, 'timing/data': 0.005032363929785788, 'timing/model': 1.3040382099570706, 'learning_rate/qwen_vl_interface': 9.987318005690992e-06, 'learning_rate/action_model': 9.987318005690992e-05, 'epoch': 0.44}) +[2026-06-19 10:15:54,419][starVLA.training.train_starvla][INFO] - Step 215, Loss: {'train/loss': 0.49775540828704834, 'train/grad_norm_pre_clip': 11.91234016418457, 'timing/data': 0.004478971008211374, 'timing/model': 1.1139430070761591, 'learning_rate/qwen_vl_interface': 9.987094639969268e-06, 'learning_rate/action_model': 9.987094639969268e-05, 'epoch': 0.44}) +[2026-06-19 10:15:58,856][starVLA.training.train_starvla][INFO] - Step 216, Loss: {'train/loss': 0.1567249596118927, 'train/grad_norm_pre_clip': 3.5970277786254883, 'timing/data': 0.005200652056373656, 'timing/model': 1.1132158460095525, 'learning_rate/qwen_vl_interface': 9.986869327006872e-06, 'learning_rate/action_model': 9.986869327006873e-05, 'epoch': 0.44}) +[2026-06-19 10:16:03,311][starVLA.training.train_starvla][INFO] - Step 217, Loss: {'train/loss': 0.3497961461544037, 'train/grad_norm_pre_clip': 2.9226322174072266, 'timing/data': 0.004643123014830053, 'timing/model': 1.118063393048942, 'learning_rate/qwen_vl_interface': 9.986642066896424e-06, 'learning_rate/action_model': 9.986642066896423e-05, 'epoch': 0.45}) +[2026-06-19 10:16:07,759][starVLA.training.train_starvla][INFO] - Step 218, Loss: {'train/loss': 0.4283221960067749, 'train/grad_norm_pre_clip': 3.4130890369415283, 'timing/data': 0.004991192021407187, 'timing/model': 1.1136301790829748, 'learning_rate/qwen_vl_interface': 9.986412859731341e-06, 'learning_rate/action_model': 9.98641285973134e-05, 'epoch': 0.45}) +[2026-06-19 10:16:12,221][starVLA.training.train_starvla][INFO] - Step 219, Loss: {'train/loss': 0.3225148618221283, 'train/grad_norm_pre_clip': 7.157883644104004, 'timing/data': 0.005041241995058954, 'timing/model': 1.1212221119785681, 'learning_rate/qwen_vl_interface': 9.986181705605839e-06, 'learning_rate/action_model': 9.986181705605839e-05, 'epoch': 0.45}) +[2026-06-19 10:16:16,687][starVLA.training.train_starvla][INFO] - Step 220, Loss: {'train/loss': 0.4847579896450043, 'train/grad_norm_pre_clip': 11.825155258178711, 'timing/data': 0.005051961052231491, 'timing/model': 1.2770585119724274, 'learning_rate/qwen_vl_interface': 9.98594860461494e-06, 'learning_rate/action_model': 9.98594860461494e-05, 'epoch': 0.45}) +[2026-06-19 10:16:20,978][starVLA.training.train_starvla][INFO] - Step 221, Loss: {'train/loss': 0.3710770606994629, 'train/grad_norm_pre_clip': 9.513915061950684, 'timing/data': 0.005036795977503061, 'timing/model': 1.112490203930065, 'learning_rate/qwen_vl_interface': 9.985713556854461e-06, 'learning_rate/action_model': 9.98571355685446e-05, 'epoch': 0.45}) +[2026-06-19 10:16:25,482][starVLA.training.train_starvla][INFO] - Step 222, Loss: {'train/loss': 0.28576213121414185, 'train/grad_norm_pre_clip': 7.2899041175842285, 'timing/data': 0.005288533982820809, 'timing/model': 1.1336587499827147, 'learning_rate/qwen_vl_interface': 9.985476562421024e-06, 'learning_rate/action_model': 9.985476562421024e-05, 'epoch': 0.46}) +[2026-06-19 10:16:29,982][starVLA.training.train_starvla][INFO] - Step 223, Loss: {'train/loss': 0.3643043637275696, 'train/grad_norm_pre_clip': 5.321681022644043, 'timing/data': 0.004985078005120158, 'timing/model': 1.1240224039647728, 'learning_rate/qwen_vl_interface': 9.985237621412047e-06, 'learning_rate/action_model': 9.985237621412046e-05, 'epoch': 0.46}) +[2026-06-19 10:16:34,427][starVLA.training.train_starvla][INFO] - Step 224, Loss: {'train/loss': 0.24731992185115814, 'train/grad_norm_pre_clip': 3.7380785942077637, 'timing/data': 0.005023646983318031, 'timing/model': 1.1160867050057277, 'learning_rate/qwen_vl_interface': 9.984996733925749e-06, 'learning_rate/action_model': 9.984996733925748e-05, 'epoch': 0.46}) +[2026-06-19 10:16:38,914][starVLA.training.train_starvla][INFO] - Step 225, Loss: {'train/loss': 0.839676022529602, 'train/grad_norm_pre_clip': 16.820789337158203, 'timing/data': 0.004613680066540837, 'timing/model': 1.113890678039752, 'learning_rate/qwen_vl_interface': 9.98475390006115e-06, 'learning_rate/action_model': 9.98475390006115e-05, 'epoch': 0.46}) +[2026-06-19 10:16:43,392][starVLA.training.train_starvla][INFO] - Step 226, Loss: {'train/loss': 0.6400364637374878, 'train/grad_norm_pre_clip': 11.516368865966797, 'timing/data': 0.005198840983211994, 'timing/model': 1.2879067349713296, 'learning_rate/qwen_vl_interface': 9.98450911991807e-06, 'learning_rate/action_model': 9.98450911991807e-05, 'epoch': 0.46}) +[2026-06-19 10:16:47,697][starVLA.training.train_starvla][INFO] - Step 227, Loss: {'train/loss': 0.3926663100719452, 'train/grad_norm_pre_clip': 7.9908976554870605, 'timing/data': 0.005058145034126937, 'timing/model': 1.115154113038443, 'learning_rate/qwen_vl_interface': 9.98426239359713e-06, 'learning_rate/action_model': 9.984262393597129e-05, 'epoch': 0.47}) +[2026-06-19 10:16:52,169][starVLA.training.train_starvla][INFO] - Step 228, Loss: {'train/loss': 0.2791127562522888, 'train/grad_norm_pre_clip': 4.282768726348877, 'timing/data': 0.005207745009101927, 'timing/model': 1.1215943730203435, 'learning_rate/qwen_vl_interface': 9.984013721199747e-06, 'learning_rate/action_model': 9.984013721199747e-05, 'epoch': 0.47}) +[2026-06-19 10:16:56,567][starVLA.training.train_starvla][INFO] - Step 229, Loss: {'train/loss': 0.3476802110671997, 'train/grad_norm_pre_clip': 2.8317513465881348, 'timing/data': 0.004939775099046528, 'timing/model': 1.1147315161069855, 'learning_rate/qwen_vl_interface': 9.983763102828143e-06, 'learning_rate/action_model': 9.983763102828143e-05, 'epoch': 0.47}) +[2026-06-19 10:17:01,014][starVLA.training.train_starvla][INFO] - Step 230, Loss: {'train/loss': 0.5700157880783081, 'train/grad_norm_pre_clip': 7.315993309020996, 'timing/data': 0.004914313089102507, 'timing/model': 1.1145479580154642, 'learning_rate/qwen_vl_interface': 9.983510538585336e-06, 'learning_rate/action_model': 9.983510538585337e-05, 'epoch': 0.47}) +[2026-06-19 10:17:05,495][starVLA.training.train_starvla][INFO] - Step 231, Loss: {'train/loss': 0.4281338155269623, 'train/grad_norm_pre_clip': 8.2987699508667, 'timing/data': 0.005037397029809654, 'timing/model': 1.1252326099202037, 'learning_rate/qwen_vl_interface': 9.98325602857515e-06, 'learning_rate/action_model': 9.98325602857515e-05, 'epoch': 0.47}) +[2026-06-19 10:17:09,951][starVLA.training.train_starvla][INFO] - Step 232, Loss: {'train/loss': 0.24282866716384888, 'train/grad_norm_pre_clip': 8.38560962677002, 'timing/data': 0.004714254988357425, 'timing/model': 1.2699515599524602, 'learning_rate/qwen_vl_interface': 9.9829995729022e-06, 'learning_rate/action_model': 9.982999572902199e-05, 'epoch': 0.48}) +[2026-06-19 10:17:14,251][starVLA.training.train_starvla][INFO] - Step 233, Loss: {'train/loss': 0.3993172347545624, 'train/grad_norm_pre_clip': 8.22278118133545, 'timing/data': 0.005189293995499611, 'timing/model': 1.1142820629756898, 'learning_rate/qwen_vl_interface': 9.982741171671905e-06, 'learning_rate/action_model': 9.982741171671906e-05, 'epoch': 0.48}) +[2026-06-19 10:17:18,716][starVLA.training.train_starvla][INFO] - Step 234, Loss: {'train/loss': 0.3415620028972626, 'train/grad_norm_pre_clip': 3.5504307746887207, 'timing/data': 0.004913361975923181, 'timing/model': 1.1215925499564037, 'learning_rate/qwen_vl_interface': 9.982480824990489e-06, 'learning_rate/action_model': 9.982480824990488e-05, 'epoch': 0.48}) +[2026-06-19 10:17:23,136][starVLA.training.train_starvla][INFO] - Step 235, Loss: {'train/loss': 0.13069066405296326, 'train/grad_norm_pre_clip': 3.6268908977508545, 'timing/data': 0.0045536760007962584, 'timing/model': 1.1169512220658362, 'learning_rate/qwen_vl_interface': 9.982218532964963e-06, 'learning_rate/action_model': 9.982218532964964e-05, 'epoch': 0.48}) +[2026-06-19 10:17:27,610][starVLA.training.train_starvla][INFO] - Step 236, Loss: {'train/loss': 0.5535985827445984, 'train/grad_norm_pre_clip': 10.178884506225586, 'timing/data': 0.005165613954886794, 'timing/model': 1.1258699289755896, 'learning_rate/qwen_vl_interface': 9.981954295703152e-06, 'learning_rate/action_model': 9.981954295703152e-05, 'epoch': 0.48}) +[2026-06-19 10:17:32,115][starVLA.training.train_starvla][INFO] - Step 237, Loss: {'train/loss': 0.07774963974952698, 'train/grad_norm_pre_clip': 2.583705186843872, 'timing/data': 0.004961973987519741, 'timing/model': 1.126713844947517, 'learning_rate/qwen_vl_interface': 9.981688113313672e-06, 'learning_rate/action_model': 9.981688113313672e-05, 'epoch': 0.49}) +[2026-06-19 10:17:36,600][starVLA.training.train_starvla][INFO] - Step 238, Loss: {'train/loss': 0.2769193947315216, 'train/grad_norm_pre_clip': 6.166096210479736, 'timing/data': 0.005005065002478659, 'timing/model': 1.2965770940063521, 'learning_rate/qwen_vl_interface': 9.98141998590594e-06, 'learning_rate/action_model': 9.981419985905939e-05, 'epoch': 0.49}) +[2026-06-19 10:17:40,893][starVLA.training.train_starvla][INFO] - Step 239, Loss: {'train/loss': 0.2302754521369934, 'train/grad_norm_pre_clip': 5.238026142120361, 'timing/data': 0.005150804994627833, 'timing/model': 1.1141236710827798, 'learning_rate/qwen_vl_interface': 9.981149913590171e-06, 'learning_rate/action_model': 9.981149913590172e-05, 'epoch': 0.49}) +[2026-06-19 10:17:45,369][starVLA.training.train_starvla][INFO] - Step 240, Loss: {'train/loss': 1.069051742553711, 'train/grad_norm_pre_clip': 17.666034698486328, 'timing/data': 0.005084292963147163, 'timing/model': 1.119249744922854, 'learning_rate/qwen_vl_interface': 9.980877896477386e-06, 'learning_rate/action_model': 9.980877896477386e-05, 'epoch': 0.49}) +[2026-06-19 10:17:49,809][starVLA.training.train_starvla][INFO] - Step 241, Loss: {'train/loss': 0.16583068668842316, 'train/grad_norm_pre_clip': 3.459634780883789, 'timing/data': 0.005003647995181382, 'timing/model': 1.119956922950223, 'learning_rate/qwen_vl_interface': 9.980603934679397e-06, 'learning_rate/action_model': 9.980603934679398e-05, 'epoch': 0.49}) +[2026-06-19 10:17:54,277][starVLA.training.train_starvla][INFO] - Step 242, Loss: {'train/loss': 0.41985926032066345, 'train/grad_norm_pre_clip': 7.1228928565979, 'timing/data': 0.00521980703342706, 'timing/model': 1.117833427968435, 'learning_rate/qwen_vl_interface': 9.980328028308824e-06, 'learning_rate/action_model': 9.980328028308824e-05, 'epoch': 0.5}) +[2026-06-19 10:17:58,757][starVLA.training.train_starvla][INFO] - Step 243, Loss: {'train/loss': 0.40231356024742126, 'train/grad_norm_pre_clip': 10.282136917114258, 'timing/data': 0.005147768999449909, 'timing/model': 1.118801640928723, 'learning_rate/qwen_vl_interface': 9.980050177479075e-06, 'learning_rate/action_model': 9.980050177479076e-05, 'epoch': 0.5}) +[2026-06-19 10:18:03,223][starVLA.training.train_starvla][INFO] - Step 244, Loss: {'train/loss': 0.20409643650054932, 'train/grad_norm_pre_clip': 3.904782772064209, 'timing/data': 0.005120744928717613, 'timing/model': 1.2768530630273744, 'learning_rate/qwen_vl_interface': 9.97977038230437e-06, 'learning_rate/action_model': 9.97977038230437e-05, 'epoch': 0.5}) +[2026-06-19 10:18:07,628][starVLA.training.train_starvla][INFO] - Step 245, Loss: {'train/loss': 0.3293236196041107, 'train/grad_norm_pre_clip': 12.102262496948242, 'timing/data': 0.004999862983822823, 'timing/model': 1.214879820938222, 'learning_rate/qwen_vl_interface': 9.97948864289972e-06, 'learning_rate/action_model': 9.97948864289972e-05, 'epoch': 0.5}) +[2026-06-19 10:18:11,933][starVLA.training.train_starvla][INFO] - Step 246, Loss: {'train/loss': 0.29446303844451904, 'train/grad_norm_pre_clip': 8.080155372619629, 'timing/data': 0.004978447919711471, 'timing/model': 1.1185435800580308, 'learning_rate/qwen_vl_interface': 9.979204959380936e-06, 'learning_rate/action_model': 9.979204959380936e-05, 'epoch': 0.51}) +[2026-06-19 10:18:16,368][starVLA.training.train_starvla][INFO] - Step 247, Loss: {'train/loss': 0.4977321922779083, 'train/grad_norm_pre_clip': 11.533405303955078, 'timing/data': 0.005054254084825516, 'timing/model': 1.1172458509681746, 'learning_rate/qwen_vl_interface': 9.978919331864631e-06, 'learning_rate/action_model': 9.97891933186463e-05, 'epoch': 0.51}) +[2026-06-19 10:18:20,860][starVLA.training.train_starvla][INFO] - Step 248, Loss: {'train/loss': 0.46198728680610657, 'train/grad_norm_pre_clip': 18.678668975830078, 'timing/data': 0.004948952933773398, 'timing/model': 1.1223927809624001, 'learning_rate/qwen_vl_interface': 9.978631760468216e-06, 'learning_rate/action_model': 9.978631760468216e-05, 'epoch': 0.51}) +[2026-06-19 10:18:25,369][starVLA.training.train_starvla][INFO] - Step 249, Loss: {'train/loss': 0.33508190512657166, 'train/grad_norm_pre_clip': 4.417479991912842, 'timing/data': 0.004895391990430653, 'timing/model': 1.1199017430189997, 'learning_rate/qwen_vl_interface': 9.9783422453099e-06, 'learning_rate/action_model': 9.9783422453099e-05, 'epoch': 0.51}) +[2026-06-19 10:18:29,872][starVLA.training.train_starvla][INFO] - Step 250, Loss: {'train/loss': 0.3100457787513733, 'train/grad_norm_pre_clip': 4.923096179962158, 'timing/data': 0.004981218953616917, 'timing/model': 1.2997387580107898, 'learning_rate/qwen_vl_interface': 9.978050786508693e-06, 'learning_rate/action_model': 9.978050786508692e-05, 'epoch': 0.51}) +[2026-06-19 10:18:34,312][starVLA.training.train_starvla][INFO] - Step 251, Loss: {'train/loss': 0.19279718399047852, 'train/grad_norm_pre_clip': 3.001378059387207, 'timing/data': 0.004890190088190138, 'timing/model': 1.2365255620097741, 'learning_rate/qwen_vl_interface': 9.9777573841844e-06, 'learning_rate/action_model': 9.977757384184401e-05, 'epoch': 0.52}) +[2026-06-19 10:18:38,638][starVLA.training.train_starvla][INFO] - Step 252, Loss: {'train/loss': 0.2475840151309967, 'train/grad_norm_pre_clip': 2.6313817501068115, 'timing/data': 0.005130614037625492, 'timing/model': 1.1259208689443767, 'learning_rate/qwen_vl_interface': 9.97746203845763e-06, 'learning_rate/action_model': 9.97746203845763e-05, 'epoch': 0.52}) +[2026-06-19 10:18:43,092][starVLA.training.train_starvla][INFO] - Step 253, Loss: {'train/loss': 0.33478859066963196, 'train/grad_norm_pre_clip': 5.6942949295043945, 'timing/data': 0.0053805779898539186, 'timing/model': 1.1191841840045527, 'learning_rate/qwen_vl_interface': 9.97716474944979e-06, 'learning_rate/action_model': 9.977164749449788e-05, 'epoch': 0.52}) +[2026-06-19 10:18:47,584][starVLA.training.train_starvla][INFO] - Step 254, Loss: {'train/loss': 0.2310744971036911, 'train/grad_norm_pre_clip': 6.939128875732422, 'timing/data': 0.004919858067296445, 'timing/model': 1.1241491689579561, 'learning_rate/qwen_vl_interface': 9.976865517283079e-06, 'learning_rate/action_model': 9.976865517283078e-05, 'epoch': 0.52}) +[2026-06-19 10:18:52,066][starVLA.training.train_starvla][INFO] - Step 255, Loss: {'train/loss': 0.10740238428115845, 'train/grad_norm_pre_clip': 2.1016547679901123, 'timing/data': 0.004994855960831046, 'timing/model': 1.1267081940313801, 'learning_rate/qwen_vl_interface': 9.976564342080505e-06, 'learning_rate/action_model': 9.976564342080504e-05, 'epoch': 0.52}) +[2026-06-19 10:18:56,536][starVLA.training.train_starvla][INFO] - Step 256, Loss: {'train/loss': 0.3955978453159332, 'train/grad_norm_pre_clip': 7.898489952087402, 'timing/data': 0.00519456306938082, 'timing/model': 1.28474875993561, 'learning_rate/qwen_vl_interface': 9.976261223965867e-06, 'learning_rate/action_model': 9.976261223965865e-05, 'epoch': 0.53}) +[2026-06-19 10:19:00,862][starVLA.training.train_starvla][INFO] - Step 257, Loss: {'train/loss': 0.19247405230998993, 'train/grad_norm_pre_clip': 4.547785758972168, 'timing/data': 0.004878920968621969, 'timing/model': 1.1248299319995567, 'learning_rate/qwen_vl_interface': 9.975956163063767e-06, 'learning_rate/action_model': 9.975956163063767e-05, 'epoch': 0.53}) +[2026-06-19 10:19:05,338][starVLA.training.train_starvla][INFO] - Step 258, Loss: {'train/loss': 0.40341219305992126, 'train/grad_norm_pre_clip': 13.09832763671875, 'timing/data': 0.004726979066617787, 'timing/model': 1.1198674809420481, 'learning_rate/qwen_vl_interface': 9.975649159499605e-06, 'learning_rate/action_model': 9.975649159499603e-05, 'epoch': 0.53}) +[2026-06-19 10:19:09,826][starVLA.training.train_starvla][INFO] - Step 259, Loss: {'train/loss': 0.2582687735557556, 'train/grad_norm_pre_clip': 4.654109001159668, 'timing/data': 0.00484197202604264, 'timing/model': 1.1189277159282938, 'learning_rate/qwen_vl_interface': 9.975340213399573e-06, 'learning_rate/action_model': 9.975340213399573e-05, 'epoch': 0.53}) +[2026-06-19 10:19:14,292][starVLA.training.train_starvla][INFO] - Step 260, Loss: {'train/loss': 0.15414217114448547, 'train/grad_norm_pre_clip': 3.998375415802002, 'timing/data': 0.004926050081849098, 'timing/model': 1.1193214089144021, 'learning_rate/qwen_vl_interface': 9.975029324890674e-06, 'learning_rate/action_model': 9.975029324890674e-05, 'epoch': 0.53}) +[2026-06-19 10:19:18,709][starVLA.training.train_starvla][INFO] - Step 261, Loss: {'train/loss': 0.4445904493331909, 'train/grad_norm_pre_clip': 9.312946319580078, 'timing/data': 0.004997541895136237, 'timing/model': 1.1173607539385557, 'learning_rate/qwen_vl_interface': 9.974716494100698e-06, 'learning_rate/action_model': 9.974716494100697e-05, 'epoch': 0.54}) +[2026-06-19 10:19:23,181][starVLA.training.train_starvla][INFO] - Step 262, Loss: {'train/loss': 0.4735372066497803, 'train/grad_norm_pre_clip': 18.79647445678711, 'timing/data': 0.005034219939261675, 'timing/model': 1.2689325200626627, 'learning_rate/qwen_vl_interface': 9.97440172115824e-06, 'learning_rate/action_model': 9.974401721158239e-05, 'epoch': 0.54}) +[2026-06-19 10:19:27,510][starVLA.training.train_starvla][INFO] - Step 263, Loss: {'train/loss': 0.3264596164226532, 'train/grad_norm_pre_clip': 7.478747367858887, 'timing/data': 0.005083605996333063, 'timing/model': 1.1242422809591517, 'learning_rate/qwen_vl_interface': 9.97408500619269e-06, 'learning_rate/action_model': 9.97408500619269e-05, 'epoch': 0.54}) +[2026-06-19 10:19:31,984][starVLA.training.train_starvla][INFO] - Step 264, Loss: {'train/loss': 0.4363854229450226, 'train/grad_norm_pre_clip': 5.139508247375488, 'timing/data': 0.0049365999875590205, 'timing/model': 1.1191548689967021, 'learning_rate/qwen_vl_interface': 9.973766349334239e-06, 'learning_rate/action_model': 9.973766349334239e-05, 'epoch': 0.54}) +[2026-06-19 10:19:36,490][starVLA.training.train_starvla][INFO] - Step 265, Loss: {'train/loss': 0.3210037648677826, 'train/grad_norm_pre_clip': 8.793815612792969, 'timing/data': 0.004724905011244118, 'timing/model': 1.1209271649131551, 'learning_rate/qwen_vl_interface': 9.973445750713875e-06, 'learning_rate/action_model': 9.973445750713875e-05, 'epoch': 0.54}) +[2026-06-19 10:19:40,994][starVLA.training.train_starvla][INFO] - Step 266, Loss: {'train/loss': 0.1556330770254135, 'train/grad_norm_pre_clip': 9.082313537597656, 'timing/data': 0.005086649092845619, 'timing/model': 1.1298526610480621, 'learning_rate/qwen_vl_interface': 9.97312321046338e-06, 'learning_rate/action_model': 9.97312321046338e-05, 'epoch': 0.55}) +[2026-06-19 10:19:45,445][starVLA.training.train_starvla][INFO] - Step 267, Loss: {'train/loss': 0.2107386291027069, 'train/grad_norm_pre_clip': 15.323704719543457, 'timing/data': 0.005011187982745469, 'timing/model': 1.1287924778880551, 'learning_rate/qwen_vl_interface': 9.972798728715343e-06, 'learning_rate/action_model': 9.972798728715344e-05, 'epoch': 0.55}) +[2026-06-19 10:19:49,920][starVLA.training.train_starvla][INFO] - Step 268, Loss: {'train/loss': 0.546417772769928, 'train/grad_norm_pre_clip': 8.584836959838867, 'timing/data': 0.004934550961479545, 'timing/model': 1.2666223680134863, 'learning_rate/qwen_vl_interface': 9.972472305603145e-06, 'learning_rate/action_model': 9.972472305603144e-05, 'epoch': 0.55}) +[2026-06-19 10:19:54,249][starVLA.training.train_starvla][INFO] - Step 269, Loss: {'train/loss': 0.6076827049255371, 'train/grad_norm_pre_clip': 17.319692611694336, 'timing/data': 0.004776434972882271, 'timing/model': 1.1286321820225567, 'learning_rate/qwen_vl_interface': 9.972143941260965e-06, 'learning_rate/action_model': 9.972143941260965e-05, 'epoch': 0.55}) +[2026-06-19 10:19:58,736][starVLA.training.train_starvla][INFO] - Step 270, Loss: {'train/loss': 0.2105022817850113, 'train/grad_norm_pre_clip': 6.246639251708984, 'timing/data': 0.0046318479580804706, 'timing/model': 1.1182461610296741, 'learning_rate/qwen_vl_interface': 9.971813635823781e-06, 'learning_rate/action_model': 9.971813635823781e-05, 'epoch': 0.55}) +[2026-06-19 10:20:03,210][starVLA.training.train_starvla][INFO] - Step 271, Loss: {'train/loss': 0.21515271067619324, 'train/grad_norm_pre_clip': 3.663574695587158, 'timing/data': 0.00571002799551934, 'timing/model': 1.1247370310593396, 'learning_rate/qwen_vl_interface': 9.971481389427372e-06, 'learning_rate/action_model': 9.971481389427371e-05, 'epoch': 0.56}) +[2026-06-19 10:20:07,684][starVLA.training.train_starvla][INFO] - Step 272, Loss: {'train/loss': 0.36337122321128845, 'train/grad_norm_pre_clip': 5.1036810874938965, 'timing/data': 0.00478710001334548, 'timing/model': 1.1169535290682688, 'learning_rate/qwen_vl_interface': 9.971147202208308e-06, 'learning_rate/action_model': 9.971147202208308e-05, 'epoch': 0.56}) +[2026-06-19 10:20:12,179][starVLA.training.train_starvla][INFO] - Step 273, Loss: {'train/loss': 0.2978536784648895, 'train/grad_norm_pre_clip': 6.117423057556152, 'timing/data': 0.005285819061100483, 'timing/model': 1.291727544972673, 'learning_rate/qwen_vl_interface': 9.970811074303962e-06, 'learning_rate/action_model': 9.970811074303962e-05, 'epoch': 0.56}) +[2026-06-19 10:20:16,561][starVLA.training.train_starvla][INFO] - Step 274, Loss: {'train/loss': 0.5759706497192383, 'train/grad_norm_pre_clip': 5.2150559425354, 'timing/data': 0.00483940402045846, 'timing/model': 1.1231512640370056, 'learning_rate/qwen_vl_interface': 9.970473005852505e-06, 'learning_rate/action_model': 9.970473005852505e-05, 'epoch': 0.56}) +[2026-06-19 10:20:21,053][starVLA.training.train_starvla][INFO] - Step 275, Loss: {'train/loss': 0.3628333806991577, 'train/grad_norm_pre_clip': 2.4691150188446045, 'timing/data': 0.0049805809976533055, 'timing/model': 1.118068565032445, 'learning_rate/qwen_vl_interface': 9.970132996992904e-06, 'learning_rate/action_model': 9.970132996992904e-05, 'epoch': 0.56}) +[2026-06-19 10:20:25,442][starVLA.training.train_starvla][INFO] - Step 276, Loss: {'train/loss': 0.44865092635154724, 'train/grad_norm_pre_clip': 5.67861795425415, 'timing/data': 0.0054591449443250895, 'timing/model': 1.120323970913887, 'learning_rate/qwen_vl_interface': 9.969791047864921e-06, 'learning_rate/action_model': 9.969791047864921e-05, 'epoch': 0.57}) +[2026-06-19 10:20:29,855][starVLA.training.train_starvla][INFO] - Step 277, Loss: {'train/loss': 0.26466378569602966, 'train/grad_norm_pre_clip': 3.359015941619873, 'timing/data': 0.005039470037445426, 'timing/model': 1.117700470960699, 'learning_rate/qwen_vl_interface': 9.969447158609122e-06, 'learning_rate/action_model': 9.96944715860912e-05, 'epoch': 0.57}) +[2026-06-19 10:20:34,316][starVLA.training.train_starvla][INFO] - Step 278, Loss: {'train/loss': 0.3387889862060547, 'train/grad_norm_pre_clip': 3.759829044342041, 'timing/data': 0.0050630709156394005, 'timing/model': 1.130885328981094, 'learning_rate/qwen_vl_interface': 9.969101329366864e-06, 'learning_rate/action_model': 9.969101329366863e-05, 'epoch': 0.57}) +[2026-06-19 10:20:38,789][starVLA.training.train_starvla][INFO] - Step 279, Loss: {'train/loss': 0.2314462661743164, 'train/grad_norm_pre_clip': 3.466442823410034, 'timing/data': 0.005438131978735328, 'timing/model': 1.2679523050319403, 'learning_rate/qwen_vl_interface': 9.968753560280305e-06, 'learning_rate/action_model': 9.968753560280304e-05, 'epoch': 0.57}) +[2026-06-19 10:20:43,122][starVLA.training.train_starvla][INFO] - Step 280, Loss: {'train/loss': 0.2726246416568756, 'train/grad_norm_pre_clip': 7.229311466217041, 'timing/data': 0.005204443936236203, 'timing/model': 1.1242386860540137, 'learning_rate/qwen_vl_interface': 9.968403851492402e-06, 'learning_rate/action_model': 9.968403851492401e-05, 'epoch': 0.58}) +[2026-06-19 10:20:47,619][starVLA.training.train_starvla][INFO] - Step 281, Loss: {'train/loss': 0.298420786857605, 'train/grad_norm_pre_clip': 8.279953956604004, 'timing/data': 0.004917674930766225, 'timing/model': 1.1166413499740884, 'learning_rate/qwen_vl_interface': 9.968052203146904e-06, 'learning_rate/action_model': 9.968052203146903e-05, 'epoch': 0.58}) +[2026-06-19 10:20:52,110][starVLA.training.train_starvla][INFO] - Step 282, Loss: {'train/loss': 0.35072091221809387, 'train/grad_norm_pre_clip': 6.31914758682251, 'timing/data': 0.005088150035589933, 'timing/model': 1.1200832689646631, 'learning_rate/qwen_vl_interface': 9.967698615388363e-06, 'learning_rate/action_model': 9.967698615388362e-05, 'epoch': 0.58}) +[2026-06-19 10:20:56,552][starVLA.training.train_starvla][INFO] - Step 283, Loss: {'train/loss': 0.10962490737438202, 'train/grad_norm_pre_clip': 6.880819320678711, 'timing/data': 0.005048281047493219, 'timing/model': 1.1275265540461987, 'learning_rate/qwen_vl_interface': 9.967343088362123e-06, 'learning_rate/action_model': 9.967343088362122e-05, 'epoch': 0.58}) +[2026-06-19 10:21:01,021][starVLA.training.train_starvla][INFO] - Step 284, Loss: {'train/loss': 0.37476781010627747, 'train/grad_norm_pre_clip': 7.757680892944336, 'timing/data': 0.005081776063889265, 'timing/model': 1.1250991020351648, 'learning_rate/qwen_vl_interface': 9.966985622214328e-06, 'learning_rate/action_model': 9.966985622214329e-05, 'epoch': 0.58}) +[2026-06-19 10:21:05,536][starVLA.training.train_starvla][INFO] - Step 285, Loss: {'train/loss': 0.07920831441879272, 'train/grad_norm_pre_clip': 2.3587796688079834, 'timing/data': 0.005130668985657394, 'timing/model': 1.2951310139615089, 'learning_rate/qwen_vl_interface': 9.966626217091922e-06, 'learning_rate/action_model': 9.966626217091922e-05, 'epoch': 0.59}) +[2026-06-19 10:21:09,851][starVLA.training.train_starvla][INFO] - Step 286, Loss: {'train/loss': 0.26501861214637756, 'train/grad_norm_pre_clip': 3.2845799922943115, 'timing/data': 0.004947759909555316, 'timing/model': 1.119847483932972, 'learning_rate/qwen_vl_interface': 9.966264873142641e-06, 'learning_rate/action_model': 9.96626487314264e-05, 'epoch': 0.59}) +[2026-06-19 10:21:14,329][starVLA.training.train_starvla][INFO] - Step 287, Loss: {'train/loss': 0.3311212658882141, 'train/grad_norm_pre_clip': 11.16004467010498, 'timing/data': 0.004993640934117138, 'timing/model': 1.1162453530123457, 'learning_rate/qwen_vl_interface': 9.965901590515018e-06, 'learning_rate/action_model': 9.965901590515018e-05, 'epoch': 0.59}) +[2026-06-19 10:21:18,817][starVLA.training.train_starvla][INFO] - Step 288, Loss: {'train/loss': 0.3800840377807617, 'train/grad_norm_pre_clip': 11.745725631713867, 'timing/data': 0.004981564008630812, 'timing/model': 1.1180773680098355, 'learning_rate/qwen_vl_interface': 9.965536369358386e-06, 'learning_rate/action_model': 9.965536369358386e-05, 'epoch': 0.59}) +[2026-06-19 10:21:23,306][starVLA.training.train_starvla][INFO] - Step 289, Loss: {'train/loss': 0.16754138469696045, 'train/grad_norm_pre_clip': 5.018815040588379, 'timing/data': 0.005221946048550308, 'timing/model': 1.1202797099249437, 'learning_rate/qwen_vl_interface': 9.965169209822874e-06, 'learning_rate/action_model': 9.965169209822874e-05, 'epoch': 0.59}) +[2026-06-19 10:21:27,771][starVLA.training.train_starvla][INFO] - Step 290, Loss: {'train/loss': 0.38617321848869324, 'train/grad_norm_pre_clip': 10.395337104797363, 'timing/data': 0.005137834930792451, 'timing/model': 1.1241973950527608, 'learning_rate/qwen_vl_interface': 9.96480011205941e-06, 'learning_rate/action_model': 9.964800112059409e-05, 'epoch': 0.6}) +[2026-06-19 10:21:32,273][starVLA.training.train_starvla][INFO] - Step 291, Loss: {'train/loss': 0.3532947897911072, 'train/grad_norm_pre_clip': 5.158811569213867, 'timing/data': 0.005095929955132306, 'timing/model': 1.2917393679963425, 'learning_rate/qwen_vl_interface': 9.964429076219711e-06, 'learning_rate/action_model': 9.964429076219711e-05, 'epoch': 0.6}) +[2026-06-19 10:21:36,580][starVLA.training.train_starvla][INFO] - Step 292, Loss: {'train/loss': 0.2434743195772171, 'train/grad_norm_pre_clip': 5.838284492492676, 'timing/data': 0.005027592065744102, 'timing/model': 1.1206220949534327, 'learning_rate/qwen_vl_interface': 9.964056102456299e-06, 'learning_rate/action_model': 9.964056102456299e-05, 'epoch': 0.6}) +[2026-06-19 10:21:41,050][starVLA.training.train_starvla][INFO] - Step 293, Loss: {'train/loss': 0.20426425337791443, 'train/grad_norm_pre_clip': 3.689948320388794, 'timing/data': 0.0049688429571688175, 'timing/model': 1.119018642930314, 'learning_rate/qwen_vl_interface': 9.963681190922489e-06, 'learning_rate/action_model': 9.963681190922488e-05, 'epoch': 0.6}) +[2026-06-19 10:21:45,545][starVLA.training.train_starvla][INFO] - Step 294, Loss: {'train/loss': 0.3913629353046417, 'train/grad_norm_pre_clip': 9.927396774291992, 'timing/data': 0.004793531959876418, 'timing/model': 1.1233931259484962, 'learning_rate/qwen_vl_interface': 9.963304341772391e-06, 'learning_rate/action_model': 9.963304341772391e-05, 'epoch': 0.6}) +[2026-06-19 10:21:49,980][starVLA.training.train_starvla][INFO] - Step 295, Loss: {'train/loss': 0.3308519423007965, 'train/grad_norm_pre_clip': 5.5354695320129395, 'timing/data': 0.005054896930232644, 'timing/model': 1.1259385240264237, 'learning_rate/qwen_vl_interface': 9.962925555160919e-06, 'learning_rate/action_model': 9.962925555160918e-05, 'epoch': 0.61}) +[2026-06-19 10:21:54,471][starVLA.training.train_starvla][INFO] - Step 296, Loss: {'train/loss': 0.33779141306877136, 'train/grad_norm_pre_clip': 8.665596008300781, 'timing/data': 0.00519361300393939, 'timing/model': 1.12275059602689, 'learning_rate/qwen_vl_interface': 9.96254483124377e-06, 'learning_rate/action_model': 9.96254483124377e-05, 'epoch': 0.61}) +[2026-06-19 10:21:58,939][starVLA.training.train_starvla][INFO] - Step 297, Loss: {'train/loss': 0.3276204764842987, 'train/grad_norm_pre_clip': 5.174367904663086, 'timing/data': 0.0049636089242994785, 'timing/model': 1.278171677957289, 'learning_rate/qwen_vl_interface': 9.96216217017745e-06, 'learning_rate/action_model': 9.962162170177451e-05, 'epoch': 0.61}) +[2026-06-19 10:22:03,253][starVLA.training.train_starvla][INFO] - Step 298, Loss: {'train/loss': 0.20747916400432587, 'train/grad_norm_pre_clip': 9.372026443481445, 'timing/data': 0.0047819060273468494, 'timing/model': 1.123486293014139, 'learning_rate/qwen_vl_interface': 9.961777572119258e-06, 'learning_rate/action_model': 9.961777572119258e-05, 'epoch': 0.61}) +[2026-06-19 10:22:07,755][starVLA.training.train_starvla][INFO] - Step 299, Loss: {'train/loss': 0.3275131285190582, 'train/grad_norm_pre_clip': 5.5572991371154785, 'timing/data': 0.005524665000848472, 'timing/model': 1.1298887309385464, 'learning_rate/qwen_vl_interface': 9.961391037227286e-06, 'learning_rate/action_model': 9.961391037227284e-05, 'epoch': 0.61}) +[2026-06-19 10:22:12,186][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 300 +[2026-06-19 10:25:10,545][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 300 +[2026-06-19 10:25:10,546][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 300 +[2026-06-19 10:28:17,027][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 300 +[2026-06-19 10:28:17,028][starVLA.training.train_starvla][INFO] - Step 300, Loss: {'train/loss': 0.38317471742630005, 'train/grad_norm_pre_clip': 8.070767402648926, 'eval/loss': 0.33237898349761963, 'eval/latency_0/loss': 0.1535312980413437, 'eval/latency_1/loss': 0.24860192835330963, 'eval/latency_2/loss': 0.3847399055957794, 'eval/latency_3/loss': 0.390020489692688, 'eval/latency_4/loss': 0.47601014375686646, 'eval/flappy/loss': 0.33237898349761963, 'eval/flappy/latency_0/loss': 0.1535312980413437, 'eval/flappy/latency_1/loss': 0.24860192835330963, 'eval/flappy/latency_2/loss': 0.3847399055957794, 'eval/flappy/latency_3/loss': 0.390020489692688, 'eval/flappy/latency_4/loss': 0.47601014375686646, 'eval/flap_precision': 0.6291666666666667, 'eval/flap_recall': 0.5490909090909091, 'eval/flap_f1': 0.5864077669902912, 'eval/flap_rate_model': 0.06666666666666667, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.5864077669902912, 'eval/latency_0/flap_precision': 0.6291666666666667, 'eval/latency_0/flap_recall': 0.5490909090909091, 'eval/latency_0/flap_f1': 0.5864077669902912, 'eval/latency_0/flap_rate_model': 0.06666666666666667, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.5864077669902912, 'timing/data': 0.005115812993608415, 'timing/model': 1.120276011992246, 'learning_rate/qwen_vl_interface': 9.961002565660419e-06, 'learning_rate/action_model': 9.961002565660419e-05, 'epoch': 0.62}) +[2026-06-19 10:28:17,028][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_300_state +[2026-06-19 10:28:17,030][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-19 10:28:20,429][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_300_state/model.safetensors +[2026-06-19 10:28:27,722][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_300_state/optimizer.bin +[2026-06-19 10:28:27,723][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_300_state/sampler.bin +[2026-06-19 10:28:27,723][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_300_state/sampler_1.bin +[2026-06-19 10:28:27,723][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_300_state/random_states_0.pkl +[2026-06-19 10:28:27,745][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 300 +[2026-06-19 10:28:27,745][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-19 10:28:27,761][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-19 10:28:32,172][starVLA.training.train_starvla][INFO] - Step 301, Loss: {'train/loss': 0.22557607293128967, 'train/grad_norm_pre_clip': 5.070694923400879, 'timing/data': 0.005443737958557904, 'timing/model': 1.1040465619880706, 'learning_rate/qwen_vl_interface': 9.96061215757835e-06, 'learning_rate/action_model': 9.960612157578349e-05, 'epoch': 0.62}) +[2026-06-19 10:28:36,612][starVLA.training.train_starvla][INFO] - Step 302, Loss: {'train/loss': 0.3301568925380707, 'train/grad_norm_pre_clip': 6.580441951751709, 'timing/data': 0.005164372967556119, 'timing/model': 1.1048571080900729, 'learning_rate/qwen_vl_interface': 9.96021981314156e-06, 'learning_rate/action_model': 9.96021981314156e-05, 'epoch': 0.62}) +[2026-06-19 10:28:41,025][starVLA.training.train_starvla][INFO] - Step 303, Loss: {'train/loss': 0.3262524902820587, 'train/grad_norm_pre_clip': 8.243804931640625, 'timing/data': 0.005306114093400538, 'timing/model': 1.262218129937537, 'learning_rate/qwen_vl_interface': 9.959825532511322e-06, 'learning_rate/action_model': 9.959825532511322e-05, 'epoch': 0.62}) +[2026-06-19 10:28:45,292][starVLA.training.train_starvla][INFO] - Step 304, Loss: {'train/loss': 0.14752385020256042, 'train/grad_norm_pre_clip': 3.1945290565490723, 'timing/data': 0.0049525980139151216, 'timing/model': 1.1033174340846017, 'learning_rate/qwen_vl_interface': 9.959429315849715e-06, 'learning_rate/action_model': 9.959429315849715e-05, 'epoch': 0.62}) +[2026-06-19 10:28:49,693][starVLA.training.train_starvla][INFO] - Step 305, Loss: {'train/loss': 0.2543475925922394, 'train/grad_norm_pre_clip': 9.957650184631348, 'timing/data': 0.005178737919777632, 'timing/model': 1.1035178940510377, 'learning_rate/qwen_vl_interface': 9.959031163319606e-06, 'learning_rate/action_model': 9.959031163319607e-05, 'epoch': 0.63}) +[2026-06-19 10:28:54,128][starVLA.training.train_starvla][INFO] - Step 306, Loss: {'train/loss': 0.3842732906341553, 'train/grad_norm_pre_clip': 9.260196685791016, 'timing/data': 0.00481899396982044, 'timing/model': 1.1087995370617136, 'learning_rate/qwen_vl_interface': 9.958631075084664e-06, 'learning_rate/action_model': 9.958631075084664e-05, 'epoch': 0.63}) +[2026-06-19 10:28:58,575][starVLA.training.train_starvla][INFO] - Step 307, Loss: {'train/loss': 0.3275666832923889, 'train/grad_norm_pre_clip': 8.621124267578125, 'timing/data': 0.004962400998920202, 'timing/model': 1.1132050709566101, 'learning_rate/qwen_vl_interface': 9.958229051309346e-06, 'learning_rate/action_model': 9.958229051309346e-05, 'epoch': 0.63}) +[2026-06-19 10:29:03,005][starVLA.training.train_starvla][INFO] - Step 308, Loss: {'train/loss': 0.45452263951301575, 'train/grad_norm_pre_clip': 8.333917617797852, 'timing/data': 0.00506318302359432, 'timing/model': 1.2679410279961303, 'learning_rate/qwen_vl_interface': 9.95782509215891e-06, 'learning_rate/action_model': 9.95782509215891e-05, 'epoch': 0.63}) +[2026-06-19 10:29:07,279][starVLA.training.train_starvla][INFO] - Step 309, Loss: {'train/loss': 0.2578601837158203, 'train/grad_norm_pre_clip': 15.202837944030762, 'timing/data': 0.004987630993127823, 'timing/model': 1.1090605369536206, 'learning_rate/qwen_vl_interface': 9.95741919779941e-06, 'learning_rate/action_model': 9.95741919779941e-05, 'epoch': 0.63}) +[2026-06-19 10:29:11,733][starVLA.training.train_starvla][INFO] - Step 310, Loss: {'train/loss': 0.4499727189540863, 'train/grad_norm_pre_clip': 7.373553276062012, 'timing/data': 0.005291519919410348, 'timing/model': 1.1111039500683546, 'learning_rate/qwen_vl_interface': 9.957011368397692e-06, 'learning_rate/action_model': 9.95701136839769e-05, 'epoch': 0.64}) +[2026-06-19 10:29:16,078][starVLA.training.train_starvla][INFO] - Step 311, Loss: {'train/loss': 0.3596418797969818, 'train/grad_norm_pre_clip': 8.968466758728027, 'timing/data': 0.00501344702206552, 'timing/model': 1.1163713249843568, 'learning_rate/qwen_vl_interface': 9.956601604121398e-06, 'learning_rate/action_model': 9.956601604121397e-05, 'epoch': 0.64}) +[2026-06-19 10:29:20,525][starVLA.training.train_starvla][INFO] - Step 312, Loss: {'train/loss': 0.20444439351558685, 'train/grad_norm_pre_clip': 6.704423904418945, 'timing/data': 0.005215447978116572, 'timing/model': 1.1205542869865894, 'learning_rate/qwen_vl_interface': 9.95618990513897e-06, 'learning_rate/action_model': 9.95618990513897e-05, 'epoch': 0.64}) +[2026-06-19 10:29:24,991][starVLA.training.train_starvla][INFO] - Step 313, Loss: {'train/loss': 0.25663095712661743, 'train/grad_norm_pre_clip': 5.043735504150391, 'timing/data': 0.005048391059972346, 'timing/model': 1.115579926990904, 'learning_rate/qwen_vl_interface': 9.95577627161964e-06, 'learning_rate/action_model': 9.95577627161964e-05, 'epoch': 0.64}) +[2026-06-19 10:29:29,459][starVLA.training.train_starvla][INFO] - Step 314, Loss: {'train/loss': 0.2690507471561432, 'train/grad_norm_pre_clip': 4.925378799438477, 'timing/data': 0.0053812540136277676, 'timing/model': 1.2889751219190657, 'learning_rate/qwen_vl_interface': 9.955360703733438e-06, 'learning_rate/action_model': 9.955360703733437e-05, 'epoch': 0.64}) +[2026-06-19 10:29:33,892][starVLA.training.train_starvla][INFO] - Step 315, Loss: {'train/loss': 0.30503687262535095, 'train/grad_norm_pre_clip': 4.454595565795898, 'timing/data': 0.0046908879885450006, 'timing/model': 1.2414090889506042, 'learning_rate/qwen_vl_interface': 9.954943201651187e-06, 'learning_rate/action_model': 9.954943201651187e-05, 'epoch': 0.65}) +[2026-06-19 10:29:38,181][starVLA.training.train_starvla][INFO] - Step 316, Loss: {'train/loss': 0.2646162807941437, 'train/grad_norm_pre_clip': 5.107668876647949, 'timing/data': 0.005398507928475738, 'timing/model': 1.1125961189391091, 'learning_rate/qwen_vl_interface': 9.954523765544507e-06, 'learning_rate/action_model': 9.954523765544508e-05, 'epoch': 0.65}) +[2026-06-19 10:29:42,649][starVLA.training.train_starvla][INFO] - Step 317, Loss: {'train/loss': 0.4016363024711609, 'train/grad_norm_pre_clip': 7.909402847290039, 'timing/data': 0.0052142629865556955, 'timing/model': 1.1279929840238765, 'learning_rate/qwen_vl_interface': 9.954102395585814e-06, 'learning_rate/action_model': 9.954102395585813e-05, 'epoch': 0.65}) +[2026-06-19 10:29:47,135][starVLA.training.train_starvla][INFO] - Step 318, Loss: {'train/loss': 0.3074420988559723, 'train/grad_norm_pre_clip': 4.904382228851318, 'timing/data': 0.0055239570792764425, 'timing/model': 1.118818359915167, 'learning_rate/qwen_vl_interface': 9.953679091948313e-06, 'learning_rate/action_model': 9.953679091948314e-05, 'epoch': 0.65}) +[2026-06-19 10:29:51,613][starVLA.training.train_starvla][INFO] - Step 319, Loss: {'train/loss': 0.15551045536994934, 'train/grad_norm_pre_clip': 4.02200174331665, 'timing/data': 0.004703272017650306, 'timing/model': 1.121515145059675, 'learning_rate/qwen_vl_interface': 9.953253854806013e-06, 'learning_rate/action_model': 9.953253854806012e-05, 'epoch': 0.66}) +[2026-06-19 10:29:56,098][starVLA.training.train_starvla][INFO] - Step 320, Loss: {'train/loss': 0.21853189170360565, 'train/grad_norm_pre_clip': 3.409451723098755, 'timing/data': 0.005352825974114239, 'timing/model': 1.2957722099963576, 'learning_rate/qwen_vl_interface': 9.95282668433371e-06, 'learning_rate/action_model': 9.952826684333709e-05, 'epoch': 0.66}) +[2026-06-19 10:30:00,416][starVLA.training.train_starvla][INFO] - Step 321, Loss: {'train/loss': 0.3012697398662567, 'train/grad_norm_pre_clip': 7.5490007400512695, 'timing/data': 0.005331136053428054, 'timing/model': 1.1204711510799825, 'learning_rate/qwen_vl_interface': 9.952397580706998e-06, 'learning_rate/action_model': 9.952397580706998e-05, 'epoch': 0.66}) +[2026-06-19 10:30:04,888][starVLA.training.train_starvla][INFO] - Step 322, Loss: {'train/loss': 0.1921246498823166, 'train/grad_norm_pre_clip': 6.2049031257629395, 'timing/data': 0.004720723023638129, 'timing/model': 1.1167974929558113, 'learning_rate/qwen_vl_interface': 9.951966544102267e-06, 'learning_rate/action_model': 9.951966544102266e-05, 'epoch': 0.66}) +[2026-06-19 10:30:09,350][starVLA.training.train_starvla][INFO] - Step 323, Loss: {'train/loss': 0.2743440270423889, 'train/grad_norm_pre_clip': 6.247211933135986, 'timing/data': 0.005040301010012627, 'timing/model': 1.1164042509626597, 'learning_rate/qwen_vl_interface': 9.951533574696696e-06, 'learning_rate/action_model': 9.951533574696697e-05, 'epoch': 0.66}) +[2026-06-19 10:30:13,797][starVLA.training.train_starvla][INFO] - Step 324, Loss: {'train/loss': 0.275617390871048, 'train/grad_norm_pre_clip': 8.289904594421387, 'timing/data': 0.005419296910986304, 'timing/model': 1.1148284780792892, 'learning_rate/qwen_vl_interface': 9.951098672668266e-06, 'learning_rate/action_model': 9.951098672668265e-05, 'epoch': 0.67}) +[2026-06-19 10:30:18,212][starVLA.training.train_starvla][INFO] - Step 325, Loss: {'train/loss': 0.23851674795150757, 'train/grad_norm_pre_clip': 4.900969982147217, 'timing/data': 0.005008865031413734, 'timing/model': 1.1136287780245766, 'learning_rate/qwen_vl_interface': 9.95066183819575e-06, 'learning_rate/action_model': 9.950661838195749e-05, 'epoch': 0.67}) +[2026-06-19 10:30:22,666][starVLA.training.train_starvla][INFO] - Step 326, Loss: {'train/loss': 0.2602548897266388, 'train/grad_norm_pre_clip': 15.221790313720703, 'timing/data': 0.005223914980888367, 'timing/model': 1.271340514998883, 'learning_rate/qwen_vl_interface': 9.95022307145871e-06, 'learning_rate/action_model': 9.950223071458708e-05, 'epoch': 0.67}) +[2026-06-19 10:30:26,980][starVLA.training.train_starvla][INFO] - Step 327, Loss: {'train/loss': 0.18757592141628265, 'train/grad_norm_pre_clip': 5.029659748077393, 'timing/data': 0.0047064319951459765, 'timing/model': 1.1275044010253623, 'learning_rate/qwen_vl_interface': 9.949782372637507e-06, 'learning_rate/action_model': 9.949782372637506e-05, 'epoch': 0.67}) +[2026-06-19 10:30:31,444][starVLA.training.train_starvla][INFO] - Step 328, Loss: {'train/loss': 0.2976171672344208, 'train/grad_norm_pre_clip': 6.224745750427246, 'timing/data': 0.005088101024739444, 'timing/model': 1.1142917909892276, 'learning_rate/qwen_vl_interface': 9.949339741913297e-06, 'learning_rate/action_model': 9.949339741913297e-05, 'epoch': 0.67}) +[2026-06-19 10:30:35,931][starVLA.training.train_starvla][INFO] - Step 329, Loss: {'train/loss': 0.32475271821022034, 'train/grad_norm_pre_clip': 13.628182411193848, 'timing/data': 0.005044965073466301, 'timing/model': 1.1184903379762545, 'learning_rate/qwen_vl_interface': 9.948895179468032e-06, 'learning_rate/action_model': 9.948895179468031e-05, 'epoch': 0.68}) +[2026-06-19 10:30:40,405][starVLA.training.train_starvla][INFO] - Step 330, Loss: {'train/loss': 0.512362539768219, 'train/grad_norm_pre_clip': 13.342552185058594, 'timing/data': 0.005259708967059851, 'timing/model': 1.1252119990531355, 'learning_rate/qwen_vl_interface': 9.948448685484448e-06, 'learning_rate/action_model': 9.948448685484447e-05, 'epoch': 0.68}) +[2026-06-19 10:30:44,838][starVLA.training.train_starvla][INFO] - Step 331, Loss: {'train/loss': 0.5187178254127502, 'train/grad_norm_pre_clip': 11.45547866821289, 'timing/data': 0.004946873988956213, 'timing/model': 1.1300500939833, 'learning_rate/qwen_vl_interface': 9.948000260146087e-06, 'learning_rate/action_model': 9.948000260146087e-05, 'epoch': 0.68}) +[2026-06-19 10:30:49,333][starVLA.training.train_starvla][INFO] - Step 332, Loss: {'train/loss': 0.16727770864963531, 'train/grad_norm_pre_clip': 6.39304780960083, 'timing/data': 0.004758390015922487, 'timing/model': 1.3120493669994175, 'learning_rate/qwen_vl_interface': 9.947549903637277e-06, 'learning_rate/action_model': 9.947549903637278e-05, 'epoch': 0.68}) +[2026-06-19 10:30:53,633][starVLA.training.train_starvla][INFO] - Step 333, Loss: {'train/loss': 0.21025289595127106, 'train/grad_norm_pre_clip': 6.02844762802124, 'timing/data': 0.00499845203012228, 'timing/model': 1.1160692140692845, 'learning_rate/qwen_vl_interface': 9.947097616143145e-06, 'learning_rate/action_model': 9.947097616143144e-05, 'epoch': 0.68}) +[2026-06-19 10:30:58,094][starVLA.training.train_starvla][INFO] - Step 334, Loss: {'train/loss': 0.5158708691596985, 'train/grad_norm_pre_clip': 14.895334243774414, 'timing/data': 0.0052580880001187325, 'timing/model': 1.1187916930066422, 'learning_rate/qwen_vl_interface': 9.946643397849607e-06, 'learning_rate/action_model': 9.946643397849607e-05, 'epoch': 0.69}) +[2026-06-19 10:31:02,602][starVLA.training.train_starvla][INFO] - Step 335, Loss: {'train/loss': 0.32164815068244934, 'train/grad_norm_pre_clip': 7.87169075012207, 'timing/data': 0.005030813976190984, 'timing/model': 1.1201846050098538, 'learning_rate/qwen_vl_interface': 9.946187248943375e-06, 'learning_rate/action_model': 9.946187248943374e-05, 'epoch': 0.69}) +[2026-06-19 10:31:07,040][starVLA.training.train_starvla][INFO] - Step 336, Loss: {'train/loss': 0.3214622735977173, 'train/grad_norm_pre_clip': 7.874912261962891, 'timing/data': 0.005291239940561354, 'timing/model': 1.1196141680702567, 'learning_rate/qwen_vl_interface': 9.945729169611957e-06, 'learning_rate/action_model': 9.945729169611958e-05, 'epoch': 0.69}) +[2026-06-19 10:31:11,484][starVLA.training.train_starvla][INFO] - Step 337, Loss: {'train/loss': 0.22940368950366974, 'train/grad_norm_pre_clip': 4.266616344451904, 'timing/data': 0.005094669060781598, 'timing/model': 1.1171982199884951, 'learning_rate/qwen_vl_interface': 9.94526916004365e-06, 'learning_rate/action_model': 9.94526916004365e-05, 'epoch': 0.69}) +[2026-06-19 10:31:15,953][starVLA.training.train_starvla][INFO] - Step 338, Loss: {'train/loss': 0.24311310052871704, 'train/grad_norm_pre_clip': 5.257404327392578, 'timing/data': 0.004707548068836331, 'timing/model': 1.2759286110522225, 'learning_rate/qwen_vl_interface': 9.944807220427549e-06, 'learning_rate/action_model': 9.944807220427548e-05, 'epoch': 0.69}) +[2026-06-19 10:31:20,255][starVLA.training.train_starvla][INFO] - Step 339, Loss: {'train/loss': 0.2838881015777588, 'train/grad_norm_pre_clip': 8.4612398147583, 'timing/data': 0.005188954994082451, 'timing/model': 1.117863396066241, 'learning_rate/qwen_vl_interface': 9.944343350953534e-06, 'learning_rate/action_model': 9.944343350953533e-05, 'epoch': 0.7}) +[2026-06-19 10:31:24,744][starVLA.training.train_starvla][INFO] - Step 340, Loss: {'train/loss': 0.5064233541488647, 'train/grad_norm_pre_clip': 10.914677619934082, 'timing/data': 0.005218129954300821, 'timing/model': 1.123200359987095, 'learning_rate/qwen_vl_interface': 9.94387755181229e-06, 'learning_rate/action_model': 9.94387755181229e-05, 'epoch': 0.7}) +[2026-06-19 10:31:29,203][starVLA.training.train_starvla][INFO] - Step 341, Loss: {'train/loss': 0.2543342113494873, 'train/grad_norm_pre_clip': 5.448864936828613, 'timing/data': 0.00515180395450443, 'timing/model': 1.115349343046546, 'learning_rate/qwen_vl_interface': 9.943409823195289e-06, 'learning_rate/action_model': 9.943409823195288e-05, 'epoch': 0.7}) +[2026-06-19 10:31:33,680][starVLA.training.train_starvla][INFO] - Step 342, Loss: {'train/loss': 0.33356767892837524, 'train/grad_norm_pre_clip': 5.37860107421875, 'timing/data': 0.005077530979178846, 'timing/model': 1.1155478740110993, 'learning_rate/qwen_vl_interface': 9.942940165294794e-06, 'learning_rate/action_model': 9.942940165294793e-05, 'epoch': 0.7}) +[2026-06-19 10:31:38,132][starVLA.training.train_starvla][INFO] - Step 343, Loss: {'train/loss': 0.11120858788490295, 'train/grad_norm_pre_clip': 3.535830020904541, 'timing/data': 0.007251153001561761, 'timing/model': 1.1206754700979218, 'learning_rate/qwen_vl_interface': 9.942468578303865e-06, 'learning_rate/action_model': 9.942468578303865e-05, 'epoch': 0.7}) +[2026-06-19 10:31:42,607][starVLA.training.train_starvla][INFO] - Step 344, Loss: {'train/loss': 0.19934718310832977, 'train/grad_norm_pre_clip': 6.5880842208862305, 'timing/data': 0.0047941820230335, 'timing/model': 1.2835055419709533, 'learning_rate/qwen_vl_interface': 9.941995062416354e-06, 'learning_rate/action_model': 9.941995062416354e-05, 'epoch': 0.71}) +[2026-06-19 10:31:46,929][starVLA.training.train_starvla][INFO] - Step 345, Loss: {'train/loss': 0.37941235303878784, 'train/grad_norm_pre_clip': 12.308234214782715, 'timing/data': 0.005236444994807243, 'timing/model': 1.1301105859456584, 'learning_rate/qwen_vl_interface': 9.941519617826906e-06, 'learning_rate/action_model': 9.941519617826906e-05, 'epoch': 0.71}) +[2026-06-19 10:31:51,412][starVLA.training.train_starvla][INFO] - Step 346, Loss: {'train/loss': 0.17452087998390198, 'train/grad_norm_pre_clip': 8.434645652770996, 'timing/data': 0.005096527980640531, 'timing/model': 1.1214866050286219, 'learning_rate/qwen_vl_interface': 9.941042244730954e-06, 'learning_rate/action_model': 9.941042244730955e-05, 'epoch': 0.71}) +[2026-06-19 10:31:55,898][starVLA.training.train_starvla][INFO] - Step 347, Loss: {'train/loss': 0.3261064291000366, 'train/grad_norm_pre_clip': 17.483856201171875, 'timing/data': 0.005277693970128894, 'timing/model': 1.1180817779386416, 'learning_rate/qwen_vl_interface': 9.940562943324736e-06, 'learning_rate/action_model': 9.940562943324735e-05, 'epoch': 0.71}) +[2026-06-19 10:32:00,363][starVLA.training.train_starvla][INFO] - Step 348, Loss: {'train/loss': 0.6140510439872742, 'train/grad_norm_pre_clip': 40.107200622558594, 'timing/data': 0.005501136998645961, 'timing/model': 1.136881956947036, 'learning_rate/qwen_vl_interface': 9.940081713805266e-06, 'learning_rate/action_model': 9.940081713805266e-05, 'epoch': 0.71}) +[2026-06-19 10:32:04,864][starVLA.training.train_starvla][INFO] - Step 349, Loss: {'train/loss': 0.2837370038032532, 'train/grad_norm_pre_clip': 13.206539154052734, 'timing/data': 0.005235944990999997, 'timing/model': 1.1284997410839424, 'learning_rate/qwen_vl_interface': 9.939598556370366e-06, 'learning_rate/action_model': 9.939598556370365e-05, 'epoch': 0.72}) +[2026-06-19 10:32:09,367][starVLA.training.train_starvla][INFO] - Step 350, Loss: {'train/loss': 0.1476072371006012, 'train/grad_norm_pre_clip': 5.324872970581055, 'timing/data': 0.005169381038285792, 'timing/model': 1.2964429999701679, 'learning_rate/qwen_vl_interface': 9.93911347121864e-06, 'learning_rate/action_model': 9.939113471218639e-05, 'epoch': 0.72}) +[2026-06-19 10:32:13,695][starVLA.training.train_starvla][INFO] - Step 351, Loss: {'train/loss': 0.27409660816192627, 'train/grad_norm_pre_clip': 13.46188735961914, 'timing/data': 0.005396277061663568, 'timing/model': 1.1275838590227067, 'learning_rate/qwen_vl_interface': 9.938626458549488e-06, 'learning_rate/action_model': 9.938626458549488e-05, 'epoch': 0.72}) +[2026-06-19 10:32:18,189][starVLA.training.train_starvla][INFO] - Step 352, Loss: {'train/loss': 0.335579514503479, 'train/grad_norm_pre_clip': 8.976510047912598, 'timing/data': 0.005150355049408972, 'timing/model': 1.1188243710203096, 'learning_rate/qwen_vl_interface': 9.938137518563106e-06, 'learning_rate/action_model': 9.938137518563106e-05, 'epoch': 0.72}) +[2026-06-19 10:32:22,608][starVLA.training.train_starvla][INFO] - Step 353, Loss: {'train/loss': 0.16468900442123413, 'train/grad_norm_pre_clip': 5.106706142425537, 'timing/data': 0.00522031111177057, 'timing/model': 1.124042922980152, 'learning_rate/qwen_vl_interface': 9.937646651460475e-06, 'learning_rate/action_model': 9.937646651460474e-05, 'epoch': 0.73}) +[2026-06-19 10:32:27,077][starVLA.training.train_starvla][INFO] - Step 354, Loss: {'train/loss': 0.2636086046695709, 'train/grad_norm_pre_clip': 5.827272891998291, 'timing/data': 0.005620189011096954, 'timing/model': 1.1179634029977024, 'learning_rate/qwen_vl_interface': 9.937153857443375e-06, 'learning_rate/action_model': 9.937153857443374e-05, 'epoch': 0.73}) +[2026-06-19 10:32:31,546][starVLA.training.train_starvla][INFO] - Step 355, Loss: {'train/loss': 0.37415847182273865, 'train/grad_norm_pre_clip': 7.973979473114014, 'timing/data': 0.004884609952569008, 'timing/model': 1.1185172660043463, 'learning_rate/qwen_vl_interface': 9.936659136714372e-06, 'learning_rate/action_model': 9.936659136714371e-05, 'epoch': 0.73}) +[2026-06-19 10:32:36,029][starVLA.training.train_starvla][INFO] - Step 356, Loss: {'train/loss': 0.2914782166481018, 'train/grad_norm_pre_clip': 5.298899173736572, 'timing/data': 0.005085351062007248, 'timing/model': 1.2962364989798516, 'learning_rate/qwen_vl_interface': 9.936162489476827e-06, 'learning_rate/action_model': 9.936162489476828e-05, 'epoch': 0.73}) +[2026-06-19 10:32:40,381][starVLA.training.train_starvla][INFO] - Step 357, Loss: {'train/loss': 0.19802893698215485, 'train/grad_norm_pre_clip': 4.879251003265381, 'timing/data': 0.0051979070995002985, 'timing/model': 1.1279718540608883, 'learning_rate/qwen_vl_interface': 9.935663915934899e-06, 'learning_rate/action_model': 9.935663915934898e-05, 'epoch': 0.73}) +[2026-06-19 10:32:44,870][starVLA.training.train_starvla][INFO] - Step 358, Loss: {'train/loss': 0.22605150938034058, 'train/grad_norm_pre_clip': 4.5174455642700195, 'timing/data': 0.005309694912284613, 'timing/model': 1.135903924005106, 'learning_rate/qwen_vl_interface': 9.935163416293525e-06, 'learning_rate/action_model': 9.935163416293525e-05, 'epoch': 0.74}) +[2026-06-19 10:32:49,340][starVLA.training.train_starvla][INFO] - Step 359, Loss: {'train/loss': 0.22850225865840912, 'train/grad_norm_pre_clip': 6.195747375488281, 'timing/data': 0.004783611977472901, 'timing/model': 1.1356979029951617, 'learning_rate/qwen_vl_interface': 9.934660990758445e-06, 'learning_rate/action_model': 9.934660990758445e-05, 'epoch': 0.74}) +[2026-06-19 10:32:53,803][starVLA.training.train_starvla][INFO] - Step 360, Loss: {'train/loss': 0.10880383104085922, 'train/grad_norm_pre_clip': 1.9802405834197998, 'timing/data': 0.005100387032143772, 'timing/model': 1.1259954810375348, 'learning_rate/qwen_vl_interface': 9.934156639536186e-06, 'learning_rate/action_model': 9.934156639536187e-05, 'epoch': 0.74}) +[2026-06-19 10:32:58,299][starVLA.training.train_starvla][INFO] - Step 361, Loss: {'train/loss': 0.2518709599971771, 'train/grad_norm_pre_clip': 6.1059889793396, 'timing/data': 0.0051945490995422006, 'timing/model': 1.1282459719805047, 'learning_rate/qwen_vl_interface': 9.93365036283407e-06, 'learning_rate/action_model': 9.93365036283407e-05, 'epoch': 0.74}) +[2026-06-19 10:33:02,772][starVLA.training.train_starvla][INFO] - Step 362, Loss: {'train/loss': 0.34417083859443665, 'train/grad_norm_pre_clip': 4.451842784881592, 'timing/data': 0.005445364047773182, 'timing/model': 1.2752109779976308, 'learning_rate/qwen_vl_interface': 9.933142160860208e-06, 'learning_rate/action_model': 9.933142160860208e-05, 'epoch': 0.74}) +[2026-06-19 10:33:07,084][starVLA.training.train_starvla][INFO] - Step 363, Loss: {'train/loss': 0.47857335209846497, 'train/grad_norm_pre_clip': 11.21125316619873, 'timing/data': 0.0051633259281516075, 'timing/model': 1.1261090940097347, 'learning_rate/qwen_vl_interface': 9.932632033823499e-06, 'learning_rate/action_model': 9.932632033823499e-05, 'epoch': 0.75}) +[2026-06-19 10:33:11,558][starVLA.training.train_starvla][INFO] - Step 364, Loss: {'train/loss': 0.411033034324646, 'train/grad_norm_pre_clip': 8.25912857055664, 'timing/data': 0.005134453997015953, 'timing/model': 1.1257202220149338, 'learning_rate/qwen_vl_interface': 9.93211998193364e-06, 'learning_rate/action_model': 9.932119981933639e-05, 'epoch': 0.75}) +[2026-06-19 10:33:16,011][starVLA.training.train_starvla][INFO] - Step 365, Loss: {'train/loss': 0.4029126763343811, 'train/grad_norm_pre_clip': 8.27713394165039, 'timing/data': 0.004691314999945462, 'timing/model': 1.1197691060369834, 'learning_rate/qwen_vl_interface': 9.931606005401117e-06, 'learning_rate/action_model': 9.931606005401116e-05, 'epoch': 0.75}) +[2026-06-19 10:33:20,496][starVLA.training.train_starvla][INFO] - Step 366, Loss: {'train/loss': 0.3066301643848419, 'train/grad_norm_pre_clip': 5.509239196777344, 'timing/data': 0.005336240981705487, 'timing/model': 1.1204227759735659, 'learning_rate/qwen_vl_interface': 9.931090104437202e-06, 'learning_rate/action_model': 9.931090104437202e-05, 'epoch': 0.75}) +[2026-06-19 10:33:24,977][starVLA.training.train_starvla][INFO] - Step 367, Loss: {'train/loss': 0.23017553985118866, 'train/grad_norm_pre_clip': 4.363545894622803, 'timing/data': 0.004666220978833735, 'timing/model': 1.120558499940671, 'learning_rate/qwen_vl_interface': 9.930572279253967e-06, 'learning_rate/action_model': 9.930572279253967e-05, 'epoch': 0.75}) +[2026-06-19 10:33:29,510][starVLA.training.train_starvla][INFO] - Step 368, Loss: {'train/loss': 0.12638632953166962, 'train/grad_norm_pre_clip': 3.3887317180633545, 'timing/data': 0.00532530399505049, 'timing/model': 1.3247074369573966, 'learning_rate/qwen_vl_interface': 9.930052530064269e-06, 'learning_rate/action_model': 9.93005253006427e-05, 'epoch': 0.76}) +[2026-06-19 10:33:33,952][starVLA.training.train_starvla][INFO] - Step 369, Loss: {'train/loss': 0.5814215540885925, 'train/grad_norm_pre_clip': 15.623977661132812, 'timing/data': 0.00495007901918143, 'timing/model': 1.2345835349988192, 'learning_rate/qwen_vl_interface': 9.929530857081757e-06, 'learning_rate/action_model': 9.929530857081757e-05, 'epoch': 0.76}) +[2026-06-19 10:33:38,279][starVLA.training.train_starvla][INFO] - Step 370, Loss: {'train/loss': 0.8287302255630493, 'train/grad_norm_pre_clip': 18.813791275024414, 'timing/data': 0.004965968895703554, 'timing/model': 1.120073840022087, 'learning_rate/qwen_vl_interface': 9.92900726052087e-06, 'learning_rate/action_model': 9.929007260520869e-05, 'epoch': 0.76}) +[2026-06-19 10:33:42,750][starVLA.training.train_starvla][INFO] - Step 371, Loss: {'train/loss': 0.4958080053329468, 'train/grad_norm_pre_clip': 15.348761558532715, 'timing/data': 0.005073745967820287, 'timing/model': 1.1281218910589814, 'learning_rate/qwen_vl_interface': 9.92848174059684e-06, 'learning_rate/action_model': 9.92848174059684e-05, 'epoch': 0.76}) +[2026-06-19 10:33:47,239][starVLA.training.train_starvla][INFO] - Step 372, Loss: {'train/loss': 0.3678101599216461, 'train/grad_norm_pre_clip': 6.41558313369751, 'timing/data': 0.005020408076234162, 'timing/model': 1.1191212281119078, 'learning_rate/qwen_vl_interface': 9.92795429752569e-06, 'learning_rate/action_model': 9.927954297525689e-05, 'epoch': 0.76}) +[2026-06-19 10:33:51,718][starVLA.training.train_starvla][INFO] - Step 373, Loss: {'train/loss': 0.54469895362854, 'train/grad_norm_pre_clip': 13.875785827636719, 'timing/data': 0.004921430954709649, 'timing/model': 1.1200821209931746, 'learning_rate/qwen_vl_interface': 9.927424931524228e-06, 'learning_rate/action_model': 9.927424931524228e-05, 'epoch': 0.77}) +[2026-06-19 10:33:56,218][starVLA.training.train_starvla][INFO] - Step 374, Loss: {'train/loss': 0.23034586012363434, 'train/grad_norm_pre_clip': 7.277536869049072, 'timing/data': 0.004948754096403718, 'timing/model': 1.294704559026286, 'learning_rate/qwen_vl_interface': 9.926893642810058e-06, 'learning_rate/action_model': 9.926893642810059e-05, 'epoch': 0.77}) +[2026-06-19 10:34:00,659][starVLA.training.train_starvla][INFO] - Step 375, Loss: {'train/loss': 0.48056548833847046, 'train/grad_norm_pre_clip': 6.536869049072266, 'timing/data': 0.0051378849893808365, 'timing/model': 1.2361802639206871, 'learning_rate/qwen_vl_interface': 9.926360431601575e-06, 'learning_rate/action_model': 9.926360431601574e-05, 'epoch': 0.77}) +[2026-06-19 10:34:04,984][starVLA.training.train_starvla][INFO] - Step 376, Loss: {'train/loss': 0.22981736063957214, 'train/grad_norm_pre_clip': 7.117458820343018, 'timing/data': 0.005375670036301017, 'timing/model': 1.1285220340359956, 'learning_rate/qwen_vl_interface': 9.92582529811796e-06, 'learning_rate/action_model': 9.925825298117959e-05, 'epoch': 0.77}) +[2026-06-19 10:34:09,450][starVLA.training.train_starvla][INFO] - Step 377, Loss: {'train/loss': 0.3100368082523346, 'train/grad_norm_pre_clip': 7.865536212921143, 'timing/data': 0.005077050998806953, 'timing/model': 1.1220191069878638, 'learning_rate/qwen_vl_interface': 9.925288242579184e-06, 'learning_rate/action_model': 9.925288242579184e-05, 'epoch': 0.77}) +[2026-06-19 10:34:13,932][starVLA.training.train_starvla][INFO] - Step 378, Loss: {'train/loss': 0.3775668442249298, 'train/grad_norm_pre_clip': 15.9486722946167, 'timing/data': 0.005384707939811051, 'timing/model': 1.1169656820129603, 'learning_rate/qwen_vl_interface': 9.924749265206015e-06, 'learning_rate/action_model': 9.924749265206015e-05, 'epoch': 0.78}) +[2026-06-19 10:34:18,409][starVLA.training.train_starvla][INFO] - Step 379, Loss: {'train/loss': 0.2903786897659302, 'train/grad_norm_pre_clip': 7.067559242248535, 'timing/data': 0.005321695003658533, 'timing/model': 1.118270116043277, 'learning_rate/qwen_vl_interface': 9.924208366220003e-06, 'learning_rate/action_model': 9.924208366220002e-05, 'epoch': 0.78}) +[2026-06-19 10:34:22,896][starVLA.training.train_starvla][INFO] - Step 380, Loss: {'train/loss': 0.2736535966396332, 'train/grad_norm_pre_clip': 3.5004024505615234, 'timing/data': 0.005007058032788336, 'timing/model': 1.2891840420197695, 'learning_rate/qwen_vl_interface': 9.923665545843492e-06, 'learning_rate/action_model': 9.923665545843491e-05, 'epoch': 0.78}) +[2026-06-19 10:34:27,218][starVLA.training.train_starvla][INFO] - Step 381, Loss: {'train/loss': 0.2392391711473465, 'train/grad_norm_pre_clip': 5.154240608215332, 'timing/data': 0.0051302199717611074, 'timing/model': 1.1247621190268546, 'learning_rate/qwen_vl_interface': 9.923120804299615e-06, 'learning_rate/action_model': 9.923120804299613e-05, 'epoch': 0.78}) +[2026-06-19 10:34:31,695][starVLA.training.train_starvla][INFO] - Step 382, Loss: {'train/loss': 0.18430620431900024, 'train/grad_norm_pre_clip': 3.3959643840789795, 'timing/data': 0.004860613029450178, 'timing/model': 1.1201588220428675, 'learning_rate/qwen_vl_interface': 9.922574141812295e-06, 'learning_rate/action_model': 9.922574141812294e-05, 'epoch': 0.78}) +[2026-06-19 10:34:36,179][starVLA.training.train_starvla][INFO] - Step 383, Loss: {'train/loss': 0.2046404629945755, 'train/grad_norm_pre_clip': 2.9509918689727783, 'timing/data': 0.005067622987553477, 'timing/model': 1.1193602910498157, 'learning_rate/qwen_vl_interface': 9.922025558606242e-06, 'learning_rate/action_model': 9.922025558606242e-05, 'epoch': 0.79}) +[2026-06-19 10:34:40,674][starVLA.training.train_starvla][INFO] - Step 384, Loss: {'train/loss': 0.20063816010951996, 'train/grad_norm_pre_clip': 6.314820289611816, 'timing/data': 0.005188573966734111, 'timing/model': 1.1361619390081614, 'learning_rate/qwen_vl_interface': 9.92147505490696e-06, 'learning_rate/action_model': 9.92147505490696e-05, 'epoch': 0.79}) +[2026-06-19 10:34:45,084][starVLA.training.train_starvla][INFO] - Step 385, Loss: {'train/loss': 0.45683813095092773, 'train/grad_norm_pre_clip': 11.708526611328125, 'timing/data': 0.005292642046697438, 'timing/model': 1.1201334970537573, 'learning_rate/qwen_vl_interface': 9.920922630940741e-06, 'learning_rate/action_model': 9.920922630940742e-05, 'epoch': 0.79}) +[2026-06-19 10:34:49,521][starVLA.training.train_starvla][INFO] - Step 386, Loss: {'train/loss': 0.33133816719055176, 'train/grad_norm_pre_clip': 12.058344841003418, 'timing/data': 0.005157153937034309, 'timing/model': 1.2405617820331827, 'learning_rate/qwen_vl_interface': 9.920368286934664e-06, 'learning_rate/action_model': 9.920368286934663e-05, 'epoch': 0.79}) +[2026-06-19 10:34:53,839][starVLA.training.train_starvla][INFO] - Step 387, Loss: {'train/loss': 0.10904251039028168, 'train/grad_norm_pre_clip': 6.104498386383057, 'timing/data': 0.005125847994349897, 'timing/model': 1.1193591949995607, 'learning_rate/qwen_vl_interface': 9.9198120231166e-06, 'learning_rate/action_model': 9.9198120231166e-05, 'epoch': 0.79}) +[2026-06-19 10:34:58,311][starVLA.training.train_starvla][INFO] - Step 388, Loss: {'train/loss': 0.4477650821208954, 'train/grad_norm_pre_clip': 14.328110694885254, 'timing/data': 0.004860128974542022, 'timing/model': 1.1170098700094968, 'learning_rate/qwen_vl_interface': 9.919253839715208e-06, 'learning_rate/action_model': 9.919253839715208e-05, 'epoch': 0.8}) +[2026-06-19 10:35:02,794][starVLA.training.train_starvla][INFO] - Step 389, Loss: {'train/loss': 0.349496990442276, 'train/grad_norm_pre_clip': 10.16331672668457, 'timing/data': 0.0050132679753005505, 'timing/model': 1.1269808600191027, 'learning_rate/qwen_vl_interface': 9.918693736959935e-06, 'learning_rate/action_model': 9.918693736959934e-05, 'epoch': 0.8}) +[2026-06-19 10:35:07,291][starVLA.training.train_starvla][INFO] - Step 390, Loss: {'train/loss': 0.49144354462623596, 'train/grad_norm_pre_clip': 12.113970756530762, 'timing/data': 0.0046303150011226535, 'timing/model': 1.1283885219600052, 'learning_rate/qwen_vl_interface': 9.918131715081019e-06, 'learning_rate/action_model': 9.918131715081018e-05, 'epoch': 0.8}) +[2026-06-19 10:35:11,757][starVLA.training.train_starvla][INFO] - Step 391, Loss: {'train/loss': 0.5189460515975952, 'train/grad_norm_pre_clip': 10.955331802368164, 'timing/data': 0.005428257980383933, 'timing/model': 1.1335149549413472, 'learning_rate/qwen_vl_interface': 9.917567774309486e-06, 'learning_rate/action_model': 9.917567774309485e-05, 'epoch': 0.8}) +[2026-06-19 10:35:16,212][starVLA.training.train_starvla][INFO] - Step 392, Loss: {'train/loss': 0.24979086220264435, 'train/grad_norm_pre_clip': 8.423419952392578, 'timing/data': 0.005199340987019241, 'timing/model': 1.2528847130015492, 'learning_rate/qwen_vl_interface': 9.91700191487715e-06, 'learning_rate/action_model': 9.91700191487715e-05, 'epoch': 0.81}) +[2026-06-19 10:35:20,552][starVLA.training.train_starvla][INFO] - Step 393, Loss: {'train/loss': 0.3090333342552185, 'train/grad_norm_pre_clip': 17.404869079589844, 'timing/data': 0.005083602969534695, 'timing/model': 1.134916861075908, 'learning_rate/qwen_vl_interface': 9.916434137016614e-06, 'learning_rate/action_model': 9.916434137016613e-05, 'epoch': 0.81}) +[2026-06-19 10:35:25,047][starVLA.training.train_starvla][INFO] - Step 394, Loss: {'train/loss': 0.16629062592983246, 'train/grad_norm_pre_clip': 4.812489986419678, 'timing/data': 0.005040280055254698, 'timing/model': 1.1283479200210422, 'learning_rate/qwen_vl_interface': 9.915864440961272e-06, 'learning_rate/action_model': 9.915864440961272e-05, 'epoch': 0.81}) +[2026-06-19 10:35:29,540][starVLA.training.train_starvla][INFO] - Step 395, Loss: {'train/loss': 0.2519082725048065, 'train/grad_norm_pre_clip': 7.0078206062316895, 'timing/data': 0.005124218063428998, 'timing/model': 1.1183387829223648, 'learning_rate/qwen_vl_interface': 9.915292826945303e-06, 'learning_rate/action_model': 9.915292826945303e-05, 'epoch': 0.81}) +[2026-06-19 10:35:34,033][starVLA.training.train_starvla][INFO] - Step 396, Loss: {'train/loss': 0.49407923221588135, 'train/grad_norm_pre_clip': 12.260651588439941, 'timing/data': 0.005163521971553564, 'timing/model': 1.1258678439771757, 'learning_rate/qwen_vl_interface': 9.914719295203676e-06, 'learning_rate/action_model': 9.914719295203676e-05, 'epoch': 0.81}) +[2026-06-19 10:35:38,512][starVLA.training.train_starvla][INFO] - Step 397, Loss: {'train/loss': 0.47260379791259766, 'train/grad_norm_pre_clip': 8.515103340148926, 'timing/data': 0.004869286087341607, 'timing/model': 1.2796937029343098, 'learning_rate/qwen_vl_interface': 9.91414384597215e-06, 'learning_rate/action_model': 9.91414384597215e-05, 'epoch': 0.82}) +[2026-06-19 10:35:42,837][starVLA.training.train_starvla][INFO] - Step 398, Loss: {'train/loss': 0.2567801773548126, 'train/grad_norm_pre_clip': 6.085897445678711, 'timing/data': 0.005230732029303908, 'timing/model': 1.1177130700089037, 'learning_rate/qwen_vl_interface': 9.913566479487267e-06, 'learning_rate/action_model': 9.913566479487267e-05, 'epoch': 0.82}) +[2026-06-19 10:35:47,333][starVLA.training.train_starvla][INFO] - Step 399, Loss: {'train/loss': 0.12656068801879883, 'train/grad_norm_pre_clip': 4.146465301513672, 'timing/data': 0.005279366043396294, 'timing/model': 1.1251165099674836, 'learning_rate/qwen_vl_interface': 9.912987195986366e-06, 'learning_rate/action_model': 9.912987195986365e-05, 'epoch': 0.82}) +[2026-06-19 10:35:51,715][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 400 +[2026-06-19 10:38:47,572][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 400 +[2026-06-19 10:38:47,572][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 400 +[2026-06-19 10:41:52,496][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 400 +[2026-06-19 10:41:52,497][starVLA.training.train_starvla][INFO] - Step 400, Loss: {'train/loss': 0.3089365065097809, 'train/grad_norm_pre_clip': 4.306354999542236, 'eval/loss': 0.3318978250026703, 'eval/latency_0/loss': 0.12808357179164886, 'eval/latency_1/loss': 0.2349090874195099, 'eval/latency_2/loss': 0.38119199872016907, 'eval/latency_3/loss': 0.38602885603904724, 'eval/latency_4/loss': 0.5185750722885132, 'eval/flappy/loss': 0.3318978250026703, 'eval/flappy/latency_0/loss': 0.12808357179164886, 'eval/flappy/latency_1/loss': 0.2349090874195099, 'eval/flappy/latency_2/loss': 0.38119199872016907, 'eval/flappy/latency_3/loss': 0.38602885603904724, 'eval/flappy/latency_4/loss': 0.5185750722885132, 'eval/flap_precision': 0.6887966804979253, 'eval/flap_recall': 0.6036363636363636, 'eval/flap_f1': 0.6434108527131782, 'eval/flap_rate_model': 0.06694444444444445, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.6434108527131782, 'eval/latency_0/flap_precision': 0.6887966804979253, 'eval/latency_0/flap_recall': 0.6036363636363636, 'eval/latency_0/flap_f1': 0.6434108527131782, 'eval/latency_0/flap_rate_model': 0.06694444444444445, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.6434108527131782, 'timing/data': 0.0051086710300296545, 'timing/model': 1.1274684340460226, 'learning_rate/qwen_vl_interface': 9.912405995707562e-06, 'learning_rate/action_model': 9.912405995707561e-05, 'epoch': 0.82}) +[2026-06-19 10:41:52,498][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_400_state +[2026-06-19 10:41:52,501][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-19 10:41:55,789][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_400_state/model.safetensors +[2026-06-19 10:42:03,011][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_400_state/optimizer.bin +[2026-06-19 10:42:03,012][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_400_state/sampler.bin +[2026-06-19 10:42:03,012][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_400_state/sampler_1.bin +[2026-06-19 10:42:03,013][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_400_state/random_states_0.pkl +[2026-06-19 10:42:03,036][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 400 +[2026-06-19 10:42:03,036][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-19 10:42:03,046][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-19 10:42:07,473][starVLA.training.train_starvla][INFO] - Step 401, Loss: {'train/loss': 0.14167246222496033, 'train/grad_norm_pre_clip': 3.7860584259033203, 'timing/data': 0.0051907640881836414, 'timing/model': 1.3080150838941336, 'learning_rate/qwen_vl_interface': 9.911822878889767e-06, 'learning_rate/action_model': 9.911822878889767e-05, 'epoch': 0.82}) +[2026-06-19 10:42:11,781][starVLA.training.train_starvla][INFO] - Step 402, Loss: {'train/loss': 0.30784890055656433, 'train/grad_norm_pre_clip': 6.1193037033081055, 'timing/data': 0.0051632189424708486, 'timing/model': 1.1558495970675722, 'learning_rate/qwen_vl_interface': 9.911237845772681e-06, 'learning_rate/action_model': 9.911237845772681e-05, 'epoch': 0.83}) +[2026-06-19 10:42:16,026][starVLA.training.train_starvla][INFO] - Step 403, Loss: {'train/loss': 0.14744454622268677, 'train/grad_norm_pre_clip': 4.21939754486084, 'timing/data': 0.005048466031439602, 'timing/model': 1.0987839580047876, 'learning_rate/qwen_vl_interface': 9.910650896596787e-06, 'learning_rate/action_model': 9.910650896596786e-05, 'epoch': 0.83}) +[2026-06-19 10:42:20,443][starVLA.training.train_starvla][INFO] - Step 404, Loss: {'train/loss': 0.18127493560314178, 'train/grad_norm_pre_clip': 3.871309757232666, 'timing/data': 0.005144575028680265, 'timing/model': 1.1046971899922937, 'learning_rate/qwen_vl_interface': 9.910062031603356e-06, 'learning_rate/action_model': 9.910062031603356e-05, 'epoch': 0.83}) +[2026-06-19 10:42:24,848][starVLA.training.train_starvla][INFO] - Step 405, Loss: {'train/loss': 0.19824562966823578, 'train/grad_norm_pre_clip': 4.933102130889893, 'timing/data': 0.005194379016757011, 'timing/model': 1.1045779050327837, 'learning_rate/qwen_vl_interface': 9.909471251034451e-06, 'learning_rate/action_model': 9.90947125103445e-05, 'epoch': 0.83}) +[2026-06-19 10:42:29,267][starVLA.training.train_starvla][INFO] - Step 406, Loss: {'train/loss': 0.34241893887519836, 'train/grad_norm_pre_clip': 11.936223030090332, 'timing/data': 0.004876391962170601, 'timing/model': 1.104108036030084, 'learning_rate/qwen_vl_interface': 9.908878555132918e-06, 'learning_rate/action_model': 9.908878555132918e-05, 'epoch': 0.83}) +[2026-06-19 10:42:33,702][starVLA.training.train_starvla][INFO] - Step 407, Loss: {'train/loss': 0.2962181866168976, 'train/grad_norm_pre_clip': 15.08159351348877, 'timing/data': 0.005123678012751043, 'timing/model': 1.278834519092925, 'learning_rate/qwen_vl_interface': 9.908283944142391e-06, 'learning_rate/action_model': 9.908283944142391e-05, 'epoch': 0.84}) +[2026-06-19 10:42:38,124][starVLA.training.train_starvla][INFO] - Step 408, Loss: {'train/loss': 0.09266588091850281, 'train/grad_norm_pre_clip': 3.708899736404419, 'timing/data': 0.0735599499894306, 'timing/model': 1.1932252020342276, 'learning_rate/qwen_vl_interface': 9.907687418307297e-06, 'learning_rate/action_model': 9.907687418307296e-05, 'epoch': 0.84}) +[2026-06-19 10:42:42,407][starVLA.training.train_starvla][INFO] - Step 409, Loss: {'train/loss': 0.28895679116249084, 'train/grad_norm_pre_clip': 5.821720600128174, 'timing/data': 0.005431176978163421, 'timing/model': 1.1136079659918323, 'learning_rate/qwen_vl_interface': 9.907088977872837e-06, 'learning_rate/action_model': 9.907088977872837e-05, 'epoch': 0.84}) +[2026-06-19 10:42:46,840][starVLA.training.train_starvla][INFO] - Step 410, Loss: {'train/loss': 0.30465009808540344, 'train/grad_norm_pre_clip': 10.142112731933594, 'timing/data': 0.005434322054497898, 'timing/model': 1.127073948038742, 'learning_rate/qwen_vl_interface': 9.906488623085015e-06, 'learning_rate/action_model': 9.906488623085015e-05, 'epoch': 0.84}) +[2026-06-19 10:42:51,295][starVLA.training.train_starvla][INFO] - Step 411, Loss: {'train/loss': 0.3288905620574951, 'train/grad_norm_pre_clip': 9.088644981384277, 'timing/data': 0.004942486062645912, 'timing/model': 1.1133976409910247, 'learning_rate/qwen_vl_interface': 9.905886354190611e-06, 'learning_rate/action_model': 9.90588635419061e-05, 'epoch': 0.84}) +[2026-06-19 10:42:55,751][starVLA.training.train_starvla][INFO] - Step 412, Loss: {'train/loss': 0.10809417068958282, 'train/grad_norm_pre_clip': 6.131795406341553, 'timing/data': 0.00540981802623719, 'timing/model': 1.1185234470758587, 'learning_rate/qwen_vl_interface': 9.905282171437192e-06, 'learning_rate/action_model': 9.905282171437192e-05, 'epoch': 0.85}) +[2026-06-19 10:43:00,238][starVLA.training.train_starvla][INFO] - Step 413, Loss: {'train/loss': 0.13991880416870117, 'train/grad_norm_pre_clip': 4.232341766357422, 'timing/data': 0.005197760066948831, 'timing/model': 1.3020204090280458, 'learning_rate/qwen_vl_interface': 9.904676075073121e-06, 'learning_rate/action_model': 9.904676075073121e-05, 'epoch': 0.85}) +[2026-06-19 10:43:04,530][starVLA.training.train_starvla][INFO] - Step 414, Loss: {'train/loss': 0.25937846302986145, 'train/grad_norm_pre_clip': 10.481745719909668, 'timing/data': 0.004921589978039265, 'timing/model': 1.120494827046059, 'learning_rate/qwen_vl_interface': 9.904068065347536e-06, 'learning_rate/action_model': 9.904068065347536e-05, 'epoch': 0.85}) +[2026-06-19 10:43:08,991][starVLA.training.train_starvla][INFO] - Step 415, Loss: {'train/loss': 0.31413739919662476, 'train/grad_norm_pre_clip': 10.796490669250488, 'timing/data': 0.005396754015237093, 'timing/model': 1.1133285419782624, 'learning_rate/qwen_vl_interface': 9.90345814251037e-06, 'learning_rate/action_model': 9.90345814251037e-05, 'epoch': 0.85}) +[2026-06-19 10:43:13,399][starVLA.training.train_starvla][INFO] - Step 416, Loss: {'train/loss': 0.536572277545929, 'train/grad_norm_pre_clip': 13.866850852966309, 'timing/data': 0.005554649047553539, 'timing/model': 1.1126805109670386, 'learning_rate/qwen_vl_interface': 9.902846306812339e-06, 'learning_rate/action_model': 9.902846306812338e-05, 'epoch': 0.85}) +[2026-06-19 10:43:17,839][starVLA.training.train_starvla][INFO] - Step 417, Loss: {'train/loss': 0.10878241807222366, 'train/grad_norm_pre_clip': 4.601665496826172, 'timing/data': 0.005424872972071171, 'timing/model': 1.130968683050014, 'learning_rate/qwen_vl_interface': 9.902232558504942e-06, 'learning_rate/action_model': 9.902232558504941e-05, 'epoch': 0.86}) +[2026-06-19 10:43:22,304][starVLA.training.train_starvla][INFO] - Step 418, Loss: {'train/loss': 0.22242848575115204, 'train/grad_norm_pre_clip': 12.228934288024902, 'timing/data': 0.00511121703311801, 'timing/model': 1.119263810920529, 'learning_rate/qwen_vl_interface': 9.901616897840471e-06, 'learning_rate/action_model': 9.90161689784047e-05, 'epoch': 0.86}) +[2026-06-19 10:43:26,759][starVLA.training.train_starvla][INFO] - Step 419, Loss: {'train/loss': 0.270283579826355, 'train/grad_norm_pre_clip': 11.127277374267578, 'timing/data': 0.00548174197319895, 'timing/model': 1.2682929220609367, 'learning_rate/qwen_vl_interface': 9.900999325072e-06, 'learning_rate/action_model': 9.900999325072001e-05, 'epoch': 0.86}) +[2026-06-19 10:43:31,042][starVLA.training.train_starvla][INFO] - Step 420, Loss: {'train/loss': 0.24190983176231384, 'train/grad_norm_pre_clip': 12.128619194030762, 'timing/data': 0.005450718104839325, 'timing/model': 1.1116807149956003, 'learning_rate/qwen_vl_interface': 9.900379840453392e-06, 'learning_rate/action_model': 9.900379840453392e-05, 'epoch': 0.86}) +[2026-06-19 10:43:35,495][starVLA.training.train_starvla][INFO] - Step 421, Loss: {'train/loss': 0.22385773062705994, 'train/grad_norm_pre_clip': 8.36878490447998, 'timing/data': 0.005405078059993684, 'timing/model': 1.114609320065938, 'learning_rate/qwen_vl_interface': 9.89975844423929e-06, 'learning_rate/action_model': 9.899758444239289e-05, 'epoch': 0.86}) +[2026-06-19 10:43:39,902][starVLA.training.train_starvla][INFO] - Step 422, Loss: {'train/loss': 0.31445932388305664, 'train/grad_norm_pre_clip': 9.302218437194824, 'timing/data': 0.00544007692951709, 'timing/model': 1.113799552898854, 'learning_rate/qwen_vl_interface': 9.899135136685131e-06, 'learning_rate/action_model': 9.89913513668513e-05, 'epoch': 0.87}) +[2026-06-19 10:43:44,352][starVLA.training.train_starvla][INFO] - Step 423, Loss: {'train/loss': 0.22662679851055145, 'train/grad_norm_pre_clip': 6.465806484222412, 'timing/data': 0.004975071060471237, 'timing/model': 1.1162830359535292, 'learning_rate/qwen_vl_interface': 9.898509918047127e-06, 'learning_rate/action_model': 9.898509918047126e-05, 'epoch': 0.87}) +[2026-06-19 10:43:48,814][starVLA.training.train_starvla][INFO] - Step 424, Loss: {'train/loss': 0.36512479186058044, 'train/grad_norm_pre_clip': 9.085944175720215, 'timing/data': 0.005318322917446494, 'timing/model': 1.1173954899422824, 'learning_rate/qwen_vl_interface': 9.897882788582286e-06, 'learning_rate/action_model': 9.897882788582287e-05, 'epoch': 0.87}) +[2026-06-19 10:43:53,266][starVLA.training.train_starvla][INFO] - Step 425, Loss: {'train/loss': 0.5461395382881165, 'train/grad_norm_pre_clip': 14.816669464111328, 'timing/data': 0.005188067094422877, 'timing/model': 1.2685092219617218, 'learning_rate/qwen_vl_interface': 9.897253748548398e-06, 'learning_rate/action_model': 9.897253748548399e-05, 'epoch': 0.87}) +[2026-06-19 10:43:57,658][starVLA.training.train_starvla][INFO] - Step 426, Loss: {'train/loss': 0.25660330057144165, 'train/grad_norm_pre_clip': 7.1970930099487305, 'timing/data': 0.005043326993472874, 'timing/model': 1.2116527559701353, 'learning_rate/qwen_vl_interface': 9.896622798204035e-06, 'learning_rate/action_model': 9.896622798204035e-05, 'epoch': 0.87}) +[2026-06-19 10:44:01,959][starVLA.training.train_starvla][INFO] - Step 427, Loss: {'train/loss': 0.33649638295173645, 'train/grad_norm_pre_clip': 5.756337642669678, 'timing/data': 0.005309955915436149, 'timing/model': 1.115559138939716, 'learning_rate/qwen_vl_interface': 9.89598993780856e-06, 'learning_rate/action_model': 9.89598993780856e-05, 'epoch': 0.88}) +[2026-06-19 10:44:06,398][starVLA.training.train_starvla][INFO] - Step 428, Loss: {'train/loss': 0.33211958408355713, 'train/grad_norm_pre_clip': 8.61389446258545, 'timing/data': 0.0052602170035243034, 'timing/model': 1.1243069740012288, 'learning_rate/qwen_vl_interface': 9.895355167622113e-06, 'learning_rate/action_model': 9.895355167622113e-05, 'epoch': 0.88}) +[2026-06-19 10:44:10,870][starVLA.training.train_starvla][INFO] - Step 429, Loss: {'train/loss': 0.2906848192214966, 'train/grad_norm_pre_clip': 5.841509819030762, 'timing/data': 0.005764012923464179, 'timing/model': 1.1171126260887831, 'learning_rate/qwen_vl_interface': 9.894718487905629e-06, 'learning_rate/action_model': 9.894718487905629e-05, 'epoch': 0.88}) +[2026-06-19 10:44:15,356][starVLA.training.train_starvla][INFO] - Step 430, Loss: {'train/loss': 0.25478121638298035, 'train/grad_norm_pre_clip': 5.864048957824707, 'timing/data': 0.005418988992460072, 'timing/model': 1.1190697270212695, 'learning_rate/qwen_vl_interface': 9.89407989892082e-06, 'learning_rate/action_model': 9.894079898920821e-05, 'epoch': 0.88}) +[2026-06-19 10:44:19,860][starVLA.training.train_starvla][INFO] - Step 431, Loss: {'train/loss': 0.13435806334018707, 'train/grad_norm_pre_clip': 3.791619300842285, 'timing/data': 0.005126088974066079, 'timing/model': 1.291604456026107, 'learning_rate/qwen_vl_interface': 9.893439400930187e-06, 'learning_rate/action_model': 9.893439400930187e-05, 'epoch': 0.89}) +[2026-06-19 10:44:24,306][starVLA.training.train_starvla][INFO] - Step 432, Loss: {'train/loss': 0.21389801800251007, 'train/grad_norm_pre_clip': 4.972012042999268, 'timing/data': 0.005437193904072046, 'timing/model': 1.2478659460321069, 'learning_rate/qwen_vl_interface': 9.892796994197014e-06, 'learning_rate/action_model': 9.892796994197014e-05, 'epoch': 0.89}) +[2026-06-19 10:44:28,618][starVLA.training.train_starvla][INFO] - Step 433, Loss: {'train/loss': 0.4643104076385498, 'train/grad_norm_pre_clip': 9.584436416625977, 'timing/data': 0.005468368995934725, 'timing/model': 1.1242400649935007, 'learning_rate/qwen_vl_interface': 9.892152678985372e-06, 'learning_rate/action_model': 9.892152678985372e-05, 'epoch': 0.89}) +[2026-06-19 10:44:33,075][starVLA.training.train_starvla][INFO] - Step 434, Loss: {'train/loss': 0.29140034317970276, 'train/grad_norm_pre_clip': 20.48760223388672, 'timing/data': 0.0055450230138376355, 'timing/model': 1.119411893072538, 'learning_rate/qwen_vl_interface': 9.891506455560111e-06, 'learning_rate/action_model': 9.891506455560111e-05, 'epoch': 0.89}) +[2026-06-19 10:44:37,562][starVLA.training.train_starvla][INFO] - Step 435, Loss: {'train/loss': 0.5545163750648499, 'train/grad_norm_pre_clip': 13.450231552124023, 'timing/data': 0.0052142220083624125, 'timing/model': 1.1239544689888135, 'learning_rate/qwen_vl_interface': 9.890858324186872e-06, 'learning_rate/action_model': 9.890858324186872e-05, 'epoch': 0.89}) +[2026-06-19 10:44:42,029][starVLA.training.train_starvla][INFO] - Step 436, Loss: {'train/loss': 0.3059079647064209, 'train/grad_norm_pre_clip': 6.980233669281006, 'timing/data': 0.005787640926428139, 'timing/model': 1.1226600500522181, 'learning_rate/qwen_vl_interface': 9.890208285132077e-06, 'learning_rate/action_model': 9.890208285132077e-05, 'epoch': 0.9}) +[2026-06-19 10:44:46,511][starVLA.training.train_starvla][INFO] - Step 437, Loss: {'train/loss': 0.3448081314563751, 'train/grad_norm_pre_clip': 7.85563850402832, 'timing/data': 0.005260169971734285, 'timing/model': 1.2792121799429879, 'learning_rate/qwen_vl_interface': 9.889556338662933e-06, 'learning_rate/action_model': 9.889556338662932e-05, 'epoch': 0.9}) +[2026-06-19 10:44:50,833][starVLA.training.train_starvla][INFO] - Step 438, Loss: {'train/loss': 0.3830872178077698, 'train/grad_norm_pre_clip': 8.23129653930664, 'timing/data': 0.005627001984976232, 'timing/model': 1.122227574000135, 'learning_rate/qwen_vl_interface': 9.888902485047427e-06, 'learning_rate/action_model': 9.888902485047427e-05, 'epoch': 0.9}) +[2026-06-19 10:44:55,322][starVLA.training.train_starvla][INFO] - Step 439, Loss: {'train/loss': 0.21116407215595245, 'train/grad_norm_pre_clip': 3.258018970489502, 'timing/data': 0.005658772075548768, 'timing/model': 1.121413475018926, 'learning_rate/qwen_vl_interface': 9.888246724554337e-06, 'learning_rate/action_model': 9.888246724554337e-05, 'epoch': 0.9}) +[2026-06-19 10:44:59,823][starVLA.training.train_starvla][INFO] - Step 440, Loss: {'train/loss': 0.2250199019908905, 'train/grad_norm_pre_clip': 3.692063093185425, 'timing/data': 0.0054160329746082425, 'timing/model': 1.1200566290644929, 'learning_rate/qwen_vl_interface': 9.887589057453222e-06, 'learning_rate/action_model': 9.88758905745322e-05, 'epoch': 0.9}) +[2026-06-19 10:45:04,303][starVLA.training.train_starvla][INFO] - Step 441, Loss: {'train/loss': 0.26237812638282776, 'train/grad_norm_pre_clip': 6.631376266479492, 'timing/data': 0.005163189023733139, 'timing/model': 1.1226008819648996, 'learning_rate/qwen_vl_interface': 9.88692948401442e-06, 'learning_rate/action_model': 9.88692948401442e-05, 'epoch': 0.91}) +[2026-06-19 10:45:08,702][starVLA.training.train_starvla][INFO] - Step 442, Loss: {'train/loss': 0.4053197503089905, 'train/grad_norm_pre_clip': 8.716270446777344, 'timing/data': 0.005625176010653377, 'timing/model': 1.117054195026867, 'learning_rate/qwen_vl_interface': 9.88626800450906e-06, 'learning_rate/action_model': 9.88626800450906e-05, 'epoch': 0.91}) +[2026-06-19 10:45:13,143][starVLA.training.train_starvla][INFO] - Step 443, Loss: {'train/loss': 0.3714711666107178, 'train/grad_norm_pre_clip': 9.09407901763916, 'timing/data': 0.00557365792337805, 'timing/model': 1.243076324928552, 'learning_rate/qwen_vl_interface': 9.885604619209052e-06, 'learning_rate/action_model': 9.885604619209051e-05, 'epoch': 0.91}) +[2026-06-19 10:45:17,471][starVLA.training.train_starvla][INFO] - Step 444, Loss: {'train/loss': 0.34163767099380493, 'train/grad_norm_pre_clip': 5.94984245300293, 'timing/data': 0.005382120958529413, 'timing/model': 1.11691737791989, 'learning_rate/qwen_vl_interface': 9.884939328387085e-06, 'learning_rate/action_model': 9.884939328387084e-05, 'epoch': 0.91}) +[2026-06-19 10:45:21,954][starVLA.training.train_starvla][INFO] - Step 445, Loss: {'train/loss': 0.2745097279548645, 'train/grad_norm_pre_clip': 6.270686149597168, 'timing/data': 0.005191272939555347, 'timing/model': 1.117486440925859, 'learning_rate/qwen_vl_interface': 9.884272132316638e-06, 'learning_rate/action_model': 9.884272132316637e-05, 'epoch': 0.91}) +[2026-06-19 10:45:26,465][starVLA.training.train_starvla][INFO] - Step 446, Loss: {'train/loss': 0.2708616852760315, 'train/grad_norm_pre_clip': 10.785408973693848, 'timing/data': 0.005388926016166806, 'timing/model': 1.1183849310036749, 'learning_rate/qwen_vl_interface': 9.883603031271968e-06, 'learning_rate/action_model': 9.883603031271969e-05, 'epoch': 0.92}) +[2026-06-19 10:45:30,935][starVLA.training.train_starvla][INFO] - Step 447, Loss: {'train/loss': 0.3931639492511749, 'train/grad_norm_pre_clip': 9.294507026672363, 'timing/data': 0.005684442003257573, 'timing/model': 1.1156778490403667, 'learning_rate/qwen_vl_interface': 9.882932025528122e-06, 'learning_rate/action_model': 9.882932025528121e-05, 'epoch': 0.92}) +[2026-06-19 10:45:35,362][starVLA.training.train_starvla][INFO] - Step 448, Loss: {'train/loss': 0.25138258934020996, 'train/grad_norm_pre_clip': 5.7956085205078125, 'timing/data': 0.004987312946468592, 'timing/model': 1.1144011289579794, 'learning_rate/qwen_vl_interface': 9.882259115360918e-06, 'learning_rate/action_model': 9.882259115360917e-05, 'epoch': 0.92}) +[2026-06-19 10:45:39,793][starVLA.training.train_starvla][INFO] - Step 449, Loss: {'train/loss': 0.14545410871505737, 'train/grad_norm_pre_clip': 3.541459798812866, 'timing/data': 0.0051446809666231275, 'timing/model': 1.2475467849289998, 'learning_rate/qwen_vl_interface': 9.881584301046968e-06, 'learning_rate/action_model': 9.881584301046968e-05, 'epoch': 0.92}) +[2026-06-19 10:45:44,104][starVLA.training.train_starvla][INFO] - Step 450, Loss: {'train/loss': 0.6860316395759583, 'train/grad_norm_pre_clip': 12.919462203979492, 'timing/data': 0.005701406975276768, 'timing/model': 1.1224988530157134, 'learning_rate/qwen_vl_interface': 9.880907582863662e-06, 'learning_rate/action_model': 9.880907582863662e-05, 'epoch': 0.92}) +[2026-06-19 10:45:48,608][starVLA.training.train_starvla][INFO] - Step 451, Loss: {'train/loss': 0.13998864591121674, 'train/grad_norm_pre_clip': 3.6721808910369873, 'timing/data': 0.005385309923440218, 'timing/model': 1.1226923980284482, 'learning_rate/qwen_vl_interface': 9.880228961089175e-06, 'learning_rate/action_model': 9.880228961089175e-05, 'epoch': 0.93}) +[2026-06-19 10:45:53,116][starVLA.training.train_starvla][INFO] - Step 452, Loss: {'train/loss': 0.3683582544326782, 'train/grad_norm_pre_clip': 7.571364879608154, 'timing/data': 0.005180648993700743, 'timing/model': 1.1425602509407327, 'learning_rate/qwen_vl_interface': 9.87954843600246e-06, 'learning_rate/action_model': 9.879548436002459e-05, 'epoch': 0.93}) +[2026-06-19 10:45:57,588][starVLA.training.train_starvla][INFO] - Step 453, Loss: {'train/loss': 0.3303641080856323, 'train/grad_norm_pre_clip': 4.4351630210876465, 'timing/data': 0.0050308729987591505, 'timing/model': 1.123717449954711, 'learning_rate/qwen_vl_interface': 9.878866007883256e-06, 'learning_rate/action_model': 9.878866007883257e-05, 'epoch': 0.93}) +[2026-06-19 10:46:02,083][starVLA.training.train_starvla][INFO] - Step 454, Loss: {'train/loss': 0.22175884246826172, 'train/grad_norm_pre_clip': 5.274360656738281, 'timing/data': 0.005692720995284617, 'timing/model': 1.3053452789317816, 'learning_rate/qwen_vl_interface': 9.878181677012084e-06, 'learning_rate/action_model': 9.878181677012085e-05, 'epoch': 0.93}) +[2026-06-19 10:46:06,410][starVLA.training.train_starvla][INFO] - Step 455, Loss: {'train/loss': 0.32092559337615967, 'train/grad_norm_pre_clip': 3.5427768230438232, 'timing/data': 0.0054572829976677895, 'timing/model': 1.1235527859535068, 'learning_rate/qwen_vl_interface': 9.877495443670247e-06, 'learning_rate/action_model': 9.877495443670246e-05, 'epoch': 0.93}) +[2026-06-19 10:46:10,912][starVLA.training.train_starvla][INFO] - Step 456, Loss: {'train/loss': 0.24598906934261322, 'train/grad_norm_pre_clip': 6.311707496643066, 'timing/data': 0.005121803958900273, 'timing/model': 1.1290536429733038, 'learning_rate/qwen_vl_interface': 9.876807308139828e-06, 'learning_rate/action_model': 9.876807308139829e-05, 'epoch': 0.94}) +[2026-06-19 10:46:15,323][starVLA.training.train_starvla][INFO] - Step 457, Loss: {'train/loss': 0.2836107611656189, 'train/grad_norm_pre_clip': 5.419698715209961, 'timing/data': 0.005749844014644623, 'timing/model': 1.1184461179655045, 'learning_rate/qwen_vl_interface': 9.876117270703695e-06, 'learning_rate/action_model': 9.876117270703694e-05, 'epoch': 0.94}) +[2026-06-19 10:46:19,775][starVLA.training.train_starvla][INFO] - Step 458, Loss: {'train/loss': 0.3001525402069092, 'train/grad_norm_pre_clip': 8.175626754760742, 'timing/data': 0.005240208003669977, 'timing/model': 1.1205165450228378, 'learning_rate/qwen_vl_interface': 9.875425331645495e-06, 'learning_rate/action_model': 9.875425331645495e-05, 'epoch': 0.94}) +[2026-06-19 10:46:24,210][starVLA.training.train_starvla][INFO] - Step 459, Loss: {'train/loss': 0.4009648263454437, 'train/grad_norm_pre_clip': 6.179261684417725, 'timing/data': 0.005864926963113248, 'timing/model': 1.1237246609525755, 'learning_rate/qwen_vl_interface': 9.874731491249658e-06, 'learning_rate/action_model': 9.874731491249657e-05, 'epoch': 0.94}) +[2026-06-19 10:46:28,690][starVLA.training.train_starvla][INFO] - Step 460, Loss: {'train/loss': 0.17783492803573608, 'train/grad_norm_pre_clip': 3.667375326156616, 'timing/data': 0.005344828008674085, 'timing/model': 1.284252888057381, 'learning_rate/qwen_vl_interface': 9.874035749801398e-06, 'learning_rate/action_model': 9.874035749801399e-05, 'epoch': 0.94}) +[2026-06-19 10:46:33,007][starVLA.training.train_starvla][INFO] - Step 461, Loss: {'train/loss': 0.2717687487602234, 'train/grad_norm_pre_clip': 5.216043949127197, 'timing/data': 0.00547090801410377, 'timing/model': 1.1193202460417524, 'learning_rate/qwen_vl_interface': 9.873338107586704e-06, 'learning_rate/action_model': 9.873338107586705e-05, 'epoch': 0.95}) +[2026-06-19 10:46:37,496][starVLA.training.train_starvla][INFO] - Step 462, Loss: {'train/loss': 0.30539771914482117, 'train/grad_norm_pre_clip': 5.23098087310791, 'timing/data': 0.005092523992061615, 'timing/model': 1.1237473209621385, 'learning_rate/qwen_vl_interface': 9.872638564892358e-06, 'learning_rate/action_model': 9.872638564892357e-05, 'epoch': 0.95}) +[2026-06-19 10:46:41,995][starVLA.training.train_starvla][INFO] - Step 463, Loss: {'train/loss': 0.28805437684059143, 'train/grad_norm_pre_clip': 5.625900745391846, 'timing/data': 0.004879639018326998, 'timing/model': 1.1287556529277936, 'learning_rate/qwen_vl_interface': 9.871937122005905e-06, 'learning_rate/action_model': 9.871937122005905e-05, 'epoch': 0.95}) +[2026-06-19 10:46:46,433][starVLA.training.train_starvla][INFO] - Step 464, Loss: {'train/loss': 0.1425619125366211, 'train/grad_norm_pre_clip': 2.563488006591797, 'timing/data': 0.005245096981525421, 'timing/model': 1.1196933700703084, 'learning_rate/qwen_vl_interface': 9.87123377921569e-06, 'learning_rate/action_model': 9.871233779215689e-05, 'epoch': 0.95}) +[2026-06-19 10:46:50,891][starVLA.training.train_starvla][INFO] - Step 465, Loss: {'train/loss': 0.15828396379947662, 'train/grad_norm_pre_clip': 2.7816085815429688, 'timing/data': 0.0054658809676766396, 'timing/model': 1.135498880990781, 'learning_rate/qwen_vl_interface': 9.870528536810825e-06, 'learning_rate/action_model': 9.870528536810825e-05, 'epoch': 0.96}) +[2026-06-19 10:46:55,382][starVLA.training.train_starvla][INFO] - Step 466, Loss: {'train/loss': 0.2629682719707489, 'train/grad_norm_pre_clip': 9.992969512939453, 'timing/data': 0.005437893094494939, 'timing/model': 1.291574064991437, 'learning_rate/qwen_vl_interface': 9.869821395081214e-06, 'learning_rate/action_model': 9.869821395081214e-05, 'epoch': 0.96}) +[2026-06-19 10:46:59,710][starVLA.training.train_starvla][INFO] - Step 467, Loss: {'train/loss': 0.28056514263153076, 'train/grad_norm_pre_clip': 7.434857368469238, 'timing/data': 0.005096781998872757, 'timing/model': 1.128481555962935, 'learning_rate/qwen_vl_interface': 9.869112354317532e-06, 'learning_rate/action_model': 9.869112354317532e-05, 'epoch': 0.96}) +[2026-06-19 10:47:04,195][starVLA.training.train_starvla][INFO] - Step 468, Loss: {'train/loss': 0.4135628938674927, 'train/grad_norm_pre_clip': 11.936931610107422, 'timing/data': 0.005086262943223119, 'timing/model': 1.1296054749982432, 'learning_rate/qwen_vl_interface': 9.86840141481124e-06, 'learning_rate/action_model': 9.868401414811239e-05, 'epoch': 0.96}) +[2026-06-19 10:47:08,702][starVLA.training.train_starvla][INFO] - Step 469, Loss: {'train/loss': 0.20111608505249023, 'train/grad_norm_pre_clip': 5.678515911102295, 'timing/data': 0.005376370041631162, 'timing/model': 1.1197150490479544, 'learning_rate/qwen_vl_interface': 9.867688576854578e-06, 'learning_rate/action_model': 9.867688576854578e-05, 'epoch': 0.96}) +[2026-06-19 10:47:13,178][starVLA.training.train_starvla][INFO] - Step 470, Loss: {'train/loss': 0.18106912076473236, 'train/grad_norm_pre_clip': 8.905372619628906, 'timing/data': 0.005515604047104716, 'timing/model': 1.1205368250375614, 'learning_rate/qwen_vl_interface': 9.86697384074057e-06, 'learning_rate/action_model': 9.866973840740568e-05, 'epoch': 0.97}) +[2026-06-19 10:47:17,621][starVLA.training.train_starvla][INFO] - Step 471, Loss: {'train/loss': 0.18252509832382202, 'train/grad_norm_pre_clip': 6.615080833435059, 'timing/data': 0.00556810200214386, 'timing/model': 1.1326506111072376, 'learning_rate/qwen_vl_interface': 9.86625720676301e-06, 'learning_rate/action_model': 9.86625720676301e-05, 'epoch': 0.97}) +[2026-06-19 10:47:22,137][starVLA.training.train_starvla][INFO] - Step 472, Loss: {'train/loss': 0.2727489173412323, 'train/grad_norm_pre_clip': 4.699528694152832, 'timing/data': 0.005242222920060158, 'timing/model': 1.32035563304089, 'learning_rate/qwen_vl_interface': 9.865538675216488e-06, 'learning_rate/action_model': 9.865538675216486e-05, 'epoch': 0.97}) +[2026-06-19 10:47:26,457][starVLA.training.train_starvla][INFO] - Step 473, Loss: {'train/loss': 0.21100078523159027, 'train/grad_norm_pre_clip': 6.051325798034668, 'timing/data': 0.004863259964622557, 'timing/model': 1.1207225460093468, 'learning_rate/qwen_vl_interface': 9.864818246396358e-06, 'learning_rate/action_model': 9.864818246396358e-05, 'epoch': 0.97}) +[2026-06-19 10:47:30,935][starVLA.training.train_starvla][INFO] - Step 474, Loss: {'train/loss': 0.21209776401519775, 'train/grad_norm_pre_clip': 12.130494117736816, 'timing/data': 0.005118667031638324, 'timing/model': 1.1210719889495522, 'learning_rate/qwen_vl_interface': 9.864095920598764e-06, 'learning_rate/action_model': 9.864095920598763e-05, 'epoch': 0.97}) +[2026-06-19 10:47:35,429][starVLA.training.train_starvla][INFO] - Step 475, Loss: {'train/loss': 0.30752667784690857, 'train/grad_norm_pre_clip': 9.714051246643066, 'timing/data': 0.0049314110074192286, 'timing/model': 1.1194558270508423, 'learning_rate/qwen_vl_interface': 9.86337169812063e-06, 'learning_rate/action_model': 9.863371698120629e-05, 'epoch': 0.98}) +[2026-06-19 10:47:39,890][starVLA.training.train_starvla][INFO] - Step 476, Loss: {'train/loss': 0.06461415439844131, 'train/grad_norm_pre_clip': 6.119316101074219, 'timing/data': 0.005303433979861438, 'timing/model': 1.1282729809172451, 'learning_rate/qwen_vl_interface': 9.862645579259648e-06, 'learning_rate/action_model': 9.862645579259648e-05, 'epoch': 0.98}) +[2026-06-19 10:47:44,368][starVLA.training.train_starvla][INFO] - Step 477, Loss: {'train/loss': 0.13446453213691711, 'train/grad_norm_pre_clip': 13.011281967163086, 'timing/data': 0.005264383042231202, 'timing/model': 1.1152596169849858, 'learning_rate/qwen_vl_interface': 9.861917564314308e-06, 'learning_rate/action_model': 9.861917564314308e-05, 'epoch': 0.98}) +[2026-06-19 10:47:48,866][starVLA.training.train_starvla][INFO] - Step 478, Loss: {'train/loss': 0.28299492597579956, 'train/grad_norm_pre_clip': 9.457742691040039, 'timing/data': 0.005171166965737939, 'timing/model': 1.2870458019897342, 'learning_rate/qwen_vl_interface': 9.861187653583862e-06, 'learning_rate/action_model': 9.861187653583862e-05, 'epoch': 0.98}) +[2026-06-19 10:47:53,183][starVLA.training.train_starvla][INFO] - Step 479, Loss: {'train/loss': 0.5497856140136719, 'train/grad_norm_pre_clip': 20.635324478149414, 'timing/data': 0.005048312013968825, 'timing/model': 1.1198319650720805, 'learning_rate/qwen_vl_interface': 9.860455847368352e-06, 'learning_rate/action_model': 9.860455847368351e-05, 'epoch': 0.98}) +[2026-06-19 10:47:57,651][starVLA.training.train_starvla][INFO] - Step 480, Loss: {'train/loss': 0.18611733615398407, 'train/grad_norm_pre_clip': 10.34455394744873, 'timing/data': 0.005045137950219214, 'timing/model': 1.1167037909617648, 'learning_rate/qwen_vl_interface': 9.859722145968595e-06, 'learning_rate/action_model': 9.859722145968595e-05, 'epoch': 0.99}) +[2026-06-19 10:48:02,062][starVLA.training.train_starvla][INFO] - Step 481, Loss: {'train/loss': 0.28196871280670166, 'train/grad_norm_pre_clip': 6.586556911468506, 'timing/data': 0.005461719934828579, 'timing/model': 1.1203202550532296, 'learning_rate/qwen_vl_interface': 9.858986549686188e-06, 'learning_rate/action_model': 9.858986549686188e-05, 'epoch': 0.99}) +[2026-06-19 10:48:06,520][starVLA.training.train_starvla][INFO] - Step 482, Loss: {'train/loss': 0.33083027601242065, 'train/grad_norm_pre_clip': 9.968198776245117, 'timing/data': 0.005259363097138703, 'timing/model': 1.1237323809182271, 'learning_rate/qwen_vl_interface': 9.85824905882351e-06, 'learning_rate/action_model': 9.858249058823509e-05, 'epoch': 0.99}) +[2026-06-19 10:48:11,001][starVLA.training.train_starvla][INFO] - Step 483, Loss: {'train/loss': 0.1991526037454605, 'train/grad_norm_pre_clip': 8.635108947753906, 'timing/data': 0.00515315099619329, 'timing/model': 1.1215140389977023, 'learning_rate/qwen_vl_interface': 9.857509673683709e-06, 'learning_rate/action_model': 9.857509673683708e-05, 'epoch': 0.99}) +[2026-06-19 10:48:15,492][starVLA.training.train_starvla][INFO] - Step 484, Loss: {'train/loss': 0.10808810591697693, 'train/grad_norm_pre_clip': 3.390244245529175, 'timing/data': 0.004886117996647954, 'timing/model': 1.2824881880078465, 'learning_rate/qwen_vl_interface': 9.856768394570725e-06, 'learning_rate/action_model': 9.856768394570723e-05, 'epoch': 0.99}) +[2026-06-19 10:48:19,796][starVLA.training.train_starvla][INFO] - Step 485, Loss: {'train/loss': 0.24619156122207642, 'train/grad_norm_pre_clip': 7.798933982849121, 'timing/data': 0.005293213063850999, 'timing/model': 1.1055167010053992, 'learning_rate/qwen_vl_interface': 9.856025221789264e-06, 'learning_rate/action_model': 9.856025221789264e-05, 'epoch': 1.0}) +[2026-06-19 10:48:24,167][starVLA.training.train_starvla][INFO] - Step 486, Loss: {'train/loss': 0.17338809370994568, 'train/grad_norm_pre_clip': 4.497951030731201, 'timing/data': 0.004340671934187412, 'timing/model': 1.1045532590942457, 'learning_rate/qwen_vl_interface': 9.85528015564482e-06, 'learning_rate/action_model': 9.85528015564482e-05, 'epoch': 1.0}) +[2026-06-19 10:48:28,865][starVLA.training.train_starvla][INFO] - Step 487, Loss: {'train/loss': 0.07120506465435028, 'train/grad_norm_pre_clip': 2.5059731006622314, 'timing/data': 0.7884513760218397, 'timing/model': 0.6487121880054474, 'learning_rate/qwen_vl_interface': 9.854533196443661e-06, 'learning_rate/action_model': 9.854533196443661e-05, 'epoch': 1.0}) +[2026-06-19 10:48:45,376][starVLA.training.train_starvla][INFO] - Step 488, Loss: {'train/loss': 0.228329598903656, 'train/grad_norm_pre_clip': 6.543917179107666, 'timing/data': 0.013036038028076291, 'timing/model': 1.107521902071312, 'learning_rate/qwen_vl_interface': 9.853784344492833e-06, 'learning_rate/action_model': 9.853784344492833e-05, 'epoch': 1.0}) +[2026-06-19 10:48:49,818][starVLA.training.train_starvla][INFO] - Step 489, Loss: {'train/loss': 0.15635734796524048, 'train/grad_norm_pre_clip': 4.608913421630859, 'timing/data': 0.012109956005588174, 'timing/model': 1.1139175849966705, 'learning_rate/qwen_vl_interface': 9.853033600100162e-06, 'learning_rate/action_model': 9.853033600100162e-05, 'epoch': 1.0}) +[2026-06-19 10:48:54,298][starVLA.training.train_starvla][INFO] - Step 490, Loss: {'train/loss': 0.36037778854370117, 'train/grad_norm_pre_clip': 8.745057106018066, 'timing/data': 0.012580880080349743, 'timing/model': 1.3021758260438219, 'learning_rate/qwen_vl_interface': 9.85228096357425e-06, 'learning_rate/action_model': 9.852280963574249e-05, 'epoch': 1.01}) +[2026-06-19 10:48:58,590][starVLA.training.train_starvla][INFO] - Step 491, Loss: {'train/loss': 0.574945330619812, 'train/grad_norm_pre_clip': 13.772653579711914, 'timing/data': 0.014002711977809668, 'timing/model': 1.1140877100406215, 'learning_rate/qwen_vl_interface': 9.851526435224475e-06, 'learning_rate/action_model': 9.851526435224474e-05, 'epoch': 1.01}) +[2026-06-19 10:49:03,047][starVLA.training.train_starvla][INFO] - Step 492, Loss: {'train/loss': 0.3336295187473297, 'train/grad_norm_pre_clip': 11.314281463623047, 'timing/data': 0.01534158899448812, 'timing/model': 1.1127982439938933, 'learning_rate/qwen_vl_interface': 9.850770015360999e-06, 'learning_rate/action_model': 9.850770015360999e-05, 'epoch': 1.01}) +[2026-06-19 10:49:07,521][starVLA.training.train_starvla][INFO] - Step 493, Loss: {'train/loss': 0.2611556053161621, 'train/grad_norm_pre_clip': 8.921157836914062, 'timing/data': 0.013112351996824145, 'timing/model': 1.1249127760529518, 'learning_rate/qwen_vl_interface': 9.850011704294755e-06, 'learning_rate/action_model': 9.850011704294755e-05, 'epoch': 1.01}) +[2026-06-19 10:49:11,956][starVLA.training.train_starvla][INFO] - Step 494, Loss: {'train/loss': 0.19771802425384521, 'train/grad_norm_pre_clip': 5.962283134460449, 'timing/data': 0.013440910028293729, 'timing/model': 1.1245519920485094, 'learning_rate/qwen_vl_interface': 9.849251502337458e-06, 'learning_rate/action_model': 9.849251502337458e-05, 'epoch': 1.01}) +[2026-06-19 10:49:16,433][starVLA.training.train_starvla][INFO] - Step 495, Loss: {'train/loss': 0.20524972677230835, 'train/grad_norm_pre_clip': 4.139707565307617, 'timing/data': 0.01374598906841129, 'timing/model': 1.1151025330182165, 'learning_rate/qwen_vl_interface': 9.848489409801598e-06, 'learning_rate/action_model': 9.848489409801598e-05, 'epoch': 1.02}) +[2026-06-19 10:49:20,899][starVLA.training.train_starvla][INFO] - Step 496, Loss: {'train/loss': 0.22490456700325012, 'train/grad_norm_pre_clip': 6.339968204498291, 'timing/data': 0.012666583992540836, 'timing/model': 1.2774163319263607, 'learning_rate/qwen_vl_interface': 9.84772542700044e-06, 'learning_rate/action_model': 9.84772542700044e-05, 'epoch': 1.02}) +[2026-06-19 10:49:25,210][starVLA.training.train_starvla][INFO] - Step 497, Loss: {'train/loss': 0.40187641978263855, 'train/grad_norm_pre_clip': 6.976327419281006, 'timing/data': 0.013555974001064897, 'timing/model': 1.1168950049905106, 'learning_rate/qwen_vl_interface': 9.846959554248032e-06, 'learning_rate/action_model': 9.84695955424803e-05, 'epoch': 1.02}) +[2026-06-19 10:49:29,690][starVLA.training.train_starvla][INFO] - Step 498, Loss: {'train/loss': 0.23578740656375885, 'train/grad_norm_pre_clip': 5.266233921051025, 'timing/data': 0.012717723962850869, 'timing/model': 1.1266261499840766, 'learning_rate/qwen_vl_interface': 9.846191791859192e-06, 'learning_rate/action_model': 9.846191791859192e-05, 'epoch': 1.02}) +[2026-06-19 10:49:34,144][starVLA.training.train_starvla][INFO] - Step 499, Loss: {'train/loss': 0.2873690128326416, 'train/grad_norm_pre_clip': 7.238549709320068, 'timing/data': 0.012194131035357714, 'timing/model': 1.1283899280242622, 'learning_rate/qwen_vl_interface': 9.845422140149518e-06, 'learning_rate/action_model': 9.845422140149519e-05, 'epoch': 1.02}) +[2026-06-19 10:49:38,615][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 500 +[2026-06-19 10:52:32,809][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 500 +[2026-06-19 10:52:32,810][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 500 +[2026-06-19 10:55:38,528][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 500 +[2026-06-19 10:55:38,529][starVLA.training.train_starvla][INFO] - Step 500, Loss: {'train/loss': 0.270133912563324, 'train/grad_norm_pre_clip': 6.465237140655518, 'eval/loss': 0.33713099360466003, 'eval/latency_0/loss': 0.12770426273345947, 'eval/latency_1/loss': 0.25080958008766174, 'eval/latency_2/loss': 0.34753331542015076, 'eval/latency_3/loss': 0.3965977430343628, 'eval/latency_4/loss': 0.5544220805168152, 'eval/flappy/loss': 0.33713099360466003, 'eval/flappy/latency_0/loss': 0.12770426273345947, 'eval/flappy/latency_1/loss': 0.25080958008766174, 'eval/flappy/latency_2/loss': 0.34753331542015076, 'eval/flappy/latency_3/loss': 0.3965977430343628, 'eval/flappy/latency_4/loss': 0.5544220805168152, 'eval/flap_precision': 0.8243243243243243, 'eval/flap_recall': 0.44363636363636366, 'eval/flap_f1': 0.5768321513002365, 'eval/flap_rate_model': 0.04111111111111111, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.5768321513002365, 'eval/latency_0/flap_precision': 0.8243243243243243, 'eval/latency_0/flap_recall': 0.44363636363636366, 'eval/latency_0/flap_f1': 0.5768321513002365, 'eval/latency_0/flap_rate_model': 0.04111111111111111, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.5768321513002365, 'timing/data': 0.012079617008566856, 'timing/model': 1.1196386660449207, 'learning_rate/qwen_vl_interface': 9.84465059943539e-06, 'learning_rate/action_model': 9.84465059943539e-05, 'epoch': 1.03}) +[2026-06-19 10:55:38,529][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_500_state +[2026-06-19 10:55:38,532][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-19 10:55:41,900][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_500_state/model.safetensors +[2026-06-19 10:55:49,060][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_500_state/optimizer.bin +[2026-06-19 10:55:49,061][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_500_state/sampler.bin +[2026-06-19 10:55:49,061][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_500_state/sampler_1.bin +[2026-06-19 10:55:49,062][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_500_state/random_states_0.pkl +[2026-06-19 10:55:49,086][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 500 +[2026-06-19 10:55:49,086][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-19 10:55:49,096][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-19 10:55:53,453][starVLA.training.train_starvla][INFO] - Step 501, Loss: {'train/loss': 0.27664873003959656, 'train/grad_norm_pre_clip': 7.003633499145508, 'timing/data': 0.01375612209085375, 'timing/model': 1.1098380229668692, 'learning_rate/qwen_vl_interface': 9.843877170033956e-06, 'learning_rate/action_model': 9.843877170033954e-05, 'epoch': 1.03}) +[2026-06-19 10:55:57,853][starVLA.training.train_starvla][INFO] - Step 502, Loss: {'train/loss': 0.19712816178798676, 'train/grad_norm_pre_clip': 4.468479156494141, 'timing/data': 0.012741290964186192, 'timing/model': 1.2544141319813207, 'learning_rate/qwen_vl_interface': 9.84310185226314e-06, 'learning_rate/action_model': 9.84310185226314e-05, 'epoch': 1.03}) +[2026-06-19 10:56:02,112][starVLA.training.train_starvla][INFO] - Step 503, Loss: {'train/loss': 0.1658157855272293, 'train/grad_norm_pre_clip': 4.584237575531006, 'timing/data': 0.013367007952183485, 'timing/model': 1.1083648269996047, 'learning_rate/qwen_vl_interface': 9.842324646441652e-06, 'learning_rate/action_model': 9.842324646441652e-05, 'epoch': 1.03}) +[2026-06-19 10:56:06,558][starVLA.training.train_starvla][INFO] - Step 504, Loss: {'train/loss': 0.14474433660507202, 'train/grad_norm_pre_clip': 3.839411973953247, 'timing/data': 0.01299585506785661, 'timing/model': 1.1117485399590805, 'learning_rate/qwen_vl_interface': 9.841545552888968e-06, 'learning_rate/action_model': 9.841545552888967e-05, 'epoch': 1.04}) +[2026-06-19 10:56:10,994][starVLA.training.train_starvla][INFO] - Step 505, Loss: {'train/loss': 0.23774662613868713, 'train/grad_norm_pre_clip': 10.411951065063477, 'timing/data': 0.013590953894890845, 'timing/model': 1.1080641770968214, 'learning_rate/qwen_vl_interface': 9.840764571925346e-06, 'learning_rate/action_model': 9.840764571925345e-05, 'epoch': 1.04}) +[2026-06-19 10:56:15,448][starVLA.training.train_starvla][INFO] - Step 506, Loss: {'train/loss': 0.1743384599685669, 'train/grad_norm_pre_clip': 6.57637357711792, 'timing/data': 0.013914998969994485, 'timing/model': 1.1106199390487745, 'learning_rate/qwen_vl_interface': 9.839981703871815e-06, 'learning_rate/action_model': 9.839981703871814e-05, 'epoch': 1.04}) +[2026-06-19 10:56:19,928][starVLA.training.train_starvla][INFO] - Step 507, Loss: {'train/loss': 0.2907508313655853, 'train/grad_norm_pre_clip': 7.943201541900635, 'timing/data': 0.013952577952295542, 'timing/model': 1.299242292996496, 'learning_rate/qwen_vl_interface': 9.839196949050185e-06, 'learning_rate/action_model': 9.839196949050185e-05, 'epoch': 1.04}) +[2026-06-19 10:56:24,236][starVLA.training.train_starvla][INFO] - Step 508, Loss: {'train/loss': 0.2503259479999542, 'train/grad_norm_pre_clip': 8.195283889770508, 'timing/data': 0.01374639302957803, 'timing/model': 1.123357219970785, 'learning_rate/qwen_vl_interface': 9.838410307783038e-06, 'learning_rate/action_model': 9.838410307783038e-05, 'epoch': 1.04}) +[2026-06-19 10:56:28,706][starVLA.training.train_starvla][INFO] - Step 509, Loss: {'train/loss': 0.15561415255069733, 'train/grad_norm_pre_clip': 8.337629318237305, 'timing/data': 0.013187543954700232, 'timing/model': 1.1128533110022545, 'learning_rate/qwen_vl_interface': 9.837621780393734e-06, 'learning_rate/action_model': 9.837621780393733e-05, 'epoch': 1.05}) +[2026-06-19 10:56:33,083][starVLA.training.train_starvla][INFO] - Step 510, Loss: {'train/loss': 0.21682731807231903, 'train/grad_norm_pre_clip': 12.243361473083496, 'timing/data': 0.012865369906648993, 'timing/model': 1.1163614549441263, 'learning_rate/qwen_vl_interface': 9.836831367206406e-06, 'learning_rate/action_model': 9.836831367206406e-05, 'epoch': 1.05}) +[2026-06-19 10:56:37,505][starVLA.training.train_starvla][INFO] - Step 511, Loss: {'train/loss': 0.6640535593032837, 'train/grad_norm_pre_clip': 20.054996490478516, 'timing/data': 0.014047579956240952, 'timing/model': 1.1205968849826604, 'learning_rate/qwen_vl_interface': 9.836039068545962e-06, 'learning_rate/action_model': 9.836039068545961e-05, 'epoch': 1.05}) +[2026-06-19 10:56:41,963][starVLA.training.train_starvla][INFO] - Step 512, Loss: {'train/loss': 0.3836362063884735, 'train/grad_norm_pre_clip': 7.413327693939209, 'timing/data': 0.01332568796351552, 'timing/model': 1.1146593890152872, 'learning_rate/qwen_vl_interface': 9.835244884738086e-06, 'learning_rate/action_model': 9.835244884738085e-05, 'epoch': 1.05}) +[2026-06-19 10:56:46,453][starVLA.training.train_starvla][INFO] - Step 513, Loss: {'train/loss': 0.2887733578681946, 'train/grad_norm_pre_clip': 11.676817893981934, 'timing/data': 0.013828291092067957, 'timing/model': 1.2881653510266915, 'learning_rate/qwen_vl_interface': 9.83444881610924e-06, 'learning_rate/action_model': 9.834448816109238e-05, 'epoch': 1.05}) +[2026-06-19 10:56:50,757][starVLA.training.train_starvla][INFO] - Step 514, Loss: {'train/loss': 0.3073905110359192, 'train/grad_norm_pre_clip': 8.07475757598877, 'timing/data': 0.01285816088784486, 'timing/model': 1.1191645280923694, 'learning_rate/qwen_vl_interface': 9.833650862986652e-06, 'learning_rate/action_model': 9.833650862986653e-05, 'epoch': 1.06}) +[2026-06-19 10:56:55,234][starVLA.training.train_starvla][INFO] - Step 515, Loss: {'train/loss': 0.4385833144187927, 'train/grad_norm_pre_clip': 12.084188461303711, 'timing/data': 0.014577387948520482, 'timing/model': 1.1124648009426892, 'learning_rate/qwen_vl_interface': 9.832851025698335e-06, 'learning_rate/action_model': 9.832851025698335e-05, 'epoch': 1.06}) +[2026-06-19 10:56:59,731][starVLA.training.train_starvla][INFO] - Step 516, Loss: {'train/loss': 0.2245638221502304, 'train/grad_norm_pre_clip': 4.52796745300293, 'timing/data': 0.012791194953024387, 'timing/model': 1.118417456978932, 'learning_rate/qwen_vl_interface': 9.832049304573072e-06, 'learning_rate/action_model': 9.832049304573072e-05, 'epoch': 1.06}) +[2026-06-19 10:57:04,131][starVLA.training.train_starvla][INFO] - Step 517, Loss: {'train/loss': 0.2721063792705536, 'train/grad_norm_pre_clip': 5.776027202606201, 'timing/data': 0.013604899984784424, 'timing/model': 1.1092932870378718, 'learning_rate/qwen_vl_interface': 9.831245699940419e-06, 'learning_rate/action_model': 9.831245699940418e-05, 'epoch': 1.06}) +[2026-06-19 10:57:08,559][starVLA.training.train_starvla][INFO] - Step 518, Loss: {'train/loss': 0.27459821105003357, 'train/grad_norm_pre_clip': 8.199884414672852, 'timing/data': 0.012732367031276226, 'timing/model': 1.1117753860307857, 'learning_rate/qwen_vl_interface': 9.830440212130707e-06, 'learning_rate/action_model': 9.830440212130708e-05, 'epoch': 1.06}) +[2026-06-19 10:57:13,050][starVLA.training.train_starvla][INFO] - Step 519, Loss: {'train/loss': 0.4094751477241516, 'train/grad_norm_pre_clip': 9.515928268432617, 'timing/data': 0.0126714309444651, 'timing/model': 1.2948431089753285, 'learning_rate/qwen_vl_interface': 9.829632841475041e-06, 'learning_rate/action_model': 9.829632841475041e-05, 'epoch': 1.07}) +[2026-06-19 10:57:17,354][starVLA.training.train_starvla][INFO] - Step 520, Loss: {'train/loss': 0.365997850894928, 'train/grad_norm_pre_clip': 10.637288093566895, 'timing/data': 0.013403918012045324, 'timing/model': 1.115020025987178, 'learning_rate/qwen_vl_interface': 9.828823588305305e-06, 'learning_rate/action_model': 9.828823588305303e-05, 'epoch': 1.07}) +[2026-06-19 10:57:21,847][starVLA.training.train_starvla][INFO] - Step 521, Loss: {'train/loss': 0.4062087833881378, 'train/grad_norm_pre_clip': 8.486966133117676, 'timing/data': 0.013682399992831051, 'timing/model': 1.1154815099434927, 'learning_rate/qwen_vl_interface': 9.828012452954146e-06, 'learning_rate/action_model': 9.828012452954145e-05, 'epoch': 1.07}) +[2026-06-19 10:57:26,315][starVLA.training.train_starvla][INFO] - Step 522, Loss: {'train/loss': 0.37834668159484863, 'train/grad_norm_pre_clip': 4.624414443969727, 'timing/data': 0.014283078024163842, 'timing/model': 1.1143544480437413, 'learning_rate/qwen_vl_interface': 9.827199435754995e-06, 'learning_rate/action_model': 9.827199435754995e-05, 'epoch': 1.07}) +[2026-06-19 10:57:30,786][starVLA.training.train_starvla][INFO] - Step 523, Loss: {'train/loss': 0.2960298955440521, 'train/grad_norm_pre_clip': 3.2728259563446045, 'timing/data': 0.013620023964904249, 'timing/model': 1.114454669994302, 'learning_rate/qwen_vl_interface': 9.826384537042052e-06, 'learning_rate/action_model': 9.826384537042052e-05, 'epoch': 1.07}) +[2026-06-19 10:57:35,182][starVLA.training.train_starvla][INFO] - Step 524, Loss: {'train/loss': 0.1843230426311493, 'train/grad_norm_pre_clip': 2.9079933166503906, 'timing/data': 0.013890445930883288, 'timing/model': 1.1222087140195072, 'learning_rate/qwen_vl_interface': 9.825567757150291e-06, 'learning_rate/action_model': 9.82556775715029e-05, 'epoch': 1.08}) +[2026-06-19 10:57:39,637][starVLA.training.train_starvla][INFO] - Step 525, Loss: {'train/loss': 0.3632262051105499, 'train/grad_norm_pre_clip': 5.760715961456299, 'timing/data': 0.014130005962215364, 'timing/model': 1.2734767779475078, 'learning_rate/qwen_vl_interface': 9.82474909641546e-06, 'learning_rate/action_model': 9.82474909641546e-05, 'epoch': 1.08}) +[2026-06-19 10:57:44,061][starVLA.training.train_starvla][INFO] - Step 526, Loss: {'train/loss': 0.3707655668258667, 'train/grad_norm_pre_clip': 5.43882942199707, 'timing/data': 0.013407587073743343, 'timing/model': 1.115500194951892, 'learning_rate/qwen_vl_interface': 9.823928555174078e-06, 'learning_rate/action_model': 9.823928555174077e-05, 'epoch': 1.08}) +[2026-06-19 10:57:48,562][starVLA.training.train_starvla][INFO] - Step 527, Loss: {'train/loss': 0.48817679286003113, 'train/grad_norm_pre_clip': 9.109733581542969, 'timing/data': 0.013253965065814555, 'timing/model': 1.1425797280389816, 'learning_rate/qwen_vl_interface': 9.82310613376344e-06, 'learning_rate/action_model': 9.82310613376344e-05, 'epoch': 1.08}) +[2026-06-19 10:57:53,027][starVLA.training.train_starvla][INFO] - Step 528, Loss: {'train/loss': 0.4883604645729065, 'train/grad_norm_pre_clip': 7.704148769378662, 'timing/data': 0.012859179987572134, 'timing/model': 1.1113059379858896, 'learning_rate/qwen_vl_interface': 9.822281832521613e-06, 'learning_rate/action_model': 9.822281832521613e-05, 'epoch': 1.08}) +[2026-06-19 10:57:57,397][starVLA.training.train_starvla][INFO] - Step 529, Loss: {'train/loss': 0.24242812395095825, 'train/grad_norm_pre_clip': 2.7167816162109375, 'timing/data': 0.013713433989323676, 'timing/model': 1.1193002130603418, 'learning_rate/qwen_vl_interface': 9.821455651787432e-06, 'learning_rate/action_model': 9.821455651787432e-05, 'epoch': 1.09}) +[2026-06-19 10:58:01,857][starVLA.training.train_starvla][INFO] - Step 530, Loss: {'train/loss': 0.3607705235481262, 'train/grad_norm_pre_clip': 6.337961673736572, 'timing/data': 0.013679076102562249, 'timing/model': 1.119898764998652, 'learning_rate/qwen_vl_interface': 9.820627591900514e-06, 'learning_rate/action_model': 9.820627591900513e-05, 'epoch': 1.09}) +[2026-06-19 10:58:06,331][starVLA.training.train_starvla][INFO] - Step 531, Loss: {'train/loss': 0.17907661199569702, 'train/grad_norm_pre_clip': 3.053341865539551, 'timing/data': 0.01400769897736609, 'timing/model': 1.2861028190236539, 'learning_rate/qwen_vl_interface': 9.819797653201238e-06, 'learning_rate/action_model': 9.819797653201238e-05, 'epoch': 1.09}) +[2026-06-19 10:58:10,645][starVLA.training.train_starvla][INFO] - Step 532, Loss: {'train/loss': 0.2734888792037964, 'train/grad_norm_pre_clip': 5.527624130249023, 'timing/data': 0.012866556993685663, 'timing/model': 1.1174249680479988, 'learning_rate/qwen_vl_interface': 9.818965836030765e-06, 'learning_rate/action_model': 9.818965836030764e-05, 'epoch': 1.09}) +[2026-06-19 10:58:15,128][starVLA.training.train_starvla][INFO] - Step 533, Loss: {'train/loss': 0.23447397351264954, 'train/grad_norm_pre_clip': 5.389881134033203, 'timing/data': 0.013646144070662558, 'timing/model': 1.1267424259567633, 'learning_rate/qwen_vl_interface': 9.81813214073102e-06, 'learning_rate/action_model': 9.81813214073102e-05, 'epoch': 1.09}) +[2026-06-19 10:58:19,642][starVLA.training.train_starvla][INFO] - Step 534, Loss: {'train/loss': 0.3838764429092407, 'train/grad_norm_pre_clip': 8.309127807617188, 'timing/data': 0.01600345002952963, 'timing/model': 1.1260599039960653, 'learning_rate/qwen_vl_interface': 9.817296567644709e-06, 'learning_rate/action_model': 9.817296567644709e-05, 'epoch': 1.1}) +[2026-06-19 10:58:24,100][starVLA.training.train_starvla][INFO] - Step 535, Loss: {'train/loss': 0.2765764892101288, 'train/grad_norm_pre_clip': 4.712152004241943, 'timing/data': 0.013111341977491975, 'timing/model': 1.1234408450545743, 'learning_rate/qwen_vl_interface': 9.816459117115296e-06, 'learning_rate/action_model': 9.816459117115297e-05, 'epoch': 1.1}) +[2026-06-19 10:58:28,574][starVLA.training.train_starvla][INFO] - Step 536, Loss: {'train/loss': 0.36538979411125183, 'train/grad_norm_pre_clip': 18.401897430419922, 'timing/data': 0.013991042040288448, 'timing/model': 1.1153656059177592, 'learning_rate/qwen_vl_interface': 9.815619789487035e-06, 'learning_rate/action_model': 9.815619789487034e-05, 'epoch': 1.1}) +[2026-06-19 10:58:33,033][starVLA.training.train_starvla][INFO] - Step 537, Loss: {'train/loss': 0.28355109691619873, 'train/grad_norm_pre_clip': 11.605774879455566, 'timing/data': 0.01400735997594893, 'timing/model': 1.275791333988309, 'learning_rate/qwen_vl_interface': 9.814778585104933e-06, 'learning_rate/action_model': 9.814778585104933e-05, 'epoch': 1.1}) +[2026-06-19 10:58:37,345][starVLA.training.train_starvla][INFO] - Step 538, Loss: {'train/loss': 0.6317718625068665, 'train/grad_norm_pre_clip': 21.266292572021484, 'timing/data': 0.013648906955495477, 'timing/model': 1.1253486040513963, 'learning_rate/qwen_vl_interface': 9.813935504314784e-06, 'learning_rate/action_model': 9.813935504314784e-05, 'epoch': 1.1}) +[2026-06-19 10:58:41,840][starVLA.training.train_starvla][INFO] - Step 539, Loss: {'train/loss': 0.33312588930130005, 'train/grad_norm_pre_clip': 14.138894081115723, 'timing/data': 0.015699479030445218, 'timing/model': 1.125904661952518, 'learning_rate/qwen_vl_interface': 9.813090547463144e-06, 'learning_rate/action_model': 9.813090547463143e-05, 'epoch': 1.11}) +[2026-06-19 10:58:46,287][starVLA.training.train_starvla][INFO] - Step 540, Loss: {'train/loss': 0.1683187633752823, 'train/grad_norm_pre_clip': 6.194839954376221, 'timing/data': 0.014470451977103949, 'timing/model': 1.123069388908334, 'learning_rate/qwen_vl_interface': 9.812243714897342e-06, 'learning_rate/action_model': 9.812243714897343e-05, 'epoch': 1.11}) +[2026-06-19 10:58:50,751][starVLA.training.train_starvla][INFO] - Step 541, Loss: {'train/loss': 0.5035234093666077, 'train/grad_norm_pre_clip': 31.05628776550293, 'timing/data': 0.012421621941030025, 'timing/model': 1.1184192310320213, 'learning_rate/qwen_vl_interface': 9.811395006965481e-06, 'learning_rate/action_model': 9.81139500696548e-05, 'epoch': 1.11}) +[2026-06-19 10:58:55,259][starVLA.training.train_starvla][INFO] - Step 542, Loss: {'train/loss': 0.22882863879203796, 'train/grad_norm_pre_clip': 6.61130428314209, 'timing/data': 0.015231100958772004, 'timing/model': 1.1227973659988493, 'learning_rate/qwen_vl_interface': 9.81054442401643e-06, 'learning_rate/action_model': 9.81054442401643e-05, 'epoch': 1.11}) +[2026-06-19 10:58:59,766][starVLA.training.train_starvla][INFO] - Step 543, Loss: {'train/loss': 0.07945863902568817, 'train/grad_norm_pre_clip': 3.8180243968963623, 'timing/data': 0.01590875699184835, 'timing/model': 1.2951701000565663, 'learning_rate/qwen_vl_interface': 9.809691966399834e-06, 'learning_rate/action_model': 9.809691966399835e-05, 'epoch': 1.12}) +[2026-06-19 10:59:04,092][starVLA.training.train_starvla][INFO] - Step 544, Loss: {'train/loss': 0.27144041657447815, 'train/grad_norm_pre_clip': 10.27269172668457, 'timing/data': 0.013639417011290789, 'timing/model': 1.125682566082105, 'learning_rate/qwen_vl_interface': 9.808837634466104e-06, 'learning_rate/action_model': 9.808837634466104e-05, 'epoch': 1.12}) +[2026-06-19 10:59:08,594][starVLA.training.train_starvla][INFO] - Step 545, Loss: {'train/loss': 0.37974774837493896, 'train/grad_norm_pre_clip': 11.096332550048828, 'timing/data': 0.014433205942623317, 'timing/model': 1.1234401599504054, 'learning_rate/qwen_vl_interface': 9.807981428566428e-06, 'learning_rate/action_model': 9.807981428566427e-05, 'epoch': 1.12}) +[2026-06-19 10:59:13,063][starVLA.training.train_starvla][INFO] - Step 546, Loss: {'train/loss': 0.15853925049304962, 'train/grad_norm_pre_clip': 7.178233623504639, 'timing/data': 0.013870457070879638, 'timing/model': 1.130887608975172, 'learning_rate/qwen_vl_interface': 9.807123349052754e-06, 'learning_rate/action_model': 9.807123349052753e-05, 'epoch': 1.12}) +[2026-06-19 10:59:17,554][starVLA.training.train_starvla][INFO] - Step 547, Loss: {'train/loss': 0.7530651092529297, 'train/grad_norm_pre_clip': 9.535487174987793, 'timing/data': 0.013398168957792222, 'timing/model': 1.123267372022383, 'learning_rate/qwen_vl_interface': 9.80626339627781e-06, 'learning_rate/action_model': 9.806263396277809e-05, 'epoch': 1.12}) +[2026-06-19 10:59:22,094][starVLA.training.train_starvla][INFO] - Step 548, Loss: {'train/loss': 0.3171099126338959, 'train/grad_norm_pre_clip': 8.352410316467285, 'timing/data': 0.01371682994067669, 'timing/model': 1.1206798510393128, 'learning_rate/qwen_vl_interface': 9.805401570595087e-06, 'learning_rate/action_model': 9.805401570595087e-05, 'epoch': 1.13}) +[2026-06-19 10:59:26,612][starVLA.training.train_starvla][INFO] - Step 549, Loss: {'train/loss': 0.15452784299850464, 'train/grad_norm_pre_clip': 5.21392297744751, 'timing/data': 0.013819682062603533, 'timing/model': 1.2877346239984035, 'learning_rate/qwen_vl_interface': 9.804537872358853e-06, 'learning_rate/action_model': 9.804537872358853e-05, 'epoch': 1.13}) +[2026-06-19 10:59:31,052][starVLA.training.train_starvla][INFO] - Step 550, Loss: {'train/loss': 0.09529796242713928, 'train/grad_norm_pre_clip': 4.004861831665039, 'timing/data': 0.015712407999671996, 'timing/model': 1.2323695870582014, 'learning_rate/qwen_vl_interface': 9.80367230192414e-06, 'learning_rate/action_model': 9.803672301924138e-05, 'epoch': 1.13}) +[2026-06-19 10:59:35,496][starVLA.training.train_starvla][INFO] - Step 551, Loss: {'train/loss': 0.16766972839832306, 'train/grad_norm_pre_clip': 5.6586995124816895, 'timing/data': 0.01381527108605951, 'timing/model': 1.1182957870187238, 'learning_rate/qwen_vl_interface': 9.80280485964675e-06, 'learning_rate/action_model': 9.80280485964675e-05, 'epoch': 1.13}) +[2026-06-19 10:59:39,962][starVLA.training.train_starvla][INFO] - Step 552, Loss: {'train/loss': 0.2081647515296936, 'train/grad_norm_pre_clip': 8.497674942016602, 'timing/data': 0.01460292807314545, 'timing/model': 1.1300253159133717, 'learning_rate/qwen_vl_interface': 9.801935545883255e-06, 'learning_rate/action_model': 9.801935545883255e-05, 'epoch': 1.13}) +[2026-06-19 10:59:44,470][starVLA.training.train_starvla][INFO] - Step 553, Loss: {'train/loss': 0.11749425530433655, 'train/grad_norm_pre_clip': 3.558131456375122, 'timing/data': 0.012901255977340043, 'timing/model': 1.1200708859832957, 'learning_rate/qwen_vl_interface': 9.801064360991e-06, 'learning_rate/action_model': 9.801064360991e-05, 'epoch': 1.14}) +[2026-06-19 10:59:48,963][starVLA.training.train_starvla][INFO] - Step 554, Loss: {'train/loss': 0.17172415554523468, 'train/grad_norm_pre_clip': 3.602766275405884, 'timing/data': 0.01447563199326396, 'timing/model': 1.1227801480563357, 'learning_rate/qwen_vl_interface': 9.800191305328095e-06, 'learning_rate/action_model': 9.800191305328094e-05, 'epoch': 1.14}) +[2026-06-19 10:59:53,464][starVLA.training.train_starvla][INFO] - Step 555, Loss: {'train/loss': 0.1454034149646759, 'train/grad_norm_pre_clip': 4.732665538787842, 'timing/data': 0.012783800950273871, 'timing/model': 1.2953952080570161, 'learning_rate/qwen_vl_interface': 9.79931637925342e-06, 'learning_rate/action_model': 9.79931637925342e-05, 'epoch': 1.14}) +[2026-06-19 10:59:57,904][starVLA.training.train_starvla][INFO] - Step 556, Loss: {'train/loss': 0.22236748039722443, 'train/grad_norm_pre_clip': 5.957350254058838, 'timing/data': 0.01448980497661978, 'timing/model': 1.242948213010095, 'learning_rate/qwen_vl_interface': 9.798439583126625e-06, 'learning_rate/action_model': 9.798439583126625e-05, 'epoch': 1.14}) +[2026-06-19 11:00:02,359][starVLA.training.train_starvla][INFO] - Step 557, Loss: {'train/loss': 0.1277962177991867, 'train/grad_norm_pre_clip': 2.5201077461242676, 'timing/data': 0.017268394003622234, 'timing/model': 1.1424140639137477, 'learning_rate/qwen_vl_interface': 9.797560917308126e-06, 'learning_rate/action_model': 9.797560917308126e-05, 'epoch': 1.14}) +[2026-06-19 11:00:06,819][starVLA.training.train_starvla][INFO] - Step 558, Loss: {'train/loss': 0.23082304000854492, 'train/grad_norm_pre_clip': 5.5067338943481445, 'timing/data': 0.013975325040519238, 'timing/model': 1.1194691200507805, 'learning_rate/qwen_vl_interface': 9.796680382159112e-06, 'learning_rate/action_model': 9.79668038215911e-05, 'epoch': 1.15}) +[2026-06-19 11:00:11,284][starVLA.training.train_starvla][INFO] - Step 559, Loss: {'train/loss': 0.21759635210037231, 'train/grad_norm_pre_clip': 5.441479206085205, 'timing/data': 0.012442409992218018, 'timing/model': 1.1197538840351626, 'learning_rate/qwen_vl_interface': 9.795797978041533e-06, 'learning_rate/action_model': 9.795797978041533e-05, 'epoch': 1.15}) +[2026-06-19 11:00:15,779][starVLA.training.train_starvla][INFO] - Step 560, Loss: {'train/loss': 0.29859089851379395, 'train/grad_norm_pre_clip': 9.75769329071045, 'timing/data': 0.013554772944189608, 'timing/model': 1.1183280609548092, 'learning_rate/qwen_vl_interface': 9.794913705318117e-06, 'learning_rate/action_model': 9.794913705318118e-05, 'epoch': 1.15}) +[2026-06-19 11:00:20,260][starVLA.training.train_starvla][INFO] - Step 561, Loss: {'train/loss': 0.11892461776733398, 'train/grad_norm_pre_clip': 3.192652940750122, 'timing/data': 0.013208032003603876, 'timing/model': 1.283243953017518, 'learning_rate/qwen_vl_interface': 9.794027564352352e-06, 'learning_rate/action_model': 9.794027564352352e-05, 'epoch': 1.15}) +[2026-06-19 11:00:24,590][starVLA.training.train_starvla][INFO] - Step 562, Loss: {'train/loss': 0.19886067509651184, 'train/grad_norm_pre_clip': 8.108552932739258, 'timing/data': 0.013225674047134817, 'timing/model': 1.1226066469680518, 'learning_rate/qwen_vl_interface': 9.793139555508501e-06, 'learning_rate/action_model': 9.7931395555085e-05, 'epoch': 1.15}) +[2026-06-19 11:00:29,086][starVLA.training.train_starvla][INFO] - Step 563, Loss: {'train/loss': 0.15150001645088196, 'train/grad_norm_pre_clip': 8.806904792785645, 'timing/data': 0.014250143081881106, 'timing/model': 1.1234637370798737, 'learning_rate/qwen_vl_interface': 9.792249679151585e-06, 'learning_rate/action_model': 9.792249679151585e-05, 'epoch': 1.16}) +[2026-06-19 11:00:33,599][starVLA.training.train_starvla][INFO] - Step 564, Loss: {'train/loss': 0.2512894868850708, 'train/grad_norm_pre_clip': 9.253201484680176, 'timing/data': 0.014034113963134587, 'timing/model': 1.1393941839924082, 'learning_rate/qwen_vl_interface': 9.791357935647403e-06, 'learning_rate/action_model': 9.791357935647403e-05, 'epoch': 1.16}) +[2026-06-19 11:00:38,105][starVLA.training.train_starvla][INFO] - Step 565, Loss: {'train/loss': 0.3418883979320526, 'train/grad_norm_pre_clip': 12.328533172607422, 'timing/data': 0.012766041909344494, 'timing/model': 1.1239478579955176, 'learning_rate/qwen_vl_interface': 9.790464325362516e-06, 'learning_rate/action_model': 9.790464325362516e-05, 'epoch': 1.16}) +[2026-06-19 11:00:42,534][starVLA.training.train_starvla][INFO] - Step 566, Loss: {'train/loss': 0.1866520792245865, 'train/grad_norm_pre_clip': 10.064162254333496, 'timing/data': 0.013371223933063447, 'timing/model': 1.122071202029474, 'learning_rate/qwen_vl_interface': 9.789568848664254e-06, 'learning_rate/action_model': 9.789568848664253e-05, 'epoch': 1.16}) +[2026-06-19 11:00:46,973][starVLA.training.train_starvla][INFO] - Step 567, Loss: {'train/loss': 0.409665584564209, 'train/grad_norm_pre_clip': 10.618468284606934, 'timing/data': 0.013590210932306945, 'timing/model': 1.2368167369859293, 'learning_rate/qwen_vl_interface': 9.788671505920712e-06, 'learning_rate/action_model': 9.788671505920711e-05, 'epoch': 1.16}) +[2026-06-19 11:00:51,319][starVLA.training.train_starvla][INFO] - Step 568, Loss: {'train/loss': 0.10622680187225342, 'train/grad_norm_pre_clip': 6.838980674743652, 'timing/data': 0.013289578026160598, 'timing/model': 1.1398065759567544, 'learning_rate/qwen_vl_interface': 9.787772297500754e-06, 'learning_rate/action_model': 9.787772297500754e-05, 'epoch': 1.17}) +[2026-06-19 11:00:55,788][starVLA.training.train_starvla][INFO] - Step 569, Loss: {'train/loss': 0.2507788836956024, 'train/grad_norm_pre_clip': 5.816616058349609, 'timing/data': 0.013780960929580033, 'timing/model': 1.1183720610570163, 'learning_rate/qwen_vl_interface': 9.78687122377401e-06, 'learning_rate/action_model': 9.78687122377401e-05, 'epoch': 1.17}) +[2026-06-19 11:01:00,299][starVLA.training.train_starvla][INFO] - Step 570, Loss: {'train/loss': 0.1601085364818573, 'train/grad_norm_pre_clip': 4.744561195373535, 'timing/data': 0.014600250986404717, 'timing/model': 1.1214629830792546, 'learning_rate/qwen_vl_interface': 9.78596828511088e-06, 'learning_rate/action_model': 9.78596828511088e-05, 'epoch': 1.17}) +[2026-06-19 11:01:04,803][starVLA.training.train_starvla][INFO] - Step 571, Loss: {'train/loss': 0.18554092943668365, 'train/grad_norm_pre_clip': 5.147105693817139, 'timing/data': 0.013824274996295571, 'timing/model': 1.1309988839784637, 'learning_rate/qwen_vl_interface': 9.785063481882527e-06, 'learning_rate/action_model': 9.785063481882527e-05, 'epoch': 1.17}) +[2026-06-19 11:01:09,235][starVLA.training.train_starvla][INFO] - Step 572, Loss: {'train/loss': 0.12408380955457687, 'train/grad_norm_pre_clip': 3.582078218460083, 'timing/data': 0.012748011038638651, 'timing/model': 1.1179774069460109, 'learning_rate/qwen_vl_interface': 9.78415681446088e-06, 'learning_rate/action_model': 9.784156814460879e-05, 'epoch': 1.17}) +[2026-06-19 11:01:13,691][starVLA.training.train_starvla][INFO] - Step 573, Loss: {'train/loss': 0.15987171232700348, 'train/grad_norm_pre_clip': 5.030451774597168, 'timing/data': 0.013187821954488754, 'timing/model': 1.2583304799627513, 'learning_rate/qwen_vl_interface': 9.783248283218638e-06, 'learning_rate/action_model': 9.783248283218638e-05, 'epoch': 1.18}) +[2026-06-19 11:01:18,007][starVLA.training.train_starvla][INFO] - Step 574, Loss: {'train/loss': 0.31950581073760986, 'train/grad_norm_pre_clip': 10.615184783935547, 'timing/data': 0.013704375014640391, 'timing/model': 1.1215169860515743, 'learning_rate/qwen_vl_interface': 9.782337888529261e-06, 'learning_rate/action_model': 9.782337888529261e-05, 'epoch': 1.18}) +[2026-06-19 11:01:22,496][starVLA.training.train_starvla][INFO] - Step 575, Loss: {'train/loss': 0.20689091086387634, 'train/grad_norm_pre_clip': 13.475370407104492, 'timing/data': 0.01447106001432985, 'timing/model': 1.1188057319959626, 'learning_rate/qwen_vl_interface': 9.781425630766978e-06, 'learning_rate/action_model': 9.781425630766979e-05, 'epoch': 1.18}) +[2026-06-19 11:01:26,998][starVLA.training.train_starvla][INFO] - Step 576, Loss: {'train/loss': 0.42728906869888306, 'train/grad_norm_pre_clip': 12.502979278564453, 'timing/data': 0.014720942010171711, 'timing/model': 1.1202949890866876, 'learning_rate/qwen_vl_interface': 9.780511510306789e-06, 'learning_rate/action_model': 9.780511510306788e-05, 'epoch': 1.18}) +[2026-06-19 11:01:31,498][starVLA.training.train_starvla][INFO] - Step 577, Loss: {'train/loss': 0.2667616307735443, 'train/grad_norm_pre_clip': 14.195562362670898, 'timing/data': 0.013982948963530362, 'timing/model': 1.1264538979157805, 'learning_rate/qwen_vl_interface': 9.779595527524447e-06, 'learning_rate/action_model': 9.779595527524447e-05, 'epoch': 1.19}) +[2026-06-19 11:01:36,004][starVLA.training.train_starvla][INFO] - Step 578, Loss: {'train/loss': 0.1999119073152542, 'train/grad_norm_pre_clip': 5.585361003875732, 'timing/data': 0.015039552003145218, 'timing/model': 1.307101206970401, 'learning_rate/qwen_vl_interface': 9.778677682796483e-06, 'learning_rate/action_model': 9.778677682796483e-05, 'epoch': 1.19}) +[2026-06-19 11:01:40,338][starVLA.training.train_starvla][INFO] - Step 579, Loss: {'train/loss': 0.18282848596572876, 'train/grad_norm_pre_clip': 7.811284065246582, 'timing/data': 0.013352606096304953, 'timing/model': 1.1225898490520194, 'learning_rate/qwen_vl_interface': 9.777757976500185e-06, 'learning_rate/action_model': 9.777757976500186e-05, 'epoch': 1.19}) +[2026-06-19 11:01:44,826][starVLA.training.train_starvla][INFO] - Step 580, Loss: {'train/loss': 0.10892324894666672, 'train/grad_norm_pre_clip': 4.867155075073242, 'timing/data': 0.014351445017382503, 'timing/model': 1.1192743830615655, 'learning_rate/qwen_vl_interface': 9.776836409013614e-06, 'learning_rate/action_model': 9.776836409013615e-05, 'epoch': 1.19}) +[2026-06-19 11:01:49,237][starVLA.training.train_starvla][INFO] - Step 581, Loss: {'train/loss': 0.10001359134912491, 'train/grad_norm_pre_clip': 2.692894220352173, 'timing/data': 0.014276825939305127, 'timing/model': 1.1204237469937652, 'learning_rate/qwen_vl_interface': 9.775912980715589e-06, 'learning_rate/action_model': 9.775912980715588e-05, 'epoch': 1.19}) +[2026-06-19 11:01:53,661][starVLA.training.train_starvla][INFO] - Step 582, Loss: {'train/loss': 0.19895055890083313, 'train/grad_norm_pre_clip': 6.533532619476318, 'timing/data': 0.01486054202541709, 'timing/model': 1.122002471005544, 'learning_rate/qwen_vl_interface': 9.774987691985694e-06, 'learning_rate/action_model': 9.774987691985694e-05, 'epoch': 1.2}) +[2026-06-19 11:01:58,129][starVLA.training.train_starvla][INFO] - Step 583, Loss: {'train/loss': 0.32431840896606445, 'train/grad_norm_pre_clip': 10.84144115447998, 'timing/data': 0.013785868068225682, 'timing/model': 1.1353300819173455, 'learning_rate/qwen_vl_interface': 9.774060543204283e-06, 'learning_rate/action_model': 9.774060543204283e-05, 'epoch': 1.2}) +[2026-06-19 11:02:02,649][starVLA.training.train_starvla][INFO] - Step 584, Loss: {'train/loss': 0.19948844611644745, 'train/grad_norm_pre_clip': 6.150760650634766, 'timing/data': 0.020171829964965582, 'timing/model': 1.2938544820062816, 'learning_rate/qwen_vl_interface': 9.773131534752472e-06, 'learning_rate/action_model': 9.773131534752472e-05, 'epoch': 1.2}) +[2026-06-19 11:02:06,966][starVLA.training.train_starvla][INFO] - Step 585, Loss: {'train/loss': 0.43872103095054626, 'train/grad_norm_pre_clip': 12.021623611450195, 'timing/data': 0.01426189101766795, 'timing/model': 1.121149315033108, 'learning_rate/qwen_vl_interface': 9.772200667012142e-06, 'learning_rate/action_model': 9.772200667012142e-05, 'epoch': 1.2}) +[2026-06-19 11:02:11,454][starVLA.training.train_starvla][INFO] - Step 586, Loss: {'train/loss': 0.08802098035812378, 'train/grad_norm_pre_clip': 4.043765068054199, 'timing/data': 0.013141116010956466, 'timing/model': 1.1227714860579, 'learning_rate/qwen_vl_interface': 9.771267940365934e-06, 'learning_rate/action_model': 9.771267940365935e-05, 'epoch': 1.2}) +[2026-06-19 11:02:15,936][starVLA.training.train_starvla][INFO] - Step 587, Loss: {'train/loss': 0.4241139888763428, 'train/grad_norm_pre_clip': 7.005833148956299, 'timing/data': 0.013784818002022803, 'timing/model': 1.1207400390412658, 'learning_rate/qwen_vl_interface': 9.770333355197262e-06, 'learning_rate/action_model': 9.77033335519726e-05, 'epoch': 1.21}) +[2026-06-19 11:02:20,432][starVLA.training.train_starvla][INFO] - Step 588, Loss: {'train/loss': 0.45492151379585266, 'train/grad_norm_pre_clip': 12.028454780578613, 'timing/data': 0.013418787042610347, 'timing/model': 1.1217029129620641, 'learning_rate/qwen_vl_interface': 9.769396911890291e-06, 'learning_rate/action_model': 9.769396911890291e-05, 'epoch': 1.21}) +[2026-06-19 11:02:24,920][starVLA.training.train_starvla][INFO] - Step 589, Loss: {'train/loss': 0.21275825798511505, 'train/grad_norm_pre_clip': 5.660711288452148, 'timing/data': 0.013875172007828951, 'timing/model': 1.1318279469851404, 'learning_rate/qwen_vl_interface': 9.768458610829965e-06, 'learning_rate/action_model': 9.768458610829965e-05, 'epoch': 1.21}) +[2026-06-19 11:02:29,445][starVLA.training.train_starvla][INFO] - Step 590, Loss: {'train/loss': 0.21407034993171692, 'train/grad_norm_pre_clip': 6.174361705780029, 'timing/data': 0.015085530001670122, 'timing/model': 1.3027096560690552, 'learning_rate/qwen_vl_interface': 9.767518452401981e-06, 'learning_rate/action_model': 9.76751845240198e-05, 'epoch': 1.21}) +[2026-06-19 11:02:33,783][starVLA.training.train_starvla][INFO] - Step 591, Loss: {'train/loss': 0.30738070607185364, 'train/grad_norm_pre_clip': 6.483307361602783, 'timing/data': 0.014413926983252168, 'timing/model': 1.1272214780328795, 'learning_rate/qwen_vl_interface': 9.766576436992802e-06, 'learning_rate/action_model': 9.766576436992801e-05, 'epoch': 1.21}) +[2026-06-19 11:02:38,278][starVLA.training.train_starvla][INFO] - Step 592, Loss: {'train/loss': 0.2701549232006073, 'train/grad_norm_pre_clip': 5.535804271697998, 'timing/data': 0.01350064598955214, 'timing/model': 1.1309550780570135, 'learning_rate/qwen_vl_interface': 9.765632564989657e-06, 'learning_rate/action_model': 9.765632564989656e-05, 'epoch': 1.22}) +[2026-06-19 11:02:42,755][starVLA.training.train_starvla][INFO] - Step 593, Loss: {'train/loss': 0.2511904239654541, 'train/grad_norm_pre_clip': 6.39338493347168, 'timing/data': 0.01383699500001967, 'timing/model': 1.1150396219454706, 'learning_rate/qwen_vl_interface': 9.764686836780533e-06, 'learning_rate/action_model': 9.764686836780532e-05, 'epoch': 1.22}) +[2026-06-19 11:02:47,261][starVLA.training.train_starvla][INFO] - Step 594, Loss: {'train/loss': 0.37471434473991394, 'train/grad_norm_pre_clip': 4.653260707855225, 'timing/data': 0.012984580011107028, 'timing/model': 1.1270822969963774, 'learning_rate/qwen_vl_interface': 9.763739252754187e-06, 'learning_rate/action_model': 9.763739252754186e-05, 'epoch': 1.22}) +[2026-06-19 11:02:51,669][starVLA.training.train_starvla][INFO] - Step 595, Loss: {'train/loss': 0.23989062011241913, 'train/grad_norm_pre_clip': 4.493363857269287, 'timing/data': 0.014345920993946493, 'timing/model': 1.1334753700066358, 'learning_rate/qwen_vl_interface': 9.76278981330013e-06, 'learning_rate/action_model': 9.762789813300131e-05, 'epoch': 1.22}) +[2026-06-19 11:02:56,178][starVLA.training.train_starvla][INFO] - Step 596, Loss: {'train/loss': 0.24356727302074432, 'train/grad_norm_pre_clip': 4.165641784667969, 'timing/data': 0.013194867060519755, 'timing/model': 1.309377979952842, 'learning_rate/qwen_vl_interface': 9.761838518808647e-06, 'learning_rate/action_model': 9.761838518808646e-05, 'epoch': 1.22}) +[2026-06-19 11:03:00,529][starVLA.training.train_starvla][INFO] - Step 597, Loss: {'train/loss': 0.20015151798725128, 'train/grad_norm_pre_clip': 7.120021820068359, 'timing/data': 0.013398401904851198, 'timing/model': 1.140909653971903, 'learning_rate/qwen_vl_interface': 9.760885369670776e-06, 'learning_rate/action_model': 9.760885369670776e-05, 'epoch': 1.23}) +[2026-06-19 11:03:05,026][starVLA.training.train_starvla][INFO] - Step 598, Loss: {'train/loss': 0.2498457431793213, 'train/grad_norm_pre_clip': 6.246856689453125, 'timing/data': 0.01341910904739052, 'timing/model': 1.1219604500802234, 'learning_rate/qwen_vl_interface': 9.759930366278319e-06, 'learning_rate/action_model': 9.759930366278318e-05, 'epoch': 1.23}) +[2026-06-19 11:03:09,508][starVLA.training.train_starvla][INFO] - Step 599, Loss: {'train/loss': 0.285295695066452, 'train/grad_norm_pre_clip': 5.034110069274902, 'timing/data': 0.013164974981918931, 'timing/model': 1.1188965949695557, 'learning_rate/qwen_vl_interface': 9.758973509023842e-06, 'learning_rate/action_model': 9.758973509023843e-05, 'epoch': 1.23}) +[2026-06-19 11:03:13,908][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 600 +[2026-06-19 11:06:09,524][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 600 +[2026-06-19 11:06:09,524][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 600 +[2026-06-19 11:09:16,931][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 600 +[2026-06-19 11:09:16,933][starVLA.training.train_starvla][INFO] - Step 600, Loss: {'train/loss': 0.2800547480583191, 'train/grad_norm_pre_clip': 4.8265910148620605, 'eval/loss': 0.32682496309280396, 'eval/latency_0/loss': 0.10706781595945358, 'eval/latency_1/loss': 0.21689856052398682, 'eval/latency_2/loss': 0.4157688319683075, 'eval/latency_3/loss': 0.3753511309623718, 'eval/latency_4/loss': 0.5046993494033813, 'eval/flappy/loss': 0.32682496309280396, 'eval/flappy/latency_0/loss': 0.10706781595945358, 'eval/flappy/latency_1/loss': 0.21689856052398682, 'eval/flappy/latency_2/loss': 0.4157688319683075, 'eval/flappy/latency_3/loss': 0.3753511309623718, 'eval/flappy/latency_4/loss': 0.5046993494033813, 'eval/flap_precision': 0.7428571428571429, 'eval/flap_recall': 0.5672727272727273, 'eval/flap_f1': 0.6432989690721649, 'eval/flap_rate_model': 0.058333333333333334, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.6432989690721649, 'eval/latency_0/flap_precision': 0.7428571428571429, 'eval/latency_0/flap_recall': 0.5672727272727273, 'eval/latency_0/flap_f1': 0.6432989690721649, 'eval/latency_0/flap_rate_model': 0.058333333333333334, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.6432989690721649, 'timing/data': 0.012542003998532891, 'timing/model': 1.1321275659138337, 'learning_rate/qwen_vl_interface': 9.758014798300677e-06, 'learning_rate/action_model': 9.758014798300676e-05, 'epoch': 1.23}) +[2026-06-19 11:09:16,933][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_600_state +[2026-06-19 11:09:16,937][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-19 11:09:20,310][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_600_state/model.safetensors +[2026-06-19 11:09:27,504][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_600_state/optimizer.bin +[2026-06-19 11:09:27,504][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_600_state/sampler.bin +[2026-06-19 11:09:27,504][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_600_state/sampler_1.bin +[2026-06-19 11:09:27,505][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_600_state/random_states_0.pkl +[2026-06-19 11:09:27,530][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 600 +[2026-06-19 11:09:27,530][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-19 11:09:27,540][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-19 11:09:32,028][starVLA.training.train_starvla][INFO] - Step 601, Loss: {'train/loss': 0.16439759731292725, 'train/grad_norm_pre_clip': 8.468353271484375, 'timing/data': 0.012862219009548426, 'timing/model': 1.1116870240075514, 'learning_rate/qwen_vl_interface': 9.757054234502909e-06, 'learning_rate/action_model': 9.757054234502909e-05, 'epoch': 1.23}) +[2026-06-19 11:09:36,441][starVLA.training.train_starvla][INFO] - Step 602, Loss: {'train/loss': 0.34699639678001404, 'train/grad_norm_pre_clip': 7.859989643096924, 'timing/data': 0.013288101996295154, 'timing/model': 1.2780380039475858, 'learning_rate/qwen_vl_interface': 9.756091818025394e-06, 'learning_rate/action_model': 9.756091818025394e-05, 'epoch': 1.24}) +[2026-06-19 11:09:40,805][starVLA.training.train_starvla][INFO] - Step 603, Loss: {'train/loss': 0.33488181233406067, 'train/grad_norm_pre_clip': 10.326505661010742, 'timing/data': 0.013761076028458774, 'timing/model': 1.209053422906436, 'learning_rate/qwen_vl_interface': 9.75512754926374e-06, 'learning_rate/action_model': 9.755127549263741e-05, 'epoch': 1.24}) +[2026-06-19 11:09:45,094][starVLA.training.train_starvla][INFO] - Step 604, Loss: {'train/loss': 0.11121740192174911, 'train/grad_norm_pre_clip': 4.224330425262451, 'timing/data': 0.013390409993007779, 'timing/model': 1.121818479965441, 'learning_rate/qwen_vl_interface': 9.754161428614325e-06, 'learning_rate/action_model': 9.754161428614324e-05, 'epoch': 1.24}) +[2026-06-19 11:09:49,488][starVLA.training.train_starvla][INFO] - Step 605, Loss: {'train/loss': 0.1215895563364029, 'train/grad_norm_pre_clip': 3.303701877593994, 'timing/data': 0.01409706100821495, 'timing/model': 1.1044121410232037, 'learning_rate/qwen_vl_interface': 9.753193456474284e-06, 'learning_rate/action_model': 9.753193456474284e-05, 'epoch': 1.24}) +[2026-06-19 11:09:53,920][starVLA.training.train_starvla][INFO] - Step 606, Loss: {'train/loss': 0.1381601244211197, 'train/grad_norm_pre_clip': 4.780540466308594, 'timing/data': 0.012264926102943718, 'timing/model': 1.111759324092418, 'learning_rate/qwen_vl_interface': 9.752223633241514e-06, 'learning_rate/action_model': 9.752223633241514e-05, 'epoch': 1.24}) +[2026-06-19 11:09:58,397][starVLA.training.train_starvla][INFO] - Step 607, Loss: {'train/loss': 0.07710503041744232, 'train/grad_norm_pre_clip': 5.067285060882568, 'timing/data': 0.013602565042674541, 'timing/model': 1.1217796279815957, 'learning_rate/qwen_vl_interface': 9.751251959314671e-06, 'learning_rate/action_model': 9.751251959314672e-05, 'epoch': 1.25}) +[2026-06-19 11:10:02,871][starVLA.training.train_starvla][INFO] - Step 608, Loss: {'train/loss': 0.300046443939209, 'train/grad_norm_pre_clip': 9.498255729675293, 'timing/data': 0.01299770106561482, 'timing/model': 1.2984461140586063, 'learning_rate/qwen_vl_interface': 9.750278435093176e-06, 'learning_rate/action_model': 9.750278435093177e-05, 'epoch': 1.25}) +[2026-06-19 11:10:07,264][starVLA.training.train_starvla][INFO] - Step 609, Loss: {'train/loss': 0.3046574890613556, 'train/grad_norm_pre_clip': 12.994891166687012, 'timing/data': 0.01376172504387796, 'timing/model': 1.2357282319571823, 'learning_rate/qwen_vl_interface': 9.749303060977205e-06, 'learning_rate/action_model': 9.749303060977205e-05, 'epoch': 1.25}) +[2026-06-19 11:10:11,530][starVLA.training.train_starvla][INFO] - Step 610, Loss: {'train/loss': 0.4845277965068817, 'train/grad_norm_pre_clip': 16.071191787719727, 'timing/data': 0.01369932305533439, 'timing/model': 1.108336808043532, 'learning_rate/qwen_vl_interface': 9.748325837367701e-06, 'learning_rate/action_model': 9.748325837367702e-05, 'epoch': 1.25}) +[2026-06-19 11:10:15,958][starVLA.training.train_starvla][INFO] - Step 611, Loss: {'train/loss': 0.32681289315223694, 'train/grad_norm_pre_clip': 12.406408309936523, 'timing/data': 0.013599226949736476, 'timing/model': 1.116081066080369, 'learning_rate/qwen_vl_interface': 9.747346764666362e-06, 'learning_rate/action_model': 9.747346764666361e-05, 'epoch': 1.25}) +[2026-06-19 11:10:20,418][starVLA.training.train_starvla][INFO] - Step 612, Loss: {'train/loss': 0.22413374483585358, 'train/grad_norm_pre_clip': 13.867388725280762, 'timing/data': 0.014338248991407454, 'timing/model': 1.1123499990208074, 'learning_rate/qwen_vl_interface': 9.746365843275648e-06, 'learning_rate/action_model': 9.746365843275646e-05, 'epoch': 1.26}) +[2026-06-19 11:10:24,887][starVLA.training.train_starvla][INFO] - Step 613, Loss: {'train/loss': 0.2092708796262741, 'train/grad_norm_pre_clip': 8.803979873657227, 'timing/data': 0.01392091903835535, 'timing/model': 1.1098811259726062, 'learning_rate/qwen_vl_interface': 9.745383073598776e-06, 'learning_rate/action_model': 9.745383073598775e-05, 'epoch': 1.26}) +[2026-06-19 11:10:29,330][starVLA.training.train_starvla][INFO] - Step 614, Loss: {'train/loss': 0.24783174693584442, 'train/grad_norm_pre_clip': 9.50412654876709, 'timing/data': 0.013336554984562099, 'timing/model': 1.2758761750301346, 'learning_rate/qwen_vl_interface': 9.744398456039729e-06, 'learning_rate/action_model': 9.744398456039729e-05, 'epoch': 1.26}) +[2026-06-19 11:10:33,625][starVLA.training.train_starvla][INFO] - Step 615, Loss: {'train/loss': 0.31845566630363464, 'train/grad_norm_pre_clip': 14.441492080688477, 'timing/data': 0.013248745002783835, 'timing/model': 1.112094484968111, 'learning_rate/qwen_vl_interface': 9.743411991003246e-06, 'learning_rate/action_model': 9.743411991003247e-05, 'epoch': 1.26}) +[2026-06-19 11:10:38,094][starVLA.training.train_starvla][INFO] - Step 616, Loss: {'train/loss': 0.23555107414722443, 'train/grad_norm_pre_clip': 9.155951499938965, 'timing/data': 0.012866412987932563, 'timing/model': 1.1162063110386953, 'learning_rate/qwen_vl_interface': 9.742423678894823e-06, 'learning_rate/action_model': 9.742423678894823e-05, 'epoch': 1.27}) +[2026-06-19 11:10:42,595][starVLA.training.train_starvla][INFO] - Step 617, Loss: {'train/loss': 0.5229216814041138, 'train/grad_norm_pre_clip': 11.240367889404297, 'timing/data': 0.013309361995197833, 'timing/model': 1.120492568006739, 'learning_rate/qwen_vl_interface': 9.741433520120722e-06, 'learning_rate/action_model': 9.74143352012072e-05, 'epoch': 1.27}) +[2026-06-19 11:10:47,104][starVLA.training.train_starvla][INFO] - Step 618, Loss: {'train/loss': 0.2535281777381897, 'train/grad_norm_pre_clip': 8.785868644714355, 'timing/data': 0.013608912006020546, 'timing/model': 1.113931112922728, 'learning_rate/qwen_vl_interface': 9.740441515087953e-06, 'learning_rate/action_model': 9.740441515087953e-05, 'epoch': 1.27}) +[2026-06-19 11:10:51,505][starVLA.training.train_starvla][INFO] - Step 619, Loss: {'train/loss': 0.2952374517917633, 'train/grad_norm_pre_clip': 3.258469581604004, 'timing/data': 0.013195326086133718, 'timing/model': 1.114271656027995, 'learning_rate/qwen_vl_interface': 9.739447664204298e-06, 'learning_rate/action_model': 9.739447664204297e-05, 'epoch': 1.27}) +[2026-06-19 11:10:55,951][starVLA.training.train_starvla][INFO] - Step 620, Loss: {'train/loss': 0.15092910826206207, 'train/grad_norm_pre_clip': 4.271928787231445, 'timing/data': 0.012324517942033708, 'timing/model': 1.2522359790746123, 'learning_rate/qwen_vl_interface': 9.73845196787829e-06, 'learning_rate/action_model': 9.738451967878289e-05, 'epoch': 1.27}) +[2026-06-19 11:11:00,259][starVLA.training.train_starvla][INFO] - Step 621, Loss: {'train/loss': 0.15273669362068176, 'train/grad_norm_pre_clip': 6.931696891784668, 'timing/data': 0.013284901971928775, 'timing/model': 1.1145678650354967, 'learning_rate/qwen_vl_interface': 9.737454426519221e-06, 'learning_rate/action_model': 9.737454426519221e-05, 'epoch': 1.28}) +[2026-06-19 11:11:04,717][starVLA.training.train_starvla][INFO] - Step 622, Loss: {'train/loss': 0.2353583574295044, 'train/grad_norm_pre_clip': 5.984917640686035, 'timing/data': 0.013179295929148793, 'timing/model': 1.1215412199962884, 'learning_rate/qwen_vl_interface': 9.736455040537145e-06, 'learning_rate/action_model': 9.736455040537144e-05, 'epoch': 1.28}) +[2026-06-19 11:11:09,199][starVLA.training.train_starvla][INFO] - Step 623, Loss: {'train/loss': 0.26963502168655396, 'train/grad_norm_pre_clip': 5.49159574508667, 'timing/data': 0.013538716011680663, 'timing/model': 1.1269028600072488, 'learning_rate/qwen_vl_interface': 9.735453810342868e-06, 'learning_rate/action_model': 9.735453810342867e-05, 'epoch': 1.28}) +[2026-06-19 11:11:13,682][starVLA.training.train_starvla][INFO] - Step 624, Loss: {'train/loss': 0.17551320791244507, 'train/grad_norm_pre_clip': 7.770176887512207, 'timing/data': 0.012681066989898682, 'timing/model': 1.1182510120561346, 'learning_rate/qwen_vl_interface': 9.734450736347958e-06, 'learning_rate/action_model': 9.73445073634796e-05, 'epoch': 1.28}) +[2026-06-19 11:11:18,098][starVLA.training.train_starvla][INFO] - Step 625, Loss: {'train/loss': 0.10316278040409088, 'train/grad_norm_pre_clip': 2.923920154571533, 'timing/data': 0.012612314079888165, 'timing/model': 1.1162756260018796, 'learning_rate/qwen_vl_interface': 9.733445818964746e-06, 'learning_rate/action_model': 9.733445818964746e-05, 'epoch': 1.28}) +[2026-06-19 11:11:22,553][starVLA.training.train_starvla][INFO] - Step 626, Loss: {'train/loss': 0.1673235446214676, 'train/grad_norm_pre_clip': 5.128851890563965, 'timing/data': 0.013020679005421698, 'timing/model': 1.2588332240702584, 'learning_rate/qwen_vl_interface': 9.732439058606311e-06, 'learning_rate/action_model': 9.732439058606312e-05, 'epoch': 1.29}) +[2026-06-19 11:11:26,853][starVLA.training.train_starvla][INFO] - Step 627, Loss: {'train/loss': 0.33403679728507996, 'train/grad_norm_pre_clip': 8.048247337341309, 'timing/data': 0.013323288992978632, 'timing/model': 1.112542951013893, 'learning_rate/qwen_vl_interface': 9.731430455686499e-06, 'learning_rate/action_model': 9.731430455686498e-05, 'epoch': 1.29}) +[2026-06-19 11:11:31,300][starVLA.training.train_starvla][INFO] - Step 628, Loss: {'train/loss': 0.1315397173166275, 'train/grad_norm_pre_clip': 5.239992141723633, 'timing/data': 0.012981684994883835, 'timing/model': 1.1120003910036758, 'learning_rate/qwen_vl_interface': 9.730420010619901e-06, 'learning_rate/action_model': 9.7304200106199e-05, 'epoch': 1.29}) +[2026-06-19 11:11:35,771][starVLA.training.train_starvla][INFO] - Step 629, Loss: {'train/loss': 0.41707441210746765, 'train/grad_norm_pre_clip': 13.67144775390625, 'timing/data': 0.013144831056706607, 'timing/model': 1.1227381869684905, 'learning_rate/qwen_vl_interface': 9.72940772382188e-06, 'learning_rate/action_model': 9.729407723821878e-05, 'epoch': 1.29}) +[2026-06-19 11:11:40,244][starVLA.training.train_starvla][INFO] - Step 630, Loss: {'train/loss': 0.16079206764698029, 'train/grad_norm_pre_clip': 4.9255170822143555, 'timing/data': 0.01304816803894937, 'timing/model': 1.1161709720036015, 'learning_rate/qwen_vl_interface': 9.728393595708542e-06, 'learning_rate/action_model': 9.728393595708542e-05, 'epoch': 1.29}) +[2026-06-19 11:11:44,720][starVLA.training.train_starvla][INFO] - Step 631, Loss: {'train/loss': 0.24160318076610565, 'train/grad_norm_pre_clip': 6.004219055175781, 'timing/data': 0.013542011030949652, 'timing/model': 1.2843173019355163, 'learning_rate/qwen_vl_interface': 9.727377626696762e-06, 'learning_rate/action_model': 9.727377626696762e-05, 'epoch': 1.3}) +[2026-06-19 11:11:49,025][starVLA.training.train_starvla][INFO] - Step 632, Loss: {'train/loss': 0.10750874131917953, 'train/grad_norm_pre_clip': 7.906558036804199, 'timing/data': 0.013315135962329805, 'timing/model': 1.1161299519008026, 'learning_rate/qwen_vl_interface': 9.726359817204165e-06, 'learning_rate/action_model': 9.726359817204166e-05, 'epoch': 1.3}) +[2026-06-19 11:11:53,516][starVLA.training.train_starvla][INFO] - Step 633, Loss: {'train/loss': 0.15822052955627441, 'train/grad_norm_pre_clip': 6.117109775543213, 'timing/data': 0.013929064967669547, 'timing/model': 1.1275097579928115, 'learning_rate/qwen_vl_interface': 9.725340167649136e-06, 'learning_rate/action_model': 9.725340167649136e-05, 'epoch': 1.3}) +[2026-06-19 11:11:57,914][starVLA.training.train_starvla][INFO] - Step 634, Loss: {'train/loss': 0.13900113105773926, 'train/grad_norm_pre_clip': 3.5269806385040283, 'timing/data': 0.013622393016703427, 'timing/model': 1.136790252989158, 'learning_rate/qwen_vl_interface': 9.72431867845081e-06, 'learning_rate/action_model': 9.72431867845081e-05, 'epoch': 1.3}) +[2026-06-19 11:12:02,350][starVLA.training.train_starvla][INFO] - Step 635, Loss: {'train/loss': 0.49134254455566406, 'train/grad_norm_pre_clip': 14.583372116088867, 'timing/data': 0.01564350409898907, 'timing/model': 1.1218654620461166, 'learning_rate/qwen_vl_interface': 9.723295350029086e-06, 'learning_rate/action_model': 9.723295350029086e-05, 'epoch': 1.3}) +[2026-06-19 11:12:06,785][starVLA.training.train_starvla][INFO] - Step 636, Loss: {'train/loss': 0.2829490005970001, 'train/grad_norm_pre_clip': 6.348171234130859, 'timing/data': 0.012874289066530764, 'timing/model': 1.1140785670140758, 'learning_rate/qwen_vl_interface': 9.722270182804616e-06, 'learning_rate/action_model': 9.722270182804614e-05, 'epoch': 1.31}) +[2026-06-19 11:12:11,249][starVLA.training.train_starvla][INFO] - Step 637, Loss: {'train/loss': 0.2462468445301056, 'train/grad_norm_pre_clip': 6.484733581542969, 'timing/data': 0.011943722027353942, 'timing/model': 1.2818420130060986, 'learning_rate/qwen_vl_interface': 9.721243177198805e-06, 'learning_rate/action_model': 9.721243177198804e-05, 'epoch': 1.31}) +[2026-06-19 11:12:15,583][starVLA.training.train_starvla][INFO] - Step 638, Loss: {'train/loss': 0.15082882344722748, 'train/grad_norm_pre_clip': 4.297821044921875, 'timing/data': 0.013132867985405028, 'timing/model': 1.1197796580381691, 'learning_rate/qwen_vl_interface': 9.720214333633817e-06, 'learning_rate/action_model': 9.720214333633817e-05, 'epoch': 1.31}) +[2026-06-19 11:12:20,065][starVLA.training.train_starvla][INFO] - Step 639, Loss: {'train/loss': 0.12062250077724457, 'train/grad_norm_pre_clip': 3.616428852081299, 'timing/data': 0.013080612989142537, 'timing/model': 1.116261724033393, 'learning_rate/qwen_vl_interface': 9.719183652532572e-06, 'learning_rate/action_model': 9.719183652532572e-05, 'epoch': 1.31}) +[2026-06-19 11:12:24,551][starVLA.training.train_starvla][INFO] - Step 640, Loss: {'train/loss': 0.17479197680950165, 'train/grad_norm_pre_clip': 6.003766059875488, 'timing/data': 0.013049107976257801, 'timing/model': 1.1258196129929274, 'learning_rate/qwen_vl_interface': 9.718151134318743e-06, 'learning_rate/action_model': 9.718151134318743e-05, 'epoch': 1.31}) +[2026-06-19 11:12:28,988][starVLA.training.train_starvla][INFO] - Step 641, Loss: {'train/loss': 0.13339045643806458, 'train/grad_norm_pre_clip': 4.871769905090332, 'timing/data': 0.012790734064765275, 'timing/model': 1.1160576139809564, 'learning_rate/qwen_vl_interface': 9.71711677941676e-06, 'learning_rate/action_model': 9.717116779416758e-05, 'epoch': 1.32}) +[2026-06-19 11:12:33,451][starVLA.training.train_starvla][INFO] - Step 642, Loss: {'train/loss': 0.13993948698043823, 'train/grad_norm_pre_clip': 3.8804843425750732, 'timing/data': 0.013243495021015406, 'timing/model': 1.1167249529389665, 'learning_rate/qwen_vl_interface': 9.716080588251805e-06, 'learning_rate/action_model': 9.716080588251805e-05, 'epoch': 1.32}) +[2026-06-19 11:12:37,970][starVLA.training.train_starvla][INFO] - Step 643, Loss: {'train/loss': 0.22109712660312653, 'train/grad_norm_pre_clip': 8.881341934204102, 'timing/data': 0.013597851037047803, 'timing/model': 1.3047274550190195, 'learning_rate/qwen_vl_interface': 9.71504256124982e-06, 'learning_rate/action_model': 9.71504256124982e-05, 'epoch': 1.32}) +[2026-06-19 11:12:42,280][starVLA.training.train_starvla][INFO] - Step 644, Loss: {'train/loss': 0.1218029037117958, 'train/grad_norm_pre_clip': 6.104808807373047, 'timing/data': 0.012481498997658491, 'timing/model': 1.1208163920091465, 'learning_rate/qwen_vl_interface': 9.714002698837499e-06, 'learning_rate/action_model': 9.714002698837498e-05, 'epoch': 1.32}) +[2026-06-19 11:12:46,749][starVLA.training.train_starvla][INFO] - Step 645, Loss: {'train/loss': 0.13477808237075806, 'train/grad_norm_pre_clip': 5.282682418823242, 'timing/data': 0.012343754991889, 'timing/model': 1.1175040639936924, 'learning_rate/qwen_vl_interface': 9.712961001442285e-06, 'learning_rate/action_model': 9.712961001442286e-05, 'epoch': 1.32}) +[2026-06-19 11:12:51,233][starVLA.training.train_starvla][INFO] - Step 646, Loss: {'train/loss': 0.18559212982654572, 'train/grad_norm_pre_clip': 6.684216499328613, 'timing/data': 0.012682699947617948, 'timing/model': 1.1183565530227497, 'learning_rate/qwen_vl_interface': 9.711917469492388e-06, 'learning_rate/action_model': 9.711917469492388e-05, 'epoch': 1.33}) +[2026-06-19 11:12:55,705][starVLA.training.train_starvla][INFO] - Step 647, Loss: {'train/loss': 0.10302236676216125, 'train/grad_norm_pre_clip': 4.566836833953857, 'timing/data': 0.01302274304907769, 'timing/model': 1.1178148579783738, 'learning_rate/qwen_vl_interface': 9.71087210341676e-06, 'learning_rate/action_model': 9.71087210341676e-05, 'epoch': 1.33}) +[2026-06-19 11:13:00,087][starVLA.training.train_starvla][INFO] - Step 648, Loss: {'train/loss': 0.09810926020145416, 'train/grad_norm_pre_clip': 4.8531107902526855, 'timing/data': 0.01369827101007104, 'timing/model': 1.122181663988158, 'learning_rate/qwen_vl_interface': 9.70982490364511e-06, 'learning_rate/action_model': 9.70982490364511e-05, 'epoch': 1.33}) +[2026-06-19 11:13:04,566][starVLA.training.train_starvla][INFO] - Step 649, Loss: {'train/loss': 0.1849023848772049, 'train/grad_norm_pre_clip': 10.813302040100098, 'timing/data': 0.012283638003282249, 'timing/model': 1.287961496040225, 'learning_rate/qwen_vl_interface': 9.708775870607905e-06, 'learning_rate/action_model': 9.708775870607904e-05, 'epoch': 1.33}) +[2026-06-19 11:13:08,906][starVLA.training.train_starvla][INFO] - Step 650, Loss: {'train/loss': 0.37378665804862976, 'train/grad_norm_pre_clip': 14.611305236816406, 'timing/data': 0.012834868975915015, 'timing/model': 1.1370104140369222, 'learning_rate/qwen_vl_interface': 9.707725004736363e-06, 'learning_rate/action_model': 9.707725004736363e-05, 'epoch': 1.34}) +[2026-06-19 11:13:13,402][starVLA.training.train_starvla][INFO] - Step 651, Loss: {'train/loss': 0.3176535665988922, 'train/grad_norm_pre_clip': 9.912697792053223, 'timing/data': 0.013037274009548128, 'timing/model': 1.1199135329807177, 'learning_rate/qwen_vl_interface': 9.706672306462454e-06, 'learning_rate/action_model': 9.706672306462454e-05, 'epoch': 1.34}) +[2026-06-19 11:13:17,916][starVLA.training.train_starvla][INFO] - Step 652, Loss: {'train/loss': 0.17642299830913544, 'train/grad_norm_pre_clip': 12.241084098815918, 'timing/data': 0.01301150000654161, 'timing/model': 1.1281106730457395, 'learning_rate/qwen_vl_interface': 9.705617776218905e-06, 'learning_rate/action_model': 9.705617776218904e-05, 'epoch': 1.34}) +[2026-06-19 11:13:22,353][starVLA.training.train_starvla][INFO] - Step 653, Loss: {'train/loss': 0.11216425895690918, 'train/grad_norm_pre_clip': 8.685484886169434, 'timing/data': 0.013141494011506438, 'timing/model': 1.1344586199847981, 'learning_rate/qwen_vl_interface': 9.70456141443919e-06, 'learning_rate/action_model': 9.704561414439188e-05, 'epoch': 1.34}) +[2026-06-19 11:13:26,882][starVLA.training.train_starvla][INFO] - Step 654, Loss: {'train/loss': 0.1406523436307907, 'train/grad_norm_pre_clip': 13.17616081237793, 'timing/data': 0.013527826056815684, 'timing/model': 1.152154966024682, 'learning_rate/qwen_vl_interface': 9.703503221557538e-06, 'learning_rate/action_model': 9.703503221557538e-05, 'epoch': 1.34}) +[2026-06-19 11:13:31,405][starVLA.training.train_starvla][INFO] - Step 655, Loss: {'train/loss': 0.3631158471107483, 'train/grad_norm_pre_clip': 15.945419311523438, 'timing/data': 0.012990796938538551, 'timing/model': 1.2995703029446304, 'learning_rate/qwen_vl_interface': 9.70244319800894e-06, 'learning_rate/action_model': 9.70244319800894e-05, 'epoch': 1.35}) +[2026-06-19 11:13:35,736][starVLA.training.train_starvla][INFO] - Step 656, Loss: {'train/loss': 0.16999319195747375, 'train/grad_norm_pre_clip': 10.211934089660645, 'timing/data': 0.013783387024886906, 'timing/model': 1.1204758590320125, 'learning_rate/qwen_vl_interface': 9.701381344229124e-06, 'learning_rate/action_model': 9.701381344229123e-05, 'epoch': 1.35}) +[2026-06-19 11:13:40,264][starVLA.training.train_starvla][INFO] - Step 657, Loss: {'train/loss': 0.3617745637893677, 'train/grad_norm_pre_clip': 16.84663200378418, 'timing/data': 0.013568183989264071, 'timing/model': 1.1250719289528206, 'learning_rate/qwen_vl_interface': 9.700317660654581e-06, 'learning_rate/action_model': 9.700317660654581e-05, 'epoch': 1.35}) +[2026-06-19 11:13:44,801][starVLA.training.train_starvla][INFO] - Step 658, Loss: {'train/loss': 0.8407184481620789, 'train/grad_norm_pre_clip': 26.40610122680664, 'timing/data': 0.012855520006269217, 'timing/model': 1.1193806080846116, 'learning_rate/qwen_vl_interface': 9.699252147722554e-06, 'learning_rate/action_model': 9.699252147722553e-05, 'epoch': 1.35}) +[2026-06-19 11:13:49,286][starVLA.training.train_starvla][INFO] - Step 659, Loss: {'train/loss': 0.5444105267524719, 'train/grad_norm_pre_clip': 17.185943603515625, 'timing/data': 0.013230746961198747, 'timing/model': 1.125907592009753, 'learning_rate/qwen_vl_interface': 9.698184805871028e-06, 'learning_rate/action_model': 9.698184805871028e-05, 'epoch': 1.35}) +[2026-06-19 11:13:53,770][starVLA.training.train_starvla][INFO] - Step 660, Loss: {'train/loss': 0.23253963887691498, 'train/grad_norm_pre_clip': 6.542031288146973, 'timing/data': 0.012782939011231065, 'timing/model': 1.1172480459790677, 'learning_rate/qwen_vl_interface': 9.697115635538753e-06, 'learning_rate/action_model': 9.697115635538753e-05, 'epoch': 1.36}) +[2026-06-19 11:13:58,281][starVLA.training.train_starvla][INFO] - Step 661, Loss: {'train/loss': 0.19893646240234375, 'train/grad_norm_pre_clip': 5.328545093536377, 'timing/data': 0.013437525019980967, 'timing/model': 1.2991125049302354, 'learning_rate/qwen_vl_interface': 9.69604463716522e-06, 'learning_rate/action_model': 9.69604463716522e-05, 'epoch': 1.36}) +[2026-06-19 11:14:02,593][starVLA.training.train_starvla][INFO] - Step 662, Loss: {'train/loss': 0.5010871887207031, 'train/grad_norm_pre_clip': 19.539827346801758, 'timing/data': 0.013515925966203213, 'timing/model': 1.1197034730575979, 'learning_rate/qwen_vl_interface': 9.69497181119068e-06, 'learning_rate/action_model': 9.694971811190681e-05, 'epoch': 1.36}) +[2026-06-19 11:14:07,065][starVLA.training.train_starvla][INFO] - Step 663, Loss: {'train/loss': 0.2442089468240738, 'train/grad_norm_pre_clip': 43.55930709838867, 'timing/data': 0.013282640022225678, 'timing/model': 1.1266637089429423, 'learning_rate/qwen_vl_interface': 9.693897158056132e-06, 'learning_rate/action_model': 9.693897158056132e-05, 'epoch': 1.36}) +[2026-06-19 11:14:11,512][starVLA.training.train_starvla][INFO] - Step 664, Loss: {'train/loss': 0.2770957350730896, 'train/grad_norm_pre_clip': 5.7176618576049805, 'timing/data': 0.01286658295430243, 'timing/model': 1.1200784150278196, 'learning_rate/qwen_vl_interface': 9.692820678203322e-06, 'learning_rate/action_model': 9.692820678203322e-05, 'epoch': 1.36}) +[2026-06-19 11:14:15,966][starVLA.training.train_starvla][INFO] - Step 665, Loss: {'train/loss': 0.33178186416625977, 'train/grad_norm_pre_clip': 15.19389820098877, 'timing/data': 0.013570565963163972, 'timing/model': 1.121415312984027, 'learning_rate/qwen_vl_interface': 9.691742372074749e-06, 'learning_rate/action_model': 9.691742372074749e-05, 'epoch': 1.37}) +[2026-06-19 11:14:20,485][starVLA.training.train_starvla][INFO] - Step 666, Loss: {'train/loss': 0.5128674507141113, 'train/grad_norm_pre_clip': 18.089942932128906, 'timing/data': 0.013451591017656028, 'timing/model': 1.1212692819535732, 'learning_rate/qwen_vl_interface': 9.690662240113669e-06, 'learning_rate/action_model': 9.690662240113668e-05, 'epoch': 1.37}) +[2026-06-19 11:14:24,993][starVLA.training.train_starvla][INFO] - Step 667, Loss: {'train/loss': 0.1665341854095459, 'train/grad_norm_pre_clip': 3.464406728744507, 'timing/data': 0.012832732987590134, 'timing/model': 1.291275159921497, 'learning_rate/qwen_vl_interface': 9.689580282764077e-06, 'learning_rate/action_model': 9.689580282764077e-05, 'epoch': 1.37}) +[2026-06-19 11:14:29,306][starVLA.training.train_starvla][INFO] - Step 668, Loss: {'train/loss': 0.26416251063346863, 'train/grad_norm_pre_clip': 4.470946311950684, 'timing/data': 0.012618153006769717, 'timing/model': 1.11667940299958, 'learning_rate/qwen_vl_interface': 9.688496500470731e-06, 'learning_rate/action_model': 9.68849650047073e-05, 'epoch': 1.37}) +[2026-06-19 11:14:33,782][starVLA.training.train_starvla][INFO] - Step 669, Loss: {'train/loss': 0.23133616149425507, 'train/grad_norm_pre_clip': 6.364943027496338, 'timing/data': 0.012480622041039169, 'timing/model': 1.1161112110130489, 'learning_rate/qwen_vl_interface': 9.68741089367913e-06, 'learning_rate/action_model': 9.687410893679129e-05, 'epoch': 1.37}) +[2026-06-19 11:14:38,235][starVLA.training.train_starvla][INFO] - Step 670, Loss: {'train/loss': 0.23821242153644562, 'train/grad_norm_pre_clip': 5.657525062561035, 'timing/data': 0.012945450958795846, 'timing/model': 1.1347033299971372, 'learning_rate/qwen_vl_interface': 9.686323462835527e-06, 'learning_rate/action_model': 9.686323462835526e-05, 'epoch': 1.38}) +[2026-06-19 11:14:42,710][starVLA.training.train_starvla][INFO] - Step 671, Loss: {'train/loss': 0.31962883472442627, 'train/grad_norm_pre_clip': 5.427786350250244, 'timing/data': 0.012865166063420475, 'timing/model': 1.1161275980994105, 'learning_rate/qwen_vl_interface': 9.685234208386922e-06, 'learning_rate/action_model': 9.685234208386921e-05, 'epoch': 1.38}) +[2026-06-19 11:14:47,207][starVLA.training.train_starvla][INFO] - Step 672, Loss: {'train/loss': 0.17684583365917206, 'train/grad_norm_pre_clip': 3.988057851791382, 'timing/data': 0.016038165078498423, 'timing/model': 1.126382165006362, 'learning_rate/qwen_vl_interface': 9.684143130781066e-06, 'learning_rate/action_model': 9.684143130781066e-05, 'epoch': 1.38}) +[2026-06-19 11:14:51,704][starVLA.training.train_starvla][INFO] - Step 673, Loss: {'train/loss': 0.18463413417339325, 'train/grad_norm_pre_clip': 2.9680614471435547, 'timing/data': 0.013050965964794159, 'timing/model': 1.2928498549154028, 'learning_rate/qwen_vl_interface': 9.683050230466462e-06, 'learning_rate/action_model': 9.683050230466462e-05, 'epoch': 1.38}) +[2026-06-19 11:14:56,129][starVLA.training.train_starvla][INFO] - Step 674, Loss: {'train/loss': 0.1069590225815773, 'train/grad_norm_pre_clip': 3.532010078430176, 'timing/data': 0.012391373980790377, 'timing/model': 1.22792231105268, 'learning_rate/qwen_vl_interface': 9.681955507892361e-06, 'learning_rate/action_model': 9.68195550789236e-05, 'epoch': 1.38}) +[2026-06-19 11:15:00,487][starVLA.training.train_starvla][INFO] - Step 675, Loss: {'train/loss': 0.19254206120967865, 'train/grad_norm_pre_clip': 5.57824182510376, 'timing/data': 0.012707683024927974, 'timing/model': 1.1373405870981514, 'learning_rate/qwen_vl_interface': 9.680858963508759e-06, 'learning_rate/action_model': 9.68085896350876e-05, 'epoch': 1.39}) +[2026-06-19 11:15:04,925][starVLA.training.train_starvla][INFO] - Step 676, Loss: {'train/loss': 0.2042199671268463, 'train/grad_norm_pre_clip': 5.211271286010742, 'timing/data': 0.01242099900264293, 'timing/model': 1.1166710529942065, 'learning_rate/qwen_vl_interface': 9.679760597766405e-06, 'learning_rate/action_model': 9.679760597766405e-05, 'epoch': 1.39}) +[2026-06-19 11:15:09,437][starVLA.training.train_starvla][INFO] - Step 677, Loss: {'train/loss': 0.13101904094219208, 'train/grad_norm_pre_clip': 6.093982696533203, 'timing/data': 0.012309582903981209, 'timing/model': 1.1296909470111132, 'learning_rate/qwen_vl_interface': 9.678660411116796e-06, 'learning_rate/action_model': 9.678660411116796e-05, 'epoch': 1.39}) +[2026-06-19 11:15:13,904][starVLA.training.train_starvla][INFO] - Step 678, Loss: {'train/loss': 0.25945016741752625, 'train/grad_norm_pre_clip': 6.795522689819336, 'timing/data': 0.014162114006467164, 'timing/model': 1.115638016955927, 'learning_rate/qwen_vl_interface': 9.677558404012178e-06, 'learning_rate/action_model': 9.677558404012177e-05, 'epoch': 1.39}) +[2026-06-19 11:15:18,366][starVLA.training.train_starvla][INFO] - Step 679, Loss: {'train/loss': 0.17357297241687775, 'train/grad_norm_pre_clip': 8.760249137878418, 'timing/data': 0.013515943894162774, 'timing/model': 1.2737565060378984, 'learning_rate/qwen_vl_interface': 9.676454576905542e-06, 'learning_rate/action_model': 9.676454576905542e-05, 'epoch': 1.39}) +[2026-06-19 11:15:22,786][starVLA.training.train_starvla][INFO] - Step 680, Loss: {'train/loss': 0.14168958365917206, 'train/grad_norm_pre_clip': 10.035694122314453, 'timing/data': 0.013588597998023033, 'timing/model': 1.2159245479851961, 'learning_rate/qwen_vl_interface': 9.675348930250633e-06, 'learning_rate/action_model': 9.675348930250632e-05, 'epoch': 1.4}) +[2026-06-19 11:15:27,129][starVLA.training.train_starvla][INFO] - Step 681, Loss: {'train/loss': 0.25649285316467285, 'train/grad_norm_pre_clip': 8.135869979858398, 'timing/data': 0.01303656690288335, 'timing/model': 1.1409505209885538, 'learning_rate/qwen_vl_interface': 9.674241464501938e-06, 'learning_rate/action_model': 9.674241464501937e-05, 'epoch': 1.4}) +[2026-06-19 11:15:31,577][starVLA.training.train_starvla][INFO] - Step 682, Loss: {'train/loss': 0.462051659822464, 'train/grad_norm_pre_clip': 11.143253326416016, 'timing/data': 0.013049338012933731, 'timing/model': 1.117688618018292, 'learning_rate/qwen_vl_interface': 9.673132180114695e-06, 'learning_rate/action_model': 9.673132180114695e-05, 'epoch': 1.4}) +[2026-06-19 11:15:36,064][starVLA.training.train_starvla][INFO] - Step 683, Loss: {'train/loss': 0.19345542788505554, 'train/grad_norm_pre_clip': 6.879632949829102, 'timing/data': 0.012355783954262733, 'timing/model': 1.1185845849104226, 'learning_rate/qwen_vl_interface': 9.67202107754489e-06, 'learning_rate/action_model': 9.672021077544889e-05, 'epoch': 1.4}) +[2026-06-19 11:15:40,588][starVLA.training.train_starvla][INFO] - Step 684, Loss: {'train/loss': 0.10862302035093307, 'train/grad_norm_pre_clip': 4.642779350280762, 'timing/data': 0.013692923006601632, 'timing/model': 1.1317186629166827, 'learning_rate/qwen_vl_interface': 9.670908157249254e-06, 'learning_rate/action_model': 9.670908157249253e-05, 'epoch': 1.4}) +[2026-06-19 11:15:45,091][starVLA.training.train_starvla][INFO] - Step 685, Loss: {'train/loss': 0.31028077006340027, 'train/grad_norm_pre_clip': 12.464494705200195, 'timing/data': 0.013377920957282186, 'timing/model': 1.3044187560444698, 'learning_rate/qwen_vl_interface': 9.669793419685267e-06, 'learning_rate/action_model': 9.669793419685266e-05, 'epoch': 1.41}) +[2026-06-19 11:15:49,521][starVLA.training.train_starvla][INFO] - Step 686, Loss: {'train/loss': 0.28008654713630676, 'train/grad_norm_pre_clip': 9.422486305236816, 'timing/data': 0.01295707793906331, 'timing/model': 1.2381889219395816, 'learning_rate/qwen_vl_interface': 9.668676865311155e-06, 'learning_rate/action_model': 9.668676865311154e-05, 'epoch': 1.41}) +[2026-06-19 11:15:53,846][starVLA.training.train_starvla][INFO] - Step 687, Loss: {'train/loss': 0.15007857978343964, 'train/grad_norm_pre_clip': 8.399105072021484, 'timing/data': 0.012929523014463484, 'timing/model': 1.1178773479769006, 'learning_rate/qwen_vl_interface': 9.667558494585891e-06, 'learning_rate/action_model': 9.66755849458589e-05, 'epoch': 1.41}) +[2026-06-19 11:15:58,376][starVLA.training.train_starvla][INFO] - Step 688, Loss: {'train/loss': 0.32262611389160156, 'train/grad_norm_pre_clip': 8.894572257995605, 'timing/data': 0.013123078970238566, 'timing/model': 1.1160043010022491, 'learning_rate/qwen_vl_interface': 9.666438307969194e-06, 'learning_rate/action_model': 9.666438307969195e-05, 'epoch': 1.41}) +[2026-06-19 11:16:02,862][starVLA.training.train_starvla][INFO] - Step 689, Loss: {'train/loss': 0.04209017753601074, 'train/grad_norm_pre_clip': 3.0426900386810303, 'timing/data': 0.014111066004261374, 'timing/model': 1.1226034150458872, 'learning_rate/qwen_vl_interface': 9.665316305921535e-06, 'learning_rate/action_model': 9.665316305921535e-05, 'epoch': 1.42}) +[2026-06-19 11:16:07,348][starVLA.training.train_starvla][INFO] - Step 690, Loss: {'train/loss': 0.11102665215730667, 'train/grad_norm_pre_clip': 4.8087592124938965, 'timing/data': 0.01347487501334399, 'timing/model': 1.1183780529536307, 'learning_rate/qwen_vl_interface': 9.664192488904123e-06, 'learning_rate/action_model': 9.664192488904122e-05, 'epoch': 1.42}) +[2026-06-19 11:16:11,833][starVLA.training.train_starvla][INFO] - Step 691, Loss: {'train/loss': 0.19161705672740936, 'train/grad_norm_pre_clip': 7.76854944229126, 'timing/data': 0.01281703996937722, 'timing/model': 1.2827593240654096, 'learning_rate/qwen_vl_interface': 9.663066857378917e-06, 'learning_rate/action_model': 9.663066857378917e-05, 'epoch': 1.42}) +[2026-06-19 11:16:16,151][starVLA.training.train_starvla][INFO] - Step 692, Loss: {'train/loss': 0.27636095881462097, 'train/grad_norm_pre_clip': 11.563769340515137, 'timing/data': 0.013292571995407343, 'timing/model': 1.1280155759304762, 'learning_rate/qwen_vl_interface': 9.661939411808623e-06, 'learning_rate/action_model': 9.661939411808622e-05, 'epoch': 1.42}) +[2026-06-19 11:16:20,639][starVLA.training.train_starvla][INFO] - Step 693, Loss: {'train/loss': 0.12144620716571808, 'train/grad_norm_pre_clip': 5.755306243896484, 'timing/data': 0.012716713012196124, 'timing/model': 1.1291123840492219, 'learning_rate/qwen_vl_interface': 9.66081015265669e-06, 'learning_rate/action_model': 9.660810152656689e-05, 'epoch': 1.42}) +[2026-06-19 11:16:25,153][starVLA.training.train_starvla][INFO] - Step 694, Loss: {'train/loss': 0.5178697109222412, 'train/grad_norm_pre_clip': 21.188343048095703, 'timing/data': 0.012394517078064382, 'timing/model': 1.1251530110603198, 'learning_rate/qwen_vl_interface': 9.659679080387314e-06, 'learning_rate/action_model': 9.659679080387313e-05, 'epoch': 1.43}) +[2026-06-19 11:16:29,649][starVLA.training.train_starvla][INFO] - Step 695, Loss: {'train/loss': 0.13139602541923523, 'train/grad_norm_pre_clip': 5.91988468170166, 'timing/data': 0.013592868926934898, 'timing/model': 1.1275208779843524, 'learning_rate/qwen_vl_interface': 9.658546195465436e-06, 'learning_rate/action_model': 9.658546195465435e-05, 'epoch': 1.43}) +[2026-06-19 11:16:34,064][starVLA.training.train_starvla][INFO] - Step 696, Loss: {'train/loss': 0.27966174483299255, 'train/grad_norm_pre_clip': 15.375279426574707, 'timing/data': 0.013170882943086326, 'timing/model': 1.1257382379844785, 'learning_rate/qwen_vl_interface': 9.65741149835674e-06, 'learning_rate/action_model': 9.657411498356741e-05, 'epoch': 1.43}) +[2026-06-19 11:16:38,495][starVLA.training.train_starvla][INFO] - Step 697, Loss: {'train/loss': 0.2516348361968994, 'train/grad_norm_pre_clip': 11.771978378295898, 'timing/data': 0.012996680103242397, 'timing/model': 1.240076250047423, 'learning_rate/qwen_vl_interface': 9.656274989527663e-06, 'learning_rate/action_model': 9.656274989527664e-05, 'epoch': 1.43}) +[2026-06-19 11:16:42,814][starVLA.training.train_starvla][INFO] - Step 698, Loss: {'train/loss': 0.2895459830760956, 'train/grad_norm_pre_clip': 10.313965797424316, 'timing/data': 0.01285885600373149, 'timing/model': 1.12162035505753, 'learning_rate/qwen_vl_interface': 9.655136669445378e-06, 'learning_rate/action_model': 9.655136669445378e-05, 'epoch': 1.43}) +[2026-06-19 11:16:47,324][starVLA.training.train_starvla][INFO] - Step 699, Loss: {'train/loss': 0.190336674451828, 'train/grad_norm_pre_clip': 8.440329551696777, 'timing/data': 0.013063043938018382, 'timing/model': 1.11862162000034, 'learning_rate/qwen_vl_interface': 9.653996538577802e-06, 'learning_rate/action_model': 9.653996538577802e-05, 'epoch': 1.44}) +[2026-06-19 11:16:51,824][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 700 +[2026-06-19 11:19:50,687][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 700 +[2026-06-19 11:19:50,687][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 700 +[2026-06-19 11:22:58,437][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 700 +[2026-06-19 11:22:58,438][starVLA.training.train_starvla][INFO] - Step 700, Loss: {'train/loss': 0.3342103660106659, 'train/grad_norm_pre_clip': 9.688658714294434, 'eval/loss': 0.3749692142009735, 'eval/latency_0/loss': 0.10484850406646729, 'eval/latency_1/loss': 0.2629852890968323, 'eval/latency_2/loss': 0.4156937599182129, 'eval/latency_3/loss': 0.4324919879436493, 'eval/latency_4/loss': 0.6445139050483704, 'eval/flappy/loss': 0.3749692142009735, 'eval/flappy/latency_0/loss': 0.10484850406646729, 'eval/flappy/latency_1/loss': 0.2629852890968323, 'eval/flappy/latency_2/loss': 0.4156937599182129, 'eval/flappy/latency_3/loss': 0.4324919879436493, 'eval/flappy/latency_4/loss': 0.6445139050483704, 'eval/flap_precision': 0.7105263157894737, 'eval/flap_recall': 0.6872727272727273, 'eval/flap_f1': 0.6987060998151571, 'eval/flap_rate_model': 0.07388888888888889, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.6987060998151571, 'eval/latency_0/flap_precision': 0.7105263157894737, 'eval/latency_0/flap_recall': 0.6872727272727273, 'eval/latency_0/flap_f1': 0.6987060998151571, 'eval/latency_0/flap_rate_model': 0.07388888888888889, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.6987060998151571, 'timing/data': 0.012164591928012669, 'timing/model': 1.1290640550432727, 'learning_rate/qwen_vl_interface': 9.652854597393604e-06, 'learning_rate/action_model': 9.652854597393603e-05, 'epoch': 1.44}) +[2026-06-19 11:22:58,438][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_700_state +[2026-06-19 11:22:58,441][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-19 11:23:01,822][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_700_state/model.safetensors +[2026-06-19 11:23:09,017][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_700_state/optimizer.bin +[2026-06-19 11:23:09,018][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_700_state/sampler.bin +[2026-06-19 11:23:09,018][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_700_state/sampler_1.bin +[2026-06-19 11:23:09,019][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_700_state/random_states_0.pkl +[2026-06-19 11:23:09,046][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 700 +[2026-06-19 11:23:09,046][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-19 11:23:09,055][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-19 11:23:13,459][starVLA.training.train_starvla][INFO] - Step 701, Loss: {'train/loss': 0.2553669214248657, 'train/grad_norm_pre_clip': 5.354673385620117, 'timing/data': 0.0144994129659608, 'timing/model': 1.0975980389630422, 'learning_rate/qwen_vl_interface': 9.651710846362192e-06, 'learning_rate/action_model': 9.651710846362192e-05, 'epoch': 1.44}) +[2026-06-19 11:23:17,862][starVLA.training.train_starvla][INFO] - Step 702, Loss: {'train/loss': 0.19876840710639954, 'train/grad_norm_pre_clip': 4.132452011108398, 'timing/data': 0.013422591960988939, 'timing/model': 1.099508999963291, 'learning_rate/qwen_vl_interface': 9.650565285953716e-06, 'learning_rate/action_model': 9.650565285953715e-05, 'epoch': 1.44}) +[2026-06-19 11:23:22,299][starVLA.training.train_starvla][INFO] - Step 703, Loss: {'train/loss': 0.24236014485359192, 'train/grad_norm_pre_clip': 4.977395057678223, 'timing/data': 0.013869793037883937, 'timing/model': 1.2904593499843031, 'learning_rate/qwen_vl_interface': 9.649417916639077e-06, 'learning_rate/action_model': 9.649417916639076e-05, 'epoch': 1.44}) +[2026-06-19 11:23:26,559][starVLA.training.train_starvla][INFO] - Step 704, Loss: {'train/loss': 0.20545639097690582, 'train/grad_norm_pre_clip': 18.326129913330078, 'timing/data': 0.012435237062163651, 'timing/model': 1.103976274956949, 'learning_rate/qwen_vl_interface': 9.648268738889911e-06, 'learning_rate/action_model': 9.648268738889911e-05, 'epoch': 1.45}) +[2026-06-19 11:23:30,991][starVLA.training.train_starvla][INFO] - Step 705, Loss: {'train/loss': 0.15422900021076202, 'train/grad_norm_pre_clip': 3.0597774982452393, 'timing/data': 0.013286639004945755, 'timing/model': 1.105071926023811, 'learning_rate/qwen_vl_interface': 9.647117753178604e-06, 'learning_rate/action_model': 9.647117753178604e-05, 'epoch': 1.45}) +[2026-06-19 11:23:35,419][starVLA.training.train_starvla][INFO] - Step 706, Loss: {'train/loss': 0.25291740894317627, 'train/grad_norm_pre_clip': 6.221273899078369, 'timing/data': 0.012985200970433652, 'timing/model': 1.1070775730768219, 'learning_rate/qwen_vl_interface': 9.645964959978284e-06, 'learning_rate/action_model': 9.645964959978284e-05, 'epoch': 1.45}) +[2026-06-19 11:23:39,814][starVLA.training.train_starvla][INFO] - Step 707, Loss: {'train/loss': 0.3916848599910736, 'train/grad_norm_pre_clip': 8.453368186950684, 'timing/data': 0.012189560919068754, 'timing/model': 1.1084194260183722, 'learning_rate/qwen_vl_interface': 9.644810359762818e-06, 'learning_rate/action_model': 9.644810359762817e-05, 'epoch': 1.45}) +[2026-06-19 11:23:44,215][starVLA.training.train_starvla][INFO] - Step 708, Loss: {'train/loss': 0.33029258251190186, 'train/grad_norm_pre_clip': 5.797595500946045, 'timing/data': 0.013278929982334375, 'timing/model': 1.1052106019342318, 'learning_rate/qwen_vl_interface': 9.64365395300682e-06, 'learning_rate/action_model': 9.643653953006819e-05, 'epoch': 1.45}) +[2026-06-19 11:23:48,660][starVLA.training.train_starvla][INFO] - Step 709, Loss: {'train/loss': 0.18404065072536469, 'train/grad_norm_pre_clip': 3.5506269931793213, 'timing/data': 0.01326705701649189, 'timing/model': 1.2819383569294587, 'learning_rate/qwen_vl_interface': 9.642495740185644e-06, 'learning_rate/action_model': 9.642495740185643e-05, 'epoch': 1.46}) +[2026-06-19 11:23:52,940][starVLA.training.train_starvla][INFO] - Step 710, Loss: {'train/loss': 0.2709748446941376, 'train/grad_norm_pre_clip': 4.438643455505371, 'timing/data': 0.013411636929959059, 'timing/model': 1.114510533050634, 'learning_rate/qwen_vl_interface': 9.641335721775388e-06, 'learning_rate/action_model': 9.641335721775387e-05, 'epoch': 1.46}) +[2026-06-19 11:23:57,378][starVLA.training.train_starvla][INFO] - Step 711, Loss: {'train/loss': 0.14822310209274292, 'train/grad_norm_pre_clip': 2.451594114303589, 'timing/data': 0.01301620400045067, 'timing/model': 1.1082503179786727, 'learning_rate/qwen_vl_interface': 9.640173898252893e-06, 'learning_rate/action_model': 9.640173898252892e-05, 'epoch': 1.46}) +[2026-06-19 11:24:01,811][starVLA.training.train_starvla][INFO] - Step 712, Loss: {'train/loss': 0.2608344852924347, 'train/grad_norm_pre_clip': 4.595765113830566, 'timing/data': 0.013493351987563074, 'timing/model': 1.1094401550944895, 'learning_rate/qwen_vl_interface': 9.639010270095738e-06, 'learning_rate/action_model': 9.639010270095739e-05, 'epoch': 1.46}) +[2026-06-19 11:24:06,268][starVLA.training.train_starvla][INFO] - Step 713, Loss: {'train/loss': 0.22808659076690674, 'train/grad_norm_pre_clip': 4.4059672355651855, 'timing/data': 0.013255700003355742, 'timing/model': 1.1098180699627846, 'learning_rate/qwen_vl_interface': 9.637844837782251e-06, 'learning_rate/action_model': 9.63784483778225e-05, 'epoch': 1.46}) +[2026-06-19 11:24:10,719][starVLA.training.train_starvla][INFO] - Step 714, Loss: {'train/loss': 0.1461622267961502, 'train/grad_norm_pre_clip': 4.228842258453369, 'timing/data': 0.014698159997351468, 'timing/model': 1.1304618639405817, 'learning_rate/qwen_vl_interface': 9.636677601791494e-06, 'learning_rate/action_model': 9.636677601791493e-05, 'epoch': 1.47}) +[2026-06-19 11:24:15,186][starVLA.training.train_starvla][INFO] - Step 715, Loss: {'train/loss': 0.12165335565805435, 'train/grad_norm_pre_clip': 2.3138294219970703, 'timing/data': 0.012845110963098705, 'timing/model': 1.2819940550252795, 'learning_rate/qwen_vl_interface': 9.635508562603273e-06, 'learning_rate/action_model': 9.635508562603272e-05, 'epoch': 1.47}) +[2026-06-19 11:24:19,497][starVLA.training.train_starvla][INFO] - Step 716, Loss: {'train/loss': 0.7999882102012634, 'train/grad_norm_pre_clip': 18.646814346313477, 'timing/data': 0.013905538013204932, 'timing/model': 1.116039537009783, 'learning_rate/qwen_vl_interface': 9.634337720698139e-06, 'learning_rate/action_model': 9.634337720698137e-05, 'epoch': 1.47}) +[2026-06-19 11:24:23,940][starVLA.training.train_starvla][INFO] - Step 717, Loss: {'train/loss': 0.3314630389213562, 'train/grad_norm_pre_clip': 17.553743362426758, 'timing/data': 0.013325342908501625, 'timing/model': 1.1201164160156623, 'learning_rate/qwen_vl_interface': 9.633165076557376e-06, 'learning_rate/action_model': 9.633165076557375e-05, 'epoch': 1.47}) +[2026-06-19 11:24:28,411][starVLA.training.train_starvla][INFO] - Step 718, Loss: {'train/loss': 0.4001843333244324, 'train/grad_norm_pre_clip': 11.921396255493164, 'timing/data': 0.01354488602373749, 'timing/model': 1.1129095299402252, 'learning_rate/qwen_vl_interface': 9.631990630663019e-06, 'learning_rate/action_model': 9.631990630663018e-05, 'epoch': 1.47}) +[2026-06-19 11:24:32,827][starVLA.training.train_starvla][INFO] - Step 719, Loss: {'train/loss': 0.9562960267066956, 'train/grad_norm_pre_clip': 21.080156326293945, 'timing/data': 0.013123458018526435, 'timing/model': 1.1141048179706559, 'learning_rate/qwen_vl_interface': 9.630814383497836e-06, 'learning_rate/action_model': 9.630814383497835e-05, 'epoch': 1.48}) +[2026-06-19 11:24:37,292][starVLA.training.train_starvla][INFO] - Step 720, Loss: {'train/loss': 0.11856692284345627, 'train/grad_norm_pre_clip': 8.832817077636719, 'timing/data': 0.013471561018377542, 'timing/model': 1.1131677980301902, 'learning_rate/qwen_vl_interface': 9.629636335545336e-06, 'learning_rate/action_model': 9.629636335545335e-05, 'epoch': 1.48}) +[2026-06-19 11:24:41,759][starVLA.training.train_starvla][INFO] - Step 721, Loss: {'train/loss': 0.09971980005502701, 'train/grad_norm_pre_clip': 4.041805267333984, 'timing/data': 0.013138153008185327, 'timing/model': 1.2804598360089585, 'learning_rate/qwen_vl_interface': 9.628456487289773e-06, 'learning_rate/action_model': 9.628456487289773e-05, 'epoch': 1.48}) +[2026-06-19 11:24:46,067][starVLA.training.train_starvla][INFO] - Step 722, Loss: {'train/loss': 0.5480586886405945, 'train/grad_norm_pre_clip': 15.193868637084961, 'timing/data': 0.012949025025591254, 'timing/model': 1.1244960970943794, 'learning_rate/qwen_vl_interface': 9.627274839216135e-06, 'learning_rate/action_model': 9.627274839216136e-05, 'epoch': 1.48}) +[2026-06-19 11:24:50,545][starVLA.training.train_starvla][INFO] - Step 723, Loss: {'train/loss': 0.12391379475593567, 'train/grad_norm_pre_clip': 4.317989826202393, 'timing/data': 0.012653187965042889, 'timing/model': 1.119882829952985, 'learning_rate/qwen_vl_interface': 9.626091391810158e-06, 'learning_rate/action_model': 9.626091391810157e-05, 'epoch': 1.48}) +[2026-06-19 11:24:54,934][starVLA.training.train_starvla][INFO] - Step 724, Loss: {'train/loss': 0.25511452555656433, 'train/grad_norm_pre_clip': 8.195158004760742, 'timing/data': 0.013511996949091554, 'timing/model': 1.114511989057064, 'learning_rate/qwen_vl_interface': 9.624906145558309e-06, 'learning_rate/action_model': 9.624906145558307e-05, 'epoch': 1.49}) +[2026-06-19 11:24:59,387][starVLA.training.train_starvla][INFO] - Step 725, Loss: {'train/loss': 0.13033457100391388, 'train/grad_norm_pre_clip': 5.571442604064941, 'timing/data': 0.01326882594730705, 'timing/model': 1.1155836180550978, 'learning_rate/qwen_vl_interface': 9.623719100947795e-06, 'learning_rate/action_model': 9.623719100947795e-05, 'epoch': 1.49}) +[2026-06-19 11:25:03,879][starVLA.training.train_starvla][INFO] - Step 726, Loss: {'train/loss': 0.2572733163833618, 'train/grad_norm_pre_clip': 5.952765464782715, 'timing/data': 0.013551888056099415, 'timing/model': 1.1186019419692457, 'learning_rate/qwen_vl_interface': 9.622530258466571e-06, 'learning_rate/action_model': 9.62253025846657e-05, 'epoch': 1.49}) +[2026-06-19 11:25:08,373][starVLA.training.train_starvla][INFO] - Step 727, Loss: {'train/loss': 0.2710776925086975, 'train/grad_norm_pre_clip': 9.713822364807129, 'timing/data': 0.014005996054038405, 'timing/model': 1.297623749007471, 'learning_rate/qwen_vl_interface': 9.62133961860332e-06, 'learning_rate/action_model': 9.62133961860332e-05, 'epoch': 1.49}) +[2026-06-19 11:25:12,729][starVLA.training.train_starvla][INFO] - Step 728, Loss: {'train/loss': 0.7483265399932861, 'train/grad_norm_pre_clip': 19.93433952331543, 'timing/data': 0.013530409894883633, 'timing/model': 1.119341738987714, 'learning_rate/qwen_vl_interface': 9.620147181847474e-06, 'learning_rate/action_model': 9.620147181847475e-05, 'epoch': 1.5}) +[2026-06-19 11:25:17,211][starVLA.training.train_starvla][INFO] - Step 729, Loss: {'train/loss': 0.23526786267757416, 'train/grad_norm_pre_clip': 6.448571681976318, 'timing/data': 0.012403373955748975, 'timing/model': 1.1171582989627495, 'learning_rate/qwen_vl_interface': 9.618952948689196e-06, 'learning_rate/action_model': 9.618952948689196e-05, 'epoch': 1.5}) +[2026-06-19 11:25:21,657][starVLA.training.train_starvla][INFO] - Step 730, Loss: {'train/loss': 0.20030702650547028, 'train/grad_norm_pre_clip': 6.257338047027588, 'timing/data': 0.013053425936959684, 'timing/model': 1.1256890170043334, 'learning_rate/qwen_vl_interface': 9.617756919619393e-06, 'learning_rate/action_model': 9.617756919619392e-05, 'epoch': 1.5}) +[2026-06-19 11:25:26,097][starVLA.training.train_starvla][INFO] - Step 731, Loss: {'train/loss': 0.10250972956418991, 'train/grad_norm_pre_clip': 3.6615688800811768, 'timing/data': 0.012691137962974608, 'timing/model': 1.1164787110174075, 'learning_rate/qwen_vl_interface': 9.616559095129701e-06, 'learning_rate/action_model': 9.6165590951297e-05, 'epoch': 1.5}) +[2026-06-19 11:25:30,573][starVLA.training.train_starvla][INFO] - Step 732, Loss: {'train/loss': 0.1440555304288864, 'train/grad_norm_pre_clip': 7.53241491317749, 'timing/data': 0.01289264706429094, 'timing/model': 1.1185228590620682, 'learning_rate/qwen_vl_interface': 9.615359475712506e-06, 'learning_rate/action_model': 9.615359475712506e-05, 'epoch': 1.5}) +[2026-06-19 11:25:35,061][starVLA.training.train_starvla][INFO] - Step 733, Loss: {'train/loss': 0.2035897821187973, 'train/grad_norm_pre_clip': 6.54819917678833, 'timing/data': 0.013230227981694043, 'timing/model': 1.2902306030737236, 'learning_rate/qwen_vl_interface': 9.614158061860924e-06, 'learning_rate/action_model': 9.614158061860924e-05, 'epoch': 1.51}) +[2026-06-19 11:25:39,354][starVLA.training.train_starvla][INFO] - Step 734, Loss: {'train/loss': 0.2431139498949051, 'train/grad_norm_pre_clip': 4.916697025299072, 'timing/data': 0.012064080103300512, 'timing/model': 1.1142933790106326, 'learning_rate/qwen_vl_interface': 9.612954854068813e-06, 'learning_rate/action_model': 9.612954854068812e-05, 'epoch': 1.51}) +[2026-06-19 11:25:43,808][starVLA.training.train_starvla][INFO] - Step 735, Loss: {'train/loss': 0.10550960153341293, 'train/grad_norm_pre_clip': 7.571264266967773, 'timing/data': 0.012689525028690696, 'timing/model': 1.115349392988719, 'learning_rate/qwen_vl_interface': 9.611749852830763e-06, 'learning_rate/action_model': 9.611749852830763e-05, 'epoch': 1.51}) +[2026-06-19 11:25:48,256][starVLA.training.train_starvla][INFO] - Step 736, Loss: {'train/loss': 0.2645242214202881, 'train/grad_norm_pre_clip': 10.39613151550293, 'timing/data': 0.01312096999026835, 'timing/model': 1.1206282620551065, 'learning_rate/qwen_vl_interface': 9.610543058642107e-06, 'learning_rate/action_model': 9.610543058642107e-05, 'epoch': 1.51}) +[2026-06-19 11:25:52,744][starVLA.training.train_starvla][INFO] - Step 737, Loss: {'train/loss': 0.28830015659332275, 'train/grad_norm_pre_clip': 10.906814575195312, 'timing/data': 0.012247945996932685, 'timing/model': 1.126798129058443, 'learning_rate/qwen_vl_interface': 9.609334471998911e-06, 'learning_rate/action_model': 9.609334471998912e-05, 'epoch': 1.51}) +[2026-06-19 11:25:57,224][starVLA.training.train_starvla][INFO] - Step 738, Loss: {'train/loss': 0.26115480065345764, 'train/grad_norm_pre_clip': 10.38032054901123, 'timing/data': 0.013389019062742591, 'timing/model': 1.1173415400553495, 'learning_rate/qwen_vl_interface': 9.60812409339798e-06, 'learning_rate/action_model': 9.60812409339798e-05, 'epoch': 1.52}) +[2026-06-19 11:26:01,686][starVLA.training.train_starvla][INFO] - Step 739, Loss: {'train/loss': 0.13062244653701782, 'train/grad_norm_pre_clip': 6.319501876831055, 'timing/data': 0.012563657015562057, 'timing/model': 1.273724662954919, 'learning_rate/qwen_vl_interface': 9.606911923336856e-06, 'learning_rate/action_model': 9.606911923336856e-05, 'epoch': 1.52}) +[2026-06-19 11:26:06,101][starVLA.training.train_starvla][INFO] - Step 740, Loss: {'train/loss': 0.3377856910228729, 'train/grad_norm_pre_clip': 7.306407928466797, 'timing/data': 0.012481782934628427, 'timing/model': 1.2226896260399371, 'learning_rate/qwen_vl_interface': 9.605697962313817e-06, 'learning_rate/action_model': 9.605697962313816e-05, 'epoch': 1.52}) +[2026-06-19 11:26:10,432][starVLA.training.train_starvla][INFO] - Step 741, Loss: {'train/loss': 0.24955689907073975, 'train/grad_norm_pre_clip': 6.518702507019043, 'timing/data': 0.01384854398202151, 'timing/model': 1.1270491629838943, 'learning_rate/qwen_vl_interface': 9.604482210827871e-06, 'learning_rate/action_model': 9.604482210827871e-05, 'epoch': 1.52}) +[2026-06-19 11:26:14,903][starVLA.training.train_starvla][INFO] - Step 742, Loss: {'train/loss': 0.25247618556022644, 'train/grad_norm_pre_clip': 7.404041767120361, 'timing/data': 0.01219088805373758, 'timing/model': 1.1217681100824848, 'learning_rate/qwen_vl_interface': 9.603264669378772e-06, 'learning_rate/action_model': 9.603264669378772e-05, 'epoch': 1.52}) +[2026-06-19 11:26:19,406][starVLA.training.train_starvla][INFO] - Step 743, Loss: {'train/loss': 0.3540791869163513, 'train/grad_norm_pre_clip': 7.044919490814209, 'timing/data': 0.013822918990626931, 'timing/model': 1.1190016539767385, 'learning_rate/qwen_vl_interface': 9.602045338467007e-06, 'learning_rate/action_model': 9.602045338467006e-05, 'epoch': 1.53}) +[2026-06-19 11:26:23,916][starVLA.training.train_starvla][INFO] - Step 744, Loss: {'train/loss': 0.33908209204673767, 'train/grad_norm_pre_clip': 7.042318344116211, 'timing/data': 0.012958644074387848, 'timing/model': 1.123242600937374, 'learning_rate/qwen_vl_interface': 9.600824218593791e-06, 'learning_rate/action_model': 9.600824218593791e-05, 'epoch': 1.53}) +[2026-06-19 11:26:28,421][starVLA.training.train_starvla][INFO] - Step 745, Loss: {'train/loss': 0.166426882147789, 'train/grad_norm_pre_clip': 5.9912848472595215, 'timing/data': 0.012643591035157442, 'timing/model': 1.292618889012374, 'learning_rate/qwen_vl_interface': 9.599601310261087e-06, 'learning_rate/action_model': 9.599601310261085e-05, 'epoch': 1.53}) +[2026-06-19 11:26:32,893][starVLA.training.train_starvla][INFO] - Step 746, Loss: {'train/loss': 0.24136878550052643, 'train/grad_norm_pre_clip': 4.508784770965576, 'timing/data': 0.013647372019477189, 'timing/model': 1.2512924809707329, 'learning_rate/qwen_vl_interface': 9.598376613971582e-06, 'learning_rate/action_model': 9.598376613971582e-05, 'epoch': 1.53}) +[2026-06-19 11:26:37,210][starVLA.training.train_starvla][INFO] - Step 747, Loss: {'train/loss': 0.26622575521469116, 'train/grad_norm_pre_clip': 8.129922866821289, 'timing/data': 0.012810498010367155, 'timing/model': 1.1284489030949771, 'learning_rate/qwen_vl_interface': 9.597150130228703e-06, 'learning_rate/action_model': 9.597150130228703e-05, 'epoch': 1.53}) +[2026-06-19 11:26:41,684][starVLA.training.train_starvla][INFO] - Step 748, Loss: {'train/loss': 0.3753740191459656, 'train/grad_norm_pre_clip': 4.196958541870117, 'timing/data': 0.011687921942211688, 'timing/model': 1.126978193060495, 'learning_rate/qwen_vl_interface': 9.595921859536612e-06, 'learning_rate/action_model': 9.595921859536611e-05, 'epoch': 1.54}) +[2026-06-19 11:26:46,168][starVLA.training.train_starvla][INFO] - Step 749, Loss: {'train/loss': 0.20376357436180115, 'train/grad_norm_pre_clip': 3.672452688217163, 'timing/data': 0.0133234360255301, 'timing/model': 1.1188186929794028, 'learning_rate/qwen_vl_interface': 9.594691802400203e-06, 'learning_rate/action_model': 9.594691802400202e-05, 'epoch': 1.54}) +[2026-06-19 11:26:50,685][starVLA.training.train_starvla][INFO] - Step 750, Loss: {'train/loss': 0.20099200308322906, 'train/grad_norm_pre_clip': 2.0103061199188232, 'timing/data': 0.01231756096240133, 'timing/model': 1.1263069359119982, 'learning_rate/qwen_vl_interface': 9.59345995932511e-06, 'learning_rate/action_model': 9.59345995932511e-05, 'epoch': 1.54}) +[2026-06-19 11:26:55,185][starVLA.training.train_starvla][INFO] - Step 751, Loss: {'train/loss': 0.2212950736284256, 'train/grad_norm_pre_clip': 4.968270301818848, 'timing/data': 0.013900137972086668, 'timing/model': 1.2847996889613569, 'learning_rate/qwen_vl_interface': 9.592226330817693e-06, 'learning_rate/action_model': 9.592226330817693e-05, 'epoch': 1.54}) +[2026-06-19 11:26:59,504][starVLA.training.train_starvla][INFO] - Step 752, Loss: {'train/loss': 0.18767665326595306, 'train/grad_norm_pre_clip': 5.355673789978027, 'timing/data': 0.012652828940190375, 'timing/model': 1.125826821080409, 'learning_rate/qwen_vl_interface': 9.590990917385052e-06, 'learning_rate/action_model': 9.590990917385052e-05, 'epoch': 1.54}) +[2026-06-19 11:27:03,982][starVLA.training.train_starvla][INFO] - Step 753, Loss: {'train/loss': 0.22862908244132996, 'train/grad_norm_pre_clip': 3.5427680015563965, 'timing/data': 0.01272917096503079, 'timing/model': 1.129212795989588, 'learning_rate/qwen_vl_interface': 9.58975371953502e-06, 'learning_rate/action_model': 9.589753719535018e-05, 'epoch': 1.55}) +[2026-06-19 11:27:08,502][starVLA.training.train_starvla][INFO] - Step 754, Loss: {'train/loss': 0.1578541100025177, 'train/grad_norm_pre_clip': 2.4311957359313965, 'timing/data': 0.012980055063962936, 'timing/model': 1.1188033720245585, 'learning_rate/qwen_vl_interface': 9.588514737776158e-06, 'learning_rate/action_model': 9.588514737776157e-05, 'epoch': 1.55}) +[2026-06-19 11:27:13,016][starVLA.training.train_starvla][INFO] - Step 755, Loss: {'train/loss': 0.44725802540779114, 'train/grad_norm_pre_clip': 10.335878372192383, 'timing/data': 0.013250319054350257, 'timing/model': 1.1368286949582398, 'learning_rate/qwen_vl_interface': 9.587273972617769e-06, 'learning_rate/action_model': 9.587273972617768e-05, 'epoch': 1.55}) +[2026-06-19 11:27:17,432][starVLA.training.train_starvla][INFO] - Step 756, Loss: {'train/loss': 0.12150661647319794, 'train/grad_norm_pre_clip': 4.163883209228516, 'timing/data': 0.012941926950588822, 'timing/model': 1.121622201986611, 'learning_rate/qwen_vl_interface': 9.586031424569882e-06, 'learning_rate/action_model': 9.58603142456988e-05, 'epoch': 1.55}) +[2026-06-19 11:27:21,865][starVLA.training.train_starvla][INFO] - Step 757, Loss: {'train/loss': 0.4559183120727539, 'train/grad_norm_pre_clip': 11.12833309173584, 'timing/data': 0.013465580996125937, 'timing/model': 1.2310417389962822, 'learning_rate/qwen_vl_interface': 9.584787094143265e-06, 'learning_rate/action_model': 9.584787094143264e-05, 'epoch': 1.55}) +[2026-06-19 11:27:26,188][starVLA.training.train_starvla][INFO] - Step 758, Loss: {'train/loss': 0.15490223467350006, 'train/grad_norm_pre_clip': 3.4947052001953125, 'timing/data': 0.012683717999607325, 'timing/model': 1.1210012120427564, 'learning_rate/qwen_vl_interface': 9.58354098184941e-06, 'learning_rate/action_model': 9.583540981849409e-05, 'epoch': 1.56}) +[2026-06-19 11:27:30,666][starVLA.training.train_starvla][INFO] - Step 759, Loss: {'train/loss': 0.3504031002521515, 'train/grad_norm_pre_clip': 6.530061721801758, 'timing/data': 0.013511377037502825, 'timing/model': 1.1266073569422588, 'learning_rate/qwen_vl_interface': 9.582293088200549e-06, 'learning_rate/action_model': 9.582293088200549e-05, 'epoch': 1.56}) +[2026-06-19 11:27:35,162][starVLA.training.train_starvla][INFO] - Step 760, Loss: {'train/loss': 0.2133350819349289, 'train/grad_norm_pre_clip': 8.146645545959473, 'timing/data': 0.012562216958031058, 'timing/model': 1.1234610490500927, 'learning_rate/qwen_vl_interface': 9.581043413709646e-06, 'learning_rate/action_model': 9.581043413709646e-05, 'epoch': 1.56}) +[2026-06-19 11:27:39,632][starVLA.training.train_starvla][INFO] - Step 761, Loss: {'train/loss': 0.1719023734331131, 'train/grad_norm_pre_clip': 4.432155132293701, 'timing/data': 0.012936377082951367, 'timing/model': 1.1180247439770028, 'learning_rate/qwen_vl_interface': 9.579791958890393e-06, 'learning_rate/action_model': 9.579791958890394e-05, 'epoch': 1.56}) +[2026-06-19 11:27:44,075][starVLA.training.train_starvla][INFO] - Step 762, Loss: {'train/loss': 0.048798199743032455, 'train/grad_norm_pre_clip': 2.400294065475464, 'timing/data': 0.013122459058649838, 'timing/model': 1.1245250519132242, 'learning_rate/qwen_vl_interface': 9.578538724257218e-06, 'learning_rate/action_model': 9.578538724257216e-05, 'epoch': 1.57}) +[2026-06-19 11:27:48,546][starVLA.training.train_starvla][INFO] - Step 763, Loss: {'train/loss': 0.22102569043636322, 'train/grad_norm_pre_clip': 6.622831344604492, 'timing/data': 0.012392483069561422, 'timing/model': 1.2542392238974571, 'learning_rate/qwen_vl_interface': 9.577283710325274e-06, 'learning_rate/action_model': 9.577283710325274e-05, 'epoch': 1.57}) +[2026-06-19 11:27:52,864][starVLA.training.train_starvla][INFO] - Step 764, Loss: {'train/loss': 0.2050376534461975, 'train/grad_norm_pre_clip': 9.000115394592285, 'timing/data': 0.012407667934894562, 'timing/model': 1.1186231399187818, 'learning_rate/qwen_vl_interface': 9.576026917610455e-06, 'learning_rate/action_model': 9.576026917610455e-05, 'epoch': 1.57}) +[2026-06-19 11:27:57,379][starVLA.training.train_starvla][INFO] - Step 765, Loss: {'train/loss': 0.17715983092784882, 'train/grad_norm_pre_clip': 8.364712715148926, 'timing/data': 0.012673516990616918, 'timing/model': 1.1310552989598364, 'learning_rate/qwen_vl_interface': 9.574768346629376e-06, 'learning_rate/action_model': 9.574768346629376e-05, 'epoch': 1.57}) +[2026-06-19 11:28:01,860][starVLA.training.train_starvla][INFO] - Step 766, Loss: {'train/loss': 0.2072482854127884, 'train/grad_norm_pre_clip': 4.768599987030029, 'timing/data': 0.012776082963682711, 'timing/model': 1.1219438520492986, 'learning_rate/qwen_vl_interface': 9.573507997899391e-06, 'learning_rate/action_model': 9.573507997899392e-05, 'epoch': 1.57}) +[2026-06-19 11:28:06,338][starVLA.training.train_starvla][INFO] - Step 767, Loss: {'train/loss': 0.11977997422218323, 'train/grad_norm_pre_clip': 6.864879131317139, 'timing/data': 0.01244438102003187, 'timing/model': 1.117894210969098, 'learning_rate/qwen_vl_interface': 9.572245871938582e-06, 'learning_rate/action_model': 9.572245871938582e-05, 'epoch': 1.58}) +[2026-06-19 11:28:10,829][starVLA.training.train_starvla][INFO] - Step 768, Loss: {'train/loss': 0.2597771883010864, 'train/grad_norm_pre_clip': 7.546469688415527, 'timing/data': 0.012779704993590713, 'timing/model': 1.2999038679990917, 'learning_rate/qwen_vl_interface': 9.570981969265761e-06, 'learning_rate/action_model': 9.570981969265761e-05, 'epoch': 1.58}) +[2026-06-19 11:28:15,150][starVLA.training.train_starvla][INFO] - Step 769, Loss: {'train/loss': 0.42498040199279785, 'train/grad_norm_pre_clip': 17.33365821838379, 'timing/data': 0.01452973100822419, 'timing/model': 1.129922364023514, 'learning_rate/qwen_vl_interface': 9.56971629040047e-06, 'learning_rate/action_model': 9.56971629040047e-05, 'epoch': 1.58}) +[2026-06-19 11:28:19,615][starVLA.training.train_starvla][INFO] - Step 770, Loss: {'train/loss': 0.27751457691192627, 'train/grad_norm_pre_clip': 6.646414279937744, 'timing/data': 0.01269545208197087, 'timing/model': 1.1243060770211741, 'learning_rate/qwen_vl_interface': 9.56844883586298e-06, 'learning_rate/action_model': 9.568448835862981e-05, 'epoch': 1.58}) +[2026-06-19 11:28:24,005][starVLA.training.train_starvla][INFO] - Step 771, Loss: {'train/loss': 0.18015484511852264, 'train/grad_norm_pre_clip': 6.702603816986084, 'timing/data': 0.014818537048995495, 'timing/model': 1.1325796929886565, 'learning_rate/qwen_vl_interface': 9.567179606174299e-06, 'learning_rate/action_model': 9.567179606174299e-05, 'epoch': 1.58}) +[2026-06-19 11:28:28,438][starVLA.training.train_starvla][INFO] - Step 772, Loss: {'train/loss': 0.08538343757390976, 'train/grad_norm_pre_clip': 3.113349676132202, 'timing/data': 0.013025609077885747, 'timing/model': 1.1253684479743242, 'learning_rate/qwen_vl_interface': 9.565908601856155e-06, 'learning_rate/action_model': 9.565908601856154e-05, 'epoch': 1.59}) +[2026-06-19 11:28:32,909][starVLA.training.train_starvla][INFO] - Step 773, Loss: {'train/loss': 0.260836124420166, 'train/grad_norm_pre_clip': 7.573538780212402, 'timing/data': 0.01297478296328336, 'timing/model': 1.1254036499885842, 'learning_rate/qwen_vl_interface': 9.56463582343101e-06, 'learning_rate/action_model': 9.56463582343101e-05, 'epoch': 1.59}) +[2026-06-19 11:28:37,409][starVLA.training.train_starvla][INFO] - Step 774, Loss: {'train/loss': 0.1535213589668274, 'train/grad_norm_pre_clip': 6.375813007354736, 'timing/data': 0.01236569497268647, 'timing/model': 1.3028618389507756, 'learning_rate/qwen_vl_interface': 9.563361271422056e-06, 'learning_rate/action_model': 9.563361271422056e-05, 'epoch': 1.59}) +[2026-06-19 11:28:41,754][starVLA.training.train_starvla][INFO] - Step 775, Loss: {'train/loss': 0.518231987953186, 'train/grad_norm_pre_clip': 11.41609001159668, 'timing/data': 0.013819537940435112, 'timing/model': 1.1326020460110158, 'learning_rate/qwen_vl_interface': 9.562084946353212e-06, 'learning_rate/action_model': 9.562084946353213e-05, 'epoch': 1.59}) +[2026-06-19 11:28:46,239][starVLA.training.train_starvla][INFO] - Step 776, Loss: {'train/loss': 0.28107258677482605, 'train/grad_norm_pre_clip': 7.923391342163086, 'timing/data': 0.013232589000836015, 'timing/model': 1.130743829999119, 'learning_rate/qwen_vl_interface': 9.560806848749132e-06, 'learning_rate/action_model': 9.560806848749132e-05, 'epoch': 1.59}) +[2026-06-19 11:28:50,755][starVLA.training.train_starvla][INFO] - Step 777, Loss: {'train/loss': 0.13666364550590515, 'train/grad_norm_pre_clip': 7.149136543273926, 'timing/data': 0.012312390957958996, 'timing/model': 1.1321547749685124, 'learning_rate/qwen_vl_interface': 9.559526979135186e-06, 'learning_rate/action_model': 9.559526979135187e-05, 'epoch': 1.6}) +[2026-06-19 11:28:55,197][starVLA.training.train_starvla][INFO] - Step 778, Loss: {'train/loss': 0.20160327851772308, 'train/grad_norm_pre_clip': 8.826179504394531, 'timing/data': 0.012350811040960252, 'timing/model': 1.1158633480081335, 'learning_rate/qwen_vl_interface': 9.558245338037486e-06, 'learning_rate/action_model': 9.558245338037486e-05, 'epoch': 1.6}) +[2026-06-19 11:28:59,632][starVLA.training.train_starvla][INFO] - Step 779, Loss: {'train/loss': 0.16168908774852753, 'train/grad_norm_pre_clip': 3.27205753326416, 'timing/data': 0.012990523944608867, 'timing/model': 1.1205472470028326, 'learning_rate/qwen_vl_interface': 9.556961925982864e-06, 'learning_rate/action_model': 9.556961925982863e-05, 'epoch': 1.6}) +[2026-06-19 11:29:04,127][starVLA.training.train_starvla][INFO] - Step 780, Loss: {'train/loss': 0.2793543040752411, 'train/grad_norm_pre_clip': 9.432287216186523, 'timing/data': 0.013530993950553238, 'timing/model': 1.2963309320621192, 'learning_rate/qwen_vl_interface': 9.555676743498882e-06, 'learning_rate/action_model': 9.555676743498881e-05, 'epoch': 1.6}) +[2026-06-19 11:29:08,442][starVLA.training.train_starvla][INFO] - Step 781, Loss: {'train/loss': 0.18515975773334503, 'train/grad_norm_pre_clip': 7.315865993499756, 'timing/data': 0.012822119984775782, 'timing/model': 1.1249722429784015, 'learning_rate/qwen_vl_interface': 9.55438979111383e-06, 'learning_rate/action_model': 9.554389791113831e-05, 'epoch': 1.6}) +[2026-06-19 11:29:12,915][starVLA.training.train_starvla][INFO] - Step 782, Loss: {'train/loss': 0.09583649784326553, 'train/grad_norm_pre_clip': 4.551312446594238, 'timing/data': 0.013096745009534061, 'timing/model': 1.120070945005864, 'learning_rate/qwen_vl_interface': 9.553101069356728e-06, 'learning_rate/action_model': 9.553101069356727e-05, 'epoch': 1.61}) +[2026-06-19 11:29:17,406][starVLA.training.train_starvla][INFO] - Step 783, Loss: {'train/loss': 0.196598082780838, 'train/grad_norm_pre_clip': 6.2721123695373535, 'timing/data': 0.012995283002965152, 'timing/model': 1.1210627700202167, 'learning_rate/qwen_vl_interface': 9.551810578757316e-06, 'learning_rate/action_model': 9.551810578757315e-05, 'epoch': 1.61}) +[2026-06-19 11:29:21,904][starVLA.training.train_starvla][INFO] - Step 784, Loss: {'train/loss': 0.24620869755744934, 'train/grad_norm_pre_clip': 9.977005004882812, 'timing/data': 0.01344522600993514, 'timing/model': 1.1258499630494043, 'learning_rate/qwen_vl_interface': 9.550518319846069e-06, 'learning_rate/action_model': 9.550518319846069e-05, 'epoch': 1.61}) +[2026-06-19 11:29:26,398][starVLA.training.train_starvla][INFO] - Step 785, Loss: {'train/loss': 0.2688750922679901, 'train/grad_norm_pre_clip': 7.1720075607299805, 'timing/data': 0.012787350919097662, 'timing/model': 1.1269540389766917, 'learning_rate/qwen_vl_interface': 9.549224293154187e-06, 'learning_rate/action_model': 9.549224293154187e-05, 'epoch': 1.61}) +[2026-06-19 11:29:30,885][starVLA.training.train_starvla][INFO] - Step 786, Loss: {'train/loss': 0.2316838502883911, 'train/grad_norm_pre_clip': 5.719647407531738, 'timing/data': 0.012540577095933259, 'timing/model': 1.2838715839898214, 'learning_rate/qwen_vl_interface': 9.547928499213593e-06, 'learning_rate/action_model': 9.547928499213593e-05, 'epoch': 1.61}) +[2026-06-19 11:29:35,225][starVLA.training.train_starvla][INFO] - Step 787, Loss: {'train/loss': 0.43456846475601196, 'train/grad_norm_pre_clip': 8.07705307006836, 'timing/data': 0.013319186982698739, 'timing/model': 1.124519611010328, 'learning_rate/qwen_vl_interface': 9.54663093855694e-06, 'learning_rate/action_model': 9.54663093855694e-05, 'epoch': 1.62}) +[2026-06-19 11:29:39,706][starVLA.training.train_starvla][INFO] - Step 788, Loss: {'train/loss': 0.1908540278673172, 'train/grad_norm_pre_clip': 4.797638416290283, 'timing/data': 0.013700324925594032, 'timing/model': 1.1250956419389695, 'learning_rate/qwen_vl_interface': 9.545331611717607e-06, 'learning_rate/action_model': 9.545331611717608e-05, 'epoch': 1.62}) +[2026-06-19 11:29:44,213][starVLA.training.train_starvla][INFO] - Step 789, Loss: {'train/loss': 0.07719884067773819, 'train/grad_norm_pre_clip': 3.3927478790283203, 'timing/data': 0.013797497027553618, 'timing/model': 1.1316127810860053, 'learning_rate/qwen_vl_interface': 9.544030519229697e-06, 'learning_rate/action_model': 9.544030519229697e-05, 'epoch': 1.62}) +[2026-06-19 11:29:48,672][starVLA.training.train_starvla][INFO] - Step 790, Loss: {'train/loss': 0.3007400631904602, 'train/grad_norm_pre_clip': 4.603348255157471, 'timing/data': 0.012948465999215841, 'timing/model': 1.12532130908221, 'learning_rate/qwen_vl_interface': 9.54272766162804e-06, 'learning_rate/action_model': 9.54272766162804e-05, 'epoch': 1.62}) +[2026-06-19 11:29:53,180][starVLA.training.train_starvla][INFO] - Step 791, Loss: {'train/loss': 0.1538086086511612, 'train/grad_norm_pre_clip': 5.76220703125, 'timing/data': 0.01275952102150768, 'timing/model': 1.1390190729871392, 'learning_rate/qwen_vl_interface': 9.54142303944819e-06, 'learning_rate/action_model': 9.541423039448191e-05, 'epoch': 1.62}) +[2026-06-19 11:29:57,730][starVLA.training.train_starvla][INFO] - Step 792, Loss: {'train/loss': 0.08683203905820847, 'train/grad_norm_pre_clip': 2.2020351886749268, 'timing/data': 0.013393005006946623, 'timing/model': 1.3273966810666025, 'learning_rate/qwen_vl_interface': 9.540116653226433e-06, 'learning_rate/action_model': 9.540116653226431e-05, 'epoch': 1.63}) +[2026-06-19 11:30:02,074][starVLA.training.train_starvla][INFO] - Step 793, Loss: {'train/loss': 0.15008842945098877, 'train/grad_norm_pre_clip': 4.297364234924316, 'timing/data': 0.012552374042570591, 'timing/model': 1.1340644609881565, 'learning_rate/qwen_vl_interface': 9.53880850349977e-06, 'learning_rate/action_model': 9.538808503499768e-05, 'epoch': 1.63}) +[2026-06-19 11:30:06,583][starVLA.training.train_starvla][INFO] - Step 794, Loss: {'train/loss': 0.16237597167491913, 'train/grad_norm_pre_clip': 3.5324583053588867, 'timing/data': 0.012795922928489745, 'timing/model': 1.1253345359582454, 'learning_rate/qwen_vl_interface': 9.537498590805934e-06, 'learning_rate/action_model': 9.537498590805933e-05, 'epoch': 1.63}) +[2026-06-19 11:30:11,035][starVLA.training.train_starvla][INFO] - Step 795, Loss: {'train/loss': 0.1923571527004242, 'train/grad_norm_pre_clip': 5.641253471374512, 'timing/data': 0.01353965897578746, 'timing/model': 1.1274178790627047, 'learning_rate/qwen_vl_interface': 9.536186915683379e-06, 'learning_rate/action_model': 9.53618691568338e-05, 'epoch': 1.63}) +[2026-06-19 11:30:15,516][starVLA.training.train_starvla][INFO] - Step 796, Loss: {'train/loss': 0.12203430384397507, 'train/grad_norm_pre_clip': 3.41996431350708, 'timing/data': 0.012829576036892831, 'timing/model': 1.123431843938306, 'learning_rate/qwen_vl_interface': 9.534873478671289e-06, 'learning_rate/action_model': 9.534873478671288e-05, 'epoch': 1.63}) +[2026-06-19 11:30:20,015][starVLA.training.train_starvla][INFO] - Step 797, Loss: {'train/loss': 0.25733399391174316, 'train/grad_norm_pre_clip': 5.833800792694092, 'timing/data': 0.012903829105198383, 'timing/model': 1.1218085610307753, 'learning_rate/qwen_vl_interface': 9.533558280309563e-06, 'learning_rate/action_model': 9.533558280309563e-05, 'epoch': 1.64}) +[2026-06-19 11:30:24,537][starVLA.training.train_starvla][INFO] - Step 798, Loss: {'train/loss': 0.1653675138950348, 'train/grad_norm_pre_clip': 4.8656206130981445, 'timing/data': 0.012988183065317571, 'timing/model': 1.3085940110031515, 'learning_rate/qwen_vl_interface': 9.532241321138832e-06, 'learning_rate/action_model': 9.532241321138833e-05, 'epoch': 1.64}) +[2026-06-19 11:30:28,876][starVLA.training.train_starvla][INFO] - Step 799, Loss: {'train/loss': 0.18242302536964417, 'train/grad_norm_pre_clip': 5.344620227813721, 'timing/data': 0.012829994084313512, 'timing/model': 1.1249606899218634, 'learning_rate/qwen_vl_interface': 9.530922601700451e-06, 'learning_rate/action_model': 9.530922601700452e-05, 'epoch': 1.64}) +[2026-06-19 11:30:33,376][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 800 +[2026-06-19 11:33:26,699][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 800 +[2026-06-19 11:33:26,699][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 800 +[2026-06-19 11:36:32,226][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 800 +[2026-06-19 11:36:32,227][starVLA.training.train_starvla][INFO] - Step 800, Loss: {'train/loss': 0.3408585488796234, 'train/grad_norm_pre_clip': 7.9026570320129395, 'eval/loss': 0.48937752842903137, 'eval/latency_0/loss': 0.10873468220233917, 'eval/latency_1/loss': 0.29328325390815735, 'eval/latency_2/loss': 0.7103605270385742, 'eval/latency_3/loss': 0.5249128341674805, 'eval/latency_4/loss': 0.7767455577850342, 'eval/flappy/loss': 0.48937752842903137, 'eval/flappy/latency_0/loss': 0.10873468220233917, 'eval/flappy/latency_1/loss': 0.29328325390815735, 'eval/flappy/latency_2/loss': 0.7103605270385742, 'eval/flappy/latency_3/loss': 0.5249128341674805, 'eval/flappy/latency_4/loss': 0.7767455577850342, 'eval/flap_precision': 0.6395348837209303, 'eval/flap_recall': 0.8, 'eval/flap_f1': 0.7108239095315024, 'eval/flap_rate_model': 0.09555555555555556, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.7108239095315024, 'eval/latency_0/flap_precision': 0.6395348837209303, 'eval/latency_0/flap_recall': 0.8, 'eval/latency_0/flap_f1': 0.7108239095315024, 'eval/latency_0/flap_rate_model': 0.09555555555555556, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.7108239095315024, 'timing/data': 0.012526371981948614, 'timing/model': 1.1226162860402837, 'learning_rate/qwen_vl_interface': 9.52960212253649e-06, 'learning_rate/action_model': 9.529602122536491e-05, 'epoch': 1.64}) +[2026-06-19 11:36:32,228][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_800_state +[2026-06-19 11:36:32,230][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-19 11:36:35,494][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_800_state/model.safetensors +[2026-06-19 11:36:42,668][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_800_state/optimizer.bin +[2026-06-19 11:36:42,669][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_800_state/sampler.bin +[2026-06-19 11:36:42,669][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_800_state/sampler_1.bin +[2026-06-19 11:36:42,669][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_800_state/random_states_0.pkl +[2026-06-19 11:36:42,693][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 800 +[2026-06-19 11:36:42,693][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-19 11:36:42,707][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-19 11:36:47,091][starVLA.training.train_starvla][INFO] - Step 801, Loss: {'train/loss': 0.0770716667175293, 'train/grad_norm_pre_clip': 3.5057337284088135, 'timing/data': 0.012942861067131162, 'timing/model': 1.2594992640661076, 'learning_rate/qwen_vl_interface': 9.528279884189754e-06, 'learning_rate/action_model': 9.528279884189754e-05, 'epoch': 1.65}) +[2026-06-19 11:36:51,445][starVLA.training.train_starvla][INFO] - Step 802, Loss: {'train/loss': 0.2880474925041199, 'train/grad_norm_pre_clip': 8.153078079223633, 'timing/data': 0.013070976012386382, 'timing/model': 1.2223227040376514, 'learning_rate/qwen_vl_interface': 9.526955887203762e-06, 'learning_rate/action_model': 9.526955887203761e-05, 'epoch': 1.65}) +[2026-06-19 11:36:55,715][starVLA.training.train_starvla][INFO] - Step 803, Loss: {'train/loss': 0.3592454791069031, 'train/grad_norm_pre_clip': 15.879953384399414, 'timing/data': 0.01264333794824779, 'timing/model': 1.1015710330102593, 'learning_rate/qwen_vl_interface': 9.525630132122758e-06, 'learning_rate/action_model': 9.525630132122758e-05, 'epoch': 1.65}) +[2026-06-19 11:37:00,119][starVLA.training.train_starvla][INFO] - Step 804, Loss: {'train/loss': 0.10297629982233047, 'train/grad_norm_pre_clip': 3.185361862182617, 'timing/data': 0.014170182053931057, 'timing/model': 1.1059918029932305, 'learning_rate/qwen_vl_interface': 9.524302619491712e-06, 'learning_rate/action_model': 9.524302619491712e-05, 'epoch': 1.65}) +[2026-06-19 11:37:04,546][starVLA.training.train_starvla][INFO] - Step 805, Loss: {'train/loss': 0.3373345136642456, 'train/grad_norm_pre_clip': 5.811807155609131, 'timing/data': 0.012609933037310839, 'timing/model': 1.1052224579034373, 'learning_rate/qwen_vl_interface': 9.522973349856315e-06, 'learning_rate/action_model': 9.522973349856315e-05, 'epoch': 1.65}) +[2026-06-19 11:37:08,984][starVLA.training.train_starvla][INFO] - Step 806, Loss: {'train/loss': 0.2770599126815796, 'train/grad_norm_pre_clip': 7.247507095336914, 'timing/data': 0.013216087012551725, 'timing/model': 1.1111938449321315, 'learning_rate/qwen_vl_interface': 9.521642323762976e-06, 'learning_rate/action_model': 9.521642323762976e-05, 'epoch': 1.66}) +[2026-06-19 11:37:13,444][starVLA.training.train_starvla][INFO] - Step 807, Loss: {'train/loss': 0.2725253701210022, 'train/grad_norm_pre_clip': 11.414261817932129, 'timing/data': 0.013653172994963825, 'timing/model': 1.2904786989092827, 'learning_rate/qwen_vl_interface': 9.520309541758833e-06, 'learning_rate/action_model': 9.520309541758833e-05, 'epoch': 1.66}) +[2026-06-19 11:37:17,869][starVLA.training.train_starvla][INFO] - Step 808, Loss: {'train/loss': 0.1823333501815796, 'train/grad_norm_pre_clip': 7.256399154663086, 'timing/data': 0.01263514906167984, 'timing/model': 1.2567467290209606, 'learning_rate/qwen_vl_interface': 9.51897500439174e-06, 'learning_rate/action_model': 9.51897500439174e-05, 'epoch': 1.66}) +[2026-06-19 11:37:22,149][starVLA.training.train_starvla][INFO] - Step 809, Loss: {'train/loss': 0.08466917276382446, 'train/grad_norm_pre_clip': 3.174696445465088, 'timing/data': 0.011689257924444973, 'timing/model': 1.1181171260541305, 'learning_rate/qwen_vl_interface': 9.517638712210277e-06, 'learning_rate/action_model': 9.517638712210276e-05, 'epoch': 1.66}) +[2026-06-19 11:37:26,623][starVLA.training.train_starvla][INFO] - Step 810, Loss: {'train/loss': 0.17234769463539124, 'train/grad_norm_pre_clip': 15.226983070373535, 'timing/data': 0.013038274017162621, 'timing/model': 1.1230476238997653, 'learning_rate/qwen_vl_interface': 9.51630066576374e-06, 'learning_rate/action_model': 9.51630066576374e-05, 'epoch': 1.66}) +[2026-06-19 11:37:31,118][starVLA.training.train_starvla][INFO] - Step 811, Loss: {'train/loss': 0.22965343296527863, 'train/grad_norm_pre_clip': 7.262146472930908, 'timing/data': 0.015802334062755108, 'timing/model': 1.1305122189223766, 'learning_rate/qwen_vl_interface': 9.514960865602154e-06, 'learning_rate/action_model': 9.514960865602153e-05, 'epoch': 1.67}) +[2026-06-19 11:37:35,593][starVLA.training.train_starvla][INFO] - Step 812, Loss: {'train/loss': 0.13461442291736603, 'train/grad_norm_pre_clip': 3.7508058547973633, 'timing/data': 0.012774500995874405, 'timing/model': 1.1204181970097125, 'learning_rate/qwen_vl_interface': 9.513619312276259e-06, 'learning_rate/action_model': 9.513619312276258e-05, 'epoch': 1.67}) +[2026-06-19 11:37:40,000][starVLA.training.train_starvla][INFO] - Step 813, Loss: {'train/loss': 0.2672545313835144, 'train/grad_norm_pre_clip': 7.62996244430542, 'timing/data': 0.013325291918590665, 'timing/model': 1.1264962379354984, 'learning_rate/qwen_vl_interface': 9.512276006337514e-06, 'learning_rate/action_model': 9.512276006337513e-05, 'epoch': 1.67}) +[2026-06-19 11:37:44,443][starVLA.training.train_starvla][INFO] - Step 814, Loss: {'train/loss': 0.27727681398391724, 'train/grad_norm_pre_clip': 8.822009086608887, 'timing/data': 0.013258030056022108, 'timing/model': 1.2546509710373357, 'learning_rate/qwen_vl_interface': 9.510930948338102e-06, 'learning_rate/action_model': 9.510930948338102e-05, 'epoch': 1.67}) +[2026-06-19 11:37:48,773][starVLA.training.train_starvla][INFO] - Step 815, Loss: {'train/loss': 0.13563816249370575, 'train/grad_norm_pre_clip': 25.641586303710938, 'timing/data': 0.013456257991492748, 'timing/model': 1.1221747380914167, 'learning_rate/qwen_vl_interface': 9.509584138830929e-06, 'learning_rate/action_model': 9.509584138830928e-05, 'epoch': 1.67}) +[2026-06-19 11:37:53,227][starVLA.training.train_starvla][INFO] - Step 816, Loss: {'train/loss': 0.17406390607357025, 'train/grad_norm_pre_clip': 10.57373046875, 'timing/data': 0.012536771944724023, 'timing/model': 1.1124105749186128, 'learning_rate/qwen_vl_interface': 9.508235578369613e-06, 'learning_rate/action_model': 9.508235578369612e-05, 'epoch': 1.68}) +[2026-06-19 11:37:57,698][starVLA.training.train_starvla][INFO] - Step 817, Loss: {'train/loss': 0.3641696572303772, 'train/grad_norm_pre_clip': 9.369215965270996, 'timing/data': 0.013439749018289149, 'timing/model': 1.118330093100667, 'learning_rate/qwen_vl_interface': 9.506885267508498e-06, 'learning_rate/action_model': 9.506885267508498e-05, 'epoch': 1.68}) +[2026-06-19 11:38:02,174][starVLA.training.train_starvla][INFO] - Step 818, Loss: {'train/loss': 0.27406543493270874, 'train/grad_norm_pre_clip': 10.8716459274292, 'timing/data': 0.01274287304840982, 'timing/model': 1.1164115520659834, 'learning_rate/qwen_vl_interface': 9.505533206802648e-06, 'learning_rate/action_model': 9.505533206802647e-05, 'epoch': 1.68}) +[2026-06-19 11:38:06,662][starVLA.training.train_starvla][INFO] - Step 819, Loss: {'train/loss': 0.17724892497062683, 'train/grad_norm_pre_clip': 6.792652130126953, 'timing/data': 0.013128662947565317, 'timing/model': 1.286132276058197, 'learning_rate/qwen_vl_interface': 9.50417939680784e-06, 'learning_rate/action_model': 9.504179396807839e-05, 'epoch': 1.68}) +[2026-06-19 11:38:10,963][starVLA.training.train_starvla][INFO] - Step 820, Loss: {'train/loss': 0.1049497202038765, 'train/grad_norm_pre_clip': 5.739870071411133, 'timing/data': 0.012023534043692052, 'timing/model': 1.1172439099755138, 'learning_rate/qwen_vl_interface': 9.502823838080578e-06, 'learning_rate/action_model': 9.502823838080578e-05, 'epoch': 1.68}) +[2026-06-19 11:38:15,427][starVLA.training.train_starvla][INFO] - Step 821, Loss: {'train/loss': 0.07592909783124924, 'train/grad_norm_pre_clip': 5.410670757293701, 'timing/data': 0.011836263001896441, 'timing/model': 1.1143179839709774, 'learning_rate/qwen_vl_interface': 9.501466531178079e-06, 'learning_rate/action_model': 9.501466531178079e-05, 'epoch': 1.69}) +[2026-06-19 11:38:19,826][starVLA.training.train_starvla][INFO] - Step 822, Loss: {'train/loss': 0.14679378271102905, 'train/grad_norm_pre_clip': 8.293157577514648, 'timing/data': 0.013902582926675677, 'timing/model': 1.1230794820003211, 'learning_rate/qwen_vl_interface': 9.50010747665828e-06, 'learning_rate/action_model': 9.500107476658281e-05, 'epoch': 1.69}) +[2026-06-19 11:38:24,305][starVLA.training.train_starvla][INFO] - Step 823, Loss: {'train/loss': 0.16806964576244354, 'train/grad_norm_pre_clip': 8.374494552612305, 'timing/data': 0.014385457034222782, 'timing/model': 1.1157580509316176, 'learning_rate/qwen_vl_interface': 9.498746675079842e-06, 'learning_rate/action_model': 9.49874667507984e-05, 'epoch': 1.69}) +[2026-06-19 11:38:28,743][starVLA.training.train_starvla][INFO] - Step 824, Loss: {'train/loss': 0.1325799971818924, 'train/grad_norm_pre_clip': 6.1668925285339355, 'timing/data': 0.012807584018446505, 'timing/model': 1.124534170026891, 'learning_rate/qwen_vl_interface': 9.49738412700213e-06, 'learning_rate/action_model': 9.49738412700213e-05, 'epoch': 1.69}) +[2026-06-19 11:38:33,204][starVLA.training.train_starvla][INFO] - Step 825, Loss: {'train/loss': 0.1876290887594223, 'train/grad_norm_pre_clip': 7.771939277648926, 'timing/data': 0.014693393022753298, 'timing/model': 1.2672108920523897, 'learning_rate/qwen_vl_interface': 9.496019832985243e-06, 'learning_rate/action_model': 9.496019832985243e-05, 'epoch': 1.69}) +[2026-06-19 11:38:37,508][starVLA.training.train_starvla][INFO] - Step 826, Loss: {'train/loss': 0.37331825494766235, 'train/grad_norm_pre_clip': 21.771560668945312, 'timing/data': 0.01340109808370471, 'timing/model': 1.1194356239866465, 'learning_rate/qwen_vl_interface': 9.49465379358999e-06, 'learning_rate/action_model': 9.49465379358999e-05, 'epoch': 1.7}) +[2026-06-19 11:38:41,982][starVLA.training.train_starvla][INFO] - Step 827, Loss: {'train/loss': 0.16082513332366943, 'train/grad_norm_pre_clip': 8.28072452545166, 'timing/data': 0.013559450046159327, 'timing/model': 1.1197679640026763, 'learning_rate/qwen_vl_interface': 9.493286009377896e-06, 'learning_rate/action_model': 9.493286009377896e-05, 'epoch': 1.7}) +[2026-06-19 11:38:46,448][starVLA.training.train_starvla][INFO] - Step 828, Loss: {'train/loss': 0.12904641032218933, 'train/grad_norm_pre_clip': 6.297342777252197, 'timing/data': 0.012545245001092553, 'timing/model': 1.1129438419593498, 'learning_rate/qwen_vl_interface': 9.491916480911208e-06, 'learning_rate/action_model': 9.491916480911207e-05, 'epoch': 1.7}) +[2026-06-19 11:38:50,860][starVLA.training.train_starvla][INFO] - Step 829, Loss: {'train/loss': 0.2545614242553711, 'train/grad_norm_pre_clip': 8.367573738098145, 'timing/data': 0.013257418060675263, 'timing/model': 1.1142399379750714, 'learning_rate/qwen_vl_interface': 9.490545208752884e-06, 'learning_rate/action_model': 9.490545208752883e-05, 'epoch': 1.7}) +[2026-06-19 11:38:55,282][starVLA.training.train_starvla][INFO] - Step 830, Loss: {'train/loss': 0.12190741300582886, 'train/grad_norm_pre_clip': 5.446449279785156, 'timing/data': 0.012923454982228577, 'timing/model': 1.1146757940296084, 'learning_rate/qwen_vl_interface': 9.489172193466603e-06, 'learning_rate/action_model': 9.489172193466603e-05, 'epoch': 1.7}) +[2026-06-19 11:38:59,751][starVLA.training.train_starvla][INFO] - Step 831, Loss: {'train/loss': 0.2729957103729248, 'train/grad_norm_pre_clip': 9.23713493347168, 'timing/data': 0.013019628007896245, 'timing/model': 1.2855288669234142, 'learning_rate/qwen_vl_interface': 9.487797435616762e-06, 'learning_rate/action_model': 9.487797435616761e-05, 'epoch': 1.71}) +[2026-06-19 11:39:04,054][starVLA.training.train_starvla][INFO] - Step 832, Loss: {'train/loss': 0.2504577934741974, 'train/grad_norm_pre_clip': 9.56435489654541, 'timing/data': 0.013169492944143713, 'timing/model': 1.1198643300449476, 'learning_rate/qwen_vl_interface': 9.486420935768472e-06, 'learning_rate/action_model': 9.486420935768472e-05, 'epoch': 1.71}) +[2026-06-19 11:39:08,526][starVLA.training.train_starvla][INFO] - Step 833, Loss: {'train/loss': 0.24335768818855286, 'train/grad_norm_pre_clip': 7.9010186195373535, 'timing/data': 0.013646131963469088, 'timing/model': 1.1222408760804683, 'learning_rate/qwen_vl_interface': 9.485042694487558e-06, 'learning_rate/action_model': 9.485042694487559e-05, 'epoch': 1.71}) +[2026-06-19 11:39:13,854][starVLA.training.train_starvla][INFO] - Latency curriculum phase 1 at step 834: active_latencies=[0, 1], dataset_size=128365 +[2026-06-19 11:39:13,855][starVLA.training.train_starvla][INFO] - Step 834, Loss: {'train/loss': 0.18437056243419647, 'train/grad_norm_pre_clip': 8.522356986999512, 'latency_curriculum/phase': 1.0, 'latency_curriculum/active_count': 2.0, 'latency_curriculum/max_active_latency': 1.0, 'latency_curriculum/dataset_size': 128365.0, 'timing/data': 0.01225715596228838, 'timing/model': 1.1281464280327782, 'learning_rate/qwen_vl_interface': 9.483662712340565e-06, 'learning_rate/action_model': 9.483662712340564e-05, 'epoch': 0.83}) +[2026-06-19 11:39:30,445][starVLA.training.train_starvla][INFO] - Step 835, Loss: {'train/loss': 0.34718453884124756, 'train/grad_norm_pre_clip': 7.2402567863464355, 'timing/data': 0.012189495959319174, 'timing/model': 1.142584340996109, 'learning_rate/qwen_vl_interface': 9.482280989894748e-06, 'learning_rate/action_model': 9.482280989894748e-05, 'epoch': 0.83}) +[2026-06-19 11:39:35,031][starVLA.training.train_starvla][INFO] - Step 836, Loss: {'train/loss': 0.23849795758724213, 'train/grad_norm_pre_clip': 14.189504623413086, 'timing/data': 0.012794594978913665, 'timing/model': 1.1558402830269188, 'learning_rate/qwen_vl_interface': 9.480897527718083e-06, 'learning_rate/action_model': 9.480897527718084e-05, 'epoch': 0.83}) +[2026-06-19 11:39:39,620][starVLA.training.train_starvla][INFO] - Step 837, Loss: {'train/loss': 0.4055750370025635, 'train/grad_norm_pre_clip': 10.146071434020996, 'timing/data': 0.012630900018848479, 'timing/model': 1.318126452038996, 'learning_rate/qwen_vl_interface': 9.479512326379263e-06, 'learning_rate/action_model': 9.479512326379264e-05, 'epoch': 0.83}) +[2026-06-19 11:39:44,032][starVLA.training.train_starvla][INFO] - Step 838, Loss: {'train/loss': 0.5729761123657227, 'train/grad_norm_pre_clip': 13.004314422607422, 'timing/data': 0.011932904948480427, 'timing/model': 1.1365313229616731, 'learning_rate/qwen_vl_interface': 9.478125386447686e-06, 'learning_rate/action_model': 9.478125386447686e-05, 'epoch': 0.84}) +[2026-06-19 11:39:48,608][starVLA.training.train_starvla][INFO] - Step 839, Loss: {'train/loss': 0.7877330780029297, 'train/grad_norm_pre_clip': 23.606096267700195, 'timing/data': 0.012719306047074497, 'timing/model': 1.1394255539635196, 'learning_rate/qwen_vl_interface': 9.476736708493472e-06, 'learning_rate/action_model': 9.476736708493472e-05, 'epoch': 0.84}) +[2026-06-19 11:39:53,109][starVLA.training.train_starvla][INFO] - Step 840, Loss: {'train/loss': 0.2702479958534241, 'train/grad_norm_pre_clip': 6.751648426055908, 'timing/data': 0.012737679993733764, 'timing/model': 1.1450611839536577, 'learning_rate/qwen_vl_interface': 9.475346293087457e-06, 'learning_rate/action_model': 9.475346293087458e-05, 'epoch': 0.84}) +[2026-06-19 11:39:57,681][starVLA.training.train_starvla][INFO] - Step 841, Loss: {'train/loss': 0.3024982511997223, 'train/grad_norm_pre_clip': 7.219858646392822, 'timing/data': 0.012986387941055, 'timing/model': 1.1491800070507452, 'learning_rate/qwen_vl_interface': 9.473954140801184e-06, 'learning_rate/action_model': 9.473954140801185e-05, 'epoch': 0.84}) +[2026-06-19 11:40:02,281][starVLA.training.train_starvla][INFO] - Step 842, Loss: {'train/loss': 0.36288222670555115, 'train/grad_norm_pre_clip': 4.323429107666016, 'timing/data': 0.013174482039175928, 'timing/model': 1.1616177910473198, 'learning_rate/qwen_vl_interface': 9.472560252206918e-06, 'learning_rate/action_model': 9.472560252206918e-05, 'epoch': 0.84}) +[2026-06-19 11:40:06,900][starVLA.training.train_starvla][INFO] - Step 843, Loss: {'train/loss': 0.3150344491004944, 'train/grad_norm_pre_clip': 5.038417816162109, 'timing/data': 0.013671902008354664, 'timing/model': 1.3207740500802174, 'learning_rate/qwen_vl_interface': 9.471164627877633e-06, 'learning_rate/action_model': 9.471164627877633e-05, 'epoch': 0.84}) +[2026-06-19 11:40:11,340][starVLA.training.train_starvla][INFO] - Step 844, Loss: {'train/loss': 0.18511684238910675, 'train/grad_norm_pre_clip': 4.493946075439453, 'timing/data': 0.01351532491389662, 'timing/model': 1.143712160992436, 'learning_rate/qwen_vl_interface': 9.469767268387016e-06, 'learning_rate/action_model': 9.469767268387016e-05, 'epoch': 0.84}) +[2026-06-19 11:40:15,941][starVLA.training.train_starvla][INFO] - Step 845, Loss: {'train/loss': 0.2568652033805847, 'train/grad_norm_pre_clip': 3.284719705581665, 'timing/data': 0.01313600898720324, 'timing/model': 1.1505071710562333, 'learning_rate/qwen_vl_interface': 9.468368174309468e-06, 'learning_rate/action_model': 9.468368174309469e-05, 'epoch': 0.84}) +[2026-06-19 11:40:20,500][starVLA.training.train_starvla][INFO] - Step 846, Loss: {'train/loss': 0.4233413636684418, 'train/grad_norm_pre_clip': 4.892098903656006, 'timing/data': 0.01291955797933042, 'timing/model': 1.147491081035696, 'learning_rate/qwen_vl_interface': 9.466967346220107e-06, 'learning_rate/action_model': 9.466967346220107e-05, 'epoch': 0.84}) +[2026-06-19 11:40:25,111][starVLA.training.train_starvla][INFO] - Step 847, Loss: {'train/loss': 0.20329053699970245, 'train/grad_norm_pre_clip': 3.509004831314087, 'timing/data': 0.012795789050869644, 'timing/model': 1.142337525030598, 'learning_rate/qwen_vl_interface': 9.465564784694759e-06, 'learning_rate/action_model': 9.465564784694759e-05, 'epoch': 0.84}) +[2026-06-19 11:40:29,734][starVLA.training.train_starvla][INFO] - Step 848, Loss: {'train/loss': 0.3996514678001404, 'train/grad_norm_pre_clip': 8.045302391052246, 'timing/data': 0.012776145013049245, 'timing/model': 1.1564050369197503, 'learning_rate/qwen_vl_interface': 9.464160490309962e-06, 'learning_rate/action_model': 9.464160490309962e-05, 'epoch': 0.85}) +[2026-06-19 11:40:34,330][starVLA.training.train_starvla][INFO] - Step 849, Loss: {'train/loss': 0.319409042596817, 'train/grad_norm_pre_clip': 5.242275714874268, 'timing/data': 0.013548664981499314, 'timing/model': 1.3200549959437922, 'learning_rate/qwen_vl_interface': 9.462754463642972e-06, 'learning_rate/action_model': 9.46275446364297e-05, 'epoch': 0.85}) +[2026-06-19 11:40:38,773][starVLA.training.train_starvla][INFO] - Step 850, Loss: {'train/loss': 0.3652125298976898, 'train/grad_norm_pre_clip': 6.551331996917725, 'timing/data': 0.012322747032158077, 'timing/model': 1.1488867020234466, 'learning_rate/qwen_vl_interface': 9.461346705271753e-06, 'learning_rate/action_model': 9.461346705271752e-05, 'epoch': 0.85}) +[2026-06-19 11:40:43,362][starVLA.training.train_starvla][INFO] - Step 851, Loss: {'train/loss': 0.3791428208351135, 'train/grad_norm_pre_clip': 4.375485420227051, 'timing/data': 0.01320859498810023, 'timing/model': 1.1439462589332834, 'learning_rate/qwen_vl_interface': 9.459937215774976e-06, 'learning_rate/action_model': 9.459937215774976e-05, 'epoch': 0.85}) +[2026-06-19 11:40:47,955][starVLA.training.train_starvla][INFO] - Step 852, Loss: {'train/loss': 0.16389361023902893, 'train/grad_norm_pre_clip': 3.2769601345062256, 'timing/data': 0.012662458000704646, 'timing/model': 1.1507916579721496, 'learning_rate/qwen_vl_interface': 9.458525995732035e-06, 'learning_rate/action_model': 9.458525995732035e-05, 'epoch': 0.85}) +[2026-06-19 11:40:52,448][starVLA.training.train_starvla][INFO] - Step 853, Loss: {'train/loss': 0.3049789071083069, 'train/grad_norm_pre_clip': 3.564856767654419, 'timing/data': 0.013660075957886875, 'timing/model': 1.1431929789250717, 'learning_rate/qwen_vl_interface': 9.457113045723027e-06, 'learning_rate/action_model': 9.457113045723026e-05, 'epoch': 0.85}) +[2026-06-19 11:40:57,018][starVLA.training.train_starvla][INFO] - Step 854, Loss: {'train/loss': 0.40908682346343994, 'train/grad_norm_pre_clip': 7.564682960510254, 'timing/data': 0.013231043005362153, 'timing/model': 1.1432189759798348, 'learning_rate/qwen_vl_interface': 9.455698366328762e-06, 'learning_rate/action_model': 9.455698366328762e-05, 'epoch': 0.85}) +[2026-06-19 11:41:01,626][starVLA.training.train_starvla][INFO] - Step 855, Loss: {'train/loss': 0.5278197526931763, 'train/grad_norm_pre_clip': 7.730119705200195, 'timing/data': 0.013022150960750878, 'timing/model': 1.3270997040672228, 'learning_rate/qwen_vl_interface': 9.454281958130761e-06, 'learning_rate/action_model': 9.454281958130761e-05, 'epoch': 0.85}) +[2026-06-19 11:41:06,050][starVLA.training.train_starvla][INFO] - Step 856, Loss: {'train/loss': 0.31351810693740845, 'train/grad_norm_pre_clip': 5.802912712097168, 'timing/data': 0.012282127980142832, 'timing/model': 1.1447136409115046, 'learning_rate/qwen_vl_interface': 9.452863821711257e-06, 'learning_rate/action_model': 9.452863821711258e-05, 'epoch': 0.85}) +[2026-06-19 11:41:10,636][starVLA.training.train_starvla][INFO] - Step 857, Loss: {'train/loss': 0.3082379400730133, 'train/grad_norm_pre_clip': 9.4241304397583, 'timing/data': 0.012490550987422466, 'timing/model': 1.1536801550537348, 'learning_rate/qwen_vl_interface': 9.451443957653195e-06, 'learning_rate/action_model': 9.451443957653195e-05, 'epoch': 0.85}) +[2026-06-19 11:41:15,242][starVLA.training.train_starvla][INFO] - Step 858, Loss: {'train/loss': 0.28108832240104675, 'train/grad_norm_pre_clip': 5.817114353179932, 'timing/data': 0.012629642966203392, 'timing/model': 1.144832496996969, 'learning_rate/qwen_vl_interface': 9.450022366540221e-06, 'learning_rate/action_model': 9.45002236654022e-05, 'epoch': 0.86}) +[2026-06-19 11:41:19,862][starVLA.training.train_starvla][INFO] - Step 859, Loss: {'train/loss': 0.37257319688796997, 'train/grad_norm_pre_clip': 5.711081504821777, 'timing/data': 0.012735850061289966, 'timing/model': 1.1512978699756786, 'learning_rate/qwen_vl_interface': 9.448599048956702e-06, 'learning_rate/action_model': 9.448599048956702e-05, 'epoch': 0.86}) +[2026-06-19 11:41:24,452][starVLA.training.train_starvla][INFO] - Step 860, Loss: {'train/loss': 0.21391762793064117, 'train/grad_norm_pre_clip': 7.021445274353027, 'timing/data': 0.013124309945851564, 'timing/model': 1.1449497680878267, 'learning_rate/qwen_vl_interface': 9.447174005487708e-06, 'learning_rate/action_model': 9.447174005487708e-05, 'epoch': 0.86}) +[2026-06-19 11:41:29,050][starVLA.training.train_starvla][INFO] - Step 861, Loss: {'train/loss': 0.3455498516559601, 'train/grad_norm_pre_clip': 8.846268653869629, 'timing/data': 0.012282263953238726, 'timing/model': 1.3258115680655465, 'learning_rate/qwen_vl_interface': 9.445747236719022e-06, 'learning_rate/action_model': 9.445747236719022e-05, 'epoch': 0.86}) +[2026-06-19 11:41:33,630][starVLA.training.train_starvla][INFO] - Step 862, Loss: {'train/loss': 0.25767406821250916, 'train/grad_norm_pre_clip': 3.277963161468506, 'timing/data': 0.012603537063114345, 'timing/model': 1.3013536119833589, 'learning_rate/qwen_vl_interface': 9.444318743237135e-06, 'learning_rate/action_model': 9.444318743237134e-05, 'epoch': 0.86}) +[2026-06-19 11:41:38,050][starVLA.training.train_starvla][INFO] - Step 863, Loss: {'train/loss': 0.3397074043750763, 'train/grad_norm_pre_clip': 5.347101211547852, 'timing/data': 0.012341944966465235, 'timing/model': 1.1438837989699095, 'learning_rate/qwen_vl_interface': 9.442888525629244e-06, 'learning_rate/action_model': 9.442888525629245e-05, 'epoch': 0.86}) +[2026-06-19 11:41:42,636][starVLA.training.train_starvla][INFO] - Step 864, Loss: {'train/loss': 0.2479463815689087, 'train/grad_norm_pre_clip': 6.11525297164917, 'timing/data': 0.012777111027389765, 'timing/model': 1.1628830780973658, 'learning_rate/qwen_vl_interface': 9.441456584483261e-06, 'learning_rate/action_model': 9.441456584483261e-05, 'epoch': 0.86}) +[2026-06-19 11:41:47,222][starVLA.training.train_starvla][INFO] - Step 865, Loss: {'train/loss': 0.2941685914993286, 'train/grad_norm_pre_clip': 4.817224502563477, 'timing/data': 0.011987684993073344, 'timing/model': 1.1441080200020224, 'learning_rate/qwen_vl_interface': 9.440022920387801e-06, 'learning_rate/action_model': 9.440022920387801e-05, 'epoch': 0.86}) +[2026-06-19 11:41:51,799][starVLA.training.train_starvla][INFO] - Step 866, Loss: {'train/loss': 0.6200323104858398, 'train/grad_norm_pre_clip': 15.377999305725098, 'timing/data': 0.012855075066909194, 'timing/model': 1.1478408209513873, 'learning_rate/qwen_vl_interface': 9.438587533932191e-06, 'learning_rate/action_model': 9.43858753393219e-05, 'epoch': 0.86}) +[2026-06-19 11:41:56,412][starVLA.training.train_starvla][INFO] - Step 867, Loss: {'train/loss': 0.6944139003753662, 'train/grad_norm_pre_clip': 16.637826919555664, 'timing/data': 0.013050104025751352, 'timing/model': 1.1521048919530585, 'learning_rate/qwen_vl_interface': 9.43715042570646e-06, 'learning_rate/action_model': 9.43715042570646e-05, 'epoch': 0.86}) +[2026-06-19 11:42:01,017][starVLA.training.train_starvla][INFO] - Step 868, Loss: {'train/loss': 0.43730127811431885, 'train/grad_norm_pre_clip': 11.36699390411377, 'timing/data': 0.01322036306373775, 'timing/model': 1.3218909000279382, 'learning_rate/qwen_vl_interface': 9.435711596301353e-06, 'learning_rate/action_model': 9.435711596301352e-05, 'epoch': 0.87}) +[2026-06-19 11:42:05,458][starVLA.training.train_starvla][INFO] - Step 869, Loss: {'train/loss': 0.3367696702480316, 'train/grad_norm_pre_clip': 6.973851203918457, 'timing/data': 0.012631392921321094, 'timing/model': 1.1459386680508032, 'learning_rate/qwen_vl_interface': 9.434271046308317e-06, 'learning_rate/action_model': 9.434271046308317e-05, 'epoch': 0.87}) +[2026-06-19 11:42:10,053][starVLA.training.train_starvla][INFO] - Step 870, Loss: {'train/loss': 0.17923787236213684, 'train/grad_norm_pre_clip': 3.380012273788452, 'timing/data': 0.01316183595918119, 'timing/model': 1.1577334279427305, 'learning_rate/qwen_vl_interface': 9.432828776319507e-06, 'learning_rate/action_model': 9.432828776319505e-05, 'epoch': 0.87}) +[2026-06-19 11:42:14,658][starVLA.training.train_starvla][INFO] - Step 871, Loss: {'train/loss': 0.2939565181732178, 'train/grad_norm_pre_clip': 4.2832865715026855, 'timing/data': 0.013297999976202846, 'timing/model': 1.1451774799497798, 'learning_rate/qwen_vl_interface': 9.431384786927785e-06, 'learning_rate/action_model': 9.431384786927786e-05, 'epoch': 0.87}) +[2026-06-19 11:42:19,239][starVLA.training.train_starvla][INFO] - Step 872, Loss: {'train/loss': 0.3030959367752075, 'train/grad_norm_pre_clip': 4.523697853088379, 'timing/data': 0.0135347651084885, 'timing/model': 1.1427878479007632, 'learning_rate/qwen_vl_interface': 9.429939078726723e-06, 'learning_rate/action_model': 9.429939078726722e-05, 'epoch': 0.87}) +[2026-06-19 11:42:23,813][starVLA.training.train_starvla][INFO] - Step 873, Loss: {'train/loss': 0.1963510811328888, 'train/grad_norm_pre_clip': 3.7164528369903564, 'timing/data': 0.012707935064099729, 'timing/model': 1.1425996359903365, 'learning_rate/qwen_vl_interface': 9.428491652310598e-06, 'learning_rate/action_model': 9.428491652310597e-05, 'epoch': 0.87}) +[2026-06-19 11:42:28,392][starVLA.training.train_starvla][INFO] - Step 874, Loss: {'train/loss': 0.2733028531074524, 'train/grad_norm_pre_clip': 6.178459167480469, 'timing/data': 0.012660531094297767, 'timing/model': 1.3081642929464579, 'learning_rate/qwen_vl_interface': 9.427042508274386e-06, 'learning_rate/action_model': 9.427042508274386e-05, 'epoch': 0.87}) +[2026-06-19 11:42:32,981][starVLA.training.train_starvla][INFO] - Step 875, Loss: {'train/loss': 0.3355368375778198, 'train/grad_norm_pre_clip': 7.9900078773498535, 'timing/data': 0.012936036917380989, 'timing/model': 1.3130297359311953, 'learning_rate/qwen_vl_interface': 9.425591647213783e-06, 'learning_rate/action_model': 9.425591647213782e-05, 'epoch': 0.87}) +[2026-06-19 11:42:37,414][starVLA.training.train_starvla][INFO] - Step 876, Loss: {'train/loss': 0.39715293049812317, 'train/grad_norm_pre_clip': 6.699317932128906, 'timing/data': 0.012596126995049417, 'timing/model': 1.1527406439417973, 'learning_rate/qwen_vl_interface': 9.424139069725177e-06, 'learning_rate/action_model': 9.424139069725178e-05, 'epoch': 0.87}) +[2026-06-19 11:42:42,018][starVLA.training.train_starvla][INFO] - Step 877, Loss: {'train/loss': 0.3715449571609497, 'train/grad_norm_pre_clip': 9.222471237182617, 'timing/data': 0.013074886985123158, 'timing/model': 1.157679332070984, 'learning_rate/qwen_vl_interface': 9.422684776405673e-06, 'learning_rate/action_model': 9.422684776405672e-05, 'epoch': 0.87}) +[2026-06-19 11:42:46,644][starVLA.training.train_starvla][INFO] - Step 878, Loss: {'train/loss': 0.29918724298477173, 'train/grad_norm_pre_clip': 6.493985652923584, 'timing/data': 0.012824741075746715, 'timing/model': 1.1552617019042373, 'learning_rate/qwen_vl_interface': 9.421228767853072e-06, 'learning_rate/action_model': 9.421228767853072e-05, 'epoch': 0.88}) +[2026-06-19 11:42:51,191][starVLA.training.train_starvla][INFO] - Step 879, Loss: {'train/loss': 0.23827768862247467, 'train/grad_norm_pre_clip': 6.21129846572876, 'timing/data': 0.01242133998312056, 'timing/model': 1.1658743679290637, 'learning_rate/qwen_vl_interface': 9.419771044665887e-06, 'learning_rate/action_model': 9.419771044665887e-05, 'epoch': 0.88}) +[2026-06-19 11:42:55,770][starVLA.training.train_starvla][INFO] - Step 880, Loss: {'train/loss': 0.36560705304145813, 'train/grad_norm_pre_clip': 7.813811302185059, 'timing/data': 0.01230203197337687, 'timing/model': 1.1536492480663583, 'learning_rate/qwen_vl_interface': 9.41831160744333e-06, 'learning_rate/action_model': 9.41831160744333e-05, 'epoch': 0.88}) +[2026-06-19 11:43:00,367][starVLA.training.train_starvla][INFO] - Step 881, Loss: {'train/loss': 0.3856441080570221, 'train/grad_norm_pre_clip': 6.04591703414917, 'timing/data': 0.013358133030124009, 'timing/model': 1.3062844100641087, 'learning_rate/qwen_vl_interface': 9.416850456785322e-06, 'learning_rate/action_model': 9.416850456785321e-05, 'epoch': 0.88}) +[2026-06-19 11:43:04,819][starVLA.training.train_starvla][INFO] - Step 882, Loss: {'train/loss': 0.28312304615974426, 'train/grad_norm_pre_clip': 4.5191473960876465, 'timing/data': 0.013528661103919148, 'timing/model': 1.1455728539731354, 'learning_rate/qwen_vl_interface': 9.415387593292489e-06, 'learning_rate/action_model': 9.415387593292488e-05, 'epoch': 0.88}) +[2026-06-19 11:43:09,423][starVLA.training.train_starvla][INFO] - Step 883, Loss: {'train/loss': 0.30988267064094543, 'train/grad_norm_pre_clip': 5.8431806564331055, 'timing/data': 0.013039977988228202, 'timing/model': 1.1492115730652586, 'learning_rate/qwen_vl_interface': 9.413923017566155e-06, 'learning_rate/action_model': 9.413923017566155e-05, 'epoch': 0.88}) +[2026-06-19 11:43:14,033][starVLA.training.train_starvla][INFO] - Step 884, Loss: {'train/loss': 0.20253366231918335, 'train/grad_norm_pre_clip': 5.808979034423828, 'timing/data': 0.01263093005400151, 'timing/model': 1.1471723899012432, 'learning_rate/qwen_vl_interface': 9.412456730208352e-06, 'learning_rate/action_model': 9.412456730208352e-05, 'epoch': 0.88}) +[2026-06-19 11:43:18,609][starVLA.training.train_starvla][INFO] - Step 885, Loss: {'train/loss': 0.34460151195526123, 'train/grad_norm_pre_clip': 8.15282917022705, 'timing/data': 0.012452890048734844, 'timing/model': 1.141742496052757, 'learning_rate/qwen_vl_interface': 9.41098873182182e-06, 'learning_rate/action_model': 9.41098873182182e-05, 'epoch': 0.88}) +[2026-06-19 11:43:23,096][starVLA.training.train_starvla][INFO] - Step 886, Loss: {'train/loss': 0.3229754567146301, 'train/grad_norm_pre_clip': 4.1683149337768555, 'timing/data': 0.012195090996101499, 'timing/model': 1.1416922860080376, 'learning_rate/qwen_vl_interface': 9.409519023009993e-06, 'learning_rate/action_model': 9.409519023009993e-05, 'epoch': 0.88}) +[2026-06-19 11:43:27,611][starVLA.training.train_starvla][INFO] - Step 887, Loss: {'train/loss': 0.2007255107164383, 'train/grad_norm_pre_clip': 2.582615613937378, 'timing/data': 0.012963453074917197, 'timing/model': 1.245133331976831, 'learning_rate/qwen_vl_interface': 9.408047604377014e-06, 'learning_rate/action_model': 9.408047604377014e-05, 'epoch': 0.88}) +[2026-06-19 11:43:32,017][starVLA.training.train_starvla][INFO] - Step 888, Loss: {'train/loss': 0.29183146357536316, 'train/grad_norm_pre_clip': 5.119619369506836, 'timing/data': 0.013018773985095322, 'timing/model': 1.1416136459447443, 'learning_rate/qwen_vl_interface': 9.40657447652773e-06, 'learning_rate/action_model': 9.406574476527731e-05, 'epoch': 0.89}) +[2026-06-19 11:43:36,635][starVLA.training.train_starvla][INFO] - Step 889, Loss: {'train/loss': 0.1754029542207718, 'train/grad_norm_pre_clip': 2.9367640018463135, 'timing/data': 0.013057925971224904, 'timing/model': 1.153303807019256, 'learning_rate/qwen_vl_interface': 9.405099640067685e-06, 'learning_rate/action_model': 9.405099640067685e-05, 'epoch': 0.89}) +[2026-06-19 11:43:41,249][starVLA.training.train_starvla][INFO] - Step 890, Loss: {'train/loss': 0.3107758164405823, 'train/grad_norm_pre_clip': 3.2338385581970215, 'timing/data': 0.013203621958382428, 'timing/model': 1.1462299081031233, 'learning_rate/qwen_vl_interface': 9.40362309560313e-06, 'learning_rate/action_model': 9.40362309560313e-05, 'epoch': 0.89}) +[2026-06-19 11:43:45,860][starVLA.training.train_starvla][INFO] - Step 891, Loss: {'train/loss': 0.20148807764053345, 'train/grad_norm_pre_clip': 5.554123878479004, 'timing/data': 0.013087196042761207, 'timing/model': 1.1592736549209803, 'learning_rate/qwen_vl_interface': 9.402144843741016e-06, 'learning_rate/action_model': 9.402144843741017e-05, 'epoch': 0.89}) +[2026-06-19 11:43:50,477][starVLA.training.train_starvla][INFO] - Step 892, Loss: {'train/loss': 0.26942628622055054, 'train/grad_norm_pre_clip': 6.513749122619629, 'timing/data': 0.013658895040862262, 'timing/model': 1.3204088860657066, 'learning_rate/qwen_vl_interface': 9.400664885088999e-06, 'learning_rate/action_model': 9.400664885088999e-05, 'epoch': 0.89}) +[2026-06-19 11:43:54,982][starVLA.training.train_starvla][INFO] - Step 893, Loss: {'train/loss': 0.159601092338562, 'train/grad_norm_pre_clip': 4.148044586181641, 'timing/data': 0.01342831295914948, 'timing/model': 1.2334271129220724, 'learning_rate/qwen_vl_interface': 9.39918322025543e-06, 'learning_rate/action_model': 9.39918322025543e-05, 'epoch': 0.89}) +[2026-06-19 11:43:59,436][starVLA.training.train_starvla][INFO] - Step 894, Loss: {'train/loss': 0.19466723501682281, 'train/grad_norm_pre_clip': 4.869960308074951, 'timing/data': 0.013215980026870966, 'timing/model': 1.157788151060231, 'learning_rate/qwen_vl_interface': 9.397699849849368e-06, 'learning_rate/action_model': 9.397699849849367e-05, 'epoch': 0.89}) +[2026-06-19 11:44:04,019][starVLA.training.train_starvla][INFO] - Step 895, Loss: {'train/loss': 0.2321922481060028, 'train/grad_norm_pre_clip': 3.984870433807373, 'timing/data': 0.013030170928686857, 'timing/model': 1.1437893609981984, 'learning_rate/qwen_vl_interface': 9.39621477448057e-06, 'learning_rate/action_model': 9.39621477448057e-05, 'epoch': 0.89}) +[2026-06-19 11:44:08,642][starVLA.training.train_starvla][INFO] - Step 896, Loss: {'train/loss': 0.2704639136791229, 'train/grad_norm_pre_clip': 13.00230884552002, 'timing/data': 0.013406897080130875, 'timing/model': 1.1536968750879169, 'learning_rate/qwen_vl_interface': 9.394727994759495e-06, 'learning_rate/action_model': 9.394727994759495e-05, 'epoch': 0.89}) +[2026-06-19 11:44:13,273][starVLA.training.train_starvla][INFO] - Step 897, Loss: {'train/loss': 0.3560202419757843, 'train/grad_norm_pre_clip': 6.34168815612793, 'timing/data': 0.012451822985894978, 'timing/model': 1.1578961689956486, 'learning_rate/qwen_vl_interface': 9.3932395112973e-06, 'learning_rate/action_model': 9.3932395112973e-05, 'epoch': 0.89}) +[2026-06-19 11:44:17,888][starVLA.training.train_starvla][INFO] - Step 898, Loss: {'train/loss': 0.20672649145126343, 'train/grad_norm_pre_clip': 7.750948429107666, 'timing/data': 0.01342946698423475, 'timing/model': 1.3260406679473817, 'learning_rate/qwen_vl_interface': 9.391749324705846e-06, 'learning_rate/action_model': 9.391749324705846e-05, 'epoch': 0.9}) +[2026-06-19 11:44:22,512][starVLA.training.train_starvla][INFO] - Step 899, Loss: {'train/loss': 0.1499580442905426, 'train/grad_norm_pre_clip': 4.868319034576416, 'timing/data': 0.012736500008031726, 'timing/model': 1.3245414789998904, 'learning_rate/qwen_vl_interface': 9.390257435597692e-06, 'learning_rate/action_model': 9.390257435597692e-05, 'epoch': 0.9}) +[2026-06-19 11:44:26,990][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 900 +[2026-06-19 11:47:23,709][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 900 +[2026-06-19 11:47:23,709][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 900 +[2026-06-19 11:50:31,979][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 900 +[2026-06-19 11:50:31,980][starVLA.training.train_starvla][INFO] - Step 900, Loss: {'train/loss': 0.5137525796890259, 'train/grad_norm_pre_clip': 8.748329162597656, 'eval/loss': 0.4107455015182495, 'eval/latency_0/loss': 0.11839628964662552, 'eval/latency_1/loss': 0.16202861070632935, 'eval/latency_2/loss': 0.5166518688201904, 'eval/latency_3/loss': 0.4903792440891266, 'eval/latency_4/loss': 0.7505027651786804, 'eval/flappy/loss': 0.4107455015182495, 'eval/flappy/latency_0/loss': 0.11839628964662552, 'eval/flappy/latency_1/loss': 0.16202861070632935, 'eval/flappy/latency_2/loss': 0.5166518688201904, 'eval/flappy/latency_3/loss': 0.4903792440891266, 'eval/flappy/latency_4/loss': 0.7505027651786804, 'eval/flap_precision': 0.6351351351351351, 'eval/flap_recall': 0.6836363636363636, 'eval/flap_f1': 0.658493870402802, 'eval/flap_rate_model': 0.08222222222222222, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.658493870402802, 'eval/latency_0/flap_precision': 0.6351351351351351, 'eval/latency_0/flap_recall': 0.6836363636363636, 'eval/latency_0/flap_f1': 0.658493870402802, 'eval/latency_0/flap_rate_model': 0.08222222222222222, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.658493870402802, 'timing/data': 0.013380912016145885, 'timing/model': 1.1566772760124877, 'learning_rate/qwen_vl_interface': 9.3887638445861e-06, 'learning_rate/action_model': 9.3887638445861e-05, 'epoch': 0.9}) +[2026-06-19 11:50:31,981][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_900_state +[2026-06-19 11:50:31,983][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-19 11:50:35,365][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_900_state/model.safetensors +[2026-06-19 11:50:42,614][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_900_state/optimizer.bin +[2026-06-19 11:50:42,614][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_900_state/sampler.bin +[2026-06-19 11:50:42,615][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_900_state/sampler_1.bin +[2026-06-19 11:50:42,615][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_900_state/random_states_0.pkl +[2026-06-19 11:50:42,636][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 900 +[2026-06-19 11:50:42,636][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-19 11:50:42,645][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-19 11:50:47,136][starVLA.training.train_starvla][INFO] - Step 901, Loss: {'train/loss': 0.31752145290374756, 'train/grad_norm_pre_clip': 18.009540557861328, 'timing/data': 0.01312567398417741, 'timing/model': 1.1226930510019884, 'learning_rate/qwen_vl_interface': 9.387268552285024e-06, 'learning_rate/action_model': 9.387268552285025e-05, 'epoch': 0.9}) +[2026-06-19 11:50:51,653][starVLA.training.train_starvla][INFO] - Step 902, Loss: {'train/loss': 0.5216633677482605, 'train/grad_norm_pre_clip': 16.561492919921875, 'timing/data': 0.013476691907271743, 'timing/model': 1.2961316630244255, 'learning_rate/qwen_vl_interface': 9.385771559309125e-06, 'learning_rate/action_model': 9.385771559309125e-05, 'epoch': 0.9}) +[2026-06-19 11:50:55,997][starVLA.training.train_starvla][INFO] - Step 903, Loss: {'train/loss': 0.07037137448787689, 'train/grad_norm_pre_clip': 3.621959686279297, 'timing/data': 0.013822582899592817, 'timing/model': 1.126480848994106, 'learning_rate/qwen_vl_interface': 9.38427286627376e-06, 'learning_rate/action_model': 9.38427286627376e-05, 'epoch': 0.9}) +[2026-06-19 11:51:00,513][starVLA.training.train_starvla][INFO] - Step 904, Loss: {'train/loss': 0.3504580557346344, 'train/grad_norm_pre_clip': 9.212297439575195, 'timing/data': 0.013393820030614734, 'timing/model': 1.1293353099608794, 'learning_rate/qwen_vl_interface': 9.382772473794985e-06, 'learning_rate/action_model': 9.382772473794985e-05, 'epoch': 0.9}) +[2026-06-19 11:51:05,026][starVLA.training.train_starvla][INFO] - Step 905, Loss: {'train/loss': 0.22627218067646027, 'train/grad_norm_pre_clip': 7.626651763916016, 'timing/data': 0.012622350011952221, 'timing/model': 1.1277046219911426, 'learning_rate/qwen_vl_interface': 9.381270382489554e-06, 'learning_rate/action_model': 9.381270382489554e-05, 'epoch': 0.9}) +[2026-06-19 11:51:09,552][starVLA.training.train_starvla][INFO] - Step 906, Loss: {'train/loss': 0.3532117009162903, 'train/grad_norm_pre_clip': 11.024369239807129, 'timing/data': 0.012264330056495965, 'timing/model': 1.1288592179771513, 'learning_rate/qwen_vl_interface': 9.37976659297492e-06, 'learning_rate/action_model': 9.379766592974921e-05, 'epoch': 0.9}) +[2026-06-19 11:51:14,128][starVLA.training.train_starvla][INFO] - Step 907, Loss: {'train/loss': 0.23436681926250458, 'train/grad_norm_pre_clip': 5.091850280761719, 'timing/data': 0.012780492077581584, 'timing/model': 1.1495653659803793, 'learning_rate/qwen_vl_interface': 9.378261105869237e-06, 'learning_rate/action_model': 9.378261105869237e-05, 'epoch': 0.9}) +[2026-06-19 11:51:18,696][starVLA.training.train_starvla][INFO] - Step 908, Loss: {'train/loss': 0.3195575177669525, 'train/grad_norm_pre_clip': 7.570698261260986, 'timing/data': 0.012451773975044489, 'timing/model': 1.3171939140884206, 'learning_rate/qwen_vl_interface': 9.376753921791348e-06, 'learning_rate/action_model': 9.376753921791348e-05, 'epoch': 0.91}) +[2026-06-19 11:51:23,098][starVLA.training.train_starvla][INFO] - Step 909, Loss: {'train/loss': 0.2839201092720032, 'train/grad_norm_pre_clip': 5.753326892852783, 'timing/data': 0.013131154933944345, 'timing/model': 1.1361295890528709, 'learning_rate/qwen_vl_interface': 9.375245041360805e-06, 'learning_rate/action_model': 9.375245041360804e-05, 'epoch': 0.91}) +[2026-06-19 11:51:27,659][starVLA.training.train_starvla][INFO] - Step 910, Loss: {'train/loss': 0.15458126366138458, 'train/grad_norm_pre_clip': 4.228327751159668, 'timing/data': 0.012938647996634245, 'timing/model': 1.1438546089921147, 'learning_rate/qwen_vl_interface': 9.37373446519785e-06, 'learning_rate/action_model': 9.373734465197849e-05, 'epoch': 0.91}) +[2026-06-19 11:51:32,122][starVLA.training.train_starvla][INFO] - Step 911, Loss: {'train/loss': 0.12638358771800995, 'train/grad_norm_pre_clip': 2.8080689907073975, 'timing/data': 0.01295074203517288, 'timing/model': 1.1422922000056133, 'learning_rate/qwen_vl_interface': 9.372222193923421e-06, 'learning_rate/action_model': 9.372222193923421e-05, 'epoch': 0.91}) +[2026-06-19 11:51:36,628][starVLA.training.train_starvla][INFO] - Step 912, Loss: {'train/loss': 0.27446508407592773, 'train/grad_norm_pre_clip': 6.7363152503967285, 'timing/data': 0.012967428076080978, 'timing/model': 1.1379053560085595, 'learning_rate/qwen_vl_interface': 9.370708228159158e-06, 'learning_rate/action_model': 9.370708228159159e-05, 'epoch': 0.91}) +[2026-06-19 11:51:41,217][starVLA.training.train_starvla][INFO] - Step 913, Loss: {'train/loss': 0.30395492911338806, 'train/grad_norm_pre_clip': 5.701106548309326, 'timing/data': 0.01793080905918032, 'timing/model': 1.1431696800282225, 'learning_rate/qwen_vl_interface': 9.369192568527398e-06, 'learning_rate/action_model': 9.369192568527398e-05, 'epoch': 0.91}) +[2026-06-19 11:51:45,810][starVLA.training.train_starvla][INFO] - Step 914, Loss: {'train/loss': 0.46312999725341797, 'train/grad_norm_pre_clip': 8.397518157958984, 'timing/data': 0.013290161965414882, 'timing/model': 1.317313874955289, 'learning_rate/qwen_vl_interface': 9.367675215651168e-06, 'learning_rate/action_model': 9.367675215651168e-05, 'epoch': 0.91}) +[2026-06-19 11:51:50,236][starVLA.training.train_starvla][INFO] - Step 915, Loss: {'train/loss': 0.4684804677963257, 'train/grad_norm_pre_clip': 10.307670593261719, 'timing/data': 0.013250084011815488, 'timing/model': 1.1541830730857328, 'learning_rate/qwen_vl_interface': 9.366156170154194e-06, 'learning_rate/action_model': 9.366156170154194e-05, 'epoch': 0.91}) +[2026-06-19 11:51:54,816][starVLA.training.train_starvla][INFO] - Step 916, Loss: {'train/loss': 0.11085471510887146, 'train/grad_norm_pre_clip': 3.0934016704559326, 'timing/data': 0.012734993011690676, 'timing/model': 1.143878304050304, 'learning_rate/qwen_vl_interface': 9.3646354326609e-06, 'learning_rate/action_model': 9.364635432660899e-05, 'epoch': 0.91}) +[2026-06-19 11:51:59,317][starVLA.training.train_starvla][INFO] - Step 917, Loss: {'train/loss': 0.28305330872535706, 'train/grad_norm_pre_clip': 6.6499128341674805, 'timing/data': 0.012327383970841765, 'timing/model': 1.1526471079559997, 'learning_rate/qwen_vl_interface': 9.3631130037964e-06, 'learning_rate/action_model': 9.363113003796401e-05, 'epoch': 0.91}) +[2026-06-19 11:52:03,851][starVLA.training.train_starvla][INFO] - Step 918, Loss: {'train/loss': 0.2920761704444885, 'train/grad_norm_pre_clip': 5.586772441864014, 'timing/data': 0.013497110921889544, 'timing/model': 1.1516789880115539, 'learning_rate/qwen_vl_interface': 9.361588884186513e-06, 'learning_rate/action_model': 9.361588884186513e-05, 'epoch': 0.92}) +[2026-06-19 11:52:08,462][starVLA.training.train_starvla][INFO] - Step 919, Loss: {'train/loss': 0.3305317759513855, 'train/grad_norm_pre_clip': 9.157379150390625, 'timing/data': 0.01399783801753074, 'timing/model': 1.1434868390206248, 'learning_rate/qwen_vl_interface': 9.360063074457744e-06, 'learning_rate/action_model': 9.360063074457743e-05, 'epoch': 0.92}) +[2026-06-19 11:52:13,037][starVLA.training.train_starvla][INFO] - Step 920, Loss: {'train/loss': 0.27373093366622925, 'train/grad_norm_pre_clip': 8.932445526123047, 'timing/data': 0.013301413971930742, 'timing/model': 1.2994472349528223, 'learning_rate/qwen_vl_interface': 9.358535575237296e-06, 'learning_rate/action_model': 9.358535575237295e-05, 'epoch': 0.92}) +[2026-06-19 11:52:17,440][starVLA.training.train_starvla][INFO] - Step 921, Loss: {'train/loss': 0.42070162296295166, 'train/grad_norm_pre_clip': 11.32827091217041, 'timing/data': 0.012565315002575517, 'timing/model': 1.1406549439998344, 'learning_rate/qwen_vl_interface': 9.357006387153065e-06, 'learning_rate/action_model': 9.357006387153065e-05, 'epoch': 0.92}) +[2026-06-19 11:52:22,021][starVLA.training.train_starvla][INFO] - Step 922, Loss: {'train/loss': 0.32335618138313293, 'train/grad_norm_pre_clip': 5.320065498352051, 'timing/data': 0.012494849972426891, 'timing/model': 1.150252536055632, 'learning_rate/qwen_vl_interface': 9.355475510833643e-06, 'learning_rate/action_model': 9.355475510833643e-05, 'epoch': 0.92}) +[2026-06-19 11:52:26,513][starVLA.training.train_starvla][INFO] - Step 923, Loss: {'train/loss': 0.26372697949409485, 'train/grad_norm_pre_clip': 6.1359453201293945, 'timing/data': 0.013023743056692183, 'timing/model': 1.1468014470301569, 'learning_rate/qwen_vl_interface': 9.353942946908318e-06, 'learning_rate/action_model': 9.353942946908317e-05, 'epoch': 0.92}) +[2026-06-19 11:52:31,044][starVLA.training.train_starvla][INFO] - Step 924, Loss: {'train/loss': 0.3000146150588989, 'train/grad_norm_pre_clip': 8.2467041015625, 'timing/data': 0.013392981956712902, 'timing/model': 1.1431498769670725, 'learning_rate/qwen_vl_interface': 9.352408696007065e-06, 'learning_rate/action_model': 9.352408696007065e-05, 'epoch': 0.92}) +[2026-06-19 11:52:35,645][starVLA.training.train_starvla][INFO] - Step 925, Loss: {'train/loss': 0.2063048630952835, 'train/grad_norm_pre_clip': 5.21058988571167, 'timing/data': 0.012930636992678046, 'timing/model': 1.1579675409011543, 'learning_rate/qwen_vl_interface': 9.350872758760558e-06, 'learning_rate/action_model': 9.350872758760559e-05, 'epoch': 0.92}) +[2026-06-19 11:52:40,239][starVLA.training.train_starvla][INFO] - Step 926, Loss: {'train/loss': 0.22453123331069946, 'train/grad_norm_pre_clip': 5.358186721801758, 'timing/data': 0.013436773093417287, 'timing/model': 1.298412236967124, 'learning_rate/qwen_vl_interface': 9.349335135800164e-06, 'learning_rate/action_model': 9.349335135800163e-05, 'epoch': 0.92}) +[2026-06-19 11:52:44,784][starVLA.training.train_starvla][INFO] - Step 927, Loss: {'train/loss': 0.37706467509269714, 'train/grad_norm_pre_clip': 5.7238287925720215, 'timing/data': 0.012234873953275383, 'timing/model': 1.265655820025131, 'learning_rate/qwen_vl_interface': 9.347795827757942e-06, 'learning_rate/action_model': 9.347795827757941e-05, 'epoch': 0.92}) +[2026-06-19 11:52:49,218][starVLA.training.train_starvla][INFO] - Step 928, Loss: {'train/loss': 0.19823215901851654, 'train/grad_norm_pre_clip': 2.935415506362915, 'timing/data': 0.012910431949421763, 'timing/model': 1.1422014149138704, 'learning_rate/qwen_vl_interface': 9.346254835266643e-06, 'learning_rate/action_model': 9.346254835266642e-05, 'epoch': 0.93}) +[2026-06-19 11:52:53,788][starVLA.training.train_starvla][INFO] - Step 929, Loss: {'train/loss': 0.2152373045682907, 'train/grad_norm_pre_clip': 3.891474485397339, 'timing/data': 0.012446376029402018, 'timing/model': 1.1452836110256612, 'learning_rate/qwen_vl_interface': 9.344712158959709e-06, 'learning_rate/action_model': 9.344712158959709e-05, 'epoch': 0.93}) +[2026-06-19 11:52:58,379][starVLA.training.train_starvla][INFO] - Step 930, Loss: {'train/loss': 0.3098149001598358, 'train/grad_norm_pre_clip': 4.592563152313232, 'timing/data': 0.019389957073144615, 'timing/model': 1.1453667300520465, 'learning_rate/qwen_vl_interface': 9.343167799471278e-06, 'learning_rate/action_model': 9.343167799471277e-05, 'epoch': 0.93}) +[2026-06-19 11:53:02,955][starVLA.training.train_starvla][INFO] - Step 931, Loss: {'train/loss': 0.38660842180252075, 'train/grad_norm_pre_clip': 7.3455586433410645, 'timing/data': 0.0136034149909392, 'timing/model': 1.1431892300024629, 'learning_rate/qwen_vl_interface': 9.341621757436177e-06, 'learning_rate/action_model': 9.341621757436177e-05, 'epoch': 0.93}) +[2026-06-19 11:53:07,445][starVLA.training.train_starvla][INFO] - Step 932, Loss: {'train/loss': 0.39335840940475464, 'train/grad_norm_pre_clip': 6.963396072387695, 'timing/data': 0.014267702936194837, 'timing/model': 1.1455722160171717, 'learning_rate/qwen_vl_interface': 9.340074033489924e-06, 'learning_rate/action_model': 9.340074033489923e-05, 'epoch': 0.93}) +[2026-06-19 11:53:11,982][starVLA.training.train_starvla][INFO] - Step 933, Loss: {'train/loss': 0.3742779791355133, 'train/grad_norm_pre_clip': 9.44147777557373, 'timing/data': 0.013483332004398108, 'timing/model': 1.2673260769806802, 'learning_rate/qwen_vl_interface': 9.338524628268734e-06, 'learning_rate/action_model': 9.338524628268733e-05, 'epoch': 0.93}) +[2026-06-19 11:53:16,401][starVLA.training.train_starvla][INFO] - Step 934, Loss: {'train/loss': 0.2087012678384781, 'train/grad_norm_pre_clip': 4.217486381530762, 'timing/data': 0.012851289007812738, 'timing/model': 1.1423576360102743, 'learning_rate/qwen_vl_interface': 9.336973542409505e-06, 'learning_rate/action_model': 9.336973542409504e-05, 'epoch': 0.93}) +[2026-06-19 11:53:20,974][starVLA.training.train_starvla][INFO] - Step 935, Loss: {'train/loss': 0.3875458836555481, 'train/grad_norm_pre_clip': 8.858777046203613, 'timing/data': 0.014367014053277671, 'timing/model': 1.148012873949483, 'learning_rate/qwen_vl_interface': 9.33542077654983e-06, 'learning_rate/action_model': 9.33542077654983e-05, 'epoch': 0.93}) +[2026-06-19 11:53:25,556][starVLA.training.train_starvla][INFO] - Step 936, Loss: {'train/loss': 0.2665683329105377, 'train/grad_norm_pre_clip': 9.324727058410645, 'timing/data': 0.01336419500876218, 'timing/model': 1.1446609359700233, 'learning_rate/qwen_vl_interface': 9.333866331327995e-06, 'learning_rate/action_model': 9.333866331327995e-05, 'epoch': 0.93}) +[2026-06-19 11:53:30,132][starVLA.training.train_starvla][INFO] - Step 937, Loss: {'train/loss': 0.37145838141441345, 'train/grad_norm_pre_clip': 8.791793823242188, 'timing/data': 0.013013942050747573, 'timing/model': 1.1430488820187747, 'learning_rate/qwen_vl_interface': 9.332310207382972e-06, 'learning_rate/action_model': 9.332310207382972e-05, 'epoch': 0.93}) +[2026-06-19 11:53:34,705][starVLA.training.train_starvla][INFO] - Step 938, Loss: {'train/loss': 0.5047566890716553, 'train/grad_norm_pre_clip': 13.293389320373535, 'timing/data': 0.012765911058522761, 'timing/model': 1.296883168979548, 'learning_rate/qwen_vl_interface': 9.330752405354426e-06, 'learning_rate/action_model': 9.330752405354426e-05, 'epoch': 0.94}) +[2026-06-19 11:53:39,331][starVLA.training.train_starvla][INFO] - Step 939, Loss: {'train/loss': 0.4776652455329895, 'train/grad_norm_pre_clip': 13.799776077270508, 'timing/data': 0.013293678988702595, 'timing/model': 1.3274299260228872, 'learning_rate/qwen_vl_interface': 9.32919292588271e-06, 'learning_rate/action_model': 9.32919292588271e-05, 'epoch': 0.94}) +[2026-06-19 11:53:43,753][starVLA.training.train_starvla][INFO] - Step 940, Loss: {'train/loss': 0.3362102210521698, 'train/grad_norm_pre_clip': 5.743569850921631, 'timing/data': 0.012976790079846978, 'timing/model': 1.141430942923762, 'learning_rate/qwen_vl_interface': 9.327631769608864e-06, 'learning_rate/action_model': 9.327631769608864e-05, 'epoch': 0.94}) +[2026-06-19 11:53:48,349][starVLA.training.train_starvla][INFO] - Step 941, Loss: {'train/loss': 0.19080345332622528, 'train/grad_norm_pre_clip': 5.33695125579834, 'timing/data': 0.012318263063207269, 'timing/model': 1.1439007279695943, 'learning_rate/qwen_vl_interface': 9.326068937174627e-06, 'learning_rate/action_model': 9.326068937174626e-05, 'epoch': 0.94}) +[2026-06-19 11:53:52,920][starVLA.training.train_starvla][INFO] - Step 942, Loss: {'train/loss': 0.28446516394615173, 'train/grad_norm_pre_clip': 5.027186393737793, 'timing/data': 0.011726335971616209, 'timing/model': 1.1407159400405362, 'learning_rate/qwen_vl_interface': 9.324504429222412e-06, 'learning_rate/action_model': 9.324504429222412e-05, 'epoch': 0.94}) +[2026-06-19 11:53:57,399][starVLA.training.train_starvla][INFO] - Step 943, Loss: {'train/loss': 0.3035092353820801, 'train/grad_norm_pre_clip': 6.544902801513672, 'timing/data': 0.011747522978112102, 'timing/model': 1.1440075759310275, 'learning_rate/qwen_vl_interface': 9.322938246395336e-06, 'learning_rate/action_model': 9.322938246395337e-05, 'epoch': 0.94}) +[2026-06-19 11:54:01,959][starVLA.training.train_starvla][INFO] - Step 944, Loss: {'train/loss': 0.3101310133934021, 'train/grad_norm_pre_clip': 6.288644313812256, 'timing/data': 0.01308691909071058, 'timing/model': 1.1434009929653257, 'learning_rate/qwen_vl_interface': 9.321370389337196e-06, 'learning_rate/action_model': 9.321370389337196e-05, 'epoch': 0.94}) +[2026-06-19 11:54:06,569][starVLA.training.train_starvla][INFO] - Step 945, Loss: {'train/loss': 0.3481697738170624, 'train/grad_norm_pre_clip': 6.893209934234619, 'timing/data': 0.012102733016945422, 'timing/model': 1.3257032969268039, 'learning_rate/qwen_vl_interface': 9.319800858692478e-06, 'learning_rate/action_model': 9.319800858692476e-05, 'epoch': 0.94}) +[2026-06-19 11:54:11,009][starVLA.training.train_starvla][INFO] - Step 946, Loss: {'train/loss': 0.2887723445892334, 'train/grad_norm_pre_clip': 3.9320809841156006, 'timing/data': 0.01274498994462192, 'timing/model': 1.16266266501043, 'learning_rate/qwen_vl_interface': 9.318229655106357e-06, 'learning_rate/action_model': 9.318229655106356e-05, 'epoch': 0.94}) +[2026-06-19 11:54:15,638][starVLA.training.train_starvla][INFO] - Step 947, Loss: {'train/loss': 0.19571535289287567, 'train/grad_norm_pre_clip': 2.786149024963379, 'timing/data': 0.01264864206314087, 'timing/model': 1.171278837020509, 'learning_rate/qwen_vl_interface': 9.316656779224693e-06, 'learning_rate/action_model': 9.316656779224693e-05, 'epoch': 0.94}) +[2026-06-19 11:54:20,229][starVLA.training.train_starvla][INFO] - Step 948, Loss: {'train/loss': 0.4181257486343384, 'train/grad_norm_pre_clip': 13.021767616271973, 'timing/data': 0.014013099018484354, 'timing/model': 1.1445479490794241, 'learning_rate/qwen_vl_interface': 9.31508223169404e-06, 'learning_rate/action_model': 9.31508223169404e-05, 'epoch': 0.95}) +[2026-06-19 11:54:24,825][starVLA.training.train_starvla][INFO] - Step 949, Loss: {'train/loss': 0.40839293599128723, 'train/grad_norm_pre_clip': 12.707304954528809, 'timing/data': 0.013243664056062698, 'timing/model': 1.1406708150170743, 'learning_rate/qwen_vl_interface': 9.313506013161633e-06, 'learning_rate/action_model': 9.313506013161633e-05, 'epoch': 0.95}) +[2026-06-19 11:54:29,337][starVLA.training.train_starvla][INFO] - Step 950, Loss: {'train/loss': 0.27504339814186096, 'train/grad_norm_pre_clip': 4.116175174713135, 'timing/data': 0.013200310058891773, 'timing/model': 1.148640825995244, 'learning_rate/qwen_vl_interface': 9.311928124275396e-06, 'learning_rate/action_model': 9.311928124275395e-05, 'epoch': 0.95}) +[2026-06-19 11:54:33,880][starVLA.training.train_starvla][INFO] - Step 951, Loss: {'train/loss': 0.27091193199157715, 'train/grad_norm_pre_clip': 2.800583600997925, 'timing/data': 0.014262170996516943, 'timing/model': 1.2509227830450982, 'learning_rate/qwen_vl_interface': 9.310348565683938e-06, 'learning_rate/action_model': 9.310348565683939e-05, 'epoch': 0.95}) +[2026-06-19 11:54:38,303][starVLA.training.train_starvla][INFO] - Step 952, Loss: {'train/loss': 0.44141799211502075, 'train/grad_norm_pre_clip': 11.654828071594238, 'timing/data': 0.012653089012019336, 'timing/model': 1.1441550679737702, 'learning_rate/qwen_vl_interface': 9.30876733803656e-06, 'learning_rate/action_model': 9.308767338036559e-05, 'epoch': 0.95}) +[2026-06-19 11:54:42,906][starVLA.training.train_starvla][INFO] - Step 953, Loss: {'train/loss': 0.1915820986032486, 'train/grad_norm_pre_clip': 4.153397083282471, 'timing/data': 0.013101937947794795, 'timing/model': 1.1481904219835997, 'learning_rate/qwen_vl_interface': 9.30718444198324e-06, 'learning_rate/action_model': 9.307184441983239e-05, 'epoch': 0.95}) +[2026-06-19 11:54:47,527][starVLA.training.train_starvla][INFO] - Step 954, Loss: {'train/loss': 0.4250679314136505, 'train/grad_norm_pre_clip': 8.320444107055664, 'timing/data': 0.01308142498601228, 'timing/model': 1.1491806919220835, 'learning_rate/qwen_vl_interface': 9.30559987817465e-06, 'learning_rate/action_model': 9.30559987817465e-05, 'epoch': 0.95}) +[2026-06-19 11:54:52,152][starVLA.training.train_starvla][INFO] - Step 955, Loss: {'train/loss': 0.38627150654792786, 'train/grad_norm_pre_clip': 4.92586088180542, 'timing/data': 0.013433326967060566, 'timing/model': 1.1657224000664428, 'learning_rate/qwen_vl_interface': 9.304013647262143e-06, 'learning_rate/action_model': 9.304013647262143e-05, 'epoch': 0.95}) +[2026-06-19 11:54:56,744][starVLA.training.train_starvla][INFO] - Step 956, Loss: {'train/loss': 0.3923930525779724, 'train/grad_norm_pre_clip': 6.813106536865234, 'timing/data': 0.012427204055711627, 'timing/model': 1.298202590085566, 'learning_rate/qwen_vl_interface': 9.302425749897761e-06, 'learning_rate/action_model': 9.302425749897761e-05, 'epoch': 0.95}) +[2026-06-19 11:55:01,175][starVLA.training.train_starvla][INFO] - Step 957, Loss: {'train/loss': 0.4434385895729065, 'train/grad_norm_pre_clip': 6.835987567901611, 'timing/data': 0.014234155998565257, 'timing/model': 1.1493568499572575, 'learning_rate/qwen_vl_interface': 9.300836186734225e-06, 'learning_rate/action_model': 9.300836186734226e-05, 'epoch': 0.95}) +[2026-06-19 11:55:05,780][starVLA.training.train_starvla][INFO] - Step 958, Loss: {'train/loss': 0.18439200520515442, 'train/grad_norm_pre_clip': 2.5351219177246094, 'timing/data': 0.012658526073209941, 'timing/model': 1.1642122839111835, 'learning_rate/qwen_vl_interface': 9.299244958424947e-06, 'learning_rate/action_model': 9.299244958424946e-05, 'epoch': 0.96}) +[2026-06-19 11:55:10,389][starVLA.training.train_starvla][INFO] - Step 959, Loss: {'train/loss': 0.2886407971382141, 'train/grad_norm_pre_clip': 5.569453716278076, 'timing/data': 0.013295976095832884, 'timing/model': 1.1572597379563376, 'learning_rate/qwen_vl_interface': 9.29765206562402e-06, 'learning_rate/action_model': 9.297652065624018e-05, 'epoch': 0.96}) +[2026-06-19 11:55:14,978][starVLA.training.train_starvla][INFO] - Step 960, Loss: {'train/loss': 0.33084917068481445, 'train/grad_norm_pre_clip': 4.698400497436523, 'timing/data': 0.012405380024574697, 'timing/model': 1.1451011369936168, 'learning_rate/qwen_vl_interface': 9.296057508986221e-06, 'learning_rate/action_model': 9.296057508986221e-05, 'epoch': 0.96}) +[2026-06-19 11:55:19,568][starVLA.training.train_starvla][INFO] - Step 961, Loss: {'train/loss': 0.3066276013851166, 'train/grad_norm_pre_clip': 7.063083648681641, 'timing/data': 0.013274195953272283, 'timing/model': 1.1548737960401922, 'learning_rate/qwen_vl_interface': 9.294461289167013e-06, 'learning_rate/action_model': 9.294461289167014e-05, 'epoch': 0.96}) +[2026-06-19 11:55:24,198][starVLA.training.train_starvla][INFO] - Step 962, Loss: {'train/loss': 0.28520792722702026, 'train/grad_norm_pre_clip': 4.4252610206604, 'timing/data': 0.013042852049693465, 'timing/model': 1.3114753270056099, 'learning_rate/qwen_vl_interface': 9.292863406822543e-06, 'learning_rate/action_model': 9.292863406822542e-05, 'epoch': 0.96}) +[2026-06-19 11:55:28,750][starVLA.training.train_starvla][INFO] - Step 963, Loss: {'train/loss': 0.1789480596780777, 'train/grad_norm_pre_clip': 2.9406721591949463, 'timing/data': 0.013526560040190816, 'timing/model': 1.2667969139292836, 'learning_rate/qwen_vl_interface': 9.291263862609638e-06, 'learning_rate/action_model': 9.291263862609638e-05, 'epoch': 0.96}) +[2026-06-19 11:55:33,188][starVLA.training.train_starvla][INFO] - Step 964, Loss: {'train/loss': 0.27862751483917236, 'train/grad_norm_pre_clip': 4.801050186157227, 'timing/data': 0.012207421008497477, 'timing/model': 1.1504081970779225, 'learning_rate/qwen_vl_interface': 9.28966265718581e-06, 'learning_rate/action_model': 9.28966265718581e-05, 'epoch': 0.96}) +[2026-06-19 11:55:37,802][starVLA.training.train_starvla][INFO] - Step 965, Loss: {'train/loss': 0.3632587790489197, 'train/grad_norm_pre_clip': 12.502066612243652, 'timing/data': 0.01389365503564477, 'timing/model': 1.1566454810090363, 'learning_rate/qwen_vl_interface': 9.288059791209256e-06, 'learning_rate/action_model': 9.288059791209256e-05, 'epoch': 0.96}) +[2026-06-19 11:55:42,383][starVLA.training.train_starvla][INFO] - Step 966, Loss: {'train/loss': 0.2555246353149414, 'train/grad_norm_pre_clip': 4.915289878845215, 'timing/data': 0.012142930994741619, 'timing/model': 1.1433000030228868, 'learning_rate/qwen_vl_interface': 9.286455265338853e-06, 'learning_rate/action_model': 9.286455265338853e-05, 'epoch': 0.96}) +[2026-06-19 11:55:46,976][starVLA.training.train_starvla][INFO] - Step 967, Loss: {'train/loss': 0.24079665541648865, 'train/grad_norm_pre_clip': 4.857773303985596, 'timing/data': 0.01358966703992337, 'timing/model': 1.1415827460587025, 'learning_rate/qwen_vl_interface': 9.284849080234161e-06, 'learning_rate/action_model': 9.28484908023416e-05, 'epoch': 0.96}) +[2026-06-19 11:55:51,564][starVLA.training.train_starvla][INFO] - Step 968, Loss: {'train/loss': 0.21576713025569916, 'train/grad_norm_pre_clip': 5.882718563079834, 'timing/data': 0.012351458077318966, 'timing/model': 1.1435852379072458, 'learning_rate/qwen_vl_interface': 9.283241236555421e-06, 'learning_rate/action_model': 9.283241236555421e-05, 'epoch': 0.97}) +[2026-06-19 11:55:56,181][starVLA.training.train_starvla][INFO] - Step 969, Loss: {'train/loss': 0.3252691626548767, 'train/grad_norm_pre_clip': 4.907985210418701, 'timing/data': 0.012565059936605394, 'timing/model': 1.3175284969620407, 'learning_rate/qwen_vl_interface': 9.281631734963557e-06, 'learning_rate/action_model': 9.281631734963557e-05, 'epoch': 0.97}) +[2026-06-19 11:56:00,637][starVLA.training.train_starvla][INFO] - Step 970, Loss: {'train/loss': 0.20752640068531036, 'train/grad_norm_pre_clip': 3.3200793266296387, 'timing/data': 0.013003212050534785, 'timing/model': 1.1525920819258317, 'learning_rate/qwen_vl_interface': 9.280020576120176e-06, 'learning_rate/action_model': 9.280020576120176e-05, 'epoch': 0.97}) +[2026-06-19 11:56:05,270][starVLA.training.train_starvla][INFO] - Step 971, Loss: {'train/loss': 0.23670370876789093, 'train/grad_norm_pre_clip': 4.668997287750244, 'timing/data': 0.013656085007824004, 'timing/model': 1.155407618964091, 'learning_rate/qwen_vl_interface': 9.278407760687562e-06, 'learning_rate/action_model': 9.278407760687562e-05, 'epoch': 0.97}) +[2026-06-19 11:56:09,906][starVLA.training.train_starvla][INFO] - Step 972, Loss: {'train/loss': 0.35538557171821594, 'train/grad_norm_pre_clip': 11.227859497070312, 'timing/data': 0.012309545069001615, 'timing/model': 1.1528965489706025, 'learning_rate/qwen_vl_interface': 9.276793289328686e-06, 'learning_rate/action_model': 9.276793289328685e-05, 'epoch': 0.97}) +[2026-06-19 11:56:14,525][starVLA.training.train_starvla][INFO] - Step 973, Loss: {'train/loss': 0.26297712326049805, 'train/grad_norm_pre_clip': 6.679120063781738, 'timing/data': 0.012601992930285633, 'timing/model': 1.1570774730062112, 'learning_rate/qwen_vl_interface': 9.27517716270719e-06, 'learning_rate/action_model': 9.275177162707191e-05, 'epoch': 0.97}) +[2026-06-19 11:56:19,148][starVLA.training.train_starvla][INFO] - Step 974, Loss: {'train/loss': 0.37246206402778625, 'train/grad_norm_pre_clip': 11.123019218444824, 'timing/data': 0.012970360112376511, 'timing/model': 1.1556785519933328, 'learning_rate/qwen_vl_interface': 9.273559381487409e-06, 'learning_rate/action_model': 9.273559381487409e-05, 'epoch': 0.97}) +[2026-06-19 11:56:23,796][starVLA.training.train_starvla][INFO] - Step 975, Loss: {'train/loss': 0.16916227340698242, 'train/grad_norm_pre_clip': 3.329163074493408, 'timing/data': 0.013479099958203733, 'timing/model': 1.327873417059891, 'learning_rate/qwen_vl_interface': 9.27193994633435e-06, 'learning_rate/action_model': 9.27193994633435e-05, 'epoch': 0.97}) +[2026-06-19 11:56:28,248][starVLA.training.train_starvla][INFO] - Step 976, Loss: {'train/loss': 0.22347652912139893, 'train/grad_norm_pre_clip': 5.753163814544678, 'timing/data': 0.012891285005025566, 'timing/model': 1.1427660620538518, 'learning_rate/qwen_vl_interface': 9.2703188579137e-06, 'learning_rate/action_model': 9.2703188579137e-05, 'epoch': 0.97}) +[2026-06-19 11:56:32,876][starVLA.training.train_starvla][INFO] - Step 977, Loss: {'train/loss': 0.39758825302124023, 'train/grad_norm_pre_clip': 7.560525894165039, 'timing/data': 0.013369896914809942, 'timing/model': 1.1468032030388713, 'learning_rate/qwen_vl_interface': 9.26869611689183e-06, 'learning_rate/action_model': 9.26869611689183e-05, 'epoch': 0.97}) +[2026-06-19 11:56:37,397][starVLA.training.train_starvla][INFO] - Step 978, Loss: {'train/loss': 0.2971368432044983, 'train/grad_norm_pre_clip': 5.055465221405029, 'timing/data': 0.013112994958646595, 'timing/model': 1.151571199996397, 'learning_rate/qwen_vl_interface': 9.267071723935784e-06, 'learning_rate/action_model': 9.267071723935784e-05, 'epoch': 0.98}) +[2026-06-19 11:56:41,898][starVLA.training.train_starvla][INFO] - Step 979, Loss: {'train/loss': 0.332210898399353, 'train/grad_norm_pre_clip': 9.462103843688965, 'timing/data': 0.013040095916949213, 'timing/model': 1.1553930849768221, 'learning_rate/qwen_vl_interface': 9.26544567971329e-06, 'learning_rate/action_model': 9.26544567971329e-05, 'epoch': 0.98}) +[2026-06-19 11:56:46,480][starVLA.training.train_starvla][INFO] - Step 980, Loss: {'train/loss': 0.3259064853191376, 'train/grad_norm_pre_clip': 9.039397239685059, 'timing/data': 0.013107546023093164, 'timing/model': 1.1455365270376205, 'learning_rate/qwen_vl_interface': 9.263817984892756e-06, 'learning_rate/action_model': 9.263817984892757e-05, 'epoch': 0.98}) +[2026-06-19 11:56:51,074][starVLA.training.train_starvla][INFO] - Step 981, Loss: {'train/loss': 0.17835675179958344, 'train/grad_norm_pre_clip': 3.1592042446136475, 'timing/data': 0.01224412804003805, 'timing/model': 1.3043160700472072, 'learning_rate/qwen_vl_interface': 9.262188640143262e-06, 'learning_rate/action_model': 9.262188640143262e-05, 'epoch': 0.98}) +[2026-06-19 11:56:55,489][starVLA.training.train_starvla][INFO] - Step 982, Loss: {'train/loss': 0.10798050463199615, 'train/grad_norm_pre_clip': 2.828303098678589, 'timing/data': 0.013383999932557344, 'timing/model': 1.1440481380559504, 'learning_rate/qwen_vl_interface': 9.260557646134572e-06, 'learning_rate/action_model': 9.260557646134572e-05, 'epoch': 0.98}) +[2026-06-19 11:57:00,100][starVLA.training.train_starvla][INFO] - Step 983, Loss: {'train/loss': 0.13800384104251862, 'train/grad_norm_pre_clip': 3.8846213817596436, 'timing/data': 0.012632564059458673, 'timing/model': 1.1479701929492876, 'learning_rate/qwen_vl_interface': 9.258925003537127e-06, 'learning_rate/action_model': 9.258925003537126e-05, 'epoch': 0.98}) +[2026-06-19 11:57:04,714][starVLA.training.train_starvla][INFO] - Step 984, Loss: {'train/loss': 0.33161601424217224, 'train/grad_norm_pre_clip': 6.5278849601745605, 'timing/data': 0.013284301036037505, 'timing/model': 1.1450364949414507, 'learning_rate/qwen_vl_interface': 9.257290713022041e-06, 'learning_rate/action_model': 9.257290713022041e-05, 'epoch': 0.98}) +[2026-06-19 11:57:09,263][starVLA.training.train_starvla][INFO] - Step 985, Loss: {'train/loss': 0.344654381275177, 'train/grad_norm_pre_clip': 3.9062440395355225, 'timing/data': 0.012613408965989947, 'timing/model': 1.1464432149659842, 'learning_rate/qwen_vl_interface': 9.255654775261112e-06, 'learning_rate/action_model': 9.255654775261112e-05, 'epoch': 0.98}) +[2026-06-19 11:57:13,847][starVLA.training.train_starvla][INFO] - Step 986, Loss: {'train/loss': 0.09369299560785294, 'train/grad_norm_pre_clip': 4.8116865158081055, 'timing/data': 0.012222509016282856, 'timing/model': 1.1415345300920308, 'learning_rate/qwen_vl_interface': 9.254017190926813e-06, 'learning_rate/action_model': 9.254017190926813e-05, 'epoch': 0.98}) +[2026-06-19 11:57:18,423][starVLA.training.train_starvla][INFO] - Step 987, Loss: {'train/loss': 0.19892804324626923, 'train/grad_norm_pre_clip': 7.197004795074463, 'timing/data': 0.01287439395673573, 'timing/model': 1.305421978002414, 'learning_rate/qwen_vl_interface': 9.25237796069229e-06, 'learning_rate/action_model': 9.25237796069229e-05, 'epoch': 0.98}) +[2026-06-19 11:57:22,843][starVLA.training.train_starvla][INFO] - Step 988, Loss: {'train/loss': 0.2280675321817398, 'train/grad_norm_pre_clip': 6.086261749267578, 'timing/data': 0.012549288920126855, 'timing/model': 1.1437503939960152, 'learning_rate/qwen_vl_interface': 9.250737085231371e-06, 'learning_rate/action_model': 9.250737085231372e-05, 'epoch': 0.99}) +[2026-06-19 11:57:27,425][starVLA.training.train_starvla][INFO] - Step 989, Loss: {'train/loss': 0.4304264485836029, 'train/grad_norm_pre_clip': 6.965105056762695, 'timing/data': 0.012985819019377232, 'timing/model': 1.1581841339357197, 'learning_rate/qwen_vl_interface': 9.24909456521856e-06, 'learning_rate/action_model': 9.249094565218559e-05, 'epoch': 0.99}) +[2026-06-19 11:57:32,052][starVLA.training.train_starvla][INFO] - Step 990, Loss: {'train/loss': 0.11781954020261765, 'train/grad_norm_pre_clip': 6.438131809234619, 'timing/data': 0.012597630033269525, 'timing/model': 1.155588921974413, 'learning_rate/qwen_vl_interface': 9.247450401329027e-06, 'learning_rate/action_model': 9.247450401329027e-05, 'epoch': 0.99}) +[2026-06-19 11:57:36,681][starVLA.training.train_starvla][INFO] - Step 991, Loss: {'train/loss': 0.156550332903862, 'train/grad_norm_pre_clip': 9.839831352233887, 'timing/data': 0.012652267003431916, 'timing/model': 1.1559623710345477, 'learning_rate/qwen_vl_interface': 9.245804594238637e-06, 'learning_rate/action_model': 9.245804594238636e-05, 'epoch': 0.99}) +[2026-06-19 11:57:41,267][starVLA.training.train_starvla][INFO] - Step 992, Loss: {'train/loss': 0.3109450936317444, 'train/grad_norm_pre_clip': 7.203096389770508, 'timing/data': 0.012530082021839917, 'timing/model': 1.1411189929349348, 'learning_rate/qwen_vl_interface': 9.244157144623909e-06, 'learning_rate/action_model': 9.244157144623909e-05, 'epoch': 0.99}) +[2026-06-19 11:57:45,867][starVLA.training.train_starvla][INFO] - Step 993, Loss: {'train/loss': 0.2814774811267853, 'train/grad_norm_pre_clip': 9.795656204223633, 'timing/data': 0.013296535005792975, 'timing/model': 1.3176500279223546, 'learning_rate/qwen_vl_interface': 9.242508053162051e-06, 'learning_rate/action_model': 9.242508053162051e-05, 'epoch': 0.99}) +[2026-06-19 11:57:50,296][starVLA.training.train_starvla][INFO] - Step 994, Loss: {'train/loss': 0.48156335949897766, 'train/grad_norm_pre_clip': 9.732427597045898, 'timing/data': 0.013204472023062408, 'timing/model': 1.1492194070015103, 'learning_rate/qwen_vl_interface': 9.240857320530944e-06, 'learning_rate/action_model': 9.240857320530943e-05, 'epoch': 0.99}) +[2026-06-19 11:57:54,900][starVLA.training.train_starvla][INFO] - Step 995, Loss: {'train/loss': 0.4062774181365967, 'train/grad_norm_pre_clip': 7.649470329284668, 'timing/data': 0.014442623010836542, 'timing/model': 1.142720574978739, 'learning_rate/qwen_vl_interface': 9.23920494740914e-06, 'learning_rate/action_model': 9.23920494740914e-05, 'epoch': 0.99}) +[2026-06-19 11:57:59,398][starVLA.training.train_starvla][INFO] - Step 996, Loss: {'train/loss': 0.1565568745136261, 'train/grad_norm_pre_clip': 7.837137222290039, 'timing/data': 0.012609828962013125, 'timing/model': 1.1403405650053173, 'learning_rate/qwen_vl_interface': 9.237550934475868e-06, 'learning_rate/action_model': 9.237550934475867e-05, 'epoch': 0.99}) +[2026-06-19 11:58:03,931][starVLA.training.train_starvla][INFO] - Step 997, Loss: {'train/loss': 0.2875710129737854, 'train/grad_norm_pre_clip': 7.079555034637451, 'timing/data': 0.012834564899094403, 'timing/model': 1.1557756758993492, 'learning_rate/qwen_vl_interface': 9.235895282411029e-06, 'learning_rate/action_model': 9.235895282411029e-05, 'epoch': 0.99}) +[2026-06-19 11:58:08,529][starVLA.training.train_starvla][INFO] - Step 998, Loss: {'train/loss': 0.38468459248542786, 'train/grad_norm_pre_clip': 8.666391372680664, 'timing/data': 0.012442894978448749, 'timing/model': 1.1434013360412791, 'learning_rate/qwen_vl_interface': 9.2342379918952e-06, 'learning_rate/action_model': 9.2342379918952e-05, 'epoch': 1.0}) +[2026-06-19 11:58:13,137][starVLA.training.train_starvla][INFO] - Step 999, Loss: {'train/loss': 0.22936248779296875, 'train/grad_norm_pre_clip': 4.61137056350708, 'timing/data': 0.013496669009327888, 'timing/model': 1.320064638974145, 'learning_rate/qwen_vl_interface': 9.232579063609631e-06, 'learning_rate/action_model': 9.232579063609631e-05, 'epoch': 1.0}) +[2026-06-19 11:58:17,581][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1000 +[2026-06-19 12:01:11,916][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1000 +[2026-06-19 12:01:11,917][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 1000 +[2026-06-19 12:04:21,084][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 1000 +[2026-06-19 12:04:21,085][starVLA.training.train_starvla][INFO] - Step 1000, Loss: {'train/loss': 0.22817085683345795, 'train/grad_norm_pre_clip': 4.335756301879883, 'eval/loss': 0.32221317291259766, 'eval/latency_0/loss': 0.10487689822912216, 'eval/latency_1/loss': 0.14220155775547028, 'eval/latency_2/loss': 0.3874324560165405, 'eval/latency_3/loss': 0.4051094651222229, 'eval/latency_4/loss': 0.5625548958778381, 'eval/flappy/loss': 0.32221317291259766, 'eval/flappy/latency_0/loss': 0.10487689822912216, 'eval/flappy/latency_1/loss': 0.14220155775547028, 'eval/flappy/latency_2/loss': 0.3874324560165405, 'eval/flappy/latency_3/loss': 0.4051094651222229, 'eval/flappy/latency_4/loss': 0.5625548958778381, 'eval/flap_precision': 0.6923076923076923, 'eval/flap_recall': 0.5890909090909091, 'eval/flap_f1': 0.6365422396856582, 'eval/flap_rate_model': 0.065, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.6365422396856582, 'eval/latency_0/flap_precision': 0.6923076923076923, 'eval/latency_0/flap_recall': 0.5890909090909091, 'eval/latency_0/flap_f1': 0.6365422396856582, 'eval/latency_0/flap_rate_model': 0.065, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.6365422396856582, 'timing/data': 0.012837655027396977, 'timing/model': 1.1556196099845693, 'learning_rate/qwen_vl_interface': 9.230918498236243e-06, 'learning_rate/action_model': 9.230918498236244e-05, 'epoch': 1.0}) +[2026-06-19 12:04:21,085][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_1000_state +[2026-06-19 12:04:21,088][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-19 12:04:24,462][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_1000_state/model.safetensors +[2026-06-19 12:04:31,720][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_1000_state/optimizer.bin +[2026-06-19 12:04:31,721][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_1000_state/sampler.bin +[2026-06-19 12:04:31,721][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_1000_state/sampler_1.bin +[2026-06-19 12:04:31,721][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_1000_state/random_states_0.pkl +[2026-06-19 12:04:31,744][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1000 +[2026-06-19 12:04:31,744][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-19 12:04:31,754][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-19 12:04:36,268][starVLA.training.train_starvla][INFO] - Step 1001, Loss: {'train/loss': 0.21720311045646667, 'train/grad_norm_pre_clip': 5.354116439819336, 'timing/data': 0.012724130065180361, 'timing/model': 1.1308792839990929, 'learning_rate/qwen_vl_interface': 9.229256296457635e-06, 'learning_rate/action_model': 9.229256296457634e-05, 'epoch': 1.0}) +[2026-06-19 12:04:40,803][starVLA.training.train_starvla][INFO] - Step 1002, Loss: {'train/loss': 0.1901385337114334, 'train/grad_norm_pre_clip': 4.4072041511535645, 'timing/data': 0.013137842994183302, 'timing/model': 1.3004085388965905, 'learning_rate/qwen_vl_interface': 9.22759245895707e-06, 'learning_rate/action_model': 9.227592458957071e-05, 'epoch': 1.0}) +[2026-06-19 12:04:45,252][starVLA.training.train_starvla][INFO] - Step 1003, Loss: {'train/loss': 0.19107098877429962, 'train/grad_norm_pre_clip': 2.5512166023254395, 'timing/data': 0.013019706006161869, 'timing/model': 1.215859308023937, 'learning_rate/qwen_vl_interface': 9.225926986418494e-06, 'learning_rate/action_model': 9.225926986418494e-05, 'epoch': 1.0}) +[2026-06-19 12:04:49,609][starVLA.training.train_starvla][INFO] - Step 1004, Loss: {'train/loss': 0.29799070954322815, 'train/grad_norm_pre_clip': 3.8970305919647217, 'timing/data': 0.013171808095648885, 'timing/model': 1.1293419030262157, 'learning_rate/qwen_vl_interface': 9.224259879526519e-06, 'learning_rate/action_model': 9.224259879526519e-05, 'epoch': 1.0}) +[2026-06-19 12:04:54,094][starVLA.training.train_starvla][INFO] - Step 1005, Loss: {'train/loss': 0.21173226833343506, 'train/grad_norm_pre_clip': 3.2231287956237793, 'timing/data': 0.012137324083596468, 'timing/model': 1.130583913065493, 'learning_rate/qwen_vl_interface': 9.222591138966427e-06, 'learning_rate/action_model': 9.222591138966426e-05, 'epoch': 1.0}) +[2026-06-19 12:04:58,631][starVLA.training.train_starvla][INFO] - Step 1006, Loss: {'train/loss': 0.18628819286823273, 'train/grad_norm_pre_clip': 2.9486899375915527, 'timing/data': 0.012871221988461912, 'timing/model': 1.145411244011484, 'learning_rate/qwen_vl_interface': 9.220920765424175e-06, 'learning_rate/action_model': 9.220920765424174e-05, 'epoch': 1.0}) +[2026-06-19 12:05:03,182][starVLA.training.train_starvla][INFO] - Step 1007, Loss: {'train/loss': 0.19315510988235474, 'train/grad_norm_pre_clip': 3.1120269298553467, 'timing/data': 0.012454451061785221, 'timing/model': 1.1363113870611414, 'learning_rate/qwen_vl_interface': 9.219248759586392e-06, 'learning_rate/action_model': 9.219248759586391e-05, 'epoch': 1.0}) +[2026-06-19 12:05:07,708][starVLA.training.train_starvla][INFO] - Step 1008, Loss: {'train/loss': 0.2592320442199707, 'train/grad_norm_pre_clip': 3.9491031169891357, 'timing/data': 0.013375014998018742, 'timing/model': 1.2962586779613048, 'learning_rate/qwen_vl_interface': 9.217575122140375e-06, 'learning_rate/action_model': 9.217575122140375e-05, 'epoch': 1.01}) +[2026-06-19 12:05:12,115][starVLA.training.train_starvla][INFO] - Step 1009, Loss: {'train/loss': 0.21151067316532135, 'train/grad_norm_pre_clip': 6.443170070648193, 'timing/data': 0.012111324933357537, 'timing/model': 1.1474165649851784, 'learning_rate/qwen_vl_interface': 9.215899853774094e-06, 'learning_rate/action_model': 9.215899853774093e-05, 'epoch': 1.01}) +[2026-06-19 12:05:16,680][starVLA.training.train_starvla][INFO] - Step 1010, Loss: {'train/loss': 0.6920514702796936, 'train/grad_norm_pre_clip': 15.040128707885742, 'timing/data': 0.013245253008790314, 'timing/model': 1.142946630017832, 'learning_rate/qwen_vl_interface': 9.214222955176189e-06, 'learning_rate/action_model': 9.214222955176189e-05, 'epoch': 1.01}) +[2026-06-19 12:05:21,200][starVLA.training.train_starvla][INFO] - Step 1011, Loss: {'train/loss': 0.37451285123825073, 'train/grad_norm_pre_clip': 5.845148086547852, 'timing/data': 0.012087634997442365, 'timing/model': 1.1464737199712545, 'learning_rate/qwen_vl_interface': 9.212544427035967e-06, 'learning_rate/action_model': 9.212544427035967e-05, 'epoch': 1.01}) +[2026-06-19 12:05:25,748][starVLA.training.train_starvla][INFO] - Step 1012, Loss: {'train/loss': 0.24146439135074615, 'train/grad_norm_pre_clip': 7.804171085357666, 'timing/data': 0.012589434045366943, 'timing/model': 1.1395861930213869, 'learning_rate/qwen_vl_interface': 9.21086427004341e-06, 'learning_rate/action_model': 9.210864270043411e-05, 'epoch': 1.01}) +[2026-06-19 12:05:30,316][starVLA.training.train_starvla][INFO] - Step 1013, Loss: {'train/loss': 0.14594751596450806, 'train/grad_norm_pre_clip': 4.573538780212402, 'timing/data': 0.013548411894589663, 'timing/model': 1.1400858039269224, 'learning_rate/qwen_vl_interface': 9.209182484889168e-06, 'learning_rate/action_model': 9.209182484889167e-05, 'epoch': 1.01}) +[2026-06-19 12:05:34,867][starVLA.training.train_starvla][INFO] - Step 1014, Loss: {'train/loss': 0.22327488660812378, 'train/grad_norm_pre_clip': 8.08250904083252, 'timing/data': 0.013576550991274416, 'timing/model': 1.2971522369189188, 'learning_rate/qwen_vl_interface': 9.207499072264557e-06, 'learning_rate/action_model': 9.207499072264558e-05, 'epoch': 1.01}) +[2026-06-19 12:05:39,259][starVLA.training.train_starvla][INFO] - Step 1015, Loss: {'train/loss': 0.42338827252388, 'train/grad_norm_pre_clip': 9.15910530090332, 'timing/data': 0.013012584066018462, 'timing/model': 1.1373448610538617, 'learning_rate/qwen_vl_interface': 9.205814032861566e-06, 'learning_rate/action_model': 9.205814032861566e-05, 'epoch': 1.01}) +[2026-06-19 12:05:43,819][starVLA.training.train_starvla][INFO] - Step 1016, Loss: {'train/loss': 0.4028592109680176, 'train/grad_norm_pre_clip': 7.605703830718994, 'timing/data': 0.013060677913017571, 'timing/model': 1.1384446610463783, 'learning_rate/qwen_vl_interface': 9.204127367372849e-06, 'learning_rate/action_model': 9.204127367372849e-05, 'epoch': 1.01}) +[2026-06-19 12:05:48,342][starVLA.training.train_starvla][INFO] - Step 1017, Loss: {'train/loss': 0.14095143973827362, 'train/grad_norm_pre_clip': 3.5332090854644775, 'timing/data': 0.01386376901064068, 'timing/model': 1.146812791004777, 'learning_rate/qwen_vl_interface': 9.202439076491732e-06, 'learning_rate/action_model': 9.202439076491732e-05, 'epoch': 1.01}) +[2026-06-19 12:05:52,937][starVLA.training.train_starvla][INFO] - Step 1018, Loss: {'train/loss': 0.19523201882839203, 'train/grad_norm_pre_clip': 4.941577911376953, 'timing/data': 0.012977832928299904, 'timing/model': 1.1531981279840693, 'learning_rate/qwen_vl_interface': 9.200749160912208e-06, 'learning_rate/action_model': 9.200749160912208e-05, 'epoch': 1.02}) +[2026-06-19 12:05:57,527][starVLA.training.train_starvla][INFO] - Step 1019, Loss: {'train/loss': 0.4003722667694092, 'train/grad_norm_pre_clip': 6.656740665435791, 'timing/data': 0.012903145980089903, 'timing/model': 1.146443537902087, 'learning_rate/qwen_vl_interface': 9.199057621328937e-06, 'learning_rate/action_model': 9.199057621328935e-05, 'epoch': 1.02}) +[2026-06-19 12:06:02,133][starVLA.training.train_starvla][INFO] - Step 1020, Loss: {'train/loss': 0.27175620198249817, 'train/grad_norm_pre_clip': 6.073362350463867, 'timing/data': 0.013236794038675725, 'timing/model': 1.3331810450181365, 'learning_rate/qwen_vl_interface': 9.197364458437244e-06, 'learning_rate/action_model': 9.197364458437245e-05, 'epoch': 1.02}) +[2026-06-19 12:06:06,555][starVLA.training.train_starvla][INFO] - Step 1021, Loss: {'train/loss': 0.4383642077445984, 'train/grad_norm_pre_clip': 8.228095054626465, 'timing/data': 0.013233614969067276, 'timing/model': 1.1428954800358042, 'learning_rate/qwen_vl_interface': 9.19566967293313e-06, 'learning_rate/action_model': 9.19566967293313e-05, 'epoch': 1.02}) +[2026-06-19 12:06:11,118][starVLA.training.train_starvla][INFO] - Step 1022, Loss: {'train/loss': 0.2205972969532013, 'train/grad_norm_pre_clip': 5.18185567855835, 'timing/data': 0.013869463931769133, 'timing/model': 1.13894731993787, 'learning_rate/qwen_vl_interface': 9.193973265513253e-06, 'learning_rate/action_model': 9.193973265513253e-05, 'epoch': 1.02}) +[2026-06-19 12:06:15,689][starVLA.training.train_starvla][INFO] - Step 1023, Loss: {'train/loss': 0.3482376039028168, 'train/grad_norm_pre_clip': 7.624456882476807, 'timing/data': 0.013098001014441252, 'timing/model': 1.1410865900106728, 'learning_rate/qwen_vl_interface': 9.192275236874946e-06, 'learning_rate/action_model': 9.192275236874946e-05, 'epoch': 1.02}) +[2026-06-19 12:06:20,186][starVLA.training.train_starvla][INFO] - Step 1024, Loss: {'train/loss': 0.46391913294792175, 'train/grad_norm_pre_clip': 7.708672046661377, 'timing/data': 0.013862645020708442, 'timing/model': 1.1390832039760426, 'learning_rate/qwen_vl_interface': 9.190575587716199e-06, 'learning_rate/action_model': 9.190575587716199e-05, 'epoch': 1.02}) +[2026-06-19 12:06:24,748][starVLA.training.train_starvla][INFO] - Step 1025, Loss: {'train/loss': 0.12066257745027542, 'train/grad_norm_pre_clip': 3.1236321926116943, 'timing/data': 0.013846819987520576, 'timing/model': 1.145377597073093, 'learning_rate/qwen_vl_interface': 9.188874318735677e-06, 'learning_rate/action_model': 9.188874318735676e-05, 'epoch': 1.02}) +[2026-06-19 12:06:29,346][starVLA.training.train_starvla][INFO] - Step 1026, Loss: {'train/loss': 0.25611358880996704, 'train/grad_norm_pre_clip': 6.122915267944336, 'timing/data': 0.01307467499282211, 'timing/model': 1.3243825889658183, 'learning_rate/qwen_vl_interface': 9.187171430632707e-06, 'learning_rate/action_model': 9.187171430632707e-05, 'epoch': 1.02}) +[2026-06-19 12:06:33,796][starVLA.training.train_starvla][INFO] - Step 1027, Loss: {'train/loss': 0.3956063687801361, 'train/grad_norm_pre_clip': 5.053912162780762, 'timing/data': 0.01422681100666523, 'timing/model': 1.1532976949820295, 'learning_rate/qwen_vl_interface': 9.185466924107283e-06, 'learning_rate/action_model': 9.185466924107282e-05, 'epoch': 1.02}) +[2026-06-19 12:06:38,376][starVLA.training.train_starvla][INFO] - Step 1028, Loss: {'train/loss': 0.21629278361797333, 'train/grad_norm_pre_clip': 4.899959564208984, 'timing/data': 0.013838412938639522, 'timing/model': 1.1405031500617042, 'learning_rate/qwen_vl_interface': 9.18376079986006e-06, 'learning_rate/action_model': 9.183760799860061e-05, 'epoch': 1.03}) +[2026-06-19 12:06:42,962][starVLA.training.train_starvla][INFO] - Step 1029, Loss: {'train/loss': 0.23389507830142975, 'train/grad_norm_pre_clip': 3.8683595657348633, 'timing/data': 0.01319599593989551, 'timing/model': 1.155240637017414, 'learning_rate/qwen_vl_interface': 9.182053058592365e-06, 'learning_rate/action_model': 9.182053058592365e-05, 'epoch': 1.03}) +[2026-06-19 12:06:47,564][starVLA.training.train_starvla][INFO] - Step 1030, Loss: {'train/loss': 0.23317943513393402, 'train/grad_norm_pre_clip': 4.7738566398620605, 'timing/data': 0.014351165038533509, 'timing/model': 1.142408701009117, 'learning_rate/qwen_vl_interface': 9.180343701006183e-06, 'learning_rate/action_model': 9.180343701006183e-05, 'epoch': 1.03}) +[2026-06-19 12:06:52,156][starVLA.training.train_starvla][INFO] - Step 1031, Loss: {'train/loss': 0.28330662846565247, 'train/grad_norm_pre_clip': 6.377046585083008, 'timing/data': 0.012689055991359055, 'timing/model': 1.155521432054229, 'learning_rate/qwen_vl_interface': 9.17863272780417e-06, 'learning_rate/action_model': 9.178632727804168e-05, 'epoch': 1.03}) +[2026-06-19 12:06:56,744][starVLA.training.train_starvla][INFO] - Step 1032, Loss: {'train/loss': 0.15649987757205963, 'train/grad_norm_pre_clip': 3.0768511295318604, 'timing/data': 0.01288629393093288, 'timing/model': 1.3218485220568255, 'learning_rate/qwen_vl_interface': 9.176920139689638e-06, 'learning_rate/action_model': 9.176920139689637e-05, 'epoch': 1.03}) +[2026-06-19 12:07:01,353][starVLA.training.train_starvla][INFO] - Step 1033, Loss: {'train/loss': 0.3599075376987457, 'train/grad_norm_pre_clip': 9.030406951904297, 'timing/data': 0.013274677097797394, 'timing/model': 1.3076702789403498, 'learning_rate/qwen_vl_interface': 9.175205937366568e-06, 'learning_rate/action_model': 9.175205937366569e-05, 'epoch': 1.03}) +[2026-06-19 12:07:05,786][starVLA.training.train_starvla][INFO] - Step 1034, Loss: {'train/loss': 0.19398564100265503, 'train/grad_norm_pre_clip': 3.3009283542633057, 'timing/data': 0.012990063056349754, 'timing/model': 1.145494700060226, 'learning_rate/qwen_vl_interface': 9.173490121539608e-06, 'learning_rate/action_model': 9.173490121539608e-05, 'epoch': 1.03}) +[2026-06-19 12:07:10,382][starVLA.training.train_starvla][INFO] - Step 1035, Loss: {'train/loss': 0.19902043044567108, 'train/grad_norm_pre_clip': 5.4542460441589355, 'timing/data': 0.01338498794939369, 'timing/model': 1.1425456730648875, 'learning_rate/qwen_vl_interface': 9.171772692914062e-06, 'learning_rate/action_model': 9.171772692914061e-05, 'epoch': 1.03}) +[2026-06-19 12:07:14,962][starVLA.training.train_starvla][INFO] - Step 1036, Loss: {'train/loss': 0.22967734932899475, 'train/grad_norm_pre_clip': 5.830562114715576, 'timing/data': 0.012780586956068873, 'timing/model': 1.1441349990200251, 'learning_rate/qwen_vl_interface': 9.1700536521959e-06, 'learning_rate/action_model': 9.170053652195899e-05, 'epoch': 1.03}) +[2026-06-19 12:07:19,582][starVLA.training.train_starvla][INFO] - Step 1037, Loss: {'train/loss': 0.3956840932369232, 'train/grad_norm_pre_clip': 7.518966197967529, 'timing/data': 0.013371272012591362, 'timing/model': 1.1558798040496185, 'learning_rate/qwen_vl_interface': 9.168333000091754e-06, 'learning_rate/action_model': 9.168333000091754e-05, 'epoch': 1.03}) +[2026-06-19 12:07:24,190][starVLA.training.train_starvla][INFO] - Step 1038, Loss: {'train/loss': 0.20171765983104706, 'train/grad_norm_pre_clip': 5.132458686828613, 'timing/data': 0.013398470007814467, 'timing/model': 1.1446463080355898, 'learning_rate/qwen_vl_interface': 9.16661073730892e-06, 'learning_rate/action_model': 9.166610737308919e-05, 'epoch': 1.04}) +[2026-06-19 12:07:28,757][starVLA.training.train_starvla][INFO] - Step 1039, Loss: {'train/loss': 0.2562594413757324, 'train/grad_norm_pre_clip': 6.685243129730225, 'timing/data': 0.013808552990667522, 'timing/model': 1.2968780869850889, 'learning_rate/qwen_vl_interface': 9.164886864555355e-06, 'learning_rate/action_model': 9.164886864555355e-05, 'epoch': 1.04}) +[2026-06-19 12:07:33,163][starVLA.training.train_starvla][INFO] - Step 1040, Loss: {'train/loss': 0.20641562342643738, 'train/grad_norm_pre_clip': 4.878001689910889, 'timing/data': 0.012468555010855198, 'timing/model': 1.1418966710334644, 'learning_rate/qwen_vl_interface': 9.163161382539678e-06, 'learning_rate/action_model': 9.163161382539678e-05, 'epoch': 1.04}) +[2026-06-19 12:07:37,736][starVLA.training.train_starvla][INFO] - Step 1041, Loss: {'train/loss': 0.19909657537937164, 'train/grad_norm_pre_clip': 6.379240989685059, 'timing/data': 0.013814495992846787, 'timing/model': 1.1493653940269724, 'learning_rate/qwen_vl_interface': 9.16143429197117e-06, 'learning_rate/action_model': 9.16143429197117e-05, 'epoch': 1.04}) +[2026-06-19 12:07:42,340][starVLA.training.train_starvla][INFO] - Step 1042, Loss: {'train/loss': 0.15313370525836945, 'train/grad_norm_pre_clip': 5.9817728996276855, 'timing/data': 0.013610235997475684, 'timing/model': 1.1488499760162085, 'learning_rate/qwen_vl_interface': 9.159705593559773e-06, 'learning_rate/action_model': 9.159705593559773e-05, 'epoch': 1.04}) +[2026-06-19 12:07:46,850][starVLA.training.train_starvla][INFO] - Step 1043, Loss: {'train/loss': 0.19911187887191772, 'train/grad_norm_pre_clip': 6.310877799987793, 'timing/data': 0.015514389961026609, 'timing/model': 1.1479933980153874, 'learning_rate/qwen_vl_interface': 9.157975288016089e-06, 'learning_rate/action_model': 9.157975288016088e-05, 'epoch': 1.04}) +[2026-06-19 12:07:51,437][starVLA.training.train_starvla][INFO] - Step 1044, Loss: {'train/loss': 0.32108184695243835, 'train/grad_norm_pre_clip': 9.026684761047363, 'timing/data': 0.012665175017900765, 'timing/model': 1.1513394169742242, 'learning_rate/qwen_vl_interface': 9.156243376051381e-06, 'learning_rate/action_model': 9.156243376051381e-05, 'epoch': 1.04}) +[2026-06-19 12:07:56,040][starVLA.training.train_starvla][INFO] - Step 1045, Loss: {'train/loss': 0.3789876699447632, 'train/grad_norm_pre_clip': 19.37908363342285, 'timing/data': 0.012590673053637147, 'timing/model': 1.3229875529650599, 'learning_rate/qwen_vl_interface': 9.154509858377574e-06, 'learning_rate/action_model': 9.154509858377574e-05, 'epoch': 1.04}) +[2026-06-19 12:08:00,469][starVLA.training.train_starvla][INFO] - Step 1046, Loss: {'train/loss': 0.1784471720457077, 'train/grad_norm_pre_clip': 5.959534168243408, 'timing/data': 0.012934044003486633, 'timing/model': 1.152232583030127, 'learning_rate/qwen_vl_interface': 9.152774735707248e-06, 'learning_rate/action_model': 9.152774735707247e-05, 'epoch': 1.04}) +[2026-06-19 12:08:05,045][starVLA.training.train_starvla][INFO] - Step 1047, Loss: {'train/loss': 0.24844898283481598, 'train/grad_norm_pre_clip': 5.580338478088379, 'timing/data': 0.012591408099979162, 'timing/model': 1.1487286700867116, 'learning_rate/qwen_vl_interface': 9.15103800875365e-06, 'learning_rate/action_model': 9.15103800875365e-05, 'epoch': 1.04}) +[2026-06-19 12:08:09,642][starVLA.training.train_starvla][INFO] - Step 1048, Loss: {'train/loss': 0.24590668082237244, 'train/grad_norm_pre_clip': 5.657140731811523, 'timing/data': 0.01374164898879826, 'timing/model': 1.1466067569563165, 'learning_rate/qwen_vl_interface': 9.149299678230681e-06, 'learning_rate/action_model': 9.149299678230681e-05, 'epoch': 1.05}) +[2026-06-19 12:08:14,246][starVLA.training.train_starvla][INFO] - Step 1049, Loss: {'train/loss': 0.21793729066848755, 'train/grad_norm_pre_clip': 5.169506072998047, 'timing/data': 0.01254793698899448, 'timing/model': 1.1474243269767612, 'learning_rate/qwen_vl_interface': 9.147559744852905e-06, 'learning_rate/action_model': 9.147559744852905e-05, 'epoch': 1.05}) +[2026-06-19 12:08:18,861][starVLA.training.train_starvla][INFO] - Step 1050, Loss: {'train/loss': 0.13046221435070038, 'train/grad_norm_pre_clip': 6.159543991088867, 'timing/data': 0.012607575976289809, 'timing/model': 1.308926068013534, 'learning_rate/qwen_vl_interface': 9.14581820933554e-06, 'learning_rate/action_model': 9.14581820933554e-05, 'epoch': 1.05}) +[2026-06-19 12:08:23,300][starVLA.training.train_starvla][INFO] - Step 1051, Loss: {'train/loss': 0.21051232516765594, 'train/grad_norm_pre_clip': 8.088692665100098, 'timing/data': 0.012788265012204647, 'timing/model': 1.1571349889272824, 'learning_rate/qwen_vl_interface': 9.14407507239447e-06, 'learning_rate/action_model': 9.14407507239447e-05, 'epoch': 1.05}) +[2026-06-19 12:08:27,896][starVLA.training.train_starvla][INFO] - Step 1052, Loss: {'train/loss': 0.20981112122535706, 'train/grad_norm_pre_clip': 3.0837323665618896, 'timing/data': 0.012821058044210076, 'timing/model': 1.1573452419834211, 'learning_rate/qwen_vl_interface': 9.142330334746227e-06, 'learning_rate/action_model': 9.142330334746227e-05, 'epoch': 1.05}) +[2026-06-19 12:08:32,490][starVLA.training.train_starvla][INFO] - Step 1053, Loss: {'train/loss': 0.7547028064727783, 'train/grad_norm_pre_clip': 18.40529441833496, 'timing/data': 0.012979810941033065, 'timing/model': 1.148310923948884, 'learning_rate/qwen_vl_interface': 9.14058399710801e-06, 'learning_rate/action_model': 9.14058399710801e-05, 'epoch': 1.05}) +[2026-06-19 12:08:37,047][starVLA.training.train_starvla][INFO] - Step 1054, Loss: {'train/loss': 0.3560020327568054, 'train/grad_norm_pre_clip': 13.3827543258667, 'timing/data': 0.013222292996942997, 'timing/model': 1.1440288230078295, 'learning_rate/qwen_vl_interface': 9.138836060197673e-06, 'learning_rate/action_model': 9.138836060197672e-05, 'epoch': 1.05}) +[2026-06-19 12:08:41,640][starVLA.training.train_starvla][INFO] - Step 1055, Loss: {'train/loss': 0.26988622546195984, 'train/grad_norm_pre_clip': 7.799154758453369, 'timing/data': 0.013234169105999172, 'timing/model': 1.152513534994796, 'learning_rate/qwen_vl_interface': 9.137086524733725e-06, 'learning_rate/action_model': 9.137086524733724e-05, 'epoch': 1.05}) +[2026-06-19 12:08:46,238][starVLA.training.train_starvla][INFO] - Step 1056, Loss: {'train/loss': 0.17814990878105164, 'train/grad_norm_pre_clip': 4.34405517578125, 'timing/data': 0.014074965030886233, 'timing/model': 1.3267381290206686, 'learning_rate/qwen_vl_interface': 9.135335391435336e-06, 'learning_rate/action_model': 9.135335391435335e-05, 'epoch': 1.05}) +[2026-06-19 12:08:50,657][starVLA.training.train_starvla][INFO] - Step 1057, Loss: {'train/loss': 0.23607808351516724, 'train/grad_norm_pre_clip': 5.9832940101623535, 'timing/data': 0.01258486695587635, 'timing/model': 1.148127791006118, 'learning_rate/qwen_vl_interface': 9.13358266102233e-06, 'learning_rate/action_model': 9.133582661022328e-05, 'epoch': 1.05}) +[2026-06-19 12:08:55,252][starVLA.training.train_starvla][INFO] - Step 1058, Loss: {'train/loss': 0.24648219347000122, 'train/grad_norm_pre_clip': 6.68474817276001, 'timing/data': 0.013411486055701971, 'timing/model': 1.1465108799748123, 'learning_rate/qwen_vl_interface': 9.131828334215185e-06, 'learning_rate/action_model': 9.131828334215185e-05, 'epoch': 1.05}) +[2026-06-19 12:08:59,765][starVLA.training.train_starvla][INFO] - Step 1059, Loss: {'train/loss': 0.2383565902709961, 'train/grad_norm_pre_clip': 11.185381889343262, 'timing/data': 0.01304014993365854, 'timing/model': 1.1495232600718737, 'learning_rate/qwen_vl_interface': 9.130072411735043e-06, 'learning_rate/action_model': 9.130072411735043e-05, 'epoch': 1.06}) +[2026-06-19 12:09:04,332][starVLA.training.train_starvla][INFO] - Step 1060, Loss: {'train/loss': 0.3119742274284363, 'train/grad_norm_pre_clip': 9.362740516662598, 'timing/data': 0.014023798983544111, 'timing/model': 1.153350356966257, 'learning_rate/qwen_vl_interface': 9.128314894303693e-06, 'learning_rate/action_model': 9.128314894303693e-05, 'epoch': 1.06}) +[2026-06-19 12:09:08,943][starVLA.training.train_starvla][INFO] - Step 1061, Loss: {'train/loss': 0.23623491823673248, 'train/grad_norm_pre_clip': 6.600555896759033, 'timing/data': 0.012211713008582592, 'timing/model': 1.1572100530611351, 'learning_rate/qwen_vl_interface': 9.126555782643592e-06, 'learning_rate/action_model': 9.126555782643591e-05, 'epoch': 1.06}) +[2026-06-19 12:09:13,521][starVLA.training.train_starvla][INFO] - Step 1062, Loss: {'train/loss': 0.1942441314458847, 'train/grad_norm_pre_clip': 5.61248254776001, 'timing/data': 0.012770081986673176, 'timing/model': 1.306901374948211, 'learning_rate/qwen_vl_interface': 9.124795077477835e-06, 'learning_rate/action_model': 9.124795077477834e-05, 'epoch': 1.06}) +[2026-06-19 12:09:17,941][starVLA.training.train_starvla][INFO] - Step 1063, Loss: {'train/loss': 0.2996845841407776, 'train/grad_norm_pre_clip': 5.968121528625488, 'timing/data': 0.013589861919172108, 'timing/model': 1.1491459540557116, 'learning_rate/qwen_vl_interface': 9.123032779530186e-06, 'learning_rate/action_model': 9.123032779530187e-05, 'epoch': 1.06}) +[2026-06-19 12:09:22,559][starVLA.training.train_starvla][INFO] - Step 1064, Loss: {'train/loss': 0.2969006896018982, 'train/grad_norm_pre_clip': 11.04738712310791, 'timing/data': 0.013747212011367083, 'timing/model': 1.1472061619861051, 'learning_rate/qwen_vl_interface': 9.12126888952506e-06, 'learning_rate/action_model': 9.12126888952506e-05, 'epoch': 1.06}) +[2026-06-19 12:09:27,069][starVLA.training.train_starvla][INFO] - Step 1065, Loss: {'train/loss': 0.2661270201206207, 'train/grad_norm_pre_clip': 8.473628997802734, 'timing/data': 0.01264025701675564, 'timing/model': 1.1441926779225469, 'learning_rate/qwen_vl_interface': 9.119503408187521e-06, 'learning_rate/action_model': 9.119503408187522e-05, 'epoch': 1.06}) +[2026-06-19 12:09:31,614][starVLA.training.train_starvla][INFO] - Step 1066, Loss: {'train/loss': 0.3734557330608368, 'train/grad_norm_pre_clip': 10.33008861541748, 'timing/data': 0.01307766605168581, 'timing/model': 1.1452922699972987, 'learning_rate/qwen_vl_interface': 9.117736336243299e-06, 'learning_rate/action_model': 9.117736336243298e-05, 'epoch': 1.06}) +[2026-06-19 12:09:36,225][starVLA.training.train_starvla][INFO] - Step 1067, Loss: {'train/loss': 0.1893303543329239, 'train/grad_norm_pre_clip': 6.053738594055176, 'timing/data': 0.012390822055749595, 'timing/model': 1.1502456500893459, 'learning_rate/qwen_vl_interface': 9.115967674418765e-06, 'learning_rate/action_model': 9.115967674418765e-05, 'epoch': 1.06}) +[2026-06-19 12:09:40,846][starVLA.training.train_starvla][INFO] - Step 1068, Loss: {'train/loss': 0.2298361361026764, 'train/grad_norm_pre_clip': 9.2184419631958, 'timing/data': 0.012620663968846202, 'timing/model': 1.3208079040050507, 'learning_rate/qwen_vl_interface': 9.114197423440949e-06, 'learning_rate/action_model': 9.11419742344095e-05, 'epoch': 1.06}) +[2026-06-19 12:09:45,293][starVLA.training.train_starvla][INFO] - Step 1069, Loss: {'train/loss': 0.23305587470531464, 'train/grad_norm_pre_clip': 6.767329692840576, 'timing/data': 0.012841204996220767, 'timing/model': 1.1537246049847454, 'learning_rate/qwen_vl_interface': 9.112425584037536e-06, 'learning_rate/action_model': 9.112425584037536e-05, 'epoch': 1.07}) +[2026-06-19 12:09:49,920][starVLA.training.train_starvla][INFO] - Step 1070, Loss: {'train/loss': 0.33483338356018066, 'train/grad_norm_pre_clip': 12.930562973022461, 'timing/data': 0.014356579049490392, 'timing/model': 1.1429715299746022, 'learning_rate/qwen_vl_interface': 9.110652156936863e-06, 'learning_rate/action_model': 9.110652156936863e-05, 'epoch': 1.07}) +[2026-06-19 12:09:54,434][starVLA.training.train_starvla][INFO] - Step 1071, Loss: {'train/loss': 0.4977920651435852, 'train/grad_norm_pre_clip': 9.334806442260742, 'timing/data': 0.013467879965901375, 'timing/model': 1.1611781060928479, 'learning_rate/qwen_vl_interface': 9.108877142867918e-06, 'learning_rate/action_model': 9.108877142867918e-05, 'epoch': 1.07}) +[2026-06-19 12:09:59,022][starVLA.training.train_starvla][INFO] - Step 1072, Loss: {'train/loss': 0.2104344516992569, 'train/grad_norm_pre_clip': 25.026033401489258, 'timing/data': 0.013766075018793344, 'timing/model': 1.1631745849736035, 'learning_rate/qwen_vl_interface': 9.107100542560342e-06, 'learning_rate/action_model': 9.107100542560342e-05, 'epoch': 1.07}) +[2026-06-19 12:10:03,664][starVLA.training.train_starvla][INFO] - Step 1073, Loss: {'train/loss': 0.43043532967567444, 'train/grad_norm_pre_clip': 8.001591682434082, 'timing/data': 0.021810061065480113, 'timing/model': 1.150180954951793, 'learning_rate/qwen_vl_interface': 9.105322356744429e-06, 'learning_rate/action_model': 9.105322356744429e-05, 'epoch': 1.07}) +[2026-06-19 12:10:08,273][starVLA.training.train_starvla][INFO] - Step 1074, Loss: {'train/loss': 0.3243202567100525, 'train/grad_norm_pre_clip': 12.487995147705078, 'timing/data': 0.013609800022095442, 'timing/model': 1.3279520249925554, 'learning_rate/qwen_vl_interface': 9.103542586151121e-06, 'learning_rate/action_model': 9.10354258615112e-05, 'epoch': 1.07}) +[2026-06-19 12:10:12,849][starVLA.training.train_starvla][INFO] - Step 1075, Loss: {'train/loss': 0.16313974559307098, 'train/grad_norm_pre_clip': 4.194576263427734, 'timing/data': 0.013150639017112553, 'timing/model': 1.2875486490083858, 'learning_rate/qwen_vl_interface': 9.101761231512017e-06, 'learning_rate/action_model': 9.101761231512017e-05, 'epoch': 1.07}) +[2026-06-19 12:10:17,282][starVLA.training.train_starvla][INFO] - Step 1076, Loss: {'train/loss': 0.29677873849868774, 'train/grad_norm_pre_clip': 6.320961952209473, 'timing/data': 0.013872693059965968, 'timing/model': 1.152550703031011, 'learning_rate/qwen_vl_interface': 9.099978293559365e-06, 'learning_rate/action_model': 9.099978293559366e-05, 'epoch': 1.07}) +[2026-06-19 12:10:21,875][starVLA.training.train_starvla][INFO] - Step 1077, Loss: {'train/loss': 0.2658977508544922, 'train/grad_norm_pre_clip': 7.930066108703613, 'timing/data': 0.012689779046922922, 'timing/model': 1.1530455189058557, 'learning_rate/qwen_vl_interface': 9.098193773026065e-06, 'learning_rate/action_model': 9.098193773026064e-05, 'epoch': 1.07}) +[2026-06-19 12:10:26,470][starVLA.training.train_starvla][INFO] - Step 1078, Loss: {'train/loss': 0.2860294580459595, 'train/grad_norm_pre_clip': 4.8795270919799805, 'timing/data': 0.013518795021809638, 'timing/model': 1.1472065950511023, 'learning_rate/qwen_vl_interface': 9.096407670645663e-06, 'learning_rate/action_model': 9.096407670645662e-05, 'epoch': 1.07}) +[2026-06-19 12:10:31,079][starVLA.training.train_starvla][INFO] - Step 1079, Loss: {'train/loss': 0.0739467516541481, 'train/grad_norm_pre_clip': 1.4711908102035522, 'timing/data': 0.01322727499064058, 'timing/model': 1.143984469003044, 'learning_rate/qwen_vl_interface': 9.094619987152357e-06, 'learning_rate/action_model': 9.094619987152358e-05, 'epoch': 1.08}) +[2026-06-19 12:10:35,588][starVLA.training.train_starvla][INFO] - Step 1080, Loss: {'train/loss': 0.22978606820106506, 'train/grad_norm_pre_clip': 6.298342227935791, 'timing/data': 0.012959587969817221, 'timing/model': 1.143215341027826, 'learning_rate/qwen_vl_interface': 9.092830723281e-06, 'learning_rate/action_model': 9.092830723281e-05, 'epoch': 1.08}) +[2026-06-19 12:10:40,121][starVLA.training.train_starvla][INFO] - Step 1081, Loss: {'train/loss': 0.1757352650165558, 'train/grad_norm_pre_clip': 3.2760796546936035, 'timing/data': 0.01303445000667125, 'timing/model': 1.2433921709889546, 'learning_rate/qwen_vl_interface': 9.09103987976709e-06, 'learning_rate/action_model': 9.09103987976709e-05, 'epoch': 1.08}) +[2026-06-19 12:10:44,551][starVLA.training.train_starvla][INFO] - Step 1082, Loss: {'train/loss': 0.26351216435432434, 'train/grad_norm_pre_clip': 9.605769157409668, 'timing/data': 0.014378743944689631, 'timing/model': 1.1438465720275417, 'learning_rate/qwen_vl_interface': 9.089247457346774e-06, 'learning_rate/action_model': 9.089247457346773e-05, 'epoch': 1.08}) +[2026-06-19 12:10:49,163][starVLA.training.train_starvla][INFO] - Step 1083, Loss: {'train/loss': 0.316726416349411, 'train/grad_norm_pre_clip': 6.549941539764404, 'timing/data': 0.013530645985156298, 'timing/model': 1.1449969939421862, 'learning_rate/qwen_vl_interface': 9.087453456756849e-06, 'learning_rate/action_model': 9.087453456756849e-05, 'epoch': 1.08}) +[2026-06-19 12:10:53,781][starVLA.training.train_starvla][INFO] - Step 1084, Loss: {'train/loss': 0.15793518722057343, 'train/grad_norm_pre_clip': 5.110330581665039, 'timing/data': 0.013557897065766156, 'timing/model': 1.1522195449797437, 'learning_rate/qwen_vl_interface': 9.085657878734759e-06, 'learning_rate/action_model': 9.085657878734758e-05, 'epoch': 1.08}) +[2026-06-19 12:10:58,397][starVLA.training.train_starvla][INFO] - Step 1085, Loss: {'train/loss': 0.1916378140449524, 'train/grad_norm_pre_clip': 5.500722885131836, 'timing/data': 0.012656804989092052, 'timing/model': 1.1514906840166077, 'learning_rate/qwen_vl_interface': 9.083860724018603e-06, 'learning_rate/action_model': 9.083860724018602e-05, 'epoch': 1.08}) +[2026-06-19 12:11:03,027][starVLA.training.train_starvla][INFO] - Step 1086, Loss: {'train/loss': 0.17553561925888062, 'train/grad_norm_pre_clip': 3.809326171875, 'timing/data': 0.012872526072897017, 'timing/model': 1.3260992569848895, 'learning_rate/qwen_vl_interface': 9.08206199334712e-06, 'learning_rate/action_model': 9.082061993347119e-05, 'epoch': 1.08}) +[2026-06-19 12:11:07,575][starVLA.training.train_starvla][INFO] - Step 1087, Loss: {'train/loss': 0.2843053936958313, 'train/grad_norm_pre_clip': 6.207118511199951, 'timing/data': 0.013498620013706386, 'timing/model': 1.2492326829815283, 'learning_rate/qwen_vl_interface': 9.0802616874597e-06, 'learning_rate/action_model': 9.0802616874597e-05, 'epoch': 1.08}) +[2026-06-19 12:11:12,005][starVLA.training.train_starvla][INFO] - Step 1088, Loss: {'train/loss': 0.28802257776260376, 'train/grad_norm_pre_clip': 11.909879684448242, 'timing/data': 0.012881022994406521, 'timing/model': 1.1461229160195217, 'learning_rate/qwen_vl_interface': 9.078459807096385e-06, 'learning_rate/action_model': 9.078459807096385e-05, 'epoch': 1.08}) +[2026-06-19 12:11:16,592][starVLA.training.train_starvla][INFO] - Step 1089, Loss: {'train/loss': 0.1381237655878067, 'train/grad_norm_pre_clip': 4.482945919036865, 'timing/data': 0.013826011097989976, 'timing/model': 1.147339090006426, 'learning_rate/qwen_vl_interface': 9.076656352997854e-06, 'learning_rate/action_model': 9.076656352997853e-05, 'epoch': 1.09}) +[2026-06-19 12:11:21,176][starVLA.training.train_starvla][INFO] - Step 1090, Loss: {'train/loss': 0.231387659907341, 'train/grad_norm_pre_clip': 16.131444931030273, 'timing/data': 0.013324378989636898, 'timing/model': 1.1446007579797879, 'learning_rate/qwen_vl_interface': 9.074851325905443e-06, 'learning_rate/action_model': 9.074851325905442e-05, 'epoch': 1.09}) +[2026-06-19 12:11:25,756][starVLA.training.train_starvla][INFO] - Step 1091, Loss: {'train/loss': 0.16359750926494598, 'train/grad_norm_pre_clip': 5.655788421630859, 'timing/data': 0.013251828029751778, 'timing/model': 1.1433891280321404, 'learning_rate/qwen_vl_interface': 9.073044726561128e-06, 'learning_rate/action_model': 9.073044726561129e-05, 'epoch': 1.09}) +[2026-06-19 12:11:30,326][starVLA.training.train_starvla][INFO] - Step 1092, Loss: {'train/loss': 0.22726720571517944, 'train/grad_norm_pre_clip': 10.595468521118164, 'timing/data': 0.014000840950757265, 'timing/model': 1.2962056420510635, 'learning_rate/qwen_vl_interface': 9.071236555707536e-06, 'learning_rate/action_model': 9.071236555707537e-05, 'epoch': 1.09}) +[2026-06-19 12:11:34,927][starVLA.training.train_starvla][INFO] - Step 1093, Loss: {'train/loss': 0.1897502839565277, 'train/grad_norm_pre_clip': 5.496670246124268, 'timing/data': 0.013377616996876895, 'timing/model': 1.3133174340473488, 'learning_rate/qwen_vl_interface': 9.069426814087937e-06, 'learning_rate/action_model': 9.069426814087937e-05, 'epoch': 1.09}) +[2026-06-19 12:11:39,363][starVLA.training.train_starvla][INFO] - Step 1094, Loss: {'train/loss': 0.3156987726688385, 'train/grad_norm_pre_clip': 8.60385799407959, 'timing/data': 0.01316837896592915, 'timing/model': 1.1461956090060994, 'learning_rate/qwen_vl_interface': 9.067615502446247e-06, 'learning_rate/action_model': 9.067615502446246e-05, 'epoch': 1.09}) +[2026-06-19 12:11:43,940][starVLA.training.train_starvla][INFO] - Step 1095, Loss: {'train/loss': 0.3019794523715973, 'train/grad_norm_pre_clip': 16.90585708618164, 'timing/data': 0.01475925196427852, 'timing/model': 1.1430161269381642, 'learning_rate/qwen_vl_interface': 9.06580262152703e-06, 'learning_rate/action_model': 9.065802621527028e-05, 'epoch': 1.09}) +[2026-06-19 12:11:48,542][starVLA.training.train_starvla][INFO] - Step 1096, Loss: {'train/loss': 0.34695887565612793, 'train/grad_norm_pre_clip': 6.186957836151123, 'timing/data': 0.012575378059409559, 'timing/model': 1.1455356139922515, 'learning_rate/qwen_vl_interface': 9.063988172075486e-06, 'learning_rate/action_model': 9.063988172075486e-05, 'epoch': 1.09}) +[2026-06-19 12:11:53,038][starVLA.training.train_starvla][INFO] - Step 1097, Loss: {'train/loss': 0.23819655179977417, 'train/grad_norm_pre_clip': 10.078485488891602, 'timing/data': 0.012656536069698632, 'timing/model': 1.1443308430025354, 'learning_rate/qwen_vl_interface': 9.062172154837475e-06, 'learning_rate/action_model': 9.062172154837475e-05, 'epoch': 1.09}) +[2026-06-19 12:11:57,583][starVLA.training.train_starvla][INFO] - Step 1098, Loss: {'train/loss': 0.3092712461948395, 'train/grad_norm_pre_clip': 7.681814670562744, 'timing/data': 0.013571075978688896, 'timing/model': 1.1462630389723927, 'learning_rate/qwen_vl_interface': 9.060354570559487e-06, 'learning_rate/action_model': 9.060354570559487e-05, 'epoch': 1.09}) +[2026-06-19 12:12:02,178][starVLA.training.train_starvla][INFO] - Step 1099, Loss: {'train/loss': 0.2121555209159851, 'train/grad_norm_pre_clip': 5.484165668487549, 'timing/data': 0.0128422100096941, 'timing/model': 1.312176095903851, 'learning_rate/qwen_vl_interface': 9.058535419988667e-06, 'learning_rate/action_model': 9.058535419988667e-05, 'epoch': 1.1}) +[2026-06-19 12:12:06,601][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1100 +[2026-06-19 12:14:59,422][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1100 +[2026-06-19 12:14:59,481][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 1100 +[2026-06-19 12:18:08,491][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 1100 +[2026-06-19 12:18:08,492][starVLA.training.train_starvla][INFO] - Step 1100, Loss: {'train/loss': 0.1454940289258957, 'train/grad_norm_pre_clip': 5.294987201690674, 'eval/loss': 0.3448493182659149, 'eval/latency_0/loss': 0.1119745522737503, 'eval/latency_1/loss': 0.13259877264499664, 'eval/latency_2/loss': 0.36670175194740295, 'eval/latency_3/loss': 0.4581182897090912, 'eval/latency_4/loss': 0.6504282355308533, 'eval/flappy/loss': 0.3448493182659149, 'eval/flappy/latency_0/loss': 0.1119745522737503, 'eval/flappy/latency_1/loss': 0.13259877264499664, 'eval/flappy/latency_2/loss': 0.36670175194740295, 'eval/flappy/latency_3/loss': 0.4581182897090912, 'eval/flappy/latency_4/loss': 0.6504282355308533, 'eval/flap_precision': 0.7093596059113301, 'eval/flap_recall': 0.5236363636363637, 'eval/flap_f1': 0.602510460251046, 'eval/flap_rate_model': 0.05638888888888889, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.602510460251046, 'eval/latency_0/flap_precision': 0.7093596059113301, 'eval/latency_0/flap_recall': 0.5236363636363637, 'eval/latency_0/flap_f1': 0.602510460251046, 'eval/latency_0/flap_rate_model': 0.05638888888888889, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.602510460251046, 'timing/data': 0.012502929079346359, 'timing/model': 1.146418921998702, 'learning_rate/qwen_vl_interface': 9.056714703872794e-06, 'learning_rate/action_model': 9.056714703872794e-05, 'epoch': 1.1}) +[2026-06-19 12:18:08,493][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_1100_state +[2026-06-19 12:18:08,496][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-19 12:18:11,875][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_1100_state/model.safetensors +[2026-06-19 12:18:19,015][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_1100_state/optimizer.bin +[2026-06-19 12:18:19,015][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_1100_state/sampler.bin +[2026-06-19 12:18:19,016][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_1100_state/sampler_1.bin +[2026-06-19 12:18:19,016][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_1100_state/random_states_0.pkl +[2026-06-19 12:18:19,039][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1100 +[2026-06-19 12:18:19,039][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-19 12:18:19,090][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-19 12:18:23,569][starVLA.training.train_starvla][INFO] - Step 1101, Loss: {'train/loss': 0.235160693526268, 'train/grad_norm_pre_clip': 4.981550693511963, 'timing/data': 0.01271004299633205, 'timing/model': 1.1309568870346993, 'learning_rate/qwen_vl_interface': 9.0548924229603e-06, 'learning_rate/action_model': 9.054892422960301e-05, 'epoch': 1.1}) +[2026-06-19 12:18:27,996][starVLA.training.train_starvla][INFO] - Step 1102, Loss: {'train/loss': 0.16383372247219086, 'train/grad_norm_pre_clip': 4.072032928466797, 'timing/data': 0.012970998999662697, 'timing/model': 1.1297643610741943, 'learning_rate/qwen_vl_interface': 9.053068578000256e-06, 'learning_rate/action_model': 9.053068578000256e-05, 'epoch': 1.1}) +[2026-06-19 12:18:32,458][starVLA.training.train_starvla][INFO] - Step 1103, Loss: {'train/loss': 0.2832050621509552, 'train/grad_norm_pre_clip': 6.5990705490112305, 'timing/data': 0.013255270081572235, 'timing/model': 1.2522730430355296, 'learning_rate/qwen_vl_interface': 9.051243169742372e-06, 'learning_rate/action_model': 9.051243169742372e-05, 'epoch': 1.1}) +[2026-06-19 12:18:36,836][starVLA.training.train_starvla][INFO] - Step 1104, Loss: {'train/loss': 0.3112899959087372, 'train/grad_norm_pre_clip': 17.75865364074707, 'timing/data': 0.012553645996376872, 'timing/model': 1.1348233120515943, 'learning_rate/qwen_vl_interface': 9.049416198937008e-06, 'learning_rate/action_model': 9.049416198937007e-05, 'epoch': 1.1}) +[2026-06-19 12:18:41,399][starVLA.training.train_starvla][INFO] - Step 1105, Loss: {'train/loss': 0.30715206265449524, 'train/grad_norm_pre_clip': 6.6360883712768555, 'timing/data': 0.012920120963826776, 'timing/model': 1.1383278430439532, 'learning_rate/qwen_vl_interface': 9.047587666335159e-06, 'learning_rate/action_model': 9.047587666335159e-05, 'epoch': 1.1}) +[2026-06-19 12:18:45,955][starVLA.training.train_starvla][INFO] - Step 1106, Loss: {'train/loss': 0.10225984454154968, 'train/grad_norm_pre_clip': 3.5479743480682373, 'timing/data': 0.013184001087211072, 'timing/model': 1.141617932007648, 'learning_rate/qwen_vl_interface': 9.045757572688469e-06, 'learning_rate/action_model': 9.04575757268847e-05, 'epoch': 1.1}) +[2026-06-19 12:18:50,508][starVLA.training.train_starvla][INFO] - Step 1107, Loss: {'train/loss': 0.2094484567642212, 'train/grad_norm_pre_clip': 7.153894901275635, 'timing/data': 0.01467773993499577, 'timing/model': 1.1422134480671957, 'learning_rate/qwen_vl_interface': 9.043925918749218e-06, 'learning_rate/action_model': 9.043925918749218e-05, 'epoch': 1.1}) +[2026-06-19 12:18:55,064][starVLA.training.train_starvla][INFO] - Step 1108, Loss: {'train/loss': 0.48465824127197266, 'train/grad_norm_pre_clip': 8.067536354064941, 'timing/data': 0.012791737914085388, 'timing/model': 1.3070245661074296, 'learning_rate/qwen_vl_interface': 9.042092705270333e-06, 'learning_rate/action_model': 9.042092705270333e-05, 'epoch': 1.1}) +[2026-06-19 12:18:59,466][starVLA.training.train_starvla][INFO] - Step 1109, Loss: {'train/loss': 0.24945902824401855, 'train/grad_norm_pre_clip': 6.321811676025391, 'timing/data': 0.013720570015721023, 'timing/model': 1.1469636450055987, 'learning_rate/qwen_vl_interface': 9.040257933005375e-06, 'learning_rate/action_model': 9.040257933005374e-05, 'epoch': 1.11}) +[2026-06-19 12:19:04,052][starVLA.training.train_starvla][INFO] - Step 1110, Loss: {'train/loss': 0.1689189374446869, 'train/grad_norm_pre_clip': 4.468410491943359, 'timing/data': 0.013183366041630507, 'timing/model': 1.1384696390014142, 'learning_rate/qwen_vl_interface': 9.038421602708552e-06, 'learning_rate/action_model': 9.038421602708552e-05, 'epoch': 1.11}) +[2026-06-19 12:19:08,647][starVLA.training.train_starvla][INFO] - Step 1111, Loss: {'train/loss': 0.36236944794654846, 'train/grad_norm_pre_clip': 5.053832530975342, 'timing/data': 0.013372407061979175, 'timing/model': 1.1423144519794732, 'learning_rate/qwen_vl_interface': 9.036583715134709e-06, 'learning_rate/action_model': 9.036583715134708e-05, 'epoch': 1.11}) +[2026-06-19 12:19:13,213][starVLA.training.train_starvla][INFO] - Step 1112, Loss: {'train/loss': 0.16481055319309235, 'train/grad_norm_pre_clip': 5.42599630355835, 'timing/data': 0.013720855000428855, 'timing/model': 1.1383475520415232, 'learning_rate/qwen_vl_interface': 9.034744271039332e-06, 'learning_rate/action_model': 9.034744271039332e-05, 'epoch': 1.11}) +[2026-06-19 12:19:17,759][starVLA.training.train_starvla][INFO] - Step 1113, Loss: {'train/loss': 0.1928461343050003, 'train/grad_norm_pre_clip': 9.16369915008545, 'timing/data': 0.012863173033110797, 'timing/model': 1.135705470922403, 'learning_rate/qwen_vl_interface': 9.03290327117855e-06, 'learning_rate/action_model': 9.032903271178549e-05, 'epoch': 1.11}) +[2026-06-19 12:19:22,323][starVLA.training.train_starvla][INFO] - Step 1114, Loss: {'train/loss': 0.407060831785202, 'train/grad_norm_pre_clip': 7.251245498657227, 'timing/data': 0.013237230014055967, 'timing/model': 1.3004089320311323, 'learning_rate/qwen_vl_interface': 9.031060716309122e-06, 'learning_rate/action_model': 9.031060716309121e-05, 'epoch': 1.11}) +[2026-06-19 12:19:26,845][starVLA.training.train_starvla][INFO] - Step 1115, Loss: {'train/loss': 0.34814873337745667, 'train/grad_norm_pre_clip': 5.5537896156311035, 'timing/data': 0.012956683058291674, 'timing/model': 1.2552228040294722, 'learning_rate/qwen_vl_interface': 9.02921660718846e-06, 'learning_rate/action_model': 9.029216607188459e-05, 'epoch': 1.11}) +[2026-06-19 12:19:31,273][starVLA.training.train_starvla][INFO] - Step 1116, Loss: {'train/loss': 0.21591317653656006, 'train/grad_norm_pre_clip': 4.890594005584717, 'timing/data': 0.013754302985034883, 'timing/model': 1.142105245962739, 'learning_rate/qwen_vl_interface': 9.027370944574604e-06, 'learning_rate/action_model': 9.027370944574604e-05, 'epoch': 1.11}) +[2026-06-19 12:19:35,856][starVLA.training.train_starvla][INFO] - Step 1117, Loss: {'train/loss': 0.30251437425613403, 'train/grad_norm_pre_clip': 7.285199165344238, 'timing/data': 0.012955793063156307, 'timing/model': 1.1380834550363943, 'learning_rate/qwen_vl_interface': 9.025523729226236e-06, 'learning_rate/action_model': 9.025523729226236e-05, 'epoch': 1.11}) +[2026-06-19 12:19:40,464][starVLA.training.train_starvla][INFO] - Step 1118, Loss: {'train/loss': 0.32768571376800537, 'train/grad_norm_pre_clip': 10.428314208984375, 'timing/data': 0.012663748930208385, 'timing/model': 1.1471008379012346, 'learning_rate/qwen_vl_interface': 9.023674961902677e-06, 'learning_rate/action_model': 9.023674961902676e-05, 'epoch': 1.11}) +[2026-06-19 12:19:45,075][starVLA.training.train_starvla][INFO] - Step 1119, Loss: {'train/loss': 0.3384532034397125, 'train/grad_norm_pre_clip': 8.683874130249023, 'timing/data': 0.01302043697796762, 'timing/model': 1.1471127000404522, 'learning_rate/qwen_vl_interface': 9.021824643363883e-06, 'learning_rate/action_model': 9.021824643363883e-05, 'epoch': 1.12}) +[2026-06-19 12:19:49,674][starVLA.training.train_starvla][INFO] - Step 1120, Loss: {'train/loss': 0.3477676510810852, 'train/grad_norm_pre_clip': 6.462311744689941, 'timing/data': 0.012443958083167672, 'timing/model': 1.313854290987365, 'learning_rate/qwen_vl_interface': 9.019972774370456e-06, 'learning_rate/action_model': 9.019972774370455e-05, 'epoch': 1.12}) +[2026-06-19 12:19:54,097][starVLA.training.train_starvla][INFO] - Step 1121, Loss: {'train/loss': 0.2792049050331116, 'train/grad_norm_pre_clip': 6.688538074493408, 'timing/data': 0.01193914597388357, 'timing/model': 1.146757475915365, 'learning_rate/qwen_vl_interface': 9.018119355683623e-06, 'learning_rate/action_model': 9.018119355683623e-05, 'epoch': 1.12}) +[2026-06-19 12:19:58,702][starVLA.training.train_starvla][INFO] - Step 1122, Loss: {'train/loss': 0.3258182108402252, 'train/grad_norm_pre_clip': 5.2064619064331055, 'timing/data': 0.014763555955141783, 'timing/model': 1.1540758489863947, 'learning_rate/qwen_vl_interface': 9.016264388065258e-06, 'learning_rate/action_model': 9.016264388065258e-05, 'epoch': 1.12}) +[2026-06-19 12:20:03,341][starVLA.training.train_starvla][INFO] - Step 1123, Loss: {'train/loss': 0.28445470333099365, 'train/grad_norm_pre_clip': 5.327445030212402, 'timing/data': 0.012832979904487729, 'timing/model': 1.1564739110181108, 'learning_rate/qwen_vl_interface': 9.014407872277869e-06, 'learning_rate/action_model': 9.014407872277869e-05, 'epoch': 1.12}) +[2026-06-19 12:20:07,936][starVLA.training.train_starvla][INFO] - Step 1124, Loss: {'train/loss': 0.1503268927335739, 'train/grad_norm_pre_clip': 3.2126801013946533, 'timing/data': 0.012131151044741273, 'timing/model': 1.1409247410483658, 'learning_rate/qwen_vl_interface': 9.012549809084595e-06, 'learning_rate/action_model': 9.012549809084595e-05, 'epoch': 1.12}) +[2026-06-19 12:20:12,543][starVLA.training.train_starvla][INFO] - Step 1125, Loss: {'train/loss': 0.29925739765167236, 'train/grad_norm_pre_clip': 5.284294128417969, 'timing/data': 0.01203178393188864, 'timing/model': 1.144406032981351, 'learning_rate/qwen_vl_interface': 9.01069019924922e-06, 'learning_rate/action_model': 9.010690199249219e-05, 'epoch': 1.12}) +[2026-06-19 12:20:17,153][starVLA.training.train_starvla][INFO] - Step 1126, Loss: {'train/loss': 0.28188222646713257, 'train/grad_norm_pre_clip': 4.0824480056762695, 'timing/data': 0.012522747041657567, 'timing/model': 1.3102552810451016, 'learning_rate/qwen_vl_interface': 9.008829043536157e-06, 'learning_rate/action_model': 9.008829043536156e-05, 'epoch': 1.12}) +[2026-06-19 12:20:21,677][starVLA.training.train_starvla][INFO] - Step 1127, Loss: {'train/loss': 0.3616194725036621, 'train/grad_norm_pre_clip': 6.334323883056641, 'timing/data': 0.012169233057647943, 'timing/model': 1.253403004957363, 'learning_rate/qwen_vl_interface': 9.006966342710458e-06, 'learning_rate/action_model': 9.006966342710457e-05, 'epoch': 1.12}) +[2026-06-19 12:20:26,087][starVLA.training.train_starvla][INFO] - Step 1128, Loss: {'train/loss': 0.24087539315223694, 'train/grad_norm_pre_clip': 3.8691558837890625, 'timing/data': 0.0118272959953174, 'timing/model': 1.1457019379595295, 'learning_rate/qwen_vl_interface': 9.005102097537809e-06, 'learning_rate/action_model': 9.005102097537809e-05, 'epoch': 1.12}) +[2026-06-19 12:20:30,715][starVLA.training.train_starvla][INFO] - Step 1129, Loss: {'train/loss': 0.3212154507637024, 'train/grad_norm_pre_clip': 8.11077880859375, 'timing/data': 0.012752779060974717, 'timing/model': 1.1511826990172267, 'learning_rate/qwen_vl_interface': 9.00323630878453e-06, 'learning_rate/action_model': 9.00323630878453e-05, 'epoch': 1.13}) +[2026-06-19 12:20:35,354][starVLA.training.train_starvla][INFO] - Step 1130, Loss: {'train/loss': 0.17831628024578094, 'train/grad_norm_pre_clip': 9.005302429199219, 'timing/data': 0.01193256489932537, 'timing/model': 1.1460411080624908, 'learning_rate/qwen_vl_interface': 9.001368977217577e-06, 'learning_rate/action_model': 9.001368977217577e-05, 'epoch': 1.13}) +[2026-06-19 12:20:39,962][starVLA.training.train_starvla][INFO] - Step 1131, Loss: {'train/loss': 0.3287815451622009, 'train/grad_norm_pre_clip': 4.856297492980957, 'timing/data': 0.017630009911954403, 'timing/model': 1.1470676669850945, 'learning_rate/qwen_vl_interface': 8.999500103604538e-06, 'learning_rate/action_model': 8.999500103604536e-05, 'epoch': 1.13}) +[2026-06-19 12:20:44,566][starVLA.training.train_starvla][INFO] - Step 1132, Loss: {'train/loss': 0.22427630424499512, 'train/grad_norm_pre_clip': 4.478969573974609, 'timing/data': 0.01167054008692503, 'timing/model': 1.3170686570229009, 'learning_rate/qwen_vl_interface': 8.997629688713639e-06, 'learning_rate/action_model': 8.997629688713638e-05, 'epoch': 1.13}) +[2026-06-19 12:20:49,000][starVLA.training.train_starvla][INFO] - Step 1133, Loss: {'train/loss': 0.3424837589263916, 'train/grad_norm_pre_clip': 5.47789192199707, 'timing/data': 0.012559617985971272, 'timing/model': 1.1502894419245422, 'learning_rate/qwen_vl_interface': 8.995757733313733e-06, 'learning_rate/action_model': 8.995757733313732e-05, 'epoch': 1.13}) +[2026-06-19 12:20:53,589][starVLA.training.train_starvla][INFO] - Step 1134, Loss: {'train/loss': 0.3969663083553314, 'train/grad_norm_pre_clip': 6.214025497436523, 'timing/data': 0.011998540023341775, 'timing/model': 1.146354913013056, 'learning_rate/qwen_vl_interface': 8.993884238174314e-06, 'learning_rate/action_model': 8.993884238174314e-05, 'epoch': 1.13}) +[2026-06-19 12:20:58,202][starVLA.training.train_starvla][INFO] - Step 1135, Loss: {'train/loss': 0.2551518678665161, 'train/grad_norm_pre_clip': 3.6514334678649902, 'timing/data': 0.013449197984300554, 'timing/model': 1.1418077879352495, 'learning_rate/qwen_vl_interface': 8.992009204065505e-06, 'learning_rate/action_model': 8.992009204065505e-05, 'epoch': 1.13}) +[2026-06-19 12:21:02,788][starVLA.training.train_starvla][INFO] - Step 1136, Loss: {'train/loss': 0.22758792340755463, 'train/grad_norm_pre_clip': 2.7538022994995117, 'timing/data': 0.012938327970914543, 'timing/model': 1.1510346119757742, 'learning_rate/qwen_vl_interface': 8.990132631758058e-06, 'learning_rate/action_model': 8.990132631758059e-05, 'epoch': 1.13}) +[2026-06-19 12:21:07,381][starVLA.training.train_starvla][INFO] - Step 1137, Loss: {'train/loss': 0.2487228959798813, 'train/grad_norm_pre_clip': 4.233521461486816, 'timing/data': 0.012323974049650133, 'timing/model': 1.1409175869775936, 'learning_rate/qwen_vl_interface': 8.988254522023364e-06, 'learning_rate/action_model': 8.988254522023363e-05, 'epoch': 1.13}) +[2026-06-19 12:21:11,981][starVLA.training.train_starvla][INFO] - Step 1138, Loss: {'train/loss': 0.24758180975914001, 'train/grad_norm_pre_clip': 5.519019603729248, 'timing/data': 0.012287994963116944, 'timing/model': 1.3239520910428837, 'learning_rate/qwen_vl_interface': 8.986374875633441e-06, 'learning_rate/action_model': 8.98637487563344e-05, 'epoch': 1.13}) +[2026-06-19 12:21:16,550][starVLA.training.train_starvla][INFO] - Step 1139, Loss: {'train/loss': 0.3903149962425232, 'train/grad_norm_pre_clip': 6.2237772941589355, 'timing/data': 0.011610865942202508, 'timing/model': 1.2604476111009717, 'learning_rate/qwen_vl_interface': 8.984493693360944e-06, 'learning_rate/action_model': 8.984493693360943e-05, 'epoch': 1.14}) +[2026-06-19 12:21:20,988][starVLA.training.train_starvla][INFO] - Step 1140, Loss: {'train/loss': 0.2058851569890976, 'train/grad_norm_pre_clip': 5.463615417480469, 'timing/data': 0.01218476693611592, 'timing/model': 1.1502887450624257, 'learning_rate/qwen_vl_interface': 8.982610975979153e-06, 'learning_rate/action_model': 8.982610975979152e-05, 'epoch': 1.14}) +[2026-06-19 12:21:25,574][starVLA.training.train_starvla][INFO] - Step 1141, Loss: {'train/loss': 0.21582521498203278, 'train/grad_norm_pre_clip': 4.725713729858398, 'timing/data': 0.012346424045972526, 'timing/model': 1.1422893620328978, 'learning_rate/qwen_vl_interface': 8.98072672426198e-06, 'learning_rate/action_model': 8.980726724261981e-05, 'epoch': 1.14}) +[2026-06-19 12:21:30,187][starVLA.training.train_starvla][INFO] - Step 1142, Loss: {'train/loss': 0.44133687019348145, 'train/grad_norm_pre_clip': 7.040428638458252, 'timing/data': 0.011747978045605123, 'timing/model': 1.152394381002523, 'learning_rate/qwen_vl_interface': 8.978840938983977e-06, 'learning_rate/action_model': 8.978840938983977e-05, 'epoch': 1.14}) +[2026-06-19 12:21:34,795][starVLA.training.train_starvla][INFO] - Step 1143, Loss: {'train/loss': 0.2469709813594818, 'train/grad_norm_pre_clip': 5.540663719177246, 'timing/data': 0.012318237102590501, 'timing/model': 1.1502302539302036, 'learning_rate/qwen_vl_interface': 8.976953620920313e-06, 'learning_rate/action_model': 8.976953620920312e-05, 'epoch': 1.14}) +[2026-06-19 12:21:39,398][starVLA.training.train_starvla][INFO] - Step 1144, Loss: {'train/loss': 0.2508592903614044, 'train/grad_norm_pre_clip': 5.650670528411865, 'timing/data': 0.01322861504741013, 'timing/model': 1.1574442059500143, 'learning_rate/qwen_vl_interface': 8.975064770846794e-06, 'learning_rate/action_model': 8.975064770846794e-05, 'epoch': 1.14}) +[2026-06-19 12:21:44,025][starVLA.training.train_starvla][INFO] - Step 1145, Loss: {'train/loss': 0.2182040810585022, 'train/grad_norm_pre_clip': 4.388484001159668, 'timing/data': 0.012383954017423093, 'timing/model': 1.3239423059858382, 'learning_rate/qwen_vl_interface': 8.973174389539856e-06, 'learning_rate/action_model': 8.973174389539855e-05, 'epoch': 1.14}) +[2026-06-19 12:21:48,439][starVLA.training.train_starvla][INFO] - Step 1146, Loss: {'train/loss': 0.2113714963197708, 'train/grad_norm_pre_clip': 4.363962173461914, 'timing/data': 0.01173804304562509, 'timing/model': 1.1401834220159799, 'learning_rate/qwen_vl_interface': 8.971282477776563e-06, 'learning_rate/action_model': 8.971282477776563e-05, 'epoch': 1.14}) +[2026-06-19 12:21:53,020][starVLA.training.train_starvla][INFO] - Step 1147, Loss: {'train/loss': 0.29413068294525146, 'train/grad_norm_pre_clip': 11.056714057922363, 'timing/data': 0.012489545973949134, 'timing/model': 1.1452803330030292, 'learning_rate/qwen_vl_interface': 8.969389036334609e-06, 'learning_rate/action_model': 8.969389036334609e-05, 'epoch': 1.14}) +[2026-06-19 12:21:57,613][starVLA.training.train_starvla][INFO] - Step 1148, Loss: {'train/loss': 0.4237114191055298, 'train/grad_norm_pre_clip': 8.255000114440918, 'timing/data': 0.013387173996306956, 'timing/model': 1.1429058710345998, 'learning_rate/qwen_vl_interface': 8.967494065992315e-06, 'learning_rate/action_model': 8.967494065992315e-05, 'epoch': 1.14}) +[2026-06-19 12:22:02,252][starVLA.training.train_starvla][INFO] - Step 1149, Loss: {'train/loss': 0.39167454838752747, 'train/grad_norm_pre_clip': 10.54248332977295, 'timing/data': 0.013558673905208707, 'timing/model': 1.1494861750397831, 'learning_rate/qwen_vl_interface': 8.965597567528634e-06, 'learning_rate/action_model': 8.965597567528633e-05, 'epoch': 1.15}) +[2026-06-19 12:22:06,886][starVLA.training.train_starvla][INFO] - Step 1150, Loss: {'train/loss': 0.4310833811759949, 'train/grad_norm_pre_clip': 16.68623161315918, 'timing/data': 0.013283846899867058, 'timing/model': 1.1506962700514123, 'learning_rate/qwen_vl_interface': 8.963699541723143e-06, 'learning_rate/action_model': 8.963699541723142e-05, 'epoch': 1.15}) +[2026-06-19 12:22:11,493][starVLA.training.train_starvla][INFO] - Step 1151, Loss: {'train/loss': 0.3145730495452881, 'train/grad_norm_pre_clip': 6.600546836853027, 'timing/data': 0.012611632002517581, 'timing/model': 1.3228713989956304, 'learning_rate/qwen_vl_interface': 8.961799989356048e-06, 'learning_rate/action_model': 8.961799989356048e-05, 'epoch': 1.15}) +[2026-06-19 12:22:15,932][starVLA.training.train_starvla][INFO] - Step 1152, Loss: {'train/loss': 0.21973088383674622, 'train/grad_norm_pre_clip': 6.223585605621338, 'timing/data': 0.012060800101608038, 'timing/model': 1.1525432580383494, 'learning_rate/qwen_vl_interface': 8.959898911208185e-06, 'learning_rate/action_model': 8.959898911208185e-05, 'epoch': 1.15}) +[2026-06-19 12:22:20,537][starVLA.training.train_starvla][INFO] - Step 1153, Loss: {'train/loss': 0.2541837990283966, 'train/grad_norm_pre_clip': 9.044015884399414, 'timing/data': 0.012955660000443459, 'timing/model': 1.1457992600044236, 'learning_rate/qwen_vl_interface': 8.957996308061015e-06, 'learning_rate/action_model': 8.957996308061014e-05, 'epoch': 1.15}) +[2026-06-19 12:22:25,055][starVLA.training.train_starvla][INFO] - Step 1154, Loss: {'train/loss': 0.1356145143508911, 'train/grad_norm_pre_clip': 3.7293734550476074, 'timing/data': 0.012460513040423393, 'timing/model': 1.1598773420555517, 'learning_rate/qwen_vl_interface': 8.956092180696626e-06, 'learning_rate/action_model': 8.956092180696625e-05, 'epoch': 1.15}) +[2026-06-19 12:22:29,564][starVLA.training.train_starvla][INFO] - Step 1155, Loss: {'train/loss': 0.5210236310958862, 'train/grad_norm_pre_clip': 13.466344833374023, 'timing/data': 0.011905414052307606, 'timing/model': 1.1440132909920067, 'learning_rate/qwen_vl_interface': 8.95418652989773e-06, 'learning_rate/action_model': 8.95418652989773e-05, 'epoch': 1.15}) +[2026-06-19 12:22:34,169][starVLA.training.train_starvla][INFO] - Step 1156, Loss: {'train/loss': 0.33015620708465576, 'train/grad_norm_pre_clip': 6.8373332023620605, 'timing/data': 0.012039560009725392, 'timing/model': 1.1477424439508468, 'learning_rate/qwen_vl_interface': 8.952279356447675e-06, 'learning_rate/action_model': 8.952279356447675e-05, 'epoch': 1.15}) +[2026-06-19 12:22:38,751][starVLA.training.train_starvla][INFO] - Step 1157, Loss: {'train/loss': 0.1956242024898529, 'train/grad_norm_pre_clip': 6.770369052886963, 'timing/data': 0.012446303036995232, 'timing/model': 1.307393421069719, 'learning_rate/qwen_vl_interface': 8.950370661130421e-06, 'learning_rate/action_model': 8.950370661130422e-05, 'epoch': 1.15}) +[2026-06-19 12:22:43,166][starVLA.training.train_starvla][INFO] - Step 1158, Loss: {'train/loss': 0.2486368715763092, 'train/grad_norm_pre_clip': 5.72235631942749, 'timing/data': 0.012374173034913838, 'timing/model': 1.1449055280536413, 'learning_rate/qwen_vl_interface': 8.948460444730566e-06, 'learning_rate/action_model': 8.948460444730566e-05, 'epoch': 1.15}) +[2026-06-19 12:22:47,753][starVLA.training.train_starvla][INFO] - Step 1159, Loss: {'train/loss': 0.6203941106796265, 'train/grad_norm_pre_clip': 12.274311065673828, 'timing/data': 0.013854943914338946, 'timing/model': 1.1444963390240446, 'learning_rate/qwen_vl_interface': 8.946548708033324e-06, 'learning_rate/action_model': 8.946548708033323e-05, 'epoch': 1.16}) +[2026-06-19 12:22:52,364][starVLA.training.train_starvla][INFO] - Step 1160, Loss: {'train/loss': 0.13920636475086212, 'train/grad_norm_pre_clip': 10.34157943725586, 'timing/data': 0.013454045983962715, 'timing/model': 1.1564761570189148, 'learning_rate/qwen_vl_interface': 8.944635451824538e-06, 'learning_rate/action_model': 8.944635451824538e-05, 'epoch': 1.16}) +[2026-06-19 12:22:56,907][starVLA.training.train_starvla][INFO] - Step 1161, Loss: {'train/loss': 0.2750488221645355, 'train/grad_norm_pre_clip': 5.494795799255371, 'timing/data': 0.012822421034798026, 'timing/model': 1.154874510015361, 'learning_rate/qwen_vl_interface': 8.942720676890677e-06, 'learning_rate/action_model': 8.942720676890677e-05, 'epoch': 1.16}) +[2026-06-19 12:23:01,507][starVLA.training.train_starvla][INFO] - Step 1162, Loss: {'train/loss': 0.1609019786119461, 'train/grad_norm_pre_clip': 3.0023863315582275, 'timing/data': 0.012636817060410976, 'timing/model': 1.1546798839699477, 'learning_rate/qwen_vl_interface': 8.940804384018833e-06, 'learning_rate/action_model': 8.940804384018832e-05, 'epoch': 1.16}) +[2026-06-19 12:23:06,109][starVLA.training.train_starvla][INFO] - Step 1163, Loss: {'train/loss': 0.191246896982193, 'train/grad_norm_pre_clip': 4.567670822143555, 'timing/data': 0.012917901040054858, 'timing/model': 1.3202010450186208, 'learning_rate/qwen_vl_interface': 8.93888657399672e-06, 'learning_rate/action_model': 8.93888657399672e-05, 'epoch': 1.16}) +[2026-06-19 12:23:10,526][starVLA.training.train_starvla][INFO] - Step 1164, Loss: {'train/loss': 0.1719803810119629, 'train/grad_norm_pre_clip': 3.7976715564727783, 'timing/data': 0.01195530197583139, 'timing/model': 1.1421841440023854, 'learning_rate/qwen_vl_interface': 8.936967247612678e-06, 'learning_rate/action_model': 8.936967247612678e-05, 'epoch': 1.16}) +[2026-06-19 12:23:15,130][starVLA.training.train_starvla][INFO] - Step 1165, Loss: {'train/loss': 0.14796605706214905, 'train/grad_norm_pre_clip': 4.29620885848999, 'timing/data': 0.011657839990220964, 'timing/model': 1.146475210902281, 'learning_rate/qwen_vl_interface': 8.93504640565567e-06, 'learning_rate/action_model': 8.93504640565567e-05, 'epoch': 1.16}) +[2026-06-19 12:23:19,754][starVLA.training.train_starvla][INFO] - Step 1166, Loss: {'train/loss': 0.14966818690299988, 'train/grad_norm_pre_clip': 4.7673540115356445, 'timing/data': 0.011847599991597235, 'timing/model': 1.156968661933206, 'learning_rate/qwen_vl_interface': 8.933124048915278e-06, 'learning_rate/action_model': 8.933124048915278e-05, 'epoch': 1.16}) +[2026-06-19 12:23:24,395][starVLA.training.train_starvla][INFO] - Step 1167, Loss: {'train/loss': 0.3738515079021454, 'train/grad_norm_pre_clip': 11.88444709777832, 'timing/data': 0.01240464299917221, 'timing/model': 1.153835783014074, 'learning_rate/qwen_vl_interface': 8.931200178181715e-06, 'learning_rate/action_model': 8.931200178181715e-05, 'epoch': 1.16}) +[2026-06-19 12:23:29,017][starVLA.training.train_starvla][INFO] - Step 1168, Loss: {'train/loss': 0.18802466988563538, 'train/grad_norm_pre_clip': 5.777780532836914, 'timing/data': 0.01259272696916014, 'timing/model': 1.146215328015387, 'learning_rate/qwen_vl_interface': 8.929274794245812e-06, 'learning_rate/action_model': 8.929274794245812e-05, 'epoch': 1.16}) +[2026-06-19 12:23:33,632][starVLA.training.train_starvla][INFO] - Step 1169, Loss: {'train/loss': 0.38503849506378174, 'train/grad_norm_pre_clip': 10.985694885253906, 'timing/data': 0.014590911916457117, 'timing/model': 1.3369159760186449, 'learning_rate/qwen_vl_interface': 8.927347897899018e-06, 'learning_rate/action_model': 8.927347897899016e-05, 'epoch': 1.17}) +[2026-06-19 12:23:38,102][starVLA.training.train_starvla][INFO] - Step 1170, Loss: {'train/loss': 0.28775671124458313, 'train/grad_norm_pre_clip': 9.15132999420166, 'timing/data': 0.01391421293374151, 'timing/model': 1.1698738520499319, 'learning_rate/qwen_vl_interface': 8.925419489933407e-06, 'learning_rate/action_model': 8.925419489933408e-05, 'epoch': 1.17}) +[2026-06-19 12:23:42,724][starVLA.training.train_starvla][INFO] - Step 1171, Loss: {'train/loss': 0.2652421295642853, 'train/grad_norm_pre_clip': 10.539647102355957, 'timing/data': 0.012582106981426477, 'timing/model': 1.1540644919732586, 'learning_rate/qwen_vl_interface': 8.923489571141679e-06, 'learning_rate/action_model': 8.923489571141679e-05, 'epoch': 1.17}) +[2026-06-19 12:23:47,268][starVLA.training.train_starvla][INFO] - Step 1172, Loss: {'train/loss': 0.09817270189523697, 'train/grad_norm_pre_clip': 4.429211616516113, 'timing/data': 0.011621432029642165, 'timing/model': 1.1437196059850976, 'learning_rate/qwen_vl_interface': 8.921558142317147e-06, 'learning_rate/action_model': 8.921558142317147e-05, 'epoch': 1.17}) +[2026-06-19 12:23:51,856][starVLA.training.train_starvla][INFO] - Step 1173, Loss: {'train/loss': 0.20409488677978516, 'train/grad_norm_pre_clip': 10.138019561767578, 'timing/data': 0.012541604926809669, 'timing/model': 1.1572446980280802, 'learning_rate/qwen_vl_interface': 8.91962520425375e-06, 'learning_rate/action_model': 8.91962520425375e-05, 'epoch': 1.17}) +[2026-06-19 12:23:56,497][starVLA.training.train_starvla][INFO] - Step 1174, Loss: {'train/loss': 0.15072724223136902, 'train/grad_norm_pre_clip': 3.8309075832366943, 'timing/data': 0.011514108977280557, 'timing/model': 1.1594188830349594, 'learning_rate/qwen_vl_interface': 8.917690757746044e-06, 'learning_rate/action_model': 8.917690757746044e-05, 'epoch': 1.17}) +[2026-06-19 12:24:01,136][starVLA.training.train_starvla][INFO] - Step 1175, Loss: {'train/loss': 0.2375328093767166, 'train/grad_norm_pre_clip': 7.059514045715332, 'timing/data': 0.012369719101116061, 'timing/model': 1.327872845926322, 'learning_rate/qwen_vl_interface': 8.915754803589208e-06, 'learning_rate/action_model': 8.915754803589209e-05, 'epoch': 1.17}) +[2026-06-19 12:24:05,580][starVLA.training.train_starvla][INFO] - Step 1176, Loss: {'train/loss': 0.4472236931324005, 'train/grad_norm_pre_clip': 13.896872520446777, 'timing/data': 0.011924547026865184, 'timing/model': 1.1539501369697973, 'learning_rate/qwen_vl_interface': 8.913817342579042e-06, 'learning_rate/action_model': 8.913817342579042e-05, 'epoch': 1.17}) +[2026-06-19 12:24:10,199][starVLA.training.train_starvla][INFO] - Step 1177, Loss: {'train/loss': 0.2840113639831543, 'train/grad_norm_pre_clip': 8.385126113891602, 'timing/data': 0.013245531008578837, 'timing/model': 1.1521630079951137, 'learning_rate/qwen_vl_interface': 8.91187837551196e-06, 'learning_rate/action_model': 8.911878375511961e-05, 'epoch': 1.17}) +[2026-06-19 12:24:14,740][starVLA.training.train_starvla][INFO] - Step 1178, Loss: {'train/loss': 0.22249947488307953, 'train/grad_norm_pre_clip': 6.614721775054932, 'timing/data': 0.012350048054940999, 'timing/model': 1.1473907079780474, 'learning_rate/qwen_vl_interface': 8.909937903185e-06, 'learning_rate/action_model': 8.909937903185e-05, 'epoch': 1.17}) +[2026-06-19 12:24:19,334][starVLA.training.train_starvla][INFO] - Step 1179, Loss: {'train/loss': 0.44686412811279297, 'train/grad_norm_pre_clip': 11.41678237915039, 'timing/data': 0.012919469038024545, 'timing/model': 1.154331799945794, 'learning_rate/qwen_vl_interface': 8.907995926395814e-06, 'learning_rate/action_model': 8.907995926395815e-05, 'epoch': 1.18}) +[2026-06-19 12:24:23,951][starVLA.training.train_starvla][INFO] - Step 1180, Loss: {'train/loss': 0.5497015714645386, 'train/grad_norm_pre_clip': 10.438831329345703, 'timing/data': 0.01232294098008424, 'timing/model': 1.160970074008219, 'learning_rate/qwen_vl_interface': 8.90605244594268e-06, 'learning_rate/action_model': 8.906052445942679e-05, 'epoch': 1.18}) +[2026-06-19 12:24:28,569][starVLA.training.train_starvla][INFO] - Step 1181, Loss: {'train/loss': 0.25035130977630615, 'train/grad_norm_pre_clip': 4.5368242263793945, 'timing/data': 0.012740388046950102, 'timing/model': 1.3266581309726462, 'learning_rate/qwen_vl_interface': 8.904107462624484e-06, 'learning_rate/action_model': 8.904107462624483e-05, 'epoch': 1.18}) +[2026-06-19 12:24:33,009][starVLA.training.train_starvla][INFO] - Step 1182, Loss: {'train/loss': 0.12387750297784805, 'train/grad_norm_pre_clip': 3.0592379570007324, 'timing/data': 0.012133423937484622, 'timing/model': 1.1487394770374522, 'learning_rate/qwen_vl_interface': 8.902160977240739e-06, 'learning_rate/action_model': 8.902160977240739e-05, 'epoch': 1.18}) +[2026-06-19 12:24:37,611][starVLA.training.train_starvla][INFO] - Step 1183, Loss: {'train/loss': 0.2793053686618805, 'train/grad_norm_pre_clip': 6.859611511230469, 'timing/data': 0.013018815079703927, 'timing/model': 1.1525114269461483, 'learning_rate/qwen_vl_interface': 8.90021299059157e-06, 'learning_rate/action_model': 8.90021299059157e-05, 'epoch': 1.18}) +[2026-06-19 12:24:42,139][starVLA.training.train_starvla][INFO] - Step 1184, Loss: {'train/loss': 0.44737210869789124, 'train/grad_norm_pre_clip': 5.474088191986084, 'timing/data': 0.012159799924120307, 'timing/model': 1.1509175769751891, 'learning_rate/qwen_vl_interface': 8.898263503477722e-06, 'learning_rate/action_model': 8.898263503477721e-05, 'epoch': 1.18}) +[2026-06-19 12:24:46,696][starVLA.training.train_starvla][INFO] - Step 1185, Loss: {'train/loss': 0.30708858370780945, 'train/grad_norm_pre_clip': 6.878727912902832, 'timing/data': 0.012042323010973632, 'timing/model': 1.1547140770126134, 'learning_rate/qwen_vl_interface': 8.896312516700553e-06, 'learning_rate/action_model': 8.896312516700554e-05, 'epoch': 1.18}) +[2026-06-19 12:24:51,320][starVLA.training.train_starvla][INFO] - Step 1186, Loss: {'train/loss': 0.2590358555316925, 'train/grad_norm_pre_clip': 5.004444599151611, 'timing/data': 0.012327907024882734, 'timing/model': 1.155269130016677, 'learning_rate/qwen_vl_interface': 8.894360031062043e-06, 'learning_rate/action_model': 8.894360031062043e-05, 'epoch': 1.18}) +[2026-06-19 12:24:55,941][starVLA.training.train_starvla][INFO] - Step 1187, Loss: {'train/loss': 0.2252611666917801, 'train/grad_norm_pre_clip': 3.798154354095459, 'timing/data': 0.012829172075726092, 'timing/model': 1.317492348025553, 'learning_rate/qwen_vl_interface': 8.892406047364783e-06, 'learning_rate/action_model': 8.892406047364784e-05, 'epoch': 1.18}) +[2026-06-19 12:25:00,362][starVLA.training.train_starvla][INFO] - Step 1188, Loss: {'train/loss': 0.16483111679553986, 'train/grad_norm_pre_clip': 3.5109829902648926, 'timing/data': 0.01315989694558084, 'timing/model': 1.144858894054778, 'learning_rate/qwen_vl_interface': 8.890450566411984e-06, 'learning_rate/action_model': 8.890450566411984e-05, 'epoch': 1.18}) +[2026-06-19 12:25:04,985][starVLA.training.train_starvla][INFO] - Step 1189, Loss: {'train/loss': 0.2530200183391571, 'train/grad_norm_pre_clip': 4.368075370788574, 'timing/data': 0.011616972042247653, 'timing/model': 1.1537933551007882, 'learning_rate/qwen_vl_interface': 8.888493589007467e-06, 'learning_rate/action_model': 8.888493589007468e-05, 'epoch': 1.19}) +[2026-06-19 12:25:09,506][starVLA.training.train_starvla][INFO] - Step 1190, Loss: {'train/loss': 0.15562517940998077, 'train/grad_norm_pre_clip': 2.4504382610321045, 'timing/data': 0.011786577058956027, 'timing/model': 1.148034940008074, 'learning_rate/qwen_vl_interface': 8.886535115955676e-06, 'learning_rate/action_model': 8.886535115955675e-05, 'epoch': 1.19}) +[2026-06-19 12:25:14,044][starVLA.training.train_starvla][INFO] - Step 1191, Loss: {'train/loss': 0.2737430930137634, 'train/grad_norm_pre_clip': 3.6561484336853027, 'timing/data': 0.012912591104395688, 'timing/model': 1.1502974050818011, 'learning_rate/qwen_vl_interface': 8.884575148061662e-06, 'learning_rate/action_model': 8.88457514806166e-05, 'epoch': 1.19}) +[2026-06-19 12:25:18,660][starVLA.training.train_starvla][INFO] - Step 1192, Loss: {'train/loss': 0.20053045451641083, 'train/grad_norm_pre_clip': 6.053457736968994, 'timing/data': 0.012038434972055256, 'timing/model': 1.1448176599806175, 'learning_rate/qwen_vl_interface': 8.882613686131095e-06, 'learning_rate/action_model': 8.882613686131093e-05, 'epoch': 1.19}) +[2026-06-19 12:25:23,247][starVLA.training.train_starvla][INFO] - Step 1193, Loss: {'train/loss': 0.21835169196128845, 'train/grad_norm_pre_clip': 3.6068055629730225, 'timing/data': 0.012834087014198303, 'timing/model': 1.3141685290029272, 'learning_rate/qwen_vl_interface': 8.880650730970254e-06, 'learning_rate/action_model': 8.880650730970254e-05, 'epoch': 1.19}) +[2026-06-19 12:25:27,683][starVLA.training.train_starvla][INFO] - Step 1194, Loss: {'train/loss': 0.37450772523880005, 'train/grad_norm_pre_clip': 8.800454139709473, 'timing/data': 0.012255673995241523, 'timing/model': 1.1505665859440342, 'learning_rate/qwen_vl_interface': 8.878686283386042e-06, 'learning_rate/action_model': 8.878686283386042e-05, 'epoch': 1.19}) +[2026-06-19 12:25:32,307][starVLA.training.train_starvla][INFO] - Step 1195, Loss: {'train/loss': 0.2433013916015625, 'train/grad_norm_pre_clip': 7.991796016693115, 'timing/data': 0.011971854022704065, 'timing/model': 1.1502734189853072, 'learning_rate/qwen_vl_interface': 8.876720344185967e-06, 'learning_rate/action_model': 8.876720344185967e-05, 'epoch': 1.19}) +[2026-06-19 12:25:36,812][starVLA.training.train_starvla][INFO] - Step 1196, Loss: {'train/loss': 0.2220725119113922, 'train/grad_norm_pre_clip': 7.833723545074463, 'timing/data': 0.011527603957802057, 'timing/model': 1.1510437260149047, 'learning_rate/qwen_vl_interface': 8.874752914178149e-06, 'learning_rate/action_model': 8.87475291417815e-05, 'epoch': 1.19}) +[2026-06-19 12:25:41,368][starVLA.training.train_starvla][INFO] - Step 1197, Loss: {'train/loss': 0.22914113104343414, 'train/grad_norm_pre_clip': 8.847319602966309, 'timing/data': 0.012100689928047359, 'timing/model': 1.150386821012944, 'learning_rate/qwen_vl_interface': 8.87278399417133e-06, 'learning_rate/action_model': 8.872783994171329e-05, 'epoch': 1.19}) +[2026-06-19 12:25:45,977][starVLA.training.train_starvla][INFO] - Step 1198, Loss: {'train/loss': 0.2806161046028137, 'train/grad_norm_pre_clip': 6.045248508453369, 'timing/data': 0.011866732966154814, 'timing/model': 1.1534569580107927, 'learning_rate/qwen_vl_interface': 8.870813584974852e-06, 'learning_rate/action_model': 8.870813584974852e-05, 'epoch': 1.19}) +[2026-06-19 12:25:50,605][starVLA.training.train_starvla][INFO] - Step 1199, Loss: {'train/loss': 0.30064234137535095, 'train/grad_norm_pre_clip': 12.647455215454102, 'timing/data': 0.01274573802947998, 'timing/model': 1.3244684339733794, 'learning_rate/qwen_vl_interface': 8.868841687398683e-06, 'learning_rate/action_model': 8.868841687398683e-05, 'epoch': 1.2}) +[2026-06-19 12:25:55,159][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1200 +[2026-06-19 12:28:48,705][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1200 +[2026-06-19 12:28:48,705][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 1200 +[2026-06-19 12:31:56,233][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 1200 +[2026-06-19 12:31:56,234][starVLA.training.train_starvla][INFO] - Step 1200, Loss: {'train/loss': 0.20545826852321625, 'train/grad_norm_pre_clip': 7.454695224761963, 'eval/loss': 0.4226562976837158, 'eval/latency_0/loss': 0.10134348273277283, 'eval/latency_1/loss': 0.15147651731967926, 'eval/latency_2/loss': 0.5419929027557373, 'eval/latency_3/loss': 0.5290266275405884, 'eval/latency_4/loss': 0.773629367351532, 'eval/flappy/loss': 0.4226562976837158, 'eval/flappy/latency_0/loss': 0.10134348273277283, 'eval/flappy/latency_1/loss': 0.15147651731967926, 'eval/flappy/latency_2/loss': 0.5419929027557373, 'eval/flappy/latency_3/loss': 0.5290266275405884, 'eval/flappy/latency_4/loss': 0.773629367351532, 'eval/flap_precision': 0.5550351288056206, 'eval/flap_recall': 0.8618181818181818, 'eval/flap_f1': 0.6752136752136751, 'eval/flap_rate_model': 0.11861111111111111, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.6752136752136751, 'eval/latency_0/flap_precision': 0.5550351288056206, 'eval/latency_0/flap_recall': 0.8618181818181818, 'eval/latency_0/flap_f1': 0.6752136752136751, 'eval/latency_0/flap_rate_model': 0.11861111111111111, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.6752136752136751, 'timing/data': 0.012111244956031442, 'timing/model': 1.2660254490328953, 'learning_rate/qwen_vl_interface': 8.86686830225339e-06, 'learning_rate/action_model': 8.866868302253389e-05, 'epoch': 1.2}) +[2026-06-19 12:31:56,235][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_1200_state +[2026-06-19 12:31:56,237][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-19 12:31:59,501][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_1200_state/model.safetensors +[2026-06-19 12:32:06,616][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_1200_state/optimizer.bin +[2026-06-19 12:32:06,617][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_1200_state/sampler.bin +[2026-06-19 12:32:06,617][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_1200_state/sampler_1.bin +[2026-06-19 12:32:06,618][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_1200_state/random_states_0.pkl +[2026-06-19 12:32:06,653][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1200 +[2026-06-19 12:32:06,653][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-19 12:32:06,674][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-19 12:32:11,081][starVLA.training.train_starvla][INFO] - Step 1201, Loss: {'train/loss': 0.16120201349258423, 'train/grad_norm_pre_clip': 5.905825614929199, 'timing/data': 0.016259449999779463, 'timing/model': 1.1282624789746478, 'learning_rate/qwen_vl_interface': 8.864893430350159e-06, 'learning_rate/action_model': 8.864893430350159e-05, 'epoch': 1.2}) +[2026-06-19 12:32:15,561][starVLA.training.train_starvla][INFO] - Step 1202, Loss: {'train/loss': 0.3738120198249817, 'train/grad_norm_pre_clip': 9.044316291809082, 'timing/data': 0.012795916991308331, 'timing/model': 1.2586405709153041, 'learning_rate/qwen_vl_interface': 8.862917072500785e-06, 'learning_rate/action_model': 8.862917072500784e-05, 'epoch': 1.2}) +[2026-06-19 12:32:19,927][starVLA.training.train_starvla][INFO] - Step 1203, Loss: {'train/loss': 0.37553977966308594, 'train/grad_norm_pre_clip': 7.332846641540527, 'timing/data': 0.012677715974859893, 'timing/model': 1.138028142042458, 'learning_rate/qwen_vl_interface': 8.860939229517674e-06, 'learning_rate/action_model': 8.860939229517674e-05, 'epoch': 1.2}) +[2026-06-19 12:32:24,452][starVLA.training.train_starvla][INFO] - Step 1204, Loss: {'train/loss': 0.230415478348732, 'train/grad_norm_pre_clip': 6.24512243270874, 'timing/data': 0.012308679986745119, 'timing/model': 1.127405826933682, 'learning_rate/qwen_vl_interface': 8.858959902213844e-06, 'learning_rate/action_model': 8.858959902213845e-05, 'epoch': 1.2}) +[2026-06-19 12:32:29,013][starVLA.training.train_starvla][INFO] - Step 1205, Loss: {'train/loss': 0.3781488537788391, 'train/grad_norm_pre_clip': 9.891733169555664, 'timing/data': 0.012360816937871277, 'timing/model': 1.1407287659822032, 'learning_rate/qwen_vl_interface': 8.856979091402921e-06, 'learning_rate/action_model': 8.85697909140292e-05, 'epoch': 1.2}) +[2026-06-19 12:32:33,506][starVLA.training.train_starvla][INFO] - Step 1206, Loss: {'train/loss': 0.24691654741764069, 'train/grad_norm_pre_clip': 5.161568641662598, 'timing/data': 0.012617369997315109, 'timing/model': 1.1424602790502831, 'learning_rate/qwen_vl_interface': 8.854996797899139e-06, 'learning_rate/action_model': 8.854996797899139e-05, 'epoch': 1.2}) +[2026-06-19 12:32:38,070][starVLA.training.train_starvla][INFO] - Step 1207, Loss: {'train/loss': 0.37602442502975464, 'train/grad_norm_pre_clip': 9.963595390319824, 'timing/data': 0.012275562970899045, 'timing/model': 1.1398509130813181, 'learning_rate/qwen_vl_interface': 8.853013022517348e-06, 'learning_rate/action_model': 8.853013022517348e-05, 'epoch': 1.2}) +[2026-06-19 12:32:42,633][starVLA.training.train_starvla][INFO] - Step 1208, Loss: {'train/loss': 0.2339421510696411, 'train/grad_norm_pre_clip': 3.9334323406219482, 'timing/data': 0.013086460065096617, 'timing/model': 1.3029362969100475, 'learning_rate/qwen_vl_interface': 8.851027766073002e-06, 'learning_rate/action_model': 8.851027766073001e-05, 'epoch': 1.2}) +[2026-06-19 12:32:47,104][starVLA.training.train_starvla][INFO] - Step 1209, Loss: {'train/loss': 0.22693093121051788, 'train/grad_norm_pre_clip': 5.900942802429199, 'timing/data': 0.012345836032181978, 'timing/model': 1.2203850229270756, 'learning_rate/qwen_vl_interface': 8.84904102938216e-06, 'learning_rate/action_model': 8.84904102938216e-05, 'epoch': 1.21}) +[2026-06-19 12:32:51,517][starVLA.training.train_starvla][INFO] - Step 1210, Loss: {'train/loss': 0.2723175585269928, 'train/grad_norm_pre_clip': 5.771205425262451, 'timing/data': 0.014709164039231837, 'timing/model': 1.1524532930925488, 'learning_rate/qwen_vl_interface': 8.8470528132615e-06, 'learning_rate/action_model': 8.8470528132615e-05, 'epoch': 1.21}) +[2026-06-19 12:32:56,062][starVLA.training.train_starvla][INFO] - Step 1211, Loss: {'train/loss': 0.17436087131500244, 'train/grad_norm_pre_clip': 4.973488807678223, 'timing/data': 0.01212170603685081, 'timing/model': 1.142423460027203, 'learning_rate/qwen_vl_interface': 8.845063118528303e-06, 'learning_rate/action_model': 8.845063118528304e-05, 'epoch': 1.21}) +[2026-06-19 12:33:00,656][starVLA.training.train_starvla][INFO] - Step 1212, Loss: {'train/loss': 0.11344042420387268, 'train/grad_norm_pre_clip': 3.681727170944214, 'timing/data': 0.012875197920948267, 'timing/model': 1.1487757950089872, 'learning_rate/qwen_vl_interface': 8.843071946000455e-06, 'learning_rate/action_model': 8.843071946000456e-05, 'epoch': 1.21}) +[2026-06-19 12:33:05,251][starVLA.training.train_starvla][INFO] - Step 1213, Loss: {'train/loss': 0.26169219613075256, 'train/grad_norm_pre_clip': 10.766144752502441, 'timing/data': 0.012989440001547337, 'timing/model': 1.1531664039939642, 'learning_rate/qwen_vl_interface': 8.841079296496451e-06, 'learning_rate/action_model': 8.84107929649645e-05, 'epoch': 1.21}) +[2026-06-19 12:33:09,760][starVLA.training.train_starvla][INFO] - Step 1214, Loss: {'train/loss': 0.301095187664032, 'train/grad_norm_pre_clip': 4.839930057525635, 'timing/data': 0.013022952945902944, 'timing/model': 1.135015859035775, 'learning_rate/qwen_vl_interface': 8.839085170835396e-06, 'learning_rate/action_model': 8.839085170835395e-05, 'epoch': 1.21}) +[2026-06-19 12:33:14,360][starVLA.training.train_starvla][INFO] - Step 1215, Loss: {'train/loss': 0.3066042959690094, 'train/grad_norm_pre_clip': 10.207146644592285, 'timing/data': 0.015424356097355485, 'timing/model': 1.3331319980788976, 'learning_rate/qwen_vl_interface': 8.837089569837e-06, 'learning_rate/action_model': 8.837089569836998e-05, 'epoch': 1.21}) +[2026-06-19 12:33:18,797][starVLA.training.train_starvla][INFO] - Step 1216, Loss: {'train/loss': 0.20652711391448975, 'train/grad_norm_pre_clip': 6.65254545211792, 'timing/data': 0.012540664058178663, 'timing/model': 1.1555685360217467, 'learning_rate/qwen_vl_interface': 8.835092494321575e-06, 'learning_rate/action_model': 8.835092494321575e-05, 'epoch': 1.21}) +[2026-06-19 12:33:23,405][starVLA.training.train_starvla][INFO] - Step 1217, Loss: {'train/loss': 0.32289934158325195, 'train/grad_norm_pre_clip': 7.596785545349121, 'timing/data': 0.013002934982068837, 'timing/model': 1.1670248119626194, 'learning_rate/qwen_vl_interface': 8.833093945110048e-06, 'learning_rate/action_model': 8.833093945110047e-05, 'epoch': 1.21}) +[2026-06-19 12:33:28,009][starVLA.training.train_starvla][INFO] - Step 1218, Loss: {'train/loss': 0.34846827387809753, 'train/grad_norm_pre_clip': 7.783318519592285, 'timing/data': 0.013001388986594975, 'timing/model': 1.1493722120067105, 'learning_rate/qwen_vl_interface': 8.831093923023946e-06, 'learning_rate/action_model': 8.831093923023945e-05, 'epoch': 1.21}) +[2026-06-19 12:33:32,618][starVLA.training.train_starvla][INFO] - Step 1219, Loss: {'train/loss': 0.11615623533725739, 'train/grad_norm_pre_clip': 3.3990848064422607, 'timing/data': 0.012294895015656948, 'timing/model': 1.1533432729775086, 'learning_rate/qwen_vl_interface': 8.8290924288854e-06, 'learning_rate/action_model': 8.8290924288854e-05, 'epoch': 1.22}) +[2026-06-19 12:33:37,203][starVLA.training.train_starvla][INFO] - Step 1220, Loss: {'train/loss': 0.23577998578548431, 'train/grad_norm_pre_clip': 4.250526428222656, 'timing/data': 0.013269355054944754, 'timing/model': 1.2996690049767494, 'learning_rate/qwen_vl_interface': 8.827089463517153e-06, 'learning_rate/action_model': 8.827089463517153e-05, 'epoch': 1.22}) +[2026-06-19 12:33:41,602][starVLA.training.train_starvla][INFO] - Step 1221, Loss: {'train/loss': 0.356985867023468, 'train/grad_norm_pre_clip': 8.275629043579102, 'timing/data': 0.012984544970095158, 'timing/model': 1.1373593569733202, 'learning_rate/qwen_vl_interface': 8.825085027742545e-06, 'learning_rate/action_model': 8.825085027742544e-05, 'epoch': 1.22}) +[2026-06-19 12:33:46,180][starVLA.training.train_starvla][INFO] - Step 1222, Loss: {'train/loss': 0.3970959186553955, 'train/grad_norm_pre_clip': 9.879996299743652, 'timing/data': 0.01307605195324868, 'timing/model': 1.1411435910267755, 'learning_rate/qwen_vl_interface': 8.823079122385523e-06, 'learning_rate/action_model': 8.823079122385523e-05, 'epoch': 1.22}) +[2026-06-19 12:33:50,757][starVLA.training.train_starvla][INFO] - Step 1223, Loss: {'train/loss': 0.3501640856266022, 'train/grad_norm_pre_clip': 6.041560173034668, 'timing/data': 0.012215637019835413, 'timing/model': 1.155727240955457, 'learning_rate/qwen_vl_interface': 8.821071748270644e-06, 'learning_rate/action_model': 8.821071748270644e-05, 'epoch': 1.22}) +[2026-06-19 12:33:55,377][starVLA.training.train_starvla][INFO] - Step 1224, Loss: {'train/loss': 0.0610186904668808, 'train/grad_norm_pre_clip': 2.5667965412139893, 'timing/data': 0.012428606045432389, 'timing/model': 1.151669950922951, 'learning_rate/qwen_vl_interface': 8.81906290622306e-06, 'learning_rate/action_model': 8.81906290622306e-05, 'epoch': 1.22}) +[2026-06-19 12:33:59,983][starVLA.training.train_starvla][INFO] - Step 1225, Loss: {'train/loss': 0.22147557139396667, 'train/grad_norm_pre_clip': 9.131704330444336, 'timing/data': 0.012890963000245392, 'timing/model': 1.142642347025685, 'learning_rate/qwen_vl_interface': 8.81705259706853e-06, 'learning_rate/action_model': 8.817052597068529e-05, 'epoch': 1.22}) +[2026-06-19 12:34:04,593][starVLA.training.train_starvla][INFO] - Step 1226, Loss: {'train/loss': 0.17597240209579468, 'train/grad_norm_pre_clip': 4.671793460845947, 'timing/data': 0.013167383032850921, 'timing/model': 1.3364300009561703, 'learning_rate/qwen_vl_interface': 8.815040821633416e-06, 'learning_rate/action_model': 8.815040821633416e-05, 'epoch': 1.22}) +[2026-06-19 12:34:09,214][starVLA.training.train_starvla][INFO] - Step 1227, Loss: {'train/loss': 0.20178565382957458, 'train/grad_norm_pre_clip': 6.736572265625, 'timing/data': 0.012519550044089556, 'timing/model': 1.3270209670299664, 'learning_rate/qwen_vl_interface': 8.813027580744685e-06, 'learning_rate/action_model': 8.813027580744684e-05, 'epoch': 1.22}) +[2026-06-19 12:34:13,617][starVLA.training.train_starvla][INFO] - Step 1228, Loss: {'train/loss': 0.36022695899009705, 'train/grad_norm_pre_clip': 11.70848560333252, 'timing/data': 0.01330798794515431, 'timing/model': 1.1417850019643083, 'learning_rate/qwen_vl_interface': 8.8110128752299e-06, 'learning_rate/action_model': 8.8110128752299e-05, 'epoch': 1.22}) +[2026-06-19 12:34:18,206][starVLA.training.train_starvla][INFO] - Step 1229, Loss: {'train/loss': 0.1177143007516861, 'train/grad_norm_pre_clip': 6.287220478057861, 'timing/data': 0.012083264999091625, 'timing/model': 1.1656708510126919, 'learning_rate/qwen_vl_interface': 8.808996705917235e-06, 'learning_rate/action_model': 8.808996705917235e-05, 'epoch': 1.23}) +[2026-06-19 12:34:22,837][starVLA.training.train_starvla][INFO] - Step 1230, Loss: {'train/loss': 0.24835054576396942, 'train/grad_norm_pre_clip': 7.833001136779785, 'timing/data': 0.013102569035254419, 'timing/model': 1.1507436410756782, 'learning_rate/qwen_vl_interface': 8.80697907363546e-06, 'learning_rate/action_model': 8.80697907363546e-05, 'epoch': 1.23}) +[2026-06-19 12:34:27,474][starVLA.training.train_starvla][INFO] - Step 1231, Loss: {'train/loss': 0.2664806544780731, 'train/grad_norm_pre_clip': 7.0672926902771, 'timing/data': 0.012756103998981416, 'timing/model': 1.163151508080773, 'learning_rate/qwen_vl_interface': 8.804959979213945e-06, 'learning_rate/action_model': 8.804959979213945e-05, 'epoch': 1.23}) +[2026-06-19 12:34:32,071][starVLA.training.train_starvla][INFO] - Step 1232, Loss: {'train/loss': 0.4323975145816803, 'train/grad_norm_pre_clip': 12.205007553100586, 'timing/data': 0.012818622984923422, 'timing/model': 1.322353065945208, 'learning_rate/qwen_vl_interface': 8.802939423482667e-06, 'learning_rate/action_model': 8.802939423482666e-05, 'epoch': 1.23}) +[2026-06-19 12:34:36,672][starVLA.training.train_starvla][INFO] - Step 1233, Loss: {'train/loss': 0.36903494596481323, 'train/grad_norm_pre_clip': 13.63418197631836, 'timing/data': 0.012224380974657834, 'timing/model': 1.3282999249640852, 'learning_rate/qwen_vl_interface': 8.800917407272198e-06, 'learning_rate/action_model': 8.800917407272198e-05, 'epoch': 1.23}) +[2026-06-19 12:34:41,091][starVLA.training.train_starvla][INFO] - Step 1234, Loss: {'train/loss': 0.28059688210487366, 'train/grad_norm_pre_clip': 9.40075969696045, 'timing/data': 0.012870764010585845, 'timing/model': 1.1488767869304866, 'learning_rate/qwen_vl_interface': 8.798893931413711e-06, 'learning_rate/action_model': 8.79889393141371e-05, 'epoch': 1.23}) +[2026-06-19 12:34:45,668][starVLA.training.train_starvla][INFO] - Step 1235, Loss: {'train/loss': 0.22275611758232117, 'train/grad_norm_pre_clip': 7.902650356292725, 'timing/data': 0.012432030984200537, 'timing/model': 1.144853907986544, 'learning_rate/qwen_vl_interface': 8.796868996738985e-06, 'learning_rate/action_model': 8.796868996738984e-05, 'epoch': 1.23}) +[2026-06-19 12:34:50,302][starVLA.training.train_starvla][INFO] - Step 1236, Loss: {'train/loss': 0.20932693779468536, 'train/grad_norm_pre_clip': 6.274613857269287, 'timing/data': 0.013868186040781438, 'timing/model': 1.142749728052877, 'learning_rate/qwen_vl_interface': 8.794842604080389e-06, 'learning_rate/action_model': 8.794842604080389e-05, 'epoch': 1.23}) +[2026-06-19 12:34:54,895][starVLA.training.train_starvla][INFO] - Step 1237, Loss: {'train/loss': 0.24182836711406708, 'train/grad_norm_pre_clip': 7.044676303863525, 'timing/data': 0.013910082983784378, 'timing/model': 1.1423256929265335, 'learning_rate/qwen_vl_interface': 8.7928147542709e-06, 'learning_rate/action_model': 8.7928147542709e-05, 'epoch': 1.23}) +[2026-06-19 12:34:59,477][starVLA.training.train_starvla][INFO] - Step 1238, Loss: {'train/loss': 0.09240935742855072, 'train/grad_norm_pre_clip': 5.340156078338623, 'timing/data': 0.013303617015480995, 'timing/model': 1.2909521770197898, 'learning_rate/qwen_vl_interface': 8.790785448144088e-06, 'learning_rate/action_model': 8.790785448144088e-05, 'epoch': 1.23}) +[2026-06-19 12:35:04,097][starVLA.training.train_starvla][INFO] - Step 1239, Loss: {'train/loss': 0.4904513359069824, 'train/grad_norm_pre_clip': 10.559032440185547, 'timing/data': 0.013126190984621644, 'timing/model': 1.3310874269809574, 'learning_rate/qwen_vl_interface': 8.788754686534126e-06, 'learning_rate/action_model': 8.788754686534127e-05, 'epoch': 1.24}) +[2026-06-19 12:35:08,527][starVLA.training.train_starvla][INFO] - Step 1240, Loss: {'train/loss': 0.1498584747314453, 'train/grad_norm_pre_clip': 4.710055351257324, 'timing/data': 0.014668955933302641, 'timing/model': 1.1555682759499177, 'learning_rate/qwen_vl_interface': 8.786722470275784e-06, 'learning_rate/action_model': 8.786722470275784e-05, 'epoch': 1.24}) +[2026-06-19 12:35:13,142][starVLA.training.train_starvla][INFO] - Step 1241, Loss: {'train/loss': 0.7612923383712769, 'train/grad_norm_pre_clip': 14.220197677612305, 'timing/data': 0.012702164007350802, 'timing/model': 1.151110231061466, 'learning_rate/qwen_vl_interface': 8.784688800204427e-06, 'learning_rate/action_model': 8.784688800204426e-05, 'epoch': 1.24}) +[2026-06-19 12:35:17,738][starVLA.training.train_starvla][INFO] - Step 1242, Loss: {'train/loss': 0.46009716391563416, 'train/grad_norm_pre_clip': 16.69182777404785, 'timing/data': 0.01375842490233481, 'timing/model': 1.1450666229939088, 'learning_rate/qwen_vl_interface': 8.782653677156022e-06, 'learning_rate/action_model': 8.782653677156022e-05, 'epoch': 1.24}) +[2026-06-19 12:35:22,339][starVLA.training.train_starvla][INFO] - Step 1243, Loss: {'train/loss': 0.2227022349834442, 'train/grad_norm_pre_clip': 7.318717956542969, 'timing/data': 0.012790811946615577, 'timing/model': 1.1570142110576853, 'learning_rate/qwen_vl_interface': 8.780617101967129e-06, 'learning_rate/action_model': 8.780617101967129e-05, 'epoch': 1.24}) +[2026-06-19 12:35:26,930][starVLA.training.train_starvla][INFO] - Step 1244, Loss: {'train/loss': 0.4393062889575958, 'train/grad_norm_pre_clip': 5.0578813552856445, 'timing/data': 0.013377538067288697, 'timing/model': 1.1475781600456685, 'learning_rate/qwen_vl_interface': 8.77857907547491e-06, 'learning_rate/action_model': 8.77857907547491e-05, 'epoch': 1.24}) +[2026-06-19 12:35:31,501][starVLA.training.train_starvla][INFO] - Step 1245, Loss: {'train/loss': 0.3980768024921417, 'train/grad_norm_pre_clip': 5.9137115478515625, 'timing/data': 0.012736140983179212, 'timing/model': 1.3003855539718643, 'learning_rate/qwen_vl_interface': 8.776539598517119e-06, 'learning_rate/action_model': 8.776539598517118e-05, 'epoch': 1.24}) +[2026-06-19 12:35:35,948][starVLA.training.train_starvla][INFO] - Step 1246, Loss: {'train/loss': 0.18472030758857727, 'train/grad_norm_pre_clip': 3.392324924468994, 'timing/data': 0.012796115013770759, 'timing/model': 1.1510104991029948, 'learning_rate/qwen_vl_interface': 8.774498671932108e-06, 'learning_rate/action_model': 8.774498671932107e-05, 'epoch': 1.24}) +[2026-06-19 12:35:40,584][starVLA.training.train_starvla][INFO] - Step 1247, Loss: {'train/loss': 0.17007741332054138, 'train/grad_norm_pre_clip': 2.5137710571289062, 'timing/data': 0.013071601977571845, 'timing/model': 1.172612338908948, 'learning_rate/qwen_vl_interface': 8.772456296558824e-06, 'learning_rate/action_model': 8.772456296558824e-05, 'epoch': 1.24}) +[2026-06-19 12:35:45,213][starVLA.training.train_starvla][INFO] - Step 1248, Loss: {'train/loss': 0.14847096800804138, 'train/grad_norm_pre_clip': 4.3487935066223145, 'timing/data': 0.012647352064959705, 'timing/model': 1.1596426060423255, 'learning_rate/qwen_vl_interface': 8.77041247323681e-06, 'learning_rate/action_model': 8.770412473236811e-05, 'epoch': 1.24}) +[2026-06-19 12:35:49,829][starVLA.training.train_starvla][INFO] - Step 1249, Loss: {'train/loss': 0.21452204883098602, 'train/grad_norm_pre_clip': 8.538260459899902, 'timing/data': 0.012605258030816913, 'timing/model': 1.148747717961669, 'learning_rate/qwen_vl_interface': 8.76836720280621e-06, 'learning_rate/action_model': 8.768367202806209e-05, 'epoch': 1.25}) +[2026-06-19 12:35:54,417][starVLA.training.train_starvla][INFO] - Step 1250, Loss: {'train/loss': 0.31522423028945923, 'train/grad_norm_pre_clip': 5.3841047286987305, 'timing/data': 0.013349923072382808, 'timing/model': 1.1477392740780488, 'learning_rate/qwen_vl_interface': 8.766320486107749e-06, 'learning_rate/action_model': 8.766320486107749e-05, 'epoch': 1.25}) +[2026-06-19 12:35:59,003][starVLA.training.train_starvla][INFO] - Step 1251, Loss: {'train/loss': 0.19889962673187256, 'train/grad_norm_pre_clip': 3.3106536865234375, 'timing/data': 0.012853232910856605, 'timing/model': 1.3053079809760675, 'learning_rate/qwen_vl_interface': 8.76427232398276e-06, 'learning_rate/action_model': 8.76427232398276e-05, 'epoch': 1.25}) +[2026-06-19 12:36:03,514][starVLA.training.train_starvla][INFO] - Step 1252, Loss: {'train/loss': 0.18250614404678345, 'train/grad_norm_pre_clip': 4.515142440795898, 'timing/data': 0.013379318988882005, 'timing/model': 1.2294550969963893, 'learning_rate/qwen_vl_interface': 8.762222717273167e-06, 'learning_rate/action_model': 8.762222717273167e-05, 'epoch': 1.25}) +[2026-06-19 12:36:07,945][starVLA.training.train_starvla][INFO] - Step 1253, Loss: {'train/loss': 0.07440225780010223, 'train/grad_norm_pre_clip': 2.208099603652954, 'timing/data': 0.012954638921655715, 'timing/model': 1.151831367984414, 'learning_rate/qwen_vl_interface': 8.76017166682148e-06, 'learning_rate/action_model': 8.760171666821481e-05, 'epoch': 1.25}) +[2026-06-19 12:36:12,551][starVLA.training.train_starvla][INFO] - Step 1254, Loss: {'train/loss': 0.51545649766922, 'train/grad_norm_pre_clip': 8.306153297424316, 'timing/data': 0.012601154972799122, 'timing/model': 1.1503330930136144, 'learning_rate/qwen_vl_interface': 8.758119173470815e-06, 'learning_rate/action_model': 8.758119173470815e-05, 'epoch': 1.25}) +[2026-06-19 12:36:17,146][starVLA.training.train_starvla][INFO] - Step 1255, Loss: {'train/loss': 0.5050718784332275, 'train/grad_norm_pre_clip': 10.617683410644531, 'timing/data': 0.012982362997718155, 'timing/model': 1.1527406319510192, 'learning_rate/qwen_vl_interface': 8.75606523806487e-06, 'learning_rate/action_model': 8.756065238064868e-05, 'epoch': 1.25}) +[2026-06-19 12:36:21,747][starVLA.training.train_starvla][INFO] - Step 1256, Loss: {'train/loss': 0.29458731412887573, 'train/grad_norm_pre_clip': 6.735457897186279, 'timing/data': 0.013439352042041719, 'timing/model': 1.1508508579572663, 'learning_rate/qwen_vl_interface': 8.75400986144794e-06, 'learning_rate/action_model': 8.75400986144794e-05, 'epoch': 1.25}) +[2026-06-19 12:36:26,359][starVLA.training.train_starvla][INFO] - Step 1257, Loss: {'train/loss': 0.18907280266284943, 'train/grad_norm_pre_clip': 7.650333881378174, 'timing/data': 0.012959318002685905, 'timing/model': 1.310768537921831, 'learning_rate/qwen_vl_interface': 8.751953044464913e-06, 'learning_rate/action_model': 8.751953044464913e-05, 'epoch': 1.25}) +[2026-06-19 12:36:30,865][starVLA.training.train_starvla][INFO] - Step 1258, Loss: {'train/loss': 0.23388832807540894, 'train/grad_norm_pre_clip': 5.1995649337768555, 'timing/data': 0.012765140039846301, 'timing/model': 1.226748613989912, 'learning_rate/qwen_vl_interface': 8.749894787961272e-06, 'learning_rate/action_model': 8.749894787961271e-05, 'epoch': 1.25}) +[2026-06-19 12:36:35,295][starVLA.training.train_starvla][INFO] - Step 1259, Loss: {'train/loss': 0.12331239879131317, 'train/grad_norm_pre_clip': 3.43058180809021, 'timing/data': 0.012529127998277545, 'timing/model': 1.1546029549790546, 'learning_rate/qwen_vl_interface': 8.747835092783086e-06, 'learning_rate/action_model': 8.747835092783085e-05, 'epoch': 1.26}) +[2026-06-19 12:36:39,843][starVLA.training.train_starvla][INFO] - Step 1260, Loss: {'train/loss': 0.12135718762874603, 'train/grad_norm_pre_clip': 3.6779706478118896, 'timing/data': 0.013584709027782083, 'timing/model': 1.1502309329807758, 'learning_rate/qwen_vl_interface': 8.745773959777018e-06, 'learning_rate/action_model': 8.745773959777017e-05, 'epoch': 1.26}) +[2026-06-19 12:36:44,450][starVLA.training.train_starvla][INFO] - Step 1261, Loss: {'train/loss': 0.2288476824760437, 'train/grad_norm_pre_clip': 4.638607501983643, 'timing/data': 0.012956344988197088, 'timing/model': 1.1457025030395016, 'learning_rate/qwen_vl_interface': 8.743711389790321e-06, 'learning_rate/action_model': 8.74371138979032e-05, 'epoch': 1.26}) +[2026-06-19 12:36:49,030][starVLA.training.train_starvla][INFO] - Step 1262, Loss: {'train/loss': 0.3326617181301117, 'train/grad_norm_pre_clip': 9.755093574523926, 'timing/data': 0.013277336023747921, 'timing/model': 1.1504099840531126, 'learning_rate/qwen_vl_interface': 8.741647383670839e-06, 'learning_rate/action_model': 8.741647383670839e-05, 'epoch': 1.26}) +[2026-06-19 12:36:53,626][starVLA.training.train_starvla][INFO] - Step 1263, Loss: {'train/loss': 0.24740362167358398, 'train/grad_norm_pre_clip': 7.785251617431641, 'timing/data': 0.0128903390141204, 'timing/model': 1.324677346041426, 'learning_rate/qwen_vl_interface': 8.739581942267011e-06, 'learning_rate/action_model': 8.739581942267011e-05, 'epoch': 1.26}) +[2026-06-19 12:36:58,046][starVLA.training.train_starvla][INFO] - Step 1264, Loss: {'train/loss': 0.31234797835350037, 'train/grad_norm_pre_clip': 8.526668548583984, 'timing/data': 0.013240159954875708, 'timing/model': 1.145791540038772, 'learning_rate/qwen_vl_interface': 8.737515066427861e-06, 'learning_rate/action_model': 8.73751506642786e-05, 'epoch': 1.26}) +[2026-06-19 12:37:02,633][starVLA.training.train_starvla][INFO] - Step 1265, Loss: {'train/loss': 0.2815049886703491, 'train/grad_norm_pre_clip': 9.682353973388672, 'timing/data': 0.012708381982520223, 'timing/model': 1.1437936349539086, 'learning_rate/qwen_vl_interface': 8.735446757003e-06, 'learning_rate/action_model': 8.735446757003e-05, 'epoch': 1.26}) +[2026-06-19 12:37:07,167][starVLA.training.train_starvla][INFO] - Step 1266, Loss: {'train/loss': 0.24427835643291473, 'train/grad_norm_pre_clip': 5.144277572631836, 'timing/data': 0.012916511972434819, 'timing/model': 1.1655786250485107, 'learning_rate/qwen_vl_interface': 8.733377014842635e-06, 'learning_rate/action_model': 8.733377014842635e-05, 'epoch': 1.26}) +[2026-06-19 12:37:11,756][starVLA.training.train_starvla][INFO] - Step 1267, Loss: {'train/loss': 0.19551964104175568, 'train/grad_norm_pre_clip': 5.439601898193359, 'timing/data': 0.012867418001405895, 'timing/model': 1.1456713990774006, 'learning_rate/qwen_vl_interface': 8.731305840797557e-06, 'learning_rate/action_model': 8.731305840797558e-05, 'epoch': 1.26}) +[2026-06-19 12:37:16,359][starVLA.training.train_starvla][INFO] - Step 1268, Loss: {'train/loss': 0.11106832325458527, 'train/grad_norm_pre_clip': 8.859185218811035, 'timing/data': 0.01274133799597621, 'timing/model': 1.1587016510311514, 'learning_rate/qwen_vl_interface': 8.72923323571915e-06, 'learning_rate/action_model': 8.72923323571915e-05, 'epoch': 1.26}) +[2026-06-19 12:37:20,964][starVLA.training.train_starvla][INFO] - Step 1269, Loss: {'train/loss': 0.5774461030960083, 'train/grad_norm_pre_clip': 9.52122688293457, 'timing/data': 0.012696092948317528, 'timing/model': 1.3172291610389948, 'learning_rate/qwen_vl_interface': 8.727159200459381e-06, 'learning_rate/action_model': 8.727159200459381e-05, 'epoch': 1.27}) +[2026-06-19 12:37:25,413][starVLA.training.train_starvla][INFO] - Step 1270, Loss: {'train/loss': 0.153122678399086, 'train/grad_norm_pre_clip': 3.3889577388763428, 'timing/data': 0.013533344957977533, 'timing/model': 1.1659888250287622, 'learning_rate/qwen_vl_interface': 8.725083735870807e-06, 'learning_rate/action_model': 8.725083735870807e-05, 'epoch': 1.27}) +[2026-06-19 12:37:30,013][starVLA.training.train_starvla][INFO] - Step 1271, Loss: {'train/loss': 0.2644892930984497, 'train/grad_norm_pre_clip': 5.319128036499023, 'timing/data': 0.013292126008309424, 'timing/model': 1.1440352590288967, 'learning_rate/qwen_vl_interface': 8.723006842806575e-06, 'learning_rate/action_model': 8.723006842806576e-05, 'epoch': 1.27}) +[2026-06-19 12:37:34,602][starVLA.training.train_starvla][INFO] - Step 1272, Loss: {'train/loss': 0.30777230858802795, 'train/grad_norm_pre_clip': 10.757728576660156, 'timing/data': 0.012922016088850796, 'timing/model': 1.14909582096152, 'learning_rate/qwen_vl_interface': 8.720928522120419e-06, 'learning_rate/action_model': 8.720928522120418e-05, 'epoch': 1.27}) +[2026-06-19 12:37:39,108][starVLA.training.train_starvla][INFO] - Step 1273, Loss: {'train/loss': 0.227349191904068, 'train/grad_norm_pre_clip': 13.197555541992188, 'timing/data': 0.012950327014550567, 'timing/model': 1.1546258450252935, 'learning_rate/qwen_vl_interface': 8.718848774666655e-06, 'learning_rate/action_model': 8.718848774666655e-05, 'epoch': 1.27}) +[2026-06-19 12:37:43,678][starVLA.training.train_starvla][INFO] - Step 1274, Loss: {'train/loss': 0.19500373303890228, 'train/grad_norm_pre_clip': 3.9108104705810547, 'timing/data': 0.01327394403051585, 'timing/model': 1.1490381199400872, 'learning_rate/qwen_vl_interface': 8.716767601300189e-06, 'learning_rate/action_model': 8.716767601300188e-05, 'epoch': 1.27}) +[2026-06-19 12:37:48,302][starVLA.training.train_starvla][INFO] - Step 1275, Loss: {'train/loss': 0.22014635801315308, 'train/grad_norm_pre_clip': 6.667755126953125, 'timing/data': 0.014269359991885722, 'timing/model': 1.3311374239856377, 'learning_rate/qwen_vl_interface': 8.714685002876514e-06, 'learning_rate/action_model': 8.714685002876513e-05, 'epoch': 1.27}) +[2026-06-19 12:37:52,722][starVLA.training.train_starvla][INFO] - Step 1276, Loss: {'train/loss': 0.2639678716659546, 'train/grad_norm_pre_clip': 6.531498908996582, 'timing/data': 0.01366808800958097, 'timing/model': 1.145605513942428, 'learning_rate/qwen_vl_interface': 8.712600980251706e-06, 'learning_rate/action_model': 8.712600980251705e-05, 'epoch': 1.27}) +[2026-06-19 12:37:57,316][starVLA.training.train_starvla][INFO] - Step 1277, Loss: {'train/loss': 0.3134738504886627, 'train/grad_norm_pre_clip': 4.908405780792236, 'timing/data': 0.01264697895385325, 'timing/model': 1.1433587719220668, 'learning_rate/qwen_vl_interface': 8.710515534282428e-06, 'learning_rate/action_model': 8.710515534282428e-05, 'epoch': 1.27}) +[2026-06-19 12:38:01,902][starVLA.training.train_starvla][INFO] - Step 1278, Loss: {'train/loss': 0.2932260036468506, 'train/grad_norm_pre_clip': 8.753902435302734, 'timing/data': 0.01358241110574454, 'timing/model': 1.1443356510717422, 'learning_rate/qwen_vl_interface': 8.708428665825931e-06, 'learning_rate/action_model': 8.708428665825931e-05, 'epoch': 1.27}) +[2026-06-19 12:38:06,517][starVLA.training.train_starvla][INFO] - Step 1279, Loss: {'train/loss': 0.2335083931684494, 'train/grad_norm_pre_clip': 6.750784873962402, 'timing/data': 0.013432202977128327, 'timing/model': 1.1538154470035806, 'learning_rate/qwen_vl_interface': 8.706340375740045e-06, 'learning_rate/action_model': 8.706340375740045e-05, 'epoch': 1.28}) +[2026-06-19 12:38:11,102][starVLA.training.train_starvla][INFO] - Step 1280, Loss: {'train/loss': 0.14539407193660736, 'train/grad_norm_pre_clip': 4.513856887817383, 'timing/data': 0.013528511044569314, 'timing/model': 1.153354504960589, 'learning_rate/qwen_vl_interface': 8.704250664883188e-06, 'learning_rate/action_model': 8.704250664883188e-05, 'epoch': 1.28}) +[2026-06-19 12:38:15,689][starVLA.training.train_starvla][INFO] - Step 1281, Loss: {'train/loss': 0.2781449556350708, 'train/grad_norm_pre_clip': 5.45053768157959, 'timing/data': 0.012338358093984425, 'timing/model': 1.3035057201050222, 'learning_rate/qwen_vl_interface': 8.70215953411436e-06, 'learning_rate/action_model': 8.70215953411436e-05, 'epoch': 1.28}) +[2026-06-19 12:38:20,291][starVLA.training.train_starvla][INFO] - Step 1282, Loss: {'train/loss': 0.111765056848526, 'train/grad_norm_pre_clip': 2.8313257694244385, 'timing/data': 0.013791935984045267, 'timing/model': 1.3058167749550194, 'learning_rate/qwen_vl_interface': 8.70006698429315e-06, 'learning_rate/action_model': 8.700066984293149e-05, 'epoch': 1.28}) +[2026-06-19 12:38:24,740][starVLA.training.train_starvla][INFO] - Step 1283, Loss: {'train/loss': 0.20809626579284668, 'train/grad_norm_pre_clip': 5.871943950653076, 'timing/data': 0.012058516964316368, 'timing/model': 1.1546064510475844, 'learning_rate/qwen_vl_interface': 8.697973016279722e-06, 'learning_rate/action_model': 8.697973016279722e-05, 'epoch': 1.28}) +[2026-06-19 12:38:29,318][starVLA.training.train_starvla][INFO] - Step 1284, Loss: {'train/loss': 0.22771139442920685, 'train/grad_norm_pre_clip': 7.147581100463867, 'timing/data': 0.01330964092630893, 'timing/model': 1.1449658060446382, 'learning_rate/qwen_vl_interface': 8.695877630934827e-06, 'learning_rate/action_model': 8.695877630934826e-05, 'epoch': 1.28}) +[2026-06-19 12:38:33,920][starVLA.training.train_starvla][INFO] - Step 1285, Loss: {'train/loss': 0.1803291141986847, 'train/grad_norm_pre_clip': 4.176711559295654, 'timing/data': 0.01211429201066494, 'timing/model': 1.1528983030002564, 'learning_rate/qwen_vl_interface': 8.693780829119801e-06, 'learning_rate/action_model': 8.693780829119801e-05, 'epoch': 1.28}) +[2026-06-19 12:38:38,504][starVLA.training.train_starvla][INFO] - Step 1286, Loss: {'train/loss': 0.13620397448539734, 'train/grad_norm_pre_clip': 4.155150890350342, 'timing/data': 0.013013125979341567, 'timing/model': 1.1459431450348347, 'learning_rate/qwen_vl_interface': 8.69168261169656e-06, 'learning_rate/action_model': 8.691682611696559e-05, 'epoch': 1.28}) +[2026-06-19 12:38:43,111][starVLA.training.train_starvla][INFO] - Step 1287, Loss: {'train/loss': 0.592512309551239, 'train/grad_norm_pre_clip': 10.401095390319824, 'timing/data': 0.012494325987063348, 'timing/model': 1.151153079001233, 'learning_rate/qwen_vl_interface': 8.689582979527599e-06, 'learning_rate/action_model': 8.689582979527599e-05, 'epoch': 1.28}) +[2026-06-19 12:38:47,727][starVLA.training.train_starvla][INFO] - Step 1288, Loss: {'train/loss': 0.15270459651947021, 'train/grad_norm_pre_clip': 3.7558131217956543, 'timing/data': 0.012465176056139171, 'timing/model': 1.3305554080288857, 'learning_rate/qwen_vl_interface': 8.687481933476e-06, 'learning_rate/action_model': 8.687481933476e-05, 'epoch': 1.28}) +[2026-06-19 12:38:52,159][starVLA.training.train_starvla][INFO] - Step 1289, Loss: {'train/loss': 0.18156953155994415, 'train/grad_norm_pre_clip': 16.780040740966797, 'timing/data': 0.01252595498226583, 'timing/model': 1.1502851149998605, 'learning_rate/qwen_vl_interface': 8.685379474405423e-06, 'learning_rate/action_model': 8.685379474405423e-05, 'epoch': 1.29}) +[2026-06-19 12:38:56,758][starVLA.training.train_starvla][INFO] - Step 1290, Loss: {'train/loss': 0.23810018599033356, 'train/grad_norm_pre_clip': 4.885016918182373, 'timing/data': 0.013088266015984118, 'timing/model': 1.1580830960301682, 'learning_rate/qwen_vl_interface': 8.68327560318011e-06, 'learning_rate/action_model': 8.68327560318011e-05, 'epoch': 1.29}) +[2026-06-19 12:39:01,366][starVLA.training.train_starvla][INFO] - Step 1291, Loss: {'train/loss': 0.500868022441864, 'train/grad_norm_pre_clip': 10.044102668762207, 'timing/data': 0.012004046002402902, 'timing/model': 1.1492836750112474, 'learning_rate/qwen_vl_interface': 8.681170320664883e-06, 'learning_rate/action_model': 8.681170320664883e-05, 'epoch': 1.29}) +[2026-06-19 12:39:05,873][starVLA.training.train_starvla][INFO] - Step 1292, Loss: {'train/loss': 0.3646894097328186, 'train/grad_norm_pre_clip': 7.791281223297119, 'timing/data': 0.01385387300979346, 'timing/model': 1.1417616490507498, 'learning_rate/qwen_vl_interface': 8.679063627725143e-06, 'learning_rate/action_model': 8.679063627725144e-05, 'epoch': 1.29}) +[2026-06-19 12:39:10,412][starVLA.training.train_starvla][INFO] - Step 1293, Loss: {'train/loss': 0.23303447663784027, 'train/grad_norm_pre_clip': 5.566094875335693, 'timing/data': 0.012556012952700257, 'timing/model': 1.1533298500580713, 'learning_rate/qwen_vl_interface': 8.676955525226876e-06, 'learning_rate/action_model': 8.676955525226875e-05, 'epoch': 1.29}) +[2026-06-19 12:39:14,998][starVLA.training.train_starvla][INFO] - Step 1294, Loss: {'train/loss': 0.3384092152118683, 'train/grad_norm_pre_clip': 9.349958419799805, 'timing/data': 0.012211762019433081, 'timing/model': 1.3210828780429438, 'learning_rate/qwen_vl_interface': 8.674846014036636e-06, 'learning_rate/action_model': 8.674846014036636e-05, 'epoch': 1.29}) +[2026-06-19 12:39:19,416][starVLA.training.train_starvla][INFO] - Step 1295, Loss: {'train/loss': 0.11509709805250168, 'train/grad_norm_pre_clip': 4.11961030960083, 'timing/data': 0.013123102951794863, 'timing/model': 1.1476321750087664, 'learning_rate/qwen_vl_interface': 8.672735095021569e-06, 'learning_rate/action_model': 8.672735095021569e-05, 'epoch': 1.29}) +[2026-06-19 12:39:24,023][starVLA.training.train_starvla][INFO] - Step 1296, Loss: {'train/loss': 0.4654754400253296, 'train/grad_norm_pre_clip': 10.312508583068848, 'timing/data': 0.012866851058788598, 'timing/model': 1.149445439921692, 'learning_rate/qwen_vl_interface': 8.670622769049393e-06, 'learning_rate/action_model': 8.670622769049392e-05, 'epoch': 1.29}) +[2026-06-19 12:39:28,614][starVLA.training.train_starvla][INFO] - Step 1297, Loss: {'train/loss': 0.26752108335494995, 'train/grad_norm_pre_clip': 6.759419918060303, 'timing/data': 0.013414330896921456, 'timing/model': 1.1497437899233773, 'learning_rate/qwen_vl_interface': 8.668509036988406e-06, 'learning_rate/action_model': 8.668509036988405e-05, 'epoch': 1.29}) +[2026-06-19 12:39:33,223][starVLA.training.train_starvla][INFO] - Step 1298, Loss: {'train/loss': 0.4537748396396637, 'train/grad_norm_pre_clip': 10.79835033416748, 'timing/data': 0.013726677978411317, 'timing/model': 1.1459430110407993, 'learning_rate/qwen_vl_interface': 8.66639389970748e-06, 'learning_rate/action_model': 8.666393899707479e-05, 'epoch': 1.29}) +[2026-06-19 12:39:37,812][starVLA.training.train_starvla][INFO] - Step 1299, Loss: {'train/loss': 0.09976699948310852, 'train/grad_norm_pre_clip': 5.07865571975708, 'timing/data': 0.0130278350552544, 'timing/model': 1.3078679349273443, 'learning_rate/qwen_vl_interface': 8.66427735807607e-06, 'learning_rate/action_model': 8.664277358076069e-05, 'epoch': 1.3}) +[2026-06-19 12:39:42,242][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1300 +[2026-06-19 12:42:32,942][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1300 +[2026-06-19 12:42:32,942][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 1300 +[2026-06-19 12:45:37,951][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 1300 +[2026-06-19 12:45:37,952][starVLA.training.train_starvla][INFO] - Step 1300, Loss: {'train/loss': 0.33239734172821045, 'train/grad_norm_pre_clip': 10.2072114944458, 'eval/loss': 0.43556690216064453, 'eval/latency_0/loss': 0.10918071866035461, 'eval/latency_1/loss': 0.16326525807380676, 'eval/latency_2/loss': 0.6156749129295349, 'eval/latency_3/loss': 0.5076268911361694, 'eval/latency_4/loss': 0.7594780325889587, 'eval/flappy/loss': 0.43556690216064453, 'eval/flappy/latency_0/loss': 0.10918071866035461, 'eval/flappy/latency_1/loss': 0.16326525807380676, 'eval/flappy/latency_2/loss': 0.6156749129295349, 'eval/flappy/latency_3/loss': 0.5076268911361694, 'eval/flappy/latency_4/loss': 0.7594780325889587, 'eval/flap_precision': 0.6612903225806451, 'eval/flap_recall': 0.7454545454545455, 'eval/flap_f1': 0.7008547008547009, 'eval/flap_rate_model': 0.08611111111111111, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.7008547008547009, 'eval/latency_0/flap_precision': 0.6612903225806451, 'eval/latency_0/flap_recall': 0.7454545454545455, 'eval/latency_0/flap_f1': 0.7008547008547009, 'eval/latency_0/flap_rate_model': 0.08611111111111111, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.7008547008547009, 'timing/data': 0.01370151003357023, 'timing/model': 1.1520237780641764, 'learning_rate/qwen_vl_interface': 8.662159412964208e-06, 'learning_rate/action_model': 8.662159412964207e-05, 'epoch': 1.3}) +[2026-06-19 12:45:37,952][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_1300_state +[2026-06-19 12:45:37,955][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-19 12:45:41,337][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_1300_state/model.safetensors +[2026-06-19 12:45:48,530][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_1300_state/optimizer.bin +[2026-06-19 12:45:48,530][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_1300_state/sampler.bin +[2026-06-19 12:45:48,530][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_1300_state/sampler_1.bin +[2026-06-19 12:45:48,531][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_1300_state/random_states_0.pkl +[2026-06-19 12:45:48,552][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1300 +[2026-06-19 12:45:48,552][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-19 12:45:48,562][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-19 12:45:53,057][starVLA.training.train_starvla][INFO] - Step 1301, Loss: {'train/loss': 0.15125368535518646, 'train/grad_norm_pre_clip': 15.281758308410645, 'timing/data': 0.016517868032678962, 'timing/model': 1.123980029951781, 'learning_rate/qwen_vl_interface': 8.660040065242499e-06, 'learning_rate/action_model': 8.660040065242498e-05, 'epoch': 1.3}) +[2026-06-19 12:45:57,581][starVLA.training.train_starvla][INFO] - Step 1302, Loss: {'train/loss': 0.11375885456800461, 'train/grad_norm_pre_clip': 4.404867172241211, 'timing/data': 0.014001928037032485, 'timing/model': 1.3064253960037604, 'learning_rate/qwen_vl_interface': 8.657919315782127e-06, 'learning_rate/action_model': 8.657919315782127e-05, 'epoch': 1.3}) +[2026-06-19 12:46:02,080][starVLA.training.train_starvla][INFO] - Step 1303, Loss: {'train/loss': 0.24471989274024963, 'train/grad_norm_pre_clip': 5.146550178527832, 'timing/data': 0.013684535049833357, 'timing/model': 1.262165528954938, 'learning_rate/qwen_vl_interface': 8.655797165454853e-06, 'learning_rate/action_model': 8.655797165454853e-05, 'epoch': 1.3}) +[2026-06-19 12:46:06,460][starVLA.training.train_starvla][INFO] - Step 1304, Loss: {'train/loss': 0.2353592813014984, 'train/grad_norm_pre_clip': 5.462836742401123, 'timing/data': 0.014366245013661683, 'timing/model': 1.1373709149193019, 'learning_rate/qwen_vl_interface': 8.653673615133013e-06, 'learning_rate/action_model': 8.653673615133013e-05, 'epoch': 1.3}) +[2026-06-19 12:46:11,012][starVLA.training.train_starvla][INFO] - Step 1305, Loss: {'train/loss': 0.2640721797943115, 'train/grad_norm_pre_clip': 6.223296165466309, 'timing/data': 0.01997766294516623, 'timing/model': 1.13087739108596, 'learning_rate/qwen_vl_interface': 8.651548665689519e-06, 'learning_rate/action_model': 8.651548665689518e-05, 'epoch': 1.3}) +[2026-06-19 12:46:15,568][starVLA.training.train_starvla][INFO] - Step 1306, Loss: {'train/loss': 0.3015223741531372, 'train/grad_norm_pre_clip': 5.129154205322266, 'timing/data': 0.014094321057200432, 'timing/model': 1.1321720549603924, 'learning_rate/qwen_vl_interface': 8.649422317997855e-06, 'learning_rate/action_model': 8.649422317997856e-05, 'epoch': 1.3}) +[2026-06-19 12:46:20,034][starVLA.training.train_starvla][INFO] - Step 1307, Loss: {'train/loss': 0.373593270778656, 'train/grad_norm_pre_clip': 8.780818939208984, 'timing/data': 0.013057132018730044, 'timing/model': 1.1391166979447007, 'learning_rate/qwen_vl_interface': 8.647294572932082e-06, 'learning_rate/action_model': 8.647294572932082e-05, 'epoch': 1.3}) +[2026-06-19 12:46:24,590][starVLA.training.train_starvla][INFO] - Step 1308, Loss: {'train/loss': 0.14586997032165527, 'train/grad_norm_pre_clip': 4.157697677612305, 'timing/data': 0.01410717994440347, 'timing/model': 1.131037960993126, 'learning_rate/qwen_vl_interface': 8.64516543136684e-06, 'learning_rate/action_model': 8.64516543136684e-05, 'epoch': 1.3}) +[2026-06-19 12:46:29,147][starVLA.training.train_starvla][INFO] - Step 1309, Loss: {'train/loss': 0.2202489823102951, 'train/grad_norm_pre_clip': 5.643213272094727, 'timing/data': 0.014164909953251481, 'timing/model': 1.2970335840946063, 'learning_rate/qwen_vl_interface': 8.643034894177335e-06, 'learning_rate/action_model': 8.643034894177335e-05, 'epoch': 1.31}) +[2026-06-19 12:46:33,611][starVLA.training.train_starvla][INFO] - Step 1310, Loss: {'train/loss': 0.32632356882095337, 'train/grad_norm_pre_clip': 8.893043518066406, 'timing/data': 0.013712074025534093, 'timing/model': 1.2079915989888832, 'learning_rate/qwen_vl_interface': 8.64090296223935e-06, 'learning_rate/action_model': 8.64090296223935e-05, 'epoch': 1.31}) +[2026-06-19 12:46:38,008][starVLA.training.train_starvla][INFO] - Step 1311, Loss: {'train/loss': 0.18020929396152496, 'train/grad_norm_pre_clip': 3.906831741333008, 'timing/data': 0.013100861920975149, 'timing/model': 1.1421465140301734, 'learning_rate/qwen_vl_interface': 8.638769636429245e-06, 'learning_rate/action_model': 8.638769636429245e-05, 'epoch': 1.31}) +[2026-06-19 12:46:42,533][starVLA.training.train_starvla][INFO] - Step 1312, Loss: {'train/loss': 0.17066001892089844, 'train/grad_norm_pre_clip': 5.026057243347168, 'timing/data': 0.014664902002550662, 'timing/model': 1.138152556028217, 'learning_rate/qwen_vl_interface': 8.636634917623946e-06, 'learning_rate/action_model': 8.636634917623944e-05, 'epoch': 1.31}) +[2026-06-19 12:46:47,099][starVLA.training.train_starvla][INFO] - Step 1313, Loss: {'train/loss': 0.2307066172361374, 'train/grad_norm_pre_clip': 8.344400405883789, 'timing/data': 0.013486062991432846, 'timing/model': 1.1503957440145314, 'learning_rate/qwen_vl_interface': 8.634498806700955e-06, 'learning_rate/action_model': 8.634498806700954e-05, 'epoch': 1.31}) +[2026-06-19 12:46:51,659][starVLA.training.train_starvla][INFO] - Step 1314, Loss: {'train/loss': 0.1905478835105896, 'train/grad_norm_pre_clip': 5.505637168884277, 'timing/data': 0.013689627987332642, 'timing/model': 1.137094366014935, 'learning_rate/qwen_vl_interface': 8.63236130453835e-06, 'learning_rate/action_model': 8.63236130453835e-05, 'epoch': 1.31}) +[2026-06-19 12:46:56,180][starVLA.training.train_starvla][INFO] - Step 1315, Loss: {'train/loss': 0.4473068118095398, 'train/grad_norm_pre_clip': 10.208706855773926, 'timing/data': 0.013390284031629562, 'timing/model': 1.1398387659573928, 'learning_rate/qwen_vl_interface': 8.630222412014777e-06, 'learning_rate/action_model': 8.630222412014776e-05, 'epoch': 1.31}) +[2026-06-19 12:47:00,741][starVLA.training.train_starvla][INFO] - Step 1316, Loss: {'train/loss': 0.20985783636569977, 'train/grad_norm_pre_clip': 6.145927429199219, 'timing/data': 0.013556619989685714, 'timing/model': 1.3035584660246968, 'learning_rate/qwen_vl_interface': 8.628082130009451e-06, 'learning_rate/action_model': 8.628082130009451e-05, 'epoch': 1.31}) +[2026-06-19 12:47:05,169][starVLA.training.train_starvla][INFO] - Step 1317, Loss: {'train/loss': 0.5878923535346985, 'train/grad_norm_pre_clip': 8.235706329345703, 'timing/data': 0.013535405974835157, 'timing/model': 1.1474294069921598, 'learning_rate/qwen_vl_interface': 8.62594045940216e-06, 'learning_rate/action_model': 8.62594045940216e-05, 'epoch': 1.31}) +[2026-06-19 12:47:09,720][starVLA.training.train_starvla][INFO] - Step 1318, Loss: {'train/loss': 0.329030305147171, 'train/grad_norm_pre_clip': 13.15878963470459, 'timing/data': 0.014722838066518307, 'timing/model': 1.139237702009268, 'learning_rate/qwen_vl_interface': 8.623797401073272e-06, 'learning_rate/action_model': 8.623797401073271e-05, 'epoch': 1.31}) +[2026-06-19 12:47:14,304][starVLA.training.train_starvla][INFO] - Step 1319, Loss: {'train/loss': 0.3015509843826294, 'train/grad_norm_pre_clip': 12.167851448059082, 'timing/data': 0.013849062961526215, 'timing/model': 1.1472350660478696, 'learning_rate/qwen_vl_interface': 8.62165295590371e-06, 'learning_rate/action_model': 8.621652955903709e-05, 'epoch': 1.32}) +[2026-06-19 12:47:18,871][starVLA.training.train_starvla][INFO] - Step 1320, Loss: {'train/loss': 0.4023624062538147, 'train/grad_norm_pre_clip': 9.224732398986816, 'timing/data': 0.013075234950520098, 'timing/model': 1.1421142159961164, 'learning_rate/qwen_vl_interface': 8.619507124774975e-06, 'learning_rate/action_model': 8.619507124774975e-05, 'epoch': 1.32}) +[2026-06-19 12:47:23,456][starVLA.training.train_starvla][INFO] - Step 1321, Loss: {'train/loss': 0.18374143540859222, 'train/grad_norm_pre_clip': 5.531968116760254, 'timing/data': 0.013121748925186694, 'timing/model': 1.3076146959792823, 'learning_rate/qwen_vl_interface': 8.61735990856914e-06, 'learning_rate/action_model': 8.617359908569139e-05, 'epoch': 1.32}) +[2026-06-19 12:47:28,044][starVLA.training.train_starvla][INFO] - Step 1322, Loss: {'train/loss': 0.13740117847919464, 'train/grad_norm_pre_clip': 2.749497175216675, 'timing/data': 0.013516796985641122, 'timing/model': 1.3132645349251106, 'learning_rate/qwen_vl_interface': 8.615211308168842e-06, 'learning_rate/action_model': 8.61521130816884e-05, 'epoch': 1.32}) +[2026-06-19 12:47:32,487][starVLA.training.train_starvla][INFO] - Step 1323, Loss: {'train/loss': 0.2057049423456192, 'train/grad_norm_pre_clip': 12.725160598754883, 'timing/data': 0.014302607043646276, 'timing/model': 1.1455357919912785, 'learning_rate/qwen_vl_interface': 8.613061324457289e-06, 'learning_rate/action_model': 8.613061324457288e-05, 'epoch': 1.32}) +[2026-06-19 12:47:37,049][starVLA.training.train_starvla][INFO] - Step 1324, Loss: {'train/loss': 0.24070316553115845, 'train/grad_norm_pre_clip': 11.335810661315918, 'timing/data': 0.014072760008275509, 'timing/model': 1.1427557419519871, 'learning_rate/qwen_vl_interface': 8.610909958318262e-06, 'learning_rate/action_model': 8.610909958318261e-05, 'epoch': 1.32}) +[2026-06-19 12:47:41,642][starVLA.training.train_starvla][INFO] - Step 1325, Loss: {'train/loss': 0.22256352007389069, 'train/grad_norm_pre_clip': 9.149456977844238, 'timing/data': 0.013601985992863774, 'timing/model': 1.1454683869378641, 'learning_rate/qwen_vl_interface': 8.608757210636101e-06, 'learning_rate/action_model': 8.608757210636101e-05, 'epoch': 1.32}) +[2026-06-19 12:47:46,226][starVLA.training.train_starvla][INFO] - Step 1326, Loss: {'train/loss': 0.7394921183586121, 'train/grad_norm_pre_clip': 17.28713035583496, 'timing/data': 0.01427438494283706, 'timing/model': 1.1422253930941224, 'learning_rate/qwen_vl_interface': 8.606603082295724e-06, 'learning_rate/action_model': 8.606603082295724e-05, 'epoch': 1.32}) +[2026-06-19 12:47:50,827][starVLA.training.train_starvla][INFO] - Step 1327, Loss: {'train/loss': 0.15881699323654175, 'train/grad_norm_pre_clip': 3.869567394256592, 'timing/data': 0.013714171014726162, 'timing/model': 1.3231302800122648, 'learning_rate/qwen_vl_interface': 8.60444757418261e-06, 'learning_rate/action_model': 8.604447574182609e-05, 'epoch': 1.32}) +[2026-06-19 12:47:55,239][starVLA.training.train_starvla][INFO] - Step 1328, Loss: {'train/loss': 0.2662360966205597, 'train/grad_norm_pre_clip': 9.672779083251953, 'timing/data': 0.014665797003544867, 'timing/model': 1.146252021077089, 'learning_rate/qwen_vl_interface': 8.602290687182805e-06, 'learning_rate/action_model': 8.602290687182806e-05, 'epoch': 1.32}) +[2026-06-19 12:47:59,802][starVLA.training.train_starvla][INFO] - Step 1329, Loss: {'train/loss': 0.13752540946006775, 'train/grad_norm_pre_clip': 4.149923324584961, 'timing/data': 0.013816759921610355, 'timing/model': 1.140476203057915, 'learning_rate/qwen_vl_interface': 8.600132422182929e-06, 'learning_rate/action_model': 8.600132422182928e-05, 'epoch': 1.33}) +[2026-06-19 12:48:04,376][starVLA.training.train_starvla][INFO] - Step 1330, Loss: {'train/loss': 0.2574869692325592, 'train/grad_norm_pre_clip': 6.35297155380249, 'timing/data': 0.013212052057497203, 'timing/model': 1.1437658050563186, 'learning_rate/qwen_vl_interface': 8.597972780070156e-06, 'learning_rate/action_model': 8.597972780070155e-05, 'epoch': 1.33}) +[2026-06-19 12:48:08,946][starVLA.training.train_starvla][INFO] - Step 1331, Loss: {'train/loss': 0.09410464763641357, 'train/grad_norm_pre_clip': 2.024538993835449, 'timing/data': 0.013104781974107027, 'timing/model': 1.1446806060848758, 'learning_rate/qwen_vl_interface': 8.595811761732239e-06, 'learning_rate/action_model': 8.595811761732238e-05, 'epoch': 1.33}) +[2026-06-19 12:48:13,543][starVLA.training.train_starvla][INFO] - Step 1332, Loss: {'train/loss': 0.3147903382778168, 'train/grad_norm_pre_clip': 5.756958484649658, 'timing/data': 0.013348640990443528, 'timing/model': 1.1539664040319622, 'learning_rate/qwen_vl_interface': 8.593649368057489e-06, 'learning_rate/action_model': 8.59364936805749e-05, 'epoch': 1.33}) +[2026-06-19 12:48:18,137][starVLA.training.train_starvla][INFO] - Step 1333, Loss: {'train/loss': 0.19353275001049042, 'train/grad_norm_pre_clip': 10.768332481384277, 'timing/data': 0.01317834691144526, 'timing/model': 1.3224622959969565, 'learning_rate/qwen_vl_interface': 8.591485599934786e-06, 'learning_rate/action_model': 8.591485599934785e-05, 'epoch': 1.33}) +[2026-06-19 12:48:22,556][starVLA.training.train_starvla][INFO] - Step 1334, Loss: {'train/loss': 0.19674380123615265, 'train/grad_norm_pre_clip': 4.172601699829102, 'timing/data': 0.013193147024139762, 'timing/model': 1.14643377403263, 'learning_rate/qwen_vl_interface': 8.58932045825357e-06, 'learning_rate/action_model': 8.589320458253569e-05, 'epoch': 1.33}) +[2026-06-19 12:48:27,142][starVLA.training.train_starvla][INFO] - Step 1335, Loss: {'train/loss': 0.26365602016448975, 'train/grad_norm_pre_clip': 5.26118278503418, 'timing/data': 0.01316913403570652, 'timing/model': 1.1436185729689896, 'learning_rate/qwen_vl_interface': 8.58715394390385e-06, 'learning_rate/action_model': 8.58715394390385e-05, 'epoch': 1.33}) +[2026-06-19 12:48:31,642][starVLA.training.train_starvla][INFO] - Step 1336, Loss: {'train/loss': 0.26431936025619507, 'train/grad_norm_pre_clip': 5.391676902770996, 'timing/data': 0.013960466952994466, 'timing/model': 1.1514764360617846, 'learning_rate/qwen_vl_interface': 8.5849860577762e-06, 'learning_rate/action_model': 8.5849860577762e-05, 'epoch': 1.33}) +[2026-06-19 12:48:36,167][starVLA.training.train_starvla][INFO] - Step 1337, Loss: {'train/loss': 0.24919432401657104, 'train/grad_norm_pre_clip': 4.993080139160156, 'timing/data': 0.013571983901783824, 'timing/model': 1.1411772649735212, 'learning_rate/qwen_vl_interface': 8.582816800761756e-06, 'learning_rate/action_model': 8.582816800761755e-05, 'epoch': 1.33}) +[2026-06-19 12:48:40,733][starVLA.training.train_starvla][INFO] - Step 1338, Loss: {'train/loss': 0.1839803159236908, 'train/grad_norm_pre_clip': 5.8752760887146, 'timing/data': 0.013687678030692041, 'timing/model': 1.1415994139388204, 'learning_rate/qwen_vl_interface': 8.580646173752214e-06, 'learning_rate/action_model': 8.580646173752213e-05, 'epoch': 1.33}) +[2026-06-19 12:48:45,366][starVLA.training.train_starvla][INFO] - Step 1339, Loss: {'train/loss': 0.17913511395454407, 'train/grad_norm_pre_clip': 3.3376808166503906, 'timing/data': 0.013527747010812163, 'timing/model': 1.3200650779763237, 'learning_rate/qwen_vl_interface': 8.578474177639841e-06, 'learning_rate/action_model': 8.578474177639842e-05, 'epoch': 1.34}) +[2026-06-19 12:48:49,812][starVLA.training.train_starvla][INFO] - Step 1340, Loss: {'train/loss': 0.19615201652050018, 'train/grad_norm_pre_clip': 3.3120810985565186, 'timing/data': 0.013391292071901262, 'timing/model': 1.1618997160112485, 'learning_rate/qwen_vl_interface': 8.576300813317463e-06, 'learning_rate/action_model': 8.576300813317462e-05, 'epoch': 1.34}) +[2026-06-19 12:48:54,423][starVLA.training.train_starvla][INFO] - Step 1341, Loss: {'train/loss': 0.16509269177913666, 'train/grad_norm_pre_clip': 3.6359851360321045, 'timing/data': 0.014446017914451659, 'timing/model': 1.1512903269613162, 'learning_rate/qwen_vl_interface': 8.574126081678463e-06, 'learning_rate/action_model': 8.574126081678462e-05, 'epoch': 1.34}) +[2026-06-19 12:48:58,922][starVLA.training.train_starvla][INFO] - Step 1342, Loss: {'train/loss': 0.3375719487667084, 'train/grad_norm_pre_clip': 12.243770599365234, 'timing/data': 0.013107386999763548, 'timing/model': 1.149009712971747, 'learning_rate/qwen_vl_interface': 8.571949983616796e-06, 'learning_rate/action_model': 8.571949983616795e-05, 'epoch': 1.34}) +[2026-06-19 12:49:03,474][starVLA.training.train_starvla][INFO] - Step 1343, Loss: {'train/loss': 0.2344377636909485, 'train/grad_norm_pre_clip': 4.34116792678833, 'timing/data': 0.014192200032994151, 'timing/model': 1.142849450930953, 'learning_rate/qwen_vl_interface': 8.569772520026971e-06, 'learning_rate/action_model': 8.569772520026971e-05, 'epoch': 1.34}) +[2026-06-19 12:49:08,098][starVLA.training.train_starvla][INFO] - Step 1344, Loss: {'train/loss': 0.22685712575912476, 'train/grad_norm_pre_clip': 5.188516139984131, 'timing/data': 0.014109416981227696, 'timing/model': 1.1527599080000073, 'learning_rate/qwen_vl_interface': 8.567593691804061e-06, 'learning_rate/action_model': 8.567593691804061e-05, 'epoch': 1.34}) +[2026-06-19 12:49:12,683][starVLA.training.train_starvla][INFO] - Step 1345, Loss: {'train/loss': 0.7363787889480591, 'train/grad_norm_pre_clip': 18.17380714416504, 'timing/data': 0.014306719996966422, 'timing/model': 1.306157677900046, 'learning_rate/qwen_vl_interface': 8.565413499843703e-06, 'learning_rate/action_model': 8.565413499843703e-05, 'epoch': 1.34}) +[2026-06-19 12:49:17,230][starVLA.training.train_starvla][INFO] - Step 1346, Loss: {'train/loss': 0.49896228313446045, 'train/grad_norm_pre_clip': 13.142574310302734, 'timing/data': 0.013600123929791152, 'timing/model': 1.265029788017273, 'learning_rate/qwen_vl_interface': 8.563231945042087e-06, 'learning_rate/action_model': 8.563231945042088e-05, 'epoch': 1.34}) +[2026-06-19 12:49:21,657][starVLA.training.train_starvla][INFO] - Step 1347, Loss: {'train/loss': 0.13739421963691711, 'train/grad_norm_pre_clip': 3.561974048614502, 'timing/data': 0.0141426200279966, 'timing/model': 1.143606228986755, 'learning_rate/qwen_vl_interface': 8.56104902829597e-06, 'learning_rate/action_model': 8.56104902829597e-05, 'epoch': 1.34}) +[2026-06-19 12:49:26,247][starVLA.training.train_starvla][INFO] - Step 1348, Loss: {'train/loss': 0.15993928909301758, 'train/grad_norm_pre_clip': 4.546494007110596, 'timing/data': 0.013580936938524246, 'timing/model': 1.14630634104833, 'learning_rate/qwen_vl_interface': 8.558864750502669e-06, 'learning_rate/action_model': 8.558864750502668e-05, 'epoch': 1.34}) +[2026-06-19 12:49:30,851][starVLA.training.train_starvla][INFO] - Step 1349, Loss: {'train/loss': 0.18891097605228424, 'train/grad_norm_pre_clip': 5.853126525878906, 'timing/data': 0.014565238961949944, 'timing/model': 1.1490510429721326, 'learning_rate/qwen_vl_interface': 8.556679112560055e-06, 'learning_rate/action_model': 8.556679112560054e-05, 'epoch': 1.35}) +[2026-06-19 12:49:35,430][starVLA.training.train_starvla][INFO] - Step 1350, Loss: {'train/loss': 0.09012347459793091, 'train/grad_norm_pre_clip': 2.9843084812164307, 'timing/data': 0.013897270080633461, 'timing/model': 1.1411593559896573, 'learning_rate/qwen_vl_interface': 8.554492115366562e-06, 'learning_rate/action_model': 8.554492115366561e-05, 'epoch': 1.35}) +[2026-06-19 12:49:39,930][starVLA.training.train_starvla][INFO] - Step 1351, Loss: {'train/loss': 0.22858478128910065, 'train/grad_norm_pre_clip': 5.401137351989746, 'timing/data': 0.013643676065839827, 'timing/model': 1.1494413228938356, 'learning_rate/qwen_vl_interface': 8.55230375982118e-06, 'learning_rate/action_model': 8.55230375982118e-05, 'epoch': 1.35}) +[2026-06-19 12:49:44,466][starVLA.training.train_starvla][INFO] - Step 1352, Loss: {'train/loss': 0.27827000617980957, 'train/grad_norm_pre_clip': 5.941426753997803, 'timing/data': 0.013293689931742847, 'timing/model': 1.2474467650754377, 'learning_rate/qwen_vl_interface': 8.550114046823463e-06, 'learning_rate/action_model': 8.550114046823463e-05, 'epoch': 1.35}) +[2026-06-19 12:49:48,893][starVLA.training.train_starvla][INFO] - Step 1353, Loss: {'train/loss': 0.18012112379074097, 'train/grad_norm_pre_clip': 4.259340286254883, 'timing/data': 0.013676622998900712, 'timing/model': 1.1432611659402028, 'learning_rate/qwen_vl_interface': 8.547922977273518e-06, 'learning_rate/action_model': 8.547922977273518e-05, 'epoch': 1.35}) +[2026-06-19 12:49:53,476][starVLA.training.train_starvla][INFO] - Step 1354, Loss: {'train/loss': 0.26808080077171326, 'train/grad_norm_pre_clip': 8.937183380126953, 'timing/data': 0.013295153039507568, 'timing/model': 1.1438146819127724, 'learning_rate/qwen_vl_interface': 8.54573055207201e-06, 'learning_rate/action_model': 8.545730552072008e-05, 'epoch': 1.35}) +[2026-06-19 12:49:58,065][starVLA.training.train_starvla][INFO] - Step 1355, Loss: {'train/loss': 0.3118745982646942, 'train/grad_norm_pre_clip': 6.427074909210205, 'timing/data': 0.013287393958307803, 'timing/model': 1.1470233929576352, 'learning_rate/qwen_vl_interface': 8.543536772120161e-06, 'learning_rate/action_model': 8.54353677212016e-05, 'epoch': 1.35}) +[2026-06-19 12:50:02,653][starVLA.training.train_starvla][INFO] - Step 1356, Loss: {'train/loss': 0.32878032326698303, 'train/grad_norm_pre_clip': 4.907170295715332, 'timing/data': 0.013758487068116665, 'timing/model': 1.1457087979651988, 'learning_rate/qwen_vl_interface': 8.541341638319754e-06, 'learning_rate/action_model': 8.541341638319754e-05, 'epoch': 1.35}) +[2026-06-19 12:50:07,263][starVLA.training.train_starvla][INFO] - Step 1357, Loss: {'train/loss': 0.22502928972244263, 'train/grad_norm_pre_clip': 6.669670104980469, 'timing/data': 0.01441270299255848, 'timing/model': 1.3253613819833845, 'learning_rate/qwen_vl_interface': 8.539145151573126e-06, 'learning_rate/action_model': 8.539145151573125e-05, 'epoch': 1.35}) +[2026-06-19 12:50:11,786][starVLA.training.train_starvla][INFO] - Step 1358, Loss: {'train/loss': 0.3131643235683441, 'train/grad_norm_pre_clip': 5.3281636238098145, 'timing/data': 0.01379517000168562, 'timing/model': 1.2423573120031506, 'learning_rate/qwen_vl_interface': 8.536947312783167e-06, 'learning_rate/action_model': 8.536947312783166e-05, 'epoch': 1.35}) +[2026-06-19 12:50:16,225][starVLA.training.train_starvla][INFO] - Step 1359, Loss: {'train/loss': 0.30940839648246765, 'train/grad_norm_pre_clip': 9.248381614685059, 'timing/data': 0.015213908045552671, 'timing/model': 1.1448627040954307, 'learning_rate/qwen_vl_interface': 8.534748122853326e-06, 'learning_rate/action_model': 8.534748122853325e-05, 'epoch': 1.36}) +[2026-06-19 12:50:20,822][starVLA.training.train_starvla][INFO] - Step 1360, Loss: {'train/loss': 0.3849502503871918, 'train/grad_norm_pre_clip': 6.5000081062316895, 'timing/data': 0.012888301978819072, 'timing/model': 1.1437888379441574, 'learning_rate/qwen_vl_interface': 8.532547582687609e-06, 'learning_rate/action_model': 8.532547582687608e-05, 'epoch': 1.36}) +[2026-06-19 12:50:25,424][starVLA.training.train_starvla][INFO] - Step 1361, Loss: {'train/loss': 0.19061534106731415, 'train/grad_norm_pre_clip': 4.706027030944824, 'timing/data': 0.013470550067722797, 'timing/model': 1.1484123789705336, 'learning_rate/qwen_vl_interface': 8.530345693190573e-06, 'learning_rate/action_model': 8.530345693190572e-05, 'epoch': 1.36}) +[2026-06-19 12:50:30,034][starVLA.training.train_starvla][INFO] - Step 1362, Loss: {'train/loss': 0.21613819897174835, 'train/grad_norm_pre_clip': 4.57231330871582, 'timing/data': 0.013982419040985405, 'timing/model': 1.145856078946963, 'learning_rate/qwen_vl_interface': 8.528142455267328e-06, 'learning_rate/action_model': 8.528142455267329e-05, 'epoch': 1.36}) +[2026-06-19 12:50:34,625][starVLA.training.train_starvla][INFO] - Step 1363, Loss: {'train/loss': 0.15295645594596863, 'train/grad_norm_pre_clip': 3.819291830062866, 'timing/data': 0.013714335043914616, 'timing/model': 1.3084988410118967, 'learning_rate/qwen_vl_interface': 8.52593786982355e-06, 'learning_rate/action_model': 8.525937869823551e-05, 'epoch': 1.36}) +[2026-06-19 12:50:39,205][starVLA.training.train_starvla][INFO] - Step 1364, Loss: {'train/loss': 0.39990994334220886, 'train/grad_norm_pre_clip': 6.422186374664307, 'timing/data': 0.013269741088151932, 'timing/model': 1.3075029080500826, 'learning_rate/qwen_vl_interface': 8.523731937765456e-06, 'learning_rate/action_model': 8.523731937765456e-05, 'epoch': 1.36}) +[2026-06-19 12:50:43,636][starVLA.training.train_starvla][INFO] - Step 1365, Loss: {'train/loss': 0.37193453311920166, 'train/grad_norm_pre_clip': 10.558782577514648, 'timing/data': 0.012201260891743004, 'timing/model': 1.151526331086643, 'learning_rate/qwen_vl_interface': 8.521524659999823e-06, 'learning_rate/action_model': 8.521524659999822e-05, 'epoch': 1.36}) +[2026-06-19 12:50:48,238][starVLA.training.train_starvla][INFO] - Step 1366, Loss: {'train/loss': 0.19453759491443634, 'train/grad_norm_pre_clip': 4.350159645080566, 'timing/data': 0.015954573987983167, 'timing/model': 1.1568307600682601, 'learning_rate/qwen_vl_interface': 8.519316037433976e-06, 'learning_rate/action_model': 8.519316037433977e-05, 'epoch': 1.36}) +[2026-06-19 12:50:52,848][starVLA.training.train_starvla][INFO] - Step 1367, Loss: {'train/loss': 0.18009738624095917, 'train/grad_norm_pre_clip': 4.7009077072143555, 'timing/data': 0.01287647895514965, 'timing/model': 1.1465776149416342, 'learning_rate/qwen_vl_interface': 8.517106070975801e-06, 'learning_rate/action_model': 8.5171060709758e-05, 'epoch': 1.36}) +[2026-06-19 12:50:57,348][starVLA.training.train_starvla][INFO] - Step 1368, Loss: {'train/loss': 0.16168321669101715, 'train/grad_norm_pre_clip': 4.17072057723999, 'timing/data': 0.013762052054516971, 'timing/model': 1.1438142210245132, 'learning_rate/qwen_vl_interface': 8.514894761533729e-06, 'learning_rate/action_model': 8.514894761533728e-05, 'epoch': 1.36}) +[2026-06-19 12:51:01,918][starVLA.training.train_starvla][INFO] - Step 1369, Loss: {'train/loss': 0.38057613372802734, 'train/grad_norm_pre_clip': 9.376947402954102, 'timing/data': 0.013493934064172208, 'timing/model': 1.14920856396202, 'learning_rate/qwen_vl_interface': 8.512682110016747e-06, 'learning_rate/action_model': 8.512682110016747e-05, 'epoch': 1.37}) +[2026-06-19 12:51:06,516][starVLA.training.train_starvla][INFO] - Step 1370, Loss: {'train/loss': 0.2321079969406128, 'train/grad_norm_pre_clip': 6.294460773468018, 'timing/data': 0.01325687300413847, 'timing/model': 1.311773846973665, 'learning_rate/qwen_vl_interface': 8.51046811733439e-06, 'learning_rate/action_model': 8.51046811733439e-05, 'epoch': 1.37}) +[2026-06-19 12:51:10,938][starVLA.training.train_starvla][INFO] - Step 1371, Loss: {'train/loss': 0.23823799192905426, 'train/grad_norm_pre_clip': 6.680778980255127, 'timing/data': 0.013088563922792673, 'timing/model': 1.1460123759461567, 'learning_rate/qwen_vl_interface': 8.50825278439675e-06, 'learning_rate/action_model': 8.508252784396749e-05, 'epoch': 1.37}) +[2026-06-19 12:51:15,540][starVLA.training.train_starvla][INFO] - Step 1372, Loss: {'train/loss': 0.25911378860473633, 'train/grad_norm_pre_clip': 6.551965713500977, 'timing/data': 0.0133711330126971, 'timing/model': 1.150581699097529, 'learning_rate/qwen_vl_interface': 8.506036112114465e-06, 'learning_rate/action_model': 8.506036112114464e-05, 'epoch': 1.37}) +[2026-06-19 12:51:20,141][starVLA.training.train_starvla][INFO] - Step 1373, Loss: {'train/loss': 0.1941790133714676, 'train/grad_norm_pre_clip': 5.8419623374938965, 'timing/data': 0.013032266055233777, 'timing/model': 1.1446245480328798, 'learning_rate/qwen_vl_interface': 8.503818101398725e-06, 'learning_rate/action_model': 8.503818101398725e-05, 'epoch': 1.37}) +[2026-06-19 12:51:24,737][starVLA.training.train_starvla][INFO] - Step 1374, Loss: {'train/loss': 0.1644975244998932, 'train/grad_norm_pre_clip': 5.429660797119141, 'timing/data': 0.01340407901443541, 'timing/model': 1.1575590210268274, 'learning_rate/qwen_vl_interface': 8.501598753161273e-06, 'learning_rate/action_model': 8.501598753161272e-05, 'epoch': 1.37}) +[2026-06-19 12:51:29,223][starVLA.training.train_starvla][INFO] - Step 1375, Loss: {'train/loss': 0.1593981385231018, 'train/grad_norm_pre_clip': 5.171716690063477, 'timing/data': 0.014021708979271352, 'timing/model': 1.1504764349665493, 'learning_rate/qwen_vl_interface': 8.499378068314395e-06, 'learning_rate/action_model': 8.499378068314395e-05, 'epoch': 1.37}) +[2026-06-19 12:51:33,797][starVLA.training.train_starvla][INFO] - Step 1376, Loss: {'train/loss': 0.2663474977016449, 'train/grad_norm_pre_clip': 5.619354248046875, 'timing/data': 0.012692888965830207, 'timing/model': 1.2433902630582452, 'learning_rate/qwen_vl_interface': 8.497156047770932e-06, 'learning_rate/action_model': 8.497156047770933e-05, 'epoch': 1.37}) +[2026-06-19 12:51:38,253][starVLA.training.train_starvla][INFO] - Step 1377, Loss: {'train/loss': 0.23166918754577637, 'train/grad_norm_pre_clip': 7.322866439819336, 'timing/data': 0.014387863921001554, 'timing/model': 1.1501484449254349, 'learning_rate/qwen_vl_interface': 8.494932692444276e-06, 'learning_rate/action_model': 8.494932692444276e-05, 'epoch': 1.37}) +[2026-06-19 12:51:42,859][starVLA.training.train_starvla][INFO] - Step 1378, Loss: {'train/loss': 0.32417032122612, 'train/grad_norm_pre_clip': 5.315924167633057, 'timing/data': 0.014238086063414812, 'timing/model': 1.1503616760019213, 'learning_rate/qwen_vl_interface': 8.492708003248362e-06, 'learning_rate/action_model': 8.492708003248362e-05, 'epoch': 1.37}) +[2026-06-19 12:51:47,483][starVLA.training.train_starvla][INFO] - Step 1379, Loss: {'train/loss': 0.22661790251731873, 'train/grad_norm_pre_clip': 7.576927185058594, 'timing/data': 0.013604223029688, 'timing/model': 1.1522266760002822, 'learning_rate/qwen_vl_interface': 8.490481981097674e-06, 'learning_rate/action_model': 8.490481981097674e-05, 'epoch': 1.38}) +[2026-06-19 12:51:52,094][starVLA.training.train_starvla][INFO] - Step 1380, Loss: {'train/loss': 0.21642984449863434, 'train/grad_norm_pre_clip': 6.814959526062012, 'timing/data': 0.013000908074900508, 'timing/model': 1.1407105069374666, 'learning_rate/qwen_vl_interface': 8.488254626907249e-06, 'learning_rate/action_model': 8.488254626907249e-05, 'epoch': 1.38}) +[2026-06-19 12:51:56,694][starVLA.training.train_starvla][INFO] - Step 1381, Loss: {'train/loss': 0.3068298399448395, 'train/grad_norm_pre_clip': 7.470322132110596, 'timing/data': 0.013140516937710345, 'timing/model': 1.307010868913494, 'learning_rate/qwen_vl_interface': 8.486025941592664e-06, 'learning_rate/action_model': 8.486025941592664e-05, 'epoch': 1.38}) +[2026-06-19 12:52:01,194][starVLA.training.train_starvla][INFO] - Step 1382, Loss: {'train/loss': 0.310297429561615, 'train/grad_norm_pre_clip': 9.300389289855957, 'timing/data': 0.012740785023197532, 'timing/model': 1.2301981910131872, 'learning_rate/qwen_vl_interface': 8.483795926070054e-06, 'learning_rate/action_model': 8.483795926070052e-05, 'epoch': 1.38}) +[2026-06-19 12:52:05,638][starVLA.training.train_starvla][INFO] - Step 1383, Loss: {'train/loss': 0.28717276453971863, 'train/grad_norm_pre_clip': 10.346061706542969, 'timing/data': 0.012839284026995301, 'timing/model': 1.153171235928312, 'learning_rate/qwen_vl_interface': 8.481564581256087e-06, 'learning_rate/action_model': 8.481564581256087e-05, 'epoch': 1.38}) +[2026-06-19 12:52:10,259][starVLA.training.train_starvla][INFO] - Step 1384, Loss: {'train/loss': 0.27671322226524353, 'train/grad_norm_pre_clip': 5.023952007293701, 'timing/data': 0.013671142980456352, 'timing/model': 1.1493576880311593, 'learning_rate/qwen_vl_interface': 8.479331908067987e-06, 'learning_rate/action_model': 8.479331908067987e-05, 'epoch': 1.38}) +[2026-06-19 12:52:14,848][starVLA.training.train_starvla][INFO] - Step 1385, Loss: {'train/loss': 0.40713703632354736, 'train/grad_norm_pre_clip': 11.210174560546875, 'timing/data': 0.012929120915941894, 'timing/model': 1.1528040069388226, 'learning_rate/qwen_vl_interface': 8.477097907423524e-06, 'learning_rate/action_model': 8.477097907423524e-05, 'epoch': 1.38}) +[2026-06-19 12:52:19,436][starVLA.training.train_starvla][INFO] - Step 1386, Loss: {'train/loss': 0.2573619484901428, 'train/grad_norm_pre_clip': 5.660366535186768, 'timing/data': 0.013672178029082716, 'timing/model': 1.1451776779722422, 'learning_rate/qwen_vl_interface': 8.474862580241006e-06, 'learning_rate/action_model': 8.474862580241006e-05, 'epoch': 1.38}) +[2026-06-19 12:52:24,042][starVLA.training.train_starvla][INFO] - Step 1387, Loss: {'train/loss': 0.3367343842983246, 'train/grad_norm_pre_clip': 9.184751510620117, 'timing/data': 0.014178061974234879, 'timing/model': 1.3215632030041888, 'learning_rate/qwen_vl_interface': 8.472625927439299e-06, 'learning_rate/action_model': 8.472625927439299e-05, 'epoch': 1.38}) +[2026-06-19 12:52:28,557][starVLA.training.train_starvla][INFO] - Step 1388, Loss: {'train/loss': 0.19841919839382172, 'train/grad_norm_pre_clip': 4.0107221603393555, 'timing/data': 0.013686398975551128, 'timing/model': 1.230033725965768, 'learning_rate/qwen_vl_interface': 8.4703879499378e-06, 'learning_rate/action_model': 8.4703879499378e-05, 'epoch': 1.38}) +[2026-06-19 12:52:32,976][starVLA.training.train_starvla][INFO] - Step 1389, Loss: {'train/loss': 0.18853892385959625, 'train/grad_norm_pre_clip': 4.92302131652832, 'timing/data': 0.01357818499673158, 'timing/model': 1.1458738319342956, 'learning_rate/qwen_vl_interface': 8.46814864865646e-06, 'learning_rate/action_model': 8.46814864865646e-05, 'epoch': 1.39}) +[2026-06-19 12:52:37,592][starVLA.training.train_starvla][INFO] - Step 1390, Loss: {'train/loss': 0.4129352867603302, 'train/grad_norm_pre_clip': 10.191593170166016, 'timing/data': 0.014377403073012829, 'timing/model': 1.1559019229607657, 'learning_rate/qwen_vl_interface': 8.465908024515771e-06, 'learning_rate/action_model': 8.46590802451577e-05, 'epoch': 1.39}) +[2026-06-19 12:52:42,202][starVLA.training.train_starvla][INFO] - Step 1391, Loss: {'train/loss': 0.3335445821285248, 'train/grad_norm_pre_clip': 8.02540111541748, 'timing/data': 0.013880733982659876, 'timing/model': 1.1427799429511651, 'learning_rate/qwen_vl_interface': 8.463666078436769e-06, 'learning_rate/action_model': 8.463666078436769e-05, 'epoch': 1.39}) +[2026-06-19 12:52:46,799][starVLA.training.train_starvla][INFO] - Step 1392, Loss: {'train/loss': 0.3291127383708954, 'train/grad_norm_pre_clip': 11.406848907470703, 'timing/data': 0.012828950071707368, 'timing/model': 1.1515236779814586, 'learning_rate/qwen_vl_interface': 8.46142281134103e-06, 'learning_rate/action_model': 8.461422811341031e-05, 'epoch': 1.39}) +[2026-06-19 12:52:51,378][starVLA.training.train_starvla][INFO] - Step 1393, Loss: {'train/loss': 0.2655354142189026, 'train/grad_norm_pre_clip': 6.3150858879089355, 'timing/data': 0.013328152941539884, 'timing/model': 1.3024114980362356, 'learning_rate/qwen_vl_interface': 8.459178224150683e-06, 'learning_rate/action_model': 8.459178224150683e-05, 'epoch': 1.39}) +[2026-06-19 12:52:55,973][starVLA.training.train_starvla][INFO] - Step 1394, Loss: {'train/loss': 0.27607133984565735, 'train/grad_norm_pre_clip': 4.826198577880859, 'timing/data': 0.012918562046252191, 'timing/model': 1.3242822249885648, 'learning_rate/qwen_vl_interface': 8.456932317788389e-06, 'learning_rate/action_model': 8.456932317788389e-05, 'epoch': 1.39}) +[2026-06-19 12:53:00,400][starVLA.training.train_starvla][INFO] - Step 1395, Loss: {'train/loss': 0.2514398396015167, 'train/grad_norm_pre_clip': 7.8012166023254395, 'timing/data': 0.012795991031453013, 'timing/model': 1.1451543120201677, 'learning_rate/qwen_vl_interface': 8.454685093177356e-06, 'learning_rate/action_model': 8.454685093177354e-05, 'epoch': 1.39}) +[2026-06-19 12:53:05,006][starVLA.training.train_starvla][INFO] - Step 1396, Loss: {'train/loss': 0.2782725691795349, 'train/grad_norm_pre_clip': 5.504852771759033, 'timing/data': 0.01344669598620385, 'timing/model': 1.1459501290228218, 'learning_rate/qwen_vl_interface': 8.452436551241328e-06, 'learning_rate/action_model': 8.452436551241328e-05, 'epoch': 1.39}) +[2026-06-19 12:53:09,587][starVLA.training.train_starvla][INFO] - Step 1397, Loss: {'train/loss': 0.1427334100008011, 'train/grad_norm_pre_clip': 6.073573112487793, 'timing/data': 0.013803299982100725, 'timing/model': 1.1421439440455288, 'learning_rate/qwen_vl_interface': 8.450186692904605e-06, 'learning_rate/action_model': 8.450186692904606e-05, 'epoch': 1.39}) +[2026-06-19 12:53:14,111][starVLA.training.train_starvla][INFO] - Step 1398, Loss: {'train/loss': 0.3508082628250122, 'train/grad_norm_pre_clip': 8.828874588012695, 'timing/data': 0.013134015956893563, 'timing/model': 1.1580969369970262, 'learning_rate/qwen_vl_interface': 8.447935519092014e-06, 'learning_rate/action_model': 8.447935519092014e-05, 'epoch': 1.39}) +[2026-06-19 12:53:18,707][starVLA.training.train_starvla][INFO] - Step 1399, Loss: {'train/loss': 0.19912593066692352, 'train/grad_norm_pre_clip': 4.6129961013793945, 'timing/data': 0.01293503597844392, 'timing/model': 1.1538250219309703, 'learning_rate/qwen_vl_interface': 8.445683030728925e-06, 'learning_rate/action_model': 8.445683030728925e-05, 'epoch': 1.4}) +[2026-06-19 12:53:23,368][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1400 +[2026-06-19 12:56:15,867][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1400 +[2026-06-19 12:56:15,867][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 1400 +[2026-06-19 12:59:23,417][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 1400 +[2026-06-19 12:59:23,419][starVLA.training.train_starvla][INFO] - Step 1400, Loss: {'train/loss': 0.354812353849411, 'train/grad_norm_pre_clip': 10.56248664855957, 'eval/loss': 0.31170588731765747, 'eval/latency_0/loss': 0.10532517731189728, 'eval/latency_1/loss': 0.1017780527472496, 'eval/latency_2/loss': 0.36665189266204834, 'eval/latency_3/loss': 0.4133698642253876, 'eval/latency_4/loss': 0.5659019351005554, 'eval/flappy/loss': 0.31170588731765747, 'eval/flappy/latency_0/loss': 0.10532517731189728, 'eval/flappy/latency_1/loss': 0.1017780527472496, 'eval/flappy/latency_2/loss': 0.36665189266204834, 'eval/flappy/latency_3/loss': 0.4133698642253876, 'eval/flappy/latency_4/loss': 0.5659019351005554, 'eval/flap_precision': 0.764102564102564, 'eval/flap_recall': 0.5418181818181819, 'eval/flap_f1': 0.6340425531914893, 'eval/flap_rate_model': 0.05416666666666667, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.6340425531914893, 'eval/latency_0/flap_precision': 0.764102564102564, 'eval/latency_0/flap_recall': 0.5418181818181819, 'eval/latency_0/flap_f1': 0.6340425531914893, 'eval/latency_0/flap_rate_model': 0.05416666666666667, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.6340425531914893, 'timing/data': 0.01706476998515427, 'timing/model': 1.333522908971645, 'learning_rate/qwen_vl_interface': 8.443429228741256e-06, 'learning_rate/action_model': 8.443429228741255e-05, 'epoch': 1.4}) +[2026-06-19 12:59:23,419][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_1400_state +[2026-06-19 12:59:23,422][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-19 12:59:26,810][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_1400_state/model.safetensors +[2026-06-19 12:59:33,927][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_1400_state/optimizer.bin +[2026-06-19 12:59:33,927][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_1400_state/sampler.bin +[2026-06-19 12:59:33,928][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_1400_state/sampler_1.bin +[2026-06-19 12:59:33,928][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_1400_state/random_states_0.pkl +[2026-06-19 12:59:33,953][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1400 +[2026-06-19 12:59:33,953][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-19 12:59:33,962][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-19 12:59:38,506][starVLA.training.train_starvla][INFO] - Step 1401, Loss: {'train/loss': 0.15550337731838226, 'train/grad_norm_pre_clip': 3.229313850402832, 'timing/data': 0.013835464953444898, 'timing/model': 1.1321961230132729, 'learning_rate/qwen_vl_interface': 8.441174114055457e-06, 'learning_rate/action_model': 8.441174114055456e-05, 'epoch': 1.4}) +[2026-06-19 12:59:43,030][starVLA.training.train_starvla][INFO] - Step 1402, Loss: {'train/loss': 0.09547275304794312, 'train/grad_norm_pre_clip': 3.664308786392212, 'timing/data': 0.013093783985823393, 'timing/model': 1.3006235939683393, 'learning_rate/qwen_vl_interface': 8.43891768759852e-06, 'learning_rate/action_model': 8.43891768759852e-05, 'epoch': 1.4}) +[2026-06-19 12:59:47,451][starVLA.training.train_starvla][INFO] - Step 1403, Loss: {'train/loss': 0.13800467550754547, 'train/grad_norm_pre_clip': 3.035672187805176, 'timing/data': 0.012292527942918241, 'timing/model': 1.2007746170274913, 'learning_rate/qwen_vl_interface': 8.436659950297978e-06, 'learning_rate/action_model': 8.436659950297977e-05, 'epoch': 1.4}) +[2026-06-19 12:59:51,808][starVLA.training.train_starvla][INFO] - Step 1404, Loss: {'train/loss': 0.43553200364112854, 'train/grad_norm_pre_clip': 18.825336456298828, 'timing/data': 0.013123979093506932, 'timing/model': 1.12702228105627, 'learning_rate/qwen_vl_interface': 8.4344009030819e-06, 'learning_rate/action_model': 8.4344009030819e-05, 'epoch': 1.4}) +[2026-06-19 12:59:56,316][starVLA.training.train_starvla][INFO] - Step 1405, Loss: {'train/loss': 0.2537800967693329, 'train/grad_norm_pre_clip': 12.859199523925781, 'timing/data': 0.013695493922568858, 'timing/model': 1.1335089760832489, 'learning_rate/qwen_vl_interface': 8.432140546878899e-06, 'learning_rate/action_model': 8.432140546878898e-05, 'epoch': 1.4}) +[2026-06-19 13:00:00,876][starVLA.training.train_starvla][INFO] - Step 1406, Loss: {'train/loss': 0.08303669840097427, 'train/grad_norm_pre_clip': 4.171656608581543, 'timing/data': 0.013269003015011549, 'timing/model': 1.1394705169368535, 'learning_rate/qwen_vl_interface': 8.429878882618116e-06, 'learning_rate/action_model': 8.429878882618116e-05, 'epoch': 1.4}) +[2026-06-19 13:00:05,412][starVLA.training.train_starvla][INFO] - Step 1407, Loss: {'train/loss': 0.25985080003738403, 'train/grad_norm_pre_clip': 8.353754043579102, 'timing/data': 0.013502240064553916, 'timing/model': 1.130273279035464, 'learning_rate/qwen_vl_interface': 8.42761591122924e-06, 'learning_rate/action_model': 8.427615911229241e-05, 'epoch': 1.4}) +[2026-06-19 13:00:09,948][starVLA.training.train_starvla][INFO] - Step 1408, Loss: {'train/loss': 0.12504130601882935, 'train/grad_norm_pre_clip': 5.06356954574585, 'timing/data': 0.012762338039465249, 'timing/model': 1.3000039050821215, 'learning_rate/qwen_vl_interface': 8.425351633642491e-06, 'learning_rate/action_model': 8.425351633642492e-05, 'epoch': 1.4}) +[2026-06-19 13:00:14,489][starVLA.training.train_starvla][INFO] - Step 1409, Loss: {'train/loss': 0.24881470203399658, 'train/grad_norm_pre_clip': 15.301371574401855, 'timing/data': 0.013069255976006389, 'timing/model': 1.2942366660572588, 'learning_rate/qwen_vl_interface': 8.423086050788632e-06, 'learning_rate/action_model': 8.423086050788631e-05, 'epoch': 1.4}) +[2026-06-19 13:00:18,870][starVLA.training.train_starvla][INFO] - Step 1410, Loss: {'train/loss': 0.13665428757667542, 'train/grad_norm_pre_clip': 4.719171047210693, 'timing/data': 0.012666203081607819, 'timing/model': 1.1368401470826939, 'learning_rate/qwen_vl_interface': 8.420819163598953e-06, 'learning_rate/action_model': 8.420819163598952e-05, 'epoch': 1.41}) +[2026-06-19 13:00:23,433][starVLA.training.train_starvla][INFO] - Step 1411, Loss: {'train/loss': 0.10729188472032547, 'train/grad_norm_pre_clip': 4.541718006134033, 'timing/data': 0.013202621950767934, 'timing/model': 1.137723686057143, 'learning_rate/qwen_vl_interface': 8.418550973005289e-06, 'learning_rate/action_model': 8.418550973005289e-05, 'epoch': 1.41}) +[2026-06-19 13:00:27,994][starVLA.training.train_starvla][INFO] - Step 1412, Loss: {'train/loss': 0.6063884496688843, 'train/grad_norm_pre_clip': 12.71124267578125, 'timing/data': 0.013370565953664482, 'timing/model': 1.1327241190010682, 'learning_rate/qwen_vl_interface': 8.416281479940005e-06, 'learning_rate/action_model': 8.416281479940005e-05, 'epoch': 1.41}) +[2026-06-19 13:00:32,472][starVLA.training.train_starvla][INFO] - Step 1413, Loss: {'train/loss': 0.25088778138160706, 'train/grad_norm_pre_clip': 9.506872177124023, 'timing/data': 0.013129222905263305, 'timing/model': 1.1506373080192134, 'learning_rate/qwen_vl_interface': 8.414010685336008e-06, 'learning_rate/action_model': 8.414010685336008e-05, 'epoch': 1.41}) +[2026-06-19 13:00:36,960][starVLA.training.train_starvla][INFO] - Step 1414, Loss: {'train/loss': 0.5484882593154907, 'train/grad_norm_pre_clip': 18.8614444732666, 'timing/data': 0.013063078979030252, 'timing/model': 1.1374582889257, 'learning_rate/qwen_vl_interface': 8.411738590126733e-06, 'learning_rate/action_model': 8.411738590126732e-05, 'epoch': 1.41}) +[2026-06-19 13:00:41,502][starVLA.training.train_starvla][INFO] - Step 1415, Loss: {'train/loss': 0.3677336275577545, 'train/grad_norm_pre_clip': 26.338153839111328, 'timing/data': 0.012912370963022113, 'timing/model': 1.2972229509614408, 'learning_rate/qwen_vl_interface': 8.409465195246152e-06, 'learning_rate/action_model': 8.409465195246152e-05, 'epoch': 1.41}) +[2026-06-19 13:00:45,908][starVLA.training.train_starvla][INFO] - Step 1416, Loss: {'train/loss': 0.23346176743507385, 'train/grad_norm_pre_clip': 8.410181999206543, 'timing/data': 0.012937974999658763, 'timing/model': 1.1370464189676568, 'learning_rate/qwen_vl_interface': 8.407190501628772e-06, 'learning_rate/action_model': 8.407190501628772e-05, 'epoch': 1.41}) +[2026-06-19 13:00:50,480][starVLA.training.train_starvla][INFO] - Step 1417, Loss: {'train/loss': 0.04698791727423668, 'train/grad_norm_pre_clip': 4.03508186340332, 'timing/data': 0.013154124026186764, 'timing/model': 1.1581776769598946, 'learning_rate/qwen_vl_interface': 8.404914510209635e-06, 'learning_rate/action_model': 8.404914510209635e-05, 'epoch': 1.41}) +[2026-06-19 13:00:55,041][starVLA.training.train_starvla][INFO] - Step 1418, Loss: {'train/loss': 0.5170143246650696, 'train/grad_norm_pre_clip': 11.973974227905273, 'timing/data': 0.01302607194520533, 'timing/model': 1.1375014069490135, 'learning_rate/qwen_vl_interface': 8.402637221924315e-06, 'learning_rate/action_model': 8.402637221924314e-05, 'epoch': 1.41}) +[2026-06-19 13:00:59,601][starVLA.training.train_starvla][INFO] - Step 1419, Loss: {'train/loss': 0.37808382511138916, 'train/grad_norm_pre_clip': 18.032630920410156, 'timing/data': 0.013595265918411314, 'timing/model': 1.1406330079771578, 'learning_rate/qwen_vl_interface': 8.400358637708919e-06, 'learning_rate/action_model': 8.400358637708918e-05, 'epoch': 1.41}) +[2026-06-19 13:01:04,199][starVLA.training.train_starvla][INFO] - Step 1420, Loss: {'train/loss': 0.34389597177505493, 'train/grad_norm_pre_clip': 12.72157096862793, 'timing/data': 0.012863431009463966, 'timing/model': 1.3218517949571833, 'learning_rate/qwen_vl_interface': 8.398078758500086e-06, 'learning_rate/action_model': 8.398078758500086e-05, 'epoch': 1.42}) +[2026-06-19 13:01:08,640][starVLA.training.train_starvla][INFO] - Step 1421, Loss: {'train/loss': 0.1845889836549759, 'train/grad_norm_pre_clip': 6.646985054016113, 'timing/data': 0.013354141032323241, 'timing/model': 1.1424456260865554, 'learning_rate/qwen_vl_interface': 8.395797585234987e-06, 'learning_rate/action_model': 8.395797585234988e-05, 'epoch': 1.42}) +[2026-06-19 13:01:13,227][starVLA.training.train_starvla][INFO] - Step 1422, Loss: {'train/loss': 0.27390965819358826, 'train/grad_norm_pre_clip': 6.060206890106201, 'timing/data': 0.013167673023417592, 'timing/model': 1.1409396020462736, 'learning_rate/qwen_vl_interface': 8.39351511885133e-06, 'learning_rate/action_model': 8.393515118851329e-05, 'epoch': 1.42}) +[2026-06-19 13:01:17,800][starVLA.training.train_starvla][INFO] - Step 1423, Loss: {'train/loss': 0.17997333407402039, 'train/grad_norm_pre_clip': 5.464931488037109, 'timing/data': 0.013145006028935313, 'timing/model': 1.1455801230622455, 'learning_rate/qwen_vl_interface': 8.391231360287348e-06, 'learning_rate/action_model': 8.391231360287347e-05, 'epoch': 1.42}) +[2026-06-19 13:01:22,341][starVLA.training.train_starvla][INFO] - Step 1424, Loss: {'train/loss': 0.24803802371025085, 'train/grad_norm_pre_clip': 7.444899559020996, 'timing/data': 0.01342795800883323, 'timing/model': 1.13866884005256, 'learning_rate/qwen_vl_interface': 8.388946310481807e-06, 'learning_rate/action_model': 8.388946310481807e-05, 'epoch': 1.42}) +[2026-06-19 13:01:26,921][starVLA.training.train_starvla][INFO] - Step 1425, Loss: {'train/loss': 0.5350378751754761, 'train/grad_norm_pre_clip': 10.415611267089844, 'timing/data': 0.013244579080492258, 'timing/model': 1.1425842460012063, 'learning_rate/qwen_vl_interface': 8.386659970374007e-06, 'learning_rate/action_model': 8.386659970374006e-05, 'epoch': 1.42}) +[2026-06-19 13:01:31,521][starVLA.training.train_starvla][INFO] - Step 1426, Loss: {'train/loss': 0.3763032555580139, 'train/grad_norm_pre_clip': 9.75097942352295, 'timing/data': 0.012723751016892493, 'timing/model': 1.316852145944722, 'learning_rate/qwen_vl_interface': 8.384372340903775e-06, 'learning_rate/action_model': 8.384372340903775e-05, 'epoch': 1.42}) +[2026-06-19 13:01:35,940][starVLA.training.train_starvla][INFO] - Step 1427, Loss: {'train/loss': 0.1604745239019394, 'train/grad_norm_pre_clip': 2.8518924713134766, 'timing/data': 0.013032515998929739, 'timing/model': 1.154089989955537, 'learning_rate/qwen_vl_interface': 8.382083423011472e-06, 'learning_rate/action_model': 8.382083423011471e-05, 'epoch': 1.42}) +[2026-06-19 13:01:40,545][starVLA.training.train_starvla][INFO] - Step 1428, Loss: {'train/loss': 0.26798126101493835, 'train/grad_norm_pre_clip': 6.601049423217773, 'timing/data': 0.013120273011736572, 'timing/model': 1.1537325599929318, 'learning_rate/qwen_vl_interface': 8.379793217637979e-06, 'learning_rate/action_model': 8.379793217637979e-05, 'epoch': 1.42}) +[2026-06-19 13:01:45,067][starVLA.training.train_starvla][INFO] - Step 1429, Loss: {'train/loss': 0.2810642421245575, 'train/grad_norm_pre_clip': 6.329930305480957, 'timing/data': 0.012706840061582625, 'timing/model': 1.1598709509707987, 'learning_rate/qwen_vl_interface': 8.377501725724719e-06, 'learning_rate/action_model': 8.377501725724719e-05, 'epoch': 1.42}) +[2026-06-19 13:01:49,624][starVLA.training.train_starvla][INFO] - Step 1430, Loss: {'train/loss': 0.09827104955911636, 'train/grad_norm_pre_clip': 2.706385374069214, 'timing/data': 0.012807431980036199, 'timing/model': 1.1422619690420106, 'learning_rate/qwen_vl_interface': 8.375208948213636e-06, 'learning_rate/action_model': 8.375208948213635e-05, 'epoch': 1.43}) +[2026-06-19 13:01:54,244][starVLA.training.train_starvla][INFO] - Step 1431, Loss: {'train/loss': 0.14218737185001373, 'train/grad_norm_pre_clip': 3.4523043632507324, 'timing/data': 0.013308439985848963, 'timing/model': 1.1496972939930856, 'learning_rate/qwen_vl_interface': 8.372914886047203e-06, 'learning_rate/action_model': 8.372914886047203e-05, 'epoch': 1.43}) +[2026-06-19 13:01:58,848][starVLA.training.train_starvla][INFO] - Step 1432, Loss: {'train/loss': 0.2703065574169159, 'train/grad_norm_pre_clip': 5.769105911254883, 'timing/data': 0.012980669969692826, 'timing/model': 1.319047242985107, 'learning_rate/qwen_vl_interface': 8.370619540168422e-06, 'learning_rate/action_model': 8.370619540168423e-05, 'epoch': 1.43}) +[2026-06-19 13:02:03,285][starVLA.training.train_starvla][INFO] - Step 1433, Loss: {'train/loss': 0.241963729262352, 'train/grad_norm_pre_clip': 4.9619059562683105, 'timing/data': 0.013423466007225215, 'timing/model': 1.147716247010976, 'learning_rate/qwen_vl_interface': 8.368322911520826e-06, 'learning_rate/action_model': 8.368322911520826e-05, 'epoch': 1.43}) +[2026-06-19 13:02:07,881][starVLA.training.train_starvla][INFO] - Step 1434, Loss: {'train/loss': 0.35046419501304626, 'train/grad_norm_pre_clip': 6.552948951721191, 'timing/data': 0.013224077061749995, 'timing/model': 1.1416319119744003, 'learning_rate/qwen_vl_interface': 8.36602500104847e-06, 'learning_rate/action_model': 8.36602500104847e-05, 'epoch': 1.43}) +[2026-06-19 13:02:12,386][starVLA.training.train_starvla][INFO] - Step 1435, Loss: {'train/loss': 0.12078488618135452, 'train/grad_norm_pre_clip': 3.479161500930786, 'timing/data': 0.01731533301062882, 'timing/model': 1.1489450010703877, 'learning_rate/qwen_vl_interface': 8.36372580969594e-06, 'learning_rate/action_model': 8.36372580969594e-05, 'epoch': 1.43}) +[2026-06-19 13:02:16,920][starVLA.training.train_starvla][INFO] - Step 1436, Loss: {'train/loss': 0.2451493889093399, 'train/grad_norm_pre_clip': 4.508773326873779, 'timing/data': 0.013301111059263349, 'timing/model': 1.1524903449462727, 'learning_rate/qwen_vl_interface': 8.361425338408344e-06, 'learning_rate/action_model': 8.361425338408343e-05, 'epoch': 1.43}) +[2026-06-19 13:02:21,512][starVLA.training.train_starvla][INFO] - Step 1437, Loss: {'train/loss': 0.08209013193845749, 'train/grad_norm_pre_clip': 3.0836942195892334, 'timing/data': 0.01360034302342683, 'timing/model': 1.1521452709566802, 'learning_rate/qwen_vl_interface': 8.359123588131319e-06, 'learning_rate/action_model': 8.359123588131319e-05, 'epoch': 1.43}) +[2026-06-19 13:02:26,094][starVLA.training.train_starvla][INFO] - Step 1438, Loss: {'train/loss': 0.3254670202732086, 'train/grad_norm_pre_clip': 8.721478462219238, 'timing/data': 0.013439033995382488, 'timing/model': 1.307351693045348, 'learning_rate/qwen_vl_interface': 8.35682055981103e-06, 'learning_rate/action_model': 8.35682055981103e-05, 'epoch': 1.43}) +[2026-06-19 13:02:30,504][starVLA.training.train_starvla][INFO] - Step 1439, Loss: {'train/loss': 0.2000749707221985, 'train/grad_norm_pre_clip': 5.155014514923096, 'timing/data': 0.013288351008668542, 'timing/model': 1.141968238982372, 'learning_rate/qwen_vl_interface': 8.354516254394163e-06, 'learning_rate/action_model': 8.354516254394162e-05, 'epoch': 1.43}) +[2026-06-19 13:02:35,084][starVLA.training.train_starvla][INFO] - Step 1440, Loss: {'train/loss': 0.23887330293655396, 'train/grad_norm_pre_clip': 6.667946815490723, 'timing/data': 0.013235362013801932, 'timing/model': 1.1445121829165146, 'learning_rate/qwen_vl_interface': 8.352210672827932e-06, 'learning_rate/action_model': 8.352210672827931e-05, 'epoch': 1.44}) +[2026-06-19 13:02:39,580][starVLA.training.train_starvla][INFO] - Step 1441, Loss: {'train/loss': 0.28373199701309204, 'train/grad_norm_pre_clip': 9.200969696044922, 'timing/data': 0.013040991965681314, 'timing/model': 1.1440955830039456, 'learning_rate/qwen_vl_interface': 8.349903816060074e-06, 'learning_rate/action_model': 8.349903816060075e-05, 'epoch': 1.44}) +[2026-06-19 13:02:44,163][starVLA.training.train_starvla][INFO] - Step 1442, Loss: {'train/loss': 0.24492070078849792, 'train/grad_norm_pre_clip': 6.224643230438232, 'timing/data': 0.013688586070202291, 'timing/model': 1.1569131860742345, 'learning_rate/qwen_vl_interface': 8.34759568503885e-06, 'learning_rate/action_model': 8.347595685038851e-05, 'epoch': 1.44}) +[2026-06-19 13:02:48,751][starVLA.training.train_starvla][INFO] - Step 1443, Loss: {'train/loss': 0.28648707270622253, 'train/grad_norm_pre_clip': 9.592947959899902, 'timing/data': 0.013494209037162364, 'timing/model': 1.141938098007813, 'learning_rate/qwen_vl_interface': 8.345286280713049e-06, 'learning_rate/action_model': 8.345286280713048e-05, 'epoch': 1.44}) +[2026-06-19 13:02:53,333][starVLA.training.train_starvla][INFO] - Step 1444, Loss: {'train/loss': 0.23089897632598877, 'train/grad_norm_pre_clip': 5.9909210205078125, 'timing/data': 0.01310121500864625, 'timing/model': 1.3026735240127891, 'learning_rate/qwen_vl_interface': 8.342975604031974e-06, 'learning_rate/action_model': 8.342975604031974e-05, 'epoch': 1.44}) +[2026-06-19 13:02:57,875][starVLA.training.train_starvla][INFO] - Step 1445, Loss: {'train/loss': 0.6144413948059082, 'train/grad_norm_pre_clip': 32.972389221191406, 'timing/data': 0.013792154961265624, 'timing/model': 1.2706200979882851, 'learning_rate/qwen_vl_interface': 8.340663655945462e-06, 'learning_rate/action_model': 8.340663655945462e-05, 'epoch': 1.44}) +[2026-06-19 13:03:02,290][starVLA.training.train_starvla][INFO] - Step 1446, Loss: {'train/loss': 0.10451344400644302, 'train/grad_norm_pre_clip': 10.052586555480957, 'timing/data': 0.01359774696175009, 'timing/model': 1.143887456972152, 'learning_rate/qwen_vl_interface': 8.338350437403866e-06, 'learning_rate/action_model': 8.338350437403865e-05, 'epoch': 1.44}) +[2026-06-19 13:03:06,872][starVLA.training.train_starvla][INFO] - Step 1447, Loss: {'train/loss': 0.25294581055641174, 'train/grad_norm_pre_clip': 7.732503414154053, 'timing/data': 0.013129985891282558, 'timing/model': 1.143693621037528, 'learning_rate/qwen_vl_interface': 8.336035949358063e-06, 'learning_rate/action_model': 8.336035949358062e-05, 'epoch': 1.44}) +[2026-06-19 13:03:11,448][starVLA.training.train_starvla][INFO] - Step 1448, Loss: {'train/loss': 0.36298269033432007, 'train/grad_norm_pre_clip': 8.617669105529785, 'timing/data': 0.01294154895003885, 'timing/model': 1.143541811965406, 'learning_rate/qwen_vl_interface': 8.333720192759448e-06, 'learning_rate/action_model': 8.333720192759448e-05, 'epoch': 1.44}) +[2026-06-19 13:03:16,044][starVLA.training.train_starvla][INFO] - Step 1449, Loss: {'train/loss': 0.12858200073242188, 'train/grad_norm_pre_clip': 4.353082656860352, 'timing/data': 0.01294733106624335, 'timing/model': 1.1500500719994307, 'learning_rate/qwen_vl_interface': 8.331403168559946e-06, 'learning_rate/action_model': 8.331403168559945e-05, 'epoch': 1.44}) +[2026-06-19 13:03:20,534][starVLA.training.train_starvla][INFO] - Step 1450, Loss: {'train/loss': 0.0657399594783783, 'train/grad_norm_pre_clip': 4.516257286071777, 'timing/data': 0.01343673700466752, 'timing/model': 1.1447555349441245, 'learning_rate/qwen_vl_interface': 8.329084877711995e-06, 'learning_rate/action_model': 8.329084877711996e-05, 'epoch': 1.45}) +[2026-06-19 13:03:25,058][starVLA.training.train_starvla][INFO] - Step 1451, Loss: {'train/loss': 0.2025023251771927, 'train/grad_norm_pre_clip': 5.910114288330078, 'timing/data': 0.012910421006381512, 'timing/model': 1.246073201065883, 'learning_rate/qwen_vl_interface': 8.32676532116856e-06, 'learning_rate/action_model': 8.326765321168559e-05, 'epoch': 1.45}) +[2026-06-19 13:03:29,480][starVLA.training.train_starvla][INFO] - Step 1452, Loss: {'train/loss': 0.21058955788612366, 'train/grad_norm_pre_clip': 13.515223503112793, 'timing/data': 0.012804110068827868, 'timing/model': 1.1474116139579564, 'learning_rate/qwen_vl_interface': 8.324444499883119e-06, 'learning_rate/action_model': 8.324444499883119e-05, 'epoch': 1.45}) +[2026-06-19 13:03:34,079][starVLA.training.train_starvla][INFO] - Step 1453, Loss: {'train/loss': 0.14866992831230164, 'train/grad_norm_pre_clip': 5.15669059753418, 'timing/data': 0.013511496013961732, 'timing/model': 1.153856590972282, 'learning_rate/qwen_vl_interface': 8.322122414809678e-06, 'learning_rate/action_model': 8.322122414809677e-05, 'epoch': 1.45}) +[2026-06-19 13:03:38,726][starVLA.training.train_starvla][INFO] - Step 1454, Loss: {'train/loss': 0.09760726243257523, 'train/grad_norm_pre_clip': 5.604989528656006, 'timing/data': 0.013064850005321205, 'timing/model': 1.1549296469893306, 'learning_rate/qwen_vl_interface': 8.319799066902755e-06, 'learning_rate/action_model': 8.319799066902755e-05, 'epoch': 1.45}) +[2026-06-19 13:03:43,310][starVLA.training.train_starvla][INFO] - Step 1455, Loss: {'train/loss': 0.3729858696460724, 'train/grad_norm_pre_clip': 9.692829132080078, 'timing/data': 0.012470450019463897, 'timing/model': 1.15075670205988, 'learning_rate/qwen_vl_interface': 8.317474457117393e-06, 'learning_rate/action_model': 8.317474457117392e-05, 'epoch': 1.45}) +[2026-06-19 13:03:47,888][starVLA.training.train_starvla][INFO] - Step 1456, Loss: {'train/loss': 0.13949929177761078, 'train/grad_norm_pre_clip': 3.7229654788970947, 'timing/data': 0.01360968197695911, 'timing/model': 1.3043964420212433, 'learning_rate/qwen_vl_interface': 8.315148586409148e-06, 'learning_rate/action_model': 8.315148586409148e-05, 'epoch': 1.45}) +[2026-06-19 13:03:52,394][starVLA.training.train_starvla][INFO] - Step 1457, Loss: {'train/loss': 0.16741956770420074, 'train/grad_norm_pre_clip': 9.177775382995605, 'timing/data': 0.013068692060187459, 'timing/model': 1.2390775079838932, 'learning_rate/qwen_vl_interface': 8.312821455734104e-06, 'learning_rate/action_model': 8.312821455734103e-05, 'epoch': 1.45}) +[2026-06-19 13:03:56,808][starVLA.training.train_starvla][INFO] - Step 1458, Loss: {'train/loss': 0.2643079161643982, 'train/grad_norm_pre_clip': 11.752453804016113, 'timing/data': 0.013034073053859174, 'timing/model': 1.143519083969295, 'learning_rate/qwen_vl_interface': 8.31049306604885e-06, 'learning_rate/action_model': 8.310493066048849e-05, 'epoch': 1.45}) +[2026-06-19 13:04:01,429][starVLA.training.train_starvla][INFO] - Step 1459, Loss: {'train/loss': 0.14525841176509857, 'train/grad_norm_pre_clip': 5.815979957580566, 'timing/data': 0.013479797984473407, 'timing/model': 1.1531250639818609, 'learning_rate/qwen_vl_interface': 8.3081634183105e-06, 'learning_rate/action_model': 8.308163418310501e-05, 'epoch': 1.45}) +[2026-06-19 13:04:06,047][starVLA.training.train_starvla][INFO] - Step 1460, Loss: {'train/loss': 0.07945435494184494, 'train/grad_norm_pre_clip': 4.443658351898193, 'timing/data': 0.01338043890427798, 'timing/model': 1.1457514190115035, 'learning_rate/qwen_vl_interface': 8.305832513476689e-06, 'learning_rate/action_model': 8.305832513476688e-05, 'epoch': 1.46}) +[2026-06-19 13:04:10,634][starVLA.training.train_starvla][INFO] - Step 1461, Loss: {'train/loss': 0.3781364858150482, 'train/grad_norm_pre_clip': 11.304495811462402, 'timing/data': 0.01324196602217853, 'timing/model': 1.1539695280371234, 'learning_rate/qwen_vl_interface': 8.303500352505557e-06, 'learning_rate/action_model': 8.303500352505557e-05, 'epoch': 1.46}) +[2026-06-19 13:04:15,245][starVLA.training.train_starvla][INFO] - Step 1462, Loss: {'train/loss': 0.11057533323764801, 'train/grad_norm_pre_clip': 4.933699607849121, 'timing/data': 0.013672497938387096, 'timing/model': 1.3172651699278504, 'learning_rate/qwen_vl_interface': 8.301166936355773e-06, 'learning_rate/action_model': 8.301166936355773e-05, 'epoch': 1.46}) +[2026-06-19 13:04:19,805][starVLA.training.train_starvla][INFO] - Step 1463, Loss: {'train/loss': 0.19191057980060577, 'train/grad_norm_pre_clip': 6.600498199462891, 'timing/data': 0.012966174981556833, 'timing/model': 1.258697826997377, 'learning_rate/qwen_vl_interface': 8.298832265986514e-06, 'learning_rate/action_model': 8.298832265986513e-05, 'epoch': 1.46}) +[2026-06-19 13:04:24,246][starVLA.training.train_starvla][INFO] - Step 1464, Loss: {'train/loss': 0.08902239054441452, 'train/grad_norm_pre_clip': 14.927236557006836, 'timing/data': 0.016785096027888358, 'timing/model': 1.1534205731004477, 'learning_rate/qwen_vl_interface': 8.296496342357475e-06, 'learning_rate/action_model': 8.296496342357474e-05, 'epoch': 1.46}) +[2026-06-19 13:04:28,846][starVLA.training.train_starvla][INFO] - Step 1465, Loss: {'train/loss': 0.3021377921104431, 'train/grad_norm_pre_clip': 7.571967601776123, 'timing/data': 0.013655468937940896, 'timing/model': 1.1610245659248903, 'learning_rate/qwen_vl_interface': 8.294159166428864e-06, 'learning_rate/action_model': 8.294159166428864e-05, 'epoch': 1.46}) +[2026-06-19 13:04:33,467][starVLA.training.train_starvla][INFO] - Step 1466, Loss: {'train/loss': 0.3428981900215149, 'train/grad_norm_pre_clip': 13.011337280273438, 'timing/data': 0.012863144977018237, 'timing/model': 1.1546609729994088, 'learning_rate/qwen_vl_interface': 8.29182073916141e-06, 'learning_rate/action_model': 8.29182073916141e-05, 'epoch': 1.46}) +[2026-06-19 13:04:38,084][starVLA.training.train_starvla][INFO] - Step 1467, Loss: {'train/loss': 0.4642728865146637, 'train/grad_norm_pre_clip': 11.857828140258789, 'timing/data': 0.013512203004211187, 'timing/model': 1.1546214040135965, 'learning_rate/qwen_vl_interface': 8.28948106151635e-06, 'learning_rate/action_model': 8.28948106151635e-05, 'epoch': 1.46}) +[2026-06-19 13:04:42,707][starVLA.training.train_starvla][INFO] - Step 1468, Loss: {'train/loss': 0.5238162875175476, 'train/grad_norm_pre_clip': 16.508474349975586, 'timing/data': 0.013178627006709576, 'timing/model': 1.3213422480039299, 'learning_rate/qwen_vl_interface': 8.287140134455435e-06, 'learning_rate/action_model': 8.287140134455435e-05, 'epoch': 1.46}) +[2026-06-19 13:04:47,193][starVLA.training.train_starvla][INFO] - Step 1469, Loss: {'train/loss': 0.14227378368377686, 'train/grad_norm_pre_clip': 7.489731788635254, 'timing/data': 0.01258400606457144, 'timing/model': 1.153833000920713, 'learning_rate/qwen_vl_interface': 8.284797958940934e-06, 'learning_rate/action_model': 8.284797958940935e-05, 'epoch': 1.46}) +[2026-06-19 13:04:51,797][starVLA.training.train_starvla][INFO] - Step 1470, Loss: {'train/loss': 0.13797536492347717, 'train/grad_norm_pre_clip': 5.363968849182129, 'timing/data': 0.01324798702262342, 'timing/model': 1.1473526710178703, 'learning_rate/qwen_vl_interface': 8.282454535935626e-06, 'learning_rate/action_model': 8.282454535935626e-05, 'epoch': 1.47}) +[2026-06-19 13:04:56,407][starVLA.training.train_starvla][INFO] - Step 1471, Loss: {'train/loss': 0.3031802773475647, 'train/grad_norm_pre_clip': 6.029438495635986, 'timing/data': 0.013158780988305807, 'timing/model': 1.1441183580318466, 'learning_rate/qwen_vl_interface': 8.280109866402805e-06, 'learning_rate/action_model': 8.280109866402806e-05, 'epoch': 1.47}) +[2026-06-19 13:05:01,012][starVLA.training.train_starvla][INFO] - Step 1472, Loss: {'train/loss': 0.22739467024803162, 'train/grad_norm_pre_clip': 6.6924309730529785, 'timing/data': 0.012664722977206111, 'timing/model': 1.148023986024782, 'learning_rate/qwen_vl_interface': 8.277763951306277e-06, 'learning_rate/action_model': 8.277763951306277e-05, 'epoch': 1.47}) +[2026-06-19 13:05:05,588][starVLA.training.train_starvla][INFO] - Step 1473, Loss: {'train/loss': 0.1953911930322647, 'train/grad_norm_pre_clip': 4.943807125091553, 'timing/data': 0.01264279696624726, 'timing/model': 1.1403135060099885, 'learning_rate/qwen_vl_interface': 8.275416791610356e-06, 'learning_rate/action_model': 8.275416791610355e-05, 'epoch': 1.47}) +[2026-06-19 13:05:10,174][starVLA.training.train_starvla][INFO] - Step 1474, Loss: {'train/loss': 0.19172148406505585, 'train/grad_norm_pre_clip': 6.380833148956299, 'timing/data': 0.013632824062369764, 'timing/model': 1.3086802510078996, 'learning_rate/qwen_vl_interface': 8.27306838827987e-06, 'learning_rate/action_model': 8.273068388279871e-05, 'epoch': 1.47}) +[2026-06-19 13:05:14,701][starVLA.training.train_starvla][INFO] - Step 1475, Loss: {'train/loss': 0.1864459216594696, 'train/grad_norm_pre_clip': 5.428098678588867, 'timing/data': 0.012991359923034906, 'timing/model': 1.2483166119782254, 'learning_rate/qwen_vl_interface': 8.270718742280162e-06, 'learning_rate/action_model': 8.270718742280162e-05, 'epoch': 1.47}) +[2026-06-19 13:05:19,150][starVLA.training.train_starvla][INFO] - Step 1476, Loss: {'train/loss': 0.23877567052841187, 'train/grad_norm_pre_clip': 5.699018478393555, 'timing/data': 0.013487934018485248, 'timing/model': 1.1529606980038807, 'learning_rate/qwen_vl_interface': 8.268367854577084e-06, 'learning_rate/action_model': 8.268367854577082e-05, 'epoch': 1.47}) +[2026-06-19 13:05:23,763][starVLA.training.train_starvla][INFO] - Step 1477, Loss: {'train/loss': 0.25977352261543274, 'train/grad_norm_pre_clip': 5.046758651733398, 'timing/data': 0.013186470954678953, 'timing/model': 1.151914150104858, 'learning_rate/qwen_vl_interface': 8.266015726136991e-06, 'learning_rate/action_model': 8.266015726136991e-05, 'epoch': 1.47}) +[2026-06-19 13:05:28,398][starVLA.training.train_starvla][INFO] - Step 1478, Loss: {'train/loss': 0.14865516126155853, 'train/grad_norm_pre_clip': 3.5338282585144043, 'timing/data': 0.013250429998151958, 'timing/model': 1.1518020599614829, 'learning_rate/qwen_vl_interface': 8.26366235792676e-06, 'learning_rate/action_model': 8.26366235792676e-05, 'epoch': 1.47}) +[2026-06-19 13:05:32,985][starVLA.training.train_starvla][INFO] - Step 1479, Loss: {'train/loss': 0.14596985280513763, 'train/grad_norm_pre_clip': 6.638970375061035, 'timing/data': 0.01241585600655526, 'timing/model': 1.1449985499493778, 'learning_rate/qwen_vl_interface': 8.261307750913769e-06, 'learning_rate/action_model': 8.261307750913769e-05, 'epoch': 1.47}) +[2026-06-19 13:05:37,575][starVLA.training.train_starvla][INFO] - Step 1480, Loss: {'train/loss': 0.269338458776474, 'train/grad_norm_pre_clip': 6.374735355377197, 'timing/data': 0.013379652984440327, 'timing/model': 1.3178866170346737, 'learning_rate/qwen_vl_interface': 8.258951906065908e-06, 'learning_rate/action_model': 8.258951906065908e-05, 'epoch': 1.48}) +[2026-06-19 13:05:42,096][starVLA.training.train_starvla][INFO] - Step 1481, Loss: {'train/loss': 0.17534790933132172, 'train/grad_norm_pre_clip': 4.587197780609131, 'timing/data': 0.012121791020035744, 'timing/model': 1.2515028109773993, 'learning_rate/qwen_vl_interface': 8.256594824351577e-06, 'learning_rate/action_model': 8.256594824351576e-05, 'epoch': 1.48}) +[2026-06-19 13:05:46,521][starVLA.training.train_starvla][INFO] - Step 1482, Loss: {'train/loss': 0.2657450735569, 'train/grad_norm_pre_clip': 11.71959400177002, 'timing/data': 0.013206162955611944, 'timing/model': 1.1430297660408542, 'learning_rate/qwen_vl_interface': 8.254236506739683e-06, 'learning_rate/action_model': 8.254236506739681e-05, 'epoch': 1.48}) +[2026-06-19 13:05:51,126][starVLA.training.train_starvla][INFO] - Step 1483, Loss: {'train/loss': 0.2155054211616516, 'train/grad_norm_pre_clip': 4.057598114013672, 'timing/data': 0.012952702003531158, 'timing/model': 1.1470935510005802, 'learning_rate/qwen_vl_interface': 8.25187695419964e-06, 'learning_rate/action_model': 8.25187695419964e-05, 'epoch': 1.48}) +[2026-06-19 13:05:55,749][starVLA.training.train_starvla][INFO] - Step 1484, Loss: {'train/loss': 0.21765068173408508, 'train/grad_norm_pre_clip': 4.682457447052002, 'timing/data': 0.012987693073228002, 'timing/model': 1.151705019059591, 'learning_rate/qwen_vl_interface': 8.249516167701373e-06, 'learning_rate/action_model': 8.249516167701374e-05, 'epoch': 1.48}) +[2026-06-19 13:06:00,352][starVLA.training.train_starvla][INFO] - Step 1485, Loss: {'train/loss': 0.09784439951181412, 'train/grad_norm_pre_clip': 5.143435478210449, 'timing/data': 0.013006636057980359, 'timing/model': 1.153491217060946, 'learning_rate/qwen_vl_interface': 8.247154148215311e-06, 'learning_rate/action_model': 8.24715414821531e-05, 'epoch': 1.48}) +[2026-06-19 13:06:04,948][starVLA.training.train_starvla][INFO] - Step 1486, Loss: {'train/loss': 0.18209706246852875, 'train/grad_norm_pre_clip': 6.371407985687256, 'timing/data': 0.013127137906849384, 'timing/model': 1.320875804987736, 'learning_rate/qwen_vl_interface': 8.24479089671239e-06, 'learning_rate/action_model': 8.24479089671239e-05, 'epoch': 1.48}) +[2026-06-19 13:06:09,468][starVLA.training.train_starvla][INFO] - Step 1487, Loss: {'train/loss': 0.35699477791786194, 'train/grad_norm_pre_clip': 14.181591033935547, 'timing/data': 0.01321895804721862, 'timing/model': 1.2418074499582872, 'learning_rate/qwen_vl_interface': 8.242426414164056e-06, 'learning_rate/action_model': 8.242426414164056e-05, 'epoch': 1.48}) +[2026-06-19 13:06:13,889][starVLA.training.train_starvla][INFO] - Step 1488, Loss: {'train/loss': 0.47021231055259705, 'train/grad_norm_pre_clip': 10.821264266967773, 'timing/data': 0.013348786975257099, 'timing/model': 1.1459643449634314, 'learning_rate/qwen_vl_interface': 8.240060701542257e-06, 'learning_rate/action_model': 8.240060701542257e-05, 'epoch': 1.48}) +[2026-06-19 13:06:18,474][starVLA.training.train_starvla][INFO] - Step 1489, Loss: {'train/loss': 0.08485870808362961, 'train/grad_norm_pre_clip': 4.67027473449707, 'timing/data': 0.01251628901809454, 'timing/model': 1.1428888199152425, 'learning_rate/qwen_vl_interface': 8.237693759819447e-06, 'learning_rate/action_model': 8.237693759819448e-05, 'epoch': 1.48}) +[2026-06-19 13:06:23,057][starVLA.training.train_starvla][INFO] - Step 1490, Loss: {'train/loss': 0.1993892788887024, 'train/grad_norm_pre_clip': 10.1199369430542, 'timing/data': 0.013111857930198312, 'timing/model': 1.1437908000079915, 'learning_rate/qwen_vl_interface': 8.235325589968587e-06, 'learning_rate/action_model': 8.235325589968587e-05, 'epoch': 1.49}) +[2026-06-19 13:06:27,650][starVLA.training.train_starvla][INFO] - Step 1491, Loss: {'train/loss': 0.1485571265220642, 'train/grad_norm_pre_clip': 4.542950630187988, 'timing/data': 0.012523583020083606, 'timing/model': 1.143674779101275, 'learning_rate/qwen_vl_interface': 8.232956192963143e-06, 'learning_rate/action_model': 8.232956192963144e-05, 'epoch': 1.49}) +[2026-06-19 13:06:32,231][starVLA.training.train_starvla][INFO] - Step 1492, Loss: {'train/loss': 0.2019360363483429, 'train/grad_norm_pre_clip': 7.848397254943848, 'timing/data': 0.01343468704726547, 'timing/model': 1.3108147060265765, 'learning_rate/qwen_vl_interface': 8.230585569777086e-06, 'learning_rate/action_model': 8.230585569777085e-05, 'epoch': 1.49}) +[2026-06-19 13:06:36,768][starVLA.training.train_starvla][INFO] - Step 1493, Loss: {'train/loss': 0.21809141337871552, 'train/grad_norm_pre_clip': 8.399097442626953, 'timing/data': 0.013164616073481739, 'timing/model': 1.2667404060484841, 'learning_rate/qwen_vl_interface': 8.228213721384885e-06, 'learning_rate/action_model': 8.228213721384885e-05, 'epoch': 1.49}) +[2026-06-19 13:06:41,192][starVLA.training.train_starvla][INFO] - Step 1494, Loss: {'train/loss': 0.15953204035758972, 'train/grad_norm_pre_clip': 4.725226402282715, 'timing/data': 0.012786149978637695, 'timing/model': 1.1448642780305818, 'learning_rate/qwen_vl_interface': 8.225840648761523e-06, 'learning_rate/action_model': 8.225840648761523e-05, 'epoch': 1.49}) +[2026-06-19 13:06:45,828][starVLA.training.train_starvla][INFO] - Step 1495, Loss: {'train/loss': 0.17522823810577393, 'train/grad_norm_pre_clip': 5.050666809082031, 'timing/data': 0.012751524918712676, 'timing/model': 1.1510215169982985, 'learning_rate/qwen_vl_interface': 8.223466352882476e-06, 'learning_rate/action_model': 8.223466352882475e-05, 'epoch': 1.49}) +[2026-06-19 13:06:50,457][starVLA.training.train_starvla][INFO] - Step 1496, Loss: {'train/loss': 0.375775545835495, 'train/grad_norm_pre_clip': 14.892256736755371, 'timing/data': 0.013205157010816038, 'timing/model': 1.1518853210145608, 'learning_rate/qwen_vl_interface': 8.22109083472373e-06, 'learning_rate/action_model': 8.22109083472373e-05, 'epoch': 1.49}) +[2026-06-19 13:06:55,053][starVLA.training.train_starvla][INFO] - Step 1497, Loss: {'train/loss': 0.29680219292640686, 'train/grad_norm_pre_clip': 21.614988327026367, 'timing/data': 0.013158367946743965, 'timing/model': 1.143712871009484, 'learning_rate/qwen_vl_interface': 8.218714095261769e-06, 'learning_rate/action_model': 8.218714095261768e-05, 'epoch': 1.49}) +[2026-06-19 13:06:59,648][starVLA.training.train_starvla][INFO] - Step 1498, Loss: {'train/loss': 0.23237691819667816, 'train/grad_norm_pre_clip': 8.319526672363281, 'timing/data': 0.014044171082787216, 'timing/model': 1.3128203980159014, 'learning_rate/qwen_vl_interface': 8.21633613547358e-06, 'learning_rate/action_model': 8.21633613547358e-05, 'epoch': 1.49}) +[2026-06-19 13:07:04,226][starVLA.training.train_starvla][INFO] - Step 1499, Loss: {'train/loss': 0.1903333067893982, 'train/grad_norm_pre_clip': 7.244655609130859, 'timing/data': 0.012790068984031677, 'timing/model': 1.2780356319854036, 'learning_rate/qwen_vl_interface': 8.213956956336656e-06, 'learning_rate/action_model': 8.213956956336656e-05, 'epoch': 1.49}) +[2026-06-19 13:07:08,664][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1500 +[2026-06-19 13:10:01,565][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1500 +[2026-06-19 13:10:01,566][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 1500 +[2026-06-19 13:13:12,342][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 1500 +[2026-06-19 13:13:12,343][starVLA.training.train_starvla][INFO] - Step 1500, Loss: {'train/loss': 0.1383492350578308, 'train/grad_norm_pre_clip': 4.560935974121094, 'eval/loss': 0.36424654722213745, 'eval/latency_0/loss': 0.10970071703195572, 'eval/latency_1/loss': 0.09576283395290375, 'eval/latency_2/loss': 0.3843526244163513, 'eval/latency_3/loss': 0.494913250207901, 'eval/latency_4/loss': 0.7329422235488892, 'eval/flappy/loss': 0.36424654722213745, 'eval/flappy/latency_0/loss': 0.10970071703195572, 'eval/flappy/latency_1/loss': 0.09576283395290375, 'eval/flappy/latency_2/loss': 0.3843526244163513, 'eval/flappy/latency_3/loss': 0.494913250207901, 'eval/flappy/latency_4/loss': 0.7329422235488892, 'eval/flap_precision': 0.7679558011049724, 'eval/flap_recall': 0.5054545454545455, 'eval/flap_f1': 0.6096491228070176, 'eval/flap_rate_model': 0.050277777777777775, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.6096491228070176, 'eval/latency_0/flap_precision': 0.7679558011049724, 'eval/latency_0/flap_recall': 0.5054545454545455, 'eval/latency_0/flap_f1': 0.6096491228070176, 'eval/latency_0/flap_rate_model': 0.050277777777777775, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.6096491228070176, 'timing/data': 0.01286287500988692, 'timing/model': 1.1531447340967134, 'learning_rate/qwen_vl_interface': 8.211576558828985e-06, 'learning_rate/action_model': 8.211576558828985e-05, 'epoch': 1.5}) +[2026-06-19 13:13:12,343][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_1500_state +[2026-06-19 13:13:12,345][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-19 13:13:15,632][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_1500_state/model.safetensors +[2026-06-19 13:13:23,224][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_1500_state/optimizer.bin +[2026-06-19 13:13:23,225][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_1500_state/sampler.bin +[2026-06-19 13:13:23,225][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_1500_state/sampler_1.bin +[2026-06-19 13:13:23,226][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_1500_state/random_states_0.pkl +[2026-06-19 13:13:23,249][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1500 +[2026-06-19 13:13:23,249][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-19 13:13:23,280][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-19 13:13:27,743][starVLA.training.train_starvla][INFO] - Step 1501, Loss: {'train/loss': 0.1093837320804596, 'train/grad_norm_pre_clip': 5.017808437347412, 'timing/data': 0.012990921037271619, 'timing/model': 1.1329190260730684, 'learning_rate/qwen_vl_interface': 8.209194943929059e-06, 'learning_rate/action_model': 8.20919494392906e-05, 'epoch': 1.5}) +[2026-06-19 13:13:32,276][starVLA.training.train_starvla][INFO] - Step 1502, Loss: {'train/loss': 0.11021135747432709, 'train/grad_norm_pre_clip': 5.29235315322876, 'timing/data': 0.012457879027351737, 'timing/model': 1.134187436895445, 'learning_rate/qwen_vl_interface': 8.20681211261587e-06, 'learning_rate/action_model': 8.20681211261587e-05, 'epoch': 1.5}) +[2026-06-19 13:13:36,833][starVLA.training.train_starvla][INFO] - Step 1503, Loss: {'train/loss': 0.24025143682956696, 'train/grad_norm_pre_clip': 5.849104881286621, 'timing/data': 0.013229946023784578, 'timing/model': 1.3093675760319456, 'learning_rate/qwen_vl_interface': 8.204428065868911e-06, 'learning_rate/action_model': 8.204428065868911e-05, 'epoch': 1.5}) +[2026-06-19 13:13:41,232][starVLA.training.train_starvla][INFO] - Step 1504, Loss: {'train/loss': 0.5237390995025635, 'train/grad_norm_pre_clip': 15.544974327087402, 'timing/data': 0.013304794090799987, 'timing/model': 1.139589975005947, 'learning_rate/qwen_vl_interface': 8.202042804668173e-06, 'learning_rate/action_model': 8.202042804668172e-05, 'epoch': 1.5}) +[2026-06-19 13:13:45,819][starVLA.training.train_starvla][INFO] - Step 1505, Loss: {'train/loss': 0.34362727403640747, 'train/grad_norm_pre_clip': 13.082817077636719, 'timing/data': 0.013193367049098015, 'timing/model': 1.1416479439940304, 'learning_rate/qwen_vl_interface': 8.199656329994145e-06, 'learning_rate/action_model': 8.199656329994145e-05, 'epoch': 1.5}) +[2026-06-19 13:13:50,251][starVLA.training.train_starvla][INFO] - Step 1506, Loss: {'train/loss': 0.348760187625885, 'train/grad_norm_pre_clip': 9.502900123596191, 'timing/data': 0.012755580013617873, 'timing/model': 1.129407296073623, 'learning_rate/qwen_vl_interface': 8.19726864282782e-06, 'learning_rate/action_model': 8.197268642827818e-05, 'epoch': 1.5}) +[2026-06-19 13:13:54,709][starVLA.training.train_starvla][INFO] - Step 1507, Loss: {'train/loss': 0.23541805148124695, 'train/grad_norm_pre_clip': 4.114806652069092, 'timing/data': 0.013275563018396497, 'timing/model': 1.134045887971297, 'learning_rate/qwen_vl_interface': 8.194879744150682e-06, 'learning_rate/action_model': 8.194879744150681e-05, 'epoch': 1.5}) +[2026-06-19 13:13:59,278][starVLA.training.train_starvla][INFO] - Step 1508, Loss: {'train/loss': 0.19840402901172638, 'train/grad_norm_pre_clip': 3.3493974208831787, 'timing/data': 0.012616959982551634, 'timing/model': 1.142898871912621, 'learning_rate/qwen_vl_interface': 8.192489634944718e-06, 'learning_rate/action_model': 8.192489634944718e-05, 'epoch': 1.5}) +[2026-06-19 13:14:03,828][starVLA.training.train_starvla][INFO] - Step 1509, Loss: {'train/loss': 0.09196220338344574, 'train/grad_norm_pre_clip': 3.924981117248535, 'timing/data': 0.01418528996873647, 'timing/model': 1.2951254529180005, 'learning_rate/qwen_vl_interface': 8.190098316192416e-06, 'learning_rate/action_model': 8.190098316192416e-05, 'epoch': 1.5}) +[2026-06-19 13:14:08,262][starVLA.training.train_starvla][INFO] - Step 1510, Loss: {'train/loss': 0.2539123594760895, 'train/grad_norm_pre_clip': 8.32514762878418, 'timing/data': 0.01275829307269305, 'timing/model': 1.1357821739511564, 'learning_rate/qwen_vl_interface': 8.18770578887675e-06, 'learning_rate/action_model': 8.18770578887675e-05, 'epoch': 1.51}) +[2026-06-19 13:14:12,826][starVLA.training.train_starvla][INFO] - Step 1511, Loss: {'train/loss': 0.144102543592453, 'train/grad_norm_pre_clip': 5.051580429077148, 'timing/data': 0.015204651048406959, 'timing/model': 1.1364704460138455, 'learning_rate/qwen_vl_interface': 8.185312053981204e-06, 'learning_rate/action_model': 8.185312053981204e-05, 'epoch': 1.51}) +[2026-06-19 13:14:17,291][starVLA.training.train_starvla][INFO] - Step 1512, Loss: {'train/loss': 0.15832403302192688, 'train/grad_norm_pre_clip': 4.238058567047119, 'timing/data': 0.015225640032440424, 'timing/model': 1.1377324960194528, 'learning_rate/qwen_vl_interface': 8.182917112489747e-06, 'learning_rate/action_model': 8.182917112489746e-05, 'epoch': 1.51}) +[2026-06-19 13:14:21,818][starVLA.training.train_starvla][INFO] - Step 1513, Loss: {'train/loss': 0.15923774242401123, 'train/grad_norm_pre_clip': 3.4861056804656982, 'timing/data': 0.014855727087706327, 'timing/model': 1.1623882099520415, 'learning_rate/qwen_vl_interface': 8.18052096538685e-06, 'learning_rate/action_model': 8.18052096538685e-05, 'epoch': 1.51}) +[2026-06-19 13:14:26,402][starVLA.training.train_starvla][INFO] - Step 1514, Loss: {'train/loss': 0.14713016152381897, 'train/grad_norm_pre_clip': 6.53960657119751, 'timing/data': 0.013901406084187329, 'timing/model': 1.1405627799686044, 'learning_rate/qwen_vl_interface': 8.178123613657483e-06, 'learning_rate/action_model': 8.178123613657482e-05, 'epoch': 1.51}) +[2026-06-19 13:14:30,962][starVLA.training.train_starvla][INFO] - Step 1515, Loss: {'train/loss': 0.23732510209083557, 'train/grad_norm_pre_clip': 6.958336353302002, 'timing/data': 0.013654618058353662, 'timing/model': 1.3121492250356823, 'learning_rate/qwen_vl_interface': 8.175725058287102e-06, 'learning_rate/action_model': 8.175725058287102e-05, 'epoch': 1.51}) +[2026-06-19 13:14:35,493][starVLA.training.train_starvla][INFO] - Step 1516, Loss: {'train/loss': 0.17655432224273682, 'train/grad_norm_pre_clip': 3.837871551513672, 'timing/data': 0.012945782043971121, 'timing/model': 1.258458325988613, 'learning_rate/qwen_vl_interface': 8.173325300261664e-06, 'learning_rate/action_model': 8.173325300261663e-05, 'epoch': 1.51}) +[2026-06-19 13:14:39,902][starVLA.training.train_starvla][INFO] - Step 1517, Loss: {'train/loss': 0.3363102972507477, 'train/grad_norm_pre_clip': 5.740531921386719, 'timing/data': 0.012921178946271539, 'timing/model': 1.1408906759461388, 'learning_rate/qwen_vl_interface': 8.170924340567617e-06, 'learning_rate/action_model': 8.170924340567616e-05, 'epoch': 1.51}) +[2026-06-19 13:14:44,473][starVLA.training.train_starvla][INFO] - Step 1518, Loss: {'train/loss': 0.2875942587852478, 'train/grad_norm_pre_clip': 8.537425994873047, 'timing/data': 0.01387696503661573, 'timing/model': 1.140149381943047, 'learning_rate/qwen_vl_interface': 8.168522180191909e-06, 'learning_rate/action_model': 8.168522180191909e-05, 'epoch': 1.51}) +[2026-06-19 13:14:49,062][starVLA.training.train_starvla][INFO] - Step 1519, Loss: {'train/loss': 0.30110806226730347, 'train/grad_norm_pre_clip': 7.260819435119629, 'timing/data': 0.01353103807196021, 'timing/model': 1.1523512289859354, 'learning_rate/qwen_vl_interface': 8.166118820121972e-06, 'learning_rate/action_model': 8.166118820121973e-05, 'epoch': 1.51}) +[2026-06-19 13:14:53,642][starVLA.training.train_starvla][INFO] - Step 1520, Loss: {'train/loss': 0.11728920042514801, 'train/grad_norm_pre_clip': 4.803163528442383, 'timing/data': 0.013154001906514168, 'timing/model': 1.136507094022818, 'learning_rate/qwen_vl_interface': 8.163714261345744e-06, 'learning_rate/action_model': 8.163714261345742e-05, 'epoch': 1.52}) +[2026-06-19 13:14:58,159][starVLA.training.train_starvla][INFO] - Step 1521, Loss: {'train/loss': 0.11432645469903946, 'train/grad_norm_pre_clip': 4.011113166809082, 'timing/data': 0.014231512905098498, 'timing/model': 1.151840440928936, 'learning_rate/qwen_vl_interface': 8.16130850485164e-06, 'learning_rate/action_model': 8.161308504851639e-05, 'epoch': 1.52}) +[2026-06-19 13:15:02,729][starVLA.training.train_starvla][INFO] - Step 1522, Loss: {'train/loss': 0.13826975226402283, 'train/grad_norm_pre_clip': 3.589017868041992, 'timing/data': 0.013480997062288225, 'timing/model': 1.2673978640232235, 'learning_rate/qwen_vl_interface': 8.158901551628579e-06, 'learning_rate/action_model': 8.158901551628579e-05, 'epoch': 1.52}) +[2026-06-19 13:15:07,145][starVLA.training.train_starvla][INFO] - Step 1523, Loss: {'train/loss': 0.4810333251953125, 'train/grad_norm_pre_clip': 9.475384712219238, 'timing/data': 0.013088035047985613, 'timing/model': 1.1412766840076074, 'learning_rate/qwen_vl_interface': 8.156493402665971e-06, 'learning_rate/action_model': 8.15649340266597e-05, 'epoch': 1.52}) +[2026-06-19 13:15:11,717][starVLA.training.train_starvla][INFO] - Step 1524, Loss: {'train/loss': 0.18475019931793213, 'train/grad_norm_pre_clip': 5.987098217010498, 'timing/data': 0.01234342297539115, 'timing/model': 1.1400420459685847, 'learning_rate/qwen_vl_interface': 8.154084058953713e-06, 'learning_rate/action_model': 8.154084058953713e-05, 'epoch': 1.52}) +[2026-06-19 13:15:16,316][starVLA.training.train_starvla][INFO] - Step 1525, Loss: {'train/loss': 0.22542205452919006, 'train/grad_norm_pre_clip': 8.22939682006836, 'timing/data': 0.013806893024593592, 'timing/model': 1.146447006962262, 'learning_rate/qwen_vl_interface': 8.151673521482193e-06, 'learning_rate/action_model': 8.151673521482194e-05, 'epoch': 1.52}) +[2026-06-19 13:15:20,920][starVLA.training.train_starvla][INFO] - Step 1526, Loss: {'train/loss': 0.23888751864433289, 'train/grad_norm_pre_clip': 7.609273433685303, 'timing/data': 0.013351799920201302, 'timing/model': 1.1481303139589727, 'learning_rate/qwen_vl_interface': 8.149261791242297e-06, 'learning_rate/action_model': 8.149261791242297e-05, 'epoch': 1.52}) +[2026-06-19 13:15:25,515][starVLA.training.train_starvla][INFO] - Step 1527, Loss: {'train/loss': 0.1443880796432495, 'train/grad_norm_pre_clip': 4.059034824371338, 'timing/data': 0.013180781970731914, 'timing/model': 1.3214997729519382, 'learning_rate/qwen_vl_interface': 8.146848869225391e-06, 'learning_rate/action_model': 8.146848869225391e-05, 'epoch': 1.52}) +[2026-06-19 13:15:30,022][starVLA.training.train_starvla][INFO] - Step 1528, Loss: {'train/loss': 0.14755043387413025, 'train/grad_norm_pre_clip': 6.914732456207275, 'timing/data': 0.01254293299280107, 'timing/model': 1.242874396033585, 'learning_rate/qwen_vl_interface': 8.14443475642334e-06, 'learning_rate/action_model': 8.144434756423339e-05, 'epoch': 1.52}) +[2026-06-19 13:15:34,430][starVLA.training.train_starvla][INFO] - Step 1529, Loss: {'train/loss': 0.3450395464897156, 'train/grad_norm_pre_clip': 19.829641342163086, 'timing/data': 0.013605805928818882, 'timing/model': 1.1407541260123253, 'learning_rate/qwen_vl_interface': 8.142019453828493e-06, 'learning_rate/action_model': 8.142019453828493e-05, 'epoch': 1.52}) +[2026-06-19 13:15:39,015][starVLA.training.train_starvla][INFO] - Step 1530, Loss: {'train/loss': 0.23904649913311005, 'train/grad_norm_pre_clip': 12.834478378295898, 'timing/data': 0.013748721918091178, 'timing/model': 1.1506890390301123, 'learning_rate/qwen_vl_interface': 8.139602962433688e-06, 'learning_rate/action_model': 8.139602962433688e-05, 'epoch': 1.53}) +[2026-06-19 13:15:43,622][starVLA.training.train_starvla][INFO] - Step 1531, Loss: {'train/loss': 0.30829593539237976, 'train/grad_norm_pre_clip': 7.45774507522583, 'timing/data': 0.013702357886359096, 'timing/model': 1.1442032610066235, 'learning_rate/qwen_vl_interface': 8.137185283232255e-06, 'learning_rate/action_model': 8.137185283232256e-05, 'epoch': 1.53}) +[2026-06-19 13:15:48,220][starVLA.training.train_starvla][INFO] - Step 1532, Loss: {'train/loss': 0.23495222628116608, 'train/grad_norm_pre_clip': 8.068740844726562, 'timing/data': 0.013429487938992679, 'timing/model': 1.150364406988956, 'learning_rate/qwen_vl_interface': 8.134766417218014e-06, 'learning_rate/action_model': 8.134766417218012e-05, 'epoch': 1.53}) +[2026-06-19 13:15:52,821][starVLA.training.train_starvla][INFO] - Step 1533, Loss: {'train/loss': 0.16416791081428528, 'train/grad_norm_pre_clip': 5.7968316078186035, 'timing/data': 0.01335325394757092, 'timing/model': 1.3236382890027016, 'learning_rate/qwen_vl_interface': 8.132346365385262e-06, 'learning_rate/action_model': 8.132346365385261e-05, 'epoch': 1.53}) +[2026-06-19 13:15:57,340][starVLA.training.train_starvla][INFO] - Step 1534, Loss: {'train/loss': 0.35573291778564453, 'train/grad_norm_pre_clip': 6.872283458709717, 'timing/data': 0.013091781991533935, 'timing/model': 1.2526054390473291, 'learning_rate/qwen_vl_interface': 8.129925128728796e-06, 'learning_rate/action_model': 8.129925128728796e-05, 'epoch': 1.53}) +[2026-06-19 13:16:01,754][starVLA.training.train_starvla][INFO] - Step 1535, Loss: {'train/loss': 0.1947796642780304, 'train/grad_norm_pre_clip': 6.871758460998535, 'timing/data': 0.014455843018367887, 'timing/model': 1.1397372739156708, 'learning_rate/qwen_vl_interface': 8.127502708243895e-06, 'learning_rate/action_model': 8.127502708243895e-05, 'epoch': 1.53}) +[2026-06-19 13:16:06,349][starVLA.training.train_starvla][INFO] - Step 1536, Loss: {'train/loss': 0.33898764848709106, 'train/grad_norm_pre_clip': 10.618742942810059, 'timing/data': 0.013436097069643438, 'timing/model': 1.1541910769883543, 'learning_rate/qwen_vl_interface': 8.125079104926324e-06, 'learning_rate/action_model': 8.125079104926323e-05, 'epoch': 1.53}) +[2026-06-19 13:16:10,940][starVLA.training.train_starvla][INFO] - Step 1537, Loss: {'train/loss': 0.05995464697480202, 'train/grad_norm_pre_clip': 2.681339740753174, 'timing/data': 0.014282861957326531, 'timing/model': 1.141202147002332, 'learning_rate/qwen_vl_interface': 8.122654319772332e-06, 'learning_rate/action_model': 8.122654319772332e-05, 'epoch': 1.53}) +[2026-06-19 13:16:15,513][starVLA.training.train_starvla][INFO] - Step 1538, Loss: {'train/loss': 0.2547304928302765, 'train/grad_norm_pre_clip': 8.067313194274902, 'timing/data': 0.01322619104757905, 'timing/model': 1.1448845829581842, 'learning_rate/qwen_vl_interface': 8.120228353778661e-06, 'learning_rate/action_model': 8.12022835377866e-05, 'epoch': 1.53}) +[2026-06-19 13:16:20,115][starVLA.training.train_starvla][INFO] - Step 1539, Loss: {'train/loss': 0.12811915576457977, 'train/grad_norm_pre_clip': 5.563538551330566, 'timing/data': 0.015804657014086843, 'timing/model': 1.3334097999613732, 'learning_rate/qwen_vl_interface': 8.117801207942528e-06, 'learning_rate/action_model': 8.117801207942528e-05, 'epoch': 1.53}) +[2026-06-19 13:16:24,538][starVLA.training.train_starvla][INFO] - Step 1540, Loss: {'train/loss': 0.1328386515378952, 'train/grad_norm_pre_clip': 5.064854145050049, 'timing/data': 0.012898840941488743, 'timing/model': 1.152989133959636, 'learning_rate/qwen_vl_interface': 8.115372883261648e-06, 'learning_rate/action_model': 8.115372883261647e-05, 'epoch': 1.54}) +[2026-06-19 13:16:29,164][starVLA.training.train_starvla][INFO] - Step 1541, Loss: {'train/loss': 0.20765815675258636, 'train/grad_norm_pre_clip': 5.829339504241943, 'timing/data': 0.0161446399288252, 'timing/model': 1.1510144940111786, 'learning_rate/qwen_vl_interface': 8.11294338073421e-06, 'learning_rate/action_model': 8.11294338073421e-05, 'epoch': 1.54}) +[2026-06-19 13:16:33,752][starVLA.training.train_starvla][INFO] - Step 1542, Loss: {'train/loss': 0.34334883093833923, 'train/grad_norm_pre_clip': 8.641185760498047, 'timing/data': 0.013829017989337444, 'timing/model': 1.146531198057346, 'learning_rate/qwen_vl_interface': 8.110512701358892e-06, 'learning_rate/action_model': 8.110512701358891e-05, 'epoch': 1.54}) +[2026-06-19 13:16:38,346][starVLA.training.train_starvla][INFO] - Step 1543, Loss: {'train/loss': 0.16048185527324677, 'train/grad_norm_pre_clip': 5.067288398742676, 'timing/data': 0.013411259045824409, 'timing/model': 1.156842372030951, 'learning_rate/qwen_vl_interface': 8.108080846134852e-06, 'learning_rate/action_model': 8.108080846134852e-05, 'epoch': 1.54}) +[2026-06-19 13:16:42,944][starVLA.training.train_starvla][INFO] - Step 1544, Loss: {'train/loss': 0.2131940871477127, 'train/grad_norm_pre_clip': 5.9729323387146, 'timing/data': 0.01317825901787728, 'timing/model': 1.1423086930299178, 'learning_rate/qwen_vl_interface': 8.105647816061734e-06, 'learning_rate/action_model': 8.105647816061735e-05, 'epoch': 1.54}) +[2026-06-19 13:16:47,554][starVLA.training.train_starvla][INFO] - Step 1545, Loss: {'train/loss': 0.11563561856746674, 'train/grad_norm_pre_clip': 4.891505718231201, 'timing/data': 0.0138265669811517, 'timing/model': 1.331385585013777, 'learning_rate/qwen_vl_interface': 8.103213612139667e-06, 'learning_rate/action_model': 8.103213612139667e-05, 'epoch': 1.54}) +[2026-06-19 13:16:52,124][starVLA.training.train_starvla][INFO] - Step 1546, Loss: {'train/loss': 0.13717471063137054, 'train/grad_norm_pre_clip': 3.619605541229248, 'timing/data': 0.012489194981753826, 'timing/model': 1.272700808942318, 'learning_rate/qwen_vl_interface': 8.100778235369262e-06, 'learning_rate/action_model': 8.10077823536926e-05, 'epoch': 1.54}) +[2026-06-19 13:16:56,566][starVLA.training.train_starvla][INFO] - Step 1547, Loss: {'train/loss': 0.20096580684185028, 'train/grad_norm_pre_clip': 13.594889640808105, 'timing/data': 0.012994057964533567, 'timing/model': 1.1582580329850316, 'learning_rate/qwen_vl_interface': 8.098341686751605e-06, 'learning_rate/action_model': 8.098341686751605e-05, 'epoch': 1.54}) +[2026-06-19 13:17:01,176][starVLA.training.train_starvla][INFO] - Step 1548, Loss: {'train/loss': 0.12712697684764862, 'train/grad_norm_pre_clip': 4.060013771057129, 'timing/data': 0.013987374026328325, 'timing/model': 1.1509699439629912, 'learning_rate/qwen_vl_interface': 8.095903967288273e-06, 'learning_rate/action_model': 8.095903967288272e-05, 'epoch': 1.54}) +[2026-06-19 13:17:05,766][starVLA.training.train_starvla][INFO] - Step 1549, Loss: {'train/loss': 0.14819008111953735, 'train/grad_norm_pre_clip': 9.683905601501465, 'timing/data': 0.012813193956390023, 'timing/model': 1.1525088660418987, 'learning_rate/qwen_vl_interface': 8.093465077981318e-06, 'learning_rate/action_model': 8.093465077981318e-05, 'epoch': 1.54}) +[2026-06-19 13:17:10,343][starVLA.training.train_starvla][INFO] - Step 1550, Loss: {'train/loss': 0.1788223534822464, 'train/grad_norm_pre_clip': 8.843829154968262, 'timing/data': 0.012901508016511798, 'timing/model': 1.1423467539716512, 'learning_rate/qwen_vl_interface': 8.091025019833276e-06, 'learning_rate/action_model': 8.091025019833276e-05, 'epoch': 1.55}) +[2026-06-19 13:17:14,925][starVLA.training.train_starvla][INFO] - Step 1551, Loss: {'train/loss': 0.18255265057086945, 'train/grad_norm_pre_clip': 4.4153571128845215, 'timing/data': 0.020057310932315886, 'timing/model': 1.1464799179229885, 'learning_rate/qwen_vl_interface': 8.088583793847165e-06, 'learning_rate/action_model': 8.088583793847164e-05, 'epoch': 1.55}) +[2026-06-19 13:17:19,521][starVLA.training.train_starvla][INFO] - Step 1552, Loss: {'train/loss': 0.09547203779220581, 'train/grad_norm_pre_clip': 4.515132904052734, 'timing/data': 0.01329174404963851, 'timing/model': 1.3247759980149567, 'learning_rate/qwen_vl_interface': 8.086141401026478e-06, 'learning_rate/action_model': 8.086141401026478e-05, 'epoch': 1.55}) +[2026-06-19 13:17:23,952][starVLA.training.train_starvla][INFO] - Step 1553, Loss: {'train/loss': 0.16366474330425262, 'train/grad_norm_pre_clip': 8.699747085571289, 'timing/data': 0.013588552945293486, 'timing/model': 1.1568545170594007, 'learning_rate/qwen_vl_interface': 8.083697842375193e-06, 'learning_rate/action_model': 8.083697842375193e-05, 'epoch': 1.55}) +[2026-06-19 13:17:28,533][starVLA.training.train_starvla][INFO] - Step 1554, Loss: {'train/loss': 0.3496500551700592, 'train/grad_norm_pre_clip': 8.51312255859375, 'timing/data': 0.012788936030119658, 'timing/model': 1.1477173520252109, 'learning_rate/qwen_vl_interface': 8.081253118897762e-06, 'learning_rate/action_model': 8.081253118897761e-05, 'epoch': 1.55}) +[2026-06-19 13:17:33,113][starVLA.training.train_starvla][INFO] - Step 1555, Loss: {'train/loss': 0.18943019211292267, 'train/grad_norm_pre_clip': 6.933098793029785, 'timing/data': 0.013002777006477118, 'timing/model': 1.1505338529823348, 'learning_rate/qwen_vl_interface': 8.07880723159912e-06, 'learning_rate/action_model': 8.07880723159912e-05, 'epoch': 1.55}) +[2026-06-19 13:17:37,691][starVLA.training.train_starvla][INFO] - Step 1556, Loss: {'train/loss': 0.07129635661840439, 'train/grad_norm_pre_clip': 4.23820686340332, 'timing/data': 0.013029137044213712, 'timing/model': 1.1425329049816355, 'learning_rate/qwen_vl_interface': 8.076360181484676e-06, 'learning_rate/action_model': 8.076360181484676e-05, 'epoch': 1.55}) +[2026-06-19 13:17:42,270][starVLA.training.train_starvla][INFO] - Step 1557, Loss: {'train/loss': 0.06842589378356934, 'train/grad_norm_pre_clip': 4.397421836853027, 'timing/data': 0.012941716937348247, 'timing/model': 1.1434981460915878, 'learning_rate/qwen_vl_interface': 8.073911969560326e-06, 'learning_rate/action_model': 8.073911969560326e-05, 'epoch': 1.55}) +[2026-06-19 13:17:46,862][starVLA.training.train_starvla][INFO] - Step 1558, Loss: {'train/loss': 0.22880896925926208, 'train/grad_norm_pre_clip': 10.937472343444824, 'timing/data': 0.012925085960887372, 'timing/model': 1.3107479059835896, 'learning_rate/qwen_vl_interface': 8.071462596832432e-06, 'learning_rate/action_model': 8.071462596832432e-05, 'epoch': 1.55}) +[2026-06-19 13:17:51,277][starVLA.training.train_starvla][INFO] - Step 1559, Loss: {'train/loss': 0.2255849391222, 'train/grad_norm_pre_clip': 7.293002128601074, 'timing/data': 0.012642544927075505, 'timing/model': 1.1480858400464058, 'learning_rate/qwen_vl_interface': 8.06901206430784e-06, 'learning_rate/action_model': 8.06901206430784e-05, 'epoch': 1.55}) +[2026-06-19 13:17:55,870][starVLA.training.train_starvla][INFO] - Step 1560, Loss: {'train/loss': 0.10026196390390396, 'train/grad_norm_pre_clip': 3.8905105590820312, 'timing/data': 0.01315093704033643, 'timing/model': 1.1566153699532151, 'learning_rate/qwen_vl_interface': 8.066560372993873e-06, 'learning_rate/action_model': 8.066560372993874e-05, 'epoch': 1.56}) +[2026-06-19 13:18:00,359][starVLA.training.train_starvla][INFO] - Step 1561, Loss: {'train/loss': 0.27421942353248596, 'train/grad_norm_pre_clip': 10.9057035446167, 'timing/data': 0.013435845961794257, 'timing/model': 1.1436734090093523, 'learning_rate/qwen_vl_interface': 8.064107523898327e-06, 'learning_rate/action_model': 8.064107523898326e-05, 'epoch': 1.56}) +[2026-06-19 13:18:04,866][starVLA.training.train_starvla][INFO] - Step 1562, Loss: {'train/loss': 0.1471669226884842, 'train/grad_norm_pre_clip': 6.17955207824707, 'timing/data': 0.012273560045287013, 'timing/model': 1.1474728930043057, 'learning_rate/qwen_vl_interface': 8.061653518029476e-06, 'learning_rate/action_model': 8.061653518029476e-05, 'epoch': 1.56}) +[2026-06-19 13:18:09,498][starVLA.training.train_starvla][INFO] - Step 1563, Loss: {'train/loss': 0.23371995985507965, 'train/grad_norm_pre_clip': 8.99444580078125, 'timing/data': 0.013815248035825789, 'timing/model': 1.165126838022843, 'learning_rate/qwen_vl_interface': 8.059198356396069e-06, 'learning_rate/action_model': 8.059198356396068e-05, 'epoch': 1.56}) +[2026-06-19 13:18:14,090][starVLA.training.train_starvla][INFO] - Step 1564, Loss: {'train/loss': 0.22442218661308289, 'train/grad_norm_pre_clip': 8.850400924682617, 'timing/data': 0.01325849094428122, 'timing/model': 1.3068501350935549, 'learning_rate/qwen_vl_interface': 8.05674204000733e-06, 'learning_rate/action_model': 8.05674204000733e-05, 'epoch': 1.56}) +[2026-06-19 13:18:18,521][starVLA.training.train_starvla][INFO] - Step 1565, Loss: {'train/loss': 0.13543908298015594, 'train/grad_norm_pre_clip': 3.3083853721618652, 'timing/data': 0.013348374050110579, 'timing/model': 1.1467465279856697, 'learning_rate/qwen_vl_interface': 8.05428456987296e-06, 'learning_rate/action_model': 8.05428456987296e-05, 'epoch': 1.56}) +[2026-06-19 13:18:23,110][starVLA.training.train_starvla][INFO] - Step 1566, Loss: {'train/loss': 0.2075776308774948, 'train/grad_norm_pre_clip': 7.3664422035217285, 'timing/data': 0.012592382961884141, 'timing/model': 1.149888526997529, 'learning_rate/qwen_vl_interface': 8.05182594700313e-06, 'learning_rate/action_model': 8.05182594700313e-05, 'epoch': 1.56}) +[2026-06-19 13:18:27,708][starVLA.training.train_starvla][INFO] - Step 1567, Loss: {'train/loss': 0.13705448806285858, 'train/grad_norm_pre_clip': 4.283749103546143, 'timing/data': 0.013486598967574537, 'timing/model': 1.1442680539330468, 'learning_rate/qwen_vl_interface': 8.049366172408484e-06, 'learning_rate/action_model': 8.049366172408485e-05, 'epoch': 1.56}) +[2026-06-19 13:18:32,260][starVLA.training.train_starvla][INFO] - Step 1568, Loss: {'train/loss': 0.14016243815422058, 'train/grad_norm_pre_clip': 6.71765661239624, 'timing/data': 0.012676113983616233, 'timing/model': 1.1594019361073151, 'learning_rate/qwen_vl_interface': 8.04690524710015e-06, 'learning_rate/action_model': 8.046905247100151e-05, 'epoch': 1.56}) +[2026-06-19 13:18:36,881][starVLA.training.train_starvla][INFO] - Step 1569, Loss: {'train/loss': 0.24104391038417816, 'train/grad_norm_pre_clip': 9.206598281860352, 'timing/data': 0.013008760986849666, 'timing/model': 1.153155029984191, 'learning_rate/qwen_vl_interface': 8.044443172089715e-06, 'learning_rate/action_model': 8.044443172089714e-05, 'epoch': 1.56}) +[2026-06-19 13:18:41,484][starVLA.training.train_starvla][INFO] - Step 1570, Loss: {'train/loss': 0.14329680800437927, 'train/grad_norm_pre_clip': 6.157890796661377, 'timing/data': 0.013419287977740169, 'timing/model': 1.3167982890736312, 'learning_rate/qwen_vl_interface': 8.041979948389248e-06, 'learning_rate/action_model': 8.041979948389248e-05, 'epoch': 1.57}) +[2026-06-19 13:18:45,940][starVLA.training.train_starvla][INFO] - Step 1571, Loss: {'train/loss': 0.22102540731430054, 'train/grad_norm_pre_clip': 5.4470720291137695, 'timing/data': 0.014666628907434642, 'timing/model': 1.1681143880123273, 'learning_rate/qwen_vl_interface': 8.039515577011287e-06, 'learning_rate/action_model': 8.039515577011288e-05, 'epoch': 1.57}) +[2026-06-19 13:18:50,570][starVLA.training.train_starvla][INFO] - Step 1572, Loss: {'train/loss': 0.24272248148918152, 'train/grad_norm_pre_clip': 11.829607009887695, 'timing/data': 0.01307769794948399, 'timing/model': 1.1437628850108013, 'learning_rate/qwen_vl_interface': 8.037050058968841e-06, 'learning_rate/action_model': 8.03705005896884e-05, 'epoch': 1.57}) +[2026-06-19 13:18:55,222][starVLA.training.train_starvla][INFO] - Step 1573, Loss: {'train/loss': 0.19553372263908386, 'train/grad_norm_pre_clip': 9.362630844116211, 'timing/data': 0.013950586086139083, 'timing/model': 1.1642032359959558, 'learning_rate/qwen_vl_interface': 8.034583395275392e-06, 'learning_rate/action_model': 8.034583395275392e-05, 'epoch': 1.57}) +[2026-06-19 13:18:59,807][starVLA.training.train_starvla][INFO] - Step 1574, Loss: {'train/loss': 0.11998213827610016, 'train/grad_norm_pre_clip': 4.76135778427124, 'timing/data': 0.013243603985756636, 'timing/model': 1.1418926210608333, 'learning_rate/qwen_vl_interface': 8.032115586944893e-06, 'learning_rate/action_model': 8.032115586944892e-05, 'epoch': 1.57}) +[2026-06-19 13:19:04,397][starVLA.training.train_starvla][INFO] - Step 1575, Loss: {'train/loss': 0.11168048530817032, 'train/grad_norm_pre_clip': 4.976623058319092, 'timing/data': 0.013142006006091833, 'timing/model': 1.142139332019724, 'learning_rate/qwen_vl_interface': 8.029646634991762e-06, 'learning_rate/action_model': 8.029646634991762e-05, 'epoch': 1.57}) +[2026-06-19 13:19:09,018][starVLA.training.train_starvla][INFO] - Step 1576, Loss: {'train/loss': 0.3706103563308716, 'train/grad_norm_pre_clip': 9.150799751281738, 'timing/data': 0.014721362036652863, 'timing/model': 1.3376982340123504, 'learning_rate/qwen_vl_interface': 8.0271765404309e-06, 'learning_rate/action_model': 8.0271765404309e-05, 'epoch': 1.57}) +[2026-06-19 13:19:13,447][starVLA.training.train_starvla][INFO] - Step 1577, Loss: {'train/loss': 0.1698959469795227, 'train/grad_norm_pre_clip': 7.802905082702637, 'timing/data': 0.013195559033192694, 'timing/model': 1.1461351049365476, 'learning_rate/qwen_vl_interface': 8.024705304277664e-06, 'learning_rate/action_model': 8.024705304277664e-05, 'epoch': 1.57}) +[2026-06-19 13:19:18,026][starVLA.training.train_starvla][INFO] - Step 1578, Loss: {'train/loss': 0.5466130971908569, 'train/grad_norm_pre_clip': 12.408208847045898, 'timing/data': 0.013596545904874802, 'timing/model': 1.1454057689988986, 'learning_rate/qwen_vl_interface': 8.022232927547888e-06, 'learning_rate/action_model': 8.022232927547886e-05, 'epoch': 1.57}) +[2026-06-19 13:19:22,545][starVLA.training.train_starvla][INFO] - Step 1579, Loss: {'train/loss': 0.1853354275226593, 'train/grad_norm_pre_clip': 7.478822231292725, 'timing/data': 0.012840612907893956, 'timing/model': 1.1471255550859496, 'learning_rate/qwen_vl_interface': 8.01975941125787e-06, 'learning_rate/action_model': 8.01975941125787e-05, 'epoch': 1.57}) +[2026-06-19 13:19:27,081][starVLA.training.train_starvla][INFO] - Step 1580, Loss: {'train/loss': 0.13769997656345367, 'train/grad_norm_pre_clip': 8.35488510131836, 'timing/data': 0.013402273994870484, 'timing/model': 1.1464270869037136, 'learning_rate/qwen_vl_interface': 8.017284756424381e-06, 'learning_rate/action_model': 8.017284756424381e-05, 'epoch': 1.58}) +[2026-06-19 13:19:31,651][starVLA.training.train_starvla][INFO] - Step 1581, Loss: {'train/loss': 0.28219282627105713, 'train/grad_norm_pre_clip': 7.476799011230469, 'timing/data': 0.013035746989771724, 'timing/model': 1.1424128580838442, 'learning_rate/qwen_vl_interface': 8.01480896406466e-06, 'learning_rate/action_model': 8.01480896406466e-05, 'epoch': 1.58}) +[2026-06-19 13:19:36,244][starVLA.training.train_starvla][INFO] - Step 1582, Loss: {'train/loss': 0.2832895815372467, 'train/grad_norm_pre_clip': 8.266678810119629, 'timing/data': 0.013382934033870697, 'timing/model': 1.3091534939594567, 'learning_rate/qwen_vl_interface': 8.01233203519641e-06, 'learning_rate/action_model': 8.012332035196409e-05, 'epoch': 1.58}) +[2026-06-19 13:19:40,658][starVLA.training.train_starvla][INFO] - Step 1583, Loss: {'train/loss': 0.1603117138147354, 'train/grad_norm_pre_clip': 7.548071384429932, 'timing/data': 0.013139626011252403, 'timing/model': 1.144570256001316, 'learning_rate/qwen_vl_interface': 8.009853970837801e-06, 'learning_rate/action_model': 8.009853970837801e-05, 'epoch': 1.58}) +[2026-06-19 13:19:45,245][starVLA.training.train_starvla][INFO] - Step 1584, Loss: {'train/loss': 0.33372044563293457, 'train/grad_norm_pre_clip': 10.647390365600586, 'timing/data': 0.01257745805196464, 'timing/model': 1.1549920960096642, 'learning_rate/qwen_vl_interface': 8.007374772007474e-06, 'learning_rate/action_model': 8.007374772007474e-05, 'epoch': 1.58}) +[2026-06-19 13:19:49,737][starVLA.training.train_starvla][INFO] - Step 1585, Loss: {'train/loss': 0.18937794864177704, 'train/grad_norm_pre_clip': 9.078651428222656, 'timing/data': 0.012711553950794041, 'timing/model': 1.1426061160163954, 'learning_rate/qwen_vl_interface': 8.004894439724533e-06, 'learning_rate/action_model': 8.004894439724533e-05, 'epoch': 1.58}) +[2026-06-19 13:19:54,299][starVLA.training.train_starvla][INFO] - Step 1586, Loss: {'train/loss': 0.27443355321884155, 'train/grad_norm_pre_clip': 6.9640607833862305, 'timing/data': 0.013329677982255816, 'timing/model': 1.1482699569314718, 'learning_rate/qwen_vl_interface': 8.002412975008552e-06, 'learning_rate/action_model': 8.002412975008552e-05, 'epoch': 1.58}) +[2026-06-19 13:19:58,907][starVLA.training.train_starvla][INFO] - Step 1587, Loss: {'train/loss': 0.292648047208786, 'train/grad_norm_pre_clip': 14.21721363067627, 'timing/data': 0.013241766951978207, 'timing/model': 1.1472363079665229, 'learning_rate/qwen_vl_interface': 7.999930378879562e-06, 'learning_rate/action_model': 7.999930378879561e-05, 'epoch': 1.58}) +[2026-06-19 13:20:03,542][starVLA.training.train_starvla][INFO] - Step 1588, Loss: {'train/loss': 0.23949812352657318, 'train/grad_norm_pre_clip': 8.888755798339844, 'timing/data': 0.013056106981821358, 'timing/model': 1.3299365870188922, 'learning_rate/qwen_vl_interface': 7.997446652358067e-06, 'learning_rate/action_model': 7.997446652358068e-05, 'epoch': 1.58}) +[2026-06-19 13:20:07,994][starVLA.training.train_starvla][INFO] - Step 1589, Loss: {'train/loss': 0.2991172969341278, 'train/grad_norm_pre_clip': 10.708462715148926, 'timing/data': 0.016969610936939716, 'timing/model': 1.1521807899698615, 'learning_rate/qwen_vl_interface': 7.994961796465035e-06, 'learning_rate/action_model': 7.994961796465035e-05, 'epoch': 1.58}) +[2026-06-19 13:20:12,597][starVLA.training.train_starvla][INFO] - Step 1590, Loss: {'train/loss': 0.2214159518480301, 'train/grad_norm_pre_clip': 5.9980950355529785, 'timing/data': 0.012869192054495215, 'timing/model': 1.1496478220215067, 'learning_rate/qwen_vl_interface': 7.992475812221894e-06, 'learning_rate/action_model': 7.992475812221894e-05, 'epoch': 1.59}) +[2026-06-19 13:20:17,109][starVLA.training.train_starvla][INFO] - Step 1591, Loss: {'train/loss': 0.3188149631023407, 'train/grad_norm_pre_clip': 8.816376686096191, 'timing/data': 0.013558155042119324, 'timing/model': 1.144986768020317, 'learning_rate/qwen_vl_interface': 7.989988700650538e-06, 'learning_rate/action_model': 7.989988700650538e-05, 'epoch': 1.59}) +[2026-06-19 13:20:21,677][starVLA.training.train_starvla][INFO] - Step 1592, Loss: {'train/loss': 0.27720677852630615, 'train/grad_norm_pre_clip': 6.920379161834717, 'timing/data': 0.012357122963294387, 'timing/model': 1.1561540419934317, 'learning_rate/qwen_vl_interface': 7.987500462773326e-06, 'learning_rate/action_model': 7.987500462773326e-05, 'epoch': 1.59}) +[2026-06-19 13:20:26,268][starVLA.training.train_starvla][INFO] - Step 1593, Loss: {'train/loss': 0.2830944359302521, 'train/grad_norm_pre_clip': 19.647369384765625, 'timing/data': 0.013336753007024527, 'timing/model': 1.1491623620968312, 'learning_rate/qwen_vl_interface': 7.985011099613078e-06, 'learning_rate/action_model': 7.985011099613077e-05, 'epoch': 1.59}) +[2026-06-19 13:20:30,870][starVLA.training.train_starvla][INFO] - Step 1594, Loss: {'train/loss': 0.29913267493247986, 'train/grad_norm_pre_clip': 4.890725612640381, 'timing/data': 0.013164539937861264, 'timing/model': 1.3198246560059488, 'learning_rate/qwen_vl_interface': 7.98252061219308e-06, 'learning_rate/action_model': 7.982520612193078e-05, 'epoch': 1.59}) +[2026-06-19 13:20:35,298][starVLA.training.train_starvla][INFO] - Step 1595, Loss: {'train/loss': 0.3237052261829376, 'train/grad_norm_pre_clip': 14.919706344604492, 'timing/data': 0.012850568047724664, 'timing/model': 1.1433559420984238, 'learning_rate/qwen_vl_interface': 7.980029001537073e-06, 'learning_rate/action_model': 7.980029001537072e-05, 'epoch': 1.59}) +[2026-06-19 13:20:39,896][starVLA.training.train_starvla][INFO] - Step 1596, Loss: {'train/loss': 0.23634673655033112, 'train/grad_norm_pre_clip': 3.9028501510620117, 'timing/data': 0.013046541949734092, 'timing/model': 1.1488287519896403, 'learning_rate/qwen_vl_interface': 7.977536268669264e-06, 'learning_rate/action_model': 7.977536268669264e-05, 'epoch': 1.59}) +[2026-06-19 13:20:44,390][starVLA.training.train_starvla][INFO] - Step 1597, Loss: {'train/loss': 0.28779953718185425, 'train/grad_norm_pre_clip': 8.972762107849121, 'timing/data': 0.013282035011798143, 'timing/model': 1.1451922989217564, 'learning_rate/qwen_vl_interface': 7.975042414614324e-06, 'learning_rate/action_model': 7.975042414614324e-05, 'epoch': 1.59}) +[2026-06-19 13:20:48,953][starVLA.training.train_starvla][INFO] - Step 1598, Loss: {'train/loss': 0.298223078250885, 'train/grad_norm_pre_clip': 4.595699310302734, 'timing/data': 0.013196393963880837, 'timing/model': 1.1617436120286584, 'learning_rate/qwen_vl_interface': 7.972547440397382e-06, 'learning_rate/action_model': 7.972547440397383e-05, 'epoch': 1.59}) +[2026-06-19 13:20:53,565][starVLA.training.train_starvla][INFO] - Step 1599, Loss: {'train/loss': 0.15988950431346893, 'train/grad_norm_pre_clip': 4.571224212646484, 'timing/data': 0.012334285071119666, 'timing/model': 1.14503000699915, 'learning_rate/qwen_vl_interface': 7.970051347044026e-06, 'learning_rate/action_model': 7.970051347044025e-05, 'epoch': 1.59}) +[2026-06-19 13:20:58,189][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1600 +[2026-06-19 13:23:52,766][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1600 +[2026-06-19 13:23:52,767][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 1600 +[2026-06-19 13:27:00,764][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 1600 +[2026-06-19 13:27:00,765][starVLA.training.train_starvla][INFO] - Step 1600, Loss: {'train/loss': 0.15200142562389374, 'train/grad_norm_pre_clip': 3.2165229320526123, 'eval/loss': 0.3200191855430603, 'eval/latency_0/loss': 0.11050744354724884, 'eval/latency_1/loss': 0.11674275249242783, 'eval/latency_2/loss': 0.42898598313331604, 'eval/latency_3/loss': 0.39353179931640625, 'eval/latency_4/loss': 0.538997232913971, 'eval/flappy/loss': 0.3200191855430603, 'eval/flappy/latency_0/loss': 0.11050744354724884, 'eval/flappy/latency_1/loss': 0.11674275249242783, 'eval/flappy/latency_2/loss': 0.42898598313331604, 'eval/flappy/latency_3/loss': 0.39353179931640625, 'eval/flappy/latency_4/loss': 0.538997232913971, 'eval/flap_precision': 0.7522123893805309, 'eval/flap_recall': 0.6181818181818182, 'eval/flap_f1': 0.6786427145708582, 'eval/flap_rate_model': 0.06277777777777778, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.6786427145708582, 'eval/latency_0/flap_precision': 0.7522123893805309, 'eval/latency_0/flap_recall': 0.6181818181818182, 'eval/latency_0/flap_f1': 0.6786427145708582, 'eval/latency_0/flap_rate_model': 0.06277777777777778, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.6786427145708582, 'timing/data': 0.013663082965649664, 'timing/model': 1.3391261639771983, 'learning_rate/qwen_vl_interface': 7.967554135580307e-06, 'learning_rate/action_model': 7.967554135580307e-05, 'epoch': 1.6}) +[2026-06-19 13:27:00,765][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_1600_state +[2026-06-19 13:27:00,768][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-19 13:27:04,162][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_1600_state/model.safetensors +[2026-06-19 13:27:11,441][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_1600_state/optimizer.bin +[2026-06-19 13:27:11,441][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_1600_state/sampler.bin +[2026-06-19 13:27:11,441][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_1600_state/sampler_1.bin +[2026-06-19 13:27:11,442][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_1600_state/random_states_0.pkl +[2026-06-19 13:27:11,467][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1600 +[2026-06-19 13:27:11,467][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-19 13:27:11,509][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-19 13:27:16,011][starVLA.training.train_starvla][INFO] - Step 1601, Loss: {'train/loss': 0.10198970884084702, 'train/grad_norm_pre_clip': 2.80666446685791, 'timing/data': 0.012585498974658549, 'timing/model': 1.2869168729521334, 'learning_rate/qwen_vl_interface': 7.965055807032734e-06, 'learning_rate/action_model': 7.965055807032735e-05, 'epoch': 1.6}) +[2026-06-19 13:27:20,450][starVLA.training.train_starvla][INFO] - Step 1602, Loss: {'train/loss': 0.22739794850349426, 'train/grad_norm_pre_clip': 5.041094779968262, 'timing/data': 0.013820199994370341, 'timing/model': 1.2224039509892464, 'learning_rate/qwen_vl_interface': 7.962556362428275e-06, 'learning_rate/action_model': 7.962556362428274e-05, 'epoch': 1.6}) +[2026-06-19 13:27:24,791][starVLA.training.train_starvla][INFO] - Step 1603, Loss: {'train/loss': 0.32344377040863037, 'train/grad_norm_pre_clip': 8.729421615600586, 'timing/data': 0.013834101962856948, 'timing/model': 1.1274308550637215, 'learning_rate/qwen_vl_interface': 7.960055802794358e-06, 'learning_rate/action_model': 7.960055802794357e-05, 'epoch': 1.6}) +[2026-06-19 13:27:29,385][starVLA.training.train_starvla][INFO] - Step 1604, Loss: {'train/loss': 0.18801118433475494, 'train/grad_norm_pre_clip': 3.6059398651123047, 'timing/data': 0.012966990005224943, 'timing/model': 1.141177403042093, 'learning_rate/qwen_vl_interface': 7.957554129158866e-06, 'learning_rate/action_model': 7.957554129158866e-05, 'epoch': 1.6}) +[2026-06-19 13:27:33,913][starVLA.training.train_starvla][INFO] - Step 1605, Loss: {'train/loss': 0.15549640357494354, 'train/grad_norm_pre_clip': 3.7095236778259277, 'timing/data': 0.011706199031323195, 'timing/model': 1.1307596859987825, 'learning_rate/qwen_vl_interface': 7.955051342550148e-06, 'learning_rate/action_model': 7.955051342550148e-05, 'epoch': 1.6}) +[2026-06-19 13:27:38,483][starVLA.training.train_starvla][INFO] - Step 1606, Loss: {'train/loss': 0.3207637071609497, 'train/grad_norm_pre_clip': 7.000325679779053, 'timing/data': 0.012620204011909664, 'timing/model': 1.138477427070029, 'learning_rate/qwen_vl_interface': 7.952547443997e-06, 'learning_rate/action_model': 7.952547443997e-05, 'epoch': 1.6}) +[2026-06-19 13:27:43,033][starVLA.training.train_starvla][INFO] - Step 1607, Loss: {'train/loss': 0.2049914002418518, 'train/grad_norm_pre_clip': 4.275685787200928, 'timing/data': 0.011112836073152721, 'timing/model': 1.3050097429659218, 'learning_rate/qwen_vl_interface': 7.950042434528681e-06, 'learning_rate/action_model': 7.950042434528682e-05, 'epoch': 1.6}) +[2026-06-19 13:27:47,418][starVLA.training.train_starvla][INFO] - Step 1608, Loss: {'train/loss': 0.1298852413892746, 'train/grad_norm_pre_clip': 5.798364162445068, 'timing/data': 0.012339292094111443, 'timing/model': 1.1344852140173316, 'learning_rate/qwen_vl_interface': 7.94753631517491e-06, 'learning_rate/action_model': 7.947536315174909e-05, 'epoch': 1.6}) +[2026-06-19 13:27:51,967][starVLA.training.train_starvla][INFO] - Step 1609, Loss: {'train/loss': 0.24316786229610443, 'train/grad_norm_pre_clip': 6.77580451965332, 'timing/data': 0.02033599396236241, 'timing/model': 1.1350996700348333, 'learning_rate/qwen_vl_interface': 7.945029086965854e-06, 'learning_rate/action_model': 7.945029086965853e-05, 'epoch': 1.6}) +[2026-06-19 13:27:56,503][starVLA.training.train_starvla][INFO] - Step 1610, Loss: {'train/loss': 0.2679691016674042, 'train/grad_norm_pre_clip': 8.13327407836914, 'timing/data': 0.011210678960196674, 'timing/model': 1.1324476590380073, 'learning_rate/qwen_vl_interface': 7.942520750932138e-06, 'learning_rate/action_model': 7.942520750932138e-05, 'epoch': 1.61}) +[2026-06-19 13:28:01,022][starVLA.training.train_starvla][INFO] - Step 1611, Loss: {'train/loss': 0.21585381031036377, 'train/grad_norm_pre_clip': 5.723505973815918, 'timing/data': 0.011924586025997996, 'timing/model': 1.1408685180358589, 'learning_rate/qwen_vl_interface': 7.940011308104848e-06, 'learning_rate/action_model': 7.940011308104848e-05, 'epoch': 1.61}) +[2026-06-19 13:28:05,585][starVLA.training.train_starvla][INFO] - Step 1612, Loss: {'train/loss': 0.18043072521686554, 'train/grad_norm_pre_clip': 5.945439338684082, 'timing/data': 0.012480514007620513, 'timing/model': 1.1498921000165865, 'learning_rate/qwen_vl_interface': 7.93750075951552e-06, 'learning_rate/action_model': 7.93750075951552e-05, 'epoch': 1.61}) +[2026-06-19 13:28:10,114][starVLA.training.train_starvla][INFO] - Step 1613, Loss: {'train/loss': 0.1814555823802948, 'train/grad_norm_pre_clip': 5.527095317840576, 'timing/data': 0.012434964999556541, 'timing/model': 1.2872434899909422, 'learning_rate/qwen_vl_interface': 7.934989106196145e-06, 'learning_rate/action_model': 7.934989106196146e-05, 'epoch': 1.61}) +[2026-06-19 13:28:14,499][starVLA.training.train_starvla][INFO] - Step 1614, Loss: {'train/loss': 0.148067906498909, 'train/grad_norm_pre_clip': 6.98646354675293, 'timing/data': 0.012612343998625875, 'timing/model': 1.1374109060270712, 'learning_rate/qwen_vl_interface': 7.93247634917917e-06, 'learning_rate/action_model': 7.93247634917917e-05, 'epoch': 1.61}) +[2026-06-19 13:28:19,088][starVLA.training.train_starvla][INFO] - Step 1615, Loss: {'train/loss': 0.17908428609371185, 'train/grad_norm_pre_clip': 7.493725299835205, 'timing/data': 0.011394489905796945, 'timing/model': 1.144454380031675, 'learning_rate/qwen_vl_interface': 7.929962489497494e-06, 'learning_rate/action_model': 7.929962489497493e-05, 'epoch': 1.61}) +[2026-06-19 13:28:23,592][starVLA.training.train_starvla][INFO] - Step 1616, Loss: {'train/loss': 0.23686820268630981, 'train/grad_norm_pre_clip': 14.972222328186035, 'timing/data': 0.011816560057923198, 'timing/model': 1.1556466680485755, 'learning_rate/qwen_vl_interface': 7.927447528184469e-06, 'learning_rate/action_model': 7.927447528184468e-05, 'epoch': 1.61}) +[2026-06-19 13:28:28,110][starVLA.training.train_starvla][INFO] - Step 1617, Loss: {'train/loss': 0.06750685721635818, 'train/grad_norm_pre_clip': 3.1119837760925293, 'timing/data': 0.012117178062908351, 'timing/model': 1.1391610649880022, 'learning_rate/qwen_vl_interface': 7.924931466273901e-06, 'learning_rate/action_model': 7.924931466273902e-05, 'epoch': 1.61}) +[2026-06-19 13:28:32,673][starVLA.training.train_starvla][INFO] - Step 1618, Loss: {'train/loss': 0.14791393280029297, 'train/grad_norm_pre_clip': 6.505038261413574, 'timing/data': 0.012109997915104032, 'timing/model': 1.1413498009787872, 'learning_rate/qwen_vl_interface': 7.92241430480005e-06, 'learning_rate/action_model': 7.922414304800049e-05, 'epoch': 1.61}) +[2026-06-19 13:28:37,236][starVLA.training.train_starvla][INFO] - Step 1619, Loss: {'train/loss': 0.20221272110939026, 'train/grad_norm_pre_clip': 6.335142612457275, 'timing/data': 0.011897105956450105, 'timing/model': 1.30424342607148, 'learning_rate/qwen_vl_interface': 7.919896044797622e-06, 'learning_rate/action_model': 7.91989604479762e-05, 'epoch': 1.61}) +[2026-06-19 13:28:41,635][starVLA.training.train_starvla][INFO] - Step 1620, Loss: {'train/loss': 0.31610307097435, 'train/grad_norm_pre_clip': 14.11928653717041, 'timing/data': 0.01132587599568069, 'timing/model': 1.1475955370115116, 'learning_rate/qwen_vl_interface': 7.91737668730178e-06, 'learning_rate/action_model': 7.91737668730178e-05, 'epoch': 1.62}) +[2026-06-19 13:28:46,222][starVLA.training.train_starvla][INFO] - Step 1621, Loss: {'train/loss': 0.09156709164381027, 'train/grad_norm_pre_clip': 4.546447277069092, 'timing/data': 0.01302925799973309, 'timing/model': 1.1397594000445679, 'learning_rate/qwen_vl_interface': 7.91485623334814e-06, 'learning_rate/action_model': 7.91485623334814e-05, 'epoch': 1.62}) +[2026-06-19 13:28:50,728][starVLA.training.train_starvla][INFO] - Step 1622, Loss: {'train/loss': 0.21150247752666473, 'train/grad_norm_pre_clip': 14.13999080657959, 'timing/data': 0.013026598026044667, 'timing/model': 1.1416563079692423, 'learning_rate/qwen_vl_interface': 7.91233468397276e-06, 'learning_rate/action_model': 7.912334683972761e-05, 'epoch': 1.62}) +[2026-06-19 13:28:55,278][starVLA.training.train_starvla][INFO] - Step 1623, Loss: {'train/loss': 0.3513457179069519, 'train/grad_norm_pre_clip': 10.730504989624023, 'timing/data': 0.012585564982146025, 'timing/model': 1.14895252592396, 'learning_rate/qwen_vl_interface': 7.90981204021216e-06, 'learning_rate/action_model': 7.90981204021216e-05, 'epoch': 1.62}) +[2026-06-19 13:28:59,880][starVLA.training.train_starvla][INFO] - Step 1624, Loss: {'train/loss': 0.07842011749744415, 'train/grad_norm_pre_clip': 3.829008102416992, 'timing/data': 0.012035412946715951, 'timing/model': 1.1468632409814745, 'learning_rate/qwen_vl_interface': 7.907288303103298e-06, 'learning_rate/action_model': 7.907288303103298e-05, 'epoch': 1.62}) +[2026-06-19 13:29:04,464][starVLA.training.train_starvla][INFO] - Step 1625, Loss: {'train/loss': 0.28563594818115234, 'train/grad_norm_pre_clip': 8.683998107910156, 'timing/data': 0.012719006044790149, 'timing/model': 1.2982991930330172, 'learning_rate/qwen_vl_interface': 7.90476347368359e-06, 'learning_rate/action_model': 7.90476347368359e-05, 'epoch': 1.62}) +[2026-06-19 13:29:09,004][starVLA.training.train_starvla][INFO] - Step 1626, Loss: {'train/loss': 0.15035825967788696, 'train/grad_norm_pre_clip': 6.428098201751709, 'timing/data': 0.012820346048101783, 'timing/model': 1.2726215720176697, 'learning_rate/qwen_vl_interface': 7.902237552990897e-06, 'learning_rate/action_model': 7.902237552990897e-05, 'epoch': 1.62}) +[2026-06-19 13:29:13,421][starVLA.training.train_starvla][INFO] - Step 1627, Loss: {'train/loss': 0.16848857700824738, 'train/grad_norm_pre_clip': 6.930466175079346, 'timing/data': 0.0126523730577901, 'timing/model': 1.1402621100423858, 'learning_rate/qwen_vl_interface': 7.89971054206353e-06, 'learning_rate/action_model': 7.89971054206353e-05, 'epoch': 1.62}) +[2026-06-19 13:29:17,985][starVLA.training.train_starvla][INFO] - Step 1628, Loss: {'train/loss': 0.16189472377300262, 'train/grad_norm_pre_clip': 14.817181587219238, 'timing/data': 0.012979915947653353, 'timing/model': 1.1396172160748392, 'learning_rate/qwen_vl_interface': 7.897182441940247e-06, 'learning_rate/action_model': 7.897182441940247e-05, 'epoch': 1.62}) +[2026-06-19 13:29:22,573][starVLA.training.train_starvla][INFO] - Step 1629, Loss: {'train/loss': 0.2807468771934509, 'train/grad_norm_pre_clip': 11.445993423461914, 'timing/data': 0.0125740539515391, 'timing/model': 1.1476151689421386, 'learning_rate/qwen_vl_interface': 7.894653253660256e-06, 'learning_rate/action_model': 7.894653253660256e-05, 'epoch': 1.62}) +[2026-06-19 13:29:27,135][starVLA.training.train_starvla][INFO] - Step 1630, Loss: {'train/loss': 0.1459222435951233, 'train/grad_norm_pre_clip': 11.154818534851074, 'timing/data': 0.012520662043243647, 'timing/model': 1.1437876109266654, 'learning_rate/qwen_vl_interface': 7.89212297826321e-06, 'learning_rate/action_model': 7.892122978263209e-05, 'epoch': 1.63}) +[2026-06-19 13:29:31,607][starVLA.training.train_starvla][INFO] - Step 1631, Loss: {'train/loss': 0.2564646899700165, 'train/grad_norm_pre_clip': 8.809464454650879, 'timing/data': 0.012422985048033297, 'timing/model': 1.1374155130470172, 'learning_rate/qwen_vl_interface': 7.889591616789206e-06, 'learning_rate/action_model': 7.889591616789206e-05, 'epoch': 1.63}) +[2026-06-19 13:29:36,113][starVLA.training.train_starvla][INFO] - Step 1632, Loss: {'train/loss': 0.2946517765522003, 'train/grad_norm_pre_clip': 11.293815612792969, 'timing/data': 0.01308106305077672, 'timing/model': 1.2433171019656584, 'learning_rate/qwen_vl_interface': 7.887059170278799e-06, 'learning_rate/action_model': 7.887059170278797e-05, 'epoch': 1.63}) +[2026-06-19 13:29:40,533][starVLA.training.train_starvla][INFO] - Step 1633, Loss: {'train/loss': 0.042947638779878616, 'train/grad_norm_pre_clip': 2.927358388900757, 'timing/data': 0.01277445605956018, 'timing/model': 1.1484750600066036, 'learning_rate/qwen_vl_interface': 7.884525639772974e-06, 'learning_rate/action_model': 7.884525639772974e-05, 'epoch': 1.63}) +[2026-06-19 13:29:45,121][starVLA.training.train_starvla][INFO] - Step 1634, Loss: {'train/loss': 0.1347332000732422, 'train/grad_norm_pre_clip': 10.367735862731934, 'timing/data': 0.01347996003460139, 'timing/model': 1.150324878981337, 'learning_rate/qwen_vl_interface': 7.881991026313176e-06, 'learning_rate/action_model': 7.881991026313175e-05, 'epoch': 1.63}) +[2026-06-19 13:29:49,690][starVLA.training.train_starvla][INFO] - Step 1635, Loss: {'train/loss': 0.2969687581062317, 'train/grad_norm_pre_clip': 11.834321022033691, 'timing/data': 0.013227356015704572, 'timing/model': 1.1400520280003548, 'learning_rate/qwen_vl_interface': 7.879455330941283e-06, 'learning_rate/action_model': 7.879455330941282e-05, 'epoch': 1.63}) +[2026-06-19 13:29:54,257][starVLA.training.train_starvla][INFO] - Step 1636, Loss: {'train/loss': 0.11349335312843323, 'train/grad_norm_pre_clip': 9.64612865447998, 'timing/data': 0.013235559104941785, 'timing/model': 1.1414772029966116, 'learning_rate/qwen_vl_interface': 7.876918554699627e-06, 'learning_rate/action_model': 7.876918554699627e-05, 'epoch': 1.63}) +[2026-06-19 13:29:58,831][starVLA.training.train_starvla][INFO] - Step 1637, Loss: {'train/loss': 0.1242017149925232, 'train/grad_norm_pre_clip': 7.458793640136719, 'timing/data': 0.013266925932839513, 'timing/model': 1.317800170974806, 'learning_rate/qwen_vl_interface': 7.874380698630982e-06, 'learning_rate/action_model': 7.874380698630982e-05, 'epoch': 1.63}) +[2026-06-19 13:30:03,422][starVLA.training.train_starvla][INFO] - Step 1638, Loss: {'train/loss': 0.18009334802627563, 'train/grad_norm_pre_clip': 10.627259254455566, 'timing/data': 0.013754885992966592, 'timing/model': 1.3175165998982266, 'learning_rate/qwen_vl_interface': 7.871841763778562e-06, 'learning_rate/action_model': 7.871841763778563e-05, 'epoch': 1.63}) +[2026-06-19 13:30:07,845][starVLA.training.train_starvla][INFO] - Step 1639, Loss: {'train/loss': 0.1679389327764511, 'train/grad_norm_pre_clip': 5.679811954498291, 'timing/data': 0.012639302061870694, 'timing/model': 1.153621191973798, 'learning_rate/qwen_vl_interface': 7.86930175118603e-06, 'learning_rate/action_model': 7.86930175118603e-05, 'epoch': 1.63}) +[2026-06-19 13:30:12,436][starVLA.training.train_starvla][INFO] - Step 1640, Loss: {'train/loss': 0.11329937726259232, 'train/grad_norm_pre_clip': 4.560925006866455, 'timing/data': 0.013260746956802905, 'timing/model': 1.145961123984307, 'learning_rate/qwen_vl_interface': 7.866760661897487e-06, 'learning_rate/action_model': 7.866760661897487e-05, 'epoch': 1.64}) +[2026-06-19 13:30:17,042][starVLA.training.train_starvla][INFO] - Step 1641, Loss: {'train/loss': 0.16271889209747314, 'train/grad_norm_pre_clip': 6.481277942657471, 'timing/data': 0.012944122077897191, 'timing/model': 1.166175946011208, 'learning_rate/qwen_vl_interface': 7.86421849695748e-06, 'learning_rate/action_model': 7.864218496957481e-05, 'epoch': 1.64}) +[2026-06-19 13:30:21,572][starVLA.training.train_starvla][INFO] - Step 1642, Loss: {'train/loss': 0.20230571925640106, 'train/grad_norm_pre_clip': 7.681417942047119, 'timing/data': 0.013453998952172697, 'timing/model': 1.1530360800679773, 'learning_rate/qwen_vl_interface': 7.861675257410998e-06, 'learning_rate/action_model': 7.861675257410997e-05, 'epoch': 1.64}) +[2026-06-19 13:30:26,123][starVLA.training.train_starvla][INFO] - Step 1643, Loss: {'train/loss': 0.10664857923984528, 'train/grad_norm_pre_clip': 6.69100284576416, 'timing/data': 0.01271714095491916, 'timing/model': 1.1415801800321788, 'learning_rate/qwen_vl_interface': 7.859130944303468e-06, 'learning_rate/action_model': 7.859130944303467e-05, 'epoch': 1.64}) +[2026-06-19 13:30:30,710][starVLA.training.train_starvla][INFO] - Step 1644, Loss: {'train/loss': 0.04627950116991997, 'train/grad_norm_pre_clip': 2.6698014736175537, 'timing/data': 0.012551046907901764, 'timing/model': 1.3222489680629224, 'learning_rate/qwen_vl_interface': 7.856585558680761e-06, 'learning_rate/action_model': 7.856585558680761e-05, 'epoch': 1.64}) +[2026-06-19 13:30:35,131][starVLA.training.train_starvla][INFO] - Step 1645, Loss: {'train/loss': 0.21686378121376038, 'train/grad_norm_pre_clip': 11.270182609558105, 'timing/data': 0.012737979996018112, 'timing/model': 1.146171994972974, 'learning_rate/qwen_vl_interface': 7.854039101589191e-06, 'learning_rate/action_model': 7.854039101589191e-05, 'epoch': 1.64}) +[2026-06-19 13:30:39,728][starVLA.training.train_starvla][INFO] - Step 1646, Loss: {'train/loss': 0.47172653675079346, 'train/grad_norm_pre_clip': 13.323911666870117, 'timing/data': 0.013102812925353646, 'timing/model': 1.1532019480364397, 'learning_rate/qwen_vl_interface': 7.851491574075509e-06, 'learning_rate/action_model': 7.851491574075509e-05, 'epoch': 1.64}) +[2026-06-19 13:30:44,345][starVLA.training.train_starvla][INFO] - Step 1647, Loss: {'train/loss': 0.20679356157779694, 'train/grad_norm_pre_clip': 7.914804935455322, 'timing/data': 0.012926487950608134, 'timing/model': 1.1439850319875404, 'learning_rate/qwen_vl_interface': 7.848942977186909e-06, 'learning_rate/action_model': 7.848942977186908e-05, 'epoch': 1.64}) +[2026-06-19 13:30:48,952][starVLA.training.train_starvla][INFO] - Step 1648, Loss: {'train/loss': 0.18367554247379303, 'train/grad_norm_pre_clip': 19.273681640625, 'timing/data': 0.013019539066590369, 'timing/model': 1.1563098159385845, 'learning_rate/qwen_vl_interface': 7.84639331197102e-06, 'learning_rate/action_model': 7.84639331197102e-05, 'epoch': 1.64}) +[2026-06-19 13:30:53,476][starVLA.training.train_starvla][INFO] - Step 1649, Loss: {'train/loss': 0.31361544132232666, 'train/grad_norm_pre_clip': 15.935699462890625, 'timing/data': 0.012818799004890025, 'timing/model': 1.1451989880297333, 'learning_rate/qwen_vl_interface': 7.843842579475918e-06, 'learning_rate/action_model': 7.843842579475918e-05, 'epoch': 1.64}) +[2026-06-19 13:30:58,063][starVLA.training.train_starvla][INFO] - Step 1650, Loss: {'train/loss': 0.38502436876296997, 'train/grad_norm_pre_clip': 12.027811050415039, 'timing/data': 0.013582214945927262, 'timing/model': 1.248342972015962, 'learning_rate/qwen_vl_interface': 7.841290780750107e-06, 'learning_rate/action_model': 7.841290780750107e-05, 'epoch': 1.65}) +[2026-06-19 13:31:02,512][starVLA.training.train_starvla][INFO] - Step 1651, Loss: {'train/loss': 0.30228835344314575, 'train/grad_norm_pre_clip': 6.485653877258301, 'timing/data': 0.0133115149801597, 'timing/model': 1.148493509972468, 'learning_rate/qwen_vl_interface': 7.838737916842538e-06, 'learning_rate/action_model': 7.838737916842538e-05, 'epoch': 1.65}) +[2026-06-19 13:31:07,087][starVLA.training.train_starvla][INFO] - Step 1652, Loss: {'train/loss': 0.10759799182415009, 'train/grad_norm_pre_clip': 3.7842321395874023, 'timing/data': 0.013368278974667192, 'timing/model': 1.1507496460108086, 'learning_rate/qwen_vl_interface': 7.836183988802596e-06, 'learning_rate/action_model': 7.836183988802595e-05, 'epoch': 1.65}) +[2026-06-19 13:31:11,669][starVLA.training.train_starvla][INFO] - Step 1653, Loss: {'train/loss': 0.2582281827926636, 'train/grad_norm_pre_clip': 6.154944896697998, 'timing/data': 0.013004072941839695, 'timing/model': 1.1466758920578286, 'learning_rate/qwen_vl_interface': 7.833628997680107e-06, 'learning_rate/action_model': 7.833628997680107e-05, 'epoch': 1.65}) +[2026-06-19 13:31:16,252][starVLA.training.train_starvla][INFO] - Step 1654, Loss: {'train/loss': 0.17932115495204926, 'train/grad_norm_pre_clip': 6.360585689544678, 'timing/data': 0.012073384947143495, 'timing/model': 1.1530493960017338, 'learning_rate/qwen_vl_interface': 7.831072944525329e-06, 'learning_rate/action_model': 7.831072944525328e-05, 'epoch': 1.65}) +[2026-06-19 13:31:20,866][starVLA.training.train_starvla][INFO] - Step 1655, Loss: {'train/loss': 0.23952731490135193, 'train/grad_norm_pre_clip': 5.8800201416015625, 'timing/data': 0.013004490057937801, 'timing/model': 1.316419003996998, 'learning_rate/qwen_vl_interface': 7.828515830388956e-06, 'learning_rate/action_model': 7.828515830388956e-05, 'epoch': 1.65}) +[2026-06-19 13:31:25,284][starVLA.training.train_starvla][INFO] - Step 1656, Loss: {'train/loss': 0.21261578798294067, 'train/grad_norm_pre_clip': 7.27155065536499, 'timing/data': 0.013923704042099416, 'timing/model': 1.1431512469425797, 'learning_rate/qwen_vl_interface': 7.825957656322129e-06, 'learning_rate/action_model': 7.825957656322128e-05, 'epoch': 1.65}) +[2026-06-19 13:31:29,895][starVLA.training.train_starvla][INFO] - Step 1657, Loss: {'train/loss': 0.12584057450294495, 'train/grad_norm_pre_clip': 6.094415664672852, 'timing/data': 0.012673950986936688, 'timing/model': 1.152193556074053, 'learning_rate/qwen_vl_interface': 7.823398423376406e-06, 'learning_rate/action_model': 7.823398423376406e-05, 'epoch': 1.65}) +[2026-06-19 13:31:34,471][starVLA.training.train_starvla][INFO] - Step 1658, Loss: {'train/loss': 0.16721145808696747, 'train/grad_norm_pre_clip': 6.613210201263428, 'timing/data': 0.013699832954443991, 'timing/model': 1.1397990150144324, 'learning_rate/qwen_vl_interface': 7.820838132603801e-06, 'learning_rate/action_model': 7.8208381326038e-05, 'epoch': 1.65}) +[2026-06-19 13:31:39,040][starVLA.training.train_starvla][INFO] - Step 1659, Loss: {'train/loss': 0.1829412281513214, 'train/grad_norm_pre_clip': 12.159311294555664, 'timing/data': 0.014089486910961568, 'timing/model': 1.1431787180481479, 'learning_rate/qwen_vl_interface': 7.818276785056748e-06, 'learning_rate/action_model': 7.818276785056749e-05, 'epoch': 1.65}) +[2026-06-19 13:31:43,620][starVLA.training.train_starvla][INFO] - Step 1660, Loss: {'train/loss': 0.15005433559417725, 'train/grad_norm_pre_clip': 8.359479904174805, 'timing/data': 0.013705785968340933, 'timing/model': 1.1413760290015489, 'learning_rate/qwen_vl_interface': 7.815714381788123e-06, 'learning_rate/action_model': 7.815714381788122e-05, 'epoch': 1.66}) +[2026-06-19 13:31:48,194][starVLA.training.train_starvla][INFO] - Step 1661, Loss: {'train/loss': 0.1502736210823059, 'train/grad_norm_pre_clip': 3.124544858932495, 'timing/data': 0.014181067002937198, 'timing/model': 1.3065098660299554, 'learning_rate/qwen_vl_interface': 7.813150923851227e-06, 'learning_rate/action_model': 7.813150923851227e-05, 'epoch': 1.66}) +[2026-06-19 13:31:52,732][starVLA.training.train_starvla][INFO] - Step 1662, Loss: {'train/loss': 0.22400273382663727, 'train/grad_norm_pre_clip': 4.748349666595459, 'timing/data': 0.012920525041408837, 'timing/model': 1.2627550279721618, 'learning_rate/qwen_vl_interface': 7.810586412299807e-06, 'learning_rate/action_model': 7.810586412299806e-05, 'epoch': 1.66}) +[2026-06-19 13:31:57,147][starVLA.training.train_starvla][INFO] - Step 1663, Loss: {'train/loss': 0.18448445200920105, 'train/grad_norm_pre_clip': 5.487415313720703, 'timing/data': 0.012930177967064083, 'timing/model': 1.1433607679791749, 'learning_rate/qwen_vl_interface': 7.808020848188034e-06, 'learning_rate/action_model': 7.808020848188034e-05, 'epoch': 1.66}) +[2026-06-19 13:32:01,745][starVLA.training.train_starvla][INFO] - Step 1664, Loss: {'train/loss': 0.4008062183856964, 'train/grad_norm_pre_clip': 9.565995216369629, 'timing/data': 0.014351614052429795, 'timing/model': 1.1548964160028845, 'learning_rate/qwen_vl_interface': 7.805454232570517e-06, 'learning_rate/action_model': 7.805454232570517e-05, 'epoch': 1.66}) +[2026-06-19 13:32:06,329][starVLA.training.train_starvla][INFO] - Step 1665, Loss: {'train/loss': 0.10591160506010056, 'train/grad_norm_pre_clip': 3.37753963470459, 'timing/data': 0.01362953009083867, 'timing/model': 1.148754506953992, 'learning_rate/qwen_vl_interface': 7.80288656650229e-06, 'learning_rate/action_model': 7.80288656650229e-05, 'epoch': 1.66}) +[2026-06-19 13:32:10,912][starVLA.training.train_starvla][INFO] - Step 1666, Loss: {'train/loss': 0.13317106664180756, 'train/grad_norm_pre_clip': 3.719219207763672, 'timing/data': 0.01288475701585412, 'timing/model': 1.143638773006387, 'learning_rate/qwen_vl_interface': 7.800317851038829e-06, 'learning_rate/action_model': 7.800317851038828e-05, 'epoch': 1.66}) +[2026-06-19 13:32:15,495][starVLA.training.train_starvla][INFO] - Step 1667, Loss: {'train/loss': 0.3324503004550934, 'train/grad_norm_pre_clip': 10.575822830200195, 'timing/data': 0.013288741931319237, 'timing/model': 1.299997644033283, 'learning_rate/qwen_vl_interface': 7.797748087236033e-06, 'learning_rate/action_model': 7.797748087236033e-05, 'epoch': 1.66}) +[2026-06-19 13:32:20,736][starVLA.training.train_starvla][INFO] - Latency curriculum phase 2 at step 1668: active_latencies=[0, 1, 2], dataset_size=201840 +[2026-06-19 13:32:20,737][starVLA.training.train_starvla][INFO] - Step 1668, Loss: {'train/loss': 0.28973713517189026, 'train/grad_norm_pre_clip': 8.847320556640625, 'latency_curriculum/phase': 2.0, 'latency_curriculum/active_count': 3.0, 'latency_curriculum/max_active_latency': 2.0, 'latency_curriculum/dataset_size': 201840.0, 'timing/data': 0.01237801497336477, 'timing/model': 1.1419348790077493, 'learning_rate/qwen_vl_interface': 7.795177276150234e-06, 'learning_rate/action_model': 7.795177276150234e-05, 'epoch': 1.06}) +[2026-06-19 13:32:37,639][starVLA.training.train_starvla][INFO] - Step 1669, Loss: {'train/loss': 0.3983432352542877, 'train/grad_norm_pre_clip': 19.42316436767578, 'timing/data': 0.018142150016501546, 'timing/model': 1.1342269629240036, 'learning_rate/qwen_vl_interface': 7.792605418838196e-06, 'learning_rate/action_model': 7.792605418838197e-05, 'epoch': 1.06}) +[2026-06-19 13:32:42,099][starVLA.training.train_starvla][INFO] - Step 1670, Loss: {'train/loss': 0.45078927278518677, 'train/grad_norm_pre_clip': 8.871378898620605, 'timing/data': 0.017424356075935066, 'timing/model': 1.1385673789773136, 'learning_rate/qwen_vl_interface': 7.790032516357116e-06, 'learning_rate/action_model': 7.790032516357117e-05, 'epoch': 1.06}) +[2026-06-19 13:32:46,627][starVLA.training.train_starvla][INFO] - Step 1671, Loss: {'train/loss': 0.5565613508224487, 'train/grad_norm_pre_clip': 14.398900032043457, 'timing/data': 0.018945461022667587, 'timing/model': 1.1433745559770614, 'learning_rate/qwen_vl_interface': 7.787458569764613e-06, 'learning_rate/action_model': 7.787458569764612e-05, 'epoch': 1.06}) +[2026-06-19 13:32:51,233][starVLA.training.train_starvla][INFO] - Step 1672, Loss: {'train/loss': 0.22577925026416779, 'train/grad_norm_pre_clip': 4.421792507171631, 'timing/data': 0.016072013997472823, 'timing/model': 1.1382849300280213, 'learning_rate/qwen_vl_interface': 7.784883580118738e-06, 'learning_rate/action_model': 7.784883580118738e-05, 'epoch': 1.06}) +[2026-06-19 13:32:55,834][starVLA.training.train_starvla][INFO] - Step 1673, Loss: {'train/loss': 0.4919285178184509, 'train/grad_norm_pre_clip': 15.590619087219238, 'timing/data': 0.01829624699894339, 'timing/model': 1.3209161859704182, 'learning_rate/qwen_vl_interface': 7.782307548477975e-06, 'learning_rate/action_model': 7.782307548477975e-05, 'epoch': 1.06}) +[2026-06-19 13:33:00,250][starVLA.training.train_starvla][INFO] - Step 1674, Loss: {'train/loss': 0.3997810482978821, 'train/grad_norm_pre_clip': 8.713021278381348, 'timing/data': 0.01922673697117716, 'timing/model': 1.1454359130002558, 'learning_rate/qwen_vl_interface': 7.779730475901233e-06, 'learning_rate/action_model': 7.779730475901232e-05, 'epoch': 1.06}) +[2026-06-19 13:33:04,843][starVLA.training.train_starvla][INFO] - Step 1675, Loss: {'train/loss': 0.18065981566905975, 'train/grad_norm_pre_clip': 4.459288597106934, 'timing/data': 0.0176051378948614, 'timing/model': 1.143446575035341, 'learning_rate/qwen_vl_interface': 7.777152363447849e-06, 'learning_rate/action_model': 7.77715236344785e-05, 'epoch': 1.06}) +[2026-06-19 13:33:09,445][starVLA.training.train_starvla][INFO] - Step 1676, Loss: {'train/loss': 0.4010618031024933, 'train/grad_norm_pre_clip': 8.407610893249512, 'timing/data': 0.018443617969751358, 'timing/model': 1.1533677659463137, 'learning_rate/qwen_vl_interface': 7.774573212177587e-06, 'learning_rate/action_model': 7.774573212177588e-05, 'epoch': 1.06}) +[2026-06-19 13:33:14,000][starVLA.training.train_starvla][INFO] - Step 1677, Loss: {'train/loss': 0.37396132946014404, 'train/grad_norm_pre_clip': 7.083704948425293, 'timing/data': 0.018638130044564605, 'timing/model': 1.1464230329729617, 'learning_rate/qwen_vl_interface': 7.77199302315064e-06, 'learning_rate/action_model': 7.77199302315064e-05, 'epoch': 1.06}) +[2026-06-19 13:33:18,616][starVLA.training.train_starvla][INFO] - Step 1678, Loss: {'train/loss': 0.29657530784606934, 'train/grad_norm_pre_clip': 4.011149883270264, 'timing/data': 0.017910705064423382, 'timing/model': 1.1468576269689947, 'learning_rate/qwen_vl_interface': 7.769411797427622e-06, 'learning_rate/action_model': 7.769411797427623e-05, 'epoch': 1.06}) +[2026-06-19 13:33:23,243][starVLA.training.train_starvla][INFO] - Step 1679, Loss: {'train/loss': 0.3363901376724243, 'train/grad_norm_pre_clip': 5.917457580566406, 'timing/data': 0.018703132052905858, 'timing/model': 1.3250753900501877, 'learning_rate/qwen_vl_interface': 7.766829536069584e-06, 'learning_rate/action_model': 7.766829536069584e-05, 'epoch': 1.06}) +[2026-06-19 13:33:27,670][starVLA.training.train_starvla][INFO] - Step 1680, Loss: {'train/loss': 0.3414912819862366, 'train/grad_norm_pre_clip': 4.323149681091309, 'timing/data': 0.018269623978994787, 'timing/model': 1.143445028923452, 'learning_rate/qwen_vl_interface': 7.764246240137991e-06, 'learning_rate/action_model': 7.764246240137991e-05, 'epoch': 1.07}) +[2026-06-19 13:33:32,246][starVLA.training.train_starvla][INFO] - Step 1681, Loss: {'train/loss': 0.2345002144575119, 'train/grad_norm_pre_clip': 4.197629928588867, 'timing/data': 0.019332772004418075, 'timing/model': 1.1416209159651771, 'learning_rate/qwen_vl_interface': 7.761661910694741e-06, 'learning_rate/action_model': 7.761661910694741e-05, 'epoch': 1.07}) +[2026-06-19 13:33:36,759][starVLA.training.train_starvla][INFO] - Step 1682, Loss: {'train/loss': 0.25629252195358276, 'train/grad_norm_pre_clip': 7.329185962677002, 'timing/data': 0.016763303079642355, 'timing/model': 1.1422812469536439, 'learning_rate/qwen_vl_interface': 7.759076548802153e-06, 'learning_rate/action_model': 7.759076548802152e-05, 'epoch': 1.07}) +[2026-06-19 13:33:41,344][starVLA.training.train_starvla][INFO] - Step 1683, Loss: {'train/loss': 0.2572467029094696, 'train/grad_norm_pre_clip': 3.3072476387023926, 'timing/data': 0.018446340109221637, 'timing/model': 1.1560867320513353, 'learning_rate/qwen_vl_interface': 7.75649015552297e-06, 'learning_rate/action_model': 7.75649015552297e-05, 'epoch': 1.07}) +[2026-06-19 13:33:45,940][starVLA.training.train_starvla][INFO] - Step 1684, Loss: {'train/loss': 0.20512251555919647, 'train/grad_norm_pre_clip': 3.4079229831695557, 'timing/data': 0.01769033691380173, 'timing/model': 1.1460081599652767, 'learning_rate/qwen_vl_interface': 7.753902731920364e-06, 'learning_rate/action_model': 7.753902731920364e-05, 'epoch': 1.07}) +[2026-06-19 13:33:50,522][starVLA.training.train_starvla][INFO] - Step 1685, Loss: {'train/loss': 0.5711147785186768, 'train/grad_norm_pre_clip': 10.196390151977539, 'timing/data': 0.018455381970852613, 'timing/model': 1.3136368580162525, 'learning_rate/qwen_vl_interface': 7.751314279057925e-06, 'learning_rate/action_model': 7.751314279057923e-05, 'epoch': 1.07}) +[2026-06-19 13:33:54,959][starVLA.training.train_starvla][INFO] - Step 1686, Loss: {'train/loss': 0.3472020924091339, 'train/grad_norm_pre_clip': 6.78325080871582, 'timing/data': 0.019390810979530215, 'timing/model': 1.150858078035526, 'learning_rate/qwen_vl_interface': 7.74872479799967e-06, 'learning_rate/action_model': 7.74872479799967e-05, 'epoch': 1.07}) +[2026-06-19 13:33:59,546][starVLA.training.train_starvla][INFO] - Step 1687, Loss: {'train/loss': 0.5755964517593384, 'train/grad_norm_pre_clip': 23.173593521118164, 'timing/data': 0.018943280912935734, 'timing/model': 1.1468581019435078, 'learning_rate/qwen_vl_interface': 7.746134289810034e-06, 'learning_rate/action_model': 7.746134289810033e-05, 'epoch': 1.07}) +[2026-06-19 13:34:04,170][starVLA.training.train_starvla][INFO] - Step 1688, Loss: {'train/loss': 0.24382953345775604, 'train/grad_norm_pre_clip': 13.495881080627441, 'timing/data': 0.018222786020487547, 'timing/model': 1.164155481965281, 'learning_rate/qwen_vl_interface': 7.74354275555388e-06, 'learning_rate/action_model': 7.74354275555388e-05, 'epoch': 1.07}) +[2026-06-19 13:34:08,781][starVLA.training.train_starvla][INFO] - Step 1689, Loss: {'train/loss': 0.16132299602031708, 'train/grad_norm_pre_clip': 4.454068183898926, 'timing/data': 0.018771874019876122, 'timing/model': 1.1572746729943901, 'learning_rate/qwen_vl_interface': 7.740950196296488e-06, 'learning_rate/action_model': 7.740950196296488e-05, 'epoch': 1.07}) +[2026-06-19 13:34:13,396][starVLA.training.train_starvla][INFO] - Step 1690, Loss: {'train/loss': 0.23824651539325714, 'train/grad_norm_pre_clip': 5.160131454467773, 'timing/data': 0.018187562003731728, 'timing/model': 1.1448729319963604, 'learning_rate/qwen_vl_interface': 7.738356613103562e-06, 'learning_rate/action_model': 7.73835661310356e-05, 'epoch': 1.07}) +[2026-06-19 13:34:17,948][starVLA.training.train_starvla][INFO] - Step 1691, Loss: {'train/loss': 0.22850745916366577, 'train/grad_norm_pre_clip': 4.6638336181640625, 'timing/data': 0.018974001053720713, 'timing/model': 1.1495320500107482, 'learning_rate/qwen_vl_interface': 7.735762007041226e-06, 'learning_rate/action_model': 7.735762007041226e-05, 'epoch': 1.07}) +[2026-06-19 13:34:22,555][starVLA.training.train_starvla][INFO] - Step 1692, Loss: {'train/loss': 0.27664774656295776, 'train/grad_norm_pre_clip': 6.418246269226074, 'timing/data': 0.01938513107597828, 'timing/model': 1.3249046209966764, 'learning_rate/qwen_vl_interface': 7.733166379176023e-06, 'learning_rate/action_model': 7.733166379176023e-05, 'epoch': 1.07}) +[2026-06-19 13:34:26,987][starVLA.training.train_starvla][INFO] - Step 1693, Loss: {'train/loss': 0.0970548465847969, 'train/grad_norm_pre_clip': 2.836515426635742, 'timing/data': 0.01956609799526632, 'timing/model': 1.1482011129846796, 'learning_rate/qwen_vl_interface': 7.730569730574921e-06, 'learning_rate/action_model': 7.73056973057492e-05, 'epoch': 1.07}) +[2026-06-19 13:34:31,613][starVLA.training.train_starvla][INFO] - Step 1694, Loss: {'train/loss': 0.4189276695251465, 'train/grad_norm_pre_clip': 6.711486339569092, 'timing/data': 0.01894370501395315, 'timing/model': 1.1501683189999312, 'learning_rate/qwen_vl_interface': 7.727972062305299e-06, 'learning_rate/action_model': 7.727972062305299e-05, 'epoch': 1.07}) +[2026-06-19 13:34:36,231][starVLA.training.train_starvla][INFO] - Step 1695, Loss: {'train/loss': 0.35816922783851624, 'train/grad_norm_pre_clip': 7.940000057220459, 'timing/data': 0.016538580995984375, 'timing/model': 1.1552198260324076, 'learning_rate/qwen_vl_interface': 7.725373375434965e-06, 'learning_rate/action_model': 7.725373375434965e-05, 'epoch': 1.07}) +[2026-06-19 13:34:40,871][starVLA.training.train_starvla][INFO] - Step 1696, Loss: {'train/loss': 0.2285134494304657, 'train/grad_norm_pre_clip': 4.207759857177734, 'timing/data': 0.01674649305641651, 'timing/model': 1.163512002909556, 'learning_rate/qwen_vl_interface': 7.722773671032141e-06, 'learning_rate/action_model': 7.722773671032142e-05, 'epoch': 1.08}) +[2026-06-19 13:34:45,455][starVLA.training.train_starvla][INFO] - Step 1697, Loss: {'train/loss': 0.22241048514842987, 'train/grad_norm_pre_clip': 4.982507228851318, 'timing/data': 0.01873640005942434, 'timing/model': 1.2964759540045634, 'learning_rate/qwen_vl_interface': 7.720172950165466e-06, 'learning_rate/action_model': 7.720172950165465e-05, 'epoch': 1.08}) +[2026-06-19 13:34:50,037][starVLA.training.train_starvla][INFO] - Step 1698, Loss: {'train/loss': 0.13452011346817017, 'train/grad_norm_pre_clip': 3.735628128051758, 'timing/data': 0.018794919014908373, 'timing/model': 1.3031236979877576, 'learning_rate/qwen_vl_interface': 7.717571213903996e-06, 'learning_rate/action_model': 7.717571213903996e-05, 'epoch': 1.08}) +[2026-06-19 13:34:54,453][starVLA.training.train_starvla][INFO] - Step 1699, Loss: {'train/loss': 0.08012424409389496, 'train/grad_norm_pre_clip': 1.7196775674819946, 'timing/data': 0.017661547055467963, 'timing/model': 1.1409360989928246, 'learning_rate/qwen_vl_interface': 7.714968463317207e-06, 'learning_rate/action_model': 7.714968463317207e-05, 'epoch': 1.08}) +[2026-06-19 13:34:59,034][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1700 +[2026-06-19 13:37:51,982][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1700 +[2026-06-19 13:37:51,983][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 1700 +[2026-06-19 13:40:59,671][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 1700 +[2026-06-19 13:40:59,672][starVLA.training.train_starvla][INFO] - Step 1700, Loss: {'train/loss': 0.39057284593582153, 'train/grad_norm_pre_clip': 8.763602256774902, 'eval/loss': 0.35199591517448425, 'eval/latency_0/loss': 0.10883688181638718, 'eval/latency_1/loss': 0.11584906280040741, 'eval/latency_2/loss': 0.30573156476020813, 'eval/latency_3/loss': 0.4863523542881012, 'eval/latency_4/loss': 0.7440370917320251, 'eval/flappy/loss': 0.35199591517448425, 'eval/flappy/latency_0/loss': 0.10883688181638718, 'eval/flappy/latency_1/loss': 0.11584906280040741, 'eval/flappy/latency_2/loss': 0.30573156476020813, 'eval/flappy/latency_3/loss': 0.4863523542881012, 'eval/flappy/latency_4/loss': 0.7440370917320251, 'eval/flap_precision': 0.8047337278106509, 'eval/flap_recall': 0.49454545454545457, 'eval/flap_f1': 0.6126126126126127, 'eval/flap_rate_model': 0.04694444444444444, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.6126126126126127, 'eval/latency_0/flap_precision': 0.8047337278106509, 'eval/latency_0/flap_recall': 0.49454545454545457, 'eval/latency_0/flap_f1': 0.6126126126126127, 'eval/latency_0/flap_rate_model': 0.04694444444444444, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.6126126126126127, 'timing/data': 0.01863392209634185, 'timing/model': 1.153759554028511, 'learning_rate/qwen_vl_interface': 7.712364699474996e-06, 'learning_rate/action_model': 7.712364699474996e-05, 'epoch': 1.08}) +[2026-06-19 13:40:59,673][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_1700_state +[2026-06-19 13:40:59,676][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-19 13:41:03,044][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_1700_state/model.safetensors +[2026-06-19 13:41:10,197][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_1700_state/optimizer.bin +[2026-06-19 13:41:10,197][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_1700_state/sampler.bin +[2026-06-19 13:41:10,198][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_1700_state/sampler_1.bin +[2026-06-19 13:41:10,198][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_1700_state/random_states_0.pkl +[2026-06-19 13:41:10,224][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1700 +[2026-06-19 13:41:10,224][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-19 13:41:10,275][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-19 13:41:14,777][starVLA.training.train_starvla][INFO] - Step 1701, Loss: {'train/loss': 0.24622191488742828, 'train/grad_norm_pre_clip': 5.204105854034424, 'timing/data': 0.01835650601424277, 'timing/model': 1.1196105469716713, 'learning_rate/qwen_vl_interface': 7.709759923447667e-06, 'learning_rate/action_model': 7.709759923447665e-05, 'epoch': 1.08}) +[2026-06-19 13:41:19,294][starVLA.training.train_starvla][INFO] - Step 1702, Loss: {'train/loss': 0.27738821506500244, 'train/grad_norm_pre_clip': 13.800424575805664, 'timing/data': 0.018398790969513357, 'timing/model': 1.3123501320369542, 'learning_rate/qwen_vl_interface': 7.707154136305945e-06, 'learning_rate/action_model': 7.707154136305945e-05, 'epoch': 1.08}) +[2026-06-19 13:41:23,643][starVLA.training.train_starvla][INFO] - Step 1703, Loss: {'train/loss': 0.2980685830116272, 'train/grad_norm_pre_clip': 14.246880531311035, 'timing/data': 0.016873531974852085, 'timing/model': 1.1214087959378958, 'learning_rate/qwen_vl_interface': 7.704547339120974e-06, 'learning_rate/action_model': 7.704547339120974e-05, 'epoch': 1.08}) +[2026-06-19 13:41:28,172][starVLA.training.train_starvla][INFO] - Step 1704, Loss: {'train/loss': 0.26467642188072205, 'train/grad_norm_pre_clip': 11.21118450164795, 'timing/data': 0.017694011912681162, 'timing/model': 1.134204534930177, 'learning_rate/qwen_vl_interface': 7.70193953296431e-06, 'learning_rate/action_model': 7.70193953296431e-05, 'epoch': 1.08}) +[2026-06-19 13:41:32,689][starVLA.training.train_starvla][INFO] - Step 1705, Loss: {'train/loss': 0.3130721151828766, 'train/grad_norm_pre_clip': 11.578940391540527, 'timing/data': 0.017096419003792107, 'timing/model': 1.1274500159779564, 'learning_rate/qwen_vl_interface': 7.69933071890792e-06, 'learning_rate/action_model': 7.699330718907919e-05, 'epoch': 1.08}) +[2026-06-19 13:41:37,197][starVLA.training.train_starvla][INFO] - Step 1706, Loss: {'train/loss': 0.23314346373081207, 'train/grad_norm_pre_clip': 9.797483444213867, 'timing/data': 0.017414530040696263, 'timing/model': 1.1388527760282159, 'learning_rate/qwen_vl_interface': 7.696720898024188e-06, 'learning_rate/action_model': 7.69672089802419e-05, 'epoch': 1.08}) +[2026-06-19 13:41:41,779][starVLA.training.train_starvla][INFO] - Step 1707, Loss: {'train/loss': 0.2670714557170868, 'train/grad_norm_pre_clip': 10.690570831298828, 'timing/data': 0.019538439926691353, 'timing/model': 1.1478998119710013, 'learning_rate/qwen_vl_interface': 7.69411007138592e-06, 'learning_rate/action_model': 7.694110071385919e-05, 'epoch': 1.08}) +[2026-06-19 13:41:46,357][starVLA.training.train_starvla][INFO] - Step 1708, Loss: {'train/loss': 0.31019312143325806, 'train/grad_norm_pre_clip': 13.349016189575195, 'timing/data': 0.017167732934467494, 'timing/model': 1.3179745440138504, 'learning_rate/qwen_vl_interface': 7.691498240066322e-06, 'learning_rate/action_model': 7.691498240066322e-05, 'epoch': 1.08}) +[2026-06-19 13:41:50,748][starVLA.training.train_starvla][INFO] - Step 1709, Loss: {'train/loss': 0.312778115272522, 'train/grad_norm_pre_clip': 7.068692684173584, 'timing/data': 0.017038207966834307, 'timing/model': 1.1428626229753718, 'learning_rate/qwen_vl_interface': 7.688885405139022e-06, 'learning_rate/action_model': 7.688885405139021e-05, 'epoch': 1.08}) +[2026-06-19 13:41:55,338][starVLA.training.train_starvla][INFO] - Step 1710, Loss: {'train/loss': 0.17826677858829498, 'train/grad_norm_pre_clip': 7.746894359588623, 'timing/data': 0.017381435027346015, 'timing/model': 1.143820398952812, 'learning_rate/qwen_vl_interface': 7.686271567678056e-06, 'learning_rate/action_model': 7.686271567678054e-05, 'epoch': 1.08}) +[2026-06-19 13:41:59,919][starVLA.training.train_starvla][INFO] - Step 1711, Loss: {'train/loss': 0.33771538734436035, 'train/grad_norm_pre_clip': 13.186553955078125, 'timing/data': 0.017611977993510664, 'timing/model': 1.1419889410026371, 'learning_rate/qwen_vl_interface': 7.683656728757873e-06, 'learning_rate/action_model': 7.683656728757872e-05, 'epoch': 1.09}) +[2026-06-19 13:42:04,431][starVLA.training.train_starvla][INFO] - Step 1712, Loss: {'train/loss': 0.3577080965042114, 'train/grad_norm_pre_clip': 33.00865936279297, 'timing/data': 0.017790801008231938, 'timing/model': 1.1517309850314632, 'learning_rate/qwen_vl_interface': 7.681040889453337e-06, 'learning_rate/action_model': 7.681040889453336e-05, 'epoch': 1.09}) +[2026-06-19 13:42:08,941][starVLA.training.train_starvla][INFO] - Step 1713, Loss: {'train/loss': 0.3452993929386139, 'train/grad_norm_pre_clip': 14.95807933807373, 'timing/data': 0.01776952890213579, 'timing/model': 1.1455987730296329, 'learning_rate/qwen_vl_interface': 7.678424050839722e-06, 'learning_rate/action_model': 7.67842405083972e-05, 'epoch': 1.09}) +[2026-06-19 13:42:13,503][starVLA.training.train_starvla][INFO] - Step 1714, Loss: {'train/loss': 0.3457014560699463, 'train/grad_norm_pre_clip': 8.908578872680664, 'timing/data': 0.017439982038922608, 'timing/model': 1.2954116350738332, 'learning_rate/qwen_vl_interface': 7.675806213992706e-06, 'learning_rate/action_model': 7.675806213992707e-05, 'epoch': 1.09}) +[2026-06-19 13:42:17,905][starVLA.training.train_starvla][INFO] - Step 1715, Loss: {'train/loss': 0.2532077729701996, 'train/grad_norm_pre_clip': 3.6120338439941406, 'timing/data': 0.024594235001131892, 'timing/model': 1.141461223945953, 'learning_rate/qwen_vl_interface': 7.673187379988388e-06, 'learning_rate/action_model': 7.673187379988388e-05, 'epoch': 1.09}) +[2026-06-19 13:42:22,479][starVLA.training.train_starvla][INFO] - Step 1716, Loss: {'train/loss': 0.271070659160614, 'train/grad_norm_pre_clip': 4.475781440734863, 'timing/data': 0.01832002494484186, 'timing/model': 1.1403290269663557, 'learning_rate/qwen_vl_interface': 7.67056754990327e-06, 'learning_rate/action_model': 7.670567549903269e-05, 'epoch': 1.09}) +[2026-06-19 13:42:27,050][starVLA.training.train_starvla][INFO] - Step 1717, Loss: {'train/loss': 0.3360609710216522, 'train/grad_norm_pre_clip': 7.349595546722412, 'timing/data': 0.019487309968098998, 'timing/model': 1.1374576970702037, 'learning_rate/qwen_vl_interface': 7.667946724814264e-06, 'learning_rate/action_model': 7.667946724814265e-05, 'epoch': 1.09}) +[2026-06-19 13:42:31,596][starVLA.training.train_starvla][INFO] - Step 1718, Loss: {'train/loss': 0.16695505380630493, 'train/grad_norm_pre_clip': 2.9848554134368896, 'timing/data': 0.019432540982961655, 'timing/model': 1.1379338910337538, 'learning_rate/qwen_vl_interface': 7.665324905798694e-06, 'learning_rate/action_model': 7.665324905798695e-05, 'epoch': 1.09}) +[2026-06-19 13:42:36,164][starVLA.training.train_starvla][INFO] - Step 1719, Loss: {'train/loss': 0.36397919058799744, 'train/grad_norm_pre_clip': 11.439579963684082, 'timing/data': 0.02048845705576241, 'timing/model': 1.1472769990796223, 'learning_rate/qwen_vl_interface': 7.66270209393429e-06, 'learning_rate/action_model': 7.66270209393429e-05, 'epoch': 1.09}) +[2026-06-19 13:42:40,726][starVLA.training.train_starvla][INFO] - Step 1720, Loss: {'train/loss': 0.23420175909996033, 'train/grad_norm_pre_clip': 9.150213241577148, 'timing/data': 0.017645985004492104, 'timing/model': 1.3068040709476918, 'learning_rate/qwen_vl_interface': 7.66007829029919e-06, 'learning_rate/action_model': 7.660078290299191e-05, 'epoch': 1.09}) +[2026-06-19 13:42:45,147][starVLA.training.train_starvla][INFO] - Step 1721, Loss: {'train/loss': 0.2786343991756439, 'train/grad_norm_pre_clip': 9.288152694702148, 'timing/data': 0.018243415979668498, 'timing/model': 1.1464964059414342, 'learning_rate/qwen_vl_interface': 7.657453495971942e-06, 'learning_rate/action_model': 7.657453495971943e-05, 'epoch': 1.09}) +[2026-06-19 13:42:49,715][starVLA.training.train_starvla][INFO] - Step 1722, Loss: {'train/loss': 0.29736074805259705, 'train/grad_norm_pre_clip': 13.378211975097656, 'timing/data': 0.017764428979717195, 'timing/model': 1.148780804942362, 'learning_rate/qwen_vl_interface': 7.6548277120315e-06, 'learning_rate/action_model': 7.6548277120315e-05, 'epoch': 1.09}) +[2026-06-19 13:42:54,235][starVLA.training.train_starvla][INFO] - Step 1723, Loss: {'train/loss': 0.25402456521987915, 'train/grad_norm_pre_clip': 12.584756851196289, 'timing/data': 0.018039067974314094, 'timing/model': 1.1445608019130304, 'learning_rate/qwen_vl_interface': 7.652200939557225e-06, 'learning_rate/action_model': 7.652200939557225e-05, 'epoch': 1.09}) +[2026-06-19 13:42:58,814][starVLA.training.train_starvla][INFO] - Step 1724, Loss: {'train/loss': 0.20922012627124786, 'train/grad_norm_pre_clip': 5.219083786010742, 'timing/data': 0.017349820001982152, 'timing/model': 1.143143426044844, 'learning_rate/qwen_vl_interface': 7.649573179628881e-06, 'learning_rate/action_model': 7.649573179628881e-05, 'epoch': 1.09}) +[2026-06-19 13:43:03,408][starVLA.training.train_starvla][INFO] - Step 1725, Loss: {'train/loss': 0.17457890510559082, 'train/grad_norm_pre_clip': 2.7499754428863525, 'timing/data': 0.01807745103724301, 'timing/model': 1.143214161042124, 'learning_rate/qwen_vl_interface': 7.646944433326645e-06, 'learning_rate/action_model': 7.646944433326643e-05, 'epoch': 1.09}) +[2026-06-19 13:43:08,011][starVLA.training.train_starvla][INFO] - Step 1726, Loss: {'train/loss': 0.28022003173828125, 'train/grad_norm_pre_clip': 7.6527252197265625, 'timing/data': 0.01816456695087254, 'timing/model': 1.3255290989764035, 'learning_rate/qwen_vl_interface': 7.64431470173109e-06, 'learning_rate/action_model': 7.64431470173109e-05, 'epoch': 1.09}) +[2026-06-19 13:43:12,431][starVLA.training.train_starvla][INFO] - Step 1727, Loss: {'train/loss': 0.4487505853176117, 'train/grad_norm_pre_clip': 26.80710220336914, 'timing/data': 0.017101914971135557, 'timing/model': 1.1489550969563425, 'learning_rate/qwen_vl_interface': 7.641683985923203e-06, 'learning_rate/action_model': 7.641683985923203e-05, 'epoch': 1.1}) +[2026-06-19 13:43:17,011][starVLA.training.train_starvla][INFO] - Step 1728, Loss: {'train/loss': 0.633299708366394, 'train/grad_norm_pre_clip': 12.3109130859375, 'timing/data': 0.01916325802449137, 'timing/model': 1.1431582629447803, 'learning_rate/qwen_vl_interface': 7.63905228698437e-06, 'learning_rate/action_model': 7.63905228698437e-05, 'epoch': 1.1}) +[2026-06-19 13:43:21,583][starVLA.training.train_starvla][INFO] - Step 1729, Loss: {'train/loss': 0.30771201848983765, 'train/grad_norm_pre_clip': 5.8957624435424805, 'timing/data': 0.01759631501045078, 'timing/model': 1.1453209299361333, 'learning_rate/qwen_vl_interface': 7.636419605996385e-06, 'learning_rate/action_model': 7.636419605996384e-05, 'epoch': 1.1}) +[2026-06-19 13:43:26,084][starVLA.training.train_starvla][INFO] - Step 1730, Loss: {'train/loss': 0.3426060676574707, 'train/grad_norm_pre_clip': 11.386238098144531, 'timing/data': 0.01763349201064557, 'timing/model': 1.1586363170063123, 'learning_rate/qwen_vl_interface': 7.633785944041444e-06, 'learning_rate/action_model': 7.633785944041444e-05, 'epoch': 1.1}) +[2026-06-19 13:43:30,663][starVLA.training.train_starvla][INFO] - Step 1731, Loss: {'train/loss': 0.4396175146102905, 'train/grad_norm_pre_clip': 10.50834846496582, 'timing/data': 0.017864251043647528, 'timing/model': 1.1426675580441952, 'learning_rate/qwen_vl_interface': 7.631151302202145e-06, 'learning_rate/action_model': 7.631151302202144e-05, 'epoch': 1.1}) +[2026-06-19 13:43:35,249][starVLA.training.train_starvla][INFO] - Step 1732, Loss: {'train/loss': 0.23294705152511597, 'train/grad_norm_pre_clip': 5.559077739715576, 'timing/data': 0.01794498600065708, 'timing/model': 1.3234145289752632, 'learning_rate/qwen_vl_interface': 7.628515681561488e-06, 'learning_rate/action_model': 7.628515681561487e-05, 'epoch': 1.1}) +[2026-06-19 13:43:39,684][starVLA.training.train_starvla][INFO] - Step 1733, Loss: {'train/loss': 0.23966869711875916, 'train/grad_norm_pre_clip': 4.755566120147705, 'timing/data': 0.01835110504180193, 'timing/model': 1.160825212020427, 'learning_rate/qwen_vl_interface': 7.62587908320288e-06, 'learning_rate/action_model': 7.62587908320288e-05, 'epoch': 1.1}) +[2026-06-19 13:43:44,288][starVLA.training.train_starvla][INFO] - Step 1734, Loss: {'train/loss': 0.156522735953331, 'train/grad_norm_pre_clip': 4.590554714202881, 'timing/data': 0.017550570075400174, 'timing/model': 1.1503680560272187, 'learning_rate/qwen_vl_interface': 7.6232415082101255e-06, 'learning_rate/action_model': 7.623241508210124e-05, 'epoch': 1.1}) +[2026-06-19 13:43:48,895][starVLA.training.train_starvla][INFO] - Step 1735, Loss: {'train/loss': 0.26883426308631897, 'train/grad_norm_pre_clip': 6.906113624572754, 'timing/data': 0.01689376204740256, 'timing/model': 1.1536468389676884, 'learning_rate/qwen_vl_interface': 7.620602957667432e-06, 'learning_rate/action_model': 7.620602957667432e-05, 'epoch': 1.1}) +[2026-06-19 13:43:53,430][starVLA.training.train_starvla][INFO] - Step 1736, Loss: {'train/loss': 0.32237258553504944, 'train/grad_norm_pre_clip': 7.6614837646484375, 'timing/data': 0.019281979999504983, 'timing/model': 1.1541768700117245, 'learning_rate/qwen_vl_interface': 7.617963432659408e-06, 'learning_rate/action_model': 7.617963432659408e-05, 'epoch': 1.1}) +[2026-06-19 13:43:58,029][starVLA.training.train_starvla][INFO] - Step 1737, Loss: {'train/loss': 0.16546718776226044, 'train/grad_norm_pre_clip': 4.572015285491943, 'timing/data': 0.01851942401845008, 'timing/model': 1.1384922610595822, 'learning_rate/qwen_vl_interface': 7.615322934271063e-06, 'learning_rate/action_model': 7.615322934271064e-05, 'epoch': 1.1}) +[2026-06-19 13:44:02,618][starVLA.training.train_starvla][INFO] - Step 1738, Loss: {'train/loss': 0.4433605670928955, 'train/grad_norm_pre_clip': 7.943424224853516, 'timing/data': 0.018303596996702254, 'timing/model': 1.3161350060254335, 'learning_rate/qwen_vl_interface': 7.6126814635878055e-06, 'learning_rate/action_model': 7.612681463587805e-05, 'epoch': 1.1}) +[2026-06-19 13:44:07,052][starVLA.training.train_starvla][INFO] - Step 1739, Loss: {'train/loss': 0.27412450313568115, 'train/grad_norm_pre_clip': 6.698486804962158, 'timing/data': 0.018095009960234165, 'timing/model': 1.1501166709931567, 'learning_rate/qwen_vl_interface': 7.6100390216954445e-06, 'learning_rate/action_model': 7.610039021695445e-05, 'epoch': 1.1}) +[2026-06-19 13:44:11,648][starVLA.training.train_starvla][INFO] - Step 1740, Loss: {'train/loss': 0.26217877864837646, 'train/grad_norm_pre_clip': 7.466832160949707, 'timing/data': 0.01817095302976668, 'timing/model': 1.1689988650614396, 'learning_rate/qwen_vl_interface': 7.607395609680189e-06, 'learning_rate/action_model': 7.60739560968019e-05, 'epoch': 1.1}) +[2026-06-19 13:44:16,241][starVLA.training.train_starvla][INFO] - Step 1741, Loss: {'train/loss': 0.10646849125623703, 'train/grad_norm_pre_clip': 5.486253261566162, 'timing/data': 0.017702693003229797, 'timing/model': 1.1676302150590345, 'learning_rate/qwen_vl_interface': 7.604751228628644e-06, 'learning_rate/action_model': 7.604751228628644e-05, 'epoch': 1.1}) +[2026-06-19 13:44:20,848][starVLA.training.train_starvla][INFO] - Step 1742, Loss: {'train/loss': 0.2520168125629425, 'train/grad_norm_pre_clip': 7.5805792808532715, 'timing/data': 0.017991138971410692, 'timing/model': 1.1480895939748734, 'learning_rate/qwen_vl_interface': 7.602105879627814e-06, 'learning_rate/action_model': 7.602105879627814e-05, 'epoch': 1.1}) +[2026-06-19 13:44:25,441][starVLA.training.train_starvla][INFO] - Step 1743, Loss: {'train/loss': 0.3624953627586365, 'train/grad_norm_pre_clip': 48.49532699584961, 'timing/data': 0.017537654959596694, 'timing/model': 1.1476515380199999, 'learning_rate/qwen_vl_interface': 7.599459563765107e-06, 'learning_rate/action_model': 7.599459563765106e-05, 'epoch': 1.11}) +[2026-06-19 13:44:30,044][starVLA.training.train_starvla][INFO] - Step 1744, Loss: {'train/loss': 0.2240174412727356, 'train/grad_norm_pre_clip': 10.125127792358398, 'timing/data': 0.01800269796513021, 'timing/model': 1.313161581987515, 'learning_rate/qwen_vl_interface': 7.596812282128319e-06, 'learning_rate/action_model': 7.596812282128319e-05, 'epoch': 1.11}) +[2026-06-19 13:44:34,579][starVLA.training.train_starvla][INFO] - Step 1745, Loss: {'train/loss': 0.3788648843765259, 'train/grad_norm_pre_clip': 11.292699813842773, 'timing/data': 0.017742156982421875, 'timing/model': 1.2508460970129818, 'learning_rate/qwen_vl_interface': 7.594164035805648e-06, 'learning_rate/action_model': 7.594164035805648e-05, 'epoch': 1.11}) +[2026-06-19 13:44:39,027][starVLA.training.train_starvla][INFO] - Step 1746, Loss: {'train/loss': 0.32038137316703796, 'train/grad_norm_pre_clip': 12.41907787322998, 'timing/data': 0.019946774933487177, 'timing/model': 1.1589259729953483, 'learning_rate/qwen_vl_interface': 7.59151482588569e-06, 'learning_rate/action_model': 7.59151482588569e-05, 'epoch': 1.11}) +[2026-06-19 13:44:43,646][starVLA.training.train_starvla][INFO] - Step 1747, Loss: {'train/loss': 0.3008665442466736, 'train/grad_norm_pre_clip': 11.461601257324219, 'timing/data': 0.01789634209126234, 'timing/model': 1.14555582893081, 'learning_rate/qwen_vl_interface': 7.588864653457433e-06, 'learning_rate/action_model': 7.588864653457432e-05, 'epoch': 1.11}) +[2026-06-19 13:44:48,268][starVLA.training.train_starvla][INFO] - Step 1748, Loss: {'train/loss': 0.22195002436637878, 'train/grad_norm_pre_clip': 6.839346885681152, 'timing/data': 0.01818163599818945, 'timing/model': 1.1552980239503086, 'learning_rate/qwen_vl_interface': 7.5862135196102634e-06, 'learning_rate/action_model': 7.586213519610263e-05, 'epoch': 1.11}) +[2026-06-19 13:44:52,882][starVLA.training.train_starvla][INFO] - Step 1749, Loss: {'train/loss': 0.32429999113082886, 'train/grad_norm_pre_clip': 12.485204696655273, 'timing/data': 0.017289039911702275, 'timing/model': 1.1554861719487235, 'learning_rate/qwen_vl_interface': 7.583561425433962e-06, 'learning_rate/action_model': 7.583561425433962e-05, 'epoch': 1.11}) +[2026-06-19 13:44:57,460][starVLA.training.train_starvla][INFO] - Step 1750, Loss: {'train/loss': 0.21075189113616943, 'train/grad_norm_pre_clip': 9.44178295135498, 'timing/data': 0.018002563039772213, 'timing/model': 1.3060087610501796, 'learning_rate/qwen_vl_interface': 7.580908372018706e-06, 'learning_rate/action_model': 7.580908372018705e-05, 'epoch': 1.11}) +[2026-06-19 13:45:01,997][starVLA.training.train_starvla][INFO] - Step 1751, Loss: {'train/loss': 0.2612987160682678, 'train/grad_norm_pre_clip': 7.317986011505127, 'timing/data': 0.018501262995414436, 'timing/model': 1.2620519679039717, 'learning_rate/qwen_vl_interface': 7.5782543604550655e-06, 'learning_rate/action_model': 7.578254360455064e-05, 'epoch': 1.11}) +[2026-06-19 13:45:06,427][starVLA.training.train_starvla][INFO] - Step 1752, Loss: {'train/loss': 0.2892703711986542, 'train/grad_norm_pre_clip': 26.047351837158203, 'timing/data': 0.01906743401195854, 'timing/model': 1.145897440961562, 'learning_rate/qwen_vl_interface': 7.575599391834e-06, 'learning_rate/action_model': 7.575599391834e-05, 'epoch': 1.11}) +[2026-06-19 13:45:11,016][starVLA.training.train_starvla][INFO] - Step 1753, Loss: {'train/loss': 0.34530919790267944, 'train/grad_norm_pre_clip': 10.571687698364258, 'timing/data': 0.01874356297776103, 'timing/model': 1.1483025159686804, 'learning_rate/qwen_vl_interface': 7.572943467246871e-06, 'learning_rate/action_model': 7.57294346724687e-05, 'epoch': 1.11}) +[2026-06-19 13:45:15,623][starVLA.training.train_starvla][INFO] - Step 1754, Loss: {'train/loss': 0.37507179379463196, 'train/grad_norm_pre_clip': 7.988367557525635, 'timing/data': 0.01881315791979432, 'timing/model': 1.153100298019126, 'learning_rate/qwen_vl_interface': 7.570286587785428e-06, 'learning_rate/action_model': 7.570286587785428e-05, 'epoch': 1.11}) +[2026-06-19 13:45:20,231][starVLA.training.train_starvla][INFO] - Step 1755, Loss: {'train/loss': 0.2709009051322937, 'train/grad_norm_pre_clip': 7.1138691902160645, 'timing/data': 0.01831724599469453, 'timing/model': 1.1548559700604528, 'learning_rate/qwen_vl_interface': 7.567628754541815e-06, 'learning_rate/action_model': 7.567628754541814e-05, 'epoch': 1.11}) +[2026-06-19 13:45:24,833][starVLA.training.train_starvla][INFO] - Step 1756, Loss: {'train/loss': 0.3735775351524353, 'train/grad_norm_pre_clip': 9.434286117553711, 'timing/data': 0.019528756965883076, 'timing/model': 1.3179249099921435, 'learning_rate/qwen_vl_interface': 7.564969968608565e-06, 'learning_rate/action_model': 7.564969968608565e-05, 'epoch': 1.11}) +[2026-06-19 13:45:29,299][starVLA.training.train_starvla][INFO] - Step 1757, Loss: {'train/loss': 0.15736646950244904, 'train/grad_norm_pre_clip': 12.077044486999512, 'timing/data': 0.020675125997513533, 'timing/model': 1.1638221250614151, 'learning_rate/qwen_vl_interface': 7.562310231078605e-06, 'learning_rate/action_model': 7.562310231078604e-05, 'epoch': 1.11}) +[2026-06-19 13:45:33,906][starVLA.training.train_starvla][INFO] - Step 1758, Loss: {'train/loss': 0.2904722988605499, 'train/grad_norm_pre_clip': 8.02686595916748, 'timing/data': 0.01762467296794057, 'timing/model': 1.1511406489880756, 'learning_rate/qwen_vl_interface': 7.559649543045252e-06, 'learning_rate/action_model': 7.559649543045252e-05, 'epoch': 1.11}) +[2026-06-19 13:45:38,440][starVLA.training.train_starvla][INFO] - Step 1759, Loss: {'train/loss': 0.36517053842544556, 'train/grad_norm_pre_clip': 12.726165771484375, 'timing/data': 0.01782614807598293, 'timing/model': 1.1511746030300856, 'learning_rate/qwen_vl_interface': 7.556987905602215e-06, 'learning_rate/action_model': 7.556987905602215e-05, 'epoch': 1.12}) +[2026-06-19 13:45:43,023][starVLA.training.train_starvla][INFO] - Step 1760, Loss: {'train/loss': 0.17714321613311768, 'train/grad_norm_pre_clip': 4.060145378112793, 'timing/data': 0.018853596062399447, 'timing/model': 1.148782067000866, 'learning_rate/qwen_vl_interface': 7.554325319843593e-06, 'learning_rate/action_model': 7.554325319843592e-05, 'epoch': 1.12}) +[2026-06-19 13:45:47,639][starVLA.training.train_starvla][INFO] - Step 1761, Loss: {'train/loss': 0.223724827170372, 'train/grad_norm_pre_clip': 7.378306865692139, 'timing/data': 0.018826341023668647, 'timing/model': 1.1417870339937508, 'learning_rate/qwen_vl_interface': 7.551661786863874e-06, 'learning_rate/action_model': 7.551661786863874e-05, 'epoch': 1.12}) +[2026-06-19 13:45:52,227][starVLA.training.train_starvla][INFO] - Step 1762, Loss: {'train/loss': 0.25526511669158936, 'train/grad_norm_pre_clip': 6.2052483558654785, 'timing/data': 0.018853156943805516, 'timing/model': 1.311225202982314, 'learning_rate/qwen_vl_interface': 7.548997307757936e-06, 'learning_rate/action_model': 7.548997307757935e-05, 'epoch': 1.12}) +[2026-06-19 13:45:56,648][starVLA.training.train_starvla][INFO] - Step 1763, Loss: {'train/loss': 0.36752432584762573, 'train/grad_norm_pre_clip': 7.423847675323486, 'timing/data': 0.018069609999656677, 'timing/model': 1.145244661020115, 'learning_rate/qwen_vl_interface': 7.546331883621044e-06, 'learning_rate/action_model': 7.546331883621044e-05, 'epoch': 1.12}) +[2026-06-19 13:46:01,248][starVLA.training.train_starvla][INFO] - Step 1764, Loss: {'train/loss': 0.22110635042190552, 'train/grad_norm_pre_clip': 5.628132343292236, 'timing/data': 0.018510331981815398, 'timing/model': 1.1454331380082294, 'learning_rate/qwen_vl_interface': 7.543665515548857e-06, 'learning_rate/action_model': 7.543665515548856e-05, 'epoch': 1.12}) +[2026-06-19 13:46:05,788][starVLA.training.train_starvla][INFO] - Step 1765, Loss: {'train/loss': 0.14218884706497192, 'train/grad_norm_pre_clip': 4.719046115875244, 'timing/data': 0.01628574903588742, 'timing/model': 1.1460379679920152, 'learning_rate/qwen_vl_interface': 7.540998204637413e-06, 'learning_rate/action_model': 7.540998204637413e-05, 'epoch': 1.12}) +[2026-06-19 13:46:10,386][starVLA.training.train_starvla][INFO] - Step 1766, Loss: {'train/loss': 0.2559484839439392, 'train/grad_norm_pre_clip': 7.0850677490234375, 'timing/data': 0.017116692033596337, 'timing/model': 1.1446104559581727, 'learning_rate/qwen_vl_interface': 7.538329951983148e-06, 'learning_rate/action_model': 7.538329951983147e-05, 'epoch': 1.12}) +[2026-06-19 13:46:14,975][starVLA.training.train_starvla][INFO] - Step 1767, Loss: {'train/loss': 0.22125588357448578, 'train/grad_norm_pre_clip': 5.747861862182617, 'timing/data': 0.018829009030014277, 'timing/model': 1.1450185789726675, 'learning_rate/qwen_vl_interface': 7.535660758682877e-06, 'learning_rate/action_model': 7.535660758682876e-05, 'epoch': 1.12}) +[2026-06-19 13:46:19,588][starVLA.training.train_starvla][INFO] - Step 1768, Loss: {'train/loss': 0.2564687132835388, 'train/grad_norm_pre_clip': 3.928969383239746, 'timing/data': 0.02003132994286716, 'timing/model': 1.3304040021030232, 'learning_rate/qwen_vl_interface': 7.532990625833806e-06, 'learning_rate/action_model': 7.532990625833805e-05, 'epoch': 1.12}) +[2026-06-19 13:46:24,022][starVLA.training.train_starvla][INFO] - Step 1769, Loss: {'train/loss': 0.2993798553943634, 'train/grad_norm_pre_clip': 7.45943021774292, 'timing/data': 0.017557590967044234, 'timing/model': 1.1445312179857865, 'learning_rate/qwen_vl_interface': 7.5303195545335225e-06, 'learning_rate/action_model': 7.530319554533522e-05, 'epoch': 1.12}) +[2026-06-19 13:46:28,601][starVLA.training.train_starvla][INFO] - Step 1770, Loss: {'train/loss': 0.3306574523448944, 'train/grad_norm_pre_clip': 14.073012351989746, 'timing/data': 0.0186409029411152, 'timing/model': 1.1406198559561744, 'learning_rate/qwen_vl_interface': 7.527647545880007e-06, 'learning_rate/action_model': 7.527647545880007e-05, 'epoch': 1.12}) +[2026-06-19 13:46:33,201][starVLA.training.train_starvla][INFO] - Step 1771, Loss: {'train/loss': 0.21394366025924683, 'train/grad_norm_pre_clip': 11.433955192565918, 'timing/data': 0.019092913018539548, 'timing/model': 1.1602222810033709, 'learning_rate/qwen_vl_interface': 7.52497460097162e-06, 'learning_rate/action_model': 7.52497460097162e-05, 'epoch': 1.12}) +[2026-06-19 13:46:37,711][starVLA.training.train_starvla][INFO] - Step 1772, Loss: {'train/loss': 0.3798692226409912, 'train/grad_norm_pre_clip': 8.202282905578613, 'timing/data': 0.017273203004151583, 'timing/model': 1.1413734870729968, 'learning_rate/qwen_vl_interface': 7.522300720907105e-06, 'learning_rate/action_model': 7.522300720907105e-05, 'epoch': 1.12}) +[2026-06-19 13:46:42,279][starVLA.training.train_starvla][INFO] - Step 1773, Loss: {'train/loss': 0.19361020624637604, 'train/grad_norm_pre_clip': 4.546109199523926, 'timing/data': 0.0173123839776963, 'timing/model': 1.146270596073009, 'learning_rate/qwen_vl_interface': 7.519625906785601e-06, 'learning_rate/action_model': 7.5196259067856e-05, 'epoch': 1.12}) +[2026-06-19 13:46:46,883][starVLA.training.train_starvla][INFO] - Step 1774, Loss: {'train/loss': 0.43238529562950134, 'train/grad_norm_pre_clip': 6.771297454833984, 'timing/data': 0.0188726190244779, 'timing/model': 1.3246483380207792, 'learning_rate/qwen_vl_interface': 7.516950159706614e-06, 'learning_rate/action_model': 7.516950159706614e-05, 'epoch': 1.13}) +[2026-06-19 13:46:51,305][starVLA.training.train_starvla][INFO] - Step 1775, Loss: {'train/loss': 0.36505386233329773, 'train/grad_norm_pre_clip': 8.520393371582031, 'timing/data': 0.018268084968440235, 'timing/model': 1.1493798410519958, 'learning_rate/qwen_vl_interface': 7.514273480770047e-06, 'learning_rate/action_model': 7.514273480770047e-05, 'epoch': 1.13}) +[2026-06-19 13:46:55,897][starVLA.training.train_starvla][INFO] - Step 1776, Loss: {'train/loss': 0.31454232335090637, 'train/grad_norm_pre_clip': 11.170153617858887, 'timing/data': 0.017716816975735128, 'timing/model': 1.1427191219991073, 'learning_rate/qwen_vl_interface': 7.511595871076181e-06, 'learning_rate/action_model': 7.511595871076181e-05, 'epoch': 1.13}) +[2026-06-19 13:47:00,468][starVLA.training.train_starvla][INFO] - Step 1777, Loss: {'train/loss': 0.1902937889099121, 'train/grad_norm_pre_clip': 5.934164047241211, 'timing/data': 0.01742434105835855, 'timing/model': 1.1419005299685523, 'learning_rate/qwen_vl_interface': 7.50891733172568e-06, 'learning_rate/action_model': 7.50891733172568e-05, 'epoch': 1.13}) +[2026-06-19 13:47:04,983][starVLA.training.train_starvla][INFO] - Step 1778, Loss: {'train/loss': 0.1754876971244812, 'train/grad_norm_pre_clip': 3.9577465057373047, 'timing/data': 0.01663936406839639, 'timing/model': 1.1496173139894381, 'learning_rate/qwen_vl_interface': 7.506237863819593e-06, 'learning_rate/action_model': 7.506237863819593e-05, 'epoch': 1.13}) +[2026-06-19 13:47:09,561][starVLA.training.train_starvla][INFO] - Step 1779, Loss: {'train/loss': 0.33655327558517456, 'train/grad_norm_pre_clip': 7.733855247497559, 'timing/data': 0.017549618962220848, 'timing/model': 1.1468017760198563, 'learning_rate/qwen_vl_interface': 7.503557468459345e-06, 'learning_rate/action_model': 7.503557468459344e-05, 'epoch': 1.13}) +[2026-06-19 13:47:14,154][starVLA.training.train_starvla][INFO] - Step 1780, Loss: {'train/loss': 0.35279086232185364, 'train/grad_norm_pre_clip': 9.046540260314941, 'timing/data': 0.01708904793485999, 'timing/model': 1.3123310450464487, 'learning_rate/qwen_vl_interface': 7.5008761467467435e-06, 'learning_rate/action_model': 7.500876146746743e-05, 'epoch': 1.13}) +[2026-06-19 13:47:18,569][starVLA.training.train_starvla][INFO] - Step 1781, Loss: {'train/loss': 0.4802088439464569, 'train/grad_norm_pre_clip': 17.418203353881836, 'timing/data': 0.018151644966565073, 'timing/model': 1.1461057640844956, 'learning_rate/qwen_vl_interface': 7.498193899783983e-06, 'learning_rate/action_model': 7.498193899783983e-05, 'epoch': 1.13}) +[2026-06-19 13:47:23,157][starVLA.training.train_starvla][INFO] - Step 1782, Loss: {'train/loss': 0.2690291106700897, 'train/grad_norm_pre_clip': 6.541931629180908, 'timing/data': 0.0187660459196195, 'timing/model': 1.1455956059508026, 'learning_rate/qwen_vl_interface': 7.495510728673631e-06, 'learning_rate/action_model': 7.495510728673632e-05, 'epoch': 1.13}) +[2026-06-19 13:47:27,652][starVLA.training.train_starvla][INFO] - Step 1783, Loss: {'train/loss': 0.3323569893836975, 'train/grad_norm_pre_clip': 15.306014060974121, 'timing/data': 0.018213342991657555, 'timing/model': 1.1430928830523044, 'learning_rate/qwen_vl_interface': 7.492826634518642e-06, 'learning_rate/action_model': 7.492826634518641e-05, 'epoch': 1.13}) +[2026-06-19 13:47:32,181][starVLA.training.train_starvla][INFO] - Step 1784, Loss: {'train/loss': 0.32523292303085327, 'train/grad_norm_pre_clip': 9.083311080932617, 'timing/data': 0.0196603259537369, 'timing/model': 1.1441541858948767, 'learning_rate/qwen_vl_interface': 7.4901416184223404e-06, 'learning_rate/action_model': 7.49014161842234e-05, 'epoch': 1.13}) +[2026-06-19 13:47:36,780][starVLA.training.train_starvla][INFO] - Step 1785, Loss: {'train/loss': 0.21934200823307037, 'train/grad_norm_pre_clip': 5.127499103546143, 'timing/data': 0.01747283397708088, 'timing/model': 1.1484407329699025, 'learning_rate/qwen_vl_interface': 7.487455681488437e-06, 'learning_rate/action_model': 7.487455681488436e-05, 'epoch': 1.13}) +[2026-06-19 13:47:41,398][starVLA.training.train_starvla][INFO] - Step 1786, Loss: {'train/loss': 0.32159602642059326, 'train/grad_norm_pre_clip': 6.349126815795898, 'timing/data': 0.018433107994496822, 'timing/model': 1.3303102400386706, 'learning_rate/qwen_vl_interface': 7.484768824821018e-06, 'learning_rate/action_model': 7.484768824821018e-05, 'epoch': 1.13}) +[2026-06-19 13:47:45,820][starVLA.training.train_starvla][INFO] - Step 1787, Loss: {'train/loss': 0.17711856961250305, 'train/grad_norm_pre_clip': 6.893588542938232, 'timing/data': 0.018137351027689874, 'timing/model': 1.14336790505331, 'learning_rate/qwen_vl_interface': 7.482081049524551e-06, 'learning_rate/action_model': 7.482081049524551e-05, 'epoch': 1.13}) +[2026-06-19 13:47:50,410][starVLA.training.train_starvla][INFO] - Step 1788, Loss: {'train/loss': 0.2988280653953552, 'train/grad_norm_pre_clip': 8.012179374694824, 'timing/data': 0.017092221998609602, 'timing/model': 1.1487339549930766, 'learning_rate/qwen_vl_interface': 7.479392356703878e-06, 'learning_rate/action_model': 7.479392356703878e-05, 'epoch': 1.13}) +[2026-06-19 13:47:55,012][starVLA.training.train_starvla][INFO] - Step 1789, Loss: {'train/loss': 0.15011121332645416, 'train/grad_norm_pre_clip': 3.4304583072662354, 'timing/data': 0.017122998018749058, 'timing/model': 1.1493946340633556, 'learning_rate/qwen_vl_interface': 7.476702747464219e-06, 'learning_rate/action_model': 7.476702747464219e-05, 'epoch': 1.13}) +[2026-06-19 13:47:59,674][starVLA.training.train_starvla][INFO] - Step 1790, Loss: {'train/loss': 0.2789885997772217, 'train/grad_norm_pre_clip': 5.3571858406066895, 'timing/data': 0.017595680081285536, 'timing/model': 1.1443948459345847, 'learning_rate/qwen_vl_interface': 7.474012222911169e-06, 'learning_rate/action_model': 7.474012222911169e-05, 'epoch': 1.14}) +[2026-06-19 13:48:04,290][starVLA.training.train_starvla][INFO] - Step 1791, Loss: {'train/loss': 0.3629116415977478, 'train/grad_norm_pre_clip': 8.860480308532715, 'timing/data': 0.019997679977677763, 'timing/model': 1.1514212279580534, 'learning_rate/qwen_vl_interface': 7.471320784150703e-06, 'learning_rate/action_model': 7.471320784150703e-05, 'epoch': 1.14}) +[2026-06-19 13:48:08,882][starVLA.training.train_starvla][INFO] - Step 1792, Loss: {'train/loss': 0.1914946436882019, 'train/grad_norm_pre_clip': 20.243759155273438, 'timing/data': 0.019557697931304574, 'timing/model': 1.3136781589128077, 'learning_rate/qwen_vl_interface': 7.468628432289169e-06, 'learning_rate/action_model': 7.468628432289168e-05, 'epoch': 1.14}) +[2026-06-19 13:48:13,402][starVLA.training.train_starvla][INFO] - Step 1793, Loss: {'train/loss': 0.30799588561058044, 'train/grad_norm_pre_clip': 6.86709451675415, 'timing/data': 0.01771280902903527, 'timing/model': 1.2400048190029338, 'learning_rate/qwen_vl_interface': 7.465935168433292e-06, 'learning_rate/action_model': 7.465935168433292e-05, 'epoch': 1.14}) +[2026-06-19 13:48:17,832][starVLA.training.train_starvla][INFO] - Step 1794, Loss: {'train/loss': 0.38568904995918274, 'train/grad_norm_pre_clip': 8.8065824508667, 'timing/data': 0.017235244973562658, 'timing/model': 1.1464879569830373, 'learning_rate/qwen_vl_interface': 7.463240993690168e-06, 'learning_rate/action_model': 7.463240993690168e-05, 'epoch': 1.14}) +[2026-06-19 13:48:22,426][starVLA.training.train_starvla][INFO] - Step 1795, Loss: {'train/loss': 0.26451560854911804, 'train/grad_norm_pre_clip': 4.780816555023193, 'timing/data': 0.017370751011185348, 'timing/model': 1.1449995770817623, 'learning_rate/qwen_vl_interface': 7.460545909167276e-06, 'learning_rate/action_model': 7.460545909167275e-05, 'epoch': 1.14}) +[2026-06-19 13:48:27,022][starVLA.training.train_starvla][INFO] - Step 1796, Loss: {'train/loss': 0.19345121085643768, 'train/grad_norm_pre_clip': 5.140735149383545, 'timing/data': 0.01683051697909832, 'timing/model': 1.1430899570696056, 'learning_rate/qwen_vl_interface': 7.457849915972459e-06, 'learning_rate/action_model': 7.457849915972458e-05, 'epoch': 1.14}) +[2026-06-19 13:48:31,601][starVLA.training.train_starvla][INFO] - Step 1797, Loss: {'train/loss': 0.18467962741851807, 'train/grad_norm_pre_clip': 4.002868175506592, 'timing/data': 0.01760573999490589, 'timing/model': 1.1428210909944028, 'learning_rate/qwen_vl_interface': 7.455153015213941e-06, 'learning_rate/action_model': 7.45515301521394e-05, 'epoch': 1.14}) +[2026-06-19 13:48:36,195][starVLA.training.train_starvla][INFO] - Step 1798, Loss: {'train/loss': 0.2832949161529541, 'train/grad_norm_pre_clip': 6.7390971183776855, 'timing/data': 0.01798625197261572, 'timing/model': 1.307629836956039, 'learning_rate/qwen_vl_interface': 7.452455208000311e-06, 'learning_rate/action_model': 7.452455208000311e-05, 'epoch': 1.14}) +[2026-06-19 13:48:40,725][starVLA.training.train_starvla][INFO] - Step 1799, Loss: {'train/loss': 0.2981902062892914, 'train/grad_norm_pre_clip': 6.905000686645508, 'timing/data': 0.01786723209079355, 'timing/model': 1.2634701339993626, 'learning_rate/qwen_vl_interface': 7.449756495440541e-06, 'learning_rate/action_model': 7.44975649544054e-05, 'epoch': 1.14}) +[2026-06-19 13:48:45,183][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1800 +[2026-06-19 13:51:40,877][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1800 +[2026-06-19 13:51:40,878][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 1800 +[2026-06-19 13:54:45,401][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 1800 +[2026-06-19 13:54:45,403][starVLA.training.train_starvla][INFO] - Step 1800, Loss: {'train/loss': 0.20422010123729706, 'train/grad_norm_pre_clip': 4.952633380889893, 'eval/loss': 0.3159283399581909, 'eval/latency_0/loss': 0.10658887773752213, 'eval/latency_1/loss': 0.09342003613710403, 'eval/latency_2/loss': 0.2254963368177414, 'eval/latency_3/loss': 0.46347206830978394, 'eval/latency_4/loss': 0.6972909569740295, 'eval/flappy/loss': 0.3159283399581909, 'eval/flappy/latency_0/loss': 0.10658887773752213, 'eval/flappy/latency_1/loss': 0.09342003613710403, 'eval/flappy/latency_2/loss': 0.2254963368177414, 'eval/flappy/latency_3/loss': 0.46347206830978394, 'eval/flappy/latency_4/loss': 0.6972909569740295, 'eval/flap_precision': 0.8117647058823529, 'eval/flap_recall': 0.5018181818181818, 'eval/flap_f1': 0.6202247191011236, 'eval/flap_rate_model': 0.04722222222222222, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.6202247191011236, 'eval/latency_0/flap_precision': 0.8117647058823529, 'eval/latency_0/flap_recall': 0.5018181818181818, 'eval/latency_0/flap_f1': 0.6202247191011236, 'eval/latency_0/flap_rate_model': 0.04722222222222222, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.6202247191011236, 'timing/data': 0.017992079025134444, 'timing/model': 1.168919151998125, 'learning_rate/qwen_vl_interface': 7.447056878643969e-06, 'learning_rate/action_model': 7.447056878643968e-05, 'epoch': 1.14}) +[2026-06-19 13:54:45,403][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_1800_state +[2026-06-19 13:54:45,405][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-19 13:54:48,767][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_1800_state/model.safetensors +[2026-06-19 13:54:56,002][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_1800_state/optimizer.bin +[2026-06-19 13:54:56,003][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_1800_state/sampler.bin +[2026-06-19 13:54:56,003][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_1800_state/sampler_1.bin +[2026-06-19 13:54:56,004][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_1800_state/random_states_0.pkl +[2026-06-19 13:54:56,028][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1800 +[2026-06-19 13:54:56,028][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-19 13:54:56,039][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-19 13:55:00,555][starVLA.training.train_starvla][INFO] - Step 1801, Loss: {'train/loss': 0.38244080543518066, 'train/grad_norm_pre_clip': 7.699162483215332, 'timing/data': 0.018261280027218163, 'timing/model': 1.1263832070399076, 'learning_rate/qwen_vl_interface': 7.444356358720303e-06, 'learning_rate/action_model': 7.444356358720302e-05, 'epoch': 1.14}) +[2026-06-19 13:55:05,054][starVLA.training.train_starvla][INFO] - Step 1802, Loss: {'train/loss': 0.41248819231987, 'train/grad_norm_pre_clip': 9.174830436706543, 'timing/data': 0.01817317900713533, 'timing/model': 1.1265813650097698, 'learning_rate/qwen_vl_interface': 7.4416549367796255e-06, 'learning_rate/action_model': 7.441654936779625e-05, 'epoch': 1.14}) +[2026-06-19 13:55:09,580][starVLA.training.train_starvla][INFO] - Step 1803, Loss: {'train/loss': 0.2679952085018158, 'train/grad_norm_pre_clip': 5.692873477935791, 'timing/data': 0.01957214600406587, 'timing/model': 1.3112069260096177, 'learning_rate/qwen_vl_interface': 7.4389526139323895e-06, 'learning_rate/action_model': 7.438952613932389e-05, 'epoch': 1.14}) +[2026-06-19 13:55:13,936][starVLA.training.train_starvla][INFO] - Step 1804, Loss: {'train/loss': 0.2903316915035248, 'train/grad_norm_pre_clip': 6.794255256652832, 'timing/data': 0.01770245097577572, 'timing/model': 1.129179161041975, 'learning_rate/qwen_vl_interface': 7.436249391289419e-06, 'learning_rate/action_model': 7.436249391289418e-05, 'epoch': 1.14}) +[2026-06-19 13:55:18,458][starVLA.training.train_starvla][INFO] - Step 1805, Loss: {'train/loss': 0.32953760027885437, 'train/grad_norm_pre_clip': 12.391585350036621, 'timing/data': 0.017019744962453842, 'timing/model': 1.1299899199511856, 'learning_rate/qwen_vl_interface': 7.433545269961906e-06, 'learning_rate/action_model': 7.433545269961906e-05, 'epoch': 1.14}) +[2026-06-19 13:55:22,981][starVLA.training.train_starvla][INFO] - Step 1806, Loss: {'train/loss': 0.3334171772003174, 'train/grad_norm_pre_clip': 8.159453392028809, 'timing/data': 0.01796503597870469, 'timing/model': 1.140667075989768, 'learning_rate/qwen_vl_interface': 7.430840251061411e-06, 'learning_rate/action_model': 7.430840251061411e-05, 'epoch': 1.15}) +[2026-06-19 13:55:27,533][starVLA.training.train_starvla][INFO] - Step 1807, Loss: {'train/loss': 0.09954936057329178, 'train/grad_norm_pre_clip': 3.3990256786346436, 'timing/data': 0.022484312998130918, 'timing/model': 1.1462283550063148, 'learning_rate/qwen_vl_interface': 7.428134335699866e-06, 'learning_rate/action_model': 7.428134335699865e-05, 'epoch': 1.15}) +[2026-06-19 13:55:32,092][starVLA.training.train_starvla][INFO] - Step 1808, Loss: {'train/loss': 0.24374200403690338, 'train/grad_norm_pre_clip': 7.553380489349365, 'timing/data': 0.018644666066393256, 'timing/model': 1.1303562640678138, 'learning_rate/qwen_vl_interface': 7.425427524989572e-06, 'learning_rate/action_model': 7.425427524989572e-05, 'epoch': 1.15}) +[2026-06-19 13:55:36,629][starVLA.training.train_starvla][INFO] - Step 1809, Loss: {'train/loss': 0.24791282415390015, 'train/grad_norm_pre_clip': 8.263717651367188, 'timing/data': 0.01972715393640101, 'timing/model': 1.2985462340293452, 'learning_rate/qwen_vl_interface': 7.422719820043194e-06, 'learning_rate/action_model': 7.422719820043193e-05, 'epoch': 1.15}) +[2026-06-19 13:55:41,017][starVLA.training.train_starvla][INFO] - Step 1810, Loss: {'train/loss': 0.2523980140686035, 'train/grad_norm_pre_clip': 9.563478469848633, 'timing/data': 0.019565860042348504, 'timing/model': 1.1411738559836522, 'learning_rate/qwen_vl_interface': 7.420011221973771e-06, 'learning_rate/action_model': 7.420011221973771e-05, 'epoch': 1.15}) +[2026-06-19 13:55:45,564][starVLA.training.train_starvla][INFO] - Step 1811, Loss: {'train/loss': 0.23883409798145294, 'train/grad_norm_pre_clip': 11.16561508178711, 'timing/data': 0.0171913739759475, 'timing/model': 1.133431322989054, 'learning_rate/qwen_vl_interface': 7.417301731894702e-06, 'learning_rate/action_model': 7.417301731894701e-05, 'epoch': 1.15}) +[2026-06-19 13:55:50,122][starVLA.training.train_starvla][INFO] - Step 1812, Loss: {'train/loss': 0.40155887603759766, 'train/grad_norm_pre_clip': 15.151307106018066, 'timing/data': 0.017127193044871092, 'timing/model': 1.1482807269785553, 'learning_rate/qwen_vl_interface': 7.414591350919756e-06, 'learning_rate/action_model': 7.414591350919756e-05, 'epoch': 1.15}) +[2026-06-19 13:55:54,612][starVLA.training.train_starvla][INFO] - Step 1813, Loss: {'train/loss': 0.1547694355249405, 'train/grad_norm_pre_clip': 5.0122456550598145, 'timing/data': 0.017684763064607978, 'timing/model': 1.1465082729700953, 'learning_rate/qwen_vl_interface': 7.4118800801630715e-06, 'learning_rate/action_model': 7.411880080163072e-05, 'epoch': 1.15}) +[2026-06-19 13:55:59,134][starVLA.training.train_starvla][INFO] - Step 1814, Loss: {'train/loss': 0.23322957754135132, 'train/grad_norm_pre_clip': 7.313562393188477, 'timing/data': 0.01924628601409495, 'timing/model': 1.145430363016203, 'learning_rate/qwen_vl_interface': 7.409167920739146e-06, 'learning_rate/action_model': 7.409167920739147e-05, 'epoch': 1.15}) +[2026-06-19 13:56:03,715][starVLA.training.train_starvla][INFO] - Step 1815, Loss: {'train/loss': 0.1582958847284317, 'train/grad_norm_pre_clip': 5.280612468719482, 'timing/data': 0.01690857403445989, 'timing/model': 1.329559339908883, 'learning_rate/qwen_vl_interface': 7.406454873762848e-06, 'learning_rate/action_model': 7.406454873762848e-05, 'epoch': 1.15}) +[2026-06-19 13:56:08,146][starVLA.training.train_starvla][INFO] - Step 1816, Loss: {'train/loss': 0.2217378318309784, 'train/grad_norm_pre_clip': 8.660656929016113, 'timing/data': 0.017320335027761757, 'timing/model': 1.146930105984211, 'learning_rate/qwen_vl_interface': 7.403740940349408e-06, 'learning_rate/action_model': 7.403740940349408e-05, 'epoch': 1.15}) +[2026-06-19 13:56:12,727][starVLA.training.train_starvla][INFO] - Step 1817, Loss: {'train/loss': 0.2960010766983032, 'train/grad_norm_pre_clip': 16.670625686645508, 'timing/data': 0.016520272009074688, 'timing/model': 1.1390751749277115, 'learning_rate/qwen_vl_interface': 7.401026121614421e-06, 'learning_rate/action_model': 7.40102612161442e-05, 'epoch': 1.15}) +[2026-06-19 13:56:17,324][starVLA.training.train_starvla][INFO] - Step 1818, Loss: {'train/loss': 0.3093149960041046, 'train/grad_norm_pre_clip': 7.14518404006958, 'timing/data': 0.016887302044779062, 'timing/model': 1.148700051009655, 'learning_rate/qwen_vl_interface': 7.398310418673846e-06, 'learning_rate/action_model': 7.398310418673846e-05, 'epoch': 1.15}) +[2026-06-19 13:56:21,848][starVLA.training.train_starvla][INFO] - Step 1819, Loss: {'train/loss': 0.18982462584972382, 'train/grad_norm_pre_clip': 8.043197631835938, 'timing/data': 0.016580270021222532, 'timing/model': 1.141910769045353, 'learning_rate/qwen_vl_interface': 7.395593832644007e-06, 'learning_rate/action_model': 7.395593832644007e-05, 'epoch': 1.15}) +[2026-06-19 13:56:26,378][starVLA.training.train_starvla][INFO] - Step 1820, Loss: {'train/loss': 0.3865242600440979, 'train/grad_norm_pre_clip': 10.981047630310059, 'timing/data': 0.01715529605280608, 'timing/model': 1.146424315054901, 'learning_rate/qwen_vl_interface': 7.39287636464159e-06, 'learning_rate/action_model': 7.39287636464159e-05, 'epoch': 1.15}) +[2026-06-19 13:56:30,977][starVLA.training.train_starvla][INFO] - Step 1821, Loss: {'train/loss': 0.3307350277900696, 'train/grad_norm_pre_clip': 9.090788841247559, 'timing/data': 0.017346681095659733, 'timing/model': 1.328979026991874, 'learning_rate/qwen_vl_interface': 7.3901580157836425e-06, 'learning_rate/action_model': 7.390158015783643e-05, 'epoch': 1.15}) +[2026-06-19 13:56:35,404][starVLA.training.train_starvla][INFO] - Step 1822, Loss: {'train/loss': 0.2904715836048126, 'train/grad_norm_pre_clip': 8.313651084899902, 'timing/data': 0.016544959973543882, 'timing/model': 1.1408640570007265, 'learning_rate/qwen_vl_interface': 7.387438787187576e-06, 'learning_rate/action_model': 7.387438787187576e-05, 'epoch': 1.16}) +[2026-06-19 13:56:39,970][starVLA.training.train_starvla][INFO] - Step 1823, Loss: {'train/loss': 0.21135035157203674, 'train/grad_norm_pre_clip': 9.32114315032959, 'timing/data': 0.01895988406613469, 'timing/model': 1.1453529460122809, 'learning_rate/qwen_vl_interface': 7.3847186799711625e-06, 'learning_rate/action_model': 7.384718679971162e-05, 'epoch': 1.16}) +[2026-06-19 13:56:44,550][starVLA.training.train_starvla][INFO] - Step 1824, Loss: {'train/loss': 0.36321893334388733, 'train/grad_norm_pre_clip': 12.427496910095215, 'timing/data': 0.016841433010995388, 'timing/model': 1.1479942729929462, 'learning_rate/qwen_vl_interface': 7.381997695252535e-06, 'learning_rate/action_model': 7.381997695252535e-05, 'epoch': 1.16}) +[2026-06-19 13:56:49,119][starVLA.training.train_starvla][INFO] - Step 1825, Loss: {'train/loss': 0.19504165649414062, 'train/grad_norm_pre_clip': 7.372631549835205, 'timing/data': 0.018217839067801833, 'timing/model': 1.1413376509444788, 'learning_rate/qwen_vl_interface': 7.379275834150186e-06, 'learning_rate/action_model': 7.379275834150186e-05, 'epoch': 1.16}) +[2026-06-19 13:56:53,596][starVLA.training.train_starvla][INFO] - Step 1826, Loss: {'train/loss': 0.22732466459274292, 'train/grad_norm_pre_clip': 6.799826145172119, 'timing/data': 0.017437738948501647, 'timing/model': 1.138252363074571, 'learning_rate/qwen_vl_interface': 7.376553097782972e-06, 'learning_rate/action_model': 7.376553097782972e-05, 'epoch': 1.16}) +[2026-06-19 13:56:58,168][starVLA.training.train_starvla][INFO] - Step 1827, Loss: {'train/loss': 0.21513649821281433, 'train/grad_norm_pre_clip': 4.911048412322998, 'timing/data': 0.018411709927022457, 'timing/model': 1.2958600759739056, 'learning_rate/qwen_vl_interface': 7.373829487270109e-06, 'learning_rate/action_model': 7.373829487270109e-05, 'epoch': 1.16}) +[2026-06-19 13:57:02,621][starVLA.training.train_starvla][INFO] - Step 1828, Loss: {'train/loss': 0.15188373625278473, 'train/grad_norm_pre_clip': 4.133571624755859, 'timing/data': 0.018584280041977763, 'timing/model': 1.1711309480015188, 'learning_rate/qwen_vl_interface': 7.371105003731165e-06, 'learning_rate/action_model': 7.371105003731164e-05, 'epoch': 1.16}) +[2026-06-19 13:57:07,227][starVLA.training.train_starvla][INFO] - Step 1829, Loss: {'train/loss': 0.21959370374679565, 'train/grad_norm_pre_clip': 5.562042236328125, 'timing/data': 0.01647718599997461, 'timing/model': 1.1535955379949883, 'learning_rate/qwen_vl_interface': 7.3683796482860755e-06, 'learning_rate/action_model': 7.368379648286075e-05, 'epoch': 1.16}) +[2026-06-19 13:57:11,830][starVLA.training.train_starvla][INFO] - Step 1830, Loss: {'train/loss': 0.3276151418685913, 'train/grad_norm_pre_clip': 6.835042953491211, 'timing/data': 0.01846508600283414, 'timing/model': 1.1481564609566703, 'learning_rate/qwen_vl_interface': 7.365653422055131e-06, 'learning_rate/action_model': 7.365653422055131e-05, 'epoch': 1.16}) +[2026-06-19 13:57:16,382][starVLA.training.train_starvla][INFO] - Step 1831, Loss: {'train/loss': 0.15612581372261047, 'train/grad_norm_pre_clip': 6.544500350952148, 'timing/data': 0.018246263032779098, 'timing/model': 1.1420950320316479, 'learning_rate/qwen_vl_interface': 7.3629263261589816e-06, 'learning_rate/action_model': 7.362926326158981e-05, 'epoch': 1.16}) +[2026-06-19 13:57:20,977][starVLA.training.train_starvla][INFO] - Step 1832, Loss: {'train/loss': 0.18525224924087524, 'train/grad_norm_pre_clip': 6.294754981994629, 'timing/data': 0.017268294002860785, 'timing/model': 1.1534644128987566, 'learning_rate/qwen_vl_interface': 7.36019836171863e-06, 'learning_rate/action_model': 7.360198361718629e-05, 'epoch': 1.16}) +[2026-06-19 13:57:25,595][starVLA.training.train_starvla][INFO] - Step 1833, Loss: {'train/loss': 0.30883175134658813, 'train/grad_norm_pre_clip': 14.459518432617188, 'timing/data': 0.01772412599530071, 'timing/model': 1.3343576459446922, 'learning_rate/qwen_vl_interface': 7.357469529855441e-06, 'learning_rate/action_model': 7.35746952985544e-05, 'epoch': 1.16}) +[2026-06-19 13:57:30,010][starVLA.training.train_starvla][INFO] - Step 1834, Loss: {'train/loss': 0.19539813697338104, 'train/grad_norm_pre_clip': 8.10987663269043, 'timing/data': 0.017437236034311354, 'timing/model': 1.1457763699581847, 'learning_rate/qwen_vl_interface': 7.354739831691134e-06, 'learning_rate/action_model': 7.354739831691135e-05, 'epoch': 1.16}) +[2026-06-19 13:57:34,611][starVLA.training.train_starvla][INFO] - Step 1835, Loss: {'train/loss': 0.13612186908721924, 'train/grad_norm_pre_clip': 5.825811862945557, 'timing/data': 0.016193526913411915, 'timing/model': 1.1576539619127288, 'learning_rate/qwen_vl_interface': 7.352009268347786e-06, 'learning_rate/action_model': 7.352009268347786e-05, 'epoch': 1.16}) +[2026-06-19 13:57:39,118][starVLA.training.train_starvla][INFO] - Step 1836, Loss: {'train/loss': 0.1313033252954483, 'train/grad_norm_pre_clip': 4.45760440826416, 'timing/data': 0.019651504000648856, 'timing/model': 1.1572986820247024, 'learning_rate/qwen_vl_interface': 7.349277840947828e-06, 'learning_rate/action_model': 7.349277840947828e-05, 'epoch': 1.16}) +[2026-06-19 13:57:43,670][starVLA.training.train_starvla][INFO] - Step 1837, Loss: {'train/loss': 0.26133298873901367, 'train/grad_norm_pre_clip': 4.965797424316406, 'timing/data': 0.01760184799786657, 'timing/model': 1.1418140679597855, 'learning_rate/qwen_vl_interface': 7.346545550614046e-06, 'learning_rate/action_model': 7.346545550614046e-05, 'epoch': 1.16}) +[2026-06-19 13:57:48,251][starVLA.training.train_starvla][INFO] - Step 1838, Loss: {'train/loss': 0.235248863697052, 'train/grad_norm_pre_clip': 5.83574914932251, 'timing/data': 0.018244453007355332, 'timing/model': 1.1433233089046553, 'learning_rate/qwen_vl_interface': 7.343812398469582e-06, 'learning_rate/action_model': 7.343812398469581e-05, 'epoch': 1.17}) +[2026-06-19 13:57:52,862][starVLA.training.train_starvla][INFO] - Step 1839, Loss: {'train/loss': 0.2321282923221588, 'train/grad_norm_pre_clip': 6.449141502380371, 'timing/data': 0.020675977924838662, 'timing/model': 1.3357247760286555, 'learning_rate/qwen_vl_interface': 7.34107838563793e-06, 'learning_rate/action_model': 7.341078385637929e-05, 'epoch': 1.17}) +[2026-06-19 13:57:57,291][starVLA.training.train_starvla][INFO] - Step 1840, Loss: {'train/loss': 0.12702630460262299, 'train/grad_norm_pre_clip': 7.595078945159912, 'timing/data': 0.01688470004592091, 'timing/model': 1.1533631919883192, 'learning_rate/qwen_vl_interface': 7.338343513242941e-06, 'learning_rate/action_model': 7.338343513242941e-05, 'epoch': 1.17}) +[2026-06-19 13:58:01,884][starVLA.training.train_starvla][INFO] - Step 1841, Loss: {'train/loss': 0.2959598898887634, 'train/grad_norm_pre_clip': 6.231290817260742, 'timing/data': 0.01802947197575122, 'timing/model': 1.1449130029650405, 'learning_rate/qwen_vl_interface': 7.335607782408818e-06, 'learning_rate/action_model': 7.335607782408818e-05, 'epoch': 1.17}) +[2026-06-19 13:58:06,472][starVLA.training.train_starvla][INFO] - Step 1842, Loss: {'train/loss': 0.2171710580587387, 'train/grad_norm_pre_clip': 7.5477213859558105, 'timing/data': 0.016906967968679965, 'timing/model': 1.1514314070809633, 'learning_rate/qwen_vl_interface': 7.3328711942601145e-06, 'learning_rate/action_model': 7.332871194260115e-05, 'epoch': 1.17}) +[2026-06-19 13:58:11,080][starVLA.training.train_starvla][INFO] - Step 1843, Loss: {'train/loss': 0.2218652069568634, 'train/grad_norm_pre_clip': 7.903654098510742, 'timing/data': 0.016832908033393323, 'timing/model': 1.1682912450050935, 'learning_rate/qwen_vl_interface': 7.33013374992174e-06, 'learning_rate/action_model': 7.33013374992174e-05, 'epoch': 1.17}) +[2026-06-19 13:58:15,663][starVLA.training.train_starvla][INFO] - Step 1844, Loss: {'train/loss': 0.33078330755233765, 'train/grad_norm_pre_clip': 10.162766456604004, 'timing/data': 0.017884701024740934, 'timing/model': 1.1447716059628874, 'learning_rate/qwen_vl_interface': 7.327395450518956e-06, 'learning_rate/action_model': 7.327395450518956e-05, 'epoch': 1.17}) +[2026-06-19 13:58:20,245][starVLA.training.train_starvla][INFO] - Step 1845, Loss: {'train/loss': 0.39641231298446655, 'train/grad_norm_pre_clip': 8.790655136108398, 'timing/data': 0.018717600964009762, 'timing/model': 1.3019594700308517, 'learning_rate/qwen_vl_interface': 7.324656297177372e-06, 'learning_rate/action_model': 7.324656297177371e-05, 'epoch': 1.17}) +[2026-06-19 13:58:24,670][starVLA.training.train_starvla][INFO] - Step 1846, Loss: {'train/loss': 0.19564858078956604, 'train/grad_norm_pre_clip': 6.400513172149658, 'timing/data': 0.018223923980258405, 'timing/model': 1.158098370069638, 'learning_rate/qwen_vl_interface': 7.32191629102295e-06, 'learning_rate/action_model': 7.32191629102295e-05, 'epoch': 1.17}) +[2026-06-19 13:58:29,242][starVLA.training.train_starvla][INFO] - Step 1847, Loss: {'train/loss': 0.6211867928504944, 'train/grad_norm_pre_clip': 14.765344619750977, 'timing/data': 0.017595025012269616, 'timing/model': 1.1458144029602408, 'learning_rate/qwen_vl_interface': 7.319175433182004e-06, 'learning_rate/action_model': 7.319175433182004e-05, 'epoch': 1.17}) +[2026-06-19 13:58:33,731][starVLA.training.train_starvla][INFO] - Step 1848, Loss: {'train/loss': 0.13084958493709564, 'train/grad_norm_pre_clip': 4.894021034240723, 'timing/data': 0.01775046798866242, 'timing/model': 1.1420385419623926, 'learning_rate/qwen_vl_interface': 7.316433724781198e-06, 'learning_rate/action_model': 7.316433724781197e-05, 'epoch': 1.17}) +[2026-06-19 13:58:38,253][starVLA.training.train_starvla][INFO] - Step 1849, Loss: {'train/loss': 0.3703453838825226, 'train/grad_norm_pre_clip': 8.4093656539917, 'timing/data': 0.02015167602803558, 'timing/model': 1.1642953120172024, 'learning_rate/qwen_vl_interface': 7.313691166947543e-06, 'learning_rate/action_model': 7.313691166947543e-05, 'epoch': 1.17}) +[2026-06-19 13:58:42,771][starVLA.training.train_starvla][INFO] - Step 1850, Loss: {'train/loss': 0.3022737205028534, 'train/grad_norm_pre_clip': 9.860127449035645, 'timing/data': 0.018573839915916324, 'timing/model': 1.1511569200083613, 'learning_rate/qwen_vl_interface': 7.310947760808401e-06, 'learning_rate/action_model': 7.3109477608084e-05, 'epoch': 1.17}) +[2026-06-19 13:58:47,368][starVLA.training.train_starvla][INFO] - Step 1851, Loss: {'train/loss': 0.3857620358467102, 'train/grad_norm_pre_clip': 12.17233943939209, 'timing/data': 0.017505284980870783, 'timing/model': 1.3192367641022429, 'learning_rate/qwen_vl_interface': 7.308203507491485e-06, 'learning_rate/action_model': 7.308203507491484e-05, 'epoch': 1.17}) +[2026-06-19 13:58:51,784][starVLA.training.train_starvla][INFO] - Step 1852, Loss: {'train/loss': 0.16474974155426025, 'train/grad_norm_pre_clip': 7.124519348144531, 'timing/data': 0.01833593298215419, 'timing/model': 1.1436194049892947, 'learning_rate/qwen_vl_interface': 7.3054584081248515e-06, 'learning_rate/action_model': 7.305458408124852e-05, 'epoch': 1.17}) +[2026-06-19 13:58:56,341][starVLA.training.train_starvla][INFO] - Step 1853, Loss: {'train/loss': 0.29740265011787415, 'train/grad_norm_pre_clip': 13.792943954467773, 'timing/data': 0.01764559408184141, 'timing/model': 1.1458516409620643, 'learning_rate/qwen_vl_interface': 7.302712463836907e-06, 'learning_rate/action_model': 7.302712463836907e-05, 'epoch': 1.18}) +[2026-06-19 13:59:00,949][starVLA.training.train_starvla][INFO] - Step 1854, Loss: {'train/loss': 0.24931767582893372, 'train/grad_norm_pre_clip': 8.343276977539062, 'timing/data': 0.017441597068682313, 'timing/model': 1.1510289140278473, 'learning_rate/qwen_vl_interface': 7.2999656757564086e-06, 'learning_rate/action_model': 7.299965675756409e-05, 'epoch': 1.18}) +[2026-06-19 13:59:05,517][starVLA.training.train_starvla][INFO] - Step 1855, Loss: {'train/loss': 0.3880421817302704, 'train/grad_norm_pre_clip': 7.987590312957764, 'timing/data': 0.01733752596192062, 'timing/model': 1.1415244180243462, 'learning_rate/qwen_vl_interface': 7.297218045012453e-06, 'learning_rate/action_model': 7.297218045012453e-05, 'epoch': 1.18}) +[2026-06-19 13:59:10,123][starVLA.training.train_starvla][INFO] - Step 1856, Loss: {'train/loss': 0.2125384509563446, 'train/grad_norm_pre_clip': 7.238096237182617, 'timing/data': 0.018077088985592127, 'timing/model': 1.327960347989574, 'learning_rate/qwen_vl_interface': 7.29446957273449e-06, 'learning_rate/action_model': 7.294469572734489e-05, 'epoch': 1.18}) +[2026-06-19 13:59:14,560][starVLA.training.train_starvla][INFO] - Step 1857, Loss: {'train/loss': 0.15900281071662903, 'train/grad_norm_pre_clip': 4.317491054534912, 'timing/data': 0.01839067996479571, 'timing/model': 1.1549370840657502, 'learning_rate/qwen_vl_interface': 7.2917202600523104e-06, 'learning_rate/action_model': 7.29172026005231e-05, 'epoch': 1.18}) +[2026-06-19 13:59:19,169][starVLA.training.train_starvla][INFO] - Step 1858, Loss: {'train/loss': 0.23868724703788757, 'train/grad_norm_pre_clip': 6.8811187744140625, 'timing/data': 0.01745990803465247, 'timing/model': 1.1504878340056166, 'learning_rate/qwen_vl_interface': 7.2889701080960565e-06, 'learning_rate/action_model': 7.288970108096056e-05, 'epoch': 1.18}) +[2026-06-19 13:59:23,763][starVLA.training.train_starvla][INFO] - Step 1859, Loss: {'train/loss': 0.2778630256652832, 'train/grad_norm_pre_clip': 6.00794792175293, 'timing/data': 0.01726029405836016, 'timing/model': 1.1442207979271188, 'learning_rate/qwen_vl_interface': 7.286219117996209e-06, 'learning_rate/action_model': 7.286219117996207e-05, 'epoch': 1.18}) +[2026-06-19 13:59:28,351][starVLA.training.train_starvla][INFO] - Step 1860, Loss: {'train/loss': 0.3247836232185364, 'train/grad_norm_pre_clip': 7.961669921875, 'timing/data': 0.01864088897127658, 'timing/model': 1.1439144070027396, 'learning_rate/qwen_vl_interface': 7.283467290883598e-06, 'learning_rate/action_model': 7.283467290883598e-05, 'epoch': 1.18}) +[2026-06-19 13:59:32,923][starVLA.training.train_starvla][INFO] - Step 1861, Loss: {'train/loss': 0.08451805263757706, 'train/grad_norm_pre_clip': 5.781470775604248, 'timing/data': 0.01953782606869936, 'timing/model': 1.1400451710214838, 'learning_rate/qwen_vl_interface': 7.2807146278893925e-06, 'learning_rate/action_model': 7.280714627889393e-05, 'epoch': 1.18}) +[2026-06-19 13:59:37,501][starVLA.training.train_starvla][INFO] - Step 1862, Loss: {'train/loss': 0.2421087622642517, 'train/grad_norm_pre_clip': 8.277669906616211, 'timing/data': 0.017829293035902083, 'timing/model': 1.301882550935261, 'learning_rate/qwen_vl_interface': 7.277961130145111e-06, 'learning_rate/action_model': 7.277961130145111e-05, 'epoch': 1.18}) +[2026-06-19 13:59:41,998][starVLA.training.train_starvla][INFO] - Step 1863, Loss: {'train/loss': 0.26056236028671265, 'train/grad_norm_pre_clip': 7.24086332321167, 'timing/data': 0.018679511034861207, 'timing/model': 1.228809230029583, 'learning_rate/qwen_vl_interface': 7.2752067987826116e-06, 'learning_rate/action_model': 7.275206798782611e-05, 'epoch': 1.18}) +[2026-06-19 13:59:46,424][starVLA.training.train_starvla][INFO] - Step 1864, Loss: {'train/loss': 0.2916465997695923, 'train/grad_norm_pre_clip': 9.0188570022583, 'timing/data': 0.017546989023685455, 'timing/model': 1.1438780101016164, 'learning_rate/qwen_vl_interface': 7.272451634934096e-06, 'learning_rate/action_model': 7.272451634934096e-05, 'epoch': 1.18}) +[2026-06-19 13:59:50,999][starVLA.training.train_starvla][INFO] - Step 1865, Loss: {'train/loss': 0.1953471153974533, 'train/grad_norm_pre_clip': 5.7427568435668945, 'timing/data': 0.017066914006136358, 'timing/model': 1.1415700030047446, 'learning_rate/qwen_vl_interface': 7.269695639732108e-06, 'learning_rate/action_model': 7.269695639732108e-05, 'epoch': 1.18}) +[2026-06-19 13:59:55,574][starVLA.training.train_starvla][INFO] - Step 1866, Loss: {'train/loss': 0.22839771211147308, 'train/grad_norm_pre_clip': 4.941433429718018, 'timing/data': 0.017196996021084487, 'timing/model': 1.1446799719706178, 'learning_rate/qwen_vl_interface': 7.266938814309533e-06, 'learning_rate/action_model': 7.266938814309533e-05, 'epoch': 1.18}) +[2026-06-19 14:00:00,145][starVLA.training.train_starvla][INFO] - Step 1867, Loss: {'train/loss': 0.2509065270423889, 'train/grad_norm_pre_clip': 6.983392238616943, 'timing/data': 0.018595747998915613, 'timing/model': 1.1417472569737583, 'learning_rate/qwen_vl_interface': 7.264181159799598e-06, 'learning_rate/action_model': 7.264181159799597e-05, 'epoch': 1.18}) +[2026-06-19 14:00:04,737][starVLA.training.train_starvla][INFO] - Step 1868, Loss: {'train/loss': 0.3065808415412903, 'train/grad_norm_pre_clip': 8.032958030700684, 'timing/data': 0.01794304104987532, 'timing/model': 1.3097053980454803, 'learning_rate/qwen_vl_interface': 7.261422677335871e-06, 'learning_rate/action_model': 7.261422677335871e-05, 'epoch': 1.18}) +[2026-06-19 14:00:09,250][starVLA.training.train_starvla][INFO] - Step 1869, Loss: {'train/loss': 0.3163440525531769, 'train/grad_norm_pre_clip': 6.551766872406006, 'timing/data': 0.018405677983537316, 'timing/model': 1.2339118709787726, 'learning_rate/qwen_vl_interface': 7.258663368052259e-06, 'learning_rate/action_model': 7.258663368052259e-05, 'epoch': 1.19}) +[2026-06-19 14:00:13,675][starVLA.training.train_starvla][INFO] - Step 1870, Loss: {'train/loss': 0.18563680350780487, 'train/grad_norm_pre_clip': 4.762993335723877, 'timing/data': 0.016989724012091756, 'timing/model': 1.1545920089120045, 'learning_rate/qwen_vl_interface': 7.2559032330830105e-06, 'learning_rate/action_model': 7.25590323308301e-05, 'epoch': 1.19}) +[2026-06-19 14:00:18,255][starVLA.training.train_starvla][INFO] - Step 1871, Loss: {'train/loss': 0.20430059731006622, 'train/grad_norm_pre_clip': 4.562358856201172, 'timing/data': 0.01802520896308124, 'timing/model': 1.1453007210511714, 'learning_rate/qwen_vl_interface': 7.253142273562712e-06, 'learning_rate/action_model': 7.253142273562712e-05, 'epoch': 1.19}) +[2026-06-19 14:00:22,840][starVLA.training.train_starvla][INFO] - Step 1872, Loss: {'train/loss': 0.13347865641117096, 'train/grad_norm_pre_clip': 4.692911148071289, 'timing/data': 0.017341763945296407, 'timing/model': 1.1453401750186458, 'learning_rate/qwen_vl_interface': 7.25038049062629e-06, 'learning_rate/action_model': 7.250380490626289e-05, 'epoch': 1.19}) +[2026-06-19 14:00:27,448][starVLA.training.train_starvla][INFO] - Step 1873, Loss: {'train/loss': 0.404969722032547, 'train/grad_norm_pre_clip': 11.032373428344727, 'timing/data': 0.017593425000086427, 'timing/model': 1.1470646210946143, 'learning_rate/qwen_vl_interface': 7.24761788540901e-06, 'learning_rate/action_model': 7.24761788540901e-05, 'epoch': 1.19}) +[2026-06-19 14:00:32,053][starVLA.training.train_starvla][INFO] - Step 1874, Loss: {'train/loss': 0.34564208984375, 'train/grad_norm_pre_clip': 5.753968715667725, 'timing/data': 0.01993495097849518, 'timing/model': 1.3266070570098236, 'learning_rate/qwen_vl_interface': 7.244854459046473e-06, 'learning_rate/action_model': 7.244854459046473e-05, 'epoch': 1.19}) +[2026-06-19 14:00:36,481][starVLA.training.train_starvla][INFO] - Step 1875, Loss: {'train/loss': 0.2905603051185608, 'train/grad_norm_pre_clip': 8.323229789733887, 'timing/data': 0.01810655405279249, 'timing/model': 1.1468377819983289, 'learning_rate/qwen_vl_interface': 7.24209021267462e-06, 'learning_rate/action_model': 7.24209021267462e-05, 'epoch': 1.19}) +[2026-06-19 14:00:41,064][starVLA.training.train_starvla][INFO] - Step 1876, Loss: {'train/loss': 0.17030034959316254, 'train/grad_norm_pre_clip': 6.263672828674316, 'timing/data': 0.017885903944261372, 'timing/model': 1.1571725759422407, 'learning_rate/qwen_vl_interface': 7.23932514742973e-06, 'learning_rate/action_model': 7.23932514742973e-05, 'epoch': 1.19}) +[2026-06-19 14:00:45,562][starVLA.training.train_starvla][INFO] - Step 1877, Loss: {'train/loss': 0.15776513516902924, 'train/grad_norm_pre_clip': 7.0219597816467285, 'timing/data': 0.017554515972733498, 'timing/model': 1.1476134909316897, 'learning_rate/qwen_vl_interface': 7.236559264448413e-06, 'learning_rate/action_model': 7.236559264448412e-05, 'epoch': 1.19}) +[2026-06-19 14:00:50,046][starVLA.training.train_starvla][INFO] - Step 1878, Loss: {'train/loss': 0.37458884716033936, 'train/grad_norm_pre_clip': 9.681999206542969, 'timing/data': 0.01861599797848612, 'timing/model': 1.146814840962179, 'learning_rate/qwen_vl_interface': 7.233792564867623e-06, 'learning_rate/action_model': 7.233792564867623e-05, 'epoch': 1.19}) +[2026-06-19 14:00:54,554][starVLA.training.train_starvla][INFO] - Step 1879, Loss: {'train/loss': 0.2840940058231354, 'train/grad_norm_pre_clip': 7.00999641418457, 'timing/data': 0.016652660095132887, 'timing/model': 1.1422343360027298, 'learning_rate/qwen_vl_interface': 7.231025049824644e-06, 'learning_rate/action_model': 7.231025049824643e-05, 'epoch': 1.19}) +[2026-06-19 14:00:59,156][starVLA.training.train_starvla][INFO] - Step 1880, Loss: {'train/loss': 0.24315741658210754, 'train/grad_norm_pre_clip': 4.467516899108887, 'timing/data': 0.017100781900808215, 'timing/model': 1.319738666061312, 'learning_rate/qwen_vl_interface': 7.228256720457097e-06, 'learning_rate/action_model': 7.228256720457096e-05, 'epoch': 1.19}) +[2026-06-19 14:01:03,575][starVLA.training.train_starvla][INFO] - Step 1881, Loss: {'train/loss': 0.2717325687408447, 'train/grad_norm_pre_clip': 15.194707870483398, 'timing/data': 0.018680437933653593, 'timing/model': 1.146765273064375, 'learning_rate/qwen_vl_interface': 7.225487577902937e-06, 'learning_rate/action_model': 7.225487577902936e-05, 'epoch': 1.19}) +[2026-06-19 14:01:08,147][starVLA.training.train_starvla][INFO] - Step 1882, Loss: {'train/loss': 0.23168130218982697, 'train/grad_norm_pre_clip': 7.660093307495117, 'timing/data': 0.016972671961411834, 'timing/model': 1.1448902470292524, 'learning_rate/qwen_vl_interface': 7.222717623300455e-06, 'learning_rate/action_model': 7.222717623300455e-05, 'epoch': 1.19}) +[2026-06-19 14:01:12,732][starVLA.training.train_starvla][INFO] - Step 1883, Loss: {'train/loss': 0.29628294706344604, 'train/grad_norm_pre_clip': 6.994107246398926, 'timing/data': 0.017506608041003346, 'timing/model': 1.143161537940614, 'learning_rate/qwen_vl_interface': 7.219946857788273e-06, 'learning_rate/action_model': 7.219946857788272e-05, 'epoch': 1.19}) +[2026-06-19 14:01:17,309][starVLA.training.train_starvla][INFO] - Step 1884, Loss: {'train/loss': 0.26936453580856323, 'train/grad_norm_pre_clip': 6.952699184417725, 'timing/data': 0.01841462100856006, 'timing/model': 1.141094468999654, 'learning_rate/qwen_vl_interface': 7.21717528250535e-06, 'learning_rate/action_model': 7.21717528250535e-05, 'epoch': 1.19}) +[2026-06-19 14:01:21,898][starVLA.training.train_starvla][INFO] - Step 1885, Loss: {'train/loss': 0.21277539432048798, 'train/grad_norm_pre_clip': 7.574492454528809, 'timing/data': 0.0172197709325701, 'timing/model': 1.3176803829846904, 'learning_rate/qwen_vl_interface': 7.214402898590974e-06, 'learning_rate/action_model': 7.214402898590973e-05, 'epoch': 1.2}) +[2026-06-19 14:01:26,322][starVLA.training.train_starvla][INFO] - Step 1886, Loss: {'train/loss': 0.10695292800664902, 'train/grad_norm_pre_clip': 4.007791519165039, 'timing/data': 0.01752142095938325, 'timing/model': 1.143085028976202, 'learning_rate/qwen_vl_interface': 7.211629707184768e-06, 'learning_rate/action_model': 7.211629707184768e-05, 'epoch': 1.2}) +[2026-06-19 14:01:30,929][starVLA.training.train_starvla][INFO] - Step 1887, Loss: {'train/loss': 0.2703646421432495, 'train/grad_norm_pre_clip': 8.726325988769531, 'timing/data': 0.018783706007525325, 'timing/model': 1.1447907319525257, 'learning_rate/qwen_vl_interface': 7.2088557094266885e-06, 'learning_rate/action_model': 7.208855709426688e-05, 'epoch': 1.2}) +[2026-06-19 14:01:35,518][starVLA.training.train_starvla][INFO] - Step 1888, Loss: {'train/loss': 0.2462003529071808, 'train/grad_norm_pre_clip': 12.469991683959961, 'timing/data': 0.018894192995503545, 'timing/model': 1.1533028709236532, 'learning_rate/qwen_vl_interface': 7.2060809064570195e-06, 'learning_rate/action_model': 7.20608090645702e-05, 'epoch': 1.2}) +[2026-06-19 14:01:40,098][starVLA.training.train_starvla][INFO] - Step 1889, Loss: {'train/loss': 0.3270893692970276, 'train/grad_norm_pre_clip': 8.980005264282227, 'timing/data': 0.018106574076227844, 'timing/model': 1.1453793910332024, 'learning_rate/qwen_vl_interface': 7.203305299416375e-06, 'learning_rate/action_model': 7.203305299416375e-05, 'epoch': 1.2}) +[2026-06-19 14:01:44,674][starVLA.training.train_starvla][INFO] - Step 1890, Loss: {'train/loss': 0.23961284756660461, 'train/grad_norm_pre_clip': 5.44351053237915, 'timing/data': 0.017903154948726296, 'timing/model': 1.1444979660445824, 'learning_rate/qwen_vl_interface': 7.200528889445707e-06, 'learning_rate/action_model': 7.200528889445706e-05, 'epoch': 1.2}) +[2026-06-19 14:01:49,254][starVLA.training.train_starvla][INFO] - Step 1891, Loss: {'train/loss': 0.07840726524591446, 'train/grad_norm_pre_clip': 3.153623104095459, 'timing/data': 0.017906980006955564, 'timing/model': 1.3069891009945422, 'learning_rate/qwen_vl_interface': 7.19775167768629e-06, 'learning_rate/action_model': 7.19775167768629e-05, 'epoch': 1.2}) +[2026-06-19 14:01:53,770][starVLA.training.train_starvla][INFO] - Step 1892, Loss: {'train/loss': 0.16183435916900635, 'train/grad_norm_pre_clip': 6.009300231933594, 'timing/data': 0.017543711001053452, 'timing/model': 1.2371322039980441, 'learning_rate/qwen_vl_interface': 7.194973665279731e-06, 'learning_rate/action_model': 7.19497366527973e-05, 'epoch': 1.2}) +[2026-06-19 14:01:58,205][starVLA.training.train_starvla][INFO] - Step 1893, Loss: {'train/loss': 0.3040845990180969, 'train/grad_norm_pre_clip': 11.698688507080078, 'timing/data': 0.01806855504401028, 'timing/model': 1.1502631459152326, 'learning_rate/qwen_vl_interface': 7.1921948533679666e-06, 'learning_rate/action_model': 7.192194853367967e-05, 'epoch': 1.2}) +[2026-06-19 14:02:02,785][starVLA.training.train_starvla][INFO] - Step 1894, Loss: {'train/loss': 0.4247058033943176, 'train/grad_norm_pre_clip': 13.052047729492188, 'timing/data': 0.018416360951960087, 'timing/model': 1.1432487039128318, 'learning_rate/qwen_vl_interface': 7.1894152430932605e-06, 'learning_rate/action_model': 7.189415243093261e-05, 'epoch': 1.2}) +[2026-06-19 14:02:07,389][starVLA.training.train_starvla][INFO] - Step 1895, Loss: {'train/loss': 0.2916630208492279, 'train/grad_norm_pre_clip': 9.223594665527344, 'timing/data': 0.018202304025180638, 'timing/model': 1.1453000759938732, 'learning_rate/qwen_vl_interface': 7.186634835598205e-06, 'learning_rate/action_model': 7.186634835598205e-05, 'epoch': 1.2}) +[2026-06-19 14:02:11,979][starVLA.training.train_starvla][INFO] - Step 1896, Loss: {'train/loss': 0.09989652782678604, 'train/grad_norm_pre_clip': 5.796417713165283, 'timing/data': 0.017724782112054527, 'timing/model': 1.1458044151077047, 'learning_rate/qwen_vl_interface': 7.183853632025724e-06, 'learning_rate/action_model': 7.183853632025723e-05, 'epoch': 1.2}) +[2026-06-19 14:02:16,567][starVLA.training.train_starvla][INFO] - Step 1897, Loss: {'train/loss': 0.2680339515209198, 'train/grad_norm_pre_clip': 10.442791938781738, 'timing/data': 0.01814136398024857, 'timing/model': 1.3027807580074295, 'learning_rate/qwen_vl_interface': 7.181071633519062e-06, 'learning_rate/action_model': 7.181071633519062e-05, 'epoch': 1.2}) +[2026-06-19 14:02:21,100][starVLA.training.train_starvla][INFO] - Step 1898, Loss: {'train/loss': 0.5162686705589294, 'train/grad_norm_pre_clip': 15.859684944152832, 'timing/data': 0.01759240694809705, 'timing/model': 1.2426390670007095, 'learning_rate/qwen_vl_interface': 7.178288841221795e-06, 'learning_rate/action_model': 7.178288841221794e-05, 'epoch': 1.2}) +[2026-06-19 14:02:25,522][starVLA.training.train_starvla][INFO] - Step 1899, Loss: {'train/loss': 0.3744140863418579, 'train/grad_norm_pre_clip': 11.11474609375, 'timing/data': 0.017660284996964037, 'timing/model': 1.143355562002398, 'learning_rate/qwen_vl_interface': 7.1755052562778215e-06, 'learning_rate/action_model': 7.175505256277822e-05, 'epoch': 1.2}) +[2026-06-19 14:02:30,139][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1900 +[2026-06-19 14:05:21,384][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1900 +[2026-06-19 14:05:21,385][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 1900 +[2026-06-19 14:08:27,752][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 1900 +[2026-06-19 14:08:27,753][starVLA.training.train_starvla][INFO] - Step 1900, Loss: {'train/loss': 0.09249638766050339, 'train/grad_norm_pre_clip': 5.052262783050537, 'eval/loss': 0.34160345792770386, 'eval/latency_0/loss': 0.11883939802646637, 'eval/latency_1/loss': 0.09605174511671066, 'eval/latency_2/loss': 0.17667633295059204, 'eval/latency_3/loss': 0.4926818609237671, 'eval/latency_4/loss': 0.8341838717460632, 'eval/flappy/loss': 0.34160345792770386, 'eval/flappy/latency_0/loss': 0.11883939802646637, 'eval/flappy/latency_1/loss': 0.09605174511671066, 'eval/flappy/latency_2/loss': 0.17667633295059204, 'eval/flappy/latency_3/loss': 0.4926818609237671, 'eval/flappy/latency_4/loss': 0.8341838717460632, 'eval/flap_precision': 0.7966101694915254, 'eval/flap_recall': 0.5127272727272727, 'eval/flap_f1': 0.6238938053097345, 'eval/flap_rate_model': 0.049166666666666664, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.6238938053097345, 'eval/latency_0/flap_precision': 0.7966101694915254, 'eval/latency_0/flap_recall': 0.5127272727272727, 'eval/latency_0/flap_f1': 0.6238938053097345, 'eval/latency_0/flap_rate_model': 0.049166666666666664, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.6238938053097345, 'timing/data': 0.017645013984292746, 'timing/model': 1.1553222679067403, 'learning_rate/qwen_vl_interface': 7.1727208798313704e-06, 'learning_rate/action_model': 7.172720879831371e-05, 'epoch': 1.2}) +[2026-06-19 14:08:27,753][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_1900_state +[2026-06-19 14:08:27,756][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-19 14:08:31,008][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_1900_state/model.safetensors +[2026-06-19 14:08:38,197][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_1900_state/optimizer.bin +[2026-06-19 14:08:38,198][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_1900_state/sampler.bin +[2026-06-19 14:08:38,198][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_1900_state/sampler_1.bin +[2026-06-19 14:08:38,199][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_1900_state/random_states_0.pkl +[2026-06-19 14:08:38,220][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1900 +[2026-06-19 14:08:38,220][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-19 14:08:38,229][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-19 14:08:42,777][starVLA.training.train_starvla][INFO] - Step 1901, Loss: {'train/loss': 0.20842918753623962, 'train/grad_norm_pre_clip': 5.215486526489258, 'timing/data': 0.018337662098929286, 'timing/model': 1.3313724560430273, 'learning_rate/qwen_vl_interface': 7.169935713026994e-06, 'learning_rate/action_model': 7.169935713026994e-05, 'epoch': 1.21}) +[2026-06-19 14:08:47,270][starVLA.training.train_starvla][INFO] - Step 1902, Loss: {'train/loss': 0.16965875029563904, 'train/grad_norm_pre_clip': 7.921555995941162, 'timing/data': 0.018739718943834305, 'timing/model': 1.2503880340373144, 'learning_rate/qwen_vl_interface': 7.1671497570095684e-06, 'learning_rate/action_model': 7.167149757009568e-05, 'epoch': 1.21}) +[2026-06-19 14:08:51,637][starVLA.training.train_starvla][INFO] - Step 1903, Loss: {'train/loss': 0.2169731855392456, 'train/grad_norm_pre_clip': 19.336519241333008, 'timing/data': 0.016788105946034193, 'timing/model': 1.1382042639888823, 'learning_rate/qwen_vl_interface': 7.164363012924295e-06, 'learning_rate/action_model': 7.164363012924294e-05, 'epoch': 1.21}) +[2026-06-19 14:08:56,131][starVLA.training.train_starvla][INFO] - Step 1904, Loss: {'train/loss': 0.3176521360874176, 'train/grad_norm_pre_clip': 12.517341613769531, 'timing/data': 0.017924004001542926, 'timing/model': 1.1366389419417828, 'learning_rate/qwen_vl_interface': 7.161575481916699e-06, 'learning_rate/action_model': 7.1615754819167e-05, 'epoch': 1.21}) +[2026-06-19 14:09:00,671][starVLA.training.train_starvla][INFO] - Step 1905, Loss: {'train/loss': 0.22932955622673035, 'train/grad_norm_pre_clip': 8.45874309539795, 'timing/data': 0.017377039999701083, 'timing/model': 1.1346884849481285, 'learning_rate/qwen_vl_interface': 7.158787165132627e-06, 'learning_rate/action_model': 7.158787165132626e-05, 'epoch': 1.21}) +[2026-06-19 14:09:05,196][starVLA.training.train_starvla][INFO] - Step 1906, Loss: {'train/loss': 0.42093613743782043, 'train/grad_norm_pre_clip': 12.39154052734375, 'timing/data': 0.016879522940143943, 'timing/model': 1.1313392339507118, 'learning_rate/qwen_vl_interface': 7.1559980637182535e-06, 'learning_rate/action_model': 7.155998063718254e-05, 'epoch': 1.21}) +[2026-06-19 14:09:09,724][starVLA.training.train_starvla][INFO] - Step 1907, Loss: {'train/loss': 0.3581703007221222, 'train/grad_norm_pre_clip': 9.48171615600586, 'timing/data': 0.01760892104357481, 'timing/model': 1.2937252899864689, 'learning_rate/qwen_vl_interface': 7.153208178820071e-06, 'learning_rate/action_model': 7.153208178820072e-05, 'epoch': 1.21}) +[2026-06-19 14:09:14,175][starVLA.training.train_starvla][INFO] - Step 1908, Loss: {'train/loss': 0.29448485374450684, 'train/grad_norm_pre_clip': 10.6785249710083, 'timing/data': 0.017393152927979827, 'timing/model': 1.2102766670286655, 'learning_rate/qwen_vl_interface': 7.150417511584898e-06, 'learning_rate/action_model': 7.150417511584898e-05, 'epoch': 1.21}) +[2026-06-19 14:09:18,556][starVLA.training.train_starvla][INFO] - Step 1909, Loss: {'train/loss': 0.14282989501953125, 'train/grad_norm_pre_clip': 10.708950996398926, 'timing/data': 0.016911529004573822, 'timing/model': 1.138749172911048, 'learning_rate/qwen_vl_interface': 7.147626063159872e-06, 'learning_rate/action_model': 7.147626063159872e-05, 'epoch': 1.21}) +[2026-06-19 14:09:23,032][starVLA.training.train_starvla][INFO] - Step 1910, Loss: {'train/loss': 0.2603628933429718, 'train/grad_norm_pre_clip': 12.552858352661133, 'timing/data': 0.01827764301560819, 'timing/model': 1.1429918908979744, 'learning_rate/qwen_vl_interface': 7.14483383469245e-06, 'learning_rate/action_model': 7.144833834692449e-05, 'epoch': 1.21}) +[2026-06-19 14:09:27,552][starVLA.training.train_starvla][INFO] - Step 1911, Loss: {'train/loss': 0.3650449812412262, 'train/grad_norm_pre_clip': 10.954238891601562, 'timing/data': 0.017822219990193844, 'timing/model': 1.1402369509451091, 'learning_rate/qwen_vl_interface': 7.142040827330411e-06, 'learning_rate/action_model': 7.142040827330411e-05, 'epoch': 1.21}) +[2026-06-19 14:09:32,137][starVLA.training.train_starvla][INFO] - Step 1912, Loss: {'train/loss': 0.3473414480686188, 'train/grad_norm_pre_clip': 11.254490852355957, 'timing/data': 0.017846444970928133, 'timing/model': 1.138567385962233, 'learning_rate/qwen_vl_interface': 7.139247042221859e-06, 'learning_rate/action_model': 7.139247042221858e-05, 'epoch': 1.21}) +[2026-06-19 14:09:36,736][starVLA.training.train_starvla][INFO] - Step 1913, Loss: {'train/loss': 0.18958374857902527, 'train/grad_norm_pre_clip': 4.953573703765869, 'timing/data': 0.01731221901718527, 'timing/model': 1.302810681052506, 'learning_rate/qwen_vl_interface': 7.136452480515208e-06, 'learning_rate/action_model': 7.136452480515208e-05, 'epoch': 1.21}) +[2026-06-19 14:09:41,149][starVLA.training.train_starvla][INFO] - Step 1914, Loss: {'train/loss': 0.1536693572998047, 'train/grad_norm_pre_clip': 5.2913665771484375, 'timing/data': 0.017522651003673673, 'timing/model': 1.1409753150073811, 'learning_rate/qwen_vl_interface': 7.133657143359203e-06, 'learning_rate/action_model': 7.133657143359202e-05, 'epoch': 1.21}) +[2026-06-19 14:09:45,714][starVLA.training.train_starvla][INFO] - Step 1915, Loss: {'train/loss': 0.43240392208099365, 'train/grad_norm_pre_clip': 12.099185943603516, 'timing/data': 0.017858595005236566, 'timing/model': 1.1379060369217768, 'learning_rate/qwen_vl_interface': 7.130861031902897e-06, 'learning_rate/action_model': 7.130861031902897e-05, 'epoch': 1.21}) +[2026-06-19 14:09:50,266][starVLA.training.train_starvla][INFO] - Step 1916, Loss: {'train/loss': 0.287477046251297, 'train/grad_norm_pre_clip': 7.114814758300781, 'timing/data': 0.018711557029746473, 'timing/model': 1.1382233399199322, 'learning_rate/qwen_vl_interface': 7.128064147295667e-06, 'learning_rate/action_model': 7.128064147295667e-05, 'epoch': 1.22}) +[2026-06-19 14:09:54,822][starVLA.training.train_starvla][INFO] - Step 1917, Loss: {'train/loss': 0.4121975600719452, 'train/grad_norm_pre_clip': 9.969666481018066, 'timing/data': 0.017928775981999934, 'timing/model': 1.138201275956817, 'learning_rate/qwen_vl_interface': 7.125266490687206e-06, 'learning_rate/action_model': 7.125266490687206e-05, 'epoch': 1.22}) +[2026-06-19 14:09:59,416][starVLA.training.train_starvla][INFO] - Step 1918, Loss: {'train/loss': 0.23485630750656128, 'train/grad_norm_pre_clip': 10.315388679504395, 'timing/data': 0.01718776300549507, 'timing/model': 1.145241146092303, 'learning_rate/qwen_vl_interface': 7.1224680632275265e-06, 'learning_rate/action_model': 7.122468063227527e-05, 'epoch': 1.22}) +[2026-06-19 14:10:04,000][starVLA.training.train_starvla][INFO] - Step 1919, Loss: {'train/loss': 0.09195084869861603, 'train/grad_norm_pre_clip': 3.2537167072296143, 'timing/data': 0.01749193691648543, 'timing/model': 1.312172353034839, 'learning_rate/qwen_vl_interface': 7.1196688660669555e-06, 'learning_rate/action_model': 7.119668866066955e-05, 'epoch': 1.22}) +[2026-06-19 14:10:08,499][starVLA.training.train_starvla][INFO] - Step 1920, Loss: {'train/loss': 0.31847044825553894, 'train/grad_norm_pre_clip': 8.383234024047852, 'timing/data': 0.018526284955441952, 'timing/model': 1.2387018030276522, 'learning_rate/qwen_vl_interface': 7.116868900356138e-06, 'learning_rate/action_model': 7.116868900356138e-05, 'epoch': 1.22}) +[2026-06-19 14:10:12,907][starVLA.training.train_starvla][INFO] - Step 1921, Loss: {'train/loss': 0.17435254156589508, 'train/grad_norm_pre_clip': 4.834111213684082, 'timing/data': 0.017378719989210367, 'timing/model': 1.1426100420067087, 'learning_rate/qwen_vl_interface': 7.114068167246033e-06, 'learning_rate/action_model': 7.114068167246033e-05, 'epoch': 1.22}) +[2026-06-19 14:10:17,487][starVLA.training.train_starvla][INFO] - Step 1922, Loss: {'train/loss': 0.39214134216308594, 'train/grad_norm_pre_clip': 13.325760841369629, 'timing/data': 0.017937969998456538, 'timing/model': 1.143463466083631, 'learning_rate/qwen_vl_interface': 7.111266667887916e-06, 'learning_rate/action_model': 7.111266667887915e-05, 'epoch': 1.22}) +[2026-06-19 14:10:22,078][starVLA.training.train_starvla][INFO] - Step 1923, Loss: {'train/loss': 0.20829549431800842, 'train/grad_norm_pre_clip': 9.977109909057617, 'timing/data': 0.01711677899584174, 'timing/model': 1.1429473620373756, 'learning_rate/qwen_vl_interface': 7.108464403433378e-06, 'learning_rate/action_model': 7.108464403433378e-05, 'epoch': 1.22}) +[2026-06-19 14:10:26,705][starVLA.training.train_starvla][INFO] - Step 1924, Loss: {'train/loss': 0.11589300632476807, 'train/grad_norm_pre_clip': 5.328619003295898, 'timing/data': 0.019084503990598023, 'timing/model': 1.1419063280336559, 'learning_rate/qwen_vl_interface': 7.105661375034324e-06, 'learning_rate/action_model': 7.105661375034324e-05, 'epoch': 1.22}) +[2026-06-19 14:10:31,309][starVLA.training.train_starvla][INFO] - Step 1925, Loss: {'train/loss': 0.3485865890979767, 'train/grad_norm_pre_clip': 10.45970344543457, 'timing/data': 0.02026330109219998, 'timing/model': 1.3078049370087683, 'learning_rate/qwen_vl_interface': 7.102857583842973e-06, 'learning_rate/action_model': 7.102857583842974e-05, 'epoch': 1.22}) +[2026-06-19 14:10:35,735][starVLA.training.train_starvla][INFO] - Step 1926, Loss: {'train/loss': 0.2943366765975952, 'train/grad_norm_pre_clip': 6.93405294418335, 'timing/data': 0.020037571899592876, 'timing/model': 1.1484768060036004, 'learning_rate/qwen_vl_interface': 7.100053031011859e-06, 'learning_rate/action_model': 7.100053031011859e-05, 'epoch': 1.22}) +[2026-06-19 14:10:40,323][starVLA.training.train_starvla][INFO] - Step 1927, Loss: {'train/loss': 0.23518647253513336, 'train/grad_norm_pre_clip': 11.316041946411133, 'timing/data': 0.01758686196990311, 'timing/model': 1.1450602440163493, 'learning_rate/qwen_vl_interface': 7.097247717693827e-06, 'learning_rate/action_model': 7.097247717693827e-05, 'epoch': 1.22}) +[2026-06-19 14:10:44,802][starVLA.training.train_starvla][INFO] - Step 1928, Loss: {'train/loss': 0.25062674283981323, 'train/grad_norm_pre_clip': 6.1543450355529785, 'timing/data': 0.017410967964679003, 'timing/model': 1.1402086101006716, 'learning_rate/qwen_vl_interface': 7.094441645042033e-06, 'learning_rate/action_model': 7.094441645042032e-05, 'epoch': 1.22}) +[2026-06-19 14:10:49,301][starVLA.training.train_starvla][INFO] - Step 1929, Loss: {'train/loss': 0.1752537190914154, 'train/grad_norm_pre_clip': 8.860275268554688, 'timing/data': 0.018627366051077843, 'timing/model': 1.1489732200279832, 'learning_rate/qwen_vl_interface': 7.09163481420995e-06, 'learning_rate/action_model': 7.09163481420995e-05, 'epoch': 1.22}) +[2026-06-19 14:10:53,812][starVLA.training.train_starvla][INFO] - Step 1930, Loss: {'train/loss': 0.23111078143119812, 'train/grad_norm_pre_clip': 10.216025352478027, 'timing/data': 0.019185815006494522, 'timing/model': 1.1475498779909685, 'learning_rate/qwen_vl_interface': 7.0888272263513595e-06, 'learning_rate/action_model': 7.088827226351359e-05, 'epoch': 1.22}) +[2026-06-19 14:10:58,394][starVLA.training.train_starvla][INFO] - Step 1931, Loss: {'train/loss': 0.41672417521476746, 'train/grad_norm_pre_clip': 8.791803359985352, 'timing/data': 0.018556348979473114, 'timing/model': 1.3072889660252258, 'learning_rate/qwen_vl_interface': 7.086018882620354e-06, 'learning_rate/action_model': 7.086018882620354e-05, 'epoch': 1.22}) +[2026-06-19 14:11:02,830][starVLA.training.train_starvla][INFO] - Step 1932, Loss: {'train/loss': 0.268375962972641, 'train/grad_norm_pre_clip': 10.938959121704102, 'timing/data': 0.019363989937119186, 'timing/model': 1.1505140240769833, 'learning_rate/qwen_vl_interface': 7.083209784171341e-06, 'learning_rate/action_model': 7.083209784171341e-05, 'epoch': 1.23}) +[2026-06-19 14:11:07,409][starVLA.training.train_starvla][INFO] - Step 1933, Loss: {'train/loss': 0.5800043940544128, 'train/grad_norm_pre_clip': 14.148673057556152, 'timing/data': 0.01886725297663361, 'timing/model': 1.1417252089595422, 'learning_rate/qwen_vl_interface': 7.080399932159029e-06, 'learning_rate/action_model': 7.080399932159029e-05, 'epoch': 1.23}) +[2026-06-19 14:11:11,984][starVLA.training.train_starvla][INFO] - Step 1934, Loss: {'train/loss': 0.090518519282341, 'train/grad_norm_pre_clip': 4.539194583892822, 'timing/data': 0.016733631025999784, 'timing/model': 1.142439577030018, 'learning_rate/qwen_vl_interface': 7.077589327738448e-06, 'learning_rate/action_model': 7.077589327738447e-05, 'epoch': 1.23}) +[2026-06-19 14:11:16,575][starVLA.training.train_starvla][INFO] - Step 1935, Loss: {'train/loss': 0.207746684551239, 'train/grad_norm_pre_clip': 6.167098045349121, 'timing/data': 0.017903328989632428, 'timing/model': 1.1434412599774078, 'learning_rate/qwen_vl_interface': 7.0747779720649265e-06, 'learning_rate/action_model': 7.074777972064927e-05, 'epoch': 1.23}) +[2026-06-19 14:11:21,147][starVLA.training.train_starvla][INFO] - Step 1936, Loss: {'train/loss': 0.24259476363658905, 'train/grad_norm_pre_clip': 9.855559349060059, 'timing/data': 0.017505382071249187, 'timing/model': 1.1421317489584908, 'learning_rate/qwen_vl_interface': 7.071965866294108e-06, 'learning_rate/action_model': 7.071965866294108e-05, 'epoch': 1.23}) +[2026-06-19 14:11:25,733][starVLA.training.train_starvla][INFO] - Step 1937, Loss: {'train/loss': 0.2575981318950653, 'train/grad_norm_pre_clip': 8.302183151245117, 'timing/data': 0.018160124076530337, 'timing/model': 1.3049506300594658, 'learning_rate/qwen_vl_interface': 7.0691530115819454e-06, 'learning_rate/action_model': 7.069153011581945e-05, 'epoch': 1.23}) +[2026-06-19 14:11:30,164][starVLA.training.train_starvla][INFO] - Step 1938, Loss: {'train/loss': 0.31739169359207153, 'train/grad_norm_pre_clip': 11.10934066772461, 'timing/data': 0.018680943991057575, 'timing/model': 1.1492620389908552, 'learning_rate/qwen_vl_interface': 7.066339409084694e-06, 'learning_rate/action_model': 7.066339409084694e-05, 'epoch': 1.23}) +[2026-06-19 14:11:34,735][starVLA.training.train_starvla][INFO] - Step 1939, Loss: {'train/loss': 0.20771971344947815, 'train/grad_norm_pre_clip': 9.51282024383545, 'timing/data': 0.01699602196458727, 'timing/model': 1.1445015319623053, 'learning_rate/qwen_vl_interface': 7.0635250599589215e-06, 'learning_rate/action_model': 7.06352505995892e-05, 'epoch': 1.23}) +[2026-06-19 14:11:39,246][starVLA.training.train_starvla][INFO] - Step 1940, Loss: {'train/loss': 0.14114424586296082, 'train/grad_norm_pre_clip': 5.76760196685791, 'timing/data': 0.017438280978240073, 'timing/model': 1.1422576749464497, 'learning_rate/qwen_vl_interface': 7.060709965361499e-06, 'learning_rate/action_model': 7.060709965361499e-05, 'epoch': 1.23}) +[2026-06-19 14:11:43,852][starVLA.training.train_starvla][INFO] - Step 1941, Loss: {'train/loss': 0.3547362685203552, 'train/grad_norm_pre_clip': 8.99118423461914, 'timing/data': 0.018470639013685286, 'timing/model': 1.1478307119105011, 'learning_rate/qwen_vl_interface': 7.057894126449606e-06, 'learning_rate/action_model': 7.057894126449606e-05, 'epoch': 1.23}) +[2026-06-19 14:11:48,467][starVLA.training.train_starvla][INFO] - Step 1942, Loss: {'train/loss': 0.2650067210197449, 'train/grad_norm_pre_clip': 10.36671257019043, 'timing/data': 0.01783252996392548, 'timing/model': 1.1466217359993607, 'learning_rate/qwen_vl_interface': 7.055077544380728e-06, 'learning_rate/action_model': 7.055077544380728e-05, 'epoch': 1.23}) +[2026-06-19 14:11:53,069][starVLA.training.train_starvla][INFO] - Step 1943, Loss: {'train/loss': 0.21863913536071777, 'train/grad_norm_pre_clip': 7.034337520599365, 'timing/data': 0.017714866087771952, 'timing/model': 1.3156800200231373, 'learning_rate/qwen_vl_interface': 7.0522602203126554e-06, 'learning_rate/action_model': 7.052260220312656e-05, 'epoch': 1.23}) +[2026-06-19 14:11:57,491][starVLA.training.train_starvla][INFO] - Step 1944, Loss: {'train/loss': 0.14602425694465637, 'train/grad_norm_pre_clip': 6.631145477294922, 'timing/data': 0.01785921200644225, 'timing/model': 1.1466046340065077, 'learning_rate/qwen_vl_interface': 7.0494421554034845e-06, 'learning_rate/action_model': 7.049442155403484e-05, 'epoch': 1.23}) +[2026-06-19 14:12:02,083][starVLA.training.train_starvla][INFO] - Step 1945, Loss: {'train/loss': 0.3599008023738861, 'train/grad_norm_pre_clip': 11.723628997802734, 'timing/data': 0.017563170986250043, 'timing/model': 1.1503307120874524, 'learning_rate/qwen_vl_interface': 7.046623350811612e-06, 'learning_rate/action_model': 7.046623350811612e-05, 'epoch': 1.23}) +[2026-06-19 14:12:06,665][starVLA.training.train_starvla][INFO] - Step 1946, Loss: {'train/loss': 0.1797487586736679, 'train/grad_norm_pre_clip': 6.713512420654297, 'timing/data': 0.01790199999231845, 'timing/model': 1.1487836679443717, 'learning_rate/qwen_vl_interface': 7.043803807695747e-06, 'learning_rate/action_model': 7.043803807695747e-05, 'epoch': 1.23}) +[2026-06-19 14:12:11,280][starVLA.training.train_starvla][INFO] - Step 1947, Loss: {'train/loss': 0.3558829426765442, 'train/grad_norm_pre_clip': 7.500076770782471, 'timing/data': 0.01708543603308499, 'timing/model': 1.1561944880522788, 'learning_rate/qwen_vl_interface': 7.040983527214893e-06, 'learning_rate/action_model': 7.040983527214893e-05, 'epoch': 1.23}) +[2026-06-19 14:12:15,867][starVLA.training.train_starvla][INFO] - Step 1948, Loss: {'train/loss': 0.3947739899158478, 'train/grad_norm_pre_clip': 12.744948387145996, 'timing/data': 0.016822070931084454, 'timing/model': 1.1435788890812546, 'learning_rate/qwen_vl_interface': 7.038162510528363e-06, 'learning_rate/action_model': 7.038162510528362e-05, 'epoch': 1.24}) +[2026-06-19 14:12:20,401][starVLA.training.train_starvla][INFO] - Step 1949, Loss: {'train/loss': 0.15130290389060974, 'train/grad_norm_pre_clip': 7.7398529052734375, 'timing/data': 0.01715585100464523, 'timing/model': 1.1535499220481142, 'learning_rate/qwen_vl_interface': 7.03534075879577e-06, 'learning_rate/action_model': 7.03534075879577e-05, 'epoch': 1.24}) +[2026-06-19 14:12:24,977][starVLA.training.train_starvla][INFO] - Step 1950, Loss: {'train/loss': 0.2237417995929718, 'train/grad_norm_pre_clip': 7.972593784332275, 'timing/data': 0.01756402791943401, 'timing/model': 1.3038420190569013, 'learning_rate/qwen_vl_interface': 7.032518273177028e-06, 'learning_rate/action_model': 7.032518273177028e-05, 'epoch': 1.24}) +[2026-06-19 14:12:29,422][starVLA.training.train_starvla][INFO] - Step 1951, Loss: {'train/loss': 0.19759105145931244, 'train/grad_norm_pre_clip': 11.108515739440918, 'timing/data': 0.01711115601938218, 'timing/model': 1.164355581975542, 'learning_rate/qwen_vl_interface': 7.029695054832357e-06, 'learning_rate/action_model': 7.029695054832357e-05, 'epoch': 1.24}) +[2026-06-19 14:12:34,014][starVLA.training.train_starvla][INFO] - Step 1952, Loss: {'train/loss': 0.1489761918783188, 'train/grad_norm_pre_clip': 5.52831506729126, 'timing/data': 0.016878767986781895, 'timing/model': 1.14873027196154, 'learning_rate/qwen_vl_interface': 7.026871104922273e-06, 'learning_rate/action_model': 7.026871104922273e-05, 'epoch': 1.24}) +[2026-06-19 14:12:38,650][starVLA.training.train_starvla][INFO] - Step 1953, Loss: {'train/loss': 0.14427611231803894, 'train/grad_norm_pre_clip': 6.552316665649414, 'timing/data': 0.018582850112579763, 'timing/model': 1.1542843710631132, 'learning_rate/qwen_vl_interface': 7.024046424607596e-06, 'learning_rate/action_model': 7.024046424607595e-05, 'epoch': 1.24}) +[2026-06-19 14:12:43,269][starVLA.training.train_starvla][INFO] - Step 1954, Loss: {'train/loss': 0.4449690580368042, 'train/grad_norm_pre_clip': 41.40354537963867, 'timing/data': 0.017575406935065985, 'timing/model': 1.1559449300402775, 'learning_rate/qwen_vl_interface': 7.021221015049445e-06, 'learning_rate/action_model': 7.021221015049446e-05, 'epoch': 1.24}) +[2026-06-19 14:12:47,891][starVLA.training.train_starvla][INFO] - Step 1955, Loss: {'train/loss': 0.3468703329563141, 'train/grad_norm_pre_clip': 5.457645893096924, 'timing/data': 0.018288132967427373, 'timing/model': 1.3280527810566127, 'learning_rate/qwen_vl_interface': 7.01839487740924e-06, 'learning_rate/action_model': 7.01839487740924e-05, 'epoch': 1.24}) +[2026-06-19 14:12:52,518][starVLA.training.train_starvla][INFO] - Step 1956, Loss: {'train/loss': 0.30974453687667847, 'train/grad_norm_pre_clip': 11.409283638000488, 'timing/data': 0.016715634032152593, 'timing/model': 1.3281839229166508, 'learning_rate/qwen_vl_interface': 7.015568012848699e-06, 'learning_rate/action_model': 7.015568012848699e-05, 'epoch': 1.24}) +[2026-06-19 14:12:56,942][starVLA.training.train_starvla][INFO] - Step 1957, Loss: {'train/loss': 0.21685035526752472, 'train/grad_norm_pre_clip': 6.325589656829834, 'timing/data': 0.017106484971009195, 'timing/model': 1.1494979460258037, 'learning_rate/qwen_vl_interface': 7.01274042252984e-06, 'learning_rate/action_model': 7.01274042252984e-05, 'epoch': 1.24}) +[2026-06-19 14:13:01,536][starVLA.training.train_starvla][INFO] - Step 1958, Loss: {'train/loss': 0.5185838341712952, 'train/grad_norm_pre_clip': 17.807201385498047, 'timing/data': 0.017062535043805838, 'timing/model': 1.1452264219988137, 'learning_rate/qwen_vl_interface': 7.009912107614978e-06, 'learning_rate/action_model': 7.009912107614978e-05, 'epoch': 1.24}) +[2026-06-19 14:13:06,158][starVLA.training.train_starvla][INFO] - Step 1959, Loss: {'train/loss': 0.17796620726585388, 'train/grad_norm_pre_clip': 15.093159675598145, 'timing/data': 0.018221343983896077, 'timing/model': 1.1467864429578185, 'learning_rate/qwen_vl_interface': 7.0070830692667244e-06, 'learning_rate/action_model': 7.007083069266724e-05, 'epoch': 1.24}) +[2026-06-19 14:13:10,741][starVLA.training.train_starvla][INFO] - Step 1960, Loss: {'train/loss': 0.3183639347553253, 'train/grad_norm_pre_clip': 13.108464241027832, 'timing/data': 0.01618655095808208, 'timing/model': 1.1440041959285736, 'learning_rate/qwen_vl_interface': 7.004253308647993e-06, 'learning_rate/action_model': 7.004253308647994e-05, 'epoch': 1.24}) +[2026-06-19 14:13:15,328][starVLA.training.train_starvla][INFO] - Step 1961, Loss: {'train/loss': 0.2727309465408325, 'train/grad_norm_pre_clip': 5.954488277435303, 'timing/data': 0.018812004011124372, 'timing/model': 1.3040958209894598, 'learning_rate/qwen_vl_interface': 7.001422826921991e-06, 'learning_rate/action_model': 7.00142282692199e-05, 'epoch': 1.24}) +[2026-06-19 14:13:19,757][starVLA.training.train_starvla][INFO] - Step 1962, Loss: {'train/loss': 0.22675473988056183, 'train/grad_norm_pre_clip': 9.303723335266113, 'timing/data': 0.017637675045989454, 'timing/model': 1.1521875039907172, 'learning_rate/qwen_vl_interface': 6.9985916252522225e-06, 'learning_rate/action_model': 6.998591625252222e-05, 'epoch': 1.24}) +[2026-06-19 14:13:24,332][starVLA.training.train_starvla][INFO] - Step 1963, Loss: {'train/loss': 0.2604750692844391, 'train/grad_norm_pre_clip': 6.087862014770508, 'timing/data': 0.018873497028835118, 'timing/model': 1.1437938319286332, 'learning_rate/qwen_vl_interface': 6.995759704802485e-06, 'learning_rate/action_model': 6.995759704802485e-05, 'epoch': 1.24}) +[2026-06-19 14:13:28,888][starVLA.training.train_starvla][INFO] - Step 1964, Loss: {'train/loss': 0.12346912175416946, 'train/grad_norm_pre_clip': 5.410841464996338, 'timing/data': 0.01781573798507452, 'timing/model': 1.1411157039692625, 'learning_rate/qwen_vl_interface': 6.992927066736876e-06, 'learning_rate/action_model': 6.992927066736875e-05, 'epoch': 1.25}) +[2026-06-19 14:13:33,456][starVLA.training.train_starvla][INFO] - Step 1965, Loss: {'train/loss': 0.2501160800457001, 'train/grad_norm_pre_clip': 9.94507884979248, 'timing/data': 0.01785916299559176, 'timing/model': 1.1416358610149473, 'learning_rate/qwen_vl_interface': 6.990093712219787e-06, 'learning_rate/action_model': 6.990093712219787e-05, 'epoch': 1.25}) +[2026-06-19 14:13:38,038][starVLA.training.train_starvla][INFO] - Step 1966, Loss: {'train/loss': 0.2034497857093811, 'train/grad_norm_pre_clip': 6.6997880935668945, 'timing/data': 0.01727107900660485, 'timing/model': 1.1430039390688762, 'learning_rate/qwen_vl_interface': 6.9872596424159e-06, 'learning_rate/action_model': 6.9872596424159e-05, 'epoch': 1.25}) +[2026-06-19 14:13:42,631][starVLA.training.train_starvla][INFO] - Step 1967, Loss: {'train/loss': 0.18436825275421143, 'train/grad_norm_pre_clip': 70.25850677490234, 'timing/data': 0.01823102100752294, 'timing/model': 1.3057413599453866, 'learning_rate/qwen_vl_interface': 6.9844248584902e-06, 'learning_rate/action_model': 6.984424858490199e-05, 'epoch': 1.25}) +[2026-06-19 14:13:47,219][starVLA.training.train_starvla][INFO] - Step 1968, Loss: {'train/loss': 0.26174408197402954, 'train/grad_norm_pre_clip': 6.848814964294434, 'timing/data': 0.016974141006357968, 'timing/model': 1.3164516550023109, 'learning_rate/qwen_vl_interface': 6.981589361607952e-06, 'learning_rate/action_model': 6.981589361607953e-05, 'epoch': 1.25}) +[2026-06-19 14:13:51,643][starVLA.training.train_starvla][INFO] - Step 1969, Loss: {'train/loss': 0.2906816303730011, 'train/grad_norm_pre_clip': 10.5778169631958, 'timing/data': 0.017741632065735757, 'timing/model': 1.1412903759628534, 'learning_rate/qwen_vl_interface': 6.9787531529347274e-06, 'learning_rate/action_model': 6.978753152934727e-05, 'epoch': 1.25}) +[2026-06-19 14:13:56,247][starVLA.training.train_starvla][INFO] - Step 1970, Loss: {'train/loss': 0.13037599623203278, 'train/grad_norm_pre_clip': 6.684443473815918, 'timing/data': 0.017830173019319773, 'timing/model': 1.1414930290775374, 'learning_rate/qwen_vl_interface': 6.97591623363638e-06, 'learning_rate/action_model': 6.97591623363638e-05, 'epoch': 1.25}) +[2026-06-19 14:14:00,839][starVLA.training.train_starvla][INFO] - Step 1971, Loss: {'train/loss': 0.15134018659591675, 'train/grad_norm_pre_clip': 6.570940017700195, 'timing/data': 0.016952092992141843, 'timing/model': 1.1429986449657008, 'learning_rate/qwen_vl_interface': 6.9730786048790644e-06, 'learning_rate/action_model': 6.973078604879064e-05, 'epoch': 1.25}) +[2026-06-19 14:14:05,346][starVLA.training.train_starvla][INFO] - Step 1972, Loss: {'train/loss': 0.17099618911743164, 'train/grad_norm_pre_clip': 4.541945457458496, 'timing/data': 0.018921438953839242, 'timing/model': 1.15487231395673, 'learning_rate/qwen_vl_interface': 6.970240267829221e-06, 'learning_rate/action_model': 6.970240267829222e-05, 'epoch': 1.25}) +[2026-06-19 14:14:09,896][starVLA.training.train_starvla][INFO] - Step 1973, Loss: {'train/loss': 0.17428861558437347, 'train/grad_norm_pre_clip': 10.863846778869629, 'timing/data': 0.018786933971568942, 'timing/model': 1.1416475389851257, 'learning_rate/qwen_vl_interface': 6.967401223653582e-06, 'learning_rate/action_model': 6.967401223653581e-05, 'epoch': 1.25}) +[2026-06-19 14:14:14,501][starVLA.training.train_starvla][INFO] - Step 1974, Loss: {'train/loss': 0.2302059829235077, 'train/grad_norm_pre_clip': 5.823570251464844, 'timing/data': 0.018193677999079227, 'timing/model': 1.3251936010783538, 'learning_rate/qwen_vl_interface': 6.9645614735191724e-06, 'learning_rate/action_model': 6.964561473519172e-05, 'epoch': 1.25}) +[2026-06-19 14:14:18,930][starVLA.training.train_starvla][INFO] - Step 1975, Loss: {'train/loss': 0.1813029795885086, 'train/grad_norm_pre_clip': 11.121757507324219, 'timing/data': 0.019039078033529222, 'timing/model': 1.1491057049715891, 'learning_rate/qwen_vl_interface': 6.961721018593305e-06, 'learning_rate/action_model': 6.961721018593305e-05, 'epoch': 1.25}) +[2026-06-19 14:14:23,540][starVLA.training.train_starvla][INFO] - Step 1976, Loss: {'train/loss': 0.3061455190181732, 'train/grad_norm_pre_clip': 10.133888244628906, 'timing/data': 0.01705829391721636, 'timing/model': 1.1445051670307294, 'learning_rate/qwen_vl_interface': 6.958879860043585e-06, 'learning_rate/action_model': 6.958879860043585e-05, 'epoch': 1.25}) +[2026-06-19 14:14:28,155][starVLA.training.train_starvla][INFO] - Step 1977, Loss: {'train/loss': 0.3888484835624695, 'train/grad_norm_pre_clip': 10.771675109863281, 'timing/data': 0.017217117943800986, 'timing/model': 1.1529987460235134, 'learning_rate/qwen_vl_interface': 6.9560379990379025e-06, 'learning_rate/action_model': 6.956037999037902e-05, 'epoch': 1.25}) +[2026-06-19 14:14:32,749][starVLA.training.train_starvla][INFO] - Step 1978, Loss: {'train/loss': 0.12096361815929413, 'train/grad_norm_pre_clip': 3.083209753036499, 'timing/data': 0.018246860010549426, 'timing/model': 1.143751123920083, 'learning_rate/qwen_vl_interface': 6.953195436744442e-06, 'learning_rate/action_model': 6.953195436744442e-05, 'epoch': 1.25}) +[2026-06-19 14:14:37,264][starVLA.training.train_starvla][INFO] - Step 1979, Loss: {'train/loss': 0.3426702320575714, 'train/grad_norm_pre_clip': 8.57146167755127, 'timing/data': 0.018028269056230783, 'timing/model': 1.153934572939761, 'learning_rate/qwen_vl_interface': 6.950352174331676e-06, 'learning_rate/action_model': 6.950352174331675e-05, 'epoch': 1.26}) +[2026-06-19 14:14:41,835][starVLA.training.train_starvla][INFO] - Step 1980, Loss: {'train/loss': 0.2569975256919861, 'train/grad_norm_pre_clip': 10.614879608154297, 'timing/data': 0.017825315007939935, 'timing/model': 1.281980522093363, 'learning_rate/qwen_vl_interface': 6.947508212968356e-06, 'learning_rate/action_model': 6.947508212968356e-05, 'epoch': 1.26}) +[2026-06-19 14:14:46,260][starVLA.training.train_starvla][INFO] - Step 1981, Loss: {'train/loss': 0.1554236263036728, 'train/grad_norm_pre_clip': 8.779994010925293, 'timing/data': 0.018589551909826696, 'timing/model': 1.1517365199979395, 'learning_rate/qwen_vl_interface': 6.9446635538235295e-06, 'learning_rate/action_model': 6.944663553823529e-05, 'epoch': 1.26}) +[2026-06-19 14:14:50,860][starVLA.training.train_starvla][INFO] - Step 1982, Loss: {'train/loss': 0.33509308099746704, 'train/grad_norm_pre_clip': 10.950370788574219, 'timing/data': 0.01807032700162381, 'timing/model': 1.148136878036894, 'learning_rate/qwen_vl_interface': 6.94181819806653e-06, 'learning_rate/action_model': 6.94181819806653e-05, 'epoch': 1.26}) +[2026-06-19 14:14:55,468][starVLA.training.train_starvla][INFO] - Step 1983, Loss: {'train/loss': 0.24590091407299042, 'train/grad_norm_pre_clip': 7.524200916290283, 'timing/data': 0.018608066020533442, 'timing/model': 1.1479067490436137, 'learning_rate/qwen_vl_interface': 6.938972146866975e-06, 'learning_rate/action_model': 6.938972146866974e-05, 'epoch': 1.26}) +[2026-06-19 14:15:00,013][starVLA.training.train_starvla][INFO] - Step 1984, Loss: {'train/loss': 0.280916690826416, 'train/grad_norm_pre_clip': 8.974148750305176, 'timing/data': 0.017389185028150678, 'timing/model': 1.1453635710058734, 'learning_rate/qwen_vl_interface': 6.936125401394769e-06, 'learning_rate/action_model': 6.936125401394768e-05, 'epoch': 1.26}) +[2026-06-19 14:15:04,610][starVLA.training.train_starvla][INFO] - Step 1985, Loss: {'train/loss': 0.17707481980323792, 'train/grad_norm_pre_clip': 6.378610610961914, 'timing/data': 0.018469686969183385, 'timing/model': 1.1450292910449207, 'learning_rate/qwen_vl_interface': 6.933277962820101e-06, 'learning_rate/action_model': 6.9332779628201e-05, 'epoch': 1.26}) +[2026-06-19 14:15:09,210][starVLA.training.train_starvla][INFO] - Step 1986, Loss: {'train/loss': 0.2725151479244232, 'train/grad_norm_pre_clip': 6.759474277496338, 'timing/data': 0.01843230298254639, 'timing/model': 1.3097866760799661, 'learning_rate/qwen_vl_interface': 6.930429832313444e-06, 'learning_rate/action_model': 6.930429832313444e-05, 'epoch': 1.26}) +[2026-06-19 14:15:13,651][starVLA.training.train_starvla][INFO] - Step 1987, Loss: {'train/loss': 0.3017858564853668, 'train/grad_norm_pre_clip': 18.27423667907715, 'timing/data': 0.017205420066602528, 'timing/model': 1.1547068579820916, 'learning_rate/qwen_vl_interface': 6.92758101104556e-06, 'learning_rate/action_model': 6.92758101104556e-05, 'epoch': 1.26}) +[2026-06-19 14:15:18,263][starVLA.training.train_starvla][INFO] - Step 1988, Loss: {'train/loss': 0.354827344417572, 'train/grad_norm_pre_clip': 12.235669136047363, 'timing/data': 0.01773795997723937, 'timing/model': 1.1504483319586143, 'learning_rate/qwen_vl_interface': 6.9247315001874885e-06, 'learning_rate/action_model': 6.924731500187488e-05, 'epoch': 1.26}) +[2026-06-19 14:15:22,780][starVLA.training.train_starvla][INFO] - Step 1989, Loss: {'train/loss': 0.40550288558006287, 'train/grad_norm_pre_clip': 14.701301574707031, 'timing/data': 0.017843819921836257, 'timing/model': 1.1476500169374049, 'learning_rate/qwen_vl_interface': 6.9218813009105605e-06, 'learning_rate/action_model': 6.92188130091056e-05, 'epoch': 1.26}) +[2026-06-19 14:15:27,316][starVLA.training.train_starvla][INFO] - Step 1990, Loss: {'train/loss': 0.34953632950782776, 'train/grad_norm_pre_clip': 9.688178062438965, 'timing/data': 0.01718643505591899, 'timing/model': 1.147499882034026, 'learning_rate/qwen_vl_interface': 6.919030414386378e-06, 'learning_rate/action_model': 6.919030414386379e-05, 'epoch': 1.26}) +[2026-06-19 14:15:31,926][starVLA.training.train_starvla][INFO] - Step 1991, Loss: {'train/loss': 0.43587782979011536, 'train/grad_norm_pre_clip': 9.275303840637207, 'timing/data': 0.017071385984309018, 'timing/model': 1.1456243520369753, 'learning_rate/qwen_vl_interface': 6.916178841786839e-06, 'learning_rate/action_model': 6.916178841786838e-05, 'epoch': 1.26}) +[2026-06-19 14:15:36,527][starVLA.training.train_starvla][INFO] - Step 1992, Loss: {'train/loss': 0.15750093758106232, 'train/grad_norm_pre_clip': 8.894577980041504, 'timing/data': 0.01818558492232114, 'timing/model': 1.3280205100309104, 'learning_rate/qwen_vl_interface': 6.913326584284115e-06, 'learning_rate/action_model': 6.913326584284115e-05, 'epoch': 1.26}) +[2026-06-19 14:15:40,953][starVLA.training.train_starvla][INFO] - Step 1993, Loss: {'train/loss': 0.13568741083145142, 'train/grad_norm_pre_clip': 6.447595119476318, 'timing/data': 0.01787417510058731, 'timing/model': 1.1531100909924135, 'learning_rate/qwen_vl_interface': 6.910473643050662e-06, 'learning_rate/action_model': 6.910473643050662e-05, 'epoch': 1.26}) +[2026-06-19 14:15:45,553][starVLA.training.train_starvla][INFO] - Step 1994, Loss: {'train/loss': 0.21285052597522736, 'train/grad_norm_pre_clip': 6.803378582000732, 'timing/data': 0.01762207201682031, 'timing/model': 1.1527394760632887, 'learning_rate/qwen_vl_interface': 6.907620019259215e-06, 'learning_rate/action_model': 6.907620019259216e-05, 'epoch': 1.26}) +[2026-06-19 14:15:50,136][starVLA.training.train_starvla][INFO] - Step 1995, Loss: {'train/loss': 0.19836769998073578, 'train/grad_norm_pre_clip': 9.707463264465332, 'timing/data': 0.017794413026422262, 'timing/model': 1.1481892640003935, 'learning_rate/qwen_vl_interface': 6.9047657140827935e-06, 'learning_rate/action_model': 6.904765714082794e-05, 'epoch': 1.27}) +[2026-06-19 14:15:54,633][starVLA.training.train_starvla][INFO] - Step 1996, Loss: {'train/loss': 0.2830777168273926, 'train/grad_norm_pre_clip': 5.533676624298096, 'timing/data': 0.018284305930137634, 'timing/model': 1.1420736759901047, 'learning_rate/qwen_vl_interface': 6.901910728694694e-06, 'learning_rate/action_model': 6.901910728694694e-05, 'epoch': 1.27}) +[2026-06-19 14:15:59,161][starVLA.training.train_starvla][INFO] - Step 1997, Loss: {'train/loss': 0.2806883454322815, 'train/grad_norm_pre_clip': 8.200189590454102, 'timing/data': 0.017397730029188097, 'timing/model': 1.1401367919752374, 'learning_rate/qwen_vl_interface': 6.899055064268489e-06, 'learning_rate/action_model': 6.899055064268489e-05, 'epoch': 1.27}) +[2026-06-19 14:16:03,772][starVLA.training.train_starvla][INFO] - Step 1998, Loss: {'train/loss': 0.23455576598644257, 'train/grad_norm_pre_clip': 6.549951553344727, 'timing/data': 0.016448662034235895, 'timing/model': 1.3295524459099397, 'learning_rate/qwen_vl_interface': 6.896198721978038e-06, 'learning_rate/action_model': 6.896198721978038e-05, 'epoch': 1.27}) +[2026-06-19 14:16:08,208][starVLA.training.train_starvla][INFO] - Step 1999, Loss: {'train/loss': 0.20369891822338104, 'train/grad_norm_pre_clip': 4.376524925231934, 'timing/data': 0.018757691024802625, 'timing/model': 1.1473251809366047, 'learning_rate/qwen_vl_interface': 6.893341702997475e-06, 'learning_rate/action_model': 6.893341702997475e-05, 'epoch': 1.27}) +[2026-06-19 14:16:12,808][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 2000 +[2026-06-19 14:19:09,441][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 2000 +[2026-06-19 14:19:09,441][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 2000 +[2026-06-19 14:22:14,144][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 2000 +[2026-06-19 14:22:14,145][starVLA.training.train_starvla][INFO] - Step 2000, Loss: {'train/loss': 0.2889645993709564, 'train/grad_norm_pre_clip': 7.524848937988281, 'eval/loss': 0.2943080961704254, 'eval/latency_0/loss': 0.12142693996429443, 'eval/latency_1/loss': 0.10680072754621506, 'eval/latency_2/loss': 0.16350215673446655, 'eval/latency_3/loss': 0.43319830298423767, 'eval/latency_4/loss': 0.6567724347114563, 'eval/flappy/loss': 0.2943080961704254, 'eval/flappy/latency_0/loss': 0.12142693996429443, 'eval/flappy/latency_1/loss': 0.10680072754621506, 'eval/flappy/latency_2/loss': 0.16350215673446655, 'eval/flappy/latency_3/loss': 0.43319830298423767, 'eval/flappy/latency_4/loss': 0.6567724347114563, 'eval/flap_precision': 0.7965116279069767, 'eval/flap_recall': 0.49818181818181817, 'eval/flap_f1': 0.6129753914988814, 'eval/flap_rate_model': 0.04777777777777778, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.6129753914988814, 'eval/latency_0/flap_precision': 0.7965116279069767, 'eval/latency_0/flap_recall': 0.49818181818181817, 'eval/latency_0/flap_f1': 0.6129753914988814, 'eval/latency_0/flap_rate_model': 0.04777777777777778, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.6129753914988814, 'timing/data': 0.018446299945935607, 'timing/model': 1.1468278339598328, 'learning_rate/qwen_vl_interface': 6.890484008501211e-06, 'learning_rate/action_model': 6.890484008501211e-05, 'epoch': 1.27}) +[2026-06-19 14:22:14,145][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_2000_state +[2026-06-19 14:22:14,148][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-19 14:22:17,530][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_2000_state/model.safetensors +[2026-06-19 14:22:24,741][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_2000_state/optimizer.bin +[2026-06-19 14:22:24,742][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_2000_state/sampler.bin +[2026-06-19 14:22:24,742][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_2000_state/sampler_1.bin +[2026-06-19 14:22:24,743][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_2000_state/random_states_0.pkl +[2026-06-19 14:22:24,767][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2000 +[2026-06-19 14:22:24,767][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-19 14:22:24,807][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-19 14:22:29,253][starVLA.training.train_starvla][INFO] - Step 2001, Loss: {'train/loss': 0.3300437927246094, 'train/grad_norm_pre_clip': 9.845537185668945, 'timing/data': 0.01787059602793306, 'timing/model': 1.1320021469146013, 'learning_rate/qwen_vl_interface': 6.887625639663941e-06, 'learning_rate/action_model': 6.88762563966394e-05, 'epoch': 1.27}) +[2026-06-19 14:22:33,731][starVLA.training.train_starvla][INFO] - Step 2002, Loss: {'train/loss': 0.3353460431098938, 'train/grad_norm_pre_clip': 5.639178276062012, 'timing/data': 0.017949196975678205, 'timing/model': 1.2619620829354972, 'learning_rate/qwen_vl_interface': 6.884766597660626e-06, 'learning_rate/action_model': 6.884766597660626e-05, 'epoch': 1.27}) +[2026-06-19 14:22:38,085][starVLA.training.train_starvla][INFO] - Step 2003, Loss: {'train/loss': 0.5241782069206238, 'train/grad_norm_pre_clip': 10.66191577911377, 'timing/data': 0.017676200019195676, 'timing/model': 1.1348310089670122, 'learning_rate/qwen_vl_interface': 6.881906883666515e-06, 'learning_rate/action_model': 6.881906883666515e-05, 'epoch': 1.27}) +[2026-06-19 14:22:42,613][starVLA.training.train_starvla][INFO] - Step 2004, Loss: {'train/loss': 0.1247948557138443, 'train/grad_norm_pre_clip': 4.023845672607422, 'timing/data': 0.016837335075251758, 'timing/model': 1.1297869499539956, 'learning_rate/qwen_vl_interface': 6.879046498857125e-06, 'learning_rate/action_model': 6.879046498857124e-05, 'epoch': 1.27}) +[2026-06-19 14:22:47,156][starVLA.training.train_starvla][INFO] - Step 2005, Loss: {'train/loss': 0.29704928398132324, 'train/grad_norm_pre_clip': 6.429655075073242, 'timing/data': 0.016244192025624216, 'timing/model': 1.1314714120235294, 'learning_rate/qwen_vl_interface': 6.876185444408255e-06, 'learning_rate/action_model': 6.876185444408255e-05, 'epoch': 1.27}) +[2026-06-19 14:22:51,626][starVLA.training.train_starvla][INFO] - Step 2006, Loss: {'train/loss': 0.33272239565849304, 'train/grad_norm_pre_clip': 5.993643283843994, 'timing/data': 0.018090485013090074, 'timing/model': 1.1389782519545406, 'learning_rate/qwen_vl_interface': 6.873323721495976e-06, 'learning_rate/action_model': 6.873323721495976e-05, 'epoch': 1.27}) +[2026-06-19 14:22:56,147][starVLA.training.train_starvla][INFO] - Step 2007, Loss: {'train/loss': 0.1759946346282959, 'train/grad_norm_pre_clip': 5.044299602508545, 'timing/data': 0.018294488079845905, 'timing/model': 1.1335044630104676, 'learning_rate/qwen_vl_interface': 6.870461331296632e-06, 'learning_rate/action_model': 6.870461331296632e-05, 'epoch': 1.27}) +[2026-06-19 14:23:00,702][starVLA.training.train_starvla][INFO] - Step 2008, Loss: {'train/loss': 0.19160984456539154, 'train/grad_norm_pre_clip': 7.339726448059082, 'timing/data': 0.017146457103081048, 'timing/model': 1.2991151100723073, 'learning_rate/qwen_vl_interface': 6.8675982749868455e-06, 'learning_rate/action_model': 6.867598274986845e-05, 'epoch': 1.27}) +[2026-06-19 14:23:05,098][starVLA.training.train_starvla][INFO] - Step 2009, Loss: {'train/loss': 0.25562024116516113, 'train/grad_norm_pre_clip': 6.251913070678711, 'timing/data': 0.017481176066212356, 'timing/model': 1.1467655469896272, 'learning_rate/qwen_vl_interface': 6.864734553743509e-06, 'learning_rate/action_model': 6.86473455374351e-05, 'epoch': 1.27}) +[2026-06-19 14:23:09,664][starVLA.training.train_starvla][INFO] - Step 2010, Loss: {'train/loss': 0.23674723505973816, 'train/grad_norm_pre_clip': 7.340991973876953, 'timing/data': 0.017462118063122034, 'timing/model': 1.143692065961659, 'learning_rate/qwen_vl_interface': 6.861870168743793e-06, 'learning_rate/action_model': 6.861870168743793e-05, 'epoch': 1.27}) +[2026-06-19 14:23:14,219][starVLA.training.train_starvla][INFO] - Step 2011, Loss: {'train/loss': 0.3278353214263916, 'train/grad_norm_pre_clip': 5.782637119293213, 'timing/data': 0.017774517997168005, 'timing/model': 1.1353606979828328, 'learning_rate/qwen_vl_interface': 6.859005121165135e-06, 'learning_rate/action_model': 6.859005121165135e-05, 'epoch': 1.28}) +[2026-06-19 14:23:18,721][starVLA.training.train_starvla][INFO] - Step 2012, Loss: {'train/loss': 0.3287189304828644, 'train/grad_norm_pre_clip': 11.023762702941895, 'timing/data': 0.01900674297939986, 'timing/model': 1.1415008319308981, 'learning_rate/qwen_vl_interface': 6.85613941218525e-06, 'learning_rate/action_model': 6.85613941218525e-05, 'epoch': 1.28}) +[2026-06-19 14:23:23,307][starVLA.training.train_starvla][INFO] - Step 2013, Loss: {'train/loss': 0.18893328309059143, 'train/grad_norm_pre_clip': 7.906788349151611, 'timing/data': 0.017575759906321764, 'timing/model': 1.157318243989721, 'learning_rate/qwen_vl_interface': 6.8532730429821215e-06, 'learning_rate/action_model': 6.853273042982122e-05, 'epoch': 1.28}) +[2026-06-19 14:23:27,892][starVLA.training.train_starvla][INFO] - Step 2014, Loss: {'train/loss': 0.34590867161750793, 'train/grad_norm_pre_clip': 6.900412559509277, 'timing/data': 0.019824433024041355, 'timing/model': 1.3325267030159011, 'learning_rate/qwen_vl_interface': 6.850406014734005e-06, 'learning_rate/action_model': 6.850406014734005e-05, 'epoch': 1.28}) +[2026-06-19 14:23:32,290][starVLA.training.train_starvla][INFO] - Step 2015, Loss: {'train/loss': 0.1555027812719345, 'train/grad_norm_pre_clip': 4.0308027267456055, 'timing/data': 0.01773727301042527, 'timing/model': 1.1385092620039359, 'learning_rate/qwen_vl_interface': 6.8475383286194295e-06, 'learning_rate/action_model': 6.84753832861943e-05, 'epoch': 1.28}) +[2026-06-19 14:23:36,852][starVLA.training.train_starvla][INFO] - Step 2016, Loss: {'train/loss': 0.1918487250804901, 'train/grad_norm_pre_clip': 4.7505879402160645, 'timing/data': 0.01840892608743161, 'timing/model': 1.1385537829482928, 'learning_rate/qwen_vl_interface': 6.844669985817192e-06, 'learning_rate/action_model': 6.844669985817192e-05, 'epoch': 1.28}) +[2026-06-19 14:23:41,402][starVLA.training.train_starvla][INFO] - Step 2017, Loss: {'train/loss': 0.25558754801750183, 'train/grad_norm_pre_clip': 6.726296424865723, 'timing/data': 0.01708473707549274, 'timing/model': 1.138127589947544, 'learning_rate/qwen_vl_interface': 6.8418009875063594e-06, 'learning_rate/action_model': 6.84180098750636e-05, 'epoch': 1.28}) +[2026-06-19 14:23:45,969][starVLA.training.train_starvla][INFO] - Step 2018, Loss: {'train/loss': 0.1999010592699051, 'train/grad_norm_pre_clip': 5.160183906555176, 'timing/data': 0.01764777407515794, 'timing/model': 1.1438063000096008, 'learning_rate/qwen_vl_interface': 6.838931334866267e-06, 'learning_rate/action_model': 6.838931334866266e-05, 'epoch': 1.28}) +[2026-06-19 14:23:50,552][starVLA.training.train_starvla][INFO] - Step 2019, Loss: {'train/loss': 0.20044589042663574, 'train/grad_norm_pre_clip': 7.518404960632324, 'timing/data': 0.017671757959760725, 'timing/model': 1.1480243219994009, 'learning_rate/qwen_vl_interface': 6.836061029076523e-06, 'learning_rate/action_model': 6.836061029076522e-05, 'epoch': 1.28}) +[2026-06-19 14:23:55,140][starVLA.training.train_starvla][INFO] - Step 2020, Loss: {'train/loss': 0.11265834420919418, 'train/grad_norm_pre_clip': 5.211441993713379, 'timing/data': 0.01750907301902771, 'timing/model': 1.3042003489099443, 'learning_rate/qwen_vl_interface': 6.833190071317e-06, 'learning_rate/action_model': 6.833190071316999e-05, 'epoch': 1.28}) +[2026-06-19 14:23:59,559][starVLA.training.train_starvla][INFO] - Step 2021, Loss: {'train/loss': 0.2383650541305542, 'train/grad_norm_pre_clip': 3.581583261489868, 'timing/data': 0.017796533997170627, 'timing/model': 1.1455335300415754, 'learning_rate/qwen_vl_interface': 6.830318462767841e-06, 'learning_rate/action_model': 6.830318462767841e-05, 'epoch': 1.28}) +[2026-06-19 14:24:04,155][starVLA.training.train_starvla][INFO] - Step 2022, Loss: {'train/loss': 0.10613644123077393, 'train/grad_norm_pre_clip': 4.9486083984375, 'timing/data': 0.017000629915855825, 'timing/model': 1.145237233955413, 'learning_rate/qwen_vl_interface': 6.827446204609456e-06, 'learning_rate/action_model': 6.827446204609456e-05, 'epoch': 1.28}) +[2026-06-19 14:24:08,667][starVLA.training.train_starvla][INFO] - Step 2023, Loss: {'train/loss': 0.29248350858688354, 'train/grad_norm_pre_clip': 11.780815124511719, 'timing/data': 0.017594660981558263, 'timing/model': 1.1444580510724336, 'learning_rate/qwen_vl_interface': 6.824573298022522e-06, 'learning_rate/action_model': 6.824573298022522e-05, 'epoch': 1.28}) +[2026-06-19 14:24:13,189][starVLA.training.train_starvla][INFO] - Step 2024, Loss: {'train/loss': 0.1622605174779892, 'train/grad_norm_pre_clip': 5.493865489959717, 'timing/data': 0.017176238005049527, 'timing/model': 1.1513160450849682, 'learning_rate/qwen_vl_interface': 6.8216997441879805e-06, 'learning_rate/action_model': 6.821699744187981e-05, 'epoch': 1.28}) +[2026-06-19 14:24:17,721][starVLA.training.train_starvla][INFO] - Step 2025, Loss: {'train/loss': 0.25837475061416626, 'train/grad_norm_pre_clip': 9.416638374328613, 'timing/data': 0.017675544950179756, 'timing/model': 1.1469602930592373, 'learning_rate/qwen_vl_interface': 6.818825544287045e-06, 'learning_rate/action_model': 6.818825544287045e-05, 'epoch': 1.28}) +[2026-06-19 14:24:22,326][starVLA.training.train_starvla][INFO] - Step 2026, Loss: {'train/loss': 0.19760379195213318, 'train/grad_norm_pre_clip': 8.007688522338867, 'timing/data': 0.017890420043841004, 'timing/model': 1.3297155159525573, 'learning_rate/qwen_vl_interface': 6.815950699501186e-06, 'learning_rate/action_model': 6.815950699501186e-05, 'epoch': 1.28}) +[2026-06-19 14:24:26,748][starVLA.training.train_starvla][INFO] - Step 2027, Loss: {'train/loss': 0.30884850025177, 'train/grad_norm_pre_clip': 8.990663528442383, 'timing/data': 0.017622785991989076, 'timing/model': 1.1488524940796196, 'learning_rate/qwen_vl_interface': 6.813075211012148e-06, 'learning_rate/action_model': 6.813075211012147e-05, 'epoch': 1.29}) +[2026-06-19 14:24:31,362][starVLA.training.train_starvla][INFO] - Step 2028, Loss: {'train/loss': 0.34827324748039246, 'train/grad_norm_pre_clip': 11.049352645874023, 'timing/data': 0.01749399397522211, 'timing/model': 1.151873201946728, 'learning_rate/qwen_vl_interface': 6.81019908000193e-06, 'learning_rate/action_model': 6.81019908000193e-05, 'epoch': 1.29}) +[2026-06-19 14:24:35,998][starVLA.training.train_starvla][INFO] - Step 2029, Loss: {'train/loss': 0.3383641541004181, 'train/grad_norm_pre_clip': 8.898517608642578, 'timing/data': 0.01736400800291449, 'timing/model': 1.1608856220263988, 'learning_rate/qwen_vl_interface': 6.807322307652806e-06, 'learning_rate/action_model': 6.807322307652806e-05, 'epoch': 1.29}) +[2026-06-19 14:24:40,595][starVLA.training.train_starvla][INFO] - Step 2030, Loss: {'train/loss': 0.33566293120384216, 'train/grad_norm_pre_clip': 11.454658508300781, 'timing/data': 0.01669709593988955, 'timing/model': 1.1544366289163008, 'learning_rate/qwen_vl_interface': 6.804444895147306e-06, 'learning_rate/action_model': 6.804444895147306e-05, 'epoch': 1.29}) +[2026-06-19 14:24:45,176][starVLA.training.train_starvla][INFO] - Step 2031, Loss: {'train/loss': 0.30974069237709045, 'train/grad_norm_pre_clip': 9.222475051879883, 'timing/data': 0.01803130900952965, 'timing/model': 1.1441091450396925, 'learning_rate/qwen_vl_interface': 6.8015668436682255e-06, 'learning_rate/action_model': 6.801566843668226e-05, 'epoch': 1.29}) +[2026-06-19 14:24:49,777][starVLA.training.train_starvla][INFO] - Step 2032, Loss: {'train/loss': 0.28268498182296753, 'train/grad_norm_pre_clip': 7.960680961608887, 'timing/data': 0.01675392792094499, 'timing/model': 1.3156344320159405, 'learning_rate/qwen_vl_interface': 6.7986881543986245e-06, 'learning_rate/action_model': 6.798688154398624e-05, 'epoch': 1.29}) +[2026-06-19 14:24:54,200][starVLA.training.train_starvla][INFO] - Step 2033, Loss: {'train/loss': 0.19550709426403046, 'train/grad_norm_pre_clip': 7.960771083831787, 'timing/data': 0.017568737035617232, 'timing/model': 1.1429546490544453, 'learning_rate/qwen_vl_interface': 6.795808828521822e-06, 'learning_rate/action_model': 6.795808828521821e-05, 'epoch': 1.29}) +[2026-06-19 14:24:58,775][starVLA.training.train_starvla][INFO] - Step 2034, Loss: {'train/loss': 0.267331600189209, 'train/grad_norm_pre_clip': 10.45228099822998, 'timing/data': 0.018839785014279187, 'timing/model': 1.15139578608796, 'learning_rate/qwen_vl_interface': 6.792928867221401e-06, 'learning_rate/action_model': 6.792928867221401e-05, 'epoch': 1.29}) +[2026-06-19 14:25:03,283][starVLA.training.train_starvla][INFO] - Step 2035, Loss: {'train/loss': 0.1361139714717865, 'train/grad_norm_pre_clip': 6.1312479972839355, 'timing/data': 0.017965298960916698, 'timing/model': 1.143141666892916, 'learning_rate/qwen_vl_interface': 6.790048271681202e-06, 'learning_rate/action_model': 6.790048271681202e-05, 'epoch': 1.29}) +[2026-06-19 14:25:07,862][starVLA.training.train_starvla][INFO] - Step 2036, Loss: {'train/loss': 0.1708739697933197, 'train/grad_norm_pre_clip': 6.737325668334961, 'timing/data': 0.01766429899726063, 'timing/model': 1.1455798120005056, 'learning_rate/qwen_vl_interface': 6.7871670430853335e-06, 'learning_rate/action_model': 6.787167043085333e-05, 'epoch': 1.29}) +[2026-06-19 14:25:12,439][starVLA.training.train_starvla][INFO] - Step 2037, Loss: {'train/loss': 0.22812239825725555, 'train/grad_norm_pre_clip': 10.888396263122559, 'timing/data': 0.01754899590741843, 'timing/model': 1.1396646701032296, 'learning_rate/qwen_vl_interface': 6.784285182618156e-06, 'learning_rate/action_model': 6.784285182618156e-05, 'epoch': 1.29}) +[2026-06-19 14:25:17,005][starVLA.training.train_starvla][INFO] - Step 2038, Loss: {'train/loss': 0.15585410594940186, 'train/grad_norm_pre_clip': 7.726358890533447, 'timing/data': 0.018389572971500456, 'timing/model': 1.305224274052307, 'learning_rate/qwen_vl_interface': 6.781402691464297e-06, 'learning_rate/action_model': 6.781402691464297e-05, 'epoch': 1.29}) +[2026-06-19 14:25:21,417][starVLA.training.train_starvla][INFO] - Step 2039, Loss: {'train/loss': 0.2690734565258026, 'train/grad_norm_pre_clip': 16.710800170898438, 'timing/data': 0.018068457022309303, 'timing/model': 1.142310378025286, 'learning_rate/qwen_vl_interface': 6.778519570808636e-06, 'learning_rate/action_model': 6.778519570808636e-05, 'epoch': 1.29}) +[2026-06-19 14:25:26,017][starVLA.training.train_starvla][INFO] - Step 2040, Loss: {'train/loss': 0.15077711641788483, 'train/grad_norm_pre_clip': 7.10125732421875, 'timing/data': 0.017626181012019515, 'timing/model': 1.1413673399947584, 'learning_rate/qwen_vl_interface': 6.775635821836317e-06, 'learning_rate/action_model': 6.775635821836317e-05, 'epoch': 1.29}) +[2026-06-19 14:25:30,622][starVLA.training.train_starvla][INFO] - Step 2041, Loss: {'train/loss': 0.09869721531867981, 'train/grad_norm_pre_clip': 2.942261219024658, 'timing/data': 0.01865485101006925, 'timing/model': 1.1454936619848013, 'learning_rate/qwen_vl_interface': 6.77275144573274e-06, 'learning_rate/action_model': 6.77275144573274e-05, 'epoch': 1.29}) +[2026-06-19 14:25:35,210][starVLA.training.train_starvla][INFO] - Step 2042, Loss: {'train/loss': 0.13111931085586548, 'train/grad_norm_pre_clip': 6.315288066864014, 'timing/data': 0.018733114004135132, 'timing/model': 1.1432504230178893, 'learning_rate/qwen_vl_interface': 6.769866443683564e-06, 'learning_rate/action_model': 6.769866443683563e-05, 'epoch': 1.29}) +[2026-06-19 14:25:39,794][starVLA.training.train_starvla][INFO] - Step 2043, Loss: {'train/loss': 0.35431525111198425, 'train/grad_norm_pre_clip': 12.341458320617676, 'timing/data': 0.01841452799271792, 'timing/model': 1.1445894819917157, 'learning_rate/qwen_vl_interface': 6.766980816874701e-06, 'learning_rate/action_model': 6.7669808168747e-05, 'epoch': 1.3}) +[2026-06-19 14:25:44,327][starVLA.training.train_starvla][INFO] - Step 2044, Loss: {'train/loss': 0.16058896481990814, 'train/grad_norm_pre_clip': 6.878413677215576, 'timing/data': 0.017869711969979107, 'timing/model': 1.144354676012881, 'learning_rate/qwen_vl_interface': 6.764094566492327e-06, 'learning_rate/action_model': 6.764094566492328e-05, 'epoch': 1.3}) +[2026-06-19 14:25:48,917][starVLA.training.train_starvla][INFO] - Step 2045, Loss: {'train/loss': 0.23968985676765442, 'train/grad_norm_pre_clip': 9.228001594543457, 'timing/data': 0.01793307496700436, 'timing/model': 1.30474047397729, 'learning_rate/qwen_vl_interface': 6.7612076937228685e-06, 'learning_rate/action_model': 6.761207693722869e-05, 'epoch': 1.3}) +[2026-06-19 14:25:53,330][starVLA.training.train_starvla][INFO] - Step 2046, Loss: {'train/loss': 0.27891889214515686, 'train/grad_norm_pre_clip': 12.45272159576416, 'timing/data': 0.01738856907468289, 'timing/model': 1.1431786840548739, 'learning_rate/qwen_vl_interface': 6.7583201997530126e-06, 'learning_rate/action_model': 6.758320199753012e-05, 'epoch': 1.3}) +[2026-06-19 14:25:57,889][starVLA.training.train_starvla][INFO] - Step 2047, Loss: {'train/loss': 0.35157066583633423, 'train/grad_norm_pre_clip': 14.311816215515137, 'timing/data': 0.0179599579423666, 'timing/model': 1.144969747052528, 'learning_rate/qwen_vl_interface': 6.755432085769693e-06, 'learning_rate/action_model': 6.755432085769694e-05, 'epoch': 1.3}) +[2026-06-19 14:26:02,479][starVLA.training.train_starvla][INFO] - Step 2048, Loss: {'train/loss': 0.2875182032585144, 'train/grad_norm_pre_clip': 9.0324125289917, 'timing/data': 0.018153926939703524, 'timing/model': 1.1469097319059074, 'learning_rate/qwen_vl_interface': 6.75254335296011e-06, 'learning_rate/action_model': 6.75254335296011e-05, 'epoch': 1.3}) +[2026-06-19 14:26:07,076][starVLA.training.train_starvla][INFO] - Step 2049, Loss: {'train/loss': 0.4019715189933777, 'train/grad_norm_pre_clip': 14.221945762634277, 'timing/data': 0.018367159995250404, 'timing/model': 1.1472035669721663, 'learning_rate/qwen_vl_interface': 6.749654002511712e-06, 'learning_rate/action_model': 6.749654002511712e-05, 'epoch': 1.3}) +[2026-06-19 14:26:11,651][starVLA.training.train_starvla][INFO] - Step 2050, Loss: {'train/loss': 0.3124650716781616, 'train/grad_norm_pre_clip': 9.292295455932617, 'timing/data': 0.017024861997924745, 'timing/model': 1.3028235929086804, 'learning_rate/qwen_vl_interface': 6.7467640356122e-06, 'learning_rate/action_model': 6.746764035612199e-05, 'epoch': 1.3}) +[2026-06-19 14:26:16,241][starVLA.training.train_starvla][INFO] - Step 2051, Loss: {'train/loss': 0.1256684958934784, 'train/grad_norm_pre_clip': 5.047618389129639, 'timing/data': 0.018241076031699777, 'timing/model': 1.3141060780035332, 'learning_rate/qwen_vl_interface': 6.74387345344953e-06, 'learning_rate/action_model': 6.74387345344953e-05, 'epoch': 1.3}) +[2026-06-19 14:26:20,665][starVLA.training.train_starvla][INFO] - Step 2052, Loss: {'train/loss': 0.3375500440597534, 'train/grad_norm_pre_clip': 11.469352722167969, 'timing/data': 0.018378299893811345, 'timing/model': 1.1504483259050176, 'learning_rate/qwen_vl_interface': 6.740982257211911e-06, 'learning_rate/action_model': 6.740982257211911e-05, 'epoch': 1.3}) +[2026-06-19 14:26:25,242][starVLA.training.train_starvla][INFO] - Step 2053, Loss: {'train/loss': 0.36068469285964966, 'train/grad_norm_pre_clip': 8.679975509643555, 'timing/data': 0.01792934292461723, 'timing/model': 1.1446521999314427, 'learning_rate/qwen_vl_interface': 6.738090448087806e-06, 'learning_rate/action_model': 6.738090448087806e-05, 'epoch': 1.3}) +[2026-06-19 14:26:29,848][starVLA.training.train_starvla][INFO] - Step 2054, Loss: {'train/loss': 0.2071501761674881, 'train/grad_norm_pre_clip': 7.373991012573242, 'timing/data': 0.0171496260445565, 'timing/model': 1.1444312171079218, 'learning_rate/qwen_vl_interface': 6.735198027265928e-06, 'learning_rate/action_model': 6.735198027265928e-05, 'epoch': 1.3}) +[2026-06-19 14:26:34,350][starVLA.training.train_starvla][INFO] - Step 2055, Loss: {'train/loss': 0.1560351401567459, 'train/grad_norm_pre_clip': 5.065420627593994, 'timing/data': 0.01945483300369233, 'timing/model': 1.1482119639404118, 'learning_rate/qwen_vl_interface': 6.7323049959352424e-06, 'learning_rate/action_model': 6.732304995935243e-05, 'epoch': 1.3}) +[2026-06-19 14:26:38,947][starVLA.training.train_starvla][INFO] - Step 2056, Loss: {'train/loss': 0.28933897614479065, 'train/grad_norm_pre_clip': 7.579813480377197, 'timing/data': 0.01776427193544805, 'timing/model': 1.1413906420348212, 'learning_rate/qwen_vl_interface': 6.729411355284962e-06, 'learning_rate/action_model': 6.729411355284963e-05, 'epoch': 1.3}) +[2026-06-19 14:26:43,554][starVLA.training.train_starvla][INFO] - Step 2057, Loss: {'train/loss': 0.14055505394935608, 'train/grad_norm_pre_clip': 4.018832683563232, 'timing/data': 0.01744280697312206, 'timing/model': 1.329704736941494, 'learning_rate/qwen_vl_interface': 6.726517106504559e-06, 'learning_rate/action_model': 6.726517106504558e-05, 'epoch': 1.3}) +[2026-06-19 14:26:47,982][starVLA.training.train_starvla][INFO] - Step 2058, Loss: {'train/loss': 0.09331931173801422, 'train/grad_norm_pre_clip': 3.044731378555298, 'timing/data': 0.016812879941426218, 'timing/model': 1.15266378398519, 'learning_rate/qwen_vl_interface': 6.723622250783745e-06, 'learning_rate/action_model': 6.723622250783744e-05, 'epoch': 1.31}) +[2026-06-19 14:26:52,595][starVLA.training.train_starvla][INFO] - Step 2059, Loss: {'train/loss': 0.21348504722118378, 'train/grad_norm_pre_clip': 5.962550163269043, 'timing/data': 0.017832892946898937, 'timing/model': 1.1511475739534944, 'learning_rate/qwen_vl_interface': 6.720726789312487e-06, 'learning_rate/action_model': 6.720726789312487e-05, 'epoch': 1.31}) +[2026-06-19 14:26:57,191][starVLA.training.train_starvla][INFO] - Step 2060, Loss: {'train/loss': 0.3288802206516266, 'train/grad_norm_pre_clip': 7.084887981414795, 'timing/data': 0.017042880062945187, 'timing/model': 1.1577165210619569, 'learning_rate/qwen_vl_interface': 6.7178307232810015e-06, 'learning_rate/action_model': 6.717830723281e-05, 'epoch': 1.31}) +[2026-06-19 14:27:01,772][starVLA.training.train_starvla][INFO] - Step 2061, Loss: {'train/loss': 0.28957265615463257, 'train/grad_norm_pre_clip': 8.140707015991211, 'timing/data': 0.018019246985204518, 'timing/model': 1.1471673339838162, 'learning_rate/qwen_vl_interface': 6.714934053879749e-06, 'learning_rate/action_model': 6.71493405387975e-05, 'epoch': 1.31}) +[2026-06-19 14:27:06,258][starVLA.training.train_starvla][INFO] - Step 2062, Loss: {'train/loss': 0.22430695593357086, 'train/grad_norm_pre_clip': 7.962959289550781, 'timing/data': 0.017333843978121877, 'timing/model': 1.1492614779854193, 'learning_rate/qwen_vl_interface': 6.712036782299442e-06, 'learning_rate/action_model': 6.712036782299443e-05, 'epoch': 1.31}) +[2026-06-19 14:27:10,842][starVLA.training.train_starvla][INFO] - Step 2063, Loss: {'train/loss': 0.21448367834091187, 'train/grad_norm_pre_clip': 7.919770240783691, 'timing/data': 0.01803942199330777, 'timing/model': 1.2916406349977478, 'learning_rate/qwen_vl_interface': 6.709138909731041e-06, 'learning_rate/action_model': 6.709138909731041e-05, 'epoch': 1.31}) +[2026-06-19 14:27:15,323][starVLA.training.train_starvla][INFO] - Step 2064, Loss: {'train/loss': 0.2254921793937683, 'train/grad_norm_pre_clip': 9.759237289428711, 'timing/data': 0.01755716698244214, 'timing/model': 1.1469832649454474, 'learning_rate/qwen_vl_interface': 6.706240437365751e-06, 'learning_rate/action_model': 6.70624043736575e-05, 'epoch': 1.31}) +[2026-06-19 14:27:19,926][starVLA.training.train_starvla][INFO] - Step 2065, Loss: {'train/loss': 0.18619723618030548, 'train/grad_norm_pre_clip': 14.67359447479248, 'timing/data': 0.02058211201801896, 'timing/model': 1.153997972025536, 'learning_rate/qwen_vl_interface': 6.703341366395023e-06, 'learning_rate/action_model': 6.703341366395023e-05, 'epoch': 1.31}) +[2026-06-19 14:27:24,500][starVLA.training.train_starvla][INFO] - Step 2066, Loss: {'train/loss': 0.1881849318742752, 'train/grad_norm_pre_clip': 11.026715278625488, 'timing/data': 0.017099608085118234, 'timing/model': 1.1420010860310867, 'learning_rate/qwen_vl_interface': 6.700441698010558e-06, 'learning_rate/action_model': 6.700441698010558e-05, 'epoch': 1.31}) +[2026-06-19 14:27:29,043][starVLA.training.train_starvla][INFO] - Step 2067, Loss: {'train/loss': 0.4028210937976837, 'train/grad_norm_pre_clip': 13.519518852233887, 'timing/data': 0.01917191396933049, 'timing/model': 1.1438995560165495, 'learning_rate/qwen_vl_interface': 6.697541433404299e-06, 'learning_rate/action_model': 6.697541433404299e-05, 'epoch': 1.31}) +[2026-06-19 14:27:33,637][starVLA.training.train_starvla][INFO] - Step 2068, Loss: {'train/loss': 0.1564052850008011, 'train/grad_norm_pre_clip': 5.742329120635986, 'timing/data': 0.017477205023169518, 'timing/model': 1.1426747960504144, 'learning_rate/qwen_vl_interface': 6.6946405737684325e-06, 'learning_rate/action_model': 6.694640573768433e-05, 'epoch': 1.31}) +[2026-06-19 14:27:38,234][starVLA.training.train_starvla][INFO] - Step 2069, Loss: {'train/loss': 0.5120100378990173, 'train/grad_norm_pre_clip': 14.371564865112305, 'timing/data': 0.01686672994401306, 'timing/model': 1.311761257937178, 'learning_rate/qwen_vl_interface': 6.691739120295397e-06, 'learning_rate/action_model': 6.691739120295397e-05, 'epoch': 1.31}) +[2026-06-19 14:27:42,661][starVLA.training.train_starvla][INFO] - Step 2070, Loss: {'train/loss': 0.2395099550485611, 'train/grad_norm_pre_clip': 6.710099220275879, 'timing/data': 0.017853736993856728, 'timing/model': 1.145204424043186, 'learning_rate/qwen_vl_interface': 6.688837074177868e-06, 'learning_rate/action_model': 6.688837074177868e-05, 'epoch': 1.31}) +[2026-06-19 14:27:47,272][starVLA.training.train_starvla][INFO] - Step 2071, Loss: {'train/loss': 0.10833465307950974, 'train/grad_norm_pre_clip': 5.032139301300049, 'timing/data': 0.016194690018892288, 'timing/model': 1.1443158630281687, 'learning_rate/qwen_vl_interface': 6.68593443660877e-06, 'learning_rate/action_model': 6.685934436608769e-05, 'epoch': 1.31}) +[2026-06-19 14:27:51,775][starVLA.training.train_starvla][INFO] - Step 2072, Loss: {'train/loss': 0.4042760133743286, 'train/grad_norm_pre_clip': 12.99601936340332, 'timing/data': 0.018009159015491605, 'timing/model': 1.1452900079311803, 'learning_rate/qwen_vl_interface': 6.6830312087812625e-06, 'learning_rate/action_model': 6.683031208781262e-05, 'epoch': 1.31}) +[2026-06-19 14:27:56,302][starVLA.training.train_starvla][INFO] - Step 2073, Loss: {'train/loss': 0.3190446197986603, 'train/grad_norm_pre_clip': 13.279924392700195, 'timing/data': 0.017980464966967702, 'timing/model': 1.1532009029760957, 'learning_rate/qwen_vl_interface': 6.680127391888754e-06, 'learning_rate/action_model': 6.680127391888754e-05, 'epoch': 1.31}) +[2026-06-19 14:28:00,907][starVLA.training.train_starvla][INFO] - Step 2074, Loss: {'train/loss': 0.24010446667671204, 'train/grad_norm_pre_clip': 7.698550224304199, 'timing/data': 0.018336078035645187, 'timing/model': 1.144755897927098, 'learning_rate/qwen_vl_interface': 6.677222987124897e-06, 'learning_rate/action_model': 6.677222987124897e-05, 'epoch': 1.32}) +[2026-06-19 14:28:05,554][starVLA.training.train_starvla][INFO] - Step 2075, Loss: {'train/loss': 0.12663403153419495, 'train/grad_norm_pre_clip': 4.861849784851074, 'timing/data': 0.017533948994241655, 'timing/model': 1.3521268599433824, 'learning_rate/qwen_vl_interface': 6.67431799568358e-06, 'learning_rate/action_model': 6.67431799568358e-05, 'epoch': 1.32}) +[2026-06-19 14:28:09,980][starVLA.training.train_starvla][INFO] - Step 2076, Loss: {'train/loss': 0.32893022894859314, 'train/grad_norm_pre_clip': 20.595088958740234, 'timing/data': 0.0179863820085302, 'timing/model': 1.151188511052169, 'learning_rate/qwen_vl_interface': 6.671412418758939e-06, 'learning_rate/action_model': 6.671412418758938e-05, 'epoch': 1.32}) +[2026-06-19 14:28:14,570][starVLA.training.train_starvla][INFO] - Step 2077, Loss: {'train/loss': 0.26377633213996887, 'train/grad_norm_pre_clip': 8.746724128723145, 'timing/data': 0.01891043095383793, 'timing/model': 1.1596685249824077, 'learning_rate/qwen_vl_interface': 6.668506257545341e-06, 'learning_rate/action_model': 6.668506257545341e-05, 'epoch': 1.32}) +[2026-06-19 14:28:19,176][starVLA.training.train_starvla][INFO] - Step 2078, Loss: {'train/loss': 0.36020857095718384, 'train/grad_norm_pre_clip': 12.563948631286621, 'timing/data': 0.01814748498145491, 'timing/model': 1.1513620929326862, 'learning_rate/qwen_vl_interface': 6.665599513237403e-06, 'learning_rate/action_model': 6.665599513237402e-05, 'epoch': 1.32}) +[2026-06-19 14:28:23,784][starVLA.training.train_starvla][INFO] - Step 2079, Loss: {'train/loss': 0.17112722992897034, 'train/grad_norm_pre_clip': 6.893033504486084, 'timing/data': 0.017154174973256886, 'timing/model': 1.1481061610393226, 'learning_rate/qwen_vl_interface': 6.662692187029978e-06, 'learning_rate/action_model': 6.662692187029978e-05, 'epoch': 1.32}) +[2026-06-19 14:28:28,411][starVLA.training.train_starvla][INFO] - Step 2080, Loss: {'train/loss': 0.1107485443353653, 'train/grad_norm_pre_clip': 5.269252300262451, 'timing/data': 0.01865167706273496, 'timing/model': 1.1641359099885449, 'learning_rate/qwen_vl_interface': 6.659784280118156e-06, 'learning_rate/action_model': 6.659784280118156e-05, 'epoch': 1.32}) +[2026-06-19 14:28:33,018][starVLA.training.train_starvla][INFO] - Step 2081, Loss: {'train/loss': 0.2940422594547272, 'train/grad_norm_pre_clip': 10.629714965820312, 'timing/data': 0.01829122193157673, 'timing/model': 1.320381238940172, 'learning_rate/qwen_vl_interface': 6.65687579369727e-06, 'learning_rate/action_model': 6.656875793697271e-05, 'epoch': 1.32}) +[2026-06-19 14:28:37,442][starVLA.training.train_starvla][INFO] - Step 2082, Loss: {'train/loss': 0.19855575263500214, 'train/grad_norm_pre_clip': 6.3858513832092285, 'timing/data': 0.018922816962003708, 'timing/model': 1.1453409940004349, 'learning_rate/qwen_vl_interface': 6.653966728962885e-06, 'learning_rate/action_model': 6.653966728962885e-05, 'epoch': 1.32}) +[2026-06-19 14:28:42,050][starVLA.training.train_starvla][INFO] - Step 2083, Loss: {'train/loss': 0.20516350865364075, 'train/grad_norm_pre_clip': 11.707620620727539, 'timing/data': 0.018612287007272243, 'timing/model': 1.1476027379976586, 'learning_rate/qwen_vl_interface': 6.651057087110813e-06, 'learning_rate/action_model': 6.651057087110813e-05, 'epoch': 1.32}) +[2026-06-19 14:28:46,672][starVLA.training.train_starvla][INFO] - Step 2084, Loss: {'train/loss': 0.19146190583705902, 'train/grad_norm_pre_clip': 7.6933794021606445, 'timing/data': 0.017906730994582176, 'timing/model': 1.1584247120190412, 'learning_rate/qwen_vl_interface': 6.6481468693370924e-06, 'learning_rate/action_model': 6.648146869337091e-05, 'epoch': 1.32}) +[2026-06-19 14:28:51,264][starVLA.training.train_starvla][INFO] - Step 2085, Loss: {'train/loss': 0.19316010177135468, 'train/grad_norm_pre_clip': 5.765833854675293, 'timing/data': 0.017584284069016576, 'timing/model': 1.1455352630000561, 'learning_rate/qwen_vl_interface': 6.645236076838009e-06, 'learning_rate/action_model': 6.645236076838009e-05, 'epoch': 1.32}) +[2026-06-19 14:28:55,864][starVLA.training.train_starvla][INFO] - Step 2086, Loss: {'train/loss': 0.16056755185127258, 'train/grad_norm_pre_clip': 6.608652114868164, 'timing/data': 0.017429943080060184, 'timing/model': 1.1460853139869869, 'learning_rate/qwen_vl_interface': 6.642324710810075e-06, 'learning_rate/action_model': 6.642324710810076e-05, 'epoch': 1.32}) +[2026-06-19 14:29:00,475][starVLA.training.train_starvla][INFO] - Step 2087, Loss: {'train/loss': 0.13344740867614746, 'train/grad_norm_pre_clip': 5.710254192352295, 'timing/data': 0.017218836001120508, 'timing/model': 1.334053992992267, 'learning_rate/qwen_vl_interface': 6.639412772450047e-06, 'learning_rate/action_model': 6.639412772450047e-05, 'epoch': 1.32}) +[2026-06-19 14:29:04,988][starVLA.training.train_starvla][INFO] - Step 2088, Loss: {'train/loss': 0.25339654088020325, 'train/grad_norm_pre_clip': 12.146929740905762, 'timing/data': 0.017894461983814836, 'timing/model': 1.225126474048011, 'learning_rate/qwen_vl_interface': 6.6365002629549104e-06, 'learning_rate/action_model': 6.63650026295491e-05, 'epoch': 1.32}) +[2026-06-19 14:29:09,427][starVLA.training.train_starvla][INFO] - Step 2089, Loss: {'train/loss': 0.18951593339443207, 'train/grad_norm_pre_clip': 6.951579570770264, 'timing/data': 0.01703953102696687, 'timing/model': 1.1565216149901971, 'learning_rate/qwen_vl_interface': 6.633587183521887e-06, 'learning_rate/action_model': 6.633587183521887e-05, 'epoch': 1.32}) +[2026-06-19 14:29:13,963][starVLA.training.train_starvla][INFO] - Step 2090, Loss: {'train/loss': 0.06649806350469589, 'train/grad_norm_pre_clip': 4.90516996383667, 'timing/data': 0.017180587048642337, 'timing/model': 1.148971420014277, 'learning_rate/qwen_vl_interface': 6.6306735353484365e-06, 'learning_rate/action_model': 6.630673535348436e-05, 'epoch': 1.33}) +[2026-06-19 14:29:18,556][starVLA.training.train_starvla][INFO] - Step 2091, Loss: {'train/loss': 0.15730386972427368, 'train/grad_norm_pre_clip': 33.8982048034668, 'timing/data': 0.01784188498277217, 'timing/model': 1.1470910230418667, 'learning_rate/qwen_vl_interface': 6.627759319632246e-06, 'learning_rate/action_model': 6.627759319632247e-05, 'epoch': 1.33}) +[2026-06-19 14:29:23,152][starVLA.training.train_starvla][INFO] - Step 2092, Loss: {'train/loss': 0.148196741938591, 'train/grad_norm_pre_clip': 10.504704475402832, 'timing/data': 0.017838437110185623, 'timing/model': 1.1451116210082546, 'learning_rate/qwen_vl_interface': 6.624844537571245e-06, 'learning_rate/action_model': 6.624844537571245e-05, 'epoch': 1.33}) +[2026-06-19 14:29:27,745][starVLA.training.train_starvla][INFO] - Step 2093, Loss: {'train/loss': 0.3194689452648163, 'train/grad_norm_pre_clip': 15.600640296936035, 'timing/data': 0.017526170006021857, 'timing/model': 1.3171152269933373, 'learning_rate/qwen_vl_interface': 6.621929190363585e-06, 'learning_rate/action_model': 6.621929190363585e-05, 'epoch': 1.33}) +[2026-06-19 14:29:32,191][starVLA.training.train_starvla][INFO] - Step 2094, Loss: {'train/loss': 0.2176932692527771, 'train/grad_norm_pre_clip': 22.82190704345703, 'timing/data': 0.018468008958734572, 'timing/model': 1.1578067129012197, 'learning_rate/qwen_vl_interface': 6.619013279207657e-06, 'learning_rate/action_model': 6.619013279207656e-05, 'epoch': 1.33}) +[2026-06-19 14:29:36,790][starVLA.training.train_starvla][INFO] - Step 2095, Loss: {'train/loss': 0.24479065835475922, 'train/grad_norm_pre_clip': 7.726989269256592, 'timing/data': 0.016239980002865195, 'timing/model': 1.1488657030276954, 'learning_rate/qwen_vl_interface': 6.616096805302083e-06, 'learning_rate/action_model': 6.616096805302082e-05, 'epoch': 1.33}) +[2026-06-19 14:29:41,413][starVLA.training.train_starvla][INFO] - Step 2096, Loss: {'train/loss': 0.3462689220905304, 'train/grad_norm_pre_clip': 11.03653335571289, 'timing/data': 0.0174456259701401, 'timing/model': 1.1486354150110856, 'learning_rate/qwen_vl_interface': 6.6131797698457135e-06, 'learning_rate/action_model': 6.613179769845713e-05, 'epoch': 1.33}) +[2026-06-19 14:29:45,921][starVLA.training.train_starvla][INFO] - Step 2097, Loss: {'train/loss': 0.26313576102256775, 'train/grad_norm_pre_clip': 8.533023834228516, 'timing/data': 0.01685529702808708, 'timing/model': 1.1420108031015843, 'learning_rate/qwen_vl_interface': 6.610262174037633e-06, 'learning_rate/action_model': 6.610262174037633e-05, 'epoch': 1.33}) +[2026-06-19 14:29:50,495][starVLA.training.train_starvla][INFO] - Step 2098, Loss: {'train/loss': 0.20800814032554626, 'train/grad_norm_pre_clip': 6.297099590301514, 'timing/data': 0.016616484965197742, 'timing/model': 1.1453773679677397, 'learning_rate/qwen_vl_interface': 6.607344019077155e-06, 'learning_rate/action_model': 6.607344019077156e-05, 'epoch': 1.33}) +[2026-06-19 14:29:55,085][starVLA.training.train_starvla][INFO] - Step 2099, Loss: {'train/loss': 0.22855150699615479, 'train/grad_norm_pre_clip': 8.629525184631348, 'timing/data': 0.01633585907984525, 'timing/model': 1.3159835980040953, 'learning_rate/qwen_vl_interface': 6.604425306163822e-06, 'learning_rate/action_model': 6.604425306163822e-05, 'epoch': 1.33}) +[2026-06-19 14:29:59,496][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 2100 +[2026-06-19 14:32:52,666][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 2100 +[2026-06-19 14:32:52,667][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 2100 +[2026-06-19 14:36:00,013][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 2100 +[2026-06-19 14:36:00,014][starVLA.training.train_starvla][INFO] - Step 2100, Loss: {'train/loss': 0.11339154094457626, 'train/grad_norm_pre_clip': 5.035484313964844, 'eval/loss': 0.31408703327178955, 'eval/latency_0/loss': 0.10274598002433777, 'eval/latency_1/loss': 0.08825726062059402, 'eval/latency_2/loss': 0.13424190878868103, 'eval/latency_3/loss': 0.49488988518714905, 'eval/latency_4/loss': 0.7647752165794373, 'eval/flappy/loss': 0.31408703327178955, 'eval/flappy/latency_0/loss': 0.10274598002433777, 'eval/flappy/latency_1/loss': 0.08825726062059402, 'eval/flappy/latency_2/loss': 0.13424190878868103, 'eval/flappy/latency_3/loss': 0.49488988518714905, 'eval/flappy/latency_4/loss': 0.7647752165794373, 'eval/flap_precision': 0.7655502392344498, 'eval/flap_recall': 0.5818181818181818, 'eval/flap_f1': 0.6611570247933884, 'eval/flap_rate_model': 0.058055555555555555, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.6611570247933884, 'eval/latency_0/flap_precision': 0.7655502392344498, 'eval/latency_0/flap_recall': 0.5818181818181818, 'eval/latency_0/flap_f1': 0.6611570247933884, 'eval/latency_0/flap_rate_model': 0.058055555555555555, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.6611570247933884, 'timing/data': 0.017610852955840528, 'timing/model': 1.143653071951121, 'learning_rate/qwen_vl_interface': 6.6015060364974045e-06, 'learning_rate/action_model': 6.601506036497404e-05, 'epoch': 1.33}) +[2026-06-19 14:36:00,014][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_2100_state +[2026-06-19 14:36:00,017][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-19 14:36:03,391][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_2100_state/model.safetensors +[2026-06-19 14:36:10,602][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_2100_state/optimizer.bin +[2026-06-19 14:36:10,602][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_2100_state/sampler.bin +[2026-06-19 14:36:10,602][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_2100_state/sampler_1.bin +[2026-06-19 14:36:10,603][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_2100_state/random_states_0.pkl +[2026-06-19 14:36:10,627][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2100 +[2026-06-19 14:36:10,627][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-19 14:36:10,666][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-19 14:36:15,200][starVLA.training.train_starvla][INFO] - Step 2101, Loss: {'train/loss': 0.18976068496704102, 'train/grad_norm_pre_clip': 13.183673858642578, 'timing/data': 0.016197446966543794, 'timing/model': 1.125950830988586, 'learning_rate/qwen_vl_interface': 6.598586211277907e-06, 'learning_rate/action_model': 6.598586211277907e-05, 'epoch': 1.33}) +[2026-06-19 14:36:19,725][starVLA.training.train_starvla][INFO] - Step 2102, Loss: {'train/loss': 0.1740942746400833, 'train/grad_norm_pre_clip': 7.841700077056885, 'timing/data': 0.016642038011923432, 'timing/model': 1.2993423910811543, 'learning_rate/qwen_vl_interface': 6.59566583170556e-06, 'learning_rate/action_model': 6.59566583170556e-05, 'epoch': 1.33}) +[2026-06-19 14:36:24,094][starVLA.training.train_starvla][INFO] - Step 2103, Loss: {'train/loss': 0.15657564997673035, 'train/grad_norm_pre_clip': 14.194109916687012, 'timing/data': 0.017005007015541196, 'timing/model': 1.1326053759548813, 'learning_rate/qwen_vl_interface': 6.59274489898082e-06, 'learning_rate/action_model': 6.592744898980819e-05, 'epoch': 1.33}) +[2026-06-19 14:36:28,615][starVLA.training.train_starvla][INFO] - Step 2104, Loss: {'train/loss': 0.1660270392894745, 'train/grad_norm_pre_clip': 10.603496551513672, 'timing/data': 0.01609777205158025, 'timing/model': 1.1307516440283507, 'learning_rate/qwen_vl_interface': 6.5898234143043716e-06, 'learning_rate/action_model': 6.589823414304371e-05, 'epoch': 1.33}) +[2026-06-19 14:36:33,136][starVLA.training.train_starvla][INFO] - Step 2105, Loss: {'train/loss': 0.16520558297634125, 'train/grad_norm_pre_clip': 9.273212432861328, 'timing/data': 0.017663034959696233, 'timing/model': 1.1279192980146036, 'learning_rate/qwen_vl_interface': 6.5869013788771264e-06, 'learning_rate/action_model': 6.586901378877126e-05, 'epoch': 1.33}) +[2026-06-19 14:36:37,709][starVLA.training.train_starvla][INFO] - Step 2106, Loss: {'train/loss': 0.22967027127742767, 'train/grad_norm_pre_clip': 9.702545166015625, 'timing/data': 0.018299655057489872, 'timing/model': 1.1511452740523964, 'learning_rate/qwen_vl_interface': 6.583978793900223e-06, 'learning_rate/action_model': 6.583978793900223e-05, 'epoch': 1.34}) +[2026-06-19 14:36:42,242][starVLA.training.train_starvla][INFO] - Step 2107, Loss: {'train/loss': 0.23219186067581177, 'train/grad_norm_pre_clip': 15.222177505493164, 'timing/data': 0.017548057017847896, 'timing/model': 1.1324573999736458, 'learning_rate/qwen_vl_interface': 6.581055660575028e-06, 'learning_rate/action_model': 6.581055660575027e-05, 'epoch': 1.34}) +[2026-06-19 14:36:46,783][starVLA.training.train_starvla][INFO] - Step 2108, Loss: {'train/loss': 0.265267550945282, 'train/grad_norm_pre_clip': 11.537802696228027, 'timing/data': 0.0185217090183869, 'timing/model': 1.297312569920905, 'learning_rate/qwen_vl_interface': 6.578131980103128e-06, 'learning_rate/action_model': 6.578131980103128e-05, 'epoch': 1.34}) +[2026-06-19 14:36:51,231][starVLA.training.train_starvla][INFO] - Step 2109, Loss: {'train/loss': 0.17893554270267487, 'train/grad_norm_pre_clip': 11.931608200073242, 'timing/data': 0.018814292969182134, 'timing/model': 1.2040231390856206, 'learning_rate/qwen_vl_interface': 6.5752077536863404e-06, 'learning_rate/action_model': 6.57520775368634e-05, 'epoch': 1.34}) +[2026-06-19 14:36:55,615][starVLA.training.train_starvla][INFO] - Step 2110, Loss: {'train/loss': 0.27354177832603455, 'train/grad_norm_pre_clip': 17.280691146850586, 'timing/data': 0.01958967710379511, 'timing/model': 1.1373188809957355, 'learning_rate/qwen_vl_interface': 6.572282982526702e-06, 'learning_rate/action_model': 6.572282982526702e-05, 'epoch': 1.34}) +[2026-06-19 14:37:00,101][starVLA.training.train_starvla][INFO] - Step 2111, Loss: {'train/loss': 0.27459320425987244, 'train/grad_norm_pre_clip': 14.724872589111328, 'timing/data': 0.017974963993765414, 'timing/model': 1.1350482490379363, 'learning_rate/qwen_vl_interface': 6.5693576678264755e-06, 'learning_rate/action_model': 6.569357667826476e-05, 'epoch': 1.34}) +[2026-06-19 14:37:04,680][starVLA.training.train_starvla][INFO] - Step 2112, Loss: {'train/loss': 0.19292522966861725, 'train/grad_norm_pre_clip': 14.07235336303711, 'timing/data': 0.022140711080282927, 'timing/model': 1.144469901919365, 'learning_rate/qwen_vl_interface': 6.566431810788148e-06, 'learning_rate/action_model': 6.566431810788148e-05, 'epoch': 1.34}) +[2026-06-19 14:37:09,252][starVLA.training.train_starvla][INFO] - Step 2113, Loss: {'train/loss': 0.1387312412261963, 'train/grad_norm_pre_clip': 8.556032180786133, 'timing/data': 0.017996008042246103, 'timing/model': 1.1375007919268683, 'learning_rate/qwen_vl_interface': 6.5635054126144306e-06, 'learning_rate/action_model': 6.56350541261443e-05, 'epoch': 1.34}) +[2026-06-19 14:37:13,821][starVLA.training.train_starvla][INFO] - Step 2114, Loss: {'train/loss': 0.2335117608308792, 'train/grad_norm_pre_clip': 9.587299346923828, 'timing/data': 0.01699249807279557, 'timing/model': 1.325314408983104, 'learning_rate/qwen_vl_interface': 6.560578474508254e-06, 'learning_rate/action_model': 6.560578474508253e-05, 'epoch': 1.34}) +[2026-06-19 14:37:18,217][starVLA.training.train_starvla][INFO] - Step 2115, Loss: {'train/loss': 0.3140580654144287, 'train/grad_norm_pre_clip': 9.309991836547852, 'timing/data': 0.01642774895299226, 'timing/model': 1.1381931570358574, 'learning_rate/qwen_vl_interface': 6.5576509976727685e-06, 'learning_rate/action_model': 6.557650997672769e-05, 'epoch': 1.34}) +[2026-06-19 14:37:22,786][starVLA.training.train_starvla][INFO] - Step 2116, Loss: {'train/loss': 0.23891925811767578, 'train/grad_norm_pre_clip': 12.291512489318848, 'timing/data': 0.017946533975191414, 'timing/model': 1.1384616689756513, 'learning_rate/qwen_vl_interface': 6.554722983311354e-06, 'learning_rate/action_model': 6.554722983311355e-05, 'epoch': 1.34}) +[2026-06-19 14:37:27,344][starVLA.training.train_starvla][INFO] - Step 2117, Loss: {'train/loss': 0.47275158762931824, 'train/grad_norm_pre_clip': 11.469931602478027, 'timing/data': 0.017915874952450395, 'timing/model': 1.1392588970484212, 'learning_rate/qwen_vl_interface': 6.551794432627605e-06, 'learning_rate/action_model': 6.551794432627605e-05, 'epoch': 1.34}) +[2026-06-19 14:37:31,823][starVLA.training.train_starvla][INFO] - Step 2118, Loss: {'train/loss': 0.15563061833381653, 'train/grad_norm_pre_clip': 8.09673023223877, 'timing/data': 0.016580611001700163, 'timing/model': 1.140804786933586, 'learning_rate/qwen_vl_interface': 6.548865346825338e-06, 'learning_rate/action_model': 6.548865346825338e-05, 'epoch': 1.34}) +[2026-06-19 14:37:36,356][starVLA.training.train_starvla][INFO] - Step 2119, Loss: {'train/loss': 0.24686315655708313, 'train/grad_norm_pre_clip': 10.534345626831055, 'timing/data': 0.016756923985667527, 'timing/model': 1.1377487629652023, 'learning_rate/qwen_vl_interface': 6.545935727108589e-06, 'learning_rate/action_model': 6.545935727108588e-05, 'epoch': 1.34}) +[2026-06-19 14:37:40,938][starVLA.training.train_starvla][INFO] - Step 2120, Loss: {'train/loss': 0.18271201848983765, 'train/grad_norm_pre_clip': 12.042542457580566, 'timing/data': 0.016019616974517703, 'timing/model': 1.3264533050823957, 'learning_rate/qwen_vl_interface': 6.543005574681613e-06, 'learning_rate/action_model': 6.543005574681613e-05, 'epoch': 1.34}) +[2026-06-19 14:37:45,355][starVLA.training.train_starvla][INFO] - Step 2121, Loss: {'train/loss': 0.1639237105846405, 'train/grad_norm_pre_clip': 8.538710594177246, 'timing/data': 0.01628680306021124, 'timing/model': 1.1400330719770864, 'learning_rate/qwen_vl_interface': 6.5400748907488886e-06, 'learning_rate/action_model': 6.540074890748888e-05, 'epoch': 1.35}) +[2026-06-19 14:37:49,918][starVLA.training.train_starvla][INFO] - Step 2122, Loss: {'train/loss': 0.1170988529920578, 'train/grad_norm_pre_clip': 6.931117057800293, 'timing/data': 0.01713385700713843, 'timing/model': 1.1404672940261662, 'learning_rate/qwen_vl_interface': 6.537143676515105e-06, 'learning_rate/action_model': 6.537143676515105e-05, 'epoch': 1.35}) +[2026-06-19 14:37:54,476][starVLA.training.train_starvla][INFO] - Step 2123, Loss: {'train/loss': 0.16358734667301178, 'train/grad_norm_pre_clip': 8.778925895690918, 'timing/data': 0.018304579076357186, 'timing/model': 1.1406673470046371, 'learning_rate/qwen_vl_interface': 6.534211933185176e-06, 'learning_rate/action_model': 6.534211933185176e-05, 'epoch': 1.35}) +[2026-06-19 14:37:59,036][starVLA.training.train_starvla][INFO] - Step 2124, Loss: {'train/loss': 0.2220253050327301, 'train/grad_norm_pre_clip': 10.752985000610352, 'timing/data': 0.018210484995506704, 'timing/model': 1.1381520109716803, 'learning_rate/qwen_vl_interface': 6.531279661964231e-06, 'learning_rate/action_model': 6.531279661964231e-05, 'epoch': 1.35}) +[2026-06-19 14:38:03,611][starVLA.training.train_starvla][INFO] - Step 2125, Loss: {'train/loss': 0.1402352899312973, 'train/grad_norm_pre_clip': 7.746518135070801, 'timing/data': 0.016338052926585078, 'timing/model': 1.1460324539802969, 'learning_rate/qwen_vl_interface': 6.528346864057616e-06, 'learning_rate/action_model': 6.528346864057615e-05, 'epoch': 1.35}) +[2026-06-19 14:38:08,172][starVLA.training.train_starvla][INFO] - Step 2126, Loss: {'train/loss': 0.3460131883621216, 'train/grad_norm_pre_clip': 9.905344009399414, 'timing/data': 0.016723661916330457, 'timing/model': 1.3059894799953327, 'learning_rate/qwen_vl_interface': 6.5254135406708906e-06, 'learning_rate/action_model': 6.52541354067089e-05, 'epoch': 1.35}) +[2026-06-19 14:38:12,600][starVLA.training.train_starvla][INFO] - Step 2127, Loss: {'train/loss': 0.1400640606880188, 'train/grad_norm_pre_clip': 7.952110767364502, 'timing/data': 0.017025458975695074, 'timing/model': 1.147696424042806, 'learning_rate/qwen_vl_interface': 6.522479693009836e-06, 'learning_rate/action_model': 6.522479693009837e-05, 'epoch': 1.35}) +[2026-06-19 14:38:17,193][starVLA.training.train_starvla][INFO] - Step 2128, Loss: {'train/loss': 0.29784977436065674, 'train/grad_norm_pre_clip': 12.365705490112305, 'timing/data': 0.018271515029482543, 'timing/model': 1.1428146639373153, 'learning_rate/qwen_vl_interface': 6.519545322280446e-06, 'learning_rate/action_model': 6.519545322280446e-05, 'epoch': 1.35}) +[2026-06-19 14:38:21,761][starVLA.training.train_starvla][INFO] - Step 2129, Loss: {'train/loss': 0.19509418308734894, 'train/grad_norm_pre_clip': 6.646749019622803, 'timing/data': 0.01857942808419466, 'timing/model': 1.1411953349597752, 'learning_rate/qwen_vl_interface': 6.5166104296889275e-06, 'learning_rate/action_model': 6.516610429688927e-05, 'epoch': 1.35}) +[2026-06-19 14:38:26,450][starVLA.training.train_starvla][INFO] - Step 2130, Loss: {'train/loss': 0.18038341403007507, 'train/grad_norm_pre_clip': 9.127530097961426, 'timing/data': 0.01849404803942889, 'timing/model': 1.1565378840314224, 'learning_rate/qwen_vl_interface': 6.513675016441707e-06, 'learning_rate/action_model': 6.513675016441707e-05, 'epoch': 1.35}) +[2026-06-19 14:38:31,057][starVLA.training.train_starvla][INFO] - Step 2131, Loss: {'train/loss': 0.25136053562164307, 'train/grad_norm_pre_clip': 10.216894149780273, 'timing/data': 0.01911076798569411, 'timing/model': 1.1397622850490734, 'learning_rate/qwen_vl_interface': 6.510739083745422e-06, 'learning_rate/action_model': 6.510739083745422e-05, 'epoch': 1.35}) +[2026-06-19 14:38:35,669][starVLA.training.train_starvla][INFO] - Step 2132, Loss: {'train/loss': 0.1203460693359375, 'train/grad_norm_pre_clip': 6.807374000549316, 'timing/data': 0.01747142698150128, 'timing/model': 1.321212434093468, 'learning_rate/qwen_vl_interface': 6.507802632806919e-06, 'learning_rate/action_model': 6.507802632806919e-05, 'epoch': 1.35}) +[2026-06-19 14:38:40,095][starVLA.training.train_starvla][INFO] - Step 2133, Loss: {'train/loss': 0.16817283630371094, 'train/grad_norm_pre_clip': 7.103250503540039, 'timing/data': 0.01911452307831496, 'timing/model': 1.1406273669563234, 'learning_rate/qwen_vl_interface': 6.5048656648332646e-06, 'learning_rate/action_model': 6.504865664833264e-05, 'epoch': 1.35}) +[2026-06-19 14:38:44,703][starVLA.training.train_starvla][INFO] - Step 2134, Loss: {'train/loss': 0.2030715048313141, 'train/grad_norm_pre_clip': 15.454164505004883, 'timing/data': 0.01693108503241092, 'timing/model': 1.1450822330079973, 'learning_rate/qwen_vl_interface': 6.501928181031737e-06, 'learning_rate/action_model': 6.501928181031737e-05, 'epoch': 1.35}) +[2026-06-19 14:38:49,314][starVLA.training.train_starvla][INFO] - Step 2135, Loss: {'train/loss': 0.16936282813549042, 'train/grad_norm_pre_clip': 8.915756225585938, 'timing/data': 0.01693901198450476, 'timing/model': 1.1424717030022293, 'learning_rate/qwen_vl_interface': 6.498990182609823e-06, 'learning_rate/action_model': 6.498990182609823e-05, 'epoch': 1.35}) +[2026-06-19 14:38:53,792][starVLA.training.train_starvla][INFO] - Step 2136, Loss: {'train/loss': 0.17825743556022644, 'train/grad_norm_pre_clip': 8.40835189819336, 'timing/data': 0.017403437988832593, 'timing/model': 1.1465601440286264, 'learning_rate/qwen_vl_interface': 6.4960516707752255e-06, 'learning_rate/action_model': 6.496051670775226e-05, 'epoch': 1.35}) +[2026-06-19 14:38:58,356][starVLA.training.train_starvla][INFO] - Step 2137, Loss: {'train/loss': 0.21219739317893982, 'train/grad_norm_pre_clip': 10.538419723510742, 'timing/data': 0.01768739300314337, 'timing/model': 1.1570942560210824, 'learning_rate/qwen_vl_interface': 6.493112646735851e-06, 'learning_rate/action_model': 6.493112646735851e-05, 'epoch': 1.36}) +[2026-06-19 14:39:02,961][starVLA.training.train_starvla][INFO] - Step 2138, Loss: {'train/loss': 0.17306740581989288, 'train/grad_norm_pre_clip': 11.429733276367188, 'timing/data': 0.018144406960345805, 'timing/model': 1.327064299955964, 'learning_rate/qwen_vl_interface': 6.490173111699824e-06, 'learning_rate/action_model': 6.490173111699825e-05, 'epoch': 1.36}) +[2026-06-19 14:39:07,379][starVLA.training.train_starvla][INFO] - Step 2139, Loss: {'train/loss': 0.18443270027637482, 'train/grad_norm_pre_clip': 6.733776092529297, 'timing/data': 0.017895507044158876, 'timing/model': 1.1525398509111255, 'learning_rate/qwen_vl_interface': 6.487233066875479e-06, 'learning_rate/action_model': 6.487233066875479e-05, 'epoch': 1.36}) +[2026-06-19 14:39:11,981][starVLA.training.train_starvla][INFO] - Step 2140, Loss: {'train/loss': 0.15171833336353302, 'train/grad_norm_pre_clip': 7.000601291656494, 'timing/data': 0.016883934964425862, 'timing/model': 1.1427176629658788, 'learning_rate/qwen_vl_interface': 6.484292513471352e-06, 'learning_rate/action_model': 6.484292513471352e-05, 'epoch': 1.36}) +[2026-06-19 14:39:16,579][starVLA.training.train_starvla][INFO] - Step 2141, Loss: {'train/loss': 0.2485656887292862, 'train/grad_norm_pre_clip': 13.114805221557617, 'timing/data': 0.017599666956812143, 'timing/model': 1.1501952249091119, 'learning_rate/qwen_vl_interface': 6.4813514526961975e-06, 'learning_rate/action_model': 6.481351452696198e-05, 'epoch': 1.36}) +[2026-06-19 14:39:21,066][starVLA.training.train_starvla][INFO] - Step 2142, Loss: {'train/loss': 0.3281056582927704, 'train/grad_norm_pre_clip': 18.05463409423828, 'timing/data': 0.017124365083873272, 'timing/model': 1.1471534800948575, 'learning_rate/qwen_vl_interface': 6.478409885758972e-06, 'learning_rate/action_model': 6.478409885758971e-05, 'epoch': 1.36}) +[2026-06-19 14:39:25,624][starVLA.training.train_starvla][INFO] - Step 2143, Loss: {'train/loss': 0.0939672589302063, 'train/grad_norm_pre_clip': 7.192903518676758, 'timing/data': 0.016841828008182347, 'timing/model': 1.1523365349275991, 'learning_rate/qwen_vl_interface': 6.475467813868844e-06, 'learning_rate/action_model': 6.475467813868844e-05, 'epoch': 1.36}) +[2026-06-19 14:39:30,234][starVLA.training.train_starvla][INFO] - Step 2144, Loss: {'train/loss': 0.15170824527740479, 'train/grad_norm_pre_clip': 12.696998596191406, 'timing/data': 0.01694215403404087, 'timing/model': 1.3295157050015405, 'learning_rate/qwen_vl_interface': 6.472525238235187e-06, 'learning_rate/action_model': 6.472525238235186e-05, 'epoch': 1.36}) +[2026-06-19 14:39:34,663][starVLA.training.train_starvla][INFO] - Step 2145, Loss: {'train/loss': 0.28532692790031433, 'train/grad_norm_pre_clip': 12.046965599060059, 'timing/data': 0.01777245698031038, 'timing/model': 1.1484459999483079, 'learning_rate/qwen_vl_interface': 6.4695821600675834e-06, 'learning_rate/action_model': 6.469582160067583e-05, 'epoch': 1.36}) +[2026-06-19 14:39:39,283][starVLA.training.train_starvla][INFO] - Step 2146, Loss: {'train/loss': 0.4021621346473694, 'train/grad_norm_pre_clip': 17.949857711791992, 'timing/data': 0.018822210025973618, 'timing/model': 1.1438527669524774, 'learning_rate/qwen_vl_interface': 6.466638580575824e-06, 'learning_rate/action_model': 6.466638580575824e-05, 'epoch': 1.36}) +[2026-06-19 14:39:43,904][starVLA.training.train_starvla][INFO] - Step 2147, Loss: {'train/loss': 0.17454245686531067, 'train/grad_norm_pre_clip': 11.854616165161133, 'timing/data': 0.01777873805258423, 'timing/model': 1.1464308960130438, 'learning_rate/qwen_vl_interface': 6.463694500969898e-06, 'learning_rate/action_model': 6.463694500969898e-05, 'epoch': 1.36}) +[2026-06-19 14:39:48,532][starVLA.training.train_starvla][INFO] - Step 2148, Loss: {'train/loss': 0.14028401672840118, 'train/grad_norm_pre_clip': 14.23000717163086, 'timing/data': 0.017782344948500395, 'timing/model': 1.1651359219104052, 'learning_rate/qwen_vl_interface': 6.460749922460008e-06, 'learning_rate/action_model': 6.460749922460008e-05, 'epoch': 1.36}) +[2026-06-19 14:39:53,037][starVLA.training.train_starvla][INFO] - Step 2149, Loss: {'train/loss': 0.2828715145587921, 'train/grad_norm_pre_clip': 15.659786224365234, 'timing/data': 0.01716650603339076, 'timing/model': 1.1448098249966279, 'learning_rate/qwen_vl_interface': 6.457804846256559e-06, 'learning_rate/action_model': 6.457804846256559e-05, 'epoch': 1.36}) +[2026-06-19 14:39:57,595][starVLA.training.train_starvla][INFO] - Step 2150, Loss: {'train/loss': 0.13214415311813354, 'train/grad_norm_pre_clip': 8.099647521972656, 'timing/data': 0.017281952081248164, 'timing/model': 1.274418073007837, 'learning_rate/qwen_vl_interface': 6.45485927357016e-06, 'learning_rate/action_model': 6.454859273570159e-05, 'epoch': 1.36}) +[2026-06-19 14:40:02,024][starVLA.training.train_starvla][INFO] - Step 2151, Loss: {'train/loss': 0.19063495099544525, 'train/grad_norm_pre_clip': 9.038339614868164, 'timing/data': 0.019009185023605824, 'timing/model': 1.1530105230631307, 'learning_rate/qwen_vl_interface': 6.451913205611626e-06, 'learning_rate/action_model': 6.451913205611626e-05, 'epoch': 1.36}) +[2026-06-19 14:40:06,606][starVLA.training.train_starvla][INFO] - Step 2152, Loss: {'train/loss': 0.47018107771873474, 'train/grad_norm_pre_clip': 22.243234634399414, 'timing/data': 0.017511181998997927, 'timing/model': 1.1413098040502518, 'learning_rate/qwen_vl_interface': 6.448966643591974e-06, 'learning_rate/action_model': 6.448966643591974e-05, 'epoch': 1.36}) +[2026-06-19 14:40:11,187][starVLA.training.train_starvla][INFO] - Step 2153, Loss: {'train/loss': 0.24204647541046143, 'train/grad_norm_pre_clip': 10.039839744567871, 'timing/data': 0.016754828044213355, 'timing/model': 1.1404762590536848, 'learning_rate/qwen_vl_interface': 6.446019588722423e-06, 'learning_rate/action_model': 6.446019588722422e-05, 'epoch': 1.37}) +[2026-06-19 14:40:15,758][starVLA.training.train_starvla][INFO] - Step 2154, Loss: {'train/loss': 0.1044972762465477, 'train/grad_norm_pre_clip': 7.454360008239746, 'timing/data': 0.017998215975239873, 'timing/model': 1.142907060100697, 'learning_rate/qwen_vl_interface': 6.443072042214398e-06, 'learning_rate/action_model': 6.443072042214397e-05, 'epoch': 1.37}) +[2026-06-19 14:40:20,335][starVLA.training.train_starvla][INFO] - Step 2155, Loss: {'train/loss': 0.05907360091805458, 'train/grad_norm_pre_clip': 7.297455787658691, 'timing/data': 0.01768548891413957, 'timing/model': 1.309604438021779, 'learning_rate/qwen_vl_interface': 6.440124005279521e-06, 'learning_rate/action_model': 6.440124005279521e-05, 'epoch': 1.37}) +[2026-06-19 14:40:24,826][starVLA.training.train_starvla][INFO] - Step 2156, Loss: {'train/loss': 0.22429649531841278, 'train/grad_norm_pre_clip': 9.790963172912598, 'timing/data': 0.017917378107085824, 'timing/model': 1.2139263880671933, 'learning_rate/qwen_vl_interface': 6.43717547912962e-06, 'learning_rate/action_model': 6.437175479129621e-05, 'epoch': 1.37}) +[2026-06-19 14:40:29,255][starVLA.training.train_starvla][INFO] - Step 2157, Loss: {'train/loss': 0.21160046756267548, 'train/grad_norm_pre_clip': 6.319003582000732, 'timing/data': 0.017362355021759868, 'timing/model': 1.1396563749294728, 'learning_rate/qwen_vl_interface': 6.434226464976725e-06, 'learning_rate/action_model': 6.434226464976724e-05, 'epoch': 1.37}) +[2026-06-19 14:40:33,779][starVLA.training.train_starvla][INFO] - Step 2158, Loss: {'train/loss': 0.46456265449523926, 'train/grad_norm_pre_clip': 9.121918678283691, 'timing/data': 0.017012706957757473, 'timing/model': 1.1430861870758235, 'learning_rate/qwen_vl_interface': 6.431276964033062e-06, 'learning_rate/action_model': 6.431276964033062e-05, 'epoch': 1.37}) +[2026-06-19 14:40:38,364][starVLA.training.train_starvla][INFO] - Step 2159, Loss: {'train/loss': 0.2315445989370346, 'train/grad_norm_pre_clip': 7.75072717666626, 'timing/data': 0.016748304944485426, 'timing/model': 1.152827191981487, 'learning_rate/qwen_vl_interface': 6.428326977511059e-06, 'learning_rate/action_model': 6.428326977511058e-05, 'epoch': 1.37}) +[2026-06-19 14:40:42,958][starVLA.training.train_starvla][INFO] - Step 2160, Loss: {'train/loss': 0.24631179869174957, 'train/grad_norm_pre_clip': 6.97604513168335, 'timing/data': 0.017176304943859577, 'timing/model': 1.1505882099736482, 'learning_rate/qwen_vl_interface': 6.425376506623344e-06, 'learning_rate/action_model': 6.425376506623344e-05, 'epoch': 1.37}) +[2026-06-19 14:40:47,561][starVLA.training.train_starvla][INFO] - Step 2161, Loss: {'train/loss': 0.3361288607120514, 'train/grad_norm_pre_clip': 16.942092895507812, 'timing/data': 0.016560437972657382, 'timing/model': 1.3256552710663527, 'learning_rate/qwen_vl_interface': 6.422425552582747e-06, 'learning_rate/action_model': 6.422425552582747e-05, 'epoch': 1.37}) +[2026-06-19 14:40:51,991][starVLA.training.train_starvla][INFO] - Step 2162, Loss: {'train/loss': 0.08785563707351685, 'train/grad_norm_pre_clip': 5.114480972290039, 'timing/data': 0.01718306599650532, 'timing/model': 1.1495669248979539, 'learning_rate/qwen_vl_interface': 6.41947411660229e-06, 'learning_rate/action_model': 6.41947411660229e-05, 'epoch': 1.37}) +[2026-06-19 14:40:56,568][starVLA.training.train_starvla][INFO] - Step 2163, Loss: {'train/loss': 0.21241746842861176, 'train/grad_norm_pre_clip': 9.400348663330078, 'timing/data': 0.01745696598663926, 'timing/model': 1.1435060889925808, 'learning_rate/qwen_vl_interface': 6.4165221998951995e-06, 'learning_rate/action_model': 6.4165221998952e-05, 'epoch': 1.37}) +[2026-06-19 14:41:01,048][starVLA.training.train_starvla][INFO] - Step 2164, Loss: {'train/loss': 0.0897173285484314, 'train/grad_norm_pre_clip': 4.895183086395264, 'timing/data': 0.018159614061005414, 'timing/model': 1.1402433349285275, 'learning_rate/qwen_vl_interface': 6.413569803674896e-06, 'learning_rate/action_model': 6.413569803674896e-05, 'epoch': 1.37}) +[2026-06-19 14:41:05,552][starVLA.training.train_starvla][INFO] - Step 2165, Loss: {'train/loss': 0.10238892585039139, 'train/grad_norm_pre_clip': 4.492772579193115, 'timing/data': 0.018096021958626807, 'timing/model': 1.1414404310053214, 'learning_rate/qwen_vl_interface': 6.410616929154997e-06, 'learning_rate/action_model': 6.410616929154996e-05, 'epoch': 1.37}) +[2026-06-19 14:41:10,157][starVLA.training.train_starvla][INFO] - Step 2166, Loss: {'train/loss': 0.2727612853050232, 'train/grad_norm_pre_clip': 8.388066291809082, 'timing/data': 0.019056426011957228, 'timing/model': 1.1538619720377028, 'learning_rate/qwen_vl_interface': 6.40766357754932e-06, 'learning_rate/action_model': 6.407663577549319e-05, 'epoch': 1.37}) +[2026-06-19 14:41:14,757][starVLA.training.train_starvla][INFO] - Step 2167, Loss: {'train/loss': 0.18218761682510376, 'train/grad_norm_pre_clip': 11.990259170532227, 'timing/data': 0.023305818089284003, 'timing/model': 1.31994011788629, 'learning_rate/qwen_vl_interface': 6.404709750071875e-06, 'learning_rate/action_model': 6.404709750071874e-05, 'epoch': 1.37}) +[2026-06-19 14:41:19,180][starVLA.training.train_starvla][INFO] - Step 2168, Loss: {'train/loss': 0.32026049494743347, 'train/grad_norm_pre_clip': 8.494701385498047, 'timing/data': 0.018959514913149178, 'timing/model': 1.1455233009764925, 'learning_rate/qwen_vl_interface': 6.401755447936869e-06, 'learning_rate/action_model': 6.401755447936869e-05, 'epoch': 1.37}) +[2026-06-19 14:41:23,779][starVLA.training.train_starvla][INFO] - Step 2169, Loss: {'train/loss': 0.25240543484687805, 'train/grad_norm_pre_clip': 7.766949653625488, 'timing/data': 0.017024325905367732, 'timing/model': 1.1538789159385487, 'learning_rate/qwen_vl_interface': 6.3988006723587036e-06, 'learning_rate/action_model': 6.398800672358703e-05, 'epoch': 1.38}) +[2026-06-19 14:41:28,359][starVLA.training.train_starvla][INFO] - Step 2170, Loss: {'train/loss': 0.1822778284549713, 'train/grad_norm_pre_clip': 7.28754186630249, 'timing/data': 0.01883395598269999, 'timing/model': 1.1416612539906055, 'learning_rate/qwen_vl_interface': 6.395845424551975e-06, 'learning_rate/action_model': 6.395845424551975e-05, 'epoch': 1.38}) +[2026-06-19 14:41:32,964][starVLA.training.train_starvla][INFO] - Step 2171, Loss: {'train/loss': 0.3640455901622772, 'train/grad_norm_pre_clip': 8.63847827911377, 'timing/data': 0.01796864508651197, 'timing/model': 1.1500545719172806, 'learning_rate/qwen_vl_interface': 6.392889705731475e-06, 'learning_rate/action_model': 6.392889705731476e-05, 'epoch': 1.38}) +[2026-06-19 14:41:37,542][starVLA.training.train_starvla][INFO] - Step 2172, Loss: {'train/loss': 0.3565713167190552, 'train/grad_norm_pre_clip': 10.499635696411133, 'timing/data': 0.01751434907782823, 'timing/model': 1.1435960680246353, 'learning_rate/qwen_vl_interface': 6.389933517112188e-06, 'learning_rate/action_model': 6.389933517112187e-05, 'epoch': 1.38}) +[2026-06-19 14:41:42,134][starVLA.training.train_starvla][INFO] - Step 2173, Loss: {'train/loss': 0.2280987650156021, 'train/grad_norm_pre_clip': 10.312883377075195, 'timing/data': 0.017179178074002266, 'timing/model': 1.3098945610690862, 'learning_rate/qwen_vl_interface': 6.386976859909293e-06, 'learning_rate/action_model': 6.386976859909292e-05, 'epoch': 1.38}) +[2026-06-19 14:41:46,639][starVLA.training.train_starvla][INFO] - Step 2174, Loss: {'train/loss': 0.1646401584148407, 'train/grad_norm_pre_clip': 9.44753360748291, 'timing/data': 0.017695770016871393, 'timing/model': 1.2307210250291973, 'learning_rate/qwen_vl_interface': 6.384019735338156e-06, 'learning_rate/action_model': 6.384019735338156e-05, 'epoch': 1.38}) +[2026-06-19 14:41:51,056][starVLA.training.train_starvla][INFO] - Step 2175, Loss: {'train/loss': 0.1644291877746582, 'train/grad_norm_pre_clip': 8.431808471679688, 'timing/data': 0.0176172029459849, 'timing/model': 1.146275995997712, 'learning_rate/qwen_vl_interface': 6.381062144614342e-06, 'learning_rate/action_model': 6.381062144614342e-05, 'epoch': 1.38}) +[2026-06-19 14:41:55,645][starVLA.training.train_starvla][INFO] - Step 2176, Loss: {'train/loss': 0.18868830800056458, 'train/grad_norm_pre_clip': 10.510809898376465, 'timing/data': 0.017301367013715208, 'timing/model': 1.1429823539219797, 'learning_rate/qwen_vl_interface': 6.378104088953605e-06, 'learning_rate/action_model': 6.378104088953605e-05, 'epoch': 1.38}) +[2026-06-19 14:42:00,238][starVLA.training.train_starvla][INFO] - Step 2177, Loss: {'train/loss': 0.24173802137374878, 'train/grad_norm_pre_clip': 10.773004531860352, 'timing/data': 0.017949581961147487, 'timing/model': 1.147636351059191, 'learning_rate/qwen_vl_interface': 6.375145569571889e-06, 'learning_rate/action_model': 6.375145569571889e-05, 'epoch': 1.38}) +[2026-06-19 14:42:04,833][starVLA.training.train_starvla][INFO] - Step 2178, Loss: {'train/loss': 0.21217861771583557, 'train/grad_norm_pre_clip': 8.933879852294922, 'timing/data': 0.020036395988427103, 'timing/model': 1.149029220920056, 'learning_rate/qwen_vl_interface': 6.372186587685329e-06, 'learning_rate/action_model': 6.372186587685329e-05, 'epoch': 1.38}) +[2026-06-19 14:42:09,419][starVLA.training.train_starvla][INFO] - Step 2179, Loss: {'train/loss': 0.12149001657962799, 'train/grad_norm_pre_clip': 5.391472339630127, 'timing/data': 0.018505513900890946, 'timing/model': 1.3088630009442568, 'learning_rate/qwen_vl_interface': 6.369227144510252e-06, 'learning_rate/action_model': 6.369227144510252e-05, 'epoch': 1.38}) +[2026-06-19 14:42:13,936][starVLA.training.train_starvla][INFO] - Step 2180, Loss: {'train/loss': 0.20936928689479828, 'train/grad_norm_pre_clip': 9.984774589538574, 'timing/data': 0.018845028011128306, 'timing/model': 1.2383906310424209, 'learning_rate/qwen_vl_interface': 6.3662672412631736e-06, 'learning_rate/action_model': 6.366267241263173e-05, 'epoch': 1.38}) +[2026-06-19 14:42:18,354][starVLA.training.train_starvla][INFO] - Step 2181, Loss: {'train/loss': 0.23383402824401855, 'train/grad_norm_pre_clip': 8.3810396194458, 'timing/data': 0.017324959975667298, 'timing/model': 1.1462458509486169, 'learning_rate/qwen_vl_interface': 6.3633068791607954e-06, 'learning_rate/action_model': 6.363306879160795e-05, 'epoch': 1.38}) +[2026-06-19 14:42:22,945][starVLA.training.train_starvla][INFO] - Step 2182, Loss: {'train/loss': 0.21182547509670258, 'train/grad_norm_pre_clip': 9.15169906616211, 'timing/data': 0.017287799040786922, 'timing/model': 1.1415024319430813, 'learning_rate/qwen_vl_interface': 6.360346059420014e-06, 'learning_rate/action_model': 6.360346059420014e-05, 'epoch': 1.38}) +[2026-06-19 14:42:27,540][starVLA.training.train_starvla][INFO] - Step 2183, Loss: {'train/loss': 0.16860896348953247, 'train/grad_norm_pre_clip': 8.519698143005371, 'timing/data': 0.0176755350548774, 'timing/model': 1.1435546710854396, 'learning_rate/qwen_vl_interface': 6.357384783257909e-06, 'learning_rate/action_model': 6.357384783257908e-05, 'epoch': 1.38}) +[2026-06-19 14:42:32,111][starVLA.training.train_starvla][INFO] - Step 2184, Loss: {'train/loss': 0.38867583870887756, 'train/grad_norm_pre_clip': 13.288172721862793, 'timing/data': 0.01841084100306034, 'timing/model': 1.1470803490374237, 'learning_rate/qwen_vl_interface': 6.354423051891748e-06, 'learning_rate/action_model': 6.354423051891749e-05, 'epoch': 1.39}) +[2026-06-19 14:42:36,710][starVLA.training.train_starvla][INFO] - Step 2185, Loss: {'train/loss': 0.28746649622917175, 'train/grad_norm_pre_clip': 9.370922088623047, 'timing/data': 0.018699271022342145, 'timing/model': 1.3241117720026523, 'learning_rate/qwen_vl_interface': 6.35146086653899e-06, 'learning_rate/action_model': 6.35146086653899e-05, 'epoch': 1.39}) +[2026-06-19 14:42:41,139][starVLA.training.train_starvla][INFO] - Step 2186, Loss: {'train/loss': 0.19470807909965515, 'train/grad_norm_pre_clip': 8.669503211975098, 'timing/data': 0.018486983026377857, 'timing/model': 1.1459699949482456, 'learning_rate/qwen_vl_interface': 6.348498228417273e-06, 'learning_rate/action_model': 6.348498228417273e-05, 'epoch': 1.39}) +[2026-06-19 14:42:45,720][starVLA.training.train_starvla][INFO] - Step 2187, Loss: {'train/loss': 0.07517752796411514, 'train/grad_norm_pre_clip': 4.961874485015869, 'timing/data': 0.01674288499634713, 'timing/model': 1.145216473028995, 'learning_rate/qwen_vl_interface': 6.34553513874443e-06, 'learning_rate/action_model': 6.34553513874443e-05, 'epoch': 1.39}) +[2026-06-19 14:42:50,211][starVLA.training.train_starvla][INFO] - Step 2188, Loss: {'train/loss': 0.3129101097583771, 'train/grad_norm_pre_clip': 9.571678161621094, 'timing/data': 0.018488488043658435, 'timing/model': 1.1470070520881563, 'learning_rate/qwen_vl_interface': 6.342571598738473e-06, 'learning_rate/action_model': 6.342571598738472e-05, 'epoch': 1.39}) +[2026-06-19 14:42:54,719][starVLA.training.train_starvla][INFO] - Step 2189, Loss: {'train/loss': 0.3154645562171936, 'train/grad_norm_pre_clip': 13.284111022949219, 'timing/data': 0.018650084966793656, 'timing/model': 1.153088531922549, 'learning_rate/qwen_vl_interface': 6.339607609617603e-06, 'learning_rate/action_model': 6.339607609617603e-05, 'epoch': 1.39}) +[2026-06-19 14:42:59,242][starVLA.training.train_starvla][INFO] - Step 2190, Loss: {'train/loss': 0.10332044214010239, 'train/grad_norm_pre_clip': 9.254790306091309, 'timing/data': 0.018312279018573463, 'timing/model': 1.1427719349740073, 'learning_rate/qwen_vl_interface': 6.336643172600202e-06, 'learning_rate/action_model': 6.336643172600202e-05, 'epoch': 1.39}) +[2026-06-19 14:43:03,842][starVLA.training.train_starvla][INFO] - Step 2191, Loss: {'train/loss': 0.15728923678398132, 'train/grad_norm_pre_clip': 6.59675407409668, 'timing/data': 0.018104276969097555, 'timing/model': 1.3119428349891677, 'learning_rate/qwen_vl_interface': 6.333678288904838e-06, 'learning_rate/action_model': 6.333678288904837e-05, 'epoch': 1.39}) +[2026-06-19 14:43:08,264][starVLA.training.train_starvla][INFO] - Step 2192, Loss: {'train/loss': 0.15469129383563995, 'train/grad_norm_pre_clip': 6.520567417144775, 'timing/data': 0.01702698797453195, 'timing/model': 1.1484152630437165, 'learning_rate/qwen_vl_interface': 6.330712959750264e-06, 'learning_rate/action_model': 6.330712959750263e-05, 'epoch': 1.39}) +[2026-06-19 14:43:12,835][starVLA.training.train_starvla][INFO] - Step 2193, Loss: {'train/loss': 0.1972053050994873, 'train/grad_norm_pre_clip': 7.15557336807251, 'timing/data': 0.018066665972582996, 'timing/model': 1.1528363020624965, 'learning_rate/qwen_vl_interface': 6.3277471863554145e-06, 'learning_rate/action_model': 6.327747186355414e-05, 'epoch': 1.39}) +[2026-06-19 14:43:17,460][starVLA.training.train_starvla][INFO] - Step 2194, Loss: {'train/loss': 0.1844935268163681, 'train/grad_norm_pre_clip': 5.2476325035095215, 'timing/data': 0.017232408979907632, 'timing/model': 1.1642808209871873, 'learning_rate/qwen_vl_interface': 6.324780969939406e-06, 'learning_rate/action_model': 6.324780969939405e-05, 'epoch': 1.39}) +[2026-06-19 14:43:22,049][starVLA.training.train_starvla][INFO] - Step 2195, Loss: {'train/loss': 0.2535094916820526, 'train/grad_norm_pre_clip': 11.449816703796387, 'timing/data': 0.016526652965694666, 'timing/model': 1.1450166209833696, 'learning_rate/qwen_vl_interface': 6.32181431172154e-06, 'learning_rate/action_model': 6.32181431172154e-05, 'epoch': 1.39}) +[2026-06-19 14:43:26,651][starVLA.training.train_starvla][INFO] - Step 2196, Loss: {'train/loss': 0.17252692580223083, 'train/grad_norm_pre_clip': 5.898573875427246, 'timing/data': 0.01732104504480958, 'timing/model': 1.3243590160273015, 'learning_rate/qwen_vl_interface': 6.318847212921295e-06, 'learning_rate/action_model': 6.318847212921295e-05, 'epoch': 1.39}) +[2026-06-19 14:43:31,246][starVLA.training.train_starvla][INFO] - Step 2197, Loss: {'train/loss': 0.4241504967212677, 'train/grad_norm_pre_clip': 55.6934814453125, 'timing/data': 0.018058441928587854, 'timing/model': 1.3019118428928778, 'learning_rate/qwen_vl_interface': 6.3158796747583364e-06, 'learning_rate/action_model': 6.315879674758336e-05, 'epoch': 1.39}) +[2026-06-19 14:43:35,657][starVLA.training.train_starvla][INFO] - Step 2198, Loss: {'train/loss': 0.2674611210823059, 'train/grad_norm_pre_clip': 10.883139610290527, 'timing/data': 0.018760218983516097, 'timing/model': 1.1435817739693448, 'learning_rate/qwen_vl_interface': 6.312911698452503e-06, 'learning_rate/action_model': 6.312911698452502e-05, 'epoch': 1.39}) +[2026-06-19 14:43:40,251][starVLA.training.train_starvla][INFO] - Step 2199, Loss: {'train/loss': 0.115867979824543, 'train/grad_norm_pre_clip': 7.394768238067627, 'timing/data': 0.01933303999248892, 'timing/model': 1.1439223219640553, 'learning_rate/qwen_vl_interface': 6.309943285223824e-06, 'learning_rate/action_model': 6.309943285223824e-05, 'epoch': 1.39}) +[2026-06-19 14:43:44,832][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 2200 +[2026-06-19 14:46:37,500][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 2200 +[2026-06-19 14:46:37,501][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 2200 +[2026-06-19 14:49:44,549][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 2200 +[2026-06-19 14:49:44,550][starVLA.training.train_starvla][INFO] - Step 2200, Loss: {'train/loss': 0.38682636618614197, 'train/grad_norm_pre_clip': 13.692280769348145, 'eval/loss': 0.33060866594314575, 'eval/latency_0/loss': 0.11533134430646896, 'eval/latency_1/loss': 0.07430507987737656, 'eval/latency_2/loss': 0.12705983221530914, 'eval/latency_3/loss': 0.49675294756889343, 'eval/latency_4/loss': 0.8541550636291504, 'eval/flappy/loss': 0.33060866594314575, 'eval/flappy/latency_0/loss': 0.11533134430646896, 'eval/flappy/latency_1/loss': 0.07430507987737656, 'eval/flappy/latency_2/loss': 0.12705983221530914, 'eval/flappy/latency_3/loss': 0.49675294756889343, 'eval/flappy/latency_4/loss': 0.8541550636291504, 'eval/flap_precision': 0.8186528497409327, 'eval/flap_recall': 0.5745454545454546, 'eval/flap_f1': 0.6752136752136753, 'eval/flap_rate_model': 0.05361111111111111, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.6752136752136753, 'eval/latency_0/flap_precision': 0.8186528497409327, 'eval/latency_0/flap_recall': 0.5745454545454546, 'eval/latency_0/flap_f1': 0.6752136752136753, 'eval/latency_0/flap_rate_model': 0.05361111111111111, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.6752136752136753, 'timing/data': 0.017709964886307716, 'timing/model': 1.147730725002475, 'learning_rate/qwen_vl_interface': 6.306974436292493e-06, 'learning_rate/action_model': 6.306974436292494e-05, 'epoch': 1.4}) +[2026-06-19 14:49:44,551][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_2200_state +[2026-06-19 14:49:44,553][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-19 14:49:47,928][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_2200_state/model.safetensors +[2026-06-19 14:49:55,125][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_2200_state/optimizer.bin +[2026-06-19 14:49:55,125][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_2200_state/sampler.bin +[2026-06-19 14:49:55,125][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_2200_state/sampler_1.bin +[2026-06-19 14:49:55,126][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_2200_state/random_states_0.pkl +[2026-06-19 14:49:55,175][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2200 +[2026-06-19 14:49:55,175][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-19 14:49:55,187][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-19 14:49:59,683][starVLA.training.train_starvla][INFO] - Step 2201, Loss: {'train/loss': 0.5571922659873962, 'train/grad_norm_pre_clip': 21.665773391723633, 'timing/data': 0.01775713195092976, 'timing/model': 1.1350690789986402, 'learning_rate/qwen_vl_interface': 6.304005152878902e-06, 'learning_rate/action_model': 6.304005152878901e-05, 'epoch': 1.4}) +[2026-06-19 14:50:04,223][starVLA.training.train_starvla][INFO] - Step 2202, Loss: {'train/loss': 0.21281512081623077, 'train/grad_norm_pre_clip': 8.611185073852539, 'timing/data': 0.01828055107034743, 'timing/model': 1.1301312110153958, 'learning_rate/qwen_vl_interface': 6.301035436203602e-06, 'learning_rate/action_model': 6.301035436203602e-05, 'epoch': 1.4}) +[2026-06-19 14:50:08,764][starVLA.training.train_starvla][INFO] - Step 2203, Loss: {'train/loss': 0.20296591520309448, 'train/grad_norm_pre_clip': 13.865019798278809, 'timing/data': 0.01759569498244673, 'timing/model': 1.3042699809884652, 'learning_rate/qwen_vl_interface': 6.298065287487337e-06, 'learning_rate/action_model': 6.298065287487337e-05, 'epoch': 1.4}) +[2026-06-19 14:50:13,128][starVLA.training.train_starvla][INFO] - Step 2204, Loss: {'train/loss': 0.15911473333835602, 'train/grad_norm_pre_clip': 10.840020179748535, 'timing/data': 0.017514970968477428, 'timing/model': 1.1349476590985432, 'learning_rate/qwen_vl_interface': 6.295094707951019e-06, 'learning_rate/action_model': 6.295094707951019e-05, 'epoch': 1.4}) +[2026-06-19 14:50:17,640][starVLA.training.train_starvla][INFO] - Step 2205, Loss: {'train/loss': 0.2751407027244568, 'train/grad_norm_pre_clip': 10.347593307495117, 'timing/data': 0.01800257503055036, 'timing/model': 1.127851678058505, 'learning_rate/qwen_vl_interface': 6.292123698815745e-06, 'learning_rate/action_model': 6.292123698815745e-05, 'epoch': 1.4}) +[2026-06-19 14:50:22,170][starVLA.training.train_starvla][INFO] - Step 2206, Loss: {'train/loss': 0.13057036697864532, 'train/grad_norm_pre_clip': 6.913029670715332, 'timing/data': 0.017821409972384572, 'timing/model': 1.1642070970265195, 'learning_rate/qwen_vl_interface': 6.289152261302783e-06, 'learning_rate/action_model': 6.289152261302784e-05, 'epoch': 1.4}) +[2026-06-19 14:50:26,730][starVLA.training.train_starvla][INFO] - Step 2207, Loss: {'train/loss': 0.21677353978157043, 'train/grad_norm_pre_clip': 7.594893455505371, 'timing/data': 0.018572058994323015, 'timing/model': 1.1456198610831052, 'learning_rate/qwen_vl_interface': 6.286180396633579e-06, 'learning_rate/action_model': 6.286180396633579e-05, 'epoch': 1.4}) +[2026-06-19 14:50:31,288][starVLA.training.train_starvla][INFO] - Step 2208, Loss: {'train/loss': 0.30214664340019226, 'train/grad_norm_pre_clip': 8.020005226135254, 'timing/data': 0.017870895913802087, 'timing/model': 1.1423161210259423, 'learning_rate/qwen_vl_interface': 6.283208106029753e-06, 'learning_rate/action_model': 6.283208106029752e-05, 'epoch': 1.4}) +[2026-06-19 14:50:35,864][starVLA.training.train_starvla][INFO] - Step 2209, Loss: {'train/loss': 0.17395846545696259, 'train/grad_norm_pre_clip': 8.99928092956543, 'timing/data': 0.022800966980867088, 'timing/model': 1.3180168339749798, 'learning_rate/qwen_vl_interface': 6.280235390713102e-06, 'learning_rate/action_model': 6.280235390713102e-05, 'epoch': 1.4}) +[2026-06-19 14:50:40,256][starVLA.training.train_starvla][INFO] - Step 2210, Loss: {'train/loss': 0.24863767623901367, 'train/grad_norm_pre_clip': 7.895306587219238, 'timing/data': 0.01679649599827826, 'timing/model': 1.1426364229992032, 'learning_rate/qwen_vl_interface': 6.277262251905597e-06, 'learning_rate/action_model': 6.277262251905597e-05, 'epoch': 1.4}) +[2026-06-19 14:50:44,829][starVLA.training.train_starvla][INFO] - Step 2211, Loss: {'train/loss': 0.16426719725131989, 'train/grad_norm_pre_clip': 5.910575866699219, 'timing/data': 0.017964678001590073, 'timing/model': 1.1383344689384103, 'learning_rate/qwen_vl_interface': 6.274288690829385e-06, 'learning_rate/action_model': 6.274288690829384e-05, 'epoch': 1.4}) +[2026-06-19 14:50:49,395][starVLA.training.train_starvla][INFO] - Step 2212, Loss: {'train/loss': 0.03356205299496651, 'train/grad_norm_pre_clip': 1.7316060066223145, 'timing/data': 0.01783847203478217, 'timing/model': 1.1447499360656366, 'learning_rate/qwen_vl_interface': 6.271314708706781e-06, 'learning_rate/action_model': 6.271314708706782e-05, 'epoch': 1.4}) +[2026-06-19 14:50:53,897][starVLA.training.train_starvla][INFO] - Step 2213, Loss: {'train/loss': 0.2058725506067276, 'train/grad_norm_pre_clip': 6.76997709274292, 'timing/data': 0.019203701987862587, 'timing/model': 1.1360582530032843, 'learning_rate/qwen_vl_interface': 6.26834030676028e-06, 'learning_rate/action_model': 6.26834030676028e-05, 'epoch': 1.4}) +[2026-06-19 14:50:58,460][starVLA.training.train_starvla][INFO] - Step 2214, Loss: {'train/loss': 0.13211394846439362, 'train/grad_norm_pre_clip': 6.971818447113037, 'timing/data': 0.017981584998779, 'timing/model': 1.1556225949898362, 'learning_rate/qwen_vl_interface': 6.265365486212545e-06, 'learning_rate/action_model': 6.265365486212545e-05, 'epoch': 1.4}) +[2026-06-19 14:51:03,039][starVLA.training.train_starvla][INFO] - Step 2215, Loss: {'train/loss': 0.23490847647190094, 'train/grad_norm_pre_clip': 8.367826461791992, 'timing/data': 0.018764629028737545, 'timing/model': 1.313886814052239, 'learning_rate/qwen_vl_interface': 6.26239024828641e-06, 'learning_rate/action_model': 6.26239024828641e-05, 'epoch': 1.4}) +[2026-06-19 14:51:07,451][starVLA.training.train_starvla][INFO] - Step 2216, Loss: {'train/loss': 0.11553849279880524, 'train/grad_norm_pre_clip': 9.653308868408203, 'timing/data': 0.017619192018173635, 'timing/model': 1.141455480013974, 'learning_rate/qwen_vl_interface': 6.259414594204888e-06, 'learning_rate/action_model': 6.259414594204888e-05, 'epoch': 1.41}) +[2026-06-19 14:51:12,018][starVLA.training.train_starvla][INFO] - Step 2217, Loss: {'train/loss': 0.19240802526474, 'train/grad_norm_pre_clip': 14.36834716796875, 'timing/data': 0.01834258495364338, 'timing/model': 1.1454875589115545, 'learning_rate/qwen_vl_interface': 6.256438525191155e-06, 'learning_rate/action_model': 6.256438525191156e-05, 'epoch': 1.41}) +[2026-06-19 14:51:16,592][starVLA.training.train_starvla][INFO] - Step 2218, Loss: {'train/loss': 0.2681504189968109, 'train/grad_norm_pre_clip': 12.508045196533203, 'timing/data': 0.017851731041446328, 'timing/model': 1.1388025910127908, 'learning_rate/qwen_vl_interface': 6.253462042468561e-06, 'learning_rate/action_model': 6.25346204246856e-05, 'epoch': 1.41}) +[2026-06-19 14:51:21,158][starVLA.training.train_starvla][INFO] - Step 2219, Loss: {'train/loss': 0.14221449196338654, 'train/grad_norm_pre_clip': 6.760333061218262, 'timing/data': 0.01897518604528159, 'timing/model': 1.1474404969485477, 'learning_rate/qwen_vl_interface': 6.2504851472606246e-06, 'learning_rate/action_model': 6.250485147260625e-05, 'epoch': 1.41}) +[2026-06-19 14:51:25,746][starVLA.training.train_starvla][INFO] - Step 2220, Loss: {'train/loss': 0.2857092320919037, 'train/grad_norm_pre_clip': 12.075883865356445, 'timing/data': 0.01717789692338556, 'timing/model': 1.146891757962294, 'learning_rate/qwen_vl_interface': 6.247507840791035e-06, 'learning_rate/action_model': 6.247507840791035e-05, 'epoch': 1.41}) +[2026-06-19 14:51:30,323][starVLA.training.train_starvla][INFO] - Step 2221, Loss: {'train/loss': 0.20658306777477264, 'train/grad_norm_pre_clip': 9.267782211303711, 'timing/data': 0.01842389500234276, 'timing/model': 1.2998782279901206, 'learning_rate/qwen_vl_interface': 6.244530124283651e-06, 'learning_rate/action_model': 6.24453012428365e-05, 'epoch': 1.41}) +[2026-06-19 14:51:34,716][starVLA.training.train_starvla][INFO] - Step 2222, Loss: {'train/loss': 0.21305567026138306, 'train/grad_norm_pre_clip': 22.187482833862305, 'timing/data': 0.016501317033544183, 'timing/model': 1.1387114849640056, 'learning_rate/qwen_vl_interface': 6.2415519989625016e-06, 'learning_rate/action_model': 6.241551998962501e-05, 'epoch': 1.41}) +[2026-06-19 14:51:39,275][starVLA.training.train_starvla][INFO] - Step 2223, Loss: {'train/loss': 0.18777216970920563, 'train/grad_norm_pre_clip': 12.65563678741455, 'timing/data': 0.017120691016316414, 'timing/model': 1.1410796520067379, 'learning_rate/qwen_vl_interface': 6.238573466051781e-06, 'learning_rate/action_model': 6.23857346605178e-05, 'epoch': 1.41}) +[2026-06-19 14:51:43,855][starVLA.training.train_starvla][INFO] - Step 2224, Loss: {'train/loss': 0.26801642775535583, 'train/grad_norm_pre_clip': 11.760077476501465, 'timing/data': 0.018101862049661577, 'timing/model': 1.1365506160072982, 'learning_rate/qwen_vl_interface': 6.2355945267758465e-06, 'learning_rate/action_model': 6.235594526775846e-05, 'epoch': 1.41}) +[2026-06-19 14:51:48,382][starVLA.training.train_starvla][INFO] - Step 2225, Loss: {'train/loss': 0.3250720500946045, 'train/grad_norm_pre_clip': 17.122695922851562, 'timing/data': 0.018319786991924047, 'timing/model': 1.1384841529652476, 'learning_rate/qwen_vl_interface': 6.232615182359231e-06, 'learning_rate/action_model': 6.232615182359231e-05, 'epoch': 1.41}) +[2026-06-19 14:51:52,976][starVLA.training.train_starvla][INFO] - Step 2226, Loss: {'train/loss': 0.17202259600162506, 'train/grad_norm_pre_clip': 4.910660743713379, 'timing/data': 0.018033484928309917, 'timing/model': 1.1382923669880256, 'learning_rate/qwen_vl_interface': 6.229635434026634e-06, 'learning_rate/action_model': 6.229635434026634e-05, 'epoch': 1.41}) +[2026-06-19 14:51:57,575][starVLA.training.train_starvla][INFO] - Step 2227, Loss: {'train/loss': 0.12228702008724213, 'train/grad_norm_pre_clip': 9.344367027282715, 'timing/data': 0.017872516997158527, 'timing/model': 1.3295828551054, 'learning_rate/qwen_vl_interface': 6.226655283002912e-06, 'learning_rate/action_model': 6.226655283002911e-05, 'epoch': 1.41}) +[2026-06-19 14:52:01,994][starVLA.training.train_starvla][INFO] - Step 2228, Loss: {'train/loss': 0.5728620886802673, 'train/grad_norm_pre_clip': 14.810065269470215, 'timing/data': 0.018508902983739972, 'timing/model': 1.1466332010459155, 'learning_rate/qwen_vl_interface': 6.223674730513096e-06, 'learning_rate/action_model': 6.223674730513096e-05, 'epoch': 1.41}) +[2026-06-19 14:52:06,554][starVLA.training.train_starvla][INFO] - Step 2229, Loss: {'train/loss': 0.10132584720849991, 'train/grad_norm_pre_clip': 13.53104019165039, 'timing/data': 0.01844989403616637, 'timing/model': 1.141057725995779, 'learning_rate/qwen_vl_interface': 6.220693777782375e-06, 'learning_rate/action_model': 6.220693777782375e-05, 'epoch': 1.41}) +[2026-06-19 14:52:11,171][starVLA.training.train_starvla][INFO] - Step 2230, Loss: {'train/loss': 0.16812515258789062, 'train/grad_norm_pre_clip': 23.254348754882812, 'timing/data': 0.018716496997512877, 'timing/model': 1.1432832339778543, 'learning_rate/qwen_vl_interface': 6.21771242603611e-06, 'learning_rate/action_model': 6.21771242603611e-05, 'epoch': 1.41}) +[2026-06-19 14:52:15,649][starVLA.training.train_starvla][INFO] - Step 2231, Loss: {'train/loss': 0.22434328496456146, 'train/grad_norm_pre_clip': 11.569439888000488, 'timing/data': 0.017365009058266878, 'timing/model': 1.1427183219930157, 'learning_rate/qwen_vl_interface': 6.21473067649982e-06, 'learning_rate/action_model': 6.21473067649982e-05, 'epoch': 1.41}) +[2026-06-19 14:52:20,179][starVLA.training.train_starvla][INFO] - Step 2232, Loss: {'train/loss': 0.19140729308128357, 'train/grad_norm_pre_clip': 8.790617942810059, 'timing/data': 0.019641687045805156, 'timing/model': 1.1442035359796137, 'learning_rate/qwen_vl_interface': 6.211748530399189e-06, 'learning_rate/action_model': 6.211748530399189e-05, 'epoch': 1.42}) +[2026-06-19 14:52:24,776][starVLA.training.train_starvla][INFO] - Step 2233, Loss: {'train/loss': 0.3339492380619049, 'train/grad_norm_pre_clip': 15.42145824432373, 'timing/data': 0.018773664953187108, 'timing/model': 1.3229690130101517, 'learning_rate/qwen_vl_interface': 6.208765988960067e-06, 'learning_rate/action_model': 6.208765988960066e-05, 'epoch': 1.42}) +[2026-06-19 14:52:29,194][starVLA.training.train_starvla][INFO] - Step 2234, Loss: {'train/loss': 0.25200986862182617, 'train/grad_norm_pre_clip': 11.717391014099121, 'timing/data': 0.01873089501168579, 'timing/model': 1.1480200450168923, 'learning_rate/qwen_vl_interface': 6.2057830534084615e-06, 'learning_rate/action_model': 6.205783053408461e-05, 'epoch': 1.42}) +[2026-06-19 14:52:33,798][starVLA.training.train_starvla][INFO] - Step 2235, Loss: {'train/loss': 0.16981126368045807, 'train/grad_norm_pre_clip': 9.294477462768555, 'timing/data': 0.018934232997708023, 'timing/model': 1.1436256839660928, 'learning_rate/qwen_vl_interface': 6.202799724970546e-06, 'learning_rate/action_model': 6.202799724970546e-05, 'epoch': 1.42}) +[2026-06-19 14:52:38,379][starVLA.training.train_starvla][INFO] - Step 2236, Loss: {'train/loss': 0.038631945848464966, 'train/grad_norm_pre_clip': 3.0992894172668457, 'timing/data': 0.017228165990673006, 'timing/model': 1.1403891299851239, 'learning_rate/qwen_vl_interface': 6.199816004872655e-06, 'learning_rate/action_model': 6.199816004872655e-05, 'epoch': 1.42}) +[2026-06-19 14:52:42,893][starVLA.training.train_starvla][INFO] - Step 2237, Loss: {'train/loss': 0.14234180748462677, 'train/grad_norm_pre_clip': 9.348892211914062, 'timing/data': 0.017999361036345363, 'timing/model': 1.1472513929475099, 'learning_rate/qwen_vl_interface': 6.196831894341281e-06, 'learning_rate/action_model': 6.19683189434128e-05, 'epoch': 1.42}) +[2026-06-19 14:52:47,376][starVLA.training.train_starvla][INFO] - Step 2238, Loss: {'train/loss': 0.180466890335083, 'train/grad_norm_pre_clip': 11.58990478515625, 'timing/data': 0.018521576072089374, 'timing/model': 1.1421411589253694, 'learning_rate/qwen_vl_interface': 6.1938473946030795e-06, 'learning_rate/action_model': 6.193847394603079e-05, 'epoch': 1.42}) +[2026-06-19 14:52:51,903][starVLA.training.train_starvla][INFO] - Step 2239, Loss: {'train/loss': 0.17073485255241394, 'train/grad_norm_pre_clip': 6.79121208190918, 'timing/data': 0.018177377060055733, 'timing/model': 1.253726682974957, 'learning_rate/qwen_vl_interface': 6.1908625068848695e-06, 'learning_rate/action_model': 6.19086250688487e-05, 'epoch': 1.42}) +[2026-06-19 14:52:56,325][starVLA.training.train_starvla][INFO] - Step 2240, Loss: {'train/loss': 0.23999814689159393, 'train/grad_norm_pre_clip': 16.14910125732422, 'timing/data': 0.0188667579786852, 'timing/model': 1.1468565030954778, 'learning_rate/qwen_vl_interface': 6.187877232413619e-06, 'learning_rate/action_model': 6.18787723241362e-05, 'epoch': 1.42}) +[2026-06-19 14:53:00,922][starVLA.training.train_starvla][INFO] - Step 2241, Loss: {'train/loss': 0.3207637667655945, 'train/grad_norm_pre_clip': 14.52874755859375, 'timing/data': 0.017240846063941717, 'timing/model': 1.150382149964571, 'learning_rate/qwen_vl_interface': 6.184891572416467e-06, 'learning_rate/action_model': 6.184891572416466e-05, 'epoch': 1.42}) +[2026-06-19 14:53:05,535][starVLA.training.train_starvla][INFO] - Step 2242, Loss: {'train/loss': 0.28585895895957947, 'train/grad_norm_pre_clip': 10.575432777404785, 'timing/data': 0.017883555963635445, 'timing/model': 1.1661077700555325, 'learning_rate/qwen_vl_interface': 6.181905528120703e-06, 'learning_rate/action_model': 6.181905528120703e-05, 'epoch': 1.42}) +[2026-06-19 14:53:10,136][starVLA.training.train_starvla][INFO] - Step 2243, Loss: {'train/loss': 0.1697506457567215, 'train/grad_norm_pre_clip': 9.745351791381836, 'timing/data': 0.017448307015001774, 'timing/model': 1.1473438369575888, 'learning_rate/qwen_vl_interface': 6.178919100753778e-06, 'learning_rate/action_model': 6.178919100753778e-05, 'epoch': 1.42}) +[2026-06-19 14:53:14,717][starVLA.training.train_starvla][INFO] - Step 2244, Loss: {'train/loss': 0.21716538071632385, 'train/grad_norm_pre_clip': 10.007768630981445, 'timing/data': 0.01704610895831138, 'timing/model': 1.3115758390631527, 'learning_rate/qwen_vl_interface': 6.175932291543299e-06, 'learning_rate/action_model': 6.175932291543298e-05, 'epoch': 1.42}) +[2026-06-19 14:53:19,259][starVLA.training.train_starvla][INFO] - Step 2245, Loss: {'train/loss': 0.2765825390815735, 'train/grad_norm_pre_clip': 12.098135948181152, 'timing/data': 0.01756206003483385, 'timing/model': 1.2689200510503724, 'learning_rate/qwen_vl_interface': 6.172945101717029e-06, 'learning_rate/action_model': 6.17294510171703e-05, 'epoch': 1.42}) +[2026-06-19 14:53:23,680][starVLA.training.train_starvla][INFO] - Step 2246, Loss: {'train/loss': 0.2980661392211914, 'train/grad_norm_pre_clip': 15.514222145080566, 'timing/data': 0.018178111989982426, 'timing/model': 1.1495189969427884, 'learning_rate/qwen_vl_interface': 6.169957532502892e-06, 'learning_rate/action_model': 6.169957532502892e-05, 'epoch': 1.42}) +[2026-06-19 14:53:28,269][starVLA.training.train_starvla][INFO] - Step 2247, Loss: {'train/loss': 0.328380286693573, 'train/grad_norm_pre_clip': 12.987157821655273, 'timing/data': 0.018246671999804676, 'timing/model': 1.142498899018392, 'learning_rate/qwen_vl_interface': 6.166969585128962e-06, 'learning_rate/action_model': 6.166969585128962e-05, 'epoch': 1.42}) +[2026-06-19 14:53:32,878][starVLA.training.train_starvla][INFO] - Step 2248, Loss: {'train/loss': 0.0368572361767292, 'train/grad_norm_pre_clip': 2.7203783988952637, 'timing/data': 0.017987378989346325, 'timing/model': 1.1468014509882778, 'learning_rate/qwen_vl_interface': 6.163981260823473e-06, 'learning_rate/action_model': 6.163981260823473e-05, 'epoch': 1.43}) +[2026-06-19 14:53:37,480][starVLA.training.train_starvla][INFO] - Step 2249, Loss: {'train/loss': 0.32549765706062317, 'train/grad_norm_pre_clip': 12.175028800964355, 'timing/data': 0.01755250198766589, 'timing/model': 1.1529379690764472, 'learning_rate/qwen_vl_interface': 6.16099256081481e-06, 'learning_rate/action_model': 6.16099256081481e-05, 'epoch': 1.43}) +[2026-06-19 14:53:42,080][starVLA.training.train_starvla][INFO] - Step 2250, Loss: {'train/loss': 0.24621173739433289, 'train/grad_norm_pre_clip': 6.6720757484436035, 'timing/data': 0.017895197961479425, 'timing/model': 1.3260232310276479, 'learning_rate/qwen_vl_interface': 6.15800348633152e-06, 'learning_rate/action_model': 6.15800348633152e-05, 'epoch': 1.43}) +[2026-06-19 14:53:46,519][starVLA.training.train_starvla][INFO] - Step 2251, Loss: {'train/loss': 0.22522954642772675, 'train/grad_norm_pre_clip': 11.522245407104492, 'timing/data': 0.017290451913140714, 'timing/model': 1.147911989944987, 'learning_rate/qwen_vl_interface': 6.155014038602293e-06, 'learning_rate/action_model': 6.155014038602292e-05, 'epoch': 1.43}) +[2026-06-19 14:53:51,103][starVLA.training.train_starvla][INFO] - Step 2252, Loss: {'train/loss': 0.21778380870819092, 'train/grad_norm_pre_clip': 13.187117576599121, 'timing/data': 0.018074330990202725, 'timing/model': 1.1423023800598457, 'learning_rate/qwen_vl_interface': 6.152024218855979e-06, 'learning_rate/action_model': 6.15202421885598e-05, 'epoch': 1.43}) +[2026-06-19 14:53:55,700][starVLA.training.train_starvla][INFO] - Step 2253, Loss: {'train/loss': 0.13852623105049133, 'train/grad_norm_pre_clip': 11.872241020202637, 'timing/data': 0.016777909942902625, 'timing/model': 1.1517942129867151, 'learning_rate/qwen_vl_interface': 6.149034028321583e-06, 'learning_rate/action_model': 6.149034028321583e-05, 'epoch': 1.43}) +[2026-06-19 14:54:00,284][starVLA.training.train_starvla][INFO] - Step 2254, Loss: {'train/loss': 0.15224172174930573, 'train/grad_norm_pre_clip': 7.437753677368164, 'timing/data': 0.017398774973116815, 'timing/model': 1.1460264220368117, 'learning_rate/qwen_vl_interface': 6.1460434682282566e-06, 'learning_rate/action_model': 6.146043468228256e-05, 'epoch': 1.43}) +[2026-06-19 14:54:04,860][starVLA.training.train_starvla][INFO] - Step 2255, Loss: {'train/loss': 0.11948519945144653, 'train/grad_norm_pre_clip': 7.758061408996582, 'timing/data': 0.017561838030815125, 'timing/model': 1.146222151000984, 'learning_rate/qwen_vl_interface': 6.143052539805308e-06, 'learning_rate/action_model': 6.143052539805308e-05, 'epoch': 1.43}) +[2026-06-19 14:54:09,465][starVLA.training.train_starvla][INFO] - Step 2256, Loss: {'train/loss': 0.1690273880958557, 'train/grad_norm_pre_clip': 6.95224142074585, 'timing/data': 0.0170464760158211, 'timing/model': 1.314276788965799, 'learning_rate/qwen_vl_interface': 6.140061244282194e-06, 'learning_rate/action_model': 6.140061244282194e-05, 'epoch': 1.43}) +[2026-06-19 14:54:14,009][starVLA.training.train_starvla][INFO] - Step 2257, Loss: {'train/loss': 0.16785390675067902, 'train/grad_norm_pre_clip': 7.499144554138184, 'timing/data': 0.017433761968277395, 'timing/model': 1.2549732010811567, 'learning_rate/qwen_vl_interface': 6.137069582888521e-06, 'learning_rate/action_model': 6.137069582888521e-05, 'epoch': 1.43}) +[2026-06-19 14:54:18,463][starVLA.training.train_starvla][INFO] - Step 2258, Loss: {'train/loss': 0.3379565477371216, 'train/grad_norm_pre_clip': 15.896065711975098, 'timing/data': 0.017381770885549486, 'timing/model': 1.1545666391029954, 'learning_rate/qwen_vl_interface': 6.13407755685405e-06, 'learning_rate/action_model': 6.13407755685405e-05, 'epoch': 1.43}) +[2026-06-19 14:54:22,965][starVLA.training.train_starvla][INFO] - Step 2259, Loss: {'train/loss': 0.10100733488798141, 'train/grad_norm_pre_clip': 5.583596706390381, 'timing/data': 0.017647861037403345, 'timing/model': 1.1418376460205764, 'learning_rate/qwen_vl_interface': 6.131085167408692e-06, 'learning_rate/action_model': 6.131085167408692e-05, 'epoch': 1.43}) +[2026-06-19 14:54:27,584][starVLA.training.train_starvla][INFO] - Step 2260, Loss: {'train/loss': 0.24881725013256073, 'train/grad_norm_pre_clip': 11.027231216430664, 'timing/data': 0.01922809996176511, 'timing/model': 1.1456254590302706, 'learning_rate/qwen_vl_interface': 6.128092415782502e-06, 'learning_rate/action_model': 6.128092415782501e-05, 'epoch': 1.43}) +[2026-06-19 14:54:32,173][starVLA.training.train_starvla][INFO] - Step 2261, Loss: {'train/loss': 0.1129695326089859, 'train/grad_norm_pre_clip': 5.189028739929199, 'timing/data': 0.017647317959927022, 'timing/model': 1.1467795359203592, 'learning_rate/qwen_vl_interface': 6.125099303205688e-06, 'learning_rate/action_model': 6.125099303205688e-05, 'epoch': 1.43}) +[2026-06-19 14:54:36,782][starVLA.training.train_starvla][INFO] - Step 2262, Loss: {'train/loss': 0.09685661643743515, 'train/grad_norm_pre_clip': 6.403448104858398, 'timing/data': 0.018627548008225858, 'timing/model': 1.2981867210473865, 'learning_rate/qwen_vl_interface': 6.122105830908607e-06, 'learning_rate/action_model': 6.122105830908606e-05, 'epoch': 1.43}) +[2026-06-19 14:54:41,205][starVLA.training.train_starvla][INFO] - Step 2263, Loss: {'train/loss': 0.33022135496139526, 'train/grad_norm_pre_clip': 10.071795463562012, 'timing/data': 0.017648960929363966, 'timing/model': 1.1443309179740027, 'learning_rate/qwen_vl_interface': 6.11911200012176e-06, 'learning_rate/action_model': 6.11911200012176e-05, 'epoch': 1.44}) +[2026-06-19 14:54:45,826][starVLA.training.train_starvla][INFO] - Step 2264, Loss: {'train/loss': 0.11114902049303055, 'train/grad_norm_pre_clip': 9.282824516296387, 'timing/data': 0.0183094049571082, 'timing/model': 1.150665328023024, 'learning_rate/qwen_vl_interface': 6.116117812075799e-06, 'learning_rate/action_model': 6.116117812075799e-05, 'epoch': 1.44}) +[2026-06-19 14:54:50,400][starVLA.training.train_starvla][INFO] - Step 2265, Loss: {'train/loss': 0.20526142418384552, 'train/grad_norm_pre_clip': 10.484647750854492, 'timing/data': 0.018157145008444786, 'timing/model': 1.1422421259339899, 'learning_rate/qwen_vl_interface': 6.113123268001524e-06, 'learning_rate/action_model': 6.113123268001524e-05, 'epoch': 1.44}) +[2026-06-19 14:54:54,876][starVLA.training.train_starvla][INFO] - Step 2266, Loss: {'train/loss': 0.22136259078979492, 'train/grad_norm_pre_clip': 10.660453796386719, 'timing/data': 0.01877374399919063, 'timing/model': 1.1423716510180384, 'learning_rate/qwen_vl_interface': 6.110128369129873e-06, 'learning_rate/action_model': 6.110128369129873e-05, 'epoch': 1.44}) +[2026-06-19 14:54:59,448][starVLA.training.train_starvla][INFO] - Step 2267, Loss: {'train/loss': 0.22871828079223633, 'train/grad_norm_pre_clip': 12.709818840026855, 'timing/data': 0.01793505495879799, 'timing/model': 1.148966173059307, 'learning_rate/qwen_vl_interface': 6.107133116691943e-06, 'learning_rate/action_model': 6.107133116691942e-05, 'epoch': 1.44}) +[2026-06-19 14:55:04,049][starVLA.training.train_starvla][INFO] - Step 2268, Loss: {'train/loss': 0.11865853518247604, 'train/grad_norm_pre_clip': 7.64796781539917, 'timing/data': 0.018975157057866454, 'timing/model': 1.3105022209929302, 'learning_rate/qwen_vl_interface': 6.104137511918963e-06, 'learning_rate/action_model': 6.104137511918963e-05, 'epoch': 1.44}) +[2026-06-19 14:55:08,463][starVLA.training.train_starvla][INFO] - Step 2269, Loss: {'train/loss': 0.13761165738105774, 'train/grad_norm_pre_clip': 9.408407211303711, 'timing/data': 0.018655681051313877, 'timing/model': 1.141167497029528, 'learning_rate/qwen_vl_interface': 6.1011415560423155e-06, 'learning_rate/action_model': 6.101141556042316e-05, 'epoch': 1.44}) +[2026-06-19 14:55:13,039][starVLA.training.train_starvla][INFO] - Step 2270, Loss: {'train/loss': 0.16140109300613403, 'train/grad_norm_pre_clip': 10.454276084899902, 'timing/data': 0.0186941270949319, 'timing/model': 1.1425898550078273, 'learning_rate/qwen_vl_interface': 6.098145250293525e-06, 'learning_rate/action_model': 6.0981452502935246e-05, 'epoch': 1.44}) +[2026-06-19 14:55:17,603][starVLA.training.train_starvla][INFO] - Step 2271, Loss: {'train/loss': 0.2370254546403885, 'train/grad_norm_pre_clip': 9.829916954040527, 'timing/data': 0.017460578004829586, 'timing/model': 1.1401534740580246, 'learning_rate/qwen_vl_interface': 6.09514859590426e-06, 'learning_rate/action_model': 6.09514859590426e-05, 'epoch': 1.44}) +[2026-06-19 14:55:22,193][starVLA.training.train_starvla][INFO] - Step 2272, Loss: {'train/loss': 0.4208988845348358, 'train/grad_norm_pre_clip': 36.034507751464844, 'timing/data': 0.01749319804366678, 'timing/model': 1.1416518399491906, 'learning_rate/qwen_vl_interface': 6.09215159410633e-06, 'learning_rate/action_model': 6.09215159410633e-05, 'epoch': 1.44}) +[2026-06-19 14:55:26,792][starVLA.training.train_starvla][INFO] - Step 2273, Loss: {'train/loss': 0.34010976552963257, 'train/grad_norm_pre_clip': 19.059932708740234, 'timing/data': 0.017822494031861424, 'timing/model': 1.1506238409783691, 'learning_rate/qwen_vl_interface': 6.089154246131692e-06, 'learning_rate/action_model': 6.089154246131692e-05, 'epoch': 1.44}) +[2026-06-19 14:55:31,385][starVLA.training.train_starvla][INFO] - Step 2274, Loss: {'train/loss': 0.156430184841156, 'train/grad_norm_pre_clip': 9.384369850158691, 'timing/data': 0.01723124203272164, 'timing/model': 1.3106234209844843, 'learning_rate/qwen_vl_interface': 6.08615655321244e-06, 'learning_rate/action_model': 6.08615655321244e-05, 'epoch': 1.44}) +[2026-06-19 14:55:35,821][starVLA.training.train_starvla][INFO] - Step 2275, Loss: {'train/loss': 0.29342222213745117, 'train/grad_norm_pre_clip': 11.564996719360352, 'timing/data': 0.018290110980160534, 'timing/model': 1.1491316109895706, 'learning_rate/qwen_vl_interface': 6.083158516580814e-06, 'learning_rate/action_model': 6.083158516580814e-05, 'epoch': 1.44}) +[2026-06-19 14:55:40,409][starVLA.training.train_starvla][INFO] - Step 2276, Loss: {'train/loss': 0.30421918630599976, 'train/grad_norm_pre_clip': 6.084876537322998, 'timing/data': 0.01915620593354106, 'timing/model': 1.1416424920316786, 'learning_rate/qwen_vl_interface': 6.080160137469192e-06, 'learning_rate/action_model': 6.0801601374691916e-05, 'epoch': 1.44}) +[2026-06-19 14:55:45,006][starVLA.training.train_starvla][INFO] - Step 2277, Loss: {'train/loss': 0.27102088928222656, 'train/grad_norm_pre_clip': 14.696115493774414, 'timing/data': 0.017867839080281556, 'timing/model': 1.1542693941155449, 'learning_rate/qwen_vl_interface': 6.077161417110097e-06, 'learning_rate/action_model': 6.0771614171100965e-05, 'epoch': 1.44}) +[2026-06-19 14:55:49,562][starVLA.training.train_starvla][INFO] - Step 2278, Loss: {'train/loss': 0.13995538651943207, 'train/grad_norm_pre_clip': 5.390113353729248, 'timing/data': 0.017735676956363022, 'timing/model': 1.148165654973127, 'learning_rate/qwen_vl_interface': 6.0741623567361864e-06, 'learning_rate/action_model': 6.074162356736186e-05, 'epoch': 1.44}) +[2026-06-19 14:55:54,166][starVLA.training.train_starvla][INFO] - Step 2279, Loss: {'train/loss': 0.10937847942113876, 'train/grad_norm_pre_clip': 7.069455623626709, 'timing/data': 0.01889395492617041, 'timing/model': 1.142267204937525, 'learning_rate/qwen_vl_interface': 6.071162957580262e-06, 'learning_rate/action_model': 6.071162957580262e-05, 'epoch': 1.45}) +[2026-06-19 14:55:58,788][starVLA.training.train_starvla][INFO] - Step 2280, Loss: {'train/loss': 0.07394855469465256, 'train/grad_norm_pre_clip': 7.011161804199219, 'timing/data': 0.017768073943443596, 'timing/model': 1.3295450810110196, 'learning_rate/qwen_vl_interface': 6.068163220875266e-06, 'learning_rate/action_model': 6.068163220875266e-05, 'epoch': 1.45}) +[2026-06-19 14:56:03,220][starVLA.training.train_starvla][INFO] - Step 2281, Loss: {'train/loss': 0.14298635721206665, 'train/grad_norm_pre_clip': 16.96076011657715, 'timing/data': 0.01712550804950297, 'timing/model': 1.1506079970858991, 'learning_rate/qwen_vl_interface': 6.065163147854273e-06, 'learning_rate/action_model': 6.065163147854273e-05, 'epoch': 1.45}) +[2026-06-19 14:56:07,837][starVLA.training.train_starvla][INFO] - Step 2282, Loss: {'train/loss': 0.2899135649204254, 'train/grad_norm_pre_clip': 8.564654350280762, 'timing/data': 0.017180321039631963, 'timing/model': 1.149979018024169, 'learning_rate/qwen_vl_interface': 6.062162739750502e-06, 'learning_rate/action_model': 6.062162739750502e-05, 'epoch': 1.45}) +[2026-06-19 14:56:12,446][starVLA.training.train_starvla][INFO] - Step 2283, Loss: {'train/loss': 0.11277914047241211, 'train/grad_norm_pre_clip': 5.946964263916016, 'timing/data': 0.01872936205472797, 'timing/model': 1.1494622740428895, 'learning_rate/qwen_vl_interface': 6.059161997797306e-06, 'learning_rate/action_model': 6.0591619977973056e-05, 'epoch': 1.45}) +[2026-06-19 14:56:16,945][starVLA.training.train_starvla][INFO] - Step 2284, Loss: {'train/loss': 0.0702856108546257, 'train/grad_norm_pre_clip': 4.420780658721924, 'timing/data': 0.019502558978274465, 'timing/model': 1.151976443012245, 'learning_rate/qwen_vl_interface': 6.056160923228178e-06, 'learning_rate/action_model': 6.056160923228178e-05, 'epoch': 1.45}) +[2026-06-19 14:56:21,498][starVLA.training.train_starvla][INFO] - Step 2285, Loss: {'train/loss': 0.22226378321647644, 'train/grad_norm_pre_clip': 8.40307903289795, 'timing/data': 0.017798732966184616, 'timing/model': 1.1502031539566815, 'learning_rate/qwen_vl_interface': 6.0531595172767455e-06, 'learning_rate/action_model': 6.0531595172767455e-05, 'epoch': 1.45}) +[2026-06-19 14:56:26,098][starVLA.training.train_starvla][INFO] - Step 2286, Loss: {'train/loss': 0.1313406229019165, 'train/grad_norm_pre_clip': 6.9275221824646, 'timing/data': 0.018986233044415712, 'timing/model': 1.3141405039932579, 'learning_rate/qwen_vl_interface': 6.050157781176774e-06, 'learning_rate/action_model': 6.050157781176774e-05, 'epoch': 1.45}) +[2026-06-19 14:56:30,526][starVLA.training.train_starvla][INFO] - Step 2287, Loss: {'train/loss': 0.14821991324424744, 'train/grad_norm_pre_clip': 6.811334133148193, 'timing/data': 0.018360863090492785, 'timing/model': 1.1445639519952238, 'learning_rate/qwen_vl_interface': 6.047155716162164e-06, 'learning_rate/action_model': 6.047155716162164e-05, 'epoch': 1.45}) +[2026-06-19 14:56:35,125][starVLA.training.train_starvla][INFO] - Step 2288, Loss: {'train/loss': 0.07682474702596664, 'train/grad_norm_pre_clip': 6.723660469055176, 'timing/data': 0.020225732936523855, 'timing/model': 1.1520565050886944, 'learning_rate/qwen_vl_interface': 6.044153323466948e-06, 'learning_rate/action_model': 6.044153323466948e-05, 'epoch': 1.45}) +[2026-06-19 14:56:39,713][starVLA.training.train_starvla][INFO] - Step 2289, Loss: {'train/loss': 0.48995059728622437, 'train/grad_norm_pre_clip': 11.133583068847656, 'timing/data': 0.019223287934437394, 'timing/model': 1.1511968319537118, 'learning_rate/qwen_vl_interface': 6.0411506043253e-06, 'learning_rate/action_model': 6.0411506043253e-05, 'epoch': 1.45}) +[2026-06-19 14:56:44,219][starVLA.training.train_starvla][INFO] - Step 2290, Loss: {'train/loss': 0.25450271368026733, 'train/grad_norm_pre_clip': 10.007586479187012, 'timing/data': 0.018806570093147457, 'timing/model': 1.1437457530992106, 'learning_rate/qwen_vl_interface': 6.038147559971523e-06, 'learning_rate/action_model': 6.038147559971523e-05, 'epoch': 1.45}) +[2026-06-19 14:56:48,773][starVLA.training.train_starvla][INFO] - Step 2291, Loss: {'train/loss': 0.25723996758461, 'train/grad_norm_pre_clip': 10.232662200927734, 'timing/data': 0.018387268064543605, 'timing/model': 1.1459732729708776, 'learning_rate/qwen_vl_interface': 6.035144191640055e-06, 'learning_rate/action_model': 6.035144191640055e-05, 'epoch': 1.45}) +[2026-06-19 14:56:53,367][starVLA.training.train_starvla][INFO] - Step 2292, Loss: {'train/loss': 0.1910981982946396, 'train/grad_norm_pre_clip': 8.1845064163208, 'timing/data': 0.017277415958233178, 'timing/model': 1.3123715399997309, 'learning_rate/qwen_vl_interface': 6.032140500565467e-06, 'learning_rate/action_model': 6.0321405005654665e-05, 'epoch': 1.45}) +[2026-06-19 14:56:57,810][starVLA.training.train_starvla][INFO] - Step 2293, Loss: {'train/loss': 0.14581885933876038, 'train/grad_norm_pre_clip': 9.579142570495605, 'timing/data': 0.019364877021871507, 'timing/model': 1.1570193419465795, 'learning_rate/qwen_vl_interface': 6.029136487982463e-06, 'learning_rate/action_model': 6.029136487982463e-05, 'epoch': 1.45}) +[2026-06-19 14:57:02,403][starVLA.training.train_starvla][INFO] - Step 2294, Loss: {'train/loss': 0.08461499959230423, 'train/grad_norm_pre_clip': 6.669648170471191, 'timing/data': 0.01820610393770039, 'timing/model': 1.1506600389257073, 'learning_rate/qwen_vl_interface': 6.02613215512588e-06, 'learning_rate/action_model': 6.0261321551258796e-05, 'epoch': 1.45}) +[2026-06-19 14:57:06,997][starVLA.training.train_starvla][INFO] - Step 2295, Loss: {'train/loss': 0.05572403967380524, 'train/grad_norm_pre_clip': 3.897329330444336, 'timing/data': 0.019010495976544917, 'timing/model': 1.1442457661032677, 'learning_rate/qwen_vl_interface': 6.023127503230684e-06, 'learning_rate/action_model': 6.023127503230683e-05, 'epoch': 1.46}) +[2026-06-19 14:57:11,575][starVLA.training.train_starvla][INFO] - Step 2296, Loss: {'train/loss': 0.1276031732559204, 'train/grad_norm_pre_clip': 6.799883842468262, 'timing/data': 0.018585167010314763, 'timing/model': 1.1534376190975308, 'learning_rate/qwen_vl_interface': 6.020122533531976e-06, 'learning_rate/action_model': 6.0201225335319755e-05, 'epoch': 1.46}) +[2026-06-19 14:57:16,078][starVLA.training.train_starvla][INFO] - Step 2297, Loss: {'train/loss': 0.22366946935653687, 'train/grad_norm_pre_clip': 11.891090393066406, 'timing/data': 0.020180235034786165, 'timing/model': 1.143360928981565, 'learning_rate/qwen_vl_interface': 6.017117247264983e-06, 'learning_rate/action_model': 6.017117247264983e-05, 'epoch': 1.46}) +[2026-06-19 14:57:20,618][starVLA.training.train_starvla][INFO] - Step 2298, Loss: {'train/loss': 0.3100700080394745, 'train/grad_norm_pre_clip': 9.027019500732422, 'timing/data': 0.018630763981491327, 'timing/model': 1.2676300809253007, 'learning_rate/qwen_vl_interface': 6.014111645665069e-06, 'learning_rate/action_model': 6.0141116456650684e-05, 'epoch': 1.46}) +[2026-06-19 14:57:25,049][starVLA.training.train_starvla][INFO] - Step 2299, Loss: {'train/loss': 0.3691639006137848, 'train/grad_norm_pre_clip': 10.137009620666504, 'timing/data': 0.01781662297435105, 'timing/model': 1.1520430330419913, 'learning_rate/qwen_vl_interface': 6.011105729967719e-06, 'learning_rate/action_model': 6.0111057299677184e-05, 'epoch': 1.46}) +[2026-06-19 14:57:29,653][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 2300 +[2026-06-19 15:00:18,307][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 2300 +[2026-06-19 15:00:18,308][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 2300 +[2026-06-19 15:03:24,329][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 2300 +[2026-06-19 15:03:24,330][starVLA.training.train_starvla][INFO] - Step 2300, Loss: {'train/loss': 0.28577232360839844, 'train/grad_norm_pre_clip': 13.864635467529297, 'eval/loss': 0.32214054465293884, 'eval/latency_0/loss': 0.10312613844871521, 'eval/latency_1/loss': 0.10372555255889893, 'eval/latency_2/loss': 0.12300629168748856, 'eval/latency_3/loss': 0.5031163692474365, 'eval/latency_4/loss': 0.7926639318466187, 'eval/flappy/loss': 0.32214054465293884, 'eval/flappy/latency_0/loss': 0.10312613844871521, 'eval/flappy/latency_1/loss': 0.10372555255889893, 'eval/flappy/latency_2/loss': 0.12300629168748856, 'eval/flappy/latency_3/loss': 0.5031163692474365, 'eval/flappy/latency_4/loss': 0.7926639318466187, 'eval/flap_precision': 0.780952380952381, 'eval/flap_recall': 0.5963636363636363, 'eval/flap_f1': 0.6762886597938145, 'eval/flap_rate_model': 0.058333333333333334, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.6762886597938145, 'eval/latency_0/flap_precision': 0.780952380952381, 'eval/latency_0/flap_recall': 0.5963636363636363, 'eval/latency_0/flap_f1': 0.6762886597938145, 'eval/latency_0/flap_rate_model': 0.058333333333333334, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.6762886597938145, 'timing/data': 0.01848335296381265, 'timing/model': 1.1411778930341825, 'learning_rate/qwen_vl_interface': 6.008099501408552e-06, 'learning_rate/action_model': 6.008099501408552e-05, 'epoch': 1.46}) +[2026-06-19 15:03:24,330][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_2300_state +[2026-06-19 15:03:24,333][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-19 15:03:27,610][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_2300_state/model.safetensors +[2026-06-19 15:03:34,820][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_2300_state/optimizer.bin +[2026-06-19 15:03:34,821][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_2300_state/sampler.bin +[2026-06-19 15:03:34,821][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_2300_state/sampler_1.bin +[2026-06-19 15:03:34,822][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_2300_state/random_states_0.pkl +[2026-06-19 15:03:34,845][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2300 +[2026-06-19 15:03:34,845][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-19 15:03:34,877][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-19 15:03:39,353][starVLA.training.train_starvla][INFO] - Step 2301, Loss: {'train/loss': 0.21097330749034882, 'train/grad_norm_pre_clip': 14.061626434326172, 'timing/data': 0.01690302393399179, 'timing/model': 1.2803330649621785, 'learning_rate/qwen_vl_interface': 6.0050929612233165e-06, 'learning_rate/action_model': 6.005092961223316e-05, 'epoch': 1.46}) +[2026-06-19 15:03:43,687][starVLA.training.train_starvla][INFO] - Step 2302, Loss: {'train/loss': 0.23688437044620514, 'train/grad_norm_pre_clip': 11.571805000305176, 'timing/data': 0.01759857893921435, 'timing/model': 1.1253213040763512, 'learning_rate/qwen_vl_interface': 6.002086110647888e-06, 'learning_rate/action_model': 6.002086110647888e-05, 'epoch': 1.46}) +[2026-06-19 15:03:48,221][starVLA.training.train_starvla][INFO] - Step 2303, Loss: {'train/loss': 0.256784051656723, 'train/grad_norm_pre_clip': 12.300324440002441, 'timing/data': 0.016925923991948366, 'timing/model': 1.1325453990139067, 'learning_rate/qwen_vl_interface': 5.999078950918267e-06, 'learning_rate/action_model': 5.999078950918267e-05, 'epoch': 1.46}) +[2026-06-19 15:03:52,679][starVLA.training.train_starvla][INFO] - Step 2304, Loss: {'train/loss': 0.38549748063087463, 'train/grad_norm_pre_clip': 12.79926586151123, 'timing/data': 0.017395713017322123, 'timing/model': 1.1395003770012408, 'learning_rate/qwen_vl_interface': 5.996071483270585e-06, 'learning_rate/action_model': 5.996071483270585e-05, 'epoch': 1.46}) +[2026-06-19 15:03:57,133][starVLA.training.train_starvla][INFO] - Step 2305, Loss: {'train/loss': 0.26238542795181274, 'train/grad_norm_pre_clip': 12.325398445129395, 'timing/data': 0.018875592038966715, 'timing/model': 1.1403661280637607, 'learning_rate/qwen_vl_interface': 5.993063708941098e-06, 'learning_rate/action_model': 5.993063708941098e-05, 'epoch': 1.46}) +[2026-06-19 15:04:01,640][starVLA.training.train_starvla][INFO] - Step 2306, Loss: {'train/loss': 0.06458945572376251, 'train/grad_norm_pre_clip': 2.946550130844116, 'timing/data': 0.01841693406458944, 'timing/model': 1.1458857170073316, 'learning_rate/qwen_vl_interface': 5.990055629166187e-06, 'learning_rate/action_model': 5.990055629166187e-05, 'epoch': 1.46}) +[2026-06-19 15:04:06,218][starVLA.training.train_starvla][INFO] - Step 2307, Loss: {'train/loss': 0.24577650427818298, 'train/grad_norm_pre_clip': 13.012487411499023, 'timing/data': 0.018508789944462478, 'timing/model': 1.3241075249388814, 'learning_rate/qwen_vl_interface': 5.987047245182363e-06, 'learning_rate/action_model': 5.987047245182362e-05, 'epoch': 1.46}) +[2026-06-19 15:04:10,609][starVLA.training.train_starvla][INFO] - Step 2308, Loss: {'train/loss': 0.17192359268665314, 'train/grad_norm_pre_clip': 12.217894554138184, 'timing/data': 0.01822766801342368, 'timing/model': 1.1471324320882559, 'learning_rate/qwen_vl_interface': 5.984038558226255e-06, 'learning_rate/action_model': 5.9840385582262546e-05, 'epoch': 1.46}) +[2026-06-19 15:04:15,162][starVLA.training.train_starvla][INFO] - Step 2309, Loss: {'train/loss': 0.06042754277586937, 'train/grad_norm_pre_clip': 3.9679605960845947, 'timing/data': 0.018772289040498435, 'timing/model': 1.136967531987466, 'learning_rate/qwen_vl_interface': 5.981029569534625e-06, 'learning_rate/action_model': 5.981029569534625e-05, 'epoch': 1.46}) +[2026-06-19 15:04:19,715][starVLA.training.train_starvla][INFO] - Step 2310, Loss: {'train/loss': 0.10484380275011063, 'train/grad_norm_pre_clip': 5.351006031036377, 'timing/data': 0.01724931609351188, 'timing/model': 1.1362602330045775, 'learning_rate/qwen_vl_interface': 5.978020280344351e-06, 'learning_rate/action_model': 5.9780202803443506e-05, 'epoch': 1.46}) +[2026-06-19 15:04:24,216][starVLA.training.train_starvla][INFO] - Step 2311, Loss: {'train/loss': 0.10980677604675293, 'train/grad_norm_pre_clip': 11.919594764709473, 'timing/data': 0.018475076067261398, 'timing/model': 1.1476200970355421, 'learning_rate/qwen_vl_interface': 5.97501069189244e-06, 'learning_rate/action_model': 5.97501069189244e-05, 'epoch': 1.47}) +[2026-06-19 15:04:28,742][starVLA.training.train_starvla][INFO] - Step 2312, Loss: {'train/loss': 0.13562199473381042, 'train/grad_norm_pre_clip': 5.613348007202148, 'timing/data': 0.019252256024628878, 'timing/model': 1.138615950010717, 'learning_rate/qwen_vl_interface': 5.97200080541602e-06, 'learning_rate/action_model': 5.97200080541602e-05, 'epoch': 1.47}) +[2026-06-19 15:04:33,318][starVLA.training.train_starvla][INFO] - Step 2313, Loss: {'train/loss': 0.2253691852092743, 'train/grad_norm_pre_clip': 12.051203727722168, 'timing/data': 0.01888103107921779, 'timing/model': 1.3233009870164096, 'learning_rate/qwen_vl_interface': 5.968990622152341e-06, 'learning_rate/action_model': 5.9689906221523404e-05, 'epoch': 1.47}) +[2026-06-19 15:04:37,719][starVLA.training.train_starvla][INFO] - Step 2314, Loss: {'train/loss': 0.27342721819877625, 'train/grad_norm_pre_clip': 38.03832244873047, 'timing/data': 0.018031246960163116, 'timing/model': 1.1439738100161776, 'learning_rate/qwen_vl_interface': 5.965980143338774e-06, 'learning_rate/action_model': 5.965980143338774e-05, 'epoch': 1.47}) +[2026-06-19 15:04:42,289][starVLA.training.train_starvla][INFO] - Step 2315, Loss: {'train/loss': 0.16921213269233704, 'train/grad_norm_pre_clip': 8.12287712097168, 'timing/data': 0.018817357951775193, 'timing/model': 1.1388915620045736, 'learning_rate/qwen_vl_interface': 5.962969370212819e-06, 'learning_rate/action_model': 5.9629693702128185e-05, 'epoch': 1.47}) +[2026-06-19 15:04:46,870][starVLA.training.train_starvla][INFO] - Step 2316, Loss: {'train/loss': 0.22458454966545105, 'train/grad_norm_pre_clip': 12.536489486694336, 'timing/data': 0.01797017897479236, 'timing/model': 1.1426833340665326, 'learning_rate/qwen_vl_interface': 5.959958304012085e-06, 'learning_rate/action_model': 5.9599583040120845e-05, 'epoch': 1.47}) +[2026-06-19 15:04:51,415][starVLA.training.train_starvla][INFO] - Step 2317, Loss: {'train/loss': 0.11869475990533829, 'train/grad_norm_pre_clip': 6.401672840118408, 'timing/data': 0.018292143009603024, 'timing/model': 1.1365272289840505, 'learning_rate/qwen_vl_interface': 5.956946945974309e-06, 'learning_rate/action_model': 5.9569469459743085e-05, 'epoch': 1.47}) +[2026-06-19 15:04:55,886][starVLA.training.train_starvla][INFO] - Step 2318, Loss: {'train/loss': 0.3953286111354828, 'train/grad_norm_pre_clip': 12.71069049835205, 'timing/data': 0.01714495604392141, 'timing/model': 1.1471221529645845, 'learning_rate/qwen_vl_interface': 5.953935297337347e-06, 'learning_rate/action_model': 5.953935297337347e-05, 'epoch': 1.47}) +[2026-06-19 15:05:00,416][starVLA.training.train_starvla][INFO] - Step 2319, Loss: {'train/loss': 0.08575539290904999, 'train/grad_norm_pre_clip': 5.432562828063965, 'timing/data': 0.017343992949463427, 'timing/model': 1.2659925629850477, 'learning_rate/qwen_vl_interface': 5.950923359339175e-06, 'learning_rate/action_model': 5.950923359339174e-05, 'epoch': 1.47}) +[2026-06-19 15:05:04,828][starVLA.training.train_starvla][INFO] - Step 2320, Loss: {'train/loss': 0.25412577390670776, 'train/grad_norm_pre_clip': 11.787150382995605, 'timing/data': 0.0186562929302454, 'timing/model': 1.1454881629906595, 'learning_rate/qwen_vl_interface': 5.947911133217886e-06, 'learning_rate/action_model': 5.947911133217886e-05, 'epoch': 1.47}) +[2026-06-19 15:05:09,384][starVLA.training.train_starvla][INFO] - Step 2321, Loss: {'train/loss': 0.11326341331005096, 'train/grad_norm_pre_clip': 7.09244966506958, 'timing/data': 0.018289305036887527, 'timing/model': 1.1406140680192038, 'learning_rate/qwen_vl_interface': 5.944898620211693e-06, 'learning_rate/action_model': 5.944898620211693e-05, 'epoch': 1.47}) +[2026-06-19 15:05:13,948][starVLA.training.train_starvla][INFO] - Step 2322, Loss: {'train/loss': 0.24954837560653687, 'train/grad_norm_pre_clip': 13.074366569519043, 'timing/data': 0.017882148968055844, 'timing/model': 1.1386102719698101, 'learning_rate/qwen_vl_interface': 5.941885821558925e-06, 'learning_rate/action_model': 5.9418858215589245e-05, 'epoch': 1.47}) +[2026-06-19 15:05:18,513][starVLA.training.train_starvla][INFO] - Step 2323, Loss: {'train/loss': 0.17347121238708496, 'train/grad_norm_pre_clip': 11.561912536621094, 'timing/data': 0.01818909402936697, 'timing/model': 1.141535526025109, 'learning_rate/qwen_vl_interface': 5.938872738498031e-06, 'learning_rate/action_model': 5.93887273849803e-05, 'epoch': 1.47}) +[2026-06-19 15:05:23,077][starVLA.training.train_starvla][INFO] - Step 2324, Loss: {'train/loss': 0.22170816361904144, 'train/grad_norm_pre_clip': 8.082714080810547, 'timing/data': 0.017529911012388766, 'timing/model': 1.30310826189816, 'learning_rate/qwen_vl_interface': 5.935859372267575e-06, 'learning_rate/action_model': 5.935859372267575e-05, 'epoch': 1.47}) +[2026-06-19 15:05:27,555][starVLA.training.train_starvla][INFO] - Step 2325, Loss: {'train/loss': 0.1452031135559082, 'train/grad_norm_pre_clip': 6.430981159210205, 'timing/data': 0.01773778593633324, 'timing/model': 1.2096495419973508, 'learning_rate/qwen_vl_interface': 5.9328457241062396e-06, 'learning_rate/action_model': 5.9328457241062394e-05, 'epoch': 1.47}) +[2026-06-19 15:05:31,974][starVLA.training.train_starvla][INFO] - Step 2326, Loss: {'train/loss': 0.29598355293273926, 'train/grad_norm_pre_clip': 10.228011131286621, 'timing/data': 0.017191419028677046, 'timing/model': 1.148005789029412, 'learning_rate/qwen_vl_interface': 5.929831795252819e-06, 'learning_rate/action_model': 5.9298317952528184e-05, 'epoch': 1.48}) +[2026-06-19 15:05:36,494][starVLA.training.train_starvla][INFO] - Step 2327, Loss: {'train/loss': 0.15960681438446045, 'train/grad_norm_pre_clip': 14.37818717956543, 'timing/data': 0.019083710038103163, 'timing/model': 1.142596025019884, 'learning_rate/qwen_vl_interface': 5.926817586946227e-06, 'learning_rate/action_model': 5.926817586946227e-05, 'epoch': 1.48}) +[2026-06-19 15:05:41,076][starVLA.training.train_starvla][INFO] - Step 2328, Loss: {'train/loss': 0.27533283829689026, 'train/grad_norm_pre_clip': 12.224120140075684, 'timing/data': 0.017241430003196, 'timing/model': 1.1510452480288222, 'learning_rate/qwen_vl_interface': 5.923803100425491e-06, 'learning_rate/action_model': 5.92380310042549e-05, 'epoch': 1.48}) +[2026-06-19 15:05:45,666][starVLA.training.train_starvla][INFO] - Step 2329, Loss: {'train/loss': 0.4443209767341614, 'train/grad_norm_pre_clip': 11.966673851013184, 'timing/data': 0.01751806796528399, 'timing/model': 1.1475666100159287, 'learning_rate/qwen_vl_interface': 5.92078833692975e-06, 'learning_rate/action_model': 5.92078833692975e-05, 'epoch': 1.48}) +[2026-06-19 15:05:50,264][starVLA.training.train_starvla][INFO] - Step 2330, Loss: {'train/loss': 0.23010672628879547, 'train/grad_norm_pre_clip': 9.735730171203613, 'timing/data': 0.017356649041175842, 'timing/model': 1.3118426679866388, 'learning_rate/qwen_vl_interface': 5.917773297698264e-06, 'learning_rate/action_model': 5.9177732976982636e-05, 'epoch': 1.48}) +[2026-06-19 15:05:54,687][starVLA.training.train_starvla][INFO] - Step 2331, Loss: {'train/loss': 0.15083938837051392, 'train/grad_norm_pre_clip': 9.691838264465332, 'timing/data': 0.018678765976801515, 'timing/model': 1.1518430060241371, 'learning_rate/qwen_vl_interface': 5.9147579839704e-06, 'learning_rate/action_model': 5.9147579839704e-05, 'epoch': 1.48}) +[2026-06-19 15:05:59,267][starVLA.training.train_starvla][INFO] - Step 2332, Loss: {'train/loss': 0.14049312472343445, 'train/grad_norm_pre_clip': 8.611392974853516, 'timing/data': 0.01899500598665327, 'timing/model': 1.1398418250028044, 'learning_rate/qwen_vl_interface': 5.911742396985637e-06, 'learning_rate/action_model': 5.9117423969856364e-05, 'epoch': 1.48}) +[2026-06-19 15:06:03,768][starVLA.training.train_starvla][INFO] - Step 2333, Loss: {'train/loss': 0.2436080127954483, 'train/grad_norm_pre_clip': 12.997438430786133, 'timing/data': 0.017449039965867996, 'timing/model': 1.1508687200257555, 'learning_rate/qwen_vl_interface': 5.908726537983571e-06, 'learning_rate/action_model': 5.908726537983571e-05, 'epoch': 1.48}) +[2026-06-19 15:06:08,277][starVLA.training.train_starvla][INFO] - Step 2334, Loss: {'train/loss': 0.15399843454360962, 'train/grad_norm_pre_clip': 7.412140846252441, 'timing/data': 0.018476722994819283, 'timing/model': 1.1445502820424736, 'learning_rate/qwen_vl_interface': 5.9057104082039085e-06, 'learning_rate/action_model': 5.905710408203908e-05, 'epoch': 1.48}) +[2026-06-19 15:06:12,853][starVLA.training.train_starvla][INFO] - Step 2335, Loss: {'train/loss': 0.2658909857273102, 'train/grad_norm_pre_clip': 7.7721171379089355, 'timing/data': 0.01846804202068597, 'timing/model': 1.1441160400863737, 'learning_rate/qwen_vl_interface': 5.902694008886466e-06, 'learning_rate/action_model': 5.902694008886466e-05, 'epoch': 1.48}) +[2026-06-19 15:06:17,447][starVLA.training.train_starvla][INFO] - Step 2336, Loss: {'train/loss': 0.1833547055721283, 'train/grad_norm_pre_clip': 6.409348487854004, 'timing/data': 0.017775521031580865, 'timing/model': 1.3091444459278136, 'learning_rate/qwen_vl_interface': 5.8996773412711694e-06, 'learning_rate/action_model': 5.899677341271169e-05, 'epoch': 1.48}) +[2026-06-19 15:06:21,863][starVLA.training.train_starvla][INFO] - Step 2337, Loss: {'train/loss': 0.2212875336408615, 'train/grad_norm_pre_clip': 11.968707084655762, 'timing/data': 0.01949247100856155, 'timing/model': 1.1420744290808216, 'learning_rate/qwen_vl_interface': 5.896660406598059e-06, 'learning_rate/action_model': 5.896660406598059e-05, 'epoch': 1.48}) +[2026-06-19 15:06:26,495][starVLA.training.train_starvla][INFO] - Step 2338, Loss: {'train/loss': 0.1385708749294281, 'train/grad_norm_pre_clip': 5.406828880310059, 'timing/data': 0.01634004502557218, 'timing/model': 1.1522661459166557, 'learning_rate/qwen_vl_interface': 5.893643206107283e-06, 'learning_rate/action_model': 5.893643206107282e-05, 'epoch': 1.48}) +[2026-06-19 15:06:31,016][starVLA.training.train_starvla][INFO] - Step 2339, Loss: {'train/loss': 0.2317228764295578, 'train/grad_norm_pre_clip': 11.857024192810059, 'timing/data': 0.02435701701324433, 'timing/model': 1.1510220809141174, 'learning_rate/qwen_vl_interface': 5.890625741039096e-06, 'learning_rate/action_model': 5.890625741039095e-05, 'epoch': 1.48}) +[2026-06-19 15:06:35,571][starVLA.training.train_starvla][INFO] - Step 2340, Loss: {'train/loss': 0.17721223831176758, 'train/grad_norm_pre_clip': 9.317501068115234, 'timing/data': 0.01786797388922423, 'timing/model': 1.14225637097843, 'learning_rate/qwen_vl_interface': 5.887608012633864e-06, 'learning_rate/action_model': 5.887608012633864e-05, 'epoch': 1.48}) +[2026-06-19 15:06:40,159][starVLA.training.train_starvla][INFO] - Step 2341, Loss: {'train/loss': 0.21952885389328003, 'train/grad_norm_pre_clip': 7.731049537658691, 'timing/data': 0.018817918957211077, 'timing/model': 1.1459675499936566, 'learning_rate/qwen_vl_interface': 5.884590022132064e-06, 'learning_rate/action_model': 5.884590022132064e-05, 'epoch': 1.48}) +[2026-06-19 15:06:44,749][starVLA.training.train_starvla][INFO] - Step 2342, Loss: {'train/loss': 0.14349202811717987, 'train/grad_norm_pre_clip': 8.052396774291992, 'timing/data': 0.017387204919941723, 'timing/model': 1.317000512033701, 'learning_rate/qwen_vl_interface': 5.881571770774278e-06, 'learning_rate/action_model': 5.881571770774278e-05, 'epoch': 1.49}) +[2026-06-19 15:06:49,168][starVLA.training.train_starvla][INFO] - Step 2343, Loss: {'train/loss': 0.08056142181158066, 'train/grad_norm_pre_clip': 4.535969257354736, 'timing/data': 0.017885521054267883, 'timing/model': 1.1492685690755025, 'learning_rate/qwen_vl_interface': 5.878553259801192e-06, 'learning_rate/action_model': 5.878553259801192e-05, 'epoch': 1.49}) +[2026-06-19 15:06:53,749][starVLA.training.train_starvla][INFO] - Step 2344, Loss: {'train/loss': 0.08487097173929214, 'train/grad_norm_pre_clip': 4.714338779449463, 'timing/data': 0.017660451936535537, 'timing/model': 1.14481688500382, 'learning_rate/qwen_vl_interface': 5.8755344904536025e-06, 'learning_rate/action_model': 5.8755344904536025e-05, 'epoch': 1.49}) +[2026-06-19 15:06:58,333][starVLA.training.train_starvla][INFO] - Step 2345, Loss: {'train/loss': 0.3095042109489441, 'train/grad_norm_pre_clip': 10.903177261352539, 'timing/data': 0.018290406093001366, 'timing/model': 1.1507765799760818, 'learning_rate/qwen_vl_interface': 5.872515463972415e-06, 'learning_rate/action_model': 5.8725154639724146e-05, 'epoch': 1.49}) +[2026-06-19 15:07:02,926][starVLA.training.train_starvla][INFO] - Step 2346, Loss: {'train/loss': 0.2131780982017517, 'train/grad_norm_pre_clip': 10.123295783996582, 'timing/data': 0.019045177032239735, 'timing/model': 1.153957166010514, 'learning_rate/qwen_vl_interface': 5.869496181598633e-06, 'learning_rate/action_model': 5.869496181598633e-05, 'epoch': 1.49}) +[2026-06-19 15:07:07,537][starVLA.training.train_starvla][INFO] - Step 2347, Loss: {'train/loss': 0.16348041594028473, 'train/grad_norm_pre_clip': 8.072420120239258, 'timing/data': 0.02449037099722773, 'timing/model': 1.1480123579967767, 'learning_rate/qwen_vl_interface': 5.866476644573371e-06, 'learning_rate/action_model': 5.866476644573371e-05, 'epoch': 1.49}) +[2026-06-19 15:07:12,116][starVLA.training.train_starvla][INFO] - Step 2348, Loss: {'train/loss': 0.2872011959552765, 'train/grad_norm_pre_clip': 38.04018783569336, 'timing/data': 0.017901582992635667, 'timing/model': 1.3018203839892522, 'learning_rate/qwen_vl_interface': 5.863456854137848e-06, 'learning_rate/action_model': 5.8634568541378474e-05, 'epoch': 1.49}) +[2026-06-19 15:07:16,630][starVLA.training.train_starvla][INFO] - Step 2349, Loss: {'train/loss': 0.415558397769928, 'train/grad_norm_pre_clip': 15.836933135986328, 'timing/data': 0.019030775991268456, 'timing/model': 1.2464227570453659, 'learning_rate/qwen_vl_interface': 5.860436811533383e-06, 'learning_rate/action_model': 5.860436811533383e-05, 'epoch': 1.49}) +[2026-06-19 15:07:21,066][starVLA.training.train_starvla][INFO] - Step 2350, Loss: {'train/loss': 0.3675572872161865, 'train/grad_norm_pre_clip': 14.587787628173828, 'timing/data': 0.017151310108602047, 'timing/model': 1.1497879659291357, 'learning_rate/qwen_vl_interface': 5.857416518001404e-06, 'learning_rate/action_model': 5.857416518001404e-05, 'epoch': 1.49}) +[2026-06-19 15:07:25,660][starVLA.training.train_starvla][INFO] - Step 2351, Loss: {'train/loss': 0.19736425578594208, 'train/grad_norm_pre_clip': 9.880751609802246, 'timing/data': 0.01817506505176425, 'timing/model': 1.1491980110295117, 'learning_rate/qwen_vl_interface': 5.854395974783439e-06, 'learning_rate/action_model': 5.8543959747834386e-05, 'epoch': 1.49}) +[2026-06-19 15:07:30,254][starVLA.training.train_starvla][INFO] - Step 2352, Loss: {'train/loss': 0.2126937210559845, 'train/grad_norm_pre_clip': 12.135225296020508, 'timing/data': 0.017712629050947726, 'timing/model': 1.1417887270217761, 'learning_rate/qwen_vl_interface': 5.851375183121119e-06, 'learning_rate/action_model': 5.8513751831211186e-05, 'epoch': 1.49}) +[2026-06-19 15:07:34,854][starVLA.training.train_starvla][INFO] - Step 2353, Loss: {'train/loss': 0.26088500022888184, 'train/grad_norm_pre_clip': 19.20050621032715, 'timing/data': 0.016939066001214087, 'timing/model': 1.142865275964141, 'learning_rate/qwen_vl_interface': 5.848354144256175e-06, 'learning_rate/action_model': 5.8483541442561754e-05, 'epoch': 1.49}) +[2026-06-19 15:07:39,486][starVLA.training.train_starvla][INFO] - Step 2354, Loss: {'train/loss': 0.2157474309206009, 'train/grad_norm_pre_clip': 12.143906593322754, 'timing/data': 0.017097768024541438, 'timing/model': 1.3024123620707542, 'learning_rate/qwen_vl_interface': 5.845332859430446e-06, 'learning_rate/action_model': 5.845332859430446e-05, 'epoch': 1.49}) +[2026-06-19 15:07:43,917][starVLA.training.train_starvla][INFO] - Step 2355, Loss: {'train/loss': 0.2958393692970276, 'train/grad_norm_pre_clip': 10.019488334655762, 'timing/data': 0.017896194010972977, 'timing/model': 1.1444614809006453, 'learning_rate/qwen_vl_interface': 5.842311329885866e-06, 'learning_rate/action_model': 5.842311329885866e-05, 'epoch': 1.49}) +[2026-06-19 15:07:48,483][starVLA.training.train_starvla][INFO] - Step 2356, Loss: {'train/loss': 0.40497973561286926, 'train/grad_norm_pre_clip': 12.400689125061035, 'timing/data': 0.017386416904628277, 'timing/model': 1.1422481629997492, 'learning_rate/qwen_vl_interface': 5.839289556864472e-06, 'learning_rate/action_model': 5.8392895568644714e-05, 'epoch': 1.49}) +[2026-06-19 15:07:52,966][starVLA.training.train_starvla][INFO] - Step 2357, Loss: {'train/loss': 0.08511289954185486, 'train/grad_norm_pre_clip': 7.745199680328369, 'timing/data': 0.018486212007701397, 'timing/model': 1.1407443899661303, 'learning_rate/qwen_vl_interface': 5.836267541608398e-06, 'learning_rate/action_model': 5.836267541608398e-05, 'epoch': 1.49}) +[2026-06-19 15:07:57,466][starVLA.training.train_starvla][INFO] - Step 2358, Loss: {'train/loss': 0.23134154081344604, 'train/grad_norm_pre_clip': 11.651463508605957, 'timing/data': 0.017590853036381304, 'timing/model': 1.1445994649548084, 'learning_rate/qwen_vl_interface': 5.833245285359887e-06, 'learning_rate/action_model': 5.833245285359886e-05, 'epoch': 1.5}) +[2026-06-19 15:08:02,001][starVLA.training.train_starvla][INFO] - Step 2359, Loss: {'train/loss': 0.19835332036018372, 'train/grad_norm_pre_clip': 10.581582069396973, 'timing/data': 0.018422254011966288, 'timing/model': 1.148185789003037, 'learning_rate/qwen_vl_interface': 5.830222789361268e-06, 'learning_rate/action_model': 5.8302227893612684e-05, 'epoch': 1.5}) +[2026-06-19 15:08:06,600][starVLA.training.train_starvla][INFO] - Step 2360, Loss: {'train/loss': 0.13534770905971527, 'train/grad_norm_pre_clip': 8.17822551727295, 'timing/data': 0.017163782031275332, 'timing/model': 1.3259099969873205, 'learning_rate/qwen_vl_interface': 5.8272000548549776e-06, 'learning_rate/action_model': 5.827200054854978e-05, 'epoch': 1.5}) +[2026-06-19 15:08:11,022][starVLA.training.train_starvla][INFO] - Step 2361, Loss: {'train/loss': 0.24176058173179626, 'train/grad_norm_pre_clip': 7.38847017288208, 'timing/data': 0.01917415193747729, 'timing/model': 1.1428163319360465, 'learning_rate/qwen_vl_interface': 5.824177083083547e-06, 'learning_rate/action_model': 5.8241770830835475e-05, 'epoch': 1.5}) +[2026-06-19 15:08:15,598][starVLA.training.train_starvla][INFO] - Step 2362, Loss: {'train/loss': 0.3490196466445923, 'train/grad_norm_pre_clip': 9.889500617980957, 'timing/data': 0.017766824923455715, 'timing/model': 1.1436721119098365, 'learning_rate/qwen_vl_interface': 5.8211538752896056e-06, 'learning_rate/action_model': 5.821153875289606e-05, 'epoch': 1.5}) +[2026-06-19 15:08:20,186][starVLA.training.train_starvla][INFO] - Step 2363, Loss: {'train/loss': 0.2306002825498581, 'train/grad_norm_pre_clip': 13.173578262329102, 'timing/data': 0.018893202068284154, 'timing/model': 1.1436284499941394, 'learning_rate/qwen_vl_interface': 5.8181304327158814e-06, 'learning_rate/action_model': 5.8181304327158814e-05, 'epoch': 1.5}) +[2026-06-19 15:08:24,772][starVLA.training.train_starvla][INFO] - Step 2364, Loss: {'train/loss': 0.301658034324646, 'train/grad_norm_pre_clip': 11.510392189025879, 'timing/data': 0.019212392042391002, 'timing/model': 1.1494256539735943, 'learning_rate/qwen_vl_interface': 5.8151067566051965e-06, 'learning_rate/action_model': 5.815106756605196e-05, 'epoch': 1.5}) +[2026-06-19 15:08:29,358][starVLA.training.train_starvla][INFO] - Step 2365, Loss: {'train/loss': 0.26128825545310974, 'train/grad_norm_pre_clip': 11.841288566589355, 'timing/data': 0.018313682987354696, 'timing/model': 1.3192652630386874, 'learning_rate/qwen_vl_interface': 5.81208284820047e-06, 'learning_rate/action_model': 5.81208284820047e-05, 'epoch': 1.5}) +[2026-06-19 15:08:33,776][starVLA.training.train_starvla][INFO] - Step 2366, Loss: {'train/loss': 0.12552256882190704, 'train/grad_norm_pre_clip': 7.187675476074219, 'timing/data': 0.01847378397360444, 'timing/model': 1.145774457952939, 'learning_rate/qwen_vl_interface': 5.809058708744716e-06, 'learning_rate/action_model': 5.809058708744716e-05, 'epoch': 1.5}) +[2026-06-19 15:08:38,390][starVLA.training.train_starvla][INFO] - Step 2367, Loss: {'train/loss': 0.1559126079082489, 'train/grad_norm_pre_clip': 4.89302396774292, 'timing/data': 0.018319108989089727, 'timing/model': 1.157637853990309, 'learning_rate/qwen_vl_interface': 5.8060343394810445e-06, 'learning_rate/action_model': 5.806034339481045e-05, 'epoch': 1.5}) +[2026-06-19 15:08:42,967][starVLA.training.train_starvla][INFO] - Step 2368, Loss: {'train/loss': 0.21582813560962677, 'train/grad_norm_pre_clip': 9.91991138458252, 'timing/data': 0.018291571061126888, 'timing/model': 1.1463559449184686, 'learning_rate/qwen_vl_interface': 5.803009741652659e-06, 'learning_rate/action_model': 5.8030097416526585e-05, 'epoch': 1.5}) +[2026-06-19 15:08:47,556][starVLA.training.train_starvla][INFO] - Step 2369, Loss: {'train/loss': 0.2692151963710785, 'train/grad_norm_pre_clip': 6.416186332702637, 'timing/data': 0.017503776005469263, 'timing/model': 1.1465129300486296, 'learning_rate/qwen_vl_interface': 5.799984916502859e-06, 'learning_rate/action_model': 5.799984916502859e-05, 'epoch': 1.5}) +[2026-06-19 15:08:52,130][starVLA.training.train_starvla][INFO] - Step 2370, Loss: {'train/loss': 0.24146687984466553, 'train/grad_norm_pre_clip': 9.121519088745117, 'timing/data': 0.017728491919115186, 'timing/model': 1.1400759420357645, 'learning_rate/qwen_vl_interface': 5.7969598652750334e-06, 'learning_rate/action_model': 5.796959865275033e-05, 'epoch': 1.5}) +[2026-06-19 15:08:56,712][starVLA.training.train_starvla][INFO] - Step 2371, Loss: {'train/loss': 0.1496875137090683, 'train/grad_norm_pre_clip': 6.839236259460449, 'timing/data': 0.01781559595838189, 'timing/model': 1.3159807329066098, 'learning_rate/qwen_vl_interface': 5.793934589212666e-06, 'learning_rate/action_model': 5.793934589212666e-05, 'epoch': 1.5}) +[2026-06-19 15:09:01,230][starVLA.training.train_starvla][INFO] - Step 2372, Loss: {'train/loss': 0.08948299288749695, 'train/grad_norm_pre_clip': 4.293489933013916, 'timing/data': 0.018517203046940267, 'timing/model': 1.2376053319312632, 'learning_rate/qwen_vl_interface': 5.7909090895593375e-06, 'learning_rate/action_model': 5.790909089559338e-05, 'epoch': 1.5}) +[2026-06-19 15:09:05,650][starVLA.training.train_starvla][INFO] - Step 2373, Loss: {'train/loss': 0.14817897975444794, 'train/grad_norm_pre_clip': 7.227459907531738, 'timing/data': 0.017527916003018618, 'timing/model': 1.1491233629640192, 'learning_rate/qwen_vl_interface': 5.7878833675587135e-06, 'learning_rate/action_model': 5.787883367558713e-05, 'epoch': 1.5}) +[2026-06-19 15:09:10,239][starVLA.training.train_starvla][INFO] - Step 2374, Loss: {'train/loss': 0.15154926478862762, 'train/grad_norm_pre_clip': 7.396459579467773, 'timing/data': 0.018936634063720703, 'timing/model': 1.151419791043736, 'learning_rate/qwen_vl_interface': 5.784857424454556e-06, 'learning_rate/action_model': 5.7848574244545563e-05, 'epoch': 1.51}) +[2026-06-19 15:09:14,839][starVLA.training.train_starvla][INFO] - Step 2375, Loss: {'train/loss': 0.19993442296981812, 'train/grad_norm_pre_clip': 10.595954895019531, 'timing/data': 0.01884747901931405, 'timing/model': 1.1460444330004975, 'learning_rate/qwen_vl_interface': 5.7818312614907125e-06, 'learning_rate/action_model': 5.781831261490712e-05, 'epoch': 1.51}) +[2026-06-19 15:09:19,417][starVLA.training.train_starvla][INFO] - Step 2376, Loss: {'train/loss': 0.15052850544452667, 'train/grad_norm_pre_clip': 8.305627822875977, 'timing/data': 0.01857549394480884, 'timing/model': 1.1462591560557485, 'learning_rate/qwen_vl_interface': 5.778804879911129e-06, 'learning_rate/action_model': 5.778804879911128e-05, 'epoch': 1.51}) +[2026-06-19 15:09:24,011][starVLA.training.train_starvla][INFO] - Step 2377, Loss: {'train/loss': 0.22486548125743866, 'train/grad_norm_pre_clip': 10.946687698364258, 'timing/data': 0.0175528860418126, 'timing/model': 1.3066409300081432, 'learning_rate/qwen_vl_interface': 5.775778280959832e-06, 'learning_rate/action_model': 5.7757782809598315e-05, 'epoch': 1.51}) +[2026-06-19 15:09:28,423][starVLA.training.train_starvla][INFO] - Step 2378, Loss: {'train/loss': 0.14963126182556152, 'train/grad_norm_pre_clip': 11.66549301147461, 'timing/data': 0.018115309067070484, 'timing/model': 1.144207076053135, 'learning_rate/qwen_vl_interface': 5.7727514658809435e-06, 'learning_rate/action_model': 5.772751465880944e-05, 'epoch': 1.51}) +[2026-06-19 15:09:33,008][starVLA.training.train_starvla][INFO] - Step 2379, Loss: {'train/loss': 0.10666428506374359, 'train/grad_norm_pre_clip': 10.038222312927246, 'timing/data': 0.018168710987083614, 'timing/model': 1.1472949380986392, 'learning_rate/qwen_vl_interface': 5.769724435918676e-06, 'learning_rate/action_model': 5.7697244359186755e-05, 'epoch': 1.51}) +[2026-06-19 15:09:37,521][starVLA.training.train_starvla][INFO] - Step 2380, Loss: {'train/loss': 0.25312283635139465, 'train/grad_norm_pre_clip': 12.286212921142578, 'timing/data': 0.018522661994211376, 'timing/model': 1.1436416520737112, 'learning_rate/qwen_vl_interface': 5.766697192317323e-06, 'learning_rate/action_model': 5.766697192317323e-05, 'epoch': 1.51}) +[2026-06-19 15:09:42,031][starVLA.training.train_starvla][INFO] - Step 2381, Loss: {'train/loss': 0.2944778800010681, 'train/grad_norm_pre_clip': 16.048696517944336, 'timing/data': 0.018490326940082014, 'timing/model': 1.1476237720344216, 'learning_rate/qwen_vl_interface': 5.763669736321271e-06, 'learning_rate/action_model': 5.763669736321271e-05, 'epoch': 1.51}) +[2026-06-19 15:09:46,570][starVLA.training.train_starvla][INFO] - Step 2382, Loss: {'train/loss': 0.15990261733531952, 'train/grad_norm_pre_clip': 5.563838005065918, 'timing/data': 0.018087988020852208, 'timing/model': 1.15115865401458, 'learning_rate/qwen_vl_interface': 5.760642069174994e-06, 'learning_rate/action_model': 5.760642069174994e-05, 'epoch': 1.51}) +[2026-06-19 15:09:51,157][starVLA.training.train_starvla][INFO] - Step 2383, Loss: {'train/loss': 0.06277889758348465, 'train/grad_norm_pre_clip': 6.345421314239502, 'timing/data': 0.019846362993121147, 'timing/model': 1.314106722944416, 'learning_rate/qwen_vl_interface': 5.757614192123051e-06, 'learning_rate/action_model': 5.757614192123051e-05, 'epoch': 1.51}) +[2026-06-19 15:09:55,582][starVLA.training.train_starvla][INFO] - Step 2384, Loss: {'train/loss': 0.3503130078315735, 'train/grad_norm_pre_clip': 24.804109573364258, 'timing/data': 0.01864502194803208, 'timing/model': 1.142490835976787, 'learning_rate/qwen_vl_interface': 5.754586106410089e-06, 'learning_rate/action_model': 5.754586106410088e-05, 'epoch': 1.51}) +[2026-06-19 15:10:00,138][starVLA.training.train_starvla][INFO] - Step 2385, Loss: {'train/loss': 0.18601572513580322, 'train/grad_norm_pre_clip': 6.197558879852295, 'timing/data': 0.018629272002726793, 'timing/model': 1.144014906953089, 'learning_rate/qwen_vl_interface': 5.751557813280838e-06, 'learning_rate/action_model': 5.751557813280838e-05, 'epoch': 1.51}) +[2026-06-19 15:10:04,713][starVLA.training.train_starvla][INFO] - Step 2386, Loss: {'train/loss': 0.12489289790391922, 'train/grad_norm_pre_clip': 6.04661750793457, 'timing/data': 0.019464373006485403, 'timing/model': 1.1520778998965397, 'learning_rate/qwen_vl_interface': 5.7485293139801146e-06, 'learning_rate/action_model': 5.7485293139801146e-05, 'epoch': 1.51}) +[2026-06-19 15:10:09,295][starVLA.training.train_starvla][INFO] - Step 2387, Loss: {'train/loss': 0.28310903906822205, 'train/grad_norm_pre_clip': 14.669711112976074, 'timing/data': 0.018191413022577763, 'timing/model': 1.1414735639700666, 'learning_rate/qwen_vl_interface': 5.745500609752821e-06, 'learning_rate/action_model': 5.745500609752821e-05, 'epoch': 1.51}) +[2026-06-19 15:10:13,870][starVLA.training.train_starvla][INFO] - Step 2388, Loss: {'train/loss': 0.14559800922870636, 'train/grad_norm_pre_clip': 10.409971237182617, 'timing/data': 0.01868333399761468, 'timing/model': 1.3090833419701084, 'learning_rate/qwen_vl_interface': 5.742471701843943e-06, 'learning_rate/action_model': 5.742471701843943e-05, 'epoch': 1.51}) +[2026-06-19 15:10:18,278][starVLA.training.train_starvla][INFO] - Step 2389, Loss: {'train/loss': 0.14973819255828857, 'train/grad_norm_pre_clip': 10.375380516052246, 'timing/data': 0.017863505985587835, 'timing/model': 1.141098075080663, 'learning_rate/qwen_vl_interface': 5.739442591498553e-06, 'learning_rate/action_model': 5.739442591498553e-05, 'epoch': 1.52}) +[2026-06-19 15:10:22,852][starVLA.training.train_starvla][INFO] - Step 2390, Loss: {'train/loss': 0.13274291157722473, 'train/grad_norm_pre_clip': 5.520857334136963, 'timing/data': 0.019446230959147215, 'timing/model': 1.1430338399950415, 'learning_rate/qwen_vl_interface': 5.7364132799617985e-06, 'learning_rate/action_model': 5.7364132799617984e-05, 'epoch': 1.52}) +[2026-06-19 15:10:27,420][starVLA.training.train_starvla][INFO] - Step 2391, Loss: {'train/loss': 0.14245183765888214, 'train/grad_norm_pre_clip': 6.578185081481934, 'timing/data': 0.018671059980988503, 'timing/model': 1.1406762569677085, 'learning_rate/qwen_vl_interface': 5.733383768478923e-06, 'learning_rate/action_model': 5.7333837684789225e-05, 'epoch': 1.52}) +[2026-06-19 15:10:32,000][starVLA.training.train_starvla][INFO] - Step 2392, Loss: {'train/loss': 0.5295029282569885, 'train/grad_norm_pre_clip': 18.171010971069336, 'timing/data': 0.01829144498333335, 'timing/model': 1.149356308975257, 'learning_rate/qwen_vl_interface': 5.730354058295234e-06, 'learning_rate/action_model': 5.730354058295234e-05, 'epoch': 1.52}) +[2026-06-19 15:10:36,580][starVLA.training.train_starvla][INFO] - Step 2393, Loss: {'train/loss': 0.2175905704498291, 'train/grad_norm_pre_clip': 12.108713150024414, 'timing/data': 0.01784379791934043, 'timing/model': 1.145644971053116, 'learning_rate/qwen_vl_interface': 5.72732415065614e-06, 'learning_rate/action_model': 5.727324150656139e-05, 'epoch': 1.52}) +[2026-06-19 15:10:41,171][starVLA.training.train_starvla][INFO] - Step 2394, Loss: {'train/loss': 0.10099340975284576, 'train/grad_norm_pre_clip': 4.149021148681641, 'timing/data': 0.020215117023326457, 'timing/model': 1.3158647749805823, 'learning_rate/qwen_vl_interface': 5.724294046807115e-06, 'learning_rate/action_model': 5.724294046807115e-05, 'epoch': 1.52}) +[2026-06-19 15:10:45,682][starVLA.training.train_starvla][INFO] - Step 2395, Loss: {'train/loss': 0.05155530571937561, 'train/grad_norm_pre_clip': 5.173791408538818, 'timing/data': 0.018311445019207895, 'timing/model': 1.2380942680174485, 'learning_rate/qwen_vl_interface': 5.7212637479937234e-06, 'learning_rate/action_model': 5.721263747993724e-05, 'epoch': 1.52}) +[2026-06-19 15:10:50,112][starVLA.training.train_starvla][INFO] - Step 2396, Loss: {'train/loss': 0.1933421790599823, 'train/grad_norm_pre_clip': 4.562290668487549, 'timing/data': 0.017850612057372928, 'timing/model': 1.1531144860200584, 'learning_rate/qwen_vl_interface': 5.718233255461607e-06, 'learning_rate/action_model': 5.718233255461607e-05, 'epoch': 1.52}) +[2026-06-19 15:10:54,703][starVLA.training.train_starvla][INFO] - Step 2397, Loss: {'train/loss': 0.11420857161283493, 'train/grad_norm_pre_clip': 6.52384614944458, 'timing/data': 0.016986513044685125, 'timing/model': 1.159914767020382, 'learning_rate/qwen_vl_interface': 5.715202570456486e-06, 'learning_rate/action_model': 5.715202570456486e-05, 'epoch': 1.52}) +[2026-06-19 15:10:59,278][starVLA.training.train_starvla][INFO] - Step 2398, Loss: {'train/loss': 0.20861463248729706, 'train/grad_norm_pre_clip': 28.081769943237305, 'timing/data': 0.018834174959920347, 'timing/model': 1.143014539964497, 'learning_rate/qwen_vl_interface': 5.7121716942241615e-06, 'learning_rate/action_model': 5.712171694224161e-05, 'epoch': 1.52}) +[2026-06-19 15:11:03,859][starVLA.training.train_starvla][INFO] - Step 2399, Loss: {'train/loss': 0.3133850693702698, 'train/grad_norm_pre_clip': 11.024986267089844, 'timing/data': 0.017253609956242144, 'timing/model': 1.1458735720952973, 'learning_rate/qwen_vl_interface': 5.709140628010509e-06, 'learning_rate/action_model': 5.7091406280105085e-05, 'epoch': 1.52}) +[2026-06-19 15:11:08,441][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 2400 +[2026-06-19 15:14:01,779][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 2400 +[2026-06-19 15:14:01,780][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 2400 +[2026-06-19 15:17:14,804][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 2400 +[2026-06-19 15:17:14,805][starVLA.training.train_starvla][INFO] - Step 2400, Loss: {'train/loss': 0.1950492262840271, 'train/grad_norm_pre_clip': 8.210260391235352, 'eval/loss': 0.3625809848308563, 'eval/latency_0/loss': 0.11488927900791168, 'eval/latency_1/loss': 0.08455581218004227, 'eval/latency_2/loss': 0.12112334370613098, 'eval/latency_3/loss': 0.5594510436058044, 'eval/latency_4/loss': 0.9503487348556519, 'eval/flappy/loss': 0.3625809848308563, 'eval/flappy/latency_0/loss': 0.11488927900791168, 'eval/flappy/latency_1/loss': 0.08455581218004227, 'eval/flappy/latency_2/loss': 0.12112334370613098, 'eval/flappy/latency_3/loss': 0.5594510436058044, 'eval/flappy/latency_4/loss': 0.9503487348556519, 'eval/flap_precision': 0.8424242424242424, 'eval/flap_recall': 0.5054545454545455, 'eval/flap_f1': 0.6318181818181818, 'eval/flap_rate_model': 0.04583333333333333, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.6318181818181818, 'eval/latency_0/flap_precision': 0.8424242424242424, 'eval/latency_0/flap_recall': 0.5054545454545455, 'eval/latency_0/flap_f1': 0.6318181818181818, 'eval/latency_0/flap_rate_model': 0.04583333333333333, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.6318181818181818, 'timing/data': 0.017788968980312347, 'timing/model': 1.3053399600321427, 'learning_rate/qwen_vl_interface': 5.706109373061489e-06, 'learning_rate/action_model': 5.706109373061489e-05, 'epoch': 1.52}) +[2026-06-19 15:17:14,805][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_2400_state +[2026-06-19 15:17:14,808][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-19 15:17:18,198][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_2400_state/model.safetensors +[2026-06-19 15:17:25,339][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_2400_state/optimizer.bin +[2026-06-19 15:17:25,339][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_2400_state/sampler.bin +[2026-06-19 15:17:25,340][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_2400_state/sampler_1.bin +[2026-06-19 15:17:25,340][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_2400_state/random_states_0.pkl +[2026-06-19 15:17:25,366][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2400 +[2026-06-19 15:17:25,367][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-19 15:17:25,418][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-19 15:17:29,886][starVLA.training.train_starvla][INFO] - Step 2401, Loss: {'train/loss': 0.0846627801656723, 'train/grad_norm_pre_clip': 8.996208190917969, 'timing/data': 0.01785906800068915, 'timing/model': 1.1343334609409794, 'learning_rate/qwen_vl_interface': 5.703077930623134e-06, 'learning_rate/action_model': 5.7030779306231344e-05, 'epoch': 1.52}) +[2026-06-19 15:17:34,418][starVLA.training.train_starvla][INFO] - Step 2402, Loss: {'train/loss': 0.18761491775512695, 'train/grad_norm_pre_clip': 9.725784301757812, 'timing/data': 0.01730459090322256, 'timing/model': 1.3018722059205174, 'learning_rate/qwen_vl_interface': 5.7000463019415565e-06, 'learning_rate/action_model': 5.7000463019415563e-05, 'epoch': 1.52}) +[2026-06-19 15:17:38,765][starVLA.training.train_starvla][INFO] - Step 2403, Loss: {'train/loss': 0.06087963283061981, 'train/grad_norm_pre_clip': 7.372532367706299, 'timing/data': 0.01608985394705087, 'timing/model': 1.129250800004229, 'learning_rate/qwen_vl_interface': 5.697014488262945e-06, 'learning_rate/action_model': 5.697014488262945e-05, 'epoch': 1.52}) +[2026-06-19 15:17:43,310][starVLA.training.train_starvla][INFO] - Step 2404, Loss: {'train/loss': 0.1537124067544937, 'train/grad_norm_pre_clip': 13.41224479675293, 'timing/data': 0.016054912004619837, 'timing/model': 1.1398539290530607, 'learning_rate/qwen_vl_interface': 5.693982490833561e-06, 'learning_rate/action_model': 5.693982490833561e-05, 'epoch': 1.52}) +[2026-06-19 15:17:47,839][starVLA.training.train_starvla][INFO] - Step 2405, Loss: {'train/loss': 0.2640479505062103, 'train/grad_norm_pre_clip': 10.372036933898926, 'timing/data': 0.01783435302786529, 'timing/model': 1.1291580339893699, 'learning_rate/qwen_vl_interface': 5.690950310899745e-06, 'learning_rate/action_model': 5.6909503108997456e-05, 'epoch': 1.53}) +[2026-06-19 15:17:52,367][starVLA.training.train_starvla][INFO] - Step 2406, Loss: {'train/loss': 0.15035638213157654, 'train/grad_norm_pre_clip': 12.644006729125977, 'timing/data': 0.017289442010223866, 'timing/model': 1.1375731009757146, 'learning_rate/qwen_vl_interface': 5.687917949707913e-06, 'learning_rate/action_model': 5.687917949707913e-05, 'epoch': 1.53}) +[2026-06-19 15:17:56,888][starVLA.training.train_starvla][INFO] - Step 2407, Loss: {'train/loss': 0.174063041806221, 'train/grad_norm_pre_clip': 10.466367721557617, 'timing/data': 0.016624837066046894, 'timing/model': 1.1314109290251508, 'learning_rate/qwen_vl_interface': 5.684885408504555e-06, 'learning_rate/action_model': 5.684885408504554e-05, 'epoch': 1.53}) +[2026-06-19 15:18:01,444][starVLA.training.train_starvla][INFO] - Step 2408, Loss: {'train/loss': 0.24899797141551971, 'train/grad_norm_pre_clip': 9.001568794250488, 'timing/data': 0.016758851008489728, 'timing/model': 1.3132959130452946, 'learning_rate/qwen_vl_interface': 5.681852688536231e-06, 'learning_rate/action_model': 5.681852688536231e-05, 'epoch': 1.53}) +[2026-06-19 15:18:05,826][starVLA.training.train_starvla][INFO] - Step 2409, Loss: {'train/loss': 0.20786234736442566, 'train/grad_norm_pre_clip': 12.272466659545898, 'timing/data': 0.016971998033113778, 'timing/model': 1.139411638956517, 'learning_rate/qwen_vl_interface': 5.67881979104958e-06, 'learning_rate/action_model': 5.67881979104958e-05, 'epoch': 1.53}) +[2026-06-19 15:18:10,387][starVLA.training.train_starvla][INFO] - Step 2410, Loss: {'train/loss': 0.46204251050949097, 'train/grad_norm_pre_clip': 17.193470001220703, 'timing/data': 0.017919045058079064, 'timing/model': 1.146832215017639, 'learning_rate/qwen_vl_interface': 5.67578671729131e-06, 'learning_rate/action_model': 5.675786717291309e-05, 'epoch': 1.53}) +[2026-06-19 15:18:14,917][starVLA.training.train_starvla][INFO] - Step 2411, Loss: {'train/loss': 0.27627381682395935, 'train/grad_norm_pre_clip': 17.108917236328125, 'timing/data': 0.01682470494415611, 'timing/model': 1.1585021150531247, 'learning_rate/qwen_vl_interface': 5.672753468508204e-06, 'learning_rate/action_model': 5.672753468508204e-05, 'epoch': 1.53}) +[2026-06-19 15:18:19,461][starVLA.training.train_starvla][INFO] - Step 2412, Loss: {'train/loss': 0.08989604562520981, 'train/grad_norm_pre_clip': 6.934350967407227, 'timing/data': 0.017178992042317986, 'timing/model': 1.1342173860175535, 'learning_rate/qwen_vl_interface': 5.6697200459471165e-06, 'learning_rate/action_model': 5.669720045947117e-05, 'epoch': 1.53}) +[2026-06-19 15:18:24,012][starVLA.training.train_starvla][INFO] - Step 2413, Loss: {'train/loss': 0.20553182065486908, 'train/grad_norm_pre_clip': 10.045926094055176, 'timing/data': 0.017146711004897952, 'timing/model': 1.135461327037774, 'learning_rate/qwen_vl_interface': 5.666686450854973e-06, 'learning_rate/action_model': 5.666686450854973e-05, 'epoch': 1.53}) +[2026-06-19 15:18:28,558][starVLA.training.train_starvla][INFO] - Step 2414, Loss: {'train/loss': 0.07356280088424683, 'train/grad_norm_pre_clip': 4.5013604164123535, 'timing/data': 0.01612998207565397, 'timing/model': 1.3022655849345028, 'learning_rate/qwen_vl_interface': 5.663652684478768e-06, 'learning_rate/action_model': 5.6636526844787676e-05, 'epoch': 1.53}) +[2026-06-19 15:18:32,951][starVLA.training.train_starvla][INFO] - Step 2415, Loss: {'train/loss': 0.14452776312828064, 'train/grad_norm_pre_clip': 10.102629661560059, 'timing/data': 0.01806740404572338, 'timing/model': 1.1411721090553328, 'learning_rate/qwen_vl_interface': 5.660618748065569e-06, 'learning_rate/action_model': 5.6606187480655695e-05, 'epoch': 1.53}) +[2026-06-19 15:18:37,516][starVLA.training.train_starvla][INFO] - Step 2416, Loss: {'train/loss': 0.2554139196872711, 'train/grad_norm_pre_clip': 10.081743240356445, 'timing/data': 0.016757583944126964, 'timing/model': 1.1453587691066787, 'learning_rate/qwen_vl_interface': 5.657584642862514e-06, 'learning_rate/action_model': 5.657584642862513e-05, 'epoch': 1.53}) +[2026-06-19 15:18:42,101][starVLA.training.train_starvla][INFO] - Step 2417, Loss: {'train/loss': 0.2942183315753937, 'train/grad_norm_pre_clip': 13.873071670532227, 'timing/data': 0.01669735205359757, 'timing/model': 1.137178171891719, 'learning_rate/qwen_vl_interface': 5.654550370116808e-06, 'learning_rate/action_model': 5.654550370116808e-05, 'epoch': 1.53}) +[2026-06-19 15:18:46,602][starVLA.training.train_starvla][INFO] - Step 2418, Loss: {'train/loss': 0.04477272927761078, 'train/grad_norm_pre_clip': 2.3904991149902344, 'timing/data': 0.016692387987859547, 'timing/model': 1.1403907479252666, 'learning_rate/qwen_vl_interface': 5.651515931075726e-06, 'learning_rate/action_model': 5.651515931075726e-05, 'epoch': 1.53}) +[2026-06-19 15:18:51,167][starVLA.training.train_starvla][INFO] - Step 2419, Loss: {'train/loss': 0.17722930014133453, 'train/grad_norm_pre_clip': 10.72164535522461, 'timing/data': 0.017867340007796884, 'timing/model': 1.1479472239734605, 'learning_rate/qwen_vl_interface': 5.6484813269866126e-06, 'learning_rate/action_model': 5.648481326986612e-05, 'epoch': 1.53}) +[2026-06-19 15:18:55,801][starVLA.training.train_starvla][INFO] - Step 2420, Loss: {'train/loss': 0.49498334527015686, 'train/grad_norm_pre_clip': 17.790996551513672, 'timing/data': 0.017265836941078305, 'timing/model': 1.3450927899684757, 'learning_rate/qwen_vl_interface': 5.645446559096875e-06, 'learning_rate/action_model': 5.645446559096875e-05, 'epoch': 1.53}) +[2026-06-19 15:19:00,223][starVLA.training.train_starvla][INFO] - Step 2421, Loss: {'train/loss': 0.13253508508205414, 'train/grad_norm_pre_clip': 9.537883758544922, 'timing/data': 0.016770974034443498, 'timing/model': 1.1412251719739288, 'learning_rate/qwen_vl_interface': 5.6424116286539965e-06, 'learning_rate/action_model': 5.642411628653996e-05, 'epoch': 1.54}) +[2026-06-19 15:19:04,824][starVLA.training.train_starvla][INFO] - Step 2422, Loss: {'train/loss': 0.37968310713768005, 'train/grad_norm_pre_clip': 20.214860916137695, 'timing/data': 0.017954142997041345, 'timing/model': 1.1463333560386673, 'learning_rate/qwen_vl_interface': 5.639376536905517e-06, 'learning_rate/action_model': 5.6393765369055175e-05, 'epoch': 1.54}) +[2026-06-19 15:19:09,398][starVLA.training.train_starvla][INFO] - Step 2423, Loss: {'train/loss': 0.7274254560470581, 'train/grad_norm_pre_clip': 17.82894515991211, 'timing/data': 0.016730144037865102, 'timing/model': 1.1352159719681367, 'learning_rate/qwen_vl_interface': 5.636341285099055e-06, 'learning_rate/action_model': 5.636341285099055e-05, 'epoch': 1.54}) +[2026-06-19 15:19:13,979][starVLA.training.train_starvla][INFO] - Step 2424, Loss: {'train/loss': 0.22151103615760803, 'train/grad_norm_pre_clip': 11.005921363830566, 'timing/data': 0.017849572002887726, 'timing/model': 1.149844550061971, 'learning_rate/qwen_vl_interface': 5.6333058744822835e-06, 'learning_rate/action_model': 5.633305874482283e-05, 'epoch': 1.54}) +[2026-06-19 15:19:18,600][starVLA.training.train_starvla][INFO] - Step 2425, Loss: {'train/loss': 0.21722598373889923, 'train/grad_norm_pre_clip': 9.330774307250977, 'timing/data': 0.02231212693732232, 'timing/model': 1.1620571980020031, 'learning_rate/qwen_vl_interface': 5.630270306302944e-06, 'learning_rate/action_model': 5.6302703063029436e-05, 'epoch': 1.54}) +[2026-06-19 15:19:23,157][starVLA.training.train_starvla][INFO] - Step 2426, Loss: {'train/loss': 0.23823367059230804, 'train/grad_norm_pre_clip': 17.600969314575195, 'timing/data': 0.01827998599037528, 'timing/model': 1.2955033299513161, 'learning_rate/qwen_vl_interface': 5.627234581808848e-06, 'learning_rate/action_model': 5.627234581808848e-05, 'epoch': 1.54}) +[2026-06-19 15:19:27,581][starVLA.training.train_starvla][INFO] - Step 2427, Loss: {'train/loss': 0.06596598029136658, 'train/grad_norm_pre_clip': 4.138229846954346, 'timing/data': 0.01874045911245048, 'timing/model': 1.1532353030052036, 'learning_rate/qwen_vl_interface': 5.624198702247863e-06, 'learning_rate/action_model': 5.624198702247863e-05, 'epoch': 1.54}) +[2026-06-19 15:19:32,144][starVLA.training.train_starvla][INFO] - Step 2428, Loss: {'train/loss': 0.21546031534671783, 'train/grad_norm_pre_clip': 14.634737014770508, 'timing/data': 0.016235598013736308, 'timing/model': 1.1409076480194926, 'learning_rate/qwen_vl_interface': 5.621162668867925e-06, 'learning_rate/action_model': 5.621162668867925e-05, 'epoch': 1.54}) +[2026-06-19 15:19:36,716][starVLA.training.train_starvla][INFO] - Step 2429, Loss: {'train/loss': 0.29857686161994934, 'train/grad_norm_pre_clip': 10.361019134521484, 'timing/data': 0.016361488960683346, 'timing/model': 1.1413573221070692, 'learning_rate/qwen_vl_interface': 5.618126482917037e-06, 'learning_rate/action_model': 5.618126482917036e-05, 'epoch': 1.54}) +[2026-06-19 15:19:41,251][starVLA.training.train_starvla][INFO] - Step 2430, Loss: {'train/loss': 0.12344025075435638, 'train/grad_norm_pre_clip': 9.296515464782715, 'timing/data': 0.016903779935091734, 'timing/model': 1.1416184920817614, 'learning_rate/qwen_vl_interface': 5.6150901456432546e-06, 'learning_rate/action_model': 5.615090145643255e-05, 'epoch': 1.54}) +[2026-06-19 15:19:45,860][starVLA.training.train_starvla][INFO] - Step 2431, Loss: {'train/loss': 0.10999304801225662, 'train/grad_norm_pre_clip': 4.785009860992432, 'timing/data': 0.016813487047329545, 'timing/model': 1.145782441017218, 'learning_rate/qwen_vl_interface': 5.612053658294705e-06, 'learning_rate/action_model': 5.6120536582947045e-05, 'epoch': 1.54}) +[2026-06-19 15:19:50,463][starVLA.training.train_starvla][INFO] - Step 2432, Loss: {'train/loss': 0.28859174251556396, 'train/grad_norm_pre_clip': 11.498143196105957, 'timing/data': 0.018288433901034296, 'timing/model': 1.3250173140550032, 'learning_rate/qwen_vl_interface': 5.609017022119572e-06, 'learning_rate/action_model': 5.609017022119573e-05, 'epoch': 1.54}) +[2026-06-19 15:19:54,891][starVLA.training.train_starvla][INFO] - Step 2433, Loss: {'train/loss': 0.37189024686813354, 'train/grad_norm_pre_clip': 12.915572166442871, 'timing/data': 0.017132901935838163, 'timing/model': 1.149515985045582, 'learning_rate/qwen_vl_interface': 5.605980238366104e-06, 'learning_rate/action_model': 5.605980238366103e-05, 'epoch': 1.54}) +[2026-06-19 15:19:59,515][starVLA.training.train_starvla][INFO] - Step 2434, Loss: {'train/loss': 0.17114844918251038, 'train/grad_norm_pre_clip': 7.605224609375, 'timing/data': 0.017051097936928272, 'timing/model': 1.1553607160458341, 'learning_rate/qwen_vl_interface': 5.602943308282606e-06, 'learning_rate/action_model': 5.602943308282605e-05, 'epoch': 1.54}) +[2026-06-19 15:20:04,133][starVLA.training.train_starvla][INFO] - Step 2435, Loss: {'train/loss': 0.12535515427589417, 'train/grad_norm_pre_clip': 8.012139320373535, 'timing/data': 0.017558719962835312, 'timing/model': 1.1459382590837777, 'learning_rate/qwen_vl_interface': 5.599906233117445e-06, 'learning_rate/action_model': 5.5999062331174446e-05, 'epoch': 1.54}) +[2026-06-19 15:20:08,666][starVLA.training.train_starvla][INFO] - Step 2436, Loss: {'train/loss': 0.21583311259746552, 'train/grad_norm_pre_clip': 7.966907978057861, 'timing/data': 0.01801437500398606, 'timing/model': 1.1519485459430143, 'learning_rate/qwen_vl_interface': 5.596869014119048e-06, 'learning_rate/action_model': 5.596869014119048e-05, 'epoch': 1.54}) +[2026-06-19 15:20:13,191][starVLA.training.train_starvla][INFO] - Step 2437, Loss: {'train/loss': 0.3268466889858246, 'train/grad_norm_pre_clip': 9.537568092346191, 'timing/data': 0.01698324701283127, 'timing/model': 1.1428840559674427, 'learning_rate/qwen_vl_interface': 5.593831652535902e-06, 'learning_rate/action_model': 5.593831652535902e-05, 'epoch': 1.55}) +[2026-06-19 15:20:17,794][starVLA.training.train_starvla][INFO] - Step 2438, Loss: {'train/loss': 0.2800356149673462, 'train/grad_norm_pre_clip': 13.593955039978027, 'timing/data': 0.017466209013946354, 'timing/model': 1.3255751439137384, 'learning_rate/qwen_vl_interface': 5.590794149616552e-06, 'learning_rate/action_model': 5.590794149616552e-05, 'epoch': 1.55}) +[2026-06-19 15:20:22,209][starVLA.training.train_starvla][INFO] - Step 2439, Loss: {'train/loss': 0.08568748831748962, 'train/grad_norm_pre_clip': 4.17785120010376, 'timing/data': 0.016765093081630766, 'timing/model': 1.1564987959573045, 'learning_rate/qwen_vl_interface': 5.587756506609601e-06, 'learning_rate/action_model': 5.5877565066096006e-05, 'epoch': 1.55}) +[2026-06-19 15:20:26,777][starVLA.training.train_starvla][INFO] - Step 2440, Loss: {'train/loss': 0.26456016302108765, 'train/grad_norm_pre_clip': 12.13991641998291, 'timing/data': 0.0180494919186458, 'timing/model': 1.1414144800510257, 'learning_rate/qwen_vl_interface': 5.584718724763707e-06, 'learning_rate/action_model': 5.584718724763707e-05, 'epoch': 1.55}) +[2026-06-19 15:20:31,366][starVLA.training.train_starvla][INFO] - Step 2441, Loss: {'train/loss': 0.2837439179420471, 'train/grad_norm_pre_clip': 14.461350440979004, 'timing/data': 0.017033216077834368, 'timing/model': 1.1482385220006108, 'learning_rate/qwen_vl_interface': 5.581680805327588e-06, 'learning_rate/action_model': 5.5816808053275885e-05, 'epoch': 1.55}) +[2026-06-19 15:20:35,975][starVLA.training.train_starvla][INFO] - Step 2442, Loss: {'train/loss': 0.17770977318286896, 'train/grad_norm_pre_clip': 8.920430183410645, 'timing/data': 0.017258981009945273, 'timing/model': 1.1588781099999323, 'learning_rate/qwen_vl_interface': 5.578642749550019e-06, 'learning_rate/action_model': 5.578642749550019e-05, 'epoch': 1.55}) +[2026-06-19 15:20:40,607][starVLA.training.train_starvla][INFO] - Step 2443, Loss: {'train/loss': 0.25485095381736755, 'train/grad_norm_pre_clip': 11.825663566589355, 'timing/data': 0.016709769028238952, 'timing/model': 1.3399884409736842, 'learning_rate/qwen_vl_interface': 5.575604558679831e-06, 'learning_rate/action_model': 5.57560455867983e-05, 'epoch': 1.55}) +[2026-06-19 15:20:45,035][starVLA.training.train_starvla][INFO] - Step 2444, Loss: {'train/loss': 0.10558202862739563, 'train/grad_norm_pre_clip': 4.544487476348877, 'timing/data': 0.017225150018930435, 'timing/model': 1.145809412933886, 'learning_rate/qwen_vl_interface': 5.5725662339659085e-06, 'learning_rate/action_model': 5.572566233965908e-05, 'epoch': 1.55}) +[2026-06-19 15:20:49,628][starVLA.training.train_starvla][INFO] - Step 2445, Loss: {'train/loss': 0.24567031860351562, 'train/grad_norm_pre_clip': 15.313959121704102, 'timing/data': 0.016562510980293155, 'timing/model': 1.145293470006436, 'learning_rate/qwen_vl_interface': 5.56952777665719e-06, 'learning_rate/action_model': 5.56952777665719e-05, 'epoch': 1.55}) +[2026-06-19 15:20:54,211][starVLA.training.train_starvla][INFO] - Step 2446, Loss: {'train/loss': 0.38165178894996643, 'train/grad_norm_pre_clip': 19.843238830566406, 'timing/data': 0.01705771090928465, 'timing/model': 1.142187320976518, 'learning_rate/qwen_vl_interface': 5.566489188002671e-06, 'learning_rate/action_model': 5.56648918800267e-05, 'epoch': 1.55}) +[2026-06-19 15:20:58,810][starVLA.training.train_starvla][INFO] - Step 2447, Loss: {'train/loss': 0.12021412700414658, 'train/grad_norm_pre_clip': 8.339924812316895, 'timing/data': 0.016905377968214452, 'timing/model': 1.1496643689461052, 'learning_rate/qwen_vl_interface': 5.563450469251403e-06, 'learning_rate/action_model': 5.563450469251402e-05, 'epoch': 1.55}) +[2026-06-19 15:21:03,400][starVLA.training.train_starvla][INFO] - Step 2448, Loss: {'train/loss': 0.30773594975471497, 'train/grad_norm_pre_clip': 10.256707191467285, 'timing/data': 0.017074243049137294, 'timing/model': 1.1485451969783753, 'learning_rate/qwen_vl_interface': 5.560411621652488e-06, 'learning_rate/action_model': 5.560411621652488e-05, 'epoch': 1.55}) +[2026-06-19 15:21:07,991][starVLA.training.train_starvla][INFO] - Step 2449, Loss: {'train/loss': 0.33166003227233887, 'train/grad_norm_pre_clip': 14.115872383117676, 'timing/data': 0.017040609964169562, 'timing/model': 1.3257483809720725, 'learning_rate/qwen_vl_interface': 5.557372646455081e-06, 'learning_rate/action_model': 5.55737264645508e-05, 'epoch': 1.55}) +[2026-06-19 15:21:12,603][starVLA.training.train_starvla][INFO] - Step 2450, Loss: {'train/loss': 0.13973116874694824, 'train/grad_norm_pre_clip': 5.868298053741455, 'timing/data': 0.017623443040065467, 'timing/model': 1.3258163690334186, 'learning_rate/qwen_vl_interface': 5.554333544908388e-06, 'learning_rate/action_model': 5.5543335449083875e-05, 'epoch': 1.55}) +[2026-06-19 15:21:17,028][starVLA.training.train_starvla][INFO] - Step 2451, Loss: {'train/loss': 0.17238235473632812, 'train/grad_norm_pre_clip': 10.844534873962402, 'timing/data': 0.018133266014046967, 'timing/model': 1.156410400988534, 'learning_rate/qwen_vl_interface': 5.551294318261671e-06, 'learning_rate/action_model': 5.5512943182616705e-05, 'epoch': 1.55}) +[2026-06-19 15:21:21,634][starVLA.training.train_starvla][INFO] - Step 2452, Loss: {'train/loss': 0.21681608259677887, 'train/grad_norm_pre_clip': 24.29270362854004, 'timing/data': 0.017131712986156344, 'timing/model': 1.1453874219441786, 'learning_rate/qwen_vl_interface': 5.548254967764239e-06, 'learning_rate/action_model': 5.5482549677642394e-05, 'epoch': 1.55}) +[2026-06-19 15:21:26,245][starVLA.training.train_starvla][INFO] - Step 2453, Loss: {'train/loss': 0.3462657928466797, 'train/grad_norm_pre_clip': 11.729354858398438, 'timing/data': 0.016271983040496707, 'timing/model': 1.150062526925467, 'learning_rate/qwen_vl_interface': 5.545215494665458e-06, 'learning_rate/action_model': 5.545215494665458e-05, 'epoch': 1.56}) +[2026-06-19 15:21:30,757][starVLA.training.train_starvla][INFO] - Step 2454, Loss: {'train/loss': 0.12904617190361023, 'train/grad_norm_pre_clip': 6.533023357391357, 'timing/data': 0.017406231025233865, 'timing/model': 1.145165344933048, 'learning_rate/qwen_vl_interface': 5.5421759002147376e-06, 'learning_rate/action_model': 5.542175900214738e-05, 'epoch': 1.56}) +[2026-06-19 15:21:35,317][starVLA.training.train_starvla][INFO] - Step 2455, Loss: {'train/loss': 0.091156966984272, 'train/grad_norm_pre_clip': 4.1299872398376465, 'timing/data': 0.018530323053710163, 'timing/model': 1.1492375530069694, 'learning_rate/qwen_vl_interface': 5.53913618566154e-06, 'learning_rate/action_model': 5.53913618566154e-05, 'epoch': 1.56}) +[2026-06-19 15:21:39,937][starVLA.training.train_starvla][INFO] - Step 2456, Loss: {'train/loss': 0.02960955537855625, 'train/grad_norm_pre_clip': 1.817463994026184, 'timing/data': 0.016643372015096247, 'timing/model': 1.337081901030615, 'learning_rate/qwen_vl_interface': 5.536096352255381e-06, 'learning_rate/action_model': 5.536096352255381e-05, 'epoch': 1.56}) +[2026-06-19 15:21:44,381][starVLA.training.train_starvla][INFO] - Step 2457, Loss: {'train/loss': 0.2973743677139282, 'train/grad_norm_pre_clip': 7.212214946746826, 'timing/data': 0.016033106949180365, 'timing/model': 1.1567949970485643, 'learning_rate/qwen_vl_interface': 5.533056401245817e-06, 'learning_rate/action_model': 5.5330564012458164e-05, 'epoch': 1.56}) +[2026-06-19 15:21:48,965][starVLA.training.train_starvla][INFO] - Step 2458, Loss: {'train/loss': 0.2361656278371811, 'train/grad_norm_pre_clip': 10.368739128112793, 'timing/data': 0.017571772914379835, 'timing/model': 1.1431130369892344, 'learning_rate/qwen_vl_interface': 5.530016333882456e-06, 'learning_rate/action_model': 5.5300163338824565e-05, 'epoch': 1.56}) +[2026-06-19 15:21:53,541][starVLA.training.train_starvla][INFO] - Step 2459, Loss: {'train/loss': 0.17722977697849274, 'train/grad_norm_pre_clip': 12.894200325012207, 'timing/data': 0.01619189907796681, 'timing/model': 1.1472280130255967, 'learning_rate/qwen_vl_interface': 5.52697615141496e-06, 'learning_rate/action_model': 5.52697615141496e-05, 'epoch': 1.56}) +[2026-06-19 15:21:58,140][starVLA.training.train_starvla][INFO] - Step 2460, Loss: {'train/loss': 0.15163768827915192, 'train/grad_norm_pre_clip': 8.618319511413574, 'timing/data': 0.017683511949144304, 'timing/model': 1.1593468090286478, 'learning_rate/qwen_vl_interface': 5.52393585509303e-06, 'learning_rate/action_model': 5.52393585509303e-05, 'epoch': 1.56}) +[2026-06-19 15:22:02,618][starVLA.training.train_starvla][INFO] - Step 2461, Loss: {'train/loss': 0.2102220207452774, 'train/grad_norm_pre_clip': 8.60398006439209, 'timing/data': 0.017207197030074894, 'timing/model': 1.1406824430450797, 'learning_rate/qwen_vl_interface': 5.520895446166418e-06, 'learning_rate/action_model': 5.5208954461664175e-05, 'epoch': 1.56}) +[2026-06-19 15:22:07,153][starVLA.training.train_starvla][INFO] - Step 2462, Loss: {'train/loss': 0.12538760900497437, 'train/grad_norm_pre_clip': 42.374019622802734, 'timing/data': 0.01763526105787605, 'timing/model': 1.2597215300193056, 'learning_rate/qwen_vl_interface': 5.517854925884919e-06, 'learning_rate/action_model': 5.517854925884919e-05, 'epoch': 1.56}) +[2026-06-19 15:22:11,583][starVLA.training.train_starvla][INFO] - Step 2463, Loss: {'train/loss': 0.2067253589630127, 'train/grad_norm_pre_clip': 7.206121444702148, 'timing/data': 0.017539512016810477, 'timing/model': 1.1440309330355376, 'learning_rate/qwen_vl_interface': 5.514814295498379e-06, 'learning_rate/action_model': 5.514814295498378e-05, 'epoch': 1.56}) +[2026-06-19 15:22:16,158][starVLA.training.train_starvla][INFO] - Step 2464, Loss: {'train/loss': 0.28466546535491943, 'train/grad_norm_pre_clip': 10.124573707580566, 'timing/data': 0.018131067976355553, 'timing/model': 1.1409981469623744, 'learning_rate/qwen_vl_interface': 5.5117735562566846e-06, 'learning_rate/action_model': 5.5117735562566844e-05, 'epoch': 1.56}) +[2026-06-19 15:22:20,747][starVLA.training.train_starvla][INFO] - Step 2465, Loss: {'train/loss': 0.22026677429676056, 'train/grad_norm_pre_clip': 12.530000686645508, 'timing/data': 0.01901794399600476, 'timing/model': 1.1450961889931932, 'learning_rate/qwen_vl_interface': 5.5087327094097685e-06, 'learning_rate/action_model': 5.508732709409769e-05, 'epoch': 1.56}) +[2026-06-19 15:22:25,321][starVLA.training.train_starvla][INFO] - Step 2466, Loss: {'train/loss': 0.1061224639415741, 'train/grad_norm_pre_clip': 7.518561840057373, 'timing/data': 0.017207612982019782, 'timing/model': 1.1426328920060769, 'learning_rate/qwen_vl_interface': 5.505691756207609e-06, 'learning_rate/action_model': 5.505691756207609e-05, 'epoch': 1.56}) +[2026-06-19 15:22:29,960][starVLA.training.train_starvla][INFO] - Step 2467, Loss: {'train/loss': 0.22434699535369873, 'train/grad_norm_pre_clip': 9.46959400177002, 'timing/data': 0.018194111995398998, 'timing/model': 1.3235750930616632, 'learning_rate/qwen_vl_interface': 5.502650697900228e-06, 'learning_rate/action_model': 5.502650697900228e-05, 'epoch': 1.56}) +[2026-06-19 15:22:34,492][starVLA.training.train_starvla][INFO] - Step 2468, Loss: {'train/loss': 0.08106043189764023, 'train/grad_norm_pre_clip': 6.65769100189209, 'timing/data': 0.01667002495378256, 'timing/model': 1.2457219929201528, 'learning_rate/qwen_vl_interface': 5.499609535737688e-06, 'learning_rate/action_model': 5.4996095357376875e-05, 'epoch': 1.57}) +[2026-06-19 15:22:38,915][starVLA.training.train_starvla][INFO] - Step 2469, Loss: {'train/loss': 0.10642890632152557, 'train/grad_norm_pre_clip': 5.8324127197265625, 'timing/data': 0.017015291028656065, 'timing/model': 1.1413819419685751, 'learning_rate/qwen_vl_interface': 5.496568270970096e-06, 'learning_rate/action_model': 5.496568270970096e-05, 'epoch': 1.57}) +[2026-06-19 15:22:43,479][starVLA.training.train_starvla][INFO] - Step 2470, Loss: {'train/loss': 0.18325850367546082, 'train/grad_norm_pre_clip': 7.4370317459106445, 'timing/data': 0.017394206020981073, 'timing/model': 1.156322245951742, 'learning_rate/qwen_vl_interface': 5.4935269048476015e-06, 'learning_rate/action_model': 5.493526904847601e-05, 'epoch': 1.57}) +[2026-06-19 15:22:48,065][starVLA.training.train_starvla][INFO] - Step 2471, Loss: {'train/loss': 0.10127761214971542, 'train/grad_norm_pre_clip': 9.933581352233887, 'timing/data': 0.01649780198931694, 'timing/model': 1.1509508040035143, 'learning_rate/qwen_vl_interface': 5.490485438620397e-06, 'learning_rate/action_model': 5.4904854386203965e-05, 'epoch': 1.57}) +[2026-06-19 15:22:52,643][starVLA.training.train_starvla][INFO] - Step 2472, Loss: {'train/loss': 0.2886962890625, 'train/grad_norm_pre_clip': 13.937889099121094, 'timing/data': 0.017443890101276338, 'timing/model': 1.1406917230924591, 'learning_rate/qwen_vl_interface': 5.487443873538712e-06, 'learning_rate/action_model': 5.4874438735387115e-05, 'epoch': 1.57}) +[2026-06-19 15:22:57,273][starVLA.training.train_starvla][INFO] - Step 2473, Loss: {'train/loss': 0.20798811316490173, 'train/grad_norm_pre_clip': 12.073372840881348, 'timing/data': 0.016760546015575528, 'timing/model': 1.3558829110115767, 'learning_rate/qwen_vl_interface': 5.484402210852819e-06, 'learning_rate/action_model': 5.484402210852818e-05, 'epoch': 1.57}) +[2026-06-19 15:23:01,706][starVLA.training.train_starvla][INFO] - Step 2474, Loss: {'train/loss': 0.26851770281791687, 'train/grad_norm_pre_clip': 11.027066230773926, 'timing/data': 0.01584190397989005, 'timing/model': 1.14949318789877, 'learning_rate/qwen_vl_interface': 5.48136045181303e-06, 'learning_rate/action_model': 5.48136045181303e-05, 'epoch': 1.57}) +[2026-06-19 15:23:06,312][starVLA.training.train_starvla][INFO] - Step 2475, Loss: {'train/loss': 0.28079891204833984, 'train/grad_norm_pre_clip': 8.367173194885254, 'timing/data': 0.01725829008501023, 'timing/model': 1.1475722810719162, 'learning_rate/qwen_vl_interface': 5.478318597669699e-06, 'learning_rate/action_model': 5.4783185976696995e-05, 'epoch': 1.57}) +[2026-06-19 15:23:10,933][starVLA.training.train_starvla][INFO] - Step 2476, Loss: {'train/loss': 0.28622567653656006, 'train/grad_norm_pre_clip': 11.719932556152344, 'timing/data': 0.01732727501075715, 'timing/model': 1.14900655800011, 'learning_rate/qwen_vl_interface': 5.475276649673218e-06, 'learning_rate/action_model': 5.4752766496732174e-05, 'epoch': 1.57}) +[2026-06-19 15:23:15,540][starVLA.training.train_starvla][INFO] - Step 2477, Loss: {'train/loss': 0.24270781874656677, 'train/grad_norm_pre_clip': 10.496137619018555, 'timing/data': 0.018673735088668764, 'timing/model': 1.1554479619953781, 'learning_rate/qwen_vl_interface': 5.472234609074014e-06, 'learning_rate/action_model': 5.472234609074014e-05, 'epoch': 1.57}) +[2026-06-19 15:23:20,188][starVLA.training.train_starvla][INFO] - Step 2478, Loss: {'train/loss': 0.14566339552402496, 'train/grad_norm_pre_clip': 7.006120681762695, 'timing/data': 0.01821210398338735, 'timing/model': 1.1444694000529125, 'learning_rate/qwen_vl_interface': 5.46919247712256e-06, 'learning_rate/action_model': 5.46919247712256e-05, 'epoch': 1.57}) +[2026-06-19 15:23:24,730][starVLA.training.train_starvla][INFO] - Step 2479, Loss: {'train/loss': 0.1310298889875412, 'train/grad_norm_pre_clip': 8.265888214111328, 'timing/data': 0.01968159608077258, 'timing/model': 1.154545999947004, 'learning_rate/qwen_vl_interface': 5.466150255069356e-06, 'learning_rate/action_model': 5.4661502550693555e-05, 'epoch': 1.57}) +[2026-06-19 15:23:29,323][starVLA.training.train_starvla][INFO] - Step 2480, Loss: {'train/loss': 0.15415266156196594, 'train/grad_norm_pre_clip': 10.248111724853516, 'timing/data': 0.018086932017467916, 'timing/model': 1.3080858290195465, 'learning_rate/qwen_vl_interface': 5.463107944164947e-06, 'learning_rate/action_model': 5.4631079441649466e-05, 'epoch': 1.57}) +[2026-06-19 15:23:33,758][starVLA.training.train_starvla][INFO] - Step 2481, Loss: {'train/loss': 0.16654114425182343, 'train/grad_norm_pre_clip': 13.783071517944336, 'timing/data': 0.018172621028497815, 'timing/model': 1.1457029459998012, 'learning_rate/qwen_vl_interface': 5.46006554565991e-06, 'learning_rate/action_model': 5.4600655456599105e-05, 'epoch': 1.57}) +[2026-06-19 15:23:38,349][starVLA.training.train_starvla][INFO] - Step 2482, Loss: {'train/loss': 0.15444180369377136, 'train/grad_norm_pre_clip': 8.002249717712402, 'timing/data': 0.018161348067224026, 'timing/model': 1.1512817268958315, 'learning_rate/qwen_vl_interface': 5.457023060804865e-06, 'learning_rate/action_model': 5.457023060804864e-05, 'epoch': 1.57}) +[2026-06-19 15:23:42,962][starVLA.training.train_starvla][INFO] - Step 2483, Loss: {'train/loss': 0.1745433807373047, 'train/grad_norm_pre_clip': 11.543634414672852, 'timing/data': 0.015991281950846314, 'timing/model': 1.142019181046635, 'learning_rate/qwen_vl_interface': 5.453980490850459e-06, 'learning_rate/action_model': 5.453980490850459e-05, 'epoch': 1.57}) +[2026-06-19 15:23:47,559][starVLA.training.train_starvla][INFO] - Step 2484, Loss: {'train/loss': 0.12710516154766083, 'train/grad_norm_pre_clip': 7.187857627868652, 'timing/data': 0.01796194294001907, 'timing/model': 1.1516874630469829, 'learning_rate/qwen_vl_interface': 5.4509378370473785e-06, 'learning_rate/action_model': 5.450937837047378e-05, 'epoch': 1.58}) +[2026-06-19 15:23:52,173][starVLA.training.train_starvla][INFO] - Step 2485, Loss: {'train/loss': 0.17957860231399536, 'train/grad_norm_pre_clip': 8.690459251403809, 'timing/data': 0.01687234896235168, 'timing/model': 1.324498123023659, 'learning_rate/qwen_vl_interface': 5.4478951006463424e-06, 'learning_rate/action_model': 5.447895100646342e-05, 'epoch': 1.58}) +[2026-06-19 15:23:56,597][starVLA.training.train_starvla][INFO] - Step 2486, Loss: {'train/loss': 0.29835453629493713, 'train/grad_norm_pre_clip': 12.856517791748047, 'timing/data': 0.01687842106912285, 'timing/model': 1.146045877947472, 'learning_rate/qwen_vl_interface': 5.444852282898106e-06, 'learning_rate/action_model': 5.444852282898106e-05, 'epoch': 1.58}) +[2026-06-19 15:24:01,199][starVLA.training.train_starvla][INFO] - Step 2487, Loss: {'train/loss': 0.13690701127052307, 'train/grad_norm_pre_clip': 6.306056022644043, 'timing/data': 0.016789627028629184, 'timing/model': 1.1496226020390168, 'learning_rate/qwen_vl_interface': 5.441809385053456e-06, 'learning_rate/action_model': 5.4418093850534555e-05, 'epoch': 1.58}) +[2026-06-19 15:24:05,794][starVLA.training.train_starvla][INFO] - Step 2488, Loss: {'train/loss': 0.12467596679925919, 'train/grad_norm_pre_clip': 6.097888946533203, 'timing/data': 0.016941700014285743, 'timing/model': 1.1502729440107942, 'learning_rate/qwen_vl_interface': 5.438766408363214e-06, 'learning_rate/action_model': 5.438766408363214e-05, 'epoch': 1.58}) +[2026-06-19 15:24:10,406][starVLA.training.train_starvla][INFO] - Step 2489, Loss: {'train/loss': 0.19748398661613464, 'train/grad_norm_pre_clip': 10.129877090454102, 'timing/data': 0.01893282902892679, 'timing/model': 1.1451982338912785, 'learning_rate/qwen_vl_interface': 5.43572335407823e-06, 'learning_rate/action_model': 5.43572335407823e-05, 'epoch': 1.58}) +[2026-06-19 15:24:15,012][starVLA.training.train_starvla][INFO] - Step 2490, Loss: {'train/loss': 0.23129108548164368, 'train/grad_norm_pre_clip': 11.069212913513184, 'timing/data': 0.017598271020688117, 'timing/model': 1.1557408940279856, 'learning_rate/qwen_vl_interface': 5.432680223449392e-06, 'learning_rate/action_model': 5.4326802234493924e-05, 'epoch': 1.58}) +[2026-06-19 15:24:19,610][starVLA.training.train_starvla][INFO] - Step 2491, Loss: {'train/loss': 0.13721057772636414, 'train/grad_norm_pre_clip': 8.75617504119873, 'timing/data': 0.017903622006997466, 'timing/model': 1.3137359149986878, 'learning_rate/qwen_vl_interface': 5.429637017727614e-06, 'learning_rate/action_model': 5.4296370177276144e-05, 'epoch': 1.58}) +[2026-06-19 15:24:24,208][starVLA.training.train_starvla][INFO] - Step 2492, Loss: {'train/loss': 0.2197885811328888, 'train/grad_norm_pre_clip': 7.9712066650390625, 'timing/data': 0.016849535051733255, 'timing/model': 1.3196456659352407, 'learning_rate/qwen_vl_interface': 5.426593738163845e-06, 'learning_rate/action_model': 5.4265937381638446e-05, 'epoch': 1.58}) +[2026-06-19 15:24:28,636][starVLA.training.train_starvla][INFO] - Step 2493, Loss: {'train/loss': 0.3252994120121002, 'train/grad_norm_pre_clip': 11.26506519317627, 'timing/data': 0.01762150297872722, 'timing/model': 1.1458018879638985, 'learning_rate/qwen_vl_interface': 5.42355038600906e-06, 'learning_rate/action_model': 5.42355038600906e-05, 'epoch': 1.58}) +[2026-06-19 15:24:33,215][starVLA.training.train_starvla][INFO] - Step 2494, Loss: {'train/loss': 0.1892121136188507, 'train/grad_norm_pre_clip': 9.754524230957031, 'timing/data': 0.016908023972064257, 'timing/model': 1.1447140839882195, 'learning_rate/qwen_vl_interface': 5.420506962514267e-06, 'learning_rate/action_model': 5.420506962514267e-05, 'epoch': 1.58}) +[2026-06-19 15:24:37,813][starVLA.training.train_starvla][INFO] - Step 2495, Loss: {'train/loss': 0.15977513790130615, 'train/grad_norm_pre_clip': 10.751517295837402, 'timing/data': 0.017187852994538844, 'timing/model': 1.1524234910029918, 'learning_rate/qwen_vl_interface': 5.4174634689305015e-06, 'learning_rate/action_model': 5.417463468930501e-05, 'epoch': 1.58}) +[2026-06-19 15:24:42,298][starVLA.training.train_starvla][INFO] - Step 2496, Loss: {'train/loss': 0.21373823285102844, 'train/grad_norm_pre_clip': 10.090208053588867, 'timing/data': 0.01669588196091354, 'timing/model': 1.1428119709016755, 'learning_rate/qwen_vl_interface': 5.41441990650883e-06, 'learning_rate/action_model': 5.4144199065088295e-05, 'epoch': 1.58}) +[2026-06-19 15:24:46,837][starVLA.training.train_starvla][INFO] - Step 2497, Loss: {'train/loss': 0.096413254737854, 'train/grad_norm_pre_clip': 6.158005714416504, 'timing/data': 0.017060084035620093, 'timing/model': 1.144772403058596, 'learning_rate/qwen_vl_interface': 5.411376276500345e-06, 'learning_rate/action_model': 5.411376276500345e-05, 'epoch': 1.58}) +[2026-06-19 15:24:51,430][starVLA.training.train_starvla][INFO] - Step 2498, Loss: {'train/loss': 0.07970119267702103, 'train/grad_norm_pre_clip': 3.1569762229919434, 'timing/data': 0.01695870200637728, 'timing/model': 1.325731401098892, 'learning_rate/qwen_vl_interface': 5.408332580156168e-06, 'learning_rate/action_model': 5.408332580156168e-05, 'epoch': 1.58}) +[2026-06-19 15:24:55,841][starVLA.training.train_starvla][INFO] - Step 2499, Loss: {'train/loss': 0.17607691884040833, 'train/grad_norm_pre_clip': 12.344592094421387, 'timing/data': 0.017293974058702588, 'timing/model': 1.1463963449932635, 'learning_rate/qwen_vl_interface': 5.40528881872745e-06, 'learning_rate/action_model': 5.40528881872745e-05, 'epoch': 1.58}) +[2026-06-19 15:25:00,411][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 2500 +[2026-06-19 15:27:56,309][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 2500 +[2026-06-19 15:27:56,309][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 2500 +[2026-06-19 15:31:04,607][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 2500 +[2026-06-19 15:31:04,608][starVLA.training.train_starvla][INFO] - Step 2500, Loss: {'train/loss': 0.23264895379543304, 'train/grad_norm_pre_clip': 11.577444076538086, 'eval/loss': 0.33103564381599426, 'eval/latency_0/loss': 0.08689964562654495, 'eval/latency_1/loss': 0.0805465504527092, 'eval/latency_2/loss': 0.11454997956752777, 'eval/latency_3/loss': 0.5145319104194641, 'eval/latency_4/loss': 0.8734319806098938, 'eval/flappy/loss': 0.33103564381599426, 'eval/flappy/latency_0/loss': 0.08689964562654495, 'eval/flappy/latency_1/loss': 0.0805465504527092, 'eval/flappy/latency_2/loss': 0.11454997956752777, 'eval/flappy/latency_3/loss': 0.5145319104194641, 'eval/flappy/latency_4/loss': 0.8734319806098938, 'eval/flap_precision': 0.6943521594684385, 'eval/flap_recall': 0.76, 'eval/flap_f1': 0.7256944444444445, 'eval/flap_rate_model': 0.08361111111111111, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.7256944444444445, 'eval/latency_0/flap_precision': 0.6943521594684385, 'eval/latency_0/flap_recall': 0.76, 'eval/latency_0/flap_f1': 0.7256944444444445, 'eval/latency_0/flap_rate_model': 0.08361111111111111, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.7256944444444445, 'timing/data': 0.016367141040973365, 'timing/model': 1.144310593022965, 'learning_rate/qwen_vl_interface': 5.402244993465364e-06, 'learning_rate/action_model': 5.402244993465364e-05, 'epoch': 1.59}) +[2026-06-19 15:31:04,609][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_2500_state +[2026-06-19 15:31:04,611][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-19 15:31:07,971][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_2500_state/model.safetensors +[2026-06-19 15:31:15,258][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_2500_state/optimizer.bin +[2026-06-19 15:31:15,259][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_2500_state/sampler.bin +[2026-06-19 15:31:15,259][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_2500_state/sampler_1.bin +[2026-06-19 15:31:15,259][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_2500_state/random_states_0.pkl +[2026-06-19 15:31:15,281][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2500 +[2026-06-19 15:31:15,282][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-19 15:31:15,305][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-19 15:31:20,690][starVLA.training.train_starvla][INFO] - Latency curriculum phase 3 at step 2501: active_latencies=[0, 1, 2, 3], dataset_size=272156 +[2026-06-19 15:31:20,691][starVLA.training.train_starvla][INFO] - Step 2501, Loss: {'train/loss': 0.2714771032333374, 'train/grad_norm_pre_clip': 11.363457679748535, 'latency_curriculum/phase': 3.0, 'latency_curriculum/active_count': 4.0, 'latency_curriculum/max_active_latency': 3.0, 'latency_curriculum/dataset_size': 272156.0, 'timing/data': 0.017408118001185358, 'timing/model': 1.1234828269807622, 'learning_rate/qwen_vl_interface': 5.399201105621111e-06, 'learning_rate/action_model': 5.39920110562111e-05, 'epoch': 1.18}) +[2026-06-19 15:31:37,619][starVLA.training.train_starvla][INFO] - Step 2502, Loss: {'train/loss': 0.8374795317649841, 'train/grad_norm_pre_clip': 19.390426635742188, 'timing/data': 0.009869833942502737, 'timing/model': 1.1279067629948258, 'learning_rate/qwen_vl_interface': 5.396157156445918e-06, 'learning_rate/action_model': 5.396157156445918e-05, 'epoch': 1.18}) +[2026-06-19 15:31:42,082][starVLA.training.train_starvla][INFO] - Step 2503, Loss: {'train/loss': 0.5896969437599182, 'train/grad_norm_pre_clip': 13.838343620300293, 'timing/data': 0.010093817953020334, 'timing/model': 1.2369194719940424, 'learning_rate/qwen_vl_interface': 5.39311314719104e-06, 'learning_rate/action_model': 5.39311314719104e-05, 'epoch': 1.18}) +[2026-06-19 15:31:46,439][starVLA.training.train_starvla][INFO] - Step 2504, Loss: {'train/loss': 0.5610668659210205, 'train/grad_norm_pre_clip': 11.486224174499512, 'timing/data': 0.009304033010266721, 'timing/model': 1.131154636037536, 'learning_rate/qwen_vl_interface': 5.390069079107755e-06, 'learning_rate/action_model': 5.390069079107755e-05, 'epoch': 1.18}) +[2026-06-19 15:31:50,984][starVLA.training.train_starvla][INFO] - Step 2505, Loss: {'train/loss': 0.35028553009033203, 'train/grad_norm_pre_clip': 11.342904090881348, 'timing/data': 0.00980831601191312, 'timing/model': 1.1327083449577913, 'learning_rate/qwen_vl_interface': 5.387024953447361e-06, 'learning_rate/action_model': 5.387024953447361e-05, 'epoch': 1.18}) +[2026-06-19 15:31:55,525][starVLA.training.train_starvla][INFO] - Step 2506, Loss: {'train/loss': 0.22272908687591553, 'train/grad_norm_pre_clip': 8.315120697021484, 'timing/data': 0.009060718002729118, 'timing/model': 1.1336676789214835, 'learning_rate/qwen_vl_interface': 5.383980771461184e-06, 'learning_rate/action_model': 5.3839807714611846e-05, 'epoch': 1.18}) +[2026-06-19 15:32:00,089][starVLA.training.train_starvla][INFO] - Step 2507, Loss: {'train/loss': 0.5146098732948303, 'train/grad_norm_pre_clip': 9.450926780700684, 'timing/data': 0.008857573964633048, 'timing/model': 1.137818600051105, 'learning_rate/qwen_vl_interface': 5.380936534400572e-06, 'learning_rate/action_model': 5.3809365344005726e-05, 'epoch': 1.18}) +[2026-06-19 15:32:04,667][starVLA.training.train_starvla][INFO] - Step 2508, Loss: {'train/loss': 0.5642725229263306, 'train/grad_norm_pre_clip': 7.780515193939209, 'timing/data': 0.009429444093257189, 'timing/model': 1.3150119689526036, 'learning_rate/qwen_vl_interface': 5.377892243516899e-06, 'learning_rate/action_model': 5.3778922435168985e-05, 'epoch': 1.18}) +[2026-06-19 15:32:09,174][starVLA.training.train_starvla][INFO] - Step 2509, Loss: {'train/loss': 0.26228049397468567, 'train/grad_norm_pre_clip': 6.847769260406494, 'timing/data': 0.010974883916787803, 'timing/model': 1.2517790159909055, 'learning_rate/qwen_vl_interface': 5.3748479000615525e-06, 'learning_rate/action_model': 5.374847900061553e-05, 'epoch': 1.18}) +[2026-06-19 15:32:13,588][starVLA.training.train_starvla][INFO] - Step 2510, Loss: {'train/loss': 0.2294887751340866, 'train/grad_norm_pre_clip': 7.477141857147217, 'timing/data': 0.009269598056562245, 'timing/model': 1.1463565239682794, 'learning_rate/qwen_vl_interface': 5.371803505285954e-06, 'learning_rate/action_model': 5.371803505285954e-05, 'epoch': 1.18}) +[2026-06-19 15:32:18,159][starVLA.training.train_starvla][INFO] - Step 2511, Loss: {'train/loss': 0.4467650353908539, 'train/grad_norm_pre_clip': 7.427365779876709, 'timing/data': 0.009097330970689654, 'timing/model': 1.142354644020088, 'learning_rate/qwen_vl_interface': 5.368759060441534e-06, 'learning_rate/action_model': 5.368759060441534e-05, 'epoch': 1.18}) +[2026-06-19 15:32:22,735][starVLA.training.train_starvla][INFO] - Step 2512, Loss: {'train/loss': 0.8110254406929016, 'train/grad_norm_pre_clip': 11.857234954833984, 'timing/data': 0.00947681104298681, 'timing/model': 1.1475310359383002, 'learning_rate/qwen_vl_interface': 5.365714566779749e-06, 'learning_rate/action_model': 5.3657145667797484e-05, 'epoch': 1.18}) +[2026-06-19 15:32:27,317][starVLA.training.train_starvla][INFO] - Step 2513, Loss: {'train/loss': 0.2007526457309723, 'train/grad_norm_pre_clip': 7.812220573425293, 'timing/data': 0.009630087995901704, 'timing/model': 1.1557992530288175, 'learning_rate/qwen_vl_interface': 5.362670025552077e-06, 'learning_rate/action_model': 5.362670025552077e-05, 'epoch': 1.18}) +[2026-06-19 15:32:31,902][starVLA.training.train_starvla][INFO] - Step 2514, Loss: {'train/loss': 0.30064159631729126, 'train/grad_norm_pre_clip': 6.434699535369873, 'timing/data': 0.010285115917213261, 'timing/model': 1.3201820249669254, 'learning_rate/qwen_vl_interface': 5.359625438010013e-06, 'learning_rate/action_model': 5.359625438010013e-05, 'epoch': 1.18}) +[2026-06-19 15:32:36,438][starVLA.training.train_starvla][INFO] - Step 2515, Loss: {'train/loss': 0.4176529347896576, 'train/grad_norm_pre_clip': 7.38638973236084, 'timing/data': 0.00949282106012106, 'timing/model': 1.2625319639919326, 'learning_rate/qwen_vl_interface': 5.356580805405075e-06, 'learning_rate/action_model': 5.356580805405075e-05, 'epoch': 1.18}) +[2026-06-19 15:32:40,837][starVLA.training.train_starvla][INFO] - Step 2516, Loss: {'train/loss': 0.29188376665115356, 'train/grad_norm_pre_clip': 8.397970199584961, 'timing/data': 0.008895232924260199, 'timing/model': 1.1393858030205593, 'learning_rate/qwen_vl_interface': 5.353536128988791e-06, 'learning_rate/action_model': 5.353536128988791e-05, 'epoch': 1.18}) +[2026-06-19 15:32:45,421][starVLA.training.train_starvla][INFO] - Step 2517, Loss: {'train/loss': 0.49733611941337585, 'train/grad_norm_pre_clip': 12.38878345489502, 'timing/data': 0.009279770078137517, 'timing/model': 1.1404865409713238, 'learning_rate/qwen_vl_interface': 5.350491410012717e-06, 'learning_rate/action_model': 5.350491410012717e-05, 'epoch': 1.18}) +[2026-06-19 15:32:50,002][starVLA.training.train_starvla][INFO] - Step 2518, Loss: {'train/loss': 0.18000458180904388, 'train/grad_norm_pre_clip': 5.690384864807129, 'timing/data': 0.009390474064275622, 'timing/model': 1.1405176400439814, 'learning_rate/qwen_vl_interface': 5.347446649728419e-06, 'learning_rate/action_model': 5.347446649728418e-05, 'epoch': 1.18}) +[2026-06-19 15:32:54,578][starVLA.training.train_starvla][INFO] - Step 2519, Loss: {'train/loss': 0.41258060932159424, 'train/grad_norm_pre_clip': 6.890435218811035, 'timing/data': 0.009038902935571969, 'timing/model': 1.140666202059947, 'learning_rate/qwen_vl_interface': 5.344401849387484e-06, 'learning_rate/action_model': 5.344401849387484e-05, 'epoch': 1.18}) +[2026-06-19 15:32:59,140][starVLA.training.train_starvla][INFO] - Step 2520, Loss: {'train/loss': 0.17968712747097015, 'train/grad_norm_pre_clip': 5.246574401855469, 'timing/data': 0.0094373399624601, 'timing/model': 1.1414667019853368, 'learning_rate/qwen_vl_interface': 5.3413570102415144e-06, 'learning_rate/action_model': 5.3413570102415144e-05, 'epoch': 1.19}) +[2026-06-19 15:33:03,732][starVLA.training.train_starvla][INFO] - Step 2521, Loss: {'train/loss': 0.33310985565185547, 'train/grad_norm_pre_clip': 11.048336029052734, 'timing/data': 0.009621568955481052, 'timing/model': 1.3210219790926203, 'learning_rate/qwen_vl_interface': 5.338312133542131e-06, 'learning_rate/action_model': 5.3383121335421305e-05, 'epoch': 1.19}) +[2026-06-19 15:33:08,153][starVLA.training.train_starvla][INFO] - Step 2522, Loss: {'train/loss': 0.2639521360397339, 'train/grad_norm_pre_clip': 5.004438400268555, 'timing/data': 0.009508875897154212, 'timing/model': 1.151336453971453, 'learning_rate/qwen_vl_interface': 5.335267220540963e-06, 'learning_rate/action_model': 5.335267220540963e-05, 'epoch': 1.19}) +[2026-06-19 15:33:12,750][starVLA.training.train_starvla][INFO] - Step 2523, Loss: {'train/loss': 0.40164414048194885, 'train/grad_norm_pre_clip': 10.301101684570312, 'timing/data': 0.009396255016326904, 'timing/model': 1.1500155669637024, 'learning_rate/qwen_vl_interface': 5.332222272489662e-06, 'learning_rate/action_model': 5.332222272489662e-05, 'epoch': 1.19}) +[2026-06-19 15:33:17,341][starVLA.training.train_starvla][INFO] - Step 2524, Loss: {'train/loss': 0.2687300443649292, 'train/grad_norm_pre_clip': 6.446675777435303, 'timing/data': 0.009906225954182446, 'timing/model': 1.1446089420933276, 'learning_rate/qwen_vl_interface': 5.3291772906398895e-06, 'learning_rate/action_model': 5.32917729063989e-05, 'epoch': 1.19}) +[2026-06-19 15:33:21,937][starVLA.training.train_starvla][INFO] - Step 2525, Loss: {'train/loss': 0.3042432963848114, 'train/grad_norm_pre_clip': 8.940523147583008, 'timing/data': 0.009091577026993036, 'timing/model': 1.1439928149338812, 'learning_rate/qwen_vl_interface': 5.326132276243327e-06, 'learning_rate/action_model': 5.3261322762433266e-05, 'epoch': 1.19}) +[2026-06-19 15:33:26,524][starVLA.training.train_starvla][INFO] - Step 2526, Loss: {'train/loss': 0.4664994776248932, 'train/grad_norm_pre_clip': 6.59037446975708, 'timing/data': 0.00902973988559097, 'timing/model': 1.145491898059845, 'learning_rate/qwen_vl_interface': 5.323087230551659e-06, 'learning_rate/action_model': 5.323087230551659e-05, 'epoch': 1.19}) +[2026-06-19 15:33:31,093][starVLA.training.train_starvla][INFO] - Step 2527, Loss: {'train/loss': 0.28492116928100586, 'train/grad_norm_pre_clip': 7.352606296539307, 'timing/data': 0.009433111059479415, 'timing/model': 1.2998453930485994, 'learning_rate/qwen_vl_interface': 5.3200421548165916e-06, 'learning_rate/action_model': 5.3200421548165914e-05, 'epoch': 1.19}) +[2026-06-19 15:33:35,522][starVLA.training.train_starvla][INFO] - Step 2528, Loss: {'train/loss': 0.24990148842334747, 'train/grad_norm_pre_clip': 4.944421291351318, 'timing/data': 0.009632335975766182, 'timing/model': 1.1425933970604092, 'learning_rate/qwen_vl_interface': 5.3169970502898405e-06, 'learning_rate/action_model': 5.31699705028984e-05, 'epoch': 1.19}) +[2026-06-19 15:33:40,104][starVLA.training.train_starvla][INFO] - Step 2529, Loss: {'train/loss': 0.20982404053211212, 'train/grad_norm_pre_clip': 5.254606246948242, 'timing/data': 0.009430269012227654, 'timing/model': 1.1466909989248961, 'learning_rate/qwen_vl_interface': 5.31395191822313e-06, 'learning_rate/action_model': 5.31395191822313e-05, 'epoch': 1.19}) +[2026-06-19 15:33:44,610][starVLA.training.train_starvla][INFO] - Step 2530, Loss: {'train/loss': 0.21887744963169098, 'train/grad_norm_pre_clip': 6.85313606262207, 'timing/data': 0.009835653007030487, 'timing/model': 1.147246022010222, 'learning_rate/qwen_vl_interface': 5.310906759868205e-06, 'learning_rate/action_model': 5.310906759868205e-05, 'epoch': 1.19}) +[2026-06-19 15:33:49,135][starVLA.training.train_starvla][INFO] - Step 2531, Loss: {'train/loss': 0.3276073634624481, 'train/grad_norm_pre_clip': 6.517692565917969, 'timing/data': 0.0095075509743765, 'timing/model': 1.1503202560124919, 'learning_rate/qwen_vl_interface': 5.30786157647681e-06, 'learning_rate/action_model': 5.30786157647681e-05, 'epoch': 1.19}) +[2026-06-19 15:33:53,739][starVLA.training.train_starvla][INFO] - Step 2532, Loss: {'train/loss': 0.6128206849098206, 'train/grad_norm_pre_clip': 7.9759135246276855, 'timing/data': 0.009696335066109896, 'timing/model': 1.14539447799325, 'learning_rate/qwen_vl_interface': 5.304816369300709e-06, 'learning_rate/action_model': 5.304816369300709e-05, 'epoch': 1.19}) +[2026-06-19 15:33:58,334][starVLA.training.train_starvla][INFO] - Step 2533, Loss: {'train/loss': 0.13082705438137054, 'train/grad_norm_pre_clip': 5.172375202178955, 'timing/data': 0.009715762920677662, 'timing/model': 1.317910384037532, 'learning_rate/qwen_vl_interface': 5.301771139591666e-06, 'learning_rate/action_model': 5.301771139591666e-05, 'epoch': 1.19}) +[2026-06-19 15:34:02,790][starVLA.training.train_starvla][INFO] - Step 2534, Loss: {'train/loss': 0.2833261787891388, 'train/grad_norm_pre_clip': 6.364388465881348, 'timing/data': 0.009549736976623535, 'timing/model': 1.154635243001394, 'learning_rate/qwen_vl_interface': 5.298725888601465e-06, 'learning_rate/action_model': 5.298725888601465e-05, 'epoch': 1.19}) +[2026-06-19 15:34:07,391][starVLA.training.train_starvla][INFO] - Step 2535, Loss: {'train/loss': 0.31168365478515625, 'train/grad_norm_pre_clip': 7.549622535705566, 'timing/data': 0.009312097099609673, 'timing/model': 1.1568107870407403, 'learning_rate/qwen_vl_interface': 5.295680617581891e-06, 'learning_rate/action_model': 5.2956806175818904e-05, 'epoch': 1.19}) +[2026-06-19 15:34:12,009][starVLA.training.train_starvla][INFO] - Step 2536, Loss: {'train/loss': 0.2941783666610718, 'train/grad_norm_pre_clip': 8.482522964477539, 'timing/data': 0.010217348928563297, 'timing/model': 1.160657444037497, 'learning_rate/qwen_vl_interface': 5.2926353277847395e-06, 'learning_rate/action_model': 5.292635327784739e-05, 'epoch': 1.19}) +[2026-06-19 15:34:16,599][starVLA.training.train_starvla][INFO] - Step 2537, Loss: {'train/loss': 0.49728628993034363, 'train/grad_norm_pre_clip': 5.381109714508057, 'timing/data': 0.009847837034612894, 'timing/model': 1.1454676979919896, 'learning_rate/qwen_vl_interface': 5.2895900204618175e-06, 'learning_rate/action_model': 5.289590020461818e-05, 'epoch': 1.19}) +[2026-06-19 15:34:21,203][starVLA.training.train_starvla][INFO] - Step 2538, Loss: {'train/loss': 0.2629833221435547, 'train/grad_norm_pre_clip': 8.163567543029785, 'timing/data': 0.009782347013242543, 'timing/model': 1.147035693982616, 'learning_rate/qwen_vl_interface': 5.286544696864933e-06, 'learning_rate/action_model': 5.2865446968649324e-05, 'epoch': 1.19}) +[2026-06-19 15:34:25,811][starVLA.training.train_starvla][INFO] - Step 2539, Loss: {'train/loss': 0.20926456153392792, 'train/grad_norm_pre_clip': 6.569118976593018, 'timing/data': 0.009560214937664568, 'timing/model': 1.314476078026928, 'learning_rate/qwen_vl_interface': 5.283499358245904e-06, 'learning_rate/action_model': 5.2834993582459034e-05, 'epoch': 1.19}) +[2026-06-19 15:34:30,251][starVLA.training.train_starvla][INFO] - Step 2540, Loss: {'train/loss': 0.4157751798629761, 'train/grad_norm_pre_clip': 7.113452911376953, 'timing/data': 0.009903708007186651, 'timing/model': 1.1495748569723219, 'learning_rate/qwen_vl_interface': 5.280454005856553e-06, 'learning_rate/action_model': 5.2804540058565534e-05, 'epoch': 1.19}) +[2026-06-19 15:34:34,847][starVLA.training.train_starvla][INFO] - Step 2541, Loss: {'train/loss': 0.38792145252227783, 'train/grad_norm_pre_clip': 12.88050651550293, 'timing/data': 0.009203820023685694, 'timing/model': 1.1508223100099713, 'learning_rate/qwen_vl_interface': 5.277408640948711e-06, 'learning_rate/action_model': 5.277408640948711e-05, 'epoch': 1.2}) +[2026-06-19 15:34:39,350][starVLA.training.train_starvla][INFO] - Step 2542, Loss: {'train/loss': 0.22326582670211792, 'train/grad_norm_pre_clip': 4.863797187805176, 'timing/data': 0.009070600033737719, 'timing/model': 1.146564738941379, 'learning_rate/qwen_vl_interface': 5.274363264774213e-06, 'learning_rate/action_model': 5.274363264774212e-05, 'epoch': 1.2}) +[2026-06-19 15:34:43,920][starVLA.training.train_starvla][INFO] - Step 2543, Loss: {'train/loss': 0.1392984539270401, 'train/grad_norm_pre_clip': 5.603970050811768, 'timing/data': 0.009569488000124693, 'timing/model': 1.1521175020607188, 'learning_rate/qwen_vl_interface': 5.271317878584897e-06, 'learning_rate/action_model': 5.2713178785848974e-05, 'epoch': 1.2}) +[2026-06-19 15:34:48,519][starVLA.training.train_starvla][INFO] - Step 2544, Loss: {'train/loss': 0.2884543836116791, 'train/grad_norm_pre_clip': 7.328149318695068, 'timing/data': 0.009474506019614637, 'timing/model': 1.1525752909947187, 'learning_rate/qwen_vl_interface': 5.268272483632606e-06, 'learning_rate/action_model': 5.268272483632606e-05, 'epoch': 1.2}) +[2026-06-19 15:34:53,135][starVLA.training.train_starvla][INFO] - Step 2545, Loss: {'train/loss': 0.1866157352924347, 'train/grad_norm_pre_clip': 5.757092475891113, 'timing/data': 0.010181708028540015, 'timing/model': 1.3128212440060452, 'learning_rate/qwen_vl_interface': 5.265227081169187e-06, 'learning_rate/action_model': 5.265227081169186e-05, 'epoch': 1.2}) +[2026-06-19 15:34:57,701][starVLA.training.train_starvla][INFO] - Step 2546, Loss: {'train/loss': 0.19332613050937653, 'train/grad_norm_pre_clip': 5.8193488121032715, 'timing/data': 0.009651781059801579, 'timing/model': 1.2715668239397928, 'learning_rate/qwen_vl_interface': 5.2621816724464895e-06, 'learning_rate/action_model': 5.262181672446489e-05, 'epoch': 1.2}) +[2026-06-19 15:35:02,147][starVLA.training.train_starvla][INFO] - Step 2547, Loss: {'train/loss': 0.24322092533111572, 'train/grad_norm_pre_clip': 5.541759490966797, 'timing/data': 0.009004181018099189, 'timing/model': 1.1564975479850546, 'learning_rate/qwen_vl_interface': 5.259136258716367e-06, 'learning_rate/action_model': 5.2591362587163675e-05, 'epoch': 1.2}) +[2026-06-19 15:35:06,762][starVLA.training.train_starvla][INFO] - Step 2548, Loss: {'train/loss': 0.34794381260871887, 'train/grad_norm_pre_clip': 9.958194732666016, 'timing/data': 0.010260457987897098, 'timing/model': 1.1453671309864148, 'learning_rate/qwen_vl_interface': 5.256090841230675e-06, 'learning_rate/action_model': 5.256090841230674e-05, 'epoch': 1.2}) +[2026-06-19 15:35:11,362][starVLA.training.train_starvla][INFO] - Step 2549, Loss: {'train/loss': 0.26160740852355957, 'train/grad_norm_pre_clip': 7.199502944946289, 'timing/data': 0.009134661057032645, 'timing/model': 1.1460813529556617, 'learning_rate/qwen_vl_interface': 5.253045421241266e-06, 'learning_rate/action_model': 5.253045421241266e-05, 'epoch': 1.2}) +[2026-06-19 15:35:15,964][starVLA.training.train_starvla][INFO] - Step 2550, Loss: {'train/loss': 0.16363239288330078, 'train/grad_norm_pre_clip': 7.423275947570801, 'timing/data': 0.008699717000126839, 'timing/model': 1.1510091070085764, 'learning_rate/qwen_vl_interface': 5.2500000000000006e-06, 'learning_rate/action_model': 5.25e-05, 'epoch': 1.2}) +[2026-06-19 15:35:20,478][starVLA.training.train_starvla][INFO] - Step 2551, Loss: {'train/loss': 0.23688863217830658, 'train/grad_norm_pre_clip': 4.866832256317139, 'timing/data': 0.009023286052979529, 'timing/model': 1.154541832045652, 'learning_rate/qwen_vl_interface': 5.246954578758735e-06, 'learning_rate/action_model': 5.246954578758735e-05, 'epoch': 1.2}) +[2026-06-19 15:35:25,019][starVLA.training.train_starvla][INFO] - Step 2552, Loss: {'train/loss': 0.30979838967323303, 'train/grad_norm_pre_clip': 9.080399513244629, 'timing/data': 0.008867241907864809, 'timing/model': 1.246055271010846, 'learning_rate/qwen_vl_interface': 5.243909158769327e-06, 'learning_rate/action_model': 5.2439091587693264e-05, 'epoch': 1.2}) +[2026-06-19 15:35:29,452][starVLA.training.train_starvla][INFO] - Step 2553, Loss: {'train/loss': 0.18240399658679962, 'train/grad_norm_pre_clip': 6.903456687927246, 'timing/data': 0.009250586968846619, 'timing/model': 1.1447315639816225, 'learning_rate/qwen_vl_interface': 5.240863741283634e-06, 'learning_rate/action_model': 5.2408637412836336e-05, 'epoch': 1.2}) +[2026-06-19 15:35:34,051][starVLA.training.train_starvla][INFO] - Step 2554, Loss: {'train/loss': 0.3178887367248535, 'train/grad_norm_pre_clip': 7.615728378295898, 'timing/data': 0.009532956057228148, 'timing/model': 1.1530870939604938, 'learning_rate/qwen_vl_interface': 5.2378183275535125e-06, 'learning_rate/action_model': 5.237818327553512e-05, 'epoch': 1.2}) +[2026-06-19 15:35:38,672][starVLA.training.train_starvla][INFO] - Step 2555, Loss: {'train/loss': 0.07146892696619034, 'train/grad_norm_pre_clip': 2.1349687576293945, 'timing/data': 0.009678398026153445, 'timing/model': 1.1489885540213436, 'learning_rate/qwen_vl_interface': 5.234772918830815e-06, 'learning_rate/action_model': 5.234772918830815e-05, 'epoch': 1.2}) +[2026-06-19 15:35:43,266][starVLA.training.train_starvla][INFO] - Step 2556, Loss: {'train/loss': 0.21915984153747559, 'train/grad_norm_pre_clip': 4.973886966705322, 'timing/data': 0.009470950928516686, 'timing/model': 1.1478508319705725, 'learning_rate/qwen_vl_interface': 5.231727516367395e-06, 'learning_rate/action_model': 5.231727516367395e-05, 'epoch': 1.2}) +[2026-06-19 15:35:47,876][starVLA.training.train_starvla][INFO] - Step 2557, Loss: {'train/loss': 0.12346101552248001, 'train/grad_norm_pre_clip': 4.3985276222229, 'timing/data': 0.009501424967311323, 'timing/model': 1.3199308979092166, 'learning_rate/qwen_vl_interface': 5.228682121415104e-06, 'learning_rate/action_model': 5.228682121415104e-05, 'epoch': 1.2}) +[2026-06-19 15:35:52,309][starVLA.training.train_starvla][INFO] - Step 2558, Loss: {'train/loss': 0.30201271176338196, 'train/grad_norm_pre_clip': 8.405606269836426, 'timing/data': 0.010052566998638213, 'timing/model': 1.1469865770777687, 'learning_rate/qwen_vl_interface': 5.225636735225789e-06, 'learning_rate/action_model': 5.225636735225788e-05, 'epoch': 1.2}) +[2026-06-19 15:35:56,917][starVLA.training.train_starvla][INFO] - Step 2559, Loss: {'train/loss': 0.26060447096824646, 'train/grad_norm_pre_clip': 4.366208553314209, 'timing/data': 0.009925815975293517, 'timing/model': 1.1520772509975359, 'learning_rate/qwen_vl_interface': 5.22259135905129e-06, 'learning_rate/action_model': 5.2225913590512895e-05, 'epoch': 1.2}) +[2026-06-19 15:36:01,552][starVLA.training.train_starvla][INFO] - Step 2560, Loss: {'train/loss': 0.3926631212234497, 'train/grad_norm_pre_clip': 8.25093936920166, 'timing/data': 0.009033980895765126, 'timing/model': 1.1624500630423427, 'learning_rate/qwen_vl_interface': 5.219545994143448e-06, 'learning_rate/action_model': 5.219545994143448e-05, 'epoch': 1.2}) +[2026-06-19 15:36:06,165][starVLA.training.train_starvla][INFO] - Step 2561, Loss: {'train/loss': 0.2819885015487671, 'train/grad_norm_pre_clip': 10.395882606506348, 'timing/data': 0.00995681807398796, 'timing/model': 1.1438051239820197, 'learning_rate/qwen_vl_interface': 5.216500641754097e-06, 'learning_rate/action_model': 5.216500641754097e-05, 'epoch': 1.2}) +[2026-06-19 15:36:10,777][starVLA.training.train_starvla][INFO] - Step 2562, Loss: {'train/loss': 0.3955066502094269, 'train/grad_norm_pre_clip': 13.755387306213379, 'timing/data': 0.009523457963950932, 'timing/model': 1.1503967950120568, 'learning_rate/qwen_vl_interface': 5.213455303135067e-06, 'learning_rate/action_model': 5.2134553031350673e-05, 'epoch': 1.2}) +[2026-06-19 15:36:15,408][starVLA.training.train_starvla][INFO] - Step 2563, Loss: {'train/loss': 0.18998093903064728, 'train/grad_norm_pre_clip': 10.371603965759277, 'timing/data': 0.009020819910801947, 'timing/model': 1.3395226629218087, 'learning_rate/qwen_vl_interface': 5.2104099795381836e-06, 'learning_rate/action_model': 5.210409979538183e-05, 'epoch': 1.21}) +[2026-06-19 15:36:19,953][starVLA.training.train_starvla][INFO] - Step 2564, Loss: {'train/loss': 0.2608245015144348, 'train/grad_norm_pre_clip': 11.296440124511719, 'timing/data': 0.009647864964790642, 'timing/model': 1.25352520798333, 'learning_rate/qwen_vl_interface': 5.207364672215261e-06, 'learning_rate/action_model': 5.2073646722152604e-05, 'epoch': 1.21}) +[2026-06-19 15:36:24,382][starVLA.training.train_starvla][INFO] - Step 2565, Loss: {'train/loss': 0.38163524866104126, 'train/grad_norm_pre_clip': 10.351557731628418, 'timing/data': 0.009438170935027301, 'timing/model': 1.149125942029059, 'learning_rate/qwen_vl_interface': 5.2043193824181095e-06, 'learning_rate/action_model': 5.2043193824181094e-05, 'epoch': 1.21}) +[2026-06-19 15:36:28,989][starVLA.training.train_starvla][INFO] - Step 2566, Loss: {'train/loss': 0.37989482283592224, 'train/grad_norm_pre_clip': 10.755776405334473, 'timing/data': 0.009189645061269403, 'timing/model': 1.148009313037619, 'learning_rate/qwen_vl_interface': 5.201274111398535e-06, 'learning_rate/action_model': 5.2012741113985356e-05, 'epoch': 1.21}) +[2026-06-19 15:36:33,601][starVLA.training.train_starvla][INFO] - Step 2567, Loss: {'train/loss': 0.13207659125328064, 'train/grad_norm_pre_clip': 4.327939510345459, 'timing/data': 0.0093726139748469, 'timing/model': 1.1467898449627683, 'learning_rate/qwen_vl_interface': 5.198228860408334e-06, 'learning_rate/action_model': 5.198228860408334e-05, 'epoch': 1.21}) +[2026-06-19 15:36:38,213][starVLA.training.train_starvla][INFO] - Step 2568, Loss: {'train/loss': 0.19386926293373108, 'train/grad_norm_pre_clip': 8.095480918884277, 'timing/data': 0.00900915008969605, 'timing/model': 1.1483373440569267, 'learning_rate/qwen_vl_interface': 5.195183630699292e-06, 'learning_rate/action_model': 5.195183630699292e-05, 'epoch': 1.21}) +[2026-06-19 15:36:42,820][starVLA.training.train_starvla][INFO] - Step 2569, Loss: {'train/loss': 0.23411007225513458, 'train/grad_norm_pre_clip': 12.1988525390625, 'timing/data': 0.010109090944752097, 'timing/model': 1.3139560869894922, 'learning_rate/qwen_vl_interface': 5.192138423523189e-06, 'learning_rate/action_model': 5.1921384235231895e-05, 'epoch': 1.21}) +[2026-06-19 15:36:47,367][starVLA.training.train_starvla][INFO] - Step 2570, Loss: {'train/loss': 0.20698970556259155, 'train/grad_norm_pre_clip': 7.496212959289551, 'timing/data': 0.009741830988787115, 'timing/model': 1.2604495369596407, 'learning_rate/qwen_vl_interface': 5.189093240131795e-06, 'learning_rate/action_model': 5.189093240131795e-05, 'epoch': 1.21}) +[2026-06-19 15:36:51,802][starVLA.training.train_starvla][INFO] - Step 2571, Loss: {'train/loss': 0.270361065864563, 'train/grad_norm_pre_clip': 8.677058219909668, 'timing/data': 0.010172666050493717, 'timing/model': 1.1491441800026223, 'learning_rate/qwen_vl_interface': 5.186048081776869e-06, 'learning_rate/action_model': 5.1860480817768685e-05, 'epoch': 1.21}) +[2026-06-19 15:36:56,394][starVLA.training.train_starvla][INFO] - Step 2572, Loss: {'train/loss': 0.12904280424118042, 'train/grad_norm_pre_clip': 8.098528861999512, 'timing/data': 0.009367519989609718, 'timing/model': 1.1475167389726266, 'learning_rate/qwen_vl_interface': 5.183002949710161e-06, 'learning_rate/action_model': 5.18300294971016e-05, 'epoch': 1.21}) +[2026-06-19 15:37:00,990][starVLA.training.train_starvla][INFO] - Step 2573, Loss: {'train/loss': 0.4273078441619873, 'train/grad_norm_pre_clip': 14.796707153320312, 'timing/data': 0.010314731975086033, 'timing/model': 1.147564257029444, 'learning_rate/qwen_vl_interface': 5.179957845183411e-06, 'learning_rate/action_model': 5.179957845183411e-05, 'epoch': 1.21}) +[2026-06-19 15:37:05,603][starVLA.training.train_starvla][INFO] - Step 2574, Loss: {'train/loss': 0.5404914617538452, 'train/grad_norm_pre_clip': 16.768321990966797, 'timing/data': 0.010267053032293916, 'timing/model': 1.1501781970728189, 'learning_rate/qwen_vl_interface': 5.176912769448343e-06, 'learning_rate/action_model': 5.176912769448343e-05, 'epoch': 1.21}) +[2026-06-19 15:37:10,216][starVLA.training.train_starvla][INFO] - Step 2575, Loss: {'train/loss': 0.27128005027770996, 'train/grad_norm_pre_clip': 8.933280944824219, 'timing/data': 0.009745298069901764, 'timing/model': 1.3299173699924722, 'learning_rate/qwen_vl_interface': 5.173867723756676e-06, 'learning_rate/action_model': 5.173867723756676e-05, 'epoch': 1.21}) +[2026-06-19 15:37:14,791][starVLA.training.train_starvla][INFO] - Step 2576, Loss: {'train/loss': 0.2643001973628998, 'train/grad_norm_pre_clip': 5.086630821228027, 'timing/data': 0.009552273899316788, 'timing/model': 1.2811437919735909, 'learning_rate/qwen_vl_interface': 5.170822709360112e-06, 'learning_rate/action_model': 5.170822709360111e-05, 'epoch': 1.21}) +[2026-06-19 15:37:19,228][starVLA.training.train_starvla][INFO] - Step 2577, Loss: {'train/loss': 0.32635363936424255, 'train/grad_norm_pre_clip': 12.76943588256836, 'timing/data': 0.009152938961051404, 'timing/model': 1.1499352380633354, 'learning_rate/qwen_vl_interface': 5.16777772751034e-06, 'learning_rate/action_model': 5.16777772751034e-05, 'epoch': 1.21}) +[2026-06-19 15:37:23,853][starVLA.training.train_starvla][INFO] - Step 2578, Loss: {'train/loss': 0.2079276740550995, 'train/grad_norm_pre_clip': 6.465079307556152, 'timing/data': 0.011336519964970648, 'timing/model': 1.155993732973002, 'learning_rate/qwen_vl_interface': 5.16473277945904e-06, 'learning_rate/action_model': 5.1647327794590396e-05, 'epoch': 1.21}) +[2026-06-19 15:37:28,445][starVLA.training.train_starvla][INFO] - Step 2579, Loss: {'train/loss': 0.1905546635389328, 'train/grad_norm_pre_clip': 5.589851379394531, 'timing/data': 0.008823871030472219, 'timing/model': 1.1473922709701583, 'learning_rate/qwen_vl_interface': 5.161687866457872e-06, 'learning_rate/action_model': 5.161687866457871e-05, 'epoch': 1.21}) +[2026-06-19 15:37:33,063][starVLA.training.train_starvla][INFO] - Step 2580, Loss: {'train/loss': 0.2799273431301117, 'train/grad_norm_pre_clip': 6.950788497924805, 'timing/data': 0.009281935985200107, 'timing/model': 1.1597473170841113, 'learning_rate/qwen_vl_interface': 5.158642989758487e-06, 'learning_rate/action_model': 5.158642989758487e-05, 'epoch': 1.21}) +[2026-06-19 15:37:37,653][starVLA.training.train_starvla][INFO] - Step 2581, Loss: {'train/loss': 0.38231077790260315, 'train/grad_norm_pre_clip': 7.841578006744385, 'timing/data': 0.009475272032432258, 'timing/model': 1.1485693940194324, 'learning_rate/qwen_vl_interface': 5.155598150612519e-06, 'learning_rate/action_model': 5.1555981506125185e-05, 'epoch': 1.21}) +[2026-06-19 15:37:42,256][starVLA.training.train_starvla][INFO] - Step 2582, Loss: {'train/loss': 0.2971627116203308, 'train/grad_norm_pre_clip': 17.867267608642578, 'timing/data': 0.00926494097802788, 'timing/model': 1.3283578179543838, 'learning_rate/qwen_vl_interface': 5.152553350271584e-06, 'learning_rate/action_model': 5.152553350271584e-05, 'epoch': 1.21}) +[2026-06-19 15:37:46,680][starVLA.training.train_starvla][INFO] - Step 2583, Loss: {'train/loss': 0.1258278787136078, 'train/grad_norm_pre_clip': 8.224858283996582, 'timing/data': 0.00988926098216325, 'timing/model': 1.1450781229650602, 'learning_rate/qwen_vl_interface': 5.149508589987286e-06, 'learning_rate/action_model': 5.1495085899872855e-05, 'epoch': 1.21}) +[2026-06-19 15:37:51,298][starVLA.training.train_starvla][INFO] - Step 2584, Loss: {'train/loss': 0.2502528131008148, 'train/grad_norm_pre_clip': 12.254562377929688, 'timing/data': 0.00939949695020914, 'timing/model': 1.156704863999039, 'learning_rate/qwen_vl_interface': 5.14646387101121e-06, 'learning_rate/action_model': 5.14646387101121e-05, 'epoch': 1.22}) +[2026-06-19 15:37:55,918][starVLA.training.train_starvla][INFO] - Step 2585, Loss: {'train/loss': 0.19881048798561096, 'train/grad_norm_pre_clip': 5.243347644805908, 'timing/data': 0.012505199061706662, 'timing/model': 1.1494151999941096, 'learning_rate/qwen_vl_interface': 5.143419194594928e-06, 'learning_rate/action_model': 5.143419194594927e-05, 'epoch': 1.22}) +[2026-06-19 15:38:00,515][starVLA.training.train_starvla][INFO] - Step 2586, Loss: {'train/loss': 0.2759411036968231, 'train/grad_norm_pre_clip': 9.813309669494629, 'timing/data': 0.009515482932329178, 'timing/model': 1.1473031889181584, 'learning_rate/qwen_vl_interface': 5.140374561989988e-06, 'learning_rate/action_model': 5.140374561989988e-05, 'epoch': 1.22}) +[2026-06-19 15:38:05,142][starVLA.training.train_starvla][INFO] - Step 2587, Loss: {'train/loss': 0.3428436517715454, 'train/grad_norm_pre_clip': 8.468432426452637, 'timing/data': 0.00886166503187269, 'timing/model': 1.1601640440057963, 'learning_rate/qwen_vl_interface': 5.137329974447924e-06, 'learning_rate/action_model': 5.137329974447924e-05, 'epoch': 1.22}) +[2026-06-19 15:38:09,755][starVLA.training.train_starvla][INFO] - Step 2588, Loss: {'train/loss': 0.4382941424846649, 'train/grad_norm_pre_clip': 7.733949661254883, 'timing/data': 0.009290880989283323, 'timing/model': 1.3278808509930968, 'learning_rate/qwen_vl_interface': 5.1342854332202516e-06, 'learning_rate/action_model': 5.1342854332202514e-05, 'epoch': 1.22}) +[2026-06-19 15:38:14,189][starVLA.training.train_starvla][INFO] - Step 2589, Loss: {'train/loss': 0.40231654047966003, 'train/grad_norm_pre_clip': 8.545133590698242, 'timing/data': 0.009563911007717252, 'timing/model': 1.1477803329471499, 'learning_rate/qwen_vl_interface': 5.131240939558467e-06, 'learning_rate/action_model': 5.131240939558467e-05, 'epoch': 1.22}) +[2026-06-19 15:38:18,792][starVLA.training.train_starvla][INFO] - Step 2590, Loss: {'train/loss': 0.28146278858184814, 'train/grad_norm_pre_clip': 7.806548595428467, 'timing/data': 0.009470424964092672, 'timing/model': 1.1510974129196256, 'learning_rate/qwen_vl_interface': 5.128196494714048e-06, 'learning_rate/action_model': 5.128196494714047e-05, 'epoch': 1.22}) +[2026-06-19 15:38:23,295][starVLA.training.train_starvla][INFO] - Step 2591, Loss: {'train/loss': 0.23245808482170105, 'train/grad_norm_pre_clip': 5.900166988372803, 'timing/data': 0.010099021019414067, 'timing/model': 1.149311888962984, 'learning_rate/qwen_vl_interface': 5.125152099938448e-06, 'learning_rate/action_model': 5.125152099938447e-05, 'epoch': 1.22}) +[2026-06-19 15:38:27,824][starVLA.training.train_starvla][INFO] - Step 2592, Loss: {'train/loss': 0.2761342525482178, 'train/grad_norm_pre_clip': 8.708725929260254, 'timing/data': 0.00952594995032996, 'timing/model': 1.1450630640611053, 'learning_rate/qwen_vl_interface': 5.122107756483104e-06, 'learning_rate/action_model': 5.122107756483103e-05, 'epoch': 1.22}) +[2026-06-19 15:38:32,418][starVLA.training.train_starvla][INFO] - Step 2593, Loss: {'train/loss': 0.3487596809864044, 'train/grad_norm_pre_clip': 7.958837032318115, 'timing/data': 0.009382765041664243, 'timing/model': 1.1507126970682293, 'learning_rate/qwen_vl_interface': 5.119063465599429e-06, 'learning_rate/action_model': 5.1190634655994285e-05, 'epoch': 1.22}) +[2026-06-19 15:38:37,046][starVLA.training.train_starvla][INFO] - Step 2594, Loss: {'train/loss': 0.3076839745044708, 'train/grad_norm_pre_clip': 7.117510795593262, 'timing/data': 0.009125236072577536, 'timing/model': 1.3236933790612966, 'learning_rate/qwen_vl_interface': 5.116019228538818e-06, 'learning_rate/action_model': 5.116019228538818e-05, 'epoch': 1.22}) +[2026-06-19 15:38:41,493][starVLA.training.train_starvla][INFO] - Step 2595, Loss: {'train/loss': 0.13473351299762726, 'train/grad_norm_pre_clip': 4.283773899078369, 'timing/data': 0.009766653995029628, 'timing/model': 1.1526468259980902, 'learning_rate/qwen_vl_interface': 5.112975046552641e-06, 'learning_rate/action_model': 5.1129750465526404e-05, 'epoch': 1.22}) +[2026-06-19 15:38:46,113][starVLA.training.train_starvla][INFO] - Step 2596, Loss: {'train/loss': 0.27853018045425415, 'train/grad_norm_pre_clip': 5.4344587326049805, 'timing/data': 0.01003982697147876, 'timing/model': 1.1549280530307442, 'learning_rate/qwen_vl_interface': 5.109930920892246e-06, 'learning_rate/action_model': 5.1099309208922464e-05, 'epoch': 1.22}) +[2026-06-19 15:38:50,728][starVLA.training.train_starvla][INFO] - Step 2597, Loss: {'train/loss': 0.19506213068962097, 'train/grad_norm_pre_clip': 4.985172271728516, 'timing/data': 0.009688875987194479, 'timing/model': 1.1575194479664788, 'learning_rate/qwen_vl_interface': 5.10688685280896e-06, 'learning_rate/action_model': 5.10688685280896e-05, 'epoch': 1.22}) +[2026-06-19 15:38:55,338][starVLA.training.train_starvla][INFO] - Step 2598, Loss: {'train/loss': 0.46234798431396484, 'train/grad_norm_pre_clip': 7.328789234161377, 'timing/data': 0.009490708005614579, 'timing/model': 1.1563640219392255, 'learning_rate/qwen_vl_interface': 5.103842843554082e-06, 'learning_rate/action_model': 5.103842843554082e-05, 'epoch': 1.22}) +[2026-06-19 15:38:59,962][starVLA.training.train_starvla][INFO] - Step 2599, Loss: {'train/loss': 0.35688531398773193, 'train/grad_norm_pre_clip': 5.908636569976807, 'timing/data': 0.009512603981420398, 'timing/model': 1.1514442609623075, 'learning_rate/qwen_vl_interface': 5.100798894378891e-06, 'learning_rate/action_model': 5.100798894378891e-05, 'epoch': 1.22}) +[2026-06-19 15:39:04,588][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 2600 +[2026-06-19 15:41:53,856][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 2600 +[2026-06-19 15:41:53,856][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 2600 +[2026-06-19 15:45:02,441][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 2600 +[2026-06-19 15:45:02,442][starVLA.training.train_starvla][INFO] - Step 2600, Loss: {'train/loss': 0.32434213161468506, 'train/grad_norm_pre_clip': 9.732488632202148, 'eval/loss': 0.23893173038959503, 'eval/latency_0/loss': 0.10391278564929962, 'eval/latency_1/loss': 0.096282459795475, 'eval/latency_2/loss': 0.22431683540344238, 'eval/latency_3/loss': 0.27963536977767944, 'eval/latency_4/loss': 0.48743292689323425, 'eval/flappy/loss': 0.23893173038959503, 'eval/flappy/latency_0/loss': 0.10391278564929962, 'eval/flappy/latency_1/loss': 0.096282459795475, 'eval/flappy/latency_2/loss': 0.22431683540344238, 'eval/flappy/latency_3/loss': 0.27963536977767944, 'eval/flappy/latency_4/loss': 0.48743292689323425, 'eval/flap_precision': 0.6765676567656765, 'eval/flap_recall': 0.7454545454545455, 'eval/flap_f1': 0.7093425605536331, 'eval/flap_rate_model': 0.08416666666666667, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.7093425605536331, 'eval/latency_0/flap_precision': 0.6765676567656765, 'eval/latency_0/flap_recall': 0.7454545454545455, 'eval/latency_0/flap_f1': 0.7093425605536331, 'eval/latency_0/flap_rate_model': 0.08416666666666667, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.7093425605536331, 'timing/data': 0.0092842779122293, 'timing/model': 1.3304220930440351, 'learning_rate/qwen_vl_interface': 5.097755006534638e-06, 'learning_rate/action_model': 5.097755006534638e-05, 'epoch': 1.22}) +[2026-06-19 15:45:02,443][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_2600_state +[2026-06-19 15:45:02,445][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-19 15:45:05,738][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_2600_state/model.safetensors +[2026-06-19 15:45:13,005][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_2600_state/optimizer.bin +[2026-06-19 15:45:13,005][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_2600_state/sampler.bin +[2026-06-19 15:45:13,005][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_2600_state/sampler_1.bin +[2026-06-19 15:45:13,006][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_2600_state/random_states_0.pkl +[2026-06-19 15:45:13,029][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2600 +[2026-06-19 15:45:13,030][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-19 15:45:13,071][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-19 15:45:17,538][starVLA.training.train_starvla][INFO] - Step 2601, Loss: {'train/loss': 0.41172221302986145, 'train/grad_norm_pre_clip': 12.81710433959961, 'timing/data': 0.009885679930448532, 'timing/model': 1.1299251120071858, 'learning_rate/qwen_vl_interface': 5.094711181272551e-06, 'learning_rate/action_model': 5.094711181272551e-05, 'epoch': 1.22}) +[2026-06-19 15:45:22,086][starVLA.training.train_starvla][INFO] - Step 2602, Loss: {'train/loss': 0.4223887324333191, 'train/grad_norm_pre_clip': 11.865710258483887, 'timing/data': 0.009321171906776726, 'timing/model': 1.1267109470209107, 'learning_rate/qwen_vl_interface': 5.091667419843833e-06, 'learning_rate/action_model': 5.091667419843833e-05, 'epoch': 1.22}) +[2026-06-19 15:45:26,629][starVLA.training.train_starvla][INFO] - Step 2603, Loss: {'train/loss': 0.4099268913269043, 'train/grad_norm_pre_clip': 6.258213520050049, 'timing/data': 0.009647792088799179, 'timing/model': 1.2950683479430154, 'learning_rate/qwen_vl_interface': 5.088623723499657e-06, 'learning_rate/action_model': 5.088623723499657e-05, 'epoch': 1.22}) +[2026-06-19 15:45:31,013][starVLA.training.train_starvla][INFO] - Step 2604, Loss: {'train/loss': 0.31436094641685486, 'train/grad_norm_pre_clip': 7.816712379455566, 'timing/data': 0.009551407070830464, 'timing/model': 1.1389543620171025, 'learning_rate/qwen_vl_interface': 5.085580093491171e-06, 'learning_rate/action_model': 5.085580093491171e-05, 'epoch': 1.22}) +[2026-06-19 15:45:35,585][starVLA.training.train_starvla][INFO] - Step 2605, Loss: {'train/loss': 0.32899895310401917, 'train/grad_norm_pre_clip': 8.827437400817871, 'timing/data': 0.009701872011646628, 'timing/model': 1.148101659025997, 'learning_rate/qwen_vl_interface': 5.0825365310695e-06, 'learning_rate/action_model': 5.082536531069499e-05, 'epoch': 1.23}) +[2026-06-19 15:45:40,064][starVLA.training.train_starvla][INFO] - Step 2606, Loss: {'train/loss': 0.12278210371732712, 'train/grad_norm_pre_clip': 4.397568225860596, 'timing/data': 0.009655228932388127, 'timing/model': 1.1538607680704445, 'learning_rate/qwen_vl_interface': 5.079493037485735e-06, 'learning_rate/action_model': 5.0794930374857344e-05, 'epoch': 1.23}) +[2026-06-19 15:45:44,538][starVLA.training.train_starvla][INFO] - Step 2607, Loss: {'train/loss': 0.1607021689414978, 'train/grad_norm_pre_clip': 3.5860354900360107, 'timing/data': 0.009721655049361289, 'timing/model': 1.1355997709324583, 'learning_rate/qwen_vl_interface': 5.076449613990941e-06, 'learning_rate/action_model': 5.076449613990941e-05, 'epoch': 1.23}) +[2026-06-19 15:45:49,074][starVLA.training.train_starvla][INFO] - Step 2608, Loss: {'train/loss': 0.2609965205192566, 'train/grad_norm_pre_clip': 5.4419474601745605, 'timing/data': 0.009847886045463383, 'timing/model': 1.1370900770416483, 'learning_rate/qwen_vl_interface': 5.073406261836157e-06, 'learning_rate/action_model': 5.0734062618361565e-05, 'epoch': 1.23}) +[2026-06-19 15:45:53,653][starVLA.training.train_starvla][INFO] - Step 2609, Loss: {'train/loss': 0.2886834144592285, 'train/grad_norm_pre_clip': 6.632470607757568, 'timing/data': 0.008941754000261426, 'timing/model': 1.3214096740121022, 'learning_rate/qwen_vl_interface': 5.070362982272386e-06, 'learning_rate/action_model': 5.070362982272385e-05, 'epoch': 1.23}) +[2026-06-19 15:45:58,049][starVLA.training.train_starvla][INFO] - Step 2610, Loss: {'train/loss': 0.13618747889995575, 'train/grad_norm_pre_clip': 5.833801746368408, 'timing/data': 0.01083080901298672, 'timing/model': 1.140733642037958, 'learning_rate/qwen_vl_interface': 5.067319776550607e-06, 'learning_rate/action_model': 5.067319776550607e-05, 'epoch': 1.23}) +[2026-06-19 15:46:02,625][starVLA.training.train_starvla][INFO] - Step 2611, Loss: {'train/loss': 0.30506590008735657, 'train/grad_norm_pre_clip': 5.994353294372559, 'timing/data': 0.009594595059752464, 'timing/model': 1.1502300190040842, 'learning_rate/qwen_vl_interface': 5.064276645921768e-06, 'learning_rate/action_model': 5.064276645921768e-05, 'epoch': 1.23}) +[2026-06-19 15:46:07,289][starVLA.training.train_starvla][INFO] - Step 2612, Loss: {'train/loss': 0.40690574049949646, 'train/grad_norm_pre_clip': 8.59706974029541, 'timing/data': 0.07562151202000678, 'timing/model': 1.1547012799419463, 'learning_rate/qwen_vl_interface': 5.061233591636787e-06, 'learning_rate/action_model': 5.061233591636787e-05, 'epoch': 1.23}) +[2026-06-19 15:46:11,823][starVLA.training.train_starvla][INFO] - Step 2613, Loss: {'train/loss': 0.2394004613161087, 'train/grad_norm_pre_clip': 5.855189323425293, 'timing/data': 0.009827125002630055, 'timing/model': 1.1409951280802488, 'learning_rate/qwen_vl_interface': 5.058190614946544e-06, 'learning_rate/action_model': 5.058190614946544e-05, 'epoch': 1.23}) +[2026-06-19 15:46:16,413][starVLA.training.train_starvla][INFO] - Step 2614, Loss: {'train/loss': 0.06905460357666016, 'train/grad_norm_pre_clip': 2.5706236362457275, 'timing/data': 0.009268073015846312, 'timing/model': 1.145160125917755, 'learning_rate/qwen_vl_interface': 5.055147717101895e-06, 'learning_rate/action_model': 5.055147717101894e-05, 'epoch': 1.23}) +[2026-06-19 15:46:21,020][starVLA.training.train_starvla][INFO] - Step 2615, Loss: {'train/loss': 0.18935903906822205, 'train/grad_norm_pre_clip': 4.905001640319824, 'timing/data': 0.009286612970754504, 'timing/model': 1.3272821840364486, 'learning_rate/qwen_vl_interface': 5.052104899353658e-06, 'learning_rate/action_model': 5.0521048993536577e-05, 'epoch': 1.23}) +[2026-06-19 15:46:25,436][starVLA.training.train_starvla][INFO] - Step 2616, Loss: {'train/loss': 0.21341004967689514, 'train/grad_norm_pre_clip': 7.591124057769775, 'timing/data': 0.009257907047867775, 'timing/model': 1.143747380003333, 'learning_rate/qwen_vl_interface': 5.049062162952622e-06, 'learning_rate/action_model': 5.0490621629526214e-05, 'epoch': 1.23}) +[2026-06-19 15:46:30,024][starVLA.training.train_starvla][INFO] - Step 2617, Loss: {'train/loss': 0.4592996835708618, 'train/grad_norm_pre_clip': 12.705140113830566, 'timing/data': 0.009338991018012166, 'timing/model': 1.1422223650151864, 'learning_rate/qwen_vl_interface': 5.046019509149541e-06, 'learning_rate/action_model': 5.046019509149541e-05, 'epoch': 1.23}) +[2026-06-19 15:46:34,627][starVLA.training.train_starvla][INFO] - Step 2618, Loss: {'train/loss': 0.2473219931125641, 'train/grad_norm_pre_clip': 7.847928047180176, 'timing/data': 0.009189978009089828, 'timing/model': 1.1449411610374227, 'learning_rate/qwen_vl_interface': 5.042976939195135e-06, 'learning_rate/action_model': 5.0429769391951355e-05, 'epoch': 1.23}) +[2026-06-19 15:46:39,210][starVLA.training.train_starvla][INFO] - Step 2619, Loss: {'train/loss': 0.3931465148925781, 'train/grad_norm_pre_clip': 8.549513816833496, 'timing/data': 0.010204900056123734, 'timing/model': 1.1441108329454437, 'learning_rate/qwen_vl_interface': 5.039934454340091e-06, 'learning_rate/action_model': 5.0399344543400907e-05, 'epoch': 1.23}) +[2026-06-19 15:46:43,821][starVLA.training.train_starvla][INFO] - Step 2620, Loss: {'train/loss': 0.1560155600309372, 'train/grad_norm_pre_clip': 8.425483703613281, 'timing/data': 0.01013377693016082, 'timing/model': 1.148990115034394, 'learning_rate/qwen_vl_interface': 5.036892055835056e-06, 'learning_rate/action_model': 5.0368920558350566e-05, 'epoch': 1.23}) +[2026-06-19 15:46:48,434][starVLA.training.train_starvla][INFO] - Step 2621, Loss: {'train/loss': 0.31360071897506714, 'train/grad_norm_pre_clip': 5.076157569885254, 'timing/data': 0.009306558989919722, 'timing/model': 1.3303726649610326, 'learning_rate/qwen_vl_interface': 5.033849744930647e-06, 'learning_rate/action_model': 5.033849744930646e-05, 'epoch': 1.23}) +[2026-06-19 15:46:52,849][starVLA.training.train_starvla][INFO] - Step 2622, Loss: {'train/loss': 0.3964136242866516, 'train/grad_norm_pre_clip': 11.0853910446167, 'timing/data': 0.010047819931060076, 'timing/model': 1.1438006849493831, 'learning_rate/qwen_vl_interface': 5.030807522877443e-06, 'learning_rate/action_model': 5.030807522877443e-05, 'epoch': 1.23}) +[2026-06-19 15:46:57,445][starVLA.training.train_starvla][INFO] - Step 2623, Loss: {'train/loss': 0.37358784675598145, 'train/grad_norm_pre_clip': 12.826759338378906, 'timing/data': 0.009708791971206665, 'timing/model': 1.1467622610507533, 'learning_rate/qwen_vl_interface': 5.027765390925987e-06, 'learning_rate/action_model': 5.027765390925987e-05, 'epoch': 1.23}) +[2026-06-19 15:47:01,978][starVLA.training.train_starvla][INFO] - Step 2624, Loss: {'train/loss': 0.24491490423679352, 'train/grad_norm_pre_clip': 8.443459510803223, 'timing/data': 0.009606110979802907, 'timing/model': 1.1654484429163858, 'learning_rate/qwen_vl_interface': 5.024723350326783e-06, 'learning_rate/action_model': 5.024723350326783e-05, 'epoch': 1.23}) +[2026-06-19 15:47:06,531][starVLA.training.train_starvla][INFO] - Step 2625, Loss: {'train/loss': 0.12547622621059418, 'train/grad_norm_pre_clip': 3.5186314582824707, 'timing/data': 0.009153542923741043, 'timing/model': 1.144717558985576, 'learning_rate/qwen_vl_interface': 5.021681402330303e-06, 'learning_rate/action_model': 5.021681402330303e-05, 'epoch': 1.23}) +[2026-06-19 15:47:11,172][starVLA.training.train_starvla][INFO] - Step 2626, Loss: {'train/loss': 0.4338972270488739, 'train/grad_norm_pre_clip': 7.5928802490234375, 'timing/data': 0.009725023992359638, 'timing/model': 1.1512502419063821, 'learning_rate/qwen_vl_interface': 5.018639548186972e-06, 'learning_rate/action_model': 5.0186395481869716e-05, 'epoch': 1.24}) +[2026-06-19 15:47:15,799][starVLA.training.train_starvla][INFO] - Step 2627, Loss: {'train/loss': 0.262116402387619, 'train/grad_norm_pre_clip': 7.525321006774902, 'timing/data': 0.009237020974978805, 'timing/model': 1.3332817821064964, 'learning_rate/qwen_vl_interface': 5.015597789147184e-06, 'learning_rate/action_model': 5.015597789147184e-05, 'epoch': 1.24}) +[2026-06-19 15:47:20,369][starVLA.training.train_starvla][INFO] - Step 2628, Loss: {'train/loss': 0.23069195449352264, 'train/grad_norm_pre_clip': 5.944731712341309, 'timing/data': 0.009420413989573717, 'timing/model': 1.2816737489774823, 'learning_rate/qwen_vl_interface': 5.012556126461291e-06, 'learning_rate/action_model': 5.012556126461291e-05, 'epoch': 1.24}) +[2026-06-19 15:47:24,803][starVLA.training.train_starvla][INFO] - Step 2629, Loss: {'train/loss': 0.1927444487810135, 'train/grad_norm_pre_clip': 7.216590404510498, 'timing/data': 0.009694405016489327, 'timing/model': 1.1456586950225756, 'learning_rate/qwen_vl_interface': 5.0095145613796044e-06, 'learning_rate/action_model': 5.0095145613796046e-05, 'epoch': 1.24}) +[2026-06-19 15:47:29,398][starVLA.training.train_starvla][INFO] - Step 2630, Loss: {'train/loss': 0.16961826384067535, 'train/grad_norm_pre_clip': 5.247096538543701, 'timing/data': 0.009479218046180904, 'timing/model': 1.1471396180568263, 'learning_rate/qwen_vl_interface': 5.0064730951524e-06, 'learning_rate/action_model': 5.0064730951524e-05, 'epoch': 1.24}) +[2026-06-19 15:47:33,991][starVLA.training.train_starvla][INFO] - Step 2631, Loss: {'train/loss': 0.28235724568367004, 'train/grad_norm_pre_clip': 5.30764627456665, 'timing/data': 0.009622226003557444, 'timing/model': 1.1465579830110073, 'learning_rate/qwen_vl_interface': 5.003431729029905e-06, 'learning_rate/action_model': 5.003431729029905e-05, 'epoch': 1.24}) +[2026-06-19 15:47:38,616][starVLA.training.train_starvla][INFO] - Step 2632, Loss: {'train/loss': 0.2722441852092743, 'train/grad_norm_pre_clip': 6.824840545654297, 'timing/data': 0.009554752963595092, 'timing/model': 1.1495010970393196, 'learning_rate/qwen_vl_interface': 5.000390464262314e-06, 'learning_rate/action_model': 5.0003904642623136e-05, 'epoch': 1.24}) +[2026-06-19 15:47:43,127][starVLA.training.train_starvla][INFO] - Step 2633, Loss: {'train/loss': 0.22516578435897827, 'train/grad_norm_pre_clip': 6.430435657501221, 'timing/data': 0.009674223023466766, 'timing/model': 1.1544167259708047, 'learning_rate/qwen_vl_interface': 4.997349302099774e-06, 'learning_rate/action_model': 4.997349302099773e-05, 'epoch': 1.24}) +[2026-06-19 15:47:47,675][starVLA.training.train_starvla][INFO] - Step 2634, Loss: {'train/loss': 0.21281111240386963, 'train/grad_norm_pre_clip': 5.030966281890869, 'timing/data': 0.009636305971071124, 'timing/model': 1.2605981869855896, 'learning_rate/qwen_vl_interface': 4.994308243792391e-06, 'learning_rate/action_model': 4.994308243792391e-05, 'epoch': 1.24}) +[2026-06-19 15:47:52,123][starVLA.training.train_starvla][INFO] - Step 2635, Loss: {'train/loss': 0.3069191575050354, 'train/grad_norm_pre_clip': 9.571650505065918, 'timing/data': 0.0094330960419029, 'timing/model': 1.1630075499415398, 'learning_rate/qwen_vl_interface': 4.991267290590233e-06, 'learning_rate/action_model': 4.991267290590233e-05, 'epoch': 1.24}) +[2026-06-19 15:47:56,749][starVLA.training.train_starvla][INFO] - Step 2636, Loss: {'train/loss': 0.5454219579696655, 'train/grad_norm_pre_clip': 10.425403594970703, 'timing/data': 0.009705750970169902, 'timing/model': 1.157409748993814, 'learning_rate/qwen_vl_interface': 4.988226443743317e-06, 'learning_rate/action_model': 4.9882264437433174e-05, 'epoch': 1.24}) +[2026-06-19 15:48:01,401][starVLA.training.train_starvla][INFO] - Step 2637, Loss: {'train/loss': 0.19418786466121674, 'train/grad_norm_pre_clip': 5.659520149230957, 'timing/data': 0.009518132079392672, 'timing/model': 1.173578087007627, 'learning_rate/qwen_vl_interface': 4.985185704501623e-06, 'learning_rate/action_model': 4.985185704501623e-05, 'epoch': 1.24}) +[2026-06-19 15:48:06,048][starVLA.training.train_starvla][INFO] - Step 2638, Loss: {'train/loss': 0.10502941161394119, 'train/grad_norm_pre_clip': 4.874574661254883, 'timing/data': 0.009689601953141391, 'timing/model': 1.1751780699705705, 'learning_rate/qwen_vl_interface': 4.982145074115082e-06, 'learning_rate/action_model': 4.982145074115082e-05, 'epoch': 1.24}) +[2026-06-19 15:48:10,686][starVLA.training.train_starvla][INFO] - Step 2639, Loss: {'train/loss': 0.26974576711654663, 'train/grad_norm_pre_clip': 5.539565563201904, 'timing/data': 0.009482000023126602, 'timing/model': 1.329156400053762, 'learning_rate/qwen_vl_interface': 4.979104553833584e-06, 'learning_rate/action_model': 4.979104553833584e-05, 'epoch': 1.24}) +[2026-06-19 15:48:15,158][starVLA.training.train_starvla][INFO] - Step 2640, Loss: {'train/loss': 0.37242937088012695, 'train/grad_norm_pre_clip': 8.566668510437012, 'timing/data': 0.009841865045018494, 'timing/model': 1.16204691096209, 'learning_rate/qwen_vl_interface': 4.9760641449069724e-06, 'learning_rate/action_model': 4.976064144906972e-05, 'epoch': 1.24}) +[2026-06-19 15:48:19,787][starVLA.training.train_starvla][INFO] - Step 2641, Loss: {'train/loss': 0.2589834928512573, 'train/grad_norm_pre_clip': 6.223586082458496, 'timing/data': 0.009965343051590025, 'timing/model': 1.1585741959279403, 'learning_rate/qwen_vl_interface': 4.973023848585042e-06, 'learning_rate/action_model': 4.973023848585042e-05, 'epoch': 1.24}) +[2026-06-19 15:48:24,558][starVLA.training.train_starvla][INFO] - Step 2642, Loss: {'train/loss': 0.33317041397094727, 'train/grad_norm_pre_clip': 24.09028434753418, 'timing/data': 0.009824148961342871, 'timing/model': 1.1543189729563892, 'learning_rate/qwen_vl_interface': 4.969983666117545e-06, 'learning_rate/action_model': 4.9699836661175446e-05, 'epoch': 1.24}) +[2026-06-19 15:48:29,182][starVLA.training.train_starvla][INFO] - Step 2643, Loss: {'train/loss': 0.1526697874069214, 'train/grad_norm_pre_clip': 5.1696977615356445, 'timing/data': 0.009983529918827116, 'timing/model': 1.156356705003418, 'learning_rate/qwen_vl_interface': 4.966943598754185e-06, 'learning_rate/action_model': 4.9669435987541854e-05, 'epoch': 1.24}) +[2026-06-19 15:48:33,803][starVLA.training.train_starvla][INFO] - Step 2644, Loss: {'train/loss': 0.30815067887306213, 'train/grad_norm_pre_clip': 7.257044792175293, 'timing/data': 0.010384054039604962, 'timing/model': 1.1510222719516605, 'learning_rate/qwen_vl_interface': 4.963903647744621e-06, 'learning_rate/action_model': 4.96390364774462e-05, 'epoch': 1.24}) +[2026-06-19 15:48:38,423][starVLA.training.train_starvla][INFO] - Step 2645, Loss: {'train/loss': 0.17996713519096375, 'train/grad_norm_pre_clip': 11.302491188049316, 'timing/data': 0.009649735991843045, 'timing/model': 1.321784483967349, 'learning_rate/qwen_vl_interface': 4.9608638143384604e-06, 'learning_rate/action_model': 4.960863814338461e-05, 'epoch': 1.24}) +[2026-06-19 15:48:42,984][starVLA.training.train_starvla][INFO] - Step 2646, Loss: {'train/loss': 0.38566821813583374, 'train/grad_norm_pre_clip': 7.649104118347168, 'timing/data': 0.01169579999987036, 'timing/model': 1.2755537090124562, 'learning_rate/qwen_vl_interface': 4.9578240997852636e-06, 'learning_rate/action_model': 4.957824099785264e-05, 'epoch': 1.24}) +[2026-06-19 15:48:47,449][starVLA.training.train_starvla][INFO] - Step 2647, Loss: {'train/loss': 0.3447047173976898, 'train/grad_norm_pre_clip': 9.725406646728516, 'timing/data': 0.009343490935862064, 'timing/model': 1.1591381020843983, 'learning_rate/qwen_vl_interface': 4.954784505334544e-06, 'learning_rate/action_model': 4.9547845053345434e-05, 'epoch': 1.24}) +[2026-06-19 15:48:52,089][starVLA.training.train_starvla][INFO] - Step 2648, Loss: {'train/loss': 0.1613432765007019, 'train/grad_norm_pre_clip': 4.821383953094482, 'timing/data': 0.00956430402584374, 'timing/model': 1.1638720610644668, 'learning_rate/qwen_vl_interface': 4.951745032235762e-06, 'learning_rate/action_model': 4.951745032235762e-05, 'epoch': 1.25}) +[2026-06-19 15:48:56,703][starVLA.training.train_starvla][INFO] - Step 2649, Loss: {'train/loss': 0.2702008783817291, 'train/grad_norm_pre_clip': 9.494553565979004, 'timing/data': 0.009379995986819267, 'timing/model': 1.1556179319741204, 'learning_rate/qwen_vl_interface': 4.94870568173833e-06, 'learning_rate/action_model': 4.9487056817383307e-05, 'epoch': 1.25}) +[2026-06-19 15:49:01,305][starVLA.training.train_starvla][INFO] - Step 2650, Loss: {'train/loss': 0.4202934801578522, 'train/grad_norm_pre_clip': 11.868165969848633, 'timing/data': 0.009603866026736796, 'timing/model': 1.1541700439993292, 'learning_rate/qwen_vl_interface': 4.945666455091614e-06, 'learning_rate/action_model': 4.9456664550916136e-05, 'epoch': 1.25}) +[2026-06-19 15:49:05,914][starVLA.training.train_starvla][INFO] - Step 2651, Loss: {'train/loss': 0.2493126094341278, 'train/grad_norm_pre_clip': 5.484055042266846, 'timing/data': 0.009736138046719134, 'timing/model': 1.144475715002045, 'learning_rate/qwen_vl_interface': 4.942627353544921e-06, 'learning_rate/action_model': 4.942627353544921e-05, 'epoch': 1.25}) +[2026-06-19 15:49:10,569][starVLA.training.train_starvla][INFO] - Step 2652, Loss: {'train/loss': 0.2830599248409271, 'train/grad_norm_pre_clip': 7.368929862976074, 'timing/data': 0.010050919954665005, 'timing/model': 1.3558352590771392, 'learning_rate/qwen_vl_interface': 4.939588378347513e-06, 'learning_rate/action_model': 4.9395883783475125e-05, 'epoch': 1.25}) +[2026-06-19 15:49:15,023][starVLA.training.train_starvla][INFO] - Step 2653, Loss: {'train/loss': 0.17255397140979767, 'train/grad_norm_pre_clip': 10.810452461242676, 'timing/data': 0.009778891922906041, 'timing/model': 1.1637598261004314, 'learning_rate/qwen_vl_interface': 4.936549530748597e-06, 'learning_rate/action_model': 4.9365495307485963e-05, 'epoch': 1.25}) +[2026-06-19 15:49:19,609][starVLA.training.train_starvla][INFO] - Step 2654, Loss: {'train/loss': 0.059579335153102875, 'train/grad_norm_pre_clip': 2.471277952194214, 'timing/data': 0.008852395927533507, 'timing/model': 1.1456067969556898, 'learning_rate/qwen_vl_interface': 4.9335108119973286e-06, 'learning_rate/action_model': 4.933510811997328e-05, 'epoch': 1.25}) +[2026-06-19 15:49:24,206][starVLA.training.train_starvla][INFO] - Step 2655, Loss: {'train/loss': 0.20363391935825348, 'train/grad_norm_pre_clip': 10.146026611328125, 'timing/data': 0.009497102000750601, 'timing/model': 1.1523866249481216, 'learning_rate/qwen_vl_interface': 4.930472223342812e-06, 'learning_rate/action_model': 4.930472223342812e-05, 'epoch': 1.25}) +[2026-06-19 15:49:28,789][starVLA.training.train_starvla][INFO] - Step 2656, Loss: {'train/loss': 0.5173392295837402, 'train/grad_norm_pre_clip': 7.391138076782227, 'timing/data': 0.009924083948135376, 'timing/model': 1.149066339014098, 'learning_rate/qwen_vl_interface': 4.927433766034093e-06, 'learning_rate/action_model': 4.9274337660340934e-05, 'epoch': 1.25}) +[2026-06-19 15:49:33,408][starVLA.training.train_starvla][INFO] - Step 2657, Loss: {'train/loss': 0.17170275747776031, 'train/grad_norm_pre_clip': 6.176392555236816, 'timing/data': 0.009386221063323319, 'timing/model': 1.1539683400187641, 'learning_rate/qwen_vl_interface': 4.92439544132017e-06, 'learning_rate/action_model': 4.92439544132017e-05, 'epoch': 1.25}) +[2026-06-19 15:49:38,019][starVLA.training.train_starvla][INFO] - Step 2658, Loss: {'train/loss': 0.2674289643764496, 'train/grad_norm_pre_clip': 7.0742716789245605, 'timing/data': 0.009449015953578055, 'timing/model': 1.318118480965495, 'learning_rate/qwen_vl_interface': 4.92135725044998e-06, 'learning_rate/action_model': 4.92135725044998e-05, 'epoch': 1.25}) +[2026-06-19 15:49:42,462][starVLA.training.train_starvla][INFO] - Step 2659, Loss: {'train/loss': 0.4330413043498993, 'train/grad_norm_pre_clip': 7.469791412353516, 'timing/data': 0.00934703298844397, 'timing/model': 1.1503526490414515, 'learning_rate/qwen_vl_interface': 4.918319194672412e-06, 'learning_rate/action_model': 4.918319194672412e-05, 'epoch': 1.25}) +[2026-06-19 15:49:47,086][starVLA.training.train_starvla][INFO] - Step 2660, Loss: {'train/loss': 0.17939762771129608, 'train/grad_norm_pre_clip': 8.773122787475586, 'timing/data': 0.009551570052281022, 'timing/model': 1.1474299329565838, 'learning_rate/qwen_vl_interface': 4.9152812752362935e-06, 'learning_rate/action_model': 4.915281275236293e-05, 'epoch': 1.25}) +[2026-06-19 15:49:51,592][starVLA.training.train_starvla][INFO] - Step 2661, Loss: {'train/loss': 0.2023930698633194, 'train/grad_norm_pre_clip': 7.87576150894165, 'timing/data': 0.009264762979000807, 'timing/model': 1.1484174060169607, 'learning_rate/qwen_vl_interface': 4.9122434933904e-06, 'learning_rate/action_model': 4.9122434933904e-05, 'epoch': 1.25}) +[2026-06-19 15:49:56,128][starVLA.training.train_starvla][INFO] - Step 2662, Loss: {'train/loss': 0.16489596664905548, 'train/grad_norm_pre_clip': 5.6326985359191895, 'timing/data': 0.010672147967852652, 'timing/model': 1.1549994410015643, 'learning_rate/qwen_vl_interface': 4.909205850383449e-06, 'learning_rate/action_model': 4.9092058503834485e-05, 'epoch': 1.25}) +[2026-06-19 15:50:00,764][starVLA.training.train_starvla][INFO] - Step 2663, Loss: {'train/loss': 0.17127645015716553, 'train/grad_norm_pre_clip': 11.25748348236084, 'timing/data': 0.009706806973554194, 'timing/model': 1.155232228920795, 'learning_rate/qwen_vl_interface': 4.906168347464098e-06, 'learning_rate/action_model': 4.906168347464098e-05, 'epoch': 1.25}) +[2026-06-19 15:50:05,366][starVLA.training.train_starvla][INFO] - Step 2664, Loss: {'train/loss': 0.21856743097305298, 'train/grad_norm_pre_clip': 5.5501604080200195, 'timing/data': 0.009121701936237514, 'timing/model': 1.3178254080703482, 'learning_rate/qwen_vl_interface': 4.903130985880955e-06, 'learning_rate/action_model': 4.903130985880954e-05, 'epoch': 1.25}) +[2026-06-19 15:50:09,821][starVLA.training.train_starvla][INFO] - Step 2665, Loss: {'train/loss': 0.17297139763832092, 'train/grad_norm_pre_clip': 5.036380767822266, 'timing/data': 0.009282779996283352, 'timing/model': 1.1559394299983978, 'learning_rate/qwen_vl_interface': 4.900093766882558e-06, 'learning_rate/action_model': 4.900093766882558e-05, 'epoch': 1.25}) +[2026-06-19 15:50:14,443][starVLA.training.train_starvla][INFO] - Step 2666, Loss: {'train/loss': 0.21125417947769165, 'train/grad_norm_pre_clip': 6.094820499420166, 'timing/data': 0.00968392996583134, 'timing/model': 1.151946173980832, 'learning_rate/qwen_vl_interface': 4.897056691717397e-06, 'learning_rate/action_model': 4.8970566917173966e-05, 'epoch': 1.25}) +[2026-06-19 15:50:19,071][starVLA.training.train_starvla][INFO] - Step 2667, Loss: {'train/loss': 0.5311073064804077, 'train/grad_norm_pre_clip': 13.014699935913086, 'timing/data': 0.009568365989252925, 'timing/model': 1.1597951679723337, 'learning_rate/qwen_vl_interface': 4.894019761633898e-06, 'learning_rate/action_model': 4.894019761633898e-05, 'epoch': 1.25}) +[2026-06-19 15:50:23,693][starVLA.training.train_starvla][INFO] - Step 2668, Loss: {'train/loss': 0.11907599121332169, 'train/grad_norm_pre_clip': 5.490999221801758, 'timing/data': 0.008851894061081111, 'timing/model': 1.1548964569810778, 'learning_rate/qwen_vl_interface': 4.8909829778804304e-06, 'learning_rate/action_model': 4.89098297788043e-05, 'epoch': 1.25}) +[2026-06-19 15:50:28,311][starVLA.training.train_starvla][INFO] - Step 2669, Loss: {'train/loss': 0.29037782549858093, 'train/grad_norm_pre_clip': 8.5232572555542, 'timing/data': 0.011280677979812026, 'timing/model': 1.15175773401279, 'learning_rate/qwen_vl_interface': 4.887946341705298e-06, 'learning_rate/action_model': 4.887946341705297e-05, 'epoch': 1.26}) +[2026-06-19 15:50:32,928][starVLA.training.train_starvla][INFO] - Step 2670, Loss: {'train/loss': 0.11476783454418182, 'train/grad_norm_pre_clip': 48.444190979003906, 'timing/data': 0.009495982900261879, 'timing/model': 1.3251081220805645, 'learning_rate/qwen_vl_interface': 4.8849098543567465e-06, 'learning_rate/action_model': 4.884909854356747e-05, 'epoch': 1.26}) +[2026-06-19 15:50:37,372][starVLA.training.train_starvla][INFO] - Step 2671, Loss: {'train/loss': 0.35100671648979187, 'train/grad_norm_pre_clip': 12.940604209899902, 'timing/data': 0.009446293930523098, 'timing/model': 1.1542305949842557, 'learning_rate/qwen_vl_interface': 4.881873517082965e-06, 'learning_rate/action_model': 4.8818735170829655e-05, 'epoch': 1.26}) +[2026-06-19 15:50:41,984][starVLA.training.train_starvla][INFO] - Step 2672, Loss: {'train/loss': 0.16456946730613708, 'train/grad_norm_pre_clip': 7.07402229309082, 'timing/data': 0.009300553938373923, 'timing/model': 1.1498722289688885, 'learning_rate/qwen_vl_interface': 4.878837331132076e-06, 'learning_rate/action_model': 4.8788373311320754e-05, 'epoch': 1.26}) +[2026-06-19 15:50:46,515][starVLA.training.train_starvla][INFO] - Step 2673, Loss: {'train/loss': 0.1947900354862213, 'train/grad_norm_pre_clip': 6.118077754974365, 'timing/data': 0.009727329015731812, 'timing/model': 1.1555922130355611, 'learning_rate/qwen_vl_interface': 4.87580129775214e-06, 'learning_rate/action_model': 4.87580129775214e-05, 'epoch': 1.26}) +[2026-06-19 15:50:51,096][starVLA.training.train_starvla][INFO] - Step 2674, Loss: {'train/loss': 0.5138908624649048, 'train/grad_norm_pre_clip': 10.448196411132812, 'timing/data': 0.009779201936908066, 'timing/model': 1.1551397789735347, 'learning_rate/qwen_vl_interface': 4.872765418191156e-06, 'learning_rate/action_model': 4.872765418191155e-05, 'epoch': 1.26}) +[2026-06-19 15:50:55,743][starVLA.training.train_starvla][INFO] - Step 2675, Loss: {'train/loss': 0.13457342982292175, 'train/grad_norm_pre_clip': 4.997495174407959, 'timing/data': 0.009616275085136294, 'timing/model': 1.1486110030673444, 'learning_rate/qwen_vl_interface': 4.8697296936970575e-06, 'learning_rate/action_model': 4.869729693697057e-05, 'epoch': 1.26}) +[2026-06-19 15:51:00,358][starVLA.training.train_starvla][INFO] - Step 2676, Loss: {'train/loss': 0.26390352845191956, 'train/grad_norm_pre_clip': 7.138612747192383, 'timing/data': 0.009927797014825046, 'timing/model': 1.312967194011435, 'learning_rate/qwen_vl_interface': 4.8666941255177185e-06, 'learning_rate/action_model': 4.8666941255177185e-05, 'epoch': 1.26}) +[2026-06-19 15:51:04,921][starVLA.training.train_starvla][INFO] - Step 2677, Loss: {'train/loss': 0.19521084427833557, 'train/grad_norm_pre_clip': 7.649680137634277, 'timing/data': 0.009926705970428884, 'timing/model': 1.274347912054509, 'learning_rate/qwen_vl_interface': 4.863658714900945e-06, 'learning_rate/action_model': 4.863658714900945e-05, 'epoch': 1.26}) +[2026-06-19 15:51:09,350][starVLA.training.train_starvla][INFO] - Step 2678, Loss: {'train/loss': 0.24175342917442322, 'train/grad_norm_pre_clip': 9.255411148071289, 'timing/data': 0.009504245012067258, 'timing/model': 1.1511115719331428, 'learning_rate/qwen_vl_interface': 4.860623463094483e-06, 'learning_rate/action_model': 4.860623463094483e-05, 'epoch': 1.26}) +[2026-06-19 15:51:13,973][starVLA.training.train_starvla][INFO] - Step 2679, Loss: {'train/loss': 0.09802518784999847, 'train/grad_norm_pre_clip': 3.185637950897217, 'timing/data': 0.009595385985448956, 'timing/model': 1.1578306970186532, 'learning_rate/qwen_vl_interface': 4.857588371346006e-06, 'learning_rate/action_model': 4.8575883713460057e-05, 'epoch': 1.26}) +[2026-06-19 15:51:18,591][starVLA.training.train_starvla][INFO] - Step 2680, Loss: {'train/loss': 0.3112630844116211, 'train/grad_norm_pre_clip': 4.715795040130615, 'timing/data': 0.009358725976198912, 'timing/model': 1.152047144016251, 'learning_rate/qwen_vl_interface': 4.854553440903127e-06, 'learning_rate/action_model': 4.854553440903126e-05, 'epoch': 1.26}) +[2026-06-19 15:51:23,230][starVLA.training.train_starvla][INFO] - Step 2681, Loss: {'train/loss': 0.24903583526611328, 'train/grad_norm_pre_clip': 6.15769624710083, 'timing/data': 0.009621887002140284, 'timing/model': 1.1623215270228684, 'learning_rate/qwen_vl_interface': 4.85151867301339e-06, 'learning_rate/action_model': 4.85151867301339e-05, 'epoch': 1.26}) +[2026-06-19 15:51:27,737][starVLA.training.train_starvla][INFO] - Step 2682, Loss: {'train/loss': 0.4637667238712311, 'train/grad_norm_pre_clip': 9.3402681350708, 'timing/data': 0.008687113993801177, 'timing/model': 1.1494811259908602, 'learning_rate/qwen_vl_interface': 4.848484068924274e-06, 'learning_rate/action_model': 4.848484068924275e-05, 'epoch': 1.26}) +[2026-06-19 15:51:32,286][starVLA.training.train_starvla][INFO] - Step 2683, Loss: {'train/loss': 0.25984588265419006, 'train/grad_norm_pre_clip': 8.059920310974121, 'timing/data': 0.009288287023082376, 'timing/model': 1.2576884320005774, 'learning_rate/qwen_vl_interface': 4.845449629883193e-06, 'learning_rate/action_model': 4.8454496298831934e-05, 'epoch': 1.26}) +[2026-06-19 15:51:36,731][starVLA.training.train_starvla][INFO] - Step 2684, Loss: {'train/loss': 0.16862048208713531, 'train/grad_norm_pre_clip': 8.359146118164062, 'timing/data': 0.009891476016491652, 'timing/model': 1.1456220229156315, 'learning_rate/qwen_vl_interface': 4.842415357137487e-06, 'learning_rate/action_model': 4.842415357137487e-05, 'epoch': 1.26}) +[2026-06-19 15:51:41,341][starVLA.training.train_starvla][INFO] - Step 2685, Loss: {'train/loss': 0.2803947329521179, 'train/grad_norm_pre_clip': 9.785421371459961, 'timing/data': 0.008905261987820268, 'timing/model': 1.148831319063902, 'learning_rate/qwen_vl_interface': 4.839381251934432e-06, 'learning_rate/action_model': 4.8393812519344316e-05, 'epoch': 1.26}) +[2026-06-19 15:51:45,931][starVLA.training.train_starvla][INFO] - Step 2686, Loss: {'train/loss': 0.11356212943792343, 'train/grad_norm_pre_clip': 6.349389553070068, 'timing/data': 0.009705892065539956, 'timing/model': 1.1444053230807185, 'learning_rate/qwen_vl_interface': 4.836347315521233e-06, 'learning_rate/action_model': 4.836347315521233e-05, 'epoch': 1.26}) +[2026-06-19 15:51:50,523][starVLA.training.train_starvla][INFO] - Step 2687, Loss: {'train/loss': 0.533907949924469, 'train/grad_norm_pre_clip': 9.939367294311523, 'timing/data': 0.00965491496026516, 'timing/model': 1.1475028810091317, 'learning_rate/qwen_vl_interface': 4.833313549145029e-06, 'learning_rate/action_model': 4.833313549145029e-05, 'epoch': 1.26}) +[2026-06-19 15:51:55,161][starVLA.training.train_starvla][INFO] - Step 2688, Loss: {'train/loss': 0.33604174852371216, 'train/grad_norm_pre_clip': 13.860628128051758, 'timing/data': 0.008875229046680033, 'timing/model': 1.3260526079684496, 'learning_rate/qwen_vl_interface': 4.830279954052884e-06, 'learning_rate/action_model': 4.8302799540528836e-05, 'epoch': 1.26}) +[2026-06-19 15:51:59,609][starVLA.training.train_starvla][INFO] - Step 2689, Loss: {'train/loss': 0.34979525208473206, 'train/grad_norm_pre_clip': 12.397239685058594, 'timing/data': 0.009104963042773306, 'timing/model': 1.1460830860305578, 'learning_rate/qwen_vl_interface': 4.827246531491797e-06, 'learning_rate/action_model': 4.827246531491797e-05, 'epoch': 1.26}) +[2026-06-19 15:52:04,218][starVLA.training.train_starvla][INFO] - Step 2690, Loss: {'train/loss': 0.397337406873703, 'train/grad_norm_pre_clip': 10.826272964477539, 'timing/data': 0.0091834970517084, 'timing/model': 1.150500415940769, 'learning_rate/qwen_vl_interface': 4.824213282708691e-06, 'learning_rate/action_model': 4.824213282708691e-05, 'epoch': 1.27}) +[2026-06-19 15:52:08,813][starVLA.training.train_starvla][INFO] - Step 2691, Loss: {'train/loss': 0.20221813023090363, 'train/grad_norm_pre_clip': 7.526240348815918, 'timing/data': 0.009321085060946643, 'timing/model': 1.1561379140475765, 'learning_rate/qwen_vl_interface': 4.821180208950422e-06, 'learning_rate/action_model': 4.821180208950421e-05, 'epoch': 1.27}) +[2026-06-19 15:52:13,417][starVLA.training.train_starvla][INFO] - Step 2692, Loss: {'train/loss': 0.1895589828491211, 'train/grad_norm_pre_clip': 5.8762078285217285, 'timing/data': 0.009361125063151121, 'timing/model': 1.156221145996824, 'learning_rate/qwen_vl_interface': 4.81814731146377e-06, 'learning_rate/action_model': 4.8181473114637695e-05, 'epoch': 1.27}) +[2026-06-19 15:52:18,039][starVLA.training.train_starvla][INFO] - Step 2693, Loss: {'train/loss': 0.37375926971435547, 'train/grad_norm_pre_clip': 11.629720687866211, 'timing/data': 0.009515899000689387, 'timing/model': 1.158400168060325, 'learning_rate/qwen_vl_interface': 4.8151145914954465e-06, 'learning_rate/action_model': 4.815114591495446e-05, 'epoch': 1.27}) +[2026-06-19 15:52:22,660][starVLA.training.train_starvla][INFO] - Step 2694, Loss: {'train/loss': 0.26116350293159485, 'train/grad_norm_pre_clip': 5.177184581756592, 'timing/data': 0.009421830996870995, 'timing/model': 1.32068951590918, 'learning_rate/qwen_vl_interface': 4.812082050292088e-06, 'learning_rate/action_model': 4.8120820502920876e-05, 'epoch': 1.27}) +[2026-06-19 15:52:27,241][starVLA.training.train_starvla][INFO] - Step 2695, Loss: {'train/loss': 0.45073291659355164, 'train/grad_norm_pre_clip': 9.902336120605469, 'timing/data': 0.00937554205302149, 'timing/model': 1.2710657520219684, 'learning_rate/qwen_vl_interface': 4.809049689100256e-06, 'learning_rate/action_model': 4.8090496891002555e-05, 'epoch': 1.27}) +[2026-06-19 15:52:31,690][starVLA.training.train_starvla][INFO] - Step 2696, Loss: {'train/loss': 0.287297785282135, 'train/grad_norm_pre_clip': 8.741640090942383, 'timing/data': 0.008624578011222184, 'timing/model': 1.1500735129229724, 'learning_rate/qwen_vl_interface': 4.80601750916644e-06, 'learning_rate/action_model': 4.8060175091664404e-05, 'epoch': 1.27}) +[2026-06-19 15:52:36,315][starVLA.training.train_starvla][INFO] - Step 2697, Loss: {'train/loss': 0.3691551685333252, 'train/grad_norm_pre_clip': 8.946976661682129, 'timing/data': 0.009769674972631037, 'timing/model': 1.1473206860246137, 'learning_rate/qwen_vl_interface': 4.802985511737057e-06, 'learning_rate/action_model': 4.802985511737057e-05, 'epoch': 1.27}) +[2026-06-19 15:52:40,966][starVLA.training.train_starvla][INFO] - Step 2698, Loss: {'train/loss': 0.12149341404438019, 'train/grad_norm_pre_clip': 5.930470943450928, 'timing/data': 0.009280838072299957, 'timing/model': 1.157961248070933, 'learning_rate/qwen_vl_interface': 4.799953698058444e-06, 'learning_rate/action_model': 4.7999536980584434e-05, 'epoch': 1.27}) +[2026-06-19 15:52:45,597][starVLA.training.train_starvla][INFO] - Step 2699, Loss: {'train/loss': 0.26099061965942383, 'train/grad_norm_pre_clip': 9.857489585876465, 'timing/data': 0.00938176002819091, 'timing/model': 1.1532150129787624, 'learning_rate/qwen_vl_interface': 4.796922069376867e-06, 'learning_rate/action_model': 4.796922069376867e-05, 'epoch': 1.27}) +[2026-06-19 15:52:50,196][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 2700 +[2026-06-19 15:55:43,845][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 2700 +[2026-06-19 15:55:43,846][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 2700 +[2026-06-19 15:58:53,242][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 2700 +[2026-06-19 15:58:53,243][starVLA.training.train_starvla][INFO] - Step 2700, Loss: {'train/loss': 0.24916359782218933, 'train/grad_norm_pre_clip': 7.0249223709106445, 'eval/loss': 0.2099471092224121, 'eval/latency_0/loss': 0.09969814866781235, 'eval/latency_1/loss': 0.07993779331445694, 'eval/latency_2/loss': 0.1557457000017166, 'eval/latency_3/loss': 0.24528393149375916, 'eval/latency_4/loss': 0.4692467153072357, 'eval/flappy/loss': 0.2099471092224121, 'eval/flappy/latency_0/loss': 0.09969814866781235, 'eval/flappy/latency_1/loss': 0.07993779331445694, 'eval/flappy/latency_2/loss': 0.1557457000017166, 'eval/flappy/latency_3/loss': 0.24528393149375916, 'eval/flappy/latency_4/loss': 0.4692467153072357, 'eval/flap_precision': 0.7583333333333333, 'eval/flap_recall': 0.6618181818181819, 'eval/flap_f1': 0.7067961165048544, 'eval/flap_rate_model': 0.06666666666666667, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.7067961165048544, 'eval/latency_0/flap_precision': 0.7583333333333333, 'eval/latency_0/flap_recall': 0.6618181818181819, 'eval/latency_0/flap_f1': 0.7067961165048544, 'eval/latency_0/flap_rate_model': 0.06666666666666667, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.7067961165048544, 'timing/data': 0.009341032011434436, 'timing/model': 1.1514292350038886, 'learning_rate/qwen_vl_interface': 4.793890626938511e-06, 'learning_rate/action_model': 4.793890626938511e-05, 'epoch': 1.27}) +[2026-06-19 15:58:53,244][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_2700_state +[2026-06-19 15:58:53,246][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-19 15:58:56,637][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_2700_state/model.safetensors +[2026-06-19 15:59:04,128][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_2700_state/optimizer.bin +[2026-06-19 15:59:04,128][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_2700_state/sampler.bin +[2026-06-19 15:59:04,129][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_2700_state/sampler_1.bin +[2026-06-19 15:59:04,129][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_2700_state/random_states_0.pkl +[2026-06-19 15:59:04,157][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2700 +[2026-06-19 15:59:04,158][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-19 15:59:04,187][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-19 15:59:08,714][starVLA.training.train_starvla][INFO] - Step 2701, Loss: {'train/loss': 0.26201897859573364, 'train/grad_norm_pre_clip': 8.56936264038086, 'timing/data': 0.010054570040665567, 'timing/model': 1.302417177008465, 'learning_rate/qwen_vl_interface': 4.7908593719894914e-06, 'learning_rate/action_model': 4.790859371989491e-05, 'epoch': 1.27}) +[2026-06-19 15:59:13,240][starVLA.training.train_starvla][INFO] - Step 2702, Loss: {'train/loss': 0.474374920129776, 'train/grad_norm_pre_clip': 10.848925590515137, 'timing/data': 0.010093888966366649, 'timing/model': 1.3051534210098907, 'learning_rate/qwen_vl_interface': 4.78782830577584e-06, 'learning_rate/action_model': 4.7878283057758396e-05, 'epoch': 1.27}) +[2026-06-19 15:59:17,612][starVLA.training.train_starvla][INFO] - Step 2703, Loss: {'train/loss': 0.13009117543697357, 'train/grad_norm_pre_clip': 5.84541654586792, 'timing/data': 0.010453931987285614, 'timing/model': 1.1301736809546128, 'learning_rate/qwen_vl_interface': 4.784797429543513e-06, 'learning_rate/action_model': 4.784797429543513e-05, 'epoch': 1.27}) +[2026-06-19 15:59:22,185][starVLA.training.train_starvla][INFO] - Step 2704, Loss: {'train/loss': 0.21522565186023712, 'train/grad_norm_pre_clip': 6.8477277755737305, 'timing/data': 0.009794753044843674, 'timing/model': 1.1319160889834166, 'learning_rate/qwen_vl_interface': 4.781766744538393e-06, 'learning_rate/action_model': 4.781766744538393e-05, 'epoch': 1.27}) +[2026-06-19 15:59:26,735][starVLA.training.train_starvla][INFO] - Step 2705, Loss: {'train/loss': 0.4002017378807068, 'train/grad_norm_pre_clip': 20.118642807006836, 'timing/data': 0.01016429194714874, 'timing/model': 1.1377622809959576, 'learning_rate/qwen_vl_interface': 4.778736252006277e-06, 'learning_rate/action_model': 4.7787362520062765e-05, 'epoch': 1.27}) +[2026-06-19 15:59:31,284][starVLA.training.train_starvla][INFO] - Step 2706, Loss: {'train/loss': 0.3542928695678711, 'train/grad_norm_pre_clip': 13.600400924682617, 'timing/data': 0.010235164081677794, 'timing/model': 1.13722278107889, 'learning_rate/qwen_vl_interface': 4.775705953192885e-06, 'learning_rate/action_model': 4.7757059531928856e-05, 'epoch': 1.27}) +[2026-06-19 15:59:35,832][starVLA.training.train_starvla][INFO] - Step 2707, Loss: {'train/loss': 0.34847670793533325, 'train/grad_norm_pre_clip': 7.062028408050537, 'timing/data': 0.01031870604492724, 'timing/model': 1.308257378055714, 'learning_rate/qwen_vl_interface': 4.7726758493438615e-06, 'learning_rate/action_model': 4.7726758493438613e-05, 'epoch': 1.27}) +[2026-06-19 15:59:40,214][starVLA.training.train_starvla][INFO] - Step 2708, Loss: {'train/loss': 0.3108997344970703, 'train/grad_norm_pre_clip': 8.187065124511719, 'timing/data': 0.00996423198375851, 'timing/model': 1.1383787100203335, 'learning_rate/qwen_vl_interface': 4.769645941704766e-06, 'learning_rate/action_model': 4.769645941704766e-05, 'epoch': 1.27}) +[2026-06-19 15:59:44,780][starVLA.training.train_starvla][INFO] - Step 2709, Loss: {'train/loss': 0.2446601241827011, 'train/grad_norm_pre_clip': 9.929805755615234, 'timing/data': 0.010605745017528534, 'timing/model': 1.1350888090673834, 'learning_rate/qwen_vl_interface': 4.766616231521081e-06, 'learning_rate/action_model': 4.7666162315210806e-05, 'epoch': 1.27}) +[2026-06-19 15:59:49,307][starVLA.training.train_starvla][INFO] - Step 2710, Loss: {'train/loss': 0.39168405532836914, 'train/grad_norm_pre_clip': 10.066298484802246, 'timing/data': 0.010648068971931934, 'timing/model': 1.142355116084218, 'learning_rate/qwen_vl_interface': 4.763586720038202e-06, 'learning_rate/action_model': 4.7635867200382014e-05, 'epoch': 1.27}) +[2026-06-19 15:59:53,898][starVLA.training.train_starvla][INFO] - Step 2711, Loss: {'train/loss': 0.3011418282985687, 'train/grad_norm_pre_clip': 9.536792755126953, 'timing/data': 0.010329817072488368, 'timing/model': 1.1463462250540033, 'learning_rate/qwen_vl_interface': 4.76055740850145e-06, 'learning_rate/action_model': 4.76055740850145e-05, 'epoch': 1.28}) +[2026-06-19 15:59:58,471][starVLA.training.train_starvla][INFO] - Step 2712, Loss: {'train/loss': 0.19700568914413452, 'train/grad_norm_pre_clip': 6.873752117156982, 'timing/data': 0.010745599051006138, 'timing/model': 1.1443293209886178, 'learning_rate/qwen_vl_interface': 4.757528298156058e-06, 'learning_rate/action_model': 4.757528298156058e-05, 'epoch': 1.28}) +[2026-06-19 16:00:03,045][starVLA.training.train_starvla][INFO] - Step 2713, Loss: {'train/loss': 0.176788330078125, 'train/grad_norm_pre_clip': 7.657266139984131, 'timing/data': 0.010223241988569498, 'timing/model': 1.3019988390151411, 'learning_rate/qwen_vl_interface': 4.754499390247181e-06, 'learning_rate/action_model': 4.754499390247181e-05, 'epoch': 1.28}) +[2026-06-19 16:00:07,630][starVLA.training.train_starvla][INFO] - Step 2714, Loss: {'train/loss': 0.30536213517189026, 'train/grad_norm_pre_clip': 8.711994171142578, 'timing/data': 0.010735715040937066, 'timing/model': 1.3198809649329633, 'learning_rate/qwen_vl_interface': 4.751470686019887e-06, 'learning_rate/action_model': 4.751470686019888e-05, 'epoch': 1.28}) +[2026-06-19 16:00:12,044][starVLA.training.train_starvla][INFO] - Step 2715, Loss: {'train/loss': 0.37265482544898987, 'train/grad_norm_pre_clip': 8.573821067810059, 'timing/data': 0.00960998295340687, 'timing/model': 1.1411252750549465, 'learning_rate/qwen_vl_interface': 4.7484421867191645e-06, 'learning_rate/action_model': 4.748442186719165e-05, 'epoch': 1.28}) +[2026-06-19 16:00:16,629][starVLA.training.train_starvla][INFO] - Step 2716, Loss: {'train/loss': 0.16596685349941254, 'train/grad_norm_pre_clip': 5.981624603271484, 'timing/data': 0.012593626044690609, 'timing/model': 1.1546710199909285, 'learning_rate/qwen_vl_interface': 4.745413893589913e-06, 'learning_rate/action_model': 4.7454138935899125e-05, 'epoch': 1.28}) +[2026-06-19 16:00:21,195][starVLA.training.train_starvla][INFO] - Step 2717, Loss: {'train/loss': 0.26856696605682373, 'train/grad_norm_pre_clip': 8.95347785949707, 'timing/data': 0.009937744005583227, 'timing/model': 1.1409818390384316, 'learning_rate/qwen_vl_interface': 4.7423858078769506e-06, 'learning_rate/action_model': 4.742385807876951e-05, 'epoch': 1.28}) +[2026-06-19 16:00:25,766][starVLA.training.train_starvla][INFO] - Step 2718, Loss: {'train/loss': 0.20078527927398682, 'train/grad_norm_pre_clip': 6.846467971801758, 'timing/data': 0.01048338005784899, 'timing/model': 1.142382059013471, 'learning_rate/qwen_vl_interface': 4.739357930825008e-06, 'learning_rate/action_model': 4.739357930825008e-05, 'epoch': 1.28}) +[2026-06-19 16:00:30,359][starVLA.training.train_starvla][INFO] - Step 2719, Loss: {'train/loss': 0.6001400351524353, 'train/grad_norm_pre_clip': 15.267949104309082, 'timing/data': 0.010169891989789903, 'timing/model': 1.1446653010789305, 'learning_rate/qwen_vl_interface': 4.736330263678731e-06, 'learning_rate/action_model': 4.736330263678731e-05, 'epoch': 1.28}) +[2026-06-19 16:00:34,953][starVLA.training.train_starvla][INFO] - Step 2720, Loss: {'train/loss': 0.3961591422557831, 'train/grad_norm_pre_clip': 9.848551750183105, 'timing/data': 0.009847869980148971, 'timing/model': 1.3183558359742165, 'learning_rate/qwen_vl_interface': 4.733302807682679e-06, 'learning_rate/action_model': 4.733302807682679e-05, 'epoch': 1.28}) +[2026-06-19 16:00:39,364][starVLA.training.train_starvla][INFO] - Step 2721, Loss: {'train/loss': 0.2770513594150543, 'train/grad_norm_pre_clip': 8.673249244689941, 'timing/data': 0.010275266016833484, 'timing/model': 1.1414438500069082, 'learning_rate/qwen_vl_interface': 4.7302755640813255e-06, 'learning_rate/action_model': 4.7302755640813256e-05, 'epoch': 1.28}) +[2026-06-19 16:00:43,943][starVLA.training.train_starvla][INFO] - Step 2722, Loss: {'train/loss': 0.25724664330482483, 'train/grad_norm_pre_clip': 6.770183086395264, 'timing/data': 0.00993646599818021, 'timing/model': 1.1474340170389041, 'learning_rate/qwen_vl_interface': 4.727248534119058e-06, 'learning_rate/action_model': 4.7272485341190574e-05, 'epoch': 1.28}) +[2026-06-19 16:00:48,532][starVLA.training.train_starvla][INFO] - Step 2723, Loss: {'train/loss': 0.3621066212654114, 'train/grad_norm_pre_clip': 11.895781517028809, 'timing/data': 0.010601822985336185, 'timing/model': 1.1496890409616753, 'learning_rate/qwen_vl_interface': 4.72422171904017e-06, 'learning_rate/action_model': 4.7242217190401697e-05, 'epoch': 1.28}) +[2026-06-19 16:00:53,145][starVLA.training.train_starvla][INFO] - Step 2724, Loss: {'train/loss': 0.27977505326271057, 'train/grad_norm_pre_clip': 12.298996925354004, 'timing/data': 0.010418624966405332, 'timing/model': 1.1483463920885697, 'learning_rate/qwen_vl_interface': 4.721195120088873e-06, 'learning_rate/action_model': 4.721195120088873e-05, 'epoch': 1.28}) +[2026-06-19 16:00:57,762][starVLA.training.train_starvla][INFO] - Step 2725, Loss: {'train/loss': 0.3140422999858856, 'train/grad_norm_pre_clip': 6.3785858154296875, 'timing/data': 0.010298799024894834, 'timing/model': 1.144866545102559, 'learning_rate/qwen_vl_interface': 4.718168738509288e-06, 'learning_rate/action_model': 4.718168738509288e-05, 'epoch': 1.28}) +[2026-06-19 16:01:02,376][starVLA.training.train_starvla][INFO] - Step 2726, Loss: {'train/loss': 0.10584631562232971, 'train/grad_norm_pre_clip': 3.3774826526641846, 'timing/data': 0.010225519072264433, 'timing/model': 1.323273352929391, 'learning_rate/qwen_vl_interface': 4.715142575545445e-06, 'learning_rate/action_model': 4.7151425755454454e-05, 'epoch': 1.28}) +[2026-06-19 16:01:06,913][starVLA.training.train_starvla][INFO] - Step 2727, Loss: {'train/loss': 0.17267824709415436, 'train/grad_norm_pre_clip': 6.98990535736084, 'timing/data': 0.010685462970286608, 'timing/model': 1.2376107870368287, 'learning_rate/qwen_vl_interface': 4.712116632441287e-06, 'learning_rate/action_model': 4.7121166324412866e-05, 'epoch': 1.28}) +[2026-06-19 16:01:11,402][starVLA.training.train_starvla][INFO] - Step 2728, Loss: {'train/loss': 0.19951678812503815, 'train/grad_norm_pre_clip': 7.856252193450928, 'timing/data': 0.010126149980351329, 'timing/model': 1.1692361750174314, 'learning_rate/qwen_vl_interface': 4.709090910440664e-06, 'learning_rate/action_model': 4.7090909104406634e-05, 'epoch': 1.28}) +[2026-06-19 16:01:16,046][starVLA.training.train_starvla][INFO] - Step 2729, Loss: {'train/loss': 0.3538447916507721, 'train/grad_norm_pre_clip': 8.381152153015137, 'timing/data': 0.010640360065735877, 'timing/model': 1.1714905980043113, 'learning_rate/qwen_vl_interface': 4.7060654107873335e-06, 'learning_rate/action_model': 4.7060654107873335e-05, 'epoch': 1.28}) +[2026-06-19 16:01:20,646][starVLA.training.train_starvla][INFO] - Step 2730, Loss: {'train/loss': 0.09659755975008011, 'train/grad_norm_pre_clip': 3.104572296142578, 'timing/data': 0.010423753992654383, 'timing/model': 1.1490549349691719, 'learning_rate/qwen_vl_interface': 4.7030401347249685e-06, 'learning_rate/action_model': 4.703040134724968e-05, 'epoch': 1.28}) +[2026-06-19 16:01:25,245][starVLA.training.train_starvla][INFO] - Step 2731, Loss: {'train/loss': 0.2615118622779846, 'train/grad_norm_pre_clip': 6.293524742126465, 'timing/data': 0.00976822292432189, 'timing/model': 1.1478159399703145, 'learning_rate/qwen_vl_interface': 4.700015083497143e-06, 'learning_rate/action_model': 4.700015083497142e-05, 'epoch': 1.28}) +[2026-06-19 16:01:29,860][starVLA.training.train_starvla][INFO] - Step 2732, Loss: {'train/loss': 0.36084797978401184, 'train/grad_norm_pre_clip': 11.520720481872559, 'timing/data': 0.009767860057763755, 'timing/model': 1.3283450240269303, 'learning_rate/qwen_vl_interface': 4.696990258347342e-06, 'learning_rate/action_model': 4.696990258347342e-05, 'epoch': 1.28}) +[2026-06-19 16:01:34,360][starVLA.training.train_starvla][INFO] - Step 2733, Loss: {'train/loss': 0.283253937959671, 'train/grad_norm_pre_clip': 10.68101978302002, 'timing/data': 0.010015161940827966, 'timing/model': 1.2148949169786647, 'learning_rate/qwen_vl_interface': 4.693965660518957e-06, 'learning_rate/action_model': 4.693965660518957e-05, 'epoch': 1.29}) +[2026-06-19 16:01:38,787][starVLA.training.train_starvla][INFO] - Step 2734, Loss: {'train/loss': 0.5654783248901367, 'train/grad_norm_pre_clip': 20.36456871032715, 'timing/data': 0.010101890075020492, 'timing/model': 1.1497250109678134, 'learning_rate/qwen_vl_interface': 4.690941291255285e-06, 'learning_rate/action_model': 4.6909412912552844e-05, 'epoch': 1.29}) +[2026-06-19 16:01:43,376][starVLA.training.train_starvla][INFO] - Step 2735, Loss: {'train/loss': 0.1496109962463379, 'train/grad_norm_pre_clip': 8.64334487915039, 'timing/data': 0.010085194953717291, 'timing/model': 1.1482005449943244, 'learning_rate/qwen_vl_interface': 4.68791715179953e-06, 'learning_rate/action_model': 4.68791715179953e-05, 'epoch': 1.29}) +[2026-06-19 16:01:47,966][starVLA.training.train_starvla][INFO] - Step 2736, Loss: {'train/loss': 0.34814178943634033, 'train/grad_norm_pre_clip': 14.45256519317627, 'timing/data': 0.010535344015806913, 'timing/model': 1.1480987379327416, 'learning_rate/qwen_vl_interface': 4.684893243394804e-06, 'learning_rate/action_model': 4.6848932433948036e-05, 'epoch': 1.29}) +[2026-06-19 16:01:52,562][starVLA.training.train_starvla][INFO] - Step 2737, Loss: {'train/loss': 0.3953290581703186, 'train/grad_norm_pre_clip': 13.446802139282227, 'timing/data': 0.01011520007159561, 'timing/model': 1.1488366540288553, 'learning_rate/qwen_vl_interface': 4.68186956728412e-06, 'learning_rate/action_model': 4.68186956728412e-05, 'epoch': 1.29}) +[2026-06-19 16:01:57,163][starVLA.training.train_starvla][INFO] - Step 2738, Loss: {'train/loss': 0.6768160462379456, 'train/grad_norm_pre_clip': 25.322452545166016, 'timing/data': 0.01021418604068458, 'timing/model': 1.3155347439460456, 'learning_rate/qwen_vl_interface': 4.678846124710395e-06, 'learning_rate/action_model': 4.678846124710394e-05, 'epoch': 1.29}) +[2026-06-19 16:02:01,661][starVLA.training.train_starvla][INFO] - Step 2739, Loss: {'train/loss': 0.14182499051094055, 'train/grad_norm_pre_clip': 8.870895385742188, 'timing/data': 0.010513223009184003, 'timing/model': 1.214700563927181, 'learning_rate/qwen_vl_interface': 4.675822916916454e-06, 'learning_rate/action_model': 4.6758229169164536e-05, 'epoch': 1.29}) +[2026-06-19 16:02:06,120][starVLA.training.train_starvla][INFO] - Step 2740, Loss: {'train/loss': 0.08053206652402878, 'train/grad_norm_pre_clip': 3.6400108337402344, 'timing/data': 0.0103109689662233, 'timing/model': 1.1497367849806324, 'learning_rate/qwen_vl_interface': 4.6727999451450236e-06, 'learning_rate/action_model': 4.672799945145023e-05, 'epoch': 1.29}) +[2026-06-19 16:02:10,735][starVLA.training.train_starvla][INFO] - Step 2741, Loss: {'train/loss': 0.37404295802116394, 'train/grad_norm_pre_clip': 10.676889419555664, 'timing/data': 0.010621020919643342, 'timing/model': 1.1524852180154994, 'learning_rate/qwen_vl_interface': 4.669777210638733e-06, 'learning_rate/action_model': 4.669777210638733e-05, 'epoch': 1.29}) +[2026-06-19 16:02:15,346][starVLA.training.train_starvla][INFO] - Step 2742, Loss: {'train/loss': 0.16914206743240356, 'train/grad_norm_pre_clip': 5.2344770431518555, 'timing/data': 0.010600092937238514, 'timing/model': 1.1455309140728787, 'learning_rate/qwen_vl_interface': 4.666754714640115e-06, 'learning_rate/action_model': 4.666754714640115e-05, 'epoch': 1.29}) +[2026-06-19 16:02:19,947][starVLA.training.train_starvla][INFO] - Step 2743, Loss: {'train/loss': 0.25314047932624817, 'train/grad_norm_pre_clip': 6.750460147857666, 'timing/data': 0.011276495992206037, 'timing/model': 1.1459464859217405, 'learning_rate/qwen_vl_interface': 4.663732458391602e-06, 'learning_rate/action_model': 4.663732458391601e-05, 'epoch': 1.29}) +[2026-06-19 16:02:24,550][starVLA.training.train_starvla][INFO] - Step 2744, Loss: {'train/loss': 0.17004750669002533, 'train/grad_norm_pre_clip': 7.368826866149902, 'timing/data': 0.010492703062482178, 'timing/model': 1.3164072070503607, 'learning_rate/qwen_vl_interface': 4.6607104431355295e-06, 'learning_rate/action_model': 4.66071044313553e-05, 'epoch': 1.29}) +[2026-06-19 16:02:29,042][starVLA.training.train_starvla][INFO] - Step 2745, Loss: {'train/loss': 0.2705359160900116, 'train/grad_norm_pre_clip': 6.510215759277344, 'timing/data': 0.010072555975057185, 'timing/model': 1.2143191139912233, 'learning_rate/qwen_vl_interface': 4.657688670114135e-06, 'learning_rate/action_model': 4.657688670114135e-05, 'epoch': 1.29}) +[2026-06-19 16:02:33,500][starVLA.training.train_starvla][INFO] - Step 2746, Loss: {'train/loss': 0.16967922449111938, 'train/grad_norm_pre_clip': 6.434675693511963, 'timing/data': 0.012416526908054948, 'timing/model': 1.1662959440145642, 'learning_rate/qwen_vl_interface': 4.654667140569555e-06, 'learning_rate/action_model': 4.6546671405695544e-05, 'epoch': 1.29}) +[2026-06-19 16:02:38,122][starVLA.training.train_starvla][INFO] - Step 2747, Loss: {'train/loss': 0.16847246885299683, 'train/grad_norm_pre_clip': 5.465282917022705, 'timing/data': 0.01074912108015269, 'timing/model': 1.1478866679826751, 'learning_rate/qwen_vl_interface': 4.651645855743825e-06, 'learning_rate/action_model': 4.651645855743825e-05, 'epoch': 1.29}) +[2026-06-19 16:02:42,731][starVLA.training.train_starvla][INFO] - Step 2748, Loss: {'train/loss': 0.2740796208381653, 'train/grad_norm_pre_clip': 5.507599353790283, 'timing/data': 0.010215696995146573, 'timing/model': 1.153672605054453, 'learning_rate/qwen_vl_interface': 4.648624816878883e-06, 'learning_rate/action_model': 4.6486248168788825e-05, 'epoch': 1.29}) +[2026-06-19 16:02:47,351][starVLA.training.train_starvla][INFO] - Step 2749, Loss: {'train/loss': 0.3053930103778839, 'train/grad_norm_pre_clip': 22.030372619628906, 'timing/data': 0.01026914909016341, 'timing/model': 1.1517745879245922, 'learning_rate/qwen_vl_interface': 4.6456040252165615e-06, 'learning_rate/action_model': 4.645604025216562e-05, 'epoch': 1.29}) +[2026-06-19 16:02:51,953][starVLA.training.train_starvla][INFO] - Step 2750, Loss: {'train/loss': 0.2816482186317444, 'train/grad_norm_pre_clip': 8.79041862487793, 'timing/data': 0.01076380698941648, 'timing/model': 1.316205281065777, 'learning_rate/qwen_vl_interface': 4.642583481998596e-06, 'learning_rate/action_model': 4.642583481998596e-05, 'epoch': 1.29}) +[2026-06-19 16:02:56,471][starVLA.training.train_starvla][INFO] - Step 2751, Loss: {'train/loss': 0.3025244176387787, 'train/grad_norm_pre_clip': 9.217543601989746, 'timing/data': 0.010574529063887894, 'timing/model': 1.2345808390527964, 'learning_rate/qwen_vl_interface': 4.639563188466616e-06, 'learning_rate/action_model': 4.6395631884666165e-05, 'epoch': 1.29}) +[2026-06-19 16:03:00,968][starVLA.training.train_starvla][INFO] - Step 2752, Loss: {'train/loss': 0.3097209930419922, 'train/grad_norm_pre_clip': 6.033722400665283, 'timing/data': 0.010438188095577061, 'timing/model': 1.1545606839936227, 'learning_rate/qwen_vl_interface': 4.636543145862152e-06, 'learning_rate/action_model': 4.6365431458621523e-05, 'epoch': 1.29}) +[2026-06-19 16:03:05,647][starVLA.training.train_starvla][INFO] - Step 2753, Loss: {'train/loss': 0.1837957352399826, 'train/grad_norm_pre_clip': 10.075704574584961, 'timing/data': 0.01163466100115329, 'timing/model': 1.1607492530019954, 'learning_rate/qwen_vl_interface': 4.633523355426629e-06, 'learning_rate/action_model': 4.6335233554266295e-05, 'epoch': 1.29}) +[2026-06-19 16:03:10,306][starVLA.training.train_starvla][INFO] - Step 2754, Loss: {'train/loss': 0.2996419668197632, 'train/grad_norm_pre_clip': 22.69639778137207, 'timing/data': 0.010211882996372879, 'timing/model': 1.147318791016005, 'learning_rate/qwen_vl_interface': 4.6305038184013675e-06, 'learning_rate/action_model': 4.6305038184013675e-05, 'epoch': 1.3}) +[2026-06-19 16:03:14,920][starVLA.training.train_starvla][INFO] - Step 2755, Loss: {'train/loss': 0.20190724730491638, 'train/grad_norm_pre_clip': 8.727228164672852, 'timing/data': 0.01111486996524036, 'timing/model': 1.1565858980175108, 'learning_rate/qwen_vl_interface': 4.627484536027589e-06, 'learning_rate/action_model': 4.627484536027588e-05, 'epoch': 1.3}) +[2026-06-19 16:03:19,525][starVLA.training.train_starvla][INFO] - Step 2756, Loss: {'train/loss': 0.1878664642572403, 'train/grad_norm_pre_clip': 11.86650562286377, 'timing/data': 0.010309655917808414, 'timing/model': 1.307630047085695, 'learning_rate/qwen_vl_interface': 4.624465509546399e-06, 'learning_rate/action_model': 4.6244655095463986e-05, 'epoch': 1.3}) +[2026-06-19 16:03:24,017][starVLA.training.train_starvla][INFO] - Step 2757, Loss: {'train/loss': 0.2849692106246948, 'train/grad_norm_pre_clip': 12.69094181060791, 'timing/data': 0.009373732027597725, 'timing/model': 1.2082215159898624, 'learning_rate/qwen_vl_interface': 4.6214467401988106e-06, 'learning_rate/action_model': 4.621446740198811e-05, 'epoch': 1.3}) +[2026-06-19 16:03:28,455][starVLA.training.train_starvla][INFO] - Step 2758, Loss: {'train/loss': 0.29918575286865234, 'train/grad_norm_pre_clip': 20.936864852905273, 'timing/data': 0.009394095977768302, 'timing/model': 1.1475930919405073, 'learning_rate/qwen_vl_interface': 4.6184282292257245e-06, 'learning_rate/action_model': 4.618428229225725e-05, 'epoch': 1.3}) +[2026-06-19 16:03:33,016][starVLA.training.train_starvla][INFO] - Step 2759, Loss: {'train/loss': 0.3440447747707367, 'train/grad_norm_pre_clip': 13.620983123779297, 'timing/data': 0.010283982031978667, 'timing/model': 1.1579479919746518, 'learning_rate/qwen_vl_interface': 4.615409977867938e-06, 'learning_rate/action_model': 4.615409977867937e-05, 'epoch': 1.3}) +[2026-06-19 16:03:37,631][starVLA.training.train_starvla][INFO] - Step 2760, Loss: {'train/loss': 0.3749081790447235, 'train/grad_norm_pre_clip': 10.804889678955078, 'timing/data': 0.010641228058375418, 'timing/model': 1.1477636019699275, 'learning_rate/qwen_vl_interface': 4.612391987366138e-06, 'learning_rate/action_model': 4.612391987366138e-05, 'epoch': 1.3}) +[2026-06-19 16:03:42,227][starVLA.training.train_starvla][INFO] - Step 2761, Loss: {'train/loss': 0.3008222281932831, 'train/grad_norm_pre_clip': 10.55845832824707, 'timing/data': 0.009583571925759315, 'timing/model': 1.1461642099311575, 'learning_rate/qwen_vl_interface': 4.609374258960908e-06, 'learning_rate/action_model': 4.6093742589609074e-05, 'epoch': 1.3}) +[2026-06-19 16:03:46,837][starVLA.training.train_starvla][INFO] - Step 2762, Loss: {'train/loss': 0.1321302205324173, 'train/grad_norm_pre_clip': 4.831158638000488, 'timing/data': 0.010328809032216668, 'timing/model': 1.3181576250353828, 'learning_rate/qwen_vl_interface': 4.60635679389272e-06, 'learning_rate/action_model': 4.60635679389272e-05, 'epoch': 1.3}) +[2026-06-19 16:03:51,315][starVLA.training.train_starvla][INFO] - Step 2763, Loss: {'train/loss': 0.1710050106048584, 'train/grad_norm_pre_clip': 5.137781143188477, 'timing/data': 0.010488095926120877, 'timing/model': 1.164264564984478, 'learning_rate/qwen_vl_interface': 4.603339593401943e-06, 'learning_rate/action_model': 4.603339593401942e-05, 'epoch': 1.3}) +[2026-06-19 16:03:55,920][starVLA.training.train_starvla][INFO] - Step 2764, Loss: {'train/loss': 0.28642600774765015, 'train/grad_norm_pre_clip': 20.632972717285156, 'timing/data': 0.010585211915895343, 'timing/model': 1.1528141979360953, 'learning_rate/qwen_vl_interface': 4.600322658728832e-06, 'learning_rate/action_model': 4.600322658728832e-05, 'epoch': 1.3}) +[2026-06-19 16:04:00,565][starVLA.training.train_starvla][INFO] - Step 2765, Loss: {'train/loss': 0.15544354915618896, 'train/grad_norm_pre_clip': 8.065027236938477, 'timing/data': 0.010842706891708076, 'timing/model': 1.1616027429699898, 'learning_rate/qwen_vl_interface': 4.597305991113537e-06, 'learning_rate/action_model': 4.597305991113537e-05, 'epoch': 1.3}) +[2026-06-19 16:04:05,107][starVLA.training.train_starvla][INFO] - Step 2766, Loss: {'train/loss': 0.4215363562107086, 'train/grad_norm_pre_clip': 10.183780670166016, 'timing/data': 0.010334842023439705, 'timing/model': 1.1497711949050426, 'learning_rate/qwen_vl_interface': 4.5942895917960935e-06, 'learning_rate/action_model': 4.5942895917960935e-05, 'epoch': 1.3}) +[2026-06-19 16:04:09,714][starVLA.training.train_starvla][INFO] - Step 2767, Loss: {'train/loss': 0.4030623137950897, 'train/grad_norm_pre_clip': 10.636068344116211, 'timing/data': 0.012521074968390167, 'timing/model': 1.152359080966562, 'learning_rate/qwen_vl_interface': 4.59127346201643e-06, 'learning_rate/action_model': 4.59127346201643e-05, 'epoch': 1.3}) +[2026-06-19 16:04:14,335][starVLA.training.train_starvla][INFO] - Step 2768, Loss: {'train/loss': 0.1279802918434143, 'train/grad_norm_pre_clip': 5.156951904296875, 'timing/data': 0.01091537403408438, 'timing/model': 1.326128341956064, 'learning_rate/qwen_vl_interface': 4.588257603014364e-06, 'learning_rate/action_model': 4.588257603014364e-05, 'epoch': 1.3}) +[2026-06-19 16:04:18,772][starVLA.training.train_starvla][INFO] - Step 2769, Loss: {'train/loss': 0.18788324296474457, 'train/grad_norm_pre_clip': 6.922463893890381, 'timing/data': 0.010447723907418549, 'timing/model': 1.1509716770378873, 'learning_rate/qwen_vl_interface': 4.585242016029602e-06, 'learning_rate/action_model': 4.5852420160296015e-05, 'epoch': 1.3}) +[2026-06-19 16:04:23,370][starVLA.training.train_starvla][INFO] - Step 2770, Loss: {'train/loss': 0.36090147495269775, 'train/grad_norm_pre_clip': 8.228497505187988, 'timing/data': 0.009626828017644584, 'timing/model': 1.1548734310781583, 'learning_rate/qwen_vl_interface': 4.582226702301737e-06, 'learning_rate/action_model': 4.5822267023017375e-05, 'epoch': 1.3}) +[2026-06-19 16:04:27,977][starVLA.training.train_starvla][INFO] - Step 2771, Loss: {'train/loss': 0.2773657739162445, 'train/grad_norm_pre_clip': 7.881589889526367, 'timing/data': 0.010006134980358183, 'timing/model': 1.146623892011121, 'learning_rate/qwen_vl_interface': 4.57921166307025e-06, 'learning_rate/action_model': 4.57921166307025e-05, 'epoch': 1.3}) +[2026-06-19 16:04:32,605][starVLA.training.train_starvla][INFO] - Step 2772, Loss: {'train/loss': 0.1853942722082138, 'train/grad_norm_pre_clip': 6.88057804107666, 'timing/data': 0.010452906019054353, 'timing/model': 1.157864102977328, 'learning_rate/qwen_vl_interface': 4.576196899574511e-06, 'learning_rate/action_model': 4.576196899574511e-05, 'epoch': 1.3}) +[2026-06-19 16:04:37,216][starVLA.training.train_starvla][INFO] - Step 2773, Loss: {'train/loss': 0.3124784827232361, 'train/grad_norm_pre_clip': 8.50256061553955, 'timing/data': 0.009871200076304376, 'timing/model': 1.1479087559273466, 'learning_rate/qwen_vl_interface': 4.573182413053775e-06, 'learning_rate/action_model': 4.5731824130537746e-05, 'epoch': 1.3}) +[2026-06-19 16:04:41,899][starVLA.training.train_starvla][INFO] - Step 2774, Loss: {'train/loss': 0.20242387056350708, 'train/grad_norm_pre_clip': 5.846654415130615, 'timing/data': 0.010369427036494017, 'timing/model': 1.3099957059603184, 'learning_rate/qwen_vl_interface': 4.570168204747183e-06, 'learning_rate/action_model': 4.570168204747183e-05, 'epoch': 1.3}) +[2026-06-19 16:04:46,483][starVLA.training.train_starvla][INFO] - Step 2775, Loss: {'train/loss': 0.11411519348621368, 'train/grad_norm_pre_clip': 4.09797477722168, 'timing/data': 0.009741490008309484, 'timing/model': 1.3011254509910941, 'learning_rate/qwen_vl_interface': 4.567154275893762e-06, 'learning_rate/action_model': 4.5671542758937624e-05, 'epoch': 1.31}) +[2026-06-19 16:04:50,929][starVLA.training.train_starvla][INFO] - Step 2776, Loss: {'train/loss': 0.30355116724967957, 'train/grad_norm_pre_clip': 8.018792152404785, 'timing/data': 0.009749001008458436, 'timing/model': 1.162251697969623, 'learning_rate/qwen_vl_interface': 4.564140627732426e-06, 'learning_rate/action_model': 4.564140627732426e-05, 'epoch': 1.31}) +[2026-06-19 16:04:55,533][starVLA.training.train_starvla][INFO] - Step 2777, Loss: {'train/loss': 0.5009494423866272, 'train/grad_norm_pre_clip': 13.344408988952637, 'timing/data': 0.01006162993144244, 'timing/model': 1.1498719869414344, 'learning_rate/qwen_vl_interface': 4.56112726150197e-06, 'learning_rate/action_model': 4.56112726150197e-05, 'epoch': 1.31}) +[2026-06-19 16:05:00,175][starVLA.training.train_starvla][INFO] - Step 2778, Loss: {'train/loss': 0.08821246027946472, 'train/grad_norm_pre_clip': 4.491791725158691, 'timing/data': 0.01045825402252376, 'timing/model': 1.1580498779658228, 'learning_rate/qwen_vl_interface': 4.558114178441076e-06, 'learning_rate/action_model': 4.558114178441076e-05, 'epoch': 1.31}) +[2026-06-19 16:05:04,781][starVLA.training.train_starvla][INFO] - Step 2779, Loss: {'train/loss': 0.20347370207309723, 'train/grad_norm_pre_clip': 6.049632549285889, 'timing/data': 0.010172988986596465, 'timing/model': 1.148683708976023, 'learning_rate/qwen_vl_interface': 4.555101379788308e-06, 'learning_rate/action_model': 4.555101379788308e-05, 'epoch': 1.31}) +[2026-06-19 16:05:09,389][starVLA.training.train_starvla][INFO] - Step 2780, Loss: {'train/loss': 0.3072282373905182, 'train/grad_norm_pre_clip': 7.961785316467285, 'timing/data': 0.010454428964294493, 'timing/model': 1.1436445079743862, 'learning_rate/qwen_vl_interface': 4.552088866782115e-06, 'learning_rate/action_model': 4.552088866782115e-05, 'epoch': 1.31}) +[2026-06-19 16:05:14,024][starVLA.training.train_starvla][INFO] - Step 2781, Loss: {'train/loss': 0.26057934761047363, 'train/grad_norm_pre_clip': 5.289051532745361, 'timing/data': 0.010696572018787265, 'timing/model': 1.3423986649140716, 'learning_rate/qwen_vl_interface': 4.5490766406608265e-06, 'learning_rate/action_model': 4.549076640660827e-05, 'epoch': 1.31}) +[2026-06-19 16:05:18,491][starVLA.training.train_starvla][INFO] - Step 2782, Loss: {'train/loss': 0.2460709661245346, 'train/grad_norm_pre_clip': 11.821372985839844, 'timing/data': 0.010211930028162897, 'timing/model': 1.1596853649243712, 'learning_rate/qwen_vl_interface': 4.546064702662655e-06, 'learning_rate/action_model': 4.546064702662655e-05, 'epoch': 1.31}) +[2026-06-19 16:05:23,096][starVLA.training.train_starvla][INFO] - Step 2783, Loss: {'train/loss': 0.22088000178337097, 'train/grad_norm_pre_clip': 10.76453685760498, 'timing/data': 0.010438670986332, 'timing/model': 1.1463339329930022, 'learning_rate/qwen_vl_interface': 4.543053054025693e-06, 'learning_rate/action_model': 4.5430530540256926e-05, 'epoch': 1.31}) +[2026-06-19 16:05:27,724][starVLA.training.train_starvla][INFO] - Step 2784, Loss: {'train/loss': 0.21957333385944366, 'train/grad_norm_pre_clip': 7.785039901733398, 'timing/data': 0.012644517002627254, 'timing/model': 1.1622704080073163, 'learning_rate/qwen_vl_interface': 4.540041695987917e-06, 'learning_rate/action_model': 4.540041695987917e-05, 'epoch': 1.31}) +[2026-06-19 16:05:32,236][starVLA.training.train_starvla][INFO] - Step 2785, Loss: {'train/loss': 0.07673550397157669, 'train/grad_norm_pre_clip': 5.836897850036621, 'timing/data': 0.009842072962783277, 'timing/model': 1.1489609510172158, 'learning_rate/qwen_vl_interface': 4.537030629787182e-06, 'learning_rate/action_model': 4.5370306297871826e-05, 'epoch': 1.31}) +[2026-06-19 16:05:36,777][starVLA.training.train_starvla][INFO] - Step 2786, Loss: {'train/loss': 0.14882461726665497, 'train/grad_norm_pre_clip': 8.259922981262207, 'timing/data': 0.010019671055488288, 'timing/model': 1.147716982057318, 'learning_rate/qwen_vl_interface': 4.534019856661226e-06, 'learning_rate/action_model': 4.534019856661226e-05, 'epoch': 1.31}) +[2026-06-19 16:05:41,375][starVLA.training.train_starvla][INFO] - Step 2787, Loss: {'train/loss': 0.314255952835083, 'train/grad_norm_pre_clip': 9.756731033325195, 'timing/data': 0.010217925067991018, 'timing/model': 1.3129075590986758, 'learning_rate/qwen_vl_interface': 4.531009377847661e-06, 'learning_rate/action_model': 4.531009377847661e-05, 'epoch': 1.31}) +[2026-06-19 16:05:45,809][starVLA.training.train_starvla][INFO] - Step 2788, Loss: {'train/loss': 0.1743392050266266, 'train/grad_norm_pre_clip': 8.8035249710083, 'timing/data': 0.01077976799570024, 'timing/model': 1.145968712051399, 'learning_rate/qwen_vl_interface': 4.527999194583982e-06, 'learning_rate/action_model': 4.527999194583981e-05, 'epoch': 1.31}) +[2026-06-19 16:05:50,397][starVLA.training.train_starvla][INFO] - Step 2789, Loss: {'train/loss': 0.4017437994480133, 'train/grad_norm_pre_clip': 16.19451332092285, 'timing/data': 0.010463037062436342, 'timing/model': 1.1526595039758831, 'learning_rate/qwen_vl_interface': 4.52498930810756e-06, 'learning_rate/action_model': 4.52498930810756e-05, 'epoch': 1.31}) +[2026-06-19 16:05:54,993][starVLA.training.train_starvla][INFO] - Step 2790, Loss: {'train/loss': 0.2779153883457184, 'train/grad_norm_pre_clip': 12.686749458312988, 'timing/data': 0.010519488016143441, 'timing/model': 1.1474759730044752, 'learning_rate/qwen_vl_interface': 4.521979719655648e-06, 'learning_rate/action_model': 4.5219797196556484e-05, 'epoch': 1.31}) +[2026-06-19 16:05:59,612][starVLA.training.train_starvla][INFO] - Step 2791, Loss: {'train/loss': 0.21743406355381012, 'train/grad_norm_pre_clip': 6.218560218811035, 'timing/data': 0.01075661601498723, 'timing/model': 1.1527652110671625, 'learning_rate/qwen_vl_interface': 4.518970430465377e-06, 'learning_rate/action_model': 4.518970430465376e-05, 'epoch': 1.31}) +[2026-06-19 16:06:04,231][starVLA.training.train_starvla][INFO] - Step 2792, Loss: {'train/loss': 0.2434031069278717, 'train/grad_norm_pre_clip': 10.037742614746094, 'timing/data': 0.010193245951086283, 'timing/model': 1.3204954449320212, 'learning_rate/qwen_vl_interface': 4.515961441773745e-06, 'learning_rate/action_model': 4.515961441773745e-05, 'epoch': 1.31}) +[2026-06-19 16:06:08,846][starVLA.training.train_starvla][INFO] - Step 2793, Loss: {'train/loss': 0.16405075788497925, 'train/grad_norm_pre_clip': 7.496862411499023, 'timing/data': 0.010175456991419196, 'timing/model': 1.3257183129899204, 'learning_rate/qwen_vl_interface': 4.5129527548176385e-06, 'learning_rate/action_model': 4.512952754817638e-05, 'epoch': 1.31}) +[2026-06-19 16:06:13,273][starVLA.training.train_starvla][INFO] - Step 2794, Loss: {'train/loss': 0.2732132375240326, 'train/grad_norm_pre_clip': 8.713143348693848, 'timing/data': 0.009701174101792276, 'timing/model': 1.1478547559818253, 'learning_rate/qwen_vl_interface': 4.509944370833813e-06, 'learning_rate/action_model': 4.509944370833813e-05, 'epoch': 1.31}) +[2026-06-19 16:06:17,867][starVLA.training.train_starvla][INFO] - Step 2795, Loss: {'train/loss': 0.4416354298591614, 'train/grad_norm_pre_clip': 12.57369613647461, 'timing/data': 0.011676020920276642, 'timing/model': 1.1670591169968247, 'learning_rate/qwen_vl_interface': 4.506936291058902e-06, 'learning_rate/action_model': 4.506936291058902e-05, 'epoch': 1.31}) +[2026-06-19 16:06:22,485][starVLA.training.train_starvla][INFO] - Step 2796, Loss: {'train/loss': 0.20679201185703278, 'train/grad_norm_pre_clip': 15.518674850463867, 'timing/data': 0.009171767043881118, 'timing/model': 1.1553353209747002, 'learning_rate/qwen_vl_interface': 4.503928516729416e-06, 'learning_rate/action_model': 4.503928516729416e-05, 'epoch': 1.32}) +[2026-06-19 16:06:27,075][starVLA.training.train_starvla][INFO] - Step 2797, Loss: {'train/loss': 0.2506030797958374, 'train/grad_norm_pre_clip': 14.157660484313965, 'timing/data': 0.010615978040732443, 'timing/model': 1.1475569510366768, 'learning_rate/qwen_vl_interface': 4.500921049081734e-06, 'learning_rate/action_model': 4.500921049081734e-05, 'epoch': 1.32}) +[2026-06-19 16:06:31,686][starVLA.training.train_starvla][INFO] - Step 2798, Loss: {'train/loss': 0.1725926250219345, 'train/grad_norm_pre_clip': 10.017999649047852, 'timing/data': 0.010480296099558473, 'timing/model': 1.3177394809899852, 'learning_rate/qwen_vl_interface': 4.497913889352113e-06, 'learning_rate/action_model': 4.4979138893521126e-05, 'epoch': 1.32}) +[2026-06-19 16:06:36,288][starVLA.training.train_starvla][INFO] - Step 2799, Loss: {'train/loss': 0.1490088254213333, 'train/grad_norm_pre_clip': 9.728569030761719, 'timing/data': 0.010483112069778144, 'timing/model': 1.3063157779397443, 'learning_rate/qwen_vl_interface': 4.494907038776684e-06, 'learning_rate/action_model': 4.4949070387766836e-05, 'epoch': 1.32}) +[2026-06-19 16:06:40,740][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 2800 +[2026-06-19 16:09:34,361][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 2800 +[2026-06-19 16:09:34,362][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 2800 +[2026-06-19 16:12:44,397][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 2800 +[2026-06-19 16:12:44,398][starVLA.training.train_starvla][INFO] - Step 2800, Loss: {'train/loss': 0.4570178985595703, 'train/grad_norm_pre_clip': 18.583820343017578, 'eval/loss': 0.20769645273685455, 'eval/latency_0/loss': 0.09375602751970291, 'eval/latency_1/loss': 0.08776798099279404, 'eval/latency_2/loss': 0.14170067012310028, 'eval/latency_3/loss': 0.21958290040493011, 'eval/latency_4/loss': 0.49400076270103455, 'eval/flappy/loss': 0.20769645273685455, 'eval/flappy/latency_0/loss': 0.09375602751970291, 'eval/flappy/latency_1/loss': 0.08776798099279404, 'eval/flappy/latency_2/loss': 0.14170067012310028, 'eval/flappy/latency_3/loss': 0.21958290040493011, 'eval/flappy/latency_4/loss': 0.49400076270103455, 'eval/flap_precision': 0.7100371747211895, 'eval/flap_recall': 0.6945454545454546, 'eval/flap_f1': 0.7022058823529411, 'eval/flap_rate_model': 0.07472222222222222, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.7022058823529411, 'eval/latency_0/flap_precision': 0.7100371747211895, 'eval/latency_0/flap_recall': 0.6945454545454546, 'eval/latency_0/flap_f1': 0.7022058823529411, 'eval/latency_0/flap_rate_model': 0.07472222222222222, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.7022058823529411, 'timing/data': 0.01035699201747775, 'timing/model': 1.1550787889864296, 'learning_rate/qwen_vl_interface': 4.49190049859145e-06, 'learning_rate/action_model': 4.49190049859145e-05, 'epoch': 1.32}) +[2026-06-19 16:12:44,399][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_2800_state +[2026-06-19 16:12:44,402][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-19 16:12:47,794][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_2800_state/model.safetensors +[2026-06-19 16:12:55,033][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_2800_state/optimizer.bin +[2026-06-19 16:12:55,034][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_2800_state/sampler.bin +[2026-06-19 16:12:55,034][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_2800_state/sampler_1.bin +[2026-06-19 16:12:55,035][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_2800_state/random_states_0.pkl +[2026-06-19 16:12:55,058][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2800 +[2026-06-19 16:12:55,058][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-19 16:12:55,087][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-19 16:12:59,618][starVLA.training.train_starvla][INFO] - Step 2801, Loss: {'train/loss': 0.6343623995780945, 'train/grad_norm_pre_clip': 16.664779663085938, 'timing/data': 0.011747266980819404, 'timing/model': 1.1428756100358441, 'learning_rate/qwen_vl_interface': 4.488894270032284e-06, 'learning_rate/action_model': 4.4888942700322834e-05, 'epoch': 1.32}) +[2026-06-19 16:13:04,145][starVLA.training.train_starvla][INFO] - Step 2802, Loss: {'train/loss': 0.41887935996055603, 'train/grad_norm_pre_clip': 15.315134048461914, 'timing/data': 0.009051847970113158, 'timing/model': 1.2904277180787176, 'learning_rate/qwen_vl_interface': 4.485888354334933e-06, 'learning_rate/action_model': 4.4858883543349334e-05, 'epoch': 1.32}) +[2026-06-19 16:13:08,510][starVLA.training.train_starvla][INFO] - Step 2803, Loss: {'train/loss': 0.2036537528038025, 'train/grad_norm_pre_clip': 8.552562713623047, 'timing/data': 0.009681974072009325, 'timing/model': 1.1307496790541336, 'learning_rate/qwen_vl_interface': 4.482882752735017e-06, 'learning_rate/action_model': 4.482882752735017e-05, 'epoch': 1.32}) +[2026-06-19 16:13:13,061][starVLA.training.train_starvla][INFO] - Step 2804, Loss: {'train/loss': 0.28694337606430054, 'train/grad_norm_pre_clip': 6.140147686004639, 'timing/data': 0.01007720606867224, 'timing/model': 1.1386313519906253, 'learning_rate/qwen_vl_interface': 4.479877466468027e-06, 'learning_rate/action_model': 4.479877466468026e-05, 'epoch': 1.32}) +[2026-06-19 16:13:17,619][starVLA.training.train_starvla][INFO] - Step 2805, Loss: {'train/loss': 0.11586792021989822, 'train/grad_norm_pre_clip': 3.8872900009155273, 'timing/data': 0.009689804981462657, 'timing/model': 1.1367228070739657, 'learning_rate/qwen_vl_interface': 4.476872496769318e-06, 'learning_rate/action_model': 4.476872496769318e-05, 'epoch': 1.32}) +[2026-06-19 16:13:22,185][starVLA.training.train_starvla][INFO] - Step 2806, Loss: {'train/loss': 0.3245336711406708, 'train/grad_norm_pre_clip': 15.321578025817871, 'timing/data': 0.009955395944416523, 'timing/model': 1.1352241530548781, 'learning_rate/qwen_vl_interface': 4.473867844874122e-06, 'learning_rate/action_model': 4.473867844874122e-05, 'epoch': 1.32}) +[2026-06-19 16:13:26,721][starVLA.training.train_starvla][INFO] - Step 2807, Loss: {'train/loss': 0.38135573267936707, 'train/grad_norm_pre_clip': 18.240758895874023, 'timing/data': 0.009746480965986848, 'timing/model': 1.1318002719199285, 'learning_rate/qwen_vl_interface': 4.470863512017539e-06, 'learning_rate/action_model': 4.470863512017539e-05, 'epoch': 1.32}) +[2026-06-19 16:13:31,283][starVLA.training.train_starvla][INFO] - Step 2808, Loss: {'train/loss': 0.2801090180873871, 'train/grad_norm_pre_clip': 15.047809600830078, 'timing/data': 0.009442902053706348, 'timing/model': 1.1425792149966583, 'learning_rate/qwen_vl_interface': 4.4678594994345344e-06, 'learning_rate/action_model': 4.467859499434534e-05, 'epoch': 1.32}) +[2026-06-19 16:13:35,850][starVLA.training.train_starvla][INFO] - Step 2809, Loss: {'train/loss': 0.17783387005329132, 'train/grad_norm_pre_clip': 7.9819135665893555, 'timing/data': 0.009954286040738225, 'timing/model': 1.3014681210042909, 'learning_rate/qwen_vl_interface': 4.4648558083599475e-06, 'learning_rate/action_model': 4.464855808359947e-05, 'epoch': 1.32}) +[2026-06-19 16:13:40,249][starVLA.training.train_starvla][INFO] - Step 2810, Loss: {'train/loss': 0.13821348547935486, 'train/grad_norm_pre_clip': 5.835389137268066, 'timing/data': 0.0087384560611099, 'timing/model': 1.1456349330255762, 'learning_rate/qwen_vl_interface': 4.461852440028479e-06, 'learning_rate/action_model': 4.461852440028479e-05, 'epoch': 1.32}) +[2026-06-19 16:13:44,805][starVLA.training.train_starvla][INFO] - Step 2811, Loss: {'train/loss': 0.1930888444185257, 'train/grad_norm_pre_clip': 10.140275955200195, 'timing/data': 0.009503046981990337, 'timing/model': 1.1408187409397215, 'learning_rate/qwen_vl_interface': 4.458849395674702e-06, 'learning_rate/action_model': 4.458849395674701e-05, 'epoch': 1.32}) +[2026-06-19 16:13:49,352][starVLA.training.train_starvla][INFO] - Step 2812, Loss: {'train/loss': 0.2897584140300751, 'train/grad_norm_pre_clip': 8.638938903808594, 'timing/data': 0.008963460102677345, 'timing/model': 1.1358608540613204, 'learning_rate/qwen_vl_interface': 4.455846676533052e-06, 'learning_rate/action_model': 4.4558466765330525e-05, 'epoch': 1.32}) +[2026-06-19 16:13:53,934][starVLA.training.train_starvla][INFO] - Step 2813, Loss: {'train/loss': 0.11005418002605438, 'train/grad_norm_pre_clip': 4.565135478973389, 'timing/data': 0.009504994959570467, 'timing/model': 1.1453307019546628, 'learning_rate/qwen_vl_interface': 4.452844283837838e-06, 'learning_rate/action_model': 4.452844283837838e-05, 'epoch': 1.32}) +[2026-06-19 16:13:58,520][starVLA.training.train_starvla][INFO] - Step 2814, Loss: {'train/loss': 0.38484787940979004, 'train/grad_norm_pre_clip': 13.980182647705078, 'timing/data': 0.009589978028088808, 'timing/model': 1.1425879980670288, 'learning_rate/qwen_vl_interface': 4.449842218823228e-06, 'learning_rate/action_model': 4.449842218823228e-05, 'epoch': 1.32}) +[2026-06-19 16:14:03,093][starVLA.training.train_starvla][INFO] - Step 2815, Loss: {'train/loss': 0.31703558564186096, 'train/grad_norm_pre_clip': 9.900687217712402, 'timing/data': 0.009994036983698606, 'timing/model': 1.29756268998608, 'learning_rate/qwen_vl_interface': 4.446840482723256e-06, 'learning_rate/action_model': 4.446840482723255e-05, 'epoch': 1.32}) +[2026-06-19 16:14:07,644][starVLA.training.train_starvla][INFO] - Step 2816, Loss: {'train/loss': 0.2406189739704132, 'train/grad_norm_pre_clip': 17.01124382019043, 'timing/data': 0.009968391037546098, 'timing/model': 1.277597252978012, 'learning_rate/qwen_vl_interface': 4.443839076771823e-06, 'learning_rate/action_model': 4.443839076771823e-05, 'epoch': 1.32}) +[2026-06-19 16:14:12,070][starVLA.training.train_starvla][INFO] - Step 2817, Loss: {'train/loss': 0.41332074999809265, 'train/grad_norm_pre_clip': 15.137259483337402, 'timing/data': 0.010100955027155578, 'timing/model': 1.1545649420004338, 'learning_rate/qwen_vl_interface': 4.4408380022026945e-06, 'learning_rate/action_model': 4.440838002202694e-05, 'epoch': 1.32}) +[2026-06-19 16:14:16,682][starVLA.training.train_starvla][INFO] - Step 2818, Loss: {'train/loss': 0.2026233673095703, 'train/grad_norm_pre_clip': 7.714494705200195, 'timing/data': 0.009606697945855558, 'timing/model': 1.1544380609411746, 'learning_rate/qwen_vl_interface': 4.437837260249499e-06, 'learning_rate/action_model': 4.437837260249499e-05, 'epoch': 1.33}) +[2026-06-19 16:14:21,290][starVLA.training.train_starvla][INFO] - Step 2819, Loss: {'train/loss': 0.4390627145767212, 'train/grad_norm_pre_clip': 11.930302619934082, 'timing/data': 0.010069004027172923, 'timing/model': 1.1493074930040166, 'learning_rate/qwen_vl_interface': 4.434836852145728e-06, 'learning_rate/action_model': 4.434836852145728e-05, 'epoch': 1.33}) +[2026-06-19 16:14:25,887][starVLA.training.train_starvla][INFO] - Step 2820, Loss: {'train/loss': 0.32465288043022156, 'train/grad_norm_pre_clip': 7.959722518920898, 'timing/data': 0.008976554963737726, 'timing/model': 1.1499868329847232, 'learning_rate/qwen_vl_interface': 4.431836779124736e-06, 'learning_rate/action_model': 4.431836779124735e-05, 'epoch': 1.33}) +[2026-06-19 16:14:30,467][starVLA.training.train_starvla][INFO] - Step 2821, Loss: {'train/loss': 0.22344490885734558, 'train/grad_norm_pre_clip': 10.01488971710205, 'timing/data': 0.01008013600949198, 'timing/model': 1.3056043960386887, 'learning_rate/qwen_vl_interface': 4.428837042419738e-06, 'learning_rate/action_model': 4.428837042419738e-05, 'epoch': 1.33}) +[2026-06-19 16:14:34,906][starVLA.training.train_starvla][INFO] - Step 2822, Loss: {'train/loss': 0.2894545793533325, 'train/grad_norm_pre_clip': 12.407144546508789, 'timing/data': 0.011191757977940142, 'timing/model': 1.1561526850564405, 'learning_rate/qwen_vl_interface': 4.425837643263815e-06, 'learning_rate/action_model': 4.425837643263814e-05, 'epoch': 1.33}) +[2026-06-19 16:14:39,539][starVLA.training.train_starvla][INFO] - Step 2823, Loss: {'train/loss': 0.3495243489742279, 'train/grad_norm_pre_clip': 10.280380249023438, 'timing/data': 0.00999927707016468, 'timing/model': 1.1631228929618374, 'learning_rate/qwen_vl_interface': 4.4228385828899045e-06, 'learning_rate/action_model': 4.4228385828899046e-05, 'epoch': 1.33}) +[2026-06-19 16:14:44,138][starVLA.training.train_starvla][INFO] - Step 2824, Loss: {'train/loss': 0.2922578454017639, 'train/grad_norm_pre_clip': 6.927113056182861, 'timing/data': 0.010260784067213535, 'timing/model': 1.1547767899464816, 'learning_rate/qwen_vl_interface': 4.419839862530809e-06, 'learning_rate/action_model': 4.419839862530809e-05, 'epoch': 1.33}) +[2026-06-19 16:14:48,772][starVLA.training.train_starvla][INFO] - Step 2825, Loss: {'train/loss': 0.1753225326538086, 'train/grad_norm_pre_clip': 7.66056489944458, 'timing/data': 0.009941592928953469, 'timing/model': 1.1597442060010508, 'learning_rate/qwen_vl_interface': 4.416841483419188e-06, 'learning_rate/action_model': 4.416841483419188e-05, 'epoch': 1.33}) +[2026-06-19 16:14:53,369][starVLA.training.train_starvla][INFO] - Step 2826, Loss: {'train/loss': 0.3099794089794159, 'train/grad_norm_pre_clip': 7.3405303955078125, 'timing/data': 0.009977984009310603, 'timing/model': 1.144438024959527, 'learning_rate/qwen_vl_interface': 4.413843446787561e-06, 'learning_rate/action_model': 4.413843446787561e-05, 'epoch': 1.33}) +[2026-06-19 16:14:57,968][starVLA.training.train_starvla][INFO] - Step 2827, Loss: {'train/loss': 0.08305096626281738, 'train/grad_norm_pre_clip': 2.6906261444091797, 'timing/data': 0.010252541047520936, 'timing/model': 1.3132452180143446, 'learning_rate/qwen_vl_interface': 4.41084575386831e-06, 'learning_rate/action_model': 4.4108457538683093e-05, 'epoch': 1.33}) +[2026-06-19 16:15:02,515][starVLA.training.train_starvla][INFO] - Step 2828, Loss: {'train/loss': 0.1299387663602829, 'train/grad_norm_pre_clip': 4.06573486328125, 'timing/data': 0.01011919102165848, 'timing/model': 1.2694023320218548, 'learning_rate/qwen_vl_interface': 4.40784840589367e-06, 'learning_rate/action_model': 4.4078484058936706e-05, 'epoch': 1.33}) +[2026-06-19 16:15:06,951][starVLA.training.train_starvla][INFO] - Step 2829, Loss: {'train/loss': 0.32770270109176636, 'train/grad_norm_pre_clip': 9.674135208129883, 'timing/data': 0.009068317012861371, 'timing/model': 1.1509859730722383, 'learning_rate/qwen_vl_interface': 4.404851404095742e-06, 'learning_rate/action_model': 4.404851404095742e-05, 'epoch': 1.33}) +[2026-06-19 16:15:11,546][starVLA.training.train_starvla][INFO] - Step 2830, Loss: {'train/loss': 0.32461580634117126, 'train/grad_norm_pre_clip': 9.835988998413086, 'timing/data': 0.009656646056100726, 'timing/model': 1.1477618869394064, 'learning_rate/qwen_vl_interface': 4.4018547497064765e-06, 'learning_rate/action_model': 4.401854749706476e-05, 'epoch': 1.33}) +[2026-06-19 16:15:16,145][starVLA.training.train_starvla][INFO] - Step 2831, Loss: {'train/loss': 0.36965450644493103, 'train/grad_norm_pre_clip': 15.146197319030762, 'timing/data': 0.009536182042211294, 'timing/model': 1.1562007040483877, 'learning_rate/qwen_vl_interface': 4.398858443957686e-06, 'learning_rate/action_model': 4.398858443957686e-05, 'epoch': 1.33}) +[2026-06-19 16:15:20,782][starVLA.training.train_starvla][INFO] - Step 2832, Loss: {'train/loss': 0.24189473688602448, 'train/grad_norm_pre_clip': 5.884593486785889, 'timing/data': 0.009328173939138651, 'timing/model': 1.1502596900099888, 'learning_rate/qwen_vl_interface': 4.395862488081038e-06, 'learning_rate/action_model': 4.395862488081038e-05, 'epoch': 1.33}) +[2026-06-19 16:15:25,390][starVLA.training.train_starvla][INFO] - Step 2833, Loss: {'train/loss': 0.3031337261199951, 'train/grad_norm_pre_clip': 12.482160568237305, 'timing/data': 0.0098622270161286, 'timing/model': 1.3286317789461464, 'learning_rate/qwen_vl_interface': 4.392866883308059e-06, 'learning_rate/action_model': 4.392866883308059e-05, 'epoch': 1.33}) +[2026-06-19 16:15:29,942][starVLA.training.train_starvla][INFO] - Step 2834, Loss: {'train/loss': 0.2615436613559723, 'train/grad_norm_pre_clip': 12.904128074645996, 'timing/data': 0.009802166954614222, 'timing/model': 1.2669246009318158, 'learning_rate/qwen_vl_interface': 4.389871630870127e-06, 'learning_rate/action_model': 4.389871630870127e-05, 'epoch': 1.33}) +[2026-06-19 16:15:34,394][starVLA.training.train_starvla][INFO] - Step 2835, Loss: {'train/loss': 0.28731366991996765, 'train/grad_norm_pre_clip': 9.157083511352539, 'timing/data': 0.009955067071132362, 'timing/model': 1.1596933320397511, 'learning_rate/qwen_vl_interface': 4.386876731998478e-06, 'learning_rate/action_model': 4.386876731998478e-05, 'epoch': 1.33}) +[2026-06-19 16:15:39,020][starVLA.training.train_starvla][INFO] - Step 2836, Loss: {'train/loss': 0.3518114686012268, 'train/grad_norm_pre_clip': 8.730057716369629, 'timing/data': 0.01046891591977328, 'timing/model': 1.1595752369612455, 'learning_rate/qwen_vl_interface': 4.3838821879242006e-06, 'learning_rate/action_model': 4.3838821879242e-05, 'epoch': 1.33}) +[2026-06-19 16:15:43,665][starVLA.training.train_starvla][INFO] - Step 2837, Loss: {'train/loss': 0.39345329999923706, 'train/grad_norm_pre_clip': 11.569853782653809, 'timing/data': 0.009960758965462446, 'timing/model': 1.1601939789252356, 'learning_rate/qwen_vl_interface': 4.38088799987824e-06, 'learning_rate/action_model': 4.38088799987824e-05, 'epoch': 1.33}) +[2026-06-19 16:15:48,253][starVLA.training.train_starvla][INFO] - Step 2838, Loss: {'train/loss': 0.24508167803287506, 'train/grad_norm_pre_clip': 7.123598575592041, 'timing/data': 0.009384877048432827, 'timing/model': 1.1547021300066262, 'learning_rate/qwen_vl_interface': 4.3778941690913935e-06, 'learning_rate/action_model': 4.3778941690913934e-05, 'epoch': 1.33}) +[2026-06-19 16:15:52,883][starVLA.training.train_starvla][INFO] - Step 2839, Loss: {'train/loss': 0.2088695615530014, 'train/grad_norm_pre_clip': 8.343636512756348, 'timing/data': 0.009819217026233673, 'timing/model': 1.3416322789853439, 'learning_rate/qwen_vl_interface': 4.374900696794311e-06, 'learning_rate/action_model': 4.374900696794311e-05, 'epoch': 1.34}) +[2026-06-19 16:15:57,408][starVLA.training.train_starvla][INFO] - Step 2840, Loss: {'train/loss': 0.2607000172138214, 'train/grad_norm_pre_clip': 9.084243774414062, 'timing/data': 0.009787265909835696, 'timing/model': 1.2512415800010785, 'learning_rate/qwen_vl_interface': 4.371907584217499e-06, 'learning_rate/action_model': 4.371907584217498e-05, 'epoch': 1.34}) +[2026-06-19 16:16:01,829][starVLA.training.train_starvla][INFO] - Step 2841, Loss: {'train/loss': 0.12627744674682617, 'train/grad_norm_pre_clip': 15.3600492477417, 'timing/data': 0.009491536999121308, 'timing/model': 1.1476235470036045, 'learning_rate/qwen_vl_interface': 4.368914832591309e-06, 'learning_rate/action_model': 4.368914832591309e-05, 'epoch': 1.34}) +[2026-06-19 16:16:06,428][starVLA.training.train_starvla][INFO] - Step 2842, Loss: {'train/loss': 0.13624845445156097, 'train/grad_norm_pre_clip': 7.54236364364624, 'timing/data': 0.00979419203940779, 'timing/model': 1.1480890340171754, 'learning_rate/qwen_vl_interface': 4.365922443145949e-06, 'learning_rate/action_model': 4.365922443145949e-05, 'epoch': 1.34}) +[2026-06-19 16:16:11,022][starVLA.training.train_starvla][INFO] - Step 2843, Loss: {'train/loss': 0.1876077502965927, 'train/grad_norm_pre_clip': 8.930691719055176, 'timing/data': 0.010741211008280516, 'timing/model': 1.1496741389855742, 'learning_rate/qwen_vl_interface': 4.362930417111479e-06, 'learning_rate/action_model': 4.362930417111479e-05, 'epoch': 1.34}) +[2026-06-19 16:16:15,610][starVLA.training.train_starvla][INFO] - Step 2844, Loss: {'train/loss': 0.15894481539726257, 'train/grad_norm_pre_clip': 6.096362113952637, 'timing/data': 0.009341255063191056, 'timing/model': 1.1468310840427876, 'learning_rate/qwen_vl_interface': 4.359938755717807e-06, 'learning_rate/action_model': 4.359938755717807e-05, 'epoch': 1.34}) +[2026-06-19 16:16:20,207][starVLA.training.train_starvla][INFO] - Step 2845, Loss: {'train/loss': 0.3650864362716675, 'train/grad_norm_pre_clip': 29.02903938293457, 'timing/data': 0.009113405016250908, 'timing/model': 1.1482216109288856, 'learning_rate/qwen_vl_interface': 4.3569474601946934e-06, 'learning_rate/action_model': 4.356947460194694e-05, 'epoch': 1.34}) +[2026-06-19 16:16:24,812][starVLA.training.train_starvla][INFO] - Step 2846, Loss: {'train/loss': 0.24530641734600067, 'train/grad_norm_pre_clip': 8.61522102355957, 'timing/data': 0.008785723010078073, 'timing/model': 1.318419124931097, 'learning_rate/qwen_vl_interface': 4.3539565317717446e-06, 'learning_rate/action_model': 4.353956531771745e-05, 'epoch': 1.34}) +[2026-06-19 16:16:29,244][starVLA.training.train_starvla][INFO] - Step 2847, Loss: {'train/loss': 0.4037202000617981, 'train/grad_norm_pre_clip': 13.775711059570312, 'timing/data': 0.009513037046417594, 'timing/model': 1.1486837270203978, 'learning_rate/qwen_vl_interface': 4.350965971678419e-06, 'learning_rate/action_model': 4.3509659716784185e-05, 'epoch': 1.34}) +[2026-06-19 16:16:33,855][starVLA.training.train_starvla][INFO] - Step 2848, Loss: {'train/loss': 0.17149926722049713, 'train/grad_norm_pre_clip': 8.22615909576416, 'timing/data': 0.009421764989383519, 'timing/model': 1.1469069509766996, 'learning_rate/qwen_vl_interface': 4.3479757811440225e-06, 'learning_rate/action_model': 4.347975781144022e-05, 'epoch': 1.34}) +[2026-06-19 16:16:38,472][starVLA.training.train_starvla][INFO] - Step 2849, Loss: {'train/loss': 0.09742847830057144, 'train/grad_norm_pre_clip': 4.367955207824707, 'timing/data': 0.00882253295276314, 'timing/model': 1.1659639569697902, 'learning_rate/qwen_vl_interface': 4.34498596139771e-06, 'learning_rate/action_model': 4.34498596139771e-05, 'epoch': 1.34}) +[2026-06-19 16:16:43,100][starVLA.training.train_starvla][INFO] - Step 2850, Loss: {'train/loss': 0.15911215543746948, 'train/grad_norm_pre_clip': 10.667093276977539, 'timing/data': 0.013863835018128157, 'timing/model': 1.1584079249296337, 'learning_rate/qwen_vl_interface': 4.341996513668483e-06, 'learning_rate/action_model': 4.3419965136684824e-05, 'epoch': 1.34}) +[2026-06-19 16:16:47,729][starVLA.training.train_starvla][INFO] - Step 2851, Loss: {'train/loss': 0.2780033349990845, 'train/grad_norm_pre_clip': 10.887114524841309, 'timing/data': 0.009281720034778118, 'timing/model': 1.152771076071076, 'learning_rate/qwen_vl_interface': 4.33900743918519e-06, 'learning_rate/action_model': 4.33900743918519e-05, 'epoch': 1.34}) +[2026-06-19 16:16:52,324][starVLA.training.train_starvla][INFO] - Step 2852, Loss: {'train/loss': 0.2225869596004486, 'train/grad_norm_pre_clip': 13.66053295135498, 'timing/data': 0.010061949957162142, 'timing/model': 1.30992739694193, 'learning_rate/qwen_vl_interface': 4.336018739176529e-06, 'learning_rate/action_model': 4.33601873917653e-05, 'epoch': 1.34}) +[2026-06-19 16:16:56,765][starVLA.training.train_starvla][INFO] - Step 2853, Loss: {'train/loss': 0.24932801723480225, 'train/grad_norm_pre_clip': 9.069683074951172, 'timing/data': 0.01020969997625798, 'timing/model': 1.1476905109593645, 'learning_rate/qwen_vl_interface': 4.333030414871041e-06, 'learning_rate/action_model': 4.33303041487104e-05, 'epoch': 1.34}) +[2026-06-19 16:17:01,391][starVLA.training.train_starvla][INFO] - Step 2854, Loss: {'train/loss': 0.23370464146137238, 'train/grad_norm_pre_clip': 13.599265098571777, 'timing/data': 0.009245302993804216, 'timing/model': 1.148659405997023, 'learning_rate/qwen_vl_interface': 4.330042467497111e-06, 'learning_rate/action_model': 4.330042467497111e-05, 'epoch': 1.34}) +[2026-06-19 16:17:05,902][starVLA.training.train_starvla][INFO] - Step 2855, Loss: {'train/loss': 0.28408852219581604, 'train/grad_norm_pre_clip': 7.842045783996582, 'timing/data': 0.010254448046907783, 'timing/model': 1.1480282880365849, 'learning_rate/qwen_vl_interface': 4.327054898282973e-06, 'learning_rate/action_model': 4.3270548982829727e-05, 'epoch': 1.34}) +[2026-06-19 16:17:10,442][starVLA.training.train_starvla][INFO] - Step 2856, Loss: {'train/loss': 0.1798095852136612, 'train/grad_norm_pre_clip': 6.944961071014404, 'timing/data': 0.00986385194119066, 'timing/model': 1.147808005916886, 'learning_rate/qwen_vl_interface': 4.324067708456703e-06, 'learning_rate/action_model': 4.3240677084567036e-05, 'epoch': 1.34}) +[2026-06-19 16:17:15,037][starVLA.training.train_starvla][INFO] - Step 2857, Loss: {'train/loss': 0.33595284819602966, 'train/grad_norm_pre_clip': 10.281168937683105, 'timing/data': 0.009896503994241357, 'timing/model': 1.1460778809851035, 'learning_rate/qwen_vl_interface': 4.321080899246225e-06, 'learning_rate/action_model': 4.321080899246225e-05, 'epoch': 1.34}) +[2026-06-19 16:17:19,675][starVLA.training.train_starvla][INFO] - Step 2858, Loss: {'train/loss': 0.35920435190200806, 'train/grad_norm_pre_clip': 10.314095497131348, 'timing/data': 0.009972017956897616, 'timing/model': 1.334842498996295, 'learning_rate/qwen_vl_interface': 4.318094471879299e-06, 'learning_rate/action_model': 4.3180944718792985e-05, 'epoch': 1.34}) +[2026-06-19 16:17:24,112][starVLA.training.train_starvla][INFO] - Step 2859, Loss: {'train/loss': 0.26274460554122925, 'train/grad_norm_pre_clip': 9.45378303527832, 'timing/data': 0.009676684974692762, 'timing/model': 1.1530234919628128, 'learning_rate/qwen_vl_interface': 4.315108427583535e-06, 'learning_rate/action_model': 4.315108427583535e-05, 'epoch': 1.34}) +[2026-06-19 16:17:28,742][starVLA.training.train_starvla][INFO] - Step 2860, Loss: {'train/loss': 0.2115856409072876, 'train/grad_norm_pre_clip': 11.116814613342285, 'timing/data': 0.009518175967969, 'timing/model': 1.1578901689499617, 'learning_rate/qwen_vl_interface': 4.312122767586381e-06, 'learning_rate/action_model': 4.3121227675863815e-05, 'epoch': 1.35}) +[2026-06-19 16:17:33,382][starVLA.training.train_starvla][INFO] - Step 2861, Loss: {'train/loss': 0.41220587491989136, 'train/grad_norm_pre_clip': 12.637402534484863, 'timing/data': 0.010371339973062277, 'timing/model': 1.1660810200264677, 'learning_rate/qwen_vl_interface': 4.3091374931151324e-06, 'learning_rate/action_model': 4.309137493115132e-05, 'epoch': 1.35}) +[2026-06-19 16:17:37,993][starVLA.training.train_starvla][INFO] - Step 2862, Loss: {'train/loss': 0.24005921185016632, 'train/grad_norm_pre_clip': 7.657349109649658, 'timing/data': 0.009589154971763492, 'timing/model': 1.1493660119595006, 'learning_rate/qwen_vl_interface': 4.306152605396921e-06, 'learning_rate/action_model': 4.306152605396921e-05, 'epoch': 1.35}) +[2026-06-19 16:17:42,575][starVLA.training.train_starvla][INFO] - Step 2863, Loss: {'train/loss': 0.3833201229572296, 'train/grad_norm_pre_clip': 8.247640609741211, 'timing/data': 0.009812055039219558, 'timing/model': 1.1446417609695345, 'learning_rate/qwen_vl_interface': 4.30316810565872e-06, 'learning_rate/action_model': 4.3031681056587206e-05, 'epoch': 1.35}) +[2026-06-19 16:17:47,190][starVLA.training.train_starvla][INFO] - Step 2864, Loss: {'train/loss': 0.23821769654750824, 'train/grad_norm_pre_clip': 6.917759895324707, 'timing/data': 0.00949404900893569, 'timing/model': 1.3158276049653068, 'learning_rate/qwen_vl_interface': 4.300183995127347e-06, 'learning_rate/action_model': 4.300183995127346e-05, 'epoch': 1.35}) +[2026-06-19 16:17:51,628][starVLA.training.train_starvla][INFO] - Step 2865, Loss: {'train/loss': 0.3140576183795929, 'train/grad_norm_pre_clip': 11.882000923156738, 'timing/data': 0.009857025928795338, 'timing/model': 1.1600334399845451, 'learning_rate/qwen_vl_interface': 4.297200275029454e-06, 'learning_rate/action_model': 4.297200275029454e-05, 'epoch': 1.35}) +[2026-06-19 16:17:56,254][starVLA.training.train_starvla][INFO] - Step 2866, Loss: {'train/loss': 0.31156042218208313, 'train/grad_norm_pre_clip': 21.308008193969727, 'timing/data': 0.009986085002310574, 'timing/model': 1.1534225170034915, 'learning_rate/qwen_vl_interface': 4.29421694659154e-06, 'learning_rate/action_model': 4.294216946591539e-05, 'epoch': 1.35}) +[2026-06-19 16:18:00,793][starVLA.training.train_starvla][INFO] - Step 2867, Loss: {'train/loss': 0.2132698893547058, 'train/grad_norm_pre_clip': 5.574558734893799, 'timing/data': 0.00945798703469336, 'timing/model': 1.1488643808988854, 'learning_rate/qwen_vl_interface': 4.2912340110399345e-06, 'learning_rate/action_model': 4.291234011039934e-05, 'epoch': 1.35}) +[2026-06-19 16:18:05,381][starVLA.training.train_starvla][INFO] - Step 2868, Loss: {'train/loss': 0.2647567689418793, 'train/grad_norm_pre_clip': 17.6802978515625, 'timing/data': 0.008896442945115268, 'timing/model': 1.1610200880095363, 'learning_rate/qwen_vl_interface': 4.2882514696008125e-06, 'learning_rate/action_model': 4.2882514696008125e-05, 'epoch': 1.35}) +[2026-06-19 16:18:10,003][starVLA.training.train_starvla][INFO] - Step 2869, Loss: {'train/loss': 0.23257245123386383, 'train/grad_norm_pre_clip': 8.715730667114258, 'timing/data': 0.009009777917526662, 'timing/model': 1.1683695720275864, 'learning_rate/qwen_vl_interface': 4.2852693235001815e-06, 'learning_rate/action_model': 4.2852693235001815e-05, 'epoch': 1.35}) +[2026-06-19 16:18:14,619][starVLA.training.train_starvla][INFO] - Step 2870, Loss: {'train/loss': 0.2520137429237366, 'train/grad_norm_pre_clip': 6.651761531829834, 'timing/data': 0.009651305968873203, 'timing/model': 1.3352393060922623, 'learning_rate/qwen_vl_interface': 4.2822875739638915e-06, 'learning_rate/action_model': 4.282287573963891e-05, 'epoch': 1.35}) +[2026-06-19 16:18:19,192][starVLA.training.train_starvla][INFO] - Step 2871, Loss: {'train/loss': 0.3444555103778839, 'train/grad_norm_pre_clip': 10.497383117675781, 'timing/data': 0.010013825027272105, 'timing/model': 1.2735160600859672, 'learning_rate/qwen_vl_interface': 4.279306222217625e-06, 'learning_rate/action_model': 4.279306222217626e-05, 'epoch': 1.35}) +[2026-06-19 16:18:23,638][starVLA.training.train_starvla][INFO] - Step 2872, Loss: {'train/loss': 0.28319111466407776, 'train/grad_norm_pre_clip': 8.371487617492676, 'timing/data': 0.00979694293346256, 'timing/model': 1.1588086509145796, 'learning_rate/qwen_vl_interface': 4.276325269486906e-06, 'learning_rate/action_model': 4.276325269486906e-05, 'epoch': 1.35}) +[2026-06-19 16:18:28,236][starVLA.training.train_starvla][INFO] - Step 2873, Loss: {'train/loss': 0.2191818505525589, 'train/grad_norm_pre_clip': 7.5093770027160645, 'timing/data': 0.009480312000960112, 'timing/model': 1.1468357510166243, 'learning_rate/qwen_vl_interface': 4.273344716997089e-06, 'learning_rate/action_model': 4.273344716997089e-05, 'epoch': 1.35}) +[2026-06-19 16:18:32,858][starVLA.training.train_starvla][INFO] - Step 2874, Loss: {'train/loss': 0.31955257058143616, 'train/grad_norm_pre_clip': 6.827648162841797, 'timing/data': 0.009802483953535557, 'timing/model': 1.159052214003168, 'learning_rate/qwen_vl_interface': 4.270364565973368e-06, 'learning_rate/action_model': 4.270364565973367e-05, 'epoch': 1.35}) +[2026-06-19 16:18:37,481][starVLA.training.train_starvla][INFO] - Step 2875, Loss: {'train/loss': 0.1983279436826706, 'train/grad_norm_pre_clip': 7.041502475738525, 'timing/data': 0.008994244039058685, 'timing/model': 1.154069401905872, 'learning_rate/qwen_vl_interface': 4.267384817640768e-06, 'learning_rate/action_model': 4.2673848176407685e-05, 'epoch': 1.35}) +[2026-06-19 16:18:42,030][starVLA.training.train_starvla][INFO] - Step 2876, Loss: {'train/loss': 0.43004110455513, 'train/grad_norm_pre_clip': 12.448336601257324, 'timing/data': 0.009233196964487433, 'timing/model': 1.1622122889384627, 'learning_rate/qwen_vl_interface': 4.264405473224155e-06, 'learning_rate/action_model': 4.2644054732241546e-05, 'epoch': 1.35}) +[2026-06-19 16:18:46,593][starVLA.training.train_starvla][INFO] - Step 2877, Loss: {'train/loss': 0.2123427391052246, 'train/grad_norm_pre_clip': 12.808295249938965, 'timing/data': 0.009800776024349034, 'timing/model': 1.2723038489930332, 'learning_rate/qwen_vl_interface': 4.261426533948221e-06, 'learning_rate/action_model': 4.2614265339482213e-05, 'epoch': 1.35}) +[2026-06-19 16:18:51,055][starVLA.training.train_starvla][INFO] - Step 2878, Loss: {'train/loss': 0.13612407445907593, 'train/grad_norm_pre_clip': 6.220202922821045, 'timing/data': 0.009387194993905723, 'timing/model': 1.1483288239687681, 'learning_rate/qwen_vl_interface': 4.258448001037499e-06, 'learning_rate/action_model': 4.2584480010374985e-05, 'epoch': 1.35}) +[2026-06-19 16:18:55,696][starVLA.training.train_starvla][INFO] - Step 2879, Loss: {'train/loss': 0.2626262903213501, 'train/grad_norm_pre_clip': 8.24143123626709, 'timing/data': 0.008838749956339598, 'timing/model': 1.15605437499471, 'learning_rate/qwen_vl_interface': 4.255469875716348e-06, 'learning_rate/action_model': 4.255469875716348e-05, 'epoch': 1.35}) +[2026-06-19 16:19:00,332][starVLA.training.train_starvla][INFO] - Step 2880, Loss: {'train/loss': 0.16184237599372864, 'train/grad_norm_pre_clip': 6.709039211273193, 'timing/data': 0.009452778031118214, 'timing/model': 1.162768448004499, 'learning_rate/qwen_vl_interface': 4.252492159208966e-06, 'learning_rate/action_model': 4.252492159208966e-05, 'epoch': 1.35}) +[2026-06-19 16:19:04,942][starVLA.training.train_starvla][INFO] - Step 2881, Loss: {'train/loss': 0.32018110156059265, 'train/grad_norm_pre_clip': 10.88703727722168, 'timing/data': 0.009518341044895351, 'timing/model': 1.1482223180355504, 'learning_rate/qwen_vl_interface': 4.2495148527393766e-06, 'learning_rate/action_model': 4.2495148527393766e-05, 'epoch': 1.35}) +[2026-06-19 16:19:09,575][starVLA.training.train_starvla][INFO] - Step 2882, Loss: {'train/loss': 0.21668611466884613, 'train/grad_norm_pre_clip': 6.534966945648193, 'timing/data': 0.010285635013133287, 'timing/model': 1.3359797559678555, 'learning_rate/qwen_vl_interface': 4.246537957531441e-06, 'learning_rate/action_model': 4.2465379575314404e-05, 'epoch': 1.36}) +[2026-06-19 16:19:13,996][starVLA.training.train_starvla][INFO] - Step 2883, Loss: {'train/loss': 0.2104227989912033, 'train/grad_norm_pre_clip': 12.980422019958496, 'timing/data': 0.010653681005351245, 'timing/model': 1.146636085002683, 'learning_rate/qwen_vl_interface': 4.243561474808846e-06, 'learning_rate/action_model': 4.2435614748088456e-05, 'epoch': 1.36}) +[2026-06-19 16:19:18,607][starVLA.training.train_starvla][INFO] - Step 2884, Loss: {'train/loss': 0.28433340787887573, 'train/grad_norm_pre_clip': 8.02110767364502, 'timing/data': 0.009820212959311903, 'timing/model': 1.1562075699912384, 'learning_rate/qwen_vl_interface': 4.240585405795112e-06, 'learning_rate/action_model': 4.240585405795112e-05, 'epoch': 1.36}) +[2026-06-19 16:19:23,199][starVLA.training.train_starvla][INFO] - Step 2885, Loss: {'train/loss': 0.20680436491966248, 'train/grad_norm_pre_clip': 8.733611106872559, 'timing/data': 0.010544232907705009, 'timing/model': 1.15139262995217, 'learning_rate/qwen_vl_interface': 4.237609751713589e-06, 'learning_rate/action_model': 4.2376097517135896e-05, 'epoch': 1.36}) +[2026-06-19 16:19:27,805][starVLA.training.train_starvla][INFO] - Step 2886, Loss: {'train/loss': 0.18137620389461517, 'train/grad_norm_pre_clip': 9.729915618896484, 'timing/data': 0.009079201961867511, 'timing/model': 1.1538974419236183, 'learning_rate/qwen_vl_interface': 4.234634513787455e-06, 'learning_rate/action_model': 4.2346345137874556e-05, 'epoch': 1.36}) +[2026-06-19 16:19:32,438][starVLA.training.train_starvla][INFO] - Step 2887, Loss: {'train/loss': 0.1254829317331314, 'train/grad_norm_pre_clip': 10.300260543823242, 'timing/data': 0.009400088107213378, 'timing/model': 1.1644737240858376, 'learning_rate/qwen_vl_interface': 4.2316596932397195e-06, 'learning_rate/action_model': 4.2316596932397195e-05, 'epoch': 1.36}) +[2026-06-19 16:19:37,037][starVLA.training.train_starvla][INFO] - Step 2888, Loss: {'train/loss': 0.22176165878772736, 'train/grad_norm_pre_clip': 9.760042190551758, 'timing/data': 0.009490968892350793, 'timing/model': 1.3131738980300725, 'learning_rate/qwen_vl_interface': 4.22868529129322e-06, 'learning_rate/action_model': 4.22868529129322e-05, 'epoch': 1.36}) +[2026-06-19 16:19:41,588][starVLA.training.train_starvla][INFO] - Step 2889, Loss: {'train/loss': 0.21286186575889587, 'train/grad_norm_pre_clip': 9.28803539276123, 'timing/data': 0.010275011998601258, 'timing/model': 1.2606780980713665, 'learning_rate/qwen_vl_interface': 4.2257113091706165e-06, 'learning_rate/action_model': 4.2257113091706165e-05, 'epoch': 1.36}) +[2026-06-19 16:19:46,014][starVLA.training.train_starvla][INFO] - Step 2890, Loss: {'train/loss': 0.21997223794460297, 'train/grad_norm_pre_clip': 11.686238288879395, 'timing/data': 0.009919947013258934, 'timing/model': 1.1461375690996647, 'learning_rate/qwen_vl_interface': 4.222737748094404e-06, 'learning_rate/action_model': 4.222737748094404e-05, 'epoch': 1.36}) +[2026-06-19 16:19:50,627][starVLA.training.train_starvla][INFO] - Step 2891, Loss: {'train/loss': 0.25603464245796204, 'train/grad_norm_pre_clip': 9.230270385742188, 'timing/data': 0.009831962990574539, 'timing/model': 1.1485141329467297, 'learning_rate/qwen_vl_interface': 4.219764609286901e-06, 'learning_rate/action_model': 4.219764609286901e-05, 'epoch': 1.36}) +[2026-06-19 16:19:55,337][starVLA.training.train_starvla][INFO] - Step 2892, Loss: {'train/loss': 0.3463113605976105, 'train/grad_norm_pre_clip': 12.947611808776855, 'timing/data': 0.010895674000494182, 'timing/model': 1.160596244968474, 'learning_rate/qwen_vl_interface': 4.21679189397025e-06, 'learning_rate/action_model': 4.21679189397025e-05, 'epoch': 1.36}) +[2026-06-19 16:19:59,937][starVLA.training.train_starvla][INFO] - Step 2893, Loss: {'train/loss': 0.24516640603542328, 'train/grad_norm_pre_clip': 9.869757652282715, 'timing/data': 0.009864793042652309, 'timing/model': 1.1488435450010002, 'learning_rate/qwen_vl_interface': 4.213819603366424e-06, 'learning_rate/action_model': 4.213819603366424e-05, 'epoch': 1.36}) +[2026-06-19 16:20:04,557][starVLA.training.train_starvla][INFO] - Step 2894, Loss: {'train/loss': 0.11864332854747772, 'train/grad_norm_pre_clip': 8.404085159301758, 'timing/data': 0.009978319983929396, 'timing/model': 1.3218058929778636, 'learning_rate/qwen_vl_interface': 4.210847738697219e-06, 'learning_rate/action_model': 4.210847738697218e-05, 'epoch': 1.36}) +[2026-06-19 16:20:09,131][starVLA.training.train_starvla][INFO] - Step 2895, Loss: {'train/loss': 0.16825978457927704, 'train/grad_norm_pre_clip': 6.669706344604492, 'timing/data': 0.009565868997015059, 'timing/model': 1.2670344669604674, 'learning_rate/qwen_vl_interface': 4.207876301184256e-06, 'learning_rate/action_model': 4.207876301184256e-05, 'epoch': 1.36}) +[2026-06-19 16:20:13,588][starVLA.training.train_starvla][INFO] - Step 2896, Loss: {'train/loss': 0.2072352170944214, 'train/grad_norm_pre_clip': 26.465057373046875, 'timing/data': 0.009322280064225197, 'timing/model': 1.153520877007395, 'learning_rate/qwen_vl_interface': 4.2049052920489825e-06, 'learning_rate/action_model': 4.2049052920489825e-05, 'epoch': 1.36}) +[2026-06-19 16:20:18,211][starVLA.training.train_starvla][INFO] - Step 2897, Loss: {'train/loss': 0.18421198427677155, 'train/grad_norm_pre_clip': 10.532221794128418, 'timing/data': 0.010103662963956594, 'timing/model': 1.1528340530348942, 'learning_rate/qwen_vl_interface': 4.201934712512666e-06, 'learning_rate/action_model': 4.2019347125126656e-05, 'epoch': 1.36}) +[2026-06-19 16:20:22,836][starVLA.training.train_starvla][INFO] - Step 2898, Loss: {'train/loss': 0.25537753105163574, 'train/grad_norm_pre_clip': 10.87871265411377, 'timing/data': 0.009463469963520765, 'timing/model': 1.152285187970847, 'learning_rate/qwen_vl_interface': 4.1989645637964e-06, 'learning_rate/action_model': 4.1989645637964e-05, 'epoch': 1.36}) +[2026-06-19 16:20:27,444][starVLA.training.train_starvla][INFO] - Step 2899, Loss: {'train/loss': 0.26396751403808594, 'train/grad_norm_pre_clip': 9.25320053100586, 'timing/data': 0.009872970054857433, 'timing/model': 1.1466195329558104, 'learning_rate/qwen_vl_interface': 4.195994847121101e-06, 'learning_rate/action_model': 4.195994847121101e-05, 'epoch': 1.36}) +[2026-06-19 16:20:32,025][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 2900 +[2026-06-19 16:23:25,801][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 2900 +[2026-06-19 16:23:25,801][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 2900 +[2026-06-19 16:26:35,576][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 2900 +[2026-06-19 16:26:35,577][starVLA.training.train_starvla][INFO] - Step 2900, Loss: {'train/loss': 0.1739852875471115, 'train/grad_norm_pre_clip': 10.12207317352295, 'eval/loss': 0.20817697048187256, 'eval/latency_0/loss': 0.10474681854248047, 'eval/latency_1/loss': 0.059448663145303726, 'eval/latency_2/loss': 0.12155165523290634, 'eval/latency_3/loss': 0.20635218918323517, 'eval/latency_4/loss': 0.5480901598930359, 'eval/flappy/loss': 0.20817697048187256, 'eval/flappy/latency_0/loss': 0.10474681854248047, 'eval/flappy/latency_1/loss': 0.059448663145303726, 'eval/flappy/latency_2/loss': 0.12155165523290634, 'eval/flappy/latency_3/loss': 0.20635218918323517, 'eval/flappy/latency_4/loss': 0.5480901598930359, 'eval/flap_precision': 0.78125, 'eval/flap_recall': 0.6363636363636364, 'eval/flap_f1': 0.7014028056112225, 'eval/flap_rate_model': 0.06222222222222222, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.7014028056112225, 'eval/latency_0/flap_precision': 0.78125, 'eval/latency_0/flap_recall': 0.6363636363636364, 'eval/latency_0/flap_f1': 0.7014028056112225, 'eval/latency_0/flap_rate_model': 0.06222222222222222, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.7014028056112225, 'timing/data': 0.009766560979187489, 'timing/model': 1.1497722229687497, 'learning_rate/qwen_vl_interface': 4.193025563707507e-06, 'learning_rate/action_model': 4.1930255637075066e-05, 'epoch': 1.36}) +[2026-06-19 16:26:35,577][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_2900_state +[2026-06-19 16:26:35,580][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-19 16:26:38,979][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_2900_state/model.safetensors +[2026-06-19 16:26:46,267][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_2900_state/optimizer.bin +[2026-06-19 16:26:46,267][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_2900_state/sampler.bin +[2026-06-19 16:26:46,267][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_2900_state/sampler_1.bin +[2026-06-19 16:26:46,268][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_2900_state/random_states_0.pkl +[2026-06-19 16:26:46,289][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2900 +[2026-06-19 16:26:46,289][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-19 16:26:46,298][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-19 16:26:50,805][starVLA.training.train_starvla][INFO] - Step 2901, Loss: {'train/loss': 0.1875559687614441, 'train/grad_norm_pre_clip': 11.664214134216309, 'timing/data': 0.009498778963461518, 'timing/model': 1.1262233550660312, 'learning_rate/qwen_vl_interface': 4.19005671477618e-06, 'learning_rate/action_model': 4.1900567147761794e-05, 'epoch': 1.36}) +[2026-06-19 16:26:55,326][starVLA.training.train_starvla][INFO] - Step 2902, Loss: {'train/loss': 0.2160925716161728, 'train/grad_norm_pre_clip': 9.126104354858398, 'timing/data': 0.009488348034210503, 'timing/model': 1.2915300590684637, 'learning_rate/qwen_vl_interface': 4.1870883015474975e-06, 'learning_rate/action_model': 4.187088301547497e-05, 'epoch': 1.36}) +[2026-06-19 16:26:59,788][starVLA.training.train_starvla][INFO] - Step 2903, Loss: {'train/loss': 0.33054065704345703, 'train/grad_norm_pre_clip': 30.23049545288086, 'timing/data': 0.010017369990237057, 'timing/model': 1.23524251091294, 'learning_rate/qwen_vl_interface': 4.1841203252416655e-06, 'learning_rate/action_model': 4.1841203252416655e-05, 'epoch': 1.37}) +[2026-06-19 16:27:04,162][starVLA.training.train_starvla][INFO] - Step 2904, Loss: {'train/loss': 0.19530297815799713, 'train/grad_norm_pre_clip': 6.718280792236328, 'timing/data': 0.009886172018013895, 'timing/model': 1.1386922359233722, 'learning_rate/qwen_vl_interface': 4.181152787078705e-06, 'learning_rate/action_model': 4.1811527870787056e-05, 'epoch': 1.37}) +[2026-06-19 16:27:08,691][starVLA.training.train_starvla][INFO] - Step 2905, Loss: {'train/loss': 0.19286958873271942, 'train/grad_norm_pre_clip': 9.44240951538086, 'timing/data': 0.010028121061623096, 'timing/model': 1.136677712900564, 'learning_rate/qwen_vl_interface': 4.178185688278462e-06, 'learning_rate/action_model': 4.1781856882784616e-05, 'epoch': 1.37}) +[2026-06-19 16:27:13,234][starVLA.training.train_starvla][INFO] - Step 2906, Loss: {'train/loss': 0.18480628728866577, 'train/grad_norm_pre_clip': 13.126147270202637, 'timing/data': 0.010436433018185198, 'timing/model': 1.1477262279950082, 'learning_rate/qwen_vl_interface': 4.175219030060595e-06, 'learning_rate/action_model': 4.175219030060595e-05, 'epoch': 1.37}) +[2026-06-19 16:27:17,813][starVLA.training.train_starvla][INFO] - Step 2907, Loss: {'train/loss': 0.2869137227535248, 'train/grad_norm_pre_clip': 9.036323547363281, 'timing/data': 0.009831540985032916, 'timing/model': 1.1376684040296823, 'learning_rate/qwen_vl_interface': 4.1722528136445875e-06, 'learning_rate/action_model': 4.172252813644588e-05, 'epoch': 1.37}) +[2026-06-19 16:27:22,413][starVLA.training.train_starvla][INFO] - Step 2908, Loss: {'train/loss': 0.2424951046705246, 'train/grad_norm_pre_clip': 11.369537353515625, 'timing/data': 0.009985584998503327, 'timing/model': 1.3223340320400894, 'learning_rate/qwen_vl_interface': 4.169287040249738e-06, 'learning_rate/action_model': 4.169287040249738e-05, 'epoch': 1.37}) +[2026-06-19 16:27:26,815][starVLA.training.train_starvla][INFO] - Step 2909, Loss: {'train/loss': 0.30832189321517944, 'train/grad_norm_pre_clip': 10.625749588012695, 'timing/data': 0.009872719994746149, 'timing/model': 1.1382696530781686, 'learning_rate/qwen_vl_interface': 4.166321711095164e-06, 'learning_rate/action_model': 4.166321711095164e-05, 'epoch': 1.37}) +[2026-06-19 16:27:31,379][starVLA.training.train_starvla][INFO] - Step 2910, Loss: {'train/loss': 0.23102077841758728, 'train/grad_norm_pre_clip': 11.595368385314941, 'timing/data': 0.009791824035346508, 'timing/model': 1.1480990210548043, 'learning_rate/qwen_vl_interface': 4.1633568273998e-06, 'learning_rate/action_model': 4.1633568273998e-05, 'epoch': 1.37}) +[2026-06-19 16:27:35,953][starVLA.training.train_starvla][INFO] - Step 2911, Loss: {'train/loss': 0.1951228380203247, 'train/grad_norm_pre_clip': 5.295166015625, 'timing/data': 0.00978265202138573, 'timing/model': 1.1412137110019103, 'learning_rate/qwen_vl_interface': 4.1603923903823995e-06, 'learning_rate/action_model': 4.160392390382399e-05, 'epoch': 1.37}) +[2026-06-19 16:27:40,550][starVLA.training.train_starvla][INFO] - Step 2912, Loss: {'train/loss': 0.2003483921289444, 'train/grad_norm_pre_clip': 5.516648769378662, 'timing/data': 0.010969645925797522, 'timing/model': 1.148479901952669, 'learning_rate/qwen_vl_interface': 4.1574284012615276e-06, 'learning_rate/action_model': 4.157428401261528e-05, 'epoch': 1.37}) +[2026-06-19 16:27:45,146][starVLA.training.train_starvla][INFO] - Step 2913, Loss: {'train/loss': 0.32894518971443176, 'train/grad_norm_pre_clip': 12.373605728149414, 'timing/data': 0.009192886995151639, 'timing/model': 1.1552798630436882, 'learning_rate/qwen_vl_interface': 4.154464861255571e-06, 'learning_rate/action_model': 4.154464861255571e-05, 'epoch': 1.37}) +[2026-06-19 16:27:49,734][starVLA.training.train_starvla][INFO] - Step 2914, Loss: {'train/loss': 0.2558536231517792, 'train/grad_norm_pre_clip': 7.116426944732666, 'timing/data': 0.010506076971068978, 'timing/model': 1.3178050010465086, 'learning_rate/qwen_vl_interface': 4.151501771582728e-06, 'learning_rate/action_model': 4.151501771582728e-05, 'epoch': 1.37}) +[2026-06-19 16:27:54,292][starVLA.training.train_starvla][INFO] - Step 2915, Loss: {'train/loss': 0.3839866816997528, 'train/grad_norm_pre_clip': 10.049678802490234, 'timing/data': 0.009573690011166036, 'timing/model': 1.27541562798433, 'learning_rate/qwen_vl_interface': 4.148539133461012e-06, 'learning_rate/action_model': 4.1485391334610116e-05, 'epoch': 1.37}) +[2026-06-19 16:27:58,724][starVLA.training.train_starvla][INFO] - Step 2916, Loss: {'train/loss': 0.270402193069458, 'train/grad_norm_pre_clip': 11.072025299072266, 'timing/data': 0.010158378980122507, 'timing/model': 1.147586602019146, 'learning_rate/qwen_vl_interface': 4.145576948108254e-06, 'learning_rate/action_model': 4.145576948108253e-05, 'epoch': 1.37}) +[2026-06-19 16:28:03,317][starVLA.training.train_starvla][INFO] - Step 2917, Loss: {'train/loss': 0.09922105818986893, 'train/grad_norm_pre_clip': 4.060013294219971, 'timing/data': 0.009497194085270166, 'timing/model': 1.152676202938892, 'learning_rate/qwen_vl_interface': 4.142615216742093e-06, 'learning_rate/action_model': 4.142615216742093e-05, 'epoch': 1.37}) +[2026-06-19 16:28:07,920][starVLA.training.train_starvla][INFO] - Step 2918, Loss: {'train/loss': 0.08339046686887741, 'train/grad_norm_pre_clip': 5.276149749755859, 'timing/data': 0.010089896968565881, 'timing/model': 1.146492455038242, 'learning_rate/qwen_vl_interface': 4.139653940579988e-06, 'learning_rate/action_model': 4.1396539405799874e-05, 'epoch': 1.37}) +[2026-06-19 16:28:12,539][starVLA.training.train_starvla][INFO] - Step 2919, Loss: {'train/loss': 0.2932519316673279, 'train/grad_norm_pre_clip': 14.401768684387207, 'timing/data': 0.010146089014597237, 'timing/model': 1.1462119480129331, 'learning_rate/qwen_vl_interface': 4.136693120839206e-06, 'learning_rate/action_model': 4.136693120839205e-05, 'epoch': 1.37}) +[2026-06-19 16:28:17,109][starVLA.training.train_starvla][INFO] - Step 2920, Loss: {'train/loss': 0.25447964668273926, 'train/grad_norm_pre_clip': 11.977662086486816, 'timing/data': 0.009918867028318346, 'timing/model': 1.1417738139862195, 'learning_rate/qwen_vl_interface': 4.133732758736828e-06, 'learning_rate/action_model': 4.133732758736828e-05, 'epoch': 1.37}) +[2026-06-19 16:28:21,727][starVLA.training.train_starvla][INFO] - Step 2921, Loss: {'train/loss': 0.0837806686758995, 'train/grad_norm_pre_clip': 3.938657760620117, 'timing/data': 0.0097152249654755, 'timing/model': 1.3422188409604132, 'learning_rate/qwen_vl_interface': 4.13077285548975e-06, 'learning_rate/action_model': 4.13077285548975e-05, 'epoch': 1.37}) +[2026-06-19 16:28:26,183][starVLA.training.train_starvla][INFO] - Step 2922, Loss: {'train/loss': 0.17311641573905945, 'train/grad_norm_pre_clip': 5.902538299560547, 'timing/data': 0.00974230992142111, 'timing/model': 1.159598487894982, 'learning_rate/qwen_vl_interface': 4.127813412314672e-06, 'learning_rate/action_model': 4.1278134123146725e-05, 'epoch': 1.37}) +[2026-06-19 16:28:30,782][starVLA.training.train_starvla][INFO] - Step 2923, Loss: {'train/loss': 0.09129220992326736, 'train/grad_norm_pre_clip': 10.028647422790527, 'timing/data': 0.009565561078488827, 'timing/model': 1.150424845982343, 'learning_rate/qwen_vl_interface': 4.124854430428113e-06, 'learning_rate/action_model': 4.1248544304281126e-05, 'epoch': 1.37}) +[2026-06-19 16:28:35,387][starVLA.training.train_starvla][INFO] - Step 2924, Loss: {'train/loss': 0.19863784313201904, 'train/grad_norm_pre_clip': 10.940650939941406, 'timing/data': 0.009559966973029077, 'timing/model': 1.1476830840110779, 'learning_rate/qwen_vl_interface': 4.121895911046397e-06, 'learning_rate/action_model': 4.121895911046396e-05, 'epoch': 1.38}) +[2026-06-19 16:28:39,987][starVLA.training.train_starvla][INFO] - Step 2925, Loss: {'train/loss': 0.6406887173652649, 'train/grad_norm_pre_clip': 26.529094696044922, 'timing/data': 0.009640245931223035, 'timing/model': 1.1532770650228485, 'learning_rate/qwen_vl_interface': 4.118937855385659e-06, 'learning_rate/action_model': 4.1189378553856584e-05, 'epoch': 1.38}) +[2026-06-19 16:28:44,565][starVLA.training.train_starvla][INFO] - Step 2926, Loss: {'train/loss': 0.10888330638408661, 'train/grad_norm_pre_clip': 8.538959503173828, 'timing/data': 0.011192564037628472, 'timing/model': 1.142980964970775, 'learning_rate/qwen_vl_interface': 4.115980264661844e-06, 'learning_rate/action_model': 4.115980264661844e-05, 'epoch': 1.38}) +[2026-06-19 16:28:49,147][starVLA.training.train_starvla][INFO] - Step 2927, Loss: {'train/loss': 0.22776994109153748, 'train/grad_norm_pre_clip': 10.006476402282715, 'timing/data': 0.010513016022741795, 'timing/model': 1.3123851899290457, 'learning_rate/qwen_vl_interface': 4.113023140090709e-06, 'learning_rate/action_model': 4.113023140090709e-05, 'epoch': 1.38}) +[2026-06-19 16:28:53,591][starVLA.training.train_starvla][INFO] - Step 2928, Loss: {'train/loss': 0.16079136729240417, 'train/grad_norm_pre_clip': 9.416836738586426, 'timing/data': 0.010134669952094555, 'timing/model': 1.1543230150127783, 'learning_rate/qwen_vl_interface': 4.110066482887812e-06, 'learning_rate/action_model': 4.110066482887812e-05, 'epoch': 1.38}) +[2026-06-19 16:28:58,189][starVLA.training.train_starvla][INFO] - Step 2929, Loss: {'train/loss': 0.10535664111375809, 'train/grad_norm_pre_clip': 7.14697265625, 'timing/data': 0.009868979104794562, 'timing/model': 1.1448709099786356, 'learning_rate/qwen_vl_interface': 4.107110294268525e-06, 'learning_rate/action_model': 4.107110294268525e-05, 'epoch': 1.38}) +[2026-06-19 16:29:02,688][starVLA.training.train_starvla][INFO] - Step 2930, Loss: {'train/loss': 0.3527335822582245, 'train/grad_norm_pre_clip': 17.928516387939453, 'timing/data': 0.009820225066505373, 'timing/model': 1.1499545199330896, 'learning_rate/qwen_vl_interface': 4.104154575448026e-06, 'learning_rate/action_model': 4.104154575448025e-05, 'epoch': 1.38}) +[2026-06-19 16:29:07,203][starVLA.training.train_starvla][INFO] - Step 2931, Loss: {'train/loss': 0.13933898508548737, 'train/grad_norm_pre_clip': 13.858793258666992, 'timing/data': 0.009815981960855424, 'timing/model': 1.14240988495294, 'learning_rate/qwen_vl_interface': 4.101199327641297e-06, 'learning_rate/action_model': 4.101199327641297e-05, 'epoch': 1.38}) +[2026-06-19 16:29:11,803][starVLA.training.train_starvla][INFO] - Step 2932, Loss: {'train/loss': 0.46439340710639954, 'train/grad_norm_pre_clip': 16.445491790771484, 'timing/data': 0.00990284897852689, 'timing/model': 1.149022739031352, 'learning_rate/qwen_vl_interface': 4.098244552063133e-06, 'learning_rate/action_model': 4.098244552063132e-05, 'epoch': 1.38}) +[2026-06-19 16:29:16,387][starVLA.training.train_starvla][INFO] - Step 2933, Loss: {'train/loss': 0.12847861647605896, 'train/grad_norm_pre_clip': 6.517320156097412, 'timing/data': 0.00955029798205942, 'timing/model': 1.3062868859851733, 'learning_rate/qwen_vl_interface': 4.0952902499281254e-06, 'learning_rate/action_model': 4.095290249928125e-05, 'epoch': 1.38}) +[2026-06-19 16:29:20,827][starVLA.training.train_starvla][INFO] - Step 2934, Loss: {'train/loss': 0.311534583568573, 'train/grad_norm_pre_clip': 15.319781303405762, 'timing/data': 0.009712587110698223, 'timing/model': 1.1484012289438397, 'learning_rate/qwen_vl_interface': 4.092336422450681e-06, 'learning_rate/action_model': 4.0923364224506804e-05, 'epoch': 1.38}) +[2026-06-19 16:29:25,427][starVLA.training.train_starvla][INFO] - Step 2935, Loss: {'train/loss': 0.1929302215576172, 'train/grad_norm_pre_clip': 9.08784294128418, 'timing/data': 0.01038158789742738, 'timing/model': 1.1452287980355322, 'learning_rate/qwen_vl_interface': 4.089383070845003e-06, 'learning_rate/action_model': 4.089383070845003e-05, 'epoch': 1.38}) +[2026-06-19 16:29:30,045][starVLA.training.train_starvla][INFO] - Step 2936, Loss: {'train/loss': 0.2670176327228546, 'train/grad_norm_pre_clip': 10.424714088439941, 'timing/data': 0.010556914028711617, 'timing/model': 1.1597750950604677, 'learning_rate/qwen_vl_interface': 4.086430196325106e-06, 'learning_rate/action_model': 4.086430196325106e-05, 'epoch': 1.38}) +[2026-06-19 16:29:34,632][starVLA.training.train_starvla][INFO] - Step 2937, Loss: {'train/loss': 0.1565847098827362, 'train/grad_norm_pre_clip': 11.062470436096191, 'timing/data': 0.009764751070179045, 'timing/model': 1.1478870789287612, 'learning_rate/qwen_vl_interface': 4.083477800104802e-06, 'learning_rate/action_model': 4.083477800104802e-05, 'epoch': 1.38}) +[2026-06-19 16:29:39,256][starVLA.training.train_starvla][INFO] - Step 2938, Loss: {'train/loss': 0.3451416492462158, 'train/grad_norm_pre_clip': 14.411580085754395, 'timing/data': 0.01012316404376179, 'timing/model': 1.1493868860416114, 'learning_rate/qwen_vl_interface': 4.080525883397711e-06, 'learning_rate/action_model': 4.080525883397711e-05, 'epoch': 1.38}) +[2026-06-19 16:29:43,889][starVLA.training.train_starvla][INFO] - Step 2939, Loss: {'train/loss': 0.11241668462753296, 'train/grad_norm_pre_clip': 13.56031322479248, 'timing/data': 0.010084462934173644, 'timing/model': 1.3300797609845176, 'learning_rate/qwen_vl_interface': 4.077574447417256e-06, 'learning_rate/action_model': 4.077574447417256e-05, 'epoch': 1.38}) +[2026-06-19 16:29:48,321][starVLA.training.train_starvla][INFO] - Step 2940, Loss: {'train/loss': 0.18861153721809387, 'train/grad_norm_pre_clip': 8.41348648071289, 'timing/data': 0.009562635910697281, 'timing/model': 1.1541440419387072, 'learning_rate/qwen_vl_interface': 4.074623493376657e-06, 'learning_rate/action_model': 4.074623493376657e-05, 'epoch': 1.38}) +[2026-06-19 16:29:52,942][starVLA.training.train_starvla][INFO] - Step 2941, Loss: {'train/loss': 0.30073782801628113, 'train/grad_norm_pre_clip': 7.65346097946167, 'timing/data': 0.01014860498253256, 'timing/model': 1.1549598609562963, 'learning_rate/qwen_vl_interface': 4.071673022488944e-06, 'learning_rate/action_model': 4.071673022488944e-05, 'epoch': 1.38}) +[2026-06-19 16:29:57,441][starVLA.training.train_starvla][INFO] - Step 2942, Loss: {'train/loss': 0.31663161516189575, 'train/grad_norm_pre_clip': 9.017399787902832, 'timing/data': 0.009735706960782409, 'timing/model': 1.1462887598900124, 'learning_rate/qwen_vl_interface': 4.06872303596694e-06, 'learning_rate/action_model': 4.06872303596694e-05, 'epoch': 1.38}) +[2026-06-19 16:30:02,054][starVLA.training.train_starvla][INFO] - Step 2943, Loss: {'train/loss': 0.19910874962806702, 'train/grad_norm_pre_clip': 15.087759971618652, 'timing/data': 0.009632723988033831, 'timing/model': 1.1598123599542305, 'learning_rate/qwen_vl_interface': 4.065773535023277e-06, 'learning_rate/action_model': 4.0657735350232766e-05, 'epoch': 1.38}) +[2026-06-19 16:30:06,693][starVLA.training.train_starvla][INFO] - Step 2944, Loss: {'train/loss': 0.14251099526882172, 'train/grad_norm_pre_clip': 7.874601364135742, 'timing/data': 0.009965520002879202, 'timing/model': 1.1592216269345954, 'learning_rate/qwen_vl_interface': 4.062824520870382e-06, 'learning_rate/action_model': 4.062824520870381e-05, 'epoch': 1.38}) +[2026-06-19 16:30:11,342][starVLA.training.train_starvla][INFO] - Step 2945, Loss: {'train/loss': 0.3682752847671509, 'train/grad_norm_pre_clip': 20.413698196411133, 'timing/data': 0.009553964948281646, 'timing/model': 1.3500654420349747, 'learning_rate/qwen_vl_interface': 4.059875994720481e-06, 'learning_rate/action_model': 4.059875994720481e-05, 'epoch': 1.39}) +[2026-06-19 16:30:15,897][starVLA.training.train_starvla][INFO] - Step 2946, Loss: {'train/loss': 0.27756303548812866, 'train/grad_norm_pre_clip': 11.571721076965332, 'timing/data': 0.009732613922096789, 'timing/model': 1.270056372974068, 'learning_rate/qwen_vl_interface': 4.056927957785605e-06, 'learning_rate/action_model': 4.056927957785605e-05, 'epoch': 1.39}) +[2026-06-19 16:30:20,332][starVLA.training.train_starvla][INFO] - Step 2947, Loss: {'train/loss': 0.20499804615974426, 'train/grad_norm_pre_clip': 6.267930030822754, 'timing/data': 0.010497619048692286, 'timing/model': 1.1499652669299394, 'learning_rate/qwen_vl_interface': 4.0539804112775785e-06, 'learning_rate/action_model': 4.053980411277578e-05, 'epoch': 1.39}) +[2026-06-19 16:30:24,957][starVLA.training.train_starvla][INFO] - Step 2948, Loss: {'train/loss': 0.30357682704925537, 'train/grad_norm_pre_clip': 8.295902252197266, 'timing/data': 0.00905872497241944, 'timing/model': 1.1592236269498244, 'learning_rate/qwen_vl_interface': 4.051033356408027e-06, 'learning_rate/action_model': 4.0510333564080264e-05, 'epoch': 1.39}) +[2026-06-19 16:30:29,601][starVLA.training.train_starvla][INFO] - Step 2949, Loss: {'train/loss': 0.16947661340236664, 'train/grad_norm_pre_clip': 6.249619007110596, 'timing/data': 0.010942222084850073, 'timing/model': 1.1558480981038883, 'learning_rate/qwen_vl_interface': 4.048086794388375e-06, 'learning_rate/action_model': 4.048086794388375e-05, 'epoch': 1.39}) +[2026-06-19 16:30:34,218][starVLA.training.train_starvla][INFO] - Step 2950, Loss: {'train/loss': 0.4985557198524475, 'train/grad_norm_pre_clip': 17.118215560913086, 'timing/data': 0.009860371937975287, 'timing/model': 1.1474881910253316, 'learning_rate/qwen_vl_interface': 4.045140726429841e-06, 'learning_rate/action_model': 4.0451407264298406e-05, 'epoch': 1.39}) +[2026-06-19 16:30:38,740][starVLA.training.train_starvla][INFO] - Step 2951, Loss: {'train/loss': 0.13140060007572174, 'train/grad_norm_pre_clip': 10.383590698242188, 'timing/data': 0.010116172954440117, 'timing/model': 1.1509195220423862, 'learning_rate/qwen_vl_interface': 4.042195153743443e-06, 'learning_rate/action_model': 4.042195153743443e-05, 'epoch': 1.39}) +[2026-06-19 16:30:43,252][starVLA.training.train_starvla][INFO] - Step 2952, Loss: {'train/loss': 0.22045408189296722, 'train/grad_norm_pre_clip': 5.03858757019043, 'timing/data': 0.009717059088870883, 'timing/model': 1.2424289899645373, 'learning_rate/qwen_vl_interface': 4.039250077539993e-06, 'learning_rate/action_model': 4.039250077539993e-05, 'epoch': 1.39}) +[2026-06-19 16:30:47,693][starVLA.training.train_starvla][INFO] - Step 2953, Loss: {'train/loss': 0.4631900489330292, 'train/grad_norm_pre_clip': 20.669261932373047, 'timing/data': 0.010492622968740761, 'timing/model': 1.155387661070563, 'learning_rate/qwen_vl_interface': 4.036305499030104e-06, 'learning_rate/action_model': 4.0363054990301036e-05, 'epoch': 1.39}) +[2026-06-19 16:30:52,336][starVLA.training.train_starvla][INFO] - Step 2954, Loss: {'train/loss': 0.32480987906455994, 'train/grad_norm_pre_clip': 18.944740295410156, 'timing/data': 0.009602176025509834, 'timing/model': 1.1561625009635463, 'learning_rate/qwen_vl_interface': 4.033361419424178e-06, 'learning_rate/action_model': 4.033361419424178e-05, 'epoch': 1.39}) +[2026-06-19 16:30:56,938][starVLA.training.train_starvla][INFO] - Step 2955, Loss: {'train/loss': 0.24963156878948212, 'train/grad_norm_pre_clip': 11.257597923278809, 'timing/data': 0.010288910008966923, 'timing/model': 1.144618854043074, 'learning_rate/qwen_vl_interface': 4.030417839932418e-06, 'learning_rate/action_model': 4.030417839932417e-05, 'epoch': 1.39}) +[2026-06-19 16:31:01,555][starVLA.training.train_starvla][INFO] - Step 2956, Loss: {'train/loss': 0.10987843573093414, 'train/grad_norm_pre_clip': 4.815155982971191, 'timing/data': 0.009467994910664856, 'timing/model': 1.1633639759384096, 'learning_rate/qwen_vl_interface': 4.027474761764815e-06, 'learning_rate/action_model': 4.027474761764814e-05, 'epoch': 1.39}) +[2026-06-19 16:31:06,167][starVLA.training.train_starvla][INFO] - Step 2957, Loss: {'train/loss': 0.32853013277053833, 'train/grad_norm_pre_clip': 8.008539199829102, 'timing/data': 0.009728402947075665, 'timing/model': 1.3177990959957242, 'learning_rate/qwen_vl_interface': 4.024532186131158e-06, 'learning_rate/action_model': 4.0245321861311575e-05, 'epoch': 1.39}) +[2026-06-19 16:31:10,694][starVLA.training.train_starvla][INFO] - Step 2958, Loss: {'train/loss': 0.23261649906635284, 'train/grad_norm_pre_clip': 6.240427017211914, 'timing/data': 0.009923830046318471, 'timing/model': 1.237560601090081, 'learning_rate/qwen_vl_interface': 4.021590114241029e-06, 'learning_rate/action_model': 4.021590114241029e-05, 'epoch': 1.39}) +[2026-06-19 16:31:15,150][starVLA.training.train_starvla][INFO] - Step 2959, Loss: {'train/loss': 0.11459547281265259, 'train/grad_norm_pre_clip': 8.25273323059082, 'timing/data': 0.009979798924177885, 'timing/model': 1.1536435639718547, 'learning_rate/qwen_vl_interface': 4.018648547303804e-06, 'learning_rate/action_model': 4.0186485473038036e-05, 'epoch': 1.39}) +[2026-06-19 16:31:19,762][starVLA.training.train_starvla][INFO] - Step 2960, Loss: {'train/loss': 0.23658783733844757, 'train/grad_norm_pre_clip': 8.552657127380371, 'timing/data': 0.009912675013765693, 'timing/model': 1.1611532929819077, 'learning_rate/qwen_vl_interface': 4.015707486528649e-06, 'learning_rate/action_model': 4.015707486528649e-05, 'epoch': 1.39}) +[2026-06-19 16:31:24,360][starVLA.training.train_starvla][INFO] - Step 2961, Loss: {'train/loss': 0.14700093865394592, 'train/grad_norm_pre_clip': 6.5802483558654785, 'timing/data': 0.009804399916902184, 'timing/model': 1.14887886994984, 'learning_rate/qwen_vl_interface': 4.012766933124523e-06, 'learning_rate/action_model': 4.012766933124523e-05, 'epoch': 1.39}) +[2026-06-19 16:31:28,964][starVLA.training.train_starvla][INFO] - Step 2962, Loss: {'train/loss': 0.5199365615844727, 'train/grad_norm_pre_clip': 18.186420440673828, 'timing/data': 0.010654251091182232, 'timing/model': 1.1516007820609957, 'learning_rate/qwen_vl_interface': 4.009826888300176e-06, 'learning_rate/action_model': 4.0098268883001755e-05, 'epoch': 1.39}) +[2026-06-19 16:31:33,592][starVLA.training.train_starvla][INFO] - Step 2963, Loss: {'train/loss': 0.10513965040445328, 'train/grad_norm_pre_clip': 4.649313449859619, 'timing/data': 0.0100835170596838, 'timing/model': 1.328321064938791, 'learning_rate/qwen_vl_interface': 4.00688735326415e-06, 'learning_rate/action_model': 4.0068873532641495e-05, 'epoch': 1.39}) +[2026-06-19 16:31:38,037][starVLA.training.train_starvla][INFO] - Step 2964, Loss: {'train/loss': 0.2170955389738083, 'train/grad_norm_pre_clip': 12.509039878845215, 'timing/data': 0.010179540957324207, 'timing/model': 1.1511545620160177, 'learning_rate/qwen_vl_interface': 4.0039483292247765e-06, 'learning_rate/action_model': 4.003948329224776e-05, 'epoch': 1.39}) +[2026-06-19 16:31:42,641][starVLA.training.train_starvla][INFO] - Step 2965, Loss: {'train/loss': 0.27166086435317993, 'train/grad_norm_pre_clip': 9.01297664642334, 'timing/data': 0.009724991978146136, 'timing/model': 1.1534832380712032, 'learning_rate/qwen_vl_interface': 4.001009817390178e-06, 'learning_rate/action_model': 4.001009817390178e-05, 'epoch': 1.39}) +[2026-06-19 16:31:47,230][starVLA.training.train_starvla][INFO] - Step 2966, Loss: {'train/loss': 0.3132038712501526, 'train/grad_norm_pre_clip': 10.091878890991211, 'timing/data': 0.009301883983425796, 'timing/model': 1.1463404200039804, 'learning_rate/qwen_vl_interface': 3.998071818968264e-06, 'learning_rate/action_model': 3.998071818968264e-05, 'epoch': 1.39}) +[2026-06-19 16:31:51,833][starVLA.training.train_starvla][INFO] - Step 2967, Loss: {'train/loss': 0.2836216986179352, 'train/grad_norm_pre_clip': 5.9075493812561035, 'timing/data': 0.009841102990321815, 'timing/model': 1.1568338589277118, 'learning_rate/qwen_vl_interface': 3.995134335166736e-06, 'learning_rate/action_model': 3.995134335166736e-05, 'epoch': 1.4}) +[2026-06-19 16:31:56,451][starVLA.training.train_starvla][INFO] - Step 2968, Loss: {'train/loss': 0.2030157744884491, 'train/grad_norm_pre_clip': 7.142405033111572, 'timing/data': 0.009907654020935297, 'timing/model': 1.1540322450455278, 'learning_rate/qwen_vl_interface': 3.992197367193083e-06, 'learning_rate/action_model': 3.992197367193082e-05, 'epoch': 1.4}) +[2026-06-19 16:32:01,068][starVLA.training.train_starvla][INFO] - Step 2969, Loss: {'train/loss': 0.11810330301523209, 'train/grad_norm_pre_clip': 7.044968605041504, 'timing/data': 0.010345956892706454, 'timing/model': 1.32531124597881, 'learning_rate/qwen_vl_interface': 3.98926091625458e-06, 'learning_rate/action_model': 3.9892609162545795e-05, 'epoch': 1.4}) +[2026-06-19 16:32:05,625][starVLA.training.train_starvla][INFO] - Step 2970, Loss: {'train/loss': 0.22504518926143646, 'train/grad_norm_pre_clip': 10.96170425415039, 'timing/data': 0.009517305996268988, 'timing/model': 1.2756766329985112, 'learning_rate/qwen_vl_interface': 3.986324983558294e-06, 'learning_rate/action_model': 3.9863249835582934e-05, 'epoch': 1.4}) +[2026-06-19 16:32:10,065][starVLA.training.train_starvla][INFO] - Step 2971, Loss: {'train/loss': 0.18531370162963867, 'train/grad_norm_pre_clip': 12.406578063964844, 'timing/data': 0.010219702031463385, 'timing/model': 1.1500752600841224, 'learning_rate/qwen_vl_interface': 3.983389570311073e-06, 'learning_rate/action_model': 3.9833895703110724e-05, 'epoch': 1.4}) +[2026-06-19 16:32:14,704][starVLA.training.train_starvla][INFO] - Step 2972, Loss: {'train/loss': 0.21622516214847565, 'train/grad_norm_pre_clip': 9.961750984191895, 'timing/data': 0.009914697031490505, 'timing/model': 1.15485357795842, 'learning_rate/qwen_vl_interface': 3.9804546777195546e-06, 'learning_rate/action_model': 3.980454677719554e-05, 'epoch': 1.4}) +[2026-06-19 16:32:19,312][starVLA.training.train_starvla][INFO] - Step 2973, Loss: {'train/loss': 0.30053189396858215, 'train/grad_norm_pre_clip': 8.608501434326172, 'timing/data': 0.009856675984337926, 'timing/model': 1.1513991479296237, 'learning_rate/qwen_vl_interface': 3.977520306990164e-06, 'learning_rate/action_model': 3.977520306990163e-05, 'epoch': 1.4}) +[2026-06-19 16:32:23,924][starVLA.training.train_starvla][INFO] - Step 2974, Loss: {'train/loss': 0.1549694985151291, 'train/grad_norm_pre_clip': 7.5889081954956055, 'timing/data': 0.00971767003647983, 'timing/model': 1.1531327530974522, 'learning_rate/qwen_vl_interface': 3.974586459329109e-06, 'learning_rate/action_model': 3.9745864593291085e-05, 'epoch': 1.4}) +[2026-06-19 16:32:28,515][starVLA.training.train_starvla][INFO] - Step 2975, Loss: {'train/loss': 0.14695750176906586, 'train/grad_norm_pre_clip': 4.959390163421631, 'timing/data': 0.009735825005918741, 'timing/model': 1.145882389973849, 'learning_rate/qwen_vl_interface': 3.9716531359423855e-06, 'learning_rate/action_model': 3.971653135942385e-05, 'epoch': 1.4}) +[2026-06-19 16:32:33,155][starVLA.training.train_starvla][INFO] - Step 2976, Loss: {'train/loss': 0.4120577275753021, 'train/grad_norm_pre_clip': 14.157498359680176, 'timing/data': 0.010390592971816659, 'timing/model': 1.3307214040542021, 'learning_rate/qwen_vl_interface': 3.968720338035769e-06, 'learning_rate/action_model': 3.968720338035769e-05, 'epoch': 1.4}) +[2026-06-19 16:32:37,587][starVLA.training.train_starvla][INFO] - Step 2977, Loss: {'train/loss': 0.2739020586013794, 'train/grad_norm_pre_clip': 15.26274585723877, 'timing/data': 0.01154504797887057, 'timing/model': 1.1484012369764969, 'learning_rate/qwen_vl_interface': 3.965788066814824e-06, 'learning_rate/action_model': 3.965788066814823e-05, 'epoch': 1.4}) +[2026-06-19 16:32:42,185][starVLA.training.train_starvla][INFO] - Step 2978, Loss: {'train/loss': 0.2468860000371933, 'train/grad_norm_pre_clip': 14.87668228149414, 'timing/data': 0.0095435039838776, 'timing/model': 1.1451959389960393, 'learning_rate/qwen_vl_interface': 3.962856323484895e-06, 'learning_rate/action_model': 3.9628563234848945e-05, 'epoch': 1.4}) +[2026-06-19 16:32:46,785][starVLA.training.train_starvla][INFO] - Step 2979, Loss: {'train/loss': 0.3495543301105499, 'train/grad_norm_pre_clip': 13.425588607788086, 'timing/data': 0.010545618017204106, 'timing/model': 1.1533050340367481, 'learning_rate/qwen_vl_interface': 3.959925109251112e-06, 'learning_rate/action_model': 3.959925109251112e-05, 'epoch': 1.4}) +[2026-06-19 16:32:51,396][starVLA.training.train_starvla][INFO] - Step 2980, Loss: {'train/loss': 0.23627136647701263, 'train/grad_norm_pre_clip': 17.467761993408203, 'timing/data': 0.010074825957417488, 'timing/model': 1.1487923139939085, 'learning_rate/qwen_vl_interface': 3.956994425318387e-06, 'learning_rate/action_model': 3.956994425318387e-05, 'epoch': 1.4}) +[2026-06-19 16:32:56,002][starVLA.training.train_starvla][INFO] - Step 2981, Loss: {'train/loss': 0.14438019692897797, 'train/grad_norm_pre_clip': 8.837894439697266, 'timing/data': 0.010228385915979743, 'timing/model': 1.1535036829300225, 'learning_rate/qwen_vl_interface': 3.9540642728914115e-06, 'learning_rate/action_model': 3.9540642728914115e-05, 'epoch': 1.4}) +[2026-06-19 16:33:00,608][starVLA.training.train_starvla][INFO] - Step 2982, Loss: {'train/loss': 0.2859226167201996, 'train/grad_norm_pre_clip': 13.48912525177002, 'timing/data': 0.010134491021744907, 'timing/model': 1.3244816170772538, 'learning_rate/qwen_vl_interface': 3.951134653174664e-06, 'learning_rate/action_model': 3.951134653174663e-05, 'epoch': 1.4}) +[2026-06-19 16:33:05,065][starVLA.training.train_starvla][INFO] - Step 2983, Loss: {'train/loss': 0.2514066696166992, 'train/grad_norm_pre_clip': 12.120179176330566, 'timing/data': 0.009573637042194605, 'timing/model': 1.1522698800545186, 'learning_rate/qwen_vl_interface': 3.948205567372397e-06, 'learning_rate/action_model': 3.948205567372397e-05, 'epoch': 1.4}) +[2026-06-19 16:33:09,689][starVLA.training.train_starvla][INFO] - Step 2984, Loss: {'train/loss': 0.24183116853237152, 'train/grad_norm_pre_clip': 13.91844654083252, 'timing/data': 0.010041891015134752, 'timing/model': 1.1522682399954647, 'learning_rate/qwen_vl_interface': 3.945277016688647e-06, 'learning_rate/action_model': 3.945277016688647e-05, 'epoch': 1.4}) +[2026-06-19 16:33:14,244][starVLA.training.train_starvla][INFO] - Step 2985, Loss: {'train/loss': 0.1583370864391327, 'train/grad_norm_pre_clip': 12.85307502746582, 'timing/data': 0.010328965028747916, 'timing/model': 1.164286470040679, 'learning_rate/qwen_vl_interface': 3.9423490023272326e-06, 'learning_rate/action_model': 3.9423490023272324e-05, 'epoch': 1.4}) +[2026-06-19 16:33:18,799][starVLA.training.train_starvla][INFO] - Step 2986, Loss: {'train/loss': 0.22392027080059052, 'train/grad_norm_pre_clip': 15.388007164001465, 'timing/data': 0.009671555017121136, 'timing/model': 1.158546069986187, 'learning_rate/qwen_vl_interface': 3.939421525491748e-06, 'learning_rate/action_model': 3.9394215254917483e-05, 'epoch': 1.4}) +[2026-06-19 16:33:23,443][starVLA.training.train_starvla][INFO] - Step 2987, Loss: {'train/loss': 0.27686020731925964, 'train/grad_norm_pre_clip': 11.330371856689453, 'timing/data': 0.010374047909863293, 'timing/model': 1.156774592003785, 'learning_rate/qwen_vl_interface': 3.9364945873855705e-06, 'learning_rate/action_model': 3.93649458738557e-05, 'epoch': 1.4}) +[2026-06-19 16:33:28,059][starVLA.training.train_starvla][INFO] - Step 2988, Loss: {'train/loss': 0.22607974708080292, 'train/grad_norm_pre_clip': 11.494580268859863, 'timing/data': 0.010010412079282105, 'timing/model': 1.335978330927901, 'learning_rate/qwen_vl_interface': 3.9335681892118536e-06, 'learning_rate/action_model': 3.933568189211854e-05, 'epoch': 1.41}) +[2026-06-19 16:33:32,516][starVLA.training.train_starvla][INFO] - Step 2989, Loss: {'train/loss': 0.1075284406542778, 'train/grad_norm_pre_clip': 7.294155597686768, 'timing/data': 0.009894750081002712, 'timing/model': 1.148850522004068, 'learning_rate/qwen_vl_interface': 3.930642332173527e-06, 'learning_rate/action_model': 3.930642332173527e-05, 'epoch': 1.41}) +[2026-06-19 16:33:37,128][starVLA.training.train_starvla][INFO] - Step 2990, Loss: {'train/loss': 0.3486453890800476, 'train/grad_norm_pre_clip': 16.93640899658203, 'timing/data': 0.009976960020139813, 'timing/model': 1.1552314120344818, 'learning_rate/qwen_vl_interface': 3.927717017473301e-06, 'learning_rate/action_model': 3.9277170174733006e-05, 'epoch': 1.41}) +[2026-06-19 16:33:41,740][starVLA.training.train_starvla][INFO] - Step 2991, Loss: {'train/loss': 0.14547288417816162, 'train/grad_norm_pre_clip': 16.4158878326416, 'timing/data': 0.009705235017463565, 'timing/model': 1.1495170290581882, 'learning_rate/qwen_vl_interface': 3.9247922463136615e-06, 'learning_rate/action_model': 3.924792246313662e-05, 'epoch': 1.41}) +[2026-06-19 16:33:46,354][starVLA.training.train_starvla][INFO] - Step 2992, Loss: {'train/loss': 0.21635578572750092, 'train/grad_norm_pre_clip': 14.912454605102539, 'timing/data': 0.01013145491015166, 'timing/model': 1.1485789109719917, 'learning_rate/qwen_vl_interface': 3.921868019896873e-06, 'learning_rate/action_model': 3.921868019896873e-05, 'epoch': 1.41}) +[2026-06-19 16:33:50,969][starVLA.training.train_starvla][INFO] - Step 2993, Loss: {'train/loss': 0.4749862551689148, 'train/grad_norm_pre_clip': 18.02393341064453, 'timing/data': 0.010220420081168413, 'timing/model': 1.1583599060541019, 'learning_rate/qwen_vl_interface': 3.918944339424974e-06, 'learning_rate/action_model': 3.9189443394249736e-05, 'epoch': 1.41}) +[2026-06-19 16:33:55,595][starVLA.training.train_starvla][INFO] - Step 2994, Loss: {'train/loss': 0.3187606930732727, 'train/grad_norm_pre_clip': 54.88065719604492, 'timing/data': 0.009833498974330723, 'timing/model': 1.3242286749882624, 'learning_rate/qwen_vl_interface': 3.9160212060997774e-06, 'learning_rate/action_model': 3.9160212060997774e-05, 'epoch': 1.41}) +[2026-06-19 16:34:00,027][starVLA.training.train_starvla][INFO] - Step 2995, Loss: {'train/loss': 0.127122700214386, 'train/grad_norm_pre_clip': 6.492264747619629, 'timing/data': 0.010000222944654524, 'timing/model': 1.1473891430068761, 'learning_rate/qwen_vl_interface': 3.913098621122875e-06, 'learning_rate/action_model': 3.913098621122875e-05, 'epoch': 1.41}) +[2026-06-19 16:34:04,665][starVLA.training.train_starvla][INFO] - Step 2996, Loss: {'train/loss': 0.2306230664253235, 'train/grad_norm_pre_clip': 8.79995346069336, 'timing/data': 0.009643884957768023, 'timing/model': 1.149221993982792, 'learning_rate/qwen_vl_interface': 3.91017658569563e-06, 'learning_rate/action_model': 3.91017658569563e-05, 'epoch': 1.41}) +[2026-06-19 16:34:09,183][starVLA.training.train_starvla][INFO] - Step 2997, Loss: {'train/loss': 0.4429936110973358, 'train/grad_norm_pre_clip': 95.26753234863281, 'timing/data': 0.009115421096794307, 'timing/model': 1.1505065719829872, 'learning_rate/qwen_vl_interface': 3.907255101019181e-06, 'learning_rate/action_model': 3.907255101019181e-05, 'epoch': 1.41}) +[2026-06-19 16:34:13,755][starVLA.training.train_starvla][INFO] - Step 2998, Loss: {'train/loss': 0.10978253185749054, 'train/grad_norm_pre_clip': 5.190329551696777, 'timing/data': 0.009623948950320482, 'timing/model': 1.1483619180507958, 'learning_rate/qwen_vl_interface': 3.904334168294442e-06, 'learning_rate/action_model': 3.9043341682944415e-05, 'epoch': 1.41}) +[2026-06-19 16:34:18,360][starVLA.training.train_starvla][INFO] - Step 2999, Loss: {'train/loss': 0.14230875670909882, 'train/grad_norm_pre_clip': 8.07446002960205, 'timing/data': 0.009432466002181172, 'timing/model': 1.1503735079895705, 'learning_rate/qwen_vl_interface': 3.901413788722093e-06, 'learning_rate/action_model': 3.901413788722093e-05, 'epoch': 1.41}) +[2026-06-19 16:34:22,993][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 3000 +[2026-06-19 16:37:20,424][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 3000 +[2026-06-19 16:37:20,424][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 3000 +[2026-06-19 16:40:31,385][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 3000 +[2026-06-19 16:40:31,386][starVLA.training.train_starvla][INFO] - Step 3000, Loss: {'train/loss': 0.1785402148962021, 'train/grad_norm_pre_clip': 5.927824974060059, 'eval/loss': 0.19521953165531158, 'eval/latency_0/loss': 0.09338618814945221, 'eval/latency_1/loss': 0.05629163235425949, 'eval/latency_2/loss': 0.11442102491855621, 'eval/latency_3/loss': 0.21182481944561005, 'eval/latency_4/loss': 0.5008140206336975, 'eval/flappy/loss': 0.19521953165531158, 'eval/flappy/latency_0/loss': 0.09338618814945221, 'eval/flappy/latency_1/loss': 0.05629163235425949, 'eval/flappy/latency_2/loss': 0.11442102491855621, 'eval/flappy/latency_3/loss': 0.21182481944561005, 'eval/flappy/latency_4/loss': 0.5008140206336975, 'eval/flap_precision': 0.6903225806451613, 'eval/flap_recall': 0.7781818181818182, 'eval/flap_f1': 0.7316239316239317, 'eval/flap_rate_model': 0.08611111111111111, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.7316239316239317, 'eval/latency_0/flap_precision': 0.6903225806451613, 'eval/latency_0/flap_recall': 0.7781818181818182, 'eval/latency_0/flap_f1': 0.7316239316239317, 'eval/latency_0/flap_rate_model': 0.08611111111111111, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.7316239316239317, 'timing/data': 0.009810257004573941, 'timing/model': 1.3375193130923435, 'learning_rate/qwen_vl_interface': 3.8984939635025966e-06, 'learning_rate/action_model': 3.898493963502596e-05, 'epoch': 1.41}) +[2026-06-19 16:40:31,387][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_3000_state +[2026-06-19 16:40:31,389][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-19 16:40:34,679][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_3000_state/model.safetensors +[2026-06-19 16:40:41,909][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_3000_state/optimizer.bin +[2026-06-19 16:40:41,910][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_3000_state/sampler.bin +[2026-06-19 16:40:41,910][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_3000_state/sampler_1.bin +[2026-06-19 16:40:41,910][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_3000_state/random_states_0.pkl +[2026-06-19 16:40:41,935][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3000 +[2026-06-19 16:40:41,935][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-19 16:40:41,955][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-19 16:40:46,420][starVLA.training.train_starvla][INFO] - Step 3001, Loss: {'train/loss': 0.3505116403102875, 'train/grad_norm_pre_clip': 10.660911560058594, 'timing/data': 0.00921503803692758, 'timing/model': 1.1247644890099764, 'learning_rate/qwen_vl_interface': 3.895574693836181e-06, 'learning_rate/action_model': 3.8955746938361804e-05, 'epoch': 1.41}) +[2026-06-19 16:40:50,939][starVLA.training.train_starvla][INFO] - Step 3002, Loss: {'train/loss': 0.2230231910943985, 'train/grad_norm_pre_clip': 33.55555725097656, 'timing/data': 0.009781396016478539, 'timing/model': 1.1260680089471862, 'learning_rate/qwen_vl_interface': 3.892655980922847e-06, 'learning_rate/action_model': 3.8926559809228466e-05, 'epoch': 1.41}) +[2026-06-19 16:40:55,539][starVLA.training.train_starvla][INFO] - Step 3003, Loss: {'train/loss': 0.29100242257118225, 'train/grad_norm_pre_clip': 13.117074966430664, 'timing/data': 0.00992104597389698, 'timing/model': 1.294329168042168, 'learning_rate/qwen_vl_interface': 3.889737825962368e-06, 'learning_rate/action_model': 3.889737825962368e-05, 'epoch': 1.41}) +[2026-06-19 16:40:59,935][starVLA.training.train_starvla][INFO] - Step 3004, Loss: {'train/loss': 0.22905951738357544, 'train/grad_norm_pre_clip': 10.72550106048584, 'timing/data': 0.00928548292722553, 'timing/model': 1.1412617929745466, 'learning_rate/qwen_vl_interface': 3.886820230154288e-06, 'learning_rate/action_model': 3.886820230154287e-05, 'epoch': 1.41}) +[2026-06-19 16:41:04,475][starVLA.training.train_starvla][INFO] - Step 3005, Loss: {'train/loss': 0.2033025026321411, 'train/grad_norm_pre_clip': 9.668158531188965, 'timing/data': 0.009468839038163424, 'timing/model': 1.1365148110780865, 'learning_rate/qwen_vl_interface': 3.8839031946979184e-06, 'learning_rate/action_model': 3.883903194697918e-05, 'epoch': 1.41}) +[2026-06-19 16:41:08,960][starVLA.training.train_starvla][INFO] - Step 3006, Loss: {'train/loss': 0.23626649379730225, 'train/grad_norm_pre_clip': 8.382911682128906, 'timing/data': 0.009993602987378836, 'timing/model': 1.13139992300421, 'learning_rate/qwen_vl_interface': 3.880986720792344e-06, 'learning_rate/action_model': 3.880986720792344e-05, 'epoch': 1.41}) +[2026-06-19 16:41:13,519][starVLA.training.train_starvla][INFO] - Step 3007, Loss: {'train/loss': 0.18998762965202332, 'train/grad_norm_pre_clip': 12.150944709777832, 'timing/data': 0.013868678943254054, 'timing/model': 1.1485187389189377, 'learning_rate/qwen_vl_interface': 3.878070809636416e-06, 'learning_rate/action_model': 3.878070809636416e-05, 'epoch': 1.41}) +[2026-06-19 16:41:18,087][starVLA.training.train_starvla][INFO] - Step 3008, Loss: {'train/loss': 0.27128273248672485, 'train/grad_norm_pre_clip': 7.88251256942749, 'timing/data': 0.009859757032245398, 'timing/model': 1.1394873610697687, 'learning_rate/qwen_vl_interface': 3.875155462428758e-06, 'learning_rate/action_model': 3.875155462428758e-05, 'epoch': 1.41}) +[2026-06-19 16:41:22,656][starVLA.training.train_starvla][INFO] - Step 3009, Loss: {'train/loss': 0.4961784780025482, 'train/grad_norm_pre_clip': 25.841665267944336, 'timing/data': 0.009802578948438168, 'timing/model': 1.3084024609997869, 'learning_rate/qwen_vl_interface': 3.872240680367754e-06, 'learning_rate/action_model': 3.872240680367754e-05, 'epoch': 1.42}) +[2026-06-19 16:41:27,046][starVLA.training.train_starvla][INFO] - Step 3010, Loss: {'train/loss': 0.12878333032131195, 'train/grad_norm_pre_clip': 11.166781425476074, 'timing/data': 0.009676915011368692, 'timing/model': 1.1393133630044758, 'learning_rate/qwen_vl_interface': 3.869326464651566e-06, 'learning_rate/action_model': 3.869326464651566e-05, 'epoch': 1.42}) +[2026-06-19 16:41:31,653][starVLA.training.train_starvla][INFO] - Step 3011, Loss: {'train/loss': 0.1542520821094513, 'train/grad_norm_pre_clip': 12.125816345214844, 'timing/data': 0.009755093953572214, 'timing/model': 1.1559654789743945, 'learning_rate/qwen_vl_interface': 3.866412816478114e-06, 'learning_rate/action_model': 3.8664128164781144e-05, 'epoch': 1.42}) +[2026-06-19 16:41:36,230][starVLA.training.train_starvla][INFO] - Step 3012, Loss: {'train/loss': 0.2112990766763687, 'train/grad_norm_pre_clip': 11.659802436828613, 'timing/data': 0.009873892064206302, 'timing/model': 1.1405769849661738, 'learning_rate/qwen_vl_interface': 3.8634997370450915e-06, 'learning_rate/action_model': 3.8634997370450914e-05, 'epoch': 1.42}) +[2026-06-19 16:41:40,760][starVLA.training.train_starvla][INFO] - Step 3013, Loss: {'train/loss': 0.10618151724338531, 'train/grad_norm_pre_clip': 3.8562405109405518, 'timing/data': 0.010165983927436173, 'timing/model': 1.1506691020913422, 'learning_rate/qwen_vl_interface': 3.860587227549956e-06, 'learning_rate/action_model': 3.860587227549955e-05, 'epoch': 1.42}) +[2026-06-19 16:41:45,346][starVLA.training.train_starvla][INFO] - Step 3014, Loss: {'train/loss': 0.21456198394298553, 'train/grad_norm_pre_clip': 8.861294746398926, 'timing/data': 0.009337036986835301, 'timing/model': 1.1460744210053235, 'learning_rate/qwen_vl_interface': 3.857675289189926e-06, 'learning_rate/action_model': 3.857675289189926e-05, 'epoch': 1.42}) +[2026-06-19 16:41:49,935][starVLA.training.train_starvla][INFO] - Step 3015, Loss: {'train/loss': 0.19825716316699982, 'train/grad_norm_pre_clip': 8.844585418701172, 'timing/data': 0.009693910018540919, 'timing/model': 1.3207990740193054, 'learning_rate/qwen_vl_interface': 3.854763923161993e-06, 'learning_rate/action_model': 3.854763923161992e-05, 'epoch': 1.42}) +[2026-06-19 16:41:54,488][starVLA.training.train_starvla][INFO] - Step 3016, Loss: {'train/loss': 0.2740459740161896, 'train/grad_norm_pre_clip': 13.698126792907715, 'timing/data': 0.009826982975937426, 'timing/model': 1.2709104380337521, 'learning_rate/qwen_vl_interface': 3.851853130662908e-06, 'learning_rate/action_model': 3.8518531306629077e-05, 'epoch': 1.42}) +[2026-06-19 16:41:58,957][starVLA.training.train_starvla][INFO] - Step 3017, Loss: {'train/loss': 0.13730967044830322, 'train/grad_norm_pre_clip': 8.744086265563965, 'timing/data': 0.015945435035973787, 'timing/model': 1.1453210600884631, 'learning_rate/qwen_vl_interface': 3.8489429128891884e-06, 'learning_rate/action_model': 3.848942912889188e-05, 'epoch': 1.42}) +[2026-06-19 16:42:03,533][starVLA.training.train_starvla][INFO] - Step 3018, Loss: {'train/loss': 0.33047986030578613, 'train/grad_norm_pre_clip': 18.78799819946289, 'timing/data': 0.009568443056195974, 'timing/model': 1.1470949520589784, 'learning_rate/qwen_vl_interface': 3.846033271037114e-06, 'learning_rate/action_model': 3.8460332710371136e-05, 'epoch': 1.42}) +[2026-06-19 16:42:08,109][starVLA.training.train_starvla][INFO] - Step 3019, Loss: {'train/loss': 0.16179679334163666, 'train/grad_norm_pre_clip': 9.442244529724121, 'timing/data': 0.011360625037923455, 'timing/model': 1.1420866950647905, 'learning_rate/qwen_vl_interface': 3.843124206302732e-06, 'learning_rate/action_model': 3.8431242063027316e-05, 'epoch': 1.42}) +[2026-06-19 16:42:12,618][starVLA.training.train_starvla][INFO] - Step 3020, Loss: {'train/loss': 0.23062600195407867, 'train/grad_norm_pre_clip': 15.13766098022461, 'timing/data': 0.009744100039824843, 'timing/model': 1.1407442759955302, 'learning_rate/qwen_vl_interface': 3.840215719881844e-06, 'learning_rate/action_model': 3.8402157198818445e-05, 'epoch': 1.42}) +[2026-06-19 16:42:17,199][starVLA.training.train_starvla][INFO] - Step 3021, Loss: {'train/loss': 0.08094137907028198, 'train/grad_norm_pre_clip': 3.877617835998535, 'timing/data': 0.00954784406349063, 'timing/model': 1.144712200970389, 'learning_rate/qwen_vl_interface': 3.837307812970022e-06, 'learning_rate/action_model': 3.837307812970022e-05, 'epoch': 1.42}) +[2026-06-19 16:42:21,789][starVLA.training.train_starvla][INFO] - Step 3022, Loss: {'train/loss': 0.27652230858802795, 'train/grad_norm_pre_clip': 10.48985481262207, 'timing/data': 0.009807055001147091, 'timing/model': 1.320908435038291, 'learning_rate/qwen_vl_interface': 3.834400486762596e-06, 'learning_rate/action_model': 3.834400486762596e-05, 'epoch': 1.42}) +[2026-06-19 16:42:26,212][starVLA.training.train_starvla][INFO] - Step 3023, Loss: {'train/loss': 0.2779209017753601, 'train/grad_norm_pre_clip': 7.525107383728027, 'timing/data': 0.009440263034775853, 'timing/model': 1.141467777080834, 'learning_rate/qwen_vl_interface': 3.8314937424546585e-06, 'learning_rate/action_model': 3.8314937424546584e-05, 'epoch': 1.42}) +[2026-06-19 16:42:30,798][starVLA.training.train_starvla][INFO] - Step 3024, Loss: {'train/loss': 0.24986803531646729, 'train/grad_norm_pre_clip': 20.346759796142578, 'timing/data': 0.009904702077619731, 'timing/model': 1.1522750520380214, 'learning_rate/qwen_vl_interface': 3.828587581241064e-06, 'learning_rate/action_model': 3.8285875812410635e-05, 'epoch': 1.42}) +[2026-06-19 16:42:35,329][starVLA.training.train_starvla][INFO] - Step 3025, Loss: {'train/loss': 0.2738227844238281, 'train/grad_norm_pre_clip': 7.370785236358643, 'timing/data': 0.00974353903438896, 'timing/model': 1.1476636469596997, 'learning_rate/qwen_vl_interface': 3.825682004316419e-06, 'learning_rate/action_model': 3.825682004316419e-05, 'epoch': 1.42}) +[2026-06-19 16:42:39,865][starVLA.training.train_starvla][INFO] - Step 3026, Loss: {'train/loss': 0.15424507856369019, 'train/grad_norm_pre_clip': 7.217535972595215, 'timing/data': 0.009870193898677826, 'timing/model': 1.1528435490326956, 'learning_rate/qwen_vl_interface': 3.822777012875104e-06, 'learning_rate/action_model': 3.822777012875103e-05, 'epoch': 1.42}) +[2026-06-19 16:42:44,578][starVLA.training.train_starvla][INFO] - Step 3027, Loss: {'train/loss': 0.19590473175048828, 'train/grad_norm_pre_clip': 10.417632102966309, 'timing/data': 0.010153315961360931, 'timing/model': 1.2687596150208265, 'learning_rate/qwen_vl_interface': 3.819872608111248e-06, 'learning_rate/action_model': 3.819872608111248e-05, 'epoch': 1.42}) +[2026-06-19 16:42:49,210][starVLA.training.train_starvla][INFO] - Step 3028, Loss: {'train/loss': 0.24244199693202972, 'train/grad_norm_pre_clip': 9.960137367248535, 'timing/data': 0.00970400299411267, 'timing/model': 1.339984836988151, 'learning_rate/qwen_vl_interface': 3.816968791218741e-06, 'learning_rate/action_model': 3.816968791218741e-05, 'epoch': 1.42}) +[2026-06-19 16:42:53,647][starVLA.training.train_starvla][INFO] - Step 3029, Loss: {'train/loss': 0.1035708636045456, 'train/grad_norm_pre_clip': 7.037193775177002, 'timing/data': 0.013895094045437872, 'timing/model': 1.1472076709615067, 'learning_rate/qwen_vl_interface': 3.8140655633912337e-06, 'learning_rate/action_model': 3.814065563391234e-05, 'epoch': 1.42}) +[2026-06-19 16:42:58,279][starVLA.training.train_starvla][INFO] - Step 3030, Loss: {'train/loss': 0.5008069276809692, 'train/grad_norm_pre_clip': 27.477821350097656, 'timing/data': 0.009346473962068558, 'timing/model': 1.1605771119939163, 'learning_rate/qwen_vl_interface': 3.811162925822133e-06, 'learning_rate/action_model': 3.811162925822132e-05, 'epoch': 1.43}) +[2026-06-19 16:43:02,872][starVLA.training.train_starvla][INFO] - Step 3031, Loss: {'train/loss': 0.47783002257347107, 'train/grad_norm_pre_clip': 55.87478256225586, 'timing/data': 0.009664529003202915, 'timing/model': 1.1511188040021807, 'learning_rate/qwen_vl_interface': 3.8082608797046045e-06, 'learning_rate/action_model': 3.8082608797046045e-05, 'epoch': 1.43}) +[2026-06-19 16:43:07,397][starVLA.training.train_starvla][INFO] - Step 3032, Loss: {'train/loss': 0.1997481733560562, 'train/grad_norm_pre_clip': 7.2241315841674805, 'timing/data': 0.009866646956652403, 'timing/model': 1.1422068390529603, 'learning_rate/qwen_vl_interface': 3.8053594262315687e-06, 'learning_rate/action_model': 3.805359426231568e-05, 'epoch': 1.43}) +[2026-06-19 16:43:12,013][starVLA.training.train_starvla][INFO] - Step 3033, Loss: {'train/loss': 0.3330005407333374, 'train/grad_norm_pre_clip': 12.567479133605957, 'timing/data': 0.009777494939044118, 'timing/model': 1.1517975539900362, 'learning_rate/qwen_vl_interface': 3.802458566595704e-06, 'learning_rate/action_model': 3.802458566595704e-05, 'epoch': 1.43}) +[2026-06-19 16:43:16,662][starVLA.training.train_starvla][INFO] - Step 3034, Loss: {'train/loss': 0.18012148141860962, 'train/grad_norm_pre_clip': 23.378366470336914, 'timing/data': 0.009710631100460887, 'timing/model': 1.3252145029837266, 'learning_rate/qwen_vl_interface': 3.7995583019894444e-06, 'learning_rate/action_model': 3.7995583019894445e-05, 'epoch': 1.43}) +[2026-06-19 16:43:21,116][starVLA.training.train_starvla][INFO] - Step 3035, Loss: {'train/loss': 0.1361154466867447, 'train/grad_norm_pre_clip': 9.457663536071777, 'timing/data': 0.00991172797512263, 'timing/model': 1.1584382490254939, 'learning_rate/qwen_vl_interface': 3.7966586336049782e-06, 'learning_rate/action_model': 3.796658633604978e-05, 'epoch': 1.43}) +[2026-06-19 16:43:25,753][starVLA.training.train_starvla][INFO] - Step 3036, Loss: {'train/loss': 0.25610119104385376, 'train/grad_norm_pre_clip': 11.36512565612793, 'timing/data': 0.009817266021855175, 'timing/model': 1.1601409770082682, 'learning_rate/qwen_vl_interface': 3.7937595626342516e-06, 'learning_rate/action_model': 3.793759562634252e-05, 'epoch': 1.43}) +[2026-06-19 16:43:30,370][starVLA.training.train_starvla][INFO] - Step 3037, Loss: {'train/loss': 0.22455580532550812, 'train/grad_norm_pre_clip': 22.30622100830078, 'timing/data': 0.009204604080878198, 'timing/model': 1.1582803839119151, 'learning_rate/qwen_vl_interface': 3.79086109026896e-06, 'learning_rate/action_model': 3.7908610902689604e-05, 'epoch': 1.43}) +[2026-06-19 16:43:34,974][starVLA.training.train_starvla][INFO] - Step 3038, Loss: {'train/loss': 0.2800193130970001, 'train/grad_norm_pre_clip': 14.743844985961914, 'timing/data': 0.010057209990918636, 'timing/model': 1.1465620009694248, 'learning_rate/qwen_vl_interface': 3.7879632177005584e-06, 'learning_rate/action_model': 3.787963217700558e-05, 'epoch': 1.43}) +[2026-06-19 16:43:39,573][starVLA.training.train_starvla][INFO] - Step 3039, Loss: {'train/loss': 0.1766320765018463, 'train/grad_norm_pre_clip': 7.9034013748168945, 'timing/data': 0.010082231019623578, 'timing/model': 1.1548641220433637, 'learning_rate/qwen_vl_interface': 3.7850659461202523e-06, 'learning_rate/action_model': 3.785065946120252e-05, 'epoch': 1.43}) +[2026-06-19 16:43:44,194][starVLA.training.train_starvla][INFO] - Step 3040, Loss: {'train/loss': 0.06591010093688965, 'train/grad_norm_pre_clip': 5.90493631362915, 'timing/data': 0.009699657908640802, 'timing/model': 1.3363195919664577, 'learning_rate/qwen_vl_interface': 3.782169276719e-06, 'learning_rate/action_model': 3.782169276719e-05, 'epoch': 1.43}) +[2026-06-19 16:43:48,616][starVLA.training.train_starvla][INFO] - Step 3041, Loss: {'train/loss': 0.1891978234052658, 'train/grad_norm_pre_clip': 9.651375770568848, 'timing/data': 0.009932660963386297, 'timing/model': 1.1452371339546517, 'learning_rate/qwen_vl_interface': 3.779273210687514e-06, 'learning_rate/action_model': 3.779273210687514e-05, 'epoch': 1.43}) +[2026-06-19 16:43:53,204][starVLA.training.train_starvla][INFO] - Step 3042, Loss: {'train/loss': 0.10797619074583054, 'train/grad_norm_pre_clip': 5.546481132507324, 'timing/data': 0.009694181964732707, 'timing/model': 1.1448948390316218, 'learning_rate/qwen_vl_interface': 3.776377749216257e-06, 'learning_rate/action_model': 3.776377749216257e-05, 'epoch': 1.43}) +[2026-06-19 16:43:57,710][starVLA.training.train_starvla][INFO] - Step 3043, Loss: {'train/loss': 0.14714547991752625, 'train/grad_norm_pre_clip': 11.18184757232666, 'timing/data': 0.009987650904804468, 'timing/model': 1.1516810189932585, 'learning_rate/qwen_vl_interface': 3.773482893495442e-06, 'learning_rate/action_model': 3.773482893495442e-05, 'epoch': 1.43}) +[2026-06-19 16:44:02,258][starVLA.training.train_starvla][INFO] - Step 3044, Loss: {'train/loss': 0.18207554519176483, 'train/grad_norm_pre_clip': 6.438564300537109, 'timing/data': 0.00993316201493144, 'timing/model': 1.1456803439650685, 'learning_rate/qwen_vl_interface': 3.7705886447150375e-06, 'learning_rate/action_model': 3.7705886447150376e-05, 'epoch': 1.43}) +[2026-06-19 16:44:06,851][starVLA.training.train_starvla][INFO] - Step 3045, Loss: {'train/loss': 0.21221186220645905, 'train/grad_norm_pre_clip': 14.686775207519531, 'timing/data': 0.010573008912615478, 'timing/model': 1.1469024390680715, 'learning_rate/qwen_vl_interface': 3.767695004064759e-06, 'learning_rate/action_model': 3.767695004064759e-05, 'epoch': 1.43}) +[2026-06-19 16:44:11,459][starVLA.training.train_starvla][INFO] - Step 3046, Loss: {'train/loss': 0.14260050654411316, 'train/grad_norm_pre_clip': 13.300567626953125, 'timing/data': 0.010490057989954948, 'timing/model': 1.3203874779865146, 'learning_rate/qwen_vl_interface': 3.7648019727340725e-06, 'learning_rate/action_model': 3.764801972734072e-05, 'epoch': 1.43}) +[2026-06-19 16:44:15,998][starVLA.training.train_starvla][INFO] - Step 3047, Loss: {'train/loss': 0.3415021598339081, 'train/grad_norm_pre_clip': 11.29234790802002, 'timing/data': 0.009490690077655017, 'timing/model': 1.2645087540149689, 'learning_rate/qwen_vl_interface': 3.761909551912196e-06, 'learning_rate/action_model': 3.761909551912196e-05, 'epoch': 1.43}) +[2026-06-19 16:44:20,419][starVLA.training.train_starvla][INFO] - Step 3048, Loss: {'train/loss': 0.24042141437530518, 'train/grad_norm_pre_clip': 8.916370391845703, 'timing/data': 0.01075159094762057, 'timing/model': 1.149605497950688, 'learning_rate/qwen_vl_interface': 3.75901774278809e-06, 'learning_rate/action_model': 3.7590177427880904e-05, 'epoch': 1.43}) +[2026-06-19 16:44:25,012][starVLA.training.train_starvla][INFO] - Step 3049, Loss: {'train/loss': 0.5110536813735962, 'train/grad_norm_pre_clip': 25.897733688354492, 'timing/data': 0.01074038993101567, 'timing/model': 1.1531995499972254, 'learning_rate/qwen_vl_interface': 3.756126546550472e-06, 'learning_rate/action_model': 3.756126546550472e-05, 'epoch': 1.43}) +[2026-06-19 16:44:29,613][starVLA.training.train_starvla][INFO] - Step 3050, Loss: {'train/loss': 0.10955079644918442, 'train/grad_norm_pre_clip': 8.84241771697998, 'timing/data': 0.009594602975994349, 'timing/model': 1.1461870960192755, 'learning_rate/qwen_vl_interface': 3.753235964387802e-06, 'learning_rate/action_model': 3.753235964387802e-05, 'epoch': 1.43}) +[2026-06-19 16:44:34,193][starVLA.training.train_starvla][INFO] - Step 3051, Loss: {'train/loss': 0.046899426728487015, 'train/grad_norm_pre_clip': 2.837329626083374, 'timing/data': 0.009605580940842628, 'timing/model': 1.1481684349710122, 'learning_rate/qwen_vl_interface': 3.750345997488289e-06, 'learning_rate/action_model': 3.7503459974882886e-05, 'epoch': 1.43}) +[2026-06-19 16:44:38,715][starVLA.training.train_starvla][INFO] - Step 3052, Loss: {'train/loss': 0.35856249928474426, 'train/grad_norm_pre_clip': 15.86587142944336, 'timing/data': 0.010137101984582841, 'timing/model': 1.1459994439501315, 'learning_rate/qwen_vl_interface': 3.7474566470398903e-06, 'learning_rate/action_model': 3.74745664703989e-05, 'epoch': 1.44}) +[2026-06-19 16:44:43,227][starVLA.training.train_starvla][INFO] - Step 3053, Loss: {'train/loss': 0.19670622050762177, 'train/grad_norm_pre_clip': 10.813837051391602, 'timing/data': 0.009684644988738, 'timing/model': 1.2369053669972345, 'learning_rate/qwen_vl_interface': 3.744567914230307e-06, 'learning_rate/action_model': 3.744567914230307e-05, 'epoch': 1.44}) +[2026-06-19 16:44:47,662][starVLA.training.train_starvla][INFO] - Step 3054, Loss: {'train/loss': 0.30828219652175903, 'train/grad_norm_pre_clip': 12.767284393310547, 'timing/data': 0.00954499002546072, 'timing/model': 1.1525833039777353, 'learning_rate/qwen_vl_interface': 3.7416798002469902e-06, 'learning_rate/action_model': 3.74167980024699e-05, 'epoch': 1.44}) +[2026-06-19 16:44:52,248][starVLA.training.train_starvla][INFO] - Step 3055, Loss: {'train/loss': 0.13652992248535156, 'train/grad_norm_pre_clip': 14.277088165283203, 'timing/data': 0.009909172891639173, 'timing/model': 1.1467521829763427, 'learning_rate/qwen_vl_interface': 3.7387923062771313e-06, 'learning_rate/action_model': 3.738792306277131e-05, 'epoch': 1.44}) +[2026-06-19 16:44:56,854][starVLA.training.train_starvla][INFO] - Step 3056, Loss: {'train/loss': 0.1858421266078949, 'train/grad_norm_pre_clip': 8.529378890991211, 'timing/data': 0.009494493948295712, 'timing/model': 1.1531166180502623, 'learning_rate/qwen_vl_interface': 3.7359054335076732e-06, 'learning_rate/action_model': 3.7359054335076734e-05, 'epoch': 1.44}) +[2026-06-19 16:45:01,432][starVLA.training.train_starvla][INFO] - Step 3057, Loss: {'train/loss': 0.1817779541015625, 'train/grad_norm_pre_clip': 17.387495040893555, 'timing/data': 0.009796125930733979, 'timing/model': 1.143549924949184, 'learning_rate/qwen_vl_interface': 3.7330191831253006e-06, 'learning_rate/action_model': 3.7330191831253e-05, 'epoch': 1.44}) +[2026-06-19 16:45:06,023][starVLA.training.train_starvla][INFO] - Step 3058, Loss: {'train/loss': 0.05853116139769554, 'train/grad_norm_pre_clip': 3.6436123847961426, 'timing/data': 0.009370411979034543, 'timing/model': 1.312219811952673, 'learning_rate/qwen_vl_interface': 3.7301335563164382e-06, 'learning_rate/action_model': 3.730133556316438e-05, 'epoch': 1.44}) +[2026-06-19 16:45:10,534][starVLA.training.train_starvla][INFO] - Step 3059, Loss: {'train/loss': 0.278902143239975, 'train/grad_norm_pre_clip': 14.527149200439453, 'timing/data': 0.01002417306881398, 'timing/model': 1.2344267540611327, 'learning_rate/qwen_vl_interface': 3.727248554267261e-06, 'learning_rate/action_model': 3.727248554267261e-05, 'epoch': 1.44}) +[2026-06-19 16:45:14,966][starVLA.training.train_starvla][INFO] - Step 3060, Loss: {'train/loss': 0.33116966485977173, 'train/grad_norm_pre_clip': 16.739151000976562, 'timing/data': 0.010113473981618881, 'timing/model': 1.151028026943095, 'learning_rate/qwen_vl_interface': 3.7243641781636838e-06, 'learning_rate/action_model': 3.7243641781636836e-05, 'epoch': 1.44}) +[2026-06-19 16:45:19,595][starVLA.training.train_starvla][INFO] - Step 3061, Loss: {'train/loss': 0.26426997780799866, 'train/grad_norm_pre_clip': 11.318304061889648, 'timing/data': 0.00965615501627326, 'timing/model': 1.1549061400583014, 'learning_rate/qwen_vl_interface': 3.721480429191364e-06, 'learning_rate/action_model': 3.721480429191364e-05, 'epoch': 1.44}) +[2026-06-19 16:45:24,200][starVLA.training.train_starvla][INFO] - Step 3062, Loss: {'train/loss': 0.1890108436346054, 'train/grad_norm_pre_clip': 14.020670890808105, 'timing/data': 0.009691074956208467, 'timing/model': 1.1542133550392464, 'learning_rate/qwen_vl_interface': 3.7185973085357045e-06, 'learning_rate/action_model': 3.718597308535705e-05, 'epoch': 1.44}) +[2026-06-19 16:45:28,818][starVLA.training.train_starvla][INFO] - Step 3063, Loss: {'train/loss': 0.025158407166600227, 'train/grad_norm_pre_clip': 2.087775230407715, 'timing/data': 0.009676225017756224, 'timing/model': 1.1602657568873838, 'learning_rate/qwen_vl_interface': 3.7157148173818443e-06, 'learning_rate/action_model': 3.715714817381844e-05, 'epoch': 1.44}) +[2026-06-19 16:45:33,415][starVLA.training.train_starvla][INFO] - Step 3064, Loss: {'train/loss': 0.1371522843837738, 'train/grad_norm_pre_clip': 7.728679180145264, 'timing/data': 0.009819729020819068, 'timing/model': 1.311678290949203, 'learning_rate/qwen_vl_interface': 3.7128329569146672e-06, 'learning_rate/action_model': 3.7128329569146666e-05, 'epoch': 1.44}) +[2026-06-19 16:45:37,850][starVLA.training.train_starvla][INFO] - Step 3065, Loss: {'train/loss': 0.12124515324831009, 'train/grad_norm_pre_clip': 21.488862991333008, 'timing/data': 0.0096478839404881, 'timing/model': 1.1480777249671519, 'learning_rate/qwen_vl_interface': 3.7099517283187977e-06, 'learning_rate/action_model': 3.7099517283187976e-05, 'epoch': 1.44}) +[2026-06-19 16:45:42,442][starVLA.training.train_starvla][INFO] - Step 3066, Loss: {'train/loss': 0.4494555592536926, 'train/grad_norm_pre_clip': 18.01586151123047, 'timing/data': 0.010282917064614594, 'timing/model': 1.1478561389958486, 'learning_rate/qwen_vl_interface': 3.7070711327786e-06, 'learning_rate/action_model': 3.7070711327785997e-05, 'epoch': 1.44}) +[2026-06-19 16:45:47,050][starVLA.training.train_starvla][INFO] - Step 3067, Loss: {'train/loss': 0.3758675456047058, 'train/grad_norm_pre_clip': 15.966438293457031, 'timing/data': 0.009483775007538497, 'timing/model': 1.1486854969989508, 'learning_rate/qwen_vl_interface': 3.704191171478179e-06, 'learning_rate/action_model': 3.704191171478179e-05, 'epoch': 1.44}) +[2026-06-19 16:45:51,675][starVLA.training.train_starvla][INFO] - Step 3068, Loss: {'train/loss': 0.4315829575061798, 'train/grad_norm_pre_clip': 44.18398666381836, 'timing/data': 0.010180309065617621, 'timing/model': 1.1573634260566905, 'learning_rate/qwen_vl_interface': 3.701311845601376e-06, 'learning_rate/action_model': 3.701311845601376e-05, 'epoch': 1.44}) +[2026-06-19 16:45:56,287][starVLA.training.train_starvla][INFO] - Step 3069, Loss: {'train/loss': 0.5727126002311707, 'train/grad_norm_pre_clip': 14.615073204040527, 'timing/data': 0.009686831966973841, 'timing/model': 1.144251647987403, 'learning_rate/qwen_vl_interface': 3.6984331563317744e-06, 'learning_rate/action_model': 3.698433156331774e-05, 'epoch': 1.44}) +[2026-06-19 16:46:00,877][starVLA.training.train_starvla][INFO] - Step 3070, Loss: {'train/loss': 0.13879208266735077, 'train/grad_norm_pre_clip': 12.45582389831543, 'timing/data': 0.009647981030866504, 'timing/model': 1.3096683389740065, 'learning_rate/qwen_vl_interface': 3.695555104852695e-06, 'learning_rate/action_model': 3.6955551048526946e-05, 'epoch': 1.44}) +[2026-06-19 16:46:05,435][starVLA.training.train_starvla][INFO] - Step 3071, Loss: {'train/loss': 0.33139026165008545, 'train/grad_norm_pre_clip': 15.282051086425781, 'timing/data': 0.010192496934905648, 'timing/model': 1.2575280470773578, 'learning_rate/qwen_vl_interface': 3.6926776923471948e-06, 'learning_rate/action_model': 3.6926776923471946e-05, 'epoch': 1.44}) +[2026-06-19 16:46:09,883][starVLA.training.train_starvla][INFO] - Step 3072, Loss: {'train/loss': 0.2595146894454956, 'train/grad_norm_pre_clip': 18.7735595703125, 'timing/data': 0.00993526098318398, 'timing/model': 1.144466912955977, 'learning_rate/qwen_vl_interface': 3.6898009199980715e-06, 'learning_rate/action_model': 3.6898009199980714e-05, 'epoch': 1.44}) +[2026-06-19 16:46:14,495][starVLA.training.train_starvla][INFO] - Step 3073, Loss: {'train/loss': 0.43875908851623535, 'train/grad_norm_pre_clip': 36.797996520996094, 'timing/data': 0.009933013003319502, 'timing/model': 1.1519238840555772, 'learning_rate/qwen_vl_interface': 3.686924788987855e-06, 'learning_rate/action_model': 3.686924788987855e-05, 'epoch': 1.45}) +[2026-06-19 16:46:19,076][starVLA.training.train_starvla][INFO] - Step 3074, Loss: {'train/loss': 0.2846599817276001, 'train/grad_norm_pre_clip': 12.547089576721191, 'timing/data': 0.010526527999900281, 'timing/model': 1.1481727720238268, 'learning_rate/qwen_vl_interface': 3.6840493004988155e-06, 'learning_rate/action_model': 3.684049300498815e-05, 'epoch': 1.45}) +[2026-06-19 16:46:23,664][starVLA.training.train_starvla][INFO] - Step 3075, Loss: {'train/loss': 0.149958997964859, 'train/grad_norm_pre_clip': 8.538960456848145, 'timing/data': 0.010537780006416142, 'timing/model': 1.1436383029213175, 'learning_rate/qwen_vl_interface': 3.681174455712958e-06, 'learning_rate/action_model': 3.681174455712958e-05, 'epoch': 1.45}) +[2026-06-19 16:46:28,301][starVLA.training.train_starvla][INFO] - Step 3076, Loss: {'train/loss': 0.09960633516311646, 'train/grad_norm_pre_clip': 9.634906768798828, 'timing/data': 0.009668165002949536, 'timing/model': 1.337394198984839, 'learning_rate/qwen_vl_interface': 3.6783002558120206e-06, 'learning_rate/action_model': 3.67830025581202e-05, 'epoch': 1.45}) +[2026-06-19 16:46:32,851][starVLA.training.train_starvla][INFO] - Step 3077, Loss: {'train/loss': 0.18294692039489746, 'train/grad_norm_pre_clip': 16.113725662231445, 'timing/data': 0.009834397002123296, 'timing/model': 1.2485954919829965, 'learning_rate/qwen_vl_interface': 3.6754267019774804e-06, 'learning_rate/action_model': 3.67542670197748e-05, 'epoch': 1.45}) +[2026-06-19 16:46:37,301][starVLA.training.train_starvla][INFO] - Step 3078, Loss: {'train/loss': 0.3290645182132721, 'train/grad_norm_pre_clip': 22.429292678833008, 'timing/data': 0.009858706034719944, 'timing/model': 1.1585748159559444, 'learning_rate/qwen_vl_interface': 3.6725537953905453e-06, 'learning_rate/action_model': 3.672553795390545e-05, 'epoch': 1.45}) +[2026-06-19 16:46:41,914][starVLA.training.train_starvla][INFO] - Step 3079, Loss: {'train/loss': 0.07464518398046494, 'train/grad_norm_pre_clip': 9.100605964660645, 'timing/data': 0.009471669094637036, 'timing/model': 1.152986790984869, 'learning_rate/qwen_vl_interface': 3.6696815372321603e-06, 'learning_rate/action_model': 3.6696815372321604e-05, 'epoch': 1.45}) +[2026-06-19 16:46:46,546][starVLA.training.train_starvla][INFO] - Step 3080, Loss: {'train/loss': 0.07785095274448395, 'train/grad_norm_pre_clip': 4.176981449127197, 'timing/data': 0.00990812701638788, 'timing/model': 1.1536519010551274, 'learning_rate/qwen_vl_interface': 3.6668099286830023e-06, 'learning_rate/action_model': 3.666809928683002e-05, 'epoch': 1.45}) +[2026-06-19 16:46:51,156][starVLA.training.train_starvla][INFO] - Step 3081, Loss: {'train/loss': 0.2886030673980713, 'train/grad_norm_pre_clip': 13.106829643249512, 'timing/data': 0.009803259978070855, 'timing/model': 1.1539620880503207, 'learning_rate/qwen_vl_interface': 3.66393897092348e-06, 'learning_rate/action_model': 3.66393897092348e-05, 'epoch': 1.45}) +[2026-06-19 16:46:55,762][starVLA.training.train_starvla][INFO] - Step 3082, Loss: {'train/loss': 0.32913440465927124, 'train/grad_norm_pre_clip': 36.920738220214844, 'timing/data': 0.009440738009288907, 'timing/model': 1.3172684849705547, 'learning_rate/qwen_vl_interface': 3.661068665133735e-06, 'learning_rate/action_model': 3.661068665133735e-05, 'epoch': 1.45}) +[2026-06-19 16:47:00,293][starVLA.training.train_starvla][INFO] - Step 3083, Loss: {'train/loss': 0.21304628252983093, 'train/grad_norm_pre_clip': 15.3811674118042, 'timing/data': 0.009492062963545322, 'timing/model': 1.2529417080804706, 'learning_rate/qwen_vl_interface': 3.6581990124936434e-06, 'learning_rate/action_model': 3.6581990124936434e-05, 'epoch': 1.45}) +[2026-06-19 16:47:04,735][starVLA.training.train_starvla][INFO] - Step 3084, Loss: {'train/loss': 0.4046493470668793, 'train/grad_norm_pre_clip': 14.281055450439453, 'timing/data': 0.010111519019119442, 'timing/model': 1.1603281720308587, 'learning_rate/qwen_vl_interface': 3.65533001418281e-06, 'learning_rate/action_model': 3.655330014182809e-05, 'epoch': 1.45}) +[2026-06-19 16:47:09,344][starVLA.training.train_starvla][INFO] - Step 3085, Loss: {'train/loss': 0.15555153787136078, 'train/grad_norm_pre_clip': 12.312335968017578, 'timing/data': 0.00922669901046902, 'timing/model': 1.1575636129127815, 'learning_rate/qwen_vl_interface': 3.652461671380572e-06, 'learning_rate/action_model': 3.652461671380572e-05, 'epoch': 1.45}) +[2026-06-19 16:47:13,964][starVLA.training.train_starvla][INFO] - Step 3086, Loss: {'train/loss': 0.2848278284072876, 'train/grad_norm_pre_clip': 9.924808502197266, 'timing/data': 0.009683353011496365, 'timing/model': 1.1507415099767968, 'learning_rate/qwen_vl_interface': 3.6495939852659966e-06, 'learning_rate/action_model': 3.649593985265997e-05, 'epoch': 1.45}) +[2026-06-19 16:47:18,587][starVLA.training.train_starvla][INFO] - Step 3087, Loss: {'train/loss': 0.33262258768081665, 'train/grad_norm_pre_clip': 19.729110717773438, 'timing/data': 0.00988644000608474, 'timing/model': 1.1474190929438919, 'learning_rate/qwen_vl_interface': 3.646726957017881e-06, 'learning_rate/action_model': 3.646726957017881e-05, 'epoch': 1.45}) +[2026-06-19 16:47:23,193][starVLA.training.train_starvla][INFO] - Step 3088, Loss: {'train/loss': 0.19966177642345428, 'train/grad_norm_pre_clip': 7.166800022125244, 'timing/data': 0.00938316504471004, 'timing/model': 1.1524526359280571, 'learning_rate/qwen_vl_interface': 3.643860587814752e-06, 'learning_rate/action_model': 3.643860587814752e-05, 'epoch': 1.45}) +[2026-06-19 16:47:27,832][starVLA.training.train_starvla][INFO] - Step 3089, Loss: {'train/loss': 0.35218125581741333, 'train/grad_norm_pre_clip': 15.519508361816406, 'timing/data': 0.010388342081569135, 'timing/model': 1.3407035149866715, 'learning_rate/qwen_vl_interface': 3.640994878834866e-06, 'learning_rate/action_model': 3.640994878834866e-05, 'epoch': 1.45}) +[2026-06-19 16:47:32,293][starVLA.training.train_starvla][INFO] - Step 3090, Loss: {'train/loss': 0.23101548850536346, 'train/grad_norm_pre_clip': 8.264638900756836, 'timing/data': 0.009598560049198568, 'timing/model': 1.1601425060071051, 'learning_rate/qwen_vl_interface': 3.6381298312562095e-06, 'learning_rate/action_model': 3.6381298312562094e-05, 'epoch': 1.45}) +[2026-06-19 16:47:36,928][starVLA.training.train_starvla][INFO] - Step 3091, Loss: {'train/loss': 0.16456589102745056, 'train/grad_norm_pre_clip': 7.990007400512695, 'timing/data': 0.009904296952299774, 'timing/model': 1.162100275978446, 'learning_rate/qwen_vl_interface': 3.635265446256493e-06, 'learning_rate/action_model': 3.6352654462564926e-05, 'epoch': 1.45}) +[2026-06-19 16:47:41,557][starVLA.training.train_starvla][INFO] - Step 3092, Loss: {'train/loss': 0.09294183552265167, 'train/grad_norm_pre_clip': 6.894495964050293, 'timing/data': 0.009895409923046827, 'timing/model': 1.151707118959166, 'learning_rate/qwen_vl_interface': 3.6324017250131565e-06, 'learning_rate/action_model': 3.6324017250131567e-05, 'epoch': 1.45}) +[2026-06-19 16:47:46,185][starVLA.training.train_starvla][INFO] - Step 3093, Loss: {'train/loss': 0.10736862570047379, 'train/grad_norm_pre_clip': 15.041441917419434, 'timing/data': 0.012311773956753314, 'timing/model': 1.1649120890069753, 'learning_rate/qwen_vl_interface': 3.6295386687033696e-06, 'learning_rate/action_model': 3.6295386687033695e-05, 'epoch': 1.45}) +[2026-06-19 16:47:50,799][starVLA.training.train_starvla][INFO] - Step 3094, Loss: {'train/loss': 0.17820808291435242, 'train/grad_norm_pre_clip': 10.729849815368652, 'timing/data': 0.010047040064819157, 'timing/model': 1.1569169060094282, 'learning_rate/qwen_vl_interface': 3.6266762785040253e-06, 'learning_rate/action_model': 3.626676278504025e-05, 'epoch': 1.46}) +[2026-06-19 16:47:55,382][starVLA.training.train_starvla][INFO] - Step 3095, Loss: {'train/loss': 0.28695377707481384, 'train/grad_norm_pre_clip': 14.813088417053223, 'timing/data': 0.009973642998374999, 'timing/model': 1.3031901370268315, 'learning_rate/qwen_vl_interface': 3.6238145555917464e-06, 'learning_rate/action_model': 3.6238145555917465e-05, 'epoch': 1.46}) +[2026-06-19 16:47:59,806][starVLA.training.train_starvla][INFO] - Step 3096, Loss: {'train/loss': 0.2849246859550476, 'train/grad_norm_pre_clip': 21.806995391845703, 'timing/data': 0.009621507953852415, 'timing/model': 1.147038767929189, 'learning_rate/qwen_vl_interface': 3.6209535011428757e-06, 'learning_rate/action_model': 3.620953501142876e-05, 'epoch': 1.46}) +[2026-06-19 16:48:04,402][starVLA.training.train_starvla][INFO] - Step 3097, Loss: {'train/loss': 0.14281156659126282, 'train/grad_norm_pre_clip': 6.606430530548096, 'timing/data': 0.009727837052196264, 'timing/model': 1.15273070090916, 'learning_rate/qwen_vl_interface': 3.6180931163334863e-06, 'learning_rate/action_model': 3.6180931163334863e-05, 'epoch': 1.46}) +[2026-06-19 16:48:08,915][starVLA.training.train_starvla][INFO] - Step 3098, Loss: {'train/loss': 0.1033715009689331, 'train/grad_norm_pre_clip': 9.467251777648926, 'timing/data': 0.009787630056962371, 'timing/model': 1.1445720529882237, 'learning_rate/qwen_vl_interface': 3.6152334023393747e-06, 'learning_rate/action_model': 3.6152334023393745e-05, 'epoch': 1.46}) +[2026-06-19 16:48:13,442][starVLA.training.train_starvla][INFO] - Step 3099, Loss: {'train/loss': 0.25160980224609375, 'train/grad_norm_pre_clip': 19.672061920166016, 'timing/data': 0.009698182926513255, 'timing/model': 1.1466084469575435, 'learning_rate/qwen_vl_interface': 3.6123743603360613e-06, 'learning_rate/action_model': 3.612374360336061e-05, 'epoch': 1.46}) +[2026-06-19 16:48:18,032][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 3100 +[2026-06-19 16:51:13,122][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 3100 +[2026-06-19 16:51:13,123][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 3100 +[2026-06-19 16:54:21,588][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 3100 +[2026-06-19 16:54:21,589][starVLA.training.train_starvla][INFO] - Step 3100, Loss: {'train/loss': 0.3171979784965515, 'train/grad_norm_pre_clip': 10.38735580444336, 'eval/loss': 0.2105204463005066, 'eval/latency_0/loss': 0.09562450647354126, 'eval/latency_1/loss': 0.06147606670856476, 'eval/latency_2/loss': 0.12012942135334015, 'eval/latency_3/loss': 0.19092097878456116, 'eval/latency_4/loss': 0.5815677046775818, 'eval/flappy/loss': 0.2105204463005066, 'eval/flappy/latency_0/loss': 0.09562450647354126, 'eval/flappy/latency_1/loss': 0.06147606670856476, 'eval/flappy/latency_2/loss': 0.12012942135334015, 'eval/flappy/latency_3/loss': 0.19092097878456116, 'eval/flappy/latency_4/loss': 0.5815677046775818, 'eval/flap_precision': 0.7450199203187251, 'eval/flap_recall': 0.68, 'eval/flap_f1': 0.7110266159695818, 'eval/flap_rate_model': 0.06972222222222223, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.7110266159695818, 'eval/latency_0/flap_precision': 0.7450199203187251, 'eval/latency_0/flap_recall': 0.68, 'eval/latency_0/flap_f1': 0.7110266159695818, 'eval/latency_0/flap_rate_model': 0.06972222222222223, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.7110266159695818, 'timing/data': 0.009675943059846759, 'timing/model': 1.1457928529707715, 'learning_rate/qwen_vl_interface': 3.6095159914987893e-06, 'learning_rate/action_model': 3.6095159914987894e-05, 'epoch': 1.46}) +[2026-06-19 16:54:21,589][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_3100_state +[2026-06-19 16:54:21,592][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-19 16:54:24,978][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_3100_state/model.safetensors +[2026-06-19 16:54:32,142][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_3100_state/optimizer.bin +[2026-06-19 16:54:32,142][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_3100_state/sampler.bin +[2026-06-19 16:54:32,142][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_3100_state/sampler_1.bin +[2026-06-19 16:54:32,143][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_3100_state/random_states_0.pkl +[2026-06-19 16:54:32,167][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3100 +[2026-06-19 16:54:32,167][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-19 16:54:32,177][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-19 16:54:36,656][starVLA.training.train_starvla][INFO] - Step 3101, Loss: {'train/loss': 0.2868494689464569, 'train/grad_norm_pre_clip': 25.9406795501709, 'timing/data': 0.009578147088177502, 'timing/model': 1.1376715080114082, 'learning_rate/qwen_vl_interface': 3.6066582970025266e-06, 'learning_rate/action_model': 3.6066582970025264e-05, 'epoch': 1.46}) +[2026-06-19 16:54:41,230][starVLA.training.train_starvla][INFO] - Step 3102, Loss: {'train/loss': 0.21169203519821167, 'train/grad_norm_pre_clip': 11.050376892089844, 'timing/data': 0.009629713953472674, 'timing/model': 1.3190125069813803, 'learning_rate/qwen_vl_interface': 3.603801278021964e-06, 'learning_rate/action_model': 3.6038012780219635e-05, 'epoch': 1.46}) +[2026-06-19 16:54:45,611][starVLA.training.train_starvla][INFO] - Step 3103, Loss: {'train/loss': 0.23553408682346344, 'train/grad_norm_pre_clip': 11.259984970092773, 'timing/data': 0.009069789084605873, 'timing/model': 1.1378995629493147, 'learning_rate/qwen_vl_interface': 3.6009449357315122e-06, 'learning_rate/action_model': 3.6009449357315125e-05, 'epoch': 1.46}) +[2026-06-19 16:54:50,175][starVLA.training.train_starvla][INFO] - Step 3104, Loss: {'train/loss': 0.3133109211921692, 'train/grad_norm_pre_clip': 13.847782135009766, 'timing/data': 0.009550143964588642, 'timing/model': 1.1358552370220423, 'learning_rate/qwen_vl_interface': 3.5980892713053084e-06, 'learning_rate/action_model': 3.598089271305308e-05, 'epoch': 1.46}) +[2026-06-19 16:54:54,743][starVLA.training.train_starvla][INFO] - Step 3105, Loss: {'train/loss': 0.10318504273891449, 'train/grad_norm_pre_clip': 4.443203926086426, 'timing/data': 0.00914459303021431, 'timing/model': 1.1422241530381143, 'learning_rate/qwen_vl_interface': 3.5952342859172063e-06, 'learning_rate/action_model': 3.595234285917206e-05, 'epoch': 1.46}) +[2026-06-19 16:54:59,252][starVLA.training.train_starvla][INFO] - Step 3106, Loss: {'train/loss': 0.2580782473087311, 'train/grad_norm_pre_clip': 12.2919921875, 'timing/data': 0.008975517004728317, 'timing/model': 1.1644054800271988, 'learning_rate/qwen_vl_interface': 3.592379980740785e-06, 'learning_rate/action_model': 3.5923799807407845e-05, 'epoch': 1.46}) +[2026-06-19 16:55:03,765][starVLA.training.train_starvla][INFO] - Step 3107, Loss: {'train/loss': 0.28908780217170715, 'train/grad_norm_pre_clip': 9.306344985961914, 'timing/data': 0.00930732802953571, 'timing/model': 1.1343106291024014, 'learning_rate/qwen_vl_interface': 3.5895263569493384e-06, 'learning_rate/action_model': 3.589526356949338e-05, 'epoch': 1.46}) +[2026-06-19 16:55:08,336][starVLA.training.train_starvla][INFO] - Step 3108, Loss: {'train/loss': 0.16668377816677094, 'train/grad_norm_pre_clip': 10.316366195678711, 'timing/data': 0.009389778948388994, 'timing/model': 1.3176335649332032, 'learning_rate/qwen_vl_interface': 3.586673415715885e-06, 'learning_rate/action_model': 3.586673415715885e-05, 'epoch': 1.46}) +[2026-06-19 16:55:12,721][starVLA.training.train_starvla][INFO] - Step 3109, Loss: {'train/loss': 0.21958027780056, 'train/grad_norm_pre_clip': 9.079522132873535, 'timing/data': 0.009066733997315168, 'timing/model': 1.1418734439648688, 'learning_rate/qwen_vl_interface': 3.5838211582131603e-06, 'learning_rate/action_model': 3.58382115821316e-05, 'epoch': 1.46}) +[2026-06-19 16:55:17,284][starVLA.training.train_starvla][INFO] - Step 3110, Loss: {'train/loss': 0.3211539685726166, 'train/grad_norm_pre_clip': 27.170902252197266, 'timing/data': 0.009301952086389065, 'timing/model': 1.1359016990754753, 'learning_rate/qwen_vl_interface': 3.580969585613622e-06, 'learning_rate/action_model': 3.580969585613622e-05, 'epoch': 1.46}) +[2026-06-19 16:55:21,851][starVLA.training.train_starvla][INFO] - Step 3111, Loss: {'train/loss': 0.2949974238872528, 'train/grad_norm_pre_clip': 28.040143966674805, 'timing/data': 0.009293299983255565, 'timing/model': 1.1471942639909685, 'learning_rate/qwen_vl_interface': 3.5781186990894423e-06, 'learning_rate/action_model': 3.578118699089442e-05, 'epoch': 1.46}) +[2026-06-19 16:55:26,432][starVLA.training.train_starvla][INFO] - Step 3112, Loss: {'train/loss': 0.14904963970184326, 'train/grad_norm_pre_clip': 9.552773475646973, 'timing/data': 0.009299807017669082, 'timing/model': 1.1459590799640864, 'learning_rate/qwen_vl_interface': 3.5752684998125113e-06, 'learning_rate/action_model': 3.5752684998125116e-05, 'epoch': 1.46}) +[2026-06-19 16:55:30,893][starVLA.training.train_starvla][INFO] - Step 3113, Loss: {'train/loss': 0.1474565714597702, 'train/grad_norm_pre_clip': 20.14957618713379, 'timing/data': 0.00962346198502928, 'timing/model': 1.1394080499885604, 'learning_rate/qwen_vl_interface': 3.5724189889544407e-06, 'learning_rate/action_model': 3.572418988954441e-05, 'epoch': 1.46}) +[2026-06-19 16:55:35,442][starVLA.training.train_starvla][INFO] - Step 3114, Loss: {'train/loss': 0.3361976146697998, 'train/grad_norm_pre_clip': 19.706941604614258, 'timing/data': 0.010176110081374645, 'timing/model': 1.2709653639467433, 'learning_rate/qwen_vl_interface': 3.569570167686556e-06, 'learning_rate/action_model': 3.569570167686556e-05, 'epoch': 1.46}) +[2026-06-19 16:55:39,860][starVLA.training.train_starvla][INFO] - Step 3115, Loss: {'train/loss': 0.22672176361083984, 'train/grad_norm_pre_clip': 22.0454044342041, 'timing/data': 0.008931168005801737, 'timing/model': 1.1469220529543236, 'learning_rate/qwen_vl_interface': 3.5667220371798995e-06, 'learning_rate/action_model': 3.566722037179899e-05, 'epoch': 1.47}) +[2026-06-19 16:55:44,462][starVLA.training.train_starvla][INFO] - Step 3116, Loss: {'train/loss': 0.38175028562545776, 'train/grad_norm_pre_clip': 15.077491760253906, 'timing/data': 0.009197207051329315, 'timing/model': 1.1380511129973456, 'learning_rate/qwen_vl_interface': 3.563874598605232e-06, 'learning_rate/action_model': 3.5638745986052316e-05, 'epoch': 1.47}) +[2026-06-19 16:55:49,035][starVLA.training.train_starvla][INFO] - Step 3117, Loss: {'train/loss': 0.31555044651031494, 'train/grad_norm_pre_clip': 9.484393119812012, 'timing/data': 0.009167684009298682, 'timing/model': 1.1444856759626418, 'learning_rate/qwen_vl_interface': 3.561027853133025e-06, 'learning_rate/action_model': 3.561027853133025e-05, 'epoch': 1.47}) +[2026-06-19 16:55:53,634][starVLA.training.train_starvla][INFO] - Step 3118, Loss: {'train/loss': 0.16758501529693604, 'train/grad_norm_pre_clip': 9.148263931274414, 'timing/data': 0.009021493955515325, 'timing/model': 1.1556922440649942, 'learning_rate/qwen_vl_interface': 3.558181801933472e-06, 'learning_rate/action_model': 3.5581818019334714e-05, 'epoch': 1.47}) +[2026-06-19 16:55:58,241][starVLA.training.train_starvla][INFO] - Step 3119, Loss: {'train/loss': 0.25188398361206055, 'train/grad_norm_pre_clip': 23.295625686645508, 'timing/data': 0.00927342101931572, 'timing/model': 1.3089897969039157, 'learning_rate/qwen_vl_interface': 3.555336446176473e-06, 'learning_rate/action_model': 3.5553364461764726e-05, 'epoch': 1.47}) +[2026-06-19 16:56:02,766][starVLA.training.train_starvla][INFO] - Step 3120, Loss: {'train/loss': 0.16519039869308472, 'train/grad_norm_pre_clip': 7.179985523223877, 'timing/data': 0.009261116967536509, 'timing/model': 1.2611529339337721, 'learning_rate/qwen_vl_interface': 3.552491787031647e-06, 'learning_rate/action_model': 3.552491787031647e-05, 'epoch': 1.47}) +[2026-06-19 16:56:07,170][starVLA.training.train_starvla][INFO] - Step 3121, Loss: {'train/loss': 0.1401771456003189, 'train/grad_norm_pre_clip': 5.744772434234619, 'timing/data': 0.00873068894725293, 'timing/model': 1.1436015849467367, 'learning_rate/qwen_vl_interface': 3.5496478256683277e-06, 'learning_rate/action_model': 3.549647825668327e-05, 'epoch': 1.47}) +[2026-06-19 16:56:11,775][starVLA.training.train_starvla][INFO] - Step 3122, Loss: {'train/loss': 0.5300289988517761, 'train/grad_norm_pre_clip': 30.092267990112305, 'timing/data': 0.00948181701824069, 'timing/model': 1.1502652229974046, 'learning_rate/qwen_vl_interface': 3.5468045632555585e-06, 'learning_rate/action_model': 3.546804563255559e-05, 'epoch': 1.47}) +[2026-06-19 16:56:16,384][starVLA.training.train_starvla][INFO] - Step 3123, Loss: {'train/loss': 0.09378639608621597, 'train/grad_norm_pre_clip': 7.168843746185303, 'timing/data': 0.009312393027357757, 'timing/model': 1.1556925129843876, 'learning_rate/qwen_vl_interface': 3.5439620009620995e-06, 'learning_rate/action_model': 3.543962000962099e-05, 'epoch': 1.47}) +[2026-06-19 16:56:20,981][starVLA.training.train_starvla][INFO] - Step 3124, Loss: {'train/loss': 0.15852434933185577, 'train/grad_norm_pre_clip': 20.708906173706055, 'timing/data': 0.009024426923133433, 'timing/model': 1.1522914359811693, 'learning_rate/qwen_vl_interface': 3.541120139956419e-06, 'learning_rate/action_model': 3.5411201399564184e-05, 'epoch': 1.47}) +[2026-06-19 16:56:25,545][starVLA.training.train_starvla][INFO] - Step 3125, Loss: {'train/loss': 0.34903737902641296, 'train/grad_norm_pre_clip': 10.576640129089355, 'timing/data': 0.009249928989447653, 'timing/model': 1.1401112650055438, 'learning_rate/qwen_vl_interface': 3.538278981406697e-06, 'learning_rate/action_model': 3.5382789814066966e-05, 'epoch': 1.47}) +[2026-06-19 16:56:30,125][starVLA.training.train_starvla][INFO] - Step 3126, Loss: {'train/loss': 0.22159768640995026, 'train/grad_norm_pre_clip': 14.462157249450684, 'timing/data': 0.009861077065579593, 'timing/model': 1.3157513740006834, 'learning_rate/qwen_vl_interface': 3.5354385264808295e-06, 'learning_rate/action_model': 3.535438526480829e-05, 'epoch': 1.47}) +[2026-06-19 16:56:34,552][starVLA.training.train_starvla][INFO] - Step 3127, Loss: {'train/loss': 0.14856427907943726, 'train/grad_norm_pre_clip': 8.40478229522705, 'timing/data': 0.009503000997938216, 'timing/model': 1.1428139860508963, 'learning_rate/qwen_vl_interface': 3.5325987763464196e-06, 'learning_rate/action_model': 3.5325987763464194e-05, 'epoch': 1.47}) +[2026-06-19 16:56:39,142][starVLA.training.train_starvla][INFO] - Step 3128, Loss: {'train/loss': 0.2859864830970764, 'train/grad_norm_pre_clip': 9.467073440551758, 'timing/data': 0.009519777027890086, 'timing/model': 1.141655724029988, 'learning_rate/qwen_vl_interface': 3.529759732170781e-06, 'learning_rate/action_model': 3.529759732170781e-05, 'epoch': 1.47}) +[2026-06-19 16:56:43,755][starVLA.training.train_starvla][INFO] - Step 3129, Loss: {'train/loss': 0.16120754182338715, 'train/grad_norm_pre_clip': 6.172126770019531, 'timing/data': 0.009054096997715533, 'timing/model': 1.1471887649968266, 'learning_rate/qwen_vl_interface': 3.5269213951209375e-06, 'learning_rate/action_model': 3.5269213951209374e-05, 'epoch': 1.47}) +[2026-06-19 16:56:48,348][starVLA.training.train_starvla][INFO] - Step 3130, Loss: {'train/loss': 0.1336069107055664, 'train/grad_norm_pre_clip': 6.808899879455566, 'timing/data': 0.008973140036687255, 'timing/model': 1.1512424609391019, 'learning_rate/qwen_vl_interface': 3.5240837663636207e-06, 'learning_rate/action_model': 3.5240837663636204e-05, 'epoch': 1.47}) +[2026-06-19 16:56:52,947][starVLA.training.train_starvla][INFO] - Step 3131, Loss: {'train/loss': 0.3113057315349579, 'train/grad_norm_pre_clip': 14.186613082885742, 'timing/data': 0.009091255953535438, 'timing/model': 1.1621676990762353, 'learning_rate/qwen_vl_interface': 3.5212468470652745e-06, 'learning_rate/action_model': 3.521246847065274e-05, 'epoch': 1.47}) +[2026-06-19 16:56:57,544][starVLA.training.train_starvla][INFO] - Step 3132, Loss: {'train/loss': 0.22533510625362396, 'train/grad_norm_pre_clip': 7.154586315155029, 'timing/data': 0.009359937044791877, 'timing/model': 1.3259598859585822, 'learning_rate/qwen_vl_interface': 3.5184106383920487e-06, 'learning_rate/action_model': 3.5184106383920486e-05, 'epoch': 1.47}) +[2026-06-19 16:57:01,977][starVLA.training.train_starvla][INFO] - Step 3133, Loss: {'train/loss': 0.193983793258667, 'train/grad_norm_pre_clip': 8.138745307922363, 'timing/data': 0.009450955083593726, 'timing/model': 1.1579535289201885, 'learning_rate/qwen_vl_interface': 3.5155751415098022e-06, 'learning_rate/action_model': 3.515575141509802e-05, 'epoch': 1.47}) +[2026-06-19 16:57:06,586][starVLA.training.train_starvla][INFO] - Step 3134, Loss: {'train/loss': 0.11847684532403946, 'train/grad_norm_pre_clip': 10.532050132751465, 'timing/data': 0.009422000031918287, 'timing/model': 1.1636569029651582, 'learning_rate/qwen_vl_interface': 3.5127403575841e-06, 'learning_rate/action_model': 3.5127403575841e-05, 'epoch': 1.47}) +[2026-06-19 16:57:11,124][starVLA.training.train_starvla][INFO] - Step 3135, Loss: {'train/loss': 0.16369862854480743, 'train/grad_norm_pre_clip': 7.679811000823975, 'timing/data': 0.009178854059427977, 'timing/model': 1.1526439529843628, 'learning_rate/qwen_vl_interface': 3.509906287780215e-06, 'learning_rate/action_model': 3.5099062877802144e-05, 'epoch': 1.47}) +[2026-06-19 16:57:15,683][starVLA.training.train_starvla][INFO] - Step 3136, Loss: {'train/loss': 0.19633512198925018, 'train/grad_norm_pre_clip': 11.742952346801758, 'timing/data': 0.009676425019279122, 'timing/model': 1.1573113669874147, 'learning_rate/qwen_vl_interface': 3.5070729332631247e-06, 'learning_rate/action_model': 3.507072933263125e-05, 'epoch': 1.47}) +[2026-06-19 16:57:20,312][starVLA.training.train_starvla][INFO] - Step 3137, Loss: {'train/loss': 0.2065393626689911, 'train/grad_norm_pre_clip': 10.79685115814209, 'timing/data': 0.009317097952589393, 'timing/model': 1.1528678779723123, 'learning_rate/qwen_vl_interface': 3.5042402951975167e-06, 'learning_rate/action_model': 3.5042402951975167e-05, 'epoch': 1.48}) +[2026-06-19 16:57:24,926][starVLA.training.train_starvla][INFO] - Step 3138, Loss: {'train/loss': 0.09053733199834824, 'train/grad_norm_pre_clip': 3.917431116104126, 'timing/data': 0.009147954056970775, 'timing/model': 1.3310172779019922, 'learning_rate/qwen_vl_interface': 3.50140837474778e-06, 'learning_rate/action_model': 3.50140837474778e-05, 'epoch': 1.48}) +[2026-06-19 16:57:29,379][starVLA.training.train_starvla][INFO] - Step 3139, Loss: {'train/loss': 0.29032084345817566, 'train/grad_norm_pre_clip': 12.624600410461426, 'timing/data': 0.00934399594552815, 'timing/model': 1.1541203639935702, 'learning_rate/qwen_vl_interface': 3.498577173078011e-06, 'learning_rate/action_model': 3.498577173078011e-05, 'epoch': 1.48}) +[2026-06-19 16:57:34,005][starVLA.training.train_starvla][INFO] - Step 3140, Loss: {'train/loss': 0.23136326670646667, 'train/grad_norm_pre_clip': 8.130542755126953, 'timing/data': 0.009702912997454405, 'timing/model': 1.1593573240097612, 'learning_rate/qwen_vl_interface': 3.495746691352008e-06, 'learning_rate/action_model': 3.495746691352008e-05, 'epoch': 1.48}) +[2026-06-19 16:57:38,632][starVLA.training.train_starvla][INFO] - Step 3141, Loss: {'train/loss': 0.2519221901893616, 'train/grad_norm_pre_clip': 15.366716384887695, 'timing/data': 0.009411607054062188, 'timing/model': 1.156890779035166, 'learning_rate/qwen_vl_interface': 3.4929169307332763e-06, 'learning_rate/action_model': 3.492916930733276e-05, 'epoch': 1.48}) +[2026-06-19 16:57:43,226][starVLA.training.train_starvla][INFO] - Step 3142, Loss: {'train/loss': 0.10622802376747131, 'train/grad_norm_pre_clip': 6.321486473083496, 'timing/data': 0.009232828044332564, 'timing/model': 1.1492740589892492, 'learning_rate/qwen_vl_interface': 3.4900878923850234e-06, 'learning_rate/action_model': 3.490087892385023e-05, 'epoch': 1.48}) +[2026-06-19 16:57:47,822][starVLA.training.train_starvla][INFO] - Step 3143, Loss: {'train/loss': 0.5354152917861938, 'train/grad_norm_pre_clip': 30.804546356201172, 'timing/data': 0.009418040979653597, 'timing/model': 1.1438119310187176, 'learning_rate/qwen_vl_interface': 3.4872595774701607e-06, 'learning_rate/action_model': 3.487259577470161e-05, 'epoch': 1.48}) +[2026-06-19 16:57:52,460][starVLA.training.train_starvla][INFO] - Step 3144, Loss: {'train/loss': 0.5058197379112244, 'train/grad_norm_pre_clip': 19.736736297607422, 'timing/data': 0.009699853020720184, 'timing/model': 1.324056004988961, 'learning_rate/qwen_vl_interface': 3.4844319871513014e-06, 'learning_rate/action_model': 3.484431987151301e-05, 'epoch': 1.48}) +[2026-06-19 16:57:56,901][starVLA.training.train_starvla][INFO] - Step 3145, Loss: {'train/loss': 0.07861338555812836, 'train/grad_norm_pre_clip': 10.655096054077148, 'timing/data': 0.009271616931073368, 'timing/model': 1.159106255043298, 'learning_rate/qwen_vl_interface': 3.4816051225907606e-06, 'learning_rate/action_model': 3.48160512259076e-05, 'epoch': 1.48}) +[2026-06-19 16:58:01,526][starVLA.training.train_starvla][INFO] - Step 3146, Loss: {'train/loss': 0.14990179240703583, 'train/grad_norm_pre_clip': 4.966337203979492, 'timing/data': 0.009351413929834962, 'timing/model': 1.1585176350781694, 'learning_rate/qwen_vl_interface': 3.478778984950556e-06, 'learning_rate/action_model': 3.478778984950556e-05, 'epoch': 1.48}) +[2026-06-19 16:58:06,078][starVLA.training.train_starvla][INFO] - Step 3147, Loss: {'train/loss': 0.14617325365543365, 'train/grad_norm_pre_clip': 14.046930313110352, 'timing/data': 0.009458604967221618, 'timing/model': 1.1466637939447537, 'learning_rate/qwen_vl_interface': 3.475953575392405e-06, 'learning_rate/action_model': 3.475953575392405e-05, 'epoch': 1.48}) +[2026-06-19 16:58:10,646][starVLA.training.train_starvla][INFO] - Step 3148, Loss: {'train/loss': 0.10151275247335434, 'train/grad_norm_pre_clip': 8.927454948425293, 'timing/data': 0.009329703985713422, 'timing/model': 1.1464754029875621, 'learning_rate/qwen_vl_interface': 3.473128895077728e-06, 'learning_rate/action_model': 3.473128895077728e-05, 'epoch': 1.48}) +[2026-06-19 16:58:15,245][starVLA.training.train_starvla][INFO] - Step 3149, Loss: {'train/loss': 0.2015535533428192, 'train/grad_norm_pre_clip': 16.361583709716797, 'timing/data': 0.009583426057361066, 'timing/model': 1.1623990500811487, 'learning_rate/qwen_vl_interface': 3.4703049451676453e-06, 'learning_rate/action_model': 3.4703049451676454e-05, 'epoch': 1.48}) +[2026-06-19 16:58:19,837][starVLA.training.train_starvla][INFO] - Step 3150, Loss: {'train/loss': 0.12359223514795303, 'train/grad_norm_pre_clip': 7.741925239562988, 'timing/data': 0.009088388993404806, 'timing/model': 1.3119830508949235, 'learning_rate/qwen_vl_interface': 3.4674817268229732e-06, 'learning_rate/action_model': 3.467481726822973e-05, 'epoch': 1.48}) +[2026-06-19 16:58:24,405][starVLA.training.train_starvla][INFO] - Step 3151, Loss: {'train/loss': 0.20973575115203857, 'train/grad_norm_pre_clip': 16.977529525756836, 'timing/data': 0.009662055061198771, 'timing/model': 1.271638744045049, 'learning_rate/qwen_vl_interface': 3.464659241204232e-06, 'learning_rate/action_model': 3.4646592412042314e-05, 'epoch': 1.48}) +[2026-06-19 16:58:28,858][starVLA.training.train_starvla][INFO] - Step 3152, Loss: {'train/loss': 0.19076456129550934, 'train/grad_norm_pre_clip': 9.671133995056152, 'timing/data': 0.009173628990538418, 'timing/model': 1.166682409006171, 'learning_rate/qwen_vl_interface': 3.4618374894716377e-06, 'learning_rate/action_model': 3.461837489471638e-05, 'epoch': 1.48}) +[2026-06-19 16:58:33,462][starVLA.training.train_starvla][INFO] - Step 3153, Loss: {'train/loss': 0.3749810755252838, 'train/grad_norm_pre_clip': 12.499445915222168, 'timing/data': 0.008963044034317136, 'timing/model': 1.153000729973428, 'learning_rate/qwen_vl_interface': 3.459016472785107e-06, 'learning_rate/action_model': 3.459016472785107e-05, 'epoch': 1.48}) +[2026-06-19 16:58:38,083][starVLA.training.train_starvla][INFO] - Step 3154, Loss: {'train/loss': 0.15508146584033966, 'train/grad_norm_pre_clip': 16.460041046142578, 'timing/data': 0.009307164000347257, 'timing/model': 1.1526951510459185, 'learning_rate/qwen_vl_interface': 3.456196192304254e-06, 'learning_rate/action_model': 3.456196192304254e-05, 'epoch': 1.48}) +[2026-06-19 16:58:42,707][starVLA.training.train_starvla][INFO] - Step 3155, Loss: {'train/loss': 0.39262595772743225, 'train/grad_norm_pre_clip': 20.224424362182617, 'timing/data': 0.008999474113807082, 'timing/model': 1.1545368760125712, 'learning_rate/qwen_vl_interface': 3.4533766491883887e-06, 'learning_rate/action_model': 3.453376649188388e-05, 'epoch': 1.48}) +[2026-06-19 16:58:47,212][starVLA.training.train_starvla][INFO] - Step 3156, Loss: {'train/loss': 0.2427343726158142, 'train/grad_norm_pre_clip': 11.517274856567383, 'timing/data': 0.00949839607346803, 'timing/model': 1.1472108969464898, 'learning_rate/qwen_vl_interface': 3.450557844596517e-06, 'learning_rate/action_model': 3.450557844596517e-05, 'epoch': 1.48}) +[2026-06-19 16:58:51,781][starVLA.training.train_starvla][INFO] - Step 3157, Loss: {'train/loss': 0.09660406410694122, 'train/grad_norm_pre_clip': 4.875244617462158, 'timing/data': 0.009495918988250196, 'timing/model': 1.2673776260344312, 'learning_rate/qwen_vl_interface': 3.447739779687345e-06, 'learning_rate/action_model': 3.4477397796873446e-05, 'epoch': 1.48}) +[2026-06-19 16:58:56,224][starVLA.training.train_starvla][INFO] - Step 3158, Loss: {'train/loss': 0.11673718690872192, 'train/grad_norm_pre_clip': 6.074470043182373, 'timing/data': 0.00950487400405109, 'timing/model': 1.1577949419151992, 'learning_rate/qwen_vl_interface': 3.4449224556192722e-06, 'learning_rate/action_model': 3.444922455619272e-05, 'epoch': 1.49}) +[2026-06-19 16:59:00,819][starVLA.training.train_starvla][INFO] - Step 3159, Loss: {'train/loss': 0.23419031500816345, 'train/grad_norm_pre_clip': 9.951249122619629, 'timing/data': 0.009558981051668525, 'timing/model': 1.1466205440228805, 'learning_rate/qwen_vl_interface': 3.442105873550395e-06, 'learning_rate/action_model': 3.442105873550395e-05, 'epoch': 1.49}) +[2026-06-19 16:59:05,417][starVLA.training.train_starvla][INFO] - Step 3160, Loss: {'train/loss': 0.19740606844425201, 'train/grad_norm_pre_clip': 13.918286323547363, 'timing/data': 0.009775654994882643, 'timing/model': 1.154206725070253, 'learning_rate/qwen_vl_interface': 3.4392900346385022e-06, 'learning_rate/action_model': 3.4392900346385024e-05, 'epoch': 1.49}) +[2026-06-19 16:59:10,022][starVLA.training.train_starvla][INFO] - Step 3161, Loss: {'train/loss': 0.15094776451587677, 'train/grad_norm_pre_clip': 7.666920185089111, 'timing/data': 0.009524657973088324, 'timing/model': 1.1503220400772989, 'learning_rate/qwen_vl_interface': 3.4364749400410796e-06, 'learning_rate/action_model': 3.43647494004108e-05, 'epoch': 1.49}) +[2026-06-19 16:59:14,623][starVLA.training.train_starvla][INFO] - Step 3162, Loss: {'train/loss': 0.2843799293041229, 'train/grad_norm_pre_clip': 23.397836685180664, 'timing/data': 0.00940956105478108, 'timing/model': 1.308343545999378, 'learning_rate/qwen_vl_interface': 3.4336605909153066e-06, 'learning_rate/action_model': 3.433660590915306e-05, 'epoch': 1.49}) +[2026-06-19 16:59:19,187][starVLA.training.train_starvla][INFO] - Step 3163, Loss: {'train/loss': 0.15840177237987518, 'train/grad_norm_pre_clip': 10.832369804382324, 'timing/data': 0.00942744396161288, 'timing/model': 1.2576287789270282, 'learning_rate/qwen_vl_interface': 3.430846988418057e-06, 'learning_rate/action_model': 3.430846988418057e-05, 'epoch': 1.49}) +[2026-06-19 16:59:23,631][starVLA.training.train_starvla][INFO] - Step 3164, Loss: {'train/loss': 0.22671209275722504, 'train/grad_norm_pre_clip': 18.375314712524414, 'timing/data': 0.009340425953269005, 'timing/model': 1.1510862449649721, 'learning_rate/qwen_vl_interface': 3.428034133705893e-06, 'learning_rate/action_model': 3.428034133705893e-05, 'epoch': 1.49}) +[2026-06-19 16:59:28,238][starVLA.training.train_starvla][INFO] - Step 3165, Loss: {'train/loss': 0.18017832934856415, 'train/grad_norm_pre_clip': 12.041890144348145, 'timing/data': 0.009216225938871503, 'timing/model': 1.155987381003797, 'learning_rate/qwen_vl_interface': 3.4252220279350763e-06, 'learning_rate/action_model': 3.425222027935076e-05, 'epoch': 1.49}) +[2026-06-19 16:59:32,876][starVLA.training.train_starvla][INFO] - Step 3166, Loss: {'train/loss': 0.2854847311973572, 'train/grad_norm_pre_clip': 11.206583976745605, 'timing/data': 0.009468684089370072, 'timing/model': 1.1579838650068268, 'learning_rate/qwen_vl_interface': 3.4224106722615547e-06, 'learning_rate/action_model': 3.422410672261554e-05, 'epoch': 1.49}) +[2026-06-19 16:59:37,522][starVLA.training.train_starvla][INFO] - Step 3167, Loss: {'train/loss': 0.20830830931663513, 'train/grad_norm_pre_clip': 13.123955726623535, 'timing/data': 0.009182180976495147, 'timing/model': 1.1912099909968674, 'learning_rate/qwen_vl_interface': 3.419600067840973e-06, 'learning_rate/action_model': 3.419600067840973e-05, 'epoch': 1.49}) +[2026-06-19 16:59:42,143][starVLA.training.train_starvla][INFO] - Step 3168, Loss: {'train/loss': 0.1801975667476654, 'train/grad_norm_pre_clip': 7.517773628234863, 'timing/data': 0.008661386906169355, 'timing/model': 1.3328582639805973, 'learning_rate/qwen_vl_interface': 3.4167902158286627e-06, 'learning_rate/action_model': 3.4167902158286626e-05, 'epoch': 1.49}) +[2026-06-19 16:59:46,594][starVLA.training.train_starvla][INFO] - Step 3169, Loss: {'train/loss': 0.2522667348384857, 'train/grad_norm_pre_clip': 19.080568313598633, 'timing/data': 0.009358840994536877, 'timing/model': 1.147809459012933, 'learning_rate/qwen_vl_interface': 3.4139811173796468e-06, 'learning_rate/action_model': 3.413981117379647e-05, 'epoch': 1.49}) +[2026-06-19 16:59:51,222][starVLA.training.train_starvla][INFO] - Step 3170, Loss: {'train/loss': 0.2770041823387146, 'train/grad_norm_pre_clip': 14.02253246307373, 'timing/data': 0.009326467989012599, 'timing/model': 1.1431732220808044, 'learning_rate/qwen_vl_interface': 3.411172773648642e-06, 'learning_rate/action_model': 3.411172773648642e-05, 'epoch': 1.49}) +[2026-06-19 16:59:55,819][starVLA.training.train_starvla][INFO] - Step 3171, Loss: {'train/loss': 0.4558309316635132, 'train/grad_norm_pre_clip': 12.305607795715332, 'timing/data': 0.009487111936323345, 'timing/model': 1.1507986320648342, 'learning_rate/qwen_vl_interface': 3.408365185790052e-06, 'learning_rate/action_model': 3.4083651857900515e-05, 'epoch': 1.49}) +[2026-06-19 17:00:00,440][starVLA.training.train_starvla][INFO] - Step 3172, Loss: {'train/loss': 0.11479005962610245, 'train/grad_norm_pre_clip': 5.691384792327881, 'timing/data': 0.009010092937387526, 'timing/model': 1.1609186860732734, 'learning_rate/qwen_vl_interface': 3.40555835495797e-06, 'learning_rate/action_model': 3.40555835495797e-05, 'epoch': 1.49}) +[2026-06-19 17:00:05,045][starVLA.training.train_starvla][INFO] - Step 3173, Loss: {'train/loss': 0.16470928490161896, 'train/grad_norm_pre_clip': 9.481409072875977, 'timing/data': 0.009446745039895177, 'timing/model': 1.145671235048212, 'learning_rate/qwen_vl_interface': 3.4027522823061763e-06, 'learning_rate/action_model': 3.4027522823061765e-05, 'epoch': 1.49}) +[2026-06-19 17:00:09,652][starVLA.training.train_starvla][INFO] - Step 3174, Loss: {'train/loss': 0.13653579354286194, 'train/grad_norm_pre_clip': 8.515944480895996, 'timing/data': 0.009485909016802907, 'timing/model': 1.3307022550143301, 'learning_rate/qwen_vl_interface': 3.399946968988143e-06, 'learning_rate/action_model': 3.399946968988143e-05, 'epoch': 1.49}) +[2026-06-19 17:00:14,215][starVLA.training.train_starvla][INFO] - Step 3175, Loss: {'train/loss': 0.06089644134044647, 'train/grad_norm_pre_clip': 3.7660598754882812, 'timing/data': 0.00916398095432669, 'timing/model': 1.2780979638919234, 'learning_rate/qwen_vl_interface': 3.397142416157028e-06, 'learning_rate/action_model': 3.397142416157028e-05, 'epoch': 1.49}) +[2026-06-19 17:00:18,659][starVLA.training.train_starvla][INFO] - Step 3176, Loss: {'train/loss': 0.31295040249824524, 'train/grad_norm_pre_clip': 23.235389709472656, 'timing/data': 0.00914652191568166, 'timing/model': 1.1552505049621686, 'learning_rate/qwen_vl_interface': 3.3943386249656766e-06, 'learning_rate/action_model': 3.394338624965677e-05, 'epoch': 1.49}) +[2026-06-19 17:00:23,273][starVLA.training.train_starvla][INFO] - Step 3177, Loss: {'train/loss': 0.2362366020679474, 'train/grad_norm_pre_clip': 13.075237274169922, 'timing/data': 0.009495145990513265, 'timing/model': 1.1504443109733984, 'learning_rate/qwen_vl_interface': 3.391535596566624e-06, 'learning_rate/action_model': 3.391535596566624e-05, 'epoch': 1.49}) +[2026-06-19 17:00:27,897][starVLA.training.train_starvla][INFO] - Step 3178, Loss: {'train/loss': 0.2076902985572815, 'train/grad_norm_pre_clip': 4.791084289550781, 'timing/data': 0.009463346912525594, 'timing/model': 1.1490216020029038, 'learning_rate/qwen_vl_interface': 3.388733332112086e-06, 'learning_rate/action_model': 3.3887333321120856e-05, 'epoch': 1.49}) +[2026-06-19 17:00:32,529][starVLA.training.train_starvla][INFO] - Step 3179, Loss: {'train/loss': 0.26798561215400696, 'train/grad_norm_pre_clip': 10.403721809387207, 'timing/data': 0.009447272052057087, 'timing/model': 1.153282154002227, 'learning_rate/qwen_vl_interface': 3.3859318327539693e-06, 'learning_rate/action_model': 3.385931832753969e-05, 'epoch': 1.5}) +[2026-06-19 17:00:37,143][starVLA.training.train_starvla][INFO] - Step 3180, Loss: {'train/loss': 0.3467038571834564, 'train/grad_norm_pre_clip': 20.641754150390625, 'timing/data': 0.009169175056740642, 'timing/model': 1.3097973199328408, 'learning_rate/qwen_vl_interface': 3.3831310996438633e-06, 'learning_rate/action_model': 3.383131099643863e-05, 'epoch': 1.5}) +[2026-06-19 17:00:41,718][starVLA.training.train_starvla][INFO] - Step 3181, Loss: {'train/loss': 0.10056687146425247, 'train/grad_norm_pre_clip': 7.110398292541504, 'timing/data': 0.00963901390787214, 'timing/model': 1.2740645740414038, 'learning_rate/qwen_vl_interface': 3.3803311339330447e-06, 'learning_rate/action_model': 3.3803311339330444e-05, 'epoch': 1.5}) +[2026-06-19 17:00:46,195][starVLA.training.train_starvla][INFO] - Step 3182, Loss: {'train/loss': 0.1575922966003418, 'train/grad_norm_pre_clip': 10.310131072998047, 'timing/data': 0.0095639229984954, 'timing/model': 1.162777898949571, 'learning_rate/qwen_vl_interface': 3.3775319367724746e-06, 'learning_rate/action_model': 3.3775319367724745e-05, 'epoch': 1.5}) +[2026-06-19 17:00:50,838][starVLA.training.train_starvla][INFO] - Step 3183, Loss: {'train/loss': 0.18342037498950958, 'train/grad_norm_pre_clip': 11.449858665466309, 'timing/data': 0.009531134041026235, 'timing/model': 1.1615685960277915, 'learning_rate/qwen_vl_interface': 3.374733509312795e-06, 'learning_rate/action_model': 3.374733509312795e-05, 'epoch': 1.5}) +[2026-06-19 17:00:55,477][starVLA.training.train_starvla][INFO] - Step 3184, Loss: {'train/loss': 0.055325623601675034, 'train/grad_norm_pre_clip': 4.699770927429199, 'timing/data': 0.009611436980776489, 'timing/model': 1.1470833849161863, 'learning_rate/qwen_vl_interface': 3.3719358527043345e-06, 'learning_rate/action_model': 3.371935852704334e-05, 'epoch': 1.5}) +[2026-06-19 17:01:00,095][starVLA.training.train_starvla][INFO] - Step 3185, Loss: {'train/loss': 0.18496918678283691, 'train/grad_norm_pre_clip': 10.285571098327637, 'timing/data': 0.009246813948266208, 'timing/model': 1.1448313699802384, 'learning_rate/qwen_vl_interface': 3.3691389680971047e-06, 'learning_rate/action_model': 3.3691389680971044e-05, 'epoch': 1.5}) +[2026-06-19 17:01:04,714][starVLA.training.train_starvla][INFO] - Step 3186, Loss: {'train/loss': 0.21900641918182373, 'train/grad_norm_pre_clip': 15.557686805725098, 'timing/data': 0.009226099005900323, 'timing/model': 1.1466058489168063, 'learning_rate/qwen_vl_interface': 3.3663428566407985e-06, 'learning_rate/action_model': 3.366342856640798e-05, 'epoch': 1.5}) +[2026-06-19 17:01:09,314][starVLA.training.train_starvla][INFO] - Step 3187, Loss: {'train/loss': 0.2430768460035324, 'train/grad_norm_pre_clip': 11.802148818969727, 'timing/data': 0.009378354996442795, 'timing/model': 1.3200434999307618, 'learning_rate/qwen_vl_interface': 3.3635475194847927e-06, 'learning_rate/action_model': 3.363547519484793e-05, 'epoch': 1.5}) +[2026-06-19 17:01:13,774][starVLA.training.train_starvla][INFO] - Step 3188, Loss: {'train/loss': 0.3769505023956299, 'train/grad_norm_pre_clip': 21.957822799682617, 'timing/data': 0.009333722060546279, 'timing/model': 1.1555898620281368, 'learning_rate/qwen_vl_interface': 3.3607529577781433e-06, 'learning_rate/action_model': 3.360752957778143e-05, 'epoch': 1.5}) +[2026-06-19 17:01:18,399][starVLA.training.train_starvla][INFO] - Step 3189, Loss: {'train/loss': 0.28396859765052795, 'train/grad_norm_pre_clip': 17.334896087646484, 'timing/data': 0.009727800963446498, 'timing/model': 1.1591141519602388, 'learning_rate/qwen_vl_interface': 3.3579591726695906e-06, 'learning_rate/action_model': 3.35795917266959e-05, 'epoch': 1.5}) +[2026-06-19 17:01:23,016][starVLA.training.train_starvla][INFO] - Step 3190, Loss: {'train/loss': 0.18506507575511932, 'train/grad_norm_pre_clip': 16.737043380737305, 'timing/data': 0.009373259032145143, 'timing/model': 1.1502773480024189, 'learning_rate/qwen_vl_interface': 3.3551661653075522e-06, 'learning_rate/action_model': 3.355166165307552e-05, 'epoch': 1.5}) +[2026-06-19 17:01:27,643][starVLA.training.train_starvla][INFO] - Step 3191, Loss: {'train/loss': 0.08808861672878265, 'train/grad_norm_pre_clip': 11.409586906433105, 'timing/data': 0.009390027029439807, 'timing/model': 1.1447328509530053, 'learning_rate/qwen_vl_interface': 3.35237393684013e-06, 'learning_rate/action_model': 3.35237393684013e-05, 'epoch': 1.5}) +[2026-06-19 17:01:32,256][starVLA.training.train_starvla][INFO] - Step 3192, Loss: {'train/loss': 0.28850555419921875, 'train/grad_norm_pre_clip': 11.129752159118652, 'timing/data': 0.010201497934758663, 'timing/model': 1.1604226749623194, 'learning_rate/qwen_vl_interface': 3.349582488415103e-06, 'learning_rate/action_model': 3.3495824884151026e-05, 'epoch': 1.5}) +[2026-06-19 17:01:36,856][starVLA.training.train_starvla][INFO] - Step 3193, Loss: {'train/loss': 0.15096493065357208, 'train/grad_norm_pre_clip': 8.964158058166504, 'timing/data': 0.010037650004960597, 'timing/model': 1.3107933709397912, 'learning_rate/qwen_vl_interface': 3.3467918211799294e-06, 'learning_rate/action_model': 3.3467918211799295e-05, 'epoch': 1.5}) +[2026-06-19 17:01:41,287][starVLA.training.train_starvla][INFO] - Step 3194, Loss: {'train/loss': 0.22853367030620575, 'train/grad_norm_pre_clip': 8.336946487426758, 'timing/data': 0.009449413977563381, 'timing/model': 1.151554983924143, 'learning_rate/qwen_vl_interface': 3.3440019362817476e-06, 'learning_rate/action_model': 3.344001936281747e-05, 'epoch': 1.5}) +[2026-06-19 17:01:45,879][starVLA.training.train_starvla][INFO] - Step 3195, Loss: {'train/loss': 0.09429796040058136, 'train/grad_norm_pre_clip': 6.223514556884766, 'timing/data': 0.009307353990152478, 'timing/model': 1.1457899879897013, 'learning_rate/qwen_vl_interface': 3.3412128348673745e-06, 'learning_rate/action_model': 3.341212834867374e-05, 'epoch': 1.5}) +[2026-06-19 17:01:50,373][starVLA.training.train_starvla][INFO] - Step 3196, Loss: {'train/loss': 0.24369487166404724, 'train/grad_norm_pre_clip': 15.953752517700195, 'timing/data': 0.009095776942558587, 'timing/model': 1.1463458850048482, 'learning_rate/qwen_vl_interface': 3.338424518083303e-06, 'learning_rate/action_model': 3.338424518083303e-05, 'epoch': 1.5}) +[2026-06-19 17:01:54,928][starVLA.training.train_starvla][INFO] - Step 3197, Loss: {'train/loss': 0.19945821166038513, 'train/grad_norm_pre_clip': 8.18701457977295, 'timing/data': 0.009756150073371828, 'timing/model': 1.1552123230649158, 'learning_rate/qwen_vl_interface': 3.3356369870757056e-06, 'learning_rate/action_model': 3.335636987075705e-05, 'epoch': 1.5}) +[2026-06-19 17:01:59,560][starVLA.training.train_starvla][INFO] - Step 3198, Loss: {'train/loss': 0.3123058080673218, 'train/grad_norm_pre_clip': 20.89899444580078, 'timing/data': 0.00927408505231142, 'timing/model': 1.1621052849804983, 'learning_rate/qwen_vl_interface': 3.3328502429904327e-06, 'learning_rate/action_model': 3.3328502429904325e-05, 'epoch': 1.5}) +[2026-06-19 17:02:04,190][starVLA.training.train_starvla][INFO] - Step 3199, Loss: {'train/loss': 0.22325342893600464, 'train/grad_norm_pre_clip': 14.799317359924316, 'timing/data': 0.009543510968796909, 'timing/model': 1.3219824089901522, 'learning_rate/qwen_vl_interface': 3.3300642869730066e-06, 'learning_rate/action_model': 3.330064286973006e-05, 'epoch': 1.5}) +[2026-06-19 17:02:08,653][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 3200 +[2026-06-19 17:05:03,343][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 3200 +[2026-06-19 17:05:03,344][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 3200 +[2026-06-19 17:08:11,951][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 3200 +[2026-06-19 17:08:11,952][starVLA.training.train_starvla][INFO] - Step 3200, Loss: {'train/loss': 0.20991571247577667, 'train/grad_norm_pre_clip': 10.355793952941895, 'eval/loss': 0.20348162949085236, 'eval/latency_0/loss': 0.08521348983049393, 'eval/latency_1/loss': 0.05524424836039543, 'eval/latency_2/loss': 0.12103023380041122, 'eval/latency_3/loss': 0.18531541526317596, 'eval/latency_4/loss': 0.5671702027320862, 'eval/flappy/loss': 0.20348162949085236, 'eval/flappy/latency_0/loss': 0.08521348983049393, 'eval/flappy/latency_1/loss': 0.05524424836039543, 'eval/flappy/latency_2/loss': 0.12103023380041122, 'eval/flappy/latency_3/loss': 0.18531541526317596, 'eval/flappy/latency_4/loss': 0.5671702027320862, 'eval/flap_precision': 0.688135593220339, 'eval/flap_recall': 0.7381818181818182, 'eval/flap_f1': 0.7122807017543858, 'eval/flap_rate_model': 0.08194444444444444, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.7122807017543858, 'eval/latency_0/flap_precision': 0.688135593220339, 'eval/latency_0/flap_recall': 0.7381818181818182, 'eval/latency_0/flap_f1': 0.7122807017543858, 'eval/latency_0/flap_rate_model': 0.08194444444444444, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.7122807017543858, 'timing/data': 0.008995436015538871, 'timing/model': 1.1620635049184784, 'learning_rate/qwen_vl_interface': 3.3272791201686294e-06, 'learning_rate/action_model': 3.327279120168629e-05, 'epoch': 1.51}) +[2026-06-19 17:08:11,952][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_3200_state +[2026-06-19 17:08:11,955][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-19 17:08:15,346][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_3200_state/model.safetensors +[2026-06-19 17:08:22,515][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_3200_state/optimizer.bin +[2026-06-19 17:08:22,515][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_3200_state/sampler.bin +[2026-06-19 17:08:22,515][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_3200_state/sampler_1.bin +[2026-06-19 17:08:22,516][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_3200_state/random_states_0.pkl +[2026-06-19 17:08:22,540][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3200 +[2026-06-19 17:08:22,540][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-19 17:08:22,566][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-19 17:08:27,095][starVLA.training.train_starvla][INFO] - Step 3201, Loss: {'train/loss': 0.2565368115901947, 'train/grad_norm_pre_clip': 9.980087280273438, 'timing/data': 0.010134428041055799, 'timing/model': 1.1318261050619185, 'learning_rate/qwen_vl_interface': 3.324494743722179e-06, 'learning_rate/action_model': 3.3244947437221785e-05, 'epoch': 1.51}) +[2026-06-19 17:08:31,617][starVLA.training.train_starvla][INFO] - Step 3202, Loss: {'train/loss': 0.3848520815372467, 'train/grad_norm_pre_clip': 15.672569274902344, 'timing/data': 0.010268400073982775, 'timing/model': 1.3057174619752914, 'learning_rate/qwen_vl_interface': 3.321711158778205e-06, 'learning_rate/action_model': 3.321711158778205e-05, 'epoch': 1.51}) +[2026-06-19 17:08:35,983][starVLA.training.train_starvla][INFO] - Step 3203, Loss: {'train/loss': 0.08687856048345566, 'train/grad_norm_pre_clip': 4.383416652679443, 'timing/data': 0.010068787029013038, 'timing/model': 1.1341804299736395, 'learning_rate/qwen_vl_interface': 3.3189283664809393e-06, 'learning_rate/action_model': 3.318928366480939e-05, 'epoch': 1.51}) +[2026-06-19 17:08:40,523][starVLA.training.train_starvla][INFO] - Step 3204, Loss: {'train/loss': 0.2586430609226227, 'train/grad_norm_pre_clip': 8.769689559936523, 'timing/data': 0.010230229003354907, 'timing/model': 1.1348830809583887, 'learning_rate/qwen_vl_interface': 3.316146367974276e-06, 'learning_rate/action_model': 3.316146367974276e-05, 'epoch': 1.51}) +[2026-06-19 17:08:45,075][starVLA.training.train_starvla][INFO] - Step 3205, Loss: {'train/loss': 0.31316694617271423, 'train/grad_norm_pre_clip': 19.250141143798828, 'timing/data': 0.00989583192858845, 'timing/model': 1.1317715420154855, 'learning_rate/qwen_vl_interface': 3.3133651644017945e-06, 'learning_rate/action_model': 3.3133651644017944e-05, 'epoch': 1.51}) +[2026-06-19 17:08:49,579][starVLA.training.train_starvla][INFO] - Step 3206, Loss: {'train/loss': 0.48428595066070557, 'train/grad_norm_pre_clip': 16.11219024658203, 'timing/data': 0.009700244991108775, 'timing/model': 1.1413197489455342, 'learning_rate/qwen_vl_interface': 3.31058475690674e-06, 'learning_rate/action_model': 3.31058475690674e-05, 'epoch': 1.51}) +[2026-06-19 17:08:54,060][starVLA.training.train_starvla][INFO] - Step 3207, Loss: {'train/loss': 0.28882554173469543, 'train/grad_norm_pre_clip': 11.506359100341797, 'timing/data': 0.01024493807926774, 'timing/model': 1.140871354029514, 'learning_rate/qwen_vl_interface': 3.3078051466320337e-06, 'learning_rate/action_model': 3.307805146632034e-05, 'epoch': 1.51}) +[2026-06-19 17:08:58,627][starVLA.training.train_starvla][INFO] - Step 3208, Loss: {'train/loss': 0.21518999338150024, 'train/grad_norm_pre_clip': 11.997965812683105, 'timing/data': 0.01112748496234417, 'timing/model': 1.3173166640335694, 'learning_rate/qwen_vl_interface': 3.3050263347202712e-06, 'learning_rate/action_model': 3.305026334720271e-05, 'epoch': 1.51}) +[2026-06-19 17:09:03,012][starVLA.training.train_starvla][INFO] - Step 3209, Loss: {'train/loss': 0.1183098703622818, 'train/grad_norm_pre_clip': 6.7832231521606445, 'timing/data': 0.009873301954939961, 'timing/model': 1.1351665679831058, 'learning_rate/qwen_vl_interface': 3.302248322313712e-06, 'learning_rate/action_model': 3.302248322313712e-05, 'epoch': 1.51}) +[2026-06-19 17:09:07,581][starVLA.training.train_starvla][INFO] - Step 3210, Loss: {'train/loss': 0.18325795233249664, 'train/grad_norm_pre_clip': 12.299090385437012, 'timing/data': 0.011178269982337952, 'timing/model': 1.1453733009984717, 'learning_rate/qwen_vl_interface': 3.299471110554295e-06, 'learning_rate/action_model': 3.299471110554295e-05, 'epoch': 1.51}) +[2026-06-19 17:09:12,167][starVLA.training.train_starvla][INFO] - Step 3211, Loss: {'train/loss': 0.08757045120000839, 'train/grad_norm_pre_clip': 3.8630335330963135, 'timing/data': 0.009670639061369002, 'timing/model': 1.1389786519575864, 'learning_rate/qwen_vl_interface': 3.296694700583626e-06, 'learning_rate/action_model': 3.296694700583626e-05, 'epoch': 1.51}) +[2026-06-19 17:09:16,762][starVLA.training.train_starvla][INFO] - Step 3212, Loss: {'train/loss': 0.08505342900753021, 'train/grad_norm_pre_clip': 5.673293113708496, 'timing/data': 0.010057786013931036, 'timing/model': 1.1483688890002668, 'learning_rate/qwen_vl_interface': 3.2939190935429833e-06, 'learning_rate/action_model': 3.293919093542983e-05, 'epoch': 1.51}) +[2026-06-19 17:09:21,335][starVLA.training.train_starvla][INFO] - Step 3213, Loss: {'train/loss': 0.1876010000705719, 'train/grad_norm_pre_clip': 13.39700698852539, 'timing/data': 0.010531344916671515, 'timing/model': 1.320165261044167, 'learning_rate/qwen_vl_interface': 3.2911442905733126e-06, 'learning_rate/action_model': 3.291144290573312e-05, 'epoch': 1.51}) +[2026-06-19 17:09:25,732][starVLA.training.train_starvla][INFO] - Step 3214, Loss: {'train/loss': 0.0965270921587944, 'train/grad_norm_pre_clip': 8.11605167388916, 'timing/data': 0.010000810027122498, 'timing/model': 1.1356395710026845, 'learning_rate/qwen_vl_interface': 3.2883702928152323e-06, 'learning_rate/action_model': 3.288370292815232e-05, 'epoch': 1.51}) +[2026-06-19 17:09:30,325][starVLA.training.train_starvla][INFO] - Step 3215, Loss: {'train/loss': 0.3224145472049713, 'train/grad_norm_pre_clip': 9.03818416595459, 'timing/data': 0.01035820902325213, 'timing/model': 1.142107376945205, 'learning_rate/qwen_vl_interface': 3.2855971014090275e-06, 'learning_rate/action_model': 3.285597101409028e-05, 'epoch': 1.51}) +[2026-06-19 17:09:34,901][starVLA.training.train_starvla][INFO] - Step 3216, Loss: {'train/loss': 0.22995591163635254, 'train/grad_norm_pre_clip': 17.855743408203125, 'timing/data': 0.010267358971759677, 'timing/model': 1.1467474559322, 'learning_rate/qwen_vl_interface': 3.2828247174946528e-06, 'learning_rate/action_model': 3.2828247174946524e-05, 'epoch': 1.51}) +[2026-06-19 17:09:39,505][starVLA.training.train_starvla][INFO] - Step 3217, Loss: {'train/loss': 0.1545536071062088, 'train/grad_norm_pre_clip': 9.730165481567383, 'timing/data': 0.009831313975155354, 'timing/model': 1.1459401280153543, 'learning_rate/qwen_vl_interface': 3.280053142211729e-06, 'learning_rate/action_model': 3.280053142211729e-05, 'epoch': 1.51}) +[2026-06-19 17:09:44,086][starVLA.training.train_starvla][INFO] - Step 3218, Loss: {'train/loss': 0.18368326127529144, 'train/grad_norm_pre_clip': 9.054330825805664, 'timing/data': 0.009496352053247392, 'timing/model': 1.1435935439076275, 'learning_rate/qwen_vl_interface': 3.2772823766995475e-06, 'learning_rate/action_model': 3.277282376699548e-05, 'epoch': 1.51}) +[2026-06-19 17:09:48,690][starVLA.training.train_starvla][INFO] - Step 3219, Loss: {'train/loss': 0.19818486273288727, 'train/grad_norm_pre_clip': 13.518579483032227, 'timing/data': 0.010176496929489076, 'timing/model': 1.321559758973308, 'learning_rate/qwen_vl_interface': 3.2745124220970646e-06, 'learning_rate/action_model': 3.274512422097065e-05, 'epoch': 1.51}) +[2026-06-19 17:09:53,293][starVLA.training.train_starvla][INFO] - Step 3220, Loss: {'train/loss': 0.20810458064079285, 'train/grad_norm_pre_clip': 11.719919204711914, 'timing/data': 0.01009709993377328, 'timing/model': 1.3340324510354549, 'learning_rate/qwen_vl_interface': 3.2717432795429057e-06, 'learning_rate/action_model': 3.271743279542905e-05, 'epoch': 1.51}) +[2026-06-19 17:09:57,702][starVLA.training.train_starvla][INFO] - Step 3221, Loss: {'train/loss': 0.40520110726356506, 'train/grad_norm_pre_clip': 14.960527420043945, 'timing/data': 0.00985402800142765, 'timing/model': 1.1420404500095174, 'learning_rate/qwen_vl_interface': 3.2689749501753578e-06, 'learning_rate/action_model': 3.2689749501753574e-05, 'epoch': 1.51}) +[2026-06-19 17:10:02,301][starVLA.training.train_starvla][INFO] - Step 3222, Loss: {'train/loss': 0.08628150820732117, 'train/grad_norm_pre_clip': 5.81040620803833, 'timing/data': 0.012460577068850398, 'timing/model': 1.1554358259309083, 'learning_rate/qwen_vl_interface': 3.2662074351323774e-06, 'learning_rate/action_model': 3.266207435132378e-05, 'epoch': 1.52}) +[2026-06-19 17:10:06,880][starVLA.training.train_starvla][INFO] - Step 3223, Loss: {'train/loss': 0.19878140091896057, 'train/grad_norm_pre_clip': 11.201517105102539, 'timing/data': 0.009615907096304, 'timing/model': 1.1518905189586803, 'learning_rate/qwen_vl_interface': 3.2634407355515875e-06, 'learning_rate/action_model': 3.2634407355515875e-05, 'epoch': 1.52}) +[2026-06-19 17:10:11,472][starVLA.training.train_starvla][INFO] - Step 3224, Loss: {'train/loss': 0.38217711448669434, 'train/grad_norm_pre_clip': 23.63259506225586, 'timing/data': 0.010031054029241204, 'timing/model': 1.1540324959205464, 'learning_rate/qwen_vl_interface': 3.2606748525702715e-06, 'learning_rate/action_model': 3.260674852570271e-05, 'epoch': 1.52}) +[2026-06-19 17:10:16,064][starVLA.training.train_starvla][INFO] - Step 3225, Loss: {'train/loss': 0.23467470705509186, 'train/grad_norm_pre_clip': 7.563326835632324, 'timing/data': 0.01012563705444336, 'timing/model': 1.1494546900503337, 'learning_rate/qwen_vl_interface': 3.2579097873253803e-06, 'learning_rate/action_model': 3.25790978732538e-05, 'epoch': 1.52}) +[2026-06-19 17:10:20,670][starVLA.training.train_starvla][INFO] - Step 3226, Loss: {'train/loss': 0.195757195353508, 'train/grad_norm_pre_clip': 7.153681755065918, 'timing/data': 0.010165507905185223, 'timing/model': 1.3241086180787534, 'learning_rate/qwen_vl_interface': 3.2551455409535274e-06, 'learning_rate/action_model': 3.255145540953528e-05, 'epoch': 1.52}) +[2026-06-19 17:10:25,105][starVLA.training.train_starvla][INFO] - Step 3227, Loss: {'train/loss': 0.11458863317966461, 'train/grad_norm_pre_clip': 11.220746040344238, 'timing/data': 0.009625246049836278, 'timing/model': 1.148532366962172, 'learning_rate/qwen_vl_interface': 3.2523821145909907e-06, 'learning_rate/action_model': 3.252382114590991e-05, 'epoch': 1.52}) +[2026-06-19 17:10:29,711][starVLA.training.train_starvla][INFO] - Step 3228, Loss: {'train/loss': 0.12443460524082184, 'train/grad_norm_pre_clip': 5.479781627655029, 'timing/data': 0.010084014968015254, 'timing/model': 1.1484230629866943, 'learning_rate/qwen_vl_interface': 3.249619509373712e-06, 'learning_rate/action_model': 3.249619509373712e-05, 'epoch': 1.52}) +[2026-06-19 17:10:34,310][starVLA.training.train_starvla][INFO] - Step 3229, Loss: {'train/loss': 0.21691633760929108, 'train/grad_norm_pre_clip': 12.302108764648438, 'timing/data': 0.009809472016058862, 'timing/model': 1.1512013600440696, 'learning_rate/qwen_vl_interface': 3.2468577264372905e-06, 'learning_rate/action_model': 3.24685772643729e-05, 'epoch': 1.52}) +[2026-06-19 17:10:38,894][starVLA.training.train_starvla][INFO] - Step 3230, Loss: {'train/loss': 0.33019211888313293, 'train/grad_norm_pre_clip': 8.330232620239258, 'timing/data': 0.009900594945065677, 'timing/model': 1.1484286850318313, 'learning_rate/qwen_vl_interface': 3.244096766916992e-06, 'learning_rate/action_model': 3.244096766916992e-05, 'epoch': 1.52}) +[2026-06-19 17:10:43,472][starVLA.training.train_starvla][INFO] - Step 3231, Loss: {'train/loss': 0.2104778289794922, 'train/grad_norm_pre_clip': 7.454969882965088, 'timing/data': 0.009205786976963282, 'timing/model': 1.1426398990442976, 'learning_rate/qwen_vl_interface': 3.241336631947743e-06, 'learning_rate/action_model': 3.2413366319477427e-05, 'epoch': 1.52}) +[2026-06-19 17:10:48,039][starVLA.training.train_starvla][INFO] - Step 3232, Loss: {'train/loss': 0.11100620031356812, 'train/grad_norm_pre_clip': 13.5201997756958, 'timing/data': 0.009511049021966755, 'timing/model': 1.2994276649551466, 'learning_rate/qwen_vl_interface': 3.238577322664131e-06, 'learning_rate/action_model': 3.238577322664131e-05, 'epoch': 1.52}) +[2026-06-19 17:10:52,531][starVLA.training.train_starvla][INFO] - Step 3233, Loss: {'train/loss': 0.1345350593328476, 'train/grad_norm_pre_clip': 7.280871391296387, 'timing/data': 0.009783825953491032, 'timing/model': 1.2108184050302953, 'learning_rate/qwen_vl_interface': 3.2358188402004026e-06, 'learning_rate/action_model': 3.2358188402004024e-05, 'epoch': 1.52}) +[2026-06-19 17:10:56,988][starVLA.training.train_starvla][INFO] - Step 3234, Loss: {'train/loss': 0.19943636655807495, 'train/grad_norm_pre_clip': 15.704001426696777, 'timing/data': 0.009304106933996081, 'timing/model': 1.170553969917819, 'learning_rate/qwen_vl_interface': 3.2330611856904675e-06, 'learning_rate/action_model': 3.233061185690468e-05, 'epoch': 1.52}) +[2026-06-19 17:11:01,526][starVLA.training.train_starvla][INFO] - Step 3235, Loss: {'train/loss': 0.08878042548894882, 'train/grad_norm_pre_clip': 4.600404262542725, 'timing/data': 0.009660647017881274, 'timing/model': 1.147972084931098, 'learning_rate/qwen_vl_interface': 3.2303043602678928e-06, 'learning_rate/action_model': 3.230304360267893e-05, 'epoch': 1.52}) +[2026-06-19 17:11:06,133][starVLA.training.train_starvla][INFO] - Step 3236, Loss: {'train/loss': 0.16733288764953613, 'train/grad_norm_pre_clip': 7.951346397399902, 'timing/data': 0.010039200074970722, 'timing/model': 1.156564934994094, 'learning_rate/qwen_vl_interface': 3.227548365065905e-06, 'learning_rate/action_model': 3.227548365065905e-05, 'epoch': 1.52}) +[2026-06-19 17:11:10,742][starVLA.training.train_starvla][INFO] - Step 3237, Loss: {'train/loss': 0.21558234095573425, 'train/grad_norm_pre_clip': 9.258478164672852, 'timing/data': 0.009890993009321392, 'timing/model': 1.1539646169403568, 'learning_rate/qwen_vl_interface': 3.2247932012173887e-06, 'learning_rate/action_model': 3.2247932012173884e-05, 'epoch': 1.52}) +[2026-06-19 17:11:15,359][starVLA.training.train_starvla][INFO] - Step 3238, Loss: {'train/loss': 0.17615759372711182, 'train/grad_norm_pre_clip': 12.137482643127441, 'timing/data': 0.009621171979233623, 'timing/model': 1.330116639030166, 'learning_rate/qwen_vl_interface': 3.2220388698548897e-06, 'learning_rate/action_model': 3.22203886985489e-05, 'epoch': 1.52}) +[2026-06-19 17:11:19,832][starVLA.training.train_starvla][INFO] - Step 3239, Loss: {'train/loss': 0.1407623291015625, 'train/grad_norm_pre_clip': 8.278048515319824, 'timing/data': 0.009841948980465531, 'timing/model': 1.1531478890683502, 'learning_rate/qwen_vl_interface': 3.21928537211061e-06, 'learning_rate/action_model': 3.21928537211061e-05, 'epoch': 1.52}) +[2026-06-19 17:11:24,443][starVLA.training.train_starvla][INFO] - Step 3240, Loss: {'train/loss': 0.21904784440994263, 'train/grad_norm_pre_clip': 15.966466903686523, 'timing/data': 0.00966245902236551, 'timing/model': 1.1520365899195895, 'learning_rate/qwen_vl_interface': 3.216532709116405e-06, 'learning_rate/action_model': 3.216532709116405e-05, 'epoch': 1.52}) +[2026-06-19 17:11:29,030][starVLA.training.train_starvla][INFO] - Step 3241, Loss: {'train/loss': 0.3491496741771698, 'train/grad_norm_pre_clip': 46.59219741821289, 'timing/data': 0.009535313001833856, 'timing/model': 1.1470290999859571, 'learning_rate/qwen_vl_interface': 3.2137808820037925e-06, 'learning_rate/action_model': 3.2137808820037924e-05, 'epoch': 1.52}) +[2026-06-19 17:11:33,553][starVLA.training.train_starvla][INFO] - Step 3242, Loss: {'train/loss': 0.27263885736465454, 'train/grad_norm_pre_clip': 21.173542022705078, 'timing/data': 0.009872396010905504, 'timing/model': 1.153687197016552, 'learning_rate/qwen_vl_interface': 3.211029891903945e-06, 'learning_rate/action_model': 3.211029891903945e-05, 'epoch': 1.52}) +[2026-06-19 17:11:38,152][starVLA.training.train_starvla][INFO] - Step 3243, Loss: {'train/loss': 0.11936963349580765, 'train/grad_norm_pre_clip': 16.881832122802734, 'timing/data': 0.010030106990598142, 'timing/model': 1.1612771359505132, 'learning_rate/qwen_vl_interface': 3.2082797399476894e-06, 'learning_rate/action_model': 3.2082797399476895e-05, 'epoch': 1.53}) +[2026-06-19 17:11:42,789][starVLA.training.train_starvla][INFO] - Step 3244, Loss: {'train/loss': 0.13842816650867462, 'train/grad_norm_pre_clip': 9.695595741271973, 'timing/data': 0.009886429063044488, 'timing/model': 1.3404579201014712, 'learning_rate/qwen_vl_interface': 3.2055304272655114e-06, 'learning_rate/action_model': 3.205530427265511e-05, 'epoch': 1.53}) +[2026-06-19 17:11:47,242][starVLA.training.train_starvla][INFO] - Step 3245, Loss: {'train/loss': 0.17908671498298645, 'train/grad_norm_pre_clip': 8.51059341430664, 'timing/data': 0.009452539030462503, 'timing/model': 1.1486825520405546, 'learning_rate/qwen_vl_interface': 3.2027819549875477e-06, 'learning_rate/action_model': 3.202781954987548e-05, 'epoch': 1.53}) +[2026-06-19 17:11:51,861][starVLA.training.train_starvla][INFO] - Step 3246, Loss: {'train/loss': 0.210804283618927, 'train/grad_norm_pre_clip': 8.532167434692383, 'timing/data': 0.00969791691750288, 'timing/model': 1.1473919739946723, 'learning_rate/qwen_vl_interface': 3.2000343242435926e-06, 'learning_rate/action_model': 3.2000343242435924e-05, 'epoch': 1.53}) +[2026-06-19 17:11:56,460][starVLA.training.train_starvla][INFO] - Step 3247, Loss: {'train/loss': 0.11470291018486023, 'train/grad_norm_pre_clip': 6.130005836486816, 'timing/data': 0.010041018016636372, 'timing/model': 1.147391195059754, 'learning_rate/qwen_vl_interface': 3.1972875361630923e-06, 'learning_rate/action_model': 3.1972875361630923e-05, 'epoch': 1.53}) +[2026-06-19 17:12:01,069][starVLA.training.train_starvla][INFO] - Step 3248, Loss: {'train/loss': 0.16763469576835632, 'train/grad_norm_pre_clip': 7.361911773681641, 'timing/data': 0.009864018997177482, 'timing/model': 1.1479530209908262, 'learning_rate/qwen_vl_interface': 3.1945415918751484e-06, 'learning_rate/action_model': 3.194541591875148e-05, 'epoch': 1.53}) +[2026-06-19 17:12:05,699][starVLA.training.train_starvla][INFO] - Step 3249, Loss: {'train/loss': 0.32756826281547546, 'train/grad_norm_pre_clip': 12.983383178710938, 'timing/data': 0.00979839893989265, 'timing/model': 1.1486731850309297, 'learning_rate/qwen_vl_interface': 3.191796492508516e-06, 'learning_rate/action_model': 3.191796492508516e-05, 'epoch': 1.53}) +[2026-06-19 17:12:10,314][starVLA.training.train_starvla][INFO] - Step 3250, Loss: {'train/loss': 0.21798278391361237, 'train/grad_norm_pre_clip': 12.229086875915527, 'timing/data': 0.00992633297573775, 'timing/model': 1.310037249000743, 'learning_rate/qwen_vl_interface': 3.189052239191599e-06, 'learning_rate/action_model': 3.189052239191599e-05, 'epoch': 1.53}) +[2026-06-19 17:12:14,881][starVLA.training.train_starvla][INFO] - Step 3251, Loss: {'train/loss': 0.3198357820510864, 'train/grad_norm_pre_clip': 17.02094841003418, 'timing/data': 0.010337718995288014, 'timing/model': 1.2865240550599992, 'learning_rate/qwen_vl_interface': 3.1863088330524577e-06, 'learning_rate/action_model': 3.186308833052458e-05, 'epoch': 1.53}) +[2026-06-19 17:12:19,311][starVLA.training.train_starvla][INFO] - Step 3252, Loss: {'train/loss': 0.3629251718521118, 'train/grad_norm_pre_clip': 14.36900806427002, 'timing/data': 0.01031209307257086, 'timing/model': 1.1487432689173147, 'learning_rate/qwen_vl_interface': 3.183566275218803e-06, 'learning_rate/action_model': 3.1835662752188026e-05, 'epoch': 1.53}) +[2026-06-19 17:12:23,902][starVLA.training.train_starvla][INFO] - Step 3253, Loss: {'train/loss': 0.1480177789926529, 'train/grad_norm_pre_clip': 15.187712669372559, 'timing/data': 0.010768144973553717, 'timing/model': 1.1469853369053453, 'learning_rate/qwen_vl_interface': 3.180824566817996e-06, 'learning_rate/action_model': 3.1808245668179956e-05, 'epoch': 1.53}) +[2026-06-19 17:12:28,512][starVLA.training.train_starvla][INFO] - Step 3254, Loss: {'train/loss': 0.12206637114286423, 'train/grad_norm_pre_clip': 7.197847366333008, 'timing/data': 0.010200188029557467, 'timing/model': 1.1525607770308852, 'learning_rate/qwen_vl_interface': 3.1780837089770524e-06, 'learning_rate/action_model': 3.1780837089770524e-05, 'epoch': 1.53}) +[2026-06-19 17:12:33,111][starVLA.training.train_starvla][INFO] - Step 3255, Loss: {'train/loss': 0.30429306626319885, 'train/grad_norm_pre_clip': 8.913599967956543, 'timing/data': 0.009706813958473504, 'timing/model': 1.1500983550213277, 'learning_rate/qwen_vl_interface': 3.17534370282263e-06, 'learning_rate/action_model': 3.1753437028226303e-05, 'epoch': 1.53}) +[2026-06-19 17:12:37,739][starVLA.training.train_starvla][INFO] - Step 3256, Loss: {'train/loss': 0.2610284388065338, 'train/grad_norm_pre_clip': 11.482831001281738, 'timing/data': 0.009592034970410168, 'timing/model': 1.1533755080308765, 'learning_rate/qwen_vl_interface': 3.172604549481046e-06, 'learning_rate/action_model': 3.1726045494810455e-05, 'epoch': 1.53}) +[2026-06-19 17:12:42,364][starVLA.training.train_starvla][INFO] - Step 3257, Loss: {'train/loss': 0.47307127714157104, 'train/grad_norm_pre_clip': 24.797792434692383, 'timing/data': 0.009470278047956526, 'timing/model': 1.3251589050050825, 'learning_rate/qwen_vl_interface': 3.169866250078261e-06, 'learning_rate/action_model': 3.169866250078261e-05, 'epoch': 1.53}) +[2026-06-19 17:12:46,816][starVLA.training.train_starvla][INFO] - Step 3258, Loss: {'train/loss': 0.19485126435756683, 'train/grad_norm_pre_clip': 10.077205657958984, 'timing/data': 0.009739384055137634, 'timing/model': 1.1533581269904971, 'learning_rate/qwen_vl_interface': 3.167128805739887e-06, 'learning_rate/action_model': 3.1671288057398866e-05, 'epoch': 1.53}) +[2026-06-19 17:12:51,466][starVLA.training.train_starvla][INFO] - Step 3259, Loss: {'train/loss': 0.2537318468093872, 'train/grad_norm_pre_clip': 9.101069450378418, 'timing/data': 0.010056137922219932, 'timing/model': 1.163814133964479, 'learning_rate/qwen_vl_interface': 3.1643922175911847e-06, 'learning_rate/action_model': 3.164392217591185e-05, 'epoch': 1.53}) +[2026-06-19 17:12:56,081][starVLA.training.train_starvla][INFO] - Step 3260, Loss: {'train/loss': 0.263489305973053, 'train/grad_norm_pre_clip': 12.785482406616211, 'timing/data': 0.0098248899448663, 'timing/model': 1.1665013349847868, 'learning_rate/qwen_vl_interface': 3.1616564867570603e-06, 'learning_rate/action_model': 3.16165648675706e-05, 'epoch': 1.53}) +[2026-06-19 17:13:00,696][starVLA.training.train_starvla][INFO] - Step 3261, Loss: {'train/loss': 0.21008871495723724, 'train/grad_norm_pre_clip': 9.120415687561035, 'timing/data': 0.009791700053028762, 'timing/model': 1.1467163420747966, 'learning_rate/qwen_vl_interface': 3.1589216143620733e-06, 'learning_rate/action_model': 3.1589216143620734e-05, 'epoch': 1.53}) +[2026-06-19 17:13:05,294][starVLA.training.train_starvla][INFO] - Step 3262, Loss: {'train/loss': 0.2760622203350067, 'train/grad_norm_pre_clip': 16.903793334960938, 'timing/data': 0.009760363958775997, 'timing/model': 1.1478855209425092, 'learning_rate/qwen_vl_interface': 3.1561876015304215e-06, 'learning_rate/action_model': 3.156187601530421e-05, 'epoch': 1.53}) +[2026-06-19 17:13:09,892][starVLA.training.train_starvla][INFO] - Step 3263, Loss: {'train/loss': 0.2856660783290863, 'train/grad_norm_pre_clip': 7.339053153991699, 'timing/data': 0.010107640060596168, 'timing/model': 1.3153713759966195, 'learning_rate/qwen_vl_interface': 3.1534544493859563e-06, 'learning_rate/action_model': 3.1534544493859566e-05, 'epoch': 1.53}) +[2026-06-19 17:13:14,438][starVLA.training.train_starvla][INFO] - Step 3264, Loss: {'train/loss': 0.23504184186458588, 'train/grad_norm_pre_clip': 10.063393592834473, 'timing/data': 0.009825360961258411, 'timing/model': 1.2451879479922354, 'learning_rate/qwen_vl_interface': 3.150722159052174e-06, 'learning_rate/action_model': 3.150722159052174e-05, 'epoch': 1.54}) +[2026-06-19 17:13:18,874][starVLA.training.train_starvla][INFO] - Step 3265, Loss: {'train/loss': 0.29919955134391785, 'train/grad_norm_pre_clip': 8.374310493469238, 'timing/data': 0.010386588051915169, 'timing/model': 1.1569887079531327, 'learning_rate/qwen_vl_interface': 3.147990731652215e-06, 'learning_rate/action_model': 3.147990731652215e-05, 'epoch': 1.54}) +[2026-06-19 17:13:23,460][starVLA.training.train_starvla][INFO] - Step 3266, Loss: {'train/loss': 0.18404777348041534, 'train/grad_norm_pre_clip': 6.828144073486328, 'timing/data': 0.00964897102676332, 'timing/model': 1.1447708000196144, 'learning_rate/qwen_vl_interface': 3.1452601683088673e-06, 'learning_rate/action_model': 3.145260168308867e-05, 'epoch': 1.54}) +[2026-06-19 17:13:28,098][starVLA.training.train_starvla][INFO] - Step 3267, Loss: {'train/loss': 0.08878976851701736, 'train/grad_norm_pre_clip': 4.306847095489502, 'timing/data': 0.00917143898550421, 'timing/model': 1.1498714060289785, 'learning_rate/qwen_vl_interface': 3.1425304701445608e-06, 'learning_rate/action_model': 3.1425304701445605e-05, 'epoch': 1.54}) +[2026-06-19 17:13:32,707][starVLA.training.train_starvla][INFO] - Step 3268, Loss: {'train/loss': 0.18826740980148315, 'train/grad_norm_pre_clip': 6.865281105041504, 'timing/data': 0.009990275953896344, 'timing/model': 1.1633152399444953, 'learning_rate/qwen_vl_interface': 3.139801638281372e-06, 'learning_rate/action_model': 3.139801638281372e-05, 'epoch': 1.54}) +[2026-06-19 17:13:37,323][starVLA.training.train_starvla][INFO] - Step 3269, Loss: {'train/loss': 0.35056957602500916, 'train/grad_norm_pre_clip': 19.201839447021484, 'timing/data': 0.009594289003871381, 'timing/model': 1.3230466459644958, 'learning_rate/qwen_vl_interface': 3.13707367384102e-06, 'learning_rate/action_model': 3.13707367384102e-05, 'epoch': 1.54}) +[2026-06-19 17:13:41,845][starVLA.training.train_starvla][INFO] - Step 3270, Loss: {'train/loss': 0.30291640758514404, 'train/grad_norm_pre_clip': 14.37177562713623, 'timing/data': 0.010457120020873845, 'timing/model': 1.234054406057112, 'learning_rate/qwen_vl_interface': 3.134346577944869e-06, 'learning_rate/action_model': 3.134346577944869e-05, 'epoch': 1.54}) +[2026-06-19 17:13:46,287][starVLA.training.train_starvla][INFO] - Step 3271, Loss: {'train/loss': 0.19950664043426514, 'train/grad_norm_pre_clip': 12.06908130645752, 'timing/data': 0.010389640927314758, 'timing/model': 1.1490464130183682, 'learning_rate/qwen_vl_interface': 3.1316203517139243e-06, 'learning_rate/action_model': 3.131620351713924e-05, 'epoch': 1.54}) +[2026-06-19 17:13:50,895][starVLA.training.train_starvla][INFO] - Step 3272, Loss: {'train/loss': 0.329820841550827, 'train/grad_norm_pre_clip': 21.603574752807617, 'timing/data': 0.00977228197734803, 'timing/model': 1.1626771399751306, 'learning_rate/qwen_vl_interface': 3.128894996268837e-06, 'learning_rate/action_model': 3.128894996268837e-05, 'epoch': 1.54}) +[2026-06-19 17:13:55,488][starVLA.training.train_starvla][INFO] - Step 3273, Loss: {'train/loss': 0.16332346200942993, 'train/grad_norm_pre_clip': 7.705524921417236, 'timing/data': 0.009568145964294672, 'timing/model': 1.147716163075529, 'learning_rate/qwen_vl_interface': 3.1261705127298934e-06, 'learning_rate/action_model': 3.1261705127298935e-05, 'epoch': 1.54}) +[2026-06-19 17:14:00,090][starVLA.training.train_starvla][INFO] - Step 3274, Loss: {'train/loss': 0.17915499210357666, 'train/grad_norm_pre_clip': 4.709717750549316, 'timing/data': 0.009712183033116162, 'timing/model': 1.1547957370057702, 'learning_rate/qwen_vl_interface': 3.123446902217028e-06, 'learning_rate/action_model': 3.1234469022170283e-05, 'epoch': 1.54}) +[2026-06-19 17:14:04,697][starVLA.training.train_starvla][INFO] - Step 3275, Loss: {'train/loss': 0.2032020390033722, 'train/grad_norm_pre_clip': 11.353927612304688, 'timing/data': 0.010221324977464974, 'timing/model': 1.3110308420145884, 'learning_rate/qwen_vl_interface': 3.1207241658498145e-06, 'learning_rate/action_model': 3.1207241658498145e-05, 'epoch': 1.54}) +[2026-06-19 17:14:09,236][starVLA.training.train_starvla][INFO] - Step 3276, Loss: {'train/loss': 0.19215308129787445, 'train/grad_norm_pre_clip': 8.138850212097168, 'timing/data': 0.00982331798877567, 'timing/model': 1.256018242915161, 'learning_rate/qwen_vl_interface': 3.1180023047474667e-06, 'learning_rate/action_model': 3.1180023047474664e-05, 'epoch': 1.54}) +[2026-06-19 17:14:13,676][starVLA.training.train_starvla][INFO] - Step 3277, Loss: {'train/loss': 0.11352324485778809, 'train/grad_norm_pre_clip': 7.922868728637695, 'timing/data': 0.010209355968981981, 'timing/model': 1.1501888540806249, 'learning_rate/qwen_vl_interface': 3.1152813200288386e-06, 'learning_rate/action_model': 3.115281320028838e-05, 'epoch': 1.54}) +[2026-06-19 17:14:18,368][starVLA.training.train_starvla][INFO] - Step 3278, Loss: {'train/loss': 0.22020722925662994, 'train/grad_norm_pre_clip': 46.13325881958008, 'timing/data': 0.07604354899376631, 'timing/model': 1.162575175985694, 'learning_rate/qwen_vl_interface': 3.1125612128124244e-06, 'learning_rate/action_model': 3.112561212812424e-05, 'epoch': 1.54}) +[2026-06-19 17:14:22,990][starVLA.training.train_starvla][INFO] - Step 3279, Loss: {'train/loss': 0.20577369630336761, 'train/grad_norm_pre_clip': 12.458364486694336, 'timing/data': 0.009392503998242319, 'timing/model': 1.1482536440016702, 'learning_rate/qwen_vl_interface': 3.1098419842163573e-06, 'learning_rate/action_model': 3.109841984216357e-05, 'epoch': 1.54}) +[2026-06-19 17:14:27,596][starVLA.training.train_starvla][INFO] - Step 3280, Loss: {'train/loss': 0.3732295334339142, 'train/grad_norm_pre_clip': 15.963193893432617, 'timing/data': 0.010135748074389994, 'timing/model': 1.1487302780151367, 'learning_rate/qwen_vl_interface': 3.10712363535841e-06, 'learning_rate/action_model': 3.10712363535841e-05, 'epoch': 1.54}) +[2026-06-19 17:14:32,198][starVLA.training.train_starvla][INFO] - Step 3281, Loss: {'train/loss': 0.06186690181493759, 'train/grad_norm_pre_clip': 5.160327434539795, 'timing/data': 0.010046549956314266, 'timing/model': 1.3198741050437093, 'learning_rate/qwen_vl_interface': 3.1044061673559926e-06, 'learning_rate/action_model': 3.1044061673559925e-05, 'epoch': 1.54}) +[2026-06-19 17:14:36,725][starVLA.training.train_starvla][INFO] - Step 3282, Loss: {'train/loss': 0.11331939697265625, 'train/grad_norm_pre_clip': 6.3419060707092285, 'timing/data': 0.010141770937480032, 'timing/model': 1.24045462498907, 'learning_rate/qwen_vl_interface': 3.1016895813261556e-06, 'learning_rate/action_model': 3.101689581326155e-05, 'epoch': 1.54}) +[2026-06-19 17:14:41,182][starVLA.training.train_starvla][INFO] - Step 3283, Loss: {'train/loss': 0.18191204965114594, 'train/grad_norm_pre_clip': 7.771256923675537, 'timing/data': 0.009746546973474324, 'timing/model': 1.1535695110214874, 'learning_rate/qwen_vl_interface': 3.098973878385581e-06, 'learning_rate/action_model': 3.098973878385581e-05, 'epoch': 1.54}) +[2026-06-19 17:14:45,772][starVLA.training.train_starvla][INFO] - Step 3284, Loss: {'train/loss': 0.19726525247097015, 'train/grad_norm_pre_clip': 8.535248756408691, 'timing/data': 0.010025267023593187, 'timing/model': 1.1502793350955471, 'learning_rate/qwen_vl_interface': 3.096259059650594e-06, 'learning_rate/action_model': 3.096259059650594e-05, 'epoch': 1.54}) +[2026-06-19 17:14:50,376][starVLA.training.train_starvla][INFO] - Step 3285, Loss: {'train/loss': 0.13977313041687012, 'train/grad_norm_pre_clip': 8.692496299743652, 'timing/data': 0.009672662010416389, 'timing/model': 1.15281921508722, 'learning_rate/qwen_vl_interface': 3.093545126237153e-06, 'learning_rate/action_model': 3.0935451262371527e-05, 'epoch': 1.54}) +[2026-06-19 17:14:54,984][starVLA.training.train_starvla][INFO] - Step 3286, Loss: {'train/loss': 0.24890370666980743, 'train/grad_norm_pre_clip': 14.394967079162598, 'timing/data': 0.009732498088851571, 'timing/model': 1.1478838039329275, 'learning_rate/qwen_vl_interface': 3.090832079260855e-06, 'learning_rate/action_model': 3.0908320792608544e-05, 'epoch': 1.55}) +[2026-06-19 17:14:59,592][starVLA.training.train_starvla][INFO] - Step 3287, Loss: {'train/loss': 0.16991348564624786, 'train/grad_norm_pre_clip': 11.134057998657227, 'timing/data': 0.009758193977177143, 'timing/model': 1.3244328189175576, 'learning_rate/qwen_vl_interface': 3.088119919836929e-06, 'learning_rate/action_model': 3.088119919836929e-05, 'epoch': 1.55}) +[2026-06-19 17:15:04,138][starVLA.training.train_starvla][INFO] - Step 3288, Loss: {'train/loss': 0.2138294130563736, 'train/grad_norm_pre_clip': 10.585152626037598, 'timing/data': 0.009929604013450444, 'timing/model': 1.2463806449668482, 'learning_rate/qwen_vl_interface': 3.0854086490802433e-06, 'learning_rate/action_model': 3.085408649080243e-05, 'epoch': 1.55}) +[2026-06-19 17:15:08,578][starVLA.training.train_starvla][INFO] - Step 3289, Loss: {'train/loss': 0.2896459698677063, 'train/grad_norm_pre_clip': 17.539527893066406, 'timing/data': 0.009814695920795202, 'timing/model': 1.1494934830116108, 'learning_rate/qwen_vl_interface': 3.082698268105299e-06, 'learning_rate/action_model': 3.082698268105299e-05, 'epoch': 1.55}) +[2026-06-19 17:15:13,183][starVLA.training.train_starvla][INFO] - Step 3290, Loss: {'train/loss': 0.182737335562706, 'train/grad_norm_pre_clip': 13.671452522277832, 'timing/data': 0.009514780016615987, 'timing/model': 1.1523731560446322, 'learning_rate/qwen_vl_interface': 3.0799887780262295e-06, 'learning_rate/action_model': 3.0799887780262296e-05, 'epoch': 1.55}) +[2026-06-19 17:15:17,785][starVLA.training.train_starvla][INFO] - Step 3291, Loss: {'train/loss': 0.26399675011634827, 'train/grad_norm_pre_clip': 17.582853317260742, 'timing/data': 0.010110449977219105, 'timing/model': 1.1511110690189525, 'learning_rate/qwen_vl_interface': 3.0772801799568054e-06, 'learning_rate/action_model': 3.0772801799568054e-05, 'epoch': 1.55}) +[2026-06-19 17:15:22,390][starVLA.training.train_starvla][INFO] - Step 3292, Loss: {'train/loss': 0.29698362946510315, 'train/grad_norm_pre_clip': 12.389494895935059, 'timing/data': 0.010064939968287945, 'timing/model': 1.1547554059652612, 'learning_rate/qwen_vl_interface': 3.07457247501043e-06, 'learning_rate/action_model': 3.0745724750104296e-05, 'epoch': 1.55}) +[2026-06-19 17:15:27,019][starVLA.training.train_starvla][INFO] - Step 3293, Loss: {'train/loss': 0.28535521030426025, 'train/grad_norm_pre_clip': 11.99459457397461, 'timing/data': 0.010137094999663532, 'timing/model': 1.3195250449934974, 'learning_rate/qwen_vl_interface': 3.0718656643001353e-06, 'learning_rate/action_model': 3.071865664300135e-05, 'epoch': 1.55}) +[2026-06-19 17:15:31,538][starVLA.training.train_starvla][INFO] - Step 3294, Loss: {'train/loss': 0.12392126768827438, 'train/grad_norm_pre_clip': 6.418717384338379, 'timing/data': 0.009883265011012554, 'timing/model': 1.236737392959185, 'learning_rate/qwen_vl_interface': 3.069159748938591e-06, 'learning_rate/action_model': 3.069159748938591e-05, 'epoch': 1.55}) +[2026-06-19 17:15:35,969][starVLA.training.train_starvla][INFO] - Step 3295, Loss: {'train/loss': 0.3603525161743164, 'train/grad_norm_pre_clip': 9.982403755187988, 'timing/data': 0.00942005799151957, 'timing/model': 1.147284192033112, 'learning_rate/qwen_vl_interface': 3.066454730038096e-06, 'learning_rate/action_model': 3.0664547300380956e-05, 'epoch': 1.55}) +[2026-06-19 17:15:40,566][starVLA.training.train_starvla][INFO] - Step 3296, Loss: {'train/loss': 0.30465763807296753, 'train/grad_norm_pre_clip': 25.59975814819336, 'timing/data': 0.01010961493011564, 'timing/model': 1.1482518800767139, 'learning_rate/qwen_vl_interface': 3.0637506087105816e-06, 'learning_rate/action_model': 3.0637506087105815e-05, 'epoch': 1.55}) +[2026-06-19 17:15:45,193][starVLA.training.train_starvla][INFO] - Step 3297, Loss: {'train/loss': 0.23239335417747498, 'train/grad_norm_pre_clip': 9.271163940429688, 'timing/data': 0.009389353916049004, 'timing/model': 1.1511938479961827, 'learning_rate/qwen_vl_interface': 3.0610473860676103e-06, 'learning_rate/action_model': 3.06104738606761e-05, 'epoch': 1.55}) +[2026-06-19 17:15:49,801][starVLA.training.train_starvla][INFO] - Step 3298, Loss: {'train/loss': 0.10547380149364471, 'train/grad_norm_pre_clip': 6.6112165451049805, 'timing/data': 0.010013608960434794, 'timing/model': 1.1590030268998817, 'learning_rate/qwen_vl_interface': 3.0583450632203748e-06, 'learning_rate/action_model': 3.058345063220375e-05, 'epoch': 1.55}) +[2026-06-19 17:15:54,448][starVLA.training.train_starvla][INFO] - Step 3299, Loss: {'train/loss': 0.20898769795894623, 'train/grad_norm_pre_clip': 10.755648612976074, 'timing/data': 0.009819571976549923, 'timing/model': 1.3313893909798935, 'learning_rate/qwen_vl_interface': 3.055643641279699e-06, 'learning_rate/action_model': 3.055643641279699e-05, 'epoch': 1.55}) +[2026-06-19 17:15:58,883][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 3300 +[2026-06-19 17:18:52,323][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 3300 +[2026-06-19 17:18:52,324][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 3300 +[2026-06-19 17:22:03,967][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 3300 +[2026-06-19 17:22:03,968][starVLA.training.train_starvla][INFO] - Step 3300, Loss: {'train/loss': 0.18135468661785126, 'train/grad_norm_pre_clip': 12.268783569335938, 'eval/loss': 0.21042923629283905, 'eval/latency_0/loss': 0.0870339423418045, 'eval/latency_1/loss': 0.049277663230895996, 'eval/latency_2/loss': 0.11902498453855515, 'eval/latency_3/loss': 0.1860395222902298, 'eval/latency_4/loss': 0.6070706248283386, 'eval/flappy/loss': 0.21042923629283905, 'eval/flappy/latency_0/loss': 0.0870339423418045, 'eval/flappy/latency_1/loss': 0.049277663230895996, 'eval/flappy/latency_2/loss': 0.11902498453855515, 'eval/flappy/latency_3/loss': 0.1860395222902298, 'eval/flappy/latency_4/loss': 0.6070706248283386, 'eval/flap_precision': 0.6721854304635762, 'eval/flap_recall': 0.7381818181818182, 'eval/flap_f1': 0.7036395147313691, 'eval/flap_rate_model': 0.08388888888888889, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.7036395147313691, 'eval/latency_0/flap_precision': 0.6721854304635762, 'eval/latency_0/flap_recall': 0.7381818181818182, 'eval/latency_0/flap_f1': 0.7036395147313691, 'eval/latency_0/flap_rate_model': 0.08388888888888889, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.7036395147313691, 'timing/data': 0.009745357907377183, 'timing/model': 1.1556828520260751, 'learning_rate/qwen_vl_interface': 3.0529431213560334e-06, 'learning_rate/action_model': 3.0529431213560334e-05, 'epoch': 1.55}) +[2026-06-19 17:22:03,968][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_3300_state +[2026-06-19 17:22:03,971][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-19 17:22:07,379][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_3300_state/model.safetensors +[2026-06-19 17:22:14,608][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_3300_state/optimizer.bin +[2026-06-19 17:22:14,609][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_3300_state/sampler.bin +[2026-06-19 17:22:14,609][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_3300_state/sampler_1.bin +[2026-06-19 17:22:14,610][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_3300_state/random_states_0.pkl +[2026-06-19 17:22:14,633][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3300 +[2026-06-19 17:22:14,633][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-19 17:22:14,663][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-19 17:22:19,163][starVLA.training.train_starvla][INFO] - Step 3301, Loss: {'train/loss': 0.10561160743236542, 'train/grad_norm_pre_clip': 7.421168804168701, 'timing/data': 0.00925027800258249, 'timing/model': 1.123808342963457, 'learning_rate/qwen_vl_interface': 3.0502435045594596e-06, 'learning_rate/action_model': 3.0502435045594595e-05, 'epoch': 1.55}) +[2026-06-19 17:22:23,663][starVLA.training.train_starvla][INFO] - Step 3302, Loss: {'train/loss': 0.14627531170845032, 'train/grad_norm_pre_clip': 5.90978479385376, 'timing/data': 0.009699546033516526, 'timing/model': 1.1249877919908613, 'learning_rate/qwen_vl_interface': 3.047544791999689e-06, 'learning_rate/action_model': 3.0475447919996892e-05, 'epoch': 1.55}) +[2026-06-19 17:22:28,195][starVLA.training.train_starvla][INFO] - Step 3303, Loss: {'train/loss': 0.16861039400100708, 'train/grad_norm_pre_clip': 8.411535263061523, 'timing/data': 0.009695970918983221, 'timing/model': 1.30238311807625, 'learning_rate/qwen_vl_interface': 3.044846984786061e-06, 'learning_rate/action_model': 3.044846984786061e-05, 'epoch': 1.55}) +[2026-06-19 17:22:32,560][starVLA.training.train_starvla][INFO] - Step 3304, Loss: {'train/loss': 0.28265082836151123, 'train/grad_norm_pre_clip': 11.314987182617188, 'timing/data': 0.009225761983543634, 'timing/model': 1.1330145390238613, 'learning_rate/qwen_vl_interface': 3.0421500840275413e-06, 'learning_rate/action_model': 3.0421500840275413e-05, 'epoch': 1.55}) +[2026-06-19 17:22:37,093][starVLA.training.train_starvla][INFO] - Step 3305, Loss: {'train/loss': 0.35054585337638855, 'train/grad_norm_pre_clip': 14.02611255645752, 'timing/data': 0.008958898950368166, 'timing/model': 1.1347549059428275, 'learning_rate/qwen_vl_interface': 3.039454090832727e-06, 'learning_rate/action_model': 3.0394540908327268e-05, 'epoch': 1.55}) +[2026-06-19 17:22:41,558][starVLA.training.train_starvla][INFO] - Step 3306, Loss: {'train/loss': 0.39904460310935974, 'train/grad_norm_pre_clip': 10.92611312866211, 'timing/data': 0.009538206038996577, 'timing/model': 1.1363178620813414, 'learning_rate/qwen_vl_interface': 3.0367590063098333e-06, 'learning_rate/action_model': 3.0367590063098333e-05, 'epoch': 1.55}) +[2026-06-19 17:22:46,088][starVLA.training.train_starvla][INFO] - Step 3307, Loss: {'train/loss': 0.1726418435573578, 'train/grad_norm_pre_clip': 10.10888385772705, 'timing/data': 0.009446321055293083, 'timing/model': 1.1458124850178137, 'learning_rate/qwen_vl_interface': 3.0340648315667115e-06, 'learning_rate/action_model': 3.0340648315667113e-05, 'epoch': 1.56}) +[2026-06-19 17:22:50,654][starVLA.training.train_starvla][INFO] - Step 3308, Loss: {'train/loss': 0.24740582704544067, 'train/grad_norm_pre_clip': 15.040778160095215, 'timing/data': 0.00926016503944993, 'timing/model': 1.1383016830077395, 'learning_rate/qwen_vl_interface': 3.0313715677108336e-06, 'learning_rate/action_model': 3.0313715677108335e-05, 'epoch': 1.56}) +[2026-06-19 17:22:55,198][starVLA.training.train_starvla][INFO] - Step 3309, Loss: {'train/loss': 0.216543510556221, 'train/grad_norm_pre_clip': 10.986711502075195, 'timing/data': 0.009373001055791974, 'timing/model': 1.2949158049887046, 'learning_rate/qwen_vl_interface': 3.0286792158492996e-06, 'learning_rate/action_model': 3.0286792158492994e-05, 'epoch': 1.56}) +[2026-06-19 17:22:59,707][starVLA.training.train_starvla][INFO] - Step 3310, Loss: {'train/loss': 0.15474598109722137, 'train/grad_norm_pre_clip': 16.22112274169922, 'timing/data': 0.009575559059157968, 'timing/model': 1.2578647539485246, 'learning_rate/qwen_vl_interface': 3.0259877770888334e-06, 'learning_rate/action_model': 3.0259877770888333e-05, 'epoch': 1.56}) +[2026-06-19 17:23:04,092][starVLA.training.train_starvla][INFO] - Step 3311, Loss: {'train/loss': 0.16564248502254486, 'train/grad_norm_pre_clip': 13.067153930664062, 'timing/data': 0.009328844025731087, 'timing/model': 1.1440295180073008, 'learning_rate/qwen_vl_interface': 3.0232972525357836e-06, 'learning_rate/action_model': 3.0232972525357834e-05, 'epoch': 1.56}) +[2026-06-19 17:23:08,655][starVLA.training.train_starvla][INFO] - Step 3312, Loss: {'train/loss': 0.2640906870365143, 'train/grad_norm_pre_clip': 12.400385856628418, 'timing/data': 0.009952042018994689, 'timing/model': 1.1382183020468801, 'learning_rate/qwen_vl_interface': 3.020607643296123e-06, 'learning_rate/action_model': 3.0206076432961226e-05, 'epoch': 1.56}) +[2026-06-19 17:23:13,204][starVLA.training.train_starvla][INFO] - Step 3313, Loss: {'train/loss': 0.2939181327819824, 'train/grad_norm_pre_clip': 21.899253845214844, 'timing/data': 0.009401845978572965, 'timing/model': 1.1384489820338786, 'learning_rate/qwen_vl_interface': 3.017918950475449e-06, 'learning_rate/action_model': 3.017918950475449e-05, 'epoch': 1.56}) +[2026-06-19 17:23:17,756][starVLA.training.train_starvla][INFO] - Step 3314, Loss: {'train/loss': 0.19334562122821808, 'train/grad_norm_pre_clip': 10.51961612701416, 'timing/data': 0.009269241942092776, 'timing/model': 1.140277402009815, 'learning_rate/qwen_vl_interface': 3.0152311751789815e-06, 'learning_rate/action_model': 3.0152311751789814e-05, 'epoch': 1.56}) +[2026-06-19 17:23:22,245][starVLA.training.train_starvla][INFO] - Step 3315, Loss: {'train/loss': 0.2183433473110199, 'train/grad_norm_pre_clip': 16.42128562927246, 'timing/data': 0.009322604048065841, 'timing/model': 1.1497070760233328, 'learning_rate/qwen_vl_interface': 3.0125443185115655e-06, 'learning_rate/action_model': 3.0125443185115655e-05, 'epoch': 1.56}) +[2026-06-19 17:23:26,770][starVLA.training.train_starvla][INFO] - Step 3316, Loss: {'train/loss': 0.08455188572406769, 'train/grad_norm_pre_clip': 5.305659770965576, 'timing/data': 0.009634971036575735, 'timing/model': 1.2512055810075253, 'learning_rate/qwen_vl_interface': 3.0098583815776615e-06, 'learning_rate/action_model': 3.0098583815776615e-05, 'epoch': 1.56}) +[2026-06-19 17:23:31,184][starVLA.training.train_starvla][INFO] - Step 3317, Loss: {'train/loss': 0.10554972290992737, 'train/grad_norm_pre_clip': 9.729127883911133, 'timing/data': 0.008880909997969866, 'timing/model': 1.152776868082583, 'learning_rate/qwen_vl_interface': 3.0071733654813595e-06, 'learning_rate/action_model': 3.0071733654813595e-05, 'epoch': 1.56}) +[2026-06-19 17:23:35,766][starVLA.training.train_starvla][INFO] - Step 3318, Loss: {'train/loss': 0.4246327877044678, 'train/grad_norm_pre_clip': 17.909975051879883, 'timing/data': 0.009227383998222649, 'timing/model': 1.1492553279967979, 'learning_rate/qwen_vl_interface': 3.0044892713263685e-06, 'learning_rate/action_model': 3.0044892713263682e-05, 'epoch': 1.56}) +[2026-06-19 17:23:40,336][starVLA.training.train_starvla][INFO] - Step 3319, Loss: {'train/loss': 0.12282239645719528, 'train/grad_norm_pre_clip': 6.411296367645264, 'timing/data': 0.009090803912840784, 'timing/model': 1.1433647070080042, 'learning_rate/qwen_vl_interface': 3.001806100216017e-06, 'learning_rate/action_model': 3.001806100216017e-05, 'epoch': 1.56}) +[2026-06-19 17:23:44,912][starVLA.training.train_starvla][INFO] - Step 3320, Loss: {'train/loss': 0.22128327190876007, 'train/grad_norm_pre_clip': 11.054975509643555, 'timing/data': 0.009472475037910044, 'timing/model': 1.1433910480700433, 'learning_rate/qwen_vl_interface': 2.9991238532532567e-06, 'learning_rate/action_model': 2.9991238532532567e-05, 'epoch': 1.56}) +[2026-06-19 17:23:49,500][starVLA.training.train_starvla][INFO] - Step 3321, Loss: {'train/loss': 0.4547045826911926, 'train/grad_norm_pre_clip': 17.058786392211914, 'timing/data': 0.008990344940684736, 'timing/model': 1.3112740439828485, 'learning_rate/qwen_vl_interface': 2.996442531540657e-06, 'learning_rate/action_model': 2.9964425315406568e-05, 'epoch': 1.56}) +[2026-06-19 17:23:53,911][starVLA.training.train_starvla][INFO] - Step 3322, Loss: {'train/loss': 0.20860742032527924, 'train/grad_norm_pre_clip': 21.09882164001465, 'timing/data': 0.009544527973048389, 'timing/model': 1.1408858590293676, 'learning_rate/qwen_vl_interface': 2.993762136180408e-06, 'learning_rate/action_model': 2.993762136180408e-05, 'epoch': 1.56}) +[2026-06-19 17:23:58,505][starVLA.training.train_starvla][INFO] - Step 3323, Loss: {'train/loss': 0.16194167733192444, 'train/grad_norm_pre_clip': 8.431160926818848, 'timing/data': 0.009260534076020122, 'timing/model': 1.1409305810229853, 'learning_rate/qwen_vl_interface': 2.9910826682743195e-06, 'learning_rate/action_model': 2.9910826682743192e-05, 'epoch': 1.56}) +[2026-06-19 17:24:03,069][starVLA.training.train_starvla][INFO] - Step 3324, Loss: {'train/loss': 0.10781942307949066, 'train/grad_norm_pre_clip': 5.974088668823242, 'timing/data': 0.009092765976674855, 'timing/model': 1.143515395000577, 'learning_rate/qwen_vl_interface': 2.9884041289238192e-06, 'learning_rate/action_model': 2.988404128923819e-05, 'epoch': 1.56}) +[2026-06-19 17:24:07,652][starVLA.training.train_starvla][INFO] - Step 3325, Loss: {'train/loss': 0.10337013751268387, 'train/grad_norm_pre_clip': 7.175891399383545, 'timing/data': 0.00995579909067601, 'timing/model': 1.147148874006234, 'learning_rate/qwen_vl_interface': 2.9857265192299557e-06, 'learning_rate/action_model': 2.9857265192299556e-05, 'epoch': 1.56}) +[2026-06-19 17:24:12,254][starVLA.training.train_starvla][INFO] - Step 3326, Loss: {'train/loss': 0.15611134469509125, 'train/grad_norm_pre_clip': 13.084081649780273, 'timing/data': 0.009306131047196686, 'timing/model': 1.1482702479697764, 'learning_rate/qwen_vl_interface': 2.9830498402933894e-06, 'learning_rate/action_model': 2.9830498402933894e-05, 'epoch': 1.56}) +[2026-06-19 17:24:16,855][starVLA.training.train_starvla][INFO] - Step 3327, Loss: {'train/loss': 0.07945086807012558, 'train/grad_norm_pre_clip': 5.083451747894287, 'timing/data': 0.009457093081437051, 'timing/model': 1.3154286929639056, 'learning_rate/qwen_vl_interface': 2.9803740932144025e-06, 'learning_rate/action_model': 2.9803740932144024e-05, 'epoch': 1.56}) +[2026-06-19 17:24:21,409][starVLA.training.train_starvla][INFO] - Step 3328, Loss: {'train/loss': 0.2566218376159668, 'train/grad_norm_pre_clip': 11.85123062133789, 'timing/data': 0.009480332024395466, 'timing/model': 1.2744855710770935, 'learning_rate/qwen_vl_interface': 2.9776992790928953e-06, 'learning_rate/action_model': 2.9776992790928952e-05, 'epoch': 1.57}) +[2026-06-19 17:24:25,832][starVLA.training.train_starvla][INFO] - Step 3329, Loss: {'train/loss': 0.13947026431560516, 'train/grad_norm_pre_clip': 8.282499313354492, 'timing/data': 0.009189037955366075, 'timing/model': 1.1468949050176889, 'learning_rate/qwen_vl_interface': 2.975025399028382e-06, 'learning_rate/action_model': 2.975025399028382e-05, 'epoch': 1.57}) +[2026-06-19 17:24:30,465][starVLA.training.train_starvla][INFO] - Step 3330, Loss: {'train/loss': 0.22878317534923553, 'train/grad_norm_pre_clip': 14.75220775604248, 'timing/data': 0.009161041001789272, 'timing/model': 1.1686311879893765, 'learning_rate/qwen_vl_interface': 2.9723524541199943e-06, 'learning_rate/action_model': 2.972352454119994e-05, 'epoch': 1.57}) +[2026-06-19 17:24:35,085][starVLA.training.train_starvla][INFO] - Step 3331, Loss: {'train/loss': 0.18102645874023438, 'train/grad_norm_pre_clip': 14.326916694641113, 'timing/data': 0.009477501036599278, 'timing/model': 1.14952751994133, 'learning_rate/qwen_vl_interface': 2.969680445466478e-06, 'learning_rate/action_model': 2.969680445466478e-05, 'epoch': 1.57}) +[2026-06-19 17:24:39,732][starVLA.training.train_starvla][INFO] - Step 3332, Loss: {'train/loss': 0.06983945518732071, 'train/grad_norm_pre_clip': 12.719423294067383, 'timing/data': 0.009467897936701775, 'timing/model': 1.1421844629803672, 'learning_rate/qwen_vl_interface': 2.9670093741661955e-06, 'learning_rate/action_model': 2.9670093741661954e-05, 'epoch': 1.57}) +[2026-06-19 17:24:44,355][starVLA.training.train_starvla][INFO] - Step 3333, Loss: {'train/loss': 0.23269449174404144, 'train/grad_norm_pre_clip': 13.034183502197266, 'timing/data': 0.0095378850819543, 'timing/model': 1.3264533240580931, 'learning_rate/qwen_vl_interface': 2.964339241317123e-06, 'learning_rate/action_model': 2.964339241317123e-05, 'epoch': 1.57}) +[2026-06-19 17:24:49,763][starVLA.training.train_starvla][INFO] - Latency curriculum phase 4 at step 3334: active_latencies=[0, 1, 2, 3, 4], dataset_size=341369 +[2026-06-19 17:24:49,764][starVLA.training.train_starvla][INFO] - Step 3334, Loss: {'train/loss': 0.10062091797590256, 'train/grad_norm_pre_clip': 7.266732215881348, 'latency_curriculum/phase': 4.0, 'latency_curriculum/active_count': 5.0, 'latency_curriculum/max_active_latency': 4.0, 'latency_curriculum/dataset_size': 341369.0, 'timing/data': 0.009463199065066874, 'timing/model': 1.268352319020778, 'learning_rate/qwen_vl_interface': 2.961670048016852e-06, 'learning_rate/action_model': 2.9616700480168518e-05, 'epoch': 1.25}) +[2026-06-19 17:25:06,556][starVLA.training.train_starvla][INFO] - Step 3335, Loss: {'train/loss': 0.30915579199790955, 'train/grad_norm_pre_clip': 12.963294982910156, 'timing/data': 0.009422674076631665, 'timing/model': 1.1365940789692104, 'learning_rate/qwen_vl_interface': 2.9590017953625854e-06, 'learning_rate/action_model': 2.9590017953625853e-05, 'epoch': 1.25}) +[2026-06-19 17:25:11,147][starVLA.training.train_starvla][INFO] - Step 3336, Loss: {'train/loss': 0.30608654022216797, 'train/grad_norm_pre_clip': 21.407800674438477, 'timing/data': 0.010333884973078966, 'timing/model': 1.1381694610463455, 'learning_rate/qwen_vl_interface': 2.9563344844511455e-06, 'learning_rate/action_model': 2.9563344844511455e-05, 'epoch': 1.25}) +[2026-06-19 17:25:15,749][starVLA.training.train_starvla][INFO] - Step 3337, Loss: {'train/loss': 0.6074445843696594, 'train/grad_norm_pre_clip': 22.2597713470459, 'timing/data': 0.009448630968108773, 'timing/model': 1.1601098470855504, 'learning_rate/qwen_vl_interface': 2.9536681163789568e-06, 'learning_rate/action_model': 2.9536681163789564e-05, 'epoch': 1.25}) +[2026-06-19 17:25:20,357][starVLA.training.train_starvla][INFO] - Step 3338, Loss: {'train/loss': 0.16596730053424835, 'train/grad_norm_pre_clip': 17.258886337280273, 'timing/data': 0.00960313598625362, 'timing/model': 1.152225418947637, 'learning_rate/qwen_vl_interface': 2.9510026922420653e-06, 'learning_rate/action_model': 2.9510026922420653e-05, 'epoch': 1.25}) +[2026-06-19 17:25:24,958][starVLA.training.train_starvla][INFO] - Step 3339, Loss: {'train/loss': 0.40814825892448425, 'train/grad_norm_pre_clip': 14.550209999084473, 'timing/data': 0.009160777903161943, 'timing/model': 1.3125346690649167, 'learning_rate/qwen_vl_interface': 2.9483382131361273e-06, 'learning_rate/action_model': 2.948338213136127e-05, 'epoch': 1.25}) +[2026-06-19 17:25:29,498][starVLA.training.train_starvla][INFO] - Step 3340, Loss: {'train/loss': 0.595031201839447, 'train/grad_norm_pre_clip': 19.7774658203125, 'timing/data': 0.009696371969766915, 'timing/model': 1.265335871023126, 'learning_rate/qwen_vl_interface': 2.9456746801564072e-06, 'learning_rate/action_model': 2.945674680156407e-05, 'epoch': 1.25}) +[2026-06-19 17:25:33,948][starVLA.training.train_starvla][INFO] - Step 3341, Loss: {'train/loss': 0.5008041858673096, 'train/grad_norm_pre_clip': 20.601362228393555, 'timing/data': 0.00898236408829689, 'timing/model': 1.1513470219215378, 'learning_rate/qwen_vl_interface': 2.9430120943977854e-06, 'learning_rate/action_model': 2.9430120943977853e-05, 'epoch': 1.25}) +[2026-06-19 17:25:38,558][starVLA.training.train_starvla][INFO] - Step 3342, Loss: {'train/loss': 0.2795427143573761, 'train/grad_norm_pre_clip': 12.9947509765625, 'timing/data': 0.00972112303134054, 'timing/model': 1.1468122620135546, 'learning_rate/qwen_vl_interface': 2.940350456954748e-06, 'learning_rate/action_model': 2.940350456954748e-05, 'epoch': 1.25}) +[2026-06-19 17:25:43,173][starVLA.training.train_starvla][INFO] - Step 3343, Loss: {'train/loss': 0.4912779927253723, 'train/grad_norm_pre_clip': 19.131486892700195, 'timing/data': 0.010121305007487535, 'timing/model': 1.1546720180194825, 'learning_rate/qwen_vl_interface': 2.9376897689213952e-06, 'learning_rate/action_model': 2.937689768921395e-05, 'epoch': 1.25}) +[2026-06-19 17:25:47,783][starVLA.training.train_starvla][INFO] - Step 3344, Loss: {'train/loss': 0.3062744736671448, 'train/grad_norm_pre_clip': 17.077791213989258, 'timing/data': 0.010197737021371722, 'timing/model': 1.1437704570125788, 'learning_rate/qwen_vl_interface': 2.9350300313914364e-06, 'learning_rate/action_model': 2.9350300313914363e-05, 'epoch': 1.25}) +[2026-06-19 17:25:52,395][starVLA.training.train_starvla][INFO] - Step 3345, Loss: {'train/loss': 0.3332328796386719, 'train/grad_norm_pre_clip': 10.439165115356445, 'timing/data': 0.009381574927829206, 'timing/model': 1.313394855009392, 'learning_rate/qwen_vl_interface': 2.9323712454581867e-06, 'learning_rate/action_model': 2.9323712454581866e-05, 'epoch': 1.25}) +[2026-06-19 17:25:56,987][starVLA.training.train_starvla][INFO] - Step 3346, Loss: {'train/loss': 0.27762970328330994, 'train/grad_norm_pre_clip': 16.031381607055664, 'timing/data': 0.0095451429951936, 'timing/model': 1.2821163469925523, 'learning_rate/qwen_vl_interface': 2.929713412214572e-06, 'learning_rate/action_model': 2.929713412214572e-05, 'epoch': 1.25}) +[2026-06-19 17:26:01,418][starVLA.training.train_starvla][INFO] - Step 3347, Loss: {'train/loss': 0.29787397384643555, 'train/grad_norm_pre_clip': 13.108612060546875, 'timing/data': 0.010191110079176724, 'timing/model': 1.149171785917133, 'learning_rate/qwen_vl_interface': 2.9270565327531294e-06, 'learning_rate/action_model': 2.927056532753129e-05, 'epoch': 1.25}) +[2026-06-19 17:26:06,032][starVLA.training.train_starvla][INFO] - Step 3348, Loss: {'train/loss': 0.21577493846416473, 'train/grad_norm_pre_clip': 22.352022171020508, 'timing/data': 0.010175339062698185, 'timing/model': 1.1572530950652435, 'learning_rate/qwen_vl_interface': 2.924400608166003e-06, 'learning_rate/action_model': 2.924400608166003e-05, 'epoch': 1.26}) +[2026-06-19 17:26:10,651][starVLA.training.train_starvla][INFO] - Step 3349, Loss: {'train/loss': 0.30953556299209595, 'train/grad_norm_pre_clip': 31.995359420776367, 'timing/data': 0.010145512060262263, 'timing/model': 1.1535948440432549, 'learning_rate/qwen_vl_interface': 2.921745639544939e-06, 'learning_rate/action_model': 2.921745639544939e-05, 'epoch': 1.26}) +[2026-06-19 17:26:15,267][starVLA.training.train_starvla][INFO] - Step 3350, Loss: {'train/loss': 0.3784090280532837, 'train/grad_norm_pre_clip': 12.710118293762207, 'timing/data': 0.00941200798843056, 'timing/model': 1.1455613360740244, 'learning_rate/qwen_vl_interface': 2.919091627981297e-06, 'learning_rate/action_model': 2.919091627981297e-05, 'epoch': 1.26}) +[2026-06-19 17:26:19,840][starVLA.training.train_starvla][INFO] - Step 3351, Loss: {'train/loss': 0.46356338262557983, 'train/grad_norm_pre_clip': 15.736713409423828, 'timing/data': 0.009558856952935457, 'timing/model': 1.1447153240442276, 'learning_rate/qwen_vl_interface': 2.9164385745660396e-06, 'learning_rate/action_model': 2.9164385745660394e-05, 'epoch': 1.26}) +[2026-06-19 17:26:24,438][starVLA.training.train_starvla][INFO] - Step 3352, Loss: {'train/loss': 0.21797344088554382, 'train/grad_norm_pre_clip': 10.015538215637207, 'timing/data': 0.010089197079651058, 'timing/model': 1.309397667995654, 'learning_rate/qwen_vl_interface': 2.9137864803897385e-06, 'learning_rate/action_model': 2.9137864803897387e-05, 'epoch': 1.26}) +[2026-06-19 17:26:28,867][starVLA.training.train_starvla][INFO] - Step 3353, Loss: {'train/loss': 0.26090019941329956, 'train/grad_norm_pre_clip': 8.499281883239746, 'timing/data': 0.00973042706027627, 'timing/model': 1.144966421998106, 'learning_rate/qwen_vl_interface': 2.9111353465425696e-06, 'learning_rate/action_model': 2.9111353465425694e-05, 'epoch': 1.26}) +[2026-06-19 17:26:33,460][starVLA.training.train_starvla][INFO] - Step 3354, Loss: {'train/loss': 0.4490754008293152, 'train/grad_norm_pre_clip': 12.774563789367676, 'timing/data': 0.009468893986195326, 'timing/model': 1.1522062889998779, 'learning_rate/qwen_vl_interface': 2.908485174114312e-06, 'learning_rate/action_model': 2.908485174114312e-05, 'epoch': 1.26}) +[2026-06-19 17:26:38,046][starVLA.training.train_starvla][INFO] - Step 3355, Loss: {'train/loss': 0.23792694509029388, 'train/grad_norm_pre_clip': 8.30878734588623, 'timing/data': 0.009899400058202446, 'timing/model': 1.1446555110160261, 'learning_rate/qwen_vl_interface': 2.9058359641943523e-06, 'learning_rate/action_model': 2.9058359641943523e-05, 'epoch': 1.26}) +[2026-06-19 17:26:42,643][starVLA.training.train_starvla][INFO] - Step 3356, Loss: {'train/loss': 0.3012658655643463, 'train/grad_norm_pre_clip': 10.957450866699219, 'timing/data': 0.00986649899277836, 'timing/model': 1.1538828309858218, 'learning_rate/qwen_vl_interface': 2.9031877178716823e-06, 'learning_rate/action_model': 2.903187717871682e-05, 'epoch': 1.26}) +[2026-06-19 17:26:47,247][starVLA.training.train_starvla][INFO] - Step 3357, Loss: {'train/loss': 0.2152702957391739, 'train/grad_norm_pre_clip': 9.857622146606445, 'timing/data': 0.01013732305727899, 'timing/model': 1.1469627160113305, 'learning_rate/qwen_vl_interface': 2.900540436234894e-06, 'learning_rate/action_model': 2.9005404362348937e-05, 'epoch': 1.26}) +[2026-06-19 17:26:51,865][starVLA.training.train_starvla][INFO] - Step 3358, Loss: {'train/loss': 0.40130630135536194, 'train/grad_norm_pre_clip': 17.826000213623047, 'timing/data': 0.008877010899595916, 'timing/model': 1.3169170960318297, 'learning_rate/qwen_vl_interface': 2.897894120372185e-06, 'learning_rate/action_model': 2.897894120372185e-05, 'epoch': 1.26}) +[2026-06-19 17:26:56,291][starVLA.training.train_starvla][INFO] - Step 3359, Loss: {'train/loss': 0.31311801075935364, 'train/grad_norm_pre_clip': 8.215410232543945, 'timing/data': 0.010971736977808177, 'timing/model': 1.1460962110431865, 'learning_rate/qwen_vl_interface': 2.8952487713713585e-06, 'learning_rate/action_model': 2.8952487713713582e-05, 'epoch': 1.26}) +[2026-06-19 17:27:00,899][starVLA.training.train_starvla][INFO] - Step 3360, Loss: {'train/loss': 0.33455654978752136, 'train/grad_norm_pre_clip': 13.236738204956055, 'timing/data': 0.010381767991930246, 'timing/model': 1.1551461570197716, 'learning_rate/qwen_vl_interface': 2.8926043903198135e-06, 'learning_rate/action_model': 2.8926043903198135e-05, 'epoch': 1.26}) +[2026-06-19 17:27:05,418][starVLA.training.train_starvla][INFO] - Step 3361, Loss: {'train/loss': 0.41634804010391235, 'train/grad_norm_pre_clip': 16.834569931030273, 'timing/data': 0.009955245070159435, 'timing/model': 1.1518318409798667, 'learning_rate/qwen_vl_interface': 2.8899609783045574e-06, 'learning_rate/action_model': 2.8899609783045573e-05, 'epoch': 1.26}) +[2026-06-19 17:27:09,963][starVLA.training.train_starvla][INFO] - Step 3362, Loss: {'train/loss': 0.38370198011398315, 'train/grad_norm_pre_clip': 17.221351623535156, 'timing/data': 0.009521553991362453, 'timing/model': 1.147212411975488, 'learning_rate/qwen_vl_interface': 2.8873185364121965e-06, 'learning_rate/action_model': 2.8873185364121963e-05, 'epoch': 1.26}) +[2026-06-19 17:27:14,575][starVLA.training.train_starvla][INFO] - Step 3363, Loss: {'train/loss': 0.19079747796058655, 'train/grad_norm_pre_clip': 10.660484313964844, 'timing/data': 0.00938271195627749, 'timing/model': 1.1497645730851218, 'learning_rate/qwen_vl_interface': 2.8846770657289386e-06, 'learning_rate/action_model': 2.8846770657289386e-05, 'epoch': 1.26}) +[2026-06-19 17:27:19,204][starVLA.training.train_starvla][INFO] - Step 3364, Loss: {'train/loss': 0.2589777410030365, 'train/grad_norm_pre_clip': 13.812707901000977, 'timing/data': 0.00947311194613576, 'timing/model': 1.3354439249960706, 'learning_rate/qwen_vl_interface': 2.882036567340593e-06, 'learning_rate/action_model': 2.8820365673405932e-05, 'epoch': 1.26}) +[2026-06-19 17:27:23,643][starVLA.training.train_starvla][INFO] - Step 3365, Loss: {'train/loss': 0.22539137303829193, 'train/grad_norm_pre_clip': 6.827558994293213, 'timing/data': 0.009552890085615218, 'timing/model': 1.1474971850402653, 'learning_rate/qwen_vl_interface': 2.8793970423325684e-06, 'learning_rate/action_model': 2.8793970423325685e-05, 'epoch': 1.26}) +[2026-06-19 17:27:28,239][starVLA.training.train_starvla][INFO] - Step 3366, Loss: {'train/loss': 0.3233368396759033, 'train/grad_norm_pre_clip': 13.41176700592041, 'timing/data': 0.00969168997835368, 'timing/model': 1.1485318009508774, 'learning_rate/qwen_vl_interface': 2.8767584917898756e-06, 'learning_rate/action_model': 2.8767584917898754e-05, 'epoch': 1.26}) +[2026-06-19 17:27:32,837][starVLA.training.train_starvla][INFO] - Step 3367, Loss: {'train/loss': 0.33806365728378296, 'train/grad_norm_pre_clip': 9.07754898071289, 'timing/data': 0.00961102801375091, 'timing/model': 1.1512137550162151, 'learning_rate/qwen_vl_interface': 2.874120916797121e-06, 'learning_rate/action_model': 2.8741209167971207e-05, 'epoch': 1.26}) +[2026-06-19 17:27:37,367][starVLA.training.train_starvla][INFO] - Step 3368, Loss: {'train/loss': 0.3138023018836975, 'train/grad_norm_pre_clip': 9.162650108337402, 'timing/data': 0.009907630970701575, 'timing/model': 1.14754214102868, 'learning_rate/qwen_vl_interface': 2.8714843184385124e-06, 'learning_rate/action_model': 2.871484318438512e-05, 'epoch': 1.26}) +[2026-06-19 17:27:41,951][starVLA.training.train_starvla][INFO] - Step 3369, Loss: {'train/loss': 0.24350057542324066, 'train/grad_norm_pre_clip': 6.522498607635498, 'timing/data': 0.010023839073255658, 'timing/model': 1.1489454599795863, 'learning_rate/qwen_vl_interface': 2.8688486977978575e-06, 'learning_rate/action_model': 2.8688486977978572e-05, 'epoch': 1.26}) +[2026-06-19 17:27:46,558][starVLA.training.train_starvla][INFO] - Step 3370, Loss: {'train/loss': 0.32082149386405945, 'train/grad_norm_pre_clip': 14.73962116241455, 'timing/data': 0.009849137044511735, 'timing/model': 1.3216417890507728, 'learning_rate/qwen_vl_interface': 2.8662140559585584e-06, 'learning_rate/action_model': 2.866214055958558e-05, 'epoch': 1.26}) +[2026-06-19 17:27:50,993][starVLA.training.train_starvla][INFO] - Step 3371, Loss: {'train/loss': 0.12331584095954895, 'train/grad_norm_pre_clip': 6.438394069671631, 'timing/data': 0.0100195889826864, 'timing/model': 1.1469961430411786, 'learning_rate/qwen_vl_interface': 2.8635803940036162e-06, 'learning_rate/action_model': 2.863580394003616e-05, 'epoch': 1.26}) +[2026-06-19 17:27:55,605][starVLA.training.train_starvla][INFO] - Step 3372, Loss: {'train/loss': 0.3967725336551666, 'train/grad_norm_pre_clip': 8.969036102294922, 'timing/data': 0.009962324984371662, 'timing/model': 1.1465342099545524, 'learning_rate/qwen_vl_interface': 2.8609477130156317e-06, 'learning_rate/action_model': 2.8609477130156314e-05, 'epoch': 1.26}) +[2026-06-19 17:28:00,197][starVLA.training.train_starvla][INFO] - Step 3373, Loss: {'train/loss': 0.39346855878829956, 'train/grad_norm_pre_clip': 13.59175968170166, 'timing/data': 0.009413955034688115, 'timing/model': 1.1496888529509306, 'learning_rate/qwen_vl_interface': 2.8583160140767997e-06, 'learning_rate/action_model': 2.8583160140767994e-05, 'epoch': 1.26}) +[2026-06-19 17:28:04,750][starVLA.training.train_starvla][INFO] - Step 3374, Loss: {'train/loss': 0.4784475862979889, 'train/grad_norm_pre_clip': 11.376269340515137, 'timing/data': 0.010083268978632987, 'timing/model': 1.1510011749342084, 'learning_rate/qwen_vl_interface': 2.8556852982689125e-06, 'learning_rate/action_model': 2.8556852982689126e-05, 'epoch': 1.27}) +[2026-06-19 17:28:09,353][starVLA.training.train_starvla][INFO] - Step 3375, Loss: {'train/loss': 0.3395560383796692, 'train/grad_norm_pre_clip': 12.329784393310547, 'timing/data': 0.009580309968441725, 'timing/model': 1.1518494880292565, 'learning_rate/qwen_vl_interface': 2.8530555666733584e-06, 'learning_rate/action_model': 2.8530555666733583e-05, 'epoch': 1.27}) +[2026-06-19 17:28:13,964][starVLA.training.train_starvla][INFO] - Step 3376, Loss: {'train/loss': 0.19700786471366882, 'train/grad_norm_pre_clip': 10.329937934875488, 'timing/data': 0.01021478301845491, 'timing/model': 1.3154987039742991, 'learning_rate/qwen_vl_interface': 2.85042682037112e-06, 'learning_rate/action_model': 2.85042682037112e-05, 'epoch': 1.27}) +[2026-06-19 17:28:18,395][starVLA.training.train_starvla][INFO] - Step 3377, Loss: {'train/loss': 0.3173808157444, 'train/grad_norm_pre_clip': 9.858592987060547, 'timing/data': 0.009938957984559238, 'timing/model': 1.1496247650356963, 'learning_rate/qwen_vl_interface': 2.8477990604427764e-06, 'learning_rate/action_model': 2.8477990604427763e-05, 'epoch': 1.27}) +[2026-06-19 17:28:23,022][starVLA.training.train_starvla][INFO] - Step 3378, Loss: {'train/loss': 0.20118194818496704, 'train/grad_norm_pre_clip': 6.621124267578125, 'timing/data': 0.009314647992141545, 'timing/model': 1.1550065020564944, 'learning_rate/qwen_vl_interface': 2.8451722879684994e-06, 'learning_rate/action_model': 2.8451722879684994e-05, 'epoch': 1.27}) +[2026-06-19 17:28:27,627][starVLA.training.train_starvla][INFO] - Step 3379, Loss: {'train/loss': 0.18651315569877625, 'train/grad_norm_pre_clip': 10.653206825256348, 'timing/data': 0.009820334962569177, 'timing/model': 1.1490537849022076, 'learning_rate/qwen_vl_interface': 2.842546504028059e-06, 'learning_rate/action_model': 2.8425465040280586e-05, 'epoch': 1.27}) +[2026-06-19 17:28:32,181][starVLA.training.train_starvla][INFO] - Step 3380, Loss: {'train/loss': 0.3205989599227905, 'train/grad_norm_pre_clip': 12.698023796081543, 'timing/data': 0.009687623009085655, 'timing/model': 1.1484667939366773, 'learning_rate/qwen_vl_interface': 2.83992170970081e-06, 'learning_rate/action_model': 2.8399217097008102e-05, 'epoch': 1.27}) +[2026-06-19 17:28:36,805][starVLA.training.train_starvla][INFO] - Step 3381, Loss: {'train/loss': 0.17925922572612762, 'train/grad_norm_pre_clip': 5.653575897216797, 'timing/data': 0.010165530955418944, 'timing/model': 1.1468938110629097, 'learning_rate/qwen_vl_interface': 2.837297906065711e-06, 'learning_rate/action_model': 2.8372979060657107e-05, 'epoch': 1.27}) +[2026-06-19 17:28:41,419][starVLA.training.train_starvla][INFO] - Step 3382, Loss: {'train/loss': 0.32674330472946167, 'train/grad_norm_pre_clip': 11.613646507263184, 'timing/data': 0.00976735504809767, 'timing/model': 1.3152733329916373, 'learning_rate/qwen_vl_interface': 2.834675094201306e-06, 'learning_rate/action_model': 2.8346750942013063e-05, 'epoch': 1.27}) +[2026-06-19 17:28:45,871][starVLA.training.train_starvla][INFO] - Step 3383, Loss: {'train/loss': 0.33067503571510315, 'train/grad_norm_pre_clip': 11.286861419677734, 'timing/data': 0.00926168099977076, 'timing/model': 1.1482950769132003, 'learning_rate/qwen_vl_interface': 2.8320532751857365e-06, 'learning_rate/action_model': 2.8320532751857366e-05, 'epoch': 1.27}) +[2026-06-19 17:28:50,479][starVLA.training.train_starvla][INFO] - Step 3384, Loss: {'train/loss': 0.16615332663059235, 'train/grad_norm_pre_clip': 3.8772037029266357, 'timing/data': 0.0099781749304384, 'timing/model': 1.1476372230099514, 'learning_rate/qwen_vl_interface': 2.829432450096731e-06, 'learning_rate/action_model': 2.8294324500967306e-05, 'epoch': 1.27}) +[2026-06-19 17:28:55,108][starVLA.training.train_starvla][INFO] - Step 3385, Loss: {'train/loss': 0.18398475646972656, 'train/grad_norm_pre_clip': 7.645740032196045, 'timing/data': 0.010117272031493485, 'timing/model': 1.1479546240298077, 'learning_rate/qwen_vl_interface': 2.8268126200116118e-06, 'learning_rate/action_model': 2.8268126200116117e-05, 'epoch': 1.27}) +[2026-06-19 17:28:59,713][starVLA.training.train_starvla][INFO] - Step 3386, Loss: {'train/loss': 0.33353549242019653, 'train/grad_norm_pre_clip': 15.967446327209473, 'timing/data': 0.010286616045050323, 'timing/model': 1.1456179689848796, 'learning_rate/qwen_vl_interface': 2.824193786007294e-06, 'learning_rate/action_model': 2.8241937860072937e-05, 'epoch': 1.27}) +[2026-06-19 17:29:04,326][starVLA.training.train_starvla][INFO] - Step 3387, Loss: {'train/loss': 0.34091290831565857, 'train/grad_norm_pre_clip': 13.93945598602295, 'timing/data': 0.009534912998788059, 'timing/model': 1.148603861918673, 'learning_rate/qwen_vl_interface': 2.821575949160279e-06, 'learning_rate/action_model': 2.8215759491602787e-05, 'epoch': 1.27}) +[2026-06-19 17:29:08,921][starVLA.training.train_starvla][INFO] - Step 3388, Loss: {'train/loss': 0.19605989754199982, 'train/grad_norm_pre_clip': 5.559347629547119, 'timing/data': 0.009886470041237772, 'timing/model': 1.3140954039990902, 'learning_rate/qwen_vl_interface': 2.8189591105466625e-06, 'learning_rate/action_model': 2.8189591105466624e-05, 'epoch': 1.27}) +[2026-06-19 17:29:13,350][starVLA.training.train_starvla][INFO] - Step 3389, Loss: {'train/loss': 0.29185807704925537, 'train/grad_norm_pre_clip': 9.87132740020752, 'timing/data': 0.009931509965099394, 'timing/model': 1.1476900180568919, 'learning_rate/qwen_vl_interface': 2.816343271242126e-06, 'learning_rate/action_model': 2.816343271242126e-05, 'epoch': 1.27}) +[2026-06-19 17:29:17,951][starVLA.training.train_starvla][INFO] - Step 3390, Loss: {'train/loss': 0.26380833983421326, 'train/grad_norm_pre_clip': 8.83696460723877, 'timing/data': 0.013673298060894012, 'timing/model': 1.147485587047413, 'learning_rate/qwen_vl_interface': 2.8137284323219467e-06, 'learning_rate/action_model': 2.8137284323219464e-05, 'epoch': 1.27}) +[2026-06-19 17:29:22,467][starVLA.training.train_starvla][INFO] - Step 3391, Loss: {'train/loss': 0.21850678324699402, 'train/grad_norm_pre_clip': 4.943713188171387, 'timing/data': 0.009643818018957973, 'timing/model': 1.1508059770567343, 'learning_rate/qwen_vl_interface': 2.81111459486098e-06, 'learning_rate/action_model': 2.81111459486098e-05, 'epoch': 1.27}) +[2026-06-19 17:29:27,042][starVLA.training.train_starvla][INFO] - Step 3392, Loss: {'train/loss': 0.22659370303153992, 'train/grad_norm_pre_clip': 9.146728515625, 'timing/data': 0.009628635016269982, 'timing/model': 1.1603126339614391, 'learning_rate/qwen_vl_interface': 2.8085017599336806e-06, 'learning_rate/action_model': 2.8085017599336806e-05, 'epoch': 1.27}) +[2026-06-19 17:29:31,645][starVLA.training.train_starvla][INFO] - Step 3393, Loss: {'train/loss': 0.4533335268497467, 'train/grad_norm_pre_clip': 10.47661018371582, 'timing/data': 0.00946784799452871, 'timing/model': 1.1534150660736486, 'learning_rate/qwen_vl_interface': 2.8058899286140828e-06, 'learning_rate/action_model': 2.805889928614083e-05, 'epoch': 1.27}) +[2026-06-19 17:29:36,281][starVLA.training.train_starvla][INFO] - Step 3394, Loss: {'train/loss': 0.18974146246910095, 'train/grad_norm_pre_clip': 6.246126174926758, 'timing/data': 0.009738057036884129, 'timing/model': 1.3160941149108112, 'learning_rate/qwen_vl_interface': 2.8032791019758127e-06, 'learning_rate/action_model': 2.8032791019758127e-05, 'epoch': 1.27}) +[2026-06-19 17:29:40,714][starVLA.training.train_starvla][INFO] - Step 3395, Loss: {'train/loss': 0.19089145958423615, 'train/grad_norm_pre_clip': 7.736780166625977, 'timing/data': 0.009602478006854653, 'timing/model': 1.1514785489998758, 'learning_rate/qwen_vl_interface': 2.800669281092084e-06, 'learning_rate/action_model': 2.8006692810920838e-05, 'epoch': 1.27}) +[2026-06-19 17:29:45,325][starVLA.training.train_starvla][INFO] - Step 3396, Loss: {'train/loss': 0.5861347913742065, 'train/grad_norm_pre_clip': 12.55454158782959, 'timing/data': 0.009105785982683301, 'timing/model': 1.155676283990033, 'learning_rate/qwen_vl_interface': 2.7980604670356938e-06, 'learning_rate/action_model': 2.7980604670356937e-05, 'epoch': 1.27}) +[2026-06-19 17:29:49,933][starVLA.training.train_starvla][INFO] - Step 3397, Loss: {'train/loss': 0.2578660845756531, 'train/grad_norm_pre_clip': 15.991650581359863, 'timing/data': 0.009843036998063326, 'timing/model': 1.1552484349813312, 'learning_rate/qwen_vl_interface': 2.795452660879027e-06, 'learning_rate/action_model': 2.7954526608790273e-05, 'epoch': 1.27}) +[2026-06-19 17:29:54,546][starVLA.training.train_starvla][INFO] - Step 3398, Loss: {'train/loss': 0.2637486755847931, 'train/grad_norm_pre_clip': 7.716123104095459, 'timing/data': 0.009659803006798029, 'timing/model': 1.1504419359844178, 'learning_rate/qwen_vl_interface': 2.7928458636940565e-06, 'learning_rate/action_model': 2.7928458636940564e-05, 'epoch': 1.27}) +[2026-06-19 17:29:59,141][starVLA.training.train_starvla][INFO] - Step 3399, Loss: {'train/loss': 0.2779954671859741, 'train/grad_norm_pre_clip': 7.06450891494751, 'timing/data': 0.010021241963841021, 'timing/model': 1.1490667239995673, 'learning_rate/qwen_vl_interface': 2.790240076552336e-06, 'learning_rate/action_model': 2.7902400765523355e-05, 'epoch': 1.27}) +[2026-06-19 17:30:03,761][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 3400 +[2026-06-19 17:32:56,776][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 3400 +[2026-06-19 17:32:56,776][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 3400 +[2026-06-19 17:36:02,874][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 3400 +[2026-06-19 17:36:02,875][starVLA.training.train_starvla][INFO] - Step 3400, Loss: {'train/loss': 0.17753693461418152, 'train/grad_norm_pre_clip': 5.282175540924072, 'eval/loss': 0.16299061477184296, 'eval/latency_0/loss': 0.0873141810297966, 'eval/latency_1/loss': 0.05843529477715492, 'eval/latency_2/loss': 0.1183767020702362, 'eval/latency_3/loss': 0.19354142248630524, 'eval/latency_4/loss': 0.3585728704929352, 'eval/flappy/loss': 0.16299061477184296, 'eval/flappy/latency_0/loss': 0.0873141810297966, 'eval/flappy/latency_1/loss': 0.05843529477715492, 'eval/flappy/latency_2/loss': 0.1183767020702362, 'eval/flappy/latency_3/loss': 0.19354142248630524, 'eval/flappy/latency_4/loss': 0.3585728704929352, 'eval/flap_precision': 0.6907216494845361, 'eval/flap_recall': 0.730909090909091, 'eval/flap_f1': 0.7102473498233216, 'eval/flap_rate_model': 0.08083333333333333, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.7102473498233216, 'eval/latency_0/flap_precision': 0.6907216494845361, 'eval/latency_0/flap_recall': 0.730909090909091, 'eval/latency_0/flap_f1': 0.7102473498233216, 'eval/latency_0/flap_rate_model': 0.08083333333333333, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.7102473498233216, 'timing/data': 0.009324038052000105, 'timing/model': 1.3078005471033975, 'learning_rate/qwen_vl_interface': 2.787635300525006e-06, 'learning_rate/action_model': 2.787635300525006e-05, 'epoch': 1.27}) +[2026-06-19 17:36:02,875][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_3400_state +[2026-06-19 17:36:02,878][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-19 17:36:06,169][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_3400_state/model.safetensors +[2026-06-19 17:36:13,383][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_3400_state/optimizer.bin +[2026-06-19 17:36:13,383][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_3400_state/sampler.bin +[2026-06-19 17:36:13,383][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_3400_state/sampler_1.bin +[2026-06-19 17:36:13,384][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_3400_state/random_states_0.pkl +[2026-06-19 17:36:13,406][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3400 +[2026-06-19 17:36:13,406][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-19 17:36:13,415][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-19 17:36:17,934][starVLA.training.train_starvla][INFO] - Step 3401, Loss: {'train/loss': 0.28881996870040894, 'train/grad_norm_pre_clip': 9.997486114501953, 'timing/data': 0.01057556294836104, 'timing/model': 1.133701249025762, 'learning_rate/qwen_vl_interface': 2.7850315366827933e-06, 'learning_rate/action_model': 2.7850315366827934e-05, 'epoch': 1.28}) +[2026-06-19 17:36:22,531][starVLA.training.train_starvla][INFO] - Step 3402, Loss: {'train/loss': 0.25722748041152954, 'train/grad_norm_pre_clip': 9.256355285644531, 'timing/data': 0.009295014082454145, 'timing/model': 1.13583240902517, 'learning_rate/qwen_vl_interface': 2.7824287860960076e-06, 'learning_rate/action_model': 2.7824287860960075e-05, 'epoch': 1.28}) +[2026-06-19 17:36:27,081][starVLA.training.train_starvla][INFO] - Step 3403, Loss: {'train/loss': 0.2727870047092438, 'train/grad_norm_pre_clip': 7.6082377433776855, 'timing/data': 0.010411246097646654, 'timing/model': 1.297009734902531, 'learning_rate/qwen_vl_interface': 2.7798270498345377e-06, 'learning_rate/action_model': 2.7798270498345376e-05, 'epoch': 1.28}) +[2026-06-19 17:36:31,475][starVLA.training.train_starvla][INFO] - Step 3404, Loss: {'train/loss': 0.18279239535331726, 'train/grad_norm_pre_clip': 7.837735652923584, 'timing/data': 0.009292764007113874, 'timing/model': 1.1471144360257313, 'learning_rate/qwen_vl_interface': 2.777226328967861e-06, 'learning_rate/action_model': 2.7772263289678607e-05, 'epoch': 1.28}) +[2026-06-19 17:36:36,042][starVLA.training.train_starvla][INFO] - Step 3405, Loss: {'train/loss': 0.1893334537744522, 'train/grad_norm_pre_clip': 6.81740665435791, 'timing/data': 0.009874077048152685, 'timing/model': 1.1406085880007595, 'learning_rate/qwen_vl_interface': 2.774626624565035e-06, 'learning_rate/action_model': 2.774626624565035e-05, 'epoch': 1.28}) +[2026-06-19 17:36:40,611][starVLA.training.train_starvla][INFO] - Step 3406, Loss: {'train/loss': 0.21742607653141022, 'train/grad_norm_pre_clip': 7.462402820587158, 'timing/data': 0.009637415991164744, 'timing/model': 1.1391659940127283, 'learning_rate/qwen_vl_interface': 2.772027937694701e-06, 'learning_rate/action_model': 2.772027937694701e-05, 'epoch': 1.28}) +[2026-06-19 17:36:45,106][starVLA.training.train_starvla][INFO] - Step 3407, Loss: {'train/loss': 0.18764062225818634, 'train/grad_norm_pre_clip': 6.060459613800049, 'timing/data': 0.01000294100958854, 'timing/model': 1.154979494982399, 'learning_rate/qwen_vl_interface': 2.7694302694250806e-06, 'learning_rate/action_model': 2.7694302694250808e-05, 'epoch': 1.28}) +[2026-06-19 17:36:49,611][starVLA.training.train_starvla][INFO] - Step 3408, Loss: {'train/loss': 0.20017960667610168, 'train/grad_norm_pre_clip': 9.340689659118652, 'timing/data': 0.009593917056918144, 'timing/model': 1.1452335459180176, 'learning_rate/qwen_vl_interface': 2.7668336208239775e-06, 'learning_rate/action_model': 2.7668336208239776e-05, 'epoch': 1.28}) +[2026-06-19 17:36:54,131][starVLA.training.train_starvla][INFO] - Step 3409, Loss: {'train/loss': 0.2514379024505615, 'train/grad_norm_pre_clip': 14.892524719238281, 'timing/data': 0.010005058953538537, 'timing/model': 1.2534203079994768, 'learning_rate/qwen_vl_interface': 2.7642379929587755e-06, 'learning_rate/action_model': 2.7642379929587754e-05, 'epoch': 1.28}) +[2026-06-19 17:36:58,577][starVLA.training.train_starvla][INFO] - Step 3410, Loss: {'train/loss': 0.1471909135580063, 'train/grad_norm_pre_clip': 6.39734411239624, 'timing/data': 0.009068591054528952, 'timing/model': 1.1427162400214002, 'learning_rate/qwen_vl_interface': 2.7616433868964387e-06, 'learning_rate/action_model': 2.7616433868964388e-05, 'epoch': 1.28}) +[2026-06-19 17:37:03,185][starVLA.training.train_starvla][INFO] - Step 3411, Loss: {'train/loss': 0.3220805525779724, 'train/grad_norm_pre_clip': 8.46319580078125, 'timing/data': 0.009724581963382661, 'timing/model': 1.1422640240052715, 'learning_rate/qwen_vl_interface': 2.7590498037035122e-06, 'learning_rate/action_model': 2.7590498037035124e-05, 'epoch': 1.28}) +[2026-06-19 17:37:07,795][starVLA.training.train_starvla][INFO] - Step 3412, Loss: {'train/loss': 0.35000482201576233, 'train/grad_norm_pre_clip': 15.929335594177246, 'timing/data': 0.009840282960794866, 'timing/model': 1.151573378010653, 'learning_rate/qwen_vl_interface': 2.7564572444461218e-06, 'learning_rate/action_model': 2.7564572444461217e-05, 'epoch': 1.28}) +[2026-06-19 17:37:12,386][starVLA.training.train_starvla][INFO] - Step 3413, Loss: {'train/loss': 0.3415144681930542, 'train/grad_norm_pre_clip': 8.78553295135498, 'timing/data': 0.00941050797700882, 'timing/model': 1.1500138529809192, 'learning_rate/qwen_vl_interface': 2.7538657101899674e-06, 'learning_rate/action_model': 2.7538657101899673e-05, 'epoch': 1.28}) +[2026-06-19 17:37:16,988][starVLA.training.train_starvla][INFO] - Step 3414, Loss: {'train/loss': 0.2774481475353241, 'train/grad_norm_pre_clip': 7.531935214996338, 'timing/data': 0.009536588098853827, 'timing/model': 1.327819949015975, 'learning_rate/qwen_vl_interface': 2.751275202000332e-06, 'learning_rate/action_model': 2.7512752020003317e-05, 'epoch': 1.28}) +[2026-06-19 17:37:21,418][starVLA.training.train_starvla][INFO] - Step 3415, Loss: {'train/loss': 0.2908443212509155, 'train/grad_norm_pre_clip': 9.781150817871094, 'timing/data': 0.009836929966695607, 'timing/model': 1.1529832330998033, 'learning_rate/qwen_vl_interface': 2.7486857209420763e-06, 'learning_rate/action_model': 2.7486857209420763e-05, 'epoch': 1.28}) +[2026-06-19 17:37:26,014][starVLA.training.train_starvla][INFO] - Step 3416, Loss: {'train/loss': 0.27333560585975647, 'train/grad_norm_pre_clip': 13.963062286376953, 'timing/data': 0.009412952000275254, 'timing/model': 1.1441107969731092, 'learning_rate/qwen_vl_interface': 2.7460972680796377e-06, 'learning_rate/action_model': 2.7460972680796375e-05, 'epoch': 1.28}) +[2026-06-19 17:37:30,614][starVLA.training.train_starvla][INFO] - Step 3417, Loss: {'train/loss': 0.2200823426246643, 'train/grad_norm_pre_clip': 7.490838527679443, 'timing/data': 0.009676660061813891, 'timing/model': 1.152819321025163, 'learning_rate/qwen_vl_interface': 2.7435098444770314e-06, 'learning_rate/action_model': 2.7435098444770315e-05, 'epoch': 1.28}) +[2026-06-19 17:37:35,225][starVLA.training.train_starvla][INFO] - Step 3418, Loss: {'train/loss': 0.12251546233892441, 'train/grad_norm_pre_clip': 6.538758754730225, 'timing/data': 0.009456790052354336, 'timing/model': 1.1623273050645366, 'learning_rate/qwen_vl_interface': 2.7409234511978495e-06, 'learning_rate/action_model': 2.7409234511978492e-05, 'epoch': 1.28}) +[2026-06-19 17:37:39,834][starVLA.training.train_starvla][INFO] - Step 3419, Loss: {'train/loss': 0.3658201992511749, 'train/grad_norm_pre_clip': 11.227808952331543, 'timing/data': 0.009401395916938782, 'timing/model': 1.1547448970377445, 'learning_rate/qwen_vl_interface': 2.7383380893052606e-06, 'learning_rate/action_model': 2.7383380893052608e-05, 'epoch': 1.28}) +[2026-06-19 17:37:44,437][starVLA.training.train_starvla][INFO] - Step 3420, Loss: {'train/loss': 0.2544572353363037, 'train/grad_norm_pre_clip': 7.724748611450195, 'timing/data': 0.009115130989812315, 'timing/model': 1.3112223820062354, 'learning_rate/qwen_vl_interface': 2.73575375986201e-06, 'learning_rate/action_model': 2.7357537598620096e-05, 'epoch': 1.28}) +[2026-06-19 17:37:49,001][starVLA.training.train_starvla][INFO] - Step 3421, Loss: {'train/loss': 0.24413368105888367, 'train/grad_norm_pre_clip': 7.419518947601318, 'timing/data': 0.009800520958378911, 'timing/model': 1.2792159899836406, 'learning_rate/qwen_vl_interface': 2.733170463930417e-06, 'learning_rate/action_model': 2.7331704639304167e-05, 'epoch': 1.28}) +[2026-06-19 17:37:53,430][starVLA.training.train_starvla][INFO] - Step 3422, Loss: {'train/loss': 0.2830122113227844, 'train/grad_norm_pre_clip': 11.060410499572754, 'timing/data': 0.00975244294386357, 'timing/model': 1.1434237860376015, 'learning_rate/qwen_vl_interface': 2.7305882025723774e-06, 'learning_rate/action_model': 2.7305882025723774e-05, 'epoch': 1.28}) +[2026-06-19 17:37:58,026][starVLA.training.train_starvla][INFO] - Step 3423, Loss: {'train/loss': 0.1115790605545044, 'train/grad_norm_pre_clip': 4.754920959472656, 'timing/data': 0.010001637041568756, 'timing/model': 1.1396412489702925, 'learning_rate/qwen_vl_interface': 2.7280069768493626e-06, 'learning_rate/action_model': 2.7280069768493628e-05, 'epoch': 1.28}) +[2026-06-19 17:38:02,605][starVLA.training.train_starvla][INFO] - Step 3424, Loss: {'train/loss': 0.34353160858154297, 'train/grad_norm_pre_clip': 11.54948902130127, 'timing/data': 0.010005873045884073, 'timing/model': 1.1476377380313352, 'learning_rate/qwen_vl_interface': 2.7254267878224144e-06, 'learning_rate/action_model': 2.7254267878224143e-05, 'epoch': 1.28}) +[2026-06-19 17:38:07,169][starVLA.training.train_starvla][INFO] - Step 3425, Loss: {'train/loss': 0.1954227238893509, 'train/grad_norm_pre_clip': 5.724894046783447, 'timing/data': 0.009496156009845436, 'timing/model': 1.143844480975531, 'learning_rate/qwen_vl_interface': 2.7228476365521524e-06, 'learning_rate/action_model': 2.722847636552152e-05, 'epoch': 1.28}) +[2026-06-19 17:38:11,761][starVLA.training.train_starvla][INFO] - Step 3426, Loss: {'train/loss': 0.24168424308300018, 'train/grad_norm_pre_clip': 13.920923233032227, 'timing/data': 0.009676344925537705, 'timing/model': 1.3198443710571155, 'learning_rate/qwen_vl_interface': 2.7202695240987674e-06, 'learning_rate/action_model': 2.7202695240987676e-05, 'epoch': 1.28}) +[2026-06-19 17:38:16,311][starVLA.training.train_starvla][INFO] - Step 3427, Loss: {'train/loss': 0.15074457228183746, 'train/grad_norm_pre_clip': 7.710803985595703, 'timing/data': 0.009940739953890443, 'timing/model': 1.2663405359489843, 'learning_rate/qwen_vl_interface': 2.717692451522025e-06, 'learning_rate/action_model': 2.7176924515220248e-05, 'epoch': 1.28}) +[2026-06-19 17:38:20,732][starVLA.training.train_starvla][INFO] - Step 3428, Loss: {'train/loss': 0.28041326999664307, 'train/grad_norm_pre_clip': 15.143375396728516, 'timing/data': 0.009483783971518278, 'timing/model': 1.1454600139986724, 'learning_rate/qwen_vl_interface': 2.7151164198812624e-06, 'learning_rate/action_model': 2.7151164198812622e-05, 'epoch': 1.29}) +[2026-06-19 17:38:25,313][starVLA.training.train_starvla][INFO] - Step 3429, Loss: {'train/loss': 0.426556795835495, 'train/grad_norm_pre_clip': 12.926898956298828, 'timing/data': 0.009702237090095878, 'timing/model': 1.1441190920304507, 'learning_rate/qwen_vl_interface': 2.712541430235388e-06, 'learning_rate/action_model': 2.712541430235388e-05, 'epoch': 1.29}) +[2026-06-19 17:38:29,890][starVLA.training.train_starvla][INFO] - Step 3430, Loss: {'train/loss': 0.21119160950183868, 'train/grad_norm_pre_clip': 11.515655517578125, 'timing/data': 0.009611875982955098, 'timing/model': 1.143396214931272, 'learning_rate/qwen_vl_interface': 2.7099674836428842e-06, 'learning_rate/action_model': 2.709967483642884e-05, 'epoch': 1.29}) +[2026-06-19 17:38:34,472][starVLA.training.train_starvla][INFO] - Step 3431, Loss: {'train/loss': 0.20556752383708954, 'train/grad_norm_pre_clip': 8.331497192382812, 'timing/data': 0.009458330925554037, 'timing/model': 1.1436860440298915, 'learning_rate/qwen_vl_interface': 2.7073945811618018e-06, 'learning_rate/action_model': 2.7073945811618017e-05, 'epoch': 1.29}) +[2026-06-19 17:38:39,059][starVLA.training.train_starvla][INFO] - Step 3432, Loss: {'train/loss': 0.48826920986175537, 'train/grad_norm_pre_clip': 29.07466697692871, 'timing/data': 0.009754622937180102, 'timing/model': 1.156541328993626, 'learning_rate/qwen_vl_interface': 2.704822723849765e-06, 'learning_rate/action_model': 2.7048227238497648e-05, 'epoch': 1.29}) +[2026-06-19 17:38:43,662][starVLA.training.train_starvla][INFO] - Step 3433, Loss: {'train/loss': 0.18313592672348022, 'train/grad_norm_pre_clip': 9.992539405822754, 'timing/data': 0.009596548043191433, 'timing/model': 1.3076140659395605, 'learning_rate/qwen_vl_interface': 2.7022519127639687e-06, 'learning_rate/action_model': 2.7022519127639682e-05, 'epoch': 1.29}) +[2026-06-19 17:38:48,097][starVLA.training.train_starvla][INFO] - Step 3434, Loss: {'train/loss': 0.2002551108598709, 'train/grad_norm_pre_clip': 13.130220413208008, 'timing/data': 0.00941878103185445, 'timing/model': 1.1467844480648637, 'learning_rate/qwen_vl_interface': 2.6996821489611713e-06, 'learning_rate/action_model': 2.699682148961171e-05, 'epoch': 1.29}) +[2026-06-19 17:38:52,712][starVLA.training.train_starvla][INFO] - Step 3435, Loss: {'train/loss': 0.3299109637737274, 'train/grad_norm_pre_clip': 13.31912612915039, 'timing/data': 0.0099628180032596, 'timing/model': 1.1464011949719861, 'learning_rate/qwen_vl_interface': 2.697113433497712e-06, 'learning_rate/action_model': 2.6971134334977115e-05, 'epoch': 1.29}) +[2026-06-19 17:38:57,304][starVLA.training.train_starvla][INFO] - Step 3436, Loss: {'train/loss': 0.3150385320186615, 'train/grad_norm_pre_clip': 11.862547874450684, 'timing/data': 0.009410802042111754, 'timing/model': 1.1508797999704257, 'learning_rate/qwen_vl_interface': 2.6945457674294863e-06, 'learning_rate/action_model': 2.694545767429486e-05, 'epoch': 1.29}) +[2026-06-19 17:39:01,885][starVLA.training.train_starvla][INFO] - Step 3437, Loss: {'train/loss': 0.42405033111572266, 'train/grad_norm_pre_clip': 32.264404296875, 'timing/data': 0.009726229007355869, 'timing/model': 1.142084961058572, 'learning_rate/qwen_vl_interface': 2.6919791518119676e-06, 'learning_rate/action_model': 2.6919791518119674e-05, 'epoch': 1.29}) +[2026-06-19 17:39:06,490][starVLA.training.train_starvla][INFO] - Step 3438, Loss: {'train/loss': 0.2129659205675125, 'train/grad_norm_pre_clip': 8.77947998046875, 'timing/data': 0.009469627984799445, 'timing/model': 1.1526300419354811, 'learning_rate/qwen_vl_interface': 2.6894135877001946e-06, 'learning_rate/action_model': 2.689413587700195e-05, 'epoch': 1.29}) +[2026-06-19 17:39:11,069][starVLA.training.train_starvla][INFO] - Step 3439, Loss: {'train/loss': 0.23766963183879852, 'train/grad_norm_pre_clip': 11.267007827758789, 'timing/data': 0.009911512956023216, 'timing/model': 1.297356029972434, 'learning_rate/qwen_vl_interface': 2.686849076148775e-06, 'learning_rate/action_model': 2.6868490761487747e-05, 'epoch': 1.29}) +[2026-06-19 17:39:15,487][starVLA.training.train_starvla][INFO] - Step 3440, Loss: {'train/loss': 0.1821822226047516, 'train/grad_norm_pre_clip': 11.29256534576416, 'timing/data': 0.009749372024089098, 'timing/model': 1.1432379829930142, 'learning_rate/qwen_vl_interface': 2.68428561821188e-06, 'learning_rate/action_model': 2.6842856182118802e-05, 'epoch': 1.29}) +[2026-06-19 17:39:20,068][starVLA.training.train_starvla][INFO] - Step 3441, Loss: {'train/loss': 0.14686144888401031, 'train/grad_norm_pre_clip': 15.078028678894043, 'timing/data': 0.00954603694844991, 'timing/model': 1.146777804940939, 'learning_rate/qwen_vl_interface': 2.681723214943253e-06, 'learning_rate/action_model': 2.681723214943253e-05, 'epoch': 1.29}) +[2026-06-19 17:39:24,582][starVLA.training.train_starvla][INFO] - Step 3442, Loss: {'train/loss': 0.3102461099624634, 'train/grad_norm_pre_clip': 15.041655540466309, 'timing/data': 0.008798748021945357, 'timing/model': 1.1520505419466645, 'learning_rate/qwen_vl_interface': 2.6791618673961996e-06, 'learning_rate/action_model': 2.6791618673961994e-05, 'epoch': 1.29}) +[2026-06-19 17:39:29,120][starVLA.training.train_starvla][INFO] - Step 3443, Loss: {'train/loss': 0.1969921886920929, 'train/grad_norm_pre_clip': 11.120804786682129, 'timing/data': 0.00929441791959107, 'timing/model': 1.145599342067726, 'learning_rate/qwen_vl_interface': 2.676601576623594e-06, 'learning_rate/action_model': 2.6766015766235937e-05, 'epoch': 1.29}) +[2026-06-19 17:39:33,714][starVLA.training.train_starvla][INFO] - Step 3444, Loss: {'train/loss': 0.30093953013420105, 'train/grad_norm_pre_clip': 14.241942405700684, 'timing/data': 0.009747063042595983, 'timing/model': 1.147728596930392, 'learning_rate/qwen_vl_interface': 2.674042343677874e-06, 'learning_rate/action_model': 2.6740423436778738e-05, 'epoch': 1.29}) +[2026-06-19 17:39:38,338][starVLA.training.train_starvla][INFO] - Step 3445, Loss: {'train/loss': 0.3394981920719147, 'train/grad_norm_pre_clip': 16.871000289916992, 'timing/data': 0.010203354991972446, 'timing/model': 1.3343539859633893, 'learning_rate/qwen_vl_interface': 2.6714841696110456e-06, 'learning_rate/action_model': 2.6714841696110456e-05, 'epoch': 1.29}) +[2026-06-19 17:39:42,783][starVLA.training.train_starvla][INFO] - Step 3446, Loss: {'train/loss': 0.1600915789604187, 'train/grad_norm_pre_clip': 10.380248069763184, 'timing/data': 0.009771697106771171, 'timing/model': 1.1438822040800005, 'learning_rate/qwen_vl_interface': 2.6689270554746743e-06, 'learning_rate/action_model': 2.6689270554746743e-05, 'epoch': 1.29}) +[2026-06-19 17:39:47,403][starVLA.training.train_starvla][INFO] - Step 3447, Loss: {'train/loss': 0.1257934868335724, 'train/grad_norm_pre_clip': 6.529753684997559, 'timing/data': 0.009348983061499894, 'timing/model': 1.1442498689284548, 'learning_rate/qwen_vl_interface': 2.6663710023198956e-06, 'learning_rate/action_model': 2.6663710023198956e-05, 'epoch': 1.29}) +[2026-06-19 17:39:52,018][starVLA.training.train_starvla][INFO] - Step 3448, Loss: {'train/loss': 0.35882532596588135, 'train/grad_norm_pre_clip': 17.691287994384766, 'timing/data': 0.009760390967130661, 'timing/model': 1.1520602660020813, 'learning_rate/qwen_vl_interface': 2.663816011197405e-06, 'learning_rate/action_model': 2.6638160111974048e-05, 'epoch': 1.29}) +[2026-06-19 17:39:56,573][starVLA.training.train_starvla][INFO] - Step 3449, Loss: {'train/loss': 0.2441588044166565, 'train/grad_norm_pre_clip': 15.261761665344238, 'timing/data': 0.009281128994189203, 'timing/model': 1.1521819139597937, 'learning_rate/qwen_vl_interface': 2.661262083157463e-06, 'learning_rate/action_model': 2.661262083157463e-05, 'epoch': 1.29}) +[2026-06-19 17:40:01,205][starVLA.training.train_starvla][INFO] - Step 3450, Loss: {'train/loss': 0.1432473212480545, 'train/grad_norm_pre_clip': 6.8280816078186035, 'timing/data': 0.00959260098170489, 'timing/model': 1.1450301859294996, 'learning_rate/qwen_vl_interface': 2.6587092192498943e-06, 'learning_rate/action_model': 2.658709219249894e-05, 'epoch': 1.29}) +[2026-06-19 17:40:05,829][starVLA.training.train_starvla][INFO] - Step 3451, Loss: {'train/loss': 0.1807655543088913, 'train/grad_norm_pre_clip': 12.041007041931152, 'timing/data': 0.0097000579116866, 'timing/model': 1.3287603328935802, 'learning_rate/qwen_vl_interface': 2.656157420524084e-06, 'learning_rate/action_model': 2.6561574205240842e-05, 'epoch': 1.29}) +[2026-06-19 17:40:10,263][starVLA.training.train_starvla][INFO] - Step 3452, Loss: {'train/loss': 0.2027217447757721, 'train/grad_norm_pre_clip': 10.31461238861084, 'timing/data': 0.00971302599646151, 'timing/model': 1.1576148089952767, 'learning_rate/qwen_vl_interface': 2.6536066880289795e-06, 'learning_rate/action_model': 2.6536066880289794e-05, 'epoch': 1.29}) +[2026-06-19 17:40:14,886][starVLA.training.train_starvla][INFO] - Step 3453, Loss: {'train/loss': 0.5114582180976868, 'train/grad_norm_pre_clip': 15.158031463623047, 'timing/data': 0.009801507112570107, 'timing/model': 1.1459838090231642, 'learning_rate/qwen_vl_interface': 2.6510570228130916e-06, 'learning_rate/action_model': 2.6510570228130914e-05, 'epoch': 1.29}) +[2026-06-19 17:40:19,516][starVLA.training.train_starvla][INFO] - Step 3454, Loss: {'train/loss': 0.2490699142217636, 'train/grad_norm_pre_clip': 15.97643756866455, 'timing/data': 0.009781294967979193, 'timing/model': 1.1435648140031844, 'learning_rate/qwen_vl_interface': 2.648508425924491e-06, 'learning_rate/action_model': 2.6485084259244907e-05, 'epoch': 1.3}) +[2026-06-19 17:40:24,064][starVLA.training.train_starvla][INFO] - Step 3455, Loss: {'train/loss': 0.1773356795310974, 'train/grad_norm_pre_clip': 10.200469970703125, 'timing/data': 0.00953477993607521, 'timing/model': 1.1444915489992127, 'learning_rate/qwen_vl_interface': 2.6459608984108093e-06, 'learning_rate/action_model': 2.645960898410809e-05, 'epoch': 1.3}) +[2026-06-19 17:40:28,697][starVLA.training.train_starvla][INFO] - Step 3456, Loss: {'train/loss': 0.2609093189239502, 'train/grad_norm_pre_clip': 17.08894157409668, 'timing/data': 0.010047670919448137, 'timing/model': 1.1635785829275846, 'learning_rate/qwen_vl_interface': 2.6434144413192414e-06, 'learning_rate/action_model': 2.6434144413192414e-05, 'epoch': 1.3}) +[2026-06-19 17:40:33,322][starVLA.training.train_starvla][INFO] - Step 3457, Loss: {'train/loss': 0.41376641392707825, 'train/grad_norm_pre_clip': 13.036958694458008, 'timing/data': 0.009581718011759222, 'timing/model': 1.3369663869962096, 'learning_rate/qwen_vl_interface': 2.6408690556965347e-06, 'learning_rate/action_model': 2.640869055696535e-05, 'epoch': 1.3}) +[2026-06-19 17:40:37,751][starVLA.training.train_starvla][INFO] - Step 3458, Loss: {'train/loss': 0.2756645381450653, 'train/grad_norm_pre_clip': 11.446657180786133, 'timing/data': 0.009447486023418605, 'timing/model': 1.1480769430054352, 'learning_rate/qwen_vl_interface': 2.6383247425890047e-06, 'learning_rate/action_model': 2.6383247425890047e-05, 'epoch': 1.3}) +[2026-06-19 17:40:42,392][starVLA.training.train_starvla][INFO] - Step 3459, Loss: {'train/loss': 0.31691670417785645, 'train/grad_norm_pre_clip': 25.564373016357422, 'timing/data': 0.009281439939513803, 'timing/model': 1.1471403669565916, 'learning_rate/qwen_vl_interface': 2.6357815030425204e-06, 'learning_rate/action_model': 2.6357815030425205e-05, 'epoch': 1.3}) +[2026-06-19 17:40:47,023][starVLA.training.train_starvla][INFO] - Step 3460, Loss: {'train/loss': 0.3702569603919983, 'train/grad_norm_pre_clip': 21.262060165405273, 'timing/data': 0.009982174960896373, 'timing/model': 1.1755322300596163, 'learning_rate/qwen_vl_interface': 2.6332393381025132e-06, 'learning_rate/action_model': 2.6332393381025133e-05, 'epoch': 1.3}) +[2026-06-19 17:40:51,658][starVLA.training.train_starvla][INFO] - Step 3461, Loss: {'train/loss': 0.5654062032699585, 'train/grad_norm_pre_clip': 20.291790008544922, 'timing/data': 0.009351635002531111, 'timing/model': 1.1495646969415247, 'learning_rate/qwen_vl_interface': 2.6306982488139704e-06, 'learning_rate/action_model': 2.6306982488139702e-05, 'epoch': 1.3}) +[2026-06-19 17:40:56,262][starVLA.training.train_starvla][INFO] - Step 3462, Loss: {'train/loss': 0.14629803597927094, 'train/grad_norm_pre_clip': 7.172219276428223, 'timing/data': 0.009539851103909314, 'timing/model': 1.1540954910451546, 'learning_rate/qwen_vl_interface': 2.628158236221438e-06, 'learning_rate/action_model': 2.628158236221438e-05, 'epoch': 1.3}) +[2026-06-19 17:41:00,888][starVLA.training.train_starvla][INFO] - Step 3463, Loss: {'train/loss': 0.18461346626281738, 'train/grad_norm_pre_clip': 7.3681159019470215, 'timing/data': 0.00951811601407826, 'timing/model': 1.3352922879857942, 'learning_rate/qwen_vl_interface': 2.6256193013690183e-06, 'learning_rate/action_model': 2.6256193013690182e-05, 'epoch': 1.3}) +[2026-06-19 17:41:05,338][starVLA.training.train_starvla][INFO] - Step 3464, Loss: {'train/loss': 0.14909259974956512, 'train/grad_norm_pre_clip': 5.891265869140625, 'timing/data': 0.009981589042581618, 'timing/model': 1.151218936080113, 'learning_rate/qwen_vl_interface': 2.623081445300373e-06, 'learning_rate/action_model': 2.6230814453003728e-05, 'epoch': 1.3}) +[2026-06-19 17:41:09,973][starVLA.training.train_starvla][INFO] - Step 3465, Loss: {'train/loss': 0.28616073727607727, 'train/grad_norm_pre_clip': 9.778620719909668, 'timing/data': 0.009528750902973115, 'timing/model': 1.1536102540558204, 'learning_rate/qwen_vl_interface': 2.6205446690587183e-06, 'learning_rate/action_model': 2.6205446690587182e-05, 'epoch': 1.3}) +[2026-06-19 17:41:14,513][starVLA.training.train_starvla][INFO] - Step 3466, Loss: {'train/loss': 0.19981664419174194, 'train/grad_norm_pre_clip': 10.324897766113281, 'timing/data': 0.009343233075924218, 'timing/model': 1.1497185949701816, 'learning_rate/qwen_vl_interface': 2.618008973686828e-06, 'learning_rate/action_model': 2.618008973686828e-05, 'epoch': 1.3}) +[2026-06-19 17:41:19,085][starVLA.training.train_starvla][INFO] - Step 3467, Loss: {'train/loss': 0.36312806606292725, 'train/grad_norm_pre_clip': 18.598392486572266, 'timing/data': 0.009909134008921683, 'timing/model': 1.1530283549800515, 'learning_rate/qwen_vl_interface': 2.615474360227028e-06, 'learning_rate/action_model': 2.615474360227028e-05, 'epoch': 1.3}) +[2026-06-19 17:41:23,734][starVLA.training.train_starvla][INFO] - Step 3468, Loss: {'train/loss': 0.4009067118167877, 'train/grad_norm_pre_clip': 16.807992935180664, 'timing/data': 0.009804458008147776, 'timing/model': 1.156111579039134, 'learning_rate/qwen_vl_interface': 2.612940829721204e-06, 'learning_rate/action_model': 2.612940829721204e-05, 'epoch': 1.3}) +[2026-06-19 17:41:28,337][starVLA.training.train_starvla][INFO] - Step 3469, Loss: {'train/loss': 0.3230459690093994, 'train/grad_norm_pre_clip': 7.272895336151123, 'timing/data': 0.009184760972857475, 'timing/model': 1.3152797420043498, 'learning_rate/qwen_vl_interface': 2.6104083832107948e-06, 'learning_rate/action_model': 2.6104083832107945e-05, 'epoch': 1.3}) +[2026-06-19 17:41:32,891][starVLA.training.train_starvla][INFO] - Step 3470, Loss: {'train/loss': 0.25227126479148865, 'train/grad_norm_pre_clip': 8.881758689880371, 'timing/data': 0.008903002948500216, 'timing/model': 1.2694854839937761, 'learning_rate/qwen_vl_interface': 2.607877021736793e-06, 'learning_rate/action_model': 2.6078770217367926e-05, 'epoch': 1.3}) +[2026-06-19 17:41:37,339][starVLA.training.train_starvla][INFO] - Step 3471, Loss: {'train/loss': 0.11525745689868927, 'train/grad_norm_pre_clip': 11.650144577026367, 'timing/data': 0.00958884705323726, 'timing/model': 1.1447939650388435, 'learning_rate/qwen_vl_interface': 2.605346746339745e-06, 'learning_rate/action_model': 2.605346746339745e-05, 'epoch': 1.3}) +[2026-06-19 17:41:41,970][starVLA.training.train_starvla][INFO] - Step 3472, Loss: {'train/loss': 0.35540375113487244, 'train/grad_norm_pre_clip': 10.372611045837402, 'timing/data': 0.009795811027288437, 'timing/model': 1.162903074058704, 'learning_rate/qwen_vl_interface': 2.602817558059753e-06, 'learning_rate/action_model': 2.6028175580597532e-05, 'epoch': 1.3}) +[2026-06-19 17:41:46,588][starVLA.training.train_starvla][INFO] - Step 3473, Loss: {'train/loss': 0.3627323806285858, 'train/grad_norm_pre_clip': 15.535446166992188, 'timing/data': 0.010287025012075901, 'timing/model': 1.156779709039256, 'learning_rate/qwen_vl_interface': 2.60028945793647e-06, 'learning_rate/action_model': 2.60028945793647e-05, 'epoch': 1.3}) +[2026-06-19 17:41:51,201][starVLA.training.train_starvla][INFO] - Step 3474, Loss: {'train/loss': 0.3254639208316803, 'train/grad_norm_pre_clip': 14.63469123840332, 'timing/data': 0.009501499007456005, 'timing/model': 1.1467723059467971, 'learning_rate/qwen_vl_interface': 2.5977624470091024e-06, 'learning_rate/action_model': 2.5977624470091023e-05, 'epoch': 1.3}) +[2026-06-19 17:41:55,724][starVLA.training.train_starvla][INFO] - Step 3475, Loss: {'train/loss': 0.2893039882183075, 'train/grad_norm_pre_clip': 12.72679615020752, 'timing/data': 0.009361214004456997, 'timing/model': 1.1586611130042002, 'learning_rate/qwen_vl_interface': 2.59523652631641e-06, 'learning_rate/action_model': 2.5952365263164097e-05, 'epoch': 1.3}) +[2026-06-19 17:42:00,289][starVLA.training.train_starvla][INFO] - Step 3476, Loss: {'train/loss': 0.29459768533706665, 'train/grad_norm_pre_clip': 14.68775749206543, 'timing/data': 0.00971322099212557, 'timing/model': 1.2528663150733337, 'learning_rate/qwen_vl_interface': 2.5927116968967024e-06, 'learning_rate/action_model': 2.5927116968967025e-05, 'epoch': 1.3}) +[2026-06-19 17:42:04,749][starVLA.training.train_starvla][INFO] - Step 3477, Loss: {'train/loss': 0.3708605468273163, 'train/grad_norm_pre_clip': 16.140459060668945, 'timing/data': 0.009813967044465244, 'timing/model': 1.151541686966084, 'learning_rate/qwen_vl_interface': 2.5901879597878417e-06, 'learning_rate/action_model': 2.5901879597878415e-05, 'epoch': 1.3}) +[2026-06-19 17:42:09,337][starVLA.training.train_starvla][INFO] - Step 3478, Loss: {'train/loss': 0.1500682830810547, 'train/grad_norm_pre_clip': 5.624704360961914, 'timing/data': 0.009520731051452458, 'timing/model': 1.1473580449819565, 'learning_rate/qwen_vl_interface': 2.5876653160272395e-06, 'learning_rate/action_model': 2.5876653160272397e-05, 'epoch': 1.3}) +[2026-06-19 17:42:13,947][starVLA.training.train_starvla][INFO] - Step 3479, Loss: {'train/loss': 0.18495865166187286, 'train/grad_norm_pre_clip': 12.026283264160156, 'timing/data': 0.009804401081055403, 'timing/model': 1.1488149350043386, 'learning_rate/qwen_vl_interface': 2.5851437666518607e-06, 'learning_rate/action_model': 2.5851437666518607e-05, 'epoch': 1.3}) +[2026-06-19 17:42:18,533][starVLA.training.train_starvla][INFO] - Step 3480, Loss: {'train/loss': 0.20639072358608246, 'train/grad_norm_pre_clip': 4.621448516845703, 'timing/data': 0.00939134694635868, 'timing/model': 1.148202640004456, 'learning_rate/qwen_vl_interface': 2.5826233126982198e-06, 'learning_rate/action_model': 2.5826233126982196e-05, 'epoch': 1.3}) +[2026-06-19 17:42:23,176][starVLA.training.train_starvla][INFO] - Step 3481, Loss: {'train/loss': 0.2749931812286377, 'train/grad_norm_pre_clip': 9.051220893859863, 'timing/data': 0.009723911061882973, 'timing/model': 1.329039804986678, 'learning_rate/qwen_vl_interface': 2.580103955202381e-06, 'learning_rate/action_model': 2.580103955202381e-05, 'epoch': 1.31}) +[2026-06-19 17:42:27,752][starVLA.training.train_starvla][INFO] - Step 3482, Loss: {'train/loss': 0.41695746779441833, 'train/grad_norm_pre_clip': 13.962074279785156, 'timing/data': 0.009704218013212085, 'timing/model': 1.2620906990487128, 'learning_rate/qwen_vl_interface': 2.577585695199954e-06, 'learning_rate/action_model': 2.5775856951999538e-05, 'epoch': 1.31}) +[2026-06-19 17:42:32,217][starVLA.training.train_starvla][INFO] - Step 3483, Loss: {'train/loss': 0.19892960786819458, 'train/grad_norm_pre_clip': 9.117942810058594, 'timing/data': 0.0099179019453004, 'timing/model': 1.1551820919848979, 'learning_rate/qwen_vl_interface': 2.575068533726101e-06, 'learning_rate/action_model': 2.575068533726101e-05, 'epoch': 1.31}) +[2026-06-19 17:42:36,848][starVLA.training.train_starvla][INFO] - Step 3484, Loss: {'train/loss': 0.2396775186061859, 'train/grad_norm_pre_clip': 8.925162315368652, 'timing/data': 0.009474423015490174, 'timing/model': 1.1500760859344155, 'learning_rate/qwen_vl_interface': 2.5725524718155333e-06, 'learning_rate/action_model': 2.572552471815533e-05, 'epoch': 1.31}) +[2026-06-19 17:42:41,461][starVLA.training.train_starvla][INFO] - Step 3485, Loss: {'train/loss': 0.26532113552093506, 'train/grad_norm_pre_clip': 13.0830659866333, 'timing/data': 0.009723140043206513, 'timing/model': 1.1539322700118646, 'learning_rate/qwen_vl_interface': 2.570037510502508e-06, 'learning_rate/action_model': 2.5700375105025083e-05, 'epoch': 1.31}) +[2026-06-19 17:42:46,065][starVLA.training.train_starvla][INFO] - Step 3486, Loss: {'train/loss': 0.11664462089538574, 'train/grad_norm_pre_clip': 7.26019287109375, 'timing/data': 0.008824288961477578, 'timing/model': 1.152313595986925, 'learning_rate/qwen_vl_interface': 2.5675236508208313e-06, 'learning_rate/action_model': 2.5675236508208313e-05, 'epoch': 1.31}) +[2026-06-19 17:42:50,701][starVLA.training.train_starvla][INFO] - Step 3487, Loss: {'train/loss': 0.337311714887619, 'train/grad_norm_pre_clip': 12.037381172180176, 'timing/data': 0.009981628973037004, 'timing/model': 1.3291094489395618, 'learning_rate/qwen_vl_interface': 2.565010893803856e-06, 'learning_rate/action_model': 2.565010893803856e-05, 'epoch': 1.31}) +[2026-06-19 17:42:55,264][starVLA.training.train_starvla][INFO] - Step 3488, Loss: {'train/loss': 0.24681033194065094, 'train/grad_norm_pre_clip': 11.169122695922852, 'timing/data': 0.009391902945935726, 'timing/model': 1.2364739460172132, 'learning_rate/qwen_vl_interface': 2.562499240484482e-06, 'learning_rate/action_model': 2.5624992404844816e-05, 'epoch': 1.31}) +[2026-06-19 17:42:59,714][starVLA.training.train_starvla][INFO] - Step 3489, Loss: {'train/loss': 0.22698098421096802, 'train/grad_norm_pre_clip': 10.988950729370117, 'timing/data': 0.009869707049801946, 'timing/model': 1.1487928200513124, 'learning_rate/qwen_vl_interface': 2.559988691895155e-06, 'learning_rate/action_model': 2.559988691895155e-05, 'epoch': 1.31}) +[2026-06-19 17:43:04,326][starVLA.training.train_starvla][INFO] - Step 3490, Loss: {'train/loss': 0.28185632824897766, 'train/grad_norm_pre_clip': 45.261539459228516, 'timing/data': 0.009514751960523427, 'timing/model': 1.1473971639061347, 'learning_rate/qwen_vl_interface': 2.5574792490678648e-06, 'learning_rate/action_model': 2.5574792490678644e-05, 'epoch': 1.31}) +[2026-06-19 17:43:08,952][starVLA.training.train_starvla][INFO] - Step 3491, Loss: {'train/loss': 0.20327068865299225, 'train/grad_norm_pre_clip': 9.75168514251709, 'timing/data': 0.009429304976947606, 'timing/model': 1.1705284080235288, 'learning_rate/qwen_vl_interface': 2.5549709130341493e-06, 'learning_rate/action_model': 2.5549709130341493e-05, 'epoch': 1.31}) +[2026-06-19 17:43:13,572][starVLA.training.train_starvla][INFO] - Step 3492, Loss: {'train/loss': 0.09722686558961868, 'train/grad_norm_pre_clip': 9.502103805541992, 'timing/data': 0.010103086940944195, 'timing/model': 1.1482089799828827, 'learning_rate/qwen_vl_interface': 2.5524636848250924e-06, 'learning_rate/action_model': 2.552463684825092e-05, 'epoch': 1.31}) +[2026-06-19 17:43:18,173][starVLA.training.train_starvla][INFO] - Step 3493, Loss: {'train/loss': 0.17496813833713531, 'train/grad_norm_pre_clip': 7.860661029815674, 'timing/data': 0.00920621503610164, 'timing/model': 1.3178457079920918, 'learning_rate/qwen_vl_interface': 2.5499575654713186e-06, 'learning_rate/action_model': 2.5499575654713185e-05, 'epoch': 1.31}) +[2026-06-19 17:43:22,713][starVLA.training.train_starvla][INFO] - Step 3494, Loss: {'train/loss': 0.17241862416267395, 'train/grad_norm_pre_clip': 12.453984260559082, 'timing/data': 0.009459491004236042, 'timing/model': 1.2566130260238424, 'learning_rate/qwen_vl_interface': 2.547452556003001e-06, 'learning_rate/action_model': 2.547452556003001e-05, 'epoch': 1.31}) +[2026-06-19 17:43:27,158][starVLA.training.train_starvla][INFO] - Step 3495, Loss: {'train/loss': 0.19317898154258728, 'train/grad_norm_pre_clip': 13.13726806640625, 'timing/data': 0.009694908978417516, 'timing/model': 1.1518784270156175, 'learning_rate/qwen_vl_interface': 2.5449486574498535e-06, 'learning_rate/action_model': 2.5449486574498534e-05, 'epoch': 1.31}) +[2026-06-19 17:43:31,815][starVLA.training.train_starvla][INFO] - Step 3496, Loss: {'train/loss': 0.2498079389333725, 'train/grad_norm_pre_clip': 13.187224388122559, 'timing/data': 0.00949236995074898, 'timing/model': 1.1701599101070315, 'learning_rate/qwen_vl_interface': 2.5424458708411335e-06, 'learning_rate/action_model': 2.5424458708411335e-05, 'epoch': 1.31}) +[2026-06-19 17:43:36,460][starVLA.training.train_starvla][INFO] - Step 3497, Loss: {'train/loss': 0.25882700085639954, 'train/grad_norm_pre_clip': 10.731551170349121, 'timing/data': 0.010021209949627519, 'timing/model': 1.1496398589806631, 'learning_rate/qwen_vl_interface': 2.5399441972056437e-06, 'learning_rate/action_model': 2.5399441972056438e-05, 'epoch': 1.31}) +[2026-06-19 17:43:41,064][starVLA.training.train_starvla][INFO] - Step 3498, Loss: {'train/loss': 0.262319952249527, 'train/grad_norm_pre_clip': 15.474617958068848, 'timing/data': 0.009553052019327879, 'timing/model': 1.1524025090038776, 'learning_rate/qwen_vl_interface': 2.537443637571726e-06, 'learning_rate/action_model': 2.537443637571726e-05, 'epoch': 1.31}) +[2026-06-19 17:43:45,683][starVLA.training.train_starvla][INFO] - Step 3499, Loss: {'train/loss': 0.28559795022010803, 'train/grad_norm_pre_clip': 8.2105073928833, 'timing/data': 0.009892228990793228, 'timing/model': 1.3170945320744067, 'learning_rate/qwen_vl_interface': 2.534944192967269e-06, 'learning_rate/action_model': 2.534944192967269e-05, 'epoch': 1.31}) +[2026-06-19 17:43:50,146][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 3500 +[2026-06-19 17:46:44,692][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 3500 +[2026-06-19 17:46:44,693][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 3500 +[2026-06-19 17:49:50,981][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 3500 +[2026-06-19 17:49:50,982][starVLA.training.train_starvla][INFO] - Step 3500, Loss: {'train/loss': 0.43447229266166687, 'train/grad_norm_pre_clip': 14.046419143676758, 'eval/loss': 0.1523071527481079, 'eval/latency_0/loss': 0.08574626594781876, 'eval/latency_1/loss': 0.04857223480939865, 'eval/latency_2/loss': 0.11506380885839462, 'eval/latency_3/loss': 0.1733899861574173, 'eval/latency_4/loss': 0.3394065797328949, 'eval/flappy/loss': 0.1523071527481079, 'eval/flappy/latency_0/loss': 0.08574626594781876, 'eval/flappy/latency_1/loss': 0.04857223480939865, 'eval/flappy/latency_2/loss': 0.11506380885839462, 'eval/flappy/latency_3/loss': 0.1733899861574173, 'eval/flappy/latency_4/loss': 0.3394065797328949, 'eval/flap_precision': 0.6816608996539792, 'eval/flap_recall': 0.7163636363636363, 'eval/flap_f1': 0.6985815602836879, 'eval/flap_rate_model': 0.08027777777777778, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.6985815602836879, 'eval/latency_0/flap_precision': 0.6816608996539792, 'eval/latency_0/flap_recall': 0.7163636363636363, 'eval/latency_0/flap_f1': 0.6985815602836879, 'eval/latency_0/flap_rate_model': 0.08027777777777778, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.6985815602836879, 'timing/data': 0.009834622032940388, 'timing/model': 1.1661462370539084, 'learning_rate/qwen_vl_interface': 2.532445864419695e-06, 'learning_rate/action_model': 2.5324458644196952e-05, 'epoch': 1.31}) +[2026-06-19 17:49:50,983][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_3500_state +[2026-06-19 17:49:50,985][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-19 17:49:54,376][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_3500_state/model.safetensors +[2026-06-19 17:50:01,580][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_3500_state/optimizer.bin +[2026-06-19 17:50:01,580][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_3500_state/sampler.bin +[2026-06-19 17:50:01,580][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_3500_state/sampler_1.bin +[2026-06-19 17:50:01,581][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_3500_state/random_states_0.pkl +[2026-06-19 17:50:01,604][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3500 +[2026-06-19 17:50:01,604][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-19 17:50:01,621][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-19 17:50:06,061][starVLA.training.train_starvla][INFO] - Step 3501, Loss: {'train/loss': 0.43558093905448914, 'train/grad_norm_pre_clip': 27.497255325317383, 'timing/data': 0.010521523072384298, 'timing/model': 1.1300967410206795, 'learning_rate/qwen_vl_interface': 2.5299486529559758e-06, 'learning_rate/action_model': 2.5299486529559756e-05, 'epoch': 1.31}) +[2026-06-19 17:50:10,524][starVLA.training.train_starvla][INFO] - Step 3502, Loss: {'train/loss': 0.23552678525447845, 'train/grad_norm_pre_clip': 12.776556968688965, 'timing/data': 0.010434234980493784, 'timing/model': 1.2143329461105168, 'learning_rate/qwen_vl_interface': 2.5274525596026202e-06, 'learning_rate/action_model': 2.52745255960262e-05, 'epoch': 1.31}) +[2026-06-19 17:50:14,898][starVLA.training.train_starvla][INFO] - Step 3503, Loss: {'train/loss': 0.37873733043670654, 'train/grad_norm_pre_clip': 11.677409172058105, 'timing/data': 0.010056196944788098, 'timing/model': 1.1318612230243161, 'learning_rate/qwen_vl_interface': 2.524957585385677e-06, 'learning_rate/action_model': 2.524957585385677e-05, 'epoch': 1.31}) +[2026-06-19 17:50:19,392][starVLA.training.train_starvla][INFO] - Step 3504, Loss: {'train/loss': 0.21417388319969177, 'train/grad_norm_pre_clip': 9.183862686157227, 'timing/data': 0.009621126926504076, 'timing/model': 1.1406709629809484, 'learning_rate/qwen_vl_interface': 2.522463731330738e-06, 'learning_rate/action_model': 2.5224637313307374e-05, 'epoch': 1.31}) +[2026-06-19 17:50:23,985][starVLA.training.train_starvla][INFO] - Step 3505, Loss: {'train/loss': 0.36357834935188293, 'train/grad_norm_pre_clip': 14.073371887207031, 'timing/data': 0.009784779977053404, 'timing/model': 1.1498388020554557, 'learning_rate/qwen_vl_interface': 2.5199709984629294e-06, 'learning_rate/action_model': 2.5199709984629292e-05, 'epoch': 1.31}) +[2026-06-19 17:50:28,595][starVLA.training.train_starvla][INFO] - Step 3506, Loss: {'train/loss': 0.2745172381401062, 'train/grad_norm_pre_clip': 13.221872329711914, 'timing/data': 0.009863840066827834, 'timing/model': 1.1561108400346711, 'learning_rate/qwen_vl_interface': 2.517479387806922e-06, 'learning_rate/action_model': 2.517479387806922e-05, 'epoch': 1.31}) +[2026-06-19 17:50:33,178][starVLA.training.train_starvla][INFO] - Step 3507, Loss: {'train/loss': 0.0806087851524353, 'train/grad_norm_pre_clip': 6.2380523681640625, 'timing/data': 0.009608781081624329, 'timing/model': 1.3141410859534517, 'learning_rate/qwen_vl_interface': 2.5149889003869217e-06, 'learning_rate/action_model': 2.5149889003869214e-05, 'epoch': 1.31}) +[2026-06-19 17:50:37,602][starVLA.training.train_starvla][INFO] - Step 3508, Loss: {'train/loss': 0.24313044548034668, 'train/grad_norm_pre_clip': 12.660135269165039, 'timing/data': 0.00967624899931252, 'timing/model': 1.1530710259685293, 'learning_rate/qwen_vl_interface': 2.5124995372266736e-06, 'learning_rate/action_model': 2.5124995372266735e-05, 'epoch': 1.32}) +[2026-06-19 17:50:42,175][starVLA.training.train_starvla][INFO] - Step 3509, Loss: {'train/loss': 0.4006635248661041, 'train/grad_norm_pre_clip': 26.188058853149414, 'timing/data': 0.009118312038481236, 'timing/model': 1.145242550992407, 'learning_rate/qwen_vl_interface': 2.5100112993494626e-06, 'learning_rate/action_model': 2.5100112993494624e-05, 'epoch': 1.32}) +[2026-06-19 17:50:46,756][starVLA.training.train_starvla][INFO] - Step 3510, Loss: {'train/loss': 0.1973181962966919, 'train/grad_norm_pre_clip': 11.895164489746094, 'timing/data': 0.00976683001499623, 'timing/model': 1.1410952670266852, 'learning_rate/qwen_vl_interface': 2.5075241877781085e-06, 'learning_rate/action_model': 2.5075241877781086e-05, 'epoch': 1.32}) +[2026-06-19 17:50:51,245][starVLA.training.train_starvla][INFO] - Step 3511, Loss: {'train/loss': 0.16378344595432281, 'train/grad_norm_pre_clip': 10.430234909057617, 'timing/data': 0.009586139000020921, 'timing/model': 1.1424709589919075, 'learning_rate/qwen_vl_interface': 2.5050382035349673e-06, 'learning_rate/action_model': 2.5050382035349674e-05, 'epoch': 1.32}) +[2026-06-19 17:50:55,715][starVLA.training.train_starvla][INFO] - Step 3512, Loss: {'train/loss': 0.13605360686779022, 'train/grad_norm_pre_clip': 8.70687484741211, 'timing/data': 0.009229343966580927, 'timing/model': 1.139865172910504, 'learning_rate/qwen_vl_interface': 2.5025533476419343e-06, 'learning_rate/action_model': 2.5025533476419343e-05, 'epoch': 1.32}) +[2026-06-19 17:51:00,222][starVLA.training.train_starvla][INFO] - Step 3513, Loss: {'train/loss': 0.2629300355911255, 'train/grad_norm_pre_clip': 12.828863143920898, 'timing/data': 0.009940134012140334, 'timing/model': 1.2497208589920774, 'learning_rate/qwen_vl_interface': 2.5000696211204394e-06, 'learning_rate/action_model': 2.5000696211204394e-05, 'epoch': 1.32}) +[2026-06-19 17:51:04,638][starVLA.training.train_starvla][INFO] - Step 3514, Loss: {'train/loss': 0.34374675154685974, 'train/grad_norm_pre_clip': 15.403167724609375, 'timing/data': 0.009003464016132057, 'timing/model': 1.1425278800306842, 'learning_rate/qwen_vl_interface': 2.4975870249914495e-06, 'learning_rate/action_model': 2.4975870249914497e-05, 'epoch': 1.32}) +[2026-06-19 17:51:09,207][starVLA.training.train_starvla][INFO] - Step 3515, Loss: {'train/loss': 0.3753287196159363, 'train/grad_norm_pre_clip': 16.537328720092773, 'timing/data': 0.009957463014870882, 'timing/model': 1.1444862730568275, 'learning_rate/qwen_vl_interface': 2.4951055602754664e-06, 'learning_rate/action_model': 2.4951055602754665e-05, 'epoch': 1.32}) +[2026-06-19 17:51:13,810][starVLA.training.train_starvla][INFO] - Step 3516, Loss: {'train/loss': 0.30344876646995544, 'train/grad_norm_pre_clip': 15.795614242553711, 'timing/data': 0.00952234398573637, 'timing/model': 1.1460660069715232, 'learning_rate/qwen_vl_interface': 2.4926252279925255e-06, 'learning_rate/action_model': 2.4926252279925254e-05, 'epoch': 1.32}) +[2026-06-19 17:51:18,384][starVLA.training.train_starvla][INFO] - Step 3517, Loss: {'train/loss': 0.320585697889328, 'train/grad_norm_pre_clip': 23.40850830078125, 'timing/data': 0.009883364080451429, 'timing/model': 1.1418578389566392, 'learning_rate/qwen_vl_interface': 2.490146029162199e-06, 'learning_rate/action_model': 2.4901460291621988e-05, 'epoch': 1.32}) +[2026-06-19 17:51:22,960][starVLA.training.train_starvla][INFO] - Step 3518, Loss: {'train/loss': 0.3842341899871826, 'train/grad_norm_pre_clip': 11.037836074829102, 'timing/data': 0.009621397941373289, 'timing/model': 1.3063230100087821, 'learning_rate/qwen_vl_interface': 2.4876679648035905e-06, 'learning_rate/action_model': 2.4876679648035903e-05, 'epoch': 1.32}) +[2026-06-19 17:51:27,482][starVLA.training.train_starvla][INFO] - Step 3519, Loss: {'train/loss': 0.28886255621910095, 'train/grad_norm_pre_clip': 13.88831615447998, 'timing/data': 0.009677647962234914, 'timing/model': 1.25199336593505, 'learning_rate/qwen_vl_interface': 2.4851910359353392e-06, 'learning_rate/action_model': 2.485191035935339e-05, 'epoch': 1.32}) +[2026-06-19 17:51:31,893][starVLA.training.train_starvla][INFO] - Step 3520, Loss: {'train/loss': 0.2442460060119629, 'train/grad_norm_pre_clip': 12.818197250366211, 'timing/data': 0.00942717504221946, 'timing/model': 1.142906955908984, 'learning_rate/qwen_vl_interface': 2.4827152435756185e-06, 'learning_rate/action_model': 2.4827152435756185e-05, 'epoch': 1.32}) +[2026-06-19 17:51:36,488][starVLA.training.train_starvla][INFO] - Step 3521, Loss: {'train/loss': 0.2868902087211609, 'train/grad_norm_pre_clip': 17.635761260986328, 'timing/data': 0.009768391028046608, 'timing/model': 1.1441228770418093, 'learning_rate/qwen_vl_interface': 2.4802405887421307e-06, 'learning_rate/action_model': 2.4802405887421306e-05, 'epoch': 1.32}) +[2026-06-19 17:51:41,111][starVLA.training.train_starvla][INFO] - Step 3522, Loss: {'train/loss': 0.09192092716693878, 'train/grad_norm_pre_clip': 5.337130069732666, 'timing/data': 0.00948546000290662, 'timing/model': 1.14447655598633, 'learning_rate/qwen_vl_interface': 2.4777670724521134e-06, 'learning_rate/action_model': 2.4777670724521134e-05, 'epoch': 1.32}) +[2026-06-19 17:51:45,699][starVLA.training.train_starvla][INFO] - Step 3523, Loss: {'train/loss': 0.31199246644973755, 'train/grad_norm_pre_clip': 20.20374298095703, 'timing/data': 0.009982643998228014, 'timing/model': 1.1431124119553715, 'learning_rate/qwen_vl_interface': 2.475294695722337e-06, 'learning_rate/action_model': 2.475294695722337e-05, 'epoch': 1.32}) +[2026-06-19 17:51:50,309][starVLA.training.train_starvla][INFO] - Step 3524, Loss: {'train/loss': 0.14299097657203674, 'train/grad_norm_pre_clip': 7.332180500030518, 'timing/data': 0.009647625964134932, 'timing/model': 1.3263540639309213, 'learning_rate/qwen_vl_interface': 2.4728234595691006e-06, 'learning_rate/action_model': 2.4728234595691005e-05, 'epoch': 1.32}) +[2026-06-19 17:51:54,836][starVLA.training.train_starvla][INFO] - Step 3525, Loss: {'train/loss': 0.1868150383234024, 'train/grad_norm_pre_clip': 11.827837944030762, 'timing/data': 0.01003645604941994, 'timing/model': 1.2543767750030383, 'learning_rate/qwen_vl_interface': 2.4703533650082368e-06, 'learning_rate/action_model': 2.4703533650082366e-05, 'epoch': 1.32}) +[2026-06-19 17:51:59,265][starVLA.training.train_starvla][INFO] - Step 3526, Loss: {'train/loss': 0.2355085015296936, 'train/grad_norm_pre_clip': 11.652604103088379, 'timing/data': 0.009747584001161158, 'timing/model': 1.1470810340251774, 'learning_rate/qwen_vl_interface': 2.4678844130551103e-06, 'learning_rate/action_model': 2.4678844130551103e-05, 'epoch': 1.32}) +[2026-06-19 17:52:03,852][starVLA.training.train_starvla][INFO] - Step 3527, Loss: {'train/loss': 0.2872023582458496, 'train/grad_norm_pre_clip': 7.865702152252197, 'timing/data': 0.009394497028551996, 'timing/model': 1.1438756890129298, 'learning_rate/qwen_vl_interface': 2.4654166047246102e-06, 'learning_rate/action_model': 2.46541660472461e-05, 'epoch': 1.32}) +[2026-06-19 17:52:08,459][starVLA.training.train_starvla][INFO] - Step 3528, Loss: {'train/loss': 0.19827812910079956, 'train/grad_norm_pre_clip': 6.5029449462890625, 'timing/data': 0.009642182965762913, 'timing/model': 1.1568928450578824, 'learning_rate/qwen_vl_interface': 2.4629499410311606e-06, 'learning_rate/action_model': 2.462949941031161e-05, 'epoch': 1.32}) +[2026-06-19 17:52:13,054][starVLA.training.train_starvla][INFO] - Step 3529, Loss: {'train/loss': 0.22839239239692688, 'train/grad_norm_pre_clip': 10.405233383178711, 'timing/data': 0.010350358090363443, 'timing/model': 1.1427457249956205, 'learning_rate/qwen_vl_interface': 2.460484422988715e-06, 'learning_rate/action_model': 2.4604844229887146e-05, 'epoch': 1.32}) +[2026-06-19 17:52:17,624][starVLA.training.train_starvla][INFO] - Step 3530, Loss: {'train/loss': 0.19052881002426147, 'train/grad_norm_pre_clip': 10.31603717803955, 'timing/data': 0.009465823066420853, 'timing/model': 1.1426304519409314, 'learning_rate/qwen_vl_interface': 2.458020051610753e-06, 'learning_rate/action_model': 2.4580200516107528e-05, 'epoch': 1.32}) +[2026-06-19 17:52:22,228][starVLA.training.train_starvla][INFO] - Step 3531, Loss: {'train/loss': 0.1531156599521637, 'train/grad_norm_pre_clip': 8.266462326049805, 'timing/data': 0.009807726019062102, 'timing/model': 1.3163035529432818, 'learning_rate/qwen_vl_interface': 2.455556827910286e-06, 'learning_rate/action_model': 2.455556827910286e-05, 'epoch': 1.32}) +[2026-06-19 17:52:26,668][starVLA.training.train_starvla][INFO] - Step 3532, Loss: {'train/loss': 0.34820181131362915, 'train/grad_norm_pre_clip': 16.11446189880371, 'timing/data': 0.00924934004433453, 'timing/model': 1.153871031012386, 'learning_rate/qwen_vl_interface': 2.453094752899853e-06, 'learning_rate/action_model': 2.453094752899853e-05, 'epoch': 1.32}) +[2026-06-19 17:52:31,366][starVLA.training.train_starvla][INFO] - Step 3533, Loss: {'train/loss': 0.20718950033187866, 'train/grad_norm_pre_clip': 7.590547561645508, 'timing/data': 0.07695360702928156, 'timing/model': 1.153651489992626, 'learning_rate/qwen_vl_interface': 2.4506338275915173e-06, 'learning_rate/action_model': 2.4506338275915172e-05, 'epoch': 1.32}) +[2026-06-19 17:52:35,960][starVLA.training.train_starvla][INFO] - Step 3534, Loss: {'train/loss': 0.2967383861541748, 'train/grad_norm_pre_clip': 13.217445373535156, 'timing/data': 0.009406420984305441, 'timing/model': 1.1420694619882852, 'learning_rate/qwen_vl_interface': 2.4481740529968735e-06, 'learning_rate/action_model': 2.4481740529968734e-05, 'epoch': 1.33}) +[2026-06-19 17:52:40,563][starVLA.training.train_starvla][INFO] - Step 3535, Loss: {'train/loss': 0.2929191589355469, 'train/grad_norm_pre_clip': 8.949967384338379, 'timing/data': 0.009223729954101145, 'timing/model': 1.152695165015757, 'learning_rate/qwen_vl_interface': 2.4457154301270427e-06, 'learning_rate/action_model': 2.4457154301270427e-05, 'epoch': 1.33}) +[2026-06-19 17:52:45,187][starVLA.training.train_starvla][INFO] - Step 3536, Loss: {'train/loss': 0.49829503893852234, 'train/grad_norm_pre_clip': 12.850749969482422, 'timing/data': 0.009139342932030559, 'timing/model': 1.1489102559862658, 'learning_rate/qwen_vl_interface': 2.443257959992671e-06, 'learning_rate/action_model': 2.443257959992671e-05, 'epoch': 1.33}) +[2026-06-19 17:52:49,825][starVLA.training.train_starvla][INFO] - Step 3537, Loss: {'train/loss': 0.19014056026935577, 'train/grad_norm_pre_clip': 17.806867599487305, 'timing/data': 0.011585868895053864, 'timing/model': 1.3398927099769935, 'learning_rate/qwen_vl_interface': 2.440801643603933e-06, 'learning_rate/action_model': 2.440801643603933e-05, 'epoch': 1.33}) +[2026-06-19 17:52:54,286][starVLA.training.train_starvla][INFO] - Step 3538, Loss: {'train/loss': 0.18755273520946503, 'train/grad_norm_pre_clip': 6.574991703033447, 'timing/data': 0.009970769053325057, 'timing/model': 1.153524647001177, 'learning_rate/qwen_vl_interface': 2.4383464819705252e-06, 'learning_rate/action_model': 2.438346481970525e-05, 'epoch': 1.33}) +[2026-06-19 17:52:58,946][starVLA.training.train_starvla][INFO] - Step 3539, Loss: {'train/loss': 0.4188830256462097, 'train/grad_norm_pre_clip': 17.153221130371094, 'timing/data': 0.012283912976272404, 'timing/model': 1.1596120150061324, 'learning_rate/qwen_vl_interface': 2.435892476101674e-06, 'learning_rate/action_model': 2.435892476101674e-05, 'epoch': 1.33}) +[2026-06-19 17:53:03,459][starVLA.training.train_starvla][INFO] - Step 3540, Loss: {'train/loss': 0.1265665590763092, 'train/grad_norm_pre_clip': 7.292163372039795, 'timing/data': 0.00954084109980613, 'timing/model': 1.1604988379403949, 'learning_rate/qwen_vl_interface': 2.433439627006128e-06, 'learning_rate/action_model': 2.4334396270061277e-05, 'epoch': 1.33}) +[2026-06-19 17:53:07,987][starVLA.training.train_starvla][INFO] - Step 3541, Loss: {'train/loss': 0.3952379524707794, 'train/grad_norm_pre_clip': 15.645636558532715, 'timing/data': 0.009995696018449962, 'timing/model': 1.1499801989411935, 'learning_rate/qwen_vl_interface': 2.4309879356921602e-06, 'learning_rate/action_model': 2.43098793569216e-05, 'epoch': 1.33}) +[2026-06-19 17:53:12,607][starVLA.training.train_starvla][INFO] - Step 3542, Loss: {'train/loss': 0.1951013207435608, 'train/grad_norm_pre_clip': 10.461785316467285, 'timing/data': 0.009462425019592047, 'timing/model': 1.147174623911269, 'learning_rate/qwen_vl_interface': 2.428537403167569e-06, 'learning_rate/action_model': 2.4285374031675687e-05, 'epoch': 1.33}) +[2026-06-19 17:53:17,213][starVLA.training.train_starvla][INFO] - Step 3543, Loss: {'train/loss': 0.24818849563598633, 'train/grad_norm_pre_clip': 16.77550506591797, 'timing/data': 0.009682941948994994, 'timing/model': 1.3269370759371668, 'learning_rate/qwen_vl_interface': 2.4260880304396765e-06, 'learning_rate/action_model': 2.4260880304396765e-05, 'epoch': 1.33}) +[2026-06-19 17:53:21,652][starVLA.training.train_starvla][INFO] - Step 3544, Loss: {'train/loss': 0.2959786653518677, 'train/grad_norm_pre_clip': 11.580885887145996, 'timing/data': 0.009643869008868933, 'timing/model': 1.1486562810605392, 'learning_rate/qwen_vl_interface': 2.4236398185153245e-06, 'learning_rate/action_model': 2.4236398185153245e-05, 'epoch': 1.33}) +[2026-06-19 17:53:26,238][starVLA.training.train_starvla][INFO] - Step 3545, Loss: {'train/loss': 0.30141592025756836, 'train/grad_norm_pre_clip': 13.607982635498047, 'timing/data': 0.009444862022064626, 'timing/model': 1.147443074034527, 'learning_rate/qwen_vl_interface': 2.4211927684008823e-06, 'learning_rate/action_model': 2.4211927684008824e-05, 'epoch': 1.33}) +[2026-06-19 17:53:30,874][starVLA.training.train_starvla][INFO] - Step 3546, Loss: {'train/loss': 0.2900504469871521, 'train/grad_norm_pre_clip': 14.135100364685059, 'timing/data': 0.009566522901877761, 'timing/model': 1.1565391139592975, 'learning_rate/qwen_vl_interface': 2.4187468811022398e-06, 'learning_rate/action_model': 2.4187468811022396e-05, 'epoch': 1.33}) +[2026-06-19 17:53:35,452][starVLA.training.train_starvla][INFO] - Step 3547, Loss: {'train/loss': 0.24478840827941895, 'train/grad_norm_pre_clip': 9.331260681152344, 'timing/data': 0.010061442037113011, 'timing/model': 1.158162135980092, 'learning_rate/qwen_vl_interface': 2.416302157624809e-06, 'learning_rate/action_model': 2.4163021576248088e-05, 'epoch': 1.33}) +[2026-06-19 17:53:40,069][starVLA.training.train_starvla][INFO] - Step 3548, Loss: {'train/loss': 0.13718506693840027, 'train/grad_norm_pre_clip': 5.224462985992432, 'timing/data': 0.009758534957654774, 'timing/model': 1.1515065958956257, 'learning_rate/qwen_vl_interface': 2.4138585989735227e-06, 'learning_rate/action_model': 2.4138585989735226e-05, 'epoch': 1.33}) +[2026-06-19 17:53:44,689][starVLA.training.train_starvla][INFO] - Step 3549, Loss: {'train/loss': 0.3474007248878479, 'train/grad_norm_pre_clip': 16.498750686645508, 'timing/data': 0.00954611599445343, 'timing/model': 1.3269891389645636, 'learning_rate/qwen_vl_interface': 2.4114162061528357e-06, 'learning_rate/action_model': 2.4114162061528354e-05, 'epoch': 1.33}) +[2026-06-19 17:53:49,137][starVLA.training.train_starvla][INFO] - Step 3550, Loss: {'train/loss': 0.20981645584106445, 'train/grad_norm_pre_clip': 62.49253463745117, 'timing/data': 0.009720740024931729, 'timing/model': 1.1592692770063877, 'learning_rate/qwen_vl_interface': 2.408974980166724e-06, 'learning_rate/action_model': 2.408974980166724e-05, 'epoch': 1.33}) +[2026-06-19 17:53:53,772][starVLA.training.train_starvla][INFO] - Step 3551, Loss: {'train/loss': 0.28194114565849304, 'train/grad_norm_pre_clip': 11.011966705322266, 'timing/data': 0.009630309999920428, 'timing/model': 1.1488217229489237, 'learning_rate/qwen_vl_interface': 2.406534922018683e-06, 'learning_rate/action_model': 2.4065349220186827e-05, 'epoch': 1.33}) +[2026-06-19 17:53:58,407][starVLA.training.train_starvla][INFO] - Step 3552, Loss: {'train/loss': 0.30223149061203003, 'train/grad_norm_pre_clip': 30.871620178222656, 'timing/data': 0.009747243952006102, 'timing/model': 1.153697333065793, 'learning_rate/qwen_vl_interface': 2.4040960327117287e-06, 'learning_rate/action_model': 2.4040960327117283e-05, 'epoch': 1.33}) +[2026-06-19 17:54:03,008][starVLA.training.train_starvla][INFO] - Step 3553, Loss: {'train/loss': 0.40154334902763367, 'train/grad_norm_pre_clip': 22.434925079345703, 'timing/data': 0.008718317956663668, 'timing/model': 1.143265272025019, 'learning_rate/qwen_vl_interface': 2.4016583132483976e-06, 'learning_rate/action_model': 2.4016583132483976e-05, 'epoch': 1.33}) +[2026-06-19 17:54:07,603][starVLA.training.train_starvla][INFO] - Step 3554, Loss: {'train/loss': 0.22963069379329681, 'train/grad_norm_pre_clip': 6.9323530197143555, 'timing/data': 0.009521574014797807, 'timing/model': 1.1481151080224663, 'learning_rate/qwen_vl_interface': 2.399221764630741e-06, 'learning_rate/action_model': 2.3992217646307406e-05, 'epoch': 1.33}) +[2026-06-19 17:54:12,200][starVLA.training.train_starvla][INFO] - Step 3555, Loss: {'train/loss': 0.5693566203117371, 'train/grad_norm_pre_clip': 39.3427848815918, 'timing/data': 0.009304239065386355, 'timing/model': 1.3185186519986019, 'learning_rate/qwen_vl_interface': 2.3967863878603334e-06, 'learning_rate/action_model': 2.3967863878603332e-05, 'epoch': 1.33}) +[2026-06-19 17:54:16,665][starVLA.training.train_starvla][INFO] - Step 3556, Loss: {'train/loss': 0.18130779266357422, 'train/grad_norm_pre_clip': 8.05081558227539, 'timing/data': 0.009572241920977831, 'timing/model': 1.1520102540962398, 'learning_rate/qwen_vl_interface': 2.394352183938266e-06, 'learning_rate/action_model': 2.3943521839382664e-05, 'epoch': 1.33}) +[2026-06-19 17:54:21,275][starVLA.training.train_starvla][INFO] - Step 3557, Loss: {'train/loss': 0.3145473301410675, 'train/grad_norm_pre_clip': 11.465058326721191, 'timing/data': 0.009601563098840415, 'timing/model': 1.1665049879811704, 'learning_rate/qwen_vl_interface': 2.39191915386515e-06, 'learning_rate/action_model': 2.39191915386515e-05, 'epoch': 1.33}) +[2026-06-19 17:54:25,834][starVLA.training.train_starvla][INFO] - Step 3558, Loss: {'train/loss': 0.36347243189811707, 'train/grad_norm_pre_clip': 14.452707290649414, 'timing/data': 0.009488599025644362, 'timing/model': 1.1633601299254224, 'learning_rate/qwen_vl_interface': 2.3894872986411095e-06, 'learning_rate/action_model': 2.3894872986411094e-05, 'epoch': 1.33}) +[2026-06-19 17:54:30,438][starVLA.training.train_starvla][INFO] - Step 3559, Loss: {'train/loss': 0.290153831243515, 'train/grad_norm_pre_clip': 11.692581176757812, 'timing/data': 0.009558952995575964, 'timing/model': 1.1686454640002921, 'learning_rate/qwen_vl_interface': 2.3870566192657903e-06, 'learning_rate/action_model': 2.38705661926579e-05, 'epoch': 1.33}) +[2026-06-19 17:54:35,072][starVLA.training.train_starvla][INFO] - Step 3560, Loss: {'train/loss': 0.3255237936973572, 'train/grad_norm_pre_clip': 12.426290512084961, 'timing/data': 0.009727058932185173, 'timing/model': 1.1615716730011627, 'learning_rate/qwen_vl_interface': 2.3846271167383515e-06, 'learning_rate/action_model': 2.3846271167383515e-05, 'epoch': 1.33}) +[2026-06-19 17:54:39,658][starVLA.training.train_starvla][INFO] - Step 3561, Loss: {'train/loss': 0.11060679703950882, 'train/grad_norm_pre_clip': 8.4908447265625, 'timing/data': 0.009533892036415637, 'timing/model': 1.3053156229434535, 'learning_rate/qwen_vl_interface': 2.382198792057471e-06, 'learning_rate/action_model': 2.3821987920574707e-05, 'epoch': 1.34}) +[2026-06-19 17:54:44,078][starVLA.training.train_starvla][INFO] - Step 3562, Loss: {'train/loss': 0.3146463930606842, 'train/grad_norm_pre_clip': 11.652177810668945, 'timing/data': 0.009966109995730221, 'timing/model': 1.1484406569506973, 'learning_rate/qwen_vl_interface': 2.3797716462213402e-06, 'learning_rate/action_model': 2.3797716462213402e-05, 'epoch': 1.34}) +[2026-06-19 17:54:48,677][starVLA.training.train_starvla][INFO] - Step 3563, Loss: {'train/loss': 0.21820655465126038, 'train/grad_norm_pre_clip': 22.633302688598633, 'timing/data': 0.010007014032453299, 'timing/model': 1.1526999530615285, 'learning_rate/qwen_vl_interface': 2.37734568022767e-06, 'learning_rate/action_model': 2.3773456802276697e-05, 'epoch': 1.34}) +[2026-06-19 17:54:53,269][starVLA.training.train_starvla][INFO] - Step 3564, Loss: {'train/loss': 0.14931008219718933, 'train/grad_norm_pre_clip': 9.179299354553223, 'timing/data': 0.00977577397134155, 'timing/model': 1.1516808699816465, 'learning_rate/qwen_vl_interface': 2.3749208950736783e-06, 'learning_rate/action_model': 2.374920895073678e-05, 'epoch': 1.34}) +[2026-06-19 17:54:57,885][starVLA.training.train_starvla][INFO] - Step 3565, Loss: {'train/loss': 0.20402202010154724, 'train/grad_norm_pre_clip': 8.479755401611328, 'timing/data': 0.009529511909931898, 'timing/model': 1.1457485160790384, 'learning_rate/qwen_vl_interface': 2.372497291756106e-06, 'learning_rate/action_model': 2.3724972917561058e-05, 'epoch': 1.34}) +[2026-06-19 17:55:02,510][starVLA.training.train_starvla][INFO] - Step 3566, Loss: {'train/loss': 0.10273677110671997, 'train/grad_norm_pre_clip': 8.325555801391602, 'timing/data': 0.00926225003786385, 'timing/model': 1.148122270940803, 'learning_rate/qwen_vl_interface': 2.3700748712712043e-06, 'learning_rate/action_model': 2.370074871271204e-05, 'epoch': 1.34}) +[2026-06-19 17:55:07,145][starVLA.training.train_starvla][INFO] - Step 3567, Loss: {'train/loss': 0.2923370599746704, 'train/grad_norm_pre_clip': 13.947896003723145, 'timing/data': 0.009749660035595298, 'timing/model': 1.3361083469353616, 'learning_rate/qwen_vl_interface': 2.3676536346147388e-06, 'learning_rate/action_model': 2.3676536346147388e-05, 'epoch': 1.34}) +[2026-06-19 17:55:11,773][starVLA.training.train_starvla][INFO] - Step 3568, Loss: {'train/loss': 0.30610647797584534, 'train/grad_norm_pre_clip': 14.039381980895996, 'timing/data': 0.010141348000615835, 'timing/model': 1.3315707510337234, 'learning_rate/qwen_vl_interface': 2.365233582781988e-06, 'learning_rate/action_model': 2.365233582781988e-05, 'epoch': 1.34}) +[2026-06-19 17:55:16,212][starVLA.training.train_starvla][INFO] - Step 3569, Loss: {'train/loss': 0.13929684460163116, 'train/grad_norm_pre_clip': 9.928020477294922, 'timing/data': 0.009488158975727856, 'timing/model': 1.147451160941273, 'learning_rate/qwen_vl_interface': 2.3628147167677444e-06, 'learning_rate/action_model': 2.3628147167677443e-05, 'epoch': 1.34}) +[2026-06-19 17:55:20,803][starVLA.training.train_starvla][INFO] - Step 3570, Loss: {'train/loss': 0.31641435623168945, 'train/grad_norm_pre_clip': 12.662467002868652, 'timing/data': 0.010052938014268875, 'timing/model': 1.1464259090134874, 'learning_rate/qwen_vl_interface': 2.3603970375663123e-06, 'learning_rate/action_model': 2.3603970375663125e-05, 'epoch': 1.34}) +[2026-06-19 17:55:25,400][starVLA.training.train_starvla][INFO] - Step 3571, Loss: {'train/loss': 0.1621258407831192, 'train/grad_norm_pre_clip': 8.407561302185059, 'timing/data': 0.009790814947336912, 'timing/model': 1.1470514999236912, 'learning_rate/qwen_vl_interface': 2.35798054617151e-06, 'learning_rate/action_model': 2.3579805461715096e-05, 'epoch': 1.34}) +[2026-06-19 17:55:29,900][starVLA.training.train_starvla][INFO] - Step 3572, Loss: {'train/loss': 0.143892303109169, 'train/grad_norm_pre_clip': 6.7445454597473145, 'timing/data': 0.009457934997044504, 'timing/model': 1.1443729320308194, 'learning_rate/qwen_vl_interface': 2.355565243576662e-06, 'learning_rate/action_model': 2.3555652435766617e-05, 'epoch': 1.34}) +[2026-06-19 17:55:34,457][starVLA.training.train_starvla][INFO] - Step 3573, Loss: {'train/loss': 0.24404869973659515, 'train/grad_norm_pre_clip': 14.783584594726562, 'timing/data': 0.009768696036189795, 'timing/model': 1.154982642037794, 'learning_rate/qwen_vl_interface': 2.35315113077461e-06, 'learning_rate/action_model': 2.3531511307746098e-05, 'epoch': 1.34}) +[2026-06-19 17:55:39,069][starVLA.training.train_starvla][INFO] - Step 3574, Loss: {'train/loss': 0.23873119056224823, 'train/grad_norm_pre_clip': 11.694908142089844, 'timing/data': 0.009716156986542046, 'timing/model': 1.3244359630625695, 'learning_rate/qwen_vl_interface': 2.3507382087577047e-06, 'learning_rate/action_model': 2.350738208757705e-05, 'epoch': 1.34}) +[2026-06-19 17:55:43,516][starVLA.training.train_starvla][INFO] - Step 3575, Loss: {'train/loss': 0.2592007517814636, 'train/grad_norm_pre_clip': 15.536076545715332, 'timing/data': 0.009497861028648913, 'timing/model': 1.1513081670273095, 'learning_rate/qwen_vl_interface': 2.348326478517807e-06, 'learning_rate/action_model': 2.3483264785178072e-05, 'epoch': 1.34}) +[2026-06-19 17:55:48,153][starVLA.training.train_starvla][INFO] - Step 3576, Loss: {'train/loss': 0.19115358591079712, 'train/grad_norm_pre_clip': 17.05504608154297, 'timing/data': 0.009487122995778918, 'timing/model': 1.1565730860456824, 'learning_rate/qwen_vl_interface': 2.3459159410462907e-06, 'learning_rate/action_model': 2.3459159410462905e-05, 'epoch': 1.34}) +[2026-06-19 17:55:52,739][starVLA.training.train_starvla][INFO] - Step 3577, Loss: {'train/loss': 0.2671109735965729, 'train/grad_norm_pre_clip': 16.9940128326416, 'timing/data': 0.009636312955990434, 'timing/model': 1.154895232990384, 'learning_rate/qwen_vl_interface': 2.343506597334031e-06, 'learning_rate/action_model': 2.343506597334031e-05, 'epoch': 1.34}) +[2026-06-19 17:55:57,373][starVLA.training.train_starvla][INFO] - Step 3578, Loss: {'train/loss': 0.43685993552207947, 'train/grad_norm_pre_clip': 17.03770637512207, 'timing/data': 0.009712432976812124, 'timing/model': 1.1471529050031677, 'learning_rate/qwen_vl_interface': 2.341098448371423e-06, 'learning_rate/action_model': 2.3410984483714227e-05, 'epoch': 1.34}) +[2026-06-19 17:56:01,971][starVLA.training.train_starvla][INFO] - Step 3579, Loss: {'train/loss': 0.2859295606613159, 'train/grad_norm_pre_clip': 10.33549690246582, 'timing/data': 0.009781338041648269, 'timing/model': 1.14639841101598, 'learning_rate/qwen_vl_interface': 2.338691495148363e-06, 'learning_rate/action_model': 2.3386914951483626e-05, 'epoch': 1.34}) +[2026-06-19 17:56:06,597][starVLA.training.train_starvla][INFO] - Step 3580, Loss: {'train/loss': 0.24118700623512268, 'train/grad_norm_pre_clip': 7.9262166023254395, 'timing/data': 0.010100375046022236, 'timing/model': 1.332718694000505, 'learning_rate/qwen_vl_interface': 2.33628573865426e-06, 'learning_rate/action_model': 2.3362857386542596e-05, 'epoch': 1.34}) +[2026-06-19 17:56:11,125][starVLA.training.train_starvla][INFO] - Step 3581, Loss: {'train/loss': 0.47797030210494995, 'train/grad_norm_pre_clip': 24.924970626831055, 'timing/data': 0.009751159930601716, 'timing/model': 1.2420127410441637, 'learning_rate/qwen_vl_interface': 2.3338811798780282e-06, 'learning_rate/action_model': 2.333881179878028e-05, 'epoch': 1.34}) +[2026-06-19 17:56:15,577][starVLA.training.train_starvla][INFO] - Step 3582, Loss: {'train/loss': 0.3173576891422272, 'train/grad_norm_pre_clip': 17.88564682006836, 'timing/data': 0.00938333896920085, 'timing/model': 1.1538302189437672, 'learning_rate/qwen_vl_interface': 2.3314778198080933e-06, 'learning_rate/action_model': 2.331477819808093e-05, 'epoch': 1.34}) +[2026-06-19 17:56:20,118][starVLA.training.train_starvla][INFO] - Step 3583, Loss: {'train/loss': 0.4349587559700012, 'train/grad_norm_pre_clip': 14.821672439575195, 'timing/data': 0.009394404012709856, 'timing/model': 1.1460910580353811, 'learning_rate/qwen_vl_interface': 2.3290756594323834e-06, 'learning_rate/action_model': 2.3290756594323836e-05, 'epoch': 1.34}) +[2026-06-19 17:56:24,716][starVLA.training.train_starvla][INFO] - Step 3584, Loss: {'train/loss': 0.20023952424526215, 'train/grad_norm_pre_clip': 9.343670845031738, 'timing/data': 0.009692355059087276, 'timing/model': 1.1446288540028036, 'learning_rate/qwen_vl_interface': 2.3266746997383378e-06, 'learning_rate/action_model': 2.3266746997383377e-05, 'epoch': 1.34}) +[2026-06-19 17:56:29,322][starVLA.training.train_starvla][INFO] - Step 3585, Loss: {'train/loss': 0.4468627870082855, 'train/grad_norm_pre_clip': 28.663482666015625, 'timing/data': 0.009984516073018312, 'timing/model': 1.148142536985688, 'learning_rate/qwen_vl_interface': 2.324274941712899e-06, 'learning_rate/action_model': 2.324274941712899e-05, 'epoch': 1.34}) +[2026-06-19 17:56:33,917][starVLA.training.train_starvla][INFO] - Step 3586, Loss: {'train/loss': 0.1455984264612198, 'train/grad_norm_pre_clip': 5.879528045654297, 'timing/data': 0.009370070998556912, 'timing/model': 1.3131174660520628, 'learning_rate/qwen_vl_interface': 2.321876386342519e-06, 'learning_rate/action_model': 2.3218763863425187e-05, 'epoch': 1.34}) +[2026-06-19 17:56:38,353][starVLA.training.train_starvla][INFO] - Step 3587, Loss: {'train/loss': 0.132695272564888, 'train/grad_norm_pre_clip': 8.681906700134277, 'timing/data': 0.00965816096868366, 'timing/model': 1.1511946379905567, 'learning_rate/qwen_vl_interface': 2.31947903461315e-06, 'learning_rate/action_model': 2.3194790346131502e-05, 'epoch': 1.34}) +[2026-06-19 17:56:42,940][starVLA.training.train_starvla][INFO] - Step 3588, Loss: {'train/loss': 0.19493509829044342, 'train/grad_norm_pre_clip': 11.323230743408203, 'timing/data': 0.009573427028954029, 'timing/model': 1.1470533659448847, 'learning_rate/qwen_vl_interface': 2.3170828875102544e-06, 'learning_rate/action_model': 2.3170828875102543e-05, 'epoch': 1.35}) +[2026-06-19 17:56:47,555][starVLA.training.train_starvla][INFO] - Step 3589, Loss: {'train/loss': 0.10866132378578186, 'train/grad_norm_pre_clip': 5.723403453826904, 'timing/data': 0.009845896973274648, 'timing/model': 1.1511651199543849, 'learning_rate/qwen_vl_interface': 2.3146879460187984e-06, 'learning_rate/action_model': 2.3146879460187982e-05, 'epoch': 1.35}) +[2026-06-19 17:56:52,085][starVLA.training.train_starvla][INFO] - Step 3590, Loss: {'train/loss': 0.1708449423313141, 'train/grad_norm_pre_clip': 12.568679809570312, 'timing/data': 0.009585649939253926, 'timing/model': 1.15412802901119, 'learning_rate/qwen_vl_interface': 2.3122942111232506e-06, 'learning_rate/action_model': 2.3122942111232506e-05, 'epoch': 1.35}) +[2026-06-19 17:56:56,607][starVLA.training.train_starvla][INFO] - Step 3591, Loss: {'train/loss': 0.1811147928237915, 'train/grad_norm_pre_clip': 13.638612747192383, 'timing/data': 0.009790780022740364, 'timing/model': 1.1632939799455926, 'learning_rate/qwen_vl_interface': 2.3099016838075857e-06, 'learning_rate/action_model': 2.3099016838075856e-05, 'epoch': 1.35}) +[2026-06-19 17:57:01,146][starVLA.training.train_starvla][INFO] - Step 3592, Loss: {'train/loss': 0.21774685382843018, 'train/grad_norm_pre_clip': 11.074481010437012, 'timing/data': 0.009594188071787357, 'timing/model': 1.2477999390102923, 'learning_rate/qwen_vl_interface': 2.307510365055282e-06, 'learning_rate/action_model': 2.307510365055282e-05, 'epoch': 1.35}) +[2026-06-19 17:57:05,593][starVLA.training.train_starvla][INFO] - Step 3593, Loss: {'train/loss': 0.3230711817741394, 'train/grad_norm_pre_clip': 11.853752136230469, 'timing/data': 0.00937544903717935, 'timing/model': 1.1547647659899667, 'learning_rate/qwen_vl_interface': 2.3051202558493196e-06, 'learning_rate/action_model': 2.3051202558493192e-05, 'epoch': 1.35}) +[2026-06-19 17:57:10,232][starVLA.training.train_starvla][INFO] - Step 3594, Loss: {'train/loss': 0.14808812737464905, 'train/grad_norm_pre_clip': 7.143833160400391, 'timing/data': 0.009907323983497918, 'timing/model': 1.1492211740696803, 'learning_rate/qwen_vl_interface': 2.302731357172182e-06, 'learning_rate/action_model': 2.302731357172182e-05, 'epoch': 1.35}) +[2026-06-19 17:57:14,848][starVLA.training.train_starvla][INFO] - Step 3595, Loss: {'train/loss': 0.29588884115219116, 'train/grad_norm_pre_clip': 17.40793800354004, 'timing/data': 0.010046688956208527, 'timing/model': 1.147646086057648, 'learning_rate/qwen_vl_interface': 2.3003436700058556e-06, 'learning_rate/action_model': 2.3003436700058554e-05, 'epoch': 1.35}) +[2026-06-19 17:57:19,485][starVLA.training.train_starvla][INFO] - Step 3596, Loss: {'train/loss': 0.48960137367248535, 'train/grad_norm_pre_clip': 18.010391235351562, 'timing/data': 0.009638298070058227, 'timing/model': 1.1559057869017124, 'learning_rate/qwen_vl_interface': 2.297957195331828e-06, 'learning_rate/action_model': 2.2979571953318276e-05, 'epoch': 1.35}) +[2026-06-19 17:57:24,110][starVLA.training.train_starvla][INFO] - Step 3597, Loss: {'train/loss': 0.3396850824356079, 'train/grad_norm_pre_clip': 11.348062515258789, 'timing/data': 0.01224470790475607, 'timing/model': 1.3326640549348667, 'learning_rate/qwen_vl_interface': 2.2955719341310903e-06, 'learning_rate/action_model': 2.2955719341310903e-05, 'epoch': 1.35}) +[2026-06-19 17:57:28,688][starVLA.training.train_starvla][INFO] - Step 3598, Loss: {'train/loss': 0.31487852334976196, 'train/grad_norm_pre_clip': 15.41188907623291, 'timing/data': 0.009274538955651224, 'timing/model': 1.282983856042847, 'learning_rate/qwen_vl_interface': 2.2931878873841304e-06, 'learning_rate/action_model': 2.2931878873841302e-05, 'epoch': 1.35}) +[2026-06-19 17:57:33,119][starVLA.training.train_starvla][INFO] - Step 3599, Loss: {'train/loss': 0.3879741132259369, 'train/grad_norm_pre_clip': 24.044496536254883, 'timing/data': 0.009204261936247349, 'timing/model': 1.147945992066525, 'learning_rate/qwen_vl_interface': 2.290805056070942e-06, 'learning_rate/action_model': 2.290805056070942e-05, 'epoch': 1.35}) +[2026-06-19 17:57:37,717][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 3600 +[2026-06-19 18:00:30,571][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 3600 +[2026-06-19 18:00:30,572][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 3600 +[2026-06-19 18:03:37,876][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 3600 +[2026-06-19 18:03:37,877][starVLA.training.train_starvla][INFO] - Step 3600, Loss: {'train/loss': 0.3393702208995819, 'train/grad_norm_pre_clip': 20.651657104492188, 'eval/loss': 0.14180520176887512, 'eval/latency_0/loss': 0.0849716067314148, 'eval/latency_1/loss': 0.046717412769794464, 'eval/latency_2/loss': 0.11461777240037918, 'eval/latency_3/loss': 0.15610358119010925, 'eval/latency_4/loss': 0.30665525794029236, 'eval/flappy/loss': 0.14180520176887512, 'eval/flappy/latency_0/loss': 0.0849716067314148, 'eval/flappy/latency_1/loss': 0.046717412769794464, 'eval/flappy/latency_2/loss': 0.11461777240037918, 'eval/flappy/latency_3/loss': 0.15610358119010925, 'eval/flappy/latency_4/loss': 0.30665525794029236, 'eval/flap_precision': 0.7148014440433214, 'eval/flap_recall': 0.72, 'eval/flap_f1': 0.717391304347826, 'eval/flap_rate_model': 0.07694444444444444, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.717391304347826, 'eval/latency_0/flap_precision': 0.7148014440433214, 'eval/latency_0/flap_recall': 0.72, 'eval/latency_0/flap_f1': 0.717391304347826, 'eval/latency_0/flap_rate_model': 0.07694444444444444, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.717391304347826, 'timing/data': 0.010487819090485573, 'timing/model': 1.1482183389598504, 'learning_rate/qwen_vl_interface': 2.288423441171016e-06, 'learning_rate/action_model': 2.288423441171016e-05, 'epoch': 1.35}) +[2026-06-19 18:03:37,878][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_3600_state +[2026-06-19 18:03:37,880][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-19 18:03:41,271][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_3600_state/model.safetensors +[2026-06-19 18:03:48,411][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_3600_state/optimizer.bin +[2026-06-19 18:03:48,411][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_3600_state/sampler.bin +[2026-06-19 18:03:48,411][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_3600_state/sampler_1.bin +[2026-06-19 18:03:48,412][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_3600_state/random_states_0.pkl +[2026-06-19 18:03:48,434][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3600 +[2026-06-19 18:03:48,434][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-19 18:03:48,444][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-19 18:03:52,972][starVLA.training.train_starvla][INFO] - Step 3601, Loss: {'train/loss': 0.12068350613117218, 'train/grad_norm_pre_clip': 8.70397663116455, 'timing/data': 0.010418054065667093, 'timing/model': 1.13537108595483, 'learning_rate/qwen_vl_interface': 2.2860430436633447e-06, 'learning_rate/action_model': 2.2860430436633447e-05, 'epoch': 1.35}) +[2026-06-19 18:03:57,516][starVLA.training.train_starvla][INFO] - Step 3602, Loss: {'train/loss': 0.106879323720932, 'train/grad_norm_pre_clip': 17.97403907775879, 'timing/data': 0.008882264024578035, 'timing/model': 1.31319032702595, 'learning_rate/qwen_vl_interface': 2.28366386452642e-06, 'learning_rate/action_model': 2.28366386452642e-05, 'epoch': 1.35}) +[2026-06-19 18:04:01,898][starVLA.training.train_starvla][INFO] - Step 3603, Loss: {'train/loss': 0.2575685679912567, 'train/grad_norm_pre_clip': 13.484320640563965, 'timing/data': 0.009521270985715091, 'timing/model': 1.1400848990306258, 'learning_rate/qwen_vl_interface': 2.2812859047382326e-06, 'learning_rate/action_model': 2.2812859047382323e-05, 'epoch': 1.35}) +[2026-06-19 18:04:06,464][starVLA.training.train_starvla][INFO] - Step 3604, Loss: {'train/loss': 0.3015730381011963, 'train/grad_norm_pre_clip': 10.83326244354248, 'timing/data': 0.009720340953208506, 'timing/model': 1.1481053499737754, 'learning_rate/qwen_vl_interface': 2.2789091652762712e-06, 'learning_rate/action_model': 2.2789091652762712e-05, 'epoch': 1.35}) +[2026-06-19 18:04:10,971][starVLA.training.train_starvla][INFO] - Step 3605, Loss: {'train/loss': 0.09314820170402527, 'train/grad_norm_pre_clip': 11.409082412719727, 'timing/data': 0.009274871088564396, 'timing/model': 1.1569406290072948, 'learning_rate/qwen_vl_interface': 2.276533647117524e-06, 'learning_rate/action_model': 2.2765336471175242e-05, 'epoch': 1.35}) +[2026-06-19 18:04:15,529][starVLA.training.train_starvla][INFO] - Step 3606, Loss: {'train/loss': 0.1681285947561264, 'train/grad_norm_pre_clip': 8.518848419189453, 'timing/data': 0.009852873045019805, 'timing/model': 1.1378040290437639, 'learning_rate/qwen_vl_interface': 2.2741593512384773e-06, 'learning_rate/action_model': 2.2741593512384772e-05, 'epoch': 1.35}) +[2026-06-19 18:04:20,106][starVLA.training.train_starvla][INFO] - Step 3607, Loss: {'train/loss': 0.18616384267807007, 'train/grad_norm_pre_clip': 13.85873794555664, 'timing/data': 0.010059132939204574, 'timing/model': 1.133255049935542, 'learning_rate/qwen_vl_interface': 2.2717862786151153e-06, 'learning_rate/action_model': 2.271786278615115e-05, 'epoch': 1.35}) +[2026-06-19 18:04:24,673][starVLA.training.train_starvla][INFO] - Step 3608, Loss: {'train/loss': 0.3454722464084625, 'train/grad_norm_pre_clip': 29.58759880065918, 'timing/data': 0.009454186074435711, 'timing/model': 1.3166033889865503, 'learning_rate/qwen_vl_interface': 2.2694144302229152e-06, 'learning_rate/action_model': 2.2694144302229152e-05, 'epoch': 1.35}) +[2026-06-19 18:04:29,089][starVLA.training.train_starvla][INFO] - Step 3609, Loss: {'train/loss': 0.36190804839134216, 'train/grad_norm_pre_clip': 16.062908172607422, 'timing/data': 0.00911778793670237, 'timing/model': 1.153685353929177, 'learning_rate/qwen_vl_interface': 2.2670438070368568e-06, 'learning_rate/action_model': 2.2670438070368566e-05, 'epoch': 1.35}) +[2026-06-19 18:04:33,669][starVLA.training.train_starvla][INFO] - Step 3610, Loss: {'train/loss': 0.18052569031715393, 'train/grad_norm_pre_clip': 9.12370777130127, 'timing/data': 0.00928963499609381, 'timing/model': 1.1408187410561368, 'learning_rate/qwen_vl_interface': 2.264674410031413e-06, 'learning_rate/action_model': 2.2646744100314127e-05, 'epoch': 1.35}) +[2026-06-19 18:04:38,279][starVLA.training.train_starvla][INFO] - Step 3611, Loss: {'train/loss': 0.10219679772853851, 'train/grad_norm_pre_clip': 7.018362522125244, 'timing/data': 0.009310006978921592, 'timing/model': 1.162433462915942, 'learning_rate/qwen_vl_interface': 2.2623062401805537e-06, 'learning_rate/action_model': 2.2623062401805535e-05, 'epoch': 1.35}) +[2026-06-19 18:04:42,788][starVLA.training.train_starvla][INFO] - Step 3612, Loss: {'train/loss': 0.16702790558338165, 'train/grad_norm_pre_clip': 10.893075942993164, 'timing/data': 0.009452536003664136, 'timing/model': 1.143953518010676, 'learning_rate/qwen_vl_interface': 2.2599392984577436e-06, 'learning_rate/action_model': 2.2599392984577438e-05, 'epoch': 1.35}) +[2026-06-19 18:04:47,287][starVLA.training.train_starvla][INFO] - Step 3613, Loss: {'train/loss': 0.11241775006055832, 'train/grad_norm_pre_clip': 7.065695762634277, 'timing/data': 0.009697227971628308, 'timing/model': 1.1402662540785968, 'learning_rate/qwen_vl_interface': 2.2575735858359436e-06, 'learning_rate/action_model': 2.257573585835944e-05, 'epoch': 1.35}) +[2026-06-19 18:04:51,813][starVLA.training.train_starvla][INFO] - Step 3614, Loss: {'train/loss': 0.3158641457557678, 'train/grad_norm_pre_clip': 22.326509475708008, 'timing/data': 0.009829516056925058, 'timing/model': 1.2656459179706872, 'learning_rate/qwen_vl_interface': 2.255209103287609e-06, 'learning_rate/action_model': 2.255209103287609e-05, 'epoch': 1.36}) +[2026-06-19 18:04:56,222][starVLA.training.train_starvla][INFO] - Step 3615, Loss: {'train/loss': 0.24054737389087677, 'train/grad_norm_pre_clip': 7.932226181030273, 'timing/data': 0.009752048994414508, 'timing/model': 1.1451844510156661, 'learning_rate/qwen_vl_interface': 2.2528458517846886e-06, 'learning_rate/action_model': 2.252845851784689e-05, 'epoch': 1.36}) +[2026-06-19 18:05:00,796][starVLA.training.train_starvla][INFO] - Step 3616, Loss: {'train/loss': 0.26811498403549194, 'train/grad_norm_pre_clip': 18.732681274414062, 'timing/data': 0.009200289961881936, 'timing/model': 1.1452815160155296, 'learning_rate/qwen_vl_interface': 2.250483832298627e-06, 'learning_rate/action_model': 2.250483832298627e-05, 'epoch': 1.36}) +[2026-06-19 18:05:05,402][starVLA.training.train_starvla][INFO] - Step 3617, Loss: {'train/loss': 0.2786581516265869, 'train/grad_norm_pre_clip': 9.7682466506958, 'timing/data': 0.00891771202441305, 'timing/model': 1.1434811440994963, 'learning_rate/qwen_vl_interface': 2.2481230458003608e-06, 'learning_rate/action_model': 2.2481230458003606e-05, 'epoch': 1.36}) +[2026-06-19 18:05:09,975][starVLA.training.train_starvla][INFO] - Step 3618, Loss: {'train/loss': 0.23658928275108337, 'train/grad_norm_pre_clip': 9.811800003051758, 'timing/data': 0.009732597973197699, 'timing/model': 1.1387726010289043, 'learning_rate/qwen_vl_interface': 2.2457634932603184e-06, 'learning_rate/action_model': 2.2457634932603183e-05, 'epoch': 1.36}) +[2026-06-19 18:05:14,556][starVLA.training.train_starvla][INFO] - Step 3619, Loss: {'train/loss': 0.09133419394493103, 'train/grad_norm_pre_clip': 7.35919189453125, 'timing/data': 0.00988721998874098, 'timing/model': 1.306709713069722, 'learning_rate/qwen_vl_interface': 2.2434051756484255e-06, 'learning_rate/action_model': 2.2434051756484253e-05, 'epoch': 1.36}) +[2026-06-19 18:05:18,976][starVLA.training.train_starvla][INFO] - Step 3620, Loss: {'train/loss': 0.17537403106689453, 'train/grad_norm_pre_clip': 9.917830467224121, 'timing/data': 0.00903151708189398, 'timing/model': 1.1525000859983265, 'learning_rate/qwen_vl_interface': 2.2410480939340945e-06, 'learning_rate/action_model': 2.2410480939340945e-05, 'epoch': 1.36}) +[2026-06-19 18:05:23,595][starVLA.training.train_starvla][INFO] - Step 3621, Loss: {'train/loss': 0.23724757134914398, 'train/grad_norm_pre_clip': 18.37830924987793, 'timing/data': 0.00987055990844965, 'timing/model': 1.1478448510169983, 'learning_rate/qwen_vl_interface': 2.238692249086234e-06, 'learning_rate/action_model': 2.2386922490862336e-05, 'epoch': 1.36}) +[2026-06-19 18:05:28,205][starVLA.training.train_starvla][INFO] - Step 3622, Loss: {'train/loss': 0.21046334505081177, 'train/grad_norm_pre_clip': 23.701704025268555, 'timing/data': 0.009858853067271411, 'timing/model': 1.1432145040016621, 'learning_rate/qwen_vl_interface': 2.2363376420732426e-06, 'learning_rate/action_model': 2.2363376420732423e-05, 'epoch': 1.36}) +[2026-06-19 18:05:32,810][starVLA.training.train_starvla][INFO] - Step 3623, Loss: {'train/loss': 0.09375244379043579, 'train/grad_norm_pre_clip': 7.9888081550598145, 'timing/data': 0.009647991973906755, 'timing/model': 1.151655999943614, 'learning_rate/qwen_vl_interface': 2.233984273863011e-06, 'learning_rate/action_model': 2.233984273863011e-05, 'epoch': 1.36}) +[2026-06-19 18:05:37,416][starVLA.training.train_starvla][INFO] - Step 3624, Loss: {'train/loss': 0.24887911975383759, 'train/grad_norm_pre_clip': 16.305273056030273, 'timing/data': 0.00958741910289973, 'timing/model': 1.150107413996011, 'learning_rate/qwen_vl_interface': 2.231632145422919e-06, 'learning_rate/action_model': 2.231632145422919e-05, 'epoch': 1.36}) +[2026-06-19 18:05:42,041][starVLA.training.train_starvla][INFO] - Step 3625, Loss: {'train/loss': 0.15097205340862274, 'train/grad_norm_pre_clip': 11.651759147644043, 'timing/data': 0.009530620998702943, 'timing/model': 1.333662598975934, 'learning_rate/qwen_vl_interface': 2.2292812577198383e-06, 'learning_rate/action_model': 2.229281257719838e-05, 'epoch': 1.36}) +[2026-06-19 18:05:46,601][starVLA.training.train_starvla][INFO] - Step 3626, Loss: {'train/loss': 0.16281461715698242, 'train/grad_norm_pre_clip': 8.990228652954102, 'timing/data': 0.009787250077351928, 'timing/model': 1.2704063509590924, 'learning_rate/qwen_vl_interface': 2.2269316117201307e-06, 'learning_rate/action_model': 2.2269316117201307e-05, 'epoch': 1.36}) +[2026-06-19 18:05:51,075][starVLA.training.train_starvla][INFO] - Step 3627, Loss: {'train/loss': 0.20973709225654602, 'train/grad_norm_pre_clip': 12.243791580200195, 'timing/data': 0.009756332030519843, 'timing/model': 1.1570689319632947, 'learning_rate/qwen_vl_interface': 2.224583208389646e-06, 'learning_rate/action_model': 2.224583208389646e-05, 'epoch': 1.36}) +[2026-06-19 18:05:55,729][starVLA.training.train_starvla][INFO] - Step 3628, Loss: {'train/loss': 0.2358947992324829, 'train/grad_norm_pre_clip': 9.927053451538086, 'timing/data': 0.009603293030522764, 'timing/model': 1.1499912360450253, 'learning_rate/qwen_vl_interface': 2.2222360486937238e-06, 'learning_rate/action_model': 2.2222360486937235e-05, 'epoch': 1.36}) +[2026-06-19 18:06:00,331][starVLA.training.train_starvla][INFO] - Step 3629, Loss: {'train/loss': 0.1953943520784378, 'train/grad_norm_pre_clip': 11.424519538879395, 'timing/data': 0.010084543959237635, 'timing/model': 1.1406479990109801, 'learning_rate/qwen_vl_interface': 2.219890133597194e-06, 'learning_rate/action_model': 2.219890133597194e-05, 'epoch': 1.36}) +[2026-06-19 18:06:04,922][starVLA.training.train_starvla][INFO] - Step 3630, Loss: {'train/loss': 0.2568834722042084, 'train/grad_norm_pre_clip': 23.11494255065918, 'timing/data': 0.009606859995983541, 'timing/model': 1.1430765650002286, 'learning_rate/qwen_vl_interface': 2.217545464064374e-06, 'learning_rate/action_model': 2.217545464064374e-05, 'epoch': 1.36}) +[2026-06-19 18:06:09,498][starVLA.training.train_starvla][INFO] - Step 3631, Loss: {'train/loss': 0.08826690167188644, 'train/grad_norm_pre_clip': 7.761810779571533, 'timing/data': 0.009597986936569214, 'timing/model': 1.1456524659879506, 'learning_rate/qwen_vl_interface': 2.215202041059067e-06, 'learning_rate/action_model': 2.215202041059067e-05, 'epoch': 1.36}) +[2026-06-19 18:06:14,081][starVLA.training.train_starvla][INFO] - Step 3632, Loss: {'train/loss': 0.19659729301929474, 'train/grad_norm_pre_clip': 16.763294219970703, 'timing/data': 0.009483380941674113, 'timing/model': 1.311972487019375, 'learning_rate/qwen_vl_interface': 2.212859865544566e-06, 'learning_rate/action_model': 2.212859865544566e-05, 'epoch': 1.36}) +[2026-06-19 18:06:18,503][starVLA.training.train_starvla][INFO] - Step 3633, Loss: {'train/loss': 0.36351582407951355, 'train/grad_norm_pre_clip': 15.974108695983887, 'timing/data': 0.009233193006366491, 'timing/model': 1.1494133420055732, 'learning_rate/qwen_vl_interface': 2.210518938483652e-06, 'learning_rate/action_model': 2.210518938483652e-05, 'epoch': 1.36}) +[2026-06-19 18:06:23,106][starVLA.training.train_starvla][INFO] - Step 3634, Loss: {'train/loss': 0.273469477891922, 'train/grad_norm_pre_clip': 15.605323791503906, 'timing/data': 0.009382819989696145, 'timing/model': 1.1524034360190853, 'learning_rate/qwen_vl_interface': 2.20817926083859e-06, 'learning_rate/action_model': 2.20817926083859e-05, 'epoch': 1.36}) +[2026-06-19 18:06:27,700][starVLA.training.train_starvla][INFO] - Step 3635, Loss: {'train/loss': 0.3362489640712738, 'train/grad_norm_pre_clip': 22.296363830566406, 'timing/data': 0.010005060001276433, 'timing/model': 1.1475196999963373, 'learning_rate/qwen_vl_interface': 2.205840833571135e-06, 'learning_rate/action_model': 2.205840833571135e-05, 'epoch': 1.36}) +[2026-06-19 18:06:32,310][starVLA.training.train_starvla][INFO] - Step 3636, Loss: {'train/loss': 0.11881185322999954, 'train/grad_norm_pre_clip': 8.174711227416992, 'timing/data': 0.010009662015363574, 'timing/model': 1.1451137270778418, 'learning_rate/qwen_vl_interface': 2.2035036576425264e-06, 'learning_rate/action_model': 2.203503657642526e-05, 'epoch': 1.36}) +[2026-06-19 18:06:36,912][starVLA.training.train_starvla][INFO] - Step 3637, Loss: {'train/loss': 0.21886315941810608, 'train/grad_norm_pre_clip': 12.571738243103027, 'timing/data': 0.009243616019375622, 'timing/model': 1.1458360709948465, 'learning_rate/qwen_vl_interface': 2.2011677340134867e-06, 'learning_rate/action_model': 2.2011677340134867e-05, 'epoch': 1.36}) +[2026-06-19 18:06:41,506][starVLA.training.train_starvla][INFO] - Step 3638, Loss: {'train/loss': 0.14532464742660522, 'train/grad_norm_pre_clip': 15.99380111694336, 'timing/data': 0.009566013002768159, 'timing/model': 1.3185786519898102, 'learning_rate/qwen_vl_interface': 2.198833063644227e-06, 'learning_rate/action_model': 2.198833063644227e-05, 'epoch': 1.36}) +[2026-06-19 18:06:45,941][starVLA.training.train_starvla][INFO] - Step 3639, Loss: {'train/loss': 0.2556096315383911, 'train/grad_norm_pre_clip': 20.43157958984375, 'timing/data': 0.00997863698285073, 'timing/model': 1.1502610640600324, 'learning_rate/qwen_vl_interface': 2.196499647494443e-06, 'learning_rate/action_model': 2.196499647494443e-05, 'epoch': 1.36}) +[2026-06-19 18:06:50,548][starVLA.training.train_starvla][INFO] - Step 3640, Loss: {'train/loss': 0.1320018470287323, 'train/grad_norm_pre_clip': 9.918553352355957, 'timing/data': 0.009977151057682931, 'timing/model': 1.1474517410388216, 'learning_rate/qwen_vl_interface': 2.1941674865233144e-06, 'learning_rate/action_model': 2.1941674865233143e-05, 'epoch': 1.36}) +[2026-06-19 18:06:55,051][starVLA.training.train_starvla][INFO] - Step 3641, Loss: {'train/loss': 0.16655932366847992, 'train/grad_norm_pre_clip': 11.436666488647461, 'timing/data': 0.009519711951725185, 'timing/model': 1.1606460789917037, 'learning_rate/qwen_vl_interface': 2.1918365816895013e-06, 'learning_rate/action_model': 2.191836581689501e-05, 'epoch': 1.37}) +[2026-06-19 18:06:59,600][starVLA.training.train_starvla][INFO] - Step 3642, Loss: {'train/loss': 0.22926412522792816, 'train/grad_norm_pre_clip': 15.338603019714355, 'timing/data': 0.00968216999899596, 'timing/model': 1.1432284499751404, 'learning_rate/qwen_vl_interface': 2.189506933951153e-06, 'learning_rate/action_model': 2.1895069339511525e-05, 'epoch': 1.37}) +[2026-06-19 18:07:04,234][starVLA.training.train_starvla][INFO] - Step 3643, Loss: {'train/loss': 0.18414710462093353, 'train/grad_norm_pre_clip': 10.414631843566895, 'timing/data': 0.008671322953887284, 'timing/model': 1.1574181620962918, 'learning_rate/qwen_vl_interface': 2.187178544265898e-06, 'learning_rate/action_model': 2.1871785442658983e-05, 'epoch': 1.37}) +[2026-06-19 18:07:08,862][starVLA.training.train_starvla][INFO] - Step 3644, Loss: {'train/loss': 0.2506541609764099, 'train/grad_norm_pre_clip': 15.826257705688477, 'timing/data': 0.009617984993383288, 'timing/model': 1.3376050629885867, 'learning_rate/qwen_vl_interface': 2.184851413590852e-06, 'learning_rate/action_model': 2.1848514135908518e-05, 'epoch': 1.37}) +[2026-06-19 18:07:13,325][starVLA.training.train_starvla][INFO] - Step 3645, Loss: {'train/loss': 0.27492523193359375, 'train/grad_norm_pre_clip': 19.778133392333984, 'timing/data': 0.009737713960930705, 'timing/model': 1.1571880580158904, 'learning_rate/qwen_vl_interface': 2.1825255428826087e-06, 'learning_rate/action_model': 2.182525542882609e-05, 'epoch': 1.37}) +[2026-06-19 18:07:17,934][starVLA.training.train_starvla][INFO] - Step 3646, Loss: {'train/loss': 0.34997162222862244, 'train/grad_norm_pre_clip': 22.347671508789062, 'timing/data': 0.00959567294921726, 'timing/model': 1.1469827049877495, 'learning_rate/qwen_vl_interface': 2.180200933097246e-06, 'learning_rate/action_model': 2.180200933097246e-05, 'epoch': 1.37}) +[2026-06-19 18:07:22,544][starVLA.training.train_starvla][INFO] - Step 3647, Loss: {'train/loss': 0.41048717498779297, 'train/grad_norm_pre_clip': 30.004539489746094, 'timing/data': 0.01024152699392289, 'timing/model': 1.1522060070419684, 'learning_rate/qwen_vl_interface': 2.1778775851903235e-06, 'learning_rate/action_model': 2.177877585190323e-05, 'epoch': 1.37}) +[2026-06-19 18:07:27,107][starVLA.training.train_starvla][INFO] - Step 3648, Loss: {'train/loss': 0.20161552727222443, 'train/grad_norm_pre_clip': 21.36147689819336, 'timing/data': 0.009510339004918933, 'timing/model': 1.1545445199590176, 'learning_rate/qwen_vl_interface': 2.175555500116881e-06, 'learning_rate/action_model': 2.175555500116881e-05, 'epoch': 1.37}) +[2026-06-19 18:07:31,753][starVLA.training.train_starvla][INFO] - Step 3649, Loss: {'train/loss': 0.2126832753419876, 'train/grad_norm_pre_clip': 15.321398735046387, 'timing/data': 0.009416030952706933, 'timing/model': 1.1585836550220847, 'learning_rate/qwen_vl_interface': 2.1732346788314404e-06, 'learning_rate/action_model': 2.1732346788314402e-05, 'epoch': 1.37}) +[2026-06-19 18:07:36,341][starVLA.training.train_starvla][INFO] - Step 3650, Loss: {'train/loss': 0.35734158754348755, 'train/grad_norm_pre_clip': 23.45352554321289, 'timing/data': 0.009093452012166381, 'timing/model': 1.3126265900209546, 'learning_rate/qwen_vl_interface': 2.170915122288006e-06, 'learning_rate/action_model': 2.170915122288006e-05, 'epoch': 1.37}) +[2026-06-19 18:07:40,778][starVLA.training.train_starvla][INFO] - Step 3651, Loss: {'train/loss': 0.15053044259548187, 'train/grad_norm_pre_clip': 6.940917491912842, 'timing/data': 0.009021935984492302, 'timing/model': 1.146655584918335, 'learning_rate/qwen_vl_interface': 2.168596831440055e-06, 'learning_rate/action_model': 2.1685968314400552e-05, 'epoch': 1.37}) +[2026-06-19 18:07:45,369][starVLA.training.train_starvla][INFO] - Step 3652, Loss: {'train/loss': 0.3043873906135559, 'train/grad_norm_pre_clip': 23.50857925415039, 'timing/data': 0.009364136960357428, 'timing/model': 1.1427781980019063, 'learning_rate/qwen_vl_interface': 2.166279807240553e-06, 'learning_rate/action_model': 2.1662798072405528e-05, 'epoch': 1.37}) +[2026-06-19 18:07:49,978][starVLA.training.train_starvla][INFO] - Step 3653, Loss: {'train/loss': 0.25086843967437744, 'train/grad_norm_pre_clip': 13.32418441772461, 'timing/data': 0.009512350079603493, 'timing/model': 1.1477749990299344, 'learning_rate/qwen_vl_interface': 2.1639640506419394e-06, 'learning_rate/action_model': 2.163964050641939e-05, 'epoch': 1.37}) +[2026-06-19 18:07:54,574][starVLA.training.train_starvla][INFO] - Step 3654, Loss: {'train/loss': 0.18260948359966278, 'train/grad_norm_pre_clip': 11.186284065246582, 'timing/data': 0.00930068304296583, 'timing/model': 1.1518341740593314, 'learning_rate/qwen_vl_interface': 2.161649562596135e-06, 'learning_rate/action_model': 2.161649562596135e-05, 'epoch': 1.37}) +[2026-06-19 18:07:59,169][starVLA.training.train_starvla][INFO] - Step 3655, Loss: {'train/loss': 0.15983355045318604, 'train/grad_norm_pre_clip': 12.02949047088623, 'timing/data': 0.00997733895201236, 'timing/model': 1.148160928976722, 'learning_rate/qwen_vl_interface': 2.1593363440545383e-06, 'learning_rate/action_model': 2.1593363440545384e-05, 'epoch': 1.37}) +[2026-06-19 18:08:03,817][starVLA.training.train_starvla][INFO] - Step 3656, Loss: {'train/loss': 0.2787867486476898, 'train/grad_norm_pre_clip': 10.413562774658203, 'timing/data': 0.009634463931433856, 'timing/model': 1.3401962439529598, 'learning_rate/qwen_vl_interface': 2.157024395968025e-06, 'learning_rate/action_model': 2.1570243959680252e-05, 'epoch': 1.37}) +[2026-06-19 18:08:08,303][starVLA.training.train_starvla][INFO] - Step 3657, Loss: {'train/loss': 0.3562055230140686, 'train/grad_norm_pre_clip': 13.796807289123535, 'timing/data': 0.00970046001020819, 'timing/model': 1.1676344909938052, 'learning_rate/qwen_vl_interface': 2.154713719286952e-06, 'learning_rate/action_model': 2.154713719286952e-05, 'epoch': 1.37}) +[2026-06-19 18:08:12,941][starVLA.training.train_starvla][INFO] - Step 3658, Loss: {'train/loss': 0.25545817613601685, 'train/grad_norm_pre_clip': 26.335620880126953, 'timing/data': 0.0093927780399099, 'timing/model': 1.1511988149723038, 'learning_rate/qwen_vl_interface': 2.152404314961149e-06, 'learning_rate/action_model': 2.152404314961149e-05, 'epoch': 1.37}) +[2026-06-19 18:08:17,482][starVLA.training.train_starvla][INFO] - Step 3659, Loss: {'train/loss': 0.09484254568815231, 'train/grad_norm_pre_clip': 8.25898265838623, 'timing/data': 0.010026619071140885, 'timing/model': 1.1481692659435794, 'learning_rate/qwen_vl_interface': 2.1500961839399264e-06, 'learning_rate/action_model': 2.1500961839399262e-05, 'epoch': 1.37}) +[2026-06-19 18:08:22,040][starVLA.training.train_starvla][INFO] - Step 3660, Loss: {'train/loss': 0.3025193214416504, 'train/grad_norm_pre_clip': 15.971449851989746, 'timing/data': 0.009949099970981479, 'timing/model': 1.1488340479554608, 'learning_rate/qwen_vl_interface': 2.1477893271720683e-06, 'learning_rate/action_model': 2.1477893271720682e-05, 'epoch': 1.37}) +[2026-06-19 18:08:26,668][starVLA.training.train_starvla][INFO] - Step 3661, Loss: {'train/loss': 0.48031654953956604, 'train/grad_norm_pre_clip': 25.411680221557617, 'timing/data': 0.010113458032719791, 'timing/model': 1.1566724529257044, 'learning_rate/qwen_vl_interface': 2.1454837456058385e-06, 'learning_rate/action_model': 2.1454837456058386e-05, 'epoch': 1.37}) +[2026-06-19 18:08:31,268][starVLA.training.train_starvla][INFO] - Step 3662, Loss: {'train/loss': 0.25219693779945374, 'train/grad_norm_pre_clip': 14.420634269714355, 'timing/data': 0.010147257940843701, 'timing/model': 1.3080322680762038, 'learning_rate/qwen_vl_interface': 2.1431794401889722e-06, 'learning_rate/action_model': 2.143179440188972e-05, 'epoch': 1.37}) +[2026-06-19 18:08:35,825][starVLA.training.train_starvla][INFO] - Step 3663, Loss: {'train/loss': 0.13401027023792267, 'train/grad_norm_pre_clip': 8.382442474365234, 'timing/data': 0.00995402003172785, 'timing/model': 1.2716669109649956, 'learning_rate/qwen_vl_interface': 2.1408764118686837e-06, 'learning_rate/action_model': 2.1408764118686838e-05, 'epoch': 1.37}) +[2026-06-19 18:08:40,344][starVLA.training.train_starvla][INFO] - Step 3664, Loss: {'train/loss': 0.25295183062553406, 'train/grad_norm_pre_clip': 11.557644844055176, 'timing/data': 0.009616966941393912, 'timing/model': 1.1530747639480978, 'learning_rate/qwen_vl_interface': 2.1385746615916598e-06, 'learning_rate/action_model': 2.1385746615916596e-05, 'epoch': 1.37}) +[2026-06-19 18:08:44,948][starVLA.training.train_starvla][INFO] - Step 3665, Loss: {'train/loss': 0.3710644245147705, 'train/grad_norm_pre_clip': 22.749656677246094, 'timing/data': 0.009330439032055438, 'timing/model': 1.14730166294612, 'learning_rate/qwen_vl_interface': 2.1362741903040636e-06, 'learning_rate/action_model': 2.1362741903040633e-05, 'epoch': 1.37}) +[2026-06-19 18:08:49,563][starVLA.training.train_starvla][INFO] - Step 3666, Loss: {'train/loss': 0.29561153054237366, 'train/grad_norm_pre_clip': 20.915401458740234, 'timing/data': 0.00980648398399353, 'timing/model': 1.1483117209281772, 'learning_rate/qwen_vl_interface': 2.1339749989515318e-06, 'learning_rate/action_model': 2.1339749989515316e-05, 'epoch': 1.37}) +[2026-06-19 18:08:54,196][starVLA.training.train_starvla][INFO] - Step 3667, Loss: {'train/loss': 0.2697139382362366, 'train/grad_norm_pre_clip': 23.001789093017578, 'timing/data': 0.00931927002966404, 'timing/model': 1.1581232140306383, 'learning_rate/qwen_vl_interface': 2.131677088479175e-06, 'learning_rate/action_model': 2.1316770884791752e-05, 'epoch': 1.37}) +[2026-06-19 18:08:58,716][starVLA.training.train_starvla][INFO] - Step 3668, Loss: {'train/loss': 0.23394520580768585, 'train/grad_norm_pre_clip': 9.962556838989258, 'timing/data': 0.009391256957314909, 'timing/model': 1.151857107062824, 'learning_rate/qwen_vl_interface': 2.1293804598315787e-06, 'learning_rate/action_model': 2.1293804598315787e-05, 'epoch': 1.38}) +[2026-06-19 18:09:03,254][starVLA.training.train_starvla][INFO] - Step 3669, Loss: {'train/loss': 0.4697688817977905, 'train/grad_norm_pre_clip': 11.028915405273438, 'timing/data': 0.009450969053432345, 'timing/model': 1.249428380979225, 'learning_rate/qwen_vl_interface': 2.1270851139527986e-06, 'learning_rate/action_model': 2.1270851139527983e-05, 'epoch': 1.38}) +[2026-06-19 18:09:07,708][starVLA.training.train_starvla][INFO] - Step 3670, Loss: {'train/loss': 0.243641659617424, 'train/grad_norm_pre_clip': 14.615585327148438, 'timing/data': 0.00916843710001558, 'timing/model': 1.153290907968767, 'learning_rate/qwen_vl_interface': 2.1247910517863653e-06, 'learning_rate/action_model': 2.124791051786365e-05, 'epoch': 1.38}) +[2026-06-19 18:09:12,338][starVLA.training.train_starvla][INFO] - Step 3671, Loss: {'train/loss': 0.18633294105529785, 'train/grad_norm_pre_clip': 9.488625526428223, 'timing/data': 0.009785000002011657, 'timing/model': 1.1564736040309072, 'learning_rate/qwen_vl_interface': 2.1224982742752814e-06, 'learning_rate/action_model': 2.1224982742752812e-05, 'epoch': 1.38}) +[2026-06-19 18:09:16,978][starVLA.training.train_starvla][INFO] - Step 3672, Loss: {'train/loss': 0.17348729074001312, 'train/grad_norm_pre_clip': 9.821494102478027, 'timing/data': 0.00940152001567185, 'timing/model': 1.1559179530013353, 'learning_rate/qwen_vl_interface': 2.1202067823620213e-06, 'learning_rate/action_model': 2.1202067823620212e-05, 'epoch': 1.38}) +[2026-06-19 18:09:21,615][starVLA.training.train_starvla][INFO] - Step 3673, Loss: {'train/loss': 0.24969308078289032, 'train/grad_norm_pre_clip': 20.383256912231445, 'timing/data': 0.009596817893907428, 'timing/model': 1.158280019997619, 'learning_rate/qwen_vl_interface': 2.117916576988531e-06, 'learning_rate/action_model': 2.117916576988531e-05, 'epoch': 1.38}) +[2026-06-19 18:09:26,253][starVLA.training.train_starvla][INFO] - Step 3674, Loss: {'train/loss': 0.19004985690116882, 'train/grad_norm_pre_clip': 9.798553466796875, 'timing/data': 0.009653022978454828, 'timing/model': 1.3332878850633278, 'learning_rate/qwen_vl_interface': 2.1156276590962257e-06, 'learning_rate/action_model': 2.1156276590962257e-05, 'epoch': 1.38}) +[2026-06-19 18:09:30,835][starVLA.training.train_starvla][INFO] - Step 3675, Loss: {'train/loss': 0.0979817807674408, 'train/grad_norm_pre_clip': 8.336230278015137, 'timing/data': 0.009999852976761758, 'timing/model': 1.2703749120701104, 'learning_rate/qwen_vl_interface': 2.113340029625994e-06, 'learning_rate/action_model': 2.1133400296259938e-05, 'epoch': 1.38}) +[2026-06-19 18:09:35,312][starVLA.training.train_starvla][INFO] - Step 3676, Loss: {'train/loss': 0.10318142920732498, 'train/grad_norm_pre_clip': 10.524434089660645, 'timing/data': 0.009639624040573835, 'timing/model': 1.167353946948424, 'learning_rate/qwen_vl_interface': 2.111053689518194e-06, 'learning_rate/action_model': 2.111053689518194e-05, 'epoch': 1.38}) +[2026-06-19 18:09:39,959][starVLA.training.train_starvla][INFO] - Step 3677, Loss: {'train/loss': 0.5717877745628357, 'train/grad_norm_pre_clip': 22.104175567626953, 'timing/data': 0.00946635298896581, 'timing/model': 1.1568262699292973, 'learning_rate/qwen_vl_interface': 2.1087686397126544e-06, 'learning_rate/action_model': 2.1087686397126545e-05, 'epoch': 1.38}) +[2026-06-19 18:09:44,582][starVLA.training.train_starvla][INFO] - Step 3678, Loss: {'train/loss': 0.24765583872795105, 'train/grad_norm_pre_clip': 19.410354614257812, 'timing/data': 0.009680156013928354, 'timing/model': 1.1560156900668517, 'learning_rate/qwen_vl_interface': 2.106484881148672e-06, 'learning_rate/action_model': 2.1064848811486718e-05, 'epoch': 1.38}) +[2026-06-19 18:09:49,176][starVLA.training.train_starvla][INFO] - Step 3679, Loss: {'train/loss': 0.3311086595058441, 'train/grad_norm_pre_clip': 20.356489181518555, 'timing/data': 0.00967320497147739, 'timing/model': 1.154789887019433, 'learning_rate/qwen_vl_interface': 2.1042024147650137e-06, 'learning_rate/action_model': 2.1042024147650137e-05, 'epoch': 1.38}) +[2026-06-19 18:09:53,816][starVLA.training.train_starvla][INFO] - Step 3680, Loss: {'train/loss': 0.29114559292793274, 'train/grad_norm_pre_clip': 10.595552444458008, 'timing/data': 0.009779721032828093, 'timing/model': 1.3340407849755138, 'learning_rate/qwen_vl_interface': 2.1019212414999163e-06, 'learning_rate/action_model': 2.101921241499916e-05, 'epoch': 1.38}) +[2026-06-19 18:09:58,332][starVLA.training.train_starvla][INFO] - Step 3681, Loss: {'train/loss': 0.1223861500620842, 'train/grad_norm_pre_clip': 12.017532348632812, 'timing/data': 0.009540974046103656, 'timing/model': 1.2300856679212302, 'learning_rate/qwen_vl_interface': 2.099641362291082e-06, 'learning_rate/action_model': 2.099641362291082e-05, 'epoch': 1.38}) +[2026-06-19 18:10:02,770][starVLA.training.train_starvla][INFO] - Step 3682, Loss: {'train/loss': 0.17237862944602966, 'train/grad_norm_pre_clip': 9.741546630859375, 'timing/data': 0.009720787056721747, 'timing/model': 1.1563229060266167, 'learning_rate/qwen_vl_interface': 2.097362778075685e-06, 'learning_rate/action_model': 2.097362778075685e-05, 'epoch': 1.38}) +[2026-06-19 18:10:07,372][starVLA.training.train_starvla][INFO] - Step 3683, Loss: {'train/loss': 0.20773671567440033, 'train/grad_norm_pre_clip': 20.81049346923828, 'timing/data': 0.0096419119508937, 'timing/model': 1.149345352081582, 'learning_rate/qwen_vl_interface': 2.0950854897903663e-06, 'learning_rate/action_model': 2.095085489790366e-05, 'epoch': 1.38}) +[2026-06-19 18:10:11,966][starVLA.training.train_starvla][INFO] - Step 3684, Loss: {'train/loss': 0.24186252057552338, 'train/grad_norm_pre_clip': 10.553801536560059, 'timing/data': 0.009873294970020652, 'timing/model': 1.145162501023151, 'learning_rate/qwen_vl_interface': 2.092809498371229e-06, 'learning_rate/action_model': 2.0928094983712292e-05, 'epoch': 1.38}) +[2026-06-19 18:10:16,549][starVLA.training.train_starvla][INFO] - Step 3685, Loss: {'train/loss': 0.24901455640792847, 'train/grad_norm_pre_clip': 21.183998107910156, 'timing/data': 0.009313961956650019, 'timing/model': 1.145688656019047, 'learning_rate/qwen_vl_interface': 2.0905348047538498e-06, 'learning_rate/action_model': 2.0905348047538497e-05, 'epoch': 1.38}) +[2026-06-19 18:10:21,170][starVLA.training.train_starvla][INFO] - Step 3686, Loss: {'train/loss': 0.29766547679901123, 'train/grad_norm_pre_clip': 11.327434539794922, 'timing/data': 0.009884958970360458, 'timing/model': 1.3362243940355256, 'learning_rate/qwen_vl_interface': 2.0882614098732686e-06, 'learning_rate/action_model': 2.0882614098732685e-05, 'epoch': 1.38}) +[2026-06-19 18:10:25,686][starVLA.training.train_starvla][INFO] - Step 3687, Loss: {'train/loss': 0.3782965838909149, 'train/grad_norm_pre_clip': 27.85715103149414, 'timing/data': 0.009867202024906874, 'timing/model': 1.239943558932282, 'learning_rate/qwen_vl_interface': 2.085989314663993e-06, 'learning_rate/action_model': 2.0859893146639927e-05, 'epoch': 1.38}) +[2026-06-19 18:10:30,146][starVLA.training.train_starvla][INFO] - Step 3688, Loss: {'train/loss': 0.1612832397222519, 'train/grad_norm_pre_clip': 6.269704341888428, 'timing/data': 0.009847142035141587, 'timing/model': 1.1566345130559057, 'learning_rate/qwen_vl_interface': 2.083718520059995e-06, 'learning_rate/action_model': 2.0837185200599946e-05, 'epoch': 1.38}) +[2026-06-19 18:10:34,739][starVLA.training.train_starvla][INFO] - Step 3689, Loss: {'train/loss': 0.3911273777484894, 'train/grad_norm_pre_clip': 24.686914443969727, 'timing/data': 0.009541873936541378, 'timing/model': 1.1528732270235196, 'learning_rate/qwen_vl_interface': 2.0814490269947117e-06, 'learning_rate/action_model': 2.0814490269947118e-05, 'epoch': 1.38}) +[2026-06-19 18:10:39,379][starVLA.training.train_starvla][INFO] - Step 3690, Loss: {'train/loss': 0.13991329073905945, 'train/grad_norm_pre_clip': 9.857394218444824, 'timing/data': 0.009815286030061543, 'timing/model': 1.155732569983229, 'learning_rate/qwen_vl_interface': 2.079180836401047e-06, 'learning_rate/action_model': 2.079180836401047e-05, 'epoch': 1.38}) +[2026-06-19 18:10:43,992][starVLA.training.train_starvla][INFO] - Step 3691, Loss: {'train/loss': 0.20369678735733032, 'train/grad_norm_pre_clip': 14.258347511291504, 'timing/data': 0.00970977998804301, 'timing/model': 1.1474482719786465, 'learning_rate/qwen_vl_interface': 2.0769139492113687e-06, 'learning_rate/action_model': 2.076913949211369e-05, 'epoch': 1.38}) +[2026-06-19 18:10:48,593][starVLA.training.train_starvla][INFO] - Step 3692, Loss: {'train/loss': 0.16281941533088684, 'train/grad_norm_pre_clip': 10.970534324645996, 'timing/data': 0.009335417998954654, 'timing/model': 1.3132843510247767, 'learning_rate/qwen_vl_interface': 2.0746483663575075e-06, 'learning_rate/action_model': 2.0746483663575076e-05, 'epoch': 1.38}) +[2026-06-19 18:10:53,108][starVLA.training.train_starvla][INFO] - Step 3693, Loss: {'train/loss': 0.07673497498035431, 'train/grad_norm_pre_clip': 6.61563777923584, 'timing/data': 0.009475572034716606, 'timing/model': 1.2379122199490666, 'learning_rate/qwen_vl_interface': 2.0723840887707595e-06, 'learning_rate/action_model': 2.0723840887707594e-05, 'epoch': 1.38}) +[2026-06-19 18:10:57,566][starVLA.training.train_starvla][INFO] - Step 3694, Loss: {'train/loss': 0.39755719900131226, 'train/grad_norm_pre_clip': 25.62896728515625, 'timing/data': 0.009808636037632823, 'timing/model': 1.1505720070563257, 'learning_rate/qwen_vl_interface': 2.0701211173818854e-06, 'learning_rate/action_model': 2.0701211173818853e-05, 'epoch': 1.39}) +[2026-06-19 18:11:02,171][starVLA.training.train_starvla][INFO] - Step 3695, Loss: {'train/loss': 0.30339640378952026, 'train/grad_norm_pre_clip': 24.927221298217773, 'timing/data': 0.010070719057694077, 'timing/model': 1.1505470300326124, 'learning_rate/qwen_vl_interface': 2.0678594531211036e-06, 'learning_rate/action_model': 2.0678594531211033e-05, 'epoch': 1.39}) +[2026-06-19 18:11:06,765][starVLA.training.train_starvla][INFO] - Step 3696, Loss: {'train/loss': 0.37197738885879517, 'train/grad_norm_pre_clip': 17.963674545288086, 'timing/data': 0.009254579083062708, 'timing/model': 1.1503151609795168, 'learning_rate/qwen_vl_interface': 2.0655990969181e-06, 'learning_rate/action_model': 2.0655990969181003e-05, 'epoch': 1.39}) +[2026-06-19 18:11:11,382][starVLA.training.train_starvla][INFO] - Step 3697, Loss: {'train/loss': 0.22262948751449585, 'train/grad_norm_pre_clip': 10.520326614379883, 'timing/data': 0.009374990011565387, 'timing/model': 1.1499285149620846, 'learning_rate/qwen_vl_interface': 2.0633400497020236e-06, 'learning_rate/action_model': 2.0633400497020232e-05, 'epoch': 1.39}) +[2026-06-19 18:11:15,974][starVLA.training.train_starvla][INFO] - Step 3698, Loss: {'train/loss': 0.18356236815452576, 'train/grad_norm_pre_clip': 16.010459899902344, 'timing/data': 0.009620303986594081, 'timing/model': 1.3090721969492733, 'learning_rate/qwen_vl_interface': 2.061082312401481e-06, 'learning_rate/action_model': 2.061082312401481e-05, 'epoch': 1.39}) +[2026-06-19 18:11:20,513][starVLA.training.train_starvla][INFO] - Step 3699, Loss: {'train/loss': 0.26969999074935913, 'train/grad_norm_pre_clip': 14.640081405639648, 'timing/data': 0.009451612015254796, 'timing/model': 1.2406698310514912, 'learning_rate/qwen_vl_interface': 2.058825885944544e-06, 'learning_rate/action_model': 2.0588258859445442e-05, 'epoch': 1.39}) +[2026-06-19 18:11:24,969][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 3700 +[2026-06-19 18:14:18,520][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 3700 +[2026-06-19 18:14:18,521][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 3700 +[2026-06-19 18:17:28,681][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 3700 +[2026-06-19 18:17:28,683][starVLA.training.train_starvla][INFO] - Step 3700, Loss: {'train/loss': 0.19426240026950836, 'train/grad_norm_pre_clip': 12.015912055969238, 'eval/loss': 0.1414996087551117, 'eval/latency_0/loss': 0.09054292738437653, 'eval/latency_1/loss': 0.04500048980116844, 'eval/latency_2/loss': 0.11610617488622665, 'eval/latency_3/loss': 0.1533738076686859, 'eval/latency_4/loss': 0.30261868238449097, 'eval/flappy/loss': 0.1414996087551117, 'eval/flappy/latency_0/loss': 0.09054292738437653, 'eval/flappy/latency_1/loss': 0.04500048980116844, 'eval/flappy/latency_2/loss': 0.11610617488622665, 'eval/flappy/latency_3/loss': 0.1533738076686859, 'eval/flappy/latency_4/loss': 0.30261868238449097, 'eval/flap_precision': 0.6572327044025157, 'eval/flap_recall': 0.76, 'eval/flap_f1': 0.7048903878583473, 'eval/flap_rate_model': 0.08833333333333333, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.7048903878583473, 'eval/latency_0/flap_precision': 0.6572327044025157, 'eval/latency_0/flap_recall': 0.76, 'eval/latency_0/flap_f1': 0.7048903878583473, 'eval/latency_0/flap_rate_model': 0.08833333333333333, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.7048903878583473, 'timing/data': 0.009521718020550907, 'timing/model': 1.153325529070571, 'learning_rate/qwen_vl_interface': 2.0565707712587447e-06, 'learning_rate/action_model': 2.0565707712587444e-05, 'epoch': 1.39}) +[2026-06-19 18:17:28,683][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_3700_state +[2026-06-19 18:17:28,686][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-19 18:17:32,088][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_3700_state/model.safetensors +[2026-06-19 18:17:39,327][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_3700_state/optimizer.bin +[2026-06-19 18:17:39,327][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_3700_state/sampler.bin +[2026-06-19 18:17:39,327][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_3700_state/sampler_1.bin +[2026-06-19 18:17:39,328][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_3700_state/random_states_0.pkl +[2026-06-19 18:17:39,353][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3700 +[2026-06-19 18:17:39,353][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-19 18:17:39,376][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-19 18:17:43,897][starVLA.training.train_starvla][INFO] - Step 3701, Loss: {'train/loss': 0.29024866223335266, 'train/grad_norm_pre_clip': 17.446163177490234, 'timing/data': 0.010146202053874731, 'timing/model': 1.1314239950152114, 'learning_rate/qwen_vl_interface': 2.0543169692710747e-06, 'learning_rate/action_model': 2.054316969271075e-05, 'epoch': 1.39}) +[2026-06-19 18:17:48,351][starVLA.training.train_starvla][INFO] - Step 3702, Loss: {'train/loss': 0.37862178683280945, 'train/grad_norm_pre_clip': 15.249061584472656, 'timing/data': 0.009710184065625072, 'timing/model': 1.1344167590141296, 'learning_rate/qwen_vl_interface': 2.0520644809079873e-06, 'learning_rate/action_model': 2.0520644809079873e-05, 'epoch': 1.39}) +[2026-06-19 18:17:52,848][starVLA.training.train_starvla][INFO] - Step 3703, Loss: {'train/loss': 0.19249175488948822, 'train/grad_norm_pre_clip': 16.22522735595703, 'timing/data': 0.009912776993587613, 'timing/model': 1.2556269220076501, 'learning_rate/qwen_vl_interface': 2.049813307095395e-06, 'learning_rate/action_model': 2.0498133070953947e-05, 'epoch': 1.39}) +[2026-06-19 18:17:57,240][starVLA.training.train_starvla][INFO] - Step 3704, Loss: {'train/loss': 0.282631516456604, 'train/grad_norm_pre_clip': 15.447813034057617, 'timing/data': 0.009994180989451706, 'timing/model': 1.143673848011531, 'learning_rate/qwen_vl_interface': 2.047563448758672e-06, 'learning_rate/action_model': 2.0475634487586717e-05, 'epoch': 1.39}) +[2026-06-19 18:18:01,822][starVLA.training.train_starvla][INFO] - Step 3705, Loss: {'train/loss': 0.23170018196105957, 'train/grad_norm_pre_clip': 10.157273292541504, 'timing/data': 0.009532010997645557, 'timing/model': 1.1322097659576684, 'learning_rate/qwen_vl_interface': 2.045314906822647e-06, 'learning_rate/action_model': 2.0453149068226467e-05, 'epoch': 1.39}) +[2026-06-19 18:18:06,409][starVLA.training.train_starvla][INFO] - Step 3706, Loss: {'train/loss': 0.12821196019649506, 'train/grad_norm_pre_clip': 12.433836936950684, 'timing/data': 0.009616624098271132, 'timing/model': 1.1466571009950712, 'learning_rate/qwen_vl_interface': 2.0430676822116124e-06, 'learning_rate/action_model': 2.0430676822116123e-05, 'epoch': 1.39}) +[2026-06-19 18:18:10,957][starVLA.training.train_starvla][INFO] - Step 3707, Loss: {'train/loss': 0.12835337221622467, 'train/grad_norm_pre_clip': 7.508581161499023, 'timing/data': 0.0101385279558599, 'timing/model': 1.140228918986395, 'learning_rate/qwen_vl_interface': 2.040821775849319e-06, 'learning_rate/action_model': 2.0408217758493188e-05, 'epoch': 1.39}) +[2026-06-19 18:18:15,510][starVLA.training.train_starvla][INFO] - Step 3708, Loss: {'train/loss': 0.18257418274879456, 'train/grad_norm_pre_clip': 10.155977249145508, 'timing/data': 0.010042100097052753, 'timing/model': 1.3029977579135448, 'learning_rate/qwen_vl_interface': 2.03857718865897e-06, 'learning_rate/action_model': 2.0385771886589702e-05, 'epoch': 1.39}) +[2026-06-19 18:18:19,900][starVLA.training.train_starvla][INFO] - Step 3709, Loss: {'train/loss': 0.3990113437175751, 'train/grad_norm_pre_clip': 17.322065353393555, 'timing/data': 0.01002535701263696, 'timing/model': 1.133765181992203, 'learning_rate/qwen_vl_interface': 2.0363339215632337e-06, 'learning_rate/action_model': 2.0363339215632334e-05, 'epoch': 1.39}) +[2026-06-19 18:18:24,481][starVLA.training.train_starvla][INFO] - Step 3710, Loss: {'train/loss': 0.30757516622543335, 'train/grad_norm_pre_clip': 28.116104125976562, 'timing/data': 0.010056251077912748, 'timing/model': 1.1407226789742708, 'learning_rate/qwen_vl_interface': 2.034091975484231e-06, 'learning_rate/action_model': 2.0340919754842307e-05, 'epoch': 1.39}) +[2026-06-19 18:18:29,067][starVLA.training.train_starvla][INFO] - Step 3711, Loss: {'train/loss': 0.3583690822124481, 'train/grad_norm_pre_clip': 13.96304702758789, 'timing/data': 0.010074003017507493, 'timing/model': 1.1454554320080206, 'learning_rate/qwen_vl_interface': 2.0318513513435415e-06, 'learning_rate/action_model': 2.0318513513435415e-05, 'epoch': 1.39}) +[2026-06-19 18:18:33,671][starVLA.training.train_starvla][INFO] - Step 3712, Loss: {'train/loss': 0.3693872392177582, 'train/grad_norm_pre_clip': 18.145166397094727, 'timing/data': 0.01041877397801727, 'timing/model': 1.1453645599540323, 'learning_rate/qwen_vl_interface': 2.0296120500622018e-06, 'learning_rate/action_model': 2.0296120500622017e-05, 'epoch': 1.39}) +[2026-06-19 18:18:38,280][starVLA.training.train_starvla][INFO] - Step 3713, Loss: {'train/loss': 0.3582156300544739, 'train/grad_norm_pre_clip': 21.367050170898438, 'timing/data': 0.010700308019295335, 'timing/model': 1.1478713330579922, 'learning_rate/qwen_vl_interface': 2.0273740725607034e-06, 'learning_rate/action_model': 2.0273740725607032e-05, 'epoch': 1.39}) +[2026-06-19 18:18:42,877][starVLA.training.train_starvla][INFO] - Step 3714, Loss: {'train/loss': 0.22652684152126312, 'train/grad_norm_pre_clip': 13.593384742736816, 'timing/data': 0.010367040988057852, 'timing/model': 1.3286364380037412, 'learning_rate/qwen_vl_interface': 2.0251374197589942e-06, 'learning_rate/action_model': 2.0251374197589943e-05, 'epoch': 1.39}) +[2026-06-19 18:18:47,384][starVLA.training.train_starvla][INFO] - Step 3715, Loss: {'train/loss': 0.13158532977104187, 'train/grad_norm_pre_clip': 9.646493911743164, 'timing/data': 0.010532399057410657, 'timing/model': 1.2384899780154228, 'learning_rate/qwen_vl_interface': 2.0229020925764785e-06, 'learning_rate/action_model': 2.0229020925764785e-05, 'epoch': 1.39}) +[2026-06-19 18:18:51,793][starVLA.training.train_starvla][INFO] - Step 3716, Loss: {'train/loss': 0.3470377027988434, 'train/grad_norm_pre_clip': 15.746170997619629, 'timing/data': 0.01027730107307434, 'timing/model': 1.147592492052354, 'learning_rate/qwen_vl_interface': 2.0206680919320145e-06, 'learning_rate/action_model': 2.020668091932014e-05, 'epoch': 1.39}) +[2026-06-19 18:18:56,391][starVLA.training.train_starvla][INFO] - Step 3717, Loss: {'train/loss': 0.4531524181365967, 'train/grad_norm_pre_clip': 21.913265228271484, 'timing/data': 0.010016156011261046, 'timing/model': 1.1423388159601018, 'learning_rate/qwen_vl_interface': 2.0184354187439164e-06, 'learning_rate/action_model': 2.0184354187439162e-05, 'epoch': 1.39}) +[2026-06-19 18:19:00,991][starVLA.training.train_starvla][INFO] - Step 3718, Loss: {'train/loss': 0.16708797216415405, 'train/grad_norm_pre_clip': 7.329251766204834, 'timing/data': 0.009911864995956421, 'timing/model': 1.1554577560164034, 'learning_rate/qwen_vl_interface': 2.016204073929949e-06, 'learning_rate/action_model': 2.016204073929949e-05, 'epoch': 1.39}) +[2026-06-19 18:19:05,605][starVLA.training.train_starvla][INFO] - Step 3719, Loss: {'train/loss': 0.361603707075119, 'train/grad_norm_pre_clip': 23.759292602539062, 'timing/data': 0.009931890992447734, 'timing/model': 1.1512284120544791, 'learning_rate/qwen_vl_interface': 2.0139740584073363e-06, 'learning_rate/action_model': 2.0139740584073364e-05, 'epoch': 1.39}) +[2026-06-19 18:19:10,202][starVLA.training.train_starvla][INFO] - Step 3720, Loss: {'train/loss': 0.07948276400566101, 'train/grad_norm_pre_clip': 5.646685600280762, 'timing/data': 0.010675022029317915, 'timing/model': 1.3112280449131504, 'learning_rate/qwen_vl_interface': 2.011745373092753e-06, 'learning_rate/action_model': 2.011745373092753e-05, 'epoch': 1.39}) +[2026-06-19 18:19:14,747][starVLA.training.train_starvla][INFO] - Step 3721, Loss: {'train/loss': 0.25606080889701843, 'train/grad_norm_pre_clip': 11.370685577392578, 'timing/data': 0.01058667991310358, 'timing/model': 1.2695947149768472, 'learning_rate/qwen_vl_interface': 2.0095180189023267e-06, 'learning_rate/action_model': 2.0095180189023265e-05, 'epoch': 1.4}) +[2026-06-19 18:19:19,173][starVLA.training.train_starvla][INFO] - Step 3722, Loss: {'train/loss': 0.2710646986961365, 'train/grad_norm_pre_clip': 11.649310111999512, 'timing/data': 0.010570666985586286, 'timing/model': 1.1529045039787889, 'learning_rate/qwen_vl_interface': 2.0072919967516393e-06, 'learning_rate/action_model': 2.0072919967516392e-05, 'epoch': 1.4}) +[2026-06-19 18:19:23,779][starVLA.training.train_starvla][INFO] - Step 3723, Loss: {'train/loss': 0.33024582266807556, 'train/grad_norm_pre_clip': 11.32726001739502, 'timing/data': 0.011239408049732447, 'timing/model': 1.1523661690298468, 'learning_rate/qwen_vl_interface': 2.0050673075557245e-06, 'learning_rate/action_model': 2.0050673075557243e-05, 'epoch': 1.4}) +[2026-06-19 18:19:28,384][starVLA.training.train_starvla][INFO] - Step 3724, Loss: {'train/loss': 0.1770385354757309, 'train/grad_norm_pre_clip': 11.910888671875, 'timing/data': 0.01038293307647109, 'timing/model': 1.1527497170027345, 'learning_rate/qwen_vl_interface': 2.0028439522290677e-06, 'learning_rate/action_model': 2.0028439522290676e-05, 'epoch': 1.4}) +[2026-06-19 18:19:33,008][starVLA.training.train_starvla][INFO] - Step 3725, Loss: {'train/loss': 0.21078425645828247, 'train/grad_norm_pre_clip': 11.726672172546387, 'timing/data': 0.01017663103993982, 'timing/model': 1.154067000024952, 'learning_rate/qwen_vl_interface': 2.000621931685606e-06, 'learning_rate/action_model': 2.000621931685606e-05, 'epoch': 1.4}) +[2026-06-19 18:19:37,616][starVLA.training.train_starvla][INFO] - Step 3726, Loss: {'train/loss': 0.20992133021354675, 'train/grad_norm_pre_clip': 9.945513725280762, 'timing/data': 0.010588072007521987, 'timing/model': 1.3310738409636542, 'learning_rate/qwen_vl_interface': 1.9984012468387285e-06, 'learning_rate/action_model': 1.9984012468387285e-05, 'epoch': 1.4}) +[2026-06-19 18:19:42,166][starVLA.training.train_starvla][INFO] - Step 3727, Loss: {'train/loss': 0.33281970024108887, 'train/grad_norm_pre_clip': 18.964736938476562, 'timing/data': 0.009799563093110919, 'timing/model': 1.2638074979186058, 'learning_rate/qwen_vl_interface': 1.9961818986012767e-06, 'learning_rate/action_model': 1.9961818986012763e-05, 'epoch': 1.4}) +[2026-06-19 18:19:46,606][starVLA.training.train_starvla][INFO] - Step 3728, Loss: {'train/loss': 0.2952732443809509, 'train/grad_norm_pre_clip': 29.853538513183594, 'timing/data': 0.010480239056050777, 'timing/model': 1.147767499089241, 'learning_rate/qwen_vl_interface': 1.9939638878855372e-06, 'learning_rate/action_model': 1.993963887885537e-05, 'epoch': 1.4}) +[2026-06-19 18:19:51,200][starVLA.training.train_starvla][INFO] - Step 3729, Loss: {'train/loss': 0.22333697974681854, 'train/grad_norm_pre_clip': 11.308883666992188, 'timing/data': 0.010588459903374314, 'timing/model': 1.144209195044823, 'learning_rate/qwen_vl_interface': 1.991747215603252e-06, 'learning_rate/action_model': 1.991747215603252e-05, 'epoch': 1.4}) +[2026-06-19 18:19:55,788][starVLA.training.train_starvla][INFO] - Step 3730, Loss: {'train/loss': 0.2713223099708557, 'train/grad_norm_pre_clip': 14.096988677978516, 'timing/data': 0.010020121932029724, 'timing/model': 1.1440175540046766, 'learning_rate/qwen_vl_interface': 1.9895318826656116e-06, 'learning_rate/action_model': 1.9895318826656114e-05, 'epoch': 1.4}) +[2026-06-19 18:20:00,374][starVLA.training.train_starvla][INFO] - Step 3731, Loss: {'train/loss': 0.256550669670105, 'train/grad_norm_pre_clip': 11.915243148803711, 'timing/data': 0.00952995102852583, 'timing/model': 1.1446086840005592, 'learning_rate/qwen_vl_interface': 1.9873178899832553e-06, 'learning_rate/action_model': 1.9873178899832554e-05, 'epoch': 1.4}) +[2026-06-19 18:20:04,977][starVLA.training.train_starvla][INFO] - Step 3732, Loss: {'train/loss': 0.16833274066448212, 'train/grad_norm_pre_clip': 11.889535903930664, 'timing/data': 0.010238592978566885, 'timing/model': 1.1600610939785838, 'learning_rate/qwen_vl_interface': 1.985105238466272e-06, 'learning_rate/action_model': 1.9851052384662722e-05, 'epoch': 1.4}) +[2026-06-19 18:20:09,606][starVLA.training.train_starvla][INFO] - Step 3733, Loss: {'train/loss': 0.22587715089321136, 'train/grad_norm_pre_clip': 14.83095932006836, 'timing/data': 0.010021197027526796, 'timing/model': 1.3306012160610408, 'learning_rate/qwen_vl_interface': 1.9828939290242e-06, 'learning_rate/action_model': 1.9828939290241998e-05, 'epoch': 1.4}) +[2026-06-19 18:20:14,054][starVLA.training.train_starvla][INFO] - Step 3734, Loss: {'train/loss': 0.2032817006111145, 'train/grad_norm_pre_clip': 9.43716812133789, 'timing/data': 0.01008982805069536, 'timing/model': 1.161885907058604, 'learning_rate/qwen_vl_interface': 1.980683962566024e-06, 'learning_rate/action_model': 1.980683962566024e-05, 'epoch': 1.4}) +[2026-06-19 18:20:18,656][starVLA.training.train_starvla][INFO] - Step 3735, Loss: {'train/loss': 0.268207848072052, 'train/grad_norm_pre_clip': 12.765628814697266, 'timing/data': 0.009592885035090148, 'timing/model': 1.1432983661070466, 'learning_rate/qwen_vl_interface': 1.9784753400001784e-06, 'learning_rate/action_model': 1.978475340000178e-05, 'epoch': 1.4}) +[2026-06-19 18:20:23,269][starVLA.training.train_starvla][INFO] - Step 3736, Loss: {'train/loss': 0.3486383855342865, 'train/grad_norm_pre_clip': 15.687233924865723, 'timing/data': 0.010464158956892788, 'timing/model': 1.1439349540742114, 'learning_rate/qwen_vl_interface': 1.9762680622345443e-06, 'learning_rate/action_model': 1.9762680622345442e-05, 'epoch': 1.4}) +[2026-06-19 18:20:27,861][starVLA.training.train_starvla][INFO] - Step 3737, Loss: {'train/loss': 0.18346664309501648, 'train/grad_norm_pre_clip': 9.274117469787598, 'timing/data': 0.009542478015646338, 'timing/model': 1.1456744730239734, 'learning_rate/qwen_vl_interface': 1.9740621301764513e-06, 'learning_rate/action_model': 1.9740621301764513e-05, 'epoch': 1.4}) +[2026-06-19 18:20:32,483][starVLA.training.train_starvla][INFO] - Step 3738, Loss: {'train/loss': 0.34292811155319214, 'train/grad_norm_pre_clip': 18.86258888244629, 'timing/data': 0.009367384016513824, 'timing/model': 1.1525477800751105, 'learning_rate/qwen_vl_interface': 1.971857544732672e-06, 'learning_rate/action_model': 1.971857544732672e-05, 'epoch': 1.4}) +[2026-06-19 18:20:37,082][starVLA.training.train_starvla][INFO] - Step 3739, Loss: {'train/loss': 0.1298619657754898, 'train/grad_norm_pre_clip': 8.791386604309082, 'timing/data': 0.009927125996910036, 'timing/model': 1.3218270969809964, 'learning_rate/qwen_vl_interface': 1.96965430680943e-06, 'learning_rate/action_model': 1.96965430680943e-05, 'epoch': 1.4}) +[2026-06-19 18:20:41,532][starVLA.training.train_starvla][INFO] - Step 3740, Loss: {'train/loss': 0.2867068648338318, 'train/grad_norm_pre_clip': 10.827861785888672, 'timing/data': 0.010031105950474739, 'timing/model': 1.1581581799546257, 'learning_rate/qwen_vl_interface': 1.967452417312393e-06, 'learning_rate/action_model': 1.9674524173123927e-05, 'epoch': 1.4}) +[2026-06-19 18:20:46,123][starVLA.training.train_starvla][INFO] - Step 3741, Loss: {'train/loss': 0.3114369213581085, 'train/grad_norm_pre_clip': 26.887109756469727, 'timing/data': 0.009735158062539995, 'timing/model': 1.146297829109244, 'learning_rate/qwen_vl_interface': 1.9652518771466746e-06, 'learning_rate/action_model': 1.9652518771466747e-05, 'epoch': 1.4}) +[2026-06-19 18:20:50,625][starVLA.training.train_starvla][INFO] - Step 3742, Loss: {'train/loss': 0.14827729761600494, 'train/grad_norm_pre_clip': 9.6415376663208, 'timing/data': 0.009636317961849272, 'timing/model': 1.143189612077549, 'learning_rate/qwen_vl_interface': 1.963052687216834e-06, 'learning_rate/action_model': 1.963052687216834e-05, 'epoch': 1.4}) +[2026-06-19 18:20:55,148][starVLA.training.train_starvla][INFO] - Step 3743, Loss: {'train/loss': 0.1169988363981247, 'train/grad_norm_pre_clip': 5.161088466644287, 'timing/data': 0.010797276976518333, 'timing/model': 1.147426958894357, 'learning_rate/qwen_vl_interface': 1.960854848426874e-06, 'learning_rate/action_model': 1.960854848426874e-05, 'epoch': 1.4}) +[2026-06-19 18:20:59,747][starVLA.training.train_starvla][INFO] - Step 3744, Loss: {'train/loss': 0.5312144756317139, 'train/grad_norm_pre_clip': 23.900522232055664, 'timing/data': 0.009668258018791676, 'timing/model': 1.14338001399301, 'learning_rate/qwen_vl_interface': 1.9586583616802446e-06, 'learning_rate/action_model': 1.9586583616802446e-05, 'epoch': 1.4}) +[2026-06-19 18:21:04,343][starVLA.training.train_starvla][INFO] - Step 3745, Loss: {'train/loss': 0.16454902291297913, 'train/grad_norm_pre_clip': 16.03527069091797, 'timing/data': 0.009353996952995658, 'timing/model': 1.3241151670226827, 'learning_rate/qwen_vl_interface': 1.956463227879838e-06, 'learning_rate/action_model': 1.9564632278798377e-05, 'epoch': 1.4}) +[2026-06-19 18:21:08,786][starVLA.training.train_starvla][INFO] - Step 3746, Loss: {'train/loss': 0.15451563894748688, 'train/grad_norm_pre_clip': 7.370743274688721, 'timing/data': 0.009012258960865438, 'timing/model': 1.1529635309707373, 'learning_rate/qwen_vl_interface': 1.95426944792799e-06, 'learning_rate/action_model': 1.95426944792799e-05, 'epoch': 1.4}) +[2026-06-19 18:21:13,431][starVLA.training.train_starvla][INFO] - Step 3747, Loss: {'train/loss': 0.16740944981575012, 'train/grad_norm_pre_clip': 8.47409439086914, 'timing/data': 0.009880269994027913, 'timing/model': 1.1466702019097283, 'learning_rate/qwen_vl_interface': 1.9520770227264817e-06, 'learning_rate/action_model': 1.9520770227264817e-05, 'epoch': 1.4}) +[2026-06-19 18:21:18,068][starVLA.training.train_starvla][INFO] - Step 3748, Loss: {'train/loss': 0.40966758131980896, 'train/grad_norm_pre_clip': 16.599525451660156, 'timing/data': 0.010241786018013954, 'timing/model': 1.1481340490281582, 'learning_rate/qwen_vl_interface': 1.9498859531765373e-06, 'learning_rate/action_model': 1.9498859531765372e-05, 'epoch': 1.41}) +[2026-06-19 18:21:22,709][starVLA.training.train_starvla][INFO] - Step 3749, Loss: {'train/loss': 0.13253532350063324, 'train/grad_norm_pre_clip': 7.831523895263672, 'timing/data': 0.010508642997592688, 'timing/model': 1.1537900570547208, 'learning_rate/qwen_vl_interface': 1.9476962401788204e-06, 'learning_rate/action_model': 1.9476962401788204e-05, 'epoch': 1.41}) +[2026-06-19 18:21:27,341][starVLA.training.train_starvla][INFO] - Step 3750, Loss: {'train/loss': 0.1456070989370346, 'train/grad_norm_pre_clip': 10.499074935913086, 'timing/data': 0.009900563978590071, 'timing/model': 1.1537986630573869, 'learning_rate/qwen_vl_interface': 1.9455078846334397e-06, 'learning_rate/action_model': 1.9455078846334395e-05, 'epoch': 1.41}) +[2026-06-19 18:21:31,979][starVLA.training.train_starvla][INFO] - Step 3751, Loss: {'train/loss': 0.18700672686100006, 'train/grad_norm_pre_clip': 10.385034561157227, 'timing/data': 0.009982276009395719, 'timing/model': 1.3378067980520427, 'learning_rate/qwen_vl_interface': 1.9433208874399465e-06, 'learning_rate/action_model': 1.9433208874399464e-05, 'epoch': 1.41}) +[2026-06-19 18:21:36,420][starVLA.training.train_starvla][INFO] - Step 3752, Loss: {'train/loss': 0.2622494697570801, 'train/grad_norm_pre_clip': 11.316642761230469, 'timing/data': 0.009644053992815316, 'timing/model': 1.14794343104586, 'learning_rate/qwen_vl_interface': 1.9411352494973314e-06, 'learning_rate/action_model': 1.9411352494973315e-05, 'epoch': 1.41}) +[2026-06-19 18:21:41,047][starVLA.training.train_starvla][INFO] - Step 3753, Loss: {'train/loss': 0.19796496629714966, 'train/grad_norm_pre_clip': 12.99061393737793, 'timing/data': 0.011475911014713347, 'timing/model': 1.153118761954829, 'learning_rate/qwen_vl_interface': 1.938950971704031e-06, 'learning_rate/action_model': 1.9389509717040308e-05, 'epoch': 1.41}) +[2026-06-19 18:21:45,580][starVLA.training.train_starvla][INFO] - Step 3754, Loss: {'train/loss': 0.20683026313781738, 'train/grad_norm_pre_clip': 10.39961051940918, 'timing/data': 0.009236083016730845, 'timing/model': 1.161937133059837, 'learning_rate/qwen_vl_interface': 1.936768054957915e-06, 'learning_rate/action_model': 1.9367680549579147e-05, 'epoch': 1.41}) +[2026-06-19 18:21:50,142][starVLA.training.train_starvla][INFO] - Step 3755, Loss: {'train/loss': 0.277927428483963, 'train/grad_norm_pre_clip': 14.131245613098145, 'timing/data': 0.009902346995659173, 'timing/model': 1.1562248930567876, 'learning_rate/qwen_vl_interface': 1.9345865001562994e-06, 'learning_rate/action_model': 1.9345865001562994e-05, 'epoch': 1.41}) +[2026-06-19 18:21:54,760][starVLA.training.train_starvla][INFO] - Step 3756, Loss: {'train/loss': 0.2762170135974884, 'train/grad_norm_pre_clip': 14.257674217224121, 'timing/data': 0.010039297980256379, 'timing/model': 1.156222140067257, 'learning_rate/qwen_vl_interface': 1.9324063081959394e-06, 'learning_rate/action_model': 1.9324063081959395e-05, 'epoch': 1.41}) +[2026-06-19 18:21:59,347][starVLA.training.train_starvla][INFO] - Step 3757, Loss: {'train/loss': 0.27697283029556274, 'train/grad_norm_pre_clip': 12.432705879211426, 'timing/data': 0.008719939040020108, 'timing/model': 1.3106598580488935, 'learning_rate/qwen_vl_interface': 1.9302274799730304e-06, 'learning_rate/action_model': 1.9302274799730306e-05, 'epoch': 1.41}) +[2026-06-19 18:22:03,903][starVLA.training.train_starvla][INFO] - Step 3758, Loss: {'train/loss': 0.1321418136358261, 'train/grad_norm_pre_clip': 9.637619018554688, 'timing/data': 0.009384279954247177, 'timing/model': 1.2789119109511375, 'learning_rate/qwen_vl_interface': 1.9280500163832055e-06, 'learning_rate/action_model': 1.9280500163832055e-05, 'epoch': 1.41}) +[2026-06-19 18:22:08,343][starVLA.training.train_starvla][INFO] - Step 3759, Loss: {'train/loss': 0.4896707236766815, 'train/grad_norm_pre_clip': 14.68276309967041, 'timing/data': 0.009250505012460053, 'timing/model': 1.1564586670137942, 'learning_rate/qwen_vl_interface': 1.925873918321537e-06, 'learning_rate/action_model': 1.925873918321537e-05, 'epoch': 1.41}) +[2026-06-19 18:22:12,956][starVLA.training.train_starvla][INFO] - Step 3760, Loss: {'train/loss': 0.4806370139122009, 'train/grad_norm_pre_clip': 23.484920501708984, 'timing/data': 0.01053497800603509, 'timing/model': 1.156435261014849, 'learning_rate/qwen_vl_interface': 1.92369918668254e-06, 'learning_rate/action_model': 1.92369918668254e-05, 'epoch': 1.41}) +[2026-06-19 18:22:17,557][starVLA.training.train_starvla][INFO] - Step 3761, Loss: {'train/loss': 0.24346764385700226, 'train/grad_norm_pre_clip': 13.600438117980957, 'timing/data': 0.009970136103220284, 'timing/model': 1.1498324590502307, 'learning_rate/qwen_vl_interface': 1.9215258223601605e-06, 'learning_rate/action_model': 1.9215258223601605e-05, 'epoch': 1.41}) +[2026-06-19 18:22:22,168][starVLA.training.train_starvla][INFO] - Step 3762, Loss: {'train/loss': 0.3516731858253479, 'train/grad_norm_pre_clip': 12.607741355895996, 'timing/data': 0.00967700406908989, 'timing/model': 1.1449529770761728, 'learning_rate/qwen_vl_interface': 1.9193538262477874e-06, 'learning_rate/action_model': 1.9193538262477874e-05, 'epoch': 1.41}) +[2026-06-19 18:22:26,706][starVLA.training.train_starvla][INFO] - Step 3763, Loss: {'train/loss': 0.18242953717708588, 'train/grad_norm_pre_clip': 11.607877731323242, 'timing/data': 0.009960641036741436, 'timing/model': 1.1615097329486161, 'learning_rate/qwen_vl_interface': 1.917183199238247e-06, 'learning_rate/action_model': 1.917183199238247e-05, 'epoch': 1.41}) +[2026-06-19 18:22:31,257][starVLA.training.train_starvla][INFO] - Step 3764, Loss: {'train/loss': 0.17141538858413696, 'train/grad_norm_pre_clip': 15.388888359069824, 'timing/data': 0.009714240091852844, 'timing/model': 1.2633247140329331, 'learning_rate/qwen_vl_interface': 1.9150139422238016e-06, 'learning_rate/action_model': 1.9150139422238018e-05, 'epoch': 1.41}) +[2026-06-19 18:22:35,702][starVLA.training.train_starvla][INFO] - Step 3765, Loss: {'train/loss': 0.15612263977527618, 'train/grad_norm_pre_clip': 11.037120819091797, 'timing/data': 0.009955885005183518, 'timing/model': 1.1499599689850584, 'learning_rate/qwen_vl_interface': 1.9128460560961513e-06, 'learning_rate/action_model': 1.9128460560961514e-05, 'epoch': 1.41}) +[2026-06-19 18:22:40,305][starVLA.training.train_starvla][INFO] - Step 3766, Loss: {'train/loss': 0.3939036726951599, 'train/grad_norm_pre_clip': 17.586435317993164, 'timing/data': 0.009846207103691995, 'timing/model': 1.1557985090184957, 'learning_rate/qwen_vl_interface': 1.910679541746432e-06, 'learning_rate/action_model': 1.9106795417464323e-05, 'epoch': 1.41}) +[2026-06-19 18:22:44,926][starVLA.training.train_starvla][INFO] - Step 3767, Loss: {'train/loss': 0.10892060399055481, 'train/grad_norm_pre_clip': 13.192697525024414, 'timing/data': 0.01008863199967891, 'timing/model': 1.1545730219222605, 'learning_rate/qwen_vl_interface': 1.9085144000652166e-06, 'learning_rate/action_model': 1.9085144000652166e-05, 'epoch': 1.41}) +[2026-06-19 18:22:49,541][starVLA.training.train_starvla][INFO] - Step 3768, Loss: {'train/loss': 0.17916838824748993, 'train/grad_norm_pre_clip': 13.535260200500488, 'timing/data': 0.009975680964998901, 'timing/model': 1.1482300950447097, 'learning_rate/qwen_vl_interface': 1.9063506319425116e-06, 'learning_rate/action_model': 1.9063506319425114e-05, 'epoch': 1.41}) +[2026-06-19 18:22:54,133][starVLA.training.train_starvla][INFO] - Step 3769, Loss: {'train/loss': 0.2073250561952591, 'train/grad_norm_pre_clip': 10.854042053222656, 'timing/data': 0.010908180032856762, 'timing/model': 1.3160401039058343, 'learning_rate/qwen_vl_interface': 1.9041882382677613e-06, 'learning_rate/action_model': 1.9041882382677612e-05, 'epoch': 1.41}) +[2026-06-19 18:22:58,584][starVLA.training.train_starvla][INFO] - Step 3770, Loss: {'train/loss': 0.10926705598831177, 'train/grad_norm_pre_clip': 7.427055358886719, 'timing/data': 0.010028808028437197, 'timing/model': 1.1604658180149272, 'learning_rate/qwen_vl_interface': 1.9020272199298452e-06, 'learning_rate/action_model': 1.902027219929845e-05, 'epoch': 1.41}) +[2026-06-19 18:23:03,197][starVLA.training.train_starvla][INFO] - Step 3771, Loss: {'train/loss': 0.12742891907691956, 'train/grad_norm_pre_clip': 13.087109565734863, 'timing/data': 0.010072155040688813, 'timing/model': 1.1653920840471983, 'learning_rate/qwen_vl_interface': 1.8998675778170736e-06, 'learning_rate/action_model': 1.8998675778170734e-05, 'epoch': 1.41}) +[2026-06-19 18:23:07,801][starVLA.training.train_starvla][INFO] - Step 3772, Loss: {'train/loss': 0.15665990114212036, 'train/grad_norm_pre_clip': 10.785232543945312, 'timing/data': 0.009854493080638349, 'timing/model': 1.1475881619844586, 'learning_rate/qwen_vl_interface': 1.8977093128171955e-06, 'learning_rate/action_model': 1.8977093128171954e-05, 'epoch': 1.41}) +[2026-06-19 18:23:12,422][starVLA.training.train_starvla][INFO] - Step 3773, Loss: {'train/loss': 0.36445462703704834, 'train/grad_norm_pre_clip': 17.261457443237305, 'timing/data': 0.00994481099769473, 'timing/model': 1.1499410460237414, 'learning_rate/qwen_vl_interface': 1.8955524258173919e-06, 'learning_rate/action_model': 1.895552425817392e-05, 'epoch': 1.41}) +[2026-06-19 18:23:17,042][starVLA.training.train_starvla][INFO] - Step 3774, Loss: {'train/loss': 0.13307377696037292, 'train/grad_norm_pre_clip': 9.807280540466309, 'timing/data': 0.010382717940956354, 'timing/model': 1.1559805350843817, 'learning_rate/qwen_vl_interface': 1.8933969177042769e-06, 'learning_rate/action_model': 1.8933969177042768e-05, 'epoch': 1.42}) +[2026-06-19 18:23:21,647][starVLA.training.train_starvla][INFO] - Step 3775, Loss: {'train/loss': 0.25123220682144165, 'train/grad_norm_pre_clip': 13.871001243591309, 'timing/data': 0.010223830002360046, 'timing/model': 1.3170748930424452, 'learning_rate/qwen_vl_interface': 1.8912427893638996e-06, 'learning_rate/action_model': 1.8912427893638996e-05, 'epoch': 1.42}) +[2026-06-19 18:23:26,195][starVLA.training.train_starvla][INFO] - Step 3776, Loss: {'train/loss': 0.20203560590744019, 'train/grad_norm_pre_clip': 13.064844131469727, 'timing/data': 0.010167742031626403, 'timing/model': 1.2548232979606837, 'learning_rate/qwen_vl_interface': 1.8890900416817398e-06, 'learning_rate/action_model': 1.8890900416817397e-05, 'epoch': 1.42}) +[2026-06-19 18:23:30,621][starVLA.training.train_starvla][INFO] - Step 3777, Loss: {'train/loss': 0.1781231015920639, 'train/grad_norm_pre_clip': 9.27495002746582, 'timing/data': 0.00952313793823123, 'timing/model': 1.149624663987197, 'learning_rate/qwen_vl_interface': 1.8869386755427117e-06, 'learning_rate/action_model': 1.8869386755427114e-05, 'epoch': 1.42}) +[2026-06-19 18:23:35,221][starVLA.training.train_starvla][INFO] - Step 3778, Loss: {'train/loss': 0.28248271346092224, 'train/grad_norm_pre_clip': 13.638511657714844, 'timing/data': 0.009780250024050474, 'timing/model': 1.146452407934703, 'learning_rate/qwen_vl_interface': 1.8847886918311597e-06, 'learning_rate/action_model': 1.8847886918311598e-05, 'epoch': 1.42}) +[2026-06-19 18:23:39,811][starVLA.training.train_starvla][INFO] - Step 3779, Loss: {'train/loss': 0.20363590121269226, 'train/grad_norm_pre_clip': 11.821496963500977, 'timing/data': 0.010129946051165462, 'timing/model': 1.1445877769729123, 'learning_rate/qwen_vl_interface': 1.8826400914308616e-06, 'learning_rate/action_model': 1.8826400914308613e-05, 'epoch': 1.42}) +[2026-06-19 18:23:44,395][starVLA.training.train_starvla][INFO] - Step 3780, Loss: {'train/loss': 0.1891826093196869, 'train/grad_norm_pre_clip': 11.343096733093262, 'timing/data': 0.009752650978043675, 'timing/model': 1.1451463989214972, 'learning_rate/qwen_vl_interface': 1.880492875225025e-06, 'learning_rate/action_model': 1.8804928752250252e-05, 'epoch': 1.42}) +[2026-06-19 18:23:49,007][starVLA.training.train_starvla][INFO] - Step 3781, Loss: {'train/loss': 0.3898681104183197, 'train/grad_norm_pre_clip': 18.136035919189453, 'timing/data': 0.010397855076007545, 'timing/model': 1.1457890019519255, 'learning_rate/qwen_vl_interface': 1.8783470440962922e-06, 'learning_rate/action_model': 1.878347044096292e-05, 'epoch': 1.42}) +[2026-06-19 18:23:53,601][starVLA.training.train_starvla][INFO] - Step 3782, Loss: {'train/loss': 0.327067494392395, 'train/grad_norm_pre_clip': 16.89457893371582, 'timing/data': 0.009778616018593311, 'timing/model': 1.3211740120314062, 'learning_rate/qwen_vl_interface': 1.8762025989267293e-06, 'learning_rate/action_model': 1.876202598926729e-05, 'epoch': 1.42}) +[2026-06-19 18:23:58,032][starVLA.training.train_starvla][INFO] - Step 3783, Loss: {'train/loss': 0.11198382824659348, 'train/grad_norm_pre_clip': 5.378623008728027, 'timing/data': 0.010201061959378421, 'timing/model': 1.1464959559962153, 'learning_rate/qwen_vl_interface': 1.874059540597839e-06, 'learning_rate/action_model': 1.874059540597839e-05, 'epoch': 1.42}) +[2026-06-19 18:24:02,651][starVLA.training.train_starvla][INFO] - Step 3784, Loss: {'train/loss': 0.31655293703079224, 'train/grad_norm_pre_clip': 18.61444664001465, 'timing/data': 0.009588047047145665, 'timing/model': 1.1528079130221158, 'learning_rate/qwen_vl_interface': 1.8719178699905508e-06, 'learning_rate/action_model': 1.8719178699905508e-05, 'epoch': 1.42}) +[2026-06-19 18:24:07,243][starVLA.training.train_starvla][INFO] - Step 3785, Loss: {'train/loss': 0.2461947202682495, 'train/grad_norm_pre_clip': 12.547304153442383, 'timing/data': 0.010429895017296076, 'timing/model': 1.1421641210326925, 'learning_rate/qwen_vl_interface': 1.8697775879852242e-06, 'learning_rate/action_model': 1.8697775879852242e-05, 'epoch': 1.42}) +[2026-06-19 18:24:11,847][starVLA.training.train_starvla][INFO] - Step 3786, Loss: {'train/loss': 0.3080936372280121, 'train/grad_norm_pre_clip': 19.219083786010742, 'timing/data': 0.00972786103375256, 'timing/model': 1.1493312780512497, 'learning_rate/qwen_vl_interface': 1.86763869546165e-06, 'learning_rate/action_model': 1.86763869546165e-05, 'epoch': 1.42}) +[2026-06-19 18:24:16,469][starVLA.training.train_starvla][INFO] - Step 3787, Loss: {'train/loss': 0.13309946656227112, 'train/grad_norm_pre_clip': 13.24649429321289, 'timing/data': 0.009793533012270927, 'timing/model': 1.1507345060817897, 'learning_rate/qwen_vl_interface': 1.865501193299045e-06, 'learning_rate/action_model': 1.8655011932990447e-05, 'epoch': 1.42}) +[2026-06-19 18:24:21,107][starVLA.training.train_starvla][INFO] - Step 3788, Loss: {'train/loss': 0.31614819169044495, 'train/grad_norm_pre_clip': 16.559165954589844, 'timing/data': 0.010061282897368073, 'timing/model': 1.3408401309279725, 'learning_rate/qwen_vl_interface': 1.8633650823760555e-06, 'learning_rate/action_model': 1.8633650823760555e-05, 'epoch': 1.42}) +[2026-06-19 18:24:25,565][starVLA.training.train_starvla][INFO] - Step 3789, Loss: {'train/loss': 0.18117381632328033, 'train/grad_norm_pre_clip': 12.858051300048828, 'timing/data': 0.010281058959662914, 'timing/model': 1.149960667011328, 'learning_rate/qwen_vl_interface': 1.8612303635707562e-06, 'learning_rate/action_model': 1.861230363570756e-05, 'epoch': 1.42}) +[2026-06-19 18:24:30,167][starVLA.training.train_starvla][INFO] - Step 3790, Loss: {'train/loss': 0.09256873279809952, 'train/grad_norm_pre_clip': 10.757080078125, 'timing/data': 0.010092634009197354, 'timing/model': 1.154430898022838, 'learning_rate/qwen_vl_interface': 1.8590970377606494e-06, 'learning_rate/action_model': 1.859097037760649e-05, 'epoch': 1.42}) +[2026-06-19 18:24:34,694][starVLA.training.train_starvla][INFO] - Step 3791, Loss: {'train/loss': 0.25975123047828674, 'train/grad_norm_pre_clip': 15.169833183288574, 'timing/data': 0.010210687993094325, 'timing/model': 1.1506730680121109, 'learning_rate/qwen_vl_interface': 1.8569651058226667e-06, 'learning_rate/action_model': 1.8569651058226666e-05, 'epoch': 1.42}) +[2026-06-19 18:24:39,252][starVLA.training.train_starvla][INFO] - Step 3792, Loss: {'train/loss': 0.16133354604244232, 'train/grad_norm_pre_clip': 4.5867018699646, 'timing/data': 0.009538771933875978, 'timing/model': 1.1540594960097224, 'learning_rate/qwen_vl_interface': 1.854834568633161e-06, 'learning_rate/action_model': 1.854834568633161e-05, 'epoch': 1.42}) +[2026-06-19 18:24:43,838][starVLA.training.train_starvla][INFO] - Step 3793, Loss: {'train/loss': 0.2928815186023712, 'train/grad_norm_pre_clip': 10.511006355285645, 'timing/data': 0.010204612044617534, 'timing/model': 1.145530715934001, 'learning_rate/qwen_vl_interface': 1.8527054270679174e-06, 'learning_rate/action_model': 1.8527054270679172e-05, 'epoch': 1.42}) +[2026-06-19 18:24:48,529][starVLA.training.train_starvla][INFO] - Step 3794, Loss: {'train/loss': 0.18990615010261536, 'train/grad_norm_pre_clip': 8.92711353302002, 'timing/data': 0.010328756994567811, 'timing/model': 1.3330831889761612, 'learning_rate/qwen_vl_interface': 1.850577682002146e-06, 'learning_rate/action_model': 1.850577682002146e-05, 'epoch': 1.42}) +[2026-06-19 18:24:52,995][starVLA.training.train_starvla][INFO] - Step 3795, Loss: {'train/loss': 0.17944054305553436, 'train/grad_norm_pre_clip': 11.506711959838867, 'timing/data': 0.009777695988304913, 'timing/model': 1.15221297100652, 'learning_rate/qwen_vl_interface': 1.848451334310482e-06, 'learning_rate/action_model': 1.848451334310482e-05, 'epoch': 1.42}) +[2026-06-19 18:24:57,641][starVLA.training.train_starvla][INFO] - Step 3796, Loss: {'train/loss': 0.319320946931839, 'train/grad_norm_pre_clip': 14.709458351135254, 'timing/data': 0.009867504006251693, 'timing/model': 1.146635175915435, 'learning_rate/qwen_vl_interface': 1.8463263848669865e-06, 'learning_rate/action_model': 1.8463263848669864e-05, 'epoch': 1.42}) +[2026-06-19 18:25:02,284][starVLA.training.train_starvla][INFO] - Step 3797, Loss: {'train/loss': 0.22505608201026917, 'train/grad_norm_pre_clip': 16.792396545410156, 'timing/data': 0.010194678092375398, 'timing/model': 1.14911292004399, 'learning_rate/qwen_vl_interface': 1.8442028345451468e-06, 'learning_rate/action_model': 1.8442028345451468e-05, 'epoch': 1.42}) +[2026-06-19 18:25:06,835][starVLA.training.train_starvla][INFO] - Step 3798, Loss: {'train/loss': 0.13246911764144897, 'train/grad_norm_pre_clip': 19.413978576660156, 'timing/data': 0.010060717002488673, 'timing/model': 1.1521516990615055, 'learning_rate/qwen_vl_interface': 1.8420806842178746e-06, 'learning_rate/action_model': 1.8420806842178745e-05, 'epoch': 1.42}) +[2026-06-19 18:25:11,440][starVLA.training.train_starvla][INFO] - Step 3799, Loss: {'train/loss': 0.18440139293670654, 'train/grad_norm_pre_clip': 12.175398826599121, 'timing/data': 0.0098503059707582, 'timing/model': 1.145372977014631, 'learning_rate/qwen_vl_interface': 1.8399599347575034e-06, 'learning_rate/action_model': 1.8399599347575033e-05, 'epoch': 1.42}) +[2026-06-19 18:25:16,042][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 3800 +[2026-06-19 18:28:11,170][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 3800 +[2026-06-19 18:28:11,170][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 3800 +[2026-06-19 18:31:17,304][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 3800 +[2026-06-19 18:31:17,306][starVLA.training.train_starvla][INFO] - Step 3800, Loss: {'train/loss': 0.23814460635185242, 'train/grad_norm_pre_clip': 14.50703239440918, 'eval/loss': 0.13955935835838318, 'eval/latency_0/loss': 0.08681293576955795, 'eval/latency_1/loss': 0.0495661236345768, 'eval/latency_2/loss': 0.10762900114059448, 'eval/latency_3/loss': 0.16018220782279968, 'eval/latency_4/loss': 0.2946990728378296, 'eval/flappy/loss': 0.13955935835838318, 'eval/flappy/latency_0/loss': 0.08681293576955795, 'eval/flappy/latency_1/loss': 0.0495661236345768, 'eval/flappy/latency_2/loss': 0.10762900114059448, 'eval/flappy/latency_3/loss': 0.16018220782279968, 'eval/flappy/latency_4/loss': 0.2946990728378296, 'eval/flap_precision': 0.7289377289377289, 'eval/flap_recall': 0.7236363636363636, 'eval/flap_f1': 0.7262773722627737, 'eval/flap_rate_model': 0.07583333333333334, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.7262773722627737, 'eval/latency_0/flap_precision': 0.7289377289377289, 'eval/latency_0/flap_recall': 0.7236363636363636, 'eval/latency_0/flap_f1': 0.7262773722627737, 'eval/latency_0/flap_rate_model': 0.07583333333333334, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.7262773722627737, 'timing/data': 0.009430850972421467, 'timing/model': 1.3195179200265557, 'learning_rate/qwen_vl_interface': 1.8378405870357943e-06, 'learning_rate/action_model': 1.837840587035794e-05, 'epoch': 1.42}) +[2026-06-19 18:31:17,306][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_3800_state +[2026-06-19 18:31:17,309][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-19 18:31:20,707][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_3800_state/model.safetensors +[2026-06-19 18:31:27,821][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_3800_state/optimizer.bin +[2026-06-19 18:31:27,821][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_3800_state/sampler.bin +[2026-06-19 18:31:27,821][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_3800_state/sampler_1.bin +[2026-06-19 18:31:27,822][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_3800_state/random_states_0.pkl +[2026-06-19 18:31:27,845][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3800 +[2026-06-19 18:31:27,845][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-19 18:31:27,940][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-19 18:31:32,493][starVLA.training.train_starvla][INFO] - Step 3801, Loss: {'train/loss': 0.2115631103515625, 'train/grad_norm_pre_clip': 9.419445037841797, 'timing/data': 0.010754744987934828, 'timing/model': 1.3241881649009883, 'learning_rate/qwen_vl_interface': 1.8357226419239316e-06, 'learning_rate/action_model': 1.8357226419239315e-05, 'epoch': 1.43}) +[2026-06-19 18:31:36,938][starVLA.training.train_starvla][INFO] - Step 3802, Loss: {'train/loss': 0.21361969411373138, 'train/grad_norm_pre_clip': 13.848898887634277, 'timing/data': 0.010236490052193403, 'timing/model': 1.2016520649194717, 'learning_rate/qwen_vl_interface': 1.8336061002925224e-06, 'learning_rate/action_model': 1.8336061002925224e-05, 'epoch': 1.43}) +[2026-06-19 18:31:41,339][starVLA.training.train_starvla][INFO] - Step 3803, Loss: {'train/loss': 0.13756495714187622, 'train/grad_norm_pre_clip': 11.84807300567627, 'timing/data': 0.01043292204849422, 'timing/model': 1.1492852510418743, 'learning_rate/qwen_vl_interface': 1.8314909630115985e-06, 'learning_rate/action_model': 1.8314909630115983e-05, 'epoch': 1.43}) +[2026-06-19 18:31:45,839][starVLA.training.train_starvla][INFO] - Step 3804, Loss: {'train/loss': 0.2648504376411438, 'train/grad_norm_pre_clip': 11.305638313293457, 'timing/data': 0.010807879967615008, 'timing/model': 1.1430147839710116, 'learning_rate/qwen_vl_interface': 1.8293772309506092e-06, 'learning_rate/action_model': 1.829377230950609e-05, 'epoch': 1.43}) +[2026-06-19 18:31:50,372][starVLA.training.train_starvla][INFO] - Step 3805, Loss: {'train/loss': 0.12238872051239014, 'train/grad_norm_pre_clip': 5.882741928100586, 'timing/data': 0.010472486959770322, 'timing/model': 1.1336284009739757, 'learning_rate/qwen_vl_interface': 1.8272649049784326e-06, 'learning_rate/action_model': 1.8272649049784323e-05, 'epoch': 1.43}) +[2026-06-19 18:31:54,908][starVLA.training.train_starvla][INFO] - Step 3806, Loss: {'train/loss': 0.26425701379776, 'train/grad_norm_pre_clip': 13.695838928222656, 'timing/data': 0.01060783106368035, 'timing/model': 1.1335538240382448, 'learning_rate/qwen_vl_interface': 1.8251539859633653e-06, 'learning_rate/action_model': 1.825153985963365e-05, 'epoch': 1.43}) +[2026-06-19 18:31:59,466][starVLA.training.train_starvla][INFO] - Step 3807, Loss: {'train/loss': 0.4398588240146637, 'train/grad_norm_pre_clip': 17.37403678894043, 'timing/data': 0.010725409025326371, 'timing/model': 1.3096230870578438, 'learning_rate/qwen_vl_interface': 1.8230444747731265e-06, 'learning_rate/action_model': 1.8230444747731264e-05, 'epoch': 1.43}) +[2026-06-19 18:32:03,864][starVLA.training.train_starvla][INFO] - Step 3808, Loss: {'train/loss': 0.30042609572410583, 'train/grad_norm_pre_clip': 27.86289405822754, 'timing/data': 0.010879407986067235, 'timing/model': 1.1388503639027476, 'learning_rate/qwen_vl_interface': 1.820936372274857e-06, 'learning_rate/action_model': 1.820936372274857e-05, 'epoch': 1.43}) +[2026-06-19 18:32:08,440][starVLA.training.train_starvla][INFO] - Step 3809, Loss: {'train/loss': 0.49719229340553284, 'train/grad_norm_pre_clip': 26.128616333007812, 'timing/data': 0.010658768936991692, 'timing/model': 1.1522843289421871, 'learning_rate/qwen_vl_interface': 1.8188296793351178e-06, 'learning_rate/action_model': 1.818829679335118e-05, 'epoch': 1.43}) +[2026-06-19 18:32:12,954][starVLA.training.train_starvla][INFO] - Step 3810, Loss: {'train/loss': 0.05927027389407158, 'train/grad_norm_pre_clip': 5.453031063079834, 'timing/data': 0.010302908020094037, 'timing/model': 1.135984548018314, 'learning_rate/qwen_vl_interface': 1.8167243968198905e-06, 'learning_rate/action_model': 1.8167243968198905e-05, 'epoch': 1.43}) +[2026-06-19 18:32:17,536][starVLA.training.train_starvla][INFO] - Step 3811, Loss: {'train/loss': 0.18946489691734314, 'train/grad_norm_pre_clip': 12.854740142822266, 'timing/data': 0.010335584986023605, 'timing/model': 1.1380156230879948, 'learning_rate/qwen_vl_interface': 1.8146205255945775e-06, 'learning_rate/action_model': 1.8146205255945777e-05, 'epoch': 1.43}) +[2026-06-19 18:32:22,106][starVLA.training.train_starvla][INFO] - Step 3812, Loss: {'train/loss': 0.32713043689727783, 'train/grad_norm_pre_clip': 14.158012390136719, 'timing/data': 0.01058110105805099, 'timing/model': 1.1364580440567806, 'learning_rate/qwen_vl_interface': 1.8125180665240007e-06, 'learning_rate/action_model': 1.8125180665240008e-05, 'epoch': 1.43}) +[2026-06-19 18:32:26,685][starVLA.training.train_starvla][INFO] - Step 3813, Loss: {'train/loss': 0.3839752972126007, 'train/grad_norm_pre_clip': 17.527311325073242, 'timing/data': 0.009716611006297171, 'timing/model': 1.3119812610093504, 'learning_rate/qwen_vl_interface': 1.8104170204724025e-06, 'learning_rate/action_model': 1.8104170204724024e-05, 'epoch': 1.43}) +[2026-06-19 18:32:31,081][starVLA.training.train_starvla][INFO] - Step 3814, Loss: {'train/loss': 0.2781873345375061, 'train/grad_norm_pre_clip': 17.976703643798828, 'timing/data': 0.010199708049185574, 'timing/model': 1.140044527943246, 'learning_rate/qwen_vl_interface': 1.8083173883034435e-06, 'learning_rate/action_model': 1.8083173883034435e-05, 'epoch': 1.43}) +[2026-06-19 18:32:35,656][starVLA.training.train_starvla][INFO] - Step 3815, Loss: {'train/loss': 0.2186058759689331, 'train/grad_norm_pre_clip': 8.863517761230469, 'timing/data': 0.010070123942568898, 'timing/model': 1.141878776019439, 'learning_rate/qwen_vl_interface': 1.8062191708802013e-06, 'learning_rate/action_model': 1.8062191708802012e-05, 'epoch': 1.43}) +[2026-06-19 18:32:40,157][starVLA.training.train_starvla][INFO] - Step 3816, Loss: {'train/loss': 0.2501266300678253, 'train/grad_norm_pre_clip': 19.70375633239746, 'timing/data': 0.010205292026512325, 'timing/model': 1.1407461100025102, 'learning_rate/qwen_vl_interface': 1.804122369065175e-06, 'learning_rate/action_model': 1.8041223690651748e-05, 'epoch': 1.43}) +[2026-06-19 18:32:44,734][starVLA.training.train_starvla][INFO] - Step 3817, Loss: {'train/loss': 0.2532273530960083, 'train/grad_norm_pre_clip': 18.931055068969727, 'timing/data': 0.010668397066183388, 'timing/model': 1.1416431500110775, 'learning_rate/qwen_vl_interface': 1.8020269837202806e-06, 'learning_rate/action_model': 1.8020269837202804e-05, 'epoch': 1.43}) +[2026-06-19 18:32:49,349][starVLA.training.train_starvla][INFO] - Step 3818, Loss: {'train/loss': 0.48503708839416504, 'train/grad_norm_pre_clip': 25.325084686279297, 'timing/data': 0.010309042991138995, 'timing/model': 1.1473628629464656, 'learning_rate/qwen_vl_interface': 1.7999330157068517e-06, 'learning_rate/action_model': 1.7999330157068515e-05, 'epoch': 1.43}) +[2026-06-19 18:32:53,936][starVLA.training.train_starvla][INFO] - Step 3819, Loss: {'train/loss': 0.27539825439453125, 'train/grad_norm_pre_clip': 14.98741340637207, 'timing/data': 0.009558416903018951, 'timing/model': 1.3153108629630879, 'learning_rate/qwen_vl_interface': 1.7978404658856393e-06, 'learning_rate/action_model': 1.7978404658856394e-05, 'epoch': 1.43}) +[2026-06-19 18:32:58,373][starVLA.training.train_starvla][INFO] - Step 3820, Loss: {'train/loss': 0.35550904273986816, 'train/grad_norm_pre_clip': 18.264768600463867, 'timing/data': 0.010618913103826344, 'timing/model': 1.1424678759649396, 'learning_rate/qwen_vl_interface': 1.7957493351168126e-06, 'learning_rate/action_model': 1.7957493351168126e-05, 'epoch': 1.43}) +[2026-06-19 18:33:02,987][starVLA.training.train_starvla][INFO] - Step 3821, Loss: {'train/loss': 0.20839066803455353, 'train/grad_norm_pre_clip': 12.27216911315918, 'timing/data': 0.009740649024024606, 'timing/model': 1.159696616930887, 'learning_rate/qwen_vl_interface': 1.7936596242599552e-06, 'learning_rate/action_model': 1.7936596242599553e-05, 'epoch': 1.43}) +[2026-06-19 18:33:07,542][starVLA.training.train_starvla][INFO] - Step 3822, Loss: {'train/loss': 0.37188097834587097, 'train/grad_norm_pre_clip': 15.884988784790039, 'timing/data': 0.00994846795219928, 'timing/model': 1.1488897639792413, 'learning_rate/qwen_vl_interface': 1.7915713341740692e-06, 'learning_rate/action_model': 1.791571334174069e-05, 'epoch': 1.43}) +[2026-06-19 18:33:12,161][starVLA.training.train_starvla][INFO] - Step 3823, Loss: {'train/loss': 0.25880423188209534, 'train/grad_norm_pre_clip': 20.027244567871094, 'timing/data': 0.010007926961407065, 'timing/model': 1.1438035260653123, 'learning_rate/qwen_vl_interface': 1.7894844657175716e-06, 'learning_rate/action_model': 1.7894844657175717e-05, 'epoch': 1.43}) +[2026-06-19 18:33:16,788][starVLA.training.train_starvla][INFO] - Step 3824, Loss: {'train/loss': 0.29148074984550476, 'train/grad_norm_pre_clip': 15.927247047424316, 'timing/data': 0.01028313604183495, 'timing/model': 1.1497829389991239, 'learning_rate/qwen_vl_interface': 1.7873990197482962e-06, 'learning_rate/action_model': 1.787399019748296e-05, 'epoch': 1.43}) +[2026-06-19 18:33:21,426][starVLA.training.train_starvla][INFO] - Step 3825, Loss: {'train/loss': 0.20841988921165466, 'train/grad_norm_pre_clip': 11.253519058227539, 'timing/data': 0.009951318032108247, 'timing/model': 1.3282756590051576, 'learning_rate/qwen_vl_interface': 1.7853149971234883e-06, 'learning_rate/action_model': 1.785314997123488e-05, 'epoch': 1.43}) +[2026-06-19 18:33:25,863][starVLA.training.train_starvla][INFO] - Step 3826, Loss: {'train/loss': 0.2791392207145691, 'train/grad_norm_pre_clip': 15.719300270080566, 'timing/data': 0.010408399975858629, 'timing/model': 1.1430260560009629, 'learning_rate/qwen_vl_interface': 1.7832323986998128e-06, 'learning_rate/action_model': 1.783232398699813e-05, 'epoch': 1.43}) +[2026-06-19 18:33:30,465][starVLA.training.train_starvla][INFO] - Step 3827, Loss: {'train/loss': 0.1338464319705963, 'train/grad_norm_pre_clip': 22.516082763671875, 'timing/data': 0.009905455051921308, 'timing/model': 1.1439006220316514, 'learning_rate/qwen_vl_interface': 1.7811512253333468e-06, 'learning_rate/action_model': 1.7811512253333467e-05, 'epoch': 1.43}) +[2026-06-19 18:33:35,049][starVLA.training.train_starvla][INFO] - Step 3828, Loss: {'train/loss': 0.3938056528568268, 'train/grad_norm_pre_clip': 22.638582229614258, 'timing/data': 0.011642497032880783, 'timing/model': 1.1482593440450728, 'learning_rate/qwen_vl_interface': 1.7790714778795823e-06, 'learning_rate/action_model': 1.779071477879582e-05, 'epoch': 1.44}) +[2026-06-19 18:33:39,557][starVLA.training.train_starvla][INFO] - Step 3829, Loss: {'train/loss': 0.21015194058418274, 'train/grad_norm_pre_clip': 8.192989349365234, 'timing/data': 0.010416870005428791, 'timing/model': 1.1467343639815226, 'learning_rate/qwen_vl_interface': 1.7769931571934246e-06, 'learning_rate/action_model': 1.7769931571934246e-05, 'epoch': 1.44}) +[2026-06-19 18:33:44,063][starVLA.training.train_starvla][INFO] - Step 3830, Loss: {'train/loss': 0.26001861691474915, 'train/grad_norm_pre_clip': 7.379389762878418, 'timing/data': 0.010446469997987151, 'timing/model': 1.1461170429829508, 'learning_rate/qwen_vl_interface': 1.7749162641291933e-06, 'learning_rate/action_model': 1.7749162641291933e-05, 'epoch': 1.44}) +[2026-06-19 18:33:48,591][starVLA.training.train_starvla][INFO] - Step 3831, Loss: {'train/loss': 0.14164048433303833, 'train/grad_norm_pre_clip': 10.307721138000488, 'timing/data': 0.010357443010434508, 'timing/model': 1.2445540950866416, 'learning_rate/qwen_vl_interface': 1.7728407995406206e-06, 'learning_rate/action_model': 1.7728407995406205e-05, 'epoch': 1.44}) +[2026-06-19 18:33:53,030][starVLA.training.train_starvla][INFO] - Step 3832, Loss: {'train/loss': 0.1433631330728531, 'train/grad_norm_pre_clip': 8.299893379211426, 'timing/data': 0.010072175064124167, 'timing/model': 1.154977745958604, 'learning_rate/qwen_vl_interface': 1.7707667642808513e-06, 'learning_rate/action_model': 1.7707667642808512e-05, 'epoch': 1.44}) +[2026-06-19 18:33:57,635][starVLA.training.train_starvla][INFO] - Step 3833, Loss: {'train/loss': 0.25024282932281494, 'train/grad_norm_pre_clip': 10.24218463897705, 'timing/data': 0.010267463978379965, 'timing/model': 1.1473121710587293, 'learning_rate/qwen_vl_interface': 1.7686941592024423e-06, 'learning_rate/action_model': 1.7686941592024423e-05, 'epoch': 1.44}) +[2026-06-19 18:34:02,230][starVLA.training.train_starvla][INFO] - Step 3834, Loss: {'train/loss': 0.26316750049591064, 'train/grad_norm_pre_clip': 12.120654106140137, 'timing/data': 0.00984944705851376, 'timing/model': 1.1491246180376038, 'learning_rate/qwen_vl_interface': 1.7666229851573662e-06, 'learning_rate/action_model': 1.7666229851573664e-05, 'epoch': 1.44}) +[2026-06-19 18:34:06,840][starVLA.training.train_starvla][INFO] - Step 3835, Loss: {'train/loss': 0.1365562528371811, 'train/grad_norm_pre_clip': 10.626558303833008, 'timing/data': 0.01018042303621769, 'timing/model': 1.1456494050798938, 'learning_rate/qwen_vl_interface': 1.7645532429970006e-06, 'learning_rate/action_model': 1.7645532429970004e-05, 'epoch': 1.44}) +[2026-06-19 18:34:11,446][starVLA.training.train_starvla][INFO] - Step 3836, Loss: {'train/loss': 0.18517588078975677, 'train/grad_norm_pre_clip': 13.51138687133789, 'timing/data': 0.009472763049416244, 'timing/model': 1.3191682480974123, 'learning_rate/qwen_vl_interface': 1.7624849335721402e-06, 'learning_rate/action_model': 1.7624849335721402e-05, 'epoch': 1.44}) +[2026-06-19 18:34:16,040][starVLA.training.train_starvla][INFO] - Step 3837, Loss: {'train/loss': 0.1535561978816986, 'train/grad_norm_pre_clip': 10.608800888061523, 'timing/data': 0.009988063015043736, 'timing/model': 1.2878191870404407, 'learning_rate/qwen_vl_interface': 1.7604180577329883e-06, 'learning_rate/action_model': 1.7604180577329883e-05, 'epoch': 1.44}) +[2026-06-19 18:34:20,492][starVLA.training.train_starvla][INFO] - Step 3838, Loss: {'train/loss': 0.15735551714897156, 'train/grad_norm_pre_clip': 8.392816543579102, 'timing/data': 0.010311260004527867, 'timing/model': 1.149308888008818, 'learning_rate/qwen_vl_interface': 1.7583526163291604e-06, 'learning_rate/action_model': 1.7583526163291604e-05, 'epoch': 1.44}) +[2026-06-19 18:34:25,106][starVLA.training.train_starvla][INFO] - Step 3839, Loss: {'train/loss': 0.3564307689666748, 'train/grad_norm_pre_clip': 25.160154342651367, 'timing/data': 0.010300104040652514, 'timing/model': 1.1460572798969224, 'learning_rate/qwen_vl_interface': 1.7562886102096807e-06, 'learning_rate/action_model': 1.7562886102096807e-05, 'epoch': 1.44}) +[2026-06-19 18:34:29,736][starVLA.training.train_starvla][INFO] - Step 3840, Loss: {'train/loss': 0.24778170883655548, 'train/grad_norm_pre_clip': 11.058509826660156, 'timing/data': 0.010373747092671692, 'timing/model': 1.1596708180150017, 'learning_rate/qwen_vl_interface': 1.7542260402229838e-06, 'learning_rate/action_model': 1.7542260402229838e-05, 'epoch': 1.44}) +[2026-06-19 18:34:34,354][starVLA.training.train_starvla][INFO] - Step 3841, Loss: {'train/loss': 0.4105110466480255, 'train/grad_norm_pre_clip': 10.867531776428223, 'timing/data': 0.010427343077026308, 'timing/model': 1.148819466936402, 'learning_rate/qwen_vl_interface': 1.7521649072169152e-06, 'learning_rate/action_model': 1.752164907216915e-05, 'epoch': 1.44}) +[2026-06-19 18:34:38,970][starVLA.training.train_starvla][INFO] - Step 3842, Loss: {'train/loss': 0.19679108262062073, 'train/grad_norm_pre_clip': 8.482881546020508, 'timing/data': 0.010243544937111437, 'timing/model': 1.1497421819949523, 'learning_rate/qwen_vl_interface': 1.7501052120387285e-06, 'learning_rate/action_model': 1.7501052120387285e-05, 'epoch': 1.44}) +[2026-06-19 18:34:43,554][starVLA.training.train_starvla][INFO] - Step 3843, Loss: {'train/loss': 0.1177670955657959, 'train/grad_norm_pre_clip': 8.37081241607666, 'timing/data': 0.010398257989436388, 'timing/model': 1.3064486070070416, 'learning_rate/qwen_vl_interface': 1.7480469555350866e-06, 'learning_rate/action_model': 1.7480469555350865e-05, 'epoch': 1.44}) +[2026-06-19 18:34:47,987][starVLA.training.train_starvla][INFO] - Step 3844, Loss: {'train/loss': 0.15730826556682587, 'train/grad_norm_pre_clip': 8.05276107788086, 'timing/data': 0.01070905092637986, 'timing/model': 1.147413837024942, 'learning_rate/qwen_vl_interface': 1.7459901385520619e-06, 'learning_rate/action_model': 1.745990138552062e-05, 'epoch': 1.44}) +[2026-06-19 18:34:52,596][starVLA.training.train_starvla][INFO] - Step 3845, Loss: {'train/loss': 0.24108706414699554, 'train/grad_norm_pre_clip': 11.157488822937012, 'timing/data': 0.010305879055522382, 'timing/model': 1.1523517929017544, 'learning_rate/qwen_vl_interface': 1.7439347619351324e-06, 'learning_rate/action_model': 1.7439347619351324e-05, 'epoch': 1.44}) +[2026-06-19 18:34:57,200][starVLA.training.train_starvla][INFO] - Step 3846, Loss: {'train/loss': 0.16688767075538635, 'train/grad_norm_pre_clip': 8.59870719909668, 'timing/data': 0.01056557206902653, 'timing/model': 1.1611238609766588, 'learning_rate/qwen_vl_interface': 1.741880826529186e-06, 'learning_rate/action_model': 1.741880826529186e-05, 'epoch': 1.44}) +[2026-06-19 18:35:01,820][starVLA.training.train_starvla][INFO] - Step 3847, Loss: {'train/loss': 0.19217966496944427, 'train/grad_norm_pre_clip': 12.202342987060547, 'timing/data': 0.00944036606233567, 'timing/model': 1.1470480910502374, 'learning_rate/qwen_vl_interface': 1.7398283331785205e-06, 'learning_rate/action_model': 1.7398283331785203e-05, 'epoch': 1.44}) +[2026-06-19 18:35:06,434][starVLA.training.train_starvla][INFO] - Step 3848, Loss: {'train/loss': 0.1932229995727539, 'train/grad_norm_pre_clip': 9.2440824508667, 'timing/data': 0.009534660959616303, 'timing/model': 1.1502282089786604, 'learning_rate/qwen_vl_interface': 1.737777282726835e-06, 'learning_rate/action_model': 1.737777282726835e-05, 'epoch': 1.44}) +[2026-06-19 18:35:11,018][starVLA.training.train_starvla][INFO] - Step 3849, Loss: {'train/loss': 0.18581673502922058, 'train/grad_norm_pre_clip': 10.246295928955078, 'timing/data': 0.010603090049698949, 'timing/model': 1.3116381450090557, 'learning_rate/qwen_vl_interface': 1.7357276760172403e-06, 'learning_rate/action_model': 1.7357276760172404e-05, 'epoch': 1.44}) +[2026-06-19 18:35:15,433][starVLA.training.train_starvla][INFO] - Step 3850, Loss: {'train/loss': 0.248386412858963, 'train/grad_norm_pre_clip': 13.357647895812988, 'timing/data': 0.010526212048716843, 'timing/model': 1.144049345050007, 'learning_rate/qwen_vl_interface': 1.733679513892252e-06, 'learning_rate/action_model': 1.7336795138922517e-05, 'epoch': 1.44}) +[2026-06-19 18:35:20,020][starVLA.training.train_starvla][INFO] - Step 3851, Loss: {'train/loss': 0.28672340512275696, 'train/grad_norm_pre_clip': 13.709209442138672, 'timing/data': 0.010364921996369958, 'timing/model': 1.146586835035123, 'learning_rate/qwen_vl_interface': 1.7316327971937926e-06, 'learning_rate/action_model': 1.7316327971937926e-05, 'epoch': 1.44}) +[2026-06-19 18:35:24,519][starVLA.training.train_starvla][INFO] - Step 3852, Loss: {'train/loss': 0.1823255717754364, 'train/grad_norm_pre_clip': 7.874217987060547, 'timing/data': 0.009635741007514298, 'timing/model': 1.1454665790079162, 'learning_rate/qwen_vl_interface': 1.7295875267631904e-06, 'learning_rate/action_model': 1.7295875267631904e-05, 'epoch': 1.44}) +[2026-06-19 18:35:29,064][starVLA.training.train_starvla][INFO] - Step 3853, Loss: {'train/loss': 0.1811588555574417, 'train/grad_norm_pre_clip': 13.912382125854492, 'timing/data': 0.00935937196481973, 'timing/model': 1.147641004063189, 'learning_rate/qwen_vl_interface': 1.7275437034411778e-06, 'learning_rate/action_model': 1.7275437034411776e-05, 'epoch': 1.44}) +[2026-06-19 18:35:33,665][starVLA.training.train_starvla][INFO] - Step 3854, Loss: {'train/loss': 0.21005815267562866, 'train/grad_norm_pre_clip': 9.99770736694336, 'timing/data': 0.009891747031360865, 'timing/model': 1.147571393987164, 'learning_rate/qwen_vl_interface': 1.7255013280678941e-06, 'learning_rate/action_model': 1.725501328067894e-05, 'epoch': 1.45}) +[2026-06-19 18:35:38,268][starVLA.training.train_starvla][INFO] - Step 3855, Loss: {'train/loss': 0.3558272421360016, 'train/grad_norm_pre_clip': 14.772323608398438, 'timing/data': 0.010707246023230255, 'timing/model': 1.3155679700430483, 'learning_rate/qwen_vl_interface': 1.7234604014828824e-06, 'learning_rate/action_model': 1.7234604014828823e-05, 'epoch': 1.45}) +[2026-06-19 18:35:42,708][starVLA.training.train_starvla][INFO] - Step 3856, Loss: {'train/loss': 0.06463116407394409, 'train/grad_norm_pre_clip': 3.5757853984832764, 'timing/data': 0.010426271939650178, 'timing/model': 1.1525180549360812, 'learning_rate/qwen_vl_interface': 1.7214209245250906e-06, 'learning_rate/action_model': 1.7214209245250905e-05, 'epoch': 1.45}) +[2026-06-19 18:35:47,317][starVLA.training.train_starvla][INFO] - Step 3857, Loss: {'train/loss': 0.2894187569618225, 'train/grad_norm_pre_clip': 12.471793174743652, 'timing/data': 0.01037179201375693, 'timing/model': 1.1563073759898543, 'learning_rate/qwen_vl_interface': 1.7193828980328723e-06, 'learning_rate/action_model': 1.719382898032872e-05, 'epoch': 1.45}) +[2026-06-19 18:35:51,901][starVLA.training.train_starvla][INFO] - Step 3858, Loss: {'train/loss': 0.10986156016588211, 'train/grad_norm_pre_clip': 3.9555275440216064, 'timing/data': 0.00948847399558872, 'timing/model': 1.1479177619330585, 'learning_rate/qwen_vl_interface': 1.71734632284398e-06, 'learning_rate/action_model': 1.7173463228439798e-05, 'epoch': 1.45}) +[2026-06-19 18:35:56,433][starVLA.training.train_starvla][INFO] - Step 3859, Loss: {'train/loss': 0.15314118564128876, 'train/grad_norm_pre_clip': 19.593502044677734, 'timing/data': 0.01023146603256464, 'timing/model': 1.1473372919717804, 'learning_rate/qwen_vl_interface': 1.7153111997955744e-06, 'learning_rate/action_model': 1.7153111997955743e-05, 'epoch': 1.45}) +[2026-06-19 18:36:01,022][starVLA.training.train_starvla][INFO] - Step 3860, Loss: {'train/loss': 0.15399490296840668, 'train/grad_norm_pre_clip': 13.814763069152832, 'timing/data': 0.010778605937957764, 'timing/model': 1.147150733973831, 'learning_rate/qwen_vl_interface': 1.7132775297242175e-06, 'learning_rate/action_model': 1.7132775297242172e-05, 'epoch': 1.45}) +[2026-06-19 18:36:05,621][starVLA.training.train_starvla][INFO] - Step 3861, Loss: {'train/loss': 0.25568005442619324, 'train/grad_norm_pre_clip': 16.34735679626465, 'timing/data': 0.010586977936327457, 'timing/model': 1.3136647139908746, 'learning_rate/qwen_vl_interface': 1.7112453134658747e-06, 'learning_rate/action_model': 1.7112453134658747e-05, 'epoch': 1.45}) +[2026-06-19 18:36:10,036][starVLA.training.train_starvla][INFO] - Step 3862, Loss: {'train/loss': 0.1658633053302765, 'train/grad_norm_pre_clip': 9.494935035705566, 'timing/data': 0.009445503004826605, 'timing/model': 1.146426509018056, 'learning_rate/qwen_vl_interface': 1.7092145518559129e-06, 'learning_rate/action_model': 1.709214551855913e-05, 'epoch': 1.45}) +[2026-06-19 18:36:14,669][starVLA.training.train_starvla][INFO] - Step 3863, Loss: {'train/loss': 0.1347178965806961, 'train/grad_norm_pre_clip': 8.597888946533203, 'timing/data': 0.01007341593503952, 'timing/model': 1.163582039065659, 'learning_rate/qwen_vl_interface': 1.7071852457291013e-06, 'learning_rate/action_model': 1.707185245729101e-05, 'epoch': 1.45}) +[2026-06-19 18:36:19,276][starVLA.training.train_starvla][INFO] - Step 3864, Loss: {'train/loss': 0.44545191526412964, 'train/grad_norm_pre_clip': 16.883392333984375, 'timing/data': 0.010089767980389297, 'timing/model': 1.1523225819692016, 'learning_rate/qwen_vl_interface': 1.7051573959196122e-06, 'learning_rate/action_model': 1.7051573959196122e-05, 'epoch': 1.45}) +[2026-06-19 18:36:23,890][starVLA.training.train_starvla][INFO] - Step 3865, Loss: {'train/loss': 0.20010589063167572, 'train/grad_norm_pre_clip': 7.511574745178223, 'timing/data': 0.009692255058325827, 'timing/model': 1.1514055631123483, 'learning_rate/qwen_vl_interface': 1.7031310032610166e-06, 'learning_rate/action_model': 1.7031310032610163e-05, 'epoch': 1.45}) +[2026-06-19 18:36:28,513][starVLA.training.train_starvla][INFO] - Step 3866, Loss: {'train/loss': 0.4382389783859253, 'train/grad_norm_pre_clip': 17.045591354370117, 'timing/data': 0.010044394060969353, 'timing/model': 1.1582006680546328, 'learning_rate/qwen_vl_interface': 1.701106068586289e-06, 'learning_rate/action_model': 1.701106068586289e-05, 'epoch': 1.45}) +[2026-06-19 18:36:33,131][starVLA.training.train_starvla][INFO] - Step 3867, Loss: {'train/loss': 0.15506961941719055, 'train/grad_norm_pre_clip': 12.156107902526855, 'timing/data': 0.010658850078471005, 'timing/model': 1.3304710669908673, 'learning_rate/qwen_vl_interface': 1.6990825927278032e-06, 'learning_rate/action_model': 1.699082592727803e-05, 'epoch': 1.45}) +[2026-06-19 18:36:37,586][starVLA.training.train_starvla][INFO] - Step 3868, Loss: {'train/loss': 0.3068109452724457, 'train/grad_norm_pre_clip': 13.594372749328613, 'timing/data': 0.010616699932143092, 'timing/model': 1.1565024759620428, 'learning_rate/qwen_vl_interface': 1.6970605765173344e-06, 'learning_rate/action_model': 1.6970605765173346e-05, 'epoch': 1.45}) +[2026-06-19 18:36:42,205][starVLA.training.train_starvla][INFO] - Step 3869, Loss: {'train/loss': 0.41210559010505676, 'train/grad_norm_pre_clip': 18.709383010864258, 'timing/data': 0.009329449967481196, 'timing/model': 1.1544524470809847, 'learning_rate/qwen_vl_interface': 1.6950400207860557e-06, 'learning_rate/action_model': 1.6950400207860557e-05, 'epoch': 1.45}) +[2026-06-19 18:36:46,726][starVLA.training.train_starvla][INFO] - Step 3870, Loss: {'train/loss': 0.23261471092700958, 'train/grad_norm_pre_clip': 13.701518058776855, 'timing/data': 0.010306240990757942, 'timing/model': 1.1592023830162361, 'learning_rate/qwen_vl_interface': 1.6930209263645412e-06, 'learning_rate/action_model': 1.6930209263645412e-05, 'epoch': 1.45}) +[2026-06-19 18:36:51,285][starVLA.training.train_starvla][INFO] - Step 3871, Loss: {'train/loss': 0.09650559723377228, 'train/grad_norm_pre_clip': 7.459986209869385, 'timing/data': 0.009721213020384312, 'timing/model': 1.1528440909460187, 'learning_rate/qwen_vl_interface': 1.6910032940827656e-06, 'learning_rate/action_model': 1.6910032940827657e-05, 'epoch': 1.45}) +[2026-06-19 18:36:55,913][starVLA.training.train_starvla][INFO] - Step 3872, Loss: {'train/loss': 0.09584227204322815, 'train/grad_norm_pre_clip': 6.394559383392334, 'timing/data': 0.009996706969104707, 'timing/model': 1.1515614830423146, 'learning_rate/qwen_vl_interface': 1.6889871247701003e-06, 'learning_rate/action_model': 1.6889871247701002e-05, 'epoch': 1.45}) +[2026-06-19 18:37:00,540][starVLA.training.train_starvla][INFO] - Step 3873, Loss: {'train/loss': 0.16577740013599396, 'train/grad_norm_pre_clip': 11.328782081604004, 'timing/data': 0.010021236026659608, 'timing/model': 1.3199389399960637, 'learning_rate/qwen_vl_interface': 1.6869724192553168e-06, 'learning_rate/action_model': 1.686972419255317e-05, 'epoch': 1.45}) +[2026-06-19 18:37:04,996][starVLA.training.train_starvla][INFO] - Step 3874, Loss: {'train/loss': 0.12723472714424133, 'train/grad_norm_pre_clip': 11.86522102355957, 'timing/data': 0.010672796051949263, 'timing/model': 1.1532136660534889, 'learning_rate/qwen_vl_interface': 1.6849591783665847e-06, 'learning_rate/action_model': 1.6849591783665847e-05, 'epoch': 1.45}) +[2026-06-19 18:37:09,610][starVLA.training.train_starvla][INFO] - Step 3875, Loss: {'train/loss': 0.10268384963274002, 'train/grad_norm_pre_clip': 10.136703491210938, 'timing/data': 0.010330880992114544, 'timing/model': 1.1478508859872818, 'learning_rate/qwen_vl_interface': 1.6829474029314713e-06, 'learning_rate/action_model': 1.6829474029314713e-05, 'epoch': 1.45}) +[2026-06-19 18:37:14,227][starVLA.training.train_starvla][INFO] - Step 3876, Loss: {'train/loss': 0.25320425629615784, 'train/grad_norm_pre_clip': 38.87125778198242, 'timing/data': 0.010800482938066125, 'timing/model': 1.1642590899718925, 'learning_rate/qwen_vl_interface': 1.6809370937769408e-06, 'learning_rate/action_model': 1.680937093776941e-05, 'epoch': 1.45}) +[2026-06-19 18:37:18,863][starVLA.training.train_starvla][INFO] - Step 3877, Loss: {'train/loss': 0.11567684262990952, 'train/grad_norm_pre_clip': 9.393158912658691, 'timing/data': 0.009750419994816184, 'timing/model': 1.1612389149377123, 'learning_rate/qwen_vl_interface': 1.678928251729356e-06, 'learning_rate/action_model': 1.6789282517293557e-05, 'epoch': 1.45}) +[2026-06-19 18:37:23,504][starVLA.training.train_starvla][INFO] - Step 3878, Loss: {'train/loss': 0.11307738721370697, 'train/grad_norm_pre_clip': 10.671616554260254, 'timing/data': 0.010810525971464813, 'timing/model': 1.1532313820207492, 'learning_rate/qwen_vl_interface': 1.6769208776144767e-06, 'learning_rate/action_model': 1.6769208776144764e-05, 'epoch': 1.45}) +[2026-06-19 18:37:28,141][starVLA.training.train_starvla][INFO] - Step 3879, Loss: {'train/loss': 0.32891565561294556, 'train/grad_norm_pre_clip': 15.528650283813477, 'timing/data': 0.010933664976619184, 'timing/model': 1.3400159180164337, 'learning_rate/qwen_vl_interface': 1.6749149722574564e-06, 'learning_rate/action_model': 1.6749149722574564e-05, 'epoch': 1.45}) +[2026-06-19 18:37:32,766][starVLA.training.train_starvla][INFO] - Step 3880, Loss: {'train/loss': 0.3872697651386261, 'train/grad_norm_pre_clip': 18.705339431762695, 'timing/data': 0.010394971002824605, 'timing/model': 1.3204131349921227, 'learning_rate/qwen_vl_interface': 1.6729105364828486e-06, 'learning_rate/action_model': 1.6729105364828484e-05, 'epoch': 1.45}) +[2026-06-19 18:37:37,201][starVLA.training.train_starvla][INFO] - Step 3881, Loss: {'train/loss': 0.2614324390888214, 'train/grad_norm_pre_clip': 17.71175765991211, 'timing/data': 0.010363854933530092, 'timing/model': 1.1477134419837967, 'learning_rate/qwen_vl_interface': 1.6709075711146e-06, 'learning_rate/action_model': 1.6709075711146e-05, 'epoch': 1.46}) +[2026-06-19 18:37:41,811][starVLA.training.train_starvla][INFO] - Step 3882, Loss: {'train/loss': 0.16693152487277985, 'train/grad_norm_pre_clip': 5.596909523010254, 'timing/data': 0.010288694989867508, 'timing/model': 1.1580646579386666, 'learning_rate/qwen_vl_interface': 1.6689060769760556e-06, 'learning_rate/action_model': 1.6689060769760557e-05, 'epoch': 1.46}) +[2026-06-19 18:37:46,421][starVLA.training.train_starvla][INFO] - Step 3883, Loss: {'train/loss': 0.06738071143627167, 'train/grad_norm_pre_clip': 11.916889190673828, 'timing/data': 0.009878352982923388, 'timing/model': 1.1485385409323499, 'learning_rate/qwen_vl_interface': 1.6669060548899528e-06, 'learning_rate/action_model': 1.6669060548899525e-05, 'epoch': 1.46}) +[2026-06-19 18:37:50,953][starVLA.training.train_starvla][INFO] - Step 3884, Loss: {'train/loss': 0.3050265908241272, 'train/grad_norm_pre_clip': 15.188512802124023, 'timing/data': 0.010661887004971504, 'timing/model': 1.1463054000632837, 'learning_rate/qwen_vl_interface': 1.6649075056784253e-06, 'learning_rate/action_model': 1.6649075056784253e-05, 'epoch': 1.46}) +[2026-06-19 18:37:55,513][starVLA.training.train_starvla][INFO] - Step 3885, Loss: {'train/loss': 0.19394691288471222, 'train/grad_norm_pre_clip': 14.488642692565918, 'timing/data': 0.009892911068163812, 'timing/model': 1.1456913399742916, 'learning_rate/qwen_vl_interface': 1.6629104301630019e-06, 'learning_rate/action_model': 1.662910430163002e-05, 'epoch': 1.46}) +[2026-06-19 18:38:00,135][starVLA.training.train_starvla][INFO] - Step 3886, Loss: {'train/loss': 0.33624589443206787, 'train/grad_norm_pre_clip': 28.974641799926758, 'timing/data': 0.010110523086041212, 'timing/model': 1.3341600330313668, 'learning_rate/qwen_vl_interface': 1.6609148291646043e-06, 'learning_rate/action_model': 1.6609148291646042e-05, 'epoch': 1.46}) +[2026-06-19 18:38:04,583][starVLA.training.train_starvla][INFO] - Step 3887, Loss: {'train/loss': 0.19828996062278748, 'train/grad_norm_pre_clip': 10.00855827331543, 'timing/data': 0.01042387099005282, 'timing/model': 1.1483797170221806, 'learning_rate/qwen_vl_interface': 1.658920703503549e-06, 'learning_rate/action_model': 1.6589207035035488e-05, 'epoch': 1.46}) +[2026-06-19 18:38:09,196][starVLA.training.train_starvla][INFO] - Step 3888, Loss: {'train/loss': 0.15925830602645874, 'train/grad_norm_pre_clip': 19.328628540039062, 'timing/data': 0.01045375899411738, 'timing/model': 1.1500344790983945, 'learning_rate/qwen_vl_interface': 1.6569280539995467e-06, 'learning_rate/action_model': 1.6569280539995467e-05, 'epoch': 1.46}) +[2026-06-19 18:38:13,744][starVLA.training.train_starvla][INFO] - Step 3889, Loss: {'train/loss': 0.44558513164520264, 'train/grad_norm_pre_clip': 21.794734954833984, 'timing/data': 0.010328813921660185, 'timing/model': 1.1454702930059284, 'learning_rate/qwen_vl_interface': 1.6549368814716977e-06, 'learning_rate/action_model': 1.6549368814716977e-05, 'epoch': 1.46}) +[2026-06-19 18:38:18,353][starVLA.training.train_starvla][INFO] - Step 3890, Loss: {'train/loss': 0.3660735785961151, 'train/grad_norm_pre_clip': 20.95484733581543, 'timing/data': 0.010338688036426902, 'timing/model': 1.1477019789163023, 'learning_rate/qwen_vl_interface': 1.6529471867385009e-06, 'learning_rate/action_model': 1.6529471867385007e-05, 'epoch': 1.46}) +[2026-06-19 18:38:22,976][starVLA.training.train_starvla][INFO] - Step 3891, Loss: {'train/loss': 0.19263552129268646, 'train/grad_norm_pre_clip': 12.073022842407227, 'timing/data': 0.010418905061669648, 'timing/model': 1.1593638360500336, 'learning_rate/qwen_vl_interface': 1.6509589706178417e-06, 'learning_rate/action_model': 1.6509589706178417e-05, 'epoch': 1.46}) +[2026-06-19 18:38:27,583][starVLA.training.train_starvla][INFO] - Step 3892, Loss: {'train/loss': 0.31647536158561707, 'train/grad_norm_pre_clip': 14.772782325744629, 'timing/data': 0.010543941985815763, 'timing/model': 1.32250476104673, 'learning_rate/qwen_vl_interface': 1.648972233927002e-06, 'learning_rate/action_model': 1.648972233927002e-05, 'epoch': 1.46}) +[2026-06-19 18:38:32,090][starVLA.training.train_starvla][INFO] - Step 3893, Loss: {'train/loss': 0.29256710410118103, 'train/grad_norm_pre_clip': 18.25123405456543, 'timing/data': 0.01014680799562484, 'timing/model': 1.2161682130536065, 'learning_rate/qwen_vl_interface': 1.646986977482654e-06, 'learning_rate/action_model': 1.6469869774826538e-05, 'epoch': 1.46}) +[2026-06-19 18:38:36,536][starVLA.training.train_starvla][INFO] - Step 3894, Loss: {'train/loss': 0.2765527367591858, 'train/grad_norm_pre_clip': 18.747751235961914, 'timing/data': 0.01055535301566124, 'timing/model': 1.1615672759944573, 'learning_rate/qwen_vl_interface': 1.6450032021008618e-06, 'learning_rate/action_model': 1.6450032021008617e-05, 'epoch': 1.46}) +[2026-06-19 18:38:41,156][starVLA.training.train_starvla][INFO] - Step 3895, Loss: {'train/loss': 0.14597555994987488, 'train/grad_norm_pre_clip': 14.148609161376953, 'timing/data': 0.012370098032988608, 'timing/model': 1.1583771569421515, 'learning_rate/qwen_vl_interface': 1.6430209085970812e-06, 'learning_rate/action_model': 1.643020908597081e-05, 'epoch': 1.46}) +[2026-06-19 18:38:45,824][starVLA.training.train_starvla][INFO] - Step 3896, Loss: {'train/loss': 0.14699317514896393, 'train/grad_norm_pre_clip': 14.37749195098877, 'timing/data': 0.009994918946176767, 'timing/model': 1.1525734539609402, 'learning_rate/qwen_vl_interface': 1.641040097786157e-06, 'learning_rate/action_model': 1.641040097786157e-05, 'epoch': 1.46}) +[2026-06-19 18:38:50,442][starVLA.training.train_starvla][INFO] - Step 3897, Loss: {'train/loss': 0.6175432205200195, 'train/grad_norm_pre_clip': 18.157636642456055, 'timing/data': 0.009647216065786779, 'timing/model': 1.1525021329289302, 'learning_rate/qwen_vl_interface': 1.6390607704823267e-06, 'learning_rate/action_model': 1.6390607704823265e-05, 'epoch': 1.46}) +[2026-06-19 18:38:55,042][starVLA.training.train_starvla][INFO] - Step 3898, Loss: {'train/loss': 0.2849879264831543, 'train/grad_norm_pre_clip': 26.103994369506836, 'timing/data': 0.009532434050925076, 'timing/model': 1.314033894916065, 'learning_rate/qwen_vl_interface': 1.6370829274992162e-06, 'learning_rate/action_model': 1.6370829274992162e-05, 'epoch': 1.46}) +[2026-06-19 18:38:59,570][starVLA.training.train_starvla][INFO] - Step 3899, Loss: {'train/loss': 0.17200106382369995, 'train/grad_norm_pre_clip': 13.64841079711914, 'timing/data': 0.010231876047328115, 'timing/model': 1.2171513610519469, 'learning_rate/qwen_vl_interface': 1.6351065696498422e-06, 'learning_rate/action_model': 1.6351065696498422e-05, 'epoch': 1.46}) +[2026-06-19 18:39:04,022][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 3900 +[2026-06-19 18:41:58,568][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 3900 +[2026-06-19 18:41:58,568][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 3900 +[2026-06-19 18:45:05,406][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 3900 +[2026-06-19 18:45:05,407][starVLA.training.train_starvla][INFO] - Step 3900, Loss: {'train/loss': 0.16306060552597046, 'train/grad_norm_pre_clip': 9.714884757995605, 'eval/loss': 0.13309217989444733, 'eval/latency_0/loss': 0.0838966816663742, 'eval/latency_1/loss': 0.04855326563119888, 'eval/latency_2/loss': 0.11000360548496246, 'eval/latency_3/loss': 0.14831608533859253, 'eval/latency_4/loss': 0.2749857008457184, 'eval/flappy/loss': 0.13309217989444733, 'eval/flappy/latency_0/loss': 0.0838966816663742, 'eval/flappy/latency_1/loss': 0.04855326563119888, 'eval/flappy/latency_2/loss': 0.11000360548496246, 'eval/flappy/latency_3/loss': 0.14831608533859253, 'eval/flappy/latency_4/loss': 0.2749857008457184, 'eval/flap_precision': 0.7279411764705882, 'eval/flap_recall': 0.72, 'eval/flap_f1': 0.7239488117001828, 'eval/flap_rate_model': 0.07555555555555556, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.7239488117001828, 'eval/latency_0/flap_precision': 0.7279411764705882, 'eval/latency_0/flap_recall': 0.72, 'eval/latency_0/flap_f1': 0.7239488117001828, 'eval/latency_0/flap_rate_model': 0.07555555555555556, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.7239488117001828, 'timing/data': 0.010105527006089687, 'timing/model': 1.146428743028082, 'learning_rate/qwen_vl_interface': 1.6331316977466117e-06, 'learning_rate/action_model': 1.6331316977466115e-05, 'epoch': 1.46}) +[2026-06-19 18:45:05,407][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_3900_state +[2026-06-19 18:45:05,410][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-19 18:45:08,700][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_3900_state/model.safetensors +[2026-06-19 18:45:15,814][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_3900_state/optimizer.bin +[2026-06-19 18:45:15,814][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_3900_state/sampler.bin +[2026-06-19 18:45:15,814][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_3900_state/sampler_1.bin +[2026-06-19 18:45:15,815][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_3900_state/random_states_0.pkl +[2026-06-19 18:45:15,836][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3900 +[2026-06-19 18:45:15,836][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-19 18:45:15,845][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-19 18:45:20,318][starVLA.training.train_starvla][INFO] - Step 3901, Loss: {'train/loss': 0.23470069468021393, 'train/grad_norm_pre_clip': 14.665162086486816, 'timing/data': 0.010196737945079803, 'timing/model': 1.1249818729702383, 'learning_rate/qwen_vl_interface': 1.6311583126013198e-06, 'learning_rate/action_model': 1.6311583126013198e-05, 'epoch': 1.46}) +[2026-06-19 18:45:24,829][starVLA.training.train_starvla][INFO] - Step 3902, Loss: {'train/loss': 0.23682130873203278, 'train/grad_norm_pre_clip': 17.64344024658203, 'timing/data': 0.010139518999494612, 'timing/model': 1.2869873639428988, 'learning_rate/qwen_vl_interface': 1.6291864150251488e-06, 'learning_rate/action_model': 1.629186415025149e-05, 'epoch': 1.46}) +[2026-06-19 18:45:29,206][starVLA.training.train_starvla][INFO] - Step 3903, Loss: {'train/loss': 0.11747437715530396, 'train/grad_norm_pre_clip': 10.31542682647705, 'timing/data': 0.010672540054656565, 'timing/model': 1.1345730000175536, 'learning_rate/qwen_vl_interface': 1.6272160058286727e-06, 'learning_rate/action_model': 1.6272160058286726e-05, 'epoch': 1.46}) +[2026-06-19 18:45:33,772][starVLA.training.train_starvla][INFO] - Step 3904, Loss: {'train/loss': 0.2959751784801483, 'train/grad_norm_pre_clip': 21.20661735534668, 'timing/data': 0.010676180943846703, 'timing/model': 1.1416473930003121, 'learning_rate/qwen_vl_interface': 1.6252470858218519e-06, 'learning_rate/action_model': 1.6252470858218517e-05, 'epoch': 1.46}) +[2026-06-19 18:45:38,290][starVLA.training.train_starvla][INFO] - Step 3905, Loss: {'train/loss': 0.33925801515579224, 'train/grad_norm_pre_clip': 21.538114547729492, 'timing/data': 0.01470111298840493, 'timing/model': 1.151879062061198, 'learning_rate/qwen_vl_interface': 1.6232796558140357e-06, 'learning_rate/action_model': 1.6232796558140357e-05, 'epoch': 1.46}) +[2026-06-19 18:45:42,826][starVLA.training.train_starvla][INFO] - Step 3906, Loss: {'train/loss': 0.30493131279945374, 'train/grad_norm_pre_clip': 20.718307495117188, 'timing/data': 0.010331348050385714, 'timing/model': 1.1327189089497551, 'learning_rate/qwen_vl_interface': 1.6213137166139592e-06, 'learning_rate/action_model': 1.621313716613959e-05, 'epoch': 1.46}) +[2026-06-19 18:45:47,382][starVLA.training.train_starvla][INFO] - Step 3907, Loss: {'train/loss': 0.24750159680843353, 'train/grad_norm_pre_clip': 14.08051586151123, 'timing/data': 0.010151447029784322, 'timing/model': 1.1408958680694923, 'learning_rate/qwen_vl_interface': 1.6193492690297463e-06, 'learning_rate/action_model': 1.6193492690297464e-05, 'epoch': 1.46}) +[2026-06-19 18:45:51,934][starVLA.training.train_starvla][INFO] - Step 3908, Loss: {'train/loss': 0.3046782910823822, 'train/grad_norm_pre_clip': 14.578306198120117, 'timing/data': 0.01088097388856113, 'timing/model': 1.3027018340071663, 'learning_rate/qwen_vl_interface': 1.6173863138689078e-06, 'learning_rate/action_model': 1.617386313868908e-05, 'epoch': 1.47}) +[2026-06-19 18:45:56,387][starVLA.training.train_starvla][INFO] - Step 3909, Loss: {'train/loss': 0.27231061458587646, 'train/grad_norm_pre_clip': 13.341917991638184, 'timing/data': 0.010391269926913083, 'timing/model': 1.2063261260045692, 'learning_rate/qwen_vl_interface': 1.6154248519383396e-06, 'learning_rate/action_model': 1.6154248519383395e-05, 'epoch': 1.47}) +[2026-06-19 18:46:00,769][starVLA.training.train_starvla][INFO] - Step 3910, Loss: {'train/loss': 0.28902313113212585, 'train/grad_norm_pre_clip': 14.747159004211426, 'timing/data': 0.01037407002877444, 'timing/model': 1.13326352101285, 'learning_rate/qwen_vl_interface': 1.6134648840443248e-06, 'learning_rate/action_model': 1.613464884044325e-05, 'epoch': 1.47}) +[2026-06-19 18:46:05,268][starVLA.training.train_starvla][INFO] - Step 3911, Loss: {'train/loss': 0.2669537663459778, 'train/grad_norm_pre_clip': 17.810096740722656, 'timing/data': 0.010599471046589315, 'timing/model': 1.1372655190061778, 'learning_rate/qwen_vl_interface': 1.6115064109925334e-06, 'learning_rate/action_model': 1.6115064109925336e-05, 'epoch': 1.47}) +[2026-06-19 18:46:09,845][starVLA.training.train_starvla][INFO] - Step 3912, Loss: {'train/loss': 0.14459362626075745, 'train/grad_norm_pre_clip': 11.165525436401367, 'timing/data': 0.010641540982760489, 'timing/model': 1.1405904779676348, 'learning_rate/qwen_vl_interface': 1.6095494335880173e-06, 'learning_rate/action_model': 1.6095494335880174e-05, 'epoch': 1.47}) +[2026-06-19 18:46:14,460][starVLA.training.train_starvla][INFO] - Step 3913, Loss: {'train/loss': 0.2958727478981018, 'train/grad_norm_pre_clip': 19.0942440032959, 'timing/data': 0.010351026081480086, 'timing/model': 1.149121611029841, 'learning_rate/qwen_vl_interface': 1.6075939526352175e-06, 'learning_rate/action_model': 1.6075939526352175e-05, 'epoch': 1.47}) +[2026-06-19 18:46:19,040][starVLA.training.train_starvla][INFO] - Step 3914, Loss: {'train/loss': 0.30119436979293823, 'train/grad_norm_pre_clip': 15.086942672729492, 'timing/data': 0.010712514980696142, 'timing/model': 1.3196585340192541, 'learning_rate/qwen_vl_interface': 1.6056399689379576e-06, 'learning_rate/action_model': 1.6056399689379575e-05, 'epoch': 1.47}) +[2026-06-19 18:46:23,481][starVLA.training.train_starvla][INFO] - Step 3915, Loss: {'train/loss': 0.17662487924098969, 'train/grad_norm_pre_clip': 17.631080627441406, 'timing/data': 0.010418441030196846, 'timing/model': 1.1459633610211313, 'learning_rate/qwen_vl_interface': 1.6036874832994473e-06, 'learning_rate/action_model': 1.603687483299447e-05, 'epoch': 1.47}) +[2026-06-19 18:46:28,098][starVLA.training.train_starvla][INFO] - Step 3916, Loss: {'train/loss': 0.30951598286628723, 'train/grad_norm_pre_clip': 11.530274391174316, 'timing/data': 0.011354377027601004, 'timing/model': 1.1478779080789536, 'learning_rate/qwen_vl_interface': 1.6017364965222796e-06, 'learning_rate/action_model': 1.6017364965222793e-05, 'epoch': 1.47}) +[2026-06-19 18:46:32,613][starVLA.training.train_starvla][INFO] - Step 3917, Loss: {'train/loss': 0.29964300990104675, 'train/grad_norm_pre_clip': 18.17545509338379, 'timing/data': 0.010756451985798776, 'timing/model': 1.148143075988628, 'learning_rate/qwen_vl_interface': 1.5997870094084305e-06, 'learning_rate/action_model': 1.5997870094084306e-05, 'epoch': 1.47}) +[2026-06-19 18:46:37,238][starVLA.training.train_starvla][INFO] - Step 3918, Loss: {'train/loss': 0.293914794921875, 'train/grad_norm_pre_clip': 11.649909973144531, 'timing/data': 0.01018951297737658, 'timing/model': 1.1522073319647461, 'learning_rate/qwen_vl_interface': 1.5978390227592616e-06, 'learning_rate/action_model': 1.5978390227592615e-05, 'epoch': 1.47}) +[2026-06-19 18:46:41,944][starVLA.training.train_starvla][INFO] - Step 3919, Loss: {'train/loss': 0.40192529559135437, 'train/grad_norm_pre_clip': 18.113880157470703, 'timing/data': 0.011601594043895602, 'timing/model': 1.151874033967033, 'learning_rate/qwen_vl_interface': 1.5958925373755164e-06, 'learning_rate/action_model': 1.5958925373755164e-05, 'epoch': 1.47}) +[2026-06-19 18:46:46,543][starVLA.training.train_starvla][INFO] - Step 3920, Loss: {'train/loss': 0.08009444922208786, 'train/grad_norm_pre_clip': 5.977238178253174, 'timing/data': 0.010074888006784022, 'timing/model': 1.330436412943527, 'learning_rate/qwen_vl_interface': 1.5939475540573213e-06, 'learning_rate/action_model': 1.5939475540573212e-05, 'epoch': 1.47}) +[2026-06-19 18:46:50,968][starVLA.training.train_starvla][INFO] - Step 3921, Loss: {'train/loss': 0.3643436133861542, 'train/grad_norm_pre_clip': 16.980134963989258, 'timing/data': 0.0104517110157758, 'timing/model': 1.150044912006706, 'learning_rate/qwen_vl_interface': 1.5920040736041867e-06, 'learning_rate/action_model': 1.5920040736041866e-05, 'epoch': 1.47}) +[2026-06-19 18:46:55,564][starVLA.training.train_starvla][INFO] - Step 3922, Loss: {'train/loss': 0.11599158495664597, 'train/grad_norm_pre_clip': 9.297792434692383, 'timing/data': 0.01019479299429804, 'timing/model': 1.1432093910407275, 'learning_rate/qwen_vl_interface': 1.5900620968150015e-06, 'learning_rate/action_model': 1.5900620968150015e-05, 'epoch': 1.47}) +[2026-06-19 18:47:00,114][starVLA.training.train_starvla][INFO] - Step 3923, Loss: {'train/loss': 0.10486482828855515, 'train/grad_norm_pre_clip': 8.322657585144043, 'timing/data': 0.010530952014960349, 'timing/model': 1.145083345938474, 'learning_rate/qwen_vl_interface': 1.5881216244880407e-06, 'learning_rate/action_model': 1.5881216244880407e-05, 'epoch': 1.47}) +[2026-06-19 18:47:04,731][starVLA.training.train_starvla][INFO] - Step 3924, Loss: {'train/loss': 0.26157259941101074, 'train/grad_norm_pre_clip': 8.413496971130371, 'timing/data': 0.010302352020516992, 'timing/model': 1.150272136903368, 'learning_rate/qwen_vl_interface': 1.5861826574209587e-06, 'learning_rate/action_model': 1.5861826574209587e-05, 'epoch': 1.47}) +[2026-06-19 18:47:09,344][starVLA.training.train_starvla][INFO] - Step 3925, Loss: {'train/loss': 0.2768508195877075, 'train/grad_norm_pre_clip': 12.844132423400879, 'timing/data': 0.010863107978366315, 'timing/model': 1.146145106991753, 'learning_rate/qwen_vl_interface': 1.5842451964107916e-06, 'learning_rate/action_model': 1.5842451964107916e-05, 'epoch': 1.47}) +[2026-06-19 18:47:13,954][starVLA.training.train_starvla][INFO] - Step 3926, Loss: {'train/loss': 0.23044082522392273, 'train/grad_norm_pre_clip': 17.74822998046875, 'timing/data': 0.010540694929659367, 'timing/model': 1.3069314510794356, 'learning_rate/qwen_vl_interface': 1.5823092422539567e-06, 'learning_rate/action_model': 1.582309242253957e-05, 'epoch': 1.47}) +[2026-06-19 18:47:18,396][starVLA.training.train_starvla][INFO] - Step 3927, Loss: {'train/loss': 0.13525527715682983, 'train/grad_norm_pre_clip': 12.167691230773926, 'timing/data': 0.010424394044093788, 'timing/model': 1.1540747369872406, 'learning_rate/qwen_vl_interface': 1.5803747957462514e-06, 'learning_rate/action_model': 1.5803747957462514e-05, 'epoch': 1.47}) +[2026-06-19 18:47:22,999][starVLA.training.train_starvla][INFO] - Step 3928, Loss: {'train/loss': 0.1413116604089737, 'train/grad_norm_pre_clip': 7.643471717834473, 'timing/data': 0.010249104932881892, 'timing/model': 1.1438343330519274, 'learning_rate/qwen_vl_interface': 1.5784418576828541e-06, 'learning_rate/action_model': 1.5784418576828543e-05, 'epoch': 1.47}) +[2026-06-19 18:47:27,547][starVLA.training.train_starvla][INFO] - Step 3929, Loss: {'train/loss': 0.2581702768802643, 'train/grad_norm_pre_clip': 17.516206741333008, 'timing/data': 0.010577494977042079, 'timing/model': 1.1569169469876215, 'learning_rate/qwen_vl_interface': 1.576510428858322e-06, 'learning_rate/action_model': 1.5765104288583217e-05, 'epoch': 1.47}) +[2026-06-19 18:47:32,141][starVLA.training.train_starvla][INFO] - Step 3930, Loss: {'train/loss': 0.44215983152389526, 'train/grad_norm_pre_clip': 29.798297882080078, 'timing/data': 0.01040830195415765, 'timing/model': 1.152687623980455, 'learning_rate/qwen_vl_interface': 1.5745805100665922e-06, 'learning_rate/action_model': 1.5745805100665923e-05, 'epoch': 1.47}) +[2026-06-19 18:47:36,735][starVLA.training.train_starvla][INFO] - Step 3931, Loss: {'train/loss': 0.23946736752986908, 'train/grad_norm_pre_clip': 15.4137601852417, 'timing/data': 0.010725036961957812, 'timing/model': 1.1456442369380966, 'learning_rate/qwen_vl_interface': 1.572652102100983e-06, 'learning_rate/action_model': 1.572652102100983e-05, 'epoch': 1.47}) +[2026-06-19 18:47:41,330][starVLA.training.train_starvla][INFO] - Step 3932, Loss: {'train/loss': 0.2088100165128708, 'train/grad_norm_pre_clip': 13.9066801071167, 'timing/data': 0.0107736149802804, 'timing/model': 1.3174887519562617, 'learning_rate/qwen_vl_interface': 1.5707252057541898e-06, 'learning_rate/action_model': 1.5707252057541896e-05, 'epoch': 1.47}) +[2026-06-19 18:47:45,761][starVLA.training.train_starvla][INFO] - Step 3933, Loss: {'train/loss': 0.24870102107524872, 'train/grad_norm_pre_clip': 14.993294715881348, 'timing/data': 0.010136925033293664, 'timing/model': 1.1515892919851467, 'learning_rate/qwen_vl_interface': 1.5687998218182849e-06, 'learning_rate/action_model': 1.5687998218182847e-05, 'epoch': 1.47}) +[2026-06-19 18:47:50,349][starVLA.training.train_starvla][INFO] - Step 3934, Loss: {'train/loss': 0.25948795676231384, 'train/grad_norm_pre_clip': 19.318233489990234, 'timing/data': 0.010412737028673291, 'timing/model': 1.1505297999829054, 'learning_rate/qwen_vl_interface': 1.5668759510847235e-06, 'learning_rate/action_model': 1.5668759510847233e-05, 'epoch': 1.48}) +[2026-06-19 18:47:54,871][starVLA.training.train_starvla][INFO] - Step 3935, Loss: {'train/loss': 0.3285526633262634, 'train/grad_norm_pre_clip': 11.095901489257812, 'timing/data': 0.01043929299339652, 'timing/model': 1.1453582129906863, 'learning_rate/qwen_vl_interface': 1.5649535943443331e-06, 'learning_rate/action_model': 1.564953594344333e-05, 'epoch': 1.48}) +[2026-06-19 18:47:59,463][starVLA.training.train_starvla][INFO] - Step 3936, Loss: {'train/loss': 0.20189128816127777, 'train/grad_norm_pre_clip': 12.014527320861816, 'timing/data': 0.010754797025583684, 'timing/model': 1.1573816509917378, 'learning_rate/qwen_vl_interface': 1.5630327523873244e-06, 'learning_rate/action_model': 1.563032752387324e-05, 'epoch': 1.48}) +[2026-06-19 18:48:04,049][starVLA.training.train_starvla][INFO] - Step 3937, Loss: {'train/loss': 0.20536074042320251, 'train/grad_norm_pre_clip': 11.296671867370605, 'timing/data': 0.01029976096469909, 'timing/model': 1.14666930900421, 'learning_rate/qwen_vl_interface': 1.561113426003282e-06, 'learning_rate/action_model': 1.561113426003282e-05, 'epoch': 1.48}) +[2026-06-19 18:48:08,673][starVLA.training.train_starvla][INFO] - Step 3938, Loss: {'train/loss': 0.22852522134780884, 'train/grad_norm_pre_clip': 11.27180004119873, 'timing/data': 0.010123159037902951, 'timing/model': 1.310845465050079, 'learning_rate/qwen_vl_interface': 1.5591956159811686e-06, 'learning_rate/action_model': 1.5591956159811687e-05, 'epoch': 1.48}) +[2026-06-19 18:48:13,097][starVLA.training.train_starvla][INFO] - Step 3939, Loss: {'train/loss': 0.15644623339176178, 'train/grad_norm_pre_clip': 12.055520057678223, 'timing/data': 0.010262347990646958, 'timing/model': 1.1508886279771104, 'learning_rate/qwen_vl_interface': 1.5572793231093235e-06, 'learning_rate/action_model': 1.5572793231093236e-05, 'epoch': 1.48}) +[2026-06-19 18:48:17,689][starVLA.training.train_starvla][INFO] - Step 3940, Loss: {'train/loss': 0.05689142271876335, 'train/grad_norm_pre_clip': 6.050436496734619, 'timing/data': 0.01076239999383688, 'timing/model': 1.146613757009618, 'learning_rate/qwen_vl_interface': 1.5553645481754626e-06, 'learning_rate/action_model': 1.5553645481754626e-05, 'epoch': 1.48}) +[2026-06-19 18:48:22,210][starVLA.training.train_starvla][INFO] - Step 3941, Loss: {'train/loss': 0.18051473796367645, 'train/grad_norm_pre_clip': 9.045699119567871, 'timing/data': 0.010405975044704974, 'timing/model': 1.1425159479258582, 'learning_rate/qwen_vl_interface': 1.5534512919666775e-06, 'learning_rate/action_model': 1.5534512919666776e-05, 'epoch': 1.48}) +[2026-06-19 18:48:26,825][starVLA.training.train_starvla][INFO] - Step 3942, Loss: {'train/loss': 0.21917495131492615, 'train/grad_norm_pre_clip': 32.76171112060547, 'timing/data': 0.010308402008377016, 'timing/model': 1.1515899659134448, 'learning_rate/qwen_vl_interface': 1.551539555269435e-06, 'learning_rate/action_model': 1.5515395552694348e-05, 'epoch': 1.48}) +[2026-06-19 18:48:31,409][starVLA.training.train_starvla][INFO] - Step 3943, Loss: {'train/loss': 0.18853648006916046, 'train/grad_norm_pre_clip': 23.02134895324707, 'timing/data': 0.010593635030090809, 'timing/model': 1.1444710840005428, 'learning_rate/qwen_vl_interface': 1.5496293388695787e-06, 'learning_rate/action_model': 1.5496293388695786e-05, 'epoch': 1.48}) +[2026-06-19 18:48:36,020][starVLA.training.train_starvla][INFO] - Step 3944, Loss: {'train/loss': 0.1296747922897339, 'train/grad_norm_pre_clip': 8.120936393737793, 'timing/data': 0.010399551945738494, 'timing/model': 1.3248953740112484, 'learning_rate/qwen_vl_interface': 1.547720643552327e-06, 'learning_rate/action_model': 1.547720643552327e-05, 'epoch': 1.48}) +[2026-06-19 18:48:40,462][starVLA.training.train_starvla][INFO] - Step 3945, Loss: {'train/loss': 0.13569921255111694, 'train/grad_norm_pre_clip': 11.406911849975586, 'timing/data': 0.012810977990739048, 'timing/model': 1.154011479113251, 'learning_rate/qwen_vl_interface': 1.5458134701022706e-06, 'learning_rate/action_model': 1.5458134701022705e-05, 'epoch': 1.48}) +[2026-06-19 18:48:45,095][starVLA.training.train_starvla][INFO] - Step 3946, Loss: {'train/loss': 0.48324984312057495, 'train/grad_norm_pre_clip': 33.19324493408203, 'timing/data': 0.010609306977130473, 'timing/model': 1.1546065340517089, 'learning_rate/qwen_vl_interface': 1.5439078193033765e-06, 'learning_rate/action_model': 1.5439078193033766e-05, 'epoch': 1.48}) +[2026-06-19 18:48:49,667][starVLA.training.train_starvla][INFO] - Step 3947, Loss: {'train/loss': 0.19173263013362885, 'train/grad_norm_pre_clip': 13.0546236038208, 'timing/data': 0.00965189398266375, 'timing/model': 1.1532401569420472, 'learning_rate/qwen_vl_interface': 1.5420036919389872e-06, 'learning_rate/action_model': 1.5420036919389873e-05, 'epoch': 1.48}) +[2026-06-19 18:48:54,291][starVLA.training.train_starvla][INFO] - Step 3948, Loss: {'train/loss': 0.04791799187660217, 'train/grad_norm_pre_clip': 4.1087565422058105, 'timing/data': 0.010512945940718055, 'timing/model': 1.1504309209994972, 'learning_rate/qwen_vl_interface': 1.540101088791817e-06, 'learning_rate/action_model': 1.5401010887918168e-05, 'epoch': 1.48}) +[2026-06-19 18:48:58,916][starVLA.training.train_starvla][INFO] - Step 3949, Loss: {'train/loss': 0.24568521976470947, 'train/grad_norm_pre_clip': 13.505324363708496, 'timing/data': 0.010760039906017482, 'timing/model': 1.1567741240141913, 'learning_rate/qwen_vl_interface': 1.5382000106439532e-06, 'learning_rate/action_model': 1.5382000106439532e-05, 'epoch': 1.48}) +[2026-06-19 18:49:03,566][starVLA.training.train_starvla][INFO] - Step 3950, Loss: {'train/loss': 0.30362945795059204, 'train/grad_norm_pre_clip': 15.365355491638184, 'timing/data': 0.010018751956522465, 'timing/model': 1.347766811028123, 'learning_rate/qwen_vl_interface': 1.5363004582768584e-06, 'learning_rate/action_model': 1.5363004582768584e-05, 'epoch': 1.48}) +[2026-06-19 18:49:08,026][starVLA.training.train_starvla][INFO] - Step 3951, Loss: {'train/loss': 0.14359456300735474, 'train/grad_norm_pre_clip': 15.39328384399414, 'timing/data': 0.010503907105885446, 'timing/model': 1.1599271870218217, 'learning_rate/qwen_vl_interface': 1.534402432471367e-06, 'learning_rate/action_model': 1.534402432471367e-05, 'epoch': 1.48}) +[2026-06-19 18:49:12,649][starVLA.training.train_starvla][INFO] - Step 3952, Loss: {'train/loss': 0.2952316403388977, 'train/grad_norm_pre_clip': 19.571813583374023, 'timing/data': 0.010260473005473614, 'timing/model': 1.1538503770716488, 'learning_rate/qwen_vl_interface': 1.532505934007685e-06, 'learning_rate/action_model': 1.532505934007685e-05, 'epoch': 1.48}) +[2026-06-19 18:49:17,233][starVLA.training.train_starvla][INFO] - Step 3953, Loss: {'train/loss': 0.19835232198238373, 'train/grad_norm_pre_clip': 14.360373497009277, 'timing/data': 0.010138461017049849, 'timing/model': 1.1457018919754773, 'learning_rate/qwen_vl_interface': 1.5306109636653915e-06, 'learning_rate/action_model': 1.5306109636653914e-05, 'epoch': 1.48}) +[2026-06-19 18:49:21,732][starVLA.training.train_starvla][INFO] - Step 3954, Loss: {'train/loss': 0.36156487464904785, 'train/grad_norm_pre_clip': 30.22895050048828, 'timing/data': 0.009976353030651808, 'timing/model': 1.1454727969830856, 'learning_rate/qwen_vl_interface': 1.5287175222234383e-06, 'learning_rate/action_model': 1.528717522223438e-05, 'epoch': 1.48}) +[2026-06-19 18:49:26,240][starVLA.training.train_starvla][INFO] - Step 3955, Loss: {'train/loss': 0.13971970975399017, 'train/grad_norm_pre_clip': 6.671319007873535, 'timing/data': 0.010669926996342838, 'timing/model': 1.145459990017116, 'learning_rate/qwen_vl_interface': 1.5268256104601457e-06, 'learning_rate/action_model': 1.5268256104601454e-05, 'epoch': 1.48}) +[2026-06-19 18:49:30,784][starVLA.training.train_starvla][INFO] - Step 3956, Loss: {'train/loss': 0.16773821413516998, 'train/grad_norm_pre_clip': 10.525137901306152, 'timing/data': 0.010824323981069028, 'timing/model': 1.2582933959783986, 'learning_rate/qwen_vl_interface': 1.5249352291532075e-06, 'learning_rate/action_model': 1.5249352291532073e-05, 'epoch': 1.48}) +[2026-06-19 18:49:35,228][starVLA.training.train_starvla][INFO] - Step 3957, Loss: {'train/loss': 0.16998109221458435, 'train/grad_norm_pre_clip': 17.779539108276367, 'timing/data': 0.010144152911379933, 'timing/model': 1.1511957179754972, 'learning_rate/qwen_vl_interface': 1.523046379079689e-06, 'learning_rate/action_model': 1.523046379079689e-05, 'epoch': 1.48}) +[2026-06-19 18:49:39,823][starVLA.training.train_starvla][INFO] - Step 3958, Loss: {'train/loss': 0.17395028471946716, 'train/grad_norm_pre_clip': 18.252933502197266, 'timing/data': 0.01043177000246942, 'timing/model': 1.1455560920294374, 'learning_rate/qwen_vl_interface': 1.5211590610160244e-06, 'learning_rate/action_model': 1.5211590610160242e-05, 'epoch': 1.48}) +[2026-06-19 18:49:44,419][starVLA.training.train_starvla][INFO] - Step 3959, Loss: {'train/loss': 0.2092309445142746, 'train/grad_norm_pre_clip': 12.623245239257812, 'timing/data': 0.010502467979677022, 'timing/model': 1.1476452460046858, 'learning_rate/qwen_vl_interface': 1.5192732757380193e-06, 'learning_rate/action_model': 1.5192732757380194e-05, 'epoch': 1.48}) +[2026-06-19 18:49:49,015][starVLA.training.train_starvla][INFO] - Step 3960, Loss: {'train/loss': 0.23425395786762238, 'train/grad_norm_pre_clip': 11.813861846923828, 'timing/data': 0.010572895989753306, 'timing/model': 1.1445442140102386, 'learning_rate/qwen_vl_interface': 1.517389024020849e-06, 'learning_rate/action_model': 1.5173890240208488e-05, 'epoch': 1.48}) +[2026-06-19 18:49:53,626][starVLA.training.train_starvla][INFO] - Step 3961, Loss: {'train/loss': 0.27427566051483154, 'train/grad_norm_pre_clip': 16.112462997436523, 'timing/data': 0.010432866984046996, 'timing/model': 1.3261004749219865, 'learning_rate/qwen_vl_interface': 1.515506306639057e-06, 'learning_rate/action_model': 1.515506306639057e-05, 'epoch': 1.49}) +[2026-06-19 18:49:58,074][starVLA.training.train_starvla][INFO] - Step 3962, Loss: {'train/loss': 0.1490367203950882, 'train/grad_norm_pre_clip': 12.036653518676758, 'timing/data': 0.01068580208811909, 'timing/model': 1.148171669919975, 'learning_rate/qwen_vl_interface': 1.5136251243665589e-06, 'learning_rate/action_model': 1.5136251243665589e-05, 'epoch': 1.49}) +[2026-06-19 18:50:02,697][starVLA.training.train_starvla][INFO] - Step 3963, Loss: {'train/loss': 0.10520465672016144, 'train/grad_norm_pre_clip': 50.73304748535156, 'timing/data': 0.010006555006839335, 'timing/model': 1.1497740230988711, 'learning_rate/qwen_vl_interface': 1.5117454779766372e-06, 'learning_rate/action_model': 1.5117454779766371e-05, 'epoch': 1.49}) +[2026-06-19 18:50:07,316][starVLA.training.train_starvla][INFO] - Step 3964, Loss: {'train/loss': 0.17254549264907837, 'train/grad_norm_pre_clip': 14.623680114746094, 'timing/data': 0.010499445954337716, 'timing/model': 1.1501008190680295, 'learning_rate/qwen_vl_interface': 1.5098673682419428e-06, 'learning_rate/action_model': 1.5098673682419429e-05, 'epoch': 1.49}) +[2026-06-19 18:50:11,935][starVLA.training.train_starvla][INFO] - Step 3965, Loss: {'train/loss': 0.18420040607452393, 'train/grad_norm_pre_clip': 8.288413047790527, 'timing/data': 0.010692934039980173, 'timing/model': 1.1578100960468873, 'learning_rate/qwen_vl_interface': 1.5079907959344977e-06, 'learning_rate/action_model': 1.5079907959344978e-05, 'epoch': 1.49}) +[2026-06-19 18:50:16,562][starVLA.training.train_starvla][INFO] - Step 3966, Loss: {'train/loss': 0.31397923827171326, 'train/grad_norm_pre_clip': 13.948212623596191, 'timing/data': 0.01089219597633928, 'timing/model': 1.150267952005379, 'learning_rate/qwen_vl_interface': 1.5061157618256873e-06, 'learning_rate/action_model': 1.5061157618256871e-05, 'epoch': 1.49}) +[2026-06-19 18:50:21,174][starVLA.training.train_starvla][INFO] - Step 3967, Loss: {'train/loss': 0.17930148541927338, 'train/grad_norm_pre_clip': 15.611422538757324, 'timing/data': 0.010255645960569382, 'timing/model': 1.3270138439256698, 'learning_rate/qwen_vl_interface': 1.5042422666862675e-06, 'learning_rate/action_model': 1.5042422666862673e-05, 'epoch': 1.49}) +[2026-06-19 18:50:25,735][starVLA.training.train_starvla][INFO] - Step 3968, Loss: {'train/loss': 0.11435554176568985, 'train/grad_norm_pre_clip': 8.922906875610352, 'timing/data': 0.010205786908045411, 'timing/model': 1.2683341320371255, 'learning_rate/qwen_vl_interface': 1.5023703112863628e-06, 'learning_rate/action_model': 1.5023703112863626e-05, 'epoch': 1.49}) +[2026-06-19 18:50:30,192][starVLA.training.train_starvla][INFO] - Step 3969, Loss: {'train/loss': 0.19135209918022156, 'train/grad_norm_pre_clip': 16.755664825439453, 'timing/data': 0.010187670006416738, 'timing/model': 1.1594556519994512, 'learning_rate/qwen_vl_interface': 1.5004998963954629e-06, 'learning_rate/action_model': 1.5004998963954628e-05, 'epoch': 1.49}) +[2026-06-19 18:50:34,810][starVLA.training.train_starvla][INFO] - Step 3970, Loss: {'train/loss': 0.30943048000335693, 'train/grad_norm_pre_clip': 16.090524673461914, 'timing/data': 0.01077324291691184, 'timing/model': 1.144959572935477, 'learning_rate/qwen_vl_interface': 1.498631022782424e-06, 'learning_rate/action_model': 1.4986310227824241e-05, 'epoch': 1.49}) +[2026-06-19 18:50:39,456][starVLA.training.train_starvla][INFO] - Step 3971, Loss: {'train/loss': 0.14209525287151337, 'train/grad_norm_pre_clip': 7.735650062561035, 'timing/data': 0.01046844501979649, 'timing/model': 1.1647994300583377, 'learning_rate/qwen_vl_interface': 1.4967636912154707e-06, 'learning_rate/action_model': 1.4967636912154706e-05, 'epoch': 1.49}) +[2026-06-19 18:50:44,088][starVLA.training.train_starvla][INFO] - Step 3972, Loss: {'train/loss': 0.16794715821743011, 'train/grad_norm_pre_clip': 5.819456577301025, 'timing/data': 0.010282268049195409, 'timing/model': 1.1490586919244379, 'learning_rate/qwen_vl_interface': 1.4948979024621915e-06, 'learning_rate/action_model': 1.4948979024621914e-05, 'epoch': 1.49}) +[2026-06-19 18:50:48,695][starVLA.training.train_starvla][INFO] - Step 3973, Loss: {'train/loss': 0.3016100823879242, 'train/grad_norm_pre_clip': 21.10791015625, 'timing/data': 0.01072072796523571, 'timing/model': 1.1529816660331562, 'learning_rate/qwen_vl_interface': 1.493033657289542e-06, 'learning_rate/action_model': 1.493033657289542e-05, 'epoch': 1.49}) +[2026-06-19 18:50:53,306][starVLA.training.train_starvla][INFO] - Step 3974, Loss: {'train/loss': 0.13898977637290955, 'train/grad_norm_pre_clip': 10.249611854553223, 'timing/data': 0.009934543981216848, 'timing/model': 1.3114028139971197, 'learning_rate/qwen_vl_interface': 1.4911709564638428e-06, 'learning_rate/action_model': 1.4911709564638427e-05, 'epoch': 1.49}) +[2026-06-19 18:50:57,753][starVLA.training.train_starvla][INFO] - Step 3975, Loss: {'train/loss': 0.0829581469297409, 'train/grad_norm_pre_clip': 5.950922012329102, 'timing/data': 0.010449265013448894, 'timing/model': 1.1475364960497245, 'learning_rate/qwen_vl_interface': 1.4893098007507817e-06, 'learning_rate/action_model': 1.4893098007507816e-05, 'epoch': 1.49}) +[2026-06-19 18:51:02,357][starVLA.training.train_starvla][INFO] - Step 3976, Loss: {'train/loss': 0.3085753321647644, 'train/grad_norm_pre_clip': 16.7351016998291, 'timing/data': 0.0103802599478513, 'timing/model': 1.1534110030625015, 'learning_rate/qwen_vl_interface': 1.487450190915406e-06, 'learning_rate/action_model': 1.487450190915406e-05, 'epoch': 1.49}) +[2026-06-19 18:51:06,985][starVLA.training.train_starvla][INFO] - Step 3977, Loss: {'train/loss': 0.09221545606851578, 'train/grad_norm_pre_clip': 10.119081497192383, 'timing/data': 0.010127644054591656, 'timing/model': 1.1512462800601497, 'learning_rate/qwen_vl_interface': 1.4855921277221336e-06, 'learning_rate/action_model': 1.4855921277221336e-05, 'epoch': 1.49}) +[2026-06-19 18:51:11,611][starVLA.training.train_starvla][INFO] - Step 3978, Loss: {'train/loss': 0.16364097595214844, 'train/grad_norm_pre_clip': 12.299385070800781, 'timing/data': 0.010521683027036488, 'timing/model': 1.1439412250183523, 'learning_rate/qwen_vl_interface': 1.4837356119347423e-06, 'learning_rate/action_model': 1.4837356119347423e-05, 'epoch': 1.49}) +[2026-06-19 18:51:16,204][starVLA.training.train_starvla][INFO] - Step 3979, Loss: {'train/loss': 0.13149648904800415, 'train/grad_norm_pre_clip': 11.621197700500488, 'timing/data': 0.010863304021768272, 'timing/model': 1.1462351368973032, 'learning_rate/qwen_vl_interface': 1.4818806443163774e-06, 'learning_rate/action_model': 1.4818806443163775e-05, 'epoch': 1.49}) +[2026-06-19 18:51:20,825][starVLA.training.train_starvla][INFO] - Step 3980, Loss: {'train/loss': 0.21970967948436737, 'train/grad_norm_pre_clip': 16.71501922607422, 'timing/data': 0.010471693938598037, 'timing/model': 1.3197061489336193, 'learning_rate/qwen_vl_interface': 1.480027225629547e-06, 'learning_rate/action_model': 1.480027225629547e-05, 'epoch': 1.49}) +[2026-06-19 18:51:25,258][starVLA.training.train_starvla][INFO] - Step 3981, Loss: {'train/loss': 0.2630661427974701, 'train/grad_norm_pre_clip': 16.318574905395508, 'timing/data': 0.010035817045718431, 'timing/model': 1.1480974140577018, 'learning_rate/qwen_vl_interface': 1.4781753566361182e-06, 'learning_rate/action_model': 1.4781753566361182e-05, 'epoch': 1.49}) +[2026-06-19 18:51:29,860][starVLA.training.train_starvla][INFO] - Step 3982, Loss: {'train/loss': 0.1544157713651657, 'train/grad_norm_pre_clip': 13.13521671295166, 'timing/data': 0.010462741018272936, 'timing/model': 1.1443604569649324, 'learning_rate/qwen_vl_interface': 1.4763250380973255e-06, 'learning_rate/action_model': 1.4763250380973254e-05, 'epoch': 1.49}) +[2026-06-19 18:51:34,363][starVLA.training.train_starvla][INFO] - Step 3983, Loss: {'train/loss': 0.3532658815383911, 'train/grad_norm_pre_clip': 24.807132720947266, 'timing/data': 0.010228631901554763, 'timing/model': 1.1475828000111505, 'learning_rate/qwen_vl_interface': 1.4744762707737662e-06, 'learning_rate/action_model': 1.4744762707737662e-05, 'epoch': 1.49}) +[2026-06-19 18:51:38,905][starVLA.training.train_starvla][INFO] - Step 3984, Loss: {'train/loss': 0.15472394227981567, 'train/grad_norm_pre_clip': 14.384005546569824, 'timing/data': 0.010458796983584762, 'timing/model': 1.1483193399617448, 'learning_rate/qwen_vl_interface': 1.4726290554253977e-06, 'learning_rate/action_model': 1.4726290554253977e-05, 'epoch': 1.49}) +[2026-06-19 18:51:43,505][starVLA.training.train_starvla][INFO] - Step 3985, Loss: {'train/loss': 0.1597263216972351, 'train/grad_norm_pre_clip': 11.987709045410156, 'timing/data': 0.010218963027000427, 'timing/model': 1.1479640699690208, 'learning_rate/qwen_vl_interface': 1.4707833928115406e-06, 'learning_rate/action_model': 1.4707833928115405e-05, 'epoch': 1.49}) +[2026-06-19 18:51:48,094][starVLA.training.train_starvla][INFO] - Step 3986, Loss: {'train/loss': 0.17436355352401733, 'train/grad_norm_pre_clip': 11.207659721374512, 'timing/data': 0.009973656968213618, 'timing/model': 1.3138799499720335, 'learning_rate/qwen_vl_interface': 1.4689392836908775e-06, 'learning_rate/action_model': 1.4689392836908774e-05, 'epoch': 1.49}) +[2026-06-19 18:51:52,551][starVLA.training.train_starvla][INFO] - Step 3987, Loss: {'train/loss': 0.4538503885269165, 'train/grad_norm_pre_clip': 24.162700653076172, 'timing/data': 0.010169262997806072, 'timing/model': 1.1497752299765125, 'learning_rate/qwen_vl_interface': 1.4670967288214514e-06, 'learning_rate/action_model': 1.4670967288214513e-05, 'epoch': 1.49}) +[2026-06-19 18:51:57,152][starVLA.training.train_starvla][INFO] - Step 3988, Loss: {'train/loss': 0.1656070202589035, 'train/grad_norm_pre_clip': 11.505088806152344, 'timing/data': 0.010430003982037306, 'timing/model': 1.1587851960211992, 'learning_rate/qwen_vl_interface': 1.4652557289606687e-06, 'learning_rate/action_model': 1.4652557289606685e-05, 'epoch': 1.5}) +[2026-06-19 18:52:01,773][starVLA.training.train_starvla][INFO] - Step 3989, Loss: {'train/loss': 0.30209577083587646, 'train/grad_norm_pre_clip': 13.163918495178223, 'timing/data': 0.010551669984124601, 'timing/model': 1.1591601429972798, 'learning_rate/qwen_vl_interface': 1.4634162848652918e-06, 'learning_rate/action_model': 1.4634162848652919e-05, 'epoch': 1.5}) +[2026-06-19 18:52:06,382][starVLA.training.train_starvla][INFO] - Step 3990, Loss: {'train/loss': 0.1611626297235489, 'train/grad_norm_pre_clip': 10.671879768371582, 'timing/data': 0.01056699501350522, 'timing/model': 1.1500117719406262, 'learning_rate/qwen_vl_interface': 1.4615783972914494e-06, 'learning_rate/action_model': 1.4615783972914493e-05, 'epoch': 1.5}) +[2026-06-19 18:52:11,017][starVLA.training.train_starvla][INFO] - Step 3991, Loss: {'train/loss': 0.3944549262523651, 'train/grad_norm_pre_clip': 16.447294235229492, 'timing/data': 0.009978789021261036, 'timing/model': 1.1605399190448225, 'learning_rate/qwen_vl_interface': 1.459742066994626e-06, 'learning_rate/action_model': 1.4597420669946261e-05, 'epoch': 1.5}) +[2026-06-19 18:52:15,652][starVLA.training.train_starvla][INFO] - Step 3992, Loss: {'train/loss': 0.10652466118335724, 'train/grad_norm_pre_clip': 8.667769432067871, 'timing/data': 0.009869870031252503, 'timing/model': 1.3270070060389116, 'learning_rate/qwen_vl_interface': 1.4579072947296687e-06, 'learning_rate/action_model': 1.4579072947296687e-05, 'epoch': 1.5}) +[2026-06-19 18:52:20,126][starVLA.training.train_starvla][INFO] - Step 3993, Loss: {'train/loss': 0.30172523856163025, 'train/grad_norm_pre_clip': 22.09567642211914, 'timing/data': 0.00978889805264771, 'timing/model': 1.1538317890372127, 'learning_rate/qwen_vl_interface': 1.4560740812507823e-06, 'learning_rate/action_model': 1.4560740812507822e-05, 'epoch': 1.5}) +[2026-06-19 18:52:24,759][starVLA.training.train_starvla][INFO] - Step 3994, Loss: {'train/loss': 0.1913090944290161, 'train/grad_norm_pre_clip': 12.327129364013672, 'timing/data': 0.01003511599265039, 'timing/model': 1.1584480489837006, 'learning_rate/qwen_vl_interface': 1.4542424273115322e-06, 'learning_rate/action_model': 1.454242427311532e-05, 'epoch': 1.5}) +[2026-06-19 18:52:29,277][starVLA.training.train_starvla][INFO] - Step 3995, Loss: {'train/loss': 0.2643570899963379, 'train/grad_norm_pre_clip': 20.36224365234375, 'timing/data': 0.009771288023330271, 'timing/model': 1.149226707057096, 'learning_rate/qwen_vl_interface': 1.4524123336648419e-06, 'learning_rate/action_model': 1.4524123336648418e-05, 'epoch': 1.5}) +[2026-06-19 18:52:33,853][starVLA.training.train_starvla][INFO] - Step 3996, Loss: {'train/loss': 0.263351708650589, 'train/grad_norm_pre_clip': 16.25257110595703, 'timing/data': 0.010043342947028577, 'timing/model': 1.1488281950587407, 'learning_rate/qwen_vl_interface': 1.4505838010629942e-06, 'learning_rate/action_model': 1.4505838010629941e-05, 'epoch': 1.5}) +[2026-06-19 18:52:38,478][starVLA.training.train_starvla][INFO] - Step 3997, Loss: {'train/loss': 0.37308040261268616, 'train/grad_norm_pre_clip': 62.74760437011719, 'timing/data': 0.009839541045948863, 'timing/model': 1.148948847898282, 'learning_rate/qwen_vl_interface': 1.4487568302576284e-06, 'learning_rate/action_model': 1.4487568302576282e-05, 'epoch': 1.5}) +[2026-06-19 18:52:43,089][starVLA.training.train_starvla][INFO] - Step 3998, Loss: {'train/loss': 0.258118212223053, 'train/grad_norm_pre_clip': 15.558939933776855, 'timing/data': 0.010430780006572604, 'timing/model': 1.3235494151012972, 'learning_rate/qwen_vl_interface': 1.446931421999746e-06, 'learning_rate/action_model': 1.4469314219997459e-05, 'epoch': 1.5}) +[2026-06-19 18:52:47,664][starVLA.training.train_starvla][INFO] - Step 3999, Loss: {'train/loss': 0.13785095512866974, 'train/grad_norm_pre_clip': 12.914201736450195, 'timing/data': 0.0101430700160563, 'timing/model': 1.2703359150327742, 'learning_rate/qwen_vl_interface': 1.4451075770397004e-06, 'learning_rate/action_model': 1.4451075770397004e-05, 'epoch': 1.5}) +[2026-06-19 18:52:52,112][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 4000 +[2026-06-19 18:55:49,211][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 4000 +[2026-06-19 18:55:49,211][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 4000 +[2026-06-19 18:58:56,111][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 4000 +[2026-06-19 18:58:56,112][starVLA.training.train_starvla][INFO] - Step 4000, Loss: {'train/loss': 0.15590634942054749, 'train/grad_norm_pre_clip': 8.516735076904297, 'eval/loss': 0.1332899034023285, 'eval/latency_0/loss': 0.08440232276916504, 'eval/latency_1/loss': 0.04830244556069374, 'eval/latency_2/loss': 0.11847317218780518, 'eval/latency_3/loss': 0.15052668750286102, 'eval/latency_4/loss': 0.2647915184497833, 'eval/flappy/loss': 0.1332899034023285, 'eval/flappy/latency_0/loss': 0.08440232276916504, 'eval/flappy/latency_1/loss': 0.04830244556069374, 'eval/flappy/latency_2/loss': 0.11847317218780518, 'eval/flappy/latency_3/loss': 0.15052668750286102, 'eval/flappy/latency_4/loss': 0.2647915184497833, 'eval/flap_precision': 0.7490636704119851, 'eval/flap_recall': 0.7272727272727273, 'eval/flap_f1': 0.7380073800738007, 'eval/flap_rate_model': 0.07416666666666667, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.7380073800738007, 'eval/latency_0/flap_precision': 0.7490636704119851, 'eval/latency_0/flap_recall': 0.7272727272727273, 'eval/latency_0/flap_f1': 0.7380073800738007, 'eval/latency_0/flap_rate_model': 0.07416666666666667, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.7380073800738007, 'timing/data': 0.010504690930247307, 'timing/model': 1.1473086629994214, 'learning_rate/qwen_vl_interface': 1.4432852961272068e-06, 'learning_rate/action_model': 1.4432852961272066e-05, 'epoch': 1.5}) +[2026-06-19 18:58:56,112][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_4000_state +[2026-06-19 18:58:56,115][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-19 18:58:59,508][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_4000_state/model.safetensors +[2026-06-19 18:59:06,699][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_4000_state/optimizer.bin +[2026-06-19 18:59:06,700][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_4000_state/sampler.bin +[2026-06-19 18:59:06,700][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_4000_state/sampler_1.bin +[2026-06-19 18:59:06,701][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_4000_state/random_states_0.pkl +[2026-06-19 18:59:06,723][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4000 +[2026-06-19 18:59:06,723][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-19 18:59:06,732][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-19 18:59:11,264][starVLA.training.train_starvla][INFO] - Step 4001, Loss: {'train/loss': 0.25082850456237793, 'train/grad_norm_pre_clip': 22.176942825317383, 'timing/data': 0.00997246103361249, 'timing/model': 1.1404668149771169, 'learning_rate/qwen_vl_interface': 1.4414645800113353e-06, 'learning_rate/action_model': 1.4414645800113352e-05, 'epoch': 1.5}) +[2026-06-19 18:59:15,798][starVLA.training.train_starvla][INFO] - Step 4002, Loss: {'train/loss': 0.2640090882778168, 'train/grad_norm_pre_clip': 13.409758567810059, 'timing/data': 0.010023958049714565, 'timing/model': 1.310037576011382, 'learning_rate/qwen_vl_interface': 1.4396454294405138e-06, 'learning_rate/action_model': 1.4396454294405136e-05, 'epoch': 1.5}) +[2026-06-19 18:59:20,342][starVLA.training.train_starvla][INFO] - Step 4003, Loss: {'train/loss': 0.19434721767902374, 'train/grad_norm_pre_clip': 12.191805839538574, 'timing/data': 0.009818313992582262, 'timing/model': 1.2981244151014835, 'learning_rate/qwen_vl_interface': 1.4378278451625271e-06, 'learning_rate/action_model': 1.4378278451625271e-05, 'epoch': 1.5}) +[2026-06-19 18:59:24,708][starVLA.training.train_starvla][INFO] - Step 4004, Loss: {'train/loss': 0.22361275553703308, 'train/grad_norm_pre_clip': 15.315472602844238, 'timing/data': 0.009827028959989548, 'timing/model': 1.1289744570385665, 'learning_rate/qwen_vl_interface': 1.4360118279245143e-06, 'learning_rate/action_model': 1.4360118279245144e-05, 'epoch': 1.5}) +[2026-06-19 18:59:29,260][starVLA.training.train_starvla][INFO] - Step 4005, Loss: {'train/loss': 0.2685416638851166, 'train/grad_norm_pre_clip': 17.77484130859375, 'timing/data': 0.010676264995709062, 'timing/model': 1.1441772490506992, 'learning_rate/qwen_vl_interface': 1.434197378472973e-06, 'learning_rate/action_model': 1.4341973784729729e-05, 'epoch': 1.5}) +[2026-06-19 18:59:33,821][starVLA.training.train_starvla][INFO] - Step 4006, Loss: {'train/loss': 0.20353934168815613, 'train/grad_norm_pre_clip': 15.613740921020508, 'timing/data': 0.009709442965686321, 'timing/model': 1.1381922019645572, 'learning_rate/qwen_vl_interface': 1.4323844975537535e-06, 'learning_rate/action_model': 1.4323844975537534e-05, 'epoch': 1.5}) +[2026-06-19 18:59:38,381][starVLA.training.train_starvla][INFO] - Step 4007, Loss: {'train/loss': 0.3500467836856842, 'train/grad_norm_pre_clip': 18.77701187133789, 'timing/data': 0.010065768961794674, 'timing/model': 1.1413613629993051, 'learning_rate/qwen_vl_interface': 1.4305731859120628e-06, 'learning_rate/action_model': 1.4305731859120628e-05, 'epoch': 1.5}) +[2026-06-19 18:59:42,957][starVLA.training.train_starvla][INFO] - Step 4008, Loss: {'train/loss': 0.43957430124282837, 'train/grad_norm_pre_clip': 23.82933235168457, 'timing/data': 0.01002646202687174, 'timing/model': 1.3191376600880176, 'learning_rate/qwen_vl_interface': 1.4287634442924647e-06, 'learning_rate/action_model': 1.4287634442924646e-05, 'epoch': 1.5}) +[2026-06-19 18:59:47,548][starVLA.training.train_starvla][INFO] - Step 4009, Loss: {'train/loss': 0.2411622405052185, 'train/grad_norm_pre_clip': 11.552257537841797, 'timing/data': 0.009362389100715518, 'timing/model': 1.3178223619470373, 'learning_rate/qwen_vl_interface': 1.4269552734388725e-06, 'learning_rate/action_model': 1.4269552734388724e-05, 'epoch': 1.5}) +[2026-06-19 18:59:51,979][starVLA.training.train_starvla][INFO] - Step 4010, Loss: {'train/loss': 0.327751487493515, 'train/grad_norm_pre_clip': 13.93937873840332, 'timing/data': 0.00990748405456543, 'timing/model': 1.1415096779819578, 'learning_rate/qwen_vl_interface': 1.4251486740945584e-06, 'learning_rate/action_model': 1.4251486740945583e-05, 'epoch': 1.5}) +[2026-06-19 18:59:56,527][starVLA.training.train_starvla][INFO] - Step 4011, Loss: {'train/loss': 0.33963075280189514, 'train/grad_norm_pre_clip': 19.832767486572266, 'timing/data': 0.01012063107918948, 'timing/model': 1.1405558559345081, 'learning_rate/qwen_vl_interface': 1.423343647002147e-06, 'learning_rate/action_model': 1.423343647002147e-05, 'epoch': 1.5}) +[2026-06-19 19:00:01,137][starVLA.training.train_starvla][INFO] - Step 4012, Loss: {'train/loss': 0.050922490656375885, 'train/grad_norm_pre_clip': 6.533125877380371, 'timing/data': 0.009749277960509062, 'timing/model': 1.1617005469743162, 'learning_rate/qwen_vl_interface': 1.421540192903617e-06, 'learning_rate/action_model': 1.421540192903617e-05, 'epoch': 1.5}) +[2026-06-19 19:00:05,744][starVLA.training.train_starvla][INFO] - Step 4013, Loss: {'train/loss': 0.30561089515686035, 'train/grad_norm_pre_clip': 20.77315330505371, 'timing/data': 0.009733816958032548, 'timing/model': 1.1456084840465337, 'learning_rate/qwen_vl_interface': 1.4197383125402997e-06, 'learning_rate/action_model': 1.4197383125402997e-05, 'epoch': 1.5}) +[2026-06-19 19:00:10,330][starVLA.training.train_starvla][INFO] - Step 4014, Loss: {'train/loss': 0.10272882133722305, 'train/grad_norm_pre_clip': 7.110091209411621, 'timing/data': 0.010027430951595306, 'timing/model': 1.2998195890104398, 'learning_rate/qwen_vl_interface': 1.4179380066528807e-06, 'learning_rate/action_model': 1.4179380066528807e-05, 'epoch': 1.51}) +[2026-06-19 19:00:14,906][starVLA.training.train_starvla][INFO] - Step 4015, Loss: {'train/loss': 0.1757308542728424, 'train/grad_norm_pre_clip': 13.925097465515137, 'timing/data': 0.009639955940656364, 'timing/model': 1.3093508679885417, 'learning_rate/qwen_vl_interface': 1.4161392759813984e-06, 'learning_rate/action_model': 1.4161392759813983e-05, 'epoch': 1.51}) +[2026-06-19 19:00:19,319][starVLA.training.train_starvla][INFO] - Step 4016, Loss: {'train/loss': 0.23558154702186584, 'train/grad_norm_pre_clip': 22.692087173461914, 'timing/data': 0.010851996019482613, 'timing/model': 1.1436666109366342, 'learning_rate/qwen_vl_interface': 1.4143421212652411e-06, 'learning_rate/action_model': 1.414342121265241e-05, 'epoch': 1.51}) +[2026-06-19 19:00:23,914][starVLA.training.train_starvla][INFO] - Step 4017, Loss: {'train/loss': 0.09899503737688065, 'train/grad_norm_pre_clip': 9.41564655303955, 'timing/data': 0.010108296992257237, 'timing/model': 1.1453063680091873, 'learning_rate/qwen_vl_interface': 1.4125465432431529e-06, 'learning_rate/action_model': 1.4125465432431529e-05, 'epoch': 1.51}) +[2026-06-19 19:00:28,521][starVLA.training.train_starvla][INFO] - Step 4018, Loss: {'train/loss': 0.39851900935173035, 'train/grad_norm_pre_clip': 19.78352928161621, 'timing/data': 0.009758461965247989, 'timing/model': 1.1536890499992296, 'learning_rate/qwen_vl_interface': 1.4107525426532273e-06, 'learning_rate/action_model': 1.4107525426532272e-05, 'epoch': 1.51}) +[2026-06-19 19:00:33,119][starVLA.training.train_starvla][INFO] - Step 4019, Loss: {'train/loss': 0.11810089647769928, 'train/grad_norm_pre_clip': 12.670957565307617, 'timing/data': 0.009391019004397094, 'timing/model': 1.1411838909843937, 'learning_rate/qwen_vl_interface': 1.4089601202329116e-06, 'learning_rate/action_model': 1.4089601202329114e-05, 'epoch': 1.51}) +[2026-06-19 19:00:37,699][starVLA.training.train_starvla][INFO] - Step 4020, Loss: {'train/loss': 0.29298365116119385, 'train/grad_norm_pre_clip': 14.952919960021973, 'timing/data': 0.009771864977665246, 'timing/model': 1.307519641937688, 'learning_rate/qwen_vl_interface': 1.407169276719e-06, 'learning_rate/action_model': 1.4071692767190001e-05, 'epoch': 1.51}) +[2026-06-19 19:00:42,311][starVLA.training.train_starvla][INFO] - Step 4021, Loss: {'train/loss': 0.15636873245239258, 'train/grad_norm_pre_clip': 11.91553783416748, 'timing/data': 0.009641984943300486, 'timing/model': 1.3070944240316749, 'learning_rate/qwen_vl_interface': 1.4053800128476426e-06, 'learning_rate/action_model': 1.4053800128476427e-05, 'epoch': 1.51}) +[2026-06-19 19:00:46,748][starVLA.training.train_starvla][INFO] - Step 4022, Loss: {'train/loss': 0.10605784505605698, 'train/grad_norm_pre_clip': 14.209214210510254, 'timing/data': 0.009733457933180034, 'timing/model': 1.1479237549938262, 'learning_rate/qwen_vl_interface': 1.403592329354338e-06, 'learning_rate/action_model': 1.403592329354338e-05, 'epoch': 1.51}) +[2026-06-19 19:00:51,377][starVLA.training.train_starvla][INFO] - Step 4023, Loss: {'train/loss': 0.16331569850444794, 'train/grad_norm_pre_clip': 10.561540603637695, 'timing/data': 0.00984429707750678, 'timing/model': 1.1439146150369197, 'learning_rate/qwen_vl_interface': 1.4018062269739354e-06, 'learning_rate/action_model': 1.4018062269739355e-05, 'epoch': 1.51}) +[2026-06-19 19:00:55,978][starVLA.training.train_starvla][INFO] - Step 4024, Loss: {'train/loss': 0.22792868316173553, 'train/grad_norm_pre_clip': 13.277180671691895, 'timing/data': 0.010017239023000002, 'timing/model': 1.1501716719940305, 'learning_rate/qwen_vl_interface': 1.4000217064406339e-06, 'learning_rate/action_model': 1.4000217064406337e-05, 'epoch': 1.51}) +[2026-06-19 19:01:00,554][starVLA.training.train_starvla][INFO] - Step 4025, Loss: {'train/loss': 0.08295790106058121, 'train/grad_norm_pre_clip': 3.220177173614502, 'timing/data': 0.009384438977576792, 'timing/model': 1.1458390440093353, 'learning_rate/qwen_vl_interface': 1.3982387684879836e-06, 'learning_rate/action_model': 1.3982387684879836e-05, 'epoch': 1.51}) +[2026-06-19 19:01:05,165][starVLA.training.train_starvla][INFO] - Step 4026, Loss: {'train/loss': 0.17775392532348633, 'train/grad_norm_pre_clip': 15.711804389953613, 'timing/data': 0.009654370020143688, 'timing/model': 1.1523323269793764, 'learning_rate/qwen_vl_interface': 1.3964574138488804e-06, 'learning_rate/action_model': 1.3964574138488803e-05, 'epoch': 1.51}) +[2026-06-19 19:01:09,750][starVLA.training.train_starvla][INFO] - Step 4027, Loss: {'train/loss': 0.16142460703849792, 'train/grad_norm_pre_clip': 8.603168487548828, 'timing/data': 0.009960448951460421, 'timing/model': 1.3054060739232227, 'learning_rate/qwen_vl_interface': 1.3946776432555743e-06, 'learning_rate/action_model': 1.3946776432555741e-05, 'epoch': 1.51}) +[2026-06-19 19:01:14,180][starVLA.training.train_starvla][INFO] - Step 4028, Loss: {'train/loss': 0.11631648987531662, 'train/grad_norm_pre_clip': 27.484115600585938, 'timing/data': 0.00958046200685203, 'timing/model': 1.1431250830646604, 'learning_rate/qwen_vl_interface': 1.3928994574396595e-06, 'learning_rate/action_model': 1.3928994574396594e-05, 'epoch': 1.51}) +[2026-06-19 19:01:18,760][starVLA.training.train_starvla][INFO] - Step 4029, Loss: {'train/loss': 0.16940923035144806, 'train/grad_norm_pre_clip': 12.126799583435059, 'timing/data': 0.010005741962231696, 'timing/model': 1.1438503579702228, 'learning_rate/qwen_vl_interface': 1.391122857132083e-06, 'learning_rate/action_model': 1.3911228571320831e-05, 'epoch': 1.51}) +[2026-06-19 19:01:23,378][starVLA.training.train_starvla][INFO] - Step 4030, Loss: {'train/loss': 0.2971990406513214, 'train/grad_norm_pre_clip': 19.226459503173828, 'timing/data': 0.009885253035463393, 'timing/model': 1.1449969710083678, 'learning_rate/qwen_vl_interface': 1.3893478430631376e-06, 'learning_rate/action_model': 1.3893478430631376e-05, 'epoch': 1.51}) +[2026-06-19 19:01:28,010][starVLA.training.train_starvla][INFO] - Step 4031, Loss: {'train/loss': 0.15507420897483826, 'train/grad_norm_pre_clip': 11.143536567687988, 'timing/data': 0.009474821970798075, 'timing/model': 1.1533613649662584, 'learning_rate/qwen_vl_interface': 1.3875744159624652e-06, 'learning_rate/action_model': 1.3875744159624652e-05, 'epoch': 1.51}) +[2026-06-19 19:01:32,602][starVLA.training.train_starvla][INFO] - Step 4032, Loss: {'train/loss': 0.2935524582862854, 'train/grad_norm_pre_clip': 17.686153411865234, 'timing/data': 0.009860267047770321, 'timing/model': 1.1449584469664842, 'learning_rate/qwen_vl_interface': 1.3858025765590525e-06, 'learning_rate/action_model': 1.3858025765590525e-05, 'epoch': 1.51}) +[2026-06-19 19:01:37,226][starVLA.training.train_starvla][INFO] - Step 4033, Loss: {'train/loss': 0.2118281126022339, 'train/grad_norm_pre_clip': 21.31253433227539, 'timing/data': 0.01019342290237546, 'timing/model': 1.322905962006189, 'learning_rate/qwen_vl_interface': 1.384032325581237e-06, 'learning_rate/action_model': 1.3840323255812371e-05, 'epoch': 1.51}) +[2026-06-19 19:01:41,759][starVLA.training.train_starvla][INFO] - Step 4034, Loss: {'train/loss': 0.16049163043498993, 'train/grad_norm_pre_clip': 9.97402286529541, 'timing/data': 0.009399262955412269, 'timing/model': 1.2432554740225896, 'learning_rate/qwen_vl_interface': 1.3822636637567025e-06, 'learning_rate/action_model': 1.3822636637567024e-05, 'epoch': 1.51}) +[2026-06-19 19:01:46,200][starVLA.training.train_starvla][INFO] - Step 4035, Loss: {'train/loss': 0.26806703209877014, 'train/grad_norm_pre_clip': 16.5447940826416, 'timing/data': 0.009925759048201144, 'timing/model': 1.1480510329129174, 'learning_rate/qwen_vl_interface': 1.3804965918124788e-06, 'learning_rate/action_model': 1.3804965918124787e-05, 'epoch': 1.51}) +[2026-06-19 19:01:50,758][starVLA.training.train_starvla][INFO] - Step 4036, Loss: {'train/loss': 0.29504209756851196, 'train/grad_norm_pre_clip': 15.542970657348633, 'timing/data': 0.009630314074456692, 'timing/model': 1.1582084510009736, 'learning_rate/qwen_vl_interface': 1.3787311104749418e-06, 'learning_rate/action_model': 1.3787311104749418e-05, 'epoch': 1.51}) +[2026-06-19 19:01:55,378][starVLA.training.train_starvla][INFO] - Step 4037, Loss: {'train/loss': 0.13129551708698273, 'train/grad_norm_pre_clip': 8.872642517089844, 'timing/data': 0.009788227966055274, 'timing/model': 1.1471324760932475, 'learning_rate/qwen_vl_interface': 1.3769672204698149e-06, 'learning_rate/action_model': 1.3769672204698147e-05, 'epoch': 1.51}) +[2026-06-19 19:02:00,000][starVLA.training.train_starvla][INFO] - Step 4038, Loss: {'train/loss': 0.21960990130901337, 'train/grad_norm_pre_clip': 11.490447044372559, 'timing/data': 0.009618664043955505, 'timing/model': 1.1527228619670495, 'learning_rate/qwen_vl_interface': 1.3752049225221664e-06, 'learning_rate/action_model': 1.3752049225221663e-05, 'epoch': 1.51}) +[2026-06-19 19:02:04,657][starVLA.training.train_starvla][INFO] - Step 4039, Loss: {'train/loss': 0.17741313576698303, 'train/grad_norm_pre_clip': 9.722824096679688, 'timing/data': 0.00995039101690054, 'timing/model': 1.3546351010445505, 'learning_rate/qwen_vl_interface': 1.3734442173564103e-06, 'learning_rate/action_model': 1.3734442173564102e-05, 'epoch': 1.51}) +[2026-06-19 19:02:09,108][starVLA.training.train_starvla][INFO] - Step 4040, Loss: {'train/loss': 0.2487543225288391, 'train/grad_norm_pre_clip': 10.079507827758789, 'timing/data': 0.0092474709963426, 'timing/model': 1.1531122219748795, 'learning_rate/qwen_vl_interface': 1.3716851056963062e-06, 'learning_rate/action_model': 1.3716851056963062e-05, 'epoch': 1.51}) +[2026-06-19 19:02:13,738][starVLA.training.train_starvla][INFO] - Step 4041, Loss: {'train/loss': 0.14463713765144348, 'train/grad_norm_pre_clip': 9.300802230834961, 'timing/data': 0.009806990972720087, 'timing/model': 1.1624073829734698, 'learning_rate/qwen_vl_interface': 1.3699275882649597e-06, 'learning_rate/action_model': 1.3699275882649596e-05, 'epoch': 1.52}) +[2026-06-19 19:02:18,308][starVLA.training.train_starvla][INFO] - Step 4042, Loss: {'train/loss': 0.21550139784812927, 'train/grad_norm_pre_clip': 12.402100563049316, 'timing/data': 0.0099472840083763, 'timing/model': 1.152362476917915, 'learning_rate/qwen_vl_interface': 1.368171665784817e-06, 'learning_rate/action_model': 1.3681716657848171e-05, 'epoch': 1.52}) +[2026-06-19 19:02:22,944][starVLA.training.train_starvla][INFO] - Step 4043, Loss: {'train/loss': 0.3340277373790741, 'train/grad_norm_pre_clip': 17.431476593017578, 'timing/data': 0.009557379060424864, 'timing/model': 1.1730154081014916, 'learning_rate/qwen_vl_interface': 1.3664173389776735e-06, 'learning_rate/action_model': 1.3664173389776735e-05, 'epoch': 1.52}) +[2026-06-19 19:02:27,551][starVLA.training.train_starvla][INFO] - Step 4044, Loss: {'train/loss': 0.21464109420776367, 'train/grad_norm_pre_clip': 12.31332015991211, 'timing/data': 0.009918213938362896, 'timing/model': 1.1475626199971884, 'learning_rate/qwen_vl_interface': 1.3646646085646658e-06, 'learning_rate/action_model': 1.3646646085646659e-05, 'epoch': 1.52}) +[2026-06-19 19:02:32,153][starVLA.training.train_starvla][INFO] - Step 4045, Loss: {'train/loss': 0.1890985369682312, 'train/grad_norm_pre_clip': 12.408374786376953, 'timing/data': 0.00963799492456019, 'timing/model': 1.3112211200641468, 'learning_rate/qwen_vl_interface': 1.3629134752662756e-06, 'learning_rate/action_model': 1.3629134752662756e-05, 'epoch': 1.52}) +[2026-06-19 19:02:36,576][starVLA.training.train_starvla][INFO] - Step 4046, Loss: {'train/loss': 0.2243587076663971, 'train/grad_norm_pre_clip': 13.492704391479492, 'timing/data': 0.009663089993409812, 'timing/model': 1.1463346489472315, 'learning_rate/qwen_vl_interface': 1.3611639398023282e-06, 'learning_rate/action_model': 1.3611639398023283e-05, 'epoch': 1.52}) +[2026-06-19 19:02:41,166][starVLA.training.train_starvla][INFO] - Step 4047, Loss: {'train/loss': 0.16959655284881592, 'train/grad_norm_pre_clip': 15.422148704528809, 'timing/data': 0.009962382027879357, 'timing/model': 1.1449619489721954, 'learning_rate/qwen_vl_interface': 1.35941600289199e-06, 'learning_rate/action_model': 1.35941600289199e-05, 'epoch': 1.52}) +[2026-06-19 19:02:45,684][starVLA.training.train_starvla][INFO] - Step 4048, Loss: {'train/loss': 0.2682022154331207, 'train/grad_norm_pre_clip': 17.88319206237793, 'timing/data': 0.009836946032010019, 'timing/model': 1.1477540569612756, 'learning_rate/qwen_vl_interface': 1.3576696652537735e-06, 'learning_rate/action_model': 1.3576696652537736e-05, 'epoch': 1.52}) +[2026-06-19 19:02:50,349][starVLA.training.train_starvla][INFO] - Step 4049, Loss: {'train/loss': 0.3735126256942749, 'train/grad_norm_pre_clip': 19.11682891845703, 'timing/data': 0.009976597037166357, 'timing/model': 1.1596571339759976, 'learning_rate/qwen_vl_interface': 1.355924927605531e-06, 'learning_rate/action_model': 1.3559249276055311e-05, 'epoch': 1.52}) +[2026-06-19 19:02:55,004][starVLA.training.train_starvla][INFO] - Step 4050, Loss: {'train/loss': 0.18260110914707184, 'train/grad_norm_pre_clip': 11.741263389587402, 'timing/data': 0.009620822966098785, 'timing/model': 1.1556241540238261, 'learning_rate/qwen_vl_interface': 1.3541817906644591e-06, 'learning_rate/action_model': 1.3541817906644591e-05, 'epoch': 1.52}) +[2026-06-19 19:02:59,620][starVLA.training.train_starvla][INFO] - Step 4051, Loss: {'train/loss': 0.216706782579422, 'train/grad_norm_pre_clip': 14.556609153747559, 'timing/data': 0.01014558202587068, 'timing/model': 1.316835750010796, 'learning_rate/qwen_vl_interface': 1.3524402551470955e-06, 'learning_rate/action_model': 1.3524402551470956e-05, 'epoch': 1.52}) +[2026-06-19 19:03:04,058][starVLA.training.train_starvla][INFO] - Step 4052, Loss: {'train/loss': 0.17589519917964935, 'train/grad_norm_pre_clip': 26.58428382873535, 'timing/data': 0.010125786997377872, 'timing/model': 1.1523435609415174, 'learning_rate/qwen_vl_interface': 1.3507003217693196e-06, 'learning_rate/action_model': 1.3507003217693196e-05, 'epoch': 1.52}) +[2026-06-19 19:03:08,660][starVLA.training.train_starvla][INFO] - Step 4053, Loss: {'train/loss': 0.12478525191545486, 'train/grad_norm_pre_clip': 13.711898803710938, 'timing/data': 0.009798016981221735, 'timing/model': 1.1470757420174778, 'learning_rate/qwen_vl_interface': 1.3489619912463515e-06, 'learning_rate/action_model': 1.3489619912463514e-05, 'epoch': 1.52}) +[2026-06-19 19:03:13,265][starVLA.training.train_starvla][INFO] - Step 4054, Loss: {'train/loss': 0.12202028930187225, 'train/grad_norm_pre_clip': 9.662975311279297, 'timing/data': 0.009974940097890794, 'timing/model': 1.1467312750173733, 'learning_rate/qwen_vl_interface': 1.3472252642927534e-06, 'learning_rate/action_model': 1.3472252642927533e-05, 'epoch': 1.52}) +[2026-06-19 19:03:17,779][starVLA.training.train_starvla][INFO] - Step 4055, Loss: {'train/loss': 0.29503515362739563, 'train/grad_norm_pre_clip': 18.151546478271484, 'timing/data': 0.011713422951288521, 'timing/model': 1.1623641810147092, 'learning_rate/qwen_vl_interface': 1.3454901416224284e-06, 'learning_rate/action_model': 1.3454901416224284e-05, 'epoch': 1.52}) +[2026-06-19 19:03:22,305][starVLA.training.train_starvla][INFO] - Step 4056, Loss: {'train/loss': 0.2689577043056488, 'train/grad_norm_pre_clip': 16.08414649963379, 'timing/data': 0.009718053974211216, 'timing/model': 1.1509345909580588, 'learning_rate/qwen_vl_interface': 1.34375662394862e-06, 'learning_rate/action_model': 1.34375662394862e-05, 'epoch': 1.52}) +[2026-06-19 19:03:26,868][starVLA.training.train_starvla][INFO] - Step 4057, Loss: {'train/loss': 0.17805129289627075, 'train/grad_norm_pre_clip': 13.11672592163086, 'timing/data': 0.011568953981623054, 'timing/model': 1.2651771940290928, 'learning_rate/qwen_vl_interface': 1.3420247119839116e-06, 'learning_rate/action_model': 1.3420247119839117e-05, 'epoch': 1.52}) +[2026-06-19 19:03:31,310][starVLA.training.train_starvla][INFO] - Step 4058, Loss: {'train/loss': 0.13993310928344727, 'train/grad_norm_pre_clip': 18.26658058166504, 'timing/data': 0.009702584007754922, 'timing/model': 1.1473679470364004, 'learning_rate/qwen_vl_interface': 1.340294406440227e-06, 'learning_rate/action_model': 1.340294406440227e-05, 'epoch': 1.52}) +[2026-06-19 19:03:35,942][starVLA.training.train_starvla][INFO] - Step 4059, Loss: {'train/loss': 0.24151617288589478, 'train/grad_norm_pre_clip': 16.231922149658203, 'timing/data': 0.009440648020245135, 'timing/model': 1.1517825790215284, 'learning_rate/qwen_vl_interface': 1.3385657080288297e-06, 'learning_rate/action_model': 1.3385657080288297e-05, 'epoch': 1.52}) +[2026-06-19 19:03:40,557][starVLA.training.train_starvla][INFO] - Step 4060, Loss: {'train/loss': 0.10879932343959808, 'train/grad_norm_pre_clip': 11.270894050598145, 'timing/data': 0.009750180994160473, 'timing/model': 1.1486594449961558, 'learning_rate/qwen_vl_interface': 1.336838617460322e-06, 'learning_rate/action_model': 1.336838617460322e-05, 'epoch': 1.52}) +[2026-06-19 19:03:45,184][starVLA.training.train_starvla][INFO] - Step 4061, Loss: {'train/loss': 0.1418587863445282, 'train/grad_norm_pre_clip': 12.005536079406738, 'timing/data': 0.009571593021973968, 'timing/model': 1.1533196889795363, 'learning_rate/qwen_vl_interface': 1.335113135444646e-06, 'learning_rate/action_model': 1.335113135444646e-05, 'epoch': 1.52}) +[2026-06-19 19:03:49,862][starVLA.training.train_starvla][INFO] - Step 4062, Loss: {'train/loss': 0.05706392973661423, 'train/grad_norm_pre_clip': 6.496501445770264, 'timing/data': 0.009685044060461223, 'timing/model': 1.3414957060012966, 'learning_rate/qwen_vl_interface': 1.3333892626910823e-06, 'learning_rate/action_model': 1.3333892626910823e-05, 'epoch': 1.52}) +[2026-06-19 19:03:54,298][starVLA.training.train_starvla][INFO] - Step 4063, Loss: {'train/loss': 0.3041197955608368, 'train/grad_norm_pre_clip': 17.715803146362305, 'timing/data': 0.009622353012673557, 'timing/model': 1.1451785250101238, 'learning_rate/qwen_vl_interface': 1.3316669999082486e-06, 'learning_rate/action_model': 1.3316669999082484e-05, 'epoch': 1.52}) +[2026-06-19 19:03:58,895][starVLA.training.train_starvla][INFO] - Step 4064, Loss: {'train/loss': 0.26640042662620544, 'train/grad_norm_pre_clip': 18.355009078979492, 'timing/data': 0.01013558404520154, 'timing/model': 1.1463888310827315, 'learning_rate/qwen_vl_interface': 1.3299463478041024e-06, 'learning_rate/action_model': 1.3299463478041024e-05, 'epoch': 1.52}) +[2026-06-19 19:04:03,499][starVLA.training.train_starvla][INFO] - Step 4065, Loss: {'train/loss': 0.08322962373495102, 'train/grad_norm_pre_clip': 7.748061180114746, 'timing/data': 0.009383957018144429, 'timing/model': 1.1549947070889175, 'learning_rate/qwen_vl_interface': 1.3282273070859396e-06, 'learning_rate/action_model': 1.3282273070859396e-05, 'epoch': 1.52}) +[2026-06-19 19:04:08,122][starVLA.training.train_starvla][INFO] - Step 4066, Loss: {'train/loss': 0.23451316356658936, 'train/grad_norm_pre_clip': 10.662515640258789, 'timing/data': 0.00994161597918719, 'timing/model': 1.146938492078334, 'learning_rate/qwen_vl_interface': 1.3265098784603924e-06, 'learning_rate/action_model': 1.3265098784603924e-05, 'epoch': 1.52}) +[2026-06-19 19:04:12,761][starVLA.training.train_starvla][INFO] - Step 4067, Loss: {'train/loss': 0.1671610325574875, 'train/grad_norm_pre_clip': 18.438373565673828, 'timing/data': 0.011900631012395024, 'timing/model': 1.1611146720824763, 'learning_rate/qwen_vl_interface': 1.324794062633431e-06, 'learning_rate/action_model': 1.324794062633431e-05, 'epoch': 1.52}) +[2026-06-19 19:04:17,384][starVLA.training.train_starvla][INFO] - Step 4068, Loss: {'train/loss': 0.1554376184940338, 'train/grad_norm_pre_clip': 13.66493034362793, 'timing/data': 0.009915814967826009, 'timing/model': 1.325495760072954, 'learning_rate/qwen_vl_interface': 1.323079860310363e-06, 'learning_rate/action_model': 1.3230798603103628e-05, 'epoch': 1.53}) +[2026-06-19 19:04:21,932][starVLA.training.train_starvla][INFO] - Step 4069, Loss: {'train/loss': 0.2872501015663147, 'train/grad_norm_pre_clip': 14.804677963256836, 'timing/data': 0.011647996958345175, 'timing/model': 1.2573229889385402, 'learning_rate/qwen_vl_interface': 1.3213672721958309e-06, 'learning_rate/action_model': 1.321367272195831e-05, 'epoch': 1.53}) +[2026-06-19 19:04:26,384][starVLA.training.train_starvla][INFO] - Step 4070, Loss: {'train/loss': 0.316910982131958, 'train/grad_norm_pre_clip': 14.73558521270752, 'timing/data': 0.010282885981723666, 'timing/model': 1.1509812630247325, 'learning_rate/qwen_vl_interface': 1.3196562989938175e-06, 'learning_rate/action_model': 1.3196562989938174e-05, 'epoch': 1.53}) +[2026-06-19 19:04:31,001][starVLA.training.train_starvla][INFO] - Step 4071, Loss: {'train/loss': 0.20224128663539886, 'train/grad_norm_pre_clip': 11.515846252441406, 'timing/data': 0.012162910075858235, 'timing/model': 1.1495067060459405, 'learning_rate/qwen_vl_interface': 1.3179469414076357e-06, 'learning_rate/action_model': 1.3179469414076356e-05, 'epoch': 1.53}) +[2026-06-19 19:04:35,611][starVLA.training.train_starvla][INFO] - Step 4072, Loss: {'train/loss': 0.20542261004447937, 'train/grad_norm_pre_clip': 11.875879287719727, 'timing/data': 0.009845172055065632, 'timing/model': 1.1519996990682557, 'learning_rate/qwen_vl_interface': 1.3162392001399401e-06, 'learning_rate/action_model': 1.3162392001399401e-05, 'epoch': 1.53}) +[2026-06-19 19:04:40,212][starVLA.training.train_starvla][INFO] - Step 4073, Loss: {'train/loss': 0.157205730676651, 'train/grad_norm_pre_clip': 13.52424430847168, 'timing/data': 0.010370910982601345, 'timing/model': 1.1496138781076297, 'learning_rate/qwen_vl_interface': 1.3145330758927182e-06, 'learning_rate/action_model': 1.3145330758927182e-05, 'epoch': 1.53}) +[2026-06-19 19:04:44,803][starVLA.training.train_starvla][INFO] - Step 4074, Loss: {'train/loss': 0.15032680332660675, 'train/grad_norm_pre_clip': 9.659638404846191, 'timing/data': 0.010040075983852148, 'timing/model': 1.1496001250343397, 'learning_rate/qwen_vl_interface': 1.3128285693672935e-06, 'learning_rate/action_model': 1.3128285693672934e-05, 'epoch': 1.53}) +[2026-06-19 19:04:49,403][starVLA.training.train_starvla][INFO] - Step 4075, Loss: {'train/loss': 0.22075310349464417, 'train/grad_norm_pre_clip': 13.375479698181152, 'timing/data': 0.00922569097019732, 'timing/model': 1.3131459089927375, 'learning_rate/qwen_vl_interface': 1.311125681264325e-06, 'learning_rate/action_model': 1.311125681264325e-05, 'epoch': 1.53}) +[2026-06-19 19:04:53,825][starVLA.training.train_starvla][INFO] - Step 4076, Loss: {'train/loss': 0.20382001996040344, 'train/grad_norm_pre_clip': 12.142016410827637, 'timing/data': 0.010036005987785757, 'timing/model': 1.1459472810383886, 'learning_rate/qwen_vl_interface': 1.309424412283803e-06, 'learning_rate/action_model': 1.3094244122838029e-05, 'epoch': 1.53}) +[2026-06-19 19:04:58,412][starVLA.training.train_starvla][INFO] - Step 4077, Loss: {'train/loss': 0.26923537254333496, 'train/grad_norm_pre_clip': 14.889673233032227, 'timing/data': 0.00979118689429015, 'timing/model': 1.1465318630216643, 'learning_rate/qwen_vl_interface': 1.307724763125057e-06, 'learning_rate/action_model': 1.307724763125057e-05, 'epoch': 1.53}) +[2026-06-19 19:05:03,057][starVLA.training.train_starvla][INFO] - Step 4078, Loss: {'train/loss': 0.11686819791793823, 'train/grad_norm_pre_clip': 10.800915718078613, 'timing/data': 0.009581603924743831, 'timing/model': 1.1622039909707382, 'learning_rate/qwen_vl_interface': 1.3060267344867478e-06, 'learning_rate/action_model': 1.3060267344867478e-05, 'epoch': 1.53}) +[2026-06-19 19:05:07,665][starVLA.training.train_starvla][INFO] - Step 4079, Loss: {'train/loss': 0.19564218819141388, 'train/grad_norm_pre_clip': 19.83374786376953, 'timing/data': 0.009826033026911318, 'timing/model': 1.1537521919235587, 'learning_rate/qwen_vl_interface': 1.3043303270668715e-06, 'learning_rate/action_model': 1.3043303270668713e-05, 'epoch': 1.53}) +[2026-06-19 19:05:12,312][starVLA.training.train_starvla][INFO] - Step 4080, Loss: {'train/loss': 0.22832676768302917, 'train/grad_norm_pre_clip': 11.79826545715332, 'timing/data': 0.012317331973463297, 'timing/model': 1.1637454968877137, 'learning_rate/qwen_vl_interface': 1.3026355415627565e-06, 'learning_rate/action_model': 1.3026355415627564e-05, 'epoch': 1.53}) +[2026-06-19 19:05:16,924][starVLA.training.train_starvla][INFO] - Step 4081, Loss: {'train/loss': 0.2348392903804779, 'train/grad_norm_pre_clip': 11.660246849060059, 'timing/data': 0.009868587017990649, 'timing/model': 1.3137737569632009, 'learning_rate/qwen_vl_interface': 1.3009423786710653e-06, 'learning_rate/action_model': 1.3009423786710653e-05, 'epoch': 1.53}) +[2026-06-19 19:05:21,375][starVLA.training.train_starvla][INFO] - Step 4082, Loss: {'train/loss': 0.055653803050518036, 'train/grad_norm_pre_clip': 15.811111450195312, 'timing/data': 0.009869177010841668, 'timing/model': 1.1566049389075488, 'learning_rate/qwen_vl_interface': 1.2992508390877933e-06, 'learning_rate/action_model': 1.2992508390877933e-05, 'epoch': 1.53}) +[2026-06-19 19:05:26,003][starVLA.training.train_starvla][INFO] - Step 4083, Loss: {'train/loss': 0.31550151109695435, 'train/grad_norm_pre_clip': 15.075516700744629, 'timing/data': 0.009630108019337058, 'timing/model': 1.1595456700306386, 'learning_rate/qwen_vl_interface': 1.297560923508269e-06, 'learning_rate/action_model': 1.297560923508269e-05, 'epoch': 1.53}) +[2026-06-19 19:05:30,530][starVLA.training.train_starvla][INFO] - Step 4084, Loss: {'train/loss': 0.2672404646873474, 'train/grad_norm_pre_clip': 19.149608612060547, 'timing/data': 0.009763889014720917, 'timing/model': 1.1627378750126809, 'learning_rate/qwen_vl_interface': 1.295872632627152e-06, 'learning_rate/action_model': 1.2958726326271519e-05, 'epoch': 1.53}) +[2026-06-19 19:05:35,080][starVLA.training.train_starvla][INFO] - Step 4085, Loss: {'train/loss': 0.2759062945842743, 'train/grad_norm_pre_clip': 13.323226928710938, 'timing/data': 0.010346627910621464, 'timing/model': 1.1475740839960054, 'learning_rate/qwen_vl_interface': 1.2941859671384364e-06, 'learning_rate/action_model': 1.2941859671384362e-05, 'epoch': 1.53}) +[2026-06-19 19:05:39,703][starVLA.training.train_starvla][INFO] - Step 4086, Loss: {'train/loss': 0.2508815824985504, 'train/grad_norm_pre_clip': 15.502120971679688, 'timing/data': 0.009957044967450202, 'timing/model': 1.1511172419413924, 'learning_rate/qwen_vl_interface': 1.2925009277354444e-06, 'learning_rate/action_model': 1.2925009277354444e-05, 'epoch': 1.53}) +[2026-06-19 19:05:44,291][starVLA.training.train_starvla][INFO] - Step 4087, Loss: {'train/loss': 0.2758607268333435, 'train/grad_norm_pre_clip': 18.012540817260742, 'timing/data': 0.009779967018403113, 'timing/model': 1.3083876910386607, 'learning_rate/qwen_vl_interface': 1.2908175151108335e-06, 'learning_rate/action_model': 1.2908175151108334e-05, 'epoch': 1.53}) +[2026-06-19 19:05:48,727][starVLA.training.train_starvla][INFO] - Step 4088, Loss: {'train/loss': 0.2099972367286682, 'train/grad_norm_pre_clip': 16.665925979614258, 'timing/data': 0.009149190969765186, 'timing/model': 1.1474098389735445, 'learning_rate/qwen_vl_interface': 1.2891357299565899e-06, 'learning_rate/action_model': 1.2891357299565898e-05, 'epoch': 1.53}) +[2026-06-19 19:05:53,337][starVLA.training.train_starvla][INFO] - Step 4089, Loss: {'train/loss': 0.14314329624176025, 'train/grad_norm_pre_clip': 13.501693725585938, 'timing/data': 0.010540161980316043, 'timing/model': 1.1486464709741995, 'learning_rate/qwen_vl_interface': 1.2874555729640333e-06, 'learning_rate/action_model': 1.2874555729640334e-05, 'epoch': 1.53}) +[2026-06-19 19:05:57,962][starVLA.training.train_starvla][INFO] - Step 4090, Loss: {'train/loss': 0.263247013092041, 'train/grad_norm_pre_clip': 11.38191032409668, 'timing/data': 0.009927626000717282, 'timing/model': 1.1533236160175875, 'learning_rate/qwen_vl_interface': 1.2857770448238124e-06, 'learning_rate/action_model': 1.2857770448238123e-05, 'epoch': 1.53}) +[2026-06-19 19:06:02,525][starVLA.training.train_starvla][INFO] - Step 4091, Loss: {'train/loss': 0.415616512298584, 'train/grad_norm_pre_clip': 15.557526588439941, 'timing/data': 0.009660679963417351, 'timing/model': 1.1596532620023936, 'learning_rate/qwen_vl_interface': 1.2841001462259067e-06, 'learning_rate/action_model': 1.2841001462259067e-05, 'epoch': 1.53}) +[2026-06-19 19:06:07,115][starVLA.training.train_starvla][INFO] - Step 4092, Loss: {'train/loss': 0.164512038230896, 'train/grad_norm_pre_clip': 13.335248947143555, 'timing/data': 0.009544358006678522, 'timing/model': 1.1463820210192353, 'learning_rate/qwen_vl_interface': 1.282424877859626e-06, 'learning_rate/action_model': 1.282424877859626e-05, 'epoch': 1.53}) +[2026-06-19 19:06:11,712][starVLA.training.train_starvla][INFO] - Step 4093, Loss: {'train/loss': 0.11902092397212982, 'train/grad_norm_pre_clip': 17.95258903503418, 'timing/data': 0.009905620943754911, 'timing/model': 1.3177443380700424, 'learning_rate/qwen_vl_interface': 1.2807512404136088e-06, 'learning_rate/action_model': 1.2807512404136087e-05, 'epoch': 1.53}) +[2026-06-19 19:06:16,160][starVLA.training.train_starvla][INFO] - Step 4094, Loss: {'train/loss': 0.23923929035663605, 'train/grad_norm_pre_clip': 38.44987106323242, 'timing/data': 0.010410759947262704, 'timing/model': 1.1531222080811858, 'learning_rate/qwen_vl_interface': 1.2790792345758255e-06, 'learning_rate/action_model': 1.2790792345758254e-05, 'epoch': 1.54}) +[2026-06-19 19:06:20,776][starVLA.training.train_starvla][INFO] - Step 4095, Loss: {'train/loss': 0.09705349802970886, 'train/grad_norm_pre_clip': 7.194062232971191, 'timing/data': 0.009798106038942933, 'timing/model': 1.1455954189877957, 'learning_rate/qwen_vl_interface': 1.2774088610335752e-06, 'learning_rate/action_model': 1.2774088610335752e-05, 'epoch': 1.54}) +[2026-06-19 19:06:25,385][starVLA.training.train_starvla][INFO] - Step 4096, Loss: {'train/loss': 0.2303847074508667, 'train/grad_norm_pre_clip': 11.123865127563477, 'timing/data': 0.009489250951446593, 'timing/model': 1.1465363629395142, 'learning_rate/qwen_vl_interface': 1.2757401204734827e-06, 'learning_rate/action_model': 1.2757401204734826e-05, 'epoch': 1.54}) +[2026-06-19 19:06:30,009][starVLA.training.train_starvla][INFO] - Step 4097, Loss: {'train/loss': 0.24711117148399353, 'train/grad_norm_pre_clip': 18.226238250732422, 'timing/data': 0.00955937395337969, 'timing/model': 1.1612781699514017, 'learning_rate/qwen_vl_interface': 1.2740730135815064e-06, 'learning_rate/action_model': 1.2740730135815062e-05, 'epoch': 1.54}) +[2026-06-19 19:06:34,657][starVLA.training.train_starvla][INFO] - Step 4098, Loss: {'train/loss': 0.1853240728378296, 'train/grad_norm_pre_clip': 11.064367294311523, 'timing/data': 0.009307836997322738, 'timing/model': 1.1504249619320035, 'learning_rate/qwen_vl_interface': 1.2724075410429298e-06, 'learning_rate/action_model': 1.2724075410429299e-05, 'epoch': 1.54}) +[2026-06-19 19:06:39,277][starVLA.training.train_starvla][INFO] - Step 4099, Loss: {'train/loss': 0.19855761528015137, 'train/grad_norm_pre_clip': 16.328645706176758, 'timing/data': 0.009950355044566095, 'timing/model': 1.3205222371034324, 'learning_rate/qwen_vl_interface': 1.2707437035423664e-06, 'learning_rate/action_model': 1.2707437035423664e-05, 'epoch': 1.54}) +[2026-06-19 19:06:43,718][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 4100 +[2026-06-19 19:09:38,120][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 4100 +[2026-06-19 19:09:38,120][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 4100 +[2026-06-19 19:12:48,840][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 4100 +[2026-06-19 19:12:48,841][starVLA.training.train_starvla][INFO] - Step 4100, Loss: {'train/loss': 0.2720748782157898, 'train/grad_norm_pre_clip': 11.717508316040039, 'eval/loss': 0.13261888921260834, 'eval/latency_0/loss': 0.08012714982032776, 'eval/latency_1/loss': 0.04959312826395035, 'eval/latency_2/loss': 0.11000950634479523, 'eval/latency_3/loss': 0.1579245626926422, 'eval/latency_4/loss': 0.26637381315231323, 'eval/flappy/loss': 0.13261888921260834, 'eval/flappy/latency_0/loss': 0.08012714982032776, 'eval/flappy/latency_1/loss': 0.04959312826395035, 'eval/flappy/latency_2/loss': 0.11000950634479523, 'eval/flappy/latency_3/loss': 0.1579245626926422, 'eval/flappy/latency_4/loss': 0.26637381315231323, 'eval/flap_precision': 0.7360594795539034, 'eval/flap_recall': 0.72, 'eval/flap_f1': 0.7279411764705882, 'eval/flap_rate_model': 0.07472222222222222, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.7279411764705882, 'eval/latency_0/flap_precision': 0.7360594795539034, 'eval/latency_0/flap_recall': 0.72, 'eval/latency_0/flap_f1': 0.7279411764705882, 'eval/latency_0/flap_rate_model': 0.07472222222222222, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.7279411764705882, 'timing/data': 0.009731604950502515, 'timing/model': 1.1496925899991766, 'learning_rate/qwen_vl_interface': 1.2690815017637573e-06, 'learning_rate/action_model': 1.2690815017637572e-05, 'epoch': 1.54}) +[2026-06-19 19:12:48,842][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_4100_state +[2026-06-19 19:12:48,845][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-19 19:12:52,247][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_4100_state/model.safetensors +[2026-06-19 19:12:59,734][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_4100_state/optimizer.bin +[2026-06-19 19:12:59,735][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_4100_state/sampler.bin +[2026-06-19 19:12:59,735][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_4100_state/sampler_1.bin +[2026-06-19 19:12:59,736][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_4100_state/random_states_0.pkl +[2026-06-19 19:12:59,759][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4100 +[2026-06-19 19:12:59,759][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-19 19:12:59,782][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-19 19:13:04,310][starVLA.training.train_starvla][INFO] - Step 4101, Loss: {'train/loss': 0.19173775613307953, 'train/grad_norm_pre_clip': 11.129652976989746, 'timing/data': 0.010830104001797736, 'timing/model': 1.1199329169467092, 'learning_rate/qwen_vl_interface': 1.2674209363903694e-06, 'learning_rate/action_model': 1.2674209363903693e-05, 'epoch': 1.54}) +[2026-06-19 19:13:08,834][starVLA.training.train_starvla][INFO] - Step 4102, Loss: {'train/loss': 0.24707502126693726, 'train/grad_norm_pre_clip': 20.07937240600586, 'timing/data': 0.009762258036062121, 'timing/model': 1.128499332931824, 'learning_rate/qwen_vl_interface': 1.2657620081048005e-06, 'learning_rate/action_model': 1.2657620081048006e-05, 'epoch': 1.54}) +[2026-06-19 19:13:13,389][starVLA.training.train_starvla][INFO] - Step 4103, Loss: {'train/loss': 0.10234080255031586, 'train/grad_norm_pre_clip': 8.721753120422363, 'timing/data': 0.010080171981826425, 'timing/model': 1.3229864500463009, 'learning_rate/qwen_vl_interface': 1.2641047175889714e-06, 'learning_rate/action_model': 1.2641047175889714e-05, 'epoch': 1.54}) +[2026-06-19 19:13:17,762][starVLA.training.train_starvla][INFO] - Step 4104, Loss: {'train/loss': 0.2794034779071808, 'train/grad_norm_pre_clip': 24.436626434326172, 'timing/data': 0.009840447921305895, 'timing/model': 1.1338336380431429, 'learning_rate/qwen_vl_interface': 1.2624490655241332e-06, 'learning_rate/action_model': 1.262449065524133e-05, 'epoch': 1.54}) +[2026-06-19 19:13:22,314][starVLA.training.train_starvla][INFO] - Step 4105, Loss: {'train/loss': 0.21698933839797974, 'train/grad_norm_pre_clip': 13.522817611694336, 'timing/data': 0.00977450201753527, 'timing/model': 1.1360155110014603, 'learning_rate/qwen_vl_interface': 1.2607950525908602e-06, 'learning_rate/action_model': 1.2607950525908601e-05, 'epoch': 1.54}) +[2026-06-19 19:13:26,868][starVLA.training.train_starvla][INFO] - Step 4106, Loss: {'train/loss': 0.16013099253177643, 'train/grad_norm_pre_clip': 8.469511985778809, 'timing/data': 0.009610448963940144, 'timing/model': 1.1400186259998009, 'learning_rate/qwen_vl_interface': 1.2591426794690575e-06, 'learning_rate/action_model': 1.2591426794690576e-05, 'epoch': 1.54}) +[2026-06-19 19:13:31,421][starVLA.training.train_starvla][INFO] - Step 4107, Loss: {'train/loss': 0.36988794803619385, 'train/grad_norm_pre_clip': 14.463738441467285, 'timing/data': 0.00960053806193173, 'timing/model': 1.1436905859736726, 'learning_rate/qwen_vl_interface': 1.25749194683795e-06, 'learning_rate/action_model': 1.25749194683795e-05, 'epoch': 1.54}) +[2026-06-19 19:13:35,977][starVLA.training.train_starvla][INFO] - Step 4108, Loss: {'train/loss': 0.3861851096153259, 'train/grad_norm_pre_clip': 15.497726440429688, 'timing/data': 0.009775026934221387, 'timing/model': 1.1363791460171342, 'learning_rate/qwen_vl_interface': 1.2558428553760933e-06, 'learning_rate/action_model': 1.2558428553760932e-05, 'epoch': 1.54}) +[2026-06-19 19:13:40,568][starVLA.training.train_starvla][INFO] - Step 4109, Loss: {'train/loss': 0.18147265911102295, 'train/grad_norm_pre_clip': 9.180336952209473, 'timing/data': 0.009841368068009615, 'timing/model': 1.3332461270038038, 'learning_rate/qwen_vl_interface': 1.2541954057613659e-06, 'learning_rate/action_model': 1.2541954057613658e-05, 'epoch': 1.54}) +[2026-06-19 19:13:44,975][starVLA.training.train_starvla][INFO] - Step 4110, Loss: {'train/loss': 0.2360648661851883, 'train/grad_norm_pre_clip': 14.60353946685791, 'timing/data': 0.009165429044514894, 'timing/model': 1.13677448395174, 'learning_rate/qwen_vl_interface': 1.2525495986709727e-06, 'learning_rate/action_model': 1.2525495986709726e-05, 'epoch': 1.54}) +[2026-06-19 19:13:49,553][starVLA.training.train_starvla][INFO] - Step 4111, Loss: {'train/loss': 0.3063024580478668, 'train/grad_norm_pre_clip': 19.402732849121094, 'timing/data': 0.009411975042894483, 'timing/model': 1.1450713940430433, 'learning_rate/qwen_vl_interface': 1.2509054347814423e-06, 'learning_rate/action_model': 1.2509054347814423e-05, 'epoch': 1.54}) +[2026-06-19 19:13:54,061][starVLA.training.train_starvla][INFO] - Step 4112, Loss: {'train/loss': 0.4382396936416626, 'train/grad_norm_pre_clip': 29.312437057495117, 'timing/data': 0.00974535197019577, 'timing/model': 1.152893577935174, 'learning_rate/qwen_vl_interface': 1.2492629147686288e-06, 'learning_rate/action_model': 1.2492629147686289e-05, 'epoch': 1.54}) +[2026-06-19 19:13:58,626][starVLA.training.train_starvla][INFO] - Step 4113, Loss: {'train/loss': 0.2692038416862488, 'train/grad_norm_pre_clip': 18.609960556030273, 'timing/data': 0.010080692009069026, 'timing/model': 1.1566065810620785, 'learning_rate/qwen_vl_interface': 1.2476220393077098e-06, 'learning_rate/action_model': 1.2476220393077098e-05, 'epoch': 1.54}) +[2026-06-19 19:14:03,207][starVLA.training.train_starvla][INFO] - Step 4114, Loss: {'train/loss': 0.430585652589798, 'train/grad_norm_pre_clip': 26.470365524291992, 'timing/data': 0.009701525908894837, 'timing/model': 1.1576606050366536, 'learning_rate/qwen_vl_interface': 1.2459828090731877e-06, 'learning_rate/action_model': 1.2459828090731876e-05, 'epoch': 1.54}) +[2026-06-19 19:14:07,781][starVLA.training.train_starvla][INFO] - Step 4115, Loss: {'train/loss': 0.32007256150245667, 'train/grad_norm_pre_clip': 17.92563819885254, 'timing/data': 0.010029328987002373, 'timing/model': 1.2979372179834172, 'learning_rate/qwen_vl_interface': 1.2443452247388876e-06, 'learning_rate/action_model': 1.2443452247388877e-05, 'epoch': 1.54}) +[2026-06-19 19:14:12,310][starVLA.training.train_starvla][INFO] - Step 4116, Loss: {'train/loss': 0.21597227454185486, 'train/grad_norm_pre_clip': 14.257020950317383, 'timing/data': 0.010365823982283473, 'timing/model': 1.2633267289493233, 'learning_rate/qwen_vl_interface': 1.2427092869779603e-06, 'learning_rate/action_model': 1.2427092869779602e-05, 'epoch': 1.54}) +[2026-06-19 19:14:16,725][starVLA.training.train_starvla][INFO] - Step 4117, Loss: {'train/loss': 0.22835879027843475, 'train/grad_norm_pre_clip': 14.076745986938477, 'timing/data': 0.00961536297108978, 'timing/model': 1.1421020579291508, 'learning_rate/qwen_vl_interface': 1.2410749964628751e-06, 'learning_rate/action_model': 1.2410749964628751e-05, 'epoch': 1.54}) +[2026-06-19 19:14:21,328][starVLA.training.train_starvla][INFO] - Step 4118, Loss: {'train/loss': 0.15031366050243378, 'train/grad_norm_pre_clip': 8.832679748535156, 'timing/data': 0.00979164894670248, 'timing/model': 1.1491515119560063, 'learning_rate/qwen_vl_interface': 1.2394423538654296e-06, 'learning_rate/action_model': 1.2394423538654296e-05, 'epoch': 1.54}) +[2026-06-19 19:14:25,925][starVLA.training.train_starvla][INFO] - Step 4119, Loss: {'train/loss': 0.3705317974090576, 'train/grad_norm_pre_clip': 35.87374496459961, 'timing/data': 0.009446055046282709, 'timing/model': 1.1421917149564251, 'learning_rate/qwen_vl_interface': 1.2378113598567394e-06, 'learning_rate/action_model': 1.2378113598567392e-05, 'epoch': 1.54}) +[2026-06-19 19:14:30,521][starVLA.training.train_starvla][INFO] - Step 4120, Loss: {'train/loss': 0.1332665979862213, 'train/grad_norm_pre_clip': 9.063817977905273, 'timing/data': 0.009935174020938575, 'timing/model': 1.1418176490115002, 'learning_rate/qwen_vl_interface': 1.2361820151072454e-06, 'learning_rate/action_model': 1.2361820151072453e-05, 'epoch': 1.54}) +[2026-06-19 19:14:35,002][starVLA.training.train_starvla][INFO] - Step 4121, Loss: {'train/loss': 0.18976706266403198, 'train/grad_norm_pre_clip': 11.225522994995117, 'timing/data': 0.010065345908515155, 'timing/model': 1.142377682030201, 'learning_rate/qwen_vl_interface': 1.23455432028671e-06, 'learning_rate/action_model': 1.23455432028671e-05, 'epoch': 1.55}) +[2026-06-19 19:14:39,526][starVLA.training.train_starvla][INFO] - Step 4122, Loss: {'train/loss': 0.2687108814716339, 'train/grad_norm_pre_clip': 17.890201568603516, 'timing/data': 0.010122815961949527, 'timing/model': 1.249724863911979, 'learning_rate/qwen_vl_interface': 1.2329282760642167e-06, 'learning_rate/action_model': 1.2329282760642166e-05, 'epoch': 1.55}) +[2026-06-19 19:14:43,953][starVLA.training.train_starvla][INFO] - Step 4123, Loss: {'train/loss': 0.32433003187179565, 'train/grad_norm_pre_clip': 17.234798431396484, 'timing/data': 0.009853948024101555, 'timing/model': 1.1476089929929003, 'learning_rate/qwen_vl_interface': 1.2313038831081716e-06, 'learning_rate/action_model': 1.2313038831081717e-05, 'epoch': 1.55}) +[2026-06-19 19:14:48,565][starVLA.training.train_starvla][INFO] - Step 4124, Loss: {'train/loss': 0.13731014728546143, 'train/grad_norm_pre_clip': 13.375247955322266, 'timing/data': 0.01016034500207752, 'timing/model': 1.1415006560273468, 'learning_rate/qwen_vl_interface': 1.2296811420863002e-06, 'learning_rate/action_model': 1.2296811420863003e-05, 'epoch': 1.55}) +[2026-06-19 19:14:53,151][starVLA.training.train_starvla][INFO] - Step 4125, Loss: {'train/loss': 0.08251746743917465, 'train/grad_norm_pre_clip': 6.115307807922363, 'timing/data': 0.009510239004157484, 'timing/model': 1.151438517961651, 'learning_rate/qwen_vl_interface': 1.2280600536656504e-06, 'learning_rate/action_model': 1.2280600536656503e-05, 'epoch': 1.55}) +[2026-06-19 19:14:57,739][starVLA.training.train_starvla][INFO] - Step 4126, Loss: {'train/loss': 0.25269800424575806, 'train/grad_norm_pre_clip': 18.049055099487305, 'timing/data': 0.010122001054696739, 'timing/model': 1.1477339760167524, 'learning_rate/qwen_vl_interface': 1.2264406185125907e-06, 'learning_rate/action_model': 1.2264406185125907e-05, 'epoch': 1.55}) +[2026-06-19 19:15:02,358][starVLA.training.train_starvla][INFO] - Step 4127, Loss: {'train/loss': 0.17864271998405457, 'train/grad_norm_pre_clip': 20.11583709716797, 'timing/data': 0.0099857960594818, 'timing/model': 1.327008927008137, 'learning_rate/qwen_vl_interface': 1.2248228372928098e-06, 'learning_rate/action_model': 1.2248228372928097e-05, 'epoch': 1.55}) +[2026-06-19 19:15:06,777][starVLA.training.train_starvla][INFO] - Step 4128, Loss: {'train/loss': 0.08306552469730377, 'train/grad_norm_pre_clip': 8.43217658996582, 'timing/data': 0.009747040108777583, 'timing/model': 1.1444159239763394, 'learning_rate/qwen_vl_interface': 1.2232067106713174e-06, 'learning_rate/action_model': 1.2232067106713174e-05, 'epoch': 1.55}) +[2026-06-19 19:15:11,383][starVLA.training.train_starvla][INFO] - Step 4129, Loss: {'train/loss': 0.15344317257404327, 'train/grad_norm_pre_clip': 13.25273323059082, 'timing/data': 0.009765962022356689, 'timing/model': 1.1447995930211619, 'learning_rate/qwen_vl_interface': 1.22159223931244e-06, 'learning_rate/action_model': 1.2215922393124399e-05, 'epoch': 1.55}) +[2026-06-19 19:15:15,996][starVLA.training.train_starvla][INFO] - Step 4130, Loss: {'train/loss': 0.19868338108062744, 'train/grad_norm_pre_clip': 14.496112823486328, 'timing/data': 0.010429550078697503, 'timing/model': 1.144672506954521, 'learning_rate/qwen_vl_interface': 1.2199794238798266e-06, 'learning_rate/action_model': 1.2199794238798264e-05, 'epoch': 1.55}) +[2026-06-19 19:15:20,581][starVLA.training.train_starvla][INFO] - Step 4131, Loss: {'train/loss': 0.19667871296405792, 'train/grad_norm_pre_clip': 13.904559135437012, 'timing/data': 0.009995943983085454, 'timing/model': 1.139387662988156, 'learning_rate/qwen_vl_interface': 1.2183682650364447e-06, 'learning_rate/action_model': 1.2183682650364447e-05, 'epoch': 1.55}) +[2026-06-19 19:15:25,159][starVLA.training.train_starvla][INFO] - Step 4132, Loss: {'train/loss': 0.28835707902908325, 'train/grad_norm_pre_clip': 17.52350616455078, 'timing/data': 0.010473121074028313, 'timing/model': 1.1441732629900798, 'learning_rate/qwen_vl_interface': 1.216758763444581e-06, 'learning_rate/action_model': 1.2167587634445809e-05, 'epoch': 1.55}) +[2026-06-19 19:15:29,751][starVLA.training.train_starvla][INFO] - Step 4133, Loss: {'train/loss': 0.16225282847881317, 'train/grad_norm_pre_clip': 12.657963752746582, 'timing/data': 0.009697081986814737, 'timing/model': 1.312379210954532, 'learning_rate/qwen_vl_interface': 1.215150919765841e-06, 'learning_rate/action_model': 1.215150919765841e-05, 'epoch': 1.55}) +[2026-06-19 19:15:34,271][starVLA.training.train_starvla][INFO] - Step 4134, Loss: {'train/loss': 0.21911770105361938, 'train/grad_norm_pre_clip': 19.182533264160156, 'timing/data': 0.009758777916431427, 'timing/model': 1.2421887340024114, 'learning_rate/qwen_vl_interface': 1.213544734661148e-06, 'learning_rate/action_model': 1.2135447346611478e-05, 'epoch': 1.55}) +[2026-06-19 19:15:38,702][starVLA.training.train_starvla][INFO] - Step 4135, Loss: {'train/loss': 0.057641856372356415, 'train/grad_norm_pre_clip': 6.219425201416016, 'timing/data': 0.009737881948240101, 'timing/model': 1.1467687119729817, 'learning_rate/qwen_vl_interface': 1.2119402087907448e-06, 'learning_rate/action_model': 1.2119402087907448e-05, 'epoch': 1.55}) +[2026-06-19 19:15:43,311][starVLA.training.train_starvla][INFO] - Step 4136, Loss: {'train/loss': 0.28721949458122253, 'train/grad_norm_pre_clip': 15.304582595825195, 'timing/data': 0.010026508942246437, 'timing/model': 1.1450879449257627, 'learning_rate/qwen_vl_interface': 1.2103373428141903e-06, 'learning_rate/action_model': 1.2103373428141903e-05, 'epoch': 1.55}) +[2026-06-19 19:15:47,902][starVLA.training.train_starvla][INFO] - Step 4137, Loss: {'train/loss': 0.10898955911397934, 'train/grad_norm_pre_clip': 8.466843605041504, 'timing/data': 0.009683717973530293, 'timing/model': 1.1479256940074265, 'learning_rate/qwen_vl_interface': 1.2087361373903633e-06, 'learning_rate/action_model': 1.2087361373903633e-05, 'epoch': 1.55}) +[2026-06-19 19:15:52,513][starVLA.training.train_starvla][INFO] - Step 4138, Loss: {'train/loss': 0.1894829422235489, 'train/grad_norm_pre_clip': 20.71937370300293, 'timing/data': 0.009747191914357245, 'timing/model': 1.1517211199970916, 'learning_rate/qwen_vl_interface': 1.2071365931774584e-06, 'learning_rate/action_model': 1.2071365931774584e-05, 'epoch': 1.55}) +[2026-06-19 19:15:57,125][starVLA.training.train_starvla][INFO] - Step 4139, Loss: {'train/loss': 0.17146772146224976, 'train/grad_norm_pre_clip': 17.069387435913086, 'timing/data': 0.009656648966483772, 'timing/model': 1.3296239179326221, 'learning_rate/qwen_vl_interface': 1.2055387108329885e-06, 'learning_rate/action_model': 1.2055387108329886e-05, 'epoch': 1.55}) +[2026-06-19 19:16:01,664][starVLA.training.train_starvla][INFO] - Step 4140, Loss: {'train/loss': 0.14969705045223236, 'train/grad_norm_pre_clip': 12.15916633605957, 'timing/data': 0.009828723967075348, 'timing/model': 1.2611712811049074, 'learning_rate/qwen_vl_interface': 1.2039424910137807e-06, 'learning_rate/action_model': 1.2039424910137806e-05, 'epoch': 1.55}) +[2026-06-19 19:16:06,088][starVLA.training.train_starvla][INFO] - Step 4141, Loss: {'train/loss': 0.2003488540649414, 'train/grad_norm_pre_clip': 13.97014045715332, 'timing/data': 0.009912910987623036, 'timing/model': 1.144270769902505, 'learning_rate/qwen_vl_interface': 1.2023479343759822e-06, 'learning_rate/action_model': 1.2023479343759823e-05, 'epoch': 1.55}) +[2026-06-19 19:16:10,689][starVLA.training.train_starvla][INFO] - Step 4142, Loss: {'train/loss': 0.3939794898033142, 'train/grad_norm_pre_clip': 25.77107810974121, 'timing/data': 0.010020493064075708, 'timing/model': 1.149647093960084, 'learning_rate/qwen_vl_interface': 1.2007550415750546e-06, 'learning_rate/action_model': 1.2007550415750545e-05, 'epoch': 1.55}) +[2026-06-19 19:16:15,305][starVLA.training.train_starvla][INFO] - Step 4143, Loss: {'train/loss': 0.2706347703933716, 'train/grad_norm_pre_clip': 11.75124454498291, 'timing/data': 0.01007333304733038, 'timing/model': 1.149972865008749, 'learning_rate/qwen_vl_interface': 1.1991638132657759e-06, 'learning_rate/action_model': 1.199163813265776e-05, 'epoch': 1.55}) +[2026-06-19 19:16:19,909][starVLA.training.train_starvla][INFO] - Step 4144, Loss: {'train/loss': 0.1835889220237732, 'train/grad_norm_pre_clip': 12.353508949279785, 'timing/data': 0.00908564100973308, 'timing/model': 1.142194948042743, 'learning_rate/qwen_vl_interface': 1.1975742501022406e-06, 'learning_rate/action_model': 1.1975742501022405e-05, 'epoch': 1.55}) +[2026-06-19 19:16:24,516][starVLA.training.train_starvla][INFO] - Step 4145, Loss: {'train/loss': 0.44634997844696045, 'train/grad_norm_pre_clip': 24.50391960144043, 'timing/data': 0.009763463051058352, 'timing/model': 1.1468741539865732, 'learning_rate/qwen_vl_interface': 1.1959863527378569e-06, 'learning_rate/action_model': 1.1959863527378567e-05, 'epoch': 1.55}) +[2026-06-19 19:16:29,128][starVLA.training.train_starvla][INFO] - Step 4146, Loss: {'train/loss': 0.24749797582626343, 'train/grad_norm_pre_clip': 16.138086318969727, 'timing/data': 0.009720982983708382, 'timing/model': 1.3265581679297611, 'learning_rate/qwen_vl_interface': 1.1944001218253505e-06, 'learning_rate/action_model': 1.1944001218253503e-05, 'epoch': 1.55}) +[2026-06-19 19:16:33,560][starVLA.training.train_starvla][INFO] - Step 4147, Loss: {'train/loss': 0.2839181423187256, 'train/grad_norm_pre_clip': 12.826908111572266, 'timing/data': 0.009866232983767986, 'timing/model': 1.151869672932662, 'learning_rate/qwen_vl_interface': 1.1928155580167608e-06, 'learning_rate/action_model': 1.1928155580167608e-05, 'epoch': 1.55}) +[2026-06-19 19:16:38,158][starVLA.training.train_starvla][INFO] - Step 4148, Loss: {'train/loss': 0.344738632440567, 'train/grad_norm_pre_clip': 13.147636413574219, 'timing/data': 0.010029238066636026, 'timing/model': 1.1530406309757382, 'learning_rate/qwen_vl_interface': 1.1912326619634412e-06, 'learning_rate/action_model': 1.1912326619634412e-05, 'epoch': 1.56}) +[2026-06-19 19:16:42,766][starVLA.training.train_starvla][INFO] - Step 4149, Loss: {'train/loss': 0.12663742899894714, 'train/grad_norm_pre_clip': 13.791001319885254, 'timing/data': 0.010042945970781147, 'timing/model': 1.160303795011714, 'learning_rate/qwen_vl_interface': 1.1896514343160632e-06, 'learning_rate/action_model': 1.189651434316063e-05, 'epoch': 1.56}) +[2026-06-19 19:16:47,390][starVLA.training.train_starvla][INFO] - Step 4150, Loss: {'train/loss': 0.0639440193772316, 'train/grad_norm_pre_clip': 5.954745769500732, 'timing/data': 0.010079319938085973, 'timing/model': 1.1505167500581592, 'learning_rate/qwen_vl_interface': 1.1880718757246062e-06, 'learning_rate/action_model': 1.1880718757246063e-05, 'epoch': 1.56}) +[2026-06-19 19:16:51,978][starVLA.training.train_starvla][INFO] - Step 4151, Loss: {'train/loss': 0.27300140261650085, 'train/grad_norm_pre_clip': 22.55867576599121, 'timing/data': 0.009966320940293372, 'timing/model': 1.143143827910535, 'learning_rate/qwen_vl_interface': 1.186493986838369e-06, 'learning_rate/action_model': 1.186493986838369e-05, 'epoch': 1.56}) +[2026-06-19 19:16:56,565][starVLA.training.train_starvla][INFO] - Step 4152, Loss: {'train/loss': 0.16969315707683563, 'train/grad_norm_pre_clip': 11.49776840209961, 'timing/data': 0.009955878020264208, 'timing/model': 1.3124571630032733, 'learning_rate/qwen_vl_interface': 1.1849177683059612e-06, 'learning_rate/action_model': 1.1849177683059612e-05, 'epoch': 1.56}) +[2026-06-19 19:17:01,003][starVLA.training.train_starvla][INFO] - Step 4153, Loss: {'train/loss': 0.11269079893827438, 'train/grad_norm_pre_clip': 12.537577629089355, 'timing/data': 0.009355832007713616, 'timing/model': 1.1443375970702618, 'learning_rate/qwen_vl_interface': 1.1833432207753075e-06, 'learning_rate/action_model': 1.1833432207753076e-05, 'epoch': 1.56}) +[2026-06-19 19:17:05,624][starVLA.training.train_starvla][INFO] - Step 4154, Loss: {'train/loss': 0.12119067460298538, 'train/grad_norm_pre_clip': 12.823892593383789, 'timing/data': 0.010010412894189358, 'timing/model': 1.1607061999384314, 'learning_rate/qwen_vl_interface': 1.1817703448936448e-06, 'learning_rate/action_model': 1.181770344893645e-05, 'epoch': 1.56}) +[2026-06-19 19:17:10,151][starVLA.training.train_starvla][INFO] - Step 4155, Loss: {'train/loss': 0.2312040776014328, 'train/grad_norm_pre_clip': 18.954116821289062, 'timing/data': 0.009207388968206942, 'timing/model': 1.1645975809078664, 'learning_rate/qwen_vl_interface': 1.1801991413075228e-06, 'learning_rate/action_model': 1.1801991413075227e-05, 'epoch': 1.56}) +[2026-06-19 19:17:14,728][starVLA.training.train_starvla][INFO] - Step 4156, Loss: {'train/loss': 0.09343019127845764, 'train/grad_norm_pre_clip': 13.711846351623535, 'timing/data': 0.009427624987438321, 'timing/model': 1.1488071730127558, 'learning_rate/qwen_vl_interface': 1.1786296106628038e-06, 'learning_rate/action_model': 1.1786296106628036e-05, 'epoch': 1.56}) +[2026-06-19 19:17:19,345][starVLA.training.train_starvla][INFO] - Step 4157, Loss: {'train/loss': 0.20575465261936188, 'train/grad_norm_pre_clip': 20.519643783569336, 'timing/data': 0.010001351009123027, 'timing/model': 1.1476447660243139, 'learning_rate/qwen_vl_interface': 1.1770617536046633e-06, 'learning_rate/action_model': 1.1770617536046632e-05, 'epoch': 1.56}) +[2026-06-19 19:17:23,987][starVLA.training.train_starvla][INFO] - Step 4158, Loss: {'train/loss': 0.2662784457206726, 'train/grad_norm_pre_clip': 12.921472549438477, 'timing/data': 0.01017317792866379, 'timing/model': 1.3394558110740036, 'learning_rate/qwen_vl_interface': 1.1754955707775873e-06, 'learning_rate/action_model': 1.1754955707775872e-05, 'epoch': 1.56}) +[2026-06-19 19:17:28,451][starVLA.training.train_starvla][INFO] - Step 4159, Loss: {'train/loss': 0.29124873876571655, 'train/grad_norm_pre_clip': 17.41378402709961, 'timing/data': 0.009955775924026966, 'timing/model': 1.1559321150416508, 'learning_rate/qwen_vl_interface': 1.1739310628253753e-06, 'learning_rate/action_model': 1.1739310628253753e-05, 'epoch': 1.56}) +[2026-06-19 19:17:33,066][starVLA.training.train_starvla][INFO] - Step 4160, Loss: {'train/loss': 0.36387985944747925, 'train/grad_norm_pre_clip': 20.548494338989258, 'timing/data': 0.009725839016027749, 'timing/model': 1.154695627046749, 'learning_rate/qwen_vl_interface': 1.172368230391136e-06, 'learning_rate/action_model': 1.1723682303911359e-05, 'epoch': 1.56}) +[2026-06-19 19:17:37,672][starVLA.training.train_starvla][INFO] - Step 4161, Loss: {'train/loss': 0.22827033698558807, 'train/grad_norm_pre_clip': 13.236753463745117, 'timing/data': 0.009706942015327513, 'timing/model': 1.1508475560694933, 'learning_rate/qwen_vl_interface': 1.1708070741172928e-06, 'learning_rate/action_model': 1.1708070741172928e-05, 'epoch': 1.56}) +[2026-06-19 19:17:42,229][starVLA.training.train_starvla][INFO] - Step 4162, Loss: {'train/loss': 0.36804863810539246, 'train/grad_norm_pre_clip': 12.969629287719727, 'timing/data': 0.009775222977623343, 'timing/model': 1.1465721660060808, 'learning_rate/qwen_vl_interface': 1.1692475946455755e-06, 'learning_rate/action_model': 1.1692475946455755e-05, 'epoch': 1.56}) +[2026-06-19 19:17:46,862][starVLA.training.train_starvla][INFO] - Step 4163, Loss: {'train/loss': 0.22574520111083984, 'train/grad_norm_pre_clip': 12.991142272949219, 'timing/data': 0.009698827983811498, 'timing/model': 1.1478015079628676, 'learning_rate/qwen_vl_interface': 1.1676897926170289e-06, 'learning_rate/action_model': 1.1676897926170289e-05, 'epoch': 1.56}) +[2026-06-19 19:17:51,449][starVLA.training.train_starvla][INFO] - Step 4164, Loss: {'train/loss': 0.12822088599205017, 'train/grad_norm_pre_clip': 17.753097534179688, 'timing/data': 0.009480723994784057, 'timing/model': 1.3108013200107962, 'learning_rate/qwen_vl_interface': 1.166133668672006e-06, 'learning_rate/action_model': 1.166133668672006e-05, 'epoch': 1.56}) +[2026-06-19 19:17:55,887][starVLA.training.train_starvla][INFO] - Step 4165, Loss: {'train/loss': 0.22311195731163025, 'train/grad_norm_pre_clip': 13.415105819702148, 'timing/data': 0.010017682914622128, 'timing/model': 1.1554068060358986, 'learning_rate/qwen_vl_interface': 1.1645792234501706e-06, 'learning_rate/action_model': 1.1645792234501706e-05, 'epoch': 1.56}) +[2026-06-19 19:18:00,496][starVLA.training.train_starvla][INFO] - Step 4166, Loss: {'train/loss': 0.23712404072284698, 'train/grad_norm_pre_clip': 21.67524528503418, 'timing/data': 0.010111096897162497, 'timing/model': 1.1604696650756523, 'learning_rate/qwen_vl_interface': 1.1630264575904972e-06, 'learning_rate/action_model': 1.1630264575904973e-05, 'epoch': 1.56}) +[2026-06-19 19:18:05,986][starVLA.training.train_starvla][INFO] - Latency curriculum phase 5 at step 4167: active_latencies=[0, 1, 2, 3, 4, 5], dataset_size=341369 +[2026-06-19 19:18:05,987][starVLA.training.train_starvla][INFO] - Step 4167, Loss: {'train/loss': 0.15531279146671295, 'train/grad_norm_pre_clip': 12.835387229919434, 'latency_curriculum/phase': 5.0, 'latency_curriculum/active_count': 6.0, 'latency_curriculum/max_active_latency': 5.0, 'latency_curriculum/dataset_size': 341369.0, 'timing/data': 0.009906212915666401, 'timing/model': 1.1540743059013039, 'learning_rate/qwen_vl_interface': 1.1614753717312683e-06, 'learning_rate/action_model': 1.1614753717312681e-05, 'epoch': 1.56}) +[2026-06-19 19:18:23,068][starVLA.training.train_starvla][INFO] - Step 4168, Loss: {'train/loss': 0.20280015468597412, 'train/grad_norm_pre_clip': 13.214567184448242, 'timing/data': 0.01724553108215332, 'timing/model': 1.137618535081856, 'learning_rate/qwen_vl_interface': 1.1599259665100767e-06, 'learning_rate/action_model': 1.1599259665100767e-05, 'epoch': 1.56}) +[2026-06-19 19:18:27,652][starVLA.training.train_starvla][INFO] - Step 4169, Loss: {'train/loss': 0.17005720734596252, 'train/grad_norm_pre_clip': 20.394100189208984, 'timing/data': 0.018376361927948892, 'timing/model': 1.1454581110738218, 'learning_rate/qwen_vl_interface': 1.1583782425638252e-06, 'learning_rate/action_model': 1.158378242563825e-05, 'epoch': 1.56}) +[2026-06-19 19:18:32,220][starVLA.training.train_starvla][INFO] - Step 4170, Loss: {'train/loss': 0.2764105200767517, 'train/grad_norm_pre_clip': 19.6364688873291, 'timing/data': 0.017168225022032857, 'timing/model': 1.3150345280300826, 'learning_rate/qwen_vl_interface': 1.1568322005287237e-06, 'learning_rate/action_model': 1.1568322005287235e-05, 'epoch': 1.56}) +[2026-06-19 19:18:36,619][starVLA.training.train_starvla][INFO] - Step 4171, Loss: {'train/loss': 0.07021449506282806, 'train/grad_norm_pre_clip': 7.016870975494385, 'timing/data': 0.019165901001542807, 'timing/model': 1.141994550009258, 'learning_rate/qwen_vl_interface': 1.1552878410402923e-06, 'learning_rate/action_model': 1.1552878410402924e-05, 'epoch': 1.56}) +[2026-06-19 19:18:41,191][starVLA.training.train_starvla][INFO] - Step 4172, Loss: {'train/loss': 0.08192577213048935, 'train/grad_norm_pre_clip': 7.159196376800537, 'timing/data': 0.016061397036537528, 'timing/model': 1.1451320379273966, 'learning_rate/qwen_vl_interface': 1.1537451647333594e-06, 'learning_rate/action_model': 1.1537451647333593e-05, 'epoch': 1.56}) +[2026-06-19 19:18:45,771][starVLA.training.train_starvla][INFO] - Step 4173, Loss: {'train/loss': 0.3312428891658783, 'train/grad_norm_pre_clip': 15.565227508544922, 'timing/data': 0.01706105994526297, 'timing/model': 1.1453167430590838, 'learning_rate/qwen_vl_interface': 1.1522041722420594e-06, 'learning_rate/action_model': 1.1522041722420594e-05, 'epoch': 1.56}) +[2026-06-19 19:18:50,242][starVLA.training.train_starvla][INFO] - Step 4174, Loss: {'train/loss': 0.29738813638687134, 'train/grad_norm_pre_clip': 24.966890335083008, 'timing/data': 0.017298191087320447, 'timing/model': 1.1434631240554154, 'learning_rate/qwen_vl_interface': 1.1506648641998365e-06, 'learning_rate/action_model': 1.1506648641998365e-05, 'epoch': 1.57}) +[2026-06-19 19:18:54,805][starVLA.training.train_starvla][INFO] - Step 4175, Loss: {'train/loss': 0.17092514038085938, 'train/grad_norm_pre_clip': 11.706826210021973, 'timing/data': 0.017859008978120983, 'timing/model': 1.1502456860616803, 'learning_rate/qwen_vl_interface': 1.1491272412394433e-06, 'learning_rate/action_model': 1.1491272412394433e-05, 'epoch': 1.57}) +[2026-06-19 19:18:59,404][starVLA.training.train_starvla][INFO] - Step 4176, Loss: {'train/loss': 0.2640831172466278, 'train/grad_norm_pre_clip': 12.21029281616211, 'timing/data': 0.01844984304625541, 'timing/model': 1.3224630610784516, 'learning_rate/qwen_vl_interface': 1.1475913039929373e-06, 'learning_rate/action_model': 1.1475913039929372e-05, 'epoch': 1.57}) +[2026-06-19 19:19:03,846][starVLA.training.train_starvla][INFO] - Step 4177, Loss: {'train/loss': 0.17098134756088257, 'train/grad_norm_pre_clip': 9.711458206176758, 'timing/data': 0.01770551106892526, 'timing/model': 1.1541552399285138, 'learning_rate/qwen_vl_interface': 1.1460570530916844e-06, 'learning_rate/action_model': 1.1460570530916844e-05, 'epoch': 1.57}) +[2026-06-19 19:19:08,443][starVLA.training.train_starvla][INFO] - Step 4178, Loss: {'train/loss': 0.1954054832458496, 'train/grad_norm_pre_clip': 15.95805549621582, 'timing/data': 0.018097321968525648, 'timing/model': 1.1519338870421052, 'learning_rate/qwen_vl_interface': 1.1445244891663581e-06, 'learning_rate/action_model': 1.1445244891663581e-05, 'epoch': 1.57}) +[2026-06-19 19:19:13,140][starVLA.training.train_starvla][INFO] - Step 4179, Loss: {'train/loss': 0.11560871452093124, 'train/grad_norm_pre_clip': 10.855276107788086, 'timing/data': 0.02272569900378585, 'timing/model': 1.1603225020226091, 'learning_rate/qwen_vl_interface': 1.1429936128469366e-06, 'learning_rate/action_model': 1.1429936128469366e-05, 'epoch': 1.57}) +[2026-06-19 19:19:17,782][starVLA.training.train_starvla][INFO] - Step 4180, Loss: {'train/loss': 0.23475505411624908, 'train/grad_norm_pre_clip': 21.319028854370117, 'timing/data': 0.02100504597183317, 'timing/model': 1.1612545000389218, 'learning_rate/qwen_vl_interface': 1.1414644247627056e-06, 'learning_rate/action_model': 1.1414644247627058e-05, 'epoch': 1.57}) +[2026-06-19 19:19:22,401][starVLA.training.train_starvla][INFO] - Step 4181, Loss: {'train/loss': 0.11498933285474777, 'train/grad_norm_pre_clip': 8.648794174194336, 'timing/data': 0.016808714950457215, 'timing/model': 1.325337283895351, 'learning_rate/qwen_vl_interface': 1.1399369255422568e-06, 'learning_rate/action_model': 1.1399369255422568e-05, 'epoch': 1.57}) +[2026-06-19 19:19:26,847][starVLA.training.train_starvla][INFO] - Step 4182, Loss: {'train/loss': 0.12082093209028244, 'train/grad_norm_pre_clip': 14.194767951965332, 'timing/data': 0.01801789505407214, 'timing/model': 1.154660871019587, 'learning_rate/qwen_vl_interface': 1.138411115813488e-06, 'learning_rate/action_model': 1.138411115813488e-05, 'epoch': 1.57}) +[2026-06-19 19:19:31,467][starVLA.training.train_starvla][INFO] - Step 4183, Loss: {'train/loss': 0.30239206552505493, 'train/grad_norm_pre_clip': 14.676318168640137, 'timing/data': 0.016196709009818733, 'timing/model': 1.1527149539906532, 'learning_rate/qwen_vl_interface': 1.1368869962036006e-06, 'learning_rate/action_model': 1.1368869962036005e-05, 'epoch': 1.57}) +[2026-06-19 19:19:36,046][starVLA.training.train_starvla][INFO] - Step 4184, Loss: {'train/loss': 0.27592459321022034, 'train/grad_norm_pre_clip': 14.599982261657715, 'timing/data': 0.017578445025719702, 'timing/model': 1.1483995369635522, 'learning_rate/qwen_vl_interface': 1.1353645673391025e-06, 'learning_rate/action_model': 1.1353645673391026e-05, 'epoch': 1.57}) +[2026-06-19 19:19:40,664][starVLA.training.train_starvla][INFO] - Step 4185, Loss: {'train/loss': 0.09939853101968765, 'train/grad_norm_pre_clip': 13.294644355773926, 'timing/data': 0.016846560058183968, 'timing/model': 1.1458205099916086, 'learning_rate/qwen_vl_interface': 1.1338438298458078e-06, 'learning_rate/action_model': 1.1338438298458078e-05, 'epoch': 1.57}) +[2026-06-19 19:19:45,285][starVLA.training.train_starvla][INFO] - Step 4186, Loss: {'train/loss': 0.10580141097307205, 'train/grad_norm_pre_clip': 12.868572235107422, 'timing/data': 0.01748693105764687, 'timing/model': 1.1530240640277043, 'learning_rate/qwen_vl_interface': 1.1323247843488345e-06, 'learning_rate/action_model': 1.1323247843488344e-05, 'epoch': 1.57}) +[2026-06-19 19:19:49,907][starVLA.training.train_starvla][INFO] - Step 4187, Loss: {'train/loss': 0.23111876845359802, 'train/grad_norm_pre_clip': 12.207746505737305, 'timing/data': 0.01687000400852412, 'timing/model': 1.3261769840028137, 'learning_rate/qwen_vl_interface': 1.130807431472603e-06, 'learning_rate/action_model': 1.130807431472603e-05, 'epoch': 1.57}) +[2026-06-19 19:19:54,521][starVLA.training.train_starvla][INFO] - Step 4188, Loss: {'train/loss': 0.11641894280910492, 'train/grad_norm_pre_clip': 7.9640607833862305, 'timing/data': 0.016940169967710972, 'timing/model': 1.3262890480691567, 'learning_rate/qwen_vl_interface': 1.1292917718408416e-06, 'learning_rate/action_model': 1.1292917718408416e-05, 'epoch': 1.57}) +[2026-06-19 19:19:58,958][starVLA.training.train_starvla][INFO] - Step 4189, Loss: {'train/loss': 0.21914200484752655, 'train/grad_norm_pre_clip': 12.603155136108398, 'timing/data': 0.01704027794767171, 'timing/model': 1.1499745809705928, 'learning_rate/qwen_vl_interface': 1.12777780607658e-06, 'learning_rate/action_model': 1.12777780607658e-05, 'epoch': 1.57}) +[2026-06-19 19:20:03,584][starVLA.training.train_starvla][INFO] - Step 4190, Loss: {'train/loss': 0.1423352211713791, 'train/grad_norm_pre_clip': 10.297834396362305, 'timing/data': 0.017094049020670354, 'timing/model': 1.157003600965254, 'learning_rate/qwen_vl_interface': 1.1262655348021523e-06, 'learning_rate/action_model': 1.1262655348021523e-05, 'epoch': 1.57}) +[2026-06-19 19:20:08,169][starVLA.training.train_starvla][INFO] - Step 4191, Loss: {'train/loss': 0.21195527911186218, 'train/grad_norm_pre_clip': 21.90856170654297, 'timing/data': 0.018149698968045413, 'timing/model': 1.1456347879720852, 'learning_rate/qwen_vl_interface': 1.1247549586391966e-06, 'learning_rate/action_model': 1.1247549586391964e-05, 'epoch': 1.57}) +[2026-06-19 19:20:12,673][starVLA.training.train_starvla][INFO] - Step 4192, Loss: {'train/loss': 0.16865618526935577, 'train/grad_norm_pre_clip': 16.14901351928711, 'timing/data': 0.018226161948405206, 'timing/model': 1.1676287200534716, 'learning_rate/qwen_vl_interface': 1.1232460782086534e-06, 'learning_rate/action_model': 1.1232460782086534e-05, 'epoch': 1.57}) +[2026-06-19 19:20:17,267][starVLA.training.train_starvla][INFO] - Step 4193, Loss: {'train/loss': 0.2701486051082611, 'train/grad_norm_pre_clip': 25.061756134033203, 'timing/data': 0.016684426926076412, 'timing/model': 1.1636383320437744, 'learning_rate/qwen_vl_interface': 1.1217388941307658e-06, 'learning_rate/action_model': 1.1217388941307658e-05, 'epoch': 1.57}) +[2026-06-19 19:20:21,878][starVLA.training.train_starvla][INFO] - Step 4194, Loss: {'train/loss': 0.15429934859275818, 'train/grad_norm_pre_clip': 13.583976745605469, 'timing/data': 0.018229992012493312, 'timing/model': 1.3255642730509862, 'learning_rate/qwen_vl_interface': 1.1202334070250807e-06, 'learning_rate/action_model': 1.1202334070250806e-05, 'epoch': 1.57}) +[2026-06-19 19:20:26,302][starVLA.training.train_starvla][INFO] - Step 4195, Loss: {'train/loss': 0.29894697666168213, 'train/grad_norm_pre_clip': 21.081592559814453, 'timing/data': 0.017329921014606953, 'timing/model': 1.1431393069215119, 'learning_rate/qwen_vl_interface': 1.1187296175104472e-06, 'learning_rate/action_model': 1.1187296175104471e-05, 'epoch': 1.57}) +[2026-06-19 19:20:30,914][starVLA.training.train_starvla][INFO] - Step 4196, Loss: {'train/loss': 0.10632138699293137, 'train/grad_norm_pre_clip': 10.114245414733887, 'timing/data': 0.01774827705230564, 'timing/model': 1.1543549950001761, 'learning_rate/qwen_vl_interface': 1.1172275262050158e-06, 'learning_rate/action_model': 1.1172275262050159e-05, 'epoch': 1.57}) +[2026-06-19 19:20:35,522][starVLA.training.train_starvla][INFO] - Step 4197, Loss: {'train/loss': 0.1738622933626175, 'train/grad_norm_pre_clip': 16.980287551879883, 'timing/data': 0.016284358920529485, 'timing/model': 1.148713579052128, 'learning_rate/qwen_vl_interface': 1.1157271337262401e-06, 'learning_rate/action_model': 1.11572713372624e-05, 'epoch': 1.57}) +[2026-06-19 19:20:40,141][starVLA.training.train_starvla][INFO] - Step 4198, Loss: {'train/loss': 0.15735411643981934, 'train/grad_norm_pre_clip': 9.903600692749023, 'timing/data': 0.017580048996023834, 'timing/model': 1.1466430969303474, 'learning_rate/qwen_vl_interface': 1.1142284406908751e-06, 'learning_rate/action_model': 1.114228440690875e-05, 'epoch': 1.57}) +[2026-06-19 19:20:44,671][starVLA.training.train_starvla][INFO] - Step 4199, Loss: {'train/loss': 0.19064494967460632, 'train/grad_norm_pre_clip': 18.34522819519043, 'timing/data': 0.01827009895350784, 'timing/model': 1.1510909920325503, 'learning_rate/qwen_vl_interface': 1.1127314477149758e-06, 'learning_rate/action_model': 1.1127314477149759e-05, 'epoch': 1.57}) +[2026-06-19 19:20:49,245][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 4200 +[2026-06-19 19:23:43,994][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 4200 +[2026-06-19 19:23:43,995][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 4200 +[2026-06-19 19:26:48,896][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 4200 +[2026-06-19 19:26:48,897][starVLA.training.train_starvla][INFO] - Step 4200, Loss: {'train/loss': 0.21192355453968048, 'train/grad_norm_pre_clip': 12.962498664855957, 'eval/loss': 0.13130345940589905, 'eval/latency_0/loss': 0.08158398419618607, 'eval/latency_1/loss': 0.04883579909801483, 'eval/latency_2/loss': 0.11173415929079056, 'eval/latency_3/loss': 0.15644505620002747, 'eval/latency_4/loss': 0.25885435938835144, 'eval/flappy/loss': 0.13130345940589905, 'eval/flappy/latency_0/loss': 0.08158398419618607, 'eval/flappy/latency_1/loss': 0.04883579909801483, 'eval/flappy/latency_2/loss': 0.11173415929079056, 'eval/flappy/latency_3/loss': 0.15644505620002747, 'eval/flappy/latency_4/loss': 0.25885435938835144, 'eval/flap_precision': 0.7030716723549488, 'eval/flap_recall': 0.7490909090909091, 'eval/flap_f1': 0.7253521126760564, 'eval/flap_rate_model': 0.08138888888888889, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.7253521126760565, 'eval/latency_0/flap_precision': 0.7030716723549488, 'eval/latency_0/flap_recall': 0.7490909090909091, 'eval/latency_0/flap_f1': 0.7253521126760564, 'eval/latency_0/flap_rate_model': 0.08138888888888889, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.7253521126760565, 'timing/data': 0.018125004949979484, 'timing/model': 1.2866988399764523, 'learning_rate/qwen_vl_interface': 1.1112361554139003e-06, 'learning_rate/action_model': 1.1112361554139004e-05, 'epoch': 1.57}) +[2026-06-19 19:26:48,897][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_4200_state +[2026-06-19 19:26:48,900][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-19 19:26:52,269][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_4200_state/model.safetensors +[2026-06-19 19:26:59,491][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_4200_state/optimizer.bin +[2026-06-19 19:26:59,492][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_4200_state/sampler.bin +[2026-06-19 19:26:59,492][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_4200_state/sampler_1.bin +[2026-06-19 19:26:59,493][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_4200_state/random_states_0.pkl +[2026-06-19 19:26:59,517][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4200 +[2026-06-19 19:26:59,517][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-19 19:26:59,528][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-19 19:27:04,088][starVLA.training.train_starvla][INFO] - Step 4201, Loss: {'train/loss': 0.23999784886837006, 'train/grad_norm_pre_clip': 15.77029800415039, 'timing/data': 0.017666762927547097, 'timing/model': 1.2967554179485887, 'learning_rate/qwen_vl_interface': 1.1097425644023065e-06, 'learning_rate/action_model': 1.1097425644023065e-05, 'epoch': 1.58}) +[2026-06-19 19:27:08,445][starVLA.training.train_starvla][INFO] - Step 4202, Loss: {'train/loss': 0.2173839509487152, 'train/grad_norm_pre_clip': 64.47758483886719, 'timing/data': 0.017196384957060218, 'timing/model': 1.1297360679600388, 'learning_rate/qwen_vl_interface': 1.1082506752941529e-06, 'learning_rate/action_model': 1.1082506752941528e-05, 'epoch': 1.58}) +[2026-06-19 19:27:12,964][starVLA.training.train_starvla][INFO] - Step 4203, Loss: {'train/loss': 0.21461345255374908, 'train/grad_norm_pre_clip': 26.427831649780273, 'timing/data': 0.019241296919062734, 'timing/model': 1.14165490900632, 'learning_rate/qwen_vl_interface': 1.1067604887027004e-06, 'learning_rate/action_model': 1.1067604887027004e-05, 'epoch': 1.58}) +[2026-06-19 19:27:17,465][starVLA.training.train_starvla][INFO] - Step 4204, Loss: {'train/loss': 0.08482697606086731, 'train/grad_norm_pre_clip': 10.629608154296875, 'timing/data': 0.01822112000081688, 'timing/model': 1.130083772004582, 'learning_rate/qwen_vl_interface': 1.1052720052405066e-06, 'learning_rate/action_model': 1.1052720052405064e-05, 'epoch': 1.58}) +[2026-06-19 19:27:22,032][starVLA.training.train_starvla][INFO] - Step 4205, Loss: {'train/loss': 0.2611420154571533, 'train/grad_norm_pre_clip': 17.018999099731445, 'timing/data': 0.017821774003095925, 'timing/model': 1.137890815967694, 'learning_rate/qwen_vl_interface': 1.1037852255194309e-06, 'learning_rate/action_model': 1.1037852255194308e-05, 'epoch': 1.58}) +[2026-06-19 19:27:26,606][starVLA.training.train_starvla][INFO] - Step 4206, Loss: {'train/loss': 0.37606316804885864, 'train/grad_norm_pre_clip': 22.438547134399414, 'timing/data': 0.016709832940250635, 'timing/model': 1.1412540359888226, 'learning_rate/qwen_vl_interface': 1.1023001501506329e-06, 'learning_rate/action_model': 1.102300150150633e-05, 'epoch': 1.58}) +[2026-06-19 19:27:31,198][starVLA.training.train_starvla][INFO] - Step 4207, Loss: {'train/loss': 0.31545886397361755, 'train/grad_norm_pre_clip': 40.056766510009766, 'timing/data': 0.01753137295600027, 'timing/model': 1.3274298569886014, 'learning_rate/qwen_vl_interface': 1.1008167797445717e-06, 'learning_rate/action_model': 1.1008167797445718e-05, 'epoch': 1.58}) +[2026-06-19 19:27:35,579][starVLA.training.train_starvla][INFO] - Step 4208, Loss: {'train/loss': 0.25559067726135254, 'train/grad_norm_pre_clip': 13.229533195495605, 'timing/data': 0.017140470095910132, 'timing/model': 1.1350634859409183, 'learning_rate/qwen_vl_interface': 1.099335114911003e-06, 'learning_rate/action_model': 1.0993351149110031e-05, 'epoch': 1.58}) +[2026-06-19 19:27:40,142][starVLA.training.train_starvla][INFO] - Step 4209, Loss: {'train/loss': 0.054698001593351364, 'train/grad_norm_pre_clip': 6.0445685386657715, 'timing/data': 0.017870665062218904, 'timing/model': 1.1377296320861205, 'learning_rate/qwen_vl_interface': 1.0978551562589846e-06, 'learning_rate/action_model': 1.0978551562589845e-05, 'epoch': 1.58}) +[2026-06-19 19:27:44,688][starVLA.training.train_starvla][INFO] - Step 4210, Loss: {'train/loss': 0.199259415268898, 'train/grad_norm_pre_clip': 12.364702224731445, 'timing/data': 0.017361130914650857, 'timing/model': 1.1389348090160638, 'learning_rate/qwen_vl_interface': 1.0963769043968708e-06, 'learning_rate/action_model': 1.0963769043968708e-05, 'epoch': 1.58}) +[2026-06-19 19:27:49,241][starVLA.training.train_starvla][INFO] - Step 4211, Loss: {'train/loss': 0.10269409418106079, 'train/grad_norm_pre_clip': 8.568338394165039, 'timing/data': 0.01716187398415059, 'timing/model': 1.1364276519743726, 'learning_rate/qwen_vl_interface': 1.0949003599323158e-06, 'learning_rate/action_model': 1.0949003599323157e-05, 'epoch': 1.58}) +[2026-06-19 19:27:53,808][starVLA.training.train_starvla][INFO] - Step 4212, Loss: {'train/loss': 0.0892462208867073, 'train/grad_norm_pre_clip': 10.393345832824707, 'timing/data': 0.017447613063268363, 'timing/model': 1.1399942819261923, 'learning_rate/qwen_vl_interface': 1.0934255234722708e-06, 'learning_rate/action_model': 1.0934255234722706e-05, 'epoch': 1.58}) +[2026-06-19 19:27:58,377][starVLA.training.train_starvla][INFO] - Step 4213, Loss: {'train/loss': 0.20748305320739746, 'train/grad_norm_pre_clip': 12.471450805664062, 'timing/data': 0.01757512998301536, 'timing/model': 1.305233674007468, 'learning_rate/qwen_vl_interface': 1.0919523956229858e-06, 'learning_rate/action_model': 1.0919523956229857e-05, 'epoch': 1.58}) +[2026-06-19 19:28:03,003][starVLA.training.train_starvla][INFO] - Step 4214, Loss: {'train/loss': 0.1120312288403511, 'train/grad_norm_pre_clip': 11.82795238494873, 'timing/data': 0.017589570954442024, 'timing/model': 1.3217968359822407, 'learning_rate/qwen_vl_interface': 1.0904809769900075e-06, 'learning_rate/action_model': 1.0904809769900073e-05, 'epoch': 1.58}) +[2026-06-19 19:28:07,398][starVLA.training.train_starvla][INFO] - Step 4215, Loss: {'train/loss': 0.07590220868587494, 'train/grad_norm_pre_clip': 6.487785816192627, 'timing/data': 0.0174139270093292, 'timing/model': 1.138670691056177, 'learning_rate/qwen_vl_interface': 1.0890112681781821e-06, 'learning_rate/action_model': 1.089011268178182e-05, 'epoch': 1.58}) +[2026-06-19 19:28:11,957][starVLA.training.train_starvla][INFO] - Step 4216, Loss: {'train/loss': 0.18757763504981995, 'train/grad_norm_pre_clip': 14.704740524291992, 'timing/data': 0.017378245946019888, 'timing/model': 1.141262739081867, 'learning_rate/qwen_vl_interface': 1.0875432697916492e-06, 'learning_rate/action_model': 1.087543269791649e-05, 'epoch': 1.58}) +[2026-06-19 19:28:16,518][starVLA.training.train_starvla][INFO] - Step 4217, Loss: {'train/loss': 0.15143251419067383, 'train/grad_norm_pre_clip': 17.63203239440918, 'timing/data': 0.017885323031805456, 'timing/model': 1.140033719013445, 'learning_rate/qwen_vl_interface': 1.0860769824338477e-06, 'learning_rate/action_model': 1.0860769824338478e-05, 'epoch': 1.58}) +[2026-06-19 19:28:20,992][starVLA.training.train_starvla][INFO] - Step 4218, Loss: {'train/loss': 0.10215447843074799, 'train/grad_norm_pre_clip': 8.40809154510498, 'timing/data': 0.018196089076809585, 'timing/model': 1.1468893089331686, 'learning_rate/qwen_vl_interface': 1.084612406707514e-06, 'learning_rate/action_model': 1.084612406707514e-05, 'epoch': 1.58}) +[2026-06-19 19:28:25,528][starVLA.training.train_starvla][INFO] - Step 4219, Loss: {'train/loss': 0.16851146519184113, 'train/grad_norm_pre_clip': 21.265872955322266, 'timing/data': 0.017475474975071847, 'timing/model': 1.1437617980409414, 'learning_rate/qwen_vl_interface': 1.0831495432146795e-06, 'learning_rate/action_model': 1.0831495432146795e-05, 'epoch': 1.58}) +[2026-06-19 19:28:30,144][starVLA.training.train_starvla][INFO] - Step 4220, Loss: {'train/loss': 0.34244370460510254, 'train/grad_norm_pre_clip': 19.220293045043945, 'timing/data': 0.016964526032097638, 'timing/model': 1.3391843929421157, 'learning_rate/qwen_vl_interface': 1.0816883925566717e-06, 'learning_rate/action_model': 1.0816883925566716e-05, 'epoch': 1.58}) +[2026-06-19 19:28:34,568][starVLA.training.train_starvla][INFO] - Step 4221, Loss: {'train/loss': 0.14116334915161133, 'train/grad_norm_pre_clip': 13.64448070526123, 'timing/data': 0.01754623802844435, 'timing/model': 1.142092310008593, 'learning_rate/qwen_vl_interface': 1.0802289553341153e-06, 'learning_rate/action_model': 1.0802289553341152e-05, 'epoch': 1.58}) +[2026-06-19 19:28:39,126][starVLA.training.train_starvla][INFO] - Step 4222, Loss: {'train/loss': 0.12396860867738724, 'train/grad_norm_pre_clip': 7.557216167449951, 'timing/data': 0.017373952083289623, 'timing/model': 1.1444403399946168, 'learning_rate/qwen_vl_interface': 1.0787712321469289e-06, 'learning_rate/action_model': 1.0787712321469288e-05, 'epoch': 1.58}) +[2026-06-19 19:28:43,702][starVLA.training.train_starvla][INFO] - Step 4223, Loss: {'train/loss': 0.24514739215373993, 'train/grad_norm_pre_clip': 18.91678237915039, 'timing/data': 0.016730064991861582, 'timing/model': 1.1456234569195658, 'learning_rate/qwen_vl_interface': 1.077315223594328e-06, 'learning_rate/action_model': 1.077315223594328e-05, 'epoch': 1.58}) +[2026-06-19 19:28:48,278][starVLA.training.train_starvla][INFO] - Step 4224, Loss: {'train/loss': 0.20719154179096222, 'train/grad_norm_pre_clip': 13.603771209716797, 'timing/data': 0.01753476506564766, 'timing/model': 1.1453617688966915, 'learning_rate/qwen_vl_interface': 1.0758609302748227e-06, 'learning_rate/action_model': 1.0758609302748227e-05, 'epoch': 1.58}) +[2026-06-19 19:28:52,770][starVLA.training.train_starvla][INFO] - Step 4225, Loss: {'train/loss': 0.14397087693214417, 'train/grad_norm_pre_clip': 13.40041732788086, 'timing/data': 0.017490137950517237, 'timing/model': 1.1432301439344883, 'learning_rate/qwen_vl_interface': 1.0744083527862182e-06, 'learning_rate/action_model': 1.074408352786218e-05, 'epoch': 1.58}) +[2026-06-19 19:28:57,317][starVLA.training.train_starvla][INFO] - Step 4226, Loss: {'train/loss': 0.33520856499671936, 'train/grad_norm_pre_clip': 14.367390632629395, 'timing/data': 0.01760818506591022, 'timing/model': 1.2684394749812782, 'learning_rate/qwen_vl_interface': 1.0729574917256144e-06, 'learning_rate/action_model': 1.0729574917256145e-05, 'epoch': 1.58}) +[2026-06-19 19:29:01,758][starVLA.training.train_starvla][INFO] - Step 4227, Loss: {'train/loss': 0.1646435409784317, 'train/grad_norm_pre_clip': 16.027563095092773, 'timing/data': 0.01744783006142825, 'timing/model': 1.1582263089949265, 'learning_rate/qwen_vl_interface': 1.0715083476894047e-06, 'learning_rate/action_model': 1.0715083476894047e-05, 'epoch': 1.58}) +[2026-06-19 19:29:06,366][starVLA.training.train_starvla][INFO] - Step 4228, Loss: {'train/loss': 0.0559275783598423, 'train/grad_norm_pre_clip': 7.59532356262207, 'timing/data': 0.017193002975545824, 'timing/model': 1.14458665298298, 'learning_rate/qwen_vl_interface': 1.0700609212732773e-06, 'learning_rate/action_model': 1.0700609212732774e-05, 'epoch': 1.59}) +[2026-06-19 19:29:11,005][starVLA.training.train_starvla][INFO] - Step 4229, Loss: {'train/loss': 0.3054342567920685, 'train/grad_norm_pre_clip': 19.353580474853516, 'timing/data': 0.016891888924874365, 'timing/model': 1.164276554947719, 'learning_rate/qwen_vl_interface': 1.0686152130722155e-06, 'learning_rate/action_model': 1.0686152130722154e-05, 'epoch': 1.59}) +[2026-06-19 19:29:15,604][starVLA.training.train_starvla][INFO] - Step 4230, Loss: {'train/loss': 0.1904771625995636, 'train/grad_norm_pre_clip': 21.400283813476562, 'timing/data': 0.017682370031252503, 'timing/model': 1.15221238497179, 'learning_rate/qwen_vl_interface': 1.0671712236804947e-06, 'learning_rate/action_model': 1.0671712236804946e-05, 'epoch': 1.59}) +[2026-06-19 19:29:20,196][starVLA.training.train_starvla][INFO] - Step 4231, Loss: {'train/loss': 0.2181796133518219, 'train/grad_norm_pre_clip': 19.073833465576172, 'timing/data': 0.016990846022963524, 'timing/model': 1.322949243010953, 'learning_rate/qwen_vl_interface': 1.0657289536916845e-06, 'learning_rate/action_model': 1.0657289536916845e-05, 'epoch': 1.59}) +[2026-06-19 19:29:24,707][starVLA.training.train_starvla][INFO] - Step 4232, Loss: {'train/loss': 0.15134304761886597, 'train/grad_norm_pre_clip': 7.6476569175720215, 'timing/data': 0.018680373090319335, 'timing/model': 1.217107928940095, 'learning_rate/qwen_vl_interface': 1.0642884036986481e-06, 'learning_rate/action_model': 1.064288403698648e-05, 'epoch': 1.59}) +[2026-06-19 19:29:29,131][starVLA.training.train_starvla][INFO] - Step 4233, Loss: {'train/loss': 0.12093176692724228, 'train/grad_norm_pre_clip': 5.166871547698975, 'timing/data': 0.0183740759966895, 'timing/model': 1.1468231070321053, 'learning_rate/qwen_vl_interface': 1.0628495742935404e-06, 'learning_rate/action_model': 1.0628495742935405e-05, 'epoch': 1.59}) +[2026-06-19 19:29:33,676][starVLA.training.train_starvla][INFO] - Step 4234, Loss: {'train/loss': 0.14292168617248535, 'train/grad_norm_pre_clip': 9.756271362304688, 'timing/data': 0.018071145052090287, 'timing/model': 1.1537582060554996, 'learning_rate/qwen_vl_interface': 1.0614124660678103e-06, 'learning_rate/action_model': 1.0614124660678102e-05, 'epoch': 1.59}) +[2026-06-19 19:29:38,269][starVLA.training.train_starvla][INFO] - Step 4235, Loss: {'train/loss': 0.15225054323673248, 'train/grad_norm_pre_clip': 10.228874206542969, 'timing/data': 0.02127561392262578, 'timing/model': 1.1498217380139977, 'learning_rate/qwen_vl_interface': 1.0599770796121988e-06, 'learning_rate/action_model': 1.0599770796121987e-05, 'epoch': 1.59}) +[2026-06-19 19:29:42,866][starVLA.training.train_starvla][INFO] - Step 4236, Loss: {'train/loss': 0.19020026922225952, 'train/grad_norm_pre_clip': 15.01335334777832, 'timing/data': 0.01805660198442638, 'timing/model': 1.147103854920715, 'learning_rate/qwen_vl_interface': 1.0585434155167394e-06, 'learning_rate/action_model': 1.0585434155167393e-05, 'epoch': 1.59}) +[2026-06-19 19:29:47,470][starVLA.training.train_starvla][INFO] - Step 4237, Loss: {'train/loss': 0.07833385467529297, 'train/grad_norm_pre_clip': 9.102252960205078, 'timing/data': 0.019136597984470427, 'timing/model': 1.3218325240304694, 'learning_rate/qwen_vl_interface': 1.0571114743707556e-06, 'learning_rate/action_model': 1.0571114743707555e-05, 'epoch': 1.59}) +[2026-06-19 19:29:52,057][starVLA.training.train_starvla][INFO] - Step 4238, Loss: {'train/loss': 0.09349499642848969, 'train/grad_norm_pre_clip': 9.563501358032227, 'timing/data': 0.016891743056476116, 'timing/model': 1.3094155099242926, 'learning_rate/qwen_vl_interface': 1.0556812567628661e-06, 'learning_rate/action_model': 1.055681256762866e-05, 'epoch': 1.59}) +[2026-06-19 19:29:56,485][starVLA.training.train_starvla][INFO] - Step 4239, Loss: {'train/loss': 0.17421726882457733, 'train/grad_norm_pre_clip': 8.268403053283691, 'timing/data': 0.017820221022702754, 'timing/model': 1.1473716460168362, 'learning_rate/qwen_vl_interface': 1.0542527632809783e-06, 'learning_rate/action_model': 1.0542527632809784e-05, 'epoch': 1.59}) +[2026-06-19 19:30:01,086][starVLA.training.train_starvla][INFO] - Step 4240, Loss: {'train/loss': 0.06822048127651215, 'train/grad_norm_pre_clip': 5.201324462890625, 'timing/data': 0.017508794902823865, 'timing/model': 1.1547223619418219, 'learning_rate/qwen_vl_interface': 1.0528259945122925e-06, 'learning_rate/action_model': 1.0528259945122925e-05, 'epoch': 1.59}) +[2026-06-19 19:30:05,691][starVLA.training.train_starvla][INFO] - Step 4241, Loss: {'train/loss': 0.14701420068740845, 'train/grad_norm_pre_clip': 22.940893173217773, 'timing/data': 0.01770304806996137, 'timing/model': 1.1464267359115183, 'learning_rate/qwen_vl_interface': 1.0514009510432988e-06, 'learning_rate/action_model': 1.0514009510432988e-05, 'epoch': 1.59}) +[2026-06-19 19:30:10,182][starVLA.training.train_starvla][INFO] - Step 4242, Loss: {'train/loss': 0.16816607117652893, 'train/grad_norm_pre_clip': 16.78184700012207, 'timing/data': 0.017055601929314435, 'timing/model': 1.1463925619609654, 'learning_rate/qwen_vl_interface': 1.0499776334597796e-06, 'learning_rate/action_model': 1.0499776334597797e-05, 'epoch': 1.59}) +[2026-06-19 19:30:14,718][starVLA.training.train_starvla][INFO] - Step 4243, Loss: {'train/loss': 0.07136446982622147, 'train/grad_norm_pre_clip': 8.984618186950684, 'timing/data': 0.018299961928278208, 'timing/model': 1.1477239740779623, 'learning_rate/qwen_vl_interface': 1.0485560423468064e-06, 'learning_rate/action_model': 1.0485560423468063e-05, 'epoch': 1.59}) +[2026-06-19 19:30:19,313][starVLA.training.train_starvla][INFO] - Step 4244, Loss: {'train/loss': 0.1504235714673996, 'train/grad_norm_pre_clip': 9.693009376525879, 'timing/data': 0.01727311103604734, 'timing/model': 1.3173320249188691, 'learning_rate/qwen_vl_interface': 1.0471361782887421e-06, 'learning_rate/action_model': 1.0471361782887421e-05, 'epoch': 1.59}) +[2026-06-19 19:30:23,740][starVLA.training.train_starvla][INFO] - Step 4245, Loss: {'train/loss': 0.3225892186164856, 'train/grad_norm_pre_clip': 26.13463020324707, 'timing/data': 0.01772934291511774, 'timing/model': 1.1481357499724254, 'learning_rate/qwen_vl_interface': 1.0457180418692385e-06, 'learning_rate/action_model': 1.0457180418692383e-05, 'epoch': 1.59}) +[2026-06-19 19:30:28,338][starVLA.training.train_starvla][INFO] - Step 4246, Loss: {'train/loss': 0.22526131570339203, 'train/grad_norm_pre_clip': 9.08125114440918, 'timing/data': 0.017275669961236417, 'timing/model': 1.1467465850291774, 'learning_rate/qwen_vl_interface': 1.0443016336712395e-06, 'learning_rate/action_model': 1.0443016336712393e-05, 'epoch': 1.59}) +[2026-06-19 19:30:32,955][starVLA.training.train_starvla][INFO] - Step 4247, Loss: {'train/loss': 0.2618500888347626, 'train/grad_norm_pre_clip': 15.73047924041748, 'timing/data': 0.0176850639982149, 'timing/model': 1.1500543019501492, 'learning_rate/qwen_vl_interface': 1.0428869542769744e-06, 'learning_rate/action_model': 1.0428869542769744e-05, 'epoch': 1.59}) +[2026-06-19 19:30:37,473][starVLA.training.train_starvla][INFO] - Step 4248, Loss: {'train/loss': 0.23250168561935425, 'train/grad_norm_pre_clip': 13.511085510253906, 'timing/data': 0.0173262000316754, 'timing/model': 1.1490299380384386, 'learning_rate/qwen_vl_interface': 1.0414740042679663e-06, 'learning_rate/action_model': 1.0414740042679662e-05, 'epoch': 1.59}) +[2026-06-19 19:30:41,969][starVLA.training.train_starvla][INFO] - Step 4249, Loss: {'train/loss': 0.2529015839099884, 'train/grad_norm_pre_clip': 19.2192325592041, 'timing/data': 0.017713902052491903, 'timing/model': 1.1443628639681265, 'learning_rate/qwen_vl_interface': 1.040062784225025e-06, 'learning_rate/action_model': 1.0400627842250249e-05, 'epoch': 1.59}) +[2026-06-19 19:30:46,513][starVLA.training.train_starvla][INFO] - Step 4250, Loss: {'train/loss': 0.1425151228904724, 'train/grad_norm_pre_clip': 6.099619388580322, 'timing/data': 0.01707048702519387, 'timing/model': 1.2603610929800197, 'learning_rate/qwen_vl_interface': 1.0386532947282496e-06, 'learning_rate/action_model': 1.0386532947282495e-05, 'epoch': 1.59}) +[2026-06-19 19:30:50,941][starVLA.training.train_starvla][INFO] - Step 4251, Loss: {'train/loss': 0.06266938149929047, 'train/grad_norm_pre_clip': 6.318929195404053, 'timing/data': 0.017065305029973388, 'timing/model': 1.1472531630424783, 'learning_rate/qwen_vl_interface': 1.0372455363570282e-06, 'learning_rate/action_model': 1.0372455363570283e-05, 'epoch': 1.59}) +[2026-06-19 19:30:55,551][starVLA.training.train_starvla][INFO] - Step 4252, Loss: {'train/loss': 0.2129080593585968, 'train/grad_norm_pre_clip': 11.243967056274414, 'timing/data': 0.016557540046051145, 'timing/model': 1.1557701010024175, 'learning_rate/qwen_vl_interface': 1.035839509690039e-06, 'learning_rate/action_model': 1.0358395096900388e-05, 'epoch': 1.59}) +[2026-06-19 19:31:00,148][starVLA.training.train_starvla][INFO] - Step 4253, Loss: {'train/loss': 0.15109783411026, 'train/grad_norm_pre_clip': 13.15438461303711, 'timing/data': 0.016831738990731537, 'timing/model': 1.1458108180668205, 'learning_rate/qwen_vl_interface': 1.0344352153052429e-06, 'learning_rate/action_model': 1.0344352153052427e-05, 'epoch': 1.59}) +[2026-06-19 19:31:04,733][starVLA.training.train_starvla][INFO] - Step 4254, Loss: {'train/loss': 0.15618771314620972, 'train/grad_norm_pre_clip': 14.184699058532715, 'timing/data': 0.016655298066325486, 'timing/model': 1.1452928729122505, 'learning_rate/qwen_vl_interface': 1.033032653779894e-06, 'learning_rate/action_model': 1.033032653779894e-05, 'epoch': 1.6}) +[2026-06-19 19:31:09,329][starVLA.training.train_starvla][INFO] - Step 4255, Loss: {'train/loss': 0.11825521290302277, 'train/grad_norm_pre_clip': 14.701507568359375, 'timing/data': 0.017866416019387543, 'timing/model': 1.149259143974632, 'learning_rate/qwen_vl_interface': 1.0316318256905325e-06, 'learning_rate/action_model': 1.0316318256905324e-05, 'epoch': 1.6}) +[2026-06-19 19:31:13,937][starVLA.training.train_starvla][INFO] - Step 4256, Loss: {'train/loss': 0.06800565868616104, 'train/grad_norm_pre_clip': 5.888883590698242, 'timing/data': 0.017116750008426607, 'timing/model': 1.311403737985529, 'learning_rate/qwen_vl_interface': 1.0302327316129856e-06, 'learning_rate/action_model': 1.0302327316129855e-05, 'epoch': 1.6}) +[2026-06-19 19:31:18,354][starVLA.training.train_starvla][INFO] - Step 4257, Loss: {'train/loss': 0.31105589866638184, 'train/grad_norm_pre_clip': 28.49814224243164, 'timing/data': 0.016408825991675258, 'timing/model': 1.1466738999588415, 'learning_rate/qwen_vl_interface': 1.028835372122368e-06, 'learning_rate/action_model': 1.0288353721223682e-05, 'epoch': 1.6}) +[2026-06-19 19:31:22,951][starVLA.training.train_starvla][INFO] - Step 4258, Loss: {'train/loss': 0.13000571727752686, 'train/grad_norm_pre_clip': 4.964537620544434, 'timing/data': 0.017215321073308587, 'timing/model': 1.1447534809121862, 'learning_rate/qwen_vl_interface': 1.0274397477930822e-06, 'learning_rate/action_model': 1.0274397477930822e-05, 'epoch': 1.6}) +[2026-06-19 19:31:27,547][starVLA.training.train_starvla][INFO] - Step 4259, Loss: {'train/loss': 0.19970233738422394, 'train/grad_norm_pre_clip': 16.265684127807617, 'timing/data': 0.017648892011493444, 'timing/model': 1.1450431779958308, 'learning_rate/qwen_vl_interface': 1.0260458591988164e-06, 'learning_rate/action_model': 1.0260458591988165e-05, 'epoch': 1.6}) +[2026-06-19 19:31:32,141][starVLA.training.train_starvla][INFO] - Step 4260, Loss: {'train/loss': 0.0967988595366478, 'train/grad_norm_pre_clip': 10.405150413513184, 'timing/data': 0.016859691007994115, 'timing/model': 1.1495630029821768, 'learning_rate/qwen_vl_interface': 1.0246537069125445e-06, 'learning_rate/action_model': 1.0246537069125444e-05, 'epoch': 1.6}) +[2026-06-19 19:31:36,757][starVLA.training.train_starvla][INFO] - Step 4261, Loss: {'train/loss': 0.18969252705574036, 'train/grad_norm_pre_clip': 50.642093658447266, 'timing/data': 0.016956816893070936, 'timing/model': 1.154306138982065, 'learning_rate/qwen_vl_interface': 1.0232632915065286e-06, 'learning_rate/action_model': 1.0232632915065285e-05, 'epoch': 1.6}) +[2026-06-19 19:31:41,338][starVLA.training.train_starvla][INFO] - Step 4262, Loss: {'train/loss': 0.4172154366970062, 'train/grad_norm_pre_clip': 23.921384811401367, 'timing/data': 0.01699182006996125, 'timing/model': 1.3113800570135936, 'learning_rate/qwen_vl_interface': 1.021874613552316e-06, 'learning_rate/action_model': 1.0218746135523157e-05, 'epoch': 1.6}) +[2026-06-19 19:31:45,750][starVLA.training.train_starvla][INFO] - Step 4263, Loss: {'train/loss': 0.20860950648784637, 'train/grad_norm_pre_clip': 24.843305587768555, 'timing/data': 0.017272182973101735, 'timing/model': 1.14334075700026, 'learning_rate/qwen_vl_interface': 1.0204876736207384e-06, 'learning_rate/action_model': 1.0204876736207384e-05, 'epoch': 1.6}) +[2026-06-19 19:31:50,356][starVLA.training.train_starvla][INFO] - Step 4264, Loss: {'train/loss': 0.21505799889564514, 'train/grad_norm_pre_clip': 7.84476900100708, 'timing/data': 0.01697852392680943, 'timing/model': 1.1515671400120482, 'learning_rate/qwen_vl_interface': 1.0191024722819163e-06, 'learning_rate/action_model': 1.0191024722819162e-05, 'epoch': 1.6}) +[2026-06-19 19:31:54,967][starVLA.training.train_starvla][INFO] - Step 4265, Loss: {'train/loss': 0.4123868942260742, 'train/grad_norm_pre_clip': 32.62789535522461, 'timing/data': 0.017045381013303995, 'timing/model': 1.151235104072839, 'learning_rate/qwen_vl_interface': 1.017719010105253e-06, 'learning_rate/action_model': 1.017719010105253e-05, 'epoch': 1.6}) +[2026-06-19 19:31:59,487][starVLA.training.train_starvla][INFO] - Step 4266, Loss: {'train/loss': 0.11217096447944641, 'train/grad_norm_pre_clip': 12.410435676574707, 'timing/data': 0.017374373972415924, 'timing/model': 1.151010627974756, 'learning_rate/qwen_vl_interface': 1.016337287659437e-06, 'learning_rate/action_model': 1.0163372876594368e-05, 'epoch': 1.6}) +[2026-06-19 19:32:04,019][starVLA.training.train_starvla][INFO] - Step 4267, Loss: {'train/loss': 0.0691339448094368, 'train/grad_norm_pre_clip': 7.628451347351074, 'timing/data': 0.017495476990006864, 'timing/model': 1.145037877955474, 'learning_rate/qwen_vl_interface': 1.0149573055124427e-06, 'learning_rate/action_model': 1.0149573055124426e-05, 'epoch': 1.6}) +[2026-06-19 19:32:08,638][starVLA.training.train_starvla][INFO] - Step 4268, Loss: {'train/loss': 0.2048041671514511, 'train/grad_norm_pre_clip': 12.114026069641113, 'timing/data': 0.018302846001461148, 'timing/model': 1.3308842809638008, 'learning_rate/qwen_vl_interface': 1.013579064231529e-06, 'learning_rate/action_model': 1.013579064231529e-05, 'epoch': 1.6}) +[2026-06-19 19:32:13,073][starVLA.training.train_starvla][INFO] - Step 4269, Loss: {'train/loss': 0.2773262560367584, 'train/grad_norm_pre_clip': 17.69350814819336, 'timing/data': 0.018919095979072154, 'timing/model': 1.149750469950959, 'learning_rate/qwen_vl_interface': 1.012202564383239e-06, 'learning_rate/action_model': 1.0122025643832388e-05, 'epoch': 1.6}) +[2026-06-19 19:32:17,645][starVLA.training.train_starvla][INFO] - Step 4270, Loss: {'train/loss': 0.3587995171546936, 'train/grad_norm_pre_clip': 25.060834884643555, 'timing/data': 0.018062502960674465, 'timing/model': 1.1462656010407954, 'learning_rate/qwen_vl_interface': 1.0108278065333978e-06, 'learning_rate/action_model': 1.0108278065333978e-05, 'epoch': 1.6}) +[2026-06-19 19:32:22,254][starVLA.training.train_starvla][INFO] - Step 4271, Loss: {'train/loss': 0.18139886856079102, 'train/grad_norm_pre_clip': 8.985037803649902, 'timing/data': 0.018171886913478374, 'timing/model': 1.151595885050483, 'learning_rate/qwen_vl_interface': 1.0094547912471183e-06, 'learning_rate/action_model': 1.0094547912471184e-05, 'epoch': 1.6}) +[2026-06-19 19:32:26,870][starVLA.training.train_starvla][INFO] - Step 4272, Loss: {'train/loss': 0.16044868528842926, 'train/grad_norm_pre_clip': 22.57065773010254, 'timing/data': 0.020264766993932426, 'timing/model': 1.1532659099902958, 'learning_rate/qwen_vl_interface': 1.0080835190887949e-06, 'learning_rate/action_model': 1.0080835190887947e-05, 'epoch': 1.6}) +[2026-06-19 19:32:31,499][starVLA.training.train_starvla][INFO] - Step 4273, Loss: {'train/loss': 0.26943129301071167, 'train/grad_norm_pre_clip': 26.018260955810547, 'timing/data': 0.018759213970042765, 'timing/model': 1.3386646800208837, 'learning_rate/qwen_vl_interface': 1.0067139906221052e-06, 'learning_rate/action_model': 1.0067139906221052e-05, 'epoch': 1.6}) +[2026-06-19 19:32:35,936][starVLA.training.train_starvla][INFO] - Step 4274, Loss: {'train/loss': 0.04686533659696579, 'train/grad_norm_pre_clip': 4.483145713806152, 'timing/data': 0.01703640096820891, 'timing/model': 1.147603873978369, 'learning_rate/qwen_vl_interface': 1.0053462064100112e-06, 'learning_rate/action_model': 1.0053462064100112e-05, 'epoch': 1.6}) +[2026-06-19 19:32:40,510][starVLA.training.train_starvla][INFO] - Step 4275, Loss: {'train/loss': 0.18621163070201874, 'train/grad_norm_pre_clip': 17.224321365356445, 'timing/data': 0.018465118017047644, 'timing/model': 1.1474291100166738, 'learning_rate/qwen_vl_interface': 1.003980167014757e-06, 'learning_rate/action_model': 1.003980167014757e-05, 'epoch': 1.6}) +[2026-06-19 19:32:45,087][starVLA.training.train_starvla][INFO] - Step 4276, Loss: {'train/loss': 0.15831680595874786, 'train/grad_norm_pre_clip': 17.42288589477539, 'timing/data': 0.018124302965588868, 'timing/model': 1.1459736140677705, 'learning_rate/qwen_vl_interface': 1.0026158729978702e-06, 'learning_rate/action_model': 1.0026158729978702e-05, 'epoch': 1.6}) +[2026-06-19 19:32:49,707][starVLA.training.train_starvla][INFO] - Step 4277, Loss: {'train/loss': 0.21687625348567963, 'train/grad_norm_pre_clip': 19.429351806640625, 'timing/data': 0.01713131705764681, 'timing/model': 1.1452325580175966, 'learning_rate/qwen_vl_interface': 1.0012533249201604e-06, 'learning_rate/action_model': 1.0012533249201602e-05, 'epoch': 1.6}) +[2026-06-19 19:32:54,328][starVLA.training.train_starvla][INFO] - Step 4278, Loss: {'train/loss': 0.2661723792552948, 'train/grad_norm_pre_clip': 28.975961685180664, 'timing/data': 0.017697650007903576, 'timing/model': 1.159970298060216, 'learning_rate/qwen_vl_interface': 9.99892523341719e-07, 'learning_rate/action_model': 9.99892523341719e-06, 'epoch': 1.6}) +[2026-06-19 19:32:58,922][starVLA.training.train_starvla][INFO] - Step 4279, Loss: {'train/loss': 0.11623776704072952, 'train/grad_norm_pre_clip': 13.765508651733398, 'timing/data': 0.018050289014354348, 'timing/model': 1.303580187028274, 'learning_rate/qwen_vl_interface': 9.98533468821922e-07, 'learning_rate/action_model': 9.98533468821922e-06, 'epoch': 1.6}) +[2026-06-19 19:33:03,522][starVLA.training.train_starvla][INFO] - Step 4280, Loss: {'train/loss': 0.052398115396499634, 'train/grad_norm_pre_clip': 4.542847633361816, 'timing/data': 0.017802671995013952, 'timing/model': 1.3155944410245866, 'learning_rate/qwen_vl_interface': 9.971761619194231e-07, 'learning_rate/action_model': 9.97176161919423e-06, 'epoch': 1.6}) +[2026-06-19 19:33:07,965][starVLA.training.train_starvla][INFO] - Step 4281, Loss: {'train/loss': 0.2175636738538742, 'train/grad_norm_pre_clip': 24.384557723999023, 'timing/data': 0.017250080942176282, 'timing/model': 1.1504196919267997, 'learning_rate/qwen_vl_interface': 9.958206031921606e-07, 'learning_rate/action_model': 9.958206031921604e-06, 'epoch': 1.61}) +[2026-06-19 19:33:12,573][starVLA.training.train_starvla][INFO] - Step 4282, Loss: {'train/loss': 0.2698710560798645, 'train/grad_norm_pre_clip': 23.131792068481445, 'timing/data': 0.017318001948297024, 'timing/model': 1.1486439510481432, 'learning_rate/qwen_vl_interface': 9.944667931973537e-07, 'learning_rate/action_model': 9.944667931973536e-06, 'epoch': 1.61}) +[2026-06-19 19:33:17,166][starVLA.training.train_starvla][INFO] - Step 4283, Loss: {'train/loss': 0.14073453843593597, 'train/grad_norm_pre_clip': 11.191747665405273, 'timing/data': 0.0181780899874866, 'timing/model': 1.145437411032617, 'learning_rate/qwen_vl_interface': 9.93114732491503e-07, 'learning_rate/action_model': 9.93114732491503e-06, 'epoch': 1.61}) +[2026-06-19 19:33:21,688][starVLA.training.train_starvla][INFO] - Step 4284, Loss: {'train/loss': 0.21303066611289978, 'train/grad_norm_pre_clip': 29.31041717529297, 'timing/data': 0.019171604071743786, 'timing/model': 1.1633740629768, 'learning_rate/qwen_vl_interface': 9.91764421630389e-07, 'learning_rate/action_model': 9.91764421630389e-06, 'epoch': 1.61}) +[2026-06-19 19:33:26,241][starVLA.training.train_starvla][INFO] - Step 4285, Loss: {'train/loss': 0.11431332677602768, 'train/grad_norm_pre_clip': 14.964561462402344, 'timing/data': 0.017902456922456622, 'timing/model': 1.1416409350931644, 'learning_rate/qwen_vl_interface': 9.904158611690732e-07, 'learning_rate/action_model': 9.90415861169073e-06, 'epoch': 1.61}) +[2026-06-19 19:33:30,848][starVLA.training.train_starvla][INFO] - Step 4286, Loss: {'train/loss': 0.39621394872665405, 'train/grad_norm_pre_clip': 20.679533004760742, 'timing/data': 0.017705676960758865, 'timing/model': 1.3267146390862763, 'learning_rate/qwen_vl_interface': 9.890690516618987e-07, 'learning_rate/action_model': 9.890690516618985e-06, 'epoch': 1.61}) +[2026-06-19 19:33:35,287][starVLA.training.train_starvla][INFO] - Step 4287, Loss: {'train/loss': 0.19275952875614166, 'train/grad_norm_pre_clip': 15.572750091552734, 'timing/data': 0.01724603201728314, 'timing/model': 1.1597228439059108, 'learning_rate/qwen_vl_interface': 9.877239936624875e-07, 'learning_rate/action_model': 9.877239936624875e-06, 'epoch': 1.61}) +[2026-06-19 19:33:39,914][starVLA.training.train_starvla][INFO] - Step 4288, Loss: {'train/loss': 0.07369635999202728, 'train/grad_norm_pre_clip': 7.229765892028809, 'timing/data': 0.01816116808913648, 'timing/model': 1.1462448820238933, 'learning_rate/qwen_vl_interface': 9.863806877237421e-07, 'learning_rate/action_model': 9.863806877237422e-06, 'epoch': 1.61}) +[2026-06-19 19:33:44,504][starVLA.training.train_starvla][INFO] - Step 4289, Loss: {'train/loss': 0.19354726374149323, 'train/grad_norm_pre_clip': 29.04437828063965, 'timing/data': 0.017859888030216098, 'timing/model': 1.1458196209277958, 'learning_rate/qwen_vl_interface': 9.850391343978454e-07, 'learning_rate/action_model': 9.850391343978454e-06, 'epoch': 1.61}) +[2026-06-19 19:33:49,104][starVLA.training.train_starvla][INFO] - Step 4290, Loss: {'train/loss': 0.29019033908843994, 'train/grad_norm_pre_clip': 26.641023635864258, 'timing/data': 0.017676013056188822, 'timing/model': 1.1571489449124783, 'learning_rate/qwen_vl_interface': 9.836993342362596e-07, 'learning_rate/action_model': 9.836993342362596e-06, 'epoch': 1.61}) +[2026-06-19 19:33:53,633][starVLA.training.train_starvla][INFO] - Step 4291, Loss: {'train/loss': 0.20775732398033142, 'train/grad_norm_pre_clip': 16.56204605102539, 'timing/data': 0.01745204092003405, 'timing/model': 1.1549950430635363, 'learning_rate/qwen_vl_interface': 9.823612877897246e-07, 'learning_rate/action_model': 9.823612877897245e-06, 'epoch': 1.61}) +[2026-06-19 19:33:58,212][starVLA.training.train_starvla][INFO] - Step 4292, Loss: {'train/loss': 0.1734285205602646, 'train/grad_norm_pre_clip': 16.907955169677734, 'timing/data': 0.018033126019872725, 'timing/model': 1.2934669670648873, 'learning_rate/qwen_vl_interface': 9.81024995608261e-07, 'learning_rate/action_model': 9.810249956082609e-06, 'epoch': 1.61}) +[2026-06-19 19:34:02,659][starVLA.training.train_starvla][INFO] - Step 4293, Loss: {'train/loss': 0.31728124618530273, 'train/grad_norm_pre_clip': 26.537981033325195, 'timing/data': 0.017606941983103752, 'timing/model': 1.1589166200719774, 'learning_rate/qwen_vl_interface': 9.796904582411685e-07, 'learning_rate/action_model': 9.796904582411684e-06, 'epoch': 1.61}) +[2026-06-19 19:34:07,286][starVLA.training.train_starvla][INFO] - Step 4294, Loss: {'train/loss': 0.6011289954185486, 'train/grad_norm_pre_clip': 38.748626708984375, 'timing/data': 0.01697656698524952, 'timing/model': 1.1524977480294183, 'learning_rate/qwen_vl_interface': 9.78357676237025e-07, 'learning_rate/action_model': 9.78357676237025e-06, 'epoch': 1.61}) +[2026-06-19 19:34:11,880][starVLA.training.train_starvla][INFO] - Step 4295, Loss: {'train/loss': 0.14425437152385712, 'train/grad_norm_pre_clip': 19.368953704833984, 'timing/data': 0.017775877960957587, 'timing/model': 1.1470953370444477, 'learning_rate/qwen_vl_interface': 9.77026650143687e-07, 'learning_rate/action_model': 9.77026650143687e-06, 'epoch': 1.61}) +[2026-06-19 19:34:16,487][starVLA.training.train_starvla][INFO] - Step 4296, Loss: {'train/loss': 0.10303230583667755, 'train/grad_norm_pre_clip': 9.6539888381958, 'timing/data': 0.017839362029917538, 'timing/model': 1.1506389500573277, 'learning_rate/qwen_vl_interface': 9.756973805082882e-07, 'learning_rate/action_model': 9.756973805082883e-06, 'epoch': 1.61}) +[2026-06-19 19:34:21,096][starVLA.training.train_starvla][INFO] - Step 4297, Loss: {'train/loss': 0.054376572370529175, 'train/grad_norm_pre_clip': 5.586243152618408, 'timing/data': 0.018556899041868746, 'timing/model': 1.3145934160565957, 'learning_rate/qwen_vl_interface': 9.743698678772434e-07, 'learning_rate/action_model': 9.743698678772433e-06, 'epoch': 1.61}) +[2026-06-19 19:34:25,593][starVLA.training.train_starvla][INFO] - Step 4298, Loss: {'train/loss': 0.25880855321884155, 'train/grad_norm_pre_clip': 15.243431091308594, 'timing/data': 0.01828386401757598, 'timing/model': 1.2155698649585247, 'learning_rate/qwen_vl_interface': 9.730441127962395e-07, 'learning_rate/action_model': 9.730441127962395e-06, 'epoch': 1.61}) +[2026-06-19 19:34:30,046][starVLA.training.train_starvla][INFO] - Step 4299, Loss: {'train/loss': 0.11687877774238586, 'train/grad_norm_pre_clip': 12.188791275024414, 'timing/data': 0.01700936898123473, 'timing/model': 1.159758482943289, 'learning_rate/qwen_vl_interface': 9.717201158102465e-07, 'learning_rate/action_model': 9.717201158102466e-06, 'epoch': 1.61}) +[2026-06-19 19:34:34,595][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 4300 +[2026-06-19 19:37:27,906][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 4300 +[2026-06-19 19:37:27,907][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 4300 +[2026-06-19 19:40:35,621][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 4300 +[2026-06-19 19:40:35,623][starVLA.training.train_starvla][INFO] - Step 4300, Loss: {'train/loss': 0.294440358877182, 'train/grad_norm_pre_clip': 26.831863403320312, 'eval/loss': 0.1312141716480255, 'eval/latency_0/loss': 0.08343665301799774, 'eval/latency_1/loss': 0.04732602834701538, 'eval/latency_2/loss': 0.11587222665548325, 'eval/latency_3/loss': 0.1533779352903366, 'eval/latency_4/loss': 0.25664255023002625, 'eval/flappy/loss': 0.1312141716480255, 'eval/flappy/latency_0/loss': 0.08343665301799774, 'eval/flappy/latency_1/loss': 0.04732602834701538, 'eval/flappy/latency_2/loss': 0.11587222665548325, 'eval/flappy/latency_3/loss': 0.1533779352903366, 'eval/flappy/latency_4/loss': 0.25664255023002625, 'eval/flap_precision': 0.7286245353159851, 'eval/flap_recall': 0.7127272727272728, 'eval/flap_f1': 0.7205882352941178, 'eval/flap_rate_model': 0.07472222222222222, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.7205882352941178, 'eval/latency_0/flap_precision': 0.7286245353159851, 'eval/latency_0/flap_recall': 0.7127272727272728, 'eval/latency_0/flap_f1': 0.7205882352941178, 'eval/latency_0/flap_rate_model': 0.07472222222222222, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.7205882352941178, 'timing/data': 0.017517450032755733, 'timing/model': 1.1485897639067844, 'learning_rate/qwen_vl_interface': 9.703978774635094e-07, 'learning_rate/action_model': 9.703978774635095e-06, 'epoch': 1.61}) +[2026-06-19 19:40:35,623][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_4300_state +[2026-06-19 19:40:35,626][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-19 19:40:38,886][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_4300_state/model.safetensors +[2026-06-19 19:40:46,133][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_4300_state/optimizer.bin +[2026-06-19 19:40:46,134][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_4300_state/sampler.bin +[2026-06-19 19:40:46,134][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_4300_state/sampler_1.bin +[2026-06-19 19:40:46,134][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_4300_state/random_states_0.pkl +[2026-06-19 19:40:46,158][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4300 +[2026-06-19 19:40:46,158][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-19 19:40:46,196][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-19 19:40:50,760][starVLA.training.train_starvla][INFO] - Step 4301, Loss: {'train/loss': 0.19641928374767303, 'train/grad_norm_pre_clip': 17.737457275390625, 'timing/data': 0.018013483029790223, 'timing/model': 1.1317368050804362, 'learning_rate/qwen_vl_interface': 9.6907739829955e-07, 'learning_rate/action_model': 9.6907739829955e-06, 'epoch': 1.61}) +[2026-06-19 19:40:55,324][starVLA.training.train_starvla][INFO] - Step 4302, Loss: {'train/loss': 0.23191989958286285, 'train/grad_norm_pre_clip': 11.45164966583252, 'timing/data': 0.01814370695501566, 'timing/model': 1.3012739589903504, 'learning_rate/qwen_vl_interface': 9.677586788611681e-07, 'learning_rate/action_model': 9.677586788611681e-06, 'epoch': 1.61}) +[2026-06-19 19:40:59,722][starVLA.training.train_starvla][INFO] - Step 4303, Loss: {'train/loss': 0.1964408904314041, 'train/grad_norm_pre_clip': 15.845046997070312, 'timing/data': 0.01724598603323102, 'timing/model': 1.138879365986213, 'learning_rate/qwen_vl_interface': 9.664417196904382e-07, 'learning_rate/action_model': 9.664417196904382e-06, 'epoch': 1.61}) +[2026-06-19 19:41:04,366][starVLA.training.train_starvla][INFO] - Step 4304, Loss: {'train/loss': 0.11801369488239288, 'train/grad_norm_pre_clip': 11.163471221923828, 'timing/data': 0.016029536025598645, 'timing/model': 1.14029470004607, 'learning_rate/qwen_vl_interface': 9.651265213287134e-07, 'learning_rate/action_model': 9.651265213287134e-06, 'epoch': 1.61}) +[2026-06-19 19:41:08,871][starVLA.training.train_starvla][INFO] - Step 4305, Loss: {'train/loss': 0.29838019609451294, 'train/grad_norm_pre_clip': 18.3173770904541, 'timing/data': 0.020126443007029593, 'timing/model': 1.1424496159888804, 'learning_rate/qwen_vl_interface': 9.638130843166218e-07, 'learning_rate/action_model': 9.638130843166219e-06, 'epoch': 1.61}) +[2026-06-19 19:41:13,413][starVLA.training.train_starvla][INFO] - Step 4306, Loss: {'train/loss': 0.30425041913986206, 'train/grad_norm_pre_clip': 16.051345825195312, 'timing/data': 0.017597669968381524, 'timing/model': 1.1425443689804524, 'learning_rate/qwen_vl_interface': 9.62501409194068e-07, 'learning_rate/action_model': 9.625014091940678e-06, 'epoch': 1.61}) +[2026-06-19 19:41:17,996][starVLA.training.train_starvla][INFO] - Step 4307, Loss: {'train/loss': 0.2195703238248825, 'train/grad_norm_pre_clip': 29.476045608520508, 'timing/data': 0.017833829042501748, 'timing/model': 1.1406335980864242, 'learning_rate/qwen_vl_interface': 9.611914965002317e-07, 'learning_rate/action_model': 9.611914965002316e-06, 'epoch': 1.61}) +[2026-06-19 19:41:22,567][starVLA.training.train_starvla][INFO] - Step 4308, Loss: {'train/loss': 0.18307653069496155, 'train/grad_norm_pre_clip': 20.979232788085938, 'timing/data': 0.017746540950611234, 'timing/model': 1.3108827440300956, 'learning_rate/qwen_vl_interface': 9.598833467735686e-07, 'learning_rate/action_model': 9.598833467735685e-06, 'epoch': 1.62}) +[2026-06-19 19:41:26,968][starVLA.training.train_starvla][INFO] - Step 4309, Loss: {'train/loss': 0.19254781305789948, 'train/grad_norm_pre_clip': 17.14654541015625, 'timing/data': 0.018392120953649282, 'timing/model': 1.1423087710281834, 'learning_rate/qwen_vl_interface': 9.585769605518097e-07, 'learning_rate/action_model': 9.585769605518097e-06, 'epoch': 1.62}) +[2026-06-19 19:41:31,546][starVLA.training.train_starvla][INFO] - Step 4310, Loss: {'train/loss': 0.38507065176963806, 'train/grad_norm_pre_clip': 26.732019424438477, 'timing/data': 0.01717877690680325, 'timing/model': 1.1395464800298214, 'learning_rate/qwen_vl_interface': 9.572723383719611e-07, 'learning_rate/action_model': 9.572723383719612e-06, 'epoch': 1.62}) +[2026-06-19 19:41:36,126][starVLA.training.train_starvla][INFO] - Step 4311, Loss: {'train/loss': 0.09825541079044342, 'train/grad_norm_pre_clip': 3.1023097038269043, 'timing/data': 0.017417364986613393, 'timing/model': 1.1423532859189436, 'learning_rate/qwen_vl_interface': 9.559694807703041e-07, 'learning_rate/action_model': 9.55969480770304e-06, 'epoch': 1.62}) +[2026-06-19 19:41:40,712][starVLA.training.train_starvla][INFO] - Step 4312, Loss: {'train/loss': 0.13429224491119385, 'train/grad_norm_pre_clip': 26.173662185668945, 'timing/data': 0.017722980002872646, 'timing/model': 1.1363285280531272, 'learning_rate/qwen_vl_interface': 9.546683882823945e-07, 'learning_rate/action_model': 9.546683882823945e-06, 'epoch': 1.62}) +[2026-06-19 19:41:45,267][starVLA.training.train_starvla][INFO] - Step 4313, Loss: {'train/loss': 0.17050020396709442, 'train/grad_norm_pre_clip': 9.927573204040527, 'timing/data': 0.018168400041759014, 'timing/model': 1.138403452001512, 'learning_rate/qwen_vl_interface': 9.533690614430608e-07, 'learning_rate/action_model': 9.533690614430608e-06, 'epoch': 1.62}) +[2026-06-19 19:41:49,832][starVLA.training.train_starvla][INFO] - Step 4314, Loss: {'train/loss': 0.20536282658576965, 'train/grad_norm_pre_clip': 11.902677536010742, 'timing/data': 0.016222856007516384, 'timing/model': 1.304526918916963, 'learning_rate/qwen_vl_interface': 9.520715007864082e-07, 'learning_rate/action_model': 9.52071500786408e-06, 'epoch': 1.62}) +[2026-06-19 19:41:54,249][starVLA.training.train_starvla][INFO] - Step 4315, Loss: {'train/loss': 0.3865360617637634, 'train/grad_norm_pre_clip': 23.5111026763916, 'timing/data': 0.020211540977470577, 'timing/model': 1.144616397912614, 'learning_rate/qwen_vl_interface': 9.50775706845814e-07, 'learning_rate/action_model': 9.50775706845814e-06, 'epoch': 1.62}) +[2026-06-19 19:41:58,836][starVLA.training.train_starvla][INFO] - Step 4316, Loss: {'train/loss': 0.16552381217479706, 'train/grad_norm_pre_clip': 14.877687454223633, 'timing/data': 0.016815607086755335, 'timing/model': 1.1442679450847208, 'learning_rate/qwen_vl_interface': 9.49481680153931e-07, 'learning_rate/action_model': 9.494816801539309e-06, 'epoch': 1.62}) +[2026-06-19 19:42:03,337][starVLA.training.train_starvla][INFO] - Step 4317, Loss: {'train/loss': 0.15815213322639465, 'train/grad_norm_pre_clip': 17.077590942382812, 'timing/data': 0.016547663952223957, 'timing/model': 1.1488674100255594, 'learning_rate/qwen_vl_interface': 9.481894212426848e-07, 'learning_rate/action_model': 9.481894212426847e-06, 'epoch': 1.62}) +[2026-06-19 19:42:07,827][starVLA.training.train_starvla][INFO] - Step 4318, Loss: {'train/loss': 0.19305232167243958, 'train/grad_norm_pre_clip': 17.028650283813477, 'timing/data': 0.017581071006134152, 'timing/model': 1.1413789879297838, 'learning_rate/qwen_vl_interface': 9.468989306432738e-07, 'learning_rate/action_model': 9.468989306432738e-06, 'epoch': 1.62}) +[2026-06-19 19:42:12,361][starVLA.training.train_starvla][INFO] - Step 4319, Loss: {'train/loss': 0.05225159972906113, 'train/grad_norm_pre_clip': 4.9612202644348145, 'timing/data': 0.017849375028163195, 'timing/model': 1.1490055720787495, 'learning_rate/qwen_vl_interface': 9.456102088861702e-07, 'learning_rate/action_model': 9.456102088861701e-06, 'epoch': 1.62}) +[2026-06-19 19:42:16,949][starVLA.training.train_starvla][INFO] - Step 4320, Loss: {'train/loss': 0.1879301220178604, 'train/grad_norm_pre_clip': 12.033791542053223, 'timing/data': 0.017321089981123805, 'timing/model': 1.3224662709981203, 'learning_rate/qwen_vl_interface': 9.443232565011188e-07, 'learning_rate/action_model': 9.443232565011187e-06, 'epoch': 1.62}) +[2026-06-19 19:42:21,375][starVLA.training.train_starvla][INFO] - Step 4321, Loss: {'train/loss': 0.1565849781036377, 'train/grad_norm_pre_clip': 10.261984825134277, 'timing/data': 0.017406887956894934, 'timing/model': 1.1513582569314167, 'learning_rate/qwen_vl_interface': 9.430380740171369e-07, 'learning_rate/action_model': 9.430380740171368e-06, 'epoch': 1.62}) +[2026-06-19 19:42:25,937][starVLA.training.train_starvla][INFO] - Step 4322, Loss: {'train/loss': 0.17796623706817627, 'train/grad_norm_pre_clip': 12.300908088684082, 'timing/data': 0.017555214930325747, 'timing/model': 1.1429727020440623, 'learning_rate/qwen_vl_interface': 9.417546619625143e-07, 'learning_rate/action_model': 9.417546619625143e-06, 'epoch': 1.62}) +[2026-06-19 19:42:30,512][starVLA.training.train_starvla][INFO] - Step 4323, Loss: {'train/loss': 0.2216910719871521, 'train/grad_norm_pre_clip': 27.62809944152832, 'timing/data': 0.017050381051376462, 'timing/model': 1.1484380350448191, 'learning_rate/qwen_vl_interface': 9.404730208648144e-07, 'learning_rate/action_model': 9.404730208648144e-06, 'epoch': 1.62}) +[2026-06-19 19:42:35,080][starVLA.training.train_starvla][INFO] - Step 4324, Loss: {'train/loss': 0.14737947285175323, 'train/grad_norm_pre_clip': 11.126279830932617, 'timing/data': 0.0168372510233894, 'timing/model': 1.1437536160228774, 'learning_rate/qwen_vl_interface': 9.3919315125087e-07, 'learning_rate/action_model': 9.3919315125087e-06, 'epoch': 1.62}) +[2026-06-19 19:42:39,691][starVLA.training.train_starvla][INFO] - Step 4325, Loss: {'train/loss': 0.02936548925936222, 'train/grad_norm_pre_clip': 4.453582286834717, 'timing/data': 0.017499624053016305, 'timing/model': 1.1562662729993463, 'learning_rate/qwen_vl_interface': 9.379150536467879e-07, 'learning_rate/action_model': 9.379150536467878e-06, 'epoch': 1.62}) +[2026-06-19 19:42:44,293][starVLA.training.train_starvla][INFO] - Step 4326, Loss: {'train/loss': 0.21667629480361938, 'train/grad_norm_pre_clip': 17.49216079711914, 'timing/data': 0.017627311986871064, 'timing/model': 1.3193064070073888, 'learning_rate/qwen_vl_interface': 9.366387285779453e-07, 'learning_rate/action_model': 9.366387285779453e-06, 'epoch': 1.62}) +[2026-06-19 19:42:48,724][starVLA.training.train_starvla][INFO] - Step 4327, Loss: {'train/loss': 0.08648130297660828, 'train/grad_norm_pre_clip': 12.25087833404541, 'timing/data': 0.017721410025842488, 'timing/model': 1.1608612870331854, 'learning_rate/qwen_vl_interface': 9.353641765689915e-07, 'learning_rate/action_model': 9.353641765689915e-06, 'epoch': 1.62}) +[2026-06-19 19:42:53,303][starVLA.training.train_starvla][INFO] - Step 4328, Loss: {'train/loss': 0.09448742121458054, 'train/grad_norm_pre_clip': 8.218198776245117, 'timing/data': 0.016944171045906842, 'timing/model': 1.1445718059549108, 'learning_rate/qwen_vl_interface': 9.340913981438462e-07, 'learning_rate/action_model': 9.340913981438463e-06, 'epoch': 1.62}) +[2026-06-19 19:42:57,818][starVLA.training.train_starvla][INFO] - Step 4329, Loss: {'train/loss': 0.2558903694152832, 'train/grad_norm_pre_clip': 21.70990753173828, 'timing/data': 0.016872685053385794, 'timing/model': 1.14447012799792, 'learning_rate/qwen_vl_interface': 9.328203938257019e-07, 'learning_rate/action_model': 9.328203938257018e-06, 'epoch': 1.62}) +[2026-06-19 19:43:02,432][starVLA.training.train_starvla][INFO] - Step 4330, Loss: {'train/loss': 0.18069195747375488, 'train/grad_norm_pre_clip': 14.754233360290527, 'timing/data': 0.01937207905575633, 'timing/model': 1.1617562810424715, 'learning_rate/qwen_vl_interface': 9.315511641370191e-07, 'learning_rate/action_model': 9.315511641370191e-06, 'epoch': 1.62}) +[2026-06-19 19:43:07,029][starVLA.training.train_starvla][INFO] - Step 4331, Loss: {'train/loss': 0.1513497531414032, 'train/grad_norm_pre_clip': 13.871817588806152, 'timing/data': 0.017373819020576775, 'timing/model': 1.1481617520330474, 'learning_rate/qwen_vl_interface': 9.302837095995306e-07, 'learning_rate/action_model': 9.302837095995306e-06, 'epoch': 1.62}) +[2026-06-19 19:43:11,627][starVLA.training.train_starvla][INFO] - Step 4332, Loss: {'train/loss': 0.25184184312820435, 'train/grad_norm_pre_clip': 16.921133041381836, 'timing/data': 0.016737098107114434, 'timing/model': 1.3119333629729226, 'learning_rate/qwen_vl_interface': 9.290180307342392e-07, 'learning_rate/action_model': 9.290180307342391e-06, 'epoch': 1.62}) +[2026-06-19 19:43:16,048][starVLA.training.train_starvla][INFO] - Step 4333, Loss: {'train/loss': 0.4128449559211731, 'train/grad_norm_pre_clip': 21.2774715423584, 'timing/data': 0.01739861210808158, 'timing/model': 1.143490688991733, 'learning_rate/qwen_vl_interface': 9.277541280614183e-07, 'learning_rate/action_model': 9.277541280614183e-06, 'epoch': 1.62}) +[2026-06-19 19:43:20,636][starVLA.training.train_starvla][INFO] - Step 4334, Loss: {'train/loss': 0.32046613097190857, 'train/grad_norm_pre_clip': 21.62022590637207, 'timing/data': 0.01804988505318761, 'timing/model': 1.1471372389933094, 'learning_rate/qwen_vl_interface': 9.264920021006093e-07, 'learning_rate/action_model': 9.264920021006093e-06, 'epoch': 1.63}) +[2026-06-19 19:43:25,231][starVLA.training.train_starvla][INFO] - Step 4335, Loss: {'train/loss': 0.1334812343120575, 'train/grad_norm_pre_clip': 11.794618606567383, 'timing/data': 0.017382774967700243, 'timing/model': 1.144047019071877, 'learning_rate/qwen_vl_interface': 9.252316533706246e-07, 'learning_rate/action_model': 9.252316533706246e-06, 'epoch': 1.63}) +[2026-06-19 19:43:29,847][starVLA.training.train_starvla][INFO] - Step 4336, Loss: {'train/loss': 0.33712413907051086, 'train/grad_norm_pre_clip': 17.570770263671875, 'timing/data': 0.016690330929122865, 'timing/model': 1.146392788970843, 'learning_rate/qwen_vl_interface': 9.239730823895469e-07, 'learning_rate/action_model': 9.23973082389547e-06, 'epoch': 1.63}) +[2026-06-19 19:43:34,428][starVLA.training.train_starvla][INFO] - Step 4337, Loss: {'train/loss': 0.22009214758872986, 'train/grad_norm_pre_clip': 17.72306251525879, 'timing/data': 0.01775613392237574, 'timing/model': 1.1517649830784649, 'learning_rate/qwen_vl_interface': 9.227162896747264e-07, 'learning_rate/action_model': 9.227162896747264e-06, 'epoch': 1.63}) +[2026-06-19 19:43:38,952][starVLA.training.train_starvla][INFO] - Step 4338, Loss: {'train/loss': 0.20031018555164337, 'train/grad_norm_pre_clip': 16.935474395751953, 'timing/data': 0.017098619951866567, 'timing/model': 1.1488128299824893, 'learning_rate/qwen_vl_interface': 9.214612757427833e-07, 'learning_rate/action_model': 9.214612757427833e-06, 'epoch': 1.63}) +[2026-06-19 19:43:43,540][starVLA.training.train_starvla][INFO] - Step 4339, Loss: {'train/loss': 0.2201225906610489, 'train/grad_norm_pre_clip': 20.615596771240234, 'timing/data': 0.01802632794715464, 'timing/model': 1.3109513130038977, 'learning_rate/qwen_vl_interface': 9.202080411096071e-07, 'learning_rate/action_model': 9.20208041109607e-06, 'epoch': 1.63}) +[2026-06-19 19:43:47,964][starVLA.training.train_starvla][INFO] - Step 4340, Loss: {'train/loss': 0.05360815301537514, 'train/grad_norm_pre_clip': 5.484149932861328, 'timing/data': 0.01772512902971357, 'timing/model': 1.1466431450098753, 'learning_rate/qwen_vl_interface': 9.189565862903543e-07, 'learning_rate/action_model': 9.189565862903543e-06, 'epoch': 1.63}) +[2026-06-19 19:43:52,539][starVLA.training.train_starvla][INFO] - Step 4341, Loss: {'train/loss': 0.24817317724227905, 'train/grad_norm_pre_clip': 16.12126350402832, 'timing/data': 0.01707099995110184, 'timing/model': 1.1464352930197492, 'learning_rate/qwen_vl_interface': 9.17706911799451e-07, 'learning_rate/action_model': 9.17706911799451e-06, 'epoch': 1.63}) +[2026-06-19 19:43:57,140][starVLA.training.train_starvla][INFO] - Step 4342, Loss: {'train/loss': 0.3192974925041199, 'train/grad_norm_pre_clip': 27.07294273376465, 'timing/data': 0.016083190916106105, 'timing/model': 1.1494953039800748, 'learning_rate/qwen_vl_interface': 9.164590181505912e-07, 'learning_rate/action_model': 9.164590181505911e-06, 'epoch': 1.63}) +[2026-06-19 19:44:01,751][starVLA.training.train_starvla][INFO] - Step 4343, Loss: {'train/loss': 0.13542668521404266, 'train/grad_norm_pre_clip': 12.608780860900879, 'timing/data': 0.01716696796938777, 'timing/model': 1.156099240994081, 'learning_rate/qwen_vl_interface': 9.15212905856738e-07, 'learning_rate/action_model': 9.15212905856738e-06, 'epoch': 1.63}) +[2026-06-19 19:44:06,368][starVLA.training.train_starvla][INFO] - Step 4344, Loss: {'train/loss': 0.11557650566101074, 'train/grad_norm_pre_clip': 9.49177074432373, 'timing/data': 0.016750243958085775, 'timing/model': 1.3403889678884298, 'learning_rate/qwen_vl_interface': 9.139685754301189e-07, 'learning_rate/action_model': 9.139685754301189e-06, 'epoch': 1.63}) +[2026-06-19 19:44:10,793][starVLA.training.train_starvla][INFO] - Step 4345, Loss: {'train/loss': 0.10373295843601227, 'train/grad_norm_pre_clip': 9.198019027709961, 'timing/data': 0.0170086839934811, 'timing/model': 1.1450431989505887, 'learning_rate/qwen_vl_interface': 9.127260273822324e-07, 'learning_rate/action_model': 9.127260273822325e-06, 'epoch': 1.63}) +[2026-06-19 19:44:15,393][starVLA.training.train_starvla][INFO] - Step 4346, Loss: {'train/loss': 0.2565739154815674, 'train/grad_norm_pre_clip': 17.887733459472656, 'timing/data': 0.01761410292237997, 'timing/model': 1.1441877600736916, 'learning_rate/qwen_vl_interface': 9.114852622238435e-07, 'learning_rate/action_model': 9.114852622238435e-06, 'epoch': 1.63}) +[2026-06-19 19:44:19,980][starVLA.training.train_starvla][INFO] - Step 4347, Loss: {'train/loss': 0.24049116671085358, 'train/grad_norm_pre_clip': 11.774791717529297, 'timing/data': 0.017874831915833056, 'timing/model': 1.1538291950710118, 'learning_rate/qwen_vl_interface': 9.102462804649827e-07, 'learning_rate/action_model': 9.102462804649827e-06, 'epoch': 1.63}) +[2026-06-19 19:44:24,574][starVLA.training.train_starvla][INFO] - Step 4348, Loss: {'train/loss': 0.2946109175682068, 'train/grad_norm_pre_clip': 17.781047821044922, 'timing/data': 0.016844766098074615, 'timing/model': 1.146220530034043, 'learning_rate/qwen_vl_interface': 9.090090826149489e-07, 'learning_rate/action_model': 9.09009082614949e-06, 'epoch': 1.63}) +[2026-06-19 19:44:29,167][starVLA.training.train_starvla][INFO] - Step 4349, Loss: {'train/loss': 0.3190591037273407, 'train/grad_norm_pre_clip': 17.52640151977539, 'timing/data': 0.017817271989770234, 'timing/model': 1.1487788920057938, 'learning_rate/qwen_vl_interface': 9.077736691823081e-07, 'learning_rate/action_model': 9.07773669182308e-06, 'epoch': 1.63}) +[2026-06-19 19:44:33,764][starVLA.training.train_starvla][INFO] - Step 4350, Loss: {'train/loss': 0.22643330693244934, 'train/grad_norm_pre_clip': 18.15375328063965, 'timing/data': 0.018272249028086662, 'timing/model': 1.3117925169644877, 'learning_rate/qwen_vl_interface': 9.065400406748912e-07, 'learning_rate/action_model': 9.06540040674891e-06, 'epoch': 1.63}) +[2026-06-19 19:44:38,354][starVLA.training.train_starvla][INFO] - Step 4351, Loss: {'train/loss': 0.2793634533882141, 'train/grad_norm_pre_clip': 16.55282974243164, 'timing/data': 0.017499140929430723, 'timing/model': 1.3128366919700056, 'learning_rate/qwen_vl_interface': 9.053081975997969e-07, 'learning_rate/action_model': 9.053081975997968e-06, 'epoch': 1.63}) +[2026-06-19 19:44:42,776][starVLA.training.train_starvla][INFO] - Step 4352, Loss: {'train/loss': 0.21234098076820374, 'train/grad_norm_pre_clip': 23.8643798828125, 'timing/data': 0.01749372307676822, 'timing/model': 1.14507639198564, 'learning_rate/qwen_vl_interface': 9.040781404633894e-07, 'learning_rate/action_model': 9.040781404633895e-06, 'epoch': 1.63}) +[2026-06-19 19:44:47,409][starVLA.training.train_starvla][INFO] - Step 4353, Loss: {'train/loss': 0.15892590582370758, 'train/grad_norm_pre_clip': 13.704556465148926, 'timing/data': 0.017501655966043472, 'timing/model': 1.1500932549824938, 'learning_rate/qwen_vl_interface': 9.028498697712982e-07, 'learning_rate/action_model': 9.028498697712981e-06, 'epoch': 1.63}) +[2026-06-19 19:44:52,021][starVLA.training.train_starvla][INFO] - Step 4354, Loss: {'train/loss': 0.2065671682357788, 'train/grad_norm_pre_clip': 12.881458282470703, 'timing/data': 0.017522930982522666, 'timing/model': 1.1498370839981362, 'learning_rate/qwen_vl_interface': 9.016233860284195e-07, 'learning_rate/action_model': 9.016233860284195e-06, 'epoch': 1.63}) +[2026-06-19 19:44:56,547][starVLA.training.train_starvla][INFO] - Step 4355, Loss: {'train/loss': 0.04908951744437218, 'train/grad_norm_pre_clip': 6.199648857116699, 'timing/data': 0.018232970964163542, 'timing/model': 1.1481986459111795, 'learning_rate/qwen_vl_interface': 9.003986897389142e-07, 'learning_rate/action_model': 9.003986897389141e-06, 'epoch': 1.63}) +[2026-06-19 19:45:01,127][starVLA.training.train_starvla][INFO] - Step 4356, Loss: {'train/loss': 0.2688322365283966, 'train/grad_norm_pre_clip': 30.747312545776367, 'timing/data': 0.01762853702530265, 'timing/model': 1.1548012940911576, 'learning_rate/qwen_vl_interface': 8.991757814062094e-07, 'learning_rate/action_model': 8.991757814062093e-06, 'epoch': 1.63}) +[2026-06-19 19:45:05,768][starVLA.training.train_starvla][INFO] - Step 4357, Loss: {'train/loss': 0.10807234793901443, 'train/grad_norm_pre_clip': 12.685608863830566, 'timing/data': 0.018875780049711466, 'timing/model': 1.330820873961784, 'learning_rate/qwen_vl_interface': 8.979546615329954e-07, 'learning_rate/action_model': 8.979546615329953e-06, 'epoch': 1.63}) +[2026-06-19 19:45:10,225][starVLA.training.train_starvla][INFO] - Step 4358, Loss: {'train/loss': 0.11369626969099045, 'train/grad_norm_pre_clip': 15.394136428833008, 'timing/data': 0.018935209955088794, 'timing/model': 1.1632698300527409, 'learning_rate/qwen_vl_interface': 8.967353306212286e-07, 'learning_rate/action_model': 8.967353306212286e-06, 'epoch': 1.63}) +[2026-06-19 19:45:14,855][starVLA.training.train_starvla][INFO] - Step 4359, Loss: {'train/loss': 0.17752611637115479, 'train/grad_norm_pre_clip': 21.23930549621582, 'timing/data': 0.018388803000561893, 'timing/model': 1.1547889650100842, 'learning_rate/qwen_vl_interface': 8.955177891721303e-07, 'learning_rate/action_model': 8.955177891721303e-06, 'epoch': 1.63}) +[2026-06-19 19:45:19,480][starVLA.training.train_starvla][INFO] - Step 4360, Loss: {'train/loss': 0.2070702165365219, 'train/grad_norm_pre_clip': 17.928871154785156, 'timing/data': 0.01676992396824062, 'timing/model': 1.1531185479834676, 'learning_rate/qwen_vl_interface': 8.943020376861858e-07, 'learning_rate/action_model': 8.943020376861857e-06, 'epoch': 1.63}) +[2026-06-19 19:45:24,121][starVLA.training.train_starvla][INFO] - Step 4361, Loss: {'train/loss': 0.13352353870868683, 'train/grad_norm_pre_clip': 7.635109901428223, 'timing/data': 0.01718448498286307, 'timing/model': 1.1552034750347957, 'learning_rate/qwen_vl_interface': 8.930880766631444e-07, 'learning_rate/action_model': 8.930880766631445e-06, 'epoch': 1.64}) +[2026-06-19 19:45:28,639][starVLA.training.train_starvla][INFO] - Step 4362, Loss: {'train/loss': 0.16642768681049347, 'train/grad_norm_pre_clip': 13.067811965942383, 'timing/data': 0.01818044704850763, 'timing/model': 1.1501949640223756, 'learning_rate/qwen_vl_interface': 8.918759066020201e-07, 'learning_rate/action_model': 8.9187590660202e-06, 'epoch': 1.64}) +[2026-06-19 19:45:33,244][starVLA.training.train_starvla][INFO] - Step 4363, Loss: {'train/loss': 0.11343912780284882, 'train/grad_norm_pre_clip': 7.896224021911621, 'timing/data': 0.01854185201227665, 'timing/model': 1.2829417450120673, 'learning_rate/qwen_vl_interface': 8.906655280010895e-07, 'learning_rate/action_model': 8.906655280010896e-06, 'epoch': 1.64}) +[2026-06-19 19:45:37,674][starVLA.training.train_starvla][INFO] - Step 4364, Loss: {'train/loss': 0.07390890270471573, 'train/grad_norm_pre_clip': 5.808444976806641, 'timing/data': 0.018090744968503714, 'timing/model': 1.1500766290118918, 'learning_rate/qwen_vl_interface': 8.894569413578938e-07, 'learning_rate/action_model': 8.894569413578937e-06, 'epoch': 1.64}) +[2026-06-19 19:45:42,280][starVLA.training.train_starvla][INFO] - Step 4365, Loss: {'train/loss': 0.2605479955673218, 'train/grad_norm_pre_clip': 12.583586692810059, 'timing/data': 0.017335748067125678, 'timing/model': 1.1478293229592964, 'learning_rate/qwen_vl_interface': 8.882501471692379e-07, 'learning_rate/action_model': 8.882501471692378e-06, 'epoch': 1.64}) +[2026-06-19 19:45:46,939][starVLA.training.train_starvla][INFO] - Step 4366, Loss: {'train/loss': 0.1855214536190033, 'train/grad_norm_pre_clip': 14.458422660827637, 'timing/data': 0.01766149897594005, 'timing/model': 1.172174290055409, 'learning_rate/qwen_vl_interface': 8.870451459311894e-07, 'learning_rate/action_model': 8.870451459311893e-06, 'epoch': 1.64}) +[2026-06-19 19:45:51,570][starVLA.training.train_starvla][INFO] - Step 4367, Loss: {'train/loss': 0.17970052361488342, 'train/grad_norm_pre_clip': 20.870067596435547, 'timing/data': 0.016853204928338528, 'timing/model': 1.1512865669792518, 'learning_rate/qwen_vl_interface': 8.85841938139077e-07, 'learning_rate/action_model': 8.85841938139077e-06, 'epoch': 1.64}) +[2026-06-19 19:45:56,187][starVLA.training.train_starvla][INFO] - Step 4368, Loss: {'train/loss': 0.24623830616474152, 'train/grad_norm_pre_clip': 11.541337013244629, 'timing/data': 0.017366586020216346, 'timing/model': 1.323634537984617, 'learning_rate/qwen_vl_interface': 8.846405242874955e-07, 'learning_rate/action_model': 8.846405242874954e-06, 'epoch': 1.64}) +[2026-06-19 19:46:00,721][starVLA.training.train_starvla][INFO] - Step 4369, Loss: {'train/loss': 0.4578922986984253, 'train/grad_norm_pre_clip': 22.48312759399414, 'timing/data': 0.017510270001366735, 'timing/model': 1.233757996931672, 'learning_rate/qwen_vl_interface': 8.834409048703004e-07, 'learning_rate/action_model': 8.834409048703005e-06, 'epoch': 1.64}) +[2026-06-19 19:46:05,172][starVLA.training.train_starvla][INFO] - Step 4370, Loss: {'train/loss': 0.18327118456363678, 'train/grad_norm_pre_clip': 25.591833114624023, 'timing/data': 0.01818541600368917, 'timing/model': 1.155719849979505, 'learning_rate/qwen_vl_interface': 8.822430803806098e-07, 'learning_rate/action_model': 8.822430803806097e-06, 'epoch': 1.64}) +[2026-06-19 19:46:09,756][starVLA.training.train_starvla][INFO] - Step 4371, Loss: {'train/loss': 0.12079937756061554, 'train/grad_norm_pre_clip': 10.604933738708496, 'timing/data': 0.017807521973736584, 'timing/model': 1.1548018580069765, 'learning_rate/qwen_vl_interface': 8.810470513108041e-07, 'learning_rate/action_model': 8.81047051310804e-06, 'epoch': 1.64}) +[2026-06-19 19:46:14,370][starVLA.training.train_starvla][INFO] - Step 4372, Loss: {'train/loss': 0.3738033175468445, 'train/grad_norm_pre_clip': 25.598102569580078, 'timing/data': 0.01705161004792899, 'timing/model': 1.1490688889753073, 'learning_rate/qwen_vl_interface': 8.798528181525262e-07, 'learning_rate/action_model': 8.798528181525261e-06, 'epoch': 1.64}) +[2026-06-19 19:46:18,989][starVLA.training.train_starvla][INFO] - Step 4373, Loss: {'train/loss': 0.11801358312368393, 'train/grad_norm_pre_clip': 15.7971830368042, 'timing/data': 0.018793957075104117, 'timing/model': 1.1474112659925595, 'learning_rate/qwen_vl_interface': 8.786603813966798e-07, 'learning_rate/action_model': 8.786603813966798e-06, 'epoch': 1.64}) +[2026-06-19 19:46:23,600][starVLA.training.train_starvla][INFO] - Step 4374, Loss: {'train/loss': 0.28425541520118713, 'train/grad_norm_pre_clip': 40.4390754699707, 'timing/data': 0.01668098196387291, 'timing/model': 1.315034663071856, 'learning_rate/qwen_vl_interface': 8.774697415334304e-07, 'learning_rate/action_model': 8.774697415334304e-06, 'epoch': 1.64}) +[2026-06-19 19:46:28,188][starVLA.training.train_starvla][INFO] - Step 4375, Loss: {'train/loss': 0.13724385201931, 'train/grad_norm_pre_clip': 14.262353897094727, 'timing/data': 0.015960043063387275, 'timing/model': 1.315037842025049, 'learning_rate/qwen_vl_interface': 8.762808990522056e-07, 'learning_rate/action_model': 8.762808990522057e-06, 'epoch': 1.64}) +[2026-06-19 19:46:32,660][starVLA.training.train_starvla][INFO] - Step 4376, Loss: {'train/loss': 0.14159350097179413, 'train/grad_norm_pre_clip': 12.695881843566895, 'timing/data': 0.0171512090601027, 'timing/model': 1.166759112966247, 'learning_rate/qwen_vl_interface': 8.750938544416934e-07, 'learning_rate/action_model': 8.750938544416934e-06, 'epoch': 1.64}) +[2026-06-19 19:46:37,299][starVLA.training.train_starvla][INFO] - Step 4377, Loss: {'train/loss': 0.17190590500831604, 'train/grad_norm_pre_clip': 13.096659660339355, 'timing/data': 0.018352788989432156, 'timing/model': 1.1525974789401516, 'learning_rate/qwen_vl_interface': 8.739086081898437e-07, 'learning_rate/action_model': 8.739086081898437e-06, 'epoch': 1.64}) +[2026-06-19 19:46:41,907][starVLA.training.train_starvla][INFO] - Step 4378, Loss: {'train/loss': 0.18141473829746246, 'train/grad_norm_pre_clip': 12.71613883972168, 'timing/data': 0.017049157992005348, 'timing/model': 1.1469113129423931, 'learning_rate/qwen_vl_interface': 8.727251607838647e-07, 'learning_rate/action_model': 8.727251607838647e-06, 'epoch': 1.64}) +[2026-06-19 19:46:46,400][starVLA.training.train_starvla][INFO] - Step 4379, Loss: {'train/loss': 0.2185421586036682, 'train/grad_norm_pre_clip': 21.4829158782959, 'timing/data': 0.017774743027985096, 'timing/model': 1.148503183037974, 'learning_rate/qwen_vl_interface': 8.715435127102281e-07, 'learning_rate/action_model': 8.715435127102281e-06, 'epoch': 1.64}) +[2026-06-19 19:46:50,934][starVLA.training.train_starvla][INFO] - Step 4380, Loss: {'train/loss': 0.14465314149856567, 'train/grad_norm_pre_clip': 12.423018455505371, 'timing/data': 0.018220840953290462, 'timing/model': 1.146987208048813, 'learning_rate/qwen_vl_interface': 8.703636644546651e-07, 'learning_rate/action_model': 8.703636644546651e-06, 'epoch': 1.64}) +[2026-06-19 19:46:55,539][starVLA.training.train_starvla][INFO] - Step 4381, Loss: {'train/loss': 0.08976035565137863, 'train/grad_norm_pre_clip': 7.745729446411133, 'timing/data': 0.016769996960647404, 'timing/model': 1.3192009560298175, 'learning_rate/qwen_vl_interface': 8.69185616502166e-07, 'learning_rate/action_model': 8.69185616502166e-06, 'epoch': 1.64}) +[2026-06-19 19:46:59,967][starVLA.training.train_starvla][INFO] - Step 4382, Loss: {'train/loss': 0.3126680254936218, 'train/grad_norm_pre_clip': 30.262191772460938, 'timing/data': 0.01889206492342055, 'timing/model': 1.146979279932566, 'learning_rate/qwen_vl_interface': 8.680093693369821e-07, 'learning_rate/action_model': 8.68009369336982e-06, 'epoch': 1.64}) +[2026-06-19 19:47:04,595][starVLA.training.train_starvla][INFO] - Step 4383, Loss: {'train/loss': 0.14612434804439545, 'train/grad_norm_pre_clip': 30.63121223449707, 'timing/data': 0.017497488996014, 'timing/model': 1.1627115809824318, 'learning_rate/qwen_vl_interface': 8.66834923442624e-07, 'learning_rate/action_model': 8.66834923442624e-06, 'epoch': 1.64}) +[2026-06-19 19:47:09,206][starVLA.training.train_starvla][INFO] - Step 4384, Loss: {'train/loss': 0.19935190677642822, 'train/grad_norm_pre_clip': 16.144580841064453, 'timing/data': 0.018738913000561297, 'timing/model': 1.1510243229568005, 'learning_rate/qwen_vl_interface': 8.65662279301863e-07, 'learning_rate/action_model': 8.656622793018628e-06, 'epoch': 1.64}) +[2026-06-19 19:47:13,829][starVLA.training.train_starvla][INFO] - Step 4385, Loss: {'train/loss': 0.21052022278308868, 'train/grad_norm_pre_clip': 18.83609390258789, 'timing/data': 0.02157702506519854, 'timing/model': 1.1584934149868786, 'learning_rate/qwen_vl_interface': 8.644914373967274e-07, 'learning_rate/action_model': 8.644914373967273e-06, 'epoch': 1.64}) +[2026-06-19 19:47:18,441][starVLA.training.train_starvla][INFO] - Step 4386, Loss: {'train/loss': 0.3065318465232849, 'train/grad_norm_pre_clip': 26.245088577270508, 'timing/data': 0.01590590097475797, 'timing/model': 1.148868966032751, 'learning_rate/qwen_vl_interface': 8.633223982085068e-07, 'learning_rate/action_model': 8.633223982085068e-06, 'epoch': 1.64}) +[2026-06-19 19:47:23,051][starVLA.training.train_starvla][INFO] - Step 4387, Loss: {'train/loss': 0.1593390703201294, 'train/grad_norm_pre_clip': 13.864668846130371, 'timing/data': 0.017445749952457845, 'timing/model': 1.3237443680409342, 'learning_rate/qwen_vl_interface': 8.621551622177499e-07, 'learning_rate/action_model': 8.621551622177499e-06, 'epoch': 1.64}) +[2026-06-19 19:47:27,485][starVLA.training.train_starvla][INFO] - Step 4388, Loss: {'train/loss': 0.5197901725769043, 'train/grad_norm_pre_clip': 23.23933982849121, 'timing/data': 0.017513945000246167, 'timing/model': 1.157621092046611, 'learning_rate/qwen_vl_interface': 8.609897299042615e-07, 'learning_rate/action_model': 8.609897299042614e-06, 'epoch': 1.65}) +[2026-06-19 19:47:32,088][starVLA.training.train_starvla][INFO] - Step 4389, Loss: {'train/loss': 0.11900139600038528, 'train/grad_norm_pre_clip': 14.168766021728516, 'timing/data': 0.01762276003137231, 'timing/model': 1.1522218749159947, 'learning_rate/qwen_vl_interface': 8.598261017471088e-07, 'learning_rate/action_model': 8.598261017471088e-06, 'epoch': 1.65}) +[2026-06-19 19:47:36,665][starVLA.training.train_starvla][INFO] - Step 4390, Loss: {'train/loss': 0.3314356505870819, 'train/grad_norm_pre_clip': 29.791107177734375, 'timing/data': 0.017648356035351753, 'timing/model': 1.1529149790294468, 'learning_rate/qwen_vl_interface': 8.586642782246133e-07, 'learning_rate/action_model': 8.586642782246131e-06, 'epoch': 1.65}) +[2026-06-19 19:47:41,258][starVLA.training.train_starvla][INFO] - Step 4391, Loss: {'train/loss': 0.2057727575302124, 'train/grad_norm_pre_clip': 18.150774002075195, 'timing/data': 0.017583576031029224, 'timing/model': 1.1448658109875396, 'learning_rate/qwen_vl_interface': 8.57504259814358e-07, 'learning_rate/action_model': 8.57504259814358e-06, 'epoch': 1.65}) +[2026-06-19 19:47:45,873][starVLA.training.train_starvla][INFO] - Step 4392, Loss: {'train/loss': 0.24239708483219147, 'train/grad_norm_pre_clip': 13.984403610229492, 'timing/data': 0.01791763794608414, 'timing/model': 1.1465973890153691, 'learning_rate/qwen_vl_interface': 8.563460469931823e-07, 'learning_rate/action_model': 8.563460469931822e-06, 'epoch': 1.65}) +[2026-06-19 19:47:50,478][starVLA.training.train_starvla][INFO] - Step 4393, Loss: {'train/loss': 0.27435481548309326, 'train/grad_norm_pre_clip': 21.49326515197754, 'timing/data': 0.0172000490128994, 'timing/model': 1.3134521530009806, 'learning_rate/qwen_vl_interface': 8.551896402371834e-07, 'learning_rate/action_model': 8.551896402371834e-06, 'epoch': 1.65}) +[2026-06-19 19:47:54,909][starVLA.training.train_starvla][INFO] - Step 4394, Loss: {'train/loss': 0.06981101632118225, 'train/grad_norm_pre_clip': 11.143123626708984, 'timing/data': 0.016955213039182127, 'timing/model': 1.1477494939463213, 'learning_rate/qwen_vl_interface': 8.540350400217177e-07, 'learning_rate/action_model': 8.540350400217176e-06, 'epoch': 1.65}) +[2026-06-19 19:47:59,510][starVLA.training.train_starvla][INFO] - Step 4395, Loss: {'train/loss': 0.17163506150245667, 'train/grad_norm_pre_clip': 24.489240646362305, 'timing/data': 0.018479528021998703, 'timing/model': 1.151701727998443, 'learning_rate/qwen_vl_interface': 8.528822468213964e-07, 'learning_rate/action_model': 8.528822468213964e-06, 'epoch': 1.65}) +[2026-06-19 19:48:04,131][starVLA.training.train_starvla][INFO] - Step 4396, Loss: {'train/loss': 0.19377633929252625, 'train/grad_norm_pre_clip': 22.20191192626953, 'timing/data': 0.0178634540643543, 'timing/model': 1.1500530330231413, 'learning_rate/qwen_vl_interface': 8.517312611100898e-07, 'learning_rate/action_model': 8.517312611100897e-06, 'epoch': 1.65}) +[2026-06-19 19:48:08,636][starVLA.training.train_starvla][INFO] - Step 4397, Loss: {'train/loss': 0.1048821359872818, 'train/grad_norm_pre_clip': 13.183367729187012, 'timing/data': 0.017952498979866505, 'timing/model': 1.1454565450549126, 'learning_rate/qwen_vl_interface': 8.505820833609249e-07, 'learning_rate/action_model': 8.50582083360925e-06, 'epoch': 1.65}) +[2026-06-19 19:48:13,202][starVLA.training.train_starvla][INFO] - Step 4398, Loss: {'train/loss': 0.15229171514511108, 'train/grad_norm_pre_clip': 11.241771697998047, 'timing/data': 0.01720176893286407, 'timing/model': 1.1451149720232934, 'learning_rate/qwen_vl_interface': 8.494347140462847e-07, 'learning_rate/action_model': 8.494347140462846e-06, 'epoch': 1.65}) +[2026-06-19 19:48:17,819][starVLA.training.train_starvla][INFO] - Step 4399, Loss: {'train/loss': 0.052235133945941925, 'train/grad_norm_pre_clip': 6.478057384490967, 'timing/data': 0.017093396047130227, 'timing/model': 1.3314359959913418, 'learning_rate/qwen_vl_interface': 8.482891536378105e-07, 'learning_rate/action_model': 8.482891536378105e-06, 'epoch': 1.65}) +[2026-06-19 19:48:22,251][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 4400 +[2026-06-19 19:51:17,345][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 4400 +[2026-06-19 19:51:17,345][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 4400 +[2026-06-19 19:54:25,019][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 4400 +[2026-06-19 19:54:25,020][starVLA.training.train_starvla][INFO] - Step 4400, Loss: {'train/loss': 0.17380599677562714, 'train/grad_norm_pre_clip': 11.739643096923828, 'eval/loss': 0.1289130449295044, 'eval/latency_0/loss': 0.08539742231369019, 'eval/latency_1/loss': 0.04778048396110535, 'eval/latency_2/loss': 0.11133348196744919, 'eval/latency_3/loss': 0.15234360098838806, 'eval/latency_4/loss': 0.248793363571167, 'eval/flappy/loss': 0.1289130449295044, 'eval/flappy/latency_0/loss': 0.08539742231369019, 'eval/flappy/latency_1/loss': 0.04778048396110535, 'eval/flappy/latency_2/loss': 0.11133348196744919, 'eval/flappy/latency_3/loss': 0.15234360098838806, 'eval/flappy/latency_4/loss': 0.248793363571167, 'eval/flap_precision': 0.7469879518072289, 'eval/flap_recall': 0.6763636363636364, 'eval/flap_f1': 0.7099236641221375, 'eval/flap_rate_model': 0.06916666666666667, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.7099236641221375, 'eval/latency_0/flap_precision': 0.7469879518072289, 'eval/latency_0/flap_recall': 0.6763636363636364, 'eval/latency_0/flap_f1': 0.7099236641221375, 'eval/latency_0/flap_rate_model': 0.06916666666666667, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.7099236641221375, 'timing/data': 0.01780349900946021, 'timing/model': 1.1436466490849853, 'learning_rate/qwen_vl_interface': 8.471454026063972e-07, 'learning_rate/action_model': 8.47145402606397e-06, 'epoch': 1.65}) +[2026-06-19 19:54:25,020][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_4400_state +[2026-06-19 19:54:25,023][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-19 19:54:28,417][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_4400_state/model.safetensors +[2026-06-19 19:54:35,668][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_4400_state/optimizer.bin +[2026-06-19 19:54:35,669][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_4400_state/sampler.bin +[2026-06-19 19:54:35,669][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_4400_state/sampler_1.bin +[2026-06-19 19:54:35,670][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_4400_state/random_states_0.pkl +[2026-06-19 19:54:35,694][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4400 +[2026-06-19 19:54:35,694][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-19 19:54:35,715][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-19 19:54:40,231][starVLA.training.train_starvla][INFO] - Step 4401, Loss: {'train/loss': 0.26212215423583984, 'train/grad_norm_pre_clip': 15.808646202087402, 'timing/data': 0.018569812062196434, 'timing/model': 1.1315126860281453, 'learning_rate/qwen_vl_interface': 8.460034614221988e-07, 'learning_rate/action_model': 8.460034614221988e-06, 'epoch': 1.65}) +[2026-06-19 19:54:44,777][starVLA.training.train_starvla][INFO] - Step 4402, Loss: {'train/loss': 0.10340810567140579, 'train/grad_norm_pre_clip': 11.291988372802734, 'timing/data': 0.0171162550104782, 'timing/model': 1.3066202739719301, 'learning_rate/qwen_vl_interface': 8.448633305546236e-07, 'learning_rate/action_model': 8.448633305546235e-06, 'epoch': 1.65}) +[2026-06-19 19:54:49,201][starVLA.training.train_starvla][INFO] - Step 4403, Loss: {'train/loss': 0.23355959355831146, 'train/grad_norm_pre_clip': 14.637150764465332, 'timing/data': 0.01702765200752765, 'timing/model': 1.198373329010792, 'learning_rate/qwen_vl_interface': 8.437250104723369e-07, 'learning_rate/action_model': 8.437250104723369e-06, 'epoch': 1.65}) +[2026-06-19 19:54:53,572][starVLA.training.train_starvla][INFO] - Step 4404, Loss: {'train/loss': 0.10451199859380722, 'train/grad_norm_pre_clip': 9.918328285217285, 'timing/data': 0.01667110004927963, 'timing/model': 1.131439384073019, 'learning_rate/qwen_vl_interface': 8.42588501643259e-07, 'learning_rate/action_model': 8.42588501643259e-06, 'epoch': 1.65}) +[2026-06-19 19:54:58,054][starVLA.training.train_starvla][INFO] - Step 4405, Loss: {'train/loss': 0.08725519478321075, 'train/grad_norm_pre_clip': 6.251135349273682, 'timing/data': 0.01682945596985519, 'timing/model': 1.1394380879355595, 'learning_rate/qwen_vl_interface': 8.414538045345655e-07, 'learning_rate/action_model': 8.414538045345654e-06, 'epoch': 1.65}) +[2026-06-19 19:55:02,612][starVLA.training.train_starvla][INFO] - Step 4406, Loss: {'train/loss': 0.15633825957775116, 'train/grad_norm_pre_clip': 16.0664119720459, 'timing/data': 0.019427125924266875, 'timing/model': 1.1399852440226823, 'learning_rate/qwen_vl_interface': 8.403209196126871e-07, 'learning_rate/action_model': 8.403209196126871e-06, 'epoch': 1.65}) +[2026-06-19 19:55:07,160][starVLA.training.train_starvla][INFO] - Step 4407, Loss: {'train/loss': 0.1831866055727005, 'train/grad_norm_pre_clip': 9.221903800964355, 'timing/data': 0.01889049494639039, 'timing/model': 1.1378177009755746, 'learning_rate/qwen_vl_interface': 8.391898473433111e-07, 'learning_rate/action_model': 8.39189847343311e-06, 'epoch': 1.65}) +[2026-06-19 19:55:11,705][starVLA.training.train_starvla][INFO] - Step 4408, Loss: {'train/loss': 0.2727201282978058, 'train/grad_norm_pre_clip': 20.349775314331055, 'timing/data': 0.01713052694685757, 'timing/model': 1.297532650991343, 'learning_rate/qwen_vl_interface': 8.380605881913778e-07, 'learning_rate/action_model': 8.380605881913779e-06, 'epoch': 1.65}) +[2026-06-19 19:55:16,127][starVLA.training.train_starvla][INFO] - Step 4409, Loss: {'train/loss': 0.14759893715381622, 'train/grad_norm_pre_clip': 14.927821159362793, 'timing/data': 0.017034274991601706, 'timing/model': 1.1455059639411047, 'learning_rate/qwen_vl_interface': 8.369331426210832e-07, 'learning_rate/action_model': 8.369331426210832e-06, 'epoch': 1.65}) +[2026-06-19 19:55:20,694][starVLA.training.train_starvla][INFO] - Step 4410, Loss: {'train/loss': 0.22538113594055176, 'train/grad_norm_pre_clip': 30.16114044189453, 'timing/data': 0.017574931029230356, 'timing/model': 1.1396558859851211, 'learning_rate/qwen_vl_interface': 8.358075110958785e-07, 'learning_rate/action_model': 8.358075110958784e-06, 'epoch': 1.65}) +[2026-06-19 19:55:25,269][starVLA.training.train_starvla][INFO] - Step 4411, Loss: {'train/loss': 0.37957680225372314, 'train/grad_norm_pre_clip': 22.30327606201172, 'timing/data': 0.01846563001163304, 'timing/model': 1.1404101239750162, 'learning_rate/qwen_vl_interface': 8.346836940784661e-07, 'learning_rate/action_model': 8.346836940784661e-06, 'epoch': 1.65}) +[2026-06-19 19:55:29,866][starVLA.training.train_starvla][INFO] - Step 4412, Loss: {'train/loss': 0.24866054952144623, 'train/grad_norm_pre_clip': 11.362442970275879, 'timing/data': 0.01814010296948254, 'timing/model': 1.1363798800157383, 'learning_rate/qwen_vl_interface': 8.335616920308061e-07, 'learning_rate/action_model': 8.33561692030806e-06, 'epoch': 1.65}) +[2026-06-19 19:55:34,460][starVLA.training.train_starvla][INFO] - Step 4413, Loss: {'train/loss': 0.20405615866184235, 'train/grad_norm_pre_clip': 11.677993774414062, 'timing/data': 0.01773763890378177, 'timing/model': 1.143072520964779, 'learning_rate/qwen_vl_interface': 8.324415054141111e-07, 'learning_rate/action_model': 8.32441505414111e-06, 'epoch': 1.65}) +[2026-06-19 19:55:38,963][starVLA.training.train_starvla][INFO] - Step 4414, Loss: {'train/loss': 0.2486286163330078, 'train/grad_norm_pre_clip': 29.895742416381836, 'timing/data': 0.018255536910146475, 'timing/model': 1.141356333042495, 'learning_rate/qwen_vl_interface': 8.313231346888476e-07, 'learning_rate/action_model': 8.313231346888476e-06, 'epoch': 1.66}) +[2026-06-19 19:55:43,534][starVLA.training.train_starvla][INFO] - Step 4415, Loss: {'train/loss': 0.17503805458545685, 'train/grad_norm_pre_clip': 20.165464401245117, 'timing/data': 0.01719110703561455, 'timing/model': 1.3125498050358146, 'learning_rate/qwen_vl_interface': 8.302065803147353e-07, 'learning_rate/action_model': 8.302065803147353e-06, 'epoch': 1.66}) +[2026-06-19 19:55:47,955][starVLA.training.train_starvla][INFO] - Step 4416, Loss: {'train/loss': 0.3780856728553772, 'train/grad_norm_pre_clip': 16.73982048034668, 'timing/data': 0.017803582013584673, 'timing/model': 1.143715361948125, 'learning_rate/qwen_vl_interface': 8.29091842750748e-07, 'learning_rate/action_model': 8.29091842750748e-06, 'epoch': 1.66}) +[2026-06-19 19:55:52,540][starVLA.training.train_starvla][INFO] - Step 4417, Loss: {'train/loss': 0.14433860778808594, 'train/grad_norm_pre_clip': 10.899112701416016, 'timing/data': 0.017204865929670632, 'timing/model': 1.1470313090831041, 'learning_rate/qwen_vl_interface': 8.279789224551116e-07, 'learning_rate/action_model': 8.279789224551116e-06, 'epoch': 1.66}) +[2026-06-19 19:55:57,152][starVLA.training.train_starvla][INFO] - Step 4418, Loss: {'train/loss': 0.41528189182281494, 'train/grad_norm_pre_clip': 25.774524688720703, 'timing/data': 0.018006405094638467, 'timing/model': 1.1608057580888271, 'learning_rate/qwen_vl_interface': 8.268678198853058e-07, 'learning_rate/action_model': 8.268678198853059e-06, 'epoch': 1.66}) +[2026-06-19 19:56:01,766][starVLA.training.train_starvla][INFO] - Step 4419, Loss: {'train/loss': 0.12165990471839905, 'train/grad_norm_pre_clip': 8.049552917480469, 'timing/data': 0.017253202968277037, 'timing/model': 1.1641248629894108, 'learning_rate/qwen_vl_interface': 8.257585354980628e-07, 'learning_rate/action_model': 8.257585354980628e-06, 'epoch': 1.66}) +[2026-06-19 19:56:06,357][starVLA.training.train_starvla][INFO] - Step 4420, Loss: {'train/loss': 0.08015692979097366, 'train/grad_norm_pre_clip': 6.543089866638184, 'timing/data': 0.016782625927589834, 'timing/model': 1.2998057869262993, 'learning_rate/qwen_vl_interface': 8.246510697493681e-07, 'learning_rate/action_model': 8.246510697493682e-06, 'epoch': 1.66}) +[2026-06-19 19:56:10,949][starVLA.training.train_starvla][INFO] - Step 4421, Loss: {'train/loss': 0.11693602055311203, 'train/grad_norm_pre_clip': 11.73083209991455, 'timing/data': 0.01709825301077217, 'timing/model': 1.3154704630142078, 'learning_rate/qwen_vl_interface': 8.235454230944584e-07, 'learning_rate/action_model': 8.235454230944584e-06, 'epoch': 1.66}) +[2026-06-19 19:56:15,390][starVLA.training.train_starvla][INFO] - Step 4422, Loss: {'train/loss': 0.06917218118906021, 'train/grad_norm_pre_clip': 6.045185089111328, 'timing/data': 0.0176788599928841, 'timing/model': 1.152032616082579, 'learning_rate/qwen_vl_interface': 8.224415959878234e-07, 'learning_rate/action_model': 8.224415959878234e-06, 'epoch': 1.66}) +[2026-06-19 19:56:19,996][starVLA.training.train_starvla][INFO] - Step 4423, Loss: {'train/loss': 0.1230340525507927, 'train/grad_norm_pre_clip': 26.79046630859375, 'timing/data': 0.018192671006545424, 'timing/model': 1.1476580230519176, 'learning_rate/qwen_vl_interface': 8.213395888832052e-07, 'learning_rate/action_model': 8.213395888832051e-06, 'epoch': 1.66}) +[2026-06-19 19:56:24,609][starVLA.training.train_starvla][INFO] - Step 4424, Loss: {'train/loss': 0.13791508972644806, 'train/grad_norm_pre_clip': 103.26885986328125, 'timing/data': 0.017108969972468913, 'timing/model': 1.1502622329862788, 'learning_rate/qwen_vl_interface': 8.202394022335958e-07, 'learning_rate/action_model': 8.202394022335957e-06, 'epoch': 1.66}) +[2026-06-19 19:56:29,213][starVLA.training.train_starvla][INFO] - Step 4425, Loss: {'train/loss': 0.16899237036705017, 'train/grad_norm_pre_clip': 19.332168579101562, 'timing/data': 0.016353878076188266, 'timing/model': 1.1523553399601951, 'learning_rate/qwen_vl_interface': 8.191410364912424e-07, 'learning_rate/action_model': 8.191410364912424e-06, 'epoch': 1.66}) +[2026-06-19 19:56:33,814][starVLA.training.train_starvla][INFO] - Step 4426, Loss: {'train/loss': 0.29462483525276184, 'train/grad_norm_pre_clip': 14.617791175842285, 'timing/data': 0.017194028943777084, 'timing/model': 1.1411198299610987, 'learning_rate/qwen_vl_interface': 8.1804449210764e-07, 'learning_rate/action_model': 8.1804449210764e-06, 'epoch': 1.66}) +[2026-06-19 19:56:38,421][starVLA.training.train_starvla][INFO] - Step 4427, Loss: {'train/loss': 0.07599244266748428, 'train/grad_norm_pre_clip': 8.804189682006836, 'timing/data': 0.01758409501053393, 'timing/model': 1.3283301580231637, 'learning_rate/qwen_vl_interface': 8.169497695335377e-07, 'learning_rate/action_model': 8.169497695335376e-06, 'epoch': 1.66}) +[2026-06-19 19:56:42,850][starVLA.training.train_starvla][INFO] - Step 4428, Loss: {'train/loss': 0.1961774080991745, 'train/grad_norm_pre_clip': 19.592430114746094, 'timing/data': 0.01663031696807593, 'timing/model': 1.1452594000147656, 'learning_rate/qwen_vl_interface': 8.158568692189342e-07, 'learning_rate/action_model': 8.158568692189342e-06, 'epoch': 1.66}) +[2026-06-19 19:56:47,470][starVLA.training.train_starvla][INFO] - Step 4429, Loss: {'train/loss': 0.38678961992263794, 'train/grad_norm_pre_clip': 24.770700454711914, 'timing/data': 0.017932264017872512, 'timing/model': 1.1541355489753187, 'learning_rate/qwen_vl_interface': 8.147657916130797e-07, 'learning_rate/action_model': 8.147657916130796e-06, 'epoch': 1.66}) +[2026-06-19 19:56:52,061][starVLA.training.train_starvla][INFO] - Step 4430, Loss: {'train/loss': 0.3087655007839203, 'train/grad_norm_pre_clip': 15.001611709594727, 'timing/data': 0.01975409302394837, 'timing/model': 1.1635408319998533, 'learning_rate/qwen_vl_interface': 8.136765371644753e-07, 'learning_rate/action_model': 8.136765371644753e-06, 'epoch': 1.66}) +[2026-06-19 19:56:56,632][starVLA.training.train_starvla][INFO] - Step 4431, Loss: {'train/loss': 0.22056609392166138, 'train/grad_norm_pre_clip': 21.424259185791016, 'timing/data': 0.018944708979688585, 'timing/model': 1.143998182960786, 'learning_rate/qwen_vl_interface': 8.125891063208713e-07, 'learning_rate/action_model': 8.125891063208712e-06, 'epoch': 1.66}) +[2026-06-19 19:57:01,236][starVLA.training.train_starvla][INFO] - Step 4432, Loss: {'train/loss': 0.3060588538646698, 'train/grad_norm_pre_clip': 22.342763900756836, 'timing/data': 0.01657891401555389, 'timing/model': 1.14970402198378, 'learning_rate/qwen_vl_interface': 8.115034995292698e-07, 'learning_rate/action_model': 8.115034995292698e-06, 'epoch': 1.66}) +[2026-06-19 19:57:05,834][starVLA.training.train_starvla][INFO] - Step 4433, Loss: {'train/loss': 0.25081387162208557, 'train/grad_norm_pre_clip': 29.94863510131836, 'timing/data': 0.01734772394411266, 'timing/model': 1.3148275680141523, 'learning_rate/qwen_vl_interface': 8.104197172359235e-07, 'learning_rate/action_model': 8.104197172359234e-06, 'epoch': 1.66}) +[2026-06-19 19:57:10,270][starVLA.training.train_starvla][INFO] - Step 4434, Loss: {'train/loss': 0.0864802673459053, 'train/grad_norm_pre_clip': 12.965700149536133, 'timing/data': 0.018758417922072113, 'timing/model': 1.1442185960477218, 'learning_rate/qwen_vl_interface': 8.093377598863334e-07, 'learning_rate/action_model': 8.093377598863334e-06, 'epoch': 1.66}) +[2026-06-19 19:57:14,954][starVLA.training.train_starvla][INFO] - Step 4435, Loss: {'train/loss': 0.06651703268289566, 'train/grad_norm_pre_clip': 22.367515563964844, 'timing/data': 0.020552044035866857, 'timing/model': 1.1455773710040376, 'learning_rate/qwen_vl_interface': 8.082576279252519e-07, 'learning_rate/action_model': 8.082576279252518e-06, 'epoch': 1.66}) +[2026-06-19 19:57:19,568][starVLA.training.train_starvla][INFO] - Step 4436, Loss: {'train/loss': 0.2276553213596344, 'train/grad_norm_pre_clip': 26.097030639648438, 'timing/data': 0.01682540203910321, 'timing/model': 1.1600103130331263, 'learning_rate/qwen_vl_interface': 8.071793217966803e-07, 'learning_rate/action_model': 8.071793217966802e-06, 'epoch': 1.66}) +[2026-06-19 19:57:24,173][starVLA.training.train_starvla][INFO] - Step 4437, Loss: {'train/loss': 0.20452412962913513, 'train/grad_norm_pre_clip': 13.680315017700195, 'timing/data': 0.016728747985325754, 'timing/model': 1.149084917968139, 'learning_rate/qwen_vl_interface': 8.061028419438691e-07, 'learning_rate/action_model': 8.06102841943869e-06, 'epoch': 1.66}) +[2026-06-19 19:57:28,797][starVLA.training.train_starvla][INFO] - Step 4438, Loss: {'train/loss': 0.1394416093826294, 'train/grad_norm_pre_clip': 15.680213928222656, 'timing/data': 0.01662916608620435, 'timing/model': 1.1551815830171108, 'learning_rate/qwen_vl_interface': 8.050281888093193e-07, 'learning_rate/action_model': 8.050281888093192e-06, 'epoch': 1.66}) +[2026-06-19 19:57:33,323][starVLA.training.train_starvla][INFO] - Step 4439, Loss: {'train/loss': 0.06344016641378403, 'train/grad_norm_pre_clip': 7.075121879577637, 'timing/data': 0.018338998896069825, 'timing/model': 1.1503443609690294, 'learning_rate/qwen_vl_interface': 8.039553628347794e-07, 'learning_rate/action_model': 8.039553628347793e-06, 'epoch': 1.66}) +[2026-06-19 19:57:37,927][starVLA.training.train_starvla][INFO] - Step 4440, Loss: {'train/loss': 0.12963643670082092, 'train/grad_norm_pre_clip': 17.45694923400879, 'timing/data': 0.01825896103400737, 'timing/model': 1.31165502499789, 'learning_rate/qwen_vl_interface': 8.028843644612483e-07, 'learning_rate/action_model': 8.028843644612482e-06, 'epoch': 1.66}) +[2026-06-19 19:57:42,351][starVLA.training.train_starvla][INFO] - Step 4441, Loss: {'train/loss': 0.260484516620636, 'train/grad_norm_pre_clip': 16.917715072631836, 'timing/data': 0.018279916024766862, 'timing/model': 1.1489802380092442, 'learning_rate/qwen_vl_interface': 8.01815194128973e-07, 'learning_rate/action_model': 8.01815194128973e-06, 'epoch': 1.67}) +[2026-06-19 19:57:46,907][starVLA.training.train_starvla][INFO] - Step 4442, Loss: {'train/loss': 0.26755115389823914, 'train/grad_norm_pre_clip': 21.250530242919922, 'timing/data': 0.01855107501614839, 'timing/model': 1.1474514169385657, 'learning_rate/qwen_vl_interface': 8.007478522774481e-07, 'learning_rate/action_model': 8.00747852277448e-06, 'epoch': 1.67}) +[2026-06-19 19:57:51,494][starVLA.training.train_starvla][INFO] - Step 4443, Loss: {'train/loss': 0.11967486143112183, 'train/grad_norm_pre_clip': 10.622766494750977, 'timing/data': 0.01804994000121951, 'timing/model': 1.1441761320456862, 'learning_rate/qwen_vl_interface': 7.996823393454197e-07, 'learning_rate/action_model': 7.996823393454196e-06, 'epoch': 1.67}) +[2026-06-19 19:57:56,098][starVLA.training.train_starvla][INFO] - Step 4444, Loss: {'train/loss': 0.045843563973903656, 'train/grad_norm_pre_clip': 3.9913692474365234, 'timing/data': 0.018702214933000505, 'timing/model': 1.1520805280888453, 'learning_rate/qwen_vl_interface': 7.986186557708773e-07, 'learning_rate/action_model': 7.986186557708771e-06, 'epoch': 1.67}) +[2026-06-19 19:58:00,716][starVLA.training.train_starvla][INFO] - Step 4445, Loss: {'train/loss': 0.12839354574680328, 'train/grad_norm_pre_clip': 17.156396865844727, 'timing/data': 0.017839065054431558, 'timing/model': 1.3173924509901553, 'learning_rate/qwen_vl_interface': 7.975568019910617e-07, 'learning_rate/action_model': 7.975568019910617e-06, 'epoch': 1.67}) +[2026-06-19 19:58:05,320][starVLA.training.train_starvla][INFO] - Step 4446, Loss: {'train/loss': 0.06750388443470001, 'train/grad_norm_pre_clip': 6.782803058624268, 'timing/data': 0.01844162796624005, 'timing/model': 1.3024365949677303, 'learning_rate/qwen_vl_interface': 7.964967784424618e-07, 'learning_rate/action_model': 7.964967784424618e-06, 'epoch': 1.67}) +[2026-06-19 19:58:09,776][starVLA.training.train_starvla][INFO] - Step 4447, Loss: {'train/loss': 0.24601349234580994, 'train/grad_norm_pre_clip': 29.978391647338867, 'timing/data': 0.017602153006009758, 'timing/model': 1.1497316710883752, 'learning_rate/qwen_vl_interface': 7.954385855608129e-07, 'learning_rate/action_model': 7.95438585560813e-06, 'epoch': 1.67}) +[2026-06-19 19:58:14,384][starVLA.training.train_starvla][INFO] - Step 4448, Loss: {'train/loss': 0.22720707952976227, 'train/grad_norm_pre_clip': 16.40255355834961, 'timing/data': 0.01851683296263218, 'timing/model': 1.1444932739250362, 'learning_rate/qwen_vl_interface': 7.943822237810974e-07, 'learning_rate/action_model': 7.943822237810975e-06, 'epoch': 1.67}) +[2026-06-19 19:58:18,977][starVLA.training.train_starvla][INFO] - Step 4449, Loss: {'train/loss': 0.17821772396564484, 'train/grad_norm_pre_clip': 21.925556182861328, 'timing/data': 0.017613095929846168, 'timing/model': 1.1468188560102135, 'learning_rate/qwen_vl_interface': 7.93327693537547e-07, 'learning_rate/action_model': 7.93327693537547e-06, 'epoch': 1.67}) +[2026-06-19 19:58:23,507][starVLA.training.train_starvla][INFO] - Step 4450, Loss: {'train/loss': 0.22829268872737885, 'train/grad_norm_pre_clip': 18.9321231842041, 'timing/data': 0.020686202100478113, 'timing/model': 1.155397942988202, 'learning_rate/qwen_vl_interface': 7.92274995263638e-07, 'learning_rate/action_model': 7.92274995263638e-06, 'epoch': 1.67}) +[2026-06-19 19:58:28,106][starVLA.training.train_starvla][INFO] - Step 4451, Loss: {'train/loss': 0.15832409262657166, 'train/grad_norm_pre_clip': 12.295856475830078, 'timing/data': 0.017188801080919802, 'timing/model': 1.1713678269879892, 'learning_rate/qwen_vl_interface': 7.912241293920958e-07, 'learning_rate/action_model': 7.912241293920959e-06, 'epoch': 1.67}) +[2026-06-19 19:58:32,726][starVLA.training.train_starvla][INFO] - Step 4452, Loss: {'train/loss': 0.03957808390259743, 'train/grad_norm_pre_clip': 7.312379360198975, 'timing/data': 0.017150353058241308, 'timing/model': 1.3243424750398844, 'learning_rate/qwen_vl_interface': 7.90175096354891e-07, 'learning_rate/action_model': 7.90175096354891e-06, 'epoch': 1.67}) +[2026-06-19 19:58:37,165][starVLA.training.train_starvla][INFO] - Step 4453, Loss: {'train/loss': 0.13276363909244537, 'train/grad_norm_pre_clip': 11.39765739440918, 'timing/data': 0.01664184406399727, 'timing/model': 1.156785857048817, 'learning_rate/qwen_vl_interface': 7.891278965832431e-07, 'learning_rate/action_model': 7.891278965832432e-06, 'epoch': 1.67}) +[2026-06-19 19:58:41,770][starVLA.training.train_starvla][INFO] - Step 4454, Loss: {'train/loss': 0.1638118326663971, 'train/grad_norm_pre_clip': 19.50364875793457, 'timing/data': 0.018251017085276544, 'timing/model': 1.1468641749816015, 'learning_rate/qwen_vl_interface': 7.880825305076138e-07, 'learning_rate/action_model': 7.880825305076139e-06, 'epoch': 1.67}) +[2026-06-19 19:58:46,376][starVLA.training.train_starvla][INFO] - Step 4455, Loss: {'train/loss': 0.18156208097934723, 'train/grad_norm_pre_clip': 21.244277954101562, 'timing/data': 0.0174907089676708, 'timing/model': 1.1454610039945692, 'learning_rate/qwen_vl_interface': 7.870389985577148e-07, 'learning_rate/action_model': 7.870389985577148e-06, 'epoch': 1.67}) +[2026-06-19 19:58:50,962][starVLA.training.train_starvla][INFO] - Step 4456, Loss: {'train/loss': 0.07553813606500626, 'train/grad_norm_pre_clip': 12.656145095825195, 'timing/data': 0.017835016013123095, 'timing/model': 1.144687918946147, 'learning_rate/qwen_vl_interface': 7.859973011625028e-07, 'learning_rate/action_model': 7.859973011625026e-06, 'epoch': 1.67}) +[2026-06-19 19:58:55,493][starVLA.training.train_starvla][INFO] - Step 4457, Loss: {'train/loss': 0.18406932055950165, 'train/grad_norm_pre_clip': 19.451812744140625, 'timing/data': 0.018907036050222814, 'timing/model': 1.1496817109873518, 'learning_rate/qwen_vl_interface': 7.849574387501801e-07, 'learning_rate/action_model': 7.8495743875018e-06, 'epoch': 1.67}) +[2026-06-19 19:59:00,074][starVLA.training.train_starvla][INFO] - Step 4458, Loss: {'train/loss': 0.10346885770559311, 'train/grad_norm_pre_clip': 13.936461448669434, 'timing/data': 0.017677362076938152, 'timing/model': 1.2906114080687985, 'learning_rate/qwen_vl_interface': 7.839194117481952e-07, 'learning_rate/action_model': 7.839194117481952e-06, 'epoch': 1.67}) +[2026-06-19 19:59:04,520][starVLA.training.train_starvla][INFO] - Step 4459, Loss: {'train/loss': 0.14665648341178894, 'train/grad_norm_pre_clip': 13.056170463562012, 'timing/data': 0.017283339984714985, 'timing/model': 1.1549123080912977, 'learning_rate/qwen_vl_interface': 7.828832205832418e-07, 'learning_rate/action_model': 7.828832205832418e-06, 'epoch': 1.67}) +[2026-06-19 19:59:09,120][starVLA.training.train_starvla][INFO] - Step 4460, Loss: {'train/loss': 0.13975372910499573, 'train/grad_norm_pre_clip': 19.788930892944336, 'timing/data': 0.017752173938788474, 'timing/model': 1.1548741040751338, 'learning_rate/qwen_vl_interface': 7.818488656812581e-07, 'learning_rate/action_model': 7.818488656812582e-06, 'epoch': 1.67}) +[2026-06-19 19:59:13,720][starVLA.training.train_starvla][INFO] - Step 4461, Loss: {'train/loss': 0.15550003945827484, 'train/grad_norm_pre_clip': 13.259611129760742, 'timing/data': 0.01669503201264888, 'timing/model': 1.1476984079927206, 'learning_rate/qwen_vl_interface': 7.808163474674291e-07, 'learning_rate/action_model': 7.80816347467429e-06, 'epoch': 1.67}) +[2026-06-19 19:59:18,284][starVLA.training.train_starvla][INFO] - Step 4462, Loss: {'train/loss': 0.08967100828886032, 'train/grad_norm_pre_clip': 49.79422378540039, 'timing/data': 0.01814557204488665, 'timing/model': 1.1503088780445978, 'learning_rate/qwen_vl_interface': 7.797856663661836e-07, 'learning_rate/action_model': 7.797856663661835e-06, 'epoch': 1.67}) +[2026-06-19 19:59:22,905][starVLA.training.train_starvla][INFO] - Step 4463, Loss: {'train/loss': 0.17186319828033447, 'train/grad_norm_pre_clip': 19.327924728393555, 'timing/data': 0.01809064205735922, 'timing/model': 1.1553131639957428, 'learning_rate/qwen_vl_interface': 7.787568228011959e-07, 'learning_rate/action_model': 7.78756822801196e-06, 'epoch': 1.67}) +[2026-06-19 19:59:27,500][starVLA.training.train_starvla][INFO] - Step 4464, Loss: {'train/loss': 0.052951667457818985, 'train/grad_norm_pre_clip': 6.634668350219727, 'timing/data': 0.018381308997049928, 'timing/model': 1.3126191470073536, 'learning_rate/qwen_vl_interface': 7.777298171953858e-07, 'learning_rate/action_model': 7.777298171953857e-06, 'epoch': 1.67}) +[2026-06-19 19:59:31,952][starVLA.training.train_starvla][INFO] - Step 4465, Loss: {'train/loss': 0.14484788477420807, 'train/grad_norm_pre_clip': 18.307113647460938, 'timing/data': 0.018806013045832515, 'timing/model': 1.1537161619635299, 'learning_rate/qwen_vl_interface': 7.767046499709144e-07, 'learning_rate/action_model': 7.767046499709143e-06, 'epoch': 1.67}) +[2026-06-19 19:59:36,605][starVLA.training.train_starvla][INFO] - Step 4466, Loss: {'train/loss': 0.2782651484012604, 'train/grad_norm_pre_clip': 18.271038055419922, 'timing/data': 0.017743488075211644, 'timing/model': 1.1506349859992042, 'learning_rate/qwen_vl_interface': 7.756813215491904e-07, 'learning_rate/action_model': 7.756813215491904e-06, 'epoch': 1.67}) +[2026-06-19 19:59:41,126][starVLA.training.train_starvla][INFO] - Step 4467, Loss: {'train/loss': 0.34498330950737, 'train/grad_norm_pre_clip': 38.86182403564453, 'timing/data': 0.018199107027612627, 'timing/model': 1.1503265359206125, 'learning_rate/qwen_vl_interface': 7.746598323508655e-07, 'learning_rate/action_model': 7.746598323508655e-06, 'epoch': 1.67}) +[2026-06-19 19:59:45,732][starVLA.training.train_starvla][INFO] - Step 4468, Loss: {'train/loss': 0.28898441791534424, 'train/grad_norm_pre_clip': 29.176647186279297, 'timing/data': 0.018007969949394464, 'timing/model': 1.1593743419507518, 'learning_rate/qwen_vl_interface': 7.736401827958351e-07, 'learning_rate/action_model': 7.73640182795835e-06, 'epoch': 1.68}) +[2026-06-19 19:59:50,388][starVLA.training.train_starvla][INFO] - Step 4469, Loss: {'train/loss': 0.09303449839353561, 'train/grad_norm_pre_clip': 7.47921895980835, 'timing/data': 0.018839405034668744, 'timing/model': 1.1562729460420087, 'learning_rate/qwen_vl_interface': 7.726223733032382e-07, 'learning_rate/action_model': 7.72622373303238e-06, 'epoch': 1.68}) +[2026-06-19 19:59:54,993][starVLA.training.train_starvla][INFO] - Step 4470, Loss: {'train/loss': 0.15664878487586975, 'train/grad_norm_pre_clip': 12.302513122558594, 'timing/data': 0.01677664101589471, 'timing/model': 1.3196837489958853, 'learning_rate/qwen_vl_interface': 7.716064042914584e-07, 'learning_rate/action_model': 7.716064042914584e-06, 'epoch': 1.68}) +[2026-06-19 19:59:59,465][starVLA.training.train_starvla][INFO] - Step 4471, Loss: {'train/loss': 0.14973321557044983, 'train/grad_norm_pre_clip': 17.167739868164062, 'timing/data': 0.017602168954908848, 'timing/model': 1.1499206690350547, 'learning_rate/qwen_vl_interface': 7.705922761781226e-07, 'learning_rate/action_model': 7.705922761781226e-06, 'epoch': 1.68}) +[2026-06-19 20:00:04,093][starVLA.training.train_starvla][INFO] - Step 4472, Loss: {'train/loss': 0.2377101331949234, 'train/grad_norm_pre_clip': 17.124855041503906, 'timing/data': 0.017028433037921786, 'timing/model': 1.1536305219633505, 'learning_rate/qwen_vl_interface': 7.695799893800996e-07, 'learning_rate/action_model': 7.695799893800996e-06, 'epoch': 1.68}) +[2026-06-19 20:00:08,700][starVLA.training.train_starvla][INFO] - Step 4473, Loss: {'train/loss': 0.07131752371788025, 'train/grad_norm_pre_clip': 6.669849395751953, 'timing/data': 0.01649013999849558, 'timing/model': 1.1566316990647465, 'learning_rate/qwen_vl_interface': 7.685695443135028e-07, 'learning_rate/action_model': 7.685695443135027e-06, 'epoch': 1.68}) +[2026-06-19 20:00:13,301][starVLA.training.train_starvla][INFO] - Step 4474, Loss: {'train/loss': 0.1447312980890274, 'train/grad_norm_pre_clip': 8.969747543334961, 'timing/data': 0.0165817360393703, 'timing/model': 1.1518341129412875, 'learning_rate/qwen_vl_interface': 7.675609413936888e-07, 'learning_rate/action_model': 7.675609413936888e-06, 'epoch': 1.68}) +[2026-06-19 20:00:17,900][starVLA.training.train_starvla][INFO] - Step 4475, Loss: {'train/loss': 0.14056748151779175, 'train/grad_norm_pre_clip': 10.07456111907959, 'timing/data': 0.0178066169610247, 'timing/model': 1.1494928350439295, 'learning_rate/qwen_vl_interface': 7.665541810352541e-07, 'learning_rate/action_model': 7.66554181035254e-06, 'epoch': 1.68}) +[2026-06-19 20:00:22,518][starVLA.training.train_starvla][INFO] - Step 4476, Loss: {'train/loss': 0.15632446110248566, 'train/grad_norm_pre_clip': 10.023049354553223, 'timing/data': 0.01843050797469914, 'timing/model': 1.3200695350533351, 'learning_rate/qwen_vl_interface': 7.655492636520415e-07, 'learning_rate/action_model': 7.655492636520414e-06, 'epoch': 1.68}) +[2026-06-19 20:00:27,041][starVLA.training.train_starvla][INFO] - Step 4477, Loss: {'train/loss': 0.17268036305904388, 'train/grad_norm_pre_clip': 14.760936737060547, 'timing/data': 0.01751518703531474, 'timing/model': 1.2353573709260672, 'learning_rate/qwen_vl_interface': 7.64546189657134e-07, 'learning_rate/action_model': 7.645461896571339e-06, 'epoch': 1.68}) +[2026-06-19 20:00:31,468][starVLA.training.train_starvla][INFO] - Step 4478, Loss: {'train/loss': 0.20336824655532837, 'train/grad_norm_pre_clip': 12.206565856933594, 'timing/data': 0.01658221404068172, 'timing/model': 1.1479259120533243, 'learning_rate/qwen_vl_interface': 7.635449594628571e-07, 'learning_rate/action_model': 7.635449594628571e-06, 'epoch': 1.68}) +[2026-06-19 20:00:36,102][starVLA.training.train_starvla][INFO] - Step 4479, Loss: {'train/loss': 0.28181928396224976, 'train/grad_norm_pre_clip': 22.35096549987793, 'timing/data': 0.01680668001063168, 'timing/model': 1.1533361700130627, 'learning_rate/qwen_vl_interface': 7.625455734807803e-07, 'learning_rate/action_model': 7.625455734807803e-06, 'epoch': 1.68}) +[2026-06-19 20:00:40,737][starVLA.training.train_starvla][INFO] - Step 4480, Loss: {'train/loss': 0.3418472707271576, 'train/grad_norm_pre_clip': 33.60166549682617, 'timing/data': 0.01736463897395879, 'timing/model': 1.1587345940060914, 'learning_rate/qwen_vl_interface': 7.615480321217116e-07, 'learning_rate/action_model': 7.615480321217117e-06, 'epoch': 1.68}) +[2026-06-19 20:00:45,380][starVLA.training.train_starvla][INFO] - Step 4481, Loss: {'train/loss': 0.15898475050926208, 'train/grad_norm_pre_clip': 20.892406463623047, 'timing/data': 0.016966294962912798, 'timing/model': 1.1514354440150782, 'learning_rate/qwen_vl_interface': 7.605523357957028e-07, 'learning_rate/action_model': 7.605523357957028e-06, 'epoch': 1.68}) +[2026-06-19 20:00:49,985][starVLA.training.train_starvla][INFO] - Step 4482, Loss: {'train/loss': 0.15759576857089996, 'train/grad_norm_pre_clip': 20.132434844970703, 'timing/data': 0.017641637939959764, 'timing/model': 1.3204984370386228, 'learning_rate/qwen_vl_interface': 7.59558484912048e-07, 'learning_rate/action_model': 7.59558484912048e-06, 'epoch': 1.68}) +[2026-06-19 20:00:54,426][starVLA.training.train_starvla][INFO] - Step 4483, Loss: {'train/loss': 0.22735555469989777, 'train/grad_norm_pre_clip': 19.086929321289062, 'timing/data': 0.016663728980347514, 'timing/model': 1.1453609940363094, 'learning_rate/qwen_vl_interface': 7.585664798792807e-07, 'learning_rate/action_model': 7.585664798792807e-06, 'epoch': 1.68}) +[2026-06-19 20:00:59,027][starVLA.training.train_starvla][INFO] - Step 4484, Loss: {'train/loss': 0.10481806844472885, 'train/grad_norm_pre_clip': 10.746337890625, 'timing/data': 0.017168047954328358, 'timing/model': 1.1516021579736844, 'learning_rate/qwen_vl_interface': 7.575763211051779e-07, 'learning_rate/action_model': 7.575763211051778e-06, 'epoch': 1.68}) +[2026-06-19 20:01:03,612][starVLA.training.train_starvla][INFO] - Step 4485, Loss: {'train/loss': 0.24824585020542145, 'train/grad_norm_pre_clip': 11.128084182739258, 'timing/data': 0.017990091000683606, 'timing/model': 1.159254887024872, 'learning_rate/qwen_vl_interface': 7.565880089967553e-07, 'learning_rate/action_model': 7.565880089967553e-06, 'epoch': 1.68}) +[2026-06-19 20:01:08,157][starVLA.training.train_starvla][INFO] - Step 4486, Loss: {'train/loss': 0.17060330510139465, 'train/grad_norm_pre_clip': 20.19631004333496, 'timing/data': 0.017635544994845986, 'timing/model': 1.170832189032808, 'learning_rate/qwen_vl_interface': 7.556015439602725e-07, 'learning_rate/action_model': 7.556015439602725e-06, 'epoch': 1.68}) +[2026-06-19 20:01:12,723][starVLA.training.train_starvla][INFO] - Step 4487, Loss: {'train/loss': 0.15709996223449707, 'train/grad_norm_pre_clip': 12.017569541931152, 'timing/data': 0.016192368930205703, 'timing/model': 1.152177492971532, 'learning_rate/qwen_vl_interface': 7.546169264012252e-07, 'learning_rate/action_model': 7.5461692640122525e-06, 'epoch': 1.68}) +[2026-06-19 20:01:17,358][starVLA.training.train_starvla][INFO] - Step 4488, Loss: {'train/loss': 0.10721314698457718, 'train/grad_norm_pre_clip': 13.385756492614746, 'timing/data': 0.016919511021114886, 'timing/model': 1.3364404740277678, 'learning_rate/qwen_vl_interface': 7.536341567243544e-07, 'learning_rate/action_model': 7.536341567243543e-06, 'epoch': 1.68}) +[2026-06-19 20:01:21,808][starVLA.training.train_starvla][INFO] - Step 4489, Loss: {'train/loss': 0.22022102773189545, 'train/grad_norm_pre_clip': 12.559554100036621, 'timing/data': 0.016622466035187244, 'timing/model': 1.151774015976116, 'learning_rate/qwen_vl_interface': 7.526532353336395e-07, 'learning_rate/action_model': 7.526532353336395e-06, 'epoch': 1.68}) +[2026-06-19 20:01:26,401][starVLA.training.train_starvla][INFO] - Step 4490, Loss: {'train/loss': 0.36455339193344116, 'train/grad_norm_pre_clip': 15.97301959991455, 'timing/data': 0.01760549598839134, 'timing/model': 1.1486633269814774, 'learning_rate/qwen_vl_interface': 7.516741626322994e-07, 'learning_rate/action_model': 7.516741626322994e-06, 'epoch': 1.68}) +[2026-06-19 20:01:31,025][starVLA.training.train_starvla][INFO] - Step 4491, Loss: {'train/loss': 0.16147777438163757, 'train/grad_norm_pre_clip': 12.973313331604004, 'timing/data': 0.016868853010237217, 'timing/model': 1.1418693179730326, 'learning_rate/qwen_vl_interface': 7.506969390227945e-07, 'learning_rate/action_model': 7.506969390227945e-06, 'epoch': 1.68}) +[2026-06-19 20:01:35,629][starVLA.training.train_starvla][INFO] - Step 4492, Loss: {'train/loss': 0.047779884189367294, 'train/grad_norm_pre_clip': 16.95459747314453, 'timing/data': 0.017397315008565784, 'timing/model': 1.1458168900571764, 'learning_rate/qwen_vl_interface': 7.497215649068244e-07, 'learning_rate/action_model': 7.497215649068244e-06, 'epoch': 1.68}) +[2026-06-19 20:01:40,217][starVLA.training.train_starvla][INFO] - Step 4493, Loss: {'train/loss': 0.21833078563213348, 'train/grad_norm_pre_clip': 24.219932556152344, 'timing/data': 0.018418081919662654, 'timing/model': 1.3062431759899482, 'learning_rate/qwen_vl_interface': 7.48748040685329e-07, 'learning_rate/action_model': 7.487480406853289e-06, 'epoch': 1.68}) +[2026-06-19 20:01:44,643][starVLA.training.train_starvla][INFO] - Step 4494, Loss: {'train/loss': 0.40926578640937805, 'train/grad_norm_pre_clip': 30.39720344543457, 'timing/data': 0.018316186033189297, 'timing/model': 1.1455667929258198, 'learning_rate/qwen_vl_interface': 7.477763667584867e-07, 'learning_rate/action_model': 7.477763667584866e-06, 'epoch': 1.69}) +[2026-06-19 20:01:49,234][starVLA.training.train_starvla][INFO] - Step 4495, Loss: {'train/loss': 0.2118629515171051, 'train/grad_norm_pre_clip': 10.788999557495117, 'timing/data': 0.01796362700406462, 'timing/model': 1.1471686600707471, 'learning_rate/qwen_vl_interface': 7.468065435257159e-07, 'learning_rate/action_model': 7.468065435257159e-06, 'epoch': 1.69}) +[2026-06-19 20:01:53,837][starVLA.training.train_starvla][INFO] - Step 4496, Loss: {'train/loss': 0.09144943207502365, 'train/grad_norm_pre_clip': 6.950227737426758, 'timing/data': 0.017354986048303545, 'timing/model': 1.1520838169381022, 'learning_rate/qwen_vl_interface': 7.458385713856745e-07, 'learning_rate/action_model': 7.458385713856744e-06, 'epoch': 1.69}) +[2026-06-19 20:01:58,445][starVLA.training.train_starvla][INFO] - Step 4497, Loss: {'train/loss': 0.28247523307800293, 'train/grad_norm_pre_clip': 13.93622875213623, 'timing/data': 0.017297358950600028, 'timing/model': 1.1505260589765385, 'learning_rate/qwen_vl_interface': 7.448724507362605e-07, 'learning_rate/action_model': 7.448724507362605e-06, 'epoch': 1.69}) +[2026-06-19 20:02:03,058][starVLA.training.train_starvla][INFO] - Step 4498, Loss: {'train/loss': 0.4566180408000946, 'train/grad_norm_pre_clip': 26.271821975708008, 'timing/data': 0.01762343104928732, 'timing/model': 1.1499140120577067, 'learning_rate/qwen_vl_interface': 7.439081819746078e-07, 'learning_rate/action_model': 7.4390818197460775e-06, 'epoch': 1.69}) +[2026-06-19 20:02:07,672][starVLA.training.train_starvla][INFO] - Step 4499, Loss: {'train/loss': 0.2546617388725281, 'train/grad_norm_pre_clip': 55.32033157348633, 'timing/data': 0.01773832703474909, 'timing/model': 1.3313775670249015, 'learning_rate/qwen_vl_interface': 7.42945765497091e-07, 'learning_rate/action_model': 7.42945765497091e-06, 'epoch': 1.69}) +[2026-06-19 20:02:12,276][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 4500 +[2026-06-19 20:05:07,227][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 4500 +[2026-06-19 20:05:07,228][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 4500 +[2026-06-19 20:08:14,723][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 4500 +[2026-06-19 20:08:14,724][starVLA.training.train_starvla][INFO] - Step 4500, Loss: {'train/loss': 0.2498902529478073, 'train/grad_norm_pre_clip': 21.4785213470459, 'eval/loss': 0.13521677255630493, 'eval/latency_0/loss': 0.08238157629966736, 'eval/latency_1/loss': 0.04772380366921425, 'eval/latency_2/loss': 0.11480560898780823, 'eval/latency_3/loss': 0.16582560539245605, 'eval/latency_4/loss': 0.2666710913181305, 'eval/flappy/loss': 0.13521677255630493, 'eval/flappy/latency_0/loss': 0.08238157629966736, 'eval/flappy/latency_1/loss': 0.04772380366921425, 'eval/flappy/latency_2/loss': 0.11480560898780823, 'eval/flappy/latency_3/loss': 0.16582560539245605, 'eval/flappy/latency_4/loss': 0.2666710913181305, 'eval/flap_precision': 0.7023411371237458, 'eval/flap_recall': 0.7636363636363637, 'eval/flap_f1': 0.7317073170731708, 'eval/flap_rate_model': 0.08305555555555555, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.7317073170731708, 'eval/latency_0/flap_precision': 0.7023411371237458, 'eval/latency_0/flap_recall': 0.7636363636363637, 'eval/latency_0/flap_f1': 0.7317073170731708, 'eval/latency_0/flap_rate_model': 0.08305555555555555, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.7317073170731708, 'timing/data': 0.017538494081236422, 'timing/model': 1.3191479850793257, 'learning_rate/qwen_vl_interface': 7.41985201699324e-07, 'learning_rate/action_model': 7.41985201699324e-06, 'epoch': 1.69}) +[2026-06-19 20:08:14,724][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_4500_state +[2026-06-19 20:08:14,727][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-19 20:08:18,129][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_4500_state/model.safetensors +[2026-06-19 20:08:25,636][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_4500_state/optimizer.bin +[2026-06-19 20:08:25,637][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_4500_state/sampler.bin +[2026-06-19 20:08:25,637][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_4500_state/sampler_1.bin +[2026-06-19 20:08:25,638][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_4500_state/random_states_0.pkl +[2026-06-19 20:08:25,660][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4500 +[2026-06-19 20:08:25,660][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-19 20:08:25,676][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-19 20:08:30,088][starVLA.training.train_starvla][INFO] - Step 4501, Loss: {'train/loss': 0.17384985089302063, 'train/grad_norm_pre_clip': 14.081230163574219, 'timing/data': 0.017928285989910364, 'timing/model': 1.126415017992258, 'learning_rate/qwen_vl_interface': 7.410264909761581e-07, 'learning_rate/action_model': 7.41026490976158e-06, 'epoch': 1.69}) +[2026-06-19 20:08:34,581][starVLA.training.train_starvla][INFO] - Step 4502, Loss: {'train/loss': 0.4580841660499573, 'train/grad_norm_pre_clip': 12.47425651550293, 'timing/data': 0.01810953300446272, 'timing/model': 1.1292965930188075, 'learning_rate/qwen_vl_interface': 7.40069633721683e-07, 'learning_rate/action_model': 7.400696337216829e-06, 'epoch': 1.69}) +[2026-06-19 20:08:39,163][starVLA.training.train_starvla][INFO] - Step 4503, Loss: {'train/loss': 0.1684442013502121, 'train/grad_norm_pre_clip': 16.971925735473633, 'timing/data': 0.02628425497096032, 'timing/model': 1.3309716830262914, 'learning_rate/qwen_vl_interface': 7.391146303292262e-07, 'learning_rate/action_model': 7.391146303292262e-06, 'epoch': 1.69}) +[2026-06-19 20:08:43,558][starVLA.training.train_starvla][INFO] - Step 4504, Loss: {'train/loss': 0.17291763424873352, 'train/grad_norm_pre_clip': 10.581643104553223, 'timing/data': 0.018051645951345563, 'timing/model': 1.1428804969182238, 'learning_rate/qwen_vl_interface': 7.381614811913538e-07, 'learning_rate/action_model': 7.381614811913537e-06, 'epoch': 1.69}) +[2026-06-19 20:08:48,108][starVLA.training.train_starvla][INFO] - Step 4505, Loss: {'train/loss': 0.16106976568698883, 'train/grad_norm_pre_clip': 17.345481872558594, 'timing/data': 0.01708043005783111, 'timing/model': 1.144467388978228, 'learning_rate/qwen_vl_interface': 7.372101866998693e-07, 'learning_rate/action_model': 7.372101866998693e-06, 'epoch': 1.69}) +[2026-06-19 20:08:52,680][starVLA.training.train_starvla][INFO] - Step 4506, Loss: {'train/loss': 0.24057124555110931, 'train/grad_norm_pre_clip': 17.17397117614746, 'timing/data': 0.018059920985251665, 'timing/model': 1.1382030219538137, 'learning_rate/qwen_vl_interface': 7.362607472458147e-07, 'learning_rate/action_model': 7.362607472458147e-06, 'epoch': 1.69}) +[2026-06-19 20:08:57,231][starVLA.training.train_starvla][INFO] - Step 4507, Loss: {'train/loss': 0.11297047883272171, 'train/grad_norm_pre_clip': 10.167266845703125, 'timing/data': 0.018313764012418687, 'timing/model': 1.1360432310029864, 'learning_rate/qwen_vl_interface': 7.353131632194682e-07, 'learning_rate/action_model': 7.3531316321946816e-06, 'epoch': 1.69}) +[2026-06-19 20:09:01,815][starVLA.training.train_starvla][INFO] - Step 4508, Loss: {'train/loss': 0.10637113451957703, 'train/grad_norm_pre_clip': 11.578441619873047, 'timing/data': 0.017195301013998687, 'timing/model': 1.137374365935102, 'learning_rate/qwen_vl_interface': 7.343674350103455e-07, 'learning_rate/action_model': 7.343674350103455e-06, 'epoch': 1.69}) +[2026-06-19 20:09:06,366][starVLA.training.train_starvla][INFO] - Step 4509, Loss: {'train/loss': 0.07639329135417938, 'train/grad_norm_pre_clip': 9.400712966918945, 'timing/data': 0.016940542962402105, 'timing/model': 1.3022323079640046, 'learning_rate/qwen_vl_interface': 7.334235630071995e-07, 'learning_rate/action_model': 7.334235630071995e-06, 'epoch': 1.69}) +[2026-06-19 20:09:10,895][starVLA.training.train_starvla][INFO] - Step 4510, Loss: {'train/loss': 0.4498867988586426, 'train/grad_norm_pre_clip': 22.932992935180664, 'timing/data': 0.01822336099576205, 'timing/model': 1.265636148978956, 'learning_rate/qwen_vl_interface': 7.324815475980207e-07, 'learning_rate/action_model': 7.324815475980207e-06, 'epoch': 1.69}) +[2026-06-19 20:09:15,326][starVLA.training.train_starvla][INFO] - Step 4511, Loss: {'train/loss': 0.14938046038150787, 'train/grad_norm_pre_clip': 19.040971755981445, 'timing/data': 0.01753370906226337, 'timing/model': 1.1468346930341795, 'learning_rate/qwen_vl_interface': 7.315413891700363e-07, 'learning_rate/action_model': 7.315413891700362e-06, 'epoch': 1.69}) +[2026-06-19 20:09:19,880][starVLA.training.train_starvla][INFO] - Step 4512, Loss: {'train/loss': 0.30138853192329407, 'train/grad_norm_pre_clip': 22.59853744506836, 'timing/data': 0.01691990008112043, 'timing/model': 1.1416524429805577, 'learning_rate/qwen_vl_interface': 7.306030881097092e-07, 'learning_rate/action_model': 7.3060308810970925e-06, 'epoch': 1.69}) +[2026-06-19 20:09:24,469][starVLA.training.train_starvla][INFO] - Step 4513, Loss: {'train/loss': 0.2665247917175293, 'train/grad_norm_pre_clip': 14.3152494430542, 'timing/data': 0.017376787960529327, 'timing/model': 1.1433979839785025, 'learning_rate/qwen_vl_interface': 7.296666448027406e-07, 'learning_rate/action_model': 7.296666448027406e-06, 'epoch': 1.69}) +[2026-06-19 20:09:29,068][starVLA.training.train_starvla][INFO] - Step 4514, Loss: {'train/loss': 0.07995340973138809, 'train/grad_norm_pre_clip': 15.185245513916016, 'timing/data': 0.017128497012890875, 'timing/model': 1.1562945219920948, 'learning_rate/qwen_vl_interface': 7.287320596340664e-07, 'learning_rate/action_model': 7.287320596340663e-06, 'epoch': 1.69}) +[2026-06-19 20:09:33,694][starVLA.training.train_starvla][INFO] - Step 4515, Loss: {'train/loss': 0.1792079210281372, 'train/grad_norm_pre_clip': 30.87801742553711, 'timing/data': 0.017195520107634366, 'timing/model': 1.3447436470305547, 'learning_rate/qwen_vl_interface': 7.277993329878593e-07, 'learning_rate/action_model': 7.277993329878594e-06, 'epoch': 1.69}) +[2026-06-19 20:09:38,117][starVLA.training.train_starvla][INFO] - Step 4516, Loss: {'train/loss': 0.17373861372470856, 'train/grad_norm_pre_clip': 14.856058120727539, 'timing/data': 0.016806841013021767, 'timing/model': 1.1495104409987107, 'learning_rate/qwen_vl_interface': 7.268684652475277e-07, 'learning_rate/action_model': 7.268684652475277e-06, 'epoch': 1.69}) +[2026-06-19 20:09:42,688][starVLA.training.train_starvla][INFO] - Step 4517, Loss: {'train/loss': 0.20310890674591064, 'train/grad_norm_pre_clip': 13.04808235168457, 'timing/data': 0.016859188908711076, 'timing/model': 1.1387143809115514, 'learning_rate/qwen_vl_interface': 7.259394567957169e-07, 'learning_rate/action_model': 7.259394567957169e-06, 'epoch': 1.69}) +[2026-06-19 20:09:47,275][starVLA.training.train_starvla][INFO] - Step 4518, Loss: {'train/loss': 0.1766851842403412, 'train/grad_norm_pre_clip': 18.597002029418945, 'timing/data': 0.016976527986116707, 'timing/model': 1.1494086389429867, 'learning_rate/qwen_vl_interface': 7.250123080143066e-07, 'learning_rate/action_model': 7.2501230801430665e-06, 'epoch': 1.69}) +[2026-06-19 20:09:51,880][starVLA.training.train_starvla][INFO] - Step 4519, Loss: {'train/loss': 0.26718869805336, 'train/grad_norm_pre_clip': 25.218416213989258, 'timing/data': 0.017330363974906504, 'timing/model': 1.159452087013051, 'learning_rate/qwen_vl_interface': 7.240870192844127e-07, 'learning_rate/action_model': 7.240870192844127e-06, 'epoch': 1.69}) +[2026-06-19 20:09:56,481][starVLA.training.train_starvla][INFO] - Step 4520, Loss: {'train/loss': 0.2560516595840454, 'train/grad_norm_pre_clip': 20.878849029541016, 'timing/data': 0.01681878208182752, 'timing/model': 1.144053908996284, 'learning_rate/qwen_vl_interface': 7.231635909863858e-07, 'learning_rate/action_model': 7.2316359098638585e-06, 'epoch': 1.69}) +[2026-06-19 20:10:01,071][starVLA.training.train_starvla][INFO] - Step 4521, Loss: {'train/loss': 0.1535469889640808, 'train/grad_norm_pre_clip': 6.637767791748047, 'timing/data': 0.018153614015318453, 'timing/model': 1.3101329780183733, 'learning_rate/qwen_vl_interface': 7.222420234998138e-07, 'learning_rate/action_model': 7.2224202349981385e-06, 'epoch': 1.7}) +[2026-06-19 20:10:05,624][starVLA.training.train_starvla][INFO] - Step 4522, Loss: {'train/loss': 0.29170581698417664, 'train/grad_norm_pre_clip': 23.224388122558594, 'timing/data': 0.01795018499251455, 'timing/model': 1.2692490209592506, 'learning_rate/qwen_vl_interface': 7.213223172035175e-07, 'learning_rate/action_model': 7.213223172035174e-06, 'epoch': 1.7}) +[2026-06-19 20:10:10,044][starVLA.training.train_starvla][INFO] - Step 4523, Loss: {'train/loss': 0.09651987999677658, 'train/grad_norm_pre_clip': 9.962016105651855, 'timing/data': 0.017523825983516872, 'timing/model': 1.1502860800828785, 'learning_rate/qwen_vl_interface': 7.204044724755528e-07, 'learning_rate/action_model': 7.204044724755528e-06, 'epoch': 1.7}) +[2026-06-19 20:10:14,648][starVLA.training.train_starvla][INFO] - Step 4524, Loss: {'train/loss': 0.16103819012641907, 'train/grad_norm_pre_clip': 17.14730453491211, 'timing/data': 0.017507164971902966, 'timing/model': 1.1510937809944153, 'learning_rate/qwen_vl_interface': 7.194884896932132e-07, 'learning_rate/action_model': 7.194884896932132e-06, 'epoch': 1.7}) +[2026-06-19 20:10:19,246][starVLA.training.train_starvla][INFO] - Step 4525, Loss: {'train/loss': 0.08007535338401794, 'train/grad_norm_pre_clip': 10.79139232635498, 'timing/data': 0.018355806940235198, 'timing/model': 1.1470448860200122, 'learning_rate/qwen_vl_interface': 7.185743692330221e-07, 'learning_rate/action_model': 7.185743692330221e-06, 'epoch': 1.7}) +[2026-06-19 20:10:23,863][starVLA.training.train_starvla][INFO] - Step 4526, Loss: {'train/loss': 0.0343766063451767, 'train/grad_norm_pre_clip': 3.8852975368499756, 'timing/data': 0.018560914904810488, 'timing/model': 1.15131917106919, 'learning_rate/qwen_vl_interface': 7.176621114707408e-07, 'learning_rate/action_model': 7.176621114707407e-06, 'epoch': 1.7}) +[2026-06-19 20:10:28,455][starVLA.training.train_starvla][INFO] - Step 4527, Loss: {'train/loss': 0.3758605420589447, 'train/grad_norm_pre_clip': 30.68902015686035, 'timing/data': 0.018667795113287866, 'timing/model': 1.3111673829844221, 'learning_rate/qwen_vl_interface': 7.167517167813642e-07, 'learning_rate/action_model': 7.167517167813642e-06, 'epoch': 1.7}) +[2026-06-19 20:10:32,998][starVLA.training.train_starvla][INFO] - Step 4528, Loss: {'train/loss': 0.26134732365608215, 'train/grad_norm_pre_clip': 31.577125549316406, 'timing/data': 0.017299107974395156, 'timing/model': 1.2593552389880642, 'learning_rate/qwen_vl_interface': 7.15843185539121e-07, 'learning_rate/action_model': 7.15843185539121e-06, 'epoch': 1.7}) +[2026-06-19 20:10:37,435][starVLA.training.train_starvla][INFO] - Step 4529, Loss: {'train/loss': 0.3432278037071228, 'train/grad_norm_pre_clip': 25.881193161010742, 'timing/data': 0.017306538065895438, 'timing/model': 1.1460409630089998, 'learning_rate/qwen_vl_interface': 7.149365181174739e-07, 'learning_rate/action_model': 7.149365181174739e-06, 'epoch': 1.7}) +[2026-06-19 20:10:42,058][starVLA.training.train_starvla][INFO] - Step 4530, Loss: {'train/loss': 0.1749938726425171, 'train/grad_norm_pre_clip': 16.81061363220215, 'timing/data': 0.017742753960192204, 'timing/model': 1.1578208160353824, 'learning_rate/qwen_vl_interface': 7.14031714889121e-07, 'learning_rate/action_model': 7.14031714889121e-06, 'epoch': 1.7}) +[2026-06-19 20:10:46,686][starVLA.training.train_starvla][INFO] - Step 4531, Loss: {'train/loss': 0.0868564173579216, 'train/grad_norm_pre_clip': 15.00294303894043, 'timing/data': 0.01782242802437395, 'timing/model': 1.1527119979728013, 'learning_rate/qwen_vl_interface': 7.131287762259907e-07, 'learning_rate/action_model': 7.131287762259907e-06, 'epoch': 1.7}) +[2026-06-19 20:10:51,299][starVLA.training.train_starvla][INFO] - Step 4532, Loss: {'train/loss': 0.19253255426883698, 'train/grad_norm_pre_clip': 14.384079933166504, 'timing/data': 0.01802367204800248, 'timing/model': 1.1613045489648357, 'learning_rate/qwen_vl_interface': 7.12227702499248e-07, 'learning_rate/action_model': 7.12227702499248e-06, 'epoch': 1.7}) +[2026-06-19 20:10:55,890][starVLA.training.train_starvla][INFO] - Step 4533, Loss: {'train/loss': 0.1511370837688446, 'train/grad_norm_pre_clip': 20.09535789489746, 'timing/data': 0.017025153967551887, 'timing/model': 1.3082280269591138, 'learning_rate/qwen_vl_interface': 7.113284940792903e-07, 'learning_rate/action_model': 7.113284940792903e-06, 'epoch': 1.7}) +[2026-06-19 20:11:00,454][starVLA.training.train_starvla][INFO] - Step 4534, Loss: {'train/loss': 0.28108200430870056, 'train/grad_norm_pre_clip': 42.0875129699707, 'timing/data': 0.016802452970296144, 'timing/model': 1.2764192960457876, 'learning_rate/qwen_vl_interface': 7.104311513357476e-07, 'learning_rate/action_model': 7.1043115133574755e-06, 'epoch': 1.7}) +[2026-06-19 20:11:04,899][starVLA.training.train_starvla][INFO] - Step 4535, Loss: {'train/loss': 0.3505769371986389, 'train/grad_norm_pre_clip': 20.894502639770508, 'timing/data': 0.016955846920609474, 'timing/model': 1.1526274179341272, 'learning_rate/qwen_vl_interface': 7.095356746374845e-07, 'learning_rate/action_model': 7.095356746374844e-06, 'epoch': 1.7}) +[2026-06-19 20:11:09,528][starVLA.training.train_starvla][INFO] - Step 4536, Loss: {'train/loss': 0.14344371855258942, 'train/grad_norm_pre_clip': 19.681777954101562, 'timing/data': 0.01676730695180595, 'timing/model': 1.149186622002162, 'learning_rate/qwen_vl_interface': 7.086420643525973e-07, 'learning_rate/action_model': 7.086420643525972e-06, 'epoch': 1.7}) +[2026-06-19 20:11:14,159][starVLA.training.train_starvla][INFO] - Step 4537, Loss: {'train/loss': 0.20911805331707, 'train/grad_norm_pre_clip': 16.08455467224121, 'timing/data': 0.016851734020747244, 'timing/model': 1.158663859940134, 'learning_rate/qwen_vl_interface': 7.077503208484151e-07, 'learning_rate/action_model': 7.0775032084841515e-06, 'epoch': 1.7}) +[2026-06-19 20:11:18,869][starVLA.training.train_starvla][INFO] - Step 4538, Loss: {'train/loss': 0.22245851159095764, 'train/grad_norm_pre_clip': 13.39286994934082, 'timing/data': 0.016688346047885716, 'timing/model': 1.1569432660471648, 'learning_rate/qwen_vl_interface': 7.068604444915003e-07, 'learning_rate/action_model': 7.068604444915003e-06, 'epoch': 1.7}) +[2026-06-19 20:11:23,472][starVLA.training.train_starvla][INFO] - Step 4539, Loss: {'train/loss': 0.11488384008407593, 'train/grad_norm_pre_clip': 14.282219886779785, 'timing/data': 0.0175785900792107, 'timing/model': 1.3194959899410605, 'learning_rate/qwen_vl_interface': 7.059724356476475e-07, 'learning_rate/action_model': 7.059724356476474e-06, 'epoch': 1.7}) +[2026-06-19 20:11:27,903][starVLA.training.train_starvla][INFO] - Step 4540, Loss: {'train/loss': 0.08978170156478882, 'train/grad_norm_pre_clip': 6.653358459472656, 'timing/data': 0.018251199973747134, 'timing/model': 1.148444204009138, 'learning_rate/qwen_vl_interface': 7.050862946818836e-07, 'learning_rate/action_model': 7.050862946818836e-06, 'epoch': 1.7}) +[2026-06-19 20:11:32,520][starVLA.training.train_starvla][INFO] - Step 4541, Loss: {'train/loss': 0.12494007498025894, 'train/grad_norm_pre_clip': 13.748682975769043, 'timing/data': 0.018857223098166287, 'timing/model': 1.1459357460262254, 'learning_rate/qwen_vl_interface': 7.042020219584673e-07, 'learning_rate/action_model': 7.042020219584673e-06, 'epoch': 1.7}) +[2026-06-19 20:11:37,138][starVLA.training.train_starvla][INFO] - Step 4542, Loss: {'train/loss': 0.3400602638721466, 'train/grad_norm_pre_clip': 23.50713539123535, 'timing/data': 0.0181510120164603, 'timing/model': 1.1560560030629858, 'learning_rate/qwen_vl_interface': 7.0331961784089e-07, 'learning_rate/action_model': 7.033196178408899e-06, 'epoch': 1.7}) +[2026-06-19 20:11:41,733][starVLA.training.train_starvla][INFO] - Step 4543, Loss: {'train/loss': 0.29748106002807617, 'train/grad_norm_pre_clip': 24.939910888671875, 'timing/data': 0.01800105592701584, 'timing/model': 1.1478677550330758, 'learning_rate/qwen_vl_interface': 7.024390826918748e-07, 'learning_rate/action_model': 7.024390826918747e-06, 'epoch': 1.7}) +[2026-06-19 20:11:46,347][starVLA.training.train_starvla][INFO] - Step 4544, Loss: {'train/loss': 0.32501691579818726, 'train/grad_norm_pre_clip': 22.748065948486328, 'timing/data': 0.018108466058038175, 'timing/model': 1.147025592974387, 'learning_rate/qwen_vl_interface': 7.01560416873376e-07, 'learning_rate/action_model': 7.015604168733759e-06, 'epoch': 1.7}) +[2026-06-19 20:11:50,979][starVLA.training.train_starvla][INFO] - Step 4545, Loss: {'train/loss': 0.2866573631763458, 'train/grad_norm_pre_clip': 29.579830169677734, 'timing/data': 0.017761340946890414, 'timing/model': 1.3362726490013301, 'learning_rate/qwen_vl_interface': 7.006836207465803e-07, 'learning_rate/action_model': 7.006836207465803e-06, 'epoch': 1.7}) +[2026-06-19 20:11:55,504][starVLA.training.train_starvla][INFO] - Step 4546, Loss: {'train/loss': 0.3215760588645935, 'train/grad_norm_pre_clip': 11.176559448242188, 'timing/data': 0.017392714973539114, 'timing/model': 1.2405357430689037, 'learning_rate/qwen_vl_interface': 6.998086946719058e-07, 'learning_rate/action_model': 6.998086946719057e-06, 'epoch': 1.7}) +[2026-06-19 20:11:59,950][starVLA.training.train_starvla][INFO] - Step 4547, Loss: {'train/loss': 0.20566841959953308, 'train/grad_norm_pre_clip': 23.072311401367188, 'timing/data': 0.016607441939413548, 'timing/model': 1.1501956150168553, 'learning_rate/qwen_vl_interface': 6.989356390090008e-07, 'learning_rate/action_model': 6.989356390090007e-06, 'epoch': 1.7}) +[2026-06-19 20:12:04,523][starVLA.training.train_starvla][INFO] - Step 4548, Loss: {'train/loss': 0.07911814004182816, 'train/grad_norm_pre_clip': 15.681517601013184, 'timing/data': 0.017885177046991885, 'timing/model': 1.1507722570095211, 'learning_rate/qwen_vl_interface': 6.980644541167459e-07, 'learning_rate/action_model': 6.980644541167458e-06, 'epoch': 1.71}) +[2026-06-19 20:12:09,150][starVLA.training.train_starvla][INFO] - Step 4549, Loss: {'train/loss': 0.2970362901687622, 'train/grad_norm_pre_clip': 26.665904998779297, 'timing/data': 0.017844025045633316, 'timing/model': 1.1537829260341823, 'learning_rate/qwen_vl_interface': 6.971951403532524e-07, 'learning_rate/action_model': 6.971951403532523e-06, 'epoch': 1.71}) +[2026-06-19 20:12:13,786][starVLA.training.train_starvla][INFO] - Step 4550, Loss: {'train/loss': 0.42391765117645264, 'train/grad_norm_pre_clip': 27.361413955688477, 'timing/data': 0.01719832408707589, 'timing/model': 1.1468619849765673, 'learning_rate/qwen_vl_interface': 6.963276980758622e-07, 'learning_rate/action_model': 6.963276980758622e-06, 'epoch': 1.71}) +[2026-06-19 20:12:18,403][starVLA.training.train_starvla][INFO] - Step 4551, Loss: {'train/loss': 0.1246432214975357, 'train/grad_norm_pre_clip': 8.327680587768555, 'timing/data': 0.018070700927637517, 'timing/model': 1.3122736709192395, 'learning_rate/qwen_vl_interface': 6.954621276411485e-07, 'learning_rate/action_model': 6.954621276411484e-06, 'epoch': 1.71}) +[2026-06-19 20:12:22,859][starVLA.training.train_starvla][INFO] - Step 4552, Loss: {'train/loss': 0.3627854287624359, 'train/grad_norm_pre_clip': 37.22109603881836, 'timing/data': 0.017705298028886318, 'timing/model': 1.1587787159951404, 'learning_rate/qwen_vl_interface': 6.945984294049136e-07, 'learning_rate/action_model': 6.9459842940491364e-06, 'epoch': 1.71}) +[2026-06-19 20:12:27,459][starVLA.training.train_starvla][INFO] - Step 4553, Loss: {'train/loss': 0.044154077768325806, 'train/grad_norm_pre_clip': 5.58748197555542, 'timing/data': 0.016931487945839763, 'timing/model': 1.1466054960619658, 'learning_rate/qwen_vl_interface': 6.937366037221919e-07, 'learning_rate/action_model': 6.937366037221918e-06, 'epoch': 1.71}) +[2026-06-19 20:12:32,079][starVLA.training.train_starvla][INFO] - Step 4554, Loss: {'train/loss': 0.17926111817359924, 'train/grad_norm_pre_clip': 13.812801361083984, 'timing/data': 0.016241599922068417, 'timing/model': 1.1554153679171577, 'learning_rate/qwen_vl_interface': 6.928766509472471e-07, 'learning_rate/action_model': 6.92876650947247e-06, 'epoch': 1.71}) +[2026-06-19 20:12:36,699][starVLA.training.train_starvla][INFO] - Step 4555, Loss: {'train/loss': 0.22410696744918823, 'train/grad_norm_pre_clip': 17.000980377197266, 'timing/data': 0.016693420009687543, 'timing/model': 1.1502427679952234, 'learning_rate/qwen_vl_interface': 6.920185714335737e-07, 'learning_rate/action_model': 6.9201857143357375e-06, 'epoch': 1.71}) +[2026-06-19 20:12:41,288][starVLA.training.train_starvla][INFO] - Step 4556, Loss: {'train/loss': 0.32401585578918457, 'train/grad_norm_pre_clip': 18.448305130004883, 'timing/data': 0.017240061075426638, 'timing/model': 1.1458807949675247, 'learning_rate/qwen_vl_interface': 6.911623655338959e-07, 'learning_rate/action_model': 6.911623655338959e-06, 'epoch': 1.71}) +[2026-06-19 20:12:45,877][starVLA.training.train_starvla][INFO] - Step 4557, Loss: {'train/loss': 0.17647969722747803, 'train/grad_norm_pre_clip': 15.086226463317871, 'timing/data': 0.016992958961054683, 'timing/model': 1.1481143740238622, 'learning_rate/qwen_vl_interface': 6.903080336001666e-07, 'learning_rate/action_model': 6.903080336001666e-06, 'epoch': 1.71}) +[2026-06-19 20:12:50,485][starVLA.training.train_starvla][INFO] - Step 4558, Loss: {'train/loss': 0.22075048089027405, 'train/grad_norm_pre_clip': 17.803245544433594, 'timing/data': 0.018253113026730716, 'timing/model': 1.325275006936863, 'learning_rate/qwen_vl_interface': 6.894555759835707e-07, 'learning_rate/action_model': 6.894555759835707e-06, 'epoch': 1.71}) +[2026-06-19 20:12:54,918][starVLA.training.train_starvla][INFO] - Step 4559, Loss: {'train/loss': 0.18665209412574768, 'train/grad_norm_pre_clip': 12.816668510437012, 'timing/data': 0.01741511607542634, 'timing/model': 1.1528951580403373, 'learning_rate/qwen_vl_interface': 6.886049930345204e-07, 'learning_rate/action_model': 6.8860499303452035e-06, 'epoch': 1.71}) +[2026-06-19 20:12:59,560][starVLA.training.train_starvla][INFO] - Step 4560, Loss: {'train/loss': 0.28159427642822266, 'train/grad_norm_pre_clip': 22.741117477416992, 'timing/data': 0.017367193941026926, 'timing/model': 1.1527838109759614, 'learning_rate/qwen_vl_interface': 6.877562851026584e-07, 'learning_rate/action_model': 6.877562851026584e-06, 'epoch': 1.71}) +[2026-06-19 20:13:04,166][starVLA.training.train_starvla][INFO] - Step 4561, Loss: {'train/loss': 0.1815330982208252, 'train/grad_norm_pre_clip': 15.479390144348145, 'timing/data': 0.017595964949578047, 'timing/model': 1.1539302519522607, 'learning_rate/qwen_vl_interface': 6.869094525368572e-07, 'learning_rate/action_model': 6.8690945253685715e-06, 'epoch': 1.71}) +[2026-06-19 20:13:08,746][starVLA.training.train_starvla][INFO] - Step 4562, Loss: {'train/loss': 0.1924305111169815, 'train/grad_norm_pre_clip': 15.182989120483398, 'timing/data': 0.01897004002239555, 'timing/model': 1.154002879979089, 'learning_rate/qwen_vl_interface': 6.860644956852168e-07, 'learning_rate/action_model': 6.860644956852168e-06, 'epoch': 1.71}) +[2026-06-19 20:13:13,348][starVLA.training.train_starvla][INFO] - Step 4563, Loss: {'train/loss': 0.250497430562973, 'train/grad_norm_pre_clip': 21.76593589782715, 'timing/data': 0.017527254996821284, 'timing/model': 1.1460386319085956, 'learning_rate/qwen_vl_interface': 6.852214148950667e-07, 'learning_rate/action_model': 6.852214148950667e-06, 'epoch': 1.71}) +[2026-06-19 20:13:17,950][starVLA.training.train_starvla][INFO] - Step 4564, Loss: {'train/loss': 0.2527025640010834, 'train/grad_norm_pre_clip': 15.600143432617188, 'timing/data': 0.016947185969911516, 'timing/model': 1.3138598379446194, 'learning_rate/qwen_vl_interface': 6.843802105129667e-07, 'learning_rate/action_model': 6.843802105129666e-06, 'epoch': 1.71}) +[2026-06-19 20:13:22,451][starVLA.training.train_starvla][INFO] - Step 4565, Loss: {'train/loss': 0.15952405333518982, 'train/grad_norm_pre_clip': 16.584396362304688, 'timing/data': 0.01747499906923622, 'timing/model': 1.1548743080347776, 'learning_rate/qwen_vl_interface': 6.835408828847035e-07, 'learning_rate/action_model': 6.8354088288470356e-06, 'epoch': 1.71}) +[2026-06-19 20:13:27,070][starVLA.training.train_starvla][INFO] - Step 4566, Loss: {'train/loss': 0.20248472690582275, 'train/grad_norm_pre_clip': 19.279781341552734, 'timing/data': 0.01836669398471713, 'timing/model': 1.15000202099327, 'learning_rate/qwen_vl_interface': 6.827034323552929e-07, 'learning_rate/action_model': 6.827034323552929e-06, 'epoch': 1.71}) +[2026-06-19 20:13:31,611][starVLA.training.train_starvla][INFO] - Step 4567, Loss: {'train/loss': 0.127637580037117, 'train/grad_norm_pre_clip': 11.692351341247559, 'timing/data': 0.015839894069358706, 'timing/model': 1.144640518934466, 'learning_rate/qwen_vl_interface': 6.818678592689794e-07, 'learning_rate/action_model': 6.818678592689794e-06, 'epoch': 1.71}) +[2026-06-19 20:13:36,213][starVLA.training.train_starvla][INFO] - Step 4568, Loss: {'train/loss': 0.2259664237499237, 'train/grad_norm_pre_clip': 17.20936393737793, 'timing/data': 0.017357656033709645, 'timing/model': 1.1471508030081168, 'learning_rate/qwen_vl_interface': 6.810341639692357e-07, 'learning_rate/action_model': 6.810341639692357e-06, 'epoch': 1.71}) +[2026-06-19 20:13:40,823][starVLA.training.train_starvla][INFO] - Step 4569, Loss: {'train/loss': 0.31029781699180603, 'train/grad_norm_pre_clip': 23.17055892944336, 'timing/data': 0.017807706026360393, 'timing/model': 1.151988280005753, 'learning_rate/qwen_vl_interface': 6.802023467987631e-07, 'learning_rate/action_model': 6.802023467987631e-06, 'epoch': 1.71}) +[2026-06-19 20:13:45,453][starVLA.training.train_starvla][INFO] - Step 4570, Loss: {'train/loss': 0.12452254444360733, 'train/grad_norm_pre_clip': 11.843422889709473, 'timing/data': 0.016998979961499572, 'timing/model': 1.3255986460717395, 'learning_rate/qwen_vl_interface': 6.793724080994884e-07, 'learning_rate/action_model': 6.793724080994884e-06, 'epoch': 1.71}) +[2026-06-19 20:13:49,909][starVLA.training.train_starvla][INFO] - Step 4571, Loss: {'train/loss': 0.31447476148605347, 'train/grad_norm_pre_clip': 25.896743774414062, 'timing/data': 0.01717100292444229, 'timing/model': 1.1477692179614678, 'learning_rate/qwen_vl_interface': 6.785443482125693e-07, 'learning_rate/action_model': 6.7854434821256915e-06, 'epoch': 1.71}) +[2026-06-19 20:13:54,488][starVLA.training.train_starvla][INFO] - Step 4572, Loss: {'train/loss': 0.09505300223827362, 'train/grad_norm_pre_clip': 10.72890567779541, 'timing/data': 0.016965948045253754, 'timing/model': 1.1455050529912114, 'learning_rate/qwen_vl_interface': 6.777181674783895e-07, 'learning_rate/action_model': 6.777181674783895e-06, 'epoch': 1.71}) +[2026-06-19 20:13:59,074][starVLA.training.train_starvla][INFO] - Step 4573, Loss: {'train/loss': 0.2228204607963562, 'train/grad_norm_pre_clip': 16.602916717529297, 'timing/data': 0.017964055994525552, 'timing/model': 1.1446275810012594, 'learning_rate/qwen_vl_interface': 6.76893866236561e-07, 'learning_rate/action_model': 6.7689386623656094e-06, 'epoch': 1.71}) +[2026-06-19 20:14:03,574][starVLA.training.train_starvla][INFO] - Step 4574, Loss: {'train/loss': 0.27103641629219055, 'train/grad_norm_pre_clip': 30.221569061279297, 'timing/data': 0.017130049993284047, 'timing/model': 1.1465477879391983, 'learning_rate/qwen_vl_interface': 6.76071444825923e-07, 'learning_rate/action_model': 6.76071444825923e-06, 'epoch': 1.72}) +[2026-06-19 20:14:08,188][starVLA.training.train_starvla][INFO] - Step 4575, Loss: {'train/loss': 0.08134865760803223, 'train/grad_norm_pre_clip': 8.507098197937012, 'timing/data': 0.01778419001493603, 'timing/model': 1.1605411130003631, 'learning_rate/qwen_vl_interface': 6.752509035845413e-07, 'learning_rate/action_model': 6.7525090358454126e-06, 'epoch': 1.72}) +[2026-06-19 20:14:12,822][starVLA.training.train_starvla][INFO] - Step 4576, Loss: {'train/loss': 0.08006499707698822, 'train/grad_norm_pre_clip': 8.052261352539062, 'timing/data': 0.016486592940054834, 'timing/model': 1.330187552026473, 'learning_rate/qwen_vl_interface': 6.744322428497095e-07, 'learning_rate/action_model': 6.744322428497095e-06, 'epoch': 1.72}) +[2026-06-19 20:14:17,261][starVLA.training.train_starvla][INFO] - Step 4577, Loss: {'train/loss': 0.4904974102973938, 'train/grad_norm_pre_clip': 35.410552978515625, 'timing/data': 0.016930105979554355, 'timing/model': 1.148089752998203, 'learning_rate/qwen_vl_interface': 6.736154629579487e-07, 'learning_rate/action_model': 6.736154629579487e-06, 'epoch': 1.72}) +[2026-06-19 20:14:21,891][starVLA.training.train_starvla][INFO] - Step 4578, Loss: {'train/loss': 0.15222595632076263, 'train/grad_norm_pre_clip': 19.89976692199707, 'timing/data': 0.01800095895305276, 'timing/model': 1.1540620860178024, 'learning_rate/qwen_vl_interface': 6.728005642450055e-07, 'learning_rate/action_model': 6.7280056424500554e-06, 'epoch': 1.72}) +[2026-06-19 20:14:26,513][starVLA.training.train_starvla][INFO] - Step 4579, Loss: {'train/loss': 0.11643090099096298, 'train/grad_norm_pre_clip': 8.866247177124023, 'timing/data': 0.018204547930508852, 'timing/model': 1.1460132399806753, 'learning_rate/qwen_vl_interface': 6.719875470458553e-07, 'learning_rate/action_model': 6.719875470458553e-06, 'epoch': 1.72}) +[2026-06-19 20:14:31,141][starVLA.training.train_starvla][INFO] - Step 4580, Loss: {'train/loss': 0.1706390678882599, 'train/grad_norm_pre_clip': 15.849614143371582, 'timing/data': 0.01683845103252679, 'timing/model': 1.1546666939975694, 'learning_rate/qwen_vl_interface': 6.711764116946972e-07, 'learning_rate/action_model': 6.7117641169469715e-06, 'epoch': 1.72}) +[2026-06-19 20:14:35,798][starVLA.training.train_starvla][INFO] - Step 4581, Loss: {'train/loss': 0.36753949522972107, 'train/grad_norm_pre_clip': 26.364830017089844, 'timing/data': 0.0181918729795143, 'timing/model': 1.1691318000666797, 'learning_rate/qwen_vl_interface': 6.703671585249587e-07, 'learning_rate/action_model': 6.703671585249586e-06, 'epoch': 1.72}) +[2026-06-19 20:14:40,425][starVLA.training.train_starvla][INFO] - Step 4582, Loss: {'train/loss': 0.07939328998327255, 'train/grad_norm_pre_clip': 5.667255878448486, 'timing/data': 0.0179634460946545, 'timing/model': 1.3338252800749615, 'learning_rate/qwen_vl_interface': 6.69559787869294e-07, 'learning_rate/action_model': 6.69559787869294e-06, 'epoch': 1.72}) +[2026-06-19 20:14:44,850][starVLA.training.train_starvla][INFO] - Step 4583, Loss: {'train/loss': 0.0907423123717308, 'train/grad_norm_pre_clip': 11.423079490661621, 'timing/data': 0.018275383044965565, 'timing/model': 1.1468087600078434, 'learning_rate/qwen_vl_interface': 6.68754300059582e-07, 'learning_rate/action_model': 6.687543000595819e-06, 'epoch': 1.72}) +[2026-06-19 20:14:49,453][starVLA.training.train_starvla][INFO] - Step 4584, Loss: {'train/loss': 0.14685173332691193, 'train/grad_norm_pre_clip': 11.03945541381836, 'timing/data': 0.01768783398438245, 'timing/model': 1.1519884320441633, 'learning_rate/qwen_vl_interface': 6.67950695426929e-07, 'learning_rate/action_model': 6.679506954269291e-06, 'epoch': 1.72}) +[2026-06-19 20:14:54,052][starVLA.training.train_starvla][INFO] - Step 4585, Loss: {'train/loss': 0.17793433368206024, 'train/grad_norm_pre_clip': 16.27923011779785, 'timing/data': 0.01738233189098537, 'timing/model': 1.1479892830830067, 'learning_rate/qwen_vl_interface': 6.671489743016651e-07, 'learning_rate/action_model': 6.671489743016651e-06, 'epoch': 1.72}) +[2026-06-19 20:14:58,602][starVLA.training.train_starvla][INFO] - Step 4586, Loss: {'train/loss': 0.17531919479370117, 'train/grad_norm_pre_clip': 14.020862579345703, 'timing/data': 0.01771594292949885, 'timing/model': 1.1476712519070134, 'learning_rate/qwen_vl_interface': 6.663491370133487e-07, 'learning_rate/action_model': 6.663491370133487e-06, 'epoch': 1.72}) +[2026-06-19 20:15:03,193][starVLA.training.train_starvla][INFO] - Step 4587, Loss: {'train/loss': 0.18401369452476501, 'train/grad_norm_pre_clip': 16.194244384765625, 'timing/data': 0.01772116997744888, 'timing/model': 1.147383262985386, 'learning_rate/qwen_vl_interface': 6.655511838907629e-07, 'learning_rate/action_model': 6.655511838907629e-06, 'epoch': 1.72}) +[2026-06-19 20:15:07,802][starVLA.training.train_starvla][INFO] - Step 4588, Loss: {'train/loss': 0.20016463100910187, 'train/grad_norm_pre_clip': 17.216201782226562, 'timing/data': 0.017721344018355012, 'timing/model': 1.3230356549611315, 'learning_rate/qwen_vl_interface': 6.647551152619149e-07, 'learning_rate/action_model': 6.647551152619148e-06, 'epoch': 1.72}) +[2026-06-19 20:15:12,256][starVLA.training.train_starvla][INFO] - Step 4589, Loss: {'train/loss': 0.271822065114975, 'train/grad_norm_pre_clip': 21.588062286376953, 'timing/data': 0.018846986000426114, 'timing/model': 1.1536183999851346, 'learning_rate/qwen_vl_interface': 6.639609314540396e-07, 'learning_rate/action_model': 6.639609314540396e-06, 'epoch': 1.72}) +[2026-06-19 20:15:16,870][starVLA.training.train_starvla][INFO] - Step 4590, Loss: {'train/loss': 0.2543400824069977, 'train/grad_norm_pre_clip': 13.355337142944336, 'timing/data': 0.01841053506359458, 'timing/model': 1.1459493309957907, 'learning_rate/qwen_vl_interface': 6.631686327935954e-07, 'learning_rate/action_model': 6.631686327935954e-06, 'epoch': 1.72}) +[2026-06-19 20:15:21,491][starVLA.training.train_starvla][INFO] - Step 4591, Loss: {'train/loss': 0.08922550082206726, 'train/grad_norm_pre_clip': 14.209964752197266, 'timing/data': 0.017591879004612565, 'timing/model': 1.1532530429540202, 'learning_rate/qwen_vl_interface': 6.623782196062665e-07, 'learning_rate/action_model': 6.623782196062664e-06, 'epoch': 1.72}) +[2026-06-19 20:15:25,989][starVLA.training.train_starvla][INFO] - Step 4592, Loss: {'train/loss': 0.4409477412700653, 'train/grad_norm_pre_clip': 25.932321548461914, 'timing/data': 0.016849567997269332, 'timing/model': 1.1504063260508701, 'learning_rate/qwen_vl_interface': 6.615896922169622e-07, 'learning_rate/action_model': 6.615896922169622e-06, 'epoch': 1.72}) +[2026-06-19 20:15:30,524][starVLA.training.train_starvla][INFO] - Step 4593, Loss: {'train/loss': 0.402432918548584, 'train/grad_norm_pre_clip': 37.70150375366211, 'timing/data': 0.017442151089198887, 'timing/model': 1.145769224036485, 'learning_rate/qwen_vl_interface': 6.608030509498156e-07, 'learning_rate/action_model': 6.608030509498156e-06, 'epoch': 1.72}) +[2026-06-19 20:15:35,123][starVLA.training.train_starvla][INFO] - Step 4594, Loss: {'train/loss': 0.1915014386177063, 'train/grad_norm_pre_clip': 12.460610389709473, 'timing/data': 0.018447092967107892, 'timing/model': 1.3194914889754727, 'learning_rate/qwen_vl_interface': 6.600182961281865e-07, 'learning_rate/action_model': 6.600182961281864e-06, 'epoch': 1.72}) +[2026-06-19 20:15:39,556][starVLA.training.train_starvla][INFO] - Step 4595, Loss: {'train/loss': 0.31166014075279236, 'train/grad_norm_pre_clip': 15.270047187805176, 'timing/data': 0.0186080620624125, 'timing/model': 1.1517809430370107, 'learning_rate/qwen_vl_interface': 6.592354280746561e-07, 'learning_rate/action_model': 6.5923542807465595e-06, 'epoch': 1.72}) +[2026-06-19 20:15:44,154][starVLA.training.train_starvla][INFO] - Step 4596, Loss: {'train/loss': 0.15946072340011597, 'train/grad_norm_pre_clip': 24.293363571166992, 'timing/data': 0.01721979700960219, 'timing/model': 1.1471885290229693, 'learning_rate/qwen_vl_interface': 6.584544471110335e-07, 'learning_rate/action_model': 6.584544471110335e-06, 'epoch': 1.72}) +[2026-06-19 20:15:48,742][starVLA.training.train_starvla][INFO] - Step 4597, Loss: {'train/loss': 0.11036413908004761, 'train/grad_norm_pre_clip': 12.777247428894043, 'timing/data': 0.017107454943470657, 'timing/model': 1.143136856961064, 'learning_rate/qwen_vl_interface': 6.576753535583495e-07, 'learning_rate/action_model': 6.576753535583495e-06, 'epoch': 1.72}) +[2026-06-19 20:15:53,232][starVLA.training.train_starvla][INFO] - Step 4598, Loss: {'train/loss': 0.09948557615280151, 'train/grad_norm_pre_clip': 16.320959091186523, 'timing/data': 0.01772683102171868, 'timing/model': 1.145527119981125, 'learning_rate/qwen_vl_interface': 6.5689814773686e-07, 'learning_rate/action_model': 6.568981477368599e-06, 'epoch': 1.72}) +[2026-06-19 20:15:57,785][starVLA.training.train_starvla][INFO] - Step 4599, Loss: {'train/loss': 0.34272798895835876, 'train/grad_norm_pre_clip': 17.356515884399414, 'timing/data': 0.01827242609579116, 'timing/model': 1.1542016729945317, 'learning_rate/qwen_vl_interface': 6.56122829966046e-07, 'learning_rate/action_model': 6.56122829966046e-06, 'epoch': 1.72}) +[2026-06-19 20:16:02,384][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 4600 +[2026-06-19 20:18:58,715][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 4600 +[2026-06-19 20:18:58,715][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 4600 +[2026-06-19 20:22:09,245][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 4600 +[2026-06-19 20:22:09,246][starVLA.training.train_starvla][INFO] - Step 4600, Loss: {'train/loss': 0.06490624696016312, 'train/grad_norm_pre_clip': 20.992839813232422, 'eval/loss': 0.1282961517572403, 'eval/latency_0/loss': 0.08242125809192657, 'eval/latency_1/loss': 0.04794701188802719, 'eval/latency_2/loss': 0.11229762434959412, 'eval/latency_3/loss': 0.15625445544719696, 'eval/latency_4/loss': 0.24379336833953857, 'eval/flappy/loss': 0.1282961517572403, 'eval/flappy/latency_0/loss': 0.08242125809192657, 'eval/flappy/latency_1/loss': 0.04794701188802719, 'eval/flappy/latency_2/loss': 0.11229762434959412, 'eval/flappy/latency_3/loss': 0.15625445544719696, 'eval/flappy/latency_4/loss': 0.24379336833953857, 'eval/flap_precision': 0.7452471482889734, 'eval/flap_recall': 0.7127272727272728, 'eval/flap_f1': 0.7286245353159851, 'eval/flap_rate_model': 0.07305555555555555, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.7286245353159851, 'eval/latency_0/flap_precision': 0.7452471482889734, 'eval/latency_0/flap_recall': 0.7127272727272728, 'eval/latency_0/flap_f1': 0.7286245353159851, 'eval/latency_0/flap_rate_model': 0.07305555555555555, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.7286245353159851, 'timing/data': 0.01726944698020816, 'timing/model': 1.3176737159956247, 'learning_rate/qwen_vl_interface': 6.553494005646101e-07, 'learning_rate/action_model': 6.553494005646101e-06, 'epoch': 1.72}) +[2026-06-19 20:22:09,247][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_4600_state +[2026-06-19 20:22:09,249][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-19 20:22:12,636][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_4600_state/model.safetensors +[2026-06-19 20:22:19,822][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_4600_state/optimizer.bin +[2026-06-19 20:22:19,823][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_4600_state/sampler.bin +[2026-06-19 20:22:19,823][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_4600_state/sampler_1.bin +[2026-06-19 20:22:19,823][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_4600_state/random_states_0.pkl +[2026-06-19 20:22:19,847][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4600 +[2026-06-19 20:22:19,847][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-19 20:22:19,873][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-19 20:22:24,371][starVLA.training.train_starvla][INFO] - Step 4601, Loss: {'train/loss': 0.09527207165956497, 'train/grad_norm_pre_clip': 13.819968223571777, 'timing/data': 0.021311502903699875, 'timing/model': 1.133574888925068, 'learning_rate/qwen_vl_interface': 6.545778598504807e-07, 'learning_rate/action_model': 6.545778598504807e-06, 'epoch': 1.73}) +[2026-06-19 20:22:28,853][starVLA.training.train_starvla][INFO] - Step 4602, Loss: {'train/loss': 0.20548014342784882, 'train/grad_norm_pre_clip': 20.270584106445312, 'timing/data': 0.020564089994877577, 'timing/model': 1.2636510449228808, 'learning_rate/qwen_vl_interface': 6.538082081408093e-07, 'learning_rate/action_model': 6.538082081408093e-06, 'epoch': 1.73}) +[2026-06-19 20:22:33,209][starVLA.training.train_starvla][INFO] - Step 4603, Loss: {'train/loss': 0.07875882834196091, 'train/grad_norm_pre_clip': 7.427894115447998, 'timing/data': 0.020828203996643424, 'timing/model': 1.128478787955828, 'learning_rate/qwen_vl_interface': 6.530404457519697e-07, 'learning_rate/action_model': 6.530404457519697e-06, 'epoch': 1.73}) +[2026-06-19 20:22:37,751][starVLA.training.train_starvla][INFO] - Step 4604, Loss: {'train/loss': 0.06687459349632263, 'train/grad_norm_pre_clip': 8.148636817932129, 'timing/data': 0.02084801101591438, 'timing/model': 1.1402513069333509, 'learning_rate/qwen_vl_interface': 6.522745729995614e-07, 'learning_rate/action_model': 6.522745729995613e-06, 'epoch': 1.73}) +[2026-06-19 20:22:42,301][starVLA.training.train_starvla][INFO] - Step 4605, Loss: {'train/loss': 0.11035604029893875, 'train/grad_norm_pre_clip': 15.760574340820312, 'timing/data': 0.021141464007087052, 'timing/model': 1.1326112999813631, 'learning_rate/qwen_vl_interface': 6.515105901984037e-07, 'learning_rate/action_model': 6.515105901984037e-06, 'epoch': 1.73}) +[2026-06-19 20:22:46,820][starVLA.training.train_starvla][INFO] - Step 4606, Loss: {'train/loss': 0.35561057925224304, 'train/grad_norm_pre_clip': 19.345623016357422, 'timing/data': 0.020790796959772706, 'timing/model': 1.1377958480734378, 'learning_rate/qwen_vl_interface': 6.50748497662543e-07, 'learning_rate/action_model': 6.507484976625429e-06, 'epoch': 1.73}) +[2026-06-19 20:22:51,400][starVLA.training.train_starvla][INFO] - Step 4607, Loss: {'train/loss': 0.0930614247918129, 'train/grad_norm_pre_clip': 9.669379234313965, 'timing/data': 0.02133242296986282, 'timing/model': 1.1445285769877955, 'learning_rate/qwen_vl_interface': 6.499882957052459e-07, 'learning_rate/action_model': 6.4998829570524586e-06, 'epoch': 1.73}) +[2026-06-19 20:22:55,974][starVLA.training.train_starvla][INFO] - Step 4608, Loss: {'train/loss': 0.18069583177566528, 'train/grad_norm_pre_clip': 16.331527709960938, 'timing/data': 0.023330571013502777, 'timing/model': 1.3079282119870186, 'learning_rate/qwen_vl_interface': 6.492299846390024e-07, 'learning_rate/action_model': 6.492299846390023e-06, 'epoch': 1.73}) +[2026-06-19 20:23:00,390][starVLA.training.train_starvla][INFO] - Step 4609, Loss: {'train/loss': 0.16243375837802887, 'train/grad_norm_pre_clip': 14.017099380493164, 'timing/data': 0.021138911950401962, 'timing/model': 1.1398973379982635, 'learning_rate/qwen_vl_interface': 6.484735647755259e-07, 'learning_rate/action_model': 6.48473564775526e-06, 'epoch': 1.73}) +[2026-06-19 20:23:04,949][starVLA.training.train_starvla][INFO] - Step 4610, Loss: {'train/loss': 0.11626115441322327, 'train/grad_norm_pre_clip': 10.679265022277832, 'timing/data': 0.02120509697124362, 'timing/model': 1.1364038129104301, 'learning_rate/qwen_vl_interface': 6.477190364257524e-07, 'learning_rate/action_model': 6.477190364257524e-06, 'epoch': 1.73}) +[2026-06-19 20:23:09,446][starVLA.training.train_starvla][INFO] - Step 4611, Loss: {'train/loss': 0.22607386112213135, 'train/grad_norm_pre_clip': 21.434083938598633, 'timing/data': 0.02240372495725751, 'timing/model': 1.1478557479567826, 'learning_rate/qwen_vl_interface': 6.469663998998393e-07, 'learning_rate/action_model': 6.469663998998393e-06, 'epoch': 1.73}) +[2026-06-19 20:23:13,974][starVLA.training.train_starvla][INFO] - Step 4612, Loss: {'train/loss': 0.13309700787067413, 'train/grad_norm_pre_clip': 15.432572364807129, 'timing/data': 0.021960483049042523, 'timing/model': 1.1395583159755915, 'learning_rate/qwen_vl_interface': 6.462156555071681e-07, 'learning_rate/action_model': 6.46215655507168e-06, 'epoch': 1.73}) +[2026-06-19 20:23:18,553][starVLA.training.train_starvla][INFO] - Step 4613, Loss: {'train/loss': 0.1844559609889984, 'train/grad_norm_pre_clip': 13.445881843566895, 'timing/data': 0.021021505002863705, 'timing/model': 1.1367022100603208, 'learning_rate/qwen_vl_interface': 6.454668035563401e-07, 'learning_rate/action_model': 6.454668035563401e-06, 'epoch': 1.73}) +[2026-06-19 20:23:23,138][starVLA.training.train_starvla][INFO] - Step 4614, Loss: {'train/loss': 0.3531567454338074, 'train/grad_norm_pre_clip': 17.311208724975586, 'timing/data': 0.02041890495456755, 'timing/model': 1.3196880370378494, 'learning_rate/qwen_vl_interface': 6.447198443551809e-07, 'learning_rate/action_model': 6.4471984435518075e-06, 'epoch': 1.73}) +[2026-06-19 20:23:27,566][starVLA.training.train_starvla][INFO] - Step 4615, Loss: {'train/loss': 0.2933248281478882, 'train/grad_norm_pre_clip': 14.017594337463379, 'timing/data': 0.020605220925062895, 'timing/model': 1.1508843860356137, 'learning_rate/qwen_vl_interface': 6.439747782107373e-07, 'learning_rate/action_model': 6.4397477821073735e-06, 'epoch': 1.73}) +[2026-06-19 20:23:32,176][starVLA.training.train_starvla][INFO] - Step 4616, Loss: {'train/loss': 0.07582063972949982, 'train/grad_norm_pre_clip': 6.137159824371338, 'timing/data': 0.020471952040679753, 'timing/model': 1.1434868819778785, 'learning_rate/qwen_vl_interface': 6.432316054292779e-07, 'learning_rate/action_model': 6.43231605429278e-06, 'epoch': 1.73}) +[2026-06-19 20:23:36,704][starVLA.training.train_starvla][INFO] - Step 4617, Loss: {'train/loss': 0.2819059491157532, 'train/grad_norm_pre_clip': 19.417316436767578, 'timing/data': 0.02090611110907048, 'timing/model': 1.1404118560021743, 'learning_rate/qwen_vl_interface': 6.424903263162924e-07, 'learning_rate/action_model': 6.424903263162923e-06, 'epoch': 1.73}) +[2026-06-19 20:23:41,266][starVLA.training.train_starvla][INFO] - Step 4618, Loss: {'train/loss': 0.1944882869720459, 'train/grad_norm_pre_clip': 13.769803047180176, 'timing/data': 0.021492307074368, 'timing/model': 1.151477882056497, 'learning_rate/qwen_vl_interface': 6.417509411764924e-07, 'learning_rate/action_model': 6.4175094117649244e-06, 'epoch': 1.73}) +[2026-06-19 20:23:45,864][starVLA.training.train_starvla][INFO] - Step 4619, Loss: {'train/loss': 0.21211865544319153, 'train/grad_norm_pre_clip': 27.27455711364746, 'timing/data': 0.021737382048740983, 'timing/model': 1.1402211859822273, 'learning_rate/qwen_vl_interface': 6.410134503138122e-07, 'learning_rate/action_model': 6.4101345031381224e-06, 'epoch': 1.73}) +[2026-06-19 20:23:50,443][starVLA.training.train_starvla][INFO] - Step 4620, Loss: {'train/loss': 0.11759204417467117, 'train/grad_norm_pre_clip': 16.678449630737305, 'timing/data': 0.021288966992869973, 'timing/model': 1.3148556730011478, 'learning_rate/qwen_vl_interface': 6.402778540314054e-07, 'learning_rate/action_model': 6.402778540314054e-06, 'epoch': 1.73}) +[2026-06-19 20:23:54,883][starVLA.training.train_starvla][INFO] - Step 4621, Loss: {'train/loss': 0.3024669289588928, 'train/grad_norm_pre_clip': 21.97547721862793, 'timing/data': 0.021017541992478073, 'timing/model': 1.1476304869865999, 'learning_rate/qwen_vl_interface': 6.39544152631649e-07, 'learning_rate/action_model': 6.3954415263164895e-06, 'epoch': 1.73}) +[2026-06-19 20:23:59,491][starVLA.training.train_starvla][INFO] - Step 4622, Loss: {'train/loss': 0.12375594675540924, 'train/grad_norm_pre_clip': 14.69864273071289, 'timing/data': 0.021387367974966764, 'timing/model': 1.149638130911626, 'learning_rate/qwen_vl_interface': 6.388123464161388e-07, 'learning_rate/action_model': 6.388123464161389e-06, 'epoch': 1.73}) +[2026-06-19 20:24:04,044][starVLA.training.train_starvla][INFO] - Step 4623, Loss: {'train/loss': 0.037817783653736115, 'train/grad_norm_pre_clip': 5.546606540679932, 'timing/data': 0.020995957078412175, 'timing/model': 1.1425012489780784, 'learning_rate/qwen_vl_interface': 6.380824356856936e-07, 'learning_rate/action_model': 6.380824356856936e-06, 'epoch': 1.73}) +[2026-06-19 20:24:08,647][starVLA.training.train_starvla][INFO] - Step 4624, Loss: {'train/loss': 0.2392575591802597, 'train/grad_norm_pre_clip': 27.051067352294922, 'timing/data': 0.021417491021566093, 'timing/model': 1.1540859049418941, 'learning_rate/qwen_vl_interface': 6.373544207403513e-07, 'learning_rate/action_model': 6.373544207403513e-06, 'epoch': 1.73}) +[2026-06-19 20:24:13,233][starVLA.training.train_starvla][INFO] - Step 4625, Loss: {'train/loss': 0.11656832695007324, 'train/grad_norm_pre_clip': 13.993165969848633, 'timing/data': 0.020699777058325708, 'timing/model': 1.1423444979591295, 'learning_rate/qwen_vl_interface': 6.366283018793723e-07, 'learning_rate/action_model': 6.366283018793723e-06, 'epoch': 1.73}) +[2026-06-19 20:24:17,868][starVLA.training.train_starvla][INFO] - Step 4626, Loss: {'train/loss': 0.2487581968307495, 'train/grad_norm_pre_clip': 8.709461212158203, 'timing/data': 0.02094550197944045, 'timing/model': 1.3391816230723634, 'learning_rate/qwen_vl_interface': 6.359040794012358e-07, 'learning_rate/action_model': 6.359040794012358e-06, 'epoch': 1.73}) +[2026-06-19 20:24:22,288][starVLA.training.train_starvla][INFO] - Step 4627, Loss: {'train/loss': 0.16644184291362762, 'train/grad_norm_pre_clip': 15.494028091430664, 'timing/data': 0.020864987978711724, 'timing/model': 1.1442964000161737, 'learning_rate/qwen_vl_interface': 6.351817536036431e-07, 'learning_rate/action_model': 6.351817536036431e-06, 'epoch': 1.73}) +[2026-06-19 20:24:26,894][starVLA.training.train_starvla][INFO] - Step 4628, Loss: {'train/loss': 0.1193917766213417, 'train/grad_norm_pre_clip': 8.978302955627441, 'timing/data': 0.020091493031941354, 'timing/model': 1.1485454390058294, 'learning_rate/qwen_vl_interface': 6.344613247835135e-07, 'learning_rate/action_model': 6.344613247835135e-06, 'epoch': 1.74}) +[2026-06-19 20:24:31,490][starVLA.training.train_starvla][INFO] - Step 4629, Loss: {'train/loss': 0.24915049970149994, 'train/grad_norm_pre_clip': 13.686715126037598, 'timing/data': 0.021446786005981266, 'timing/model': 1.1450095149921253, 'learning_rate/qwen_vl_interface': 6.337427932369898e-07, 'learning_rate/action_model': 6.337427932369898e-06, 'epoch': 1.74}) +[2026-06-19 20:24:36,079][starVLA.training.train_starvla][INFO] - Step 4630, Loss: {'train/loss': 0.23415164649486542, 'train/grad_norm_pre_clip': 25.190765380859375, 'timing/data': 0.021962802042253315, 'timing/model': 1.153391876956448, 'learning_rate/qwen_vl_interface': 6.330261592594328e-07, 'learning_rate/action_model': 6.330261592594327e-06, 'epoch': 1.74}) +[2026-06-19 20:24:40,679][starVLA.training.train_starvla][INFO] - Step 4631, Loss: {'train/loss': 0.1238999217748642, 'train/grad_norm_pre_clip': 19.559560775756836, 'timing/data': 0.021361729013733566, 'timing/model': 1.154957243008539, 'learning_rate/qwen_vl_interface': 6.323114231454229e-07, 'learning_rate/action_model': 6.323114231454229e-06, 'epoch': 1.74}) +[2026-06-19 20:24:45,279][starVLA.training.train_starvla][INFO] - Step 4632, Loss: {'train/loss': 0.1338282972574234, 'train/grad_norm_pre_clip': 10.5565767288208, 'timing/data': 0.022536140982992947, 'timing/model': 1.1502219040412456, 'learning_rate/qwen_vl_interface': 6.315985851887618e-07, 'learning_rate/action_model': 6.315985851887618e-06, 'epoch': 1.74}) +[2026-06-19 20:24:49,876][starVLA.training.train_starvla][INFO] - Step 4633, Loss: {'train/loss': 0.15813203155994415, 'train/grad_norm_pre_clip': 12.82598876953125, 'timing/data': 0.021933097043074667, 'timing/model': 1.3142942569684237, 'learning_rate/qwen_vl_interface': 6.308876456824697e-07, 'learning_rate/action_model': 6.3088764568246975e-06, 'epoch': 1.74}) +[2026-06-19 20:24:54,316][starVLA.training.train_starvla][INFO] - Step 4634, Loss: {'train/loss': 0.21355706453323364, 'train/grad_norm_pre_clip': 13.510714530944824, 'timing/data': 0.02126622200012207, 'timing/model': 1.1611758369253948, 'learning_rate/qwen_vl_interface': 6.301786049187875e-07, 'learning_rate/action_model': 6.3017860491878745e-06, 'epoch': 1.74}) +[2026-06-19 20:24:58,950][starVLA.training.train_starvla][INFO] - Step 4635, Loss: {'train/loss': 0.15069448947906494, 'train/grad_norm_pre_clip': 29.220664978027344, 'timing/data': 0.021117537980899215, 'timing/model': 1.1472191479988396, 'learning_rate/qwen_vl_interface': 6.294714631891755e-07, 'learning_rate/action_model': 6.294714631891754e-06, 'epoch': 1.74}) +[2026-06-19 20:25:03,558][starVLA.training.train_starvla][INFO] - Step 4636, Loss: {'train/loss': 0.10142023861408234, 'train/grad_norm_pre_clip': 13.904815673828125, 'timing/data': 0.021411373047158122, 'timing/model': 1.1492254380136728, 'learning_rate/qwen_vl_interface': 6.287662207843115e-07, 'learning_rate/action_model': 6.287662207843114e-06, 'epoch': 1.74}) +[2026-06-19 20:25:08,115][starVLA.training.train_starvla][INFO] - Step 4637, Loss: {'train/loss': 0.2067490965127945, 'train/grad_norm_pre_clip': 10.85057258605957, 'timing/data': 0.021345002111047506, 'timing/model': 1.154895311105065, 'learning_rate/qwen_vl_interface': 6.280628779940956e-07, 'learning_rate/action_model': 6.280628779940956e-06, 'epoch': 1.74}) +[2026-06-19 20:25:12,733][starVLA.training.train_starvla][INFO] - Step 4638, Loss: {'train/loss': 0.10272099077701569, 'train/grad_norm_pre_clip': 10.539921760559082, 'timing/data': 0.021244930918328464, 'timing/model': 1.1570993980858475, 'learning_rate/qwen_vl_interface': 6.273614351076445e-07, 'learning_rate/action_model': 6.273614351076445e-06, 'epoch': 1.74}) +[2026-06-19 20:25:17,369][starVLA.training.train_starvla][INFO] - Step 4639, Loss: {'train/loss': 0.22510655224323273, 'train/grad_norm_pre_clip': 17.32982635498047, 'timing/data': 0.02174727595411241, 'timing/model': 1.3513085950398818, 'learning_rate/qwen_vl_interface': 6.266618924132951e-07, 'learning_rate/action_model': 6.266618924132951e-06, 'epoch': 1.74}) +[2026-06-19 20:25:21,810][starVLA.training.train_starvla][INFO] - Step 4640, Loss: {'train/loss': 0.4020518958568573, 'train/grad_norm_pre_clip': 21.687368392944336, 'timing/data': 0.021160202100872993, 'timing/model': 1.1515266860369593, 'learning_rate/qwen_vl_interface': 6.25964250198603e-07, 'learning_rate/action_model': 6.2596425019860294e-06, 'epoch': 1.74}) +[2026-06-19 20:25:26,437][starVLA.training.train_starvla][INFO] - Step 4641, Loss: {'train/loss': 0.18096835911273956, 'train/grad_norm_pre_clip': 23.31078338623047, 'timing/data': 0.021194008993916214, 'timing/model': 1.143066007993184, 'learning_rate/qwen_vl_interface': 6.25268508750342e-07, 'learning_rate/action_model': 6.25268508750342e-06, 'epoch': 1.74}) +[2026-06-19 20:25:30,975][starVLA.training.train_starvla][INFO] - Step 4642, Loss: {'train/loss': 0.4437454640865326, 'train/grad_norm_pre_clip': 24.235809326171875, 'timing/data': 0.020542641985230148, 'timing/model': 1.1463619159767404, 'learning_rate/qwen_vl_interface': 6.245746683545062e-07, 'learning_rate/action_model': 6.245746683545062e-06, 'epoch': 1.74}) +[2026-06-19 20:25:35,596][starVLA.training.train_starvla][INFO] - Step 4643, Loss: {'train/loss': 0.031876713037490845, 'train/grad_norm_pre_clip': 4.594974517822266, 'timing/data': 0.02128505299333483, 'timing/model': 1.1554497890174389, 'learning_rate/qwen_vl_interface': 6.238827292963063e-07, 'learning_rate/action_model': 6.238827292963063e-06, 'epoch': 1.74}) +[2026-06-19 20:25:40,214][starVLA.training.train_starvla][INFO] - Step 4644, Loss: {'train/loss': 0.1560247838497162, 'train/grad_norm_pre_clip': 14.782835960388184, 'timing/data': 0.021452625980600715, 'timing/model': 1.1552455929340795, 'learning_rate/qwen_vl_interface': 6.231926918601731e-07, 'learning_rate/action_model': 6.2319269186017305e-06, 'epoch': 1.74}) +[2026-06-19 20:25:44,824][starVLA.training.train_starvla][INFO] - Step 4645, Loss: {'train/loss': 0.26987823843955994, 'train/grad_norm_pre_clip': 18.198402404785156, 'timing/data': 0.02055805001873523, 'timing/model': 1.3207526380429044, 'learning_rate/qwen_vl_interface': 6.22504556329754e-07, 'learning_rate/action_model': 6.22504556329754e-06, 'epoch': 1.74}) +[2026-06-19 20:25:49,257][starVLA.training.train_starvla][INFO] - Step 4646, Loss: {'train/loss': 0.37808889150619507, 'train/grad_norm_pre_clip': 23.216625213623047, 'timing/data': 0.021117256954312325, 'timing/model': 1.1506905650021508, 'learning_rate/qwen_vl_interface': 6.218183229879164e-07, 'learning_rate/action_model': 6.218183229879164e-06, 'epoch': 1.74}) +[2026-06-19 20:25:53,843][starVLA.training.train_starvla][INFO] - Step 4647, Loss: {'train/loss': 0.23214982450008392, 'train/grad_norm_pre_clip': 17.298343658447266, 'timing/data': 0.021282334113493562, 'timing/model': 1.1468287700554356, 'learning_rate/qwen_vl_interface': 6.211339921167447e-07, 'learning_rate/action_model': 6.211339921167446e-06, 'epoch': 1.74}) +[2026-06-19 20:25:58,470][starVLA.training.train_starvla][INFO] - Step 4648, Loss: {'train/loss': 0.2348468005657196, 'train/grad_norm_pre_clip': 10.64925479888916, 'timing/data': 0.02046336606144905, 'timing/model': 1.1570635149255395, 'learning_rate/qwen_vl_interface': 6.204515639975415e-07, 'learning_rate/action_model': 6.204515639975414e-06, 'epoch': 1.74}) +[2026-06-19 20:26:03,110][starVLA.training.train_starvla][INFO] - Step 4649, Loss: {'train/loss': 0.18907344341278076, 'train/grad_norm_pre_clip': 17.559234619140625, 'timing/data': 0.020862575038336217, 'timing/model': 1.1582059140782803, 'learning_rate/qwen_vl_interface': 6.197710389108262e-07, 'learning_rate/action_model': 6.197710389108262e-06, 'epoch': 1.74}) +[2026-06-19 20:26:07,707][starVLA.training.train_starvla][INFO] - Step 4650, Loss: {'train/loss': 0.12057037651538849, 'train/grad_norm_pre_clip': 10.692718505859375, 'timing/data': 0.02162304800003767, 'timing/model': 1.147895872942172, 'learning_rate/qwen_vl_interface': 6.190924171363378e-07, 'learning_rate/action_model': 6.190924171363378e-06, 'epoch': 1.74}) +[2026-06-19 20:26:12,245][starVLA.training.train_starvla][INFO] - Step 4651, Loss: {'train/loss': 0.3809612989425659, 'train/grad_norm_pre_clip': 26.38181495666504, 'timing/data': 0.02629081206396222, 'timing/model': 1.1680602219421417, 'learning_rate/qwen_vl_interface': 6.184156989530326e-07, 'learning_rate/action_model': 6.184156989530325e-06, 'epoch': 1.74}) +[2026-06-19 20:26:16,865][starVLA.training.train_starvla][INFO] - Step 4652, Loss: {'train/loss': 0.22753660380840302, 'train/grad_norm_pre_clip': 26.652189254760742, 'timing/data': 0.02102987701073289, 'timing/model': 1.3312959980685264, 'learning_rate/qwen_vl_interface': 6.17740884639083e-07, 'learning_rate/action_model': 6.177408846390829e-06, 'epoch': 1.74}) +[2026-06-19 20:26:21,287][starVLA.training.train_starvla][INFO] - Step 4653, Loss: {'train/loss': 0.3213317394256592, 'train/grad_norm_pre_clip': 21.66725730895996, 'timing/data': 0.021774086984805763, 'timing/model': 1.1463758909376338, 'learning_rate/qwen_vl_interface': 6.170679744718798e-07, 'learning_rate/action_model': 6.1706797447187975e-06, 'epoch': 1.74}) +[2026-06-19 20:26:25,852][starVLA.training.train_starvla][INFO] - Step 4654, Loss: {'train/loss': 0.17011664807796478, 'train/grad_norm_pre_clip': 17.42003631591797, 'timing/data': 0.021052652038633823, 'timing/model': 1.1431869089137763, 'learning_rate/qwen_vl_interface': 6.163969687280311e-07, 'learning_rate/action_model': 6.163969687280311e-06, 'epoch': 1.75}) +[2026-06-19 20:26:30,441][starVLA.training.train_starvla][INFO] - Step 4655, Loss: {'train/loss': 0.29220396280288696, 'train/grad_norm_pre_clip': 15.88931941986084, 'timing/data': 0.021608427981846035, 'timing/model': 1.1446948880329728, 'learning_rate/qwen_vl_interface': 6.157278676833627e-07, 'learning_rate/action_model': 6.157278676833627e-06, 'epoch': 1.75}) +[2026-06-19 20:26:35,048][starVLA.training.train_starvla][INFO] - Step 4656, Loss: {'train/loss': 0.2290031760931015, 'train/grad_norm_pre_clip': 17.305753707885742, 'timing/data': 0.020954024977982044, 'timing/model': 1.1562724099494517, 'learning_rate/qwen_vl_interface': 6.150606716129155e-07, 'learning_rate/action_model': 6.150606716129154e-06, 'epoch': 1.75}) +[2026-06-19 20:26:39,682][starVLA.training.train_starvla][INFO] - Step 4657, Loss: {'train/loss': 0.2305207997560501, 'train/grad_norm_pre_clip': 17.661876678466797, 'timing/data': 0.020804119063541293, 'timing/model': 1.340909396065399, 'learning_rate/qwen_vl_interface': 6.143953807909497e-07, 'learning_rate/action_model': 6.143953807909496e-06, 'epoch': 1.75}) +[2026-06-19 20:26:44,130][starVLA.training.train_starvla][INFO] - Step 4658, Loss: {'train/loss': 0.17631033062934875, 'train/grad_norm_pre_clip': 22.758867263793945, 'timing/data': 0.020939130103215575, 'timing/model': 1.1506263629999012, 'learning_rate/qwen_vl_interface': 6.137319954909407e-07, 'learning_rate/action_model': 6.137319954909406e-06, 'epoch': 1.75}) +[2026-06-19 20:26:48,736][starVLA.training.train_starvla][INFO] - Step 4659, Loss: {'train/loss': 0.24703939259052277, 'train/grad_norm_pre_clip': 23.13180160522461, 'timing/data': 0.02125276403967291, 'timing/model': 1.1512783060315996, 'learning_rate/qwen_vl_interface': 6.130705159855815e-07, 'learning_rate/action_model': 6.130705159855814e-06, 'epoch': 1.75}) +[2026-06-19 20:26:53,336][starVLA.training.train_starvla][INFO] - Step 4660, Loss: {'train/loss': 0.11911208182573318, 'train/grad_norm_pre_clip': 31.037912368774414, 'timing/data': 0.02128940401598811, 'timing/model': 1.153171642916277, 'learning_rate/qwen_vl_interface': 6.124109425467805e-07, 'learning_rate/action_model': 6.124109425467805e-06, 'epoch': 1.75}) +[2026-06-19 20:26:57,940][starVLA.training.train_starvla][INFO] - Step 4661, Loss: {'train/loss': 0.367814838886261, 'train/grad_norm_pre_clip': 25.605182647705078, 'timing/data': 0.02113192796241492, 'timing/model': 1.1508890150580555, 'learning_rate/qwen_vl_interface': 6.117532754456642e-07, 'learning_rate/action_model': 6.117532754456642e-06, 'epoch': 1.75}) +[2026-06-19 20:27:02,556][starVLA.training.train_starvla][INFO] - Step 4662, Loss: {'train/loss': 0.17403876781463623, 'train/grad_norm_pre_clip': 15.80652904510498, 'timing/data': 0.021425365936011076, 'timing/model': 1.1509151170030236, 'learning_rate/qwen_vl_interface': 6.110975149525746e-07, 'learning_rate/action_model': 6.110975149525746e-06, 'epoch': 1.75}) +[2026-06-19 20:27:07,186][starVLA.training.train_starvla][INFO] - Step 4663, Loss: {'train/loss': 0.23737625777721405, 'train/grad_norm_pre_clip': 10.85733699798584, 'timing/data': 0.021000833017751575, 'timing/model': 1.3212472830200568, 'learning_rate/qwen_vl_interface': 6.104436613370692e-07, 'learning_rate/action_model': 6.104436613370692e-06, 'epoch': 1.75}) +[2026-06-19 20:27:11,792][starVLA.training.train_starvla][INFO] - Step 4664, Loss: {'train/loss': 0.14593705534934998, 'train/grad_norm_pre_clip': 17.02823257446289, 'timing/data': 0.0214033069787547, 'timing/model': 1.3230081559158862, 'learning_rate/qwen_vl_interface': 6.09791714867924e-07, 'learning_rate/action_model': 6.0979171486792396e-06, 'epoch': 1.75}) +[2026-06-19 20:27:16,233][starVLA.training.train_starvla][INFO] - Step 4665, Loss: {'train/loss': 0.11908108741044998, 'train/grad_norm_pre_clip': 8.528472900390625, 'timing/data': 0.02111366600729525, 'timing/model': 1.1521138289244846, 'learning_rate/qwen_vl_interface': 6.091416758131287e-07, 'learning_rate/action_model': 6.091416758131286e-06, 'epoch': 1.75}) +[2026-06-19 20:27:20,862][starVLA.training.train_starvla][INFO] - Step 4666, Loss: {'train/loss': 0.18491509556770325, 'train/grad_norm_pre_clip': 14.980956077575684, 'timing/data': 0.02034537901636213, 'timing/model': 1.1487224920419976, 'learning_rate/qwen_vl_interface': 6.084935444398895e-07, 'learning_rate/action_model': 6.084935444398894e-06, 'epoch': 1.75}) +[2026-06-19 20:27:25,473][starVLA.training.train_starvla][INFO] - Step 4667, Loss: {'train/loss': 0.2233763188123703, 'train/grad_norm_pre_clip': 15.51646614074707, 'timing/data': 0.020902476040646434, 'timing/model': 1.1469158210093156, 'learning_rate/qwen_vl_interface': 6.078473210146294e-07, 'learning_rate/action_model': 6.078473210146294e-06, 'epoch': 1.75}) +[2026-06-19 20:27:29,978][starVLA.training.train_starvla][INFO] - Step 4668, Loss: {'train/loss': 0.07984282076358795, 'train/grad_norm_pre_clip': 9.713497161865234, 'timing/data': 0.020747140049934387, 'timing/model': 1.147763081942685, 'learning_rate/qwen_vl_interface': 6.072030058029862e-07, 'learning_rate/action_model': 6.072030058029861e-06, 'epoch': 1.75}) +[2026-06-19 20:27:34,551][starVLA.training.train_starvla][INFO] - Step 4669, Loss: {'train/loss': 0.10352315753698349, 'train/grad_norm_pre_clip': 13.826995849609375, 'timing/data': 0.020913321059197187, 'timing/model': 1.1544399349950254, 'learning_rate/qwen_vl_interface': 6.065605990698136e-07, 'learning_rate/action_model': 6.065605990698136e-06, 'epoch': 1.75}) +[2026-06-19 20:27:39,155][starVLA.training.train_starvla][INFO] - Step 4670, Loss: {'train/loss': 0.08153969794511795, 'train/grad_norm_pre_clip': 8.775945663452148, 'timing/data': 0.02105077204760164, 'timing/model': 1.31811698700767, 'learning_rate/qwen_vl_interface': 6.059201010791811e-07, 'learning_rate/action_model': 6.059201010791811e-06, 'epoch': 1.75}) +[2026-06-19 20:27:43,614][starVLA.training.train_starvla][INFO] - Step 4671, Loss: {'train/loss': 0.09602861851453781, 'train/grad_norm_pre_clip': 22.868736267089844, 'timing/data': 0.02129550208337605, 'timing/model': 1.1524152620695531, 'learning_rate/qwen_vl_interface': 6.052815120943721e-07, 'learning_rate/action_model': 6.052815120943721e-06, 'epoch': 1.75}) +[2026-06-19 20:27:48,240][starVLA.training.train_starvla][INFO] - Step 4672, Loss: {'train/loss': 0.29574787616729736, 'train/grad_norm_pre_clip': 22.883577346801758, 'timing/data': 0.02066417597234249, 'timing/model': 1.1499594600172713, 'learning_rate/qwen_vl_interface': 6.046448323778874e-07, 'learning_rate/action_model': 6.046448323778874e-06, 'epoch': 1.75}) +[2026-06-19 20:27:52,844][starVLA.training.train_starvla][INFO] - Step 4673, Loss: {'train/loss': 0.1864035278558731, 'train/grad_norm_pre_clip': 11.409812927246094, 'timing/data': 0.020667292061261833, 'timing/model': 1.1481770819518715, 'learning_rate/qwen_vl_interface': 6.040100621914419e-07, 'learning_rate/action_model': 6.040100621914418e-06, 'epoch': 1.75}) +[2026-06-19 20:27:57,461][starVLA.training.train_starvla][INFO] - Step 4674, Loss: {'train/loss': 0.3570335805416107, 'train/grad_norm_pre_clip': 21.473949432373047, 'timing/data': 0.02098234393633902, 'timing/model': 1.1566714009968564, 'learning_rate/qwen_vl_interface': 6.033772017959652e-07, 'learning_rate/action_model': 6.033772017959652e-06, 'epoch': 1.75}) +[2026-06-19 20:28:01,999][starVLA.training.train_starvla][INFO] - Step 4675, Loss: {'train/loss': 0.17472253739833832, 'train/grad_norm_pre_clip': 8.37251091003418, 'timing/data': 0.02166190999560058, 'timing/model': 1.1692103210370988, 'learning_rate/qwen_vl_interface': 6.027462514516022e-07, 'learning_rate/action_model': 6.0274625145160214e-06, 'epoch': 1.75}) +[2026-06-19 20:28:06,573][starVLA.training.train_starvla][INFO] - Step 4676, Loss: {'train/loss': 0.06366344541311264, 'train/grad_norm_pre_clip': 6.356656551361084, 'timing/data': 0.021368828020058572, 'timing/model': 1.2897156259277835, 'learning_rate/qwen_vl_interface': 6.021172114177139e-07, 'learning_rate/action_model': 6.021172114177139e-06, 'epoch': 1.75}) +[2026-06-19 20:28:11,017][starVLA.training.train_starvla][INFO] - Step 4677, Loss: {'train/loss': 0.14471393823623657, 'train/grad_norm_pre_clip': 12.324004173278809, 'timing/data': 0.021425351966172457, 'timing/model': 1.1539202210260555, 'learning_rate/qwen_vl_interface': 6.014900819528734e-07, 'learning_rate/action_model': 6.0149008195287335e-06, 'epoch': 1.75}) +[2026-06-19 20:28:15,648][starVLA.training.train_starvla][INFO] - Step 4678, Loss: {'train/loss': 0.19687481224536896, 'train/grad_norm_pre_clip': 19.264297485351562, 'timing/data': 0.021084815030917525, 'timing/model': 1.1582155270734802, 'learning_rate/qwen_vl_interface': 6.008648633148707e-07, 'learning_rate/action_model': 6.008648633148707e-06, 'epoch': 1.75}) +[2026-06-19 20:28:20,266][starVLA.training.train_starvla][INFO] - Step 4679, Loss: {'train/loss': 0.13836170732975006, 'train/grad_norm_pre_clip': 20.59642219543457, 'timing/data': 0.021598209044896066, 'timing/model': 1.156286692013964, 'learning_rate/qwen_vl_interface': 6.002415557607097e-07, 'learning_rate/action_model': 6.002415557607096e-06, 'epoch': 1.75}) +[2026-06-19 20:28:24,882][starVLA.training.train_starvla][INFO] - Step 4680, Loss: {'train/loss': 0.13988648355007172, 'train/grad_norm_pre_clip': 14.17779541015625, 'timing/data': 0.020740021020174026, 'timing/model': 1.1549209310906008, 'learning_rate/qwen_vl_interface': 5.996201595466088e-07, 'learning_rate/action_model': 5.996201595466088e-06, 'epoch': 1.75}) +[2026-06-19 20:28:29,496][starVLA.training.train_starvla][INFO] - Step 4681, Loss: {'train/loss': 0.15821479260921478, 'train/grad_norm_pre_clip': 14.769196510314941, 'timing/data': 0.020633280975744128, 'timing/model': 1.3213011500192806, 'learning_rate/qwen_vl_interface': 5.990006749279998e-07, 'learning_rate/action_model': 5.990006749279998e-06, 'epoch': 1.76}) +[2026-06-19 20:28:34,003][starVLA.training.train_starvla][INFO] - Step 4682, Loss: {'train/loss': 0.16261860728263855, 'train/grad_norm_pre_clip': 14.202799797058105, 'timing/data': 0.020925568998791277, 'timing/model': 1.210593490046449, 'learning_rate/qwen_vl_interface': 5.983831021595296e-07, 'learning_rate/action_model': 5.9838310215952956e-06, 'epoch': 1.76}) +[2026-06-19 20:28:38,428][starVLA.training.train_starvla][INFO] - Step 4683, Loss: {'train/loss': 0.19868628680706024, 'train/grad_norm_pre_clip': 15.491765022277832, 'timing/data': 0.020982322981581092, 'timing/model': 1.147499002981931, 'learning_rate/qwen_vl_interface': 5.977674414950595e-07, 'learning_rate/action_model': 5.9776744149505955e-06, 'epoch': 1.76}) +[2026-06-19 20:28:42,990][starVLA.training.train_starvla][INFO] - Step 4684, Loss: {'train/loss': 0.23655860126018524, 'train/grad_norm_pre_clip': 21.946569442749023, 'timing/data': 0.020950436941348016, 'timing/model': 1.1639022470917553, 'learning_rate/qwen_vl_interface': 5.971536931876635e-07, 'learning_rate/action_model': 5.971536931876634e-06, 'epoch': 1.76}) +[2026-06-19 20:28:47,601][starVLA.training.train_starvla][INFO] - Step 4685, Loss: {'train/loss': 0.1856255829334259, 'train/grad_norm_pre_clip': 12.425747871398926, 'timing/data': 0.02106269507203251, 'timing/model': 1.1523720240220428, 'learning_rate/qwen_vl_interface': 5.965418574896315e-07, 'learning_rate/action_model': 5.9654185748963145e-06, 'epoch': 1.76}) +[2026-06-19 20:28:52,189][starVLA.training.train_starvla][INFO] - Step 4686, Loss: {'train/loss': 0.17779304087162018, 'train/grad_norm_pre_clip': 24.605106353759766, 'timing/data': 0.020984083996154368, 'timing/model': 1.1455417020479217, 'learning_rate/qwen_vl_interface': 5.959319346524646e-07, 'learning_rate/action_model': 5.959319346524646e-06, 'epoch': 1.76}) +[2026-06-19 20:28:56,777][starVLA.training.train_starvla][INFO] - Step 4687, Loss: {'train/loss': 0.18978196382522583, 'train/grad_norm_pre_clip': 9.912936210632324, 'timing/data': 0.02163235389161855, 'timing/model': 1.3121066930470988, 'learning_rate/qwen_vl_interface': 5.953239249268804e-07, 'learning_rate/action_model': 5.953239249268804e-06, 'epoch': 1.76}) +[2026-06-19 20:29:01,203][starVLA.training.train_starvla][INFO] - Step 4688, Loss: {'train/loss': 0.30508163571357727, 'train/grad_norm_pre_clip': 19.1606502532959, 'timing/data': 0.020944515941664577, 'timing/model': 1.145019495044835, 'learning_rate/qwen_vl_interface': 5.947178285628081e-07, 'learning_rate/action_model': 5.94717828562808e-06, 'epoch': 1.76}) +[2026-06-19 20:29:05,836][starVLA.training.train_starvla][INFO] - Step 4689, Loss: {'train/loss': 0.041606608778238297, 'train/grad_norm_pre_clip': 3.2994394302368164, 'timing/data': 0.021344744018279016, 'timing/model': 1.167357315076515, 'learning_rate/qwen_vl_interface': 5.941136458093907e-07, 'learning_rate/action_model': 5.941136458093907e-06, 'epoch': 1.76}) +[2026-06-19 20:29:10,379][starVLA.training.train_starvla][INFO] - Step 4690, Loss: {'train/loss': 0.2852344810962677, 'train/grad_norm_pre_clip': 19.11713409423828, 'timing/data': 0.020857600960880518, 'timing/model': 1.1502753149252385, 'learning_rate/qwen_vl_interface': 5.935113769149855e-07, 'learning_rate/action_model': 5.935113769149856e-06, 'epoch': 1.76}) +[2026-06-19 20:29:14,959][starVLA.training.train_starvla][INFO] - Step 4691, Loss: {'train/loss': 0.09367844462394714, 'train/grad_norm_pre_clip': 4.396571636199951, 'timing/data': 0.020905570941977203, 'timing/model': 1.1509978770045564, 'learning_rate/qwen_vl_interface': 5.92911022127163e-07, 'learning_rate/action_model': 5.929110221271629e-06, 'epoch': 1.76}) +[2026-06-19 20:29:19,584][starVLA.training.train_starvla][INFO] - Step 4692, Loss: {'train/loss': 0.129682257771492, 'train/grad_norm_pre_clip': 21.660497665405273, 'timing/data': 0.020699569955468178, 'timing/model': 1.1602996089495718, 'learning_rate/qwen_vl_interface': 5.92312581692705e-07, 'learning_rate/action_model': 5.923125816927049e-06, 'epoch': 1.76}) +[2026-06-19 20:29:24,208][starVLA.training.train_starvla][INFO] - Step 4693, Loss: {'train/loss': 0.1124548465013504, 'train/grad_norm_pre_clip': 14.203935623168945, 'timing/data': 0.02011010400019586, 'timing/model': 1.31717995100189, 'learning_rate/qwen_vl_interface': 5.917160558576088e-07, 'learning_rate/action_model': 5.917160558576088e-06, 'epoch': 1.76}) +[2026-06-19 20:29:28,684][starVLA.training.train_starvla][INFO] - Step 4694, Loss: {'train/loss': 0.2149953991174698, 'train/grad_norm_pre_clip': 21.645753860473633, 'timing/data': 0.021125642932020128, 'timing/model': 1.150314314989373, 'learning_rate/qwen_vl_interface': 5.911214448670829e-07, 'learning_rate/action_model': 5.911214448670828e-06, 'epoch': 1.76}) +[2026-06-19 20:29:33,288][starVLA.training.train_starvla][INFO] - Step 4695, Loss: {'train/loss': 0.1486620157957077, 'train/grad_norm_pre_clip': 13.576072692871094, 'timing/data': 0.020700267050415277, 'timing/model': 1.1566691419575363, 'learning_rate/qwen_vl_interface': 5.905287489655499e-07, 'learning_rate/action_model': 5.905287489655499e-06, 'epoch': 1.76}) +[2026-06-19 20:29:37,981][starVLA.training.train_starvla][INFO] - Step 4696, Loss: {'train/loss': 0.11786799132823944, 'train/grad_norm_pre_clip': 10.701478958129883, 'timing/data': 0.0291191169526428, 'timing/model': 1.1468420670134947, 'learning_rate/qwen_vl_interface': 5.899379683966445e-07, 'learning_rate/action_model': 5.899379683966445e-06, 'epoch': 1.76}) +[2026-06-19 20:29:42,577][starVLA.training.train_starvla][INFO] - Step 4697, Loss: {'train/loss': 0.36054155230522156, 'train/grad_norm_pre_clip': 21.19307518005371, 'timing/data': 0.020887207007035613, 'timing/model': 1.1566564090317115, 'learning_rate/qwen_vl_interface': 5.893491034032142e-07, 'learning_rate/action_model': 5.893491034032142e-06, 'epoch': 1.76}) +[2026-06-19 20:29:47,163][starVLA.training.train_starvla][INFO] - Step 4698, Loss: {'train/loss': 0.1779177337884903, 'train/grad_norm_pre_clip': 6.9566497802734375, 'timing/data': 0.02093424799386412, 'timing/model': 1.1440541109768674, 'learning_rate/qwen_vl_interface': 5.887621542273196e-07, 'learning_rate/action_model': 5.887621542273196e-06, 'epoch': 1.76}) +[2026-06-19 20:29:51,788][starVLA.training.train_starvla][INFO] - Step 4699, Loss: {'train/loss': 0.42535924911499023, 'train/grad_norm_pre_clip': 26.361324310302734, 'timing/data': 0.021264094044454396, 'timing/model': 1.3239115469623357, 'learning_rate/qwen_vl_interface': 5.881771211102327e-07, 'learning_rate/action_model': 5.881771211102327e-06, 'epoch': 1.76}) +[2026-06-19 20:29:56,221][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 4700 +[2026-06-19 20:32:49,049][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 4700 +[2026-06-19 20:32:49,049][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 4700 +[2026-06-19 20:35:58,343][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 4700 +[2026-06-19 20:35:58,344][starVLA.training.train_starvla][INFO] - Step 4700, Loss: {'train/loss': 0.16386643052101135, 'train/grad_norm_pre_clip': 14.829197883605957, 'eval/loss': 0.1291632056236267, 'eval/latency_0/loss': 0.08184707909822464, 'eval/latency_1/loss': 0.04792596772313118, 'eval/latency_2/loss': 0.11157472431659698, 'eval/latency_3/loss': 0.15793000161647797, 'eval/latency_4/loss': 0.2478434294462204, 'eval/flappy/loss': 0.1291632056236267, 'eval/flappy/latency_0/loss': 0.08184707909822464, 'eval/flappy/latency_1/loss': 0.04792596772313118, 'eval/flappy/latency_2/loss': 0.11157472431659698, 'eval/flappy/latency_3/loss': 0.15793000161647797, 'eval/flappy/latency_4/loss': 0.2478434294462204, 'eval/flap_precision': 0.7177700348432056, 'eval/flap_recall': 0.7490909090909091, 'eval/flap_f1': 0.7330960854092528, 'eval/flap_rate_model': 0.07972222222222222, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.7330960854092528, 'eval/latency_0/flap_precision': 0.7177700348432056, 'eval/latency_0/flap_recall': 0.7490909090909091, 'eval/latency_0/flap_f1': 0.7330960854092528, 'eval/latency_0/flap_rate_model': 0.07972222222222222, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.7330960854092528, 'timing/data': 0.02106367493979633, 'timing/model': 1.1457935150247067, 'learning_rate/qwen_vl_interface': 5.875940042924391e-07, 'learning_rate/action_model': 5.875940042924391e-06, 'epoch': 1.76}) +[2026-06-19 20:35:58,345][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_4700_state +[2026-06-19 20:35:58,347][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-19 20:36:01,724][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_4700_state/model.safetensors +[2026-06-19 20:36:08,912][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_4700_state/optimizer.bin +[2026-06-19 20:36:08,912][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_4700_state/sampler.bin +[2026-06-19 20:36:08,912][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_4700_state/sampler_1.bin +[2026-06-19 20:36:08,913][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_4700_state/random_states_0.pkl +[2026-06-19 20:36:08,934][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4700 +[2026-06-19 20:36:08,935][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-19 20:36:08,969][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-19 20:36:13,480][starVLA.training.train_starvla][INFO] - Step 4701, Loss: {'train/loss': 0.23913173377513885, 'train/grad_norm_pre_clip': 19.012895584106445, 'timing/data': 0.020786488079465926, 'timing/model': 1.1278974850429222, 'learning_rate/qwen_vl_interface': 5.870128040136368e-07, 'learning_rate/action_model': 5.870128040136367e-06, 'epoch': 1.76}) +[2026-06-19 20:36:17,989][starVLA.training.train_starvla][INFO] - Step 4702, Loss: {'train/loss': 0.3893912434577942, 'train/grad_norm_pre_clip': 22.91116714477539, 'timing/data': 0.01702280400786549, 'timing/model': 1.2956253089942038, 'learning_rate/qwen_vl_interface': 5.864335205127332e-07, 'learning_rate/action_model': 5.864335205127332e-06, 'epoch': 1.76}) +[2026-06-19 20:36:22,378][starVLA.training.train_starvla][INFO] - Step 4703, Loss: {'train/loss': 0.19154652953147888, 'train/grad_norm_pre_clip': 17.607345581054688, 'timing/data': 0.018293994944542646, 'timing/model': 1.1343231499195099, 'learning_rate/qwen_vl_interface': 5.858561540278513e-07, 'learning_rate/action_model': 5.858561540278513e-06, 'epoch': 1.76}) +[2026-06-19 20:36:26,954][starVLA.training.train_starvla][INFO] - Step 4704, Loss: {'train/loss': 0.0819864571094513, 'train/grad_norm_pre_clip': 9.195436477661133, 'timing/data': 0.018376418971456587, 'timing/model': 1.13758348103147, 'learning_rate/qwen_vl_interface': 5.852807047963244e-07, 'learning_rate/action_model': 5.852807047963243e-06, 'epoch': 1.76}) +[2026-06-19 20:36:31,468][starVLA.training.train_starvla][INFO] - Step 4705, Loss: {'train/loss': 0.07045046240091324, 'train/grad_norm_pre_clip': 7.36595344543457, 'timing/data': 0.017481651972047985, 'timing/model': 1.1390008729649708, 'learning_rate/qwen_vl_interface': 5.847071730546974e-07, 'learning_rate/action_model': 5.847071730546974e-06, 'epoch': 1.76}) +[2026-06-19 20:36:36,085][starVLA.training.train_starvla][INFO] - Step 4706, Loss: {'train/loss': 0.13544988632202148, 'train/grad_norm_pre_clip': 13.362278938293457, 'timing/data': 0.017481668037362397, 'timing/model': 1.131785186007619, 'learning_rate/qwen_vl_interface': 5.841355590387291e-07, 'learning_rate/action_model': 5.841355590387291e-06, 'epoch': 1.76}) +[2026-06-19 20:36:40,642][starVLA.training.train_starvla][INFO] - Step 4707, Loss: {'train/loss': 0.1225336492061615, 'train/grad_norm_pre_clip': 17.04286766052246, 'timing/data': 0.01808469300158322, 'timing/model': 1.1368384070228785, 'learning_rate/qwen_vl_interface': 5.835658629833866e-07, 'learning_rate/action_model': 5.8356586298338654e-06, 'epoch': 1.76}) +[2026-06-19 20:36:45,209][starVLA.training.train_starvla][INFO] - Step 4708, Loss: {'train/loss': 0.06991998851299286, 'train/grad_norm_pre_clip': 5.534377098083496, 'timing/data': 0.017424289952032268, 'timing/model': 1.3190933580044657, 'learning_rate/qwen_vl_interface': 5.829980851228512e-07, 'learning_rate/action_model': 5.829980851228512e-06, 'epoch': 1.77}) +[2026-06-19 20:36:49,594][starVLA.training.train_starvla][INFO] - Step 4709, Loss: {'train/loss': 0.22812846302986145, 'train/grad_norm_pre_clip': 28.99696922302246, 'timing/data': 0.01765374094247818, 'timing/model': 1.1414902199758217, 'learning_rate/qwen_vl_interface': 5.824322256905155e-07, 'learning_rate/action_model': 5.824322256905154e-06, 'epoch': 1.77}) +[2026-06-19 20:36:54,158][starVLA.training.train_starvla][INFO] - Step 4710, Loss: {'train/loss': 0.0855741947889328, 'train/grad_norm_pre_clip': 10.688240051269531, 'timing/data': 0.01889575505629182, 'timing/model': 1.149095342028886, 'learning_rate/qwen_vl_interface': 5.818682849189818e-07, 'learning_rate/action_model': 5.818682849189817e-06, 'epoch': 1.77}) +[2026-06-19 20:36:58,711][starVLA.training.train_starvla][INFO] - Step 4711, Loss: {'train/loss': 0.08654960989952087, 'train/grad_norm_pre_clip': 12.377152442932129, 'timing/data': 0.01772724895272404, 'timing/model': 1.1470919459825382, 'learning_rate/qwen_vl_interface': 5.813062630400653e-07, 'learning_rate/action_model': 5.813062630400653e-06, 'epoch': 1.77}) +[2026-06-19 20:37:03,217][starVLA.training.train_starvla][INFO] - Step 4712, Loss: {'train/loss': 0.27206850051879883, 'train/grad_norm_pre_clip': 15.380260467529297, 'timing/data': 0.019173408043570817, 'timing/model': 1.1413436599541456, 'learning_rate/qwen_vl_interface': 5.807461602847922e-07, 'learning_rate/action_model': 5.807461602847922e-06, 'epoch': 1.77}) +[2026-06-19 20:37:07,774][starVLA.training.train_starvla][INFO] - Step 4713, Loss: {'train/loss': 0.4077397584915161, 'train/grad_norm_pre_clip': 26.261043548583984, 'timing/data': 0.018648730008862913, 'timing/model': 1.1455530960811302, 'learning_rate/qwen_vl_interface': 5.801879768833998e-07, 'learning_rate/action_model': 5.801879768833998e-06, 'epoch': 1.77}) +[2026-06-19 20:37:12,371][starVLA.training.train_starvla][INFO] - Step 4714, Loss: {'train/loss': 0.2013835459947586, 'train/grad_norm_pre_clip': 15.561394691467285, 'timing/data': 0.021705094957724214, 'timing/model': 1.3253788339206949, 'learning_rate/qwen_vl_interface': 5.796317130653363e-07, 'learning_rate/action_model': 5.796317130653363e-06, 'epoch': 1.77}) +[2026-06-19 20:37:16,784][starVLA.training.train_starvla][INFO] - Step 4715, Loss: {'train/loss': 0.1414032131433487, 'train/grad_norm_pre_clip': 5.482113361358643, 'timing/data': 0.018134842975996435, 'timing/model': 1.144938823999837, 'learning_rate/qwen_vl_interface': 5.790773690592592e-07, 'learning_rate/action_model': 5.7907736905925925e-06, 'epoch': 1.77}) +[2026-06-19 20:37:21,384][starVLA.training.train_starvla][INFO] - Step 4716, Loss: {'train/loss': 0.07882481068372726, 'train/grad_norm_pre_clip': 24.00424575805664, 'timing/data': 0.01831526798196137, 'timing/model': 1.147969749988988, 'learning_rate/qwen_vl_interface': 5.7852494509304e-07, 'learning_rate/action_model': 5.7852494509304e-06, 'epoch': 1.77}) +[2026-06-19 20:37:25,966][starVLA.training.train_starvla][INFO] - Step 4717, Loss: {'train/loss': 0.26093778014183044, 'train/grad_norm_pre_clip': 16.721023559570312, 'timing/data': 0.018264838960021734, 'timing/model': 1.1496681360295042, 'learning_rate/qwen_vl_interface': 5.779744413937592e-07, 'learning_rate/action_model': 5.779744413937592e-06, 'epoch': 1.77}) +[2026-06-19 20:37:30,550][starVLA.training.train_starvla][INFO] - Step 4718, Loss: {'train/loss': 0.1329115778207779, 'train/grad_norm_pre_clip': 15.494169235229492, 'timing/data': 0.018679081927984953, 'timing/model': 1.154538718983531, 'learning_rate/qwen_vl_interface': 5.774258581877073e-07, 'learning_rate/action_model': 5.774258581877073e-06, 'epoch': 1.77}) +[2026-06-19 20:37:35,137][starVLA.training.train_starvla][INFO] - Step 4719, Loss: {'train/loss': 0.30141934752464294, 'train/grad_norm_pre_clip': 45.75420379638672, 'timing/data': 0.017960401019081473, 'timing/model': 1.149924093973823, 'learning_rate/qwen_vl_interface': 5.768791957003866e-07, 'learning_rate/action_model': 5.768791957003865e-06, 'epoch': 1.77}) +[2026-06-19 20:37:39,709][starVLA.training.train_starvla][INFO] - Step 4720, Loss: {'train/loss': 0.16125372052192688, 'train/grad_norm_pre_clip': 9.818235397338867, 'timing/data': 0.018241468933410943, 'timing/model': 1.3092334099346772, 'learning_rate/qwen_vl_interface': 5.763344541565091e-07, 'learning_rate/action_model': 5.7633445415650905e-06, 'epoch': 1.77}) +[2026-06-19 20:37:44,140][starVLA.training.train_starvla][INFO] - Step 4721, Loss: {'train/loss': 0.08125781267881393, 'train/grad_norm_pre_clip': 16.307083129882812, 'timing/data': 0.01843053102493286, 'timing/model': 1.1433242830680683, 'learning_rate/qwen_vl_interface': 5.75791633779998e-07, 'learning_rate/action_model': 5.75791633779998e-06, 'epoch': 1.77}) +[2026-06-19 20:37:48,725][starVLA.training.train_starvla][INFO] - Step 4722, Loss: {'train/loss': 0.4213225245475769, 'train/grad_norm_pre_clip': 26.695592880249023, 'timing/data': 0.01963155798148364, 'timing/model': 1.1522223150823265, 'learning_rate/qwen_vl_interface': 5.752507347939857e-07, 'learning_rate/action_model': 5.752507347939858e-06, 'epoch': 1.77}) +[2026-06-19 20:37:53,349][starVLA.training.train_starvla][INFO] - Step 4723, Loss: {'train/loss': 0.2879592180252075, 'train/grad_norm_pre_clip': 18.30925750732422, 'timing/data': 0.018650030018761754, 'timing/model': 1.1555214889813215, 'learning_rate/qwen_vl_interface': 5.74711757420816e-07, 'learning_rate/action_model': 5.747117574208159e-06, 'epoch': 1.77}) +[2026-06-19 20:37:57,888][starVLA.training.train_starvla][INFO] - Step 4724, Loss: {'train/loss': 0.19486530125141144, 'train/grad_norm_pre_clip': 14.175802230834961, 'timing/data': 0.018426131922751665, 'timing/model': 1.1457442939281464, 'learning_rate/qwen_vl_interface': 5.74174701882042e-07, 'learning_rate/action_model': 5.74174701882042e-06, 'epoch': 1.77}) +[2026-06-19 20:38:02,495][starVLA.training.train_starvla][INFO] - Step 4725, Loss: {'train/loss': 0.279882550239563, 'train/grad_norm_pre_clip': 15.898576736450195, 'timing/data': 0.018625637982040644, 'timing/model': 1.1490953449392691, 'learning_rate/qwen_vl_interface': 5.736395683984262e-07, 'learning_rate/action_model': 5.736395683984261e-06, 'epoch': 1.77}) +[2026-06-19 20:38:07,128][starVLA.training.train_starvla][INFO] - Step 4726, Loss: {'train/loss': 0.2555390000343323, 'train/grad_norm_pre_clip': 22.796138763427734, 'timing/data': 0.018826835905201733, 'timing/model': 1.316694350913167, 'learning_rate/qwen_vl_interface': 5.731063571899421e-07, 'learning_rate/action_model': 5.731063571899421e-06, 'epoch': 1.77}) +[2026-06-19 20:38:11,554][starVLA.training.train_starvla][INFO] - Step 4727, Loss: {'train/loss': 0.2629951536655426, 'train/grad_norm_pre_clip': 19.091205596923828, 'timing/data': 0.01858648902270943, 'timing/model': 1.1556286789709702, 'learning_rate/qwen_vl_interface': 5.72575068475773e-07, 'learning_rate/action_model': 5.72575068475773e-06, 'epoch': 1.77}) +[2026-06-19 20:38:16,150][starVLA.training.train_starvla][INFO] - Step 4728, Loss: {'train/loss': 0.11326868087053299, 'train/grad_norm_pre_clip': 16.68683433532715, 'timing/data': 0.019417635980062187, 'timing/model': 1.1441988279111683, 'learning_rate/qwen_vl_interface': 5.720457024743115e-07, 'learning_rate/action_model': 5.720457024743115e-06, 'epoch': 1.77}) +[2026-06-19 20:38:20,737][starVLA.training.train_starvla][INFO] - Step 4729, Loss: {'train/loss': 0.1014833077788353, 'train/grad_norm_pre_clip': 17.896686553955078, 'timing/data': 0.01935931306798011, 'timing/model': 1.1430056400131434, 'learning_rate/qwen_vl_interface': 5.715182594031601e-07, 'learning_rate/action_model': 5.7151825940316e-06, 'epoch': 1.77}) +[2026-06-19 20:38:25,236][starVLA.training.train_starvla][INFO] - Step 4730, Loss: {'train/loss': 0.6260999441146851, 'train/grad_norm_pre_clip': 23.71709442138672, 'timing/data': 0.021495311055332422, 'timing/model': 1.151090212049894, 'learning_rate/qwen_vl_interface': 5.709927394791305e-07, 'learning_rate/action_model': 5.709927394791305e-06, 'epoch': 1.77}) +[2026-06-19 20:38:29,774][starVLA.training.train_starvla][INFO] - Step 4731, Loss: {'train/loss': 0.23740559816360474, 'train/grad_norm_pre_clip': 20.569286346435547, 'timing/data': 0.01833728502970189, 'timing/model': 1.1535181149374694, 'learning_rate/qwen_vl_interface': 5.70469142918244e-07, 'learning_rate/action_model': 5.704691429182441e-06, 'epoch': 1.77}) +[2026-06-19 20:38:34,379][starVLA.training.train_starvla][INFO] - Step 4732, Loss: {'train/loss': 0.11459149420261383, 'train/grad_norm_pre_clip': 14.2979736328125, 'timing/data': 0.017707080929540098, 'timing/model': 1.320711191976443, 'learning_rate/qwen_vl_interface': 5.699474699357317e-07, 'learning_rate/action_model': 5.699474699357317e-06, 'epoch': 1.77}) +[2026-06-19 20:38:38,793][starVLA.training.train_starvla][INFO] - Step 4733, Loss: {'train/loss': 0.15069927275180817, 'train/grad_norm_pre_clip': 15.466286659240723, 'timing/data': 0.01839490607380867, 'timing/model': 1.1453571860911325, 'learning_rate/qwen_vl_interface': 5.694277207460331e-07, 'learning_rate/action_model': 5.694277207460332e-06, 'epoch': 1.77}) +[2026-06-19 20:38:43,378][starVLA.training.train_starvla][INFO] - Step 4734, Loss: {'train/loss': 0.22248831391334534, 'train/grad_norm_pre_clip': 18.79408073425293, 'timing/data': 0.02092009806074202, 'timing/model': 1.1482020859839395, 'learning_rate/qwen_vl_interface': 5.689098955627982e-07, 'learning_rate/action_model': 5.689098955627982e-06, 'epoch': 1.78}) +[2026-06-19 20:38:47,960][starVLA.training.train_starvla][INFO] - Step 4735, Loss: {'train/loss': 0.198625847697258, 'train/grad_norm_pre_clip': 13.528658866882324, 'timing/data': 0.018782928003929555, 'timing/model': 1.1484741199528798, 'learning_rate/qwen_vl_interface': 5.683939945988853e-07, 'learning_rate/action_model': 5.683939945988853e-06, 'epoch': 1.78}) +[2026-06-19 20:38:52,559][starVLA.training.train_starvla][INFO] - Step 4736, Loss: {'train/loss': 0.087483249604702, 'train/grad_norm_pre_clip': 8.626812934875488, 'timing/data': 0.019278640975244343, 'timing/model': 1.1455844319425523, 'learning_rate/qwen_vl_interface': 5.678800180663608e-07, 'learning_rate/action_model': 5.678800180663609e-06, 'epoch': 1.78}) +[2026-06-19 20:38:57,158][starVLA.training.train_starvla][INFO] - Step 4737, Loss: {'train/loss': 0.24722784757614136, 'train/grad_norm_pre_clip': 20.881126403808594, 'timing/data': 0.019220602000132203, 'timing/model': 1.3166879329364747, 'learning_rate/qwen_vl_interface': 5.673679661765018e-07, 'learning_rate/action_model': 5.6736796617650175e-06, 'epoch': 1.78}) +[2026-06-19 20:39:01,588][starVLA.training.train_starvla][INFO] - Step 4738, Loss: {'train/loss': 0.2842349112033844, 'train/grad_norm_pre_clip': 36.07576370239258, 'timing/data': 0.018077571992762387, 'timing/model': 1.145633367006667, 'learning_rate/qwen_vl_interface': 5.668578391397938e-07, 'learning_rate/action_model': 5.668578391397937e-06, 'epoch': 1.78}) +[2026-06-19 20:39:06,183][starVLA.training.train_starvla][INFO] - Step 4739, Loss: {'train/loss': 0.2525298595428467, 'train/grad_norm_pre_clip': 23.505186080932617, 'timing/data': 0.017650478053838015, 'timing/model': 1.152182613965124, 'learning_rate/qwen_vl_interface': 5.663496371659303e-07, 'learning_rate/action_model': 5.663496371659303e-06, 'epoch': 1.78}) +[2026-06-19 20:39:10,782][starVLA.training.train_starvla][INFO] - Step 4740, Loss: {'train/loss': 0.23142209649085999, 'train/grad_norm_pre_clip': 17.038240432739258, 'timing/data': 0.019564175978302956, 'timing/model': 1.1542348640505224, 'learning_rate/qwen_vl_interface': 5.658433604638141e-07, 'learning_rate/action_model': 5.65843360463814e-06, 'epoch': 1.78}) +[2026-06-19 20:39:15,400][starVLA.training.train_starvla][INFO] - Step 4741, Loss: {'train/loss': 0.2537972033023834, 'train/grad_norm_pre_clip': 19.006580352783203, 'timing/data': 0.019244999042712152, 'timing/model': 1.156846665078774, 'learning_rate/qwen_vl_interface': 5.653390092415559e-07, 'learning_rate/action_model': 5.653390092415559e-06, 'epoch': 1.78}) +[2026-06-19 20:39:20,024][starVLA.training.train_starvla][INFO] - Step 4742, Loss: {'train/loss': 0.23461945354938507, 'train/grad_norm_pre_clip': 13.405107498168945, 'timing/data': 0.018397795036435127, 'timing/model': 1.147765783010982, 'learning_rate/qwen_vl_interface': 5.648365837064759e-07, 'learning_rate/action_model': 5.648365837064759e-06, 'epoch': 1.78}) +[2026-06-19 20:39:24,622][starVLA.training.train_starvla][INFO] - Step 4743, Loss: {'train/loss': 0.2998114228248596, 'train/grad_norm_pre_clip': 19.803743362426758, 'timing/data': 0.018177712918259203, 'timing/model': 1.3086941690417007, 'learning_rate/qwen_vl_interface': 5.643360840651025e-07, 'learning_rate/action_model': 5.643360840651025e-06, 'epoch': 1.78}) +[2026-06-19 20:39:29,232][starVLA.training.train_starvla][INFO] - Step 4744, Loss: {'train/loss': 0.2009638249874115, 'train/grad_norm_pre_clip': 20.40715217590332, 'timing/data': 0.018755590077489614, 'timing/model': 1.3150049319956452, 'learning_rate/qwen_vl_interface': 5.638375105231719e-07, 'learning_rate/action_model': 5.638375105231719e-06, 'epoch': 1.78}) +[2026-06-19 20:39:33,654][starVLA.training.train_starvla][INFO] - Step 4745, Loss: {'train/loss': 0.11221808195114136, 'train/grad_norm_pre_clip': 13.174910545349121, 'timing/data': 0.01791740395128727, 'timing/model': 1.1493804439669475, 'learning_rate/qwen_vl_interface': 5.633408632856294e-07, 'learning_rate/action_model': 5.6334086328562945e-06, 'epoch': 1.78}) +[2026-06-19 20:39:38,256][starVLA.training.train_starvla][INFO] - Step 4746, Loss: {'train/loss': 0.24385473132133484, 'train/grad_norm_pre_clip': 72.31214904785156, 'timing/data': 0.018362465081736445, 'timing/model': 1.1453718079719692, 'learning_rate/qwen_vl_interface': 5.628461425566268e-07, 'learning_rate/action_model': 5.6284614255662675e-06, 'epoch': 1.78}) +[2026-06-19 20:39:42,864][starVLA.training.train_starvla][INFO] - Step 4747, Loss: {'train/loss': 0.2804848849773407, 'train/grad_norm_pre_clip': 20.842605590820312, 'timing/data': 0.018092819955199957, 'timing/model': 1.1457923600682989, 'learning_rate/qwen_vl_interface': 5.62353348539526e-07, 'learning_rate/action_model': 5.623533485395259e-06, 'epoch': 1.78}) +[2026-06-19 20:39:47,357][starVLA.training.train_starvla][INFO] - Step 4748, Loss: {'train/loss': 0.14642786979675293, 'train/grad_norm_pre_clip': 18.207103729248047, 'timing/data': 0.01895613397937268, 'timing/model': 1.147345416015014, 'learning_rate/qwen_vl_interface': 5.618624814368956e-07, 'learning_rate/action_model': 5.618624814368956e-06, 'epoch': 1.78}) +[2026-06-19 20:39:51,906][starVLA.training.train_starvla][INFO] - Step 4749, Loss: {'train/loss': 0.29992473125457764, 'train/grad_norm_pre_clip': 12.569561958312988, 'timing/data': 0.01832251693122089, 'timing/model': 1.1437338989926502, 'learning_rate/qwen_vl_interface': 5.613735414505126e-07, 'learning_rate/action_model': 5.613735414505125e-06, 'epoch': 1.78}) +[2026-06-19 20:39:56,504][starVLA.training.train_starvla][INFO] - Step 4750, Loss: {'train/loss': 0.3034801185131073, 'train/grad_norm_pre_clip': 20.564146041870117, 'timing/data': 0.01898598193656653, 'timing/model': 1.3193877270678058, 'learning_rate/qwen_vl_interface': 5.608865287813617e-07, 'learning_rate/action_model': 5.608865287813616e-06, 'epoch': 1.78}) +[2026-06-19 20:40:00,936][starVLA.training.train_starvla][INFO] - Step 4751, Loss: {'train/loss': 0.22722336649894714, 'train/grad_norm_pre_clip': 15.813864707946777, 'timing/data': 0.01779008295852691, 'timing/model': 1.1455040730070323, 'learning_rate/qwen_vl_interface': 5.604014436296362e-07, 'learning_rate/action_model': 5.604014436296362e-06, 'epoch': 1.78}) +[2026-06-19 20:40:05,538][starVLA.training.train_starvla][INFO] - Step 4752, Loss: {'train/loss': 0.3770109713077545, 'train/grad_norm_pre_clip': 21.972431182861328, 'timing/data': 0.01926581107545644, 'timing/model': 1.1473981260787696, 'learning_rate/qwen_vl_interface': 5.599182861947348e-07, 'learning_rate/action_model': 5.599182861947347e-06, 'epoch': 1.78}) +[2026-06-19 20:40:10,143][starVLA.training.train_starvla][INFO] - Step 4753, Loss: {'train/loss': 0.057747915387153625, 'train/grad_norm_pre_clip': 6.469921112060547, 'timing/data': 0.01841210399288684, 'timing/model': 1.1488180810119957, 'learning_rate/qwen_vl_interface': 5.594370566752665e-07, 'learning_rate/action_model': 5.5943705667526645e-06, 'epoch': 1.78}) +[2026-06-19 20:40:14,737][starVLA.training.train_starvla][INFO] - Step 4754, Loss: {'train/loss': 0.3423556089401245, 'train/grad_norm_pre_clip': 23.080904006958008, 'timing/data': 0.018528582993894815, 'timing/model': 1.1460415159817785, 'learning_rate/qwen_vl_interface': 5.589577552690456e-07, 'learning_rate/action_model': 5.589577552690456e-06, 'epoch': 1.78}) +[2026-06-19 20:40:19,223][starVLA.training.train_starvla][INFO] - Step 4755, Loss: {'train/loss': 0.11889366805553436, 'train/grad_norm_pre_clip': 14.322529792785645, 'timing/data': 0.01772097998764366, 'timing/model': 1.1487698599230498, 'learning_rate/qwen_vl_interface': 5.584803821730961e-07, 'learning_rate/action_model': 5.58480382173096e-06, 'epoch': 1.78}) +[2026-06-19 20:40:23,781][starVLA.training.train_starvla][INFO] - Step 4756, Loss: {'train/loss': 0.07168154418468475, 'train/grad_norm_pre_clip': 8.869093894958496, 'timing/data': 0.0246307710185647, 'timing/model': 1.2713366250973195, 'learning_rate/qwen_vl_interface': 5.580049375836466e-07, 'learning_rate/action_model': 5.580049375836466e-06, 'epoch': 1.78}) +[2026-06-19 20:40:28,218][starVLA.training.train_starvla][INFO] - Step 4757, Loss: {'train/loss': 0.21695774793624878, 'train/grad_norm_pre_clip': 10.448748588562012, 'timing/data': 0.018839479074813426, 'timing/model': 1.150524433935061, 'learning_rate/qwen_vl_interface': 5.575314216961356e-07, 'learning_rate/action_model': 5.575314216961356e-06, 'epoch': 1.78}) +[2026-06-19 20:40:32,820][starVLA.training.train_starvla][INFO] - Step 4758, Loss: {'train/loss': 0.29362669587135315, 'train/grad_norm_pre_clip': 20.059101104736328, 'timing/data': 0.021689218003302813, 'timing/model': 1.146913085016422, 'learning_rate/qwen_vl_interface': 5.570598347052073e-07, 'learning_rate/action_model': 5.570598347052072e-06, 'epoch': 1.78}) +[2026-06-19 20:40:37,418][starVLA.training.train_starvla][INFO] - Step 4759, Loss: {'train/loss': 0.16289927065372467, 'train/grad_norm_pre_clip': 26.2204647064209, 'timing/data': 0.01912881399039179, 'timing/model': 1.1564459579531103, 'learning_rate/qwen_vl_interface': 5.565901768047126e-07, 'learning_rate/action_model': 5.5659017680471255e-06, 'epoch': 1.78}) +[2026-06-19 20:40:41,965][starVLA.training.train_starvla][INFO] - Step 4760, Loss: {'train/loss': 0.10241823643445969, 'train/grad_norm_pre_clip': 11.159700393676758, 'timing/data': 0.01904444198589772, 'timing/model': 1.1461503030732274, 'learning_rate/qwen_vl_interface': 5.561224481877105e-07, 'learning_rate/action_model': 5.561224481877105e-06, 'epoch': 1.78}) +[2026-06-19 20:40:46,573][starVLA.training.train_starvla][INFO] - Step 4761, Loss: {'train/loss': 0.11090922355651855, 'train/grad_norm_pre_clip': 10.06694507598877, 'timing/data': 0.01918327994644642, 'timing/model': 1.1462993349414319, 'learning_rate/qwen_vl_interface': 5.55656649046467e-07, 'learning_rate/action_model': 5.556566490464669e-06, 'epoch': 1.79}) +[2026-06-19 20:40:51,181][starVLA.training.train_starvla][INFO] - Step 4762, Loss: {'train/loss': 0.18519370257854462, 'train/grad_norm_pre_clip': 29.43399429321289, 'timing/data': 0.018833121983334422, 'timing/model': 1.3217506629880518, 'learning_rate/qwen_vl_interface': 5.551927795724544e-07, 'learning_rate/action_model': 5.551927795724543e-06, 'epoch': 1.79}) +[2026-06-19 20:40:55,616][starVLA.training.train_starvla][INFO] - Step 4763, Loss: {'train/loss': 0.4005691111087799, 'train/grad_norm_pre_clip': 31.11564064025879, 'timing/data': 0.019043853040784597, 'timing/model': 1.1462148539721966, 'learning_rate/qwen_vl_interface': 5.547308399563515e-07, 'learning_rate/action_model': 5.547308399563514e-06, 'epoch': 1.79}) +[2026-06-19 20:41:00,218][starVLA.training.train_starvla][INFO] - Step 4764, Loss: {'train/loss': 0.19895264506340027, 'train/grad_norm_pre_clip': 16.4822998046875, 'timing/data': 0.018640702008269727, 'timing/model': 1.1454039140371606, 'learning_rate/qwen_vl_interface': 5.542708303880442e-07, 'learning_rate/action_model': 5.542708303880442e-06, 'epoch': 1.79}) +[2026-06-19 20:41:04,745][starVLA.training.train_starvla][INFO] - Step 4765, Loss: {'train/loss': 0.18743430078029633, 'train/grad_norm_pre_clip': 15.19084358215332, 'timing/data': 0.018552941037341952, 'timing/model': 1.1520399180008098, 'learning_rate/qwen_vl_interface': 5.538127510566258e-07, 'learning_rate/action_model': 5.538127510566258e-06, 'epoch': 1.79}) +[2026-06-19 20:41:09,294][starVLA.training.train_starvla][INFO] - Step 4766, Loss: {'train/loss': 0.2153186798095703, 'train/grad_norm_pre_clip': 17.2492733001709, 'timing/data': 0.020003373036161065, 'timing/model': 1.1527509440202266, 'learning_rate/qwen_vl_interface': 5.533566021503944e-07, 'learning_rate/action_model': 5.533566021503943e-06, 'epoch': 1.79}) +[2026-06-19 20:41:13,901][starVLA.training.train_starvla][INFO] - Step 4767, Loss: {'train/loss': 0.21061651408672333, 'train/grad_norm_pre_clip': 17.284435272216797, 'timing/data': 0.01880839397199452, 'timing/model': 1.1524589599575847, 'learning_rate/qwen_vl_interface': 5.529023838568566e-07, 'learning_rate/action_model': 5.529023838568566e-06, 'epoch': 1.79}) +[2026-06-19 20:41:18,524][starVLA.training.train_starvla][INFO] - Step 4768, Loss: {'train/loss': 0.32737383246421814, 'train/grad_norm_pre_clip': 20.649974822998047, 'timing/data': 0.018692029058001935, 'timing/model': 1.3208771171048284, 'learning_rate/qwen_vl_interface': 5.524500963627239e-07, 'learning_rate/action_model': 5.524500963627239e-06, 'epoch': 1.79}) +[2026-06-19 20:41:22,970][starVLA.training.train_starvla][INFO] - Step 4769, Loss: {'train/loss': 0.14453919231891632, 'train/grad_norm_pre_clip': 14.212541580200195, 'timing/data': 0.018494384945370257, 'timing/model': 1.152193233021535, 'learning_rate/qwen_vl_interface': 5.519997398539145e-07, 'learning_rate/action_model': 5.519997398539145e-06, 'epoch': 1.79}) +[2026-06-19 20:41:27,586][starVLA.training.train_starvla][INFO] - Step 4770, Loss: {'train/loss': 0.16981381177902222, 'train/grad_norm_pre_clip': 13.58181095123291, 'timing/data': 0.022525540087372065, 'timing/model': 1.1457272949628532, 'learning_rate/qwen_vl_interface': 5.515513145155531e-07, 'learning_rate/action_model': 5.515513145155531e-06, 'epoch': 1.79}) +[2026-06-19 20:41:32,174][starVLA.training.train_starvla][INFO] - Step 4771, Loss: {'train/loss': 0.17974701523780823, 'train/grad_norm_pre_clip': 13.698724746704102, 'timing/data': 0.020365793025121093, 'timing/model': 1.1476267910329625, 'learning_rate/qwen_vl_interface': 5.511048205319702e-07, 'learning_rate/action_model': 5.511048205319701e-06, 'epoch': 1.79}) +[2026-06-19 20:41:36,773][starVLA.training.train_starvla][INFO] - Step 4772, Loss: {'train/loss': 0.11532474309206009, 'train/grad_norm_pre_clip': 10.196478843688965, 'timing/data': 0.019918760983273387, 'timing/model': 1.1491056489758193, 'learning_rate/qwen_vl_interface': 5.506602580867034e-07, 'learning_rate/action_model': 5.506602580867034e-06, 'epoch': 1.79}) +[2026-06-19 20:41:41,382][starVLA.training.train_starvla][INFO] - Step 4773, Loss: {'train/loss': 0.052418798208236694, 'train/grad_norm_pre_clip': 7.445201396942139, 'timing/data': 0.018714474979788065, 'timing/model': 1.154545309022069, 'learning_rate/qwen_vl_interface': 5.502176273624942e-07, 'learning_rate/action_model': 5.502176273624942e-06, 'epoch': 1.79}) +[2026-06-19 20:41:46,015][starVLA.training.train_starvla][INFO] - Step 4774, Loss: {'train/loss': 0.2007703185081482, 'train/grad_norm_pre_clip': 16.75261688232422, 'timing/data': 0.019811334903351963, 'timing/model': 1.3430768009275198, 'learning_rate/qwen_vl_interface': 5.497769285412921e-07, 'learning_rate/action_model': 5.497769285412921e-06, 'epoch': 1.79}) +[2026-06-19 20:41:50,443][starVLA.training.train_starvla][INFO] - Step 4775, Loss: {'train/loss': 0.2431049644947052, 'train/grad_norm_pre_clip': 21.68973731994629, 'timing/data': 0.018322131014429033, 'timing/model': 1.148027500952594, 'learning_rate/qwen_vl_interface': 5.493381618042519e-07, 'learning_rate/action_model': 5.493381618042518e-06, 'epoch': 1.79}) +[2026-06-19 20:41:55,036][starVLA.training.train_starvla][INFO] - Step 4776, Loss: {'train/loss': 0.13702645897865295, 'train/grad_norm_pre_clip': 16.691850662231445, 'timing/data': 0.01880881702527404, 'timing/model': 1.1505427700467408, 'learning_rate/qwen_vl_interface': 5.489013273317343e-07, 'learning_rate/action_model': 5.489013273317343e-06, 'epoch': 1.79}) +[2026-06-19 20:41:59,539][starVLA.training.train_starvla][INFO] - Step 4777, Loss: {'train/loss': 0.09234587103128433, 'train/grad_norm_pre_clip': 9.505351066589355, 'timing/data': 0.017980252974666655, 'timing/model': 1.1537245019571856, 'learning_rate/qwen_vl_interface': 5.484664253033042e-07, 'learning_rate/action_model': 5.484664253033042e-06, 'epoch': 1.79}) +[2026-06-19 20:42:04,055][starVLA.training.train_starvla][INFO] - Step 4778, Loss: {'train/loss': 0.1451777070760727, 'train/grad_norm_pre_clip': 20.914443969726562, 'timing/data': 0.019061045022681355, 'timing/model': 1.1522160910535604, 'learning_rate/qwen_vl_interface': 5.480334558977351e-07, 'learning_rate/action_model': 5.480334558977351e-06, 'epoch': 1.79}) +[2026-06-19 20:42:08,603][starVLA.training.train_starvla][INFO] - Step 4779, Loss: {'train/loss': 0.3717300593852997, 'train/grad_norm_pre_clip': 32.97748565673828, 'timing/data': 0.01867240492720157, 'timing/model': 1.153561312938109, 'learning_rate/qwen_vl_interface': 5.476024192930029e-07, 'learning_rate/action_model': 5.47602419293003e-06, 'epoch': 1.79}) +[2026-06-19 20:42:13,187][starVLA.training.train_starvla][INFO] - Step 4780, Loss: {'train/loss': 0.12601065635681152, 'train/grad_norm_pre_clip': 14.099862098693848, 'timing/data': 0.018712721997871995, 'timing/model': 1.3113395309774205, 'learning_rate/qwen_vl_interface': 5.471733156662914e-07, 'learning_rate/action_model': 5.4717331566629135e-06, 'epoch': 1.79}) +[2026-06-19 20:42:17,636][starVLA.training.train_starvla][INFO] - Step 4781, Loss: {'train/loss': 0.04147220775485039, 'train/grad_norm_pre_clip': 4.850362777709961, 'timing/data': 0.025039776926860213, 'timing/model': 1.1533374339342117, 'learning_rate/qwen_vl_interface': 5.467461451939882e-07, 'learning_rate/action_model': 5.467461451939882e-06, 'epoch': 1.79}) +[2026-06-19 20:42:22,240][starVLA.training.train_starvla][INFO] - Step 4782, Loss: {'train/loss': 0.19850847125053406, 'train/grad_norm_pre_clip': 13.25180721282959, 'timing/data': 0.019395472016185522, 'timing/model': 1.1503853210015222, 'learning_rate/qwen_vl_interface': 5.463209080516873e-07, 'learning_rate/action_model': 5.463209080516873e-06, 'epoch': 1.79}) +[2026-06-19 20:42:26,840][starVLA.training.train_starvla][INFO] - Step 4783, Loss: {'train/loss': 0.2453448325395584, 'train/grad_norm_pre_clip': 19.130319595336914, 'timing/data': 0.018773896037600935, 'timing/model': 1.148337662103586, 'learning_rate/qwen_vl_interface': 5.458976044141875e-07, 'learning_rate/action_model': 5.458976044141875e-06, 'epoch': 1.79}) +[2026-06-19 20:42:31,378][starVLA.training.train_starvla][INFO] - Step 4784, Loss: {'train/loss': 0.12516267597675323, 'train/grad_norm_pre_clip': 20.2219181060791, 'timing/data': 0.01938854600302875, 'timing/model': 1.1515680140582845, 'learning_rate/qwen_vl_interface': 5.454762344554931e-07, 'learning_rate/action_model': 5.4547623445549316e-06, 'epoch': 1.79}) +[2026-06-19 20:42:35,976][starVLA.training.train_starvla][INFO] - Step 4785, Loss: {'train/loss': 0.29353946447372437, 'train/grad_norm_pre_clip': 30.369789123535156, 'timing/data': 0.018833876005373895, 'timing/model': 1.1648565609939396, 'learning_rate/qwen_vl_interface': 5.450567983488137e-07, 'learning_rate/action_model': 5.450567983488137e-06, 'epoch': 1.79}) +[2026-06-19 20:42:40,584][starVLA.training.train_starvla][INFO] - Step 4786, Loss: {'train/loss': 0.16617855429649353, 'train/grad_norm_pre_clip': 16.06584358215332, 'timing/data': 0.01892221800517291, 'timing/model': 1.3120013140141964, 'learning_rate/qwen_vl_interface': 5.446392962665629e-07, 'learning_rate/action_model': 5.4463929626656296e-06, 'epoch': 1.79}) +[2026-06-19 20:42:45,016][starVLA.training.train_starvla][INFO] - Step 4787, Loss: {'train/loss': 0.2492002546787262, 'train/grad_norm_pre_clip': 29.582416534423828, 'timing/data': 0.018764174077659845, 'timing/model': 1.146389817004092, 'learning_rate/qwen_vl_interface': 5.442237283803605e-07, 'learning_rate/action_model': 5.442237283803604e-06, 'epoch': 1.79}) +[2026-06-19 20:42:49,624][starVLA.training.train_starvla][INFO] - Step 4788, Loss: {'train/loss': 0.11283554136753082, 'train/grad_norm_pre_clip': 7.308964252471924, 'timing/data': 0.019351319991983473, 'timing/model': 1.1476675359299406, 'learning_rate/qwen_vl_interface': 5.43810094861031e-07, 'learning_rate/action_model': 5.43810094861031e-06, 'epoch': 1.8}) +[2026-06-19 20:42:54,224][starVLA.training.train_starvla][INFO] - Step 4789, Loss: {'train/loss': 0.13508348166942596, 'train/grad_norm_pre_clip': 9.685248374938965, 'timing/data': 0.018713425961323082, 'timing/model': 1.1576444919919595, 'learning_rate/qwen_vl_interface': 5.433983958786027e-07, 'learning_rate/action_model': 5.433983958786027e-06, 'epoch': 1.8}) +[2026-06-19 20:42:58,862][starVLA.training.train_starvla][INFO] - Step 4790, Loss: {'train/loss': 0.11620417982339859, 'train/grad_norm_pre_clip': 17.893068313598633, 'timing/data': 0.018905999953858554, 'timing/model': 1.1754368129186332, 'learning_rate/qwen_vl_interface': 5.429886316023102e-07, 'learning_rate/action_model': 5.429886316023101e-06, 'epoch': 1.8}) +[2026-06-19 20:43:03,380][starVLA.training.train_starvla][INFO] - Step 4791, Loss: {'train/loss': 0.11293984949588776, 'train/grad_norm_pre_clip': 8.468485832214355, 'timing/data': 0.01806345593649894, 'timing/model': 1.1553835879312828, 'learning_rate/qwen_vl_interface': 5.425808022005916e-07, 'learning_rate/action_model': 5.425808022005916e-06, 'epoch': 1.8}) +[2026-06-19 20:43:07,943][starVLA.training.train_starvla][INFO] - Step 4792, Loss: {'train/loss': 0.1583854854106903, 'train/grad_norm_pre_clip': 16.1968936920166, 'timing/data': 0.019391252077184618, 'timing/model': 1.2780797310406342, 'learning_rate/qwen_vl_interface': 5.421749078410903e-07, 'learning_rate/action_model': 5.421749078410903e-06, 'epoch': 1.8}) +[2026-06-19 20:43:12,391][starVLA.training.train_starvla][INFO] - Step 4793, Loss: {'train/loss': 0.2160884290933609, 'train/grad_norm_pre_clip': 14.793668746948242, 'timing/data': 0.01836452109273523, 'timing/model': 1.1462401719763875, 'learning_rate/qwen_vl_interface': 5.417709486906547e-07, 'learning_rate/action_model': 5.417709486906547e-06, 'epoch': 1.8}) +[2026-06-19 20:43:16,992][starVLA.training.train_starvla][INFO] - Step 4794, Loss: {'train/loss': 0.20693524181842804, 'train/grad_norm_pre_clip': 27.512136459350586, 'timing/data': 0.018749778042547405, 'timing/model': 1.1445452950429171, 'learning_rate/qwen_vl_interface': 5.413689249153367e-07, 'learning_rate/action_model': 5.413689249153366e-06, 'epoch': 1.8}) +[2026-06-19 20:43:21,578][starVLA.training.train_starvla][INFO] - Step 4795, Loss: {'train/loss': 0.11526710540056229, 'train/grad_norm_pre_clip': 11.702753067016602, 'timing/data': 0.01861263602040708, 'timing/model': 1.1459133819444105, 'learning_rate/qwen_vl_interface': 5.409688366803931e-07, 'learning_rate/action_model': 5.409688366803931e-06, 'epoch': 1.8}) +[2026-06-19 20:43:26,147][starVLA.training.train_starvla][INFO] - Step 4796, Loss: {'train/loss': 0.10536666959524155, 'train/grad_norm_pre_clip': 9.220293045043945, 'timing/data': 0.0190277659567073, 'timing/model': 1.1479550240328535, 'learning_rate/qwen_vl_interface': 5.405706841502853e-07, 'learning_rate/action_model': 5.4057068415028525e-06, 'epoch': 1.8}) +[2026-06-19 20:43:30,772][starVLA.training.train_starvla][INFO] - Step 4797, Loss: {'train/loss': 0.15987655520439148, 'train/grad_norm_pre_clip': 30.96704864501953, 'timing/data': 0.01897133095189929, 'timing/model': 1.1682069629896432, 'learning_rate/qwen_vl_interface': 5.401744674886781e-07, 'learning_rate/action_model': 5.40174467488678e-06, 'epoch': 1.8}) +[2026-06-19 20:43:35,449][starVLA.training.train_starvla][INFO] - Step 4798, Loss: {'train/loss': 0.14072714745998383, 'train/grad_norm_pre_clip': 9.413905143737793, 'timing/data': 0.01818812801502645, 'timing/model': 1.3286195079563186, 'learning_rate/qwen_vl_interface': 5.397801868584418e-07, 'learning_rate/action_model': 5.397801868584417e-06, 'epoch': 1.8}) +[2026-06-19 20:43:39,912][starVLA.training.train_starvla][INFO] - Step 4799, Loss: {'train/loss': 0.30264702439308167, 'train/grad_norm_pre_clip': 17.37540626525879, 'timing/data': 0.01881990407127887, 'timing/model': 1.1561195309041068, 'learning_rate/qwen_vl_interface': 5.393878424216502e-07, 'learning_rate/action_model': 5.393878424216501e-06, 'epoch': 1.8}) +[2026-06-19 20:43:44,516][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 4800 +[2026-06-19 20:46:36,318][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 4800 +[2026-06-19 20:46:36,319][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 4800 +[2026-06-19 20:49:37,897][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 4800 +[2026-06-19 20:49:37,899][starVLA.training.train_starvla][INFO] - Step 4800, Loss: {'train/loss': 0.18621060252189636, 'train/grad_norm_pre_clip': 20.752859115600586, 'eval/loss': 0.12793102860450745, 'eval/latency_0/loss': 0.08091391623020172, 'eval/latency_1/loss': 0.04826418682932854, 'eval/latency_2/loss': 0.11461400240659714, 'eval/latency_3/loss': 0.1542598307132721, 'eval/latency_4/loss': 0.24245373904705048, 'eval/flappy/loss': 0.12793102860450745, 'eval/flappy/latency_0/loss': 0.08091391623020172, 'eval/flappy/latency_1/loss': 0.04826418682932854, 'eval/flappy/latency_2/loss': 0.11461400240659714, 'eval/flappy/latency_3/loss': 0.1542598307132721, 'eval/flappy/latency_4/loss': 0.24245373904705048, 'eval/flap_precision': 0.7406015037593985, 'eval/flap_recall': 0.7163636363636363, 'eval/flap_f1': 0.7282809611829945, 'eval/flap_rate_model': 0.07388888888888889, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.7282809611829945, 'eval/latency_0/flap_precision': 0.7406015037593985, 'eval/latency_0/flap_recall': 0.7163636363636363, 'eval/latency_0/flap_f1': 0.7282809611829945, 'eval/latency_0/flap_rate_model': 0.07388888888888889, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.7282809611829945, 'timing/data': 0.019204873009584844, 'timing/model': 1.1451370459981263, 'learning_rate/qwen_vl_interface': 5.389974343395813e-07, 'learning_rate/action_model': 5.389974343395812e-06, 'epoch': 1.8}) +[2026-06-19 20:49:37,899][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_4800_state +[2026-06-19 20:49:37,902][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-19 20:49:41,168][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_4800_state/model.safetensors +[2026-06-19 20:49:48,329][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_4800_state/optimizer.bin +[2026-06-19 20:49:48,330][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_4800_state/sampler.bin +[2026-06-19 20:49:48,330][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_4800_state/sampler_1.bin +[2026-06-19 20:49:48,331][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_4800_state/random_states_0.pkl +[2026-06-19 20:49:48,352][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4800 +[2026-06-19 20:49:48,352][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-19 20:49:48,362][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-19 20:49:52,903][starVLA.training.train_starvla][INFO] - Step 4801, Loss: {'train/loss': 0.1008177399635315, 'train/grad_norm_pre_clip': 9.53171443939209, 'timing/data': 0.020621406030841172, 'timing/model': 1.1359173009404913, 'learning_rate/qwen_vl_interface': 5.38608962772717e-07, 'learning_rate/action_model': 5.38608962772717e-06, 'epoch': 1.8}) +[2026-06-19 20:49:57,377][starVLA.training.train_starvla][INFO] - Step 4802, Loss: {'train/loss': 0.15667788684368134, 'train/grad_norm_pre_clip': 15.765616416931152, 'timing/data': 0.017033134936355054, 'timing/model': 1.133967906003818, 'learning_rate/qwen_vl_interface': 5.382224278807429e-07, 'learning_rate/action_model': 5.382224278807429e-06, 'epoch': 1.8}) +[2026-06-19 20:50:01,941][starVLA.training.train_starvla][INFO] - Step 4803, Loss: {'train/loss': 0.2854824364185333, 'train/grad_norm_pre_clip': 19.475479125976562, 'timing/data': 0.01625471597071737, 'timing/model': 1.3276511579751968, 'learning_rate/qwen_vl_interface': 5.378378298225496e-07, 'learning_rate/action_model': 5.378378298225496e-06, 'epoch': 1.8}) +[2026-06-19 20:50:06,340][starVLA.training.train_starvla][INFO] - Step 4804, Loss: {'train/loss': 0.12732236087322235, 'train/grad_norm_pre_clip': 5.697295188903809, 'timing/data': 0.01717345998622477, 'timing/model': 1.1363536840071902, 'learning_rate/qwen_vl_interface': 5.374551687562306e-07, 'learning_rate/action_model': 5.374551687562306e-06, 'epoch': 1.8}) +[2026-06-19 20:50:10,874][starVLA.training.train_starvla][INFO] - Step 4805, Loss: {'train/loss': 0.1536005139350891, 'train/grad_norm_pre_clip': 6.614789962768555, 'timing/data': 0.016918414970859885, 'timing/model': 1.1323918850393966, 'learning_rate/qwen_vl_interface': 5.37074444839083e-07, 'learning_rate/action_model': 5.370744448390829e-06, 'epoch': 1.8}) +[2026-06-19 20:50:15,411][starVLA.training.train_starvla][INFO] - Step 4806, Loss: {'train/loss': 0.3023560643196106, 'train/grad_norm_pre_clip': 22.416223526000977, 'timing/data': 0.0165611490374431, 'timing/model': 1.1323379679815844, 'learning_rate/qwen_vl_interface': 5.366956582276085e-07, 'learning_rate/action_model': 5.366956582276084e-06, 'epoch': 1.8}) +[2026-06-19 20:50:19,852][starVLA.training.train_starvla][INFO] - Step 4807, Loss: {'train/loss': 0.10715270787477493, 'train/grad_norm_pre_clip': 12.062297821044922, 'timing/data': 0.01593696407508105, 'timing/model': 1.134765757014975, 'learning_rate/qwen_vl_interface': 5.363188090775122e-07, 'learning_rate/action_model': 5.363188090775122e-06, 'epoch': 1.8}) +[2026-06-19 20:50:24,388][starVLA.training.train_starvla][INFO] - Step 4808, Loss: {'train/loss': 0.052345067262649536, 'train/grad_norm_pre_clip': 5.573636531829834, 'timing/data': 0.0168205559020862, 'timing/model': 1.1456942160148174, 'learning_rate/qwen_vl_interface': 5.359438975437018e-07, 'learning_rate/action_model': 5.359438975437017e-06, 'epoch': 1.8}) +[2026-06-19 20:50:28,954][starVLA.training.train_starvla][INFO] - Step 4809, Loss: {'train/loss': 0.29530152678489685, 'train/grad_norm_pre_clip': 17.549394607543945, 'timing/data': 0.01679932198021561, 'timing/model': 1.3165512169944122, 'learning_rate/qwen_vl_interface': 5.355709237802895e-07, 'learning_rate/action_model': 5.3557092378028955e-06, 'epoch': 1.8}) +[2026-06-19 20:50:33,345][starVLA.training.train_starvla][INFO] - Step 4810, Loss: {'train/loss': 0.07797594368457794, 'train/grad_norm_pre_clip': 11.000816345214844, 'timing/data': 0.018392323050647974, 'timing/model': 1.1415842770365998, 'learning_rate/qwen_vl_interface': 5.351998879405911e-07, 'learning_rate/action_model': 5.351998879405911e-06, 'epoch': 1.8}) +[2026-06-19 20:50:37,921][starVLA.training.train_starvla][INFO] - Step 4811, Loss: {'train/loss': 0.13674095273017883, 'train/grad_norm_pre_clip': 12.648868560791016, 'timing/data': 0.017990238033235073, 'timing/model': 1.144256453961134, 'learning_rate/qwen_vl_interface': 5.348307901771256e-07, 'learning_rate/action_model': 5.348307901771256e-06, 'epoch': 1.8}) +[2026-06-19 20:50:42,502][starVLA.training.train_starvla][INFO] - Step 4812, Loss: {'train/loss': 0.12237591296434402, 'train/grad_norm_pre_clip': 10.554496765136719, 'timing/data': 0.016832150053232908, 'timing/model': 1.1483561299974099, 'learning_rate/qwen_vl_interface': 5.344636306416147e-07, 'learning_rate/action_model': 5.344636306416146e-06, 'epoch': 1.8}) +[2026-06-19 20:50:47,071][starVLA.training.train_starvla][INFO] - Step 4813, Loss: {'train/loss': 0.1465877741575241, 'train/grad_norm_pre_clip': 17.48050880432129, 'timing/data': 0.017696023918688297, 'timing/model': 1.1522383550181985, 'learning_rate/qwen_vl_interface': 5.340984094849836e-07, 'learning_rate/action_model': 5.340984094849836e-06, 'epoch': 1.8}) +[2026-06-19 20:50:51,566][starVLA.training.train_starvla][INFO] - Step 4814, Loss: {'train/loss': 0.2128903716802597, 'train/grad_norm_pre_clip': 17.67336654663086, 'timing/data': 0.01756301603745669, 'timing/model': 1.1365048539591953, 'learning_rate/qwen_vl_interface': 5.337351268573606e-07, 'learning_rate/action_model': 5.337351268573606e-06, 'epoch': 1.81}) +[2026-06-19 20:50:56,101][starVLA.training.train_starvla][INFO] - Step 4815, Loss: {'train/loss': 0.10643132030963898, 'train/grad_norm_pre_clip': 11.963102340698242, 'timing/data': 0.017613299074582756, 'timing/model': 1.2848653949331492, 'learning_rate/qwen_vl_interface': 5.33373782908079e-07, 'learning_rate/action_model': 5.333737829080789e-06, 'epoch': 1.81}) +[2026-06-19 20:51:00,495][starVLA.training.train_starvla][INFO] - Step 4816, Loss: {'train/loss': 0.2841367721557617, 'train/grad_norm_pre_clip': 24.84120750427246, 'timing/data': 0.018231036025099456, 'timing/model': 1.1397082750918344, 'learning_rate/qwen_vl_interface': 5.330143777856718e-07, 'learning_rate/action_model': 5.330143777856718e-06, 'epoch': 1.81}) +[2026-06-19 20:51:05,083][starVLA.training.train_starvla][INFO] - Step 4817, Loss: {'train/loss': 0.13055233657360077, 'train/grad_norm_pre_clip': 15.07437801361084, 'timing/data': 0.017599993967451155, 'timing/model': 1.1462576619815081, 'learning_rate/qwen_vl_interface': 5.326569116378786e-07, 'learning_rate/action_model': 5.326569116378786e-06, 'epoch': 1.81}) +[2026-06-19 20:51:09,666][starVLA.training.train_starvla][INFO] - Step 4818, Loss: {'train/loss': 0.13212993741035461, 'train/grad_norm_pre_clip': 16.07562255859375, 'timing/data': 0.018709111027419567, 'timing/model': 1.1414302189368755, 'learning_rate/qwen_vl_interface': 5.323013846116387e-07, 'learning_rate/action_model': 5.323013846116386e-06, 'epoch': 1.81}) +[2026-06-19 20:51:14,228][starVLA.training.train_starvla][INFO] - Step 4819, Loss: {'train/loss': 0.1452489048242569, 'train/grad_norm_pre_clip': 12.794965744018555, 'timing/data': 0.017553737969137728, 'timing/model': 1.1502597259823233, 'learning_rate/qwen_vl_interface': 5.31947796853097e-07, 'learning_rate/action_model': 5.3194779685309696e-06, 'epoch': 1.81}) +[2026-06-19 20:51:18,805][starVLA.training.train_starvla][INFO] - Step 4820, Loss: {'train/loss': 0.44096624851226807, 'train/grad_norm_pre_clip': 52.959716796875, 'timing/data': 0.017563804984092712, 'timing/model': 1.1408326150849462, 'learning_rate/qwen_vl_interface': 5.315961485075992e-07, 'learning_rate/action_model': 5.315961485075993e-06, 'epoch': 1.81}) +[2026-06-19 20:51:23,432][starVLA.training.train_starvla][INFO] - Step 4821, Loss: {'train/loss': 0.13891643285751343, 'train/grad_norm_pre_clip': 11.647384643554688, 'timing/data': 0.024051825050264597, 'timing/model': 1.2938157290918753, 'learning_rate/qwen_vl_interface': 5.312464397196958e-07, 'learning_rate/action_model': 5.312464397196958e-06, 'epoch': 1.81}) +[2026-06-19 20:51:27,846][starVLA.training.train_starvla][INFO] - Step 4822, Loss: {'train/loss': 0.2895270586013794, 'train/grad_norm_pre_clip': 15.804941177368164, 'timing/data': 0.016926120035350323, 'timing/model': 1.143591589992866, 'learning_rate/qwen_vl_interface': 5.308986706331377e-07, 'learning_rate/action_model': 5.308986706331376e-06, 'epoch': 1.81}) +[2026-06-19 20:51:32,450][starVLA.training.train_starvla][INFO] - Step 4823, Loss: {'train/loss': 0.14733454585075378, 'train/grad_norm_pre_clip': 16.99561309814453, 'timing/data': 0.017486097058281302, 'timing/model': 1.1446758579695597, 'learning_rate/qwen_vl_interface': 5.305528413908797e-07, 'learning_rate/action_model': 5.305528413908798e-06, 'epoch': 1.81}) +[2026-06-19 20:51:36,958][starVLA.training.train_starvla][INFO] - Step 4824, Loss: {'train/loss': 0.3799605071544647, 'train/grad_norm_pre_clip': 16.498714447021484, 'timing/data': 0.017632217961363494, 'timing/model': 1.1440081729087979, 'learning_rate/qwen_vl_interface': 5.302089521350799e-07, 'learning_rate/action_model': 5.3020895213507996e-06, 'epoch': 1.81}) +[2026-06-19 20:51:41,483][starVLA.training.train_starvla][INFO] - Step 4825, Loss: {'train/loss': 0.09881766885519028, 'train/grad_norm_pre_clip': 8.805397033691406, 'timing/data': 0.01775269105564803, 'timing/model': 1.1522726149996743, 'learning_rate/qwen_vl_interface': 5.298670030070977e-07, 'learning_rate/action_model': 5.298670030070976e-06, 'epoch': 1.81}) +[2026-06-19 20:51:46,079][starVLA.training.train_starvla][INFO] - Step 4826, Loss: {'train/loss': 0.23523667454719543, 'train/grad_norm_pre_clip': 19.34558868408203, 'timing/data': 0.016769676934927702, 'timing/model': 1.149653500993736, 'learning_rate/qwen_vl_interface': 5.295269941474957e-07, 'learning_rate/action_model': 5.295269941474957e-06, 'epoch': 1.81}) +[2026-06-19 20:51:50,684][starVLA.training.train_starvla][INFO] - Step 4827, Loss: {'train/loss': 0.16952386498451233, 'train/grad_norm_pre_clip': 15.537766456604004, 'timing/data': 0.018078409950248897, 'timing/model': 1.324009453994222, 'learning_rate/qwen_vl_interface': 5.291889256960384e-07, 'learning_rate/action_model': 5.291889256960384e-06, 'epoch': 1.81}) +[2026-06-19 20:51:55,108][starVLA.training.train_starvla][INFO] - Step 4828, Loss: {'train/loss': 0.2276928871870041, 'train/grad_norm_pre_clip': 19.67597770690918, 'timing/data': 0.017069805064238608, 'timing/model': 1.1420263070613146, 'learning_rate/qwen_vl_interface': 5.288527977916929e-07, 'learning_rate/action_model': 5.288527977916929e-06, 'epoch': 1.81}) +[2026-06-19 20:51:59,713][starVLA.training.train_starvla][INFO] - Step 4829, Loss: {'train/loss': 0.2774013876914978, 'train/grad_norm_pre_clip': 29.83613395690918, 'timing/data': 0.01787679991684854, 'timing/model': 1.1535400039283559, 'learning_rate/qwen_vl_interface': 5.28518610572629e-07, 'learning_rate/action_model': 5.28518610572629e-06, 'epoch': 1.81}) +[2026-06-19 20:52:04,292][starVLA.training.train_starvla][INFO] - Step 4830, Loss: {'train/loss': 0.36747121810913086, 'train/grad_norm_pre_clip': 58.571964263916016, 'timing/data': 0.017666978994384408, 'timing/model': 1.1418787870788947, 'learning_rate/qwen_vl_interface': 5.281863641762184e-07, 'learning_rate/action_model': 5.281863641762184e-06, 'epoch': 1.81}) +[2026-06-19 20:52:08,779][starVLA.training.train_starvla][INFO] - Step 4831, Loss: {'train/loss': 0.253101110458374, 'train/grad_norm_pre_clip': 20.20256805419922, 'timing/data': 0.01886206795461476, 'timing/model': 1.1488416959764436, 'learning_rate/qwen_vl_interface': 5.278560587390355e-07, 'learning_rate/action_model': 5.278560587390355e-06, 'epoch': 1.81}) +[2026-06-19 20:52:13,303][starVLA.training.train_starvla][INFO] - Step 4832, Loss: {'train/loss': 0.08570639044046402, 'train/grad_norm_pre_clip': 10.027698516845703, 'timing/data': 0.016956627019681036, 'timing/model': 1.1424558939179406, 'learning_rate/qwen_vl_interface': 5.275276943968557e-07, 'learning_rate/action_model': 5.275276943968557e-06, 'epoch': 1.81}) +[2026-06-19 20:52:17,895][starVLA.training.train_starvla][INFO] - Step 4833, Loss: {'train/loss': 0.12580662965774536, 'train/grad_norm_pre_clip': 9.256105422973633, 'timing/data': 0.017555098049342632, 'timing/model': 1.3209568549646065, 'learning_rate/qwen_vl_interface': 5.272012712846566e-07, 'learning_rate/action_model': 5.2720127128465654e-06, 'epoch': 1.81}) +[2026-06-19 20:52:22,313][starVLA.training.train_starvla][INFO] - Step 4834, Loss: {'train/loss': 0.23814257979393005, 'train/grad_norm_pre_clip': 32.28611755371094, 'timing/data': 0.01835166907403618, 'timing/model': 1.143489405978471, 'learning_rate/qwen_vl_interface': 5.268767895366198e-07, 'learning_rate/action_model': 5.268767895366198e-06, 'epoch': 1.81}) +[2026-06-19 20:52:26,889][starVLA.training.train_starvla][INFO] - Step 4835, Loss: {'train/loss': 0.2307533621788025, 'train/grad_norm_pre_clip': 11.268499374389648, 'timing/data': 0.01974258804693818, 'timing/model': 1.14587848901283, 'learning_rate/qwen_vl_interface': 5.265542492861268e-07, 'learning_rate/action_model': 5.2655424928612676e-06, 'epoch': 1.81}) +[2026-06-19 20:52:31,473][starVLA.training.train_starvla][INFO] - Step 4836, Loss: {'train/loss': 0.20462776720523834, 'train/grad_norm_pre_clip': 20.30350685119629, 'timing/data': 0.01783921103924513, 'timing/model': 1.1496468980330974, 'learning_rate/qwen_vl_interface': 5.262336506657617e-07, 'learning_rate/action_model': 5.2623365066576165e-06, 'epoch': 1.81}) +[2026-06-19 20:52:36,051][starVLA.training.train_starvla][INFO] - Step 4837, Loss: {'train/loss': 0.16547797620296478, 'train/grad_norm_pre_clip': 15.655994415283203, 'timing/data': 0.015945485094562173, 'timing/model': 1.150426103034988, 'learning_rate/qwen_vl_interface': 5.259149938073103e-07, 'learning_rate/action_model': 5.259149938073102e-06, 'epoch': 1.81}) +[2026-06-19 20:52:40,663][starVLA.training.train_starvla][INFO] - Step 4838, Loss: {'train/loss': 0.24769552052021027, 'train/grad_norm_pre_clip': 18.531143188476562, 'timing/data': 0.016027920064516366, 'timing/model': 1.161653561051935, 'learning_rate/qwen_vl_interface': 5.255982788417609e-07, 'learning_rate/action_model': 5.255982788417608e-06, 'epoch': 1.81}) +[2026-06-19 20:52:45,236][starVLA.training.train_starvla][INFO] - Step 4839, Loss: {'train/loss': 0.1932474821805954, 'train/grad_norm_pre_clip': 18.323360443115234, 'timing/data': 0.01546536199748516, 'timing/model': 1.3046948839910328, 'learning_rate/qwen_vl_interface': 5.252835058993029e-07, 'learning_rate/action_model': 5.252835058993029e-06, 'epoch': 1.81}) +[2026-06-19 20:52:49,662][starVLA.training.train_starvla][INFO] - Step 4840, Loss: {'train/loss': 0.35520198941230774, 'train/grad_norm_pre_clip': 20.765396118164062, 'timing/data': 0.016262885998003185, 'timing/model': 1.1563245470169932, 'learning_rate/qwen_vl_interface': 5.249706751093274e-07, 'learning_rate/action_model': 5.249706751093273e-06, 'epoch': 1.81}) +[2026-06-19 20:52:54,272][starVLA.training.train_starvla][INFO] - Step 4841, Loss: {'train/loss': 0.4071175754070282, 'train/grad_norm_pre_clip': 22.266887664794922, 'timing/data': 0.01774290599860251, 'timing/model': 1.163994487025775, 'learning_rate/qwen_vl_interface': 5.246597866004273e-07, 'learning_rate/action_model': 5.246597866004273e-06, 'epoch': 1.82}) +[2026-06-19 20:52:58,805][starVLA.training.train_starvla][INFO] - Step 4842, Loss: {'train/loss': 0.22293734550476074, 'train/grad_norm_pre_clip': 15.204216957092285, 'timing/data': 0.01779117190744728, 'timing/model': 1.1416032240958884, 'learning_rate/qwen_vl_interface': 5.243508405003975e-07, 'learning_rate/action_model': 5.243508405003975e-06, 'epoch': 1.82}) +[2026-06-19 20:53:03,431][starVLA.training.train_starvla][INFO] - Step 4843, Loss: {'train/loss': 0.31090056896209717, 'train/grad_norm_pre_clip': 18.4289608001709, 'timing/data': 0.016528567066416144, 'timing/model': 1.1620392909971997, 'learning_rate/qwen_vl_interface': 5.240438369362336e-07, 'learning_rate/action_model': 5.2404383693623366e-06, 'epoch': 1.82}) +[2026-06-19 20:53:08,030][starVLA.training.train_starvla][INFO] - Step 4844, Loss: {'train/loss': 0.32074183225631714, 'train/grad_norm_pre_clip': 22.636903762817383, 'timing/data': 0.017012540018185973, 'timing/model': 1.1530256300466135, 'learning_rate/qwen_vl_interface': 5.237387760341335e-07, 'learning_rate/action_model': 5.237387760341335e-06, 'epoch': 1.82}) +[2026-06-19 20:53:12,627][starVLA.training.train_starvla][INFO] - Step 4845, Loss: {'train/loss': 0.032169852405786514, 'train/grad_norm_pre_clip': 5.1689453125, 'timing/data': 0.017740673036314547, 'timing/model': 1.3184219000395387, 'learning_rate/qwen_vl_interface': 5.234356579194963e-07, 'learning_rate/action_model': 5.234356579194963e-06, 'epoch': 1.82}) +[2026-06-19 20:53:17,064][starVLA.training.train_starvla][INFO] - Step 4846, Loss: {'train/loss': 0.25654998421669006, 'train/grad_norm_pre_clip': 20.62955093383789, 'timing/data': 0.01853848807513714, 'timing/model': 1.1552834259346128, 'learning_rate/qwen_vl_interface': 5.231344827169218e-07, 'learning_rate/action_model': 5.231344827169217e-06, 'epoch': 1.82}) +[2026-06-19 20:53:21,655][starVLA.training.train_starvla][INFO] - Step 4847, Loss: {'train/loss': 0.07537903636693954, 'train/grad_norm_pre_clip': 7.018333911895752, 'timing/data': 0.017396549927070737, 'timing/model': 1.145225549931638, 'learning_rate/qwen_vl_interface': 5.228352505502119e-07, 'learning_rate/action_model': 5.228352505502119e-06, 'epoch': 1.82}) +[2026-06-19 20:53:26,263][starVLA.training.train_starvla][INFO] - Step 4848, Loss: {'train/loss': 0.14434437453746796, 'train/grad_norm_pre_clip': 15.88896656036377, 'timing/data': 0.016738249105401337, 'timing/model': 1.1591090619331226, 'learning_rate/qwen_vl_interface': 5.225379615423701e-07, 'learning_rate/action_model': 5.2253796154237e-06, 'epoch': 1.82}) +[2026-06-19 20:53:30,883][starVLA.training.train_starvla][INFO] - Step 4849, Loss: {'train/loss': 0.18677079677581787, 'train/grad_norm_pre_clip': 16.32274627685547, 'timing/data': 0.01575388095807284, 'timing/model': 1.1428367471089587, 'learning_rate/qwen_vl_interface': 5.222426158156008e-07, 'learning_rate/action_model': 5.222426158156008e-06, 'epoch': 1.82}) +[2026-06-19 20:53:35,485][starVLA.training.train_starvla][INFO] - Step 4850, Loss: {'train/loss': 0.0865924134850502, 'train/grad_norm_pre_clip': 8.986632347106934, 'timing/data': 0.016486985026858747, 'timing/model': 1.1425685930298641, 'learning_rate/qwen_vl_interface': 5.219492134913084e-07, 'learning_rate/action_model': 5.219492134913084e-06, 'epoch': 1.82}) +[2026-06-19 20:53:40,035][starVLA.training.train_starvla][INFO] - Step 4851, Loss: {'train/loss': 0.3986274003982544, 'train/grad_norm_pre_clip': 26.283687591552734, 'timing/data': 0.0182719569420442, 'timing/model': 1.1443112769629806, 'learning_rate/qwen_vl_interface': 5.216577546901008e-07, 'learning_rate/action_model': 5.216577546901007e-06, 'epoch': 1.82}) +[2026-06-19 20:53:44,655][starVLA.training.train_starvla][INFO] - Step 4852, Loss: {'train/loss': 0.10241764038801193, 'train/grad_norm_pre_clip': 13.377646446228027, 'timing/data': 0.016739654005505145, 'timing/model': 1.3132796009304002, 'learning_rate/qwen_vl_interface': 5.213682395317852e-07, 'learning_rate/action_model': 5.213682395317851e-06, 'epoch': 1.82}) +[2026-06-19 20:53:49,081][starVLA.training.train_starvla][INFO] - Step 4853, Loss: {'train/loss': 0.2649690508842468, 'train/grad_norm_pre_clip': 19.43951416015625, 'timing/data': 0.016671992023475468, 'timing/model': 1.1461998970480636, 'learning_rate/qwen_vl_interface': 5.2108066813537e-07, 'learning_rate/action_model': 5.2108066813537e-06, 'epoch': 1.82}) +[2026-06-19 20:53:53,692][starVLA.training.train_starvla][INFO] - Step 4854, Loss: {'train/loss': 0.13724417984485626, 'train/grad_norm_pre_clip': 13.03152084350586, 'timing/data': 0.01607307605445385, 'timing/model': 1.1521513659972697, 'learning_rate/qwen_vl_interface': 5.207950406190654e-07, 'learning_rate/action_model': 5.207950406190653e-06, 'epoch': 1.82}) +[2026-06-19 20:53:58,309][starVLA.training.train_starvla][INFO] - Step 4855, Loss: {'train/loss': 0.09106631577014923, 'train/grad_norm_pre_clip': 19.0264949798584, 'timing/data': 0.016433016047813, 'timing/model': 1.1593262100359425, 'learning_rate/qwen_vl_interface': 5.205113571002818e-07, 'learning_rate/action_model': 5.205113571002818e-06, 'epoch': 1.82}) +[2026-06-19 20:54:02,895][starVLA.training.train_starvla][INFO] - Step 4856, Loss: {'train/loss': 0.12335601449012756, 'train/grad_norm_pre_clip': 8.599184036254883, 'timing/data': 0.017511868034489453, 'timing/model': 1.1461668950505555, 'learning_rate/qwen_vl_interface': 5.202296176956308e-07, 'learning_rate/action_model': 5.202296176956308e-06, 'epoch': 1.82}) +[2026-06-19 20:54:07,482][starVLA.training.train_starvla][INFO] - Step 4857, Loss: {'train/loss': 0.21509641408920288, 'train/grad_norm_pre_clip': 19.138248443603516, 'timing/data': 0.017020671046338975, 'timing/model': 1.3100741499802098, 'learning_rate/qwen_vl_interface': 5.19949822520925e-07, 'learning_rate/action_model': 5.19949822520925e-06, 'epoch': 1.82}) +[2026-06-19 20:54:12,065][starVLA.training.train_starvla][INFO] - Step 4858, Loss: {'train/loss': 0.05646247789263725, 'train/grad_norm_pre_clip': 9.81747055053711, 'timing/data': 0.017178395064547658, 'timing/model': 1.3114811549894512, 'learning_rate/qwen_vl_interface': 5.196719716911777e-07, 'learning_rate/action_model': 5.1967197169117764e-06, 'epoch': 1.82}) +[2026-06-19 20:54:16,489][starVLA.training.train_starvla][INFO] - Step 4859, Loss: {'train/loss': 0.1507253795862198, 'train/grad_norm_pre_clip': 17.87099838256836, 'timing/data': 0.017526900046505034, 'timing/model': 1.150894035003148, 'learning_rate/qwen_vl_interface': 5.193960653206029e-07, 'learning_rate/action_model': 5.193960653206029e-06, 'epoch': 1.82}) +[2026-06-19 20:54:21,086][starVLA.training.train_starvla][INFO] - Step 4860, Loss: {'train/loss': 0.14884841442108154, 'train/grad_norm_pre_clip': 9.543272018432617, 'timing/data': 0.01831644505728036, 'timing/model': 1.1466284899506718, 'learning_rate/qwen_vl_interface': 5.191221035226144e-07, 'learning_rate/action_model': 5.191221035226144e-06, 'epoch': 1.82}) +[2026-06-19 20:54:25,710][starVLA.training.train_starvla][INFO] - Step 4861, Loss: {'train/loss': 0.0795341432094574, 'train/grad_norm_pre_clip': 6.120978355407715, 'timing/data': 0.01842371397651732, 'timing/model': 1.1459501800127327, 'learning_rate/qwen_vl_interface': 5.188500864098292e-07, 'learning_rate/action_model': 5.1885008640982915e-06, 'epoch': 1.82}) +[2026-06-19 20:54:30,313][starVLA.training.train_starvla][INFO] - Step 4862, Loss: {'train/loss': 0.19035708904266357, 'train/grad_norm_pre_clip': 15.308528900146484, 'timing/data': 0.01763179199770093, 'timing/model': 1.1487878100015223, 'learning_rate/qwen_vl_interface': 5.185800140940613e-07, 'learning_rate/action_model': 5.185800140940613e-06, 'epoch': 1.82}) +[2026-06-19 20:54:34,909][starVLA.training.train_starvla][INFO] - Step 4863, Loss: {'train/loss': 0.2793383002281189, 'train/grad_norm_pre_clip': 20.625333786010742, 'timing/data': 0.017917896970175207, 'timing/model': 1.321029750048183, 'learning_rate/qwen_vl_interface': 5.183118866863288e-07, 'learning_rate/action_model': 5.1831188668632875e-06, 'epoch': 1.82}) +[2026-06-19 20:54:39,357][starVLA.training.train_starvla][INFO] - Step 4864, Loss: {'train/loss': 0.15218323469161987, 'train/grad_norm_pre_clip': 14.427790641784668, 'timing/data': 0.016860755975358188, 'timing/model': 1.1530976200010628, 'learning_rate/qwen_vl_interface': 5.180457042968484e-07, 'learning_rate/action_model': 5.1804570429684835e-06, 'epoch': 1.82}) +[2026-06-19 20:54:43,977][starVLA.training.train_starvla][INFO] - Step 4865, Loss: {'train/loss': 0.2778216600418091, 'train/grad_norm_pre_clip': 21.828519821166992, 'timing/data': 0.01990132499486208, 'timing/model': 1.1611193779390305, 'learning_rate/qwen_vl_interface': 5.17781467035037e-07, 'learning_rate/action_model': 5.177814670350371e-06, 'epoch': 1.82}) +[2026-06-19 20:54:48,556][starVLA.training.train_starvla][INFO] - Step 4866, Loss: {'train/loss': 0.1717631220817566, 'train/grad_norm_pre_clip': 12.020539283752441, 'timing/data': 0.018902882002294064, 'timing/model': 1.149416088941507, 'learning_rate/qwen_vl_interface': 5.175191750095133e-07, 'learning_rate/action_model': 5.175191750095132e-06, 'epoch': 1.82}) +[2026-06-19 20:54:53,180][starVLA.training.train_starvla][INFO] - Step 4867, Loss: {'train/loss': 0.12251175940036774, 'train/grad_norm_pre_clip': 13.5789213180542, 'timing/data': 0.017638028948567808, 'timing/model': 1.1499440500047058, 'learning_rate/qwen_vl_interface': 5.172588283280953e-07, 'learning_rate/action_model': 5.172588283280953e-06, 'epoch': 1.82}) +[2026-06-19 20:54:57,776][starVLA.training.train_starvla][INFO] - Step 4868, Loss: {'train/loss': 0.2080889493227005, 'train/grad_norm_pre_clip': 18.460758209228516, 'timing/data': 0.019149744068272412, 'timing/model': 1.1499005409423262, 'learning_rate/qwen_vl_interface': 5.170004270978019e-07, 'learning_rate/action_model': 5.170004270978018e-06, 'epoch': 1.83}) +[2026-06-19 20:55:02,387][starVLA.training.train_starvla][INFO] - Step 4869, Loss: {'train/loss': 0.3309805989265442, 'train/grad_norm_pre_clip': 21.13397216796875, 'timing/data': 0.017993489978834987, 'timing/model': 1.325769840972498, 'learning_rate/qwen_vl_interface': 5.167439714248512e-07, 'learning_rate/action_model': 5.1674397142485116e-06, 'epoch': 1.83}) +[2026-06-19 20:55:07,003][starVLA.training.train_starvla][INFO] - Step 4870, Loss: {'train/loss': 0.09497224539518356, 'train/grad_norm_pre_clip': 7.94087553024292, 'timing/data': 0.017369990004226565, 'timing/model': 1.3227688840124756, 'learning_rate/qwen_vl_interface': 5.164894614146633e-07, 'learning_rate/action_model': 5.164894614146634e-06, 'epoch': 1.83}) +[2026-06-19 20:55:11,450][starVLA.training.train_starvla][INFO] - Step 4871, Loss: {'train/loss': 0.19143494963645935, 'train/grad_norm_pre_clip': 14.534173965454102, 'timing/data': 0.017899267957545817, 'timing/model': 1.1502922220388427, 'learning_rate/qwen_vl_interface': 5.16236897171858e-07, 'learning_rate/action_model': 5.16236897171858e-06, 'epoch': 1.83}) +[2026-06-19 20:55:16,055][starVLA.training.train_starvla][INFO] - Step 4872, Loss: {'train/loss': 0.19865936040878296, 'train/grad_norm_pre_clip': 14.312586784362793, 'timing/data': 0.018108309945091605, 'timing/model': 1.1495182960061356, 'learning_rate/qwen_vl_interface': 5.15986278800254e-07, 'learning_rate/action_model': 5.15986278800254e-06, 'epoch': 1.83}) +[2026-06-19 20:55:20,673][starVLA.training.train_starvla][INFO] - Step 4873, Loss: {'train/loss': 0.2354232370853424, 'train/grad_norm_pre_clip': 9.210007667541504, 'timing/data': 0.01826293906196952, 'timing/model': 1.1480130080599338, 'learning_rate/qwen_vl_interface': 5.157376064028719e-07, 'learning_rate/action_model': 5.157376064028718e-06, 'epoch': 1.83}) +[2026-06-19 20:55:25,174][starVLA.training.train_starvla][INFO] - Step 4874, Loss: {'train/loss': 0.1045655608177185, 'train/grad_norm_pre_clip': 7.0048980712890625, 'timing/data': 0.01844126603100449, 'timing/model': 1.143158908933401, 'learning_rate/qwen_vl_interface': 5.154908800819307e-07, 'learning_rate/action_model': 5.154908800819306e-06, 'epoch': 1.83}) +[2026-06-19 20:55:29,676][starVLA.training.train_starvla][INFO] - Step 4875, Loss: {'train/loss': 0.16824862360954285, 'train/grad_norm_pre_clip': 11.062185287475586, 'timing/data': 0.017213392071425915, 'timing/model': 1.1460355389863253, 'learning_rate/qwen_vl_interface': 5.152460999388504e-07, 'learning_rate/action_model': 5.152460999388503e-06, 'epoch': 1.83}) +[2026-06-19 20:55:34,216][starVLA.training.train_starvla][INFO] - Step 4876, Loss: {'train/loss': 0.3224124610424042, 'train/grad_norm_pre_clip': 32.76121520996094, 'timing/data': 0.017900782055221498, 'timing/model': 1.2490992810344324, 'learning_rate/qwen_vl_interface': 5.150032660742517e-07, 'learning_rate/action_model': 5.150032660742517e-06, 'epoch': 1.83}) +[2026-06-19 20:55:38,648][starVLA.training.train_starvla][INFO] - Step 4877, Loss: {'train/loss': 0.17407064139842987, 'train/grad_norm_pre_clip': 14.83801555633545, 'timing/data': 0.01664331299252808, 'timing/model': 1.1510374440113083, 'learning_rate/qwen_vl_interface': 5.147623785879535e-07, 'learning_rate/action_model': 5.147623785879534e-06, 'epoch': 1.83}) +[2026-06-19 20:55:43,242][starVLA.training.train_starvla][INFO] - Step 4878, Loss: {'train/loss': 0.14010733366012573, 'train/grad_norm_pre_clip': 12.006842613220215, 'timing/data': 0.016287832986563444, 'timing/model': 1.1553875559475273, 'learning_rate/qwen_vl_interface': 5.145234375789758e-07, 'learning_rate/action_model': 5.145234375789758e-06, 'epoch': 1.83}) +[2026-06-19 20:55:48,008][starVLA.training.train_starvla][INFO] - Step 4879, Loss: {'train/loss': 0.2286709100008011, 'train/grad_norm_pre_clip': 15.68450927734375, 'timing/data': 0.016721304040402174, 'timing/model': 1.152902880916372, 'learning_rate/qwen_vl_interface': 5.142864431455387e-07, 'learning_rate/action_model': 5.1428644314553875e-06, 'epoch': 1.83}) +[2026-06-19 20:55:52,622][starVLA.training.train_starvla][INFO] - Step 4880, Loss: {'train/loss': 0.1464112251996994, 'train/grad_norm_pre_clip': 21.468875885009766, 'timing/data': 0.016344948089681566, 'timing/model': 1.1659936580108479, 'learning_rate/qwen_vl_interface': 5.140513953850612e-07, 'learning_rate/action_model': 5.140513953850612e-06, 'epoch': 1.83}) +[2026-06-19 20:55:57,220][starVLA.training.train_starvla][INFO] - Step 4881, Loss: {'train/loss': 0.23649384081363678, 'train/grad_norm_pre_clip': 20.37803840637207, 'timing/data': 0.0159151719417423, 'timing/model': 1.3193579909857363, 'learning_rate/qwen_vl_interface': 5.138182943941623e-07, 'learning_rate/action_model': 5.138182943941623e-06, 'epoch': 1.83}) +[2026-06-19 20:56:01,763][starVLA.training.train_starvla][INFO] - Step 4882, Loss: {'train/loss': 0.4206187129020691, 'train/grad_norm_pre_clip': 33.213504791259766, 'timing/data': 0.016858650953508914, 'timing/model': 1.2553312010131776, 'learning_rate/qwen_vl_interface': 5.135871402686615e-07, 'learning_rate/action_model': 5.135871402686615e-06, 'epoch': 1.83}) +[2026-06-19 20:56:06,210][starVLA.training.train_starvla][INFO] - Step 4883, Loss: {'train/loss': 0.12730266153812408, 'train/grad_norm_pre_clip': 15.653076171875, 'timing/data': 0.017502780072391033, 'timing/model': 1.1640886940294877, 'learning_rate/qwen_vl_interface': 5.133579331035777e-07, 'learning_rate/action_model': 5.133579331035777e-06, 'epoch': 1.83}) +[2026-06-19 20:56:10,806][starVLA.training.train_starvla][INFO] - Step 4884, Loss: {'train/loss': 0.18108311295509338, 'train/grad_norm_pre_clip': 15.0501127243042, 'timing/data': 0.017528604017570615, 'timing/model': 1.1478797290474176, 'learning_rate/qwen_vl_interface': 5.131306729931284e-07, 'learning_rate/action_model': 5.1313067299312836e-06, 'epoch': 1.83}) +[2026-06-19 20:56:15,441][starVLA.training.train_starvla][INFO] - Step 4885, Loss: {'train/loss': 0.15726520121097565, 'train/grad_norm_pre_clip': 15.93820858001709, 'timing/data': 0.01998046599328518, 'timing/model': 1.1686421179911122, 'learning_rate/qwen_vl_interface': 5.129053600307326e-07, 'learning_rate/action_model': 5.129053600307326e-06, 'epoch': 1.83}) +[2026-06-19 20:56:20,053][starVLA.training.train_starvla][INFO] - Step 4886, Loss: {'train/loss': 0.0888320803642273, 'train/grad_norm_pre_clip': 15.946173667907715, 'timing/data': 0.0176154290093109, 'timing/model': 1.1455160070909187, 'learning_rate/qwen_vl_interface': 5.126819943090078e-07, 'learning_rate/action_model': 5.1268199430900775e-06, 'epoch': 1.83}) +[2026-06-19 20:56:24,665][starVLA.training.train_starvla][INFO] - Step 4887, Loss: {'train/loss': 0.11609823256731033, 'train/grad_norm_pre_clip': 11.948588371276855, 'timing/data': 0.02009155403356999, 'timing/model': 1.328720960067585, 'learning_rate/qwen_vl_interface': 5.124605759197706e-07, 'learning_rate/action_model': 5.124605759197706e-06, 'epoch': 1.83}) +[2026-06-19 20:56:29,224][starVLA.training.train_starvla][INFO] - Step 4888, Loss: {'train/loss': 0.24374808371067047, 'train/grad_norm_pre_clip': 24.34841537475586, 'timing/data': 0.017936620977707207, 'timing/model': 1.2693976379232481, 'learning_rate/qwen_vl_interface': 5.122411049540383e-07, 'learning_rate/action_model': 5.1224110495403825e-06, 'epoch': 1.83}) +[2026-06-19 20:56:33,670][starVLA.training.train_starvla][INFO] - Step 4889, Loss: {'train/loss': 0.23679964244365692, 'train/grad_norm_pre_clip': 19.86180305480957, 'timing/data': 0.018403704976662993, 'timing/model': 1.1467843640130013, 'learning_rate/qwen_vl_interface': 5.120235815020269e-07, 'learning_rate/action_model': 5.120235815020269e-06, 'epoch': 1.83}) +[2026-06-19 20:56:38,285][starVLA.training.train_starvla][INFO] - Step 4890, Loss: {'train/loss': 0.1640814244747162, 'train/grad_norm_pre_clip': 13.094941139221191, 'timing/data': 0.016998067032545805, 'timing/model': 1.1435402060160413, 'learning_rate/qwen_vl_interface': 5.118080056531524e-07, 'learning_rate/action_model': 5.118080056531524e-06, 'epoch': 1.83}) +[2026-06-19 20:56:42,882][starVLA.training.train_starvla][INFO] - Step 4891, Loss: {'train/loss': 0.11726120859384537, 'train/grad_norm_pre_clip': 22.873998641967773, 'timing/data': 0.01812497992068529, 'timing/model': 1.15332176303491, 'learning_rate/qwen_vl_interface': 5.115943774960295e-07, 'learning_rate/action_model': 5.115943774960295e-06, 'epoch': 1.83}) +[2026-06-19 20:56:47,480][starVLA.training.train_starvla][INFO] - Step 4892, Loss: {'train/loss': 0.1953113079071045, 'train/grad_norm_pre_clip': 16.94223403930664, 'timing/data': 0.01774808601476252, 'timing/model': 1.1541122989729047, 'learning_rate/qwen_vl_interface': 5.113826971184727e-07, 'learning_rate/action_model': 5.113826971184727e-06, 'epoch': 1.83}) +[2026-06-19 20:56:52,087][starVLA.training.train_starvla][INFO] - Step 4893, Loss: {'train/loss': 0.08923157304525375, 'train/grad_norm_pre_clip': 8.19950008392334, 'timing/data': 0.01817640697117895, 'timing/model': 1.1485538170672953, 'learning_rate/qwen_vl_interface': 5.111729646074954e-07, 'learning_rate/action_model': 5.111729646074954e-06, 'epoch': 1.83}) +[2026-06-19 20:56:56,712][starVLA.training.train_starvla][INFO] - Step 4894, Loss: {'train/loss': 0.13742958009243011, 'train/grad_norm_pre_clip': 17.094152450561523, 'timing/data': 0.017793159931898117, 'timing/model': 1.328732882044278, 'learning_rate/qwen_vl_interface': 5.10965180049312e-07, 'learning_rate/action_model': 5.10965180049312e-06, 'epoch': 1.84}) +[2026-06-19 20:57:01,156][starVLA.training.train_starvla][INFO] - Step 4895, Loss: {'train/loss': 0.03157869353890419, 'train/grad_norm_pre_clip': 5.688024044036865, 'timing/data': 0.01864997297525406, 'timing/model': 1.1477641679812223, 'learning_rate/qwen_vl_interface': 5.107593435293336e-07, 'learning_rate/action_model': 5.107593435293336e-06, 'epoch': 1.84}) +[2026-06-19 20:57:05,757][starVLA.training.train_starvla][INFO] - Step 4896, Loss: {'train/loss': 0.29366952180862427, 'train/grad_norm_pre_clip': 20.068655014038086, 'timing/data': 0.017591975978575647, 'timing/model': 1.1488417220534757, 'learning_rate/qwen_vl_interface': 5.105554551321724e-07, 'learning_rate/action_model': 5.105554551321723e-06, 'epoch': 1.84}) +[2026-06-19 20:57:10,371][starVLA.training.train_starvla][INFO] - Step 4897, Loss: {'train/loss': 0.2338222712278366, 'train/grad_norm_pre_clip': 19.488935470581055, 'timing/data': 0.017764079966582358, 'timing/model': 1.1514405569760129, 'learning_rate/qwen_vl_interface': 5.103535149416389e-07, 'learning_rate/action_model': 5.1035351494163885e-06, 'epoch': 1.84}) +[2026-06-19 20:57:14,989][starVLA.training.train_starvla][INFO] - Step 4898, Loss: {'train/loss': 0.06227422505617142, 'train/grad_norm_pre_clip': 8.977731704711914, 'timing/data': 0.01912901201285422, 'timing/model': 1.1426437180489302, 'learning_rate/qwen_vl_interface': 5.101535230407436e-07, 'learning_rate/action_model': 5.101535230407436e-06, 'epoch': 1.84}) +[2026-06-19 20:57:19,587][starVLA.training.train_starvla][INFO] - Step 4899, Loss: {'train/loss': 0.2150987982749939, 'train/grad_norm_pre_clip': 15.932999610900879, 'timing/data': 0.017384764971211553, 'timing/model': 1.1449393869843334, 'learning_rate/qwen_vl_interface': 5.099554795116956e-07, 'learning_rate/action_model': 5.0995547951169565e-06, 'epoch': 1.84}) +[2026-06-19 20:57:24,206][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 4900 +[2026-06-19 21:00:17,588][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 4900 +[2026-06-19 21:00:17,588][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 4900 +[2026-06-19 21:03:28,872][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 4900 +[2026-06-19 21:03:28,874][starVLA.training.train_starvla][INFO] - Step 4900, Loss: {'train/loss': 0.14107556641101837, 'train/grad_norm_pre_clip': 19.781665802001953, 'eval/loss': 0.12699159979820251, 'eval/latency_0/loss': 0.08099435269832611, 'eval/latency_1/loss': 0.047102369368076324, 'eval/latency_2/loss': 0.11346220225095749, 'eval/latency_3/loss': 0.15542304515838623, 'eval/latency_4/loss': 0.2391013354063034, 'eval/flappy/loss': 0.12699159979820251, 'eval/flappy/latency_0/loss': 0.08099435269832611, 'eval/flappy/latency_1/loss': 0.047102369368076324, 'eval/flappy/latency_2/loss': 0.11346220225095749, 'eval/flappy/latency_3/loss': 0.15542304515838623, 'eval/flappy/latency_4/loss': 0.2391013354063034, 'eval/flap_precision': 0.7216117216117216, 'eval/flap_recall': 0.7163636363636363, 'eval/flap_f1': 0.718978102189781, 'eval/flap_rate_model': 0.07583333333333334, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.718978102189781, 'eval/latency_0/flap_precision': 0.7216117216117216, 'eval/latency_0/flap_recall': 0.7163636363636363, 'eval/latency_0/flap_f1': 0.718978102189781, 'eval/latency_0/flap_rate_model': 0.07583333333333334, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.718978102189781, 'timing/data': 0.019040224025957286, 'timing/model': 1.3186401210259646, 'learning_rate/qwen_vl_interface': 5.097593844359024e-07, 'learning_rate/action_model': 5.0975938443590245e-06, 'epoch': 1.84}) +[2026-06-19 21:03:28,874][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_4900_state +[2026-06-19 21:03:28,877][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-19 21:03:32,264][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_4900_state/model.safetensors +[2026-06-19 21:03:39,785][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_4900_state/optimizer.bin +[2026-06-19 21:03:39,786][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_4900_state/sampler.bin +[2026-06-19 21:03:39,786][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_4900_state/sampler_1.bin +[2026-06-19 21:03:39,787][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_4900_state/random_states_0.pkl +[2026-06-19 21:03:39,810][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4900 +[2026-06-19 21:03:39,810][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-19 21:03:39,840][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-19 21:03:44,304][starVLA.training.train_starvla][INFO] - Step 4901, Loss: {'train/loss': 0.1378696709871292, 'train/grad_norm_pre_clip': 17.09958839416504, 'timing/data': 0.01789720298256725, 'timing/model': 1.124515444971621, 'learning_rate/qwen_vl_interface': 5.095652378939718e-07, 'learning_rate/action_model': 5.095652378939717e-06, 'epoch': 1.84}) +[2026-06-19 21:03:48,837][starVLA.training.train_starvla][INFO] - Step 4902, Loss: {'train/loss': 0.2638920843601227, 'train/grad_norm_pre_clip': 15.670701026916504, 'timing/data': 0.017051036935299635, 'timing/model': 1.1390883080894127, 'learning_rate/qwen_vl_interface': 5.093730399657101e-07, 'learning_rate/action_model': 5.0937303996571e-06, 'epoch': 1.84}) +[2026-06-19 21:03:53,361][starVLA.training.train_starvla][INFO] - Step 4903, Loss: {'train/loss': 0.34739670157432556, 'train/grad_norm_pre_clip': 18.31582260131836, 'timing/data': 0.017701214994303882, 'timing/model': 1.2971245370572433, 'learning_rate/qwen_vl_interface': 5.091827907301227e-07, 'learning_rate/action_model': 5.091827907301227e-06, 'epoch': 1.84}) +[2026-06-19 21:03:57,726][starVLA.training.train_starvla][INFO] - Step 4904, Loss: {'train/loss': 0.17167887091636658, 'train/grad_norm_pre_clip': 17.089397430419922, 'timing/data': 0.018545996979810297, 'timing/model': 1.1284225219860673, 'learning_rate/qwen_vl_interface': 5.089944902654134e-07, 'learning_rate/action_model': 5.089944902654134e-06, 'epoch': 1.84}) +[2026-06-19 21:04:02,243][starVLA.training.train_starvla][INFO] - Step 4905, Loss: {'train/loss': 0.18126586079597473, 'train/grad_norm_pre_clip': 13.88349723815918, 'timing/data': 0.0192023889394477, 'timing/model': 1.1300775710260496, 'learning_rate/qwen_vl_interface': 5.088081386489856e-07, 'learning_rate/action_model': 5.088081386489857e-06, 'epoch': 1.84}) +[2026-06-19 21:04:06,771][starVLA.training.train_starvla][INFO] - Step 4906, Loss: {'train/loss': 0.12495169043540955, 'train/grad_norm_pre_clip': 10.116466522216797, 'timing/data': 0.018950286088511348, 'timing/model': 1.138523559900932, 'learning_rate/qwen_vl_interface': 5.086237359574419e-07, 'learning_rate/action_model': 5.086237359574419e-06, 'epoch': 1.84}) +[2026-06-19 21:04:11,325][starVLA.training.train_starvla][INFO] - Step 4907, Loss: {'train/loss': 0.14956887066364288, 'train/grad_norm_pre_clip': 14.431385040283203, 'timing/data': 0.018893323955126107, 'timing/model': 1.1379277569940314, 'learning_rate/qwen_vl_interface': 5.084412822665825e-07, 'learning_rate/action_model': 5.084412822665825e-06, 'epoch': 1.84}) +[2026-06-19 21:04:15,882][starVLA.training.train_starvla][INFO] - Step 4908, Loss: {'train/loss': 0.1949121505022049, 'train/grad_norm_pre_clip': 15.012730598449707, 'timing/data': 0.017994508030824363, 'timing/model': 1.1358586830319837, 'learning_rate/qwen_vl_interface': 5.082607776514081e-07, 'learning_rate/action_model': 5.082607776514081e-06, 'epoch': 1.84}) +[2026-06-19 21:04:20,437][starVLA.training.train_starvla][INFO] - Step 4909, Loss: {'train/loss': 0.1702539622783661, 'train/grad_norm_pre_clip': 16.868946075439453, 'timing/data': 0.015977807925082743, 'timing/model': 1.3020788290305063, 'learning_rate/qwen_vl_interface': 5.080822221861167e-07, 'learning_rate/action_model': 5.080822221861166e-06, 'epoch': 1.84}) +[2026-06-19 21:04:24,945][starVLA.training.train_starvla][INFO] - Step 4910, Loss: {'train/loss': 0.23888497054576874, 'train/grad_norm_pre_clip': 19.624801635742188, 'timing/data': 0.016255680937319994, 'timing/model': 1.2584608459146693, 'learning_rate/qwen_vl_interface': 5.07905615944106e-07, 'learning_rate/action_model': 5.0790561594410595e-06, 'epoch': 1.84}) +[2026-06-19 21:04:29,343][starVLA.training.train_starvla][INFO] - Step 4911, Loss: {'train/loss': 0.07845891267061234, 'train/grad_norm_pre_clip': 9.003445625305176, 'timing/data': 0.016813850030303, 'timing/model': 1.137966020964086, 'learning_rate/qwen_vl_interface': 5.077309589979716e-07, 'learning_rate/action_model': 5.077309589979715e-06, 'epoch': 1.84}) +[2026-06-19 21:04:33,935][starVLA.training.train_starvla][INFO] - Step 4912, Loss: {'train/loss': 0.20079807937145233, 'train/grad_norm_pre_clip': 24.359909057617188, 'timing/data': 0.01600928499829024, 'timing/model': 1.1514546379912645, 'learning_rate/qwen_vl_interface': 5.075582514195089e-07, 'learning_rate/action_model': 5.0755825141950895e-06, 'epoch': 1.84}) +[2026-06-19 21:04:38,546][starVLA.training.train_starvla][INFO] - Step 4913, Loss: {'train/loss': 0.2047107070684433, 'train/grad_norm_pre_clip': 12.099120140075684, 'timing/data': 0.017797125969082117, 'timing/model': 1.1464409589534625, 'learning_rate/qwen_vl_interface': 5.073874932797115e-07, 'learning_rate/action_model': 5.0738749327971155e-06, 'epoch': 1.84}) +[2026-06-19 21:04:43,143][starVLA.training.train_starvla][INFO] - Step 4914, Loss: {'train/loss': 0.20798715949058533, 'train/grad_norm_pre_clip': 14.08934211730957, 'timing/data': 0.018266086000949144, 'timing/model': 1.1464129050727934, 'learning_rate/qwen_vl_interface': 5.072186846487713e-07, 'learning_rate/action_model': 5.072186846487713e-06, 'epoch': 1.84}) +[2026-06-19 21:04:47,732][starVLA.training.train_starvla][INFO] - Step 4915, Loss: {'train/loss': 0.04285592958331108, 'train/grad_norm_pre_clip': 5.0823540687561035, 'timing/data': 0.01798054901883006, 'timing/model': 1.319883243064396, 'learning_rate/qwen_vl_interface': 5.070518255960796e-07, 'learning_rate/action_model': 5.070518255960795e-06, 'epoch': 1.84}) +[2026-06-19 21:04:52,247][starVLA.training.train_starvla][INFO] - Step 4916, Loss: {'train/loss': 0.30242085456848145, 'train/grad_norm_pre_clip': 19.46874237060547, 'timing/data': 0.017982607008889318, 'timing/model': 1.2357701590517536, 'learning_rate/qwen_vl_interface': 5.068869161902256e-07, 'learning_rate/action_model': 5.068869161902256e-06, 'epoch': 1.84}) +[2026-06-19 21:04:56,666][starVLA.training.train_starvla][INFO] - Step 4917, Loss: {'train/loss': 0.25236019492149353, 'train/grad_norm_pre_clip': 24.40330696105957, 'timing/data': 0.022290633991360664, 'timing/model': 1.1467137429863214, 'learning_rate/qwen_vl_interface': 5.067239564989968e-07, 'learning_rate/action_model': 5.067239564989968e-06, 'epoch': 1.84}) +[2026-06-19 21:05:01,245][starVLA.training.train_starvla][INFO] - Step 4918, Loss: {'train/loss': 0.24184849858283997, 'train/grad_norm_pre_clip': 11.279180526733398, 'timing/data': 0.019737533992156386, 'timing/model': 1.1573176749516279, 'learning_rate/qwen_vl_interface': 5.065629465893808e-07, 'learning_rate/action_model': 5.065629465893807e-06, 'epoch': 1.84}) +[2026-06-19 21:05:05,866][starVLA.training.train_starvla][INFO] - Step 4919, Loss: {'train/loss': 0.2254611700773239, 'train/grad_norm_pre_clip': 45.9481315612793, 'timing/data': 0.022196665056981146, 'timing/model': 1.1620668040122837, 'learning_rate/qwen_vl_interface': 5.064038865275625e-07, 'learning_rate/action_model': 5.0640388652756245e-06, 'epoch': 1.84}) +[2026-06-19 21:05:10,459][starVLA.training.train_starvla][INFO] - Step 4920, Loss: {'train/loss': 0.2623347342014313, 'train/grad_norm_pre_clip': 25.58147621154785, 'timing/data': 0.017832187935709953, 'timing/model': 1.1461052800295874, 'learning_rate/qwen_vl_interface': 5.062467763789242e-07, 'learning_rate/action_model': 5.062467763789242e-06, 'epoch': 1.84}) +[2026-06-19 21:05:15,029][starVLA.training.train_starvla][INFO] - Step 4921, Loss: {'train/loss': 0.18703223764896393, 'train/grad_norm_pre_clip': 19.396574020385742, 'timing/data': 0.017860893975012004, 'timing/model': 1.3080278229899704, 'learning_rate/qwen_vl_interface': 5.060916162080496e-07, 'learning_rate/action_model': 5.060916162080496e-06, 'epoch': 1.85}) +[2026-06-19 21:05:19,448][starVLA.training.train_starvla][INFO] - Step 4922, Loss: {'train/loss': 0.09917428344488144, 'train/grad_norm_pre_clip': 12.499076843261719, 'timing/data': 0.0179328239755705, 'timing/model': 1.1490390630206093, 'learning_rate/qwen_vl_interface': 5.059384060787182e-07, 'learning_rate/action_model': 5.059384060787183e-06, 'epoch': 1.85}) +[2026-06-19 21:05:24,029][starVLA.training.train_starvla][INFO] - Step 4923, Loss: {'train/loss': 0.2578219175338745, 'train/grad_norm_pre_clip': 12.900177001953125, 'timing/data': 0.01740338304080069, 'timing/model': 1.1419263910502195, 'learning_rate/qwen_vl_interface': 5.05787146053909e-07, 'learning_rate/action_model': 5.0578714605390904e-06, 'epoch': 1.85}) +[2026-06-19 21:05:28,531][starVLA.training.train_starvla][INFO] - Step 4924, Loss: {'train/loss': 0.4094894826412201, 'train/grad_norm_pre_clip': 19.430612564086914, 'timing/data': 0.017574994009919465, 'timing/model': 1.1435484939720482, 'learning_rate/qwen_vl_interface': 5.056378361957998e-07, 'learning_rate/action_model': 5.056378361957998e-06, 'epoch': 1.85}) +[2026-06-19 21:05:33,025][starVLA.training.train_starvla][INFO] - Step 4925, Loss: {'train/loss': 0.13747401535511017, 'train/grad_norm_pre_clip': 11.620129585266113, 'timing/data': 0.018214648007415235, 'timing/model': 1.1441812999546528, 'learning_rate/qwen_vl_interface': 5.05490476565766e-07, 'learning_rate/action_model': 5.054904765657659e-06, 'epoch': 1.85}) +[2026-06-19 21:05:37,558][starVLA.training.train_starvla][INFO] - Step 4926, Loss: {'train/loss': 0.09082327783107758, 'train/grad_norm_pre_clip': 16.95698356628418, 'timing/data': 0.018189493915997446, 'timing/model': 1.1482220840407535, 'learning_rate/qwen_vl_interface': 5.053450672243811e-07, 'learning_rate/action_model': 5.0534506722438114e-06, 'epoch': 1.85}) +[2026-06-19 21:05:42,149][starVLA.training.train_starvla][INFO] - Step 4927, Loss: {'train/loss': 0.23049050569534302, 'train/grad_norm_pre_clip': 28.872390747070312, 'timing/data': 0.018394752987660468, 'timing/model': 1.3189776149811223, 'learning_rate/qwen_vl_interface': 5.052016082314184e-07, 'learning_rate/action_model': 5.052016082314184e-06, 'epoch': 1.85}) +[2026-06-19 21:05:46,593][starVLA.training.train_starvla][INFO] - Step 4928, Loss: {'train/loss': 0.0824943482875824, 'train/grad_norm_pre_clip': 10.654383659362793, 'timing/data': 0.018110907054506242, 'timing/model': 1.1486587659455836, 'learning_rate/qwen_vl_interface': 5.050600996458473e-07, 'learning_rate/action_model': 5.050600996458473e-06, 'epoch': 1.85}) +[2026-06-19 21:05:51,173][starVLA.training.train_starvla][INFO] - Step 4929, Loss: {'train/loss': 0.1930205225944519, 'train/grad_norm_pre_clip': 10.063904762268066, 'timing/data': 0.018114587990567088, 'timing/model': 1.1454313909634948, 'learning_rate/qwen_vl_interface': 5.049205415258373e-07, 'learning_rate/action_model': 5.049205415258373e-06, 'epoch': 1.85}) +[2026-06-19 21:05:55,784][starVLA.training.train_starvla][INFO] - Step 4930, Loss: {'train/loss': 0.2318461686372757, 'train/grad_norm_pre_clip': 9.387422561645508, 'timing/data': 0.017591863055713475, 'timing/model': 1.1484981519170105, 'learning_rate/qwen_vl_interface': 5.047829339287557e-07, 'learning_rate/action_model': 5.047829339287557e-06, 'epoch': 1.85}) +[2026-06-19 21:06:00,364][starVLA.training.train_starvla][INFO] - Step 4931, Loss: {'train/loss': 0.14988406002521515, 'train/grad_norm_pre_clip': 9.20035457611084, 'timing/data': 0.016055856947787106, 'timing/model': 1.1434132040012628, 'learning_rate/qwen_vl_interface': 5.046472769111668e-07, 'learning_rate/action_model': 5.046472769111667e-06, 'epoch': 1.85}) +[2026-06-19 21:06:04,965][starVLA.training.train_starvla][INFO] - Step 4932, Loss: {'train/loss': 0.1867690086364746, 'train/grad_norm_pre_clip': 16.63341522216797, 'timing/data': 0.01675696996971965, 'timing/model': 1.3109495480312034, 'learning_rate/qwen_vl_interface': 5.045135705288353e-07, 'learning_rate/action_model': 5.0451357052883525e-06, 'epoch': 1.85}) +[2026-06-19 21:06:09,372][starVLA.training.train_starvla][INFO] - Step 4933, Loss: {'train/loss': 0.24721170961856842, 'train/grad_norm_pre_clip': 16.552913665771484, 'timing/data': 0.016594737069681287, 'timing/model': 1.144510563928634, 'learning_rate/qwen_vl_interface': 5.043818148367221e-07, 'learning_rate/action_model': 5.043818148367221e-06, 'epoch': 1.85}) +[2026-06-19 21:06:13,957][starVLA.training.train_starvla][INFO] - Step 4934, Loss: {'train/loss': 0.19681698083877563, 'train/grad_norm_pre_clip': 15.002717971801758, 'timing/data': 0.01773544494062662, 'timing/model': 1.1485437939409167, 'learning_rate/qwen_vl_interface': 5.042520098889869e-07, 'learning_rate/action_model': 5.0425200988898685e-06, 'epoch': 1.85}) +[2026-06-19 21:06:18,533][starVLA.training.train_starvla][INFO] - Step 4935, Loss: {'train/loss': 0.23832544684410095, 'train/grad_norm_pre_clip': 23.347183227539062, 'timing/data': 0.01764217298477888, 'timing/model': 1.1466072850162163, 'learning_rate/qwen_vl_interface': 5.041241557389879e-07, 'learning_rate/action_model': 5.041241557389879e-06, 'epoch': 1.85}) +[2026-06-19 21:06:23,145][starVLA.training.train_starvla][INFO] - Step 4936, Loss: {'train/loss': 0.0719757080078125, 'train/grad_norm_pre_clip': 7.969968318939209, 'timing/data': 0.017166705918498337, 'timing/model': 1.1584614340681583, 'learning_rate/qwen_vl_interface': 5.039982524392814e-07, 'learning_rate/action_model': 5.039982524392814e-06, 'epoch': 1.85}) +[2026-06-19 21:06:27,735][starVLA.training.train_starvla][INFO] - Step 4937, Loss: {'train/loss': 0.32442012429237366, 'train/grad_norm_pre_clip': 34.76947784423828, 'timing/data': 0.017827136092819273, 'timing/model': 1.1599279959918931, 'learning_rate/qwen_vl_interface': 5.038743000416207e-07, 'learning_rate/action_model': 5.0387430004162064e-06, 'epoch': 1.85}) +[2026-06-19 21:06:32,332][starVLA.training.train_starvla][INFO] - Step 4938, Loss: {'train/loss': 0.23436923325061798, 'train/grad_norm_pre_clip': 20.037246704101562, 'timing/data': 0.018097370979376137, 'timing/model': 1.3236516630277038, 'learning_rate/qwen_vl_interface': 5.037522985969588e-07, 'learning_rate/action_model': 5.037522985969587e-06, 'epoch': 1.85}) +[2026-06-19 21:06:36,921][starVLA.training.train_starvla][INFO] - Step 4939, Loss: {'train/loss': 0.13516178727149963, 'train/grad_norm_pre_clip': 10.047167778015137, 'timing/data': 0.017853218945674598, 'timing/model': 1.3138853770215064, 'learning_rate/qwen_vl_interface': 5.036322481554451e-07, 'learning_rate/action_model': 5.036322481554451e-06, 'epoch': 1.85}) +[2026-06-19 21:06:41,352][starVLA.training.train_starvla][INFO] - Step 4940, Loss: {'train/loss': 0.3382946252822876, 'train/grad_norm_pre_clip': 14.352545738220215, 'timing/data': 0.0179639799753204, 'timing/model': 1.1494202059693635, 'learning_rate/qwen_vl_interface': 5.035141487664285e-07, 'learning_rate/action_model': 5.035141487664285e-06, 'epoch': 1.85}) +[2026-06-19 21:06:45,945][starVLA.training.train_starvla][INFO] - Step 4941, Loss: {'train/loss': 0.4020456075668335, 'train/grad_norm_pre_clip': 18.53411102294922, 'timing/data': 0.017760068061761558, 'timing/model': 1.1527266299817711, 'learning_rate/qwen_vl_interface': 5.03398000478455e-07, 'learning_rate/action_model': 5.03398000478455e-06, 'epoch': 1.85}) +[2026-06-19 21:06:50,549][starVLA.training.train_starvla][INFO] - Step 4942, Loss: {'train/loss': 0.17659418284893036, 'train/grad_norm_pre_clip': 10.543144226074219, 'timing/data': 0.01747515401802957, 'timing/model': 1.1544846160104498, 'learning_rate/qwen_vl_interface': 5.032838033392686e-07, 'learning_rate/action_model': 5.032838033392685e-06, 'epoch': 1.85}) +[2026-06-19 21:06:55,042][starVLA.training.train_starvla][INFO] - Step 4943, Loss: {'train/loss': 0.2030653953552246, 'train/grad_norm_pre_clip': 17.446718215942383, 'timing/data': 0.018382713082246482, 'timing/model': 1.1496976730413735, 'learning_rate/qwen_vl_interface': 5.031715573958111e-07, 'learning_rate/action_model': 5.031715573958111e-06, 'epoch': 1.85}) +[2026-06-19 21:06:59,613][starVLA.training.train_starvla][INFO] - Step 4944, Loss: {'train/loss': 0.3086978793144226, 'train/grad_norm_pre_clip': 20.934110641479492, 'timing/data': 0.017859108978882432, 'timing/model': 1.1463846750557423, 'learning_rate/qwen_vl_interface': 5.030612626942235e-07, 'learning_rate/action_model': 5.030612626942235e-06, 'epoch': 1.85}) +[2026-06-19 21:07:04,215][starVLA.training.train_starvla][INFO] - Step 4945, Loss: {'train/loss': 0.39263659715652466, 'train/grad_norm_pre_clip': 31.403099060058594, 'timing/data': 0.017472959007136524, 'timing/model': 1.3181345600169152, 'learning_rate/qwen_vl_interface': 5.029529192798433e-07, 'learning_rate/action_model': 5.029529192798433e-06, 'epoch': 1.85}) +[2026-06-19 21:07:08,656][starVLA.training.train_starvla][INFO] - Step 4946, Loss: {'train/loss': 0.2389877438545227, 'train/grad_norm_pre_clip': 18.5737247467041, 'timing/data': 0.016362341004423797, 'timing/model': 1.1592178780119866, 'learning_rate/qwen_vl_interface': 5.028465271972058e-07, 'learning_rate/action_model': 5.028465271972058e-06, 'epoch': 1.85}) +[2026-06-19 21:07:13,238][starVLA.training.train_starvla][INFO] - Step 4947, Loss: {'train/loss': 0.42397135496139526, 'train/grad_norm_pre_clip': 32.3898811340332, 'timing/data': 0.017138910945504904, 'timing/model': 1.151701228111051, 'learning_rate/qwen_vl_interface': 5.027420864900457e-07, 'learning_rate/action_model': 5.027420864900456e-06, 'epoch': 1.85}) +[2026-06-19 21:07:17,863][starVLA.training.train_starvla][INFO] - Step 4948, Loss: {'train/loss': 0.31280940771102905, 'train/grad_norm_pre_clip': 20.114255905151367, 'timing/data': 0.01574686006642878, 'timing/model': 1.1614784490084276, 'learning_rate/qwen_vl_interface': 5.026395972012943e-07, 'learning_rate/action_model': 5.0263959720129436e-06, 'epoch': 1.86}) +[2026-06-19 21:07:22,478][starVLA.training.train_starvla][INFO] - Step 4949, Loss: {'train/loss': 0.17638899385929108, 'train/grad_norm_pre_clip': 19.73284912109375, 'timing/data': 0.01710275001823902, 'timing/model': 1.1609581550583243, 'learning_rate/qwen_vl_interface': 5.025390593730809e-07, 'learning_rate/action_model': 5.025390593730809e-06, 'epoch': 1.86}) +[2026-06-19 21:07:27,001][starVLA.training.train_starvla][INFO] - Step 4950, Loss: {'train/loss': 0.19136105477809906, 'train/grad_norm_pre_clip': 11.915974617004395, 'timing/data': 0.018333886982873082, 'timing/model': 1.1484814169816673, 'learning_rate/qwen_vl_interface': 5.024404730467325e-07, 'learning_rate/action_model': 5.024404730467325e-06, 'epoch': 1.86}) +[2026-06-19 21:07:31,655][starVLA.training.train_starvla][INFO] - Step 4951, Loss: {'train/loss': 0.11944439262151718, 'train/grad_norm_pre_clip': 8.62289810180664, 'timing/data': 0.017204446019604802, 'timing/model': 1.2955557469977066, 'learning_rate/qwen_vl_interface': 5.023438382627749e-07, 'learning_rate/action_model': 5.0234383826277486e-06, 'epoch': 1.86}) +[2026-06-19 21:07:36,088][starVLA.training.train_starvla][INFO] - Step 4952, Loss: {'train/loss': 0.3242640793323517, 'train/grad_norm_pre_clip': 32.565975189208984, 'timing/data': 0.019881571060977876, 'timing/model': 1.1486669270088896, 'learning_rate/qwen_vl_interface': 5.022491550609311e-07, 'learning_rate/action_model': 5.022491550609311e-06, 'epoch': 1.86}) +[2026-06-19 21:07:40,733][starVLA.training.train_starvla][INFO] - Step 4953, Loss: {'train/loss': 0.11101385205984116, 'train/grad_norm_pre_clip': 11.24067211151123, 'timing/data': 0.01624071307014674, 'timing/model': 1.1543766530230641, 'learning_rate/qwen_vl_interface': 5.021564234801213e-07, 'learning_rate/action_model': 5.021564234801213e-06, 'epoch': 1.86}) +[2026-06-19 21:07:45,341][starVLA.training.train_starvla][INFO] - Step 4954, Loss: {'train/loss': 0.16195155680179596, 'train/grad_norm_pre_clip': 14.053423881530762, 'timing/data': 0.017159480950795114, 'timing/model': 1.144461845047772, 'learning_rate/qwen_vl_interface': 5.020656435584636e-07, 'learning_rate/action_model': 5.020656435584635e-06, 'epoch': 1.86}) +[2026-06-19 21:07:49,954][starVLA.training.train_starvla][INFO] - Step 4955, Loss: {'train/loss': 0.2686580419540405, 'train/grad_norm_pre_clip': 13.51729679107666, 'timing/data': 0.01673611206933856, 'timing/model': 1.1468731029890478, 'learning_rate/qwen_vl_interface': 5.019768153332752e-07, 'learning_rate/action_model': 5.019768153332752e-06, 'epoch': 1.86}) +[2026-06-19 21:07:54,579][starVLA.training.train_starvla][INFO] - Step 4956, Loss: {'train/loss': 0.3302392065525055, 'train/grad_norm_pre_clip': 23.875408172607422, 'timing/data': 0.01675384200643748, 'timing/model': 1.3350333240814507, 'learning_rate/qwen_vl_interface': 5.018899388410693e-07, 'learning_rate/action_model': 5.0188993884106935e-06, 'epoch': 1.86}) +[2026-06-19 21:07:59,099][starVLA.training.train_starvla][INFO] - Step 4957, Loss: {'train/loss': 0.11376087367534637, 'train/grad_norm_pre_clip': 12.222844123840332, 'timing/data': 0.01721141196321696, 'timing/model': 1.2274283199803904, 'learning_rate/qwen_vl_interface': 5.018050141175576e-07, 'learning_rate/action_model': 5.018050141175575e-06, 'epoch': 1.86}) +[2026-06-19 21:08:03,545][starVLA.training.train_starvla][INFO] - Step 4958, Loss: {'train/loss': 0.05552123114466667, 'train/grad_norm_pre_clip': 6.84769344329834, 'timing/data': 0.015941611025482416, 'timing/model': 1.1568051929352805, 'learning_rate/qwen_vl_interface': 5.017220411976496e-07, 'learning_rate/action_model': 5.0172204119764965e-06, 'epoch': 1.86}) +[2026-06-19 21:08:08,094][starVLA.training.train_starvla][INFO] - Step 4959, Loss: {'train/loss': 0.2613823115825653, 'train/grad_norm_pre_clip': 18.901105880737305, 'timing/data': 0.0169381300220266, 'timing/model': 1.1479759600479156, 'learning_rate/qwen_vl_interface': 5.016410201154523e-07, 'learning_rate/action_model': 5.016410201154523e-06, 'epoch': 1.86}) +[2026-06-19 21:08:12,689][starVLA.training.train_starvla][INFO] - Step 4960, Loss: {'train/loss': 0.16266444325447083, 'train/grad_norm_pre_clip': 14.125540733337402, 'timing/data': 0.015687534934841096, 'timing/model': 1.1530101620592177, 'learning_rate/qwen_vl_interface': 5.015619509042709e-07, 'learning_rate/action_model': 5.015619509042709e-06, 'epoch': 1.86}) +[2026-06-19 21:08:17,287][starVLA.training.train_starvla][INFO] - Step 4961, Loss: {'train/loss': 0.05702732130885124, 'train/grad_norm_pre_clip': 10.621880531311035, 'timing/data': 0.01692512701265514, 'timing/model': 1.1514938700711355, 'learning_rate/qwen_vl_interface': 5.014848335966067e-07, 'learning_rate/action_model': 5.014848335966067e-06, 'epoch': 1.86}) +[2026-06-19 21:08:21,873][starVLA.training.train_starvla][INFO] - Step 4962, Loss: {'train/loss': 0.11632712185382843, 'train/grad_norm_pre_clip': 9.64327335357666, 'timing/data': 0.01596759504172951, 'timing/model': 1.3073769690236077, 'learning_rate/qwen_vl_interface': 5.014096682241608e-07, 'learning_rate/action_model': 5.014096682241608e-06, 'epoch': 1.86}) +[2026-06-19 21:08:26,395][starVLA.training.train_starvla][INFO] - Step 4963, Loss: {'train/loss': 0.17411287128925323, 'train/grad_norm_pre_clip': 16.34644317626953, 'timing/data': 0.01842820004094392, 'timing/model': 1.221532488008961, 'learning_rate/qwen_vl_interface': 5.013364548178299e-07, 'learning_rate/action_model': 5.013364548178299e-06, 'epoch': 1.86}) +[2026-06-19 21:08:30,820][starVLA.training.train_starvla][INFO] - Step 4964, Loss: {'train/loss': 0.16174577176570892, 'train/grad_norm_pre_clip': 17.98016357421875, 'timing/data': 0.017319557955488563, 'timing/model': 1.1483598239719868, 'learning_rate/qwen_vl_interface': 5.012651934077101e-07, 'learning_rate/action_model': 5.0126519340771e-06, 'epoch': 1.86}) +[2026-06-19 21:08:35,380][starVLA.training.train_starvla][INFO] - Step 4965, Loss: {'train/loss': 0.29862111806869507, 'train/grad_norm_pre_clip': 23.975671768188477, 'timing/data': 0.01728462101891637, 'timing/model': 1.148774049943313, 'learning_rate/qwen_vl_interface': 5.01195884023094e-07, 'learning_rate/action_model': 5.0119588402309395e-06, 'epoch': 1.86}) +[2026-06-19 21:08:39,994][starVLA.training.train_starvla][INFO] - Step 4966, Loss: {'train/loss': 0.16190017759799957, 'train/grad_norm_pre_clip': 13.925931930541992, 'timing/data': 0.01695474295411259, 'timing/model': 1.1500663600163534, 'learning_rate/qwen_vl_interface': 5.011285266924714e-07, 'learning_rate/action_model': 5.011285266924714e-06, 'epoch': 1.86}) +[2026-06-19 21:08:44,604][starVLA.training.train_starvla][INFO] - Step 4967, Loss: {'train/loss': 0.19195769727230072, 'train/grad_norm_pre_clip': 24.678552627563477, 'timing/data': 0.016597220906987786, 'timing/model': 1.1531310511054471, 'learning_rate/qwen_vl_interface': 5.010631214435312e-07, 'learning_rate/action_model': 5.010631214435312e-06, 'epoch': 1.86}) +[2026-06-19 21:08:49,210][starVLA.training.train_starvla][INFO] - Step 4968, Loss: {'train/loss': 0.043529752641916275, 'train/grad_norm_pre_clip': 5.893505096435547, 'timing/data': 0.016463648062199354, 'timing/model': 1.3088925780029967, 'learning_rate/qwen_vl_interface': 5.009996683031586e-07, 'learning_rate/action_model': 5.009996683031586e-06, 'epoch': 1.86}) +[2026-06-19 21:08:53,710][starVLA.training.train_starvla][INFO] - Step 4969, Loss: {'train/loss': 0.25592902302742004, 'train/grad_norm_pre_clip': 15.316372871398926, 'timing/data': 0.017089233035221696, 'timing/model': 1.219343462958932, 'learning_rate/qwen_vl_interface': 5.009381672974375e-07, 'learning_rate/action_model': 5.009381672974375e-06, 'epoch': 1.86}) +[2026-06-19 21:08:58,127][starVLA.training.train_starvla][INFO] - Step 4970, Loss: {'train/loss': 0.08866088837385178, 'train/grad_norm_pre_clip': 8.046687126159668, 'timing/data': 0.017206485965289176, 'timing/model': 1.146734384004958, 'learning_rate/qwen_vl_interface': 5.008786184516475e-07, 'learning_rate/action_model': 5.008786184516474e-06, 'epoch': 1.86}) +[2026-06-19 21:09:02,676][starVLA.training.train_starvla][INFO] - Step 4971, Loss: {'train/loss': 0.16597144305706024, 'train/grad_norm_pre_clip': 22.212953567504883, 'timing/data': 0.0168504579924047, 'timing/model': 1.1445536710089073, 'learning_rate/qwen_vl_interface': 5.008210217902676e-07, 'learning_rate/action_model': 5.008210217902676e-06, 'epoch': 1.86}) +[2026-06-19 21:09:07,276][starVLA.training.train_starvla][INFO] - Step 4972, Loss: {'train/loss': 0.13673323392868042, 'train/grad_norm_pre_clip': 13.739386558532715, 'timing/data': 0.01763566303998232, 'timing/model': 1.1525402049301192, 'learning_rate/qwen_vl_interface': 5.007653773369737e-07, 'learning_rate/action_model': 5.007653773369736e-06, 'epoch': 1.86}) +[2026-06-19 21:09:11,879][starVLA.training.train_starvla][INFO] - Step 4973, Loss: {'train/loss': 0.14286765456199646, 'train/grad_norm_pre_clip': 25.081775665283203, 'timing/data': 0.017674450995400548, 'timing/model': 1.147819940932095, 'learning_rate/qwen_vl_interface': 5.007116851146386e-07, 'learning_rate/action_model': 5.007116851146386e-06, 'epoch': 1.86}) +[2026-06-19 21:09:16,490][starVLA.training.train_starvla][INFO] - Step 4974, Loss: {'train/loss': 0.1925821602344513, 'train/grad_norm_pre_clip': 13.999175071716309, 'timing/data': 0.01811465702485293, 'timing/model': 1.32359489495866, 'learning_rate/qwen_vl_interface': 5.006599451453334e-07, 'learning_rate/action_model': 5.006599451453334e-06, 'epoch': 1.87}) +[2026-06-19 21:09:20,997][starVLA.training.train_starvla][INFO] - Step 4975, Loss: {'train/loss': 0.3927392363548279, 'train/grad_norm_pre_clip': 28.268047332763672, 'timing/data': 0.017126424005255103, 'timing/model': 1.2142451079562306, 'learning_rate/qwen_vl_interface': 5.006101574503268e-07, 'learning_rate/action_model': 5.006101574503267e-06, 'epoch': 1.87}) +[2026-06-19 21:09:25,426][starVLA.training.train_starvla][INFO] - Step 4976, Loss: {'train/loss': 0.23198047280311584, 'train/grad_norm_pre_clip': 9.448660850524902, 'timing/data': 0.016962729045189917, 'timing/model': 1.1498389720218256, 'learning_rate/qwen_vl_interface': 5.005623220500839e-07, 'learning_rate/action_model': 5.005623220500839e-06, 'epoch': 1.87}) +[2026-06-19 21:09:30,020][starVLA.training.train_starvla][INFO] - Step 4977, Loss: {'train/loss': 0.15234379470348358, 'train/grad_norm_pre_clip': 14.99095344543457, 'timing/data': 0.01727745297830552, 'timing/model': 1.152046275907196, 'learning_rate/qwen_vl_interface': 5.005164389642687e-07, 'learning_rate/action_model': 5.005164389642687e-06, 'epoch': 1.87}) +[2026-06-19 21:09:34,615][starVLA.training.train_starvla][INFO] - Step 4978, Loss: {'train/loss': 0.42563900351524353, 'train/grad_norm_pre_clip': 28.785207748413086, 'timing/data': 0.017445974051952362, 'timing/model': 1.1516857490641996, 'learning_rate/qwen_vl_interface': 5.004725082117423e-07, 'learning_rate/action_model': 5.004725082117422e-06, 'epoch': 1.87}) +[2026-06-19 21:09:39,219][starVLA.training.train_starvla][INFO] - Step 4979, Loss: {'train/loss': 0.2598297894001007, 'train/grad_norm_pre_clip': 17.761920928955078, 'timing/data': 0.01893318397924304, 'timing/model': 1.1594814429990947, 'learning_rate/qwen_vl_interface': 5.004305298105616e-07, 'learning_rate/action_model': 5.004305298105616e-06, 'epoch': 1.87}) +[2026-06-19 21:09:43,811][starVLA.training.train_starvla][INFO] - Step 4980, Loss: {'train/loss': 0.26450347900390625, 'train/grad_norm_pre_clip': 13.386337280273438, 'timing/data': 0.01752834499347955, 'timing/model': 1.3140363149577752, 'learning_rate/qwen_vl_interface': 5.003905037779834e-07, 'learning_rate/action_model': 5.0039050377798345e-06, 'epoch': 1.87}) +[2026-06-19 21:09:48,305][starVLA.training.train_starvla][INFO] - Step 4981, Loss: {'train/loss': 0.3602021634578705, 'train/grad_norm_pre_clip': 38.66647720336914, 'timing/data': 0.016938387998379767, 'timing/model': 1.2159085730090737, 'learning_rate/qwen_vl_interface': 5.003524301304608e-07, 'learning_rate/action_model': 5.003524301304608e-06, 'epoch': 1.87}) +[2026-06-19 21:09:52,734][starVLA.training.train_starvla][INFO] - Step 4982, Loss: {'train/loss': 0.1277254819869995, 'train/grad_norm_pre_clip': 19.888229370117188, 'timing/data': 0.017883616033941507, 'timing/model': 1.152290905942209, 'learning_rate/qwen_vl_interface': 5.003163088836445e-07, 'learning_rate/action_model': 5.003163088836446e-06, 'epoch': 1.87}) +[2026-06-19 21:09:57,295][starVLA.training.train_starvla][INFO] - Step 4983, Loss: {'train/loss': 0.05416005104780197, 'train/grad_norm_pre_clip': 5.865408897399902, 'timing/data': 0.01799689303152263, 'timing/model': 1.146383623010479, 'learning_rate/qwen_vl_interface': 5.002821400523818e-07, 'learning_rate/action_model': 5.002821400523818e-06, 'epoch': 1.87}) +[2026-06-19 21:10:01,913][starVLA.training.train_starvla][INFO] - Step 4984, Loss: {'train/loss': 0.2995058000087738, 'train/grad_norm_pre_clip': 60.65419006347656, 'timing/data': 0.018462080974131823, 'timing/model': 1.1532271379837766, 'learning_rate/qwen_vl_interface': 5.002499236507192e-07, 'learning_rate/action_model': 5.0024992365071925e-06, 'epoch': 1.87}) +[2026-06-19 21:10:06,538][starVLA.training.train_starvla][INFO] - Step 4985, Loss: {'train/loss': 0.14775291085243225, 'train/grad_norm_pre_clip': 10.107160568237305, 'timing/data': 0.01696249796077609, 'timing/model': 1.1659848890267313, 'learning_rate/qwen_vl_interface': 5.002196596918993e-07, 'learning_rate/action_model': 5.002196596918993e-06, 'epoch': 1.87}) +[2026-06-19 21:10:11,151][starVLA.training.train_starvla][INFO] - Step 4986, Loss: {'train/loss': 0.092397041618824, 'train/grad_norm_pre_clip': 10.680670738220215, 'timing/data': 0.01725651603192091, 'timing/model': 1.3274938589893281, 'learning_rate/qwen_vl_interface': 5.001913481883623e-07, 'learning_rate/action_model': 5.001913481883623e-06, 'epoch': 1.87}) +[2026-06-19 21:10:15,657][starVLA.training.train_starvla][INFO] - Step 4987, Loss: {'train/loss': 0.11532542109489441, 'train/grad_norm_pre_clip': 15.445293426513672, 'timing/data': 0.01750264095608145, 'timing/model': 1.2172970440005884, 'learning_rate/qwen_vl_interface': 5.001649891517459e-07, 'learning_rate/action_model': 5.001649891517459e-06, 'epoch': 1.87}) +[2026-06-19 21:10:20,100][starVLA.training.train_starvla][INFO] - Step 4988, Loss: {'train/loss': 0.1867087483406067, 'train/grad_norm_pre_clip': 26.51919937133789, 'timing/data': 0.017847193055786192, 'timing/model': 1.1529032259713858, 'learning_rate/qwen_vl_interface': 5.001405825928858e-07, 'learning_rate/action_model': 5.0014058259288575e-06, 'epoch': 1.87}) +[2026-06-19 21:10:24,669][starVLA.training.train_starvla][INFO] - Step 4989, Loss: {'train/loss': 0.07445884495973587, 'train/grad_norm_pre_clip': 13.79100513458252, 'timing/data': 0.016796619980596006, 'timing/model': 1.1576897030463442, 'learning_rate/qwen_vl_interface': 5.001181285218143e-07, 'learning_rate/action_model': 5.0011812852181434e-06, 'epoch': 1.87}) +[2026-06-19 21:10:29,271][starVLA.training.train_starvla][INFO] - Step 4990, Loss: {'train/loss': 0.18990913033485413, 'train/grad_norm_pre_clip': 25.930387496948242, 'timing/data': 0.017493125051259995, 'timing/model': 1.1463393700541928, 'learning_rate/qwen_vl_interface': 5.000976269477613e-07, 'learning_rate/action_model': 5.0009762694776125e-06, 'epoch': 1.87}) +[2026-06-19 21:10:33,897][starVLA.training.train_starvla][INFO] - Step 4991, Loss: {'train/loss': 0.2516540586948395, 'train/grad_norm_pre_clip': 14.265019416809082, 'timing/data': 0.01655830000527203, 'timing/model': 1.1523645829875022, 'learning_rate/qwen_vl_interface': 5.000790778791542e-07, 'learning_rate/action_model': 5.000790778791543e-06, 'epoch': 1.87}) +[2026-06-19 21:10:38,522][starVLA.training.train_starvla][INFO] - Step 4992, Loss: {'train/loss': 0.2718396484851837, 'train/grad_norm_pre_clip': 25.602033615112305, 'timing/data': 0.016998704057186842, 'timing/model': 1.3269442369928584, 'learning_rate/qwen_vl_interface': 5.000624813236185e-07, 'learning_rate/action_model': 5.000624813236185e-06, 'epoch': 1.87}) +[2026-06-19 21:10:43,034][starVLA.training.train_starvla][INFO] - Step 4993, Loss: {'train/loss': 0.32061755657196045, 'train/grad_norm_pre_clip': 24.382709503173828, 'timing/data': 0.017128924024291337, 'timing/model': 1.222479976946488, 'learning_rate/qwen_vl_interface': 5.000478372879752e-07, 'learning_rate/action_model': 5.000478372879752e-06, 'epoch': 1.87}) +[2026-06-19 21:10:47,478][starVLA.training.train_starvla][INFO] - Step 4994, Loss: {'train/loss': 0.23502935469150543, 'train/grad_norm_pre_clip': 19.29374122619629, 'timing/data': 0.017091927002184093, 'timing/model': 1.157576333032921, 'learning_rate/qwen_vl_interface': 5.000351457782452e-07, 'learning_rate/action_model': 5.000351457782451e-06, 'epoch': 1.87}) +[2026-06-19 21:10:52,004][starVLA.training.train_starvla][INFO] - Step 4995, Loss: {'train/loss': 0.32380351424217224, 'train/grad_norm_pre_clip': 16.507932662963867, 'timing/data': 0.016842862009070814, 'timing/model': 1.1528372539905831, 'learning_rate/qwen_vl_interface': 5.000244067996445e-07, 'learning_rate/action_model': 5.000244067996445e-06, 'epoch': 1.87}) +[2026-06-19 21:10:56,577][starVLA.training.train_starvla][INFO] - Step 4996, Loss: {'train/loss': 0.20047695934772491, 'train/grad_norm_pre_clip': 14.109020233154297, 'timing/data': 0.01747054303996265, 'timing/model': 1.1506371160503477, 'learning_rate/qwen_vl_interface': 5.000156203565884e-07, 'learning_rate/action_model': 5.000156203565883e-06, 'epoch': 1.87}) +[2026-06-19 21:11:01,159][starVLA.training.train_starvla][INFO] - Step 4997, Loss: {'train/loss': 0.12323866039514542, 'train/grad_norm_pre_clip': 19.238262176513672, 'timing/data': 0.01726228801999241, 'timing/model': 1.1429292750544846, 'learning_rate/qwen_vl_interface': 5.000087864526878e-07, 'learning_rate/action_model': 5.000087864526878e-06, 'epoch': 1.87}) +[2026-06-19 21:11:05,777][starVLA.training.train_starvla][INFO] - Step 4998, Loss: {'train/loss': 0.17176860570907593, 'train/grad_norm_pre_clip': 13.032869338989258, 'timing/data': 0.017029643058776855, 'timing/model': 1.335681346943602, 'learning_rate/qwen_vl_interface': 5.000039050907523e-07, 'learning_rate/action_model': 5.000039050907523e-06, 'epoch': 1.87}) +[2026-06-19 21:11:10,208][starVLA.training.train_starvla][INFO] - Step 4999, Loss: {'train/loss': 0.20029623806476593, 'train/grad_norm_pre_clip': 22.75592613220215, 'timing/data': 0.016422607004642487, 'timing/model': 1.1453431049594656, 'learning_rate/qwen_vl_interface': 5.000009762727885e-07, 'learning_rate/action_model': 5.000009762727884e-06, 'epoch': 1.87}) +[2026-06-19 21:11:14,817][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 5000 +[2026-06-19 21:14:11,449][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 5000 +[2026-06-19 21:14:11,450][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 5000 +[2026-06-19 21:17:18,786][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 5000 +[2026-06-19 21:17:18,787][starVLA.training.train_starvla][INFO] - Step 5000, Loss: {'train/loss': 0.15397807955741882, 'train/grad_norm_pre_clip': 14.903421401977539, 'eval/loss': 0.1269863396883011, 'eval/latency_0/loss': 0.08197925984859467, 'eval/latency_1/loss': 0.04751233011484146, 'eval/latency_2/loss': 0.112636037170887, 'eval/latency_3/loss': 0.15563225746154785, 'eval/latency_4/loss': 0.23842492699623108, 'eval/flappy/loss': 0.1269863396883011, 'eval/flappy/latency_0/loss': 0.08197925984859467, 'eval/flappy/latency_1/loss': 0.04751233011484146, 'eval/flappy/latency_2/loss': 0.112636037170887, 'eval/flappy/latency_3/loss': 0.15563225746154785, 'eval/flappy/latency_4/loss': 0.23842492699623108, 'eval/flap_precision': 0.7376425855513308, 'eval/flap_recall': 0.7054545454545454, 'eval/flap_f1': 0.7211895910780669, 'eval/flap_rate_model': 0.07305555555555555, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.7211895910780669, 'eval/latency_0/flap_precision': 0.7376425855513308, 'eval/latency_0/flap_recall': 0.7054545454545454, 'eval/latency_0/flap_f1': 0.7211895910780669, 'eval/latency_0/flap_rate_model': 0.07305555555555555, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.7211895910780669, 'timing/data': 0.01687914796639234, 'timing/model': 1.1559716889169067, 'learning_rate/qwen_vl_interface': 5e-07, 'learning_rate/action_model': 4.9999999999999996e-06, 'epoch': 1.87}) +[2026-06-19 21:17:18,787][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_5000_state +[2026-06-19 21:17:18,790][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-19 21:17:22,185][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_5000_state/model.safetensors +[2026-06-19 21:17:29,293][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_5000_state/optimizer.bin +[2026-06-19 21:17:29,293][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_5000_state/sampler.bin +[2026-06-19 21:17:29,293][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_5000_state/sampler_1.bin +[2026-06-19 21:17:29,294][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints/steps_5000_state/random_states_0.pkl +[2026-06-19 21:17:29,315][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 5000 +[2026-06-19 21:17:29,316][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-19 21:17:29,329][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-19 21:17:29,329][starVLA.training.train_starvla][INFO] - Training complete. Final training state is saved under /workspace/starVLA/results/Checkpoints/flappy_curriculum_cumulative_40ep_per_latency/checkpoints +[2026-06-19 21:17:32,905][starVLA.training.train_starvla][INFO] - ... and that's all, folks! diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_curriculum_cumulative_40ep_per_latency/summary.jsonl b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_curriculum_cumulative_40ep_per_latency/summary.jsonl new file mode 100644 index 0000000000000000000000000000000000000000..ffa1ed16fa7cd08f0c83739d659e9d2a8bd4e827 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_curriculum_cumulative_40ep_per_latency/summary.jsonl @@ -0,0 +1,50 @@ +{"steps": 100} +{"steps": 200} +{"steps": 300} +{"steps": 400} +{"steps": 500} +{"steps": 600} +{"steps": 700} +{"steps": 800} +{"steps": 900} +{"steps": 1000} +{"steps": 1100} +{"steps": 1200} +{"steps": 1300} +{"steps": 1400} +{"steps": 1500} +{"steps": 1600} +{"steps": 1700} +{"steps": 1800} +{"steps": 1900} +{"steps": 2000} +{"steps": 2100} +{"steps": 2200} +{"steps": 2300} +{"steps": 2400} +{"steps": 2500} +{"steps": 2600} +{"steps": 2700} +{"steps": 2800} +{"steps": 2900} +{"steps": 3000} +{"steps": 3100} +{"steps": 3200} +{"steps": 3300} +{"steps": 3400} +{"steps": 3500} +{"steps": 3600} +{"steps": 3700} +{"steps": 3800} +{"steps": 3900} +{"steps": 4000} +{"steps": 4100} +{"steps": 4200} +{"steps": 4300} +{"steps": 4400} +{"steps": 4500} +{"steps": 4600} +{"steps": 4700} +{"steps": 4800} +{"steps": 4900} +{"steps": 5000} diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_5000_model.safetensors b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_5000_model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..2b0c37b9113313f874acb4fd768efb02cad9b705 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_5000_model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4a6b8739fd13e8de24eaacf3f1e133ea6792e01f66cd919ed980d744c0a52e4e +size 9138230516 diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_curriculum_exclusive_40ep_per_latency/config.full.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_curriculum_exclusive_40ep_per_latency/config.full.yaml new file mode 100644 index 0000000000000000000000000000000000000000..40e253e8100a87aadde720e05dcb4efa5f42f1e7 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_curriculum_exclusive_40ep_per_latency/config.full.yaml @@ -0,0 +1,224 @@ +framework: + name: QwenOFT + qwenvl: + base_vlm: /workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + attn_implementation: flash_attention_2 + enable_gradient_checkpointing: true + action_model: + action_model_type: MLP + action_dim: 7 + action_hidden_dim: 2560 + future_action_window_size: 0 + past_action_window_size: 0 + loss_type: discrete_ce + state_dim: 7 + action_horizon: 1 + action_env_dim: 2 +datasets: + vla_data: + dataset_py: lerobot_datasets + include_state: true + data_root_dir: /workspace/starVLA/data/flappy_mixed_latency_40ep_per_lat + data_mix: flappy_mixed_latency_train__bridge + eval_data_mix: flappy_mixed_latency_train__bridge__val + custom_mixtures_path: null + action_type: discrete + sequential_step_sampling: true + eval_sequential_step_sampling: false + num_workers: 4 + eval_num_workers: 0 + shuffle: true + action_balance: + enabled: false + strategy: balanced_epoch + action_key: action_id + target_flap_fraction: 0.3 + noop_id: 0 + flap_id: 1 + latency_curriculum: + enabled: true + strategy: exclusive + latencies: + - 0 + - 1 + - 2 + - 3 + - 4 + phase_steps: null + per_device_batch_size: 32 + load_all_data_for_training: true + obs_image_size: null + video_backend: torchvision_av +dataset: + source_hf: '' + config_name: null + source_subdir: null + converted_name: flappy_mixed_latency_train + single_source_hf: '' + mixed_source_hf: '' + single_converted_name: flappy_train + mixed_converted_name: flappy_mixed_latency_train + single_latency_filter: null + mixed_latency_filter: null + force_download: false + setup_force: false + skip_verification: false + verify_rows: 200 + max_episodes: null + episodes_per_latency: null + latency_filter: null + debug_subset: + enabled: false + max_episodes: 5 + suffix: debug +base_model: + repo_id: Qwen/Qwen3-VL-4B-Instruct +initialization: + checkpoint_local_dir: playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_hf_repo_id: StarVLA/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_filename: checkpoints/steps_5000_pytorch_model.pt +trainer: + max_train_steps: 5000 + num_warmup_steps: 100 + save_interval: 100 + eval_interval: 100 + eval_num_batches: 100 + per_latency_eval_num_batches: null + eval_action_classification: true + eval_action_classification_interval: null + cc_f1_tolerance: 1 + learning_rate: + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + action_model: 0.0001 + lr_scheduler_type: cosine_with_min_lr + scheduler_specific_kwargs: + min_lr: 1.0e-06 + freeze_modules: '' + loss_scale: + vla: 1.0 + vlm: 0.1 + max_grad_norm: 1.0 + weight_decay: 0.0 + logging_frequency: 1 + gradient_clipping: 1.0 + gradient_accumulation_steps: 4 + distributed_backend: none + is_resume: true + pretrained_checkpoint: /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_4100_state + resume_step: 4100 + reload_modules: null + optimizer: + name: AdamW + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + weight_decay: 1.0e-08 + fused: false + save_format: pt +workspace_dir: WORKSPACE_DIR +run_root_dir: /workspace/starVLA/results/Checkpoints +seed: 42 +wandb_entity: dongqianyu99-zhejiang-university +wandb_project: starVLA_rl_games +auth: + env_file: null + hf_token_env: HF_TOKEN + wandb_api_key_env: WANDB_API_KEY +paths: + run_root_dir: results/Checkpoints + dataset_local_dir: playground/Datasets/rl_games + dataset_cache_dir: null + base_model_dir: playground/Pretrained_models/Qwen3-VL-4B-Instruct + accelerate_config: starVLA/config/deepseeds/deepspeed_zero2.yaml +launch: + use_accelerate: true + gpus: null + num_processes: 1 + dry_run: false +conda: + enabled: true + env_name: null +rl_games: + model_alias: openvla + env_eval: + image_size: 224 + frameskip: 1 + seed: 42 + fixed_episode_seeds: true + latency_seed_stride: 0 + task_seed_stride: 0 + task_description: '' + enabled: true + distributed_mode: none + vectorized: + enabled: false + batch_size: 1 + latency: + prompt_map_path: /workspace/starVLA/data/flappy_mixed_latency_40ep_per_lat/flappy_mixed_latency_train__bridge/latency_prompt_map.json + mode: mixed + values: + - 0 + - 1 + - 2 + - 3 + - 4 + mid_train: + enabled: false + interval_steps: 100 + latencies: + - 0 + - 1 + - 2 + - 3 + - 4 + - 5 + num_episodes: 5 + max_steps_per_episode: 3600 + post_train: + enabled: false + latencies: + - 0 + - 1 + - 2 + - 3 + - 4 + num_episodes: 5 + max_steps_per_episode: 3600 + task: flappy + initialization_mode: bridge + action_carrier: bridge +model: openvla +env: flappy +init: bridge +bridge_base_model: + repo_id: + openvla: Qwen/Qwen3-VL-4B-Instruct + pi0: StarVLA/Qwen2.5-VL-3B-Instruct-Action + pi05: Qwen/Qwen3-VL-4B-Instruct + gr00t: Qwen/Qwen3-VL-4B-Instruct + local_dir: + openvla: playground/Pretrained_models/Qwen3-VL-4B-Instruct + pi0: playground/Pretrained_models/Qwen2.5-VL-3B-Instruct-Action + pi05: playground/Pretrained_models/Qwen3-VL-4B-Instruct + gr00t: playground/Pretrained_models/Qwen3-VL-4B-Instruct +mode: curriculum_exclusive +checkpoint: + load: auto + hf_repo_id: null + save_best_model: false + save_pt_file: false + local: + keep_last_n: 1 + sync: + enabled: false + repo_id: null + keep_last_n: 0 + sync_every_n_checkpoints: 1 + resume_policy: local_latest +run_id: flappy_curriculum_exclusive_40ep_per_latency +output_dir: /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency +config_yaml: null +is_debug: false +version_id: '0.21' diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_curriculum_exclusive_40ep_per_latency/config.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_curriculum_exclusive_40ep_per_latency/config.yaml new file mode 100644 index 0000000000000000000000000000000000000000..837e5a33b1f931129d0eac53ff9e7dd07f35a167 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_curriculum_exclusive_40ep_per_latency/config.yaml @@ -0,0 +1,83 @@ +checkpoint: + local: + keep_last_n: 1 + save_best_model: false + save_pt_file: false + sync: + enabled: false + keep_last_n: 0 + repo_id: null +datasets: + vla_data: + data_mix: flappy_mixed_latency_train__bridge + dataset_py: lerobot_datasets + eval_data_mix: flappy_mixed_latency_train__bridge__val + latency_curriculum: + enabled: true + latencies: + - 0 + - 1 + - 2 + - 3 + - 4 + phase_steps: null + strategy: exclusive + obs_image_size: null + per_device_batch_size: 32 +framework: + action_model: + action_dim: 7 + action_env_dim: 2 + action_hidden_dim: 2560 + action_horizon: 1 + action_model_type: MLP + loss_type: discrete_ce + name: QwenOFT + qwenvl: + attn_implementation: flash_attention_2 + base_vlm: /workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + enable_gradient_checkpointing: true +output_dir: /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency +rl_games: + env_eval: + enabled: true + mid_train: + enabled: false + interval_steps: 100 + task: flappy +run_id: flappy_curriculum_exclusive_40ep_per_latency +run_root_dir: /workspace/starVLA/results/Checkpoints +seed: 42 +trainer: + distributed_backend: none + eval_action_classification: true + eval_action_classification_interval: null + eval_interval: 100 + eval_num_batches: 100 + freeze_modules: '' + gradient_accumulation_steps: 4 + gradient_clipping: 1.0 + is_resume: true + learning_rate: + action_model: 0.0001 + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + logging_frequency: 1 + lr_scheduler_type: cosine_with_min_lr + max_train_steps: 5000 + num_warmup_steps: 100 + optimizer: + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + fused: false + weight_decay: 1.0e-08 + per_latency_eval_num_batches: null + pretrained_checkpoint: /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_4100_state + resume_step: 4100 + save_interval: 100 + scheduler_specific_kwargs: + min_lr: 1.0e-06 +wandb_entity: dongqianyu99-zhejiang-university +wandb_project: starVLA_rl_games diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_curriculum_exclusive_40ep_per_latency/dataset_statistics.json b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_curriculum_exclusive_40ep_per_latency/dataset_statistics.json new file mode 100644 index 0000000000000000000000000000000000000000..00c5bd8bf289b8861d281fa80c419f1969ba9d24 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_curriculum_exclusive_40ep_per_latency/dataset_statistics.json @@ -0,0 +1,127 @@ +{ + "new_embodiment": { + "action": { + "mean": [ + 0.6633877158164978, + 0.3366122841835022, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.4729187786579132, + 0.4729187786579132, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "mask": [ + true, + true, + true, + true, + true, + true, + true + ] + }, + "state": { + "mean": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ] + }, + "num_transitions": 341369, + "num_trajectories": 200 + } +} \ No newline at end of file diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_curriculum_exclusive_40ep_per_latency/dataset_statistics_eval.json b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_curriculum_exclusive_40ep_per_latency/dataset_statistics_eval.json new file mode 100644 index 0000000000000000000000000000000000000000..83b9bfe2a5a1d63a6bf5004f3b884056fb467ea8 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_curriculum_exclusive_40ep_per_latency/dataset_statistics_eval.json @@ -0,0 +1,127 @@ +{ + "new_embodiment": { + "action": { + "mean": [ + 0.8398444652557373, + 0.1601555496454239, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.3672837018966675, + 0.3672837018966675, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "mask": [ + true, + true, + true, + true, + true, + true, + true + ] + }, + "state": { + "mean": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ] + }, + "num_transitions": 360000, + "num_trajectories": 100 + } +} \ No newline at end of file diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_curriculum_exclusive_40ep_per_latency/hydra/.hydra/config.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_curriculum_exclusive_40ep_per_latency/hydra/.hydra/config.yaml new file mode 100644 index 0000000000000000000000000000000000000000..3f6555525e68c67819f3093b43d233caffbc098f --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_curriculum_exclusive_40ep_per_latency/hydra/.hydra/config.yaml @@ -0,0 +1,222 @@ +framework: + qwenvl: + base_vlm: /workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + attn_implementation: flash_attention_2 + enable_gradient_checkpointing: true + action_model: + state_dim: 7 + loss_type: discrete_ce + action_horizon: 1 + future_action_window_size: 0 + past_action_window_size: 0 + action_dim: 7 + action_env_dim: 2 + name: QwenOFT +datasets: + vla_data: + dataset_py: lerobot_datasets + include_state: true + data_root_dir: /workspace/starVLA/data/flappy_mixed_latency_40ep_per_lat + data_mix: flappy_mixed_latency_train__bridge + eval_data_mix: flappy_mixed_latency_train__bridge__val + custom_mixtures_path: null + action_type: discrete + sequential_step_sampling: true + eval_sequential_step_sampling: false + num_workers: 4 + eval_num_workers: 0 + shuffle: true + action_balance: + enabled: false + strategy: balanced_epoch + action_key: action_id + target_flap_fraction: 0.3 + noop_id: 0 + flap_id: 1 + latency_curriculum: + enabled: true + strategy: exclusive + latencies: + - 0 + - 1 + - 2 + - 3 + - 4 + phase_steps: null + per_device_batch_size: 32 + load_all_data_for_training: true + obs_image_size: null + video_backend: torchvision_av +dataset: + source_hf: ${dataset.mixed_source_hf} + config_name: null + source_subdir: null + converted_name: ${dataset.mixed_converted_name} + single_source_hf: '' + mixed_source_hf: '' + single_converted_name: flappy_train + mixed_converted_name: flappy_mixed_latency_train + single_latency_filter: null + mixed_latency_filter: null + force_download: false + setup_force: false + skip_verification: false + verify_rows: 200 + max_episodes: null + episodes_per_latency: null + latency_filter: ${dataset.mixed_latency_filter} + debug_subset: + enabled: false + max_episodes: 5 + suffix: debug +base_model: + repo_id: ${bridge_base_model.repo_id.${model}} +initialization: + checkpoint_local_dir: playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_hf_repo_id: StarVLA/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_filename: checkpoints/steps_5000_pytorch_model.pt +trainer: + max_train_steps: 5000 + num_warmup_steps: 100 + save_interval: 100 + eval_interval: 100 + eval_num_batches: 100 + per_latency_eval_num_batches: null + eval_action_classification: true + eval_action_classification_interval: null + cc_f1_tolerance: 1 + learning_rate: + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + action_model: 0.0001 + lr_scheduler_type: cosine_with_min_lr + scheduler_specific_kwargs: + min_lr: 1.0e-06 + freeze_modules: '' + loss_scale: + vla: 1.0 + vlm: 0.1 + max_grad_norm: 1.0 + weight_decay: 0.0 + logging_frequency: 1 + gradient_clipping: 1.0 + gradient_accumulation_steps: 4 + distributed_backend: none + is_resume: true + pretrained_checkpoint: /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_4100_state + resume_step: 4100 + reload_modules: null + optimizer: + name: AdamW + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + weight_decay: 1.0e-08 + fused: false + save_format: pt +workspace_dir: WORKSPACE_DIR +run_root_dir: /workspace/starVLA/results/Checkpoints +seed: 42 +wandb_entity: dongqianyu99-zhejiang-university +wandb_project: starVLA_rl_games +auth: + env_file: null + hf_token_env: HF_TOKEN + wandb_api_key_env: WANDB_API_KEY +paths: + run_root_dir: results/Checkpoints + dataset_local_dir: playground/Datasets/rl_games + dataset_cache_dir: null + base_model_dir: ${bridge_base_model.local_dir.${model}} + accelerate_config: starVLA/config/deepseeds/deepspeed_zero2.yaml +launch: + use_accelerate: true + gpus: null + num_processes: 1 + dry_run: false +conda: + enabled: true + env_name: null +rl_games: + model_alias: openvla + env_eval: + image_size: 224 + frameskip: 1 + seed: 42 + fixed_episode_seeds: true + latency_seed_stride: 0 + task_seed_stride: 0 + task_description: '' + enabled: true + distributed_mode: none + vectorized: + enabled: false + batch_size: 1 + latency: + prompt_map_path: /workspace/starVLA/data/flappy_mixed_latency_40ep_per_lat/flappy_mixed_latency_train__bridge/latency_prompt_map.json + mode: mixed + values: + - 0 + - 1 + - 2 + - 3 + - 4 + mid_train: + enabled: false + interval_steps: 100 + latencies: + - 0 + - 1 + - 2 + - 3 + - 4 + - 5 + num_episodes: 5 + max_steps_per_episode: 3600 + post_train: + enabled: false + latencies: + - 0 + - 1 + - 2 + - 3 + - 4 + num_episodes: 5 + max_steps_per_episode: 3600 + task: flappy + initialization_mode: bridge + action_carrier: bridge +model: openvla +env: flappy +init: bridge +bridge_base_model: + repo_id: + openvla: Qwen/Qwen3-VL-4B-Instruct + pi0: StarVLA/Qwen2.5-VL-3B-Instruct-Action + pi05: Qwen/Qwen3-VL-4B-Instruct + gr00t: Qwen/Qwen3-VL-4B-Instruct + local_dir: + openvla: playground/Pretrained_models/Qwen3-VL-4B-Instruct + pi0: playground/Pretrained_models/Qwen2.5-VL-3B-Instruct-Action + pi05: playground/Pretrained_models/Qwen3-VL-4B-Instruct + gr00t: playground/Pretrained_models/Qwen3-VL-4B-Instruct +mode: curriculum_exclusive +checkpoint: + load: auto + hf_repo_id: null + save_best_model: false + save_pt_file: false + local: + keep_last_n: 1 + sync: + enabled: false + repo_id: null + keep_last_n: 0 + sync_every_n_checkpoints: 1 + resume_policy: local_latest +run_id: flappy_curriculum_exclusive_40ep_per_latency +output_dir: null +config_yaml: null +is_debug: false +version_id: '0.21' diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_curriculum_exclusive_40ep_per_latency/hydra/.hydra/hydra.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_curriculum_exclusive_40ep_per_latency/hydra/.hydra/hydra.yaml new file mode 100644 index 0000000000000000000000000000000000000000..c4d90f59717271954d70bf2422b35224d64d8434 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_curriculum_exclusive_40ep_per_latency/hydra/.hydra/hydra.yaml @@ -0,0 +1,255 @@ +hydra: + run: + dir: ${run_root_dir}/${run_id}/hydra + sweep: + dir: multirun/${now:%Y-%m-%d}/${now:%H-%M-%S} + subdir: ${hydra.job.num} + launcher: + _target_: hydra._internal.core_plugins.basic_launcher.BasicLauncher + sweeper: + _target_: hydra._internal.core_plugins.basic_sweeper.BasicSweeper + max_batch_size: null + params: null + help: + app_name: ${hydra.job.name} + header: '${hydra.help.app_name} is powered by Hydra. + + ' + footer: 'Powered by Hydra (https://hydra.cc) + + Use --hydra-help to view Hydra specific help + + ' + template: '${hydra.help.header} + + == Configuration groups == + + Compose your configuration from those groups (group=option) + + + $APP_CONFIG_GROUPS + + + == Config == + + Override anything in the config (foo.bar=value) + + + $CONFIG + + + ${hydra.help.footer} + + ' + hydra_help: + template: 'Hydra (${hydra.runtime.version}) + + See https://hydra.cc for more info. + + + == Flags == + + $FLAGS_HELP + + + == Configuration groups == + + Compose your configuration from those groups (For example, append hydra/job_logging=disabled + to command line) + + + $HYDRA_CONFIG_GROUPS + + + Use ''--cfg hydra'' to Show the Hydra config. + + ' + hydra_help: ??? + hydra_logging: + version: 1 + formatters: + simple: + format: '[%(asctime)s][HYDRA] %(message)s' + handlers: + console: + class: logging.StreamHandler + formatter: simple + stream: ext://sys.stdout + root: + level: INFO + handlers: + - console + loggers: + logging_example: + level: DEBUG + disable_existing_loggers: false + job_logging: + version: 1 + formatters: + simple: + format: '[%(asctime)s][%(name)s][%(levelname)s] - %(message)s' + handlers: + console: + class: logging.StreamHandler + formatter: simple + stream: ext://sys.stdout + file: + class: logging.FileHandler + formatter: simple + filename: ${hydra.runtime.output_dir}/${hydra.job.name}.log + root: + level: INFO + handlers: + - console + - file + disable_existing_loggers: false + env: {} + mode: RUN + searchpath: [] + callbacks: {} + output_subdir: .hydra + overrides: + hydra: + - hydra.mode=RUN + task: + - model=openvla + - env=flappy + - init=bridge + - mode=curriculum_exclusive + - run_id=flappy_curriculum_exclusive_40ep_per_latency + - run_root_dir=/workspace/starVLA/results/Checkpoints + - seed=42 + - wandb_entity=dongqianyu99-zhejiang-university + - wandb_project=starVLA_rl_games + - rl_games.env_eval.enabled=true + - checkpoint.sync.enabled=false + - checkpoint.sync.keep_last_n=0 + - checkpoint.local.keep_last_n=1 + - checkpoint.save_best_model=false + - checkpoint.save_pt_file=false + - trainer.is_resume=true + - trainer.pretrained_checkpoint=/workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_4100_state + - trainer.resume_step=4100 + - trainer.max_train_steps=5000 + - trainer.num_warmup_steps=100 + - trainer.save_interval=100 + - trainer.eval_interval=100 + - trainer.eval_num_batches=100 + - trainer.logging_frequency=1 + - trainer.gradient_accumulation_steps=4 + - trainer.distributed_backend=none + - trainer.learning_rate.base=2e-05 + - trainer.learning_rate.qwen_vl_interface=1e-05 + - trainer.learning_rate.action_model=0.0001 + - trainer.lr_scheduler_type=cosine_with_min_lr + - trainer.scheduler_specific_kwargs.min_lr=1e-06 + - trainer.loss_scale.vla=1.0 + - trainer.loss_scale.vlm=0.1 + - trainer.max_grad_norm=1.0 + - trainer.weight_decay=0.0 + - trainer.gradient_clipping=1.0 + - trainer.optimizer.name=AdamW + - trainer.optimizer.betas=[0.9,0.95] + - trainer.optimizer.eps=1e-08 + - trainer.optimizer.weight_decay=1e-08 + - trainer.optimizer.fused=false + - trainer.save_format=pt + - framework.name=QwenOFT + - framework.qwenvl.attn_implementation=flash_attention_2 + - framework.qwenvl.enable_gradient_checkpointing=true + - framework.action_model.action_dim=7 + - framework.action_model.action_env_dim=2 + - framework.action_model.state_dim=7 + - framework.action_model.loss_type=discrete_ce + - framework.action_model.action_horizon=1 + - framework.action_model.future_action_window_size=0 + - framework.action_model.past_action_window_size=0 + - datasets.vla_data.include_state=true + - datasets.vla_data.action_type=discrete + - datasets.vla_data.sequential_step_sampling=true + - datasets.vla_data.shuffle=true + - datasets.vla_data.action_balance.enabled=false + - datasets.vla_data.action_balance.strategy=balanced_epoch + - datasets.vla_data.action_balance.action_key=action_id + - datasets.vla_data.action_balance.target_flap_fraction=0.3 + - datasets.vla_data.action_balance.noop_id=0 + - datasets.vla_data.action_balance.flap_id=1 + - datasets.vla_data.latency_curriculum.enabled=true + - datasets.vla_data.latency_curriculum.strategy=exclusive + - datasets.vla_data.latency_curriculum.latencies=[0,1,2,3,4] + - datasets.vla_data.per_device_batch_size=32 + - datasets.vla_data.load_all_data_for_training=true + - datasets.vla_data.video_backend=torchvision_av + - datasets.vla_data.data_root_dir=/workspace/starVLA/data/flappy_mixed_latency_40ep_per_lat + - datasets.vla_data.data_mix=flappy_mixed_latency_train__bridge + - datasets.vla_data.eval_data_mix=flappy_mixed_latency_train__bridge__val + - framework.qwenvl.base_vlm=/workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + - rl_games.task=flappy + - rl_games.model_alias=openvla + - rl_games.initialization_mode=bridge + - rl_games.action_carrier=bridge + - rl_games.env_eval.distributed_mode=none + - rl_games.env_eval.latency.mode=mixed + - rl_games.env_eval.frameskip=1 + - rl_games.env_eval.image_size=224 + - rl_games.env_eval.seed=42 + - rl_games.env_eval.fixed_episode_seeds=true + - rl_games.env_eval.latency_seed_stride=0 + - rl_games.env_eval.task_seed_stride=0 + - rl_games.env_eval.latency.values=[0,1,2,3,4] + - rl_games.env_eval.latency.prompt_map_path=/workspace/starVLA/data/flappy_mixed_latency_40ep_per_lat/flappy_mixed_latency_train__bridge/latency_prompt_map.json + - rl_games.env_eval.mid_train.enabled=false + - rl_games.env_eval.mid_train.interval_steps=100 + - rl_games.env_eval.mid_train.num_episodes=5 + - rl_games.env_eval.mid_train.max_steps_per_episode=3600 + - rl_games.env_eval.mid_train.latencies=[0,1,2,3,4,5] + - rl_games.env_eval.post_train.enabled=false + - rl_games.env_eval.post_train.num_episodes=5 + - rl_games.env_eval.post_train.max_steps_per_episode=3600 + - rl_games.env_eval.post_train.latencies=[0,1,2,3,4] + job: + name: train_starvla_hydra + chdir: false + override_dirname: checkpoint.local.keep_last_n=1,checkpoint.save_best_model=false,checkpoint.save_pt_file=false,checkpoint.sync.enabled=false,checkpoint.sync.keep_last_n=0,datasets.vla_data.action_balance.action_key=action_id,datasets.vla_data.action_balance.enabled=false,datasets.vla_data.action_balance.flap_id=1,datasets.vla_data.action_balance.noop_id=0,datasets.vla_data.action_balance.strategy=balanced_epoch,datasets.vla_data.action_balance.target_flap_fraction=0.3,datasets.vla_data.action_type=discrete,datasets.vla_data.data_mix=flappy_mixed_latency_train__bridge,datasets.vla_data.data_root_dir=/workspace/starVLA/data/flappy_mixed_latency_40ep_per_lat,datasets.vla_data.eval_data_mix=flappy_mixed_latency_train__bridge__val,datasets.vla_data.include_state=true,datasets.vla_data.latency_curriculum.enabled=true,datasets.vla_data.latency_curriculum.latencies=[0,1,2,3,4],datasets.vla_data.latency_curriculum.strategy=exclusive,datasets.vla_data.load_all_data_for_training=true,datasets.vla_data.per_device_batch_size=32,datasets.vla_data.sequential_step_sampling=true,datasets.vla_data.shuffle=true,datasets.vla_data.video_backend=torchvision_av,env=flappy,framework.action_model.action_dim=7,framework.action_model.action_env_dim=2,framework.action_model.action_horizon=1,framework.action_model.future_action_window_size=0,framework.action_model.loss_type=discrete_ce,framework.action_model.past_action_window_size=0,framework.action_model.state_dim=7,framework.name=QwenOFT,framework.qwenvl.attn_implementation=flash_attention_2,framework.qwenvl.base_vlm=/workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct,framework.qwenvl.enable_gradient_checkpointing=true,init=bridge,mode=curriculum_exclusive,model=openvla,rl_games.action_carrier=bridge,rl_games.env_eval.distributed_mode=none,rl_games.env_eval.enabled=true,rl_games.env_eval.fixed_episode_seeds=true,rl_games.env_eval.frameskip=1,rl_games.env_eval.image_size=224,rl_games.env_eval.latency.mode=mixed,rl_games.env_eval.latency.prompt_map_path=/workspace/starVLA/data/flappy_mixed_latency_40ep_per_lat/flappy_mixed_latency_train__bridge/latency_prompt_map.json,rl_games.env_eval.latency.values=[0,1,2,3,4],rl_games.env_eval.latency_seed_stride=0,rl_games.env_eval.mid_train.enabled=false,rl_games.env_eval.mid_train.interval_steps=100,rl_games.env_eval.mid_train.latencies=[0,1,2,3,4,5],rl_games.env_eval.mid_train.max_steps_per_episode=3600,rl_games.env_eval.mid_train.num_episodes=5,rl_games.env_eval.post_train.enabled=false,rl_games.env_eval.post_train.latencies=[0,1,2,3,4],rl_games.env_eval.post_train.max_steps_per_episode=3600,rl_games.env_eval.post_train.num_episodes=5,rl_games.env_eval.seed=42,rl_games.env_eval.task_seed_stride=0,rl_games.initialization_mode=bridge,rl_games.model_alias=openvla,rl_games.task=flappy,run_id=flappy_curriculum_exclusive_40ep_per_latency,run_root_dir=/workspace/starVLA/results/Checkpoints,seed=42,trainer.distributed_backend=none,trainer.eval_interval=100,trainer.eval_num_batches=100,trainer.gradient_accumulation_steps=4,trainer.gradient_clipping=1.0,trainer.is_resume=true,trainer.learning_rate.action_model=0.0001,trainer.learning_rate.base=2e-05,trainer.learning_rate.qwen_vl_interface=1e-05,trainer.logging_frequency=1,trainer.loss_scale.vla=1.0,trainer.loss_scale.vlm=0.1,trainer.lr_scheduler_type=cosine_with_min_lr,trainer.max_grad_norm=1.0,trainer.max_train_steps=5000,trainer.num_warmup_steps=100,trainer.optimizer.betas=[0.9,0.95],trainer.optimizer.eps=1e-08,trainer.optimizer.fused=false,trainer.optimizer.name=AdamW,trainer.optimizer.weight_decay=1e-08,trainer.pretrained_checkpoint=/workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_4100_state,trainer.resume_step=4100,trainer.save_format=pt,trainer.save_interval=100,trainer.scheduler_specific_kwargs.min_lr=1e-06,trainer.weight_decay=0.0,wandb_entity=dongqianyu99-zhejiang-university,wandb_project=starVLA_rl_games + id: ??? + num: ??? + config_name: train + env_set: {} + env_copy: [] + config: + override_dirname: + kv_sep: '=' + item_sep: ',' + exclude_keys: [] + runtime: + version: 1.3.3 + version_base: '1.1' + cwd: /workspace/starVLA + config_sources: + - path: hydra.conf + schema: pkg + provider: hydra + - path: /workspace/starVLA/examples/rl_games/config + schema: file + provider: main + - path: '' + schema: structured + provider: schema + output_dir: /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/hydra + choices: + cross_task_setup: null + checkpoint: default + mode: curriculum_exclusive + init: bridge + env: flappy + model: openvla + hydra/env: default + hydra/callbacks: null + hydra/job_logging: default + hydra/hydra_logging: default + hydra/hydra_help: default + hydra/help: default + hydra/sweeper: basic + hydra/launcher: basic + hydra/output: default + verbose: false diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_curriculum_exclusive_40ep_per_latency/hydra/.hydra/overrides.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_curriculum_exclusive_40ep_per_latency/hydra/.hydra/overrides.yaml new file mode 100644 index 0000000000000000000000000000000000000000..61133b827f29c294049711328309405f9b266e50 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_curriculum_exclusive_40ep_per_latency/hydra/.hydra/overrides.yaml @@ -0,0 +1,95 @@ +- model=openvla +- env=flappy +- init=bridge +- mode=curriculum_exclusive +- run_id=flappy_curriculum_exclusive_40ep_per_latency +- run_root_dir=/workspace/starVLA/results/Checkpoints +- seed=42 +- wandb_entity=dongqianyu99-zhejiang-university +- wandb_project=starVLA_rl_games +- rl_games.env_eval.enabled=true +- checkpoint.sync.enabled=false +- checkpoint.sync.keep_last_n=0 +- checkpoint.local.keep_last_n=1 +- checkpoint.save_best_model=false +- checkpoint.save_pt_file=false +- trainer.is_resume=true +- trainer.pretrained_checkpoint=/workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_4100_state +- trainer.resume_step=4100 +- trainer.max_train_steps=5000 +- trainer.num_warmup_steps=100 +- trainer.save_interval=100 +- trainer.eval_interval=100 +- trainer.eval_num_batches=100 +- trainer.logging_frequency=1 +- trainer.gradient_accumulation_steps=4 +- trainer.distributed_backend=none +- trainer.learning_rate.base=2e-05 +- trainer.learning_rate.qwen_vl_interface=1e-05 +- trainer.learning_rate.action_model=0.0001 +- trainer.lr_scheduler_type=cosine_with_min_lr +- trainer.scheduler_specific_kwargs.min_lr=1e-06 +- trainer.loss_scale.vla=1.0 +- trainer.loss_scale.vlm=0.1 +- trainer.max_grad_norm=1.0 +- trainer.weight_decay=0.0 +- trainer.gradient_clipping=1.0 +- trainer.optimizer.name=AdamW +- trainer.optimizer.betas=[0.9,0.95] +- trainer.optimizer.eps=1e-08 +- trainer.optimizer.weight_decay=1e-08 +- trainer.optimizer.fused=false +- trainer.save_format=pt +- framework.name=QwenOFT +- framework.qwenvl.attn_implementation=flash_attention_2 +- framework.qwenvl.enable_gradient_checkpointing=true +- framework.action_model.action_dim=7 +- framework.action_model.action_env_dim=2 +- framework.action_model.state_dim=7 +- framework.action_model.loss_type=discrete_ce +- framework.action_model.action_horizon=1 +- framework.action_model.future_action_window_size=0 +- framework.action_model.past_action_window_size=0 +- datasets.vla_data.include_state=true +- datasets.vla_data.action_type=discrete +- datasets.vla_data.sequential_step_sampling=true +- datasets.vla_data.shuffle=true +- datasets.vla_data.action_balance.enabled=false +- datasets.vla_data.action_balance.strategy=balanced_epoch +- datasets.vla_data.action_balance.action_key=action_id +- datasets.vla_data.action_balance.target_flap_fraction=0.3 +- datasets.vla_data.action_balance.noop_id=0 +- datasets.vla_data.action_balance.flap_id=1 +- datasets.vla_data.latency_curriculum.enabled=true +- datasets.vla_data.latency_curriculum.strategy=exclusive +- datasets.vla_data.latency_curriculum.latencies=[0,1,2,3,4] +- datasets.vla_data.per_device_batch_size=32 +- datasets.vla_data.load_all_data_for_training=true +- datasets.vla_data.video_backend=torchvision_av +- datasets.vla_data.data_root_dir=/workspace/starVLA/data/flappy_mixed_latency_40ep_per_lat +- datasets.vla_data.data_mix=flappy_mixed_latency_train__bridge +- datasets.vla_data.eval_data_mix=flappy_mixed_latency_train__bridge__val +- framework.qwenvl.base_vlm=/workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct +- rl_games.task=flappy +- rl_games.model_alias=openvla +- rl_games.initialization_mode=bridge +- rl_games.action_carrier=bridge +- rl_games.env_eval.distributed_mode=none +- rl_games.env_eval.latency.mode=mixed +- rl_games.env_eval.frameskip=1 +- rl_games.env_eval.image_size=224 +- rl_games.env_eval.seed=42 +- rl_games.env_eval.fixed_episode_seeds=true +- rl_games.env_eval.latency_seed_stride=0 +- rl_games.env_eval.task_seed_stride=0 +- rl_games.env_eval.latency.values=[0,1,2,3,4] +- rl_games.env_eval.latency.prompt_map_path=/workspace/starVLA/data/flappy_mixed_latency_40ep_per_lat/flappy_mixed_latency_train__bridge/latency_prompt_map.json +- rl_games.env_eval.mid_train.enabled=false +- rl_games.env_eval.mid_train.interval_steps=100 +- rl_games.env_eval.mid_train.num_episodes=5 +- rl_games.env_eval.mid_train.max_steps_per_episode=3600 +- rl_games.env_eval.mid_train.latencies=[0,1,2,3,4,5] +- rl_games.env_eval.post_train.enabled=false +- rl_games.env_eval.post_train.num_episodes=5 +- rl_games.env_eval.post_train.max_steps_per_episode=3600 +- rl_games.env_eval.post_train.latencies=[0,1,2,3,4] diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_curriculum_exclusive_40ep_per_latency/hydra/train_starvla_hydra.log b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_curriculum_exclusive_40ep_per_latency/hydra/train_starvla_hydra.log new file mode 100644 index 0000000000000000000000000000000000000000..5c738e212d3e35134c99152d52642b2ff17ac3a3 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_curriculum_exclusive_40ep_per_latency/hydra/train_starvla_hydra.log @@ -0,0 +1,6059 @@ +[2026-06-19 15:28:39,979][starVLA.training.train_starvla][INFO] - VLA Training :: Warming Up +[2026-06-19 15:28:39,992][starVLA.training.train_starvla][INFO] - ✅ Configuration wrapped for access tracking +[2026-06-19 15:28:51,198][starVLA.training.train_starvla][INFO] - Preloaded model checkpoint before Accelerator/DeepSpeed initialization: /workspace/starVLA/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt +[2026-06-19 15:28:51,247][starVLA.training.train_starvla][INFO] - Creating VLA Dataset with Mixture `flappy_mixed_latency_train__bridge` +[2026-06-19 15:28:51,981][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Behavior (data_config): ['BEHAVIOR_challenge'] +[2026-06-19 15:28:51,982][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from DOMINO: ['robotwin'] +[2026-06-19 15:28:51,982][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from DOMINO (data_config): [] +[2026-06-19 15:28:51,982][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from DOMINO (data_config): ['domino', 'domino_clean', 'domino_random', 'domino_cotrain'] +[2026-06-19 15:28:51,983][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Franka: ['custom_robot_config', 'demo_sim_franka_delta_joints', 'SO101'] +[2026-06-19 15:28:51,983][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Franka (data_config): [] +[2026-06-19 15:28:51,983][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Franka (data_config): ['custom_dataset', 'custom_dataset_2', 'demo_sim_pick_place', 'SO101_pick'] +[2026-06-19 15:28:51,983][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from LIBERO: ['libero_franka'] +[2026-06-19 15:28:51,983][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from LIBERO (data_config): [] +[2026-06-19 15:28:51,983][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from LIBERO (data_config): ['libero_all', 'libero_goal', 'multi_robot'] +[2026-06-19 15:28:51,984][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from RoboChallenge_table30v2: ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-19 15:28:51,984][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from RoboChallenge_table30v2 (data_config): ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-19 15:28:51,984][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from RoboChallenge_table30v2 (data_config): ['robochallenge_table30v2_shred_paper', 'robochallenge_table30v2_ur5_all', 'robochallenge_table30v2_arx5_all', 'robochallenge_table30v2_dosw1_all'] +[2026-06-19 15:28:51,984][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_365: ['panda_omron_robocasa365'] +[2026-06-19 15:28:51,985][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_365 (data_config): [] +[2026-06-19 15:28:51,985][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_365 (data_config): ['robocasa365_open_drawer_target_human', 'robocasa365_atomic_target_human_all', 'robocasa365_composite_target_human_all', 'robocasa365_target_human_all'] +[2026-06-19 15:28:51,985][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_tabletop: ['fourier_gr1_arms_waist'] +[2026-06-19 15:28:51,985][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_tabletop (data_config): [] +[2026-06-19 15:28:51,985][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_tabletop (data_config): ['fourier_gr1_unified_1000'] +[2026-06-19 15:28:51,986][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robotwin: ['robotwin', 'robotwin50', 'arx_x5'] +[2026-06-19 15:28:51,986][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robotwin (data_config): [] +[2026-06-19 15:28:51,986][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robotwin (data_config): ['robotwin_all', 'robotwin_all_50', 'robotwin', 'robotwin_task1', 'robotwin_task2', 'arx_x5'] +[2026-06-19 15:28:51,987][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from SimplerEnv: ['oxe_droid', 'oxe_bridge', 'oxe_rt1'] +[2026-06-19 15:28:51,987][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from SimplerEnv (data_config): [] +[2026-06-19 15:28:51,987][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from SimplerEnv (data_config): ['bridge', 'bridge_rt_1'] +[2026-06-19 15:28:51,987][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from VLA-Arena: ['vla_arena_franka'] +[2026-06-19 15:28:51,987][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from VLA-Arena (data_config): [] +[2026-06-19 15:28:51,987][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from VLA-Arena (data_config): ['vla_arena_L0_S', 'vla_arena_L0_M', 'vla_arena_L0_L'] +[2026-06-19 15:28:51,988][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from rl_games: ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-19 15:28:51,988][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from rl_games (data_config): ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-19 15:28:51,988][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from rl_games (data_config): ['flappy_train', 'flappy_train__bridge', 'flappy_mixed_latency_train', 'flappy_mixed_latency_train__bridge', 'demon_attack_train', 'demon_attack_train__bridge', 'demon_attack_mixed_latency_train', 'demon_attack_mixed_latency_train__bridge', 'deadly_corridor_train', 'deadly_corridor_train__bridge', 'deadly_corridor_mixed_latency_train', 'deadly_corridor_mixed_latency_train__bridge'] +[2026-06-19 15:28:52,084][starVLA.training.train_starvla][INFO] - Creating VLA Eval Dataset with Mixture `flappy_mixed_latency_train__bridge__val` +[2026-06-19 15:28:52,144][starVLA.training.train_starvla][INFO] - 📝 Full config saved at /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/config.full.yaml +[2026-06-19 15:28:52,152][starVLA.training.train_starvla][INFO] - 📊 Accessed config snapshot saved at /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/config.yaml +[2026-06-19 15:28:52,152][starVLA.training.train_starvla][INFO] - No pretrained checkpoint provided. Starting training from scratch. +[2026-06-19 15:32:24,788][starVLA.training.train_starvla][INFO] - VLA Training :: Warming Up +[2026-06-19 15:32:24,800][starVLA.training.train_starvla][INFO] - ✅ Configuration wrapped for access tracking +[2026-06-19 15:32:35,641][starVLA.training.train_starvla][INFO] - Preloaded model checkpoint before Accelerator/DeepSpeed initialization: /workspace/starVLA/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt +[2026-06-19 15:32:35,689][starVLA.training.train_starvla][INFO] - Creating VLA Dataset with Mixture `flappy_mixed_latency_train__bridge` +[2026-06-19 15:32:36,459][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Behavior (data_config): ['BEHAVIOR_challenge'] +[2026-06-19 15:32:36,460][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from DOMINO: ['robotwin'] +[2026-06-19 15:32:36,460][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from DOMINO (data_config): [] +[2026-06-19 15:32:36,460][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from DOMINO (data_config): ['domino', 'domino_clean', 'domino_random', 'domino_cotrain'] +[2026-06-19 15:32:36,460][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Franka: ['custom_robot_config', 'demo_sim_franka_delta_joints', 'SO101'] +[2026-06-19 15:32:36,461][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Franka (data_config): [] +[2026-06-19 15:32:36,461][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Franka (data_config): ['custom_dataset', 'custom_dataset_2', 'demo_sim_pick_place', 'SO101_pick'] +[2026-06-19 15:32:36,461][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from LIBERO: ['libero_franka'] +[2026-06-19 15:32:36,461][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from LIBERO (data_config): [] +[2026-06-19 15:32:36,461][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from LIBERO (data_config): ['libero_all', 'libero_goal', 'multi_robot'] +[2026-06-19 15:32:36,462][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from RoboChallenge_table30v2: ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-19 15:32:36,462][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from RoboChallenge_table30v2 (data_config): ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-19 15:32:36,462][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from RoboChallenge_table30v2 (data_config): ['robochallenge_table30v2_shred_paper', 'robochallenge_table30v2_ur5_all', 'robochallenge_table30v2_arx5_all', 'robochallenge_table30v2_dosw1_all'] +[2026-06-19 15:32:36,462][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_365: ['panda_omron_robocasa365'] +[2026-06-19 15:32:36,462][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_365 (data_config): [] +[2026-06-19 15:32:36,462][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_365 (data_config): ['robocasa365_open_drawer_target_human', 'robocasa365_atomic_target_human_all', 'robocasa365_composite_target_human_all', 'robocasa365_target_human_all'] +[2026-06-19 15:32:36,463][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_tabletop: ['fourier_gr1_arms_waist'] +[2026-06-19 15:32:36,463][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_tabletop (data_config): [] +[2026-06-19 15:32:36,463][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_tabletop (data_config): ['fourier_gr1_unified_1000'] +[2026-06-19 15:32:36,464][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robotwin: ['robotwin', 'robotwin50', 'arx_x5'] +[2026-06-19 15:32:36,464][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robotwin (data_config): [] +[2026-06-19 15:32:36,464][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robotwin (data_config): ['robotwin_all', 'robotwin_all_50', 'robotwin', 'robotwin_task1', 'robotwin_task2', 'arx_x5'] +[2026-06-19 15:32:36,464][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from SimplerEnv: ['oxe_droid', 'oxe_bridge', 'oxe_rt1'] +[2026-06-19 15:32:36,464][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from SimplerEnv (data_config): [] +[2026-06-19 15:32:36,465][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from SimplerEnv (data_config): ['bridge', 'bridge_rt_1'] +[2026-06-19 15:32:36,465][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from VLA-Arena: ['vla_arena_franka'] +[2026-06-19 15:32:36,465][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from VLA-Arena (data_config): [] +[2026-06-19 15:32:36,465][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from VLA-Arena (data_config): ['vla_arena_L0_S', 'vla_arena_L0_M', 'vla_arena_L0_L'] +[2026-06-19 15:32:36,465][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from rl_games: ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-19 15:32:36,465][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from rl_games (data_config): ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-19 15:32:36,465][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from rl_games (data_config): ['flappy_train', 'flappy_train__bridge', 'flappy_mixed_latency_train', 'flappy_mixed_latency_train__bridge', 'demon_attack_train', 'demon_attack_train__bridge', 'demon_attack_mixed_latency_train', 'demon_attack_mixed_latency_train__bridge', 'deadly_corridor_train', 'deadly_corridor_train__bridge', 'deadly_corridor_mixed_latency_train', 'deadly_corridor_mixed_latency_train__bridge'] +[2026-06-19 15:32:36,560][starVLA.training.train_starvla][INFO] - Creating VLA Eval Dataset with Mixture `flappy_mixed_latency_train__bridge__val` +[2026-06-19 15:32:36,620][starVLA.training.train_starvla][INFO] - 📝 Full config saved at /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/config.full.yaml +[2026-06-19 15:32:36,628][starVLA.training.train_starvla][INFO] - 📊 Accessed config snapshot saved at /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/config.yaml +[2026-06-19 15:32:36,628][starVLA.training.train_starvla][INFO] - No pretrained checkpoint provided. Starting training from scratch. +[2026-06-19 15:32:41,050][starVLA.training.train_starvla][INFO] - ***** Training Configuration ***** +[2026-06-19 15:32:41,051][starVLA.training.train_starvla][INFO] - Total optimization steps = 5000 +[2026-06-19 15:32:41,051][starVLA.training.train_starvla][INFO] - Per device batch size = 32 +[2026-06-19 15:32:41,052][starVLA.training.train_starvla][INFO] - Gradient accumulation steps = 4 +[2026-06-19 15:32:41,052][starVLA.training.train_starvla][INFO] - Total batch size = 128 +[2026-06-19 15:32:41,572][starVLA.training.train_starvla][INFO] - Latency curriculum phase 0 at step 0: active_latencies=[0], dataset_size=62322 +[2026-06-19 15:33:11,252][starVLA.training.train_starvla][INFO] - Step 1, Loss: {'train/loss': 0.6945967674255371, 'train/grad_norm_pre_clip': 7.346078872680664, 'timing/data': 0.03042016737163067, 'timing/model': 2.2849616296589375, 'learning_rate/qwen_vl_interface': 1.0000000000000001e-07, 'learning_rate/action_model': 1.0000000000000002e-06, 'epoch': 0.0}) +[2026-06-19 15:33:20,124][starVLA.training.train_starvla][INFO] - Step 2, Loss: {'train/loss': 0.6955437660217285, 'train/grad_norm_pre_clip': 8.868399620056152, 'timing/data': 0.02744140289723873, 'timing/model': 2.255193565040827, 'learning_rate/qwen_vl_interface': 2.0000000000000002e-07, 'learning_rate/action_model': 2.0000000000000003e-06, 'epoch': 0.0}) +[2026-06-19 15:33:29,033][starVLA.training.train_starvla][INFO] - Step 3, Loss: {'train/loss': 0.6885901689529419, 'train/grad_norm_pre_clip': 5.4768853187561035, 'timing/data': 0.025778260082006454, 'timing/model': 2.270680269226432, 'learning_rate/qwen_vl_interface': 3.0000000000000004e-07, 'learning_rate/action_model': 3e-06, 'epoch': 0.01}) +[2026-06-19 15:33:37,923][starVLA.training.train_starvla][INFO] - Step 4, Loss: {'train/loss': 0.6850428581237793, 'train/grad_norm_pre_clip': 2.712930679321289, 'timing/data': 0.027093589305877686, 'timing/model': 2.258874362334609, 'learning_rate/qwen_vl_interface': 4.0000000000000003e-07, 'learning_rate/action_model': 4.000000000000001e-06, 'epoch': 0.01}) +[2026-06-19 15:33:46,827][starVLA.training.train_starvla][INFO] - Step 5, Loss: {'train/loss': 0.631145715713501, 'train/grad_norm_pre_clip': 9.425503730773926, 'timing/data': 0.025235697627067566, 'timing/model': 2.261479675769806, 'learning_rate/qwen_vl_interface': 5.000000000000001e-07, 'learning_rate/action_model': 5e-06, 'epoch': 0.01}) +[2026-06-19 15:33:55,623][starVLA.training.train_starvla][INFO] - Step 6, Loss: {'train/loss': 0.6355599761009216, 'train/grad_norm_pre_clip': 4.684020519256592, 'timing/data': 0.021707145497202873, 'timing/model': 2.244677033275366, 'learning_rate/qwen_vl_interface': 6.000000000000001e-07, 'learning_rate/action_model': 6e-06, 'epoch': 0.01}) +[2026-06-19 15:34:04,525][starVLA.training.train_starvla][INFO] - Step 7, Loss: {'train/loss': 0.6129920482635498, 'train/grad_norm_pre_clip': 3.8059709072113037, 'timing/data': 0.025997459888458252, 'timing/model': 2.2511363495141268, 'learning_rate/qwen_vl_interface': 7.000000000000001e-07, 'learning_rate/action_model': 7.000000000000001e-06, 'epoch': 0.01}) +[2026-06-19 15:34:13,426][starVLA.training.train_starvla][INFO] - Step 8, Loss: {'train/loss': 0.4997468590736389, 'train/grad_norm_pre_clip': 6.843386173248291, 'timing/data': 0.02775745838880539, 'timing/model': 2.252879222854972, 'learning_rate/qwen_vl_interface': 8.000000000000001e-07, 'learning_rate/action_model': 8.000000000000001e-06, 'epoch': 0.02}) +[2026-06-19 15:34:22,325][starVLA.training.train_starvla][INFO] - Step 9, Loss: {'train/loss': 0.4775608777999878, 'train/grad_norm_pre_clip': 4.926837921142578, 'timing/data': 0.023146219551563263, 'timing/model': 2.2561705075204372, 'learning_rate/qwen_vl_interface': 9.000000000000001e-07, 'learning_rate/action_model': 9e-06, 'epoch': 0.02}) +[2026-06-19 15:34:31,232][starVLA.training.train_starvla][INFO] - Step 10, Loss: {'train/loss': 0.3784622251987457, 'train/grad_norm_pre_clip': 5.370218276977539, 'timing/data': 0.02745535969734192, 'timing/model': 2.259959375485778, 'learning_rate/qwen_vl_interface': 1.0000000000000002e-06, 'learning_rate/action_model': 1e-05, 'epoch': 0.02}) +[2026-06-19 15:34:40,132][starVLA.training.train_starvla][INFO] - Step 11, Loss: {'train/loss': 0.6051262021064758, 'train/grad_norm_pre_clip': 1.4687607288360596, 'timing/data': 0.0232562106102705, 'timing/model': 2.259601827710867, 'learning_rate/qwen_vl_interface': 1.1e-06, 'learning_rate/action_model': 1.1000000000000001e-05, 'epoch': 0.02}) +[2026-06-19 15:34:49,048][starVLA.training.train_starvla][INFO] - Step 12, Loss: {'train/loss': 0.574729859828949, 'train/grad_norm_pre_clip': 1.4916166067123413, 'timing/data': 0.028059562668204308, 'timing/model': 2.268360098823905, 'learning_rate/qwen_vl_interface': 1.2000000000000002e-06, 'learning_rate/action_model': 1.2e-05, 'epoch': 0.02}) +[2026-06-19 15:34:57,936][starVLA.training.train_starvla][INFO] - Step 13, Loss: {'train/loss': 0.5814814567565918, 'train/grad_norm_pre_clip': 2.0199363231658936, 'timing/data': 0.02331196703016758, 'timing/model': 2.2590834368020296, 'learning_rate/qwen_vl_interface': 1.3e-06, 'learning_rate/action_model': 1.3000000000000001e-05, 'epoch': 0.03}) +[2026-06-19 15:35:06,838][starVLA.training.train_starvla][INFO] - Step 14, Loss: {'train/loss': 0.434064656496048, 'train/grad_norm_pre_clip': 0.1913110464811325, 'timing/data': 0.025850558653473854, 'timing/model': 2.2548608239740133, 'learning_rate/qwen_vl_interface': 1.4000000000000001e-06, 'learning_rate/action_model': 1.4000000000000001e-05, 'epoch': 0.03}) +[2026-06-19 15:35:15,930][starVLA.training.train_starvla][INFO] - Step 15, Loss: {'train/loss': 0.4351596236228943, 'train/grad_norm_pre_clip': 0.2616838812828064, 'timing/data': 0.02329672873020172, 'timing/model': 2.2544103171676397, 'learning_rate/qwen_vl_interface': 1.5e-06, 'learning_rate/action_model': 1.5e-05, 'epoch': 0.03}) +[2026-06-19 15:35:24,835][starVLA.training.train_starvla][INFO] - Step 16, Loss: {'train/loss': 0.5476454496383667, 'train/grad_norm_pre_clip': 2.0602190494537354, 'timing/data': 0.026572491973638535, 'timing/model': 2.2475198544561863, 'learning_rate/qwen_vl_interface': 1.6000000000000001e-06, 'learning_rate/action_model': 1.6000000000000003e-05, 'epoch': 0.03}) +[2026-06-19 15:35:33,739][starVLA.training.train_starvla][INFO] - Step 17, Loss: {'train/loss': 0.549913763999939, 'train/grad_norm_pre_clip': 2.0925629138946533, 'timing/data': 0.02339363656938076, 'timing/model': 2.2657360192388296, 'learning_rate/qwen_vl_interface': 1.7000000000000002e-06, 'learning_rate/action_model': 1.7000000000000003e-05, 'epoch': 0.03}) +[2026-06-19 15:35:42,636][starVLA.training.train_starvla][INFO] - Step 18, Loss: {'train/loss': 0.4336744546890259, 'train/grad_norm_pre_clip': 0.3937773108482361, 'timing/data': 0.023112259805202484, 'timing/model': 2.2588647548109293, 'learning_rate/qwen_vl_interface': 1.8000000000000001e-06, 'learning_rate/action_model': 1.8e-05, 'epoch': 0.04}) +[2026-06-19 15:35:51,539][starVLA.training.train_starvla][INFO] - Step 19, Loss: {'train/loss': 0.37763768434524536, 'train/grad_norm_pre_clip': 0.6796740889549255, 'timing/data': 0.02349247597157955, 'timing/model': 2.2631367668509483, 'learning_rate/qwen_vl_interface': 1.9000000000000002e-06, 'learning_rate/action_model': 1.9e-05, 'epoch': 0.04}) +[2026-06-19 15:36:00,437][starVLA.training.train_starvla][INFO] - Step 20, Loss: {'train/loss': 0.43263015151023865, 'train/grad_norm_pre_clip': 0.20647621154785156, 'timing/data': 0.02509714663028717, 'timing/model': 2.257848309352994, 'learning_rate/qwen_vl_interface': 2.0000000000000003e-06, 'learning_rate/action_model': 2e-05, 'epoch': 0.04}) +[2026-06-19 15:36:09,339][starVLA.training.train_starvla][INFO] - Step 21, Loss: {'train/loss': 0.43247583508491516, 'train/grad_norm_pre_clip': 0.18520694971084595, 'timing/data': 0.023297838866710663, 'timing/model': 2.2616442162543535, 'learning_rate/qwen_vl_interface': 2.1000000000000002e-06, 'learning_rate/action_model': 2.1e-05, 'epoch': 0.04}) +[2026-06-19 15:36:18,240][starVLA.training.train_starvla][INFO] - Step 22, Loss: {'train/loss': 0.58979731798172, 'train/grad_norm_pre_clip': 2.3797757625579834, 'timing/data': 0.023433785885572433, 'timing/model': 2.26207129098475, 'learning_rate/qwen_vl_interface': 2.2e-06, 'learning_rate/action_model': 2.2000000000000003e-05, 'epoch': 0.05}) +[2026-06-19 15:36:27,136][starVLA.training.train_starvla][INFO] - Step 23, Loss: {'train/loss': 0.3364630937576294, 'train/grad_norm_pre_clip': 2.145015001296997, 'timing/data': 0.02430792711675167, 'timing/model': 2.2541636507958174, 'learning_rate/qwen_vl_interface': 2.3000000000000004e-06, 'learning_rate/action_model': 2.3000000000000003e-05, 'epoch': 0.05}) +[2026-06-19 15:36:36,034][starVLA.training.train_starvla][INFO] - Step 24, Loss: {'train/loss': 0.33679935336112976, 'train/grad_norm_pre_clip': 2.140519618988037, 'timing/data': 0.024974867701530457, 'timing/model': 2.250041740015149, 'learning_rate/qwen_vl_interface': 2.4000000000000003e-06, 'learning_rate/action_model': 2.4e-05, 'epoch': 0.05}) +[2026-06-19 15:36:44,935][starVLA.training.train_starvla][INFO] - Step 25, Loss: {'train/loss': 0.5821484327316284, 'train/grad_norm_pre_clip': 2.257704734802246, 'timing/data': 0.022577665746212006, 'timing/model': 2.2545661963522434, 'learning_rate/qwen_vl_interface': 2.5e-06, 'learning_rate/action_model': 2.5e-05, 'epoch': 0.05}) +[2026-06-19 15:36:53,841][starVLA.training.train_starvla][INFO] - Step 26, Loss: {'train/loss': 0.4313305914402008, 'train/grad_norm_pre_clip': 0.42392829060554504, 'timing/data': 0.027734078466892242, 'timing/model': 2.2545901089906693, 'learning_rate/qwen_vl_interface': 2.6e-06, 'learning_rate/action_model': 2.6000000000000002e-05, 'epoch': 0.05}) +[2026-06-19 15:37:02,746][starVLA.training.train_starvla][INFO] - Step 27, Loss: {'train/loss': 0.5317792296409607, 'train/grad_norm_pre_clip': 1.3857035636901855, 'timing/data': 0.02319354936480522, 'timing/model': 2.2624086886644363, 'learning_rate/qwen_vl_interface': 2.7000000000000004e-06, 'learning_rate/action_model': 2.7000000000000002e-05, 'epoch': 0.06}) +[2026-06-19 15:37:11,640][starVLA.training.train_starvla][INFO] - Step 28, Loss: {'train/loss': 0.4811341166496277, 'train/grad_norm_pre_clip': 0.4244998097419739, 'timing/data': 0.024894502013921738, 'timing/model': 2.2577932607382536, 'learning_rate/qwen_vl_interface': 2.8000000000000003e-06, 'learning_rate/action_model': 2.8000000000000003e-05, 'epoch': 0.06}) +[2026-06-19 15:37:20,542][starVLA.training.train_starvla][INFO] - Step 29, Loss: {'train/loss': 0.34654679894447327, 'train/grad_norm_pre_clip': 2.7458317279815674, 'timing/data': 0.022969212383031845, 'timing/model': 2.2610884830355644, 'learning_rate/qwen_vl_interface': 2.9e-06, 'learning_rate/action_model': 2.9e-05, 'epoch': 0.06}) +[2026-06-19 15:37:29,437][starVLA.training.train_starvla][INFO] - Step 30, Loss: {'train/loss': 0.34564629197120667, 'train/grad_norm_pre_clip': 2.7625250816345215, 'timing/data': 0.02542291209101677, 'timing/model': 2.251186903566122, 'learning_rate/qwen_vl_interface': 3e-06, 'learning_rate/action_model': 3e-05, 'epoch': 0.06}) +[2026-06-19 15:37:38,339][starVLA.training.train_starvla][INFO] - Step 31, Loss: {'train/loss': 0.574243426322937, 'train/grad_norm_pre_clip': 2.654330015182495, 'timing/data': 0.02291986159980297, 'timing/model': 2.259325046092272, 'learning_rate/qwen_vl_interface': 3.1000000000000004e-06, 'learning_rate/action_model': 3.1e-05, 'epoch': 0.06}) +[2026-06-19 15:37:47,241][starVLA.training.train_starvla][INFO] - Step 32, Loss: {'train/loss': 0.6752209663391113, 'train/grad_norm_pre_clip': 3.9440529346466064, 'timing/data': 0.025251615792512894, 'timing/model': 2.2584387864917517, 'learning_rate/qwen_vl_interface': 3.2000000000000003e-06, 'learning_rate/action_model': 3.2000000000000005e-05, 'epoch': 0.07}) +[2026-06-19 15:37:56,134][starVLA.training.train_starvla][INFO] - Step 33, Loss: {'train/loss': 0.4286845922470093, 'train/grad_norm_pre_clip': 0.9891082644462585, 'timing/data': 0.023576844483613968, 'timing/model': 2.2597753293812275, 'learning_rate/qwen_vl_interface': 3.3000000000000006e-06, 'learning_rate/action_model': 3.3e-05, 'epoch': 0.07}) +[2026-06-19 15:38:05,035][starVLA.training.train_starvla][INFO] - Step 34, Loss: {'train/loss': 0.47788673639297485, 'train/grad_norm_pre_clip': 0.7628512978553772, 'timing/data': 0.025381583720445633, 'timing/model': 2.256828121840954, 'learning_rate/qwen_vl_interface': 3.4000000000000005e-06, 'learning_rate/action_model': 3.4000000000000007e-05, 'epoch': 0.07}) +[2026-06-19 15:38:13,937][starVLA.training.train_starvla][INFO] - Step 35, Loss: {'train/loss': 0.5597577095031738, 'train/grad_norm_pre_clip': 2.106518507003784, 'timing/data': 0.02318442054092884, 'timing/model': 2.258608853444457, 'learning_rate/qwen_vl_interface': 3.5e-06, 'learning_rate/action_model': 3.5e-05, 'epoch': 0.07}) +[2026-06-19 15:38:22,837][starVLA.training.train_starvla][INFO] - Step 36, Loss: {'train/loss': 0.4748092293739319, 'train/grad_norm_pre_clip': 1.5869379043579102, 'timing/data': 0.024646524339914322, 'timing/model': 2.25376246124506, 'learning_rate/qwen_vl_interface': 3.6000000000000003e-06, 'learning_rate/action_model': 3.6e-05, 'epoch': 0.07}) +[2026-06-19 15:38:31,737][starVLA.training.train_starvla][INFO] - Step 37, Loss: {'train/loss': 0.5123044848442078, 'train/grad_norm_pre_clip': 1.0200375318527222, 'timing/data': 0.022513827309012413, 'timing/model': 2.259210307151079, 'learning_rate/qwen_vl_interface': 3.7e-06, 'learning_rate/action_model': 3.7e-05, 'epoch': 0.08}) +[2026-06-19 15:38:40,638][starVLA.training.train_starvla][INFO] - Step 38, Loss: {'train/loss': 0.5393093824386597, 'train/grad_norm_pre_clip': 2.8042008876800537, 'timing/data': 0.025038519874215126, 'timing/model': 2.2536362092942, 'learning_rate/qwen_vl_interface': 3.8000000000000005e-06, 'learning_rate/action_model': 3.8e-05, 'epoch': 0.08}) +[2026-06-19 15:38:49,535][starVLA.training.train_starvla][INFO] - Step 39, Loss: {'train/loss': 0.4548751413822174, 'train/grad_norm_pre_clip': 4.540852069854736, 'timing/data': 0.022667912766337395, 'timing/model': 2.2571836095303297, 'learning_rate/qwen_vl_interface': 3.900000000000001e-06, 'learning_rate/action_model': 3.9000000000000006e-05, 'epoch': 0.08}) +[2026-06-19 15:38:58,440][starVLA.training.train_starvla][INFO] - Step 40, Loss: {'train/loss': 0.4736689627170563, 'train/grad_norm_pre_clip': 2.9519314765930176, 'timing/data': 0.02615506388247013, 'timing/model': 2.252634394913912, 'learning_rate/qwen_vl_interface': 4.000000000000001e-06, 'learning_rate/action_model': 4e-05, 'epoch': 0.08}) +[2026-06-19 15:39:07,337][starVLA.training.train_starvla][INFO] - Step 41, Loss: {'train/loss': 0.5600329637527466, 'train/grad_norm_pre_clip': 8.264060974121094, 'timing/data': 0.02324017696082592, 'timing/model': 2.254600090906024, 'learning_rate/qwen_vl_interface': 4.1e-06, 'learning_rate/action_model': 4.1e-05, 'epoch': 0.08}) +[2026-06-19 15:39:16,239][starVLA.training.train_starvla][INFO] - Step 42, Loss: {'train/loss': 0.43262606859207153, 'train/grad_norm_pre_clip': 3.048243999481201, 'timing/data': 0.025238830596208572, 'timing/model': 2.255137350410223, 'learning_rate/qwen_vl_interface': 4.2000000000000004e-06, 'learning_rate/action_model': 4.2e-05, 'epoch': 0.09}) +[2026-06-19 15:39:25,136][starVLA.training.train_starvla][INFO] - Step 43, Loss: {'train/loss': 0.34700343012809753, 'train/grad_norm_pre_clip': 3.5963997840881348, 'timing/data': 0.02676745131611824, 'timing/model': 2.2515796199440956, 'learning_rate/qwen_vl_interface': 4.3e-06, 'learning_rate/action_model': 4.3e-05, 'epoch': 0.09}) +[2026-06-19 15:39:34,037][starVLA.training.train_starvla][INFO] - Step 44, Loss: {'train/loss': 0.547275185585022, 'train/grad_norm_pre_clip': 9.033526420593262, 'timing/data': 0.024945536628365517, 'timing/model': 2.2607589177787304, 'learning_rate/qwen_vl_interface': 4.4e-06, 'learning_rate/action_model': 4.4000000000000006e-05, 'epoch': 0.09}) +[2026-06-19 15:39:42,937][starVLA.training.train_starvla][INFO] - Step 45, Loss: {'train/loss': 0.3164074122905731, 'train/grad_norm_pre_clip': 7.686972618103027, 'timing/data': 0.023438584059476852, 'timing/model': 2.260577468201518, 'learning_rate/qwen_vl_interface': 4.5e-06, 'learning_rate/action_model': 4.5e-05, 'epoch': 0.09}) +[2026-06-19 15:39:51,832][starVLA.training.train_starvla][INFO] - Step 46, Loss: {'train/loss': 0.536998450756073, 'train/grad_norm_pre_clip': 5.586982250213623, 'timing/data': 0.02387423627078533, 'timing/model': 2.254136897623539, 'learning_rate/qwen_vl_interface': 4.600000000000001e-06, 'learning_rate/action_model': 4.600000000000001e-05, 'epoch': 0.09}) +[2026-06-19 15:40:00,731][starVLA.training.train_starvla][INFO] - Step 47, Loss: {'train/loss': 0.3424138128757477, 'train/grad_norm_pre_clip': 4.131486892700195, 'timing/data': 0.02366884984076023, 'timing/model': 2.2534201610833406, 'learning_rate/qwen_vl_interface': 4.7e-06, 'learning_rate/action_model': 4.7e-05, 'epoch': 0.1}) +[2026-06-19 15:40:09,637][starVLA.training.train_starvla][INFO] - Step 48, Loss: {'train/loss': 0.22585999965667725, 'train/grad_norm_pre_clip': 8.147828102111816, 'timing/data': 0.024348512291908264, 'timing/model': 2.260107483714819, 'learning_rate/qwen_vl_interface': 4.800000000000001e-06, 'learning_rate/action_model': 4.8e-05, 'epoch': 0.1}) +[2026-06-19 15:40:18,534][starVLA.training.train_starvla][INFO] - Step 49, Loss: {'train/loss': 0.5078151822090149, 'train/grad_norm_pre_clip': 16.503520965576172, 'timing/data': 0.023803098127245903, 'timing/model': 2.255965195596218, 'learning_rate/qwen_vl_interface': 4.9000000000000005e-06, 'learning_rate/action_model': 4.9e-05, 'epoch': 0.1}) +[2026-06-19 15:40:27,439][starVLA.training.train_starvla][INFO] - Step 50, Loss: {'train/loss': 0.29411354660987854, 'train/grad_norm_pre_clip': 5.612155914306641, 'timing/data': 0.024374360218644142, 'timing/model': 2.2624345794320107, 'learning_rate/qwen_vl_interface': 5e-06, 'learning_rate/action_model': 5e-05, 'epoch': 0.1}) +[2026-06-19 15:40:36,335][starVLA.training.train_starvla][INFO] - Step 51, Loss: {'train/loss': 0.5689382553100586, 'train/grad_norm_pre_clip': 12.196976661682129, 'timing/data': 0.022948408499360085, 'timing/model': 2.2587036564946175, 'learning_rate/qwen_vl_interface': 5.1e-06, 'learning_rate/action_model': 5.1000000000000006e-05, 'epoch': 0.1}) +[2026-06-19 15:40:45,237][starVLA.training.train_starvla][INFO] - Step 52, Loss: {'train/loss': 0.500743567943573, 'train/grad_norm_pre_clip': 6.350891590118408, 'timing/data': 0.022954769432544708, 'timing/model': 2.2574007380753756, 'learning_rate/qwen_vl_interface': 5.2e-06, 'learning_rate/action_model': 5.2000000000000004e-05, 'epoch': 0.11}) +[2026-06-19 15:40:54,140][starVLA.training.train_starvla][INFO] - Step 53, Loss: {'train/loss': 0.4424315392971039, 'train/grad_norm_pre_clip': 9.57518196105957, 'timing/data': 0.02442334219813347, 'timing/model': 2.258442861959338, 'learning_rate/qwen_vl_interface': 5.300000000000001e-06, 'learning_rate/action_model': 5.300000000000001e-05, 'epoch': 0.11}) +[2026-06-19 15:41:03,040][starVLA.training.train_starvla][INFO] - Step 54, Loss: {'train/loss': 0.4245353639125824, 'train/grad_norm_pre_clip': 13.736879348754883, 'timing/data': 0.02578243613243103, 'timing/model': 2.2563571091741323, 'learning_rate/qwen_vl_interface': 5.400000000000001e-06, 'learning_rate/action_model': 5.4000000000000005e-05, 'epoch': 0.11}) +[2026-06-19 15:41:11,939][starVLA.training.train_starvla][INFO] - Step 55, Loss: {'train/loss': 0.41599345207214355, 'train/grad_norm_pre_clip': 22.48100471496582, 'timing/data': 0.02658930793404579, 'timing/model': 2.256739541888237, 'learning_rate/qwen_vl_interface': 5.500000000000001e-06, 'learning_rate/action_model': 5.500000000000001e-05, 'epoch': 0.11}) +[2026-06-19 15:41:20,842][starVLA.training.train_starvla][INFO] - Step 56, Loss: {'train/loss': 0.504814863204956, 'train/grad_norm_pre_clip': 15.842777252197266, 'timing/data': 0.022355694323778152, 'timing/model': 2.2626805007457733, 'learning_rate/qwen_vl_interface': 5.600000000000001e-06, 'learning_rate/action_model': 5.6000000000000006e-05, 'epoch': 0.12}) +[2026-06-19 15:41:29,934][starVLA.training.train_starvla][INFO] - Step 57, Loss: {'train/loss': 0.3421284854412079, 'train/grad_norm_pre_clip': 11.772478103637695, 'timing/data': 0.02336641401052475, 'timing/model': 2.253676975145936, 'learning_rate/qwen_vl_interface': 5.7e-06, 'learning_rate/action_model': 5.6999999999999996e-05, 'epoch': 0.12}) +[2026-06-19 15:41:38,835][starVLA.training.train_starvla][INFO] - Step 58, Loss: {'train/loss': 0.8125484585762024, 'train/grad_norm_pre_clip': 58.09130859375, 'timing/data': 0.025685708969831467, 'timing/model': 2.252162467688322, 'learning_rate/qwen_vl_interface': 5.8e-06, 'learning_rate/action_model': 5.8e-05, 'epoch': 0.12}) +[2026-06-19 15:41:47,739][starVLA.training.train_starvla][INFO] - Step 59, Loss: {'train/loss': 0.5980346202850342, 'train/grad_norm_pre_clip': 42.459651947021484, 'timing/data': 0.025003256276249886, 'timing/model': 2.2546266485005617, 'learning_rate/qwen_vl_interface': 5.9e-06, 'learning_rate/action_model': 5.9e-05, 'epoch': 0.12}) +[2026-06-19 15:41:56,637][starVLA.training.train_starvla][INFO] - Step 60, Loss: {'train/loss': 0.4861774742603302, 'train/grad_norm_pre_clip': 34.58059310913086, 'timing/data': 0.022810552269220352, 'timing/model': 2.2581412829458714, 'learning_rate/qwen_vl_interface': 6e-06, 'learning_rate/action_model': 6e-05, 'epoch': 0.12}) +[2026-06-19 15:42:05,543][starVLA.training.train_starvla][INFO] - Step 61, Loss: {'train/loss': 0.2884911000728607, 'train/grad_norm_pre_clip': 6.389272689819336, 'timing/data': 0.025707878172397614, 'timing/model': 2.2607114389538765, 'learning_rate/qwen_vl_interface': 6.1e-06, 'learning_rate/action_model': 6.1e-05, 'epoch': 0.13}) +[2026-06-19 15:42:14,441][starVLA.training.train_starvla][INFO] - Step 62, Loss: {'train/loss': 0.5471611022949219, 'train/grad_norm_pre_clip': 13.952592849731445, 'timing/data': 0.02308131568133831, 'timing/model': 2.258032951503992, 'learning_rate/qwen_vl_interface': 6.200000000000001e-06, 'learning_rate/action_model': 6.2e-05, 'epoch': 0.13}) +[2026-06-19 15:42:23,339][starVLA.training.train_starvla][INFO] - Step 63, Loss: {'train/loss': 0.3655402958393097, 'train/grad_norm_pre_clip': 22.896102905273438, 'timing/data': 0.026015853509306908, 'timing/model': 2.2571964357048273, 'learning_rate/qwen_vl_interface': 6.300000000000001e-06, 'learning_rate/action_model': 6.3e-05, 'epoch': 0.13}) +[2026-06-19 15:42:32,243][starVLA.training.train_starvla][INFO] - Step 64, Loss: {'train/loss': 0.38208407163619995, 'train/grad_norm_pre_clip': 24.342575073242188, 'timing/data': 0.0235119741410017, 'timing/model': 2.260997327044606, 'learning_rate/qwen_vl_interface': 6.4000000000000006e-06, 'learning_rate/action_model': 6.400000000000001e-05, 'epoch': 0.13}) +[2026-06-19 15:42:41,235][starVLA.training.train_starvla][INFO] - Step 65, Loss: {'train/loss': 0.41074517369270325, 'train/grad_norm_pre_clip': 12.494775772094727, 'timing/data': 0.02435818314552307, 'timing/model': 2.2542289197444916, 'learning_rate/qwen_vl_interface': 6.5000000000000004e-06, 'learning_rate/action_model': 6.500000000000001e-05, 'epoch': 0.13}) +[2026-06-19 15:42:50,135][starVLA.training.train_starvla][INFO] - Step 66, Loss: {'train/loss': 0.4084881842136383, 'train/grad_norm_pre_clip': 9.657155990600586, 'timing/data': 0.023125186562538147, 'timing/model': 2.2554241698235273, 'learning_rate/qwen_vl_interface': 6.600000000000001e-06, 'learning_rate/action_model': 6.6e-05, 'epoch': 0.14}) +[2026-06-19 15:42:59,036][starVLA.training.train_starvla][INFO] - Step 67, Loss: {'train/loss': 0.3180418610572815, 'train/grad_norm_pre_clip': 9.524004936218262, 'timing/data': 0.024639850482344627, 'timing/model': 2.255958618596196, 'learning_rate/qwen_vl_interface': 6.700000000000001e-06, 'learning_rate/action_model': 6.7e-05, 'epoch': 0.14}) +[2026-06-19 15:43:07,933][starVLA.training.train_starvla][INFO] - Step 68, Loss: {'train/loss': 0.44835221767425537, 'train/grad_norm_pre_clip': 12.33663558959961, 'timing/data': 0.02275681123137474, 'timing/model': 2.2564999666064978, 'learning_rate/qwen_vl_interface': 6.800000000000001e-06, 'learning_rate/action_model': 6.800000000000001e-05, 'epoch': 0.14}) +[2026-06-19 15:43:16,840][starVLA.training.train_starvla][INFO] - Step 69, Loss: {'train/loss': 0.3120841681957245, 'train/grad_norm_pre_clip': 16.220043182373047, 'timing/data': 0.024356592446565628, 'timing/model': 2.257504340261221, 'learning_rate/qwen_vl_interface': 6.9e-06, 'learning_rate/action_model': 6.9e-05, 'epoch': 0.14}) +[2026-06-19 15:43:25,736][starVLA.training.train_starvla][INFO] - Step 70, Loss: {'train/loss': 0.17795483767986298, 'train/grad_norm_pre_clip': 5.064012050628662, 'timing/data': 0.02284671925008297, 'timing/model': 2.2586924098432064, 'learning_rate/qwen_vl_interface': 7e-06, 'learning_rate/action_model': 7e-05, 'epoch': 0.14}) +[2026-06-19 15:43:34,641][starVLA.training.train_starvla][INFO] - Step 71, Loss: {'train/loss': 0.19550344347953796, 'train/grad_norm_pre_clip': 3.222506046295166, 'timing/data': 0.02532576397061348, 'timing/model': 2.2579908445477486, 'learning_rate/qwen_vl_interface': 7.100000000000001e-06, 'learning_rate/action_model': 7.1e-05, 'epoch': 0.15}) +[2026-06-19 15:43:43,535][starVLA.training.train_starvla][INFO] - Step 72, Loss: {'train/loss': 0.39800721406936646, 'train/grad_norm_pre_clip': 11.090755462646484, 'timing/data': 0.023112287744879723, 'timing/model': 2.25242305919528, 'learning_rate/qwen_vl_interface': 7.2000000000000005e-06, 'learning_rate/action_model': 7.2e-05, 'epoch': 0.15}) +[2026-06-19 15:43:52,435][starVLA.training.train_starvla][INFO] - Step 73, Loss: {'train/loss': 0.427999347448349, 'train/grad_norm_pre_clip': 7.981045246124268, 'timing/data': 0.025638828054070473, 'timing/model': 2.2553723994642496, 'learning_rate/qwen_vl_interface': 7.3e-06, 'learning_rate/action_model': 7.3e-05, 'epoch': 0.15}) +[2026-06-19 15:44:01,337][starVLA.training.train_starvla][INFO] - Step 74, Loss: {'train/loss': 0.41020992398262024, 'train/grad_norm_pre_clip': 9.842045783996582, 'timing/data': 0.02325427532196045, 'timing/model': 2.258063605055213, 'learning_rate/qwen_vl_interface': 7.4e-06, 'learning_rate/action_model': 7.4e-05, 'epoch': 0.15}) +[2026-06-19 15:44:10,237][starVLA.training.train_starvla][INFO] - Step 75, Loss: {'train/loss': 0.46814265847206116, 'train/grad_norm_pre_clip': 17.513540267944336, 'timing/data': 0.02514273300766945, 'timing/model': 2.2613710816949606, 'learning_rate/qwen_vl_interface': 7.500000000000001e-06, 'learning_rate/action_model': 7.500000000000001e-05, 'epoch': 0.15}) +[2026-06-19 15:44:19,138][starVLA.training.train_starvla][INFO] - Step 76, Loss: {'train/loss': 0.4082649052143097, 'train/grad_norm_pre_clip': 12.701233863830566, 'timing/data': 0.022903408855199814, 'timing/model': 2.264679016545415, 'learning_rate/qwen_vl_interface': 7.600000000000001e-06, 'learning_rate/action_model': 7.6e-05, 'epoch': 0.16}) +[2026-06-19 15:44:28,038][starVLA.training.train_starvla][INFO] - Step 77, Loss: {'train/loss': 0.3956528604030609, 'train/grad_norm_pre_clip': 4.9382243156433105, 'timing/data': 0.025522852316498756, 'timing/model': 2.257892668247223, 'learning_rate/qwen_vl_interface': 7.7e-06, 'learning_rate/action_model': 7.7e-05, 'epoch': 0.16}) +[2026-06-19 15:44:36,939][starVLA.training.train_starvla][INFO] - Step 78, Loss: {'train/loss': 0.4617392420768738, 'train/grad_norm_pre_clip': 4.977255821228027, 'timing/data': 0.023699792101979256, 'timing/model': 2.257911991328001, 'learning_rate/qwen_vl_interface': 7.800000000000002e-06, 'learning_rate/action_model': 7.800000000000001e-05, 'epoch': 0.16}) +[2026-06-19 15:44:45,834][starVLA.training.train_starvla][INFO] - Step 79, Loss: {'train/loss': 0.2904415726661682, 'train/grad_norm_pre_clip': 7.751798152923584, 'timing/data': 0.02477557770907879, 'timing/model': 2.255737069994211, 'learning_rate/qwen_vl_interface': 7.9e-06, 'learning_rate/action_model': 7.900000000000001e-05, 'epoch': 0.16}) +[2026-06-19 15:44:54,738][starVLA.training.train_starvla][INFO] - Step 80, Loss: {'train/loss': 0.2999514937400818, 'train/grad_norm_pre_clip': 11.956693649291992, 'timing/data': 0.023061875253915787, 'timing/model': 2.260291423648596, 'learning_rate/qwen_vl_interface': 8.000000000000001e-06, 'learning_rate/action_model': 8e-05, 'epoch': 0.16}) +[2026-06-19 15:45:03,639][starVLA.training.train_starvla][INFO] - Step 81, Loss: {'train/loss': 0.3936704099178314, 'train/grad_norm_pre_clip': 14.184341430664062, 'timing/data': 0.024180015549063683, 'timing/model': 2.255974691361189, 'learning_rate/qwen_vl_interface': 8.1e-06, 'learning_rate/action_model': 8.1e-05, 'epoch': 0.17}) +[2026-06-19 15:45:12,537][starVLA.training.train_starvla][INFO] - Step 82, Loss: {'train/loss': 0.3642851710319519, 'train/grad_norm_pre_clip': 11.318883895874023, 'timing/data': 0.02266860194504261, 'timing/model': 2.25802587158978, 'learning_rate/qwen_vl_interface': 8.2e-06, 'learning_rate/action_model': 8.2e-05, 'epoch': 0.17}) +[2026-06-19 15:45:21,435][starVLA.training.train_starvla][INFO] - Step 83, Loss: {'train/loss': 0.4221312999725342, 'train/grad_norm_pre_clip': 10.991941452026367, 'timing/data': 0.02536277286708355, 'timing/model': 2.2540348805487156, 'learning_rate/qwen_vl_interface': 8.3e-06, 'learning_rate/action_model': 8.3e-05, 'epoch': 0.17}) +[2026-06-19 15:45:30,339][starVLA.training.train_starvla][INFO] - Step 84, Loss: {'train/loss': 0.4050699770450592, 'train/grad_norm_pre_clip': 10.126909255981445, 'timing/data': 0.02286878041923046, 'timing/model': 2.2572621181607246, 'learning_rate/qwen_vl_interface': 8.400000000000001e-06, 'learning_rate/action_model': 8.4e-05, 'epoch': 0.17}) +[2026-06-19 15:45:39,241][starVLA.training.train_starvla][INFO] - Step 85, Loss: {'train/loss': 0.37714388966560364, 'train/grad_norm_pre_clip': 11.936652183532715, 'timing/data': 0.025415582582354546, 'timing/model': 2.25617016479373, 'learning_rate/qwen_vl_interface': 8.5e-06, 'learning_rate/action_model': 8.5e-05, 'epoch': 0.17}) +[2026-06-19 15:45:48,142][starVLA.training.train_starvla][INFO] - Step 86, Loss: {'train/loss': 0.6872298121452332, 'train/grad_norm_pre_clip': 10.607147216796875, 'timing/data': 0.021923372521996498, 'timing/model': 2.2611496038734913, 'learning_rate/qwen_vl_interface': 8.6e-06, 'learning_rate/action_model': 8.6e-05, 'epoch': 0.18}) +[2026-06-19 15:45:57,040][starVLA.training.train_starvla][INFO] - Step 87, Loss: {'train/loss': 0.3955375850200653, 'train/grad_norm_pre_clip': 8.903698921203613, 'timing/data': 0.024191653355956078, 'timing/model': 2.255447532981634, 'learning_rate/qwen_vl_interface': 8.700000000000001e-06, 'learning_rate/action_model': 8.7e-05, 'epoch': 0.18}) +[2026-06-19 15:46:05,938][starVLA.training.train_starvla][INFO] - Step 88, Loss: {'train/loss': 0.46702349185943604, 'train/grad_norm_pre_clip': 3.9244749546051025, 'timing/data': 0.021301528438925743, 'timing/model': 2.2585099767893553, 'learning_rate/qwen_vl_interface': 8.8e-06, 'learning_rate/action_model': 8.800000000000001e-05, 'epoch': 0.18}) +[2026-06-19 15:46:14,842][starVLA.training.train_starvla][INFO] - Step 89, Loss: {'train/loss': 0.515399158000946, 'train/grad_norm_pre_clip': 10.655876159667969, 'timing/data': 0.024167336523532867, 'timing/model': 2.2505815029144287, 'learning_rate/qwen_vl_interface': 8.900000000000001e-06, 'learning_rate/action_model': 8.900000000000001e-05, 'epoch': 0.18}) +[2026-06-19 15:46:23,741][starVLA.training.train_starvla][INFO] - Step 90, Loss: {'train/loss': 0.5327528119087219, 'train/grad_norm_pre_clip': 14.144841194152832, 'timing/data': 0.021285241469740868, 'timing/model': 2.2574937772005796, 'learning_rate/qwen_vl_interface': 9e-06, 'learning_rate/action_model': 9e-05, 'epoch': 0.18}) +[2026-06-19 15:46:32,635][starVLA.training.train_starvla][INFO] - Step 91, Loss: {'train/loss': 0.49778124690055847, 'train/grad_norm_pre_clip': 6.309598445892334, 'timing/data': 0.02505272813141346, 'timing/model': 2.2489318661391735, 'learning_rate/qwen_vl_interface': 9.100000000000001e-06, 'learning_rate/action_model': 9.1e-05, 'epoch': 0.19}) +[2026-06-19 15:46:41,540][starVLA.training.train_starvla][INFO] - Step 92, Loss: {'train/loss': 0.4059441089630127, 'train/grad_norm_pre_clip': 4.5190582275390625, 'timing/data': 0.02240157499909401, 'timing/model': 2.260097123682499, 'learning_rate/qwen_vl_interface': 9.200000000000002e-06, 'learning_rate/action_model': 9.200000000000001e-05, 'epoch': 0.19}) +[2026-06-19 15:46:50,440][starVLA.training.train_starvla][INFO] - Step 93, Loss: {'train/loss': 0.38436073064804077, 'train/grad_norm_pre_clip': 8.583765029907227, 'timing/data': 0.024022769182920456, 'timing/model': 2.2585365436971188, 'learning_rate/qwen_vl_interface': 9.3e-06, 'learning_rate/action_model': 9.300000000000001e-05, 'epoch': 0.19}) +[2026-06-19 15:46:59,335][starVLA.training.train_starvla][INFO] - Step 94, Loss: {'train/loss': 0.2820519506931305, 'train/grad_norm_pre_clip': 10.592371940612793, 'timing/data': 0.022682402282953262, 'timing/model': 2.2545564398169518, 'learning_rate/qwen_vl_interface': 9.4e-06, 'learning_rate/action_model': 9.4e-05, 'epoch': 0.19}) +[2026-06-19 15:47:08,239][starVLA.training.train_starvla][INFO] - Step 95, Loss: {'train/loss': 0.10186849534511566, 'train/grad_norm_pre_clip': 6.314637660980225, 'timing/data': 0.03530353121459484, 'timing/model': 2.2426743172109127, 'learning_rate/qwen_vl_interface': 9.5e-06, 'learning_rate/action_model': 9.5e-05, 'epoch': 0.2}) +[2026-06-19 15:47:17,142][starVLA.training.train_starvla][INFO] - Step 96, Loss: {'train/loss': 0.3539377748966217, 'train/grad_norm_pre_clip': 16.13762855529785, 'timing/data': 0.00938127189874649, 'timing/model': 2.2795571740716696, 'learning_rate/qwen_vl_interface': 9.600000000000001e-06, 'learning_rate/action_model': 9.6e-05, 'epoch': 0.2}) +[2026-06-19 15:47:26,037][starVLA.training.train_starvla][INFO] - Step 97, Loss: {'train/loss': 0.6481843590736389, 'train/grad_norm_pre_clip': 14.937919616699219, 'timing/data': 0.01923169195652008, 'timing/model': 2.2665000706911087, 'learning_rate/qwen_vl_interface': 9.7e-06, 'learning_rate/action_model': 9.7e-05, 'epoch': 0.2}) +[2026-06-19 15:47:34,932][starVLA.training.train_starvla][INFO] - Step 98, Loss: {'train/loss': 0.500943124294281, 'train/grad_norm_pre_clip': 24.16336441040039, 'timing/data': 0.021546365693211555, 'timing/model': 2.255038805305958, 'learning_rate/qwen_vl_interface': 9.800000000000001e-06, 'learning_rate/action_model': 9.8e-05, 'epoch': 0.2}) +[2026-06-19 15:47:44,037][starVLA.training.train_starvla][INFO] - Step 99, Loss: {'train/loss': 0.3457077443599701, 'train/grad_norm_pre_clip': 18.229949951171875, 'timing/data': 0.022923031821846962, 'timing/model': 2.4560511633753777, 'learning_rate/qwen_vl_interface': 9.9e-06, 'learning_rate/action_model': 9.900000000000001e-05, 'epoch': 0.2}) +[2026-06-19 15:47:52,935][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 100 +[2026-06-19 15:56:05,722][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 100 +[2026-06-19 15:56:05,723][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 100 +[2026-06-19 15:56:06,820][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_openvla/include -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_openvla/include -fPIC -c /tmp/tmpdifyl1xh/test.c -o /tmp/tmpdifyl1xh/test.o +[2026-06-19 15:56:06,849][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat /tmp/tmpdifyl1xh/test.o -laio -o /tmp/tmpdifyl1xh/a.out +[2026-06-19 15:56:07,298][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_openvla/include -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_openvla/include -fPIC -c /tmp/tmpb4rwd_hm/test.c -o /tmp/tmpb4rwd_hm/test.o +[2026-06-19 15:56:07,326][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat /tmp/tmpb4rwd_hm/test.o -L/usr/local/cuda -L/usr/local/cuda/lib64 -lcufile -o /tmp/tmpb4rwd_hm/a.out +[2026-06-19 15:56:07,388][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_openvla/include -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_openvla/include -fPIC -c /tmp/tmpfl0tnmn0/test.c -o /tmp/tmpfl0tnmn0/test.o +[2026-06-19 15:56:07,405][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat /tmp/tmpfl0tnmn0/test.o -laio -o /tmp/tmpfl0tnmn0/a.out +[2026-06-19 16:04:57,624][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 100 +[2026-06-19 16:04:57,627][starVLA.training.train_starvla][INFO] - Step 100, Loss: {'train/loss': 0.5159574747085571, 'train/grad_norm_pre_clip': 14.607379913330078, 'eval/loss': 0.42659515142440796, 'eval/latency_0/loss': 0.38316306471824646, 'eval/latency_1/loss': 0.39301952719688416, 'eval/latency_2/loss': 0.45331746339797974, 'eval/latency_3/loss': 0.42638900876045227, 'eval/latency_4/loss': 0.4730066955089569, 'eval/flappy/loss': 0.42659515142440796, 'eval/flappy/latency_0/loss': 0.38316306471824646, 'eval/flappy/latency_1/loss': 0.39301952719688416, 'eval/flappy/latency_2/loss': 0.45331746339797974, 'eval/flappy/latency_3/loss': 0.42638900876045227, 'eval/flappy/latency_4/loss': 0.4730066955089569, 'eval/flap_precision': 0.0, 'eval/flap_recall': 0.0, 'eval/flap_f1': 0.0, 'eval/flap_rate_model': 0.0, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.0, 'eval/latency_0/flap_precision': 0.0, 'eval/latency_0/flap_recall': 0.0, 'eval/latency_0/flap_f1': 0.0, 'eval/latency_0/flap_rate_model': 0.0, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.0, 'timing/data': 0.023137357085943222, 'timing/model': 2.254923151805997, 'learning_rate/qwen_vl_interface': 1e-05, 'learning_rate/action_model': 0.0001, 'epoch': 0.21}) +[2026-06-19 16:04:57,627][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_100_state +[2026-06-19 16:04:57,659][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-19 16:05:03,550][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_100_state/model.safetensors +[2026-06-19 16:05:12,108][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_100_state/optimizer.bin +[2026-06-19 16:05:12,110][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_100_state/sampler.bin +[2026-06-19 16:05:12,110][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_100_state/sampler_1.bin +[2026-06-19 16:05:12,112][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_100_state/random_states_0.pkl +[2026-06-19 16:05:12,117][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 100 +[2026-06-19 16:05:12,117][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-19 16:05:12,131][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-19 16:05:22,816][starVLA.training.train_starvla][INFO] - Step 101, Loss: {'train/loss': 0.5338240265846252, 'train/grad_norm_pre_clip': 6.357398986816406, 'timing/data': 0.03007565066218376, 'timing/model': 2.7643566988408566, 'learning_rate/qwen_vl_interface': 9.999999023727212e-06, 'learning_rate/action_model': 9.99999902372721e-05, 'epoch': 0.21}) +[2026-06-19 16:05:33,322][starVLA.training.train_starvla][INFO] - Step 102, Loss: {'train/loss': 0.3617720305919647, 'train/grad_norm_pre_clip': 4.542844772338867, 'timing/data': 0.047377027571201324, 'timing/model': 2.639927299693227, 'learning_rate/qwen_vl_interface': 9.999996094909247e-06, 'learning_rate/action_model': 9.999996094909248e-05, 'epoch': 0.21}) +[2026-06-19 16:05:44,018][starVLA.training.train_starvla][INFO] - Step 103, Loss: {'train/loss': 0.5649276375770569, 'train/grad_norm_pre_clip': 4.071011066436768, 'timing/data': 0.03129131905734539, 'timing/model': 2.7430212032049894, 'learning_rate/qwen_vl_interface': 9.999991213547314e-06, 'learning_rate/action_model': 9.999991213547313e-05, 'epoch': 0.21}) +[2026-06-19 16:05:54,630][starVLA.training.train_starvla][INFO] - Step 104, Loss: {'train/loss': 0.407948762178421, 'train/grad_norm_pre_clip': 2.8354392051696777, 'timing/data': 0.026727547869086266, 'timing/model': 2.6790901105850935, 'learning_rate/qwen_vl_interface': 9.999984379643413e-06, 'learning_rate/action_model': 9.999984379643412e-05, 'epoch': 0.21}) +[2026-06-19 16:06:05,330][starVLA.training.train_starvla][INFO] - Step 105, Loss: {'train/loss': 0.24092106521129608, 'train/grad_norm_pre_clip': 2.7314846515655518, 'timing/data': 0.024505412206053734, 'timing/model': 2.6778330076485872, 'learning_rate/qwen_vl_interface': 9.999975593200357e-06, 'learning_rate/action_model': 9.999975593200357e-05, 'epoch': 0.22}) +[2026-06-19 16:06:15,856][starVLA.training.train_starvla][INFO] - Step 106, Loss: {'train/loss': 0.17661114037036896, 'train/grad_norm_pre_clip': 7.515818119049072, 'timing/data': 0.02381967194378376, 'timing/model': 2.6538405679166317, 'learning_rate/qwen_vl_interface': 9.999964854221756e-06, 'learning_rate/action_model': 9.999964854221754e-05, 'epoch': 0.22}) +[2026-06-19 16:06:26,714][starVLA.training.train_starvla][INFO] - Step 107, Loss: {'train/loss': 0.3097705841064453, 'train/grad_norm_pre_clip': 11.684271812438965, 'timing/data': 0.02511615864932537, 'timing/model': 2.748378224670887, 'learning_rate/qwen_vl_interface': 9.999952162712027e-06, 'learning_rate/action_model': 9.999952162712026e-05, 'epoch': 0.22}) +[2026-06-19 16:06:37,406][starVLA.training.train_starvla][INFO] - Step 108, Loss: {'train/loss': 0.2130942940711975, 'train/grad_norm_pre_clip': 6.347006797790527, 'timing/data': 0.02535402402281761, 'timing/model': 2.7386925127357244, 'learning_rate/qwen_vl_interface': 9.999937518676382e-06, 'learning_rate/action_model': 9.999937518676382e-05, 'epoch': 0.22}) +[2026-06-19 16:06:47,935][starVLA.training.train_starvla][INFO] - Step 109, Loss: {'train/loss': 0.24227233231067657, 'train/grad_norm_pre_clip': 22.044025421142578, 'timing/data': 0.026577720418572426, 'timing/model': 2.663030644878745, 'learning_rate/qwen_vl_interface': 9.999920922120846e-06, 'learning_rate/action_model': 9.999920922120846e-05, 'epoch': 0.22}) +[2026-06-19 16:06:58,477][starVLA.training.train_starvla][INFO] - Step 110, Loss: {'train/loss': 0.5614456534385681, 'train/grad_norm_pre_clip': 17.47772789001465, 'timing/data': 0.023285267874598503, 'timing/model': 2.7054583244025707, 'learning_rate/qwen_vl_interface': 9.999902373052239e-06, 'learning_rate/action_model': 9.999902373052239e-05, 'epoch': 0.23}) +[2026-06-19 16:07:08,933][starVLA.training.train_starvla][INFO] - Step 111, Loss: {'train/loss': 0.5437918305397034, 'train/grad_norm_pre_clip': 17.47494888305664, 'timing/data': 0.024176178500056267, 'timing/model': 2.6582125294953585, 'learning_rate/qwen_vl_interface': 9.999881871478188e-06, 'learning_rate/action_model': 9.999881871478187e-05, 'epoch': 0.23}) +[2026-06-19 16:07:19,433][starVLA.training.train_starvla][INFO] - Step 112, Loss: {'train/loss': 0.3291398286819458, 'train/grad_norm_pre_clip': 12.853412628173828, 'timing/data': 0.023923519998788834, 'timing/model': 2.661665689200163, 'learning_rate/qwen_vl_interface': 9.999859417407115e-06, 'learning_rate/action_model': 9.999859417407115e-05, 'epoch': 0.23}) +[2026-06-19 16:07:30,036][starVLA.training.train_starvla][INFO] - Step 113, Loss: {'train/loss': 0.6315219402313232, 'train/grad_norm_pre_clip': 31.225534439086914, 'timing/data': 0.02539067529141903, 'timing/model': 2.6725499276071787, 'learning_rate/qwen_vl_interface': 9.999835010848254e-06, 'learning_rate/action_model': 9.999835010848254e-05, 'epoch': 0.23}) +[2026-06-19 16:07:40,747][starVLA.training.train_starvla][INFO] - Step 114, Loss: {'train/loss': 1.0265998840332031, 'train/grad_norm_pre_clip': 44.06726837158203, 'timing/data': 0.023213278502225876, 'timing/model': 2.8724638503044844, 'learning_rate/qwen_vl_interface': 9.999808651811639e-06, 'learning_rate/action_model': 9.999808651811639e-05, 'epoch': 0.23}) +[2026-06-19 16:07:51,235][starVLA.training.train_starvla][INFO] - Step 115, Loss: {'train/loss': 0.9469782114028931, 'train/grad_norm_pre_clip': 45.2721061706543, 'timing/data': 0.024994248524308205, 'timing/model': 2.6589425913989544, 'learning_rate/qwen_vl_interface': 9.999780340308101e-06, 'learning_rate/action_model': 9.999780340308101e-05, 'epoch': 0.24}) +[2026-06-19 16:08:01,835][starVLA.training.train_starvla][INFO] - Step 116, Loss: {'train/loss': 0.547874391078949, 'train/grad_norm_pre_clip': 21.66370964050293, 'timing/data': 0.023383265361189842, 'timing/model': 2.6620280612260103, 'learning_rate/qwen_vl_interface': 9.999750076349282e-06, 'learning_rate/action_model': 9.999750076349282e-05, 'epoch': 0.24}) +[2026-06-19 16:08:12,345][starVLA.training.train_starvla][INFO] - Step 117, Loss: {'train/loss': 0.38447603583335876, 'train/grad_norm_pre_clip': 4.639749050140381, 'timing/data': 0.023029979318380356, 'timing/model': 2.6695593539625406, 'learning_rate/qwen_vl_interface': 9.999717859947618e-06, 'learning_rate/action_model': 9.999717859947618e-05, 'epoch': 0.24}) +[2026-06-19 16:08:23,151][starVLA.training.train_starvla][INFO] - Step 118, Loss: {'train/loss': 0.4099132716655731, 'train/grad_norm_pre_clip': 8.075307846069336, 'timing/data': 0.024149008095264435, 'timing/model': 2.6955801714211702, 'learning_rate/qwen_vl_interface': 9.999683691116356e-06, 'learning_rate/action_model': 9.999683691116355e-05, 'epoch': 0.24}) +[2026-06-19 16:08:33,975][starVLA.training.train_starvla][INFO] - Step 119, Loss: {'train/loss': 0.4590858519077301, 'train/grad_norm_pre_clip': 17.829618453979492, 'timing/data': 0.022814081981778145, 'timing/model': 2.74520000256598, 'learning_rate/qwen_vl_interface': 9.999647569869539e-06, 'learning_rate/action_model': 9.99964756986954e-05, 'epoch': 0.24}) +[2026-06-19 16:08:44,760][starVLA.training.train_starvla][INFO] - Step 120, Loss: {'train/loss': 0.41201624274253845, 'train/grad_norm_pre_clip': 8.578435897827148, 'timing/data': 0.02873854711651802, 'timing/model': 2.7134629022330046, 'learning_rate/qwen_vl_interface': 9.999609496222017e-06, 'learning_rate/action_model': 9.999609496222017e-05, 'epoch': 0.25}) +[2026-06-19 16:08:55,342][starVLA.training.train_starvla][INFO] - Step 121, Loss: {'train/loss': 0.22532837092876434, 'train/grad_norm_pre_clip': 5.460809707641602, 'timing/data': 0.021359950304031372, 'timing/model': 2.6664873976260424, 'learning_rate/qwen_vl_interface': 9.99956947018944e-06, 'learning_rate/action_model': 9.999569470189439e-05, 'epoch': 0.25}) +[2026-06-19 16:09:06,135][starVLA.training.train_starvla][INFO] - Step 122, Loss: {'train/loss': 0.23402826488018036, 'train/grad_norm_pre_clip': 6.603276252746582, 'timing/data': 0.022118791937828064, 'timing/model': 2.6646455824375153, 'learning_rate/qwen_vl_interface': 9.99952749178826e-06, 'learning_rate/action_model': 9.999527491788259e-05, 'epoch': 0.25}) +[2026-06-19 16:09:16,636][starVLA.training.train_starvla][INFO] - Step 123, Loss: {'train/loss': 0.23936711251735687, 'train/grad_norm_pre_clip': 3.268901824951172, 'timing/data': 0.022052211686968803, 'timing/model': 2.6628844402730465, 'learning_rate/qwen_vl_interface': 9.999483561035732e-06, 'learning_rate/action_model': 9.999483561035732e-05, 'epoch': 0.25}) +[2026-06-19 16:09:27,133][starVLA.training.train_starvla][INFO] - Step 124, Loss: {'train/loss': 0.28711017966270447, 'train/grad_norm_pre_clip': 9.510815620422363, 'timing/data': 0.021987512707710266, 'timing/model': 2.6620364412665367, 'learning_rate/qwen_vl_interface': 9.999437677949917e-06, 'learning_rate/action_model': 9.999437677949917e-05, 'epoch': 0.25}) +[2026-06-19 16:09:37,732][starVLA.training.train_starvla][INFO] - Step 125, Loss: {'train/loss': 0.6977466940879822, 'train/grad_norm_pre_clip': 28.31946563720703, 'timing/data': 0.022232361137866974, 'timing/model': 2.675522578880191, 'learning_rate/qwen_vl_interface': 9.999389842549674e-06, 'learning_rate/action_model': 9.999389842549673e-05, 'epoch': 0.26}) +[2026-06-19 16:09:48,234][starVLA.training.train_starvla][INFO] - Step 126, Loss: {'train/loss': 0.4517640471458435, 'train/grad_norm_pre_clip': 28.316926956176758, 'timing/data': 0.026442663744091988, 'timing/model': 2.654759583994746, 'learning_rate/qwen_vl_interface': 9.999340054854669e-06, 'learning_rate/action_model': 9.999340054854667e-05, 'epoch': 0.26}) +[2026-06-19 16:09:58,736][starVLA.training.train_starvla][INFO] - Step 127, Loss: {'train/loss': 0.6223576664924622, 'train/grad_norm_pre_clip': 18.3880672454834, 'timing/data': 0.021818991750478745, 'timing/model': 2.659948343411088, 'learning_rate/qwen_vl_interface': 9.999288314885362e-06, 'learning_rate/action_model': 9.999288314885362e-05, 'epoch': 0.26}) +[2026-06-19 16:10:09,234][starVLA.training.train_starvla][INFO] - Step 128, Loss: {'train/loss': 0.2481713593006134, 'train/grad_norm_pre_clip': 7.209381580352783, 'timing/data': 0.026783807203173637, 'timing/model': 2.6557261515408754, 'learning_rate/qwen_vl_interface': 9.999234622663027e-06, 'learning_rate/action_model': 9.999234622663028e-05, 'epoch': 0.26}) +[2026-06-19 16:10:19,737][starVLA.training.train_starvla][INFO] - Step 129, Loss: {'train/loss': 0.29359185695648193, 'train/grad_norm_pre_clip': 14.853249549865723, 'timing/data': 0.0232577882707119, 'timing/model': 2.662820253521204, 'learning_rate/qwen_vl_interface': 9.999178978209732e-06, 'learning_rate/action_model': 9.999178978209732e-05, 'epoch': 0.26}) +[2026-06-19 16:10:30,517][starVLA.training.train_starvla][INFO] - Step 130, Loss: {'train/loss': 0.5525147914886475, 'train/grad_norm_pre_clip': 11.567943572998047, 'timing/data': 0.022806163877248764, 'timing/model': 2.739080334082246, 'learning_rate/qwen_vl_interface': 9.999121381548354e-06, 'learning_rate/action_model': 9.999121381548353e-05, 'epoch': 0.27}) +[2026-06-19 16:10:41,136][starVLA.training.train_starvla][INFO] - Step 131, Loss: {'train/loss': 0.4779205918312073, 'train/grad_norm_pre_clip': 10.617030143737793, 'timing/data': 0.02298128232359886, 'timing/model': 2.7089122030884027, 'learning_rate/qwen_vl_interface': 9.999061832702563e-06, 'learning_rate/action_model': 9.999061832702563e-05, 'epoch': 0.27}) +[2026-06-19 16:10:51,804][starVLA.training.train_starvla][INFO] - Step 132, Loss: {'train/loss': 0.31635573506355286, 'train/grad_norm_pre_clip': 6.068423271179199, 'timing/data': 0.021988393738865852, 'timing/model': 2.758499626070261, 'learning_rate/qwen_vl_interface': 9.999000331696842e-06, 'learning_rate/action_model': 9.999000331696841e-05, 'epoch': 0.27}) +[2026-06-19 16:11:02,326][starVLA.training.train_starvla][INFO] - Step 133, Loss: {'train/loss': 0.4652940630912781, 'train/grad_norm_pre_clip': 11.255694389343262, 'timing/data': 0.023280268535017967, 'timing/model': 2.7055369336158037, 'learning_rate/qwen_vl_interface': 9.99893687855647e-06, 'learning_rate/action_model': 9.99893687855647e-05, 'epoch': 0.27}) +[2026-06-19 16:11:12,845][starVLA.training.train_starvla][INFO] - Step 134, Loss: {'train/loss': 0.3971138000488281, 'train/grad_norm_pre_clip': 5.9241251945495605, 'timing/data': 0.02229570783674717, 'timing/model': 2.667926747351885, 'learning_rate/qwen_vl_interface': 9.998871473307529e-06, 'learning_rate/action_model': 9.99887147330753e-05, 'epoch': 0.28}) +[2026-06-19 16:11:23,417][starVLA.training.train_starvla][INFO] - Step 135, Loss: {'train/loss': 0.3240598440170288, 'train/grad_norm_pre_clip': 5.370147228240967, 'timing/data': 0.022898243740200996, 'timing/model': 2.7435363363474607, 'learning_rate/qwen_vl_interface': 9.998804115976907e-06, 'learning_rate/action_model': 9.998804115976908e-05, 'epoch': 0.28}) +[2026-06-19 16:11:34,015][starVLA.training.train_starvla][INFO] - Step 136, Loss: {'train/loss': 0.27878454327583313, 'train/grad_norm_pre_clip': 3.672877311706543, 'timing/data': 0.02236664853990078, 'timing/model': 2.7382790073752403, 'learning_rate/qwen_vl_interface': 9.99873480659229e-06, 'learning_rate/action_model': 9.99873480659229e-05, 'epoch': 0.28}) +[2026-06-19 16:11:44,539][starVLA.training.train_starvla][INFO] - Step 137, Loss: {'train/loss': 0.30227479338645935, 'train/grad_norm_pre_clip': 3.2876179218292236, 'timing/data': 0.02339542657136917, 'timing/model': 2.662985870614648, 'learning_rate/qwen_vl_interface': 9.99866354518217e-06, 'learning_rate/action_model': 9.99866354518217e-05, 'epoch': 0.28}) +[2026-06-19 16:11:55,033][starVLA.training.train_starvla][INFO] - Step 138, Loss: {'train/loss': 0.27841171622276306, 'train/grad_norm_pre_clip': 6.33939266204834, 'timing/data': 0.02248089574277401, 'timing/model': 2.659131208434701, 'learning_rate/qwen_vl_interface': 9.99859033177584e-06, 'learning_rate/action_model': 9.99859033177584e-05, 'epoch': 0.28}) +[2026-06-19 16:12:05,604][starVLA.training.train_starvla][INFO] - Step 139, Loss: {'train/loss': 0.3675801753997803, 'train/grad_norm_pre_clip': 4.7740325927734375, 'timing/data': 0.022503525018692017, 'timing/model': 2.7239978574216366, 'learning_rate/qwen_vl_interface': 9.998515166403394e-06, 'learning_rate/action_model': 9.998515166403394e-05, 'epoch': 0.29}) +[2026-06-19 16:12:16,126][starVLA.training.train_starvla][INFO] - Step 140, Loss: {'train/loss': 0.34922945499420166, 'train/grad_norm_pre_clip': 6.070805072784424, 'timing/data': 0.03330528363585472, 'timing/model': 2.734779754653573, 'learning_rate/qwen_vl_interface': 9.99843804909573e-06, 'learning_rate/action_model': 9.99843804909573e-05, 'epoch': 0.29}) +[2026-06-19 16:12:26,642][starVLA.training.train_starvla][INFO] - Step 141, Loss: {'train/loss': 0.4024295210838318, 'train/grad_norm_pre_clip': 19.005231857299805, 'timing/data': 0.02597789838910103, 'timing/model': 2.640378190204501, 'learning_rate/qwen_vl_interface': 9.998358979884548e-06, 'learning_rate/action_model': 9.998358979884547e-05, 'epoch': 0.29}) +[2026-06-19 16:12:37,429][starVLA.training.train_starvla][INFO] - Step 142, Loss: {'train/loss': 0.4112786054611206, 'train/grad_norm_pre_clip': 14.471362113952637, 'timing/data': 0.030690666288137436, 'timing/model': 2.763130208477378, 'learning_rate/qwen_vl_interface': 9.998277958802352e-06, 'learning_rate/action_model': 9.998277958802351e-05, 'epoch': 0.29}) +[2026-06-19 16:12:48,005][starVLA.training.train_starvla][INFO] - Step 143, Loss: {'train/loss': 0.2075985223054886, 'train/grad_norm_pre_clip': 13.237858772277832, 'timing/data': 0.024209266528487206, 'timing/model': 2.7262070439755917, 'learning_rate/qwen_vl_interface': 9.998194985882442e-06, 'learning_rate/action_model': 9.998194985882442e-05, 'epoch': 0.29}) +[2026-06-19 16:12:58,513][starVLA.training.train_starvla][INFO] - Step 144, Loss: {'train/loss': 0.558600664138794, 'train/grad_norm_pre_clip': 16.685949325561523, 'timing/data': 0.021844493225216866, 'timing/model': 2.735606275498867, 'learning_rate/qwen_vl_interface': 9.99811006115893e-06, 'learning_rate/action_model': 9.998110061158931e-05, 'epoch': 0.3}) +[2026-06-19 16:13:09,039][starVLA.training.train_starvla][INFO] - Step 145, Loss: {'train/loss': 0.8104312419891357, 'train/grad_norm_pre_clip': 30.894195556640625, 'timing/data': 0.022407349199056625, 'timing/model': 2.661363882943988, 'learning_rate/qwen_vl_interface': 9.998023184666726e-06, 'learning_rate/action_model': 9.998023184666726e-05, 'epoch': 0.3}) +[2026-06-19 16:13:19,535][starVLA.training.train_starvla][INFO] - Step 146, Loss: {'train/loss': 0.7721033096313477, 'train/grad_norm_pre_clip': 22.328907012939453, 'timing/data': 0.02249191887676716, 'timing/model': 2.656167032197118, 'learning_rate/qwen_vl_interface': 9.997934356441538e-06, 'learning_rate/action_model': 9.997934356441537e-05, 'epoch': 0.3}) +[2026-06-19 16:13:30,040][starVLA.training.train_starvla][INFO] - Step 147, Loss: {'train/loss': 0.7353147268295288, 'train/grad_norm_pre_clip': 28.474037170410156, 'timing/data': 0.023026112467050552, 'timing/model': 2.6625527013093233, 'learning_rate/qwen_vl_interface': 9.997843576519879e-06, 'learning_rate/action_model': 9.997843576519878e-05, 'epoch': 0.3}) +[2026-06-19 16:13:40,536][starVLA.training.train_starvla][INFO] - Step 148, Loss: {'train/loss': 0.3453528583049774, 'train/grad_norm_pre_clip': 10.651527404785156, 'timing/data': 0.02241622470319271, 'timing/model': 2.6565249878913164, 'learning_rate/qwen_vl_interface': 9.99775084493907e-06, 'learning_rate/action_model': 9.997750844939069e-05, 'epoch': 0.3}) +[2026-06-19 16:13:51,112][starVLA.training.train_starvla][INFO] - Step 149, Loss: {'train/loss': 0.26427319645881653, 'train/grad_norm_pre_clip': 5.2773237228393555, 'timing/data': 0.021711990237236023, 'timing/model': 2.737695001065731, 'learning_rate/qwen_vl_interface': 9.997656161737227e-06, 'learning_rate/action_model': 9.997656161737226e-05, 'epoch': 0.31}) +[2026-06-19 16:14:01,719][starVLA.training.train_starvla][INFO] - Step 150, Loss: {'train/loss': 0.2593134343624115, 'train/grad_norm_pre_clip': 11.513336181640625, 'timing/data': 0.023128217086195946, 'timing/model': 2.7070106249302626, 'learning_rate/qwen_vl_interface': 9.99755952695327e-06, 'learning_rate/action_model': 9.997559526953269e-05, 'epoch': 0.31}) +[2026-06-19 16:14:12,236][starVLA.training.train_starvla][INFO] - Step 151, Loss: {'train/loss': 0.3414633870124817, 'train/grad_norm_pre_clip': 8.231610298156738, 'timing/data': 0.021840078756213188, 'timing/model': 2.652756802737713, 'learning_rate/qwen_vl_interface': 9.99746094062692e-06, 'learning_rate/action_model': 9.99746094062692e-05, 'epoch': 0.31}) +[2026-06-19 16:14:22,745][starVLA.training.train_starvla][INFO] - Step 152, Loss: {'train/loss': 0.30822286009788513, 'train/grad_norm_pre_clip': 3.259775400161743, 'timing/data': 0.021768830716609955, 'timing/model': 2.671121794730425, 'learning_rate/qwen_vl_interface': 9.997360402798706e-06, 'learning_rate/action_model': 9.997360402798705e-05, 'epoch': 0.31}) +[2026-06-19 16:14:33,518][starVLA.training.train_starvla][INFO] - Step 153, Loss: {'train/loss': 0.3702337145805359, 'train/grad_norm_pre_clip': 4.661111831665039, 'timing/data': 0.02324146218597889, 'timing/model': 2.762380050495267, 'learning_rate/qwen_vl_interface': 9.997257913509955e-06, 'learning_rate/action_model': 9.997257913509955e-05, 'epoch': 0.31}) +[2026-06-19 16:14:44,184][starVLA.training.train_starvla][INFO] - Step 154, Loss: {'train/loss': 0.48155513405799866, 'train/grad_norm_pre_clip': 7.406219959259033, 'timing/data': 0.03315076604485512, 'timing/model': 2.705164197832346, 'learning_rate/qwen_vl_interface': 9.997153472802794e-06, 'learning_rate/action_model': 9.997153472802795e-05, 'epoch': 0.32}) +[2026-06-19 16:14:54,723][starVLA.training.train_starvla][INFO] - Step 155, Loss: {'train/loss': 0.32020607590675354, 'train/grad_norm_pre_clip': 10.420280456542969, 'timing/data': 0.03314545936882496, 'timing/model': 2.733019633218646, 'learning_rate/qwen_vl_interface': 9.997047080720157e-06, 'learning_rate/action_model': 9.997047080720158e-05, 'epoch': 0.32}) +[2026-06-19 16:15:05,303][starVLA.training.train_starvla][INFO] - Step 156, Loss: {'train/loss': 0.2545555531978607, 'train/grad_norm_pre_clip': 4.1741509437561035, 'timing/data': 0.024766022339463234, 'timing/model': 2.7246317360550165, 'learning_rate/qwen_vl_interface': 9.996938737305777e-06, 'learning_rate/action_model': 9.996938737305776e-05, 'epoch': 0.32}) +[2026-06-19 16:15:16,031][starVLA.training.train_starvla][INFO] - Step 157, Loss: {'train/loss': 0.17051821947097778, 'train/grad_norm_pre_clip': 7.73253870010376, 'timing/data': 0.05588253028690815, 'timing/model': 2.738913342356682, 'learning_rate/qwen_vl_interface': 9.996828442604189e-06, 'learning_rate/action_model': 9.99682844260419e-05, 'epoch': 0.32}) +[2026-06-19 16:15:26,631][starVLA.training.train_starvla][INFO] - Step 158, Loss: {'train/loss': 0.19280672073364258, 'train/grad_norm_pre_clip': 6.372618198394775, 'timing/data': 0.027592312544584274, 'timing/model': 2.6730305291712284, 'learning_rate/qwen_vl_interface': 9.996716196660733e-06, 'learning_rate/action_model': 9.996716196660731e-05, 'epoch': 0.32}) +[2026-06-19 16:15:37,237][starVLA.training.train_starvla][INFO] - Step 159, Loss: {'train/loss': 0.2874855697154999, 'train/grad_norm_pre_clip': 5.1953277587890625, 'timing/data': 0.027061210945248604, 'timing/model': 2.6628002133220434, 'learning_rate/qwen_vl_interface': 9.996601999521546e-06, 'learning_rate/action_model': 9.996601999521546e-05, 'epoch': 0.33}) +[2026-06-19 16:15:47,737][starVLA.training.train_starvla][INFO] - Step 160, Loss: {'train/loss': 0.3926021456718445, 'train/grad_norm_pre_clip': 9.697309494018555, 'timing/data': 0.023700617253780365, 'timing/model': 2.66275642439723, 'learning_rate/qwen_vl_interface': 9.996485851233572e-06, 'learning_rate/action_model': 9.996485851233572e-05, 'epoch': 0.33}) +[2026-06-19 16:15:58,236][starVLA.training.train_starvla][INFO] - Step 161, Loss: {'train/loss': 0.24899743497371674, 'train/grad_norm_pre_clip': 7.929901599884033, 'timing/data': 0.021772626787424088, 'timing/model': 2.663888903334737, 'learning_rate/qwen_vl_interface': 9.996367751844556e-06, 'learning_rate/action_model': 9.996367751844555e-05, 'epoch': 0.33}) +[2026-06-19 16:16:08,832][starVLA.training.train_starvla][INFO] - Step 162, Loss: {'train/loss': 0.3384433388710022, 'train/grad_norm_pre_clip': 6.332489490509033, 'timing/data': 0.02760297991335392, 'timing/model': 2.6581401638686657, 'learning_rate/qwen_vl_interface': 9.996247701403043e-06, 'learning_rate/action_model': 9.996247701403042e-05, 'epoch': 0.33}) +[2026-06-19 16:16:19,334][starVLA.training.train_starvla][INFO] - Step 163, Loss: {'train/loss': 0.4829704165458679, 'train/grad_norm_pre_clip': 7.5690412521362305, 'timing/data': 0.023714778944849968, 'timing/model': 2.652372647076845, 'learning_rate/qwen_vl_interface': 9.99612569995838e-06, 'learning_rate/action_model': 9.99612569995838e-05, 'epoch': 0.33}) +[2026-06-19 16:16:30,114][starVLA.training.train_starvla][INFO] - Step 164, Loss: {'train/loss': 0.6113607883453369, 'train/grad_norm_pre_clip': 19.652517318725586, 'timing/data': 0.029110031202435493, 'timing/model': 2.7465410977602005, 'learning_rate/qwen_vl_interface': 9.99600174756072e-06, 'learning_rate/action_model': 9.99600174756072e-05, 'epoch': 0.34}) +[2026-06-19 16:16:40,720][starVLA.training.train_starvla][INFO] - Step 165, Loss: {'train/loss': 0.31758442521095276, 'train/grad_norm_pre_clip': 11.629300117492676, 'timing/data': 0.022838866338133812, 'timing/model': 2.750806003808975, 'learning_rate/qwen_vl_interface': 9.995875844261012e-06, 'learning_rate/action_model': 9.995875844261013e-05, 'epoch': 0.34}) +[2026-06-19 16:16:51,227][starVLA.training.train_starvla][INFO] - Step 166, Loss: {'train/loss': 0.3115226626396179, 'train/grad_norm_pre_clip': 8.202302932739258, 'timing/data': 0.022969206795096397, 'timing/model': 2.655673651024699, 'learning_rate/qwen_vl_interface': 9.995747990111015e-06, 'learning_rate/action_model': 9.995747990111013e-05, 'epoch': 0.34}) +[2026-06-19 16:17:01,728][starVLA.training.train_starvla][INFO] - Step 167, Loss: {'train/loss': 0.44599178433418274, 'train/grad_norm_pre_clip': 13.229584693908691, 'timing/data': 0.02174999564886093, 'timing/model': 2.661515550687909, 'learning_rate/qwen_vl_interface': 9.99561818516328e-06, 'learning_rate/action_model': 9.99561818516328e-05, 'epoch': 0.34}) +[2026-06-19 16:17:12,228][starVLA.training.train_starvla][INFO] - Step 168, Loss: {'train/loss': 0.16252844035625458, 'train/grad_norm_pre_clip': 2.4962663650512695, 'timing/data': 0.021324224770069122, 'timing/model': 2.6587897520512342, 'learning_rate/qwen_vl_interface': 9.995486429471166e-06, 'learning_rate/action_model': 9.995486429471166e-05, 'epoch': 0.35}) +[2026-06-19 16:17:22,759][starVLA.training.train_starvla][INFO] - Step 169, Loss: {'train/loss': 0.6849890351295471, 'train/grad_norm_pre_clip': 19.768739700317383, 'timing/data': 0.022155780345201492, 'timing/model': 2.690986957401037, 'learning_rate/qwen_vl_interface': 9.995352723088833e-06, 'learning_rate/action_model': 9.995352723088834e-05, 'epoch': 0.35}) +[2026-06-19 16:17:33,236][starVLA.training.train_starvla][INFO] - Step 170, Loss: {'train/loss': 0.31537365913391113, 'train/grad_norm_pre_clip': 10.26108455657959, 'timing/data': 0.02078414335846901, 'timing/model': 2.665176872164011, 'learning_rate/qwen_vl_interface': 9.995217066071245e-06, 'learning_rate/action_model': 9.995217066071246e-05, 'epoch': 0.35}) +[2026-06-19 16:17:43,837][starVLA.training.train_starvla][INFO] - Step 171, Loss: {'train/loss': 0.40230631828308105, 'train/grad_norm_pre_clip': 10.869258880615234, 'timing/data': 0.021423636004328728, 'timing/model': 2.6718046497553587, 'learning_rate/qwen_vl_interface': 9.995079458474163e-06, 'learning_rate/action_model': 9.995079458474164e-05, 'epoch': 0.35}) +[2026-06-19 16:17:54,354][starVLA.training.train_starvla][INFO] - Step 172, Loss: {'train/loss': 0.2637704312801361, 'train/grad_norm_pre_clip': 3.577345609664917, 'timing/data': 0.02167869172990322, 'timing/model': 2.682652272284031, 'learning_rate/qwen_vl_interface': 9.994939900354154e-06, 'learning_rate/action_model': 9.994939900354154e-05, 'epoch': 0.35}) +[2026-06-19 16:18:04,915][starVLA.training.train_starvla][INFO] - Step 173, Loss: {'train/loss': 0.2450912445783615, 'train/grad_norm_pre_clip': 4.255972385406494, 'timing/data': 0.02096453681588173, 'timing/model': 2.6929998211562634, 'learning_rate/qwen_vl_interface': 9.994798391768582e-06, 'learning_rate/action_model': 9.994798391768582e-05, 'epoch': 0.36}) +[2026-06-19 16:18:15,509][starVLA.training.train_starvla][INFO] - Step 174, Loss: {'train/loss': 0.39680737257003784, 'train/grad_norm_pre_clip': 11.363875389099121, 'timing/data': 0.021196072921156883, 'timing/model': 2.7373928911983967, 'learning_rate/qwen_vl_interface': 9.99465493277562e-06, 'learning_rate/action_model': 9.994654932775618e-05, 'epoch': 0.36}) +[2026-06-19 16:18:25,930][starVLA.training.train_starvla][INFO] - Step 175, Loss: {'train/loss': 0.393686980009079, 'train/grad_norm_pre_clip': 11.410432815551758, 'timing/data': 0.021473176777362823, 'timing/model': 2.659523794427514, 'learning_rate/qwen_vl_interface': 9.994509523434235e-06, 'learning_rate/action_model': 9.994509523434235e-05, 'epoch': 0.36}) +[2026-06-19 16:18:36,429][starVLA.training.train_starvla][INFO] - Step 176, Loss: {'train/loss': 0.3764641582965851, 'train/grad_norm_pre_clip': 11.402998924255371, 'timing/data': 0.022325247526168823, 'timing/model': 2.652000291272998, 'learning_rate/qwen_vl_interface': 9.994362163804202e-06, 'learning_rate/action_model': 9.9943621638042e-05, 'epoch': 0.36}) +[2026-06-19 16:18:46,928][starVLA.training.train_starvla][INFO] - Step 177, Loss: {'train/loss': 0.3757359981536865, 'train/grad_norm_pre_clip': 6.57354211807251, 'timing/data': 0.021535128355026245, 'timing/model': 2.6604631058871746, 'learning_rate/qwen_vl_interface': 9.994212853946092e-06, 'learning_rate/action_model': 9.994212853946092e-05, 'epoch': 0.36}) +[2026-06-19 16:18:57,502][starVLA.training.train_starvla][INFO] - Step 178, Loss: {'train/loss': 0.2948210835456848, 'train/grad_norm_pre_clip': 3.7723195552825928, 'timing/data': 0.021984044462442398, 'timing/model': 2.7295566983520985, 'learning_rate/qwen_vl_interface': 9.994061593921283e-06, 'learning_rate/action_model': 9.994061593921282e-05, 'epoch': 0.37}) +[2026-06-19 16:19:07,932][starVLA.training.train_starvla][INFO] - Step 179, Loss: {'train/loss': 0.3984367847442627, 'train/grad_norm_pre_clip': 7.750726222991943, 'timing/data': 0.020792124792933464, 'timing/model': 2.6248808205127716, 'learning_rate/qwen_vl_interface': 9.993908383791951e-06, 'learning_rate/action_model': 9.993908383791951e-05, 'epoch': 0.37}) +[2026-06-19 16:19:18,533][starVLA.training.train_starvla][INFO] - Step 180, Loss: {'train/loss': 0.402901828289032, 'train/grad_norm_pre_clip': 6.298789024353027, 'timing/data': 0.02139047533273697, 'timing/model': 2.6787289530038834, 'learning_rate/qwen_vl_interface': 9.993753223621076e-06, 'learning_rate/action_model': 9.993753223621077e-05, 'epoch': 0.37}) +[2026-06-19 16:19:29,129][starVLA.training.train_starvla][INFO] - Step 181, Loss: {'train/loss': 0.15080136060714722, 'train/grad_norm_pre_clip': 2.2505860328674316, 'timing/data': 0.02198033779859543, 'timing/model': 2.6770880725234747, 'learning_rate/qwen_vl_interface': 9.99359611347244e-06, 'learning_rate/action_model': 9.993596113472439e-05, 'epoch': 0.37}) +[2026-06-19 16:19:39,631][starVLA.training.train_starvla][INFO] - Step 182, Loss: {'train/loss': 0.3551597595214844, 'train/grad_norm_pre_clip': 7.168659687042236, 'timing/data': 0.02110184170305729, 'timing/model': 2.641401855275035, 'learning_rate/qwen_vl_interface': 9.99343705341062e-06, 'learning_rate/action_model': 9.99343705341062e-05, 'epoch': 0.37}) +[2026-06-19 16:19:50,131][starVLA.training.train_starvla][INFO] - Step 183, Loss: {'train/loss': 0.3009962737560272, 'train/grad_norm_pre_clip': 3.670942544937134, 'timing/data': 0.021635526791214943, 'timing/model': 2.6605749521404505, 'learning_rate/qwen_vl_interface': 9.993276043501004e-06, 'learning_rate/action_model': 9.993276043501004e-05, 'epoch': 0.38}) +[2026-06-19 16:20:00,626][starVLA.training.train_starvla][INFO] - Step 184, Loss: {'train/loss': 0.2549295723438263, 'train/grad_norm_pre_clip': 5.107835292816162, 'timing/data': 0.02178049273788929, 'timing/model': 2.6597036067396402, 'learning_rate/qwen_vl_interface': 9.993113083809776e-06, 'learning_rate/action_model': 9.993113083809775e-05, 'epoch': 0.38}) +[2026-06-19 16:20:11,134][starVLA.training.train_starvla][INFO] - Step 185, Loss: {'train/loss': 0.438249796628952, 'train/grad_norm_pre_clip': 6.022160053253174, 'timing/data': 0.020655550062656403, 'timing/model': 2.6656458880752325, 'learning_rate/qwen_vl_interface': 9.992948174403921e-06, 'learning_rate/action_model': 9.992948174403922e-05, 'epoch': 0.38}) +[2026-06-19 16:20:21,629][starVLA.training.train_starvla][INFO] - Step 186, Loss: {'train/loss': 0.24351760745048523, 'train/grad_norm_pre_clip': 7.578587055206299, 'timing/data': 0.021702194586396217, 'timing/model': 2.6578476540744305, 'learning_rate/qwen_vl_interface': 9.99278131535123e-06, 'learning_rate/action_model': 9.992781315351229e-05, 'epoch': 0.38}) +[2026-06-19 16:20:32,153][starVLA.training.train_starvla][INFO] - Step 187, Loss: {'train/loss': 0.2725242078304291, 'train/grad_norm_pre_clip': 4.488133907318115, 'timing/data': 0.021792935207486153, 'timing/model': 2.6844253353774548, 'learning_rate/qwen_vl_interface': 9.992612506720289e-06, 'learning_rate/action_model': 9.99261250672029e-05, 'epoch': 0.38}) +[2026-06-19 16:20:42,741][starVLA.training.train_starvla][INFO] - Step 188, Loss: {'train/loss': 0.12259765714406967, 'train/grad_norm_pre_clip': 6.000981330871582, 'timing/data': 0.02131003327667713, 'timing/model': 2.673050757497549, 'learning_rate/qwen_vl_interface': 9.992441748580492e-06, 'learning_rate/action_model': 9.99244174858049e-05, 'epoch': 0.39}) +[2026-06-19 16:20:53,329][starVLA.training.train_starvla][INFO] - Step 189, Loss: {'train/loss': 0.1415766328573227, 'train/grad_norm_pre_clip': 4.372497081756592, 'timing/data': 0.021712077781558037, 'timing/model': 2.6564910877496004, 'learning_rate/qwen_vl_interface': 9.992269041002028e-06, 'learning_rate/action_model': 9.992269041002027e-05, 'epoch': 0.39}) +[2026-06-19 16:21:03,829][starVLA.training.train_starvla][INFO] - Step 190, Loss: {'train/loss': 0.7776826620101929, 'train/grad_norm_pre_clip': 31.950708389282227, 'timing/data': 0.020711032673716545, 'timing/model': 2.663301929831505, 'learning_rate/qwen_vl_interface': 9.992094384055894e-06, 'learning_rate/action_model': 9.992094384055895e-05, 'epoch': 0.39}) +[2026-06-19 16:21:14,329][starVLA.training.train_starvla][INFO] - Step 191, Loss: {'train/loss': 0.39868783950805664, 'train/grad_norm_pre_clip': 14.791869163513184, 'timing/data': 0.02134375460445881, 'timing/model': 2.663606509566307, 'learning_rate/qwen_vl_interface': 9.991917777813884e-06, 'learning_rate/action_model': 9.991917777813884e-05, 'epoch': 0.39}) +[2026-06-19 16:21:24,831][starVLA.training.train_starvla][INFO] - Step 192, Loss: {'train/loss': 0.5079382061958313, 'train/grad_norm_pre_clip': 17.64641571044922, 'timing/data': 0.020777493715286255, 'timing/model': 2.662285203114152, 'learning_rate/qwen_vl_interface': 9.991739222348593e-06, 'learning_rate/action_model': 9.991739222348592e-05, 'epoch': 0.39}) +[2026-06-19 16:21:35,334][starVLA.training.train_starvla][INFO] - Step 193, Loss: {'train/loss': 0.18176265060901642, 'train/grad_norm_pre_clip': 9.978610038757324, 'timing/data': 0.022203052416443825, 'timing/model': 2.6588808111846447, 'learning_rate/qwen_vl_interface': 9.991558717733418e-06, 'learning_rate/action_model': 9.991558717733417e-05, 'epoch': 0.4}) +[2026-06-19 16:21:45,835][starVLA.training.train_starvla][INFO] - Step 194, Loss: {'train/loss': 0.22322435677051544, 'train/grad_norm_pre_clip': 11.89747142791748, 'timing/data': 0.020508870482444763, 'timing/model': 2.6709314845502377, 'learning_rate/qwen_vl_interface': 9.99137626404256e-06, 'learning_rate/action_model': 9.991376264042559e-05, 'epoch': 0.4}) +[2026-06-19 16:21:56,326][starVLA.training.train_starvla][INFO] - Step 195, Loss: {'train/loss': 0.5741075277328491, 'train/grad_norm_pre_clip': 12.980138778686523, 'timing/data': 0.02250825986266136, 'timing/model': 2.6524911113083363, 'learning_rate/qwen_vl_interface': 9.991191861351015e-06, 'learning_rate/action_model': 9.991191861351014e-05, 'epoch': 0.4}) +[2026-06-19 16:22:06,829][starVLA.training.train_starvla][INFO] - Step 196, Loss: {'train/loss': 0.34296804666519165, 'train/grad_norm_pre_clip': 12.42869758605957, 'timing/data': 0.021007096394896507, 'timing/model': 2.6620438881218433, 'learning_rate/qwen_vl_interface': 9.991005509734587e-06, 'learning_rate/action_model': 9.991005509734587e-05, 'epoch': 0.4}) +[2026-06-19 16:22:17,327][starVLA.training.train_starvla][INFO] - Step 197, Loss: {'train/loss': 0.5871556401252747, 'train/grad_norm_pre_clip': 23.609209060668945, 'timing/data': 0.02188515104353428, 'timing/model': 2.6563490610569715, 'learning_rate/qwen_vl_interface': 9.990817209269878e-06, 'learning_rate/action_model': 9.990817209269878e-05, 'epoch': 0.4}) +[2026-06-19 16:22:27,825][starVLA.training.train_starvla][INFO] - Step 198, Loss: {'train/loss': 0.43904435634613037, 'train/grad_norm_pre_clip': 18.425838470458984, 'timing/data': 0.021353380754590034, 'timing/model': 2.6601243503391743, 'learning_rate/qwen_vl_interface': 9.990626960034291e-06, 'learning_rate/action_model': 9.990626960034291e-05, 'epoch': 0.41}) +[2026-06-19 16:22:38,327][starVLA.training.train_starvla][INFO] - Step 199, Loss: {'train/loss': 0.37527644634246826, 'train/grad_norm_pre_clip': 9.005928039550781, 'timing/data': 0.020776960998773575, 'timing/model': 2.657538264989853, 'learning_rate/qwen_vl_interface': 9.990434762106029e-06, 'learning_rate/action_model': 9.990434762106028e-05, 'epoch': 0.41}) +[2026-06-19 16:22:48,826][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 200 +[2026-06-19 16:30:45,314][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 200 +[2026-06-19 16:30:45,315][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 200 +[2026-06-19 16:39:40,695][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 200 +[2026-06-19 16:39:40,697][starVLA.training.train_starvla][INFO] - Step 200, Loss: {'train/loss': 0.26912251114845276, 'train/grad_norm_pre_clip': 3.557455539703369, 'eval/loss': 0.3451746702194214, 'eval/latency_0/loss': 0.18209244310855865, 'eval/latency_1/loss': 0.26337119936943054, 'eval/latency_2/loss': 0.38965559005737305, 'eval/latency_3/loss': 0.4123360812664032, 'eval/latency_4/loss': 0.4717881977558136, 'eval/flappy/loss': 0.3451746702194214, 'eval/flappy/latency_0/loss': 0.18209244310855865, 'eval/flappy/latency_1/loss': 0.26337119936943054, 'eval/flappy/latency_2/loss': 0.38965559005737305, 'eval/flappy/latency_3/loss': 0.4123360812664032, 'eval/flappy/latency_4/loss': 0.4717881977558136, 'eval/flap_precision': 0.881578947368421, 'eval/flap_recall': 0.24363636363636362, 'eval/flap_f1': 0.3817663817663818, 'eval/flap_rate_model': 0.021111111111111112, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.3817663817663818, 'eval/latency_0/flap_precision': 0.881578947368421, 'eval/latency_0/flap_recall': 0.24363636363636362, 'eval/latency_0/flap_f1': 0.3817663817663818, 'eval/latency_0/flap_rate_model': 0.021111111111111112, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.3817663817663818, 'timing/data': 0.0208280012011528, 'timing/model': 2.6626138854771852, 'learning_rate/qwen_vl_interface': 9.990240615564099e-06, 'learning_rate/action_model': 9.990240615564099e-05, 'epoch': 0.41}) +[2026-06-19 16:39:40,697][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_200_state +[2026-06-19 16:39:40,702][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-19 16:39:48,931][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_200_state/model.safetensors +[2026-06-19 16:40:03,824][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_200_state/optimizer.bin +[2026-06-19 16:40:03,825][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_200_state/sampler.bin +[2026-06-19 16:40:03,825][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_200_state/sampler_1.bin +[2026-06-19 16:40:03,826][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_200_state/random_states_0.pkl +[2026-06-19 16:40:03,948][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 200 +[2026-06-19 16:40:03,949][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-19 16:40:03,964][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-19 16:40:14,829][starVLA.training.train_starvla][INFO] - Step 201, Loss: {'train/loss': 0.28580960631370544, 'train/grad_norm_pre_clip': 3.297569751739502, 'timing/data': 0.029610004276037216, 'timing/model': 2.7587364073842764, 'learning_rate/qwen_vl_interface': 9.990044520488305e-06, 'learning_rate/action_model': 9.990044520488305e-05, 'epoch': 0.41}) +[2026-06-19 16:40:26,179][starVLA.training.train_starvla][INFO] - Step 202, Loss: {'train/loss': 0.5920882225036621, 'train/grad_norm_pre_clip': 10.330618858337402, 'timing/data': 0.02677007019519806, 'timing/model': 3.209307599812746, 'learning_rate/qwen_vl_interface': 9.989846476959258e-06, 'learning_rate/action_model': 9.989846476959257e-05, 'epoch': 0.41}) +[2026-06-19 16:40:37,023][starVLA.training.train_starvla][INFO] - Step 203, Loss: {'train/loss': 0.41879406571388245, 'train/grad_norm_pre_clip': 6.8997416496276855, 'timing/data': 0.029308302327990532, 'timing/model': 2.7650349009782076, 'learning_rate/qwen_vl_interface': 9.989646485058361e-06, 'learning_rate/action_model': 9.989646485058361e-05, 'epoch': 0.42}) +[2026-06-19 16:40:47,829][starVLA.training.train_starvla][INFO] - Step 204, Loss: {'train/loss': 0.810757040977478, 'train/grad_norm_pre_clip': 13.94235897064209, 'timing/data': 0.030241208150982857, 'timing/model': 2.762645261362195, 'learning_rate/qwen_vl_interface': 9.98944454486783e-06, 'learning_rate/action_model': 9.989444544867829e-05, 'epoch': 0.42}) +[2026-06-19 16:40:58,709][starVLA.training.train_starvla][INFO] - Step 205, Loss: {'train/loss': 0.598392128944397, 'train/grad_norm_pre_clip': 10.711764335632324, 'timing/data': 0.029871776700019836, 'timing/model': 2.7439567875117064, 'learning_rate/qwen_vl_interface': 9.989240656470667e-06, 'learning_rate/action_model': 9.989240656470666e-05, 'epoch': 0.42}) +[2026-06-19 16:41:09,819][starVLA.training.train_starvla][INFO] - Step 206, Loss: {'train/loss': 0.29544782638549805, 'train/grad_norm_pre_clip': 2.1382906436920166, 'timing/data': 0.03034638799726963, 'timing/model': 2.7403350938111544, 'learning_rate/qwen_vl_interface': 9.98903481995069e-06, 'learning_rate/action_model': 9.989034819950689e-05, 'epoch': 0.42}) +[2026-06-19 16:41:20,629][starVLA.training.train_starvla][INFO] - Step 207, Loss: {'train/loss': 0.36988189816474915, 'train/grad_norm_pre_clip': 3.5945239067077637, 'timing/data': 0.029202532023191452, 'timing/model': 2.7531706783920527, 'learning_rate/qwen_vl_interface': 9.988827035392506e-06, 'learning_rate/action_model': 9.988827035392507e-05, 'epoch': 0.43}) +[2026-06-19 16:41:31,525][starVLA.training.train_starvla][INFO] - Step 208, Loss: {'train/loss': 0.34772905707359314, 'train/grad_norm_pre_clip': 6.612998008728027, 'timing/data': 0.028984956443309784, 'timing/model': 2.751130882650614, 'learning_rate/qwen_vl_interface': 9.988617302881529e-06, 'learning_rate/action_model': 9.988617302881529e-05, 'epoch': 0.43}) +[2026-06-19 16:41:42,324][starVLA.training.train_starvla][INFO] - Step 209, Loss: {'train/loss': 0.2853602170944214, 'train/grad_norm_pre_clip': 5.222613334655762, 'timing/data': 0.029286062344908714, 'timing/model': 2.751373115926981, 'learning_rate/qwen_vl_interface': 9.988405622503972e-06, 'learning_rate/action_model': 9.988405622503971e-05, 'epoch': 0.43}) +[2026-06-19 16:41:53,526][starVLA.training.train_starvla][INFO] - Step 210, Loss: {'train/loss': 0.21651823818683624, 'train/grad_norm_pre_clip': 5.21579122543335, 'timing/data': 0.02908383309841156, 'timing/model': 2.7561364490538836, 'learning_rate/qwen_vl_interface': 9.988191994346848e-06, 'learning_rate/action_model': 9.988191994346848e-05, 'epoch': 0.43}) +[2026-06-19 16:42:04,328][starVLA.training.train_starvla][INFO] - Step 211, Loss: {'train/loss': 0.3493710160255432, 'train/grad_norm_pre_clip': 4.6989641189575195, 'timing/data': 0.029214398935437202, 'timing/model': 2.757403090596199, 'learning_rate/qwen_vl_interface': 9.987976418497974e-06, 'learning_rate/action_model': 9.987976418497972e-05, 'epoch': 0.43}) +[2026-06-19 16:42:15,127][starVLA.training.train_starvla][INFO] - Step 212, Loss: {'train/loss': 0.13885943591594696, 'train/grad_norm_pre_clip': 5.144724369049072, 'timing/data': 0.029605308547616005, 'timing/model': 2.730126064270735, 'learning_rate/qwen_vl_interface': 9.987758895045964e-06, 'learning_rate/action_model': 9.987758895045963e-05, 'epoch': 0.44}) +[2026-06-19 16:42:25,910][starVLA.training.train_starvla][INFO] - Step 213, Loss: {'train/loss': 0.28581562638282776, 'train/grad_norm_pre_clip': 9.403271675109863, 'timing/data': 0.02952318824827671, 'timing/model': 2.7435527611523867, 'learning_rate/qwen_vl_interface': 9.98753942408023e-06, 'learning_rate/action_model': 9.98753942408023e-05, 'epoch': 0.44}) +[2026-06-19 16:42:36,821][starVLA.training.train_starvla][INFO] - Step 214, Loss: {'train/loss': 0.3142140805721283, 'train/grad_norm_pre_clip': 3.6483418941497803, 'timing/data': 0.029051152989268303, 'timing/model': 2.7635899391025305, 'learning_rate/qwen_vl_interface': 9.987318005690992e-06, 'learning_rate/action_model': 9.987318005690992e-05, 'epoch': 0.44}) +[2026-06-19 16:42:47,619][starVLA.training.train_starvla][INFO] - Step 215, Loss: {'train/loss': 0.4453405439853668, 'train/grad_norm_pre_clip': 16.140121459960938, 'timing/data': 0.029132824391126633, 'timing/model': 2.7590491250157356, 'learning_rate/qwen_vl_interface': 9.987094639969268e-06, 'learning_rate/action_model': 9.987094639969268e-05, 'epoch': 0.44}) +[2026-06-19 16:42:58,421][starVLA.training.train_starvla][INFO] - Step 216, Loss: {'train/loss': 0.19730855524539948, 'train/grad_norm_pre_clip': 5.93105936050415, 'timing/data': 0.02873462811112404, 'timing/model': 2.751403596252203, 'learning_rate/qwen_vl_interface': 9.986869327006872e-06, 'learning_rate/action_model': 9.986869327006873e-05, 'epoch': 0.44}) +[2026-06-19 16:43:09,230][starVLA.training.train_starvla][INFO] - Step 217, Loss: {'train/loss': 0.3311622738838196, 'train/grad_norm_pre_clip': 5.884889125823975, 'timing/data': 0.029425200074911118, 'timing/model': 2.7562177050858736, 'learning_rate/qwen_vl_interface': 9.986642066896424e-06, 'learning_rate/action_model': 9.986642066896423e-05, 'epoch': 0.45}) +[2026-06-19 16:43:20,126][starVLA.training.train_starvla][INFO] - Step 218, Loss: {'train/loss': 0.5166257619857788, 'train/grad_norm_pre_clip': 9.398880958557129, 'timing/data': 0.028045158833265305, 'timing/model': 2.771878531202674, 'learning_rate/qwen_vl_interface': 9.986412859731341e-06, 'learning_rate/action_model': 9.98641285973134e-05, 'epoch': 0.45}) +[2026-06-19 16:43:30,927][starVLA.training.train_starvla][INFO] - Step 219, Loss: {'train/loss': 0.33676978945732117, 'train/grad_norm_pre_clip': 10.318428039550781, 'timing/data': 0.026718124747276306, 'timing/model': 2.759132694453001, 'learning_rate/qwen_vl_interface': 9.986181705605839e-06, 'learning_rate/action_model': 9.986181705605839e-05, 'epoch': 0.45}) +[2026-06-19 16:43:41,725][starVLA.training.train_starvla][INFO] - Step 220, Loss: {'train/loss': 0.5034211874008179, 'train/grad_norm_pre_clip': 15.624155044555664, 'timing/data': 0.027913350611925125, 'timing/model': 2.7663886938244104, 'learning_rate/qwen_vl_interface': 9.98594860461494e-06, 'learning_rate/action_model': 9.98594860461494e-05, 'epoch': 0.45}) +[2026-06-19 16:43:52,518][starVLA.training.train_starvla][INFO] - Step 221, Loss: {'train/loss': 0.289277583360672, 'train/grad_norm_pre_clip': 6.672054290771484, 'timing/data': 0.0262681283056736, 'timing/model': 2.753267891705036, 'learning_rate/qwen_vl_interface': 9.985713556854461e-06, 'learning_rate/action_model': 9.98571355685446e-05, 'epoch': 0.45}) +[2026-06-19 16:44:03,423][starVLA.training.train_starvla][INFO] - Step 222, Loss: {'train/loss': 0.2566379904747009, 'train/grad_norm_pre_clip': 2.5838284492492676, 'timing/data': 0.029448289424180984, 'timing/model': 2.755535079166293, 'learning_rate/qwen_vl_interface': 9.985476562421024e-06, 'learning_rate/action_model': 9.985476562421024e-05, 'epoch': 0.46}) +[2026-06-19 16:44:14,215][starVLA.training.train_starvla][INFO] - Step 223, Loss: {'train/loss': 0.4044649600982666, 'train/grad_norm_pre_clip': 5.1862640380859375, 'timing/data': 0.02862315997481346, 'timing/model': 2.744358090683818, 'learning_rate/qwen_vl_interface': 9.985237621412047e-06, 'learning_rate/action_model': 9.985237621412046e-05, 'epoch': 0.46}) +[2026-06-19 16:44:25,022][starVLA.training.train_starvla][INFO] - Step 224, Loss: {'train/loss': 0.27304741740226746, 'train/grad_norm_pre_clip': 5.245738506317139, 'timing/data': 0.02802066132426262, 'timing/model': 2.750350685790181, 'learning_rate/qwen_vl_interface': 9.984996733925749e-06, 'learning_rate/action_model': 9.984996733925748e-05, 'epoch': 0.46}) +[2026-06-19 16:44:35,827][starVLA.training.train_starvla][INFO] - Step 225, Loss: {'train/loss': 0.7172096967697144, 'train/grad_norm_pre_clip': 17.071992874145508, 'timing/data': 0.027916260063648224, 'timing/model': 2.7604536693543196, 'learning_rate/qwen_vl_interface': 9.98475390006115e-06, 'learning_rate/action_model': 9.98475390006115e-05, 'epoch': 0.46}) +[2026-06-19 16:44:46,722][starVLA.training.train_starvla][INFO] - Step 226, Loss: {'train/loss': 0.5289368629455566, 'train/grad_norm_pre_clip': 8.709330558776855, 'timing/data': 0.02921300195157528, 'timing/model': 2.7481038067489862, 'learning_rate/qwen_vl_interface': 9.98450911991807e-06, 'learning_rate/action_model': 9.98450911991807e-05, 'epoch': 0.46}) +[2026-06-19 16:44:57,513][starVLA.training.train_starvla][INFO] - Step 227, Loss: {'train/loss': 0.41168615221977234, 'train/grad_norm_pre_clip': 5.767687797546387, 'timing/data': 0.028357213363051414, 'timing/model': 2.745194526389241, 'learning_rate/qwen_vl_interface': 9.98426239359713e-06, 'learning_rate/action_model': 9.984262393597129e-05, 'epoch': 0.47}) +[2026-06-19 16:45:08,323][starVLA.training.train_starvla][INFO] - Step 228, Loss: {'train/loss': 0.36657387018203735, 'train/grad_norm_pre_clip': 7.952930927276611, 'timing/data': 0.027283156290650368, 'timing/model': 2.7534146644175053, 'learning_rate/qwen_vl_interface': 9.984013721199747e-06, 'learning_rate/action_model': 9.984013721199747e-05, 'epoch': 0.47}) +[2026-06-19 16:45:19,120][starVLA.training.train_starvla][INFO] - Step 229, Loss: {'train/loss': 0.3759744465351105, 'train/grad_norm_pre_clip': 7.439939498901367, 'timing/data': 0.028635399416089058, 'timing/model': 2.752382043749094, 'learning_rate/qwen_vl_interface': 9.983763102828143e-06, 'learning_rate/action_model': 9.983763102828143e-05, 'epoch': 0.47}) +[2026-06-19 16:45:29,977][starVLA.training.train_starvla][INFO] - Step 230, Loss: {'train/loss': 0.5032948851585388, 'train/grad_norm_pre_clip': 6.01268196105957, 'timing/data': 0.028398193418979645, 'timing/model': 2.709077535197139, 'learning_rate/qwen_vl_interface': 9.983510538585336e-06, 'learning_rate/action_model': 9.983510538585337e-05, 'epoch': 0.47}) +[2026-06-19 16:45:40,724][starVLA.training.train_starvla][INFO] - Step 231, Loss: {'train/loss': 0.32767200469970703, 'train/grad_norm_pre_clip': 6.3004655838012695, 'timing/data': 0.028584832325577736, 'timing/model': 2.7348406985402107, 'learning_rate/qwen_vl_interface': 9.98325602857515e-06, 'learning_rate/action_model': 9.98325602857515e-05, 'epoch': 0.47}) +[2026-06-19 16:45:51,346][starVLA.training.train_starvla][INFO] - Step 232, Loss: {'train/loss': 0.18877671658992767, 'train/grad_norm_pre_clip': 3.1849069595336914, 'timing/data': 0.02874743938446045, 'timing/model': 2.661972312256694, 'learning_rate/qwen_vl_interface': 9.9829995729022e-06, 'learning_rate/action_model': 9.982999572902199e-05, 'epoch': 0.48}) +[2026-06-19 16:46:02,078][starVLA.training.train_starvla][INFO] - Step 233, Loss: {'train/loss': 0.31208333373069763, 'train/grad_norm_pre_clip': 5.042578220367432, 'timing/data': 0.027858871966600418, 'timing/model': 2.7022147476673126, 'learning_rate/qwen_vl_interface': 9.982741171671905e-06, 'learning_rate/action_model': 9.982741171671906e-05, 'epoch': 0.48}) +[2026-06-19 16:46:12,928][starVLA.training.train_starvla][INFO] - Step 234, Loss: {'train/loss': 0.3924262225627899, 'train/grad_norm_pre_clip': 8.514647483825684, 'timing/data': 0.027766302227973938, 'timing/model': 2.707292089238763, 'learning_rate/qwen_vl_interface': 9.982480824990489e-06, 'learning_rate/action_model': 9.982480824990488e-05, 'epoch': 0.48}) +[2026-06-19 16:46:23,613][starVLA.training.train_starvla][INFO] - Step 235, Loss: {'train/loss': 0.08957509696483612, 'train/grad_norm_pre_clip': 2.9527831077575684, 'timing/data': 0.0271841399371624, 'timing/model': 2.751454019919038, 'learning_rate/qwen_vl_interface': 9.982218532964963e-06, 'learning_rate/action_model': 9.982218532964964e-05, 'epoch': 0.48}) +[2026-06-19 16:46:34,166][starVLA.training.train_starvla][INFO] - Step 236, Loss: {'train/loss': 0.5091214776039124, 'train/grad_norm_pre_clip': 13.538387298583984, 'timing/data': 0.02753954567015171, 'timing/model': 2.6791934054344893, 'learning_rate/qwen_vl_interface': 9.981954295703152e-06, 'learning_rate/action_model': 9.981954295703152e-05, 'epoch': 0.48}) +[2026-06-19 16:46:44,736][starVLA.training.train_starvla][INFO] - Step 237, Loss: {'train/loss': 0.15012460947036743, 'train/grad_norm_pre_clip': 8.372705459594727, 'timing/data': 0.027788901701569557, 'timing/model': 2.7126544900238514, 'learning_rate/qwen_vl_interface': 9.981688113313672e-06, 'learning_rate/action_model': 9.981688113313672e-05, 'epoch': 0.49}) +[2026-06-19 16:46:55,401][starVLA.training.train_starvla][INFO] - Step 238, Loss: {'train/loss': 0.14387479424476624, 'train/grad_norm_pre_clip': 3.522040605545044, 'timing/data': 0.02830163575708866, 'timing/model': 2.710457531735301, 'learning_rate/qwen_vl_interface': 9.98141998590594e-06, 'learning_rate/action_model': 9.981419985905939e-05, 'epoch': 0.49}) +[2026-06-19 16:47:05,959][starVLA.training.train_starvla][INFO] - Step 239, Loss: {'train/loss': 0.14454860985279083, 'train/grad_norm_pre_clip': 4.2317633628845215, 'timing/data': 0.027671193704009056, 'timing/model': 2.672554997727275, 'learning_rate/qwen_vl_interface': 9.981149913590171e-06, 'learning_rate/action_model': 9.981149913590172e-05, 'epoch': 0.49}) +[2026-06-19 16:47:16,827][starVLA.training.train_starvla][INFO] - Step 240, Loss: {'train/loss': 0.5751661658287048, 'train/grad_norm_pre_clip': 20.004634857177734, 'timing/data': 0.026967788115143776, 'timing/model': 2.757739584892988, 'learning_rate/qwen_vl_interface': 9.980877896477386e-06, 'learning_rate/action_model': 9.980877896477386e-05, 'epoch': 0.49}) +[2026-06-19 16:47:27,624][starVLA.training.train_starvla][INFO] - Step 241, Loss: {'train/loss': 0.22173647582530975, 'train/grad_norm_pre_clip': 9.621342658996582, 'timing/data': 0.027745531871914864, 'timing/model': 2.763225795701146, 'learning_rate/qwen_vl_interface': 9.980603934679397e-06, 'learning_rate/action_model': 9.980603934679398e-05, 'epoch': 0.49}) +[2026-06-19 16:47:38,353][starVLA.training.train_starvla][INFO] - Step 242, Loss: {'train/loss': 0.3564516007900238, 'train/grad_norm_pre_clip': 4.6104960441589355, 'timing/data': 0.030280208215117455, 'timing/model': 2.6805585995316505, 'learning_rate/qwen_vl_interface': 9.980328028308824e-06, 'learning_rate/action_model': 9.980328028308824e-05, 'epoch': 0.5}) +[2026-06-19 16:47:49,115][starVLA.training.train_starvla][INFO] - Step 243, Loss: {'train/loss': 0.3625032603740692, 'train/grad_norm_pre_clip': 11.246158599853516, 'timing/data': 0.036518730223178864, 'timing/model': 2.7212023343890905, 'learning_rate/qwen_vl_interface': 9.980050177479075e-06, 'learning_rate/action_model': 9.980050177479076e-05, 'epoch': 0.5}) +[2026-06-19 16:48:00,097][starVLA.training.train_starvla][INFO] - Step 244, Loss: {'train/loss': 0.22516541182994843, 'train/grad_norm_pre_clip': 7.558163166046143, 'timing/data': 0.03710746392607689, 'timing/model': 2.749689105898142, 'learning_rate/qwen_vl_interface': 9.97977038230437e-06, 'learning_rate/action_model': 9.97977038230437e-05, 'epoch': 0.5}) +[2026-06-19 16:48:10,920][starVLA.training.train_starvla][INFO] - Step 245, Loss: {'train/loss': 0.2473784238100052, 'train/grad_norm_pre_clip': 7.970814228057861, 'timing/data': 0.03817902319133282, 'timing/model': 2.741717530414462, 'learning_rate/qwen_vl_interface': 9.97948864289972e-06, 'learning_rate/action_model': 9.97948864289972e-05, 'epoch': 0.5}) +[2026-06-19 16:48:21,722][starVLA.training.train_starvla][INFO] - Step 246, Loss: {'train/loss': 0.20805741846561432, 'train/grad_norm_pre_clip': 2.9238619804382324, 'timing/data': 0.02513529360294342, 'timing/model': 2.765618335455656, 'learning_rate/qwen_vl_interface': 9.979204959380936e-06, 'learning_rate/action_model': 9.979204959380936e-05, 'epoch': 0.51}) +[2026-06-19 16:48:32,514][starVLA.training.train_starvla][INFO] - Step 247, Loss: {'train/loss': 0.41083380579948425, 'train/grad_norm_pre_clip': 10.532510757446289, 'timing/data': 0.025088269263505936, 'timing/model': 2.751701733097434, 'learning_rate/qwen_vl_interface': 9.978919331864631e-06, 'learning_rate/action_model': 9.97891933186463e-05, 'epoch': 0.51}) +[2026-06-19 16:48:43,313][starVLA.training.train_starvla][INFO] - Step 248, Loss: {'train/loss': 0.34155982732772827, 'train/grad_norm_pre_clip': 5.31627082824707, 'timing/data': 0.029238024726510048, 'timing/model': 2.750352757051587, 'learning_rate/qwen_vl_interface': 9.978631760468216e-06, 'learning_rate/action_model': 9.978631760468216e-05, 'epoch': 0.51}) +[2026-06-19 16:48:54,122][starVLA.training.train_starvla][INFO] - Step 249, Loss: {'train/loss': 0.4204465448856354, 'train/grad_norm_pre_clip': 7.156818389892578, 'timing/data': 0.025161493569612503, 'timing/model': 2.7614192347973585, 'learning_rate/qwen_vl_interface': 9.9783422453099e-06, 'learning_rate/action_model': 9.9783422453099e-05, 'epoch': 0.51}) +[2026-06-19 16:49:04,922][starVLA.training.train_starvla][INFO] - Step 250, Loss: {'train/loss': 0.26498299837112427, 'train/grad_norm_pre_clip': 5.533157825469971, 'timing/data': 0.02481113001704216, 'timing/model': 2.7789121326059103, 'learning_rate/qwen_vl_interface': 9.978050786508693e-06, 'learning_rate/action_model': 9.978050786508692e-05, 'epoch': 0.51}) +[2026-06-19 16:49:15,623][starVLA.training.train_starvla][INFO] - Step 251, Loss: {'train/loss': 0.2280978262424469, 'train/grad_norm_pre_clip': 4.167520046234131, 'timing/data': 0.024758808314800262, 'timing/model': 2.745280595496297, 'learning_rate/qwen_vl_interface': 9.9777573841844e-06, 'learning_rate/action_model': 9.977757384184401e-05, 'epoch': 0.52}) +[2026-06-19 16:49:26,315][starVLA.training.train_starvla][INFO] - Step 252, Loss: {'train/loss': 0.2686191499233246, 'train/grad_norm_pre_clip': 4.626786231994629, 'timing/data': 0.02476695366203785, 'timing/model': 2.764582108706236, 'learning_rate/qwen_vl_interface': 9.97746203845763e-06, 'learning_rate/action_model': 9.97746203845763e-05, 'epoch': 0.52}) +[2026-06-19 16:49:37,030][starVLA.training.train_starvla][INFO] - Step 253, Loss: {'train/loss': 0.2483784258365631, 'train/grad_norm_pre_clip': 4.598902702331543, 'timing/data': 0.025104479864239693, 'timing/model': 2.7101229280233383, 'learning_rate/qwen_vl_interface': 9.97716474944979e-06, 'learning_rate/action_model': 9.977164749449788e-05, 'epoch': 0.52}) +[2026-06-19 16:49:47,924][starVLA.training.train_starvla][INFO] - Step 254, Loss: {'train/loss': 0.16467058658599854, 'train/grad_norm_pre_clip': 3.132411003112793, 'timing/data': 0.03187369368970394, 'timing/model': 2.7715731244534254, 'learning_rate/qwen_vl_interface': 9.976865517283079e-06, 'learning_rate/action_model': 9.976865517283078e-05, 'epoch': 0.52}) +[2026-06-19 16:49:58,615][starVLA.training.train_starvla][INFO] - Step 255, Loss: {'train/loss': 0.11976176500320435, 'train/grad_norm_pre_clip': 1.774872064590454, 'timing/data': 0.025685491040349007, 'timing/model': 2.73117315210402, 'learning_rate/qwen_vl_interface': 9.976564342080505e-06, 'learning_rate/action_model': 9.976564342080504e-05, 'epoch': 0.52}) +[2026-06-19 16:50:09,413][starVLA.training.train_starvla][INFO] - Step 256, Loss: {'train/loss': 0.3679201900959015, 'train/grad_norm_pre_clip': 9.487663269042969, 'timing/data': 0.02460586279630661, 'timing/model': 2.7610592246055603, 'learning_rate/qwen_vl_interface': 9.976261223965867e-06, 'learning_rate/action_model': 9.976261223965865e-05, 'epoch': 0.53}) +[2026-06-19 16:50:20,168][starVLA.training.train_starvla][INFO] - Step 257, Loss: {'train/loss': 0.1849796622991562, 'train/grad_norm_pre_clip': 5.928753852844238, 'timing/data': 0.024995898827910423, 'timing/model': 2.7069375533610582, 'learning_rate/qwen_vl_interface': 9.975956163063767e-06, 'learning_rate/action_model': 9.975956163063767e-05, 'epoch': 0.53}) +[2026-06-19 16:50:30,746][starVLA.training.train_starvla][INFO] - Step 258, Loss: {'train/loss': 0.364168643951416, 'train/grad_norm_pre_clip': 13.401302337646484, 'timing/data': 0.024953022599220276, 'timing/model': 2.6935855969786644, 'learning_rate/qwen_vl_interface': 9.975649159499605e-06, 'learning_rate/action_model': 9.975649159499603e-05, 'epoch': 0.53}) +[2026-06-19 16:50:41,341][starVLA.training.train_starvla][INFO] - Step 259, Loss: {'train/loss': 0.3050956726074219, 'train/grad_norm_pre_clip': 6.1946892738342285, 'timing/data': 0.025110622867941856, 'timing/model': 2.659030247479677, 'learning_rate/qwen_vl_interface': 9.975340213399573e-06, 'learning_rate/action_model': 9.975340213399573e-05, 'epoch': 0.53}) +[2026-06-19 16:50:51,956][starVLA.training.train_starvla][INFO] - Step 260, Loss: {'train/loss': 0.10888149589300156, 'train/grad_norm_pre_clip': 9.204935073852539, 'timing/data': 0.024772971868515015, 'timing/model': 2.7102768681943417, 'learning_rate/qwen_vl_interface': 9.975029324890674e-06, 'learning_rate/action_model': 9.975029324890674e-05, 'epoch': 0.53}) +[2026-06-19 16:51:02,620][starVLA.training.train_starvla][INFO] - Step 261, Loss: {'train/loss': 0.551238477230072, 'train/grad_norm_pre_clip': 20.4012393951416, 'timing/data': 0.025036007165908813, 'timing/model': 2.7798324059695005, 'learning_rate/qwen_vl_interface': 9.974716494100698e-06, 'learning_rate/action_model': 9.974716494100697e-05, 'epoch': 0.54}) +[2026-06-19 16:51:13,394][starVLA.training.train_starvla][INFO] - Step 262, Loss: {'train/loss': 0.5950942635536194, 'train/grad_norm_pre_clip': 36.405433654785156, 'timing/data': 0.024574315175414085, 'timing/model': 2.743707451969385, 'learning_rate/qwen_vl_interface': 9.97440172115824e-06, 'learning_rate/action_model': 9.974401721158239e-05, 'epoch': 0.54}) +[2026-06-19 16:51:23,915][starVLA.training.train_starvla][INFO] - Step 263, Loss: {'train/loss': 0.2140079289674759, 'train/grad_norm_pre_clip': 8.809545516967773, 'timing/data': 0.024654923006892204, 'timing/model': 2.7351194228976965, 'learning_rate/qwen_vl_interface': 9.97408500619269e-06, 'learning_rate/action_model': 9.97408500619269e-05, 'epoch': 0.54}) +[2026-06-19 16:51:34,624][starVLA.training.train_starvla][INFO] - Step 264, Loss: {'train/loss': 0.40735453367233276, 'train/grad_norm_pre_clip': 10.965362548828125, 'timing/data': 0.025157548487186432, 'timing/model': 2.74282736890018, 'learning_rate/qwen_vl_interface': 9.973766349334239e-06, 'learning_rate/action_model': 9.973766349334239e-05, 'epoch': 0.54}) +[2026-06-19 16:51:45,414][starVLA.training.train_starvla][INFO] - Step 265, Loss: {'train/loss': 0.2761439085006714, 'train/grad_norm_pre_clip': 8.71479606628418, 'timing/data': 0.024140313267707825, 'timing/model': 2.74881162494421, 'learning_rate/qwen_vl_interface': 9.973445750713875e-06, 'learning_rate/action_model': 9.973445750713875e-05, 'epoch': 0.54}) +[2026-06-19 16:51:56,116][starVLA.training.train_starvla][INFO] - Step 266, Loss: {'train/loss': 0.21103425323963165, 'train/grad_norm_pre_clip': 6.994565010070801, 'timing/data': 0.02433905005455017, 'timing/model': 2.729257693514228, 'learning_rate/qwen_vl_interface': 9.97312321046338e-06, 'learning_rate/action_model': 9.97312321046338e-05, 'epoch': 0.55}) +[2026-06-19 16:52:06,822][starVLA.training.train_starvla][INFO] - Step 267, Loss: {'train/loss': 0.28025591373443604, 'train/grad_norm_pre_clip': 13.807369232177734, 'timing/data': 0.024548422545194626, 'timing/model': 2.734622772783041, 'learning_rate/qwen_vl_interface': 9.972798728715343e-06, 'learning_rate/action_model': 9.972798728715344e-05, 'epoch': 0.55}) +[2026-06-19 16:52:17,622][starVLA.training.train_starvla][INFO] - Step 268, Loss: {'train/loss': 0.4473382234573364, 'train/grad_norm_pre_clip': 13.748684883117676, 'timing/data': 0.02478746883571148, 'timing/model': 2.7665868550539017, 'learning_rate/qwen_vl_interface': 9.972472305603145e-06, 'learning_rate/action_model': 9.972472305603144e-05, 'epoch': 0.55}) +[2026-06-19 16:52:28,347][starVLA.training.train_starvla][INFO] - Step 269, Loss: {'train/loss': 0.3865281343460083, 'train/grad_norm_pre_clip': 13.554683685302734, 'timing/data': 0.02447197027504444, 'timing/model': 2.685180328786373, 'learning_rate/qwen_vl_interface': 9.972143941260965e-06, 'learning_rate/action_model': 9.972143941260965e-05, 'epoch': 0.55}) +[2026-06-19 16:52:38,943][starVLA.training.train_starvla][INFO] - Step 270, Loss: {'train/loss': 0.34532690048217773, 'train/grad_norm_pre_clip': 10.081531524658203, 'timing/data': 0.025485031306743622, 'timing/model': 2.66139917075634, 'learning_rate/qwen_vl_interface': 9.971813635823781e-06, 'learning_rate/action_model': 9.971813635823781e-05, 'epoch': 0.55}) +[2026-06-19 16:52:49,716][starVLA.training.train_starvla][INFO] - Step 271, Loss: {'train/loss': 0.3051231801509857, 'train/grad_norm_pre_clip': 12.251531600952148, 'timing/data': 0.02521948330104351, 'timing/model': 2.7678025998175144, 'learning_rate/qwen_vl_interface': 9.971481389427372e-06, 'learning_rate/action_model': 9.971481389427371e-05, 'epoch': 0.56}) +[2026-06-19 16:53:00,342][starVLA.training.train_starvla][INFO] - Step 272, Loss: {'train/loss': 0.38970237970352173, 'train/grad_norm_pre_clip': 10.08929443359375, 'timing/data': 0.02390860766172409, 'timing/model': 2.6620124634355307, 'learning_rate/qwen_vl_interface': 9.971147202208308e-06, 'learning_rate/action_model': 9.971147202208308e-05, 'epoch': 0.56}) +[2026-06-19 16:53:10,940][starVLA.training.train_starvla][INFO] - Step 273, Loss: {'train/loss': 0.28397002816200256, 'train/grad_norm_pre_clip': 10.760152816772461, 'timing/data': 0.024810083210468292, 'timing/model': 2.6850306186825037, 'learning_rate/qwen_vl_interface': 9.970811074303962e-06, 'learning_rate/action_model': 9.970811074303962e-05, 'epoch': 0.56}) +[2026-06-19 16:53:21,541][starVLA.training.train_starvla][INFO] - Step 274, Loss: {'train/loss': 0.6443268060684204, 'train/grad_norm_pre_clip': 11.400175094604492, 'timing/data': 0.02476660907268524, 'timing/model': 2.6601641681045294, 'learning_rate/qwen_vl_interface': 9.970473005852505e-06, 'learning_rate/action_model': 9.970473005852505e-05, 'epoch': 0.56}) +[2026-06-19 16:53:32,060][starVLA.training.train_starvla][INFO] - Step 275, Loss: {'train/loss': 0.3461037278175354, 'train/grad_norm_pre_clip': 4.520620822906494, 'timing/data': 0.02396981418132782, 'timing/model': 2.680232724174857, 'learning_rate/qwen_vl_interface': 9.970132996992904e-06, 'learning_rate/action_model': 9.970132996992904e-05, 'epoch': 0.56}) +[2026-06-19 16:53:42,537][starVLA.training.train_starvla][INFO] - Step 276, Loss: {'train/loss': 0.5407920479774475, 'train/grad_norm_pre_clip': 10.889683723449707, 'timing/data': 0.02464163862168789, 'timing/model': 2.6591184493154287, 'learning_rate/qwen_vl_interface': 9.969791047864921e-06, 'learning_rate/action_model': 9.969791047864921e-05, 'epoch': 0.57}) +[2026-06-19 16:53:53,040][starVLA.training.train_starvla][INFO] - Step 277, Loss: {'train/loss': 0.22613640129566193, 'train/grad_norm_pre_clip': 3.2001943588256836, 'timing/data': 0.024421347305178642, 'timing/model': 2.658170899376273, 'learning_rate/qwen_vl_interface': 9.969447158609122e-06, 'learning_rate/action_model': 9.96944715860912e-05, 'epoch': 0.57}) +[2026-06-19 16:54:03,821][starVLA.training.train_starvla][INFO] - Step 278, Loss: {'train/loss': 0.33439013361930847, 'train/grad_norm_pre_clip': 3.799901008605957, 'timing/data': 0.025218788534402847, 'timing/model': 2.7574162762612104, 'learning_rate/qwen_vl_interface': 9.969101329366864e-06, 'learning_rate/action_model': 9.969101329366863e-05, 'epoch': 0.57}) +[2026-06-19 16:54:14,623][starVLA.training.train_starvla][INFO] - Step 279, Loss: {'train/loss': 0.25564295053482056, 'train/grad_norm_pre_clip': 4.447198867797852, 'timing/data': 0.024845369160175323, 'timing/model': 2.768640147522092, 'learning_rate/qwen_vl_interface': 9.968753560280305e-06, 'learning_rate/action_model': 9.968753560280304e-05, 'epoch': 0.57}) +[2026-06-19 16:54:25,421][starVLA.training.train_starvla][INFO] - Step 280, Loss: {'train/loss': 0.328645795583725, 'train/grad_norm_pre_clip': 4.2406439781188965, 'timing/data': 0.02471660077571869, 'timing/model': 2.7589716743677855, 'learning_rate/qwen_vl_interface': 9.968403851492402e-06, 'learning_rate/action_model': 9.968403851492401e-05, 'epoch': 0.58}) +[2026-06-19 16:54:36,210][starVLA.training.train_starvla][INFO] - Step 281, Loss: {'train/loss': 0.2495889514684677, 'train/grad_norm_pre_clip': 6.4293742179870605, 'timing/data': 0.024773817509412766, 'timing/model': 2.757402354851365, 'learning_rate/qwen_vl_interface': 9.968052203146904e-06, 'learning_rate/action_model': 9.968052203146903e-05, 'epoch': 0.58}) +[2026-06-19 16:54:46,924][starVLA.training.train_starvla][INFO] - Step 282, Loss: {'train/loss': 0.29618215560913086, 'train/grad_norm_pre_clip': 5.107324600219727, 'timing/data': 0.024907665327191353, 'timing/model': 2.765978528186679, 'learning_rate/qwen_vl_interface': 9.967698615388363e-06, 'learning_rate/action_model': 9.967698615388362e-05, 'epoch': 0.58}) +[2026-06-19 16:54:57,621][starVLA.training.train_starvla][INFO] - Step 283, Loss: {'train/loss': 0.07288990914821625, 'train/grad_norm_pre_clip': 4.540794849395752, 'timing/data': 0.018534613773226738, 'timing/model': 2.749843480065465, 'learning_rate/qwen_vl_interface': 9.967343088362123e-06, 'learning_rate/action_model': 9.967343088362122e-05, 'epoch': 0.58}) +[2026-06-19 16:55:08,315][starVLA.training.train_starvla][INFO] - Step 284, Loss: {'train/loss': 0.26435840129852295, 'train/grad_norm_pre_clip': 5.256845951080322, 'timing/data': 0.018346745520830154, 'timing/model': 2.7423221841454506, 'learning_rate/qwen_vl_interface': 9.966985622214328e-06, 'learning_rate/action_model': 9.966985622214329e-05, 'epoch': 0.58}) +[2026-06-19 16:55:19,012][starVLA.training.train_starvla][INFO] - Step 285, Loss: {'train/loss': 0.13639381527900696, 'train/grad_norm_pre_clip': 9.183269500732422, 'timing/data': 0.019190004095435143, 'timing/model': 2.7390620447695255, 'learning_rate/qwen_vl_interface': 9.966626217091922e-06, 'learning_rate/action_model': 9.966626217091922e-05, 'epoch': 0.59}) +[2026-06-19 16:55:29,818][starVLA.training.train_starvla][INFO] - Step 286, Loss: {'train/loss': 0.3011937141418457, 'train/grad_norm_pre_clip': 4.375170707702637, 'timing/data': 0.01892310567200184, 'timing/model': 2.7568173464387655, 'learning_rate/qwen_vl_interface': 9.966264873142641e-06, 'learning_rate/action_model': 9.96626487314264e-05, 'epoch': 0.59}) +[2026-06-19 16:55:40,623][starVLA.training.train_starvla][INFO] - Step 287, Loss: {'train/loss': 0.3720093071460724, 'train/grad_norm_pre_clip': 9.480889320373535, 'timing/data': 0.02182215265929699, 'timing/model': 2.7603103276342154, 'learning_rate/qwen_vl_interface': 9.965901590515018e-06, 'learning_rate/action_model': 9.965901590515018e-05, 'epoch': 0.59}) +[2026-06-19 16:55:51,429][starVLA.training.train_starvla][INFO] - Step 288, Loss: {'train/loss': 0.31748607754707336, 'train/grad_norm_pre_clip': 9.951638221740723, 'timing/data': 0.026339204981923103, 'timing/model': 2.7618885971605778, 'learning_rate/qwen_vl_interface': 9.965536369358386e-06, 'learning_rate/action_model': 9.965536369358386e-05, 'epoch': 0.59}) +[2026-06-19 16:56:02,211][starVLA.training.train_starvla][INFO] - Step 289, Loss: {'train/loss': 0.22220729291439056, 'train/grad_norm_pre_clip': 9.867378234863281, 'timing/data': 0.02103562466800213, 'timing/model': 2.762203959748149, 'learning_rate/qwen_vl_interface': 9.965169209822874e-06, 'learning_rate/action_model': 9.965169209822874e-05, 'epoch': 0.59}) +[2026-06-19 16:56:13,223][starVLA.training.train_starvla][INFO] - Step 290, Loss: {'train/loss': 0.34745922684669495, 'train/grad_norm_pre_clip': 10.566567420959473, 'timing/data': 0.02039075829088688, 'timing/model': 2.7721532844007015, 'learning_rate/qwen_vl_interface': 9.96480011205941e-06, 'learning_rate/action_model': 9.964800112059409e-05, 'epoch': 0.6}) +[2026-06-19 16:56:23,864][starVLA.training.train_starvla][INFO] - Step 291, Loss: {'train/loss': 0.28516507148742676, 'train/grad_norm_pre_clip': 7.008796691894531, 'timing/data': 0.018420256674289703, 'timing/model': 2.658810505643487, 'learning_rate/qwen_vl_interface': 9.964429076219711e-06, 'learning_rate/action_model': 9.964429076219711e-05, 'epoch': 0.6}) +[2026-06-19 16:56:34,726][starVLA.training.train_starvla][INFO] - Step 292, Loss: {'train/loss': 0.21059779822826385, 'train/grad_norm_pre_clip': 10.034452438354492, 'timing/data': 0.03025425225496292, 'timing/model': 2.751010099425912, 'learning_rate/qwen_vl_interface': 9.964056102456299e-06, 'learning_rate/action_model': 9.964056102456299e-05, 'epoch': 0.6}) +[2026-06-19 16:56:45,535][starVLA.training.train_starvla][INFO] - Step 293, Loss: {'train/loss': 0.23004066944122314, 'train/grad_norm_pre_clip': 5.332183837890625, 'timing/data': 0.02014482580125332, 'timing/model': 2.7896196600049734, 'learning_rate/qwen_vl_interface': 9.963681190922489e-06, 'learning_rate/action_model': 9.963681190922488e-05, 'epoch': 0.6}) +[2026-06-19 16:56:56,412][starVLA.training.train_starvla][INFO] - Step 294, Loss: {'train/loss': 0.33346614241600037, 'train/grad_norm_pre_clip': 6.212222099304199, 'timing/data': 0.020702091977000237, 'timing/model': 2.7720476519316435, 'learning_rate/qwen_vl_interface': 9.963304341772391e-06, 'learning_rate/action_model': 9.963304341772391e-05, 'epoch': 0.6}) +[2026-06-19 16:57:07,223][starVLA.training.train_starvla][INFO] - Step 295, Loss: {'train/loss': 0.3179146647453308, 'train/grad_norm_pre_clip': 7.932426929473877, 'timing/data': 0.020528191700577736, 'timing/model': 2.7748608738183975, 'learning_rate/qwen_vl_interface': 9.962925555160919e-06, 'learning_rate/action_model': 9.962925555160918e-05, 'epoch': 0.61}) +[2026-06-19 16:57:18,020][starVLA.training.train_starvla][INFO] - Step 296, Loss: {'train/loss': 0.4111824631690979, 'train/grad_norm_pre_clip': 12.637194633483887, 'timing/data': 0.02215026505291462, 'timing/model': 2.761319901794195, 'learning_rate/qwen_vl_interface': 9.96254483124377e-06, 'learning_rate/action_model': 9.96254483124377e-05, 'epoch': 0.61}) +[2026-06-19 16:57:28,722][starVLA.training.train_starvla][INFO] - Step 297, Loss: {'train/loss': 0.32587477564811707, 'train/grad_norm_pre_clip': 7.880000114440918, 'timing/data': 0.021388504654169083, 'timing/model': 2.7422010712325573, 'learning_rate/qwen_vl_interface': 9.96216217017745e-06, 'learning_rate/action_model': 9.962162170177451e-05, 'epoch': 0.61}) +[2026-06-19 16:57:39,516][starVLA.training.train_starvla][INFO] - Step 298, Loss: {'train/loss': 0.23712508380413055, 'train/grad_norm_pre_clip': 7.702201843261719, 'timing/data': 0.026245616376399994, 'timing/model': 2.7524723932147026, 'learning_rate/qwen_vl_interface': 9.961777572119258e-06, 'learning_rate/action_model': 9.961777572119258e-05, 'epoch': 0.61}) +[2026-06-19 16:57:50,311][starVLA.training.train_starvla][INFO] - Step 299, Loss: {'train/loss': 0.36964911222457886, 'train/grad_norm_pre_clip': 3.2596275806427, 'timing/data': 0.019185231998562813, 'timing/model': 2.761995308101177, 'learning_rate/qwen_vl_interface': 9.961391037227286e-06, 'learning_rate/action_model': 9.961391037227284e-05, 'epoch': 0.61}) +[2026-06-19 16:58:01,124][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 300 +[2026-06-19 17:05:55,813][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 300 +[2026-06-19 17:05:55,814][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 300 +[2026-06-19 17:14:29,819][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 300 +[2026-06-19 17:14:29,821][starVLA.training.train_starvla][INFO] - Step 300, Loss: {'train/loss': 0.3593957722187042, 'train/grad_norm_pre_clip': 3.2650222778320312, 'eval/loss': 0.3182315528392792, 'eval/latency_0/loss': 0.1624845713376999, 'eval/latency_1/loss': 0.2307625114917755, 'eval/latency_2/loss': 0.38436880707740784, 'eval/latency_3/loss': 0.361742228269577, 'eval/latency_4/loss': 0.44240981340408325, 'eval/flappy/loss': 0.3182315528392792, 'eval/flappy/latency_0/loss': 0.1624845713376999, 'eval/flappy/latency_1/loss': 0.2307625114917755, 'eval/flappy/latency_2/loss': 0.38436880707740784, 'eval/flappy/latency_3/loss': 0.361742228269577, 'eval/flappy/latency_4/loss': 0.44240981340408325, 'eval/flap_precision': 0.6694560669456067, 'eval/flap_recall': 0.5818181818181818, 'eval/flap_f1': 0.622568093385214, 'eval/flap_rate_model': 0.06638888888888889, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.622568093385214, 'eval/latency_0/flap_precision': 0.6694560669456067, 'eval/latency_0/flap_recall': 0.5818181818181818, 'eval/latency_0/flap_f1': 0.622568093385214, 'eval/latency_0/flap_rate_model': 0.06638888888888889, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.622568093385214, 'timing/data': 0.018904779106378555, 'timing/model': 2.7927132844924927, 'learning_rate/qwen_vl_interface': 9.961002565660419e-06, 'learning_rate/action_model': 9.961002565660419e-05, 'epoch': 0.62}) +[2026-06-19 17:14:29,822][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_300_state +[2026-06-19 17:14:29,848][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-19 17:14:35,633][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_300_state/model.safetensors +[2026-06-19 17:14:42,815][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_300_state/optimizer.bin +[2026-06-19 17:14:42,816][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_300_state/sampler.bin +[2026-06-19 17:14:42,816][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_300_state/sampler_1.bin +[2026-06-19 17:14:42,818][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_300_state/random_states_0.pkl +[2026-06-19 17:14:43,028][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 300 +[2026-06-19 17:14:43,028][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-19 17:14:43,043][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-19 17:14:53,642][starVLA.training.train_starvla][INFO] - Step 301, Loss: {'train/loss': 0.27709153294563293, 'train/grad_norm_pre_clip': 5.500699996948242, 'timing/data': 0.01955350674688816, 'timing/model': 2.6981573533266783, 'learning_rate/qwen_vl_interface': 9.96061215757835e-06, 'learning_rate/action_model': 9.960612157578349e-05, 'epoch': 0.62}) +[2026-06-19 17:15:04,501][starVLA.training.train_starvla][INFO] - Step 302, Loss: {'train/loss': 0.3034593164920807, 'train/grad_norm_pre_clip': 3.592505693435669, 'timing/data': 0.0191886518150568, 'timing/model': 2.7332127578556538, 'learning_rate/qwen_vl_interface': 9.96021981314156e-06, 'learning_rate/action_model': 9.96021981314156e-05, 'epoch': 0.62}) +[2026-06-19 17:15:14,930][starVLA.training.train_starvla][INFO] - Step 303, Loss: {'train/loss': 0.31527549028396606, 'train/grad_norm_pre_clip': 5.929819107055664, 'timing/data': 0.018667427822947502, 'timing/model': 2.664628593251109, 'learning_rate/qwen_vl_interface': 9.959825532511322e-06, 'learning_rate/action_model': 9.959825532511322e-05, 'epoch': 0.62}) +[2026-06-19 17:15:25,439][starVLA.training.train_starvla][INFO] - Step 304, Loss: {'train/loss': 0.18740446865558624, 'train/grad_norm_pre_clip': 3.901078224182129, 'timing/data': 0.02545289695262909, 'timing/model': 2.663080992177129, 'learning_rate/qwen_vl_interface': 9.959429315849715e-06, 'learning_rate/action_model': 9.959429315849715e-05, 'epoch': 0.62}) +[2026-06-19 17:15:36,029][starVLA.training.train_starvla][INFO] - Step 305, Loss: {'train/loss': 0.24223367869853973, 'train/grad_norm_pre_clip': 5.713326930999756, 'timing/data': 0.019722186028957367, 'timing/model': 2.6622545775026083, 'learning_rate/qwen_vl_interface': 9.959031163319606e-06, 'learning_rate/action_model': 9.959031163319607e-05, 'epoch': 0.63}) +[2026-06-19 17:15:46,534][starVLA.training.train_starvla][INFO] - Step 306, Loss: {'train/loss': 0.3599264919757843, 'train/grad_norm_pre_clip': 10.56810188293457, 'timing/data': 0.018570391461253166, 'timing/model': 2.6646230462938547, 'learning_rate/qwen_vl_interface': 9.958631075084664e-06, 'learning_rate/action_model': 9.958631075084664e-05, 'epoch': 0.63}) +[2026-06-19 17:15:57,212][starVLA.training.train_starvla][INFO] - Step 307, Loss: {'train/loss': 0.29761388897895813, 'train/grad_norm_pre_clip': 4.773687839508057, 'timing/data': 0.01867920719087124, 'timing/model': 2.7420943025499582, 'learning_rate/qwen_vl_interface': 9.958229051309346e-06, 'learning_rate/action_model': 9.958229051309346e-05, 'epoch': 0.63}) +[2026-06-19 17:16:07,847][starVLA.training.train_starvla][INFO] - Step 308, Loss: {'train/loss': 0.3586433231830597, 'train/grad_norm_pre_clip': 10.918169021606445, 'timing/data': 0.018268311396241188, 'timing/model': 2.6799419820308685, 'learning_rate/qwen_vl_interface': 9.95782509215891e-06, 'learning_rate/action_model': 9.95782509215891e-05, 'epoch': 0.63}) +[2026-06-19 17:16:18,401][starVLA.training.train_starvla][INFO] - Step 309, Loss: {'train/loss': 0.2474258989095688, 'train/grad_norm_pre_clip': 5.137218952178955, 'timing/data': 0.018404021859169006, 'timing/model': 2.689106846228242, 'learning_rate/qwen_vl_interface': 9.95741919779941e-06, 'learning_rate/action_model': 9.95741919779941e-05, 'epoch': 0.63}) +[2026-06-19 17:16:28,833][starVLA.training.train_starvla][INFO] - Step 310, Loss: {'train/loss': 0.49056652188301086, 'train/grad_norm_pre_clip': 8.592036247253418, 'timing/data': 0.018485188484191895, 'timing/model': 2.6621045768260956, 'learning_rate/qwen_vl_interface': 9.957011368397692e-06, 'learning_rate/action_model': 9.95701136839769e-05, 'epoch': 0.64}) +[2026-06-19 17:16:39,431][starVLA.training.train_starvla][INFO] - Step 311, Loss: {'train/loss': 0.3336552381515503, 'train/grad_norm_pre_clip': 11.496557235717773, 'timing/data': 0.018596375361084938, 'timing/model': 2.6953318286687136, 'learning_rate/qwen_vl_interface': 9.956601604121398e-06, 'learning_rate/action_model': 9.956601604121397e-05, 'epoch': 0.64}) +[2026-06-19 17:16:49,938][starVLA.training.train_starvla][INFO] - Step 312, Loss: {'train/loss': 0.259672611951828, 'train/grad_norm_pre_clip': 9.294967651367188, 'timing/data': 0.019011421129107475, 'timing/model': 2.6647625993937254, 'learning_rate/qwen_vl_interface': 9.95618990513897e-06, 'learning_rate/action_model': 9.95618990513897e-05, 'epoch': 0.64}) +[2026-06-19 17:17:00,438][starVLA.training.train_starvla][INFO] - Step 313, Loss: {'train/loss': 0.2694905698299408, 'train/grad_norm_pre_clip': 5.999553203582764, 'timing/data': 0.01852451078593731, 'timing/model': 2.663400111719966, 'learning_rate/qwen_vl_interface': 9.95577627161964e-06, 'learning_rate/action_model': 9.95577627161964e-05, 'epoch': 0.64}) +[2026-06-19 17:17:11,038][starVLA.training.train_starvla][INFO] - Step 314, Loss: {'train/loss': 0.2303234338760376, 'train/grad_norm_pre_clip': 4.923952102661133, 'timing/data': 0.019011331722140312, 'timing/model': 2.665385005995631, 'learning_rate/qwen_vl_interface': 9.955360703733438e-06, 'learning_rate/action_model': 9.955360703733437e-05, 'epoch': 0.64}) +[2026-06-19 17:17:21,539][starVLA.training.train_starvla][INFO] - Step 315, Loss: {'train/loss': 0.31279730796813965, 'train/grad_norm_pre_clip': 7.550579071044922, 'timing/data': 0.01879340410232544, 'timing/model': 2.666637111455202, 'learning_rate/qwen_vl_interface': 9.954943201651187e-06, 'learning_rate/action_model': 9.954943201651187e-05, 'epoch': 0.65}) +[2026-06-19 17:17:32,035][starVLA.training.train_starvla][INFO] - Step 316, Loss: {'train/loss': 0.30702871084213257, 'train/grad_norm_pre_clip': 9.533317565917969, 'timing/data': 0.018743056803941727, 'timing/model': 2.6661006659269333, 'learning_rate/qwen_vl_interface': 9.954523765544507e-06, 'learning_rate/action_model': 9.954523765544508e-05, 'epoch': 0.65}) +[2026-06-19 17:17:42,538][starVLA.training.train_starvla][INFO] - Step 317, Loss: {'train/loss': 0.3773520290851593, 'train/grad_norm_pre_clip': 9.231888771057129, 'timing/data': 0.018451247364282608, 'timing/model': 2.664065431803465, 'learning_rate/qwen_vl_interface': 9.954102395585814e-06, 'learning_rate/action_model': 9.954102395585813e-05, 'epoch': 0.65}) +[2026-06-19 17:17:53,038][starVLA.training.train_starvla][INFO] - Step 318, Loss: {'train/loss': 0.32099753618240356, 'train/grad_norm_pre_clip': 8.281829833984375, 'timing/data': 0.018692396581172943, 'timing/model': 2.6664033755660057, 'learning_rate/qwen_vl_interface': 9.953679091948313e-06, 'learning_rate/action_model': 9.953679091948314e-05, 'epoch': 0.65}) +[2026-06-19 17:18:03,540][starVLA.training.train_starvla][INFO] - Step 319, Loss: {'train/loss': 0.1463533341884613, 'train/grad_norm_pre_clip': 3.5631301403045654, 'timing/data': 0.018352871760725975, 'timing/model': 2.6632392406463623, 'learning_rate/qwen_vl_interface': 9.953253854806013e-06, 'learning_rate/action_model': 9.953253854806012e-05, 'epoch': 0.66}) +[2026-06-19 17:18:14,036][starVLA.training.train_starvla][INFO] - Step 320, Loss: {'train/loss': 0.2495434582233429, 'train/grad_norm_pre_clip': 10.22552490234375, 'timing/data': 0.025937847793102264, 'timing/model': 2.6560838874429464, 'learning_rate/qwen_vl_interface': 9.95282668433371e-06, 'learning_rate/action_model': 9.952826684333709e-05, 'epoch': 0.66}) +[2026-06-19 17:18:24,774][starVLA.training.train_starvla][INFO] - Step 321, Loss: {'train/loss': 0.3306224048137665, 'train/grad_norm_pre_clip': 6.736478805541992, 'timing/data': 0.018655065447092056, 'timing/model': 2.8987502176314592, 'learning_rate/qwen_vl_interface': 9.952397580706998e-06, 'learning_rate/action_model': 9.952397580706998e-05, 'epoch': 0.66}) +[2026-06-19 17:18:35,235][starVLA.training.train_starvla][INFO] - Step 322, Loss: {'train/loss': 0.37876400351524353, 'train/grad_norm_pre_clip': 16.689002990722656, 'timing/data': 0.018637925386428833, 'timing/model': 2.6620442494750023, 'learning_rate/qwen_vl_interface': 9.951966544102267e-06, 'learning_rate/action_model': 9.951966544102266e-05, 'epoch': 0.66}) +[2026-06-19 17:18:45,937][starVLA.training.train_starvla][INFO] - Step 323, Loss: {'train/loss': 0.33358234167099, 'train/grad_norm_pre_clip': 12.22751235961914, 'timing/data': 0.018817495554685593, 'timing/model': 2.6950925271958113, 'learning_rate/qwen_vl_interface': 9.951533574696696e-06, 'learning_rate/action_model': 9.951533574696697e-05, 'epoch': 0.66}) +[2026-06-19 17:18:56,609][starVLA.training.train_starvla][INFO] - Step 324, Loss: {'train/loss': 0.18748906254768372, 'train/grad_norm_pre_clip': 5.258973598480225, 'timing/data': 0.018725886940956116, 'timing/model': 2.7540414836257696, 'learning_rate/qwen_vl_interface': 9.951098672668266e-06, 'learning_rate/action_model': 9.951098672668265e-05, 'epoch': 0.67}) +[2026-06-19 17:19:07,309][starVLA.training.train_starvla][INFO] - Step 325, Loss: {'train/loss': 0.18955250084400177, 'train/grad_norm_pre_clip': 4.25917387008667, 'timing/data': 0.018054502084851265, 'timing/model': 2.7593132350593805, 'learning_rate/qwen_vl_interface': 9.95066183819575e-06, 'learning_rate/action_model': 9.950661838195749e-05, 'epoch': 0.67}) +[2026-06-19 17:19:18,018][starVLA.training.train_starvla][INFO] - Step 326, Loss: {'train/loss': 0.39115408062934875, 'train/grad_norm_pre_clip': 20.667129516601562, 'timing/data': 0.01911221258342266, 'timing/model': 2.7457534447312355, 'learning_rate/qwen_vl_interface': 9.95022307145871e-06, 'learning_rate/action_model': 9.950223071458708e-05, 'epoch': 0.67}) +[2026-06-19 17:19:28,721][starVLA.training.train_starvla][INFO] - Step 327, Loss: {'train/loss': 0.2597735822200775, 'train/grad_norm_pre_clip': 9.74138355255127, 'timing/data': 0.018628304824233055, 'timing/model': 2.7607773281633854, 'learning_rate/qwen_vl_interface': 9.949782372637507e-06, 'learning_rate/action_model': 9.949782372637506e-05, 'epoch': 0.67}) +[2026-06-19 17:19:39,502][starVLA.training.train_starvla][INFO] - Step 328, Loss: {'train/loss': 0.3582678437232971, 'train/grad_norm_pre_clip': 19.721288681030273, 'timing/data': 0.023975899443030357, 'timing/model': 2.7414315585047007, 'learning_rate/qwen_vl_interface': 9.949339741913297e-06, 'learning_rate/action_model': 9.949339741913297e-05, 'epoch': 0.67}) +[2026-06-19 17:19:50,436][starVLA.training.train_starvla][INFO] - Step 329, Loss: {'train/loss': 0.16680891811847687, 'train/grad_norm_pre_clip': 5.738879203796387, 'timing/data': 0.019042501226067543, 'timing/model': 2.7126374151557684, 'learning_rate/qwen_vl_interface': 9.948895179468032e-06, 'learning_rate/action_model': 9.948895179468031e-05, 'epoch': 0.68}) +[2026-06-19 17:20:01,138][starVLA.training.train_starvla][INFO] - Step 330, Loss: {'train/loss': 0.3644218146800995, 'train/grad_norm_pre_clip': 8.98378849029541, 'timing/data': 0.022390855476260185, 'timing/model': 2.6607485320419073, 'learning_rate/qwen_vl_interface': 9.948448685484448e-06, 'learning_rate/action_model': 9.948448685484447e-05, 'epoch': 0.68}) +[2026-06-19 17:20:11,637][starVLA.training.train_starvla][INFO] - Step 331, Loss: {'train/loss': 0.41601744294166565, 'train/grad_norm_pre_clip': 10.143158912658691, 'timing/data': 0.018489869311451912, 'timing/model': 2.6647342406213284, 'learning_rate/qwen_vl_interface': 9.948000260146087e-06, 'learning_rate/action_model': 9.948000260146087e-05, 'epoch': 0.68}) +[2026-06-19 17:20:22,138][starVLA.training.train_starvla][INFO] - Step 332, Loss: {'train/loss': 0.16971178352832794, 'train/grad_norm_pre_clip': 6.318087577819824, 'timing/data': 0.01827343925833702, 'timing/model': 2.6666046120226383, 'learning_rate/qwen_vl_interface': 9.947549903637277e-06, 'learning_rate/action_model': 9.947549903637278e-05, 'epoch': 0.68}) +[2026-06-19 17:20:32,632][starVLA.training.train_starvla][INFO] - Step 333, Loss: {'train/loss': 0.16223907470703125, 'train/grad_norm_pre_clip': 5.760888576507568, 'timing/data': 0.018888581544160843, 'timing/model': 2.6595423873513937, 'learning_rate/qwen_vl_interface': 9.947097616143145e-06, 'learning_rate/action_model': 9.947097616143144e-05, 'epoch': 0.68}) +[2026-06-19 17:20:43,139][starVLA.training.train_starvla][INFO] - Step 334, Loss: {'train/loss': 0.38493674993515015, 'train/grad_norm_pre_clip': 11.706459999084473, 'timing/data': 0.018561137840151787, 'timing/model': 2.6679613273590803, 'learning_rate/qwen_vl_interface': 9.946643397849607e-06, 'learning_rate/action_model': 9.946643397849607e-05, 'epoch': 0.69}) +[2026-06-19 17:20:53,634][starVLA.training.train_starvla][INFO] - Step 335, Loss: {'train/loss': 0.32167527079582214, 'train/grad_norm_pre_clip': 11.260202407836914, 'timing/data': 0.018531136214733124, 'timing/model': 2.6614310909062624, 'learning_rate/qwen_vl_interface': 9.946187248943375e-06, 'learning_rate/action_model': 9.946187248943374e-05, 'epoch': 0.69}) +[2026-06-19 17:21:04,219][starVLA.training.train_starvla][INFO] - Step 336, Loss: {'train/loss': 0.22115814685821533, 'train/grad_norm_pre_clip': 5.355254650115967, 'timing/data': 0.03006283938884735, 'timing/model': 2.7315399646759033, 'learning_rate/qwen_vl_interface': 9.945729169611957e-06, 'learning_rate/action_model': 9.945729169611958e-05, 'epoch': 0.69}) +[2026-06-19 17:21:14,736][starVLA.training.train_starvla][INFO] - Step 337, Loss: {'train/loss': 0.21177564561367035, 'train/grad_norm_pre_clip': 4.113237380981445, 'timing/data': 0.02062741294503212, 'timing/model': 2.6581140719354153, 'learning_rate/qwen_vl_interface': 9.94526916004365e-06, 'learning_rate/action_model': 9.94526916004365e-05, 'epoch': 0.69}) +[2026-06-19 17:21:25,236][starVLA.training.train_starvla][INFO] - Step 338, Loss: {'train/loss': 0.16648174822330475, 'train/grad_norm_pre_clip': 5.3068060874938965, 'timing/data': 0.01850680075585842, 'timing/model': 2.6659586783498526, 'learning_rate/qwen_vl_interface': 9.944807220427549e-06, 'learning_rate/action_model': 9.944807220427548e-05, 'epoch': 0.69}) +[2026-06-19 17:21:35,919][starVLA.training.train_starvla][INFO] - Step 339, Loss: {'train/loss': 0.22329244017601013, 'train/grad_norm_pre_clip': 7.986429214477539, 'timing/data': 0.022378506138920784, 'timing/model': 2.7429263219237328, 'learning_rate/qwen_vl_interface': 9.944343350953534e-06, 'learning_rate/action_model': 9.944343350953533e-05, 'epoch': 0.7}) +[2026-06-19 17:21:46,541][starVLA.training.train_starvla][INFO] - Step 340, Loss: {'train/loss': 0.44771164655685425, 'train/grad_norm_pre_clip': 15.263097763061523, 'timing/data': 0.01955850049853325, 'timing/model': 2.6651749312877655, 'learning_rate/qwen_vl_interface': 9.94387755181229e-06, 'learning_rate/action_model': 9.94387755181229e-05, 'epoch': 0.7}) +[2026-06-19 17:21:57,236][starVLA.training.train_starvla][INFO] - Step 341, Loss: {'train/loss': 0.2129071205854416, 'train/grad_norm_pre_clip': 6.619846343994141, 'timing/data': 0.018450750038027763, 'timing/model': 2.6663605216890574, 'learning_rate/qwen_vl_interface': 9.943409823195289e-06, 'learning_rate/action_model': 9.943409823195288e-05, 'epoch': 0.7}) +[2026-06-19 17:22:07,933][starVLA.training.train_starvla][INFO] - Step 342, Loss: {'train/loss': 0.3315819501876831, 'train/grad_norm_pre_clip': 8.117989540100098, 'timing/data': 0.021795041859149933, 'timing/model': 2.7062215264886618, 'learning_rate/qwen_vl_interface': 9.942940165294794e-06, 'learning_rate/action_model': 9.942940165294793e-05, 'epoch': 0.7}) +[2026-06-19 17:22:18,440][starVLA.training.train_starvla][INFO] - Step 343, Loss: {'train/loss': 0.08056126534938812, 'train/grad_norm_pre_clip': 2.8139827251434326, 'timing/data': 0.01930360309779644, 'timing/model': 2.664995277300477, 'learning_rate/qwen_vl_interface': 9.942468578303865e-06, 'learning_rate/action_model': 9.942468578303865e-05, 'epoch': 0.7}) +[2026-06-19 17:22:28,965][starVLA.training.train_starvla][INFO] - Step 344, Loss: {'train/loss': 0.21024706959724426, 'train/grad_norm_pre_clip': 8.031124114990234, 'timing/data': 0.019207382574677467, 'timing/model': 2.692873803898692, 'learning_rate/qwen_vl_interface': 9.941995062416354e-06, 'learning_rate/action_model': 9.941995062416354e-05, 'epoch': 0.71}) +[2026-06-19 17:22:39,537][starVLA.training.train_starvla][INFO] - Step 345, Loss: {'train/loss': 0.22110417485237122, 'train/grad_norm_pre_clip': 8.915519714355469, 'timing/data': 0.02011181227862835, 'timing/model': 2.6736235711723566, 'learning_rate/qwen_vl_interface': 9.941519617826906e-06, 'learning_rate/action_model': 9.941519617826906e-05, 'epoch': 0.71}) +[2026-06-19 17:22:50,036][starVLA.training.train_starvla][INFO] - Step 346, Loss: {'train/loss': 0.1865888088941574, 'train/grad_norm_pre_clip': 10.046953201293945, 'timing/data': 0.0131077840924263, 'timing/model': 2.673929139971733, 'learning_rate/qwen_vl_interface': 9.941042244730954e-06, 'learning_rate/action_model': 9.941042244730955e-05, 'epoch': 0.71}) +[2026-06-19 17:23:00,537][starVLA.training.train_starvla][INFO] - Step 347, Loss: {'train/loss': 0.260331392288208, 'train/grad_norm_pre_clip': 15.987944602966309, 'timing/data': 0.013071365654468536, 'timing/model': 2.666130594909191, 'learning_rate/qwen_vl_interface': 9.940562943324736e-06, 'learning_rate/action_model': 9.940562943324735e-05, 'epoch': 0.71}) +[2026-06-19 17:23:11,308][starVLA.training.train_starvla][INFO] - Step 348, Loss: {'train/loss': 0.5343088507652283, 'train/grad_norm_pre_clip': 23.176494598388672, 'timing/data': 0.01329452358186245, 'timing/model': 2.9413338992744684, 'learning_rate/qwen_vl_interface': 9.940081713805266e-06, 'learning_rate/action_model': 9.940081713805266e-05, 'epoch': 0.71}) +[2026-06-19 17:23:21,738][starVLA.training.train_starvla][INFO] - Step 349, Loss: {'train/loss': 0.21123884618282318, 'train/grad_norm_pre_clip': 12.07716178894043, 'timing/data': 0.012771647423505783, 'timing/model': 2.674422414973378, 'learning_rate/qwen_vl_interface': 9.939598556370366e-06, 'learning_rate/action_model': 9.939598556370365e-05, 'epoch': 0.72}) +[2026-06-19 17:23:32,320][starVLA.training.train_starvla][INFO] - Step 350, Loss: {'train/loss': 0.18658296763896942, 'train/grad_norm_pre_clip': 7.931232929229736, 'timing/data': 0.015393851324915886, 'timing/model': 2.7472420409321785, 'learning_rate/qwen_vl_interface': 9.93911347121864e-06, 'learning_rate/action_model': 9.939113471218639e-05, 'epoch': 0.72}) +[2026-06-19 17:23:42,939][starVLA.training.train_starvla][INFO] - Step 351, Loss: {'train/loss': 0.4604804217815399, 'train/grad_norm_pre_clip': 22.491777420043945, 'timing/data': 0.013751668855547905, 'timing/model': 2.6640292555093765, 'learning_rate/qwen_vl_interface': 9.938626458549488e-06, 'learning_rate/action_model': 9.938626458549488e-05, 'epoch': 0.72}) +[2026-06-19 17:23:53,435][starVLA.training.train_starvla][INFO] - Step 352, Loss: {'train/loss': 0.41651207208633423, 'train/grad_norm_pre_clip': 16.454694747924805, 'timing/data': 0.012377625331282616, 'timing/model': 2.673125995323062, 'learning_rate/qwen_vl_interface': 9.938137518563106e-06, 'learning_rate/action_model': 9.938137518563106e-05, 'epoch': 0.72}) +[2026-06-19 17:24:03,936][starVLA.training.train_starvla][INFO] - Step 353, Loss: {'train/loss': 0.13792327046394348, 'train/grad_norm_pre_clip': 6.461341857910156, 'timing/data': 0.013119736686348915, 'timing/model': 2.666557192802429, 'learning_rate/qwen_vl_interface': 9.937646651460475e-06, 'learning_rate/action_model': 9.937646651460474e-05, 'epoch': 0.73}) +[2026-06-19 17:24:14,441][starVLA.training.train_starvla][INFO] - Step 354, Loss: {'train/loss': 0.2127649337053299, 'train/grad_norm_pre_clip': 5.054922103881836, 'timing/data': 0.01293489895761013, 'timing/model': 2.675265856087208, 'learning_rate/qwen_vl_interface': 9.937153857443375e-06, 'learning_rate/action_model': 9.937153857443374e-05, 'epoch': 0.73}) +[2026-06-19 17:24:24,937][starVLA.training.train_starvla][INFO] - Step 355, Loss: {'train/loss': 0.35519546270370483, 'train/grad_norm_pre_clip': 24.73820686340332, 'timing/data': 0.012773068621754646, 'timing/model': 2.669891996309161, 'learning_rate/qwen_vl_interface': 9.936659136714372e-06, 'learning_rate/action_model': 9.936659136714371e-05, 'epoch': 0.73}) +[2026-06-19 17:24:35,538][starVLA.training.train_starvla][INFO] - Step 356, Loss: {'train/loss': 0.3427318334579468, 'train/grad_norm_pre_clip': 11.765227317810059, 'timing/data': 0.013090498745441437, 'timing/model': 2.628370588645339, 'learning_rate/qwen_vl_interface': 9.936162489476827e-06, 'learning_rate/action_model': 9.936162489476828e-05, 'epoch': 0.73}) +[2026-06-19 17:24:46,073][starVLA.training.train_starvla][INFO] - Step 357, Loss: {'train/loss': 0.10452527552843094, 'train/grad_norm_pre_clip': 3.8160908222198486, 'timing/data': 0.01261891983449459, 'timing/model': 2.704850049689412, 'learning_rate/qwen_vl_interface': 9.935663915934899e-06, 'learning_rate/action_model': 9.935663915934898e-05, 'epoch': 0.73}) +[2026-06-19 17:24:56,539][starVLA.training.train_starvla][INFO] - Step 358, Loss: {'train/loss': 0.21668970584869385, 'train/grad_norm_pre_clip': 9.986796379089355, 'timing/data': 0.013012640178203583, 'timing/model': 2.6724574118852615, 'learning_rate/qwen_vl_interface': 9.935163416293525e-06, 'learning_rate/action_model': 9.935163416293525e-05, 'epoch': 0.74}) +[2026-06-19 17:25:07,116][starVLA.training.train_starvla][INFO] - Step 359, Loss: {'train/loss': 0.29975488781929016, 'train/grad_norm_pre_clip': 7.159776210784912, 'timing/data': 0.012578433379530907, 'timing/model': 2.716273946687579, 'learning_rate/qwen_vl_interface': 9.934660990758445e-06, 'learning_rate/action_model': 9.934660990758445e-05, 'epoch': 0.74}) +[2026-06-19 17:25:18,018][starVLA.training.train_starvla][INFO] - Step 360, Loss: {'train/loss': 0.08879700303077698, 'train/grad_norm_pre_clip': 9.283373832702637, 'timing/data': 0.015520857647061348, 'timing/model': 2.7597369085997343, 'learning_rate/qwen_vl_interface': 9.934156639536186e-06, 'learning_rate/action_model': 9.934156639536187e-05, 'epoch': 0.74}) +[2026-06-19 17:25:28,817][starVLA.training.train_starvla][INFO] - Step 361, Loss: {'train/loss': 0.2464449405670166, 'train/grad_norm_pre_clip': 8.74374771118164, 'timing/data': 0.012576233595609665, 'timing/model': 2.7756368182599545, 'learning_rate/qwen_vl_interface': 9.93365036283407e-06, 'learning_rate/action_model': 9.93365036283407e-05, 'epoch': 0.74}) +[2026-06-19 17:25:39,409][starVLA.training.train_starvla][INFO] - Step 362, Loss: {'train/loss': 0.393926203250885, 'train/grad_norm_pre_clip': 13.672608375549316, 'timing/data': 0.013100847601890564, 'timing/model': 2.6966144274920225, 'learning_rate/qwen_vl_interface': 9.933142160860208e-06, 'learning_rate/action_model': 9.933142160860208e-05, 'epoch': 0.74}) +[2026-06-19 17:25:50,019][starVLA.training.train_starvla][INFO] - Step 363, Loss: {'train/loss': 0.4881184697151184, 'train/grad_norm_pre_clip': 12.106826782226562, 'timing/data': 0.01273009181022644, 'timing/model': 2.7699767015874386, 'learning_rate/qwen_vl_interface': 9.932632033823499e-06, 'learning_rate/action_model': 9.932632033823499e-05, 'epoch': 0.75}) +[2026-06-19 17:26:00,612][starVLA.training.train_starvla][INFO] - Step 364, Loss: {'train/loss': 0.4967104494571686, 'train/grad_norm_pre_clip': 11.334589004516602, 'timing/data': 0.012926461175084114, 'timing/model': 2.742352604866028, 'learning_rate/qwen_vl_interface': 9.93211998193364e-06, 'learning_rate/action_model': 9.932119981933639e-05, 'epoch': 0.75}) +[2026-06-19 17:26:11,203][starVLA.training.train_starvla][INFO] - Step 365, Loss: {'train/loss': 0.517767071723938, 'train/grad_norm_pre_clip': 13.201346397399902, 'timing/data': 0.012600811198353767, 'timing/model': 2.727300440892577, 'learning_rate/qwen_vl_interface': 9.931606005401117e-06, 'learning_rate/action_model': 9.931606005401116e-05, 'epoch': 0.75}) +[2026-06-19 17:26:21,838][starVLA.training.train_starvla][INFO] - Step 366, Loss: {'train/loss': 0.3683288097381592, 'train/grad_norm_pre_clip': 8.30107307434082, 'timing/data': 0.012874562293291092, 'timing/model': 2.6821697540581226, 'learning_rate/qwen_vl_interface': 9.931090104437202e-06, 'learning_rate/action_model': 9.931090104437202e-05, 'epoch': 0.75}) +[2026-06-19 17:26:32,336][starVLA.training.train_starvla][INFO] - Step 367, Loss: {'train/loss': 0.20526017248630524, 'train/grad_norm_pre_clip': 4.7205586433410645, 'timing/data': 0.0130278579890728, 'timing/model': 2.6707402765750885, 'learning_rate/qwen_vl_interface': 9.930572279253967e-06, 'learning_rate/action_model': 9.930572279253967e-05, 'epoch': 0.75}) +[2026-06-19 17:26:43,032][starVLA.training.train_starvla][INFO] - Step 368, Loss: {'train/loss': 0.11697161942720413, 'train/grad_norm_pre_clip': 3.6596202850341797, 'timing/data': 0.013148920610547066, 'timing/model': 2.665426168590784, 'learning_rate/qwen_vl_interface': 9.930052530064269e-06, 'learning_rate/action_model': 9.93005253006427e-05, 'epoch': 0.76}) +[2026-06-19 17:26:53,683][starVLA.training.train_starvla][INFO] - Step 369, Loss: {'train/loss': 0.5545100569725037, 'train/grad_norm_pre_clip': 14.029191970825195, 'timing/data': 0.012750821188092232, 'timing/model': 2.6688686292618513, 'learning_rate/qwen_vl_interface': 9.929530857081757e-06, 'learning_rate/action_model': 9.929530857081757e-05, 'epoch': 0.76}) +[2026-06-19 17:27:04,237][starVLA.training.train_starvla][INFO] - Step 370, Loss: {'train/loss': 0.7857765555381775, 'train/grad_norm_pre_clip': 15.88582706451416, 'timing/data': 0.012843849137425423, 'timing/model': 2.669313207268715, 'learning_rate/qwen_vl_interface': 9.92900726052087e-06, 'learning_rate/action_model': 9.929007260520869e-05, 'epoch': 0.76}) +[2026-06-19 17:27:14,795][starVLA.training.train_starvla][INFO] - Step 371, Loss: {'train/loss': 0.5775384306907654, 'train/grad_norm_pre_clip': 16.294757843017578, 'timing/data': 0.013067960739135742, 'timing/model': 2.727777497842908, 'learning_rate/qwen_vl_interface': 9.92848174059684e-06, 'learning_rate/action_model': 9.92848174059684e-05, 'epoch': 0.76}) +[2026-06-19 17:27:25,335][starVLA.training.train_starvla][INFO] - Step 372, Loss: {'train/loss': 0.28281477093696594, 'train/grad_norm_pre_clip': 4.1154351234436035, 'timing/data': 0.01865389570593834, 'timing/model': 2.6897223256528378, 'learning_rate/qwen_vl_interface': 9.92795429752569e-06, 'learning_rate/action_model': 9.927954297525689e-05, 'epoch': 0.76}) +[2026-06-19 17:27:35,837][starVLA.training.train_starvla][INFO] - Step 373, Loss: {'train/loss': 0.5605051517486572, 'train/grad_norm_pre_clip': 11.991711616516113, 'timing/data': 0.012878190726041794, 'timing/model': 2.6686496548354626, 'learning_rate/qwen_vl_interface': 9.927424931524228e-06, 'learning_rate/action_model': 9.927424931524228e-05, 'epoch': 0.77}) +[2026-06-19 17:27:46,416][starVLA.training.train_starvla][INFO] - Step 374, Loss: {'train/loss': 0.19736534357070923, 'train/grad_norm_pre_clip': 17.764863967895508, 'timing/data': 0.013088177889585495, 'timing/model': 2.7459849137812853, 'learning_rate/qwen_vl_interface': 9.926893642810058e-06, 'learning_rate/action_model': 9.926893642810059e-05, 'epoch': 0.77}) +[2026-06-19 17:27:57,126][starVLA.training.train_starvla][INFO] - Step 375, Loss: {'train/loss': 0.513561487197876, 'train/grad_norm_pre_clip': 4.966762542724609, 'timing/data': 0.01586572639644146, 'timing/model': 2.7669575307518244, 'learning_rate/qwen_vl_interface': 9.926360431601575e-06, 'learning_rate/action_model': 9.926360431601574e-05, 'epoch': 0.77}) +[2026-06-19 17:28:07,915][starVLA.training.train_starvla][INFO] - Step 376, Loss: {'train/loss': 0.22580429911613464, 'train/grad_norm_pre_clip': 6.501821041107178, 'timing/data': 0.013417895883321762, 'timing/model': 2.7682867515832186, 'learning_rate/qwen_vl_interface': 9.92582529811796e-06, 'learning_rate/action_model': 9.925825298117959e-05, 'epoch': 0.77}) +[2026-06-19 17:28:18,638][starVLA.training.train_starvla][INFO] - Step 377, Loss: {'train/loss': 0.3047262728214264, 'train/grad_norm_pre_clip': 8.329646110534668, 'timing/data': 0.012707684189081192, 'timing/model': 2.698184607550502, 'learning_rate/qwen_vl_interface': 9.925288242579184e-06, 'learning_rate/action_model': 9.925288242579184e-05, 'epoch': 0.77}) +[2026-06-19 17:28:29,237][starVLA.training.train_starvla][INFO] - Step 378, Loss: {'train/loss': 0.3488764464855194, 'train/grad_norm_pre_clip': 8.831257820129395, 'timing/data': 0.01615268364548683, 'timing/model': 2.6834990456700325, 'learning_rate/qwen_vl_interface': 9.924749265206015e-06, 'learning_rate/action_model': 9.924749265206015e-05, 'epoch': 0.78}) +[2026-06-19 17:28:39,935][starVLA.training.train_starvla][INFO] - Step 379, Loss: {'train/loss': 0.2682000994682312, 'train/grad_norm_pre_clip': 5.645657062530518, 'timing/data': 0.01300082914531231, 'timing/model': 2.667026847600937, 'learning_rate/qwen_vl_interface': 9.924208366220003e-06, 'learning_rate/action_model': 9.924208366220002e-05, 'epoch': 0.78}) +[2026-06-19 17:28:50,437][starVLA.training.train_starvla][INFO] - Step 380, Loss: {'train/loss': 0.2534142732620239, 'train/grad_norm_pre_clip': 3.390883684158325, 'timing/data': 0.015018075704574585, 'timing/model': 2.6721181459724903, 'learning_rate/qwen_vl_interface': 9.923665545843492e-06, 'learning_rate/action_model': 9.923665545843491e-05, 'epoch': 0.78}) +[2026-06-19 17:29:01,112][starVLA.training.train_starvla][INFO] - Step 381, Loss: {'train/loss': 0.2161930352449417, 'train/grad_norm_pre_clip': 3.28253436088562, 'timing/data': 0.012650564312934875, 'timing/model': 2.7746136356145144, 'learning_rate/qwen_vl_interface': 9.923120804299615e-06, 'learning_rate/action_model': 9.923120804299613e-05, 'epoch': 0.78}) +[2026-06-19 17:29:11,738][starVLA.training.train_starvla][INFO] - Step 382, Loss: {'train/loss': 0.1723778247833252, 'train/grad_norm_pre_clip': 3.519373655319214, 'timing/data': 0.012718664482235909, 'timing/model': 2.6981114987283945, 'learning_rate/qwen_vl_interface': 9.922574141812295e-06, 'learning_rate/action_model': 9.922574141812294e-05, 'epoch': 0.78}) +[2026-06-19 17:29:22,376][starVLA.training.train_starvla][INFO] - Step 383, Loss: {'train/loss': 0.1468333899974823, 'train/grad_norm_pre_clip': 3.758831739425659, 'timing/data': 0.012595225125551224, 'timing/model': 2.7445021718740463, 'learning_rate/qwen_vl_interface': 9.922025558606242e-06, 'learning_rate/action_model': 9.922025558606242e-05, 'epoch': 0.79}) +[2026-06-19 17:29:32,960][starVLA.training.train_starvla][INFO] - Step 384, Loss: {'train/loss': 0.21101073920726776, 'train/grad_norm_pre_clip': 8.625394821166992, 'timing/data': 0.012846063822507858, 'timing/model': 2.6907127741724253, 'learning_rate/qwen_vl_interface': 9.92147505490696e-06, 'learning_rate/action_model': 9.92147505490696e-05, 'epoch': 0.79}) +[2026-06-19 17:29:43,475][starVLA.training.train_starvla][INFO] - Step 385, Loss: {'train/loss': 0.4489753246307373, 'train/grad_norm_pre_clip': 10.969736099243164, 'timing/data': 0.012777542695403099, 'timing/model': 2.692952912300825, 'learning_rate/qwen_vl_interface': 9.920922630940741e-06, 'learning_rate/action_model': 9.920922630940742e-05, 'epoch': 0.79}) +[2026-06-19 17:29:54,017][starVLA.training.train_starvla][INFO] - Step 386, Loss: {'train/loss': 0.26142778992652893, 'train/grad_norm_pre_clip': 10.814041137695312, 'timing/data': 0.01311630941927433, 'timing/model': 2.722005743533373, 'learning_rate/qwen_vl_interface': 9.920368286934664e-06, 'learning_rate/action_model': 9.920368286934663e-05, 'epoch': 0.79}) +[2026-06-19 17:30:05,022][starVLA.training.train_starvla][INFO] - Step 387, Loss: {'train/loss': 0.25385797023773193, 'train/grad_norm_pre_clip': 15.836365699768066, 'timing/data': 0.012431109324097633, 'timing/model': 2.779267417266965, 'learning_rate/qwen_vl_interface': 9.9198120231166e-06, 'learning_rate/action_model': 9.9198120231166e-05, 'epoch': 0.79}) +[2026-06-19 17:30:15,706][starVLA.training.train_starvla][INFO] - Step 388, Loss: {'train/loss': 0.41389670968055725, 'train/grad_norm_pre_clip': 15.084014892578125, 'timing/data': 0.013084780424833298, 'timing/model': 2.7552270721644163, 'learning_rate/qwen_vl_interface': 9.919253839715208e-06, 'learning_rate/action_model': 9.919253839715208e-05, 'epoch': 0.8}) +[2026-06-19 17:30:26,314][starVLA.training.train_starvla][INFO] - Step 389, Loss: {'train/loss': 0.3916192948818207, 'train/grad_norm_pre_clip': 11.519621849060059, 'timing/data': 0.013171140104532242, 'timing/model': 2.77497305162251, 'learning_rate/qwen_vl_interface': 9.918693736959935e-06, 'learning_rate/action_model': 9.918693736959934e-05, 'epoch': 0.8}) +[2026-06-19 17:30:37,084][starVLA.training.train_starvla][INFO] - Step 390, Loss: {'train/loss': 0.5421770811080933, 'train/grad_norm_pre_clip': 12.978519439697266, 'timing/data': 0.012946143746376038, 'timing/model': 2.817740695551038, 'learning_rate/qwen_vl_interface': 9.918131715081019e-06, 'learning_rate/action_model': 9.918131715081018e-05, 'epoch': 0.8}) +[2026-06-19 17:30:47,891][starVLA.training.train_starvla][INFO] - Step 391, Loss: {'train/loss': 0.5658846497535706, 'train/grad_norm_pre_clip': 18.978496551513672, 'timing/data': 0.013650564476847649, 'timing/model': 2.7545709498226643, 'learning_rate/qwen_vl_interface': 9.917567774309486e-06, 'learning_rate/action_model': 9.917567774309485e-05, 'epoch': 0.8}) +[2026-06-19 17:30:58,712][starVLA.training.train_starvla][INFO] - Step 392, Loss: {'train/loss': 0.2990366220474243, 'train/grad_norm_pre_clip': 9.129392623901367, 'timing/data': 0.012634115293622017, 'timing/model': 2.7653480917215347, 'learning_rate/qwen_vl_interface': 9.91700191487715e-06, 'learning_rate/action_model': 9.91700191487715e-05, 'epoch': 0.81}) +[2026-06-19 17:31:09,324][starVLA.training.train_starvla][INFO] - Step 393, Loss: {'train/loss': 0.14067164063453674, 'train/grad_norm_pre_clip': 14.325122833251953, 'timing/data': 0.013183960691094398, 'timing/model': 2.7024447601288557, 'learning_rate/qwen_vl_interface': 9.916434137016614e-06, 'learning_rate/action_model': 9.916434137016613e-05, 'epoch': 0.81}) +[2026-06-19 17:31:20,089][starVLA.training.train_starvla][INFO] - Step 394, Loss: {'train/loss': 0.1601932793855667, 'train/grad_norm_pre_clip': 3.4014954566955566, 'timing/data': 0.013415444642305374, 'timing/model': 2.741217838600278, 'learning_rate/qwen_vl_interface': 9.915864440961272e-06, 'learning_rate/action_model': 9.915864440961272e-05, 'epoch': 0.81}) +[2026-06-19 17:31:30,740][starVLA.training.train_starvla][INFO] - Step 395, Loss: {'train/loss': 0.331128865480423, 'train/grad_norm_pre_clip': 10.048826217651367, 'timing/data': 0.01263696700334549, 'timing/model': 2.6973031759262085, 'learning_rate/qwen_vl_interface': 9.915292826945303e-06, 'learning_rate/action_model': 9.915292826945303e-05, 'epoch': 0.81}) +[2026-06-19 17:31:41,237][starVLA.training.train_starvla][INFO] - Step 396, Loss: {'train/loss': 0.4833582937717438, 'train/grad_norm_pre_clip': 13.2106294631958, 'timing/data': 0.013472987338900566, 'timing/model': 2.667365115135908, 'learning_rate/qwen_vl_interface': 9.914719295203676e-06, 'learning_rate/action_model': 9.914719295203676e-05, 'epoch': 0.81}) +[2026-06-19 17:31:51,735][starVLA.training.train_starvla][INFO] - Step 397, Loss: {'train/loss': 0.4543439447879791, 'train/grad_norm_pre_clip': 8.434627532958984, 'timing/data': 0.013651296496391296, 'timing/model': 2.6681811679154634, 'learning_rate/qwen_vl_interface': 9.91414384597215e-06, 'learning_rate/action_model': 9.91414384597215e-05, 'epoch': 0.82}) +[2026-06-19 17:32:02,435][starVLA.training.train_starvla][INFO] - Step 398, Loss: {'train/loss': 0.25856855511665344, 'train/grad_norm_pre_clip': 4.089259147644043, 'timing/data': 0.013555856421589851, 'timing/model': 2.65583873167634, 'learning_rate/qwen_vl_interface': 9.913566479487267e-06, 'learning_rate/action_model': 9.913566479487267e-05, 'epoch': 0.82}) +[2026-06-19 17:32:12,934][starVLA.training.train_starvla][INFO] - Step 399, Loss: {'train/loss': 0.21189412474632263, 'train/grad_norm_pre_clip': 10.264260292053223, 'timing/data': 0.012872850522398949, 'timing/model': 2.6717280577868223, 'learning_rate/qwen_vl_interface': 9.912987195986366e-06, 'learning_rate/action_model': 9.912987195986365e-05, 'epoch': 0.82}) +[2026-06-19 17:32:23,435][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 400 +[2026-06-19 17:40:07,941][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 400 +[2026-06-19 17:40:07,943][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 400 +[2026-06-19 17:48:33,119][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 400 +[2026-06-19 17:48:33,121][starVLA.training.train_starvla][INFO] - Step 400, Loss: {'train/loss': 0.3283444941043854, 'train/grad_norm_pre_clip': 5.19657039642334, 'eval/loss': 0.3579845130443573, 'eval/latency_0/loss': 0.1523195058107376, 'eval/latency_1/loss': 0.24638350307941437, 'eval/latency_2/loss': 0.4310617744922638, 'eval/latency_3/loss': 0.4108182489871979, 'eval/latency_4/loss': 0.5372603535652161, 'eval/flappy/loss': 0.3579845130443573, 'eval/flappy/latency_0/loss': 0.1523195058107376, 'eval/flappy/latency_1/loss': 0.24638350307941437, 'eval/flappy/latency_2/loss': 0.4310617744922638, 'eval/flappy/latency_3/loss': 0.4108182489871979, 'eval/flappy/latency_4/loss': 0.5372603535652161, 'eval/flap_precision': 0.5878787878787879, 'eval/flap_recall': 0.7054545454545454, 'eval/flap_f1': 0.6413223140495867, 'eval/flap_rate_model': 0.09166666666666666, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.6413223140495867, 'eval/latency_0/flap_precision': 0.5878787878787879, 'eval/latency_0/flap_recall': 0.7054545454545454, 'eval/latency_0/flap_f1': 0.6413223140495867, 'eval/latency_0/flap_rate_model': 0.09166666666666666, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.6413223140495867, 'timing/data': 0.012103823944926262, 'timing/model': 2.6704719364643097, 'learning_rate/qwen_vl_interface': 9.912405995707562e-06, 'learning_rate/action_model': 9.912405995707561e-05, 'epoch': 0.82}) +[2026-06-19 17:48:33,121][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_400_state +[2026-06-19 17:48:33,147][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-19 17:48:39,009][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_400_state/model.safetensors +[2026-06-19 17:48:47,077][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_400_state/optimizer.bin +[2026-06-19 17:48:47,078][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_400_state/sampler.bin +[2026-06-19 17:48:47,078][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_400_state/sampler_1.bin +[2026-06-19 17:48:47,079][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_400_state/random_states_0.pkl +[2026-06-19 17:48:47,227][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 400 +[2026-06-19 17:48:47,227][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-19 17:48:47,241][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-19 17:48:58,006][starVLA.training.train_starvla][INFO] - Step 401, Loss: {'train/loss': 0.1800493746995926, 'train/grad_norm_pre_clip': 7.208378314971924, 'timing/data': 0.011388903483748436, 'timing/model': 2.7619039565324783, 'learning_rate/qwen_vl_interface': 9.911822878889767e-06, 'learning_rate/action_model': 9.911822878889767e-05, 'epoch': 0.82}) +[2026-06-19 17:49:08,701][starVLA.training.train_starvla][INFO] - Step 402, Loss: {'train/loss': 0.31351491808891296, 'train/grad_norm_pre_clip': 4.396163463592529, 'timing/data': 0.01191253587603569, 'timing/model': 2.7700683642178774, 'learning_rate/qwen_vl_interface': 9.911237845772681e-06, 'learning_rate/action_model': 9.911237845772681e-05, 'epoch': 0.83}) +[2026-06-19 17:49:19,727][starVLA.training.train_starvla][INFO] - Step 403, Loss: {'train/loss': 0.15567591786384583, 'train/grad_norm_pre_clip': 4.015517234802246, 'timing/data': 0.013620378449559212, 'timing/model': 2.7069633044302464, 'learning_rate/qwen_vl_interface': 9.910650896596787e-06, 'learning_rate/action_model': 9.910650896596786e-05, 'epoch': 0.83}) +[2026-06-19 17:49:30,511][starVLA.training.train_starvla][INFO] - Step 404, Loss: {'train/loss': 0.21189740300178528, 'train/grad_norm_pre_clip': 5.915843963623047, 'timing/data': 0.012097276747226715, 'timing/model': 2.7625146452337503, 'learning_rate/qwen_vl_interface': 9.910062031603356e-06, 'learning_rate/action_model': 9.910062031603356e-05, 'epoch': 0.83}) +[2026-06-19 17:49:41,315][starVLA.training.train_starvla][INFO] - Step 405, Loss: {'train/loss': 0.2687646448612213, 'train/grad_norm_pre_clip': 7.954106330871582, 'timing/data': 0.012339431792497635, 'timing/model': 2.7675769068300724, 'learning_rate/qwen_vl_interface': 9.909471251034451e-06, 'learning_rate/action_model': 9.90947125103445e-05, 'epoch': 0.83}) +[2026-06-19 17:49:52,121][starVLA.training.train_starvla][INFO] - Step 406, Loss: {'train/loss': 0.4915916323661804, 'train/grad_norm_pre_clip': 15.80085277557373, 'timing/data': 0.01785684935748577, 'timing/model': 2.7588146086782217, 'learning_rate/qwen_vl_interface': 9.908878555132918e-06, 'learning_rate/action_model': 9.908878555132918e-05, 'epoch': 0.83}) +[2026-06-19 17:50:02,916][starVLA.training.train_starvla][INFO] - Step 407, Loss: {'train/loss': 0.295369029045105, 'train/grad_norm_pre_clip': 6.534071922302246, 'timing/data': 0.012441590428352356, 'timing/model': 2.766256244853139, 'learning_rate/qwen_vl_interface': 9.908283944142391e-06, 'learning_rate/action_model': 9.908283944142391e-05, 'epoch': 0.84}) +[2026-06-19 17:50:13,721][starVLA.training.train_starvla][INFO] - Step 408, Loss: {'train/loss': 0.15954118967056274, 'train/grad_norm_pre_clip': 5.245645999908447, 'timing/data': 0.01214379258453846, 'timing/model': 2.773993579670787, 'learning_rate/qwen_vl_interface': 9.907687418307297e-06, 'learning_rate/action_model': 9.907687418307296e-05, 'epoch': 0.84}) +[2026-06-19 17:50:24,524][starVLA.training.train_starvla][INFO] - Step 409, Loss: {'train/loss': 0.2770024538040161, 'train/grad_norm_pre_clip': 9.256956100463867, 'timing/data': 0.016383005306124687, 'timing/model': 2.768140198662877, 'learning_rate/qwen_vl_interface': 9.907088977872837e-06, 'learning_rate/action_model': 9.907088977872837e-05, 'epoch': 0.84}) +[2026-06-19 17:50:35,312][starVLA.training.train_starvla][INFO] - Step 410, Loss: {'train/loss': 0.30986925959587097, 'train/grad_norm_pre_clip': 10.852357864379883, 'timing/data': 0.012285692617297173, 'timing/model': 2.7529112678021193, 'learning_rate/qwen_vl_interface': 9.906488623085015e-06, 'learning_rate/action_model': 9.906488623085015e-05, 'epoch': 0.84}) +[2026-06-19 17:50:46,118][starVLA.training.train_starvla][INFO] - Step 411, Loss: {'train/loss': 0.3267079293727875, 'train/grad_norm_pre_clip': 10.431591033935547, 'timing/data': 0.01194656640291214, 'timing/model': 2.7660030853003263, 'learning_rate/qwen_vl_interface': 9.905886354190611e-06, 'learning_rate/action_model': 9.90588635419061e-05, 'epoch': 0.84}) +[2026-06-19 17:50:56,909][starVLA.training.train_starvla][INFO] - Step 412, Loss: {'train/loss': 0.21388007700443268, 'train/grad_norm_pre_clip': 9.898435592651367, 'timing/data': 0.01207338459789753, 'timing/model': 2.7659902032464743, 'learning_rate/qwen_vl_interface': 9.905282171437192e-06, 'learning_rate/action_model': 9.905282171437192e-05, 'epoch': 0.85}) +[2026-06-19 17:51:07,622][starVLA.training.train_starvla][INFO] - Step 413, Loss: {'train/loss': 0.1865009367465973, 'train/grad_norm_pre_clip': 6.615798473358154, 'timing/data': 0.011407453566789627, 'timing/model': 2.780659969896078, 'learning_rate/qwen_vl_interface': 9.904676075073121e-06, 'learning_rate/action_model': 9.904676075073121e-05, 'epoch': 0.85}) +[2026-06-19 17:51:18,421][starVLA.training.train_starvla][INFO] - Step 414, Loss: {'train/loss': 0.23824404180049896, 'train/grad_norm_pre_clip': 6.7938151359558105, 'timing/data': 0.012361492961645126, 'timing/model': 2.7715347353368998, 'learning_rate/qwen_vl_interface': 9.904068065347536e-06, 'learning_rate/action_model': 9.904068065347536e-05, 'epoch': 0.85}) +[2026-06-19 17:51:29,218][starVLA.training.train_starvla][INFO] - Step 415, Loss: {'train/loss': 0.38374894857406616, 'train/grad_norm_pre_clip': 11.487491607666016, 'timing/data': 0.01204424537718296, 'timing/model': 2.7692481111735106, 'learning_rate/qwen_vl_interface': 9.90345814251037e-06, 'learning_rate/action_model': 9.90345814251037e-05, 'epoch': 0.85}) +[2026-06-19 17:51:39,919][starVLA.training.train_starvla][INFO] - Step 416, Loss: {'train/loss': 0.5134698152542114, 'train/grad_norm_pre_clip': 12.280498504638672, 'timing/data': 0.01239084079861641, 'timing/model': 2.7507786508649588, 'learning_rate/qwen_vl_interface': 9.902846306812339e-06, 'learning_rate/action_model': 9.902846306812338e-05, 'epoch': 0.85}) +[2026-06-19 17:51:50,700][starVLA.training.train_starvla][INFO] - Step 417, Loss: {'train/loss': 0.18683400750160217, 'train/grad_norm_pre_clip': 8.833602905273438, 'timing/data': 0.012030184268951416, 'timing/model': 2.7515852600336075, 'learning_rate/qwen_vl_interface': 9.902232558504942e-06, 'learning_rate/action_model': 9.902232558504941e-05, 'epoch': 0.86}) +[2026-06-19 17:52:01,519][starVLA.training.train_starvla][INFO] - Step 418, Loss: {'train/loss': 0.17825999855995178, 'train/grad_norm_pre_clip': 6.447485446929932, 'timing/data': 0.012925995513796806, 'timing/model': 2.7853137757629156, 'learning_rate/qwen_vl_interface': 9.901616897840471e-06, 'learning_rate/action_model': 9.90161689784047e-05, 'epoch': 0.86}) +[2026-06-19 17:52:12,523][starVLA.training.train_starvla][INFO] - Step 419, Loss: {'train/loss': 0.2234756201505661, 'train/grad_norm_pre_clip': 6.342789649963379, 'timing/data': 0.012832676991820335, 'timing/model': 2.7731641717255116, 'learning_rate/qwen_vl_interface': 9.900999325072e-06, 'learning_rate/action_model': 9.900999325072001e-05, 'epoch': 0.86}) +[2026-06-19 17:52:23,324][starVLA.training.train_starvla][INFO] - Step 420, Loss: {'train/loss': 0.32562190294265747, 'train/grad_norm_pre_clip': 6.9986958503723145, 'timing/data': 0.012439390644431114, 'timing/model': 2.772628638893366, 'learning_rate/qwen_vl_interface': 9.900379840453392e-06, 'learning_rate/action_model': 9.900379840453392e-05, 'epoch': 0.86}) +[2026-06-19 17:52:34,114][starVLA.training.train_starvla][INFO] - Step 421, Loss: {'train/loss': 0.13403168320655823, 'train/grad_norm_pre_clip': 3.364814519882202, 'timing/data': 0.012254022061824799, 'timing/model': 2.7622786443680525, 'learning_rate/qwen_vl_interface': 9.89975844423929e-06, 'learning_rate/action_model': 9.899758444239289e-05, 'epoch': 0.86}) +[2026-06-19 17:52:45,037][starVLA.training.train_starvla][INFO] - Step 422, Loss: {'train/loss': 0.23516146838665009, 'train/grad_norm_pre_clip': 8.400290489196777, 'timing/data': 0.012149123474955559, 'timing/model': 2.8957971706986427, 'learning_rate/qwen_vl_interface': 9.899135136685131e-06, 'learning_rate/action_model': 9.89913513668513e-05, 'epoch': 0.87}) +[2026-06-19 17:52:55,922][starVLA.training.train_starvla][INFO] - Step 423, Loss: {'train/loss': 0.2447776049375534, 'train/grad_norm_pre_clip': 5.545234680175781, 'timing/data': 0.012150663882493973, 'timing/model': 2.7647532485425472, 'learning_rate/qwen_vl_interface': 9.898509918047127e-06, 'learning_rate/action_model': 9.898509918047126e-05, 'epoch': 0.87}) +[2026-06-19 17:53:06,725][starVLA.training.train_starvla][INFO] - Step 424, Loss: {'train/loss': 0.33179599046707153, 'train/grad_norm_pre_clip': 9.881490707397461, 'timing/data': 0.012246783822774887, 'timing/model': 2.783193776383996, 'learning_rate/qwen_vl_interface': 9.897882788582286e-06, 'learning_rate/action_model': 9.897882788582287e-05, 'epoch': 0.87}) +[2026-06-19 17:53:17,514][starVLA.training.train_starvla][INFO] - Step 425, Loss: {'train/loss': 0.511569619178772, 'train/grad_norm_pre_clip': 12.956796646118164, 'timing/data': 0.012460272759199142, 'timing/model': 2.7665425706654787, 'learning_rate/qwen_vl_interface': 9.897253748548398e-06, 'learning_rate/action_model': 9.897253748548399e-05, 'epoch': 0.87}) +[2026-06-19 17:53:28,323][starVLA.training.train_starvla][INFO] - Step 426, Loss: {'train/loss': 0.2079925835132599, 'train/grad_norm_pre_clip': 6.153552532196045, 'timing/data': 0.012425899505615234, 'timing/model': 2.7668065428733826, 'learning_rate/qwen_vl_interface': 9.896622798204035e-06, 'learning_rate/action_model': 9.896622798204035e-05, 'epoch': 0.87}) +[2026-06-19 17:53:39,087][starVLA.training.train_starvla][INFO] - Step 427, Loss: {'train/loss': 0.4430357813835144, 'train/grad_norm_pre_clip': 8.72072696685791, 'timing/data': 0.012668909505009651, 'timing/model': 2.7313386593014, 'learning_rate/qwen_vl_interface': 9.89598993780856e-06, 'learning_rate/action_model': 9.89598993780856e-05, 'epoch': 0.88}) +[2026-06-19 17:53:49,923][starVLA.training.train_starvla][INFO] - Step 428, Loss: {'train/loss': 0.37807363271713257, 'train/grad_norm_pre_clip': 11.469555854797363, 'timing/data': 0.012688487768173218, 'timing/model': 2.7725333273410797, 'learning_rate/qwen_vl_interface': 9.895355167622113e-06, 'learning_rate/action_model': 9.895355167622113e-05, 'epoch': 0.88}) +[2026-06-19 17:54:00,722][starVLA.training.train_starvla][INFO] - Step 429, Loss: {'train/loss': 0.3362572491168976, 'train/grad_norm_pre_clip': 8.991414070129395, 'timing/data': 0.012237101793289185, 'timing/model': 2.7784754429012537, 'learning_rate/qwen_vl_interface': 9.894718487905629e-06, 'learning_rate/action_model': 9.894718487905629e-05, 'epoch': 0.88}) +[2026-06-19 17:54:11,512][starVLA.training.train_starvla][INFO] - Step 430, Loss: {'train/loss': 0.29945382475852966, 'train/grad_norm_pre_clip': 7.574662208557129, 'timing/data': 0.012237364426255226, 'timing/model': 2.7618664670735598, 'learning_rate/qwen_vl_interface': 9.89407989892082e-06, 'learning_rate/action_model': 9.894079898920821e-05, 'epoch': 0.88}) +[2026-06-19 17:54:22,319][starVLA.training.train_starvla][INFO] - Step 431, Loss: {'train/loss': 0.13963495194911957, 'train/grad_norm_pre_clip': 6.023941516876221, 'timing/data': 0.01305600255727768, 'timing/model': 2.7718212883919477, 'learning_rate/qwen_vl_interface': 9.893439400930187e-06, 'learning_rate/action_model': 9.893439400930187e-05, 'epoch': 0.89}) +[2026-06-19 17:54:33,119][starVLA.training.train_starvla][INFO] - Step 432, Loss: {'train/loss': 0.21275298297405243, 'train/grad_norm_pre_clip': 4.907741069793701, 'timing/data': 0.012675555422902107, 'timing/model': 2.7727851513773203, 'learning_rate/qwen_vl_interface': 9.892796994197014e-06, 'learning_rate/action_model': 9.892796994197014e-05, 'epoch': 0.89}) +[2026-06-19 17:54:43,819][starVLA.training.train_starvla][INFO] - Step 433, Loss: {'train/loss': 0.509277880191803, 'train/grad_norm_pre_clip': 14.191023826599121, 'timing/data': 0.012392781674861908, 'timing/model': 2.75860914029181, 'learning_rate/qwen_vl_interface': 9.892152678985372e-06, 'learning_rate/action_model': 9.892152678985372e-05, 'epoch': 0.89}) +[2026-06-19 17:54:54,719][starVLA.training.train_starvla][INFO] - Step 434, Loss: {'train/loss': 0.2770678699016571, 'train/grad_norm_pre_clip': 11.538189888000488, 'timing/data': 0.01232961192727089, 'timing/model': 2.7767995055764914, 'learning_rate/qwen_vl_interface': 9.891506455560111e-06, 'learning_rate/action_model': 9.891506455560111e-05, 'epoch': 0.89}) +[2026-06-19 17:55:05,512][starVLA.training.train_starvla][INFO] - Step 435, Loss: {'train/loss': 0.6148986220359802, 'train/grad_norm_pre_clip': 21.191997528076172, 'timing/data': 0.012336591258645058, 'timing/model': 2.7683522813022137, 'learning_rate/qwen_vl_interface': 9.890858324186872e-06, 'learning_rate/action_model': 9.890858324186872e-05, 'epoch': 0.89}) +[2026-06-19 17:55:16,320][starVLA.training.train_starvla][INFO] - Step 436, Loss: {'train/loss': 0.34033241868019104, 'train/grad_norm_pre_clip': 11.036275863647461, 'timing/data': 0.01215062290430069, 'timing/model': 2.7695304192602634, 'learning_rate/qwen_vl_interface': 9.890208285132077e-06, 'learning_rate/action_model': 9.890208285132077e-05, 'epoch': 0.9}) +[2026-06-19 17:55:27,114][starVLA.training.train_starvla][INFO] - Step 437, Loss: {'train/loss': 0.39131689071655273, 'train/grad_norm_pre_clip': 13.606972694396973, 'timing/data': 0.012363912537693977, 'timing/model': 2.7739132307469845, 'learning_rate/qwen_vl_interface': 9.889556338662933e-06, 'learning_rate/action_model': 9.889556338662932e-05, 'epoch': 0.9}) +[2026-06-19 17:55:37,922][starVLA.training.train_starvla][INFO] - Step 438, Loss: {'train/loss': 0.37991398572921753, 'train/grad_norm_pre_clip': 9.753085136413574, 'timing/data': 0.012350652366876602, 'timing/model': 2.7734989561140537, 'learning_rate/qwen_vl_interface': 9.888902485047427e-06, 'learning_rate/action_model': 9.888902485047427e-05, 'epoch': 0.9}) +[2026-06-19 17:55:48,719][starVLA.training.train_starvla][INFO] - Step 439, Loss: {'train/loss': 0.21542418003082275, 'train/grad_norm_pre_clip': 4.036428451538086, 'timing/data': 0.012501951307058334, 'timing/model': 2.767434822395444, 'learning_rate/qwen_vl_interface': 9.888246724554337e-06, 'learning_rate/action_model': 9.888246724554337e-05, 'epoch': 0.9}) +[2026-06-19 17:55:59,519][starVLA.training.train_starvla][INFO] - Step 440, Loss: {'train/loss': 0.24318404495716095, 'train/grad_norm_pre_clip': 3.3662314414978027, 'timing/data': 0.012298442423343658, 'timing/model': 2.767823714762926, 'learning_rate/qwen_vl_interface': 9.887589057453222e-06, 'learning_rate/action_model': 9.88758905745322e-05, 'epoch': 0.9}) +[2026-06-19 17:56:10,324][starVLA.training.train_starvla][INFO] - Step 441, Loss: {'train/loss': 0.2896217107772827, 'train/grad_norm_pre_clip': 7.4682488441467285, 'timing/data': 0.01220402680337429, 'timing/model': 2.775626989081502, 'learning_rate/qwen_vl_interface': 9.88692948401442e-06, 'learning_rate/action_model': 9.88692948401442e-05, 'epoch': 0.91}) +[2026-06-19 17:56:21,325][starVLA.training.train_starvla][INFO] - Step 442, Loss: {'train/loss': 0.3608107566833496, 'train/grad_norm_pre_clip': 10.521245002746582, 'timing/data': 0.013010075315833092, 'timing/model': 2.7785165198147297, 'learning_rate/qwen_vl_interface': 9.88626800450906e-06, 'learning_rate/action_model': 9.88626800450906e-05, 'epoch': 0.91}) +[2026-06-19 17:56:32,112][starVLA.training.train_starvla][INFO] - Step 443, Loss: {'train/loss': 0.32730159163475037, 'train/grad_norm_pre_clip': 9.038580894470215, 'timing/data': 0.012136915698647499, 'timing/model': 2.7591095194220543, 'learning_rate/qwen_vl_interface': 9.885604619209052e-06, 'learning_rate/action_model': 9.885604619209051e-05, 'epoch': 0.91}) +[2026-06-19 17:56:42,918][starVLA.training.train_starvla][INFO] - Step 444, Loss: {'train/loss': 0.28165918588638306, 'train/grad_norm_pre_clip': 4.28943395614624, 'timing/data': 0.0122891366481781, 'timing/model': 2.7654931396245956, 'learning_rate/qwen_vl_interface': 9.884939328387085e-06, 'learning_rate/action_model': 9.884939328387084e-05, 'epoch': 0.91}) +[2026-06-19 17:56:53,720][starVLA.training.train_starvla][INFO] - Step 445, Loss: {'train/loss': 0.2542828917503357, 'train/grad_norm_pre_clip': 5.029804706573486, 'timing/data': 0.012438913807272911, 'timing/model': 2.766885180026293, 'learning_rate/qwen_vl_interface': 9.884272132316638e-06, 'learning_rate/action_model': 9.884272132316637e-05, 'epoch': 0.91}) +[2026-06-19 17:57:04,525][starVLA.training.train_starvla][INFO] - Step 446, Loss: {'train/loss': 0.2609791159629822, 'train/grad_norm_pre_clip': 5.313543319702148, 'timing/data': 0.011913441121578217, 'timing/model': 2.7697945833206177, 'learning_rate/qwen_vl_interface': 9.883603031271968e-06, 'learning_rate/action_model': 9.883603031271969e-05, 'epoch': 0.92}) +[2026-06-19 17:57:15,326][starVLA.training.train_starvla][INFO] - Step 447, Loss: {'train/loss': 0.29274222254753113, 'train/grad_norm_pre_clip': 8.718134880065918, 'timing/data': 0.012558303773403168, 'timing/model': 2.7612322755157948, 'learning_rate/qwen_vl_interface': 9.882932025528122e-06, 'learning_rate/action_model': 9.882932025528121e-05, 'epoch': 0.92}) +[2026-06-19 17:57:26,118][starVLA.training.train_starvla][INFO] - Step 448, Loss: {'train/loss': 0.2505960762500763, 'train/grad_norm_pre_clip': 7.3244218826293945, 'timing/data': 0.012192508205771446, 'timing/model': 2.767565941438079, 'learning_rate/qwen_vl_interface': 9.882259115360918e-06, 'learning_rate/action_model': 9.882259115360917e-05, 'epoch': 0.92}) +[2026-06-19 17:57:37,044][starVLA.training.train_starvla][INFO] - Step 449, Loss: {'train/loss': 0.15240198373794556, 'train/grad_norm_pre_clip': 2.8741986751556396, 'timing/data': 0.012674951925873756, 'timing/model': 2.885906234383583, 'learning_rate/qwen_vl_interface': 9.881584301046968e-06, 'learning_rate/action_model': 9.881584301046968e-05, 'epoch': 0.92}) +[2026-06-19 17:57:47,925][starVLA.training.train_starvla][INFO] - Step 450, Loss: {'train/loss': 0.6582474112510681, 'train/grad_norm_pre_clip': 16.121936798095703, 'timing/data': 0.01240575686097145, 'timing/model': 2.7761662527918816, 'learning_rate/qwen_vl_interface': 9.880907582863662e-06, 'learning_rate/action_model': 9.880907582863662e-05, 'epoch': 0.92}) +[2026-06-19 17:57:58,722][starVLA.training.train_starvla][INFO] - Step 451, Loss: {'train/loss': 0.19396109879016876, 'train/grad_norm_pre_clip': 7.843143939971924, 'timing/data': 0.012265877798199654, 'timing/model': 2.7664063163101673, 'learning_rate/qwen_vl_interface': 9.880228961089175e-06, 'learning_rate/action_model': 9.880228961089175e-05, 'epoch': 0.93}) +[2026-06-19 17:58:09,525][starVLA.training.train_starvla][INFO] - Step 452, Loss: {'train/loss': 0.2619982957839966, 'train/grad_norm_pre_clip': 13.698404312133789, 'timing/data': 0.011949751526117325, 'timing/model': 2.771607231348753, 'learning_rate/qwen_vl_interface': 9.87954843600246e-06, 'learning_rate/action_model': 9.879548436002459e-05, 'epoch': 0.93}) +[2026-06-19 17:58:20,427][starVLA.training.train_starvla][INFO] - Step 453, Loss: {'train/loss': 0.29657837748527527, 'train/grad_norm_pre_clip': 8.88918685913086, 'timing/data': 0.012558607384562492, 'timing/model': 2.770864760503173, 'learning_rate/qwen_vl_interface': 9.878866007883256e-06, 'learning_rate/action_model': 9.878866007883257e-05, 'epoch': 0.93}) +[2026-06-19 17:58:31,227][starVLA.training.train_starvla][INFO] - Step 454, Loss: {'train/loss': 0.23924244940280914, 'train/grad_norm_pre_clip': 9.579360008239746, 'timing/data': 0.013098355382680893, 'timing/model': 2.7745615243911743, 'learning_rate/qwen_vl_interface': 9.878181677012084e-06, 'learning_rate/action_model': 9.878181677012085e-05, 'epoch': 0.93}) +[2026-06-19 17:58:42,025][starVLA.training.train_starvla][INFO] - Step 455, Loss: {'train/loss': 0.2767929434776306, 'train/grad_norm_pre_clip': 10.289191246032715, 'timing/data': 0.012306887656450272, 'timing/model': 2.7683049645274878, 'learning_rate/qwen_vl_interface': 9.877495443670247e-06, 'learning_rate/action_model': 9.877495443670246e-05, 'epoch': 0.93}) +[2026-06-19 17:58:52,826][starVLA.training.train_starvla][INFO] - Step 456, Loss: {'train/loss': 0.17052742838859558, 'train/grad_norm_pre_clip': 5.964609146118164, 'timing/data': 0.013050558045506477, 'timing/model': 2.7683435697108507, 'learning_rate/qwen_vl_interface': 9.876807308139828e-06, 'learning_rate/action_model': 9.876807308139829e-05, 'epoch': 0.94}) +[2026-06-19 17:59:03,574][starVLA.training.train_starvla][INFO] - Step 457, Loss: {'train/loss': 0.26647764444351196, 'train/grad_norm_pre_clip': 5.977085590362549, 'timing/data': 0.011446680873632431, 'timing/model': 2.744833432137966, 'learning_rate/qwen_vl_interface': 9.876117270703695e-06, 'learning_rate/action_model': 9.876117270703694e-05, 'epoch': 0.94}) +[2026-06-19 17:59:14,425][starVLA.training.train_starvla][INFO] - Step 458, Loss: {'train/loss': 0.25050702691078186, 'train/grad_norm_pre_clip': 3.8507819175720215, 'timing/data': 0.012084882706403732, 'timing/model': 2.7775603737682104, 'learning_rate/qwen_vl_interface': 9.875425331645495e-06, 'learning_rate/action_model': 9.875425331645495e-05, 'epoch': 0.94}) +[2026-06-19 17:59:25,229][starVLA.training.train_starvla][INFO] - Step 459, Loss: {'train/loss': 0.5304377675056458, 'train/grad_norm_pre_clip': 16.197473526000977, 'timing/data': 0.012188421562314034, 'timing/model': 2.7767816483974457, 'learning_rate/qwen_vl_interface': 9.874731491249658e-06, 'learning_rate/action_model': 9.874731491249657e-05, 'epoch': 0.94}) +[2026-06-19 17:59:36,124][starVLA.training.train_starvla][INFO] - Step 460, Loss: {'train/loss': 0.32340383529663086, 'train/grad_norm_pre_clip': 8.920331954956055, 'timing/data': 0.01253269799053669, 'timing/model': 2.7704919166862965, 'learning_rate/qwen_vl_interface': 9.874035749801398e-06, 'learning_rate/action_model': 9.874035749801399e-05, 'epoch': 0.94}) +[2026-06-19 17:59:47,011][starVLA.training.train_starvla][INFO] - Step 461, Loss: {'train/loss': 0.35504353046417236, 'train/grad_norm_pre_clip': 11.973381042480469, 'timing/data': 0.01234113983809948, 'timing/model': 2.7579250056296587, 'learning_rate/qwen_vl_interface': 9.873338107586704e-06, 'learning_rate/action_model': 9.873338107586705e-05, 'epoch': 0.95}) +[2026-06-19 17:59:57,822][starVLA.training.train_starvla][INFO] - Step 462, Loss: {'train/loss': 0.31260207295417786, 'train/grad_norm_pre_clip': 6.048887729644775, 'timing/data': 0.012547226622700691, 'timing/model': 2.7696694526821375, 'learning_rate/qwen_vl_interface': 9.872638564892358e-06, 'learning_rate/action_model': 9.872638564892357e-05, 'epoch': 0.95}) +[2026-06-19 18:00:08,624][starVLA.training.train_starvla][INFO] - Step 463, Loss: {'train/loss': 0.3066123425960541, 'train/grad_norm_pre_clip': 13.884530067443848, 'timing/data': 0.012081844732165337, 'timing/model': 2.7747532576322556, 'learning_rate/qwen_vl_interface': 9.871937122005905e-06, 'learning_rate/action_model': 9.871937122005905e-05, 'epoch': 0.95}) +[2026-06-19 18:00:19,427][starVLA.training.train_starvla][INFO] - Step 464, Loss: {'train/loss': 0.19021904468536377, 'train/grad_norm_pre_clip': 3.318652868270874, 'timing/data': 0.012257149443030357, 'timing/model': 2.7715832628309727, 'learning_rate/qwen_vl_interface': 9.87123377921569e-06, 'learning_rate/action_model': 9.871233779215689e-05, 'epoch': 0.95}) +[2026-06-19 18:00:30,222][starVLA.training.train_starvla][INFO] - Step 465, Loss: {'train/loss': 0.22297504544258118, 'train/grad_norm_pre_clip': 5.066164016723633, 'timing/data': 0.012261692434549332, 'timing/model': 2.7692954763770103, 'learning_rate/qwen_vl_interface': 9.870528536810825e-06, 'learning_rate/action_model': 9.870528536810825e-05, 'epoch': 0.96}) +[2026-06-19 18:00:41,023][starVLA.training.train_starvla][INFO] - Step 466, Loss: {'train/loss': 0.2769460678100586, 'train/grad_norm_pre_clip': 5.358778476715088, 'timing/data': 0.012207843363285065, 'timing/model': 2.7832721807062626, 'learning_rate/qwen_vl_interface': 9.869821395081214e-06, 'learning_rate/action_model': 9.869821395081214e-05, 'epoch': 0.96}) +[2026-06-19 18:00:51,826][starVLA.training.train_starvla][INFO] - Step 467, Loss: {'train/loss': 0.2577836215496063, 'train/grad_norm_pre_clip': 3.1033754348754883, 'timing/data': 0.011963555589318275, 'timing/model': 2.785787995904684, 'learning_rate/qwen_vl_interface': 9.869112354317532e-06, 'learning_rate/action_model': 9.869112354317532e-05, 'epoch': 0.96}) +[2026-06-19 18:01:02,621][starVLA.training.train_starvla][INFO] - Step 468, Loss: {'train/loss': 0.4257735013961792, 'train/grad_norm_pre_clip': 5.495494842529297, 'timing/data': 0.012353118509054184, 'timing/model': 2.7668709345161915, 'learning_rate/qwen_vl_interface': 9.86840141481124e-06, 'learning_rate/action_model': 9.868401414811239e-05, 'epoch': 0.96}) +[2026-06-19 18:01:13,620][starVLA.training.train_starvla][INFO] - Step 469, Loss: {'train/loss': 0.2130039483308792, 'train/grad_norm_pre_clip': 4.822216987609863, 'timing/data': 0.011831184849143028, 'timing/model': 2.778499746695161, 'learning_rate/qwen_vl_interface': 9.867688576854578e-06, 'learning_rate/action_model': 9.867688576854578e-05, 'epoch': 0.96}) +[2026-06-19 18:01:24,414][starVLA.training.train_starvla][INFO] - Step 470, Loss: {'train/loss': 0.24155043065547943, 'train/grad_norm_pre_clip': 4.648077964782715, 'timing/data': 0.011811360716819763, 'timing/model': 2.7693831603974104, 'learning_rate/qwen_vl_interface': 9.86697384074057e-06, 'learning_rate/action_model': 9.866973840740568e-05, 'epoch': 0.97}) +[2026-06-19 18:01:35,222][starVLA.training.train_starvla][INFO] - Step 471, Loss: {'train/loss': 0.17220920324325562, 'train/grad_norm_pre_clip': 4.098660945892334, 'timing/data': 0.0122237429022789, 'timing/model': 2.7726066913455725, 'learning_rate/qwen_vl_interface': 9.86625720676301e-06, 'learning_rate/action_model': 9.86625720676301e-05, 'epoch': 0.97}) +[2026-06-19 18:01:46,023][starVLA.training.train_starvla][INFO] - Step 472, Loss: {'train/loss': 0.24892255663871765, 'train/grad_norm_pre_clip': 2.9431326389312744, 'timing/data': 0.012090794742107391, 'timing/model': 2.7693607322871685, 'learning_rate/qwen_vl_interface': 9.865538675216488e-06, 'learning_rate/action_model': 9.865538675216486e-05, 'epoch': 0.97}) +[2026-06-19 18:01:56,827][starVLA.training.train_starvla][INFO] - Step 473, Loss: {'train/loss': 0.18289512395858765, 'train/grad_norm_pre_clip': 3.973400354385376, 'timing/data': 0.01231149211525917, 'timing/model': 2.7726160008460283, 'learning_rate/qwen_vl_interface': 9.864818246396358e-06, 'learning_rate/action_model': 9.864818246396358e-05, 'epoch': 0.97}) +[2026-06-19 18:02:07,623][starVLA.training.train_starvla][INFO] - Step 474, Loss: {'train/loss': 0.1573455035686493, 'train/grad_norm_pre_clip': 7.635137557983398, 'timing/data': 0.012549439445137978, 'timing/model': 2.7672897838056087, 'learning_rate/qwen_vl_interface': 9.864095920598764e-06, 'learning_rate/action_model': 9.864095920598763e-05, 'epoch': 0.97}) +[2026-06-19 18:02:18,425][starVLA.training.train_starvla][INFO] - Step 475, Loss: {'train/loss': 0.2468852400779724, 'train/grad_norm_pre_clip': 11.364233016967773, 'timing/data': 0.012535568326711655, 'timing/model': 2.7725375574082136, 'learning_rate/qwen_vl_interface': 9.86337169812063e-06, 'learning_rate/action_model': 9.863371698120629e-05, 'epoch': 0.98}) +[2026-06-19 18:02:29,329][starVLA.training.train_starvla][INFO] - Step 476, Loss: {'train/loss': 0.08471474796533585, 'train/grad_norm_pre_clip': 19.63138198852539, 'timing/data': 0.012406112626194954, 'timing/model': 2.8746301401406527, 'learning_rate/qwen_vl_interface': 9.862645579259648e-06, 'learning_rate/action_model': 9.862645579259648e-05, 'epoch': 0.98}) +[2026-06-19 18:02:40,230][starVLA.training.train_starvla][INFO] - Step 477, Loss: {'train/loss': 0.18970394134521484, 'train/grad_norm_pre_clip': 10.572940826416016, 'timing/data': 0.012459030374884605, 'timing/model': 2.7779807541519403, 'learning_rate/qwen_vl_interface': 9.861917564314308e-06, 'learning_rate/action_model': 9.861917564314308e-05, 'epoch': 0.98}) +[2026-06-19 18:02:51,122][starVLA.training.train_starvla][INFO] - Step 478, Loss: {'train/loss': 0.5971349477767944, 'train/grad_norm_pre_clip': 28.79168128967285, 'timing/data': 0.012464333325624466, 'timing/model': 2.7736414205282927, 'learning_rate/qwen_vl_interface': 9.861187653583862e-06, 'learning_rate/action_model': 9.861187653583862e-05, 'epoch': 0.98}) +[2026-06-19 18:03:01,926][starVLA.training.train_starvla][INFO] - Step 479, Loss: {'train/loss': 0.6020956039428711, 'train/grad_norm_pre_clip': 18.976762771606445, 'timing/data': 0.012182474136352539, 'timing/model': 2.773840930312872, 'learning_rate/qwen_vl_interface': 9.860455847368352e-06, 'learning_rate/action_model': 9.860455847368351e-05, 'epoch': 0.98}) +[2026-06-19 18:03:12,724][starVLA.training.train_starvla][INFO] - Step 480, Loss: {'train/loss': 0.06031307205557823, 'train/grad_norm_pre_clip': 3.360717296600342, 'timing/data': 0.012461161240935326, 'timing/model': 2.7722709141671658, 'learning_rate/qwen_vl_interface': 9.859722145968595e-06, 'learning_rate/action_model': 9.859722145968595e-05, 'epoch': 0.99}) +[2026-06-19 18:03:23,523][starVLA.training.train_starvla][INFO] - Step 481, Loss: {'train/loss': 0.3703251779079437, 'train/grad_norm_pre_clip': 11.688996315002441, 'timing/data': 0.012042045593261719, 'timing/model': 2.779571918770671, 'learning_rate/qwen_vl_interface': 9.858986549686188e-06, 'learning_rate/action_model': 9.858986549686188e-05, 'epoch': 0.99}) +[2026-06-19 18:03:34,325][starVLA.training.train_starvla][INFO] - Step 482, Loss: {'train/loss': 0.32053491473197937, 'train/grad_norm_pre_clip': 26.065963745117188, 'timing/data': 0.012589208781719208, 'timing/model': 2.7720103226602077, 'learning_rate/qwen_vl_interface': 9.85824905882351e-06, 'learning_rate/action_model': 9.858249058823509e-05, 'epoch': 0.99}) +[2026-06-19 18:03:45,125][starVLA.training.train_starvla][INFO] - Step 483, Loss: {'train/loss': 0.3744065463542938, 'train/grad_norm_pre_clip': 16.108131408691406, 'timing/data': 0.011791747063398361, 'timing/model': 2.7720906920731068, 'learning_rate/qwen_vl_interface': 9.857509673683709e-06, 'learning_rate/action_model': 9.857509673683708e-05, 'epoch': 0.99}) +[2026-06-19 18:03:55,924][starVLA.training.train_starvla][INFO] - Step 484, Loss: {'train/loss': 0.1151052713394165, 'train/grad_norm_pre_clip': 5.657758712768555, 'timing/data': 0.011775979772210121, 'timing/model': 2.772748274728656, 'learning_rate/qwen_vl_interface': 9.856768394570725e-06, 'learning_rate/action_model': 9.856768394570723e-05, 'epoch': 0.99}) +[2026-06-19 18:04:06,725][starVLA.training.train_starvla][INFO] - Step 485, Loss: {'train/loss': 0.26509544253349304, 'train/grad_norm_pre_clip': 12.817876815795898, 'timing/data': 0.012450622394680977, 'timing/model': 2.7825934179127216, 'learning_rate/qwen_vl_interface': 9.856025221789264e-06, 'learning_rate/action_model': 9.856025221789264e-05, 'epoch': 1.0}) +[2026-06-19 18:04:17,524][starVLA.training.train_starvla][INFO] - Step 486, Loss: {'train/loss': 0.1565287858247757, 'train/grad_norm_pre_clip': 9.28158187866211, 'timing/data': 0.011820260435342789, 'timing/model': 2.7719248421490192, 'learning_rate/qwen_vl_interface': 9.85528015564482e-06, 'learning_rate/action_model': 9.85528015564482e-05, 'epoch': 1.0}) +[2026-06-19 18:04:28,350][starVLA.training.train_starvla][INFO] - Step 487, Loss: {'train/loss': 0.05924367159605026, 'train/grad_norm_pre_clip': 3.3355114459991455, 'timing/data': 1.0971672926098108, 'timing/model': 1.7135899495333433, 'learning_rate/qwen_vl_interface': 9.854533196443661e-06, 'learning_rate/action_model': 9.854533196443661e-05, 'epoch': 1.0}) +[2026-06-19 18:04:59,831][starVLA.training.train_starvla][INFO] - Step 488, Loss: {'train/loss': 0.4002188742160797, 'train/grad_norm_pre_clip': 19.228174209594727, 'timing/data': 0.028800638392567635, 'timing/model': 2.7652703430503607, 'learning_rate/qwen_vl_interface': 9.853784344492833e-06, 'learning_rate/action_model': 9.853784344492833e-05, 'epoch': 1.0}) +[2026-06-19 18:05:10,719][starVLA.training.train_starvla][INFO] - Step 489, Loss: {'train/loss': 0.15046504139900208, 'train/grad_norm_pre_clip': 13.652653694152832, 'timing/data': 0.027447305619716644, 'timing/model': 2.754874359816313, 'learning_rate/qwen_vl_interface': 9.853033600100162e-06, 'learning_rate/action_model': 9.853033600100162e-05, 'epoch': 1.0}) +[2026-06-19 18:05:21,511][starVLA.training.train_starvla][INFO] - Step 490, Loss: {'train/loss': 0.3121340870857239, 'train/grad_norm_pre_clip': 12.455655097961426, 'timing/data': 0.025923412293195724, 'timing/model': 2.748982695862651, 'learning_rate/qwen_vl_interface': 9.85228096357425e-06, 'learning_rate/action_model': 9.852280963574249e-05, 'epoch': 1.01}) +[2026-06-19 18:05:32,307][starVLA.training.train_starvla][INFO] - Step 491, Loss: {'train/loss': 0.45241793990135193, 'train/grad_norm_pre_clip': 14.69688892364502, 'timing/data': 0.022771921008825302, 'timing/model': 2.747906332835555, 'learning_rate/qwen_vl_interface': 9.851526435224475e-06, 'learning_rate/action_model': 9.851526435224474e-05, 'epoch': 1.01}) +[2026-06-19 18:05:43,320][starVLA.training.train_starvla][INFO] - Step 492, Loss: {'train/loss': 0.39492738246917725, 'train/grad_norm_pre_clip': 11.687771797180176, 'timing/data': 0.028777210041880608, 'timing/model': 2.754397500306368, 'learning_rate/qwen_vl_interface': 9.850770015360999e-06, 'learning_rate/action_model': 9.850770015360999e-05, 'epoch': 1.01}) +[2026-06-19 18:05:54,122][starVLA.training.train_starvla][INFO] - Step 493, Loss: {'train/loss': 0.1912427693605423, 'train/grad_norm_pre_clip': 4.304666519165039, 'timing/data': 0.022891629487276077, 'timing/model': 2.7614712342619896, 'learning_rate/qwen_vl_interface': 9.850011704294755e-06, 'learning_rate/action_model': 9.850011704294755e-05, 'epoch': 1.01}) +[2026-06-19 18:06:05,063][starVLA.training.train_starvla][INFO] - Step 494, Loss: {'train/loss': 0.18785981833934784, 'train/grad_norm_pre_clip': 5.646547317504883, 'timing/data': 0.025759074836969376, 'timing/model': 2.7550087440758944, 'learning_rate/qwen_vl_interface': 9.849251502337458e-06, 'learning_rate/action_model': 9.849251502337458e-05, 'epoch': 1.01}) +[2026-06-19 18:06:16,003][starVLA.training.train_starvla][INFO] - Step 495, Loss: {'train/loss': 0.21133165061473846, 'train/grad_norm_pre_clip': 5.16396951675415, 'timing/data': 0.03263893723487854, 'timing/model': 2.830378418788314, 'learning_rate/qwen_vl_interface': 9.848489409801598e-06, 'learning_rate/action_model': 9.848489409801598e-05, 'epoch': 1.02}) +[2026-06-19 18:06:26,821][starVLA.training.train_starvla][INFO] - Step 496, Loss: {'train/loss': 0.23849239945411682, 'train/grad_norm_pre_clip': 6.026435375213623, 'timing/data': 0.023232046514749527, 'timing/model': 2.757514448836446, 'learning_rate/qwen_vl_interface': 9.84772542700044e-06, 'learning_rate/action_model': 9.84772542700044e-05, 'epoch': 1.02}) +[2026-06-19 18:06:37,621][starVLA.training.train_starvla][INFO] - Step 497, Loss: {'train/loss': 0.36039552092552185, 'train/grad_norm_pre_clip': 4.257076263427734, 'timing/data': 0.02450033836066723, 'timing/model': 2.751059915870428, 'learning_rate/qwen_vl_interface': 9.846959554248032e-06, 'learning_rate/action_model': 9.84695955424803e-05, 'epoch': 1.02}) +[2026-06-19 18:06:48,415][starVLA.training.train_starvla][INFO] - Step 498, Loss: {'train/loss': 0.21585452556610107, 'train/grad_norm_pre_clip': 4.151798725128174, 'timing/data': 0.029405003413558006, 'timing/model': 2.7553582079708576, 'learning_rate/qwen_vl_interface': 9.846191791859192e-06, 'learning_rate/action_model': 9.846191791859192e-05, 'epoch': 1.02}) +[2026-06-19 18:06:59,218][starVLA.training.train_starvla][INFO] - Step 499, Loss: {'train/loss': 0.24835512042045593, 'train/grad_norm_pre_clip': 5.209338665008545, 'timing/data': 0.026234911754727364, 'timing/model': 2.7544660102576017, 'learning_rate/qwen_vl_interface': 9.845422140149518e-06, 'learning_rate/action_model': 9.845422140149519e-05, 'epoch': 1.02}) +[2026-06-19 18:07:10,022][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 500 +[2026-06-19 18:14:44,931][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 500 +[2026-06-19 18:14:44,932][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 500 +[2026-06-19 18:23:09,620][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 500 +[2026-06-19 18:23:09,622][starVLA.training.train_starvla][INFO] - Step 500, Loss: {'train/loss': 0.3152812123298645, 'train/grad_norm_pre_clip': 8.229748725891113, 'eval/loss': 0.3090752363204956, 'eval/latency_0/loss': 0.12985940277576447, 'eval/latency_1/loss': 0.2266828417778015, 'eval/latency_2/loss': 0.33017057180404663, 'eval/latency_3/loss': 0.3567911684513092, 'eval/latency_4/loss': 0.493667334318161, 'eval/flappy/loss': 0.3090752363204956, 'eval/flappy/latency_0/loss': 0.12985940277576447, 'eval/flappy/latency_1/loss': 0.2266828417778015, 'eval/flappy/latency_2/loss': 0.33017057180404663, 'eval/flappy/latency_3/loss': 0.3567911684513092, 'eval/flappy/latency_4/loss': 0.493667334318161, 'eval/flap_precision': 0.8421052631578947, 'eval/flap_recall': 0.46545454545454545, 'eval/flap_f1': 0.5995316159250585, 'eval/flap_rate_model': 0.042222222222222223, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.5995316159250585, 'eval/latency_0/flap_precision': 0.8421052631578947, 'eval/latency_0/flap_recall': 0.46545454545454545, 'eval/latency_0/flap_f1': 0.5995316159250585, 'eval/latency_0/flap_rate_model': 0.042222222222222223, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.5995316159250585, 'timing/data': 0.02364226058125496, 'timing/model': 2.7634841445833445, 'learning_rate/qwen_vl_interface': 9.84465059943539e-06, 'learning_rate/action_model': 9.84465059943539e-05, 'epoch': 1.03}) +[2026-06-19 18:23:09,623][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_500_state +[2026-06-19 18:23:09,649][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-19 18:23:15,154][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_500_state/model.safetensors +[2026-06-19 18:23:22,356][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_500_state/optimizer.bin +[2026-06-19 18:23:22,357][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_500_state/sampler.bin +[2026-06-19 18:23:22,357][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_500_state/sampler_1.bin +[2026-06-19 18:23:22,358][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_500_state/random_states_0.pkl +[2026-06-19 18:23:22,612][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 500 +[2026-06-19 18:23:22,613][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-19 18:23:22,627][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-19 18:23:33,410][starVLA.training.train_starvla][INFO] - Step 501, Loss: {'train/loss': 0.3260020613670349, 'train/grad_norm_pre_clip': 7.794963359832764, 'timing/data': 0.02393270470201969, 'timing/model': 2.7590560875833035, 'learning_rate/qwen_vl_interface': 9.843877170033956e-06, 'learning_rate/action_model': 9.843877170033954e-05, 'epoch': 1.03}) +[2026-06-19 18:23:44,009][starVLA.training.train_starvla][INFO] - Step 502, Loss: {'train/loss': 0.1764415204524994, 'train/grad_norm_pre_clip': 7.135547637939453, 'timing/data': 0.0224679633975029, 'timing/model': 2.7412298563867807, 'learning_rate/qwen_vl_interface': 9.84310185226314e-06, 'learning_rate/action_model': 9.84310185226314e-05, 'epoch': 1.03}) +[2026-06-19 18:23:54,710][starVLA.training.train_starvla][INFO] - Step 503, Loss: {'train/loss': 0.17468519508838654, 'train/grad_norm_pre_clip': 5.797946453094482, 'timing/data': 0.02282682992517948, 'timing/model': 2.7590944450348616, 'learning_rate/qwen_vl_interface': 9.842324646441652e-06, 'learning_rate/action_model': 9.842324646441652e-05, 'epoch': 1.03}) +[2026-06-19 18:24:05,505][starVLA.training.train_starvla][INFO] - Step 504, Loss: {'train/loss': 0.13999220728874207, 'train/grad_norm_pre_clip': 3.1861789226531982, 'timing/data': 0.022321585565805435, 'timing/model': 2.7598878387361765, 'learning_rate/qwen_vl_interface': 9.841545552888968e-06, 'learning_rate/action_model': 9.841545552888967e-05, 'epoch': 1.04}) +[2026-06-19 18:24:16,213][starVLA.training.train_starvla][INFO] - Step 505, Loss: {'train/loss': 0.26643386483192444, 'train/grad_norm_pre_clip': 9.84546947479248, 'timing/data': 0.02361239306628704, 'timing/model': 2.755502112209797, 'learning_rate/qwen_vl_interface': 9.840764571925346e-06, 'learning_rate/action_model': 9.840764571925345e-05, 'epoch': 1.04}) +[2026-06-19 18:24:26,905][starVLA.training.train_starvla][INFO] - Step 506, Loss: {'train/loss': 0.20502348244190216, 'train/grad_norm_pre_clip': 9.34613037109375, 'timing/data': 0.02801552042365074, 'timing/model': 2.725606506690383, 'learning_rate/qwen_vl_interface': 9.839981703871815e-06, 'learning_rate/action_model': 9.839981703871814e-05, 'epoch': 1.04}) +[2026-06-19 18:24:37,913][starVLA.training.train_starvla][INFO] - Step 507, Loss: {'train/loss': 0.3129970133304596, 'train/grad_norm_pre_clip': 14.446449279785156, 'timing/data': 0.025344738736748695, 'timing/model': 2.7640809677541256, 'learning_rate/qwen_vl_interface': 9.839196949050185e-06, 'learning_rate/action_model': 9.839196949050185e-05, 'epoch': 1.04}) +[2026-06-19 18:24:48,709][starVLA.training.train_starvla][INFO] - Step 508, Loss: {'train/loss': 0.30793046951293945, 'train/grad_norm_pre_clip': 11.564813613891602, 'timing/data': 0.02681894414126873, 'timing/model': 2.748752696439624, 'learning_rate/qwen_vl_interface': 9.838410307783038e-06, 'learning_rate/action_model': 9.838410307783038e-05, 'epoch': 1.04}) +[2026-06-19 18:24:59,408][starVLA.training.train_starvla][INFO] - Step 509, Loss: {'train/loss': 0.17028959095478058, 'train/grad_norm_pre_clip': 9.694350242614746, 'timing/data': 0.022381365299224854, 'timing/model': 2.7510789949446917, 'learning_rate/qwen_vl_interface': 9.837621780393734e-06, 'learning_rate/action_model': 9.837621780393733e-05, 'epoch': 1.05}) +[2026-06-19 18:25:10,297][starVLA.training.train_starvla][INFO] - Step 510, Loss: {'train/loss': 0.25506970286369324, 'train/grad_norm_pre_clip': 7.41010856628418, 'timing/data': 0.022622501477599144, 'timing/model': 2.839185370132327, 'learning_rate/qwen_vl_interface': 9.836831367206406e-06, 'learning_rate/action_model': 9.836831367206406e-05, 'epoch': 1.05}) +[2026-06-19 18:25:21,112][starVLA.training.train_starvla][INFO] - Step 511, Loss: {'train/loss': 0.623613178730011, 'train/grad_norm_pre_clip': 14.926711082458496, 'timing/data': 0.022113587707281113, 'timing/model': 2.7598704360425472, 'learning_rate/qwen_vl_interface': 9.836039068545962e-06, 'learning_rate/action_model': 9.836039068545961e-05, 'epoch': 1.05}) +[2026-06-19 18:25:31,791][starVLA.training.train_starvla][INFO] - Step 512, Loss: {'train/loss': 0.297836035490036, 'train/grad_norm_pre_clip': 5.5445475578308105, 'timing/data': 0.02298590913414955, 'timing/model': 2.7314511239528656, 'learning_rate/qwen_vl_interface': 9.835244884738086e-06, 'learning_rate/action_model': 9.835244884738085e-05, 'epoch': 1.05}) +[2026-06-19 18:25:42,507][starVLA.training.train_starvla][INFO] - Step 513, Loss: {'train/loss': 0.24040713906288147, 'train/grad_norm_pre_clip': 9.594143867492676, 'timing/data': 0.022504352033138275, 'timing/model': 2.7722566202282906, 'learning_rate/qwen_vl_interface': 9.83444881610924e-06, 'learning_rate/action_model': 9.834448816109238e-05, 'epoch': 1.05}) +[2026-06-19 18:25:53,364][starVLA.training.train_starvla][INFO] - Step 514, Loss: {'train/loss': 0.3059457242488861, 'train/grad_norm_pre_clip': 8.121133804321289, 'timing/data': 0.022622019052505493, 'timing/model': 2.9418067894876003, 'learning_rate/qwen_vl_interface': 9.833650862986652e-06, 'learning_rate/action_model': 9.833650862986653e-05, 'epoch': 1.06}) +[2026-06-19 18:26:04,209][starVLA.training.train_starvla][INFO] - Step 515, Loss: {'train/loss': 0.4557589888572693, 'train/grad_norm_pre_clip': 14.869263648986816, 'timing/data': 0.022636720910668373, 'timing/model': 2.7757066767662764, 'learning_rate/qwen_vl_interface': 9.832851025698335e-06, 'learning_rate/action_model': 9.832851025698335e-05, 'epoch': 1.06}) +[2026-06-19 18:26:14,812][starVLA.training.train_starvla][INFO] - Step 516, Loss: {'train/loss': 0.2500070631504059, 'train/grad_norm_pre_clip': 15.783557891845703, 'timing/data': 0.021869990974664688, 'timing/model': 2.720569718629122, 'learning_rate/qwen_vl_interface': 9.832049304573072e-06, 'learning_rate/action_model': 9.832049304573072e-05, 'epoch': 1.06}) +[2026-06-19 18:26:25,613][starVLA.training.train_starvla][INFO] - Step 517, Loss: {'train/loss': 0.23537491261959076, 'train/grad_norm_pre_clip': 4.69284200668335, 'timing/data': 0.022808941081166267, 'timing/model': 2.7569565437734127, 'learning_rate/qwen_vl_interface': 9.831245699940419e-06, 'learning_rate/action_model': 9.831245699940418e-05, 'epoch': 1.06}) +[2026-06-19 18:26:36,432][starVLA.training.train_starvla][INFO] - Step 518, Loss: {'train/loss': 0.22162304818630219, 'train/grad_norm_pre_clip': 6.461904048919678, 'timing/data': 0.022604791447520256, 'timing/model': 2.7128314282745123, 'learning_rate/qwen_vl_interface': 9.830440212130707e-06, 'learning_rate/action_model': 9.830440212130708e-05, 'epoch': 1.06}) +[2026-06-19 18:26:47,119][starVLA.training.train_starvla][INFO] - Step 519, Loss: {'train/loss': 0.36689814925193787, 'train/grad_norm_pre_clip': 8.010037422180176, 'timing/data': 0.023260803893208504, 'timing/model': 2.7551790680736303, 'learning_rate/qwen_vl_interface': 9.829632841475041e-06, 'learning_rate/action_model': 9.829632841475041e-05, 'epoch': 1.07}) +[2026-06-19 18:26:57,918][starVLA.training.train_starvla][INFO] - Step 520, Loss: {'train/loss': 0.2150816023349762, 'train/grad_norm_pre_clip': 6.011682987213135, 'timing/data': 0.02233683504164219, 'timing/model': 2.7683440186083317, 'learning_rate/qwen_vl_interface': 9.828823588305305e-06, 'learning_rate/action_model': 9.828823588305303e-05, 'epoch': 1.07}) +[2026-06-19 18:27:08,721][starVLA.training.train_starvla][INFO] - Step 521, Loss: {'train/loss': 0.31778842210769653, 'train/grad_norm_pre_clip': 6.4356160163879395, 'timing/data': 0.022572705522179604, 'timing/model': 2.7700425144284964, 'learning_rate/qwen_vl_interface': 9.828012452954146e-06, 'learning_rate/action_model': 9.828012452954145e-05, 'epoch': 1.07}) +[2026-06-19 18:27:19,413][starVLA.training.train_starvla][INFO] - Step 522, Loss: {'train/loss': 0.3766616880893707, 'train/grad_norm_pre_clip': 6.367791175842285, 'timing/data': 0.02221931517124176, 'timing/model': 2.7146436981856823, 'learning_rate/qwen_vl_interface': 9.827199435754995e-06, 'learning_rate/action_model': 9.827199435754995e-05, 'epoch': 1.07}) +[2026-06-19 18:27:30,218][starVLA.training.train_starvla][INFO] - Step 523, Loss: {'train/loss': 0.2908064126968384, 'train/grad_norm_pre_clip': 4.565536975860596, 'timing/data': 0.020903652533888817, 'timing/model': 2.7672092001885176, 'learning_rate/qwen_vl_interface': 9.826384537042052e-06, 'learning_rate/action_model': 9.826384537042052e-05, 'epoch': 1.07}) +[2026-06-19 18:27:40,920][starVLA.training.train_starvla][INFO] - Step 524, Loss: {'train/loss': 0.15479254722595215, 'train/grad_norm_pre_clip': 4.237898826599121, 'timing/data': 0.023332374170422554, 'timing/model': 2.7891862988471985, 'learning_rate/qwen_vl_interface': 9.825567757150291e-06, 'learning_rate/action_model': 9.82556775715029e-05, 'epoch': 1.08}) +[2026-06-19 18:27:51,621][starVLA.training.train_starvla][INFO] - Step 525, Loss: {'train/loss': 0.3058893382549286, 'train/grad_norm_pre_clip': 5.83422327041626, 'timing/data': 0.02161778323352337, 'timing/model': 2.7249219454824924, 'learning_rate/qwen_vl_interface': 9.82474909641546e-06, 'learning_rate/action_model': 9.82474909641546e-05, 'epoch': 1.08}) +[2026-06-19 18:28:02,414][starVLA.training.train_starvla][INFO] - Step 526, Loss: {'train/loss': 0.3194541037082672, 'train/grad_norm_pre_clip': 4.670807361602783, 'timing/data': 0.023324783891439438, 'timing/model': 2.7556002270430326, 'learning_rate/qwen_vl_interface': 9.823928555174078e-06, 'learning_rate/action_model': 9.823928555174077e-05, 'epoch': 1.08}) +[2026-06-19 18:28:13,214][starVLA.training.train_starvla][INFO] - Step 527, Loss: {'train/loss': 0.47203657031059265, 'train/grad_norm_pre_clip': 8.167013168334961, 'timing/data': 0.025029582902789116, 'timing/model': 2.763641808182001, 'learning_rate/qwen_vl_interface': 9.82310613376344e-06, 'learning_rate/action_model': 9.82310613376344e-05, 'epoch': 1.08}) +[2026-06-19 18:28:24,017][starVLA.training.train_starvla][INFO] - Step 528, Loss: {'train/loss': 0.36098524928092957, 'train/grad_norm_pre_clip': 6.636391639709473, 'timing/data': 0.02256925217807293, 'timing/model': 2.7549760285764933, 'learning_rate/qwen_vl_interface': 9.822281832521613e-06, 'learning_rate/action_model': 9.822281832521613e-05, 'epoch': 1.08}) +[2026-06-19 18:28:34,807][starVLA.training.train_starvla][INFO] - Step 529, Loss: {'train/loss': 0.2257951945066452, 'train/grad_norm_pre_clip': 5.073829650878906, 'timing/data': 0.025179963558912277, 'timing/model': 2.7438988368958235, 'learning_rate/qwen_vl_interface': 9.821455651787432e-06, 'learning_rate/action_model': 9.821455651787432e-05, 'epoch': 1.09}) +[2026-06-19 18:28:45,718][starVLA.training.train_starvla][INFO] - Step 530, Loss: {'train/loss': 0.268410861492157, 'train/grad_norm_pre_clip': 3.1395833492279053, 'timing/data': 0.023126116022467613, 'timing/model': 2.741414275020361, 'learning_rate/qwen_vl_interface': 9.820627591900514e-06, 'learning_rate/action_model': 9.820627591900513e-05, 'epoch': 1.09}) +[2026-06-19 18:28:56,519][starVLA.training.train_starvla][INFO] - Step 531, Loss: {'train/loss': 0.2573440968990326, 'train/grad_norm_pre_clip': 8.255439758300781, 'timing/data': 0.02270607091486454, 'timing/model': 2.7698151040822268, 'learning_rate/qwen_vl_interface': 9.819797653201238e-06, 'learning_rate/action_model': 9.819797653201238e-05, 'epoch': 1.09}) +[2026-06-19 18:29:07,220][starVLA.training.train_starvla][INFO] - Step 532, Loss: {'train/loss': 0.3506838381290436, 'train/grad_norm_pre_clip': 9.470800399780273, 'timing/data': 0.025173671543598175, 'timing/model': 2.7418259643018246, 'learning_rate/qwen_vl_interface': 9.818965836030765e-06, 'learning_rate/action_model': 9.818965836030764e-05, 'epoch': 1.09}) +[2026-06-19 18:29:18,156][starVLA.training.train_starvla][INFO] - Step 533, Loss: {'train/loss': 0.2803981602191925, 'train/grad_norm_pre_clip': 7.845792770385742, 'timing/data': 0.0257264357060194, 'timing/model': 2.9761997368186712, 'learning_rate/qwen_vl_interface': 9.81813214073102e-06, 'learning_rate/action_model': 9.81813214073102e-05, 'epoch': 1.09}) +[2026-06-19 18:29:28,921][starVLA.training.train_starvla][INFO] - Step 534, Loss: {'train/loss': 0.439352422952652, 'train/grad_norm_pre_clip': 12.323283195495605, 'timing/data': 0.0225471630692482, 'timing/model': 2.772530607879162, 'learning_rate/qwen_vl_interface': 9.817296567644709e-06, 'learning_rate/action_model': 9.817296567644709e-05, 'epoch': 1.1}) +[2026-06-19 18:29:39,720][starVLA.training.train_starvla][INFO] - Step 535, Loss: {'train/loss': 0.25854992866516113, 'train/grad_norm_pre_clip': 5.17099142074585, 'timing/data': 0.02273509092628956, 'timing/model': 2.768476678058505, 'learning_rate/qwen_vl_interface': 9.816459117115296e-06, 'learning_rate/action_model': 9.816459117115297e-05, 'epoch': 1.1}) +[2026-06-19 18:29:50,507][starVLA.training.train_starvla][INFO] - Step 536, Loss: {'train/loss': 0.4394424855709076, 'train/grad_norm_pre_clip': 10.711019515991211, 'timing/data': 0.022677328437566757, 'timing/model': 2.7498339787125587, 'learning_rate/qwen_vl_interface': 9.815619789487035e-06, 'learning_rate/action_model': 9.815619789487034e-05, 'epoch': 1.1}) +[2026-06-19 18:30:01,311][starVLA.training.train_starvla][INFO] - Step 537, Loss: {'train/loss': 0.32266509532928467, 'train/grad_norm_pre_clip': 13.763830184936523, 'timing/data': 0.02533862367272377, 'timing/model': 2.7383672017604113, 'learning_rate/qwen_vl_interface': 9.814778585104933e-06, 'learning_rate/action_model': 9.814778585104933e-05, 'epoch': 1.1}) +[2026-06-19 18:30:12,021][starVLA.training.train_starvla][INFO] - Step 538, Loss: {'train/loss': 0.68889981508255, 'train/grad_norm_pre_clip': 16.1527099609375, 'timing/data': 0.0252546314150095, 'timing/model': 2.7650691475719213, 'learning_rate/qwen_vl_interface': 9.813935504314784e-06, 'learning_rate/action_model': 9.813935504314784e-05, 'epoch': 1.1}) +[2026-06-19 18:30:22,721][starVLA.training.train_starvla][INFO] - Step 539, Loss: {'train/loss': 0.5240870118141174, 'train/grad_norm_pre_clip': 22.600143432617188, 'timing/data': 0.022743284702301025, 'timing/model': 2.772767862305045, 'learning_rate/qwen_vl_interface': 9.813090547463144e-06, 'learning_rate/action_model': 9.813090547463143e-05, 'epoch': 1.11}) +[2026-06-19 18:30:33,498][starVLA.training.train_starvla][INFO] - Step 540, Loss: {'train/loss': 0.2671268880367279, 'train/grad_norm_pre_clip': 10.007970809936523, 'timing/data': 0.022788437083363533, 'timing/model': 2.7467330526560545, 'learning_rate/qwen_vl_interface': 9.812243714897342e-06, 'learning_rate/action_model': 9.812243714897343e-05, 'epoch': 1.11}) +[2026-06-19 18:30:44,312][starVLA.training.train_starvla][INFO] - Step 541, Loss: {'train/loss': 0.5012385249137878, 'train/grad_norm_pre_clip': 10.825936317443848, 'timing/data': 0.023277647793293, 'timing/model': 2.7451821360737085, 'learning_rate/qwen_vl_interface': 9.811395006965481e-06, 'learning_rate/action_model': 9.81139500696548e-05, 'epoch': 1.11}) +[2026-06-19 18:30:55,122][starVLA.training.train_starvla][INFO] - Step 542, Loss: {'train/loss': 0.2152719646692276, 'train/grad_norm_pre_clip': 3.8320586681365967, 'timing/data': 0.022580446675419807, 'timing/model': 2.771016323938966, 'learning_rate/qwen_vl_interface': 9.81054442401643e-06, 'learning_rate/action_model': 9.81054442401643e-05, 'epoch': 1.11}) +[2026-06-19 18:31:05,926][starVLA.training.train_starvla][INFO] - Step 543, Loss: {'train/loss': 0.18142342567443848, 'train/grad_norm_pre_clip': 5.594169616699219, 'timing/data': 0.02296116016805172, 'timing/model': 2.7633736338466406, 'learning_rate/qwen_vl_interface': 9.809691966399834e-06, 'learning_rate/action_model': 9.809691966399835e-05, 'epoch': 1.12}) +[2026-06-19 18:31:16,719][starVLA.training.train_starvla][INFO] - Step 544, Loss: {'train/loss': 0.21777211129665375, 'train/grad_norm_pre_clip': 6.3661041259765625, 'timing/data': 0.0232646856456995, 'timing/model': 2.752994643524289, 'learning_rate/qwen_vl_interface': 9.808837634466104e-06, 'learning_rate/action_model': 9.808837634466104e-05, 'epoch': 1.12}) +[2026-06-19 18:31:27,523][starVLA.training.train_starvla][INFO] - Step 545, Loss: {'train/loss': 0.3172628581523895, 'train/grad_norm_pre_clip': 7.581732273101807, 'timing/data': 0.022645892575383186, 'timing/model': 2.7638092134147882, 'learning_rate/qwen_vl_interface': 9.807981428566428e-06, 'learning_rate/action_model': 9.807981428566427e-05, 'epoch': 1.12}) +[2026-06-19 18:31:38,325][starVLA.training.train_starvla][INFO] - Step 546, Loss: {'train/loss': 0.1662718653678894, 'train/grad_norm_pre_clip': 3.7706031799316406, 'timing/data': 0.022936608642339706, 'timing/model': 2.7591838855296373, 'learning_rate/qwen_vl_interface': 9.807123349052754e-06, 'learning_rate/action_model': 9.807123349052753e-05, 'epoch': 1.12}) +[2026-06-19 18:31:49,123][starVLA.training.train_starvla][INFO] - Step 547, Loss: {'train/loss': 0.5387831330299377, 'train/grad_norm_pre_clip': 8.778151512145996, 'timing/data': 0.02508203499019146, 'timing/model': 2.7546137012541294, 'learning_rate/qwen_vl_interface': 9.80626339627781e-06, 'learning_rate/action_model': 9.806263396277809e-05, 'epoch': 1.12}) +[2026-06-19 18:31:59,922][starVLA.training.train_starvla][INFO] - Step 548, Loss: {'train/loss': 0.3191976845264435, 'train/grad_norm_pre_clip': 4.602490425109863, 'timing/data': 0.023094942793250084, 'timing/model': 2.7558237351477146, 'learning_rate/qwen_vl_interface': 9.805401570595087e-06, 'learning_rate/action_model': 9.805401570595087e-05, 'epoch': 1.13}) +[2026-06-19 18:32:10,922][starVLA.training.train_starvla][INFO] - Step 549, Loss: {'train/loss': 0.12694385647773743, 'train/grad_norm_pre_clip': 3.4885053634643555, 'timing/data': 0.022997034713625908, 'timing/model': 2.758256694301963, 'learning_rate/qwen_vl_interface': 9.804537872358853e-06, 'learning_rate/action_model': 9.804537872358853e-05, 'epoch': 1.13}) +[2026-06-19 18:32:21,724][starVLA.training.train_starvla][INFO] - Step 550, Loss: {'train/loss': 0.11318317800760269, 'train/grad_norm_pre_clip': 3.193547010421753, 'timing/data': 0.022263862192630768, 'timing/model': 2.7609872967004776, 'learning_rate/qwen_vl_interface': 9.80367230192414e-06, 'learning_rate/action_model': 9.803672301924138e-05, 'epoch': 1.13}) +[2026-06-19 18:32:32,521][starVLA.training.train_starvla][INFO] - Step 551, Loss: {'train/loss': 0.20883691310882568, 'train/grad_norm_pre_clip': 8.224003791809082, 'timing/data': 0.02255120873451233, 'timing/model': 2.755954135209322, 'learning_rate/qwen_vl_interface': 9.80280485964675e-06, 'learning_rate/action_model': 9.80280485964675e-05, 'epoch': 1.13}) +[2026-06-19 18:32:43,357][starVLA.training.train_starvla][INFO] - Step 552, Loss: {'train/loss': 0.22747869789600372, 'train/grad_norm_pre_clip': 6.029112339019775, 'timing/data': 0.022468727082014084, 'timing/model': 2.8030209708958864, 'learning_rate/qwen_vl_interface': 9.801935545883255e-06, 'learning_rate/action_model': 9.801935545883255e-05, 'epoch': 1.13}) +[2026-06-19 18:32:54,108][starVLA.training.train_starvla][INFO] - Step 553, Loss: {'train/loss': 0.14178603887557983, 'train/grad_norm_pre_clip': 4.860816955566406, 'timing/data': 0.02256595529615879, 'timing/model': 2.746867708861828, 'learning_rate/qwen_vl_interface': 9.801064360991e-06, 'learning_rate/action_model': 9.801064360991e-05, 'epoch': 1.14}) +[2026-06-19 18:33:04,813][starVLA.training.train_starvla][INFO] - Step 554, Loss: {'train/loss': 0.22741791605949402, 'train/grad_norm_pre_clip': 7.074727535247803, 'timing/data': 0.023552782833576202, 'timing/model': 2.7492250222712755, 'learning_rate/qwen_vl_interface': 9.800191305328095e-06, 'learning_rate/action_model': 9.800191305328094e-05, 'epoch': 1.14}) +[2026-06-19 18:33:15,514][starVLA.training.train_starvla][INFO] - Step 555, Loss: {'train/loss': 0.11150728911161423, 'train/grad_norm_pre_clip': 5.308433532714844, 'timing/data': 0.025021200999617577, 'timing/model': 2.751987362280488, 'learning_rate/qwen_vl_interface': 9.79931637925342e-06, 'learning_rate/action_model': 9.79931637925342e-05, 'epoch': 1.14}) +[2026-06-19 18:33:26,319][starVLA.training.train_starvla][INFO] - Step 556, Loss: {'train/loss': 0.18876872956752777, 'train/grad_norm_pre_clip': 6.076939105987549, 'timing/data': 0.023082615807652473, 'timing/model': 2.7611463237553835, 'learning_rate/qwen_vl_interface': 9.798439583126625e-06, 'learning_rate/action_model': 9.798439583126625e-05, 'epoch': 1.14}) +[2026-06-19 18:33:37,021][starVLA.training.train_starvla][INFO] - Step 557, Loss: {'train/loss': 0.1862991750240326, 'train/grad_norm_pre_clip': 6.96172571182251, 'timing/data': 0.02234194613993168, 'timing/model': 2.770391035825014, 'learning_rate/qwen_vl_interface': 9.797560917308126e-06, 'learning_rate/action_model': 9.797560917308126e-05, 'epoch': 1.14}) +[2026-06-19 18:33:47,814][starVLA.training.train_starvla][INFO] - Step 558, Loss: {'train/loss': 0.24818813800811768, 'train/grad_norm_pre_clip': 7.743868827819824, 'timing/data': 0.02290172688663006, 'timing/model': 2.759137876331806, 'learning_rate/qwen_vl_interface': 9.796680382159112e-06, 'learning_rate/action_model': 9.79668038215911e-05, 'epoch': 1.15}) +[2026-06-19 18:33:58,626][starVLA.training.train_starvla][INFO] - Step 559, Loss: {'train/loss': 0.2709193825721741, 'train/grad_norm_pre_clip': 10.302102088928223, 'timing/data': 0.02479858137667179, 'timing/model': 2.7760881558060646, 'learning_rate/qwen_vl_interface': 9.795797978041533e-06, 'learning_rate/action_model': 9.795797978041533e-05, 'epoch': 1.15}) +[2026-06-19 18:34:09,611][starVLA.training.train_starvla][INFO] - Step 560, Loss: {'train/loss': 0.34042513370513916, 'train/grad_norm_pre_clip': 21.37569236755371, 'timing/data': 0.022703679278492928, 'timing/model': 2.716972693800926, 'learning_rate/qwen_vl_interface': 9.794913705318117e-06, 'learning_rate/action_model': 9.794913705318118e-05, 'epoch': 1.15}) +[2026-06-19 18:34:20,223][starVLA.training.train_starvla][INFO] - Step 561, Loss: {'train/loss': 0.22460836172103882, 'train/grad_norm_pre_clip': 11.092327117919922, 'timing/data': 0.022458352148532867, 'timing/model': 2.750696474686265, 'learning_rate/qwen_vl_interface': 9.794027564352352e-06, 'learning_rate/action_model': 9.794027564352352e-05, 'epoch': 1.15}) +[2026-06-19 18:34:30,914][starVLA.training.train_starvla][INFO] - Step 562, Loss: {'train/loss': 0.15077197551727295, 'train/grad_norm_pre_clip': 6.395081043243408, 'timing/data': 0.02297680638730526, 'timing/model': 2.7337195612490177, 'learning_rate/qwen_vl_interface': 9.793139555508501e-06, 'learning_rate/action_model': 9.7931395555085e-05, 'epoch': 1.15}) +[2026-06-19 18:34:41,620][starVLA.training.train_starvla][INFO] - Step 563, Loss: {'train/loss': 0.1567850112915039, 'train/grad_norm_pre_clip': 6.8709001541137695, 'timing/data': 0.022507639601826668, 'timing/model': 2.746469035744667, 'learning_rate/qwen_vl_interface': 9.792249679151585e-06, 'learning_rate/action_model': 9.792249679151585e-05, 'epoch': 1.16}) +[2026-06-19 18:34:52,495][starVLA.training.train_starvla][INFO] - Step 564, Loss: {'train/loss': 0.28047215938568115, 'train/grad_norm_pre_clip': 6.7852277755737305, 'timing/data': 0.022502567619085312, 'timing/model': 2.7397600132972, 'learning_rate/qwen_vl_interface': 9.791357935647403e-06, 'learning_rate/action_model': 9.791357935647403e-05, 'epoch': 1.16}) +[2026-06-19 18:35:03,317][starVLA.training.train_starvla][INFO] - Step 565, Loss: {'train/loss': 0.3006685674190521, 'train/grad_norm_pre_clip': 11.394636154174805, 'timing/data': 0.02281917631626129, 'timing/model': 2.7583206836134195, 'learning_rate/qwen_vl_interface': 9.790464325362516e-06, 'learning_rate/action_model': 9.790464325362516e-05, 'epoch': 1.16}) +[2026-06-19 18:35:14,023][starVLA.training.train_starvla][INFO] - Step 566, Loss: {'train/loss': 0.18372708559036255, 'train/grad_norm_pre_clip': 6.3827033042907715, 'timing/data': 0.022171441465616226, 'timing/model': 2.717533690854907, 'learning_rate/qwen_vl_interface': 9.789568848664254e-06, 'learning_rate/action_model': 9.789568848664253e-05, 'epoch': 1.16}) +[2026-06-19 18:35:24,826][starVLA.training.train_starvla][INFO] - Step 567, Loss: {'train/loss': 0.32837340235710144, 'train/grad_norm_pre_clip': 7.21914005279541, 'timing/data': 0.022557277232408524, 'timing/model': 2.7626999504864216, 'learning_rate/qwen_vl_interface': 9.788671505920712e-06, 'learning_rate/action_model': 9.788671505920711e-05, 'epoch': 1.16}) +[2026-06-19 18:35:35,629][starVLA.training.train_starvla][INFO] - Step 568, Loss: {'train/loss': 0.15634281933307648, 'train/grad_norm_pre_clip': 8.907649993896484, 'timing/data': 0.022390779107809067, 'timing/model': 2.764729280024767, 'learning_rate/qwen_vl_interface': 9.787772297500754e-06, 'learning_rate/action_model': 9.787772297500754e-05, 'epoch': 1.17}) +[2026-06-19 18:35:46,516][starVLA.training.train_starvla][INFO] - Step 569, Loss: {'train/loss': 0.23772931098937988, 'train/grad_norm_pre_clip': 4.929779529571533, 'timing/data': 0.022358177229762077, 'timing/model': 2.7539956122636795, 'learning_rate/qwen_vl_interface': 9.78687122377401e-06, 'learning_rate/action_model': 9.78687122377401e-05, 'epoch': 1.17}) +[2026-06-19 18:35:57,318][starVLA.training.train_starvla][INFO] - Step 570, Loss: {'train/loss': 0.14921914041042328, 'train/grad_norm_pre_clip': 4.151749610900879, 'timing/data': 0.021974531933665276, 'timing/model': 2.7659574542194605, 'learning_rate/qwen_vl_interface': 9.78596828511088e-06, 'learning_rate/action_model': 9.78596828511088e-05, 'epoch': 1.17}) +[2026-06-19 18:36:08,013][starVLA.training.train_starvla][INFO] - Step 571, Loss: {'train/loss': 0.2015024721622467, 'train/grad_norm_pre_clip': 3.8432250022888184, 'timing/data': 0.02243584580719471, 'timing/model': 2.732551896944642, 'learning_rate/qwen_vl_interface': 9.785063481882527e-06, 'learning_rate/action_model': 9.785063481882527e-05, 'epoch': 1.17}) +[2026-06-19 18:36:18,822][starVLA.training.train_starvla][INFO] - Step 572, Loss: {'train/loss': 0.14011284708976746, 'train/grad_norm_pre_clip': 4.724226474761963, 'timing/data': 0.022824201732873917, 'timing/model': 2.7617385778576136, 'learning_rate/qwen_vl_interface': 9.78415681446088e-06, 'learning_rate/action_model': 9.784156814460879e-05, 'epoch': 1.17}) +[2026-06-19 18:36:29,619][starVLA.training.train_starvla][INFO] - Step 573, Loss: {'train/loss': 0.15422382950782776, 'train/grad_norm_pre_clip': 7.214071750640869, 'timing/data': 0.020354807376861572, 'timing/model': 2.767392935231328, 'learning_rate/qwen_vl_interface': 9.783248283218638e-06, 'learning_rate/action_model': 9.783248283218638e-05, 'epoch': 1.18}) +[2026-06-19 18:36:40,422][starVLA.training.train_starvla][INFO] - Step 574, Loss: {'train/loss': 0.2602923512458801, 'train/grad_norm_pre_clip': 11.402376174926758, 'timing/data': 0.02346813678741455, 'timing/model': 2.769837884232402, 'learning_rate/qwen_vl_interface': 9.782337888529261e-06, 'learning_rate/action_model': 9.782337888529261e-05, 'epoch': 1.18}) +[2026-06-19 18:36:51,436][starVLA.training.train_starvla][INFO] - Step 575, Loss: {'train/loss': 0.1900513768196106, 'train/grad_norm_pre_clip': 10.244409561157227, 'timing/data': 0.022903909906744957, 'timing/model': 2.976129710674286, 'learning_rate/qwen_vl_interface': 9.781425630766978e-06, 'learning_rate/action_model': 9.781425630766979e-05, 'epoch': 1.18}) +[2026-06-19 18:37:02,215][starVLA.training.train_starvla][INFO] - Step 576, Loss: {'train/loss': 0.6403857469558716, 'train/grad_norm_pre_clip': 27.779741287231445, 'timing/data': 0.02321251854300499, 'timing/model': 2.766953656449914, 'learning_rate/qwen_vl_interface': 9.780511510306789e-06, 'learning_rate/action_model': 9.780511510306788e-05, 'epoch': 1.18}) +[2026-06-19 18:37:12,912][starVLA.training.train_starvla][INFO] - Step 577, Loss: {'train/loss': 0.3520515263080597, 'train/grad_norm_pre_clip': 15.772262573242188, 'timing/data': 0.022627783939242363, 'timing/model': 2.7489196956157684, 'learning_rate/qwen_vl_interface': 9.779595527524447e-06, 'learning_rate/action_model': 9.779595527524447e-05, 'epoch': 1.19}) +[2026-06-19 18:37:23,638][starVLA.training.train_starvla][INFO] - Step 578, Loss: {'train/loss': 0.19932617247104645, 'train/grad_norm_pre_clip': 7.742005348205566, 'timing/data': 0.022811001166701317, 'timing/model': 2.7108913380652666, 'learning_rate/qwen_vl_interface': 9.778677682796483e-06, 'learning_rate/action_model': 9.778677682796483e-05, 'epoch': 1.19}) +[2026-06-19 18:37:34,436][starVLA.training.train_starvla][INFO] - Step 579, Loss: {'train/loss': 0.160657599568367, 'train/grad_norm_pre_clip': 22.86210060119629, 'timing/data': 0.022941987961530685, 'timing/model': 2.7111655194312334, 'learning_rate/qwen_vl_interface': 9.777757976500185e-06, 'learning_rate/action_model': 9.777757976500186e-05, 'epoch': 1.19}) +[2026-06-19 18:37:45,323][starVLA.training.train_starvla][INFO] - Step 580, Loss: {'train/loss': 0.11063850671052933, 'train/grad_norm_pre_clip': 5.867618083953857, 'timing/data': 0.0226057730615139, 'timing/model': 2.7585145197808743, 'learning_rate/qwen_vl_interface': 9.776836409013614e-06, 'learning_rate/action_model': 9.776836409013615e-05, 'epoch': 1.19}) +[2026-06-19 18:37:56,014][starVLA.training.train_starvla][INFO] - Step 581, Loss: {'train/loss': 0.11073929071426392, 'train/grad_norm_pre_clip': 4.197635173797607, 'timing/data': 0.02238532342016697, 'timing/model': 2.7565426621586084, 'learning_rate/qwen_vl_interface': 9.775912980715589e-06, 'learning_rate/action_model': 9.775912980715588e-05, 'epoch': 1.19}) +[2026-06-19 18:38:06,825][starVLA.training.train_starvla][INFO] - Step 582, Loss: {'train/loss': 0.11391517519950867, 'train/grad_norm_pre_clip': 6.940207481384277, 'timing/data': 0.022916706278920174, 'timing/model': 2.7704577557742596, 'learning_rate/qwen_vl_interface': 9.774987691985694e-06, 'learning_rate/action_model': 9.774987691985694e-05, 'epoch': 1.2}) +[2026-06-19 18:38:17,620][starVLA.training.train_starvla][INFO] - Step 583, Loss: {'train/loss': 0.3214358389377594, 'train/grad_norm_pre_clip': 8.325640678405762, 'timing/data': 0.02162942849099636, 'timing/model': 2.7519032955169678, 'learning_rate/qwen_vl_interface': 9.774060543204283e-06, 'learning_rate/action_model': 9.774060543204283e-05, 'epoch': 1.2}) +[2026-06-19 18:38:28,295][starVLA.training.train_starvla][INFO] - Step 584, Loss: {'train/loss': 0.27724504470825195, 'train/grad_norm_pre_clip': 8.015800476074219, 'timing/data': 0.02213551104068756, 'timing/model': 2.7003358993679285, 'learning_rate/qwen_vl_interface': 9.773131534752472e-06, 'learning_rate/action_model': 9.773131534752472e-05, 'epoch': 1.2}) +[2026-06-19 18:38:38,917][starVLA.training.train_starvla][INFO] - Step 585, Loss: {'train/loss': 0.37864720821380615, 'train/grad_norm_pre_clip': 10.7173433303833, 'timing/data': 0.023289792239665985, 'timing/model': 2.7400084398686886, 'learning_rate/qwen_vl_interface': 9.772200667012142e-06, 'learning_rate/action_model': 9.772200667012142e-05, 'epoch': 1.2}) +[2026-06-19 18:38:49,676][starVLA.training.train_starvla][INFO] - Step 586, Loss: {'train/loss': 0.13835908472537994, 'train/grad_norm_pre_clip': 4.095727443695068, 'timing/data': 0.023649636656045914, 'timing/model': 2.7269520219415426, 'learning_rate/qwen_vl_interface': 9.771267940365934e-06, 'learning_rate/action_model': 9.771267940365935e-05, 'epoch': 1.2}) +[2026-06-19 18:39:00,282][starVLA.training.train_starvla][INFO] - Step 587, Loss: {'train/loss': 0.4345376491546631, 'train/grad_norm_pre_clip': 9.805890083312988, 'timing/data': 0.02297322452068329, 'timing/model': 2.675237601622939, 'learning_rate/qwen_vl_interface': 9.770333355197262e-06, 'learning_rate/action_model': 9.77033335519726e-05, 'epoch': 1.21}) +[2026-06-19 18:39:11,020][starVLA.training.train_starvla][INFO] - Step 588, Loss: {'train/loss': 0.348875492811203, 'train/grad_norm_pre_clip': 6.988464832305908, 'timing/data': 0.022651419043540955, 'timing/model': 2.767089158296585, 'learning_rate/qwen_vl_interface': 9.769396911890291e-06, 'learning_rate/action_model': 9.769396911890291e-05, 'epoch': 1.21}) +[2026-06-19 18:39:21,657][starVLA.training.train_starvla][INFO] - Step 589, Loss: {'train/loss': 0.22592943906784058, 'train/grad_norm_pre_clip': 11.657254219055176, 'timing/data': 0.022640898823738098, 'timing/model': 2.6838019639253616, 'learning_rate/qwen_vl_interface': 9.768458610829965e-06, 'learning_rate/action_model': 9.768458610829965e-05, 'epoch': 1.21}) +[2026-06-19 18:39:32,319][starVLA.training.train_starvla][INFO] - Step 590, Loss: {'train/loss': 0.3710549473762512, 'train/grad_norm_pre_clip': 11.269479751586914, 'timing/data': 0.022608527913689613, 'timing/model': 2.723831342533231, 'learning_rate/qwen_vl_interface': 9.767518452401981e-06, 'learning_rate/action_model': 9.76751845240198e-05, 'epoch': 1.21}) +[2026-06-19 18:39:43,118][starVLA.training.train_starvla][INFO] - Step 591, Loss: {'train/loss': 0.26011592149734497, 'train/grad_norm_pre_clip': 69.31007385253906, 'timing/data': 0.022886432707309723, 'timing/model': 2.7564159464091063, 'learning_rate/qwen_vl_interface': 9.766576436992802e-06, 'learning_rate/action_model': 9.766576436992801e-05, 'epoch': 1.21}) +[2026-06-19 18:39:53,909][starVLA.training.train_starvla][INFO] - Step 592, Loss: {'train/loss': 0.28383588790893555, 'train/grad_norm_pre_clip': 4.814090251922607, 'timing/data': 0.02301786094903946, 'timing/model': 2.7442942168563604, 'learning_rate/qwen_vl_interface': 9.765632564989657e-06, 'learning_rate/action_model': 9.765632564989656e-05, 'epoch': 1.22}) +[2026-06-19 18:40:04,724][starVLA.training.train_starvla][INFO] - Step 593, Loss: {'train/loss': 0.2851371467113495, 'train/grad_norm_pre_clip': 7.473323822021484, 'timing/data': 0.0226603876799345, 'timing/model': 2.760679639875889, 'learning_rate/qwen_vl_interface': 9.764686836780533e-06, 'learning_rate/action_model': 9.764686836780532e-05, 'epoch': 1.22}) +[2026-06-19 18:40:15,521][starVLA.training.train_starvla][INFO] - Step 594, Loss: {'train/loss': 0.3975091874599457, 'train/grad_norm_pre_clip': 4.572579860687256, 'timing/data': 0.021788114681839943, 'timing/model': 2.7631042785942554, 'learning_rate/qwen_vl_interface': 9.763739252754187e-06, 'learning_rate/action_model': 9.763739252754186e-05, 'epoch': 1.22}) +[2026-06-19 18:40:26,522][starVLA.training.train_starvla][INFO] - Step 595, Loss: {'train/loss': 0.22272059321403503, 'train/grad_norm_pre_clip': 3.4903903007507324, 'timing/data': 0.02174973674118519, 'timing/model': 2.7628598641604185, 'learning_rate/qwen_vl_interface': 9.76278981330013e-06, 'learning_rate/action_model': 9.762789813300131e-05, 'epoch': 1.22}) +[2026-06-19 18:40:37,325][starVLA.training.train_starvla][INFO] - Step 596, Loss: {'train/loss': 0.29752659797668457, 'train/grad_norm_pre_clip': 6.410109043121338, 'timing/data': 0.02252080664038658, 'timing/model': 2.763255435973406, 'learning_rate/qwen_vl_interface': 9.761838518808647e-06, 'learning_rate/action_model': 9.761838518808646e-05, 'epoch': 1.22}) +[2026-06-19 18:40:48,122][starVLA.training.train_starvla][INFO] - Step 597, Loss: {'train/loss': 0.18193179368972778, 'train/grad_norm_pre_clip': 5.363231658935547, 'timing/data': 0.02231418713927269, 'timing/model': 2.75930480286479, 'learning_rate/qwen_vl_interface': 9.760885369670776e-06, 'learning_rate/action_model': 9.760885369670776e-05, 'epoch': 1.23}) +[2026-06-19 18:40:58,983][starVLA.training.train_starvla][INFO] - Step 598, Loss: {'train/loss': 0.20752288401126862, 'train/grad_norm_pre_clip': 4.857449531555176, 'timing/data': 0.021891653537750244, 'timing/model': 2.8460621666163206, 'learning_rate/qwen_vl_interface': 9.759930366278319e-06, 'learning_rate/action_model': 9.759930366278318e-05, 'epoch': 1.23}) +[2026-06-19 18:41:09,614][starVLA.training.train_starvla][INFO] - Step 599, Loss: {'train/loss': 0.3124578893184662, 'train/grad_norm_pre_clip': 7.377992153167725, 'timing/data': 0.02246265485882759, 'timing/model': 2.736379185691476, 'learning_rate/qwen_vl_interface': 9.758973509023842e-06, 'learning_rate/action_model': 9.758973509023843e-05, 'epoch': 1.23}) +[2026-06-19 18:41:20,419][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 600 +[2026-06-19 18:48:48,101][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 600 +[2026-06-19 18:48:48,102][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 600 +[2026-06-19 18:56:52,218][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 600 +[2026-06-19 18:56:52,220][starVLA.training.train_starvla][INFO] - Step 600, Loss: {'train/loss': 0.2167454957962036, 'train/grad_norm_pre_clip': 5.979025840759277, 'eval/loss': 0.353803813457489, 'eval/latency_0/loss': 0.10812118649482727, 'eval/latency_1/loss': 0.2261960357427597, 'eval/latency_2/loss': 0.46917322278022766, 'eval/latency_3/loss': 0.39634236693382263, 'eval/latency_4/loss': 0.551274836063385, 'eval/flappy/loss': 0.353803813457489, 'eval/flappy/latency_0/loss': 0.10812118649482727, 'eval/flappy/latency_1/loss': 0.2261960357427597, 'eval/flappy/latency_2/loss': 0.46917322278022766, 'eval/flappy/latency_3/loss': 0.39634236693382263, 'eval/flappy/latency_4/loss': 0.551274836063385, 'eval/flap_precision': 0.7619047619047619, 'eval/flap_recall': 0.64, 'eval/flap_f1': 0.6956521739130435, 'eval/flap_rate_model': 0.06416666666666666, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.6956521739130435, 'eval/latency_0/flap_precision': 0.7619047619047619, 'eval/latency_0/flap_recall': 0.64, 'eval/latency_0/flap_f1': 0.6956521739130435, 'eval/latency_0/flap_rate_model': 0.06416666666666666, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.6956521739130435, 'timing/data': 0.022390248253941536, 'timing/model': 2.7536254730075598, 'learning_rate/qwen_vl_interface': 9.758014798300677e-06, 'learning_rate/action_model': 9.758014798300676e-05, 'epoch': 1.23}) +[2026-06-19 18:56:52,221][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_600_state +[2026-06-19 18:56:52,242][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-19 18:56:57,704][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_600_state/model.safetensors +[2026-06-19 18:57:04,918][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_600_state/optimizer.bin +[2026-06-19 18:57:04,918][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_600_state/sampler.bin +[2026-06-19 18:57:04,918][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_600_state/sampler_1.bin +[2026-06-19 18:57:04,919][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_600_state/random_states_0.pkl +[2026-06-19 18:57:05,147][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 600 +[2026-06-19 18:57:05,148][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-19 18:57:05,161][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-19 18:57:15,908][starVLA.training.train_starvla][INFO] - Step 601, Loss: {'train/loss': 0.1528681516647339, 'train/grad_norm_pre_clip': 6.834046840667725, 'timing/data': 0.024920348078012466, 'timing/model': 2.7402530293911695, 'learning_rate/qwen_vl_interface': 9.757054234502909e-06, 'learning_rate/action_model': 9.757054234502909e-05, 'epoch': 1.23}) +[2026-06-19 18:57:26,609][starVLA.training.train_starvla][INFO] - Step 602, Loss: {'train/loss': 0.3367694616317749, 'train/grad_norm_pre_clip': 4.844418525695801, 'timing/data': 0.023871861398220062, 'timing/model': 2.765289794653654, 'learning_rate/qwen_vl_interface': 9.756091818025394e-06, 'learning_rate/action_model': 9.756091818025394e-05, 'epoch': 1.24}) +[2026-06-19 18:57:37,255][starVLA.training.train_starvla][INFO] - Step 603, Loss: {'train/loss': 0.2489713430404663, 'train/grad_norm_pre_clip': 11.106623649597168, 'timing/data': 0.02351902425289154, 'timing/model': 2.7123716063797474, 'learning_rate/qwen_vl_interface': 9.75512754926374e-06, 'learning_rate/action_model': 9.755127549263741e-05, 'epoch': 1.24}) +[2026-06-19 18:57:47,982][starVLA.training.train_starvla][INFO] - Step 604, Loss: {'train/loss': 0.08962303400039673, 'train/grad_norm_pre_clip': 5.000922679901123, 'timing/data': 0.024510834366083145, 'timing/model': 2.732318803668022, 'learning_rate/qwen_vl_interface': 9.754161428614325e-06, 'learning_rate/action_model': 9.754161428614324e-05, 'epoch': 1.24}) +[2026-06-19 18:57:58,707][starVLA.training.train_starvla][INFO] - Step 605, Loss: {'train/loss': 0.1335173398256302, 'train/grad_norm_pre_clip': 7.333714008331299, 'timing/data': 0.023359283804893494, 'timing/model': 2.7564951945096254, 'learning_rate/qwen_vl_interface': 9.753193456474284e-06, 'learning_rate/action_model': 9.753193456474284e-05, 'epoch': 1.24}) +[2026-06-19 18:58:09,568][starVLA.training.train_starvla][INFO] - Step 606, Loss: {'train/loss': 0.07149512320756912, 'train/grad_norm_pre_clip': 6.531147003173828, 'timing/data': 0.024194005876779556, 'timing/model': 3.006785374134779, 'learning_rate/qwen_vl_interface': 9.752223633241514e-06, 'learning_rate/action_model': 9.752223633241514e-05, 'epoch': 1.24}) +[2026-06-19 18:58:20,318][starVLA.training.train_starvla][INFO] - Step 607, Loss: {'train/loss': 0.039340790361166, 'train/grad_norm_pre_clip': 3.3989956378936768, 'timing/data': 0.031504638493061066, 'timing/model': 2.736892979592085, 'learning_rate/qwen_vl_interface': 9.751251959314671e-06, 'learning_rate/action_model': 9.751251959314672e-05, 'epoch': 1.25}) +[2026-06-19 18:58:31,253][starVLA.training.train_starvla][INFO] - Step 608, Loss: {'train/loss': 0.27047374844551086, 'train/grad_norm_pre_clip': 10.314629554748535, 'timing/data': 0.024739069864153862, 'timing/model': 2.761436177417636, 'learning_rate/qwen_vl_interface': 9.750278435093176e-06, 'learning_rate/action_model': 9.750278435093177e-05, 'epoch': 1.25}) +[2026-06-19 18:58:42,111][starVLA.training.train_starvla][INFO] - Step 609, Loss: {'train/loss': 0.18927033245563507, 'train/grad_norm_pre_clip': 10.192915916442871, 'timing/data': 0.03508126735687256, 'timing/model': 2.745258539915085, 'learning_rate/qwen_vl_interface': 9.749303060977205e-06, 'learning_rate/action_model': 9.749303060977205e-05, 'epoch': 1.25}) +[2026-06-19 18:58:53,119][starVLA.training.train_starvla][INFO] - Step 610, Loss: {'train/loss': 0.5789158344268799, 'train/grad_norm_pre_clip': 14.555747032165527, 'timing/data': 0.022856542840600014, 'timing/model': 2.7374199628829956, 'learning_rate/qwen_vl_interface': 9.748325837367701e-06, 'learning_rate/action_model': 9.748325837367702e-05, 'epoch': 1.25}) +[2026-06-19 18:59:03,911][starVLA.training.train_starvla][INFO] - Step 611, Loss: {'train/loss': 0.27229130268096924, 'train/grad_norm_pre_clip': 12.98223876953125, 'timing/data': 0.023381514474749565, 'timing/model': 2.7507935017347336, 'learning_rate/qwen_vl_interface': 9.747346764666362e-06, 'learning_rate/action_model': 9.747346764666361e-05, 'epoch': 1.25}) +[2026-06-19 18:59:14,716][starVLA.training.train_starvla][INFO] - Step 612, Loss: {'train/loss': 0.2933291494846344, 'train/grad_norm_pre_clip': 19.49646759033203, 'timing/data': 0.025291673839092255, 'timing/model': 2.7539442740380764, 'learning_rate/qwen_vl_interface': 9.746365843275648e-06, 'learning_rate/action_model': 9.746365843275646e-05, 'epoch': 1.26}) +[2026-06-19 18:59:25,520][starVLA.training.train_starvla][INFO] - Step 613, Loss: {'train/loss': 0.2394610494375229, 'train/grad_norm_pre_clip': 16.20095443725586, 'timing/data': 0.02339145354926586, 'timing/model': 2.7607923336327076, 'learning_rate/qwen_vl_interface': 9.745383073598776e-06, 'learning_rate/action_model': 9.745383073598775e-05, 'epoch': 1.26}) +[2026-06-19 18:59:36,408][starVLA.training.train_starvla][INFO] - Step 614, Loss: {'train/loss': 0.20422013103961945, 'train/grad_norm_pre_clip': 12.951881408691406, 'timing/data': 0.023103900253772736, 'timing/model': 2.762902496382594, 'learning_rate/qwen_vl_interface': 9.744398456039729e-06, 'learning_rate/action_model': 9.744398456039729e-05, 'epoch': 1.26}) +[2026-06-19 18:59:47,123][starVLA.training.train_starvla][INFO] - Step 615, Loss: {'train/loss': 0.3696673512458801, 'train/grad_norm_pre_clip': 18.702136993408203, 'timing/data': 0.023050939664244652, 'timing/model': 2.760583830997348, 'learning_rate/qwen_vl_interface': 9.743411991003246e-06, 'learning_rate/action_model': 9.743411991003247e-05, 'epoch': 1.26}) +[2026-06-19 18:59:57,920][starVLA.training.train_starvla][INFO] - Step 616, Loss: {'train/loss': 0.19222389161586761, 'train/grad_norm_pre_clip': 7.821043491363525, 'timing/data': 0.02333650551736355, 'timing/model': 2.7648476511240005, 'learning_rate/qwen_vl_interface': 9.742423678894823e-06, 'learning_rate/action_model': 9.742423678894823e-05, 'epoch': 1.27}) +[2026-06-19 19:00:08,723][starVLA.training.train_starvla][INFO] - Step 617, Loss: {'train/loss': 0.346493124961853, 'train/grad_norm_pre_clip': 96.19966125488281, 'timing/data': 0.023154817521572113, 'timing/model': 2.784108394756913, 'learning_rate/qwen_vl_interface': 9.741433520120722e-06, 'learning_rate/action_model': 9.74143352012072e-05, 'epoch': 1.27}) +[2026-06-19 19:00:19,511][starVLA.training.train_starvla][INFO] - Step 618, Loss: {'train/loss': 0.2416522204875946, 'train/grad_norm_pre_clip': 9.54736614227295, 'timing/data': 0.023225588724017143, 'timing/model': 2.751596327871084, 'learning_rate/qwen_vl_interface': 9.740441515087953e-06, 'learning_rate/action_model': 9.740441515087953e-05, 'epoch': 1.27}) +[2026-06-19 19:00:30,308][starVLA.training.train_starvla][INFO] - Step 619, Loss: {'train/loss': 0.2622641623020172, 'train/grad_norm_pre_clip': 11.489924430847168, 'timing/data': 0.0233104657381773, 'timing/model': 2.75495582818985, 'learning_rate/qwen_vl_interface': 9.739447664204298e-06, 'learning_rate/action_model': 9.739447664204297e-05, 'epoch': 1.27}) +[2026-06-19 19:00:41,019][starVLA.training.train_starvla][INFO] - Step 620, Loss: {'train/loss': 0.17680101096630096, 'train/grad_norm_pre_clip': 9.40743350982666, 'timing/data': 0.023152967914938927, 'timing/model': 2.773841256275773, 'learning_rate/qwen_vl_interface': 9.73845196787829e-06, 'learning_rate/action_model': 9.738451967878289e-05, 'epoch': 1.27}) +[2026-06-19 19:00:51,821][starVLA.training.train_starvla][INFO] - Step 621, Loss: {'train/loss': 0.2094082236289978, 'train/grad_norm_pre_clip': 8.27950668334961, 'timing/data': 0.023540381342172623, 'timing/model': 2.773902002722025, 'learning_rate/qwen_vl_interface': 9.737454426519221e-06, 'learning_rate/action_model': 9.737454426519221e-05, 'epoch': 1.28}) +[2026-06-19 19:01:02,621][starVLA.training.train_starvla][INFO] - Step 622, Loss: {'train/loss': 0.25710535049438477, 'train/grad_norm_pre_clip': 5.772334098815918, 'timing/data': 0.022871199995279312, 'timing/model': 2.755657698959112, 'learning_rate/qwen_vl_interface': 9.736455040537145e-06, 'learning_rate/action_model': 9.736455040537144e-05, 'epoch': 1.28}) +[2026-06-19 19:01:13,320][starVLA.training.train_starvla][INFO] - Step 623, Loss: {'train/loss': 0.24674181640148163, 'train/grad_norm_pre_clip': 4.437570095062256, 'timing/data': 0.02360789105296135, 'timing/model': 2.7270277980715036, 'learning_rate/qwen_vl_interface': 9.735453810342868e-06, 'learning_rate/action_model': 9.735453810342867e-05, 'epoch': 1.28}) +[2026-06-19 19:01:24,121][starVLA.training.train_starvla][INFO] - Step 624, Loss: {'train/loss': 0.1992778182029724, 'train/grad_norm_pre_clip': 3.788081407546997, 'timing/data': 0.023007988929748535, 'timing/model': 2.763454919680953, 'learning_rate/qwen_vl_interface': 9.734450736347958e-06, 'learning_rate/action_model': 9.73445073634796e-05, 'epoch': 1.28}) +[2026-06-19 19:01:34,918][starVLA.training.train_starvla][INFO] - Step 625, Loss: {'train/loss': 0.11456457525491714, 'train/grad_norm_pre_clip': 2.978931188583374, 'timing/data': 0.023770421743392944, 'timing/model': 2.7661187555640936, 'learning_rate/qwen_vl_interface': 9.733445818964746e-06, 'learning_rate/action_model': 9.733445818964746e-05, 'epoch': 1.28}) +[2026-06-19 19:01:45,917][starVLA.training.train_starvla][INFO] - Step 626, Loss: {'train/loss': 0.11937594413757324, 'train/grad_norm_pre_clip': 5.4280219078063965, 'timing/data': 0.0231120977550745, 'timing/model': 2.7527493871748447, 'learning_rate/qwen_vl_interface': 9.732439058606311e-06, 'learning_rate/action_model': 9.732439058606312e-05, 'epoch': 1.29}) +[2026-06-19 19:01:56,722][starVLA.training.train_starvla][INFO] - Step 627, Loss: {'train/loss': 0.30690813064575195, 'train/grad_norm_pre_clip': 12.222121238708496, 'timing/data': 0.023533031344413757, 'timing/model': 2.7709788270294666, 'learning_rate/qwen_vl_interface': 9.731430455686499e-06, 'learning_rate/action_model': 9.731430455686498e-05, 'epoch': 1.29}) +[2026-06-19 19:02:07,420][starVLA.training.train_starvla][INFO] - Step 628, Loss: {'train/loss': 0.14338867366313934, 'train/grad_norm_pre_clip': 3.2938313484191895, 'timing/data': 0.02323664352297783, 'timing/model': 2.7414021361619234, 'learning_rate/qwen_vl_interface': 9.730420010619901e-06, 'learning_rate/action_model': 9.7304200106199e-05, 'epoch': 1.29}) +[2026-06-19 19:02:18,282][starVLA.training.train_starvla][INFO] - Step 629, Loss: {'train/loss': 0.34616774320602417, 'train/grad_norm_pre_clip': 11.476892471313477, 'timing/data': 0.0234097708016634, 'timing/model': 2.8214501813054085, 'learning_rate/qwen_vl_interface': 9.72940772382188e-06, 'learning_rate/action_model': 9.729407723821878e-05, 'epoch': 1.29}) +[2026-06-19 19:02:29,116][starVLA.training.train_starvla][INFO] - Step 630, Loss: {'train/loss': 0.16976255178451538, 'train/grad_norm_pre_clip': 7.180722236633301, 'timing/data': 0.02341473661363125, 'timing/model': 2.7585377749055624, 'learning_rate/qwen_vl_interface': 9.728393595708542e-06, 'learning_rate/action_model': 9.728393595708542e-05, 'epoch': 1.29}) +[2026-06-19 19:02:39,925][starVLA.training.train_starvla][INFO] - Step 631, Loss: {'train/loss': 0.3215578496456146, 'train/grad_norm_pre_clip': 8.234594345092773, 'timing/data': 0.023255908861756325, 'timing/model': 2.7624582201242447, 'learning_rate/qwen_vl_interface': 9.727377626696762e-06, 'learning_rate/action_model': 9.727377626696762e-05, 'epoch': 1.3}) +[2026-06-19 19:02:50,712][starVLA.training.train_starvla][INFO] - Step 632, Loss: {'train/loss': 0.14458686113357544, 'train/grad_norm_pre_clip': 5.913578510284424, 'timing/data': 0.02282620035111904, 'timing/model': 2.747922595590353, 'learning_rate/qwen_vl_interface': 9.726359817204165e-06, 'learning_rate/action_model': 9.726359817204166e-05, 'epoch': 1.3}) +[2026-06-19 19:03:01,420][starVLA.training.train_starvla][INFO] - Step 633, Loss: {'train/loss': 0.2096107006072998, 'train/grad_norm_pre_clip': 6.593381881713867, 'timing/data': 0.022919919341802597, 'timing/model': 2.748737746849656, 'learning_rate/qwen_vl_interface': 9.725340167649136e-06, 'learning_rate/action_model': 9.725340167649136e-05, 'epoch': 1.3}) +[2026-06-19 19:03:12,224][starVLA.training.train_starvla][INFO] - Step 634, Loss: {'train/loss': 0.14769116044044495, 'train/grad_norm_pre_clip': 6.12263298034668, 'timing/data': 0.022844763472676277, 'timing/model': 2.7693885657936335, 'learning_rate/qwen_vl_interface': 9.72431867845081e-06, 'learning_rate/action_model': 9.72431867845081e-05, 'epoch': 1.3}) +[2026-06-19 19:03:22,912][starVLA.training.train_starvla][INFO] - Step 635, Loss: {'train/loss': 0.5702125430107117, 'train/grad_norm_pre_clip': 9.899269104003906, 'timing/data': 0.02304350771009922, 'timing/model': 2.7338941898196936, 'learning_rate/qwen_vl_interface': 9.723295350029086e-06, 'learning_rate/action_model': 9.723295350029086e-05, 'epoch': 1.3}) +[2026-06-19 19:03:33,722][starVLA.training.train_starvla][INFO] - Step 636, Loss: {'train/loss': 0.22093188762664795, 'train/grad_norm_pre_clip': 6.483609199523926, 'timing/data': 0.023273015394806862, 'timing/model': 2.7572231739759445, 'learning_rate/qwen_vl_interface': 9.722270182804616e-06, 'learning_rate/action_model': 9.722270182804614e-05, 'epoch': 1.31}) +[2026-06-19 19:03:44,638][starVLA.training.train_starvla][INFO] - Step 637, Loss: {'train/loss': 0.30428117513656616, 'train/grad_norm_pre_clip': 8.109766960144043, 'timing/data': 0.023022066801786423, 'timing/model': 2.7161534894257784, 'learning_rate/qwen_vl_interface': 9.721243177198805e-06, 'learning_rate/action_model': 9.721243177198804e-05, 'epoch': 1.31}) +[2026-06-19 19:03:55,424][starVLA.training.train_starvla][INFO] - Step 638, Loss: {'train/loss': 0.1786588430404663, 'train/grad_norm_pre_clip': 3.2537379264831543, 'timing/data': 0.02280053310096264, 'timing/model': 2.75979539193213, 'learning_rate/qwen_vl_interface': 9.720214333633817e-06, 'learning_rate/action_model': 9.720214333633817e-05, 'epoch': 1.31}) +[2026-06-19 19:04:06,115][starVLA.training.train_starvla][INFO] - Step 639, Loss: {'train/loss': 0.12351522594690323, 'train/grad_norm_pre_clip': 4.838022708892822, 'timing/data': 0.02304054982960224, 'timing/model': 2.7356272656470537, 'learning_rate/qwen_vl_interface': 9.719183652532572e-06, 'learning_rate/action_model': 9.719183652532572e-05, 'epoch': 1.31}) +[2026-06-19 19:04:16,911][starVLA.training.train_starvla][INFO] - Step 640, Loss: {'train/loss': 0.16274596750736237, 'train/grad_norm_pre_clip': 5.047049045562744, 'timing/data': 0.02282639406621456, 'timing/model': 2.7516735829412937, 'learning_rate/qwen_vl_interface': 9.718151134318743e-06, 'learning_rate/action_model': 9.718151134318743e-05, 'epoch': 1.31}) +[2026-06-19 19:04:27,822][starVLA.training.train_starvla][INFO] - Step 641, Loss: {'train/loss': 0.15757423639297485, 'train/grad_norm_pre_clip': 4.9469499588012695, 'timing/data': 0.023150527849793434, 'timing/model': 2.7588372696191072, 'learning_rate/qwen_vl_interface': 9.71711677941676e-06, 'learning_rate/action_model': 9.717116779416758e-05, 'epoch': 1.32}) +[2026-06-19 19:04:38,624][starVLA.training.train_starvla][INFO] - Step 642, Loss: {'train/loss': 0.2358780950307846, 'train/grad_norm_pre_clip': 6.6486334800720215, 'timing/data': 0.022996611893177032, 'timing/model': 2.7672615833580494, 'learning_rate/qwen_vl_interface': 9.716080588251805e-06, 'learning_rate/action_model': 9.716080588251805e-05, 'epoch': 1.32}) +[2026-06-19 19:04:49,424][starVLA.training.train_starvla][INFO] - Step 643, Loss: {'train/loss': 0.1848086267709732, 'train/grad_norm_pre_clip': 4.738509654998779, 'timing/data': 0.023191459476947784, 'timing/model': 2.761920304968953, 'learning_rate/qwen_vl_interface': 9.71504256124982e-06, 'learning_rate/action_model': 9.71504256124982e-05, 'epoch': 1.32}) +[2026-06-19 19:05:00,225][starVLA.training.train_starvla][INFO] - Step 644, Loss: {'train/loss': 0.11530743539333344, 'train/grad_norm_pre_clip': 4.057904243469238, 'timing/data': 0.023382429033517838, 'timing/model': 2.7615096755325794, 'learning_rate/qwen_vl_interface': 9.714002698837499e-06, 'learning_rate/action_model': 9.714002698837498e-05, 'epoch': 1.32}) +[2026-06-19 19:05:11,022][starVLA.training.train_starvla][INFO] - Step 645, Loss: {'train/loss': 0.1340167224407196, 'train/grad_norm_pre_clip': 4.259034156799316, 'timing/data': 0.023618847131729126, 'timing/model': 2.758938269689679, 'learning_rate/qwen_vl_interface': 9.712961001442285e-06, 'learning_rate/action_model': 9.712961001442286e-05, 'epoch': 1.32}) +[2026-06-19 19:05:21,829][starVLA.training.train_starvla][INFO] - Step 646, Loss: {'train/loss': 0.1393989622592926, 'train/grad_norm_pre_clip': 21.97491455078125, 'timing/data': 0.02311103232204914, 'timing/model': 2.7755783945322037, 'learning_rate/qwen_vl_interface': 9.711917469492388e-06, 'learning_rate/action_model': 9.711917469492388e-05, 'epoch': 1.33}) +[2026-06-19 19:05:32,724][starVLA.training.train_starvla][INFO] - Step 647, Loss: {'train/loss': 0.09189971536397934, 'train/grad_norm_pre_clip': 4.353814601898193, 'timing/data': 0.023568248376250267, 'timing/model': 2.766459448263049, 'learning_rate/qwen_vl_interface': 9.71087210341676e-06, 'learning_rate/action_model': 9.71087210341676e-05, 'epoch': 1.33}) +[2026-06-19 19:05:43,522][starVLA.training.train_starvla][INFO] - Step 648, Loss: {'train/loss': 0.16057361662387848, 'train/grad_norm_pre_clip': 6.244725227355957, 'timing/data': 0.023202523589134216, 'timing/model': 2.759469533339143, 'learning_rate/qwen_vl_interface': 9.70982490364511e-06, 'learning_rate/action_model': 9.70982490364511e-05, 'epoch': 1.33}) +[2026-06-19 19:05:54,322][starVLA.training.train_starvla][INFO] - Step 649, Loss: {'train/loss': 0.17547717690467834, 'train/grad_norm_pre_clip': 7.056075096130371, 'timing/data': 0.023217391222715378, 'timing/model': 2.766142074018717, 'learning_rate/qwen_vl_interface': 9.708775870607905e-06, 'learning_rate/action_model': 9.708775870607904e-05, 'epoch': 1.33}) +[2026-06-19 19:06:05,124][starVLA.training.train_starvla][INFO] - Step 650, Loss: {'train/loss': 0.27655601501464844, 'train/grad_norm_pre_clip': 11.115291595458984, 'timing/data': 0.023165574297308922, 'timing/model': 2.7613416593521833, 'learning_rate/qwen_vl_interface': 9.707725004736363e-06, 'learning_rate/action_model': 9.707725004736363e-05, 'epoch': 1.34}) +[2026-06-19 19:06:15,930][starVLA.training.train_starvla][INFO] - Step 651, Loss: {'train/loss': 0.31855300068855286, 'train/grad_norm_pre_clip': 20.169687271118164, 'timing/data': 0.023652557283639908, 'timing/model': 2.763258844614029, 'learning_rate/qwen_vl_interface': 9.706672306462454e-06, 'learning_rate/action_model': 9.706672306462454e-05, 'epoch': 1.34}) +[2026-06-19 19:06:27,014][starVLA.training.train_starvla][INFO] - Step 652, Loss: {'train/loss': 0.11315039545297623, 'train/grad_norm_pre_clip': 7.4725542068481445, 'timing/data': 0.023099776357412338, 'timing/model': 2.950513243675232, 'learning_rate/qwen_vl_interface': 9.705617776218905e-06, 'learning_rate/action_model': 9.705617776218904e-05, 'epoch': 1.34}) +[2026-06-19 19:06:37,826][starVLA.training.train_starvla][INFO] - Step 653, Loss: {'train/loss': 0.08772408962249756, 'train/grad_norm_pre_clip': 10.419126510620117, 'timing/data': 0.02366056852042675, 'timing/model': 2.7623454108834267, 'learning_rate/qwen_vl_interface': 9.70456141443919e-06, 'learning_rate/action_model': 9.704561414439188e-05, 'epoch': 1.34}) +[2026-06-19 19:06:48,626][starVLA.training.train_starvla][INFO] - Step 654, Loss: {'train/loss': 0.15321514010429382, 'train/grad_norm_pre_clip': 12.436710357666016, 'timing/data': 0.023653170093894005, 'timing/model': 2.7618611119687557, 'learning_rate/qwen_vl_interface': 9.703503221557538e-06, 'learning_rate/action_model': 9.703503221557538e-05, 'epoch': 1.34}) +[2026-06-19 19:06:59,420][starVLA.training.train_starvla][INFO] - Step 655, Loss: {'train/loss': 0.2875070571899414, 'train/grad_norm_pre_clip': 11.741839408874512, 'timing/data': 0.022648602724075317, 'timing/model': 2.7610355857759714, 'learning_rate/qwen_vl_interface': 9.70244319800894e-06, 'learning_rate/action_model': 9.70244319800894e-05, 'epoch': 1.35}) +[2026-06-19 19:07:10,322][starVLA.training.train_starvla][INFO] - Step 656, Loss: {'train/loss': 0.06575867533683777, 'train/grad_norm_pre_clip': 5.324944019317627, 'timing/data': 0.023095404729247093, 'timing/model': 2.7915336173027754, 'learning_rate/qwen_vl_interface': 9.701381344229124e-06, 'learning_rate/action_model': 9.701381344229123e-05, 'epoch': 1.35}) +[2026-06-19 19:07:21,120][starVLA.training.train_starvla][INFO] - Step 657, Loss: {'train/loss': 0.45830440521240234, 'train/grad_norm_pre_clip': 30.29107666015625, 'timing/data': 0.023037977516651154, 'timing/model': 2.7573625668883324, 'learning_rate/qwen_vl_interface': 9.700317660654581e-06, 'learning_rate/action_model': 9.700317660654581e-05, 'epoch': 1.35}) +[2026-06-19 19:07:31,922][starVLA.training.train_starvla][INFO] - Step 658, Loss: {'train/loss': 0.7723957300186157, 'train/grad_norm_pre_clip': 27.224599838256836, 'timing/data': 0.023518983274698257, 'timing/model': 2.7585170920938253, 'learning_rate/qwen_vl_interface': 9.699252147722554e-06, 'learning_rate/action_model': 9.699252147722553e-05, 'epoch': 1.35}) +[2026-06-19 19:07:42,724][starVLA.training.train_starvla][INFO] - Step 659, Loss: {'train/loss': 0.5681145787239075, 'train/grad_norm_pre_clip': 18.19942283630371, 'timing/data': 0.02350408211350441, 'timing/model': 2.7584031857550144, 'learning_rate/qwen_vl_interface': 9.698184805871028e-06, 'learning_rate/action_model': 9.698184805871028e-05, 'epoch': 1.35}) +[2026-06-19 19:07:53,624][starVLA.training.train_starvla][INFO] - Step 660, Loss: {'train/loss': 0.18128114938735962, 'train/grad_norm_pre_clip': 8.769113540649414, 'timing/data': 0.02357783168554306, 'timing/model': 2.775781571865082, 'learning_rate/qwen_vl_interface': 9.697115635538753e-06, 'learning_rate/action_model': 9.697115635538753e-05, 'epoch': 1.36}) +[2026-06-19 19:08:04,421][starVLA.training.train_starvla][INFO] - Step 661, Loss: {'train/loss': 0.19362252950668335, 'train/grad_norm_pre_clip': 8.003217697143555, 'timing/data': 0.023303527384996414, 'timing/model': 2.770585360005498, 'learning_rate/qwen_vl_interface': 9.69604463716522e-06, 'learning_rate/action_model': 9.69604463716522e-05, 'epoch': 1.36}) +[2026-06-19 19:08:15,221][starVLA.training.train_starvla][INFO] - Step 662, Loss: {'train/loss': 0.5643110275268555, 'train/grad_norm_pre_clip': 26.9722843170166, 'timing/data': 0.023310596123337746, 'timing/model': 2.7619647439569235, 'learning_rate/qwen_vl_interface': 9.69497181119068e-06, 'learning_rate/action_model': 9.694971811190681e-05, 'epoch': 1.36}) +[2026-06-19 19:08:26,023][starVLA.training.train_starvla][INFO] - Step 663, Loss: {'train/loss': 0.1933601051568985, 'train/grad_norm_pre_clip': 9.82336711883545, 'timing/data': 0.021981772035360336, 'timing/model': 2.7627654187381268, 'learning_rate/qwen_vl_interface': 9.693897158056132e-06, 'learning_rate/action_model': 9.693897158056132e-05, 'epoch': 1.36}) +[2026-06-19 19:08:36,822][starVLA.training.train_starvla][INFO] - Step 664, Loss: {'train/loss': 0.3158564269542694, 'train/grad_norm_pre_clip': 11.595159530639648, 'timing/data': 0.025175223127007484, 'timing/model': 2.768733123317361, 'learning_rate/qwen_vl_interface': 9.692820678203322e-06, 'learning_rate/action_model': 9.692820678203322e-05, 'epoch': 1.36}) +[2026-06-19 19:08:47,623][starVLA.training.train_starvla][INFO] - Step 665, Loss: {'train/loss': 0.518484890460968, 'train/grad_norm_pre_clip': 24.908588409423828, 'timing/data': 0.023574965074658394, 'timing/model': 2.7598807867616415, 'learning_rate/qwen_vl_interface': 9.691742372074749e-06, 'learning_rate/action_model': 9.691742372074749e-05, 'epoch': 1.37}) +[2026-06-19 19:08:58,413][starVLA.training.train_starvla][INFO] - Step 666, Loss: {'train/loss': 0.6048749089241028, 'train/grad_norm_pre_clip': 17.068546295166016, 'timing/data': 0.023912519216537476, 'timing/model': 2.747084490954876, 'learning_rate/qwen_vl_interface': 9.690662240113669e-06, 'learning_rate/action_model': 9.690662240113668e-05, 'epoch': 1.37}) +[2026-06-19 19:09:09,226][starVLA.training.train_starvla][INFO] - Step 667, Loss: {'train/loss': 0.11817041039466858, 'train/grad_norm_pre_clip': 4.315395355224609, 'timing/data': 0.02346961572766304, 'timing/model': 2.77296594157815, 'learning_rate/qwen_vl_interface': 9.689580282764077e-06, 'learning_rate/action_model': 9.689580282764077e-05, 'epoch': 1.37}) +[2026-06-19 19:09:20,025][starVLA.training.train_starvla][INFO] - Step 668, Loss: {'train/loss': 0.2569692134857178, 'train/grad_norm_pre_clip': 5.828307628631592, 'timing/data': 0.02388611249625683, 'timing/model': 2.7572303488850594, 'learning_rate/qwen_vl_interface': 9.688496500470731e-06, 'learning_rate/action_model': 9.68849650047073e-05, 'epoch': 1.37}) +[2026-06-19 19:09:30,820][starVLA.training.train_starvla][INFO] - Step 669, Loss: {'train/loss': 0.12734153866767883, 'train/grad_norm_pre_clip': 11.368842124938965, 'timing/data': 0.0235143955796957, 'timing/model': 2.7538378834724426, 'learning_rate/qwen_vl_interface': 9.68741089367913e-06, 'learning_rate/action_model': 9.687410893679129e-05, 'epoch': 1.37}) +[2026-06-19 19:09:41,623][starVLA.training.train_starvla][INFO] - Step 670, Loss: {'train/loss': 0.1783275455236435, 'train/grad_norm_pre_clip': 8.812482833862305, 'timing/data': 0.023537665605545044, 'timing/model': 2.770926147699356, 'learning_rate/qwen_vl_interface': 9.686323462835527e-06, 'learning_rate/action_model': 9.686323462835526e-05, 'epoch': 1.38}) +[2026-06-19 19:09:52,423][starVLA.training.train_starvla][INFO] - Step 671, Loss: {'train/loss': 0.2907727062702179, 'train/grad_norm_pre_clip': 8.531895637512207, 'timing/data': 0.02375752292573452, 'timing/model': 2.7585706505924463, 'learning_rate/qwen_vl_interface': 9.685234208386922e-06, 'learning_rate/action_model': 9.685234208386921e-05, 'epoch': 1.38}) +[2026-06-19 19:10:03,413][starVLA.training.train_starvla][INFO] - Step 672, Loss: {'train/loss': 0.18801066279411316, 'train/grad_norm_pre_clip': 5.583706378936768, 'timing/data': 0.023870259523391724, 'timing/model': 2.752725848928094, 'learning_rate/qwen_vl_interface': 9.684143130781066e-06, 'learning_rate/action_model': 9.684143130781066e-05, 'epoch': 1.38}) +[2026-06-19 19:10:14,214][starVLA.training.train_starvla][INFO] - Step 673, Loss: {'train/loss': 0.16555307805538177, 'train/grad_norm_pre_clip': 2.872779369354248, 'timing/data': 0.023604366928339005, 'timing/model': 2.7666255813091993, 'learning_rate/qwen_vl_interface': 9.683050230466462e-06, 'learning_rate/action_model': 9.683050230466462e-05, 'epoch': 1.38}) +[2026-06-19 19:10:25,023][starVLA.training.train_starvla][INFO] - Step 674, Loss: {'train/loss': 0.10723760724067688, 'train/grad_norm_pre_clip': 5.440916061401367, 'timing/data': 0.02387598156929016, 'timing/model': 2.7606032975018024, 'learning_rate/qwen_vl_interface': 9.681955507892361e-06, 'learning_rate/action_model': 9.68195550789236e-05, 'epoch': 1.38}) +[2026-06-19 19:10:35,987][starVLA.training.train_starvla][INFO] - Step 675, Loss: {'train/loss': 0.15284153819084167, 'train/grad_norm_pre_clip': 4.684335708618164, 'timing/data': 0.029036130756139755, 'timing/model': 2.81192534416914, 'learning_rate/qwen_vl_interface': 9.680858963508759e-06, 'learning_rate/action_model': 9.68085896350876e-05, 'epoch': 1.39}) +[2026-06-19 19:10:46,825][starVLA.training.train_starvla][INFO] - Step 676, Loss: {'train/loss': 0.20454084873199463, 'train/grad_norm_pre_clip': 7.970993995666504, 'timing/data': 0.023114623501896858, 'timing/model': 2.7613019105046988, 'learning_rate/qwen_vl_interface': 9.679760597766405e-06, 'learning_rate/action_model': 9.679760597766405e-05, 'epoch': 1.39}) +[2026-06-19 19:10:57,621][starVLA.training.train_starvla][INFO] - Step 677, Loss: {'train/loss': 0.25327709317207336, 'train/grad_norm_pre_clip': 12.712312698364258, 'timing/data': 0.02282455563545227, 'timing/model': 2.7897784262895584, 'learning_rate/qwen_vl_interface': 9.678660411116796e-06, 'learning_rate/action_model': 9.678660411116796e-05, 'epoch': 1.39}) +[2026-06-19 19:11:08,422][starVLA.training.train_starvla][INFO] - Step 678, Loss: {'train/loss': 0.3339143693447113, 'train/grad_norm_pre_clip': 12.754923820495605, 'timing/data': 0.02327193319797516, 'timing/model': 2.7570399548858404, 'learning_rate/qwen_vl_interface': 9.677558404012178e-06, 'learning_rate/action_model': 9.677558404012177e-05, 'epoch': 1.39}) +[2026-06-19 19:11:19,441][starVLA.training.train_starvla][INFO] - Step 679, Loss: {'train/loss': 0.2029220312833786, 'train/grad_norm_pre_clip': 8.159093856811523, 'timing/data': 0.023310832679271698, 'timing/model': 2.976737892255187, 'learning_rate/qwen_vl_interface': 9.676454576905542e-06, 'learning_rate/action_model': 9.676454576905542e-05, 'epoch': 1.39}) +[2026-06-19 19:11:30,318][starVLA.training.train_starvla][INFO] - Step 680, Loss: {'train/loss': 0.16376058757305145, 'train/grad_norm_pre_clip': 7.676896572113037, 'timing/data': 0.023355728015303612, 'timing/model': 2.7534774355590343, 'learning_rate/qwen_vl_interface': 9.675348930250633e-06, 'learning_rate/action_model': 9.675348930250632e-05, 'epoch': 1.4}) +[2026-06-19 19:11:41,121][starVLA.training.train_starvla][INFO] - Step 681, Loss: {'train/loss': 0.2423870712518692, 'train/grad_norm_pre_clip': 7.113092422485352, 'timing/data': 0.02311927080154419, 'timing/model': 2.7611973844468594, 'learning_rate/qwen_vl_interface': 9.674241464501938e-06, 'learning_rate/action_model': 9.674241464501937e-05, 'epoch': 1.4}) +[2026-06-19 19:11:51,924][starVLA.training.train_starvla][INFO] - Step 682, Loss: {'train/loss': 0.2762572169303894, 'train/grad_norm_pre_clip': 5.430168628692627, 'timing/data': 0.02313690446317196, 'timing/model': 2.7600038535892963, 'learning_rate/qwen_vl_interface': 9.673132180114695e-06, 'learning_rate/action_model': 9.673132180114695e-05, 'epoch': 1.4}) +[2026-06-19 19:12:02,925][starVLA.training.train_starvla][INFO] - Step 683, Loss: {'train/loss': 0.18710245192050934, 'train/grad_norm_pre_clip': 7.071233749389648, 'timing/data': 0.022648587822914124, 'timing/model': 2.738731851801276, 'learning_rate/qwen_vl_interface': 9.67202107754489e-06, 'learning_rate/action_model': 9.672021077544889e-05, 'epoch': 1.4}) +[2026-06-19 19:12:13,721][starVLA.training.train_starvla][INFO] - Step 684, Loss: {'train/loss': 0.1264570951461792, 'train/grad_norm_pre_clip': 5.258248329162598, 'timing/data': 0.023346900939941406, 'timing/model': 2.756592432036996, 'learning_rate/qwen_vl_interface': 9.670908157249254e-06, 'learning_rate/action_model': 9.670908157249253e-05, 'epoch': 1.4}) +[2026-06-19 19:12:24,520][starVLA.training.train_starvla][INFO] - Step 685, Loss: {'train/loss': 0.2604981064796448, 'train/grad_norm_pre_clip': 7.968754768371582, 'timing/data': 0.023375850170850754, 'timing/model': 2.757177697494626, 'learning_rate/qwen_vl_interface': 9.669793419685267e-06, 'learning_rate/action_model': 9.669793419685266e-05, 'epoch': 1.41}) +[2026-06-19 19:12:35,323][starVLA.training.train_starvla][INFO] - Step 686, Loss: {'train/loss': 0.3575630486011505, 'train/grad_norm_pre_clip': 7.205647945404053, 'timing/data': 0.022798769176006317, 'timing/model': 2.7605832125991583, 'learning_rate/qwen_vl_interface': 9.668676865311155e-06, 'learning_rate/action_model': 9.668676865311154e-05, 'epoch': 1.41}) +[2026-06-19 19:12:46,223][starVLA.training.train_starvla][INFO] - Step 687, Loss: {'train/loss': 0.11765839904546738, 'train/grad_norm_pre_clip': 6.438945293426514, 'timing/data': 0.023370401933789253, 'timing/model': 2.7627859096974134, 'learning_rate/qwen_vl_interface': 9.667558494585891e-06, 'learning_rate/action_model': 9.66755849458589e-05, 'epoch': 1.41}) +[2026-06-19 19:12:57,010][starVLA.training.train_starvla][INFO] - Step 688, Loss: {'train/loss': 0.27938568592071533, 'train/grad_norm_pre_clip': 8.376493453979492, 'timing/data': 0.023486560210585594, 'timing/model': 2.747392848134041, 'learning_rate/qwen_vl_interface': 9.666438307969194e-06, 'learning_rate/action_model': 9.666438307969195e-05, 'epoch': 1.41}) +[2026-06-19 19:13:07,823][starVLA.training.train_starvla][INFO] - Step 689, Loss: {'train/loss': 0.07928533107042313, 'train/grad_norm_pre_clip': 3.1296706199645996, 'timing/data': 0.023345479741692543, 'timing/model': 2.764398742467165, 'learning_rate/qwen_vl_interface': 9.665316305921535e-06, 'learning_rate/action_model': 9.665316305921535e-05, 'epoch': 1.42}) +[2026-06-19 19:13:18,623][starVLA.training.train_starvla][INFO] - Step 690, Loss: {'train/loss': 0.21655625104904175, 'train/grad_norm_pre_clip': 7.630858421325684, 'timing/data': 0.023604191839694977, 'timing/model': 2.7664660420268774, 'learning_rate/qwen_vl_interface': 9.664192488904123e-06, 'learning_rate/action_model': 9.664192488904122e-05, 'epoch': 1.42}) +[2026-06-19 19:13:29,421][starVLA.training.train_starvla][INFO] - Step 691, Loss: {'train/loss': 0.2104869782924652, 'train/grad_norm_pre_clip': 7.021061420440674, 'timing/data': 0.02391492761671543, 'timing/model': 2.760440358892083, 'learning_rate/qwen_vl_interface': 9.663066857378917e-06, 'learning_rate/action_model': 9.663066857378917e-05, 'epoch': 1.42}) +[2026-06-19 19:13:40,222][starVLA.training.train_starvla][INFO] - Step 692, Loss: {'train/loss': 0.24104152619838715, 'train/grad_norm_pre_clip': 9.145027160644531, 'timing/data': 0.0232378039509058, 'timing/model': 2.755166271701455, 'learning_rate/qwen_vl_interface': 9.661939411808623e-06, 'learning_rate/action_model': 9.661939411808622e-05, 'epoch': 1.42}) +[2026-06-19 19:13:51,028][starVLA.training.train_starvla][INFO] - Step 693, Loss: {'train/loss': 0.15481635928153992, 'train/grad_norm_pre_clip': 4.340765953063965, 'timing/data': 0.023920457810163498, 'timing/model': 2.7600800674408674, 'learning_rate/qwen_vl_interface': 9.66081015265669e-06, 'learning_rate/action_model': 9.660810152656689e-05, 'epoch': 1.42}) +[2026-06-19 19:14:01,921][starVLA.training.train_starvla][INFO] - Step 694, Loss: {'train/loss': 0.26904234290122986, 'train/grad_norm_pre_clip': 5.7386088371276855, 'timing/data': 0.023399053141474724, 'timing/model': 2.7605098988860846, 'learning_rate/qwen_vl_interface': 9.659679080387314e-06, 'learning_rate/action_model': 9.659679080387313e-05, 'epoch': 1.43}) +[2026-06-19 19:14:12,838][starVLA.training.train_starvla][INFO] - Step 695, Loss: {'train/loss': 0.1149793192744255, 'train/grad_norm_pre_clip': 5.05486536026001, 'timing/data': 0.023305296897888184, 'timing/model': 2.7124289367347956, 'learning_rate/qwen_vl_interface': 9.658546195465436e-06, 'learning_rate/action_model': 9.658546195465435e-05, 'epoch': 1.43}) +[2026-06-19 19:14:23,713][starVLA.training.train_starvla][INFO] - Step 696, Loss: {'train/loss': 0.22934022545814514, 'train/grad_norm_pre_clip': 12.402290344238281, 'timing/data': 0.02326003462076187, 'timing/model': 2.7684781160205603, 'learning_rate/qwen_vl_interface': 9.65741149835674e-06, 'learning_rate/action_model': 9.657411498356741e-05, 'epoch': 1.43}) +[2026-06-19 19:14:34,423][starVLA.training.train_starvla][INFO] - Step 697, Loss: {'train/loss': 0.2868211269378662, 'train/grad_norm_pre_clip': 9.849251747131348, 'timing/data': 0.023233452811837196, 'timing/model': 2.7496463004499674, 'learning_rate/qwen_vl_interface': 9.656274989527663e-06, 'learning_rate/action_model': 9.656274989527664e-05, 'epoch': 1.43}) +[2026-06-19 19:14:45,450][starVLA.training.train_starvla][INFO] - Step 698, Loss: {'train/loss': 0.28326770663261414, 'train/grad_norm_pre_clip': 5.4192070960998535, 'timing/data': 0.023160988464951515, 'timing/model': 3.004452830180526, 'learning_rate/qwen_vl_interface': 9.655136669445378e-06, 'learning_rate/action_model': 9.655136669445378e-05, 'epoch': 1.43}) +[2026-06-19 19:14:56,113][starVLA.training.train_starvla][INFO] - Step 699, Loss: {'train/loss': 0.19748464226722717, 'train/grad_norm_pre_clip': 5.205318450927734, 'timing/data': 0.02354966104030609, 'timing/model': 2.7336227484047413, 'learning_rate/qwen_vl_interface': 9.653996538577802e-06, 'learning_rate/action_model': 9.653996538577802e-05, 'epoch': 1.44}) +[2026-06-19 19:15:06,912][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 700 +[2026-06-19 19:22:30,002][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 700 +[2026-06-19 19:22:30,003][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 700 +[2026-06-19 19:30:27,133][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 700 +[2026-06-19 19:30:27,136][starVLA.training.train_starvla][INFO] - Step 700, Loss: {'train/loss': 0.2822785973548889, 'train/grad_norm_pre_clip': 6.0742034912109375, 'eval/loss': 0.38110387325286865, 'eval/latency_0/loss': 0.11006680876016617, 'eval/latency_1/loss': 0.2350303679704666, 'eval/latency_2/loss': 0.4750823676586151, 'eval/latency_3/loss': 0.42741385102272034, 'eval/latency_4/loss': 0.6400138139724731, 'eval/flappy/loss': 0.38110387325286865, 'eval/flappy/latency_0/loss': 0.11006680876016617, 'eval/flappy/latency_1/loss': 0.2350303679704666, 'eval/flappy/latency_2/loss': 0.4750823676586151, 'eval/flappy/latency_3/loss': 0.42741385102272034, 'eval/flappy/latency_4/loss': 0.6400138139724731, 'eval/flap_precision': 0.6138888888888889, 'eval/flap_recall': 0.8036363636363636, 'eval/flap_f1': 0.6960629921259842, 'eval/flap_rate_model': 0.1, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.6960629921259842, 'eval/latency_0/flap_precision': 0.6138888888888889, 'eval/latency_0/flap_recall': 0.8036363636363636, 'eval/latency_0/flap_f1': 0.6960629921259842, 'eval/latency_0/flap_rate_model': 0.1, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.6960629921259842, 'timing/data': 0.023279614746570587, 'timing/model': 2.7515396382659674, 'learning_rate/qwen_vl_interface': 9.652854597393604e-06, 'learning_rate/action_model': 9.652854597393603e-05, 'epoch': 1.44}) +[2026-06-19 19:30:27,136][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_700_state +[2026-06-19 19:30:27,141][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-19 19:30:32,682][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_700_state/model.safetensors +[2026-06-19 19:30:40,710][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_700_state/optimizer.bin +[2026-06-19 19:30:40,711][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_700_state/sampler.bin +[2026-06-19 19:30:40,711][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_700_state/sampler_1.bin +[2026-06-19 19:30:40,712][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_700_state/random_states_0.pkl +[2026-06-19 19:30:40,834][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 700 +[2026-06-19 19:30:40,834][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-19 19:30:40,848][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-19 19:30:51,613][starVLA.training.train_starvla][INFO] - Step 701, Loss: {'train/loss': 0.34348681569099426, 'train/grad_norm_pre_clip': 6.106568813323975, 'timing/data': 0.02478216402232647, 'timing/model': 2.767452172935009, 'learning_rate/qwen_vl_interface': 9.651710846362192e-06, 'learning_rate/action_model': 9.651710846362192e-05, 'epoch': 1.44}) +[2026-06-19 19:31:02,414][starVLA.training.train_starvla][INFO] - Step 702, Loss: {'train/loss': 0.14742960035800934, 'train/grad_norm_pre_clip': 4.725160598754883, 'timing/data': 0.023307841271162033, 'timing/model': 2.765811437740922, 'learning_rate/qwen_vl_interface': 9.650565285953716e-06, 'learning_rate/action_model': 9.650565285953715e-05, 'epoch': 1.44}) +[2026-06-19 19:31:13,117][starVLA.training.train_starvla][INFO] - Step 703, Loss: {'train/loss': 0.2719106674194336, 'train/grad_norm_pre_clip': 6.222769260406494, 'timing/data': 0.023307081311941147, 'timing/model': 2.7440487891435623, 'learning_rate/qwen_vl_interface': 9.649417916639077e-06, 'learning_rate/action_model': 9.649417916639076e-05, 'epoch': 1.44}) +[2026-06-19 19:31:23,918][starVLA.training.train_starvla][INFO] - Step 704, Loss: {'train/loss': 0.14495916664600372, 'train/grad_norm_pre_clip': 3.524913787841797, 'timing/data': 0.02336888015270233, 'timing/model': 2.7682104744017124, 'learning_rate/qwen_vl_interface': 9.648268738889911e-06, 'learning_rate/action_model': 9.648268738889911e-05, 'epoch': 1.45}) +[2026-06-19 19:31:34,712][starVLA.training.train_starvla][INFO] - Step 705, Loss: {'train/loss': 0.15830187499523163, 'train/grad_norm_pre_clip': 2.8606460094451904, 'timing/data': 0.022923963144421577, 'timing/model': 2.752171376720071, 'learning_rate/qwen_vl_interface': 9.647117753178604e-06, 'learning_rate/action_model': 9.647117753178604e-05, 'epoch': 1.45}) +[2026-06-19 19:31:45,735][starVLA.training.train_starvla][INFO] - Step 706, Loss: {'train/loss': 0.2657551169395447, 'train/grad_norm_pre_clip': 7.2724690437316895, 'timing/data': 0.028451362624764442, 'timing/model': 2.763987535610795, 'learning_rate/qwen_vl_interface': 9.645964959978284e-06, 'learning_rate/action_model': 9.645964959978284e-05, 'epoch': 1.45}) +[2026-06-19 19:31:56,620][starVLA.training.train_starvla][INFO] - Step 707, Loss: {'train/loss': 0.47210779786109924, 'train/grad_norm_pre_clip': 15.61213493347168, 'timing/data': 0.029590819031000137, 'timing/model': 2.7388569097965956, 'learning_rate/qwen_vl_interface': 9.644810359762818e-06, 'learning_rate/action_model': 9.644810359762817e-05, 'epoch': 1.45}) +[2026-06-19 19:32:07,418][starVLA.training.train_starvla][INFO] - Step 708, Loss: {'train/loss': 0.3626343011856079, 'train/grad_norm_pre_clip': 8.878658294677734, 'timing/data': 0.02342607080936432, 'timing/model': 2.758426371961832, 'learning_rate/qwen_vl_interface': 9.64365395300682e-06, 'learning_rate/action_model': 9.643653953006819e-05, 'epoch': 1.45}) +[2026-06-19 19:32:18,219][starVLA.training.train_starvla][INFO] - Step 709, Loss: {'train/loss': 0.19565033912658691, 'train/grad_norm_pre_clip': 3.593019962310791, 'timing/data': 0.02286173589527607, 'timing/model': 2.7687899190932512, 'learning_rate/qwen_vl_interface': 9.642495740185644e-06, 'learning_rate/action_model': 9.642495740185643e-05, 'epoch': 1.46}) +[2026-06-19 19:32:29,020][starVLA.training.train_starvla][INFO] - Step 710, Loss: {'train/loss': 0.2441745400428772, 'train/grad_norm_pre_clip': 6.134877681732178, 'timing/data': 0.023194724693894386, 'timing/model': 2.7616905495524406, 'learning_rate/qwen_vl_interface': 9.641335721775388e-06, 'learning_rate/action_model': 9.641335721775387e-05, 'epoch': 1.46}) +[2026-06-19 19:32:39,821][starVLA.training.train_starvla][INFO] - Step 711, Loss: {'train/loss': 0.14585313200950623, 'train/grad_norm_pre_clip': 3.0344905853271484, 'timing/data': 0.023389698937535286, 'timing/model': 2.7613435368984938, 'learning_rate/qwen_vl_interface': 9.640173898252893e-06, 'learning_rate/action_model': 9.640173898252892e-05, 'epoch': 1.46}) +[2026-06-19 19:32:50,621][starVLA.training.train_starvla][INFO] - Step 712, Loss: {'train/loss': 0.2547018229961395, 'train/grad_norm_pre_clip': 4.666520118713379, 'timing/data': 0.022824475541710854, 'timing/model': 2.758881686255336, 'learning_rate/qwen_vl_interface': 9.639010270095738e-06, 'learning_rate/action_model': 9.639010270095739e-05, 'epoch': 1.46}) +[2026-06-19 19:33:01,585][starVLA.training.train_starvla][INFO] - Step 713, Loss: {'train/loss': 0.22659003734588623, 'train/grad_norm_pre_clip': 6.4198527336120605, 'timing/data': 0.023532100021839142, 'timing/model': 2.922297690063715, 'learning_rate/qwen_vl_interface': 9.637844837782251e-06, 'learning_rate/action_model': 9.63784483778225e-05, 'epoch': 1.46}) +[2026-06-19 19:33:12,425][starVLA.training.train_starvla][INFO] - Step 714, Loss: {'train/loss': 0.19660498201847076, 'train/grad_norm_pre_clip': 5.784811019897461, 'timing/data': 0.023566318675875664, 'timing/model': 2.7641696613281965, 'learning_rate/qwen_vl_interface': 9.636677601791494e-06, 'learning_rate/action_model': 9.636677601791493e-05, 'epoch': 1.47}) +[2026-06-19 19:33:23,312][starVLA.training.train_starvla][INFO] - Step 715, Loss: {'train/loss': 0.12409981340169907, 'train/grad_norm_pre_clip': 3.1824820041656494, 'timing/data': 0.02359509840607643, 'timing/model': 2.7501702196896076, 'learning_rate/qwen_vl_interface': 9.635508562603273e-06, 'learning_rate/action_model': 9.635508562603272e-05, 'epoch': 1.47}) +[2026-06-19 19:33:34,113][starVLA.training.train_starvla][INFO] - Step 716, Loss: {'train/loss': 0.6542242765426636, 'train/grad_norm_pre_clip': 14.3995361328125, 'timing/data': 0.023353029042482376, 'timing/model': 2.7536891624331474, 'learning_rate/qwen_vl_interface': 9.634337720698139e-06, 'learning_rate/action_model': 9.634337720698137e-05, 'epoch': 1.47}) +[2026-06-19 19:33:45,125][starVLA.training.train_starvla][INFO] - Step 717, Loss: {'train/loss': 0.20476563274860382, 'train/grad_norm_pre_clip': 7.225001811981201, 'timing/data': 0.023408278822898865, 'timing/model': 2.778021674603224, 'learning_rate/qwen_vl_interface': 9.633165076557376e-06, 'learning_rate/action_model': 9.633165076557375e-05, 'epoch': 1.47}) +[2026-06-19 19:33:55,924][starVLA.training.train_starvla][INFO] - Step 718, Loss: {'train/loss': 0.27804285287857056, 'train/grad_norm_pre_clip': 7.150699615478516, 'timing/data': 0.023423438891768456, 'timing/model': 2.769856685772538, 'learning_rate/qwen_vl_interface': 9.631990630663019e-06, 'learning_rate/action_model': 9.631990630663018e-05, 'epoch': 1.47}) +[2026-06-19 19:34:06,727][starVLA.training.train_starvla][INFO] - Step 719, Loss: {'train/loss': 0.5659098625183105, 'train/grad_norm_pre_clip': 13.022526741027832, 'timing/data': 0.02553301677107811, 'timing/model': 2.7541558761149645, 'learning_rate/qwen_vl_interface': 9.630814383497836e-06, 'learning_rate/action_model': 9.630814383497835e-05, 'epoch': 1.48}) +[2026-06-19 19:34:17,521][starVLA.training.train_starvla][INFO] - Step 720, Loss: {'train/loss': 0.1081572100520134, 'train/grad_norm_pre_clip': 2.899064064025879, 'timing/data': 0.023106982931494713, 'timing/model': 2.7607570625841618, 'learning_rate/qwen_vl_interface': 9.629636335545336e-06, 'learning_rate/action_model': 9.629636335545335e-05, 'epoch': 1.48}) +[2026-06-19 19:34:28,523][starVLA.training.train_starvla][INFO] - Step 721, Loss: {'train/loss': 0.14898639917373657, 'train/grad_norm_pre_clip': 4.644529819488525, 'timing/data': 0.024212129414081573, 'timing/model': 2.757185272872448, 'learning_rate/qwen_vl_interface': 9.628456487289773e-06, 'learning_rate/action_model': 9.628456487289773e-05, 'epoch': 1.48}) +[2026-06-19 19:34:39,325][starVLA.training.train_starvla][INFO] - Step 722, Loss: {'train/loss': 0.37300097942352295, 'train/grad_norm_pre_clip': 10.138298988342285, 'timing/data': 0.022899555042386055, 'timing/model': 2.7581675723195076, 'learning_rate/qwen_vl_interface': 9.627274839216135e-06, 'learning_rate/action_model': 9.627274839216136e-05, 'epoch': 1.48}) +[2026-06-19 19:34:50,236][starVLA.training.train_starvla][INFO] - Step 723, Loss: {'train/loss': 0.12955781817436218, 'train/grad_norm_pre_clip': 5.883690357208252, 'timing/data': 0.024159589782357216, 'timing/model': 2.759309194982052, 'learning_rate/qwen_vl_interface': 9.626091391810158e-06, 'learning_rate/action_model': 9.626091391810157e-05, 'epoch': 1.48}) +[2026-06-19 19:35:01,117][starVLA.training.train_starvla][INFO] - Step 724, Loss: {'train/loss': 0.21125704050064087, 'train/grad_norm_pre_clip': 6.275787353515625, 'timing/data': 0.02320045977830887, 'timing/model': 2.75666013173759, 'learning_rate/qwen_vl_interface': 9.624906145558309e-06, 'learning_rate/action_model': 9.624906145558307e-05, 'epoch': 1.49}) +[2026-06-19 19:35:11,923][starVLA.training.train_starvla][INFO] - Step 725, Loss: {'train/loss': 0.1335754096508026, 'train/grad_norm_pre_clip': 4.3163228034973145, 'timing/data': 0.023962443694472313, 'timing/model': 2.764565048739314, 'learning_rate/qwen_vl_interface': 9.623719100947795e-06, 'learning_rate/action_model': 9.623719100947795e-05, 'epoch': 1.49}) +[2026-06-19 19:35:22,729][starVLA.training.train_starvla][INFO] - Step 726, Loss: {'train/loss': 0.2870453894138336, 'train/grad_norm_pre_clip': 10.470602035522461, 'timing/data': 0.024017203599214554, 'timing/model': 2.7625406216830015, 'learning_rate/qwen_vl_interface': 9.622530258466571e-06, 'learning_rate/action_model': 9.62253025846657e-05, 'epoch': 1.49}) +[2026-06-19 19:35:33,626][starVLA.training.train_starvla][INFO] - Step 727, Loss: {'train/loss': 0.2080572247505188, 'train/grad_norm_pre_clip': 6.259969234466553, 'timing/data': 0.02380678616464138, 'timing/model': 2.7631193455308676, 'learning_rate/qwen_vl_interface': 9.62133961860332e-06, 'learning_rate/action_model': 9.62133961860332e-05, 'epoch': 1.49}) +[2026-06-19 19:35:44,427][starVLA.training.train_starvla][INFO] - Step 728, Loss: {'train/loss': 0.4419296979904175, 'train/grad_norm_pre_clip': 13.911166191101074, 'timing/data': 0.024175940081477165, 'timing/model': 2.758716568350792, 'learning_rate/qwen_vl_interface': 9.620147181847474e-06, 'learning_rate/action_model': 9.620147181847475e-05, 'epoch': 1.5}) +[2026-06-19 19:35:55,223][starVLA.training.train_starvla][INFO] - Step 729, Loss: {'train/loss': 0.17662060260772705, 'train/grad_norm_pre_clip': 5.085665225982666, 'timing/data': 0.02360398694872856, 'timing/model': 2.7591055650264025, 'learning_rate/qwen_vl_interface': 9.618952948689196e-06, 'learning_rate/action_model': 9.618952948689196e-05, 'epoch': 1.5}) +[2026-06-19 19:36:06,024][starVLA.training.train_starvla][INFO] - Step 730, Loss: {'train/loss': 0.1426129937171936, 'train/grad_norm_pre_clip': 5.556804656982422, 'timing/data': 0.023097220808267593, 'timing/model': 2.7584443539381027, 'learning_rate/qwen_vl_interface': 9.617756919619393e-06, 'learning_rate/action_model': 9.617756919619392e-05, 'epoch': 1.5}) +[2026-06-19 19:36:16,826][starVLA.training.train_starvla][INFO] - Step 731, Loss: {'train/loss': 0.12906093895435333, 'train/grad_norm_pre_clip': 5.601204872131348, 'timing/data': 0.023262672126293182, 'timing/model': 2.762513481080532, 'learning_rate/qwen_vl_interface': 9.616559095129701e-06, 'learning_rate/action_model': 9.6165590951297e-05, 'epoch': 1.5}) +[2026-06-19 19:36:27,624][starVLA.training.train_starvla][INFO] - Step 732, Loss: {'train/loss': 0.156679168343544, 'train/grad_norm_pre_clip': 7.770115852355957, 'timing/data': 0.023282110691070557, 'timing/model': 2.758666144683957, 'learning_rate/qwen_vl_interface': 9.615359475712506e-06, 'learning_rate/action_model': 9.615359475712506e-05, 'epoch': 1.5}) +[2026-06-19 19:36:38,430][starVLA.training.train_starvla][INFO] - Step 733, Loss: {'train/loss': 0.14842703938484192, 'train/grad_norm_pre_clip': 10.29772663116455, 'timing/data': 0.02343468740582466, 'timing/model': 2.771987186744809, 'learning_rate/qwen_vl_interface': 9.614158061860924e-06, 'learning_rate/action_model': 9.614158061860924e-05, 'epoch': 1.51}) +[2026-06-19 19:36:49,430][starVLA.training.train_starvla][INFO] - Step 734, Loss: {'train/loss': 0.2103140950202942, 'train/grad_norm_pre_clip': 8.46744155883789, 'timing/data': 0.03892333805561066, 'timing/model': 2.7454709243029356, 'learning_rate/qwen_vl_interface': 9.612954854068813e-06, 'learning_rate/action_model': 9.612954854068812e-05, 'epoch': 1.51}) +[2026-06-19 19:37:00,227][starVLA.training.train_starvla][INFO] - Step 735, Loss: {'train/loss': 0.23879961669445038, 'train/grad_norm_pre_clip': 13.802791595458984, 'timing/data': 0.023579854518175125, 'timing/model': 2.74229877255857, 'learning_rate/qwen_vl_interface': 9.611749852830763e-06, 'learning_rate/action_model': 9.611749852830763e-05, 'epoch': 1.51}) +[2026-06-19 19:37:11,024][starVLA.training.train_starvla][INFO] - Step 736, Loss: {'train/loss': 0.31114035844802856, 'train/grad_norm_pre_clip': 18.94083023071289, 'timing/data': 0.023450369015336037, 'timing/model': 2.755672425031662, 'learning_rate/qwen_vl_interface': 9.610543058642107e-06, 'learning_rate/action_model': 9.610543058642107e-05, 'epoch': 1.51}) +[2026-06-19 19:37:21,825][starVLA.training.train_starvla][INFO] - Step 737, Loss: {'train/loss': 0.19696103036403656, 'train/grad_norm_pre_clip': 7.344390869140625, 'timing/data': 0.023458588868379593, 'timing/model': 2.758467610925436, 'learning_rate/qwen_vl_interface': 9.609334471998911e-06, 'learning_rate/action_model': 9.609334471998912e-05, 'epoch': 1.51}) +[2026-06-19 19:37:32,625][starVLA.training.train_starvla][INFO] - Step 738, Loss: {'train/loss': 0.24214933812618256, 'train/grad_norm_pre_clip': 10.366589546203613, 'timing/data': 0.02347508817911148, 'timing/model': 2.7561448011547327, 'learning_rate/qwen_vl_interface': 9.60812409339798e-06, 'learning_rate/action_model': 9.60812409339798e-05, 'epoch': 1.52}) +[2026-06-19 19:37:43,427][starVLA.training.train_starvla][INFO] - Step 739, Loss: {'train/loss': 0.1447131484746933, 'train/grad_norm_pre_clip': 5.0246477127075195, 'timing/data': 0.02343648672103882, 'timing/model': 2.7619541361927986, 'learning_rate/qwen_vl_interface': 9.606911923336856e-06, 'learning_rate/action_model': 9.606911923336856e-05, 'epoch': 1.52}) +[2026-06-19 19:37:54,229][starVLA.training.train_starvla][INFO] - Step 740, Loss: {'train/loss': 0.40511465072631836, 'train/grad_norm_pre_clip': 9.99147891998291, 'timing/data': 0.023812076076865196, 'timing/model': 2.761736363172531, 'learning_rate/qwen_vl_interface': 9.605697962313817e-06, 'learning_rate/action_model': 9.605697962313816e-05, 'epoch': 1.52}) +[2026-06-19 19:38:05,128][starVLA.training.train_starvla][INFO] - Step 741, Loss: {'train/loss': 0.22945263981819153, 'train/grad_norm_pre_clip': 5.036732196807861, 'timing/data': 0.023778080940246582, 'timing/model': 2.7615941017866135, 'learning_rate/qwen_vl_interface': 9.604482210827871e-06, 'learning_rate/action_model': 9.604482210827871e-05, 'epoch': 1.52}) +[2026-06-19 19:38:16,028][starVLA.training.train_starvla][INFO] - Step 742, Loss: {'train/loss': 0.2277870923280716, 'train/grad_norm_pre_clip': 12.508380889892578, 'timing/data': 0.023354066535830498, 'timing/model': 2.7566575407981873, 'learning_rate/qwen_vl_interface': 9.603264669378772e-06, 'learning_rate/action_model': 9.603264669378772e-05, 'epoch': 1.52}) +[2026-06-19 19:38:26,928][starVLA.training.train_starvla][INFO] - Step 743, Loss: {'train/loss': 0.3534919321537018, 'train/grad_norm_pre_clip': 7.434678554534912, 'timing/data': 0.02327575907111168, 'timing/model': 2.762443797662854, 'learning_rate/qwen_vl_interface': 9.602045338467007e-06, 'learning_rate/action_model': 9.602045338467006e-05, 'epoch': 1.53}) +[2026-06-19 19:38:37,718][starVLA.training.train_starvla][INFO] - Step 744, Loss: {'train/loss': 0.25770050287246704, 'train/grad_norm_pre_clip': 21.56616973876953, 'timing/data': 0.023674514144659042, 'timing/model': 2.753891944885254, 'learning_rate/qwen_vl_interface': 9.600824218593791e-06, 'learning_rate/action_model': 9.600824218593791e-05, 'epoch': 1.53}) +[2026-06-19 19:38:48,526][starVLA.training.train_starvla][INFO] - Step 745, Loss: {'train/loss': 0.111415334045887, 'train/grad_norm_pre_clip': 6.39248514175415, 'timing/data': 0.023322608321905136, 'timing/model': 2.766769964247942, 'learning_rate/qwen_vl_interface': 9.599601310261087e-06, 'learning_rate/action_model': 9.599601310261085e-05, 'epoch': 1.53}) +[2026-06-19 19:38:59,329][starVLA.training.train_starvla][INFO] - Step 746, Loss: {'train/loss': 0.20529836416244507, 'train/grad_norm_pre_clip': 4.979381561279297, 'timing/data': 0.02597229741513729, 'timing/model': 2.7666800431907177, 'learning_rate/qwen_vl_interface': 9.598376613971582e-06, 'learning_rate/action_model': 9.598376613971582e-05, 'epoch': 1.53}) +[2026-06-19 19:39:10,226][starVLA.training.train_starvla][INFO] - Step 747, Loss: {'train/loss': 0.25072571635246277, 'train/grad_norm_pre_clip': 5.039496421813965, 'timing/data': 0.022868674248456955, 'timing/model': 2.765912877395749, 'learning_rate/qwen_vl_interface': 9.597150130228703e-06, 'learning_rate/action_model': 9.597150130228703e-05, 'epoch': 1.53}) +[2026-06-19 19:39:21,026][starVLA.training.train_starvla][INFO] - Step 748, Loss: {'train/loss': 0.3403940200805664, 'train/grad_norm_pre_clip': 5.55670166015625, 'timing/data': 0.023199094459414482, 'timing/model': 2.7603550516068935, 'learning_rate/qwen_vl_interface': 9.595921859536612e-06, 'learning_rate/action_model': 9.595921859536611e-05, 'epoch': 1.54}) +[2026-06-19 19:39:31,824][starVLA.training.train_starvla][INFO] - Step 749, Loss: {'train/loss': 0.16604480147361755, 'train/grad_norm_pre_clip': 2.902505397796631, 'timing/data': 0.022914787754416466, 'timing/model': 2.761599527671933, 'learning_rate/qwen_vl_interface': 9.594691802400203e-06, 'learning_rate/action_model': 9.594691802400202e-05, 'epoch': 1.54}) +[2026-06-19 19:39:42,627][starVLA.training.train_starvla][INFO] - Step 750, Loss: {'train/loss': 0.17847086489200592, 'train/grad_norm_pre_clip': 3.6817617416381836, 'timing/data': 0.023145833984017372, 'timing/model': 2.76344246044755, 'learning_rate/qwen_vl_interface': 9.59345995932511e-06, 'learning_rate/action_model': 9.59345995932511e-05, 'epoch': 1.54}) +[2026-06-19 19:39:53,526][starVLA.training.train_starvla][INFO] - Step 751, Loss: {'train/loss': 0.20817457139492035, 'train/grad_norm_pre_clip': 5.011748790740967, 'timing/data': 0.02300950512290001, 'timing/model': 2.7592801731079817, 'learning_rate/qwen_vl_interface': 9.592226330817693e-06, 'learning_rate/action_model': 9.592226330817693e-05, 'epoch': 1.54}) +[2026-06-19 19:40:04,326][starVLA.training.train_starvla][INFO] - Step 752, Loss: {'train/loss': 0.17466673254966736, 'train/grad_norm_pre_clip': 6.441939830780029, 'timing/data': 0.022866690531373024, 'timing/model': 2.765149449929595, 'learning_rate/qwen_vl_interface': 9.590990917385052e-06, 'learning_rate/action_model': 9.590990917385052e-05, 'epoch': 1.54}) +[2026-06-19 19:40:15,122][starVLA.training.train_starvla][INFO] - Step 753, Loss: {'train/loss': 0.1876225620508194, 'train/grad_norm_pre_clip': 5.801787853240967, 'timing/data': 0.023597421124577522, 'timing/model': 2.754383757710457, 'learning_rate/qwen_vl_interface': 9.58975371953502e-06, 'learning_rate/action_model': 9.589753719535018e-05, 'epoch': 1.55}) +[2026-06-19 19:40:25,926][starVLA.training.train_starvla][INFO] - Step 754, Loss: {'train/loss': 0.1408623307943344, 'train/grad_norm_pre_clip': 4.055177688598633, 'timing/data': 0.02284776046872139, 'timing/model': 2.771339427679777, 'learning_rate/qwen_vl_interface': 9.588514737776158e-06, 'learning_rate/action_model': 9.588514737776157e-05, 'epoch': 1.55}) +[2026-06-19 19:40:36,730][starVLA.training.train_starvla][INFO] - Step 755, Loss: {'train/loss': 0.3125779926776886, 'train/grad_norm_pre_clip': 11.585591316223145, 'timing/data': 0.022602083161473274, 'timing/model': 2.7679143752902746, 'learning_rate/qwen_vl_interface': 9.587273972617769e-06, 'learning_rate/action_model': 9.587273972617768e-05, 'epoch': 1.55}) +[2026-06-19 19:40:47,623][starVLA.training.train_starvla][INFO] - Step 756, Loss: {'train/loss': 0.16064371168613434, 'train/grad_norm_pre_clip': 4.96071195602417, 'timing/data': 0.024054545909166336, 'timing/model': 2.7507556099444628, 'learning_rate/qwen_vl_interface': 9.586031424569882e-06, 'learning_rate/action_model': 9.58603142456988e-05, 'epoch': 1.55}) +[2026-06-19 19:40:58,426][starVLA.training.train_starvla][INFO] - Step 757, Loss: {'train/loss': 0.3680846691131592, 'train/grad_norm_pre_clip': 11.521751403808594, 'timing/data': 0.022500567138195038, 'timing/model': 2.761379161849618, 'learning_rate/qwen_vl_interface': 9.584787094143265e-06, 'learning_rate/action_model': 9.584787094143264e-05, 'epoch': 1.55}) +[2026-06-19 19:41:09,223][starVLA.training.train_starvla][INFO] - Step 758, Loss: {'train/loss': 0.1612005978822708, 'train/grad_norm_pre_clip': 5.882193565368652, 'timing/data': 0.02351614460349083, 'timing/model': 2.754492772743106, 'learning_rate/qwen_vl_interface': 9.58354098184941e-06, 'learning_rate/action_model': 9.583540981849409e-05, 'epoch': 1.56}) +[2026-06-19 19:41:20,028][starVLA.training.train_starvla][INFO] - Step 759, Loss: {'train/loss': 0.37806808948516846, 'train/grad_norm_pre_clip': 9.217093467712402, 'timing/data': 0.0228252150118351, 'timing/model': 2.7655201479792595, 'learning_rate/qwen_vl_interface': 9.582293088200549e-06, 'learning_rate/action_model': 9.582293088200549e-05, 'epoch': 1.56}) +[2026-06-19 19:41:30,922][starVLA.training.train_starvla][INFO] - Step 760, Loss: {'train/loss': 0.19558170437812805, 'train/grad_norm_pre_clip': 6.344268321990967, 'timing/data': 0.02348385564982891, 'timing/model': 2.7667607627809048, 'learning_rate/qwen_vl_interface': 9.581043413709646e-06, 'learning_rate/action_model': 9.581043413709646e-05, 'epoch': 1.56}) +[2026-06-19 19:41:41,711][starVLA.training.train_starvla][INFO] - Step 761, Loss: {'train/loss': 0.13223735988140106, 'train/grad_norm_pre_clip': 38.247657775878906, 'timing/data': 0.03005925565958023, 'timing/model': 2.739907294511795, 'learning_rate/qwen_vl_interface': 9.579791958890393e-06, 'learning_rate/action_model': 9.579791958890394e-05, 'epoch': 1.56}) +[2026-06-19 19:41:52,514][starVLA.training.train_starvla][INFO] - Step 762, Loss: {'train/loss': 0.05642542243003845, 'train/grad_norm_pre_clip': 2.47847318649292, 'timing/data': 0.02396424673497677, 'timing/model': 2.760804170742631, 'learning_rate/qwen_vl_interface': 9.578538724257218e-06, 'learning_rate/action_model': 9.578538724257216e-05, 'epoch': 1.57}) +[2026-06-19 19:42:03,314][starVLA.training.train_starvla][INFO] - Step 763, Loss: {'train/loss': 0.19885671138763428, 'train/grad_norm_pre_clip': 8.026532173156738, 'timing/data': 0.023541828617453575, 'timing/model': 2.7541248705238104, 'learning_rate/qwen_vl_interface': 9.577283710325274e-06, 'learning_rate/action_model': 9.577283710325274e-05, 'epoch': 1.57}) +[2026-06-19 19:42:13,928][starVLA.training.train_starvla][INFO] - Step 764, Loss: {'train/loss': 0.18759426474571228, 'train/grad_norm_pre_clip': 6.194331645965576, 'timing/data': 0.02383839339017868, 'timing/model': 2.7117466051131487, 'learning_rate/qwen_vl_interface': 9.576026917610455e-06, 'learning_rate/action_model': 9.576026917610455e-05, 'epoch': 1.57}) +[2026-06-19 19:42:24,818][starVLA.training.train_starvla][INFO] - Step 765, Loss: {'train/loss': 0.20813551545143127, 'train/grad_norm_pre_clip': 14.352299690246582, 'timing/data': 0.023712266236543655, 'timing/model': 2.75390007160604, 'learning_rate/qwen_vl_interface': 9.574768346629376e-06, 'learning_rate/action_model': 9.574768346629376e-05, 'epoch': 1.57}) +[2026-06-19 19:42:35,512][starVLA.training.train_starvla][INFO] - Step 766, Loss: {'train/loss': 0.2400321364402771, 'train/grad_norm_pre_clip': 10.786895751953125, 'timing/data': 0.023817984387278557, 'timing/model': 2.743711454793811, 'learning_rate/qwen_vl_interface': 9.573507997899391e-06, 'learning_rate/action_model': 9.573507997899392e-05, 'epoch': 1.57}) +[2026-06-19 19:42:46,201][starVLA.training.train_starvla][INFO] - Step 767, Loss: {'train/loss': 0.14683328568935394, 'train/grad_norm_pre_clip': 8.413485527038574, 'timing/data': 0.023410677909851074, 'timing/model': 2.7205981593579054, 'learning_rate/qwen_vl_interface': 9.572245871938582e-06, 'learning_rate/action_model': 9.572245871938582e-05, 'epoch': 1.58}) +[2026-06-19 19:42:56,922][starVLA.training.train_starvla][INFO] - Step 768, Loss: {'train/loss': 0.24672269821166992, 'train/grad_norm_pre_clip': 12.508917808532715, 'timing/data': 0.023888753727078438, 'timing/model': 2.7682319581508636, 'learning_rate/qwen_vl_interface': 9.570981969265761e-06, 'learning_rate/action_model': 9.570981969265761e-05, 'epoch': 1.58}) +[2026-06-19 19:43:07,724][starVLA.training.train_starvla][INFO] - Step 769, Loss: {'train/loss': 0.388606458902359, 'train/grad_norm_pre_clip': 9.348607063293457, 'timing/data': 0.0240864809602499, 'timing/model': 2.75707995146513, 'learning_rate/qwen_vl_interface': 9.56971629040047e-06, 'learning_rate/action_model': 9.56971629040047e-05, 'epoch': 1.58}) +[2026-06-19 19:43:18,529][starVLA.training.train_starvla][INFO] - Step 770, Loss: {'train/loss': 0.2654132544994354, 'train/grad_norm_pre_clip': 9.552607536315918, 'timing/data': 0.024264907464385033, 'timing/model': 2.764968778938055, 'learning_rate/qwen_vl_interface': 9.56844883586298e-06, 'learning_rate/action_model': 9.568448835862981e-05, 'epoch': 1.58}) +[2026-06-19 19:43:29,324][starVLA.training.train_starvla][INFO] - Step 771, Loss: {'train/loss': 0.1426013857126236, 'train/grad_norm_pre_clip': 6.838987827301025, 'timing/data': 0.02343771979212761, 'timing/model': 2.7637051045894623, 'learning_rate/qwen_vl_interface': 9.567179606174299e-06, 'learning_rate/action_model': 9.567179606174299e-05, 'epoch': 1.58}) +[2026-06-19 19:43:40,225][starVLA.training.train_starvla][INFO] - Step 772, Loss: {'train/loss': 0.12828142940998077, 'train/grad_norm_pre_clip': 5.929286003112793, 'timing/data': 0.02300342358648777, 'timing/model': 2.7591252271085978, 'learning_rate/qwen_vl_interface': 9.565908601856155e-06, 'learning_rate/action_model': 9.565908601856154e-05, 'epoch': 1.59}) +[2026-06-19 19:43:51,021][starVLA.training.train_starvla][INFO] - Step 773, Loss: {'train/loss': 0.22320687770843506, 'train/grad_norm_pre_clip': 5.7223029136657715, 'timing/data': 0.023042574524879456, 'timing/model': 2.7754053436219692, 'learning_rate/qwen_vl_interface': 9.56463582343101e-06, 'learning_rate/action_model': 9.56463582343101e-05, 'epoch': 1.59}) +[2026-06-19 19:44:01,823][starVLA.training.train_starvla][INFO] - Step 774, Loss: {'train/loss': 0.19481724500656128, 'train/grad_norm_pre_clip': 7.099625110626221, 'timing/data': 0.023406660184264183, 'timing/model': 2.7565414365381002, 'learning_rate/qwen_vl_interface': 9.563361271422056e-06, 'learning_rate/action_model': 9.563361271422056e-05, 'epoch': 1.59}) +[2026-06-19 19:44:12,628][starVLA.training.train_starvla][INFO] - Step 775, Loss: {'train/loss': 0.40953749418258667, 'train/grad_norm_pre_clip': 9.253047943115234, 'timing/data': 0.02303512580692768, 'timing/model': 2.762664707377553, 'learning_rate/qwen_vl_interface': 9.562084946353212e-06, 'learning_rate/action_model': 9.562084946353213e-05, 'epoch': 1.59}) +[2026-06-19 19:44:23,422][starVLA.training.train_starvla][INFO] - Step 776, Loss: {'train/loss': 0.21280036866664886, 'train/grad_norm_pre_clip': 5.722093105316162, 'timing/data': 0.023683270439505577, 'timing/model': 2.7584147844463587, 'learning_rate/qwen_vl_interface': 9.560806848749132e-06, 'learning_rate/action_model': 9.560806848749132e-05, 'epoch': 1.59}) +[2026-06-19 19:44:34,220][starVLA.training.train_starvla][INFO] - Step 777, Loss: {'train/loss': 0.11163254827260971, 'train/grad_norm_pre_clip': 3.0844035148620605, 'timing/data': 0.024054337292909622, 'timing/model': 2.7550156116485596, 'learning_rate/qwen_vl_interface': 9.559526979135186e-06, 'learning_rate/action_model': 9.559526979135187e-05, 'epoch': 1.6}) +[2026-06-19 19:44:45,020][starVLA.training.train_starvla][INFO] - Step 778, Loss: {'train/loss': 0.2552483081817627, 'train/grad_norm_pre_clip': 8.199423789978027, 'timing/data': 0.023871086537837982, 'timing/model': 2.7529168892651796, 'learning_rate/qwen_vl_interface': 9.558245338037486e-06, 'learning_rate/action_model': 9.558245338037486e-05, 'epoch': 1.6}) +[2026-06-19 19:44:55,823][starVLA.training.train_starvla][INFO] - Step 779, Loss: {'train/loss': 0.10213737934827805, 'train/grad_norm_pre_clip': 3.7555365562438965, 'timing/data': 0.02368753030896187, 'timing/model': 2.7527167443186045, 'learning_rate/qwen_vl_interface': 9.556961925982864e-06, 'learning_rate/action_model': 9.556961925982863e-05, 'epoch': 1.6}) +[2026-06-19 19:45:06,628][starVLA.training.train_starvla][INFO] - Step 780, Loss: {'train/loss': 0.41407153010368347, 'train/grad_norm_pre_clip': 11.997309684753418, 'timing/data': 0.02301339991390705, 'timing/model': 2.7646741550415754, 'learning_rate/qwen_vl_interface': 9.555676743498882e-06, 'learning_rate/action_model': 9.555676743498881e-05, 'epoch': 1.6}) +[2026-06-19 19:45:17,521][starVLA.training.train_starvla][INFO] - Step 781, Loss: {'train/loss': 0.2434287667274475, 'train/grad_norm_pre_clip': 8.384162902832031, 'timing/data': 0.02274647355079651, 'timing/model': 2.757335903123021, 'learning_rate/qwen_vl_interface': 9.55438979111383e-06, 'learning_rate/action_model': 9.554389791113831e-05, 'epoch': 1.6}) +[2026-06-19 19:45:28,328][starVLA.training.train_starvla][INFO] - Step 782, Loss: {'train/loss': 0.12879052758216858, 'train/grad_norm_pre_clip': 9.129517555236816, 'timing/data': 0.023902008309960365, 'timing/model': 2.764198863878846, 'learning_rate/qwen_vl_interface': 9.553101069356728e-06, 'learning_rate/action_model': 9.553101069356727e-05, 'epoch': 1.61}) +[2026-06-19 19:45:39,123][starVLA.training.train_starvla][INFO] - Step 783, Loss: {'train/loss': 0.16659927368164062, 'train/grad_norm_pre_clip': 4.400563716888428, 'timing/data': 0.023216238245368004, 'timing/model': 2.758644862100482, 'learning_rate/qwen_vl_interface': 9.551810578757316e-06, 'learning_rate/action_model': 9.551810578757315e-05, 'epoch': 1.61}) +[2026-06-19 19:45:49,923][starVLA.training.train_starvla][INFO] - Step 784, Loss: {'train/loss': 0.1845281720161438, 'train/grad_norm_pre_clip': 4.891262531280518, 'timing/data': 0.023713761940598488, 'timing/model': 2.7522959765046835, 'learning_rate/qwen_vl_interface': 9.550518319846069e-06, 'learning_rate/action_model': 9.550518319846069e-05, 'epoch': 1.61}) +[2026-06-19 19:46:00,818][starVLA.training.train_starvla][INFO] - Step 785, Loss: {'train/loss': 0.198032945394516, 'train/grad_norm_pre_clip': 7.022421360015869, 'timing/data': 0.023718303069472313, 'timing/model': 2.750729875639081, 'learning_rate/qwen_vl_interface': 9.549224293154187e-06, 'learning_rate/action_model': 9.549224293154187e-05, 'epoch': 1.61}) +[2026-06-19 19:46:11,624][starVLA.training.train_starvla][INFO] - Step 786, Loss: {'train/loss': 0.1512417495250702, 'train/grad_norm_pre_clip': 7.3186936378479, 'timing/data': 0.023566143587231636, 'timing/model': 2.75427214987576, 'learning_rate/qwen_vl_interface': 9.547928499213593e-06, 'learning_rate/action_model': 9.547928499213593e-05, 'epoch': 1.61}) +[2026-06-19 19:46:22,429][starVLA.training.train_starvla][INFO] - Step 787, Loss: {'train/loss': 0.2988007664680481, 'train/grad_norm_pre_clip': 8.489110946655273, 'timing/data': 0.022952506318688393, 'timing/model': 2.7621151600033045, 'learning_rate/qwen_vl_interface': 9.54663093855694e-06, 'learning_rate/action_model': 9.54663093855694e-05, 'epoch': 1.62}) +[2026-06-19 19:46:33,323][starVLA.training.train_starvla][INFO] - Step 788, Loss: {'train/loss': 0.19120244681835175, 'train/grad_norm_pre_clip': 6.685885906219482, 'timing/data': 0.022561945021152496, 'timing/model': 2.750538121908903, 'learning_rate/qwen_vl_interface': 9.545331611717607e-06, 'learning_rate/action_model': 9.545331611717608e-05, 'epoch': 1.62}) +[2026-06-19 19:46:44,225][starVLA.training.train_starvla][INFO] - Step 789, Loss: {'train/loss': 0.10213574022054672, 'train/grad_norm_pre_clip': 4.210313320159912, 'timing/data': 0.023702632635831833, 'timing/model': 2.7600769083946943, 'learning_rate/qwen_vl_interface': 9.544030519229697e-06, 'learning_rate/action_model': 9.544030519229697e-05, 'epoch': 1.62}) +[2026-06-19 19:46:55,027][starVLA.training.train_starvla][INFO] - Step 790, Loss: {'train/loss': 0.28876835107803345, 'train/grad_norm_pre_clip': 15.201372146606445, 'timing/data': 0.025313831865787506, 'timing/model': 2.7578150890767574, 'learning_rate/qwen_vl_interface': 9.54272766162804e-06, 'learning_rate/action_model': 9.54272766162804e-05, 'epoch': 1.62}) +[2026-06-19 19:47:05,946][starVLA.training.train_starvla][INFO] - Step 791, Loss: {'train/loss': 0.15684542059898376, 'train/grad_norm_pre_clip': 5.9530792236328125, 'timing/data': 0.022540409117937088, 'timing/model': 2.7801859956234694, 'learning_rate/qwen_vl_interface': 9.54142303944819e-06, 'learning_rate/action_model': 9.541423039448191e-05, 'epoch': 1.62}) +[2026-06-19 19:47:16,524][starVLA.training.train_starvla][INFO] - Step 792, Loss: {'train/loss': 0.07184617966413498, 'train/grad_norm_pre_clip': 3.0759522914886475, 'timing/data': 0.02440071478486061, 'timing/model': 2.740175446495414, 'learning_rate/qwen_vl_interface': 9.540116653226433e-06, 'learning_rate/action_model': 9.540116653226431e-05, 'epoch': 1.63}) +[2026-06-19 19:47:27,206][starVLA.training.train_starvla][INFO] - Step 793, Loss: {'train/loss': 0.09987396746873856, 'train/grad_norm_pre_clip': 3.2601780891418457, 'timing/data': 0.024365674704313278, 'timing/model': 2.7219447363168, 'learning_rate/qwen_vl_interface': 9.53880850349977e-06, 'learning_rate/action_model': 9.538808503499768e-05, 'epoch': 1.63}) +[2026-06-19 19:47:37,715][starVLA.training.train_starvla][INFO] - Step 794, Loss: {'train/loss': 0.18149879574775696, 'train/grad_norm_pre_clip': 8.618123054504395, 'timing/data': 0.02275681309401989, 'timing/model': 2.7212968971580267, 'learning_rate/qwen_vl_interface': 9.537498590805934e-06, 'learning_rate/action_model': 9.537498590805933e-05, 'epoch': 1.63}) +[2026-06-19 19:47:48,521][starVLA.training.train_starvla][INFO] - Step 795, Loss: {'train/loss': 0.25716790556907654, 'train/grad_norm_pre_clip': 10.600346565246582, 'timing/data': 0.023040011525154114, 'timing/model': 2.759790102019906, 'learning_rate/qwen_vl_interface': 9.536186915683379e-06, 'learning_rate/action_model': 9.53618691568338e-05, 'epoch': 1.63}) +[2026-06-19 19:47:59,200][starVLA.training.train_starvla][INFO] - Step 796, Loss: {'train/loss': 0.20866982638835907, 'train/grad_norm_pre_clip': 7.5046610832214355, 'timing/data': 0.02350170724093914, 'timing/model': 2.749783020466566, 'learning_rate/qwen_vl_interface': 9.534873478671289e-06, 'learning_rate/action_model': 9.534873478671288e-05, 'epoch': 1.63}) +[2026-06-19 19:48:10,021][starVLA.training.train_starvla][INFO] - Step 797, Loss: {'train/loss': 0.33376580476760864, 'train/grad_norm_pre_clip': 9.388632774353027, 'timing/data': 0.02347295545041561, 'timing/model': 2.7569283805787563, 'learning_rate/qwen_vl_interface': 9.533558280309563e-06, 'learning_rate/action_model': 9.533558280309563e-05, 'epoch': 1.64}) +[2026-06-19 19:48:20,716][starVLA.training.train_starvla][INFO] - Step 798, Loss: {'train/loss': 0.18889881670475006, 'train/grad_norm_pre_clip': 5.026923179626465, 'timing/data': 0.022383350878953934, 'timing/model': 2.739168170839548, 'learning_rate/qwen_vl_interface': 9.532241321138832e-06, 'learning_rate/action_model': 9.532241321138833e-05, 'epoch': 1.64}) +[2026-06-19 19:48:31,523][starVLA.training.train_starvla][INFO] - Step 799, Loss: {'train/loss': 0.19068942964076996, 'train/grad_norm_pre_clip': 7.652830123901367, 'timing/data': 0.022938821464776993, 'timing/model': 2.7633868493139744, 'learning_rate/qwen_vl_interface': 9.530922601700451e-06, 'learning_rate/action_model': 9.530922601700452e-05, 'epoch': 1.64}) +[2026-06-19 19:48:42,323][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 800 +[2026-06-19 19:56:04,971][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 800 +[2026-06-19 19:56:04,972][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 800 +[2026-06-19 20:04:02,583][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 800 +[2026-06-19 20:04:02,585][starVLA.training.train_starvla][INFO] - Step 800, Loss: {'train/loss': 0.21609947085380554, 'train/grad_norm_pre_clip': 6.463833332061768, 'eval/loss': 0.45406338572502136, 'eval/latency_0/loss': 0.08855976909399033, 'eval/latency_1/loss': 0.27098342776298523, 'eval/latency_2/loss': 0.6330444812774658, 'eval/latency_3/loss': 0.49481356143951416, 'eval/latency_4/loss': 0.753740668296814, 'eval/flappy/loss': 0.45406338572502136, 'eval/flappy/latency_0/loss': 0.08855976909399033, 'eval/flappy/latency_1/loss': 0.27098342776298523, 'eval/flappy/latency_2/loss': 0.6330444812774658, 'eval/flappy/latency_3/loss': 0.49481356143951416, 'eval/flappy/latency_4/loss': 0.753740668296814, 'eval/flap_precision': 0.6666666666666666, 'eval/flap_recall': 0.8, 'eval/flap_f1': 0.7272727272727272, 'eval/flap_rate_model': 0.09166666666666666, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.7272727272727272, 'eval/latency_0/flap_precision': 0.6666666666666666, 'eval/latency_0/flap_recall': 0.8, 'eval/latency_0/flap_f1': 0.7272727272727272, 'eval/latency_0/flap_rate_model': 0.09166666666666666, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.7272727272727272, 'timing/data': 0.022427691146731377, 'timing/model': 2.7597809433937073, 'learning_rate/qwen_vl_interface': 9.52960212253649e-06, 'learning_rate/action_model': 9.529602122536491e-05, 'epoch': 1.64}) +[2026-06-19 20:04:02,585][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_800_state +[2026-06-19 20:04:02,605][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-19 20:04:07,974][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_800_state/model.safetensors +[2026-06-19 20:04:15,140][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_800_state/optimizer.bin +[2026-06-19 20:04:15,141][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_800_state/sampler.bin +[2026-06-19 20:04:15,141][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_800_state/sampler_1.bin +[2026-06-19 20:04:15,142][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_800_state/random_states_0.pkl +[2026-06-19 20:04:15,369][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 800 +[2026-06-19 20:04:15,369][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-19 20:04:15,383][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-19 20:04:26,098][starVLA.training.train_starvla][INFO] - Step 801, Loss: {'train/loss': 0.10086522996425629, 'train/grad_norm_pre_clip': 3.3820366859436035, 'timing/data': 0.02395695634186268, 'timing/model': 2.7428292371332645, 'learning_rate/qwen_vl_interface': 9.528279884189754e-06, 'learning_rate/action_model': 9.528279884189754e-05, 'epoch': 1.65}) +[2026-06-19 20:04:36,809][starVLA.training.train_starvla][INFO] - Step 802, Loss: {'train/loss': 0.2603665292263031, 'train/grad_norm_pre_clip': 11.445144653320312, 'timing/data': 0.02278195135295391, 'timing/model': 2.7613572478294373, 'learning_rate/qwen_vl_interface': 9.526955887203762e-06, 'learning_rate/action_model': 9.526955887203761e-05, 'epoch': 1.65}) +[2026-06-19 20:04:47,508][starVLA.training.train_starvla][INFO] - Step 803, Loss: {'train/loss': 0.344818651676178, 'train/grad_norm_pre_clip': 6.566898345947266, 'timing/data': 0.02314615808427334, 'timing/model': 2.705281114205718, 'learning_rate/qwen_vl_interface': 9.525630132122758e-06, 'learning_rate/action_model': 9.525630132122758e-05, 'epoch': 1.65}) +[2026-06-19 20:04:58,105][starVLA.training.train_starvla][INFO] - Step 804, Loss: {'train/loss': 0.09518200159072876, 'train/grad_norm_pre_clip': 2.7018020153045654, 'timing/data': 0.021595316007733345, 'timing/model': 2.7522868551313877, 'learning_rate/qwen_vl_interface': 9.524302619491712e-06, 'learning_rate/action_model': 9.524302619491712e-05, 'epoch': 1.65}) +[2026-06-19 20:05:08,609][starVLA.training.train_starvla][INFO] - Step 805, Loss: {'train/loss': 0.2344181090593338, 'train/grad_norm_pre_clip': 4.307932376861572, 'timing/data': 0.021107342094182968, 'timing/model': 2.73854704387486, 'learning_rate/qwen_vl_interface': 9.522973349856315e-06, 'learning_rate/action_model': 9.522973349856315e-05, 'epoch': 1.65}) +[2026-06-19 20:05:19,219][starVLA.training.train_starvla][INFO] - Step 806, Loss: {'train/loss': 0.2806694209575653, 'train/grad_norm_pre_clip': 9.381634712219238, 'timing/data': 0.02460792288184166, 'timing/model': 2.76838356629014, 'learning_rate/qwen_vl_interface': 9.521642323762976e-06, 'learning_rate/action_model': 9.521642323762976e-05, 'epoch': 1.66}) +[2026-06-19 20:05:29,735][starVLA.training.train_starvla][INFO] - Step 807, Loss: {'train/loss': 0.3155812621116638, 'train/grad_norm_pre_clip': 13.028937339782715, 'timing/data': 0.02240057662129402, 'timing/model': 2.6400401033461094, 'learning_rate/qwen_vl_interface': 9.520309541758833e-06, 'learning_rate/action_model': 9.520309541758833e-05, 'epoch': 1.66}) +[2026-06-19 20:05:40,433][starVLA.training.train_starvla][INFO] - Step 808, Loss: {'train/loss': 0.22789297997951508, 'train/grad_norm_pre_clip': 6.7308807373046875, 'timing/data': 0.028313687071204185, 'timing/model': 2.627854235470295, 'learning_rate/qwen_vl_interface': 9.51897500439174e-06, 'learning_rate/action_model': 9.51897500439174e-05, 'epoch': 1.66}) +[2026-06-19 20:05:51,125][starVLA.training.train_starvla][INFO] - Step 809, Loss: {'train/loss': 0.12797383964061737, 'train/grad_norm_pre_clip': 3.728128433227539, 'timing/data': 0.02475271001458168, 'timing/model': 2.7071969714015722, 'learning_rate/qwen_vl_interface': 9.517638712210277e-06, 'learning_rate/action_model': 9.517638712210276e-05, 'epoch': 1.66}) +[2026-06-19 20:06:01,725][starVLA.training.train_starvla][INFO] - Step 810, Loss: {'train/loss': 0.13070406019687653, 'train/grad_norm_pre_clip': 5.521783828735352, 'timing/data': 0.023953018710017204, 'timing/model': 2.702673574909568, 'learning_rate/qwen_vl_interface': 9.51630066576374e-06, 'learning_rate/action_model': 9.51630066576374e-05, 'epoch': 1.66}) +[2026-06-19 20:06:12,382][starVLA.training.train_starvla][INFO] - Step 811, Loss: {'train/loss': 0.26217347383499146, 'train/grad_norm_pre_clip': 7.518341064453125, 'timing/data': 0.022085731849074364, 'timing/model': 2.7321597021073103, 'learning_rate/qwen_vl_interface': 9.514960865602154e-06, 'learning_rate/action_model': 9.514960865602153e-05, 'epoch': 1.67}) +[2026-06-19 20:06:23,129][starVLA.training.train_starvla][INFO] - Step 812, Loss: {'train/loss': 0.14441174268722534, 'train/grad_norm_pre_clip': 5.46359920501709, 'timing/data': 0.025329511612653732, 'timing/model': 2.7636701110750437, 'learning_rate/qwen_vl_interface': 9.513619312276259e-06, 'learning_rate/action_model': 9.513619312276258e-05, 'epoch': 1.67}) +[2026-06-19 20:06:34,027][starVLA.training.train_starvla][INFO] - Step 813, Loss: {'train/loss': 0.26472094655036926, 'train/grad_norm_pre_clip': 6.639620304107666, 'timing/data': 0.023731021210551262, 'timing/model': 2.760912384837866, 'learning_rate/qwen_vl_interface': 9.512276006337514e-06, 'learning_rate/action_model': 9.512276006337513e-05, 'epoch': 1.67}) +[2026-06-19 20:06:44,820][starVLA.training.train_starvla][INFO] - Step 814, Loss: {'train/loss': 0.18452724814414978, 'train/grad_norm_pre_clip': 6.3259382247924805, 'timing/data': 0.02252921462059021, 'timing/model': 2.7820269279181957, 'learning_rate/qwen_vl_interface': 9.510930948338102e-06, 'learning_rate/action_model': 9.510930948338102e-05, 'epoch': 1.67}) +[2026-06-19 20:06:55,818][starVLA.training.train_starvla][INFO] - Step 815, Loss: {'train/loss': 0.08988868445158005, 'train/grad_norm_pre_clip': 4.61065673828125, 'timing/data': 0.021896863356232643, 'timing/model': 2.7591190692037344, 'learning_rate/qwen_vl_interface': 9.509584138830929e-06, 'learning_rate/action_model': 9.509584138830928e-05, 'epoch': 1.67}) +[2026-06-19 20:07:06,617][starVLA.training.train_starvla][INFO] - Step 816, Loss: {'train/loss': 0.1897294670343399, 'train/grad_norm_pre_clip': 11.66673755645752, 'timing/data': 0.02189680188894272, 'timing/model': 2.760356016457081, 'learning_rate/qwen_vl_interface': 9.508235578369613e-06, 'learning_rate/action_model': 9.508235578369612e-05, 'epoch': 1.68}) +[2026-06-19 20:07:17,317][starVLA.training.train_starvla][INFO] - Step 817, Loss: {'train/loss': 0.33195823431015015, 'train/grad_norm_pre_clip': 13.658226013183594, 'timing/data': 0.020823556929826736, 'timing/model': 2.741885369643569, 'learning_rate/qwen_vl_interface': 9.506885267508498e-06, 'learning_rate/action_model': 9.506885267508498e-05, 'epoch': 1.68}) +[2026-06-19 20:07:28,187][starVLA.training.train_starvla][INFO] - Step 818, Loss: {'train/loss': 0.28400689363479614, 'train/grad_norm_pre_clip': 10.533695220947266, 'timing/data': 0.021370507776737213, 'timing/model': 2.830861544236541, 'learning_rate/qwen_vl_interface': 9.505533206802648e-06, 'learning_rate/action_model': 9.505533206802647e-05, 'epoch': 1.68}) +[2026-06-19 20:07:38,920][starVLA.training.train_starvla][INFO] - Step 819, Loss: {'train/loss': 0.14429159462451935, 'train/grad_norm_pre_clip': 6.872230052947998, 'timing/data': 0.023262036964297295, 'timing/model': 2.758209766820073, 'learning_rate/qwen_vl_interface': 9.50417939680784e-06, 'learning_rate/action_model': 9.504179396807839e-05, 'epoch': 1.68}) +[2026-06-19 20:07:49,714][starVLA.training.train_starvla][INFO] - Step 820, Loss: {'train/loss': 0.09768955409526825, 'train/grad_norm_pre_clip': 5.580332279205322, 'timing/data': 0.02097926288843155, 'timing/model': 2.755199994891882, 'learning_rate/qwen_vl_interface': 9.502823838080578e-06, 'learning_rate/action_model': 9.502823838080578e-05, 'epoch': 1.68}) +[2026-06-19 20:08:00,512][starVLA.training.train_starvla][INFO] - Step 821, Loss: {'train/loss': 0.09464909881353378, 'train/grad_norm_pre_clip': 6.477362155914307, 'timing/data': 0.02077576518058777, 'timing/model': 2.7555271331220865, 'learning_rate/qwen_vl_interface': 9.501466531178079e-06, 'learning_rate/action_model': 9.501466531178079e-05, 'epoch': 1.69}) +[2026-06-19 20:08:11,553][starVLA.training.train_starvla][INFO] - Step 822, Loss: {'train/loss': 0.20930705964565277, 'train/grad_norm_pre_clip': 19.8013916015625, 'timing/data': 0.020740805193781853, 'timing/model': 2.9982025045901537, 'learning_rate/qwen_vl_interface': 9.50010747665828e-06, 'learning_rate/action_model': 9.500107476658281e-05, 'epoch': 1.69}) +[2026-06-19 20:08:22,318][starVLA.training.train_starvla][INFO] - Step 823, Loss: {'train/loss': 0.1540071964263916, 'train/grad_norm_pre_clip': 8.381778717041016, 'timing/data': 0.020793205127120018, 'timing/model': 2.7677690908312798, 'learning_rate/qwen_vl_interface': 9.498746675079842e-06, 'learning_rate/action_model': 9.49874667507984e-05, 'epoch': 1.69}) +[2026-06-19 20:08:33,120][starVLA.training.train_starvla][INFO] - Step 824, Loss: {'train/loss': 0.18683305382728577, 'train/grad_norm_pre_clip': 8.545365333557129, 'timing/data': 0.02073005586862564, 'timing/model': 2.7589147817343473, 'learning_rate/qwen_vl_interface': 9.49738412700213e-06, 'learning_rate/action_model': 9.49738412700213e-05, 'epoch': 1.69}) +[2026-06-19 20:08:43,815][starVLA.training.train_starvla][INFO] - Step 825, Loss: {'train/loss': 0.22749504446983337, 'train/grad_norm_pre_clip': 12.767333030700684, 'timing/data': 0.021451296284794807, 'timing/model': 2.735170552507043, 'learning_rate/qwen_vl_interface': 9.496019832985243e-06, 'learning_rate/action_model': 9.496019832985243e-05, 'epoch': 1.69}) +[2026-06-19 20:08:54,812][starVLA.training.train_starvla][INFO] - Step 826, Loss: {'train/loss': 0.4013521075248718, 'train/grad_norm_pre_clip': 16.05940818786621, 'timing/data': 0.02184395305812359, 'timing/model': 2.7299076337367296, 'learning_rate/qwen_vl_interface': 9.49465379358999e-06, 'learning_rate/action_model': 9.49465379358999e-05, 'epoch': 1.7}) +[2026-06-19 20:09:05,519][starVLA.training.train_starvla][INFO] - Step 827, Loss: {'train/loss': 0.19589661061763763, 'train/grad_norm_pre_clip': 12.691173553466797, 'timing/data': 0.021755054593086243, 'timing/model': 2.7445508297532797, 'learning_rate/qwen_vl_interface': 9.493286009377896e-06, 'learning_rate/action_model': 9.493286009377896e-05, 'epoch': 1.7}) +[2026-06-19 20:09:16,217][starVLA.training.train_starvla][INFO] - Step 828, Loss: {'train/loss': 0.15060749650001526, 'train/grad_norm_pre_clip': 9.346224784851074, 'timing/data': 0.021746356040239334, 'timing/model': 2.752116745337844, 'learning_rate/qwen_vl_interface': 9.491916480911208e-06, 'learning_rate/action_model': 9.491916480911207e-05, 'epoch': 1.7}) +[2026-06-19 20:09:26,919][starVLA.training.train_starvla][INFO] - Step 829, Loss: {'train/loss': 0.18024137616157532, 'train/grad_norm_pre_clip': 8.216164588928223, 'timing/data': 0.023220518603920937, 'timing/model': 2.745752388611436, 'learning_rate/qwen_vl_interface': 9.490545208752884e-06, 'learning_rate/action_model': 9.490545208752883e-05, 'epoch': 1.7}) +[2026-06-19 20:09:37,811][starVLA.training.train_starvla][INFO] - Step 830, Loss: {'train/loss': 0.10714339464902878, 'train/grad_norm_pre_clip': 4.3153605461120605, 'timing/data': 0.021889662370085716, 'timing/model': 2.7497165873646736, 'learning_rate/qwen_vl_interface': 9.489172193466603e-06, 'learning_rate/action_model': 9.489172193466603e-05, 'epoch': 1.7}) +[2026-06-19 20:09:48,617][starVLA.training.train_starvla][INFO] - Step 831, Loss: {'train/loss': 0.24281717836856842, 'train/grad_norm_pre_clip': 9.591702461242676, 'timing/data': 0.02163584716618061, 'timing/model': 2.7696299832314253, 'learning_rate/qwen_vl_interface': 9.487797435616762e-06, 'learning_rate/action_model': 9.487797435616761e-05, 'epoch': 1.71}) +[2026-06-19 20:09:59,421][starVLA.training.train_starvla][INFO] - Step 832, Loss: {'train/loss': 0.27874377369880676, 'train/grad_norm_pre_clip': 7.568379878997803, 'timing/data': 0.021657206118106842, 'timing/model': 2.7627150155603886, 'learning_rate/qwen_vl_interface': 9.486420935768472e-06, 'learning_rate/action_model': 9.486420935768472e-05, 'epoch': 1.71}) +[2026-06-19 20:10:10,124][starVLA.training.train_starvla][INFO] - Step 833, Loss: {'train/loss': 0.1946297436952591, 'train/grad_norm_pre_clip': 7.816272735595703, 'timing/data': 0.022054022178053856, 'timing/model': 2.7475689463317394, 'learning_rate/qwen_vl_interface': 9.485042694487558e-06, 'learning_rate/action_model': 9.485042694487559e-05, 'epoch': 1.71}) +[2026-06-19 20:10:22,442][starVLA.training.train_starvla][INFO] - Latency curriculum phase 1 at step 834: active_latencies=[1], dataset_size=66043 +[2026-06-19 20:10:22,444][starVLA.training.train_starvla][INFO] - Step 834, Loss: {'train/loss': 0.33146676421165466, 'train/grad_norm_pre_clip': 8.970559120178223, 'latency_curriculum/phase': 1.0, 'latency_curriculum/active_count': 1.0, 'latency_curriculum/max_active_latency': 1.0, 'latency_curriculum/dataset_size': 66043.0, 'timing/data': 0.021510278806090355, 'timing/model': 2.7606374006718397, 'learning_rate/qwen_vl_interface': 9.483662712340565e-06, 'learning_rate/action_model': 9.483662712340564e-05, 'epoch': 1.62}) +[2026-06-19 20:10:54,186][starVLA.training.train_starvla][INFO] - Step 835, Loss: {'train/loss': 0.3744208514690399, 'train/grad_norm_pre_clip': 8.866756439208984, 'timing/data': 0.006402086466550827, 'timing/model': 2.748621951788664, 'learning_rate/qwen_vl_interface': 9.482280989894748e-06, 'learning_rate/action_model': 9.482280989894748e-05, 'epoch': 1.62}) +[2026-06-19 20:11:04,737][starVLA.training.train_starvla][INFO] - Step 836, Loss: {'train/loss': 0.8382613658905029, 'train/grad_norm_pre_clip': 17.592622756958008, 'timing/data': 0.0060623325407505035, 'timing/model': 2.6788589898496866, 'learning_rate/qwen_vl_interface': 9.480897527718083e-06, 'learning_rate/action_model': 9.480897527718084e-05, 'epoch': 1.62}) +[2026-06-19 20:11:15,508][starVLA.training.train_starvla][INFO] - Step 837, Loss: {'train/loss': 0.5357578992843628, 'train/grad_norm_pre_clip': 12.664338111877441, 'timing/data': 0.00663948617875576, 'timing/model': 2.85344997048378, 'learning_rate/qwen_vl_interface': 9.479512326379263e-06, 'learning_rate/action_model': 9.479512326379264e-05, 'epoch': 1.62}) +[2026-06-19 20:11:26,113][starVLA.training.train_starvla][INFO] - Step 838, Loss: {'train/loss': 0.38644421100616455, 'train/grad_norm_pre_clip': 6.142669200897217, 'timing/data': 0.008815927430987358, 'timing/model': 2.747267484664917, 'learning_rate/qwen_vl_interface': 9.478125386447686e-06, 'learning_rate/action_model': 9.478125386447686e-05, 'epoch': 1.62}) +[2026-06-19 20:11:36,738][starVLA.training.train_starvla][INFO] - Step 839, Loss: {'train/loss': 0.3755718767642975, 'train/grad_norm_pre_clip': 3.7997419834136963, 'timing/data': 0.006570706143975258, 'timing/model': 2.6803602389991283, 'learning_rate/qwen_vl_interface': 9.476736708493472e-06, 'learning_rate/action_model': 9.476736708493472e-05, 'epoch': 1.63}) +[2026-06-19 20:11:47,437][starVLA.training.train_starvla][INFO] - Step 840, Loss: {'train/loss': 0.3612805902957916, 'train/grad_norm_pre_clip': 4.449556350708008, 'timing/data': 0.006241628900170326, 'timing/model': 2.675823552533984, 'learning_rate/qwen_vl_interface': 9.475346293087457e-06, 'learning_rate/action_model': 9.475346293087458e-05, 'epoch': 1.63}) +[2026-06-19 20:11:57,938][starVLA.training.train_starvla][INFO] - Step 841, Loss: {'train/loss': 0.4639926552772522, 'train/grad_norm_pre_clip': 6.910965442657471, 'timing/data': 0.006343066692352295, 'timing/model': 2.6736238058656454, 'learning_rate/qwen_vl_interface': 9.473954140801184e-06, 'learning_rate/action_model': 9.473954140801185e-05, 'epoch': 1.63}) +[2026-06-19 20:12:08,615][starVLA.training.train_starvla][INFO] - Step 842, Loss: {'train/loss': 0.6067440509796143, 'train/grad_norm_pre_clip': 10.079658508300781, 'timing/data': 0.006118215620517731, 'timing/model': 2.7556566447019577, 'learning_rate/qwen_vl_interface': 9.472560252206918e-06, 'learning_rate/action_model': 9.472560252206918e-05, 'epoch': 1.63}) +[2026-06-19 20:12:19,135][starVLA.training.train_starvla][INFO] - Step 843, Loss: {'train/loss': 0.5063860416412354, 'train/grad_norm_pre_clip': 5.3624162673950195, 'timing/data': 0.0063760727643966675, 'timing/model': 2.6760701164603233, 'learning_rate/qwen_vl_interface': 9.471164627877633e-06, 'learning_rate/action_model': 9.471164627877633e-05, 'epoch': 1.63}) +[2026-06-19 20:12:29,819][starVLA.training.train_starvla][INFO] - Step 844, Loss: {'train/loss': 0.36287522315979004, 'train/grad_norm_pre_clip': 2.63432240486145, 'timing/data': 0.006534019485116005, 'timing/model': 2.7585062477737665, 'learning_rate/qwen_vl_interface': 9.469767268387016e-06, 'learning_rate/action_model': 9.469767268387016e-05, 'epoch': 1.64}) +[2026-06-19 20:12:40,397][starVLA.training.train_starvla][INFO] - Step 845, Loss: {'train/loss': 0.35064879059791565, 'train/grad_norm_pre_clip': 2.633406639099121, 'timing/data': 0.006496770307421684, 'timing/model': 2.7325678486377, 'learning_rate/qwen_vl_interface': 9.468368174309468e-06, 'learning_rate/action_model': 9.468368174309469e-05, 'epoch': 1.64}) +[2026-06-19 20:12:50,839][starVLA.training.train_starvla][INFO] - Step 846, Loss: {'train/loss': 0.8288307189941406, 'train/grad_norm_pre_clip': 11.533313751220703, 'timing/data': 0.00643790140748024, 'timing/model': 2.676914243027568, 'learning_rate/qwen_vl_interface': 9.466967346220107e-06, 'learning_rate/action_model': 9.466967346220107e-05, 'epoch': 1.64}) +[2026-06-19 20:13:01,436][starVLA.training.train_starvla][INFO] - Step 847, Loss: {'train/loss': 0.5302761793136597, 'train/grad_norm_pre_clip': 6.023563861846924, 'timing/data': 0.006274798884987831, 'timing/model': 2.6995697673410177, 'learning_rate/qwen_vl_interface': 9.465564784694759e-06, 'learning_rate/action_model': 9.465564784694759e-05, 'epoch': 1.64}) +[2026-06-19 20:13:11,942][starVLA.training.train_starvla][INFO] - Step 848, Loss: {'train/loss': 0.6947939991950989, 'train/grad_norm_pre_clip': 11.215921401977539, 'timing/data': 0.005948865786194801, 'timing/model': 2.6764693427830935, 'learning_rate/qwen_vl_interface': 9.464160490309962e-06, 'learning_rate/action_model': 9.464160490309962e-05, 'epoch': 1.64}) +[2026-06-19 20:13:22,640][starVLA.training.train_starvla][INFO] - Step 849, Loss: {'train/loss': 0.4839482605457306, 'train/grad_norm_pre_clip': 3.8943088054656982, 'timing/data': 0.006176413968205452, 'timing/model': 2.676745466887951, 'learning_rate/qwen_vl_interface': 9.462754463642972e-06, 'learning_rate/action_model': 9.46275446364297e-05, 'epoch': 1.65}) +[2026-06-19 20:13:33,319][starVLA.training.train_starvla][INFO] - Step 850, Loss: {'train/loss': 0.4250158369541168, 'train/grad_norm_pre_clip': 11.706741333007812, 'timing/data': 0.0061660706996917725, 'timing/model': 2.7780008912086487, 'learning_rate/qwen_vl_interface': 9.461346705271753e-06, 'learning_rate/action_model': 9.461346705271752e-05, 'epoch': 1.65}) +[2026-06-19 20:13:43,935][starVLA.training.train_starvla][INFO] - Step 851, Loss: {'train/loss': 0.3618190586566925, 'train/grad_norm_pre_clip': 3.730656147003174, 'timing/data': 0.00650496780872345, 'timing/model': 2.6722838766872883, 'learning_rate/qwen_vl_interface': 9.459937215774976e-06, 'learning_rate/action_model': 9.459937215774976e-05, 'epoch': 1.65}) +[2026-06-19 20:13:54,437][starVLA.training.train_starvla][INFO] - Step 852, Loss: {'train/loss': 0.33013802766799927, 'train/grad_norm_pre_clip': 2.1036107540130615, 'timing/data': 0.006303930655121803, 'timing/model': 2.67838173545897, 'learning_rate/qwen_vl_interface': 9.458525995732035e-06, 'learning_rate/action_model': 9.458525995732035e-05, 'epoch': 1.65}) +[2026-06-19 20:14:04,938][starVLA.training.train_starvla][INFO] - Step 853, Loss: {'train/loss': 0.48117774724960327, 'train/grad_norm_pre_clip': 5.687211990356445, 'timing/data': 0.0064369086176157, 'timing/model': 2.676261166110635, 'learning_rate/qwen_vl_interface': 9.457113045723027e-06, 'learning_rate/action_model': 9.457113045723026e-05, 'epoch': 1.65}) +[2026-06-19 20:14:15,604][starVLA.training.train_starvla][INFO] - Step 854, Loss: {'train/loss': 0.29678821563720703, 'train/grad_norm_pre_clip': 3.265991449356079, 'timing/data': 0.006502058357000351, 'timing/model': 2.735991196706891, 'learning_rate/qwen_vl_interface': 9.455698366328762e-06, 'learning_rate/action_model': 9.455698366328762e-05, 'epoch': 1.66}) +[2026-06-19 20:14:26,142][starVLA.training.train_starvla][INFO] - Step 855, Loss: {'train/loss': 0.36169755458831787, 'train/grad_norm_pre_clip': 8.618858337402344, 'timing/data': 0.006303021684288979, 'timing/model': 2.683619311079383, 'learning_rate/qwen_vl_interface': 9.454281958130761e-06, 'learning_rate/action_model': 9.454281958130761e-05, 'epoch': 1.66}) +[2026-06-19 20:14:36,742][starVLA.training.train_starvla][INFO] - Step 856, Loss: {'train/loss': 0.5011985301971436, 'train/grad_norm_pre_clip': 8.744213104248047, 'timing/data': 0.006169211119413376, 'timing/model': 2.677538573741913, 'learning_rate/qwen_vl_interface': 9.452863821711257e-06, 'learning_rate/action_model': 9.452863821711258e-05, 'epoch': 1.66}) +[2026-06-19 20:14:47,442][starVLA.training.train_starvla][INFO] - Step 857, Loss: {'train/loss': 0.3379295766353607, 'train/grad_norm_pre_clip': 7.091867446899414, 'timing/data': 0.006326610222458839, 'timing/model': 2.687579119578004, 'learning_rate/qwen_vl_interface': 9.451443957653195e-06, 'learning_rate/action_model': 9.451443957653195e-05, 'epoch': 1.66}) +[2026-06-19 20:14:57,937][starVLA.training.train_starvla][INFO] - Step 858, Loss: {'train/loss': 0.16748565435409546, 'train/grad_norm_pre_clip': 4.798769950866699, 'timing/data': 0.006324639543890953, 'timing/model': 2.675612820312381, 'learning_rate/qwen_vl_interface': 9.450022366540221e-06, 'learning_rate/action_model': 9.45002236654022e-05, 'epoch': 1.66}) +[2026-06-19 20:15:08,638][starVLA.training.train_starvla][INFO] - Step 859, Loss: {'train/loss': 0.35381650924682617, 'train/grad_norm_pre_clip': 7.946413993835449, 'timing/data': 0.006206892430782318, 'timing/model': 2.651596011593938, 'learning_rate/qwen_vl_interface': 9.448599048956702e-06, 'learning_rate/action_model': 9.448599048956702e-05, 'epoch': 1.66}) +[2026-06-19 20:15:19,446][starVLA.training.train_starvla][INFO] - Step 860, Loss: {'train/loss': 0.7619301080703735, 'train/grad_norm_pre_clip': 19.256391525268555, 'timing/data': 0.008929751813411713, 'timing/model': 2.8782746251672506, 'learning_rate/qwen_vl_interface': 9.447174005487708e-06, 'learning_rate/action_model': 9.447174005487708e-05, 'epoch': 1.67}) +[2026-06-19 20:15:30,040][starVLA.training.train_starvla][INFO] - Step 861, Loss: {'train/loss': 0.49481987953186035, 'train/grad_norm_pre_clip': 9.765365600585938, 'timing/data': 0.006479017436504364, 'timing/model': 2.701024355366826, 'learning_rate/qwen_vl_interface': 9.445747236719022e-06, 'learning_rate/action_model': 9.445747236719022e-05, 'epoch': 1.67}) +[2026-06-19 20:15:40,540][starVLA.training.train_starvla][INFO] - Step 862, Loss: {'train/loss': 0.490894079208374, 'train/grad_norm_pre_clip': 14.711136817932129, 'timing/data': 0.006462067365646362, 'timing/model': 2.675803592428565, 'learning_rate/qwen_vl_interface': 9.444318743237135e-06, 'learning_rate/action_model': 9.444318743237134e-05, 'epoch': 1.67}) +[2026-06-19 20:15:51,041][starVLA.training.train_starvla][INFO] - Step 863, Loss: {'train/loss': 0.4931413233280182, 'train/grad_norm_pre_clip': 8.896780014038086, 'timing/data': 0.006469115614891052, 'timing/model': 2.6787819284945726, 'learning_rate/qwen_vl_interface': 9.442888525629244e-06, 'learning_rate/action_model': 9.442888525629245e-05, 'epoch': 1.67}) +[2026-06-19 20:16:01,622][starVLA.training.train_starvla][INFO] - Step 864, Loss: {'train/loss': 0.5042452216148376, 'train/grad_norm_pre_clip': 13.03678035736084, 'timing/data': 0.006684955209493637, 'timing/model': 2.7091843597590923, 'learning_rate/qwen_vl_interface': 9.441456584483261e-06, 'learning_rate/action_model': 9.441456584483261e-05, 'epoch': 1.67}) +[2026-06-19 20:16:12,143][starVLA.training.train_starvla][INFO] - Step 865, Loss: {'train/loss': 0.42220383882522583, 'train/grad_norm_pre_clip': 9.135519981384277, 'timing/data': 0.006463639438152313, 'timing/model': 2.677152242511511, 'learning_rate/qwen_vl_interface': 9.440022920387801e-06, 'learning_rate/action_model': 9.440022920387801e-05, 'epoch': 1.68}) +[2026-06-19 20:16:23,026][starVLA.training.train_starvla][INFO] - Step 866, Loss: {'train/loss': 0.4060588479042053, 'train/grad_norm_pre_clip': 4.940544605255127, 'timing/data': 0.006374811753630638, 'timing/model': 2.77677390165627, 'learning_rate/qwen_vl_interface': 9.438587533932191e-06, 'learning_rate/action_model': 9.43858753393219e-05, 'epoch': 1.68}) +[2026-06-19 20:16:33,743][starVLA.training.train_starvla][INFO] - Step 867, Loss: {'train/loss': 0.2681605815887451, 'train/grad_norm_pre_clip': 4.6188740730285645, 'timing/data': 0.0069650523364543915, 'timing/model': 2.7050329260528088, 'learning_rate/qwen_vl_interface': 9.43715042570646e-06, 'learning_rate/action_model': 9.43715042570646e-05, 'epoch': 1.68}) +[2026-06-19 20:16:44,247][starVLA.training.train_starvla][INFO] - Step 868, Loss: {'train/loss': 0.2789464592933655, 'train/grad_norm_pre_clip': 3.940656900405884, 'timing/data': 0.006709607318043709, 'timing/model': 2.6758943293243647, 'learning_rate/qwen_vl_interface': 9.435711596301353e-06, 'learning_rate/action_model': 9.435711596301352e-05, 'epoch': 1.68}) +[2026-06-19 20:16:54,820][starVLA.training.train_starvla][INFO] - Step 869, Loss: {'train/loss': 0.3307233452796936, 'train/grad_norm_pre_clip': 5.079006671905518, 'timing/data': 0.0063648521900177, 'timing/model': 2.749607052654028, 'learning_rate/qwen_vl_interface': 9.434271046308317e-06, 'learning_rate/action_model': 9.434271046308317e-05, 'epoch': 1.68}) +[2026-06-19 20:17:05,529][starVLA.training.train_starvla][INFO] - Step 870, Loss: {'train/loss': 0.3965666890144348, 'train/grad_norm_pre_clip': 5.38018798828125, 'timing/data': 0.00641166977584362, 'timing/model': 2.7152775432914495, 'learning_rate/qwen_vl_interface': 9.432828776319507e-06, 'learning_rate/action_model': 9.432828776319505e-05, 'epoch': 1.69}) +[2026-06-19 20:17:16,245][starVLA.training.train_starvla][INFO] - Step 871, Loss: {'train/loss': 0.16188232600688934, 'train/grad_norm_pre_clip': 2.990657329559326, 'timing/data': 0.006872791796922684, 'timing/model': 2.65884917601943, 'learning_rate/qwen_vl_interface': 9.431384786927785e-06, 'learning_rate/action_model': 9.431384786927786e-05, 'epoch': 1.69}) +[2026-06-19 20:17:27,144][starVLA.training.train_starvla][INFO] - Step 872, Loss: {'train/loss': 0.27282479405403137, 'train/grad_norm_pre_clip': 5.370522975921631, 'timing/data': 0.006469998508691788, 'timing/model': 2.706283250823617, 'learning_rate/qwen_vl_interface': 9.429939078726723e-06, 'learning_rate/action_model': 9.429939078726722e-05, 'epoch': 1.69}) +[2026-06-19 20:17:37,747][starVLA.training.train_starvla][INFO] - Step 873, Loss: {'train/loss': 0.42478862404823303, 'train/grad_norm_pre_clip': 7.9482245445251465, 'timing/data': 0.0066946763545274734, 'timing/model': 2.6955981217324734, 'learning_rate/qwen_vl_interface': 9.428491652310598e-06, 'learning_rate/action_model': 9.428491652310597e-05, 'epoch': 1.69}) +[2026-06-19 20:17:48,374][starVLA.training.train_starvla][INFO] - Step 874, Loss: {'train/loss': 0.23120011389255524, 'train/grad_norm_pre_clip': 3.7123422622680664, 'timing/data': 0.006297890096902847, 'timing/model': 2.7388812843710184, 'learning_rate/qwen_vl_interface': 9.427042508274386e-06, 'learning_rate/action_model': 9.427042508274386e-05, 'epoch': 1.69}) +[2026-06-19 20:17:59,198][starVLA.training.train_starvla][INFO] - Step 875, Loss: {'train/loss': 0.2959907352924347, 'train/grad_norm_pre_clip': 5.674790382385254, 'timing/data': 0.006600197404623032, 'timing/model': 2.7490931898355484, 'learning_rate/qwen_vl_interface': 9.425591647213783e-06, 'learning_rate/action_model': 9.425591647213782e-05, 'epoch': 1.7}) +[2026-06-19 20:18:09,794][starVLA.training.train_starvla][INFO] - Step 876, Loss: {'train/loss': 0.3800484836101532, 'train/grad_norm_pre_clip': 11.743876457214355, 'timing/data': 0.006817035377025604, 'timing/model': 2.7187995854765177, 'learning_rate/qwen_vl_interface': 9.424139069725177e-06, 'learning_rate/action_model': 9.424139069725178e-05, 'epoch': 1.7}) +[2026-06-19 20:18:20,432][starVLA.training.train_starvla][INFO] - Step 877, Loss: {'train/loss': 0.4166170656681061, 'train/grad_norm_pre_clip': 5.409682273864746, 'timing/data': 0.0064798686653375626, 'timing/model': 2.7179446406662464, 'learning_rate/qwen_vl_interface': 9.422684776405673e-06, 'learning_rate/action_model': 9.422684776405672e-05, 'epoch': 1.7}) +[2026-06-19 20:18:31,051][starVLA.training.train_starvla][INFO] - Step 878, Loss: {'train/loss': 0.18297871947288513, 'train/grad_norm_pre_clip': 3.491655111312866, 'timing/data': 0.0069589316844940186, 'timing/model': 2.64611772634089, 'learning_rate/qwen_vl_interface': 9.421228767853072e-06, 'learning_rate/action_model': 9.421228767853072e-05, 'epoch': 1.7}) +[2026-06-19 20:18:41,644][starVLA.training.train_starvla][INFO] - Step 879, Loss: {'train/loss': 0.3349725604057312, 'train/grad_norm_pre_clip': 6.129434108734131, 'timing/data': 0.006519706919789314, 'timing/model': 2.694838974624872, 'learning_rate/qwen_vl_interface': 9.419771044665887e-06, 'learning_rate/action_model': 9.419771044665887e-05, 'epoch': 1.7}) +[2026-06-19 20:18:52,223][starVLA.training.train_starvla][INFO] - Step 880, Loss: {'train/loss': 0.35346218943595886, 'train/grad_norm_pre_clip': 6.988980293273926, 'timing/data': 0.0066915955394506454, 'timing/model': 2.752234425395727, 'learning_rate/qwen_vl_interface': 9.41831160744333e-06, 'learning_rate/action_model': 9.41831160744333e-05, 'epoch': 1.71}) +[2026-06-19 20:19:02,977][starVLA.training.train_starvla][INFO] - Step 881, Loss: {'train/loss': 0.27185559272766113, 'train/grad_norm_pre_clip': 5.1149115562438965, 'timing/data': 0.006802944466471672, 'timing/model': 2.7310260608792305, 'learning_rate/qwen_vl_interface': 9.416850456785322e-06, 'learning_rate/action_model': 9.416850456785321e-05, 'epoch': 1.71}) +[2026-06-19 20:19:13,542][starVLA.training.train_starvla][INFO] - Step 882, Loss: {'train/loss': 0.28770914673805237, 'train/grad_norm_pre_clip': 5.398738861083984, 'timing/data': 0.006308861076831818, 'timing/model': 2.6719316709786654, 'learning_rate/qwen_vl_interface': 9.415387593292489e-06, 'learning_rate/action_model': 9.415387593292488e-05, 'epoch': 1.71}) +[2026-06-19 20:19:24,365][starVLA.training.train_starvla][INFO] - Step 883, Loss: {'train/loss': 0.24661974608898163, 'train/grad_norm_pre_clip': 5.931349754333496, 'timing/data': 0.006928851827979088, 'timing/model': 2.9192027878016233, 'learning_rate/qwen_vl_interface': 9.413923017566155e-06, 'learning_rate/action_model': 9.413923017566155e-05, 'epoch': 1.71}) +[2026-06-19 20:19:34,955][starVLA.training.train_starvla][INFO] - Step 884, Loss: {'train/loss': 0.35628077387809753, 'train/grad_norm_pre_clip': 9.464120864868164, 'timing/data': 0.006335331127047539, 'timing/model': 2.675397027283907, 'learning_rate/qwen_vl_interface': 9.412456730208352e-06, 'learning_rate/action_model': 9.412456730208352e-05, 'epoch': 1.71}) +[2026-06-19 20:19:45,723][starVLA.training.train_starvla][INFO] - Step 885, Loss: {'train/loss': 0.25552162528038025, 'train/grad_norm_pre_clip': 4.895392894744873, 'timing/data': 0.0064004287123680115, 'timing/model': 2.757988665252924, 'learning_rate/qwen_vl_interface': 9.41098873182182e-06, 'learning_rate/action_model': 9.41098873182182e-05, 'epoch': 1.72}) +[2026-06-19 20:19:56,416][starVLA.training.train_starvla][INFO] - Step 886, Loss: {'train/loss': 0.22901657223701477, 'train/grad_norm_pre_clip': 5.68200159072876, 'timing/data': 0.006802916526794434, 'timing/model': 2.747970836237073, 'learning_rate/qwen_vl_interface': 9.409519023009993e-06, 'learning_rate/action_model': 9.409519023009993e-05, 'epoch': 1.72}) +[2026-06-19 20:20:07,117][starVLA.training.train_starvla][INFO] - Step 887, Loss: {'train/loss': 0.3102121353149414, 'train/grad_norm_pre_clip': 6.726599216461182, 'timing/data': 0.006340140476822853, 'timing/model': 2.721505381166935, 'learning_rate/qwen_vl_interface': 9.408047604377014e-06, 'learning_rate/action_model': 9.408047604377014e-05, 'epoch': 1.72}) +[2026-06-19 20:20:17,867][starVLA.training.train_starvla][INFO] - Step 888, Loss: {'train/loss': 0.26026856899261475, 'train/grad_norm_pre_clip': 9.489614486694336, 'timing/data': 0.006769435480237007, 'timing/model': 2.716065624728799, 'learning_rate/qwen_vl_interface': 9.40657447652773e-06, 'learning_rate/action_model': 9.406574476527731e-05, 'epoch': 1.72}) +[2026-06-19 20:20:28,424][starVLA.training.train_starvla][INFO] - Step 889, Loss: {'train/loss': 0.2976091802120209, 'train/grad_norm_pre_clip': 5.63503885269165, 'timing/data': 0.006518948823213577, 'timing/model': 2.7067429833114147, 'learning_rate/qwen_vl_interface': 9.405099640067685e-06, 'learning_rate/action_model': 9.405099640067685e-05, 'epoch': 1.72}) +[2026-06-19 20:20:39,123][starVLA.training.train_starvla][INFO] - Step 890, Loss: {'train/loss': 0.250385046005249, 'train/grad_norm_pre_clip': 4.2530622482299805, 'timing/data': 0.006596194580197334, 'timing/model': 2.7352848518639803, 'learning_rate/qwen_vl_interface': 9.40362309560313e-06, 'learning_rate/action_model': 9.40362309560313e-05, 'epoch': 1.72}) +[2026-06-19 20:20:49,918][starVLA.training.train_starvla][INFO] - Step 891, Loss: {'train/loss': 0.2704126238822937, 'train/grad_norm_pre_clip': 7.432651996612549, 'timing/data': 0.006685836240649223, 'timing/model': 2.7726931162178516, 'learning_rate/qwen_vl_interface': 9.402144843741016e-06, 'learning_rate/action_model': 9.402144843741017e-05, 'epoch': 1.73}) +[2026-06-19 20:21:00,547][starVLA.training.train_starvla][INFO] - Step 892, Loss: {'train/loss': 0.19250625371932983, 'train/grad_norm_pre_clip': 3.2523679733276367, 'timing/data': 0.006893722340464592, 'timing/model': 2.682945279404521, 'learning_rate/qwen_vl_interface': 9.400664885088999e-06, 'learning_rate/action_model': 9.400664885088999e-05, 'epoch': 1.73}) +[2026-06-19 20:21:11,143][starVLA.training.train_starvla][INFO] - Step 893, Loss: {'train/loss': 0.37514233589172363, 'train/grad_norm_pre_clip': 41.688697814941406, 'timing/data': 0.006744585931301117, 'timing/model': 2.6784712933003902, 'learning_rate/qwen_vl_interface': 9.39918322025543e-06, 'learning_rate/action_model': 9.39918322025543e-05, 'epoch': 1.73}) +[2026-06-19 20:21:21,642][starVLA.training.train_starvla][INFO] - Step 894, Loss: {'train/loss': 0.29917582869529724, 'train/grad_norm_pre_clip': 12.006102561950684, 'timing/data': 0.006846444681286812, 'timing/model': 2.675053345039487, 'learning_rate/qwen_vl_interface': 9.397699849849368e-06, 'learning_rate/action_model': 9.397699849849367e-05, 'epoch': 1.73}) +[2026-06-19 20:21:32,242][starVLA.training.train_starvla][INFO] - Step 895, Loss: {'train/loss': 0.456746906042099, 'train/grad_norm_pre_clip': 12.757256507873535, 'timing/data': 0.005906354635953903, 'timing/model': 2.6756019946187735, 'learning_rate/qwen_vl_interface': 9.39621477448057e-06, 'learning_rate/action_model': 9.39621477448057e-05, 'epoch': 1.73}) +[2026-06-19 20:21:42,742][starVLA.training.train_starvla][INFO] - Step 896, Loss: {'train/loss': 0.21176207065582275, 'train/grad_norm_pre_clip': 5.4253339767456055, 'timing/data': 0.00658181868493557, 'timing/model': 2.6755866073071957, 'learning_rate/qwen_vl_interface': 9.394727994759495e-06, 'learning_rate/action_model': 9.394727994759495e-05, 'epoch': 1.74}) +[2026-06-19 20:21:53,241][starVLA.training.train_starvla][INFO] - Step 897, Loss: {'train/loss': 0.3084631562232971, 'train/grad_norm_pre_clip': 8.753410339355469, 'timing/data': 0.00639474019408226, 'timing/model': 2.6750660482794046, 'learning_rate/qwen_vl_interface': 9.3932395112973e-06, 'learning_rate/action_model': 9.3932395112973e-05, 'epoch': 1.74}) +[2026-06-19 20:22:03,747][starVLA.training.train_starvla][INFO] - Step 898, Loss: {'train/loss': 0.3258253335952759, 'train/grad_norm_pre_clip': 6.0641045570373535, 'timing/data': 0.006638623774051666, 'timing/model': 2.6768009159713984, 'learning_rate/qwen_vl_interface': 9.391749324705846e-06, 'learning_rate/action_model': 9.391749324705846e-05, 'epoch': 1.74}) +[2026-06-19 20:22:14,385][starVLA.training.train_starvla][INFO] - Step 899, Loss: {'train/loss': 0.166045144200325, 'train/grad_norm_pre_clip': 4.391674041748047, 'timing/data': 0.00677204504609108, 'timing/model': 2.7135145496577024, 'learning_rate/qwen_vl_interface': 9.390257435597692e-06, 'learning_rate/action_model': 9.390257435597692e-05, 'epoch': 1.74}) +[2026-06-19 20:22:25,186][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 900 +[2026-06-19 20:30:00,904][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 900 +[2026-06-19 20:30:00,905][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 900 +[2026-06-19 20:38:05,832][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 900 +[2026-06-19 20:38:05,834][starVLA.training.train_starvla][INFO] - Step 900, Loss: {'train/loss': 0.3533267676830292, 'train/grad_norm_pre_clip': 9.018933296203613, 'eval/loss': 0.5533662438392639, 'eval/latency_0/loss': 0.22694499790668488, 'eval/latency_1/loss': 0.21900151669979095, 'eval/latency_2/loss': 0.7825118899345398, 'eval/latency_3/loss': 0.618074357509613, 'eval/latency_4/loss': 0.8950756192207336, 'eval/flappy/loss': 0.5533662438392639, 'eval/flappy/latency_0/loss': 0.22694499790668488, 'eval/flappy/latency_1/loss': 0.21900151669979095, 'eval/flappy/latency_2/loss': 0.7825118899345398, 'eval/flappy/latency_3/loss': 0.618074357509613, 'eval/flappy/latency_4/loss': 0.8950756192207336, 'eval/flap_precision': 0.37943262411347517, 'eval/flap_recall': 0.7781818181818182, 'eval/flap_f1': 0.5101311084624552, 'eval/flap_rate_model': 0.15666666666666668, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.5101311084624552, 'eval/latency_0/flap_precision': 0.37943262411347517, 'eval/latency_0/flap_recall': 0.7781818181818182, 'eval/latency_0/flap_f1': 0.5101311084624552, 'eval/latency_0/flap_rate_model': 0.15666666666666668, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.5101311084624552, 'timing/data': 0.006567869335412979, 'timing/model': 2.7383348383009434, 'learning_rate/qwen_vl_interface': 9.3887638445861e-06, 'learning_rate/action_model': 9.3887638445861e-05, 'epoch': 1.74}) +[2026-06-19 20:38:05,834][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_900_state +[2026-06-19 20:38:05,854][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-19 20:38:11,681][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_900_state/model.safetensors +[2026-06-19 20:38:20,035][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_900_state/optimizer.bin +[2026-06-19 20:38:20,036][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_900_state/sampler.bin +[2026-06-19 20:38:20,036][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_900_state/sampler_1.bin +[2026-06-19 20:38:20,037][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_900_state/random_states_0.pkl +[2026-06-19 20:38:20,163][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 900 +[2026-06-19 20:38:20,163][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-19 20:38:20,177][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-19 20:38:30,618][starVLA.training.train_starvla][INFO] - Step 901, Loss: {'train/loss': 0.3371836245059967, 'train/grad_norm_pre_clip': 11.348271369934082, 'timing/data': 0.006495906040072441, 'timing/model': 2.673030698671937, 'learning_rate/qwen_vl_interface': 9.387268552285024e-06, 'learning_rate/action_model': 9.387268552285025e-05, 'epoch': 1.75}) +[2026-06-19 20:38:41,124][starVLA.training.train_starvla][INFO] - Step 902, Loss: {'train/loss': 0.2182508409023285, 'train/grad_norm_pre_clip': 9.552127838134766, 'timing/data': 0.006373347714543343, 'timing/model': 2.6714552864432335, 'learning_rate/qwen_vl_interface': 9.385771559309125e-06, 'learning_rate/action_model': 9.385771559309125e-05, 'epoch': 1.75}) +[2026-06-19 20:38:51,625][starVLA.training.train_starvla][INFO] - Step 903, Loss: {'train/loss': 0.28095918893814087, 'train/grad_norm_pre_clip': 8.159647941589355, 'timing/data': 0.006411818787455559, 'timing/model': 2.6779274214059114, 'learning_rate/qwen_vl_interface': 9.38427286627376e-06, 'learning_rate/action_model': 9.38427286627376e-05, 'epoch': 1.75}) +[2026-06-19 20:39:02,125][starVLA.training.train_starvla][INFO] - Step 904, Loss: {'train/loss': 0.2246043086051941, 'train/grad_norm_pre_clip': 7.129412651062012, 'timing/data': 0.0064012352377176285, 'timing/model': 2.6686848048120737, 'learning_rate/qwen_vl_interface': 9.382772473794985e-06, 'learning_rate/action_model': 9.382772473794985e-05, 'epoch': 1.75}) +[2026-06-19 20:39:12,624][starVLA.training.train_starvla][INFO] - Step 905, Loss: {'train/loss': 0.24489419162273407, 'train/grad_norm_pre_clip': 3.5082473754882812, 'timing/data': 0.006265690550208092, 'timing/model': 2.6750249918550253, 'learning_rate/qwen_vl_interface': 9.381270382489554e-06, 'learning_rate/action_model': 9.381270382489554e-05, 'epoch': 1.75}) +[2026-06-19 20:39:23,126][starVLA.training.train_starvla][INFO] - Step 906, Loss: {'train/loss': 0.2920949161052704, 'train/grad_norm_pre_clip': 11.011174201965332, 'timing/data': 0.006560616195201874, 'timing/model': 2.6727284770458937, 'learning_rate/qwen_vl_interface': 9.37976659297492e-06, 'learning_rate/action_model': 9.379766592974921e-05, 'epoch': 1.76}) +[2026-06-19 20:39:33,630][starVLA.training.train_starvla][INFO] - Step 907, Loss: {'train/loss': 0.4523627758026123, 'train/grad_norm_pre_clip': 13.48035717010498, 'timing/data': 0.006270868703722954, 'timing/model': 2.6808320451527834, 'learning_rate/qwen_vl_interface': 9.378261105869237e-06, 'learning_rate/action_model': 9.378261105869237e-05, 'epoch': 1.76}) +[2026-06-19 20:39:44,135][starVLA.training.train_starvla][INFO] - Step 908, Loss: {'train/loss': 0.4096146821975708, 'train/grad_norm_pre_clip': 11.57959270477295, 'timing/data': 0.006352096796035767, 'timing/model': 2.683545842766762, 'learning_rate/qwen_vl_interface': 9.376753921791348e-06, 'learning_rate/action_model': 9.376753921791348e-05, 'epoch': 1.76}) +[2026-06-19 20:39:54,626][starVLA.training.train_starvla][INFO] - Step 909, Loss: {'train/loss': 0.08408927917480469, 'train/grad_norm_pre_clip': 5.768840312957764, 'timing/data': 0.006556658074259758, 'timing/model': 2.6734473183751106, 'learning_rate/qwen_vl_interface': 9.375245041360805e-06, 'learning_rate/action_model': 9.375245041360804e-05, 'epoch': 1.76}) +[2026-06-19 20:40:05,267][starVLA.training.train_starvla][INFO] - Step 910, Loss: {'train/loss': 0.16960763931274414, 'train/grad_norm_pre_clip': 4.233395576477051, 'timing/data': 0.006339048966765404, 'timing/model': 2.8123105000704527, 'learning_rate/qwen_vl_interface': 9.37373446519785e-06, 'learning_rate/action_model': 9.373734465197849e-05, 'epoch': 1.76}) +[2026-06-19 20:40:15,731][starVLA.training.train_starvla][INFO] - Step 911, Loss: {'train/loss': 0.19139565527439117, 'train/grad_norm_pre_clip': 4.434180736541748, 'timing/data': 0.0066170841455459595, 'timing/model': 2.6738503128290176, 'learning_rate/qwen_vl_interface': 9.372222193923421e-06, 'learning_rate/action_model': 9.372222193923421e-05, 'epoch': 1.77}) +[2026-06-19 20:40:26,230][starVLA.training.train_starvla][INFO] - Step 912, Loss: {'train/loss': 0.13197602331638336, 'train/grad_norm_pre_clip': 2.6878671646118164, 'timing/data': 0.006507657468318939, 'timing/model': 2.670220522210002, 'learning_rate/qwen_vl_interface': 9.370708228159158e-06, 'learning_rate/action_model': 9.370708228159159e-05, 'epoch': 1.77}) +[2026-06-19 20:40:36,735][starVLA.training.train_starvla][INFO] - Step 913, Loss: {'train/loss': 0.0932302176952362, 'train/grad_norm_pre_clip': 3.1586520671844482, 'timing/data': 0.006467597559094429, 'timing/model': 2.6808516904711723, 'learning_rate/qwen_vl_interface': 9.369192568527398e-06, 'learning_rate/action_model': 9.369192568527398e-05, 'epoch': 1.77}) +[2026-06-19 20:40:47,234][starVLA.training.train_starvla][INFO] - Step 914, Loss: {'train/loss': 0.3427131772041321, 'train/grad_norm_pre_clip': 5.708178997039795, 'timing/data': 0.006145920604467392, 'timing/model': 2.6593558341264725, 'learning_rate/qwen_vl_interface': 9.367675215651168e-06, 'learning_rate/action_model': 9.367675215651168e-05, 'epoch': 1.77}) +[2026-06-19 20:40:57,732][starVLA.training.train_starvla][INFO] - Step 915, Loss: {'train/loss': 0.2831612825393677, 'train/grad_norm_pre_clip': 5.0797224044799805, 'timing/data': 0.006527647376060486, 'timing/model': 2.674520570784807, 'learning_rate/qwen_vl_interface': 9.366156170154194e-06, 'learning_rate/action_model': 9.366156170154194e-05, 'epoch': 1.77}) +[2026-06-19 20:41:08,436][starVLA.training.train_starvla][INFO] - Step 916, Loss: {'train/loss': 0.25463157892227173, 'train/grad_norm_pre_clip': 6.207554817199707, 'timing/data': 0.006476987153291702, 'timing/model': 2.688238749280572, 'learning_rate/qwen_vl_interface': 9.3646354326609e-06, 'learning_rate/action_model': 9.364635432660899e-05, 'epoch': 1.78}) +[2026-06-19 20:41:19,034][starVLA.training.train_starvla][INFO] - Step 917, Loss: {'train/loss': 0.291856974363327, 'train/grad_norm_pre_clip': 6.737911224365234, 'timing/data': 0.006336197257041931, 'timing/model': 2.6480020377784967, 'learning_rate/qwen_vl_interface': 9.3631130037964e-06, 'learning_rate/action_model': 9.363113003796401e-05, 'epoch': 1.78}) +[2026-06-19 20:41:29,537][starVLA.training.train_starvla][INFO] - Step 918, Loss: {'train/loss': 0.2155803143978119, 'train/grad_norm_pre_clip': 5.326330184936523, 'timing/data': 0.006372828036546707, 'timing/model': 2.6813456919044256, 'learning_rate/qwen_vl_interface': 9.361588884186513e-06, 'learning_rate/action_model': 9.361588884186513e-05, 'epoch': 1.78}) +[2026-06-19 20:41:40,033][starVLA.training.train_starvla][INFO] - Step 919, Loss: {'train/loss': 0.27644655108451843, 'train/grad_norm_pre_clip': 9.564435005187988, 'timing/data': 0.00665164552628994, 'timing/model': 2.6744093019515276, 'learning_rate/qwen_vl_interface': 9.360063074457744e-06, 'learning_rate/action_model': 9.360063074457743e-05, 'epoch': 1.78}) +[2026-06-19 20:41:50,534][starVLA.training.train_starvla][INFO] - Step 920, Loss: {'train/loss': 0.10883063077926636, 'train/grad_norm_pre_clip': 2.858656406402588, 'timing/data': 0.006676545366644859, 'timing/model': 2.6778085436671972, 'learning_rate/qwen_vl_interface': 9.358535575237296e-06, 'learning_rate/action_model': 9.358535575237295e-05, 'epoch': 1.78}) +[2026-06-19 20:42:01,036][starVLA.training.train_starvla][INFO] - Step 921, Loss: {'train/loss': 0.3042827844619751, 'train/grad_norm_pre_clip': 8.289085388183594, 'timing/data': 0.006407875567674637, 'timing/model': 2.674484459683299, 'learning_rate/qwen_vl_interface': 9.357006387153065e-06, 'learning_rate/action_model': 9.357006387153065e-05, 'epoch': 1.79}) +[2026-06-19 20:42:11,736][starVLA.training.train_starvla][INFO] - Step 922, Loss: {'train/loss': 0.21782028675079346, 'train/grad_norm_pre_clip': 6.139126300811768, 'timing/data': 0.006321638822555542, 'timing/model': 2.6735761780291796, 'learning_rate/qwen_vl_interface': 9.355475510833643e-06, 'learning_rate/action_model': 9.355475510833643e-05, 'epoch': 1.79}) +[2026-06-19 20:42:22,235][starVLA.training.train_starvla][INFO] - Step 923, Loss: {'train/loss': 0.22426941990852356, 'train/grad_norm_pre_clip': 6.383862018585205, 'timing/data': 0.00676509365439415, 'timing/model': 2.6792070753872395, 'learning_rate/qwen_vl_interface': 9.353942946908318e-06, 'learning_rate/action_model': 9.353942946908317e-05, 'epoch': 1.79}) +[2026-06-19 20:42:32,739][starVLA.training.train_starvla][INFO] - Step 924, Loss: {'train/loss': 0.6404010057449341, 'train/grad_norm_pre_clip': 18.000534057617188, 'timing/data': 0.006270959973335266, 'timing/model': 2.6751272697001696, 'learning_rate/qwen_vl_interface': 9.352408696007065e-06, 'learning_rate/action_model': 9.352408696007065e-05, 'epoch': 1.79}) +[2026-06-19 20:42:43,234][starVLA.training.train_starvla][INFO] - Step 925, Loss: {'train/loss': 0.18212269246578217, 'train/grad_norm_pre_clip': 5.776640892028809, 'timing/data': 0.00636286661028862, 'timing/model': 2.678081376478076, 'learning_rate/qwen_vl_interface': 9.350872758760558e-06, 'learning_rate/action_model': 9.350872758760559e-05, 'epoch': 1.79}) +[2026-06-19 20:42:53,735][starVLA.training.train_starvla][INFO] - Step 926, Loss: {'train/loss': 0.26552140712738037, 'train/grad_norm_pre_clip': 4.9281744956970215, 'timing/data': 0.006556795910000801, 'timing/model': 2.674774130806327, 'learning_rate/qwen_vl_interface': 9.349335135800164e-06, 'learning_rate/action_model': 9.349335135800163e-05, 'epoch': 1.79}) +[2026-06-19 20:43:04,237][starVLA.training.train_starvla][INFO] - Step 927, Loss: {'train/loss': 0.189948171377182, 'train/grad_norm_pre_clip': 3.927666664123535, 'timing/data': 0.006491554901003838, 'timing/model': 2.677942017093301, 'learning_rate/qwen_vl_interface': 9.347795827757942e-06, 'learning_rate/action_model': 9.347795827757941e-05, 'epoch': 1.8}) +[2026-06-19 20:43:14,738][starVLA.training.train_starvla][INFO] - Step 928, Loss: {'train/loss': 0.3228802978992462, 'train/grad_norm_pre_clip': 6.274033546447754, 'timing/data': 0.006617486476898193, 'timing/model': 2.6771101783961058, 'learning_rate/qwen_vl_interface': 9.346254835266643e-06, 'learning_rate/action_model': 9.346254835266642e-05, 'epoch': 1.8}) +[2026-06-19 20:43:25,251][starVLA.training.train_starvla][INFO] - Step 929, Loss: {'train/loss': 0.18493734300136566, 'train/grad_norm_pre_clip': 3.69948673248291, 'timing/data': 0.006178760901093483, 'timing/model': 2.6919648107141256, 'learning_rate/qwen_vl_interface': 9.344712158959709e-06, 'learning_rate/action_model': 9.344712158959709e-05, 'epoch': 1.8}) +[2026-06-19 20:43:35,736][starVLA.training.train_starvla][INFO] - Step 930, Loss: {'train/loss': 0.22967402637004852, 'train/grad_norm_pre_clip': 6.067958831787109, 'timing/data': 0.006502848118543625, 'timing/model': 2.6767923310399055, 'learning_rate/qwen_vl_interface': 9.343167799471278e-06, 'learning_rate/action_model': 9.343167799471277e-05, 'epoch': 1.8}) +[2026-06-19 20:43:46,239][starVLA.training.train_starvla][INFO] - Step 931, Loss: {'train/loss': 0.19477461278438568, 'train/grad_norm_pre_clip': 6.34150505065918, 'timing/data': 0.006297210231423378, 'timing/model': 2.6775319036096334, 'learning_rate/qwen_vl_interface': 9.341621757436177e-06, 'learning_rate/action_model': 9.341621757436177e-05, 'epoch': 1.8}) +[2026-06-19 20:43:56,819][starVLA.training.train_starvla][INFO] - Step 932, Loss: {'train/loss': 0.16854338347911835, 'train/grad_norm_pre_clip': 7.7161865234375, 'timing/data': 0.006267501041293144, 'timing/model': 2.76248012855649, 'learning_rate/qwen_vl_interface': 9.340074033489924e-06, 'learning_rate/action_model': 9.340074033489923e-05, 'epoch': 1.81}) +[2026-06-19 20:44:07,591][starVLA.training.train_starvla][INFO] - Step 933, Loss: {'train/loss': 0.4701997637748718, 'train/grad_norm_pre_clip': 9.147111892700195, 'timing/data': 0.006481407210230827, 'timing/model': 2.8565420042723417, 'learning_rate/qwen_vl_interface': 9.338524628268734e-06, 'learning_rate/action_model': 9.338524628268733e-05, 'epoch': 1.81}) +[2026-06-19 20:44:18,034][starVLA.training.train_starvla][INFO] - Step 934, Loss: {'train/loss': 0.29971757531166077, 'train/grad_norm_pre_clip': 5.093752384185791, 'timing/data': 0.0066477954387664795, 'timing/model': 2.6781097017228603, 'learning_rate/qwen_vl_interface': 9.336973542409505e-06, 'learning_rate/action_model': 9.336973542409504e-05, 'epoch': 1.81}) +[2026-06-19 20:44:28,538][starVLA.training.train_starvla][INFO] - Step 935, Loss: {'train/loss': 0.12308089435100555, 'train/grad_norm_pre_clip': 4.559631824493408, 'timing/data': 0.006839415058493614, 'timing/model': 2.6453819796442986, 'learning_rate/qwen_vl_interface': 9.33542077654983e-06, 'learning_rate/action_model': 9.33542077654983e-05, 'epoch': 1.81}) +[2026-06-19 20:44:39,039][starVLA.training.train_starvla][INFO] - Step 936, Loss: {'train/loss': 0.3213178515434265, 'train/grad_norm_pre_clip': 10.763530731201172, 'timing/data': 0.006488228216767311, 'timing/model': 2.6839830856770277, 'learning_rate/qwen_vl_interface': 9.333866331327995e-06, 'learning_rate/action_model': 9.333866331327995e-05, 'epoch': 1.81}) +[2026-06-19 20:44:49,538][starVLA.training.train_starvla][INFO] - Step 937, Loss: {'train/loss': 0.32302984595298767, 'train/grad_norm_pre_clip': 18.44063949584961, 'timing/data': 0.006447995081543922, 'timing/model': 2.662161162123084, 'learning_rate/qwen_vl_interface': 9.332310207382972e-06, 'learning_rate/action_model': 9.332310207382972e-05, 'epoch': 1.82}) +[2026-06-19 20:45:00,033][starVLA.training.train_starvla][INFO] - Step 938, Loss: {'train/loss': 0.22445711493492126, 'train/grad_norm_pre_clip': 6.609992980957031, 'timing/data': 0.006501555442810059, 'timing/model': 2.6745751220732927, 'learning_rate/qwen_vl_interface': 9.330752405354426e-06, 'learning_rate/action_model': 9.330752405354426e-05, 'epoch': 1.82}) +[2026-06-19 20:45:10,538][starVLA.training.train_starvla][INFO] - Step 939, Loss: {'train/loss': 0.31235000491142273, 'train/grad_norm_pre_clip': 33.66093444824219, 'timing/data': 0.006311006844043732, 'timing/model': 2.679971730336547, 'learning_rate/qwen_vl_interface': 9.32919292588271e-06, 'learning_rate/action_model': 9.32919292588271e-05, 'epoch': 1.82}) +[2026-06-19 20:45:21,036][starVLA.training.train_starvla][INFO] - Step 940, Loss: {'train/loss': 0.2255188524723053, 'train/grad_norm_pre_clip': 8.60241413116455, 'timing/data': 0.006576476618647575, 'timing/model': 2.680364776402712, 'learning_rate/qwen_vl_interface': 9.327631769608864e-06, 'learning_rate/action_model': 9.327631769608864e-05, 'epoch': 1.82}) +[2026-06-19 20:45:31,535][starVLA.training.train_starvla][INFO] - Step 941, Loss: {'train/loss': 0.17597989737987518, 'train/grad_norm_pre_clip': 3.7351412773132324, 'timing/data': 0.006421409547328949, 'timing/model': 2.6785105168819427, 'learning_rate/qwen_vl_interface': 9.326068937174627e-06, 'learning_rate/action_model': 9.326068937174626e-05, 'epoch': 1.82}) +[2026-06-19 20:45:42,035][starVLA.training.train_starvla][INFO] - Step 942, Loss: {'train/loss': 0.13333284854888916, 'train/grad_norm_pre_clip': 3.5342230796813965, 'timing/data': 0.006822902709245682, 'timing/model': 2.6797743029892445, 'learning_rate/qwen_vl_interface': 9.324504429222412e-06, 'learning_rate/action_model': 9.324504429222412e-05, 'epoch': 1.83}) +[2026-06-19 20:45:52,534][starVLA.training.train_starvla][INFO] - Step 943, Loss: {'train/loss': 0.12046646326780319, 'train/grad_norm_pre_clip': 4.032931804656982, 'timing/data': 0.006524067372083664, 'timing/model': 2.674296362325549, 'learning_rate/qwen_vl_interface': 9.322938246395336e-06, 'learning_rate/action_model': 9.322938246395337e-05, 'epoch': 1.83}) +[2026-06-19 20:46:03,037][starVLA.training.train_starvla][INFO] - Step 944, Loss: {'train/loss': 0.2879601716995239, 'train/grad_norm_pre_clip': 9.38398265838623, 'timing/data': 0.006773723289370537, 'timing/model': 2.6746677979826927, 'learning_rate/qwen_vl_interface': 9.321370389337196e-06, 'learning_rate/action_model': 9.321370389337196e-05, 'epoch': 1.83}) +[2026-06-19 20:46:13,638][starVLA.training.train_starvla][INFO] - Step 945, Loss: {'train/loss': 0.41627559065818787, 'train/grad_norm_pre_clip': 18.176406860351562, 'timing/data': 0.006604423746466637, 'timing/model': 2.678257828578353, 'learning_rate/qwen_vl_interface': 9.319800858692478e-06, 'learning_rate/action_model': 9.319800858692476e-05, 'epoch': 1.83}) +[2026-06-19 20:46:24,136][starVLA.training.train_starvla][INFO] - Step 946, Loss: {'train/loss': 0.5355917811393738, 'train/grad_norm_pre_clip': 13.944583892822266, 'timing/data': 0.006209706887602806, 'timing/model': 2.6764103434979916, 'learning_rate/qwen_vl_interface': 9.318229655106357e-06, 'learning_rate/action_model': 9.318229655106356e-05, 'epoch': 1.83}) +[2026-06-19 20:46:34,635][starVLA.training.train_starvla][INFO] - Step 947, Loss: {'train/loss': 0.47993671894073486, 'train/grad_norm_pre_clip': 16.457082748413086, 'timing/data': 0.006476115435361862, 'timing/model': 2.6767880097031593, 'learning_rate/qwen_vl_interface': 9.316656779224693e-06, 'learning_rate/action_model': 9.316656779224693e-05, 'epoch': 1.84}) +[2026-06-19 20:46:45,135][starVLA.training.train_starvla][INFO] - Step 948, Loss: {'train/loss': 0.268056720495224, 'train/grad_norm_pre_clip': 18.173660278320312, 'timing/data': 0.006169786676764488, 'timing/model': 2.672571413218975, 'learning_rate/qwen_vl_interface': 9.31508223169404e-06, 'learning_rate/action_model': 9.31508223169404e-05, 'epoch': 1.84}) +[2026-06-19 20:46:55,635][starVLA.training.train_starvla][INFO] - Step 949, Loss: {'train/loss': 0.2302970141172409, 'train/grad_norm_pre_clip': 5.6351494789123535, 'timing/data': 0.00632532499730587, 'timing/model': 2.675518963485956, 'learning_rate/qwen_vl_interface': 9.313506013161633e-06, 'learning_rate/action_model': 9.313506013161633e-05, 'epoch': 1.84}) +[2026-06-19 20:47:06,132][starVLA.training.train_starvla][INFO] - Step 950, Loss: {'train/loss': 0.10304994881153107, 'train/grad_norm_pre_clip': 4.9550042152404785, 'timing/data': 0.0065742600709199905, 'timing/model': 2.6738293692469597, 'learning_rate/qwen_vl_interface': 9.311928124275396e-06, 'learning_rate/action_model': 9.311928124275395e-05, 'epoch': 1.84}) +[2026-06-19 20:47:16,635][starVLA.training.train_starvla][INFO] - Step 951, Loss: {'train/loss': 0.07668030261993408, 'train/grad_norm_pre_clip': 2.603861093521118, 'timing/data': 0.006742170080542564, 'timing/model': 2.6756487619131804, 'learning_rate/qwen_vl_interface': 9.310348565683938e-06, 'learning_rate/action_model': 9.310348565683939e-05, 'epoch': 1.84}) +[2026-06-19 20:47:27,140][starVLA.training.train_starvla][INFO] - Step 952, Loss: {'train/loss': 0.44649210572242737, 'train/grad_norm_pre_clip': 8.757769584655762, 'timing/data': 0.006627602502703667, 'timing/model': 2.677182672545314, 'learning_rate/qwen_vl_interface': 9.30876733803656e-06, 'learning_rate/action_model': 9.308767338036559e-05, 'epoch': 1.85}) +[2026-06-19 20:47:37,634][starVLA.training.train_starvla][INFO] - Step 953, Loss: {'train/loss': 0.22028186917304993, 'train/grad_norm_pre_clip': 6.549337387084961, 'timing/data': 0.006550421938300133, 'timing/model': 2.6780744679272175, 'learning_rate/qwen_vl_interface': 9.30718444198324e-06, 'learning_rate/action_model': 9.307184441983239e-05, 'epoch': 1.85}) +[2026-06-19 20:47:48,130][starVLA.training.train_starvla][INFO] - Step 954, Loss: {'train/loss': 0.1959715485572815, 'train/grad_norm_pre_clip': 5.364976406097412, 'timing/data': 0.006667891517281532, 'timing/model': 2.670629734173417, 'learning_rate/qwen_vl_interface': 9.30559987817465e-06, 'learning_rate/action_model': 9.30559987817465e-05, 'epoch': 1.85}) +[2026-06-19 20:47:58,633][starVLA.training.train_starvla][INFO] - Step 955, Loss: {'train/loss': 0.28474611043930054, 'train/grad_norm_pre_clip': 8.999275207519531, 'timing/data': 0.006748739629983902, 'timing/model': 2.677089635282755, 'learning_rate/qwen_vl_interface': 9.304013647262143e-06, 'learning_rate/action_model': 9.304013647262143e-05, 'epoch': 1.85}) +[2026-06-19 20:48:09,268][starVLA.training.train_starvla][INFO] - Step 956, Loss: {'train/loss': 0.45026400685310364, 'train/grad_norm_pre_clip': 14.31983470916748, 'timing/data': 0.0065604206174612045, 'timing/model': 2.8123889435082674, 'learning_rate/qwen_vl_interface': 9.302425749897761e-06, 'learning_rate/action_model': 9.302425749897761e-05, 'epoch': 1.85}) +[2026-06-19 20:48:19,734][starVLA.training.train_starvla][INFO] - Step 957, Loss: {'train/loss': 0.36018139123916626, 'train/grad_norm_pre_clip': 9.32719898223877, 'timing/data': 0.006498970091342926, 'timing/model': 2.676388945430517, 'learning_rate/qwen_vl_interface': 9.300836186734225e-06, 'learning_rate/action_model': 9.300836186734226e-05, 'epoch': 1.85}) +[2026-06-19 20:48:30,238][starVLA.training.train_starvla][INFO] - Step 958, Loss: {'train/loss': 0.315730482339859, 'train/grad_norm_pre_clip': 6.449740886688232, 'timing/data': 0.006553519517183304, 'timing/model': 2.6807380951941013, 'learning_rate/qwen_vl_interface': 9.299244958424947e-06, 'learning_rate/action_model': 9.299244958424946e-05, 'epoch': 1.86}) +[2026-06-19 20:48:40,736][starVLA.training.train_starvla][INFO] - Step 959, Loss: {'train/loss': 0.13982556760311127, 'train/grad_norm_pre_clip': 3.823962688446045, 'timing/data': 0.006602451205253601, 'timing/model': 2.6791643537580967, 'learning_rate/qwen_vl_interface': 9.29765206562402e-06, 'learning_rate/action_model': 9.297652065624018e-05, 'epoch': 1.86}) +[2026-06-19 20:48:51,234][starVLA.training.train_starvla][INFO] - Step 960, Loss: {'train/loss': 0.23243871331214905, 'train/grad_norm_pre_clip': 6.432033061981201, 'timing/data': 0.006332531571388245, 'timing/model': 2.663897916674614, 'learning_rate/qwen_vl_interface': 9.296057508986221e-06, 'learning_rate/action_model': 9.296057508986221e-05, 'epoch': 1.86}) +[2026-06-19 20:49:01,733][starVLA.training.train_starvla][INFO] - Step 961, Loss: {'train/loss': 0.13551151752471924, 'train/grad_norm_pre_clip': 5.644729137420654, 'timing/data': 0.006402730941772461, 'timing/model': 2.6774174124002457, 'learning_rate/qwen_vl_interface': 9.294461289167013e-06, 'learning_rate/action_model': 9.294461289167014e-05, 'epoch': 1.86}) +[2026-06-19 20:49:12,236][starVLA.training.train_starvla][INFO] - Step 962, Loss: {'train/loss': 0.34880581498146057, 'train/grad_norm_pre_clip': 8.66279411315918, 'timing/data': 0.006196331232786179, 'timing/model': 2.6796375140547752, 'learning_rate/qwen_vl_interface': 9.292863406822543e-06, 'learning_rate/action_model': 9.292863406822542e-05, 'epoch': 1.86}) +[2026-06-19 20:49:22,733][starVLA.training.train_starvla][INFO] - Step 963, Loss: {'train/loss': 0.13734567165374756, 'train/grad_norm_pre_clip': 2.889232635498047, 'timing/data': 0.0064129289239645, 'timing/model': 2.679013716056943, 'learning_rate/qwen_vl_interface': 9.291263862609638e-06, 'learning_rate/action_model': 9.291263862609638e-05, 'epoch': 1.87}) +[2026-06-19 20:49:33,236][starVLA.training.train_starvla][INFO] - Step 964, Loss: {'train/loss': 0.16995982825756073, 'train/grad_norm_pre_clip': 7.435939311981201, 'timing/data': 0.0062861014157533646, 'timing/model': 2.6770998686552048, 'learning_rate/qwen_vl_interface': 9.28966265718581e-06, 'learning_rate/action_model': 9.28966265718581e-05, 'epoch': 1.87}) +[2026-06-19 20:49:43,735][starVLA.training.train_starvla][INFO] - Step 965, Loss: {'train/loss': 0.4037601947784424, 'train/grad_norm_pre_clip': 7.078357219696045, 'timing/data': 0.006431169807910919, 'timing/model': 2.6723173540085554, 'learning_rate/qwen_vl_interface': 9.288059791209256e-06, 'learning_rate/action_model': 9.288059791209256e-05, 'epoch': 1.87}) +[2026-06-19 20:49:54,236][starVLA.training.train_starvla][INFO] - Step 966, Loss: {'train/loss': 0.3459594249725342, 'train/grad_norm_pre_clip': 6.5960693359375, 'timing/data': 0.006572628393769264, 'timing/model': 2.6784389168024063, 'learning_rate/qwen_vl_interface': 9.286455265338853e-06, 'learning_rate/action_model': 9.286455265338853e-05, 'epoch': 1.87}) +[2026-06-19 20:50:04,732][starVLA.training.train_starvla][INFO] - Step 967, Loss: {'train/loss': 0.27339455485343933, 'train/grad_norm_pre_clip': 7.148406028747559, 'timing/data': 0.006436629220843315, 'timing/model': 2.6785712651908398, 'learning_rate/qwen_vl_interface': 9.284849080234161e-06, 'learning_rate/action_model': 9.28484908023416e-05, 'epoch': 1.87}) +[2026-06-19 20:50:15,334][starVLA.training.train_starvla][INFO] - Step 968, Loss: {'train/loss': 0.21489226818084717, 'train/grad_norm_pre_clip': 5.070303440093994, 'timing/data': 0.006456870585680008, 'timing/model': 2.673477068543434, 'learning_rate/qwen_vl_interface': 9.283241236555421e-06, 'learning_rate/action_model': 9.283241236555421e-05, 'epoch': 1.88}) +[2026-06-19 20:50:25,837][starVLA.training.train_starvla][INFO] - Step 969, Loss: {'train/loss': 0.11812455207109451, 'train/grad_norm_pre_clip': 4.361490249633789, 'timing/data': 0.006521226838231087, 'timing/model': 2.681313367560506, 'learning_rate/qwen_vl_interface': 9.281631734963557e-06, 'learning_rate/action_model': 9.281631734963557e-05, 'epoch': 1.88}) +[2026-06-19 20:50:36,434][starVLA.training.train_starvla][INFO] - Step 970, Loss: {'train/loss': 0.23058514297008514, 'train/grad_norm_pre_clip': 13.70643138885498, 'timing/data': 0.0066697970032691956, 'timing/model': 2.678551709279418, 'learning_rate/qwen_vl_interface': 9.280020576120176e-06, 'learning_rate/action_model': 9.280020576120176e-05, 'epoch': 1.88}) +[2026-06-19 20:50:46,938][starVLA.training.train_starvla][INFO] - Step 971, Loss: {'train/loss': 0.38117414712905884, 'train/grad_norm_pre_clip': 9.87643051147461, 'timing/data': 0.0066994186490774155, 'timing/model': 2.6754464358091354, 'learning_rate/qwen_vl_interface': 9.278407760687562e-06, 'learning_rate/action_model': 9.278407760687562e-05, 'epoch': 1.88}) +[2026-06-19 20:50:57,436][starVLA.training.train_starvla][INFO] - Step 972, Loss: {'train/loss': 0.20480148494243622, 'train/grad_norm_pre_clip': 6.573485374450684, 'timing/data': 0.006536176428198814, 'timing/model': 2.6751162577420473, 'learning_rate/qwen_vl_interface': 9.276793289328686e-06, 'learning_rate/action_model': 9.276793289328685e-05, 'epoch': 1.88}) +[2026-06-19 20:51:07,932][starVLA.training.train_starvla][INFO] - Step 973, Loss: {'train/loss': 0.17130650579929352, 'train/grad_norm_pre_clip': 8.936175346374512, 'timing/data': 0.006350748240947723, 'timing/model': 2.6724194046109915, 'learning_rate/qwen_vl_interface': 9.27517716270719e-06, 'learning_rate/action_model': 9.275177162707191e-05, 'epoch': 1.89}) +[2026-06-19 20:51:18,431][starVLA.training.train_starvla][INFO] - Step 974, Loss: {'train/loss': 0.23435066640377045, 'train/grad_norm_pre_clip': 4.9910173416137695, 'timing/data': 0.0065884068608284, 'timing/model': 2.676848476752639, 'learning_rate/qwen_vl_interface': 9.273559381487409e-06, 'learning_rate/action_model': 9.273559381487409e-05, 'epoch': 1.89}) +[2026-06-19 20:51:28,933][starVLA.training.train_starvla][INFO] - Step 975, Loss: {'train/loss': 0.16982817649841309, 'train/grad_norm_pre_clip': 4.096660614013672, 'timing/data': 0.006388157606124878, 'timing/model': 2.6759582478553057, 'learning_rate/qwen_vl_interface': 9.27193994633435e-06, 'learning_rate/action_model': 9.27193994633435e-05, 'epoch': 1.89}) +[2026-06-19 20:51:39,434][starVLA.training.train_starvla][INFO] - Step 976, Loss: {'train/loss': 0.5923959612846375, 'train/grad_norm_pre_clip': 12.988954544067383, 'timing/data': 0.006533566862344742, 'timing/model': 2.679122844710946, 'learning_rate/qwen_vl_interface': 9.2703188579137e-06, 'learning_rate/action_model': 9.2703188579137e-05, 'epoch': 1.89}) +[2026-06-19 20:51:49,929][starVLA.training.train_starvla][INFO] - Step 977, Loss: {'train/loss': 0.08866468816995621, 'train/grad_norm_pre_clip': 3.3142237663269043, 'timing/data': 0.006526026874780655, 'timing/model': 2.6741462368518114, 'learning_rate/qwen_vl_interface': 9.26869611689183e-06, 'learning_rate/action_model': 9.26869611689183e-05, 'epoch': 1.89}) +[2026-06-19 20:52:00,433][starVLA.training.train_starvla][INFO] - Step 978, Loss: {'train/loss': 0.16602613031864166, 'train/grad_norm_pre_clip': 5.217551231384277, 'timing/data': 0.006424957886338234, 'timing/model': 2.6771117951720953, 'learning_rate/qwen_vl_interface': 9.267071723935784e-06, 'learning_rate/action_model': 9.267071723935784e-05, 'epoch': 1.9}) +[2026-06-19 20:52:11,066][starVLA.training.train_starvla][INFO] - Step 979, Loss: {'train/loss': 0.32395774126052856, 'train/grad_norm_pre_clip': 11.549181938171387, 'timing/data': 0.006508955731987953, 'timing/model': 2.8114550597965717, 'learning_rate/qwen_vl_interface': 9.26544567971329e-06, 'learning_rate/action_model': 9.26544567971329e-05, 'epoch': 1.9}) +[2026-06-19 20:52:21,533][starVLA.training.train_starvla][INFO] - Step 980, Loss: {'train/loss': 0.2515409588813782, 'train/grad_norm_pre_clip': 9.9373779296875, 'timing/data': 0.006232140585780144, 'timing/model': 2.6822672225534916, 'learning_rate/qwen_vl_interface': 9.263817984892756e-06, 'learning_rate/action_model': 9.263817984892757e-05, 'epoch': 1.9}) +[2026-06-19 20:52:32,036][starVLA.training.train_starvla][INFO] - Step 981, Loss: {'train/loss': 0.1944645345211029, 'train/grad_norm_pre_clip': 4.493185997009277, 'timing/data': 0.00679672509431839, 'timing/model': 2.678425271064043, 'learning_rate/qwen_vl_interface': 9.262188640143262e-06, 'learning_rate/action_model': 9.262188640143262e-05, 'epoch': 1.9}) +[2026-06-19 20:52:42,531][starVLA.training.train_starvla][INFO] - Step 982, Loss: {'train/loss': 0.22716355323791504, 'train/grad_norm_pre_clip': 8.4376220703125, 'timing/data': 0.006478818133473396, 'timing/model': 2.6739269979298115, 'learning_rate/qwen_vl_interface': 9.260557646134572e-06, 'learning_rate/action_model': 9.260557646134572e-05, 'epoch': 1.9}) +[2026-06-19 20:52:53,032][starVLA.training.train_starvla][INFO] - Step 983, Loss: {'train/loss': 0.1474127620458603, 'train/grad_norm_pre_clip': 6.235175132751465, 'timing/data': 0.006530884653329849, 'timing/model': 2.6592752058058977, 'learning_rate/qwen_vl_interface': 9.258925003537127e-06, 'learning_rate/action_model': 9.258925003537126e-05, 'epoch': 1.91}) +[2026-06-19 20:53:03,631][starVLA.training.train_starvla][INFO] - Step 984, Loss: {'train/loss': 0.09529019147157669, 'train/grad_norm_pre_clip': 4.978886127471924, 'timing/data': 0.005792103707790375, 'timing/model': 2.6814622320234776, 'learning_rate/qwen_vl_interface': 9.257290713022041e-06, 'learning_rate/action_model': 9.257290713022041e-05, 'epoch': 1.91}) +[2026-06-19 20:53:14,506][starVLA.training.train_starvla][INFO] - Step 985, Loss: {'train/loss': 0.11261941492557526, 'train/grad_norm_pre_clip': 4.9441375732421875, 'timing/data': 0.02034185081720352, 'timing/model': 2.69033544883132, 'learning_rate/qwen_vl_interface': 9.255654775261112e-06, 'learning_rate/action_model': 9.255654775261112e-05, 'epoch': 1.91}) +[2026-06-19 20:53:25,032][starVLA.training.train_starvla][INFO] - Step 986, Loss: {'train/loss': 0.12908193469047546, 'train/grad_norm_pre_clip': 4.492911338806152, 'timing/data': 0.0066389646381139755, 'timing/model': 2.6743732411414385, 'learning_rate/qwen_vl_interface': 9.254017190926813e-06, 'learning_rate/action_model': 9.254017190926813e-05, 'epoch': 1.91}) +[2026-06-19 20:53:35,534][starVLA.training.train_starvla][INFO] - Step 987, Loss: {'train/loss': 0.24581047892570496, 'train/grad_norm_pre_clip': 7.234626770019531, 'timing/data': 0.006551783531904221, 'timing/model': 2.6746341306716204, 'learning_rate/qwen_vl_interface': 9.25237796069229e-06, 'learning_rate/action_model': 9.25237796069229e-05, 'epoch': 1.91}) +[2026-06-19 20:53:46,031][starVLA.training.train_starvla][INFO] - Step 988, Loss: {'train/loss': 0.14856982231140137, 'train/grad_norm_pre_clip': 7.841958999633789, 'timing/data': 0.006506634876132011, 'timing/model': 2.6708561163395643, 'learning_rate/qwen_vl_interface': 9.250737085231371e-06, 'learning_rate/action_model': 9.250737085231372e-05, 'epoch': 1.91}) +[2026-06-19 20:53:56,539][starVLA.training.train_starvla][INFO] - Step 989, Loss: {'train/loss': 0.10071127861738205, 'train/grad_norm_pre_clip': 3.770677089691162, 'timing/data': 0.006313009187579155, 'timing/model': 2.678333781659603, 'learning_rate/qwen_vl_interface': 9.24909456521856e-06, 'learning_rate/action_model': 9.249094565218559e-05, 'epoch': 1.92}) +[2026-06-19 20:54:07,032][starVLA.training.train_starvla][INFO] - Step 990, Loss: {'train/loss': 0.5004912614822388, 'train/grad_norm_pre_clip': 12.531779289245605, 'timing/data': 0.006682932376861572, 'timing/model': 2.676859376952052, 'learning_rate/qwen_vl_interface': 9.247450401329027e-06, 'learning_rate/action_model': 9.247450401329027e-05, 'epoch': 1.92}) +[2026-06-19 20:54:17,633][starVLA.training.train_starvla][INFO] - Step 991, Loss: {'train/loss': 0.2797532081604004, 'train/grad_norm_pre_clip': 10.243867874145508, 'timing/data': 0.005983561277389526, 'timing/model': 2.6763159837573767, 'learning_rate/qwen_vl_interface': 9.245804594238637e-06, 'learning_rate/action_model': 9.245804594238636e-05, 'epoch': 1.92}) +[2026-06-19 20:54:28,135][starVLA.training.train_starvla][INFO] - Step 992, Loss: {'train/loss': 0.15191620588302612, 'train/grad_norm_pre_clip': 7.92747163772583, 'timing/data': 0.006800781935453415, 'timing/model': 2.674654694274068, 'learning_rate/qwen_vl_interface': 9.244157144623909e-06, 'learning_rate/action_model': 9.244157144623909e-05, 'epoch': 1.92}) +[2026-06-19 20:54:38,635][starVLA.training.train_starvla][INFO] - Step 993, Loss: {'train/loss': 0.39503371715545654, 'train/grad_norm_pre_clip': 12.510636329650879, 'timing/data': 0.006018847227096558, 'timing/model': 2.673886423930526, 'learning_rate/qwen_vl_interface': 9.242508053162051e-06, 'learning_rate/action_model': 9.242508053162051e-05, 'epoch': 1.92}) +[2026-06-19 20:54:49,139][starVLA.training.train_starvla][INFO] - Step 994, Loss: {'train/loss': 0.3318037688732147, 'train/grad_norm_pre_clip': 11.924153327941895, 'timing/data': 0.005931911990046501, 'timing/model': 2.6802918575704098, 'learning_rate/qwen_vl_interface': 9.240857320530944e-06, 'learning_rate/action_model': 9.240857320530943e-05, 'epoch': 1.93}) +[2026-06-19 20:54:59,640][starVLA.training.train_starvla][INFO] - Step 995, Loss: {'train/loss': 0.20737320184707642, 'train/grad_norm_pre_clip': 5.039819240570068, 'timing/data': 0.0057675521820783615, 'timing/model': 2.67776039801538, 'learning_rate/qwen_vl_interface': 9.23920494740914e-06, 'learning_rate/action_model': 9.23920494740914e-05, 'epoch': 1.93}) +[2026-06-19 20:55:10,135][starVLA.training.train_starvla][INFO] - Step 996, Loss: {'train/loss': 0.12686263024806976, 'train/grad_norm_pre_clip': 5.415556907653809, 'timing/data': 0.0059694889932870865, 'timing/model': 2.6759918220341206, 'learning_rate/qwen_vl_interface': 9.237550934475868e-06, 'learning_rate/action_model': 9.237550934475867e-05, 'epoch': 1.93}) +[2026-06-19 20:55:20,638][starVLA.training.train_starvla][INFO] - Step 997, Loss: {'train/loss': 0.3037131130695343, 'train/grad_norm_pre_clip': 9.741388320922852, 'timing/data': 0.0057890526950359344, 'timing/model': 2.6765787433832884, 'learning_rate/qwen_vl_interface': 9.235895282411029e-06, 'learning_rate/action_model': 9.235895282411029e-05, 'epoch': 1.93}) +[2026-06-19 20:55:31,237][starVLA.training.train_starvla][INFO] - Step 998, Loss: {'train/loss': 0.07758699357509613, 'train/grad_norm_pre_clip': 2.7572367191314697, 'timing/data': 0.005760852247476578, 'timing/model': 2.6946956887841225, 'learning_rate/qwen_vl_interface': 9.2342379918952e-06, 'learning_rate/action_model': 9.2342379918952e-05, 'epoch': 1.93}) +[2026-06-19 20:55:41,735][starVLA.training.train_starvla][INFO] - Step 999, Loss: {'train/loss': 0.06279980391263962, 'train/grad_norm_pre_clip': 2.5929830074310303, 'timing/data': 0.006612034514546394, 'timing/model': 2.6768440175801516, 'learning_rate/qwen_vl_interface': 9.232579063609631e-06, 'learning_rate/action_model': 9.232579063609631e-05, 'epoch': 1.94}) +[2026-06-19 20:55:52,238][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1000 +[2026-06-19 21:03:22,165][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1000 +[2026-06-19 21:03:22,166][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 1000 +[2026-06-19 21:11:30,338][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 1000 +[2026-06-19 21:11:30,339][starVLA.training.train_starvla][INFO] - Step 1000, Loss: {'train/loss': 0.2721678912639618, 'train/grad_norm_pre_clip': 8.071146965026855, 'eval/loss': 0.43164902925491333, 'eval/latency_0/loss': 0.22479307651519775, 'eval/latency_1/loss': 0.09873497486114502, 'eval/latency_2/loss': 0.5096372365951538, 'eval/latency_3/loss': 0.5431572198867798, 'eval/latency_4/loss': 0.7775148153305054, 'eval/flappy/loss': 0.43164902925491333, 'eval/flappy/latency_0/loss': 0.22479307651519775, 'eval/flappy/latency_1/loss': 0.09873497486114502, 'eval/flappy/latency_2/loss': 0.5096372365951538, 'eval/flappy/latency_3/loss': 0.5431572198867798, 'eval/flappy/latency_4/loss': 0.7775148153305054, 'eval/flap_precision': 0.44857142857142857, 'eval/flap_recall': 0.5709090909090909, 'eval/flap_f1': 0.5024, 'eval/flap_rate_model': 0.09722222222222222, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.5024, 'eval/latency_0/flap_precision': 0.44857142857142857, 'eval/latency_0/flap_recall': 0.5709090909090909, 'eval/latency_0/flap_f1': 0.5024, 'eval/latency_0/flap_rate_model': 0.09722222222222222, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.5024, 'timing/data': 0.006607852876186371, 'timing/model': 2.6776755824685097, 'learning_rate/qwen_vl_interface': 9.230918498236243e-06, 'learning_rate/action_model': 9.230918498236244e-05, 'epoch': 1.94}) +[2026-06-19 21:11:30,340][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_1000_state +[2026-06-19 21:11:30,344][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-19 21:11:36,135][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_1000_state/model.safetensors +[2026-06-19 21:11:50,373][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_1000_state/optimizer.bin +[2026-06-19 21:11:50,373][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_1000_state/sampler.bin +[2026-06-19 21:11:50,373][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_1000_state/sampler_1.bin +[2026-06-19 21:11:50,374][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_1000_state/random_states_0.pkl +[2026-06-19 21:11:50,502][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1000 +[2026-06-19 21:11:50,502][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-19 21:11:50,516][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-19 21:12:01,304][starVLA.training.train_starvla][INFO] - Step 1001, Loss: {'train/loss': 0.2990710437297821, 'train/grad_norm_pre_clip': 8.519952774047852, 'timing/data': 0.006667189300060272, 'timing/model': 2.766077868640423, 'learning_rate/qwen_vl_interface': 9.229256296457635e-06, 'learning_rate/action_model': 9.229256296457634e-05, 'epoch': 1.94}) +[2026-06-19 21:12:12,004][starVLA.training.train_starvla][INFO] - Step 1002, Loss: {'train/loss': 0.253338485956192, 'train/grad_norm_pre_clip': 6.133780479431152, 'timing/data': 0.006444063037633896, 'timing/model': 2.7491198237985373, 'learning_rate/qwen_vl_interface': 9.22759245895707e-06, 'learning_rate/action_model': 9.227592458957071e-05, 'epoch': 1.94}) +[2026-06-19 21:12:22,733][starVLA.training.train_starvla][INFO] - Step 1003, Loss: {'train/loss': 0.08649710565805435, 'train/grad_norm_pre_clip': 3.627134084701538, 'timing/data': 0.006239084526896477, 'timing/model': 2.705345068126917, 'learning_rate/qwen_vl_interface': 9.225926986418494e-06, 'learning_rate/action_model': 9.225926986418494e-05, 'epoch': 1.94}) +[2026-06-19 21:12:33,518][starVLA.training.train_starvla][INFO] - Step 1004, Loss: {'train/loss': 0.18944935500621796, 'train/grad_norm_pre_clip': 9.019564628601074, 'timing/data': 0.006979405879974365, 'timing/model': 2.7616046871989965, 'learning_rate/qwen_vl_interface': 9.224259879526519e-06, 'learning_rate/action_model': 9.224259879526519e-05, 'epoch': 1.95}) +[2026-06-19 21:12:44,308][starVLA.training.train_starvla][INFO] - Step 1005, Loss: {'train/loss': 0.21287748217582703, 'train/grad_norm_pre_clip': 8.22262954711914, 'timing/data': 0.006372151896357536, 'timing/model': 2.766285791993141, 'learning_rate/qwen_vl_interface': 9.222591138966427e-06, 'learning_rate/action_model': 9.222591138966426e-05, 'epoch': 1.95}) +[2026-06-19 21:12:55,018][starVLA.training.train_starvla][INFO] - Step 1006, Loss: {'train/loss': 0.24091388285160065, 'train/grad_norm_pre_clip': 8.669740676879883, 'timing/data': 0.00648144818842411, 'timing/model': 2.7573751732707024, 'learning_rate/qwen_vl_interface': 9.220920765424175e-06, 'learning_rate/action_model': 9.220920765424174e-05, 'epoch': 1.95}) +[2026-06-19 21:13:05,825][starVLA.training.train_starvla][INFO] - Step 1007, Loss: {'train/loss': 0.23394989967346191, 'train/grad_norm_pre_clip': 10.691128730773926, 'timing/data': 0.006203491240739822, 'timing/model': 2.785448445007205, 'learning_rate/qwen_vl_interface': 9.219248759586392e-06, 'learning_rate/action_model': 9.219248759586391e-05, 'epoch': 1.95}) +[2026-06-19 21:13:16,723][starVLA.training.train_starvla][INFO] - Step 1008, Loss: {'train/loss': 0.1571449488401413, 'train/grad_norm_pre_clip': 12.34255313873291, 'timing/data': 0.006440099328756332, 'timing/model': 2.788423929363489, 'learning_rate/qwen_vl_interface': 9.217575122140375e-06, 'learning_rate/action_model': 9.217575122140375e-05, 'epoch': 1.95}) +[2026-06-19 21:13:27,617][starVLA.training.train_starvla][INFO] - Step 1009, Loss: {'train/loss': 0.317313551902771, 'train/grad_norm_pre_clip': 8.90037727355957, 'timing/data': 0.006891267374157906, 'timing/model': 2.7535543646663427, 'learning_rate/qwen_vl_interface': 9.215899853774094e-06, 'learning_rate/action_model': 9.215899853774093e-05, 'epoch': 1.96}) +[2026-06-19 21:13:38,402][starVLA.training.train_starvla][INFO] - Step 1010, Loss: {'train/loss': 0.4333055317401886, 'train/grad_norm_pre_clip': 8.024016380310059, 'timing/data': 0.006558569148182869, 'timing/model': 2.6931818779557943, 'learning_rate/qwen_vl_interface': 9.214222955176189e-06, 'learning_rate/action_model': 9.214222955176189e-05, 'epoch': 1.96}) +[2026-06-19 21:13:49,123][starVLA.training.train_starvla][INFO] - Step 1011, Loss: {'train/loss': 0.17692852020263672, 'train/grad_norm_pre_clip': 6.93530797958374, 'timing/data': 0.006463529542088509, 'timing/model': 2.783540090546012, 'learning_rate/qwen_vl_interface': 9.212544427035967e-06, 'learning_rate/action_model': 9.212544427035967e-05, 'epoch': 1.96}) +[2026-06-19 21:13:59,911][starVLA.training.train_starvla][INFO] - Step 1012, Loss: {'train/loss': 0.2149277627468109, 'train/grad_norm_pre_clip': 5.2415080070495605, 'timing/data': 0.00641120970249176, 'timing/model': 2.713674169033766, 'learning_rate/qwen_vl_interface': 9.21086427004341e-06, 'learning_rate/action_model': 9.210864270043411e-05, 'epoch': 1.96}) +[2026-06-19 21:14:10,723][starVLA.training.train_starvla][INFO] - Step 1013, Loss: {'train/loss': 0.05072127282619476, 'train/grad_norm_pre_clip': 2.7729952335357666, 'timing/data': 0.006293078884482384, 'timing/model': 2.7762990929186344, 'learning_rate/qwen_vl_interface': 9.209182484889168e-06, 'learning_rate/action_model': 9.209182484889167e-05, 'epoch': 1.96}) +[2026-06-19 21:14:21,513][starVLA.training.train_starvla][INFO] - Step 1014, Loss: {'train/loss': 0.2014809101819992, 'train/grad_norm_pre_clip': 9.652641296386719, 'timing/data': 0.006377698853611946, 'timing/model': 2.745253572240472, 'learning_rate/qwen_vl_interface': 9.207499072264557e-06, 'learning_rate/action_model': 9.207499072264558e-05, 'epoch': 1.97}) +[2026-06-19 21:14:32,326][starVLA.training.train_starvla][INFO] - Step 1015, Loss: {'train/loss': 0.2218322902917862, 'train/grad_norm_pre_clip': 5.657140731811523, 'timing/data': 0.006304563954472542, 'timing/model': 2.7844413928687572, 'learning_rate/qwen_vl_interface': 9.205814032861566e-06, 'learning_rate/action_model': 9.205814032861566e-05, 'epoch': 1.97}) +[2026-06-19 21:14:43,129][starVLA.training.train_starvla][INFO] - Step 1016, Loss: {'train/loss': 0.21573233604431152, 'train/grad_norm_pre_clip': 9.099689483642578, 'timing/data': 0.006215333938598633, 'timing/model': 2.7091459278017282, 'learning_rate/qwen_vl_interface': 9.204127367372849e-06, 'learning_rate/action_model': 9.204127367372849e-05, 'epoch': 1.97}) +[2026-06-19 21:14:54,020][starVLA.training.train_starvla][INFO] - Step 1017, Loss: {'train/loss': 0.3156384527683258, 'train/grad_norm_pre_clip': 6.117032051086426, 'timing/data': 0.006135793402791023, 'timing/model': 2.76936399564147, 'learning_rate/qwen_vl_interface': 9.202439076491732e-06, 'learning_rate/action_model': 9.202439076491732e-05, 'epoch': 1.97}) +[2026-06-19 21:15:04,825][starVLA.training.train_starvla][INFO] - Step 1018, Loss: {'train/loss': 0.4103311002254486, 'train/grad_norm_pre_clip': 9.256604194641113, 'timing/data': 0.006576647982001305, 'timing/model': 2.7857513185590506, 'learning_rate/qwen_vl_interface': 9.200749160912208e-06, 'learning_rate/action_model': 9.200749160912208e-05, 'epoch': 1.97}) +[2026-06-19 21:15:15,726][starVLA.training.train_starvla][INFO] - Step 1019, Loss: {'train/loss': 0.37209418416023254, 'train/grad_norm_pre_clip': 8.272799491882324, 'timing/data': 0.0062821246683597565, 'timing/model': 2.776115408167243, 'learning_rate/qwen_vl_interface': 9.199057621328937e-06, 'learning_rate/action_model': 9.199057621328935e-05, 'epoch': 1.97}) +[2026-06-19 21:15:26,632][starVLA.training.train_starvla][INFO] - Step 1020, Loss: {'train/loss': 0.13128647208213806, 'train/grad_norm_pre_clip': 3.2052571773529053, 'timing/data': 0.00628192164003849, 'timing/model': 2.7847704216837883, 'learning_rate/qwen_vl_interface': 9.197364458437244e-06, 'learning_rate/action_model': 9.197364458437245e-05, 'epoch': 1.98}) +[2026-06-19 21:15:37,690][starVLA.training.train_starvla][INFO] - Step 1021, Loss: {'train/loss': 0.26941126585006714, 'train/grad_norm_pre_clip': 7.256925106048584, 'timing/data': 0.006286123767495155, 'timing/model': 2.9405664708465338, 'learning_rate/qwen_vl_interface': 9.19566967293313e-06, 'learning_rate/action_model': 9.19566967293313e-05, 'epoch': 1.98}) +[2026-06-19 21:15:48,417][starVLA.training.train_starvla][INFO] - Step 1022, Loss: {'train/loss': 0.09975899755954742, 'train/grad_norm_pre_clip': 6.178145885467529, 'timing/data': 0.006508907303214073, 'timing/model': 2.7674276158213615, 'learning_rate/qwen_vl_interface': 9.193973265513253e-06, 'learning_rate/action_model': 9.193973265513253e-05, 'epoch': 1.98}) +[2026-06-19 21:15:59,314][starVLA.training.train_starvla][INFO] - Step 1023, Loss: {'train/loss': 0.14890646934509277, 'train/grad_norm_pre_clip': 5.123253345489502, 'timing/data': 0.006376123055815697, 'timing/model': 2.764331804588437, 'learning_rate/qwen_vl_interface': 9.192275236874946e-06, 'learning_rate/action_model': 9.192275236874946e-05, 'epoch': 1.98}) +[2026-06-19 21:16:10,127][starVLA.training.train_starvla][INFO] - Step 1024, Loss: {'train/loss': 0.2502192556858063, 'train/grad_norm_pre_clip': 4.704198360443115, 'timing/data': 0.006277894601225853, 'timing/model': 2.774001007899642, 'learning_rate/qwen_vl_interface': 9.190575587716199e-06, 'learning_rate/action_model': 9.190575587716199e-05, 'epoch': 1.98}) +[2026-06-19 21:16:21,028][starVLA.training.train_starvla][INFO] - Step 1025, Loss: {'train/loss': 0.22603489458560944, 'train/grad_norm_pre_clip': 7.349737644195557, 'timing/data': 0.006485790014266968, 'timing/model': 2.7546987514942884, 'learning_rate/qwen_vl_interface': 9.188874318735677e-06, 'learning_rate/action_model': 9.188874318735676e-05, 'epoch': 1.99}) +[2026-06-19 21:16:31,927][starVLA.training.train_starvla][INFO] - Step 1026, Loss: {'train/loss': 0.19690202176570892, 'train/grad_norm_pre_clip': 6.573780536651611, 'timing/data': 0.006446046754717827, 'timing/model': 2.777674589306116, 'learning_rate/qwen_vl_interface': 9.187171430632707e-06, 'learning_rate/action_model': 9.187171430632707e-05, 'epoch': 1.99}) +[2026-06-19 21:16:42,815][starVLA.training.train_starvla][INFO] - Step 1027, Loss: {'train/loss': 0.2934895157814026, 'train/grad_norm_pre_clip': 9.830978393554688, 'timing/data': 0.006282564252614975, 'timing/model': 2.7622810937464237, 'learning_rate/qwen_vl_interface': 9.185466924107283e-06, 'learning_rate/action_model': 9.185466924107282e-05, 'epoch': 1.99}) +[2026-06-19 21:16:53,624][starVLA.training.train_starvla][INFO] - Step 1028, Loss: {'train/loss': 0.27163419127464294, 'train/grad_norm_pre_clip': 11.819774627685547, 'timing/data': 0.006433723494410515, 'timing/model': 2.784181861206889, 'learning_rate/qwen_vl_interface': 9.18376079986006e-06, 'learning_rate/action_model': 9.183760799860061e-05, 'epoch': 1.99}) +[2026-06-19 21:17:04,522][starVLA.training.train_starvla][INFO] - Step 1029, Loss: {'train/loss': 0.11614811420440674, 'train/grad_norm_pre_clip': 3.929888963699341, 'timing/data': 0.006193552166223526, 'timing/model': 2.782911876216531, 'learning_rate/qwen_vl_interface': 9.182053058592365e-06, 'learning_rate/action_model': 9.182053058592365e-05, 'epoch': 1.99}) +[2026-06-19 21:17:15,267][starVLA.training.train_starvla][INFO] - Step 1030, Loss: {'train/loss': 0.20293766260147095, 'train/grad_norm_pre_clip': 4.778825759887695, 'timing/data': 0.006289804354310036, 'timing/model': 2.7284423876553774, 'learning_rate/qwen_vl_interface': 9.180343701006183e-06, 'learning_rate/action_model': 9.180343701006183e-05, 'epoch': 2.0}) +[2026-06-19 21:17:26,129][starVLA.training.train_starvla][INFO] - Step 1031, Loss: {'train/loss': 0.15366333723068237, 'train/grad_norm_pre_clip': 6.5560808181762695, 'timing/data': 0.00635175034403801, 'timing/model': 2.77830358594656, 'learning_rate/qwen_vl_interface': 9.17863272780417e-06, 'learning_rate/action_model': 9.178632727804168e-05, 'epoch': 2.0}) +[2026-06-19 21:17:37,020][starVLA.training.train_starvla][INFO] - Step 1032, Loss: {'train/loss': 0.2108803242444992, 'train/grad_norm_pre_clip': 5.787728786468506, 'timing/data': 0.006192674860358238, 'timing/model': 2.770894592627883, 'learning_rate/qwen_vl_interface': 9.176920139689638e-06, 'learning_rate/action_model': 9.176920139689637e-05, 'epoch': 2.0}) +[2026-06-19 21:17:47,851][starVLA.training.train_starvla][INFO] - Step 1033, Loss: {'train/loss': 0.27134400606155396, 'train/grad_norm_pre_clip': 8.60086441040039, 'timing/data': 0.0056053511798381805, 'timing/model': 2.700032165274024, 'learning_rate/qwen_vl_interface': 9.175205937366568e-06, 'learning_rate/action_model': 9.175205937366569e-05, 'epoch': 2.0}) +[2026-06-19 21:17:58,664][starVLA.training.train_starvla][INFO] - Step 1034, Loss: {'train/loss': 0.35684096813201904, 'train/grad_norm_pre_clip': 13.928919792175293, 'timing/data': 0.00641765259206295, 'timing/model': 2.7171368058770895, 'learning_rate/qwen_vl_interface': 9.173490121539608e-06, 'learning_rate/action_model': 9.173490121539608e-05, 'epoch': 2.0}) +[2026-06-19 21:18:09,429][starVLA.training.train_starvla][INFO] - Step 1035, Loss: {'train/loss': 0.23378461599349976, 'train/grad_norm_pre_clip': 8.871358871459961, 'timing/data': 0.006615187972784042, 'timing/model': 2.7133269906044006, 'learning_rate/qwen_vl_interface': 9.171772692914062e-06, 'learning_rate/action_model': 9.171772692914061e-05, 'epoch': 2.01}) +[2026-06-19 21:18:20,220][starVLA.training.train_starvla][INFO] - Step 1036, Loss: {'train/loss': 0.37815722823143005, 'train/grad_norm_pre_clip': 9.33322525024414, 'timing/data': 0.006547996774315834, 'timing/model': 2.7705815955996513, 'learning_rate/qwen_vl_interface': 9.1700536521959e-06, 'learning_rate/action_model': 9.170053652195899e-05, 'epoch': 2.01}) +[2026-06-19 21:18:31,033][starVLA.training.train_starvla][INFO] - Step 1037, Loss: {'train/loss': 0.15299329161643982, 'train/grad_norm_pre_clip': 6.109523773193359, 'timing/data': 0.006557280197739601, 'timing/model': 2.793856820091605, 'learning_rate/qwen_vl_interface': 9.168333000091754e-06, 'learning_rate/action_model': 9.168333000091754e-05, 'epoch': 2.01}) +[2026-06-19 21:18:41,934][starVLA.training.train_starvla][INFO] - Step 1038, Loss: {'train/loss': 0.26006215810775757, 'train/grad_norm_pre_clip': 12.941841125488281, 'timing/data': 0.006431831046938896, 'timing/model': 2.7836742382496595, 'learning_rate/qwen_vl_interface': 9.16661073730892e-06, 'learning_rate/action_model': 9.166610737308919e-05, 'epoch': 2.01}) +[2026-06-19 21:18:52,827][starVLA.training.train_starvla][INFO] - Step 1039, Loss: {'train/loss': 0.1365443468093872, 'train/grad_norm_pre_clip': 4.184664249420166, 'timing/data': 0.006342584267258644, 'timing/model': 2.78357058390975, 'learning_rate/qwen_vl_interface': 9.164886864555355e-06, 'learning_rate/action_model': 9.164886864555355e-05, 'epoch': 2.01}) +[2026-06-19 21:19:03,719][starVLA.training.train_starvla][INFO] - Step 1040, Loss: {'train/loss': 0.1020842120051384, 'train/grad_norm_pre_clip': 7.348331451416016, 'timing/data': 0.006505640223622322, 'timing/model': 2.780178990215063, 'learning_rate/qwen_vl_interface': 9.163161382539678e-06, 'learning_rate/action_model': 9.163161382539678e-05, 'epoch': 2.02}) +[2026-06-19 21:19:14,517][starVLA.training.train_starvla][INFO] - Step 1041, Loss: {'train/loss': 0.1924351453781128, 'train/grad_norm_pre_clip': 9.997008323669434, 'timing/data': 0.006344091147184372, 'timing/model': 2.7700520139187574, 'learning_rate/qwen_vl_interface': 9.16143429197117e-06, 'learning_rate/action_model': 9.16143429197117e-05, 'epoch': 2.02}) +[2026-06-19 21:19:25,315][starVLA.training.train_starvla][INFO] - Step 1042, Loss: {'train/loss': 0.12239819765090942, 'train/grad_norm_pre_clip': 2.4463024139404297, 'timing/data': 0.006463611498475075, 'timing/model': 2.770884621888399, 'learning_rate/qwen_vl_interface': 9.159705593559773e-06, 'learning_rate/action_model': 9.159705593559773e-05, 'epoch': 2.02}) +[2026-06-19 21:19:36,116][starVLA.training.train_starvla][INFO] - Step 1043, Loss: {'train/loss': 0.14074760675430298, 'train/grad_norm_pre_clip': 5.009562969207764, 'timing/data': 0.006353141739964485, 'timing/model': 2.777240661904216, 'learning_rate/qwen_vl_interface': 9.157975288016089e-06, 'learning_rate/action_model': 9.157975288016088e-05, 'epoch': 2.02}) +[2026-06-19 21:19:47,067][starVLA.training.train_starvla][INFO] - Step 1044, Loss: {'train/loss': 0.2861430048942566, 'train/grad_norm_pre_clip': 9.939454078674316, 'timing/data': 0.006204981356859207, 'timing/model': 2.9204237554222345, 'learning_rate/qwen_vl_interface': 9.156243376051381e-06, 'learning_rate/action_model': 9.156243376051381e-05, 'epoch': 2.02}) +[2026-06-19 21:19:57,929][starVLA.training.train_starvla][INFO] - Step 1045, Loss: {'train/loss': 0.1422307789325714, 'train/grad_norm_pre_clip': 5.029151439666748, 'timing/data': 0.006319722160696983, 'timing/model': 2.7821081150323153, 'learning_rate/qwen_vl_interface': 9.154509858377574e-06, 'learning_rate/action_model': 9.154509858377574e-05, 'epoch': 2.03}) +[2026-06-19 21:20:08,819][starVLA.training.train_starvla][INFO] - Step 1046, Loss: {'train/loss': 0.4013984203338623, 'train/grad_norm_pre_clip': 15.175430297851562, 'timing/data': 0.005971986800432205, 'timing/model': 2.762326804921031, 'learning_rate/qwen_vl_interface': 9.152774735707248e-06, 'learning_rate/action_model': 9.152774735707247e-05, 'epoch': 2.03}) +[2026-06-19 21:20:19,625][starVLA.training.train_starvla][INFO] - Step 1047, Loss: {'train/loss': 0.045344073325395584, 'train/grad_norm_pre_clip': 5.083835601806641, 'timing/data': 0.006211832165718079, 'timing/model': 2.7939777802675962, 'learning_rate/qwen_vl_interface': 9.15103800875365e-06, 'learning_rate/action_model': 9.15103800875365e-05, 'epoch': 2.03}) +[2026-06-19 21:20:30,435][starVLA.training.train_starvla][INFO] - Step 1048, Loss: {'train/loss': 0.15393225848674774, 'train/grad_norm_pre_clip': 12.01545524597168, 'timing/data': 0.006383037194609642, 'timing/model': 2.7129559870809317, 'learning_rate/qwen_vl_interface': 9.149299678230681e-06, 'learning_rate/action_model': 9.149299678230681e-05, 'epoch': 2.03}) +[2026-06-19 21:20:41,046][starVLA.training.train_starvla][INFO] - Step 1049, Loss: {'train/loss': 0.2557983100414276, 'train/grad_norm_pre_clip': 6.953004837036133, 'timing/data': 0.005801644176244736, 'timing/model': 2.703309452161193, 'learning_rate/qwen_vl_interface': 9.147559744852905e-06, 'learning_rate/action_model': 9.147559744852905e-05, 'epoch': 2.03}) +[2026-06-19 21:20:51,757][starVLA.training.train_starvla][INFO] - Step 1050, Loss: {'train/loss': 0.056970685720443726, 'train/grad_norm_pre_clip': 6.139730453491211, 'timing/data': 0.006456214934587479, 'timing/model': 2.7130364049226046, 'learning_rate/qwen_vl_interface': 9.14581820933554e-06, 'learning_rate/action_model': 9.14581820933554e-05, 'epoch': 2.04}) +[2026-06-19 21:21:02,615][starVLA.training.train_starvla][INFO] - Step 1051, Loss: {'train/loss': 0.27927690744400024, 'train/grad_norm_pre_clip': 12.975057601928711, 'timing/data': 0.006454583257436752, 'timing/model': 2.7673205584287643, 'learning_rate/qwen_vl_interface': 9.14407507239447e-06, 'learning_rate/action_model': 9.14407507239447e-05, 'epoch': 2.04}) +[2026-06-19 21:21:13,416][starVLA.training.train_starvla][INFO] - Step 1052, Loss: {'train/loss': 0.2034875452518463, 'train/grad_norm_pre_clip': 7.946272373199463, 'timing/data': 0.006535263732075691, 'timing/model': 2.7719184532761574, 'learning_rate/qwen_vl_interface': 9.142330334746227e-06, 'learning_rate/action_model': 9.142330334746227e-05, 'epoch': 2.04}) +[2026-06-19 21:21:24,108][starVLA.training.train_starvla][INFO] - Step 1053, Loss: {'train/loss': 0.18098869919776917, 'train/grad_norm_pre_clip': 8.863764762878418, 'timing/data': 0.006599795073270798, 'timing/model': 2.7523896396160126, 'learning_rate/qwen_vl_interface': 9.14058399710801e-06, 'learning_rate/action_model': 9.14058399710801e-05, 'epoch': 2.04}) +[2026-06-19 21:21:34,921][starVLA.training.train_starvla][INFO] - Step 1054, Loss: {'train/loss': 0.0761197879910469, 'train/grad_norm_pre_clip': 8.348834991455078, 'timing/data': 0.006284087896347046, 'timing/model': 2.7785918712615967, 'learning_rate/qwen_vl_interface': 9.138836060197673e-06, 'learning_rate/action_model': 9.138836060197672e-05, 'epoch': 2.04}) +[2026-06-19 21:21:45,731][starVLA.training.train_starvla][INFO] - Step 1055, Loss: {'train/loss': 0.30172979831695557, 'train/grad_norm_pre_clip': 13.134965896606445, 'timing/data': 0.006426556035876274, 'timing/model': 2.7788491379469633, 'learning_rate/qwen_vl_interface': 9.137086524733725e-06, 'learning_rate/action_model': 9.137086524733724e-05, 'epoch': 2.04}) +[2026-06-19 21:21:56,727][starVLA.training.train_starvla][INFO] - Step 1056, Loss: {'train/loss': 0.4414159953594208, 'train/grad_norm_pre_clip': 13.312040328979492, 'timing/data': 0.007009411230683327, 'timing/model': 2.7866785507649183, 'learning_rate/qwen_vl_interface': 9.135335391435336e-06, 'learning_rate/action_model': 9.135335391435335e-05, 'epoch': 2.05}) +[2026-06-19 21:22:07,626][starVLA.training.train_starvla][INFO] - Step 1057, Loss: {'train/loss': 0.3586456775665283, 'train/grad_norm_pre_clip': 8.924849510192871, 'timing/data': 0.006429867818951607, 'timing/model': 2.7752635963261127, 'learning_rate/qwen_vl_interface': 9.13358266102233e-06, 'learning_rate/action_model': 9.133582661022328e-05, 'epoch': 2.05}) +[2026-06-19 21:22:18,529][starVLA.training.train_starvla][INFO] - Step 1058, Loss: {'train/loss': 0.14006933569908142, 'train/grad_norm_pre_clip': 5.384145736694336, 'timing/data': 0.006545323878526688, 'timing/model': 2.776426311582327, 'learning_rate/qwen_vl_interface': 9.131828334215185e-06, 'learning_rate/action_model': 9.131828334215185e-05, 'epoch': 2.05}) +[2026-06-19 21:22:29,346][starVLA.training.train_starvla][INFO] - Step 1059, Loss: {'train/loss': 0.26480862498283386, 'train/grad_norm_pre_clip': 9.563055992126465, 'timing/data': 0.006467174738645554, 'timing/model': 2.6938522178679705, 'learning_rate/qwen_vl_interface': 9.130072411735043e-06, 'learning_rate/action_model': 9.130072411735043e-05, 'epoch': 2.05}) +[2026-06-19 21:22:40,118][starVLA.training.train_starvla][INFO] - Step 1060, Loss: {'train/loss': 0.05071096867322922, 'train/grad_norm_pre_clip': 3.6481144428253174, 'timing/data': 0.0066632237285375595, 'timing/model': 2.761899644508958, 'learning_rate/qwen_vl_interface': 9.128314894303693e-06, 'learning_rate/action_model': 9.128314894303693e-05, 'epoch': 2.05}) +[2026-06-19 21:22:50,908][starVLA.training.train_starvla][INFO] - Step 1061, Loss: {'train/loss': 0.2545582950115204, 'train/grad_norm_pre_clip': 6.066177845001221, 'timing/data': 0.006504625082015991, 'timing/model': 2.7668225578963757, 'learning_rate/qwen_vl_interface': 9.126555782643592e-06, 'learning_rate/action_model': 9.126555782643591e-05, 'epoch': 2.06}) +[2026-06-19 21:23:01,713][starVLA.training.train_starvla][INFO] - Step 1062, Loss: {'train/loss': 0.30490565299987793, 'train/grad_norm_pre_clip': 10.707058906555176, 'timing/data': 0.006388215348124504, 'timing/model': 2.768663937225938, 'learning_rate/qwen_vl_interface': 9.124795077477835e-06, 'learning_rate/action_model': 9.124795077477834e-05, 'epoch': 2.06}) +[2026-06-19 21:23:12,431][starVLA.training.train_starvla][INFO] - Step 1063, Loss: {'train/loss': 0.22895975410938263, 'train/grad_norm_pre_clip': 8.75396728515625, 'timing/data': 0.0063063595443964005, 'timing/model': 2.7597362305969, 'learning_rate/qwen_vl_interface': 9.123032779530186e-06, 'learning_rate/action_model': 9.123032779530187e-05, 'epoch': 2.06}) +[2026-06-19 21:23:23,218][starVLA.training.train_starvla][INFO] - Step 1064, Loss: {'train/loss': 0.06707895547151566, 'train/grad_norm_pre_clip': 3.0283091068267822, 'timing/data': 0.0064401403069496155, 'timing/model': 2.7644773721694946, 'learning_rate/qwen_vl_interface': 9.12126888952506e-06, 'learning_rate/action_model': 9.12126888952506e-05, 'epoch': 2.06}) +[2026-06-19 21:23:33,938][starVLA.training.train_starvla][INFO] - Step 1065, Loss: {'train/loss': 0.10333827137947083, 'train/grad_norm_pre_clip': 4.597219944000244, 'timing/data': 0.006566287949681282, 'timing/model': 2.719378164038062, 'learning_rate/qwen_vl_interface': 9.119503408187521e-06, 'learning_rate/action_model': 9.119503408187522e-05, 'epoch': 2.06}) +[2026-06-19 21:23:44,626][starVLA.training.train_starvla][INFO] - Step 1066, Loss: {'train/loss': 0.3712020814418793, 'train/grad_norm_pre_clip': 12.940486907958984, 'timing/data': 0.006571236997842789, 'timing/model': 2.737284243106842, 'learning_rate/qwen_vl_interface': 9.117736336243299e-06, 'learning_rate/action_model': 9.117736336243298e-05, 'epoch': 2.07}) +[2026-06-19 21:23:55,665][starVLA.training.train_starvla][INFO] - Step 1067, Loss: {'train/loss': 0.24741563200950623, 'train/grad_norm_pre_clip': 5.896001815795898, 'timing/data': 0.006259791553020477, 'timing/model': 2.920951971784234, 'learning_rate/qwen_vl_interface': 9.115967674418765e-06, 'learning_rate/action_model': 9.115967674418765e-05, 'epoch': 2.07}) +[2026-06-19 21:24:06,526][starVLA.training.train_starvla][INFO] - Step 1068, Loss: {'train/loss': 0.3730211853981018, 'train/grad_norm_pre_clip': 7.260675430297852, 'timing/data': 0.006257399916648865, 'timing/model': 2.784333413466811, 'learning_rate/qwen_vl_interface': 9.114197423440949e-06, 'learning_rate/action_model': 9.11419742344095e-05, 'epoch': 2.07}) +[2026-06-19 21:24:17,420][starVLA.training.train_starvla][INFO] - Step 1069, Loss: {'train/loss': 0.25947698950767517, 'train/grad_norm_pre_clip': 9.699645042419434, 'timing/data': 0.0061954353004693985, 'timing/model': 2.770145930349827, 'learning_rate/qwen_vl_interface': 9.112425584037536e-06, 'learning_rate/action_model': 9.112425584037536e-05, 'epoch': 2.07}) +[2026-06-19 21:24:28,161][starVLA.training.train_starvla][INFO] - Step 1070, Loss: {'train/loss': 0.1647605001926422, 'train/grad_norm_pre_clip': 6.411410331726074, 'timing/data': 0.0063459910452365875, 'timing/model': 2.7089293990284204, 'learning_rate/qwen_vl_interface': 9.110652156936863e-06, 'learning_rate/action_model': 9.110652156936863e-05, 'epoch': 2.07}) +[2026-06-19 21:24:39,027][starVLA.training.train_starvla][INFO] - Step 1071, Loss: {'train/loss': 0.23154810070991516, 'train/grad_norm_pre_clip': 4.4407477378845215, 'timing/data': 0.006324805319309235, 'timing/model': 2.7553551718592644, 'learning_rate/qwen_vl_interface': 9.108877142867918e-06, 'learning_rate/action_model': 9.108877142867918e-05, 'epoch': 2.08}) +[2026-06-19 21:24:49,930][starVLA.training.train_starvla][INFO] - Step 1072, Loss: {'train/loss': 0.1368187814950943, 'train/grad_norm_pre_clip': 4.023007392883301, 'timing/data': 0.006413530558347702, 'timing/model': 2.779235264286399, 'learning_rate/qwen_vl_interface': 9.107100542560342e-06, 'learning_rate/action_model': 9.107100542560342e-05, 'epoch': 2.08}) +[2026-06-19 21:25:00,831][starVLA.training.train_starvla][INFO] - Step 1073, Loss: {'train/loss': 0.18635576963424683, 'train/grad_norm_pre_clip': 5.51793098449707, 'timing/data': 0.0063341520726680756, 'timing/model': 2.7795881535857916, 'learning_rate/qwen_vl_interface': 9.105322356744429e-06, 'learning_rate/action_model': 9.105322356744429e-05, 'epoch': 2.08}) +[2026-06-19 21:25:11,727][starVLA.training.train_starvla][INFO] - Step 1074, Loss: {'train/loss': 0.10273712128400803, 'train/grad_norm_pre_clip': 4.267868995666504, 'timing/data': 0.00647393986582756, 'timing/model': 2.7873909696936607, 'learning_rate/qwen_vl_interface': 9.103542586151121e-06, 'learning_rate/action_model': 9.10354258615112e-05, 'epoch': 2.08}) +[2026-06-19 21:25:22,630][starVLA.training.train_starvla][INFO] - Step 1075, Loss: {'train/loss': 0.23405440151691437, 'train/grad_norm_pre_clip': 4.959403038024902, 'timing/data': 0.006334694102406502, 'timing/model': 2.773442042991519, 'learning_rate/qwen_vl_interface': 9.101761231512017e-06, 'learning_rate/action_model': 9.101761231512017e-05, 'epoch': 2.08}) +[2026-06-19 21:25:33,528][starVLA.training.train_starvla][INFO] - Step 1076, Loss: {'train/loss': 0.14289367198944092, 'train/grad_norm_pre_clip': 5.908406734466553, 'timing/data': 0.006662437692284584, 'timing/model': 2.789286198094487, 'learning_rate/qwen_vl_interface': 9.099978293559365e-06, 'learning_rate/action_model': 9.099978293559366e-05, 'epoch': 2.09}) +[2026-06-19 21:25:44,428][starVLA.training.train_starvla][INFO] - Step 1077, Loss: {'train/loss': 0.3743169605731964, 'train/grad_norm_pre_clip': 14.556595802307129, 'timing/data': 0.006271002814173698, 'timing/model': 2.78183076903224, 'learning_rate/qwen_vl_interface': 9.098193773026065e-06, 'learning_rate/action_model': 9.098193773026064e-05, 'epoch': 2.09}) +[2026-06-19 21:25:55,221][starVLA.training.train_starvla][INFO] - Step 1078, Loss: {'train/loss': 0.09145710617303848, 'train/grad_norm_pre_clip': 4.286182403564453, 'timing/data': 0.0066183097660541534, 'timing/model': 2.7824036981910467, 'learning_rate/qwen_vl_interface': 9.096407670645663e-06, 'learning_rate/action_model': 9.096407670645662e-05, 'epoch': 2.09}) +[2026-06-19 21:26:06,122][starVLA.training.train_starvla][INFO] - Step 1079, Loss: {'train/loss': 0.2449515163898468, 'train/grad_norm_pre_clip': 8.185591697692871, 'timing/data': 0.006399860605597496, 'timing/model': 2.783253213390708, 'learning_rate/qwen_vl_interface': 9.094619987152357e-06, 'learning_rate/action_model': 9.094619987152358e-05, 'epoch': 2.09}) +[2026-06-19 21:26:16,933][starVLA.training.train_starvla][INFO] - Step 1080, Loss: {'train/loss': 0.09607022255659103, 'train/grad_norm_pre_clip': 4.385643482208252, 'timing/data': 0.006580399349331856, 'timing/model': 2.7928120009601116, 'learning_rate/qwen_vl_interface': 9.092830723281e-06, 'learning_rate/action_model': 9.092830723281e-05, 'epoch': 2.09}) +[2026-06-19 21:26:27,830][starVLA.training.train_starvla][INFO] - Step 1081, Loss: {'train/loss': 0.182406947016716, 'train/grad_norm_pre_clip': 11.160584449768066, 'timing/data': 0.006464539095759392, 'timing/model': 2.779907301068306, 'learning_rate/qwen_vl_interface': 9.09103987976709e-06, 'learning_rate/action_model': 9.09103987976709e-05, 'epoch': 2.1}) +[2026-06-19 21:26:38,698][starVLA.training.train_starvla][INFO] - Step 1082, Loss: {'train/loss': 0.28936663269996643, 'train/grad_norm_pre_clip': 10.594725608825684, 'timing/data': 0.006161222234368324, 'timing/model': 2.7482266407459974, 'learning_rate/qwen_vl_interface': 9.089247457346774e-06, 'learning_rate/action_model': 9.089247457346773e-05, 'epoch': 2.1}) +[2026-06-19 21:26:49,427][starVLA.training.train_starvla][INFO] - Step 1083, Loss: {'train/loss': 0.3446485996246338, 'train/grad_norm_pre_clip': 8.269856452941895, 'timing/data': 0.0063463244587183, 'timing/model': 2.7591898441314697, 'learning_rate/qwen_vl_interface': 9.087453456756849e-06, 'learning_rate/action_model': 9.087453456756849e-05, 'epoch': 2.1}) +[2026-06-19 21:27:00,335][starVLA.training.train_starvla][INFO] - Step 1084, Loss: {'train/loss': 0.16954170167446136, 'train/grad_norm_pre_clip': 12.288139343261719, 'timing/data': 0.006113555282354355, 'timing/model': 2.7780986335128546, 'learning_rate/qwen_vl_interface': 9.085657878734759e-06, 'learning_rate/action_model': 9.085657878734758e-05, 'epoch': 2.1}) +[2026-06-19 21:27:11,116][starVLA.training.train_starvla][INFO] - Step 1085, Loss: {'train/loss': 0.2547057271003723, 'train/grad_norm_pre_clip': 11.533114433288574, 'timing/data': 0.006574351340532303, 'timing/model': 2.7480538561940193, 'learning_rate/qwen_vl_interface': 9.083860724018603e-06, 'learning_rate/action_model': 9.083860724018602e-05, 'epoch': 2.1}) +[2026-06-19 21:27:21,922][starVLA.training.train_starvla][INFO] - Step 1086, Loss: {'train/loss': 0.3490062654018402, 'train/grad_norm_pre_clip': 10.811058044433594, 'timing/data': 0.006264032796025276, 'timing/model': 2.7691628374159336, 'learning_rate/qwen_vl_interface': 9.08206199334712e-06, 'learning_rate/action_model': 9.082061993347119e-05, 'epoch': 2.1}) +[2026-06-19 21:27:32,826][starVLA.training.train_starvla][INFO] - Step 1087, Loss: {'train/loss': 0.30857887864112854, 'train/grad_norm_pre_clip': 10.554393768310547, 'timing/data': 0.006329502910375595, 'timing/model': 2.781798841431737, 'learning_rate/qwen_vl_interface': 9.0802616874597e-06, 'learning_rate/action_model': 9.0802616874597e-05, 'epoch': 2.11}) +[2026-06-19 21:27:43,729][starVLA.training.train_starvla][INFO] - Step 1088, Loss: {'train/loss': 0.1319383680820465, 'train/grad_norm_pre_clip': 5.622751712799072, 'timing/data': 0.006306694820523262, 'timing/model': 2.77774665504694, 'learning_rate/qwen_vl_interface': 9.078459807096385e-06, 'learning_rate/action_model': 9.078459807096385e-05, 'epoch': 2.11}) +[2026-06-19 21:27:54,627][starVLA.training.train_starvla][INFO] - Step 1089, Loss: {'train/loss': 0.1676263064146042, 'train/grad_norm_pre_clip': 6.863338470458984, 'timing/data': 0.0062351468950510025, 'timing/model': 2.7769354563206434, 'learning_rate/qwen_vl_interface': 9.076656352997854e-06, 'learning_rate/action_model': 9.076656352997853e-05, 'epoch': 2.11}) +[2026-06-19 21:28:05,668][starVLA.training.train_starvla][INFO] - Step 1090, Loss: {'train/loss': 0.05661791190505028, 'train/grad_norm_pre_clip': 3.6825780868530273, 'timing/data': 0.0061707571148872375, 'timing/model': 2.9265116918832064, 'learning_rate/qwen_vl_interface': 9.074851325905443e-06, 'learning_rate/action_model': 9.074851325905442e-05, 'epoch': 2.11}) +[2026-06-19 21:28:16,529][starVLA.training.train_starvla][INFO] - Step 1091, Loss: {'train/loss': 0.13404078781604767, 'train/grad_norm_pre_clip': 5.184169769287109, 'timing/data': 0.006441492587327957, 'timing/model': 2.7782289441674948, 'learning_rate/qwen_vl_interface': 9.073044726561128e-06, 'learning_rate/action_model': 9.073044726561129e-05, 'epoch': 2.11}) +[2026-06-19 21:28:27,419][starVLA.training.train_starvla][INFO] - Step 1092, Loss: {'train/loss': 0.28413164615631104, 'train/grad_norm_pre_clip': 14.958589553833008, 'timing/data': 0.006495082750916481, 'timing/model': 2.7752971313893795, 'learning_rate/qwen_vl_interface': 9.071236555707536e-06, 'learning_rate/action_model': 9.071236555707537e-05, 'epoch': 2.12}) +[2026-06-19 21:28:38,227][starVLA.training.train_starvla][INFO] - Step 1093, Loss: {'train/loss': 0.08883653581142426, 'train/grad_norm_pre_clip': 7.108341693878174, 'timing/data': 0.006230335682630539, 'timing/model': 2.777075197547674, 'learning_rate/qwen_vl_interface': 9.069426814087937e-06, 'learning_rate/action_model': 9.069426814087937e-05, 'epoch': 2.12}) +[2026-06-19 21:28:49,125][starVLA.training.train_starvla][INFO] - Step 1094, Loss: {'train/loss': 0.1917516589164734, 'train/grad_norm_pre_clip': 7.270793914794922, 'timing/data': 0.006664242595434189, 'timing/model': 2.7638146840035915, 'learning_rate/qwen_vl_interface': 9.067615502446247e-06, 'learning_rate/action_model': 9.067615502446246e-05, 'epoch': 2.12}) +[2026-06-19 21:29:00,033][starVLA.training.train_starvla][INFO] - Step 1095, Loss: {'train/loss': 0.2089015394449234, 'train/grad_norm_pre_clip': 7.5414652824401855, 'timing/data': 0.006273586302995682, 'timing/model': 2.7810592837631702, 'learning_rate/qwen_vl_interface': 9.06580262152703e-06, 'learning_rate/action_model': 9.065802621527028e-05, 'epoch': 2.12}) +[2026-06-19 21:29:10,927][starVLA.training.train_starvla][INFO] - Step 1096, Loss: {'train/loss': 0.08818389475345612, 'train/grad_norm_pre_clip': 3.965406656265259, 'timing/data': 0.006351795047521591, 'timing/model': 2.779906375333667, 'learning_rate/qwen_vl_interface': 9.063988172075486e-06, 'learning_rate/action_model': 9.063988172075486e-05, 'epoch': 2.12}) +[2026-06-19 21:29:21,829][starVLA.training.train_starvla][INFO] - Step 1097, Loss: {'train/loss': 0.1015242263674736, 'train/grad_norm_pre_clip': 4.564703941345215, 'timing/data': 0.006406523287296295, 'timing/model': 2.7690389193594456, 'learning_rate/qwen_vl_interface': 9.062172154837475e-06, 'learning_rate/action_model': 9.062172154837475e-05, 'epoch': 2.13}) +[2026-06-19 21:29:32,713][starVLA.training.train_starvla][INFO] - Step 1098, Loss: {'train/loss': 0.07383298128843307, 'train/grad_norm_pre_clip': 3.874713897705078, 'timing/data': 0.00660187192261219, 'timing/model': 2.7624131999909878, 'learning_rate/qwen_vl_interface': 9.060354570559487e-06, 'learning_rate/action_model': 9.060354570559487e-05, 'epoch': 2.13}) +[2026-06-19 21:29:43,468][starVLA.training.train_starvla][INFO] - Step 1099, Loss: {'train/loss': 0.2947351038455963, 'train/grad_norm_pre_clip': 8.858241081237793, 'timing/data': 0.006499221548438072, 'timing/model': 2.7171242274343967, 'learning_rate/qwen_vl_interface': 9.058535419988667e-06, 'learning_rate/action_model': 9.058535419988667e-05, 'epoch': 2.13}) +[2026-06-19 21:29:54,309][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1100 +[2026-06-19 21:37:28,729][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1100 +[2026-06-19 21:37:28,730][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 1100 +[2026-06-19 21:45:37,617][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 1100 +[2026-06-19 21:45:37,691][starVLA.training.train_starvla][INFO] - Step 1100, Loss: {'train/loss': 0.36353206634521484, 'train/grad_norm_pre_clip': 15.07550048828125, 'eval/loss': 0.4852042496204376, 'eval/latency_0/loss': 0.2236316055059433, 'eval/latency_1/loss': 0.10387441515922546, 'eval/latency_2/loss': 0.6341453194618225, 'eval/latency_3/loss': 0.5895885229110718, 'eval/latency_4/loss': 0.8626297116279602, 'eval/flappy/loss': 0.4852042496204376, 'eval/flappy/latency_0/loss': 0.2236316055059433, 'eval/flappy/latency_1/loss': 0.10387441515922546, 'eval/flappy/latency_2/loss': 0.6341453194618225, 'eval/flappy/latency_3/loss': 0.5895885229110718, 'eval/flappy/latency_4/loss': 0.8626297116279602, 'eval/flap_precision': 0.5637860082304527, 'eval/flap_recall': 0.49818181818181817, 'eval/flap_f1': 0.528957528957529, 'eval/flap_rate_model': 0.0675, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.528957528957529, 'eval/latency_0/flap_precision': 0.5637860082304527, 'eval/latency_0/flap_recall': 0.49818181818181817, 'eval/latency_0/flap_f1': 0.528957528957529, 'eval/latency_0/flap_rate_model': 0.0675, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.528957528957529, 'timing/data': 0.006747839972376823, 'timing/model': 2.767455283552408, 'learning_rate/qwen_vl_interface': 9.056714703872794e-06, 'learning_rate/action_model': 9.056714703872794e-05, 'epoch': 2.13}) +[2026-06-19 21:45:37,692][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_1100_state +[2026-06-19 21:45:37,713][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-19 21:45:43,535][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_1100_state/model.safetensors +[2026-06-19 21:45:51,827][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_1100_state/optimizer.bin +[2026-06-19 21:45:51,828][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_1100_state/sampler.bin +[2026-06-19 21:45:51,828][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_1100_state/sampler_1.bin +[2026-06-19 21:45:51,829][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_1100_state/random_states_0.pkl +[2026-06-19 21:45:51,949][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1100 +[2026-06-19 21:45:51,949][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-19 21:45:51,963][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-19 21:46:02,458][starVLA.training.train_starvla][INFO] - Step 1101, Loss: {'train/loss': 0.3815770149230957, 'train/grad_norm_pre_clip': 9.637086868286133, 'timing/data': 0.006759444251656532, 'timing/model': 2.675308560952544, 'learning_rate/qwen_vl_interface': 9.0548924229603e-06, 'learning_rate/action_model': 9.054892422960301e-05, 'epoch': 2.13}) +[2026-06-19 21:46:13,031][starVLA.training.train_starvla][INFO] - Step 1102, Loss: {'train/loss': 0.24112820625305176, 'train/grad_norm_pre_clip': 11.149089813232422, 'timing/data': 0.00660526379942894, 'timing/model': 2.628524821251631, 'learning_rate/qwen_vl_interface': 9.053068578000256e-06, 'learning_rate/action_model': 9.053068578000256e-05, 'epoch': 2.14}) +[2026-06-19 21:46:23,645][starVLA.training.train_starvla][INFO] - Step 1103, Loss: {'train/loss': 0.061558838933706284, 'train/grad_norm_pre_clip': 3.4382238388061523, 'timing/data': 0.006793085485696793, 'timing/model': 2.648902392014861, 'learning_rate/qwen_vl_interface': 9.051243169742372e-06, 'learning_rate/action_model': 9.051243169742372e-05, 'epoch': 2.14}) +[2026-06-19 21:46:34,363][starVLA.training.train_starvla][INFO] - Step 1104, Loss: {'train/loss': 0.3334570527076721, 'train/grad_norm_pre_clip': 10.364910125732422, 'timing/data': 0.006579726934432983, 'timing/model': 2.7091908901929855, 'learning_rate/qwen_vl_interface': 9.049416198937008e-06, 'learning_rate/action_model': 9.049416198937007e-05, 'epoch': 2.14}) +[2026-06-19 21:46:45,107][starVLA.training.train_starvla][INFO] - Step 1105, Loss: {'train/loss': 0.1290983408689499, 'train/grad_norm_pre_clip': 4.714689254760742, 'timing/data': 0.00657365657389164, 'timing/model': 2.7270649671554565, 'learning_rate/qwen_vl_interface': 9.047587666335159e-06, 'learning_rate/action_model': 9.047587666335159e-05, 'epoch': 2.14}) +[2026-06-19 21:46:55,804][starVLA.training.train_starvla][INFO] - Step 1106, Loss: {'train/loss': 0.2827480733394623, 'train/grad_norm_pre_clip': 8.031508445739746, 'timing/data': 0.0068090856075286865, 'timing/model': 2.75769792124629, 'learning_rate/qwen_vl_interface': 9.045757572688469e-06, 'learning_rate/action_model': 9.04575757268847e-05, 'epoch': 2.14}) +[2026-06-19 21:47:06,477][starVLA.training.train_starvla][INFO] - Step 1107, Loss: {'train/loss': 0.432634174823761, 'train/grad_norm_pre_clip': 7.966538429260254, 'timing/data': 0.006541037932038307, 'timing/model': 2.7196948789060116, 'learning_rate/qwen_vl_interface': 9.043925918749218e-06, 'learning_rate/action_model': 9.043925918749218e-05, 'epoch': 2.15}) +[2026-06-19 21:47:17,122][starVLA.training.train_starvla][INFO] - Step 1108, Loss: {'train/loss': 0.10154814273118973, 'train/grad_norm_pre_clip': 6.754366397857666, 'timing/data': 0.007133379578590393, 'timing/model': 2.701644688844681, 'learning_rate/qwen_vl_interface': 9.042092705270333e-06, 'learning_rate/action_model': 9.042092705270333e-05, 'epoch': 2.15}) +[2026-06-19 21:47:27,659][starVLA.training.train_starvla][INFO] - Step 1109, Loss: {'train/loss': 0.24638891220092773, 'train/grad_norm_pre_clip': 6.666226863861084, 'timing/data': 0.006453877314925194, 'timing/model': 2.6962926629930735, 'learning_rate/qwen_vl_interface': 9.040257933005375e-06, 'learning_rate/action_model': 9.040257933005374e-05, 'epoch': 2.15}) +[2026-06-19 21:47:38,603][starVLA.training.train_starvla][INFO] - Step 1110, Loss: {'train/loss': 0.14043539762496948, 'train/grad_norm_pre_clip': 4.969446659088135, 'timing/data': 0.0067856647074222565, 'timing/model': 2.759311543777585, 'learning_rate/qwen_vl_interface': 9.038421602708552e-06, 'learning_rate/action_model': 9.038421602708552e-05, 'epoch': 2.15}) +[2026-06-19 21:47:49,412][starVLA.training.train_starvla][INFO] - Step 1111, Loss: {'train/loss': 0.26425501704216003, 'train/grad_norm_pre_clip': 5.808713912963867, 'timing/data': 0.006854094564914703, 'timing/model': 2.7746038381010294, 'learning_rate/qwen_vl_interface': 9.036583715134709e-06, 'learning_rate/action_model': 9.036583715134708e-05, 'epoch': 2.15}) +[2026-06-19 21:47:59,940][starVLA.training.train_starvla][INFO] - Step 1112, Loss: {'train/loss': 0.15405143797397614, 'train/grad_norm_pre_clip': 6.039797306060791, 'timing/data': 0.0066501665860414505, 'timing/model': 2.6725106425583363, 'learning_rate/qwen_vl_interface': 9.034744271039332e-06, 'learning_rate/action_model': 9.034744271039332e-05, 'epoch': 2.16}) +[2026-06-19 21:48:10,642][starVLA.training.train_starvla][INFO] - Step 1113, Loss: {'train/loss': 0.1583472192287445, 'train/grad_norm_pre_clip': 3.7852537631988525, 'timing/data': 0.0065675172954797745, 'timing/model': 2.7766335513442755, 'learning_rate/qwen_vl_interface': 9.03290327117855e-06, 'learning_rate/action_model': 9.032903271178549e-05, 'epoch': 2.16}) +[2026-06-19 21:48:21,414][starVLA.training.train_starvla][INFO] - Step 1114, Loss: {'train/loss': 0.1456444263458252, 'train/grad_norm_pre_clip': 5.687941551208496, 'timing/data': 0.006568968296051025, 'timing/model': 2.7537812031805515, 'learning_rate/qwen_vl_interface': 9.031060716309122e-06, 'learning_rate/action_model': 9.031060716309121e-05, 'epoch': 2.16}) +[2026-06-19 21:48:32,120][starVLA.training.train_starvla][INFO] - Step 1115, Loss: {'train/loss': 0.08755947649478912, 'train/grad_norm_pre_clip': 3.546522855758667, 'timing/data': 0.006756763905286789, 'timing/model': 2.7614686209708452, 'learning_rate/qwen_vl_interface': 9.02921660718846e-06, 'learning_rate/action_model': 9.029216607188459e-05, 'epoch': 2.16}) +[2026-06-19 21:48:42,819][starVLA.training.train_starvla][INFO] - Step 1116, Loss: {'train/loss': 0.2869127094745636, 'train/grad_norm_pre_clip': 8.584134101867676, 'timing/data': 0.006691113114356995, 'timing/model': 2.7608184050768614, 'learning_rate/qwen_vl_interface': 9.027370944574604e-06, 'learning_rate/action_model': 9.027370944574604e-05, 'epoch': 2.16}) +[2026-06-19 21:48:53,622][starVLA.training.train_starvla][INFO] - Step 1117, Loss: {'train/loss': 0.11300602555274963, 'train/grad_norm_pre_clip': 3.833725690841675, 'timing/data': 0.006683053448796272, 'timing/model': 2.7803819812834263, 'learning_rate/qwen_vl_interface': 9.025523729226236e-06, 'learning_rate/action_model': 9.025523729226236e-05, 'epoch': 2.16}) +[2026-06-19 21:49:04,321][starVLA.training.train_starvla][INFO] - Step 1118, Loss: {'train/loss': 0.2984031140804291, 'train/grad_norm_pre_clip': 7.993099689483643, 'timing/data': 0.00664585642516613, 'timing/model': 2.7620604299008846, 'learning_rate/qwen_vl_interface': 9.023674961902677e-06, 'learning_rate/action_model': 9.023674961902676e-05, 'epoch': 2.17}) +[2026-06-19 21:49:15,022][starVLA.training.train_starvla][INFO] - Step 1119, Loss: {'train/loss': 0.2392996996641159, 'train/grad_norm_pre_clip': 7.507402420043945, 'timing/data': 0.006750546395778656, 'timing/model': 2.7794691510498524, 'learning_rate/qwen_vl_interface': 9.021824643363883e-06, 'learning_rate/action_model': 9.021824643363883e-05, 'epoch': 2.17}) +[2026-06-19 21:49:25,912][starVLA.training.train_starvla][INFO] - Step 1120, Loss: {'train/loss': 0.11924947053194046, 'train/grad_norm_pre_clip': 7.614011764526367, 'timing/data': 0.00689772330224514, 'timing/model': 2.774283442646265, 'learning_rate/qwen_vl_interface': 9.019972774370456e-06, 'learning_rate/action_model': 9.019972774370455e-05, 'epoch': 2.17}) +[2026-06-19 21:49:36,879][starVLA.training.train_starvla][INFO] - Step 1121, Loss: {'train/loss': 0.24239346385002136, 'train/grad_norm_pre_clip': 8.089818954467773, 'timing/data': 0.00674639455974102, 'timing/model': 2.7330951057374477, 'learning_rate/qwen_vl_interface': 9.018119355683623e-06, 'learning_rate/action_model': 9.018119355683623e-05, 'epoch': 2.17}) +[2026-06-19 21:49:47,680][starVLA.training.train_starvla][INFO] - Step 1122, Loss: {'train/loss': 0.19133828580379486, 'train/grad_norm_pre_clip': 6.457376480102539, 'timing/data': 0.006949510425329208, 'timing/model': 2.7428926788270473, 'learning_rate/qwen_vl_interface': 9.016264388065258e-06, 'learning_rate/action_model': 9.016264388065258e-05, 'epoch': 2.17}) +[2026-06-19 21:49:58,524][starVLA.training.train_starvla][INFO] - Step 1123, Loss: {'train/loss': 0.24113643169403076, 'train/grad_norm_pre_clip': 12.061075210571289, 'timing/data': 0.006570488214492798, 'timing/model': 2.7904721032828093, 'learning_rate/qwen_vl_interface': 9.014407872277869e-06, 'learning_rate/action_model': 9.014407872277869e-05, 'epoch': 2.18}) +[2026-06-19 21:50:09,214][starVLA.training.train_starvla][INFO] - Step 1124, Loss: {'train/loss': 0.2571621239185333, 'train/grad_norm_pre_clip': 13.33632755279541, 'timing/data': 0.006875794380903244, 'timing/model': 2.771964680403471, 'learning_rate/qwen_vl_interface': 9.012549809084595e-06, 'learning_rate/action_model': 9.012549809084595e-05, 'epoch': 2.18}) +[2026-06-19 21:50:19,920][starVLA.training.train_starvla][INFO] - Step 1125, Loss: {'train/loss': 0.14707522094249725, 'train/grad_norm_pre_clip': 5.996098041534424, 'timing/data': 0.0066826362162828445, 'timing/model': 2.7519282400608063, 'learning_rate/qwen_vl_interface': 9.01069019924922e-06, 'learning_rate/action_model': 9.010690199249219e-05, 'epoch': 2.18}) +[2026-06-19 21:50:30,522][starVLA.training.train_starvla][INFO] - Step 1126, Loss: {'train/loss': 0.490020751953125, 'train/grad_norm_pre_clip': 16.139205932617188, 'timing/data': 0.006413968279957771, 'timing/model': 2.745178559795022, 'learning_rate/qwen_vl_interface': 9.008829043536157e-06, 'learning_rate/action_model': 9.008829043536156e-05, 'epoch': 2.18}) +[2026-06-19 21:50:41,225][starVLA.training.train_starvla][INFO] - Step 1127, Loss: {'train/loss': 0.11850821226835251, 'train/grad_norm_pre_clip': 7.6041741371154785, 'timing/data': 0.006700614467263222, 'timing/model': 2.7091756742447615, 'learning_rate/qwen_vl_interface': 9.006966342710458e-06, 'learning_rate/action_model': 9.006966342710457e-05, 'epoch': 2.18}) +[2026-06-19 21:50:52,328][starVLA.training.train_starvla][INFO] - Step 1128, Loss: {'train/loss': 0.35244670510292053, 'train/grad_norm_pre_clip': 17.006587982177734, 'timing/data': 0.006520206108689308, 'timing/model': 2.991378154605627, 'learning_rate/qwen_vl_interface': 9.005102097537809e-06, 'learning_rate/action_model': 9.005102097537809e-05, 'epoch': 2.19}) +[2026-06-19 21:51:03,118][starVLA.training.train_starvla][INFO] - Step 1129, Loss: {'train/loss': 0.24196389317512512, 'train/grad_norm_pre_clip': 11.94610595703125, 'timing/data': 0.006862953305244446, 'timing/model': 2.757343504577875, 'learning_rate/qwen_vl_interface': 9.00323630878453e-06, 'learning_rate/action_model': 9.00323630878453e-05, 'epoch': 2.19}) +[2026-06-19 21:51:13,921][starVLA.training.train_starvla][INFO] - Step 1130, Loss: {'train/loss': 0.062441080808639526, 'train/grad_norm_pre_clip': 5.764161109924316, 'timing/data': 0.006667574867606163, 'timing/model': 2.781699925661087, 'learning_rate/qwen_vl_interface': 9.001368977217577e-06, 'learning_rate/action_model': 9.001368977217577e-05, 'epoch': 2.19}) +[2026-06-19 21:51:24,620][starVLA.training.train_starvla][INFO] - Step 1131, Loss: {'train/loss': 0.16408255696296692, 'train/grad_norm_pre_clip': 10.674535751342773, 'timing/data': 0.006935251876711845, 'timing/model': 2.7559838965535164, 'learning_rate/qwen_vl_interface': 8.999500103604538e-06, 'learning_rate/action_model': 8.999500103604536e-05, 'epoch': 2.19}) +[2026-06-19 21:51:35,324][starVLA.training.train_starvla][INFO] - Step 1132, Loss: {'train/loss': 0.3561324179172516, 'train/grad_norm_pre_clip': 11.12852668762207, 'timing/data': 0.006753154098987579, 'timing/model': 2.7108229734003544, 'learning_rate/qwen_vl_interface': 8.997629688713639e-06, 'learning_rate/action_model': 8.997629688713638e-05, 'epoch': 2.19}) +[2026-06-19 21:51:46,141][starVLA.training.train_starvla][INFO] - Step 1133, Loss: {'train/loss': 0.2599785625934601, 'train/grad_norm_pre_clip': 8.284392356872559, 'timing/data': 0.0066169071942567825, 'timing/model': 2.7121373414993286, 'learning_rate/qwen_vl_interface': 8.995757733313733e-06, 'learning_rate/action_model': 8.995757733313732e-05, 'epoch': 2.2}) +[2026-06-19 21:51:57,018][starVLA.training.train_starvla][INFO] - Step 1134, Loss: {'train/loss': 0.34159281849861145, 'train/grad_norm_pre_clip': 8.022284507751465, 'timing/data': 0.0066333673894405365, 'timing/model': 2.7729001473635435, 'learning_rate/qwen_vl_interface': 8.993884238174314e-06, 'learning_rate/action_model': 8.993884238174314e-05, 'epoch': 2.2}) +[2026-06-19 21:52:07,712][starVLA.training.train_starvla][INFO] - Step 1135, Loss: {'train/loss': 0.13872209191322327, 'train/grad_norm_pre_clip': 6.328365325927734, 'timing/data': 0.00672687403857708, 'timing/model': 2.7647465597838163, 'learning_rate/qwen_vl_interface': 8.992009204065505e-06, 'learning_rate/action_model': 8.992009204065505e-05, 'epoch': 2.2}) +[2026-06-19 21:52:18,518][starVLA.training.train_starvla][INFO] - Step 1136, Loss: {'train/loss': 0.15743909776210785, 'train/grad_norm_pre_clip': 8.267866134643555, 'timing/data': 0.0067627765238285065, 'timing/model': 2.775212947279215, 'learning_rate/qwen_vl_interface': 8.990132631758058e-06, 'learning_rate/action_model': 8.990132631758059e-05, 'epoch': 2.2}) +[2026-06-19 21:52:29,322][starVLA.training.train_starvla][INFO] - Step 1137, Loss: {'train/loss': 0.15962433815002441, 'train/grad_norm_pre_clip': 6.490076541900635, 'timing/data': 0.006742943078279495, 'timing/model': 2.7749222312122583, 'learning_rate/qwen_vl_interface': 8.988254522023364e-06, 'learning_rate/action_model': 8.988254522023363e-05, 'epoch': 2.2}) +[2026-06-19 21:52:40,000][starVLA.training.train_starvla][INFO] - Step 1138, Loss: {'train/loss': 0.13981489837169647, 'train/grad_norm_pre_clip': 3.777937412261963, 'timing/data': 0.006429538130760193, 'timing/model': 2.7524391561746597, 'learning_rate/qwen_vl_interface': 8.986374875633441e-06, 'learning_rate/action_model': 8.98637487563344e-05, 'epoch': 2.21}) +[2026-06-19 21:52:50,711][starVLA.training.train_starvla][INFO] - Step 1139, Loss: {'train/loss': 0.2399263083934784, 'train/grad_norm_pre_clip': 7.687924385070801, 'timing/data': 0.005833825096487999, 'timing/model': 2.763267068192363, 'learning_rate/qwen_vl_interface': 8.984493693360944e-06, 'learning_rate/action_model': 8.984493693360943e-05, 'epoch': 2.21}) +[2026-06-19 21:53:01,602][starVLA.training.train_starvla][INFO] - Step 1140, Loss: {'train/loss': 0.14495576918125153, 'train/grad_norm_pre_clip': 3.367316722869873, 'timing/data': 0.006578076630830765, 'timing/model': 2.7550559230148792, 'learning_rate/qwen_vl_interface': 8.982610975979153e-06, 'learning_rate/action_model': 8.982610975979152e-05, 'epoch': 2.21}) +[2026-06-19 21:53:12,321][starVLA.training.train_starvla][INFO] - Step 1141, Loss: {'train/loss': 0.15834130346775055, 'train/grad_norm_pre_clip': 6.017948150634766, 'timing/data': 0.006539728492498398, 'timing/model': 2.779608752578497, 'learning_rate/qwen_vl_interface': 8.98072672426198e-06, 'learning_rate/action_model': 8.980726724261981e-05, 'epoch': 2.21}) +[2026-06-19 21:53:23,025][starVLA.training.train_starvla][INFO] - Step 1142, Loss: {'train/loss': 0.18251806497573853, 'train/grad_norm_pre_clip': 5.795281410217285, 'timing/data': 0.0067795440554618835, 'timing/model': 2.743872782215476, 'learning_rate/qwen_vl_interface': 8.978840938983977e-06, 'learning_rate/action_model': 8.978840938983977e-05, 'epoch': 2.21}) +[2026-06-19 21:53:33,791][starVLA.training.train_starvla][INFO] - Step 1143, Loss: {'train/loss': 0.12325756251811981, 'train/grad_norm_pre_clip': 3.6750829219818115, 'timing/data': 0.0069343820214271545, 'timing/model': 2.7542449701577425, 'learning_rate/qwen_vl_interface': 8.976953620920313e-06, 'learning_rate/action_model': 8.976953620920312e-05, 'epoch': 2.22}) +[2026-06-19 21:53:44,621][starVLA.training.train_starvla][INFO] - Step 1144, Loss: {'train/loss': 0.1887388527393341, 'train/grad_norm_pre_clip': 8.849418640136719, 'timing/data': 0.006653284654021263, 'timing/model': 2.784931832924485, 'learning_rate/qwen_vl_interface': 8.975064770846794e-06, 'learning_rate/action_model': 8.975064770846794e-05, 'epoch': 2.22}) +[2026-06-19 21:53:55,410][starVLA.training.train_starvla][INFO] - Step 1145, Loss: {'train/loss': 0.21266013383865356, 'train/grad_norm_pre_clip': 6.220707893371582, 'timing/data': 0.006769714877009392, 'timing/model': 2.7645891699939966, 'learning_rate/qwen_vl_interface': 8.973174389539856e-06, 'learning_rate/action_model': 8.973174389539855e-05, 'epoch': 2.22}) +[2026-06-19 21:54:06,219][starVLA.training.train_starvla][INFO] - Step 1146, Loss: {'train/loss': 0.13545896112918854, 'train/grad_norm_pre_clip': 4.946743488311768, 'timing/data': 0.005801023915410042, 'timing/model': 2.78199427947402, 'learning_rate/qwen_vl_interface': 8.971282477776563e-06, 'learning_rate/action_model': 8.971282477776563e-05, 'epoch': 2.22}) +[2026-06-19 21:54:16,940][starVLA.training.train_starvla][INFO] - Step 1147, Loss: {'train/loss': 0.12402522563934326, 'train/grad_norm_pre_clip': 4.950476169586182, 'timing/data': 0.00619950145483017, 'timing/model': 2.7140585482120514, 'learning_rate/qwen_vl_interface': 8.969389036334609e-06, 'learning_rate/action_model': 8.969389036334609e-05, 'epoch': 2.22}) +[2026-06-19 21:54:27,718][starVLA.training.train_starvla][INFO] - Step 1148, Loss: {'train/loss': 0.0899936854839325, 'train/grad_norm_pre_clip': 3.6920602321624756, 'timing/data': 0.006734514608979225, 'timing/model': 2.7703594136983156, 'learning_rate/qwen_vl_interface': 8.967494065992315e-06, 'learning_rate/action_model': 8.967494065992315e-05, 'epoch': 2.22}) +[2026-06-19 21:54:38,439][starVLA.training.train_starvla][INFO] - Step 1149, Loss: {'train/loss': 0.14568662643432617, 'train/grad_norm_pre_clip': 6.009234428405762, 'timing/data': 0.006764274090528488, 'timing/model': 2.7127819415181875, 'learning_rate/qwen_vl_interface': 8.965597567528634e-06, 'learning_rate/action_model': 8.965597567528633e-05, 'epoch': 2.23}) +[2026-06-19 21:54:49,314][starVLA.training.train_starvla][INFO] - Step 1150, Loss: {'train/loss': 0.11003979295492172, 'train/grad_norm_pre_clip': 5.659023761749268, 'timing/data': 0.00646720826625824, 'timing/model': 2.780081467702985, 'learning_rate/qwen_vl_interface': 8.963699541723143e-06, 'learning_rate/action_model': 8.963699541723142e-05, 'epoch': 2.23}) +[2026-06-19 21:55:00,149][starVLA.training.train_starvla][INFO] - Step 1151, Loss: {'train/loss': 0.4664593040943146, 'train/grad_norm_pre_clip': 16.556568145751953, 'timing/data': 0.005876453593373299, 'timing/model': 2.871108079329133, 'learning_rate/qwen_vl_interface': 8.961799989356048e-06, 'learning_rate/action_model': 8.961799989356048e-05, 'epoch': 2.23}) +[2026-06-19 21:55:10,807][starVLA.training.train_starvla][INFO] - Step 1152, Loss: {'train/loss': 0.11661013215780258, 'train/grad_norm_pre_clip': 7.222687244415283, 'timing/data': 0.0059416815638542175, 'timing/model': 2.7420289013534784, 'learning_rate/qwen_vl_interface': 8.959898911208185e-06, 'learning_rate/action_model': 8.959898911208185e-05, 'epoch': 2.23}) +[2026-06-19 21:55:21,621][starVLA.training.train_starvla][INFO] - Step 1153, Loss: {'train/loss': 0.18191441893577576, 'train/grad_norm_pre_clip': 6.131049633026123, 'timing/data': 0.006571425125002861, 'timing/model': 2.7767046820372343, 'learning_rate/qwen_vl_interface': 8.957996308061015e-06, 'learning_rate/action_model': 8.957996308061014e-05, 'epoch': 2.23}) +[2026-06-19 21:55:32,314][starVLA.training.train_starvla][INFO] - Step 1154, Loss: {'train/loss': 0.1550832986831665, 'train/grad_norm_pre_clip': 10.286359786987305, 'timing/data': 0.006141010671854019, 'timing/model': 2.7779957968741655, 'learning_rate/qwen_vl_interface': 8.956092180696626e-06, 'learning_rate/action_model': 8.956092180696625e-05, 'epoch': 2.24}) +[2026-06-19 21:55:43,020][starVLA.training.train_starvla][INFO] - Step 1155, Loss: {'train/loss': 0.1259460300207138, 'train/grad_norm_pre_clip': 6.166821002960205, 'timing/data': 0.006128361448645592, 'timing/model': 2.7666639909148216, 'learning_rate/qwen_vl_interface': 8.95418652989773e-06, 'learning_rate/action_model': 8.95418652989773e-05, 'epoch': 2.24}) +[2026-06-19 21:55:53,880][starVLA.training.train_starvla][INFO] - Step 1156, Loss: {'train/loss': 0.16443303227424622, 'train/grad_norm_pre_clip': 9.961315155029297, 'timing/data': 0.006547104567289352, 'timing/model': 2.7449831441044807, 'learning_rate/qwen_vl_interface': 8.952279356447675e-06, 'learning_rate/action_model': 8.952279356447675e-05, 'epoch': 2.24}) +[2026-06-19 21:56:04,611][starVLA.training.train_starvla][INFO] - Step 1157, Loss: {'train/loss': 0.0774182677268982, 'train/grad_norm_pre_clip': 7.52142333984375, 'timing/data': 0.006716081872582436, 'timing/model': 2.7201864048838615, 'learning_rate/qwen_vl_interface': 8.950370661130421e-06, 'learning_rate/action_model': 8.950370661130422e-05, 'epoch': 2.24}) +[2026-06-19 21:56:15,242][starVLA.training.train_starvla][INFO] - Step 1158, Loss: {'train/loss': 0.43090248107910156, 'train/grad_norm_pre_clip': 14.279502868652344, 'timing/data': 0.0066410843282938, 'timing/model': 2.687697548419237, 'learning_rate/qwen_vl_interface': 8.948460444730566e-06, 'learning_rate/action_model': 8.948460444730566e-05, 'epoch': 2.24}) +[2026-06-19 21:56:26,015][starVLA.training.train_starvla][INFO] - Step 1159, Loss: {'train/loss': 0.10882294923067093, 'train/grad_norm_pre_clip': 6.9942402839660645, 'timing/data': 0.006108580157160759, 'timing/model': 2.781001679599285, 'learning_rate/qwen_vl_interface': 8.946548708033324e-06, 'learning_rate/action_model': 8.946548708033323e-05, 'epoch': 2.25}) +[2026-06-19 21:56:36,720][starVLA.training.train_starvla][INFO] - Step 1160, Loss: {'train/loss': 0.19627633690834045, 'train/grad_norm_pre_clip': 15.202190399169922, 'timing/data': 0.0064102280884981155, 'timing/model': 2.7855953443795443, 'learning_rate/qwen_vl_interface': 8.944635451824538e-06, 'learning_rate/action_model': 8.944635451824538e-05, 'epoch': 2.25}) +[2026-06-19 21:56:47,352][starVLA.training.train_starvla][INFO] - Step 1161, Loss: {'train/loss': 0.04173244163393974, 'train/grad_norm_pre_clip': 2.0080573558807373, 'timing/data': 0.006514174863696098, 'timing/model': 2.6797067523002625, 'learning_rate/qwen_vl_interface': 8.942720676890677e-06, 'learning_rate/action_model': 8.942720676890677e-05, 'epoch': 2.25}) +[2026-06-19 21:56:57,941][starVLA.training.train_starvla][INFO] - Step 1162, Loss: {'train/loss': 0.304621160030365, 'train/grad_norm_pre_clip': 17.948671340942383, 'timing/data': 0.006709994748234749, 'timing/model': 2.6767910700291395, 'learning_rate/qwen_vl_interface': 8.940804384018833e-06, 'learning_rate/action_model': 8.940804384018832e-05, 'epoch': 2.25}) +[2026-06-19 21:57:08,910][starVLA.training.train_starvla][INFO] - Step 1163, Loss: {'train/loss': 0.03094124235212803, 'train/grad_norm_pre_clip': 1.9088088274002075, 'timing/data': 0.006432346999645233, 'timing/model': 2.7687625885009766, 'learning_rate/qwen_vl_interface': 8.93888657399672e-06, 'learning_rate/action_model': 8.93888657399672e-05, 'epoch': 2.25}) +[2026-06-19 21:57:19,672][starVLA.training.train_starvla][INFO] - Step 1164, Loss: {'train/loss': 0.192828968167305, 'train/grad_norm_pre_clip': 10.585870742797852, 'timing/data': 0.006647935137152672, 'timing/model': 2.7249883599579334, 'learning_rate/qwen_vl_interface': 8.936967247612678e-06, 'learning_rate/action_model': 8.936967247612678e-05, 'epoch': 2.26}) +[2026-06-19 21:57:30,420][starVLA.training.train_starvla][INFO] - Step 1165, Loss: {'train/loss': 0.0826563686132431, 'train/grad_norm_pre_clip': 4.927108287811279, 'timing/data': 0.006458137184381485, 'timing/model': 2.7565873246639967, 'learning_rate/qwen_vl_interface': 8.93504640565567e-06, 'learning_rate/action_model': 8.93504640565567e-05, 'epoch': 2.26}) +[2026-06-19 21:57:41,195][starVLA.training.train_starvla][INFO] - Step 1166, Loss: {'train/loss': 0.1649753302335739, 'train/grad_norm_pre_clip': 11.068206787109375, 'timing/data': 0.006699513643980026, 'timing/model': 2.7638213578611612, 'learning_rate/qwen_vl_interface': 8.933124048915278e-06, 'learning_rate/action_model': 8.933124048915278e-05, 'epoch': 2.26}) +[2026-06-19 21:57:51,800][starVLA.training.train_starvla][INFO] - Step 1167, Loss: {'train/loss': 0.16520650684833527, 'train/grad_norm_pre_clip': 16.336896896362305, 'timing/data': 0.006554365158081055, 'timing/model': 2.730261769145727, 'learning_rate/qwen_vl_interface': 8.931200178181715e-06, 'learning_rate/action_model': 8.931200178181715e-05, 'epoch': 2.26}) +[2026-06-19 21:58:02,522][starVLA.training.train_starvla][INFO] - Step 1168, Loss: {'train/loss': 0.049067508429288864, 'train/grad_norm_pre_clip': 12.984491348266602, 'timing/data': 0.006590254604816437, 'timing/model': 2.7777411956340075, 'learning_rate/qwen_vl_interface': 8.929274794245812e-06, 'learning_rate/action_model': 8.929274794245812e-05, 'epoch': 2.26}) +[2026-06-19 21:58:13,254][starVLA.training.train_starvla][INFO] - Step 1169, Loss: {'train/loss': 0.31094077229499817, 'train/grad_norm_pre_clip': 13.623383522033691, 'timing/data': 0.00669395737349987, 'timing/model': 2.7167893201112747, 'learning_rate/qwen_vl_interface': 8.927347897899018e-06, 'learning_rate/action_model': 8.927347897899016e-05, 'epoch': 2.27}) +[2026-06-19 21:58:23,823][starVLA.training.train_starvla][INFO] - Step 1170, Loss: {'train/loss': 0.3318755328655243, 'train/grad_norm_pre_clip': 42.5531120300293, 'timing/data': 0.006572246551513672, 'timing/model': 2.7526896726340055, 'learning_rate/qwen_vl_interface': 8.925419489933407e-06, 'learning_rate/action_model': 8.925419489933408e-05, 'epoch': 2.27}) +[2026-06-19 21:58:34,524][starVLA.training.train_starvla][INFO] - Step 1171, Loss: {'train/loss': 0.07765349745750427, 'train/grad_norm_pre_clip': 6.812607288360596, 'timing/data': 0.006680864840745926, 'timing/model': 2.710389144718647, 'learning_rate/qwen_vl_interface': 8.923489571141679e-06, 'learning_rate/action_model': 8.923489571141679e-05, 'epoch': 2.27}) +[2026-06-19 21:58:45,313][starVLA.training.train_starvla][INFO] - Step 1172, Loss: {'train/loss': 0.23533737659454346, 'train/grad_norm_pre_clip': 14.09490966796875, 'timing/data': 0.006727451458573341, 'timing/model': 2.775957642123103, 'learning_rate/qwen_vl_interface': 8.921558142317147e-06, 'learning_rate/action_model': 8.921558142317147e-05, 'epoch': 2.27}) +[2026-06-19 21:58:55,919][starVLA.training.train_starvla][INFO] - Step 1173, Loss: {'train/loss': 0.27839136123657227, 'train/grad_norm_pre_clip': 12.128270149230957, 'timing/data': 0.006801333278417587, 'timing/model': 2.7480758894234896, 'learning_rate/qwen_vl_interface': 8.91962520425375e-06, 'learning_rate/action_model': 8.91962520425375e-05, 'epoch': 2.27}) +[2026-06-19 21:59:06,663][starVLA.training.train_starvla][INFO] - Step 1174, Loss: {'train/loss': 0.40691104531288147, 'train/grad_norm_pre_clip': 13.255921363830566, 'timing/data': 0.006601724773645401, 'timing/model': 2.8450736347585917, 'learning_rate/qwen_vl_interface': 8.917690757746044e-06, 'learning_rate/action_model': 8.917690757746044e-05, 'epoch': 2.28}) +[2026-06-19 21:59:17,340][starVLA.training.train_starvla][INFO] - Step 1175, Loss: {'train/loss': 0.5296177864074707, 'train/grad_norm_pre_clip': 31.2700138092041, 'timing/data': 0.006815331056714058, 'timing/model': 2.698268612846732, 'learning_rate/qwen_vl_interface': 8.915754803589208e-06, 'learning_rate/action_model': 8.915754803589209e-05, 'epoch': 2.28}) +[2026-06-19 21:59:27,844][starVLA.training.train_starvla][INFO] - Step 1176, Loss: {'train/loss': 0.09571776539087296, 'train/grad_norm_pre_clip': 5.731822967529297, 'timing/data': 0.006717804819345474, 'timing/model': 2.6757171507924795, 'learning_rate/qwen_vl_interface': 8.913817342579042e-06, 'learning_rate/action_model': 8.913817342579042e-05, 'epoch': 2.28}) +[2026-06-19 21:59:38,344][starVLA.training.train_starvla][INFO] - Step 1177, Loss: {'train/loss': 0.2816718518733978, 'train/grad_norm_pre_clip': 13.803377151489258, 'timing/data': 0.006824754178524017, 'timing/model': 2.6764813903719187, 'learning_rate/qwen_vl_interface': 8.91187837551196e-06, 'learning_rate/action_model': 8.911878375511961e-05, 'epoch': 2.28}) +[2026-06-19 21:59:48,848][starVLA.training.train_starvla][INFO] - Step 1178, Loss: {'train/loss': 0.13111427426338196, 'train/grad_norm_pre_clip': 7.88117790222168, 'timing/data': 0.006973378360271454, 'timing/model': 2.654498590156436, 'learning_rate/qwen_vl_interface': 8.909937903185e-06, 'learning_rate/action_model': 8.909937903185e-05, 'epoch': 2.28}) +[2026-06-19 21:59:59,343][starVLA.training.train_starvla][INFO] - Step 1179, Loss: {'train/loss': 0.08262178301811218, 'train/grad_norm_pre_clip': 4.553964138031006, 'timing/data': 0.006783151999115944, 'timing/model': 2.677205104380846, 'learning_rate/qwen_vl_interface': 8.907995926395814e-06, 'learning_rate/action_model': 8.907995926395815e-05, 'epoch': 2.29}) +[2026-06-19 22:00:09,840][starVLA.training.train_starvla][INFO] - Step 1180, Loss: {'train/loss': 0.05752638727426529, 'train/grad_norm_pre_clip': 6.113757133483887, 'timing/data': 0.006692053750157356, 'timing/model': 2.674075808376074, 'learning_rate/qwen_vl_interface': 8.90605244594268e-06, 'learning_rate/action_model': 8.906052445942679e-05, 'epoch': 2.29}) +[2026-06-19 22:00:20,344][starVLA.training.train_starvla][INFO] - Step 1181, Loss: {'train/loss': 0.18294771015644073, 'train/grad_norm_pre_clip': 9.11444091796875, 'timing/data': 0.006960630416870117, 'timing/model': 2.6799581423401833, 'learning_rate/qwen_vl_interface': 8.904107462624484e-06, 'learning_rate/action_model': 8.904107462624483e-05, 'epoch': 2.29}) +[2026-06-19 22:00:30,847][starVLA.training.train_starvla][INFO] - Step 1182, Loss: {'train/loss': 0.3962898552417755, 'train/grad_norm_pre_clip': 7.904050350189209, 'timing/data': 0.006820132955908775, 'timing/model': 2.678094232454896, 'learning_rate/qwen_vl_interface': 8.902160977240739e-06, 'learning_rate/action_model': 8.902160977240739e-05, 'epoch': 2.29}) +[2026-06-19 22:00:41,341][starVLA.training.train_starvla][INFO] - Step 1183, Loss: {'train/loss': 0.2351311594247818, 'train/grad_norm_pre_clip': 8.332719802856445, 'timing/data': 0.006865711882710457, 'timing/model': 2.675833337008953, 'learning_rate/qwen_vl_interface': 8.90021299059157e-06, 'learning_rate/action_model': 8.90021299059157e-05, 'epoch': 2.29}) +[2026-06-19 22:00:51,846][starVLA.training.train_starvla][INFO] - Step 1184, Loss: {'train/loss': 0.20389653742313385, 'train/grad_norm_pre_clip': 7.3481950759887695, 'timing/data': 0.006851194426417351, 'timing/model': 2.65653670206666, 'learning_rate/qwen_vl_interface': 8.898263503477722e-06, 'learning_rate/action_model': 8.898263503477721e-05, 'epoch': 2.29}) +[2026-06-19 22:01:02,345][starVLA.training.train_starvla][INFO] - Step 1185, Loss: {'train/loss': 0.09578597545623779, 'train/grad_norm_pre_clip': 7.713089466094971, 'timing/data': 0.006831621751189232, 'timing/model': 2.6796251479536295, 'learning_rate/qwen_vl_interface': 8.896312516700553e-06, 'learning_rate/action_model': 8.896312516700554e-05, 'epoch': 2.3}) +[2026-06-19 22:01:12,944][starVLA.training.train_starvla][INFO] - Step 1186, Loss: {'train/loss': 0.13352198898792267, 'train/grad_norm_pre_clip': 6.7578535079956055, 'timing/data': 0.0068616922944784164, 'timing/model': 2.6744847036898136, 'learning_rate/qwen_vl_interface': 8.894360031062043e-06, 'learning_rate/action_model': 8.894360031062043e-05, 'epoch': 2.3}) +[2026-06-19 22:01:23,443][starVLA.training.train_starvla][INFO] - Step 1187, Loss: {'train/loss': 0.2828826606273651, 'train/grad_norm_pre_clip': 7.614752769470215, 'timing/data': 0.006703384220600128, 'timing/model': 2.6772382613271475, 'learning_rate/qwen_vl_interface': 8.892406047364783e-06, 'learning_rate/action_model': 8.892406047364784e-05, 'epoch': 2.3}) +[2026-06-19 22:01:33,951][starVLA.training.train_starvla][INFO] - Step 1188, Loss: {'train/loss': 0.1409757137298584, 'train/grad_norm_pre_clip': 6.296018600463867, 'timing/data': 0.006543567404150963, 'timing/model': 2.6834009923040867, 'learning_rate/qwen_vl_interface': 8.890450566411984e-06, 'learning_rate/action_model': 8.890450566411984e-05, 'epoch': 2.3}) +[2026-06-19 22:01:44,439][starVLA.training.train_starvla][INFO] - Step 1189, Loss: {'train/loss': 0.232759490609169, 'train/grad_norm_pre_clip': 6.261069297790527, 'timing/data': 0.006806293502449989, 'timing/model': 2.6738127283751965, 'learning_rate/qwen_vl_interface': 8.888493589007467e-06, 'learning_rate/action_model': 8.888493589007468e-05, 'epoch': 2.3}) +[2026-06-19 22:01:54,945][starVLA.training.train_starvla][INFO] - Step 1190, Loss: {'train/loss': 0.07297711819410324, 'train/grad_norm_pre_clip': 3.268237352371216, 'timing/data': 0.006407897919416428, 'timing/model': 2.6771493572741747, 'learning_rate/qwen_vl_interface': 8.886535115955676e-06, 'learning_rate/action_model': 8.886535115955675e-05, 'epoch': 2.31}) +[2026-06-19 22:02:05,441][starVLA.training.train_starvla][INFO] - Step 1191, Loss: {'train/loss': 0.18864791095256805, 'train/grad_norm_pre_clip': 4.719696044921875, 'timing/data': 0.006519315764307976, 'timing/model': 2.6775583270937204, 'learning_rate/qwen_vl_interface': 8.884575148061662e-06, 'learning_rate/action_model': 8.88457514806166e-05, 'epoch': 2.31}) +[2026-06-19 22:02:16,133][starVLA.training.train_starvla][INFO] - Step 1192, Loss: {'train/loss': 0.18391016125679016, 'train/grad_norm_pre_clip': 6.74661922454834, 'timing/data': 0.0065335463732481, 'timing/model': 2.715366391465068, 'learning_rate/qwen_vl_interface': 8.882613686131095e-06, 'learning_rate/action_model': 8.882613686131093e-05, 'epoch': 2.31}) +[2026-06-19 22:02:26,779][starVLA.training.train_starvla][INFO] - Step 1193, Loss: {'train/loss': 0.1469944417476654, 'train/grad_norm_pre_clip': 3.9932587146759033, 'timing/data': 0.0066374558955430984, 'timing/model': 2.717869373038411, 'learning_rate/qwen_vl_interface': 8.880650730970254e-06, 'learning_rate/action_model': 8.880650730970254e-05, 'epoch': 2.31}) +[2026-06-19 22:02:37,529][starVLA.training.train_starvla][INFO] - Step 1194, Loss: {'train/loss': 0.4571206271648407, 'train/grad_norm_pre_clip': 9.364331245422363, 'timing/data': 0.006728284060955048, 'timing/model': 2.709623672068119, 'learning_rate/qwen_vl_interface': 8.878686283386042e-06, 'learning_rate/action_model': 8.878686283386042e-05, 'epoch': 2.31}) +[2026-06-19 22:02:48,368][starVLA.training.train_starvla][INFO] - Step 1195, Loss: {'train/loss': 0.16554683446884155, 'train/grad_norm_pre_clip': 9.573542594909668, 'timing/data': 0.0069047920405864716, 'timing/model': 2.7214375343173742, 'learning_rate/qwen_vl_interface': 8.876720344185967e-06, 'learning_rate/action_model': 8.876720344185967e-05, 'epoch': 2.32}) +[2026-06-19 22:02:59,176][starVLA.training.train_starvla][INFO] - Step 1196, Loss: {'train/loss': 0.3148607313632965, 'train/grad_norm_pre_clip': 5.804676532745361, 'timing/data': 0.006444726139307022, 'timing/model': 2.729569710791111, 'learning_rate/qwen_vl_interface': 8.874752914178149e-06, 'learning_rate/action_model': 8.87475291417815e-05, 'epoch': 2.32}) +[2026-06-19 22:03:09,983][starVLA.training.train_starvla][INFO] - Step 1197, Loss: {'train/loss': 0.05071099475026131, 'train/grad_norm_pre_clip': 3.0996253490448, 'timing/data': 0.00690014474093914, 'timing/model': 2.8362331204116344, 'learning_rate/qwen_vl_interface': 8.87278399417133e-06, 'learning_rate/action_model': 8.872783994171329e-05, 'epoch': 2.32}) +[2026-06-19 22:03:20,681][starVLA.training.train_starvla][INFO] - Step 1198, Loss: {'train/loss': 0.17134009301662445, 'train/grad_norm_pre_clip': 7.967342853546143, 'timing/data': 0.00666157528758049, 'timing/model': 2.728479055687785, 'learning_rate/qwen_vl_interface': 8.870813584974852e-06, 'learning_rate/action_model': 8.870813584974852e-05, 'epoch': 2.32}) +[2026-06-19 22:03:31,426][starVLA.training.train_starvla][INFO] - Step 1199, Loss: {'train/loss': 0.3957805633544922, 'train/grad_norm_pre_clip': 8.414802551269531, 'timing/data': 0.006744613870978355, 'timing/model': 2.790283877402544, 'learning_rate/qwen_vl_interface': 8.868841687398683e-06, 'learning_rate/action_model': 8.868841687398683e-05, 'epoch': 2.32}) +[2026-06-19 22:03:42,166][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1200 +[2026-06-19 22:11:14,487][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1200 +[2026-06-19 22:11:14,488][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 1200 +[2026-06-19 22:19:24,724][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 1200 +[2026-06-19 22:19:24,726][starVLA.training.train_starvla][INFO] - Step 1200, Loss: {'train/loss': 0.09764586389064789, 'train/grad_norm_pre_clip': 4.887906551361084, 'eval/loss': 0.511775016784668, 'eval/latency_0/loss': 0.21061931550502777, 'eval/latency_1/loss': 0.09444146603345871, 'eval/latency_2/loss': 0.6737809777259827, 'eval/latency_3/loss': 0.6513047814369202, 'eval/latency_4/loss': 0.9167462587356567, 'eval/flappy/loss': 0.511775016784668, 'eval/flappy/latency_0/loss': 0.21061931550502777, 'eval/flappy/latency_1/loss': 0.09444146603345871, 'eval/flappy/latency_2/loss': 0.6737809777259827, 'eval/flappy/latency_3/loss': 0.6513047814369202, 'eval/flappy/latency_4/loss': 0.9167462587356567, 'eval/flap_precision': 0.44141689373297005, 'eval/flap_recall': 0.5890909090909091, 'eval/flap_f1': 0.5046728971962616, 'eval/flap_rate_model': 0.10194444444444445, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.5046728971962616, 'eval/latency_0/flap_precision': 0.44141689373297005, 'eval/latency_0/flap_recall': 0.5890909090909091, 'eval/latency_0/flap_f1': 0.5046728971962616, 'eval/latency_0/flap_rate_model': 0.10194444444444445, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.5046728971962616, 'timing/data': 0.0067559946328401566, 'timing/model': 2.7310044821351767, 'learning_rate/qwen_vl_interface': 8.86686830225339e-06, 'learning_rate/action_model': 8.866868302253389e-05, 'epoch': 2.33}) +[2026-06-19 22:19:24,726][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_1200_state +[2026-06-19 22:19:24,754][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-19 22:19:30,497][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_1200_state/model.safetensors +[2026-06-19 22:19:44,600][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_1200_state/optimizer.bin +[2026-06-19 22:19:44,601][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_1200_state/sampler.bin +[2026-06-19 22:19:44,601][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_1200_state/sampler_1.bin +[2026-06-19 22:19:44,602][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_1200_state/random_states_0.pkl +[2026-06-19 22:19:44,728][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1200 +[2026-06-19 22:19:44,729][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-19 22:19:44,742][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-19 22:19:55,427][starVLA.training.train_starvla][INFO] - Step 1201, Loss: {'train/loss': 0.14781662821769714, 'train/grad_norm_pre_clip': 5.020946025848389, 'timing/data': 0.006346499547362328, 'timing/model': 2.677792703732848, 'learning_rate/qwen_vl_interface': 8.864893430350159e-06, 'learning_rate/action_model': 8.864893430350159e-05, 'epoch': 2.33}) +[2026-06-19 22:20:06,006][starVLA.training.train_starvla][INFO] - Step 1202, Loss: {'train/loss': 0.17902682721614838, 'train/grad_norm_pre_clip': 5.4994378089904785, 'timing/data': 0.005925122648477554, 'timing/model': 2.7356724496930838, 'learning_rate/qwen_vl_interface': 8.862917072500785e-06, 'learning_rate/action_model': 8.862917072500784e-05, 'epoch': 2.33}) +[2026-06-19 22:20:16,830][starVLA.training.train_starvla][INFO] - Step 1203, Loss: {'train/loss': 0.2719801068305969, 'train/grad_norm_pre_clip': 6.8553948402404785, 'timing/data': 0.006656864657998085, 'timing/model': 2.7018646597862244, 'learning_rate/qwen_vl_interface': 8.860939229517674e-06, 'learning_rate/action_model': 8.860939229517674e-05, 'epoch': 2.33}) +[2026-06-19 22:20:27,454][starVLA.training.train_starvla][INFO] - Step 1204, Loss: {'train/loss': 0.24386535584926605, 'train/grad_norm_pre_clip': 10.213166236877441, 'timing/data': 0.005554208531975746, 'timing/model': 2.719858216121793, 'learning_rate/qwen_vl_interface': 8.858959902213844e-06, 'learning_rate/action_model': 8.858959902213845e-05, 'epoch': 2.33}) +[2026-06-19 22:20:38,134][starVLA.training.train_starvla][INFO] - Step 1205, Loss: {'train/loss': 0.143144428730011, 'train/grad_norm_pre_clip': 4.949001312255859, 'timing/data': 0.00659879669547081, 'timing/model': 2.6954826414585114, 'learning_rate/qwen_vl_interface': 8.856979091402921e-06, 'learning_rate/action_model': 8.85697909140292e-05, 'epoch': 2.34}) +[2026-06-19 22:20:48,638][starVLA.training.train_starvla][INFO] - Step 1206, Loss: {'train/loss': 0.29390233755111694, 'train/grad_norm_pre_clip': 7.964905261993408, 'timing/data': 0.006521187722682953, 'timing/model': 2.675114918500185, 'learning_rate/qwen_vl_interface': 8.854996797899139e-06, 'learning_rate/action_model': 8.854996797899139e-05, 'epoch': 2.34}) +[2026-06-19 22:20:59,157][starVLA.training.train_starvla][INFO] - Step 1207, Loss: {'train/loss': 0.17167377471923828, 'train/grad_norm_pre_clip': 4.91272497177124, 'timing/data': 0.005768716335296631, 'timing/model': 2.697210965678096, 'learning_rate/qwen_vl_interface': 8.853013022517348e-06, 'learning_rate/action_model': 8.853013022517348e-05, 'epoch': 2.34}) +[2026-06-19 22:21:09,940][starVLA.training.train_starvla][INFO] - Step 1208, Loss: {'train/loss': 0.21078120172023773, 'train/grad_norm_pre_clip': 10.851481437683105, 'timing/data': 0.008545635268092155, 'timing/model': 2.6882190946489573, 'learning_rate/qwen_vl_interface': 8.851027766073002e-06, 'learning_rate/action_model': 8.851027766073001e-05, 'epoch': 2.34}) +[2026-06-19 22:21:20,717][starVLA.training.train_starvla][INFO] - Step 1209, Loss: {'train/loss': 0.20140200853347778, 'train/grad_norm_pre_clip': 7.659865856170654, 'timing/data': 0.006710944697260857, 'timing/model': 2.7799574434757233, 'learning_rate/qwen_vl_interface': 8.84904102938216e-06, 'learning_rate/action_model': 8.84904102938216e-05, 'epoch': 2.34}) +[2026-06-19 22:21:31,310][starVLA.training.train_starvla][INFO] - Step 1210, Loss: {'train/loss': 0.23518021404743195, 'train/grad_norm_pre_clip': 6.309005260467529, 'timing/data': 0.006605144590139389, 'timing/model': 2.708439040929079, 'learning_rate/qwen_vl_interface': 8.8470528132615e-06, 'learning_rate/action_model': 8.8470528132615e-05, 'epoch': 2.35}) +[2026-06-19 22:21:42,315][starVLA.training.train_starvla][INFO] - Step 1211, Loss: {'train/loss': 0.14366939663887024, 'train/grad_norm_pre_clip': 5.873898506164551, 'timing/data': 0.0063402000814676285, 'timing/model': 2.7741783875972033, 'learning_rate/qwen_vl_interface': 8.845063118528303e-06, 'learning_rate/action_model': 8.845063118528304e-05, 'epoch': 2.35}) +[2026-06-19 22:21:53,037][starVLA.training.train_starvla][INFO] - Step 1212, Loss: {'train/loss': 0.30403363704681396, 'train/grad_norm_pre_clip': 12.03905963897705, 'timing/data': 0.00645308755338192, 'timing/model': 2.7111448626965284, 'learning_rate/qwen_vl_interface': 8.843071946000455e-06, 'learning_rate/action_model': 8.843071946000456e-05, 'epoch': 2.35}) +[2026-06-19 22:22:03,814][starVLA.training.train_starvla][INFO] - Step 1213, Loss: {'train/loss': 0.32470354437828064, 'train/grad_norm_pre_clip': 17.263416290283203, 'timing/data': 0.006475446745753288, 'timing/model': 2.7726063672453165, 'learning_rate/qwen_vl_interface': 8.841079296496451e-06, 'learning_rate/action_model': 8.84107929649645e-05, 'epoch': 2.35}) +[2026-06-19 22:22:14,612][starVLA.training.train_starvla][INFO] - Step 1214, Loss: {'train/loss': 0.2481350153684616, 'train/grad_norm_pre_clip': 9.387040138244629, 'timing/data': 0.006339637562632561, 'timing/model': 2.849205456674099, 'learning_rate/qwen_vl_interface': 8.839085170835396e-06, 'learning_rate/action_model': 8.839085170835395e-05, 'epoch': 2.35}) +[2026-06-19 22:22:25,239][starVLA.training.train_starvla][INFO] - Step 1215, Loss: {'train/loss': 0.12904463708400726, 'train/grad_norm_pre_clip': 3.2185754776000977, 'timing/data': 0.006468236446380615, 'timing/model': 2.702204482629895, 'learning_rate/qwen_vl_interface': 8.837089569837e-06, 'learning_rate/action_model': 8.837089569836998e-05, 'epoch': 2.35}) +[2026-06-19 22:22:35,739][starVLA.training.train_starvla][INFO] - Step 1216, Loss: {'train/loss': 0.19773419201374054, 'train/grad_norm_pre_clip': 5.737751007080078, 'timing/data': 0.006343070417642593, 'timing/model': 2.6742529962211847, 'learning_rate/qwen_vl_interface': 8.835092494321575e-06, 'learning_rate/action_model': 8.835092494321575e-05, 'epoch': 2.36}) +[2026-06-19 22:22:46,238][starVLA.training.train_starvla][INFO] - Step 1217, Loss: {'train/loss': 0.14602965116500854, 'train/grad_norm_pre_clip': 4.282960891723633, 'timing/data': 0.006397968158125877, 'timing/model': 2.6778391487896442, 'learning_rate/qwen_vl_interface': 8.833093945110048e-06, 'learning_rate/action_model': 8.833093945110047e-05, 'epoch': 2.36}) +[2026-06-19 22:22:56,950][starVLA.training.train_starvla][INFO] - Step 1218, Loss: {'train/loss': 0.18048255145549774, 'train/grad_norm_pre_clip': 4.966446399688721, 'timing/data': 0.006621817126870155, 'timing/model': 2.640742300078273, 'learning_rate/qwen_vl_interface': 8.831093923023946e-06, 'learning_rate/action_model': 8.831093923023945e-05, 'epoch': 2.36}) +[2026-06-19 22:23:07,481][starVLA.training.train_starvla][INFO] - Step 1219, Loss: {'train/loss': 0.16882005333900452, 'train/grad_norm_pre_clip': 7.613949298858643, 'timing/data': 0.006312558427453041, 'timing/model': 2.714973183348775, 'learning_rate/qwen_vl_interface': 8.8290924288854e-06, 'learning_rate/action_model': 8.8290924288854e-05, 'epoch': 2.36}) +[2026-06-19 22:23:17,942][starVLA.training.train_starvla][INFO] - Step 1220, Loss: {'train/loss': 0.14769349992275238, 'train/grad_norm_pre_clip': 6.107253074645996, 'timing/data': 0.006348719820380211, 'timing/model': 2.6845339350402355, 'learning_rate/qwen_vl_interface': 8.827089463517153e-06, 'learning_rate/action_model': 8.827089463517153e-05, 'epoch': 2.36}) +[2026-06-19 22:23:28,540][starVLA.training.train_starvla][INFO] - Step 1221, Loss: {'train/loss': 0.14975528419017792, 'train/grad_norm_pre_clip': 8.317562103271484, 'timing/data': 0.006325749680399895, 'timing/model': 2.6722280718386173, 'learning_rate/qwen_vl_interface': 8.825085027742545e-06, 'learning_rate/action_model': 8.825085027742544e-05, 'epoch': 2.37}) +[2026-06-19 22:23:39,238][starVLA.training.train_starvla][INFO] - Step 1222, Loss: {'train/loss': 0.17480091750621796, 'train/grad_norm_pre_clip': 10.606436729431152, 'timing/data': 0.006455596536397934, 'timing/model': 2.6714437399059534, 'learning_rate/qwen_vl_interface': 8.823079122385523e-06, 'learning_rate/action_model': 8.823079122385523e-05, 'epoch': 2.37}) +[2026-06-19 22:23:49,740][starVLA.training.train_starvla][INFO] - Step 1223, Loss: {'train/loss': 0.2713044583797455, 'train/grad_norm_pre_clip': 14.38650131225586, 'timing/data': 0.006468538194894791, 'timing/model': 2.6783786732703447, 'learning_rate/qwen_vl_interface': 8.821071748270644e-06, 'learning_rate/action_model': 8.821071748270644e-05, 'epoch': 2.37}) +[2026-06-19 22:24:00,458][starVLA.training.train_starvla][INFO] - Step 1224, Loss: {'train/loss': 0.17915846407413483, 'train/grad_norm_pre_clip': 10.249011993408203, 'timing/data': 0.006359351798892021, 'timing/model': 2.892487982288003, 'learning_rate/qwen_vl_interface': 8.81906290622306e-06, 'learning_rate/action_model': 8.81906290622306e-05, 'epoch': 2.37}) +[2026-06-19 22:24:11,143][starVLA.training.train_starvla][INFO] - Step 1225, Loss: {'train/loss': 0.08026205003261566, 'train/grad_norm_pre_clip': 4.068319320678711, 'timing/data': 0.006497954949736595, 'timing/model': 2.7026221454143524, 'learning_rate/qwen_vl_interface': 8.81705259706853e-06, 'learning_rate/action_model': 8.817052597068529e-05, 'epoch': 2.37}) +[2026-06-19 22:24:21,815][starVLA.training.train_starvla][INFO] - Step 1226, Loss: {'train/loss': 0.04072560369968414, 'train/grad_norm_pre_clip': 3.506603717803955, 'timing/data': 0.006511947140097618, 'timing/model': 2.7301229182630777, 'learning_rate/qwen_vl_interface': 8.815040821633416e-06, 'learning_rate/action_model': 8.815040821633416e-05, 'epoch': 2.38}) +[2026-06-19 22:24:32,443][starVLA.training.train_starvla][INFO] - Step 1227, Loss: {'train/loss': 0.13814549148082733, 'train/grad_norm_pre_clip': 13.508907318115234, 'timing/data': 0.006698984652757645, 'timing/model': 2.68065476231277, 'learning_rate/qwen_vl_interface': 8.813027580744685e-06, 'learning_rate/action_model': 8.813027580744684e-05, 'epoch': 2.38}) +[2026-06-19 22:24:43,015][starVLA.training.train_starvla][INFO] - Step 1228, Loss: {'train/loss': 0.3098078966140747, 'train/grad_norm_pre_clip': 17.089719772338867, 'timing/data': 0.006268680095672607, 'timing/model': 2.7465052474290133, 'learning_rate/qwen_vl_interface': 8.8110128752299e-06, 'learning_rate/action_model': 8.8110128752299e-05, 'epoch': 2.38}) +[2026-06-19 22:24:53,808][starVLA.training.train_starvla][INFO] - Step 1229, Loss: {'train/loss': 0.06071055307984352, 'train/grad_norm_pre_clip': 4.943051338195801, 'timing/data': 0.006316939368844032, 'timing/model': 2.772567531093955, 'learning_rate/qwen_vl_interface': 8.808996705917235e-06, 'learning_rate/action_model': 8.808996705917235e-05, 'epoch': 2.38}) +[2026-06-19 22:25:04,528][starVLA.training.train_starvla][INFO] - Step 1230, Loss: {'train/loss': 0.1494779735803604, 'train/grad_norm_pre_clip': 8.504426002502441, 'timing/data': 0.006557527929544449, 'timing/model': 2.7581172194331884, 'learning_rate/qwen_vl_interface': 8.80697907363546e-06, 'learning_rate/action_model': 8.80697907363546e-05, 'epoch': 2.38}) +[2026-06-19 22:25:15,222][starVLA.training.train_starvla][INFO] - Step 1231, Loss: {'train/loss': 0.27992552518844604, 'train/grad_norm_pre_clip': 14.640761375427246, 'timing/data': 0.00634189136326313, 'timing/model': 2.7071163188666105, 'learning_rate/qwen_vl_interface': 8.804959979213945e-06, 'learning_rate/action_model': 8.804959979213945e-05, 'epoch': 2.39}) +[2026-06-19 22:25:25,740][starVLA.training.train_starvla][INFO] - Step 1232, Loss: {'train/loss': 0.1095711886882782, 'train/grad_norm_pre_clip': 16.001314163208008, 'timing/data': 0.0066143181174993515, 'timing/model': 2.681463573127985, 'learning_rate/qwen_vl_interface': 8.802939423482667e-06, 'learning_rate/action_model': 8.802939423482666e-05, 'epoch': 2.39}) +[2026-06-19 22:25:36,241][starVLA.training.train_starvla][INFO] - Step 1233, Loss: {'train/loss': 0.026284663006663322, 'train/grad_norm_pre_clip': 3.024714946746826, 'timing/data': 0.006549978628754616, 'timing/model': 2.677898593246937, 'learning_rate/qwen_vl_interface': 8.800917407272198e-06, 'learning_rate/action_model': 8.800917407272198e-05, 'epoch': 2.39}) +[2026-06-19 22:25:46,741][starVLA.training.train_starvla][INFO] - Step 1234, Loss: {'train/loss': 0.07917297631502151, 'train/grad_norm_pre_clip': 12.208919525146484, 'timing/data': 0.006240202113986015, 'timing/model': 2.6713291071355343, 'learning_rate/qwen_vl_interface': 8.798893931413711e-06, 'learning_rate/action_model': 8.79889393141371e-05, 'epoch': 2.39}) +[2026-06-19 22:25:57,347][starVLA.training.train_starvla][INFO] - Step 1235, Loss: {'train/loss': 0.015910297632217407, 'train/grad_norm_pre_clip': 1.8457013368606567, 'timing/data': 0.006152238696813583, 'timing/model': 2.699224015697837, 'learning_rate/qwen_vl_interface': 8.796868996738985e-06, 'learning_rate/action_model': 8.796868996738984e-05, 'epoch': 2.39}) +[2026-06-19 22:26:08,040][starVLA.training.train_starvla][INFO] - Step 1236, Loss: {'train/loss': 0.24814081192016602, 'train/grad_norm_pre_clip': 28.932165145874023, 'timing/data': 0.0063532814383506775, 'timing/model': 2.7144128903746605, 'learning_rate/qwen_vl_interface': 8.794842604080389e-06, 'learning_rate/action_model': 8.794842604080389e-05, 'epoch': 2.4}) +[2026-06-19 22:26:18,745][starVLA.training.train_starvla][INFO] - Step 1237, Loss: {'train/loss': 0.12121827900409698, 'train/grad_norm_pre_clip': 17.06837272644043, 'timing/data': 0.00634412094950676, 'timing/model': 2.635992903262377, 'learning_rate/qwen_vl_interface': 8.7928147542709e-06, 'learning_rate/action_model': 8.7928147542709e-05, 'epoch': 2.4}) +[2026-06-19 22:26:29,344][starVLA.training.train_starvla][INFO] - Step 1238, Loss: {'train/loss': 0.10452894866466522, 'train/grad_norm_pre_clip': 8.557167053222656, 'timing/data': 0.006081581115722656, 'timing/model': 2.6801438126713037, 'learning_rate/qwen_vl_interface': 8.790785448144088e-06, 'learning_rate/action_model': 8.790785448144088e-05, 'epoch': 2.4}) +[2026-06-19 22:26:40,066][starVLA.training.train_starvla][INFO] - Step 1239, Loss: {'train/loss': 0.41354167461395264, 'train/grad_norm_pre_clip': 20.314697265625, 'timing/data': 0.006340719759464264, 'timing/model': 2.7240780740976334, 'learning_rate/qwen_vl_interface': 8.788754686534126e-06, 'learning_rate/action_model': 8.788754686534127e-05, 'epoch': 2.4}) +[2026-06-19 22:26:50,645][starVLA.training.train_starvla][INFO] - Step 1240, Loss: {'train/loss': 0.025482047349214554, 'train/grad_norm_pre_clip': 5.511460304260254, 'timing/data': 0.006206089630723, 'timing/model': 2.6574059668928385, 'learning_rate/qwen_vl_interface': 8.786722470275784e-06, 'learning_rate/action_model': 8.786722470275784e-05, 'epoch': 2.4}) +[2026-06-19 22:27:01,242][starVLA.training.train_starvla][INFO] - Step 1241, Loss: {'train/loss': 0.019006092101335526, 'train/grad_norm_pre_clip': 5.88983154296875, 'timing/data': 0.006455808877944946, 'timing/model': 2.70717304199934, 'learning_rate/qwen_vl_interface': 8.784688800204427e-06, 'learning_rate/action_model': 8.784688800204426e-05, 'epoch': 2.41}) +[2026-06-19 22:27:11,923][starVLA.training.train_starvla][INFO] - Step 1242, Loss: {'train/loss': 0.5547042489051819, 'train/grad_norm_pre_clip': 23.95797348022461, 'timing/data': 0.006739115342497826, 'timing/model': 2.7085852958261967, 'learning_rate/qwen_vl_interface': 8.782653677156022e-06, 'learning_rate/action_model': 8.782653677156022e-05, 'epoch': 2.41}) +[2026-06-19 22:27:22,743][starVLA.training.train_starvla][INFO] - Step 1243, Loss: {'train/loss': 0.2713925242424011, 'train/grad_norm_pre_clip': 20.30454444885254, 'timing/data': 0.0067824311554431915, 'timing/model': 2.713411334902048, 'learning_rate/qwen_vl_interface': 8.780617101967129e-06, 'learning_rate/action_model': 8.780617101967129e-05, 'epoch': 2.41}) +[2026-06-19 22:27:33,447][starVLA.training.train_starvla][INFO] - Step 1244, Loss: {'train/loss': 0.5635338425636292, 'train/grad_norm_pre_clip': 25.52276611328125, 'timing/data': 0.0062981266528368, 'timing/model': 2.648417266085744, 'learning_rate/qwen_vl_interface': 8.77857907547491e-06, 'learning_rate/action_model': 8.77857907547491e-05, 'epoch': 2.41}) +[2026-06-19 22:27:43,943][starVLA.training.train_starvla][INFO] - Step 1245, Loss: {'train/loss': 0.08647136390209198, 'train/grad_norm_pre_clip': 8.654216766357422, 'timing/data': 0.006304681301116943, 'timing/model': 2.6695045828819275, 'learning_rate/qwen_vl_interface': 8.776539598517119e-06, 'learning_rate/action_model': 8.776539598517118e-05, 'epoch': 2.41}) +[2026-06-19 22:27:54,559][starVLA.training.train_starvla][INFO] - Step 1246, Loss: {'train/loss': 0.17494888603687286, 'train/grad_norm_pre_clip': 15.451736450195312, 'timing/data': 0.006405100226402283, 'timing/model': 2.6720031164586544, 'learning_rate/qwen_vl_interface': 8.774498671932108e-06, 'learning_rate/action_model': 8.774498671932107e-05, 'epoch': 2.41}) +[2026-06-19 22:28:05,344][starVLA.training.train_starvla][INFO] - Step 1247, Loss: {'train/loss': 0.03891146928071976, 'train/grad_norm_pre_clip': 8.015178680419922, 'timing/data': 0.006346270442008972, 'timing/model': 2.702634323388338, 'learning_rate/qwen_vl_interface': 8.772456296558824e-06, 'learning_rate/action_model': 8.772456296558824e-05, 'epoch': 2.42}) +[2026-06-19 22:28:15,845][starVLA.training.train_starvla][INFO] - Step 1248, Loss: {'train/loss': 0.30987393856048584, 'train/grad_norm_pre_clip': 8.802821159362793, 'timing/data': 0.005974693223834038, 'timing/model': 2.65868304297328, 'learning_rate/qwen_vl_interface': 8.77041247323681e-06, 'learning_rate/action_model': 8.770412473236811e-05, 'epoch': 2.42}) +[2026-06-19 22:28:26,543][starVLA.training.train_starvla][INFO] - Step 1249, Loss: {'train/loss': 0.10901893675327301, 'train/grad_norm_pre_clip': 7.322055339813232, 'timing/data': 0.00651976652443409, 'timing/model': 2.675166131928563, 'learning_rate/qwen_vl_interface': 8.76836720280621e-06, 'learning_rate/action_model': 8.768367202806209e-05, 'epoch': 2.42}) +[2026-06-19 22:28:37,044][starVLA.training.train_starvla][INFO] - Step 1250, Loss: {'train/loss': 0.09081894904375076, 'train/grad_norm_pre_clip': 9.432104110717773, 'timing/data': 0.006435459479689598, 'timing/model': 2.680253576487303, 'learning_rate/qwen_vl_interface': 8.766320486107749e-06, 'learning_rate/action_model': 8.766320486107749e-05, 'epoch': 2.42}) +[2026-06-19 22:28:47,546][starVLA.training.train_starvla][INFO] - Step 1251, Loss: {'train/loss': 0.1982961893081665, 'train/grad_norm_pre_clip': 7.149750232696533, 'timing/data': 0.006409700959920883, 'timing/model': 2.675108067691326, 'learning_rate/qwen_vl_interface': 8.76427232398276e-06, 'learning_rate/action_model': 8.76427232398276e-05, 'epoch': 2.42}) +[2026-06-19 22:28:58,191][starVLA.training.train_starvla][INFO] - Step 1252, Loss: {'train/loss': 0.16635006666183472, 'train/grad_norm_pre_clip': 14.4402437210083, 'timing/data': 0.00659773126244545, 'timing/model': 2.705250795930624, 'learning_rate/qwen_vl_interface': 8.762222717273167e-06, 'learning_rate/action_model': 8.762222717273167e-05, 'epoch': 2.43}) +[2026-06-19 22:29:08,845][starVLA.training.train_starvla][INFO] - Step 1253, Loss: {'train/loss': 0.43229857087135315, 'train/grad_norm_pre_clip': 13.286412239074707, 'timing/data': 0.006514828652143478, 'timing/model': 2.69959457218647, 'learning_rate/qwen_vl_interface': 8.76017166682148e-06, 'learning_rate/action_model': 8.760171666821481e-05, 'epoch': 2.43}) +[2026-06-19 22:29:19,354][starVLA.training.train_starvla][INFO] - Step 1254, Loss: {'train/loss': 0.14416778087615967, 'train/grad_norm_pre_clip': 12.108421325683594, 'timing/data': 0.0061357151716947556, 'timing/model': 2.667231908068061, 'learning_rate/qwen_vl_interface': 8.758119173470815e-06, 'learning_rate/action_model': 8.758119173470815e-05, 'epoch': 2.43}) +[2026-06-19 22:29:29,955][starVLA.training.train_starvla][INFO] - Step 1255, Loss: {'train/loss': 0.39997178316116333, 'train/grad_norm_pre_clip': 8.678153038024902, 'timing/data': 0.006243454292416573, 'timing/model': 2.689644768834114, 'learning_rate/qwen_vl_interface': 8.75606523806487e-06, 'learning_rate/action_model': 8.756065238064868e-05, 'epoch': 2.43}) +[2026-06-19 22:29:40,617][starVLA.training.train_starvla][INFO] - Step 1256, Loss: {'train/loss': 0.14037157595157623, 'train/grad_norm_pre_clip': 6.920300006866455, 'timing/data': 0.006442179903388023, 'timing/model': 2.749778635799885, 'learning_rate/qwen_vl_interface': 8.75400986144794e-06, 'learning_rate/action_model': 8.75400986144794e-05, 'epoch': 2.43}) +[2026-06-19 22:29:51,167][starVLA.training.train_starvla][INFO] - Step 1257, Loss: {'train/loss': 0.3023490011692047, 'train/grad_norm_pre_clip': 10.012414932250977, 'timing/data': 0.006215944886207581, 'timing/model': 2.6862280890345573, 'learning_rate/qwen_vl_interface': 8.751953044464913e-06, 'learning_rate/action_model': 8.751953044464913e-05, 'epoch': 2.44}) +[2026-06-19 22:30:01,669][starVLA.training.train_starvla][INFO] - Step 1258, Loss: {'train/loss': 0.17101624608039856, 'train/grad_norm_pre_clip': 4.21258544921875, 'timing/data': 0.00630619190633297, 'timing/model': 2.7029916923493147, 'learning_rate/qwen_vl_interface': 8.749894787961272e-06, 'learning_rate/action_model': 8.749894787961271e-05, 'epoch': 2.44}) +[2026-06-19 22:30:12,405][starVLA.training.train_starvla][INFO] - Step 1259, Loss: {'train/loss': 0.24296891689300537, 'train/grad_norm_pre_clip': 7.912667274475098, 'timing/data': 0.006688546389341354, 'timing/model': 2.7716569546610117, 'learning_rate/qwen_vl_interface': 8.747835092783086e-06, 'learning_rate/action_model': 8.747835092783085e-05, 'epoch': 2.44}) +[2026-06-19 22:30:23,424][starVLA.training.train_starvla][INFO] - Step 1260, Loss: {'train/loss': 0.208940789103508, 'train/grad_norm_pre_clip': 6.411944389343262, 'timing/data': 0.006379192695021629, 'timing/model': 2.7762621734291315, 'learning_rate/qwen_vl_interface': 8.745773959777018e-06, 'learning_rate/action_model': 8.745773959777017e-05, 'epoch': 2.44}) +[2026-06-19 22:30:34,320][starVLA.training.train_starvla][INFO] - Step 1261, Loss: {'train/loss': 0.14088167250156403, 'train/grad_norm_pre_clip': 3.4197609424591064, 'timing/data': 0.006449732929468155, 'timing/model': 2.77008206397295, 'learning_rate/qwen_vl_interface': 8.743711389790321e-06, 'learning_rate/action_model': 8.74371138979032e-05, 'epoch': 2.44}) +[2026-06-19 22:30:44,976][starVLA.training.train_starvla][INFO] - Step 1262, Loss: {'train/loss': 0.05905279889702797, 'train/grad_norm_pre_clip': 1.904463768005371, 'timing/data': 0.006407111883163452, 'timing/model': 2.7275796104222536, 'learning_rate/qwen_vl_interface': 8.741647383670839e-06, 'learning_rate/action_model': 8.741647383670839e-05, 'epoch': 2.45}) +[2026-06-19 22:30:55,728][starVLA.training.train_starvla][INFO] - Step 1263, Loss: {'train/loss': 0.18993419408798218, 'train/grad_norm_pre_clip': 6.999081611633301, 'timing/data': 0.0061370134353637695, 'timing/model': 2.74481825158, 'learning_rate/qwen_vl_interface': 8.739581942267011e-06, 'learning_rate/action_model': 8.739581942267011e-05, 'epoch': 2.45}) +[2026-06-19 22:31:06,435][starVLA.training.train_starvla][INFO] - Step 1264, Loss: {'train/loss': 0.1819130927324295, 'train/grad_norm_pre_clip': 6.294799327850342, 'timing/data': 0.0064380597323179245, 'timing/model': 2.713233971968293, 'learning_rate/qwen_vl_interface': 8.737515066427861e-06, 'learning_rate/action_model': 8.73751506642786e-05, 'epoch': 2.45}) +[2026-06-19 22:31:17,121][starVLA.training.train_starvla][INFO] - Step 1265, Loss: {'train/loss': 0.19041292369365692, 'train/grad_norm_pre_clip': 7.633578777313232, 'timing/data': 0.0063058920204639435, 'timing/model': 2.7511536069214344, 'learning_rate/qwen_vl_interface': 8.735446757003e-06, 'learning_rate/action_model': 8.735446757003e-05, 'epoch': 2.45}) +[2026-06-19 22:31:27,904][starVLA.training.train_starvla][INFO] - Step 1266, Loss: {'train/loss': 0.1919524073600769, 'train/grad_norm_pre_clip': 6.783097267150879, 'timing/data': 0.00638497993350029, 'timing/model': 2.763526290655136, 'learning_rate/qwen_vl_interface': 8.733377014842635e-06, 'learning_rate/action_model': 8.733377014842635e-05, 'epoch': 2.45}) +[2026-06-19 22:31:38,759][starVLA.training.train_starvla][INFO] - Step 1267, Loss: {'train/loss': 0.1660171002149582, 'train/grad_norm_pre_clip': 7.0600385665893555, 'timing/data': 0.006342459470033646, 'timing/model': 2.9447672069072723, 'learning_rate/qwen_vl_interface': 8.731305840797557e-06, 'learning_rate/action_model': 8.731305840797558e-05, 'epoch': 2.46}) +[2026-06-19 22:31:49,527][starVLA.training.train_starvla][INFO] - Step 1268, Loss: {'train/loss': 0.2892625331878662, 'train/grad_norm_pre_clip': 8.83155632019043, 'timing/data': 0.006332309916615486, 'timing/model': 2.7632314320653677, 'learning_rate/qwen_vl_interface': 8.72923323571915e-06, 'learning_rate/action_model': 8.72923323571915e-05, 'epoch': 2.46}) +[2026-06-19 22:32:00,312][starVLA.training.train_starvla][INFO] - Step 1269, Loss: {'train/loss': 0.17798104882240295, 'train/grad_norm_pre_clip': 6.742486953735352, 'timing/data': 0.006364760920405388, 'timing/model': 2.7420197799801826, 'learning_rate/qwen_vl_interface': 8.727159200459381e-06, 'learning_rate/action_model': 8.727159200459381e-05, 'epoch': 2.46}) +[2026-06-19 22:32:10,978][starVLA.training.train_starvla][INFO] - Step 1270, Loss: {'train/loss': 0.12086788564920425, 'train/grad_norm_pre_clip': 4.77813720703125, 'timing/data': 0.006309822201728821, 'timing/model': 2.6886600125581026, 'learning_rate/qwen_vl_interface': 8.725083735870807e-06, 'learning_rate/action_model': 8.725083735870807e-05, 'epoch': 2.46}) +[2026-06-19 22:32:21,721][starVLA.training.train_starvla][INFO] - Step 1271, Loss: {'train/loss': 0.07596120238304138, 'train/grad_norm_pre_clip': 6.975674152374268, 'timing/data': 0.006457390263676643, 'timing/model': 2.748730655759573, 'learning_rate/qwen_vl_interface': 8.723006842806575e-06, 'learning_rate/action_model': 8.723006842806576e-05, 'epoch': 2.46}) +[2026-06-19 22:32:32,328][starVLA.training.train_starvla][INFO] - Step 1272, Loss: {'train/loss': 0.25427693128585815, 'train/grad_norm_pre_clip': 13.66321086883545, 'timing/data': 0.006515879184007645, 'timing/model': 2.7130702156573534, 'learning_rate/qwen_vl_interface': 8.720928522120419e-06, 'learning_rate/action_model': 8.720928522120418e-05, 'epoch': 2.47}) +[2026-06-19 22:32:43,216][starVLA.training.train_starvla][INFO] - Step 1273, Loss: {'train/loss': 0.13043104112148285, 'train/grad_norm_pre_clip': 9.829690933227539, 'timing/data': 0.0064207520335912704, 'timing/model': 2.7764098662883043, 'learning_rate/qwen_vl_interface': 8.718848774666655e-06, 'learning_rate/action_model': 8.718848774666655e-05, 'epoch': 2.47}) +[2026-06-19 22:32:53,803][starVLA.training.train_starvla][INFO] - Step 1274, Loss: {'train/loss': 0.1099136620759964, 'train/grad_norm_pre_clip': 3.5904130935668945, 'timing/data': 0.007157683372497559, 'timing/model': 2.7339066676795483, 'learning_rate/qwen_vl_interface': 8.716767601300189e-06, 'learning_rate/action_model': 8.716767601300188e-05, 'epoch': 2.47}) +[2026-06-19 22:33:04,529][starVLA.training.train_starvla][INFO] - Step 1275, Loss: {'train/loss': 0.3053632974624634, 'train/grad_norm_pre_clip': 16.238431930541992, 'timing/data': 0.013215813785791397, 'timing/model': 2.75438248552382, 'learning_rate/qwen_vl_interface': 8.714685002876514e-06, 'learning_rate/action_model': 8.714685002876513e-05, 'epoch': 2.47}) +[2026-06-19 22:33:15,143][starVLA.training.train_starvla][INFO] - Step 1276, Loss: {'train/loss': 0.12036563456058502, 'train/grad_norm_pre_clip': 5.267298221588135, 'timing/data': 0.006529510021209717, 'timing/model': 2.679829539731145, 'learning_rate/qwen_vl_interface': 8.712600980251706e-06, 'learning_rate/action_model': 8.712600980251705e-05, 'epoch': 2.47}) +[2026-06-19 22:33:25,742][starVLA.training.train_starvla][INFO] - Step 1277, Loss: {'train/loss': 0.211900532245636, 'train/grad_norm_pre_clip': 11.465205192565918, 'timing/data': 0.006588198244571686, 'timing/model': 2.677385525777936, 'learning_rate/qwen_vl_interface': 8.710515534282428e-06, 'learning_rate/action_model': 8.710515534282428e-05, 'epoch': 2.47}) +[2026-06-19 22:33:36,246][starVLA.training.train_starvla][INFO] - Step 1278, Loss: {'train/loss': 0.15257732570171356, 'train/grad_norm_pre_clip': 14.591828346252441, 'timing/data': 0.006170852109789848, 'timing/model': 2.677434118464589, 'learning_rate/qwen_vl_interface': 8.708428665825931e-06, 'learning_rate/action_model': 8.708428665825931e-05, 'epoch': 2.48}) +[2026-06-19 22:33:46,952][starVLA.training.train_starvla][INFO] - Step 1279, Loss: {'train/loss': 0.018437741324305534, 'train/grad_norm_pre_clip': 1.349651575088501, 'timing/data': 0.006458688527345657, 'timing/model': 2.68394198641181, 'learning_rate/qwen_vl_interface': 8.706340375740045e-06, 'learning_rate/action_model': 8.706340375740045e-05, 'epoch': 2.48}) +[2026-06-19 22:33:57,683][starVLA.training.train_starvla][INFO] - Step 1280, Loss: {'train/loss': 0.39597415924072266, 'train/grad_norm_pre_clip': 16.754867553710938, 'timing/data': 0.006502177566289902, 'timing/model': 2.7337237037718296, 'learning_rate/qwen_vl_interface': 8.704250664883188e-06, 'learning_rate/action_model': 8.704250664883188e-05, 'epoch': 2.48}) +[2026-06-19 22:34:08,315][starVLA.training.train_starvla][INFO] - Step 1281, Loss: {'train/loss': 0.1930878460407257, 'train/grad_norm_pre_clip': 12.29040241241455, 'timing/data': 0.006534690037369728, 'timing/model': 2.746875951066613, 'learning_rate/qwen_vl_interface': 8.70215953411436e-06, 'learning_rate/action_model': 8.70215953411436e-05, 'epoch': 2.48}) +[2026-06-19 22:34:18,947][starVLA.training.train_starvla][INFO] - Step 1282, Loss: {'train/loss': 0.2854868173599243, 'train/grad_norm_pre_clip': 15.831772804260254, 'timing/data': 0.006513338536024094, 'timing/model': 2.639052202925086, 'learning_rate/qwen_vl_interface': 8.70006698429315e-06, 'learning_rate/action_model': 8.700066984293149e-05, 'epoch': 2.48}) +[2026-06-19 22:34:29,575][starVLA.training.train_starvla][INFO] - Step 1283, Loss: {'train/loss': 0.16994071006774902, 'train/grad_norm_pre_clip': 9.677271842956543, 'timing/data': 0.00623181089758873, 'timing/model': 2.730059629306197, 'learning_rate/qwen_vl_interface': 8.697973016279722e-06, 'learning_rate/action_model': 8.697973016279722e-05, 'epoch': 2.49}) +[2026-06-19 22:34:40,145][starVLA.training.train_starvla][INFO] - Step 1284, Loss: {'train/loss': 0.15835760533809662, 'train/grad_norm_pre_clip': 7.3650736808776855, 'timing/data': 0.006466979160904884, 'timing/model': 2.703278163447976, 'learning_rate/qwen_vl_interface': 8.695877630934827e-06, 'learning_rate/action_model': 8.695877630934826e-05, 'epoch': 2.49}) +[2026-06-19 22:34:50,647][starVLA.training.train_starvla][INFO] - Step 1285, Loss: {'train/loss': 0.18151520192623138, 'train/grad_norm_pre_clip': 10.937174797058105, 'timing/data': 0.006391981616616249, 'timing/model': 2.6774550769478083, 'learning_rate/qwen_vl_interface': 8.693780829119801e-06, 'learning_rate/action_model': 8.693780829119801e-05, 'epoch': 2.49}) +[2026-06-19 22:35:01,151][starVLA.training.train_starvla][INFO] - Step 1286, Loss: {'train/loss': 0.23820282518863678, 'train/grad_norm_pre_clip': 10.43736743927002, 'timing/data': 0.006246780976653099, 'timing/model': 2.6819611825048923, 'learning_rate/qwen_vl_interface': 8.69168261169656e-06, 'learning_rate/action_model': 8.691682611696559e-05, 'epoch': 2.49}) +[2026-06-19 22:35:12,012][starVLA.training.train_starvla][INFO] - Step 1287, Loss: {'train/loss': 0.21415607631206512, 'train/grad_norm_pre_clip': 8.780905723571777, 'timing/data': 0.006417060270905495, 'timing/model': 2.771892659366131, 'learning_rate/qwen_vl_interface': 8.689582979527599e-06, 'learning_rate/action_model': 8.689582979527599e-05, 'epoch': 2.49}) +[2026-06-19 22:35:22,786][starVLA.training.train_starvla][INFO] - Step 1288, Loss: {'train/loss': 0.07652367651462555, 'train/grad_norm_pre_clip': 6.771851062774658, 'timing/data': 0.006503278389573097, 'timing/model': 2.735985115170479, 'learning_rate/qwen_vl_interface': 8.687481933476e-06, 'learning_rate/action_model': 8.687481933476e-05, 'epoch': 2.5}) +[2026-06-19 22:35:33,425][starVLA.training.train_starvla][INFO] - Step 1289, Loss: {'train/loss': 0.04815477505326271, 'train/grad_norm_pre_clip': 4.289219379425049, 'timing/data': 0.006335001438856125, 'timing/model': 2.7697471976280212, 'learning_rate/qwen_vl_interface': 8.685379474405423e-06, 'learning_rate/action_model': 8.685379474405423e-05, 'epoch': 2.5}) +[2026-06-19 22:35:44,189][starVLA.training.train_starvla][INFO] - Step 1290, Loss: {'train/loss': 0.17569246888160706, 'train/grad_norm_pre_clip': 8.355905532836914, 'timing/data': 0.006658194586634636, 'timing/model': 2.8208022825419903, 'learning_rate/qwen_vl_interface': 8.68327560318011e-06, 'learning_rate/action_model': 8.68327560318011e-05, 'epoch': 2.5}) +[2026-06-19 22:35:54,642][starVLA.training.train_starvla][INFO] - Step 1291, Loss: {'train/loss': 0.04587435722351074, 'train/grad_norm_pre_clip': 2.0600457191467285, 'timing/data': 0.006178632378578186, 'timing/model': 2.6462030950933695, 'learning_rate/qwen_vl_interface': 8.681170320664883e-06, 'learning_rate/action_model': 8.681170320664883e-05, 'epoch': 2.5}) +[2026-06-19 22:36:05,171][starVLA.training.train_starvla][INFO] - Step 1292, Loss: {'train/loss': 0.11356107145547867, 'train/grad_norm_pre_clip': 7.844808101654053, 'timing/data': 0.006572335958480835, 'timing/model': 2.70066193677485, 'learning_rate/qwen_vl_interface': 8.679063627725143e-06, 'learning_rate/action_model': 8.679063627725144e-05, 'epoch': 2.5}) +[2026-06-19 22:36:15,827][starVLA.training.train_starvla][INFO] - Step 1293, Loss: {'train/loss': 0.29335451126098633, 'train/grad_norm_pre_clip': 7.49007511138916, 'timing/data': 0.006338771432638168, 'timing/model': 2.785566095262766, 'learning_rate/qwen_vl_interface': 8.676955525226876e-06, 'learning_rate/action_model': 8.676955525226875e-05, 'epoch': 2.51}) +[2026-06-19 22:36:26,529][starVLA.training.train_starvla][INFO] - Step 1294, Loss: {'train/loss': 0.2236914336681366, 'train/grad_norm_pre_clip': 6.875609874725342, 'timing/data': 0.006535060703754425, 'timing/model': 2.7165672071278095, 'learning_rate/qwen_vl_interface': 8.674846014036636e-06, 'learning_rate/action_model': 8.674846014036636e-05, 'epoch': 2.51}) +[2026-06-19 22:36:37,224][starVLA.training.train_starvla][INFO] - Step 1295, Loss: {'train/loss': 0.42506346106529236, 'train/grad_norm_pre_clip': 13.103010177612305, 'timing/data': 0.006473759189248085, 'timing/model': 2.7549974471330643, 'learning_rate/qwen_vl_interface': 8.672735095021569e-06, 'learning_rate/action_model': 8.672735095021569e-05, 'epoch': 2.51}) +[2026-06-19 22:36:47,920][starVLA.training.train_starvla][INFO] - Step 1296, Loss: {'train/loss': 0.1516299992799759, 'train/grad_norm_pre_clip': 9.494017601013184, 'timing/data': 0.006459688767790794, 'timing/model': 2.7482659965753555, 'learning_rate/qwen_vl_interface': 8.670622769049393e-06, 'learning_rate/action_model': 8.670622769049392e-05, 'epoch': 2.51}) +[2026-06-19 22:36:58,592][starVLA.training.train_starvla][INFO] - Step 1297, Loss: {'train/loss': 0.20076340436935425, 'train/grad_norm_pre_clip': 7.134530067443848, 'timing/data': 0.006513968110084534, 'timing/model': 2.7198889926075935, 'learning_rate/qwen_vl_interface': 8.668509036988406e-06, 'learning_rate/action_model': 8.668509036988405e-05, 'epoch': 2.51}) +[2026-06-19 22:37:09,165][starVLA.training.train_starvla][INFO] - Step 1298, Loss: {'train/loss': 0.19496934115886688, 'train/grad_norm_pre_clip': 11.09653377532959, 'timing/data': 0.00645730085670948, 'timing/model': 2.7350337486714125, 'learning_rate/qwen_vl_interface': 8.66639389970748e-06, 'learning_rate/action_model': 8.666393899707479e-05, 'epoch': 2.52}) +[2026-06-19 22:37:19,828][starVLA.training.train_starvla][INFO] - Step 1299, Loss: {'train/loss': 0.11059501767158508, 'train/grad_norm_pre_clip': 6.576317310333252, 'timing/data': 0.006509875878691673, 'timing/model': 2.7607362028211355, 'learning_rate/qwen_vl_interface': 8.66427735807607e-06, 'learning_rate/action_model': 8.664277358076069e-05, 'epoch': 2.52}) +[2026-06-19 22:37:30,626][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1300 +[2026-06-19 22:45:03,111][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1300 +[2026-06-19 22:45:03,112][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 1300 +[2026-06-19 22:53:08,124][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 1300 +[2026-06-19 22:53:08,125][starVLA.training.train_starvla][INFO] - Step 1300, Loss: {'train/loss': 0.16634057462215424, 'train/grad_norm_pre_clip': 9.174789428710938, 'eval/loss': 0.5068615078926086, 'eval/latency_0/loss': 0.27233749628067017, 'eval/latency_1/loss': 0.06847261637449265, 'eval/latency_2/loss': 0.5646015405654907, 'eval/latency_3/loss': 0.6636732220649719, 'eval/latency_4/loss': 0.9654171466827393, 'eval/flappy/loss': 0.5068615078926086, 'eval/flappy/latency_0/loss': 0.27233749628067017, 'eval/flappy/latency_1/loss': 0.06847261637449265, 'eval/flappy/latency_2/loss': 0.5646015405654907, 'eval/flappy/latency_3/loss': 0.6636732220649719, 'eval/flappy/latency_4/loss': 0.9654171466827393, 'eval/flap_precision': 0.49169435215946844, 'eval/flap_recall': 0.5381818181818182, 'eval/flap_f1': 0.513888888888889, 'eval/flap_rate_model': 0.08361111111111111, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.513888888888889, 'eval/latency_0/flap_precision': 0.49169435215946844, 'eval/latency_0/flap_recall': 0.5381818181818182, 'eval/latency_0/flap_f1': 0.513888888888889, 'eval/latency_0/flap_rate_model': 0.08361111111111111, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.513888888888889, 'timing/data': 0.006487170234322548, 'timing/model': 2.759625729173422, 'learning_rate/qwen_vl_interface': 8.662159412964208e-06, 'learning_rate/action_model': 8.662159412964207e-05, 'epoch': 2.52}) +[2026-06-19 22:53:08,126][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_1300_state +[2026-06-19 22:53:08,152][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-19 22:53:13,945][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_1300_state/model.safetensors +[2026-06-19 22:53:21,962][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_1300_state/optimizer.bin +[2026-06-19 22:53:21,962][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_1300_state/sampler.bin +[2026-06-19 22:53:21,963][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_1300_state/sampler_1.bin +[2026-06-19 22:53:21,964][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_1300_state/random_states_0.pkl +[2026-06-19 22:53:22,108][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1300 +[2026-06-19 22:53:22,108][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-19 22:53:22,122][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-19 22:53:33,008][starVLA.training.train_starvla][INFO] - Step 1301, Loss: {'train/loss': 0.14634107053279877, 'train/grad_norm_pre_clip': 2.827075242996216, 'timing/data': 0.006645489484071732, 'timing/model': 2.7753951847553253, 'learning_rate/qwen_vl_interface': 8.660040065242499e-06, 'learning_rate/action_model': 8.660040065242498e-05, 'epoch': 2.52}) +[2026-06-19 22:53:43,692][starVLA.training.train_starvla][INFO] - Step 1302, Loss: {'train/loss': 0.08310742676258087, 'train/grad_norm_pre_clip': 3.1902096271514893, 'timing/data': 0.0065501779317855835, 'timing/model': 2.7310020960867405, 'learning_rate/qwen_vl_interface': 8.657919315782127e-06, 'learning_rate/action_model': 8.657919315782127e-05, 'epoch': 2.52}) +[2026-06-19 22:53:54,512][starVLA.training.train_starvla][INFO] - Step 1303, Loss: {'train/loss': 0.06831508874893188, 'train/grad_norm_pre_clip': 4.578146934509277, 'timing/data': 0.006870191544294357, 'timing/model': 2.777754409238696, 'learning_rate/qwen_vl_interface': 8.655797165454853e-06, 'learning_rate/action_model': 8.655797165454853e-05, 'epoch': 2.53}) +[2026-06-19 22:54:05,326][starVLA.training.train_starvla][INFO] - Step 1304, Loss: {'train/loss': 0.09945143759250641, 'train/grad_norm_pre_clip': 3.392630100250244, 'timing/data': 0.009130099788308144, 'timing/model': 2.778495578095317, 'learning_rate/qwen_vl_interface': 8.653673615133013e-06, 'learning_rate/action_model': 8.653673615133013e-05, 'epoch': 2.53}) +[2026-06-19 22:54:16,191][starVLA.training.train_starvla][INFO] - Step 1305, Loss: {'train/loss': 0.12832671403884888, 'train/grad_norm_pre_clip': 8.49085807800293, 'timing/data': 0.007535016164183617, 'timing/model': 2.7464555706828833, 'learning_rate/qwen_vl_interface': 8.651548665689519e-06, 'learning_rate/action_model': 8.651548665689518e-05, 'epoch': 2.53}) +[2026-06-19 22:54:26,961][starVLA.training.train_starvla][INFO] - Step 1306, Loss: {'train/loss': 0.058250464498996735, 'train/grad_norm_pre_clip': 6.205978870391846, 'timing/data': 0.008062079548835754, 'timing/model': 2.704955590888858, 'learning_rate/qwen_vl_interface': 8.649422317997855e-06, 'learning_rate/action_model': 8.649422317997856e-05, 'epoch': 2.53}) +[2026-06-19 22:54:37,718][starVLA.training.train_starvla][INFO] - Step 1307, Loss: {'train/loss': 0.04910305142402649, 'train/grad_norm_pre_clip': 2.0560569763183594, 'timing/data': 0.006433280184864998, 'timing/model': 2.7152189053595066, 'learning_rate/qwen_vl_interface': 8.647294572932082e-06, 'learning_rate/action_model': 8.647294572932082e-05, 'epoch': 2.53}) +[2026-06-19 22:54:48,514][starVLA.training.train_starvla][INFO] - Step 1308, Loss: {'train/loss': 0.2047642469406128, 'train/grad_norm_pre_clip': 6.410258769989014, 'timing/data': 0.0065391287207603455, 'timing/model': 2.7669937033206224, 'learning_rate/qwen_vl_interface': 8.64516543136684e-06, 'learning_rate/action_model': 8.64516543136684e-05, 'epoch': 2.54}) +[2026-06-19 22:54:59,317][starVLA.training.train_starvla][INFO] - Step 1309, Loss: {'train/loss': 0.3489081859588623, 'train/grad_norm_pre_clip': 11.872015953063965, 'timing/data': 0.00831579603254795, 'timing/model': 2.7703970447182655, 'learning_rate/qwen_vl_interface': 8.643034894177335e-06, 'learning_rate/action_model': 8.643034894177335e-05, 'epoch': 2.54}) +[2026-06-19 22:55:10,459][starVLA.training.train_starvla][INFO] - Step 1310, Loss: {'train/loss': 0.12924841046333313, 'train/grad_norm_pre_clip': 6.603099346160889, 'timing/data': 0.0077821556478738785, 'timing/model': 3.0092960577458143, 'learning_rate/qwen_vl_interface': 8.64090296223935e-06, 'learning_rate/action_model': 8.64090296223935e-05, 'epoch': 2.54}) +[2026-06-19 22:55:21,269][starVLA.training.train_starvla][INFO] - Step 1311, Loss: {'train/loss': 0.15900491178035736, 'train/grad_norm_pre_clip': 10.171762466430664, 'timing/data': 0.006311411038041115, 'timing/model': 2.7197124641388655, 'learning_rate/qwen_vl_interface': 8.638769636429245e-06, 'learning_rate/action_model': 8.638769636429245e-05, 'epoch': 2.54}) +[2026-06-19 22:55:32,118][starVLA.training.train_starvla][INFO] - Step 1312, Loss: {'train/loss': 0.052709806710481644, 'train/grad_norm_pre_clip': 2.3973448276519775, 'timing/data': 0.006353020668029785, 'timing/model': 2.7774661108851433, 'learning_rate/qwen_vl_interface': 8.636634917623946e-06, 'learning_rate/action_model': 8.636634917623944e-05, 'epoch': 2.54}) +[2026-06-19 22:55:42,916][starVLA.training.train_starvla][INFO] - Step 1313, Loss: {'train/loss': 0.2519795298576355, 'train/grad_norm_pre_clip': 6.414036750793457, 'timing/data': 0.0065965354442596436, 'timing/model': 2.7616355288773775, 'learning_rate/qwen_vl_interface': 8.634498806700955e-06, 'learning_rate/action_model': 8.634498806700954e-05, 'epoch': 2.54}) +[2026-06-19 22:55:53,927][starVLA.training.train_starvla][INFO] - Step 1314, Loss: {'train/loss': 0.1514001339673996, 'train/grad_norm_pre_clip': 6.6662211418151855, 'timing/data': 0.006424501538276672, 'timing/model': 2.7883837912231684, 'learning_rate/qwen_vl_interface': 8.63236130453835e-06, 'learning_rate/action_model': 8.63236130453835e-05, 'epoch': 2.55}) +[2026-06-19 22:56:04,822][starVLA.training.train_starvla][INFO] - Step 1315, Loss: {'train/loss': 0.3071836829185486, 'train/grad_norm_pre_clip': 11.46737289428711, 'timing/data': 0.0067298151552677155, 'timing/model': 2.7694631293416023, 'learning_rate/qwen_vl_interface': 8.630222412014777e-06, 'learning_rate/action_model': 8.630222412014776e-05, 'epoch': 2.55}) +[2026-06-19 22:56:15,731][starVLA.training.train_starvla][INFO] - Step 1316, Loss: {'train/loss': 0.29013821482658386, 'train/grad_norm_pre_clip': 9.966020584106445, 'timing/data': 0.006552359089255333, 'timing/model': 2.7790112048387527, 'learning_rate/qwen_vl_interface': 8.628082130009451e-06, 'learning_rate/action_model': 8.628082130009451e-05, 'epoch': 2.55}) +[2026-06-19 22:56:26,600][starVLA.training.train_starvla][INFO] - Step 1317, Loss: {'train/loss': 0.043714798986911774, 'train/grad_norm_pre_clip': 2.434683322906494, 'timing/data': 0.006426246836781502, 'timing/model': 2.7565468847751617, 'learning_rate/qwen_vl_interface': 8.62594045940216e-06, 'learning_rate/action_model': 8.62594045940216e-05, 'epoch': 2.55}) +[2026-06-19 22:56:37,627][starVLA.training.train_starvla][INFO] - Step 1318, Loss: {'train/loss': 0.05449893698096275, 'train/grad_norm_pre_clip': 3.647324800491333, 'timing/data': 0.006732193753123283, 'timing/model': 2.762502895668149, 'learning_rate/qwen_vl_interface': 8.623797401073272e-06, 'learning_rate/action_model': 8.623797401073271e-05, 'epoch': 2.55}) +[2026-06-19 22:56:48,529][starVLA.training.train_starvla][INFO] - Step 1319, Loss: {'train/loss': 0.17856737971305847, 'train/grad_norm_pre_clip': 7.269644260406494, 'timing/data': 0.006614286452531815, 'timing/model': 2.7866586055606604, 'learning_rate/qwen_vl_interface': 8.62165295590371e-06, 'learning_rate/action_model': 8.621652955903709e-05, 'epoch': 2.56}) +[2026-06-19 22:56:59,420][starVLA.training.train_starvla][INFO] - Step 1320, Loss: {'train/loss': 0.19175784289836884, 'train/grad_norm_pre_clip': 10.251526832580566, 'timing/data': 0.00664885900914669, 'timing/model': 2.764042077586055, 'learning_rate/qwen_vl_interface': 8.619507124774975e-06, 'learning_rate/action_model': 8.619507124774975e-05, 'epoch': 2.56}) +[2026-06-19 22:57:10,460][starVLA.training.train_starvla][INFO] - Step 1321, Loss: {'train/loss': 0.23831884562969208, 'train/grad_norm_pre_clip': 14.091158866882324, 'timing/data': 0.00682266429066658, 'timing/model': 3.015653820708394, 'learning_rate/qwen_vl_interface': 8.61735990856914e-06, 'learning_rate/action_model': 8.617359908569139e-05, 'epoch': 2.56}) +[2026-06-19 22:57:21,319][starVLA.training.train_starvla][INFO] - Step 1322, Loss: {'train/loss': 0.31922420859336853, 'train/grad_norm_pre_clip': 11.701216697692871, 'timing/data': 0.006753865629434586, 'timing/model': 2.766921304166317, 'learning_rate/qwen_vl_interface': 8.615211308168842e-06, 'learning_rate/action_model': 8.61521130816884e-05, 'epoch': 2.56}) +[2026-06-19 22:57:32,023][starVLA.training.train_starvla][INFO] - Step 1323, Loss: {'train/loss': 0.24458445608615875, 'train/grad_norm_pre_clip': 6.346066474914551, 'timing/data': 0.00666830874979496, 'timing/model': 2.758294928818941, 'learning_rate/qwen_vl_interface': 8.613061324457289e-06, 'learning_rate/action_model': 8.613061324457288e-05, 'epoch': 2.56}) +[2026-06-19 22:57:42,935][starVLA.training.train_starvla][INFO] - Step 1324, Loss: {'train/loss': 0.09937464445829391, 'train/grad_norm_pre_clip': 7.4268035888671875, 'timing/data': 0.0065541863441467285, 'timing/model': 2.7804014328867197, 'learning_rate/qwen_vl_interface': 8.610909958318262e-06, 'learning_rate/action_model': 8.610909958318261e-05, 'epoch': 2.57}) +[2026-06-19 22:57:53,833][starVLA.training.train_starvla][INFO] - Step 1325, Loss: {'train/loss': 0.09624581784009933, 'train/grad_norm_pre_clip': 4.809853553771973, 'timing/data': 0.006417721509933472, 'timing/model': 2.783301930874586, 'learning_rate/qwen_vl_interface': 8.608757210636101e-06, 'learning_rate/action_model': 8.608757210636101e-05, 'epoch': 2.57}) +[2026-06-19 22:58:04,721][starVLA.training.train_starvla][INFO] - Step 1326, Loss: {'train/loss': 0.04423320293426514, 'train/grad_norm_pre_clip': 4.533619403839111, 'timing/data': 0.00644647516310215, 'timing/model': 2.7635678686201572, 'learning_rate/qwen_vl_interface': 8.606603082295724e-06, 'learning_rate/action_model': 8.606603082295724e-05, 'epoch': 2.57}) +[2026-06-19 22:58:15,523][starVLA.training.train_starvla][INFO] - Step 1327, Loss: {'train/loss': 0.38377854228019714, 'train/grad_norm_pre_clip': 11.28753662109375, 'timing/data': 0.006711507216095924, 'timing/model': 2.7826670445501804, 'learning_rate/qwen_vl_interface': 8.60444757418261e-06, 'learning_rate/action_model': 8.604447574182609e-05, 'epoch': 2.57}) +[2026-06-19 22:58:26,436][starVLA.training.train_starvla][INFO] - Step 1328, Loss: {'train/loss': 0.09177741408348083, 'train/grad_norm_pre_clip': 4.055384635925293, 'timing/data': 0.006703045219182968, 'timing/model': 2.78291648440063, 'learning_rate/qwen_vl_interface': 8.602290687182805e-06, 'learning_rate/action_model': 8.602290687182806e-05, 'epoch': 2.57}) +[2026-06-19 22:58:37,338][starVLA.training.train_starvla][INFO] - Step 1329, Loss: {'train/loss': 0.2647174596786499, 'train/grad_norm_pre_clip': 10.178759574890137, 'timing/data': 0.006401028484106064, 'timing/model': 2.7836366929113865, 'learning_rate/qwen_vl_interface': 8.600132422182929e-06, 'learning_rate/action_model': 8.600132422182928e-05, 'epoch': 2.58}) +[2026-06-19 22:58:48,218][starVLA.training.train_starvla][INFO] - Step 1330, Loss: {'train/loss': 0.19193682074546814, 'train/grad_norm_pre_clip': 5.908178329467773, 'timing/data': 0.006531799212098122, 'timing/model': 2.7667695824056864, 'learning_rate/qwen_vl_interface': 8.597972780070156e-06, 'learning_rate/action_model': 8.597972780070155e-05, 'epoch': 2.58}) +[2026-06-19 22:58:59,029][starVLA.training.train_starvla][INFO] - Step 1331, Loss: {'train/loss': 0.1698819249868393, 'train/grad_norm_pre_clip': 8.786758422851562, 'timing/data': 0.006578516215085983, 'timing/model': 2.7744532953947783, 'learning_rate/qwen_vl_interface': 8.595811761732239e-06, 'learning_rate/action_model': 8.595811761732238e-05, 'epoch': 2.58}) +[2026-06-19 22:59:09,933][starVLA.training.train_starvla][INFO] - Step 1332, Loss: {'train/loss': 0.30867937207221985, 'train/grad_norm_pre_clip': 8.17162799835205, 'timing/data': 0.006802825257182121, 'timing/model': 2.7830499596893787, 'learning_rate/qwen_vl_interface': 8.593649368057489e-06, 'learning_rate/action_model': 8.59364936805749e-05, 'epoch': 2.58}) +[2026-06-19 22:59:20,932][starVLA.training.train_starvla][INFO] - Step 1333, Loss: {'train/loss': 0.1332542896270752, 'train/grad_norm_pre_clip': 10.68317985534668, 'timing/data': 0.0064429547637701035, 'timing/model': 2.778276974335313, 'learning_rate/qwen_vl_interface': 8.591485599934786e-06, 'learning_rate/action_model': 8.591485599934785e-05, 'epoch': 2.58}) +[2026-06-19 22:59:31,831][starVLA.training.train_starvla][INFO] - Step 1334, Loss: {'train/loss': 0.17808648943901062, 'train/grad_norm_pre_clip': 4.565822601318359, 'timing/data': 0.006551288068294525, 'timing/model': 2.775162035599351, 'learning_rate/qwen_vl_interface': 8.58932045825357e-06, 'learning_rate/action_model': 8.589320458253569e-05, 'epoch': 2.59}) +[2026-06-19 22:59:42,736][starVLA.training.train_starvla][INFO] - Step 1335, Loss: {'train/loss': 0.18086643517017365, 'train/grad_norm_pre_clip': 8.147249221801758, 'timing/data': 0.006259661167860031, 'timing/model': 2.791401155292988, 'learning_rate/qwen_vl_interface': 8.58715394390385e-06, 'learning_rate/action_model': 8.58715394390385e-05, 'epoch': 2.59}) +[2026-06-19 22:59:53,632][starVLA.training.train_starvla][INFO] - Step 1336, Loss: {'train/loss': 0.2589464783668518, 'train/grad_norm_pre_clip': 6.536508083343506, 'timing/data': 0.006406849250197411, 'timing/model': 2.7826959919184446, 'learning_rate/qwen_vl_interface': 8.5849860577762e-06, 'learning_rate/action_model': 8.5849860577762e-05, 'epoch': 2.59}) +[2026-06-19 23:00:04,532][starVLA.training.train_starvla][INFO] - Step 1337, Loss: {'train/loss': 0.3489092290401459, 'train/grad_norm_pre_clip': 10.530860900878906, 'timing/data': 0.006911272183060646, 'timing/model': 2.7852940522134304, 'learning_rate/qwen_vl_interface': 8.582816800761756e-06, 'learning_rate/action_model': 8.582816800761755e-05, 'epoch': 2.59}) +[2026-06-19 23:00:15,434][starVLA.training.train_starvla][INFO] - Step 1338, Loss: {'train/loss': 0.11726274341344833, 'train/grad_norm_pre_clip': 5.786782264709473, 'timing/data': 0.007118241861462593, 'timing/model': 2.775971567258239, 'learning_rate/qwen_vl_interface': 8.580646173752214e-06, 'learning_rate/action_model': 8.580646173752213e-05, 'epoch': 2.59}) +[2026-06-19 23:00:26,333][starVLA.training.train_starvla][INFO] - Step 1339, Loss: {'train/loss': 0.18712279200553894, 'train/grad_norm_pre_clip': 5.281761169433594, 'timing/data': 0.006993712857365608, 'timing/model': 2.7808633800596, 'learning_rate/qwen_vl_interface': 8.578474177639841e-06, 'learning_rate/action_model': 8.578474177639842e-05, 'epoch': 2.6}) +[2026-06-19 23:00:37,230][starVLA.training.train_starvla][INFO] - Step 1340, Loss: {'train/loss': 0.1338861584663391, 'train/grad_norm_pre_clip': 4.6844916343688965, 'timing/data': 0.006441429257392883, 'timing/model': 2.7762136328965425, 'learning_rate/qwen_vl_interface': 8.576300813317463e-06, 'learning_rate/action_model': 8.576300813317462e-05, 'epoch': 2.6}) +[2026-06-19 23:00:48,133][starVLA.training.train_starvla][INFO] - Step 1341, Loss: {'train/loss': 0.18879400193691254, 'train/grad_norm_pre_clip': 5.419119834899902, 'timing/data': 0.006561927497386932, 'timing/model': 2.7772768698632717, 'learning_rate/qwen_vl_interface': 8.574126081678463e-06, 'learning_rate/action_model': 8.574126081678462e-05, 'epoch': 2.6}) +[2026-06-19 23:00:59,031][starVLA.training.train_starvla][INFO] - Step 1342, Loss: {'train/loss': 0.2489887773990631, 'train/grad_norm_pre_clip': 6.226896286010742, 'timing/data': 0.006934063509106636, 'timing/model': 2.775758735835552, 'learning_rate/qwen_vl_interface': 8.571949983616796e-06, 'learning_rate/action_model': 8.571949983616795e-05, 'epoch': 2.6}) +[2026-06-19 23:01:09,931][starVLA.training.train_starvla][INFO] - Step 1343, Loss: {'train/loss': 0.2533261179924011, 'train/grad_norm_pre_clip': 4.540677547454834, 'timing/data': 0.005632886663079262, 'timing/model': 2.771003257483244, 'learning_rate/qwen_vl_interface': 8.569772520026971e-06, 'learning_rate/action_model': 8.569772520026971e-05, 'epoch': 2.6}) +[2026-06-19 23:01:20,973][starVLA.training.train_starvla][INFO] - Step 1344, Loss: {'train/loss': 0.15035510063171387, 'train/grad_norm_pre_clip': 6.093794345855713, 'timing/data': 0.005891554057598114, 'timing/model': 2.918525503948331, 'learning_rate/qwen_vl_interface': 8.567593691804061e-06, 'learning_rate/action_model': 8.567593691804061e-05, 'epoch': 2.6}) +[2026-06-19 23:01:31,831][starVLA.training.train_starvla][INFO] - Step 1345, Loss: {'train/loss': 0.22871151566505432, 'train/grad_norm_pre_clip': 5.991972923278809, 'timing/data': 0.005966523662209511, 'timing/model': 2.7765401788055897, 'learning_rate/qwen_vl_interface': 8.565413499843703e-06, 'learning_rate/action_model': 8.565413499843703e-05, 'epoch': 2.61}) +[2026-06-19 23:01:42,733][starVLA.training.train_starvla][INFO] - Step 1346, Loss: {'train/loss': 0.11008656024932861, 'train/grad_norm_pre_clip': 3.0231375694274902, 'timing/data': 0.006733007729053497, 'timing/model': 2.7747299652546644, 'learning_rate/qwen_vl_interface': 8.563231945042087e-06, 'learning_rate/action_model': 8.563231945042088e-05, 'epoch': 2.61}) +[2026-06-19 23:01:53,630][starVLA.training.train_starvla][INFO] - Step 1347, Loss: {'train/loss': 0.12575119733810425, 'train/grad_norm_pre_clip': 2.9166171550750732, 'timing/data': 0.006530458107590675, 'timing/model': 2.769827239215374, 'learning_rate/qwen_vl_interface': 8.56104902829597e-06, 'learning_rate/action_model': 8.56104902829597e-05, 'epoch': 2.61}) +[2026-06-19 23:02:04,529][starVLA.training.train_starvla][INFO] - Step 1348, Loss: {'train/loss': 0.13033868372440338, 'train/grad_norm_pre_clip': 2.75221848487854, 'timing/data': 0.006598047912120819, 'timing/model': 2.754943497478962, 'learning_rate/qwen_vl_interface': 8.558864750502669e-06, 'learning_rate/action_model': 8.558864750502668e-05, 'epoch': 2.61}) +[2026-06-19 23:02:15,432][starVLA.training.train_starvla][INFO] - Step 1349, Loss: {'train/loss': 0.10687819868326187, 'train/grad_norm_pre_clip': 2.771298408508301, 'timing/data': 0.005530858412384987, 'timing/model': 2.7835433278232813, 'learning_rate/qwen_vl_interface': 8.556679112560055e-06, 'learning_rate/action_model': 8.556679112560054e-05, 'epoch': 2.61}) +[2026-06-19 23:02:27,435][starVLA.training.train_starvla][INFO] - Step 1350, Loss: {'train/loss': 0.10406547039747238, 'train/grad_norm_pre_clip': 3.8904242515563965, 'timing/data': 1.3279251884669065, 'timing/model': 2.4630784280598164, 'learning_rate/qwen_vl_interface': 8.554492115366562e-06, 'learning_rate/action_model': 8.554492115366561e-05, 'epoch': 2.62}) +[2026-06-19 23:02:59,528][starVLA.training.train_starvla][INFO] - Step 1351, Loss: {'train/loss': 0.18557719886302948, 'train/grad_norm_pre_clip': 9.058348655700684, 'timing/data': 0.01231287606060505, 'timing/model': 2.7701946925371885, 'learning_rate/qwen_vl_interface': 8.55230375982118e-06, 'learning_rate/action_model': 8.55230375982118e-05, 'epoch': 2.62}) +[2026-06-19 23:03:10,343][starVLA.training.train_starvla][INFO] - Step 1352, Loss: {'train/loss': 0.289346843957901, 'train/grad_norm_pre_clip': 9.381610870361328, 'timing/data': 0.01211208663880825, 'timing/model': 2.687143301591277, 'learning_rate/qwen_vl_interface': 8.550114046823463e-06, 'learning_rate/action_model': 8.550114046823463e-05, 'epoch': 2.62}) +[2026-06-19 23:03:21,133][starVLA.training.train_starvla][INFO] - Step 1353, Loss: {'train/loss': 0.252089262008667, 'train/grad_norm_pre_clip': 10.395166397094727, 'timing/data': 0.012082168832421303, 'timing/model': 2.7149283830076456, 'learning_rate/qwen_vl_interface': 8.547922977273518e-06, 'learning_rate/action_model': 8.547922977273518e-05, 'epoch': 2.62}) +[2026-06-19 23:03:31,931][starVLA.training.train_starvla][INFO] - Step 1354, Loss: {'train/loss': 0.0902262032032013, 'train/grad_norm_pre_clip': 10.700078964233398, 'timing/data': 0.013824671506881714, 'timing/model': 2.7334864996373653, 'learning_rate/qwen_vl_interface': 8.54573055207201e-06, 'learning_rate/action_model': 8.545730552072008e-05, 'epoch': 2.62}) +[2026-06-19 23:03:42,828][starVLA.training.train_starvla][INFO] - Step 1355, Loss: {'train/loss': 0.048797328025102615, 'train/grad_norm_pre_clip': 2.8502721786499023, 'timing/data': 0.014151116833090782, 'timing/model': 2.7683969773352146, 'learning_rate/qwen_vl_interface': 8.543536772120161e-06, 'learning_rate/action_model': 8.54353677212016e-05, 'epoch': 2.63}) +[2026-06-19 23:03:53,721][starVLA.training.train_starvla][INFO] - Step 1356, Loss: {'train/loss': 0.13622182607650757, 'train/grad_norm_pre_clip': 5.838896751403809, 'timing/data': 0.01209251768887043, 'timing/model': 2.7575125694274902, 'learning_rate/qwen_vl_interface': 8.541341638319754e-06, 'learning_rate/action_model': 8.541341638319754e-05, 'epoch': 2.63}) +[2026-06-19 23:04:04,638][starVLA.training.train_starvla][INFO] - Step 1357, Loss: {'train/loss': 0.17504344880580902, 'train/grad_norm_pre_clip': 6.874162197113037, 'timing/data': 0.012673512101173401, 'timing/model': 2.7800577208399773, 'learning_rate/qwen_vl_interface': 8.539145151573126e-06, 'learning_rate/action_model': 8.539145151573125e-05, 'epoch': 2.63}) +[2026-06-19 23:04:15,530][starVLA.training.train_starvla][INFO] - Step 1358, Loss: {'train/loss': 0.04755702614784241, 'train/grad_norm_pre_clip': 4.17603063583374, 'timing/data': 0.012257594615221024, 'timing/model': 2.7713774610310793, 'learning_rate/qwen_vl_interface': 8.536947312783167e-06, 'learning_rate/action_model': 8.536947312783166e-05, 'epoch': 2.63}) +[2026-06-19 23:04:26,429][starVLA.training.train_starvla][INFO] - Step 1359, Loss: {'train/loss': 0.21100452542304993, 'train/grad_norm_pre_clip': 21.320877075195312, 'timing/data': 0.012397903949022293, 'timing/model': 2.7675023078918457, 'learning_rate/qwen_vl_interface': 8.534748122853326e-06, 'learning_rate/action_model': 8.534748122853325e-05, 'epoch': 2.63}) +[2026-06-19 23:04:37,429][starVLA.training.train_starvla][INFO] - Step 1360, Loss: {'train/loss': 0.46028396487236023, 'train/grad_norm_pre_clip': 13.571944236755371, 'timing/data': 0.01244698278605938, 'timing/model': 2.7793543115258217, 'learning_rate/qwen_vl_interface': 8.532547582687609e-06, 'learning_rate/action_model': 8.532547582687608e-05, 'epoch': 2.64}) +[2026-06-19 23:04:48,330][starVLA.training.train_starvla][INFO] - Step 1361, Loss: {'train/loss': 0.0527416467666626, 'train/grad_norm_pre_clip': 8.54625129699707, 'timing/data': 0.01228206418454647, 'timing/model': 2.766971779987216, 'learning_rate/qwen_vl_interface': 8.530345693190573e-06, 'learning_rate/action_model': 8.530345693190572e-05, 'epoch': 2.64}) +[2026-06-19 23:04:59,230][starVLA.training.train_starvla][INFO] - Step 1362, Loss: {'train/loss': 0.1445501297712326, 'train/grad_norm_pre_clip': 8.458934783935547, 'timing/data': 0.012075034901499748, 'timing/model': 2.778892908245325, 'learning_rate/qwen_vl_interface': 8.528142455267328e-06, 'learning_rate/action_model': 8.528142455267329e-05, 'epoch': 2.64}) +[2026-06-19 23:05:10,131][starVLA.training.train_starvla][INFO] - Step 1363, Loss: {'train/loss': 0.18203160166740417, 'train/grad_norm_pre_clip': 8.92088508605957, 'timing/data': 0.01148507371544838, 'timing/model': 2.7701130639761686, 'learning_rate/qwen_vl_interface': 8.52593786982355e-06, 'learning_rate/action_model': 8.525937869823551e-05, 'epoch': 2.64}) +[2026-06-19 23:05:21,032][starVLA.training.train_starvla][INFO] - Step 1364, Loss: {'train/loss': 0.1913902312517166, 'train/grad_norm_pre_clip': 18.522472381591797, 'timing/data': 0.011619850993156433, 'timing/model': 2.773907136172056, 'learning_rate/qwen_vl_interface': 8.523731937765456e-06, 'learning_rate/action_model': 8.523731937765456e-05, 'epoch': 2.64}) +[2026-06-19 23:05:31,950][starVLA.training.train_starvla][INFO] - Step 1365, Loss: {'train/loss': 0.06837503612041473, 'train/grad_norm_pre_clip': 6.583115100860596, 'timing/data': 0.011330824345350266, 'timing/model': 2.7927592117339373, 'learning_rate/qwen_vl_interface': 8.521524659999823e-06, 'learning_rate/action_model': 8.521524659999822e-05, 'epoch': 2.65}) +[2026-06-19 23:05:42,833][starVLA.training.train_starvla][INFO] - Step 1366, Loss: {'train/loss': 0.4316628575325012, 'train/grad_norm_pre_clip': 13.812376976013184, 'timing/data': 0.012163503095507622, 'timing/model': 2.771804306656122, 'learning_rate/qwen_vl_interface': 8.519316037433976e-06, 'learning_rate/action_model': 8.519316037433977e-05, 'epoch': 2.65}) +[2026-06-19 23:05:53,730][starVLA.training.train_starvla][INFO] - Step 1367, Loss: {'train/loss': 0.19251026213169098, 'train/grad_norm_pre_clip': 12.08154296875, 'timing/data': 0.012125926092267036, 'timing/model': 2.768050706014037, 'learning_rate/qwen_vl_interface': 8.517106070975801e-06, 'learning_rate/action_model': 8.5171060709758e-05, 'epoch': 2.65}) +[2026-06-19 23:06:04,632][starVLA.training.train_starvla][INFO] - Step 1368, Loss: {'train/loss': 0.2786266505718231, 'train/grad_norm_pre_clip': 14.138290405273438, 'timing/data': 0.011973068118095398, 'timing/model': 2.7749608997255564, 'learning_rate/qwen_vl_interface': 8.514894761533729e-06, 'learning_rate/action_model': 8.514894761533728e-05, 'epoch': 2.65}) +[2026-06-19 23:06:15,517][starVLA.training.train_starvla][INFO] - Step 1369, Loss: {'train/loss': 0.04217807948589325, 'train/grad_norm_pre_clip': 4.28759765625, 'timing/data': 0.011936727911233902, 'timing/model': 2.7642494402825832, 'learning_rate/qwen_vl_interface': 8.512682110016747e-06, 'learning_rate/action_model': 8.512682110016747e-05, 'epoch': 2.65}) +[2026-06-19 23:06:26,331][starVLA.training.train_starvla][INFO] - Step 1370, Loss: {'train/loss': 0.25421780347824097, 'train/grad_norm_pre_clip': 12.407452583312988, 'timing/data': 0.012386882677674294, 'timing/model': 2.7527227755635977, 'learning_rate/qwen_vl_interface': 8.51046811733439e-06, 'learning_rate/action_model': 8.51046811733439e-05, 'epoch': 2.66}) +[2026-06-19 23:06:37,333][starVLA.training.train_starvla][INFO] - Step 1371, Loss: {'train/loss': 0.3117377758026123, 'train/grad_norm_pre_clip': 14.385468482971191, 'timing/data': 0.012260064482688904, 'timing/model': 2.715114476159215, 'learning_rate/qwen_vl_interface': 8.50825278439675e-06, 'learning_rate/action_model': 8.508252784396749e-05, 'epoch': 2.66}) +[2026-06-19 23:06:48,009][starVLA.training.train_starvla][INFO] - Step 1372, Loss: {'train/loss': 0.23440876603126526, 'train/grad_norm_pre_clip': 15.239870071411133, 'timing/data': 0.0118921659886837, 'timing/model': 2.7608801517635584, 'learning_rate/qwen_vl_interface': 8.506036112114465e-06, 'learning_rate/action_model': 8.506036112114464e-05, 'epoch': 2.66}) +[2026-06-19 23:06:58,834][starVLA.training.train_starvla][INFO] - Step 1373, Loss: {'train/loss': 0.17898401618003845, 'train/grad_norm_pre_clip': 9.509818077087402, 'timing/data': 0.01226850226521492, 'timing/model': 2.794745907187462, 'learning_rate/qwen_vl_interface': 8.503818101398725e-06, 'learning_rate/action_model': 8.503818101398725e-05, 'epoch': 2.66}) +[2026-06-19 23:07:09,525][starVLA.training.train_starvla][INFO] - Step 1374, Loss: {'train/loss': 0.260092556476593, 'train/grad_norm_pre_clip': 18.804502487182617, 'timing/data': 0.012116635218262672, 'timing/model': 2.74297814629972, 'learning_rate/qwen_vl_interface': 8.501598753161273e-06, 'learning_rate/action_model': 8.501598753161272e-05, 'epoch': 2.66}) +[2026-06-19 23:07:20,421][starVLA.training.train_starvla][INFO] - Step 1375, Loss: {'train/loss': 0.18116018176078796, 'train/grad_norm_pre_clip': 8.743950843811035, 'timing/data': 0.012785818427801132, 'timing/model': 2.758863279595971, 'learning_rate/qwen_vl_interface': 8.499378068314395e-06, 'learning_rate/action_model': 8.499378068314395e-05, 'epoch': 2.66}) +[2026-06-19 23:07:31,324][starVLA.training.train_starvla][INFO] - Step 1376, Loss: {'train/loss': 0.12109344452619553, 'train/grad_norm_pre_clip': 7.769097328186035, 'timing/data': 0.012110542505979538, 'timing/model': 2.768281841650605, 'learning_rate/qwen_vl_interface': 8.497156047770932e-06, 'learning_rate/action_model': 8.497156047770933e-05, 'epoch': 2.67}) +[2026-06-19 23:07:42,222][starVLA.training.train_starvla][INFO] - Step 1377, Loss: {'train/loss': 0.1335730254650116, 'train/grad_norm_pre_clip': 6.560985088348389, 'timing/data': 0.012672960758209229, 'timing/model': 2.7670121639966965, 'learning_rate/qwen_vl_interface': 8.494932692444276e-06, 'learning_rate/action_model': 8.494932692444276e-05, 'epoch': 2.67}) +[2026-06-19 23:07:53,132][starVLA.training.train_starvla][INFO] - Step 1378, Loss: {'train/loss': 0.13922742009162903, 'train/grad_norm_pre_clip': 8.231939315795898, 'timing/data': 0.012318143621087074, 'timing/model': 2.77344523742795, 'learning_rate/qwen_vl_interface': 8.492708003248362e-06, 'learning_rate/action_model': 8.492708003248362e-05, 'epoch': 2.67}) +[2026-06-19 23:08:04,237][starVLA.training.train_starvla][INFO] - Step 1379, Loss: {'train/loss': 0.16772957146167755, 'train/grad_norm_pre_clip': 7.063072681427002, 'timing/data': 0.01213003322482109, 'timing/model': 2.7920732125639915, 'learning_rate/qwen_vl_interface': 8.490481981097674e-06, 'learning_rate/action_model': 8.490481981097674e-05, 'epoch': 2.67}) +[2026-06-19 23:08:15,131][starVLA.training.train_starvla][INFO] - Step 1380, Loss: {'train/loss': 0.151582732796669, 'train/grad_norm_pre_clip': 10.979941368103027, 'timing/data': 0.012234874069690704, 'timing/model': 2.776039509102702, 'learning_rate/qwen_vl_interface': 8.488254626907249e-06, 'learning_rate/action_model': 8.488254626907249e-05, 'epoch': 2.67}) +[2026-06-19 23:08:26,038][starVLA.training.train_starvla][INFO] - Step 1381, Loss: {'train/loss': 0.10657594352960587, 'train/grad_norm_pre_clip': 6.920506000518799, 'timing/data': 0.012452561408281326, 'timing/model': 2.775512306019664, 'learning_rate/qwen_vl_interface': 8.486025941592664e-06, 'learning_rate/action_model': 8.486025941592664e-05, 'epoch': 2.68}) +[2026-06-19 23:08:36,938][starVLA.training.train_starvla][INFO] - Step 1382, Loss: {'train/loss': 0.35654106736183167, 'train/grad_norm_pre_clip': 16.183042526245117, 'timing/data': 0.012036904692649841, 'timing/model': 2.7838187124580145, 'learning_rate/qwen_vl_interface': 8.483795926070054e-06, 'learning_rate/action_model': 8.483795926070052e-05, 'epoch': 2.68}) +[2026-06-19 23:08:48,038][starVLA.training.train_starvla][INFO] - Step 1383, Loss: {'train/loss': 0.22776079177856445, 'train/grad_norm_pre_clip': 8.981063842773438, 'timing/data': 0.012231102213263512, 'timing/model': 2.7756108976900578, 'learning_rate/qwen_vl_interface': 8.481564581256087e-06, 'learning_rate/action_model': 8.481564581256087e-05, 'epoch': 2.68}) +[2026-06-19 23:08:58,920][starVLA.training.train_starvla][INFO] - Step 1384, Loss: {'train/loss': 0.11307558417320251, 'train/grad_norm_pre_clip': 3.8202900886535645, 'timing/data': 0.012007536366581917, 'timing/model': 2.765232030302286, 'learning_rate/qwen_vl_interface': 8.479331908067987e-06, 'learning_rate/action_model': 8.479331908067987e-05, 'epoch': 2.68}) +[2026-06-19 23:09:09,728][starVLA.training.train_starvla][INFO] - Step 1385, Loss: {'train/loss': 0.27871307730674744, 'train/grad_norm_pre_clip': 7.7124481201171875, 'timing/data': 0.011861605569720268, 'timing/model': 2.7734502591192722, 'learning_rate/qwen_vl_interface': 8.477097907423524e-06, 'learning_rate/action_model': 8.477097907423524e-05, 'epoch': 2.68}) +[2026-06-19 23:09:20,699][starVLA.training.train_starvla][INFO] - Step 1386, Loss: {'train/loss': 0.0954013392329216, 'train/grad_norm_pre_clip': 5.133251190185547, 'timing/data': 0.012298792600631714, 'timing/model': 2.83647402189672, 'learning_rate/qwen_vl_interface': 8.474862580241006e-06, 'learning_rate/action_model': 8.474862580241006e-05, 'epoch': 2.69}) +[2026-06-19 23:09:31,534][starVLA.training.train_starvla][INFO] - Step 1387, Loss: {'train/loss': 0.07396382093429565, 'train/grad_norm_pre_clip': 5.899219512939453, 'timing/data': 0.012207342311739922, 'timing/model': 2.7726638689637184, 'learning_rate/qwen_vl_interface': 8.472625927439299e-06, 'learning_rate/action_model': 8.472625927439299e-05, 'epoch': 2.69}) +[2026-06-19 23:09:42,435][starVLA.training.train_starvla][INFO] - Step 1388, Loss: {'train/loss': 0.15580902993679047, 'train/grad_norm_pre_clip': 9.078539848327637, 'timing/data': 0.01223871111869812, 'timing/model': 2.783756673336029, 'learning_rate/qwen_vl_interface': 8.4703879499378e-06, 'learning_rate/action_model': 8.4703879499378e-05, 'epoch': 2.69}) +[2026-06-19 23:09:53,333][starVLA.training.train_starvla][INFO] - Step 1389, Loss: {'train/loss': 0.03753362223505974, 'train/grad_norm_pre_clip': 2.9160401821136475, 'timing/data': 0.012326421216130257, 'timing/model': 2.7742445692420006, 'learning_rate/qwen_vl_interface': 8.46814864865646e-06, 'learning_rate/action_model': 8.46814864865646e-05, 'epoch': 2.69}) +[2026-06-19 23:10:04,470][starVLA.training.train_starvla][INFO] - Step 1390, Loss: {'train/loss': 0.2205103486776352, 'train/grad_norm_pre_clip': 8.63653564453125, 'timing/data': 0.012132642790675163, 'timing/model': 3.0047612413764, 'learning_rate/qwen_vl_interface': 8.465908024515771e-06, 'learning_rate/action_model': 8.46590802451577e-05, 'epoch': 2.69}) +[2026-06-19 23:10:15,318][starVLA.training.train_starvla][INFO] - Step 1391, Loss: {'train/loss': 0.18452490866184235, 'train/grad_norm_pre_clip': 11.544214248657227, 'timing/data': 0.012192010879516602, 'timing/model': 2.764955470338464, 'learning_rate/qwen_vl_interface': 8.463666078436769e-06, 'learning_rate/action_model': 8.463666078436769e-05, 'epoch': 2.7}) +[2026-06-19 23:10:26,227][starVLA.training.train_starvla][INFO] - Step 1392, Loss: {'train/loss': 0.28280338644981384, 'train/grad_norm_pre_clip': 10.154221534729004, 'timing/data': 0.012088144198060036, 'timing/model': 2.7632741276174784, 'learning_rate/qwen_vl_interface': 8.46142281134103e-06, 'learning_rate/action_model': 8.461422811341031e-05, 'epoch': 2.7}) +[2026-06-19 23:10:37,137][starVLA.training.train_starvla][INFO] - Step 1393, Loss: {'train/loss': 0.07249103486537933, 'train/grad_norm_pre_clip': 5.666153907775879, 'timing/data': 0.012098325416445732, 'timing/model': 2.774726439267397, 'learning_rate/qwen_vl_interface': 8.459178224150683e-06, 'learning_rate/action_model': 8.459178224150683e-05, 'epoch': 2.7}) +[2026-06-19 23:10:48,033][starVLA.training.train_starvla][INFO] - Step 1394, Loss: {'train/loss': 0.178879052400589, 'train/grad_norm_pre_clip': 9.176228523254395, 'timing/data': 0.01223452016711235, 'timing/model': 2.775097317993641, 'learning_rate/qwen_vl_interface': 8.456932317788389e-06, 'learning_rate/action_model': 8.456932317788389e-05, 'epoch': 2.7}) +[2026-06-19 23:10:58,933][starVLA.training.train_starvla][INFO] - Step 1395, Loss: {'train/loss': 0.20967631042003632, 'train/grad_norm_pre_clip': 8.81981372833252, 'timing/data': 0.012159030884504318, 'timing/model': 2.7757035084068775, 'learning_rate/qwen_vl_interface': 8.454685093177356e-06, 'learning_rate/action_model': 8.454685093177354e-05, 'epoch': 2.7}) +[2026-06-19 23:11:09,837][starVLA.training.train_starvla][INFO] - Step 1396, Loss: {'train/loss': 0.05461525917053223, 'train/grad_norm_pre_clip': 3.594377040863037, 'timing/data': 0.012011125683784485, 'timing/model': 2.7802654188126326, 'learning_rate/qwen_vl_interface': 8.452436551241328e-06, 'learning_rate/action_model': 8.452436551241328e-05, 'epoch': 2.71}) +[2026-06-19 23:11:20,731][starVLA.training.train_starvla][INFO] - Step 1397, Loss: {'train/loss': 0.20260757207870483, 'train/grad_norm_pre_clip': 10.168628692626953, 'timing/data': 0.01196250319480896, 'timing/model': 2.776035264134407, 'learning_rate/qwen_vl_interface': 8.450186692904605e-06, 'learning_rate/action_model': 8.450186692904606e-05, 'epoch': 2.71}) +[2026-06-19 23:11:31,731][starVLA.training.train_starvla][INFO] - Step 1398, Loss: {'train/loss': 0.33515840768814087, 'train/grad_norm_pre_clip': 16.06574249267578, 'timing/data': 0.012148460373282433, 'timing/model': 2.7622209787368774, 'learning_rate/qwen_vl_interface': 8.447935519092014e-06, 'learning_rate/action_model': 8.447935519092014e-05, 'epoch': 2.71}) +[2026-06-19 23:11:42,631][starVLA.training.train_starvla][INFO] - Step 1399, Loss: {'train/loss': 0.045050811022520065, 'train/grad_norm_pre_clip': 5.975949287414551, 'timing/data': 0.012276072055101395, 'timing/model': 2.7665696758776903, 'learning_rate/qwen_vl_interface': 8.445683030728925e-06, 'learning_rate/action_model': 8.445683030728925e-05, 'epoch': 2.71}) +[2026-06-19 23:11:53,521][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1400 +[2026-06-19 23:19:22,090][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1400 +[2026-06-19 23:19:22,091][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 1400 +[2026-06-19 23:27:25,213][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 1400 +[2026-06-19 23:27:25,215][starVLA.training.train_starvla][INFO] - Step 1400, Loss: {'train/loss': 0.08644905686378479, 'train/grad_norm_pre_clip': 10.299456596374512, 'eval/loss': 0.569129228591919, 'eval/latency_0/loss': 0.3193189203739166, 'eval/latency_1/loss': 0.08381705731153488, 'eval/latency_2/loss': 0.617101788520813, 'eval/latency_3/loss': 0.766507625579834, 'eval/latency_4/loss': 1.0628267526626587, 'eval/flappy/loss': 0.569129228591919, 'eval/flappy/latency_0/loss': 0.3193189203739166, 'eval/flappy/latency_1/loss': 0.08381705731153488, 'eval/flappy/latency_2/loss': 0.617101788520813, 'eval/flappy/latency_3/loss': 0.766507625579834, 'eval/flappy/latency_4/loss': 1.0628267526626587, 'eval/flap_precision': 0.5110294117647058, 'eval/flap_recall': 0.5054545454545455, 'eval/flap_f1': 0.5082266910420474, 'eval/flap_rate_model': 0.07555555555555556, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.5082266910420474, 'eval/latency_0/flap_precision': 0.5110294117647058, 'eval/latency_0/flap_recall': 0.5054545454545455, 'eval/latency_0/flap_f1': 0.5082266910420474, 'eval/latency_0/flap_rate_model': 0.07555555555555556, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.5082266910420474, 'timing/data': 0.012186102569103241, 'timing/model': 2.7760502733290195, 'learning_rate/qwen_vl_interface': 8.443429228741256e-06, 'learning_rate/action_model': 8.443429228741255e-05, 'epoch': 2.71}) +[2026-06-19 23:27:25,215][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_1400_state +[2026-06-19 23:27:25,237][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-19 23:27:30,986][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_1400_state/model.safetensors +[2026-06-19 23:27:45,007][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_1400_state/optimizer.bin +[2026-06-19 23:27:45,007][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_1400_state/sampler.bin +[2026-06-19 23:27:45,007][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_1400_state/sampler_1.bin +[2026-06-19 23:27:45,009][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_1400_state/random_states_0.pkl +[2026-06-19 23:27:45,134][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1400 +[2026-06-19 23:27:45,135][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-19 23:27:45,149][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-19 23:27:56,025][starVLA.training.train_starvla][INFO] - Step 1401, Loss: {'train/loss': 0.2810784876346588, 'train/grad_norm_pre_clip': 16.888227462768555, 'timing/data': 0.012876000255346298, 'timing/model': 2.7735819090157747, 'learning_rate/qwen_vl_interface': 8.441174114055457e-06, 'learning_rate/action_model': 8.441174114055456e-05, 'epoch': 2.72}) +[2026-06-19 23:28:07,153][starVLA.training.train_starvla][INFO] - Step 1402, Loss: {'train/loss': 0.21251752972602844, 'train/grad_norm_pre_clip': 11.202861785888672, 'timing/data': 0.0130413006991148, 'timing/model': 3.0011064000427723, 'learning_rate/qwen_vl_interface': 8.43891768759852e-06, 'learning_rate/action_model': 8.43891768759852e-05, 'epoch': 2.72}) +[2026-06-19 23:28:18,029][starVLA.training.train_starvla][INFO] - Step 1403, Loss: {'train/loss': 0.4586424231529236, 'train/grad_norm_pre_clip': 15.463888168334961, 'timing/data': 0.012438224628567696, 'timing/model': 2.780009800568223, 'learning_rate/qwen_vl_interface': 8.436659950297978e-06, 'learning_rate/action_model': 8.436659950297977e-05, 'epoch': 2.72}) +[2026-06-19 23:28:28,928][starVLA.training.train_starvla][INFO] - Step 1404, Loss: {'train/loss': 0.17607997357845306, 'train/grad_norm_pre_clip': 12.799053192138672, 'timing/data': 0.013266626745462418, 'timing/model': 2.76764908246696, 'learning_rate/qwen_vl_interface': 8.4344009030819e-06, 'learning_rate/action_model': 8.4344009030819e-05, 'epoch': 2.72}) +[2026-06-19 23:28:39,528][starVLA.training.train_starvla][INFO] - Step 1405, Loss: {'train/loss': 0.22353023290634155, 'train/grad_norm_pre_clip': 12.69201374053955, 'timing/data': 0.012388724833726883, 'timing/model': 2.7514331359416246, 'learning_rate/qwen_vl_interface': 8.432140546878899e-06, 'learning_rate/action_model': 8.432140546878898e-05, 'epoch': 2.72}) +[2026-06-19 23:28:50,620][starVLA.training.train_starvla][INFO] - Step 1406, Loss: {'train/loss': 0.10056553035974503, 'train/grad_norm_pre_clip': 5.081244945526123, 'timing/data': 0.012687981128692627, 'timing/model': 2.709259131923318, 'learning_rate/qwen_vl_interface': 8.429878882618116e-06, 'learning_rate/action_model': 8.429878882618116e-05, 'epoch': 2.73}) +[2026-06-19 23:29:01,536][starVLA.training.train_starvla][INFO] - Step 1407, Loss: {'train/loss': 0.15519596636295319, 'train/grad_norm_pre_clip': 11.54000186920166, 'timing/data': 0.02285558357834816, 'timing/model': 2.745227877050638, 'learning_rate/qwen_vl_interface': 8.42761591122924e-06, 'learning_rate/action_model': 8.427615911229241e-05, 'epoch': 2.73}) +[2026-06-19 23:29:12,246][starVLA.training.train_starvla][INFO] - Step 1408, Loss: {'train/loss': 0.05227020010352135, 'train/grad_norm_pre_clip': 3.7590408325195312, 'timing/data': 0.013158967718482018, 'timing/model': 2.6867347229272127, 'learning_rate/qwen_vl_interface': 8.425351633642491e-06, 'learning_rate/action_model': 8.425351633642492e-05, 'epoch': 2.73}) +[2026-06-19 23:29:22,747][starVLA.training.train_starvla][INFO] - Step 1409, Loss: {'train/loss': 0.17940078675746918, 'train/grad_norm_pre_clip': 11.255916595458984, 'timing/data': 0.012697571888566017, 'timing/model': 2.666469158604741, 'learning_rate/qwen_vl_interface': 8.423086050788632e-06, 'learning_rate/action_model': 8.423086050788631e-05, 'epoch': 2.73}) +[2026-06-19 23:29:33,431][starVLA.training.train_starvla][INFO] - Step 1410, Loss: {'train/loss': 0.12711194157600403, 'train/grad_norm_pre_clip': 8.933703422546387, 'timing/data': 0.012454517185688019, 'timing/model': 2.718618931248784, 'learning_rate/qwen_vl_interface': 8.420819163598953e-06, 'learning_rate/action_model': 8.420819163598952e-05, 'epoch': 2.73}) +[2026-06-19 23:29:44,331][starVLA.training.train_starvla][INFO] - Step 1411, Loss: {'train/loss': 0.11736109852790833, 'train/grad_norm_pre_clip': 9.467523574829102, 'timing/data': 0.012566976249217987, 'timing/model': 2.77031890489161, 'learning_rate/qwen_vl_interface': 8.418550973005289e-06, 'learning_rate/action_model': 8.418550973005289e-05, 'epoch': 2.73}) +[2026-06-19 23:29:55,219][starVLA.training.train_starvla][INFO] - Step 1412, Loss: {'train/loss': 0.16678164899349213, 'train/grad_norm_pre_clip': 12.522843360900879, 'timing/data': 0.01227838546037674, 'timing/model': 2.7875315975397825, 'learning_rate/qwen_vl_interface': 8.416281479940005e-06, 'learning_rate/action_model': 8.416281479940005e-05, 'epoch': 2.74}) +[2026-06-19 23:30:05,956][starVLA.training.train_starvla][INFO] - Step 1413, Loss: {'train/loss': 0.1177753135561943, 'train/grad_norm_pre_clip': 8.701865196228027, 'timing/data': 0.01260945200920105, 'timing/model': 2.6451484318822622, 'learning_rate/qwen_vl_interface': 8.414010685336008e-06, 'learning_rate/action_model': 8.414010685336008e-05, 'epoch': 2.74}) +[2026-06-19 23:30:16,863][starVLA.training.train_starvla][INFO] - Step 1414, Loss: {'train/loss': 0.17316611111164093, 'train/grad_norm_pre_clip': 18.055200576782227, 'timing/data': 0.012413596734404564, 'timing/model': 2.702479986473918, 'learning_rate/qwen_vl_interface': 8.411738590126733e-06, 'learning_rate/action_model': 8.411738590126732e-05, 'epoch': 2.74}) +[2026-06-19 23:30:27,621][starVLA.training.train_starvla][INFO] - Step 1415, Loss: {'train/loss': 0.12508438527584076, 'train/grad_norm_pre_clip': 12.687675476074219, 'timing/data': 0.012583581730723381, 'timing/model': 2.7608055397868156, 'learning_rate/qwen_vl_interface': 8.409465195246152e-06, 'learning_rate/action_model': 8.409465195246152e-05, 'epoch': 2.74}) +[2026-06-19 23:30:38,388][starVLA.training.train_starvla][INFO] - Step 1416, Loss: {'train/loss': 0.13560177385807037, 'train/grad_norm_pre_clip': 5.885149955749512, 'timing/data': 0.012545771896839142, 'timing/model': 2.7353635542094707, 'learning_rate/qwen_vl_interface': 8.407190501628772e-06, 'learning_rate/action_model': 8.407190501628772e-05, 'epoch': 2.74}) +[2026-06-19 23:30:49,149][starVLA.training.train_starvla][INFO] - Step 1417, Loss: {'train/loss': 0.2293185293674469, 'train/grad_norm_pre_clip': 7.834150791168213, 'timing/data': 0.012548364698886871, 'timing/model': 2.6936973053961992, 'learning_rate/qwen_vl_interface': 8.404914510209635e-06, 'learning_rate/action_model': 8.404914510209635e-05, 'epoch': 2.75}) +[2026-06-19 23:30:59,801][starVLA.training.train_starvla][INFO] - Step 1418, Loss: {'train/loss': 0.16384164988994598, 'train/grad_norm_pre_clip': 9.914412498474121, 'timing/data': 0.012584922835230827, 'timing/model': 2.721123669296503, 'learning_rate/qwen_vl_interface': 8.402637221924315e-06, 'learning_rate/action_model': 8.402637221924314e-05, 'epoch': 2.75}) +[2026-06-19 23:31:10,467][starVLA.training.train_starvla][INFO] - Step 1419, Loss: {'train/loss': 0.0856839269399643, 'train/grad_norm_pre_clip': 5.639561176300049, 'timing/data': 0.012296067550778389, 'timing/model': 2.6930060777813196, 'learning_rate/qwen_vl_interface': 8.400358637708919e-06, 'learning_rate/action_model': 8.400358637708918e-05, 'epoch': 2.75}) +[2026-06-19 23:31:21,324][starVLA.training.train_starvla][INFO] - Step 1420, Loss: {'train/loss': 0.3476860821247101, 'train/grad_norm_pre_clip': 9.866313934326172, 'timing/data': 0.012413574382662773, 'timing/model': 2.7745719123631716, 'learning_rate/qwen_vl_interface': 8.398078758500086e-06, 'learning_rate/action_model': 8.398078758500086e-05, 'epoch': 2.75}) +[2026-06-19 23:31:32,038][starVLA.training.train_starvla][INFO] - Step 1421, Loss: {'train/loss': 0.019684478640556335, 'train/grad_norm_pre_clip': 1.2856885194778442, 'timing/data': 0.012469926849007607, 'timing/model': 2.728654956445098, 'learning_rate/qwen_vl_interface': 8.395797585234987e-06, 'learning_rate/action_model': 8.395797585234988e-05, 'epoch': 2.75}) +[2026-06-19 23:31:42,931][starVLA.training.train_starvla][INFO] - Step 1422, Loss: {'train/loss': 0.19782128930091858, 'train/grad_norm_pre_clip': 11.634883880615234, 'timing/data': 0.012424465268850327, 'timing/model': 2.778281595557928, 'learning_rate/qwen_vl_interface': 8.39351511885133e-06, 'learning_rate/action_model': 8.393515118851329e-05, 'epoch': 2.76}) +[2026-06-19 23:31:53,830][starVLA.training.train_starvla][INFO] - Step 1423, Loss: {'train/loss': 0.11941283196210861, 'train/grad_norm_pre_clip': 4.73797082901001, 'timing/data': 0.01289748027920723, 'timing/model': 2.7650505676865578, 'learning_rate/qwen_vl_interface': 8.391231360287348e-06, 'learning_rate/action_model': 8.391231360287347e-05, 'epoch': 2.76}) +[2026-06-19 23:32:04,736][starVLA.training.train_starvla][INFO] - Step 1424, Loss: {'train/loss': 0.25248590111732483, 'train/grad_norm_pre_clip': 7.952041149139404, 'timing/data': 0.012667372822761536, 'timing/model': 2.7728365808725357, 'learning_rate/qwen_vl_interface': 8.388946310481807e-06, 'learning_rate/action_model': 8.388946310481807e-05, 'epoch': 2.76}) +[2026-06-19 23:32:15,733][starVLA.training.train_starvla][INFO] - Step 1425, Loss: {'train/loss': 0.04455486685037613, 'train/grad_norm_pre_clip': 4.251411437988281, 'timing/data': 0.012887287884950638, 'timing/model': 2.7912406399846077, 'learning_rate/qwen_vl_interface': 8.386659970374007e-06, 'learning_rate/action_model': 8.386659970374006e-05, 'epoch': 2.76}) +[2026-06-19 23:32:26,636][starVLA.training.train_starvla][INFO] - Step 1426, Loss: {'train/loss': 0.11133027821779251, 'train/grad_norm_pre_clip': 6.757406711578369, 'timing/data': 0.01261216402053833, 'timing/model': 2.7664343398064375, 'learning_rate/qwen_vl_interface': 8.384372340903775e-06, 'learning_rate/action_model': 8.384372340903775e-05, 'epoch': 2.76}) +[2026-06-19 23:32:37,533][starVLA.training.train_starvla][INFO] - Step 1427, Loss: {'train/loss': 0.10291063785552979, 'train/grad_norm_pre_clip': 13.397905349731445, 'timing/data': 0.01206211932003498, 'timing/model': 2.7665267176926136, 'learning_rate/qwen_vl_interface': 8.382083423011472e-06, 'learning_rate/action_model': 8.382083423011471e-05, 'epoch': 2.77}) +[2026-06-19 23:32:48,435][starVLA.training.train_starvla][INFO] - Step 1428, Loss: {'train/loss': 0.1322629451751709, 'train/grad_norm_pre_clip': 13.522815704345703, 'timing/data': 0.012504171580076218, 'timing/model': 2.7797277197241783, 'learning_rate/qwen_vl_interface': 8.379793217637979e-06, 'learning_rate/action_model': 8.379793217637979e-05, 'epoch': 2.77}) +[2026-06-19 23:32:59,335][starVLA.training.train_starvla][INFO] - Step 1429, Loss: {'train/loss': 0.2934275269508362, 'train/grad_norm_pre_clip': 11.435909271240234, 'timing/data': 0.012398354709148407, 'timing/model': 2.772073382511735, 'learning_rate/qwen_vl_interface': 8.377501725724719e-06, 'learning_rate/action_model': 8.377501725724719e-05, 'epoch': 2.77}) +[2026-06-19 23:33:10,230][starVLA.training.train_starvla][INFO] - Step 1430, Loss: {'train/loss': 0.07676155120134354, 'train/grad_norm_pre_clip': 6.0267815589904785, 'timing/data': 0.011875161901116371, 'timing/model': 2.7656743600964546, 'learning_rate/qwen_vl_interface': 8.375208948213636e-06, 'learning_rate/action_model': 8.375208948213635e-05, 'epoch': 2.77}) +[2026-06-19 23:33:21,138][starVLA.training.train_starvla][INFO] - Step 1431, Loss: {'train/loss': 0.329112708568573, 'train/grad_norm_pre_clip': 13.783997535705566, 'timing/data': 0.013393282890319824, 'timing/model': 2.786612842231989, 'learning_rate/qwen_vl_interface': 8.372914886047203e-06, 'learning_rate/action_model': 8.372914886047203e-05, 'epoch': 2.77}) +[2026-06-19 23:33:32,035][starVLA.training.train_starvla][INFO] - Step 1432, Loss: {'train/loss': 0.3416786789894104, 'train/grad_norm_pre_clip': 22.261962890625, 'timing/data': 0.013046298176050186, 'timing/model': 2.765119591727853, 'learning_rate/qwen_vl_interface': 8.370619540168422e-06, 'learning_rate/action_model': 8.370619540168423e-05, 'epoch': 2.78}) +[2026-06-19 23:33:43,127][starVLA.training.train_starvla][INFO] - Step 1433, Loss: {'train/loss': 0.10042090713977814, 'train/grad_norm_pre_clip': 7.1350932121276855, 'timing/data': 0.013145169243216515, 'timing/model': 2.765238620340824, 'learning_rate/qwen_vl_interface': 8.368322911520826e-06, 'learning_rate/action_model': 8.368322911520826e-05, 'epoch': 2.78}) +[2026-06-19 23:33:54,033][starVLA.training.train_starvla][INFO] - Step 1434, Loss: {'train/loss': 0.24907147884368896, 'train/grad_norm_pre_clip': 12.833024978637695, 'timing/data': 0.012546025216579437, 'timing/model': 2.7726934608072042, 'learning_rate/qwen_vl_interface': 8.36602500104847e-06, 'learning_rate/action_model': 8.36602500104847e-05, 'epoch': 2.78}) +[2026-06-19 23:34:04,923][starVLA.training.train_starvla][INFO] - Step 1435, Loss: {'train/loss': 0.10758595913648605, 'train/grad_norm_pre_clip': 6.7966108322143555, 'timing/data': 0.012461384758353233, 'timing/model': 2.7659065891057253, 'learning_rate/qwen_vl_interface': 8.36372580969594e-06, 'learning_rate/action_model': 8.36372580969594e-05, 'epoch': 2.78}) +[2026-06-19 23:34:15,837][starVLA.training.train_starvla][INFO] - Step 1436, Loss: {'train/loss': 0.2200348675251007, 'train/grad_norm_pre_clip': 10.803685188293457, 'timing/data': 0.012566592544317245, 'timing/model': 2.7771712355315685, 'learning_rate/qwen_vl_interface': 8.361425338408344e-06, 'learning_rate/action_model': 8.361425338408343e-05, 'epoch': 2.78}) +[2026-06-19 23:34:26,735][starVLA.training.train_starvla][INFO] - Step 1437, Loss: {'train/loss': 0.48038673400878906, 'train/grad_norm_pre_clip': 15.87341022491455, 'timing/data': 0.013183444738388062, 'timing/model': 2.769062152132392, 'learning_rate/qwen_vl_interface': 8.359123588131319e-06, 'learning_rate/action_model': 8.359123588131319e-05, 'epoch': 2.79}) +[2026-06-19 23:34:37,628][starVLA.training.train_starvla][INFO] - Step 1438, Loss: {'train/loss': 0.33719098567962646, 'train/grad_norm_pre_clip': 16.579914093017578, 'timing/data': 0.012946870177984238, 'timing/model': 2.767431028187275, 'learning_rate/qwen_vl_interface': 8.35682055981103e-06, 'learning_rate/action_model': 8.35682055981103e-05, 'epoch': 2.79}) +[2026-06-19 23:34:48,535][starVLA.training.train_starvla][INFO] - Step 1439, Loss: {'train/loss': 0.12654532492160797, 'train/grad_norm_pre_clip': 6.237405300140381, 'timing/data': 0.013069000095129013, 'timing/model': 2.7744005117565393, 'learning_rate/qwen_vl_interface': 8.354516254394163e-06, 'learning_rate/action_model': 8.354516254394162e-05, 'epoch': 2.79}) +[2026-06-19 23:34:59,431][starVLA.training.train_starvla][INFO] - Step 1440, Loss: {'train/loss': 0.13523755967617035, 'train/grad_norm_pre_clip': 7.4785075187683105, 'timing/data': 0.012408846989274025, 'timing/model': 2.7625248599797487, 'learning_rate/qwen_vl_interface': 8.352210672827932e-06, 'learning_rate/action_model': 8.352210672827931e-05, 'epoch': 2.79}) +[2026-06-19 23:35:10,432][starVLA.training.train_starvla][INFO] - Step 1441, Loss: {'train/loss': 0.19781847298145294, 'train/grad_norm_pre_clip': 5.884148120880127, 'timing/data': 0.012642383575439453, 'timing/model': 2.765017155557871, 'learning_rate/qwen_vl_interface': 8.349903816060074e-06, 'learning_rate/action_model': 8.349903816060075e-05, 'epoch': 2.79}) +[2026-06-19 23:35:21,335][starVLA.training.train_starvla][INFO] - Step 1442, Loss: {'train/loss': 0.16732150316238403, 'train/grad_norm_pre_clip': 7.002877235412598, 'timing/data': 0.012745941057801247, 'timing/model': 2.76904483884573, 'learning_rate/qwen_vl_interface': 8.34759568503885e-06, 'learning_rate/action_model': 8.347595685038851e-05, 'epoch': 2.79}) +[2026-06-19 23:35:32,219][starVLA.training.train_starvla][INFO] - Step 1443, Loss: {'train/loss': 0.1871192306280136, 'train/grad_norm_pre_clip': 6.232399940490723, 'timing/data': 0.013112157583236694, 'timing/model': 2.758812988176942, 'learning_rate/qwen_vl_interface': 8.345286280713049e-06, 'learning_rate/action_model': 8.345286280713048e-05, 'epoch': 2.8}) +[2026-06-19 23:35:43,138][starVLA.training.train_starvla][INFO] - Step 1444, Loss: {'train/loss': 0.026365263387560844, 'train/grad_norm_pre_clip': 1.7641175985336304, 'timing/data': 0.013084936887025833, 'timing/model': 2.784998182207346, 'learning_rate/qwen_vl_interface': 8.342975604031974e-06, 'learning_rate/action_model': 8.342975604031974e-05, 'epoch': 2.8}) +[2026-06-19 23:35:54,026][starVLA.training.train_starvla][INFO] - Step 1445, Loss: {'train/loss': 0.12516413629055023, 'train/grad_norm_pre_clip': 5.14547061920166, 'timing/data': 0.01283595897257328, 'timing/model': 2.7642604038119316, 'learning_rate/qwen_vl_interface': 8.340663655945462e-06, 'learning_rate/action_model': 8.340663655945462e-05, 'epoch': 2.8}) +[2026-06-19 23:36:04,936][starVLA.training.train_starvla][INFO] - Step 1446, Loss: {'train/loss': 0.09977921098470688, 'train/grad_norm_pre_clip': 5.03214693069458, 'timing/data': 0.014773387461900711, 'timing/model': 2.77062164619565, 'learning_rate/qwen_vl_interface': 8.338350437403866e-06, 'learning_rate/action_model': 8.338350437403865e-05, 'epoch': 2.8}) +[2026-06-19 23:36:15,828][starVLA.training.train_starvla][INFO] - Step 1447, Loss: {'train/loss': 0.05258764699101448, 'train/grad_norm_pre_clip': 2.7335212230682373, 'timing/data': 0.013040337711572647, 'timing/model': 2.7747904285788536, 'learning_rate/qwen_vl_interface': 8.336035949358063e-06, 'learning_rate/action_model': 8.336035949358062e-05, 'epoch': 2.8}) +[2026-06-19 23:36:26,960][starVLA.training.train_starvla][INFO] - Step 1448, Loss: {'train/loss': 0.18962004780769348, 'train/grad_norm_pre_clip': 7.363434791564941, 'timing/data': 0.012276485562324524, 'timing/model': 2.9976042602211237, 'learning_rate/qwen_vl_interface': 8.333720192759448e-06, 'learning_rate/action_model': 8.333720192759448e-05, 'epoch': 2.81}) +[2026-06-19 23:36:37,764][starVLA.training.train_starvla][INFO] - Step 1449, Loss: {'train/loss': 0.38128387928009033, 'train/grad_norm_pre_clip': 14.300926208496094, 'timing/data': 0.01248125359416008, 'timing/model': 2.7016200236976147, 'learning_rate/qwen_vl_interface': 8.331403168559946e-06, 'learning_rate/action_model': 8.331403168559945e-05, 'epoch': 2.81}) +[2026-06-19 23:36:48,630][starVLA.training.train_starvla][INFO] - Step 1450, Loss: {'train/loss': 0.06736814975738525, 'train/grad_norm_pre_clip': 14.017552375793457, 'timing/data': 0.011998139321804047, 'timing/model': 2.779549891129136, 'learning_rate/qwen_vl_interface': 8.329084877711995e-06, 'learning_rate/action_model': 8.329084877711996e-05, 'epoch': 2.81}) +[2026-06-19 23:36:59,426][starVLA.training.train_starvla][INFO] - Step 1451, Loss: {'train/loss': 0.07308795303106308, 'train/grad_norm_pre_clip': 2.391908645629883, 'timing/data': 0.012255627661943436, 'timing/model': 2.7609923984855413, 'learning_rate/qwen_vl_interface': 8.32676532116856e-06, 'learning_rate/action_model': 8.326765321168559e-05, 'epoch': 2.81}) +[2026-06-19 23:37:10,522][starVLA.training.train_starvla][INFO] - Step 1452, Loss: {'train/loss': 0.14877665042877197, 'train/grad_norm_pre_clip': 8.829833030700684, 'timing/data': 0.01237693801522255, 'timing/model': 2.7788331639021635, 'learning_rate/qwen_vl_interface': 8.324444499883119e-06, 'learning_rate/action_model': 8.324444499883119e-05, 'epoch': 2.81}) +[2026-06-19 23:37:21,429][starVLA.training.train_starvla][INFO] - Step 1453, Loss: {'train/loss': 0.07782068103551865, 'train/grad_norm_pre_clip': 5.605858325958252, 'timing/data': 0.01246727630496025, 'timing/model': 2.7712425850331783, 'learning_rate/qwen_vl_interface': 8.322122414809678e-06, 'learning_rate/action_model': 8.322122414809677e-05, 'epoch': 2.82}) +[2026-06-19 23:37:32,334][starVLA.training.train_starvla][INFO] - Step 1454, Loss: {'train/loss': 0.19408617913722992, 'train/grad_norm_pre_clip': 8.26841926574707, 'timing/data': 0.012334136292338371, 'timing/model': 2.773323206230998, 'learning_rate/qwen_vl_interface': 8.319799066902755e-06, 'learning_rate/action_model': 8.319799066902755e-05, 'epoch': 2.82}) +[2026-06-19 23:37:43,222][starVLA.training.train_starvla][INFO] - Step 1455, Loss: {'train/loss': 0.08564821630716324, 'train/grad_norm_pre_clip': 7.58897590637207, 'timing/data': 0.012290915474295616, 'timing/model': 2.755162151530385, 'learning_rate/qwen_vl_interface': 8.317474457117393e-06, 'learning_rate/action_model': 8.317474457117392e-05, 'epoch': 2.82}) +[2026-06-19 23:37:54,332][starVLA.training.train_starvla][INFO] - Step 1456, Loss: {'train/loss': 0.10181085765361786, 'train/grad_norm_pre_clip': 4.561152458190918, 'timing/data': 0.012865303084254265, 'timing/model': 2.764729021117091, 'learning_rate/qwen_vl_interface': 8.315148586409148e-06, 'learning_rate/action_model': 8.315148586409148e-05, 'epoch': 2.82}) +[2026-06-19 23:38:05,233][starVLA.training.train_starvla][INFO] - Step 1457, Loss: {'train/loss': 0.4031858444213867, 'train/grad_norm_pre_clip': 14.070077896118164, 'timing/data': 0.012781241908669472, 'timing/model': 2.780546536669135, 'learning_rate/qwen_vl_interface': 8.312821455734104e-06, 'learning_rate/action_model': 8.312821455734103e-05, 'epoch': 2.82}) +[2026-06-19 23:38:16,131][starVLA.training.train_starvla][INFO] - Step 1458, Loss: {'train/loss': 0.22627420723438263, 'train/grad_norm_pre_clip': 9.042252540588379, 'timing/data': 0.0129015501588583, 'timing/model': 2.7715740725398064, 'learning_rate/qwen_vl_interface': 8.31049306604885e-06, 'learning_rate/action_model': 8.310493066048849e-05, 'epoch': 2.83}) +[2026-06-19 23:38:27,257][starVLA.training.train_starvla][INFO] - Step 1459, Loss: {'train/loss': 0.15361671149730682, 'train/grad_norm_pre_clip': 7.00474214553833, 'timing/data': 0.012566691264510155, 'timing/model': 2.996565617620945, 'learning_rate/qwen_vl_interface': 8.3081634183105e-06, 'learning_rate/action_model': 8.308163418310501e-05, 'epoch': 2.83}) +[2026-06-19 23:38:38,134][starVLA.training.train_starvla][INFO] - Step 1460, Loss: {'train/loss': 0.08337290585041046, 'train/grad_norm_pre_clip': 3.8060333728790283, 'timing/data': 0.01242523081600666, 'timing/model': 2.7793813291937113, 'learning_rate/qwen_vl_interface': 8.305832513476689e-06, 'learning_rate/action_model': 8.305832513476688e-05, 'epoch': 2.83}) +[2026-06-19 23:38:49,023][starVLA.training.train_starvla][INFO] - Step 1461, Loss: {'train/loss': 0.07059614360332489, 'train/grad_norm_pre_clip': 3.708143711090088, 'timing/data': 0.012872567400336266, 'timing/model': 2.7706927359104156, 'learning_rate/qwen_vl_interface': 8.303500352505557e-06, 'learning_rate/action_model': 8.303500352505557e-05, 'epoch': 2.83}) +[2026-06-19 23:38:59,834][starVLA.training.train_starvla][INFO] - Step 1462, Loss: {'train/loss': 0.0667375847697258, 'train/grad_norm_pre_clip': 19.486854553222656, 'timing/data': 0.012887796387076378, 'timing/model': 2.7527555618435144, 'learning_rate/qwen_vl_interface': 8.301166936355773e-06, 'learning_rate/action_model': 8.301166936355773e-05, 'epoch': 2.83}) +[2026-06-19 23:39:10,732][starVLA.training.train_starvla][INFO] - Step 1463, Loss: {'train/loss': 0.05733616277575493, 'train/grad_norm_pre_clip': 2.8772246837615967, 'timing/data': 0.012762688100337982, 'timing/model': 2.7855215948075056, 'learning_rate/qwen_vl_interface': 8.298832265986514e-06, 'learning_rate/action_model': 8.298832265986513e-05, 'epoch': 2.84}) +[2026-06-19 23:39:21,625][starVLA.training.train_starvla][INFO] - Step 1464, Loss: {'train/loss': 0.16741319000720978, 'train/grad_norm_pre_clip': 13.544400215148926, 'timing/data': 0.012483000755310059, 'timing/model': 2.7719653509557247, 'learning_rate/qwen_vl_interface': 8.296496342357475e-06, 'learning_rate/action_model': 8.296496342357474e-05, 'epoch': 2.84}) +[2026-06-19 23:39:32,432][starVLA.training.train_starvla][INFO] - Step 1465, Loss: {'train/loss': 0.1173849031329155, 'train/grad_norm_pre_clip': 7.010702610015869, 'timing/data': 0.012730127200484276, 'timing/model': 2.7697503864765167, 'learning_rate/qwen_vl_interface': 8.294159166428864e-06, 'learning_rate/action_model': 8.294159166428864e-05, 'epoch': 2.84}) +[2026-06-19 23:39:43,236][starVLA.training.train_starvla][INFO] - Step 1466, Loss: {'train/loss': 0.2924193739891052, 'train/grad_norm_pre_clip': 11.698824882507324, 'timing/data': 0.012472297996282578, 'timing/model': 2.7756313141435385, 'learning_rate/qwen_vl_interface': 8.29182073916141e-06, 'learning_rate/action_model': 8.29182073916141e-05, 'epoch': 2.84}) +[2026-06-19 23:39:54,132][starVLA.training.train_starvla][INFO] - Step 1467, Loss: {'train/loss': 0.19360533356666565, 'train/grad_norm_pre_clip': 11.918675422668457, 'timing/data': 0.012402258813381195, 'timing/model': 2.768506109714508, 'learning_rate/qwen_vl_interface': 8.28948106151635e-06, 'learning_rate/action_model': 8.28948106151635e-05, 'epoch': 2.84}) +[2026-06-19 23:40:05,029][starVLA.training.train_starvla][INFO] - Step 1468, Loss: {'train/loss': 0.12837406992912292, 'train/grad_norm_pre_clip': 9.699746131896973, 'timing/data': 0.012506719678640366, 'timing/model': 2.7616092041134834, 'learning_rate/qwen_vl_interface': 8.287140134455435e-06, 'learning_rate/action_model': 8.287140134455435e-05, 'epoch': 2.85}) +[2026-06-19 23:40:15,948][starVLA.training.train_starvla][INFO] - Step 1469, Loss: {'train/loss': 0.05899069085717201, 'train/grad_norm_pre_clip': 4.711939811706543, 'timing/data': 0.012523636221885681, 'timing/model': 2.7812106627970934, 'learning_rate/qwen_vl_interface': 8.284797958940934e-06, 'learning_rate/action_model': 8.284797958940935e-05, 'epoch': 2.85}) +[2026-06-19 23:40:26,834][starVLA.training.train_starvla][INFO] - Step 1470, Loss: {'train/loss': 0.17323529720306396, 'train/grad_norm_pre_clip': 9.883670806884766, 'timing/data': 0.01253516599535942, 'timing/model': 2.7709958404302597, 'learning_rate/qwen_vl_interface': 8.282454535935626e-06, 'learning_rate/action_model': 8.282454535935626e-05, 'epoch': 2.85}) +[2026-06-19 23:40:37,829][starVLA.training.train_starvla][INFO] - Step 1471, Loss: {'train/loss': 0.15117621421813965, 'train/grad_norm_pre_clip': 10.004716873168945, 'timing/data': 0.012426625937223434, 'timing/model': 2.7729618716984987, 'learning_rate/qwen_vl_interface': 8.280109866402805e-06, 'learning_rate/action_model': 8.280109866402806e-05, 'epoch': 2.85}) +[2026-06-19 23:40:48,734][starVLA.training.train_starvla][INFO] - Step 1472, Loss: {'train/loss': 0.27317386865615845, 'train/grad_norm_pre_clip': 13.244982719421387, 'timing/data': 0.012480827048420906, 'timing/model': 2.7714674063026905, 'learning_rate/qwen_vl_interface': 8.277763951306277e-06, 'learning_rate/action_model': 8.277763951306277e-05, 'epoch': 2.85}) +[2026-06-19 23:40:59,636][starVLA.training.train_starvla][INFO] - Step 1473, Loss: {'train/loss': 0.09255527704954147, 'train/grad_norm_pre_clip': 6.052670955657959, 'timing/data': 0.012524055317044258, 'timing/model': 2.775230774655938, 'learning_rate/qwen_vl_interface': 8.275416791610356e-06, 'learning_rate/action_model': 8.275416791610355e-05, 'epoch': 2.85}) +[2026-06-19 23:41:10,529][starVLA.training.train_starvla][INFO] - Step 1474, Loss: {'train/loss': 0.14297884702682495, 'train/grad_norm_pre_clip': 9.713521957397461, 'timing/data': 0.01219770684838295, 'timing/model': 2.7569981049746275, 'learning_rate/qwen_vl_interface': 8.27306838827987e-06, 'learning_rate/action_model': 8.273068388279871e-05, 'epoch': 2.86}) +[2026-06-19 23:41:21,438][starVLA.training.train_starvla][INFO] - Step 1475, Loss: {'train/loss': 0.13810855150222778, 'train/grad_norm_pre_clip': 8.426998138427734, 'timing/data': 0.029270758852362633, 'timing/model': 2.74614280462265, 'learning_rate/qwen_vl_interface': 8.270718742280162e-06, 'learning_rate/action_model': 8.270718742280162e-05, 'epoch': 2.86}) +[2026-06-19 23:41:32,322][starVLA.training.train_starvla][INFO] - Step 1476, Loss: {'train/loss': 0.28796470165252686, 'train/grad_norm_pre_clip': 18.933422088623047, 'timing/data': 0.012824200093746185, 'timing/model': 2.7606403082609177, 'learning_rate/qwen_vl_interface': 8.268367854577084e-06, 'learning_rate/action_model': 8.268367854577082e-05, 'epoch': 2.86}) +[2026-06-19 23:41:43,149][starVLA.training.train_starvla][INFO] - Step 1477, Loss: {'train/loss': 0.28786852955818176, 'train/grad_norm_pre_clip': 28.99149513244629, 'timing/data': 0.013197619467973709, 'timing/model': 2.6932559572160244, 'learning_rate/qwen_vl_interface': 8.266015726136991e-06, 'learning_rate/action_model': 8.266015726136991e-05, 'epoch': 2.86}) +[2026-06-19 23:41:53,934][starVLA.training.train_starvla][INFO] - Step 1478, Loss: {'train/loss': 0.04120069369673729, 'train/grad_norm_pre_clip': 3.192188262939453, 'timing/data': 0.012220235541462898, 'timing/model': 2.76497302390635, 'learning_rate/qwen_vl_interface': 8.26366235792676e-06, 'learning_rate/action_model': 8.26366235792676e-05, 'epoch': 2.86}) +[2026-06-19 23:42:05,023][starVLA.training.train_starvla][INFO] - Step 1479, Loss: {'train/loss': 0.08326012641191483, 'train/grad_norm_pre_clip': 7.604618072509766, 'timing/data': 0.01249624416232109, 'timing/model': 2.759931392967701, 'learning_rate/qwen_vl_interface': 8.261307750913769e-06, 'learning_rate/action_model': 8.261307750913769e-05, 'epoch': 2.87}) +[2026-06-19 23:42:15,929][starVLA.training.train_starvla][INFO] - Step 1480, Loss: {'train/loss': 0.1912539303302765, 'train/grad_norm_pre_clip': 5.765068531036377, 'timing/data': 0.012662872672080994, 'timing/model': 2.775570210069418, 'learning_rate/qwen_vl_interface': 8.258951906065908e-06, 'learning_rate/action_model': 8.258951906065908e-05, 'epoch': 2.87}) +[2026-06-19 23:42:26,834][starVLA.training.train_starvla][INFO] - Step 1481, Loss: {'train/loss': 0.062086157500743866, 'train/grad_norm_pre_clip': 3.137481689453125, 'timing/data': 0.012646310031414032, 'timing/model': 2.767894558608532, 'learning_rate/qwen_vl_interface': 8.256594824351577e-06, 'learning_rate/action_model': 8.256594824351576e-05, 'epoch': 2.87}) +[2026-06-19 23:42:37,728][starVLA.training.train_starvla][INFO] - Step 1482, Loss: {'train/loss': 0.12412574142217636, 'train/grad_norm_pre_clip': 6.428948879241943, 'timing/data': 0.012495292350649834, 'timing/model': 2.7674030121415854, 'learning_rate/qwen_vl_interface': 8.254236506739683e-06, 'learning_rate/action_model': 8.254236506739681e-05, 'epoch': 2.87}) +[2026-06-19 23:42:48,630][starVLA.training.train_starvla][INFO] - Step 1483, Loss: {'train/loss': 0.08933938294649124, 'train/grad_norm_pre_clip': 8.666231155395508, 'timing/data': 0.012407811358571053, 'timing/model': 2.7740573808550835, 'learning_rate/qwen_vl_interface': 8.25187695419964e-06, 'learning_rate/action_model': 8.25187695419964e-05, 'epoch': 2.87}) +[2026-06-19 23:42:59,534][starVLA.training.train_starvla][INFO] - Step 1484, Loss: {'train/loss': 0.18729470670223236, 'train/grad_norm_pre_clip': 10.019632339477539, 'timing/data': 0.01263594813644886, 'timing/model': 2.7810318637639284, 'learning_rate/qwen_vl_interface': 8.249516167701373e-06, 'learning_rate/action_model': 8.249516167701374e-05, 'epoch': 2.88}) +[2026-06-19 23:43:10,438][starVLA.training.train_starvla][INFO] - Step 1485, Loss: {'train/loss': 0.09763534367084503, 'train/grad_norm_pre_clip': 5.6483001708984375, 'timing/data': 0.013039303943514824, 'timing/model': 2.7793686725199223, 'learning_rate/qwen_vl_interface': 8.247154148215311e-06, 'learning_rate/action_model': 8.24715414821531e-05, 'epoch': 2.88}) +[2026-06-19 23:43:21,386][starVLA.training.train_starvla][INFO] - Step 1486, Loss: {'train/loss': 0.03593505918979645, 'train/grad_norm_pre_clip': 3.882720470428467, 'timing/data': 0.012852698564529419, 'timing/model': 2.9217119328677654, 'learning_rate/qwen_vl_interface': 8.24479089671239e-06, 'learning_rate/action_model': 8.24479089671239e-05, 'epoch': 2.88}) +[2026-06-19 23:43:32,236][starVLA.training.train_starvla][INFO] - Step 1487, Loss: {'train/loss': 0.12133734673261642, 'train/grad_norm_pre_clip': 4.239729881286621, 'timing/data': 0.012358441948890686, 'timing/model': 2.770317506045103, 'learning_rate/qwen_vl_interface': 8.242426414164056e-06, 'learning_rate/action_model': 8.242426414164056e-05, 'epoch': 2.88}) +[2026-06-19 23:43:43,134][starVLA.training.train_starvla][INFO] - Step 1488, Loss: {'train/loss': 0.16986845433712006, 'train/grad_norm_pre_clip': 8.829377174377441, 'timing/data': 0.012518182396888733, 'timing/model': 2.7713300734758377, 'learning_rate/qwen_vl_interface': 8.240060701542257e-06, 'learning_rate/action_model': 8.240060701542257e-05, 'epoch': 2.88}) +[2026-06-19 23:43:53,960][starVLA.training.train_starvla][INFO] - Step 1489, Loss: {'train/loss': 0.19267231225967407, 'train/grad_norm_pre_clip': 9.402512550354004, 'timing/data': 0.013199705630540848, 'timing/model': 2.7022559214383364, 'learning_rate/qwen_vl_interface': 8.237693759819447e-06, 'learning_rate/action_model': 8.237693759819448e-05, 'epoch': 2.89}) +[2026-06-19 23:44:05,126][starVLA.training.train_starvla][INFO] - Step 1490, Loss: {'train/loss': 0.33106839656829834, 'train/grad_norm_pre_clip': 8.253472328186035, 'timing/data': 0.012933043763041496, 'timing/model': 2.773221952840686, 'learning_rate/qwen_vl_interface': 8.235325589968587e-06, 'learning_rate/action_model': 8.235325589968587e-05, 'epoch': 2.89}) +[2026-06-19 23:44:16,029][starVLA.training.train_starvla][INFO] - Step 1491, Loss: {'train/loss': 0.06405520439147949, 'train/grad_norm_pre_clip': 4.5097455978393555, 'timing/data': 0.012373751029372215, 'timing/model': 2.774335991591215, 'learning_rate/qwen_vl_interface': 8.232956192963143e-06, 'learning_rate/action_model': 8.232956192963144e-05, 'epoch': 2.89}) +[2026-06-19 23:44:26,939][starVLA.training.train_starvla][INFO] - Step 1492, Loss: {'train/loss': 0.01819464936852455, 'train/grad_norm_pre_clip': 1.2847294807434082, 'timing/data': 0.014000486582517624, 'timing/model': 2.7736371904611588, 'learning_rate/qwen_vl_interface': 8.230585569777086e-06, 'learning_rate/action_model': 8.230585569777085e-05, 'epoch': 2.89}) +[2026-06-19 23:44:37,835][starVLA.training.train_starvla][INFO] - Step 1493, Loss: {'train/loss': 0.3686767816543579, 'train/grad_norm_pre_clip': 16.121559143066406, 'timing/data': 0.012591669335961342, 'timing/model': 2.773161517456174, 'learning_rate/qwen_vl_interface': 8.228213721384885e-06, 'learning_rate/action_model': 8.228213721384885e-05, 'epoch': 2.89}) +[2026-06-19 23:44:48,937][starVLA.training.train_starvla][INFO] - Step 1494, Loss: {'train/loss': 0.1817319542169571, 'train/grad_norm_pre_clip': 8.756819725036621, 'timing/data': 0.012669045478105545, 'timing/model': 2.773391069844365, 'learning_rate/qwen_vl_interface': 8.225840648761523e-06, 'learning_rate/action_model': 8.225840648761523e-05, 'epoch': 2.9}) +[2026-06-19 23:44:59,833][starVLA.training.train_starvla][INFO] - Step 1495, Loss: {'train/loss': 0.12813588976860046, 'train/grad_norm_pre_clip': 7.463328838348389, 'timing/data': 0.01222672127187252, 'timing/model': 2.7692232336848974, 'learning_rate/qwen_vl_interface': 8.223466352882476e-06, 'learning_rate/action_model': 8.223466352882475e-05, 'epoch': 2.9}) +[2026-06-19 23:45:10,735][starVLA.training.train_starvla][INFO] - Step 1496, Loss: {'train/loss': 0.222346231341362, 'train/grad_norm_pre_clip': 23.70722770690918, 'timing/data': 0.013200141489505768, 'timing/model': 2.781566431745887, 'learning_rate/qwen_vl_interface': 8.22109083472373e-06, 'learning_rate/action_model': 8.22109083472373e-05, 'epoch': 2.9}) +[2026-06-19 23:45:21,526][starVLA.training.train_starvla][INFO] - Step 1497, Loss: {'train/loss': 0.043324898928403854, 'train/grad_norm_pre_clip': 4.076433181762695, 'timing/data': 0.01262475736439228, 'timing/model': 2.7717323675751686, 'learning_rate/qwen_vl_interface': 8.218714095261769e-06, 'learning_rate/action_model': 8.218714095261768e-05, 'epoch': 2.9}) +[2026-06-19 23:45:32,533][starVLA.training.train_starvla][INFO] - Step 1498, Loss: {'train/loss': 0.3435248136520386, 'train/grad_norm_pre_clip': 9.39307689666748, 'timing/data': 0.012440595775842667, 'timing/model': 2.7729667518287897, 'learning_rate/qwen_vl_interface': 8.21633613547358e-06, 'learning_rate/action_model': 8.21633613547358e-05, 'epoch': 2.9}) +[2026-06-19 23:45:43,426][starVLA.training.train_starvla][INFO] - Step 1499, Loss: {'train/loss': 0.08074773102998734, 'train/grad_norm_pre_clip': 7.704392433166504, 'timing/data': 0.012888791039586067, 'timing/model': 2.770430486649275, 'learning_rate/qwen_vl_interface': 8.213956956336656e-06, 'learning_rate/action_model': 8.213956956336656e-05, 'epoch': 2.91}) +[2026-06-19 23:45:54,228][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1500 +[2026-06-19 23:53:25,510][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1500 +[2026-06-19 23:53:25,511][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 1500 +[2026-06-20 00:01:35,146][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 1500 +[2026-06-20 00:01:35,148][starVLA.training.train_starvla][INFO] - Step 1500, Loss: {'train/loss': 0.17557279765605927, 'train/grad_norm_pre_clip': 10.047070503234863, 'eval/loss': 0.6158952116966248, 'eval/latency_0/loss': 0.3299492597579956, 'eval/latency_1/loss': 0.07114863395690918, 'eval/latency_2/loss': 0.7669627070426941, 'eval/latency_3/loss': 0.7864687442779541, 'eval/latency_4/loss': 1.118989109992981, 'eval/flappy/loss': 0.6158952116966248, 'eval/flappy/latency_0/loss': 0.3299492597579956, 'eval/flappy/latency_1/loss': 0.07114863395690918, 'eval/flappy/latency_2/loss': 0.7669627070426941, 'eval/flappy/latency_3/loss': 0.7864687442779541, 'eval/flappy/latency_4/loss': 1.118989109992981, 'eval/flap_precision': 0.43956043956043955, 'eval/flap_recall': 0.5818181818181818, 'eval/flap_f1': 0.5007824726134585, 'eval/flap_rate_model': 0.10111111111111111, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.5007824726134585, 'eval/latency_0/flap_precision': 0.43956043956043955, 'eval/latency_0/flap_recall': 0.5818181818181818, 'eval/latency_0/flap_f1': 0.5007824726134585, 'eval/latency_0/flap_rate_model': 0.10111111111111111, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.5007824726134585, 'timing/data': 0.012529198080301285, 'timing/model': 2.7497948613017797, 'learning_rate/qwen_vl_interface': 8.211576558828985e-06, 'learning_rate/action_model': 8.211576558828985e-05, 'epoch': 2.91}) +[2026-06-20 00:01:35,149][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_1500_state +[2026-06-20 00:01:35,153][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-20 00:01:40,929][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_1500_state/model.safetensors +[2026-06-20 00:01:54,914][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_1500_state/optimizer.bin +[2026-06-20 00:01:54,915][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_1500_state/sampler.bin +[2026-06-20 00:01:54,915][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_1500_state/sampler_1.bin +[2026-06-20 00:01:54,916][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_1500_state/random_states_0.pkl +[2026-06-20 00:01:55,041][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1500 +[2026-06-20 00:01:55,041][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-20 00:01:55,056][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-20 00:02:05,914][starVLA.training.train_starvla][INFO] - Step 1501, Loss: {'train/loss': 0.18817228078842163, 'train/grad_norm_pre_clip': 7.545243740081787, 'timing/data': 0.013712216168642044, 'timing/model': 2.7648255340754986, 'learning_rate/qwen_vl_interface': 8.209194943929059e-06, 'learning_rate/action_model': 8.20919494392906e-05, 'epoch': 2.91}) +[2026-06-20 00:02:16,819][starVLA.training.train_starvla][INFO] - Step 1502, Loss: {'train/loss': 0.14517523348331451, 'train/grad_norm_pre_clip': 9.528417587280273, 'timing/data': 0.01266772672533989, 'timing/model': 2.765629217028618, 'learning_rate/qwen_vl_interface': 8.20681211261587e-06, 'learning_rate/action_model': 8.20681211261587e-05, 'epoch': 2.91}) +[2026-06-20 00:02:27,618][starVLA.training.train_starvla][INFO] - Step 1503, Loss: {'train/loss': 0.10347912460565567, 'train/grad_norm_pre_clip': 5.617247581481934, 'timing/data': 0.012634756043553352, 'timing/model': 2.7459528259932995, 'learning_rate/qwen_vl_interface': 8.204428065868911e-06, 'learning_rate/action_model': 8.204428065868911e-05, 'epoch': 2.91}) +[2026-06-20 00:02:38,528][starVLA.training.train_starvla][INFO] - Step 1504, Loss: {'train/loss': 0.05343504250049591, 'train/grad_norm_pre_clip': 5.0942230224609375, 'timing/data': 0.013543818145990372, 'timing/model': 2.7616496551781893, 'learning_rate/qwen_vl_interface': 8.202042804668173e-06, 'learning_rate/action_model': 8.202042804668172e-05, 'epoch': 2.91}) +[2026-06-20 00:02:49,428][starVLA.training.train_starvla][INFO] - Step 1505, Loss: {'train/loss': 0.15109561383724213, 'train/grad_norm_pre_clip': 4.982647895812988, 'timing/data': 0.012741556391119957, 'timing/model': 2.7718147318810225, 'learning_rate/qwen_vl_interface': 8.199656329994145e-06, 'learning_rate/action_model': 8.199656329994145e-05, 'epoch': 2.92}) +[2026-06-20 00:03:00,635][starVLA.training.train_starvla][INFO] - Step 1506, Loss: {'train/loss': 0.1544555276632309, 'train/grad_norm_pre_clip': 5.429764747619629, 'timing/data': 0.019722023978829384, 'timing/model': 2.770451372489333, 'learning_rate/qwen_vl_interface': 8.19726864282782e-06, 'learning_rate/action_model': 8.197268642827818e-05, 'epoch': 2.92}) +[2026-06-20 00:03:11,531][starVLA.training.train_starvla][INFO] - Step 1507, Loss: {'train/loss': 0.08425800502300262, 'train/grad_norm_pre_clip': 4.872903347015381, 'timing/data': 0.01811612956225872, 'timing/model': 2.7538796178996563, 'learning_rate/qwen_vl_interface': 8.194879744150682e-06, 'learning_rate/action_model': 8.194879744150681e-05, 'epoch': 2.92}) +[2026-06-20 00:03:22,432][starVLA.training.train_starvla][INFO] - Step 1508, Loss: {'train/loss': 0.28179478645324707, 'train/grad_norm_pre_clip': 24.36124610900879, 'timing/data': 0.013347366824746132, 'timing/model': 2.776561602950096, 'learning_rate/qwen_vl_interface': 8.192489634944718e-06, 'learning_rate/action_model': 8.192489634944718e-05, 'epoch': 2.92}) +[2026-06-20 00:03:33,544][starVLA.training.train_starvla][INFO] - Step 1509, Loss: {'train/loss': 0.2321835458278656, 'train/grad_norm_pre_clip': 9.519989013671875, 'timing/data': 0.013418847694993019, 'timing/model': 2.9959923308342695, 'learning_rate/qwen_vl_interface': 8.190098316192416e-06, 'learning_rate/action_model': 8.190098316192416e-05, 'epoch': 2.92}) +[2026-06-20 00:03:44,426][starVLA.training.train_starvla][INFO] - Step 1510, Loss: {'train/loss': 0.07249096035957336, 'train/grad_norm_pre_clip': 5.096002101898193, 'timing/data': 0.012422429397702217, 'timing/model': 2.7717049326747656, 'learning_rate/qwen_vl_interface': 8.18770578887675e-06, 'learning_rate/action_model': 8.18770578887675e-05, 'epoch': 2.93}) +[2026-06-20 00:03:55,314][starVLA.training.train_starvla][INFO] - Step 1511, Loss: {'train/loss': 0.20932459831237793, 'train/grad_norm_pre_clip': 9.326399803161621, 'timing/data': 0.01811296120285988, 'timing/model': 2.7474953588098288, 'learning_rate/qwen_vl_interface': 8.185312053981204e-06, 'learning_rate/action_model': 8.185312053981204e-05, 'epoch': 2.93}) +[2026-06-20 00:04:06,227][starVLA.training.train_starvla][INFO] - Step 1512, Loss: {'train/loss': 0.22117775678634644, 'train/grad_norm_pre_clip': 12.967777252197266, 'timing/data': 0.014000024646520615, 'timing/model': 2.772226043045521, 'learning_rate/qwen_vl_interface': 8.182917112489747e-06, 'learning_rate/action_model': 8.182917112489746e-05, 'epoch': 2.93}) +[2026-06-20 00:04:17,044][starVLA.training.train_starvla][INFO] - Step 1513, Loss: {'train/loss': 0.06469880044460297, 'train/grad_norm_pre_clip': 2.7017099857330322, 'timing/data': 0.012484954670071602, 'timing/model': 2.7175096347928047, 'learning_rate/qwen_vl_interface': 8.18052096538685e-06, 'learning_rate/action_model': 8.18052096538685e-05, 'epoch': 2.93}) +[2026-06-20 00:04:27,924][starVLA.training.train_starvla][INFO] - Step 1514, Loss: {'train/loss': 0.09022803604602814, 'train/grad_norm_pre_clip': 6.980447292327881, 'timing/data': 0.01252678781747818, 'timing/model': 2.76796024851501, 'learning_rate/qwen_vl_interface': 8.178123613657483e-06, 'learning_rate/action_model': 8.178123613657482e-05, 'epoch': 2.93}) +[2026-06-20 00:04:38,724][starVLA.training.train_starvla][INFO] - Step 1515, Loss: {'train/loss': 0.130357027053833, 'train/grad_norm_pre_clip': 9.700384140014648, 'timing/data': 0.012669717893004417, 'timing/model': 2.748522127047181, 'learning_rate/qwen_vl_interface': 8.175725058287102e-06, 'learning_rate/action_model': 8.175725058287102e-05, 'epoch': 2.94}) +[2026-06-20 00:04:49,629][starVLA.training.train_starvla][INFO] - Step 1516, Loss: {'train/loss': 0.1389709711074829, 'train/grad_norm_pre_clip': 9.6027193069458, 'timing/data': 0.012212961912155151, 'timing/model': 2.773659586906433, 'learning_rate/qwen_vl_interface': 8.173325300261664e-06, 'learning_rate/action_model': 8.173325300261663e-05, 'epoch': 2.94}) +[2026-06-20 00:05:00,424][starVLA.training.train_starvla][INFO] - Step 1517, Loss: {'train/loss': 0.25973570346832275, 'train/grad_norm_pre_clip': 14.262505531311035, 'timing/data': 0.012657107785344124, 'timing/model': 2.7621892727911472, 'learning_rate/qwen_vl_interface': 8.170924340567617e-06, 'learning_rate/action_model': 8.170924340567616e-05, 'epoch': 2.94}) +[2026-06-20 00:05:11,222][starVLA.training.train_starvla][INFO] - Step 1518, Loss: {'train/loss': 0.1821543425321579, 'train/grad_norm_pre_clip': 15.29802417755127, 'timing/data': 0.012677555903792381, 'timing/model': 2.749108435586095, 'learning_rate/qwen_vl_interface': 8.168522180191909e-06, 'learning_rate/action_model': 8.168522180191909e-05, 'epoch': 2.94}) +[2026-06-20 00:05:22,130][starVLA.training.train_starvla][INFO] - Step 1519, Loss: {'train/loss': 0.33655983209609985, 'train/grad_norm_pre_clip': 12.747614860534668, 'timing/data': 0.01272694580256939, 'timing/model': 2.780101403594017, 'learning_rate/qwen_vl_interface': 8.166118820121972e-06, 'learning_rate/action_model': 8.166118820121973e-05, 'epoch': 2.94}) +[2026-06-20 00:05:33,025][starVLA.training.train_starvla][INFO] - Step 1520, Loss: {'train/loss': 0.10810714215040207, 'train/grad_norm_pre_clip': 5.44248104095459, 'timing/data': 0.011787975206971169, 'timing/model': 2.768288256600499, 'learning_rate/qwen_vl_interface': 8.163714261345744e-06, 'learning_rate/action_model': 8.163714261345742e-05, 'epoch': 2.95}) +[2026-06-20 00:05:44,030][starVLA.training.train_starvla][INFO] - Step 1521, Loss: {'train/loss': 0.08212081342935562, 'train/grad_norm_pre_clip': 9.570003509521484, 'timing/data': 0.012442318722605705, 'timing/model': 2.7787906359881163, 'learning_rate/qwen_vl_interface': 8.16130850485164e-06, 'learning_rate/action_model': 8.161308504851639e-05, 'epoch': 2.95}) +[2026-06-20 00:05:54,732][starVLA.training.train_starvla][INFO] - Step 1522, Loss: {'train/loss': 0.14998455345630646, 'train/grad_norm_pre_clip': 9.475973129272461, 'timing/data': 0.011847052723169327, 'timing/model': 2.7365706879645586, 'learning_rate/qwen_vl_interface': 8.158901551628579e-06, 'learning_rate/action_model': 8.158901551628579e-05, 'epoch': 2.95}) +[2026-06-20 00:06:05,522][starVLA.training.train_starvla][INFO] - Step 1523, Loss: {'train/loss': 0.21611979603767395, 'train/grad_norm_pre_clip': 14.944380760192871, 'timing/data': 0.012244163081049919, 'timing/model': 2.7416718993335962, 'learning_rate/qwen_vl_interface': 8.156493402665971e-06, 'learning_rate/action_model': 8.15649340266597e-05, 'epoch': 2.95}) +[2026-06-20 00:06:16,434][starVLA.training.train_starvla][INFO] - Step 1524, Loss: {'train/loss': 0.3620782494544983, 'train/grad_norm_pre_clip': 16.677888870239258, 'timing/data': 0.013122810050845146, 'timing/model': 2.7633962277323008, 'learning_rate/qwen_vl_interface': 8.154084058953713e-06, 'learning_rate/action_model': 8.154084058953713e-05, 'epoch': 2.95}) +[2026-06-20 00:06:27,329][starVLA.training.train_starvla][INFO] - Step 1525, Loss: {'train/loss': 0.24372275173664093, 'train/grad_norm_pre_clip': 14.630154609680176, 'timing/data': 0.01172141544520855, 'timing/model': 2.769909054040909, 'learning_rate/qwen_vl_interface': 8.151673521482193e-06, 'learning_rate/action_model': 8.151673521482194e-05, 'epoch': 2.96}) +[2026-06-20 00:06:38,236][starVLA.training.train_starvla][INFO] - Step 1526, Loss: {'train/loss': 0.1281196027994156, 'train/grad_norm_pre_clip': 11.362210273742676, 'timing/data': 0.011929355561733246, 'timing/model': 2.7774323411285877, 'learning_rate/qwen_vl_interface': 8.149261791242297e-06, 'learning_rate/action_model': 8.149261791242297e-05, 'epoch': 2.96}) +[2026-06-20 00:06:49,133][starVLA.training.train_starvla][INFO] - Step 1527, Loss: {'train/loss': 0.1307302713394165, 'train/grad_norm_pre_clip': 20.49481773376465, 'timing/data': 0.01147046685218811, 'timing/model': 2.777706490829587, 'learning_rate/qwen_vl_interface': 8.146848869225391e-06, 'learning_rate/action_model': 8.146848869225391e-05, 'epoch': 2.96}) +[2026-06-20 00:07:00,033][starVLA.training.train_starvla][INFO] - Step 1528, Loss: {'train/loss': 0.01938731037080288, 'train/grad_norm_pre_clip': 8.054075241088867, 'timing/data': 0.01202431134879589, 'timing/model': 2.7667603939771652, 'learning_rate/qwen_vl_interface': 8.14443475642334e-06, 'learning_rate/action_model': 8.144434756423339e-05, 'epoch': 2.96}) +[2026-06-20 00:07:10,922][starVLA.training.train_starvla][INFO] - Step 1529, Loss: {'train/loss': 0.32063770294189453, 'train/grad_norm_pre_clip': 11.857958793640137, 'timing/data': 0.012347178533673286, 'timing/model': 2.759898593649268, 'learning_rate/qwen_vl_interface': 8.142019453828493e-06, 'learning_rate/action_model': 8.142019453828493e-05, 'epoch': 2.96}) +[2026-06-20 00:07:21,830][starVLA.training.train_starvla][INFO] - Step 1530, Loss: {'train/loss': 0.08552972972393036, 'train/grad_norm_pre_clip': 5.913455963134766, 'timing/data': 0.01179080456495285, 'timing/model': 2.772245056927204, 'learning_rate/qwen_vl_interface': 8.139602962433688e-06, 'learning_rate/action_model': 8.139602962433688e-05, 'epoch': 2.97}) +[2026-06-20 00:07:32,728][starVLA.training.train_starvla][INFO] - Step 1531, Loss: {'train/loss': 0.14796404540538788, 'train/grad_norm_pre_clip': 6.459455966949463, 'timing/data': 0.011995924636721611, 'timing/model': 2.7610753905028105, 'learning_rate/qwen_vl_interface': 8.137185283232255e-06, 'learning_rate/action_model': 8.137185283232256e-05, 'epoch': 2.97}) +[2026-06-20 00:07:43,732][starVLA.training.train_starvla][INFO] - Step 1532, Loss: {'train/loss': 0.06396894156932831, 'train/grad_norm_pre_clip': 6.001483917236328, 'timing/data': 0.011854706332087517, 'timing/model': 2.7824228331446648, 'learning_rate/qwen_vl_interface': 8.134766417218014e-06, 'learning_rate/action_model': 8.134766417218012e-05, 'epoch': 2.97}) +[2026-06-20 00:07:54,630][starVLA.training.train_starvla][INFO] - Step 1533, Loss: {'train/loss': 0.04274323210120201, 'train/grad_norm_pre_clip': 3.435554265975952, 'timing/data': 0.01160808838903904, 'timing/model': 2.768083281815052, 'learning_rate/qwen_vl_interface': 8.132346365385262e-06, 'learning_rate/action_model': 8.132346365385261e-05, 'epoch': 2.97}) +[2026-06-20 00:08:05,534][starVLA.training.train_starvla][INFO] - Step 1534, Loss: {'train/loss': 0.043578553944826126, 'train/grad_norm_pre_clip': 8.618144989013672, 'timing/data': 0.012772735208272934, 'timing/model': 2.7653911132365465, 'learning_rate/qwen_vl_interface': 8.129925128728796e-06, 'learning_rate/action_model': 8.129925128728796e-05, 'epoch': 2.97}) +[2026-06-20 00:08:16,434][starVLA.training.train_starvla][INFO] - Step 1535, Loss: {'train/loss': 0.08948424458503723, 'train/grad_norm_pre_clip': 6.577389717102051, 'timing/data': 0.012719525024294853, 'timing/model': 2.771313039585948, 'learning_rate/qwen_vl_interface': 8.127502708243895e-06, 'learning_rate/action_model': 8.127502708243895e-05, 'epoch': 2.98}) +[2026-06-20 00:08:27,333][starVLA.training.train_starvla][INFO] - Step 1536, Loss: {'train/loss': 0.10548237711191177, 'train/grad_norm_pre_clip': 7.873223304748535, 'timing/data': 0.012752695009112358, 'timing/model': 2.771128538995981, 'learning_rate/qwen_vl_interface': 8.125079104926324e-06, 'learning_rate/action_model': 8.125079104926323e-05, 'epoch': 2.98}) +[2026-06-20 00:08:38,234][starVLA.training.train_starvla][INFO] - Step 1537, Loss: {'train/loss': 0.13093999028205872, 'train/grad_norm_pre_clip': 8.228355407714844, 'timing/data': 0.012523038312792778, 'timing/model': 2.76616295799613, 'learning_rate/qwen_vl_interface': 8.122654319772332e-06, 'learning_rate/action_model': 8.122654319772332e-05, 'epoch': 2.98}) +[2026-06-20 00:08:49,130][starVLA.training.train_starvla][INFO] - Step 1538, Loss: {'train/loss': 0.04752953350543976, 'train/grad_norm_pre_clip': 5.180408477783203, 'timing/data': 0.012256298214197159, 'timing/model': 2.7660071663558483, 'learning_rate/qwen_vl_interface': 8.120228353778661e-06, 'learning_rate/action_model': 8.12022835377866e-05, 'epoch': 2.98}) +[2026-06-20 00:09:00,032][starVLA.training.train_starvla][INFO] - Step 1539, Loss: {'train/loss': 0.12155651301145554, 'train/grad_norm_pre_clip': 9.148992538452148, 'timing/data': 0.012325489893555641, 'timing/model': 2.7724366933107376, 'learning_rate/qwen_vl_interface': 8.117801207942528e-06, 'learning_rate/action_model': 8.117801207942528e-05, 'epoch': 2.98}) +[2026-06-20 00:09:11,127][starVLA.training.train_starvla][INFO] - Step 1540, Loss: {'train/loss': 0.21858197450637817, 'train/grad_norm_pre_clip': 8.553328514099121, 'timing/data': 0.012700976803898811, 'timing/model': 2.7638652715831995, 'learning_rate/qwen_vl_interface': 8.115372883261648e-06, 'learning_rate/action_model': 8.115372883261647e-05, 'epoch': 2.98}) +[2026-06-20 00:09:22,036][starVLA.training.train_starvla][INFO] - Step 1541, Loss: {'train/loss': 0.09457176178693771, 'train/grad_norm_pre_clip': 5.9048309326171875, 'timing/data': 0.013310050591826439, 'timing/model': 2.7760524693876505, 'learning_rate/qwen_vl_interface': 8.11294338073421e-06, 'learning_rate/action_model': 8.11294338073421e-05, 'epoch': 2.99}) +[2026-06-20 00:09:32,924][starVLA.training.train_starvla][INFO] - Step 1542, Loss: {'train/loss': 0.14741474390029907, 'train/grad_norm_pre_clip': 9.91748332977295, 'timing/data': 0.012444177642464638, 'timing/model': 2.7731389086693525, 'learning_rate/qwen_vl_interface': 8.110512701358892e-06, 'learning_rate/action_model': 8.110512701358891e-05, 'epoch': 2.99}) +[2026-06-20 00:09:43,843][starVLA.training.train_starvla][INFO] - Step 1543, Loss: {'train/loss': 0.3018459975719452, 'train/grad_norm_pre_clip': 12.213857650756836, 'timing/data': 0.012984374538064003, 'timing/model': 2.776221811771393, 'learning_rate/qwen_vl_interface': 8.108080846134852e-06, 'learning_rate/action_model': 8.108080846134852e-05, 'epoch': 2.99}) +[2026-06-20 00:09:54,733][starVLA.training.train_starvla][INFO] - Step 1544, Loss: {'train/loss': 0.1300380975008011, 'train/grad_norm_pre_clip': 6.808666229248047, 'timing/data': 0.012937463819980621, 'timing/model': 2.7717049457132816, 'learning_rate/qwen_vl_interface': 8.105647816061734e-06, 'learning_rate/action_model': 8.105647816061735e-05, 'epoch': 2.99}) +[2026-06-20 00:10:05,634][starVLA.training.train_starvla][INFO] - Step 1545, Loss: {'train/loss': 0.19666464626789093, 'train/grad_norm_pre_clip': 10.102404594421387, 'timing/data': 0.012750957161188126, 'timing/model': 2.7710479889065027, 'learning_rate/qwen_vl_interface': 8.103213612139667e-06, 'learning_rate/action_model': 8.103213612139667e-05, 'epoch': 2.99}) +[2026-06-20 00:10:16,526][starVLA.training.train_starvla][INFO] - Step 1546, Loss: {'train/loss': 0.39797070622444153, 'train/grad_norm_pre_clip': 11.980997085571289, 'timing/data': 0.012756215408444405, 'timing/model': 2.7621735334396362, 'learning_rate/qwen_vl_interface': 8.100778235369262e-06, 'learning_rate/action_model': 8.10077823536926e-05, 'epoch': 3.0}) +[2026-06-20 00:10:27,433][starVLA.training.train_starvla][INFO] - Step 1547, Loss: {'train/loss': 0.16480572521686554, 'train/grad_norm_pre_clip': 10.75795841217041, 'timing/data': 0.012689417228102684, 'timing/model': 2.7538537960499525, 'learning_rate/qwen_vl_interface': 8.098341686751605e-06, 'learning_rate/action_model': 8.098341686751605e-05, 'epoch': 3.0}) +[2026-06-20 00:10:38,434][starVLA.training.train_starvla][INFO] - Step 1548, Loss: {'train/loss': 0.32873424887657166, 'train/grad_norm_pre_clip': 18.305143356323242, 'timing/data': 0.012827044352889061, 'timing/model': 2.7703753393143415, 'learning_rate/qwen_vl_interface': 8.095903967288273e-06, 'learning_rate/action_model': 8.095903967288272e-05, 'epoch': 3.0}) +[2026-06-20 00:10:49,334][starVLA.training.train_starvla][INFO] - Step 1549, Loss: {'train/loss': 0.226674884557724, 'train/grad_norm_pre_clip': 12.941751480102539, 'timing/data': 0.012574266642332077, 'timing/model': 2.7674783878028393, 'learning_rate/qwen_vl_interface': 8.093465077981318e-06, 'learning_rate/action_model': 8.093465077981318e-05, 'epoch': 3.0}) +[2026-06-20 00:11:00,236][starVLA.training.train_starvla][INFO] - Step 1550, Loss: {'train/loss': 0.14409996569156647, 'train/grad_norm_pre_clip': 5.4363226890563965, 'timing/data': 0.012358369305729866, 'timing/model': 2.772750247269869, 'learning_rate/qwen_vl_interface': 8.091025019833276e-06, 'learning_rate/action_model': 8.091025019833276e-05, 'epoch': 3.0}) +[2026-06-20 00:11:11,129][starVLA.training.train_starvla][INFO] - Step 1551, Loss: {'train/loss': 0.09093114733695984, 'train/grad_norm_pre_clip': 6.113832950592041, 'timing/data': 0.012181572616100311, 'timing/model': 2.775249542668462, 'learning_rate/qwen_vl_interface': 8.088583793847165e-06, 'learning_rate/action_model': 8.088583793847164e-05, 'epoch': 3.01}) +[2026-06-20 00:11:22,036][starVLA.training.train_starvla][INFO] - Step 1552, Loss: {'train/loss': 0.2594541907310486, 'train/grad_norm_pre_clip': 8.333245277404785, 'timing/data': 0.012625608593225479, 'timing/model': 2.768317559733987, 'learning_rate/qwen_vl_interface': 8.086141401026478e-06, 'learning_rate/action_model': 8.086141401026478e-05, 'epoch': 3.01}) +[2026-06-20 00:11:32,935][starVLA.training.train_starvla][INFO] - Step 1553, Loss: {'train/loss': 0.238800510764122, 'train/grad_norm_pre_clip': 8.507122993469238, 'timing/data': 0.012473257258534431, 'timing/model': 2.771077850833535, 'learning_rate/qwen_vl_interface': 8.083697842375193e-06, 'learning_rate/action_model': 8.083697842375193e-05, 'epoch': 3.01}) +[2026-06-20 00:11:43,835][starVLA.training.train_starvla][INFO] - Step 1554, Loss: {'train/loss': 0.18016758561134338, 'train/grad_norm_pre_clip': 8.136149406433105, 'timing/data': 0.01251181960105896, 'timing/model': 2.768751110881567, 'learning_rate/qwen_vl_interface': 8.081253118897762e-06, 'learning_rate/action_model': 8.081253118897761e-05, 'epoch': 3.01}) +[2026-06-20 00:11:54,960][starVLA.training.train_starvla][INFO] - Step 1555, Loss: {'train/loss': 0.16357144713401794, 'train/grad_norm_pre_clip': 7.521374702453613, 'timing/data': 0.01264747604727745, 'timing/model': 2.99325736425817, 'learning_rate/qwen_vl_interface': 8.07880723159912e-06, 'learning_rate/action_model': 8.07880723159912e-05, 'epoch': 3.01}) +[2026-06-20 00:12:05,832][starVLA.training.train_starvla][INFO] - Step 1556, Loss: {'train/loss': 0.11793187260627747, 'train/grad_norm_pre_clip': 3.957188129425049, 'timing/data': 0.012228172272443771, 'timing/model': 2.7876976560801268, 'learning_rate/qwen_vl_interface': 8.076360181484676e-06, 'learning_rate/action_model': 8.076360181484676e-05, 'epoch': 3.02}) +[2026-06-20 00:12:16,722][starVLA.training.train_starvla][INFO] - Step 1557, Loss: {'train/loss': 0.09560885280370712, 'train/grad_norm_pre_clip': 3.4751644134521484, 'timing/data': 0.012491628527641296, 'timing/model': 2.7598756477236748, 'learning_rate/qwen_vl_interface': 8.073911969560326e-06, 'learning_rate/action_model': 8.073911969560326e-05, 'epoch': 3.02}) +[2026-06-20 00:12:27,630][starVLA.training.train_starvla][INFO] - Step 1558, Loss: {'train/loss': 0.16234581172466278, 'train/grad_norm_pre_clip': 3.936305522918701, 'timing/data': 0.012423619627952576, 'timing/model': 2.770045191049576, 'learning_rate/qwen_vl_interface': 8.071462596832432e-06, 'learning_rate/action_model': 8.071462596832432e-05, 'epoch': 3.02}) +[2026-06-20 00:12:38,729][starVLA.training.train_starvla][INFO] - Step 1559, Loss: {'train/loss': 0.21380968391895294, 'train/grad_norm_pre_clip': 10.26782512664795, 'timing/data': 0.011981803923845291, 'timing/model': 2.751138797029853, 'learning_rate/qwen_vl_interface': 8.06901206430784e-06, 'learning_rate/action_model': 8.06901206430784e-05, 'epoch': 3.02}) +[2026-06-20 00:12:49,623][starVLA.training.train_starvla][INFO] - Step 1560, Loss: {'train/loss': 0.5311309099197388, 'train/grad_norm_pre_clip': 10.591680526733398, 'timing/data': 0.012273980304598808, 'timing/model': 2.7683570347726345, 'learning_rate/qwen_vl_interface': 8.066560372993873e-06, 'learning_rate/action_model': 8.066560372993874e-05, 'epoch': 3.02}) +[2026-06-20 00:13:00,424][starVLA.training.train_starvla][INFO] - Step 1561, Loss: {'train/loss': 0.2168906182050705, 'train/grad_norm_pre_clip': 7.836094379425049, 'timing/data': 0.012265700846910477, 'timing/model': 2.756728971377015, 'learning_rate/qwen_vl_interface': 8.064107523898327e-06, 'learning_rate/action_model': 8.064107523898326e-05, 'epoch': 3.03}) +[2026-06-20 00:13:11,138][starVLA.training.train_starvla][INFO] - Step 1562, Loss: {'train/loss': 0.12013892084360123, 'train/grad_norm_pre_clip': 5.910186767578125, 'timing/data': 0.01250009797513485, 'timing/model': 2.7167370673269033, 'learning_rate/qwen_vl_interface': 8.061653518029476e-06, 'learning_rate/action_model': 8.061653518029476e-05, 'epoch': 3.03}) +[2026-06-20 00:13:22,056][starVLA.training.train_starvla][INFO] - Step 1563, Loss: {'train/loss': 0.15241238474845886, 'train/grad_norm_pre_clip': 5.372791290283203, 'timing/data': 0.012463709339499474, 'timing/model': 2.6487367805093527, 'learning_rate/qwen_vl_interface': 8.059198356396069e-06, 'learning_rate/action_model': 8.059198356396068e-05, 'epoch': 3.03}) +[2026-06-20 00:13:32,832][starVLA.training.train_starvla][INFO] - Step 1564, Loss: {'train/loss': 0.3137371838092804, 'train/grad_norm_pre_clip': 9.340458869934082, 'timing/data': 0.012315720319747925, 'timing/model': 2.772265223786235, 'learning_rate/qwen_vl_interface': 8.05674204000733e-06, 'learning_rate/action_model': 8.05674204000733e-05, 'epoch': 3.03}) +[2026-06-20 00:13:43,730][starVLA.training.train_starvla][INFO] - Step 1565, Loss: {'train/loss': 0.17785392701625824, 'train/grad_norm_pre_clip': 5.088750839233398, 'timing/data': 0.012346860021352768, 'timing/model': 2.794573375955224, 'learning_rate/qwen_vl_interface': 8.05428456987296e-06, 'learning_rate/action_model': 8.05428456987296e-05, 'epoch': 3.03}) +[2026-06-20 00:13:54,862][starVLA.training.train_starvla][INFO] - Step 1566, Loss: {'train/loss': 0.1383511871099472, 'train/grad_norm_pre_clip': 5.329689025878906, 'timing/data': 0.012598037719726562, 'timing/model': 3.0052702128887177, 'learning_rate/qwen_vl_interface': 8.05182594700313e-06, 'learning_rate/action_model': 8.05182594700313e-05, 'epoch': 3.04}) +[2026-06-20 00:14:05,554][starVLA.training.train_starvla][INFO] - Step 1567, Loss: {'train/loss': 0.0736556127667427, 'train/grad_norm_pre_clip': 2.6732091903686523, 'timing/data': 0.012515045702457428, 'timing/model': 2.6875117253512144, 'learning_rate/qwen_vl_interface': 8.049366172408484e-06, 'learning_rate/action_model': 8.049366172408485e-05, 'epoch': 3.04}) +[2026-06-20 00:14:16,422][starVLA.training.train_starvla][INFO] - Step 1568, Loss: {'train/loss': 0.21607251465320587, 'train/grad_norm_pre_clip': 12.761844635009766, 'timing/data': 0.01245848648250103, 'timing/model': 2.76187826693058, 'learning_rate/qwen_vl_interface': 8.04690524710015e-06, 'learning_rate/action_model': 8.046905247100151e-05, 'epoch': 3.04}) +[2026-06-20 00:14:27,332][starVLA.training.train_starvla][INFO] - Step 1569, Loss: {'train/loss': 0.10736259818077087, 'train/grad_norm_pre_clip': 6.178288459777832, 'timing/data': 0.012599356472492218, 'timing/model': 2.766768442466855, 'learning_rate/qwen_vl_interface': 8.044443172089715e-06, 'learning_rate/action_model': 8.044443172089714e-05, 'epoch': 3.04}) +[2026-06-20 00:14:38,232][starVLA.training.train_starvla][INFO] - Step 1570, Loss: {'train/loss': 0.19381237030029297, 'train/grad_norm_pre_clip': 6.3864970207214355, 'timing/data': 0.01269555650651455, 'timing/model': 2.7679596189409494, 'learning_rate/qwen_vl_interface': 8.041979948389248e-06, 'learning_rate/action_model': 8.041979948389248e-05, 'epoch': 3.04}) +[2026-06-20 00:14:49,130][starVLA.training.train_starvla][INFO] - Step 1571, Loss: {'train/loss': 0.22977198660373688, 'train/grad_norm_pre_clip': 8.319549560546875, 'timing/data': 0.012688955292105675, 'timing/model': 2.7764029670506716, 'learning_rate/qwen_vl_interface': 8.039515577011287e-06, 'learning_rate/action_model': 8.039515577011288e-05, 'epoch': 3.04}) +[2026-06-20 00:15:00,032][starVLA.training.train_starvla][INFO] - Step 1572, Loss: {'train/loss': 0.38779255747795105, 'train/grad_norm_pre_clip': 8.94431209564209, 'timing/data': 0.011602578684687614, 'timing/model': 2.767856627702713, 'learning_rate/qwen_vl_interface': 8.037050058968841e-06, 'learning_rate/action_model': 8.03705005896884e-05, 'epoch': 3.05}) +[2026-06-20 00:15:10,840][starVLA.training.train_starvla][INFO] - Step 1573, Loss: {'train/loss': 0.19153985381126404, 'train/grad_norm_pre_clip': 8.161651611328125, 'timing/data': 0.012537015601992607, 'timing/model': 2.7826510667800903, 'learning_rate/qwen_vl_interface': 8.034583395275392e-06, 'learning_rate/action_model': 8.034583395275392e-05, 'epoch': 3.05}) +[2026-06-20 00:15:21,733][starVLA.training.train_starvla][INFO] - Step 1574, Loss: {'train/loss': 0.16520148515701294, 'train/grad_norm_pre_clip': 5.168887615203857, 'timing/data': 0.012591637670993805, 'timing/model': 2.770131889730692, 'learning_rate/qwen_vl_interface': 8.032115586944893e-06, 'learning_rate/action_model': 8.032115586944892e-05, 'epoch': 3.05}) +[2026-06-20 00:15:32,622][starVLA.training.train_starvla][INFO] - Step 1575, Loss: {'train/loss': 0.1029246523976326, 'train/grad_norm_pre_clip': 4.1756672859191895, 'timing/data': 0.01242939755320549, 'timing/model': 2.757812649011612, 'learning_rate/qwen_vl_interface': 8.029646634991762e-06, 'learning_rate/action_model': 8.029646634991762e-05, 'epoch': 3.05}) +[2026-06-20 00:15:43,530][starVLA.training.train_starvla][INFO] - Step 1576, Loss: {'train/loss': 0.08278274536132812, 'train/grad_norm_pre_clip': 3.7209956645965576, 'timing/data': 0.01265941746532917, 'timing/model': 2.767346015200019, 'learning_rate/qwen_vl_interface': 8.0271765404309e-06, 'learning_rate/action_model': 8.0271765404309e-05, 'epoch': 3.05}) +[2026-06-20 00:15:54,283][starVLA.training.train_starvla][INFO] - Step 1577, Loss: {'train/loss': 0.11326855421066284, 'train/grad_norm_pre_clip': 6.109038829803467, 'timing/data': 0.012438749894499779, 'timing/model': 2.7173672523349524, 'learning_rate/qwen_vl_interface': 8.024705304277664e-06, 'learning_rate/action_model': 8.024705304277664e-05, 'epoch': 3.06}) +[2026-06-20 00:16:05,130][starVLA.training.train_starvla][INFO] - Step 1578, Loss: {'train/loss': 0.19388115406036377, 'train/grad_norm_pre_clip': 3.870013952255249, 'timing/data': 0.012558218091726303, 'timing/model': 2.7768636848777533, 'learning_rate/qwen_vl_interface': 8.022232927547888e-06, 'learning_rate/action_model': 8.022232927547886e-05, 'epoch': 3.06}) +[2026-06-20 00:16:16,029][starVLA.training.train_starvla][INFO] - Step 1579, Loss: {'train/loss': 0.2694814205169678, 'train/grad_norm_pre_clip': 11.712347030639648, 'timing/data': 0.012370636686682701, 'timing/model': 2.775134241208434, 'learning_rate/qwen_vl_interface': 8.01975941125787e-06, 'learning_rate/action_model': 8.01975941125787e-05, 'epoch': 3.06}) +[2026-06-20 00:16:26,695][starVLA.training.train_starvla][INFO] - Step 1580, Loss: {'train/loss': 0.11915755271911621, 'train/grad_norm_pre_clip': 4.9690632820129395, 'timing/data': 0.012915344908833504, 'timing/model': 2.74086357280612, 'learning_rate/qwen_vl_interface': 8.017284756424381e-06, 'learning_rate/action_model': 8.017284756424381e-05, 'epoch': 3.06}) +[2026-06-20 00:16:37,301][starVLA.training.train_starvla][INFO] - Step 1581, Loss: {'train/loss': 0.177574023604393, 'train/grad_norm_pre_clip': 5.395419597625732, 'timing/data': 0.012571388855576515, 'timing/model': 2.7467325162142515, 'learning_rate/qwen_vl_interface': 8.01480896406466e-06, 'learning_rate/action_model': 8.01480896406466e-05, 'epoch': 3.06}) +[2026-06-20 00:16:48,131][starVLA.training.train_starvla][INFO] - Step 1582, Loss: {'train/loss': 0.134691521525383, 'train/grad_norm_pre_clip': 6.581813335418701, 'timing/data': 0.012119323015213013, 'timing/model': 2.7671530805528164, 'learning_rate/qwen_vl_interface': 8.01233203519641e-06, 'learning_rate/action_model': 8.012332035196409e-05, 'epoch': 3.07}) +[2026-06-20 00:16:59,034][starVLA.training.train_starvla][INFO] - Step 1583, Loss: {'train/loss': 0.27990180253982544, 'train/grad_norm_pre_clip': 15.23835277557373, 'timing/data': 0.011950481683015823, 'timing/model': 2.7678301725536585, 'learning_rate/qwen_vl_interface': 8.009853970837801e-06, 'learning_rate/action_model': 8.009853970837801e-05, 'epoch': 3.07}) +[2026-06-20 00:17:09,937][starVLA.training.train_starvla][INFO] - Step 1584, Loss: {'train/loss': 0.23299458622932434, 'train/grad_norm_pre_clip': 7.803154468536377, 'timing/data': 0.01252223551273346, 'timing/model': 2.7766508869826794, 'learning_rate/qwen_vl_interface': 8.007374772007474e-06, 'learning_rate/action_model': 8.007374772007474e-05, 'epoch': 3.07}) +[2026-06-20 00:17:20,826][starVLA.training.train_starvla][INFO] - Step 1585, Loss: {'train/loss': 0.2369326651096344, 'train/grad_norm_pre_clip': 14.78237533569336, 'timing/data': 0.01245974749326706, 'timing/model': 2.765360014513135, 'learning_rate/qwen_vl_interface': 8.004894439724533e-06, 'learning_rate/action_model': 8.004894439724533e-05, 'epoch': 3.07}) +[2026-06-20 00:17:31,920][starVLA.training.train_starvla][INFO] - Step 1586, Loss: {'train/loss': 0.17547522485256195, 'train/grad_norm_pre_clip': 7.07669734954834, 'timing/data': 0.012466054409742355, 'timing/model': 2.7617146242409945, 'learning_rate/qwen_vl_interface': 8.002412975008552e-06, 'learning_rate/action_model': 8.002412975008552e-05, 'epoch': 3.07}) +[2026-06-20 00:17:42,832][starVLA.training.train_starvla][INFO] - Step 1587, Loss: {'train/loss': 0.2827952206134796, 'train/grad_norm_pre_clip': 9.032753944396973, 'timing/data': 0.012456918135285378, 'timing/model': 2.770562356337905, 'learning_rate/qwen_vl_interface': 7.999930378879562e-06, 'learning_rate/action_model': 7.999930378879561e-05, 'epoch': 3.08}) +[2026-06-20 00:17:53,734][starVLA.training.train_starvla][INFO] - Step 1588, Loss: {'train/loss': 0.12422157824039459, 'train/grad_norm_pre_clip': 7.509862899780273, 'timing/data': 0.012397440150380135, 'timing/model': 2.7723468467593193, 'learning_rate/qwen_vl_interface': 7.997446652358067e-06, 'learning_rate/action_model': 7.997446652358068e-05, 'epoch': 3.08}) +[2026-06-20 00:18:04,621][starVLA.training.train_starvla][INFO] - Step 1589, Loss: {'train/loss': 0.3460017144680023, 'train/grad_norm_pre_clip': 11.66649055480957, 'timing/data': 0.012544376775622368, 'timing/model': 2.7624103110283613, 'learning_rate/qwen_vl_interface': 7.994961796465035e-06, 'learning_rate/action_model': 7.994961796465035e-05, 'epoch': 3.08}) +[2026-06-20 00:18:15,533][starVLA.training.train_starvla][INFO] - Step 1590, Loss: {'train/loss': 0.12285570800304413, 'train/grad_norm_pre_clip': 8.738100051879883, 'timing/data': 0.012569865211844444, 'timing/model': 2.772186268121004, 'learning_rate/qwen_vl_interface': 7.992475812221894e-06, 'learning_rate/action_model': 7.992475812221894e-05, 'epoch': 3.08}) +[2026-06-20 00:18:26,432][starVLA.training.train_starvla][INFO] - Step 1591, Loss: {'train/loss': 0.26206260919570923, 'train/grad_norm_pre_clip': 9.821499824523926, 'timing/data': 0.012484429404139519, 'timing/model': 2.7826434150338173, 'learning_rate/qwen_vl_interface': 7.989988700650538e-06, 'learning_rate/action_model': 7.989988700650538e-05, 'epoch': 3.08}) +[2026-06-20 00:18:37,334][starVLA.training.train_starvla][INFO] - Step 1592, Loss: {'train/loss': 0.20522719621658325, 'train/grad_norm_pre_clip': 12.192110061645508, 'timing/data': 0.012427156791090965, 'timing/model': 2.780314415693283, 'learning_rate/qwen_vl_interface': 7.987500462773326e-06, 'learning_rate/action_model': 7.987500462773326e-05, 'epoch': 3.09}) +[2026-06-20 00:18:48,378][starVLA.training.train_starvla][INFO] - Step 1593, Loss: {'train/loss': 0.19883808493614197, 'train/grad_norm_pre_clip': 7.304538726806641, 'timing/data': 0.012495234608650208, 'timing/model': 2.919620832428336, 'learning_rate/qwen_vl_interface': 7.985011099613078e-06, 'learning_rate/action_model': 7.985011099613077e-05, 'epoch': 3.09}) +[2026-06-20 00:18:59,233][starVLA.training.train_starvla][INFO] - Step 1594, Loss: {'train/loss': 0.037080395966768265, 'train/grad_norm_pre_clip': 3.937055826187134, 'timing/data': 0.01234503835439682, 'timing/model': 2.7823844235390425, 'learning_rate/qwen_vl_interface': 7.98252061219308e-06, 'learning_rate/action_model': 7.982520612193078e-05, 'epoch': 3.09}) +[2026-06-20 00:19:10,133][starVLA.training.train_starvla][INFO] - Step 1595, Loss: {'train/loss': 0.19059067964553833, 'train/grad_norm_pre_clip': 6.446018218994141, 'timing/data': 0.012916464358568192, 'timing/model': 2.7670911960303783, 'learning_rate/qwen_vl_interface': 7.980029001537073e-06, 'learning_rate/action_model': 7.980029001537072e-05, 'epoch': 3.09}) +[2026-06-20 00:19:21,033][starVLA.training.train_starvla][INFO] - Step 1596, Loss: {'train/loss': 0.1310233771800995, 'train/grad_norm_pre_clip': 6.621772766113281, 'timing/data': 0.012446386739611626, 'timing/model': 2.775255749002099, 'learning_rate/qwen_vl_interface': 7.977536268669264e-06, 'learning_rate/action_model': 7.977536268669264e-05, 'epoch': 3.09}) +[2026-06-20 00:19:32,138][starVLA.training.train_starvla][INFO] - Step 1597, Loss: {'train/loss': 0.09204135090112686, 'train/grad_norm_pre_clip': 4.4184250831604, 'timing/data': 0.012526016682386398, 'timing/model': 2.7501829527318478, 'learning_rate/qwen_vl_interface': 7.975042414614324e-06, 'learning_rate/action_model': 7.975042414614324e-05, 'epoch': 3.1}) +[2026-06-20 00:19:43,025][starVLA.training.train_starvla][INFO] - Step 1598, Loss: {'train/loss': 0.1619686633348465, 'train/grad_norm_pre_clip': 7.4547438621521, 'timing/data': 0.012389488518238068, 'timing/model': 2.767670525237918, 'learning_rate/qwen_vl_interface': 7.972547440397382e-06, 'learning_rate/action_model': 7.972547440397383e-05, 'epoch': 3.1}) +[2026-06-20 00:19:53,929][starVLA.training.train_starvla][INFO] - Step 1599, Loss: {'train/loss': 0.24414995312690735, 'train/grad_norm_pre_clip': 12.105621337890625, 'timing/data': 0.012396767735481262, 'timing/model': 2.768016252666712, 'learning_rate/qwen_vl_interface': 7.970051347044026e-06, 'learning_rate/action_model': 7.970051347044025e-05, 'epoch': 3.1}) +[2026-06-20 00:20:04,949][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1600 +[2026-06-20 00:27:35,557][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1600 +[2026-06-20 00:27:35,558][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 1600 +[2026-06-20 00:35:43,723][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 1600 +[2026-06-20 00:35:43,725][starVLA.training.train_starvla][INFO] - Step 1600, Loss: {'train/loss': 0.06723684817552567, 'train/grad_norm_pre_clip': 5.182672500610352, 'eval/loss': 0.5442596673965454, 'eval/latency_0/loss': 0.27376726269721985, 'eval/latency_1/loss': 0.07464274764060974, 'eval/latency_2/loss': 0.6618261337280273, 'eval/latency_3/loss': 0.714319109916687, 'eval/latency_4/loss': 0.9927335381507874, 'eval/flappy/loss': 0.5442596673965454, 'eval/flappy/latency_0/loss': 0.27376726269721985, 'eval/flappy/latency_1/loss': 0.07464274764060974, 'eval/flappy/latency_2/loss': 0.6618261337280273, 'eval/flappy/latency_3/loss': 0.714319109916687, 'eval/flappy/latency_4/loss': 0.9927335381507874, 'eval/flap_precision': 0.4713375796178344, 'eval/flap_recall': 0.5381818181818182, 'eval/flap_f1': 0.502546689303905, 'eval/flap_rate_model': 0.08722222222222223, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.502546689303905, 'eval/latency_0/flap_precision': 0.4713375796178344, 'eval/latency_0/flap_recall': 0.5381818181818182, 'eval/latency_0/flap_f1': 0.502546689303905, 'eval/latency_0/flap_rate_model': 0.08722222222222223, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.502546689303905, 'timing/data': 0.036785850301384926, 'timing/model': 2.7322384901344776, 'learning_rate/qwen_vl_interface': 7.967554135580307e-06, 'learning_rate/action_model': 7.967554135580307e-05, 'epoch': 3.1}) +[2026-06-20 00:35:43,725][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_1600_state +[2026-06-20 00:35:43,748][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-20 00:35:49,498][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_1600_state/model.safetensors +[2026-06-20 00:36:03,556][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_1600_state/optimizer.bin +[2026-06-20 00:36:03,557][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_1600_state/sampler.bin +[2026-06-20 00:36:03,557][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_1600_state/sampler_1.bin +[2026-06-20 00:36:03,558][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_1600_state/random_states_0.pkl +[2026-06-20 00:36:03,676][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1600 +[2026-06-20 00:36:03,676][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-20 00:36:03,690][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-20 00:36:14,420][starVLA.training.train_starvla][INFO] - Step 1601, Loss: {'train/loss': 0.20328624546527863, 'train/grad_norm_pre_clip': 11.157981872558594, 'timing/data': 0.01365046389400959, 'timing/model': 2.7783844843506813, 'learning_rate/qwen_vl_interface': 7.965055807032734e-06, 'learning_rate/action_model': 7.965055807032735e-05, 'epoch': 3.1}) +[2026-06-20 00:36:25,327][starVLA.training.train_starvla][INFO] - Step 1602, Loss: {'train/loss': 0.13924755156040192, 'train/grad_norm_pre_clip': 9.011073112487793, 'timing/data': 0.012162942439317703, 'timing/model': 2.7757488042116165, 'learning_rate/qwen_vl_interface': 7.962556362428275e-06, 'learning_rate/action_model': 7.962556362428274e-05, 'epoch': 3.1}) +[2026-06-20 00:36:36,424][starVLA.training.train_starvla][INFO] - Step 1603, Loss: {'train/loss': 0.10320346802473068, 'train/grad_norm_pre_clip': 7.326256275177002, 'timing/data': 0.01333775743842125, 'timing/model': 2.7652614656835794, 'learning_rate/qwen_vl_interface': 7.960055802794358e-06, 'learning_rate/action_model': 7.960055802794357e-05, 'epoch': 3.11}) +[2026-06-20 00:36:47,330][starVLA.training.train_starvla][INFO] - Step 1604, Loss: {'train/loss': 0.05407487601041794, 'train/grad_norm_pre_clip': 4.766706466674805, 'timing/data': 0.012495707720518112, 'timing/model': 2.774857424199581, 'learning_rate/qwen_vl_interface': 7.957554129158866e-06, 'learning_rate/action_model': 7.957554129158866e-05, 'epoch': 3.11}) +[2026-06-20 00:36:58,224][starVLA.training.train_starvla][INFO] - Step 1605, Loss: {'train/loss': 0.3030020594596863, 'train/grad_norm_pre_clip': 9.280981063842773, 'timing/data': 0.01272672601044178, 'timing/model': 2.7715632002800703, 'learning_rate/qwen_vl_interface': 7.955051342550148e-06, 'learning_rate/action_model': 7.955051342550148e-05, 'epoch': 3.11}) +[2026-06-20 00:37:09,140][starVLA.training.train_starvla][INFO] - Step 1606, Loss: {'train/loss': 0.16345299780368805, 'train/grad_norm_pre_clip': 5.014131546020508, 'timing/data': 0.0173899345099926, 'timing/model': 2.794134948402643, 'learning_rate/qwen_vl_interface': 7.952547443997e-06, 'learning_rate/action_model': 7.952547443997e-05, 'epoch': 3.11}) +[2026-06-20 00:37:20,018][starVLA.training.train_starvla][INFO] - Step 1607, Loss: {'train/loss': 0.15752023458480835, 'train/grad_norm_pre_clip': 11.903221130371094, 'timing/data': 0.012837130576372147, 'timing/model': 2.7714774683117867, 'learning_rate/qwen_vl_interface': 7.950042434528681e-06, 'learning_rate/action_model': 7.950042434528682e-05, 'epoch': 3.11}) +[2026-06-20 00:37:30,717][starVLA.training.train_starvla][INFO] - Step 1608, Loss: {'train/loss': 0.3369658589363098, 'train/grad_norm_pre_clip': 12.922658920288086, 'timing/data': 0.013627186417579651, 'timing/model': 2.752010265365243, 'learning_rate/qwen_vl_interface': 7.94753631517491e-06, 'learning_rate/action_model': 7.947536315174909e-05, 'epoch': 3.12}) +[2026-06-20 00:37:41,530][starVLA.training.train_starvla][INFO] - Step 1609, Loss: {'train/loss': 0.06131141632795334, 'train/grad_norm_pre_clip': 7.889542579650879, 'timing/data': 0.02148282155394554, 'timing/model': 2.7751188687980175, 'learning_rate/qwen_vl_interface': 7.945029086965854e-06, 'learning_rate/action_model': 7.945029086965853e-05, 'epoch': 3.12}) +[2026-06-20 00:37:52,431][starVLA.training.train_starvla][INFO] - Step 1610, Loss: {'train/loss': 0.05510694906115532, 'train/grad_norm_pre_clip': 8.669212341308594, 'timing/data': 0.015276636928319931, 'timing/model': 2.767492040991783, 'learning_rate/qwen_vl_interface': 7.942520750932138e-06, 'learning_rate/action_model': 7.942520750932138e-05, 'epoch': 3.12}) +[2026-06-20 00:38:03,333][starVLA.training.train_starvla][INFO] - Step 1611, Loss: {'train/loss': 0.10183728486299515, 'train/grad_norm_pre_clip': 4.042280197143555, 'timing/data': 0.01228460855782032, 'timing/model': 2.7669917568564415, 'learning_rate/qwen_vl_interface': 7.940011308104848e-06, 'learning_rate/action_model': 7.940011308104848e-05, 'epoch': 3.12}) +[2026-06-20 00:38:14,238][starVLA.training.train_starvla][INFO] - Step 1612, Loss: {'train/loss': 0.27341797947883606, 'train/grad_norm_pre_clip': 16.060379028320312, 'timing/data': 0.013668054714798927, 'timing/model': 2.765748519450426, 'learning_rate/qwen_vl_interface': 7.93750075951552e-06, 'learning_rate/action_model': 7.93750075951552e-05, 'epoch': 3.12}) +[2026-06-20 00:38:25,135][starVLA.training.train_starvla][INFO] - Step 1613, Loss: {'train/loss': 0.06445370614528656, 'train/grad_norm_pre_clip': 4.411264419555664, 'timing/data': 0.012779584154486656, 'timing/model': 2.7914701644331217, 'learning_rate/qwen_vl_interface': 7.934989106196145e-06, 'learning_rate/action_model': 7.934989106196146e-05, 'epoch': 3.13}) +[2026-06-20 00:38:36,286][starVLA.training.train_starvla][INFO] - Step 1614, Loss: {'train/loss': 0.15977509319782257, 'train/grad_norm_pre_clip': 9.748635292053223, 'timing/data': 0.013273587450385094, 'timing/model': 3.029162462800741, 'learning_rate/qwen_vl_interface': 7.93247634917917e-06, 'learning_rate/action_model': 7.93247634917917e-05, 'epoch': 3.13}) +[2026-06-20 00:38:47,133][starVLA.training.train_starvla][INFO] - Step 1615, Loss: {'train/loss': 0.21590650081634521, 'train/grad_norm_pre_clip': 10.927168846130371, 'timing/data': 0.012872835621237755, 'timing/model': 2.798023486509919, 'learning_rate/qwen_vl_interface': 7.929962489497494e-06, 'learning_rate/action_model': 7.929962489497493e-05, 'epoch': 3.13}) +[2026-06-20 00:38:58,037][starVLA.training.train_starvla][INFO] - Step 1616, Loss: {'train/loss': 0.07001453638076782, 'train/grad_norm_pre_clip': 5.7305006980896, 'timing/data': 0.012987921014428139, 'timing/model': 2.7835738472640514, 'learning_rate/qwen_vl_interface': 7.927447528184469e-06, 'learning_rate/action_model': 7.927447528184468e-05, 'epoch': 3.13}) +[2026-06-20 00:39:08,920][starVLA.training.train_starvla][INFO] - Step 1617, Loss: {'train/loss': 0.07867283374071121, 'train/grad_norm_pre_clip': 7.665501117706299, 'timing/data': 0.012833373621106148, 'timing/model': 2.759227328002453, 'learning_rate/qwen_vl_interface': 7.924931466273901e-06, 'learning_rate/action_model': 7.924931466273902e-05, 'epoch': 3.13}) +[2026-06-20 00:39:19,854][starVLA.training.train_starvla][INFO] - Step 1618, Loss: {'train/loss': 0.34428927302360535, 'train/grad_norm_pre_clip': 8.42178726196289, 'timing/data': 0.012892311438918114, 'timing/model': 2.6679230760782957, 'learning_rate/qwen_vl_interface': 7.92241430480005e-06, 'learning_rate/action_model': 7.922414304800049e-05, 'epoch': 3.14}) +[2026-06-20 00:39:30,729][starVLA.training.train_starvla][INFO] - Step 1619, Loss: {'train/loss': 0.015355014242231846, 'train/grad_norm_pre_clip': 1.296874761581421, 'timing/data': 0.013212600722908974, 'timing/model': 2.756807181984186, 'learning_rate/qwen_vl_interface': 7.919896044797622e-06, 'learning_rate/action_model': 7.91989604479762e-05, 'epoch': 3.14}) +[2026-06-20 00:39:41,631][starVLA.training.train_starvla][INFO] - Step 1620, Loss: {'train/loss': 0.1875590831041336, 'train/grad_norm_pre_clip': 10.255620956420898, 'timing/data': 0.013253189623355865, 'timing/model': 2.762992037460208, 'learning_rate/qwen_vl_interface': 7.91737668730178e-06, 'learning_rate/action_model': 7.91737668730178e-05, 'epoch': 3.14}) +[2026-06-20 00:39:52,536][starVLA.training.train_starvla][INFO] - Step 1621, Loss: {'train/loss': 0.07580892741680145, 'train/grad_norm_pre_clip': 9.618205070495605, 'timing/data': 0.01313742808997631, 'timing/model': 2.764205252751708, 'learning_rate/qwen_vl_interface': 7.91485623334814e-06, 'learning_rate/action_model': 7.91485623334814e-05, 'epoch': 3.14}) +[2026-06-20 00:40:03,444][starVLA.training.train_starvla][INFO] - Step 1622, Loss: {'train/loss': 0.2436159998178482, 'train/grad_norm_pre_clip': 13.007986068725586, 'timing/data': 0.0131370909512043, 'timing/model': 2.7225766256451607, 'learning_rate/qwen_vl_interface': 7.91233468397276e-06, 'learning_rate/action_model': 7.912334683972761e-05, 'epoch': 3.14}) +[2026-06-20 00:40:14,331][starVLA.training.train_starvla][INFO] - Step 1623, Loss: {'train/loss': 0.18986661732196808, 'train/grad_norm_pre_clip': 6.814924716949463, 'timing/data': 0.013258630409836769, 'timing/model': 2.765294961631298, 'learning_rate/qwen_vl_interface': 7.90981204021216e-06, 'learning_rate/action_model': 7.90981204021216e-05, 'epoch': 3.15}) +[2026-06-20 00:40:25,214][starVLA.training.train_starvla][INFO] - Step 1624, Loss: {'train/loss': 0.19452634453773499, 'train/grad_norm_pre_clip': 7.048431873321533, 'timing/data': 0.013184290379285812, 'timing/model': 2.7577924616634846, 'learning_rate/qwen_vl_interface': 7.907288303103298e-06, 'learning_rate/action_model': 7.907288303103298e-05, 'epoch': 3.15}) +[2026-06-20 00:40:36,303][starVLA.training.train_starvla][INFO] - Step 1625, Loss: {'train/loss': 0.2689211368560791, 'train/grad_norm_pre_clip': 11.09523868560791, 'timing/data': 0.013508176431059837, 'timing/model': 3.0610270518809557, 'learning_rate/qwen_vl_interface': 7.90476347368359e-06, 'learning_rate/action_model': 7.90476347368359e-05, 'epoch': 3.15}) +[2026-06-20 00:40:47,033][starVLA.training.train_starvla][INFO] - Step 1626, Loss: {'train/loss': 0.24935635924339294, 'train/grad_norm_pre_clip': 14.751718521118164, 'timing/data': 0.012459827587008476, 'timing/model': 2.772365378215909, 'learning_rate/qwen_vl_interface': 7.902237552990897e-06, 'learning_rate/action_model': 7.902237552990897e-05, 'epoch': 3.15}) +[2026-06-20 00:40:57,936][starVLA.training.train_starvla][INFO] - Step 1627, Loss: {'train/loss': 0.2074582576751709, 'train/grad_norm_pre_clip': 10.492403030395508, 'timing/data': 0.012768084183335304, 'timing/model': 2.768609445542097, 'learning_rate/qwen_vl_interface': 7.89971054206353e-06, 'learning_rate/action_model': 7.89971054206353e-05, 'epoch': 3.15}) +[2026-06-20 00:41:08,834][starVLA.training.train_starvla][INFO] - Step 1628, Loss: {'train/loss': 0.03664226457476616, 'train/grad_norm_pre_clip': 4.9006123542785645, 'timing/data': 0.012775424867868423, 'timing/model': 2.777515532448888, 'learning_rate/qwen_vl_interface': 7.897182441940247e-06, 'learning_rate/action_model': 7.897182441940247e-05, 'epoch': 3.16}) +[2026-06-20 00:41:19,734][starVLA.training.train_starvla][INFO] - Step 1629, Loss: {'train/loss': 0.19778966903686523, 'train/grad_norm_pre_clip': 13.081168174743652, 'timing/data': 0.012572389096021652, 'timing/model': 2.7688549235463142, 'learning_rate/qwen_vl_interface': 7.894653253660256e-06, 'learning_rate/action_model': 7.894653253660256e-05, 'epoch': 3.16}) +[2026-06-20 00:41:30,637][starVLA.training.train_starvla][INFO] - Step 1630, Loss: {'train/loss': 0.06241079419851303, 'train/grad_norm_pre_clip': 4.8577880859375, 'timing/data': 0.012609856203198433, 'timing/model': 2.775541413575411, 'learning_rate/qwen_vl_interface': 7.89212297826321e-06, 'learning_rate/action_model': 7.892122978263209e-05, 'epoch': 3.16}) +[2026-06-20 00:41:41,534][starVLA.training.train_starvla][INFO] - Step 1631, Loss: {'train/loss': 0.07158653438091278, 'train/grad_norm_pre_clip': 5.414783954620361, 'timing/data': 0.012321308255195618, 'timing/model': 2.7667857445776463, 'learning_rate/qwen_vl_interface': 7.889591616789206e-06, 'learning_rate/action_model': 7.889591616789206e-05, 'epoch': 3.16}) +[2026-06-20 00:41:52,434][starVLA.training.train_starvla][INFO] - Step 1632, Loss: {'train/loss': 0.09362737834453583, 'train/grad_norm_pre_clip': 10.796358108520508, 'timing/data': 0.012786824256181717, 'timing/model': 2.773184238001704, 'learning_rate/qwen_vl_interface': 7.887059170278799e-06, 'learning_rate/action_model': 7.887059170278797e-05, 'epoch': 3.16}) +[2026-06-20 00:42:03,330][starVLA.training.train_starvla][INFO] - Step 1633, Loss: {'train/loss': 0.160355806350708, 'train/grad_norm_pre_clip': 12.354934692382812, 'timing/data': 0.012312488630414009, 'timing/model': 2.769199701026082, 'learning_rate/qwen_vl_interface': 7.884525639772974e-06, 'learning_rate/action_model': 7.884525639772974e-05, 'epoch': 3.16}) +[2026-06-20 00:42:14,223][starVLA.training.train_starvla][INFO] - Step 1634, Loss: {'train/loss': 0.049191251397132874, 'train/grad_norm_pre_clip': 3.476076126098633, 'timing/data': 0.012581195682287216, 'timing/model': 2.7566421926021576, 'learning_rate/qwen_vl_interface': 7.881991026313176e-06, 'learning_rate/action_model': 7.881991026313175e-05, 'epoch': 3.17}) +[2026-06-20 00:42:25,138][starVLA.training.train_starvla][INFO] - Step 1635, Loss: {'train/loss': 0.16225583851337433, 'train/grad_norm_pre_clip': 7.45708703994751, 'timing/data': 0.012585138902068138, 'timing/model': 2.774564154446125, 'learning_rate/qwen_vl_interface': 7.879455330941283e-06, 'learning_rate/action_model': 7.879455330941282e-05, 'epoch': 3.17}) +[2026-06-20 00:42:36,039][starVLA.training.train_starvla][INFO] - Step 1636, Loss: {'train/loss': 0.17665185034275055, 'train/grad_norm_pre_clip': 8.12295150756836, 'timing/data': 0.012598935514688492, 'timing/model': 2.7771558947861195, 'learning_rate/qwen_vl_interface': 7.876918554699627e-06, 'learning_rate/action_model': 7.876918554699627e-05, 'epoch': 3.17}) +[2026-06-20 00:42:47,026][starVLA.training.train_starvla][INFO] - Step 1637, Loss: {'train/loss': 0.11357160657644272, 'train/grad_norm_pre_clip': 6.123090744018555, 'timing/data': 0.012571597471833229, 'timing/model': 2.7552189752459526, 'learning_rate/qwen_vl_interface': 7.874380698630982e-06, 'learning_rate/action_model': 7.874380698630982e-05, 'epoch': 3.17}) +[2026-06-20 00:42:57,939][starVLA.training.train_starvla][INFO] - Step 1638, Loss: {'train/loss': 0.08957672119140625, 'train/grad_norm_pre_clip': 9.973672866821289, 'timing/data': 0.012599855661392212, 'timing/model': 2.78986157104373, 'learning_rate/qwen_vl_interface': 7.871841763778562e-06, 'learning_rate/action_model': 7.871841763778563e-05, 'epoch': 3.17}) +[2026-06-20 00:43:08,836][starVLA.training.train_starvla][INFO] - Step 1639, Loss: {'train/loss': 0.057437628507614136, 'train/grad_norm_pre_clip': 4.370642185211182, 'timing/data': 0.012393360957503319, 'timing/model': 2.7758481595665216, 'learning_rate/qwen_vl_interface': 7.86930175118603e-06, 'learning_rate/action_model': 7.86930175118603e-05, 'epoch': 3.18}) +[2026-06-20 00:43:19,735][starVLA.training.train_starvla][INFO] - Step 1640, Loss: {'train/loss': 0.19600734114646912, 'train/grad_norm_pre_clip': 7.684906005859375, 'timing/data': 0.01260821707546711, 'timing/model': 2.769096329808235, 'learning_rate/qwen_vl_interface': 7.866760661897487e-06, 'learning_rate/action_model': 7.866760661897487e-05, 'epoch': 3.18}) +[2026-06-20 00:43:30,636][starVLA.training.train_starvla][INFO] - Step 1641, Loss: {'train/loss': 0.39182019233703613, 'train/grad_norm_pre_clip': 13.307161331176758, 'timing/data': 0.012616856023669243, 'timing/model': 2.770321087911725, 'learning_rate/qwen_vl_interface': 7.86421849695748e-06, 'learning_rate/action_model': 7.864218496957481e-05, 'epoch': 3.18}) +[2026-06-20 00:43:41,534][starVLA.training.train_starvla][INFO] - Step 1642, Loss: {'train/loss': 0.132360577583313, 'train/grad_norm_pre_clip': 9.430876731872559, 'timing/data': 0.012421578168869019, 'timing/model': 2.776438344269991, 'learning_rate/qwen_vl_interface': 7.861675257410998e-06, 'learning_rate/action_model': 7.861675257410997e-05, 'epoch': 3.18}) +[2026-06-20 00:43:52,436][starVLA.training.train_starvla][INFO] - Step 1643, Loss: {'train/loss': 0.24225451052188873, 'train/grad_norm_pre_clip': 8.997909545898438, 'timing/data': 0.012703252956271172, 'timing/model': 2.768798204138875, 'learning_rate/qwen_vl_interface': 7.859130944303468e-06, 'learning_rate/action_model': 7.859130944303467e-05, 'epoch': 3.18}) +[2026-06-20 00:44:03,323][starVLA.training.train_starvla][INFO] - Step 1644, Loss: {'train/loss': 0.35112521052360535, 'train/grad_norm_pre_clip': 9.35057258605957, 'timing/data': 0.012363169342279434, 'timing/model': 2.7725033834576607, 'learning_rate/qwen_vl_interface': 7.856585558680761e-06, 'learning_rate/action_model': 7.856585558680761e-05, 'epoch': 3.19}) +[2026-06-20 00:44:14,440][starVLA.training.train_starvla][INFO] - Step 1645, Loss: {'train/loss': 0.12939324975013733, 'train/grad_norm_pre_clip': 6.369220733642578, 'timing/data': 0.012625446543097496, 'timing/model': 2.7724444456398487, 'learning_rate/qwen_vl_interface': 7.854039101589191e-06, 'learning_rate/action_model': 7.854039101589191e-05, 'epoch': 3.19}) +[2026-06-20 00:44:25,337][starVLA.training.train_starvla][INFO] - Step 1646, Loss: {'train/loss': 0.04675007611513138, 'train/grad_norm_pre_clip': 4.610432147979736, 'timing/data': 0.01293238252401352, 'timing/model': 2.773635381832719, 'learning_rate/qwen_vl_interface': 7.851491574075509e-06, 'learning_rate/action_model': 7.851491574075509e-05, 'epoch': 3.19}) +[2026-06-20 00:44:36,231][starVLA.training.train_starvla][INFO] - Step 1647, Loss: {'train/loss': 0.09509650617837906, 'train/grad_norm_pre_clip': 6.756292819976807, 'timing/data': 0.01272602565586567, 'timing/model': 2.773675512522459, 'learning_rate/qwen_vl_interface': 7.848942977186909e-06, 'learning_rate/action_model': 7.848942977186908e-05, 'epoch': 3.19}) +[2026-06-20 00:44:47,138][starVLA.training.train_starvla][INFO] - Step 1648, Loss: {'train/loss': 0.12499605864286423, 'train/grad_norm_pre_clip': 6.841418743133545, 'timing/data': 0.012583546340465546, 'timing/model': 2.7684633787721395, 'learning_rate/qwen_vl_interface': 7.84639331197102e-06, 'learning_rate/action_model': 7.84639331197102e-05, 'epoch': 3.19}) +[2026-06-20 00:44:58,039][starVLA.training.train_starvla][INFO] - Step 1649, Loss: {'train/loss': 0.11536403000354767, 'train/grad_norm_pre_clip': 6.355405807495117, 'timing/data': 0.012879250571131706, 'timing/model': 2.771611263975501, 'learning_rate/qwen_vl_interface': 7.843842579475918e-06, 'learning_rate/action_model': 7.843842579475918e-05, 'epoch': 3.2}) +[2026-06-20 00:45:09,028][starVLA.training.train_starvla][INFO] - Step 1650, Loss: {'train/loss': 0.2520141303539276, 'train/grad_norm_pre_clip': 8.824153900146484, 'timing/data': 0.025870494544506073, 'timing/model': 2.8492873441427946, 'learning_rate/qwen_vl_interface': 7.841290780750107e-06, 'learning_rate/action_model': 7.841290780750107e-05, 'epoch': 3.2}) +[2026-06-20 00:45:19,936][starVLA.training.train_starvla][INFO] - Step 1651, Loss: {'train/loss': 0.09146439284086227, 'train/grad_norm_pre_clip': 7.803628444671631, 'timing/data': 0.0238001998513937, 'timing/model': 2.754150738939643, 'learning_rate/qwen_vl_interface': 7.838737916842538e-06, 'learning_rate/action_model': 7.838737916842538e-05, 'epoch': 3.2}) +[2026-06-20 00:45:31,037][starVLA.training.train_starvla][INFO] - Step 1652, Loss: {'train/loss': 0.04837670549750328, 'train/grad_norm_pre_clip': 4.734090805053711, 'timing/data': 0.01918851211667061, 'timing/model': 2.962666179984808, 'learning_rate/qwen_vl_interface': 7.836183988802596e-06, 'learning_rate/action_model': 7.836183988802595e-05, 'epoch': 3.2}) +[2026-06-20 00:45:41,930][starVLA.training.train_starvla][INFO] - Step 1653, Loss: {'train/loss': 0.08839984983205795, 'train/grad_norm_pre_clip': 6.839529991149902, 'timing/data': 0.015744280070066452, 'timing/model': 2.767629411071539, 'learning_rate/qwen_vl_interface': 7.833628997680107e-06, 'learning_rate/action_model': 7.833628997680107e-05, 'epoch': 3.2}) +[2026-06-20 00:45:52,729][starVLA.training.train_starvla][INFO] - Step 1654, Loss: {'train/loss': 0.09566767513751984, 'train/grad_norm_pre_clip': 2.4718799591064453, 'timing/data': 0.012911960482597351, 'timing/model': 2.749249678105116, 'learning_rate/qwen_vl_interface': 7.831072944525329e-06, 'learning_rate/action_model': 7.831072944525328e-05, 'epoch': 3.21}) +[2026-06-20 00:46:03,632][starVLA.training.train_starvla][INFO] - Step 1655, Loss: {'train/loss': 0.08764476329088211, 'train/grad_norm_pre_clip': 7.403408527374268, 'timing/data': 0.012797091156244278, 'timing/model': 2.7799631394445896, 'learning_rate/qwen_vl_interface': 7.828515830388956e-06, 'learning_rate/action_model': 7.828515830388956e-05, 'epoch': 3.21}) +[2026-06-20 00:46:14,734][starVLA.training.train_starvla][INFO] - Step 1656, Loss: {'train/loss': 0.19702139496803284, 'train/grad_norm_pre_clip': 11.342302322387695, 'timing/data': 0.012848788872361183, 'timing/model': 2.7427568696439266, 'learning_rate/qwen_vl_interface': 7.825957656322129e-06, 'learning_rate/action_model': 7.825957656322128e-05, 'epoch': 3.21}) +[2026-06-20 00:46:25,627][starVLA.training.train_starvla][INFO] - Step 1657, Loss: {'train/loss': 0.1199539378285408, 'train/grad_norm_pre_clip': 10.17071533203125, 'timing/data': 0.012582750990986824, 'timing/model': 2.7740325164049864, 'learning_rate/qwen_vl_interface': 7.823398423376406e-06, 'learning_rate/action_model': 7.823398423376406e-05, 'epoch': 3.21}) +[2026-06-20 00:46:36,533][starVLA.training.train_starvla][INFO] - Step 1658, Loss: {'train/loss': 0.24974524974822998, 'train/grad_norm_pre_clip': 17.51268768310547, 'timing/data': 0.012434493750333786, 'timing/model': 2.782679120078683, 'learning_rate/qwen_vl_interface': 7.820838132603801e-06, 'learning_rate/action_model': 7.8208381326038e-05, 'epoch': 3.21}) +[2026-06-20 00:46:47,435][starVLA.training.train_starvla][INFO] - Step 1659, Loss: {'train/loss': 0.12982125580310822, 'train/grad_norm_pre_clip': 8.075108528137207, 'timing/data': 0.012997109442949295, 'timing/model': 2.7745831795036793, 'learning_rate/qwen_vl_interface': 7.818276785056748e-06, 'learning_rate/action_model': 7.818276785056749e-05, 'epoch': 3.22}) +[2026-06-20 00:46:58,534][starVLA.training.train_starvla][INFO] - Step 1660, Loss: {'train/loss': 0.11255612969398499, 'train/grad_norm_pre_clip': 5.878717422485352, 'timing/data': 0.012497320771217346, 'timing/model': 2.771396128460765, 'learning_rate/qwen_vl_interface': 7.815714381788123e-06, 'learning_rate/action_model': 7.815714381788122e-05, 'epoch': 3.22}) +[2026-06-20 00:47:09,432][starVLA.training.train_starvla][INFO] - Step 1661, Loss: {'train/loss': 0.11455963551998138, 'train/grad_norm_pre_clip': 8.070233345031738, 'timing/data': 0.013232145458459854, 'timing/model': 2.767699582502246, 'learning_rate/qwen_vl_interface': 7.813150923851227e-06, 'learning_rate/action_model': 7.813150923851227e-05, 'epoch': 3.22}) +[2026-06-20 00:47:20,324][starVLA.training.train_starvla][INFO] - Step 1662, Loss: {'train/loss': 0.10702956467866898, 'train/grad_norm_pre_clip': 8.070598602294922, 'timing/data': 0.013050626963376999, 'timing/model': 2.759899288415909, 'learning_rate/qwen_vl_interface': 7.810586412299807e-06, 'learning_rate/action_model': 7.810586412299806e-05, 'epoch': 3.22}) +[2026-06-20 00:47:31,237][starVLA.training.train_starvla][INFO] - Step 1663, Loss: {'train/loss': 0.06898822635412216, 'train/grad_norm_pre_clip': 5.2528910636901855, 'timing/data': 0.013194935396313667, 'timing/model': 2.781460005789995, 'learning_rate/qwen_vl_interface': 7.808020848188034e-06, 'learning_rate/action_model': 7.808020848188034e-05, 'epoch': 3.22}) +[2026-06-20 00:47:42,237][starVLA.training.train_starvla][INFO] - Step 1664, Loss: {'train/loss': 0.10753658413887024, 'train/grad_norm_pre_clip': 6.961932182312012, 'timing/data': 0.013479480519890785, 'timing/model': 2.7748408168554306, 'learning_rate/qwen_vl_interface': 7.805454232570517e-06, 'learning_rate/action_model': 7.805454232570517e-05, 'epoch': 3.23}) +[2026-06-20 00:47:52,962][starVLA.training.train_starvla][INFO] - Step 1665, Loss: {'train/loss': 0.0912800058722496, 'train/grad_norm_pre_clip': 6.38529634475708, 'timing/data': 0.013299951329827309, 'timing/model': 2.691704139113426, 'learning_rate/qwen_vl_interface': 7.80288656650229e-06, 'learning_rate/action_model': 7.80288656650229e-05, 'epoch': 3.23}) +[2026-06-20 00:48:03,836][starVLA.training.train_starvla][INFO] - Step 1666, Loss: {'train/loss': 0.01937827281653881, 'train/grad_norm_pre_clip': 1.9651317596435547, 'timing/data': 0.012942695990204811, 'timing/model': 2.780941516160965, 'learning_rate/qwen_vl_interface': 7.800317851038829e-06, 'learning_rate/action_model': 7.800317851038828e-05, 'epoch': 3.23}) +[2026-06-20 00:48:14,727][starVLA.training.train_starvla][INFO] - Step 1667, Loss: {'train/loss': 0.2107294499874115, 'train/grad_norm_pre_clip': 17.3631649017334, 'timing/data': 0.011691145598888397, 'timing/model': 2.762810932472348, 'learning_rate/qwen_vl_interface': 7.797748087236033e-06, 'learning_rate/action_model': 7.797748087236033e-05, 'epoch': 3.23}) +[2026-06-20 00:48:26,746][starVLA.training.train_starvla][INFO] - Latency curriculum phase 2 at step 1668: active_latencies=[2], dataset_size=73475 +[2026-06-20 00:48:26,748][starVLA.training.train_starvla][INFO] - Step 1668, Loss: {'train/loss': 0.19401337206363678, 'train/grad_norm_pre_clip': 18.038488388061523, 'latency_curriculum/phase': 2.0, 'latency_curriculum/active_count': 1.0, 'latency_curriculum/max_active_latency': 2.0, 'latency_curriculum/dataset_size': 73475.0, 'timing/data': 0.012541720643639565, 'timing/model': 2.76892114803195, 'learning_rate/qwen_vl_interface': 7.795177276150234e-06, 'learning_rate/action_model': 7.795177276150234e-05, 'epoch': 2.91}) +[2026-06-20 00:48:58,528][starVLA.training.train_starvla][INFO] - Step 1669, Loss: {'train/loss': 1.0919080972671509, 'train/grad_norm_pre_clip': 28.966571807861328, 'timing/data': 0.017889751121401787, 'timing/model': 2.767488880082965, 'learning_rate/qwen_vl_interface': 7.792605418838196e-06, 'learning_rate/action_model': 7.792605418838197e-05, 'epoch': 2.91}) +[2026-06-20 00:49:09,320][starVLA.training.train_starvla][INFO] - Step 1670, Loss: {'train/loss': 1.0034193992614746, 'train/grad_norm_pre_clip': 21.355911254882812, 'timing/data': 0.01809135638177395, 'timing/model': 2.755879720672965, 'learning_rate/qwen_vl_interface': 7.790032516357116e-06, 'learning_rate/action_model': 7.790032516357117e-05, 'epoch': 2.91}) +[2026-06-20 00:49:20,130][starVLA.training.train_starvla][INFO] - Step 1671, Loss: {'train/loss': 1.0694911479949951, 'train/grad_norm_pre_clip': 21.130748748779297, 'timing/data': 0.020463107153773308, 'timing/model': 2.76248118840158, 'learning_rate/qwen_vl_interface': 7.787458569764613e-06, 'learning_rate/action_model': 7.787458569764612e-05, 'epoch': 2.91}) +[2026-06-20 00:49:31,032][starVLA.training.train_starvla][INFO] - Step 1672, Loss: {'train/loss': 0.8156435489654541, 'train/grad_norm_pre_clip': 15.025460243225098, 'timing/data': 0.02032174915075302, 'timing/model': 2.763341637328267, 'learning_rate/qwen_vl_interface': 7.784883580118738e-06, 'learning_rate/action_model': 7.784883580118738e-05, 'epoch': 2.91}) +[2026-06-20 00:49:41,932][starVLA.training.train_starvla][INFO] - Step 1673, Loss: {'train/loss': 0.3810659646987915, 'train/grad_norm_pre_clip': 9.877211570739746, 'timing/data': 0.012259107083082199, 'timing/model': 2.77419313788414, 'learning_rate/qwen_vl_interface': 7.782307548477975e-06, 'learning_rate/action_model': 7.782307548477975e-05, 'epoch': 2.91}) +[2026-06-20 00:49:52,835][starVLA.training.train_starvla][INFO] - Step 1674, Loss: {'train/loss': 0.7408220171928406, 'train/grad_norm_pre_clip': 15.866827011108398, 'timing/data': 0.014167102053761482, 'timing/model': 2.7712381444871426, 'learning_rate/qwen_vl_interface': 7.779730475901233e-06, 'learning_rate/action_model': 7.779730475901232e-05, 'epoch': 2.92}) +[2026-06-20 00:50:03,935][starVLA.training.train_starvla][INFO] - Step 1675, Loss: {'train/loss': 0.618423342704773, 'train/grad_norm_pre_clip': 13.109477043151855, 'timing/data': 0.006913518533110619, 'timing/model': 2.8060504347085953, 'learning_rate/qwen_vl_interface': 7.777152363447849e-06, 'learning_rate/action_model': 7.77715236344785e-05, 'epoch': 2.92}) +[2026-06-20 00:50:14,835][starVLA.training.train_starvla][INFO] - Step 1676, Loss: {'train/loss': 0.35396936535835266, 'train/grad_norm_pre_clip': 5.150400161743164, 'timing/data': 0.007599933072924614, 'timing/model': 2.788618667051196, 'learning_rate/qwen_vl_interface': 7.774573212177587e-06, 'learning_rate/action_model': 7.774573212177588e-05, 'epoch': 2.92}) +[2026-06-20 00:50:25,738][starVLA.training.train_starvla][INFO] - Step 1677, Loss: {'train/loss': 0.4975329637527466, 'train/grad_norm_pre_clip': 8.51413631439209, 'timing/data': 0.006778525188565254, 'timing/model': 2.7957110181450844, 'learning_rate/qwen_vl_interface': 7.77199302315064e-06, 'learning_rate/action_model': 7.77199302315064e-05, 'epoch': 2.92}) +[2026-06-20 00:50:36,616][starVLA.training.train_starvla][INFO] - Step 1678, Loss: {'train/loss': 0.4799434542655945, 'train/grad_norm_pre_clip': 13.359488487243652, 'timing/data': 0.006815515458583832, 'timing/model': 2.7682665158063173, 'learning_rate/qwen_vl_interface': 7.769411797427622e-06, 'learning_rate/action_model': 7.769411797427623e-05, 'epoch': 2.92}) +[2026-06-20 00:50:47,533][starVLA.training.train_starvla][INFO] - Step 1679, Loss: {'train/loss': 0.5822833180427551, 'train/grad_norm_pre_clip': 5.859084606170654, 'timing/data': 0.006838055327534676, 'timing/model': 2.7707194704562426, 'learning_rate/qwen_vl_interface': 7.766829536069584e-06, 'learning_rate/action_model': 7.766829536069584e-05, 'epoch': 2.92}) +[2026-06-20 00:50:58,436][starVLA.training.train_starvla][INFO] - Step 1680, Loss: {'train/loss': 0.5887859463691711, 'train/grad_norm_pre_clip': 5.584522724151611, 'timing/data': 0.006881868466734886, 'timing/model': 2.7797954734414816, 'learning_rate/qwen_vl_interface': 7.764246240137991e-06, 'learning_rate/action_model': 7.764246240137991e-05, 'epoch': 2.93}) +[2026-06-20 00:51:09,226][starVLA.training.train_starvla][INFO] - Step 1681, Loss: {'train/loss': 0.5091512799263, 'train/grad_norm_pre_clip': 7.060197830200195, 'timing/data': 0.0068670157343149185, 'timing/model': 2.7493074759840965, 'learning_rate/qwen_vl_interface': 7.761661910694741e-06, 'learning_rate/action_model': 7.761661910694741e-05, 'epoch': 2.93}) +[2026-06-20 00:51:20,131][starVLA.training.train_starvla][INFO] - Step 1682, Loss: {'train/loss': 0.446854829788208, 'train/grad_norm_pre_clip': 3.925626754760742, 'timing/data': 0.006883123889565468, 'timing/model': 2.78587506711483, 'learning_rate/qwen_vl_interface': 7.759076548802153e-06, 'learning_rate/action_model': 7.759076548802152e-05, 'epoch': 2.93}) +[2026-06-20 00:51:31,033][starVLA.training.train_starvla][INFO] - Step 1683, Loss: {'train/loss': 0.49332454800605774, 'train/grad_norm_pre_clip': 5.591362953186035, 'timing/data': 0.006989356130361557, 'timing/model': 2.7614777088165283, 'learning_rate/qwen_vl_interface': 7.75649015552297e-06, 'learning_rate/action_model': 7.75649015552297e-05, 'epoch': 2.93}) +[2026-06-20 00:51:41,929][starVLA.training.train_starvla][INFO] - Step 1684, Loss: {'train/loss': 0.4973584711551666, 'train/grad_norm_pre_clip': 5.524012565612793, 'timing/data': 0.006801215931773186, 'timing/model': 2.772365240380168, 'learning_rate/qwen_vl_interface': 7.753902731920364e-06, 'learning_rate/action_model': 7.753902731920364e-05, 'epoch': 2.93}) +[2026-06-20 00:51:52,838][starVLA.training.train_starvla][INFO] - Step 1685, Loss: {'train/loss': 0.45824238657951355, 'train/grad_norm_pre_clip': 6.644413948059082, 'timing/data': 0.0066762082278728485, 'timing/model': 2.7826955895870924, 'learning_rate/qwen_vl_interface': 7.751314279057925e-06, 'learning_rate/action_model': 7.751314279057923e-05, 'epoch': 2.94}) +[2026-06-20 00:52:03,731][starVLA.training.train_starvla][INFO] - Step 1686, Loss: {'train/loss': 0.4389786720275879, 'train/grad_norm_pre_clip': 3.388577699661255, 'timing/data': 0.006562408059835434, 'timing/model': 2.7803417406976223, 'learning_rate/qwen_vl_interface': 7.74872479799967e-06, 'learning_rate/action_model': 7.74872479799967e-05, 'epoch': 2.94}) +[2026-06-20 00:52:14,823][starVLA.training.train_starvla][INFO] - Step 1687, Loss: {'train/loss': 0.47005099058151245, 'train/grad_norm_pre_clip': 4.25949764251709, 'timing/data': 0.007037732750177383, 'timing/model': 2.769944431260228, 'learning_rate/qwen_vl_interface': 7.746134289810034e-06, 'learning_rate/action_model': 7.746134289810033e-05, 'epoch': 2.94}) +[2026-06-20 00:52:25,730][starVLA.training.train_starvla][INFO] - Step 1688, Loss: {'train/loss': 0.44012171030044556, 'train/grad_norm_pre_clip': 6.256351947784424, 'timing/data': 0.006803575903177261, 'timing/model': 2.7758141923695803, 'learning_rate/qwen_vl_interface': 7.74354275555388e-06, 'learning_rate/action_model': 7.74354275555388e-05, 'epoch': 2.94}) +[2026-06-20 00:52:36,650][starVLA.training.train_starvla][INFO] - Step 1689, Loss: {'train/loss': 0.41189974546432495, 'train/grad_norm_pre_clip': 2.7780187129974365, 'timing/data': 0.006892092525959015, 'timing/model': 2.6973058562725782, 'learning_rate/qwen_vl_interface': 7.740950196296488e-06, 'learning_rate/action_model': 7.740950196296488e-05, 'epoch': 2.94}) +[2026-06-20 00:52:47,393][starVLA.training.train_starvla][INFO] - Step 1690, Loss: {'train/loss': 0.4731658101081848, 'train/grad_norm_pre_clip': 3.417994976043701, 'timing/data': 0.006615595892071724, 'timing/model': 2.822653690353036, 'learning_rate/qwen_vl_interface': 7.738356613103562e-06, 'learning_rate/action_model': 7.73835661310356e-05, 'epoch': 2.94}) +[2026-06-20 00:52:58,235][starVLA.training.train_starvla][INFO] - Step 1691, Loss: {'train/loss': 0.45275411009788513, 'train/grad_norm_pre_clip': 3.5393083095550537, 'timing/data': 0.00682472437620163, 'timing/model': 2.779273984953761, 'learning_rate/qwen_vl_interface': 7.735762007041226e-06, 'learning_rate/action_model': 7.735762007041226e-05, 'epoch': 2.95}) +[2026-06-20 00:53:09,131][starVLA.training.train_starvla][INFO] - Step 1692, Loss: {'train/loss': 0.5582844614982605, 'train/grad_norm_pre_clip': 4.555156707763672, 'timing/data': 0.0066202059388160706, 'timing/model': 2.7807965204119682, 'learning_rate/qwen_vl_interface': 7.733166379176023e-06, 'learning_rate/action_model': 7.733166379176023e-05, 'epoch': 2.95}) +[2026-06-20 00:53:20,031][starVLA.training.train_starvla][INFO] - Step 1693, Loss: {'train/loss': 0.478667676448822, 'train/grad_norm_pre_clip': 6.043013095855713, 'timing/data': 0.006876453757286072, 'timing/model': 2.7697532698512077, 'learning_rate/qwen_vl_interface': 7.730569730574921e-06, 'learning_rate/action_model': 7.73056973057492e-05, 'epoch': 2.95}) +[2026-06-20 00:53:31,165][starVLA.training.train_starvla][INFO] - Step 1694, Loss: {'train/loss': 0.5396417379379272, 'train/grad_norm_pre_clip': 4.620471477508545, 'timing/data': 0.006920244544744492, 'timing/model': 3.0102025847882032, 'learning_rate/qwen_vl_interface': 7.727972062305299e-06, 'learning_rate/action_model': 7.727972062305299e-05, 'epoch': 2.95}) +[2026-06-20 00:53:42,032][starVLA.training.train_starvla][INFO] - Step 1695, Loss: {'train/loss': 0.32180824875831604, 'train/grad_norm_pre_clip': 4.196526527404785, 'timing/data': 0.006639726459980011, 'timing/model': 2.7747154626995325, 'learning_rate/qwen_vl_interface': 7.725373375434965e-06, 'learning_rate/action_model': 7.725373375434965e-05, 'epoch': 2.95}) +[2026-06-20 00:53:52,919][starVLA.training.train_starvla][INFO] - Step 1696, Loss: {'train/loss': 0.31353914737701416, 'train/grad_norm_pre_clip': 5.217731475830078, 'timing/data': 0.007325800135731697, 'timing/model': 2.7635132055729628, 'learning_rate/qwen_vl_interface': 7.722773671032141e-06, 'learning_rate/action_model': 7.722773671032142e-05, 'epoch': 2.95}) +[2026-06-20 00:54:03,828][starVLA.training.train_starvla][INFO] - Step 1697, Loss: {'train/loss': 0.3788731098175049, 'train/grad_norm_pre_clip': 7.727225303649902, 'timing/data': 0.0066404081881046295, 'timing/model': 2.7766465973109007, 'learning_rate/qwen_vl_interface': 7.720172950165466e-06, 'learning_rate/action_model': 7.720172950165465e-05, 'epoch': 2.96}) +[2026-06-20 00:54:14,733][starVLA.training.train_starvla][INFO] - Step 1698, Loss: {'train/loss': 0.28354424238204956, 'train/grad_norm_pre_clip': 5.100313186645508, 'timing/data': 0.006512967869639397, 'timing/model': 2.776481283828616, 'learning_rate/qwen_vl_interface': 7.717571213903996e-06, 'learning_rate/action_model': 7.717571213903996e-05, 'epoch': 2.96}) +[2026-06-20 00:54:25,637][starVLA.training.train_starvla][INFO] - Step 1699, Loss: {'train/loss': 0.4102720618247986, 'train/grad_norm_pre_clip': 7.201212406158447, 'timing/data': 0.006721945479512215, 'timing/model': 2.7805443052202463, 'learning_rate/qwen_vl_interface': 7.714968463317207e-06, 'learning_rate/action_model': 7.714968463317207e-05, 'epoch': 2.96}) +[2026-06-20 00:54:36,523][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1700 +[2026-06-20 01:01:52,531][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1700 +[2026-06-20 01:01:52,532][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 1700 +[2026-06-20 01:10:02,571][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 1700 +[2026-06-20 01:10:02,573][starVLA.training.train_starvla][INFO] - Step 1700, Loss: {'train/loss': 0.4431453049182892, 'train/grad_norm_pre_clip': 9.018235206604004, 'eval/loss': 0.3297713100910187, 'eval/latency_0/loss': 0.2722429037094116, 'eval/latency_1/loss': 0.23673421144485474, 'eval/latency_2/loss': 0.2087756246328354, 'eval/latency_3/loss': 0.3637951612472534, 'eval/latency_4/loss': 0.5742392539978027, 'eval/flappy/loss': 0.3297713100910187, 'eval/flappy/latency_0/loss': 0.2722429037094116, 'eval/flappy/latency_1/loss': 0.23673421144485474, 'eval/flappy/latency_2/loss': 0.2087756246328354, 'eval/flappy/latency_3/loss': 0.3637951612472534, 'eval/flappy/latency_4/loss': 0.5742392539978027, 'eval/flap_precision': 0.8155339805825242, 'eval/flap_recall': 0.3054545454545455, 'eval/flap_f1': 0.4444444444444444, 'eval/flap_rate_model': 0.02861111111111111, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.4444444444444444, 'eval/latency_0/flap_precision': 0.8155339805825242, 'eval/latency_0/flap_recall': 0.3054545454545455, 'eval/latency_0/flap_f1': 0.4444444444444444, 'eval/latency_0/flap_rate_model': 0.02861111111111111, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.4444444444444444, 'timing/data': 0.006587797775864601, 'timing/model': 2.7665694523602724, 'learning_rate/qwen_vl_interface': 7.712364699474996e-06, 'learning_rate/action_model': 7.712364699474996e-05, 'epoch': 2.96}) +[2026-06-20 01:10:02,573][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_1700_state +[2026-06-20 01:10:02,593][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-20 01:10:08,359][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_1700_state/model.safetensors +[2026-06-20 01:10:22,475][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_1700_state/optimizer.bin +[2026-06-20 01:10:22,475][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_1700_state/sampler.bin +[2026-06-20 01:10:22,475][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_1700_state/sampler_1.bin +[2026-06-20 01:10:22,476][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_1700_state/random_states_0.pkl +[2026-06-20 01:10:22,602][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1700 +[2026-06-20 01:10:22,602][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-20 01:10:22,616][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-20 01:10:33,471][starVLA.training.train_starvla][INFO] - Step 1701, Loss: {'train/loss': 0.33523595333099365, 'train/grad_norm_pre_clip': 8.526941299438477, 'timing/data': 0.0067055001854896545, 'timing/model': 2.741091478615999, 'learning_rate/qwen_vl_interface': 7.709759923447667e-06, 'learning_rate/action_model': 7.709759923447665e-05, 'epoch': 2.96}) +[2026-06-20 01:10:44,316][starVLA.training.train_starvla][INFO] - Step 1702, Loss: {'train/loss': 0.5006207823753357, 'train/grad_norm_pre_clip': 11.127304077148438, 'timing/data': 0.006896987557411194, 'timing/model': 2.7798878867179155, 'learning_rate/qwen_vl_interface': 7.707154136305945e-06, 'learning_rate/action_model': 7.707154136305945e-05, 'epoch': 2.97}) +[2026-06-20 01:10:55,226][starVLA.training.train_starvla][INFO] - Step 1703, Loss: {'train/loss': 0.3157358467578888, 'train/grad_norm_pre_clip': 11.460321426391602, 'timing/data': 0.0067494530230760574, 'timing/model': 2.778006913140416, 'learning_rate/qwen_vl_interface': 7.704547339120974e-06, 'learning_rate/action_model': 7.704547339120974e-05, 'epoch': 2.97}) +[2026-06-20 01:11:06,126][starVLA.training.train_starvla][INFO] - Step 1704, Loss: {'train/loss': 0.5311195254325867, 'train/grad_norm_pre_clip': 9.457274436950684, 'timing/data': 0.006996387615799904, 'timing/model': 2.776684360578656, 'learning_rate/qwen_vl_interface': 7.70193953296431e-06, 'learning_rate/action_model': 7.70193953296431e-05, 'epoch': 2.97}) +[2026-06-20 01:11:16,945][starVLA.training.train_starvla][INFO] - Step 1705, Loss: {'train/loss': 0.4573401212692261, 'train/grad_norm_pre_clip': 6.377574443817139, 'timing/data': 0.006696011871099472, 'timing/model': 2.704307295382023, 'learning_rate/qwen_vl_interface': 7.69933071890792e-06, 'learning_rate/action_model': 7.699330718907919e-05, 'epoch': 2.97}) +[2026-06-20 01:11:27,839][starVLA.training.train_starvla][INFO] - Step 1706, Loss: {'train/loss': 0.2314566671848297, 'train/grad_norm_pre_clip': 4.64073371887207, 'timing/data': 0.007365334779024124, 'timing/model': 2.7924488969147205, 'learning_rate/qwen_vl_interface': 7.696720898024188e-06, 'learning_rate/action_model': 7.69672089802419e-05, 'epoch': 2.97}) +[2026-06-20 01:11:38,630][starVLA.training.train_starvla][INFO] - Step 1707, Loss: {'train/loss': 0.2788800597190857, 'train/grad_norm_pre_clip': 9.142740249633789, 'timing/data': 0.006659060716629028, 'timing/model': 2.756029825657606, 'learning_rate/qwen_vl_interface': 7.69411007138592e-06, 'learning_rate/action_model': 7.694110071385919e-05, 'epoch': 2.97}) +[2026-06-20 01:11:49,427][starVLA.training.train_starvla][INFO] - Step 1708, Loss: {'train/loss': 0.4099445343017578, 'train/grad_norm_pre_clip': 6.479133605957031, 'timing/data': 0.006881719455122948, 'timing/model': 2.7876728251576424, 'learning_rate/qwen_vl_interface': 7.691498240066322e-06, 'learning_rate/action_model': 7.691498240066322e-05, 'epoch': 2.98}) +[2026-06-20 01:12:00,198][starVLA.training.train_starvla][INFO] - Step 1709, Loss: {'train/loss': 0.27294713258743286, 'train/grad_norm_pre_clip': 10.28071117401123, 'timing/data': 0.006934858858585358, 'timing/model': 2.767147980630398, 'learning_rate/qwen_vl_interface': 7.688885405139022e-06, 'learning_rate/action_model': 7.688885405139021e-05, 'epoch': 2.98}) +[2026-06-20 01:12:10,920][starVLA.training.train_starvla][INFO] - Step 1710, Loss: {'train/loss': 0.26915261149406433, 'train/grad_norm_pre_clip': 7.589756488800049, 'timing/data': 0.006796220317482948, 'timing/model': 2.7745490595698357, 'learning_rate/qwen_vl_interface': 7.686271567678056e-06, 'learning_rate/action_model': 7.686271567678054e-05, 'epoch': 2.98}) +[2026-06-20 01:12:21,763][starVLA.training.train_starvla][INFO] - Step 1711, Loss: {'train/loss': 0.4016384482383728, 'train/grad_norm_pre_clip': 19.057931900024414, 'timing/data': 0.006712459027767181, 'timing/model': 2.717582706362009, 'learning_rate/qwen_vl_interface': 7.683656728757873e-06, 'learning_rate/action_model': 7.683656728757872e-05, 'epoch': 2.98}) +[2026-06-20 01:12:32,619][starVLA.training.train_starvla][INFO] - Step 1712, Loss: {'train/loss': 0.352212518453598, 'train/grad_norm_pre_clip': 12.631278038024902, 'timing/data': 0.0066058821976184845, 'timing/model': 2.771025475114584, 'learning_rate/qwen_vl_interface': 7.681040889453337e-06, 'learning_rate/action_model': 7.681040889453336e-05, 'epoch': 2.98}) +[2026-06-20 01:12:43,555][starVLA.training.train_starvla][INFO] - Step 1713, Loss: {'train/loss': 0.412747323513031, 'train/grad_norm_pre_clip': 11.871106147766113, 'timing/data': 0.006424134597182274, 'timing/model': 2.751080885529518, 'learning_rate/qwen_vl_interface': 7.678424050839722e-06, 'learning_rate/action_model': 7.67842405083972e-05, 'epoch': 2.98}) +[2026-06-20 01:12:54,630][starVLA.training.train_starvla][INFO] - Step 1714, Loss: {'train/loss': 0.1435859352350235, 'train/grad_norm_pre_clip': 5.775922775268555, 'timing/data': 0.00684179924428463, 'timing/model': 2.7763417325913906, 'learning_rate/qwen_vl_interface': 7.675806213992706e-06, 'learning_rate/action_model': 7.675806213992707e-05, 'epoch': 2.99}) +[2026-06-20 01:13:05,516][starVLA.training.train_starvla][INFO] - Step 1715, Loss: {'train/loss': 0.2329806536436081, 'train/grad_norm_pre_clip': 9.400606155395508, 'timing/data': 0.006703000515699387, 'timing/model': 2.76271247677505, 'learning_rate/qwen_vl_interface': 7.673187379988388e-06, 'learning_rate/action_model': 7.673187379988388e-05, 'epoch': 2.99}) +[2026-06-20 01:13:16,321][starVLA.training.train_starvla][INFO] - Step 1716, Loss: {'train/loss': 0.26370662450790405, 'train/grad_norm_pre_clip': 7.732304573059082, 'timing/data': 0.006779739633202553, 'timing/model': 2.7661143969744444, 'learning_rate/qwen_vl_interface': 7.67056754990327e-06, 'learning_rate/action_model': 7.670567549903269e-05, 'epoch': 2.99}) +[2026-06-20 01:13:27,231][starVLA.training.train_starvla][INFO] - Step 1717, Loss: {'train/loss': 0.38873541355133057, 'train/grad_norm_pre_clip': 22.145198822021484, 'timing/data': 0.006850259378552437, 'timing/model': 2.784690383821726, 'learning_rate/qwen_vl_interface': 7.667946724814264e-06, 'learning_rate/action_model': 7.667946724814265e-05, 'epoch': 2.99}) +[2026-06-20 01:13:38,101][starVLA.training.train_starvla][INFO] - Step 1718, Loss: {'train/loss': 0.41059115529060364, 'train/grad_norm_pre_clip': 13.208842277526855, 'timing/data': 0.006811659783124924, 'timing/model': 2.7489351965487003, 'learning_rate/qwen_vl_interface': 7.665324905798694e-06, 'learning_rate/action_model': 7.665324905798695e-05, 'epoch': 2.99}) +[2026-06-20 01:13:48,931][starVLA.training.train_starvla][INFO] - Step 1719, Loss: {'train/loss': 0.377341091632843, 'train/grad_norm_pre_clip': 18.743541717529297, 'timing/data': 0.006766010075807571, 'timing/model': 2.7861117497086525, 'learning_rate/qwen_vl_interface': 7.66270209393429e-06, 'learning_rate/action_model': 7.66270209393429e-05, 'epoch': 2.99}) +[2026-06-20 01:13:59,732][starVLA.training.train_starvla][INFO] - Step 1720, Loss: {'train/loss': 0.2536274194717407, 'train/grad_norm_pre_clip': 9.915481567382812, 'timing/data': 0.006539672613143921, 'timing/model': 2.7832018863409758, 'learning_rate/qwen_vl_interface': 7.66007829029919e-06, 'learning_rate/action_model': 7.660078290299191e-05, 'epoch': 3.0}) +[2026-06-20 01:14:10,633][starVLA.training.train_starvla][INFO] - Step 1721, Loss: {'train/loss': 0.3890725076198578, 'train/grad_norm_pre_clip': 25.031660079956055, 'timing/data': 0.006586020812392235, 'timing/model': 2.779841335490346, 'learning_rate/qwen_vl_interface': 7.657453495971942e-06, 'learning_rate/action_model': 7.657453495971943e-05, 'epoch': 3.0}) +[2026-06-20 01:14:21,735][starVLA.training.train_starvla][INFO] - Step 1722, Loss: {'train/loss': 0.21292519569396973, 'train/grad_norm_pre_clip': 7.637755393981934, 'timing/data': 0.0066533517092466354, 'timing/model': 2.78747015632689, 'learning_rate/qwen_vl_interface': 7.6548277120315e-06, 'learning_rate/action_model': 7.6548277120315e-05, 'epoch': 3.0}) +[2026-06-20 01:14:32,633][starVLA.training.train_starvla][INFO] - Step 1723, Loss: {'train/loss': 0.2840009331703186, 'train/grad_norm_pre_clip': 10.925625801086426, 'timing/data': 0.006814751774072647, 'timing/model': 2.7716563642024994, 'learning_rate/qwen_vl_interface': 7.652200939557225e-06, 'learning_rate/action_model': 7.652200939557225e-05, 'epoch': 3.0}) +[2026-06-20 01:14:43,532][starVLA.training.train_starvla][INFO] - Step 1724, Loss: {'train/loss': 0.2536010444164276, 'train/grad_norm_pre_clip': 11.945406913757324, 'timing/data': 0.008694302290678024, 'timing/model': 2.7828944250941277, 'learning_rate/qwen_vl_interface': 7.649573179628881e-06, 'learning_rate/action_model': 7.649573179628881e-05, 'epoch': 3.0}) +[2026-06-20 01:14:54,647][starVLA.training.train_starvla][INFO] - Step 1725, Loss: {'train/loss': 0.21814116835594177, 'train/grad_norm_pre_clip': 8.237343788146973, 'timing/data': 0.006750361993908882, 'timing/model': 2.990356156602502, 'learning_rate/qwen_vl_interface': 7.646944433326645e-06, 'learning_rate/action_model': 7.646944433326643e-05, 'epoch': 3.01}) +[2026-06-20 01:15:05,525][starVLA.training.train_starvla][INFO] - Step 1726, Loss: {'train/loss': 0.3235319256782532, 'train/grad_norm_pre_clip': 14.968111991882324, 'timing/data': 0.006349476054310799, 'timing/model': 2.7689922880381346, 'learning_rate/qwen_vl_interface': 7.64431470173109e-06, 'learning_rate/action_model': 7.64431470173109e-05, 'epoch': 3.01}) +[2026-06-20 01:15:16,425][starVLA.training.train_starvla][INFO] - Step 1727, Loss: {'train/loss': 0.42691078782081604, 'train/grad_norm_pre_clip': 24.36567497253418, 'timing/data': 0.006823420524597168, 'timing/model': 2.767051365226507, 'learning_rate/qwen_vl_interface': 7.641683985923203e-06, 'learning_rate/action_model': 7.641683985923203e-05, 'epoch': 3.01}) +[2026-06-20 01:15:27,164][starVLA.training.train_starvla][INFO] - Step 1728, Loss: {'train/loss': 0.30137884616851807, 'train/grad_norm_pre_clip': 12.614429473876953, 'timing/data': 0.006703503429889679, 'timing/model': 2.7297101728618145, 'learning_rate/qwen_vl_interface': 7.63905228698437e-06, 'learning_rate/action_model': 7.63905228698437e-05, 'epoch': 3.01}) +[2026-06-20 01:15:38,140][starVLA.training.train_starvla][INFO] - Step 1729, Loss: {'train/loss': 0.25864124298095703, 'train/grad_norm_pre_clip': 8.299049377441406, 'timing/data': 0.006513943895697594, 'timing/model': 2.761590253561735, 'learning_rate/qwen_vl_interface': 7.636419605996385e-06, 'learning_rate/action_model': 7.636419605996384e-05, 'epoch': 3.01}) +[2026-06-20 01:15:49,032][starVLA.training.train_starvla][INFO] - Step 1730, Loss: {'train/loss': 0.5095328092575073, 'train/grad_norm_pre_clip': 16.028669357299805, 'timing/data': 0.006538063287734985, 'timing/model': 2.781830832362175, 'learning_rate/qwen_vl_interface': 7.633785944041444e-06, 'learning_rate/action_model': 7.633785944041444e-05, 'epoch': 3.01}) +[2026-06-20 01:15:59,839][starVLA.training.train_starvla][INFO] - Step 1731, Loss: {'train/loss': 0.252465158700943, 'train/grad_norm_pre_clip': 11.988344192504883, 'timing/data': 0.006411226466298103, 'timing/model': 2.788757808506489, 'learning_rate/qwen_vl_interface': 7.631151302202145e-06, 'learning_rate/action_model': 7.631151302202144e-05, 'epoch': 3.02}) +[2026-06-20 01:16:10,631][starVLA.training.train_starvla][INFO] - Step 1732, Loss: {'train/loss': 0.28412926197052, 'train/grad_norm_pre_clip': 16.182716369628906, 'timing/data': 0.006557483226060867, 'timing/model': 2.7742337435483932, 'learning_rate/qwen_vl_interface': 7.628515681561488e-06, 'learning_rate/action_model': 7.628515681561487e-05, 'epoch': 3.02}) +[2026-06-20 01:16:21,342][starVLA.training.train_starvla][INFO] - Step 1733, Loss: {'train/loss': 0.4004385471343994, 'train/grad_norm_pre_clip': 12.630765914916992, 'timing/data': 0.006388895213603973, 'timing/model': 2.7607522383332253, 'learning_rate/qwen_vl_interface': 7.62587908320288e-06, 'learning_rate/action_model': 7.62587908320288e-05, 'epoch': 3.02}) +[2026-06-20 01:16:32,221][starVLA.training.train_starvla][INFO] - Step 1734, Loss: {'train/loss': 0.2634162902832031, 'train/grad_norm_pre_clip': 17.99195671081543, 'timing/data': 0.006781639531254768, 'timing/model': 2.761162854731083, 'learning_rate/qwen_vl_interface': 7.6232415082101255e-06, 'learning_rate/action_model': 7.623241508210124e-05, 'epoch': 3.02}) +[2026-06-20 01:16:43,134][starVLA.training.train_starvla][INFO] - Step 1735, Loss: {'train/loss': 0.16462570428848267, 'train/grad_norm_pre_clip': 10.90760326385498, 'timing/data': 0.006653971970081329, 'timing/model': 2.7760686855763197, 'learning_rate/qwen_vl_interface': 7.620602957667432e-06, 'learning_rate/action_model': 7.620602957667432e-05, 'epoch': 3.02}) +[2026-06-20 01:16:54,037][starVLA.training.train_starvla][INFO] - Step 1736, Loss: {'train/loss': 0.4432390034198761, 'train/grad_norm_pre_clip': 12.086881637573242, 'timing/data': 0.006517285481095314, 'timing/model': 2.788231836631894, 'learning_rate/qwen_vl_interface': 7.617963432659408e-06, 'learning_rate/action_model': 7.617963432659408e-05, 'epoch': 3.02}) +[2026-06-20 01:17:04,934][starVLA.training.train_starvla][INFO] - Step 1737, Loss: {'train/loss': 0.33676785230636597, 'train/grad_norm_pre_clip': 10.567813873291016, 'timing/data': 0.006595361977815628, 'timing/model': 2.776802035048604, 'learning_rate/qwen_vl_interface': 7.615322934271063e-06, 'learning_rate/action_model': 7.615322934271064e-05, 'epoch': 3.03}) +[2026-06-20 01:17:15,833][starVLA.training.train_starvla][INFO] - Step 1738, Loss: {'train/loss': 0.32980313897132874, 'train/grad_norm_pre_clip': 16.203519821166992, 'timing/data': 0.006582435220479965, 'timing/model': 2.770166840404272, 'learning_rate/qwen_vl_interface': 7.6126814635878055e-06, 'learning_rate/action_model': 7.612681463587805e-05, 'epoch': 3.03}) +[2026-06-20 01:17:26,736][starVLA.training.train_starvla][INFO] - Step 1739, Loss: {'train/loss': 0.3253912925720215, 'train/grad_norm_pre_clip': 9.299145698547363, 'timing/data': 0.006469812244176865, 'timing/model': 2.7782827019691467, 'learning_rate/qwen_vl_interface': 7.6100390216954445e-06, 'learning_rate/action_model': 7.610039021695445e-05, 'epoch': 3.03}) +[2026-06-20 01:17:37,634][starVLA.training.train_starvla][INFO] - Step 1740, Loss: {'train/loss': 0.27883002161979675, 'train/grad_norm_pre_clip': 11.552301406860352, 'timing/data': 0.006635161116719246, 'timing/model': 2.7787366043776274, 'learning_rate/qwen_vl_interface': 7.607395609680189e-06, 'learning_rate/action_model': 7.60739560968019e-05, 'epoch': 3.03}) +[2026-06-20 01:17:48,635][starVLA.training.train_starvla][INFO] - Step 1741, Loss: {'train/loss': 0.3365994095802307, 'train/grad_norm_pre_clip': 14.241546630859375, 'timing/data': 0.006394524127244949, 'timing/model': 2.7818338200449944, 'learning_rate/qwen_vl_interface': 7.604751228628644e-06, 'learning_rate/action_model': 7.604751228628644e-05, 'epoch': 3.03}) +[2026-06-20 01:17:59,537][starVLA.training.train_starvla][INFO] - Step 1742, Loss: {'train/loss': 0.25858405232429504, 'train/grad_norm_pre_clip': 5.938195705413818, 'timing/data': 0.006820252165198326, 'timing/model': 2.7778384871780872, 'learning_rate/qwen_vl_interface': 7.602105879627814e-06, 'learning_rate/action_model': 7.602105879627814e-05, 'epoch': 3.03}) +[2026-06-20 01:18:10,423][starVLA.training.train_starvla][INFO] - Step 1743, Loss: {'train/loss': 0.5461624264717102, 'train/grad_norm_pre_clip': 13.990606307983398, 'timing/data': 0.006700489670038223, 'timing/model': 2.7649640403687954, 'learning_rate/qwen_vl_interface': 7.599459563765107e-06, 'learning_rate/action_model': 7.599459563765106e-05, 'epoch': 3.04}) +[2026-06-20 01:18:21,316][starVLA.training.train_starvla][INFO] - Step 1744, Loss: {'train/loss': 0.3575925827026367, 'train/grad_norm_pre_clip': 8.63084602355957, 'timing/data': 0.006683379411697388, 'timing/model': 2.761576920747757, 'learning_rate/qwen_vl_interface': 7.596812282128319e-06, 'learning_rate/action_model': 7.596812282128319e-05, 'epoch': 3.04}) +[2026-06-20 01:18:32,022][starVLA.training.train_starvla][INFO] - Step 1745, Loss: {'train/loss': 0.3449099063873291, 'train/grad_norm_pre_clip': 12.477035522460938, 'timing/data': 0.006603270769119263, 'timing/model': 2.764409529045224, 'learning_rate/qwen_vl_interface': 7.594164035805648e-06, 'learning_rate/action_model': 7.594164035805648e-05, 'epoch': 3.04}) +[2026-06-20 01:18:42,936][starVLA.training.train_starvla][INFO] - Step 1746, Loss: {'train/loss': 0.2014603614807129, 'train/grad_norm_pre_clip': 5.984549522399902, 'timing/data': 0.006871931254863739, 'timing/model': 2.773236095905304, 'learning_rate/qwen_vl_interface': 7.59151482588569e-06, 'learning_rate/action_model': 7.59151482588569e-05, 'epoch': 3.04}) +[2026-06-20 01:18:53,834][starVLA.training.train_starvla][INFO] - Step 1747, Loss: {'train/loss': 0.38757792115211487, 'train/grad_norm_pre_clip': 10.460644721984863, 'timing/data': 0.00671728141605854, 'timing/model': 2.776690097525716, 'learning_rate/qwen_vl_interface': 7.588864653457433e-06, 'learning_rate/action_model': 7.588864653457432e-05, 'epoch': 3.04}) +[2026-06-20 01:19:04,734][starVLA.training.train_starvla][INFO] - Step 1748, Loss: {'train/loss': 0.25540557503700256, 'train/grad_norm_pre_clip': 10.026741027832031, 'timing/data': 0.006958620622754097, 'timing/model': 2.779420668259263, 'learning_rate/qwen_vl_interface': 7.5862135196102634e-06, 'learning_rate/action_model': 7.586213519610263e-05, 'epoch': 3.05}) +[2026-06-20 01:19:15,635][starVLA.training.train_starvla][INFO] - Step 1749, Loss: {'train/loss': 0.28916770219802856, 'train/grad_norm_pre_clip': 7.226225852966309, 'timing/data': 0.006452484056353569, 'timing/model': 2.774568459019065, 'learning_rate/qwen_vl_interface': 7.583561425433962e-06, 'learning_rate/action_model': 7.583561425433962e-05, 'epoch': 3.05}) +[2026-06-20 01:19:26,536][starVLA.training.train_starvla][INFO] - Step 1750, Loss: {'train/loss': 0.3066606819629669, 'train/grad_norm_pre_clip': 8.806401252746582, 'timing/data': 0.006853632628917694, 'timing/model': 2.7782230153679848, 'learning_rate/qwen_vl_interface': 7.580908372018706e-06, 'learning_rate/action_model': 7.580908372018705e-05, 'epoch': 3.05}) +[2026-06-20 01:19:37,431][starVLA.training.train_starvla][INFO] - Step 1751, Loss: {'train/loss': 0.4927896559238434, 'train/grad_norm_pre_clip': 12.16141128540039, 'timing/data': 0.006818441674113274, 'timing/model': 2.7715018317103386, 'learning_rate/qwen_vl_interface': 7.5782543604550655e-06, 'learning_rate/action_model': 7.578254360455064e-05, 'epoch': 3.05}) +[2026-06-20 01:19:48,441][starVLA.training.train_starvla][INFO] - Step 1752, Loss: {'train/loss': 0.3291303813457489, 'train/grad_norm_pre_clip': 11.298881530761719, 'timing/data': 0.006723182275891304, 'timing/model': 2.7209668438881636, 'learning_rate/qwen_vl_interface': 7.575599391834e-06, 'learning_rate/action_model': 7.575599391834e-05, 'epoch': 3.05}) +[2026-06-20 01:19:59,330][starVLA.training.train_starvla][INFO] - Step 1753, Loss: {'train/loss': 0.23883557319641113, 'train/grad_norm_pre_clip': 14.282971382141113, 'timing/data': 0.006665712222456932, 'timing/model': 2.767683584243059, 'learning_rate/qwen_vl_interface': 7.572943467246871e-06, 'learning_rate/action_model': 7.57294346724687e-05, 'epoch': 3.05}) +[2026-06-20 01:20:10,237][starVLA.training.train_starvla][INFO] - Step 1754, Loss: {'train/loss': 0.24536767601966858, 'train/grad_norm_pre_clip': 6.217752456665039, 'timing/data': 0.006544401869177818, 'timing/model': 2.780166583135724, 'learning_rate/qwen_vl_interface': 7.570286587785428e-06, 'learning_rate/action_model': 7.570286587785428e-05, 'epoch': 3.06}) +[2026-06-20 01:20:21,133][starVLA.training.train_starvla][INFO] - Step 1755, Loss: {'train/loss': 0.30278483033180237, 'train/grad_norm_pre_clip': 10.696438789367676, 'timing/data': 0.006711272522807121, 'timing/model': 2.7714860290288925, 'learning_rate/qwen_vl_interface': 7.567628754541815e-06, 'learning_rate/action_model': 7.567628754541814e-05, 'epoch': 3.06}) +[2026-06-20 01:20:32,033][starVLA.training.train_starvla][INFO] - Step 1756, Loss: {'train/loss': 0.23780131340026855, 'train/grad_norm_pre_clip': 5.883386135101318, 'timing/data': 0.006682712584733963, 'timing/model': 2.785250633955002, 'learning_rate/qwen_vl_interface': 7.564969968608565e-06, 'learning_rate/action_model': 7.564969968608565e-05, 'epoch': 3.06}) +[2026-06-20 01:20:42,934][starVLA.training.train_starvla][INFO] - Step 1757, Loss: {'train/loss': 0.26229923963546753, 'train/grad_norm_pre_clip': 6.720148086547852, 'timing/data': 0.006633361801505089, 'timing/model': 2.779263410717249, 'learning_rate/qwen_vl_interface': 7.562310231078605e-06, 'learning_rate/action_model': 7.562310231078604e-05, 'epoch': 3.06}) +[2026-06-20 01:20:53,834][starVLA.training.train_starvla][INFO] - Step 1758, Loss: {'train/loss': 0.2650616466999054, 'train/grad_norm_pre_clip': 6.2917914390563965, 'timing/data': 0.0066271815448999405, 'timing/model': 2.779542997479439, 'learning_rate/qwen_vl_interface': 7.559649543045252e-06, 'learning_rate/action_model': 7.559649543045252e-05, 'epoch': 3.06}) +[2026-06-20 01:21:04,730][starVLA.training.train_starvla][INFO] - Step 1759, Loss: {'train/loss': 0.1814732700586319, 'train/grad_norm_pre_clip': 7.4221649169921875, 'timing/data': 0.0067337993532419205, 'timing/model': 2.7697859834879637, 'learning_rate/qwen_vl_interface': 7.556987905602215e-06, 'learning_rate/action_model': 7.556987905602215e-05, 'epoch': 3.06}) +[2026-06-20 01:21:15,756][starVLA.training.train_starvla][INFO] - Step 1760, Loss: {'train/loss': 0.18929055333137512, 'train/grad_norm_pre_clip': 7.420587062835693, 'timing/data': 0.006693379953503609, 'timing/model': 2.6925433818250895, 'learning_rate/qwen_vl_interface': 7.554325319843593e-06, 'learning_rate/action_model': 7.554325319843592e-05, 'epoch': 3.07}) +[2026-06-20 01:21:26,534][starVLA.training.train_starvla][INFO] - Step 1761, Loss: {'train/loss': 0.2553398609161377, 'train/grad_norm_pre_clip': 16.3226318359375, 'timing/data': 0.007067708298563957, 'timing/model': 2.77399561367929, 'learning_rate/qwen_vl_interface': 7.551661786863874e-06, 'learning_rate/action_model': 7.551661786863874e-05, 'epoch': 3.07}) +[2026-06-20 01:21:37,430][starVLA.training.train_starvla][INFO] - Step 1762, Loss: {'train/loss': 0.4532863199710846, 'train/grad_norm_pre_clip': 10.733624458312988, 'timing/data': 0.0067181698977947235, 'timing/model': 2.764398481696844, 'learning_rate/qwen_vl_interface': 7.548997307757936e-06, 'learning_rate/action_model': 7.548997307757935e-05, 'epoch': 3.07}) +[2026-06-20 01:21:48,335][starVLA.training.train_starvla][INFO] - Step 1763, Loss: {'train/loss': 0.17258363962173462, 'train/grad_norm_pre_clip': 6.549299716949463, 'timing/data': 0.006708161905407906, 'timing/model': 2.7846126928925514, 'learning_rate/qwen_vl_interface': 7.546331883621044e-06, 'learning_rate/action_model': 7.546331883621044e-05, 'epoch': 3.07}) +[2026-06-20 01:21:59,424][starVLA.training.train_starvla][INFO] - Step 1764, Loss: {'train/loss': 0.14893656969070435, 'train/grad_norm_pre_clip': 9.843993186950684, 'timing/data': 0.0064738038927316666, 'timing/model': 2.7636644318699837, 'learning_rate/qwen_vl_interface': 7.543665515548857e-06, 'learning_rate/action_model': 7.543665515548856e-05, 'epoch': 3.07}) +[2026-06-20 01:22:10,336][starVLA.training.train_starvla][INFO] - Step 1765, Loss: {'train/loss': 0.2776840925216675, 'train/grad_norm_pre_clip': 8.67840576171875, 'timing/data': 0.007118677720427513, 'timing/model': 2.77889290638268, 'learning_rate/qwen_vl_interface': 7.540998204637413e-06, 'learning_rate/action_model': 7.540998204637413e-05, 'epoch': 3.07}) +[2026-06-20 01:22:21,227][starVLA.training.train_starvla][INFO] - Step 1766, Loss: {'train/loss': 0.3422660529613495, 'train/grad_norm_pre_clip': 12.404377937316895, 'timing/data': 0.00691135972738266, 'timing/model': 2.774487627670169, 'learning_rate/qwen_vl_interface': 7.538329951983148e-06, 'learning_rate/action_model': 7.538329951983147e-05, 'epoch': 3.08}) +[2026-06-20 01:22:32,186][starVLA.training.train_starvla][INFO] - Step 1767, Loss: {'train/loss': 0.2595571279525757, 'train/grad_norm_pre_clip': 12.69318962097168, 'timing/data': 0.006498392671346664, 'timing/model': 2.8427182864397764, 'learning_rate/qwen_vl_interface': 7.535660758682877e-06, 'learning_rate/action_model': 7.535660758682876e-05, 'epoch': 3.08}) +[2026-06-20 01:22:43,022][starVLA.training.train_starvla][INFO] - Step 1768, Loss: {'train/loss': 0.2326858639717102, 'train/grad_norm_pre_clip': 9.245427131652832, 'timing/data': 0.0065917037427425385, 'timing/model': 2.7757613211870193, 'learning_rate/qwen_vl_interface': 7.532990625833806e-06, 'learning_rate/action_model': 7.532990625833805e-05, 'epoch': 3.08}) +[2026-06-20 01:22:53,925][starVLA.training.train_starvla][INFO] - Step 1769, Loss: {'train/loss': 0.3255525827407837, 'train/grad_norm_pre_clip': 16.88981056213379, 'timing/data': 0.006615191698074341, 'timing/model': 2.7642061058431864, 'learning_rate/qwen_vl_interface': 7.5303195545335225e-06, 'learning_rate/action_model': 7.530319554533522e-05, 'epoch': 3.08}) +[2026-06-20 01:23:04,829][starVLA.training.train_starvla][INFO] - Step 1770, Loss: {'train/loss': 0.18610091507434845, 'train/grad_norm_pre_clip': 6.971606731414795, 'timing/data': 0.006909308955073357, 'timing/model': 2.772667208686471, 'learning_rate/qwen_vl_interface': 7.527647545880007e-06, 'learning_rate/action_model': 7.527647545880007e-05, 'epoch': 3.08}) +[2026-06-20 01:23:15,963][starVLA.training.train_starvla][INFO] - Step 1771, Loss: {'train/loss': 0.3209623098373413, 'train/grad_norm_pre_clip': 36.7432975769043, 'timing/data': 0.0065808724611997604, 'timing/model': 3.0126794166862965, 'learning_rate/qwen_vl_interface': 7.52497460097162e-06, 'learning_rate/action_model': 7.52497460097162e-05, 'epoch': 3.09}) +[2026-06-20 01:23:26,627][starVLA.training.train_starvla][INFO] - Step 1772, Loss: {'train/loss': 0.15145935118198395, 'train/grad_norm_pre_clip': 6.02217960357666, 'timing/data': 0.0065526943653821945, 'timing/model': 2.7428869493305683, 'learning_rate/qwen_vl_interface': 7.522300720907105e-06, 'learning_rate/action_model': 7.522300720907105e-05, 'epoch': 3.09}) +[2026-06-20 01:23:37,434][starVLA.training.train_starvla][INFO] - Step 1773, Loss: {'train/loss': 0.22968491911888123, 'train/grad_norm_pre_clip': 13.068802833557129, 'timing/data': 0.006929460912942886, 'timing/model': 2.783156981691718, 'learning_rate/qwen_vl_interface': 7.519625906785601e-06, 'learning_rate/action_model': 7.5196259067856e-05, 'epoch': 3.09}) +[2026-06-20 01:23:48,335][starVLA.training.train_starvla][INFO] - Step 1774, Loss: {'train/loss': 0.38060566782951355, 'train/grad_norm_pre_clip': 19.478322982788086, 'timing/data': 0.006626971065998077, 'timing/model': 2.7860806696116924, 'learning_rate/qwen_vl_interface': 7.516950159706614e-06, 'learning_rate/action_model': 7.516950159706614e-05, 'epoch': 3.09}) +[2026-06-20 01:23:59,236][starVLA.training.train_starvla][INFO] - Step 1775, Loss: {'train/loss': 0.41997259855270386, 'train/grad_norm_pre_clip': 13.977797508239746, 'timing/data': 0.006639812141656876, 'timing/model': 2.77411731146276, 'learning_rate/qwen_vl_interface': 7.514273480770047e-06, 'learning_rate/action_model': 7.514273480770047e-05, 'epoch': 3.09}) +[2026-06-20 01:24:10,140][starVLA.training.train_starvla][INFO] - Step 1776, Loss: {'train/loss': 0.24473704397678375, 'train/grad_norm_pre_clip': 10.735710144042969, 'timing/data': 0.006817679852247238, 'timing/model': 2.780557556077838, 'learning_rate/qwen_vl_interface': 7.511595871076181e-06, 'learning_rate/action_model': 7.511595871076181e-05, 'epoch': 3.09}) +[2026-06-20 01:24:21,042][starVLA.training.train_starvla][INFO] - Step 1777, Loss: {'train/loss': 0.3586130142211914, 'train/grad_norm_pre_clip': 22.644329071044922, 'timing/data': 0.006689403206110001, 'timing/model': 2.7883927430957556, 'learning_rate/qwen_vl_interface': 7.50891733172568e-06, 'learning_rate/action_model': 7.50891733172568e-05, 'epoch': 3.1}) +[2026-06-20 01:24:31,940][starVLA.training.train_starvla][INFO] - Step 1778, Loss: {'train/loss': 0.1917628049850464, 'train/grad_norm_pre_clip': 14.826385498046875, 'timing/data': 0.00708804652094841, 'timing/model': 2.777318501845002, 'learning_rate/qwen_vl_interface': 7.506237863819593e-06, 'learning_rate/action_model': 7.506237863819593e-05, 'epoch': 3.1}) +[2026-06-20 01:24:42,941][starVLA.training.train_starvla][INFO] - Step 1779, Loss: {'train/loss': 0.09853797405958176, 'train/grad_norm_pre_clip': 8.733010292053223, 'timing/data': 0.0069605764001607895, 'timing/model': 2.7615573927760124, 'learning_rate/qwen_vl_interface': 7.503557468459345e-06, 'learning_rate/action_model': 7.503557468459344e-05, 'epoch': 3.1}) +[2026-06-20 01:24:53,825][starVLA.training.train_starvla][INFO] - Step 1780, Loss: {'train/loss': 0.18400782346725464, 'train/grad_norm_pre_clip': 16.647275924682617, 'timing/data': 0.0067427996546030045, 'timing/model': 2.7776910960674286, 'learning_rate/qwen_vl_interface': 7.5008761467467435e-06, 'learning_rate/action_model': 7.500876146746743e-05, 'epoch': 3.1}) +[2026-06-20 01:25:04,731][starVLA.training.train_starvla][INFO] - Step 1781, Loss: {'train/loss': 0.26998135447502136, 'train/grad_norm_pre_clip': 10.766366004943848, 'timing/data': 0.0066310204565525055, 'timing/model': 2.784453922882676, 'learning_rate/qwen_vl_interface': 7.498193899783983e-06, 'learning_rate/action_model': 7.498193899783983e-05, 'epoch': 3.1}) +[2026-06-20 01:25:15,634][starVLA.training.train_starvla][INFO] - Step 1782, Loss: {'train/loss': 0.11397050321102142, 'train/grad_norm_pre_clip': 12.016782760620117, 'timing/data': 0.006581543013453484, 'timing/model': 2.7740036491304636, 'learning_rate/qwen_vl_interface': 7.495510728673631e-06, 'learning_rate/action_model': 7.495510728673632e-05, 'epoch': 3.1}) +[2026-06-20 01:25:26,533][starVLA.training.train_starvla][INFO] - Step 1783, Loss: {'train/loss': 0.2530721127986908, 'train/grad_norm_pre_clip': 14.933066368103027, 'timing/data': 0.006754299625754356, 'timing/model': 2.7766070924699306, 'learning_rate/qwen_vl_interface': 7.492826634518642e-06, 'learning_rate/action_model': 7.492826634518641e-05, 'epoch': 3.11}) +[2026-06-20 01:25:37,426][starVLA.training.train_starvla][INFO] - Step 1784, Loss: {'train/loss': 0.15633049607276917, 'train/grad_norm_pre_clip': 6.825434684753418, 'timing/data': 0.0067864470183849335, 'timing/model': 2.7786365784704685, 'learning_rate/qwen_vl_interface': 7.4901416184223404e-06, 'learning_rate/action_model': 7.49014161842234e-05, 'epoch': 3.11}) +[2026-06-20 01:25:48,335][starVLA.training.train_starvla][INFO] - Step 1785, Loss: {'train/loss': 0.20993183553218842, 'train/grad_norm_pre_clip': 12.269219398498535, 'timing/data': 0.006870400160551071, 'timing/model': 2.7773711513727903, 'learning_rate/qwen_vl_interface': 7.487455681488437e-06, 'learning_rate/action_model': 7.487455681488436e-05, 'epoch': 3.11}) +[2026-06-20 01:25:59,235][starVLA.training.train_starvla][INFO] - Step 1786, Loss: {'train/loss': 0.19265802204608917, 'train/grad_norm_pre_clip': 10.730772018432617, 'timing/data': 0.006458263844251633, 'timing/model': 2.7812960781157017, 'learning_rate/qwen_vl_interface': 7.484768824821018e-06, 'learning_rate/action_model': 7.484768824821018e-05, 'epoch': 3.11}) +[2026-06-20 01:26:10,336][starVLA.training.train_starvla][INFO] - Step 1787, Loss: {'train/loss': 0.2173989862203598, 'train/grad_norm_pre_clip': 9.39089298248291, 'timing/data': 0.006423423066735268, 'timing/model': 2.7900166176259518, 'learning_rate/qwen_vl_interface': 7.482081049524551e-06, 'learning_rate/action_model': 7.482081049524551e-05, 'epoch': 3.11}) +[2026-06-20 01:26:21,235][starVLA.training.train_starvla][INFO] - Step 1788, Loss: {'train/loss': 0.4838270843029022, 'train/grad_norm_pre_clip': 22.56813621520996, 'timing/data': 0.0065806955099105835, 'timing/model': 2.7766163647174835, 'learning_rate/qwen_vl_interface': 7.479392356703878e-06, 'learning_rate/action_model': 7.479392356703878e-05, 'epoch': 3.11}) +[2026-06-20 01:26:32,140][starVLA.training.train_starvla][INFO] - Step 1789, Loss: {'train/loss': 0.20732012391090393, 'train/grad_norm_pre_clip': 12.485930442810059, 'timing/data': 0.006647162139415741, 'timing/model': 2.7774885315448046, 'learning_rate/qwen_vl_interface': 7.476702747464219e-06, 'learning_rate/action_model': 7.476702747464219e-05, 'epoch': 3.12}) +[2026-06-20 01:26:43,037][starVLA.training.train_starvla][INFO] - Step 1790, Loss: {'train/loss': 0.11584552377462387, 'train/grad_norm_pre_clip': 6.583060264587402, 'timing/data': 0.006696011871099472, 'timing/model': 2.777367353439331, 'learning_rate/qwen_vl_interface': 7.474012222911169e-06, 'learning_rate/action_model': 7.474012222911169e-05, 'epoch': 3.12}) +[2026-06-20 01:26:53,931][starVLA.training.train_starvla][INFO] - Step 1791, Loss: {'train/loss': 0.20000998675823212, 'train/grad_norm_pre_clip': 13.517830848693848, 'timing/data': 0.006636790931224823, 'timing/model': 2.7750839442014694, 'learning_rate/qwen_vl_interface': 7.471320784150703e-06, 'learning_rate/action_model': 7.471320784150703e-05, 'epoch': 3.12}) +[2026-06-20 01:27:04,831][starVLA.training.train_starvla][INFO] - Step 1792, Loss: {'train/loss': 0.21613220870494843, 'train/grad_norm_pre_clip': 11.130853652954102, 'timing/data': 0.006783090531826019, 'timing/model': 2.7704835683107376, 'learning_rate/qwen_vl_interface': 7.468628432289169e-06, 'learning_rate/action_model': 7.468628432289168e-05, 'epoch': 3.12}) +[2026-06-20 01:27:15,724][starVLA.training.train_starvla][INFO] - Step 1793, Loss: {'train/loss': 0.17357318103313446, 'train/grad_norm_pre_clip': 11.68710994720459, 'timing/data': 0.006655832752585411, 'timing/model': 2.772850325331092, 'learning_rate/qwen_vl_interface': 7.465935168433292e-06, 'learning_rate/action_model': 7.465935168433292e-05, 'epoch': 3.12}) +[2026-06-20 01:27:26,769][starVLA.training.train_starvla][INFO] - Step 1794, Loss: {'train/loss': 0.13377682864665985, 'train/grad_norm_pre_clip': 8.789917945861816, 'timing/data': 0.006479984149336815, 'timing/model': 2.9129148460924625, 'learning_rate/qwen_vl_interface': 7.463240993690168e-06, 'learning_rate/action_model': 7.463240993690168e-05, 'epoch': 3.13}) +[2026-06-20 01:27:37,634][starVLA.training.train_starvla][INFO] - Step 1795, Loss: {'train/loss': 0.1728857308626175, 'train/grad_norm_pre_clip': 8.63211441040039, 'timing/data': 0.006788000464439392, 'timing/model': 2.7866989709436893, 'learning_rate/qwen_vl_interface': 7.460545909167276e-06, 'learning_rate/action_model': 7.460545909167275e-05, 'epoch': 3.13}) +[2026-06-20 01:27:48,529][starVLA.training.train_starvla][INFO] - Step 1796, Loss: {'train/loss': 0.13845278322696686, 'train/grad_norm_pre_clip': 8.778450012207031, 'timing/data': 0.006752749904990196, 'timing/model': 2.7797037698328495, 'learning_rate/qwen_vl_interface': 7.457849915972459e-06, 'learning_rate/action_model': 7.457849915972458e-05, 'epoch': 3.13}) +[2026-06-20 01:27:59,416][starVLA.training.train_starvla][INFO] - Step 1797, Loss: {'train/loss': 0.20028820633888245, 'train/grad_norm_pre_clip': 7.581376075744629, 'timing/data': 0.006686439737677574, 'timing/model': 2.763934990391135, 'learning_rate/qwen_vl_interface': 7.455153015213941e-06, 'learning_rate/action_model': 7.45515301521394e-05, 'epoch': 3.13}) +[2026-06-20 01:28:10,528][starVLA.training.train_starvla][INFO] - Step 1798, Loss: {'train/loss': 0.2522231638431549, 'train/grad_norm_pre_clip': 11.179845809936523, 'timing/data': 0.0068742986768484116, 'timing/model': 2.7426175996661186, 'learning_rate/qwen_vl_interface': 7.452455208000311e-06, 'learning_rate/action_model': 7.452455208000311e-05, 'epoch': 3.13}) +[2026-06-20 01:28:21,434][starVLA.training.train_starvla][INFO] - Step 1799, Loss: {'train/loss': 0.23265352845191956, 'train/grad_norm_pre_clip': 10.525537490844727, 'timing/data': 0.006742110475897789, 'timing/model': 2.7787639647722244, 'learning_rate/qwen_vl_interface': 7.449756495440541e-06, 'learning_rate/action_model': 7.44975649544054e-05, 'epoch': 3.13}) +[2026-06-20 01:28:32,328][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1800 +[2026-06-20 01:36:02,918][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1800 +[2026-06-20 01:36:02,919][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 1800 +[2026-06-20 01:44:09,620][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 1800 +[2026-06-20 01:44:09,622][starVLA.training.train_starvla][INFO] - Step 1800, Loss: {'train/loss': 0.12118784338235855, 'train/grad_norm_pre_clip': 8.165597915649414, 'eval/loss': 0.5819608569145203, 'eval/latency_0/loss': 0.5532543063163757, 'eval/latency_1/loss': 0.6144655346870422, 'eval/latency_2/loss': 0.1350693553686142, 'eval/latency_3/loss': 0.6277448534965515, 'eval/latency_4/loss': 1.0057079792022705, 'eval/flappy/loss': 0.5819608569145203, 'eval/flappy/latency_0/loss': 0.5532543063163757, 'eval/flappy/latency_1/loss': 0.6144655346870422, 'eval/flappy/latency_2/loss': 0.1350693553686142, 'eval/flappy/latency_3/loss': 0.6277448534965515, 'eval/flappy/latency_4/loss': 1.0057079792022705, 'eval/flap_precision': 0.8181818181818182, 'eval/flap_recall': 0.29454545454545455, 'eval/flap_f1': 0.43315508021390375, 'eval/flap_rate_model': 0.0275, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.43315508021390375, 'eval/latency_0/flap_precision': 0.8181818181818182, 'eval/latency_0/flap_recall': 0.29454545454545455, 'eval/latency_0/flap_f1': 0.43315508021390375, 'eval/latency_0/flap_rate_model': 0.0275, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.43315508021390375, 'timing/data': 0.007138947024941444, 'timing/model': 2.7751083448529243, 'learning_rate/qwen_vl_interface': 7.447056878643969e-06, 'learning_rate/action_model': 7.447056878643968e-05, 'epoch': 3.14}) +[2026-06-20 01:44:09,623][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_1800_state +[2026-06-20 01:44:09,652][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-20 01:44:15,406][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_1800_state/model.safetensors +[2026-06-20 01:44:29,615][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_1800_state/optimizer.bin +[2026-06-20 01:44:29,616][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_1800_state/sampler.bin +[2026-06-20 01:44:29,616][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_1800_state/sampler_1.bin +[2026-06-20 01:44:29,617][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_1800_state/random_states_0.pkl +[2026-06-20 01:44:29,742][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1800 +[2026-06-20 01:44:29,743][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-20 01:44:29,757][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-20 01:44:40,624][starVLA.training.train_starvla][INFO] - Step 1801, Loss: {'train/loss': 0.19023551046848297, 'train/grad_norm_pre_clip': 11.684358596801758, 'timing/data': 0.006743868812918663, 'timing/model': 2.776481119915843, 'learning_rate/qwen_vl_interface': 7.444356358720303e-06, 'learning_rate/action_model': 7.444356358720302e-05, 'epoch': 3.14}) +[2026-06-20 01:44:51,724][starVLA.training.train_starvla][INFO] - Step 1802, Loss: {'train/loss': 0.2363886684179306, 'train/grad_norm_pre_clip': 12.504449844360352, 'timing/data': 0.007227135822176933, 'timing/model': 2.729621522128582, 'learning_rate/qwen_vl_interface': 7.4416549367796255e-06, 'learning_rate/action_model': 7.441654936779625e-05, 'epoch': 3.14}) +[2026-06-20 01:45:02,625][starVLA.training.train_starvla][INFO] - Step 1803, Loss: {'train/loss': 0.34190985560417175, 'train/grad_norm_pre_clip': 16.882709503173828, 'timing/data': 0.006745390594005585, 'timing/model': 2.775274343788624, 'learning_rate/qwen_vl_interface': 7.4389526139323895e-06, 'learning_rate/action_model': 7.438952613932389e-05, 'epoch': 3.14}) +[2026-06-20 01:45:13,526][starVLA.training.train_starvla][INFO] - Step 1804, Loss: {'train/loss': 0.40901824831962585, 'train/grad_norm_pre_clip': 14.321038246154785, 'timing/data': 0.006933476775884628, 'timing/model': 2.7770624067634344, 'learning_rate/qwen_vl_interface': 7.436249391289419e-06, 'learning_rate/action_model': 7.436249391289418e-05, 'epoch': 3.14}) +[2026-06-20 01:45:24,424][starVLA.training.train_starvla][INFO] - Step 1805, Loss: {'train/loss': 0.17197586596012115, 'train/grad_norm_pre_clip': 14.668868064880371, 'timing/data': 0.006789378821849823, 'timing/model': 2.7804233003407717, 'learning_rate/qwen_vl_interface': 7.433545269961906e-06, 'learning_rate/action_model': 7.433545269961906e-05, 'epoch': 3.14}) +[2026-06-20 01:45:35,429][starVLA.training.train_starvla][INFO] - Step 1806, Loss: {'train/loss': 0.29166051745414734, 'train/grad_norm_pre_clip': 9.65592098236084, 'timing/data': 0.006711339578032494, 'timing/model': 2.7733054533600807, 'learning_rate/qwen_vl_interface': 7.430840251061411e-06, 'learning_rate/action_model': 7.430840251061411e-05, 'epoch': 3.15}) +[2026-06-20 01:45:46,333][starVLA.training.train_starvla][INFO] - Step 1807, Loss: {'train/loss': 0.15967531502246857, 'train/grad_norm_pre_clip': 9.211124420166016, 'timing/data': 0.0067754629999399185, 'timing/model': 2.784130172803998, 'learning_rate/qwen_vl_interface': 7.428134335699866e-06, 'learning_rate/action_model': 7.428134335699865e-05, 'epoch': 3.15}) +[2026-06-20 01:45:57,306][starVLA.training.train_starvla][INFO] - Step 1808, Loss: {'train/loss': 0.12399638444185257, 'train/grad_norm_pre_clip': 3.3081862926483154, 'timing/data': 0.015874339267611504, 'timing/model': 2.8271854501217604, 'learning_rate/qwen_vl_interface': 7.425427524989572e-06, 'learning_rate/action_model': 7.425427524989572e-05, 'epoch': 3.15}) +[2026-06-20 01:46:08,127][starVLA.training.train_starvla][INFO] - Step 1809, Loss: {'train/loss': 0.15910819172859192, 'train/grad_norm_pre_clip': 8.495495796203613, 'timing/data': 0.00687565840780735, 'timing/model': 2.7708818949759007, 'learning_rate/qwen_vl_interface': 7.422719820043194e-06, 'learning_rate/action_model': 7.422719820043193e-05, 'epoch': 3.15}) +[2026-06-20 01:46:19,026][starVLA.training.train_starvla][INFO] - Step 1810, Loss: {'train/loss': 0.2610308825969696, 'train/grad_norm_pre_clip': 9.829415321350098, 'timing/data': 0.007427385076880455, 'timing/model': 2.7698656674474478, 'learning_rate/qwen_vl_interface': 7.420011221973771e-06, 'learning_rate/action_model': 7.420011221973771e-05, 'epoch': 3.15}) +[2026-06-20 01:46:29,931][starVLA.training.train_starvla][INFO] - Step 1811, Loss: {'train/loss': 0.16318443417549133, 'train/grad_norm_pre_clip': 6.529964447021484, 'timing/data': 0.0068531278520822525, 'timing/model': 2.7767084687948227, 'learning_rate/qwen_vl_interface': 7.417301731894702e-06, 'learning_rate/action_model': 7.417301731894701e-05, 'epoch': 3.15}) +[2026-06-20 01:46:40,830][starVLA.training.train_starvla][INFO] - Step 1812, Loss: {'train/loss': 0.3270198404788971, 'train/grad_norm_pre_clip': 21.602800369262695, 'timing/data': 0.006715133786201477, 'timing/model': 2.774384232237935, 'learning_rate/qwen_vl_interface': 7.414591350919756e-06, 'learning_rate/action_model': 7.414591350919756e-05, 'epoch': 3.16}) +[2026-06-20 01:46:51,723][starVLA.training.train_starvla][INFO] - Step 1813, Loss: {'train/loss': 0.25245344638824463, 'train/grad_norm_pre_clip': 17.78245735168457, 'timing/data': 0.007078228518366814, 'timing/model': 2.765754349529743, 'learning_rate/qwen_vl_interface': 7.4118800801630715e-06, 'learning_rate/action_model': 7.411880080163072e-05, 'epoch': 3.16}) +[2026-06-20 01:47:02,831][starVLA.training.train_starvla][INFO] - Step 1814, Loss: {'train/loss': 0.38281115889549255, 'train/grad_norm_pre_clip': 13.656255722045898, 'timing/data': 0.007142307236790657, 'timing/model': 2.7754369117319584, 'learning_rate/qwen_vl_interface': 7.409167920739146e-06, 'learning_rate/action_model': 7.409167920739147e-05, 'epoch': 3.16}) +[2026-06-20 01:47:13,734][starVLA.training.train_starvla][INFO] - Step 1815, Loss: {'train/loss': 0.26822587847709656, 'train/grad_norm_pre_clip': 6.73706579208374, 'timing/data': 0.0067916084080934525, 'timing/model': 2.7774577010422945, 'learning_rate/qwen_vl_interface': 7.406454873762848e-06, 'learning_rate/action_model': 7.406454873762848e-05, 'epoch': 3.16}) +[2026-06-20 01:47:24,624][starVLA.training.train_starvla][INFO] - Step 1816, Loss: {'train/loss': 0.16344623267650604, 'train/grad_norm_pre_clip': 9.54320240020752, 'timing/data': 0.00932200439274311, 'timing/model': 2.763054648414254, 'learning_rate/qwen_vl_interface': 7.403740940349408e-06, 'learning_rate/action_model': 7.403740940349408e-05, 'epoch': 3.16}) +[2026-06-20 01:47:35,672][starVLA.training.train_starvla][INFO] - Step 1817, Loss: {'train/loss': 0.4211585223674774, 'train/grad_norm_pre_clip': 28.986560821533203, 'timing/data': 0.006927847862243652, 'timing/model': 2.915982685983181, 'learning_rate/qwen_vl_interface': 7.401026121614421e-06, 'learning_rate/action_model': 7.40102612161442e-05, 'epoch': 3.17}) +[2026-06-20 01:47:46,521][starVLA.training.train_starvla][INFO] - Step 1818, Loss: {'train/loss': 0.19267120957374573, 'train/grad_norm_pre_clip': 5.9209089279174805, 'timing/data': 0.006713312119245529, 'timing/model': 2.766648329794407, 'learning_rate/qwen_vl_interface': 7.398310418673846e-06, 'learning_rate/action_model': 7.398310418673846e-05, 'epoch': 3.17}) +[2026-06-20 01:47:57,435][starVLA.training.train_starvla][INFO] - Step 1819, Loss: {'train/loss': 0.31797805428504944, 'train/grad_norm_pre_clip': 9.422690391540527, 'timing/data': 0.006715072318911552, 'timing/model': 2.777576047927141, 'learning_rate/qwen_vl_interface': 7.395593832644007e-06, 'learning_rate/action_model': 7.395593832644007e-05, 'epoch': 3.17}) +[2026-06-20 01:48:08,337][starVLA.training.train_starvla][INFO] - Step 1820, Loss: {'train/loss': 0.3539661765098572, 'train/grad_norm_pre_clip': 7.759190559387207, 'timing/data': 0.007623100653290749, 'timing/model': 2.7743043415248394, 'learning_rate/qwen_vl_interface': 7.39287636464159e-06, 'learning_rate/action_model': 7.39287636464159e-05, 'epoch': 3.17}) +[2026-06-20 01:48:19,084][starVLA.training.train_starvla][INFO] - Step 1821, Loss: {'train/loss': 0.09523077309131622, 'train/grad_norm_pre_clip': 4.036437511444092, 'timing/data': 0.006804430857300758, 'timing/model': 2.753251239657402, 'learning_rate/qwen_vl_interface': 7.3901580157836425e-06, 'learning_rate/action_model': 7.390158015783643e-05, 'epoch': 3.17}) +[2026-06-20 01:48:29,923][starVLA.training.train_starvla][INFO] - Step 1822, Loss: {'train/loss': 0.16786201298236847, 'train/grad_norm_pre_clip': 7.395824909210205, 'timing/data': 0.006853299215435982, 'timing/model': 2.7742943186312914, 'learning_rate/qwen_vl_interface': 7.387438787187576e-06, 'learning_rate/action_model': 7.387438787187576e-05, 'epoch': 3.17}) +[2026-06-20 01:48:40,643][starVLA.training.train_starvla][INFO] - Step 1823, Loss: {'train/loss': 0.1856360137462616, 'train/grad_norm_pre_clip': 9.144452095031738, 'timing/data': 0.0066452231258153915, 'timing/model': 2.7235597502440214, 'learning_rate/qwen_vl_interface': 7.3847186799711625e-06, 'learning_rate/action_model': 7.384718679971162e-05, 'epoch': 3.18}) +[2026-06-20 01:48:51,537][starVLA.training.train_starvla][INFO] - Step 1824, Loss: {'train/loss': 0.20392902195453644, 'train/grad_norm_pre_clip': 21.523235321044922, 'timing/data': 0.010137215256690979, 'timing/model': 2.7831526808440685, 'learning_rate/qwen_vl_interface': 7.381997695252535e-06, 'learning_rate/action_model': 7.381997695252535e-05, 'epoch': 3.18}) +[2026-06-20 01:49:02,262][starVLA.training.train_starvla][INFO] - Step 1825, Loss: {'train/loss': 0.2039998173713684, 'train/grad_norm_pre_clip': 8.485487937927246, 'timing/data': 0.008393602445721626, 'timing/model': 2.7164078764617443, 'learning_rate/qwen_vl_interface': 7.379275834150186e-06, 'learning_rate/action_model': 7.379275834150186e-05, 'epoch': 3.18}) +[2026-06-20 01:49:13,006][starVLA.training.train_starvla][INFO] - Step 1826, Loss: {'train/loss': 0.12214502692222595, 'train/grad_norm_pre_clip': 7.1926984786987305, 'timing/data': 0.006517251953482628, 'timing/model': 2.760243324562907, 'learning_rate/qwen_vl_interface': 7.376553097782972e-06, 'learning_rate/action_model': 7.376553097782972e-05, 'epoch': 3.18}) +[2026-06-20 01:49:23,727][starVLA.training.train_starvla][INFO] - Step 1827, Loss: {'train/loss': 0.24053773283958435, 'train/grad_norm_pre_clip': 7.137540340423584, 'timing/data': 0.006544651463627815, 'timing/model': 2.761100707575679, 'learning_rate/qwen_vl_interface': 7.373829487270109e-06, 'learning_rate/action_model': 7.373829487270109e-05, 'epoch': 3.18}) +[2026-06-20 01:49:34,553][starVLA.training.train_starvla][INFO] - Step 1828, Loss: {'train/loss': 0.24108527600765228, 'train/grad_norm_pre_clip': 8.787053108215332, 'timing/data': 0.00670391321182251, 'timing/model': 2.729300055652857, 'learning_rate/qwen_vl_interface': 7.371105003731165e-06, 'learning_rate/action_model': 7.371105003731164e-05, 'epoch': 3.18}) +[2026-06-20 01:49:45,536][starVLA.training.train_starvla][INFO] - Step 1829, Loss: {'train/loss': 0.12174822390079498, 'train/grad_norm_pre_clip': 5.869236469268799, 'timing/data': 0.006823640316724777, 'timing/model': 2.7728549670428038, 'learning_rate/qwen_vl_interface': 7.3683796482860755e-06, 'learning_rate/action_model': 7.368379648286075e-05, 'epoch': 3.19}) +[2026-06-20 01:49:56,438][starVLA.training.train_starvla][INFO] - Step 1830, Loss: {'train/loss': 0.21673721075057983, 'train/grad_norm_pre_clip': 8.337907791137695, 'timing/data': 0.0067545827478170395, 'timing/model': 2.7764096278697252, 'learning_rate/qwen_vl_interface': 7.365653422055131e-06, 'learning_rate/action_model': 7.365653422055131e-05, 'epoch': 3.19}) +[2026-06-20 01:50:07,256][starVLA.training.train_starvla][INFO] - Step 1831, Loss: {'train/loss': 0.2685593366622925, 'train/grad_norm_pre_clip': 11.717620849609375, 'timing/data': 0.0065520331263542175, 'timing/model': 2.698334952816367, 'learning_rate/qwen_vl_interface': 7.3629263261589816e-06, 'learning_rate/action_model': 7.362926326158981e-05, 'epoch': 3.19}) +[2026-06-20 01:50:18,140][starVLA.training.train_starvla][INFO] - Step 1832, Loss: {'train/loss': 0.4010348618030548, 'train/grad_norm_pre_clip': 12.04171085357666, 'timing/data': 0.006560411304235458, 'timing/model': 2.801247203722596, 'learning_rate/qwen_vl_interface': 7.36019836171863e-06, 'learning_rate/action_model': 7.360198361718629e-05, 'epoch': 3.19}) +[2026-06-20 01:50:29,038][starVLA.training.train_starvla][INFO] - Step 1833, Loss: {'train/loss': 0.18135115504264832, 'train/grad_norm_pre_clip': 13.121712684631348, 'timing/data': 0.006605964154005051, 'timing/model': 2.7799534127116203, 'learning_rate/qwen_vl_interface': 7.357469529855441e-06, 'learning_rate/action_model': 7.35746952985544e-05, 'epoch': 3.19}) +[2026-06-20 01:50:39,936][starVLA.training.train_starvla][INFO] - Step 1834, Loss: {'train/loss': 0.3910203278064728, 'train/grad_norm_pre_clip': 20.757274627685547, 'timing/data': 0.006544243544340134, 'timing/model': 2.7704428900033236, 'learning_rate/qwen_vl_interface': 7.354739831691134e-06, 'learning_rate/action_model': 7.354739831691135e-05, 'epoch': 3.19}) +[2026-06-20 01:50:50,839][starVLA.training.train_starvla][INFO] - Step 1835, Loss: {'train/loss': 0.1999765783548355, 'train/grad_norm_pre_clip': 8.327908515930176, 'timing/data': 0.006585359573364258, 'timing/model': 2.784641943871975, 'learning_rate/qwen_vl_interface': 7.352009268347786e-06, 'learning_rate/action_model': 7.352009268347786e-05, 'epoch': 3.2}) +[2026-06-20 01:51:01,638][starVLA.training.train_starvla][INFO] - Step 1836, Loss: {'train/loss': 0.22138577699661255, 'train/grad_norm_pre_clip': 15.362008094787598, 'timing/data': 0.0062406230717897415, 'timing/model': 2.7787781171500683, 'learning_rate/qwen_vl_interface': 7.349277840947828e-06, 'learning_rate/action_model': 7.349277840947828e-05, 'epoch': 3.2}) +[2026-06-20 01:51:12,728][starVLA.training.train_starvla][INFO] - Step 1837, Loss: {'train/loss': 0.2320849597454071, 'train/grad_norm_pre_clip': 15.59887981414795, 'timing/data': 0.006730282679200172, 'timing/model': 2.7754253670573235, 'learning_rate/qwen_vl_interface': 7.346545550614046e-06, 'learning_rate/action_model': 7.346545550614046e-05, 'epoch': 3.2}) +[2026-06-20 01:51:23,425][starVLA.training.train_starvla][INFO] - Step 1838, Loss: {'train/loss': 0.25977960228919983, 'train/grad_norm_pre_clip': 15.742077827453613, 'timing/data': 0.005960388109087944, 'timing/model': 2.747914122417569, 'learning_rate/qwen_vl_interface': 7.343812398469582e-06, 'learning_rate/action_model': 7.343812398469581e-05, 'epoch': 3.2}) +[2026-06-20 01:51:34,522][starVLA.training.train_starvla][INFO] - Step 1839, Loss: {'train/loss': 0.15384802222251892, 'train/grad_norm_pre_clip': 5.252108573913574, 'timing/data': 0.006628651171922684, 'timing/model': 2.768367800861597, 'learning_rate/qwen_vl_interface': 7.34107838563793e-06, 'learning_rate/action_model': 7.341078385637929e-05, 'epoch': 3.2}) +[2026-06-20 01:51:45,437][starVLA.training.train_starvla][INFO] - Step 1840, Loss: {'train/loss': 0.18434035778045654, 'train/grad_norm_pre_clip': 8.601783752441406, 'timing/data': 0.006909798830747604, 'timing/model': 2.7775984462350607, 'learning_rate/qwen_vl_interface': 7.338343513242941e-06, 'learning_rate/action_model': 7.338343513242941e-05, 'epoch': 3.21}) +[2026-06-20 01:51:56,316][starVLA.training.train_starvla][INFO] - Step 1841, Loss: {'train/loss': 0.1911722868680954, 'train/grad_norm_pre_clip': 11.305957794189453, 'timing/data': 0.006717221811413765, 'timing/model': 2.755534114316106, 'learning_rate/qwen_vl_interface': 7.335607782408818e-06, 'learning_rate/action_model': 7.335607782408818e-05, 'epoch': 3.21}) +[2026-06-20 01:52:07,126][starVLA.training.train_starvla][INFO] - Step 1842, Loss: {'train/loss': 0.15856611728668213, 'train/grad_norm_pre_clip': 18.97975730895996, 'timing/data': 0.006112877279520035, 'timing/model': 2.76954210922122, 'learning_rate/qwen_vl_interface': 7.3328711942601145e-06, 'learning_rate/action_model': 7.332871194260115e-05, 'epoch': 3.21}) +[2026-06-20 01:52:18,035][starVLA.training.train_starvla][INFO] - Step 1843, Loss: {'train/loss': 0.1744503378868103, 'train/grad_norm_pre_clip': 6.73590087890625, 'timing/data': 0.0065455809235572815, 'timing/model': 2.7737288866192102, 'learning_rate/qwen_vl_interface': 7.33013374992174e-06, 'learning_rate/action_model': 7.33013374992174e-05, 'epoch': 3.21}) +[2026-06-20 01:52:29,070][starVLA.training.train_starvla][INFO] - Step 1844, Loss: {'train/loss': 0.12216203659772873, 'train/grad_norm_pre_clip': 18.49892234802246, 'timing/data': 0.006646938621997833, 'timing/model': 2.908642878755927, 'learning_rate/qwen_vl_interface': 7.327395450518956e-06, 'learning_rate/action_model': 7.327395450518956e-05, 'epoch': 3.21}) +[2026-06-20 01:52:39,926][starVLA.training.train_starvla][INFO] - Step 1845, Loss: {'train/loss': 0.2695045471191406, 'train/grad_norm_pre_clip': 16.845197677612305, 'timing/data': 0.006784131750464439, 'timing/model': 2.766638819128275, 'learning_rate/qwen_vl_interface': 7.324656297177372e-06, 'learning_rate/action_model': 7.324656297177371e-05, 'epoch': 3.21}) +[2026-06-20 01:52:50,837][starVLA.training.train_starvla][INFO] - Step 1846, Loss: {'train/loss': 0.21494518220424652, 'train/grad_norm_pre_clip': 9.759364128112793, 'timing/data': 0.006853310391306877, 'timing/model': 2.78002374060452, 'learning_rate/qwen_vl_interface': 7.32191629102295e-06, 'learning_rate/action_model': 7.32191629102295e-05, 'epoch': 3.22}) +[2026-06-20 01:53:01,734][starVLA.training.train_starvla][INFO] - Step 1847, Loss: {'train/loss': 0.14963901042938232, 'train/grad_norm_pre_clip': 18.973682403564453, 'timing/data': 0.006594853475689888, 'timing/model': 2.7829019520431757, 'learning_rate/qwen_vl_interface': 7.319175433182004e-06, 'learning_rate/action_model': 7.319175433182004e-05, 'epoch': 3.22}) +[2026-06-20 01:53:12,835][starVLA.training.train_starvla][INFO] - Step 1848, Loss: {'train/loss': 0.17667417228221893, 'train/grad_norm_pre_clip': 7.9512858390808105, 'timing/data': 0.006736528128385544, 'timing/model': 2.7556301206350327, 'learning_rate/qwen_vl_interface': 7.316433724781198e-06, 'learning_rate/action_model': 7.316433724781197e-05, 'epoch': 3.22}) +[2026-06-20 01:53:23,734][starVLA.training.train_starvla][INFO] - Step 1849, Loss: {'train/loss': 0.1892988085746765, 'train/grad_norm_pre_clip': 16.22827911376953, 'timing/data': 0.006622312590479851, 'timing/model': 2.786240851506591, 'learning_rate/qwen_vl_interface': 7.313691166947543e-06, 'learning_rate/action_model': 7.313691166947543e-05, 'epoch': 3.22}) +[2026-06-20 01:53:34,636][starVLA.training.train_starvla][INFO] - Step 1850, Loss: {'train/loss': 0.24267347157001495, 'train/grad_norm_pre_clip': 12.148523330688477, 'timing/data': 0.0066461581736803055, 'timing/model': 2.776886433362961, 'learning_rate/qwen_vl_interface': 7.310947760808401e-06, 'learning_rate/action_model': 7.3109477608084e-05, 'epoch': 3.22}) +[2026-06-20 01:53:45,535][starVLA.training.train_starvla][INFO] - Step 1851, Loss: {'train/loss': 0.27133315801620483, 'train/grad_norm_pre_clip': 15.807448387145996, 'timing/data': 0.006838735193014145, 'timing/model': 2.7706042788922787, 'learning_rate/qwen_vl_interface': 7.308203507491485e-06, 'learning_rate/action_model': 7.308203507491484e-05, 'epoch': 3.22}) +[2026-06-20 01:53:56,637][starVLA.training.train_starvla][INFO] - Step 1852, Loss: {'train/loss': 0.20136238634586334, 'train/grad_norm_pre_clip': 35.117610931396484, 'timing/data': 0.006668616086244583, 'timing/model': 2.7773511316627264, 'learning_rate/qwen_vl_interface': 7.3054584081248515e-06, 'learning_rate/action_model': 7.305458408124852e-05, 'epoch': 3.23}) +[2026-06-20 01:54:07,536][starVLA.training.train_starvla][INFO] - Step 1853, Loss: {'train/loss': 0.1279653012752533, 'train/grad_norm_pre_clip': 11.715038299560547, 'timing/data': 0.006961703300476074, 'timing/model': 2.7745915800333023, 'learning_rate/qwen_vl_interface': 7.302712463836907e-06, 'learning_rate/action_model': 7.302712463836907e-05, 'epoch': 3.23}) +[2026-06-20 01:54:18,434][starVLA.training.train_starvla][INFO] - Step 1854, Loss: {'train/loss': 0.16916462779045105, 'train/grad_norm_pre_clip': 10.225646018981934, 'timing/data': 0.006944412365555763, 'timing/model': 2.7748897075653076, 'learning_rate/qwen_vl_interface': 7.2999656757564086e-06, 'learning_rate/action_model': 7.299965675756409e-05, 'epoch': 3.23}) +[2026-06-20 01:54:29,334][starVLA.training.train_starvla][INFO] - Step 1855, Loss: {'train/loss': 0.17904230952262878, 'train/grad_norm_pre_clip': 13.840543746948242, 'timing/data': 0.006817067041993141, 'timing/model': 2.7753479331731796, 'learning_rate/qwen_vl_interface': 7.297218045012453e-06, 'learning_rate/action_model': 7.297218045012453e-05, 'epoch': 3.23}) +[2026-06-20 01:54:40,237][starVLA.training.train_starvla][INFO] - Step 1856, Loss: {'train/loss': 0.08987747877836227, 'train/grad_norm_pre_clip': 8.44717025756836, 'timing/data': 0.0063672177493572235, 'timing/model': 2.772938009351492, 'learning_rate/qwen_vl_interface': 7.29446957273449e-06, 'learning_rate/action_model': 7.294469572734489e-05, 'epoch': 3.23}) +[2026-06-20 01:54:51,126][starVLA.training.train_starvla][INFO] - Step 1857, Loss: {'train/loss': 0.2420014590024948, 'train/grad_norm_pre_clip': 15.1500825881958, 'timing/data': 0.006763394922018051, 'timing/model': 2.7658938094973564, 'learning_rate/qwen_vl_interface': 7.2917202600523104e-06, 'learning_rate/action_model': 7.29172026005231e-05, 'epoch': 3.24}) +[2026-06-20 01:55:02,001][starVLA.training.train_starvla][INFO] - Step 1858, Loss: {'train/loss': 0.1679040491580963, 'train/grad_norm_pre_clip': 18.33736228942871, 'timing/data': 0.006962785497307777, 'timing/model': 2.744438460096717, 'learning_rate/qwen_vl_interface': 7.2889701080960565e-06, 'learning_rate/action_model': 7.288970108096056e-05, 'epoch': 3.24}) +[2026-06-20 01:55:12,841][starVLA.training.train_starvla][INFO] - Step 1859, Loss: {'train/loss': 0.19641955196857452, 'train/grad_norm_pre_clip': 11.124974250793457, 'timing/data': 0.0064484477043151855, 'timing/model': 2.7921394277364016, 'learning_rate/qwen_vl_interface': 7.286219117996209e-06, 'learning_rate/action_model': 7.286219117996207e-05, 'epoch': 3.24}) +[2026-06-20 01:55:23,737][starVLA.training.train_starvla][INFO] - Step 1860, Loss: {'train/loss': 0.16388577222824097, 'train/grad_norm_pre_clip': 13.344654083251953, 'timing/data': 0.006766596809029579, 'timing/model': 2.779392495751381, 'learning_rate/qwen_vl_interface': 7.283467290883598e-06, 'learning_rate/action_model': 7.283467290883598e-05, 'epoch': 3.24}) +[2026-06-20 01:55:34,628][starVLA.training.train_starvla][INFO] - Step 1861, Loss: {'train/loss': 0.3207511901855469, 'train/grad_norm_pre_clip': 19.593294143676758, 'timing/data': 0.006889475509524345, 'timing/model': 2.775252550840378, 'learning_rate/qwen_vl_interface': 7.2807146278893925e-06, 'learning_rate/action_model': 7.280714627889393e-05, 'epoch': 3.24}) +[2026-06-20 01:55:45,532][starVLA.training.train_starvla][INFO] - Step 1862, Loss: {'train/loss': 0.15924878418445587, 'train/grad_norm_pre_clip': 10.627544403076172, 'timing/data': 0.006509607657790184, 'timing/model': 2.772675411775708, 'learning_rate/qwen_vl_interface': 7.277961130145111e-06, 'learning_rate/action_model': 7.277961130145111e-05, 'epoch': 3.24}) +[2026-06-20 01:55:56,422][starVLA.training.train_starvla][INFO] - Step 1863, Loss: {'train/loss': 0.28060582280158997, 'train/grad_norm_pre_clip': 18.46246337890625, 'timing/data': 0.0067817047238349915, 'timing/model': 2.7649637330323458, 'learning_rate/qwen_vl_interface': 7.2752067987826116e-06, 'learning_rate/action_model': 7.275206798782611e-05, 'epoch': 3.25}) +[2026-06-20 01:56:07,335][starVLA.training.train_starvla][INFO] - Step 1864, Loss: {'train/loss': 0.22920498251914978, 'train/grad_norm_pre_clip': 12.995610237121582, 'timing/data': 0.006889022886753082, 'timing/model': 2.7725104615092278, 'learning_rate/qwen_vl_interface': 7.272451634934096e-06, 'learning_rate/action_model': 7.272451634934096e-05, 'epoch': 3.25}) +[2026-06-20 01:56:18,234][starVLA.training.train_starvla][INFO] - Step 1865, Loss: {'train/loss': 0.31030601263046265, 'train/grad_norm_pre_clip': 15.22749137878418, 'timing/data': 0.006913373246788979, 'timing/model': 2.773431707173586, 'learning_rate/qwen_vl_interface': 7.269695639732108e-06, 'learning_rate/action_model': 7.269695639732108e-05, 'epoch': 3.25}) +[2026-06-20 01:56:29,135][starVLA.training.train_starvla][INFO] - Step 1866, Loss: {'train/loss': 0.1580737680196762, 'train/grad_norm_pre_clip': 10.181836128234863, 'timing/data': 0.006584446877241135, 'timing/model': 2.774770924821496, 'learning_rate/qwen_vl_interface': 7.266938814309533e-06, 'learning_rate/action_model': 7.266938814309533e-05, 'epoch': 3.25}) +[2026-06-20 01:56:40,051][starVLA.training.train_starvla][INFO] - Step 1867, Loss: {'train/loss': 0.19796624779701233, 'train/grad_norm_pre_clip': 10.035213470458984, 'timing/data': 0.006815474480390549, 'timing/model': 2.722898168489337, 'learning_rate/qwen_vl_interface': 7.264181159799598e-06, 'learning_rate/action_model': 7.264181159799597e-05, 'epoch': 3.25}) +[2026-06-20 01:56:50,935][starVLA.training.train_starvla][INFO] - Step 1868, Loss: {'train/loss': 0.09046632051467896, 'train/grad_norm_pre_clip': 4.978296279907227, 'timing/data': 0.0069765932857990265, 'timing/model': 2.775930354371667, 'learning_rate/qwen_vl_interface': 7.261422677335871e-06, 'learning_rate/action_model': 7.261422677335871e-05, 'epoch': 3.25}) +[2026-06-20 01:57:01,833][starVLA.training.train_starvla][INFO] - Step 1869, Loss: {'train/loss': 0.23864111304283142, 'train/grad_norm_pre_clip': 9.6336669921875, 'timing/data': 0.006790323182940483, 'timing/model': 2.7791254967451096, 'learning_rate/qwen_vl_interface': 7.258663368052259e-06, 'learning_rate/action_model': 7.258663368052259e-05, 'epoch': 3.26}) +[2026-06-20 01:57:12,736][starVLA.training.train_starvla][INFO] - Step 1870, Loss: {'train/loss': 0.17112767696380615, 'train/grad_norm_pre_clip': 7.923403263092041, 'timing/data': 0.006620364263653755, 'timing/model': 2.7819322794675827, 'learning_rate/qwen_vl_interface': 7.2559032330830105e-06, 'learning_rate/action_model': 7.25590323308301e-05, 'epoch': 3.26}) +[2026-06-20 01:57:23,632][starVLA.training.train_starvla][INFO] - Step 1871, Loss: {'train/loss': 0.4450521171092987, 'train/grad_norm_pre_clip': 24.86465835571289, 'timing/data': 0.0071138907223939896, 'timing/model': 2.774232616648078, 'learning_rate/qwen_vl_interface': 7.253142273562712e-06, 'learning_rate/action_model': 7.253142273562712e-05, 'epoch': 3.26}) +[2026-06-20 01:57:34,534][starVLA.training.train_starvla][INFO] - Step 1872, Loss: {'train/loss': 0.13248570263385773, 'train/grad_norm_pre_clip': 5.90959358215332, 'timing/data': 0.00663374550640583, 'timing/model': 2.773607697337866, 'learning_rate/qwen_vl_interface': 7.25038049062629e-06, 'learning_rate/action_model': 7.250380490626289e-05, 'epoch': 3.26}) +[2026-06-20 01:57:45,428][starVLA.training.train_starvla][INFO] - Step 1873, Loss: {'train/loss': 0.25037476420402527, 'train/grad_norm_pre_clip': 7.540397644042969, 'timing/data': 0.006361939013004303, 'timing/model': 2.76522359251976, 'learning_rate/qwen_vl_interface': 7.24761788540901e-06, 'learning_rate/action_model': 7.24761788540901e-05, 'epoch': 3.26}) +[2026-06-20 01:57:56,335][starVLA.training.train_starvla][INFO] - Step 1874, Loss: {'train/loss': 0.20906133949756622, 'train/grad_norm_pre_clip': 10.380157470703125, 'timing/data': 0.006473947316408157, 'timing/model': 2.777367888018489, 'learning_rate/qwen_vl_interface': 7.244854459046473e-06, 'learning_rate/action_model': 7.244854459046473e-05, 'epoch': 3.26}) +[2026-06-20 01:58:07,425][starVLA.training.train_starvla][INFO] - Step 1875, Loss: {'train/loss': 0.2889327108860016, 'train/grad_norm_pre_clip': 18.819190979003906, 'timing/data': 0.0066422149538993835, 'timing/model': 2.7688572853803635, 'learning_rate/qwen_vl_interface': 7.24209021267462e-06, 'learning_rate/action_model': 7.24209021267462e-05, 'epoch': 3.27}) +[2026-06-20 01:58:18,331][starVLA.training.train_starvla][INFO] - Step 1876, Loss: {'train/loss': 0.20949052274227142, 'train/grad_norm_pre_clip': 6.033623695373535, 'timing/data': 0.006840642541646957, 'timing/model': 2.7754224985837936, 'learning_rate/qwen_vl_interface': 7.23932514742973e-06, 'learning_rate/action_model': 7.23932514742973e-05, 'epoch': 3.27}) +[2026-06-20 01:58:29,235][starVLA.training.train_starvla][INFO] - Step 1877, Loss: {'train/loss': 0.31284210085868835, 'train/grad_norm_pre_clip': 21.595678329467773, 'timing/data': 0.006394959986209869, 'timing/model': 2.7802314646542072, 'learning_rate/qwen_vl_interface': 7.236559264448413e-06, 'learning_rate/action_model': 7.236559264448412e-05, 'epoch': 3.27}) +[2026-06-20 01:58:40,134][starVLA.training.train_starvla][INFO] - Step 1878, Loss: {'train/loss': 0.31959396600723267, 'train/grad_norm_pre_clip': 17.40937614440918, 'timing/data': 0.006676435470581055, 'timing/model': 2.7832358311861753, 'learning_rate/qwen_vl_interface': 7.233792564867623e-06, 'learning_rate/action_model': 7.233792564867623e-05, 'epoch': 3.27}) +[2026-06-20 01:58:51,037][starVLA.training.train_starvla][INFO] - Step 1879, Loss: {'train/loss': 0.28462547063827515, 'train/grad_norm_pre_clip': 11.54720401763916, 'timing/data': 0.006684105843305588, 'timing/model': 2.781349955126643, 'learning_rate/qwen_vl_interface': 7.231025049824644e-06, 'learning_rate/action_model': 7.231025049824643e-05, 'epoch': 3.27}) +[2026-06-20 01:59:01,935][starVLA.training.train_starvla][INFO] - Step 1880, Loss: {'train/loss': 0.12719260156154633, 'train/grad_norm_pre_clip': 5.9788103103637695, 'timing/data': 0.006453877314925194, 'timing/model': 2.7794814128428698, 'learning_rate/qwen_vl_interface': 7.228256720457097e-06, 'learning_rate/action_model': 7.228256720457096e-05, 'epoch': 3.28}) +[2026-06-20 01:59:12,832][starVLA.training.train_starvla][INFO] - Step 1881, Loss: {'train/loss': 0.15855064988136292, 'train/grad_norm_pre_clip': 12.152239799499512, 'timing/data': 0.006861051544547081, 'timing/model': 2.7855267506092787, 'learning_rate/qwen_vl_interface': 7.225487577902937e-06, 'learning_rate/action_model': 7.225487577902936e-05, 'epoch': 3.28}) +[2026-06-20 01:59:23,722][starVLA.training.train_starvla][INFO] - Step 1882, Loss: {'train/loss': 0.2331591546535492, 'train/grad_norm_pre_clip': 12.066515922546387, 'timing/data': 0.006766434758901596, 'timing/model': 2.7673808578401804, 'learning_rate/qwen_vl_interface': 7.222717623300455e-06, 'learning_rate/action_model': 7.222717623300455e-05, 'epoch': 3.28}) +[2026-06-20 01:59:34,724][starVLA.training.train_starvla][INFO] - Step 1883, Loss: {'train/loss': 0.13097141683101654, 'train/grad_norm_pre_clip': 10.959005355834961, 'timing/data': 0.006712734699249268, 'timing/model': 2.778419667854905, 'learning_rate/qwen_vl_interface': 7.219946857788273e-06, 'learning_rate/action_model': 7.219946857788272e-05, 'epoch': 3.28}) +[2026-06-20 01:59:45,631][starVLA.training.train_starvla][INFO] - Step 1884, Loss: {'train/loss': 0.13363485038280487, 'train/grad_norm_pre_clip': 8.034001350402832, 'timing/data': 0.006743812933564186, 'timing/model': 2.7812867760658264, 'learning_rate/qwen_vl_interface': 7.21717528250535e-06, 'learning_rate/action_model': 7.21717528250535e-05, 'epoch': 3.28}) +[2026-06-20 01:59:56,532][starVLA.training.train_starvla][INFO] - Step 1885, Loss: {'train/loss': 0.3626720607280731, 'train/grad_norm_pre_clip': 17.891191482543945, 'timing/data': 0.006653934717178345, 'timing/model': 2.7865680251270533, 'learning_rate/qwen_vl_interface': 7.214402898590974e-06, 'learning_rate/action_model': 7.214402898590973e-05, 'epoch': 3.28}) +[2026-06-20 02:00:07,431][starVLA.training.train_starvla][INFO] - Step 1886, Loss: {'train/loss': 0.15875791013240814, 'train/grad_norm_pre_clip': 8.294260025024414, 'timing/data': 0.006826063618063927, 'timing/model': 2.7836433444172144, 'learning_rate/qwen_vl_interface': 7.211629707184768e-06, 'learning_rate/action_model': 7.211629707184768e-05, 'epoch': 3.29}) +[2026-06-20 02:00:18,332][starVLA.training.train_starvla][INFO] - Step 1887, Loss: {'train/loss': 0.18867869675159454, 'train/grad_norm_pre_clip': 6.483005523681641, 'timing/data': 0.0068656522780656815, 'timing/model': 2.783139578998089, 'learning_rate/qwen_vl_interface': 7.2088557094266885e-06, 'learning_rate/action_model': 7.208855709426688e-05, 'epoch': 3.29}) +[2026-06-20 02:00:29,230][starVLA.training.train_starvla][INFO] - Step 1888, Loss: {'train/loss': 0.24298487603664398, 'train/grad_norm_pre_clip': 11.407617568969727, 'timing/data': 0.006862511858344078, 'timing/model': 2.772513296455145, 'learning_rate/qwen_vl_interface': 7.2060809064570195e-06, 'learning_rate/action_model': 7.20608090645702e-05, 'epoch': 3.29}) +[2026-06-20 02:00:40,116][starVLA.training.train_starvla][INFO] - Step 1889, Loss: {'train/loss': 0.12751798331737518, 'train/grad_norm_pre_clip': 8.447094917297363, 'timing/data': 0.00666443444788456, 'timing/model': 2.7606726326048374, 'learning_rate/qwen_vl_interface': 7.203305299416375e-06, 'learning_rate/action_model': 7.203305299416375e-05, 'epoch': 3.29}) +[2026-06-20 02:00:51,161][starVLA.training.train_starvla][INFO] - Step 1890, Loss: {'train/loss': 0.23009835183620453, 'train/grad_norm_pre_clip': 13.497589111328125, 'timing/data': 0.006653403863310814, 'timing/model': 3.01873935572803, 'learning_rate/qwen_vl_interface': 7.200528889445707e-06, 'learning_rate/action_model': 7.200528889445706e-05, 'epoch': 3.29}) +[2026-06-20 02:01:02,028][starVLA.training.train_starvla][INFO] - Step 1891, Loss: {'train/loss': 0.28957757353782654, 'train/grad_norm_pre_clip': 14.355897903442383, 'timing/data': 0.006352154538035393, 'timing/model': 2.781669309362769, 'learning_rate/qwen_vl_interface': 7.19775167768629e-06, 'learning_rate/action_model': 7.19775167768629e-05, 'epoch': 3.29}) +[2026-06-20 02:01:12,929][starVLA.training.train_starvla][INFO] - Step 1892, Loss: {'train/loss': 0.21831287443637848, 'train/grad_norm_pre_clip': 11.561792373657227, 'timing/data': 0.006666572764515877, 'timing/model': 2.787210987880826, 'learning_rate/qwen_vl_interface': 7.194973665279731e-06, 'learning_rate/action_model': 7.19497366527973e-05, 'epoch': 3.3}) +[2026-06-20 02:01:23,825][starVLA.training.train_starvla][INFO] - Step 1893, Loss: {'train/loss': 0.1668338179588318, 'train/grad_norm_pre_clip': 11.167726516723633, 'timing/data': 0.006902672350406647, 'timing/model': 2.7774665262550116, 'learning_rate/qwen_vl_interface': 7.1921948533679666e-06, 'learning_rate/action_model': 7.192194853367967e-05, 'epoch': 3.3}) +[2026-06-20 02:01:34,929][starVLA.training.train_starvla][INFO] - Step 1894, Loss: {'train/loss': 0.216019868850708, 'train/grad_norm_pre_clip': 13.603313446044922, 'timing/data': 0.00664837472140789, 'timing/model': 2.780125653371215, 'learning_rate/qwen_vl_interface': 7.1894152430932605e-06, 'learning_rate/action_model': 7.189415243093261e-05, 'epoch': 3.3}) +[2026-06-20 02:01:45,822][starVLA.training.train_starvla][INFO] - Step 1895, Loss: {'train/loss': 0.15725070238113403, 'train/grad_norm_pre_clip': 8.770258903503418, 'timing/data': 0.006876152008771896, 'timing/model': 2.772060927003622, 'learning_rate/qwen_vl_interface': 7.186634835598205e-06, 'learning_rate/action_model': 7.186634835598205e-05, 'epoch': 3.3}) +[2026-06-20 02:01:56,736][starVLA.training.train_starvla][INFO] - Step 1896, Loss: {'train/loss': 0.33007606863975525, 'train/grad_norm_pre_clip': 13.092987060546875, 'timing/data': 0.006920319050550461, 'timing/model': 2.7792424391955137, 'learning_rate/qwen_vl_interface': 7.183853632025724e-06, 'learning_rate/action_model': 7.183853632025723e-05, 'epoch': 3.3}) +[2026-06-20 02:02:07,636][starVLA.training.train_starvla][INFO] - Step 1897, Loss: {'train/loss': 0.1797112226486206, 'train/grad_norm_pre_clip': 9.992034912109375, 'timing/data': 0.006608955562114716, 'timing/model': 2.78221601434052, 'learning_rate/qwen_vl_interface': 7.181071633519062e-06, 'learning_rate/action_model': 7.181071633519062e-05, 'epoch': 3.3}) +[2026-06-20 02:02:18,728][starVLA.training.train_starvla][INFO] - Step 1898, Loss: {'train/loss': 0.2687896490097046, 'train/grad_norm_pre_clip': 12.022222518920898, 'timing/data': 0.006826221942901611, 'timing/model': 2.7665734123438597, 'learning_rate/qwen_vl_interface': 7.178288841221795e-06, 'learning_rate/action_model': 7.178288841221794e-05, 'epoch': 3.31}) +[2026-06-20 02:02:29,636][starVLA.training.train_starvla][INFO] - Step 1899, Loss: {'train/loss': 0.21641284227371216, 'train/grad_norm_pre_clip': 20.054157257080078, 'timing/data': 0.006510872393846512, 'timing/model': 2.7758058477193117, 'learning_rate/qwen_vl_interface': 7.1755052562778215e-06, 'learning_rate/action_model': 7.175505256277822e-05, 'epoch': 3.31}) +[2026-06-20 02:02:40,536][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1900 +[2026-06-20 02:10:05,282][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1900 +[2026-06-20 02:10:05,283][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 1900 +[2026-06-20 02:17:57,620][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 1900 +[2026-06-20 02:17:57,621][starVLA.training.train_starvla][INFO] - Step 1900, Loss: {'train/loss': 0.12041661143302917, 'train/grad_norm_pre_clip': 6.773932933807373, 'eval/loss': 0.4758014380931854, 'eval/latency_0/loss': 0.45019084215164185, 'eval/latency_1/loss': 0.44531822204589844, 'eval/latency_2/loss': 0.14004534482955933, 'eval/latency_3/loss': 0.5105194449424744, 'eval/latency_4/loss': 0.8532195091247559, 'eval/flappy/loss': 0.4758014380931854, 'eval/flappy/latency_0/loss': 0.45019084215164185, 'eval/flappy/latency_1/loss': 0.44531822204589844, 'eval/flappy/latency_2/loss': 0.14004534482955933, 'eval/flappy/latency_3/loss': 0.5105194449424744, 'eval/flappy/latency_4/loss': 0.8532195091247559, 'eval/flap_precision': 0.8404255319148937, 'eval/flap_recall': 0.2872727272727273, 'eval/flap_f1': 0.4281842818428184, 'eval/flap_rate_model': 0.026111111111111113, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.4281842818428184, 'eval/latency_0/flap_precision': 0.8404255319148937, 'eval/latency_0/flap_recall': 0.2872727272727273, 'eval/latency_0/flap_f1': 0.4281842818428184, 'eval/latency_0/flap_rate_model': 0.026111111111111113, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.4281842818428184, 'timing/data': 0.006788343191146851, 'timing/model': 2.7826026044785976, 'learning_rate/qwen_vl_interface': 7.1727208798313704e-06, 'learning_rate/action_model': 7.172720879831371e-05, 'epoch': 3.31}) +[2026-06-20 02:17:57,622][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_1900_state +[2026-06-20 02:17:57,643][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-20 02:18:03,452][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_1900_state/model.safetensors +[2026-06-20 02:18:17,626][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_1900_state/optimizer.bin +[2026-06-20 02:18:17,627][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_1900_state/sampler.bin +[2026-06-20 02:18:17,628][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_1900_state/sampler_1.bin +[2026-06-20 02:18:17,629][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_1900_state/random_states_0.pkl +[2026-06-20 02:18:17,753][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1900 +[2026-06-20 02:18:17,754][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-20 02:18:17,768][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-20 02:18:28,522][starVLA.training.train_starvla][INFO] - Step 1901, Loss: {'train/loss': 0.16623082756996155, 'train/grad_norm_pre_clip': 8.921217918395996, 'timing/data': 0.0068061016499996185, 'timing/model': 2.777747731655836, 'learning_rate/qwen_vl_interface': 7.169935713026994e-06, 'learning_rate/action_model': 7.169935713026994e-05, 'epoch': 3.31}) +[2026-06-20 02:18:39,417][starVLA.training.train_starvla][INFO] - Step 1902, Loss: {'train/loss': 0.15624161064624786, 'train/grad_norm_pre_clip': 8.726998329162598, 'timing/data': 0.0064808763563632965, 'timing/model': 2.7693373076617718, 'learning_rate/qwen_vl_interface': 7.1671497570095684e-06, 'learning_rate/action_model': 7.167149757009568e-05, 'epoch': 3.31}) +[2026-06-20 02:18:50,496][starVLA.training.train_starvla][INFO] - Step 1903, Loss: {'train/loss': 0.2097526341676712, 'train/grad_norm_pre_clip': 30.34453010559082, 'timing/data': 0.006449094042181969, 'timing/model': 2.948841020464897, 'learning_rate/qwen_vl_interface': 7.164363012924295e-06, 'learning_rate/action_model': 7.164363012924294e-05, 'epoch': 3.32}) +[2026-06-20 02:19:01,208][starVLA.training.train_starvla][INFO] - Step 1904, Loss: {'train/loss': 0.11148753017187119, 'train/grad_norm_pre_clip': 5.932083606719971, 'timing/data': 0.006608832627534866, 'timing/model': 2.766591940075159, 'learning_rate/qwen_vl_interface': 7.161575481916699e-06, 'learning_rate/action_model': 7.1615754819167e-05, 'epoch': 3.32}) +[2026-06-20 02:19:11,969][starVLA.training.train_starvla][INFO] - Step 1905, Loss: {'train/loss': 0.20503045618534088, 'train/grad_norm_pre_clip': 13.33576488494873, 'timing/data': 0.006342725828289986, 'timing/model': 2.7195078060030937, 'learning_rate/qwen_vl_interface': 7.158787165132627e-06, 'learning_rate/action_model': 7.158787165132626e-05, 'epoch': 3.32}) +[2026-06-20 02:19:22,725][starVLA.training.train_starvla][INFO] - Step 1906, Loss: {'train/loss': 0.32628124952316284, 'train/grad_norm_pre_clip': 22.47044563293457, 'timing/data': 0.006578454747796059, 'timing/model': 2.78630362264812, 'learning_rate/qwen_vl_interface': 7.1559980637182535e-06, 'learning_rate/action_model': 7.155998063718254e-05, 'epoch': 3.32}) +[2026-06-20 02:19:33,751][starVLA.training.train_starvla][INFO] - Step 1907, Loss: {'train/loss': 0.2566687762737274, 'train/grad_norm_pre_clip': 10.090494155883789, 'timing/data': 0.0065340157598257065, 'timing/model': 2.6471471153199673, 'learning_rate/qwen_vl_interface': 7.153208178820071e-06, 'learning_rate/action_model': 7.153208178820072e-05, 'epoch': 3.32}) +[2026-06-20 02:19:44,417][starVLA.training.train_starvla][INFO] - Step 1908, Loss: {'train/loss': 0.3461039066314697, 'train/grad_norm_pre_clip': 17.754154205322266, 'timing/data': 0.00715736486017704, 'timing/model': 2.744452089071274, 'learning_rate/qwen_vl_interface': 7.150417511584898e-06, 'learning_rate/action_model': 7.150417511584898e-05, 'epoch': 3.32}) +[2026-06-20 02:19:55,303][starVLA.training.train_starvla][INFO] - Step 1909, Loss: {'train/loss': 0.20532165467739105, 'train/grad_norm_pre_clip': 20.027973175048828, 'timing/data': 0.006561374291777611, 'timing/model': 2.760974034667015, 'learning_rate/qwen_vl_interface': 7.147626063159872e-06, 'learning_rate/action_model': 7.147626063159872e-05, 'epoch': 3.33}) +[2026-06-20 02:20:06,137][starVLA.training.train_starvla][INFO] - Step 1910, Loss: {'train/loss': 0.33667945861816406, 'train/grad_norm_pre_clip': 15.758013725280762, 'timing/data': 0.006417853757739067, 'timing/model': 2.7878577429801226, 'learning_rate/qwen_vl_interface': 7.14483383469245e-06, 'learning_rate/action_model': 7.144833834692449e-05, 'epoch': 3.33}) +[2026-06-20 02:20:17,042][starVLA.training.train_starvla][INFO] - Step 1911, Loss: {'train/loss': 0.19245801866054535, 'train/grad_norm_pre_clip': 8.519420623779297, 'timing/data': 0.0071622151881456375, 'timing/model': 2.761652646586299, 'learning_rate/qwen_vl_interface': 7.142040827330411e-06, 'learning_rate/action_model': 7.142040827330411e-05, 'epoch': 3.33}) +[2026-06-20 02:20:27,820][starVLA.training.train_starvla][INFO] - Step 1912, Loss: {'train/loss': 0.14339837431907654, 'train/grad_norm_pre_clip': 12.534210205078125, 'timing/data': 0.006647711619734764, 'timing/model': 2.770549051463604, 'learning_rate/qwen_vl_interface': 7.139247042221859e-06, 'learning_rate/action_model': 7.139247042221858e-05, 'epoch': 3.33}) +[2026-06-20 02:20:38,721][starVLA.training.train_starvla][INFO] - Step 1913, Loss: {'train/loss': 0.08636902272701263, 'train/grad_norm_pre_clip': 3.8721840381622314, 'timing/data': 0.00665036216378212, 'timing/model': 2.7668061200529337, 'learning_rate/qwen_vl_interface': 7.136452480515208e-06, 'learning_rate/action_model': 7.136452480515208e-05, 'epoch': 3.33}) +[2026-06-20 02:20:49,630][starVLA.training.train_starvla][INFO] - Step 1914, Loss: {'train/loss': 0.16228531301021576, 'train/grad_norm_pre_clip': 7.099661827087402, 'timing/data': 0.006254374980926514, 'timing/model': 2.7775413002818823, 'learning_rate/qwen_vl_interface': 7.133657143359203e-06, 'learning_rate/action_model': 7.133657143359202e-05, 'epoch': 3.33}) +[2026-06-20 02:21:00,630][starVLA.training.train_starvla][INFO] - Step 1915, Loss: {'train/loss': 0.08784526586532593, 'train/grad_norm_pre_clip': 4.322089195251465, 'timing/data': 0.006471535190939903, 'timing/model': 2.776825614273548, 'learning_rate/qwen_vl_interface': 7.130861031902897e-06, 'learning_rate/action_model': 7.130861031902897e-05, 'epoch': 3.34}) +[2026-06-20 02:21:11,518][starVLA.training.train_starvla][INFO] - Step 1916, Loss: {'train/loss': 0.10665053129196167, 'train/grad_norm_pre_clip': 7.479077339172363, 'timing/data': 0.006572730839252472, 'timing/model': 2.765496000647545, 'learning_rate/qwen_vl_interface': 7.128064147295667e-06, 'learning_rate/action_model': 7.128064147295667e-05, 'epoch': 3.34}) +[2026-06-20 02:21:22,430][starVLA.training.train_starvla][INFO] - Step 1917, Loss: {'train/loss': 0.3717236816883087, 'train/grad_norm_pre_clip': 14.571558952331543, 'timing/data': 0.006608480587601662, 'timing/model': 2.775009099394083, 'learning_rate/qwen_vl_interface': 7.125266490687206e-06, 'learning_rate/action_model': 7.125266490687206e-05, 'epoch': 3.34}) +[2026-06-20 02:21:33,330][starVLA.training.train_starvla][INFO] - Step 1918, Loss: {'train/loss': 0.1167808249592781, 'train/grad_norm_pre_clip': 8.841174125671387, 'timing/data': 0.006471715867519379, 'timing/model': 2.774447640404105, 'learning_rate/qwen_vl_interface': 7.1224680632275265e-06, 'learning_rate/action_model': 7.122468063227527e-05, 'epoch': 3.34}) +[2026-06-20 02:21:44,231][starVLA.training.train_starvla][INFO] - Step 1919, Loss: {'train/loss': 0.11901520937681198, 'train/grad_norm_pre_clip': 9.488201141357422, 'timing/data': 0.006327196955680847, 'timing/model': 2.7829234041273594, 'learning_rate/qwen_vl_interface': 7.1196688660669555e-06, 'learning_rate/action_model': 7.119668866066955e-05, 'epoch': 3.34}) +[2026-06-20 02:21:55,130][starVLA.training.train_starvla][INFO] - Step 1920, Loss: {'train/loss': 0.14476190507411957, 'train/grad_norm_pre_clip': 13.139514923095703, 'timing/data': 0.00651131384074688, 'timing/model': 2.7742053009569645, 'learning_rate/qwen_vl_interface': 7.116868900356138e-06, 'learning_rate/action_model': 7.116868900356138e-05, 'epoch': 3.34}) +[2026-06-20 02:22:06,033][starVLA.training.train_starvla][INFO] - Step 1921, Loss: {'train/loss': 0.06343743950128555, 'train/grad_norm_pre_clip': 6.013578414916992, 'timing/data': 0.006475117057561874, 'timing/model': 2.7744599282741547, 'learning_rate/qwen_vl_interface': 7.114068167246033e-06, 'learning_rate/action_model': 7.114068167246033e-05, 'epoch': 3.35}) +[2026-06-20 02:22:16,933][starVLA.training.train_starvla][INFO] - Step 1922, Loss: {'train/loss': 0.1771944910287857, 'train/grad_norm_pre_clip': 12.132187843322754, 'timing/data': 0.006427966058254242, 'timing/model': 2.783766482025385, 'learning_rate/qwen_vl_interface': 7.111266667887916e-06, 'learning_rate/action_model': 7.111266667887915e-05, 'epoch': 3.35}) +[2026-06-20 02:22:27,826][starVLA.training.train_starvla][INFO] - Step 1923, Loss: {'train/loss': 0.21656596660614014, 'train/grad_norm_pre_clip': 14.85682201385498, 'timing/data': 0.006224935874342918, 'timing/model': 2.768193358555436, 'learning_rate/qwen_vl_interface': 7.108464403433378e-06, 'learning_rate/action_model': 7.108464403433378e-05, 'epoch': 3.35}) +[2026-06-20 02:22:38,736][starVLA.training.train_starvla][INFO] - Step 1924, Loss: {'train/loss': 0.2768654525279999, 'train/grad_norm_pre_clip': 20.317079544067383, 'timing/data': 0.006184207275509834, 'timing/model': 2.779292855411768, 'learning_rate/qwen_vl_interface': 7.105661375034324e-06, 'learning_rate/action_model': 7.105661375034324e-05, 'epoch': 3.35}) +[2026-06-20 02:22:49,619][starVLA.training.train_starvla][INFO] - Step 1925, Loss: {'train/loss': 0.07874570041894913, 'train/grad_norm_pre_clip': 9.54081916809082, 'timing/data': 0.006302503868937492, 'timing/model': 2.7605855241417885, 'learning_rate/qwen_vl_interface': 7.102857583842973e-06, 'learning_rate/action_model': 7.102857583842974e-05, 'epoch': 3.35}) +[2026-06-20 02:23:00,634][starVLA.training.train_starvla][INFO] - Step 1926, Loss: {'train/loss': 0.052122268825769424, 'train/grad_norm_pre_clip': 6.598580360412598, 'timing/data': 0.006586723029613495, 'timing/model': 2.786275612190366, 'learning_rate/qwen_vl_interface': 7.100053031011859e-06, 'learning_rate/action_model': 7.100053031011859e-05, 'epoch': 3.36}) +[2026-06-20 02:23:11,530][starVLA.training.train_starvla][INFO] - Step 1927, Loss: {'train/loss': 0.08540385961532593, 'train/grad_norm_pre_clip': 7.011022567749023, 'timing/data': 0.006238466128706932, 'timing/model': 2.774986194446683, 'learning_rate/qwen_vl_interface': 7.097247717693827e-06, 'learning_rate/action_model': 7.097247717693827e-05, 'epoch': 3.36}) +[2026-06-20 02:23:22,357][starVLA.training.train_starvla][INFO] - Step 1928, Loss: {'train/loss': 0.37180981040000916, 'train/grad_norm_pre_clip': 17.40233612060547, 'timing/data': 0.005736600607633591, 'timing/model': 2.70166277885437, 'learning_rate/qwen_vl_interface': 7.094441645042033e-06, 'learning_rate/action_model': 7.094441645042032e-05, 'epoch': 3.36}) +[2026-06-20 02:23:33,133][starVLA.training.train_starvla][INFO] - Step 1929, Loss: {'train/loss': 0.15167813003063202, 'train/grad_norm_pre_clip': 16.803482055664062, 'timing/data': 0.006408127024769783, 'timing/model': 2.778811464086175, 'learning_rate/qwen_vl_interface': 7.09163481420995e-06, 'learning_rate/action_model': 7.09163481420995e-05, 'epoch': 3.36}) +[2026-06-20 02:23:44,035][starVLA.training.train_starvla][INFO] - Step 1930, Loss: {'train/loss': 0.2674461603164673, 'train/grad_norm_pre_clip': 12.65896224975586, 'timing/data': 0.0064860135316848755, 'timing/model': 2.7816898450255394, 'learning_rate/qwen_vl_interface': 7.0888272263513595e-06, 'learning_rate/action_model': 7.088827226351359e-05, 'epoch': 3.36}) +[2026-06-20 02:23:54,931][starVLA.training.train_starvla][INFO] - Step 1931, Loss: {'train/loss': 0.07791285961866379, 'train/grad_norm_pre_clip': 6.943277835845947, 'timing/data': 0.006524493917822838, 'timing/model': 2.7687330432236195, 'learning_rate/qwen_vl_interface': 7.086018882620354e-06, 'learning_rate/action_model': 7.086018882620354e-05, 'epoch': 3.36}) +[2026-06-20 02:24:05,832][starVLA.training.train_starvla][INFO] - Step 1932, Loss: {'train/loss': 0.17771080136299133, 'train/grad_norm_pre_clip': 13.035889625549316, 'timing/data': 0.0063857827335596085, 'timing/model': 2.7777241691946983, 'learning_rate/qwen_vl_interface': 7.083209784171341e-06, 'learning_rate/action_model': 7.083209784171341e-05, 'epoch': 3.37}) +[2026-06-20 02:24:16,738][starVLA.training.train_starvla][INFO] - Step 1933, Loss: {'train/loss': 0.22255730628967285, 'train/grad_norm_pre_clip': 20.210695266723633, 'timing/data': 0.006608262658119202, 'timing/model': 2.78296628408134, 'learning_rate/qwen_vl_interface': 7.080399932159029e-06, 'learning_rate/action_model': 7.080399932159029e-05, 'epoch': 3.37}) +[2026-06-20 02:24:27,834][starVLA.training.train_starvla][INFO] - Step 1934, Loss: {'train/loss': 0.13227961957454681, 'train/grad_norm_pre_clip': 9.707557678222656, 'timing/data': 0.006515564396977425, 'timing/model': 2.7770154047757387, 'learning_rate/qwen_vl_interface': 7.077589327738448e-06, 'learning_rate/action_model': 7.077589327738447e-05, 'epoch': 3.37}) +[2026-06-20 02:24:38,733][starVLA.training.train_starvla][INFO] - Step 1935, Loss: {'train/loss': 0.12048397958278656, 'train/grad_norm_pre_clip': 13.418431282043457, 'timing/data': 0.00652313232421875, 'timing/model': 2.7799567710608244, 'learning_rate/qwen_vl_interface': 7.0747779720649265e-06, 'learning_rate/action_model': 7.074777972064927e-05, 'epoch': 3.37}) +[2026-06-20 02:24:49,633][starVLA.training.train_starvla][INFO] - Step 1936, Loss: {'train/loss': 0.10764959454536438, 'train/grad_norm_pre_clip': 7.338595390319824, 'timing/data': 0.006495783105492592, 'timing/model': 2.7783460784703493, 'learning_rate/qwen_vl_interface': 7.071965866294108e-06, 'learning_rate/action_model': 7.071965866294108e-05, 'epoch': 3.37}) +[2026-06-20 02:25:00,534][starVLA.training.train_starvla][INFO] - Step 1937, Loss: {'train/loss': 0.2808903157711029, 'train/grad_norm_pre_clip': 15.954017639160156, 'timing/data': 0.006403524428606033, 'timing/model': 2.783717393875122, 'learning_rate/qwen_vl_interface': 7.0691530115819454e-06, 'learning_rate/action_model': 7.069153011581945e-05, 'epoch': 3.37}) +[2026-06-20 02:25:11,423][starVLA.training.train_starvla][INFO] - Step 1938, Loss: {'train/loss': 0.08402400463819504, 'train/grad_norm_pre_clip': 7.617893695831299, 'timing/data': 0.006386784836649895, 'timing/model': 2.768479436635971, 'learning_rate/qwen_vl_interface': 7.066339409084694e-06, 'learning_rate/action_model': 7.066339409084694e-05, 'epoch': 3.38}) +[2026-06-20 02:25:22,337][starVLA.training.train_starvla][INFO] - Step 1939, Loss: {'train/loss': 0.18400493264198303, 'train/grad_norm_pre_clip': 14.749869346618652, 'timing/data': 0.006599823012948036, 'timing/model': 2.7798548825085163, 'learning_rate/qwen_vl_interface': 7.0635250599589215e-06, 'learning_rate/action_model': 7.06352505995892e-05, 'epoch': 3.38}) +[2026-06-20 02:25:33,233][starVLA.training.train_starvla][INFO] - Step 1940, Loss: {'train/loss': 0.27831950783729553, 'train/grad_norm_pre_clip': 15.735225677490234, 'timing/data': 0.0065527018159627914, 'timing/model': 2.781417664140463, 'learning_rate/qwen_vl_interface': 7.060709965361499e-06, 'learning_rate/action_model': 7.060709965361499e-05, 'epoch': 3.38}) +[2026-06-20 02:25:44,122][starVLA.training.train_starvla][INFO] - Step 1941, Loss: {'train/loss': 0.1793634593486786, 'train/grad_norm_pre_clip': 12.44010066986084, 'timing/data': 0.006561905145645142, 'timing/model': 2.7595856562256813, 'learning_rate/qwen_vl_interface': 7.057894126449606e-06, 'learning_rate/action_model': 7.057894126449606e-05, 'epoch': 3.38}) +[2026-06-20 02:25:55,135][starVLA.training.train_starvla][INFO] - Step 1942, Loss: {'train/loss': 0.31977447867393494, 'train/grad_norm_pre_clip': 25.632205963134766, 'timing/data': 0.006678661331534386, 'timing/model': 2.7751149348914623, 'learning_rate/qwen_vl_interface': 7.055077544380728e-06, 'learning_rate/action_model': 7.055077544380728e-05, 'epoch': 3.38}) +[2026-06-20 02:26:06,029][starVLA.training.train_starvla][INFO] - Step 1943, Loss: {'train/loss': 0.3661572337150574, 'train/grad_norm_pre_clip': 26.272415161132812, 'timing/data': 0.006398333236575127, 'timing/model': 2.7701923958957195, 'learning_rate/qwen_vl_interface': 7.0522602203126554e-06, 'learning_rate/action_model': 7.052260220312656e-05, 'epoch': 3.38}) +[2026-06-20 02:26:16,934][starVLA.training.train_starvla][INFO] - Step 1944, Loss: {'train/loss': 0.10035019367933273, 'train/grad_norm_pre_clip': 8.167330741882324, 'timing/data': 0.006383756175637245, 'timing/model': 2.7749176286160946, 'learning_rate/qwen_vl_interface': 7.0494421554034845e-06, 'learning_rate/action_model': 7.049442155403484e-05, 'epoch': 3.39}) +[2026-06-20 02:26:27,834][starVLA.training.train_starvla][INFO] - Step 1945, Loss: {'train/loss': 0.20853547751903534, 'train/grad_norm_pre_clip': 12.807429313659668, 'timing/data': 0.006557963788509369, 'timing/model': 2.7874814961105585, 'learning_rate/qwen_vl_interface': 7.046623350811612e-06, 'learning_rate/action_model': 7.046623350811612e-05, 'epoch': 3.39}) +[2026-06-20 02:26:38,732][starVLA.training.train_starvla][INFO] - Step 1946, Loss: {'train/loss': 0.28504809737205505, 'train/grad_norm_pre_clip': 12.157838821411133, 'timing/data': 0.00664990209043026, 'timing/model': 2.7752708327025175, 'learning_rate/qwen_vl_interface': 7.043803807695747e-06, 'learning_rate/action_model': 7.043803807695747e-05, 'epoch': 3.39}) +[2026-06-20 02:26:49,522][starVLA.training.train_starvla][INFO] - Step 1947, Loss: {'train/loss': 0.16711516678333282, 'train/grad_norm_pre_clip': 11.525553703308105, 'timing/data': 0.006463745608925819, 'timing/model': 2.745207415893674, 'learning_rate/qwen_vl_interface': 7.040983527214893e-06, 'learning_rate/action_model': 7.040983527214893e-05, 'epoch': 3.39}) +[2026-06-20 02:27:00,126][starVLA.training.train_starvla][INFO] - Step 1948, Loss: {'train/loss': 0.16505223512649536, 'train/grad_norm_pre_clip': 11.619194984436035, 'timing/data': 0.00633537583053112, 'timing/model': 2.7407271321862936, 'learning_rate/qwen_vl_interface': 7.038162510528363e-06, 'learning_rate/action_model': 7.038162510528362e-05, 'epoch': 3.39}) +[2026-06-20 02:27:11,069][starVLA.training.train_starvla][INFO] - Step 1949, Loss: {'train/loss': 0.10875318944454193, 'train/grad_norm_pre_clip': 6.98398494720459, 'timing/data': 0.006272885948419571, 'timing/model': 2.9929781518876553, 'learning_rate/qwen_vl_interface': 7.03534075879577e-06, 'learning_rate/action_model': 7.03534075879577e-05, 'epoch': 3.4}) +[2026-06-20 02:27:21,932][starVLA.training.train_starvla][INFO] - Step 1950, Loss: {'train/loss': 0.20970378816127777, 'train/grad_norm_pre_clip': 14.004779815673828, 'timing/data': 0.006651081144809723, 'timing/model': 2.786197615787387, 'learning_rate/qwen_vl_interface': 7.032518273177028e-06, 'learning_rate/action_model': 7.032518273177028e-05, 'epoch': 3.4}) +[2026-06-20 02:27:32,830][starVLA.training.train_starvla][INFO] - Step 1951, Loss: {'train/loss': 0.13579335808753967, 'train/grad_norm_pre_clip': 12.30936336517334, 'timing/data': 0.006418503820896149, 'timing/model': 2.7720390800386667, 'learning_rate/qwen_vl_interface': 7.029695054832357e-06, 'learning_rate/action_model': 7.029695054832357e-05, 'epoch': 3.4}) +[2026-06-20 02:27:43,735][starVLA.training.train_starvla][INFO] - Step 1952, Loss: {'train/loss': 0.16125498712062836, 'train/grad_norm_pre_clip': 25.532760620117188, 'timing/data': 0.006400877609848976, 'timing/model': 2.7824776601046324, 'learning_rate/qwen_vl_interface': 7.026871104922273e-06, 'learning_rate/action_model': 7.026871104922273e-05, 'epoch': 3.4}) +[2026-06-20 02:27:54,832][starVLA.training.train_starvla][INFO] - Step 1953, Loss: {'train/loss': 0.22423064708709717, 'train/grad_norm_pre_clip': 18.257251739501953, 'timing/data': 0.006372248753905296, 'timing/model': 2.7756111212074757, 'learning_rate/qwen_vl_interface': 7.024046424607596e-06, 'learning_rate/action_model': 7.024046424607595e-05, 'epoch': 3.4}) +[2026-06-20 02:28:05,735][starVLA.training.train_starvla][INFO] - Step 1954, Loss: {'train/loss': 0.14149940013885498, 'train/grad_norm_pre_clip': 11.62570571899414, 'timing/data': 0.006232360377907753, 'timing/model': 2.773576395586133, 'learning_rate/qwen_vl_interface': 7.021221015049445e-06, 'learning_rate/action_model': 7.021221015049446e-05, 'epoch': 3.4}) +[2026-06-20 02:28:16,635][starVLA.training.train_starvla][INFO] - Step 1955, Loss: {'train/loss': 0.3037463128566742, 'train/grad_norm_pre_clip': 17.2684383392334, 'timing/data': 0.006342217326164246, 'timing/model': 2.7857550866901875, 'learning_rate/qwen_vl_interface': 7.01839487740924e-06, 'learning_rate/action_model': 7.01839487740924e-05, 'epoch': 3.41}) +[2026-06-20 02:28:27,520][starVLA.training.train_starvla][INFO] - Step 1956, Loss: {'train/loss': 0.13094022870063782, 'train/grad_norm_pre_clip': 10.149837493896484, 'timing/data': 0.006599416956305504, 'timing/model': 2.7627522833645344, 'learning_rate/qwen_vl_interface': 7.015568012848699e-06, 'learning_rate/action_model': 7.015568012848699e-05, 'epoch': 3.41}) +[2026-06-20 02:28:38,638][starVLA.training.train_starvla][INFO] - Step 1957, Loss: {'train/loss': 0.3382137417793274, 'train/grad_norm_pre_clip': 18.73556900024414, 'timing/data': 0.006507327780127525, 'timing/model': 2.778288146480918, 'learning_rate/qwen_vl_interface': 7.01274042252984e-06, 'learning_rate/action_model': 7.01274042252984e-05, 'epoch': 3.41}) +[2026-06-20 02:28:49,531][starVLA.training.train_starvla][INFO] - Step 1958, Loss: {'train/loss': 0.29695719480514526, 'train/grad_norm_pre_clip': 23.630847930908203, 'timing/data': 0.0065287258476018906, 'timing/model': 2.7744500152766705, 'learning_rate/qwen_vl_interface': 7.009912107614978e-06, 'learning_rate/action_model': 7.009912107614978e-05, 'epoch': 3.41}) +[2026-06-20 02:29:00,433][starVLA.training.train_starvla][INFO] - Step 1959, Loss: {'train/loss': 0.03389207646250725, 'train/grad_norm_pre_clip': 3.1606972217559814, 'timing/data': 0.006622936576604843, 'timing/model': 2.7729290574789047, 'learning_rate/qwen_vl_interface': 7.0070830692667244e-06, 'learning_rate/action_model': 7.007083069266724e-05, 'epoch': 3.41}) +[2026-06-20 02:29:11,567][starVLA.training.train_starvla][INFO] - Step 1960, Loss: {'train/loss': 0.23894110321998596, 'train/grad_norm_pre_clip': 11.92245864868164, 'timing/data': 0.006669443100690842, 'timing/model': 3.0187681932002306, 'learning_rate/qwen_vl_interface': 7.004253308647993e-06, 'learning_rate/action_model': 7.004253308647994e-05, 'epoch': 3.41}) +[2026-06-20 02:29:22,434][starVLA.training.train_starvla][INFO] - Step 1961, Loss: {'train/loss': 0.13416551053524017, 'train/grad_norm_pre_clip': 10.076051712036133, 'timing/data': 0.006656654179096222, 'timing/model': 2.776605624705553, 'learning_rate/qwen_vl_interface': 7.001422826921991e-06, 'learning_rate/action_model': 7.00142282692199e-05, 'epoch': 3.42}) +[2026-06-20 02:29:33,337][starVLA.training.train_starvla][INFO] - Step 1962, Loss: {'train/loss': 0.195668026804924, 'train/grad_norm_pre_clip': 13.681099891662598, 'timing/data': 0.0065874867141246796, 'timing/model': 2.777815580368042, 'learning_rate/qwen_vl_interface': 6.9985916252522225e-06, 'learning_rate/action_model': 6.998591625252222e-05, 'epoch': 3.42}) +[2026-06-20 02:29:44,235][starVLA.training.train_starvla][INFO] - Step 1963, Loss: {'train/loss': 0.3871583938598633, 'train/grad_norm_pre_clip': 20.338953018188477, 'timing/data': 0.006615547463297844, 'timing/model': 2.78835098259151, 'learning_rate/qwen_vl_interface': 6.995759704802485e-06, 'learning_rate/action_model': 6.995759704802485e-05, 'epoch': 3.42}) +[2026-06-20 02:29:55,265][starVLA.training.train_starvla][INFO] - Step 1964, Loss: {'train/loss': 0.0832812562584877, 'train/grad_norm_pre_clip': 12.92591667175293, 'timing/data': 0.006383197382092476, 'timing/model': 2.774273967370391, 'learning_rate/qwen_vl_interface': 6.992927066736876e-06, 'learning_rate/action_model': 6.992927066736875e-05, 'epoch': 3.42}) +[2026-06-20 02:30:06,137][starVLA.training.train_starvla][INFO] - Step 1965, Loss: {'train/loss': 0.308237761259079, 'train/grad_norm_pre_clip': 19.199514389038086, 'timing/data': 0.006261307746171951, 'timing/model': 2.775501864030957, 'learning_rate/qwen_vl_interface': 6.990093712219787e-06, 'learning_rate/action_model': 6.990093712219787e-05, 'epoch': 3.42}) +[2026-06-20 02:30:17,039][starVLA.training.train_starvla][INFO] - Step 1966, Loss: {'train/loss': 0.22195588052272797, 'train/grad_norm_pre_clip': 20.246379852294922, 'timing/data': 0.006758164614439011, 'timing/model': 2.7789509128779173, 'learning_rate/qwen_vl_interface': 6.9872596424159e-06, 'learning_rate/action_model': 6.9872596424159e-05, 'epoch': 3.42}) +[2026-06-20 02:30:27,936][starVLA.training.train_starvla][INFO] - Step 1967, Loss: {'train/loss': 0.28131988644599915, 'train/grad_norm_pre_clip': 16.16592788696289, 'timing/data': 0.006833365187048912, 'timing/model': 2.7778596002608538, 'learning_rate/qwen_vl_interface': 6.9844248584902e-06, 'learning_rate/action_model': 6.984424858490199e-05, 'epoch': 3.43}) +[2026-06-20 02:30:38,822][starVLA.training.train_starvla][INFO] - Step 1968, Loss: {'train/loss': 0.07624862343072891, 'train/grad_norm_pre_clip': 9.490707397460938, 'timing/data': 0.006302708759903908, 'timing/model': 2.764594931155443, 'learning_rate/qwen_vl_interface': 6.981589361607952e-06, 'learning_rate/action_model': 6.981589361607953e-05, 'epoch': 3.43}) +[2026-06-20 02:30:49,739][starVLA.training.train_starvla][INFO] - Step 1969, Loss: {'train/loss': 0.21919438242912292, 'train/grad_norm_pre_clip': 13.951997756958008, 'timing/data': 0.006869934499263763, 'timing/model': 2.781056236475706, 'learning_rate/qwen_vl_interface': 6.9787531529347274e-06, 'learning_rate/action_model': 6.978753152934727e-05, 'epoch': 3.43}) +[2026-06-20 02:31:00,625][starVLA.training.train_starvla][INFO] - Step 1970, Loss: {'train/loss': 0.14234106242656708, 'train/grad_norm_pre_clip': 6.633638858795166, 'timing/data': 0.006427528336644173, 'timing/model': 2.784823175519705, 'learning_rate/qwen_vl_interface': 6.97591623363638e-06, 'learning_rate/action_model': 6.97591623363638e-05, 'epoch': 3.43}) +[2026-06-20 02:31:11,530][starVLA.training.train_starvla][INFO] - Step 1971, Loss: {'train/loss': 0.13752861320972443, 'train/grad_norm_pre_clip': 7.39617919921875, 'timing/data': 0.006512828171253204, 'timing/model': 2.785837806761265, 'learning_rate/qwen_vl_interface': 6.9730786048790644e-06, 'learning_rate/action_model': 6.973078604879064e-05, 'epoch': 3.43}) +[2026-06-20 02:31:22,427][starVLA.training.train_starvla][INFO] - Step 1972, Loss: {'train/loss': 0.19627344608306885, 'train/grad_norm_pre_clip': 21.438648223876953, 'timing/data': 0.006397400051355362, 'timing/model': 2.7692142613232136, 'learning_rate/qwen_vl_interface': 6.970240267829221e-06, 'learning_rate/action_model': 6.970240267829222e-05, 'epoch': 3.44}) +[2026-06-20 02:31:33,331][starVLA.training.train_starvla][INFO] - Step 1973, Loss: {'train/loss': 0.11012299358844757, 'train/grad_norm_pre_clip': 12.75657844543457, 'timing/data': 0.006596658378839493, 'timing/model': 2.7792130801826715, 'learning_rate/qwen_vl_interface': 6.967401223653582e-06, 'learning_rate/action_model': 6.967401223653581e-05, 'epoch': 3.44}) +[2026-06-20 02:31:44,237][starVLA.training.train_starvla][INFO] - Step 1974, Loss: {'train/loss': 0.31832027435302734, 'train/grad_norm_pre_clip': 16.494094848632812, 'timing/data': 0.006591856479644775, 'timing/model': 2.772545652464032, 'learning_rate/qwen_vl_interface': 6.9645614735191724e-06, 'learning_rate/action_model': 6.964561473519172e-05, 'epoch': 3.44}) +[2026-06-20 02:31:55,140][starVLA.training.train_starvla][INFO] - Step 1975, Loss: {'train/loss': 0.2620448172092438, 'train/grad_norm_pre_clip': 14.066671371459961, 'timing/data': 0.006643565371632576, 'timing/model': 2.7820527143776417, 'learning_rate/qwen_vl_interface': 6.961721018593305e-06, 'learning_rate/action_model': 6.961721018593305e-05, 'epoch': 3.44}) +[2026-06-20 02:32:06,034][starVLA.training.train_starvla][INFO] - Step 1976, Loss: {'train/loss': 0.26059266924858093, 'train/grad_norm_pre_clip': 244.87692260742188, 'timing/data': 0.006686335429549217, 'timing/model': 2.7809363398700953, 'learning_rate/qwen_vl_interface': 6.958879860043585e-06, 'learning_rate/action_model': 6.958879860043585e-05, 'epoch': 3.44}) +[2026-06-20 02:32:16,933][starVLA.training.train_starvla][INFO] - Step 1977, Loss: {'train/loss': 0.23388642072677612, 'train/grad_norm_pre_clip': 8.2330904006958, 'timing/data': 0.006396329030394554, 'timing/model': 2.7896887101233006, 'learning_rate/qwen_vl_interface': 6.9560379990379025e-06, 'learning_rate/action_model': 6.956037999037902e-05, 'epoch': 3.44}) +[2026-06-20 02:32:27,841][starVLA.training.train_starvla][INFO] - Step 1978, Loss: {'train/loss': 0.1401969939470291, 'train/grad_norm_pre_clip': 7.187943935394287, 'timing/data': 0.006695887073874474, 'timing/model': 2.7827292010188103, 'learning_rate/qwen_vl_interface': 6.953195436744442e-06, 'learning_rate/action_model': 6.953195436744442e-05, 'epoch': 3.45}) +[2026-06-20 02:32:38,740][starVLA.training.train_starvla][INFO] - Step 1979, Loss: {'train/loss': 0.057992056012153625, 'train/grad_norm_pre_clip': 7.1363911628723145, 'timing/data': 0.0065643396228551865, 'timing/model': 2.778232455253601, 'learning_rate/qwen_vl_interface': 6.950352174331676e-06, 'learning_rate/action_model': 6.950352174331675e-05, 'epoch': 3.45}) +[2026-06-20 02:32:49,832][starVLA.training.train_starvla][INFO] - Step 1980, Loss: {'train/loss': 0.19632351398468018, 'train/grad_norm_pre_clip': 9.868864059448242, 'timing/data': 0.006657598540186882, 'timing/model': 2.773508032783866, 'learning_rate/qwen_vl_interface': 6.947508212968356e-06, 'learning_rate/action_model': 6.947508212968356e-05, 'epoch': 3.45}) +[2026-06-20 02:33:00,739][starVLA.training.train_starvla][INFO] - Step 1981, Loss: {'train/loss': 0.14321818947792053, 'train/grad_norm_pre_clip': 6.749695301055908, 'timing/data': 0.0066928863525390625, 'timing/model': 2.779994575306773, 'learning_rate/qwen_vl_interface': 6.9446635538235295e-06, 'learning_rate/action_model': 6.944663553823529e-05, 'epoch': 3.45}) +[2026-06-20 02:33:11,632][starVLA.training.train_starvla][INFO] - Step 1982, Loss: {'train/loss': 0.15570950508117676, 'train/grad_norm_pre_clip': 10.371009826660156, 'timing/data': 0.006627356633543968, 'timing/model': 2.768203489482403, 'learning_rate/qwen_vl_interface': 6.94181819806653e-06, 'learning_rate/action_model': 6.94181819806653e-05, 'epoch': 3.45}) +[2026-06-20 02:33:22,528][starVLA.training.train_starvla][INFO] - Step 1983, Loss: {'train/loss': 0.05184436962008476, 'train/grad_norm_pre_clip': 2.9766933917999268, 'timing/data': 0.006504889577627182, 'timing/model': 2.7670414745807648, 'learning_rate/qwen_vl_interface': 6.938972146866975e-06, 'learning_rate/action_model': 6.938972146866974e-05, 'epoch': 3.45}) +[2026-06-20 02:33:33,428][starVLA.training.train_starvla][INFO] - Step 1984, Loss: {'train/loss': 0.20814214646816254, 'train/grad_norm_pre_clip': 13.773103713989258, 'timing/data': 0.006620679050683975, 'timing/model': 2.781363297253847, 'learning_rate/qwen_vl_interface': 6.936125401394769e-06, 'learning_rate/action_model': 6.936125401394768e-05, 'epoch': 3.46}) +[2026-06-20 02:33:44,329][starVLA.training.train_starvla][INFO] - Step 1985, Loss: {'train/loss': 0.28127267956733704, 'train/grad_norm_pre_clip': 14.705610275268555, 'timing/data': 0.006290651857852936, 'timing/model': 2.7703140880912542, 'learning_rate/qwen_vl_interface': 6.933277962820101e-06, 'learning_rate/action_model': 6.9332779628201e-05, 'epoch': 3.46}) +[2026-06-20 02:33:55,224][starVLA.training.train_starvla][INFO] - Step 1986, Loss: {'train/loss': 0.1501571387052536, 'train/grad_norm_pre_clip': 11.449899673461914, 'timing/data': 0.006477199494838715, 'timing/model': 2.764255115762353, 'learning_rate/qwen_vl_interface': 6.930429832313444e-06, 'learning_rate/action_model': 6.930429832313444e-05, 'epoch': 3.46}) +[2026-06-20 02:34:06,279][starVLA.training.train_starvla][INFO] - Step 1987, Loss: {'train/loss': 0.11268139630556107, 'train/grad_norm_pre_clip': 7.013158321380615, 'timing/data': 0.00651898980140686, 'timing/model': 2.921937020495534, 'learning_rate/qwen_vl_interface': 6.92758101104556e-06, 'learning_rate/action_model': 6.92758101104556e-05, 'epoch': 3.46}) +[2026-06-20 02:34:17,134][starVLA.training.train_starvla][INFO] - Step 1988, Loss: {'train/loss': 0.11582875996828079, 'train/grad_norm_pre_clip': 11.916479110717773, 'timing/data': 0.0066098980605602264, 'timing/model': 2.775471257045865, 'learning_rate/qwen_vl_interface': 6.9247315001874885e-06, 'learning_rate/action_model': 6.924731500187488e-05, 'epoch': 3.46}) +[2026-06-20 02:34:28,039][starVLA.training.train_starvla][INFO] - Step 1989, Loss: {'train/loss': 0.0901784598827362, 'train/grad_norm_pre_clip': 10.673752784729004, 'timing/data': 0.006939584389328957, 'timing/model': 2.7766842301934958, 'learning_rate/qwen_vl_interface': 6.9218813009105605e-06, 'learning_rate/action_model': 6.92188130091056e-05, 'epoch': 3.47}) +[2026-06-20 02:34:38,937][starVLA.training.train_starvla][INFO] - Step 1990, Loss: {'train/loss': 0.1184859648346901, 'train/grad_norm_pre_clip': 12.28909683227539, 'timing/data': 0.006735425442457199, 'timing/model': 2.779280910268426, 'learning_rate/qwen_vl_interface': 6.919030414386378e-06, 'learning_rate/action_model': 6.919030414386379e-05, 'epoch': 3.47}) +[2026-06-20 02:34:50,031][starVLA.training.train_starvla][INFO] - Step 1991, Loss: {'train/loss': 0.27369049191474915, 'train/grad_norm_pre_clip': 21.96187400817871, 'timing/data': 0.006545448675751686, 'timing/model': 2.7229246757924557, 'learning_rate/qwen_vl_interface': 6.916178841786839e-06, 'learning_rate/action_model': 6.916178841786838e-05, 'epoch': 3.47}) +[2026-06-20 02:35:00,921][starVLA.training.train_starvla][INFO] - Step 1992, Loss: {'train/loss': 0.09459090977907181, 'train/grad_norm_pre_clip': 8.317337036132812, 'timing/data': 0.006517980247735977, 'timing/model': 2.7624148968607187, 'learning_rate/qwen_vl_interface': 6.913326584284115e-06, 'learning_rate/action_model': 6.913326584284115e-05, 'epoch': 3.47}) +[2026-06-20 02:35:11,821][starVLA.training.train_starvla][INFO] - Step 1993, Loss: {'train/loss': 0.11377205699682236, 'train/grad_norm_pre_clip': 15.41740894317627, 'timing/data': 0.006599698215723038, 'timing/model': 2.7730416283011436, 'learning_rate/qwen_vl_interface': 6.910473643050662e-06, 'learning_rate/action_model': 6.910473643050662e-05, 'epoch': 3.47}) +[2026-06-20 02:35:22,636][starVLA.training.train_starvla][INFO] - Step 1994, Loss: {'train/loss': 0.2485571652650833, 'train/grad_norm_pre_clip': 20.85654067993164, 'timing/data': 0.006620179861783981, 'timing/model': 2.781179716810584, 'learning_rate/qwen_vl_interface': 6.907620019259215e-06, 'learning_rate/action_model': 6.907620019259216e-05, 'epoch': 3.47}) +[2026-06-20 02:35:33,738][starVLA.training.train_starvla][INFO] - Step 1995, Loss: {'train/loss': 0.15593478083610535, 'train/grad_norm_pre_clip': 11.223484992980957, 'timing/data': 0.006608389317989349, 'timing/model': 2.7764559648931026, 'learning_rate/qwen_vl_interface': 6.9047657140827935e-06, 'learning_rate/action_model': 6.904765714082794e-05, 'epoch': 3.48}) +[2026-06-20 02:35:44,623][starVLA.training.train_starvla][INFO] - Step 1996, Loss: {'train/loss': 0.19953683018684387, 'train/grad_norm_pre_clip': 10.61208724975586, 'timing/data': 0.006520461291074753, 'timing/model': 2.7732053622603416, 'learning_rate/qwen_vl_interface': 6.901910728694694e-06, 'learning_rate/action_model': 6.901910728694694e-05, 'epoch': 3.48}) +[2026-06-20 02:35:55,535][starVLA.training.train_starvla][INFO] - Step 1997, Loss: {'train/loss': 0.18115046620368958, 'train/grad_norm_pre_clip': 16.212276458740234, 'timing/data': 0.00662476010620594, 'timing/model': 2.7886350825428963, 'learning_rate/qwen_vl_interface': 6.899055064268489e-06, 'learning_rate/action_model': 6.899055064268489e-05, 'epoch': 3.48}) +[2026-06-20 02:36:06,428][starVLA.training.train_starvla][INFO] - Step 1998, Loss: {'train/loss': 0.13145850598812103, 'train/grad_norm_pre_clip': 22.355815887451172, 'timing/data': 0.006519218906760216, 'timing/model': 2.782277222722769, 'learning_rate/qwen_vl_interface': 6.896198721978038e-06, 'learning_rate/action_model': 6.896198721978038e-05, 'epoch': 3.48}) +[2026-06-20 02:36:17,400][starVLA.training.train_starvla][INFO] - Step 1999, Loss: {'train/loss': 0.26778313517570496, 'train/grad_norm_pre_clip': 13.845410346984863, 'timing/data': 0.005864478647708893, 'timing/model': 2.745328690856695, 'learning_rate/qwen_vl_interface': 6.893341702997475e-06, 'learning_rate/action_model': 6.893341702997475e-05, 'epoch': 3.48}) +[2026-06-20 02:36:28,228][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 2000 +[2026-06-20 02:43:56,828][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 2000 +[2026-06-20 02:43:56,829][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 2000 +[2026-06-20 02:51:50,735][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 2000 +[2026-06-20 02:51:50,737][starVLA.training.train_starvla][INFO] - Step 2000, Loss: {'train/loss': 0.08859306573867798, 'train/grad_norm_pre_clip': 15.014080047607422, 'eval/loss': 0.754784345626831, 'eval/latency_0/loss': 0.726469099521637, 'eval/latency_1/loss': 0.8597943186759949, 'eval/latency_2/loss': 0.10399628430604935, 'eval/latency_3/loss': 0.838905394077301, 'eval/latency_4/loss': 1.2850592136383057, 'eval/flappy/loss': 0.754784345626831, 'eval/flappy/latency_0/loss': 0.726469099521637, 'eval/flappy/latency_1/loss': 0.8597943186759949, 'eval/flappy/latency_2/loss': 0.10399628430604935, 'eval/flappy/latency_3/loss': 0.838905394077301, 'eval/flappy/latency_4/loss': 1.2850592136383057, 'eval/flap_precision': 0.8674698795180723, 'eval/flap_recall': 0.26181818181818184, 'eval/flap_f1': 0.4022346368715084, 'eval/flap_rate_model': 0.023055555555555555, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.40223463687150834, 'eval/latency_0/flap_precision': 0.8674698795180723, 'eval/latency_0/flap_recall': 0.26181818181818184, 'eval/latency_0/flap_f1': 0.4022346368715084, 'eval/latency_0/flap_rate_model': 0.023055555555555555, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.40223463687150834, 'timing/data': 0.006511827930808067, 'timing/model': 2.770959371700883, 'learning_rate/qwen_vl_interface': 6.890484008501211e-06, 'learning_rate/action_model': 6.890484008501211e-05, 'epoch': 3.48}) +[2026-06-20 02:51:50,738][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_2000_state +[2026-06-20 02:51:50,742][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-20 02:51:56,468][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_2000_state/model.safetensors +[2026-06-20 02:52:10,453][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_2000_state/optimizer.bin +[2026-06-20 02:52:10,454][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_2000_state/sampler.bin +[2026-06-20 02:52:10,454][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_2000_state/sampler_1.bin +[2026-06-20 02:52:10,455][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_2000_state/random_states_0.pkl +[2026-06-20 02:52:10,581][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2000 +[2026-06-20 02:52:10,582][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-20 02:52:10,596][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-20 02:52:21,422][starVLA.training.train_starvla][INFO] - Step 2001, Loss: {'train/loss': 0.34038758277893066, 'train/grad_norm_pre_clip': 16.714109420776367, 'timing/data': 0.00791860744357109, 'timing/model': 2.77689684741199, 'learning_rate/qwen_vl_interface': 6.887625639663941e-06, 'learning_rate/action_model': 6.88762563966394e-05, 'epoch': 3.49}) +[2026-06-20 02:52:32,324][starVLA.training.train_starvla][INFO] - Step 2002, Loss: {'train/loss': 0.1856895089149475, 'train/grad_norm_pre_clip': 13.107693672180176, 'timing/data': 0.007403342053294182, 'timing/model': 2.78618904389441, 'learning_rate/qwen_vl_interface': 6.884766597660626e-06, 'learning_rate/action_model': 6.884766597660626e-05, 'epoch': 3.49}) +[2026-06-20 02:52:43,129][starVLA.training.train_starvla][INFO] - Step 2003, Loss: {'train/loss': 0.061191149055957794, 'train/grad_norm_pre_clip': 8.673636436462402, 'timing/data': 0.012505292892456055, 'timing/model': 2.7772104125469923, 'learning_rate/qwen_vl_interface': 6.881906883666515e-06, 'learning_rate/action_model': 6.881906883666515e-05, 'epoch': 3.49}) +[2026-06-20 02:52:53,926][starVLA.training.train_starvla][INFO] - Step 2004, Loss: {'train/loss': 0.29038551449775696, 'train/grad_norm_pre_clip': 20.99344253540039, 'timing/data': 0.0065123531967401505, 'timing/model': 2.789084941148758, 'learning_rate/qwen_vl_interface': 6.879046498857125e-06, 'learning_rate/action_model': 6.879046498857124e-05, 'epoch': 3.49}) +[2026-06-20 02:53:04,827][starVLA.training.train_starvla][INFO] - Step 2005, Loss: {'train/loss': 0.0813935399055481, 'train/grad_norm_pre_clip': 5.460483074188232, 'timing/data': 0.0064750127494335175, 'timing/model': 2.7814588993787766, 'learning_rate/qwen_vl_interface': 6.876185444408255e-06, 'learning_rate/action_model': 6.876185444408255e-05, 'epoch': 3.49}) +[2026-06-20 02:53:15,622][starVLA.training.train_starvla][INFO] - Step 2006, Loss: {'train/loss': 0.03028198517858982, 'train/grad_norm_pre_clip': 3.1607046127319336, 'timing/data': 0.006840521469712257, 'timing/model': 2.7714751921594143, 'learning_rate/qwen_vl_interface': 6.873323721495976e-06, 'learning_rate/action_model': 6.873323721495976e-05, 'epoch': 3.49}) +[2026-06-20 02:53:26,723][starVLA.training.train_starvla][INFO] - Step 2007, Loss: {'train/loss': 0.15524274110794067, 'train/grad_norm_pre_clip': 13.000077247619629, 'timing/data': 0.006856158375740051, 'timing/model': 2.7843121886253357, 'learning_rate/qwen_vl_interface': 6.870461331296632e-06, 'learning_rate/action_model': 6.870461331296632e-05, 'epoch': 3.5}) +[2026-06-20 02:53:37,626][starVLA.training.train_starvla][INFO] - Step 2008, Loss: {'train/loss': 0.19087444245815277, 'train/grad_norm_pre_clip': 17.267972946166992, 'timing/data': 0.00665656104683876, 'timing/model': 2.7756563927978277, 'learning_rate/qwen_vl_interface': 6.8675982749868455e-06, 'learning_rate/action_model': 6.867598274986845e-05, 'epoch': 3.5}) +[2026-06-20 02:53:48,526][starVLA.training.train_starvla][INFO] - Step 2009, Loss: {'train/loss': 0.2439776211977005, 'train/grad_norm_pre_clip': 30.073211669921875, 'timing/data': 0.006838241592049599, 'timing/model': 2.777217708528042, 'learning_rate/qwen_vl_interface': 6.864734553743509e-06, 'learning_rate/action_model': 6.86473455374351e-05, 'epoch': 3.5}) +[2026-06-20 02:53:59,511][starVLA.training.train_starvla][INFO] - Step 2010, Loss: {'train/loss': 0.1411060243844986, 'train/grad_norm_pre_clip': 10.097511291503906, 'timing/data': 0.006712010130286217, 'timing/model': 2.853216752409935, 'learning_rate/qwen_vl_interface': 6.861870168743793e-06, 'learning_rate/action_model': 6.861870168743793e-05, 'epoch': 3.5}) +[2026-06-20 02:54:10,426][starVLA.training.train_starvla][INFO] - Step 2011, Loss: {'train/loss': 0.2674989700317383, 'train/grad_norm_pre_clip': 27.260025024414062, 'timing/data': 0.00710984505712986, 'timing/model': 2.7715351413935423, 'learning_rate/qwen_vl_interface': 6.859005121165135e-06, 'learning_rate/action_model': 6.859005121165135e-05, 'epoch': 3.5}) +[2026-06-20 02:54:21,333][starVLA.training.train_starvla][INFO] - Step 2012, Loss: {'train/loss': 0.1421816647052765, 'train/grad_norm_pre_clip': 13.660046577453613, 'timing/data': 0.006470423191785812, 'timing/model': 2.781450442969799, 'learning_rate/qwen_vl_interface': 6.85613941218525e-06, 'learning_rate/action_model': 6.85613941218525e-05, 'epoch': 3.51}) +[2026-06-20 02:54:32,150][starVLA.training.train_starvla][INFO] - Step 2013, Loss: {'train/loss': 0.12015513330698013, 'train/grad_norm_pre_clip': 9.649894714355469, 'timing/data': 0.006659381091594696, 'timing/model': 2.6987229231745005, 'learning_rate/qwen_vl_interface': 6.8532730429821215e-06, 'learning_rate/action_model': 6.853273042982122e-05, 'epoch': 3.51}) +[2026-06-20 02:54:43,308][starVLA.training.train_starvla][INFO] - Step 2014, Loss: {'train/loss': 0.1865609884262085, 'train/grad_norm_pre_clip': 13.686820030212402, 'timing/data': 0.0065382421016693115, 'timing/model': 3.059966119006276, 'learning_rate/qwen_vl_interface': 6.850406014734005e-06, 'learning_rate/action_model': 6.850406014734005e-05, 'epoch': 3.51}) +[2026-06-20 02:54:53,944][starVLA.training.train_starvla][INFO] - Step 2015, Loss: {'train/loss': 0.29761171340942383, 'train/grad_norm_pre_clip': 14.47510051727295, 'timing/data': 0.006636079400777817, 'timing/model': 2.76442039757967, 'learning_rate/qwen_vl_interface': 6.8475383286194295e-06, 'learning_rate/action_model': 6.84753832861943e-05, 'epoch': 3.51}) +[2026-06-20 02:55:04,779][starVLA.training.train_starvla][INFO] - Step 2016, Loss: {'train/loss': 0.07143963128328323, 'train/grad_norm_pre_clip': 7.102839469909668, 'timing/data': 0.006831377744674683, 'timing/model': 2.7317831274122, 'learning_rate/qwen_vl_interface': 6.844669985817192e-06, 'learning_rate/action_model': 6.844669985817192e-05, 'epoch': 3.51}) +[2026-06-20 02:55:15,624][starVLA.training.train_starvla][INFO] - Step 2017, Loss: {'train/loss': 0.11309275031089783, 'train/grad_norm_pre_clip': 7.522976875305176, 'timing/data': 0.006718391552567482, 'timing/model': 2.7700291834771633, 'learning_rate/qwen_vl_interface': 6.8418009875063594e-06, 'learning_rate/action_model': 6.84180098750636e-05, 'epoch': 3.51}) +[2026-06-20 02:55:26,530][starVLA.training.train_starvla][INFO] - Step 2018, Loss: {'train/loss': 0.1962045580148697, 'train/grad_norm_pre_clip': 16.44817352294922, 'timing/data': 0.006682399660348892, 'timing/model': 2.771289575845003, 'learning_rate/qwen_vl_interface': 6.838931334866267e-06, 'learning_rate/action_model': 6.838931334866266e-05, 'epoch': 3.52}) +[2026-06-20 02:55:37,437][starVLA.training.train_starvla][INFO] - Step 2019, Loss: {'train/loss': 0.03267326205968857, 'train/grad_norm_pre_clip': 5.341914653778076, 'timing/data': 0.0066481102257966995, 'timing/model': 2.7810969464480877, 'learning_rate/qwen_vl_interface': 6.836061029076523e-06, 'learning_rate/action_model': 6.836061029076522e-05, 'epoch': 3.52}) +[2026-06-20 02:55:48,335][starVLA.training.train_starvla][INFO] - Step 2020, Loss: {'train/loss': 0.15953077375888824, 'train/grad_norm_pre_clip': 11.206814765930176, 'timing/data': 0.006825568154454231, 'timing/model': 2.770672056823969, 'learning_rate/qwen_vl_interface': 6.833190071317e-06, 'learning_rate/action_model': 6.833190071316999e-05, 'epoch': 3.52}) +[2026-06-20 02:55:59,228][starVLA.training.train_starvla][INFO] - Step 2021, Loss: {'train/loss': 0.2213222086429596, 'train/grad_norm_pre_clip': 13.45390510559082, 'timing/data': 0.0065672919154167175, 'timing/model': 2.768911173567176, 'learning_rate/qwen_vl_interface': 6.830318462767841e-06, 'learning_rate/action_model': 6.830318462767841e-05, 'epoch': 3.52}) +[2026-06-20 02:56:10,235][starVLA.training.train_starvla][INFO] - Step 2022, Loss: {'train/loss': 0.22779187560081482, 'train/grad_norm_pre_clip': 15.587806701660156, 'timing/data': 0.007148945704102516, 'timing/model': 2.769264604896307, 'learning_rate/qwen_vl_interface': 6.827446204609456e-06, 'learning_rate/action_model': 6.827446204609456e-05, 'epoch': 3.52}) +[2026-06-20 02:56:21,139][starVLA.training.train_starvla][INFO] - Step 2023, Loss: {'train/loss': 0.051434680819511414, 'train/grad_norm_pre_clip': 5.062228679656982, 'timing/data': 0.006594952195882797, 'timing/model': 2.7790480349212885, 'learning_rate/qwen_vl_interface': 6.824573298022522e-06, 'learning_rate/action_model': 6.824573298022522e-05, 'epoch': 3.52}) +[2026-06-20 02:56:32,034][starVLA.training.train_starvla][INFO] - Step 2024, Loss: {'train/loss': 0.2901996672153473, 'train/grad_norm_pre_clip': 49.9079475402832, 'timing/data': 0.007027396932244301, 'timing/model': 2.7853765934705734, 'learning_rate/qwen_vl_interface': 6.8216997441879805e-06, 'learning_rate/action_model': 6.821699744187981e-05, 'epoch': 3.53}) +[2026-06-20 02:56:42,940][starVLA.training.train_starvla][INFO] - Step 2025, Loss: {'train/loss': 0.1522100418806076, 'train/grad_norm_pre_clip': 7.967951774597168, 'timing/data': 0.006752878427505493, 'timing/model': 2.7781974375247955, 'learning_rate/qwen_vl_interface': 6.818825544287045e-06, 'learning_rate/action_model': 6.818825544287045e-05, 'epoch': 3.53}) +[2026-06-20 02:56:53,829][starVLA.training.train_starvla][INFO] - Step 2026, Loss: {'train/loss': 0.13258206844329834, 'train/grad_norm_pre_clip': 8.54732894897461, 'timing/data': 0.006444994360208511, 'timing/model': 2.7669792845845222, 'learning_rate/qwen_vl_interface': 6.815950699501186e-06, 'learning_rate/action_model': 6.815950699501186e-05, 'epoch': 3.53}) +[2026-06-20 02:57:04,731][starVLA.training.train_starvla][INFO] - Step 2027, Loss: {'train/loss': 0.11347896605730057, 'train/grad_norm_pre_clip': 7.266505718231201, 'timing/data': 0.006647482514381409, 'timing/model': 2.7731227539479733, 'learning_rate/qwen_vl_interface': 6.813075211012148e-06, 'learning_rate/action_model': 6.813075211012147e-05, 'epoch': 3.53}) +[2026-06-20 02:57:15,625][starVLA.training.train_starvla][INFO] - Step 2028, Loss: {'train/loss': 0.23075251281261444, 'train/grad_norm_pre_clip': 20.49020767211914, 'timing/data': 0.006529552862048149, 'timing/model': 2.7691545374691486, 'learning_rate/qwen_vl_interface': 6.81019908000193e-06, 'learning_rate/action_model': 6.81019908000193e-05, 'epoch': 3.53}) +[2026-06-20 02:57:26,538][starVLA.training.train_starvla][INFO] - Step 2029, Loss: {'train/loss': 0.100528284907341, 'train/grad_norm_pre_clip': 8.615925788879395, 'timing/data': 0.006844738498330116, 'timing/model': 2.7765925843268633, 'learning_rate/qwen_vl_interface': 6.807322307652806e-06, 'learning_rate/action_model': 6.807322307652806e-05, 'epoch': 3.53}) +[2026-06-20 02:57:37,640][starVLA.training.train_starvla][INFO] - Step 2030, Loss: {'train/loss': 0.3707636892795563, 'train/grad_norm_pre_clip': 19.02582550048828, 'timing/data': 0.006858270615339279, 'timing/model': 2.7787277717143297, 'learning_rate/qwen_vl_interface': 6.804444895147306e-06, 'learning_rate/action_model': 6.804444895147306e-05, 'epoch': 3.54}) +[2026-06-20 02:57:48,540][starVLA.training.train_starvla][INFO] - Step 2031, Loss: {'train/loss': 0.2918863296508789, 'train/grad_norm_pre_clip': 12.002359390258789, 'timing/data': 0.006375495344400406, 'timing/model': 2.78039313480258, 'learning_rate/qwen_vl_interface': 6.8015668436682255e-06, 'learning_rate/action_model': 6.801566843668226e-05, 'epoch': 3.54}) +[2026-06-20 02:57:59,443][starVLA.training.train_starvla][INFO] - Step 2032, Loss: {'train/loss': 0.16997946798801422, 'train/grad_norm_pre_clip': 11.49817180633545, 'timing/data': 0.0067157503217458725, 'timing/model': 2.778474610298872, 'learning_rate/qwen_vl_interface': 6.7986881543986245e-06, 'learning_rate/action_model': 6.798688154398624e-05, 'epoch': 3.54}) +[2026-06-20 02:58:10,337][starVLA.training.train_starvla][INFO] - Step 2033, Loss: {'train/loss': 0.0641133040189743, 'train/grad_norm_pre_clip': 5.355755805969238, 'timing/data': 0.013820497319102287, 'timing/model': 2.7674849927425385, 'learning_rate/qwen_vl_interface': 6.795808828521822e-06, 'learning_rate/action_model': 6.795808828521821e-05, 'epoch': 3.54}) +[2026-06-20 02:58:21,232][starVLA.training.train_starvla][INFO] - Step 2034, Loss: {'train/loss': 0.27999815344810486, 'train/grad_norm_pre_clip': 19.369625091552734, 'timing/data': 0.006850028410553932, 'timing/model': 2.774926559999585, 'learning_rate/qwen_vl_interface': 6.792928867221401e-06, 'learning_rate/action_model': 6.792928867221401e-05, 'epoch': 3.54}) +[2026-06-20 02:58:32,144][starVLA.training.train_starvla][INFO] - Step 2035, Loss: {'train/loss': 0.1137123852968216, 'train/grad_norm_pre_clip': 13.275386810302734, 'timing/data': 0.011479070410132408, 'timing/model': 2.7841376326978207, 'learning_rate/qwen_vl_interface': 6.790048271681202e-06, 'learning_rate/action_model': 6.790048271681202e-05, 'epoch': 3.55}) +[2026-06-20 02:58:42,857][starVLA.training.train_starvla][INFO] - Step 2036, Loss: {'train/loss': 0.21314740180969238, 'train/grad_norm_pre_clip': 17.608951568603516, 'timing/data': 0.006749199703335762, 'timing/model': 2.6789299864321947, 'learning_rate/qwen_vl_interface': 6.7871670430853335e-06, 'learning_rate/action_model': 6.787167043085333e-05, 'epoch': 3.55}) +[2026-06-20 02:58:53,899][starVLA.training.train_starvla][INFO] - Step 2037, Loss: {'train/loss': 0.16782952845096588, 'train/grad_norm_pre_clip': 9.274520874023438, 'timing/data': 0.0063629113137722015, 'timing/model': 2.94434517249465, 'learning_rate/qwen_vl_interface': 6.784285182618156e-06, 'learning_rate/action_model': 6.784285182618156e-05, 'epoch': 3.55}) +[2026-06-20 02:59:04,632][starVLA.training.train_starvla][INFO] - Step 2038, Loss: {'train/loss': 0.13096095621585846, 'train/grad_norm_pre_clip': 10.604225158691406, 'timing/data': 0.006542712450027466, 'timing/model': 2.7703922241926193, 'learning_rate/qwen_vl_interface': 6.781402691464297e-06, 'learning_rate/action_model': 6.781402691464297e-05, 'epoch': 3.55}) +[2026-06-20 02:59:15,533][starVLA.training.train_starvla][INFO] - Step 2039, Loss: {'train/loss': 0.2334921807050705, 'train/grad_norm_pre_clip': 8.588998794555664, 'timing/data': 0.0068430472165346146, 'timing/model': 2.7759566958993673, 'learning_rate/qwen_vl_interface': 6.778519570808636e-06, 'learning_rate/action_model': 6.778519570808636e-05, 'epoch': 3.55}) +[2026-06-20 02:59:26,425][starVLA.training.train_starvla][INFO] - Step 2040, Loss: {'train/loss': 0.07591146230697632, 'train/grad_norm_pre_clip': 6.045179843902588, 'timing/data': 0.006678169593214989, 'timing/model': 2.7726489137858152, 'learning_rate/qwen_vl_interface': 6.775635821836317e-06, 'learning_rate/action_model': 6.775635821836317e-05, 'epoch': 3.55}) +[2026-06-20 02:59:37,543][starVLA.training.train_starvla][INFO] - Step 2041, Loss: {'train/loss': 0.32902318239212036, 'train/grad_norm_pre_clip': 21.527406692504883, 'timing/data': 0.006786299869418144, 'timing/model': 2.7440011091530323, 'learning_rate/qwen_vl_interface': 6.77275144573274e-06, 'learning_rate/action_model': 6.77275144573274e-05, 'epoch': 3.56}) +[2026-06-20 02:59:48,435][starVLA.training.train_starvla][INFO] - Step 2042, Loss: {'train/loss': 0.0816943347454071, 'train/grad_norm_pre_clip': 5.243033409118652, 'timing/data': 0.006549639627337456, 'timing/model': 2.77648582495749, 'learning_rate/qwen_vl_interface': 6.769866443683564e-06, 'learning_rate/action_model': 6.769866443683563e-05, 'epoch': 3.56}) +[2026-06-20 02:59:59,332][starVLA.training.train_starvla][INFO] - Step 2043, Loss: {'train/loss': 0.14011605083942413, 'train/grad_norm_pre_clip': 7.430380821228027, 'timing/data': 0.0068779196590185165, 'timing/model': 2.7678203247487545, 'learning_rate/qwen_vl_interface': 6.766980816874701e-06, 'learning_rate/action_model': 6.7669808168747e-05, 'epoch': 3.56}) +[2026-06-20 03:00:10,238][starVLA.training.train_starvla][INFO] - Step 2044, Loss: {'train/loss': 0.19206702709197998, 'train/grad_norm_pre_clip': 10.59689998626709, 'timing/data': 0.008635180070996284, 'timing/model': 2.7755828574299812, 'learning_rate/qwen_vl_interface': 6.764094566492327e-06, 'learning_rate/action_model': 6.764094566492328e-05, 'epoch': 3.56}) +[2026-06-20 03:00:21,326][starVLA.training.train_starvla][INFO] - Step 2045, Loss: {'train/loss': 0.12691953778266907, 'train/grad_norm_pre_clip': 8.752217292785645, 'timing/data': 0.006541410461068153, 'timing/model': 2.7634330056607723, 'learning_rate/qwen_vl_interface': 6.7612076937228685e-06, 'learning_rate/action_model': 6.761207693722869e-05, 'epoch': 3.56}) +[2026-06-20 03:00:32,238][starVLA.training.train_starvla][INFO] - Step 2046, Loss: {'train/loss': 0.08705180883407593, 'train/grad_norm_pre_clip': 7.1269450187683105, 'timing/data': 0.006524901837110519, 'timing/model': 2.773818677291274, 'learning_rate/qwen_vl_interface': 6.7583201997530126e-06, 'learning_rate/action_model': 6.758320199753012e-05, 'epoch': 3.56}) +[2026-06-20 03:00:43,138][starVLA.training.train_starvla][INFO] - Step 2047, Loss: {'train/loss': 0.20299306511878967, 'train/grad_norm_pre_clip': 11.151874542236328, 'timing/data': 0.006720291450619698, 'timing/model': 2.7800269443541765, 'learning_rate/qwen_vl_interface': 6.755432085769693e-06, 'learning_rate/action_model': 6.755432085769694e-05, 'epoch': 3.57}) +[2026-06-20 03:00:54,035][starVLA.training.train_starvla][INFO] - Step 2048, Loss: {'train/loss': 0.10760026425123215, 'train/grad_norm_pre_clip': 6.258491516113281, 'timing/data': 0.006828980520367622, 'timing/model': 2.7690778989344835, 'learning_rate/qwen_vl_interface': 6.75254335296011e-06, 'learning_rate/action_model': 6.75254335296011e-05, 'epoch': 3.57}) +[2026-06-20 03:01:05,037][starVLA.training.train_starvla][INFO] - Step 2049, Loss: {'train/loss': 0.0491412989795208, 'train/grad_norm_pre_clip': 5.127586841583252, 'timing/data': 0.006600851193070412, 'timing/model': 2.7737515717744827, 'learning_rate/qwen_vl_interface': 6.749654002511712e-06, 'learning_rate/action_model': 6.749654002511712e-05, 'epoch': 3.57}) +[2026-06-20 03:01:15,937][starVLA.training.train_starvla][INFO] - Step 2050, Loss: {'train/loss': 0.1024264246225357, 'train/grad_norm_pre_clip': 6.719748497009277, 'timing/data': 0.006658952683210373, 'timing/model': 2.77693635225296, 'learning_rate/qwen_vl_interface': 6.7467640356122e-06, 'learning_rate/action_model': 6.746764035612199e-05, 'epoch': 3.57}) +[2026-06-20 03:01:26,839][starVLA.training.train_starvla][INFO] - Step 2051, Loss: {'train/loss': 0.1170383095741272, 'train/grad_norm_pre_clip': 9.077643394470215, 'timing/data': 0.006717700511217117, 'timing/model': 2.7803170196712017, 'learning_rate/qwen_vl_interface': 6.74387345344953e-06, 'learning_rate/action_model': 6.74387345344953e-05, 'epoch': 3.57}) +[2026-06-20 03:01:37,735][starVLA.training.train_starvla][INFO] - Step 2052, Loss: {'train/loss': 0.18597568571567535, 'train/grad_norm_pre_clip': 15.867692947387695, 'timing/data': 0.006792828440666199, 'timing/model': 2.7855718061327934, 'learning_rate/qwen_vl_interface': 6.740982257211911e-06, 'learning_rate/action_model': 6.740982257211911e-05, 'epoch': 3.57}) +[2026-06-20 03:01:48,634][starVLA.training.train_starvla][INFO] - Step 2053, Loss: {'train/loss': 0.09843610227108002, 'train/grad_norm_pre_clip': 11.388859748840332, 'timing/data': 0.00654580257833004, 'timing/model': 2.7733753845095634, 'learning_rate/qwen_vl_interface': 6.738090448087806e-06, 'learning_rate/action_model': 6.738090448087806e-05, 'epoch': 3.58}) +[2026-06-20 03:01:59,540][starVLA.training.train_starvla][INFO] - Step 2054, Loss: {'train/loss': 0.10356967896223068, 'train/grad_norm_pre_clip': 11.121692657470703, 'timing/data': 0.006549334153532982, 'timing/model': 2.784157508984208, 'learning_rate/qwen_vl_interface': 6.735198027265928e-06, 'learning_rate/action_model': 6.735198027265928e-05, 'epoch': 3.58}) +[2026-06-20 03:02:10,423][starVLA.training.train_starvla][INFO] - Step 2055, Loss: {'train/loss': 0.13173332810401917, 'train/grad_norm_pre_clip': 7.91864013671875, 'timing/data': 0.0068823471665382385, 'timing/model': 2.763787405565381, 'learning_rate/qwen_vl_interface': 6.7323049959352424e-06, 'learning_rate/action_model': 6.732304995935243e-05, 'epoch': 3.58}) +[2026-06-20 03:02:21,334][starVLA.training.train_starvla][INFO] - Step 2056, Loss: {'train/loss': 0.0668766126036644, 'train/grad_norm_pre_clip': 6.56643009185791, 'timing/data': 0.006888618692755699, 'timing/model': 2.7784322444349527, 'learning_rate/qwen_vl_interface': 6.729411355284962e-06, 'learning_rate/action_model': 6.729411355284963e-05, 'epoch': 3.58}) +[2026-06-20 03:02:32,233][starVLA.training.train_starvla][INFO] - Step 2057, Loss: {'train/loss': 0.20893682539463043, 'train/grad_norm_pre_clip': 11.873074531555176, 'timing/data': 0.0068978797644376755, 'timing/model': 2.7776134945452213, 'learning_rate/qwen_vl_interface': 6.726517106504559e-06, 'learning_rate/action_model': 6.726517106504558e-05, 'epoch': 3.58}) +[2026-06-20 03:02:43,033][starVLA.training.train_starvla][INFO] - Step 2058, Loss: {'train/loss': 0.07219672948122025, 'train/grad_norm_pre_clip': 7.235660552978516, 'timing/data': 0.006460173055529594, 'timing/model': 2.761385580524802, 'learning_rate/qwen_vl_interface': 6.723622250783745e-06, 'learning_rate/action_model': 6.723622250783744e-05, 'epoch': 3.59}) +[2026-06-20 03:02:53,932][starVLA.training.train_starvla][INFO] - Step 2059, Loss: {'train/loss': 0.3770943880081177, 'train/grad_norm_pre_clip': 21.84417152404785, 'timing/data': 0.009004272520542145, 'timing/model': 2.776163775473833, 'learning_rate/qwen_vl_interface': 6.720726789312487e-06, 'learning_rate/action_model': 6.720726789312487e-05, 'epoch': 3.59}) +[2026-06-20 03:03:04,933][starVLA.training.train_starvla][INFO] - Step 2060, Loss: {'train/loss': 0.19698971509933472, 'train/grad_norm_pre_clip': 13.281747817993164, 'timing/data': 0.006928928196430206, 'timing/model': 2.7778394762426615, 'learning_rate/qwen_vl_interface': 6.7178307232810015e-06, 'learning_rate/action_model': 6.717830723281e-05, 'epoch': 3.59}) +[2026-06-20 03:03:15,833][starVLA.training.train_starvla][INFO] - Step 2061, Loss: {'train/loss': 0.11695108562707901, 'train/grad_norm_pre_clip': 9.79274845123291, 'timing/data': 0.006868269294500351, 'timing/model': 2.770563129335642, 'learning_rate/qwen_vl_interface': 6.714934053879749e-06, 'learning_rate/action_model': 6.71493405387975e-05, 'epoch': 3.59}) +[2026-06-20 03:03:26,730][starVLA.training.train_starvla][INFO] - Step 2062, Loss: {'train/loss': 0.2789333760738373, 'train/grad_norm_pre_clip': 25.108047485351562, 'timing/data': 0.006598029285669327, 'timing/model': 2.773021785542369, 'learning_rate/qwen_vl_interface': 6.712036782299442e-06, 'learning_rate/action_model': 6.712036782299443e-05, 'epoch': 3.59}) +[2026-06-20 03:03:37,628][starVLA.training.train_starvla][INFO] - Step 2063, Loss: {'train/loss': 0.02852323092520237, 'train/grad_norm_pre_clip': 4.725509166717529, 'timing/data': 0.006723528727889061, 'timing/model': 2.7735594008117914, 'learning_rate/qwen_vl_interface': 6.709138909731041e-06, 'learning_rate/action_model': 6.709138909731041e-05, 'epoch': 3.59}) +[2026-06-20 03:03:48,525][starVLA.training.train_starvla][INFO] - Step 2064, Loss: {'train/loss': 0.07987692952156067, 'train/grad_norm_pre_clip': 9.456483840942383, 'timing/data': 0.006491182371973991, 'timing/model': 2.766907023265958, 'learning_rate/qwen_vl_interface': 6.706240437365751e-06, 'learning_rate/action_model': 6.70624043736575e-05, 'epoch': 3.6}) +[2026-06-20 03:03:59,411][starVLA.training.train_starvla][INFO] - Step 2065, Loss: {'train/loss': 0.1219780221581459, 'train/grad_norm_pre_clip': 10.185935974121094, 'timing/data': 0.006790578365325928, 'timing/model': 2.758206557482481, 'learning_rate/qwen_vl_interface': 6.703341366395023e-06, 'learning_rate/action_model': 6.703341366395023e-05, 'epoch': 3.6}) +[2026-06-20 03:04:10,126][starVLA.training.train_starvla][INFO] - Step 2066, Loss: {'train/loss': 0.20551057159900665, 'train/grad_norm_pre_clip': 19.705936431884766, 'timing/data': 0.006724020466208458, 'timing/model': 2.768871469423175, 'learning_rate/qwen_vl_interface': 6.700441698010558e-06, 'learning_rate/action_model': 6.700441698010558e-05, 'epoch': 3.6}) +[2026-06-20 03:04:21,020][starVLA.training.train_starvla][INFO] - Step 2067, Loss: {'train/loss': 0.09673172235488892, 'train/grad_norm_pre_clip': 11.301695823669434, 'timing/data': 0.006457531824707985, 'timing/model': 2.7781528793275356, 'learning_rate/qwen_vl_interface': 6.697541433404299e-06, 'learning_rate/action_model': 6.697541433404299e-05, 'epoch': 3.6}) +[2026-06-20 03:04:32,136][starVLA.training.train_starvla][INFO] - Step 2068, Loss: {'train/loss': 0.02211950346827507, 'train/grad_norm_pre_clip': 2.377990245819092, 'timing/data': 0.006790075451135635, 'timing/model': 2.784909203648567, 'learning_rate/qwen_vl_interface': 6.6946405737684325e-06, 'learning_rate/action_model': 6.694640573768433e-05, 'epoch': 3.6}) +[2026-06-20 03:04:43,024][starVLA.training.train_starvla][INFO] - Step 2069, Loss: {'train/loss': 0.1454307585954666, 'train/grad_norm_pre_clip': 14.276208877563477, 'timing/data': 0.006610509008169174, 'timing/model': 2.778186187148094, 'learning_rate/qwen_vl_interface': 6.691739120295397e-06, 'learning_rate/action_model': 6.691739120295397e-05, 'epoch': 3.6}) +[2026-06-20 03:04:53,923][starVLA.training.train_starvla][INFO] - Step 2070, Loss: {'train/loss': 0.21827343106269836, 'train/grad_norm_pre_clip': 27.500900268554688, 'timing/data': 0.006648838520050049, 'timing/model': 2.774920815601945, 'learning_rate/qwen_vl_interface': 6.688837074177868e-06, 'learning_rate/action_model': 6.688837074177868e-05, 'epoch': 3.61}) +[2026-06-20 03:05:04,755][starVLA.training.train_starvla][INFO] - Step 2071, Loss: {'train/loss': 0.08205116540193558, 'train/grad_norm_pre_clip': 7.219945907592773, 'timing/data': 0.008731931447982788, 'timing/model': 2.7003444097936153, 'learning_rate/qwen_vl_interface': 6.68593443660877e-06, 'learning_rate/action_model': 6.685934436608769e-05, 'epoch': 3.61}) +[2026-06-20 03:05:15,524][starVLA.training.train_starvla][INFO] - Step 2072, Loss: {'train/loss': 0.08463609218597412, 'train/grad_norm_pre_clip': 8.608846664428711, 'timing/data': 0.006355790421366692, 'timing/model': 2.7522968407720327, 'learning_rate/qwen_vl_interface': 6.6830312087812625e-06, 'learning_rate/action_model': 6.683031208781262e-05, 'epoch': 3.61}) +[2026-06-20 03:05:26,435][starVLA.training.train_starvla][INFO] - Step 2073, Loss: {'train/loss': 0.08687091618776321, 'train/grad_norm_pre_clip': 8.88976001739502, 'timing/data': 0.0066342782229185104, 'timing/model': 2.7774225808680058, 'learning_rate/qwen_vl_interface': 6.680127391888754e-06, 'learning_rate/action_model': 6.680127391888754e-05, 'epoch': 3.61}) +[2026-06-20 03:05:37,234][starVLA.training.train_starvla][INFO] - Step 2074, Loss: {'train/loss': 0.10539861023426056, 'train/grad_norm_pre_clip': 12.438026428222656, 'timing/data': 0.006592500954866409, 'timing/model': 2.776531707495451, 'learning_rate/qwen_vl_interface': 6.677222987124897e-06, 'learning_rate/action_model': 6.677222987124897e-05, 'epoch': 3.61}) +[2026-06-20 03:05:47,932][starVLA.training.train_starvla][INFO] - Step 2075, Loss: {'train/loss': 0.24815703928470612, 'train/grad_norm_pre_clip': 23.456838607788086, 'timing/data': 0.008716892451047897, 'timing/model': 2.771707398816943, 'learning_rate/qwen_vl_interface': 6.67431799568358e-06, 'learning_rate/action_model': 6.67431799568358e-05, 'epoch': 3.61}) +[2026-06-20 03:05:58,908][starVLA.training.train_starvla][INFO] - Step 2076, Loss: {'train/loss': 0.06135650351643562, 'train/grad_norm_pre_clip': 9.274198532104492, 'timing/data': 0.006698818877339363, 'timing/model': 2.758001644164324, 'learning_rate/qwen_vl_interface': 6.671412418758939e-06, 'learning_rate/action_model': 6.671412418758938e-05, 'epoch': 3.62}) +[2026-06-20 03:06:09,729][starVLA.training.train_starvla][INFO] - Step 2077, Loss: {'train/loss': 0.15956538915634155, 'train/grad_norm_pre_clip': 19.408735275268555, 'timing/data': 0.006777767091989517, 'timing/model': 2.7738185226917267, 'learning_rate/qwen_vl_interface': 6.668506257545341e-06, 'learning_rate/action_model': 6.668506257545341e-05, 'epoch': 3.62}) +[2026-06-20 03:06:20,628][starVLA.training.train_starvla][INFO] - Step 2078, Loss: {'train/loss': 0.021988222375512123, 'train/grad_norm_pre_clip': 4.682260036468506, 'timing/data': 0.0065982770174741745, 'timing/model': 2.7704642955213785, 'learning_rate/qwen_vl_interface': 6.665599513237403e-06, 'learning_rate/action_model': 6.665599513237402e-05, 'epoch': 3.62}) +[2026-06-20 03:06:31,530][starVLA.training.train_starvla][INFO] - Step 2079, Loss: {'train/loss': 0.27412912249565125, 'train/grad_norm_pre_clip': 21.118408203125, 'timing/data': 0.006425440311431885, 'timing/model': 2.7789182886481285, 'learning_rate/qwen_vl_interface': 6.662692187029978e-06, 'learning_rate/action_model': 6.662692187029978e-05, 'epoch': 3.62}) +[2026-06-20 03:06:42,424][starVLA.training.train_starvla][INFO] - Step 2080, Loss: {'train/loss': 0.10296288132667542, 'train/grad_norm_pre_clip': 15.837898254394531, 'timing/data': 0.0070932842791080475, 'timing/model': 2.764056660234928, 'learning_rate/qwen_vl_interface': 6.659784280118156e-06, 'learning_rate/action_model': 6.659784280118156e-05, 'epoch': 3.62}) +[2026-06-20 03:06:53,326][starVLA.training.train_starvla][INFO] - Step 2081, Loss: {'train/loss': 0.15405146777629852, 'train/grad_norm_pre_clip': 25.461597442626953, 'timing/data': 0.006947426125407219, 'timing/model': 2.774808442220092, 'learning_rate/qwen_vl_interface': 6.65687579369727e-06, 'learning_rate/action_model': 6.656875793697271e-05, 'epoch': 3.63}) +[2026-06-20 03:07:04,224][starVLA.training.train_starvla][INFO] - Step 2082, Loss: {'train/loss': 0.14500942826271057, 'train/grad_norm_pre_clip': 11.495637893676758, 'timing/data': 0.007279632613062859, 'timing/model': 2.770858971402049, 'learning_rate/qwen_vl_interface': 6.653966728962885e-06, 'learning_rate/action_model': 6.653966728962885e-05, 'epoch': 3.63}) +[2026-06-20 03:07:15,388][starVLA.training.train_starvla][INFO] - Step 2083, Loss: {'train/loss': 0.3090623915195465, 'train/grad_norm_pre_clip': 22.63943099975586, 'timing/data': 0.0064795613288879395, 'timing/model': 3.0331309512257576, 'learning_rate/qwen_vl_interface': 6.651057087110813e-06, 'learning_rate/action_model': 6.651057087110813e-05, 'epoch': 3.63}) +[2026-06-20 03:07:26,230][starVLA.training.train_starvla][INFO] - Step 2084, Loss: {'train/loss': 0.11525306105613708, 'train/grad_norm_pre_clip': 19.036846160888672, 'timing/data': 0.0067926570773124695, 'timing/model': 2.794223051518202, 'learning_rate/qwen_vl_interface': 6.6481468693370924e-06, 'learning_rate/action_model': 6.648146869337091e-05, 'epoch': 3.63}) +[2026-06-20 03:07:37,019][starVLA.training.train_starvla][INFO] - Step 2085, Loss: {'train/loss': 0.11429524421691895, 'train/grad_norm_pre_clip': 17.740083694458008, 'timing/data': 0.006490468978881836, 'timing/model': 2.739966321736574, 'learning_rate/qwen_vl_interface': 6.645236076838009e-06, 'learning_rate/action_model': 6.645236076838009e-05, 'epoch': 3.63}) +[2026-06-20 03:07:47,756][starVLA.training.train_starvla][INFO] - Step 2086, Loss: {'train/loss': 0.0906221866607666, 'train/grad_norm_pre_clip': 16.751867294311523, 'timing/data': 0.006455998867750168, 'timing/model': 2.687077507376671, 'learning_rate/qwen_vl_interface': 6.642324710810075e-06, 'learning_rate/action_model': 6.642324710810076e-05, 'epoch': 3.63}) +[2026-06-20 03:07:58,831][starVLA.training.train_starvla][INFO] - Step 2087, Loss: {'train/loss': 0.44655081629753113, 'train/grad_norm_pre_clip': 31.100875854492188, 'timing/data': 0.00656706839799881, 'timing/model': 2.7643247973173857, 'learning_rate/qwen_vl_interface': 6.639412772450047e-06, 'learning_rate/action_model': 6.639412772450047e-05, 'epoch': 3.64}) +[2026-06-20 03:08:09,722][starVLA.training.train_starvla][INFO] - Step 2088, Loss: {'train/loss': 0.2548808157444, 'train/grad_norm_pre_clip': 19.796628952026367, 'timing/data': 0.006413906812667847, 'timing/model': 2.766797633841634, 'learning_rate/qwen_vl_interface': 6.6365002629549104e-06, 'learning_rate/action_model': 6.63650026295491e-05, 'epoch': 3.64}) +[2026-06-20 03:08:20,533][starVLA.training.train_starvla][INFO] - Step 2089, Loss: {'train/loss': 0.03374338895082474, 'train/grad_norm_pre_clip': 10.216595649719238, 'timing/data': 0.006547696888446808, 'timing/model': 2.7691588327288628, 'learning_rate/qwen_vl_interface': 6.633587183521887e-06, 'learning_rate/action_model': 6.633587183521887e-05, 'epoch': 3.64}) +[2026-06-20 03:08:31,635][starVLA.training.train_starvla][INFO] - Step 2090, Loss: {'train/loss': 0.09784262627363205, 'train/grad_norm_pre_clip': 13.228099822998047, 'timing/data': 0.006736963987350464, 'timing/model': 2.7845102455466986, 'learning_rate/qwen_vl_interface': 6.6306735353484365e-06, 'learning_rate/action_model': 6.630673535348436e-05, 'epoch': 3.64}) +[2026-06-20 03:08:42,741][starVLA.training.train_starvla][INFO] - Step 2091, Loss: {'train/loss': 0.02609347179532051, 'train/grad_norm_pre_clip': 5.105035781860352, 'timing/data': 0.006393188610672951, 'timing/model': 2.7808568347245455, 'learning_rate/qwen_vl_interface': 6.627759319632246e-06, 'learning_rate/action_model': 6.627759319632247e-05, 'epoch': 3.64}) +[2026-06-20 03:08:53,632][starVLA.training.train_starvla][INFO] - Step 2092, Loss: {'train/loss': 0.011884383857250214, 'train/grad_norm_pre_clip': 1.8911560773849487, 'timing/data': 0.006482848897576332, 'timing/model': 2.7799791526049376, 'learning_rate/qwen_vl_interface': 6.624844537571245e-06, 'learning_rate/action_model': 6.624844537571245e-05, 'epoch': 3.64}) +[2026-06-20 03:09:04,475][starVLA.training.train_starvla][INFO] - Step 2093, Loss: {'train/loss': 0.10670223087072372, 'train/grad_norm_pre_clip': 20.53021240234375, 'timing/data': 0.00654245913028717, 'timing/model': 2.7171008866280317, 'learning_rate/qwen_vl_interface': 6.621929190363585e-06, 'learning_rate/action_model': 6.621929190363585e-05, 'epoch': 3.65}) +[2026-06-20 03:09:15,578][starVLA.training.train_starvla][INFO] - Step 2094, Loss: {'train/loss': 0.18081551790237427, 'train/grad_norm_pre_clip': 19.92546844482422, 'timing/data': 0.0067208074033260345, 'timing/model': 3.0358771588653326, 'learning_rate/qwen_vl_interface': 6.619013279207657e-06, 'learning_rate/action_model': 6.619013279207656e-05, 'epoch': 3.65}) +[2026-06-20 03:09:26,421][starVLA.training.train_starvla][INFO] - Step 2095, Loss: {'train/loss': 0.22676925361156464, 'train/grad_norm_pre_clip': 15.90440559387207, 'timing/data': 0.006470849737524986, 'timing/model': 2.7710988614708185, 'learning_rate/qwen_vl_interface': 6.616096805302083e-06, 'learning_rate/action_model': 6.616096805302082e-05, 'epoch': 3.65}) +[2026-06-20 03:09:37,212][starVLA.training.train_starvla][INFO] - Step 2096, Loss: {'train/loss': 0.14599892497062683, 'train/grad_norm_pre_clip': 20.438100814819336, 'timing/data': 0.006570205092430115, 'timing/model': 2.7473954372107983, 'learning_rate/qwen_vl_interface': 6.6131797698457135e-06, 'learning_rate/action_model': 6.613179769845713e-05, 'epoch': 3.65}) +[2026-06-20 03:09:48,028][starVLA.training.train_starvla][INFO] - Step 2097, Loss: {'train/loss': 0.07399947941303253, 'train/grad_norm_pre_clip': 20.45889663696289, 'timing/data': 0.0065875668078660965, 'timing/model': 2.7806056942790747, 'learning_rate/qwen_vl_interface': 6.610262174037633e-06, 'learning_rate/action_model': 6.610262174037633e-05, 'epoch': 3.65}) +[2026-06-20 03:09:58,936][starVLA.training.train_starvla][INFO] - Step 2098, Loss: {'train/loss': 0.10207931697368622, 'train/grad_norm_pre_clip': 8.541787147521973, 'timing/data': 0.006736954674124718, 'timing/model': 2.7764749005436897, 'learning_rate/qwen_vl_interface': 6.607344019077155e-06, 'learning_rate/action_model': 6.607344019077156e-05, 'epoch': 3.65}) +[2026-06-20 03:10:09,824][starVLA.training.train_starvla][INFO] - Step 2099, Loss: {'train/loss': 0.22565972805023193, 'train/grad_norm_pre_clip': 20.93874740600586, 'timing/data': 0.00667402520775795, 'timing/model': 2.776000641286373, 'learning_rate/qwen_vl_interface': 6.604425306163822e-06, 'learning_rate/action_model': 6.604425306163822e-05, 'epoch': 3.66}) +[2026-06-20 03:10:20,626][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 2100 +[2026-06-20 03:17:54,598][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 2100 +[2026-06-20 03:17:54,599][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 2100 +[2026-06-20 03:26:02,917][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 2100 +[2026-06-20 03:26:02,918][starVLA.training.train_starvla][INFO] - Step 2100, Loss: {'train/loss': 0.2849810719490051, 'train/grad_norm_pre_clip': 16.499683380126953, 'eval/loss': 0.8111832141876221, 'eval/latency_0/loss': 0.7412520051002502, 'eval/latency_1/loss': 0.9115340709686279, 'eval/latency_2/loss': 0.13106027245521545, 'eval/latency_3/loss': 0.9135148525238037, 'eval/latency_4/loss': 1.3993126153945923, 'eval/flappy/loss': 0.8111832141876221, 'eval/flappy/latency_0/loss': 0.7412520051002502, 'eval/flappy/latency_1/loss': 0.9115340709686279, 'eval/flappy/latency_2/loss': 0.13106027245521545, 'eval/flappy/latency_3/loss': 0.9135148525238037, 'eval/flappy/latency_4/loss': 1.3993126153945923, 'eval/flap_precision': 0.84375, 'eval/flap_recall': 0.29454545454545455, 'eval/flap_f1': 0.43665768194070087, 'eval/flap_rate_model': 0.02666666666666667, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.4366576819407009, 'eval/latency_0/flap_precision': 0.84375, 'eval/latency_0/flap_recall': 0.29454545454545455, 'eval/latency_0/flap_f1': 0.43665768194070087, 'eval/latency_0/flap_rate_model': 0.02666666666666667, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.4366576819407009, 'timing/data': 0.0066908858716487885, 'timing/model': 2.7342790123075247, 'learning_rate/qwen_vl_interface': 6.6015060364974045e-06, 'learning_rate/action_model': 6.601506036497404e-05, 'epoch': 3.66}) +[2026-06-20 03:26:02,919][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_2100_state +[2026-06-20 03:26:02,939][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-20 03:26:08,659][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_2100_state/model.safetensors +[2026-06-20 03:26:22,765][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_2100_state/optimizer.bin +[2026-06-20 03:26:22,766][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_2100_state/sampler.bin +[2026-06-20 03:26:22,766][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_2100_state/sampler_1.bin +[2026-06-20 03:26:22,767][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_2100_state/random_states_0.pkl +[2026-06-20 03:26:22,890][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2100 +[2026-06-20 03:26:22,891][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-20 03:26:22,904][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-20 03:26:33,719][starVLA.training.train_starvla][INFO] - Step 2101, Loss: {'train/loss': 0.2629120945930481, 'train/grad_norm_pre_clip': 19.452537536621094, 'timing/data': 0.006789524108171463, 'timing/model': 2.7770158778876066, 'learning_rate/qwen_vl_interface': 6.598586211277907e-06, 'learning_rate/action_model': 6.598586211277907e-05, 'epoch': 3.66}) +[2026-06-20 03:26:44,715][starVLA.training.train_starvla][INFO] - Step 2102, Loss: {'train/loss': 0.1434406340122223, 'train/grad_norm_pre_clip': 13.721574783325195, 'timing/data': 0.009454846382141113, 'timing/model': 2.867719389498234, 'learning_rate/qwen_vl_interface': 6.59566583170556e-06, 'learning_rate/action_model': 6.59566583170556e-05, 'epoch': 3.66}) +[2026-06-20 03:26:55,523][starVLA.training.train_starvla][INFO] - Step 2103, Loss: {'train/loss': 0.1322384774684906, 'train/grad_norm_pre_clip': 14.067678451538086, 'timing/data': 0.006792942062020302, 'timing/model': 2.772768095135689, 'learning_rate/qwen_vl_interface': 6.59274489898082e-06, 'learning_rate/action_model': 6.592744898980819e-05, 'epoch': 3.66}) +[2026-06-20 03:27:06,429][starVLA.training.train_starvla][INFO] - Step 2104, Loss: {'train/loss': 0.049104832112789154, 'train/grad_norm_pre_clip': 6.943136692047119, 'timing/data': 0.0068083517253398895, 'timing/model': 2.777483630925417, 'learning_rate/qwen_vl_interface': 6.5898234143043716e-06, 'learning_rate/action_model': 6.589823414304371e-05, 'epoch': 3.67}) +[2026-06-20 03:27:17,326][starVLA.training.train_starvla][INFO] - Step 2105, Loss: {'train/loss': 0.07916465401649475, 'train/grad_norm_pre_clip': 6.242753505706787, 'timing/data': 0.0067119114100933075, 'timing/model': 2.78407278098166, 'learning_rate/qwen_vl_interface': 6.5869013788771264e-06, 'learning_rate/action_model': 6.586901378877126e-05, 'epoch': 3.67}) +[2026-06-20 03:27:28,226][starVLA.training.train_starvla][INFO] - Step 2106, Loss: {'train/loss': 0.13542549312114716, 'train/grad_norm_pre_clip': 11.5911226272583, 'timing/data': 0.007004180923104286, 'timing/model': 2.770827289670706, 'learning_rate/qwen_vl_interface': 6.583978793900223e-06, 'learning_rate/action_model': 6.583978793900223e-05, 'epoch': 3.67}) +[2026-06-20 03:27:39,129][starVLA.training.train_starvla][INFO] - Step 2107, Loss: {'train/loss': 0.13253679871559143, 'train/grad_norm_pre_clip': 12.276402473449707, 'timing/data': 0.006415726616978645, 'timing/model': 2.7780886627733707, 'learning_rate/qwen_vl_interface': 6.581055660575028e-06, 'learning_rate/action_model': 6.581055660575027e-05, 'epoch': 3.67}) +[2026-06-20 03:27:50,229][starVLA.training.train_starvla][INFO] - Step 2108, Loss: {'train/loss': 0.16294708847999573, 'train/grad_norm_pre_clip': 19.795942306518555, 'timing/data': 0.006712963804602623, 'timing/model': 2.9799644947052, 'learning_rate/qwen_vl_interface': 6.578131980103128e-06, 'learning_rate/action_model': 6.578131980103128e-05, 'epoch': 3.67}) +[2026-06-20 03:28:00,750][starVLA.training.train_starvla][INFO] - Step 2109, Loss: {'train/loss': 0.05129053816199303, 'train/grad_norm_pre_clip': 4.4204840660095215, 'timing/data': 0.006639473140239716, 'timing/model': 2.6766087375581264, 'learning_rate/qwen_vl_interface': 6.5752077536863404e-06, 'learning_rate/action_model': 6.57520775368634e-05, 'epoch': 3.67}) +[2026-06-20 03:28:11,819][starVLA.training.train_starvla][INFO] - Step 2110, Loss: {'train/loss': 0.321296364068985, 'train/grad_norm_pre_clip': 19.867759704589844, 'timing/data': 0.006526485085487366, 'timing/model': 2.7589281145483255, 'learning_rate/qwen_vl_interface': 6.572282982526702e-06, 'learning_rate/action_model': 6.572282982526702e-05, 'epoch': 3.68}) +[2026-06-20 03:28:22,727][starVLA.training.train_starvla][INFO] - Step 2111, Loss: {'train/loss': 0.17133566737174988, 'train/grad_norm_pre_clip': 19.904052734375, 'timing/data': 0.0065750852227211, 'timing/model': 2.775150103494525, 'learning_rate/qwen_vl_interface': 6.5693576678264755e-06, 'learning_rate/action_model': 6.569357667826476e-05, 'epoch': 3.68}) +[2026-06-20 03:28:33,619][starVLA.training.train_starvla][INFO] - Step 2112, Loss: {'train/loss': 0.08129531145095825, 'train/grad_norm_pre_clip': 13.808307647705078, 'timing/data': 0.006714534014463425, 'timing/model': 2.7600972428917885, 'learning_rate/qwen_vl_interface': 6.566431810788148e-06, 'learning_rate/action_model': 6.566431810788148e-05, 'epoch': 3.68}) +[2026-06-20 03:28:44,532][starVLA.training.train_starvla][INFO] - Step 2113, Loss: {'train/loss': 0.31568294763565063, 'train/grad_norm_pre_clip': 26.148550033569336, 'timing/data': 0.006576744839549065, 'timing/model': 2.7800523675978184, 'learning_rate/qwen_vl_interface': 6.5635054126144306e-06, 'learning_rate/action_model': 6.56350541261443e-05, 'epoch': 3.68}) +[2026-06-20 03:28:55,534][starVLA.training.train_starvla][INFO] - Step 2114, Loss: {'train/loss': 0.12884435057640076, 'train/grad_norm_pre_clip': 13.247220039367676, 'timing/data': 0.006755081936717033, 'timing/model': 2.7789074759930372, 'learning_rate/qwen_vl_interface': 6.560578474508254e-06, 'learning_rate/action_model': 6.560578474508253e-05, 'epoch': 3.68}) +[2026-06-20 03:29:06,430][starVLA.training.train_starvla][INFO] - Step 2115, Loss: {'train/loss': 0.03972427919507027, 'train/grad_norm_pre_clip': 4.0478434562683105, 'timing/data': 0.0067911092191934586, 'timing/model': 2.7747416887432337, 'learning_rate/qwen_vl_interface': 6.5576509976727685e-06, 'learning_rate/action_model': 6.557650997672769e-05, 'epoch': 3.68}) +[2026-06-20 03:29:17,328][starVLA.training.train_starvla][INFO] - Step 2116, Loss: {'train/loss': 0.1448691040277481, 'train/grad_norm_pre_clip': 9.544382095336914, 'timing/data': 0.00659436360001564, 'timing/model': 2.772132957354188, 'learning_rate/qwen_vl_interface': 6.554722983311354e-06, 'learning_rate/action_model': 6.554722983311355e-05, 'epoch': 3.69}) +[2026-06-20 03:29:28,138][starVLA.training.train_starvla][INFO] - Step 2117, Loss: {'train/loss': 0.07973352819681168, 'train/grad_norm_pre_clip': 9.858851432800293, 'timing/data': 0.006825853139162064, 'timing/model': 2.7196264434605837, 'learning_rate/qwen_vl_interface': 6.551794432627605e-06, 'learning_rate/action_model': 6.551794432627605e-05, 'epoch': 3.69}) +[2026-06-20 03:29:38,744][starVLA.training.train_starvla][INFO] - Step 2118, Loss: {'train/loss': 0.04031120613217354, 'train/grad_norm_pre_clip': 4.556032180786133, 'timing/data': 0.006751002743840218, 'timing/model': 2.720447802916169, 'learning_rate/qwen_vl_interface': 6.548865346825338e-06, 'learning_rate/action_model': 6.548865346825338e-05, 'epoch': 3.69}) +[2026-06-20 03:29:49,379][starVLA.training.train_starvla][INFO] - Step 2119, Loss: {'train/loss': 0.062376365065574646, 'train/grad_norm_pre_clip': 8.080179214477539, 'timing/data': 0.006524752825498581, 'timing/model': 2.6705034263432026, 'learning_rate/qwen_vl_interface': 6.545935727108589e-06, 'learning_rate/action_model': 6.545935727108588e-05, 'epoch': 3.69}) +[2026-06-20 03:29:59,971][starVLA.training.train_starvla][INFO] - Step 2120, Loss: {'train/loss': 0.08823563903570175, 'train/grad_norm_pre_clip': 8.178947448730469, 'timing/data': 0.006692012771964073, 'timing/model': 2.693321343511343, 'learning_rate/qwen_vl_interface': 6.543005574681613e-06, 'learning_rate/action_model': 6.543005574681613e-05, 'epoch': 3.69}) +[2026-06-20 03:30:10,635][starVLA.training.train_starvla][INFO] - Step 2121, Loss: {'train/loss': 0.09643101692199707, 'train/grad_norm_pre_clip': 11.76615047454834, 'timing/data': 0.006442602723836899, 'timing/model': 2.7180968169122934, 'learning_rate/qwen_vl_interface': 6.5400748907488886e-06, 'learning_rate/action_model': 6.540074890748888e-05, 'epoch': 3.69}) +[2026-06-20 03:30:21,451][starVLA.training.train_starvla][INFO] - Step 2122, Loss: {'train/loss': 0.06148651987314224, 'train/grad_norm_pre_clip': 6.047503471374512, 'timing/data': 0.0067067015916109085, 'timing/model': 2.697457604110241, 'learning_rate/qwen_vl_interface': 6.537143676515105e-06, 'learning_rate/action_model': 6.537143676515105e-05, 'epoch': 3.7}) +[2026-06-20 03:30:32,317][starVLA.training.train_starvla][INFO] - Step 2123, Loss: {'train/loss': 0.15411794185638428, 'train/grad_norm_pre_clip': 18.28849220275879, 'timing/data': 0.006470033898949623, 'timing/model': 2.7640001364052296, 'learning_rate/qwen_vl_interface': 6.534211933185176e-06, 'learning_rate/action_model': 6.534211933185176e-05, 'epoch': 3.7}) +[2026-06-20 03:30:43,218][starVLA.training.train_starvla][INFO] - Step 2124, Loss: {'train/loss': 0.11561821401119232, 'train/grad_norm_pre_clip': 20.309667587280273, 'timing/data': 0.006772181019186974, 'timing/model': 2.7650317028164864, 'learning_rate/qwen_vl_interface': 6.531279661964231e-06, 'learning_rate/action_model': 6.531279661964231e-05, 'epoch': 3.7}) +[2026-06-20 03:30:54,231][starVLA.training.train_starvla][INFO] - Step 2125, Loss: {'train/loss': 0.1932513415813446, 'train/grad_norm_pre_clip': 15.80919075012207, 'timing/data': 0.0068858228623867035, 'timing/model': 2.8879132121801376, 'learning_rate/qwen_vl_interface': 6.528346864057616e-06, 'learning_rate/action_model': 6.528346864057615e-05, 'epoch': 3.7}) +[2026-06-20 03:31:05,128][starVLA.training.train_starvla][INFO] - Step 2126, Loss: {'train/loss': 0.08805480599403381, 'train/grad_norm_pre_clip': 8.555890083312988, 'timing/data': 0.007037008181214333, 'timing/model': 2.7796628642827272, 'learning_rate/qwen_vl_interface': 6.5254135406708906e-06, 'learning_rate/action_model': 6.52541354067089e-05, 'epoch': 3.7}) +[2026-06-20 03:31:16,036][starVLA.training.train_starvla][INFO] - Step 2127, Loss: {'train/loss': 0.3853214681148529, 'train/grad_norm_pre_clip': 20.27421760559082, 'timing/data': 0.006842900067567825, 'timing/model': 2.7865866906940937, 'learning_rate/qwen_vl_interface': 6.522479693009836e-06, 'learning_rate/action_model': 6.522479693009837e-05, 'epoch': 3.71}) +[2026-06-20 03:31:26,933][starVLA.training.train_starvla][INFO] - Step 2128, Loss: {'train/loss': 0.5754218697547913, 'train/grad_norm_pre_clip': 34.9937858581543, 'timing/data': 0.0067754220217466354, 'timing/model': 2.7729286272078753, 'learning_rate/qwen_vl_interface': 6.519545322280446e-06, 'learning_rate/action_model': 6.519545322280446e-05, 'epoch': 3.71}) +[2026-06-20 03:31:37,825][starVLA.training.train_starvla][INFO] - Step 2129, Loss: {'train/loss': 0.17846226692199707, 'train/grad_norm_pre_clip': 16.044239044189453, 'timing/data': 0.007639611139893532, 'timing/model': 2.715163415297866, 'learning_rate/qwen_vl_interface': 6.5166104296889275e-06, 'learning_rate/action_model': 6.516610429688927e-05, 'epoch': 3.71}) +[2026-06-20 03:31:48,734][starVLA.training.train_starvla][INFO] - Step 2130, Loss: {'train/loss': 0.027533821761608124, 'train/grad_norm_pre_clip': 5.1684250831604, 'timing/data': 0.006932999938726425, 'timing/model': 2.77936845459044, 'learning_rate/qwen_vl_interface': 6.513675016441707e-06, 'learning_rate/action_model': 6.513675016441707e-05, 'epoch': 3.71}) +[2026-06-20 03:31:59,630][starVLA.training.train_starvla][INFO] - Step 2131, Loss: {'train/loss': 0.067172110080719, 'train/grad_norm_pre_clip': 15.690053939819336, 'timing/data': 0.006837211549282074, 'timing/model': 2.778632650151849, 'learning_rate/qwen_vl_interface': 6.510739083745422e-06, 'learning_rate/action_model': 6.510739083745422e-05, 'epoch': 3.71}) +[2026-06-20 03:32:10,537][starVLA.training.train_starvla][INFO] - Step 2132, Loss: {'train/loss': 0.0946057140827179, 'train/grad_norm_pre_clip': 18.55881118774414, 'timing/data': 0.011051541194319725, 'timing/model': 2.774297023192048, 'learning_rate/qwen_vl_interface': 6.507802632806919e-06, 'learning_rate/action_model': 6.507802632806919e-05, 'epoch': 3.71}) +[2026-06-20 03:32:21,434][starVLA.training.train_starvla][INFO] - Step 2133, Loss: {'train/loss': 0.013017714023590088, 'train/grad_norm_pre_clip': 2.0630297660827637, 'timing/data': 0.008272029459476471, 'timing/model': 2.7748613245785236, 'learning_rate/qwen_vl_interface': 6.5048656648332646e-06, 'learning_rate/action_model': 6.504865664833264e-05, 'epoch': 3.72}) +[2026-06-20 03:32:32,330][starVLA.training.train_starvla][INFO] - Step 2134, Loss: {'train/loss': 0.24276374280452728, 'train/grad_norm_pre_clip': 18.443588256835938, 'timing/data': 0.006941311061382294, 'timing/model': 2.774234227836132, 'learning_rate/qwen_vl_interface': 6.501928181031737e-06, 'learning_rate/action_model': 6.501928181031737e-05, 'epoch': 3.72}) +[2026-06-20 03:32:43,234][starVLA.training.train_starvla][INFO] - Step 2135, Loss: {'train/loss': 0.27539873123168945, 'train/grad_norm_pre_clip': 24.25104522705078, 'timing/data': 0.006578294560313225, 'timing/model': 2.7753221932798624, 'learning_rate/qwen_vl_interface': 6.498990182609823e-06, 'learning_rate/action_model': 6.498990182609823e-05, 'epoch': 3.72}) +[2026-06-20 03:32:54,135][starVLA.training.train_starvla][INFO] - Step 2136, Loss: {'train/loss': 0.18176890909671783, 'train/grad_norm_pre_clip': 17.6603946685791, 'timing/data': 0.006626792252063751, 'timing/model': 2.7785799093544483, 'learning_rate/qwen_vl_interface': 6.4960516707752255e-06, 'learning_rate/action_model': 6.496051670775226e-05, 'epoch': 3.72}) +[2026-06-20 03:33:05,135][starVLA.training.train_starvla][INFO] - Step 2137, Loss: {'train/loss': 0.20157186686992645, 'train/grad_norm_pre_clip': 14.410087585449219, 'timing/data': 0.007347963750362396, 'timing/model': 2.7834356389939785, 'learning_rate/qwen_vl_interface': 6.493112646735851e-06, 'learning_rate/action_model': 6.493112646735851e-05, 'epoch': 3.72}) +[2026-06-20 03:33:16,030][starVLA.training.train_starvla][INFO] - Step 2138, Loss: {'train/loss': 0.012411218136548996, 'train/grad_norm_pre_clip': 1.866410493850708, 'timing/data': 0.006705552339553833, 'timing/model': 2.7772949002683163, 'learning_rate/qwen_vl_interface': 6.490173111699824e-06, 'learning_rate/action_model': 6.490173111699825e-05, 'epoch': 3.72}) +[2026-06-20 03:33:26,929][starVLA.training.train_starvla][INFO] - Step 2139, Loss: {'train/loss': 0.5839923620223999, 'train/grad_norm_pre_clip': 27.609107971191406, 'timing/data': 0.006618114188313484, 'timing/model': 2.772883800789714, 'learning_rate/qwen_vl_interface': 6.487233066875479e-06, 'learning_rate/action_model': 6.487233066875479e-05, 'epoch': 3.73}) +[2026-06-20 03:33:37,835][starVLA.training.train_starvla][INFO] - Step 2140, Loss: {'train/loss': 0.11697833240032196, 'train/grad_norm_pre_clip': 15.374773979187012, 'timing/data': 0.006599225103855133, 'timing/model': 2.783594239503145, 'learning_rate/qwen_vl_interface': 6.484292513471352e-06, 'learning_rate/action_model': 6.484292513471352e-05, 'epoch': 3.73}) +[2026-06-20 03:33:48,731][starVLA.training.train_starvla][INFO] - Step 2141, Loss: {'train/loss': 0.30546560883522034, 'train/grad_norm_pre_clip': 18.807594299316406, 'timing/data': 0.006396804004907608, 'timing/model': 2.787929095327854, 'learning_rate/qwen_vl_interface': 6.4813514526961975e-06, 'learning_rate/action_model': 6.481351452696198e-05, 'epoch': 3.73}) +[2026-06-20 03:33:59,630][starVLA.training.train_starvla][INFO] - Step 2142, Loss: {'train/loss': 0.20643654465675354, 'train/grad_norm_pre_clip': 21.458803176879883, 'timing/data': 0.007097907364368439, 'timing/model': 2.7837721332907677, 'learning_rate/qwen_vl_interface': 6.478409885758972e-06, 'learning_rate/action_model': 6.478409885758971e-05, 'epoch': 3.73}) +[2026-06-20 03:34:10,534][starVLA.training.train_starvla][INFO] - Step 2143, Loss: {'train/loss': 0.16633319854736328, 'train/grad_norm_pre_clip': 16.906883239746094, 'timing/data': 0.0067055243998765945, 'timing/model': 2.778615104034543, 'learning_rate/qwen_vl_interface': 6.475467813868844e-06, 'learning_rate/action_model': 6.475467813868844e-05, 'epoch': 3.73}) +[2026-06-20 03:34:21,437][starVLA.training.train_starvla][INFO] - Step 2144, Loss: {'train/loss': 0.10118145495653152, 'train/grad_norm_pre_clip': 9.120614051818848, 'timing/data': 0.0066380929201841354, 'timing/model': 2.7714267931878567, 'learning_rate/qwen_vl_interface': 6.472525238235187e-06, 'learning_rate/action_model': 6.472525238235186e-05, 'epoch': 3.74}) +[2026-06-20 03:34:32,533][starVLA.training.train_starvla][INFO] - Step 2145, Loss: {'train/loss': 0.23312167823314667, 'train/grad_norm_pre_clip': 15.806109428405762, 'timing/data': 0.006811739876866341, 'timing/model': 2.7769965045154095, 'learning_rate/qwen_vl_interface': 6.4695821600675834e-06, 'learning_rate/action_model': 6.469582160067583e-05, 'epoch': 3.74}) +[2026-06-20 03:34:43,323][starVLA.training.train_starvla][INFO] - Step 2146, Loss: {'train/loss': 0.13943539559841156, 'train/grad_norm_pre_clip': 13.23800277709961, 'timing/data': 0.0067070722579956055, 'timing/model': 2.7788782864809036, 'learning_rate/qwen_vl_interface': 6.466638580575824e-06, 'learning_rate/action_model': 6.466638580575824e-05, 'epoch': 3.74}) +[2026-06-20 03:34:54,226][starVLA.training.train_starvla][INFO] - Step 2147, Loss: {'train/loss': 0.15683449804782867, 'train/grad_norm_pre_clip': 12.314726829528809, 'timing/data': 0.0066214147955179214, 'timing/model': 2.7722248025238514, 'learning_rate/qwen_vl_interface': 6.463694500969898e-06, 'learning_rate/action_model': 6.463694500969898e-05, 'epoch': 3.74}) +[2026-06-20 03:35:05,134][starVLA.training.train_starvla][INFO] - Step 2148, Loss: {'train/loss': 0.0340433344244957, 'train/grad_norm_pre_clip': 4.818573951721191, 'timing/data': 0.006770901381969452, 'timing/model': 2.7726496253162622, 'learning_rate/qwen_vl_interface': 6.460749922460008e-06, 'learning_rate/action_model': 6.460749922460008e-05, 'epoch': 3.74}) +[2026-06-20 03:35:16,036][starVLA.training.train_starvla][INFO] - Step 2149, Loss: {'train/loss': 0.10832097381353378, 'train/grad_norm_pre_clip': 9.556519508361816, 'timing/data': 0.006760852411389351, 'timing/model': 2.7802547980099916, 'learning_rate/qwen_vl_interface': 6.457804846256559e-06, 'learning_rate/action_model': 6.457804846256559e-05, 'epoch': 3.74}) +[2026-06-20 03:35:26,927][starVLA.training.train_starvla][INFO] - Step 2150, Loss: {'train/loss': 0.0876900777220726, 'train/grad_norm_pre_clip': 12.266098022460938, 'timing/data': 0.006766309961676598, 'timing/model': 2.7791193556040525, 'learning_rate/qwen_vl_interface': 6.45485927357016e-06, 'learning_rate/action_model': 6.454859273570159e-05, 'epoch': 3.75}) +[2026-06-20 03:35:37,835][starVLA.training.train_starvla][INFO] - Step 2151, Loss: {'train/loss': 0.272504985332489, 'train/grad_norm_pre_clip': 15.638129234313965, 'timing/data': 0.0069213081151247025, 'timing/model': 2.7761194556951523, 'learning_rate/qwen_vl_interface': 6.451913205611626e-06, 'learning_rate/action_model': 6.451913205611626e-05, 'epoch': 3.75}) +[2026-06-20 03:35:48,868][starVLA.training.train_starvla][INFO] - Step 2152, Loss: {'train/loss': 0.06653693318367004, 'train/grad_norm_pre_clip': 8.322014808654785, 'timing/data': 0.006596185266971588, 'timing/model': 2.9094027187675238, 'learning_rate/qwen_vl_interface': 6.448966643591974e-06, 'learning_rate/action_model': 6.448966643591974e-05, 'epoch': 3.75}) +[2026-06-20 03:35:59,721][starVLA.training.train_starvla][INFO] - Step 2153, Loss: {'train/loss': 0.14188389480113983, 'train/grad_norm_pre_clip': 13.221331596374512, 'timing/data': 0.006796324625611305, 'timing/model': 2.7736322209239006, 'learning_rate/qwen_vl_interface': 6.446019588722423e-06, 'learning_rate/action_model': 6.446019588722422e-05, 'epoch': 3.75}) +[2026-06-20 03:36:10,459][starVLA.training.train_starvla][INFO] - Step 2154, Loss: {'train/loss': 0.21422308683395386, 'train/grad_norm_pre_clip': 13.522614479064941, 'timing/data': 0.006668547168374062, 'timing/model': 2.6507758889347315, 'learning_rate/qwen_vl_interface': 6.443072042214398e-06, 'learning_rate/action_model': 6.443072042214397e-05, 'epoch': 3.75}) +[2026-06-20 03:36:21,323][starVLA.training.train_starvla][INFO] - Step 2155, Loss: {'train/loss': 0.23792575299739838, 'train/grad_norm_pre_clip': 21.963930130004883, 'timing/data': 0.006893735378980637, 'timing/model': 2.7783897686749697, 'learning_rate/qwen_vl_interface': 6.440124005279521e-06, 'learning_rate/action_model': 6.440124005279521e-05, 'epoch': 3.75}) +[2026-06-20 03:36:32,437][starVLA.training.train_starvla][INFO] - Step 2156, Loss: {'train/loss': 0.234098881483078, 'train/grad_norm_pre_clip': 17.20138168334961, 'timing/data': 0.006736645475029945, 'timing/model': 2.7425351347774267, 'learning_rate/qwen_vl_interface': 6.43717547912962e-06, 'learning_rate/action_model': 6.437175479129621e-05, 'epoch': 3.76}) +[2026-06-20 03:36:43,335][starVLA.training.train_starvla][INFO] - Step 2157, Loss: {'train/loss': 0.2266838401556015, 'train/grad_norm_pre_clip': 8.785462379455566, 'timing/data': 0.006869485601782799, 'timing/model': 2.7787340246140957, 'learning_rate/qwen_vl_interface': 6.434226464976725e-06, 'learning_rate/action_model': 6.434226464976724e-05, 'epoch': 3.76}) +[2026-06-20 03:36:54,234][starVLA.training.train_starvla][INFO] - Step 2158, Loss: {'train/loss': 0.14965257048606873, 'train/grad_norm_pre_clip': 11.769576072692871, 'timing/data': 0.00686364620923996, 'timing/model': 2.773648042231798, 'learning_rate/qwen_vl_interface': 6.431276964033062e-06, 'learning_rate/action_model': 6.431276964033062e-05, 'epoch': 3.76}) +[2026-06-20 03:37:05,133][starVLA.training.train_starvla][INFO] - Step 2159, Loss: {'train/loss': 0.3633001744747162, 'train/grad_norm_pre_clip': 14.859750747680664, 'timing/data': 0.006667489185929298, 'timing/model': 2.777055637910962, 'learning_rate/qwen_vl_interface': 6.428326977511059e-06, 'learning_rate/action_model': 6.428326977511058e-05, 'epoch': 3.76}) +[2026-06-20 03:37:16,225][starVLA.training.train_starvla][INFO] - Step 2160, Loss: {'train/loss': 0.21008484065532684, 'train/grad_norm_pre_clip': 14.632917404174805, 'timing/data': 0.0066284313797950745, 'timing/model': 2.7776266038417816, 'learning_rate/qwen_vl_interface': 6.425376506623344e-06, 'learning_rate/action_model': 6.425376506623344e-05, 'epoch': 3.76}) +[2026-06-20 03:37:27,126][starVLA.training.train_starvla][INFO] - Step 2161, Loss: {'train/loss': 0.14532668888568878, 'train/grad_norm_pre_clip': 13.564667701721191, 'timing/data': 0.006706638261675835, 'timing/model': 2.7718198653310537, 'learning_rate/qwen_vl_interface': 6.422425552582747e-06, 'learning_rate/action_model': 6.422425552582747e-05, 'epoch': 3.76}) +[2026-06-20 03:37:38,024][starVLA.training.train_starvla][INFO] - Step 2162, Loss: {'train/loss': 0.2254127860069275, 'train/grad_norm_pre_clip': 12.382807731628418, 'timing/data': 0.006667228415608406, 'timing/model': 2.7771326191723347, 'learning_rate/qwen_vl_interface': 6.41947411660229e-06, 'learning_rate/action_model': 6.41947411660229e-05, 'epoch': 3.77}) +[2026-06-20 03:37:48,936][starVLA.training.train_starvla][INFO] - Step 2163, Loss: {'train/loss': 0.11926685273647308, 'train/grad_norm_pre_clip': 6.834718227386475, 'timing/data': 0.006847115233540535, 'timing/model': 2.778470726683736, 'learning_rate/qwen_vl_interface': 6.4165221998951995e-06, 'learning_rate/action_model': 6.4165221998952e-05, 'epoch': 3.77}) +[2026-06-20 03:37:59,933][starVLA.training.train_starvla][INFO] - Step 2164, Loss: {'train/loss': 0.09670160710811615, 'train/grad_norm_pre_clip': 8.999585151672363, 'timing/data': 0.006572699174284935, 'timing/model': 2.7761848401278257, 'learning_rate/qwen_vl_interface': 6.413569803674896e-06, 'learning_rate/action_model': 6.413569803674896e-05, 'epoch': 3.77}) +[2026-06-20 03:38:10,834][starVLA.training.train_starvla][INFO] - Step 2165, Loss: {'train/loss': 0.07932227849960327, 'train/grad_norm_pre_clip': 5.13422155380249, 'timing/data': 0.006775198504328728, 'timing/model': 2.7774994149804115, 'learning_rate/qwen_vl_interface': 6.410616929154997e-06, 'learning_rate/action_model': 6.410616929154996e-05, 'epoch': 3.77}) +[2026-06-20 03:38:21,734][starVLA.training.train_starvla][INFO] - Step 2166, Loss: {'train/loss': 0.1470184624195099, 'train/grad_norm_pre_clip': 10.518903732299805, 'timing/data': 0.006642298772931099, 'timing/model': 2.777672691270709, 'learning_rate/qwen_vl_interface': 6.40766357754932e-06, 'learning_rate/action_model': 6.407663577549319e-05, 'epoch': 3.77}) +[2026-06-20 03:38:32,630][starVLA.training.train_starvla][INFO] - Step 2167, Loss: {'train/loss': 0.11176218837499619, 'train/grad_norm_pre_clip': 9.08107852935791, 'timing/data': 0.006710348650813103, 'timing/model': 2.770457945764065, 'learning_rate/qwen_vl_interface': 6.404709750071875e-06, 'learning_rate/action_model': 6.404709750071874e-05, 'epoch': 3.78}) +[2026-06-20 03:38:43,533][starVLA.training.train_starvla][INFO] - Step 2168, Loss: {'train/loss': 0.07241597771644592, 'train/grad_norm_pre_clip': 4.525272369384766, 'timing/data': 0.00647728331387043, 'timing/model': 2.7781082801520824, 'learning_rate/qwen_vl_interface': 6.401755447936869e-06, 'learning_rate/action_model': 6.401755447936869e-05, 'epoch': 3.78}) +[2026-06-20 03:38:54,433][starVLA.training.train_starvla][INFO] - Step 2169, Loss: {'train/loss': 0.19746163487434387, 'train/grad_norm_pre_clip': 16.909543991088867, 'timing/data': 0.006650660187005997, 'timing/model': 2.775579361245036, 'learning_rate/qwen_vl_interface': 6.3988006723587036e-06, 'learning_rate/action_model': 6.398800672358703e-05, 'epoch': 3.78}) +[2026-06-20 03:39:05,334][starVLA.training.train_starvla][INFO] - Step 2170, Loss: {'train/loss': 0.15320812165737152, 'train/grad_norm_pre_clip': 6.697765827178955, 'timing/data': 0.006659990176558495, 'timing/model': 2.7776538878679276, 'learning_rate/qwen_vl_interface': 6.395845424551975e-06, 'learning_rate/action_model': 6.395845424551975e-05, 'epoch': 3.78}) +[2026-06-20 03:39:16,232][starVLA.training.train_starvla][INFO] - Step 2171, Loss: {'train/loss': 0.07507973909378052, 'train/grad_norm_pre_clip': 5.331212043762207, 'timing/data': 0.006768835708498955, 'timing/model': 2.7775288242846727, 'learning_rate/qwen_vl_interface': 6.392889705731475e-06, 'learning_rate/action_model': 6.392889705731476e-05, 'epoch': 3.78}) +[2026-06-20 03:39:27,135][starVLA.training.train_starvla][INFO] - Step 2172, Loss: {'train/loss': 0.24528968334197998, 'train/grad_norm_pre_clip': 12.186738967895508, 'timing/data': 0.0068885888904333115, 'timing/model': 2.7779833916574717, 'learning_rate/qwen_vl_interface': 6.389933517112188e-06, 'learning_rate/action_model': 6.389933517112187e-05, 'epoch': 3.78}) +[2026-06-20 03:39:38,032][starVLA.training.train_starvla][INFO] - Step 2173, Loss: {'train/loss': 0.2948925197124481, 'train/grad_norm_pre_clip': 19.6929931640625, 'timing/data': 0.006555140018463135, 'timing/model': 2.7732247672975063, 'learning_rate/qwen_vl_interface': 6.386976859909293e-06, 'learning_rate/action_model': 6.386976859909292e-05, 'epoch': 3.79}) +[2026-06-20 03:39:48,934][starVLA.training.train_starvla][INFO] - Step 2174, Loss: {'train/loss': 0.1382654458284378, 'train/grad_norm_pre_clip': 13.700448989868164, 'timing/data': 0.0069130174815654755, 'timing/model': 2.7797627933323383, 'learning_rate/qwen_vl_interface': 6.384019735338156e-06, 'learning_rate/action_model': 6.384019735338156e-05, 'epoch': 3.79}) +[2026-06-20 03:39:59,932][starVLA.training.train_starvla][INFO] - Step 2175, Loss: {'train/loss': 0.052395254373550415, 'train/grad_norm_pre_clip': 6.77156925201416, 'timing/data': 0.00664764828979969, 'timing/model': 2.803058188408613, 'learning_rate/qwen_vl_interface': 6.381062144614342e-06, 'learning_rate/action_model': 6.381062144614342e-05, 'epoch': 3.79}) +[2026-06-20 03:40:10,820][starVLA.training.train_starvla][INFO] - Step 2176, Loss: {'train/loss': 0.11528286337852478, 'train/grad_norm_pre_clip': 29.791349411010742, 'timing/data': 0.007034946233034134, 'timing/model': 2.7576323617249727, 'learning_rate/qwen_vl_interface': 6.378104088953605e-06, 'learning_rate/action_model': 6.378104088953605e-05, 'epoch': 3.79}) +[2026-06-20 03:40:21,734][starVLA.training.train_starvla][INFO] - Step 2177, Loss: {'train/loss': 0.2604919672012329, 'train/grad_norm_pre_clip': 21.185935974121094, 'timing/data': 0.006929235532879829, 'timing/model': 2.7762485574930906, 'learning_rate/qwen_vl_interface': 6.375145569571889e-06, 'learning_rate/action_model': 6.375145569571889e-05, 'epoch': 3.79}) +[2026-06-20 03:40:32,630][starVLA.training.train_starvla][INFO] - Step 2178, Loss: {'train/loss': 0.22767692804336548, 'train/grad_norm_pre_clip': 14.347724914550781, 'timing/data': 0.006646057590842247, 'timing/model': 2.7713314332067966, 'learning_rate/qwen_vl_interface': 6.372186587685329e-06, 'learning_rate/action_model': 6.372186587685329e-05, 'epoch': 3.79}) +[2026-06-20 03:40:43,534][starVLA.training.train_starvla][INFO] - Step 2179, Loss: {'train/loss': 0.18136952817440033, 'train/grad_norm_pre_clip': 20.848039627075195, 'timing/data': 0.00656476803123951, 'timing/model': 2.7767394464462996, 'learning_rate/qwen_vl_interface': 6.369227144510252e-06, 'learning_rate/action_model': 6.369227144510252e-05, 'epoch': 3.8}) +[2026-06-20 03:40:54,433][starVLA.training.train_starvla][INFO] - Step 2180, Loss: {'train/loss': 0.027506910264492035, 'train/grad_norm_pre_clip': 3.298217535018921, 'timing/data': 0.006672928109765053, 'timing/model': 2.783412976190448, 'learning_rate/qwen_vl_interface': 6.3662672412631736e-06, 'learning_rate/action_model': 6.366267241263173e-05, 'epoch': 3.8}) +[2026-06-20 03:41:05,331][starVLA.training.train_starvla][INFO] - Step 2181, Loss: {'train/loss': 0.14257405698299408, 'train/grad_norm_pre_clip': 8.319087982177734, 'timing/data': 0.006921717897057533, 'timing/model': 2.7743527814745903, 'learning_rate/qwen_vl_interface': 6.3633068791607954e-06, 'learning_rate/action_model': 6.363306879160795e-05, 'epoch': 3.8}) +[2026-06-20 03:41:16,234][starVLA.training.train_starvla][INFO] - Step 2182, Loss: {'train/loss': 0.08377314358949661, 'train/grad_norm_pre_clip': 6.498588562011719, 'timing/data': 0.006751619279384613, 'timing/model': 2.7746637668460608, 'learning_rate/qwen_vl_interface': 6.360346059420014e-06, 'learning_rate/action_model': 6.360346059420014e-05, 'epoch': 3.8}) +[2026-06-20 03:41:27,336][starVLA.training.train_starvla][INFO] - Step 2183, Loss: {'train/loss': 0.19768789410591125, 'train/grad_norm_pre_clip': 21.46225929260254, 'timing/data': 0.006720108911395073, 'timing/model': 2.778943222016096, 'learning_rate/qwen_vl_interface': 6.357384783257909e-06, 'learning_rate/action_model': 6.357384783257908e-05, 'epoch': 3.8}) +[2026-06-20 03:41:38,233][starVLA.training.train_starvla][INFO] - Step 2184, Loss: {'train/loss': 0.19830197095870972, 'train/grad_norm_pre_clip': 20.456966400146484, 'timing/data': 0.006846016272902489, 'timing/model': 2.7719227261841297, 'learning_rate/qwen_vl_interface': 6.354423051891748e-06, 'learning_rate/action_model': 6.354423051891749e-05, 'epoch': 3.8}) +[2026-06-20 03:41:49,125][starVLA.training.train_starvla][INFO] - Step 2185, Loss: {'train/loss': 0.206764817237854, 'train/grad_norm_pre_clip': 13.628625869750977, 'timing/data': 0.00722898356616497, 'timing/model': 2.7675489503890276, 'learning_rate/qwen_vl_interface': 6.35146086653899e-06, 'learning_rate/action_model': 6.35146086653899e-05, 'epoch': 3.81}) +[2026-06-20 03:42:00,027][starVLA.training.train_starvla][INFO] - Step 2186, Loss: {'train/loss': 0.04876933991909027, 'train/grad_norm_pre_clip': 5.9391632080078125, 'timing/data': 0.0071541927754879, 'timing/model': 2.773572213947773, 'learning_rate/qwen_vl_interface': 6.348498228417273e-06, 'learning_rate/action_model': 6.348498228417273e-05, 'epoch': 3.81}) +[2026-06-20 03:42:10,939][starVLA.training.train_starvla][INFO] - Step 2187, Loss: {'train/loss': 0.32847145199775696, 'train/grad_norm_pre_clip': 22.018009185791016, 'timing/data': 0.007036454975605011, 'timing/model': 2.7766129802912474, 'learning_rate/qwen_vl_interface': 6.34553513874443e-06, 'learning_rate/action_model': 6.34553513874443e-05, 'epoch': 3.81}) +[2026-06-20 03:42:21,837][starVLA.training.train_starvla][INFO] - Step 2188, Loss: {'train/loss': 0.06768926978111267, 'train/grad_norm_pre_clip': 13.07395076751709, 'timing/data': 0.006840420886874199, 'timing/model': 2.7810500226914883, 'learning_rate/qwen_vl_interface': 6.342571598738473e-06, 'learning_rate/action_model': 6.342571598738472e-05, 'epoch': 3.81}) +[2026-06-20 03:42:32,740][starVLA.training.train_starvla][INFO] - Step 2189, Loss: {'train/loss': 0.09824207425117493, 'train/grad_norm_pre_clip': 8.572423934936523, 'timing/data': 0.007184924557805061, 'timing/model': 2.775705862790346, 'learning_rate/qwen_vl_interface': 6.339607609617603e-06, 'learning_rate/action_model': 6.339607609617603e-05, 'epoch': 3.81}) +[2026-06-20 03:42:43,624][starVLA.training.train_starvla][INFO] - Step 2190, Loss: {'train/loss': 0.04982898011803627, 'train/grad_norm_pre_clip': 4.070724010467529, 'timing/data': 0.006867187097668648, 'timing/model': 2.7603000048547983, 'learning_rate/qwen_vl_interface': 6.336643172600202e-06, 'learning_rate/action_model': 6.336643172600202e-05, 'epoch': 3.82}) +[2026-06-20 03:42:54,639][starVLA.training.train_starvla][INFO] - Step 2191, Loss: {'train/loss': 0.3361192047595978, 'train/grad_norm_pre_clip': 18.667316436767578, 'timing/data': 0.006872009485960007, 'timing/model': 2.7771693523973227, 'learning_rate/qwen_vl_interface': 6.333678288904838e-06, 'learning_rate/action_model': 6.333678288904837e-05, 'epoch': 3.82}) +[2026-06-20 03:43:05,536][starVLA.training.train_starvla][INFO] - Step 2192, Loss: {'train/loss': 0.1671338975429535, 'train/grad_norm_pre_clip': 11.810441017150879, 'timing/data': 0.006824338808655739, 'timing/model': 2.7780580390244722, 'learning_rate/qwen_vl_interface': 6.330712959750264e-06, 'learning_rate/action_model': 6.330712959750263e-05, 'epoch': 3.82}) +[2026-06-20 03:43:16,433][starVLA.training.train_starvla][INFO] - Step 2193, Loss: {'train/loss': 0.046911388635635376, 'train/grad_norm_pre_clip': 9.248369216918945, 'timing/data': 0.006914755329489708, 'timing/model': 2.7699420005083084, 'learning_rate/qwen_vl_interface': 6.3277471863554145e-06, 'learning_rate/action_model': 6.327747186355414e-05, 'epoch': 3.82}) +[2026-06-20 03:43:27,338][starVLA.training.train_starvla][INFO] - Step 2194, Loss: {'train/loss': 0.16407153010368347, 'train/grad_norm_pre_clip': 19.785869598388672, 'timing/data': 0.007070435211062431, 'timing/model': 2.777441008016467, 'learning_rate/qwen_vl_interface': 6.324780969939406e-06, 'learning_rate/action_model': 6.324780969939405e-05, 'epoch': 3.82}) +[2026-06-20 03:43:38,236][starVLA.training.train_starvla][INFO] - Step 2195, Loss: {'train/loss': 0.09366846084594727, 'train/grad_norm_pre_clip': 7.557528972625732, 'timing/data': 0.006706269457936287, 'timing/model': 2.7781064864248037, 'learning_rate/qwen_vl_interface': 6.32181431172154e-06, 'learning_rate/action_model': 6.32181431172154e-05, 'epoch': 3.82}) +[2026-06-20 03:43:49,135][starVLA.training.train_starvla][INFO] - Step 2196, Loss: {'train/loss': 0.07592269033193588, 'train/grad_norm_pre_clip': 11.770668983459473, 'timing/data': 0.006819449365139008, 'timing/model': 2.781644871458411, 'learning_rate/qwen_vl_interface': 6.318847212921295e-06, 'learning_rate/action_model': 6.318847212921295e-05, 'epoch': 3.83}) +[2026-06-20 03:44:00,037][starVLA.training.train_starvla][INFO] - Step 2197, Loss: {'train/loss': 0.3105352818965912, 'train/grad_norm_pre_clip': 19.504093170166016, 'timing/data': 0.0066678207367658615, 'timing/model': 2.7772227935492992, 'learning_rate/qwen_vl_interface': 6.3158796747583364e-06, 'learning_rate/action_model': 6.315879674758336e-05, 'epoch': 3.83}) +[2026-06-20 03:44:11,167][starVLA.training.train_starvla][INFO] - Step 2198, Loss: {'train/loss': 0.2851191759109497, 'train/grad_norm_pre_clip': 20.26673126220703, 'timing/data': 0.006670428439974785, 'timing/model': 3.011510908603668, 'learning_rate/qwen_vl_interface': 6.312911698452503e-06, 'learning_rate/action_model': 6.312911698452502e-05, 'epoch': 3.83}) +[2026-06-20 03:44:22,034][starVLA.training.train_starvla][INFO] - Step 2199, Loss: {'train/loss': 0.26611465215682983, 'train/grad_norm_pre_clip': 27.292282104492188, 'timing/data': 0.006495123729109764, 'timing/model': 2.7726846802979708, 'learning_rate/qwen_vl_interface': 6.309943285223824e-06, 'learning_rate/action_model': 6.309943285223824e-05, 'epoch': 3.83}) +[2026-06-20 03:44:32,934][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 2200 +[2026-06-20 03:52:05,741][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 2200 +[2026-06-20 03:52:05,742][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 2200 +[2026-06-20 04:00:17,923][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 2200 +[2026-06-20 04:00:17,925][starVLA.training.train_starvla][INFO] - Step 2200, Loss: {'train/loss': 0.12370163947343826, 'train/grad_norm_pre_clip': 13.962154388427734, 'eval/loss': 0.7131090760231018, 'eval/latency_0/loss': 0.6616060733795166, 'eval/latency_1/loss': 0.7963295578956604, 'eval/latency_2/loss': 0.09046684205532074, 'eval/latency_3/loss': 0.8106390237808228, 'eval/latency_4/loss': 1.2451344728469849, 'eval/flappy/loss': 0.7131090760231018, 'eval/flappy/latency_0/loss': 0.6616060733795166, 'eval/flappy/latency_1/loss': 0.7963295578956604, 'eval/flappy/latency_2/loss': 0.09046684205532074, 'eval/flappy/latency_3/loss': 0.8106390237808228, 'eval/flappy/latency_4/loss': 1.2451344728469849, 'eval/flap_precision': 0.8222222222222222, 'eval/flap_recall': 0.2690909090909091, 'eval/flap_f1': 0.4054794520547946, 'eval/flap_rate_model': 0.025, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.4054794520547946, 'eval/latency_0/flap_precision': 0.8222222222222222, 'eval/latency_0/flap_recall': 0.2690909090909091, 'eval/latency_0/flap_f1': 0.4054794520547946, 'eval/latency_0/flap_rate_model': 0.025, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.4054794520547946, 'timing/data': 0.006635960191488266, 'timing/model': 2.7763685137033463, 'learning_rate/qwen_vl_interface': 6.306974436292493e-06, 'learning_rate/action_model': 6.306974436292494e-05, 'epoch': 3.83}) +[2026-06-20 04:00:17,925][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_2200_state +[2026-06-20 04:00:17,946][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-20 04:00:24,190][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_2200_state/model.safetensors +[2026-06-20 04:00:38,256][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_2200_state/optimizer.bin +[2026-06-20 04:00:38,257][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_2200_state/sampler.bin +[2026-06-20 04:00:38,257][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_2200_state/sampler_1.bin +[2026-06-20 04:00:38,258][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_2200_state/random_states_0.pkl +[2026-06-20 04:00:38,384][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2200 +[2026-06-20 04:00:38,385][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-20 04:00:38,398][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-20 04:00:49,218][starVLA.training.train_starvla][INFO] - Step 2201, Loss: {'train/loss': 0.14082440733909607, 'train/grad_norm_pre_clip': 16.267301559448242, 'timing/data': 0.006581161171197891, 'timing/model': 2.783090738579631, 'learning_rate/qwen_vl_interface': 6.304005152878902e-06, 'learning_rate/action_model': 6.304005152878901e-05, 'epoch': 3.83}) +[2026-06-20 04:01:00,132][starVLA.training.train_starvla][INFO] - Step 2202, Loss: {'train/loss': 0.11101607978343964, 'train/grad_norm_pre_clip': 9.010367393493652, 'timing/data': 0.00658310204744339, 'timing/model': 2.7165530435740948, 'learning_rate/qwen_vl_interface': 6.301035436203602e-06, 'learning_rate/action_model': 6.301035436203602e-05, 'epoch': 3.84}) +[2026-06-20 04:01:11,026][starVLA.training.train_starvla][INFO] - Step 2203, Loss: {'train/loss': 0.2457277625799179, 'train/grad_norm_pre_clip': 10.117630958557129, 'timing/data': 0.006474463269114494, 'timing/model': 2.7800758313387632, 'learning_rate/qwen_vl_interface': 6.298065287487337e-06, 'learning_rate/action_model': 6.298065287487337e-05, 'epoch': 3.84}) +[2026-06-20 04:01:21,925][starVLA.training.train_starvla][INFO] - Step 2204, Loss: {'train/loss': 0.1217990592122078, 'train/grad_norm_pre_clip': 20.256572723388672, 'timing/data': 0.006417384371161461, 'timing/model': 2.777856158092618, 'learning_rate/qwen_vl_interface': 6.295094707951019e-06, 'learning_rate/action_model': 6.295094707951019e-05, 'epoch': 3.84}) +[2026-06-20 04:01:32,821][starVLA.training.train_starvla][INFO] - Step 2205, Loss: {'train/loss': 0.3524519205093384, 'train/grad_norm_pre_clip': 15.554078102111816, 'timing/data': 0.006621280685067177, 'timing/model': 2.779722949489951, 'learning_rate/qwen_vl_interface': 6.292123698815745e-06, 'learning_rate/action_model': 6.292123698815745e-05, 'epoch': 3.84}) +[2026-06-20 04:01:43,715][starVLA.training.train_starvla][INFO] - Step 2206, Loss: {'train/loss': 0.13587415218353271, 'train/grad_norm_pre_clip': 10.41031265258789, 'timing/data': 0.00649353489279747, 'timing/model': 2.7707158159464598, 'learning_rate/qwen_vl_interface': 6.289152261302783e-06, 'learning_rate/action_model': 6.289152261302784e-05, 'epoch': 3.84}) +[2026-06-20 04:01:54,519][starVLA.training.train_starvla][INFO] - Step 2207, Loss: {'train/loss': 0.23850499093532562, 'train/grad_norm_pre_clip': 13.940826416015625, 'timing/data': 0.006770968437194824, 'timing/model': 2.770568870007992, 'learning_rate/qwen_vl_interface': 6.286180396633579e-06, 'learning_rate/action_model': 6.286180396633579e-05, 'epoch': 3.84}) +[2026-06-20 04:02:05,427][starVLA.training.train_starvla][INFO] - Step 2208, Loss: {'train/loss': 0.18768525123596191, 'train/grad_norm_pre_clip': 16.122623443603516, 'timing/data': 0.006816167384386063, 'timing/model': 2.7725152280181646, 'learning_rate/qwen_vl_interface': 6.283208106029753e-06, 'learning_rate/action_model': 6.283208106029752e-05, 'epoch': 3.85}) +[2026-06-20 04:02:16,327][starVLA.training.train_starvla][INFO] - Step 2209, Loss: {'train/loss': 0.13192430138587952, 'train/grad_norm_pre_clip': 25.626243591308594, 'timing/data': 0.006726572290062904, 'timing/model': 2.7740590255707502, 'learning_rate/qwen_vl_interface': 6.280235390713102e-06, 'learning_rate/action_model': 6.280235390713102e-05, 'epoch': 3.85}) +[2026-06-20 04:02:27,427][starVLA.training.train_starvla][INFO] - Step 2210, Loss: {'train/loss': 0.15987488627433777, 'train/grad_norm_pre_clip': 11.899003028869629, 'timing/data': 0.006605302914977074, 'timing/model': 2.7712082155048847, 'learning_rate/qwen_vl_interface': 6.277262251905597e-06, 'learning_rate/action_model': 6.277262251905597e-05, 'epoch': 3.85}) +[2026-06-20 04:02:38,335][starVLA.training.train_starvla][INFO] - Step 2211, Loss: {'train/loss': 0.06560767441987991, 'train/grad_norm_pre_clip': 3.7520973682403564, 'timing/data': 0.006614742800593376, 'timing/model': 2.783809771761298, 'learning_rate/qwen_vl_interface': 6.274288690829385e-06, 'learning_rate/action_model': 6.274288690829384e-05, 'epoch': 3.85}) +[2026-06-20 04:02:49,227][starVLA.training.train_starvla][INFO] - Step 2212, Loss: {'train/loss': 0.12923669815063477, 'train/grad_norm_pre_clip': 10.254097938537598, 'timing/data': 0.006913136690855026, 'timing/model': 2.781802363693714, 'learning_rate/qwen_vl_interface': 6.271314708706781e-06, 'learning_rate/action_model': 6.271314708706782e-05, 'epoch': 3.85}) +[2026-06-20 04:03:00,126][starVLA.training.train_starvla][INFO] - Step 2213, Loss: {'train/loss': 0.13836929202079773, 'train/grad_norm_pre_clip': 12.5086030960083, 'timing/data': 0.0065629128366708755, 'timing/model': 2.7817705050110817, 'learning_rate/qwen_vl_interface': 6.26834030676028e-06, 'learning_rate/action_model': 6.26834030676028e-05, 'epoch': 3.86}) +[2026-06-20 04:03:11,017][starVLA.training.train_starvla][INFO] - Step 2214, Loss: {'train/loss': 0.08838292211294174, 'train/grad_norm_pre_clip': 6.313057899475098, 'timing/data': 0.006767751649022102, 'timing/model': 2.776121212169528, 'learning_rate/qwen_vl_interface': 6.265365486212545e-06, 'learning_rate/action_model': 6.265365486212545e-05, 'epoch': 3.86}) +[2026-06-20 04:03:21,929][starVLA.training.train_starvla][INFO] - Step 2215, Loss: {'train/loss': 0.15716448426246643, 'train/grad_norm_pre_clip': 8.105684280395508, 'timing/data': 0.00648491270840168, 'timing/model': 2.7787271179258823, 'learning_rate/qwen_vl_interface': 6.26239024828641e-06, 'learning_rate/action_model': 6.26239024828641e-05, 'epoch': 3.86}) +[2026-06-20 04:03:32,831][starVLA.training.train_starvla][INFO] - Step 2216, Loss: {'train/loss': 0.23540230095386505, 'train/grad_norm_pre_clip': 14.150040626525879, 'timing/data': 0.006632721051573753, 'timing/model': 2.7909392043948174, 'learning_rate/qwen_vl_interface': 6.259414594204888e-06, 'learning_rate/action_model': 6.259414594204888e-05, 'epoch': 3.86}) +[2026-06-20 04:03:43,735][starVLA.training.train_starvla][INFO] - Step 2217, Loss: {'train/loss': 0.2696365416049957, 'train/grad_norm_pre_clip': 16.0770263671875, 'timing/data': 0.006498953327536583, 'timing/model': 2.7804601918905973, 'learning_rate/qwen_vl_interface': 6.256438525191155e-06, 'learning_rate/action_model': 6.256438525191156e-05, 'epoch': 3.86}) +[2026-06-20 04:03:54,632][starVLA.training.train_starvla][INFO] - Step 2218, Loss: {'train/loss': 0.17154867947101593, 'train/grad_norm_pre_clip': 14.509273529052734, 'timing/data': 0.006968997418880463, 'timing/model': 2.7771412692964077, 'learning_rate/qwen_vl_interface': 6.253462042468561e-06, 'learning_rate/action_model': 6.25346204246856e-05, 'epoch': 3.86}) +[2026-06-20 04:04:05,527][starVLA.training.train_starvla][INFO] - Step 2219, Loss: {'train/loss': 0.07843359559774399, 'train/grad_norm_pre_clip': 14.471808433532715, 'timing/data': 0.006531471386551857, 'timing/model': 2.7675211410969496, 'learning_rate/qwen_vl_interface': 6.2504851472606246e-06, 'learning_rate/action_model': 6.250485147260625e-05, 'epoch': 3.87}) +[2026-06-20 04:04:16,432][starVLA.training.train_starvla][INFO] - Step 2220, Loss: {'train/loss': 0.0768728032708168, 'train/grad_norm_pre_clip': 4.809507369995117, 'timing/data': 0.00626792386174202, 'timing/model': 2.7767447866499424, 'learning_rate/qwen_vl_interface': 6.247507840791035e-06, 'learning_rate/action_model': 6.247507840791035e-05, 'epoch': 3.87}) +[2026-06-20 04:04:27,348][starVLA.training.train_starvla][INFO] - Step 2221, Loss: {'train/loss': 0.11854200810194016, 'train/grad_norm_pre_clip': 8.415366172790527, 'timing/data': 0.006518524140119553, 'timing/model': 2.7239037696272135, 'learning_rate/qwen_vl_interface': 6.244530124283651e-06, 'learning_rate/action_model': 6.24453012428365e-05, 'epoch': 3.87}) +[2026-06-20 04:04:38,204][starVLA.training.train_starvla][INFO] - Step 2222, Loss: {'train/loss': 0.21164479851722717, 'train/grad_norm_pre_clip': 9.445307731628418, 'timing/data': 0.006700800731778145, 'timing/model': 2.7637701611965895, 'learning_rate/qwen_vl_interface': 6.2415519989625016e-06, 'learning_rate/action_model': 6.241551998962501e-05, 'epoch': 3.87}) +[2026-06-20 04:04:48,948][starVLA.training.train_starvla][INFO] - Step 2223, Loss: {'train/loss': 0.1643553525209427, 'train/grad_norm_pre_clip': 8.596611022949219, 'timing/data': 0.00645078532397747, 'timing/model': 2.702542271465063, 'learning_rate/qwen_vl_interface': 6.238573466051781e-06, 'learning_rate/action_model': 6.23857346605178e-05, 'epoch': 3.87}) +[2026-06-20 04:04:59,820][starVLA.training.train_starvla][INFO] - Step 2224, Loss: {'train/loss': 0.18903739750385284, 'train/grad_norm_pre_clip': 9.943099021911621, 'timing/data': 0.006798580288887024, 'timing/model': 2.77269191108644, 'learning_rate/qwen_vl_interface': 6.2355945267758465e-06, 'learning_rate/action_model': 6.235594526775846e-05, 'epoch': 3.87}) +[2026-06-20 04:05:10,721][starVLA.training.train_starvla][INFO] - Step 2225, Loss: {'train/loss': 0.29267770051956177, 'train/grad_norm_pre_clip': 13.115431785583496, 'timing/data': 0.006699882447719574, 'timing/model': 2.7613488603383303, 'learning_rate/qwen_vl_interface': 6.232615182359231e-06, 'learning_rate/action_model': 6.232615182359231e-05, 'epoch': 3.88}) +[2026-06-20 04:05:21,630][starVLA.training.train_starvla][INFO] - Step 2226, Loss: {'train/loss': 0.12362205237150192, 'train/grad_norm_pre_clip': 12.010883331298828, 'timing/data': 0.0064775943756103516, 'timing/model': 2.772120689973235, 'learning_rate/qwen_vl_interface': 6.229635434026634e-06, 'learning_rate/action_model': 6.229635434026634e-05, 'epoch': 3.88}) +[2026-06-20 04:05:32,531][starVLA.training.train_starvla][INFO] - Step 2227, Loss: {'train/loss': 0.043296344578266144, 'train/grad_norm_pre_clip': 6.102853298187256, 'timing/data': 0.006782993674278259, 'timing/model': 2.7842434979975224, 'learning_rate/qwen_vl_interface': 6.226655283002912e-06, 'learning_rate/action_model': 6.226655283002911e-05, 'epoch': 3.88}) +[2026-06-20 04:05:43,433][starVLA.training.train_starvla][INFO] - Step 2228, Loss: {'train/loss': 0.029152169823646545, 'train/grad_norm_pre_clip': 2.6924564838409424, 'timing/data': 0.006666840985417366, 'timing/model': 2.7794034890830517, 'learning_rate/qwen_vl_interface': 6.223674730513096e-06, 'learning_rate/action_model': 6.223674730513096e-05, 'epoch': 3.88}) +[2026-06-20 04:05:54,537][starVLA.training.train_starvla][INFO] - Step 2229, Loss: {'train/loss': 0.07972402125597, 'train/grad_norm_pre_clip': 10.884870529174805, 'timing/data': 0.007164526730775833, 'timing/model': 2.778120331466198, 'learning_rate/qwen_vl_interface': 6.220693777782375e-06, 'learning_rate/action_model': 6.220693777782375e-05, 'epoch': 3.88}) +[2026-06-20 04:06:05,435][starVLA.training.train_starvla][INFO] - Step 2230, Loss: {'train/loss': 0.06933610141277313, 'train/grad_norm_pre_clip': 8.619635581970215, 'timing/data': 0.006650539115071297, 'timing/model': 2.7805059906095266, 'learning_rate/qwen_vl_interface': 6.21771242603611e-06, 'learning_rate/action_model': 6.21771242603611e-05, 'epoch': 3.88}) +[2026-06-20 04:06:16,326][starVLA.training.train_starvla][INFO] - Step 2231, Loss: {'train/loss': 0.21514081954956055, 'train/grad_norm_pre_clip': 10.45163631439209, 'timing/data': 0.006709462031722069, 'timing/model': 2.7673431430011988, 'learning_rate/qwen_vl_interface': 6.21473067649982e-06, 'learning_rate/action_model': 6.21473067649982e-05, 'epoch': 3.89}) +[2026-06-20 04:06:27,237][starVLA.training.train_starvla][INFO] - Step 2232, Loss: {'train/loss': 0.13173027336597443, 'train/grad_norm_pre_clip': 15.939555168151855, 'timing/data': 0.006741922348737717, 'timing/model': 2.7841232419013977, 'learning_rate/qwen_vl_interface': 6.211748530399189e-06, 'learning_rate/action_model': 6.211748530399189e-05, 'epoch': 3.89}) +[2026-06-20 04:06:38,129][starVLA.training.train_starvla][INFO] - Step 2233, Loss: {'train/loss': 0.12153694778680801, 'train/grad_norm_pre_clip': 7.712298393249512, 'timing/data': 0.006677292287349701, 'timing/model': 2.773766342550516, 'learning_rate/qwen_vl_interface': 6.208765988960067e-06, 'learning_rate/action_model': 6.208765988960066e-05, 'epoch': 3.89}) +[2026-06-20 04:06:49,034][starVLA.training.train_starvla][INFO] - Step 2234, Loss: {'train/loss': 0.2666884958744049, 'train/grad_norm_pre_clip': 20.949867248535156, 'timing/data': 0.006812049075961113, 'timing/model': 2.789355091750622, 'learning_rate/qwen_vl_interface': 6.2057830534084615e-06, 'learning_rate/action_model': 6.205783053408461e-05, 'epoch': 3.89}) +[2026-06-20 04:06:59,937][starVLA.training.train_starvla][INFO] - Step 2235, Loss: {'train/loss': 0.06252428144216537, 'train/grad_norm_pre_clip': 7.662460803985596, 'timing/data': 0.006812639534473419, 'timing/model': 2.785560382530093, 'learning_rate/qwen_vl_interface': 6.202799724970546e-06, 'learning_rate/action_model': 6.202799724970546e-05, 'epoch': 3.89}) +[2026-06-20 04:07:10,836][starVLA.training.train_starvla][INFO] - Step 2236, Loss: {'train/loss': 0.16437940299510956, 'train/grad_norm_pre_clip': 20.153099060058594, 'timing/data': 0.008704360574483871, 'timing/model': 2.766025273129344, 'learning_rate/qwen_vl_interface': 6.199816004872655e-06, 'learning_rate/action_model': 6.199816004872655e-05, 'epoch': 3.9}) +[2026-06-20 04:07:21,838][starVLA.training.train_starvla][INFO] - Step 2237, Loss: {'train/loss': 0.15517547726631165, 'train/grad_norm_pre_clip': 16.88385772705078, 'timing/data': 0.015001213178038597, 'timing/model': 2.772429406642914, 'learning_rate/qwen_vl_interface': 6.196831894341281e-06, 'learning_rate/action_model': 6.19683189434128e-05, 'epoch': 3.9}) +[2026-06-20 04:07:32,723][starVLA.training.train_starvla][INFO] - Step 2238, Loss: {'train/loss': 0.2046191245317459, 'train/grad_norm_pre_clip': 13.71955680847168, 'timing/data': 0.00670253299176693, 'timing/model': 2.7754326686263084, 'learning_rate/qwen_vl_interface': 6.1938473946030795e-06, 'learning_rate/action_model': 6.193847394603079e-05, 'epoch': 3.9}) +[2026-06-20 04:07:43,633][starVLA.training.train_starvla][INFO] - Step 2239, Loss: {'train/loss': 0.1254889816045761, 'train/grad_norm_pre_clip': 12.53183650970459, 'timing/data': 0.0064123161137104034, 'timing/model': 2.7727493178099394, 'learning_rate/qwen_vl_interface': 6.1908625068848695e-06, 'learning_rate/action_model': 6.19086250688487e-05, 'epoch': 3.9}) +[2026-06-20 04:07:54,525][starVLA.training.train_starvla][INFO] - Step 2240, Loss: {'train/loss': 0.2094724327325821, 'train/grad_norm_pre_clip': 16.96273422241211, 'timing/data': 0.0065347012132406235, 'timing/model': 2.7731229551136494, 'learning_rate/qwen_vl_interface': 6.187877232413619e-06, 'learning_rate/action_model': 6.18787723241362e-05, 'epoch': 3.9}) +[2026-06-20 04:08:05,190][starVLA.training.train_starvla][INFO] - Step 2241, Loss: {'train/loss': 0.11807580292224884, 'train/grad_norm_pre_clip': 13.647419929504395, 'timing/data': 0.00652630440890789, 'timing/model': 2.7101518400013447, 'learning_rate/qwen_vl_interface': 6.184891572416467e-06, 'learning_rate/action_model': 6.184891572416466e-05, 'epoch': 3.9}) +[2026-06-20 04:08:15,754][starVLA.training.train_starvla][INFO] - Step 2242, Loss: {'train/loss': 0.09212255477905273, 'train/grad_norm_pre_clip': 9.840909957885742, 'timing/data': 0.0015685223042964935, 'timing/model': 2.6744820680469275, 'learning_rate/qwen_vl_interface': 6.181905528120703e-06, 'learning_rate/action_model': 6.181905528120703e-05, 'epoch': 3.91}) +[2026-06-20 04:08:46,547][starVLA.training.train_starvla][INFO] - Step 2243, Loss: {'train/loss': 0.11693014204502106, 'train/grad_norm_pre_clip': 8.378212928771973, 'timing/data': 0.0005909856408834457, 'timing/model': 2.710228683426976, 'learning_rate/qwen_vl_interface': 6.178919100753778e-06, 'learning_rate/action_model': 6.178919100753778e-05, 'epoch': 3.91}) +[2026-06-20 04:08:57,425][starVLA.training.train_starvla][INFO] - Step 2244, Loss: {'train/loss': 0.2407134473323822, 'train/grad_norm_pre_clip': 31.47600555419922, 'timing/data': 0.000577414408326149, 'timing/model': 2.791861327365041, 'learning_rate/qwen_vl_interface': 6.175932291543299e-06, 'learning_rate/action_model': 6.175932291543298e-05, 'epoch': 3.91}) +[2026-06-20 04:09:08,330][starVLA.training.train_starvla][INFO] - Step 2245, Loss: {'train/loss': 0.15411704778671265, 'train/grad_norm_pre_clip': 14.209538459777832, 'timing/data': 0.000586213544011116, 'timing/model': 2.7860539574176073, 'learning_rate/qwen_vl_interface': 6.172945101717029e-06, 'learning_rate/action_model': 6.17294510171703e-05, 'epoch': 3.91}) +[2026-06-20 04:09:19,053][starVLA.training.train_starvla][INFO] - Step 2246, Loss: {'train/loss': 0.1218484416604042, 'train/grad_norm_pre_clip': 17.070354461669922, 'timing/data': 0.0007015727460384369, 'timing/model': 2.6440648678690195, 'learning_rate/qwen_vl_interface': 6.169957532502892e-06, 'learning_rate/action_model': 6.169957532502892e-05, 'epoch': 3.91}) +[2026-06-20 04:09:29,826][starVLA.training.train_starvla][INFO] - Step 2247, Loss: {'train/loss': 0.17832054197788239, 'train/grad_norm_pre_clip': 17.47808074951172, 'timing/data': 0.0005764123052358627, 'timing/model': 2.7929553482681513, 'learning_rate/qwen_vl_interface': 6.166969585128962e-06, 'learning_rate/action_model': 6.166969585128962e-05, 'epoch': 3.91}) +[2026-06-20 04:09:40,927][starVLA.training.train_starvla][INFO] - Step 2248, Loss: {'train/loss': 0.24077141284942627, 'train/grad_norm_pre_clip': 24.292049407958984, 'timing/data': 0.008417744189500809, 'timing/model': 2.7649235669523478, 'learning_rate/qwen_vl_interface': 6.163981260823473e-06, 'learning_rate/action_model': 6.163981260823473e-05, 'epoch': 3.92}) +[2026-06-20 04:09:51,832][starVLA.training.train_starvla][INFO] - Step 2249, Loss: {'train/loss': 0.08301874250173569, 'train/grad_norm_pre_clip': 6.720280647277832, 'timing/data': 0.006592772901058197, 'timing/model': 2.7798666320741177, 'learning_rate/qwen_vl_interface': 6.16099256081481e-06, 'learning_rate/action_model': 6.16099256081481e-05, 'epoch': 3.92}) +[2026-06-20 04:10:02,728][starVLA.training.train_starvla][INFO] - Step 2250, Loss: {'train/loss': 0.11008954048156738, 'train/grad_norm_pre_clip': 14.734241485595703, 'timing/data': 0.006918266415596008, 'timing/model': 2.7808152362704277, 'learning_rate/qwen_vl_interface': 6.15800348633152e-06, 'learning_rate/action_model': 6.15800348633152e-05, 'epoch': 3.92}) +[2026-06-20 04:10:13,627][starVLA.training.train_starvla][INFO] - Step 2251, Loss: {'train/loss': 0.13412970304489136, 'train/grad_norm_pre_clip': 14.606502532958984, 'timing/data': 0.006597163155674934, 'timing/model': 2.775301354005933, 'learning_rate/qwen_vl_interface': 6.155014038602293e-06, 'learning_rate/action_model': 6.155014038602292e-05, 'epoch': 3.92}) +[2026-06-20 04:10:24,730][starVLA.training.train_starvla][INFO] - Step 2252, Loss: {'train/loss': 0.14913469552993774, 'train/grad_norm_pre_clip': 26.967937469482422, 'timing/data': 0.006563860923051834, 'timing/model': 2.784315975382924, 'learning_rate/qwen_vl_interface': 6.152024218855979e-06, 'learning_rate/action_model': 6.15202421885598e-05, 'epoch': 3.92}) +[2026-06-20 04:10:35,621][starVLA.training.train_starvla][INFO] - Step 2253, Loss: {'train/loss': 0.16612625122070312, 'train/grad_norm_pre_clip': 10.66148567199707, 'timing/data': 0.009410148486495018, 'timing/model': 2.7760292403399944, 'learning_rate/qwen_vl_interface': 6.149034028321583e-06, 'learning_rate/action_model': 6.149034028321583e-05, 'epoch': 3.92}) +[2026-06-20 04:10:46,532][starVLA.training.train_starvla][INFO] - Step 2254, Loss: {'train/loss': 0.05775515362620354, 'train/grad_norm_pre_clip': 9.844670295715332, 'timing/data': 0.006645850837230682, 'timing/model': 2.7770653646439314, 'learning_rate/qwen_vl_interface': 6.1460434682282566e-06, 'learning_rate/action_model': 6.146043468228256e-05, 'epoch': 3.93}) +[2026-06-20 04:10:57,649][starVLA.training.train_starvla][INFO] - Step 2255, Loss: {'train/loss': 0.28690215945243835, 'train/grad_norm_pre_clip': 22.134559631347656, 'timing/data': 0.00667923130095005, 'timing/model': 2.9958452843129635, 'learning_rate/qwen_vl_interface': 6.143052539805308e-06, 'learning_rate/action_model': 6.143052539805308e-05, 'epoch': 3.93}) +[2026-06-20 04:11:08,517][starVLA.training.train_starvla][INFO] - Step 2256, Loss: {'train/loss': 0.1508358120918274, 'train/grad_norm_pre_clip': 16.348472595214844, 'timing/data': 0.00670100562274456, 'timing/model': 2.761839099228382, 'learning_rate/qwen_vl_interface': 6.140061244282194e-06, 'learning_rate/action_model': 6.140061244282194e-05, 'epoch': 3.93}) +[2026-06-20 04:11:19,329][starVLA.training.train_starvla][INFO] - Step 2257, Loss: {'train/loss': 0.13770483434200287, 'train/grad_norm_pre_clip': 7.7140398025512695, 'timing/data': 0.006778912618756294, 'timing/model': 2.7734742369502783, 'learning_rate/qwen_vl_interface': 6.137069582888521e-06, 'learning_rate/action_model': 6.137069582888521e-05, 'epoch': 3.93}) +[2026-06-20 04:11:30,153][starVLA.training.train_starvla][INFO] - Step 2258, Loss: {'train/loss': 0.16873936355113983, 'train/grad_norm_pre_clip': 13.063393592834473, 'timing/data': 0.006475944072008133, 'timing/model': 2.7004698459059, 'learning_rate/qwen_vl_interface': 6.13407755685405e-06, 'learning_rate/action_model': 6.13407755685405e-05, 'epoch': 3.93}) +[2026-06-20 04:11:41,023][starVLA.training.train_starvla][INFO] - Step 2259, Loss: {'train/loss': 0.26742762327194214, 'train/grad_norm_pre_clip': 20.260557174682617, 'timing/data': 0.007620671764016151, 'timing/model': 2.7701273430138826, 'learning_rate/qwen_vl_interface': 6.131085167408692e-06, 'learning_rate/action_model': 6.131085167408692e-05, 'epoch': 3.94}) +[2026-06-20 04:11:51,938][starVLA.training.train_starvla][INFO] - Step 2260, Loss: {'train/loss': 0.06284605711698532, 'train/grad_norm_pre_clip': 8.011581420898438, 'timing/data': 0.007542332634329796, 'timing/model': 2.7824399080127478, 'learning_rate/qwen_vl_interface': 6.128092415782502e-06, 'learning_rate/action_model': 6.128092415782501e-05, 'epoch': 3.94}) +[2026-06-20 04:12:02,833][starVLA.training.train_starvla][INFO] - Step 2261, Loss: {'train/loss': 0.33964353799819946, 'train/grad_norm_pre_clip': 11.390408515930176, 'timing/data': 0.006632694974541664, 'timing/model': 2.777098845690489, 'learning_rate/qwen_vl_interface': 6.125099303205688e-06, 'learning_rate/action_model': 6.125099303205688e-05, 'epoch': 3.94}) +[2026-06-20 04:12:13,733][starVLA.training.train_starvla][INFO] - Step 2262, Loss: {'train/loss': 0.24872998893260956, 'train/grad_norm_pre_clip': 21.533174514770508, 'timing/data': 0.006814951077103615, 'timing/model': 2.775906950235367, 'learning_rate/qwen_vl_interface': 6.122105830908607e-06, 'learning_rate/action_model': 6.122105830908606e-05, 'epoch': 3.94}) +[2026-06-20 04:12:24,621][starVLA.training.train_starvla][INFO] - Step 2263, Loss: {'train/loss': 0.061155397444963455, 'train/grad_norm_pre_clip': 10.185956954956055, 'timing/data': 0.006711512804031372, 'timing/model': 2.767005255445838, 'learning_rate/qwen_vl_interface': 6.11911200012176e-06, 'learning_rate/action_model': 6.11911200012176e-05, 'epoch': 3.94}) +[2026-06-20 04:12:35,534][starVLA.training.train_starvla][INFO] - Step 2264, Loss: {'train/loss': 0.10043564438819885, 'train/grad_norm_pre_clip': 13.225363731384277, 'timing/data': 0.006992992013692856, 'timing/model': 2.7723628487437963, 'learning_rate/qwen_vl_interface': 6.116117812075799e-06, 'learning_rate/action_model': 6.116117812075799e-05, 'epoch': 3.94}) +[2026-06-20 04:12:46,438][starVLA.training.train_starvla][INFO] - Step 2265, Loss: {'train/loss': 0.0851326659321785, 'train/grad_norm_pre_clip': 9.317597389221191, 'timing/data': 0.006678752601146698, 'timing/model': 2.785448309034109, 'learning_rate/qwen_vl_interface': 6.113123268001524e-06, 'learning_rate/action_model': 6.113123268001524e-05, 'epoch': 3.95}) +[2026-06-20 04:12:57,333][starVLA.training.train_starvla][INFO] - Step 2266, Loss: {'train/loss': 0.0343017615377903, 'train/grad_norm_pre_clip': 3.947922706604004, 'timing/data': 0.007047818973660469, 'timing/model': 2.7750906236469746, 'learning_rate/qwen_vl_interface': 6.110128369129873e-06, 'learning_rate/action_model': 6.110128369129873e-05, 'epoch': 3.95}) +[2026-06-20 04:13:08,336][starVLA.training.train_starvla][INFO] - Step 2267, Loss: {'train/loss': 0.19741636514663696, 'train/grad_norm_pre_clip': 7.307820796966553, 'timing/data': 0.006981790065765381, 'timing/model': 2.7768796235322952, 'learning_rate/qwen_vl_interface': 6.107133116691943e-06, 'learning_rate/action_model': 6.107133116691942e-05, 'epoch': 3.95}) +[2026-06-20 04:13:19,232][starVLA.training.train_starvla][INFO] - Step 2268, Loss: {'train/loss': 0.027691392228007317, 'train/grad_norm_pre_clip': 2.956163167953491, 'timing/data': 0.006727844476699829, 'timing/model': 2.7871886417269707, 'learning_rate/qwen_vl_interface': 6.104137511918963e-06, 'learning_rate/action_model': 6.104137511918963e-05, 'epoch': 3.95}) +[2026-06-20 04:13:30,118][starVLA.training.train_starvla][INFO] - Step 2269, Loss: {'train/loss': 0.3081822693347931, 'train/grad_norm_pre_clip': 14.500913619995117, 'timing/data': 0.0067880116403102875, 'timing/model': 2.7632129713892937, 'learning_rate/qwen_vl_interface': 6.1011415560423155e-06, 'learning_rate/action_model': 6.101141556042316e-05, 'epoch': 3.95}) +[2026-06-20 04:13:40,927][starVLA.training.train_starvla][INFO] - Step 2270, Loss: {'train/loss': 0.1748560667037964, 'train/grad_norm_pre_clip': 12.762742042541504, 'timing/data': 0.006984978914260864, 'timing/model': 2.7813737392425537, 'learning_rate/qwen_vl_interface': 6.098145250293525e-06, 'learning_rate/action_model': 6.0981452502935246e-05, 'epoch': 3.95}) +[2026-06-20 04:13:51,833][starVLA.training.train_starvla][INFO] - Step 2271, Loss: {'train/loss': 0.22478143870830536, 'train/grad_norm_pre_clip': 13.424336433410645, 'timing/data': 0.00659356452524662, 'timing/model': 2.7746741883456707, 'learning_rate/qwen_vl_interface': 6.09514859590426e-06, 'learning_rate/action_model': 6.09514859590426e-05, 'epoch': 3.96}) +[2026-06-20 04:14:02,732][starVLA.training.train_starvla][INFO] - Step 2272, Loss: {'train/loss': 0.13238555192947388, 'train/grad_norm_pre_clip': 11.501971244812012, 'timing/data': 0.006580382585525513, 'timing/model': 2.7778562735766172, 'learning_rate/qwen_vl_interface': 6.09215159410633e-06, 'learning_rate/action_model': 6.09215159410633e-05, 'epoch': 3.96}) +[2026-06-20 04:14:13,631][starVLA.training.train_starvla][INFO] - Step 2273, Loss: {'train/loss': 0.07040218263864517, 'train/grad_norm_pre_clip': 7.035827159881592, 'timing/data': 0.006588175892829895, 'timing/model': 2.7731853164732456, 'learning_rate/qwen_vl_interface': 6.089154246131692e-06, 'learning_rate/action_model': 6.089154246131692e-05, 'epoch': 3.96}) +[2026-06-20 04:14:24,537][starVLA.training.train_starvla][INFO] - Step 2274, Loss: {'train/loss': 0.11018478870391846, 'train/grad_norm_pre_clip': 8.959946632385254, 'timing/data': 0.006955400109291077, 'timing/model': 2.775963855907321, 'learning_rate/qwen_vl_interface': 6.08615655321244e-06, 'learning_rate/action_model': 6.08615655321244e-05, 'epoch': 3.96}) +[2026-06-20 04:14:35,635][starVLA.training.train_starvla][INFO] - Step 2275, Loss: {'train/loss': 0.14395567774772644, 'train/grad_norm_pre_clip': 9.468009948730469, 'timing/data': 0.006822213530540466, 'timing/model': 2.7790796905755997, 'learning_rate/qwen_vl_interface': 6.083158516580814e-06, 'learning_rate/action_model': 6.083158516580814e-05, 'epoch': 3.96}) +[2026-06-20 04:14:46,542][starVLA.training.train_starvla][INFO] - Step 2276, Loss: {'train/loss': 0.11453663557767868, 'train/grad_norm_pre_clip': 9.344584465026855, 'timing/data': 0.0070309992879629135, 'timing/model': 2.789556499570608, 'learning_rate/qwen_vl_interface': 6.080160137469192e-06, 'learning_rate/action_model': 6.0801601374691916e-05, 'epoch': 3.96}) +[2026-06-20 04:14:57,435][starVLA.training.train_starvla][INFO] - Step 2277, Loss: {'train/loss': 0.12944354116916656, 'train/grad_norm_pre_clip': 12.180830001831055, 'timing/data': 0.006886033341288567, 'timing/model': 2.7753981351852417, 'learning_rate/qwen_vl_interface': 6.077161417110097e-06, 'learning_rate/action_model': 6.0771614171100965e-05, 'epoch': 3.97}) +[2026-06-20 04:15:08,323][starVLA.training.train_starvla][INFO] - Step 2278, Loss: {'train/loss': 0.04682726413011551, 'train/grad_norm_pre_clip': 5.447670936584473, 'timing/data': 0.006657397374510765, 'timing/model': 2.7621857579797506, 'learning_rate/qwen_vl_interface': 6.0741623567361864e-06, 'learning_rate/action_model': 6.074162356736186e-05, 'epoch': 3.97}) +[2026-06-20 04:15:19,234][starVLA.training.train_starvla][INFO] - Step 2279, Loss: {'train/loss': 0.07428163290023804, 'train/grad_norm_pre_clip': 8.042256355285645, 'timing/data': 0.006945991888642311, 'timing/model': 2.7772473245859146, 'learning_rate/qwen_vl_interface': 6.071162957580262e-06, 'learning_rate/action_model': 6.071162957580262e-05, 'epoch': 3.97}) +[2026-06-20 04:15:30,137][starVLA.training.train_starvla][INFO] - Step 2280, Loss: {'train/loss': 0.1363855004310608, 'train/grad_norm_pre_clip': 10.771050453186035, 'timing/data': 0.006659423932433128, 'timing/model': 2.780841287225485, 'learning_rate/qwen_vl_interface': 6.068163220875266e-06, 'learning_rate/action_model': 6.068163220875266e-05, 'epoch': 3.97}) +[2026-06-20 04:15:41,023][starVLA.training.train_starvla][INFO] - Step 2281, Loss: {'train/loss': 0.22140148282051086, 'train/grad_norm_pre_clip': 22.340478897094727, 'timing/data': 0.006579097360372543, 'timing/model': 2.7742737252265215, 'learning_rate/qwen_vl_interface': 6.065163147854273e-06, 'learning_rate/action_model': 6.065163147854273e-05, 'epoch': 3.97}) +[2026-06-20 04:15:52,055][starVLA.training.train_starvla][INFO] - Step 2282, Loss: {'train/loss': 0.14790908992290497, 'train/grad_norm_pre_clip': 15.539799690246582, 'timing/data': 0.006659107282757759, 'timing/model': 2.89748540148139, 'learning_rate/qwen_vl_interface': 6.062162739750502e-06, 'learning_rate/action_model': 6.062162739750502e-05, 'epoch': 3.98}) +[2026-06-20 04:16:02,937][starVLA.training.train_starvla][INFO] - Step 2283, Loss: {'train/loss': 0.13811464607715607, 'train/grad_norm_pre_clip': 13.238125801086426, 'timing/data': 0.006752964109182358, 'timing/model': 2.7791588120162487, 'learning_rate/qwen_vl_interface': 6.059161997797306e-06, 'learning_rate/action_model': 6.0591619977973056e-05, 'epoch': 3.98}) +[2026-06-20 04:16:13,838][starVLA.training.train_starvla][INFO] - Step 2284, Loss: {'train/loss': 0.13212716579437256, 'train/grad_norm_pre_clip': 8.04012680053711, 'timing/data': 0.006640275940299034, 'timing/model': 2.782309517264366, 'learning_rate/qwen_vl_interface': 6.056160923228178e-06, 'learning_rate/action_model': 6.056160923228178e-05, 'epoch': 3.98}) +[2026-06-20 04:16:24,739][starVLA.training.train_starvla][INFO] - Step 2285, Loss: {'train/loss': 0.27456310391426086, 'train/grad_norm_pre_clip': 14.471025466918945, 'timing/data': 0.006945982575416565, 'timing/model': 2.7940402273088694, 'learning_rate/qwen_vl_interface': 6.0531595172767455e-06, 'learning_rate/action_model': 6.0531595172767455e-05, 'epoch': 3.98}) +[2026-06-20 04:16:35,638][starVLA.training.train_starvla][INFO] - Step 2286, Loss: {'train/loss': 0.086053766310215, 'train/grad_norm_pre_clip': 11.196139335632324, 'timing/data': 0.006679447367787361, 'timing/model': 2.7793893981724977, 'learning_rate/qwen_vl_interface': 6.050157781176774e-06, 'learning_rate/action_model': 6.050157781176774e-05, 'epoch': 3.98}) +[2026-06-20 04:16:46,536][starVLA.training.train_starvla][INFO] - Step 2287, Loss: {'train/loss': 0.04484737664461136, 'train/grad_norm_pre_clip': 7.9372477531433105, 'timing/data': 0.006808146834373474, 'timing/model': 2.7756960075348616, 'learning_rate/qwen_vl_interface': 6.047155716162164e-06, 'learning_rate/action_model': 6.047155716162164e-05, 'epoch': 3.98}) +[2026-06-20 04:16:57,436][starVLA.training.train_starvla][INFO] - Step 2288, Loss: {'train/loss': 0.09749103337526321, 'train/grad_norm_pre_clip': 7.148461818695068, 'timing/data': 0.006589785218238831, 'timing/model': 2.7765954937785864, 'learning_rate/qwen_vl_interface': 6.044153323466948e-06, 'learning_rate/action_model': 6.044153323466948e-05, 'epoch': 3.99}) +[2026-06-20 04:17:08,339][starVLA.training.train_starvla][INFO] - Step 2289, Loss: {'train/loss': 0.2868479788303375, 'train/grad_norm_pre_clip': 13.53475284576416, 'timing/data': 0.006534267216920853, 'timing/model': 2.781230978667736, 'learning_rate/qwen_vl_interface': 6.0411506043253e-06, 'learning_rate/action_model': 6.0411506043253e-05, 'epoch': 3.99}) +[2026-06-20 04:17:19,233][starVLA.training.train_starvla][INFO] - Step 2290, Loss: {'train/loss': 0.11027836054563522, 'train/grad_norm_pre_clip': 6.359996318817139, 'timing/data': 0.006836526095867157, 'timing/model': 2.7733471300452948, 'learning_rate/qwen_vl_interface': 6.038147559971523e-06, 'learning_rate/action_model': 6.038147559971523e-05, 'epoch': 3.99}) +[2026-06-20 04:17:30,137][starVLA.training.train_starvla][INFO] - Step 2291, Loss: {'train/loss': 0.12499592453241348, 'train/grad_norm_pre_clip': 20.401891708374023, 'timing/data': 0.007960520684719086, 'timing/model': 2.774943647906184, 'learning_rate/qwen_vl_interface': 6.035144191640055e-06, 'learning_rate/action_model': 6.035144191640055e-05, 'epoch': 3.99}) +[2026-06-20 04:17:41,039][starVLA.training.train_starvla][INFO] - Step 2292, Loss: {'train/loss': 0.14677411317825317, 'train/grad_norm_pre_clip': 10.385839462280273, 'timing/data': 0.006899746134877205, 'timing/model': 2.78085102327168, 'learning_rate/qwen_vl_interface': 6.032140500565467e-06, 'learning_rate/action_model': 6.0321405005654665e-05, 'epoch': 3.99}) +[2026-06-20 04:17:51,932][starVLA.training.train_starvla][INFO] - Step 2293, Loss: {'train/loss': 0.11555179208517075, 'train/grad_norm_pre_clip': 8.775211334228516, 'timing/data': 0.006635969504714012, 'timing/model': 2.7825517673045397, 'learning_rate/qwen_vl_interface': 6.029136487982463e-06, 'learning_rate/action_model': 6.029136487982463e-05, 'epoch': 3.99}) +[2026-06-20 04:18:03,034][starVLA.training.train_starvla][INFO] - Step 2294, Loss: {'train/loss': 0.04714256897568703, 'train/grad_norm_pre_clip': 9.123623847961426, 'timing/data': 0.006607959046959877, 'timing/model': 2.7646887097507715, 'learning_rate/qwen_vl_interface': 6.02613215512588e-06, 'learning_rate/action_model': 6.0261321551258796e-05, 'epoch': 4.0}) +[2026-06-20 04:18:13,936][starVLA.training.train_starvla][INFO] - Step 2295, Loss: {'train/loss': 0.09090609848499298, 'train/grad_norm_pre_clip': 12.240134239196777, 'timing/data': 0.006478147581219673, 'timing/model': 2.7772403843700886, 'learning_rate/qwen_vl_interface': 6.023127503230684e-06, 'learning_rate/action_model': 6.023127503230683e-05, 'epoch': 4.0}) +[2026-06-20 04:18:24,840][starVLA.training.train_starvla][INFO] - Step 2296, Loss: {'train/loss': 0.1840762197971344, 'train/grad_norm_pre_clip': 8.865229606628418, 'timing/data': 0.01516716368496418, 'timing/model': 2.7737898733466864, 'learning_rate/qwen_vl_interface': 6.020122533531976e-06, 'learning_rate/action_model': 6.0201225335319755e-05, 'epoch': 4.0}) +[2026-06-20 04:18:35,738][starVLA.training.train_starvla][INFO] - Step 2297, Loss: {'train/loss': 0.05649994686245918, 'train/grad_norm_pre_clip': 11.585912704467773, 'timing/data': 0.006549138575792313, 'timing/model': 2.778078665956855, 'learning_rate/qwen_vl_interface': 6.017117247264983e-06, 'learning_rate/action_model': 6.017117247264983e-05, 'epoch': 4.0}) +[2026-06-20 04:18:46,636][starVLA.training.train_starvla][INFO] - Step 2298, Loss: {'train/loss': 0.10194046795368195, 'train/grad_norm_pre_clip': 9.828392028808594, 'timing/data': 0.006634168326854706, 'timing/model': 2.7757155634462833, 'learning_rate/qwen_vl_interface': 6.014111645665069e-06, 'learning_rate/action_model': 6.0141116456650684e-05, 'epoch': 4.0}) +[2026-06-20 04:18:57,538][starVLA.training.train_starvla][INFO] - Step 2299, Loss: {'train/loss': 0.0409870445728302, 'train/grad_norm_pre_clip': 4.352397441864014, 'timing/data': 0.006295371800661087, 'timing/model': 2.777529591694474, 'learning_rate/qwen_vl_interface': 6.011105729967719e-06, 'learning_rate/action_model': 6.0111057299677184e-05, 'epoch': 4.01}) +[2026-06-20 04:19:08,430][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 2300 +[2026-06-20 04:26:38,963][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 2300 +[2026-06-20 04:26:38,964][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 2300 +[2026-06-20 04:34:52,017][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 2300 +[2026-06-20 04:34:52,018][starVLA.training.train_starvla][INFO] - Step 2300, Loss: {'train/loss': 0.17556963860988617, 'train/grad_norm_pre_clip': 11.891397476196289, 'eval/loss': 0.7980558276176453, 'eval/latency_0/loss': 0.7230067253112793, 'eval/latency_1/loss': 0.903584897518158, 'eval/latency_2/loss': 0.08916199207305908, 'eval/latency_3/loss': 0.9061099290847778, 'eval/latency_4/loss': 1.4108617305755615, 'eval/flappy/loss': 0.7980558276176453, 'eval/flappy/latency_0/loss': 0.7230067253112793, 'eval/flappy/latency_1/loss': 0.903584897518158, 'eval/flappy/latency_2/loss': 0.08916199207305908, 'eval/flappy/latency_3/loss': 0.9061099290847778, 'eval/flappy/latency_4/loss': 1.4108617305755615, 'eval/flap_precision': 0.8297872340425532, 'eval/flap_recall': 0.28363636363636363, 'eval/flap_f1': 0.4227642276422764, 'eval/flap_rate_model': 0.026111111111111113, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.4227642276422764, 'eval/latency_0/flap_precision': 0.8297872340425532, 'eval/latency_0/flap_recall': 0.28363636363636363, 'eval/latency_0/flap_f1': 0.4227642276422764, 'eval/latency_0/flap_rate_model': 0.026111111111111113, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.4227642276422764, 'timing/data': 0.006610468029975891, 'timing/model': 2.773167798295617, 'learning_rate/qwen_vl_interface': 6.008099501408552e-06, 'learning_rate/action_model': 6.008099501408552e-05, 'epoch': 4.01}) +[2026-06-20 04:34:52,019][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_2300_state +[2026-06-20 04:34:52,023][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-20 04:34:57,904][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_2300_state/model.safetensors +[2026-06-20 04:35:05,144][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_2300_state/optimizer.bin +[2026-06-20 04:35:05,145][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_2300_state/sampler.bin +[2026-06-20 04:35:05,145][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_2300_state/sampler_1.bin +[2026-06-20 04:35:05,146][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_2300_state/random_states_0.pkl +[2026-06-20 04:35:05,261][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2300 +[2026-06-20 04:35:05,261][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-20 04:35:05,275][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-20 04:35:16,333][starVLA.training.train_starvla][INFO] - Step 2301, Loss: {'train/loss': 0.05613628402352333, 'train/grad_norm_pre_clip': 9.862112998962402, 'timing/data': 0.005590241402387619, 'timing/model': 2.7852727323770523, 'learning_rate/qwen_vl_interface': 6.0050929612233165e-06, 'learning_rate/action_model': 6.005092961223316e-05, 'epoch': 4.01}) +[2026-06-20 04:35:27,235][starVLA.training.train_starvla][INFO] - Step 2302, Loss: {'train/loss': 0.11598845571279526, 'train/grad_norm_pre_clip': 9.291013717651367, 'timing/data': 0.007602643221616745, 'timing/model': 2.7812775913625956, 'learning_rate/qwen_vl_interface': 6.002086110647888e-06, 'learning_rate/action_model': 6.002086110647888e-05, 'epoch': 4.01}) +[2026-06-20 04:35:38,336][starVLA.training.train_starvla][INFO] - Step 2303, Loss: {'train/loss': 0.11275006085634232, 'train/grad_norm_pre_clip': 18.989669799804688, 'timing/data': 0.0066657401621341705, 'timing/model': 2.718859739601612, 'learning_rate/qwen_vl_interface': 5.999078950918267e-06, 'learning_rate/action_model': 5.999078950918267e-05, 'epoch': 4.01}) +[2026-06-20 04:35:49,231][starVLA.training.train_starvla][INFO] - Step 2304, Loss: {'train/loss': 0.033069659024477005, 'train/grad_norm_pre_clip': 9.270312309265137, 'timing/data': 0.0066224802285432816, 'timing/model': 2.7732586823403835, 'learning_rate/qwen_vl_interface': 5.996071483270585e-06, 'learning_rate/action_model': 5.996071483270585e-05, 'epoch': 4.01}) +[2026-06-20 04:36:00,134][starVLA.training.train_starvla][INFO] - Step 2305, Loss: {'train/loss': 0.14567142724990845, 'train/grad_norm_pre_clip': 14.893364906311035, 'timing/data': 0.0065265316516160965, 'timing/model': 2.781241722404957, 'learning_rate/qwen_vl_interface': 5.993063708941098e-06, 'learning_rate/action_model': 5.993063708941098e-05, 'epoch': 4.02}) +[2026-06-20 04:36:11,018][starVLA.training.train_starvla][INFO] - Step 2306, Loss: {'train/loss': 0.021804794669151306, 'train/grad_norm_pre_clip': 3.5819101333618164, 'timing/data': 0.006702879443764687, 'timing/model': 2.774491436779499, 'learning_rate/qwen_vl_interface': 5.990055629166187e-06, 'learning_rate/action_model': 5.990055629166187e-05, 'epoch': 4.02}) +[2026-06-20 04:36:21,924][starVLA.training.train_starvla][INFO] - Step 2307, Loss: {'train/loss': 0.09772183746099472, 'train/grad_norm_pre_clip': 12.511630058288574, 'timing/data': 0.006702320650219917, 'timing/model': 2.774689557030797, 'learning_rate/qwen_vl_interface': 5.987047245182363e-06, 'learning_rate/action_model': 5.987047245182362e-05, 'epoch': 4.02}) +[2026-06-20 04:36:32,816][starVLA.training.train_starvla][INFO] - Step 2308, Loss: {'train/loss': 0.1428801417350769, 'train/grad_norm_pre_clip': 24.21009635925293, 'timing/data': 0.006824048236012459, 'timing/model': 2.7690206840634346, 'learning_rate/qwen_vl_interface': 5.984038558226255e-06, 'learning_rate/action_model': 5.9840385582262546e-05, 'epoch': 4.02}) +[2026-06-20 04:36:43,626][starVLA.training.train_starvla][INFO] - Step 2309, Loss: {'train/loss': 0.01579046994447708, 'train/grad_norm_pre_clip': 2.9369301795959473, 'timing/data': 0.006663858890533447, 'timing/model': 2.77688717469573, 'learning_rate/qwen_vl_interface': 5.981029569534625e-06, 'learning_rate/action_model': 5.981029569534625e-05, 'epoch': 4.02}) +[2026-06-20 04:36:54,478][starVLA.training.train_starvla][INFO] - Step 2310, Loss: {'train/loss': 0.17011526226997375, 'train/grad_norm_pre_clip': 15.287089347839355, 'timing/data': 0.006987638771533966, 'timing/model': 2.7407999262213707, 'learning_rate/qwen_vl_interface': 5.978020280344351e-06, 'learning_rate/action_model': 5.9780202803443506e-05, 'epoch': 4.02}) +[2026-06-20 04:37:05,324][starVLA.training.train_starvla][INFO] - Step 2311, Loss: {'train/loss': 0.25078046321868896, 'train/grad_norm_pre_clip': 24.54135513305664, 'timing/data': 0.0064360518008470535, 'timing/model': 2.7791089061647654, 'learning_rate/qwen_vl_interface': 5.97501069189244e-06, 'learning_rate/action_model': 5.97501069189244e-05, 'epoch': 4.03}) +[2026-06-20 04:37:16,226][starVLA.training.train_starvla][INFO] - Step 2312, Loss: {'train/loss': 0.11695438623428345, 'train/grad_norm_pre_clip': 9.15456771850586, 'timing/data': 0.00668356753885746, 'timing/model': 2.7742754835635424, 'learning_rate/qwen_vl_interface': 5.97200080541602e-06, 'learning_rate/action_model': 5.97200080541602e-05, 'epoch': 4.03}) +[2026-06-20 04:37:27,130][starVLA.training.train_starvla][INFO] - Step 2313, Loss: {'train/loss': 0.041022978723049164, 'train/grad_norm_pre_clip': 17.910655975341797, 'timing/data': 0.006600100547075272, 'timing/model': 2.7767988480627537, 'learning_rate/qwen_vl_interface': 5.968990622152341e-06, 'learning_rate/action_model': 5.9689906221523404e-05, 'epoch': 4.03}) +[2026-06-20 04:37:38,031][starVLA.training.train_starvla][INFO] - Step 2314, Loss: {'train/loss': 0.08551871031522751, 'train/grad_norm_pre_clip': 16.369226455688477, 'timing/data': 0.006602982059121132, 'timing/model': 2.7746564131230116, 'learning_rate/qwen_vl_interface': 5.965980143338774e-06, 'learning_rate/action_model': 5.965980143338774e-05, 'epoch': 4.03}) +[2026-06-20 04:37:48,925][starVLA.training.train_starvla][INFO] - Step 2315, Loss: {'train/loss': 0.12335880100727081, 'train/grad_norm_pre_clip': 135.76702880859375, 'timing/data': 0.006796481087803841, 'timing/model': 2.7678621988743544, 'learning_rate/qwen_vl_interface': 5.962969370212819e-06, 'learning_rate/action_model': 5.9629693702128185e-05, 'epoch': 4.03}) +[2026-06-20 04:37:59,968][starVLA.training.train_starvla][INFO] - Step 2316, Loss: {'train/loss': 0.3584597408771515, 'train/grad_norm_pre_clip': 32.5564079284668, 'timing/data': 0.00694916769862175, 'timing/model': 2.778825404122472, 'learning_rate/qwen_vl_interface': 5.959958304012085e-06, 'learning_rate/action_model': 5.9599583040120845e-05, 'epoch': 4.03}) +[2026-06-20 04:38:10,651][starVLA.training.train_starvla][INFO] - Step 2317, Loss: {'train/loss': 0.15369178354740143, 'train/grad_norm_pre_clip': 30.782014846801758, 'timing/data': 0.006447359919548035, 'timing/model': 2.6532324086874723, 'learning_rate/qwen_vl_interface': 5.956946945974309e-06, 'learning_rate/action_model': 5.9569469459743085e-05, 'epoch': 4.04}) +[2026-06-20 04:38:21,579][starVLA.training.train_starvla][INFO] - Step 2318, Loss: {'train/loss': 0.03779572993516922, 'train/grad_norm_pre_clip': 7.438839912414551, 'timing/data': 0.006531480699777603, 'timing/model': 2.9019216392189264, 'learning_rate/qwen_vl_interface': 5.953935297337347e-06, 'learning_rate/action_model': 5.953935297337347e-05, 'epoch': 4.04}) +[2026-06-20 04:38:32,349][starVLA.training.train_starvla][INFO] - Step 2319, Loss: {'train/loss': 0.2104591727256775, 'train/grad_norm_pre_clip': 18.5015926361084, 'timing/data': 0.006600182503461838, 'timing/model': 2.7045502550899982, 'learning_rate/qwen_vl_interface': 5.950923359339175e-06, 'learning_rate/action_model': 5.950923359339174e-05, 'epoch': 4.04}) +[2026-06-20 04:38:43,118][starVLA.training.train_starvla][INFO] - Step 2320, Loss: {'train/loss': 0.523860514163971, 'train/grad_norm_pre_clip': 19.723363876342773, 'timing/data': 0.006452880799770355, 'timing/model': 2.765082571655512, 'learning_rate/qwen_vl_interface': 5.947911133217886e-06, 'learning_rate/action_model': 5.947911133217886e-05, 'epoch': 4.04}) +[2026-06-20 04:38:53,889][starVLA.training.train_starvla][INFO] - Step 2321, Loss: {'train/loss': 0.148894265294075, 'train/grad_norm_pre_clip': 12.659723281860352, 'timing/data': 0.006809080019593239, 'timing/model': 2.7387323919683695, 'learning_rate/qwen_vl_interface': 5.944898620211693e-06, 'learning_rate/action_model': 5.944898620211693e-05, 'epoch': 4.04}) +[2026-06-20 04:39:04,836][starVLA.training.train_starvla][INFO] - Step 2322, Loss: {'train/loss': 0.1929086446762085, 'train/grad_norm_pre_clip': 18.898757934570312, 'timing/data': 0.006562460213899612, 'timing/model': 2.7209298722445965, 'learning_rate/qwen_vl_interface': 5.941885821558925e-06, 'learning_rate/action_model': 5.9418858215589245e-05, 'epoch': 4.05}) +[2026-06-20 04:39:15,624][starVLA.training.train_starvla][INFO] - Step 2323, Loss: {'train/loss': 0.025837164372205734, 'train/grad_norm_pre_clip': 4.805911064147949, 'timing/data': 0.006673598662018776, 'timing/model': 2.775490403175354, 'learning_rate/qwen_vl_interface': 5.938872738498031e-06, 'learning_rate/action_model': 5.93887273849803e-05, 'epoch': 4.05}) +[2026-06-20 04:39:26,254][starVLA.training.train_starvla][INFO] - Step 2324, Loss: {'train/loss': 0.11548706144094467, 'train/grad_norm_pre_clip': 10.964041709899902, 'timing/data': 0.006669407710433006, 'timing/model': 2.6965720802545547, 'learning_rate/qwen_vl_interface': 5.935859372267575e-06, 'learning_rate/action_model': 5.935859372267575e-05, 'epoch': 4.05}) +[2026-06-20 04:39:37,033][starVLA.training.train_starvla][INFO] - Step 2325, Loss: {'train/loss': 0.060858651995658875, 'train/grad_norm_pre_clip': 8.353705406188965, 'timing/data': 0.006740720942616463, 'timing/model': 2.7877871859818697, 'learning_rate/qwen_vl_interface': 5.9328457241062396e-06, 'learning_rate/action_model': 5.9328457241062394e-05, 'epoch': 4.05}) +[2026-06-20 04:39:48,130][starVLA.training.train_starvla][INFO] - Step 2326, Loss: {'train/loss': 0.08740440756082535, 'train/grad_norm_pre_clip': 9.242222785949707, 'timing/data': 0.006719667464494705, 'timing/model': 2.772726748138666, 'learning_rate/qwen_vl_interface': 5.929831795252819e-06, 'learning_rate/action_model': 5.9298317952528184e-05, 'epoch': 4.05}) +[2026-06-20 04:39:59,032][starVLA.training.train_starvla][INFO] - Step 2327, Loss: {'train/loss': 0.233995258808136, 'train/grad_norm_pre_clip': 21.560457229614258, 'timing/data': 0.006617151200771332, 'timing/model': 2.775183754041791, 'learning_rate/qwen_vl_interface': 5.926817586946227e-06, 'learning_rate/action_model': 5.926817586946227e-05, 'epoch': 4.05}) +[2026-06-20 04:40:09,923][starVLA.training.train_starvla][INFO] - Step 2328, Loss: {'train/loss': 0.11385197192430496, 'train/grad_norm_pre_clip': 10.080092430114746, 'timing/data': 0.006651628762483597, 'timing/model': 2.7673861291259527, 'learning_rate/qwen_vl_interface': 5.923803100425491e-06, 'learning_rate/action_model': 5.92380310042549e-05, 'epoch': 4.06}) +[2026-06-20 04:40:20,833][starVLA.training.train_starvla][INFO] - Step 2329, Loss: {'train/loss': 0.07903898507356644, 'train/grad_norm_pre_clip': 8.72570514678955, 'timing/data': 0.006752068176865578, 'timing/model': 2.773963524028659, 'learning_rate/qwen_vl_interface': 5.92078833692975e-06, 'learning_rate/action_model': 5.92078833692975e-05, 'epoch': 4.06}) +[2026-06-20 04:40:31,832][starVLA.training.train_starvla][INFO] - Step 2330, Loss: {'train/loss': 0.05761871114373207, 'train/grad_norm_pre_clip': 22.243776321411133, 'timing/data': 0.006621452048420906, 'timing/model': 2.7866856902837753, 'learning_rate/qwen_vl_interface': 5.917773297698264e-06, 'learning_rate/action_model': 5.9177732976982636e-05, 'epoch': 4.06}) +[2026-06-20 04:40:42,735][starVLA.training.train_starvla][INFO] - Step 2331, Loss: {'train/loss': 0.04107361286878586, 'train/grad_norm_pre_clip': 4.450350284576416, 'timing/data': 0.00685855932533741, 'timing/model': 2.782278139144182, 'learning_rate/qwen_vl_interface': 5.9147579839704e-06, 'learning_rate/action_model': 5.9147579839704e-05, 'epoch': 4.06}) +[2026-06-20 04:40:53,629][starVLA.training.train_starvla][INFO] - Step 2332, Loss: {'train/loss': 0.25882741808891296, 'train/grad_norm_pre_clip': 21.51420783996582, 'timing/data': 0.006976006552577019, 'timing/model': 2.776135738939047, 'learning_rate/qwen_vl_interface': 5.911742396985637e-06, 'learning_rate/action_model': 5.9117423969856364e-05, 'epoch': 4.06}) +[2026-06-20 04:41:04,332][starVLA.training.train_starvla][INFO] - Step 2333, Loss: {'train/loss': 0.05855261906981468, 'train/grad_norm_pre_clip': 8.685850143432617, 'timing/data': 0.006614841520786285, 'timing/model': 2.7738351430743933, 'learning_rate/qwen_vl_interface': 5.908726537983571e-06, 'learning_rate/action_model': 5.908726537983571e-05, 'epoch': 4.06}) +[2026-06-20 04:41:15,226][starVLA.training.train_starvla][INFO] - Step 2334, Loss: {'train/loss': 0.12174046039581299, 'train/grad_norm_pre_clip': 16.44234275817871, 'timing/data': 0.00656525231897831, 'timing/model': 2.7818776108324528, 'learning_rate/qwen_vl_interface': 5.9057104082039085e-06, 'learning_rate/action_model': 5.905710408203908e-05, 'epoch': 4.07}) +[2026-06-20 04:41:26,135][starVLA.training.train_starvla][INFO] - Step 2335, Loss: {'train/loss': 0.05782518908381462, 'train/grad_norm_pre_clip': 11.485840797424316, 'timing/data': 0.006667150184512138, 'timing/model': 2.781610105186701, 'learning_rate/qwen_vl_interface': 5.902694008886466e-06, 'learning_rate/action_model': 5.902694008886466e-05, 'epoch': 4.07}) +[2026-06-20 04:41:36,947][starVLA.training.train_starvla][INFO] - Step 2336, Loss: {'train/loss': 0.09880120307207108, 'train/grad_norm_pre_clip': 20.100688934326172, 'timing/data': 0.006848396733403206, 'timing/model': 2.7260928079485893, 'learning_rate/qwen_vl_interface': 5.8996773412711694e-06, 'learning_rate/action_model': 5.899677341271169e-05, 'epoch': 4.07}) +[2026-06-20 04:41:47,947][starVLA.training.train_starvla][INFO] - Step 2337, Loss: {'train/loss': 0.16063186526298523, 'train/grad_norm_pre_clip': 19.338159561157227, 'timing/data': 0.0067462008446455, 'timing/model': 3.0158140156418085, 'learning_rate/qwen_vl_interface': 5.896660406598059e-06, 'learning_rate/action_model': 5.896660406598059e-05, 'epoch': 4.07}) +[2026-06-20 04:41:58,834][starVLA.training.train_starvla][INFO] - Step 2338, Loss: {'train/loss': 0.17040880024433136, 'train/grad_norm_pre_clip': 15.062195777893066, 'timing/data': 0.0067381300032138824, 'timing/model': 2.7785597946494818, 'learning_rate/qwen_vl_interface': 5.893643206107283e-06, 'learning_rate/action_model': 5.893643206107282e-05, 'epoch': 4.07}) +[2026-06-20 04:42:09,736][starVLA.training.train_starvla][INFO] - Step 2339, Loss: {'train/loss': 0.17557011544704437, 'train/grad_norm_pre_clip': 14.613560676574707, 'timing/data': 0.007570130750536919, 'timing/model': 2.777138886973262, 'learning_rate/qwen_vl_interface': 5.890625741039096e-06, 'learning_rate/action_model': 5.890625741039095e-05, 'epoch': 4.07}) +[2026-06-20 04:42:20,646][starVLA.training.train_starvla][INFO] - Step 2340, Loss: {'train/loss': 0.4062695801258087, 'train/grad_norm_pre_clip': 29.752094268798828, 'timing/data': 0.0065480489283800125, 'timing/model': 2.78559154458344, 'learning_rate/qwen_vl_interface': 5.887608012633864e-06, 'learning_rate/action_model': 5.887608012633864e-05, 'epoch': 4.08}) +[2026-06-20 04:42:31,492][starVLA.training.train_starvla][INFO] - Step 2341, Loss: {'train/loss': 0.13105592131614685, 'train/grad_norm_pre_clip': 20.645944595336914, 'timing/data': 0.006735699251294136, 'timing/model': 2.7090934831649065, 'learning_rate/qwen_vl_interface': 5.884590022132064e-06, 'learning_rate/action_model': 5.884590022132064e-05, 'epoch': 4.08}) +[2026-06-20 04:42:42,082][starVLA.training.train_starvla][INFO] - Step 2342, Loss: {'train/loss': 0.07321061939001083, 'train/grad_norm_pre_clip': 12.257654190063477, 'timing/data': 0.006754670292139053, 'timing/model': 2.7163856849074364, 'learning_rate/qwen_vl_interface': 5.881571770774278e-06, 'learning_rate/action_model': 5.881571770774278e-05, 'epoch': 4.08}) +[2026-06-20 04:42:52,740][starVLA.training.train_starvla][INFO] - Step 2343, Loss: {'train/loss': 0.11918353289365768, 'train/grad_norm_pre_clip': 10.142727851867676, 'timing/data': 0.006907988339662552, 'timing/model': 2.79717861674726, 'learning_rate/qwen_vl_interface': 5.878553259801192e-06, 'learning_rate/action_model': 5.878553259801192e-05, 'epoch': 4.08}) +[2026-06-20 04:43:03,637][starVLA.training.train_starvla][INFO] - Step 2344, Loss: {'train/loss': 0.09567790478467941, 'train/grad_norm_pre_clip': 12.579485893249512, 'timing/data': 0.0068815965205430984, 'timing/model': 2.7812409847974777, 'learning_rate/qwen_vl_interface': 5.8755344904536025e-06, 'learning_rate/action_model': 5.8755344904536025e-05, 'epoch': 4.08}) +[2026-06-20 04:43:14,233][starVLA.training.train_starvla][INFO] - Step 2345, Loss: {'train/loss': 0.22788041830062866, 'train/grad_norm_pre_clip': 16.614656448364258, 'timing/data': 0.006776459515094757, 'timing/model': 2.757483234629035, 'learning_rate/qwen_vl_interface': 5.872515463972415e-06, 'learning_rate/action_model': 5.8725154639724146e-05, 'epoch': 4.09}) +[2026-06-20 04:43:25,036][starVLA.training.train_starvla][INFO] - Step 2346, Loss: {'train/loss': 0.22338372468948364, 'train/grad_norm_pre_clip': 12.267091751098633, 'timing/data': 0.007087435573339462, 'timing/model': 2.77640195004642, 'learning_rate/qwen_vl_interface': 5.869496181598633e-06, 'learning_rate/action_model': 5.869496181598633e-05, 'epoch': 4.09}) +[2026-06-20 04:43:35,925][starVLA.training.train_starvla][INFO] - Step 2347, Loss: {'train/loss': 0.21422690153121948, 'train/grad_norm_pre_clip': 14.14632797241211, 'timing/data': 0.00690164789557457, 'timing/model': 2.7680179104208946, 'learning_rate/qwen_vl_interface': 5.866476644573371e-06, 'learning_rate/action_model': 5.866476644573371e-05, 'epoch': 4.09}) +[2026-06-20 04:43:46,834][starVLA.training.train_starvla][INFO] - Step 2348, Loss: {'train/loss': 0.023254087194800377, 'train/grad_norm_pre_clip': 3.7649452686309814, 'timing/data': 0.006305484101176262, 'timing/model': 2.7706727553159, 'learning_rate/qwen_vl_interface': 5.863456854137848e-06, 'learning_rate/action_model': 5.8634568541378474e-05, 'epoch': 4.09}) +[2026-06-20 04:43:57,921][starVLA.training.train_starvla][INFO] - Step 2349, Loss: {'train/loss': 0.105342335999012, 'train/grad_norm_pre_clip': 10.908504486083984, 'timing/data': 0.006764499470591545, 'timing/model': 2.7691392097622156, 'learning_rate/qwen_vl_interface': 5.860436811533383e-06, 'learning_rate/action_model': 5.860436811533383e-05, 'epoch': 4.09}) +[2026-06-20 04:44:08,736][starVLA.training.train_starvla][INFO] - Step 2350, Loss: {'train/loss': 0.023818567395210266, 'train/grad_norm_pre_clip': 5.099541187286377, 'timing/data': 0.009190572425723076, 'timing/model': 2.774480026215315, 'learning_rate/qwen_vl_interface': 5.857416518001404e-06, 'learning_rate/action_model': 5.857416518001404e-05, 'epoch': 4.09}) +[2026-06-20 04:44:19,633][starVLA.training.train_starvla][INFO] - Step 2351, Loss: {'train/loss': 0.01494934968650341, 'train/grad_norm_pre_clip': 2.2678821086883545, 'timing/data': 0.006663905456662178, 'timing/model': 2.7784920521080494, 'learning_rate/qwen_vl_interface': 5.854395974783439e-06, 'learning_rate/action_model': 5.8543959747834386e-05, 'epoch': 4.1}) +[2026-06-20 04:44:30,524][starVLA.training.train_starvla][INFO] - Step 2352, Loss: {'train/loss': 0.1854572743177414, 'train/grad_norm_pre_clip': 14.825457572937012, 'timing/data': 0.006757434457540512, 'timing/model': 2.7766397017985582, 'learning_rate/qwen_vl_interface': 5.851375183121119e-06, 'learning_rate/action_model': 5.8513751831211186e-05, 'epoch': 4.1}) +[2026-06-20 04:44:41,437][starVLA.training.train_starvla][INFO] - Step 2353, Loss: {'train/loss': 0.11809724569320679, 'train/grad_norm_pre_clip': 10.04238224029541, 'timing/data': 0.006574276834726334, 'timing/model': 2.776046793907881, 'learning_rate/qwen_vl_interface': 5.848354144256175e-06, 'learning_rate/action_model': 5.8483541442561754e-05, 'epoch': 4.1}) +[2026-06-20 04:44:52,321][starVLA.training.train_starvla][INFO] - Step 2354, Loss: {'train/loss': 0.3877201974391937, 'train/grad_norm_pre_clip': 33.28348159790039, 'timing/data': 0.006609244272112846, 'timing/model': 2.7665242329239845, 'learning_rate/qwen_vl_interface': 5.845332859430446e-06, 'learning_rate/action_model': 5.845332859430446e-05, 'epoch': 4.1}) +[2026-06-20 04:45:03,136][starVLA.training.train_starvla][INFO] - Step 2355, Loss: {'train/loss': 0.1599549800157547, 'train/grad_norm_pre_clip': 9.612469673156738, 'timing/data': 0.006644250825047493, 'timing/model': 2.7767031770199537, 'learning_rate/qwen_vl_interface': 5.842311329885866e-06, 'learning_rate/action_model': 5.842311329885866e-05, 'epoch': 4.1}) +[2026-06-20 04:45:14,289][starVLA.training.train_starvla][INFO] - Step 2356, Loss: {'train/loss': 0.13678675889968872, 'train/grad_norm_pre_clip': 9.028108596801758, 'timing/data': 0.007072350010275841, 'timing/model': 3.0290134754031897, 'learning_rate/qwen_vl_interface': 5.839289556864472e-06, 'learning_rate/action_model': 5.8392895568644714e-05, 'epoch': 4.1}) +[2026-06-20 04:45:25,142][starVLA.training.train_starvla][INFO] - Step 2357, Loss: {'train/loss': 0.026056697592139244, 'train/grad_norm_pre_clip': 3.9798026084899902, 'timing/data': 0.007179057225584984, 'timing/model': 2.7866472583264112, 'learning_rate/qwen_vl_interface': 5.836267541608398e-06, 'learning_rate/action_model': 5.836267541608398e-05, 'epoch': 4.11}) +[2026-06-20 04:45:36,038][starVLA.training.train_starvla][INFO] - Step 2358, Loss: {'train/loss': 0.2770375609397888, 'train/grad_norm_pre_clip': 15.448328018188477, 'timing/data': 0.007081527262926102, 'timing/model': 2.7727703768759966, 'learning_rate/qwen_vl_interface': 5.833245285359887e-06, 'learning_rate/action_model': 5.833245285359886e-05, 'epoch': 4.11}) +[2026-06-20 04:45:46,932][starVLA.training.train_starvla][INFO] - Step 2359, Loss: {'train/loss': 0.1553257703781128, 'train/grad_norm_pre_clip': 9.113811492919922, 'timing/data': 0.007005399093031883, 'timing/model': 2.772148597985506, 'learning_rate/qwen_vl_interface': 5.830222789361268e-06, 'learning_rate/action_model': 5.8302227893612684e-05, 'epoch': 4.11}) +[2026-06-20 04:45:58,032][starVLA.training.train_starvla][INFO] - Step 2360, Loss: {'train/loss': 0.08537841588258743, 'train/grad_norm_pre_clip': 7.404664516448975, 'timing/data': 0.007091136649250984, 'timing/model': 2.7364912796765566, 'learning_rate/qwen_vl_interface': 5.8272000548549776e-06, 'learning_rate/action_model': 5.827200054854978e-05, 'epoch': 4.11}) +[2026-06-20 04:46:08,935][starVLA.training.train_starvla][INFO] - Step 2361, Loss: {'train/loss': 0.1764031946659088, 'train/grad_norm_pre_clip': 15.206047058105469, 'timing/data': 0.006987139582633972, 'timing/model': 2.771485447883606, 'learning_rate/qwen_vl_interface': 5.824177083083547e-06, 'learning_rate/action_model': 5.8241770830835475e-05, 'epoch': 4.11}) +[2026-06-20 04:46:19,828][starVLA.training.train_starvla][INFO] - Step 2362, Loss: {'train/loss': 0.057487018406391144, 'train/grad_norm_pre_clip': 8.447968482971191, 'timing/data': 0.007040880620479584, 'timing/model': 2.774069629609585, 'learning_rate/qwen_vl_interface': 5.8211538752896056e-06, 'learning_rate/action_model': 5.821153875289606e-05, 'epoch': 4.11}) +[2026-06-20 04:46:30,737][starVLA.training.train_starvla][INFO] - Step 2363, Loss: {'train/loss': 0.1701173186302185, 'train/grad_norm_pre_clip': 20.27176284790039, 'timing/data': 0.007047818973660469, 'timing/model': 2.780194191262126, 'learning_rate/qwen_vl_interface': 5.8181304327158814e-06, 'learning_rate/action_model': 5.8181304327158814e-05, 'epoch': 4.12}) +[2026-06-20 04:46:41,622][starVLA.training.train_starvla][INFO] - Step 2364, Loss: {'train/loss': 0.03600915148854256, 'train/grad_norm_pre_clip': 7.67948579788208, 'timing/data': 0.008276505395770073, 'timing/model': 2.7518856693059206, 'learning_rate/qwen_vl_interface': 5.8151067566051965e-06, 'learning_rate/action_model': 5.815106756605196e-05, 'epoch': 4.12}) +[2026-06-20 04:46:52,521][starVLA.training.train_starvla][INFO] - Step 2365, Loss: {'train/loss': 0.010372587479650974, 'train/grad_norm_pre_clip': 2.2062320709228516, 'timing/data': 0.0062882136553525925, 'timing/model': 2.768134256824851, 'learning_rate/qwen_vl_interface': 5.81208284820047e-06, 'learning_rate/action_model': 5.81208284820047e-05, 'epoch': 4.12}) +[2026-06-20 04:47:03,334][starVLA.training.train_starvla][INFO] - Step 2366, Loss: {'train/loss': 0.2675115168094635, 'train/grad_norm_pre_clip': 16.14897918701172, 'timing/data': 0.007051745429635048, 'timing/model': 2.774725917726755, 'learning_rate/qwen_vl_interface': 5.809058708744716e-06, 'learning_rate/action_model': 5.809058708744716e-05, 'epoch': 4.12}) +[2026-06-20 04:47:14,118][starVLA.training.train_starvla][INFO] - Step 2367, Loss: {'train/loss': 0.21295678615570068, 'train/grad_norm_pre_clip': 16.783035278320312, 'timing/data': 0.006548730656504631, 'timing/model': 2.7424028608947992, 'learning_rate/qwen_vl_interface': 5.8060343394810445e-06, 'learning_rate/action_model': 5.806034339481045e-05, 'epoch': 4.12}) +[2026-06-20 04:47:24,857][starVLA.training.train_starvla][INFO] - Step 2368, Loss: {'train/loss': 0.230222687125206, 'train/grad_norm_pre_clip': 15.287303924560547, 'timing/data': 0.0066344719380140305, 'timing/model': 2.6474629566073418, 'learning_rate/qwen_vl_interface': 5.803009741652659e-06, 'learning_rate/action_model': 5.8030097416526585e-05, 'epoch': 4.13}) +[2026-06-20 04:47:35,627][starVLA.training.train_starvla][INFO] - Step 2369, Loss: {'train/loss': 0.09612400084733963, 'train/grad_norm_pre_clip': 7.941519260406494, 'timing/data': 0.007040208205580711, 'timing/model': 2.749167224392295, 'learning_rate/qwen_vl_interface': 5.799984916502859e-06, 'learning_rate/action_model': 5.799984916502859e-05, 'epoch': 4.13}) +[2026-06-20 04:47:46,322][starVLA.training.train_starvla][INFO] - Step 2370, Loss: {'train/loss': 0.024237707257270813, 'train/grad_norm_pre_clip': 3.601461410522461, 'timing/data': 0.0068485066294670105, 'timing/model': 2.7653351109474897, 'learning_rate/qwen_vl_interface': 5.7969598652750334e-06, 'learning_rate/action_model': 5.796959865275033e-05, 'epoch': 4.13}) +[2026-06-20 04:47:56,854][starVLA.training.train_starvla][INFO] - Step 2371, Loss: {'train/loss': 0.12748242914676666, 'train/grad_norm_pre_clip': 7.392415523529053, 'timing/data': 0.006737038493156433, 'timing/model': 2.675332561135292, 'learning_rate/qwen_vl_interface': 5.793934589212666e-06, 'learning_rate/action_model': 5.793934589212666e-05, 'epoch': 4.13}) +[2026-06-20 04:48:07,723][starVLA.training.train_starvla][INFO] - Step 2372, Loss: {'train/loss': 0.1083858534693718, 'train/grad_norm_pre_clip': 17.883352279663086, 'timing/data': 0.006766609847545624, 'timing/model': 2.774299431592226, 'learning_rate/qwen_vl_interface': 5.7909090895593375e-06, 'learning_rate/action_model': 5.790909089559338e-05, 'epoch': 4.13}) +[2026-06-20 04:48:18,623][starVLA.training.train_starvla][INFO] - Step 2373, Loss: {'train/loss': 0.23829609155654907, 'train/grad_norm_pre_clip': 20.505399703979492, 'timing/data': 0.006520163267850876, 'timing/model': 2.7715825475752354, 'learning_rate/qwen_vl_interface': 5.7878833675587135e-06, 'learning_rate/action_model': 5.787883367558713e-05, 'epoch': 4.13}) +[2026-06-20 04:48:29,401][starVLA.training.train_starvla][INFO] - Step 2374, Loss: {'train/loss': 0.2807323634624481, 'train/grad_norm_pre_clip': 21.777711868286133, 'timing/data': 0.006596019491553307, 'timing/model': 2.7380868159234524, 'learning_rate/qwen_vl_interface': 5.784857424454556e-06, 'learning_rate/action_model': 5.7848574244545563e-05, 'epoch': 4.14}) +[2026-06-20 04:48:40,251][starVLA.training.train_starvla][INFO] - Step 2375, Loss: {'train/loss': 0.11206502467393875, 'train/grad_norm_pre_clip': 5.980494976043701, 'timing/data': 0.006883155554533005, 'timing/model': 3.0044102650135756, 'learning_rate/qwen_vl_interface': 5.7818312614907125e-06, 'learning_rate/action_model': 5.781831261490712e-05, 'epoch': 4.14}) +[2026-06-20 04:48:50,853][starVLA.training.train_starvla][INFO] - Step 2376, Loss: {'train/loss': 0.10262031853199005, 'train/grad_norm_pre_clip': 16.82750129699707, 'timing/data': 0.0066369082778692245, 'timing/model': 2.6771503183990717, 'learning_rate/qwen_vl_interface': 5.778804879911129e-06, 'learning_rate/action_model': 5.778804879911128e-05, 'epoch': 4.14}) +[2026-06-20 04:49:01,555][starVLA.training.train_starvla][INFO] - Step 2377, Loss: {'train/loss': 0.13915100693702698, 'train/grad_norm_pre_clip': 12.727182388305664, 'timing/data': 0.0065926797688007355, 'timing/model': 2.6444878596812487, 'learning_rate/qwen_vl_interface': 5.775778280959832e-06, 'learning_rate/action_model': 5.7757782809598315e-05, 'epoch': 4.14}) +[2026-06-20 04:49:12,360][starVLA.training.train_starvla][INFO] - Step 2378, Loss: {'train/loss': 0.08197443187236786, 'train/grad_norm_pre_clip': 10.31649112701416, 'timing/data': 0.006767241284251213, 'timing/model': 2.715112352743745, 'learning_rate/qwen_vl_interface': 5.7727514658809435e-06, 'learning_rate/action_model': 5.772751465880944e-05, 'epoch': 4.14}) +[2026-06-20 04:49:23,291][starVLA.training.train_starvla][INFO] - Step 2379, Loss: {'train/loss': 0.027089644223451614, 'train/grad_norm_pre_clip': 3.087794303894043, 'timing/data': 0.006680388003587723, 'timing/model': 2.7053892835974693, 'learning_rate/qwen_vl_interface': 5.769724435918676e-06, 'learning_rate/action_model': 5.7697244359186755e-05, 'epoch': 4.14}) +[2026-06-20 04:49:34,063][starVLA.training.train_starvla][INFO] - Step 2380, Loss: {'train/loss': 0.10105925798416138, 'train/grad_norm_pre_clip': 13.847381591796875, 'timing/data': 0.0064947400242090225, 'timing/model': 2.7063127383589745, 'learning_rate/qwen_vl_interface': 5.766697192317323e-06, 'learning_rate/action_model': 5.766697192317323e-05, 'epoch': 4.15}) +[2026-06-20 04:49:44,933][starVLA.training.train_starvla][INFO] - Step 2381, Loss: {'train/loss': 0.11587274819612503, 'train/grad_norm_pre_clip': 6.641198635101318, 'timing/data': 0.006560510024428368, 'timing/model': 2.774038165807724, 'learning_rate/qwen_vl_interface': 5.763669736321271e-06, 'learning_rate/action_model': 5.763669736321271e-05, 'epoch': 4.15}) +[2026-06-20 04:49:55,828][starVLA.training.train_starvla][INFO] - Step 2382, Loss: {'train/loss': 0.0463036485016346, 'train/grad_norm_pre_clip': 9.098451614379883, 'timing/data': 0.006696099415421486, 'timing/model': 2.782315282151103, 'learning_rate/qwen_vl_interface': 5.760642069174994e-06, 'learning_rate/action_model': 5.760642069174994e-05, 'epoch': 4.15}) +[2026-06-20 04:50:06,735][starVLA.training.train_starvla][INFO] - Step 2383, Loss: {'train/loss': 0.0488312728703022, 'train/grad_norm_pre_clip': 7.954839706420898, 'timing/data': 0.0066628772765398026, 'timing/model': 2.7796845324337482, 'learning_rate/qwen_vl_interface': 5.757614192123051e-06, 'learning_rate/action_model': 5.757614192123051e-05, 'epoch': 4.15}) +[2026-06-20 04:50:17,630][starVLA.training.train_starvla][INFO] - Step 2384, Loss: {'train/loss': 0.023966332897543907, 'train/grad_norm_pre_clip': 11.784860610961914, 'timing/data': 0.006551900878548622, 'timing/model': 2.770134247839451, 'learning_rate/qwen_vl_interface': 5.754586106410089e-06, 'learning_rate/action_model': 5.754586106410088e-05, 'epoch': 4.15}) +[2026-06-20 04:50:28,523][starVLA.training.train_starvla][INFO] - Step 2385, Loss: {'train/loss': 0.1140584796667099, 'train/grad_norm_pre_clip': 10.98414134979248, 'timing/data': 0.006448289379477501, 'timing/model': 2.7650052551180124, 'learning_rate/qwen_vl_interface': 5.751557813280838e-06, 'learning_rate/action_model': 5.751557813280838e-05, 'epoch': 4.15}) +[2026-06-20 04:50:39,434][starVLA.training.train_starvla][INFO] - Step 2386, Loss: {'train/loss': 0.19435983896255493, 'train/grad_norm_pre_clip': 34.67481994628906, 'timing/data': 0.006719918921589851, 'timing/model': 2.7787156719714403, 'learning_rate/qwen_vl_interface': 5.7485293139801146e-06, 'learning_rate/action_model': 5.7485293139801146e-05, 'epoch': 4.16}) +[2026-06-20 04:50:50,536][starVLA.training.train_starvla][INFO] - Step 2387, Loss: {'train/loss': 0.15827220678329468, 'train/grad_norm_pre_clip': 18.7952823638916, 'timing/data': 0.006683835759758949, 'timing/model': 2.77950500510633, 'learning_rate/qwen_vl_interface': 5.745500609752821e-06, 'learning_rate/action_model': 5.745500609752821e-05, 'epoch': 4.16}) +[2026-06-20 04:51:01,431][starVLA.training.train_starvla][INFO] - Step 2388, Loss: {'train/loss': 0.05692579597234726, 'train/grad_norm_pre_clip': 11.579059600830078, 'timing/data': 0.006444947794079781, 'timing/model': 2.7773613445460796, 'learning_rate/qwen_vl_interface': 5.742471701843943e-06, 'learning_rate/action_model': 5.742471701843943e-05, 'epoch': 4.16}) +[2026-06-20 04:51:12,333][starVLA.training.train_starvla][INFO] - Step 2389, Loss: {'train/loss': 0.3986881971359253, 'train/grad_norm_pre_clip': 28.56842803955078, 'timing/data': 0.006461316719651222, 'timing/model': 2.7843186631798744, 'learning_rate/qwen_vl_interface': 5.739442591498553e-06, 'learning_rate/action_model': 5.739442591498553e-05, 'epoch': 4.16}) +[2026-06-20 04:51:23,236][starVLA.training.train_starvla][INFO] - Step 2390, Loss: {'train/loss': 0.050592049956321716, 'train/grad_norm_pre_clip': 6.747501850128174, 'timing/data': 0.006650187075138092, 'timing/model': 2.7796565983444452, 'learning_rate/qwen_vl_interface': 5.7364132799617985e-06, 'learning_rate/action_model': 5.7364132799617984e-05, 'epoch': 4.16}) +[2026-06-20 04:51:34,128][starVLA.training.train_starvla][INFO] - Step 2391, Loss: {'train/loss': 0.1965169459581375, 'train/grad_norm_pre_clip': 19.44631004333496, 'timing/data': 0.006438247859477997, 'timing/model': 2.775200594216585, 'learning_rate/qwen_vl_interface': 5.733383768478923e-06, 'learning_rate/action_model': 5.7333837684789225e-05, 'epoch': 4.17}) +[2026-06-20 04:51:45,033][starVLA.training.train_starvla][INFO] - Step 2392, Loss: {'train/loss': 0.112244613468647, 'train/grad_norm_pre_clip': 24.155418395996094, 'timing/data': 0.006638256832957268, 'timing/model': 2.772193206474185, 'learning_rate/qwen_vl_interface': 5.730354058295234e-06, 'learning_rate/action_model': 5.730354058295234e-05, 'epoch': 4.17}) +[2026-06-20 04:51:55,922][starVLA.training.train_starvla][INFO] - Step 2393, Loss: {'train/loss': 0.09367241710424423, 'train/grad_norm_pre_clip': 6.629029273986816, 'timing/data': 0.0066474974155426025, 'timing/model': 2.773880748078227, 'learning_rate/qwen_vl_interface': 5.72732415065614e-06, 'learning_rate/action_model': 5.727324150656139e-05, 'epoch': 4.17}) +[2026-06-20 04:52:07,043][starVLA.training.train_starvla][INFO] - Step 2394, Loss: {'train/loss': 0.11253292858600616, 'train/grad_norm_pre_clip': 17.723094940185547, 'timing/data': 0.006620857864618301, 'timing/model': 2.9900296851992607, 'learning_rate/qwen_vl_interface': 5.724294046807115e-06, 'learning_rate/action_model': 5.724294046807115e-05, 'epoch': 4.17}) +[2026-06-20 04:52:17,940][starVLA.training.train_starvla][INFO] - Step 2395, Loss: {'train/loss': 0.06908655166625977, 'train/grad_norm_pre_clip': 10.510076522827148, 'timing/data': 0.006664615124464035, 'timing/model': 2.7801355104893446, 'learning_rate/qwen_vl_interface': 5.7212637479937234e-06, 'learning_rate/action_model': 5.721263747993724e-05, 'epoch': 4.17}) +[2026-06-20 04:52:28,837][starVLA.training.train_starvla][INFO] - Step 2396, Loss: {'train/loss': 0.1377526819705963, 'train/grad_norm_pre_clip': 11.917459487915039, 'timing/data': 0.006512777879834175, 'timing/model': 2.780346566811204, 'learning_rate/qwen_vl_interface': 5.718233255461607e-06, 'learning_rate/action_model': 5.718233255461607e-05, 'epoch': 4.17}) +[2026-06-20 04:52:39,735][starVLA.training.train_starvla][INFO] - Step 2397, Loss: {'train/loss': 0.1354372203350067, 'train/grad_norm_pre_clip': 15.822611808776855, 'timing/data': 0.006836105138063431, 'timing/model': 2.7793052662163973, 'learning_rate/qwen_vl_interface': 5.715202570456486e-06, 'learning_rate/action_model': 5.715202570456486e-05, 'epoch': 4.18}) +[2026-06-20 04:52:50,835][starVLA.training.train_starvla][INFO] - Step 2398, Loss: {'train/loss': 0.07766734808683395, 'train/grad_norm_pre_clip': 7.684608459472656, 'timing/data': 0.006697148084640503, 'timing/model': 2.7556112930178642, 'learning_rate/qwen_vl_interface': 5.7121716942241615e-06, 'learning_rate/action_model': 5.712171694224161e-05, 'epoch': 4.18}) +[2026-06-20 04:53:01,734][starVLA.training.train_starvla][INFO] - Step 2399, Loss: {'train/loss': 0.14370806515216827, 'train/grad_norm_pre_clip': 14.993456840515137, 'timing/data': 0.006660623475909233, 'timing/model': 2.770260103046894, 'learning_rate/qwen_vl_interface': 5.709140628010509e-06, 'learning_rate/action_model': 5.7091406280105085e-05, 'epoch': 4.18}) +[2026-06-20 04:53:12,622][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 2400 +[2026-06-20 05:00:44,473][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 2400 +[2026-06-20 05:00:44,474][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 2400 +[2026-06-20 05:08:53,221][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 2400 +[2026-06-20 05:08:53,223][starVLA.training.train_starvla][INFO] - Step 2400, Loss: {'train/loss': 0.07145217806100845, 'train/grad_norm_pre_clip': 6.9147047996521, 'eval/loss': 0.6978841423988342, 'eval/latency_0/loss': 0.5987222790718079, 'eval/latency_1/loss': 0.7185165882110596, 'eval/latency_2/loss': 0.10862225294113159, 'eval/latency_3/loss': 0.8283945918083191, 'eval/latency_4/loss': 1.2724143266677856, 'eval/flappy/loss': 0.6978841423988342, 'eval/flappy/latency_0/loss': 0.5987222790718079, 'eval/flappy/latency_1/loss': 0.7185165882110596, 'eval/flappy/latency_2/loss': 0.10862225294113159, 'eval/flappy/latency_3/loss': 0.8283945918083191, 'eval/flappy/latency_4/loss': 1.2724143266677856, 'eval/flap_precision': 0.8260869565217391, 'eval/flap_recall': 0.27636363636363637, 'eval/flap_f1': 0.4141689373297002, 'eval/flap_rate_model': 0.025555555555555557, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.4141689373297002, 'eval/latency_0/flap_precision': 0.8260869565217391, 'eval/latency_0/flap_recall': 0.27636363636363637, 'eval/latency_0/flap_f1': 0.4141689373297002, 'eval/latency_0/flap_rate_model': 0.025555555555555557, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.4141689373297002, 'timing/data': 0.006251990795135498, 'timing/model': 2.7766919042915106, 'learning_rate/qwen_vl_interface': 5.706109373061489e-06, 'learning_rate/action_model': 5.706109373061489e-05, 'epoch': 4.18}) +[2026-06-20 05:08:53,223][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_2400_state +[2026-06-20 05:08:53,243][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-20 05:08:59,042][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_2400_state/model.safetensors +[2026-06-20 05:09:07,113][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_2400_state/optimizer.bin +[2026-06-20 05:09:07,113][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_2400_state/sampler.bin +[2026-06-20 05:09:07,114][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_2400_state/sampler_1.bin +[2026-06-20 05:09:07,115][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_2400_state/random_states_0.pkl +[2026-06-20 05:09:07,252][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2400 +[2026-06-20 05:09:07,253][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-20 05:09:07,267][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-20 05:09:18,121][starVLA.training.train_starvla][INFO] - Step 2401, Loss: {'train/loss': 0.024917537346482277, 'train/grad_norm_pre_clip': 4.091004848480225, 'timing/data': 0.007035421207547188, 'timing/model': 2.7872654292732477, 'learning_rate/qwen_vl_interface': 5.703077930623134e-06, 'learning_rate/action_model': 5.7030779306231344e-05, 'epoch': 4.18}) +[2026-06-20 05:09:28,922][starVLA.training.train_starvla][INFO] - Step 2402, Loss: {'train/loss': 0.06167161092162132, 'train/grad_norm_pre_clip': 10.728036880493164, 'timing/data': 0.0068229492753744125, 'timing/model': 2.777076480910182, 'learning_rate/qwen_vl_interface': 5.7000463019415565e-06, 'learning_rate/action_model': 5.7000463019415563e-05, 'epoch': 4.18}) +[2026-06-20 05:09:40,028][starVLA.training.train_starvla][INFO] - Step 2403, Loss: {'train/loss': 0.02434574067592621, 'train/grad_norm_pre_clip': 3.7796342372894287, 'timing/data': 0.006825810298323631, 'timing/model': 2.7821966484189034, 'learning_rate/qwen_vl_interface': 5.697014488262945e-06, 'learning_rate/action_model': 5.697014488262945e-05, 'epoch': 4.19}) +[2026-06-20 05:09:50,925][starVLA.training.train_starvla][INFO] - Step 2404, Loss: {'train/loss': 0.0552656315267086, 'train/grad_norm_pre_clip': 17.74817657470703, 'timing/data': 0.006956826895475388, 'timing/model': 2.769589951261878, 'learning_rate/qwen_vl_interface': 5.693982490833561e-06, 'learning_rate/action_model': 5.693982490833561e-05, 'epoch': 4.19}) +[2026-06-20 05:10:01,826][starVLA.training.train_starvla][INFO] - Step 2405, Loss: {'train/loss': 0.03354918584227562, 'train/grad_norm_pre_clip': 7.1235551834106445, 'timing/data': 0.006667371839284897, 'timing/model': 2.7770099323242903, 'learning_rate/qwen_vl_interface': 5.690950310899745e-06, 'learning_rate/action_model': 5.6909503108997456e-05, 'epoch': 4.19}) +[2026-06-20 05:10:12,725][starVLA.training.train_starvla][INFO] - Step 2406, Loss: {'train/loss': 0.1907852739095688, 'train/grad_norm_pre_clip': 21.16190528869629, 'timing/data': 0.006842439994215965, 'timing/model': 2.780054187402129, 'learning_rate/qwen_vl_interface': 5.687917949707913e-06, 'learning_rate/action_model': 5.687917949707913e-05, 'epoch': 4.19}) +[2026-06-20 05:10:23,631][starVLA.training.train_starvla][INFO] - Step 2407, Loss: {'train/loss': 0.026973938569426537, 'train/grad_norm_pre_clip': 6.208418369293213, 'timing/data': 0.006995311006903648, 'timing/model': 2.777704542502761, 'learning_rate/qwen_vl_interface': 5.684885408504555e-06, 'learning_rate/action_model': 5.684885408504554e-05, 'epoch': 4.19}) +[2026-06-20 05:10:34,531][starVLA.training.train_starvla][INFO] - Step 2408, Loss: {'train/loss': 0.04199861362576485, 'train/grad_norm_pre_clip': 12.713780403137207, 'timing/data': 0.0068937186151742935, 'timing/model': 2.78061287291348, 'learning_rate/qwen_vl_interface': 5.681852688536231e-06, 'learning_rate/action_model': 5.681852688536231e-05, 'epoch': 4.19}) +[2026-06-20 05:10:45,429][starVLA.training.train_starvla][INFO] - Step 2409, Loss: {'train/loss': 0.24723416566848755, 'train/grad_norm_pre_clip': 17.66143035888672, 'timing/data': 0.006704730913043022, 'timing/model': 2.7780907694250345, 'learning_rate/qwen_vl_interface': 5.67881979104958e-06, 'learning_rate/action_model': 5.67881979104958e-05, 'epoch': 4.2}) +[2026-06-20 05:10:56,542][starVLA.training.train_starvla][INFO] - Step 2410, Loss: {'train/loss': 0.12200355529785156, 'train/grad_norm_pre_clip': 15.888178825378418, 'timing/data': 0.006807491183280945, 'timing/model': 2.986876228824258, 'learning_rate/qwen_vl_interface': 5.67578671729131e-06, 'learning_rate/action_model': 5.675786717291309e-05, 'epoch': 4.2}) +[2026-06-20 05:11:07,429][starVLA.training.train_starvla][INFO] - Step 2411, Loss: {'train/loss': 0.084938645362854, 'train/grad_norm_pre_clip': 23.27936363220215, 'timing/data': 0.006857149302959442, 'timing/model': 2.777425328269601, 'learning_rate/qwen_vl_interface': 5.672753468508204e-06, 'learning_rate/action_model': 5.672753468508204e-05, 'epoch': 4.2}) +[2026-06-20 05:11:18,331][starVLA.training.train_starvla][INFO] - Step 2412, Loss: {'train/loss': 0.3409789800643921, 'train/grad_norm_pre_clip': 47.58074951171875, 'timing/data': 0.006886400282382965, 'timing/model': 2.774601710960269, 'learning_rate/qwen_vl_interface': 5.6697200459471165e-06, 'learning_rate/action_model': 5.669720045947117e-05, 'epoch': 4.2}) +[2026-06-20 05:11:29,218][starVLA.training.train_starvla][INFO] - Step 2413, Loss: {'train/loss': 0.04332756623625755, 'train/grad_norm_pre_clip': 13.737772941589355, 'timing/data': 0.006885571405291557, 'timing/model': 2.764965884387493, 'learning_rate/qwen_vl_interface': 5.666686450854973e-06, 'learning_rate/action_model': 5.666686450854973e-05, 'epoch': 4.2}) +[2026-06-20 05:11:40,231][starVLA.training.train_starvla][INFO] - Step 2414, Loss: {'train/loss': 0.3073243200778961, 'train/grad_norm_pre_clip': 42.47039031982422, 'timing/data': 0.007043387740850449, 'timing/model': 2.7554248068481684, 'learning_rate/qwen_vl_interface': 5.663652684478768e-06, 'learning_rate/action_model': 5.6636526844787676e-05, 'epoch': 4.21}) +[2026-06-20 05:11:51,129][starVLA.training.train_starvla][INFO] - Step 2415, Loss: {'train/loss': 0.2818077802658081, 'train/grad_norm_pre_clip': 39.54948806762695, 'timing/data': 0.006883159279823303, 'timing/model': 2.772600008174777, 'learning_rate/qwen_vl_interface': 5.660618748065569e-06, 'learning_rate/action_model': 5.6606187480655695e-05, 'epoch': 4.21}) +[2026-06-20 05:12:02,036][starVLA.training.train_starvla][INFO] - Step 2416, Loss: {'train/loss': 0.05310801789164543, 'train/grad_norm_pre_clip': 12.515059471130371, 'timing/data': 0.006882810965180397, 'timing/model': 2.7793789077550173, 'learning_rate/qwen_vl_interface': 5.657584642862514e-06, 'learning_rate/action_model': 5.657584642862513e-05, 'epoch': 4.21}) +[2026-06-20 05:12:12,931][starVLA.training.train_starvla][INFO] - Step 2417, Loss: {'train/loss': 0.06872217357158661, 'train/grad_norm_pre_clip': 14.319795608520508, 'timing/data': 0.006815899163484573, 'timing/model': 2.776407513767481, 'learning_rate/qwen_vl_interface': 5.654550370116808e-06, 'learning_rate/action_model': 5.654550370116808e-05, 'epoch': 4.21}) +[2026-06-20 05:12:23,824][starVLA.training.train_starvla][INFO] - Step 2418, Loss: {'train/loss': 0.06148013845086098, 'train/grad_norm_pre_clip': 12.262544631958008, 'timing/data': 0.006851350888609886, 'timing/model': 2.7673617899417877, 'learning_rate/qwen_vl_interface': 5.651515931075726e-06, 'learning_rate/action_model': 5.651515931075726e-05, 'epoch': 4.21}) +[2026-06-20 05:12:34,739][starVLA.training.train_starvla][INFO] - Step 2419, Loss: {'train/loss': 0.023322194814682007, 'train/grad_norm_pre_clip': 5.228662967681885, 'timing/data': 0.007705790922045708, 'timing/model': 2.7809895165264606, 'learning_rate/qwen_vl_interface': 5.6484813269866126e-06, 'learning_rate/action_model': 5.648481326986612e-05, 'epoch': 4.21}) +[2026-06-20 05:12:45,633][starVLA.training.train_starvla][INFO] - Step 2420, Loss: {'train/loss': 0.07095029950141907, 'train/grad_norm_pre_clip': 12.974847793579102, 'timing/data': 0.006652692332863808, 'timing/model': 2.7869507018476725, 'learning_rate/qwen_vl_interface': 5.645446559096875e-06, 'learning_rate/action_model': 5.645446559096875e-05, 'epoch': 4.22}) +[2026-06-20 05:12:56,530][starVLA.training.train_starvla][INFO] - Step 2421, Loss: {'train/loss': 0.22560769319534302, 'train/grad_norm_pre_clip': 20.815217971801758, 'timing/data': 0.006701691076159477, 'timing/model': 2.774256968870759, 'learning_rate/qwen_vl_interface': 5.6424116286539965e-06, 'learning_rate/action_model': 5.642411628653996e-05, 'epoch': 4.22}) +[2026-06-20 05:13:07,525][starVLA.training.train_starvla][INFO] - Step 2422, Loss: {'train/loss': 0.19728614389896393, 'train/grad_norm_pre_clip': 18.677385330200195, 'timing/data': 0.006843419745564461, 'timing/model': 2.7482282295823097, 'learning_rate/qwen_vl_interface': 5.639376536905517e-06, 'learning_rate/action_model': 5.6393765369055175e-05, 'epoch': 4.22}) +[2026-06-20 05:13:18,317][starVLA.training.train_starvla][INFO] - Step 2423, Loss: {'train/loss': 0.20785465836524963, 'train/grad_norm_pre_clip': 15.379855155944824, 'timing/data': 0.0068241506814956665, 'timing/model': 2.766258131712675, 'learning_rate/qwen_vl_interface': 5.636341285099055e-06, 'learning_rate/action_model': 5.636341285099055e-05, 'epoch': 4.22}) +[2026-06-20 05:13:29,135][starVLA.training.train_starvla][INFO] - Step 2424, Loss: {'train/loss': 0.08086718618869781, 'train/grad_norm_pre_clip': 19.18305015563965, 'timing/data': 0.006866341456770897, 'timing/model': 2.783875120803714, 'learning_rate/qwen_vl_interface': 5.6333058744822835e-06, 'learning_rate/action_model': 5.633305874482283e-05, 'epoch': 4.22}) +[2026-06-20 05:13:40,036][starVLA.training.train_starvla][INFO] - Step 2425, Loss: {'train/loss': 0.07764676213264465, 'train/grad_norm_pre_clip': 21.840072631835938, 'timing/data': 0.007087908685207367, 'timing/model': 2.774914937093854, 'learning_rate/qwen_vl_interface': 5.630270306302944e-06, 'learning_rate/action_model': 5.6302703063029436e-05, 'epoch': 4.22}) +[2026-06-20 05:13:50,935][starVLA.training.train_starvla][INFO] - Step 2426, Loss: {'train/loss': 0.1096249371767044, 'train/grad_norm_pre_clip': 18.411901473999023, 'timing/data': 0.006806200370192528, 'timing/model': 2.7745981980115175, 'learning_rate/qwen_vl_interface': 5.627234581808848e-06, 'learning_rate/action_model': 5.627234581808848e-05, 'epoch': 4.23}) +[2026-06-20 05:14:01,819][starVLA.training.train_starvla][INFO] - Step 2427, Loss: {'train/loss': 0.01739632710814476, 'train/grad_norm_pre_clip': 6.392757892608643, 'timing/data': 0.007041437551379204, 'timing/model': 2.7473208010196686, 'learning_rate/qwen_vl_interface': 5.624198702247863e-06, 'learning_rate/action_model': 5.624198702247863e-05, 'epoch': 4.23}) +[2026-06-20 05:14:12,625][starVLA.training.train_starvla][INFO] - Step 2428, Loss: {'train/loss': 0.1427307277917862, 'train/grad_norm_pre_clip': 12.1959228515625, 'timing/data': 0.007084926590323448, 'timing/model': 2.766566025093198, 'learning_rate/qwen_vl_interface': 5.621162668867925e-06, 'learning_rate/action_model': 5.621162668867925e-05, 'epoch': 4.23}) +[2026-06-20 05:14:23,752][starVLA.training.train_starvla][INFO] - Step 2429, Loss: {'train/loss': 0.31220006942749023, 'train/grad_norm_pre_clip': 26.511306762695312, 'timing/data': 0.0067124925553798676, 'timing/model': 2.995985809713602, 'learning_rate/qwen_vl_interface': 5.618126482917037e-06, 'learning_rate/action_model': 5.618126482917036e-05, 'epoch': 4.23}) +[2026-06-20 05:14:34,628][starVLA.training.train_starvla][INFO] - Step 2430, Loss: {'train/loss': 0.10007363557815552, 'train/grad_norm_pre_clip': 20.306041717529297, 'timing/data': 0.007006298750638962, 'timing/model': 2.7704470455646515, 'learning_rate/qwen_vl_interface': 5.6150901456432546e-06, 'learning_rate/action_model': 5.615090145643255e-05, 'epoch': 4.23}) +[2026-06-20 05:14:45,536][starVLA.training.train_starvla][INFO] - Step 2431, Loss: {'train/loss': 0.2965543568134308, 'train/grad_norm_pre_clip': 35.81508255004883, 'timing/data': 0.006934536620974541, 'timing/model': 2.7763632629066706, 'learning_rate/qwen_vl_interface': 5.612053658294705e-06, 'learning_rate/action_model': 5.6120536582947045e-05, 'epoch': 4.24}) +[2026-06-20 05:14:56,437][starVLA.training.train_starvla][INFO] - Step 2432, Loss: {'train/loss': 0.1522424817085266, 'train/grad_norm_pre_clip': 25.83692169189453, 'timing/data': 0.007047936320304871, 'timing/model': 2.7743040807545185, 'learning_rate/qwen_vl_interface': 5.609017022119572e-06, 'learning_rate/action_model': 5.609017022119573e-05, 'epoch': 4.24}) +[2026-06-20 05:15:07,533][starVLA.training.train_starvla][INFO] - Step 2433, Loss: {'train/loss': 0.0995500385761261, 'train/grad_norm_pre_clip': 9.753339767456055, 'timing/data': 0.006661813706159592, 'timing/model': 2.746204163879156, 'learning_rate/qwen_vl_interface': 5.605980238366104e-06, 'learning_rate/action_model': 5.605980238366103e-05, 'epoch': 4.24}) +[2026-06-20 05:15:18,536][starVLA.training.train_starvla][INFO] - Step 2434, Loss: {'train/loss': 0.10638932138681412, 'train/grad_norm_pre_clip': 10.228731155395508, 'timing/data': 0.009150449186563492, 'timing/model': 2.76649909093976, 'learning_rate/qwen_vl_interface': 5.602943308282606e-06, 'learning_rate/action_model': 5.602943308282605e-05, 'epoch': 4.24}) +[2026-06-20 05:15:29,427][starVLA.training.train_starvla][INFO] - Step 2435, Loss: {'train/loss': 0.02076314389705658, 'train/grad_norm_pre_clip': 3.472949743270874, 'timing/data': 0.00667693093419075, 'timing/model': 2.76253517344594, 'learning_rate/qwen_vl_interface': 5.599906233117445e-06, 'learning_rate/action_model': 5.5999062331174446e-05, 'epoch': 4.24}) +[2026-06-20 05:15:40,335][starVLA.training.train_starvla][INFO] - Step 2436, Loss: {'train/loss': 0.034191060811281204, 'train/grad_norm_pre_clip': 7.6931891441345215, 'timing/data': 0.006826469674706459, 'timing/model': 2.778384068980813, 'learning_rate/qwen_vl_interface': 5.596869014119048e-06, 'learning_rate/action_model': 5.596869014119048e-05, 'epoch': 4.24}) +[2026-06-20 05:15:51,236][starVLA.training.train_starvla][INFO] - Step 2437, Loss: {'train/loss': 0.010826239362359047, 'train/grad_norm_pre_clip': 1.744454264640808, 'timing/data': 0.006593622267246246, 'timing/model': 2.78080734051764, 'learning_rate/qwen_vl_interface': 5.593831652535902e-06, 'learning_rate/action_model': 5.593831652535902e-05, 'epoch': 4.25}) +[2026-06-20 05:16:02,137][starVLA.training.train_starvla][INFO] - Step 2438, Loss: {'train/loss': 0.1372976452112198, 'train/grad_norm_pre_clip': 18.17857551574707, 'timing/data': 0.006816931068897247, 'timing/model': 2.772815266624093, 'learning_rate/qwen_vl_interface': 5.590794149616552e-06, 'learning_rate/action_model': 5.590794149616552e-05, 'epoch': 4.25}) +[2026-06-20 05:16:13,029][starVLA.training.train_starvla][INFO] - Step 2439, Loss: {'train/loss': 0.07363414019346237, 'train/grad_norm_pre_clip': 21.188283920288086, 'timing/data': 0.007164627313613892, 'timing/model': 2.767864027991891, 'learning_rate/qwen_vl_interface': 5.587756506609601e-06, 'learning_rate/action_model': 5.5877565066096006e-05, 'epoch': 4.25}) +[2026-06-20 05:16:23,938][starVLA.training.train_starvla][INFO] - Step 2440, Loss: {'train/loss': 0.12086605280637741, 'train/grad_norm_pre_clip': 7.726025581359863, 'timing/data': 0.006657762452960014, 'timing/model': 2.777110928669572, 'learning_rate/qwen_vl_interface': 5.584718724763707e-06, 'learning_rate/action_model': 5.584718724763707e-05, 'epoch': 4.25}) +[2026-06-20 05:16:35,036][starVLA.training.train_starvla][INFO] - Step 2441, Loss: {'train/loss': 0.10189221054315567, 'train/grad_norm_pre_clip': 13.443830490112305, 'timing/data': 0.0069170501083135605, 'timing/model': 2.779405266046524, 'learning_rate/qwen_vl_interface': 5.581680805327588e-06, 'learning_rate/action_model': 5.5816808053275885e-05, 'epoch': 4.25}) +[2026-06-20 05:16:45,938][starVLA.training.train_starvla][INFO] - Step 2442, Loss: {'train/loss': 0.0041649192571640015, 'train/grad_norm_pre_clip': 0.7079266309738159, 'timing/data': 0.006451914086937904, 'timing/model': 2.776994038373232, 'learning_rate/qwen_vl_interface': 5.578642749550019e-06, 'learning_rate/action_model': 5.578642749550019e-05, 'epoch': 4.25}) +[2026-06-20 05:16:56,835][starVLA.training.train_starvla][INFO] - Step 2443, Loss: {'train/loss': 0.4248368442058563, 'train/grad_norm_pre_clip': 35.48908233642578, 'timing/data': 0.006991470232605934, 'timing/model': 2.7763572204858065, 'learning_rate/qwen_vl_interface': 5.575604558679831e-06, 'learning_rate/action_model': 5.57560455867983e-05, 'epoch': 4.26}) +[2026-06-20 05:17:07,733][starVLA.training.train_starvla][INFO] - Step 2444, Loss: {'train/loss': 0.19758976995944977, 'train/grad_norm_pre_clip': 21.551191329956055, 'timing/data': 0.006770879030227661, 'timing/model': 2.776200482621789, 'learning_rate/qwen_vl_interface': 5.5725662339659085e-06, 'learning_rate/action_model': 5.572566233965908e-05, 'epoch': 4.26}) +[2026-06-20 05:17:18,637][starVLA.training.train_starvla][INFO] - Step 2445, Loss: {'train/loss': 0.18142865598201752, 'train/grad_norm_pre_clip': 14.049074172973633, 'timing/data': 0.007035480812191963, 'timing/model': 2.7763657551258802, 'learning_rate/qwen_vl_interface': 5.56952777665719e-06, 'learning_rate/action_model': 5.56952777665719e-05, 'epoch': 4.26}) +[2026-06-20 05:17:29,533][starVLA.training.train_starvla][INFO] - Step 2446, Loss: {'train/loss': 0.25039753317832947, 'train/grad_norm_pre_clip': 19.357107162475586, 'timing/data': 0.006685512140393257, 'timing/model': 2.7728689406067133, 'learning_rate/qwen_vl_interface': 5.566489188002671e-06, 'learning_rate/action_model': 5.56648918800267e-05, 'epoch': 4.26}) +[2026-06-20 05:17:40,419][starVLA.training.train_starvla][INFO] - Step 2447, Loss: {'train/loss': 0.5413644313812256, 'train/grad_norm_pre_clip': 53.183162689208984, 'timing/data': 0.0065683238208293915, 'timing/model': 2.7630624379962683, 'learning_rate/qwen_vl_interface': 5.563450469251403e-06, 'learning_rate/action_model': 5.563450469251402e-05, 'epoch': 4.26}) +[2026-06-20 05:17:51,457][starVLA.training.train_starvla][INFO] - Step 2448, Loss: {'train/loss': 0.0589732751250267, 'train/grad_norm_pre_clip': 7.621187210083008, 'timing/data': 0.006997261196374893, 'timing/model': 3.0048043467104435, 'learning_rate/qwen_vl_interface': 5.560411621652488e-06, 'learning_rate/action_model': 5.560411621652488e-05, 'epoch': 4.26}) +[2026-06-20 05:18:02,143][starVLA.training.train_starvla][INFO] - Step 2449, Loss: {'train/loss': 0.1418963074684143, 'train/grad_norm_pre_clip': 16.44276237487793, 'timing/data': 0.006954768672585487, 'timing/model': 2.722943300381303, 'learning_rate/qwen_vl_interface': 5.557372646455081e-06, 'learning_rate/action_model': 5.55737264645508e-05, 'epoch': 4.27}) +[2026-06-20 05:18:13,036][starVLA.training.train_starvla][INFO] - Step 2450, Loss: {'train/loss': 0.2349729984998703, 'train/grad_norm_pre_clip': 25.696468353271484, 'timing/data': 0.006507864221930504, 'timing/model': 2.776410998776555, 'learning_rate/qwen_vl_interface': 5.554333544908388e-06, 'learning_rate/action_model': 5.5543335449083875e-05, 'epoch': 4.27}) +[2026-06-20 05:18:23,916][starVLA.training.train_starvla][INFO] - Step 2451, Loss: {'train/loss': 0.11699175834655762, 'train/grad_norm_pre_clip': 13.332406997680664, 'timing/data': 0.006630100309848785, 'timing/model': 2.75414227694273, 'learning_rate/qwen_vl_interface': 5.551294318261671e-06, 'learning_rate/action_model': 5.5512943182616705e-05, 'epoch': 4.27}) +[2026-06-20 05:18:34,940][starVLA.training.train_starvla][INFO] - Step 2452, Loss: {'train/loss': 0.11142449080944061, 'train/grad_norm_pre_clip': 10.169522285461426, 'timing/data': 0.006620222702622414, 'timing/model': 2.7582469172775745, 'learning_rate/qwen_vl_interface': 5.548254967764239e-06, 'learning_rate/action_model': 5.5482549677642394e-05, 'epoch': 4.27}) +[2026-06-20 05:18:45,835][starVLA.training.train_starvla][INFO] - Step 2453, Loss: {'train/loss': 0.12232078611850739, 'train/grad_norm_pre_clip': 15.233670234680176, 'timing/data': 0.006561903282999992, 'timing/model': 2.7845140118151903, 'learning_rate/qwen_vl_interface': 5.545215494665458e-06, 'learning_rate/action_model': 5.545215494665458e-05, 'epoch': 4.27}) +[2026-06-20 05:18:56,737][starVLA.training.train_starvla][INFO] - Step 2454, Loss: {'train/loss': 0.027756180614233017, 'train/grad_norm_pre_clip': 5.806852340698242, 'timing/data': 0.006470015272498131, 'timing/model': 2.772237965837121, 'learning_rate/qwen_vl_interface': 5.5421759002147376e-06, 'learning_rate/action_model': 5.542175900214738e-05, 'epoch': 4.28}) +[2026-06-20 05:19:07,628][starVLA.training.train_starvla][INFO] - Step 2455, Loss: {'train/loss': 0.3905738294124603, 'train/grad_norm_pre_clip': 28.87627410888672, 'timing/data': 0.006475264206528664, 'timing/model': 2.7805113717913628, 'learning_rate/qwen_vl_interface': 5.53913618566154e-06, 'learning_rate/action_model': 5.53913618566154e-05, 'epoch': 4.28}) +[2026-06-20 05:19:18,374][starVLA.training.train_starvla][INFO] - Step 2456, Loss: {'train/loss': 0.24643349647521973, 'train/grad_norm_pre_clip': 21.43980598449707, 'timing/data': 0.0067735705524683, 'timing/model': 2.670743051916361, 'learning_rate/qwen_vl_interface': 5.536096352255381e-06, 'learning_rate/action_model': 5.536096352255381e-05, 'epoch': 4.28}) +[2026-06-20 05:19:29,131][starVLA.training.train_starvla][INFO] - Step 2457, Loss: {'train/loss': 0.1481950879096985, 'train/grad_norm_pre_clip': 23.053266525268555, 'timing/data': 0.006609752774238586, 'timing/model': 2.7828438505530357, 'learning_rate/qwen_vl_interface': 5.533056401245817e-06, 'learning_rate/action_model': 5.5330564012458164e-05, 'epoch': 4.28}) +[2026-06-20 05:19:40,027][starVLA.training.train_starvla][INFO] - Step 2458, Loss: {'train/loss': 0.210909903049469, 'train/grad_norm_pre_clip': 25.378822326660156, 'timing/data': 0.006558453664183617, 'timing/model': 2.776512401178479, 'learning_rate/qwen_vl_interface': 5.530016333882456e-06, 'learning_rate/action_model': 5.5300163338824565e-05, 'epoch': 4.28}) +[2026-06-20 05:19:50,917][starVLA.training.train_starvla][INFO] - Step 2459, Loss: {'train/loss': 0.23853714764118195, 'train/grad_norm_pre_clip': 20.738828659057617, 'timing/data': 0.006866959854960442, 'timing/model': 2.75980562902987, 'learning_rate/qwen_vl_interface': 5.52697615141496e-06, 'learning_rate/action_model': 5.52697615141496e-05, 'epoch': 4.28}) +[2026-06-20 05:20:01,936][starVLA.training.train_starvla][INFO] - Step 2460, Loss: {'train/loss': 0.07878294587135315, 'train/grad_norm_pre_clip': 10.41518783569336, 'timing/data': 0.00690176896750927, 'timing/model': 2.783516589552164, 'learning_rate/qwen_vl_interface': 5.52393585509303e-06, 'learning_rate/action_model': 5.52393585509303e-05, 'epoch': 4.29}) +[2026-06-20 05:20:12,838][starVLA.training.train_starvla][INFO] - Step 2461, Loss: {'train/loss': 0.033014848828315735, 'train/grad_norm_pre_clip': 2.7042617797851562, 'timing/data': 0.006661210209131241, 'timing/model': 2.7747976444661617, 'learning_rate/qwen_vl_interface': 5.520895446166418e-06, 'learning_rate/action_model': 5.5208954461664175e-05, 'epoch': 4.29}) +[2026-06-20 05:20:23,737][starVLA.training.train_starvla][INFO] - Step 2462, Loss: {'train/loss': 0.0621316060423851, 'train/grad_norm_pre_clip': 8.810948371887207, 'timing/data': 0.0068375468254089355, 'timing/model': 2.7789836935698986, 'learning_rate/qwen_vl_interface': 5.517854925884919e-06, 'learning_rate/action_model': 5.517854925884919e-05, 'epoch': 4.29}) +[2026-06-20 05:20:34,636][starVLA.training.train_starvla][INFO] - Step 2463, Loss: {'train/loss': 0.03227084502577782, 'train/grad_norm_pre_clip': 5.0244574546813965, 'timing/data': 0.006764799356460571, 'timing/model': 2.7773457542061806, 'learning_rate/qwen_vl_interface': 5.514814295498379e-06, 'learning_rate/action_model': 5.514814295498378e-05, 'epoch': 4.29}) +[2026-06-20 05:20:45,526][starVLA.training.train_starvla][INFO] - Step 2464, Loss: {'train/loss': 0.11663509905338287, 'train/grad_norm_pre_clip': 14.526805877685547, 'timing/data': 0.007074229419231415, 'timing/model': 2.7662472631782293, 'learning_rate/qwen_vl_interface': 5.5117735562566846e-06, 'learning_rate/action_model': 5.5117735562566844e-05, 'epoch': 4.29}) +[2026-06-20 05:20:56,436][starVLA.training.train_starvla][INFO] - Step 2465, Loss: {'train/loss': 0.11147372424602509, 'train/grad_norm_pre_clip': 12.994699478149414, 'timing/data': 0.00698784738779068, 'timing/model': 2.7847845181822777, 'learning_rate/qwen_vl_interface': 5.5087327094097685e-06, 'learning_rate/action_model': 5.508732709409769e-05, 'epoch': 4.29}) +[2026-06-20 05:21:07,255][starVLA.training.train_starvla][INFO] - Step 2466, Loss: {'train/loss': 0.14774517714977264, 'train/grad_norm_pre_clip': 15.707748413085938, 'timing/data': 0.006760682910680771, 'timing/model': 2.697114944458008, 'learning_rate/qwen_vl_interface': 5.505691756207609e-06, 'learning_rate/action_model': 5.505691756207609e-05, 'epoch': 4.3}) +[2026-06-20 05:21:18,156][starVLA.training.train_starvla][INFO] - Step 2467, Loss: {'train/loss': 0.2160029411315918, 'train/grad_norm_pre_clip': 15.673542976379395, 'timing/data': 0.006845071911811829, 'timing/model': 2.9874188397079706, 'learning_rate/qwen_vl_interface': 5.502650697900228e-06, 'learning_rate/action_model': 5.502650697900228e-05, 'epoch': 4.3}) +[2026-06-20 05:21:29,038][starVLA.training.train_starvla][INFO] - Step 2468, Loss: {'train/loss': 0.15934889018535614, 'train/grad_norm_pre_clip': 17.78899383544922, 'timing/data': 0.006579732522368431, 'timing/model': 2.7782823387533426, 'learning_rate/qwen_vl_interface': 5.499609535737688e-06, 'learning_rate/action_model': 5.4996095357376875e-05, 'epoch': 4.3}) +[2026-06-20 05:21:39,937][starVLA.training.train_starvla][INFO] - Step 2469, Loss: {'train/loss': 0.11072225868701935, 'train/grad_norm_pre_clip': 9.47830867767334, 'timing/data': 0.006835548207163811, 'timing/model': 2.7785909473896027, 'learning_rate/qwen_vl_interface': 5.496568270970096e-06, 'learning_rate/action_model': 5.496568270970096e-05, 'epoch': 4.3}) +[2026-06-20 05:21:50,837][starVLA.training.train_starvla][INFO] - Step 2470, Loss: {'train/loss': 0.13796091079711914, 'train/grad_norm_pre_clip': 14.394742965698242, 'timing/data': 0.0064044538885355, 'timing/model': 2.77526817843318, 'learning_rate/qwen_vl_interface': 5.4935269048476015e-06, 'learning_rate/action_model': 5.493526904847601e-05, 'epoch': 4.3}) +[2026-06-20 05:22:01,936][starVLA.training.train_starvla][INFO] - Step 2471, Loss: {'train/loss': 0.41644346714019775, 'train/grad_norm_pre_clip': 20.64332389831543, 'timing/data': 0.006936907768249512, 'timing/model': 2.7466315925121307, 'learning_rate/qwen_vl_interface': 5.490485438620397e-06, 'learning_rate/action_model': 5.4904854386203965e-05, 'epoch': 4.3}) +[2026-06-20 05:22:12,838][starVLA.training.train_starvla][INFO] - Step 2472, Loss: {'train/loss': 0.44670477509498596, 'train/grad_norm_pre_clip': 18.979232788085938, 'timing/data': 0.007064979523420334, 'timing/model': 2.7767934687435627, 'learning_rate/qwen_vl_interface': 5.487443873538712e-06, 'learning_rate/action_model': 5.4874438735387115e-05, 'epoch': 4.31}) +[2026-06-20 05:22:23,738][starVLA.training.train_starvla][INFO] - Step 2473, Loss: {'train/loss': 0.06854412704706192, 'train/grad_norm_pre_clip': 6.356279373168945, 'timing/data': 0.006640158593654633, 'timing/model': 2.7789758294820786, 'learning_rate/qwen_vl_interface': 5.484402210852819e-06, 'learning_rate/action_model': 5.484402210852818e-05, 'epoch': 4.31}) +[2026-06-20 05:22:34,629][starVLA.training.train_starvla][INFO] - Step 2474, Loss: {'train/loss': 0.07665464282035828, 'train/grad_norm_pre_clip': 11.675085067749023, 'timing/data': 0.007106197997927666, 'timing/model': 2.7659571189433336, 'learning_rate/qwen_vl_interface': 5.48136045181303e-06, 'learning_rate/action_model': 5.48136045181303e-05, 'epoch': 4.31}) +[2026-06-20 05:22:45,539][starVLA.training.train_starvla][INFO] - Step 2475, Loss: {'train/loss': 0.17470858991146088, 'train/grad_norm_pre_clip': 12.734417915344238, 'timing/data': 0.006934475153684616, 'timing/model': 2.7805818300694227, 'learning_rate/qwen_vl_interface': 5.478318597669699e-06, 'learning_rate/action_model': 5.4783185976696995e-05, 'epoch': 4.31}) +[2026-06-20 05:22:56,443][starVLA.training.train_starvla][INFO] - Step 2476, Loss: {'train/loss': 0.05816904082894325, 'train/grad_norm_pre_clip': 4.071351528167725, 'timing/data': 0.00679740309715271, 'timing/model': 2.7845322247594595, 'learning_rate/qwen_vl_interface': 5.475276649673218e-06, 'learning_rate/action_model': 5.4752766496732174e-05, 'epoch': 4.31}) +[2026-06-20 05:23:07,336][starVLA.training.train_starvla][INFO] - Step 2477, Loss: {'train/loss': 0.059113532304763794, 'train/grad_norm_pre_clip': 6.967360973358154, 'timing/data': 0.006974227726459503, 'timing/model': 2.772651769220829, 'learning_rate/qwen_vl_interface': 5.472234609074014e-06, 'learning_rate/action_model': 5.472234609074014e-05, 'epoch': 4.32}) +[2026-06-20 05:23:18,239][starVLA.training.train_starvla][INFO] - Step 2478, Loss: {'train/loss': 0.21571342647075653, 'train/grad_norm_pre_clip': 14.654877662658691, 'timing/data': 0.006886297836899757, 'timing/model': 2.7804248724132776, 'learning_rate/qwen_vl_interface': 5.46919247712256e-06, 'learning_rate/action_model': 5.46919247712256e-05, 'epoch': 4.32}) +[2026-06-20 05:23:29,331][starVLA.training.train_starvla][INFO] - Step 2479, Loss: {'train/loss': 0.06578674912452698, 'train/grad_norm_pre_clip': 8.160483360290527, 'timing/data': 0.006727330386638641, 'timing/model': 2.767209691926837, 'learning_rate/qwen_vl_interface': 5.466150255069356e-06, 'learning_rate/action_model': 5.4661502550693555e-05, 'epoch': 4.32}) +[2026-06-20 05:23:40,235][starVLA.training.train_starvla][INFO] - Step 2480, Loss: {'train/loss': 0.3698047399520874, 'train/grad_norm_pre_clip': 14.808704376220703, 'timing/data': 0.006838247179985046, 'timing/model': 2.775766694918275, 'learning_rate/qwen_vl_interface': 5.463107944164947e-06, 'learning_rate/action_model': 5.4631079441649466e-05, 'epoch': 4.32}) +[2026-06-20 05:23:51,140][starVLA.training.train_starvla][INFO] - Step 2481, Loss: {'train/loss': 0.0665295273065567, 'train/grad_norm_pre_clip': 8.232959747314453, 'timing/data': 0.0067427512258291245, 'timing/model': 2.7796167209744453, 'learning_rate/qwen_vl_interface': 5.46006554565991e-06, 'learning_rate/action_model': 5.4600655456599105e-05, 'epoch': 4.32}) +[2026-06-20 05:24:02,031][starVLA.training.train_starvla][INFO] - Step 2482, Loss: {'train/loss': 0.08232177793979645, 'train/grad_norm_pre_clip': 6.965965270996094, 'timing/data': 0.006862428039312363, 'timing/model': 2.769957162439823, 'learning_rate/qwen_vl_interface': 5.457023060804865e-06, 'learning_rate/action_model': 5.457023060804864e-05, 'epoch': 4.32}) +[2026-06-20 05:24:12,937][starVLA.training.train_starvla][INFO] - Step 2483, Loss: {'train/loss': 0.1186835914850235, 'train/grad_norm_pre_clip': 10.126099586486816, 'timing/data': 0.007768390700221062, 'timing/model': 2.774670507758856, 'learning_rate/qwen_vl_interface': 5.453980490850459e-06, 'learning_rate/action_model': 5.453980490850459e-05, 'epoch': 4.33}) +[2026-06-20 05:24:23,836][starVLA.training.train_starvla][INFO] - Step 2484, Loss: {'train/loss': 0.04756554588675499, 'train/grad_norm_pre_clip': 5.592818737030029, 'timing/data': 0.006556842476129532, 'timing/model': 2.7785701900720596, 'learning_rate/qwen_vl_interface': 5.4509378370473785e-06, 'learning_rate/action_model': 5.450937837047378e-05, 'epoch': 4.33}) +[2026-06-20 05:24:34,739][starVLA.training.train_starvla][INFO] - Step 2485, Loss: {'train/loss': 0.03534625470638275, 'train/grad_norm_pre_clip': 2.6030445098876953, 'timing/data': 0.006876219063997269, 'timing/model': 2.7769238594919443, 'learning_rate/qwen_vl_interface': 5.4478951006463424e-06, 'learning_rate/action_model': 5.447895100646342e-05, 'epoch': 4.33}) +[2026-06-20 05:24:45,885][starVLA.training.train_starvla][INFO] - Step 2486, Loss: {'train/loss': 0.11797765642404556, 'train/grad_norm_pre_clip': 9.503827095031738, 'timing/data': 0.007726710289716721, 'timing/model': 3.026995414867997, 'learning_rate/qwen_vl_interface': 5.444852282898106e-06, 'learning_rate/action_model': 5.444852282898106e-05, 'epoch': 4.33}) +[2026-06-20 05:24:56,736][starVLA.training.train_starvla][INFO] - Step 2487, Loss: {'train/loss': 0.040509019047021866, 'train/grad_norm_pre_clip': 3.6468687057495117, 'timing/data': 0.008550692349672318, 'timing/model': 2.7682747561484575, 'learning_rate/qwen_vl_interface': 5.441809385053456e-06, 'learning_rate/action_model': 5.4418093850534555e-05, 'epoch': 4.33}) +[2026-06-20 05:25:07,637][starVLA.training.train_starvla][INFO] - Step 2488, Loss: {'train/loss': 0.231672465801239, 'train/grad_norm_pre_clip': 30.56604766845703, 'timing/data': 0.006566954776644707, 'timing/model': 2.777998935431242, 'learning_rate/qwen_vl_interface': 5.438766408363214e-06, 'learning_rate/action_model': 5.438766408363214e-05, 'epoch': 4.33}) +[2026-06-20 05:25:18,544][starVLA.training.train_starvla][INFO] - Step 2489, Loss: {'train/loss': 0.2979671061038971, 'train/grad_norm_pre_clip': 13.024054527282715, 'timing/data': 0.006756547838449478, 'timing/model': 2.7811239156872034, 'learning_rate/qwen_vl_interface': 5.43572335407823e-06, 'learning_rate/action_model': 5.43572335407823e-05, 'epoch': 4.34}) +[2026-06-20 05:25:29,635][starVLA.training.train_starvla][INFO] - Step 2490, Loss: {'train/loss': 0.011096610687673092, 'train/grad_norm_pre_clip': 2.3361198902130127, 'timing/data': 0.006918907165527344, 'timing/model': 2.7420919686555862, 'learning_rate/qwen_vl_interface': 5.432680223449392e-06, 'learning_rate/action_model': 5.4326802234493924e-05, 'epoch': 4.34}) +[2026-06-20 05:25:40,538][starVLA.training.train_starvla][INFO] - Step 2491, Loss: {'train/loss': 0.08915618807077408, 'train/grad_norm_pre_clip': 9.48166275024414, 'timing/data': 0.0070070866495370865, 'timing/model': 2.7752685621380806, 'learning_rate/qwen_vl_interface': 5.429637017727614e-06, 'learning_rate/action_model': 5.4296370177276144e-05, 'epoch': 4.34}) +[2026-06-20 05:25:51,435][starVLA.training.train_starvla][INFO] - Step 2492, Loss: {'train/loss': 0.07106426358222961, 'train/grad_norm_pre_clip': 7.490193843841553, 'timing/data': 0.006933547556400299, 'timing/model': 2.775838477537036, 'learning_rate/qwen_vl_interface': 5.426593738163845e-06, 'learning_rate/action_model': 5.4265937381638446e-05, 'epoch': 4.34}) +[2026-06-20 05:26:02,338][starVLA.training.train_starvla][INFO] - Step 2493, Loss: {'train/loss': 0.048720333725214005, 'train/grad_norm_pre_clip': 9.264886856079102, 'timing/data': 0.00681900791823864, 'timing/model': 2.7820366490632296, 'learning_rate/qwen_vl_interface': 5.42355038600906e-06, 'learning_rate/action_model': 5.42355038600906e-05, 'epoch': 4.34}) +[2026-06-20 05:26:13,235][starVLA.training.train_starvla][INFO] - Step 2494, Loss: {'train/loss': 0.07209008932113647, 'train/grad_norm_pre_clip': 13.922618865966797, 'timing/data': 0.006844161078333855, 'timing/model': 2.7730112317949533, 'learning_rate/qwen_vl_interface': 5.420506962514267e-06, 'learning_rate/action_model': 5.420506962514267e-05, 'epoch': 4.34}) +[2026-06-20 05:26:24,140][starVLA.training.train_starvla][INFO] - Step 2495, Loss: {'train/loss': 0.11753726750612259, 'train/grad_norm_pre_clip': 36.31583786010742, 'timing/data': 0.0068406518548727036, 'timing/model': 2.777652896940708, 'learning_rate/qwen_vl_interface': 5.4174634689305015e-06, 'learning_rate/action_model': 5.417463468930501e-05, 'epoch': 4.35}) +[2026-06-20 05:26:35,038][starVLA.training.train_starvla][INFO] - Step 2496, Loss: {'train/loss': 0.1383914053440094, 'train/grad_norm_pre_clip': 10.58372688293457, 'timing/data': 0.006718531250953674, 'timing/model': 2.781420649960637, 'learning_rate/qwen_vl_interface': 5.41441990650883e-06, 'learning_rate/action_model': 5.4144199065088295e-05, 'epoch': 4.35}) +[2026-06-20 05:26:45,837][starVLA.training.train_starvla][INFO] - Step 2497, Loss: {'train/loss': 0.16033697128295898, 'train/grad_norm_pre_clip': 22.164478302001953, 'timing/data': 0.006910430267453194, 'timing/model': 2.7745818123221397, 'learning_rate/qwen_vl_interface': 5.411376276500345e-06, 'learning_rate/action_model': 5.411376276500345e-05, 'epoch': 4.35}) +[2026-06-20 05:26:56,938][starVLA.training.train_starvla][INFO] - Step 2498, Loss: {'train/loss': 0.14783960580825806, 'train/grad_norm_pre_clip': 18.783342361450195, 'timing/data': 0.0069909002631902695, 'timing/model': 2.789002787321806, 'learning_rate/qwen_vl_interface': 5.408332580156168e-06, 'learning_rate/action_model': 5.408332580156168e-05, 'epoch': 4.35}) +[2026-06-20 05:27:07,835][starVLA.training.train_starvla][INFO] - Step 2499, Loss: {'train/loss': 0.040859807282686234, 'train/grad_norm_pre_clip': 6.96409273147583, 'timing/data': 0.006779730319976807, 'timing/model': 2.775810645893216, 'learning_rate/qwen_vl_interface': 5.40528881872745e-06, 'learning_rate/action_model': 5.40528881872745e-05, 'epoch': 4.35}) +[2026-06-20 05:27:18,734][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 2500 +[2026-06-20 05:34:48,546][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 2500 +[2026-06-20 05:34:48,548][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 2500 +[2026-06-20 05:43:06,691][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 2500 +[2026-06-20 05:43:06,693][starVLA.training.train_starvla][INFO] - Step 2500, Loss: {'train/loss': 0.05944886803627014, 'train/grad_norm_pre_clip': 11.95466423034668, 'eval/loss': 0.7203582525253296, 'eval/latency_0/loss': 0.6037023067474365, 'eval/latency_1/loss': 0.7820838689804077, 'eval/latency_2/loss': 0.10153037309646606, 'eval/latency_3/loss': 0.8713173270225525, 'eval/latency_4/loss': 1.2821773290634155, 'eval/flappy/loss': 0.7203582525253296, 'eval/flappy/latency_0/loss': 0.6037023067474365, 'eval/flappy/latency_1/loss': 0.7820838689804077, 'eval/flappy/latency_2/loss': 0.10153037309646606, 'eval/flappy/latency_3/loss': 0.8713173270225525, 'eval/flappy/latency_4/loss': 1.2821773290634155, 'eval/flap_precision': 0.8588235294117647, 'eval/flap_recall': 0.26545454545454544, 'eval/flap_f1': 0.4055555555555555, 'eval/flap_rate_model': 0.02361111111111111, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.40555555555555545, 'eval/latency_0/flap_precision': 0.8588235294117647, 'eval/latency_0/flap_recall': 0.26545454545454544, 'eval/latency_0/flap_f1': 0.4055555555555555, 'eval/latency_0/flap_rate_model': 0.02361111111111111, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.40555555555555545, 'timing/data': 0.006723739206790924, 'timing/model': 2.79520177654922, 'learning_rate/qwen_vl_interface': 5.402244993465364e-06, 'learning_rate/action_model': 5.402244993465364e-05, 'epoch': 4.36}) +[2026-06-20 05:43:06,693][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_2500_state +[2026-06-20 05:43:06,723][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-20 05:43:12,622][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_2500_state/model.safetensors +[2026-06-20 05:43:27,111][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_2500_state/optimizer.bin +[2026-06-20 05:43:27,112][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_2500_state/sampler.bin +[2026-06-20 05:43:27,112][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_2500_state/sampler_1.bin +[2026-06-20 05:43:27,113][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_2500_state/random_states_0.pkl +[2026-06-20 05:43:27,247][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2500 +[2026-06-20 05:43:27,247][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-20 05:43:27,261][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-20 05:43:39,238][starVLA.training.train_starvla][INFO] - Latency curriculum phase 3 at step 2501: active_latencies=[3], dataset_size=70316 +[2026-06-20 05:43:39,240][starVLA.training.train_starvla][INFO] - Step 2501, Loss: {'train/loss': 0.3569756746292114, 'train/grad_norm_pre_clip': 18.017606735229492, 'latency_curriculum/phase': 3.0, 'latency_curriculum/active_count': 1.0, 'latency_curriculum/max_active_latency': 3.0, 'latency_curriculum/dataset_size': 70316.0, 'timing/data': 0.007049407809972763, 'timing/model': 2.7749308962374926, 'learning_rate/qwen_vl_interface': 5.399201105621111e-06, 'learning_rate/action_model': 5.39920110562111e-05, 'epoch': 4.55}) +[2026-06-20 05:44:11,126][starVLA.training.train_starvla][INFO] - Step 2502, Loss: {'train/loss': 0.5476489067077637, 'train/grad_norm_pre_clip': 18.902935028076172, 'timing/data': 0.012073235586285591, 'timing/model': 2.7242672257125378, 'learning_rate/qwen_vl_interface': 5.396157156445918e-06, 'learning_rate/action_model': 5.396157156445918e-05, 'epoch': 4.55}) +[2026-06-20 05:44:21,932][starVLA.training.train_starvla][INFO] - Step 2503, Loss: {'train/loss': 1.4703986644744873, 'train/grad_norm_pre_clip': 19.36479377746582, 'timing/data': 0.012619368731975555, 'timing/model': 2.778757568448782, 'learning_rate/qwen_vl_interface': 5.39311314719104e-06, 'learning_rate/action_model': 5.39311314719104e-05, 'epoch': 4.56}) +[2026-06-20 05:44:32,829][starVLA.training.train_starvla][INFO] - Step 2504, Loss: {'train/loss': 0.5368564128875732, 'train/grad_norm_pre_clip': 31.53093719482422, 'timing/data': 0.013223003596067429, 'timing/model': 2.756748666986823, 'learning_rate/qwen_vl_interface': 5.390069079107755e-06, 'learning_rate/action_model': 5.390069079107755e-05, 'epoch': 4.56}) +[2026-06-20 05:44:43,726][starVLA.training.train_starvla][INFO] - Step 2505, Loss: {'train/loss': 1.5434335470199585, 'train/grad_norm_pre_clip': 24.92537498474121, 'timing/data': 0.014615170657634735, 'timing/model': 2.763055568560958, 'learning_rate/qwen_vl_interface': 5.387024953447361e-06, 'learning_rate/action_model': 5.387024953447361e-05, 'epoch': 4.56}) +[2026-06-20 05:44:54,448][starVLA.training.train_starvla][INFO] - Step 2506, Loss: {'train/loss': 1.1371344327926636, 'train/grad_norm_pre_clip': 21.786754608154297, 'timing/data': 0.015964116901159286, 'timing/model': 2.694583121687174, 'learning_rate/qwen_vl_interface': 5.383980771461184e-06, 'learning_rate/action_model': 5.3839807714611846e-05, 'epoch': 4.56}) +[2026-06-20 05:45:05,222][starVLA.training.train_starvla][INFO] - Step 2507, Loss: {'train/loss': 0.9428720474243164, 'train/grad_norm_pre_clip': 14.56038761138916, 'timing/data': 0.014947216957807541, 'timing/model': 2.7617995645850897, 'learning_rate/qwen_vl_interface': 5.380936534400572e-06, 'learning_rate/action_model': 5.3809365344005726e-05, 'epoch': 4.56}) +[2026-06-20 05:45:16,052][starVLA.training.train_starvla][INFO] - Step 2508, Loss: {'train/loss': 1.9661449193954468, 'train/grad_norm_pre_clip': 35.64896011352539, 'timing/data': 0.010264154523611069, 'timing/model': 2.7884453032165766, 'learning_rate/qwen_vl_interface': 5.377892243516899e-06, 'learning_rate/action_model': 5.3778922435168985e-05, 'epoch': 4.57}) +[2026-06-20 05:45:26,934][starVLA.training.train_starvla][INFO] - Step 2509, Loss: {'train/loss': 0.8461973667144775, 'train/grad_norm_pre_clip': 18.546527862548828, 'timing/data': 0.006726538762450218, 'timing/model': 2.7734159231185913, 'learning_rate/qwen_vl_interface': 5.3748479000615525e-06, 'learning_rate/action_model': 5.374847900061553e-05, 'epoch': 4.57}) +[2026-06-20 05:45:38,032][starVLA.training.train_starvla][INFO] - Step 2510, Loss: {'train/loss': 0.8335842490196228, 'train/grad_norm_pre_clip': 15.993629455566406, 'timing/data': 0.007845539599657059, 'timing/model': 2.772508416324854, 'learning_rate/qwen_vl_interface': 5.371803505285954e-06, 'learning_rate/action_model': 5.371803505285954e-05, 'epoch': 4.57}) +[2026-06-20 05:45:48,926][starVLA.training.train_starvla][INFO] - Step 2511, Loss: {'train/loss': 0.7951415777206421, 'train/grad_norm_pre_clip': 23.778614044189453, 'timing/data': 0.0068031903356313705, 'timing/model': 2.7696796245872974, 'learning_rate/qwen_vl_interface': 5.368759060441534e-06, 'learning_rate/action_model': 5.368759060441534e-05, 'epoch': 4.57}) +[2026-06-20 05:45:59,824][starVLA.training.train_starvla][INFO] - Step 2512, Loss: {'train/loss': 1.2760311365127563, 'train/grad_norm_pre_clip': 16.818979263305664, 'timing/data': 0.010190287604928017, 'timing/model': 2.76345637999475, 'learning_rate/qwen_vl_interface': 5.365714566779749e-06, 'learning_rate/action_model': 5.3657145667797484e-05, 'epoch': 4.57}) +[2026-06-20 05:46:10,731][starVLA.training.train_starvla][INFO] - Step 2513, Loss: {'train/loss': 1.1265498399734497, 'train/grad_norm_pre_clip': 18.325023651123047, 'timing/data': 0.0065522827208042145, 'timing/model': 2.7744860257953405, 'learning_rate/qwen_vl_interface': 5.362670025552077e-06, 'learning_rate/action_model': 5.362670025552077e-05, 'epoch': 4.57}) +[2026-06-20 05:46:21,634][starVLA.training.train_starvla][INFO] - Step 2514, Loss: {'train/loss': 0.7217670679092407, 'train/grad_norm_pre_clip': 12.153273582458496, 'timing/data': 0.006600171327590942, 'timing/model': 2.7756409142166376, 'learning_rate/qwen_vl_interface': 5.359625438010013e-06, 'learning_rate/action_model': 5.359625438010013e-05, 'epoch': 4.58}) +[2026-06-20 05:46:32,539][starVLA.training.train_starvla][INFO] - Step 2515, Loss: {'train/loss': 0.519062340259552, 'train/grad_norm_pre_clip': 13.934962272644043, 'timing/data': 0.006697973236441612, 'timing/model': 2.7811323571950197, 'learning_rate/qwen_vl_interface': 5.356580805405075e-06, 'learning_rate/action_model': 5.356580805405075e-05, 'epoch': 4.58}) +[2026-06-20 05:46:43,434][starVLA.training.train_starvla][INFO] - Step 2516, Loss: {'train/loss': 0.6717763543128967, 'train/grad_norm_pre_clip': 25.716751098632812, 'timing/data': 0.006631731986999512, 'timing/model': 2.7713139187544584, 'learning_rate/qwen_vl_interface': 5.353536128988791e-06, 'learning_rate/action_model': 5.353536128988791e-05, 'epoch': 4.58}) +[2026-06-20 05:46:54,547][starVLA.training.train_starvla][INFO] - Step 2517, Loss: {'train/loss': 0.7970531582832336, 'train/grad_norm_pre_clip': 16.603683471679688, 'timing/data': 0.006466412916779518, 'timing/model': 2.984185680747032, 'learning_rate/qwen_vl_interface': 5.350491410012717e-06, 'learning_rate/action_model': 5.350491410012717e-05, 'epoch': 4.58}) +[2026-06-20 05:47:05,439][starVLA.training.train_starvla][INFO] - Step 2518, Loss: {'train/loss': 0.5307009220123291, 'train/grad_norm_pre_clip': 13.529768943786621, 'timing/data': 0.006681552156805992, 'timing/model': 2.7804886735975742, 'learning_rate/qwen_vl_interface': 5.347446649728419e-06, 'learning_rate/action_model': 5.347446649728418e-05, 'epoch': 4.58}) +[2026-06-20 05:47:16,339][starVLA.training.train_starvla][INFO] - Step 2519, Loss: {'train/loss': 0.3368700444698334, 'train/grad_norm_pre_clip': 7.582131862640381, 'timing/data': 0.006606759503483772, 'timing/model': 2.7818354927003384, 'learning_rate/qwen_vl_interface': 5.344401849387484e-06, 'learning_rate/action_model': 5.344401849387484e-05, 'epoch': 4.59}) +[2026-06-20 05:47:27,223][starVLA.training.train_starvla][INFO] - Step 2520, Loss: {'train/loss': 0.3894509971141815, 'train/grad_norm_pre_clip': 8.469015121459961, 'timing/data': 0.006777109578251839, 'timing/model': 2.768631236627698, 'learning_rate/qwen_vl_interface': 5.3413570102415144e-06, 'learning_rate/action_model': 5.3413570102415144e-05, 'epoch': 4.59}) +[2026-06-20 05:47:38,334][starVLA.training.train_starvla][INFO] - Step 2521, Loss: {'train/loss': 0.41156336665153503, 'train/grad_norm_pre_clip': 10.33731746673584, 'timing/data': 0.0067963916808366776, 'timing/model': 2.7525886986404657, 'learning_rate/qwen_vl_interface': 5.338312133542131e-06, 'learning_rate/action_model': 5.3383121335421305e-05, 'epoch': 4.59}) +[2026-06-20 05:47:49,236][starVLA.training.train_starvla][INFO] - Step 2522, Loss: {'train/loss': 0.5860933065414429, 'train/grad_norm_pre_clip': 8.234692573547363, 'timing/data': 0.006761528551578522, 'timing/model': 2.788206009194255, 'learning_rate/qwen_vl_interface': 5.335267220540963e-06, 'learning_rate/action_model': 5.335267220540963e-05, 'epoch': 4.59}) +[2026-06-20 05:48:00,030][starVLA.training.train_starvla][INFO] - Step 2523, Loss: {'train/loss': 0.406431645154953, 'train/grad_norm_pre_clip': 4.721336364746094, 'timing/data': 0.006901970133185387, 'timing/model': 2.7837440632283688, 'learning_rate/qwen_vl_interface': 5.332222272489662e-06, 'learning_rate/action_model': 5.332222272489662e-05, 'epoch': 4.59}) +[2026-06-20 05:48:10,936][starVLA.training.train_starvla][INFO] - Step 2524, Loss: {'train/loss': 0.2578800916671753, 'train/grad_norm_pre_clip': 3.890002727508545, 'timing/data': 0.006603943184018135, 'timing/model': 2.7824607267975807, 'learning_rate/qwen_vl_interface': 5.3291772906398895e-06, 'learning_rate/action_model': 5.32917729063989e-05, 'epoch': 4.59}) +[2026-06-20 05:48:21,832][starVLA.training.train_starvla][INFO] - Step 2525, Loss: {'train/loss': 0.6545909643173218, 'train/grad_norm_pre_clip': 9.171764373779297, 'timing/data': 0.006461042910814285, 'timing/model': 2.775920193642378, 'learning_rate/qwen_vl_interface': 5.326132276243327e-06, 'learning_rate/action_model': 5.3261322762433266e-05, 'epoch': 4.6}) +[2026-06-20 05:48:32,709][starVLA.training.train_starvla][INFO] - Step 2526, Loss: {'train/loss': 0.2880322337150574, 'train/grad_norm_pre_clip': 4.726602077484131, 'timing/data': 0.006528463214635849, 'timing/model': 2.7505147140473127, 'learning_rate/qwen_vl_interface': 5.323087230551659e-06, 'learning_rate/action_model': 5.323087230551659e-05, 'epoch': 4.6}) +[2026-06-20 05:48:43,526][starVLA.training.train_starvla][INFO] - Step 2527, Loss: {'train/loss': 0.6528596878051758, 'train/grad_norm_pre_clip': 7.5851521492004395, 'timing/data': 0.00654793344438076, 'timing/model': 2.772868476808071, 'learning_rate/qwen_vl_interface': 5.3200421548165916e-06, 'learning_rate/action_model': 5.3200421548165914e-05, 'epoch': 4.6}) +[2026-06-20 05:48:54,397][starVLA.training.train_starvla][INFO] - Step 2528, Loss: {'train/loss': 0.6264409422874451, 'train/grad_norm_pre_clip': 8.945762634277344, 'timing/data': 0.006909869611263275, 'timing/model': 2.7385332081466913, 'learning_rate/qwen_vl_interface': 5.3169970502898405e-06, 'learning_rate/action_model': 5.31699705028984e-05, 'epoch': 4.6}) +[2026-06-20 05:49:05,427][starVLA.training.train_starvla][INFO] - Step 2529, Loss: {'train/loss': 0.605849027633667, 'train/grad_norm_pre_clip': 6.484311580657959, 'timing/data': 0.006435444578528404, 'timing/model': 2.770636934787035, 'learning_rate/qwen_vl_interface': 5.31395191822313e-06, 'learning_rate/action_model': 5.31395191822313e-05, 'epoch': 4.6}) +[2026-06-20 05:49:16,329][starVLA.training.train_starvla][INFO] - Step 2530, Loss: {'train/loss': 0.5259069204330444, 'train/grad_norm_pre_clip': 6.318321228027344, 'timing/data': 0.006640581414103508, 'timing/model': 2.775910774245858, 'learning_rate/qwen_vl_interface': 5.310906759868205e-06, 'learning_rate/action_model': 5.310906759868205e-05, 'epoch': 4.61}) +[2026-06-20 05:49:27,229][starVLA.training.train_starvla][INFO] - Step 2531, Loss: {'train/loss': 0.41059035062789917, 'train/grad_norm_pre_clip': 3.87802791595459, 'timing/data': 0.006612082943320274, 'timing/model': 2.7706415876746178, 'learning_rate/qwen_vl_interface': 5.30786157647681e-06, 'learning_rate/action_model': 5.30786157647681e-05, 'epoch': 4.61}) +[2026-06-20 05:49:38,061][starVLA.training.train_starvla][INFO] - Step 2532, Loss: {'train/loss': 0.4390503764152527, 'train/grad_norm_pre_clip': 3.9935953617095947, 'timing/data': 0.006480943411588669, 'timing/model': 2.7046023309230804, 'learning_rate/qwen_vl_interface': 5.304816369300709e-06, 'learning_rate/action_model': 5.304816369300709e-05, 'epoch': 4.61}) +[2026-06-20 05:49:48,929][starVLA.training.train_starvla][INFO] - Step 2533, Loss: {'train/loss': 0.415066123008728, 'train/grad_norm_pre_clip': 5.899432182312012, 'timing/data': 0.006512582302093506, 'timing/model': 2.7785650063306093, 'learning_rate/qwen_vl_interface': 5.301771139591666e-06, 'learning_rate/action_model': 5.301771139591666e-05, 'epoch': 4.61}) +[2026-06-20 05:49:59,829][starVLA.training.train_starvla][INFO] - Step 2534, Loss: {'train/loss': 0.4816826283931732, 'train/grad_norm_pre_clip': 6.08451509475708, 'timing/data': 0.006818540394306183, 'timing/model': 2.7715252488851547, 'learning_rate/qwen_vl_interface': 5.298725888601465e-06, 'learning_rate/action_model': 5.298725888601465e-05, 'epoch': 4.61}) +[2026-06-20 05:50:10,728][starVLA.training.train_starvla][INFO] - Step 2535, Loss: {'train/loss': 0.32042115926742554, 'train/grad_norm_pre_clip': 3.404013156890869, 'timing/data': 0.0065898653119802475, 'timing/model': 2.7743356097489595, 'learning_rate/qwen_vl_interface': 5.295680617581891e-06, 'learning_rate/action_model': 5.2956806175818904e-05, 'epoch': 4.61}) +[2026-06-20 05:50:21,844][starVLA.training.train_starvla][INFO] - Step 2536, Loss: {'train/loss': 0.323971152305603, 'train/grad_norm_pre_clip': 2.2929675579071045, 'timing/data': 0.00663875974714756, 'timing/model': 2.9914360120892525, 'learning_rate/qwen_vl_interface': 5.2926353277847395e-06, 'learning_rate/action_model': 5.292635327784739e-05, 'epoch': 4.62}) +[2026-06-20 05:50:32,738][starVLA.training.train_starvla][INFO] - Step 2537, Loss: {'train/loss': 0.4718455374240875, 'train/grad_norm_pre_clip': 4.457642078399658, 'timing/data': 0.006339496001601219, 'timing/model': 2.7755809780210257, 'learning_rate/qwen_vl_interface': 5.2895900204618175e-06, 'learning_rate/action_model': 5.289590020461818e-05, 'epoch': 4.62}) +[2026-06-20 05:50:43,641][starVLA.training.train_starvla][INFO] - Step 2538, Loss: {'train/loss': 0.46956828236579895, 'train/grad_norm_pre_clip': 3.4807333946228027, 'timing/data': 0.006682179868221283, 'timing/model': 2.7918641213327646, 'learning_rate/qwen_vl_interface': 5.286544696864933e-06, 'learning_rate/action_model': 5.2865446968649324e-05, 'epoch': 4.62}) +[2026-06-20 05:50:54,537][starVLA.training.train_starvla][INFO] - Step 2539, Loss: {'train/loss': 0.39324697852134705, 'train/grad_norm_pre_clip': 5.138810634613037, 'timing/data': 0.006537472829222679, 'timing/model': 2.777271192520857, 'learning_rate/qwen_vl_interface': 5.283499358245904e-06, 'learning_rate/action_model': 5.2834993582459034e-05, 'epoch': 4.62}) +[2026-06-20 05:51:05,638][starVLA.training.train_starvla][INFO] - Step 2540, Loss: {'train/loss': 0.3764902353286743, 'train/grad_norm_pre_clip': 3.930351495742798, 'timing/data': 0.006905639544129372, 'timing/model': 2.7474511358886957, 'learning_rate/qwen_vl_interface': 5.280454005856553e-06, 'learning_rate/action_model': 5.2804540058565534e-05, 'epoch': 4.62}) +[2026-06-20 05:51:16,536][starVLA.training.train_starvla][INFO] - Step 2541, Loss: {'train/loss': 0.519363284111023, 'train/grad_norm_pre_clip': 22.457597732543945, 'timing/data': 0.006495414301753044, 'timing/model': 2.7766079399734735, 'learning_rate/qwen_vl_interface': 5.277408640948711e-06, 'learning_rate/action_model': 5.277408640948711e-05, 'epoch': 4.63}) +[2026-06-20 05:51:27,442][starVLA.training.train_starvla][INFO] - Step 2542, Loss: {'train/loss': 0.3520437777042389, 'train/grad_norm_pre_clip': 4.215785026550293, 'timing/data': 0.006557101383805275, 'timing/model': 2.783134412020445, 'learning_rate/qwen_vl_interface': 5.274363264774213e-06, 'learning_rate/action_model': 5.274363264774212e-05, 'epoch': 4.63}) +[2026-06-20 05:51:38,335][starVLA.training.train_starvla][INFO] - Step 2543, Loss: {'train/loss': 0.5039405226707458, 'train/grad_norm_pre_clip': 3.85092830657959, 'timing/data': 0.006763959303498268, 'timing/model': 2.770856147632003, 'learning_rate/qwen_vl_interface': 5.271317878584897e-06, 'learning_rate/action_model': 5.2713178785848974e-05, 'epoch': 4.63}) +[2026-06-20 05:51:49,238][starVLA.training.train_starvla][INFO] - Step 2544, Loss: {'train/loss': 0.5429818630218506, 'train/grad_norm_pre_clip': 5.481074333190918, 'timing/data': 0.00647040456533432, 'timing/model': 2.7877153251320124, 'learning_rate/qwen_vl_interface': 5.268272483632606e-06, 'learning_rate/action_model': 5.268272483632606e-05, 'epoch': 4.63}) +[2026-06-20 05:52:00,137][starVLA.training.train_starvla][INFO] - Step 2545, Loss: {'train/loss': 0.3200232684612274, 'train/grad_norm_pre_clip': 3.833596706390381, 'timing/data': 0.006831429898738861, 'timing/model': 2.7774422094225883, 'learning_rate/qwen_vl_interface': 5.265227081169187e-06, 'learning_rate/action_model': 5.265227081169186e-05, 'epoch': 4.63}) +[2026-06-20 05:52:11,039][starVLA.training.train_starvla][INFO] - Step 2546, Loss: {'train/loss': 0.3805001676082611, 'train/grad_norm_pre_clip': 4.432682037353516, 'timing/data': 0.00661410391330719, 'timing/model': 2.7738020289689302, 'learning_rate/qwen_vl_interface': 5.2621816724464895e-06, 'learning_rate/action_model': 5.262181672446489e-05, 'epoch': 4.63}) +[2026-06-20 05:52:21,929][starVLA.training.train_starvla][INFO] - Step 2547, Loss: {'train/loss': 0.5080825090408325, 'train/grad_norm_pre_clip': 5.170691967010498, 'timing/data': 0.006721280515193939, 'timing/model': 2.766366360709071, 'learning_rate/qwen_vl_interface': 5.259136258716367e-06, 'learning_rate/action_model': 5.2591362587163675e-05, 'epoch': 4.64}) +[2026-06-20 05:52:33,040][starVLA.training.train_starvla][INFO] - Step 2548, Loss: {'train/loss': 0.5331747531890869, 'train/grad_norm_pre_clip': 4.7806715965271, 'timing/data': 0.006629250943660736, 'timing/model': 2.7884492985904217, 'learning_rate/qwen_vl_interface': 5.256090841230675e-06, 'learning_rate/action_model': 5.256090841230674e-05, 'epoch': 4.64}) +[2026-06-20 05:52:43,936][starVLA.training.train_starvla][INFO] - Step 2549, Loss: {'train/loss': 0.26478251814842224, 'train/grad_norm_pre_clip': 3.303896427154541, 'timing/data': 0.00680253840982914, 'timing/model': 2.7809001859277487, 'learning_rate/qwen_vl_interface': 5.253045421241266e-06, 'learning_rate/action_model': 5.253045421241266e-05, 'epoch': 4.64}) +[2026-06-20 05:52:54,838][starVLA.training.train_starvla][INFO] - Step 2550, Loss: {'train/loss': 0.5394920706748962, 'train/grad_norm_pre_clip': 5.775993824005127, 'timing/data': 0.0065555088222026825, 'timing/model': 2.7769970558583736, 'learning_rate/qwen_vl_interface': 5.2500000000000006e-06, 'learning_rate/action_model': 5.25e-05, 'epoch': 4.64}) +[2026-06-20 05:53:05,939][starVLA.training.train_starvla][INFO] - Step 2551, Loss: {'train/loss': 0.4868692457675934, 'train/grad_norm_pre_clip': 4.320923328399658, 'timing/data': 0.006928848102688789, 'timing/model': 2.7765448596328497, 'learning_rate/qwen_vl_interface': 5.246954578758735e-06, 'learning_rate/action_model': 5.246954578758735e-05, 'epoch': 4.64}) +[2026-06-20 05:53:16,843][starVLA.training.train_starvla][INFO] - Step 2552, Loss: {'train/loss': 0.3751709461212158, 'train/grad_norm_pre_clip': 3.9903273582458496, 'timing/data': 0.00665237195789814, 'timing/model': 2.779132902622223, 'learning_rate/qwen_vl_interface': 5.243909158769327e-06, 'learning_rate/action_model': 5.2439091587693264e-05, 'epoch': 4.65}) +[2026-06-20 05:53:27,722][starVLA.training.train_starvla][INFO] - Step 2553, Loss: {'train/loss': 0.5217124819755554, 'train/grad_norm_pre_clip': 6.759532928466797, 'timing/data': 0.006292074918746948, 'timing/model': 2.761711983010173, 'learning_rate/qwen_vl_interface': 5.240863741283634e-06, 'learning_rate/action_model': 5.2408637412836336e-05, 'epoch': 4.65}) +[2026-06-20 05:53:38,541][starVLA.training.train_starvla][INFO] - Step 2554, Loss: {'train/loss': 0.5592164397239685, 'train/grad_norm_pre_clip': 4.178248405456543, 'timing/data': 0.007215335965156555, 'timing/model': 2.7687208876013756, 'learning_rate/qwen_vl_interface': 5.2378183275535125e-06, 'learning_rate/action_model': 5.237818327553512e-05, 'epoch': 4.65}) +[2026-06-20 05:53:49,676][starVLA.training.train_starvla][INFO] - Step 2555, Loss: {'train/loss': 0.3317832052707672, 'train/grad_norm_pre_clip': 4.061176776885986, 'timing/data': 0.006793348118662834, 'timing/model': 3.0206208005547523, 'learning_rate/qwen_vl_interface': 5.234772918830815e-06, 'learning_rate/action_model': 5.234772918830815e-05, 'epoch': 4.65}) +[2026-06-20 05:54:00,428][starVLA.training.train_starvla][INFO] - Step 2556, Loss: {'train/loss': 0.3805009722709656, 'train/grad_norm_pre_clip': 5.552830219268799, 'timing/data': 0.0065750498324632645, 'timing/model': 2.763672299683094, 'learning_rate/qwen_vl_interface': 5.231727516367395e-06, 'learning_rate/action_model': 5.231727516367395e-05, 'epoch': 4.65}) +[2026-06-20 05:54:11,340][starVLA.training.train_starvla][INFO] - Step 2557, Loss: {'train/loss': 0.35878661274909973, 'train/grad_norm_pre_clip': 4.605045318603516, 'timing/data': 0.006416983902454376, 'timing/model': 2.7761820070445538, 'learning_rate/qwen_vl_interface': 5.228682121415104e-06, 'learning_rate/action_model': 5.228682121415104e-05, 'epoch': 4.65}) +[2026-06-20 05:54:22,236][starVLA.training.train_starvla][INFO] - Step 2558, Loss: {'train/loss': 0.39845141768455505, 'train/grad_norm_pre_clip': 4.878006458282471, 'timing/data': 0.0065882205963134766, 'timing/model': 2.771899012848735, 'learning_rate/qwen_vl_interface': 5.225636735225789e-06, 'learning_rate/action_model': 5.225636735225788e-05, 'epoch': 4.66}) +[2026-06-20 05:54:33,337][starVLA.training.train_starvla][INFO] - Step 2559, Loss: {'train/loss': 0.403668612241745, 'train/grad_norm_pre_clip': 7.016496181488037, 'timing/data': 0.009081315249204636, 'timing/model': 2.745854778215289, 'learning_rate/qwen_vl_interface': 5.22259135905129e-06, 'learning_rate/action_model': 5.2225913590512895e-05, 'epoch': 4.66}) +[2026-06-20 05:54:44,238][starVLA.training.train_starvla][INFO] - Step 2560, Loss: {'train/loss': 0.532317578792572, 'train/grad_norm_pre_clip': 5.335728645324707, 'timing/data': 0.009628329426050186, 'timing/model': 2.769036205485463, 'learning_rate/qwen_vl_interface': 5.219545994143448e-06, 'learning_rate/action_model': 5.219545994143448e-05, 'epoch': 4.66}) +[2026-06-20 05:54:55,139][starVLA.training.train_starvla][INFO] - Step 2561, Loss: {'train/loss': 0.4741041362285614, 'train/grad_norm_pre_clip': 4.600339889526367, 'timing/data': 0.006586482748389244, 'timing/model': 2.7756484244018793, 'learning_rate/qwen_vl_interface': 5.216500641754097e-06, 'learning_rate/action_model': 5.216500641754097e-05, 'epoch': 4.66}) +[2026-06-20 05:55:06,025][starVLA.training.train_starvla][INFO] - Step 2562, Loss: {'train/loss': 0.21322974562644958, 'train/grad_norm_pre_clip': 4.763965606689453, 'timing/data': 0.00653168186545372, 'timing/model': 2.761613069102168, 'learning_rate/qwen_vl_interface': 5.213455303135067e-06, 'learning_rate/action_model': 5.2134553031350673e-05, 'epoch': 4.66}) +[2026-06-20 05:55:16,938][starVLA.training.train_starvla][INFO] - Step 2563, Loss: {'train/loss': 0.31708651781082153, 'train/grad_norm_pre_clip': 5.67076301574707, 'timing/data': 0.006780281662940979, 'timing/model': 2.7747201696038246, 'learning_rate/qwen_vl_interface': 5.2104099795381836e-06, 'learning_rate/action_model': 5.210409979538183e-05, 'epoch': 4.67}) +[2026-06-20 05:55:27,831][starVLA.training.train_starvla][INFO] - Step 2564, Loss: {'train/loss': 0.226236492395401, 'train/grad_norm_pre_clip': 3.784134864807129, 'timing/data': 0.006852639839053154, 'timing/model': 2.7698569279164076, 'learning_rate/qwen_vl_interface': 5.207364672215261e-06, 'learning_rate/action_model': 5.2073646722152604e-05, 'epoch': 4.67}) +[2026-06-20 05:55:38,733][starVLA.training.train_starvla][INFO] - Step 2565, Loss: {'train/loss': 0.14519983530044556, 'train/grad_norm_pre_clip': 3.1065657138824463, 'timing/data': 0.006821921095252037, 'timing/model': 2.7664159201085567, 'learning_rate/qwen_vl_interface': 5.2043193824181095e-06, 'learning_rate/action_model': 5.2043193824181094e-05, 'epoch': 4.67}) +[2026-06-20 05:55:49,632][starVLA.training.train_starvla][INFO] - Step 2566, Loss: {'train/loss': 0.3190349042415619, 'train/grad_norm_pre_clip': 3.935164451599121, 'timing/data': 0.0066660307347774506, 'timing/model': 2.770406434312463, 'learning_rate/qwen_vl_interface': 5.201274111398535e-06, 'learning_rate/action_model': 5.2012741113985356e-05, 'epoch': 4.67}) +[2026-06-20 05:56:00,729][starVLA.training.train_starvla][INFO] - Step 2567, Loss: {'train/loss': 0.28218016028404236, 'train/grad_norm_pre_clip': 5.040124416351318, 'timing/data': 0.0069073885679244995, 'timing/model': 2.758971620351076, 'learning_rate/qwen_vl_interface': 5.198228860408334e-06, 'learning_rate/action_model': 5.198228860408334e-05, 'epoch': 4.67}) +[2026-06-20 05:56:11,640][starVLA.training.train_starvla][INFO] - Step 2568, Loss: {'train/loss': 0.5111947655677795, 'train/grad_norm_pre_clip': 6.655897617340088, 'timing/data': 0.00664704293012619, 'timing/model': 2.778561744838953, 'learning_rate/qwen_vl_interface': 5.195183630699292e-06, 'learning_rate/action_model': 5.195183630699292e-05, 'epoch': 4.67}) +[2026-06-20 05:56:22,538][starVLA.training.train_starvla][INFO] - Step 2569, Loss: {'train/loss': 0.5667089819908142, 'train/grad_norm_pre_clip': 6.266008377075195, 'timing/data': 0.006570290774106979, 'timing/model': 2.7767521888017654, 'learning_rate/qwen_vl_interface': 5.192138423523189e-06, 'learning_rate/action_model': 5.1921384235231895e-05, 'epoch': 4.68}) +[2026-06-20 05:56:33,435][starVLA.training.train_starvla][INFO] - Step 2570, Loss: {'train/loss': 0.3961339592933655, 'train/grad_norm_pre_clip': 5.037969589233398, 'timing/data': 0.00683102011680603, 'timing/model': 2.7737378776073456, 'learning_rate/qwen_vl_interface': 5.189093240131795e-06, 'learning_rate/action_model': 5.189093240131795e-05, 'epoch': 4.68}) +[2026-06-20 05:56:44,340][starVLA.training.train_starvla][INFO] - Step 2571, Loss: {'train/loss': 0.5144164562225342, 'train/grad_norm_pre_clip': 10.41146469116211, 'timing/data': 0.006666790693998337, 'timing/model': 2.7889370750635862, 'learning_rate/qwen_vl_interface': 5.186048081776869e-06, 'learning_rate/action_model': 5.1860480817768685e-05, 'epoch': 4.68}) +[2026-06-20 05:56:55,240][starVLA.training.train_starvla][INFO] - Step 2572, Loss: {'train/loss': 0.4346056282520294, 'train/grad_norm_pre_clip': 6.0138163566589355, 'timing/data': 0.006810219958424568, 'timing/model': 2.784315289929509, 'learning_rate/qwen_vl_interface': 5.183002949710161e-06, 'learning_rate/action_model': 5.18300294971016e-05, 'epoch': 4.68}) +[2026-06-20 05:57:06,133][starVLA.training.train_starvla][INFO] - Step 2573, Loss: {'train/loss': 0.5025306940078735, 'train/grad_norm_pre_clip': 6.927627086639404, 'timing/data': 0.006842589005827904, 'timing/model': 2.7686829399317503, 'learning_rate/qwen_vl_interface': 5.179957845183411e-06, 'learning_rate/action_model': 5.179957845183411e-05, 'epoch': 4.68}) +[2026-06-20 05:57:17,255][starVLA.training.train_starvla][INFO] - Step 2574, Loss: {'train/loss': 0.4555656313896179, 'train/grad_norm_pre_clip': 5.024550437927246, 'timing/data': 0.006637291982769966, 'timing/model': 2.9946592561900616, 'learning_rate/qwen_vl_interface': 5.176912769448343e-06, 'learning_rate/action_model': 5.176912769448343e-05, 'epoch': 4.69}) +[2026-06-20 05:57:28,136][starVLA.training.train_starvla][INFO] - Step 2575, Loss: {'train/loss': 0.48256802558898926, 'train/grad_norm_pre_clip': 5.332505702972412, 'timing/data': 0.006620530039072037, 'timing/model': 2.772273601964116, 'learning_rate/qwen_vl_interface': 5.173867723756676e-06, 'learning_rate/action_model': 5.173867723756676e-05, 'epoch': 4.69}) +[2026-06-20 05:57:39,039][starVLA.training.train_starvla][INFO] - Step 2576, Loss: {'train/loss': 0.4079565107822418, 'train/grad_norm_pre_clip': 4.952523708343506, 'timing/data': 0.0067950524389743805, 'timing/model': 2.7804114427417517, 'learning_rate/qwen_vl_interface': 5.170822709360112e-06, 'learning_rate/action_model': 5.170822709360111e-05, 'epoch': 4.69}) +[2026-06-20 05:57:49,937][starVLA.training.train_starvla][INFO] - Step 2577, Loss: {'train/loss': 0.4399031102657318, 'train/grad_norm_pre_clip': 4.6288862228393555, 'timing/data': 0.006743039935827255, 'timing/model': 2.7750925067812204, 'learning_rate/qwen_vl_interface': 5.16777772751034e-06, 'learning_rate/action_model': 5.16777772751034e-05, 'epoch': 4.69}) +[2026-06-20 05:58:01,041][starVLA.training.train_starvla][INFO] - Step 2578, Loss: {'train/loss': 0.3427622318267822, 'train/grad_norm_pre_clip': 6.224424839019775, 'timing/data': 0.006746601313352585, 'timing/model': 2.7562450747936964, 'learning_rate/qwen_vl_interface': 5.16473277945904e-06, 'learning_rate/action_model': 5.1647327794590396e-05, 'epoch': 4.69}) +[2026-06-20 05:58:11,937][starVLA.training.train_starvla][INFO] - Step 2579, Loss: {'train/loss': 0.3640914559364319, 'train/grad_norm_pre_clip': 9.97828197479248, 'timing/data': 0.006623281165957451, 'timing/model': 2.7698061298578978, 'learning_rate/qwen_vl_interface': 5.161687866457872e-06, 'learning_rate/action_model': 5.161687866457871e-05, 'epoch': 4.69}) +[2026-06-20 05:58:22,831][starVLA.training.train_starvla][INFO] - Step 2580, Loss: {'train/loss': 0.389825701713562, 'train/grad_norm_pre_clip': 4.718049049377441, 'timing/data': 0.006775667890906334, 'timing/model': 2.7666782960295677, 'learning_rate/qwen_vl_interface': 5.158642989758487e-06, 'learning_rate/action_model': 5.158642989758487e-05, 'epoch': 4.7}) +[2026-06-20 05:58:33,740][starVLA.training.train_starvla][INFO] - Step 2581, Loss: {'train/loss': 0.3570209741592407, 'train/grad_norm_pre_clip': 4.261622905731201, 'timing/data': 0.006495404988527298, 'timing/model': 2.7827100586146116, 'learning_rate/qwen_vl_interface': 5.155598150612519e-06, 'learning_rate/action_model': 5.1555981506125185e-05, 'epoch': 4.7}) +[2026-06-20 05:58:44,639][starVLA.training.train_starvla][INFO] - Step 2582, Loss: {'train/loss': 0.3737250864505768, 'train/grad_norm_pre_clip': 4.552599906921387, 'timing/data': 0.006603049114346504, 'timing/model': 2.786442391574383, 'learning_rate/qwen_vl_interface': 5.152553350271584e-06, 'learning_rate/action_model': 5.152553350271584e-05, 'epoch': 4.7}) +[2026-06-20 05:58:55,543][starVLA.training.train_starvla][INFO] - Step 2583, Loss: {'train/loss': 0.4201798737049103, 'train/grad_norm_pre_clip': 6.745172023773193, 'timing/data': 0.006500011309981346, 'timing/model': 2.7902894876897335, 'learning_rate/qwen_vl_interface': 5.149508589987286e-06, 'learning_rate/action_model': 5.1495085899872855e-05, 'epoch': 4.7}) +[2026-06-20 05:59:06,435][starVLA.training.train_starvla][INFO] - Step 2584, Loss: {'train/loss': 0.3325313329696655, 'train/grad_norm_pre_clip': 5.438939094543457, 'timing/data': 0.006458071991801262, 'timing/model': 2.7735145166516304, 'learning_rate/qwen_vl_interface': 5.14646387101121e-06, 'learning_rate/action_model': 5.14646387101121e-05, 'epoch': 4.7}) +[2026-06-20 05:59:17,336][starVLA.training.train_starvla][INFO] - Step 2585, Loss: {'train/loss': 0.32628828287124634, 'train/grad_norm_pre_clip': 7.010700702667236, 'timing/data': 0.0068497806787490845, 'timing/model': 2.7786866184324026, 'learning_rate/qwen_vl_interface': 5.143419194594928e-06, 'learning_rate/action_model': 5.143419194594927e-05, 'epoch': 4.71}) +[2026-06-20 05:59:28,439][starVLA.training.train_starvla][INFO] - Step 2586, Loss: {'train/loss': 0.4155566096305847, 'train/grad_norm_pre_clip': 4.76569128036499, 'timing/data': 0.006509939208626747, 'timing/model': 2.7823785841464996, 'learning_rate/qwen_vl_interface': 5.140374561989988e-06, 'learning_rate/action_model': 5.140374561989988e-05, 'epoch': 4.71}) +[2026-06-20 05:59:39,310][starVLA.training.train_starvla][INFO] - Step 2587, Loss: {'train/loss': 0.48857221007347107, 'train/grad_norm_pre_clip': 6.596163749694824, 'timing/data': 0.00665426068007946, 'timing/model': 2.7474953085184097, 'learning_rate/qwen_vl_interface': 5.137329974447924e-06, 'learning_rate/action_model': 5.137329974447924e-05, 'epoch': 4.71}) +[2026-06-20 05:59:50,138][starVLA.training.train_starvla][INFO] - Step 2588, Loss: {'train/loss': 0.3069554269313812, 'train/grad_norm_pre_clip': 4.85524845123291, 'timing/data': 0.006818048655986786, 'timing/model': 2.7829739563167095, 'learning_rate/qwen_vl_interface': 5.1342854332202516e-06, 'learning_rate/action_model': 5.1342854332202514e-05, 'epoch': 4.71}) +[2026-06-20 06:00:01,034][starVLA.training.train_starvla][INFO] - Step 2589, Loss: {'train/loss': 0.17476551234722137, 'train/grad_norm_pre_clip': 4.067751407623291, 'timing/data': 0.006729321554303169, 'timing/model': 2.7858431227505207, 'learning_rate/qwen_vl_interface': 5.131240939558467e-06, 'learning_rate/action_model': 5.131240939558467e-05, 'epoch': 4.71}) +[2026-06-20 06:00:11,929][starVLA.training.train_starvla][INFO] - Step 2590, Loss: {'train/loss': 0.276075541973114, 'train/grad_norm_pre_clip': 5.193172931671143, 'timing/data': 0.00677834078669548, 'timing/model': 2.7741681840270758, 'learning_rate/qwen_vl_interface': 5.128196494714048e-06, 'learning_rate/action_model': 5.128196494714047e-05, 'epoch': 4.71}) +[2026-06-20 06:00:22,829][starVLA.training.train_starvla][INFO] - Step 2591, Loss: {'train/loss': 0.354774534702301, 'train/grad_norm_pre_clip': 6.453070640563965, 'timing/data': 0.006579693406820297, 'timing/model': 2.7721610069274902, 'learning_rate/qwen_vl_interface': 5.125152099938448e-06, 'learning_rate/action_model': 5.125152099938447e-05, 'epoch': 4.72}) +[2026-06-20 06:00:33,735][starVLA.training.train_starvla][INFO] - Step 2592, Loss: {'train/loss': 0.2781505584716797, 'train/grad_norm_pre_clip': 3.9645488262176514, 'timing/data': 0.0071176569908857346, 'timing/model': 2.78240854665637, 'learning_rate/qwen_vl_interface': 5.122107756483104e-06, 'learning_rate/action_model': 5.122107756483103e-05, 'epoch': 4.72}) +[2026-06-20 06:00:44,843][starVLA.training.train_starvla][INFO] - Step 2593, Loss: {'train/loss': 0.21089549362659454, 'train/grad_norm_pre_clip': 3.2559268474578857, 'timing/data': 0.006547670811414719, 'timing/model': 2.981303211301565, 'learning_rate/qwen_vl_interface': 5.119063465599429e-06, 'learning_rate/action_model': 5.1190634655994285e-05, 'epoch': 4.72}) +[2026-06-20 06:00:55,739][starVLA.training.train_starvla][INFO] - Step 2594, Loss: {'train/loss': 0.28490766882896423, 'train/grad_norm_pre_clip': 9.484779357910156, 'timing/data': 0.006909407675266266, 'timing/model': 2.774339970201254, 'learning_rate/qwen_vl_interface': 5.116019228538818e-06, 'learning_rate/action_model': 5.116019228538818e-05, 'epoch': 4.72}) +[2026-06-20 06:01:06,516][starVLA.training.train_starvla][INFO] - Step 2595, Loss: {'train/loss': 0.21735525131225586, 'train/grad_norm_pre_clip': 4.109523296356201, 'timing/data': 0.006841979920864105, 'timing/model': 2.705671340227127, 'learning_rate/qwen_vl_interface': 5.112975046552641e-06, 'learning_rate/action_model': 5.1129750465526404e-05, 'epoch': 4.72}) +[2026-06-20 06:01:17,328][starVLA.training.train_starvla][INFO] - Step 2596, Loss: {'train/loss': 0.24420426785945892, 'train/grad_norm_pre_clip': 6.387289524078369, 'timing/data': 0.007003797218203545, 'timing/model': 2.7739424519240856, 'learning_rate/qwen_vl_interface': 5.109930920892246e-06, 'learning_rate/action_model': 5.1099309208922464e-05, 'epoch': 4.73}) +[2026-06-20 06:01:28,335][starVLA.training.train_starvla][INFO] - Step 2597, Loss: {'train/loss': 0.2939472496509552, 'train/grad_norm_pre_clip': 12.670228004455566, 'timing/data': 0.006934259086847305, 'timing/model': 2.7360179480165243, 'learning_rate/qwen_vl_interface': 5.10688685280896e-06, 'learning_rate/action_model': 5.10688685280896e-05, 'epoch': 4.73}) +[2026-06-20 06:01:39,132][starVLA.training.train_starvla][INFO] - Step 2598, Loss: {'train/loss': 0.27525630593299866, 'train/grad_norm_pre_clip': 5.887885093688965, 'timing/data': 0.006798561662435532, 'timing/model': 2.752077104523778, 'learning_rate/qwen_vl_interface': 5.103842843554082e-06, 'learning_rate/action_model': 5.103842843554082e-05, 'epoch': 4.73}) +[2026-06-20 06:01:49,926][starVLA.training.train_starvla][INFO] - Step 2599, Loss: {'train/loss': 0.12374941259622574, 'train/grad_norm_pre_clip': 5.0093302726745605, 'timing/data': 0.006566785275936127, 'timing/model': 2.753258775919676, 'learning_rate/qwen_vl_interface': 5.100798894378891e-06, 'learning_rate/action_model': 5.100798894378891e-05, 'epoch': 4.73}) +[2026-06-20 06:02:00,834][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 2600 +[2026-06-20 06:09:32,298][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 2600 +[2026-06-20 06:09:32,299][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 2600 +[2026-06-20 06:17:41,415][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 2600 +[2026-06-20 06:17:41,417][starVLA.training.train_starvla][INFO] - Step 2600, Loss: {'train/loss': 0.2228391468524933, 'train/grad_norm_pre_clip': 3.346384048461914, 'eval/loss': 0.2954998314380646, 'eval/latency_0/loss': 0.27991625666618347, 'eval/latency_1/loss': 0.34117308259010315, 'eval/latency_2/loss': 0.26624199748039246, 'eval/latency_3/loss': 0.18155436217784882, 'eval/latency_4/loss': 0.39828237891197205, 'eval/flappy/loss': 0.2954998314380646, 'eval/flappy/latency_0/loss': 0.27991625666618347, 'eval/flappy/latency_1/loss': 0.34117308259010315, 'eval/flappy/latency_2/loss': 0.26624199748039246, 'eval/flappy/latency_3/loss': 0.18155436217784882, 'eval/flappy/latency_4/loss': 0.39828237891197205, 'eval/flap_precision': 0.8387096774193549, 'eval/flap_recall': 0.1890909090909091, 'eval/flap_f1': 0.3086053412462908, 'eval/flap_rate_model': 0.017222222222222222, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.3086053412462908, 'eval/latency_0/flap_precision': 0.8387096774193549, 'eval/latency_0/flap_recall': 0.1890909090909091, 'eval/latency_0/flap_f1': 0.3086053412462908, 'eval/latency_0/flap_rate_model': 0.017222222222222222, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.3086053412462908, 'timing/data': 0.00655217282474041, 'timing/model': 2.7817288525402546, 'learning_rate/qwen_vl_interface': 5.097755006534638e-06, 'learning_rate/action_model': 5.097755006534638e-05, 'epoch': 4.73}) +[2026-06-20 06:17:41,417][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_2600_state +[2026-06-20 06:17:41,438][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-20 06:17:47,239][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_2600_state/model.safetensors +[2026-06-20 06:18:01,313][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_2600_state/optimizer.bin +[2026-06-20 06:18:01,313][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_2600_state/sampler.bin +[2026-06-20 06:18:01,313][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_2600_state/sampler_1.bin +[2026-06-20 06:18:01,314][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_2600_state/random_states_0.pkl +[2026-06-20 06:18:01,442][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2600 +[2026-06-20 06:18:01,442][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-20 06:18:01,456][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-20 06:18:12,326][starVLA.training.train_starvla][INFO] - Step 2601, Loss: {'train/loss': 0.2690171003341675, 'train/grad_norm_pre_clip': 5.981978416442871, 'timing/data': 0.006976539269089699, 'timing/model': 2.781080447137356, 'learning_rate/qwen_vl_interface': 5.094711181272551e-06, 'learning_rate/action_model': 5.094711181272551e-05, 'epoch': 4.73}) +[2026-06-20 06:18:23,426][starVLA.training.train_starvla][INFO] - Step 2602, Loss: {'train/loss': 0.7719182968139648, 'train/grad_norm_pre_clip': 15.842336654663086, 'timing/data': 0.00688081793487072, 'timing/model': 2.7876512594521046, 'learning_rate/qwen_vl_interface': 5.091667419843833e-06, 'learning_rate/action_model': 5.091667419843833e-05, 'epoch': 4.74}) +[2026-06-20 06:18:34,330][starVLA.training.train_starvla][INFO] - Step 2603, Loss: {'train/loss': 0.31909123063087463, 'train/grad_norm_pre_clip': 7.106058597564697, 'timing/data': 0.006697390228509903, 'timing/model': 2.783922925591469, 'learning_rate/qwen_vl_interface': 5.088623723499657e-06, 'learning_rate/action_model': 5.088623723499657e-05, 'epoch': 4.74}) +[2026-06-20 06:18:45,234][starVLA.training.train_starvla][INFO] - Step 2604, Loss: {'train/loss': 0.5638740658760071, 'train/grad_norm_pre_clip': 8.420454978942871, 'timing/data': 0.007149484008550644, 'timing/model': 2.7816905807703733, 'learning_rate/qwen_vl_interface': 5.085580093491171e-06, 'learning_rate/action_model': 5.085580093491171e-05, 'epoch': 4.74}) +[2026-06-20 06:18:56,130][starVLA.training.train_starvla][INFO] - Step 2605, Loss: {'train/loss': 0.5119425654411316, 'train/grad_norm_pre_clip': 9.982476234436035, 'timing/data': 0.00676305964589119, 'timing/model': 2.7750147487968206, 'learning_rate/qwen_vl_interface': 5.0825365310695e-06, 'learning_rate/action_model': 5.082536531069499e-05, 'epoch': 4.74}) +[2026-06-20 06:19:07,229][starVLA.training.train_starvla][INFO] - Step 2606, Loss: {'train/loss': 0.23207169771194458, 'train/grad_norm_pre_clip': 7.182315826416016, 'timing/data': 0.006803169846534729, 'timing/model': 2.7688449807465076, 'learning_rate/qwen_vl_interface': 5.079493037485735e-06, 'learning_rate/action_model': 5.0794930374857344e-05, 'epoch': 4.74}) +[2026-06-20 06:19:18,134][starVLA.training.train_starvla][INFO] - Step 2607, Loss: {'train/loss': 0.3048470616340637, 'train/grad_norm_pre_clip': 6.373602390289307, 'timing/data': 0.006571725010871887, 'timing/model': 2.778711672872305, 'learning_rate/qwen_vl_interface': 5.076449613990941e-06, 'learning_rate/action_model': 5.076449613990941e-05, 'epoch': 4.75}) +[2026-06-20 06:19:29,031][starVLA.training.train_starvla][INFO] - Step 2608, Loss: {'train/loss': 0.32157769799232483, 'train/grad_norm_pre_clip': 8.180216789245605, 'timing/data': 0.006803326308727264, 'timing/model': 2.769995380192995, 'learning_rate/qwen_vl_interface': 5.073406261836157e-06, 'learning_rate/action_model': 5.0734062618361565e-05, 'epoch': 4.75}) +[2026-06-20 06:19:40,159][starVLA.training.train_starvla][INFO] - Step 2609, Loss: {'train/loss': 0.14289483428001404, 'train/grad_norm_pre_clip': 6.876204967498779, 'timing/data': 0.0066605303436517715, 'timing/model': 2.9980302825570107, 'learning_rate/qwen_vl_interface': 5.070362982272386e-06, 'learning_rate/action_model': 5.070362982272385e-05, 'epoch': 4.75}) +[2026-06-20 06:19:51,033][starVLA.training.train_starvla][INFO] - Step 2610, Loss: {'train/loss': 0.1352194845676422, 'train/grad_norm_pre_clip': 4.414078235626221, 'timing/data': 0.006591591984033585, 'timing/model': 2.7746312748640776, 'learning_rate/qwen_vl_interface': 5.067319776550607e-06, 'learning_rate/action_model': 5.067319776550607e-05, 'epoch': 4.75}) +[2026-06-20 06:20:01,938][starVLA.training.train_starvla][INFO] - Step 2611, Loss: {'train/loss': 0.4273209273815155, 'train/grad_norm_pre_clip': 12.975607872009277, 'timing/data': 0.006555778905749321, 'timing/model': 2.778271496295929, 'learning_rate/qwen_vl_interface': 5.064276645921768e-06, 'learning_rate/action_model': 5.064276645921768e-05, 'epoch': 4.75}) +[2026-06-20 06:20:12,829][starVLA.training.train_starvla][INFO] - Step 2612, Loss: {'train/loss': 0.307813823223114, 'train/grad_norm_pre_clip': 7.418641567230225, 'timing/data': 0.006641792133450508, 'timing/model': 2.7747296262532473, 'learning_rate/qwen_vl_interface': 5.061233591636787e-06, 'learning_rate/action_model': 5.061233591636787e-05, 'epoch': 4.75}) +[2026-06-20 06:20:23,739][starVLA.training.train_starvla][INFO] - Step 2613, Loss: {'train/loss': 0.46787846088409424, 'train/grad_norm_pre_clip': 8.330792427062988, 'timing/data': 0.006515141576528549, 'timing/model': 2.778143748641014, 'learning_rate/qwen_vl_interface': 5.058190614946544e-06, 'learning_rate/action_model': 5.058190614946544e-05, 'epoch': 4.76}) +[2026-06-20 06:20:34,639][starVLA.training.train_starvla][INFO] - Step 2614, Loss: {'train/loss': 0.25077706575393677, 'train/grad_norm_pre_clip': 11.30126667022705, 'timing/data': 0.006651604548096657, 'timing/model': 2.781429436057806, 'learning_rate/qwen_vl_interface': 5.055147717101895e-06, 'learning_rate/action_model': 5.055147717101894e-05, 'epoch': 4.76}) +[2026-06-20 06:20:45,530][starVLA.training.train_starvla][INFO] - Step 2615, Loss: {'train/loss': 0.3931778371334076, 'train/grad_norm_pre_clip': 8.418876647949219, 'timing/data': 0.006906198337674141, 'timing/model': 2.774685787037015, 'learning_rate/qwen_vl_interface': 5.052104899353658e-06, 'learning_rate/action_model': 5.0521048993536577e-05, 'epoch': 4.76}) +[2026-06-20 06:20:56,439][starVLA.training.train_starvla][INFO] - Step 2616, Loss: {'train/loss': 0.16327832639217377, 'train/grad_norm_pre_clip': 8.270843505859375, 'timing/data': 0.006735790520906448, 'timing/model': 2.7775369472801685, 'learning_rate/qwen_vl_interface': 5.049062162952622e-06, 'learning_rate/action_model': 5.0490621629526214e-05, 'epoch': 4.76}) +[2026-06-20 06:21:07,337][starVLA.training.train_starvla][INFO] - Step 2617, Loss: {'train/loss': 0.3723294734954834, 'train/grad_norm_pre_clip': 9.066883087158203, 'timing/data': 0.006916576996445656, 'timing/model': 2.775234231725335, 'learning_rate/qwen_vl_interface': 5.046019509149541e-06, 'learning_rate/action_model': 5.046019509149541e-05, 'epoch': 4.76}) +[2026-06-20 06:21:18,231][starVLA.training.train_starvla][INFO] - Step 2618, Loss: {'train/loss': 0.3049897253513336, 'train/grad_norm_pre_clip': 6.6554484367370605, 'timing/data': 0.006598861888051033, 'timing/model': 2.7706017196178436, 'learning_rate/qwen_vl_interface': 5.042976939195135e-06, 'learning_rate/action_model': 5.0429769391951355e-05, 'epoch': 4.77}) +[2026-06-20 06:21:29,139][starVLA.training.train_starvla][INFO] - Step 2619, Loss: {'train/loss': 0.3336437940597534, 'train/grad_norm_pre_clip': 6.094396591186523, 'timing/data': 0.006440211087465286, 'timing/model': 2.777420397847891, 'learning_rate/qwen_vl_interface': 5.039934454340091e-06, 'learning_rate/action_model': 5.0399344543400907e-05, 'epoch': 4.77}) +[2026-06-20 06:21:40,040][starVLA.training.train_starvla][INFO] - Step 2620, Loss: {'train/loss': 0.18930961191654205, 'train/grad_norm_pre_clip': 7.61920166015625, 'timing/data': 0.006842959672212601, 'timing/model': 2.7743958812206984, 'learning_rate/qwen_vl_interface': 5.036892055835056e-06, 'learning_rate/action_model': 5.0368920558350566e-05, 'epoch': 4.77}) +[2026-06-20 06:21:51,039][starVLA.training.train_starvla][INFO] - Step 2621, Loss: {'train/loss': 0.5582055449485779, 'train/grad_norm_pre_clip': 7.6819610595703125, 'timing/data': 0.006795819848775864, 'timing/model': 2.7740366738289595, 'learning_rate/qwen_vl_interface': 5.033849744930647e-06, 'learning_rate/action_model': 5.033849744930646e-05, 'epoch': 4.77}) +[2026-06-20 06:22:01,946][starVLA.training.train_starvla][INFO] - Step 2622, Loss: {'train/loss': 0.21771183609962463, 'train/grad_norm_pre_clip': 7.329986572265625, 'timing/data': 0.006653541699051857, 'timing/model': 2.7790250554680824, 'learning_rate/qwen_vl_interface': 5.030807522877443e-06, 'learning_rate/action_model': 5.030807522877443e-05, 'epoch': 4.77}) +[2026-06-20 06:22:12,830][starVLA.training.train_starvla][INFO] - Step 2623, Loss: {'train/loss': 0.5751827359199524, 'train/grad_norm_pre_clip': 11.952942848205566, 'timing/data': 0.007012350484728813, 'timing/model': 2.767543962225318, 'learning_rate/qwen_vl_interface': 5.027765390925987e-06, 'learning_rate/action_model': 5.027765390925987e-05, 'epoch': 4.77}) +[2026-06-20 06:22:23,701][starVLA.training.train_starvla][INFO] - Step 2624, Loss: {'train/loss': 0.37792661786079407, 'train/grad_norm_pre_clip': 13.098001480102539, 'timing/data': 0.0066851601004600525, 'timing/model': 2.7407376393675804, 'learning_rate/qwen_vl_interface': 5.024723350326783e-06, 'learning_rate/action_model': 5.024723350326783e-05, 'epoch': 4.78}) +[2026-06-20 06:22:34,522][starVLA.training.train_starvla][INFO] - Step 2625, Loss: {'train/loss': 0.31997185945510864, 'train/grad_norm_pre_clip': 9.758109092712402, 'timing/data': 0.0069396402686834335, 'timing/model': 2.7483622692525387, 'learning_rate/qwen_vl_interface': 5.021681402330303e-06, 'learning_rate/action_model': 5.021681402330303e-05, 'epoch': 4.78}) +[2026-06-20 06:22:45,230][starVLA.training.train_starvla][INFO] - Step 2626, Loss: {'train/loss': 0.32183292508125305, 'train/grad_norm_pre_clip': 9.581570625305176, 'timing/data': 0.006531484425067902, 'timing/model': 2.7636975198984146, 'learning_rate/qwen_vl_interface': 5.018639548186972e-06, 'learning_rate/action_model': 5.0186395481869716e-05, 'epoch': 4.78}) +[2026-06-20 06:22:56,128][starVLA.training.train_starvla][INFO] - Step 2627, Loss: {'train/loss': 0.22257192432880402, 'train/grad_norm_pre_clip': 5.931558132171631, 'timing/data': 0.006553571671247482, 'timing/model': 2.7618684843182564, 'learning_rate/qwen_vl_interface': 5.015597789147184e-06, 'learning_rate/action_model': 5.015597789147184e-05, 'epoch': 4.78}) +[2026-06-20 06:23:07,034][starVLA.training.train_starvla][INFO] - Step 2628, Loss: {'train/loss': 0.2699838876724243, 'train/grad_norm_pre_clip': 5.957004547119141, 'timing/data': 0.0069085992872715, 'timing/model': 2.7816576194018126, 'learning_rate/qwen_vl_interface': 5.012556126461291e-06, 'learning_rate/action_model': 5.012556126461291e-05, 'epoch': 4.78}) +[2026-06-20 06:23:17,763][starVLA.training.train_starvla][INFO] - Step 2629, Loss: {'train/loss': 0.10163120180368423, 'train/grad_norm_pre_clip': 3.5901968479156494, 'timing/data': 0.006503472104668617, 'timing/model': 2.655371382832527, 'learning_rate/qwen_vl_interface': 5.0095145613796044e-06, 'learning_rate/action_model': 5.0095145613796046e-05, 'epoch': 4.79}) +[2026-06-20 06:23:28,527][starVLA.training.train_starvla][INFO] - Step 2630, Loss: {'train/loss': 0.2920598089694977, 'train/grad_norm_pre_clip': 7.113335132598877, 'timing/data': 0.0066219717264175415, 'timing/model': 2.7478492353111506, 'learning_rate/qwen_vl_interface': 5.0064730951524e-06, 'learning_rate/action_model': 5.0064730951524e-05, 'epoch': 4.79}) +[2026-06-20 06:23:39,428][starVLA.training.train_starvla][INFO] - Step 2631, Loss: {'train/loss': 0.24935349822044373, 'train/grad_norm_pre_clip': 7.03803825378418, 'timing/data': 0.007036076858639717, 'timing/model': 2.77436507306993, 'learning_rate/qwen_vl_interface': 5.003431729029905e-06, 'learning_rate/action_model': 5.003431729029905e-05, 'epoch': 4.79}) +[2026-06-20 06:23:50,372][starVLA.training.train_starvla][INFO] - Step 2632, Loss: {'train/loss': 0.29850196838378906, 'train/grad_norm_pre_clip': 8.016339302062988, 'timing/data': 0.006845392286777496, 'timing/model': 2.915572475641966, 'learning_rate/qwen_vl_interface': 5.000390464262314e-06, 'learning_rate/action_model': 5.0003904642623136e-05, 'epoch': 4.79}) +[2026-06-20 06:24:01,238][starVLA.training.train_starvla][INFO] - Step 2633, Loss: {'train/loss': 0.29033711552619934, 'train/grad_norm_pre_clip': 13.801413536071777, 'timing/data': 0.006889421492815018, 'timing/model': 2.783773770555854, 'learning_rate/qwen_vl_interface': 4.997349302099774e-06, 'learning_rate/action_model': 4.997349302099773e-05, 'epoch': 4.79}) +[2026-06-20 06:24:12,127][starVLA.training.train_starvla][INFO] - Step 2634, Loss: {'train/loss': 0.16468016803264618, 'train/grad_norm_pre_clip': 5.753456115722656, 'timing/data': 0.006863569840788841, 'timing/model': 2.760172162204981, 'learning_rate/qwen_vl_interface': 4.994308243792391e-06, 'learning_rate/action_model': 4.994308243792391e-05, 'epoch': 4.79}) +[2026-06-20 06:24:23,026][starVLA.training.train_starvla][INFO] - Step 2635, Loss: {'train/loss': 0.4754055142402649, 'train/grad_norm_pre_clip': 8.699311256408691, 'timing/data': 0.006792118772864342, 'timing/model': 2.7720088195055723, 'learning_rate/qwen_vl_interface': 4.991267290590233e-06, 'learning_rate/action_model': 4.991267290590233e-05, 'epoch': 4.8}) +[2026-06-20 06:24:33,940][starVLA.training.train_starvla][INFO] - Step 2636, Loss: {'train/loss': 0.42082664370536804, 'train/grad_norm_pre_clip': 9.977675437927246, 'timing/data': 0.006981506943702698, 'timing/model': 2.7559142652899027, 'learning_rate/qwen_vl_interface': 4.988226443743317e-06, 'learning_rate/action_model': 4.9882264437433174e-05, 'epoch': 4.8}) +[2026-06-20 06:24:44,841][starVLA.training.train_starvla][INFO] - Step 2637, Loss: {'train/loss': 0.1619795560836792, 'train/grad_norm_pre_clip': 6.401666164398193, 'timing/data': 0.006864190101623535, 'timing/model': 2.7849768064916134, 'learning_rate/qwen_vl_interface': 4.985185704501623e-06, 'learning_rate/action_model': 4.985185704501623e-05, 'epoch': 4.8}) +[2026-06-20 06:24:55,753][starVLA.training.train_starvla][INFO] - Step 2638, Loss: {'train/loss': 0.3620726466178894, 'train/grad_norm_pre_clip': 10.01500415802002, 'timing/data': 0.006949508562684059, 'timing/model': 2.776805592700839, 'learning_rate/qwen_vl_interface': 4.982145074115082e-06, 'learning_rate/action_model': 4.982145074115082e-05, 'epoch': 4.8}) +[2026-06-20 11:51:45,848][starVLA.training.train_starvla][INFO] - VLA Training :: Warming Up +[2026-06-20 11:51:45,861][starVLA.training.train_starvla][INFO] - ✅ Configuration wrapped for access tracking +[2026-06-20 11:51:47,790][starVLA.training.train_starvla][INFO] - Creating VLA Dataset with Mixture `flappy_mixed_latency_train__bridge` +[2026-06-20 11:51:48,344][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Behavior (data_config): ['BEHAVIOR_challenge'] +[2026-06-20 11:51:48,345][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from DOMINO: ['robotwin'] +[2026-06-20 11:51:48,345][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from DOMINO (data_config): [] +[2026-06-20 11:51:48,345][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from DOMINO (data_config): ['domino', 'domino_clean', 'domino_random', 'domino_cotrain'] +[2026-06-20 11:51:48,345][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Franka: ['custom_robot_config', 'demo_sim_franka_delta_joints', 'SO101'] +[2026-06-20 11:51:48,345][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Franka (data_config): [] +[2026-06-20 11:51:48,345][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Franka (data_config): ['custom_dataset', 'custom_dataset_2', 'demo_sim_pick_place', 'SO101_pick'] +[2026-06-20 11:51:48,345][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from LIBERO: ['libero_franka'] +[2026-06-20 11:51:48,345][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from LIBERO (data_config): [] +[2026-06-20 11:51:48,345][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from LIBERO (data_config): ['libero_all', 'libero_goal', 'multi_robot'] +[2026-06-20 11:51:48,345][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from RoboChallenge_table30v2: ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-20 11:51:48,346][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from RoboChallenge_table30v2 (data_config): ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-20 11:51:48,346][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from RoboChallenge_table30v2 (data_config): ['robochallenge_table30v2_shred_paper', 'robochallenge_table30v2_ur5_all', 'robochallenge_table30v2_arx5_all', 'robochallenge_table30v2_dosw1_all'] +[2026-06-20 11:51:48,346][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_365: ['panda_omron_robocasa365'] +[2026-06-20 11:51:48,346][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_365 (data_config): [] +[2026-06-20 11:51:48,346][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_365 (data_config): ['robocasa365_open_drawer_target_human', 'robocasa365_atomic_target_human_all', 'robocasa365_composite_target_human_all', 'robocasa365_target_human_all'] +[2026-06-20 11:51:48,346][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_tabletop: ['fourier_gr1_arms_waist'] +[2026-06-20 11:51:48,346][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_tabletop (data_config): [] +[2026-06-20 11:51:48,346][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_tabletop (data_config): ['fourier_gr1_unified_1000'] +[2026-06-20 11:51:48,346][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robotwin: ['robotwin', 'robotwin50', 'arx_x5'] +[2026-06-20 11:51:48,346][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robotwin (data_config): [] +[2026-06-20 11:51:48,346][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robotwin (data_config): ['robotwin_all', 'robotwin_all_50', 'robotwin', 'robotwin_task1', 'robotwin_task2', 'arx_x5'] +[2026-06-20 11:51:48,347][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from SimplerEnv: ['oxe_droid', 'oxe_bridge', 'oxe_rt1'] +[2026-06-20 11:51:48,347][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from SimplerEnv (data_config): [] +[2026-06-20 11:51:48,347][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from SimplerEnv (data_config): ['bridge', 'bridge_rt_1'] +[2026-06-20 11:51:48,347][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from VLA-Arena: ['vla_arena_franka'] +[2026-06-20 11:51:48,347][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from VLA-Arena (data_config): [] +[2026-06-20 11:51:48,347][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from VLA-Arena (data_config): ['vla_arena_L0_S', 'vla_arena_L0_M', 'vla_arena_L0_L'] +[2026-06-20 11:51:48,347][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from rl_games: ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-20 11:51:48,347][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from rl_games (data_config): ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-20 11:51:48,347][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from rl_games (data_config): ['flappy_train', 'flappy_train__bridge', 'flappy_mixed_latency_train', 'flappy_mixed_latency_train__bridge', 'demon_attack_train', 'demon_attack_train__bridge', 'demon_attack_mixed_latency_train', 'demon_attack_mixed_latency_train__bridge', 'deadly_corridor_train', 'deadly_corridor_train__bridge', 'deadly_corridor_mixed_latency_train', 'deadly_corridor_mixed_latency_train__bridge'] +[2026-06-20 11:51:48,423][starVLA.training.train_starvla][INFO] - Creating VLA Eval Dataset with Mixture `flappy_mixed_latency_train__bridge__val` +[2026-06-20 11:51:48,481][starVLA.training.train_starvla][INFO] - 📝 Full config saved at /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/config.full.yaml +[2026-06-20 11:51:48,489][starVLA.training.train_starvla][INFO] - 📊 Accessed config snapshot saved at /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/config.yaml +[2026-06-20 11:51:48,489][starVLA.training.train_starvla][INFO] - Will resume full training state from checkpoint: /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_2600_state, steps: 2600 +[2026-06-20 11:51:50,128][accelerate.accelerator][INFO] - Loading states from /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_2600_state +[2026-06-20 11:52:01,689][accelerate.checkpointing][INFO] - All model weights loaded successfully +[2026-06-20 11:52:24,579][accelerate.checkpointing][INFO] - All optimizer states loaded successfully +[2026-06-20 11:52:24,579][accelerate.checkpointing][INFO] - All scheduler states loaded successfully +[2026-06-20 11:52:24,579][accelerate.checkpointing][INFO] - All dataloader sampler states loaded successfully +[2026-06-20 11:52:24,581][accelerate.checkpointing][INFO] - All random states loaded successfully +[2026-06-20 11:52:25,727][accelerate.accelerator][INFO] - Loading in 0 custom states +[2026-06-20 11:54:38,340][starVLA.training.train_starvla][INFO] - VLA Training :: Warming Up +[2026-06-20 11:54:38,353][starVLA.training.train_starvla][INFO] - ✅ Configuration wrapped for access tracking +[2026-06-20 11:54:40,050][starVLA.training.train_starvla][INFO] - Creating VLA Dataset with Mixture `flappy_mixed_latency_train__bridge` +[2026-06-20 11:54:40,598][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Behavior (data_config): ['BEHAVIOR_challenge'] +[2026-06-20 11:54:40,598][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from DOMINO: ['robotwin'] +[2026-06-20 11:54:40,598][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from DOMINO (data_config): [] +[2026-06-20 11:54:40,598][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from DOMINO (data_config): ['domino', 'domino_clean', 'domino_random', 'domino_cotrain'] +[2026-06-20 11:54:40,598][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Franka: ['custom_robot_config', 'demo_sim_franka_delta_joints', 'SO101'] +[2026-06-20 11:54:40,598][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Franka (data_config): [] +[2026-06-20 11:54:40,598][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Franka (data_config): ['custom_dataset', 'custom_dataset_2', 'demo_sim_pick_place', 'SO101_pick'] +[2026-06-20 11:54:40,599][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from LIBERO: ['libero_franka'] +[2026-06-20 11:54:40,599][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from LIBERO (data_config): [] +[2026-06-20 11:54:40,599][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from LIBERO (data_config): ['libero_all', 'libero_goal', 'multi_robot'] +[2026-06-20 11:54:40,599][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from RoboChallenge_table30v2: ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-20 11:54:40,599][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from RoboChallenge_table30v2 (data_config): ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-20 11:54:40,599][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from RoboChallenge_table30v2 (data_config): ['robochallenge_table30v2_shred_paper', 'robochallenge_table30v2_ur5_all', 'robochallenge_table30v2_arx5_all', 'robochallenge_table30v2_dosw1_all'] +[2026-06-20 11:54:40,599][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_365: ['panda_omron_robocasa365'] +[2026-06-20 11:54:40,599][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_365 (data_config): [] +[2026-06-20 11:54:40,599][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_365 (data_config): ['robocasa365_open_drawer_target_human', 'robocasa365_atomic_target_human_all', 'robocasa365_composite_target_human_all', 'robocasa365_target_human_all'] +[2026-06-20 11:54:40,599][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_tabletop: ['fourier_gr1_arms_waist'] +[2026-06-20 11:54:40,600][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_tabletop (data_config): [] +[2026-06-20 11:54:40,600][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_tabletop (data_config): ['fourier_gr1_unified_1000'] +[2026-06-20 11:54:40,600][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robotwin: ['robotwin', 'robotwin50', 'arx_x5'] +[2026-06-20 11:54:40,600][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robotwin (data_config): [] +[2026-06-20 11:54:40,600][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robotwin (data_config): ['robotwin_all', 'robotwin_all_50', 'robotwin', 'robotwin_task1', 'robotwin_task2', 'arx_x5'] +[2026-06-20 11:54:40,600][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from SimplerEnv: ['oxe_droid', 'oxe_bridge', 'oxe_rt1'] +[2026-06-20 11:54:40,601][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from SimplerEnv (data_config): [] +[2026-06-20 11:54:40,601][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from SimplerEnv (data_config): ['bridge', 'bridge_rt_1'] +[2026-06-20 11:54:40,601][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from VLA-Arena: ['vla_arena_franka'] +[2026-06-20 11:54:40,601][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from VLA-Arena (data_config): [] +[2026-06-20 11:54:40,601][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from VLA-Arena (data_config): ['vla_arena_L0_S', 'vla_arena_L0_M', 'vla_arena_L0_L'] +[2026-06-20 11:54:40,601][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from rl_games: ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-20 11:54:40,601][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from rl_games (data_config): ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-20 11:54:40,601][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from rl_games (data_config): ['flappy_train', 'flappy_train__bridge', 'flappy_mixed_latency_train', 'flappy_mixed_latency_train__bridge', 'demon_attack_train', 'demon_attack_train__bridge', 'demon_attack_mixed_latency_train', 'demon_attack_mixed_latency_train__bridge', 'deadly_corridor_train', 'deadly_corridor_train__bridge', 'deadly_corridor_mixed_latency_train', 'deadly_corridor_mixed_latency_train__bridge'] +[2026-06-20 11:54:40,675][starVLA.training.train_starvla][INFO] - Creating VLA Eval Dataset with Mixture `flappy_mixed_latency_train__bridge__val` +[2026-06-20 11:54:40,734][starVLA.training.train_starvla][INFO] - 📝 Full config saved at /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/config.full.yaml +[2026-06-20 11:54:40,741][starVLA.training.train_starvla][INFO] - 📊 Accessed config snapshot saved at /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/config.yaml +[2026-06-20 11:54:40,741][starVLA.training.train_starvla][INFO] - Will resume full training state from checkpoint: /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_2600_state, steps: 2600 +[2026-06-20 11:54:49,354][accelerate.accelerator][INFO] - Loading states from /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_2600_state +[2026-06-20 11:54:51,501][accelerate.checkpointing][INFO] - All model weights loaded successfully +[2026-06-20 11:55:03,391][accelerate.checkpointing][INFO] - All optimizer states loaded successfully +[2026-06-20 11:55:03,391][accelerate.checkpointing][INFO] - All scheduler states loaded successfully +[2026-06-20 11:55:03,391][accelerate.checkpointing][INFO] - All dataloader sampler states loaded successfully +[2026-06-20 11:55:03,393][accelerate.checkpointing][INFO] - All random states loaded successfully +[2026-06-20 11:55:04,412][accelerate.accelerator][INFO] - Loading in 0 custom states +[2026-06-20 11:55:05,980][starVLA.training.train_starvla][INFO] - ***** Training Configuration ***** +[2026-06-20 11:55:05,981][starVLA.training.train_starvla][INFO] - Total optimization steps = 5000 +[2026-06-20 11:55:05,981][starVLA.training.train_starvla][INFO] - Per device batch size = 32 +[2026-06-20 11:55:05,981][starVLA.training.train_starvla][INFO] - Gradient accumulation steps = 4 +[2026-06-20 11:55:05,981][starVLA.training.train_starvla][INFO] - Total batch size = 128 +[2026-06-20 11:55:06,491][starVLA.training.train_starvla][INFO] - Latency curriculum phase 3 at step 2600: active_latencies=[3], dataset_size=70316 +[2026-06-20 11:55:31,414][starVLA.training.train_starvla][INFO] - Step 2601, Loss: {'train/loss': 0.3728872835636139, 'train/grad_norm_pre_clip': 8.663826942443848, 'timing/data': 0.021483693039044738, 'timing/model': 0.9766329469857737, 'learning_rate/qwen_vl_interface': 1.0000000000000001e-07, 'learning_rate/action_model': 1.0000000000000002e-06, 'epoch': 4.73}) +[2026-06-20 11:55:35,136][starVLA.training.train_starvla][INFO] - Step 2602, Loss: {'train/loss': 0.3899080753326416, 'train/grad_norm_pre_clip': 8.076006889343262, 'timing/data': 0.0061061769956722856, 'timing/model': 0.9200065089389682, 'learning_rate/qwen_vl_interface': 2.0000000000000002e-07, 'learning_rate/action_model': 2.0000000000000003e-06, 'epoch': 4.74}) +[2026-06-20 11:55:38,789][starVLA.training.train_starvla][INFO] - Step 2603, Loss: {'train/loss': 0.23440055549144745, 'train/grad_norm_pre_clip': 8.546422004699707, 'timing/data': 0.007126935059204698, 'timing/model': 0.9353129339870065, 'learning_rate/qwen_vl_interface': 3.0000000000000004e-07, 'learning_rate/action_model': 3e-06, 'epoch': 4.74}) +[2026-06-20 11:55:42,353][starVLA.training.train_starvla][INFO] - Step 2604, Loss: {'train/loss': 0.42370253801345825, 'train/grad_norm_pre_clip': 9.898092269897461, 'timing/data': 0.006474834983237088, 'timing/model': 0.9237156879389659, 'learning_rate/qwen_vl_interface': 4.0000000000000003e-07, 'learning_rate/action_model': 4.000000000000001e-06, 'epoch': 4.74}) +[2026-06-20 11:55:46,031][starVLA.training.train_starvla][INFO] - Step 2605, Loss: {'train/loss': 0.23064297437667847, 'train/grad_norm_pre_clip': 9.244189262390137, 'timing/data': 0.006603346089832485, 'timing/model': 0.9703926220536232, 'learning_rate/qwen_vl_interface': 5.000000000000001e-07, 'learning_rate/action_model': 5e-06, 'epoch': 4.74}) +[2026-06-20 11:55:49,676][starVLA.training.train_starvla][INFO] - Step 2606, Loss: {'train/loss': 0.3400792181491852, 'train/grad_norm_pre_clip': 10.420202255249023, 'timing/data': 0.006735825911164284, 'timing/model': 0.9238284439779818, 'learning_rate/qwen_vl_interface': 6.000000000000001e-07, 'learning_rate/action_model': 6e-06, 'epoch': 4.74}) +[2026-06-20 11:55:53,309][starVLA.training.train_starvla][INFO] - Step 2607, Loss: {'train/loss': 0.2753751277923584, 'train/grad_norm_pre_clip': 13.369217872619629, 'timing/data': 0.006883693975396454, 'timing/model': 0.93912208895199, 'learning_rate/qwen_vl_interface': 7.000000000000001e-07, 'learning_rate/action_model': 7.000000000000001e-06, 'epoch': 4.75}) +[2026-06-20 11:55:57,091][starVLA.training.train_starvla][INFO] - Step 2608, Loss: {'train/loss': 0.4454250633716583, 'train/grad_norm_pre_clip': 11.241199493408203, 'timing/data': 0.006920855026692152, 'timing/model': 0.972777861985378, 'learning_rate/qwen_vl_interface': 8.000000000000001e-07, 'learning_rate/action_model': 8.000000000000001e-06, 'epoch': 4.75}) +[2026-06-20 11:56:00,769][starVLA.training.train_starvla][INFO] - Step 2609, Loss: {'train/loss': 0.2379763126373291, 'train/grad_norm_pre_clip': 5.6223673820495605, 'timing/data': 0.005792067036963999, 'timing/model': 0.9280517800943926, 'learning_rate/qwen_vl_interface': 9.000000000000001e-07, 'learning_rate/action_model': 9e-06, 'epoch': 4.75}) +[2026-06-20 11:56:04,324][starVLA.training.train_starvla][INFO] - Step 2610, Loss: {'train/loss': 0.43693652749061584, 'train/grad_norm_pre_clip': 7.790552139282227, 'timing/data': 0.005796027020551264, 'timing/model': 0.9303883239626884, 'learning_rate/qwen_vl_interface': 1.0000000000000002e-06, 'learning_rate/action_model': 1e-05, 'epoch': 4.75}) +[2026-06-20 11:56:07,898][starVLA.training.train_starvla][INFO] - Step 2611, Loss: {'train/loss': 0.2927325665950775, 'train/grad_norm_pre_clip': 6.714746952056885, 'timing/data': 0.005702267051674426, 'timing/model': 0.9325813360046595, 'learning_rate/qwen_vl_interface': 1.1e-06, 'learning_rate/action_model': 1.1000000000000001e-05, 'epoch': 4.75}) +[2026-06-20 11:56:11,493][starVLA.training.train_starvla][INFO] - Step 2612, Loss: {'train/loss': 0.34054404497146606, 'train/grad_norm_pre_clip': 9.017688751220703, 'timing/data': 0.007562312996014953, 'timing/model': 0.9424068629741669, 'learning_rate/qwen_vl_interface': 1.2000000000000002e-06, 'learning_rate/action_model': 1.2e-05, 'epoch': 4.75}) +[2026-06-20 11:56:15,074][starVLA.training.train_starvla][INFO] - Step 2613, Loss: {'train/loss': 0.3139592111110687, 'train/grad_norm_pre_clip': 6.4429779052734375, 'timing/data': 0.005557497963309288, 'timing/model': 0.934285327908583, 'learning_rate/qwen_vl_interface': 1.3e-06, 'learning_rate/action_model': 1.3000000000000001e-05, 'epoch': 4.76}) +[2026-06-20 11:56:18,669][starVLA.training.train_starvla][INFO] - Step 2614, Loss: {'train/loss': 0.31499239802360535, 'train/grad_norm_pre_clip': 6.96127986907959, 'timing/data': 0.005562537000514567, 'timing/model': 0.9362638629972935, 'learning_rate/qwen_vl_interface': 1.4000000000000001e-06, 'learning_rate/action_model': 1.4000000000000001e-05, 'epoch': 4.76}) +[2026-06-20 11:56:22,239][starVLA.training.train_starvla][INFO] - Step 2615, Loss: {'train/loss': 0.5251842141151428, 'train/grad_norm_pre_clip': 10.597312927246094, 'timing/data': 0.0056102980161085725, 'timing/model': 0.9321257210103795, 'learning_rate/qwen_vl_interface': 1.5e-06, 'learning_rate/action_model': 1.5e-05, 'epoch': 4.76}) +[2026-06-20 11:56:25,851][starVLA.training.train_starvla][INFO] - Step 2616, Loss: {'train/loss': 0.27008286118507385, 'train/grad_norm_pre_clip': 10.791695594787598, 'timing/data': 0.006789725041016936, 'timing/model': 0.9524349140701815, 'learning_rate/qwen_vl_interface': 1.6000000000000001e-06, 'learning_rate/action_model': 1.6000000000000003e-05, 'epoch': 4.76}) +[2026-06-20 11:56:29,548][starVLA.training.train_starvla][INFO] - Step 2617, Loss: {'train/loss': 0.4836253225803375, 'train/grad_norm_pre_clip': 12.611417770385742, 'timing/data': 0.005731885903514922, 'timing/model': 0.9587840590393171, 'learning_rate/qwen_vl_interface': 1.7000000000000002e-06, 'learning_rate/action_model': 1.7000000000000003e-05, 'epoch': 4.76}) +[2026-06-20 11:56:33,281][starVLA.training.train_starvla][INFO] - Step 2618, Loss: {'train/loss': 0.22861723601818085, 'train/grad_norm_pre_clip': 7.656860828399658, 'timing/data': 0.0057412669993937016, 'timing/model': 0.9742401520488784, 'learning_rate/qwen_vl_interface': 1.8000000000000001e-06, 'learning_rate/action_model': 1.8e-05, 'epoch': 4.77}) +[2026-06-20 11:56:36,973][starVLA.training.train_starvla][INFO] - Step 2619, Loss: {'train/loss': 0.23733331263065338, 'train/grad_norm_pre_clip': 4.985769271850586, 'timing/data': 0.005662197945639491, 'timing/model': 0.9603004519594833, 'learning_rate/qwen_vl_interface': 1.9000000000000002e-06, 'learning_rate/action_model': 1.9e-05, 'epoch': 4.77}) +[2026-06-20 11:56:40,698][starVLA.training.train_starvla][INFO] - Step 2620, Loss: {'train/loss': 0.5004906058311462, 'train/grad_norm_pre_clip': 13.484793663024902, 'timing/data': 0.007588682929053903, 'timing/model': 0.9874203390209004, 'learning_rate/qwen_vl_interface': 2.0000000000000003e-06, 'learning_rate/action_model': 2e-05, 'epoch': 4.77}) +[2026-06-20 11:56:44,407][starVLA.training.train_starvla][INFO] - Step 2621, Loss: {'train/loss': 0.28170618414878845, 'train/grad_norm_pre_clip': 9.583516120910645, 'timing/data': 0.005650986917316914, 'timing/model': 0.9749398750718683, 'learning_rate/qwen_vl_interface': 2.1000000000000002e-06, 'learning_rate/action_model': 2.1e-05, 'epoch': 4.77}) +[2026-06-20 11:56:48,112][starVLA.training.train_starvla][INFO] - Step 2622, Loss: {'train/loss': 0.4446536600589752, 'train/grad_norm_pre_clip': 10.981334686279297, 'timing/data': 0.00595458596944809, 'timing/model': 0.9624244930455461, 'learning_rate/qwen_vl_interface': 2.2e-06, 'learning_rate/action_model': 2.2000000000000003e-05, 'epoch': 4.77}) +[2026-06-20 11:56:51,845][starVLA.training.train_starvla][INFO] - Step 2623, Loss: {'train/loss': 0.35182926058769226, 'train/grad_norm_pre_clip': 7.361926078796387, 'timing/data': 0.0057876770151779056, 'timing/model': 0.9720442879479378, 'learning_rate/qwen_vl_interface': 2.3000000000000004e-06, 'learning_rate/action_model': 2.3000000000000003e-05, 'epoch': 4.77}) +[2026-06-20 11:56:55,506][starVLA.training.train_starvla][INFO] - Step 2624, Loss: {'train/loss': 0.4390282928943634, 'train/grad_norm_pre_clip': 9.068682670593262, 'timing/data': 0.007546083070337772, 'timing/model': 0.9482595300069079, 'learning_rate/qwen_vl_interface': 2.4000000000000003e-06, 'learning_rate/action_model': 2.4e-05, 'epoch': 4.78}) +[2026-06-20 11:56:59,080][starVLA.training.train_starvla][INFO] - Step 2625, Loss: {'train/loss': 0.4582104980945587, 'train/grad_norm_pre_clip': 9.403414726257324, 'timing/data': 0.0005910278996452689, 'timing/model': 0.9343156209215522, 'learning_rate/qwen_vl_interface': 2.5e-06, 'learning_rate/action_model': 2.5e-05, 'epoch': 4.78}) +[2026-06-20 11:57:02,679][starVLA.training.train_starvla][INFO] - Step 2626, Loss: {'train/loss': 0.35566797852516174, 'train/grad_norm_pre_clip': 8.803519248962402, 'timing/data': 0.0005259880563244224, 'timing/model': 0.963380063069053, 'learning_rate/qwen_vl_interface': 2.6e-06, 'learning_rate/action_model': 2.6000000000000002e-05, 'epoch': 4.78}) +[2026-06-20 11:57:06,354][starVLA.training.train_starvla][INFO] - Step 2627, Loss: {'train/loss': 0.2912960350513458, 'train/grad_norm_pre_clip': 6.449276924133301, 'timing/data': 0.0005129589699208736, 'timing/model': 0.96109544695355, 'learning_rate/qwen_vl_interface': 2.7000000000000004e-06, 'learning_rate/action_model': 2.7000000000000002e-05, 'epoch': 4.78}) +[2026-06-20 11:57:09,977][starVLA.training.train_starvla][INFO] - Step 2628, Loss: {'train/loss': 0.3523256778717041, 'train/grad_norm_pre_clip': 6.47239875793457, 'timing/data': 0.0005451380275189877, 'timing/model': 0.9339514169842005, 'learning_rate/qwen_vl_interface': 2.8000000000000003e-06, 'learning_rate/action_model': 2.8000000000000003e-05, 'epoch': 4.78}) +[2026-06-20 11:57:13,565][starVLA.training.train_starvla][INFO] - Step 2629, Loss: {'train/loss': 0.21771463751792908, 'train/grad_norm_pre_clip': 6.195363521575928, 'timing/data': 0.0005274990107864141, 'timing/model': 0.9407559999963269, 'learning_rate/qwen_vl_interface': 2.9e-06, 'learning_rate/action_model': 2.9e-05, 'epoch': 4.79}) +[2026-06-20 11:57:17,163][starVLA.training.train_starvla][INFO] - Step 2630, Loss: {'train/loss': 0.2185688018798828, 'train/grad_norm_pre_clip': 8.442734718322754, 'timing/data': 0.0005954989464953542, 'timing/model': 0.9415608170675114, 'learning_rate/qwen_vl_interface': 3e-06, 'learning_rate/action_model': 3e-05, 'epoch': 4.79}) +[2026-06-20 11:57:20,728][starVLA.training.train_starvla][INFO] - Step 2631, Loss: {'train/loss': 0.44052818417549133, 'train/grad_norm_pre_clip': 11.706724166870117, 'timing/data': 0.0005335089517757297, 'timing/model': 0.9332029040670022, 'learning_rate/qwen_vl_interface': 3.1000000000000004e-06, 'learning_rate/action_model': 3.1e-05, 'epoch': 4.79}) +[2026-06-20 11:57:24,302][starVLA.training.train_starvla][INFO] - Step 2632, Loss: {'train/loss': 0.22594322264194489, 'train/grad_norm_pre_clip': 5.622068405151367, 'timing/data': 0.0005300280172377825, 'timing/model': 0.9365283250808716, 'learning_rate/qwen_vl_interface': 3.2000000000000003e-06, 'learning_rate/action_model': 3.2000000000000005e-05, 'epoch': 4.79}) +[2026-06-20 11:57:27,878][starVLA.training.train_starvla][INFO] - Step 2633, Loss: {'train/loss': 0.24065625667572021, 'train/grad_norm_pre_clip': 15.09248161315918, 'timing/data': 0.0005128179909661412, 'timing/model': 0.9336765410844237, 'learning_rate/qwen_vl_interface': 3.3000000000000006e-06, 'learning_rate/action_model': 3.3e-05, 'epoch': 4.79}) +[2026-06-20 11:57:31,471][starVLA.training.train_starvla][INFO] - Step 2634, Loss: {'train/loss': 0.3759472370147705, 'train/grad_norm_pre_clip': 5.888494491577148, 'timing/data': 0.0005218989681452513, 'timing/model': 0.9333851200062782, 'learning_rate/qwen_vl_interface': 3.4000000000000005e-06, 'learning_rate/action_model': 3.4000000000000007e-05, 'epoch': 4.79}) +[2026-06-20 11:57:35,251][starVLA.training.train_starvla][INFO] - Step 2635, Loss: {'train/loss': 0.19558534026145935, 'train/grad_norm_pre_clip': 3.9824912548065186, 'timing/data': 0.0005323289660736918, 'timing/model': 0.9363514920696616, 'learning_rate/qwen_vl_interface': 3.5e-06, 'learning_rate/action_model': 3.5e-05, 'epoch': 4.8}) +[2026-06-20 11:57:38,855][starVLA.training.train_starvla][INFO] - Step 2636, Loss: {'train/loss': 0.492594450712204, 'train/grad_norm_pre_clip': 7.043511390686035, 'timing/data': 0.0005171890370547771, 'timing/model': 0.932405699044466, 'learning_rate/qwen_vl_interface': 3.6000000000000003e-06, 'learning_rate/action_model': 3.6e-05, 'epoch': 4.8}) +[2026-06-20 11:57:42,446][starVLA.training.train_starvla][INFO] - Step 2637, Loss: {'train/loss': 0.2112005352973938, 'train/grad_norm_pre_clip': 10.71523666381836, 'timing/data': 0.0005247190129011869, 'timing/model': 0.9314627799903974, 'learning_rate/qwen_vl_interface': 3.7e-06, 'learning_rate/action_model': 3.7e-05, 'epoch': 4.8}) +[2026-06-20 11:57:46,111][starVLA.training.train_starvla][INFO] - Step 2638, Loss: {'train/loss': 0.3672385811805725, 'train/grad_norm_pre_clip': 7.342958450317383, 'timing/data': 0.0005366289988160133, 'timing/model': 0.9470225940458477, 'learning_rate/qwen_vl_interface': 3.8000000000000005e-06, 'learning_rate/action_model': 3.8e-05, 'epoch': 4.8}) +[2026-06-20 11:57:49,953][starVLA.training.train_starvla][INFO] - Step 2639, Loss: {'train/loss': 0.2658303380012512, 'train/grad_norm_pre_clip': 9.348015785217285, 'timing/data': 0.0005313190631568432, 'timing/model': 1.0001118018990383, 'learning_rate/qwen_vl_interface': 3.900000000000001e-06, 'learning_rate/action_model': 3.9000000000000006e-05, 'epoch': 4.8}) +[2026-06-20 11:57:53,703][starVLA.training.train_starvla][INFO] - Step 2640, Loss: {'train/loss': 0.42109617590904236, 'train/grad_norm_pre_clip': 7.633195877075195, 'timing/data': 0.0005186980124562979, 'timing/model': 0.9839471959276125, 'learning_rate/qwen_vl_interface': 4.000000000000001e-06, 'learning_rate/action_model': 4e-05, 'epoch': 4.81}) +[2026-06-20 11:57:57,451][starVLA.training.train_starvla][INFO] - Step 2641, Loss: {'train/loss': 0.2628897428512573, 'train/grad_norm_pre_clip': 6.455465316772461, 'timing/data': 0.0005186189664527774, 'timing/model': 0.995736627955921, 'learning_rate/qwen_vl_interface': 4.1e-06, 'learning_rate/action_model': 4.1e-05, 'epoch': 4.81}) +[2026-06-20 11:58:01,190][starVLA.training.train_starvla][INFO] - Step 2642, Loss: {'train/loss': 0.22716295719146729, 'train/grad_norm_pre_clip': 17.479324340820312, 'timing/data': 0.0005271489499136806, 'timing/model': 0.9287187900627032, 'learning_rate/qwen_vl_interface': 4.2000000000000004e-06, 'learning_rate/action_model': 4.2e-05, 'epoch': 4.81}) +[2026-06-20 11:58:04,884][starVLA.training.train_starvla][INFO] - Step 2643, Loss: {'train/loss': 0.35273152589797974, 'train/grad_norm_pre_clip': 7.5311079025268555, 'timing/data': 0.0005170579534024, 'timing/model': 0.9342242259299383, 'learning_rate/qwen_vl_interface': 4.3e-06, 'learning_rate/action_model': 4.3e-05, 'epoch': 4.81}) +[2026-06-20 11:58:08,568][starVLA.training.train_starvla][INFO] - Step 2644, Loss: {'train/loss': 0.3176897168159485, 'train/grad_norm_pre_clip': 16.85755729675293, 'timing/data': 0.0005230490351095796, 'timing/model': 0.9357888519298285, 'learning_rate/qwen_vl_interface': 4.4e-06, 'learning_rate/action_model': 4.4000000000000006e-05, 'epoch': 4.81}) +[2026-06-20 11:58:12,340][starVLA.training.train_starvla][INFO] - Step 2645, Loss: {'train/loss': 0.30770939588546753, 'train/grad_norm_pre_clip': 11.50604248046875, 'timing/data': 0.0005205090856179595, 'timing/model': 1.000458282069303, 'learning_rate/qwen_vl_interface': 4.5e-06, 'learning_rate/action_model': 4.5e-05, 'epoch': 4.81}) +[2026-06-20 11:58:16,029][starVLA.training.train_starvla][INFO] - Step 2646, Loss: {'train/loss': 0.3949663043022156, 'train/grad_norm_pre_clip': 10.763465881347656, 'timing/data': 0.0005302589852362871, 'timing/model': 0.9335894329706207, 'learning_rate/qwen_vl_interface': 4.600000000000001e-06, 'learning_rate/action_model': 4.600000000000001e-05, 'epoch': 4.82}) +[2026-06-20 11:58:19,623][starVLA.training.train_starvla][INFO] - Step 2647, Loss: {'train/loss': 0.18487249314785004, 'train/grad_norm_pre_clip': 4.318317890167236, 'timing/data': 0.0005253989947959781, 'timing/model': 0.9448841359699145, 'learning_rate/qwen_vl_interface': 4.7e-06, 'learning_rate/action_model': 4.7e-05, 'epoch': 4.82}) +[2026-06-20 11:58:23,193][starVLA.training.train_starvla][INFO] - Step 2648, Loss: {'train/loss': 0.1683511734008789, 'train/grad_norm_pre_clip': 9.557085037231445, 'timing/data': 0.0005261789774522185, 'timing/model': 0.9320558439940214, 'learning_rate/qwen_vl_interface': 4.800000000000001e-06, 'learning_rate/action_model': 4.8e-05, 'epoch': 4.82}) +[2026-06-20 11:58:26,802][starVLA.training.train_starvla][INFO] - Step 2649, Loss: {'train/loss': 0.26073789596557617, 'train/grad_norm_pre_clip': 6.576919078826904, 'timing/data': 0.0005173190729692578, 'timing/model': 0.9340370990103111, 'learning_rate/qwen_vl_interface': 4.9000000000000005e-06, 'learning_rate/action_model': 4.9e-05, 'epoch': 4.82}) +[2026-06-20 11:58:30,365][starVLA.training.train_starvla][INFO] - Step 2650, Loss: {'train/loss': 0.22445148229599, 'train/grad_norm_pre_clip': 9.350979804992676, 'timing/data': 0.0005265789804980159, 'timing/model': 0.9324354099808261, 'learning_rate/qwen_vl_interface': 5e-06, 'learning_rate/action_model': 5e-05, 'epoch': 4.82}) +[2026-06-20 11:58:33,941][starVLA.training.train_starvla][INFO] - Step 2651, Loss: {'train/loss': 0.19667549431324005, 'train/grad_norm_pre_clip': 5.540965557098389, 'timing/data': 0.0005472189513966441, 'timing/model': 0.9334354870952666, 'learning_rate/qwen_vl_interface': 5.1e-06, 'learning_rate/action_model': 5.1000000000000006e-05, 'epoch': 4.83}) +[2026-06-20 11:58:37,518][starVLA.training.train_starvla][INFO] - Step 2652, Loss: {'train/loss': 0.24227921664714813, 'train/grad_norm_pre_clip': 9.926383018493652, 'timing/data': 0.0005342880031093955, 'timing/model': 0.9430364130530506, 'learning_rate/qwen_vl_interface': 5.2e-06, 'learning_rate/action_model': 5.2000000000000004e-05, 'epoch': 4.83}) +[2026-06-20 11:58:41,090][starVLA.training.train_starvla][INFO] - Step 2653, Loss: {'train/loss': 0.30720198154449463, 'train/grad_norm_pre_clip': 8.855467796325684, 'timing/data': 0.0005183590110391378, 'timing/model': 0.9306625319877639, 'learning_rate/qwen_vl_interface': 5.300000000000001e-06, 'learning_rate/action_model': 5.300000000000001e-05, 'epoch': 4.83}) +[2026-06-20 11:58:44,664][starVLA.training.train_starvla][INFO] - Step 2654, Loss: {'train/loss': 0.2576315402984619, 'train/grad_norm_pre_clip': 5.940279483795166, 'timing/data': 0.0007189579773694277, 'timing/model': 0.9424821230350062, 'learning_rate/qwen_vl_interface': 5.400000000000001e-06, 'learning_rate/action_model': 5.4000000000000005e-05, 'epoch': 4.83}) +[2026-06-20 11:58:48,234][starVLA.training.train_starvla][INFO] - Step 2655, Loss: {'train/loss': 0.2971765398979187, 'train/grad_norm_pre_clip': 6.384650707244873, 'timing/data': 0.0005311189452186227, 'timing/model': 0.9299427709775046, 'learning_rate/qwen_vl_interface': 5.500000000000001e-06, 'learning_rate/action_model': 5.500000000000001e-05, 'epoch': 4.83}) +[2026-06-20 11:58:51,801][starVLA.training.train_starvla][INFO] - Step 2656, Loss: {'train/loss': 0.3210330903530121, 'train/grad_norm_pre_clip': 6.95345401763916, 'timing/data': 0.0005271990085020661, 'timing/model': 0.9319455140503123, 'learning_rate/qwen_vl_interface': 5.600000000000001e-06, 'learning_rate/action_model': 5.6000000000000006e-05, 'epoch': 4.83}) +[2026-06-20 11:58:55,382][starVLA.training.train_starvla][INFO] - Step 2657, Loss: {'train/loss': 0.38684362173080444, 'train/grad_norm_pre_clip': 9.21744441986084, 'timing/data': 0.0005215290002524853, 'timing/model': 0.9324378530727699, 'learning_rate/qwen_vl_interface': 5.7e-06, 'learning_rate/action_model': 5.6999999999999996e-05, 'epoch': 4.84}) +[2026-06-20 11:58:58,959][starVLA.training.train_starvla][INFO] - Step 2658, Loss: {'train/loss': 0.317688912153244, 'train/grad_norm_pre_clip': 7.6874680519104, 'timing/data': 0.0005337380571290851, 'timing/model': 0.9328520209528506, 'learning_rate/qwen_vl_interface': 5.8e-06, 'learning_rate/action_model': 5.8e-05, 'epoch': 4.84}) +[2026-06-20 11:59:02,526][starVLA.training.train_starvla][INFO] - Step 2659, Loss: {'train/loss': 0.22228805720806122, 'train/grad_norm_pre_clip': 9.507344245910645, 'timing/data': 0.0005283490754663944, 'timing/model': 0.9325290210545063, 'learning_rate/qwen_vl_interface': 5.9e-06, 'learning_rate/action_model': 5.9e-05, 'epoch': 4.84}) +[2026-06-20 11:59:06,115][starVLA.training.train_starvla][INFO] - Step 2660, Loss: {'train/loss': 0.38314953446388245, 'train/grad_norm_pre_clip': 14.038697242736816, 'timing/data': 0.0005304189398884773, 'timing/model': 0.9337261660257354, 'learning_rate/qwen_vl_interface': 6e-06, 'learning_rate/action_model': 6e-05, 'epoch': 4.84}) +[2026-06-20 11:59:09,724][starVLA.training.train_starvla][INFO] - Step 2661, Loss: {'train/loss': 0.11407235264778137, 'train/grad_norm_pre_clip': 4.350198268890381, 'timing/data': 0.0005181389860808849, 'timing/model': 0.9461638770299032, 'learning_rate/qwen_vl_interface': 6.1e-06, 'learning_rate/action_model': 6.1e-05, 'epoch': 4.84}) +[2026-06-20 11:59:13,412][starVLA.training.train_starvla][INFO] - Step 2662, Loss: {'train/loss': 0.2132267951965332, 'train/grad_norm_pre_clip': 6.925409317016602, 'timing/data': 0.0005273390561342239, 'timing/model': 0.9345661520492285, 'learning_rate/qwen_vl_interface': 6.200000000000001e-06, 'learning_rate/action_model': 6.2e-05, 'epoch': 4.85}) +[2026-06-20 11:59:16,998][starVLA.training.train_starvla][INFO] - Step 2663, Loss: {'train/loss': 0.1449657827615738, 'train/grad_norm_pre_clip': 5.6122002601623535, 'timing/data': 0.0005182090681046247, 'timing/model': 0.9343111210037023, 'learning_rate/qwen_vl_interface': 6.300000000000001e-06, 'learning_rate/action_model': 6.3e-05, 'epoch': 4.85}) +[2026-06-20 11:59:20,573][starVLA.training.train_starvla][INFO] - Step 2664, Loss: {'train/loss': 0.08252395689487457, 'train/grad_norm_pre_clip': 5.258727550506592, 'timing/data': 0.0008446179563179612, 'timing/model': 0.9301772790495306, 'learning_rate/qwen_vl_interface': 6.4000000000000006e-06, 'learning_rate/action_model': 6.400000000000001e-05, 'epoch': 4.85}) +[2026-06-20 11:59:24,245][starVLA.training.train_starvla][INFO] - Step 2665, Loss: {'train/loss': 0.22147226333618164, 'train/grad_norm_pre_clip': 7.659328937530518, 'timing/data': 0.0005638189613819122, 'timing/model': 0.9307639779290184, 'learning_rate/qwen_vl_interface': 6.5000000000000004e-06, 'learning_rate/action_model': 6.500000000000001e-05, 'epoch': 4.85}) +[2026-06-20 11:59:27,980][starVLA.training.train_starvla][INFO] - Step 2666, Loss: {'train/loss': 0.3017566204071045, 'train/grad_norm_pre_clip': 15.236042022705078, 'timing/data': 0.0009138580644503236, 'timing/model': 0.9290819509187713, 'learning_rate/qwen_vl_interface': 6.600000000000001e-06, 'learning_rate/action_model': 6.6e-05, 'epoch': 4.85}) +[2026-06-20 11:59:31,681][starVLA.training.train_starvla][INFO] - Step 2667, Loss: {'train/loss': 0.2503458857536316, 'train/grad_norm_pre_clip': 8.525744438171387, 'timing/data': 0.0007093590684235096, 'timing/model': 0.9388445060467348, 'learning_rate/qwen_vl_interface': 6.700000000000001e-06, 'learning_rate/action_model': 6.7e-05, 'epoch': 4.85}) +[2026-06-20 11:59:35,382][starVLA.training.train_starvla][INFO] - Step 2668, Loss: {'train/loss': 0.19131022691726685, 'train/grad_norm_pre_clip': 8.542712211608887, 'timing/data': 0.0005407289136201143, 'timing/model': 0.9918339629657567, 'learning_rate/qwen_vl_interface': 6.800000000000001e-06, 'learning_rate/action_model': 6.800000000000001e-05, 'epoch': 4.86}) +[2026-06-20 11:59:39,073][starVLA.training.train_starvla][INFO] - Step 2669, Loss: {'train/loss': 0.21744874119758606, 'train/grad_norm_pre_clip': 14.697195053100586, 'timing/data': 0.0005246889777481556, 'timing/model': 0.9837764099938795, 'learning_rate/qwen_vl_interface': 6.9e-06, 'learning_rate/action_model': 6.9e-05, 'epoch': 4.86}) +[2026-06-20 11:59:42,697][starVLA.training.train_starvla][INFO] - Step 2670, Loss: {'train/loss': 0.22613829374313354, 'train/grad_norm_pre_clip': 8.770519256591797, 'timing/data': 0.0005223890766501427, 'timing/model': 0.9315716909477487, 'learning_rate/qwen_vl_interface': 7e-06, 'learning_rate/action_model': 7e-05, 'epoch': 4.86}) +[2026-06-20 11:59:46,262][starVLA.training.train_starvla][INFO] - Step 2671, Loss: {'train/loss': 0.17380847036838531, 'train/grad_norm_pre_clip': 5.2317023277282715, 'timing/data': 0.0005219290032982826, 'timing/model': 0.9336383929476142, 'learning_rate/qwen_vl_interface': 7.100000000000001e-06, 'learning_rate/action_model': 7.1e-05, 'epoch': 4.86}) +[2026-06-20 11:59:49,957][starVLA.training.train_starvla][INFO] - Step 2672, Loss: {'train/loss': 0.14956100285053253, 'train/grad_norm_pre_clip': 11.998397827148438, 'timing/data': 0.0005243091145530343, 'timing/model': 0.9661989680025727, 'learning_rate/qwen_vl_interface': 7.2000000000000005e-06, 'learning_rate/action_model': 7.2e-05, 'epoch': 4.86}) +[2026-06-20 11:59:53,583][starVLA.training.train_starvla][INFO] - Step 2673, Loss: {'train/loss': 0.1346888393163681, 'train/grad_norm_pre_clip': 8.078850746154785, 'timing/data': 0.0006366379093378782, 'timing/model': 0.9316510659409687, 'learning_rate/qwen_vl_interface': 7.3e-06, 'learning_rate/action_model': 7.3e-05, 'epoch': 4.87}) +[2026-06-20 11:59:57,141][starVLA.training.train_starvla][INFO] - Step 2674, Loss: {'train/loss': 0.36074376106262207, 'train/grad_norm_pre_clip': 12.830755233764648, 'timing/data': 0.0005294890142977238, 'timing/model': 0.9328570640645921, 'learning_rate/qwen_vl_interface': 7.4e-06, 'learning_rate/action_model': 7.4e-05, 'epoch': 4.87}) +[2026-06-20 12:00:00,765][starVLA.training.train_starvla][INFO] - Step 2675, Loss: {'train/loss': 0.4838016927242279, 'train/grad_norm_pre_clip': 11.788893699645996, 'timing/data': 0.0005405290285125375, 'timing/model': 0.9583394229412079, 'learning_rate/qwen_vl_interface': 7.500000000000001e-06, 'learning_rate/action_model': 7.500000000000001e-05, 'epoch': 4.87}) +[2026-06-20 12:00:04,448][starVLA.training.train_starvla][INFO] - Step 2676, Loss: {'train/loss': 0.3605077564716339, 'train/grad_norm_pre_clip': 16.283653259277344, 'timing/data': 0.0005248680245131254, 'timing/model': 0.966671462985687, 'learning_rate/qwen_vl_interface': 7.600000000000001e-06, 'learning_rate/action_model': 7.6e-05, 'epoch': 4.87}) +[2026-06-20 12:00:08,386][starVLA.training.train_starvla][INFO] - Step 2677, Loss: {'train/loss': 0.2881101071834564, 'train/grad_norm_pre_clip': 19.6063232421875, 'timing/data': 0.0010385880013927817, 'timing/model': 1.0595178860239685, 'learning_rate/qwen_vl_interface': 7.7e-06, 'learning_rate/action_model': 7.7e-05, 'epoch': 4.87}) +[2026-06-20 12:00:12,461][starVLA.training.train_starvla][INFO] - Step 2678, Loss: {'train/loss': 0.1882663518190384, 'train/grad_norm_pre_clip': 10.115285873413086, 'timing/data': 0.0011695569846779108, 'timing/model': 1.0636039560195059, 'learning_rate/qwen_vl_interface': 7.800000000000002e-06, 'learning_rate/action_model': 7.800000000000001e-05, 'epoch': 4.87}) +[2026-06-20 12:00:16,533][starVLA.training.train_starvla][INFO] - Step 2679, Loss: {'train/loss': 0.17439506947994232, 'train/grad_norm_pre_clip': 12.452128410339355, 'timing/data': 0.001110267941839993, 'timing/model': 1.0630841549718753, 'learning_rate/qwen_vl_interface': 7.9e-06, 'learning_rate/action_model': 7.900000000000001e-05, 'epoch': 4.88}) +[2026-06-20 12:00:20,501][starVLA.training.train_starvla][INFO] - Step 2680, Loss: {'train/loss': 0.24025477468967438, 'train/grad_norm_pre_clip': 32.37049865722656, 'timing/data': 0.001350577105768025, 'timing/model': 0.9586804859573022, 'learning_rate/qwen_vl_interface': 8.000000000000001e-06, 'learning_rate/action_model': 8e-05, 'epoch': 4.88}) +[2026-06-20 12:00:24,230][starVLA.training.train_starvla][INFO] - Step 2681, Loss: {'train/loss': 0.4197632968425751, 'train/grad_norm_pre_clip': 18.596372604370117, 'timing/data': 0.0005274489521980286, 'timing/model': 0.9675123339984566, 'learning_rate/qwen_vl_interface': 8.1e-06, 'learning_rate/action_model': 8.1e-05, 'epoch': 4.88}) +[2026-06-20 12:00:27,913][starVLA.training.train_starvla][INFO] - Step 2682, Loss: {'train/loss': 0.5178141593933105, 'train/grad_norm_pre_clip': 17.936756134033203, 'timing/data': 0.0005353880114853382, 'timing/model': 0.9625493749044836, 'learning_rate/qwen_vl_interface': 8.2e-06, 'learning_rate/action_model': 8.2e-05, 'epoch': 4.88}) +[2026-06-20 12:00:31,624][starVLA.training.train_starvla][INFO] - Step 2683, Loss: {'train/loss': 0.34104636311531067, 'train/grad_norm_pre_clip': 17.832014083862305, 'timing/data': 0.0005211880197748542, 'timing/model': 0.9781508280429989, 'learning_rate/qwen_vl_interface': 8.3e-06, 'learning_rate/action_model': 8.3e-05, 'epoch': 4.88}) +[2026-06-20 12:00:35,323][starVLA.training.train_starvla][INFO] - Step 2684, Loss: {'train/loss': 0.18485985696315765, 'train/grad_norm_pre_clip': 9.23586368560791, 'timing/data': 0.0005261389305815101, 'timing/model': 0.9665005139540881, 'learning_rate/qwen_vl_interface': 8.400000000000001e-06, 'learning_rate/action_model': 8.4e-05, 'epoch': 4.89}) +[2026-06-20 12:00:39,044][starVLA.training.train_starvla][INFO] - Step 2685, Loss: {'train/loss': 0.36908259987831116, 'train/grad_norm_pre_clip': 13.117729187011719, 'timing/data': 0.0005349690327420831, 'timing/model': 0.9575774039840326, 'learning_rate/qwen_vl_interface': 8.5e-06, 'learning_rate/action_model': 8.5e-05, 'epoch': 4.89}) +[2026-06-20 12:00:42,790][starVLA.training.train_starvla][INFO] - Step 2686, Loss: {'train/loss': 0.2898798882961273, 'train/grad_norm_pre_clip': 12.52411937713623, 'timing/data': 0.000533999060280621, 'timing/model': 0.9875930319540203, 'learning_rate/qwen_vl_interface': 8.6e-06, 'learning_rate/action_model': 8.6e-05, 'epoch': 4.89}) +[2026-06-20 12:00:46,504][starVLA.training.train_starvla][INFO] - Step 2687, Loss: {'train/loss': 0.3058619201183319, 'train/grad_norm_pre_clip': 14.221665382385254, 'timing/data': 0.0005250689573585987, 'timing/model': 0.9627290801145136, 'learning_rate/qwen_vl_interface': 8.700000000000001e-06, 'learning_rate/action_model': 8.7e-05, 'epoch': 4.89}) +[2026-06-20 12:00:50,205][starVLA.training.train_starvla][INFO] - Step 2688, Loss: {'train/loss': 0.3266960680484772, 'train/grad_norm_pre_clip': 7.599016189575195, 'timing/data': 0.0005143689922988415, 'timing/model': 0.962772389058955, 'learning_rate/qwen_vl_interface': 8.8e-06, 'learning_rate/action_model': 8.800000000000001e-05, 'epoch': 4.89}) +[2026-06-20 12:00:54,166][starVLA.training.train_starvla][INFO] - Step 2689, Loss: {'train/loss': 0.18073229491710663, 'train/grad_norm_pre_clip': 9.841980934143066, 'timing/data': 0.0005203790497034788, 'timing/model': 0.9311221910174936, 'learning_rate/qwen_vl_interface': 8.900000000000001e-06, 'learning_rate/action_model': 8.900000000000001e-05, 'epoch': 4.89}) +[2026-06-20 12:00:57,804][starVLA.training.train_starvla][INFO] - Step 2690, Loss: {'train/loss': 0.22007225453853607, 'train/grad_norm_pre_clip': 8.006376266479492, 'timing/data': 0.0005209980299696326, 'timing/model': 0.9291059960378334, 'learning_rate/qwen_vl_interface': 9e-06, 'learning_rate/action_model': 9e-05, 'epoch': 4.9}) +[2026-06-20 12:01:01,483][starVLA.training.train_starvla][INFO] - Step 2691, Loss: {'train/loss': 0.27220723032951355, 'train/grad_norm_pre_clip': 12.154268264770508, 'timing/data': 0.0005175590049475431, 'timing/model': 0.9341173720313236, 'learning_rate/qwen_vl_interface': 9.100000000000001e-06, 'learning_rate/action_model': 9.1e-05, 'epoch': 4.9}) +[2026-06-20 12:01:05,222][starVLA.training.train_starvla][INFO] - Step 2692, Loss: {'train/loss': 0.2250114381313324, 'train/grad_norm_pre_clip': 12.620795249938965, 'timing/data': 0.0005477690137922764, 'timing/model': 0.9709057250292972, 'learning_rate/qwen_vl_interface': 9.200000000000002e-06, 'learning_rate/action_model': 9.200000000000001e-05, 'epoch': 4.9}) +[2026-06-20 12:01:08,947][starVLA.training.train_starvla][INFO] - Step 2693, Loss: {'train/loss': 0.342499315738678, 'train/grad_norm_pre_clip': 15.877462387084961, 'timing/data': 0.0005172889214009047, 'timing/model': 0.9631508520105854, 'learning_rate/qwen_vl_interface': 9.3e-06, 'learning_rate/action_model': 9.300000000000001e-05, 'epoch': 4.9}) +[2026-06-20 12:01:12,714][starVLA.training.train_starvla][INFO] - Step 2694, Loss: {'train/loss': 0.15674535930156708, 'train/grad_norm_pre_clip': 10.886423110961914, 'timing/data': 0.0005338690243661404, 'timing/model': 0.960469008074142, 'learning_rate/qwen_vl_interface': 9.4e-06, 'learning_rate/action_model': 9.4e-05, 'epoch': 4.9}) +[2026-06-20 12:01:16,395][starVLA.training.train_starvla][INFO] - Step 2695, Loss: {'train/loss': 0.17933638393878937, 'train/grad_norm_pre_clip': 16.293642044067383, 'timing/data': 0.0005281590856611729, 'timing/model': 0.9638773689512163, 'learning_rate/qwen_vl_interface': 9.5e-06, 'learning_rate/action_model': 9.5e-05, 'epoch': 4.91}) +[2026-06-20 12:01:20,218][starVLA.training.train_starvla][INFO] - Step 2696, Loss: {'train/loss': 0.19367922842502594, 'train/grad_norm_pre_clip': 8.270737648010254, 'timing/data': 0.0006522489711642265, 'timing/model': 0.9649702260503545, 'learning_rate/qwen_vl_interface': 9.600000000000001e-06, 'learning_rate/action_model': 9.6e-05, 'epoch': 4.91}) +[2026-06-20 12:01:24,003][starVLA.training.train_starvla][INFO] - Step 2697, Loss: {'train/loss': 0.530794620513916, 'train/grad_norm_pre_clip': 21.35256576538086, 'timing/data': 0.0005293990252539515, 'timing/model': 1.0115243160398677, 'learning_rate/qwen_vl_interface': 9.7e-06, 'learning_rate/action_model': 9.7e-05, 'epoch': 4.91}) +[2026-06-20 12:01:27,760][starVLA.training.train_starvla][INFO] - Step 2698, Loss: {'train/loss': 0.2155282348394394, 'train/grad_norm_pre_clip': 11.494537353515625, 'timing/data': 0.0005235089920461178, 'timing/model': 0.9679340380243957, 'learning_rate/qwen_vl_interface': 9.800000000000001e-06, 'learning_rate/action_model': 9.8e-05, 'epoch': 4.91}) +[2026-06-20 12:01:31,454][starVLA.training.train_starvla][INFO] - Step 2699, Loss: {'train/loss': 0.46554556488990784, 'train/grad_norm_pre_clip': 11.56495475769043, 'timing/data': 0.0005289489636197686, 'timing/model': 0.9298826050944626, 'learning_rate/qwen_vl_interface': 9.9e-06, 'learning_rate/action_model': 9.900000000000001e-05, 'epoch': 4.91}) +[2026-06-20 12:01:35,136][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 2700 +[2026-06-20 12:07:57,095][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 2700 +[2026-06-20 12:07:57,096][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 2700 +[2026-06-20 12:07:58,222][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_openvla/include -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_openvla/include -fPIC -c /tmp/tmp8n97sqm5/test.c -o /tmp/tmp8n97sqm5/test.o +[2026-06-20 12:07:58,249][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat /tmp/tmp8n97sqm5/test.o -laio -o /tmp/tmp8n97sqm5/a.out +[2026-06-20 12:07:58,666][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_openvla/include -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_openvla/include -fPIC -c /tmp/tmpz4k1fxh1/test.c -o /tmp/tmpz4k1fxh1/test.o +[2026-06-20 12:07:58,699][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat /tmp/tmpz4k1fxh1/test.o -L/usr/local/cuda -L/usr/local/cuda/lib64 -lcufile -o /tmp/tmpz4k1fxh1/a.out +[2026-06-20 12:07:58,744][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_openvla/include -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_openvla/include -fPIC -c /tmp/tmpimkl4hds/test.c -o /tmp/tmpimkl4hds/test.o +[2026-06-20 12:07:58,777][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat /tmp/tmpimkl4hds/test.o -laio -o /tmp/tmpimkl4hds/a.out +[2026-06-20 12:14:35,098][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 2700 +[2026-06-20 12:14:35,099][starVLA.training.train_starvla][INFO] - Step 2700, Loss: {'train/loss': 0.18552616238594055, 'train/grad_norm_pre_clip': 5.106449127197266, 'eval/loss': 0.31538140773773193, 'eval/latency_0/loss': 0.3155497610569, 'eval/latency_1/loss': 0.37630459666252136, 'eval/latency_2/loss': 0.37506383657455444, 'eval/latency_3/loss': 0.13413143157958984, 'eval/latency_4/loss': 0.3554246425628662, 'eval/flappy/loss': 0.31538140773773193, 'eval/flappy/latency_0/loss': 0.3155497610569, 'eval/flappy/latency_1/loss': 0.37630459666252136, 'eval/flappy/latency_2/loss': 0.37506383657455444, 'eval/flappy/latency_3/loss': 0.13413143157958984, 'eval/flappy/latency_4/loss': 0.3554246425628662, 'eval/flap_precision': 0.3223684210526316, 'eval/flap_recall': 0.1781818181818182, 'eval/flap_f1': 0.2295081967213115, 'eval/flap_rate_model': 0.042222222222222223, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.2295081967213115, 'eval/latency_0/flap_precision': 0.3223684210526316, 'eval/latency_0/flap_recall': 0.1781818181818182, 'eval/latency_0/flap_f1': 0.2295081967213115, 'eval/latency_0/flap_rate_model': 0.042222222222222223, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.2295081967213115, 'timing/data': 0.0005335190799087286, 'timing/model': 0.9936799440765753, 'learning_rate/qwen_vl_interface': 1e-05, 'learning_rate/action_model': 0.0001, 'epoch': 4.91}) +[2026-06-20 12:14:35,100][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_2700_state +[2026-06-20 12:14:35,104][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-20 12:14:43,641][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_2700_state/model.safetensors +[2026-06-20 12:15:03,510][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_2700_state/optimizer.bin +[2026-06-20 12:15:03,511][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_2700_state/sampler.bin +[2026-06-20 12:15:03,511][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_2700_state/sampler_1.bin +[2026-06-20 12:15:03,512][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_2700_state/random_states_0.pkl +[2026-06-20 12:15:06,344][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2700 +[2026-06-20 12:15:06,345][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-20 12:15:06,360][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-20 12:15:11,153][starVLA.training.train_starvla][INFO] - Step 2701, Loss: {'train/loss': 0.22072528302669525, 'train/grad_norm_pre_clip': 13.451971054077148, 'timing/data': 0.0006237169727683067, 'timing/model': 1.4185683859977871, 'learning_rate/qwen_vl_interface': 9.999999023727212e-06, 'learning_rate/action_model': 9.99999902372721e-05, 'epoch': 4.92}) +[2026-06-20 12:15:16,256][starVLA.training.train_starvla][INFO] - Step 2702, Loss: {'train/loss': 0.2698257863521576, 'train/grad_norm_pre_clip': 27.0367374420166, 'timing/data': 0.0005708970129489899, 'timing/model': 1.648981487029232, 'learning_rate/qwen_vl_interface': 9.999996094909247e-06, 'learning_rate/action_model': 9.999996094909248e-05, 'epoch': 4.92}) +[2026-06-20 12:15:21,333][starVLA.training.train_starvla][INFO] - Step 2703, Loss: {'train/loss': 0.1628388613462448, 'train/grad_norm_pre_clip': 9.537579536437988, 'timing/data': 0.0005171369994059205, 'timing/model': 1.2878563980339095, 'learning_rate/qwen_vl_interface': 9.999991213547314e-06, 'learning_rate/action_model': 9.999991213547313e-05, 'epoch': 4.92}) +[2026-06-20 12:15:26,591][starVLA.training.train_starvla][INFO] - Step 2704, Loss: {'train/loss': 0.27546268701553345, 'train/grad_norm_pre_clip': 13.285028457641602, 'timing/data': 0.0005452680634334683, 'timing/model': 1.2880846669431776, 'learning_rate/qwen_vl_interface': 9.999984379643413e-06, 'learning_rate/action_model': 9.999984379643412e-05, 'epoch': 4.92}) +[2026-06-20 12:15:31,880][starVLA.training.train_starvla][INFO] - Step 2705, Loss: {'train/loss': 0.5638366937637329, 'train/grad_norm_pre_clip': 15.927667617797852, 'timing/data': 0.0005309080006554723, 'timing/model': 1.3064465669449419, 'learning_rate/qwen_vl_interface': 9.999975593200357e-06, 'learning_rate/action_model': 9.999975593200357e-05, 'epoch': 4.92}) +[2026-06-20 12:15:37,155][starVLA.training.train_starvla][INFO] - Step 2706, Loss: {'train/loss': 0.14366847276687622, 'train/grad_norm_pre_clip': 5.875300407409668, 'timing/data': 0.0005379869835451245, 'timing/model': 1.2925274570006877, 'learning_rate/qwen_vl_interface': 9.999964854221756e-06, 'learning_rate/action_model': 9.999964854221754e-05, 'epoch': 4.93}) +[2026-06-20 12:15:42,440][starVLA.training.train_starvla][INFO] - Step 2707, Loss: {'train/loss': 0.2372702956199646, 'train/grad_norm_pre_clip': 13.305136680603027, 'timing/data': 0.0005308770341798663, 'timing/model': 1.5316304609877989, 'learning_rate/qwen_vl_interface': 9.999952162712027e-06, 'learning_rate/action_model': 9.999952162712026e-05, 'epoch': 4.93}) +[2026-06-20 12:15:47,696][starVLA.training.train_starvla][INFO] - Step 2708, Loss: {'train/loss': 0.29142096638679504, 'train/grad_norm_pre_clip': 14.062357902526855, 'timing/data': 0.0005535170203074813, 'timing/model': 1.4885992339113727, 'learning_rate/qwen_vl_interface': 9.999937518676382e-06, 'learning_rate/action_model': 9.999937518676382e-05, 'epoch': 4.93}) +[2026-06-20 12:15:52,399][starVLA.training.train_starvla][INFO] - Step 2709, Loss: {'train/loss': 0.27202603220939636, 'train/grad_norm_pre_clip': 9.834410667419434, 'timing/data': 0.0005703369388356805, 'timing/model': 1.1960756139596924, 'learning_rate/qwen_vl_interface': 9.999920922120846e-06, 'learning_rate/action_model': 9.999920922120846e-05, 'epoch': 4.93}) +[2026-06-20 12:15:57,152][starVLA.training.train_starvla][INFO] - Step 2710, Loss: {'train/loss': 0.3388103246688843, 'train/grad_norm_pre_clip': 11.864477157592773, 'timing/data': 0.0005493770586326718, 'timing/model': 1.1807280040811747, 'learning_rate/qwen_vl_interface': 9.999902373052239e-06, 'learning_rate/action_model': 9.999902373052239e-05, 'epoch': 4.93}) +[2026-06-20 12:16:02,429][starVLA.training.train_starvla][INFO] - Step 2711, Loss: {'train/loss': 0.19518360495567322, 'train/grad_norm_pre_clip': 13.0238676071167, 'timing/data': 0.0005199869628995657, 'timing/model': 1.2996686500264332, 'learning_rate/qwen_vl_interface': 9.999881871478188e-06, 'learning_rate/action_model': 9.999881871478187e-05, 'epoch': 4.93}) +[2026-06-20 12:16:07,227][starVLA.training.train_starvla][INFO] - Step 2712, Loss: {'train/loss': 0.349265992641449, 'train/grad_norm_pre_clip': 12.399462699890137, 'timing/data': 0.0005270279943943024, 'timing/model': 1.197929585003294, 'learning_rate/qwen_vl_interface': 9.999859417407115e-06, 'learning_rate/action_model': 9.999859417407115e-05, 'epoch': 4.94}) +[2026-06-20 12:16:12,020][starVLA.training.train_starvla][INFO] - Step 2713, Loss: {'train/loss': 0.1465606838464737, 'train/grad_norm_pre_clip': 7.304201126098633, 'timing/data': 0.0005420870147645473, 'timing/model': 1.4228478539735079, 'learning_rate/qwen_vl_interface': 9.999835010848254e-06, 'learning_rate/action_model': 9.999835010848254e-05, 'epoch': 4.94}) +[2026-06-20 12:16:16,575][starVLA.training.train_starvla][INFO] - Step 2714, Loss: {'train/loss': 0.18338538706302643, 'train/grad_norm_pre_clip': 8.556649208068848, 'timing/data': 0.0005519170081242919, 'timing/model': 1.1838409149786457, 'learning_rate/qwen_vl_interface': 9.999808651811639e-06, 'learning_rate/action_model': 9.999808651811639e-05, 'epoch': 4.94}) +[2026-06-20 12:16:21,375][starVLA.training.train_starvla][INFO] - Step 2715, Loss: {'train/loss': 0.14067448675632477, 'train/grad_norm_pre_clip': 7.049117565155029, 'timing/data': 0.0005290779517963529, 'timing/model': 1.2038994240574539, 'learning_rate/qwen_vl_interface': 9.999780340308101e-06, 'learning_rate/action_model': 9.999780340308101e-05, 'epoch': 4.94}) +[2026-06-20 12:16:26,188][starVLA.training.train_starvla][INFO] - Step 2716, Loss: {'train/loss': 0.3216809928417206, 'train/grad_norm_pre_clip': 11.50394058227539, 'timing/data': 0.0005522570572793484, 'timing/model': 1.189691097009927, 'learning_rate/qwen_vl_interface': 9.999750076349282e-06, 'learning_rate/action_model': 9.999750076349282e-05, 'epoch': 4.94}) +[2026-06-20 12:16:30,980][starVLA.training.train_starvla][INFO] - Step 2717, Loss: {'train/loss': 0.13538363575935364, 'train/grad_norm_pre_clip': 7.361138820648193, 'timing/data': 0.0005534969968721271, 'timing/model': 1.188956486992538, 'learning_rate/qwen_vl_interface': 9.999717859947618e-06, 'learning_rate/action_model': 9.999717859947618e-05, 'epoch': 4.95}) +[2026-06-20 12:16:35,820][starVLA.training.train_starvla][INFO] - Step 2718, Loss: {'train/loss': 0.264010488986969, 'train/grad_norm_pre_clip': 12.815896034240723, 'timing/data': 0.0005528971087187529, 'timing/model': 1.2035943520022556, 'learning_rate/qwen_vl_interface': 9.999683691116356e-06, 'learning_rate/action_model': 9.999683691116355e-05, 'epoch': 4.95}) +[2026-06-20 12:16:40,635][starVLA.training.train_starvla][INFO] - Step 2719, Loss: {'train/loss': 0.2291252762079239, 'train/grad_norm_pre_clip': 12.099652290344238, 'timing/data': 0.0005533780204132199, 'timing/model': 1.4327230879571289, 'learning_rate/qwen_vl_interface': 9.999647569869539e-06, 'learning_rate/action_model': 9.99964756986954e-05, 'epoch': 4.95}) +[2026-06-20 12:16:45,603][starVLA.training.train_starvla][INFO] - Step 2720, Loss: {'train/loss': 0.19295670092105865, 'train/grad_norm_pre_clip': 10.183094024658203, 'timing/data': 0.0005550869973376393, 'timing/model': 1.2011597460368648, 'learning_rate/qwen_vl_interface': 9.999609496222017e-06, 'learning_rate/action_model': 9.999609496222017e-05, 'epoch': 4.95}) +[2026-06-20 12:16:50,424][starVLA.training.train_starvla][INFO] - Step 2721, Loss: {'train/loss': 0.23234987258911133, 'train/grad_norm_pre_clip': 10.218464851379395, 'timing/data': 0.0008328059921041131, 'timing/model': 1.1776069209445268, 'learning_rate/qwen_vl_interface': 9.99956947018944e-06, 'learning_rate/action_model': 9.999569470189439e-05, 'epoch': 4.95}) +[2026-06-20 12:16:55,183][starVLA.training.train_starvla][INFO] - Step 2722, Loss: {'train/loss': 0.0826093852519989, 'train/grad_norm_pre_clip': 7.652498245239258, 'timing/data': 0.0005347880069166422, 'timing/model': 1.1764268920524046, 'learning_rate/qwen_vl_interface': 9.99952749178826e-06, 'learning_rate/action_model': 9.999527491788259e-05, 'epoch': 4.96}) +[2026-06-20 12:17:00,008][starVLA.training.train_starvla][INFO] - Step 2723, Loss: {'train/loss': 0.09675373136997223, 'train/grad_norm_pre_clip': 7.643002510070801, 'timing/data': 0.0005489980103448033, 'timing/model': 1.1824604870053008, 'learning_rate/qwen_vl_interface': 9.999483561035732e-06, 'learning_rate/action_model': 9.999483561035732e-05, 'epoch': 4.96}) +[2026-06-20 12:17:04,796][starVLA.training.train_starvla][INFO] - Step 2724, Loss: {'train/loss': 0.24471287429332733, 'train/grad_norm_pre_clip': 11.590651512145996, 'timing/data': 0.0005605770274996758, 'timing/model': 1.1781415219884366, 'learning_rate/qwen_vl_interface': 9.999437677949917e-06, 'learning_rate/action_model': 9.999437677949917e-05, 'epoch': 4.96}) +[2026-06-20 12:17:09,572][starVLA.training.train_starvla][INFO] - Step 2725, Loss: {'train/loss': 0.25887560844421387, 'train/grad_norm_pre_clip': 13.377161026000977, 'timing/data': 0.0005766670219600201, 'timing/model': 1.4220161950215697, 'learning_rate/qwen_vl_interface': 9.999389842549674e-06, 'learning_rate/action_model': 9.999389842549673e-05, 'epoch': 4.96}) +[2026-06-20 12:17:14,334][starVLA.training.train_starvla][INFO] - Step 2726, Loss: {'train/loss': 0.0555117130279541, 'train/grad_norm_pre_clip': 8.22169017791748, 'timing/data': 0.0005381179507821798, 'timing/model': 1.3864576460327953, 'learning_rate/qwen_vl_interface': 9.999340054854669e-06, 'learning_rate/action_model': 9.999340054854667e-05, 'epoch': 4.96}) +[2026-06-20 12:17:18,953][starVLA.training.train_starvla][INFO] - Step 2727, Loss: {'train/loss': 0.1468210518360138, 'train/grad_norm_pre_clip': 7.314894676208496, 'timing/data': 0.0005371769657358527, 'timing/model': 1.218056768993847, 'learning_rate/qwen_vl_interface': 9.999288314885362e-06, 'learning_rate/action_model': 9.999288314885362e-05, 'epoch': 4.96}) +[2026-06-20 12:17:23,847][starVLA.training.train_starvla][INFO] - Step 2728, Loss: {'train/loss': 0.17657196521759033, 'train/grad_norm_pre_clip': 14.362394332885742, 'timing/data': 0.0005401970120146871, 'timing/model': 1.1691814389778301, 'learning_rate/qwen_vl_interface': 9.999234622663027e-06, 'learning_rate/action_model': 9.999234622663028e-05, 'epoch': 4.97}) +[2026-06-20 12:17:28,659][starVLA.training.train_starvla][INFO] - Step 2729, Loss: {'train/loss': 0.24158860743045807, 'train/grad_norm_pre_clip': 9.582598686218262, 'timing/data': 0.0006611670833081007, 'timing/model': 1.1881332460325211, 'learning_rate/qwen_vl_interface': 9.999178978209732e-06, 'learning_rate/action_model': 9.999178978209732e-05, 'epoch': 4.97}) +[2026-06-20 12:17:33,449][starVLA.training.train_starvla][INFO] - Step 2730, Loss: {'train/loss': 0.2934989631175995, 'train/grad_norm_pre_clip': 18.01160430908203, 'timing/data': 0.0006098470184952021, 'timing/model': 1.1675556139089167, 'learning_rate/qwen_vl_interface': 9.999121381548354e-06, 'learning_rate/action_model': 9.999121381548353e-05, 'epoch': 4.97}) +[2026-06-20 12:17:38,224][starVLA.training.train_starvla][INFO] - Step 2731, Loss: {'train/loss': 0.4322647154331207, 'train/grad_norm_pre_clip': 21.92798614501953, 'timing/data': 0.0005491781048476696, 'timing/model': 1.409463283023797, 'learning_rate/qwen_vl_interface': 9.999061832702563e-06, 'learning_rate/action_model': 9.999061832702563e-05, 'epoch': 4.97}) +[2026-06-20 12:17:43,040][starVLA.training.train_starvla][INFO] - Step 2732, Loss: {'train/loss': 0.25545239448547363, 'train/grad_norm_pre_clip': 29.632577896118164, 'timing/data': 0.0005362980300560594, 'timing/model': 1.4566218330292031, 'learning_rate/qwen_vl_interface': 9.999000331696842e-06, 'learning_rate/action_model': 9.999000331696841e-05, 'epoch': 4.97}) +[2026-06-20 12:17:47,597][starVLA.training.train_starvla][INFO] - Step 2733, Loss: {'train/loss': 0.14772483706474304, 'train/grad_norm_pre_clip': 11.95134162902832, 'timing/data': 0.0005433179903775454, 'timing/model': 1.199149169959128, 'learning_rate/qwen_vl_interface': 9.99893687855647e-06, 'learning_rate/action_model': 9.99893687855647e-05, 'epoch': 4.98}) +[2026-06-20 12:17:52,368][starVLA.training.train_starvla][INFO] - Step 2734, Loss: {'train/loss': 0.27809271216392517, 'train/grad_norm_pre_clip': 17.445104598999023, 'timing/data': 0.0007323760073632002, 'timing/model': 1.1941653749672696, 'learning_rate/qwen_vl_interface': 9.998871473307529e-06, 'learning_rate/action_model': 9.99887147330753e-05, 'epoch': 4.98}) +[2026-06-20 12:17:57,177][starVLA.training.train_starvla][INFO] - Step 2735, Loss: {'train/loss': 0.22244195640087128, 'train/grad_norm_pre_clip': 12.081043243408203, 'timing/data': 0.0005497180391103029, 'timing/model': 1.1705727650551125, 'learning_rate/qwen_vl_interface': 9.998804115976907e-06, 'learning_rate/action_model': 9.998804115976908e-05, 'epoch': 4.98}) +[2026-06-20 12:18:02,012][starVLA.training.train_starvla][INFO] - Step 2736, Loss: {'train/loss': 0.10650895535945892, 'train/grad_norm_pre_clip': 7.133247375488281, 'timing/data': 0.000540897948667407, 'timing/model': 1.1736443439731374, 'learning_rate/qwen_vl_interface': 9.99873480659229e-06, 'learning_rate/action_model': 9.99873480659229e-05, 'epoch': 4.98}) +[2026-06-20 12:18:06,857][starVLA.training.train_starvla][INFO] - Step 2737, Loss: {'train/loss': 0.1590258628129959, 'train/grad_norm_pre_clip': 10.104109764099121, 'timing/data': 0.0005671980325132608, 'timing/model': 1.4185910429805517, 'learning_rate/qwen_vl_interface': 9.99866354518217e-06, 'learning_rate/action_model': 9.99866354518217e-05, 'epoch': 4.98}) +[2026-06-20 12:18:11,457][starVLA.training.train_starvla][INFO] - Step 2738, Loss: {'train/loss': 0.2987954020500183, 'train/grad_norm_pre_clip': 12.833685874938965, 'timing/data': 0.0005742770154029131, 'timing/model': 1.1733283709036186, 'learning_rate/qwen_vl_interface': 9.99859033177584e-06, 'learning_rate/action_model': 9.99859033177584e-05, 'epoch': 4.98}) +[2026-06-20 12:18:16,406][starVLA.training.train_starvla][INFO] - Step 2739, Loss: {'train/loss': 0.3977431058883667, 'train/grad_norm_pre_clip': 19.322284698486328, 'timing/data': 0.0005431370809674263, 'timing/model': 1.2023558219661936, 'learning_rate/qwen_vl_interface': 9.998515166403394e-06, 'learning_rate/action_model': 9.998515166403394e-05, 'epoch': 4.99}) +[2026-06-20 12:18:21,360][starVLA.training.train_starvla][INFO] - Step 2740, Loss: {'train/loss': 0.19601894915103912, 'train/grad_norm_pre_clip': 18.238372802734375, 'timing/data': 0.000545116956345737, 'timing/model': 1.2215729410527274, 'learning_rate/qwen_vl_interface': 9.99843804909573e-06, 'learning_rate/action_model': 9.99843804909573e-05, 'epoch': 4.99}) +[2026-06-20 12:18:26,299][starVLA.training.train_starvla][INFO] - Step 2741, Loss: {'train/loss': 0.18724685907363892, 'train/grad_norm_pre_clip': 10.146873474121094, 'timing/data': 0.0005483470158651471, 'timing/model': 1.1891450979746878, 'learning_rate/qwen_vl_interface': 9.998358979884548e-06, 'learning_rate/action_model': 9.998358979884547e-05, 'epoch': 4.99}) +[2026-06-20 12:18:31,040][starVLA.training.train_starvla][INFO] - Step 2742, Loss: {'train/loss': 0.18755726516246796, 'train/grad_norm_pre_clip': 7.666799068450928, 'timing/data': 0.0005503669381141663, 'timing/model': 1.1761291490402073, 'learning_rate/qwen_vl_interface': 9.998277958802352e-06, 'learning_rate/action_model': 9.998277958802351e-05, 'epoch': 4.99}) +[2026-06-20 12:18:35,769][starVLA.training.train_starvla][INFO] - Step 2743, Loss: {'train/loss': 0.24527958035469055, 'train/grad_norm_pre_clip': 12.889720916748047, 'timing/data': 0.0005346480756998062, 'timing/model': 1.3830781809519976, 'learning_rate/qwen_vl_interface': 9.998194985882442e-06, 'learning_rate/action_model': 9.998194985882442e-05, 'epoch': 4.99}) +[2026-06-20 12:18:40,287][starVLA.training.train_starvla][INFO] - Step 2744, Loss: {'train/loss': 0.25910985469818115, 'train/grad_norm_pre_clip': 11.152078628540039, 'timing/data': 0.0005488470196723938, 'timing/model': 1.173013495048508, 'learning_rate/qwen_vl_interface': 9.99811006115893e-06, 'learning_rate/action_model': 9.998110061158931e-05, 'epoch': 5.0}) +[2026-06-20 12:18:45,077][starVLA.training.train_starvla][INFO] - Step 2745, Loss: {'train/loss': 0.18813081085681915, 'train/grad_norm_pre_clip': 6.551059246063232, 'timing/data': 0.0005480078980326653, 'timing/model': 1.1713782451115549, 'learning_rate/qwen_vl_interface': 9.998023184666726e-06, 'learning_rate/action_model': 9.998023184666726e-05, 'epoch': 5.0}) +[2026-06-20 12:18:49,859][starVLA.training.train_starvla][INFO] - Step 2746, Loss: {'train/loss': 0.15334412455558777, 'train/grad_norm_pre_clip': 6.829064846038818, 'timing/data': 0.0005446180002763867, 'timing/model': 1.1766168220201507, 'learning_rate/qwen_vl_interface': 9.997934356441538e-06, 'learning_rate/action_model': 9.997934356441537e-05, 'epoch': 5.0}) +[2026-06-20 12:18:54,665][starVLA.training.train_starvla][INFO] - Step 2747, Loss: {'train/loss': 0.30294734239578247, 'train/grad_norm_pre_clip': 12.396835327148438, 'timing/data': 0.0005541580030694604, 'timing/model': 1.1968213450163603, 'learning_rate/qwen_vl_interface': 9.997843576519879e-06, 'learning_rate/action_model': 9.997843576519878e-05, 'epoch': 5.0}) +[2026-06-20 12:18:59,475][starVLA.training.train_starvla][INFO] - Step 2748, Loss: {'train/loss': 0.14663085341453552, 'train/grad_norm_pre_clip': 7.434058666229248, 'timing/data': 0.0005582680460065603, 'timing/model': 1.1928139840019867, 'learning_rate/qwen_vl_interface': 9.99775084493907e-06, 'learning_rate/action_model': 9.997750844939069e-05, 'epoch': 5.0}) +[2026-06-20 12:19:04,229][starVLA.training.train_starvla][INFO] - Step 2749, Loss: {'train/loss': 0.2562229037284851, 'train/grad_norm_pre_clip': 16.577611923217773, 'timing/data': 0.0005390880396589637, 'timing/model': 1.3914265159983188, 'learning_rate/qwen_vl_interface': 9.997656161737227e-06, 'learning_rate/action_model': 9.997656161737226e-05, 'epoch': 5.0}) +[2026-06-20 12:19:08,784][starVLA.training.train_starvla][INFO] - Step 2750, Loss: {'train/loss': 0.26022306084632874, 'train/grad_norm_pre_clip': 12.167411804199219, 'timing/data': 0.0005450879689306021, 'timing/model': 1.1949362279847264, 'learning_rate/qwen_vl_interface': 9.99755952695327e-06, 'learning_rate/action_model': 9.997559526953269e-05, 'epoch': 5.01}) +[2026-06-20 12:19:13,567][starVLA.training.train_starvla][INFO] - Step 2751, Loss: {'train/loss': 0.38704514503479004, 'train/grad_norm_pre_clip': 13.061151504516602, 'timing/data': 0.0005567470798268914, 'timing/model': 1.172692195046693, 'learning_rate/qwen_vl_interface': 9.99746094062692e-06, 'learning_rate/action_model': 9.99746094062692e-05, 'epoch': 5.01}) +[2026-06-20 12:19:18,376][starVLA.training.train_starvla][INFO] - Step 2752, Loss: {'train/loss': 0.07564505934715271, 'train/grad_norm_pre_clip': 4.127475261688232, 'timing/data': 0.0005636280402541161, 'timing/model': 1.199621869949624, 'learning_rate/qwen_vl_interface': 9.997360402798706e-06, 'learning_rate/action_model': 9.997360402798705e-05, 'epoch': 5.01}) +[2026-06-20 12:19:23,176][starVLA.training.train_starvla][INFO] - Step 2753, Loss: {'train/loss': 0.28830966353416443, 'train/grad_norm_pre_clip': 15.117514610290527, 'timing/data': 0.0007279770215973258, 'timing/model': 1.188841218012385, 'learning_rate/qwen_vl_interface': 9.997257913509955e-06, 'learning_rate/action_model': 9.997257913509955e-05, 'epoch': 5.01}) +[2026-06-20 12:19:27,999][starVLA.training.train_starvla][INFO] - Step 2754, Loss: {'train/loss': 0.28523513674736023, 'train/grad_norm_pre_clip': 14.382322311401367, 'timing/data': 0.0006285769632086158, 'timing/model': 1.4287538679782301, 'learning_rate/qwen_vl_interface': 9.997153472802794e-06, 'learning_rate/action_model': 9.997153472802795e-05, 'epoch': 5.01}) +[2026-06-20 12:19:32,543][starVLA.training.train_starvla][INFO] - Step 2755, Loss: {'train/loss': 0.08545518666505814, 'train/grad_norm_pre_clip': 5.524876117706299, 'timing/data': 0.0007662270218133926, 'timing/model': 1.1844316680217162, 'learning_rate/qwen_vl_interface': 9.997047080720157e-06, 'learning_rate/action_model': 9.997047080720158e-05, 'epoch': 5.02}) +[2026-06-20 12:19:37,312][starVLA.training.train_starvla][INFO] - Step 2756, Loss: {'train/loss': 0.11345000565052032, 'train/grad_norm_pre_clip': 6.2668538093566895, 'timing/data': 0.0005356080364435911, 'timing/model': 1.178046126035042, 'learning_rate/qwen_vl_interface': 9.996938737305777e-06, 'learning_rate/action_model': 9.996938737305776e-05, 'epoch': 5.02}) +[2026-06-20 12:19:42,116][starVLA.training.train_starvla][INFO] - Step 2757, Loss: {'train/loss': 0.22546303272247314, 'train/grad_norm_pre_clip': 14.322531700134277, 'timing/data': 0.0005443680565804243, 'timing/model': 1.1768084419891238, 'learning_rate/qwen_vl_interface': 9.996828442604189e-06, 'learning_rate/action_model': 9.99682844260419e-05, 'epoch': 5.02}) +[2026-06-20 12:19:46,877][starVLA.training.train_starvla][INFO] - Step 2758, Loss: {'train/loss': 0.2183210551738739, 'train/grad_norm_pre_clip': 26.068910598754883, 'timing/data': 0.00054399692453444, 'timing/model': 1.2022534030256793, 'learning_rate/qwen_vl_interface': 9.996716196660733e-06, 'learning_rate/action_model': 9.996716196660731e-05, 'epoch': 5.02}) +[2026-06-20 12:19:51,672][starVLA.training.train_starvla][INFO] - Step 2759, Loss: {'train/loss': 0.11982163041830063, 'train/grad_norm_pre_clip': 12.990180015563965, 'timing/data': 0.000521437032148242, 'timing/model': 1.166447746916674, 'learning_rate/qwen_vl_interface': 9.996601999521546e-06, 'learning_rate/action_model': 9.996601999521546e-05, 'epoch': 5.02}) +[2026-06-20 12:19:56,492][starVLA.training.train_starvla][INFO] - Step 2760, Loss: {'train/loss': 0.15619735419750214, 'train/grad_norm_pre_clip': 10.212644577026367, 'timing/data': 0.0005472080083563924, 'timing/model': 1.4202810220886022, 'learning_rate/qwen_vl_interface': 9.996485851233572e-06, 'learning_rate/action_model': 9.996485851233572e-05, 'epoch': 5.02}) +[2026-06-20 12:20:01,654][starVLA.training.train_starvla][INFO] - Step 2761, Loss: {'train/loss': 0.22816194593906403, 'train/grad_norm_pre_clip': 19.277938842773438, 'timing/data': 0.0014163339510560036, 'timing/model': 1.366715913056396, 'learning_rate/qwen_vl_interface': 9.996367751844556e-06, 'learning_rate/action_model': 9.996367751844555e-05, 'epoch': 5.03}) +[2026-06-20 12:20:07,119][starVLA.training.train_starvla][INFO] - Step 2762, Loss: {'train/loss': 0.40078020095825195, 'train/grad_norm_pre_clip': 19.101844787597656, 'timing/data': 0.0010167760774493217, 'timing/model': 1.3563574610743672, 'learning_rate/qwen_vl_interface': 9.996247701403043e-06, 'learning_rate/action_model': 9.996247701403042e-05, 'epoch': 5.03}) +[2026-06-20 12:20:12,711][starVLA.training.train_starvla][INFO] - Step 2763, Loss: {'train/loss': 0.14521053433418274, 'train/grad_norm_pre_clip': 31.18706512451172, 'timing/data': 0.0011343950172886252, 'timing/model': 1.3548361100256443, 'learning_rate/qwen_vl_interface': 9.99612569995838e-06, 'learning_rate/action_model': 9.99612569995838e-05, 'epoch': 5.03}) +[2026-06-20 12:20:18,182][starVLA.training.train_starvla][INFO] - Step 2764, Loss: {'train/loss': 0.7988835573196411, 'train/grad_norm_pre_clip': 25.08171844482422, 'timing/data': 0.0014433839824050665, 'timing/model': 1.3585188969736919, 'learning_rate/qwen_vl_interface': 9.99600174756072e-06, 'learning_rate/action_model': 9.99600174756072e-05, 'epoch': 5.03}) +[2026-06-20 12:20:23,729][starVLA.training.train_starvla][INFO] - Step 2765, Loss: {'train/loss': 0.30164000391960144, 'train/grad_norm_pre_clip': 31.570438385009766, 'timing/data': 0.0010281760478392243, 'timing/model': 1.3541754790348932, 'learning_rate/qwen_vl_interface': 9.995875844261012e-06, 'learning_rate/action_model': 9.995875844261013e-05, 'epoch': 5.03}) +[2026-06-20 12:20:29,266][starVLA.training.train_starvla][INFO] - Step 2766, Loss: {'train/loss': 0.21730150282382965, 'train/grad_norm_pre_clip': 16.502002716064453, 'timing/data': 0.0010219960240647197, 'timing/model': 1.6600445199292153, 'learning_rate/qwen_vl_interface': 9.995747990111015e-06, 'learning_rate/action_model': 9.995747990111013e-05, 'epoch': 5.04}) +[2026-06-20 12:20:34,478][starVLA.training.train_starvla][INFO] - Step 2767, Loss: {'train/loss': 0.4988503158092499, 'train/grad_norm_pre_clip': 32.66389846801758, 'timing/data': 0.001015554997138679, 'timing/model': 1.3482989389449358, 'learning_rate/qwen_vl_interface': 9.99561818516328e-06, 'learning_rate/action_model': 9.99561818516328e-05, 'epoch': 5.04}) +[2026-06-20 12:20:40,044][starVLA.training.train_starvla][INFO] - Step 2768, Loss: {'train/loss': 0.1775301694869995, 'train/grad_norm_pre_clip': 15.307939529418945, 'timing/data': 0.0010167560540139675, 'timing/model': 1.3499835829716176, 'learning_rate/qwen_vl_interface': 9.995486429471166e-06, 'learning_rate/action_model': 9.995486429471166e-05, 'epoch': 5.04}) +[2026-06-20 12:20:45,591][starVLA.training.train_starvla][INFO] - Step 2769, Loss: {'train/loss': 0.14108498394489288, 'train/grad_norm_pre_clip': 12.356536865234375, 'timing/data': 0.001380982925184071, 'timing/model': 1.3617383379023522, 'learning_rate/qwen_vl_interface': 9.995352723088833e-06, 'learning_rate/action_model': 9.995352723088834e-05, 'epoch': 5.04}) +[2026-06-20 12:20:50,742][starVLA.training.train_starvla][INFO] - Step 2770, Loss: {'train/loss': 0.34914425015449524, 'train/grad_norm_pre_clip': 19.412206649780273, 'timing/data': 0.0007763769244775176, 'timing/model': 1.190681834006682, 'learning_rate/qwen_vl_interface': 9.995217066071245e-06, 'learning_rate/action_model': 9.995217066071246e-05, 'epoch': 5.04}) +[2026-06-20 12:20:55,494][starVLA.training.train_starvla][INFO] - Step 2771, Loss: {'train/loss': 0.18583030998706818, 'train/grad_norm_pre_clip': 10.064627647399902, 'timing/data': 0.0006149769760668278, 'timing/model': 1.4163826720323414, 'learning_rate/qwen_vl_interface': 9.995079458474163e-06, 'learning_rate/action_model': 9.995079458474164e-05, 'epoch': 5.04}) +[2026-06-20 12:21:00,179][starVLA.training.train_starvla][INFO] - Step 2772, Loss: {'train/loss': 0.14097417891025543, 'train/grad_norm_pre_clip': 11.515395164489746, 'timing/data': 0.0005232769763097167, 'timing/model': 1.3421620009467006, 'learning_rate/qwen_vl_interface': 9.994939900354154e-06, 'learning_rate/action_model': 9.994939900354154e-05, 'epoch': 5.05}) +[2026-06-20 12:21:05,410][starVLA.training.train_starvla][INFO] - Step 2773, Loss: {'train/loss': 0.14064835011959076, 'train/grad_norm_pre_clip': 7.789729595184326, 'timing/data': 0.001238614902831614, 'timing/model': 1.353372095967643, 'learning_rate/qwen_vl_interface': 9.994798391768582e-06, 'learning_rate/action_model': 9.994798391768582e-05, 'epoch': 5.05}) +[2026-06-20 12:21:10,363][starVLA.training.train_starvla][INFO] - Step 2774, Loss: {'train/loss': 0.25080087780952454, 'train/grad_norm_pre_clip': 24.13416862487793, 'timing/data': 0.0005583380116149783, 'timing/model': 1.230735106044449, 'learning_rate/qwen_vl_interface': 9.99465493277562e-06, 'learning_rate/action_model': 9.994654932775618e-05, 'epoch': 5.05}) +[2026-06-20 12:21:15,208][starVLA.training.train_starvla][INFO] - Step 2775, Loss: {'train/loss': 0.23584574460983276, 'train/grad_norm_pre_clip': 12.649971961975098, 'timing/data': 0.0005313979927450418, 'timing/model': 1.1696493090130389, 'learning_rate/qwen_vl_interface': 9.994509523434235e-06, 'learning_rate/action_model': 9.994509523434235e-05, 'epoch': 5.05}) +[2026-06-20 12:21:20,009][starVLA.training.train_starvla][INFO] - Step 2776, Loss: {'train/loss': 0.33275726437568665, 'train/grad_norm_pre_clip': 13.28165340423584, 'timing/data': 0.0005177579587325454, 'timing/model': 1.1680327630601823, 'learning_rate/qwen_vl_interface': 9.994362163804202e-06, 'learning_rate/action_model': 9.9943621638042e-05, 'epoch': 5.05}) +[2026-06-20 12:21:24,792][starVLA.training.train_starvla][INFO] - Step 2777, Loss: {'train/loss': 0.35946688055992126, 'train/grad_norm_pre_clip': 18.459041595458984, 'timing/data': 0.0005186179187148809, 'timing/model': 1.4270627810619771, 'learning_rate/qwen_vl_interface': 9.994212853946092e-06, 'learning_rate/action_model': 9.994212853946092e-05, 'epoch': 5.06}) +[2026-06-20 12:21:29,525][starVLA.training.train_starvla][INFO] - Step 2778, Loss: {'train/loss': 0.11874312907457352, 'train/grad_norm_pre_clip': 8.36754322052002, 'timing/data': 0.0005390379810705781, 'timing/model': 1.356122893979773, 'learning_rate/qwen_vl_interface': 9.994061593921283e-06, 'learning_rate/action_model': 9.994061593921282e-05, 'epoch': 5.06}) +[2026-06-20 12:21:34,056][starVLA.training.train_starvla][INFO] - Step 2779, Loss: {'train/loss': 0.26004669070243835, 'train/grad_norm_pre_clip': 10.37458324432373, 'timing/data': 0.0005437179934233427, 'timing/model': 1.1728873089887202, 'learning_rate/qwen_vl_interface': 9.993908383791951e-06, 'learning_rate/action_model': 9.993908383791951e-05, 'epoch': 5.06}) +[2026-06-20 12:21:38,848][starVLA.training.train_starvla][INFO] - Step 2780, Loss: {'train/loss': 0.37393608689308167, 'train/grad_norm_pre_clip': 14.283358573913574, 'timing/data': 0.0005329570267349482, 'timing/model': 1.174408940016292, 'learning_rate/qwen_vl_interface': 9.993753223621076e-06, 'learning_rate/action_model': 9.993753223621077e-05, 'epoch': 5.06}) +[2026-06-20 12:21:43,608][starVLA.training.train_starvla][INFO] - Step 2781, Loss: {'train/loss': 0.10555193573236465, 'train/grad_norm_pre_clip': 4.415225982666016, 'timing/data': 0.0005352679872885346, 'timing/model': 1.1703786159632728, 'learning_rate/qwen_vl_interface': 9.99359611347244e-06, 'learning_rate/action_model': 9.993596113472439e-05, 'epoch': 5.06}) +[2026-06-20 12:21:48,421][starVLA.training.train_starvla][INFO] - Step 2782, Loss: {'train/loss': 0.3124861419200897, 'train/grad_norm_pre_clip': 11.668413162231445, 'timing/data': 0.0005186080234125257, 'timing/model': 1.1873115029884502, 'learning_rate/qwen_vl_interface': 9.99343705341062e-06, 'learning_rate/action_model': 9.99343705341062e-05, 'epoch': 5.06}) +[2026-06-20 12:21:53,304][starVLA.training.train_starvla][INFO] - Step 2783, Loss: {'train/loss': 0.3463212251663208, 'train/grad_norm_pre_clip': 10.784469604492188, 'timing/data': 0.000556856975890696, 'timing/model': 1.5298120719380677, 'learning_rate/qwen_vl_interface': 9.993276043501004e-06, 'learning_rate/action_model': 9.993276043501004e-05, 'epoch': 5.07}) +[2026-06-20 12:21:57,899][starVLA.training.train_starvla][INFO] - Step 2784, Loss: {'train/loss': 0.29811012744903564, 'train/grad_norm_pre_clip': 12.307194709777832, 'timing/data': 0.0005766779650002718, 'timing/model': 1.2044172699097544, 'learning_rate/qwen_vl_interface': 9.993113083809776e-06, 'learning_rate/action_model': 9.993113083809775e-05, 'epoch': 5.07}) +[2026-06-20 12:22:02,699][starVLA.training.train_starvla][INFO] - Step 2785, Loss: {'train/loss': 0.22706355154514313, 'train/grad_norm_pre_clip': 10.599502563476562, 'timing/data': 0.0005416879430413246, 'timing/model': 1.180645624990575, 'learning_rate/qwen_vl_interface': 9.992948174403921e-06, 'learning_rate/action_model': 9.992948174403922e-05, 'epoch': 5.07}) +[2026-06-20 12:22:07,529][starVLA.training.train_starvla][INFO] - Step 2786, Loss: {'train/loss': 0.1014709547162056, 'train/grad_norm_pre_clip': 5.406336784362793, 'timing/data': 0.000549658085219562, 'timing/model': 1.205089895054698, 'learning_rate/qwen_vl_interface': 9.99278131535123e-06, 'learning_rate/action_model': 9.992781315351229e-05, 'epoch': 5.07}) +[2026-06-20 12:22:12,315][starVLA.training.train_starvla][INFO] - Step 2787, Loss: {'train/loss': 0.06575390696525574, 'train/grad_norm_pre_clip': 24.20122528076172, 'timing/data': 0.0005457779625430703, 'timing/model': 1.1840484481072053, 'learning_rate/qwen_vl_interface': 9.992612506720289e-06, 'learning_rate/action_model': 9.99261250672029e-05, 'epoch': 5.07}) +[2026-06-20 12:22:17,153][starVLA.training.train_starvla][INFO] - Step 2788, Loss: {'train/loss': 0.17201058566570282, 'train/grad_norm_pre_clip': 5.751764297485352, 'timing/data': 0.0005878070369362831, 'timing/model': 1.1890982560580596, 'learning_rate/qwen_vl_interface': 9.992441748580492e-06, 'learning_rate/action_model': 9.99244174858049e-05, 'epoch': 5.08}) +[2026-06-20 12:22:21,965][starVLA.training.train_starvla][INFO] - Step 2789, Loss: {'train/loss': 0.322407066822052, 'train/grad_norm_pre_clip': 11.34830379486084, 'timing/data': 0.000566388014703989, 'timing/model': 1.4256083959480748, 'learning_rate/qwen_vl_interface': 9.992269041002028e-06, 'learning_rate/action_model': 9.992269041002027e-05, 'epoch': 5.08}) +[2026-06-20 12:22:26,527][starVLA.training.train_starvla][INFO] - Step 2790, Loss: {'train/loss': 0.30398234724998474, 'train/grad_norm_pre_clip': 21.876745223999023, 'timing/data': 0.0005792479496449232, 'timing/model': 1.1901537070516497, 'learning_rate/qwen_vl_interface': 9.992094384055894e-06, 'learning_rate/action_model': 9.992094384055895e-05, 'epoch': 5.08}) +[2026-06-20 12:22:31,316][starVLA.training.train_starvla][INFO] - Step 2791, Loss: {'train/loss': 0.18873965740203857, 'train/grad_norm_pre_clip': 8.645974159240723, 'timing/data': 0.000545907998457551, 'timing/model': 1.1806394529994577, 'learning_rate/qwen_vl_interface': 9.991917777813884e-06, 'learning_rate/action_model': 9.991917777813884e-05, 'epoch': 5.08}) +[2026-06-20 12:22:36,105][starVLA.training.train_starvla][INFO] - Step 2792, Loss: {'train/loss': 0.1828513741493225, 'train/grad_norm_pre_clip': 10.629242897033691, 'timing/data': 0.0006810670020058751, 'timing/model': 1.1923213630216196, 'learning_rate/qwen_vl_interface': 9.991739222348593e-06, 'learning_rate/action_model': 9.991739222348592e-05, 'epoch': 5.08}) +[2026-06-20 12:22:40,943][starVLA.training.train_starvla][INFO] - Step 2793, Loss: {'train/loss': 0.07838615775108337, 'train/grad_norm_pre_clip': 5.40256929397583, 'timing/data': 0.0005474078934639692, 'timing/model': 1.2067736620083451, 'learning_rate/qwen_vl_interface': 9.991558717733418e-06, 'learning_rate/action_model': 9.991558717733417e-05, 'epoch': 5.08}) +[2026-06-20 12:22:45,751][starVLA.training.train_starvla][INFO] - Step 2794, Loss: {'train/loss': 0.44517308473587036, 'train/grad_norm_pre_clip': 11.258270263671875, 'timing/data': 0.0005462979897856712, 'timing/model': 1.1808966940734535, 'learning_rate/qwen_vl_interface': 9.99137626404256e-06, 'learning_rate/action_model': 9.991376264042559e-05, 'epoch': 5.09}) +[2026-06-20 12:22:50,532][starVLA.training.train_starvla][INFO] - Step 2795, Loss: {'train/loss': 0.21197538077831268, 'train/grad_norm_pre_clip': 15.225217819213867, 'timing/data': 0.0005433079786598682, 'timing/model': 1.4213095960440114, 'learning_rate/qwen_vl_interface': 9.991191861351015e-06, 'learning_rate/action_model': 9.991191861351014e-05, 'epoch': 5.09}) +[2026-06-20 12:22:55,268][starVLA.training.train_starvla][INFO] - Step 2796, Loss: {'train/loss': 0.1329202651977539, 'train/grad_norm_pre_clip': 9.051836013793945, 'timing/data': 0.0007750160293653607, 'timing/model': 1.3684134170180187, 'learning_rate/qwen_vl_interface': 9.991005509734587e-06, 'learning_rate/action_model': 9.991005509734587e-05, 'epoch': 5.09}) +[2026-06-20 12:22:59,821][starVLA.training.train_starvla][INFO] - Step 2797, Loss: {'train/loss': 0.16072167456150055, 'train/grad_norm_pre_clip': 12.028908729553223, 'timing/data': 0.0007024270016700029, 'timing/model': 1.1753142580855638, 'learning_rate/qwen_vl_interface': 9.990817209269878e-06, 'learning_rate/action_model': 9.990817209269878e-05, 'epoch': 5.09}) +[2026-06-20 12:23:04,760][starVLA.training.train_starvla][INFO] - Step 2798, Loss: {'train/loss': 0.22230283915996552, 'train/grad_norm_pre_clip': 14.382644653320312, 'timing/data': 0.0005170170916244388, 'timing/model': 1.3534998840186745, 'learning_rate/qwen_vl_interface': 9.990626960034291e-06, 'learning_rate/action_model': 9.990626960034291e-05, 'epoch': 5.09}) +[2026-06-20 12:23:09,990][starVLA.training.train_starvla][INFO] - Step 2799, Loss: {'train/loss': 0.12716025114059448, 'train/grad_norm_pre_clip': 13.059749603271484, 'timing/data': 0.0005965980235487223, 'timing/model': 1.3008850410114974, 'learning_rate/qwen_vl_interface': 9.990434762106029e-06, 'learning_rate/action_model': 9.990434762106028e-05, 'epoch': 5.1}) +[2026-06-20 12:23:15,146][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 2800 +[2026-06-20 12:29:39,033][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 2800 +[2026-06-20 12:29:39,034][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 2800 +[2026-06-20 12:36:03,282][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 2800 +[2026-06-20 12:36:03,284][starVLA.training.train_starvla][INFO] - Step 2800, Loss: {'train/loss': 0.14258092641830444, 'train/grad_norm_pre_clip': 8.264623641967773, 'eval/loss': 0.3806423842906952, 'eval/latency_0/loss': 0.3316071927547455, 'eval/latency_1/loss': 0.48363661766052246, 'eval/latency_2/loss': 0.4859537184238434, 'eval/latency_3/loss': 0.13114827871322632, 'eval/latency_4/loss': 0.43782559037208557, 'eval/flappy/loss': 0.3806423842906952, 'eval/flappy/latency_0/loss': 0.3316071927547455, 'eval/flappy/latency_1/loss': 0.48363661766052246, 'eval/flappy/latency_2/loss': 0.4859537184238434, 'eval/flappy/latency_3/loss': 0.13114827871322632, 'eval/flappy/latency_4/loss': 0.43782559037208557, 'eval/flap_precision': 0.3235294117647059, 'eval/flap_recall': 0.16, 'eval/flap_f1': 0.21411192214111924, 'eval/flap_rate_model': 0.03777777777777778, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.21411192214111924, 'eval/latency_0/flap_precision': 0.3235294117647059, 'eval/latency_0/flap_recall': 0.16, 'eval/latency_0/flap_f1': 0.21411192214111924, 'eval/latency_0/flap_rate_model': 0.03777777777777778, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.21411192214111924, 'timing/data': 0.0005297580501064658, 'timing/model': 1.182574253063649, 'learning_rate/qwen_vl_interface': 9.990240615564099e-06, 'learning_rate/action_model': 9.990240615564099e-05, 'epoch': 5.1}) +[2026-06-20 12:36:03,285][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_2800_state +[2026-06-20 12:36:03,289][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-20 12:36:12,279][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_2800_state/model.safetensors +[2026-06-20 12:36:26,628][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_2800_state/optimizer.bin +[2026-06-20 12:36:26,628][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_2800_state/sampler.bin +[2026-06-20 12:36:26,628][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_2800_state/sampler_1.bin +[2026-06-20 12:36:26,629][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_2800_state/random_states_0.pkl +[2026-06-20 12:36:28,923][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2800 +[2026-06-20 12:36:28,923][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-20 12:36:28,937][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-20 12:36:33,677][starVLA.training.train_starvla][INFO] - Step 2801, Loss: {'train/loss': 0.18294788897037506, 'train/grad_norm_pre_clip': 13.691869735717773, 'timing/data': 0.0005636679707095027, 'timing/model': 1.166176536004059, 'learning_rate/qwen_vl_interface': 9.990044520488305e-06, 'learning_rate/action_model': 9.990044520488305e-05, 'epoch': 5.1}) +[2026-06-20 12:36:38,493][starVLA.training.train_starvla][INFO] - Step 2802, Loss: {'train/loss': 0.2366085797548294, 'train/grad_norm_pre_clip': 16.20461654663086, 'timing/data': 0.0005210380768403411, 'timing/model': 1.4432341509964317, 'learning_rate/qwen_vl_interface': 9.989846476959258e-06, 'learning_rate/action_model': 9.989846476959257e-05, 'epoch': 5.1}) +[2026-06-20 12:36:43,263][starVLA.training.train_starvla][INFO] - Step 2803, Loss: {'train/loss': 0.11497364938259125, 'train/grad_norm_pre_clip': 10.456953048706055, 'timing/data': 0.0005242580082267523, 'timing/model': 1.4453552659833804, 'learning_rate/qwen_vl_interface': 9.989646485058361e-06, 'learning_rate/action_model': 9.989646485058361e-05, 'epoch': 5.1}) +[2026-06-20 12:36:47,780][starVLA.training.train_starvla][INFO] - Step 2804, Loss: {'train/loss': 0.05975043401122093, 'train/grad_norm_pre_clip': 4.490545749664307, 'timing/data': 0.0005211319075897336, 'timing/model': 1.178145905956626, 'learning_rate/qwen_vl_interface': 9.98944454486783e-06, 'learning_rate/action_model': 9.989444544867829e-05, 'epoch': 5.1}) +[2026-06-20 12:36:52,552][starVLA.training.train_starvla][INFO] - Step 2805, Loss: {'train/loss': 0.23795929551124573, 'train/grad_norm_pre_clip': 7.328043460845947, 'timing/data': 0.0005533820949494839, 'timing/model': 1.1785071140620857, 'learning_rate/qwen_vl_interface': 9.989240656470667e-06, 'learning_rate/action_model': 9.989240656470666e-05, 'epoch': 5.11}) +[2026-06-20 12:36:57,368][starVLA.training.train_starvla][INFO] - Step 2806, Loss: {'train/loss': 0.09234005957841873, 'train/grad_norm_pre_clip': 14.012133598327637, 'timing/data': 0.0005408709403127432, 'timing/model': 1.170127185061574, 'learning_rate/qwen_vl_interface': 9.98903481995069e-06, 'learning_rate/action_model': 9.989034819950689e-05, 'epoch': 5.11}) +[2026-06-20 12:37:02,198][starVLA.training.train_starvla][INFO] - Step 2807, Loss: {'train/loss': 0.15961307287216187, 'train/grad_norm_pre_clip': 6.585079193115234, 'timing/data': 0.0005426909774541855, 'timing/model': 1.1775631789350882, 'learning_rate/qwen_vl_interface': 9.988827035392506e-06, 'learning_rate/action_model': 9.988827035392507e-05, 'epoch': 5.11}) +[2026-06-20 12:37:07,024][starVLA.training.train_starvla][INFO] - Step 2808, Loss: {'train/loss': 0.3118802607059479, 'train/grad_norm_pre_clip': 16.5885009765625, 'timing/data': 0.0005425310228019953, 'timing/model': 1.454698177985847, 'learning_rate/qwen_vl_interface': 9.988617302881529e-06, 'learning_rate/action_model': 9.988617302881529e-05, 'epoch': 5.11}) +[2026-06-20 12:37:11,543][starVLA.training.train_starvla][INFO] - Step 2809, Loss: {'train/loss': 0.11873725056648254, 'train/grad_norm_pre_clip': 4.967164516448975, 'timing/data': 0.0005359509959816933, 'timing/model': 1.1696779070189223, 'learning_rate/qwen_vl_interface': 9.988405622503972e-06, 'learning_rate/action_model': 9.988405622503971e-05, 'epoch': 5.11}) +[2026-06-20 12:37:16,340][starVLA.training.train_starvla][INFO] - Step 2810, Loss: {'train/loss': 0.289132684469223, 'train/grad_norm_pre_clip': 12.998836517333984, 'timing/data': 0.000545092043466866, 'timing/model': 1.173917631036602, 'learning_rate/qwen_vl_interface': 9.988191994346848e-06, 'learning_rate/action_model': 9.988191994346848e-05, 'epoch': 5.12}) +[2026-06-20 12:37:21,129][starVLA.training.train_starvla][INFO] - Step 2811, Loss: {'train/loss': 0.2320336103439331, 'train/grad_norm_pre_clip': 6.923661231994629, 'timing/data': 0.0005888718878850341, 'timing/model': 1.170678652008064, 'learning_rate/qwen_vl_interface': 9.987976418497974e-06, 'learning_rate/action_model': 9.987976418497972e-05, 'epoch': 5.12}) +[2026-06-20 12:37:25,880][starVLA.training.train_starvla][INFO] - Step 2812, Loss: {'train/loss': 0.2130458503961563, 'train/grad_norm_pre_clip': 16.425195693969727, 'timing/data': 0.0005191819509491324, 'timing/model': 1.186798113049008, 'learning_rate/qwen_vl_interface': 9.987758895045964e-06, 'learning_rate/action_model': 9.987758895045963e-05, 'epoch': 5.12}) +[2026-06-20 12:37:30,624][starVLA.training.train_starvla][INFO] - Step 2813, Loss: {'train/loss': 0.13455979526042938, 'train/grad_norm_pre_clip': 9.501726150512695, 'timing/data': 0.0005462620174512267, 'timing/model': 1.1863731140037999, 'learning_rate/qwen_vl_interface': 9.98753942408023e-06, 'learning_rate/action_model': 9.98753942408023e-05, 'epoch': 5.12}) +[2026-06-20 12:37:35,407][starVLA.training.train_starvla][INFO] - Step 2814, Loss: {'train/loss': 0.19428449869155884, 'train/grad_norm_pre_clip': 15.154706954956055, 'timing/data': 0.0005305620143190026, 'timing/model': 1.4394205230055377, 'learning_rate/qwen_vl_interface': 9.987318005690992e-06, 'learning_rate/action_model': 9.987318005690992e-05, 'epoch': 5.12}) +[2026-06-20 12:37:39,943][starVLA.training.train_starvla][INFO] - Step 2815, Loss: {'train/loss': 0.37850669026374817, 'train/grad_norm_pre_clip': 22.856945037841797, 'timing/data': 0.0005158919375389814, 'timing/model': 1.174683740013279, 'learning_rate/qwen_vl_interface': 9.987094639969268e-06, 'learning_rate/action_model': 9.987094639969268e-05, 'epoch': 5.12}) +[2026-06-20 12:37:44,839][starVLA.training.train_starvla][INFO] - Step 2816, Loss: {'train/loss': 0.2827688455581665, 'train/grad_norm_pre_clip': 16.460289001464844, 'timing/data': 0.0006330820033326745, 'timing/model': 1.1851580509683117, 'learning_rate/qwen_vl_interface': 9.986869327006872e-06, 'learning_rate/action_model': 9.986869327006873e-05, 'epoch': 5.13}) +[2026-06-20 12:37:49,633][starVLA.training.train_starvla][INFO] - Step 2817, Loss: {'train/loss': 0.12412600964307785, 'train/grad_norm_pre_clip': 11.73154067993164, 'timing/data': 0.0008285320363938808, 'timing/model': 1.1694845830788836, 'learning_rate/qwen_vl_interface': 9.986642066896424e-06, 'learning_rate/action_model': 9.986642066896423e-05, 'epoch': 5.13}) +[2026-06-20 12:37:54,552][starVLA.training.train_starvla][INFO] - Step 2818, Loss: {'train/loss': 0.21711690723896027, 'train/grad_norm_pre_clip': 15.417732238769531, 'timing/data': 0.0005393419414758682, 'timing/model': 1.240114816930145, 'learning_rate/qwen_vl_interface': 9.986412859731341e-06, 'learning_rate/action_model': 9.98641285973134e-05, 'epoch': 5.13}) +[2026-06-20 12:37:59,480][starVLA.training.train_starvla][INFO] - Step 2819, Loss: {'train/loss': 0.1414383500814438, 'train/grad_norm_pre_clip': 5.428592205047607, 'timing/data': 0.0005344609962776303, 'timing/model': 1.1641116960672662, 'learning_rate/qwen_vl_interface': 9.986181705605839e-06, 'learning_rate/action_model': 9.986181705605839e-05, 'epoch': 5.13}) +[2026-06-20 12:38:04,252][starVLA.training.train_starvla][INFO] - Step 2820, Loss: {'train/loss': 0.23653832077980042, 'train/grad_norm_pre_clip': 12.32818603515625, 'timing/data': 0.0005270920228213072, 'timing/model': 1.4043775460449979, 'learning_rate/qwen_vl_interface': 9.98594860461494e-06, 'learning_rate/action_model': 9.98594860461494e-05, 'epoch': 5.13}) +[2026-06-20 12:38:08,779][starVLA.training.train_starvla][INFO] - Step 2821, Loss: {'train/loss': 0.1809346228837967, 'train/grad_norm_pre_clip': 6.856722831726074, 'timing/data': 0.0005381309892982244, 'timing/model': 1.168085443903692, 'learning_rate/qwen_vl_interface': 9.985713556854461e-06, 'learning_rate/action_model': 9.98571355685446e-05, 'epoch': 5.14}) +[2026-06-20 12:38:13,534][starVLA.training.train_starvla][INFO] - Step 2822, Loss: {'train/loss': 0.4362505078315735, 'train/grad_norm_pre_clip': 18.88014793395996, 'timing/data': 0.0005352910375222564, 'timing/model': 1.1784974579932168, 'learning_rate/qwen_vl_interface': 9.985476562421024e-06, 'learning_rate/action_model': 9.985476562421024e-05, 'epoch': 5.14}) +[2026-06-20 12:38:18,263][starVLA.training.train_starvla][INFO] - Step 2823, Loss: {'train/loss': 0.16845163702964783, 'train/grad_norm_pre_clip': 11.029733657836914, 'timing/data': 0.0005418609362095594, 'timing/model': 1.175857488065958, 'learning_rate/qwen_vl_interface': 9.985237621412047e-06, 'learning_rate/action_model': 9.985237621412046e-05, 'epoch': 5.14}) +[2026-06-20 12:38:23,012][starVLA.training.train_starvla][INFO] - Step 2824, Loss: {'train/loss': 0.09026007354259491, 'train/grad_norm_pre_clip': 4.610846519470215, 'timing/data': 0.0005349810235202312, 'timing/model': 1.1699847941054031, 'learning_rate/qwen_vl_interface': 9.984996733925749e-06, 'learning_rate/action_model': 9.984996733925748e-05, 'epoch': 5.14}) +[2026-06-20 12:38:27,841][starVLA.training.train_starvla][INFO] - Step 2825, Loss: {'train/loss': 0.11562637984752655, 'train/grad_norm_pre_clip': 14.596061706542969, 'timing/data': 0.0005366209661588073, 'timing/model': 1.19238631404005, 'learning_rate/qwen_vl_interface': 9.98475390006115e-06, 'learning_rate/action_model': 9.98475390006115e-05, 'epoch': 5.14}) +[2026-06-20 12:38:32,619][starVLA.training.train_starvla][INFO] - Step 2826, Loss: {'train/loss': 0.13314004242420197, 'train/grad_norm_pre_clip': 15.44979190826416, 'timing/data': 0.0005211810348555446, 'timing/model': 1.4051742190495133, 'learning_rate/qwen_vl_interface': 9.98450911991807e-06, 'learning_rate/action_model': 9.98450911991807e-05, 'epoch': 5.14}) +[2026-06-20 12:38:37,162][starVLA.training.train_starvla][INFO] - Step 2827, Loss: {'train/loss': 0.29602760076522827, 'train/grad_norm_pre_clip': 14.958196640014648, 'timing/data': 0.0005202509928494692, 'timing/model': 1.1692462710198015, 'learning_rate/qwen_vl_interface': 9.98426239359713e-06, 'learning_rate/action_model': 9.984262393597129e-05, 'epoch': 5.15}) +[2026-06-20 12:38:41,935][starVLA.training.train_starvla][INFO] - Step 2828, Loss: {'train/loss': 0.1486562192440033, 'train/grad_norm_pre_clip': 9.335599899291992, 'timing/data': 0.0005352810258045793, 'timing/model': 1.1879996430361643, 'learning_rate/qwen_vl_interface': 9.984013721199747e-06, 'learning_rate/action_model': 9.984013721199747e-05, 'epoch': 5.15}) +[2026-06-20 12:38:46,718][starVLA.training.train_starvla][INFO] - Step 2829, Loss: {'train/loss': 0.4305303692817688, 'train/grad_norm_pre_clip': 23.03856658935547, 'timing/data': 0.0005145809845998883, 'timing/model': 1.1816495599923655, 'learning_rate/qwen_vl_interface': 9.983763102828143e-06, 'learning_rate/action_model': 9.983763102828143e-05, 'epoch': 5.15}) +[2026-06-20 12:38:51,447][starVLA.training.train_starvla][INFO] - Step 2830, Loss: {'train/loss': 0.22627201676368713, 'train/grad_norm_pre_clip': 8.982501029968262, 'timing/data': 0.0005177110433578491, 'timing/model': 1.1701832260005176, 'learning_rate/qwen_vl_interface': 9.983510538585336e-06, 'learning_rate/action_model': 9.983510538585337e-05, 'epoch': 5.15}) +[2026-06-20 12:38:56,228][starVLA.training.train_starvla][INFO] - Step 2831, Loss: {'train/loss': 0.18712125718593597, 'train/grad_norm_pre_clip': 16.595571517944336, 'timing/data': 0.0005178609862923622, 'timing/model': 1.1715238000033423, 'learning_rate/qwen_vl_interface': 9.98325602857515e-06, 'learning_rate/action_model': 9.98325602857515e-05, 'epoch': 5.15}) +[2026-06-20 12:39:01,023][starVLA.training.train_starvla][INFO] - Step 2832, Loss: {'train/loss': 0.34010106325149536, 'train/grad_norm_pre_clip': 12.277470588684082, 'timing/data': 0.0005151809891685843, 'timing/model': 1.4094179680105299, 'learning_rate/qwen_vl_interface': 9.9829995729022e-06, 'learning_rate/action_model': 9.982999572902199e-05, 'epoch': 5.16}) +[2026-06-20 12:39:05,555][starVLA.training.train_starvla][INFO] - Step 2833, Loss: {'train/loss': 0.13855195045471191, 'train/grad_norm_pre_clip': 10.175836563110352, 'timing/data': 0.0005432110046967864, 'timing/model': 1.171604268020019, 'learning_rate/qwen_vl_interface': 9.982741171671905e-06, 'learning_rate/action_model': 9.982741171671906e-05, 'epoch': 5.16}) +[2026-06-20 12:39:10,327][starVLA.training.train_starvla][INFO] - Step 2834, Loss: {'train/loss': 0.2583894729614258, 'train/grad_norm_pre_clip': 18.159160614013672, 'timing/data': 0.000527941039763391, 'timing/model': 1.1740554150892422, 'learning_rate/qwen_vl_interface': 9.982480824990489e-06, 'learning_rate/action_model': 9.982480824990488e-05, 'epoch': 5.16}) +[2026-06-20 12:39:15,083][starVLA.training.train_starvla][INFO] - Step 2835, Loss: {'train/loss': 0.15902218222618103, 'train/grad_norm_pre_clip': 9.333231925964355, 'timing/data': 0.0007270309142768383, 'timing/model': 1.1858600819250569, 'learning_rate/qwen_vl_interface': 9.982218532964963e-06, 'learning_rate/action_model': 9.982218532964964e-05, 'epoch': 5.16}) +[2026-06-20 12:39:19,864][starVLA.training.train_starvla][INFO] - Step 2836, Loss: {'train/loss': 0.1557912528514862, 'train/grad_norm_pre_clip': 7.730020523071289, 'timing/data': 0.0005097710527479649, 'timing/model': 1.174022743012756, 'learning_rate/qwen_vl_interface': 9.981954295703152e-06, 'learning_rate/action_model': 9.981954295703152e-05, 'epoch': 5.16}) +[2026-06-20 12:39:24,647][starVLA.training.train_starvla][INFO] - Step 2837, Loss: {'train/loss': 0.2576598823070526, 'train/grad_norm_pre_clip': 13.42088794708252, 'timing/data': 0.0005158110288903117, 'timing/model': 1.17411554697901, 'learning_rate/qwen_vl_interface': 9.981688113313672e-06, 'learning_rate/action_model': 9.981688113313672e-05, 'epoch': 5.16}) +[2026-06-20 12:39:29,437][starVLA.training.train_starvla][INFO] - Step 2838, Loss: {'train/loss': 0.35125210881233215, 'train/grad_norm_pre_clip': 14.755807876586914, 'timing/data': 0.0006033909739926457, 'timing/model': 1.4227099979761988, 'learning_rate/qwen_vl_interface': 9.98141998590594e-06, 'learning_rate/action_model': 9.981419985905939e-05, 'epoch': 5.17}) +[2026-06-20 12:39:34,169][starVLA.training.train_starvla][INFO] - Step 2839, Loss: {'train/loss': 0.10559254884719849, 'train/grad_norm_pre_clip': 4.230405807495117, 'timing/data': 0.0006639410275965929, 'timing/model': 1.3628108779666945, 'learning_rate/qwen_vl_interface': 9.981149913590171e-06, 'learning_rate/action_model': 9.981149913590172e-05, 'epoch': 5.17}) +[2026-06-20 12:39:38,740][starVLA.training.train_starvla][INFO] - Step 2840, Loss: {'train/loss': 0.2402823269367218, 'train/grad_norm_pre_clip': 8.83093547821045, 'timing/data': 0.0005855700001120567, 'timing/model': 1.1991885469760746, 'learning_rate/qwen_vl_interface': 9.980877896477386e-06, 'learning_rate/action_model': 9.980877896477386e-05, 'epoch': 5.17}) +[2026-06-20 12:39:43,524][starVLA.training.train_starvla][INFO] - Step 2841, Loss: {'train/loss': 0.17867574095726013, 'train/grad_norm_pre_clip': 9.48828411102295, 'timing/data': 0.000528520904481411, 'timing/model': 1.1888177840737626, 'learning_rate/qwen_vl_interface': 9.980603934679397e-06, 'learning_rate/action_model': 9.980603934679398e-05, 'epoch': 5.17}) +[2026-06-20 12:39:48,307][starVLA.training.train_starvla][INFO] - Step 2842, Loss: {'train/loss': 0.10665042698383331, 'train/grad_norm_pre_clip': 6.64341402053833, 'timing/data': 0.0005107809556648135, 'timing/model': 1.1809850169811398, 'learning_rate/qwen_vl_interface': 9.980328028308824e-06, 'learning_rate/action_model': 9.980328028308824e-05, 'epoch': 5.17}) +[2026-06-20 12:39:53,073][starVLA.training.train_starvla][INFO] - Step 2843, Loss: {'train/loss': 0.1784052550792694, 'train/grad_norm_pre_clip': 7.96221923828125, 'timing/data': 0.0005207509966567159, 'timing/model': 1.1720367489615455, 'learning_rate/qwen_vl_interface': 9.980050177479075e-06, 'learning_rate/action_model': 9.980050177479076e-05, 'epoch': 5.18}) +[2026-06-20 12:39:57,851][starVLA.training.train_starvla][INFO] - Step 2844, Loss: {'train/loss': 0.2214810699224472, 'train/grad_norm_pre_clip': 14.387094497680664, 'timing/data': 0.0005893909838050604, 'timing/model': 1.4142363190185279, 'learning_rate/qwen_vl_interface': 9.97977038230437e-06, 'learning_rate/action_model': 9.97977038230437e-05, 'epoch': 5.18}) +[2026-06-20 12:40:02,634][starVLA.training.train_starvla][INFO] - Step 2845, Loss: {'train/loss': 0.5433669090270996, 'train/grad_norm_pre_clip': 19.099140167236328, 'timing/data': 0.0007209610193967819, 'timing/model': 1.4165768598904833, 'learning_rate/qwen_vl_interface': 9.97948864289972e-06, 'learning_rate/action_model': 9.97948864289972e-05, 'epoch': 5.18}) +[2026-06-20 12:40:07,369][starVLA.training.train_starvla][INFO] - Step 2846, Loss: {'train/loss': 0.29239949584007263, 'train/grad_norm_pre_clip': 14.93934440612793, 'timing/data': 0.0006659500068053603, 'timing/model': 1.1888918870827183, 'learning_rate/qwen_vl_interface': 9.979204959380936e-06, 'learning_rate/action_model': 9.979204959380936e-05, 'epoch': 5.18}) +[2026-06-20 12:40:12,137][starVLA.training.train_starvla][INFO] - Step 2847, Loss: {'train/loss': 0.3157753050327301, 'train/grad_norm_pre_clip': 15.026482582092285, 'timing/data': 0.0005262310151010752, 'timing/model': 1.168379732989706, 'learning_rate/qwen_vl_interface': 9.978919331864631e-06, 'learning_rate/action_model': 9.97891933186463e-05, 'epoch': 5.18}) +[2026-06-20 12:40:16,931][starVLA.training.train_starvla][INFO] - Step 2848, Loss: {'train/loss': 0.17582902312278748, 'train/grad_norm_pre_clip': 11.821761131286621, 'timing/data': 0.0005259009776636958, 'timing/model': 1.1814015429699793, 'learning_rate/qwen_vl_interface': 9.978631760468216e-06, 'learning_rate/action_model': 9.978631760468216e-05, 'epoch': 5.18}) +[2026-06-20 12:40:21,725][starVLA.training.train_starvla][INFO] - Step 2849, Loss: {'train/loss': 0.29372042417526245, 'train/grad_norm_pre_clip': 15.071139335632324, 'timing/data': 0.0005125700263306499, 'timing/model': 1.1765307440655306, 'learning_rate/qwen_vl_interface': 9.9783422453099e-06, 'learning_rate/action_model': 9.9783422453099e-05, 'epoch': 5.19}) +[2026-06-20 12:40:26,544][starVLA.training.train_starvla][INFO] - Step 2850, Loss: {'train/loss': 0.48226577043533325, 'train/grad_norm_pre_clip': 15.649797439575195, 'timing/data': 0.0005877200746908784, 'timing/model': 1.4500793379265815, 'learning_rate/qwen_vl_interface': 9.978050786508693e-06, 'learning_rate/action_model': 9.978050786508692e-05, 'epoch': 5.19}) +[2026-06-20 12:40:31,114][starVLA.training.train_starvla][INFO] - Step 2851, Loss: {'train/loss': 0.37262672185897827, 'train/grad_norm_pre_clip': 9.185877799987793, 'timing/data': 0.0007769509684294462, 'timing/model': 1.1986569740111008, 'learning_rate/qwen_vl_interface': 9.9777573841844e-06, 'learning_rate/action_model': 9.977757384184401e-05, 'epoch': 5.19}) +[2026-06-20 12:40:35,949][starVLA.training.train_starvla][INFO] - Step 2852, Loss: {'train/loss': 0.14263132214546204, 'train/grad_norm_pre_clip': 10.354263305664062, 'timing/data': 0.0005563109880313277, 'timing/model': 1.1872336579253897, 'learning_rate/qwen_vl_interface': 9.97746203845763e-06, 'learning_rate/action_model': 9.97746203845763e-05, 'epoch': 5.19}) +[2026-06-20 12:40:40,759][starVLA.training.train_starvla][INFO] - Step 2853, Loss: {'train/loss': 0.3693561553955078, 'train/grad_norm_pre_clip': 11.831499099731445, 'timing/data': 0.0005470110336318612, 'timing/model': 1.184919674997218, 'learning_rate/qwen_vl_interface': 9.97716474944979e-06, 'learning_rate/action_model': 9.977164749449788e-05, 'epoch': 5.19}) +[2026-06-20 12:40:45,665][starVLA.training.train_starvla][INFO] - Step 2854, Loss: {'train/loss': 0.16230522096157074, 'train/grad_norm_pre_clip': 8.699881553649902, 'timing/data': 0.0005347100086510181, 'timing/model': 1.2399449739605188, 'learning_rate/qwen_vl_interface': 9.976865517283079e-06, 'learning_rate/action_model': 9.976865517283078e-05, 'epoch': 5.2}) +[2026-06-20 12:40:50,456][starVLA.training.train_starvla][INFO] - Step 2855, Loss: {'train/loss': 0.2855905294418335, 'train/grad_norm_pre_clip': 15.416592597961426, 'timing/data': 0.0005410910816863179, 'timing/model': 1.1741502600489184, 'learning_rate/qwen_vl_interface': 9.976564342080505e-06, 'learning_rate/action_model': 9.976564342080504e-05, 'epoch': 5.2}) +[2026-06-20 12:40:55,336][starVLA.training.train_starvla][INFO] - Step 2856, Loss: {'train/loss': 0.25598374009132385, 'train/grad_norm_pre_clip': 9.67849063873291, 'timing/data': 0.0005384599789977074, 'timing/model': 1.4496821139473468, 'learning_rate/qwen_vl_interface': 9.976261223965867e-06, 'learning_rate/action_model': 9.976261223965865e-05, 'epoch': 5.2}) +[2026-06-20 12:40:59,999][starVLA.training.train_starvla][INFO] - Step 2857, Loss: {'train/loss': 0.15010693669319153, 'train/grad_norm_pre_clip': 10.245810508728027, 'timing/data': 0.0005421310197561979, 'timing/model': 1.2436967469984666, 'learning_rate/qwen_vl_interface': 9.975956163063767e-06, 'learning_rate/action_model': 9.975956163063767e-05, 'epoch': 5.2}) +[2026-06-20 12:41:04,791][starVLA.training.train_starvla][INFO] - Step 2858, Loss: {'train/loss': 0.2596295177936554, 'train/grad_norm_pre_clip': 9.954523086547852, 'timing/data': 0.0005138199776411057, 'timing/model': 1.1899722249945626, 'learning_rate/qwen_vl_interface': 9.975649159499605e-06, 'learning_rate/action_model': 9.975649159499603e-05, 'epoch': 5.2}) +[2026-06-20 12:41:09,587][starVLA.training.train_starvla][INFO] - Step 2859, Loss: {'train/loss': 0.1634996235370636, 'train/grad_norm_pre_clip': 3.7743828296661377, 'timing/data': 0.0005184210604056716, 'timing/model': 1.1843554399674758, 'learning_rate/qwen_vl_interface': 9.975340213399573e-06, 'learning_rate/action_model': 9.975340213399573e-05, 'epoch': 5.2}) +[2026-06-20 12:41:14,350][starVLA.training.train_starvla][INFO] - Step 2860, Loss: {'train/loss': 0.17782801389694214, 'train/grad_norm_pre_clip': 6.620753765106201, 'timing/data': 0.0005802310770377517, 'timing/model': 1.185230762930587, 'learning_rate/qwen_vl_interface': 9.975029324890674e-06, 'learning_rate/action_model': 9.975029324890674e-05, 'epoch': 5.21}) +[2026-06-20 12:41:19,127][starVLA.training.train_starvla][INFO] - Step 2861, Loss: {'train/loss': 0.10434567183256149, 'train/grad_norm_pre_clip': 4.750254154205322, 'timing/data': 0.0005838810466229916, 'timing/model': 1.1925272130174562, 'learning_rate/qwen_vl_interface': 9.974716494100698e-06, 'learning_rate/action_model': 9.974716494100697e-05, 'epoch': 5.21}) +[2026-06-20 12:41:23,903][starVLA.training.train_starvla][INFO] - Step 2862, Loss: {'train/loss': 0.26895439624786377, 'train/grad_norm_pre_clip': 9.412445068359375, 'timing/data': 0.0005773910088464618, 'timing/model': 1.406223596073687, 'learning_rate/qwen_vl_interface': 9.97440172115824e-06, 'learning_rate/action_model': 9.974401721158239e-05, 'epoch': 5.21}) +[2026-06-20 12:41:28,447][starVLA.training.train_starvla][INFO] - Step 2863, Loss: {'train/loss': 0.19742752611637115, 'train/grad_norm_pre_clip': 7.477267742156982, 'timing/data': 0.0005728208925575018, 'timing/model': 1.1759929460240528, 'learning_rate/qwen_vl_interface': 9.97408500619269e-06, 'learning_rate/action_model': 9.97408500619269e-05, 'epoch': 5.21}) +[2026-06-20 12:41:33,265][starVLA.training.train_starvla][INFO] - Step 2864, Loss: {'train/loss': 0.12508590519428253, 'train/grad_norm_pre_clip': 8.544050216674805, 'timing/data': 0.0005435399943962693, 'timing/model': 1.1788102119462565, 'learning_rate/qwen_vl_interface': 9.973766349334239e-06, 'learning_rate/action_model': 9.973766349334239e-05, 'epoch': 5.21}) +[2026-06-20 12:41:38,097][starVLA.training.train_starvla][INFO] - Step 2865, Loss: {'train/loss': 0.21281698346138, 'train/grad_norm_pre_clip': 6.911544322967529, 'timing/data': 0.0005272200796753168, 'timing/model': 1.1772830459522083, 'learning_rate/qwen_vl_interface': 9.973445750713875e-06, 'learning_rate/action_model': 9.973445750713875e-05, 'epoch': 5.22}) +[2026-06-20 12:41:42,907][starVLA.training.train_starvla][INFO] - Step 2866, Loss: {'train/loss': 0.1055869311094284, 'train/grad_norm_pre_clip': 5.327216625213623, 'timing/data': 0.0005245200591161847, 'timing/model': 1.1778577010845765, 'learning_rate/qwen_vl_interface': 9.97312321046338e-06, 'learning_rate/action_model': 9.97312321046338e-05, 'epoch': 5.22}) +[2026-06-20 12:41:47,748][starVLA.training.train_starvla][INFO] - Step 2867, Loss: {'train/loss': 0.23659026622772217, 'train/grad_norm_pre_clip': 11.568239212036133, 'timing/data': 0.0005323300138115883, 'timing/model': 1.4533150249626487, 'learning_rate/qwen_vl_interface': 9.972798728715343e-06, 'learning_rate/action_model': 9.972798728715344e-05, 'epoch': 5.22}) +[2026-06-20 12:41:52,328][starVLA.training.train_starvla][INFO] - Step 2868, Loss: {'train/loss': 0.7930717468261719, 'train/grad_norm_pre_clip': 20.228052139282227, 'timing/data': 0.0005435000639408827, 'timing/model': 1.182526177028194, 'learning_rate/qwen_vl_interface': 9.972472305603145e-06, 'learning_rate/action_model': 9.972472305603144e-05, 'epoch': 5.22}) +[2026-06-20 12:41:57,149][starVLA.training.train_starvla][INFO] - Step 2869, Loss: {'train/loss': 0.27420541644096375, 'train/grad_norm_pre_clip': 15.058921813964844, 'timing/data': 0.0005235300632193685, 'timing/model': 1.198877974995412, 'learning_rate/qwen_vl_interface': 9.972143941260965e-06, 'learning_rate/action_model': 9.972143941260965e-05, 'epoch': 5.22}) +[2026-06-20 12:42:01,989][starVLA.training.train_starvla][INFO] - Step 2870, Loss: {'train/loss': 0.4792429208755493, 'train/grad_norm_pre_clip': 14.344749450683594, 'timing/data': 0.000535380095243454, 'timing/model': 1.1748040809761733, 'learning_rate/qwen_vl_interface': 9.971813635823781e-06, 'learning_rate/action_model': 9.971813635823781e-05, 'epoch': 5.22}) +[2026-06-20 12:42:06,758][starVLA.training.train_starvla][INFO] - Step 2871, Loss: {'train/loss': 0.35492631793022156, 'train/grad_norm_pre_clip': 26.35102653503418, 'timing/data': 0.0005494500510394573, 'timing/model': 1.17895434203092, 'learning_rate/qwen_vl_interface': 9.971481389427372e-06, 'learning_rate/action_model': 9.971481389427371e-05, 'epoch': 5.23}) +[2026-06-20 12:42:12,061][starVLA.training.train_starvla][INFO] - Step 2872, Loss: {'train/loss': 0.7534573674201965, 'train/grad_norm_pre_clip': 31.110923767089844, 'timing/data': 0.0012034709798172116, 'timing/model': 1.3807551809586585, 'learning_rate/qwen_vl_interface': 9.971147202208308e-06, 'learning_rate/action_model': 9.971147202208308e-05, 'epoch': 5.23}) +[2026-06-20 12:42:17,227][starVLA.training.train_starvla][INFO] - Step 2873, Loss: {'train/loss': 0.285905659198761, 'train/grad_norm_pre_clip': 15.765110969543457, 'timing/data': 0.0005408300785347819, 'timing/model': 1.4366321549750865, 'learning_rate/qwen_vl_interface': 9.970811074303962e-06, 'learning_rate/action_model': 9.970811074303962e-05, 'epoch': 5.23}) +[2026-06-20 12:42:21,837][starVLA.training.train_starvla][INFO] - Step 2874, Loss: {'train/loss': 0.1447981894016266, 'train/grad_norm_pre_clip': 6.197227954864502, 'timing/data': 0.0005569200729951262, 'timing/model': 1.1858836920000613, 'learning_rate/qwen_vl_interface': 9.970473005852505e-06, 'learning_rate/action_model': 9.970473005852505e-05, 'epoch': 5.23}) +[2026-06-20 12:42:27,551][starVLA.training.train_starvla][INFO] - Step 2875, Loss: {'train/loss': 0.15466450154781342, 'train/grad_norm_pre_clip': 6.892392158508301, 'timing/data': 0.0005597699200734496, 'timing/model': 1.176775227067992, 'learning_rate/qwen_vl_interface': 9.970132996992904e-06, 'learning_rate/action_model': 9.970132996992904e-05, 'epoch': 5.23}) +[2026-06-20 12:42:32,379][starVLA.training.train_starvla][INFO] - Step 2876, Loss: {'train/loss': 0.34206998348236084, 'train/grad_norm_pre_clip': 9.58084487915039, 'timing/data': 0.0005223810439929366, 'timing/model': 1.1707219809759408, 'learning_rate/qwen_vl_interface': 9.969791047864921e-06, 'learning_rate/action_model': 9.969791047864921e-05, 'epoch': 5.24}) +[2026-06-20 12:42:37,186][starVLA.training.train_starvla][INFO] - Step 2877, Loss: {'train/loss': 0.35798707604408264, 'train/grad_norm_pre_clip': 45.370872497558594, 'timing/data': 0.0005354300374165177, 'timing/model': 1.1959745839703828, 'learning_rate/qwen_vl_interface': 9.969447158609122e-06, 'learning_rate/action_model': 9.96944715860912e-05, 'epoch': 5.24}) +[2026-06-20 12:42:42,332][starVLA.training.train_starvla][INFO] - Step 2878, Loss: {'train/loss': 0.2141696661710739, 'train/grad_norm_pre_clip': 6.886309623718262, 'timing/data': 0.0007361200405284762, 'timing/model': 1.1931690579513088, 'learning_rate/qwen_vl_interface': 9.969101329366864e-06, 'learning_rate/action_model': 9.969101329366863e-05, 'epoch': 5.24}) +[2026-06-20 12:42:47,831][starVLA.training.train_starvla][INFO] - Step 2879, Loss: {'train/loss': 0.26228344440460205, 'train/grad_norm_pre_clip': 8.578025817871094, 'timing/data': 0.000608351081609726, 'timing/model': 1.4522226160624996, 'learning_rate/qwen_vl_interface': 9.968753560280305e-06, 'learning_rate/action_model': 9.968753560280304e-05, 'epoch': 5.24}) +[2026-06-20 12:42:52,512][starVLA.training.train_starvla][INFO] - Step 2880, Loss: {'train/loss': 0.2279382050037384, 'train/grad_norm_pre_clip': 10.032456398010254, 'timing/data': 0.0005299900658428669, 'timing/model': 1.2221266659907997, 'learning_rate/qwen_vl_interface': 9.968403851492402e-06, 'learning_rate/action_model': 9.968403851492401e-05, 'epoch': 5.24}) +[2026-06-20 12:42:57,431][starVLA.training.train_starvla][INFO] - Step 2881, Loss: {'train/loss': 0.290778249502182, 'train/grad_norm_pre_clip': 14.509855270385742, 'timing/data': 0.0005228800000622869, 'timing/model': 1.1769611430354416, 'learning_rate/qwen_vl_interface': 9.968052203146904e-06, 'learning_rate/action_model': 9.968052203146903e-05, 'epoch': 5.24}) +[2026-06-20 12:43:02,557][starVLA.training.train_starvla][INFO] - Step 2882, Loss: {'train/loss': 0.1216619685292244, 'train/grad_norm_pre_clip': 5.199209213256836, 'timing/data': 0.0005157500272616744, 'timing/model': 1.1691017169505358, 'learning_rate/qwen_vl_interface': 9.967698615388363e-06, 'learning_rate/action_model': 9.967698615388362e-05, 'epoch': 5.25}) +[2026-06-20 12:43:07,954][starVLA.training.train_starvla][INFO] - Step 2883, Loss: {'train/loss': 0.3422481417655945, 'train/grad_norm_pre_clip': 18.247278213500977, 'timing/data': 0.0005456010112538934, 'timing/model': 1.1950329539831728, 'learning_rate/qwen_vl_interface': 9.967343088362123e-06, 'learning_rate/action_model': 9.967343088362122e-05, 'epoch': 5.25}) +[2026-06-20 12:43:12,759][starVLA.training.train_starvla][INFO] - Step 2884, Loss: {'train/loss': 0.12774182856082916, 'train/grad_norm_pre_clip': 5.8553972244262695, 'timing/data': 0.0008442699909210205, 'timing/model': 1.4484229710651562, 'learning_rate/qwen_vl_interface': 9.966985622214328e-06, 'learning_rate/action_model': 9.966985622214329e-05, 'epoch': 5.25}) +[2026-06-20 12:43:17,665][starVLA.training.train_starvla][INFO] - Step 2885, Loss: {'train/loss': 0.31978586316108704, 'train/grad_norm_pre_clip': 19.1691837310791, 'timing/data': 0.0005116709508001804, 'timing/model': 1.3832166700158268, 'learning_rate/qwen_vl_interface': 9.966626217091922e-06, 'learning_rate/action_model': 9.966626217091922e-05, 'epoch': 5.25}) +[2026-06-20 12:43:22,759][starVLA.training.train_starvla][INFO] - Step 2886, Loss: {'train/loss': 0.19421522319316864, 'train/grad_norm_pre_clip': 7.243786811828613, 'timing/data': 0.0005404700059443712, 'timing/model': 1.2194572109729052, 'learning_rate/qwen_vl_interface': 9.966264873142641e-06, 'learning_rate/action_model': 9.96626487314264e-05, 'epoch': 5.25}) +[2026-06-20 12:43:27,636][starVLA.training.train_starvla][INFO] - Step 2887, Loss: {'train/loss': 0.3678395748138428, 'train/grad_norm_pre_clip': 10.185614585876465, 'timing/data': 0.0005839000223204494, 'timing/model': 1.223188761039637, 'learning_rate/qwen_vl_interface': 9.965901590515018e-06, 'learning_rate/action_model': 9.965901590515018e-05, 'epoch': 5.26}) +[2026-06-20 12:43:32,679][starVLA.training.train_starvla][INFO] - Step 2888, Loss: {'train/loss': 0.14923709630966187, 'train/grad_norm_pre_clip': 9.933525085449219, 'timing/data': 0.0005893809720873833, 'timing/model': 1.181496528093703, 'learning_rate/qwen_vl_interface': 9.965536369358386e-06, 'learning_rate/action_model': 9.965536369358386e-05, 'epoch': 5.26}) +[2026-06-20 12:43:37,835][starVLA.training.train_starvla][INFO] - Step 2889, Loss: {'train/loss': 0.2992059290409088, 'train/grad_norm_pre_clip': 17.698406219482422, 'timing/data': 0.0005131600191816688, 'timing/model': 1.173906594980508, 'learning_rate/qwen_vl_interface': 9.965169209822874e-06, 'learning_rate/action_model': 9.965169209822874e-05, 'epoch': 5.26}) +[2026-06-20 12:43:43,040][starVLA.training.train_starvla][INFO] - Step 2890, Loss: {'train/loss': 0.17916016280651093, 'train/grad_norm_pre_clip': 8.874777793884277, 'timing/data': 0.0005102999275550246, 'timing/model': 1.7746175449574366, 'learning_rate/qwen_vl_interface': 9.96480011205941e-06, 'learning_rate/action_model': 9.964800112059409e-05, 'epoch': 5.26}) +[2026-06-20 12:43:47,844][starVLA.training.train_starvla][INFO] - Step 2891, Loss: {'train/loss': 0.3454863131046295, 'train/grad_norm_pre_clip': 15.847078323364258, 'timing/data': 0.0005164089379832149, 'timing/model': 1.3998476041015238, 'learning_rate/qwen_vl_interface': 9.964429076219711e-06, 'learning_rate/action_model': 9.964429076219711e-05, 'epoch': 5.26}) +[2026-06-20 12:43:53,260][starVLA.training.train_starvla][INFO] - Step 2892, Loss: {'train/loss': 0.17109152674674988, 'train/grad_norm_pre_clip': 7.7013983726501465, 'timing/data': 0.000591559917666018, 'timing/model': 1.7915688889333978, 'learning_rate/qwen_vl_interface': 9.964056102456299e-06, 'learning_rate/action_model': 9.964056102456299e-05, 'epoch': 5.26}) +[2026-06-20 12:43:58,007][starVLA.training.train_starvla][INFO] - Step 2893, Loss: {'train/loss': 0.2255840301513672, 'train/grad_norm_pre_clip': 20.132898330688477, 'timing/data': 0.0005891310283914208, 'timing/model': 1.1927345829317346, 'learning_rate/qwen_vl_interface': 9.963681190922489e-06, 'learning_rate/action_model': 9.963681190922488e-05, 'epoch': 5.27}) +[2026-06-20 12:44:02,770][starVLA.training.train_starvla][INFO] - Step 2894, Loss: {'train/loss': 0.1633952111005783, 'train/grad_norm_pre_clip': 13.70093822479248, 'timing/data': 0.0005974399391561747, 'timing/model': 1.19351152703166, 'learning_rate/qwen_vl_interface': 9.963304341772391e-06, 'learning_rate/action_model': 9.963304341772391e-05, 'epoch': 5.27}) +[2026-06-20 12:44:07,548][starVLA.training.train_starvla][INFO] - Step 2895, Loss: {'train/loss': 0.3455149531364441, 'train/grad_norm_pre_clip': 9.560409545898438, 'timing/data': 0.0005774699384346604, 'timing/model': 1.1697518969886005, 'learning_rate/qwen_vl_interface': 9.962925555160919e-06, 'learning_rate/action_model': 9.962925555160918e-05, 'epoch': 5.27}) +[2026-06-20 12:44:12,401][starVLA.training.train_starvla][INFO] - Step 2896, Loss: {'train/loss': 0.19806243479251862, 'train/grad_norm_pre_clip': 18.255884170532227, 'timing/data': 0.0006013009697198868, 'timing/model': 1.4642126959515736, 'learning_rate/qwen_vl_interface': 9.96254483124377e-06, 'learning_rate/action_model': 9.96254483124377e-05, 'epoch': 5.27}) +[2026-06-20 12:44:17,023][starVLA.training.train_starvla][INFO] - Step 2897, Loss: {'train/loss': 0.14015811681747437, 'train/grad_norm_pre_clip': 13.955283164978027, 'timing/data': 0.0005198309663683176, 'timing/model': 1.2266099089756608, 'learning_rate/qwen_vl_interface': 9.96216217017745e-06, 'learning_rate/action_model': 9.962162170177451e-05, 'epoch': 5.27}) +[2026-06-20 12:44:21,891][starVLA.training.train_starvla][INFO] - Step 2898, Loss: {'train/loss': 0.43943628668785095, 'train/grad_norm_pre_clip': 16.046409606933594, 'timing/data': 0.0005784400273114443, 'timing/model': 1.22443271195516, 'learning_rate/qwen_vl_interface': 9.961777572119258e-06, 'learning_rate/action_model': 9.961777572119258e-05, 'epoch': 5.28}) +[2026-06-20 12:44:26,772][starVLA.training.train_starvla][INFO] - Step 2899, Loss: {'train/loss': 0.17653271555900574, 'train/grad_norm_pre_clip': 11.67265510559082, 'timing/data': 0.0005271700210869312, 'timing/model': 1.2150305249961093, 'learning_rate/qwen_vl_interface': 9.961391037227286e-06, 'learning_rate/action_model': 9.961391037227284e-05, 'epoch': 5.28}) +[2026-06-20 12:44:31,881][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 2900 +[2026-06-20 12:50:59,958][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 2900 +[2026-06-20 12:50:59,959][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 2900 +[2026-06-20 12:57:50,799][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 2900 +[2026-06-20 12:57:50,801][starVLA.training.train_starvla][INFO] - Step 2900, Loss: {'train/loss': 0.22193565964698792, 'train/grad_norm_pre_clip': 12.546204566955566, 'eval/loss': 0.35967332124710083, 'eval/latency_0/loss': 0.3792562484741211, 'eval/latency_1/loss': 0.4502732753753662, 'eval/latency_2/loss': 0.4382082521915436, 'eval/latency_3/loss': 0.11563324183225632, 'eval/latency_4/loss': 0.38881465792655945, 'eval/flappy/loss': 0.35967332124710083, 'eval/flappy/latency_0/loss': 0.3792562484741211, 'eval/flappy/latency_1/loss': 0.4502732753753662, 'eval/flappy/latency_2/loss': 0.4382082521915436, 'eval/flappy/latency_3/loss': 0.11563324183225632, 'eval/flappy/latency_4/loss': 0.38881465792655945, 'eval/flap_precision': 0.16352201257861634, 'eval/flap_recall': 0.09454545454545454, 'eval/flap_f1': 0.11981566820276497, 'eval/flap_rate_model': 0.04416666666666667, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.11981566820276497, 'eval/latency_0/flap_precision': 0.16352201257861634, 'eval/latency_0/flap_recall': 0.09454545454545454, 'eval/latency_0/flap_f1': 0.11981566820276497, 'eval/latency_0/flap_rate_model': 0.04416666666666667, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.11981566820276497, 'timing/data': 0.0012840699637308717, 'timing/model': 1.3827720979461446, 'learning_rate/qwen_vl_interface': 9.961002565660419e-06, 'learning_rate/action_model': 9.961002565660419e-05, 'epoch': 5.28}) +[2026-06-20 12:57:50,802][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_2900_state +[2026-06-20 12:57:50,805][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-20 12:57:59,849][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_2900_state/model.safetensors +[2026-06-20 12:58:13,710][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_2900_state/optimizer.bin +[2026-06-20 12:58:13,711][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_2900_state/sampler.bin +[2026-06-20 12:58:13,711][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_2900_state/sampler_1.bin +[2026-06-20 12:58:13,712][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_2900_state/random_states_0.pkl +[2026-06-20 12:58:15,712][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2900 +[2026-06-20 12:58:15,713][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-20 12:58:15,726][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-20 12:58:20,477][starVLA.training.train_starvla][INFO] - Step 2901, Loss: {'train/loss': 0.08485101163387299, 'train/grad_norm_pre_clip': 8.112321853637695, 'timing/data': 0.0005397299537435174, 'timing/model': 1.1589841219829395, 'learning_rate/qwen_vl_interface': 9.96061215757835e-06, 'learning_rate/action_model': 9.960612157578349e-05, 'epoch': 5.28}) +[2026-06-20 12:58:25,334][starVLA.training.train_starvla][INFO] - Step 2902, Loss: {'train/loss': 0.05349842086434364, 'train/grad_norm_pre_clip': 3.314965009689331, 'timing/data': 0.0005619489820674062, 'timing/model': 1.2241199569543824, 'learning_rate/qwen_vl_interface': 9.96021981314156e-06, 'learning_rate/action_model': 9.96021981314156e-05, 'epoch': 5.28}) +[2026-06-20 12:58:30,160][starVLA.training.train_starvla][INFO] - Step 2903, Loss: {'train/loss': 0.34999269247055054, 'train/grad_norm_pre_clip': 14.934783935546875, 'timing/data': 0.0005380589282140136, 'timing/model': 1.460109717096202, 'learning_rate/qwen_vl_interface': 9.959825532511322e-06, 'learning_rate/action_model': 9.959825532511322e-05, 'epoch': 5.28}) +[2026-06-20 12:58:34,765][starVLA.training.train_starvla][INFO] - Step 2904, Loss: {'train/loss': 0.19956538081169128, 'train/grad_norm_pre_clip': 12.722735404968262, 'timing/data': 0.0005452090408653021, 'timing/model': 1.2273038809653372, 'learning_rate/qwen_vl_interface': 9.959429315849715e-06, 'learning_rate/action_model': 9.959429315849715e-05, 'epoch': 5.29}) +[2026-06-20 12:58:39,720][starVLA.training.train_starvla][INFO] - Step 2905, Loss: {'train/loss': 0.12098637968301773, 'train/grad_norm_pre_clip': 20.417259216308594, 'timing/data': 0.0005384590476751328, 'timing/model': 1.2338700640248135, 'learning_rate/qwen_vl_interface': 9.959031163319606e-06, 'learning_rate/action_model': 9.959031163319607e-05, 'epoch': 5.29}) +[2026-06-20 12:58:44,604][starVLA.training.train_starvla][INFO] - Step 2906, Loss: {'train/loss': 0.1540135145187378, 'train/grad_norm_pre_clip': 17.350549697875977, 'timing/data': 0.0005329590057954192, 'timing/model': 1.2322871719952673, 'learning_rate/qwen_vl_interface': 9.958631075084664e-06, 'learning_rate/action_model': 9.958631075084664e-05, 'epoch': 5.29}) +[2026-06-20 12:58:49,466][starVLA.training.train_starvla][INFO] - Step 2907, Loss: {'train/loss': 0.1652112454175949, 'train/grad_norm_pre_clip': 13.51411247253418, 'timing/data': 0.0005324389785528183, 'timing/model': 1.1673880169400945, 'learning_rate/qwen_vl_interface': 9.958229051309346e-06, 'learning_rate/action_model': 9.958229051309346e-05, 'epoch': 5.29}) +[2026-06-20 12:58:54,267][starVLA.training.train_starvla][INFO] - Step 2908, Loss: {'train/loss': 0.052737146615982056, 'train/grad_norm_pre_clip': 4.556467533111572, 'timing/data': 0.0005160690052434802, 'timing/model': 1.470852316939272, 'learning_rate/qwen_vl_interface': 9.95782509215891e-06, 'learning_rate/action_model': 9.95782509215891e-05, 'epoch': 5.29}) +[2026-06-20 12:58:58,887][starVLA.training.train_starvla][INFO] - Step 2909, Loss: {'train/loss': 0.18513648211956024, 'train/grad_norm_pre_clip': 13.107041358947754, 'timing/data': 0.0005137689877301455, 'timing/model': 1.1713899109745398, 'learning_rate/qwen_vl_interface': 9.95741919779941e-06, 'learning_rate/action_model': 9.95741919779941e-05, 'epoch': 5.3}) +[2026-06-20 12:59:03,797][starVLA.training.train_starvla][INFO] - Step 2910, Loss: {'train/loss': 0.22629518806934357, 'train/grad_norm_pre_clip': 7.9281816482543945, 'timing/data': 0.0005043790442869067, 'timing/model': 1.2195305060595274, 'learning_rate/qwen_vl_interface': 9.957011368397692e-06, 'learning_rate/action_model': 9.95701136839769e-05, 'epoch': 5.3}) +[2026-06-20 12:59:08,584][starVLA.training.train_starvla][INFO] - Step 2911, Loss: {'train/loss': 0.052024852484464645, 'train/grad_norm_pre_clip': 4.265850067138672, 'timing/data': 0.0005231490358710289, 'timing/model': 1.186110523995012, 'learning_rate/qwen_vl_interface': 9.956601604121398e-06, 'learning_rate/action_model': 9.956601604121397e-05, 'epoch': 5.3}) +[2026-06-20 12:59:13,448][starVLA.training.train_starvla][INFO] - Step 2912, Loss: {'train/loss': 0.19069327414035797, 'train/grad_norm_pre_clip': 21.205814361572266, 'timing/data': 0.0005099389236420393, 'timing/model': 1.2296229869825765, 'learning_rate/qwen_vl_interface': 9.95618990513897e-06, 'learning_rate/action_model': 9.95618990513897e-05, 'epoch': 5.3}) +[2026-06-20 12:59:18,263][starVLA.training.train_starvla][INFO] - Step 2913, Loss: {'train/loss': 0.16380912065505981, 'train/grad_norm_pre_clip': 13.260992050170898, 'timing/data': 0.0005023989360779524, 'timing/model': 1.1809973650379106, 'learning_rate/qwen_vl_interface': 9.95577627161964e-06, 'learning_rate/action_model': 9.95577627161964e-05, 'epoch': 5.3}) +[2026-06-20 12:59:23,154][starVLA.training.train_starvla][INFO] - Step 2914, Loss: {'train/loss': 0.20930522680282593, 'train/grad_norm_pre_clip': 16.5673828125, 'timing/data': 0.0005206490168347955, 'timing/model': 1.494265549001284, 'learning_rate/qwen_vl_interface': 9.955360703733438e-06, 'learning_rate/action_model': 9.955360703733437e-05, 'epoch': 5.3}) +[2026-06-20 12:59:27,851][starVLA.training.train_starvla][INFO] - Step 2915, Loss: {'train/loss': 0.053423330187797546, 'train/grad_norm_pre_clip': 7.711750030517578, 'timing/data': 0.0005159089341759682, 'timing/model': 1.1856017869431525, 'learning_rate/qwen_vl_interface': 9.954943201651187e-06, 'learning_rate/action_model': 9.954943201651187e-05, 'epoch': 5.31}) +[2026-06-20 12:59:32,740][starVLA.training.train_starvla][INFO] - Step 2916, Loss: {'train/loss': 0.24238882958889008, 'train/grad_norm_pre_clip': 23.420122146606445, 'timing/data': 0.0005110390484333038, 'timing/model': 1.1909005920169875, 'learning_rate/qwen_vl_interface': 9.954523765544507e-06, 'learning_rate/action_model': 9.954523765544508e-05, 'epoch': 5.31}) +[2026-06-20 12:59:37,655][starVLA.training.train_starvla][INFO] - Step 2917, Loss: {'train/loss': 0.2724495232105255, 'train/grad_norm_pre_clip': 16.313491821289062, 'timing/data': 0.0005145389586687088, 'timing/model': 1.2023831360274926, 'learning_rate/qwen_vl_interface': 9.954102395585814e-06, 'learning_rate/action_model': 9.954102395585813e-05, 'epoch': 5.31}) +[2026-06-20 12:59:42,634][starVLA.training.train_starvla][INFO] - Step 2918, Loss: {'train/loss': 0.10586486011743546, 'train/grad_norm_pre_clip': 8.197659492492676, 'timing/data': 0.0005223589250817895, 'timing/model': 1.2352585589978844, 'learning_rate/qwen_vl_interface': 9.953679091948313e-06, 'learning_rate/action_model': 9.953679091948314e-05, 'epoch': 5.31}) +[2026-06-20 12:59:47,536][starVLA.training.train_starvla][INFO] - Step 2919, Loss: {'train/loss': 0.27409160137176514, 'train/grad_norm_pre_clip': 39.41169357299805, 'timing/data': 0.0005371689330786467, 'timing/model': 1.2409792939433828, 'learning_rate/qwen_vl_interface': 9.953253854806013e-06, 'learning_rate/action_model': 9.953253854806012e-05, 'epoch': 5.31}) +[2026-06-20 12:59:52,504][starVLA.training.train_starvla][INFO] - Step 2920, Loss: {'train/loss': 0.18032579123973846, 'train/grad_norm_pre_clip': 16.201839447021484, 'timing/data': 0.0005391589365899563, 'timing/model': 1.4692869029240683, 'learning_rate/qwen_vl_interface': 9.95282668433371e-06, 'learning_rate/action_model': 9.952826684333709e-05, 'epoch': 5.32}) +[2026-06-20 12:59:57,103][starVLA.training.train_starvla][INFO] - Step 2921, Loss: {'train/loss': 0.11749356985092163, 'train/grad_norm_pre_clip': 9.369136810302734, 'timing/data': 0.0005388690624386072, 'timing/model': 1.1884292060276493, 'learning_rate/qwen_vl_interface': 9.952397580706998e-06, 'learning_rate/action_model': 9.952397580706998e-05, 'epoch': 5.32}) +[2026-06-20 13:00:01,941][starVLA.training.train_starvla][INFO] - Step 2922, Loss: {'train/loss': 0.2527315616607666, 'train/grad_norm_pre_clip': 19.569700241088867, 'timing/data': 0.0005368690472096205, 'timing/model': 1.1879329329822212, 'learning_rate/qwen_vl_interface': 9.951966544102267e-06, 'learning_rate/action_model': 9.951966544102266e-05, 'epoch': 5.32}) +[2026-06-20 13:00:06,916][starVLA.training.train_starvla][INFO] - Step 2923, Loss: {'train/loss': 0.08348532766103745, 'train/grad_norm_pre_clip': 8.981203079223633, 'timing/data': 0.0005338189657777548, 'timing/model': 1.1853672829456627, 'learning_rate/qwen_vl_interface': 9.951533574696696e-06, 'learning_rate/action_model': 9.951533574696697e-05, 'epoch': 5.32}) +[2026-06-20 13:00:11,829][starVLA.training.train_starvla][INFO] - Step 2924, Loss: {'train/loss': 0.20349037647247314, 'train/grad_norm_pre_clip': 17.54881477355957, 'timing/data': 0.0005247090011835098, 'timing/model': 1.1869057159638032, 'learning_rate/qwen_vl_interface': 9.951098672668266e-06, 'learning_rate/action_model': 9.951098672668265e-05, 'epoch': 5.32}) +[2026-06-20 13:00:16,771][starVLA.training.train_starvla][INFO] - Step 2925, Loss: {'train/loss': 0.09492518752813339, 'train/grad_norm_pre_clip': 16.515087127685547, 'timing/data': 0.0005267079686746001, 'timing/model': 1.4664039129856974, 'learning_rate/qwen_vl_interface': 9.95066183819575e-06, 'learning_rate/action_model': 9.950661838195749e-05, 'epoch': 5.32}) +[2026-06-20 13:00:21,392][starVLA.training.train_starvla][INFO] - Step 2926, Loss: {'train/loss': 0.14401614665985107, 'train/grad_norm_pre_clip': 19.686250686645508, 'timing/data': 0.0005400090012699366, 'timing/model': 1.2409236039966345, 'learning_rate/qwen_vl_interface': 9.95022307145871e-06, 'learning_rate/action_model': 9.950223071458708e-05, 'epoch': 5.33}) +[2026-06-20 13:00:26,219][starVLA.training.train_starvla][INFO] - Step 2927, Loss: {'train/loss': 0.17288662493228912, 'train/grad_norm_pre_clip': 27.68585968017578, 'timing/data': 0.0005230490351095796, 'timing/model': 1.1742405779659748, 'learning_rate/qwen_vl_interface': 9.949782372637507e-06, 'learning_rate/action_model': 9.949782372637506e-05, 'epoch': 5.33}) +[2026-06-20 13:00:31,100][starVLA.training.train_starvla][INFO] - Step 2928, Loss: {'train/loss': 0.13934443891048431, 'train/grad_norm_pre_clip': 14.069083213806152, 'timing/data': 0.0005148090422153473, 'timing/model': 1.2422849930590019, 'learning_rate/qwen_vl_interface': 9.949339741913297e-06, 'learning_rate/action_model': 9.949339741913297e-05, 'epoch': 5.33}) +[2026-06-20 13:00:35,868][starVLA.training.train_starvla][INFO] - Step 2929, Loss: {'train/loss': 0.3143250346183777, 'train/grad_norm_pre_clip': 13.557605743408203, 'timing/data': 0.0005262689664959908, 'timing/model': 1.174186710966751, 'learning_rate/qwen_vl_interface': 9.948895179468032e-06, 'learning_rate/action_model': 9.948895179468031e-05, 'epoch': 5.33}) +[2026-06-20 13:00:40,799][starVLA.training.train_starvla][INFO] - Step 2930, Loss: {'train/loss': 0.13028202950954437, 'train/grad_norm_pre_clip': 11.181344032287598, 'timing/data': 0.0005315890302881598, 'timing/model': 1.224606508971192, 'learning_rate/qwen_vl_interface': 9.948448685484448e-06, 'learning_rate/action_model': 9.948448685484447e-05, 'epoch': 5.33}) +[2026-06-20 13:00:45,648][starVLA.training.train_starvla][INFO] - Step 2931, Loss: {'train/loss': 0.03524122014641762, 'train/grad_norm_pre_clip': 4.5728983879089355, 'timing/data': 0.000524279079400003, 'timing/model': 1.4278749279910699, 'learning_rate/qwen_vl_interface': 9.948000260146087e-06, 'learning_rate/action_model': 9.948000260146087e-05, 'epoch': 5.34}) +[2026-06-20 13:00:50,310][starVLA.training.train_starvla][INFO] - Step 2932, Loss: {'train/loss': 0.21626879274845123, 'train/grad_norm_pre_clip': 12.137025833129883, 'timing/data': 0.0005121190333738923, 'timing/model': 1.2330814050510526, 'learning_rate/qwen_vl_interface': 9.947549903637277e-06, 'learning_rate/action_model': 9.947549903637278e-05, 'epoch': 5.34}) +[2026-06-20 13:00:55,212][starVLA.training.train_starvla][INFO] - Step 2933, Loss: {'train/loss': 0.26838386058807373, 'train/grad_norm_pre_clip': 12.521307945251465, 'timing/data': 0.000548967975191772, 'timing/model': 1.2293522589607164, 'learning_rate/qwen_vl_interface': 9.947097616143145e-06, 'learning_rate/action_model': 9.947097616143144e-05, 'epoch': 5.34}) +[2026-06-20 13:01:00,224][starVLA.training.train_starvla][INFO] - Step 2934, Loss: {'train/loss': 0.24825501441955566, 'train/grad_norm_pre_clip': 13.948663711547852, 'timing/data': 0.0006566180381923914, 'timing/model': 1.272268270025961, 'learning_rate/qwen_vl_interface': 9.946643397849607e-06, 'learning_rate/action_model': 9.946643397849607e-05, 'epoch': 5.34}) +[2026-06-20 13:01:05,108][starVLA.training.train_starvla][INFO] - Step 2935, Loss: {'train/loss': 0.23738203942775726, 'train/grad_norm_pre_clip': 13.261850357055664, 'timing/data': 0.0005384590476751328, 'timing/model': 1.1725709110032767, 'learning_rate/qwen_vl_interface': 9.946187248943375e-06, 'learning_rate/action_model': 9.946187248943374e-05, 'epoch': 5.34}) +[2026-06-20 13:01:09,961][starVLA.training.train_starvla][INFO] - Step 2936, Loss: {'train/loss': 0.3501184582710266, 'train/grad_norm_pre_clip': 21.392223358154297, 'timing/data': 0.0007514189928770065, 'timing/model': 1.1847953639226034, 'learning_rate/qwen_vl_interface': 9.945729169611957e-06, 'learning_rate/action_model': 9.945729169611958e-05, 'epoch': 5.34}) +[2026-06-20 13:01:15,108][starVLA.training.train_starvla][INFO] - Step 2937, Loss: {'train/loss': 0.18751783668994904, 'train/grad_norm_pre_clip': 11.367030143737793, 'timing/data': 0.0005371390143409371, 'timing/model': 1.4977357329335064, 'learning_rate/qwen_vl_interface': 9.94526916004365e-06, 'learning_rate/action_model': 9.94526916004365e-05, 'epoch': 5.35}) +[2026-06-20 13:01:19,780][starVLA.training.train_starvla][INFO] - Step 2938, Loss: {'train/loss': 0.3454590141773224, 'train/grad_norm_pre_clip': 26.358272552490234, 'timing/data': 0.0005297290626913309, 'timing/model': 1.2328969009686261, 'learning_rate/qwen_vl_interface': 9.944807220427549e-06, 'learning_rate/action_model': 9.944807220427548e-05, 'epoch': 5.35}) +[2026-06-20 13:01:24,738][starVLA.training.train_starvla][INFO] - Step 2939, Loss: {'train/loss': 0.18121124804019928, 'train/grad_norm_pre_clip': 19.432376861572266, 'timing/data': 0.000581658910959959, 'timing/model': 1.2314834808930755, 'learning_rate/qwen_vl_interface': 9.944343350953534e-06, 'learning_rate/action_model': 9.944343350953533e-05, 'epoch': 5.35}) +[2026-06-20 13:01:29,594][starVLA.training.train_starvla][INFO] - Step 2940, Loss: {'train/loss': 0.159423828125, 'train/grad_norm_pre_clip': 8.960820198059082, 'timing/data': 0.000569639028981328, 'timing/model': 1.1817583550000563, 'learning_rate/qwen_vl_interface': 9.94387755181229e-06, 'learning_rate/action_model': 9.94387755181229e-05, 'epoch': 5.35}) +[2026-06-20 13:01:34,927][starVLA.training.train_starvla][INFO] - Step 2941, Loss: {'train/loss': 0.1794959306716919, 'train/grad_norm_pre_clip': 9.708344459533691, 'timing/data': 0.000550568918697536, 'timing/model': 1.1836559569928795, 'learning_rate/qwen_vl_interface': 9.943409823195289e-06, 'learning_rate/action_model': 9.943409823195288e-05, 'epoch': 5.35}) +[2026-06-20 13:01:39,924][starVLA.training.train_starvla][INFO] - Step 2942, Loss: {'train/loss': 0.271170049905777, 'train/grad_norm_pre_clip': 14.503120422363281, 'timing/data': 0.0006924390327185392, 'timing/model': 1.6003661380382255, 'learning_rate/qwen_vl_interface': 9.942940165294794e-06, 'learning_rate/action_model': 9.942940165294793e-05, 'epoch': 5.36}) +[2026-06-20 13:01:45,000][starVLA.training.train_starvla][INFO] - Step 2943, Loss: {'train/loss': 0.07749494910240173, 'train/grad_norm_pre_clip': 4.488255977630615, 'timing/data': 0.0005333790322765708, 'timing/model': 1.383521564071998, 'learning_rate/qwen_vl_interface': 9.942468578303865e-06, 'learning_rate/action_model': 9.942468578303865e-05, 'epoch': 5.36}) +[2026-06-20 13:01:50,433][starVLA.training.train_starvla][INFO] - Step 2944, Loss: {'train/loss': 0.14056764543056488, 'train/grad_norm_pre_clip': 6.996706962585449, 'timing/data': 0.0005242889747023582, 'timing/model': 1.267208459903486, 'learning_rate/qwen_vl_interface': 9.941995062416354e-06, 'learning_rate/action_model': 9.941995062416354e-05, 'epoch': 5.36}) +[2026-06-20 13:01:56,175][starVLA.training.train_starvla][INFO] - Step 2945, Loss: {'train/loss': 0.35372835397720337, 'train/grad_norm_pre_clip': 20.8686580657959, 'timing/data': 0.0005872590700164437, 'timing/model': 1.7212655020412058, 'learning_rate/qwen_vl_interface': 9.941519617826906e-06, 'learning_rate/action_model': 9.941519617826906e-05, 'epoch': 5.36}) +[2026-06-20 13:02:01,119][starVLA.training.train_starvla][INFO] - Step 2946, Loss: {'train/loss': 0.41843920946121216, 'train/grad_norm_pre_clip': 12.990681648254395, 'timing/data': 0.0005193880060687661, 'timing/model': 1.1806631960207596, 'learning_rate/qwen_vl_interface': 9.941042244730954e-06, 'learning_rate/action_model': 9.941042244730955e-05, 'epoch': 5.36}) +[2026-06-20 13:02:06,568][starVLA.training.train_starvla][INFO] - Step 2947, Loss: {'train/loss': 0.14822682738304138, 'train/grad_norm_pre_clip': 10.709379196166992, 'timing/data': 0.0007345289923250675, 'timing/model': 1.3039721639361233, 'learning_rate/qwen_vl_interface': 9.940562943324736e-06, 'learning_rate/action_model': 9.940562943324735e-05, 'epoch': 5.36}) +[2026-06-20 13:02:11,457][starVLA.training.train_starvla][INFO] - Step 2948, Loss: {'train/loss': 0.37875598669052124, 'train/grad_norm_pre_clip': 22.316463470458984, 'timing/data': 0.0005466780858114362, 'timing/model': 1.4580510539235547, 'learning_rate/qwen_vl_interface': 9.940081713805266e-06, 'learning_rate/action_model': 9.940081713805266e-05, 'epoch': 5.37}) +[2026-06-20 13:02:16,261][starVLA.training.train_starvla][INFO] - Step 2949, Loss: {'train/loss': 0.1426979899406433, 'train/grad_norm_pre_clip': 8.58749008178711, 'timing/data': 0.0006271390011534095, 'timing/model': 1.3576705809682608, 'learning_rate/qwen_vl_interface': 9.939598556370366e-06, 'learning_rate/action_model': 9.939598556370365e-05, 'epoch': 5.37}) +[2026-06-20 13:02:21,175][starVLA.training.train_starvla][INFO] - Step 2950, Loss: {'train/loss': 0.10077013820409775, 'train/grad_norm_pre_clip': 6.721641540527344, 'timing/data': 0.0005382489180192351, 'timing/model': 1.3638440440408885, 'learning_rate/qwen_vl_interface': 9.93911347121864e-06, 'learning_rate/action_model': 9.939113471218639e-05, 'epoch': 5.37}) +[2026-06-20 13:02:26,700][starVLA.training.train_starvla][INFO] - Step 2951, Loss: {'train/loss': 0.12719886004924774, 'train/grad_norm_pre_clip': 9.529473304748535, 'timing/data': 0.000563359004445374, 'timing/model': 1.8611723510548472, 'learning_rate/qwen_vl_interface': 9.938626458549488e-06, 'learning_rate/action_model': 9.938626458549488e-05, 'epoch': 5.37}) +[2026-06-20 13:02:32,299][starVLA.training.train_starvla][INFO] - Step 2952, Loss: {'train/loss': 0.3173404932022095, 'train/grad_norm_pre_clip': 19.423452377319336, 'timing/data': 0.000829698983579874, 'timing/model': 1.2626731599448249, 'learning_rate/qwen_vl_interface': 9.938137518563106e-06, 'learning_rate/action_model': 9.938137518563106e-05, 'epoch': 5.37}) +[2026-06-20 13:02:37,475][starVLA.training.train_starvla][INFO] - Step 2953, Loss: {'train/loss': 0.23664800822734833, 'train/grad_norm_pre_clip': 10.427611351013184, 'timing/data': 0.000643999082967639, 'timing/model': 1.179912423947826, 'learning_rate/qwen_vl_interface': 9.937646651460475e-06, 'learning_rate/action_model': 9.937646651460474e-05, 'epoch': 5.38}) +[2026-06-20 13:02:42,319][starVLA.training.train_starvla][INFO] - Step 2954, Loss: {'train/loss': 0.14226698875427246, 'train/grad_norm_pre_clip': 8.739557266235352, 'timing/data': 0.0005267789820209146, 'timing/model': 1.412429903051816, 'learning_rate/qwen_vl_interface': 9.937153857443375e-06, 'learning_rate/action_model': 9.937153857443374e-05, 'epoch': 5.38}) +[2026-06-20 13:02:47,105][starVLA.training.train_starvla][INFO] - Step 2955, Loss: {'train/loss': 0.19579915702342987, 'train/grad_norm_pre_clip': 8.771257400512695, 'timing/data': 0.007908785017207265, 'timing/model': 1.1927182140061632, 'learning_rate/qwen_vl_interface': 9.936659136714372e-06, 'learning_rate/action_model': 9.936659136714371e-05, 'epoch': 5.38}) +[2026-06-20 13:02:52,451][starVLA.training.train_starvla][INFO] - Step 2956, Loss: {'train/loss': 0.03990619257092476, 'train/grad_norm_pre_clip': 2.8685452938079834, 'timing/data': 0.028441152069717646, 'timing/model': 1.3430475438944995, 'learning_rate/qwen_vl_interface': 9.936162489476827e-06, 'learning_rate/action_model': 9.936162489476828e-05, 'epoch': 5.38}) +[2026-06-20 13:02:57,627][starVLA.training.train_starvla][INFO] - Step 2957, Loss: {'train/loss': 0.09487304836511612, 'train/grad_norm_pre_clip': 6.0327911376953125, 'timing/data': 0.02716252498794347, 'timing/model': 1.2560709270182997, 'learning_rate/qwen_vl_interface': 9.935663915934899e-06, 'learning_rate/action_model': 9.935663915934898e-05, 'epoch': 5.38}) +[2026-06-20 13:03:02,604][starVLA.training.train_starvla][INFO] - Step 2958, Loss: {'train/loss': 0.16975222527980804, 'train/grad_norm_pre_clip': 11.199204444885254, 'timing/data': 0.02700806502252817, 'timing/model': 1.2417137019801885, 'learning_rate/qwen_vl_interface': 9.935163416293525e-06, 'learning_rate/action_model': 9.935163416293525e-05, 'epoch': 5.38}) +[2026-06-20 13:03:08,482][starVLA.training.train_starvla][INFO] - Step 2959, Loss: {'train/loss': 0.16823241114616394, 'train/grad_norm_pre_clip': 9.411001205444336, 'timing/data': 0.018496612086892128, 'timing/model': 1.6071080670226365, 'learning_rate/qwen_vl_interface': 9.934660990758445e-06, 'learning_rate/action_model': 9.934660990758445e-05, 'epoch': 5.39}) +[2026-06-20 13:03:13,953][starVLA.training.train_starvla][INFO] - Step 2960, Loss: {'train/loss': 0.22299741208553314, 'train/grad_norm_pre_clip': 20.033531188964844, 'timing/data': 0.019532250007614493, 'timing/model': 1.554763937019743, 'learning_rate/qwen_vl_interface': 9.934156639536186e-06, 'learning_rate/action_model': 9.934156639536187e-05, 'epoch': 5.39}) +[2026-06-20 13:03:18,541][starVLA.training.train_starvla][INFO] - Step 2961, Loss: {'train/loss': 0.0933733880519867, 'train/grad_norm_pre_clip': 8.33241081237793, 'timing/data': 0.017505954951047897, 'timing/model': 1.2170449120458215, 'learning_rate/qwen_vl_interface': 9.93365036283407e-06, 'learning_rate/action_model': 9.93365036283407e-05, 'epoch': 5.39}) +[2026-06-20 13:03:23,694][starVLA.training.train_starvla][INFO] - Step 2962, Loss: {'train/loss': 0.1696798950433731, 'train/grad_norm_pre_clip': 10.21670150756836, 'timing/data': 0.017491224105469882, 'timing/model': 1.2700029410189018, 'learning_rate/qwen_vl_interface': 9.933142160860208e-06, 'learning_rate/action_model': 9.933142160860208e-05, 'epoch': 5.39}) +[2026-06-20 13:03:28,661][starVLA.training.train_starvla][INFO] - Step 2963, Loss: {'train/loss': 0.05588407814502716, 'train/grad_norm_pre_clip': 4.222412109375, 'timing/data': 0.018191243056207895, 'timing/model': 1.312677831039764, 'learning_rate/qwen_vl_interface': 9.932632033823499e-06, 'learning_rate/action_model': 9.932632033823499e-05, 'epoch': 5.39}) +[2026-06-20 13:03:34,102][starVLA.training.train_starvla][INFO] - Step 2964, Loss: {'train/loss': 0.11669608950614929, 'train/grad_norm_pre_clip': 30.590660095214844, 'timing/data': 0.016652185935527086, 'timing/model': 1.205453139031306, 'learning_rate/qwen_vl_interface': 9.93211998193364e-06, 'learning_rate/action_model': 9.932119981933639e-05, 'epoch': 5.4}) +[2026-06-20 13:03:39,387][starVLA.training.train_starvla][INFO] - Step 2965, Loss: {'train/loss': 0.15539191663265228, 'train/grad_norm_pre_clip': 11.007851600646973, 'timing/data': 0.023346192901954055, 'timing/model': 1.2070954800583422, 'learning_rate/qwen_vl_interface': 9.931606005401117e-06, 'learning_rate/action_model': 9.931606005401116e-05, 'epoch': 5.4}) +[2026-06-20 13:03:44,792][starVLA.training.train_starvla][INFO] - Step 2966, Loss: {'train/loss': 0.2219029814004898, 'train/grad_norm_pre_clip': 17.617645263671875, 'timing/data': 0.016311415936797857, 'timing/model': 1.5714602520456538, 'learning_rate/qwen_vl_interface': 9.931090104437202e-06, 'learning_rate/action_model': 9.931090104437202e-05, 'epoch': 5.4}) +[2026-06-20 13:03:49,964][starVLA.training.train_starvla][INFO] - Step 2967, Loss: {'train/loss': 0.038256511092185974, 'train/grad_norm_pre_clip': 4.072727203369141, 'timing/data': 0.016795714967884123, 'timing/model': 1.4447039270307869, 'learning_rate/qwen_vl_interface': 9.930572279253967e-06, 'learning_rate/action_model': 9.930572279253967e-05, 'epoch': 5.4}) +[2026-06-20 13:03:55,059][starVLA.training.train_starvla][INFO] - Step 2968, Loss: {'train/loss': 0.02754882536828518, 'train/grad_norm_pre_clip': 2.5837416648864746, 'timing/data': 0.018940540961921215, 'timing/model': 1.2358331329887733, 'learning_rate/qwen_vl_interface': 9.930052530064269e-06, 'learning_rate/action_model': 9.93005253006427e-05, 'epoch': 5.4}) +[2026-06-20 13:04:00,623][starVLA.training.train_starvla][INFO] - Step 2969, Loss: {'train/loss': 0.175492063164711, 'train/grad_norm_pre_clip': 10.352469444274902, 'timing/data': 0.021743174991570413, 'timing/model': 1.2959862659918144, 'learning_rate/qwen_vl_interface': 9.929530857081757e-06, 'learning_rate/action_model': 9.929530857081757e-05, 'epoch': 5.4}) +[2026-06-20 13:04:05,715][starVLA.training.train_starvla][INFO] - Step 2970, Loss: {'train/loss': 0.31299611926078796, 'train/grad_norm_pre_clip': 31.265701293945312, 'timing/data': 0.017686343984678388, 'timing/model': 1.2724192730383947, 'learning_rate/qwen_vl_interface': 9.92900726052087e-06, 'learning_rate/action_model': 9.929007260520869e-05, 'epoch': 5.41}) +[2026-06-20 13:04:11,114][starVLA.training.train_starvla][INFO] - Step 2971, Loss: {'train/loss': 0.15982523560523987, 'train/grad_norm_pre_clip': 19.62776756286621, 'timing/data': 0.022587702958844602, 'timing/model': 1.405945994076319, 'learning_rate/qwen_vl_interface': 9.92848174059684e-06, 'learning_rate/action_model': 9.92848174059684e-05, 'epoch': 5.41}) +[2026-06-20 13:04:16,699][starVLA.training.train_starvla][INFO] - Step 2972, Loss: {'train/loss': 0.15794731676578522, 'train/grad_norm_pre_clip': 13.774443626403809, 'timing/data': 0.018350922036916018, 'timing/model': 1.4430182829964906, 'learning_rate/qwen_vl_interface': 9.92795429752569e-06, 'learning_rate/action_model': 9.927954297525689e-05, 'epoch': 5.41}) +[2026-06-20 13:04:21,483][starVLA.training.train_starvla][INFO] - Step 2973, Loss: {'train/loss': 0.1464628428220749, 'train/grad_norm_pre_clip': 10.93990707397461, 'timing/data': 0.021473336033523083, 'timing/model': 1.224684719927609, 'learning_rate/qwen_vl_interface': 9.927424931524228e-06, 'learning_rate/action_model': 9.927424931524228e-05, 'epoch': 5.41}) +[2026-06-20 13:04:26,659][starVLA.training.train_starvla][INFO] - Step 2974, Loss: {'train/loss': 0.09922550618648529, 'train/grad_norm_pre_clip': 10.392342567443848, 'timing/data': 0.02251681301277131, 'timing/model': 1.2886742149712518, 'learning_rate/qwen_vl_interface': 9.926893642810058e-06, 'learning_rate/action_model': 9.926893642810059e-05, 'epoch': 5.41}) +[2026-06-20 13:04:31,733][starVLA.training.train_starvla][INFO] - Step 2975, Loss: {'train/loss': 0.12139231711626053, 'train/grad_norm_pre_clip': 13.312413215637207, 'timing/data': 0.01858299202285707, 'timing/model': 1.237966899992898, 'learning_rate/qwen_vl_interface': 9.926360431601575e-06, 'learning_rate/action_model': 9.926360431601574e-05, 'epoch': 5.42}) +[2026-06-20 13:04:37,034][starVLA.training.train_starvla][INFO] - Step 2976, Loss: {'train/loss': 0.1397903710603714, 'train/grad_norm_pre_clip': 18.903362274169922, 'timing/data': 0.016815705923363566, 'timing/model': 1.2374145070789382, 'learning_rate/qwen_vl_interface': 9.92582529811796e-06, 'learning_rate/action_model': 9.925825298117959e-05, 'epoch': 5.42}) +[2026-06-20 13:04:42,439][starVLA.training.train_starvla][INFO] - Step 2977, Loss: {'train/loss': 0.2688731253147125, 'train/grad_norm_pre_clip': 20.13686180114746, 'timing/data': 0.01659128605388105, 'timing/model': 1.2852107839426026, 'learning_rate/qwen_vl_interface': 9.925288242579184e-06, 'learning_rate/action_model': 9.925288242579184e-05, 'epoch': 5.42}) +[2026-06-20 13:04:47,756][starVLA.training.train_starvla][INFO] - Step 2978, Loss: {'train/loss': 0.3141503930091858, 'train/grad_norm_pre_clip': 18.5780086517334, 'timing/data': 0.016864925040863454, 'timing/model': 1.5069968410534784, 'learning_rate/qwen_vl_interface': 9.924749265206015e-06, 'learning_rate/action_model': 9.924749265206015e-05, 'epoch': 5.42}) +[2026-06-20 13:04:52,580][starVLA.training.train_starvla][INFO] - Step 2979, Loss: {'train/loss': 0.6195382475852966, 'train/grad_norm_pre_clip': 26.75477409362793, 'timing/data': 0.017142715048976243, 'timing/model': 1.257722795009613, 'learning_rate/qwen_vl_interface': 9.924208366220003e-06, 'learning_rate/action_model': 9.924208366220002e-05, 'epoch': 5.42}) +[2026-06-20 13:04:57,940][starVLA.training.train_starvla][INFO] - Step 2980, Loss: {'train/loss': 0.22866079211235046, 'train/grad_norm_pre_clip': 25.88548469543457, 'timing/data': 0.015438548056408763, 'timing/model': 1.3335062250262126, 'learning_rate/qwen_vl_interface': 9.923665545843492e-06, 'learning_rate/action_model': 9.923665545843491e-05, 'epoch': 5.42}) +[2026-06-20 13:05:03,168][starVLA.training.train_starvla][INFO] - Step 2981, Loss: {'train/loss': 0.04625383019447327, 'train/grad_norm_pre_clip': 11.399273872375488, 'timing/data': 0.019051460083574057, 'timing/model': 1.255543045932427, 'learning_rate/qwen_vl_interface': 9.923120804299615e-06, 'learning_rate/action_model': 9.923120804299613e-05, 'epoch': 5.43}) +[2026-06-20 13:05:08,691][starVLA.training.train_starvla][INFO] - Step 2982, Loss: {'train/loss': 0.1453237384557724, 'train/grad_norm_pre_clip': 17.894445419311523, 'timing/data': 0.050207495922222733, 'timing/model': 1.2087117210030556, 'learning_rate/qwen_vl_interface': 9.922574141812295e-06, 'learning_rate/action_model': 9.922574141812294e-05, 'epoch': 5.43}) +[2026-06-20 13:05:13,681][starVLA.training.train_starvla][INFO] - Step 2983, Loss: {'train/loss': 0.33594435453414917, 'train/grad_norm_pre_clip': 21.66338348388672, 'timing/data': 0.006354626966640353, 'timing/model': 1.330213455017656, 'learning_rate/qwen_vl_interface': 9.922025558606242e-06, 'learning_rate/action_model': 9.922025558606242e-05, 'epoch': 5.43}) +[2026-06-20 13:05:18,696][starVLA.training.train_starvla][INFO] - Step 2984, Loss: {'train/loss': 0.19373148679733276, 'train/grad_norm_pre_clip': 17.711990356445312, 'timing/data': 0.019501530099660158, 'timing/model': 1.4615085970144719, 'learning_rate/qwen_vl_interface': 9.92147505490696e-06, 'learning_rate/action_model': 9.92147505490696e-05, 'epoch': 5.43}) +[2026-06-20 13:05:23,536][starVLA.training.train_starvla][INFO] - Step 2985, Loss: {'train/loss': 0.21301984786987305, 'train/grad_norm_pre_clip': 25.33880615234375, 'timing/data': 0.02084072690922767, 'timing/model': 1.2364828349091113, 'learning_rate/qwen_vl_interface': 9.920922630940741e-06, 'learning_rate/action_model': 9.920922630940742e-05, 'epoch': 5.43}) +[2026-06-20 13:05:28,888][starVLA.training.train_starvla][INFO] - Step 2986, Loss: {'train/loss': 0.17950814962387085, 'train/grad_norm_pre_clip': 20.78676414489746, 'timing/data': 0.01710853399708867, 'timing/model': 1.5832333069993183, 'learning_rate/qwen_vl_interface': 9.920368286934664e-06, 'learning_rate/action_model': 9.920368286934663e-05, 'epoch': 5.44}) +[2026-06-20 13:05:34,433][starVLA.training.train_starvla][INFO] - Step 2987, Loss: {'train/loss': 0.2733600437641144, 'train/grad_norm_pre_clip': 19.456331253051758, 'timing/data': 0.028162430971860886, 'timing/model': 1.2764619459630921, 'learning_rate/qwen_vl_interface': 9.9198120231166e-06, 'learning_rate/action_model': 9.9198120231166e-05, 'epoch': 5.44}) +[2026-06-20 13:05:39,642][starVLA.training.train_starvla][INFO] - Step 2988, Loss: {'train/loss': 0.29560530185699463, 'train/grad_norm_pre_clip': 26.957901000976562, 'timing/data': 0.011059527052566409, 'timing/model': 1.2506350670009851, 'learning_rate/qwen_vl_interface': 9.919253839715208e-06, 'learning_rate/action_model': 9.919253839715208e-05, 'epoch': 5.44}) +[2026-06-20 13:05:44,522][starVLA.training.train_starvla][INFO] - Step 2989, Loss: {'train/loss': 0.14134147763252258, 'train/grad_norm_pre_clip': 10.425869941711426, 'timing/data': 0.016816835035569966, 'timing/model': 1.1775339579908177, 'learning_rate/qwen_vl_interface': 9.918693736959935e-06, 'learning_rate/action_model': 9.918693736959934e-05, 'epoch': 5.44}) +[2026-06-20 13:05:49,391][starVLA.training.train_starvla][INFO] - Step 2990, Loss: {'train/loss': 0.4415806829929352, 'train/grad_norm_pre_clip': 16.651702880859375, 'timing/data': 0.015851726988330483, 'timing/model': 1.329389106016606, 'learning_rate/qwen_vl_interface': 9.918131715081019e-06, 'learning_rate/action_model': 9.918131715081018e-05, 'epoch': 5.44}) +[2026-06-20 13:05:53,947][starVLA.training.train_starvla][INFO] - Step 2991, Loss: {'train/loss': 0.34563344717025757, 'train/grad_norm_pre_clip': 14.269312858581543, 'timing/data': 0.018604621989652514, 'timing/model': 1.1786169219994918, 'learning_rate/qwen_vl_interface': 9.917567774309486e-06, 'learning_rate/action_model': 9.917567774309485e-05, 'epoch': 5.44}) +[2026-06-20 13:05:58,727][starVLA.training.train_starvla][INFO] - Step 2992, Loss: {'train/loss': 0.07572969794273376, 'train/grad_norm_pre_clip': 5.261978626251221, 'timing/data': 0.017696022987365723, 'timing/model': 1.1974556999048218, 'learning_rate/qwen_vl_interface': 9.91700191487715e-06, 'learning_rate/action_model': 9.91700191487715e-05, 'epoch': 5.45}) +[2026-06-20 13:06:03,611][starVLA.training.train_starvla][INFO] - Step 2993, Loss: {'train/loss': 0.2430771142244339, 'train/grad_norm_pre_clip': 11.61305046081543, 'timing/data': 0.018135791993699968, 'timing/model': 1.1932298979954794, 'learning_rate/qwen_vl_interface': 9.916434137016614e-06, 'learning_rate/action_model': 9.916434137016613e-05, 'epoch': 5.45}) +[2026-06-20 13:06:08,777][starVLA.training.train_starvla][INFO] - Step 2994, Loss: {'train/loss': 0.11487090587615967, 'train/grad_norm_pre_clip': 7.6246538162231445, 'timing/data': 0.01644886506255716, 'timing/model': 1.3267049540299922, 'learning_rate/qwen_vl_interface': 9.915864440961272e-06, 'learning_rate/action_model': 9.915864440961272e-05, 'epoch': 5.45}) +[2026-06-20 13:06:13,807][starVLA.training.train_starvla][INFO] - Step 2995, Loss: {'train/loss': 0.11773817986249924, 'train/grad_norm_pre_clip': 7.170829772949219, 'timing/data': 0.042199680930934846, 'timing/model': 1.491635065060109, 'learning_rate/qwen_vl_interface': 9.915292826945303e-06, 'learning_rate/action_model': 9.915292826945303e-05, 'epoch': 5.45}) +[2026-06-20 13:06:18,879][starVLA.training.train_starvla][INFO] - Step 2996, Loss: {'train/loss': 0.18842892348766327, 'train/grad_norm_pre_clip': 11.889495849609375, 'timing/data': 0.011636556009761989, 'timing/model': 1.3377401758916676, 'learning_rate/qwen_vl_interface': 9.914719295203676e-06, 'learning_rate/action_model': 9.914719295203676e-05, 'epoch': 5.45}) +[2026-06-20 13:06:24,446][starVLA.training.train_starvla][INFO] - Step 2997, Loss: {'train/loss': 0.11263086646795273, 'train/grad_norm_pre_clip': 8.735101699829102, 'timing/data': 0.01393047091551125, 'timing/model': 1.2302729811053723, 'learning_rate/qwen_vl_interface': 9.91414384597215e-06, 'learning_rate/action_model': 9.91414384597215e-05, 'epoch': 5.46}) +[2026-06-20 13:06:29,441][starVLA.training.train_starvla][INFO] - Step 2998, Loss: {'train/loss': 0.11962109804153442, 'train/grad_norm_pre_clip': 5.835721492767334, 'timing/data': 0.011949484003707767, 'timing/model': 1.1744753170059994, 'learning_rate/qwen_vl_interface': 9.913566479487267e-06, 'learning_rate/action_model': 9.913566479487267e-05, 'epoch': 5.46}) +[2026-06-20 13:06:34,393][starVLA.training.train_starvla][INFO] - Step 2999, Loss: {'train/loss': 0.17725419998168945, 'train/grad_norm_pre_clip': 11.718061447143555, 'timing/data': 0.012649463023990393, 'timing/model': 1.2477307099616155, 'learning_rate/qwen_vl_interface': 9.912987195986366e-06, 'learning_rate/action_model': 9.912987195986365e-05, 'epoch': 5.46}) +[2026-06-20 13:06:39,602][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 3000 +[2026-06-20 13:12:50,651][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 3000 +[2026-06-20 13:12:50,652][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 3000 +[2026-06-20 13:19:19,543][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 3000 +[2026-06-20 13:19:19,545][starVLA.training.train_starvla][INFO] - Step 3000, Loss: {'train/loss': 0.19053764641284943, 'train/grad_norm_pre_clip': 8.53667163848877, 'eval/loss': 0.3816286325454712, 'eval/latency_0/loss': 0.37554094195365906, 'eval/latency_1/loss': 0.46909892559051514, 'eval/latency_2/loss': 0.49875620007514954, 'eval/latency_3/loss': 0.09705671668052673, 'eval/latency_4/loss': 0.43397924304008484, 'eval/flappy/loss': 0.3816286325454712, 'eval/flappy/latency_0/loss': 0.37554094195365906, 'eval/flappy/latency_1/loss': 0.46909892559051514, 'eval/flappy/latency_2/loss': 0.49875620007514954, 'eval/flappy/latency_3/loss': 0.09705671668052673, 'eval/flappy/latency_4/loss': 0.43397924304008484, 'eval/flap_precision': 0.14814814814814814, 'eval/flap_recall': 0.13090909090909092, 'eval/flap_f1': 0.138996138996139, 'eval/flap_rate_model': 0.0675, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.138996138996139, 'eval/latency_0/flap_precision': 0.14814814814814814, 'eval/latency_0/flap_recall': 0.13090909090909092, 'eval/latency_0/flap_f1': 0.138996138996139, 'eval/latency_0/flap_rate_model': 0.0675, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.138996138996139, 'timing/data': 0.010501778102479875, 'timing/model': 1.5835641149897128, 'learning_rate/qwen_vl_interface': 9.912405995707562e-06, 'learning_rate/action_model': 9.912405995707561e-05, 'epoch': 5.46}) +[2026-06-20 13:19:19,545][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_3000_state +[2026-06-20 13:19:19,549][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-20 13:19:29,785][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_3000_state/model.safetensors +[2026-06-20 13:19:46,157][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_3000_state/optimizer.bin +[2026-06-20 13:19:46,157][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_3000_state/sampler.bin +[2026-06-20 13:19:46,157][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_3000_state/sampler_1.bin +[2026-06-20 13:19:46,158][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_3000_state/random_states_0.pkl +[2026-06-20 13:19:47,231][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3000 +[2026-06-20 13:19:47,231][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-20 13:19:47,245][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-20 13:19:52,227][starVLA.training.train_starvla][INFO] - Step 3001, Loss: {'train/loss': 0.13613170385360718, 'train/grad_norm_pre_clip': 6.197278022766113, 'timing/data': 0.016904444084502757, 'timing/model': 1.5415578370448202, 'learning_rate/qwen_vl_interface': 9.911822878889767e-06, 'learning_rate/action_model': 9.911822878889767e-05, 'epoch': 5.46}) +[2026-06-20 13:19:56,876][starVLA.training.train_starvla][INFO] - Step 3002, Loss: {'train/loss': 0.36827579140663147, 'train/grad_norm_pre_clip': 11.530111312866211, 'timing/data': 0.011479011969640851, 'timing/model': 1.3307397189782932, 'learning_rate/qwen_vl_interface': 9.911237845772681e-06, 'learning_rate/action_model': 9.911237845772681e-05, 'epoch': 5.46}) +[2026-06-20 13:20:01,946][starVLA.training.train_starvla][INFO] - Step 3003, Loss: {'train/loss': 0.06212420389056206, 'train/grad_norm_pre_clip': 5.503949165344238, 'timing/data': 0.01783826993778348, 'timing/model': 1.2227924970211461, 'learning_rate/qwen_vl_interface': 9.910650896596787e-06, 'learning_rate/action_model': 9.910650896596786e-05, 'epoch': 5.47}) +[2026-06-20 13:20:06,777][starVLA.training.train_starvla][INFO] - Step 3004, Loss: {'train/loss': 0.04251173511147499, 'train/grad_norm_pre_clip': 3.3611340522766113, 'timing/data': 0.011082604993134737, 'timing/model': 1.1680818729801103, 'learning_rate/qwen_vl_interface': 9.910062031603356e-06, 'learning_rate/action_model': 9.910062031603356e-05, 'epoch': 5.47}) +[2026-06-20 13:20:11,821][starVLA.training.train_starvla][INFO] - Step 3005, Loss: {'train/loss': 0.18385626375675201, 'train/grad_norm_pre_clip': 12.492008209228516, 'timing/data': 0.01550005201715976, 'timing/model': 1.2287730240495875, 'learning_rate/qwen_vl_interface': 9.909471251034451e-06, 'learning_rate/action_model': 9.90947125103445e-05, 'epoch': 5.47}) +[2026-06-20 13:20:16,795][starVLA.training.train_starvla][INFO] - Step 3006, Loss: {'train/loss': 0.10099714994430542, 'train/grad_norm_pre_clip': 4.4031662940979, 'timing/data': 0.011374573921784759, 'timing/model': 1.1634406800149009, 'learning_rate/qwen_vl_interface': 9.908878555132918e-06, 'learning_rate/action_model': 9.908878555132918e-05, 'epoch': 5.47}) +[2026-06-20 13:20:21,722][starVLA.training.train_starvla][INFO] - Step 3007, Loss: {'train/loss': 0.10365147143602371, 'train/grad_norm_pre_clip': 10.632144927978516, 'timing/data': 0.015480443951673806, 'timing/model': 1.4669171019922942, 'learning_rate/qwen_vl_interface': 9.908283944142391e-06, 'learning_rate/action_model': 9.908283944142391e-05, 'epoch': 5.47}) +[2026-06-20 13:20:26,498][starVLA.training.train_starvla][INFO] - Step 3008, Loss: {'train/loss': 0.0700826644897461, 'train/grad_norm_pre_clip': 6.285645961761475, 'timing/data': 0.014623738010413945, 'timing/model': 1.214727601967752, 'learning_rate/qwen_vl_interface': 9.907687418307297e-06, 'learning_rate/action_model': 9.907687418307296e-05, 'epoch': 5.48}) +[2026-06-20 13:20:31,737][starVLA.training.train_starvla][INFO] - Step 3009, Loss: {'train/loss': 0.09967028349637985, 'train/grad_norm_pre_clip': 7.951142311096191, 'timing/data': 0.011188204982317984, 'timing/model': 1.3489569169469178, 'learning_rate/qwen_vl_interface': 9.907088977872837e-06, 'learning_rate/action_model': 9.907088977872837e-05, 'epoch': 5.48}) +[2026-06-20 13:20:36,649][starVLA.training.train_starvla][INFO] - Step 3010, Loss: {'train/loss': 0.23219913244247437, 'train/grad_norm_pre_clip': 12.003857612609863, 'timing/data': 0.012992497067898512, 'timing/model': 1.1979236479382962, 'learning_rate/qwen_vl_interface': 9.906488623085015e-06, 'learning_rate/action_model': 9.906488623085015e-05, 'epoch': 5.48}) +[2026-06-20 13:20:41,752][starVLA.training.train_starvla][INFO] - Step 3011, Loss: {'train/loss': 0.14499731361865997, 'train/grad_norm_pre_clip': 13.831395149230957, 'timing/data': 0.011295935022644699, 'timing/model': 1.2108064390486106, 'learning_rate/qwen_vl_interface': 9.905886354190611e-06, 'learning_rate/action_model': 9.90588635419061e-05, 'epoch': 5.48}) +[2026-06-20 13:20:46,603][starVLA.training.train_starvla][INFO] - Step 3012, Loss: {'train/loss': 0.25058257579803467, 'train/grad_norm_pre_clip': 22.593738555908203, 'timing/data': 0.010959357023239136, 'timing/model': 1.201729392982088, 'learning_rate/qwen_vl_interface': 9.905282171437192e-06, 'learning_rate/action_model': 9.905282171437192e-05, 'epoch': 5.48}) +[2026-06-20 13:20:51,789][starVLA.training.train_starvla][INFO] - Step 3013, Loss: {'train/loss': 0.16794726252555847, 'train/grad_norm_pre_clip': 11.923050880432129, 'timing/data': 0.01427349098958075, 'timing/model': 1.637800617958419, 'learning_rate/qwen_vl_interface': 9.904676075073121e-06, 'learning_rate/action_model': 9.904676075073121e-05, 'epoch': 5.48}) +[2026-06-20 13:20:56,741][starVLA.training.train_starvla][INFO] - Step 3014, Loss: {'train/loss': 0.34430065751075745, 'train/grad_norm_pre_clip': 19.588056564331055, 'timing/data': 0.18235493102110922, 'timing/model': 1.383905170019716, 'learning_rate/qwen_vl_interface': 9.904068065347536e-06, 'learning_rate/action_model': 9.904068065347536e-05, 'epoch': 5.49}) +[2026-06-20 13:21:01,869][starVLA.training.train_starvla][INFO] - Step 3015, Loss: {'train/loss': 0.10033266991376877, 'train/grad_norm_pre_clip': 8.643365859985352, 'timing/data': 0.011242885957472026, 'timing/model': 1.2081957960035652, 'learning_rate/qwen_vl_interface': 9.90345814251037e-06, 'learning_rate/action_model': 9.90345814251037e-05, 'epoch': 5.49}) +[2026-06-20 13:21:06,813][starVLA.training.train_starvla][INFO] - Step 3016, Loss: {'train/loss': 0.1421986073255539, 'train/grad_norm_pre_clip': 6.053017616271973, 'timing/data': 0.010163272032514215, 'timing/model': 1.176758558023721, 'learning_rate/qwen_vl_interface': 9.902846306812339e-06, 'learning_rate/action_model': 9.902846306812338e-05, 'epoch': 5.49}) +[2026-06-20 13:21:11,619][starVLA.training.train_starvla][INFO] - Step 3017, Loss: {'train/loss': 0.09797987341880798, 'train/grad_norm_pre_clip': 13.016395568847656, 'timing/data': 0.010750568937510252, 'timing/model': 1.179282204946503, 'learning_rate/qwen_vl_interface': 9.902232558504942e-06, 'learning_rate/action_model': 9.902232558504941e-05, 'epoch': 5.49}) +[2026-06-20 13:21:16,722][starVLA.training.train_starvla][INFO] - Step 3018, Loss: {'train/loss': 0.03979640454053879, 'train/grad_norm_pre_clip': 3.543510913848877, 'timing/data': 0.013357377029024065, 'timing/model': 1.2199871420161799, 'learning_rate/qwen_vl_interface': 9.901616897840471e-06, 'learning_rate/action_model': 9.90161689784047e-05, 'epoch': 5.49}) +[2026-06-20 13:21:21,588][starVLA.training.train_starvla][INFO] - Step 3019, Loss: {'train/loss': 0.291007399559021, 'train/grad_norm_pre_clip': 20.723947525024414, 'timing/data': 0.013547535985708237, 'timing/model': 1.4178150469670072, 'learning_rate/qwen_vl_interface': 9.900999325072e-06, 'learning_rate/action_model': 9.900999325072001e-05, 'epoch': 5.5}) +[2026-06-20 13:21:26,391][starVLA.training.train_starvla][INFO] - Step 3020, Loss: {'train/loss': 0.12639811635017395, 'train/grad_norm_pre_clip': 17.257680892944336, 'timing/data': 0.011249867035076022, 'timing/model': 1.4508607650641352, 'learning_rate/qwen_vl_interface': 9.900379840453392e-06, 'learning_rate/action_model': 9.900379840453392e-05, 'epoch': 5.5}) +[2026-06-20 13:21:31,300][starVLA.training.train_starvla][INFO] - Step 3021, Loss: {'train/loss': 0.08614238351583481, 'train/grad_norm_pre_clip': 13.774682998657227, 'timing/data': 0.011898573953658342, 'timing/model': 1.2060645269230008, 'learning_rate/qwen_vl_interface': 9.89975844423929e-06, 'learning_rate/action_model': 9.899758444239289e-05, 'epoch': 5.5}) +[2026-06-20 13:21:36,095][starVLA.training.train_starvla][INFO] - Step 3022, Loss: {'train/loss': 0.018174834549427032, 'train/grad_norm_pre_clip': 1.5829850435256958, 'timing/data': 0.010996668017469347, 'timing/model': 1.1825729259289801, 'learning_rate/qwen_vl_interface': 9.899135136685131e-06, 'learning_rate/action_model': 9.89913513668513e-05, 'epoch': 5.5}) +[2026-06-20 13:21:41,024][starVLA.training.train_starvla][INFO] - Step 3023, Loss: {'train/loss': 0.3388861119747162, 'train/grad_norm_pre_clip': 13.115781784057617, 'timing/data': 0.012565801036544144, 'timing/model': 1.2373778000473976, 'learning_rate/qwen_vl_interface': 9.898509918047127e-06, 'learning_rate/action_model': 9.898509918047126e-05, 'epoch': 5.5}) +[2026-06-20 13:21:46,079][starVLA.training.train_starvla][INFO] - Step 3024, Loss: {'train/loss': 0.033621106296777725, 'train/grad_norm_pre_clip': 3.5154519081115723, 'timing/data': 0.011519956053234637, 'timing/model': 1.2558911229716614, 'learning_rate/qwen_vl_interface': 9.897882788582286e-06, 'learning_rate/action_model': 9.897882788582287e-05, 'epoch': 5.5}) +[2026-06-20 13:21:51,352][starVLA.training.train_starvla][INFO] - Step 3025, Loss: {'train/loss': 0.12088149785995483, 'train/grad_norm_pre_clip': 8.1708984375, 'timing/data': 0.013144769938662648, 'timing/model': 1.847440231940709, 'learning_rate/qwen_vl_interface': 9.897253748548398e-06, 'learning_rate/action_model': 9.897253748548399e-05, 'epoch': 5.51}) +[2026-06-20 13:21:56,244][starVLA.training.train_starvla][INFO] - Step 3026, Loss: {'train/loss': 0.22997979819774628, 'train/grad_norm_pre_clip': 15.41966438293457, 'timing/data': 0.011780045926570892, 'timing/model': 1.438559078029357, 'learning_rate/qwen_vl_interface': 9.896622798204035e-06, 'learning_rate/action_model': 9.896622798204035e-05, 'epoch': 5.51}) +[2026-06-20 13:22:01,108][starVLA.training.train_starvla][INFO] - Step 3027, Loss: {'train/loss': 0.24998874962329865, 'train/grad_norm_pre_clip': 11.514182090759277, 'timing/data': 0.013180478941649199, 'timing/model': 1.3219465779839084, 'learning_rate/qwen_vl_interface': 9.89598993780856e-06, 'learning_rate/action_model': 9.89598993780856e-05, 'epoch': 5.51}) +[2026-06-20 13:22:06,022][starVLA.training.train_starvla][INFO] - Step 3028, Loss: {'train/loss': 0.21595199406147003, 'train/grad_norm_pre_clip': 15.099786758422852, 'timing/data': 0.011601537931710482, 'timing/model': 1.1941932139452547, 'learning_rate/qwen_vl_interface': 9.895355167622113e-06, 'learning_rate/action_model': 9.895355167622113e-05, 'epoch': 5.51}) +[2026-06-20 13:22:10,898][starVLA.training.train_starvla][INFO] - Step 3029, Loss: {'train/loss': 0.07393696159124374, 'train/grad_norm_pre_clip': 8.011595726013184, 'timing/data': 0.011198028922080994, 'timing/model': 1.1771616189507768, 'learning_rate/qwen_vl_interface': 9.894718487905629e-06, 'learning_rate/action_model': 9.894718487905629e-05, 'epoch': 5.51}) +[2026-06-20 13:22:15,855][starVLA.training.train_starvla][INFO] - Step 3030, Loss: {'train/loss': 0.06169252097606659, 'train/grad_norm_pre_clip': 6.930784702301025, 'timing/data': 0.012244053999893367, 'timing/model': 1.1926236769650131, 'learning_rate/qwen_vl_interface': 9.89407989892082e-06, 'learning_rate/action_model': 9.894079898920821e-05, 'epoch': 5.52}) +[2026-06-20 13:22:21,008][starVLA.training.train_starvla][INFO] - Step 3031, Loss: {'train/loss': 0.13663063943386078, 'train/grad_norm_pre_clip': 13.942240715026855, 'timing/data': 0.013912096968851984, 'timing/model': 1.5209267339669168, 'learning_rate/qwen_vl_interface': 9.893439400930187e-06, 'learning_rate/action_model': 9.893439400930187e-05, 'epoch': 5.52}) +[2026-06-20 13:22:25,906][starVLA.training.train_starvla][INFO] - Step 3032, Loss: {'train/loss': 0.2742788791656494, 'train/grad_norm_pre_clip': 18.990331649780273, 'timing/data': 0.01331051904708147, 'timing/model': 1.33209049096331, 'learning_rate/qwen_vl_interface': 9.892796994197014e-06, 'learning_rate/action_model': 9.892796994197014e-05, 'epoch': 5.52}) +[2026-06-20 13:22:31,084][starVLA.training.train_starvla][INFO] - Step 3033, Loss: {'train/loss': 0.18651451170444489, 'train/grad_norm_pre_clip': 19.460622787475586, 'timing/data': 0.012604112969711423, 'timing/model': 1.30664975696709, 'learning_rate/qwen_vl_interface': 9.892152678985372e-06, 'learning_rate/action_model': 9.892152678985372e-05, 'epoch': 5.52}) +[2026-06-20 13:22:36,024][starVLA.training.train_starvla][INFO] - Step 3034, Loss: {'train/loss': 0.16781164705753326, 'train/grad_norm_pre_clip': 10.045421600341797, 'timing/data': 0.012118285987526178, 'timing/model': 1.188722364953719, 'learning_rate/qwen_vl_interface': 9.891506455560111e-06, 'learning_rate/action_model': 9.891506455560111e-05, 'epoch': 5.52}) +[2026-06-20 13:22:41,145][starVLA.training.train_starvla][INFO] - Step 3035, Loss: {'train/loss': 0.166065976023674, 'train/grad_norm_pre_clip': 17.057138442993164, 'timing/data': 0.010821310919709504, 'timing/model': 1.2914340629940853, 'learning_rate/qwen_vl_interface': 9.890858324186872e-06, 'learning_rate/action_model': 9.890858324186872e-05, 'epoch': 5.52}) +[2026-06-20 13:22:46,064][starVLA.training.train_starvla][INFO] - Step 3036, Loss: {'train/loss': 0.1590777337551117, 'train/grad_norm_pre_clip': 8.65599250793457, 'timing/data': 0.017116492963396013, 'timing/model': 1.228425742010586, 'learning_rate/qwen_vl_interface': 9.890208285132077e-06, 'learning_rate/action_model': 9.890208285132077e-05, 'epoch': 5.53}) +[2026-06-20 13:22:51,070][starVLA.training.train_starvla][INFO] - Step 3037, Loss: {'train/loss': 0.15937191247940063, 'train/grad_norm_pre_clip': 10.402688980102539, 'timing/data': 0.011583478073589504, 'timing/model': 1.414479712024331, 'learning_rate/qwen_vl_interface': 9.889556338662933e-06, 'learning_rate/action_model': 9.889556338662932e-05, 'epoch': 5.53}) +[2026-06-20 13:22:55,954][starVLA.training.train_starvla][INFO] - Step 3038, Loss: {'train/loss': 0.048163365572690964, 'train/grad_norm_pre_clip': 5.308138847351074, 'timing/data': 0.011971706990152597, 'timing/model': 1.2186603191075847, 'learning_rate/qwen_vl_interface': 9.888902485047427e-06, 'learning_rate/action_model': 9.888902485047427e-05, 'epoch': 5.53}) +[2026-06-20 13:23:00,953][starVLA.training.train_starvla][INFO] - Step 3039, Loss: {'train/loss': 0.20249687135219574, 'train/grad_norm_pre_clip': 12.264002799987793, 'timing/data': 0.014585944940336049, 'timing/model': 1.2248175790300593, 'learning_rate/qwen_vl_interface': 9.888246724554337e-06, 'learning_rate/action_model': 9.888246724554337e-05, 'epoch': 5.53}) +[2026-06-20 13:23:06,007][starVLA.training.train_starvla][INFO] - Step 3040, Loss: {'train/loss': 0.14249378442764282, 'train/grad_norm_pre_clip': 12.49387264251709, 'timing/data': 0.016592907020822167, 'timing/model': 1.2460361429257318, 'learning_rate/qwen_vl_interface': 9.887589057453222e-06, 'learning_rate/action_model': 9.88758905745322e-05, 'epoch': 5.53}) +[2026-06-20 13:23:10,951][starVLA.training.train_starvla][INFO] - Step 3041, Loss: {'train/loss': 0.21527519822120667, 'train/grad_norm_pre_clip': 15.340920448303223, 'timing/data': 0.011272580013610423, 'timing/model': 1.1988284479593858, 'learning_rate/qwen_vl_interface': 9.88692948401442e-06, 'learning_rate/action_model': 9.88692948401442e-05, 'epoch': 5.54}) +[2026-06-20 13:23:16,025][starVLA.training.train_starvla][INFO] - Step 3042, Loss: {'train/loss': 0.1967834085226059, 'train/grad_norm_pre_clip': 11.861944198608398, 'timing/data': 0.013843599008396268, 'timing/model': 1.2299146259902045, 'learning_rate/qwen_vl_interface': 9.88626800450906e-06, 'learning_rate/action_model': 9.88626800450906e-05, 'epoch': 5.54}) +[2026-06-20 13:23:21,157][starVLA.training.train_starvla][INFO] - Step 3043, Loss: {'train/loss': 0.2166910171508789, 'train/grad_norm_pre_clip': 13.827960968017578, 'timing/data': 0.01417465799022466, 'timing/model': 1.5196871550288051, 'learning_rate/qwen_vl_interface': 9.885604619209052e-06, 'learning_rate/action_model': 9.885604619209051e-05, 'epoch': 5.54}) +[2026-06-20 13:23:25,819][starVLA.training.train_starvla][INFO] - Step 3044, Loss: {'train/loss': 0.3311970829963684, 'train/grad_norm_pre_clip': 41.58323669433594, 'timing/data': 0.011305510066449642, 'timing/model': 1.2624975959770381, 'learning_rate/qwen_vl_interface': 9.884939328387085e-06, 'learning_rate/action_model': 9.884939328387084e-05, 'epoch': 5.54}) +[2026-06-20 13:23:30,825][starVLA.training.train_starvla][INFO] - Step 3045, Loss: {'train/loss': 0.09357759356498718, 'train/grad_norm_pre_clip': 6.420935153961182, 'timing/data': 0.011514729936607182, 'timing/model': 1.2004847029456869, 'learning_rate/qwen_vl_interface': 9.884272132316638e-06, 'learning_rate/action_model': 9.884272132316637e-05, 'epoch': 5.54}) +[2026-06-20 13:23:35,733][starVLA.training.train_starvla][INFO] - Step 3046, Loss: {'train/loss': 0.10603103786706924, 'train/grad_norm_pre_clip': 9.799097061157227, 'timing/data': 0.010273225023411214, 'timing/model': 1.1972349419957027, 'learning_rate/qwen_vl_interface': 9.883603031271968e-06, 'learning_rate/action_model': 9.883603031271969e-05, 'epoch': 5.54}) +[2026-06-20 13:23:40,932][starVLA.training.train_starvla][INFO] - Step 3047, Loss: {'train/loss': 0.15179434418678284, 'train/grad_norm_pre_clip': 16.58452796936035, 'timing/data': 0.011849318980239332, 'timing/model': 1.1907363459467888, 'learning_rate/qwen_vl_interface': 9.882932025528122e-06, 'learning_rate/action_model': 9.882932025528121e-05, 'epoch': 5.55}) +[2026-06-20 13:23:45,978][starVLA.training.train_starvla][INFO] - Step 3048, Loss: {'train/loss': 0.04473142698407173, 'train/grad_norm_pre_clip': 6.41350793838501, 'timing/data': 0.01132102101109922, 'timing/model': 1.435623886063695, 'learning_rate/qwen_vl_interface': 9.882259115360918e-06, 'learning_rate/action_model': 9.882259115360917e-05, 'epoch': 5.55}) +[2026-06-20 13:23:50,610][starVLA.training.train_starvla][INFO] - Step 3049, Loss: {'train/loss': 0.22048300504684448, 'train/grad_norm_pre_clip': 12.578326225280762, 'timing/data': 0.01273876498453319, 'timing/model': 1.2081001510377973, 'learning_rate/qwen_vl_interface': 9.881584301046968e-06, 'learning_rate/action_model': 9.881584301046968e-05, 'epoch': 5.55}) +[2026-06-20 13:23:55,668][starVLA.training.train_starvla][INFO] - Step 3050, Loss: {'train/loss': 0.11123208701610565, 'train/grad_norm_pre_clip': 8.929841995239258, 'timing/data': 0.011149471974931657, 'timing/model': 1.1960488869808614, 'learning_rate/qwen_vl_interface': 9.880907582863662e-06, 'learning_rate/action_model': 9.880907582863662e-05, 'epoch': 5.55}) +[2026-06-20 13:24:00,947][starVLA.training.train_starvla][INFO] - Step 3051, Loss: {'train/loss': 0.1284801959991455, 'train/grad_norm_pre_clip': 6.7715678215026855, 'timing/data': 0.013726260978728533, 'timing/model': 1.4509397639194503, 'learning_rate/qwen_vl_interface': 9.880228961089175e-06, 'learning_rate/action_model': 9.880228961089175e-05, 'epoch': 5.55}) +[2026-06-20 13:24:06,020][starVLA.training.train_starvla][INFO] - Step 3052, Loss: {'train/loss': 0.20814023911952972, 'train/grad_norm_pre_clip': 11.132547378540039, 'timing/data': 0.011225512018427253, 'timing/model': 1.1982827199390158, 'learning_rate/qwen_vl_interface': 9.87954843600246e-06, 'learning_rate/action_model': 9.879548436002459e-05, 'epoch': 5.56}) +[2026-06-20 13:24:11,254][starVLA.training.train_starvla][INFO] - Step 3053, Loss: {'train/loss': 0.1730794906616211, 'train/grad_norm_pre_clip': 9.339232444763184, 'timing/data': 0.01134941098280251, 'timing/model': 1.2083762699039653, 'learning_rate/qwen_vl_interface': 9.878866007883256e-06, 'learning_rate/action_model': 9.878866007883257e-05, 'epoch': 5.56}) +[2026-06-20 13:24:16,642][starVLA.training.train_starvla][INFO] - Step 3054, Loss: {'train/loss': 0.07755157351493835, 'train/grad_norm_pre_clip': 5.593517303466797, 'timing/data': 0.012478537973947823, 'timing/model': 1.7852356450166553, 'learning_rate/qwen_vl_interface': 9.878181677012084e-06, 'learning_rate/action_model': 9.878181677012085e-05, 'epoch': 5.56}) +[2026-06-20 13:24:21,215][starVLA.training.train_starvla][INFO] - Step 3055, Loss: {'train/loss': 0.13180376589298248, 'train/grad_norm_pre_clip': 8.88236141204834, 'timing/data': 0.011681880918331444, 'timing/model': 1.1822483598953113, 'learning_rate/qwen_vl_interface': 9.877495443670247e-06, 'learning_rate/action_model': 9.877495443670246e-05, 'epoch': 5.56}) +[2026-06-20 13:24:26,688][starVLA.training.train_starvla][INFO] - Step 3056, Loss: {'train/loss': 0.17705906927585602, 'train/grad_norm_pre_clip': 12.475643157958984, 'timing/data': 0.010513225104659796, 'timing/model': 1.3042018620762974, 'learning_rate/qwen_vl_interface': 9.876807308139828e-06, 'learning_rate/action_model': 9.876807308139829e-05, 'epoch': 5.56}) +[2026-06-20 13:24:31,881][starVLA.training.train_starvla][INFO] - Step 3057, Loss: {'train/loss': 0.17323628067970276, 'train/grad_norm_pre_clip': 8.842272758483887, 'timing/data': 0.013371934997849166, 'timing/model': 1.303361979080364, 'learning_rate/qwen_vl_interface': 9.876117270703695e-06, 'learning_rate/action_model': 9.876117270703694e-05, 'epoch': 5.56}) +[2026-06-20 13:24:36,911][starVLA.training.train_starvla][INFO] - Step 3058, Loss: {'train/loss': 0.03351685032248497, 'train/grad_norm_pre_clip': 1.9390404224395752, 'timing/data': 0.013185865012928843, 'timing/model': 1.2275446529965848, 'learning_rate/qwen_vl_interface': 9.875425331645495e-06, 'learning_rate/action_model': 9.875425331645495e-05, 'epoch': 5.57}) +[2026-06-20 13:24:41,956][starVLA.training.train_starvla][INFO] - Step 3059, Loss: {'train/loss': 0.18280673027038574, 'train/grad_norm_pre_clip': 10.537301063537598, 'timing/data': 0.010971143958158791, 'timing/model': 1.1830915419850498, 'learning_rate/qwen_vl_interface': 9.874731491249658e-06, 'learning_rate/action_model': 9.874731491249657e-05, 'epoch': 5.57}) +[2026-06-20 13:24:46,940][starVLA.training.train_starvla][INFO] - Step 3060, Loss: {'train/loss': 0.14874674379825592, 'train/grad_norm_pre_clip': 9.428691864013672, 'timing/data': 0.025317035033367574, 'timing/model': 1.492675757035613, 'learning_rate/qwen_vl_interface': 9.874035749801398e-06, 'learning_rate/action_model': 9.874035749801399e-05, 'epoch': 5.57}) +[2026-06-20 13:24:51,921][starVLA.training.train_starvla][INFO] - Step 3061, Loss: {'train/loss': 0.15609240531921387, 'train/grad_norm_pre_clip': 10.10861873626709, 'timing/data': 0.012251259991899133, 'timing/model': 1.2147991380188614, 'learning_rate/qwen_vl_interface': 9.873338107586704e-06, 'learning_rate/action_model': 9.873338107586705e-05, 'epoch': 5.57}) +[2026-06-20 13:24:57,210][starVLA.training.train_starvla][INFO] - Step 3062, Loss: {'train/loss': 0.33435994386672974, 'train/grad_norm_pre_clip': 8.032007217407227, 'timing/data': 0.011468362994492054, 'timing/model': 1.1845126779517159, 'learning_rate/qwen_vl_interface': 9.872638564892358e-06, 'learning_rate/action_model': 9.872638564892357e-05, 'epoch': 5.57}) +[2026-06-20 13:25:02,349][starVLA.training.train_starvla][INFO] - Step 3063, Loss: {'train/loss': 0.14159895479679108, 'train/grad_norm_pre_clip': 7.475249290466309, 'timing/data': 0.014961507986299694, 'timing/model': 1.242757610976696, 'learning_rate/qwen_vl_interface': 9.871937122005905e-06, 'learning_rate/action_model': 9.871937122005905e-05, 'epoch': 5.58}) +[2026-06-20 13:25:07,660][starVLA.training.train_starvla][INFO] - Step 3064, Loss: {'train/loss': 0.027999280020594597, 'train/grad_norm_pre_clip': 2.9339306354522705, 'timing/data': 0.011552693089470267, 'timing/model': 1.3678730389801785, 'learning_rate/qwen_vl_interface': 9.87123377921569e-06, 'learning_rate/action_model': 9.871233779215689e-05, 'epoch': 5.58}) +[2026-06-20 13:25:12,720][starVLA.training.train_starvla][INFO] - Step 3065, Loss: {'train/loss': 0.11904959380626678, 'train/grad_norm_pre_clip': 7.816854953765869, 'timing/data': 0.011364683974534273, 'timing/model': 1.2215285639977083, 'learning_rate/qwen_vl_interface': 9.870528536810825e-06, 'learning_rate/action_model': 9.870528536810825e-05, 'epoch': 5.58}) +[2026-06-20 13:25:17,799][starVLA.training.train_starvla][INFO] - Step 3066, Loss: {'train/loss': 0.208317369222641, 'train/grad_norm_pre_clip': 19.198944091796875, 'timing/data': 0.0116356429643929, 'timing/model': 1.506167521001771, 'learning_rate/qwen_vl_interface': 9.869821395081214e-06, 'learning_rate/action_model': 9.869821395081214e-05, 'epoch': 5.58}) +[2026-06-20 13:25:22,649][starVLA.training.train_starvla][INFO] - Step 3067, Loss: {'train/loss': 0.0778198093175888, 'train/grad_norm_pre_clip': 9.298876762390137, 'timing/data': 0.013527614995837212, 'timing/model': 1.2454754329519346, 'learning_rate/qwen_vl_interface': 9.869112354317532e-06, 'learning_rate/action_model': 9.869112354317532e-05, 'epoch': 5.58}) +[2026-06-20 13:25:27,694][starVLA.training.train_starvla][INFO] - Step 3068, Loss: {'train/loss': 0.10574354231357574, 'train/grad_norm_pre_clip': 10.085785865783691, 'timing/data': 0.014458341058343649, 'timing/model': 1.2319144109496847, 'learning_rate/qwen_vl_interface': 9.86840141481124e-06, 'learning_rate/action_model': 9.868401414811239e-05, 'epoch': 5.58}) +[2026-06-20 13:25:33,066][starVLA.training.train_starvla][INFO] - Step 3069, Loss: {'train/loss': 0.10542966425418854, 'train/grad_norm_pre_clip': 10.231301307678223, 'timing/data': 0.012841127929277718, 'timing/model': 1.4181424470152706, 'learning_rate/qwen_vl_interface': 9.867688576854578e-06, 'learning_rate/action_model': 9.867688576854578e-05, 'epoch': 5.59}) +[2026-06-20 13:25:37,979][starVLA.training.train_starvla][INFO] - Step 3070, Loss: {'train/loss': 0.14550364017486572, 'train/grad_norm_pre_clip': 21.78631019592285, 'timing/data': 0.011620362987741828, 'timing/model': 1.2158765320200473, 'learning_rate/qwen_vl_interface': 9.86697384074057e-06, 'learning_rate/action_model': 9.866973840740568e-05, 'epoch': 5.59}) +[2026-06-20 13:25:42,884][starVLA.training.train_starvla][INFO] - Step 3071, Loss: {'train/loss': 0.1932278722524643, 'train/grad_norm_pre_clip': 4.039425849914551, 'timing/data': 0.017990277032367885, 'timing/model': 1.2354911540169269, 'learning_rate/qwen_vl_interface': 9.86625720676301e-06, 'learning_rate/action_model': 9.86625720676301e-05, 'epoch': 5.59}) +[2026-06-20 13:25:47,987][starVLA.training.train_starvla][INFO] - Step 3072, Loss: {'train/loss': 0.2805006504058838, 'train/grad_norm_pre_clip': 26.255409240722656, 'timing/data': 0.015343238017521799, 'timing/model': 1.4932284749811515, 'learning_rate/qwen_vl_interface': 9.865538675216488e-06, 'learning_rate/action_model': 9.865538675216486e-05, 'epoch': 5.59}) +[2026-06-20 13:25:52,868][starVLA.training.train_starvla][INFO] - Step 3073, Loss: {'train/loss': 0.010982242412865162, 'train/grad_norm_pre_clip': 1.7975772619247437, 'timing/data': 0.012746429070830345, 'timing/model': 1.2246328020701185, 'learning_rate/qwen_vl_interface': 9.864818246396358e-06, 'learning_rate/action_model': 9.864818246396358e-05, 'epoch': 5.59}) +[2026-06-20 13:25:58,047][starVLA.training.train_starvla][INFO] - Step 3074, Loss: {'train/loss': 0.11205843836069107, 'train/grad_norm_pre_clip': 13.367505073547363, 'timing/data': 0.012956207967363298, 'timing/model': 1.252513692015782, 'learning_rate/qwen_vl_interface': 9.864095920598764e-06, 'learning_rate/action_model': 9.864095920598763e-05, 'epoch': 5.6}) +[2026-06-20 13:26:03,001][starVLA.training.train_starvla][INFO] - Step 3075, Loss: {'train/loss': 0.3414159417152405, 'train/grad_norm_pre_clip': 17.760042190551758, 'timing/data': 0.012632219004444778, 'timing/model': 1.2019743479322642, 'learning_rate/qwen_vl_interface': 9.86337169812063e-06, 'learning_rate/action_model': 9.863371698120629e-05, 'epoch': 5.6}) +[2026-06-20 13:26:07,840][starVLA.training.train_starvla][INFO] - Step 3076, Loss: {'train/loss': 0.13466626405715942, 'train/grad_norm_pre_clip': 7.473911762237549, 'timing/data': 0.012911207973957062, 'timing/model': 1.197818943997845, 'learning_rate/qwen_vl_interface': 9.862645579259648e-06, 'learning_rate/action_model': 9.862645579259648e-05, 'epoch': 5.6}) +[2026-06-20 13:26:12,885][starVLA.training.train_starvla][INFO] - Step 3077, Loss: {'train/loss': 0.11088915914297104, 'train/grad_norm_pre_clip': 6.633611679077148, 'timing/data': 0.011626093997620046, 'timing/model': 1.2281561859417707, 'learning_rate/qwen_vl_interface': 9.861917564314308e-06, 'learning_rate/action_model': 9.861917564314308e-05, 'epoch': 5.6}) +[2026-06-20 13:26:17,748][starVLA.training.train_starvla][INFO] - Step 3078, Loss: {'train/loss': 0.10672590881586075, 'train/grad_norm_pre_clip': 15.156054496765137, 'timing/data': 0.011368674924597144, 'timing/model': 1.4341844080481678, 'learning_rate/qwen_vl_interface': 9.861187653583862e-06, 'learning_rate/action_model': 9.861187653583862e-05, 'epoch': 5.6}) +[2026-06-20 13:26:22,324][starVLA.training.train_starvla][INFO] - Step 3079, Loss: {'train/loss': 0.18180866539478302, 'train/grad_norm_pre_clip': 13.34056282043457, 'timing/data': 0.011720383074134588, 'timing/model': 1.177059149951674, 'learning_rate/qwen_vl_interface': 9.860455847368352e-06, 'learning_rate/action_model': 9.860455847368351e-05, 'epoch': 5.6}) +[2026-06-20 13:26:27,255][starVLA.training.train_starvla][INFO] - Step 3080, Loss: {'train/loss': 0.19998100399971008, 'train/grad_norm_pre_clip': 22.01089096069336, 'timing/data': 0.01423557405360043, 'timing/model': 1.2014807749073952, 'learning_rate/qwen_vl_interface': 9.859722145968595e-06, 'learning_rate/action_model': 9.859722145968595e-05, 'epoch': 5.61}) +[2026-06-20 13:26:32,088][starVLA.training.train_starvla][INFO] - Step 3081, Loss: {'train/loss': 0.06994982808828354, 'train/grad_norm_pre_clip': 10.019118309020996, 'timing/data': 0.011144766001962125, 'timing/model': 1.1859164870111272, 'learning_rate/qwen_vl_interface': 9.858986549686188e-06, 'learning_rate/action_model': 9.858986549686188e-05, 'epoch': 5.61}) +[2026-06-20 13:26:36,924][starVLA.training.train_starvla][INFO] - Step 3082, Loss: {'train/loss': 0.3769880533218384, 'train/grad_norm_pre_clip': 12.440704345703125, 'timing/data': 0.01136285497341305, 'timing/model': 1.2055639589671046, 'learning_rate/qwen_vl_interface': 9.85824905882351e-06, 'learning_rate/action_model': 9.858249058823509e-05, 'epoch': 5.61}) +[2026-06-20 13:26:41,863][starVLA.training.train_starvla][INFO] - Step 3083, Loss: {'train/loss': 0.1849372833967209, 'train/grad_norm_pre_clip': 14.860264778137207, 'timing/data': 0.011431065038777888, 'timing/model': 1.2215502969920635, 'learning_rate/qwen_vl_interface': 9.857509673683709e-06, 'learning_rate/action_model': 9.857509673683708e-05, 'epoch': 5.61}) +[2026-06-20 13:26:46,776][starVLA.training.train_starvla][INFO] - Step 3084, Loss: {'train/loss': 0.1476985514163971, 'train/grad_norm_pre_clip': 11.352641105651855, 'timing/data': 0.013801465975120664, 'timing/model': 1.4336601659888402, 'learning_rate/qwen_vl_interface': 9.856768394570725e-06, 'learning_rate/action_model': 9.856768394570723e-05, 'epoch': 5.61}) +[2026-06-20 13:26:51,439][starVLA.training.train_starvla][INFO] - Step 3085, Loss: {'train/loss': 0.18567292392253876, 'train/grad_norm_pre_clip': 7.14225959777832, 'timing/data': 0.01181943400297314, 'timing/model': 1.1819001539843157, 'learning_rate/qwen_vl_interface': 9.856025221789264e-06, 'learning_rate/action_model': 9.856025221789264e-05, 'epoch': 5.62}) +[2026-06-20 13:26:56,328][starVLA.training.train_starvla][INFO] - Step 3086, Loss: {'train/loss': 0.03876679018139839, 'train/grad_norm_pre_clip': 3.363090753555298, 'timing/data': 0.011841693078167737, 'timing/model': 1.1849847909761593, 'learning_rate/qwen_vl_interface': 9.85528015564482e-06, 'learning_rate/action_model': 9.85528015564482e-05, 'epoch': 5.62}) +[2026-06-20 13:27:01,369][starVLA.training.train_starvla][INFO] - Step 3087, Loss: {'train/loss': 0.1470639407634735, 'train/grad_norm_pre_clip': 11.007747650146484, 'timing/data': 0.014723562984727323, 'timing/model': 1.2343703389633447, 'learning_rate/qwen_vl_interface': 9.854533196443661e-06, 'learning_rate/action_model': 9.854533196443661e-05, 'epoch': 5.62}) +[2026-06-20 13:27:06,436][starVLA.training.train_starvla][INFO] - Step 3088, Loss: {'train/loss': 0.033199094235897064, 'train/grad_norm_pre_clip': 2.25113844871521, 'timing/data': 0.011716043925844133, 'timing/model': 1.198080311063677, 'learning_rate/qwen_vl_interface': 9.853784344492833e-06, 'learning_rate/action_model': 9.853784344492833e-05, 'epoch': 5.62}) +[2026-06-20 13:27:11,302][starVLA.training.train_starvla][INFO] - Step 3089, Loss: {'train/loss': 0.1978251188993454, 'train/grad_norm_pre_clip': 11.142056465148926, 'timing/data': 0.015210650977678597, 'timing/model': 1.1914816070348024, 'learning_rate/qwen_vl_interface': 9.853033600100162e-06, 'learning_rate/action_model': 9.853033600100162e-05, 'epoch': 5.62}) +[2026-06-20 13:27:16,147][starVLA.training.train_starvla][INFO] - Step 3090, Loss: {'train/loss': 0.12621931731700897, 'train/grad_norm_pre_clip': 10.761423110961914, 'timing/data': 0.012122962973080575, 'timing/model': 1.4678065070183948, 'learning_rate/qwen_vl_interface': 9.85228096357425e-06, 'learning_rate/action_model': 9.852280963574249e-05, 'epoch': 5.62}) +[2026-06-20 13:27:21,128][starVLA.training.train_starvla][INFO] - Step 3091, Loss: {'train/loss': 0.2949501872062683, 'train/grad_norm_pre_clip': 18.333505630493164, 'timing/data': 0.01123198599088937, 'timing/model': 1.4061874760081992, 'learning_rate/qwen_vl_interface': 9.851526435224475e-06, 'learning_rate/action_model': 9.851526435224474e-05, 'epoch': 5.63}) +[2026-06-20 13:27:25,691][starVLA.training.train_starvla][INFO] - Step 3092, Loss: {'train/loss': 0.10690207034349442, 'train/grad_norm_pre_clip': 9.987529754638672, 'timing/data': 0.011931143002584577, 'timing/model': 1.1902753800386563, 'learning_rate/qwen_vl_interface': 9.850770015360999e-06, 'learning_rate/action_model': 9.850770015360999e-05, 'epoch': 5.63}) +[2026-06-20 13:27:30,554][starVLA.training.train_starvla][INFO] - Step 3093, Loss: {'train/loss': 0.32474222779273987, 'train/grad_norm_pre_clip': 18.131786346435547, 'timing/data': 0.012098904000595212, 'timing/model': 1.196481584920548, 'learning_rate/qwen_vl_interface': 9.850011704294755e-06, 'learning_rate/action_model': 9.850011704294755e-05, 'epoch': 5.63}) +[2026-06-20 13:27:35,782][starVLA.training.train_starvla][INFO] - Step 3094, Loss: {'train/loss': 0.02391902171075344, 'train/grad_norm_pre_clip': 2.451601028442383, 'timing/data': 0.013900986057706177, 'timing/model': 1.4675107939401641, 'learning_rate/qwen_vl_interface': 9.849251502337458e-06, 'learning_rate/action_model': 9.849251502337458e-05, 'epoch': 5.63}) +[2026-06-20 13:27:41,039][starVLA.training.train_starvla][INFO] - Step 3095, Loss: {'train/loss': 0.07333626598119736, 'train/grad_norm_pre_clip': 3.5008387565612793, 'timing/data': 0.012632010970264673, 'timing/model': 1.1920409130398184, 'learning_rate/qwen_vl_interface': 9.848489409801598e-06, 'learning_rate/action_model': 9.848489409801598e-05, 'epoch': 5.63}) +[2026-06-20 13:27:45,960][starVLA.training.train_starvla][INFO] - Step 3096, Loss: {'train/loss': 0.09616869688034058, 'train/grad_norm_pre_clip': 10.773356437683105, 'timing/data': 0.01145753706805408, 'timing/model': 1.45246917498298, 'learning_rate/qwen_vl_interface': 9.84772542700044e-06, 'learning_rate/action_model': 9.84772542700044e-05, 'epoch': 5.64}) +[2026-06-20 13:27:51,319][starVLA.training.train_starvla][INFO] - Step 3097, Loss: {'train/loss': 0.08691611886024475, 'train/grad_norm_pre_clip': 6.020015716552734, 'timing/data': 0.01273098192177713, 'timing/model': 1.4865277860080823, 'learning_rate/qwen_vl_interface': 9.846959554248032e-06, 'learning_rate/action_model': 9.84695955424803e-05, 'epoch': 5.64}) +[2026-06-20 13:27:56,184][starVLA.training.train_starvla][INFO] - Step 3098, Loss: {'train/loss': 0.13085585832595825, 'train/grad_norm_pre_clip': 11.024792671203613, 'timing/data': 0.013088209903798997, 'timing/model': 1.2490565719781443, 'learning_rate/qwen_vl_interface': 9.846191791859192e-06, 'learning_rate/action_model': 9.846191791859192e-05, 'epoch': 5.64}) +[2026-06-20 13:28:01,529][starVLA.training.train_starvla][INFO] - Step 3099, Loss: {'train/loss': 0.21586725115776062, 'train/grad_norm_pre_clip': 15.355889320373535, 'timing/data': 0.011629496002569795, 'timing/model': 1.33581790397875, 'learning_rate/qwen_vl_interface': 9.845422140149518e-06, 'learning_rate/action_model': 9.845422140149519e-05, 'epoch': 5.64}) +[2026-06-20 13:28:06,600][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 3100 +[2026-06-20 13:34:30,467][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 3100 +[2026-06-20 13:34:30,468][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 3100 +[2026-06-20 13:40:54,306][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 3100 +[2026-06-20 13:40:54,308][starVLA.training.train_starvla][INFO] - Step 3100, Loss: {'train/loss': 0.1625230461359024, 'train/grad_norm_pre_clip': 13.324094772338867, 'eval/loss': 0.4880538880825043, 'eval/latency_0/loss': 0.4345249533653259, 'eval/latency_1/loss': 0.6582164168357849, 'eval/latency_2/loss': 0.637844979763031, 'eval/latency_3/loss': 0.12518835067749023, 'eval/latency_4/loss': 0.5376055836677551, 'eval/flappy/loss': 0.4880538880825043, 'eval/flappy/latency_0/loss': 0.4345249533653259, 'eval/flappy/latency_1/loss': 0.6582164168357849, 'eval/flappy/latency_2/loss': 0.637844979763031, 'eval/flappy/latency_3/loss': 0.12518835067749023, 'eval/flappy/latency_4/loss': 0.5376055836677551, 'eval/flap_precision': 0.3135593220338983, 'eval/flap_recall': 0.13454545454545455, 'eval/flap_f1': 0.18829516539440203, 'eval/flap_rate_model': 0.03277777777777778, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.18829516539440203, 'eval/latency_0/flap_precision': 0.3135593220338983, 'eval/latency_0/flap_recall': 0.13454545454545455, 'eval/latency_0/flap_f1': 0.18829516539440203, 'eval/latency_0/flap_rate_model': 0.03277777777777778, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.18829516539440203, 'timing/data': 0.011382956989109516, 'timing/model': 1.2350776849780232, 'learning_rate/qwen_vl_interface': 9.84465059943539e-06, 'learning_rate/action_model': 9.84465059943539e-05, 'epoch': 5.64}) +[2026-06-20 13:40:54,308][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_3100_state +[2026-06-20 13:40:54,312][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-20 13:41:13,048][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_3100_state/model.safetensors +[2026-06-20 13:41:30,492][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_3100_state/optimizer.bin +[2026-06-20 13:41:30,492][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_3100_state/sampler.bin +[2026-06-20 13:41:30,492][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_3100_state/sampler_1.bin +[2026-06-20 13:41:30,493][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_3100_state/random_states_0.pkl +[2026-06-20 13:41:31,945][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3100 +[2026-06-20 13:41:31,945][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-20 13:41:31,959][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-20 13:41:36,756][starVLA.training.train_starvla][INFO] - Step 3101, Loss: {'train/loss': 0.19410698115825653, 'train/grad_norm_pre_clip': 11.216485977172852, 'timing/data': 0.011911942041479051, 'timing/model': 1.1592136579565704, 'learning_rate/qwen_vl_interface': 9.843877170033956e-06, 'learning_rate/action_model': 9.843877170033954e-05, 'epoch': 5.64}) +[2026-06-20 13:41:41,485][starVLA.training.train_starvla][INFO] - Step 3102, Loss: {'train/loss': 0.048377297818660736, 'train/grad_norm_pre_clip': 17.83953285217285, 'timing/data': 0.011571462964639068, 'timing/model': 1.3998168159741908, 'learning_rate/qwen_vl_interface': 9.84310185226314e-06, 'learning_rate/action_model': 9.84310185226314e-05, 'epoch': 5.65}) +[2026-06-20 13:41:46,120][starVLA.training.train_starvla][INFO] - Step 3103, Loss: {'train/loss': 0.07372260093688965, 'train/grad_norm_pre_clip': 9.948348999023438, 'timing/data': 0.017151725944131613, 'timing/model': 1.2200605779653415, 'learning_rate/qwen_vl_interface': 9.842324646441652e-06, 'learning_rate/action_model': 9.842324646441652e-05, 'epoch': 5.65}) +[2026-06-20 13:41:50,993][starVLA.training.train_starvla][INFO] - Step 3104, Loss: {'train/loss': 0.05269517004489899, 'train/grad_norm_pre_clip': 8.6284761428833, 'timing/data': 0.011899831937626004, 'timing/model': 1.18272211891599, 'learning_rate/qwen_vl_interface': 9.841545552888968e-06, 'learning_rate/action_model': 9.841545552888967e-05, 'epoch': 5.65}) +[2026-06-20 13:41:55,856][starVLA.training.train_starvla][INFO] - Step 3105, Loss: {'train/loss': 0.2615777254104614, 'train/grad_norm_pre_clip': 12.922093391418457, 'timing/data': 0.011377874063327909, 'timing/model': 1.1951347109861672, 'learning_rate/qwen_vl_interface': 9.840764571925346e-06, 'learning_rate/action_model': 9.840764571925345e-05, 'epoch': 5.65}) +[2026-06-20 13:42:00,668][starVLA.training.train_starvla][INFO] - Step 3106, Loss: {'train/loss': 0.12543904781341553, 'train/grad_norm_pre_clip': 16.430416107177734, 'timing/data': 0.011692333035171032, 'timing/model': 1.170869478955865, 'learning_rate/qwen_vl_interface': 9.839981703871815e-06, 'learning_rate/action_model': 9.839981703871814e-05, 'epoch': 5.65}) +[2026-06-20 13:42:05,418][starVLA.training.train_starvla][INFO] - Step 3107, Loss: {'train/loss': 0.2232871949672699, 'train/grad_norm_pre_clip': 17.07534408569336, 'timing/data': 0.013137758010998368, 'timing/model': 1.1806816200260073, 'learning_rate/qwen_vl_interface': 9.839196949050185e-06, 'learning_rate/action_model': 9.839196949050185e-05, 'epoch': 5.66}) +[2026-06-20 13:42:10,254][starVLA.training.train_starvla][INFO] - Step 3108, Loss: {'train/loss': 0.2677016854286194, 'train/grad_norm_pre_clip': 18.338905334472656, 'timing/data': 0.01118349505122751, 'timing/model': 1.4417847950244322, 'learning_rate/qwen_vl_interface': 9.838410307783038e-06, 'learning_rate/action_model': 9.838410307783038e-05, 'epoch': 5.66}) +[2026-06-20 13:42:14,835][starVLA.training.train_starvla][INFO] - Step 3109, Loss: {'train/loss': 0.07012468576431274, 'train/grad_norm_pre_clip': 15.00636100769043, 'timing/data': 0.014044736046344042, 'timing/model': 1.187770911026746, 'learning_rate/qwen_vl_interface': 9.837621780393734e-06, 'learning_rate/action_model': 9.837621780393733e-05, 'epoch': 5.66}) +[2026-06-20 13:42:19,728][starVLA.training.train_starvla][INFO] - Step 3110, Loss: {'train/loss': 0.4067729115486145, 'train/grad_norm_pre_clip': 18.261890411376953, 'timing/data': 0.014376695035025477, 'timing/model': 1.2711862380383536, 'learning_rate/qwen_vl_interface': 9.836831367206406e-06, 'learning_rate/action_model': 9.836831367206406e-05, 'epoch': 5.66}) +[2026-06-20 13:42:24,811][starVLA.training.train_starvla][INFO] - Step 3111, Loss: {'train/loss': 0.023389562964439392, 'train/grad_norm_pre_clip': 3.206918239593506, 'timing/data': 0.01710062602069229, 'timing/model': 1.2394392209826037, 'learning_rate/qwen_vl_interface': 9.836039068545962e-06, 'learning_rate/action_model': 9.836039068545961e-05, 'epoch': 5.66}) +[2026-06-20 13:42:29,716][starVLA.training.train_starvla][INFO] - Step 3112, Loss: {'train/loss': 0.22619110345840454, 'train/grad_norm_pre_clip': 19.140626907348633, 'timing/data': 0.011150915059261024, 'timing/model': 1.1694009830243886, 'learning_rate/qwen_vl_interface': 9.835244884738086e-06, 'learning_rate/action_model': 9.835244884738085e-05, 'epoch': 5.66}) +[2026-06-20 13:42:34,679][starVLA.training.train_starvla][INFO] - Step 3113, Loss: {'train/loss': 0.2576438784599304, 'train/grad_norm_pre_clip': 17.5583438873291, 'timing/data': 0.013055417919531465, 'timing/model': 1.209650448989123, 'learning_rate/qwen_vl_interface': 9.83444881610924e-06, 'learning_rate/action_model': 9.834448816109238e-05, 'epoch': 5.67}) +[2026-06-20 13:42:39,520][starVLA.training.train_starvla][INFO] - Step 3114, Loss: {'train/loss': 0.04318558797240257, 'train/grad_norm_pre_clip': 5.547231674194336, 'timing/data': 0.011313884053379297, 'timing/model': 1.4386672569671646, 'learning_rate/qwen_vl_interface': 9.833650862986652e-06, 'learning_rate/action_model': 9.833650862986653e-05, 'epoch': 5.67}) +[2026-06-20 13:42:44,262][starVLA.training.train_starvla][INFO] - Step 3115, Loss: {'train/loss': 0.1641186624765396, 'train/grad_norm_pre_clip': 11.856980323791504, 'timing/data': 0.012456769938580692, 'timing/model': 1.2105082490015775, 'learning_rate/qwen_vl_interface': 9.832851025698335e-06, 'learning_rate/action_model': 9.832851025698335e-05, 'epoch': 5.67}) +[2026-06-20 13:42:49,237][starVLA.training.train_starvla][INFO] - Step 3116, Loss: {'train/loss': 0.030665621161460876, 'train/grad_norm_pre_clip': 4.358555793762207, 'timing/data': 0.01159639295656234, 'timing/model': 1.1868270459817722, 'learning_rate/qwen_vl_interface': 9.832049304573072e-06, 'learning_rate/action_model': 9.832049304573072e-05, 'epoch': 5.67}) +[2026-06-20 13:42:54,441][starVLA.training.train_starvla][INFO] - Step 3117, Loss: {'train/loss': 0.1548043042421341, 'train/grad_norm_pre_clip': 17.83754539489746, 'timing/data': 0.014471794944256544, 'timing/model': 1.408987155999057, 'learning_rate/qwen_vl_interface': 9.831245699940419e-06, 'learning_rate/action_model': 9.831245699940418e-05, 'epoch': 5.67}) +[2026-06-20 13:42:59,521][starVLA.training.train_starvla][INFO] - Step 3118, Loss: {'train/loss': 0.10708185285329819, 'train/grad_norm_pre_clip': 15.247824668884277, 'timing/data': 0.012446600943803787, 'timing/model': 1.3023270739940926, 'learning_rate/qwen_vl_interface': 9.830440212130707e-06, 'learning_rate/action_model': 9.830440212130708e-05, 'epoch': 5.68}) +[2026-06-20 13:43:04,436][starVLA.training.train_starvla][INFO] - Step 3119, Loss: {'train/loss': 0.04857255145907402, 'train/grad_norm_pre_clip': 9.442766189575195, 'timing/data': 0.011717743007466197, 'timing/model': 1.3104430410312489, 'learning_rate/qwen_vl_interface': 9.829632841475041e-06, 'learning_rate/action_model': 9.829632841475041e-05, 'epoch': 5.68}) +[2026-06-20 13:43:10,119][starVLA.training.train_starvla][INFO] - Step 3120, Loss: {'train/loss': 0.13553079962730408, 'train/grad_norm_pre_clip': 14.940738677978516, 'timing/data': 0.011453984072431922, 'timing/model': 1.5952189040835947, 'learning_rate/qwen_vl_interface': 9.828823588305305e-06, 'learning_rate/action_model': 9.828823588305303e-05, 'epoch': 5.68}) +[2026-06-20 13:43:14,931][starVLA.training.train_starvla][INFO] - Step 3121, Loss: {'train/loss': 0.2585063576698303, 'train/grad_norm_pre_clip': 19.385007858276367, 'timing/data': 0.011968711973167956, 'timing/model': 1.2217275929870084, 'learning_rate/qwen_vl_interface': 9.828012452954146e-06, 'learning_rate/action_model': 9.828012452954145e-05, 'epoch': 5.68}) +[2026-06-20 13:43:20,248][starVLA.training.train_starvla][INFO] - Step 3122, Loss: {'train/loss': 0.1911715865135193, 'train/grad_norm_pre_clip': 15.71951675415039, 'timing/data': 0.014140396029688418, 'timing/model': 1.1954123240429908, 'learning_rate/qwen_vl_interface': 9.827199435754995e-06, 'learning_rate/action_model': 9.827199435754995e-05, 'epoch': 5.68}) +[2026-06-20 13:43:25,112][starVLA.training.train_starvla][INFO] - Step 3123, Loss: {'train/loss': 0.2032734751701355, 'train/grad_norm_pre_clip': 15.314949989318848, 'timing/data': 0.011185295064933598, 'timing/model': 1.192004879936576, 'learning_rate/qwen_vl_interface': 9.826384537042052e-06, 'learning_rate/action_model': 9.826384537042052e-05, 'epoch': 5.68}) +[2026-06-20 13:43:30,055][starVLA.training.train_starvla][INFO] - Step 3124, Loss: {'train/loss': 0.35140344500541687, 'train/grad_norm_pre_clip': 23.823348999023438, 'timing/data': 0.01147612405475229, 'timing/model': 1.3246003340464085, 'learning_rate/qwen_vl_interface': 9.825567757150291e-06, 'learning_rate/action_model': 9.82556775715029e-05, 'epoch': 5.69}) +[2026-06-20 13:43:34,992][starVLA.training.train_starvla][INFO] - Step 3125, Loss: {'train/loss': 0.2647741734981537, 'train/grad_norm_pre_clip': 18.77605438232422, 'timing/data': 0.013542757951654494, 'timing/model': 1.243511799024418, 'learning_rate/qwen_vl_interface': 9.82474909641546e-06, 'learning_rate/action_model': 9.82474909641546e-05, 'epoch': 5.69}) +[2026-06-20 13:43:40,153][starVLA.training.train_starvla][INFO] - Step 3126, Loss: {'train/loss': 0.21127338707447052, 'train/grad_norm_pre_clip': 24.076313018798828, 'timing/data': 0.013005040003918111, 'timing/model': 1.6553996379952878, 'learning_rate/qwen_vl_interface': 9.823928555174078e-06, 'learning_rate/action_model': 9.823928555174077e-05, 'epoch': 5.69}) +[2026-06-20 13:43:44,828][starVLA.training.train_starvla][INFO] - Step 3127, Loss: {'train/loss': 0.04006994515657425, 'train/grad_norm_pre_clip': 6.412690162658691, 'timing/data': 0.024616802926175296, 'timing/model': 1.2288499489659443, 'learning_rate/qwen_vl_interface': 9.82310613376344e-06, 'learning_rate/action_model': 9.82310613376344e-05, 'epoch': 5.69}) +[2026-06-20 13:43:49,672][starVLA.training.train_starvla][INFO] - Step 3128, Loss: {'train/loss': 0.259031742811203, 'train/grad_norm_pre_clip': 24.67473793029785, 'timing/data': 0.01016548799816519, 'timing/model': 1.1855299780145288, 'learning_rate/qwen_vl_interface': 9.822281832521613e-06, 'learning_rate/action_model': 9.822281832521613e-05, 'epoch': 5.69}) +[2026-06-20 13:43:54,587][starVLA.training.train_starvla][INFO] - Step 3129, Loss: {'train/loss': 0.5164490938186646, 'train/grad_norm_pre_clip': 18.723590850830078, 'timing/data': 0.011962143005803227, 'timing/model': 1.1986694680526853, 'learning_rate/qwen_vl_interface': 9.821455651787432e-06, 'learning_rate/action_model': 9.821455651787432e-05, 'epoch': 5.7}) +[2026-06-20 13:43:59,793][starVLA.training.train_starvla][INFO] - Step 3130, Loss: {'train/loss': 0.051448360085487366, 'train/grad_norm_pre_clip': 6.127295017242432, 'timing/data': 0.011999742011539638, 'timing/model': 1.377480329014361, 'learning_rate/qwen_vl_interface': 9.820627591900514e-06, 'learning_rate/action_model': 9.820627591900513e-05, 'epoch': 5.7}) +[2026-06-20 13:44:04,667][starVLA.training.train_starvla][INFO] - Step 3131, Loss: {'train/loss': 0.21646158397197723, 'train/grad_norm_pre_clip': 24.206876754760742, 'timing/data': 0.012455020099878311, 'timing/model': 1.2017907409463078, 'learning_rate/qwen_vl_interface': 9.819797653201238e-06, 'learning_rate/action_model': 9.819797653201238e-05, 'epoch': 5.7}) +[2026-06-20 13:44:09,520][starVLA.training.train_starvla][INFO] - Step 3132, Loss: {'train/loss': 0.06502962112426758, 'train/grad_norm_pre_clip': 8.162653923034668, 'timing/data': 0.012897499022074044, 'timing/model': 1.4571988920215517, 'learning_rate/qwen_vl_interface': 9.818965836030765e-06, 'learning_rate/action_model': 9.818965836030764e-05, 'epoch': 5.7}) +[2026-06-20 13:44:14,181][starVLA.training.train_starvla][INFO] - Step 3133, Loss: {'train/loss': 0.22348587214946747, 'train/grad_norm_pre_clip': 61.641231536865234, 'timing/data': 0.01139775407500565, 'timing/model': 1.2298053649719805, 'learning_rate/qwen_vl_interface': 9.81813214073102e-06, 'learning_rate/action_model': 9.81813214073102e-05, 'epoch': 5.7}) +[2026-06-20 13:44:19,037][starVLA.training.train_starvla][INFO] - Step 3134, Loss: {'train/loss': 0.0808405950665474, 'train/grad_norm_pre_clip': 12.967718124389648, 'timing/data': 0.011341263889335096, 'timing/model': 1.2091988219181076, 'learning_rate/qwen_vl_interface': 9.817296567644709e-06, 'learning_rate/action_model': 9.817296567644709e-05, 'epoch': 5.7}) +[2026-06-20 13:44:23,888][starVLA.training.train_starvla][INFO] - Step 3135, Loss: {'train/loss': 0.16803012788295746, 'train/grad_norm_pre_clip': 10.89002799987793, 'timing/data': 0.011981313000433147, 'timing/model': 1.2015539960702881, 'learning_rate/qwen_vl_interface': 9.816459117115296e-06, 'learning_rate/action_model': 9.816459117115297e-05, 'epoch': 5.71}) +[2026-06-20 13:44:28,841][starVLA.training.train_starvla][INFO] - Step 3136, Loss: {'train/loss': 0.04062064737081528, 'train/grad_norm_pre_clip': 4.264141082763672, 'timing/data': 0.011521084001287818, 'timing/model': 1.2415262319846079, 'learning_rate/qwen_vl_interface': 9.815619789487035e-06, 'learning_rate/action_model': 9.815619789487034e-05, 'epoch': 5.71}) +[2026-06-20 13:44:33,932][starVLA.training.train_starvla][INFO] - Step 3137, Loss: {'train/loss': 0.3315454125404358, 'train/grad_norm_pre_clip': 13.44291877746582, 'timing/data': 0.016086669056676328, 'timing/model': 1.2427577510243282, 'learning_rate/qwen_vl_interface': 9.814778585104933e-06, 'learning_rate/action_model': 9.814778585104933e-05, 'epoch': 5.71}) +[2026-06-20 13:44:38,972][starVLA.training.train_starvla][INFO] - Step 3138, Loss: {'train/loss': 0.35766685009002686, 'train/grad_norm_pre_clip': 21.748802185058594, 'timing/data': 0.013184508075937629, 'timing/model': 1.4936304669827223, 'learning_rate/qwen_vl_interface': 9.813935504314784e-06, 'learning_rate/action_model': 9.813935504314784e-05, 'epoch': 5.71}) +[2026-06-20 13:44:43,641][starVLA.training.train_starvla][INFO] - Step 3139, Loss: {'train/loss': 0.4942879378795624, 'train/grad_norm_pre_clip': 16.98537254333496, 'timing/data': 0.006181049975566566, 'timing/model': 1.1815936549101025, 'learning_rate/qwen_vl_interface': 9.813090547463144e-06, 'learning_rate/action_model': 9.813090547463143e-05, 'epoch': 5.71}) +[2026-06-20 13:44:48,597][starVLA.training.train_starvla][INFO] - Step 3140, Loss: {'train/loss': 0.15463677048683167, 'train/grad_norm_pre_clip': 14.369742393493652, 'timing/data': 0.006218030001036823, 'timing/model': 1.2975507320370525, 'learning_rate/qwen_vl_interface': 9.812243714897342e-06, 'learning_rate/action_model': 9.812243714897343e-05, 'epoch': 5.72}) +[2026-06-20 13:44:53,477][starVLA.training.train_starvla][INFO] - Step 3141, Loss: {'train/loss': 0.07982466369867325, 'train/grad_norm_pre_clip': 7.299292087554932, 'timing/data': 0.00607198104262352, 'timing/model': 1.1929807510459796, 'learning_rate/qwen_vl_interface': 9.811395006965481e-06, 'learning_rate/action_model': 9.81139500696548e-05, 'epoch': 5.72}) +[2026-06-20 13:44:58,400][starVLA.training.train_starvla][INFO] - Step 3142, Loss: {'train/loss': 0.22846275568008423, 'train/grad_norm_pre_clip': 11.609786987304688, 'timing/data': 0.006227490957826376, 'timing/model': 1.2558227360714227, 'learning_rate/qwen_vl_interface': 9.81054442401643e-06, 'learning_rate/action_model': 9.81054442401643e-05, 'epoch': 5.72}) +[2026-06-20 13:45:03,341][starVLA.training.train_starvla][INFO] - Step 3143, Loss: {'train/loss': 0.048171646893024445, 'train/grad_norm_pre_clip': 5.223119258880615, 'timing/data': 0.00556843401864171, 'timing/model': 1.1795403059804812, 'learning_rate/qwen_vl_interface': 9.809691966399834e-06, 'learning_rate/action_model': 9.809691966399835e-05, 'epoch': 5.72}) +[2026-06-20 13:45:08,433][starVLA.training.train_starvla][INFO] - Step 3144, Loss: {'train/loss': 0.2169361412525177, 'train/grad_norm_pre_clip': 15.439229011535645, 'timing/data': 0.009829307091422379, 'timing/model': 1.6089091809699312, 'learning_rate/qwen_vl_interface': 9.808837634466104e-06, 'learning_rate/action_model': 9.808837634466104e-05, 'epoch': 5.72}) +[2026-06-20 13:45:13,344][starVLA.training.train_starvla][INFO] - Step 3145, Loss: {'train/loss': 0.3369269371032715, 'train/grad_norm_pre_clip': 11.230364799499512, 'timing/data': 0.0063543799333274364, 'timing/model': 1.4148854160448536, 'learning_rate/qwen_vl_interface': 9.807981428566428e-06, 'learning_rate/action_model': 9.807981428566427e-05, 'epoch': 5.73}) +[2026-06-20 13:45:18,307][starVLA.training.train_starvla][INFO] - Step 3146, Loss: {'train/loss': 0.1233929693698883, 'train/grad_norm_pre_clip': 7.590638160705566, 'timing/data': 0.005611375090666115, 'timing/model': 1.2132182359928265, 'learning_rate/qwen_vl_interface': 9.807123349052754e-06, 'learning_rate/action_model': 9.807123349052753e-05, 'epoch': 5.73}) +[2026-06-20 13:45:23,272][starVLA.training.train_starvla][INFO] - Step 3147, Loss: {'train/loss': 0.08465836197137833, 'train/grad_norm_pre_clip': 8.257462501525879, 'timing/data': 0.0058394529623910785, 'timing/model': 1.2151517579331994, 'learning_rate/qwen_vl_interface': 9.80626339627781e-06, 'learning_rate/action_model': 9.806263396277809e-05, 'epoch': 5.73}) +[2026-06-20 13:45:28,360][starVLA.training.train_starvla][INFO] - Step 3148, Loss: {'train/loss': 0.2603682279586792, 'train/grad_norm_pre_clip': 11.628384590148926, 'timing/data': 0.007571183028630912, 'timing/model': 1.3561992059694603, 'learning_rate/qwen_vl_interface': 9.805401570595087e-06, 'learning_rate/action_model': 9.805401570595087e-05, 'epoch': 5.73}) +[2026-06-20 13:45:33,303][starVLA.training.train_starvla][INFO] - Step 3149, Loss: {'train/loss': 0.23571588099002838, 'train/grad_norm_pre_clip': 17.387374877929688, 'timing/data': 0.006692679016850889, 'timing/model': 1.2256351719843224, 'learning_rate/qwen_vl_interface': 9.804537872358853e-06, 'learning_rate/action_model': 9.804537872358853e-05, 'epoch': 5.73}) +[2026-06-20 13:46:23,412][starVLA.training.train_starvla][INFO] - Step 3150, Loss: {'train/loss': 0.39822283387184143, 'train/grad_norm_pre_clip': 12.315004348754883, 'timing/data': 0.0005431859754025936, 'timing/model': 1.5117202639812604, 'learning_rate/qwen_vl_interface': 9.80367230192414e-06, 'learning_rate/action_model': 9.803672301924138e-05, 'epoch': 5.73}) +[2026-06-20 13:46:28,141][starVLA.training.train_starvla][INFO] - Step 3151, Loss: {'train/loss': 0.20233766734600067, 'train/grad_norm_pre_clip': 23.07877540588379, 'timing/data': 0.006615360034629703, 'timing/model': 1.3969438839703798, 'learning_rate/qwen_vl_interface': 9.80280485964675e-06, 'learning_rate/action_model': 9.80280485964675e-05, 'epoch': 5.74}) +[2026-06-20 13:46:32,667][starVLA.training.train_starvla][INFO] - Step 3152, Loss: {'train/loss': 0.08850046992301941, 'train/grad_norm_pre_clip': 7.006867408752441, 'timing/data': 0.006512042018584907, 'timing/model': 1.1856666820822284, 'learning_rate/qwen_vl_interface': 9.801935545883255e-06, 'learning_rate/action_model': 9.801935545883255e-05, 'epoch': 5.74}) +[2026-06-20 13:46:37,388][starVLA.training.train_starvla][INFO] - Step 3153, Loss: {'train/loss': 0.11520922929048538, 'train/grad_norm_pre_clip': 7.8633646965026855, 'timing/data': 0.011993579100817442, 'timing/model': 1.1670597509946674, 'learning_rate/qwen_vl_interface': 9.801064360991e-06, 'learning_rate/action_model': 9.801064360991e-05, 'epoch': 5.74}) +[2026-06-20 13:46:42,129][starVLA.training.train_starvla][INFO] - Step 3154, Loss: {'train/loss': 0.07440092414617538, 'train/grad_norm_pre_clip': 6.341396331787109, 'timing/data': 0.012027518940158188, 'timing/model': 1.1682482910109684, 'learning_rate/qwen_vl_interface': 9.800191305328095e-06, 'learning_rate/action_model': 9.800191305328094e-05, 'epoch': 5.74}) +[2026-06-20 13:46:46,922][starVLA.training.train_starvla][INFO] - Step 3155, Loss: {'train/loss': 0.1033429503440857, 'train/grad_norm_pre_clip': 4.438661098480225, 'timing/data': 0.012121889041736722, 'timing/model': 1.1773874739883468, 'learning_rate/qwen_vl_interface': 9.79931637925342e-06, 'learning_rate/action_model': 9.79931637925342e-05, 'epoch': 5.74}) +[2026-06-20 13:46:51,688][starVLA.training.train_starvla][INFO] - Step 3156, Loss: {'train/loss': 0.17625117301940918, 'train/grad_norm_pre_clip': 10.955156326293945, 'timing/data': 0.011998290079645813, 'timing/model': 1.4102799280080944, 'learning_rate/qwen_vl_interface': 9.798439583126625e-06, 'learning_rate/action_model': 9.798439583126625e-05, 'epoch': 5.75}) +[2026-06-20 13:46:56,188][starVLA.training.train_starvla][INFO] - Step 3157, Loss: {'train/loss': 0.2663078010082245, 'train/grad_norm_pre_clip': 13.390167236328125, 'timing/data': 0.012014930020086467, 'timing/model': 1.171674839919433, 'learning_rate/qwen_vl_interface': 9.797560917308126e-06, 'learning_rate/action_model': 9.797560917308126e-05, 'epoch': 5.75}) +[2026-06-20 13:47:00,935][starVLA.training.train_starvla][INFO] - Step 3158, Loss: {'train/loss': 0.24393607676029205, 'train/grad_norm_pre_clip': 15.090191841125488, 'timing/data': 0.012279147980734706, 'timing/model': 1.166463976027444, 'learning_rate/qwen_vl_interface': 9.796680382159112e-06, 'learning_rate/action_model': 9.79668038215911e-05, 'epoch': 5.75}) +[2026-06-20 13:47:05,685][starVLA.training.train_starvla][INFO] - Step 3159, Loss: {'train/loss': 0.21098878979682922, 'train/grad_norm_pre_clip': 6.578527927398682, 'timing/data': 0.012132450006902218, 'timing/model': 1.1857608909485862, 'learning_rate/qwen_vl_interface': 9.795797978041533e-06, 'learning_rate/action_model': 9.795797978041533e-05, 'epoch': 5.75}) +[2026-06-20 13:47:10,463][starVLA.training.train_starvla][INFO] - Step 3160, Loss: {'train/loss': 0.04299738258123398, 'train/grad_norm_pre_clip': 3.262147903442383, 'timing/data': 0.012213020003400743, 'timing/model': 1.1796366819180548, 'learning_rate/qwen_vl_interface': 9.794913705318117e-06, 'learning_rate/action_model': 9.794913705318118e-05, 'epoch': 5.75}) +[2026-06-20 13:47:15,249][starVLA.training.train_starvla][INFO] - Step 3161, Loss: {'train/loss': 0.19507259130477905, 'train/grad_norm_pre_clip': 11.309762954711914, 'timing/data': 0.012416218989528716, 'timing/model': 1.175628271070309, 'learning_rate/qwen_vl_interface': 9.794027564352352e-06, 'learning_rate/action_model': 9.794027564352352e-05, 'epoch': 5.75}) +[2026-06-20 13:47:20,007][starVLA.training.train_starvla][INFO] - Step 3162, Loss: {'train/loss': 0.07280024886131287, 'train/grad_norm_pre_clip': 10.600898742675781, 'timing/data': 0.012158469995483756, 'timing/model': 1.3950370809761807, 'learning_rate/qwen_vl_interface': 9.793139555508501e-06, 'learning_rate/action_model': 9.7931395555085e-05, 'epoch': 5.76}) +[2026-06-20 13:47:24,570][starVLA.training.train_starvla][INFO] - Step 3163, Loss: {'train/loss': 0.19211052358150482, 'train/grad_norm_pre_clip': 7.332980155944824, 'timing/data': 0.012034671031869948, 'timing/model': 1.172989456099458, 'learning_rate/qwen_vl_interface': 9.792249679151585e-06, 'learning_rate/action_model': 9.792249679151585e-05, 'epoch': 5.76}) +[2026-06-20 13:47:29,446][starVLA.training.train_starvla][INFO] - Step 3164, Loss: {'train/loss': 0.10558668524026871, 'train/grad_norm_pre_clip': 5.443112850189209, 'timing/data': 0.02156651800032705, 'timing/model': 1.2075847249943763, 'learning_rate/qwen_vl_interface': 9.791357935647403e-06, 'learning_rate/action_model': 9.791357935647403e-05, 'epoch': 5.76}) +[2026-06-20 13:47:34,245][starVLA.training.train_starvla][INFO] - Step 3165, Loss: {'train/loss': 0.05322946608066559, 'train/grad_norm_pre_clip': 4.794557094573975, 'timing/data': 0.011907833046279848, 'timing/model': 1.1794793310109526, 'learning_rate/qwen_vl_interface': 9.790464325362516e-06, 'learning_rate/action_model': 9.790464325362516e-05, 'epoch': 5.76}) +[2026-06-20 13:47:39,071][starVLA.training.train_starvla][INFO] - Step 3166, Loss: {'train/loss': 0.105836421251297, 'train/grad_norm_pre_clip': 7.0659708976745605, 'timing/data': 0.012948106043040752, 'timing/model': 1.169184574042447, 'learning_rate/qwen_vl_interface': 9.789568848664254e-06, 'learning_rate/action_model': 9.789568848664253e-05, 'epoch': 5.76}) +[2026-06-20 13:47:43,893][starVLA.training.train_starvla][INFO] - Step 3167, Loss: {'train/loss': 0.03867898881435394, 'train/grad_norm_pre_clip': 4.114421844482422, 'timing/data': 0.011868253001011908, 'timing/model': 1.1895496129291132, 'learning_rate/qwen_vl_interface': 9.788671505920712e-06, 'learning_rate/action_model': 9.788671505920711e-05, 'epoch': 5.77}) +[2026-06-20 13:47:49,179][starVLA.training.train_starvla][INFO] - Step 3168, Loss: {'train/loss': 0.0688161626458168, 'train/grad_norm_pre_clip': 10.802742958068848, 'timing/data': 0.019822109024971724, 'timing/model': 1.6783656169427559, 'learning_rate/qwen_vl_interface': 9.787772297500754e-06, 'learning_rate/action_model': 9.787772297500754e-05, 'epoch': 5.77}) +[2026-06-20 13:47:54,383][starVLA.training.train_starvla][INFO] - Step 3169, Loss: {'train/loss': 0.3794093430042267, 'train/grad_norm_pre_clip': 16.895835876464844, 'timing/data': 0.019025383982807398, 'timing/model': 1.331856093951501, 'learning_rate/qwen_vl_interface': 9.78687122377401e-06, 'learning_rate/action_model': 9.78687122377401e-05, 'epoch': 5.77}) +[2026-06-20 13:47:59,752][starVLA.training.train_starvla][INFO] - Step 3170, Loss: {'train/loss': 0.17552511394023895, 'train/grad_norm_pre_clip': 10.98235034942627, 'timing/data': 0.019285122980363667, 'timing/model': 1.3667149010580033, 'learning_rate/qwen_vl_interface': 9.78596828511088e-06, 'learning_rate/action_model': 9.78596828511088e-05, 'epoch': 5.77}) +[2026-06-20 13:48:05,003][starVLA.training.train_starvla][INFO] - Step 3171, Loss: {'train/loss': 0.10033329576253891, 'train/grad_norm_pre_clip': 8.572196006774902, 'timing/data': 0.013912722934037447, 'timing/model': 1.1752672380534932, 'learning_rate/qwen_vl_interface': 9.785063481882527e-06, 'learning_rate/action_model': 9.785063481882527e-05, 'epoch': 5.77}) +[2026-06-20 13:48:09,977][starVLA.training.train_starvla][INFO] - Step 3172, Loss: {'train/loss': 0.07773502171039581, 'train/grad_norm_pre_clip': 7.01971960067749, 'timing/data': 0.013394396984949708, 'timing/model': 1.2535355910658836, 'learning_rate/qwen_vl_interface': 9.78415681446088e-06, 'learning_rate/action_model': 9.784156814460879e-05, 'epoch': 5.77}) +[2026-06-20 13:48:14,964][starVLA.training.train_starvla][INFO] - Step 3173, Loss: {'train/loss': 0.1832333356142044, 'train/grad_norm_pre_clip': 12.36798095703125, 'timing/data': 0.01731644500978291, 'timing/model': 1.2479524279478937, 'learning_rate/qwen_vl_interface': 9.783248283218638e-06, 'learning_rate/action_model': 9.783248283218638e-05, 'epoch': 5.78}) +[2026-06-20 13:48:19,940][starVLA.training.train_starvla][INFO] - Step 3174, Loss: {'train/loss': 0.102208711206913, 'train/grad_norm_pre_clip': 8.641210556030273, 'timing/data': 0.012331533012911677, 'timing/model': 1.4609744580229744, 'learning_rate/qwen_vl_interface': 9.782337888529261e-06, 'learning_rate/action_model': 9.782337888529261e-05, 'epoch': 5.78}) +[2026-06-20 13:48:24,605][starVLA.training.train_starvla][INFO] - Step 3175, Loss: {'train/loss': 0.17214997112751007, 'train/grad_norm_pre_clip': 15.25583267211914, 'timing/data': 0.011569787049666047, 'timing/model': 1.1995343010639772, 'learning_rate/qwen_vl_interface': 9.781425630766978e-06, 'learning_rate/action_model': 9.781425630766979e-05, 'epoch': 5.78}) +[2026-06-20 13:48:29,906][starVLA.training.train_starvla][INFO] - Step 3176, Loss: {'train/loss': 0.11271986365318298, 'train/grad_norm_pre_clip': 16.660066604614258, 'timing/data': 0.012961090076714754, 'timing/model': 1.2989305260125548, 'learning_rate/qwen_vl_interface': 9.780511510306789e-06, 'learning_rate/action_model': 9.780511510306788e-05, 'epoch': 5.78}) +[2026-06-20 13:48:34,745][starVLA.training.train_starvla][INFO] - Step 3177, Loss: {'train/loss': 0.15997231006622314, 'train/grad_norm_pre_clip': 25.57311248779297, 'timing/data': 0.011964016011916101, 'timing/model': 1.1902963579632342, 'learning_rate/qwen_vl_interface': 9.779595527524447e-06, 'learning_rate/action_model': 9.779595527524447e-05, 'epoch': 5.78}) +[2026-06-20 13:48:39,647][starVLA.training.train_starvla][INFO] - Step 3178, Loss: {'train/loss': 0.09052268415689468, 'train/grad_norm_pre_clip': 11.053271293640137, 'timing/data': 0.012681582011282444, 'timing/model': 1.238055294030346, 'learning_rate/qwen_vl_interface': 9.778677682796483e-06, 'learning_rate/action_model': 9.778677682796483e-05, 'epoch': 5.79}) +[2026-06-20 13:48:44,705][starVLA.training.train_starvla][INFO] - Step 3179, Loss: {'train/loss': 0.07144705206155777, 'train/grad_norm_pre_clip': 10.984807968139648, 'timing/data': 0.011217089951969683, 'timing/model': 1.2117353290086612, 'learning_rate/qwen_vl_interface': 9.777757976500185e-06, 'learning_rate/action_model': 9.777757976500186e-05, 'epoch': 5.79}) +[2026-06-20 13:48:49,773][starVLA.training.train_starvla][INFO] - Step 3180, Loss: {'train/loss': 0.20370031893253326, 'train/grad_norm_pre_clip': 12.121996879577637, 'timing/data': 0.013928284985013306, 'timing/model': 1.5253082760609686, 'learning_rate/qwen_vl_interface': 9.776836409013614e-06, 'learning_rate/action_model': 9.776836409013615e-05, 'epoch': 5.79}) +[2026-06-20 13:48:54,712][starVLA.training.train_starvla][INFO] - Step 3181, Loss: {'train/loss': 0.03347166255116463, 'train/grad_norm_pre_clip': 5.596256732940674, 'timing/data': 0.011768157011829317, 'timing/model': 1.2580251889303327, 'learning_rate/qwen_vl_interface': 9.775912980715589e-06, 'learning_rate/action_model': 9.775912980715588e-05, 'epoch': 5.79}) +[2026-06-20 13:48:59,893][starVLA.training.train_starvla][INFO] - Step 3182, Loss: {'train/loss': 0.3742486834526062, 'train/grad_norm_pre_clip': 14.917435646057129, 'timing/data': 0.012173914932645857, 'timing/model': 1.3283916799118742, 'learning_rate/qwen_vl_interface': 9.774987691985694e-06, 'learning_rate/action_model': 9.774987691985694e-05, 'epoch': 5.79}) +[2026-06-20 13:49:05,817][starVLA.training.train_starvla][INFO] - Step 3183, Loss: {'train/loss': 0.38172635436058044, 'train/grad_norm_pre_clip': 43.64640426635742, 'timing/data': 0.010940762935206294, 'timing/model': 1.4647583180340007, 'learning_rate/qwen_vl_interface': 9.774060543204283e-06, 'learning_rate/action_model': 9.774060543204283e-05, 'epoch': 5.79}) +[2026-06-20 13:49:10,599][starVLA.training.train_starvla][INFO] - Step 3184, Loss: {'train/loss': 0.18317975103855133, 'train/grad_norm_pre_clip': 9.85590934753418, 'timing/data': 0.011609468958340585, 'timing/model': 1.1874915849184617, 'learning_rate/qwen_vl_interface': 9.773131534752472e-06, 'learning_rate/action_model': 9.773131534752472e-05, 'epoch': 5.8}) +[2026-06-20 13:49:15,629][starVLA.training.train_starvla][INFO] - Step 3185, Loss: {'train/loss': 0.30956289172172546, 'train/grad_norm_pre_clip': 13.972548484802246, 'timing/data': 0.01154865010175854, 'timing/model': 1.3088224190287292, 'learning_rate/qwen_vl_interface': 9.772200667012142e-06, 'learning_rate/action_model': 9.772200667012142e-05, 'epoch': 5.8}) +[2026-06-20 13:49:20,485][starVLA.training.train_starvla][INFO] - Step 3186, Loss: {'train/loss': 0.10070929676294327, 'train/grad_norm_pre_clip': 8.309845924377441, 'timing/data': 0.011046431958675385, 'timing/model': 1.4592583049088717, 'learning_rate/qwen_vl_interface': 9.771267940365934e-06, 'learning_rate/action_model': 9.771267940365935e-05, 'epoch': 5.8}) +[2026-06-20 13:49:25,851][starVLA.training.train_starvla][INFO] - Step 3187, Loss: {'train/loss': 0.5867688059806824, 'train/grad_norm_pre_clip': 18.891036987304688, 'timing/data': 0.011345751001499593, 'timing/model': 1.5508948740316555, 'learning_rate/qwen_vl_interface': 9.770333355197262e-06, 'learning_rate/action_model': 9.77033335519726e-05, 'epoch': 5.8}) +[2026-06-20 13:49:30,862][starVLA.training.train_starvla][INFO] - Step 3188, Loss: {'train/loss': 0.12234382331371307, 'train/grad_norm_pre_clip': 12.30612850189209, 'timing/data': 0.01594715705141425, 'timing/model': 1.242245116038248, 'learning_rate/qwen_vl_interface': 9.769396911890291e-06, 'learning_rate/action_model': 9.769396911890291e-05, 'epoch': 5.8}) +[2026-06-20 13:49:35,899][starVLA.training.train_starvla][INFO] - Step 3189, Loss: {'train/loss': 0.2773950397968292, 'train/grad_norm_pre_clip': 17.487430572509766, 'timing/data': 0.012188785942271352, 'timing/model': 1.3199476260924712, 'learning_rate/qwen_vl_interface': 9.768458610829965e-06, 'learning_rate/action_model': 9.768458610829965e-05, 'epoch': 5.81}) +[2026-06-20 13:49:40,716][starVLA.training.train_starvla][INFO] - Step 3190, Loss: {'train/loss': 0.1604171097278595, 'train/grad_norm_pre_clip': 7.360955715179443, 'timing/data': 0.011817009071819484, 'timing/model': 1.174070770968683, 'learning_rate/qwen_vl_interface': 9.767518452401981e-06, 'learning_rate/action_model': 9.76751845240198e-05, 'epoch': 5.81}) +[2026-06-20 13:49:45,902][starVLA.training.train_starvla][INFO] - Step 3191, Loss: {'train/loss': 0.14364472031593323, 'train/grad_norm_pre_clip': 7.777548789978027, 'timing/data': 0.012254136963747442, 'timing/model': 1.1696541249984875, 'learning_rate/qwen_vl_interface': 9.766576436992802e-06, 'learning_rate/action_model': 9.766576436992801e-05, 'epoch': 5.81}) +[2026-06-20 13:49:51,191][starVLA.training.train_starvla][INFO] - Step 3192, Loss: {'train/loss': 0.19233867526054382, 'train/grad_norm_pre_clip': 7.740134239196777, 'timing/data': 0.012736525037325919, 'timing/model': 1.436286878073588, 'learning_rate/qwen_vl_interface': 9.765632564989657e-06, 'learning_rate/action_model': 9.765632564989656e-05, 'epoch': 5.81}) +[2026-06-20 13:49:56,366][starVLA.training.train_starvla][INFO] - Step 3193, Loss: {'train/loss': 0.10482753813266754, 'train/grad_norm_pre_clip': 4.347268104553223, 'timing/data': 0.015103782992810011, 'timing/model': 1.3266926440410316, 'learning_rate/qwen_vl_interface': 9.764686836780533e-06, 'learning_rate/action_model': 9.764686836780532e-05, 'epoch': 5.81}) +[2026-06-20 13:50:01,240][starVLA.training.train_starvla][INFO] - Step 3194, Loss: {'train/loss': 0.25706061720848083, 'train/grad_norm_pre_clip': 14.481354713439941, 'timing/data': 0.011830700095742941, 'timing/model': 1.1840279020834714, 'learning_rate/qwen_vl_interface': 9.763739252754187e-06, 'learning_rate/action_model': 9.763739252754186e-05, 'epoch': 5.81}) +[2026-06-20 13:50:06,071][starVLA.training.train_starvla][INFO] - Step 3195, Loss: {'train/loss': 0.19906175136566162, 'train/grad_norm_pre_clip': 8.719017028808594, 'timing/data': 0.01152247202116996, 'timing/model': 1.2102784180315211, 'learning_rate/qwen_vl_interface': 9.76278981330013e-06, 'learning_rate/action_model': 9.762789813300131e-05, 'epoch': 5.82}) +[2026-06-20 13:50:11,437][starVLA.training.train_starvla][INFO] - Step 3196, Loss: {'train/loss': 0.3634912967681885, 'train/grad_norm_pre_clip': 13.050212860107422, 'timing/data': 0.013371662003919482, 'timing/model': 1.2050386760383844, 'learning_rate/qwen_vl_interface': 9.761838518808647e-06, 'learning_rate/action_model': 9.761838518808646e-05, 'epoch': 5.82}) +[2026-06-20 13:50:16,296][starVLA.training.train_starvla][INFO] - Step 3197, Loss: {'train/loss': 0.14210771024227142, 'train/grad_norm_pre_clip': 8.306825637817383, 'timing/data': 0.012236117967404425, 'timing/model': 1.1940994589822367, 'learning_rate/qwen_vl_interface': 9.760885369670776e-06, 'learning_rate/action_model': 9.760885369670776e-05, 'epoch': 5.82}) +[2026-06-20 13:50:22,497][starVLA.training.train_starvla][INFO] - Step 3198, Loss: {'train/loss': 0.20970793068408966, 'train/grad_norm_pre_clip': 9.010666847229004, 'timing/data': 0.011645311024039984, 'timing/model': 1.933611816028133, 'learning_rate/qwen_vl_interface': 9.759930366278319e-06, 'learning_rate/action_model': 9.759930366278318e-05, 'epoch': 5.82}) +[2026-06-20 13:50:27,444][starVLA.training.train_starvla][INFO] - Step 3199, Loss: {'train/loss': 0.06672219187021255, 'train/grad_norm_pre_clip': 3.719175100326538, 'timing/data': 0.012126208981499076, 'timing/model': 1.357603101991117, 'learning_rate/qwen_vl_interface': 9.758973509023842e-06, 'learning_rate/action_model': 9.758973509023843e-05, 'epoch': 5.82}) +[2026-06-20 13:50:32,290][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 3200 +[2026-06-20 13:56:52,222][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 3200 +[2026-06-20 13:56:52,223][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 3200 +[2026-06-20 14:03:40,974][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 3200 +[2026-06-20 14:03:40,976][starVLA.training.train_starvla][INFO] - Step 3200, Loss: {'train/loss': 0.2517833411693573, 'train/grad_norm_pre_clip': 8.276689529418945, 'eval/loss': 0.4486616253852844, 'eval/latency_0/loss': 0.4610968828201294, 'eval/latency_1/loss': 0.6197479367256165, 'eval/latency_2/loss': 0.5347290635108948, 'eval/latency_3/loss': 0.11858775466680527, 'eval/latency_4/loss': 0.47328606247901917, 'eval/flappy/loss': 0.4486616253852844, 'eval/flappy/latency_0/loss': 0.4610968828201294, 'eval/flappy/latency_1/loss': 0.6197479367256165, 'eval/flappy/latency_2/loss': 0.5347290635108948, 'eval/flappy/latency_3/loss': 0.11858775466680527, 'eval/flappy/latency_4/loss': 0.47328606247901917, 'eval/flap_precision': 0.25263157894736843, 'eval/flap_recall': 0.08727272727272728, 'eval/flap_f1': 0.12972972972972974, 'eval/flap_rate_model': 0.02638888888888889, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.12972972972972974, 'eval/latency_0/flap_precision': 0.25263157894736843, 'eval/latency_0/flap_recall': 0.08727272727272728, 'eval/latency_0/flap_f1': 0.12972972972972974, 'eval/latency_0/flap_rate_model': 0.02638888888888889, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.12972972972972974, 'timing/data': 0.01743332203477621, 'timing/model': 1.1923999179853126, 'learning_rate/qwen_vl_interface': 9.758014798300677e-06, 'learning_rate/action_model': 9.758014798300676e-05, 'epoch': 5.83}) +[2026-06-20 14:03:40,977][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_3200_state +[2026-06-20 14:03:40,981][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-20 14:04:00,998][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_3200_state/model.safetensors +[2026-06-20 14:04:18,393][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_3200_state/optimizer.bin +[2026-06-20 14:04:18,393][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_3200_state/sampler.bin +[2026-06-20 14:04:18,393][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_3200_state/sampler_1.bin +[2026-06-20 14:04:18,395][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_3200_state/random_states_0.pkl +[2026-06-20 14:04:19,289][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3200 +[2026-06-20 14:04:19,290][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-20 14:04:19,303][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-20 14:04:24,548][starVLA.training.train_starvla][INFO] - Step 3201, Loss: {'train/loss': 0.11268799006938934, 'train/grad_norm_pre_clip': 7.027853488922119, 'timing/data': 0.014376027043908834, 'timing/model': 1.1490362560143694, 'learning_rate/qwen_vl_interface': 9.757054234502909e-06, 'learning_rate/action_model': 9.757054234502909e-05, 'epoch': 5.83}) +[2026-06-20 14:04:29,432][starVLA.training.train_starvla][INFO] - Step 3202, Loss: {'train/loss': 0.4185985028743744, 'train/grad_norm_pre_clip': 15.094035148620605, 'timing/data': 0.012483114027418196, 'timing/model': 1.5289915589382872, 'learning_rate/qwen_vl_interface': 9.756091818025394e-06, 'learning_rate/action_model': 9.756091818025394e-05, 'epoch': 5.83}) +[2026-06-20 14:04:34,209][starVLA.training.train_starvla][INFO] - Step 3203, Loss: {'train/loss': 0.48802927136421204, 'train/grad_norm_pre_clip': 15.306517601013184, 'timing/data': 0.011751117068342865, 'timing/model': 1.3833139849593863, 'learning_rate/qwen_vl_interface': 9.75512754926374e-06, 'learning_rate/action_model': 9.755127549263741e-05, 'epoch': 5.83}) +[2026-06-20 14:04:38,828][starVLA.training.train_starvla][INFO] - Step 3204, Loss: {'train/loss': 0.20156215131282806, 'train/grad_norm_pre_clip': 6.961760520935059, 'timing/data': 0.011864255997352302, 'timing/model': 1.1934320310829207, 'learning_rate/qwen_vl_interface': 9.754161428614325e-06, 'learning_rate/action_model': 9.754161428614324e-05, 'epoch': 5.83}) +[2026-06-20 14:04:44,349][starVLA.training.train_starvla][INFO] - Step 3205, Loss: {'train/loss': 0.05366571247577667, 'train/grad_norm_pre_clip': 3.221625566482544, 'timing/data': 0.01266068301629275, 'timing/model': 1.5548616679152474, 'learning_rate/qwen_vl_interface': 9.753193456474284e-06, 'learning_rate/action_model': 9.753193456474284e-05, 'epoch': 5.83}) +[2026-06-20 14:04:49,740][starVLA.training.train_starvla][INFO] - Step 3206, Loss: {'train/loss': 0.15325020253658295, 'train/grad_norm_pre_clip': 8.820396423339844, 'timing/data': 0.01694104797206819, 'timing/model': 1.186933042947203, 'learning_rate/qwen_vl_interface': 9.752223633241514e-06, 'learning_rate/action_model': 9.752223633241514e-05, 'epoch': 5.84}) +[2026-06-20 14:04:54,892][starVLA.training.train_starvla][INFO] - Step 3207, Loss: {'train/loss': 0.31908750534057617, 'train/grad_norm_pre_clip': 16.10158920288086, 'timing/data': 0.011306997970677912, 'timing/model': 1.1812692990060896, 'learning_rate/qwen_vl_interface': 9.751251959314671e-06, 'learning_rate/action_model': 9.751251959314672e-05, 'epoch': 5.84}) +[2026-06-20 14:05:00,285][starVLA.training.train_starvla][INFO] - Step 3208, Loss: {'train/loss': 0.1544746607542038, 'train/grad_norm_pre_clip': 6.387684345245361, 'timing/data': 0.019537477986887097, 'timing/model': 1.4850588319823146, 'learning_rate/qwen_vl_interface': 9.750278435093176e-06, 'learning_rate/action_model': 9.750278435093177e-05, 'epoch': 5.84}) +[2026-06-20 14:05:05,288][starVLA.training.train_starvla][INFO] - Step 3209, Loss: {'train/loss': 0.043150220066308975, 'train/grad_norm_pre_clip': 4.185895919799805, 'timing/data': 0.0147380760172382, 'timing/model': 1.4623506279895082, 'learning_rate/qwen_vl_interface': 9.749303060977205e-06, 'learning_rate/action_model': 9.749303060977205e-05, 'epoch': 5.84}) +[2026-06-20 14:05:09,975][starVLA.training.train_starvla][INFO] - Step 3210, Loss: {'train/loss': 0.11260344833135605, 'train/grad_norm_pre_clip': 8.737004280090332, 'timing/data': 0.011246409034356475, 'timing/model': 1.1736224460182711, 'learning_rate/qwen_vl_interface': 9.748325837367701e-06, 'learning_rate/action_model': 9.748325837367702e-05, 'epoch': 5.84}) +[2026-06-20 14:05:15,583][starVLA.training.train_starvla][INFO] - Step 3211, Loss: {'train/loss': 0.10821807384490967, 'train/grad_norm_pre_clip': 10.09687328338623, 'timing/data': 0.02647298201918602, 'timing/model': 1.3003089539706707, 'learning_rate/qwen_vl_interface': 9.747346764666362e-06, 'learning_rate/action_model': 9.747346764666361e-05, 'epoch': 5.85}) +[2026-06-20 14:05:20,852][starVLA.training.train_starvla][INFO] - Step 3212, Loss: {'train/loss': 0.06203000247478485, 'train/grad_norm_pre_clip': 4.164025783538818, 'timing/data': 0.012794623035006225, 'timing/model': 1.4081971700070426, 'learning_rate/qwen_vl_interface': 9.746365843275648e-06, 'learning_rate/action_model': 9.746365843275646e-05, 'epoch': 5.85}) +[2026-06-20 14:05:25,821][starVLA.training.train_starvla][INFO] - Step 3213, Loss: {'train/loss': 0.2932603061199188, 'train/grad_norm_pre_clip': 22.31366539001465, 'timing/data': 0.013622700003907084, 'timing/model': 1.2406810800312087, 'learning_rate/qwen_vl_interface': 9.745383073598776e-06, 'learning_rate/action_model': 9.745383073598775e-05, 'epoch': 5.85}) +[2026-06-20 14:05:31,208][starVLA.training.train_starvla][INFO] - Step 3214, Loss: {'train/loss': 0.05128594487905502, 'train/grad_norm_pre_clip': 4.5319952964782715, 'timing/data': 0.011621527024544775, 'timing/model': 1.7139098659390584, 'learning_rate/qwen_vl_interface': 9.744398456039729e-06, 'learning_rate/action_model': 9.744398456039729e-05, 'epoch': 5.85}) +[2026-06-20 14:05:36,040][starVLA.training.train_starvla][INFO] - Step 3215, Loss: {'train/loss': 0.08501284569501877, 'train/grad_norm_pre_clip': 6.708944797515869, 'timing/data': 0.01137862796895206, 'timing/model': 1.2665031820069999, 'learning_rate/qwen_vl_interface': 9.743411991003246e-06, 'learning_rate/action_model': 9.743411991003247e-05, 'epoch': 5.85}) +[2026-06-20 14:05:41,093][starVLA.training.train_starvla][INFO] - Step 3216, Loss: {'train/loss': 0.3442423939704895, 'train/grad_norm_pre_clip': 17.02141761779785, 'timing/data': 0.011455868021585047, 'timing/model': 1.1895209680078551, 'learning_rate/qwen_vl_interface': 9.742423678894823e-06, 'learning_rate/action_model': 9.742423678894823e-05, 'epoch': 5.85}) +[2026-06-20 14:05:46,167][starVLA.training.train_starvla][INFO] - Step 3217, Loss: {'train/loss': 0.2038590908050537, 'train/grad_norm_pre_clip': 14.61421012878418, 'timing/data': 0.010851160972379148, 'timing/model': 1.1754003920359537, 'learning_rate/qwen_vl_interface': 9.741433520120722e-06, 'learning_rate/action_model': 9.74143352012072e-05, 'epoch': 5.86}) +[2026-06-20 14:05:51,208][starVLA.training.train_starvla][INFO] - Step 3218, Loss: {'train/loss': 0.1384476274251938, 'train/grad_norm_pre_clip': 19.153980255126953, 'timing/data': 0.011478229076601565, 'timing/model': 1.195721619995311, 'learning_rate/qwen_vl_interface': 9.740441515087953e-06, 'learning_rate/action_model': 9.740441515087953e-05, 'epoch': 5.86}) +[2026-06-20 14:05:56,158][starVLA.training.train_starvla][INFO] - Step 3219, Loss: {'train/loss': 0.11059294641017914, 'train/grad_norm_pre_clip': 9.349197387695312, 'timing/data': 0.01357462105806917, 'timing/model': 1.2467975190375, 'learning_rate/qwen_vl_interface': 9.739447664204298e-06, 'learning_rate/action_model': 9.739447664204297e-05, 'epoch': 5.86}) +[2026-06-20 14:06:01,147][starVLA.training.train_starvla][INFO] - Step 3220, Loss: {'train/loss': 0.11954491585493088, 'train/grad_norm_pre_clip': 11.476821899414062, 'timing/data': 0.011573317926377058, 'timing/model': 1.3870886079967022, 'learning_rate/qwen_vl_interface': 9.73845196787829e-06, 'learning_rate/action_model': 9.738451967878289e-05, 'epoch': 5.86}) +[2026-06-20 14:06:06,004][starVLA.training.train_starvla][INFO] - Step 3221, Loss: {'train/loss': 0.1692914366722107, 'train/grad_norm_pre_clip': 14.810797691345215, 'timing/data': 0.013332750997506082, 'timing/model': 1.2423424200387672, 'learning_rate/qwen_vl_interface': 9.737454426519221e-06, 'learning_rate/action_model': 9.737454426519221e-05, 'epoch': 5.86}) +[2026-06-20 14:06:10,983][starVLA.training.train_starvla][INFO] - Step 3222, Loss: {'train/loss': 0.13431768119335175, 'train/grad_norm_pre_clip': 8.227206230163574, 'timing/data': 0.01169925695285201, 'timing/model': 1.2946286019869149, 'learning_rate/qwen_vl_interface': 9.736455040537145e-06, 'learning_rate/action_model': 9.736455040537144e-05, 'epoch': 5.87}) +[2026-06-20 14:06:15,804][starVLA.training.train_starvla][INFO] - Step 3223, Loss: {'train/loss': 0.2722458839416504, 'train/grad_norm_pre_clip': 26.030414581298828, 'timing/data': 0.011721207993105054, 'timing/model': 1.1798316539498046, 'learning_rate/qwen_vl_interface': 9.735453810342868e-06, 'learning_rate/action_model': 9.735453810342867e-05, 'epoch': 5.87}) +[2026-06-20 14:06:20,853][starVLA.training.train_starvla][INFO] - Step 3224, Loss: {'train/loss': 0.08094362169504166, 'train/grad_norm_pre_clip': 9.85335636138916, 'timing/data': 0.011758996988646686, 'timing/model': 1.264658278087154, 'learning_rate/qwen_vl_interface': 9.734450736347958e-06, 'learning_rate/action_model': 9.73445073634796e-05, 'epoch': 5.87}) +[2026-06-20 14:06:25,934][starVLA.training.train_starvla][INFO] - Step 3225, Loss: {'train/loss': 0.10861551761627197, 'train/grad_norm_pre_clip': 10.6046724319458, 'timing/data': 0.013067702995613217, 'timing/model': 1.2806205820525065, 'learning_rate/qwen_vl_interface': 9.733445818964746e-06, 'learning_rate/action_model': 9.733445818964746e-05, 'epoch': 5.87}) +[2026-06-20 14:06:30,985][starVLA.training.train_starvla][INFO] - Step 3226, Loss: {'train/loss': 0.16013363003730774, 'train/grad_norm_pre_clip': 17.30336570739746, 'timing/data': 0.012488135020248592, 'timing/model': 1.507757477927953, 'learning_rate/qwen_vl_interface': 9.732439058606311e-06, 'learning_rate/action_model': 9.732439058606312e-05, 'epoch': 5.87}) +[2026-06-20 14:06:35,804][starVLA.training.train_starvla][INFO] - Step 3227, Loss: {'train/loss': 0.21656958758831024, 'train/grad_norm_pre_clip': 11.840788841247559, 'timing/data': 0.011566587956622243, 'timing/model': 1.3099989020265639, 'learning_rate/qwen_vl_interface': 9.731430455686499e-06, 'learning_rate/action_model': 9.731430455686498e-05, 'epoch': 5.87}) +[2026-06-20 14:06:40,815][starVLA.training.train_starvla][INFO] - Step 3228, Loss: {'train/loss': 0.13459859788417816, 'train/grad_norm_pre_clip': 10.805737495422363, 'timing/data': 0.011417587986215949, 'timing/model': 1.2459044790593907, 'learning_rate/qwen_vl_interface': 9.730420010619901e-06, 'learning_rate/action_model': 9.7304200106199e-05, 'epoch': 5.88}) +[2026-06-20 14:06:45,654][starVLA.training.train_starvla][INFO] - Step 3229, Loss: {'train/loss': 0.22718608379364014, 'train/grad_norm_pre_clip': 12.797584533691406, 'timing/data': 0.011495167971588671, 'timing/model': 1.1853155320277438, 'learning_rate/qwen_vl_interface': 9.72940772382188e-06, 'learning_rate/action_model': 9.729407723821878e-05, 'epoch': 5.88}) +[2026-06-20 14:06:50,636][starVLA.training.train_starvla][INFO] - Step 3230, Loss: {'train/loss': 0.02255832776427269, 'train/grad_norm_pre_clip': 2.9361636638641357, 'timing/data': 0.011423029005527496, 'timing/model': 1.1886547040194273, 'learning_rate/qwen_vl_interface': 9.728393595708542e-06, 'learning_rate/action_model': 9.728393595708542e-05, 'epoch': 5.88}) +[2026-06-20 14:06:55,708][starVLA.training.train_starvla][INFO] - Step 3231, Loss: {'train/loss': 0.0322241373360157, 'train/grad_norm_pre_clip': 7.7934112548828125, 'timing/data': 0.011242989916354418, 'timing/model': 1.5699277119711041, 'learning_rate/qwen_vl_interface': 9.727377626696762e-06, 'learning_rate/action_model': 9.727377626696762e-05, 'epoch': 5.88}) +[2026-06-20 14:07:00,348][starVLA.training.train_starvla][INFO] - Step 3232, Loss: {'train/loss': 0.04869227856397629, 'train/grad_norm_pre_clip': 4.855114459991455, 'timing/data': 0.01078879099804908, 'timing/model': 1.19635989703238, 'learning_rate/qwen_vl_interface': 9.726359817204165e-06, 'learning_rate/action_model': 9.726359817204166e-05, 'epoch': 5.88}) +[2026-06-20 14:07:05,533][starVLA.training.train_starvla][INFO] - Step 3233, Loss: {'train/loss': 0.1371813863515854, 'train/grad_norm_pre_clip': 8.796388626098633, 'timing/data': 0.01166328799445182, 'timing/model': 1.268873549066484, 'learning_rate/qwen_vl_interface': 9.725340167649136e-06, 'learning_rate/action_model': 9.725340167649136e-05, 'epoch': 5.89}) +[2026-06-20 14:07:10,829][starVLA.training.train_starvla][INFO] - Step 3234, Loss: {'train/loss': 0.12164420634508133, 'train/grad_norm_pre_clip': 14.89236831665039, 'timing/data': 0.011331529007293284, 'timing/model': 1.1819650980178267, 'learning_rate/qwen_vl_interface': 9.72431867845081e-06, 'learning_rate/action_model': 9.72431867845081e-05, 'epoch': 5.89}) +[2026-06-20 14:07:15,784][starVLA.training.train_starvla][INFO] - Step 3235, Loss: {'train/loss': 0.18825678527355194, 'train/grad_norm_pre_clip': 10.670035362243652, 'timing/data': 0.012703413958661258, 'timing/model': 1.1837488829623908, 'learning_rate/qwen_vl_interface': 9.723295350029086e-06, 'learning_rate/action_model': 9.723295350029086e-05, 'epoch': 5.89}) +[2026-06-20 14:07:20,692][starVLA.training.train_starvla][INFO] - Step 3236, Loss: {'train/loss': 0.16232441365718842, 'train/grad_norm_pre_clip': 15.760770797729492, 'timing/data': 0.011444818926975131, 'timing/model': 1.1834776770556346, 'learning_rate/qwen_vl_interface': 9.722270182804616e-06, 'learning_rate/action_model': 9.722270182804614e-05, 'epoch': 5.89}) +[2026-06-20 14:07:25,620][starVLA.training.train_starvla][INFO] - Step 3237, Loss: {'train/loss': 0.09821365028619766, 'train/grad_norm_pre_clip': 11.011554718017578, 'timing/data': 0.011043010046705604, 'timing/model': 1.4267803579568863, 'learning_rate/qwen_vl_interface': 9.721243177198805e-06, 'learning_rate/action_model': 9.721243177198804e-05, 'epoch': 5.89}) +[2026-06-20 14:07:30,499][starVLA.training.train_starvla][INFO] - Step 3238, Loss: {'train/loss': 0.07156156003475189, 'train/grad_norm_pre_clip': 7.372698783874512, 'timing/data': 0.014290558057837188, 'timing/model': 1.274112863931805, 'learning_rate/qwen_vl_interface': 9.720214333633817e-06, 'learning_rate/action_model': 9.720214333633817e-05, 'epoch': 5.89}) +[2026-06-20 14:07:35,531][starVLA.training.train_starvla][INFO] - Step 3239, Loss: {'train/loss': 0.06513093411922455, 'train/grad_norm_pre_clip': 14.652656555175781, 'timing/data': 0.0060414280742406845, 'timing/model': 1.1760291720274836, 'learning_rate/qwen_vl_interface': 9.719183652532572e-06, 'learning_rate/action_model': 9.719183652532572e-05, 'epoch': 5.9}) +[2026-06-20 14:07:40,696][starVLA.training.train_starvla][INFO] - Step 3240, Loss: {'train/loss': 0.14905664324760437, 'train/grad_norm_pre_clip': 12.789680480957031, 'timing/data': 0.007942700991407037, 'timing/model': 1.2790744820376858, 'learning_rate/qwen_vl_interface': 9.718151134318743e-06, 'learning_rate/action_model': 9.718151134318743e-05, 'epoch': 5.9}) +[2026-06-20 14:07:45,747][starVLA.training.train_starvla][INFO] - Step 3241, Loss: {'train/loss': 0.04673564061522484, 'train/grad_norm_pre_clip': 6.697552680969238, 'timing/data': 0.006155267939902842, 'timing/model': 1.223428165889345, 'learning_rate/qwen_vl_interface': 9.71711677941676e-06, 'learning_rate/action_model': 9.717116779416758e-05, 'epoch': 5.9}) +[2026-06-20 14:07:50,572][starVLA.training.train_starvla][INFO] - Step 3242, Loss: {'train/loss': 0.29218271374702454, 'train/grad_norm_pre_clip': 13.330758094787598, 'timing/data': 0.018138063955120742, 'timing/model': 1.210273084929213, 'learning_rate/qwen_vl_interface': 9.716080588251805e-06, 'learning_rate/action_model': 9.716080588251805e-05, 'epoch': 5.9}) +[2026-06-20 14:07:55,517][starVLA.training.train_starvla][INFO] - Step 3243, Loss: {'train/loss': 0.46631717681884766, 'train/grad_norm_pre_clip': 17.123689651489258, 'timing/data': 0.006205827929079533, 'timing/model': 1.4105805340223014, 'learning_rate/qwen_vl_interface': 9.71504256124982e-06, 'learning_rate/action_model': 9.71504256124982e-05, 'epoch': 5.9}) +[2026-06-20 14:08:00,799][starVLA.training.train_starvla][INFO] - Step 3244, Loss: {'train/loss': 0.23985245823860168, 'train/grad_norm_pre_clip': 17.549421310424805, 'timing/data': 0.023797984002158046, 'timing/model': 1.7607768239686266, 'learning_rate/qwen_vl_interface': 9.714002698837499e-06, 'learning_rate/action_model': 9.714002698837498e-05, 'epoch': 5.91}) +[2026-06-20 14:08:06,146][starVLA.training.train_starvla][INFO] - Step 3245, Loss: {'train/loss': 0.21753084659576416, 'train/grad_norm_pre_clip': 17.375490188598633, 'timing/data': 0.01721154898405075, 'timing/model': 1.29580756300129, 'learning_rate/qwen_vl_interface': 9.712961001442285e-06, 'learning_rate/action_model': 9.712961001442286e-05, 'epoch': 5.91}) +[2026-06-20 14:08:11,178][starVLA.training.train_starvla][INFO] - Step 3246, Loss: {'train/loss': 0.3365713953971863, 'train/grad_norm_pre_clip': 13.407150268554688, 'timing/data': 0.01712061802390963, 'timing/model': 1.214036121033132, 'learning_rate/qwen_vl_interface': 9.711917469492388e-06, 'learning_rate/action_model': 9.711917469492388e-05, 'epoch': 5.91}) +[2026-06-20 14:08:16,344][starVLA.training.train_starvla][INFO] - Step 3247, Loss: {'train/loss': 0.21717245876789093, 'train/grad_norm_pre_clip': 20.004314422607422, 'timing/data': 0.021250772988423705, 'timing/model': 1.282475857064128, 'learning_rate/qwen_vl_interface': 9.71087210341676e-06, 'learning_rate/action_model': 9.71087210341676e-05, 'epoch': 5.91}) +[2026-06-20 14:08:21,627][starVLA.training.train_starvla][INFO] - Step 3248, Loss: {'train/loss': 0.11442078649997711, 'train/grad_norm_pre_clip': 9.754549980163574, 'timing/data': 0.018144604982808232, 'timing/model': 1.4383083500433713, 'learning_rate/qwen_vl_interface': 9.70982490364511e-06, 'learning_rate/action_model': 9.70982490364511e-05, 'epoch': 5.91}) +[2026-06-20 14:08:26,682][starVLA.training.train_starvla][INFO] - Step 3249, Loss: {'train/loss': 0.134622260928154, 'train/grad_norm_pre_clip': 11.83079719543457, 'timing/data': 0.02397603401914239, 'timing/model': 1.5201596460537985, 'learning_rate/qwen_vl_interface': 9.708775870607905e-06, 'learning_rate/action_model': 9.708775870607904e-05, 'epoch': 5.91}) +[2026-06-20 14:08:31,760][starVLA.training.train_starvla][INFO] - Step 3250, Loss: {'train/loss': 0.1141790822148323, 'train/grad_norm_pre_clip': 6.353628635406494, 'timing/data': 0.0174639179604128, 'timing/model': 1.276935244910419, 'learning_rate/qwen_vl_interface': 9.707725004736363e-06, 'learning_rate/action_model': 9.707725004736363e-05, 'epoch': 5.92}) +[2026-06-20 14:08:36,939][starVLA.training.train_starvla][INFO] - Step 3251, Loss: {'train/loss': 0.33985722064971924, 'train/grad_norm_pre_clip': 20.312061309814453, 'timing/data': 0.015995912021026015, 'timing/model': 1.2780710140941665, 'learning_rate/qwen_vl_interface': 9.706672306462454e-06, 'learning_rate/action_model': 9.706672306462454e-05, 'epoch': 5.92}) +[2026-06-20 14:08:41,989][starVLA.training.train_starvla][INFO] - Step 3252, Loss: {'train/loss': 0.3129095435142517, 'train/grad_norm_pre_clip': 9.309109687805176, 'timing/data': 0.024205232970416546, 'timing/model': 1.2150123619940132, 'learning_rate/qwen_vl_interface': 9.705617776218905e-06, 'learning_rate/action_model': 9.705617776218904e-05, 'epoch': 5.92}) +[2026-06-20 14:08:46,964][starVLA.training.train_starvla][INFO] - Step 3253, Loss: {'train/loss': 0.05934209004044533, 'train/grad_norm_pre_clip': 3.529608726501465, 'timing/data': 0.012670625001192093, 'timing/model': 1.2660912290448323, 'learning_rate/qwen_vl_interface': 9.70456141443919e-06, 'learning_rate/action_model': 9.704561414439188e-05, 'epoch': 5.92}) +[2026-06-20 14:08:51,940][starVLA.training.train_starvla][INFO] - Step 3254, Loss: {'train/loss': 0.0966743677854538, 'train/grad_norm_pre_clip': 5.507063388824463, 'timing/data': 0.011738377972505987, 'timing/model': 1.1951128079090267, 'learning_rate/qwen_vl_interface': 9.703503221557538e-06, 'learning_rate/action_model': 9.703503221557538e-05, 'epoch': 5.92}) +[2026-06-20 14:08:57,084][starVLA.training.train_starvla][INFO] - Step 3255, Loss: {'train/loss': 0.13871586322784424, 'train/grad_norm_pre_clip': 5.41738224029541, 'timing/data': 0.01122356008272618, 'timing/model': 1.4875717120012268, 'learning_rate/qwen_vl_interface': 9.70244319800894e-06, 'learning_rate/action_model': 9.70244319800894e-05, 'epoch': 5.93}) +[2026-06-20 14:09:01,864][starVLA.training.train_starvla][INFO] - Step 3256, Loss: {'train/loss': 0.04265923053026199, 'train/grad_norm_pre_clip': 2.7679195404052734, 'timing/data': 0.011433467967435718, 'timing/model': 1.1904056689236313, 'learning_rate/qwen_vl_interface': 9.701381344229124e-06, 'learning_rate/action_model': 9.701381344229123e-05, 'epoch': 5.93}) +[2026-06-20 14:09:07,004][starVLA.training.train_starvla][INFO] - Step 3257, Loss: {'train/loss': 0.24107816815376282, 'train/grad_norm_pre_clip': 11.373896598815918, 'timing/data': 0.011678498005494475, 'timing/model': 1.1831528359325603, 'learning_rate/qwen_vl_interface': 9.700317660654581e-06, 'learning_rate/action_model': 9.700317660654581e-05, 'epoch': 5.93}) +[2026-06-20 14:09:11,996][starVLA.training.train_starvla][INFO] - Step 3258, Loss: {'train/loss': 0.14189693331718445, 'train/grad_norm_pre_clip': 6.787549018859863, 'timing/data': 0.01181298797018826, 'timing/model': 1.1790445040678605, 'learning_rate/qwen_vl_interface': 9.699252147722554e-06, 'learning_rate/action_model': 9.699252147722553e-05, 'epoch': 5.93}) +[2026-06-20 14:09:16,980][starVLA.training.train_starvla][INFO] - Step 3259, Loss: {'train/loss': 0.04011260345578194, 'train/grad_norm_pre_clip': 3.3867595195770264, 'timing/data': 0.012062667054124177, 'timing/model': 1.2338963099755347, 'learning_rate/qwen_vl_interface': 9.698184805871028e-06, 'learning_rate/action_model': 9.698184805871028e-05, 'epoch': 5.93}) +[2026-06-20 14:09:22,044][starVLA.training.train_starvla][INFO] - Step 3260, Loss: {'train/loss': 0.033587746322155, 'train/grad_norm_pre_clip': 2.1287343502044678, 'timing/data': 0.011694867978803813, 'timing/model': 1.2486270299414173, 'learning_rate/qwen_vl_interface': 9.697115635538753e-06, 'learning_rate/action_model': 9.697115635538753e-05, 'epoch': 5.93}) +[2026-06-20 14:09:27,013][starVLA.training.train_starvla][INFO] - Step 3261, Loss: {'train/loss': 0.12833839654922485, 'train/grad_norm_pre_clip': 7.197160720825195, 'timing/data': 0.011729717953130603, 'timing/model': 1.4385344641050324, 'learning_rate/qwen_vl_interface': 9.69604463716522e-06, 'learning_rate/action_model': 9.69604463716522e-05, 'epoch': 5.94}) +[2026-06-20 14:09:31,692][starVLA.training.train_starvla][INFO] - Step 3262, Loss: {'train/loss': 0.2622123658657074, 'train/grad_norm_pre_clip': 7.8014960289001465, 'timing/data': 0.011562488973140717, 'timing/model': 1.1807721849763766, 'learning_rate/qwen_vl_interface': 9.69497181119068e-06, 'learning_rate/action_model': 9.694971811190681e-05, 'epoch': 5.94}) +[2026-06-20 14:09:36,747][starVLA.training.train_starvla][INFO] - Step 3263, Loss: {'train/loss': 0.06745021045207977, 'train/grad_norm_pre_clip': 5.4567766189575195, 'timing/data': 0.015880684019066393, 'timing/model': 1.2402887949720025, 'learning_rate/qwen_vl_interface': 9.693897158056132e-06, 'learning_rate/action_model': 9.693897158056132e-05, 'epoch': 5.94}) +[2026-06-20 14:09:41,752][starVLA.training.train_starvla][INFO] - Step 3264, Loss: {'train/loss': 0.21162380278110504, 'train/grad_norm_pre_clip': 11.342657089233398, 'timing/data': 0.01683914102613926, 'timing/model': 1.2583047120133415, 'learning_rate/qwen_vl_interface': 9.692820678203322e-06, 'learning_rate/action_model': 9.692820678203322e-05, 'epoch': 5.94}) +[2026-06-20 14:09:46,797][starVLA.training.train_starvla][INFO] - Step 3265, Loss: {'train/loss': 0.09534203261137009, 'train/grad_norm_pre_clip': 7.809260368347168, 'timing/data': 0.011840746970847249, 'timing/model': 1.1839034910080954, 'learning_rate/qwen_vl_interface': 9.691742372074749e-06, 'learning_rate/action_model': 9.691742372074749e-05, 'epoch': 5.94}) +[2026-06-20 14:09:51,847][starVLA.training.train_starvla][INFO] - Step 3266, Loss: {'train/loss': 0.02679455652832985, 'train/grad_norm_pre_clip': 2.3781816959381104, 'timing/data': 0.01233184605371207, 'timing/model': 1.3076982910279185, 'learning_rate/qwen_vl_interface': 9.690662240113669e-06, 'learning_rate/action_model': 9.690662240113668e-05, 'epoch': 5.95}) +[2026-06-20 14:09:57,008][starVLA.training.train_starvla][INFO] - Step 3267, Loss: {'train/loss': 0.13878943026065826, 'train/grad_norm_pre_clip': 14.186551094055176, 'timing/data': 0.017964276019483805, 'timing/model': 1.478435204946436, 'learning_rate/qwen_vl_interface': 9.689580282764077e-06, 'learning_rate/action_model': 9.689580282764077e-05, 'epoch': 5.95}) +[2026-06-20 14:10:01,832][starVLA.training.train_starvla][INFO] - Step 3268, Loss: {'train/loss': 0.15497712790966034, 'train/grad_norm_pre_clip': 16.494495391845703, 'timing/data': 0.013524541980586946, 'timing/model': 1.3834770950488746, 'learning_rate/qwen_vl_interface': 9.688496500470731e-06, 'learning_rate/action_model': 9.68849650047073e-05, 'epoch': 5.95}) +[2026-06-20 14:10:06,550][starVLA.training.train_starvla][INFO] - Step 3269, Loss: {'train/loss': 0.29936543107032776, 'train/grad_norm_pre_clip': 46.21840286254883, 'timing/data': 0.011780656990595162, 'timing/model': 1.2191710119368508, 'learning_rate/qwen_vl_interface': 9.68741089367913e-06, 'learning_rate/action_model': 9.687410893679129e-05, 'epoch': 5.95}) +[2026-06-20 14:10:11,720][starVLA.training.train_starvla][INFO] - Step 3270, Loss: {'train/loss': 0.20703385770320892, 'train/grad_norm_pre_clip': 19.8165340423584, 'timing/data': 0.021986082079820335, 'timing/model': 1.2222773829707876, 'learning_rate/qwen_vl_interface': 9.686323462835527e-06, 'learning_rate/action_model': 9.686323462835526e-05, 'epoch': 5.95}) +[2026-06-20 14:10:17,038][starVLA.training.train_starvla][INFO] - Step 3271, Loss: {'train/loss': 0.3215935230255127, 'train/grad_norm_pre_clip': 13.869832992553711, 'timing/data': 0.010869030957110226, 'timing/model': 1.1844427499454468, 'learning_rate/qwen_vl_interface': 9.685234208386922e-06, 'learning_rate/action_model': 9.685234208386921e-05, 'epoch': 5.95}) +[2026-06-20 14:10:21,964][starVLA.training.train_starvla][INFO] - Step 3272, Loss: {'train/loss': 0.25359079241752625, 'train/grad_norm_pre_clip': 15.96039867401123, 'timing/data': 0.011231438955292106, 'timing/model': 1.2740233140066266, 'learning_rate/qwen_vl_interface': 9.684143130781066e-06, 'learning_rate/action_model': 9.684143130781066e-05, 'epoch': 5.96}) +[2026-06-20 14:10:27,556][starVLA.training.train_starvla][INFO] - Step 3273, Loss: {'train/loss': 0.43209901452064514, 'train/grad_norm_pre_clip': 29.405986785888672, 'timing/data': 0.01143331010825932, 'timing/model': 1.43388642789796, 'learning_rate/qwen_vl_interface': 9.683050230466462e-06, 'learning_rate/action_model': 9.683050230466462e-05, 'epoch': 5.96}) +[2026-06-20 14:10:32,406][starVLA.training.train_starvla][INFO] - Step 3274, Loss: {'train/loss': 0.3895108997821808, 'train/grad_norm_pre_clip': 18.242639541625977, 'timing/data': 0.01167862897273153, 'timing/model': 1.4405871060444042, 'learning_rate/qwen_vl_interface': 9.681955507892361e-06, 'learning_rate/action_model': 9.68195550789236e-05, 'epoch': 5.96}) +[2026-06-20 14:10:37,528][starVLA.training.train_starvla][INFO] - Step 3275, Loss: {'train/loss': 0.11122709512710571, 'train/grad_norm_pre_clip': 11.091216087341309, 'timing/data': 0.012519685085862875, 'timing/model': 1.2339035029290244, 'learning_rate/qwen_vl_interface': 9.680858963508759e-06, 'learning_rate/action_model': 9.68085896350876e-05, 'epoch': 5.96}) +[2026-06-20 14:10:42,595][starVLA.training.train_starvla][INFO] - Step 3276, Loss: {'train/loss': 0.08677268773317337, 'train/grad_norm_pre_clip': 8.327300071716309, 'timing/data': 0.01405191095545888, 'timing/model': 1.2282371029723436, 'learning_rate/qwen_vl_interface': 9.679760597766405e-06, 'learning_rate/action_model': 9.679760597766405e-05, 'epoch': 5.96}) +[2026-06-20 14:10:48,260][starVLA.training.train_starvla][INFO] - Step 3277, Loss: {'train/loss': 0.1390577256679535, 'train/grad_norm_pre_clip': 8.784123420715332, 'timing/data': 0.011658349074423313, 'timing/model': 1.2619320660596713, 'learning_rate/qwen_vl_interface': 9.678660411116796e-06, 'learning_rate/action_model': 9.678660411116796e-05, 'epoch': 5.97}) +[2026-06-20 14:10:53,505][starVLA.training.train_starvla][INFO] - Step 3278, Loss: {'train/loss': 0.12737439572811127, 'train/grad_norm_pre_clip': 13.330289840698242, 'timing/data': 0.011886567925103009, 'timing/model': 1.259041148936376, 'learning_rate/qwen_vl_interface': 9.677558404012178e-06, 'learning_rate/action_model': 9.677558404012177e-05, 'epoch': 5.97}) +[2026-06-20 14:10:58,825][starVLA.training.train_starvla][INFO] - Step 3279, Loss: {'train/loss': 0.18390412628650665, 'train/grad_norm_pre_clip': 6.499790668487549, 'timing/data': 0.012526044971309602, 'timing/model': 1.4713104480179027, 'learning_rate/qwen_vl_interface': 9.676454576905542e-06, 'learning_rate/action_model': 9.676454576905542e-05, 'epoch': 5.97}) +[2026-06-20 14:11:03,628][starVLA.training.train_starvla][INFO] - Step 3280, Loss: {'train/loss': 0.10427875071763992, 'train/grad_norm_pre_clip': 7.101532459259033, 'timing/data': 0.011101901065558195, 'timing/model': 1.1807436590315774, 'learning_rate/qwen_vl_interface': 9.675348930250633e-06, 'learning_rate/action_model': 9.675348930250632e-05, 'epoch': 5.97}) +[2026-06-20 14:11:08,663][starVLA.training.train_starvla][INFO] - Step 3281, Loss: {'train/loss': 0.20047736167907715, 'train/grad_norm_pre_clip': 8.353198051452637, 'timing/data': 0.011390899075195193, 'timing/model': 1.28516955091618, 'learning_rate/qwen_vl_interface': 9.674241464501938e-06, 'learning_rate/action_model': 9.674241464501937e-05, 'epoch': 5.97}) +[2026-06-20 14:11:13,724][starVLA.training.train_starvla][INFO] - Step 3282, Loss: {'train/loss': 0.12511275708675385, 'train/grad_norm_pre_clip': 4.201874256134033, 'timing/data': 0.011270519928075373, 'timing/model': 1.3104138039052486, 'learning_rate/qwen_vl_interface': 9.673132180114695e-06, 'learning_rate/action_model': 9.673132180114695e-05, 'epoch': 5.97}) +[2026-06-20 14:11:18,895][starVLA.training.train_starvla][INFO] - Step 3283, Loss: {'train/loss': 0.27197298407554626, 'train/grad_norm_pre_clip': 9.818015098571777, 'timing/data': 0.01132917997892946, 'timing/model': 1.2858981620520353, 'learning_rate/qwen_vl_interface': 9.67202107754489e-06, 'learning_rate/action_model': 9.672021077544889e-05, 'epoch': 5.98}) +[2026-06-20 14:11:24,044][starVLA.training.train_starvla][INFO] - Step 3284, Loss: {'train/loss': 0.13454459607601166, 'train/grad_norm_pre_clip': 7.814175128936768, 'timing/data': 0.01336464297492057, 'timing/model': 1.2358271901030093, 'learning_rate/qwen_vl_interface': 9.670908157249254e-06, 'learning_rate/action_model': 9.670908157249253e-05, 'epoch': 5.98}) +[2026-06-20 14:11:28,934][starVLA.training.train_starvla][INFO] - Step 3285, Loss: {'train/loss': 0.10843672603368759, 'train/grad_norm_pre_clip': 7.485348224639893, 'timing/data': 0.01729316997807473, 'timing/model': 1.4393829619511962, 'learning_rate/qwen_vl_interface': 9.669793419685267e-06, 'learning_rate/action_model': 9.669793419685266e-05, 'epoch': 5.98}) +[2026-06-20 14:11:33,614][starVLA.training.train_starvla][INFO] - Step 3286, Loss: {'train/loss': 0.09093528985977173, 'train/grad_norm_pre_clip': 4.4504170417785645, 'timing/data': 0.01128369104117155, 'timing/model': 1.1769864519592375, 'learning_rate/qwen_vl_interface': 9.668676865311155e-06, 'learning_rate/action_model': 9.668676865311154e-05, 'epoch': 5.98}) +[2026-06-20 14:11:38,644][starVLA.training.train_starvla][INFO] - Step 3287, Loss: {'train/loss': 0.06467979401350021, 'train/grad_norm_pre_clip': 3.8676984310150146, 'timing/data': 0.012053597951307893, 'timing/model': 1.18890749104321, 'learning_rate/qwen_vl_interface': 9.667558494585891e-06, 'learning_rate/action_model': 9.66755849458589e-05, 'epoch': 5.98}) +[2026-06-20 14:11:43,556][starVLA.training.train_starvla][INFO] - Step 3288, Loss: {'train/loss': 0.14166930317878723, 'train/grad_norm_pre_clip': 7.800753593444824, 'timing/data': 0.013297043973580003, 'timing/model': 1.2224990649847314, 'learning_rate/qwen_vl_interface': 9.666438307969194e-06, 'learning_rate/action_model': 9.666438307969195e-05, 'epoch': 5.99}) +[2026-06-20 14:11:48,574][starVLA.training.train_starvla][INFO] - Step 3289, Loss: {'train/loss': 0.233231782913208, 'train/grad_norm_pre_clip': 9.018108367919922, 'timing/data': 0.012399145984090865, 'timing/model': 1.207518678973429, 'learning_rate/qwen_vl_interface': 9.665316305921535e-06, 'learning_rate/action_model': 9.665316305921535e-05, 'epoch': 5.99}) +[2026-06-20 14:11:53,505][starVLA.training.train_starvla][INFO] - Step 3290, Loss: {'train/loss': 0.5142180919647217, 'train/grad_norm_pre_clip': 19.45972442626953, 'timing/data': 0.011708578909747303, 'timing/model': 1.203711484093219, 'learning_rate/qwen_vl_interface': 9.664192488904123e-06, 'learning_rate/action_model': 9.664192488904122e-05, 'epoch': 5.99}) +[2026-06-20 14:11:58,387][starVLA.training.train_starvla][INFO] - Step 3291, Loss: {'train/loss': 0.0921262726187706, 'train/grad_norm_pre_clip': 9.59306526184082, 'timing/data': 0.013733222032897174, 'timing/model': 1.4588148270267993, 'learning_rate/qwen_vl_interface': 9.663066857378917e-06, 'learning_rate/action_model': 9.663066857378917e-05, 'epoch': 5.99}) +[2026-06-20 14:12:03,186][starVLA.training.train_starvla][INFO] - Step 3292, Loss: {'train/loss': 0.3495364487171173, 'train/grad_norm_pre_clip': 9.117620468139648, 'timing/data': 0.013294802978634834, 'timing/model': 1.2788399420678616, 'learning_rate/qwen_vl_interface': 9.661939411808623e-06, 'learning_rate/action_model': 9.661939411808622e-05, 'epoch': 5.99}) +[2026-06-20 14:12:08,052][starVLA.training.train_starvla][INFO] - Step 3293, Loss: {'train/loss': 0.24924597144126892, 'train/grad_norm_pre_clip': 11.106454849243164, 'timing/data': 0.015168807003647089, 'timing/model': 1.206256728968583, 'learning_rate/qwen_vl_interface': 9.66081015265669e-06, 'learning_rate/action_model': 9.660810152656689e-05, 'epoch': 5.99}) +[2026-06-20 14:12:13,431][starVLA.training.train_starvla][INFO] - Step 3294, Loss: {'train/loss': 0.19998995959758759, 'train/grad_norm_pre_clip': 9.325854301452637, 'timing/data': 0.011482520028948784, 'timing/model': 1.4936581699876115, 'learning_rate/qwen_vl_interface': 9.659679080387314e-06, 'learning_rate/action_model': 9.659679080387313e-05, 'epoch': 6.0}) +[2026-06-20 14:12:18,300][starVLA.training.train_starvla][INFO] - Step 3295, Loss: {'train/loss': 0.2351517379283905, 'train/grad_norm_pre_clip': 9.131805419921875, 'timing/data': 0.011557868914678693, 'timing/model': 1.1780141639756039, 'learning_rate/qwen_vl_interface': 9.658546195465436e-06, 'learning_rate/action_model': 9.658546195465435e-05, 'epoch': 6.0}) +[2026-06-20 14:12:23,719][starVLA.training.train_starvla][INFO] - Step 3296, Loss: {'train/loss': 0.04678524285554886, 'train/grad_norm_pre_clip': 2.9535415172576904, 'timing/data': 0.015582734951749444, 'timing/model': 1.5641915630549192, 'learning_rate/qwen_vl_interface': 9.65741149835674e-06, 'learning_rate/action_model': 9.657411498356741e-05, 'epoch': 6.0}) +[2026-06-20 14:12:28,641][starVLA.training.train_starvla][INFO] - Step 3297, Loss: {'train/loss': 0.03044411540031433, 'train/grad_norm_pre_clip': 2.4174270629882812, 'timing/data': 0.011792468023486435, 'timing/model': 1.254079867969267, 'learning_rate/qwen_vl_interface': 9.656274989527663e-06, 'learning_rate/action_model': 9.656274989527664e-05, 'epoch': 6.0}) +[2026-06-20 14:12:33,628][starVLA.training.train_starvla][INFO] - Step 3298, Loss: {'train/loss': 0.08328339457511902, 'train/grad_norm_pre_clip': 4.96774959564209, 'timing/data': 0.011370041058398783, 'timing/model': 1.1935779911000282, 'learning_rate/qwen_vl_interface': 9.655136669445378e-06, 'learning_rate/action_model': 9.655136669445378e-05, 'epoch': 6.0}) +[2026-06-20 14:12:38,799][starVLA.training.train_starvla][INFO] - Step 3299, Loss: {'train/loss': 0.04996780678629875, 'train/grad_norm_pre_clip': 3.817108631134033, 'timing/data': 0.011145989992655814, 'timing/model': 1.447019062936306, 'learning_rate/qwen_vl_interface': 9.653996538577802e-06, 'learning_rate/action_model': 9.653996538577802e-05, 'epoch': 6.01}) +[2026-06-20 14:12:43,908][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 3300 +[2026-06-20 14:19:04,865][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 3300 +[2026-06-20 14:19:04,866][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 3300 +[2026-06-20 14:25:37,030][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 3300 +[2026-06-20 14:25:37,032][starVLA.training.train_starvla][INFO] - Step 3300, Loss: {'train/loss': 0.06320258975028992, 'train/grad_norm_pre_clip': 4.919093132019043, 'eval/loss': 0.45420268177986145, 'eval/latency_0/loss': 0.4356570839881897, 'eval/latency_1/loss': 0.624192476272583, 'eval/latency_2/loss': 0.5948716998100281, 'eval/latency_3/loss': 0.09301449358463287, 'eval/latency_4/loss': 0.47945961356163025, 'eval/flappy/loss': 0.45420268177986145, 'eval/flappy/latency_0/loss': 0.4356570839881897, 'eval/flappy/latency_1/loss': 0.624192476272583, 'eval/flappy/latency_2/loss': 0.5948716998100281, 'eval/flappy/latency_3/loss': 0.09301449358463287, 'eval/flappy/latency_4/loss': 0.47945961356163025, 'eval/flap_precision': 0.18604651162790697, 'eval/flap_recall': 0.11636363636363636, 'eval/flap_f1': 0.14317673378076062, 'eval/flap_rate_model': 0.04777777777777778, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.14317673378076062, 'eval/latency_0/flap_precision': 0.18604651162790697, 'eval/latency_0/flap_recall': 0.11636363636363636, 'eval/latency_0/flap_f1': 0.14317673378076062, 'eval/latency_0/flap_rate_model': 0.04777777777777778, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.14317673378076062, 'timing/data': 0.014304949901998043, 'timing/model': 1.2539571420056745, 'learning_rate/qwen_vl_interface': 9.652854597393604e-06, 'learning_rate/action_model': 9.652854597393603e-05, 'epoch': 6.01}) +[2026-06-20 14:25:37,032][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_3300_state +[2026-06-20 14:25:37,036][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-20 14:25:54,730][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_3300_state/model.safetensors +[2026-06-20 14:26:19,779][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_3300_state/optimizer.bin +[2026-06-20 14:26:19,780][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_3300_state/sampler.bin +[2026-06-20 14:26:19,780][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_3300_state/sampler_1.bin +[2026-06-20 14:26:19,781][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_3300_state/random_states_0.pkl +[2026-06-20 14:26:21,155][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3300 +[2026-06-20 14:26:21,155][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-20 14:26:21,168][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-20 14:26:25,927][starVLA.training.train_starvla][INFO] - Step 3301, Loss: {'train/loss': 0.2184608429670334, 'train/grad_norm_pre_clip': 13.149882316589355, 'timing/data': 0.011886010062880814, 'timing/model': 1.1745404310058802, 'learning_rate/qwen_vl_interface': 9.651710846362192e-06, 'learning_rate/action_model': 9.651710846362192e-05, 'epoch': 6.01}) +[2026-06-20 14:26:30,722][starVLA.training.train_starvla][INFO] - Step 3302, Loss: {'train/loss': 0.11925794184207916, 'train/grad_norm_pre_clip': 6.893258094787598, 'timing/data': 0.011450069956481457, 'timing/model': 1.1707806120393798, 'learning_rate/qwen_vl_interface': 9.650565285953716e-06, 'learning_rate/action_model': 9.650565285953715e-05, 'epoch': 6.01}) +[2026-06-20 14:26:35,606][starVLA.training.train_starvla][INFO] - Step 3303, Loss: {'train/loss': 0.23206137120723724, 'train/grad_norm_pre_clip': 8.755240440368652, 'timing/data': 0.013153998064808547, 'timing/model': 1.4636419920716435, 'learning_rate/qwen_vl_interface': 9.649417916639077e-06, 'learning_rate/action_model': 9.649417916639076e-05, 'epoch': 6.01}) +[2026-06-20 14:26:40,407][starVLA.training.train_starvla][INFO] - Step 3304, Loss: {'train/loss': 0.17593728005886078, 'train/grad_norm_pre_clip': 9.083892822265625, 'timing/data': 0.011668590013869107, 'timing/model': 1.250950479065068, 'learning_rate/qwen_vl_interface': 9.648268738889911e-06, 'learning_rate/action_model': 9.648268738889911e-05, 'epoch': 6.01}) +[2026-06-20 14:26:45,239][starVLA.training.train_starvla][INFO] - Step 3305, Loss: {'train/loss': 0.21886548399925232, 'train/grad_norm_pre_clip': 10.586700439453125, 'timing/data': 0.010897001018747687, 'timing/model': 1.1655361319426447, 'learning_rate/qwen_vl_interface': 9.647117753178604e-06, 'learning_rate/action_model': 9.647117753178604e-05, 'epoch': 6.02}) +[2026-06-20 14:26:50,020][starVLA.training.train_starvla][INFO] - Step 3306, Loss: {'train/loss': 0.028491009026765823, 'train/grad_norm_pre_clip': 2.061391592025757, 'timing/data': 0.011791030061431229, 'timing/model': 1.1719039760064334, 'learning_rate/qwen_vl_interface': 9.645964959978284e-06, 'learning_rate/action_model': 9.645964959978284e-05, 'epoch': 6.02}) +[2026-06-20 14:26:54,768][starVLA.training.train_starvla][INFO] - Step 3307, Loss: {'train/loss': 0.11124611645936966, 'train/grad_norm_pre_clip': 8.523788452148438, 'timing/data': 0.012115549063310027, 'timing/model': 1.1785935988882557, 'learning_rate/qwen_vl_interface': 9.644810359762818e-06, 'learning_rate/action_model': 9.644810359762817e-05, 'epoch': 6.02}) +[2026-06-20 14:26:59,627][starVLA.training.train_starvla][INFO] - Step 3308, Loss: {'train/loss': 0.16034454107284546, 'train/grad_norm_pre_clip': 15.38046646118164, 'timing/data': 0.01735909399576485, 'timing/model': 1.2272115820087492, 'learning_rate/qwen_vl_interface': 9.64365395300682e-06, 'learning_rate/action_model': 9.643653953006819e-05, 'epoch': 6.02}) +[2026-06-20 14:27:04,486][starVLA.training.train_starvla][INFO] - Step 3309, Loss: {'train/loss': 0.12646979093551636, 'train/grad_norm_pre_clip': 8.997007369995117, 'timing/data': 0.011703527998179197, 'timing/model': 1.4592264669481665, 'learning_rate/qwen_vl_interface': 9.642495740185644e-06, 'learning_rate/action_model': 9.642495740185643e-05, 'epoch': 6.02}) +[2026-06-20 14:27:09,315][starVLA.training.train_starvla][INFO] - Step 3310, Loss: {'train/loss': 0.010566214099526405, 'train/grad_norm_pre_clip': 2.229170560836792, 'timing/data': 0.012103907996788621, 'timing/model': 1.289144217967987, 'learning_rate/qwen_vl_interface': 9.641335721775388e-06, 'learning_rate/action_model': 9.641335721775387e-05, 'epoch': 6.03}) +[2026-06-20 14:27:14,243][starVLA.training.train_starvla][INFO] - Step 3311, Loss: {'train/loss': 0.17269013822078705, 'train/grad_norm_pre_clip': 11.852250099182129, 'timing/data': 0.011430868995375931, 'timing/model': 1.1857249879976735, 'learning_rate/qwen_vl_interface': 9.640173898252893e-06, 'learning_rate/action_model': 9.640173898252892e-05, 'epoch': 6.03}) +[2026-06-20 14:27:19,236][starVLA.training.train_starvla][INFO] - Step 3312, Loss: {'train/loss': 0.10666123777627945, 'train/grad_norm_pre_clip': 15.192362785339355, 'timing/data': 0.012263547978363931, 'timing/model': 1.1796157429926097, 'learning_rate/qwen_vl_interface': 9.639010270095738e-06, 'learning_rate/action_model': 9.639010270095739e-05, 'epoch': 6.03}) +[2026-06-20 14:27:24,151][starVLA.training.train_starvla][INFO] - Step 3313, Loss: {'train/loss': 0.3903907239437103, 'train/grad_norm_pre_clip': 21.984724044799805, 'timing/data': 0.01189417706336826, 'timing/model': 1.1764021769631654, 'learning_rate/qwen_vl_interface': 9.637844837782251e-06, 'learning_rate/action_model': 9.63784483778225e-05, 'epoch': 6.03}) +[2026-06-20 14:27:28,981][starVLA.training.train_starvla][INFO] - Step 3314, Loss: {'train/loss': 0.05470145121216774, 'train/grad_norm_pre_clip': 7.307964324951172, 'timing/data': 0.011888386914506555, 'timing/model': 1.1862916279351339, 'learning_rate/qwen_vl_interface': 9.636677601791494e-06, 'learning_rate/action_model': 9.636677601791493e-05, 'epoch': 6.03}) +[2026-06-20 14:27:34,050][starVLA.training.train_starvla][INFO] - Step 3315, Loss: {'train/loss': 0.3164503574371338, 'train/grad_norm_pre_clip': 12.891372680664062, 'timing/data': 0.011481927940621972, 'timing/model': 1.5118452809983864, 'learning_rate/qwen_vl_interface': 9.635508562603273e-06, 'learning_rate/action_model': 9.635508562603272e-05, 'epoch': 6.03}) +[2026-06-20 14:27:38,792][starVLA.training.train_starvla][INFO] - Step 3316, Loss: {'train/loss': 0.08768559247255325, 'train/grad_norm_pre_clip': 12.062042236328125, 'timing/data': 0.011477808002382517, 'timing/model': 1.173007630975917, 'learning_rate/qwen_vl_interface': 9.634337720698139e-06, 'learning_rate/action_model': 9.634337720698137e-05, 'epoch': 6.04}) +[2026-06-20 14:27:45,135][starVLA.training.train_starvla][INFO] - Step 3317, Loss: {'train/loss': 0.10693369805812836, 'train/grad_norm_pre_clip': 10.847405433654785, 'timing/data': 0.011234577978029847, 'timing/model': 1.912502521998249, 'learning_rate/qwen_vl_interface': 9.633165076557376e-06, 'learning_rate/action_model': 9.633165076557375e-05, 'epoch': 6.04}) +[2026-06-20 14:27:50,145][starVLA.training.train_starvla][INFO] - Step 3318, Loss: {'train/loss': 0.2642504870891571, 'train/grad_norm_pre_clip': 19.991846084594727, 'timing/data': 0.011850486975163221, 'timing/model': 1.2411437600385398, 'learning_rate/qwen_vl_interface': 9.631990630663019e-06, 'learning_rate/action_model': 9.631990630663018e-05, 'epoch': 6.04}) +[2026-06-20 14:27:55,068][starVLA.training.train_starvla][INFO] - Step 3319, Loss: {'train/loss': 0.06889673322439194, 'train/grad_norm_pre_clip': 8.86396598815918, 'timing/data': 0.011573827010579407, 'timing/model': 1.195020123035647, 'learning_rate/qwen_vl_interface': 9.630814383497836e-06, 'learning_rate/action_model': 9.630814383497835e-05, 'epoch': 6.04}) +[2026-06-20 14:28:00,040][starVLA.training.train_starvla][INFO] - Step 3320, Loss: {'train/loss': 0.16999469697475433, 'train/grad_norm_pre_clip': 17.35920524597168, 'timing/data': 0.012290295911952853, 'timing/model': 1.258224700926803, 'learning_rate/qwen_vl_interface': 9.629636335545336e-06, 'learning_rate/action_model': 9.629636335545335e-05, 'epoch': 6.04}) +[2026-06-20 14:28:05,172][starVLA.training.train_starvla][INFO] - Step 3321, Loss: {'train/loss': 0.2631710171699524, 'train/grad_norm_pre_clip': 15.095664024353027, 'timing/data': 0.013863214058801532, 'timing/model': 1.5676589290378615, 'learning_rate/qwen_vl_interface': 9.628456487289773e-06, 'learning_rate/action_model': 9.628456487289773e-05, 'epoch': 6.05}) +[2026-06-20 14:28:09,998][starVLA.training.train_starvla][INFO] - Step 3322, Loss: {'train/loss': 0.02178337797522545, 'train/grad_norm_pre_clip': 4.700344562530518, 'timing/data': 0.011482047033496201, 'timing/model': 1.253380257054232, 'learning_rate/qwen_vl_interface': 9.627274839216135e-06, 'learning_rate/action_model': 9.627274839216136e-05, 'epoch': 6.05}) +[2026-06-20 14:28:14,924][starVLA.training.train_starvla][INFO] - Step 3323, Loss: {'train/loss': 0.19596312940120697, 'train/grad_norm_pre_clip': 16.158279418945312, 'timing/data': 0.011323975981213152, 'timing/model': 1.183557890006341, 'learning_rate/qwen_vl_interface': 9.626091391810158e-06, 'learning_rate/action_model': 9.626091391810157e-05, 'epoch': 6.05}) +[2026-06-20 14:28:20,015][starVLA.training.train_starvla][INFO] - Step 3324, Loss: {'train/loss': 0.03883315995335579, 'train/grad_norm_pre_clip': 6.426018238067627, 'timing/data': 0.01781981799285859, 'timing/model': 1.2140092049958184, 'learning_rate/qwen_vl_interface': 9.624906145558309e-06, 'learning_rate/action_model': 9.624906145558307e-05, 'epoch': 6.05}) +[2026-06-20 14:28:24,921][starVLA.training.train_starvla][INFO] - Step 3325, Loss: {'train/loss': 0.05669252201914787, 'train/grad_norm_pre_clip': 6.888862609863281, 'timing/data': 0.01218639500439167, 'timing/model': 1.1861461099470034, 'learning_rate/qwen_vl_interface': 9.623719100947795e-06, 'learning_rate/action_model': 9.623719100947795e-05, 'epoch': 6.05}) +[2026-06-20 14:28:29,765][starVLA.training.train_starvla][INFO] - Step 3326, Loss: {'train/loss': 0.1919640451669693, 'train/grad_norm_pre_clip': 11.39360523223877, 'timing/data': 0.011366725084371865, 'timing/model': 1.1758589060045779, 'learning_rate/qwen_vl_interface': 9.622530258466571e-06, 'learning_rate/action_model': 9.62253025846657e-05, 'epoch': 6.05}) +[2026-06-20 14:28:34,653][starVLA.training.train_starvla][INFO] - Step 3327, Loss: {'train/loss': 0.12280359864234924, 'train/grad_norm_pre_clip': 15.771286010742188, 'timing/data': 0.011750714969821274, 'timing/model': 1.451670749927871, 'learning_rate/qwen_vl_interface': 9.62133961860332e-06, 'learning_rate/action_model': 9.62133961860332e-05, 'epoch': 6.06}) +[2026-06-20 14:28:39,221][starVLA.training.train_starvla][INFO] - Step 3328, Loss: {'train/loss': 0.23394525051116943, 'train/grad_norm_pre_clip': 14.939471244812012, 'timing/data': 0.011390154948458076, 'timing/model': 1.1907190119381994, 'learning_rate/qwen_vl_interface': 9.620147181847474e-06, 'learning_rate/action_model': 9.620147181847475e-05, 'epoch': 6.06}) +[2026-06-20 14:28:44,176][starVLA.training.train_starvla][INFO] - Step 3329, Loss: {'train/loss': 0.011351536028087139, 'train/grad_norm_pre_clip': 1.2279021739959717, 'timing/data': 0.014015331980772316, 'timing/model': 1.3143027869518846, 'learning_rate/qwen_vl_interface': 9.618952948689196e-06, 'learning_rate/action_model': 9.618952948689196e-05, 'epoch': 6.06}) +[2026-06-20 14:28:49,041][starVLA.training.train_starvla][INFO] - Step 3330, Loss: {'train/loss': 0.0923663079738617, 'train/grad_norm_pre_clip': 9.5435152053833, 'timing/data': 0.012037324020639062, 'timing/model': 1.186277834000066, 'learning_rate/qwen_vl_interface': 9.617756919619393e-06, 'learning_rate/action_model': 9.617756919619392e-05, 'epoch': 6.06}) +[2026-06-20 14:28:54,051][starVLA.training.train_starvla][INFO] - Step 3331, Loss: {'train/loss': 0.17894820868968964, 'train/grad_norm_pre_clip': 12.052729606628418, 'timing/data': 0.012434144038707018, 'timing/model': 1.213575000059791, 'learning_rate/qwen_vl_interface': 9.616559095129701e-06, 'learning_rate/action_model': 9.6165590951297e-05, 'epoch': 6.06}) +[2026-06-20 14:28:58,943][starVLA.training.train_starvla][INFO] - Step 3332, Loss: {'train/loss': 0.03783080354332924, 'train/grad_norm_pre_clip': 8.083690643310547, 'timing/data': 0.011654644971713424, 'timing/model': 1.2143363009672612, 'learning_rate/qwen_vl_interface': 9.615359475712506e-06, 'learning_rate/action_model': 9.615359475712506e-05, 'epoch': 6.07}) +[2026-06-20 14:29:03,968][starVLA.training.train_starvla][INFO] - Step 3333, Loss: {'train/loss': 0.044533099979162216, 'train/grad_norm_pre_clip': 7.168249130249023, 'timing/data': 0.01761387602891773, 'timing/model': 1.5360217010602355, 'learning_rate/qwen_vl_interface': 9.614158061860924e-06, 'learning_rate/action_model': 9.614158061860924e-05, 'epoch': 6.07}) +[2026-06-20 14:29:12,172][starVLA.training.train_starvla][INFO] - Latency curriculum phase 4 at step 3334: active_latencies=[4], dataset_size=69213 +[2026-06-20 14:29:12,173][starVLA.training.train_starvla][INFO] - Step 3334, Loss: {'train/loss': 0.135001078248024, 'train/grad_norm_pre_clip': 11.42141342163086, 'latency_curriculum/phase': 4.0, 'latency_curriculum/active_count': 1.0, 'latency_curriculum/max_active_latency': 4.0, 'latency_curriculum/dataset_size': 69213.0, 'timing/data': 0.013841720996424556, 'timing/model': 1.1852380570489913, 'learning_rate/qwen_vl_interface': 9.612954854068813e-06, 'learning_rate/action_model': 9.612954854068812e-05, 'epoch': 6.17}) +[2026-06-20 14:29:42,243][starVLA.training.train_starvla][INFO] - Step 3335, Loss: {'train/loss': 1.3162665367126465, 'train/grad_norm_pre_clip': 28.731353759765625, 'timing/data': 0.0166415759595111, 'timing/model': 1.1705948999151587, 'learning_rate/qwen_vl_interface': 9.611749852830763e-06, 'learning_rate/action_model': 9.611749852830763e-05, 'epoch': 6.17}) +[2026-06-20 14:29:47,315][starVLA.training.train_starvla][INFO] - Step 3336, Loss: {'train/loss': 1.5278282165527344, 'train/grad_norm_pre_clip': 36.23982238769531, 'timing/data': 0.011321592959575355, 'timing/model': 1.424351701978594, 'learning_rate/qwen_vl_interface': 9.610543058642107e-06, 'learning_rate/action_model': 9.610543058642107e-05, 'epoch': 6.17}) +[2026-06-20 14:29:53,076][starVLA.training.train_starvla][INFO] - Step 3337, Loss: {'train/loss': 0.6547421813011169, 'train/grad_norm_pre_clip': 16.039392471313477, 'timing/data': 0.013425381039269269, 'timing/model': 1.5337286009453237, 'learning_rate/qwen_vl_interface': 9.609334471998911e-06, 'learning_rate/action_model': 9.609334471998912e-05, 'epoch': 6.17}) +[2026-06-20 14:29:57,940][starVLA.training.train_starvla][INFO] - Step 3338, Loss: {'train/loss': 1.1075634956359863, 'train/grad_norm_pre_clip': 17.23139190673828, 'timing/data': 0.01424319890793413, 'timing/model': 1.1970228119753301, 'learning_rate/qwen_vl_interface': 9.60812409339798e-06, 'learning_rate/action_model': 9.60812409339798e-05, 'epoch': 6.17}) +[2026-06-20 14:30:02,731][starVLA.training.train_starvla][INFO] - Step 3339, Loss: {'train/loss': 1.6405614614486694, 'train/grad_norm_pre_clip': 17.21381950378418, 'timing/data': 0.012201342033222318, 'timing/model': 1.412988172029145, 'learning_rate/qwen_vl_interface': 9.606911923336856e-06, 'learning_rate/action_model': 9.606911923336856e-05, 'epoch': 6.18}) +[2026-06-20 14:30:07,564][starVLA.training.train_starvla][INFO] - Step 3340, Loss: {'train/loss': 0.8664974570274353, 'train/grad_norm_pre_clip': 16.379919052124023, 'timing/data': 0.01221064105629921, 'timing/model': 1.4014464420033619, 'learning_rate/qwen_vl_interface': 9.605697962313817e-06, 'learning_rate/action_model': 9.605697962313816e-05, 'epoch': 6.18}) +[2026-06-20 14:30:12,276][starVLA.training.train_starvla][INFO] - Step 3341, Loss: {'train/loss': 1.285516619682312, 'train/grad_norm_pre_clip': 16.89234161376953, 'timing/data': 0.011701822048053145, 'timing/model': 1.2950522989267483, 'learning_rate/qwen_vl_interface': 9.604482210827871e-06, 'learning_rate/action_model': 9.604482210827871e-05, 'epoch': 6.18}) +[2026-06-20 14:30:17,492][starVLA.training.train_starvla][INFO] - Step 3342, Loss: {'train/loss': 0.36142921447753906, 'train/grad_norm_pre_clip': 9.721741676330566, 'timing/data': 0.011544881970621645, 'timing/model': 1.4406276270747185, 'learning_rate/qwen_vl_interface': 9.603264669378772e-06, 'learning_rate/action_model': 9.603264669378772e-05, 'epoch': 6.18}) +[2026-06-20 14:30:22,484][starVLA.training.train_starvla][INFO] - Step 3343, Loss: {'train/loss': 0.35732167959213257, 'train/grad_norm_pre_clip': 11.436737060546875, 'timing/data': 0.011596341035328805, 'timing/model': 1.2805754679720849, 'learning_rate/qwen_vl_interface': 9.602045338467007e-06, 'learning_rate/action_model': 9.602045338467006e-05, 'epoch': 6.18}) +[2026-06-20 14:30:27,399][starVLA.training.train_starvla][INFO] - Step 3344, Loss: {'train/loss': 0.4366070330142975, 'train/grad_norm_pre_clip': 11.473382949829102, 'timing/data': 0.01739582302980125, 'timing/model': 1.1959329660749063, 'learning_rate/qwen_vl_interface': 9.600824218593791e-06, 'learning_rate/action_model': 9.600824218593791e-05, 'epoch': 6.18}) +[2026-06-20 14:30:32,486][starVLA.training.train_starvla][INFO] - Step 3345, Loss: {'train/loss': 0.5065674781799316, 'train/grad_norm_pre_clip': 8.634780883789062, 'timing/data': 0.01273564004804939, 'timing/model': 1.3164231799310073, 'learning_rate/qwen_vl_interface': 9.599601310261087e-06, 'learning_rate/action_model': 9.599601310261085e-05, 'epoch': 6.19}) +[2026-06-20 14:30:37,502][starVLA.training.train_starvla][INFO] - Step 3346, Loss: {'train/loss': 0.4374825060367584, 'train/grad_norm_pre_clip': 6.419034957885742, 'timing/data': 0.013218799023889005, 'timing/model': 1.4956120200222358, 'learning_rate/qwen_vl_interface': 9.598376613971582e-06, 'learning_rate/action_model': 9.598376613971582e-05, 'epoch': 6.19}) +[2026-06-20 14:30:42,420][starVLA.training.train_starvla][INFO] - Step 3347, Loss: {'train/loss': 0.3130064904689789, 'train/grad_norm_pre_clip': 5.955646514892578, 'timing/data': 0.013689858024008572, 'timing/model': 1.2332438830053434, 'learning_rate/qwen_vl_interface': 9.597150130228703e-06, 'learning_rate/action_model': 9.597150130228703e-05, 'epoch': 6.19}) +[2026-06-20 14:30:47,502][starVLA.training.train_starvla][INFO] - Step 3348, Loss: {'train/loss': 0.39746007323265076, 'train/grad_norm_pre_clip': 7.382796287536621, 'timing/data': 0.013046048930846155, 'timing/model': 1.227996755973436, 'learning_rate/qwen_vl_interface': 9.595921859536612e-06, 'learning_rate/action_model': 9.595921859536611e-05, 'epoch': 6.19}) +[2026-06-20 14:30:52,504][starVLA.training.train_starvla][INFO] - Step 3349, Loss: {'train/loss': 0.5624818205833435, 'train/grad_norm_pre_clip': 9.355195045471191, 'timing/data': 0.012839920003898442, 'timing/model': 1.227851542062126, 'learning_rate/qwen_vl_interface': 9.594691802400203e-06, 'learning_rate/action_model': 9.594691802400202e-05, 'epoch': 6.19}) +[2026-06-20 14:30:57,381][starVLA.training.train_starvla][INFO] - Step 3350, Loss: {'train/loss': 0.442714661359787, 'train/grad_norm_pre_clip': 5.5872368812561035, 'timing/data': 0.011666840990073979, 'timing/model': 1.1761694129090756, 'learning_rate/qwen_vl_interface': 9.59345995932511e-06, 'learning_rate/action_model': 9.59345995932511e-05, 'epoch': 6.2}) +[2026-06-20 14:31:02,305][starVLA.training.train_starvla][INFO] - Step 3351, Loss: {'train/loss': 0.4719182848930359, 'train/grad_norm_pre_clip': 16.867778778076172, 'timing/data': 0.011687899939715862, 'timing/model': 1.4294880820671096, 'learning_rate/qwen_vl_interface': 9.592226330817693e-06, 'learning_rate/action_model': 9.592226330817693e-05, 'epoch': 6.2}) +[2026-06-20 14:31:07,048][starVLA.training.train_starvla][INFO] - Step 3352, Loss: {'train/loss': 0.31080740690231323, 'train/grad_norm_pre_clip': 4.7320990562438965, 'timing/data': 0.011449411045759916, 'timing/model': 1.248581237043254, 'learning_rate/qwen_vl_interface': 9.590990917385052e-06, 'learning_rate/action_model': 9.590990917385052e-05, 'epoch': 6.2}) +[2026-06-20 14:31:11,947][starVLA.training.train_starvla][INFO] - Step 3353, Loss: {'train/loss': 0.43078479170799255, 'train/grad_norm_pre_clip': 5.326832294464111, 'timing/data': 0.011870309943333268, 'timing/model': 1.199960744008422, 'learning_rate/qwen_vl_interface': 9.58975371953502e-06, 'learning_rate/action_model': 9.589753719535018e-05, 'epoch': 6.2}) +[2026-06-20 14:31:16,927][starVLA.training.train_starvla][INFO] - Step 3354, Loss: {'train/loss': 0.42662471532821655, 'train/grad_norm_pre_clip': 4.571209907531738, 'timing/data': 0.011682520038448274, 'timing/model': 1.1940465900115669, 'learning_rate/qwen_vl_interface': 9.588514737776158e-06, 'learning_rate/action_model': 9.588514737776157e-05, 'epoch': 6.2}) +[2026-06-20 14:31:22,080][starVLA.training.train_starvla][INFO] - Step 3355, Loss: {'train/loss': 0.4232085347175598, 'train/grad_norm_pre_clip': 4.137982368469238, 'timing/data': 0.013525997055694461, 'timing/model': 1.5291945500066504, 'learning_rate/qwen_vl_interface': 9.587273972617769e-06, 'learning_rate/action_model': 9.587273972617768e-05, 'epoch': 6.2}) +[2026-06-20 14:31:27,006][starVLA.training.train_starvla][INFO] - Step 3356, Loss: {'train/loss': 0.5351240038871765, 'train/grad_norm_pre_clip': 5.958725929260254, 'timing/data': 0.01138471090234816, 'timing/model': 1.2099000179441646, 'learning_rate/qwen_vl_interface': 9.586031424569882e-06, 'learning_rate/action_model': 9.58603142456988e-05, 'epoch': 6.21}) +[2026-06-20 14:31:32,496][starVLA.training.train_starvla][INFO] - Step 3357, Loss: {'train/loss': 0.3592938482761383, 'train/grad_norm_pre_clip': 4.300039291381836, 'timing/data': 0.015124354045838118, 'timing/model': 1.7226777239702642, 'learning_rate/qwen_vl_interface': 9.584787094143265e-06, 'learning_rate/action_model': 9.584787094143264e-05, 'epoch': 6.21}) +[2026-06-20 14:31:37,260][starVLA.training.train_starvla][INFO] - Step 3358, Loss: {'train/loss': 0.4408870041370392, 'train/grad_norm_pre_clip': 5.29179048538208, 'timing/data': 0.013474366976879537, 'timing/model': 1.2541860439814627, 'learning_rate/qwen_vl_interface': 9.58354098184941e-06, 'learning_rate/action_model': 9.583540981849409e-05, 'epoch': 6.21}) +[2026-06-20 14:31:42,538][starVLA.training.train_starvla][INFO] - Step 3359, Loss: {'train/loss': 0.5204142332077026, 'train/grad_norm_pre_clip': 5.66981840133667, 'timing/data': 0.015156024019233882, 'timing/model': 1.2790529169142246, 'learning_rate/qwen_vl_interface': 9.582293088200549e-06, 'learning_rate/action_model': 9.582293088200549e-05, 'epoch': 6.21}) +[2026-06-20 14:31:47,976][starVLA.training.train_starvla][INFO] - Step 3360, Loss: {'train/loss': 0.420636385679245, 'train/grad_norm_pre_clip': 5.359379768371582, 'timing/data': 0.01115222997032106, 'timing/model': 1.2019223279785365, 'learning_rate/qwen_vl_interface': 9.581043413709646e-06, 'learning_rate/action_model': 9.581043413709646e-05, 'epoch': 6.21}) +[2026-06-20 14:31:53,040][starVLA.training.train_starvla][INFO] - Step 3361, Loss: {'train/loss': 0.24560189247131348, 'train/grad_norm_pre_clip': 3.308509111404419, 'timing/data': 0.011707638972438872, 'timing/model': 1.181400632020086, 'learning_rate/qwen_vl_interface': 9.579791958890393e-06, 'learning_rate/action_model': 9.579791958890394e-05, 'epoch': 6.22}) +[2026-06-20 14:31:58,013][starVLA.training.train_starvla][INFO] - Step 3362, Loss: {'train/loss': 0.3088667690753937, 'train/grad_norm_pre_clip': 2.2349159717559814, 'timing/data': 0.011465169955044985, 'timing/model': 1.18708007899113, 'learning_rate/qwen_vl_interface': 9.578538724257218e-06, 'learning_rate/action_model': 9.578538724257216e-05, 'epoch': 6.22}) +[2026-06-20 14:32:02,896][starVLA.training.train_starvla][INFO] - Step 3363, Loss: {'train/loss': 0.240461066365242, 'train/grad_norm_pre_clip': 4.2447509765625, 'timing/data': 0.011758467997424304, 'timing/model': 1.4456461480585858, 'learning_rate/qwen_vl_interface': 9.577283710325274e-06, 'learning_rate/action_model': 9.577283710325274e-05, 'epoch': 6.22}) +[2026-06-20 14:32:07,647][starVLA.training.train_starvla][INFO] - Step 3364, Loss: {'train/loss': 0.261344850063324, 'train/grad_norm_pre_clip': 4.761045932769775, 'timing/data': 0.011750368983484805, 'timing/model': 1.1940832230029628, 'learning_rate/qwen_vl_interface': 9.576026917610455e-06, 'learning_rate/action_model': 9.576026917610455e-05, 'epoch': 6.22}) +[2026-06-20 14:32:12,804][starVLA.training.train_starvla][INFO] - Step 3365, Loss: {'train/loss': 0.3129911720752716, 'train/grad_norm_pre_clip': 16.11353302001953, 'timing/data': 0.011660809046588838, 'timing/model': 1.2391695000696927, 'learning_rate/qwen_vl_interface': 9.574768346629376e-06, 'learning_rate/action_model': 9.574768346629376e-05, 'epoch': 6.22}) +[2026-06-20 14:32:17,648][starVLA.training.train_starvla][INFO] - Step 3366, Loss: {'train/loss': 0.24946030974388123, 'train/grad_norm_pre_clip': 4.738988399505615, 'timing/data': 0.011748058954253793, 'timing/model': 1.178823816939257, 'learning_rate/qwen_vl_interface': 9.573507997899391e-06, 'learning_rate/action_model': 9.573507997899392e-05, 'epoch': 6.22}) +[2026-06-20 14:32:22,457][starVLA.training.train_starvla][INFO] - Step 3367, Loss: {'train/loss': 0.21946868300437927, 'train/grad_norm_pre_clip': 5.470561504364014, 'timing/data': 0.012038668035529554, 'timing/model': 1.1865754829486832, 'learning_rate/qwen_vl_interface': 9.572245871938582e-06, 'learning_rate/action_model': 9.572245871938582e-05, 'epoch': 6.23}) +[2026-06-20 14:32:27,250][starVLA.training.train_starvla][INFO] - Step 3368, Loss: {'train/loss': 0.4001021385192871, 'train/grad_norm_pre_clip': 9.431039810180664, 'timing/data': 0.012338116997852921, 'timing/model': 1.193723547970876, 'learning_rate/qwen_vl_interface': 9.570981969265761e-06, 'learning_rate/action_model': 9.570981969265761e-05, 'epoch': 6.23}) +[2026-06-20 14:32:32,088][starVLA.training.train_starvla][INFO] - Step 3369, Loss: {'train/loss': 0.36526861786842346, 'train/grad_norm_pre_clip': 14.14981746673584, 'timing/data': 0.012677925988100469, 'timing/model': 1.4497041570721194, 'learning_rate/qwen_vl_interface': 9.56971629040047e-06, 'learning_rate/action_model': 9.56971629040047e-05, 'epoch': 6.23}) +[2026-06-20 14:32:36,899][starVLA.training.train_starvla][INFO] - Step 3370, Loss: {'train/loss': 0.4766339659690857, 'train/grad_norm_pre_clip': 6.46803617477417, 'timing/data': 0.011977397953160107, 'timing/model': 1.1933531870599836, 'learning_rate/qwen_vl_interface': 9.56844883586298e-06, 'learning_rate/action_model': 9.568448835862981e-05, 'epoch': 6.23}) +[2026-06-20 14:32:41,732][starVLA.training.train_starvla][INFO] - Step 3371, Loss: {'train/loss': 0.3282594084739685, 'train/grad_norm_pre_clip': 8.164704322814941, 'timing/data': 0.0005270700203254819, 'timing/model': 1.180711310938932, 'learning_rate/qwen_vl_interface': 9.567179606174299e-06, 'learning_rate/action_model': 9.567179606174299e-05, 'epoch': 6.23}) +[2026-06-20 14:32:46,604][starVLA.training.train_starvla][INFO] - Step 3372, Loss: {'train/loss': 0.4264049232006073, 'train/grad_norm_pre_clip': 6.393074035644531, 'timing/data': 0.0005851290188729763, 'timing/model': 1.1938045140122995, 'learning_rate/qwen_vl_interface': 9.565908601856155e-06, 'learning_rate/action_model': 9.565908601856154e-05, 'epoch': 6.24}) +[2026-06-20 14:32:51,492][starVLA.training.train_starvla][INFO] - Step 3373, Loss: {'train/loss': 0.15571154654026031, 'train/grad_norm_pre_clip': 4.491219520568848, 'timing/data': 0.0005357490153983235, 'timing/model': 1.1947068019071594, 'learning_rate/qwen_vl_interface': 9.56463582343101e-06, 'learning_rate/action_model': 9.56463582343101e-05, 'epoch': 6.24}) +[2026-06-20 14:32:56,280][starVLA.training.train_starvla][INFO] - Step 3374, Loss: {'train/loss': 0.27924367785453796, 'train/grad_norm_pre_clip': 27.938573837280273, 'timing/data': 0.0005440489621832967, 'timing/model': 1.175137817976065, 'learning_rate/qwen_vl_interface': 9.563361271422056e-06, 'learning_rate/action_model': 9.563361271422056e-05, 'epoch': 6.24}) +[2026-06-20 14:33:01,116][starVLA.training.train_starvla][INFO] - Step 3375, Loss: {'train/loss': 0.43924835324287415, 'train/grad_norm_pre_clip': 10.28370189666748, 'timing/data': 0.0005482690175995231, 'timing/model': 1.4480210830224678, 'learning_rate/qwen_vl_interface': 9.562084946353212e-06, 'learning_rate/action_model': 9.562084946353213e-05, 'epoch': 6.24}) +[2026-06-20 14:33:05,721][starVLA.training.train_starvla][INFO] - Step 3376, Loss: {'train/loss': 0.3374059796333313, 'train/grad_norm_pre_clip': 9.68843936920166, 'timing/data': 0.0006095689022913575, 'timing/model': 1.178197773057036, 'learning_rate/qwen_vl_interface': 9.560806848749132e-06, 'learning_rate/action_model': 9.560806848749132e-05, 'epoch': 6.24}) +[2026-06-20 14:33:10,616][starVLA.training.train_starvla][INFO] - Step 3377, Loss: {'train/loss': 0.22322754561901093, 'train/grad_norm_pre_clip': 4.336058616638184, 'timing/data': 0.0005837189964950085, 'timing/model': 1.188595262938179, 'learning_rate/qwen_vl_interface': 9.559526979135186e-06, 'learning_rate/action_model': 9.559526979135187e-05, 'epoch': 6.25}) +[2026-06-20 14:33:15,517][starVLA.training.train_starvla][INFO] - Step 3378, Loss: {'train/loss': 0.18443991243839264, 'train/grad_norm_pre_clip': 5.314150333404541, 'timing/data': 0.0005460289539769292, 'timing/model': 1.2013700579991564, 'learning_rate/qwen_vl_interface': 9.558245338037486e-06, 'learning_rate/action_model': 9.558245338037486e-05, 'epoch': 6.25}) +[2026-06-20 14:33:20,361][starVLA.training.train_starvla][INFO] - Step 3379, Loss: {'train/loss': 0.3322743773460388, 'train/grad_norm_pre_clip': 6.768147945404053, 'timing/data': 0.0005235490389168262, 'timing/model': 1.1790033309953287, 'learning_rate/qwen_vl_interface': 9.556961925982864e-06, 'learning_rate/action_model': 9.556961925982863e-05, 'epoch': 6.25}) +[2026-06-20 14:33:25,258][starVLA.training.train_starvla][INFO] - Step 3380, Loss: {'train/loss': 0.25901302695274353, 'train/grad_norm_pre_clip': 5.523934364318848, 'timing/data': 0.0005653889384120703, 'timing/model': 1.4356132749235258, 'learning_rate/qwen_vl_interface': 9.555676743498882e-06, 'learning_rate/action_model': 9.555676743498881e-05, 'epoch': 6.25}) +[2026-06-20 14:33:30,168][starVLA.training.train_starvla][INFO] - Step 3381, Loss: {'train/loss': 0.14288032054901123, 'train/grad_norm_pre_clip': 4.850433826446533, 'timing/data': 0.0009147980017587543, 'timing/model': 1.4532645490253344, 'learning_rate/qwen_vl_interface': 9.55438979111383e-06, 'learning_rate/action_model': 9.554389791113831e-05, 'epoch': 6.25}) +[2026-06-20 14:33:34,897][starVLA.training.train_starvla][INFO] - Step 3382, Loss: {'train/loss': 0.2602565884590149, 'train/grad_norm_pre_clip': 4.190783977508545, 'timing/data': 0.0005669790552929044, 'timing/model': 1.1824519649380818, 'learning_rate/qwen_vl_interface': 9.553101069356728e-06, 'learning_rate/action_model': 9.553101069356727e-05, 'epoch': 6.25}) +[2026-06-20 14:33:39,972][starVLA.training.train_starvla][INFO] - Step 3383, Loss: {'train/loss': 0.3341367840766907, 'train/grad_norm_pre_clip': 6.307971477508545, 'timing/data': 0.0005527589237317443, 'timing/model': 1.2366177079966292, 'learning_rate/qwen_vl_interface': 9.551810578757316e-06, 'learning_rate/action_model': 9.551810578757315e-05, 'epoch': 6.26}) +[2026-06-20 14:33:44,864][starVLA.training.train_starvla][INFO] - Step 3384, Loss: {'train/loss': 0.16170509159564972, 'train/grad_norm_pre_clip': 3.032061815261841, 'timing/data': 0.0005261090118438005, 'timing/model': 1.218787633930333, 'learning_rate/qwen_vl_interface': 9.550518319846069e-06, 'learning_rate/action_model': 9.550518319846069e-05, 'epoch': 6.26}) +[2026-06-20 14:33:49,866][starVLA.training.train_starvla][INFO] - Step 3385, Loss: {'train/loss': 0.17144808173179626, 'train/grad_norm_pre_clip': 7.973244667053223, 'timing/data': 0.0005912280175834894, 'timing/model': 1.2288599429884925, 'learning_rate/qwen_vl_interface': 9.549224293154187e-06, 'learning_rate/action_model': 9.549224293154187e-05, 'epoch': 6.26}) +[2026-06-20 14:33:54,773][starVLA.training.train_starvla][INFO] - Step 3386, Loss: {'train/loss': 0.49180978536605835, 'train/grad_norm_pre_clip': 7.9353766441345215, 'timing/data': 0.00051029899623245, 'timing/model': 1.4420291119022295, 'learning_rate/qwen_vl_interface': 9.547928499213593e-06, 'learning_rate/action_model': 9.547928499213593e-05, 'epoch': 6.26}) +[2026-06-20 14:33:59,403][starVLA.training.train_starvla][INFO] - Step 3387, Loss: {'train/loss': 0.2772437334060669, 'train/grad_norm_pre_clip': 6.893413543701172, 'timing/data': 0.0005234790733084083, 'timing/model': 1.1765402480959892, 'learning_rate/qwen_vl_interface': 9.54663093855694e-06, 'learning_rate/action_model': 9.54663093855694e-05, 'epoch': 6.26}) +[2026-06-20 14:34:04,250][starVLA.training.train_starvla][INFO] - Step 3388, Loss: {'train/loss': 0.1981021612882614, 'train/grad_norm_pre_clip': 5.726367473602295, 'timing/data': 0.0006798390531912446, 'timing/model': 1.17592341999989, 'learning_rate/qwen_vl_interface': 9.545331611717607e-06, 'learning_rate/action_model': 9.545331611717608e-05, 'epoch': 6.27}) +[2026-06-20 14:34:09,208][starVLA.training.train_starvla][INFO] - Step 3389, Loss: {'train/loss': 0.4277828335762024, 'train/grad_norm_pre_clip': 10.421865463256836, 'timing/data': 0.0005303190555423498, 'timing/model': 1.178972965106368, 'learning_rate/qwen_vl_interface': 9.544030519229697e-06, 'learning_rate/action_model': 9.544030519229697e-05, 'epoch': 6.27}) +[2026-06-20 14:34:14,145][starVLA.training.train_starvla][INFO] - Step 3390, Loss: {'train/loss': 0.19741642475128174, 'train/grad_norm_pre_clip': 4.615640163421631, 'timing/data': 0.0005815789336338639, 'timing/model': 1.2602154129417613, 'learning_rate/qwen_vl_interface': 9.54272766162804e-06, 'learning_rate/action_model': 9.54272766162804e-05, 'epoch': 6.27}) +[2026-06-20 14:34:19,048][starVLA.training.train_starvla][INFO] - Step 3391, Loss: {'train/loss': 0.2460847795009613, 'train/grad_norm_pre_clip': 5.216846942901611, 'timing/data': 0.0006155089940875769, 'timing/model': 1.222660448984243, 'learning_rate/qwen_vl_interface': 9.54142303944819e-06, 'learning_rate/action_model': 9.541423039448191e-05, 'epoch': 6.27}) +[2026-06-20 14:34:23,927][starVLA.training.train_starvla][INFO] - Step 3392, Loss: {'train/loss': 0.19516578316688538, 'train/grad_norm_pre_clip': 5.752118110656738, 'timing/data': 0.0006965290522202849, 'timing/model': 1.4243093410041183, 'learning_rate/qwen_vl_interface': 9.540116653226433e-06, 'learning_rate/action_model': 9.540116653226431e-05, 'epoch': 6.27}) +[2026-06-20 14:34:28,481][starVLA.training.train_starvla][INFO] - Step 3393, Loss: {'train/loss': 0.1848689764738083, 'train/grad_norm_pre_clip': 6.212395191192627, 'timing/data': 0.0005301890196278691, 'timing/model': 1.1743739600060508, 'learning_rate/qwen_vl_interface': 9.53880850349977e-06, 'learning_rate/action_model': 9.538808503499768e-05, 'epoch': 6.27}) +[2026-06-20 14:34:33,366][starVLA.training.train_starvla][INFO] - Step 3394, Loss: {'train/loss': 0.2700224816799164, 'train/grad_norm_pre_clip': 6.1344146728515625, 'timing/data': 0.0005282689817249775, 'timing/model': 1.2257224660133943, 'learning_rate/qwen_vl_interface': 9.537498590805934e-06, 'learning_rate/action_model': 9.537498590805933e-05, 'epoch': 6.28}) +[2026-06-20 14:34:38,336][starVLA.training.train_starvla][INFO] - Step 3395, Loss: {'train/loss': 0.17837516963481903, 'train/grad_norm_pre_clip': 3.762704372406006, 'timing/data': 0.0005398490466177464, 'timing/model': 1.1826992169953883, 'learning_rate/qwen_vl_interface': 9.536186915683379e-06, 'learning_rate/action_model': 9.53618691568338e-05, 'epoch': 6.28}) +[2026-06-20 14:34:43,682][starVLA.training.train_starvla][INFO] - Step 3396, Loss: {'train/loss': 0.43648606538772583, 'train/grad_norm_pre_clip': 9.733945846557617, 'timing/data': 0.0005386790726333857, 'timing/model': 1.354985491023399, 'learning_rate/qwen_vl_interface': 9.534873478671289e-06, 'learning_rate/action_model': 9.534873478671288e-05, 'epoch': 6.28}) +[2026-06-20 14:34:48,875][starVLA.training.train_starvla][INFO] - Step 3397, Loss: {'train/loss': 0.17373311519622803, 'train/grad_norm_pre_clip': 4.477955341339111, 'timing/data': 0.0005379889626055956, 'timing/model': 1.2693792980862781, 'learning_rate/qwen_vl_interface': 9.533558280309563e-06, 'learning_rate/action_model': 9.533558280309563e-05, 'epoch': 6.28}) +[2026-06-20 14:34:54,071][starVLA.training.train_starvla][INFO] - Step 3398, Loss: {'train/loss': 0.332407146692276, 'train/grad_norm_pre_clip': 7.573073863983154, 'timing/data': 0.0005983590381219983, 'timing/model': 1.4715626299148425, 'learning_rate/qwen_vl_interface': 9.532241321138832e-06, 'learning_rate/action_model': 9.532241321138833e-05, 'epoch': 6.28}) +[2026-06-20 14:34:58,943][starVLA.training.train_starvla][INFO] - Step 3399, Loss: {'train/loss': 0.3759814500808716, 'train/grad_norm_pre_clip': 5.15875244140625, 'timing/data': 0.0005259389290586114, 'timing/model': 1.3655318199889734, 'learning_rate/qwen_vl_interface': 9.530922601700451e-06, 'learning_rate/action_model': 9.530922601700452e-05, 'epoch': 6.29}) +[2026-06-20 14:35:03,891][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 3400 +[2026-06-20 14:41:09,015][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 3400 +[2026-06-20 14:41:09,016][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 3400 +[2026-06-20 14:47:35,355][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 3400 +[2026-06-20 14:47:35,356][starVLA.training.train_starvla][INFO] - Step 3400, Loss: {'train/loss': 0.21670116484165192, 'train/grad_norm_pre_clip': 8.300854682922363, 'eval/loss': 0.42283061146736145, 'eval/latency_0/loss': 0.49275320768356323, 'eval/latency_1/loss': 0.6147928237915039, 'eval/latency_2/loss': 0.5544512271881104, 'eval/latency_3/loss': 0.2550557851791382, 'eval/latency_4/loss': 0.1771211177110672, 'eval/flappy/loss': 0.42283061146736145, 'eval/flappy/latency_0/loss': 0.49275320768356323, 'eval/flappy/latency_1/loss': 0.6147928237915039, 'eval/flappy/latency_2/loss': 0.5544512271881104, 'eval/flappy/latency_3/loss': 0.2550557851791382, 'eval/flappy/latency_4/loss': 0.1771211177110672, 'eval/flap_precision': 0.2876712328767123, 'eval/flap_recall': 0.07636363636363637, 'eval/flap_f1': 0.1206896551724138, 'eval/flap_rate_model': 0.020277777777777777, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.12068965517241381, 'eval/latency_0/flap_precision': 0.2876712328767123, 'eval/latency_0/flap_recall': 0.07636363636363637, 'eval/latency_0/flap_f1': 0.1206896551724138, 'eval/latency_0/flap_rate_model': 0.020277777777777777, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.12068965517241381, 'timing/data': 0.000533608952537179, 'timing/model': 1.2379204059252515, 'learning_rate/qwen_vl_interface': 9.52960212253649e-06, 'learning_rate/action_model': 9.529602122536491e-05, 'epoch': 6.29}) +[2026-06-20 14:47:35,357][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_3400_state +[2026-06-20 14:47:35,363][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-20 14:47:49,735][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_3400_state/model.safetensors +[2026-06-20 14:48:05,278][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_3400_state/optimizer.bin +[2026-06-20 14:48:05,279][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_3400_state/sampler.bin +[2026-06-20 14:48:05,279][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_3400_state/sampler_1.bin +[2026-06-20 14:48:05,280][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_3400_state/random_states_0.pkl +[2026-06-20 14:48:07,075][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3400 +[2026-06-20 14:48:07,075][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-20 14:48:07,089][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-20 14:48:11,974][starVLA.training.train_starvla][INFO] - Step 3401, Loss: {'train/loss': 0.14456862211227417, 'train/grad_norm_pre_clip': 8.88994026184082, 'timing/data': 0.0006380579434335232, 'timing/model': 1.4114027899922803, 'learning_rate/qwen_vl_interface': 9.528279884189754e-06, 'learning_rate/action_model': 9.528279884189754e-05, 'epoch': 6.29}) +[2026-06-20 14:48:17,076][starVLA.training.train_starvla][INFO] - Step 3402, Loss: {'train/loss': 0.45091313123703003, 'train/grad_norm_pre_clip': 15.274184226989746, 'timing/data': 0.0005075479857623577, 'timing/model': 1.41583457600791, 'learning_rate/qwen_vl_interface': 9.526955887203762e-06, 'learning_rate/action_model': 9.526955887203761e-05, 'epoch': 6.29}) +[2026-06-20 14:48:21,810][starVLA.training.train_starvla][INFO] - Step 3403, Loss: {'train/loss': 0.2728499174118042, 'train/grad_norm_pre_clip': 9.631449699401855, 'timing/data': 0.000562497996725142, 'timing/model': 1.1839947709813714, 'learning_rate/qwen_vl_interface': 9.525630132122758e-06, 'learning_rate/action_model': 9.525630132122758e-05, 'epoch': 6.29}) +[2026-06-20 14:48:26,719][starVLA.training.train_starvla][INFO] - Step 3404, Loss: {'train/loss': 0.35773974657058716, 'train/grad_norm_pre_clip': 13.751073837280273, 'timing/data': 0.0005557679105550051, 'timing/model': 1.1687139400746673, 'learning_rate/qwen_vl_interface': 9.524302619491712e-06, 'learning_rate/action_model': 9.524302619491712e-05, 'epoch': 6.3}) +[2026-06-20 14:48:31,631][starVLA.training.train_starvla][INFO] - Step 3405, Loss: {'train/loss': 0.20826944708824158, 'train/grad_norm_pre_clip': 8.022165298461914, 'timing/data': 0.0005071089835837483, 'timing/model': 1.218240445945412, 'learning_rate/qwen_vl_interface': 9.522973349856315e-06, 'learning_rate/action_model': 9.522973349856315e-05, 'epoch': 6.3}) +[2026-06-20 14:48:36,668][starVLA.training.train_starvla][INFO] - Step 3406, Loss: {'train/loss': 0.3688795268535614, 'train/grad_norm_pre_clip': 6.359481334686279, 'timing/data': 0.0005376989720389247, 'timing/model': 1.2326008160598576, 'learning_rate/qwen_vl_interface': 9.521642323762976e-06, 'learning_rate/action_model': 9.521642323762976e-05, 'epoch': 6.3}) +[2026-06-20 14:48:41,644][starVLA.training.train_starvla][INFO] - Step 3407, Loss: {'train/loss': 0.19796305894851685, 'train/grad_norm_pre_clip': 4.442756652832031, 'timing/data': 0.0005355179309844971, 'timing/model': 1.5544851660961285, 'learning_rate/qwen_vl_interface': 9.520309541758833e-06, 'learning_rate/action_model': 9.520309541758833e-05, 'epoch': 6.3}) +[2026-06-20 14:48:46,875][starVLA.training.train_starvla][INFO] - Step 3408, Loss: {'train/loss': 0.32645711302757263, 'train/grad_norm_pre_clip': 6.196913242340088, 'timing/data': 0.0005195790436118841, 'timing/model': 1.2635813880478963, 'learning_rate/qwen_vl_interface': 9.51897500439174e-06, 'learning_rate/action_model': 9.51897500439174e-05, 'epoch': 6.3}) +[2026-06-20 14:48:51,844][starVLA.training.train_starvla][INFO] - Step 3409, Loss: {'train/loss': 0.22231680154800415, 'train/grad_norm_pre_clip': 4.485314846038818, 'timing/data': 0.0006125579820945859, 'timing/model': 1.1711625640746206, 'learning_rate/qwen_vl_interface': 9.517638712210277e-06, 'learning_rate/action_model': 9.517638712210276e-05, 'epoch': 6.3}) +[2026-06-20 14:48:56,901][starVLA.training.train_starvla][INFO] - Step 3410, Loss: {'train/loss': 0.19912663102149963, 'train/grad_norm_pre_clip': 5.849781036376953, 'timing/data': 0.0005356980254873633, 'timing/model': 1.1747764829779044, 'learning_rate/qwen_vl_interface': 9.51630066576374e-06, 'learning_rate/action_model': 9.51630066576374e-05, 'epoch': 6.31}) +[2026-06-20 14:49:01,735][starVLA.training.train_starvla][INFO] - Step 3411, Loss: {'train/loss': 0.3533095419406891, 'train/grad_norm_pre_clip': 6.448474407196045, 'timing/data': 0.0005357789341360331, 'timing/model': 1.2234798909630626, 'learning_rate/qwen_vl_interface': 9.514960865602154e-06, 'learning_rate/action_model': 9.514960865602153e-05, 'epoch': 6.31}) +[2026-06-20 14:49:06,884][starVLA.training.train_starvla][INFO] - Step 3412, Loss: {'train/loss': 0.3491267263889313, 'train/grad_norm_pre_clip': 7.051959991455078, 'timing/data': 0.0005318480543792248, 'timing/model': 1.1842579650692642, 'learning_rate/qwen_vl_interface': 9.513619312276259e-06, 'learning_rate/action_model': 9.513619312276258e-05, 'epoch': 6.31}) +[2026-06-20 14:49:12,183][starVLA.training.train_starvla][INFO] - Step 3413, Loss: {'train/loss': 0.3430456519126892, 'train/grad_norm_pre_clip': 7.031896591186523, 'timing/data': 0.0005322289653122425, 'timing/model': 1.5158518749522045, 'learning_rate/qwen_vl_interface': 9.512276006337514e-06, 'learning_rate/action_model': 9.512276006337513e-05, 'epoch': 6.31}) +[2026-06-20 14:49:17,084][starVLA.training.train_starvla][INFO] - Step 3414, Loss: {'train/loss': 0.31579136848449707, 'train/grad_norm_pre_clip': 5.041324615478516, 'timing/data': 0.0005413980688899755, 'timing/model': 1.3785493320319802, 'learning_rate/qwen_vl_interface': 9.510930948338102e-06, 'learning_rate/action_model': 9.510930948338102e-05, 'epoch': 6.31}) +[2026-06-20 14:49:22,115][starVLA.training.train_starvla][INFO] - Step 3415, Loss: {'train/loss': 0.23450422286987305, 'train/grad_norm_pre_clip': 6.2189507484436035, 'timing/data': 0.0005392789607867599, 'timing/model': 1.1920019490644336, 'learning_rate/qwen_vl_interface': 9.509584138830929e-06, 'learning_rate/action_model': 9.509584138830928e-05, 'epoch': 6.32}) +[2026-06-20 14:49:27,227][starVLA.training.train_starvla][INFO] - Step 3416, Loss: {'train/loss': 0.3507675528526306, 'train/grad_norm_pre_clip': 6.333509922027588, 'timing/data': 0.0005469090538099408, 'timing/model': 1.3191119359107688, 'learning_rate/qwen_vl_interface': 9.508235578369613e-06, 'learning_rate/action_model': 9.508235578369612e-05, 'epoch': 6.32}) +[2026-06-20 14:49:32,375][starVLA.training.train_starvla][INFO] - Step 3417, Loss: {'train/loss': 0.3007163405418396, 'train/grad_norm_pre_clip': 8.78411865234375, 'timing/data': 0.0062580929370597005, 'timing/model': 1.2987118599703535, 'learning_rate/qwen_vl_interface': 9.506885267508498e-06, 'learning_rate/action_model': 9.506885267508498e-05, 'epoch': 6.32}) +[2026-06-20 14:49:37,678][starVLA.training.train_starvla][INFO] - Step 3418, Loss: {'train/loss': 0.2684513032436371, 'train/grad_norm_pre_clip': 7.918431282043457, 'timing/data': 0.008847285993397236, 'timing/model': 1.537029659957625, 'learning_rate/qwen_vl_interface': 9.505533206802648e-06, 'learning_rate/action_model': 9.505533206802647e-05, 'epoch': 6.32}) +[2026-06-20 14:49:43,064][starVLA.training.train_starvla][INFO] - Step 3419, Loss: {'train/loss': 0.2698175311088562, 'train/grad_norm_pre_clip': 7.104245185852051, 'timing/data': 0.007500559906475246, 'timing/model': 1.5479534569894895, 'learning_rate/qwen_vl_interface': 9.50417939680784e-06, 'learning_rate/action_model': 9.504179396807839e-05, 'epoch': 6.32}) +[2026-06-20 14:49:48,003][starVLA.training.train_starvla][INFO] - Step 3420, Loss: {'train/loss': 0.17995931208133698, 'train/grad_norm_pre_clip': 4.303799152374268, 'timing/data': 0.005995613988488913, 'timing/model': 1.1756256450898945, 'learning_rate/qwen_vl_interface': 9.502823838080578e-06, 'learning_rate/action_model': 9.502823838080578e-05, 'epoch': 6.32}) +[2026-06-20 14:49:52,987][starVLA.training.train_starvla][INFO] - Step 3421, Loss: {'train/loss': 0.2147970348596573, 'train/grad_norm_pre_clip': 4.181838035583496, 'timing/data': 0.005752643919549882, 'timing/model': 1.2011489340802655, 'learning_rate/qwen_vl_interface': 9.501466531178079e-06, 'learning_rate/action_model': 9.501466531178079e-05, 'epoch': 6.33}) +[2026-06-20 14:49:58,151][starVLA.training.train_starvla][INFO] - Step 3422, Loss: {'train/loss': 0.24783837795257568, 'train/grad_norm_pre_clip': 7.213555812835693, 'timing/data': 0.0060915740905329585, 'timing/model': 1.2759787009563297, 'learning_rate/qwen_vl_interface': 9.50010747665828e-06, 'learning_rate/action_model': 9.500107476658281e-05, 'epoch': 6.33}) +[2026-06-20 14:50:03,083][starVLA.training.train_starvla][INFO] - Step 3423, Loss: {'train/loss': 0.3503386080265045, 'train/grad_norm_pre_clip': 5.76378870010376, 'timing/data': 0.0060316938906908035, 'timing/model': 1.1810063049197197, 'learning_rate/qwen_vl_interface': 9.498746675079842e-06, 'learning_rate/action_model': 9.49874667507984e-05, 'epoch': 6.33}) +[2026-06-20 14:50:08,142][starVLA.training.train_starvla][INFO] - Step 3424, Loss: {'train/loss': 0.27000686526298523, 'train/grad_norm_pre_clip': 6.8752360343933105, 'timing/data': 0.0061997330049052835, 'timing/model': 1.2215073759434745, 'learning_rate/qwen_vl_interface': 9.49738412700213e-06, 'learning_rate/action_model': 9.49738412700213e-05, 'epoch': 6.33}) +[2026-06-20 14:50:13,093][starVLA.training.train_starvla][INFO] - Step 3425, Loss: {'train/loss': 0.2524598240852356, 'train/grad_norm_pre_clip': 5.52160120010376, 'timing/data': 0.005965774995274842, 'timing/model': 1.4451701119542122, 'learning_rate/qwen_vl_interface': 9.496019832985243e-06, 'learning_rate/action_model': 9.496019832985243e-05, 'epoch': 6.33}) +[2026-06-20 14:50:17,789][starVLA.training.train_starvla][INFO] - Step 3426, Loss: {'train/loss': 0.3220329284667969, 'train/grad_norm_pre_clip': 6.654609203338623, 'timing/data': 0.006317492923699319, 'timing/model': 1.1861997679807246, 'learning_rate/qwen_vl_interface': 9.49465379358999e-06, 'learning_rate/action_model': 9.49465379358999e-05, 'epoch': 6.34}) +[2026-06-20 14:50:22,739][starVLA.training.train_starvla][INFO] - Step 3427, Loss: {'train/loss': 0.24569547176361084, 'train/grad_norm_pre_clip': 15.7769193649292, 'timing/data': 0.006247034063562751, 'timing/model': 1.2470843020128086, 'learning_rate/qwen_vl_interface': 9.493286009377896e-06, 'learning_rate/action_model': 9.493286009377896e-05, 'epoch': 6.34}) +[2026-06-20 14:50:27,855][starVLA.training.train_starvla][INFO] - Step 3428, Loss: {'train/loss': 0.17180009186267853, 'train/grad_norm_pre_clip': 4.8125505447387695, 'timing/data': 0.006916610989719629, 'timing/model': 1.2385848939884454, 'learning_rate/qwen_vl_interface': 9.491916480911208e-06, 'learning_rate/action_model': 9.491916480911207e-05, 'epoch': 6.34}) +[2026-06-20 14:50:32,873][starVLA.training.train_starvla][INFO] - Step 3429, Loss: {'train/loss': 0.11945110559463501, 'train/grad_norm_pre_clip': 3.254976511001587, 'timing/data': 0.005946954013779759, 'timing/model': 1.2812240190105513, 'learning_rate/qwen_vl_interface': 9.490545208752884e-06, 'learning_rate/action_model': 9.490545208752883e-05, 'epoch': 6.34}) +[2026-06-20 14:50:37,877][starVLA.training.train_starvla][INFO] - Step 3430, Loss: {'train/loss': 0.2949806749820709, 'train/grad_norm_pre_clip': 7.905581474304199, 'timing/data': 0.006294882972724736, 'timing/model': 1.1974911720026284, 'learning_rate/qwen_vl_interface': 9.489172193466603e-06, 'learning_rate/action_model': 9.489172193466603e-05, 'epoch': 6.34}) +[2026-06-20 14:50:43,033][starVLA.training.train_starvla][INFO] - Step 3431, Loss: {'train/loss': 0.2111004739999771, 'train/grad_norm_pre_clip': 8.518927574157715, 'timing/data': 0.007497879909351468, 'timing/model': 1.6505347680067644, 'learning_rate/qwen_vl_interface': 9.487797435616762e-06, 'learning_rate/action_model': 9.487797435616761e-05, 'epoch': 6.35}) +[2026-06-20 14:50:48,165][starVLA.training.train_starvla][INFO] - Step 3432, Loss: {'train/loss': 0.19912569224834442, 'train/grad_norm_pre_clip': 6.898385047912598, 'timing/data': 0.006562761962413788, 'timing/model': 1.4776319930097088, 'learning_rate/qwen_vl_interface': 9.486420935768472e-06, 'learning_rate/action_model': 9.486420935768472e-05, 'epoch': 6.35}) +[2026-06-20 14:50:52,912][starVLA.training.train_starvla][INFO] - Step 3433, Loss: {'train/loss': 0.21341858804225922, 'train/grad_norm_pre_clip': 6.996338367462158, 'timing/data': 0.00683902099262923, 'timing/model': 1.231247907038778, 'learning_rate/qwen_vl_interface': 9.485042694487558e-06, 'learning_rate/action_model': 9.485042694487559e-05, 'epoch': 6.35}) +[2026-06-20 14:50:57,973][starVLA.training.train_starvla][INFO] - Step 3434, Loss: {'train/loss': 0.2022755742073059, 'train/grad_norm_pre_clip': 9.024794578552246, 'timing/data': 0.006259093061089516, 'timing/model': 1.177242291974835, 'learning_rate/qwen_vl_interface': 9.483662712340565e-06, 'learning_rate/action_model': 9.483662712340564e-05, 'epoch': 6.35}) +[2026-06-20 14:51:03,084][starVLA.training.train_starvla][INFO] - Step 3435, Loss: {'train/loss': 0.29750779271125793, 'train/grad_norm_pre_clip': 6.891419410705566, 'timing/data': 0.006312972982414067, 'timing/model': 1.1952356600668281, 'learning_rate/qwen_vl_interface': 9.482280989894748e-06, 'learning_rate/action_model': 9.482280989894748e-05, 'epoch': 6.35}) +[2026-06-20 14:51:08,076][starVLA.training.train_starvla][INFO] - Step 3436, Loss: {'train/loss': 0.14848016202449799, 'train/grad_norm_pre_clip': 4.842453956604004, 'timing/data': 0.007741688983514905, 'timing/model': 1.22292547498364, 'learning_rate/qwen_vl_interface': 9.480897527718083e-06, 'learning_rate/action_model': 9.480897527718084e-05, 'epoch': 6.35}) +[2026-06-20 14:51:12,872][starVLA.training.train_starvla][INFO] - Step 3437, Loss: {'train/loss': 0.2603996992111206, 'train/grad_norm_pre_clip': 7.366706371307373, 'timing/data': 0.006444182945415378, 'timing/model': 1.184276978019625, 'learning_rate/qwen_vl_interface': 9.479512326379263e-06, 'learning_rate/action_model': 9.479512326379264e-05, 'epoch': 6.36}) +[2026-06-20 14:51:17,848][starVLA.training.train_starvla][INFO] - Step 3438, Loss: {'train/loss': 0.506151020526886, 'train/grad_norm_pre_clip': 9.12466049194336, 'timing/data': 0.006015223916620016, 'timing/model': 1.3981911879964173, 'learning_rate/qwen_vl_interface': 9.478125386447686e-06, 'learning_rate/action_model': 9.478125386447686e-05, 'epoch': 6.36}) +[2026-06-20 14:51:22,548][starVLA.training.train_starvla][INFO] - Step 3439, Loss: {'train/loss': 0.32008740305900574, 'train/grad_norm_pre_clip': 8.625502586364746, 'timing/data': 0.005516765057109296, 'timing/model': 1.24507625203114, 'learning_rate/qwen_vl_interface': 9.476736708493472e-06, 'learning_rate/action_model': 9.476736708493472e-05, 'epoch': 6.36}) +[2026-06-20 14:51:31,169][starVLA.training.train_starvla][INFO] - Step 3440, Loss: {'train/loss': 0.20952361822128296, 'train/grad_norm_pre_clip': 9.310546875, 'timing/data': 0.005642445059493184, 'timing/model': 1.3386491859564558, 'learning_rate/qwen_vl_interface': 9.475346293087457e-06, 'learning_rate/action_model': 9.475346293087458e-05, 'epoch': 6.36}) +[2026-06-20 14:51:36,539][starVLA.training.train_starvla][INFO] - Step 3441, Loss: {'train/loss': 0.30474215745925903, 'train/grad_norm_pre_clip': 6.880227565765381, 'timing/data': 0.0058433940866962075, 'timing/model': 1.279759394004941, 'learning_rate/qwen_vl_interface': 9.473954140801184e-06, 'learning_rate/action_model': 9.473954140801185e-05, 'epoch': 6.36}) +[2026-06-20 14:51:42,093][starVLA.training.train_starvla][INFO] - Step 3442, Loss: {'train/loss': 0.17311853170394897, 'train/grad_norm_pre_clip': 19.13472557067871, 'timing/data': 0.006481813034042716, 'timing/model': 1.4842007990228012, 'learning_rate/qwen_vl_interface': 9.472560252206918e-06, 'learning_rate/action_model': 9.472560252206918e-05, 'epoch': 6.37}) +[2026-06-20 14:51:47,363][starVLA.training.train_starvla][INFO] - Step 3443, Loss: {'train/loss': 0.21838170289993286, 'train/grad_norm_pre_clip': 7.078705310821533, 'timing/data': 0.006107563036493957, 'timing/model': 1.4498509489931166, 'learning_rate/qwen_vl_interface': 9.471164627877633e-06, 'learning_rate/action_model': 9.471164627877633e-05, 'epoch': 6.37}) +[2026-06-20 14:51:52,240][starVLA.training.train_starvla][INFO] - Step 3444, Loss: {'train/loss': 0.1843482106924057, 'train/grad_norm_pre_clip': 5.203549861907959, 'timing/data': 0.006154422997497022, 'timing/model': 1.401458490989171, 'learning_rate/qwen_vl_interface': 9.469767268387016e-06, 'learning_rate/action_model': 9.469767268387016e-05, 'epoch': 6.37}) +[2026-06-20 14:51:57,996][starVLA.training.train_starvla][INFO] - Step 3445, Loss: {'train/loss': 0.16494964063167572, 'train/grad_norm_pre_clip': 7.315157413482666, 'timing/data': 0.006026812945492566, 'timing/model': 1.3851957520237193, 'learning_rate/qwen_vl_interface': 9.468368174309468e-06, 'learning_rate/action_model': 9.468368174309469e-05, 'epoch': 6.37}) +[2026-06-20 14:52:03,320][starVLA.training.train_starvla][INFO] - Step 3446, Loss: {'train/loss': 0.10001388937234879, 'train/grad_norm_pre_clip': 5.850983142852783, 'timing/data': 0.006071503972634673, 'timing/model': 1.3036812409991398, 'learning_rate/qwen_vl_interface': 9.466967346220107e-06, 'learning_rate/action_model': 9.466967346220107e-05, 'epoch': 6.37}) +[2026-06-20 14:52:08,175][starVLA.training.train_starvla][INFO] - Step 3447, Loss: {'train/loss': 0.11689461022615433, 'train/grad_norm_pre_clip': 5.272213459014893, 'timing/data': 0.007287890068255365, 'timing/model': 1.190201269928366, 'learning_rate/qwen_vl_interface': 9.465564784694759e-06, 'learning_rate/action_model': 9.465564784694759e-05, 'epoch': 6.37}) +[2026-06-20 14:52:13,294][starVLA.training.train_starvla][INFO] - Step 3448, Loss: {'train/loss': 0.13978572189807892, 'train/grad_norm_pre_clip': 3.7070870399475098, 'timing/data': 0.006745991064235568, 'timing/model': 1.3961482379818335, 'learning_rate/qwen_vl_interface': 9.464160490309962e-06, 'learning_rate/action_model': 9.464160490309962e-05, 'epoch': 6.38}) +[2026-06-20 14:52:18,450][starVLA.training.train_starvla][INFO] - Step 3449, Loss: {'train/loss': 0.18134091794490814, 'train/grad_norm_pre_clip': 6.792766571044922, 'timing/data': 0.006072254036553204, 'timing/model': 1.5784418320981786, 'learning_rate/qwen_vl_interface': 9.462754463642972e-06, 'learning_rate/action_model': 9.46275446364297e-05, 'epoch': 6.38}) +[2026-06-20 14:52:23,143][starVLA.training.train_starvla][INFO] - Step 3450, Loss: {'train/loss': 0.30661606788635254, 'train/grad_norm_pre_clip': 10.589229583740234, 'timing/data': 0.006328093004412949, 'timing/model': 1.1884286110289395, 'learning_rate/qwen_vl_interface': 9.461346705271753e-06, 'learning_rate/action_model': 9.461346705271752e-05, 'epoch': 6.38}) +[2026-06-20 14:52:28,908][starVLA.training.train_starvla][INFO] - Step 3451, Loss: {'train/loss': 0.22092926502227783, 'train/grad_norm_pre_clip': 6.456974506378174, 'timing/data': 0.00854763702955097, 'timing/model': 1.4256270540645346, 'learning_rate/qwen_vl_interface': 9.459937215774976e-06, 'learning_rate/action_model': 9.459937215774976e-05, 'epoch': 6.38}) +[2026-06-20 14:52:33,812][starVLA.training.train_starvla][INFO] - Step 3452, Loss: {'train/loss': 0.09827043116092682, 'train/grad_norm_pre_clip': 6.327447891235352, 'timing/data': 0.006310073076747358, 'timing/model': 1.1954050189815462, 'learning_rate/qwen_vl_interface': 9.458525995732035e-06, 'learning_rate/action_model': 9.458525995732035e-05, 'epoch': 6.38}) +[2026-06-20 14:52:38,858][starVLA.training.train_starvla][INFO] - Step 3453, Loss: {'train/loss': 0.08937720954418182, 'train/grad_norm_pre_clip': 4.807157516479492, 'timing/data': 0.007106291013769805, 'timing/model': 1.2448955529835075, 'learning_rate/qwen_vl_interface': 9.457113045723027e-06, 'learning_rate/action_model': 9.457113045723026e-05, 'epoch': 6.39}) +[2026-06-20 14:52:43,837][starVLA.training.train_starvla][INFO] - Step 3454, Loss: {'train/loss': 0.4275033473968506, 'train/grad_norm_pre_clip': 8.022622108459473, 'timing/data': 0.008744605933316052, 'timing/model': 1.1943278999533504, 'learning_rate/qwen_vl_interface': 9.455698366328762e-06, 'learning_rate/action_model': 9.455698366328762e-05, 'epoch': 6.39}) +[2026-06-20 14:52:48,704][starVLA.training.train_starvla][INFO] - Step 3455, Loss: {'train/loss': 0.29658082127571106, 'train/grad_norm_pre_clip': 12.04468822479248, 'timing/data': 0.0062359030125662684, 'timing/model': 1.4463547120103613, 'learning_rate/qwen_vl_interface': 9.454281958130761e-06, 'learning_rate/action_model': 9.454281958130761e-05, 'epoch': 6.39}) +[2026-06-20 14:52:53,609][starVLA.training.train_starvla][INFO] - Step 3456, Loss: {'train/loss': 0.16248519718647003, 'train/grad_norm_pre_clip': 11.479574203491211, 'timing/data': 0.00612423405982554, 'timing/model': 1.1922835630830377, 'learning_rate/qwen_vl_interface': 9.452863821711257e-06, 'learning_rate/action_model': 9.452863821711258e-05, 'epoch': 6.39}) +[2026-06-20 14:52:58,830][starVLA.training.train_starvla][INFO] - Step 3457, Loss: {'train/loss': 0.4243856370449066, 'train/grad_norm_pre_clip': 6.762670040130615, 'timing/data': 0.006414092960767448, 'timing/model': 1.2981956340372562, 'learning_rate/qwen_vl_interface': 9.451443957653195e-06, 'learning_rate/action_model': 9.451443957653195e-05, 'epoch': 6.39}) +[2026-06-20 14:53:03,879][starVLA.training.train_starvla][INFO] - Step 3458, Loss: {'train/loss': 0.22262060642242432, 'train/grad_norm_pre_clip': 5.908664703369141, 'timing/data': 0.00557761499658227, 'timing/model': 1.2638947550440207, 'learning_rate/qwen_vl_interface': 9.450022366540221e-06, 'learning_rate/action_model': 9.45002236654022e-05, 'epoch': 6.4}) +[2026-06-20 14:53:08,711][starVLA.training.train_starvla][INFO] - Step 3459, Loss: {'train/loss': 0.34274813532829285, 'train/grad_norm_pre_clip': 10.821563720703125, 'timing/data': 0.006230072001926601, 'timing/model': 1.253495002980344, 'learning_rate/qwen_vl_interface': 9.448599048956702e-06, 'learning_rate/action_model': 9.448599048956702e-05, 'epoch': 6.4}) +[2026-06-20 14:53:13,736][starVLA.training.train_starvla][INFO] - Step 3460, Loss: {'train/loss': 0.3849234879016876, 'train/grad_norm_pre_clip': 19.04593849182129, 'timing/data': 0.005656814086250961, 'timing/model': 1.3082560820039362, 'learning_rate/qwen_vl_interface': 9.447174005487708e-06, 'learning_rate/action_model': 9.447174005487708e-05, 'epoch': 6.4}) +[2026-06-20 14:53:19,128][starVLA.training.train_starvla][INFO] - Step 3461, Loss: {'train/loss': 0.2774921655654907, 'train/grad_norm_pre_clip': 11.170002937316895, 'timing/data': 0.006014705984853208, 'timing/model': 1.8614607310155407, 'learning_rate/qwen_vl_interface': 9.445747236719022e-06, 'learning_rate/action_model': 9.445747236719022e-05, 'epoch': 6.4}) +[2026-06-20 14:53:24,004][starVLA.training.train_starvla][INFO] - Step 3462, Loss: {'train/loss': 0.22717520594596863, 'train/grad_norm_pre_clip': 6.691308975219727, 'timing/data': 0.0063126360764727, 'timing/model': 1.320134405978024, 'learning_rate/qwen_vl_interface': 9.444318743237135e-06, 'learning_rate/action_model': 9.444318743237134e-05, 'epoch': 6.4}) +[2026-06-20 14:53:29,323][starVLA.training.train_starvla][INFO] - Step 3463, Loss: {'train/loss': 0.22863605618476868, 'train/grad_norm_pre_clip': 7.777898788452148, 'timing/data': 0.006224256008863449, 'timing/model': 1.205151841044426, 'learning_rate/qwen_vl_interface': 9.442888525629244e-06, 'learning_rate/action_model': 9.442888525629245e-05, 'epoch': 6.4}) +[2026-06-20 14:53:34,601][starVLA.training.train_starvla][INFO] - Step 3464, Loss: {'train/loss': 0.35032737255096436, 'train/grad_norm_pre_clip': 9.918401718139648, 'timing/data': 0.006291256053373218, 'timing/model': 1.3642765309195966, 'learning_rate/qwen_vl_interface': 9.441456584483261e-06, 'learning_rate/action_model': 9.441456584483261e-05, 'epoch': 6.41}) +[2026-06-20 14:53:39,612][starVLA.training.train_starvla][INFO] - Step 3465, Loss: {'train/loss': 0.10410171747207642, 'train/grad_norm_pre_clip': 6.311360836029053, 'timing/data': 0.005619397037662566, 'timing/model': 1.1781660809647292, 'learning_rate/qwen_vl_interface': 9.440022920387801e-06, 'learning_rate/action_model': 9.440022920387801e-05, 'epoch': 6.41}) +[2026-06-20 14:53:44,814][starVLA.training.train_starvla][INFO] - Step 3466, Loss: {'train/loss': 0.10299687832593918, 'train/grad_norm_pre_clip': 7.216930389404297, 'timing/data': 0.006774875917471945, 'timing/model': 1.2538974840426818, 'learning_rate/qwen_vl_interface': 9.438587533932191e-06, 'learning_rate/action_model': 9.43858753393219e-05, 'epoch': 6.41}) +[2026-06-20 14:53:50,129][starVLA.training.train_starvla][INFO] - Step 3467, Loss: {'train/loss': 0.33675840497016907, 'train/grad_norm_pre_clip': 9.735884666442871, 'timing/data': 0.0068296759855002165, 'timing/model': 1.5327803599648178, 'learning_rate/qwen_vl_interface': 9.43715042570646e-06, 'learning_rate/action_model': 9.43715042570646e-05, 'epoch': 6.41}) +[2026-06-20 14:53:55,433][starVLA.training.train_starvla][INFO] - Step 3468, Loss: {'train/loss': 0.2967986464500427, 'train/grad_norm_pre_clip': 8.151650428771973, 'timing/data': 0.006635425961576402, 'timing/model': 1.338755492004566, 'learning_rate/qwen_vl_interface': 9.435711596301353e-06, 'learning_rate/action_model': 9.435711596301352e-05, 'epoch': 6.41}) +[2026-06-20 14:54:00,515][starVLA.training.train_starvla][INFO] - Step 3469, Loss: {'train/loss': 0.26481127738952637, 'train/grad_norm_pre_clip': 10.766650199890137, 'timing/data': 0.005950026912614703, 'timing/model': 1.2461425489746034, 'learning_rate/qwen_vl_interface': 9.434271046308317e-06, 'learning_rate/action_model': 9.434271046308317e-05, 'epoch': 6.42}) +[2026-06-20 14:54:05,632][starVLA.training.train_starvla][INFO] - Step 3470, Loss: {'train/loss': 0.08580360561609268, 'train/grad_norm_pre_clip': 4.350168228149414, 'timing/data': 0.007498934981413186, 'timing/model': 1.2443552169715986, 'learning_rate/qwen_vl_interface': 9.432828776319507e-06, 'learning_rate/action_model': 9.432828776319505e-05, 'epoch': 6.42}) +[2026-06-20 14:54:11,220][starVLA.training.train_starvla][INFO] - Step 3471, Loss: {'train/loss': 0.08589053153991699, 'train/grad_norm_pre_clip': 4.470390319824219, 'timing/data': 0.0059951760340482, 'timing/model': 1.565574851934798, 'learning_rate/qwen_vl_interface': 9.431384786927785e-06, 'learning_rate/action_model': 9.431384786927786e-05, 'epoch': 6.42}) +[2026-06-20 14:54:16,438][starVLA.training.train_starvla][INFO] - Step 3472, Loss: {'train/loss': 0.2455257773399353, 'train/grad_norm_pre_clip': 15.862956047058105, 'timing/data': 0.006209545070305467, 'timing/model': 1.333796309074387, 'learning_rate/qwen_vl_interface': 9.429939078726723e-06, 'learning_rate/action_model': 9.429939078726722e-05, 'epoch': 6.42}) +[2026-06-20 14:54:21,450][starVLA.training.train_starvla][INFO] - Step 3473, Loss: {'train/loss': 0.3702321946620941, 'train/grad_norm_pre_clip': 9.564416885375977, 'timing/data': 0.00616146600805223, 'timing/model': 1.3969956699293107, 'learning_rate/qwen_vl_interface': 9.428491652310598e-06, 'learning_rate/action_model': 9.428491652310597e-05, 'epoch': 6.42}) +[2026-06-20 14:54:26,513][starVLA.training.train_starvla][INFO] - Step 3474, Loss: {'train/loss': 0.1630544811487198, 'train/grad_norm_pre_clip': 5.322054862976074, 'timing/data': 0.005557555938139558, 'timing/model': 1.2890899860067293, 'learning_rate/qwen_vl_interface': 9.427042508274386e-06, 'learning_rate/action_model': 9.427042508274386e-05, 'epoch': 6.42}) +[2026-06-20 14:54:31,672][starVLA.training.train_starvla][INFO] - Step 3475, Loss: {'train/loss': 0.10414590686559677, 'train/grad_norm_pre_clip': 3.009639263153076, 'timing/data': 0.005562486010603607, 'timing/model': 1.2483460549265146, 'learning_rate/qwen_vl_interface': 9.425591647213783e-06, 'learning_rate/action_model': 9.425591647213782e-05, 'epoch': 6.43}) +[2026-06-20 14:54:36,626][starVLA.training.train_starvla][INFO] - Step 3476, Loss: {'train/loss': 0.24069708585739136, 'train/grad_norm_pre_clip': 8.211150169372559, 'timing/data': 0.0005435399943962693, 'timing/model': 1.2461736859986559, 'learning_rate/qwen_vl_interface': 9.424139069725177e-06, 'learning_rate/action_model': 9.424139069725178e-05, 'epoch': 6.43}) +[2026-06-20 14:54:41,539][starVLA.training.train_starvla][INFO] - Step 3477, Loss: {'train/loss': 0.27169016003608704, 'train/grad_norm_pre_clip': 8.452743530273438, 'timing/data': 0.0008739989716559649, 'timing/model': 1.1854431029642, 'learning_rate/qwen_vl_interface': 9.422684776405673e-06, 'learning_rate/action_model': 9.422684776405672e-05, 'epoch': 6.43}) +[2026-06-20 14:54:46,653][starVLA.training.train_starvla][INFO] - Step 3478, Loss: {'train/loss': 0.2551262080669403, 'train/grad_norm_pre_clip': 9.017059326171875, 'timing/data': 0.000541210058145225, 'timing/model': 1.441131537896581, 'learning_rate/qwen_vl_interface': 9.421228767853072e-06, 'learning_rate/action_model': 9.421228767853072e-05, 'epoch': 6.43}) +[2026-06-20 14:54:51,289][starVLA.training.train_starvla][INFO] - Step 3479, Loss: {'train/loss': 0.1526976227760315, 'train/grad_norm_pre_clip': 8.730827331542969, 'timing/data': 0.0005480590043589473, 'timing/model': 1.1966473960783333, 'learning_rate/qwen_vl_interface': 9.419771044665887e-06, 'learning_rate/action_model': 9.419771044665887e-05, 'epoch': 6.43}) +[2026-06-20 14:54:56,521][starVLA.training.train_starvla][INFO] - Step 3480, Loss: {'train/loss': 0.05558501183986664, 'train/grad_norm_pre_clip': 2.8857860565185547, 'timing/data': 0.0005468199960887432, 'timing/model': 1.2996073809918016, 'learning_rate/qwen_vl_interface': 9.41831160744333e-06, 'learning_rate/action_model': 9.41831160744333e-05, 'epoch': 6.44}) +[2026-06-20 14:55:01,483][starVLA.training.train_starvla][INFO] - Step 3481, Loss: {'train/loss': 0.20721150934696198, 'train/grad_norm_pre_clip': 10.749894142150879, 'timing/data': 0.000554880010895431, 'timing/model': 1.1962067689746618, 'learning_rate/qwen_vl_interface': 9.416850456785322e-06, 'learning_rate/action_model': 9.416850456785321e-05, 'epoch': 6.44}) +[2026-06-20 14:55:06,352][starVLA.training.train_starvla][INFO] - Step 3482, Loss: {'train/loss': 0.1906047761440277, 'train/grad_norm_pre_clip': 7.621182441711426, 'timing/data': 0.0006502789910882711, 'timing/model': 1.2326173610053957, 'learning_rate/qwen_vl_interface': 9.415387593292489e-06, 'learning_rate/action_model': 9.415387593292488e-05, 'epoch': 6.44}) +[2026-06-20 14:55:11,247][starVLA.training.train_starvla][INFO] - Step 3483, Loss: {'train/loss': 0.19298399984836578, 'train/grad_norm_pre_clip': 5.9060893058776855, 'timing/data': 0.000660030054859817, 'timing/model': 1.1804399869870394, 'learning_rate/qwen_vl_interface': 9.413923017566155e-06, 'learning_rate/action_model': 9.413923017566155e-05, 'epoch': 6.44}) +[2026-06-20 14:55:16,152][starVLA.training.train_starvla][INFO] - Step 3484, Loss: {'train/loss': 0.3133313059806824, 'train/grad_norm_pre_clip': 9.19665241241455, 'timing/data': 0.0006160400807857513, 'timing/model': 1.5102599740494043, 'learning_rate/qwen_vl_interface': 9.412456730208352e-06, 'learning_rate/action_model': 9.412456730208352e-05, 'epoch': 6.44}) +[2026-06-20 14:55:20,797][starVLA.training.train_starvla][INFO] - Step 3485, Loss: {'train/loss': 0.3185654878616333, 'train/grad_norm_pre_clip': 9.392183303833008, 'timing/data': 0.0005485799629241228, 'timing/model': 1.2469455830287188, 'learning_rate/qwen_vl_interface': 9.41098873182182e-06, 'learning_rate/action_model': 9.41098873182182e-05, 'epoch': 6.45}) +[2026-06-20 14:55:25,796][starVLA.training.train_starvla][INFO] - Step 3486, Loss: {'train/loss': 0.3210577964782715, 'train/grad_norm_pre_clip': 11.639309883117676, 'timing/data': 0.0008865090785548091, 'timing/model': 1.201599286054261, 'learning_rate/qwen_vl_interface': 9.409519023009993e-06, 'learning_rate/action_model': 9.409519023009993e-05, 'epoch': 6.45}) +[2026-06-20 14:55:30,677][starVLA.training.train_starvla][INFO] - Step 3487, Loss: {'train/loss': 0.24104177951812744, 'train/grad_norm_pre_clip': 10.556137084960938, 'timing/data': 0.0005352389998733997, 'timing/model': 1.19376099598594, 'learning_rate/qwen_vl_interface': 9.408047604377014e-06, 'learning_rate/action_model': 9.408047604377014e-05, 'epoch': 6.45}) +[2026-06-20 14:55:35,652][starVLA.training.train_starvla][INFO] - Step 3488, Loss: {'train/loss': 0.15468500554561615, 'train/grad_norm_pre_clip': 9.630354881286621, 'timing/data': 0.0005545789608731866, 'timing/model': 1.2254800179507583, 'learning_rate/qwen_vl_interface': 9.40657447652773e-06, 'learning_rate/action_model': 9.406574476527731e-05, 'epoch': 6.45}) +[2026-06-20 14:55:40,512][starVLA.training.train_starvla][INFO] - Step 3489, Loss: {'train/loss': 0.07617954164743423, 'train/grad_norm_pre_clip': 3.936479091644287, 'timing/data': 0.000621460028924048, 'timing/model': 1.178749356069602, 'learning_rate/qwen_vl_interface': 9.405099640067685e-06, 'learning_rate/action_model': 9.405099640067685e-05, 'epoch': 6.45}) +[2026-06-20 14:55:45,460][starVLA.training.train_starvla][INFO] - Step 3490, Loss: {'train/loss': 0.07486264407634735, 'train/grad_norm_pre_clip': 3.165529489517212, 'timing/data': 0.0005618189461529255, 'timing/model': 1.4294502929551527, 'learning_rate/qwen_vl_interface': 9.40362309560313e-06, 'learning_rate/action_model': 9.40362309560313e-05, 'epoch': 6.45}) +[2026-06-20 14:55:50,197][starVLA.training.train_starvla][INFO] - Step 3491, Loss: {'train/loss': 0.36717087030410767, 'train/grad_norm_pre_clip': 9.114066123962402, 'timing/data': 0.0005485899746417999, 'timing/model': 1.2260826260317117, 'learning_rate/qwen_vl_interface': 9.402144843741016e-06, 'learning_rate/action_model': 9.402144843741017e-05, 'epoch': 6.46}) +[2026-06-20 14:55:55,412][starVLA.training.train_starvla][INFO] - Step 3492, Loss: {'train/loss': 0.23741444945335388, 'train/grad_norm_pre_clip': 5.339022636413574, 'timing/data': 0.0005449799355119467, 'timing/model': 1.3300058339955285, 'learning_rate/qwen_vl_interface': 9.400664885088999e-06, 'learning_rate/action_model': 9.400664885088999e-05, 'epoch': 6.46}) +[2026-06-20 14:56:00,584][starVLA.training.train_starvla][INFO] - Step 3493, Loss: {'train/loss': 0.1680341213941574, 'train/grad_norm_pre_clip': 7.894122123718262, 'timing/data': 0.0005527290049940348, 'timing/model': 1.2547794519923627, 'learning_rate/qwen_vl_interface': 9.39918322025543e-06, 'learning_rate/action_model': 9.39918322025543e-05, 'epoch': 6.46}) +[2026-06-20 14:56:05,441][starVLA.training.train_starvla][INFO] - Step 3494, Loss: {'train/loss': 0.20987942814826965, 'train/grad_norm_pre_clip': 7.007767677307129, 'timing/data': 0.0006146889645606279, 'timing/model': 1.2228552360320464, 'learning_rate/qwen_vl_interface': 9.397699849849368e-06, 'learning_rate/action_model': 9.397699849849367e-05, 'epoch': 6.46}) +[2026-06-20 14:56:10,392][starVLA.training.train_starvla][INFO] - Step 3495, Loss: {'train/loss': 0.14452476799488068, 'train/grad_norm_pre_clip': 8.769722938537598, 'timing/data': 0.0006315700011327863, 'timing/model': 1.4633648280287161, 'learning_rate/qwen_vl_interface': 9.39621477448057e-06, 'learning_rate/action_model': 9.39621477448057e-05, 'epoch': 6.46}) +[2026-06-20 14:56:15,304][starVLA.training.train_starvla][INFO] - Step 3496, Loss: {'train/loss': 0.23256361484527588, 'train/grad_norm_pre_clip': 6.1635589599609375, 'timing/data': 0.0006250690203160048, 'timing/model': 1.343301157001406, 'learning_rate/qwen_vl_interface': 9.394727994759495e-06, 'learning_rate/action_model': 9.394727994759495e-05, 'epoch': 6.47}) +[2026-06-20 14:56:19,900][starVLA.training.train_starvla][INFO] - Step 3497, Loss: {'train/loss': 0.20564313232898712, 'train/grad_norm_pre_clip': 7.697408676147461, 'timing/data': 0.0005517989629879594, 'timing/model': 1.1813103139866143, 'learning_rate/qwen_vl_interface': 9.3932395112973e-06, 'learning_rate/action_model': 9.3932395112973e-05, 'epoch': 6.47}) +[2026-06-20 14:56:24,736][starVLA.training.train_starvla][INFO] - Step 3498, Loss: {'train/loss': 0.10562295466661453, 'train/grad_norm_pre_clip': 5.829117298126221, 'timing/data': 0.0005290890112519264, 'timing/model': 1.17937205103226, 'learning_rate/qwen_vl_interface': 9.391749324705846e-06, 'learning_rate/action_model': 9.391749324705846e-05, 'epoch': 6.47}) +[2026-06-20 14:56:29,770][starVLA.training.train_starvla][INFO] - Step 3499, Loss: {'train/loss': 0.07228115946054459, 'train/grad_norm_pre_clip': 4.483569145202637, 'timing/data': 0.0005277699092403054, 'timing/model': 1.4233213760890067, 'learning_rate/qwen_vl_interface': 9.390257435597692e-06, 'learning_rate/action_model': 9.390257435597692e-05, 'epoch': 6.47}) +[2026-06-20 14:56:34,942][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 3500 +[2026-06-20 15:02:52,637][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 3500 +[2026-06-20 15:02:52,638][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 3500 +[2026-06-20 15:09:33,544][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 3500 +[2026-06-20 15:09:33,546][starVLA.training.train_starvla][INFO] - Step 3500, Loss: {'train/loss': 0.25327977538108826, 'train/grad_norm_pre_clip': 7.17431640625, 'eval/loss': 0.4384521245956421, 'eval/latency_0/loss': 0.532110333442688, 'eval/latency_1/loss': 0.6374090313911438, 'eval/latency_2/loss': 0.6209284663200378, 'eval/latency_3/loss': 0.24226665496826172, 'eval/latency_4/loss': 0.13574057817459106, 'eval/flappy/loss': 0.4384521245956421, 'eval/flappy/latency_0/loss': 0.532110333442688, 'eval/flappy/latency_1/loss': 0.6374090313911438, 'eval/flappy/latency_2/loss': 0.6209284663200378, 'eval/flappy/latency_3/loss': 0.24226665496826172, 'eval/flappy/latency_4/loss': 0.13574057817459106, 'eval/flap_precision': 0.18404907975460122, 'eval/flap_recall': 0.10909090909090909, 'eval/flap_f1': 0.136986301369863, 'eval/flap_rate_model': 0.04527777777777778, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.136986301369863, 'eval/latency_0/flap_precision': 0.18404907975460122, 'eval/latency_0/flap_recall': 0.10909090909090909, 'eval/latency_0/flap_f1': 0.136986301369863, 'eval/latency_0/flap_rate_model': 0.04527777777777778, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.136986301369863, 'timing/data': 0.0005355689208954573, 'timing/model': 1.3602272958960384, 'learning_rate/qwen_vl_interface': 9.3887638445861e-06, 'learning_rate/action_model': 9.3887638445861e-05, 'epoch': 6.47}) +[2026-06-20 15:09:33,547][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_3500_state +[2026-06-20 15:09:33,551][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-20 15:09:46,513][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_3500_state/model.safetensors +[2026-06-20 15:10:04,475][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_3500_state/optimizer.bin +[2026-06-20 15:10:04,476][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_3500_state/sampler.bin +[2026-06-20 15:10:04,476][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_3500_state/sampler_1.bin +[2026-06-20 15:10:04,477][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_3500_state/random_states_0.pkl +[2026-06-20 15:10:06,127][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3500 +[2026-06-20 15:10:06,128][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-20 15:10:06,142][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-20 15:10:11,052][starVLA.training.train_starvla][INFO] - Step 3501, Loss: {'train/loss': 0.19732187688350677, 'train/grad_norm_pre_clip': 7.933377265930176, 'timing/data': 0.0005414190236479044, 'timing/model': 1.226062349975109, 'learning_rate/qwen_vl_interface': 9.387268552285024e-06, 'learning_rate/action_model': 9.387268552285025e-05, 'epoch': 6.47}) +[2026-06-20 15:10:15,829][starVLA.training.train_starvla][INFO] - Step 3502, Loss: {'train/loss': 0.09269855916500092, 'train/grad_norm_pre_clip': 51.37782669067383, 'timing/data': 0.0005413390463218093, 'timing/model': 1.4223632579669356, 'learning_rate/qwen_vl_interface': 9.385771559309125e-06, 'learning_rate/action_model': 9.385771559309125e-05, 'epoch': 6.48}) +[2026-06-20 15:10:20,410][starVLA.training.train_starvla][INFO] - Step 3503, Loss: {'train/loss': 0.1817762404680252, 'train/grad_norm_pre_clip': 6.998424530029297, 'timing/data': 0.0005827690474689007, 'timing/model': 1.1696373670129105, 'learning_rate/qwen_vl_interface': 9.38427286627376e-06, 'learning_rate/action_model': 9.38427286627376e-05, 'epoch': 6.48}) +[2026-06-20 15:10:25,243][starVLA.training.train_starvla][INFO] - Step 3504, Loss: {'train/loss': 0.09250421077013016, 'train/grad_norm_pre_clip': 6.666118144989014, 'timing/data': 0.000522137968800962, 'timing/model': 1.2148408049251884, 'learning_rate/qwen_vl_interface': 9.382772473794985e-06, 'learning_rate/action_model': 9.382772473794985e-05, 'epoch': 6.48}) +[2026-06-20 15:10:30,069][starVLA.training.train_starvla][INFO] - Step 3505, Loss: {'train/loss': 0.38931164145469666, 'train/grad_norm_pre_clip': 13.925652503967285, 'timing/data': 0.0007756480481475592, 'timing/model': 1.1787917719921097, 'learning_rate/qwen_vl_interface': 9.381270382489554e-06, 'learning_rate/action_model': 9.381270382489554e-05, 'epoch': 6.48}) +[2026-06-20 15:10:35,036][starVLA.training.train_starvla][INFO] - Step 3506, Loss: {'train/loss': 0.14292629063129425, 'train/grad_norm_pre_clip': 5.55343770980835, 'timing/data': 0.0005696779116988182, 'timing/model': 1.183921196963638, 'learning_rate/qwen_vl_interface': 9.37976659297492e-06, 'learning_rate/action_model': 9.379766592974921e-05, 'epoch': 6.48}) +[2026-06-20 15:10:40,179][starVLA.training.train_starvla][INFO] - Step 3507, Loss: {'train/loss': 0.37774422764778137, 'train/grad_norm_pre_clip': 11.572476387023926, 'timing/data': 0.0006751390174031258, 'timing/model': 1.2510421960614622, 'learning_rate/qwen_vl_interface': 9.378261105869237e-06, 'learning_rate/action_model': 9.378261105869237e-05, 'epoch': 6.49}) +[2026-06-20 15:10:45,037][starVLA.training.train_starvla][INFO] - Step 3508, Loss: {'train/loss': 0.1694616675376892, 'train/grad_norm_pre_clip': 12.386011123657227, 'timing/data': 0.0007923688972368836, 'timing/model': 1.465476143057458, 'learning_rate/qwen_vl_interface': 9.376753921791348e-06, 'learning_rate/action_model': 9.376753921791348e-05, 'epoch': 6.49}) +[2026-06-20 15:10:49,635][starVLA.training.train_starvla][INFO] - Step 3509, Loss: {'train/loss': 0.18516799807548523, 'train/grad_norm_pre_clip': 9.277673721313477, 'timing/data': 0.0006634790915995836, 'timing/model': 1.1678685629740357, 'learning_rate/qwen_vl_interface': 9.375245041360805e-06, 'learning_rate/action_model': 9.375245041360804e-05, 'epoch': 6.49}) +[2026-06-20 15:10:54,597][starVLA.training.train_starvla][INFO] - Step 3510, Loss: {'train/loss': 0.29846489429473877, 'train/grad_norm_pre_clip': 7.301971912384033, 'timing/data': 0.0005303090438246727, 'timing/model': 1.1763248419156298, 'learning_rate/qwen_vl_interface': 9.37373446519785e-06, 'learning_rate/action_model': 9.373734465197849e-05, 'epoch': 6.49}) +[2026-06-20 15:10:59,499][starVLA.training.train_starvla][INFO] - Step 3511, Loss: {'train/loss': 0.3363395035266876, 'train/grad_norm_pre_clip': 9.8040189743042, 'timing/data': 0.0005367789417505264, 'timing/model': 1.2373180449940264, 'learning_rate/qwen_vl_interface': 9.372222193923421e-06, 'learning_rate/action_model': 9.372222193923421e-05, 'epoch': 6.49}) +[2026-06-20 15:11:04,356][starVLA.training.train_starvla][INFO] - Step 3512, Loss: {'train/loss': 0.18423664569854736, 'train/grad_norm_pre_clip': 8.132378578186035, 'timing/data': 0.0005312979919835925, 'timing/model': 1.2220697860466316, 'learning_rate/qwen_vl_interface': 9.370708228159158e-06, 'learning_rate/action_model': 9.370708228159159e-05, 'epoch': 6.49}) +[2026-06-20 15:11:09,687][starVLA.training.train_starvla][INFO] - Step 3513, Loss: {'train/loss': 0.1335029900074005, 'train/grad_norm_pre_clip': 7.121514797210693, 'timing/data': 0.0005942980060353875, 'timing/model': 1.249923661001958, 'learning_rate/qwen_vl_interface': 9.369192568527398e-06, 'learning_rate/action_model': 9.369192568527398e-05, 'epoch': 6.5}) +[2026-06-20 15:11:14,971][starVLA.training.train_starvla][INFO] - Step 3514, Loss: {'train/loss': 0.32480284571647644, 'train/grad_norm_pre_clip': 10.585193634033203, 'timing/data': 0.0005694690626114607, 'timing/model': 1.72301344294101, 'learning_rate/qwen_vl_interface': 9.367675215651168e-06, 'learning_rate/action_model': 9.367675215651168e-05, 'epoch': 6.5}) +[2026-06-20 15:11:19,870][starVLA.training.train_starvla][INFO] - Step 3515, Loss: {'train/loss': 0.1260048896074295, 'train/grad_norm_pre_clip': 12.777932167053223, 'timing/data': 0.0005522689316421747, 'timing/model': 1.3149110319791362, 'learning_rate/qwen_vl_interface': 9.366156170154194e-06, 'learning_rate/action_model': 9.366156170154194e-05, 'epoch': 6.5}) +[2026-06-20 15:11:24,724][starVLA.training.train_starvla][INFO] - Step 3516, Loss: {'train/loss': 0.24008579552173615, 'train/grad_norm_pre_clip': 9.991236686706543, 'timing/data': 0.0006616179598495364, 'timing/model': 1.1910665630130097, 'learning_rate/qwen_vl_interface': 9.3646354326609e-06, 'learning_rate/action_model': 9.364635432660899e-05, 'epoch': 6.5}) +[2026-06-20 15:11:29,709][starVLA.training.train_starvla][INFO] - Step 3517, Loss: {'train/loss': 0.1661134511232376, 'train/grad_norm_pre_clip': 10.727482795715332, 'timing/data': 0.0007230080664157867, 'timing/model': 1.2435052039800212, 'learning_rate/qwen_vl_interface': 9.3631130037964e-06, 'learning_rate/action_model': 9.363113003796401e-05, 'epoch': 6.5}) +[2026-06-20 15:11:34,712][starVLA.training.train_starvla][INFO] - Step 3518, Loss: {'train/loss': 0.10573313385248184, 'train/grad_norm_pre_clip': 9.096447944641113, 'timing/data': 0.0005558590637519956, 'timing/model': 1.2220430079614744, 'learning_rate/qwen_vl_interface': 9.361588884186513e-06, 'learning_rate/action_model': 9.361588884186513e-05, 'epoch': 6.51}) +[2026-06-20 15:11:39,844][starVLA.training.train_starvla][INFO] - Step 3519, Loss: {'train/loss': 0.04367608577013016, 'train/grad_norm_pre_clip': 3.9536001682281494, 'timing/data': 0.0005372990854084492, 'timing/model': 1.386283201049082, 'learning_rate/qwen_vl_interface': 9.360063074457744e-06, 'learning_rate/action_model': 9.360063074457743e-05, 'epoch': 6.51}) +[2026-06-20 15:11:44,922][starVLA.training.train_starvla][INFO] - Step 3520, Loss: {'train/loss': 0.2177729606628418, 'train/grad_norm_pre_clip': 7.887760162353516, 'timing/data': 0.0005344889359548688, 'timing/model': 1.553898285026662, 'learning_rate/qwen_vl_interface': 9.358535575237296e-06, 'learning_rate/action_model': 9.358535575237295e-05, 'epoch': 6.51}) +[2026-06-20 15:11:50,043][starVLA.training.train_starvla][INFO] - Step 3521, Loss: {'train/loss': 0.14771611988544464, 'train/grad_norm_pre_clip': 7.965354919433594, 'timing/data': 0.0005341090727597475, 'timing/model': 1.5800908439559862, 'learning_rate/qwen_vl_interface': 9.357006387153065e-06, 'learning_rate/action_model': 9.357006387153065e-05, 'epoch': 6.51}) +[2026-06-20 15:11:54,683][starVLA.training.train_starvla][INFO] - Step 3522, Loss: {'train/loss': 0.09298063069581985, 'train/grad_norm_pre_clip': 7.39120626449585, 'timing/data': 0.0005338890478014946, 'timing/model': 1.201102014980279, 'learning_rate/qwen_vl_interface': 9.355475510833643e-06, 'learning_rate/action_model': 9.355475510833643e-05, 'epoch': 6.51}) +[2026-06-20 15:11:59,651][starVLA.training.train_starvla][INFO] - Step 3523, Loss: {'train/loss': 0.10229838639497757, 'train/grad_norm_pre_clip': 5.691985130310059, 'timing/data': 0.0005316589958965778, 'timing/model': 1.2351735960692167, 'learning_rate/qwen_vl_interface': 9.353942946908318e-06, 'learning_rate/action_model': 9.353942946908317e-05, 'epoch': 6.52}) +[2026-06-20 15:12:04,999][starVLA.training.train_starvla][INFO] - Step 3524, Loss: {'train/loss': 0.07693780213594437, 'train/grad_norm_pre_clip': 4.1879448890686035, 'timing/data': 0.0005750289419665933, 'timing/model': 1.573982466943562, 'learning_rate/qwen_vl_interface': 9.352408696007065e-06, 'learning_rate/action_model': 9.352408696007065e-05, 'epoch': 6.52}) +[2026-06-20 15:12:10,128][starVLA.training.train_starvla][INFO] - Step 3525, Loss: {'train/loss': 0.14820104837417603, 'train/grad_norm_pre_clip': 14.629494667053223, 'timing/data': 0.0005334590096026659, 'timing/model': 1.3323010930325836, 'learning_rate/qwen_vl_interface': 9.350872758760558e-06, 'learning_rate/action_model': 9.350872758760559e-05, 'epoch': 6.52}) +[2026-06-20 15:12:17,547][starVLA.training.train_starvla][INFO] - Step 3526, Loss: {'train/loss': 0.07407272607088089, 'train/grad_norm_pre_clip': 4.6127610206604, 'timing/data': 0.000629558926448226, 'timing/model': 1.4864522439893335, 'learning_rate/qwen_vl_interface': 9.349335135800164e-06, 'learning_rate/action_model': 9.349335135800163e-05, 'epoch': 6.52}) +[2026-06-20 15:12:22,496][starVLA.training.train_starvla][INFO] - Step 3527, Loss: {'train/loss': 0.13240200281143188, 'train/grad_norm_pre_clip': 8.827372550964355, 'timing/data': 0.0007566580316051841, 'timing/model': 1.2727689900202677, 'learning_rate/qwen_vl_interface': 9.347795827757942e-06, 'learning_rate/action_model': 9.347795827757941e-05, 'epoch': 6.52}) +[2026-06-20 15:12:27,440][starVLA.training.train_starvla][INFO] - Step 3528, Loss: {'train/loss': 0.289279580116272, 'train/grad_norm_pre_clip': 18.10749626159668, 'timing/data': 0.0007090379949659109, 'timing/model': 1.2192505379207432, 'learning_rate/qwen_vl_interface': 9.346254835266643e-06, 'learning_rate/action_model': 9.346254835266642e-05, 'epoch': 6.52}) +[2026-06-20 15:12:32,535][starVLA.training.train_starvla][INFO] - Step 3529, Loss: {'train/loss': 0.170735165476799, 'train/grad_norm_pre_clip': 10.393783569335938, 'timing/data': 0.0007225680164992809, 'timing/model': 1.2650130729889497, 'learning_rate/qwen_vl_interface': 9.344712158959709e-06, 'learning_rate/action_model': 9.344712158959709e-05, 'epoch': 6.53}) +[2026-06-20 15:12:37,383][starVLA.training.train_starvla][INFO] - Step 3530, Loss: {'train/loss': 0.2357388734817505, 'train/grad_norm_pre_clip': 10.562273025512695, 'timing/data': 0.0005383290117606521, 'timing/model': 1.1802325500175357, 'learning_rate/qwen_vl_interface': 9.343167799471278e-06, 'learning_rate/action_model': 9.343167799471277e-05, 'epoch': 6.53}) +[2026-06-20 15:12:42,240][starVLA.training.train_starvla][INFO] - Step 3531, Loss: {'train/loss': 0.3389416038990021, 'train/grad_norm_pre_clip': 20.076025009155273, 'timing/data': 0.0006314089987426996, 'timing/model': 1.1838490510126576, 'learning_rate/qwen_vl_interface': 9.341621757436177e-06, 'learning_rate/action_model': 9.341621757436177e-05, 'epoch': 6.53}) +[2026-06-20 15:12:47,136][starVLA.training.train_starvla][INFO] - Step 3532, Loss: {'train/loss': 0.41098707914352417, 'train/grad_norm_pre_clip': 18.456212997436523, 'timing/data': 0.0005522489082068205, 'timing/model': 1.3939769739517942, 'learning_rate/qwen_vl_interface': 9.340074033489924e-06, 'learning_rate/action_model': 9.340074033489923e-05, 'epoch': 6.53}) +[2026-06-20 15:12:52,081][starVLA.training.train_starvla][INFO] - Step 3533, Loss: {'train/loss': 0.1232328787446022, 'train/grad_norm_pre_clip': 4.433931350708008, 'timing/data': 0.0005423990078270435, 'timing/model': 1.2620476299198344, 'learning_rate/qwen_vl_interface': 9.338524628268734e-06, 'learning_rate/action_model': 9.338524628268733e-05, 'epoch': 6.53}) +[2026-06-20 15:12:56,945][starVLA.training.train_starvla][INFO] - Step 3534, Loss: {'train/loss': 0.12284141033887863, 'train/grad_norm_pre_clip': 2.032363176345825, 'timing/data': 0.0005377280758693814, 'timing/model': 1.207615999970585, 'learning_rate/qwen_vl_interface': 9.336973542409505e-06, 'learning_rate/action_model': 9.336973542409504e-05, 'epoch': 6.54}) +[2026-06-20 15:13:01,961][starVLA.training.train_starvla][INFO] - Step 3535, Loss: {'train/loss': 0.37644556164741516, 'train/grad_norm_pre_clip': 17.446842193603516, 'timing/data': 0.0005820989608764648, 'timing/model': 1.277401759987697, 'learning_rate/qwen_vl_interface': 9.33542077654983e-06, 'learning_rate/action_model': 9.33542077654983e-05, 'epoch': 6.54}) +[2026-06-20 15:13:06,833][starVLA.training.train_starvla][INFO] - Step 3536, Loss: {'train/loss': 0.2737351655960083, 'train/grad_norm_pre_clip': 16.807889938354492, 'timing/data': 0.000543248956091702, 'timing/model': 1.2152434579329565, 'learning_rate/qwen_vl_interface': 9.333866331327995e-06, 'learning_rate/action_model': 9.333866331327995e-05, 'epoch': 6.54}) +[2026-06-20 15:13:11,752][starVLA.training.train_starvla][INFO] - Step 3537, Loss: {'train/loss': 0.1404353529214859, 'train/grad_norm_pre_clip': 8.5327787399292, 'timing/data': 0.0005105279851704836, 'timing/model': 1.2422773439902812, 'learning_rate/qwen_vl_interface': 9.332310207382972e-06, 'learning_rate/action_model': 9.332310207382972e-05, 'epoch': 6.54}) +[2026-06-20 15:13:16,750][starVLA.training.train_starvla][INFO] - Step 3538, Loss: {'train/loss': 0.10088451206684113, 'train/grad_norm_pre_clip': 9.820709228515625, 'timing/data': 0.0005184389883652329, 'timing/model': 1.4280529730021954, 'learning_rate/qwen_vl_interface': 9.330752405354426e-06, 'learning_rate/action_model': 9.330752405354426e-05, 'epoch': 6.54}) +[2026-06-20 15:13:21,849][starVLA.training.train_starvla][INFO] - Step 3539, Loss: {'train/loss': 0.05270172283053398, 'train/grad_norm_pre_clip': 3.492903470993042, 'timing/data': 0.0006430780049413443, 'timing/model': 1.1923998720012605, 'learning_rate/qwen_vl_interface': 9.32919292588271e-06, 'learning_rate/action_model': 9.32919292588271e-05, 'epoch': 6.54}) +[2026-06-20 15:13:26,913][starVLA.training.train_starvla][INFO] - Step 3540, Loss: {'train/loss': 0.19874729216098785, 'train/grad_norm_pre_clip': 6.215974807739258, 'timing/data': 0.0005403979448601604, 'timing/model': 1.3390975380316377, 'learning_rate/qwen_vl_interface': 9.327631769608864e-06, 'learning_rate/action_model': 9.327631769608864e-05, 'epoch': 6.55}) +[2026-06-20 15:13:31,975][starVLA.training.train_starvla][INFO] - Step 3541, Loss: {'train/loss': 0.17511829733848572, 'train/grad_norm_pre_clip': 6.818627834320068, 'timing/data': 0.0005776690086349845, 'timing/model': 1.2689113150117919, 'learning_rate/qwen_vl_interface': 9.326068937174627e-06, 'learning_rate/action_model': 9.326068937174626e-05, 'epoch': 6.55}) +[2026-06-20 15:13:38,568][starVLA.training.train_starvla][INFO] - Step 3542, Loss: {'train/loss': 0.2167711704969406, 'train/grad_norm_pre_clip': 6.760990142822266, 'timing/data': 0.0005433089099824429, 'timing/model': 1.3148404769599438, 'learning_rate/qwen_vl_interface': 9.324504429222412e-06, 'learning_rate/action_model': 9.324504429222412e-05, 'epoch': 6.55}) +[2026-06-20 15:13:43,721][starVLA.training.train_starvla][INFO] - Step 3543, Loss: {'train/loss': 0.056075241416692734, 'train/grad_norm_pre_clip': 5.156228065490723, 'timing/data': 0.0012399869738146663, 'timing/model': 1.510383661952801, 'learning_rate/qwen_vl_interface': 9.322938246395336e-06, 'learning_rate/action_model': 9.322938246395337e-05, 'epoch': 6.55}) +[2026-06-20 15:13:48,428][starVLA.training.train_starvla][INFO] - Step 3544, Loss: {'train/loss': 0.2325882911682129, 'train/grad_norm_pre_clip': 6.219805717468262, 'timing/data': 0.0005409790901467204, 'timing/model': 1.2249152680160478, 'learning_rate/qwen_vl_interface': 9.321370389337196e-06, 'learning_rate/action_model': 9.321370389337196e-05, 'epoch': 6.55}) +[2026-06-20 15:13:53,746][starVLA.training.train_starvla][INFO] - Step 3545, Loss: {'train/loss': 0.10763116925954819, 'train/grad_norm_pre_clip': 5.7229156494140625, 'timing/data': 0.0005201390013098717, 'timing/model': 1.2490790409501642, 'learning_rate/qwen_vl_interface': 9.319800858692478e-06, 'learning_rate/action_model': 9.319800858692476e-05, 'epoch': 6.56}) +[2026-06-20 15:13:58,955][starVLA.training.train_starvla][INFO] - Step 3546, Loss: {'train/loss': 0.27691060304641724, 'train/grad_norm_pre_clip': 15.906261444091797, 'timing/data': 0.0005301290657371283, 'timing/model': 1.2258875409606844, 'learning_rate/qwen_vl_interface': 9.318229655106357e-06, 'learning_rate/action_model': 9.318229655106356e-05, 'epoch': 6.56}) +[2026-06-20 15:14:04,002][starVLA.training.train_starvla][INFO] - Step 3547, Loss: {'train/loss': 0.15541140735149384, 'train/grad_norm_pre_clip': 15.533454895019531, 'timing/data': 0.0005189590156078339, 'timing/model': 1.2294627010123804, 'learning_rate/qwen_vl_interface': 9.316656779224693e-06, 'learning_rate/action_model': 9.316656779224693e-05, 'epoch': 6.56}) +[2026-06-20 15:14:08,968][starVLA.training.train_starvla][INFO] - Step 3548, Loss: {'train/loss': 0.11548090726137161, 'train/grad_norm_pre_clip': 7.009515762329102, 'timing/data': 0.0005164090543985367, 'timing/model': 1.2265790150268003, 'learning_rate/qwen_vl_interface': 9.31508223169404e-06, 'learning_rate/action_model': 9.31508223169404e-05, 'epoch': 6.56}) +[2026-06-20 15:14:14,179][starVLA.training.train_starvla][INFO] - Step 3549, Loss: {'train/loss': 0.09831352531909943, 'train/grad_norm_pre_clip': 5.9388532638549805, 'timing/data': 0.0005839880323037505, 'timing/model': 1.4371102509321645, 'learning_rate/qwen_vl_interface': 9.313506013161633e-06, 'learning_rate/action_model': 9.313506013161633e-05, 'epoch': 6.56}) +[2026-06-20 15:14:18,976][starVLA.training.train_starvla][INFO] - Step 3550, Loss: {'train/loss': 0.2538737654685974, 'train/grad_norm_pre_clip': 9.821147918701172, 'timing/data': 0.0005378290079534054, 'timing/model': 1.2411738589871675, 'learning_rate/qwen_vl_interface': 9.311928124275396e-06, 'learning_rate/action_model': 9.311928124275395e-05, 'epoch': 6.57}) +[2026-06-20 15:14:24,125][starVLA.training.train_starvla][INFO] - Step 3551, Loss: {'train/loss': 0.21880924701690674, 'train/grad_norm_pre_clip': 14.142720222473145, 'timing/data': 0.0006588289979845285, 'timing/model': 1.2345427219988778, 'learning_rate/qwen_vl_interface': 9.310348565683938e-06, 'learning_rate/action_model': 9.310348565683939e-05, 'epoch': 6.57}) +[2026-06-20 15:14:29,108][starVLA.training.train_starvla][INFO] - Step 3552, Loss: {'train/loss': 0.07193176448345184, 'train/grad_norm_pre_clip': 4.7072882652282715, 'timing/data': 0.0005561579018831253, 'timing/model': 1.2364138460252434, 'learning_rate/qwen_vl_interface': 9.30876733803656e-06, 'learning_rate/action_model': 9.308767338036559e-05, 'epoch': 6.57}) +[2026-06-20 15:14:33,995][starVLA.training.train_starvla][INFO] - Step 3553, Loss: {'train/loss': 0.1259964406490326, 'train/grad_norm_pre_clip': 11.31324291229248, 'timing/data': 0.0005525280721485615, 'timing/model': 1.2474844479002059, 'learning_rate/qwen_vl_interface': 9.30718444198324e-06, 'learning_rate/action_model': 9.307184441983239e-05, 'epoch': 6.57}) +[2026-06-20 15:14:38,846][starVLA.training.train_starvla][INFO] - Step 3554, Loss: {'train/loss': 0.2689003050327301, 'train/grad_norm_pre_clip': 9.021657943725586, 'timing/data': 0.0008526579476892948, 'timing/model': 1.1858617769321427, 'learning_rate/qwen_vl_interface': 9.30559987817465e-06, 'learning_rate/action_model': 9.30559987817465e-05, 'epoch': 6.57}) +[2026-06-20 15:14:43,863][starVLA.training.train_starvla][INFO] - Step 3555, Loss: {'train/loss': 0.22977232933044434, 'train/grad_norm_pre_clip': 12.39754867553711, 'timing/data': 0.0006538279121741652, 'timing/model': 1.538010539021343, 'learning_rate/qwen_vl_interface': 9.304013647262143e-06, 'learning_rate/action_model': 9.304013647262143e-05, 'epoch': 6.57}) +[2026-06-20 15:14:48,638][starVLA.training.train_starvla][INFO] - Step 3556, Loss: {'train/loss': 0.25168320536613464, 'train/grad_norm_pre_clip': 11.651642799377441, 'timing/data': 0.0005604089237749577, 'timing/model': 1.3321375689702109, 'learning_rate/qwen_vl_interface': 9.302425749897761e-06, 'learning_rate/action_model': 9.302425749897761e-05, 'epoch': 6.58}) +[2026-06-20 15:14:53,583][starVLA.training.train_starvla][INFO] - Step 3557, Loss: {'train/loss': 0.3569336235523224, 'train/grad_norm_pre_clip': 14.820572853088379, 'timing/data': 0.0005168690113350749, 'timing/model': 1.2255274700000882, 'learning_rate/qwen_vl_interface': 9.300836186734225e-06, 'learning_rate/action_model': 9.300836186734226e-05, 'epoch': 6.58}) +[2026-06-20 15:14:58,610][starVLA.training.train_starvla][INFO] - Step 3558, Loss: {'train/loss': 0.2872318923473358, 'train/grad_norm_pre_clip': 18.34307861328125, 'timing/data': 0.0008498469833284616, 'timing/model': 1.2183561050333083, 'learning_rate/qwen_vl_interface': 9.299244958424947e-06, 'learning_rate/action_model': 9.299244958424946e-05, 'epoch': 6.58}) +[2026-06-20 15:15:03,655][starVLA.training.train_starvla][INFO] - Step 3559, Loss: {'train/loss': 0.1053585633635521, 'train/grad_norm_pre_clip': 3.8846545219421387, 'timing/data': 0.0006377180106937885, 'timing/model': 1.2249734989600256, 'learning_rate/qwen_vl_interface': 9.29765206562402e-06, 'learning_rate/action_model': 9.297652065624018e-05, 'epoch': 6.58}) +[2026-06-20 15:15:08,655][starVLA.training.train_starvla][INFO] - Step 3560, Loss: {'train/loss': 0.23293349146842957, 'train/grad_norm_pre_clip': 10.74483585357666, 'timing/data': 0.0005431489553302526, 'timing/model': 1.4777131960727274, 'learning_rate/qwen_vl_interface': 9.296057508986221e-06, 'learning_rate/action_model': 9.296057508986221e-05, 'epoch': 6.58}) +[2026-06-20 15:15:13,604][starVLA.training.train_starvla][INFO] - Step 3561, Loss: {'train/loss': 0.17389151453971863, 'train/grad_norm_pre_clip': 7.414223670959473, 'timing/data': 0.000519038992933929, 'timing/model': 1.4390565330395475, 'learning_rate/qwen_vl_interface': 9.294461289167013e-06, 'learning_rate/action_model': 9.294461289167014e-05, 'epoch': 6.59}) +[2026-06-20 15:15:18,363][starVLA.training.train_starvla][INFO] - Step 3562, Loss: {'train/loss': 0.12420906126499176, 'train/grad_norm_pre_clip': 11.341472625732422, 'timing/data': 0.0005473979981616139, 'timing/model': 1.1843786059180275, 'learning_rate/qwen_vl_interface': 9.292863406822543e-06, 'learning_rate/action_model': 9.292863406822542e-05, 'epoch': 6.59}) +[2026-06-20 15:15:23,203][starVLA.training.train_starvla][INFO] - Step 3563, Loss: {'train/loss': 0.21189294755458832, 'train/grad_norm_pre_clip': 8.467986106872559, 'timing/data': 0.0006926290225237608, 'timing/model': 1.1798762050457299, 'learning_rate/qwen_vl_interface': 9.291263862609638e-06, 'learning_rate/action_model': 9.291263862609638e-05, 'epoch': 6.59}) +[2026-06-20 15:15:28,129][starVLA.training.train_starvla][INFO] - Step 3564, Loss: {'train/loss': 0.09772532433271408, 'train/grad_norm_pre_clip': 5.49806022644043, 'timing/data': 0.0005558190168812871, 'timing/model': 1.2071248100837693, 'learning_rate/qwen_vl_interface': 9.28966265718581e-06, 'learning_rate/action_model': 9.28966265718581e-05, 'epoch': 6.59}) +[2026-06-20 15:15:33,053][starVLA.training.train_starvla][INFO] - Step 3565, Loss: {'train/loss': 0.19510436058044434, 'train/grad_norm_pre_clip': 5.594989776611328, 'timing/data': 0.0007726580370217562, 'timing/model': 1.2196641600457951, 'learning_rate/qwen_vl_interface': 9.288059791209256e-06, 'learning_rate/action_model': 9.288059791209256e-05, 'epoch': 6.59}) +[2026-06-20 15:15:38,061][starVLA.training.train_starvla][INFO] - Step 3566, Loss: {'train/loss': 0.19350719451904297, 'train/grad_norm_pre_clip': 13.855690956115723, 'timing/data': 0.000636928016319871, 'timing/model': 1.4535637990338728, 'learning_rate/qwen_vl_interface': 9.286455265338853e-06, 'learning_rate/action_model': 9.286455265338853e-05, 'epoch': 6.59}) +[2026-06-20 15:15:42,969][starVLA.training.train_starvla][INFO] - Step 3567, Loss: {'train/loss': 0.11034929752349854, 'train/grad_norm_pre_clip': 5.476318836212158, 'timing/data': 0.0005351389991119504, 'timing/model': 1.4477695700479671, 'learning_rate/qwen_vl_interface': 9.284849080234161e-06, 'learning_rate/action_model': 9.28484908023416e-05, 'epoch': 6.6}) +[2026-06-20 15:15:47,720][starVLA.training.train_starvla][INFO] - Step 3568, Loss: {'train/loss': 0.16631251573562622, 'train/grad_norm_pre_clip': 7.129632472991943, 'timing/data': 0.0005646890494972467, 'timing/model': 1.2510489519918337, 'learning_rate/qwen_vl_interface': 9.283241236555421e-06, 'learning_rate/action_model': 9.283241236555421e-05, 'epoch': 6.6}) +[2026-06-20 15:15:52,746][starVLA.training.train_starvla][INFO] - Step 3569, Loss: {'train/loss': 0.21470975875854492, 'train/grad_norm_pre_clip': 57.810394287109375, 'timing/data': 0.00058858806733042, 'timing/model': 1.2399681260576472, 'learning_rate/qwen_vl_interface': 9.281631734963557e-06, 'learning_rate/action_model': 9.281631734963557e-05, 'epoch': 6.6}) +[2026-06-20 15:15:57,704][starVLA.training.train_starvla][INFO] - Step 3570, Loss: {'train/loss': 0.16133075952529907, 'train/grad_norm_pre_clip': 6.959965705871582, 'timing/data': 0.0007604080019518733, 'timing/model': 1.2454288930166513, 'learning_rate/qwen_vl_interface': 9.280020576120176e-06, 'learning_rate/action_model': 9.280020576120176e-05, 'epoch': 6.6}) +[2026-06-20 15:16:02,756][starVLA.training.train_starvla][INFO] - Step 3571, Loss: {'train/loss': 0.19612181186676025, 'train/grad_norm_pre_clip': 6.151658058166504, 'timing/data': 0.0005170479416847229, 'timing/model': 1.2705720840021968, 'learning_rate/qwen_vl_interface': 9.278407760687562e-06, 'learning_rate/action_model': 9.278407760687562e-05, 'epoch': 6.6}) +[2026-06-20 15:16:07,796][starVLA.training.train_starvla][INFO] - Step 3572, Loss: {'train/loss': 0.17315752804279327, 'train/grad_norm_pre_clip': 6.601064205169678, 'timing/data': 0.0009153969585895538, 'timing/model': 1.4565656820777804, 'learning_rate/qwen_vl_interface': 9.276793289328686e-06, 'learning_rate/action_model': 9.276793289328685e-05, 'epoch': 6.61}) +[2026-06-20 15:16:12,932][starVLA.training.train_starvla][INFO] - Step 3573, Loss: {'train/loss': 0.10594304651021957, 'train/grad_norm_pre_clip': 6.987273693084717, 'timing/data': 0.0009271380258724093, 'timing/model': 1.2799891870236024, 'learning_rate/qwen_vl_interface': 9.27517716270719e-06, 'learning_rate/action_model': 9.275177162707191e-05, 'epoch': 6.61}) +[2026-06-20 15:16:18,080][starVLA.training.train_starvla][INFO] - Step 3574, Loss: {'train/loss': 0.07484778761863708, 'train/grad_norm_pre_clip': 3.6432888507843018, 'timing/data': 0.0006235589971765876, 'timing/model': 1.1891897459281608, 'learning_rate/qwen_vl_interface': 9.273559381487409e-06, 'learning_rate/action_model': 9.273559381487409e-05, 'epoch': 6.61}) +[2026-06-20 15:16:23,060][starVLA.training.train_starvla][INFO] - Step 3575, Loss: {'train/loss': 0.07807332277297974, 'train/grad_norm_pre_clip': 10.641034126281738, 'timing/data': 0.0005518179386854172, 'timing/model': 1.273270079982467, 'learning_rate/qwen_vl_interface': 9.27193994633435e-06, 'learning_rate/action_model': 9.27193994633435e-05, 'epoch': 6.61}) +[2026-06-20 15:16:28,196][starVLA.training.train_starvla][INFO] - Step 3576, Loss: {'train/loss': 0.09895317256450653, 'train/grad_norm_pre_clip': 5.766268730163574, 'timing/data': 0.0006235280307009816, 'timing/model': 1.2016773340292275, 'learning_rate/qwen_vl_interface': 9.2703188579137e-06, 'learning_rate/action_model': 9.2703188579137e-05, 'epoch': 6.61}) +[2026-06-20 15:16:34,037][starVLA.training.train_starvla][INFO] - Step 3577, Loss: {'train/loss': 0.4088498055934906, 'train/grad_norm_pre_clip': 14.10886001586914, 'timing/data': 0.0012533770641312003, 'timing/model': 1.3007967830635607, 'learning_rate/qwen_vl_interface': 9.26869611689183e-06, 'learning_rate/action_model': 9.26869611689183e-05, 'epoch': 6.62}) +[2026-06-20 15:16:39,242][starVLA.training.train_starvla][INFO] - Step 3578, Loss: {'train/loss': 0.11426010727882385, 'train/grad_norm_pre_clip': 6.744184970855713, 'timing/data': 0.0005534089868888259, 'timing/model': 1.5318251539720222, 'learning_rate/qwen_vl_interface': 9.267071723935784e-06, 'learning_rate/action_model': 9.267071723935784e-05, 'epoch': 6.62}) +[2026-06-20 15:16:44,176][starVLA.training.train_starvla][INFO] - Step 3579, Loss: {'train/loss': 0.2463371455669403, 'train/grad_norm_pre_clip': 8.608275413513184, 'timing/data': 0.0005437580402940512, 'timing/model': 1.218467530910857, 'learning_rate/qwen_vl_interface': 9.26544567971329e-06, 'learning_rate/action_model': 9.26544567971329e-05, 'epoch': 6.62}) +[2026-06-20 15:16:49,385][starVLA.training.train_starvla][INFO] - Step 3580, Loss: {'train/loss': 0.07197839766740799, 'train/grad_norm_pre_clip': 3.9519975185394287, 'timing/data': 0.00092911790125072, 'timing/model': 1.1930677170166746, 'learning_rate/qwen_vl_interface': 9.263817984892756e-06, 'learning_rate/action_model': 9.263817984892757e-05, 'epoch': 6.62}) +[2026-06-20 15:16:54,685][starVLA.training.train_starvla][INFO] - Step 3581, Loss: {'train/loss': 0.10510287433862686, 'train/grad_norm_pre_clip': 7.409726142883301, 'timing/data': 0.0005255080759525299, 'timing/model': 1.2322223839582875, 'learning_rate/qwen_vl_interface': 9.262188640143262e-06, 'learning_rate/action_model': 9.262188640143262e-05, 'epoch': 6.62}) +[2026-06-20 15:16:59,795][starVLA.training.train_starvla][INFO] - Step 3582, Loss: {'train/loss': 0.1276671588420868, 'train/grad_norm_pre_clip': 8.779417991638184, 'timing/data': 0.0005481790285557508, 'timing/model': 1.2410550430649891, 'learning_rate/qwen_vl_interface': 9.260557646134572e-06, 'learning_rate/action_model': 9.260557646134572e-05, 'epoch': 6.62}) +[2026-06-20 15:17:04,908][starVLA.training.train_starvla][INFO] - Step 3583, Loss: {'train/loss': 0.17151029407978058, 'train/grad_norm_pre_clip': 8.158964157104492, 'timing/data': 0.0062526550609618425, 'timing/model': 1.2406143030384555, 'learning_rate/qwen_vl_interface': 9.258925003537127e-06, 'learning_rate/action_model': 9.258925003537126e-05, 'epoch': 6.63}) +[2026-06-20 15:17:09,996][starVLA.training.train_starvla][INFO] - Step 3584, Loss: {'train/loss': 0.33749446272850037, 'train/grad_norm_pre_clip': 11.378189086914062, 'timing/data': 0.0060624959878623486, 'timing/model': 1.5774803259409964, 'learning_rate/qwen_vl_interface': 9.257290713022041e-06, 'learning_rate/action_model': 9.257290713022041e-05, 'epoch': 6.63}) +[2026-06-20 15:17:14,723][starVLA.training.train_starvla][INFO] - Step 3585, Loss: {'train/loss': 0.03562771901488304, 'train/grad_norm_pre_clip': 4.572016716003418, 'timing/data': 0.007346083060838282, 'timing/model': 1.2497832980006933, 'learning_rate/qwen_vl_interface': 9.255654775261112e-06, 'learning_rate/action_model': 9.255654775261112e-05, 'epoch': 6.63}) +[2026-06-20 15:17:19,859][starVLA.training.train_starvla][INFO] - Step 3586, Loss: {'train/loss': 0.04114537686109543, 'train/grad_norm_pre_clip': 3.231902837753296, 'timing/data': 0.006146066007204354, 'timing/model': 1.1795533798867837, 'learning_rate/qwen_vl_interface': 9.254017190926813e-06, 'learning_rate/action_model': 9.254017190926813e-05, 'epoch': 6.63}) +[2026-06-20 15:17:25,081][starVLA.training.train_starvla][INFO] - Step 3587, Loss: {'train/loss': 0.27550381422042847, 'train/grad_norm_pre_clip': 6.815603733062744, 'timing/data': 0.006175536895170808, 'timing/model': 1.210657846997492, 'learning_rate/qwen_vl_interface': 9.25237796069229e-06, 'learning_rate/action_model': 9.25237796069229e-05, 'epoch': 6.63}) +[2026-06-20 15:17:30,074][starVLA.training.train_starvla][INFO] - Step 3588, Loss: {'train/loss': 0.13675789535045624, 'train/grad_norm_pre_clip': 3.4271504878997803, 'timing/data': 0.006761995027773082, 'timing/model': 1.2532882349332795, 'learning_rate/qwen_vl_interface': 9.250737085231371e-06, 'learning_rate/action_model': 9.250737085231372e-05, 'epoch': 6.64}) +[2026-06-20 15:17:35,700][starVLA.training.train_starvla][INFO] - Step 3589, Loss: {'train/loss': 0.11382944881916046, 'train/grad_norm_pre_clip': 9.297930717468262, 'timing/data': 0.0232430559117347, 'timing/model': 1.273906736052595, 'learning_rate/qwen_vl_interface': 9.24909456521856e-06, 'learning_rate/action_model': 9.249094565218559e-05, 'epoch': 6.64}) +[2026-06-20 15:17:41,059][starVLA.training.train_starvla][INFO] - Step 3590, Loss: {'train/loss': 0.06281809508800507, 'train/grad_norm_pre_clip': 7.578958034515381, 'timing/data': 0.0064143650233745575, 'timing/model': 1.500633054994978, 'learning_rate/qwen_vl_interface': 9.247450401329027e-06, 'learning_rate/action_model': 9.247450401329027e-05, 'epoch': 6.64}) +[2026-06-20 15:17:46,138][starVLA.training.train_starvla][INFO] - Step 3591, Loss: {'train/loss': 0.3627660274505615, 'train/grad_norm_pre_clip': 12.823492050170898, 'timing/data': 0.006158706033602357, 'timing/model': 1.3904727098997682, 'learning_rate/qwen_vl_interface': 9.245804594238637e-06, 'learning_rate/action_model': 9.245804594238636e-05, 'epoch': 6.64}) +[2026-06-20 15:17:51,335][starVLA.training.train_starvla][INFO] - Step 3592, Loss: {'train/loss': 0.1619131863117218, 'train/grad_norm_pre_clip': 11.515377044677734, 'timing/data': 0.006280414992943406, 'timing/model': 1.1794243829790503, 'learning_rate/qwen_vl_interface': 9.244157144623909e-06, 'learning_rate/action_model': 9.244157144623909e-05, 'epoch': 6.64}) +[2026-06-20 15:17:56,332][starVLA.training.train_starvla][INFO] - Step 3593, Loss: {'train/loss': 0.09678216278553009, 'train/grad_norm_pre_clip': 8.22866153717041, 'timing/data': 0.006214414956048131, 'timing/model': 1.2671273960731924, 'learning_rate/qwen_vl_interface': 9.242508053162051e-06, 'learning_rate/action_model': 9.242508053162051e-05, 'epoch': 6.64}) +[2026-06-20 15:18:01,409][starVLA.training.train_starvla][INFO] - Step 3594, Loss: {'train/loss': 0.1937614232301712, 'train/grad_norm_pre_clip': 5.045252323150635, 'timing/data': 0.007702102069742978, 'timing/model': 1.2347674310440198, 'learning_rate/qwen_vl_interface': 9.240857320530944e-06, 'learning_rate/action_model': 9.240857320530943e-05, 'epoch': 6.65}) +[2026-06-20 15:18:06,396][starVLA.training.train_starvla][INFO] - Step 3595, Loss: {'train/loss': 0.24226360023021698, 'train/grad_norm_pre_clip': 8.431694030761719, 'timing/data': 0.006696465075947344, 'timing/model': 1.4280801359564066, 'learning_rate/qwen_vl_interface': 9.23920494740914e-06, 'learning_rate/action_model': 9.23920494740914e-05, 'epoch': 6.65}) +[2026-06-20 15:18:11,304][starVLA.training.train_starvla][INFO] - Step 3596, Loss: {'train/loss': 0.1096838191151619, 'train/grad_norm_pre_clip': 9.84489917755127, 'timing/data': 0.006222016061656177, 'timing/model': 1.3599791750311852, 'learning_rate/qwen_vl_interface': 9.237550934475868e-06, 'learning_rate/action_model': 9.237550934475867e-05, 'epoch': 6.65}) +[2026-06-20 15:18:16,083][starVLA.training.train_starvla][INFO] - Step 3597, Loss: {'train/loss': 0.11734078824520111, 'train/grad_norm_pre_clip': 14.76175308227539, 'timing/data': 0.0061555749271065, 'timing/model': 1.3142571109347045, 'learning_rate/qwen_vl_interface': 9.235895282411029e-06, 'learning_rate/action_model': 9.235895282411029e-05, 'epoch': 6.65}) +[2026-06-20 15:18:20,982][starVLA.training.train_starvla][INFO] - Step 3598, Loss: {'train/loss': 0.2983798384666443, 'train/grad_norm_pre_clip': 15.13048267364502, 'timing/data': 0.006465393933467567, 'timing/model': 1.1770262810168788, 'learning_rate/qwen_vl_interface': 9.2342379918952e-06, 'learning_rate/action_model': 9.2342379918952e-05, 'epoch': 6.65}) +[2026-06-20 15:18:25,791][starVLA.training.train_starvla][INFO] - Step 3599, Loss: {'train/loss': 0.20829005539417267, 'train/grad_norm_pre_clip': 10.915257453918457, 'timing/data': 0.0063526660669595, 'timing/model': 1.1812206299509853, 'learning_rate/qwen_vl_interface': 9.232579063609631e-06, 'learning_rate/action_model': 9.232579063609631e-05, 'epoch': 6.66}) +[2026-06-20 15:18:30,667][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 3600 +[2026-06-20 15:24:39,073][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 3600 +[2026-06-20 15:24:39,074][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 3600 +[2026-06-20 15:31:19,283][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 3600 +[2026-06-20 15:31:19,285][starVLA.training.train_starvla][INFO] - Step 3600, Loss: {'train/loss': 0.1352245956659317, 'train/grad_norm_pre_clip': 7.836069107055664, 'eval/loss': 0.6460870504379272, 'eval/latency_0/loss': 0.8204459547996521, 'eval/latency_1/loss': 0.9291410446166992, 'eval/latency_2/loss': 0.9157472848892212, 'eval/latency_3/loss': 0.4210902750492096, 'eval/latency_4/loss': 0.11732111126184464, 'eval/flappy/loss': 0.6460870504379272, 'eval/flappy/latency_0/loss': 0.8204459547996521, 'eval/flappy/latency_1/loss': 0.9291410446166992, 'eval/flappy/latency_2/loss': 0.9157472848892212, 'eval/flappy/latency_3/loss': 0.4210902750492096, 'eval/flappy/latency_4/loss': 0.11732111126184464, 'eval/flap_precision': 0.13953488372093023, 'eval/flap_recall': 0.08727272727272728, 'eval/flap_f1': 0.10738255033557048, 'eval/flap_rate_model': 0.04777777777777778, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.1073825503355705, 'eval/latency_0/flap_precision': 0.13953488372093023, 'eval/latency_0/flap_recall': 0.08727272727272728, 'eval/latency_0/flap_f1': 0.10738255033557048, 'eval/latency_0/flap_rate_model': 0.04777777777777778, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.1073825503355705, 'timing/data': 0.006951244082301855, 'timing/model': 1.1841752809705213, 'learning_rate/qwen_vl_interface': 9.230918498236243e-06, 'learning_rate/action_model': 9.230918498236244e-05, 'epoch': 6.66}) +[2026-06-20 15:31:19,285][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_3600_state +[2026-06-20 15:31:19,289][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-20 15:31:44,270][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_3600_state/model.safetensors +[2026-06-20 15:32:00,905][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_3600_state/optimizer.bin +[2026-06-20 15:32:00,906][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_3600_state/sampler.bin +[2026-06-20 15:32:00,906][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_3600_state/sampler_1.bin +[2026-06-20 15:32:00,907][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_3600_state/random_states_0.pkl +[2026-06-20 15:32:02,654][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3600 +[2026-06-20 15:32:02,654][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-20 15:32:02,668][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-20 15:32:07,426][starVLA.training.train_starvla][INFO] - Step 3601, Loss: {'train/loss': 0.1928839087486267, 'train/grad_norm_pre_clip': 10.095403671264648, 'timing/data': 0.00600107095669955, 'timing/model': 1.1747568589635193, 'learning_rate/qwen_vl_interface': 9.229256296457635e-06, 'learning_rate/action_model': 9.229256296457634e-05, 'epoch': 6.66}) +[2026-06-20 15:32:12,160][starVLA.training.train_starvla][INFO] - Step 3602, Loss: {'train/loss': 0.04925866425037384, 'train/grad_norm_pre_clip': 3.967911720275879, 'timing/data': 0.006094609969295561, 'timing/model': 1.1622205231105909, 'learning_rate/qwen_vl_interface': 9.22759245895707e-06, 'learning_rate/action_model': 9.227592458957071e-05, 'epoch': 6.66}) +[2026-06-20 15:32:17,040][starVLA.training.train_starvla][INFO] - Step 3603, Loss: {'train/loss': 0.35631388425827026, 'train/grad_norm_pre_clip': 15.390301704406738, 'timing/data': 0.005903771962039173, 'timing/model': 1.524852687958628, 'learning_rate/qwen_vl_interface': 9.225926986418494e-06, 'learning_rate/action_model': 9.225926986418494e-05, 'epoch': 6.66}) +[2026-06-20 15:32:21,747][starVLA.training.train_starvla][INFO] - Step 3604, Loss: {'train/loss': 0.2525824010372162, 'train/grad_norm_pre_clip': 23.491928100585938, 'timing/data': 0.006193690001964569, 'timing/model': 1.2926780119305477, 'learning_rate/qwen_vl_interface': 9.224259879526519e-06, 'learning_rate/action_model': 9.224259879526519e-05, 'epoch': 6.67}) +[2026-06-20 15:32:26,712][starVLA.training.train_starvla][INFO] - Step 3605, Loss: {'train/loss': 0.18444599211215973, 'train/grad_norm_pre_clip': 6.826834201812744, 'timing/data': 0.0061201510252431035, 'timing/model': 1.1687150449724868, 'learning_rate/qwen_vl_interface': 9.222591138966427e-06, 'learning_rate/action_model': 9.222591138966426e-05, 'epoch': 6.67}) +[2026-06-20 15:32:31,592][starVLA.training.train_starvla][INFO] - Step 3606, Loss: {'train/loss': 0.2322072982788086, 'train/grad_norm_pre_clip': 8.732542991638184, 'timing/data': 0.007472119992598891, 'timing/model': 1.2279095359845087, 'learning_rate/qwen_vl_interface': 9.220920765424175e-06, 'learning_rate/action_model': 9.220920765424174e-05, 'epoch': 6.67}) +[2026-06-20 15:32:36,474][starVLA.training.train_starvla][INFO] - Step 3607, Loss: {'train/loss': 0.09229252487421036, 'train/grad_norm_pre_clip': 6.68802547454834, 'timing/data': 0.008421880076639354, 'timing/model': 1.190293559920974, 'learning_rate/qwen_vl_interface': 9.219248759586392e-06, 'learning_rate/action_model': 9.219248759586391e-05, 'epoch': 6.67}) +[2026-06-20 15:32:41,480][starVLA.training.train_starvla][INFO] - Step 3608, Loss: {'train/loss': 0.2222660779953003, 'train/grad_norm_pre_clip': 8.754643440246582, 'timing/data': 0.006526560056954622, 'timing/model': 1.4398803730728105, 'learning_rate/qwen_vl_interface': 9.217575122140375e-06, 'learning_rate/action_model': 9.217575122140375e-05, 'epoch': 6.67}) +[2026-06-20 15:32:46,296][starVLA.training.train_starvla][INFO] - Step 3609, Loss: {'train/loss': 0.20627570152282715, 'train/grad_norm_pre_clip': 12.182596206665039, 'timing/data': 0.006587340030819178, 'timing/model': 1.394506986020133, 'learning_rate/qwen_vl_interface': 9.215899853774094e-06, 'learning_rate/action_model': 9.215899853774093e-05, 'epoch': 6.67}) +[2026-06-20 15:32:50,995][starVLA.training.train_starvla][INFO] - Step 3610, Loss: {'train/loss': 0.0954926460981369, 'train/grad_norm_pre_clip': 5.557653903961182, 'timing/data': 0.008003639988601208, 'timing/model': 1.191742143011652, 'learning_rate/qwen_vl_interface': 9.214222955176189e-06, 'learning_rate/action_model': 9.214222955176189e-05, 'epoch': 6.68}) +[2026-06-20 15:32:55,779][starVLA.training.train_starvla][INFO] - Step 3611, Loss: {'train/loss': 0.3583817481994629, 'train/grad_norm_pre_clip': 20.71999740600586, 'timing/data': 0.00652427994646132, 'timing/model': 1.182552641024813, 'learning_rate/qwen_vl_interface': 9.212544427035967e-06, 'learning_rate/action_model': 9.212544427035967e-05, 'epoch': 6.68}) +[2026-06-20 15:33:00,620][starVLA.training.train_starvla][INFO] - Step 3612, Loss: {'train/loss': 0.17160895466804504, 'train/grad_norm_pre_clip': 16.30912208557129, 'timing/data': 0.006026659975759685, 'timing/model': 1.1763118390226737, 'learning_rate/qwen_vl_interface': 9.21086427004341e-06, 'learning_rate/action_model': 9.210864270043411e-05, 'epoch': 6.68}) +[2026-06-20 15:33:05,559][starVLA.training.train_starvla][INFO] - Step 3613, Loss: {'train/loss': 0.0994853675365448, 'train/grad_norm_pre_clip': 5.8757004737854, 'timing/data': 0.006346768932417035, 'timing/model': 1.2638777419924736, 'learning_rate/qwen_vl_interface': 9.209182484889168e-06, 'learning_rate/action_model': 9.209182484889167e-05, 'epoch': 6.68}) +[2026-06-20 15:33:10,821][starVLA.training.train_starvla][INFO] - Step 3614, Loss: {'train/loss': 0.040997497737407684, 'train/grad_norm_pre_clip': 2.492914915084839, 'timing/data': 0.006063408916816115, 'timing/model': 1.5683822539867833, 'learning_rate/qwen_vl_interface': 9.207499072264557e-06, 'learning_rate/action_model': 9.207499072264558e-05, 'epoch': 6.68}) +[2026-06-20 15:33:15,543][starVLA.training.train_starvla][INFO] - Step 3615, Loss: {'train/loss': 0.0878167450428009, 'train/grad_norm_pre_clip': 11.828450202941895, 'timing/data': 0.005996599909849465, 'timing/model': 1.1786683030659333, 'learning_rate/qwen_vl_interface': 9.205814032861566e-06, 'learning_rate/action_model': 9.205814032861566e-05, 'epoch': 6.69}) +[2026-06-20 15:33:20,588][starVLA.training.train_starvla][INFO] - Step 3616, Loss: {'train/loss': 0.05427585542201996, 'train/grad_norm_pre_clip': 3.2125816345214844, 'timing/data': 0.00647672894410789, 'timing/model': 1.2232238370925188, 'learning_rate/qwen_vl_interface': 9.204127367372849e-06, 'learning_rate/action_model': 9.204127367372849e-05, 'epoch': 6.69}) +[2026-06-20 15:33:25,344][starVLA.training.train_starvla][INFO] - Step 3617, Loss: {'train/loss': 0.13773588836193085, 'train/grad_norm_pre_clip': 9.74302864074707, 'timing/data': 0.0081450289580971, 'timing/model': 1.1756999309873208, 'learning_rate/qwen_vl_interface': 9.202439076491732e-06, 'learning_rate/action_model': 9.202439076491732e-05, 'epoch': 6.69}) +[2026-06-20 15:33:30,485][starVLA.training.train_starvla][INFO] - Step 3618, Loss: {'train/loss': 0.44200441241264343, 'train/grad_norm_pre_clip': 42.14463806152344, 'timing/data': 0.006029719021171331, 'timing/model': 1.2185335430549458, 'learning_rate/qwen_vl_interface': 9.200749160912208e-06, 'learning_rate/action_model': 9.200749160912208e-05, 'epoch': 6.69}) +[2026-06-20 15:33:35,611][starVLA.training.train_starvla][INFO] - Step 3619, Loss: {'train/loss': 0.10778346657752991, 'train/grad_norm_pre_clip': 14.578831672668457, 'timing/data': 0.006101658917032182, 'timing/model': 1.3353602020069957, 'learning_rate/qwen_vl_interface': 9.199057621328937e-06, 'learning_rate/action_model': 9.199057621328935e-05, 'epoch': 6.69}) +[2026-06-20 15:33:40,544][starVLA.training.train_starvla][INFO] - Step 3620, Loss: {'train/loss': 0.08307112753391266, 'train/grad_norm_pre_clip': 6.357573986053467, 'timing/data': 0.0062234289944171906, 'timing/model': 1.5239869139622897, 'learning_rate/qwen_vl_interface': 9.197364458437244e-06, 'learning_rate/action_model': 9.197364458437245e-05, 'epoch': 6.69}) +[2026-06-20 15:33:45,722][starVLA.training.train_starvla][INFO] - Step 3621, Loss: {'train/loss': 0.583545446395874, 'train/grad_norm_pre_clip': 16.249847412109375, 'timing/data': 0.005993808037601411, 'timing/model': 1.5427723809843883, 'learning_rate/qwen_vl_interface': 9.19566967293313e-06, 'learning_rate/action_model': 9.19566967293313e-05, 'epoch': 6.7}) +[2026-06-20 15:33:50,387][starVLA.training.train_starvla][INFO] - Step 3622, Loss: {'train/loss': 0.1895880252122879, 'train/grad_norm_pre_clip': 15.01622486114502, 'timing/data': 0.005934978951700032, 'timing/model': 1.2072995089692995, 'learning_rate/qwen_vl_interface': 9.193973265513253e-06, 'learning_rate/action_model': 9.193973265513253e-05, 'epoch': 6.7}) +[2026-06-20 15:33:55,247][starVLA.training.train_starvla][INFO] - Step 3623, Loss: {'train/loss': 0.3347873389720917, 'train/grad_norm_pre_clip': 16.084260940551758, 'timing/data': 0.006048017996363342, 'timing/model': 1.1952372640371323, 'learning_rate/qwen_vl_interface': 9.192275236874946e-06, 'learning_rate/action_model': 9.192275236874946e-05, 'epoch': 6.7}) +[2026-06-20 15:34:00,112][starVLA.training.train_starvla][INFO] - Step 3624, Loss: {'train/loss': 0.11155693233013153, 'train/grad_norm_pre_clip': 8.263819694519043, 'timing/data': 0.006183768040500581, 'timing/model': 1.188520984025672, 'learning_rate/qwen_vl_interface': 9.190575587716199e-06, 'learning_rate/action_model': 9.190575587716199e-05, 'epoch': 6.7}) +[2026-06-20 15:34:05,020][starVLA.training.train_starvla][INFO] - Step 3625, Loss: {'train/loss': 0.12279034405946732, 'train/grad_norm_pre_clip': 13.732274055480957, 'timing/data': 0.006180228083394468, 'timing/model': 1.223369016079232, 'learning_rate/qwen_vl_interface': 9.188874318735677e-06, 'learning_rate/action_model': 9.188874318735676e-05, 'epoch': 6.7}) +[2026-06-20 15:34:10,011][starVLA.training.train_starvla][INFO] - Step 3626, Loss: {'train/loss': 0.055667389184236526, 'train/grad_norm_pre_clip': 8.192599296569824, 'timing/data': 0.006320999003946781, 'timing/model': 1.6173249329440296, 'learning_rate/qwen_vl_interface': 9.187171430632707e-06, 'learning_rate/action_model': 9.187171430632707e-05, 'epoch': 6.71}) +[2026-06-20 15:34:14,637][starVLA.training.train_starvla][INFO] - Step 3627, Loss: {'train/loss': 0.3814454674720764, 'train/grad_norm_pre_clip': 16.59626007080078, 'timing/data': 0.006464409059844911, 'timing/model': 1.1801906909095123, 'learning_rate/qwen_vl_interface': 9.185466924107283e-06, 'learning_rate/action_model': 9.185466924107282e-05, 'epoch': 6.71}) +[2026-06-20 15:34:19,516][starVLA.training.train_starvla][INFO] - Step 3628, Loss: {'train/loss': 0.30390068888664246, 'train/grad_norm_pre_clip': 17.329099655151367, 'timing/data': 0.006207857979461551, 'timing/model': 1.1917175580747426, 'learning_rate/qwen_vl_interface': 9.18376079986006e-06, 'learning_rate/action_model': 9.183760799860061e-05, 'epoch': 6.71}) +[2026-06-20 15:34:24,452][starVLA.training.train_starvla][INFO] - Step 3629, Loss: {'train/loss': 0.06433048099279404, 'train/grad_norm_pre_clip': 7.95141077041626, 'timing/data': 0.006180929020047188, 'timing/model': 1.2042105350410566, 'learning_rate/qwen_vl_interface': 9.182053058592365e-06, 'learning_rate/action_model': 9.182053058592365e-05, 'epoch': 6.71}) +[2026-06-20 15:34:29,317][starVLA.training.train_starvla][INFO] - Step 3630, Loss: {'train/loss': 0.07886946946382523, 'train/grad_norm_pre_clip': 8.01224422454834, 'timing/data': 0.008585478062741458, 'timing/model': 1.2590564070269465, 'learning_rate/qwen_vl_interface': 9.180343701006183e-06, 'learning_rate/action_model': 9.180343701006183e-05, 'epoch': 6.71}) +[2026-06-20 15:34:34,153][starVLA.training.train_starvla][INFO] - Step 3631, Loss: {'train/loss': 0.08913163840770721, 'train/grad_norm_pre_clip': 7.927730083465576, 'timing/data': 0.006388277979567647, 'timing/model': 1.1805220870301127, 'learning_rate/qwen_vl_interface': 9.17863272780417e-06, 'learning_rate/action_model': 9.178632727804168e-05, 'epoch': 6.72}) +[2026-06-20 15:34:39,141][starVLA.training.train_starvla][INFO] - Step 3632, Loss: {'train/loss': 0.09586123377084732, 'train/grad_norm_pre_clip': 7.542025566101074, 'timing/data': 0.00650308805052191, 'timing/model': 1.4340301509946585, 'learning_rate/qwen_vl_interface': 9.176920139689638e-06, 'learning_rate/action_model': 9.176920139689637e-05, 'epoch': 6.72}) +[2026-06-20 15:34:43,766][starVLA.training.train_starvla][INFO] - Step 3633, Loss: {'train/loss': 0.028782276436686516, 'train/grad_norm_pre_clip': 2.1855204105377197, 'timing/data': 0.005559989018365741, 'timing/model': 1.2408307569567114, 'learning_rate/qwen_vl_interface': 9.175205937366568e-06, 'learning_rate/action_model': 9.175205937366569e-05, 'epoch': 6.72}) +[2026-06-20 15:34:49,428][starVLA.training.train_starvla][INFO] - Step 3634, Loss: {'train/loss': 0.1247577965259552, 'train/grad_norm_pre_clip': 7.121351718902588, 'timing/data': 0.0061537070432677865, 'timing/model': 1.3023317399201915, 'learning_rate/qwen_vl_interface': 9.173490121539608e-06, 'learning_rate/action_model': 9.173490121539608e-05, 'epoch': 6.72}) +[2026-06-20 15:34:54,781][starVLA.training.train_starvla][INFO] - Step 3635, Loss: {'train/loss': 0.037487857043743134, 'train/grad_norm_pre_clip': 5.23213005065918, 'timing/data': 0.006647436995990574, 'timing/model': 1.220016292994842, 'learning_rate/qwen_vl_interface': 9.171772692914062e-06, 'learning_rate/action_model': 9.171772692914061e-05, 'epoch': 6.72}) +[2026-06-20 15:35:00,177][starVLA.training.train_starvla][INFO] - Step 3636, Loss: {'train/loss': 0.2515570819377899, 'train/grad_norm_pre_clip': 12.648019790649414, 'timing/data': 0.006488417042419314, 'timing/model': 1.3048168059904128, 'learning_rate/qwen_vl_interface': 9.1700536521959e-06, 'learning_rate/action_model': 9.170053652195899e-05, 'epoch': 6.72}) +[2026-06-20 15:35:05,378][starVLA.training.train_starvla][INFO] - Step 3637, Loss: {'train/loss': 0.02207268960773945, 'train/grad_norm_pre_clip': 1.4322940111160278, 'timing/data': 0.006443158024922013, 'timing/model': 1.1889316389570013, 'learning_rate/qwen_vl_interface': 9.168333000091754e-06, 'learning_rate/action_model': 9.168333000091754e-05, 'epoch': 6.73}) +[2026-06-20 15:35:10,267][starVLA.training.train_starvla][INFO] - Step 3638, Loss: {'train/loss': 0.4306042492389679, 'train/grad_norm_pre_clip': 17.264516830444336, 'timing/data': 0.007188156945630908, 'timing/model': 1.4613795620389283, 'learning_rate/qwen_vl_interface': 9.16661073730892e-06, 'learning_rate/action_model': 9.166610737308919e-05, 'epoch': 6.73}) +[2026-06-20 15:35:14,904][starVLA.training.train_starvla][INFO] - Step 3639, Loss: {'train/loss': 0.0952674001455307, 'train/grad_norm_pre_clip': 15.665936470031738, 'timing/data': 0.00614159693941474, 'timing/model': 1.1842410690151155, 'learning_rate/qwen_vl_interface': 9.164886864555355e-06, 'learning_rate/action_model': 9.164886864555355e-05, 'epoch': 6.73}) +[2026-06-20 15:35:20,006][starVLA.training.train_starvla][INFO] - Step 3640, Loss: {'train/loss': 0.20338894426822662, 'train/grad_norm_pre_clip': 15.156386375427246, 'timing/data': 0.006817416986450553, 'timing/model': 1.1850343310507014, 'learning_rate/qwen_vl_interface': 9.163161382539678e-06, 'learning_rate/action_model': 9.163161382539678e-05, 'epoch': 6.73}) +[2026-06-20 15:35:25,008][starVLA.training.train_starvla][INFO] - Step 3641, Loss: {'train/loss': 0.02829648368060589, 'train/grad_norm_pre_clip': 3.3861727714538574, 'timing/data': 0.00625706696882844, 'timing/model': 1.1814544150838628, 'learning_rate/qwen_vl_interface': 9.16143429197117e-06, 'learning_rate/action_model': 9.16143429197117e-05, 'epoch': 6.73}) +[2026-06-20 15:35:30,060][starVLA.training.train_starvla][INFO] - Step 3642, Loss: {'train/loss': 0.10967008024454117, 'train/grad_norm_pre_clip': 15.142308235168457, 'timing/data': 0.006280296947807074, 'timing/model': 1.2726450399495661, 'learning_rate/qwen_vl_interface': 9.159705593559773e-06, 'learning_rate/action_model': 9.159705593559773e-05, 'epoch': 6.74}) +[2026-06-20 15:35:34,993][starVLA.training.train_starvla][INFO] - Step 3643, Loss: {'train/loss': 0.28302469849586487, 'train/grad_norm_pre_clip': 12.236518859863281, 'timing/data': 0.006268527009524405, 'timing/model': 1.1779713629512116, 'learning_rate/qwen_vl_interface': 9.157975288016089e-06, 'learning_rate/action_model': 9.157975288016088e-05, 'epoch': 6.74}) +[2026-06-20 15:35:40,023][starVLA.training.train_starvla][INFO] - Step 3644, Loss: {'train/loss': 0.12567313015460968, 'train/grad_norm_pre_clip': 15.112638473510742, 'timing/data': 0.017925889929756522, 'timing/model': 1.5288958339951932, 'learning_rate/qwen_vl_interface': 9.156243376051381e-06, 'learning_rate/action_model': 9.156243376051381e-05, 'epoch': 6.74}) +[2026-06-20 15:35:44,840][starVLA.training.train_starvla][INFO] - Step 3645, Loss: {'train/loss': 0.36311259865760803, 'train/grad_norm_pre_clip': 20.6561279296875, 'timing/data': 0.006236547022126615, 'timing/model': 1.1826003780588508, 'learning_rate/qwen_vl_interface': 9.154509858377574e-06, 'learning_rate/action_model': 9.154509858377574e-05, 'epoch': 6.74}) +[2026-06-20 15:35:49,889][starVLA.training.train_starvla][INFO] - Step 3646, Loss: {'train/loss': 0.17970401048660278, 'train/grad_norm_pre_clip': 10.255823135375977, 'timing/data': 0.009569744928739965, 'timing/model': 1.2663665929576382, 'learning_rate/qwen_vl_interface': 9.152774735707248e-06, 'learning_rate/action_model': 9.152774735707247e-05, 'epoch': 6.74}) +[2026-06-20 15:35:54,833][starVLA.training.train_starvla][INFO] - Step 3647, Loss: {'train/loss': 0.2528822124004364, 'train/grad_norm_pre_clip': 11.665087699890137, 'timing/data': 0.005966595956124365, 'timing/model': 1.2124342189636081, 'learning_rate/qwen_vl_interface': 9.15103800875365e-06, 'learning_rate/action_model': 9.15103800875365e-05, 'epoch': 6.74}) +[2026-06-20 15:35:59,781][starVLA.training.train_starvla][INFO] - Step 3648, Loss: {'train/loss': 0.16538332402706146, 'train/grad_norm_pre_clip': 13.890137672424316, 'timing/data': 0.005833386909216642, 'timing/model': 1.2502197499852628, 'learning_rate/qwen_vl_interface': 9.149299678230681e-06, 'learning_rate/action_model': 9.149299678230681e-05, 'epoch': 6.75}) +[2026-06-20 15:36:04,781][starVLA.training.train_starvla][INFO] - Step 3649, Loss: {'train/loss': 0.2781918942928314, 'train/grad_norm_pre_clip': 21.274080276489258, 'timing/data': 0.007771495031192899, 'timing/model': 1.2164703039452434, 'learning_rate/qwen_vl_interface': 9.147559744852905e-06, 'learning_rate/action_model': 9.147559744852905e-05, 'epoch': 6.75}) +[2026-06-20 15:36:10,219][starVLA.training.train_starvla][INFO] - Step 3650, Loss: {'train/loss': 0.11959919333457947, 'train/grad_norm_pre_clip': 5.910766124725342, 'timing/data': 0.006149146007373929, 'timing/model': 1.4796031879959628, 'learning_rate/qwen_vl_interface': 9.14581820933554e-06, 'learning_rate/action_model': 9.14581820933554e-05, 'epoch': 6.75}) +[2026-06-20 15:36:15,514][starVLA.training.train_starvla][INFO] - Step 3651, Loss: {'train/loss': 0.13007259368896484, 'train/grad_norm_pre_clip': 15.314351081848145, 'timing/data': 0.005989124998450279, 'timing/model': 1.5879729759180918, 'learning_rate/qwen_vl_interface': 9.14407507239447e-06, 'learning_rate/action_model': 9.14407507239447e-05, 'epoch': 6.75}) +[2026-06-20 15:36:20,265][starVLA.training.train_starvla][INFO] - Step 3652, Loss: {'train/loss': 0.3878692090511322, 'train/grad_norm_pre_clip': 26.86692237854004, 'timing/data': 0.005632455926388502, 'timing/model': 1.2564351009204984, 'learning_rate/qwen_vl_interface': 9.142330334746227e-06, 'learning_rate/action_model': 9.142330334746227e-05, 'epoch': 6.75}) +[2026-06-20 15:36:25,236][starVLA.training.train_starvla][INFO] - Step 3653, Loss: {'train/loss': 0.24418631196022034, 'train/grad_norm_pre_clip': 15.185637474060059, 'timing/data': 0.006070396048016846, 'timing/model': 1.2044335769023746, 'learning_rate/qwen_vl_interface': 9.14058399710801e-06, 'learning_rate/action_model': 9.14058399710801e-05, 'epoch': 6.76}) +[2026-06-20 15:36:30,489][starVLA.training.train_starvla][INFO] - Step 3654, Loss: {'train/loss': 0.21301238238811493, 'train/grad_norm_pre_clip': 12.09538745880127, 'timing/data': 0.0072059560334309936, 'timing/model': 1.2222099030623212, 'learning_rate/qwen_vl_interface': 9.138836060197673e-06, 'learning_rate/action_model': 9.138836060197672e-05, 'epoch': 6.76}) +[2026-06-20 15:36:35,968][starVLA.training.train_starvla][INFO] - Step 3655, Loss: {'train/loss': 0.20998886227607727, 'train/grad_norm_pre_clip': 13.101238250732422, 'timing/data': 0.007720025023445487, 'timing/model': 1.4196810730500147, 'learning_rate/qwen_vl_interface': 9.137086524733725e-06, 'learning_rate/action_model': 9.137086524733724e-05, 'epoch': 6.76}) +[2026-06-20 15:36:41,011][starVLA.training.train_starvla][INFO] - Step 3656, Loss: {'train/loss': 0.3333814740180969, 'train/grad_norm_pre_clip': 11.308058738708496, 'timing/data': 0.005623106029815972, 'timing/model': 1.223859597928822, 'learning_rate/qwen_vl_interface': 9.135335391435336e-06, 'learning_rate/action_model': 9.135335391435335e-05, 'epoch': 6.76}) +[2026-06-20 15:36:46,375][starVLA.training.train_starvla][INFO] - Step 3657, Loss: {'train/loss': 0.08801637589931488, 'train/grad_norm_pre_clip': 8.062482833862305, 'timing/data': 0.005784025997854769, 'timing/model': 1.3922994369640946, 'learning_rate/qwen_vl_interface': 9.13358266102233e-06, 'learning_rate/action_model': 9.133582661022328e-05, 'epoch': 6.76}) +[2026-06-20 15:36:51,192][starVLA.training.train_starvla][INFO] - Step 3658, Loss: {'train/loss': 0.22604960203170776, 'train/grad_norm_pre_clip': 16.323781967163086, 'timing/data': 0.0062226951122283936, 'timing/model': 1.205826479010284, 'learning_rate/qwen_vl_interface': 9.131828334215185e-06, 'learning_rate/action_model': 9.131828334215185e-05, 'epoch': 6.76}) +[2026-06-20 15:36:56,172][starVLA.training.train_starvla][INFO] - Step 3659, Loss: {'train/loss': 0.04311297461390495, 'train/grad_norm_pre_clip': 4.271501064300537, 'timing/data': 0.005564645980484784, 'timing/model': 1.188564357929863, 'learning_rate/qwen_vl_interface': 9.130072411735043e-06, 'learning_rate/action_model': 9.130072411735043e-05, 'epoch': 6.77}) +[2026-06-20 15:37:01,029][starVLA.training.train_starvla][INFO] - Step 3660, Loss: {'train/loss': 0.09705265611410141, 'train/grad_norm_pre_clip': 12.878371238708496, 'timing/data': 0.006251135026104748, 'timing/model': 1.1760275029810145, 'learning_rate/qwen_vl_interface': 9.128314894303693e-06, 'learning_rate/action_model': 9.128314894303693e-05, 'epoch': 6.77}) +[2026-06-20 15:37:05,908][starVLA.training.train_starvla][INFO] - Step 3661, Loss: {'train/loss': 0.2612088918685913, 'train/grad_norm_pre_clip': 11.554611206054688, 'timing/data': 0.006486486061476171, 'timing/model': 1.1848800929728895, 'learning_rate/qwen_vl_interface': 9.126555782643592e-06, 'learning_rate/action_model': 9.126555782643591e-05, 'epoch': 6.77}) +[2026-06-20 15:37:10,776][starVLA.training.train_starvla][INFO] - Step 3662, Loss: {'train/loss': 0.1719789355993271, 'train/grad_norm_pre_clip': 11.27314281463623, 'timing/data': 0.006360423984006047, 'timing/model': 1.4631679900921881, 'learning_rate/qwen_vl_interface': 9.124795077477835e-06, 'learning_rate/action_model': 9.124795077477834e-05, 'epoch': 6.77}) +[2026-06-20 15:37:15,552][starVLA.training.train_starvla][INFO] - Step 3663, Loss: {'train/loss': 0.21390584111213684, 'train/grad_norm_pre_clip': 12.746817588806152, 'timing/data': 0.0060538859106600285, 'timing/model': 1.2855560060124844, 'learning_rate/qwen_vl_interface': 9.123032779530186e-06, 'learning_rate/action_model': 9.123032779530187e-05, 'epoch': 6.77}) +[2026-06-20 15:37:20,497][starVLA.training.train_starvla][INFO] - Step 3664, Loss: {'train/loss': 0.12491431832313538, 'train/grad_norm_pre_clip': 10.653339385986328, 'timing/data': 0.006490744999609888, 'timing/model': 1.1813250969862565, 'learning_rate/qwen_vl_interface': 9.12126888952506e-06, 'learning_rate/action_model': 9.12126888952506e-05, 'epoch': 6.78}) +[2026-06-20 15:37:25,521][starVLA.training.train_starvla][INFO] - Step 3665, Loss: {'train/loss': 0.06593064963817596, 'train/grad_norm_pre_clip': 6.357564449310303, 'timing/data': 0.008966512978076935, 'timing/model': 1.2339730790117756, 'learning_rate/qwen_vl_interface': 9.119503408187521e-06, 'learning_rate/action_model': 9.119503408187522e-05, 'epoch': 6.78}) +[2026-06-20 15:37:30,384][starVLA.training.train_starvla][INFO] - Step 3666, Loss: {'train/loss': 0.32275909185409546, 'train/grad_norm_pre_clip': 11.703073501586914, 'timing/data': 0.00614046398550272, 'timing/model': 1.2017906429246068, 'learning_rate/qwen_vl_interface': 9.117736336243299e-06, 'learning_rate/action_model': 9.117736336243298e-05, 'epoch': 6.78}) +[2026-06-20 15:37:35,372][starVLA.training.train_starvla][INFO] - Step 3667, Loss: {'train/loss': 0.2317340075969696, 'train/grad_norm_pre_clip': 20.157655715942383, 'timing/data': 0.005717315012589097, 'timing/model': 1.2282305069966242, 'learning_rate/qwen_vl_interface': 9.115967674418765e-06, 'learning_rate/action_model': 9.115967674418765e-05, 'epoch': 6.78}) +[2026-06-20 15:37:40,332][starVLA.training.train_starvla][INFO] - Step 3668, Loss: {'train/loss': 0.3557969033718109, 'train/grad_norm_pre_clip': 13.513304710388184, 'timing/data': 0.006093695061281323, 'timing/model': 1.4516681409440935, 'learning_rate/qwen_vl_interface': 9.114197423440949e-06, 'learning_rate/action_model': 9.11419742344095e-05, 'epoch': 6.78}) +[2026-06-20 15:37:45,292][starVLA.training.train_starvla][INFO] - Step 3669, Loss: {'train/loss': 0.08604246377944946, 'train/grad_norm_pre_clip': 9.335515975952148, 'timing/data': 0.005784125067293644, 'timing/model': 1.2612447920255363, 'learning_rate/qwen_vl_interface': 9.112425584037536e-06, 'learning_rate/action_model': 9.112425584037536e-05, 'epoch': 6.79}) +[2026-06-20 15:37:50,212][starVLA.training.train_starvla][INFO] - Step 3670, Loss: {'train/loss': 0.23431551456451416, 'train/grad_norm_pre_clip': 8.711603164672852, 'timing/data': 0.0060118649853393435, 'timing/model': 1.202990829013288, 'learning_rate/qwen_vl_interface': 9.110652156936863e-06, 'learning_rate/action_model': 9.110652156936863e-05, 'epoch': 6.79}) +[2026-06-20 15:37:55,096][starVLA.training.train_starvla][INFO] - Step 3671, Loss: {'train/loss': 0.11077509820461273, 'train/grad_norm_pre_clip': 6.473695755004883, 'timing/data': 0.007825692999176681, 'timing/model': 1.235081837978214, 'learning_rate/qwen_vl_interface': 9.108877142867918e-06, 'learning_rate/action_model': 9.108877142867918e-05, 'epoch': 6.79}) +[2026-06-20 15:38:00,013][starVLA.training.train_starvla][INFO] - Step 3672, Loss: {'train/loss': 0.37896305322647095, 'train/grad_norm_pre_clip': 26.79150390625, 'timing/data': 0.006098745972849429, 'timing/model': 1.1933513029944152, 'learning_rate/qwen_vl_interface': 9.107100542560342e-06, 'learning_rate/action_model': 9.107100542560342e-05, 'epoch': 6.79}) +[2026-06-20 15:38:04,992][starVLA.training.train_starvla][INFO] - Step 3673, Loss: {'train/loss': 0.05228741094470024, 'train/grad_norm_pre_clip': 4.77401065826416, 'timing/data': 0.006287214928306639, 'timing/model': 1.186952137039043, 'learning_rate/qwen_vl_interface': 9.105322356744429e-06, 'learning_rate/action_model': 9.105322356744429e-05, 'epoch': 6.79}) +[2026-06-20 15:38:09,988][starVLA.training.train_starvla][INFO] - Step 3674, Loss: {'train/loss': 0.2884938418865204, 'train/grad_norm_pre_clip': 13.920431137084961, 'timing/data': 0.005633795051835477, 'timing/model': 1.5374680398963392, 'learning_rate/qwen_vl_interface': 9.103542586151121e-06, 'learning_rate/action_model': 9.10354258615112e-05, 'epoch': 6.79}) +[2026-06-20 15:38:14,808][starVLA.training.train_starvla][INFO] - Step 3675, Loss: {'train/loss': 0.14726921916007996, 'train/grad_norm_pre_clip': 7.753912448883057, 'timing/data': 0.005682035000063479, 'timing/model': 1.191890837973915, 'learning_rate/qwen_vl_interface': 9.101761231512017e-06, 'learning_rate/action_model': 9.101761231512017e-05, 'epoch': 6.8}) +[2026-06-20 15:38:20,240][starVLA.training.train_starvla][INFO] - Step 3676, Loss: {'train/loss': 0.06542831659317017, 'train/grad_norm_pre_clip': 5.691974639892578, 'timing/data': 0.006146733998320997, 'timing/model': 1.743688193964772, 'learning_rate/qwen_vl_interface': 9.099978293559365e-06, 'learning_rate/action_model': 9.099978293559366e-05, 'epoch': 6.8}) +[2026-06-20 15:38:25,579][starVLA.training.train_starvla][INFO] - Step 3677, Loss: {'train/loss': 0.15019211173057556, 'train/grad_norm_pre_clip': 8.00128173828125, 'timing/data': 0.006472783046774566, 'timing/model': 1.3534294180572033, 'learning_rate/qwen_vl_interface': 9.098193773026065e-06, 'learning_rate/action_model': 9.098193773026064e-05, 'epoch': 6.8}) +[2026-06-20 15:38:30,505][starVLA.training.train_starvla][INFO] - Step 3678, Loss: {'train/loss': 0.2002420872449875, 'train/grad_norm_pre_clip': 11.518580436706543, 'timing/data': 0.006773454020731151, 'timing/model': 1.2690048340009525, 'learning_rate/qwen_vl_interface': 9.096407670645663e-06, 'learning_rate/action_model': 9.096407670645662e-05, 'epoch': 6.8}) +[2026-06-20 15:38:35,622][starVLA.training.train_starvla][INFO] - Step 3679, Loss: {'train/loss': 0.20009569823741913, 'train/grad_norm_pre_clip': 5.985422134399414, 'timing/data': 0.0062564139952883124, 'timing/model': 1.4292501169256866, 'learning_rate/qwen_vl_interface': 9.094619987152357e-06, 'learning_rate/action_model': 9.094619987152358e-05, 'epoch': 6.8}) +[2026-06-20 15:38:40,594][starVLA.training.train_starvla][INFO] - Step 3680, Loss: {'train/loss': 0.06354337930679321, 'train/grad_norm_pre_clip': 6.173710346221924, 'timing/data': 0.006521663046441972, 'timing/model': 1.3573621339164674, 'learning_rate/qwen_vl_interface': 9.092830723281e-06, 'learning_rate/action_model': 9.092830723281e-05, 'epoch': 6.81}) +[2026-06-20 15:38:45,572][starVLA.training.train_starvla][INFO] - Step 3681, Loss: {'train/loss': 0.08555340021848679, 'train/grad_norm_pre_clip': 5.847056865692139, 'timing/data': 0.007396673085168004, 'timing/model': 1.3427281558979303, 'learning_rate/qwen_vl_interface': 9.09103987976709e-06, 'learning_rate/action_model': 9.09103987976709e-05, 'epoch': 6.81}) +[2026-06-20 15:38:50,635][starVLA.training.train_starvla][INFO] - Step 3682, Loss: {'train/loss': 0.12817354500293732, 'train/grad_norm_pre_clip': 8.795022010803223, 'timing/data': 0.007091023027896881, 'timing/model': 1.2374375580111519, 'learning_rate/qwen_vl_interface': 9.089247457346774e-06, 'learning_rate/action_model': 9.089247457346773e-05, 'epoch': 6.81}) +[2026-06-20 15:38:55,602][starVLA.training.train_starvla][INFO] - Step 3683, Loss: {'train/loss': 0.38847485184669495, 'train/grad_norm_pre_clip': 16.152538299560547, 'timing/data': 0.0006753990892320871, 'timing/model': 1.1850906290346757, 'learning_rate/qwen_vl_interface': 9.087453456756849e-06, 'learning_rate/action_model': 9.087453456756849e-05, 'epoch': 6.81}) +[2026-06-20 15:39:00,559][starVLA.training.train_starvla][INFO] - Step 3684, Loss: {'train/loss': 0.27297547459602356, 'train/grad_norm_pre_clip': 11.820740699768066, 'timing/data': 0.0005320790223777294, 'timing/model': 1.2423463389277458, 'learning_rate/qwen_vl_interface': 9.085657878734759e-06, 'learning_rate/action_model': 9.085657878734758e-05, 'epoch': 6.81}) +[2026-06-20 15:39:05,495][starVLA.training.train_starvla][INFO] - Step 3685, Loss: {'train/loss': 0.173561692237854, 'train/grad_norm_pre_clip': 11.548561096191406, 'timing/data': 0.0007049590349197388, 'timing/model': 1.4523241620045155, 'learning_rate/qwen_vl_interface': 9.083860724018603e-06, 'learning_rate/action_model': 9.083860724018602e-05, 'epoch': 6.81}) +[2026-06-20 15:39:10,310][starVLA.training.train_starvla][INFO] - Step 3686, Loss: {'train/loss': 0.35153821110725403, 'train/grad_norm_pre_clip': 12.388704299926758, 'timing/data': 0.0005479400279000401, 'timing/model': 1.4143502380466089, 'learning_rate/qwen_vl_interface': 9.08206199334712e-06, 'learning_rate/action_model': 9.082061993347119e-05, 'epoch': 6.82}) +[2026-06-20 15:39:14,877][starVLA.training.train_starvla][INFO] - Step 3687, Loss: {'train/loss': 0.23000912368297577, 'train/grad_norm_pre_clip': 9.62258529663086, 'timing/data': 0.0005624790210276842, 'timing/model': 1.1964487050427124, 'learning_rate/qwen_vl_interface': 9.0802616874597e-06, 'learning_rate/action_model': 9.0802616874597e-05, 'epoch': 6.82}) +[2026-06-20 15:39:19,700][starVLA.training.train_starvla][INFO] - Step 3688, Loss: {'train/loss': 0.25693950057029724, 'train/grad_norm_pre_clip': 6.716993808746338, 'timing/data': 0.0005953800864517689, 'timing/model': 1.1984436810016632, 'learning_rate/qwen_vl_interface': 9.078459807096385e-06, 'learning_rate/action_model': 9.078459807096385e-05, 'epoch': 6.82}) +[2026-06-20 15:39:24,574][starVLA.training.train_starvla][INFO] - Step 3689, Loss: {'train/loss': 0.22375772893428802, 'train/grad_norm_pre_clip': 9.902831077575684, 'timing/data': 0.0005699590547010303, 'timing/model': 1.1990516110090539, 'learning_rate/qwen_vl_interface': 9.076656352997854e-06, 'learning_rate/action_model': 9.076656352997853e-05, 'epoch': 6.82}) +[2026-06-20 15:39:29,402][starVLA.training.train_starvla][INFO] - Step 3690, Loss: {'train/loss': 0.08779607713222504, 'train/grad_norm_pre_clip': 4.2954182624816895, 'timing/data': 0.000575389014557004, 'timing/model': 1.184259927016683, 'learning_rate/qwen_vl_interface': 9.074851325905443e-06, 'learning_rate/action_model': 9.074851325905442e-05, 'epoch': 6.82}) +[2026-06-20 15:39:34,197][starVLA.training.train_starvla][INFO] - Step 3691, Loss: {'train/loss': 0.0671328753232956, 'train/grad_norm_pre_clip': 2.739262342453003, 'timing/data': 0.000706399092450738, 'timing/model': 1.4278027929831296, 'learning_rate/qwen_vl_interface': 9.073044726561128e-06, 'learning_rate/action_model': 9.073044726561129e-05, 'epoch': 6.83}) +[2026-06-20 15:39:39,182][starVLA.training.train_starvla][INFO] - Step 3692, Loss: {'train/loss': 0.18536652624607086, 'train/grad_norm_pre_clip': 9.597211837768555, 'timing/data': 0.0005182389868423343, 'timing/model': 1.49762938299682, 'learning_rate/qwen_vl_interface': 9.071236555707536e-06, 'learning_rate/action_model': 9.071236555707537e-05, 'epoch': 6.83}) +[2026-06-20 15:39:43,895][starVLA.training.train_starvla][INFO] - Step 3693, Loss: {'train/loss': 0.07019809633493423, 'train/grad_norm_pre_clip': 4.219405651092529, 'timing/data': 0.0005120689747855067, 'timing/model': 1.2254243809729815, 'learning_rate/qwen_vl_interface': 9.069426814087937e-06, 'learning_rate/action_model': 9.069426814087937e-05, 'epoch': 6.83}) +[2026-06-20 15:39:48,734][starVLA.training.train_starvla][INFO] - Step 3694, Loss: {'train/loss': 0.11371705681085587, 'train/grad_norm_pre_clip': 12.747726440429688, 'timing/data': 0.0005351689178496599, 'timing/model': 1.180030420073308, 'learning_rate/qwen_vl_interface': 9.067615502446247e-06, 'learning_rate/action_model': 9.067615502446246e-05, 'epoch': 6.83}) +[2026-06-20 15:39:53,768][starVLA.training.train_starvla][INFO] - Step 3695, Loss: {'train/loss': 0.09996519982814789, 'train/grad_norm_pre_clip': 7.820932865142822, 'timing/data': 0.0005869900342077017, 'timing/model': 1.273796456051059, 'learning_rate/qwen_vl_interface': 9.06580262152703e-06, 'learning_rate/action_model': 9.065802621527028e-05, 'epoch': 6.83}) +[2026-06-20 15:39:58,724][starVLA.training.train_starvla][INFO] - Step 3696, Loss: {'train/loss': 0.24679329991340637, 'train/grad_norm_pre_clip': 11.913620948791504, 'timing/data': 0.0005113499937579036, 'timing/model': 1.2779067909577861, 'learning_rate/qwen_vl_interface': 9.063988172075486e-06, 'learning_rate/action_model': 9.063988172075486e-05, 'epoch': 6.84}) +[2026-06-20 15:40:03,672][starVLA.training.train_starvla][INFO] - Step 3697, Loss: {'train/loss': 0.04284385219216347, 'train/grad_norm_pre_clip': 3.102586030960083, 'timing/data': 0.0008214390836656094, 'timing/model': 1.5055909139337018, 'learning_rate/qwen_vl_interface': 9.062172154837475e-06, 'learning_rate/action_model': 9.062172154837475e-05, 'epoch': 6.84}) +[2026-06-20 15:40:08,449][starVLA.training.train_starvla][INFO] - Step 3698, Loss: {'train/loss': 0.05113670602440834, 'train/grad_norm_pre_clip': 6.030794143676758, 'timing/data': 0.000655268901027739, 'timing/model': 1.3089509649435058, 'learning_rate/qwen_vl_interface': 9.060354570559487e-06, 'learning_rate/action_model': 9.060354570559487e-05, 'epoch': 6.84}) +[2026-06-20 15:40:13,613][starVLA.training.train_starvla][INFO] - Step 3699, Loss: {'train/loss': 0.16795045137405396, 'train/grad_norm_pre_clip': 5.601712226867676, 'timing/data': 0.0005460790125653148, 'timing/model': 1.2806887269252911, 'learning_rate/qwen_vl_interface': 9.058535419988667e-06, 'learning_rate/action_model': 9.058535419988667e-05, 'epoch': 6.84}) +[2026-06-20 15:40:18,676][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 3700 +[2026-06-20 15:46:18,239][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 3700 +[2026-06-20 15:46:18,240][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 3700 +[2026-06-20 15:53:11,526][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 3700 +[2026-06-20 15:53:11,528][starVLA.training.train_starvla][INFO] - Step 3700, Loss: {'train/loss': 0.16339410841464996, 'train/grad_norm_pre_clip': 6.466350555419922, 'eval/loss': 0.5426910519599915, 'eval/latency_0/loss': 0.6505677700042725, 'eval/latency_1/loss': 0.8124854564666748, 'eval/latency_2/loss': 0.7908450365066528, 'eval/latency_3/loss': 0.34226498007774353, 'eval/latency_4/loss': 0.0895874947309494, 'eval/flappy/loss': 0.5426910519599915, 'eval/flappy/latency_0/loss': 0.6505677700042725, 'eval/flappy/latency_1/loss': 0.8124854564666748, 'eval/flappy/latency_2/loss': 0.7908450365066528, 'eval/flappy/latency_3/loss': 0.34226498007774353, 'eval/flappy/latency_4/loss': 0.0895874947309494, 'eval/flap_precision': 0.1987179487179487, 'eval/flap_recall': 0.11272727272727273, 'eval/flap_f1': 0.14385150812064965, 'eval/flap_rate_model': 0.043333333333333335, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.14385150812064965, 'eval/latency_0/flap_precision': 0.1987179487179487, 'eval/latency_0/flap_recall': 0.11272727272727273, 'eval/latency_0/flap_f1': 0.14385150812064965, 'eval/latency_0/flap_rate_model': 0.043333333333333335, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.14385150812064965, 'timing/data': 0.0005360000068321824, 'timing/model': 1.2412250220077112, 'learning_rate/qwen_vl_interface': 9.056714703872794e-06, 'learning_rate/action_model': 9.056714703872794e-05, 'epoch': 6.84}) +[2026-06-20 15:53:11,528][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_3700_state +[2026-06-20 15:53:11,533][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-20 15:53:18,900][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_3700_state/model.safetensors +[2026-06-20 15:53:32,699][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_3700_state/optimizer.bin +[2026-06-20 15:53:32,699][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_3700_state/sampler.bin +[2026-06-20 15:53:32,700][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_3700_state/sampler_1.bin +[2026-06-20 15:53:32,701][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_3700_state/random_states_0.pkl +[2026-06-20 15:53:35,221][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3700 +[2026-06-20 15:53:35,222][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-20 15:53:35,235][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-20 15:53:40,057][starVLA.training.train_starvla][INFO] - Step 3701, Loss: {'train/loss': 0.3972029387950897, 'train/grad_norm_pre_clip': 15.935707092285156, 'timing/data': 0.0005555689567700028, 'timing/model': 1.2136217129882425, 'learning_rate/qwen_vl_interface': 9.0548924229603e-06, 'learning_rate/action_model': 9.054892422960301e-05, 'epoch': 6.84}) +[2026-06-20 15:53:44,831][starVLA.training.train_starvla][INFO] - Step 3702, Loss: {'train/loss': 0.17094244062900543, 'train/grad_norm_pre_clip': 8.938023567199707, 'timing/data': 0.0005338700721040368, 'timing/model': 1.2326583960093558, 'learning_rate/qwen_vl_interface': 9.053068578000256e-06, 'learning_rate/action_model': 9.053068578000256e-05, 'epoch': 6.85}) +[2026-06-20 15:53:49,513][starVLA.training.train_starvla][INFO] - Step 3703, Loss: {'train/loss': 0.07408047467470169, 'train/grad_norm_pre_clip': 4.847179889678955, 'timing/data': 0.0005803389940410852, 'timing/model': 1.347028866992332, 'learning_rate/qwen_vl_interface': 9.051243169742372e-06, 'learning_rate/action_model': 9.051243169742372e-05, 'epoch': 6.85}) +[2026-06-20 15:53:54,219][starVLA.training.train_starvla][INFO] - Step 3704, Loss: {'train/loss': 0.09402991831302643, 'train/grad_norm_pre_clip': 7.465030670166016, 'timing/data': 0.0006072490941733122, 'timing/model': 1.1625328999944031, 'learning_rate/qwen_vl_interface': 9.049416198937008e-06, 'learning_rate/action_model': 9.049416198937007e-05, 'epoch': 6.85}) +[2026-06-20 15:53:58,986][starVLA.training.train_starvla][INFO] - Step 3705, Loss: {'train/loss': 0.03823089972138405, 'train/grad_norm_pre_clip': 7.676388740539551, 'timing/data': 0.0005181089509278536, 'timing/model': 1.1647611639928073, 'learning_rate/qwen_vl_interface': 9.047587666335159e-06, 'learning_rate/action_model': 9.047587666335159e-05, 'epoch': 6.85}) +[2026-06-20 15:54:03,780][starVLA.training.train_starvla][INFO] - Step 3706, Loss: {'train/loss': 0.08350737392902374, 'train/grad_norm_pre_clip': 8.211663246154785, 'timing/data': 0.0007659289985895157, 'timing/model': 1.1865142630413175, 'learning_rate/qwen_vl_interface': 9.045757572688469e-06, 'learning_rate/action_model': 9.04575757268847e-05, 'epoch': 6.85}) +[2026-06-20 15:54:08,558][starVLA.training.train_starvla][INFO] - Step 3707, Loss: {'train/loss': 0.08289063721895218, 'train/grad_norm_pre_clip': 15.421634674072266, 'timing/data': 0.0005786489928141236, 'timing/model': 1.1696121520362794, 'learning_rate/qwen_vl_interface': 9.043925918749218e-06, 'learning_rate/action_model': 9.043925918749218e-05, 'epoch': 6.86}) +[2026-06-20 15:54:13,367][starVLA.training.train_starvla][INFO] - Step 3708, Loss: {'train/loss': 0.04527299851179123, 'train/grad_norm_pre_clip': 6.94079065322876, 'timing/data': 0.0005552690709009767, 'timing/model': 1.4154531629756093, 'learning_rate/qwen_vl_interface': 9.042092705270333e-06, 'learning_rate/action_model': 9.042092705270333e-05, 'epoch': 6.86}) +[2026-06-20 15:54:17,959][starVLA.training.train_starvla][INFO] - Step 3709, Loss: {'train/loss': 0.2837490439414978, 'train/grad_norm_pre_clip': 14.316662788391113, 'timing/data': 0.0005186990601941943, 'timing/model': 1.1908932990627363, 'learning_rate/qwen_vl_interface': 9.040257933005375e-06, 'learning_rate/action_model': 9.040257933005374e-05, 'epoch': 6.86}) +[2026-06-20 15:54:22,804][starVLA.training.train_starvla][INFO] - Step 3710, Loss: {'train/loss': 0.19284968078136444, 'train/grad_norm_pre_clip': 14.08856201171875, 'timing/data': 0.0005258589517325163, 'timing/model': 1.190283578936942, 'learning_rate/qwen_vl_interface': 9.038421602708552e-06, 'learning_rate/action_model': 9.038421602708552e-05, 'epoch': 6.86}) +[2026-06-20 15:54:27,601][starVLA.training.train_starvla][INFO] - Step 3711, Loss: {'train/loss': 0.16279298067092896, 'train/grad_norm_pre_clip': 7.868061065673828, 'timing/data': 0.0005416489439085126, 'timing/model': 1.1832492399262264, 'learning_rate/qwen_vl_interface': 9.036583715134709e-06, 'learning_rate/action_model': 9.036583715134708e-05, 'epoch': 6.86}) +[2026-06-20 15:54:32,408][starVLA.training.train_starvla][INFO] - Step 3712, Loss: {'train/loss': 0.29160720109939575, 'train/grad_norm_pre_clip': 14.892437934875488, 'timing/data': 0.0005277489544823766, 'timing/model': 1.1733346660621464, 'learning_rate/qwen_vl_interface': 9.034744271039332e-06, 'learning_rate/action_model': 9.034744271039332e-05, 'epoch': 6.86}) +[2026-06-20 15:54:37,168][starVLA.training.train_starvla][INFO] - Step 3713, Loss: {'train/loss': 0.34173721075057983, 'train/grad_norm_pre_clip': 14.8019380569458, 'timing/data': 0.0005414499901235104, 'timing/model': 1.1736167339840904, 'learning_rate/qwen_vl_interface': 9.03290327117855e-06, 'learning_rate/action_model': 9.032903271178549e-05, 'epoch': 6.87}) +[2026-06-20 15:54:41,912][starVLA.training.train_starvla][INFO] - Step 3714, Loss: {'train/loss': 0.04500892385840416, 'train/grad_norm_pre_clip': 3.399303674697876, 'timing/data': 0.0005273589631542563, 'timing/model': 1.377082620980218, 'learning_rate/qwen_vl_interface': 9.031060716309122e-06, 'learning_rate/action_model': 9.031060716309121e-05, 'epoch': 6.87}) +[2026-06-20 15:54:46,457][starVLA.training.train_starvla][INFO] - Step 3715, Loss: {'train/loss': 0.199669748544693, 'train/grad_norm_pre_clip': 12.884012222290039, 'timing/data': 0.000531108002178371, 'timing/model': 1.1758298160275444, 'learning_rate/qwen_vl_interface': 9.02921660718846e-06, 'learning_rate/action_model': 9.029216607188459e-05, 'epoch': 6.87}) +[2026-06-20 15:54:51,324][starVLA.training.train_starvla][INFO] - Step 3716, Loss: {'train/loss': 0.20981411635875702, 'train/grad_norm_pre_clip': 13.362964630126953, 'timing/data': 0.0005442089168354869, 'timing/model': 1.1940479909535497, 'learning_rate/qwen_vl_interface': 9.027370944574604e-06, 'learning_rate/action_model': 9.027370944574604e-05, 'epoch': 6.87}) +[2026-06-20 15:54:56,144][starVLA.training.train_starvla][INFO] - Step 3717, Loss: {'train/loss': 0.0972554162144661, 'train/grad_norm_pre_clip': 6.3999552726745605, 'timing/data': 0.000546099036000669, 'timing/model': 1.1935439810622483, 'learning_rate/qwen_vl_interface': 9.025523729226236e-06, 'learning_rate/action_model': 9.025523729226236e-05, 'epoch': 6.87}) +[2026-06-20 15:55:00,965][starVLA.training.train_starvla][INFO] - Step 3718, Loss: {'train/loss': 0.08776181936264038, 'train/grad_norm_pre_clip': 9.414565086364746, 'timing/data': 0.0005279090255498886, 'timing/model': 1.1934032399440184, 'learning_rate/qwen_vl_interface': 9.023674961902677e-06, 'learning_rate/action_model': 9.023674961902676e-05, 'epoch': 6.88}) +[2026-06-20 15:55:05,756][starVLA.training.train_starvla][INFO] - Step 3719, Loss: {'train/loss': 0.09729939699172974, 'train/grad_norm_pre_clip': 9.091464042663574, 'timing/data': 0.0005291590932756662, 'timing/model': 1.1764836789807305, 'learning_rate/qwen_vl_interface': 9.021824643363883e-06, 'learning_rate/action_model': 9.021824643363883e-05, 'epoch': 6.88}) +[2026-06-20 15:55:10,532][starVLA.training.train_starvla][INFO] - Step 3720, Loss: {'train/loss': 0.21908342838287354, 'train/grad_norm_pre_clip': 10.332945823669434, 'timing/data': 0.00052209896966815, 'timing/model': 1.400743555976078, 'learning_rate/qwen_vl_interface': 9.019972774370456e-06, 'learning_rate/action_model': 9.019972774370455e-05, 'epoch': 6.88}) +[2026-06-20 15:55:15,122][starVLA.training.train_starvla][INFO] - Step 3721, Loss: {'train/loss': 0.06605537235736847, 'train/grad_norm_pre_clip': 6.031994342803955, 'timing/data': 0.0005270990077406168, 'timing/model': 1.1754398570628837, 'learning_rate/qwen_vl_interface': 9.018119355683623e-06, 'learning_rate/action_model': 9.018119355683623e-05, 'epoch': 6.88}) +[2026-06-20 15:55:19,976][starVLA.training.train_starvla][INFO] - Step 3722, Loss: {'train/loss': 0.16700470447540283, 'train/grad_norm_pre_clip': 10.035266876220703, 'timing/data': 0.000523668946698308, 'timing/model': 1.1935356030007824, 'learning_rate/qwen_vl_interface': 9.016264388065258e-06, 'learning_rate/action_model': 9.016264388065258e-05, 'epoch': 6.88}) +[2026-06-20 15:55:24,805][starVLA.training.train_starvla][INFO] - Step 3723, Loss: {'train/loss': 0.2444387525320053, 'train/grad_norm_pre_clip': 9.246280670166016, 'timing/data': 0.0005297290626913309, 'timing/model': 1.1938880210509524, 'learning_rate/qwen_vl_interface': 9.014407872277869e-06, 'learning_rate/action_model': 9.014407872277869e-05, 'epoch': 6.89}) +[2026-06-20 15:55:29,613][starVLA.training.train_starvla][INFO] - Step 3724, Loss: {'train/loss': 0.11259331554174423, 'train/grad_norm_pre_clip': 6.526053428649902, 'timing/data': 0.0005308700492605567, 'timing/model': 1.179163426044397, 'learning_rate/qwen_vl_interface': 9.012549809084595e-06, 'learning_rate/action_model': 9.012549809084595e-05, 'epoch': 6.89}) +[2026-06-20 15:55:34,548][starVLA.training.train_starvla][INFO] - Step 3725, Loss: {'train/loss': 0.07917369157075882, 'train/grad_norm_pre_clip': 9.862738609313965, 'timing/data': 0.0005344190867617726, 'timing/model': 1.4848940409719944, 'learning_rate/qwen_vl_interface': 9.01069019924922e-06, 'learning_rate/action_model': 9.010690199249219e-05, 'epoch': 6.89}) +[2026-06-20 15:55:39,220][starVLA.training.train_starvla][INFO] - Step 3726, Loss: {'train/loss': 0.28116127848625183, 'train/grad_norm_pre_clip': 10.484341621398926, 'timing/data': 0.0005238789599388838, 'timing/model': 1.1796310720965266, 'learning_rate/qwen_vl_interface': 9.008829043536157e-06, 'learning_rate/action_model': 9.008829043536156e-05, 'epoch': 6.89}) +[2026-06-20 15:55:44,125][starVLA.training.train_starvla][INFO] - Step 3727, Loss: {'train/loss': 0.07012999802827835, 'train/grad_norm_pre_clip': 5.245060920715332, 'timing/data': 0.0005414489423856139, 'timing/model': 1.1783659430220723, 'learning_rate/qwen_vl_interface': 9.006966342710458e-06, 'learning_rate/action_model': 9.006966342710457e-05, 'epoch': 6.89}) +[2026-06-20 15:55:48,916][starVLA.training.train_starvla][INFO] - Step 3728, Loss: {'train/loss': 0.0359792560338974, 'train/grad_norm_pre_clip': 3.664246082305908, 'timing/data': 0.0005382290109992027, 'timing/model': 1.1812800859333947, 'learning_rate/qwen_vl_interface': 9.005102097537809e-06, 'learning_rate/action_model': 9.005102097537809e-05, 'epoch': 6.89}) +[2026-06-20 15:55:53,717][starVLA.training.train_starvla][INFO] - Step 3729, Loss: {'train/loss': 0.1674998253583908, 'train/grad_norm_pre_clip': 12.143112182617188, 'timing/data': 0.0005385789554566145, 'timing/model': 1.1809074450284243, 'learning_rate/qwen_vl_interface': 9.00323630878453e-06, 'learning_rate/action_model': 9.00323630878453e-05, 'epoch': 6.9}) +[2026-06-20 15:55:58,617][starVLA.training.train_starvla][INFO] - Step 3730, Loss: {'train/loss': 0.0487620085477829, 'train/grad_norm_pre_clip': 4.066675662994385, 'timing/data': 0.0005394690670073032, 'timing/model': 1.1811794139211997, 'learning_rate/qwen_vl_interface': 9.001368977217577e-06, 'learning_rate/action_model': 9.001368977217577e-05, 'epoch': 6.9}) +[2026-06-20 15:56:03,540][starVLA.training.train_starvla][INFO] - Step 3731, Loss: {'train/loss': 0.19988465309143066, 'train/grad_norm_pre_clip': 11.084634780883789, 'timing/data': 0.0005333690205588937, 'timing/model': 1.4864139159908518, 'learning_rate/qwen_vl_interface': 8.999500103604538e-06, 'learning_rate/action_model': 8.999500103604536e-05, 'epoch': 6.9}) +[2026-06-20 15:56:08,227][starVLA.training.train_starvla][INFO] - Step 3732, Loss: {'train/loss': 0.07737909257411957, 'train/grad_norm_pre_clip': 7.034862041473389, 'timing/data': 0.0005310389678925276, 'timing/model': 1.1794260239694268, 'learning_rate/qwen_vl_interface': 8.997629688713639e-06, 'learning_rate/action_model': 8.997629688713638e-05, 'epoch': 6.9}) +[2026-06-20 15:56:13,180][starVLA.training.train_starvla][INFO] - Step 3733, Loss: {'train/loss': 0.15226629376411438, 'train/grad_norm_pre_clip': 10.511744499206543, 'timing/data': 0.0005444390699267387, 'timing/model': 1.1814452180406079, 'learning_rate/qwen_vl_interface': 8.995757733313733e-06, 'learning_rate/action_model': 8.995757733313732e-05, 'epoch': 6.9}) +[2026-06-20 15:56:18,096][starVLA.training.train_starvla][INFO] - Step 3734, Loss: {'train/loss': 0.23472963273525238, 'train/grad_norm_pre_clip': 13.64599323272705, 'timing/data': 0.0005187789211049676, 'timing/model': 1.1776934339432046, 'learning_rate/qwen_vl_interface': 8.993884238174314e-06, 'learning_rate/action_model': 8.993884238174314e-05, 'epoch': 6.91}) +[2026-06-20 15:56:22,964][starVLA.training.train_starvla][INFO] - Step 3735, Loss: {'train/loss': 0.195625901222229, 'train/grad_norm_pre_clip': 10.742276191711426, 'timing/data': 0.0005394190084189177, 'timing/model': 1.1803312179399654, 'learning_rate/qwen_vl_interface': 8.992009204065505e-06, 'learning_rate/action_model': 8.992009204065505e-05, 'epoch': 6.91}) +[2026-06-20 15:56:27,869][starVLA.training.train_starvla][INFO] - Step 3736, Loss: {'train/loss': 0.040301453322172165, 'train/grad_norm_pre_clip': 4.045953273773193, 'timing/data': 0.000537988031283021, 'timing/model': 1.177964802016504, 'learning_rate/qwen_vl_interface': 8.990132631758058e-06, 'learning_rate/action_model': 8.990132631758059e-05, 'epoch': 6.91}) +[2026-06-20 15:56:32,769][starVLA.training.train_starvla][INFO] - Step 3737, Loss: {'train/loss': 0.06395886093378067, 'train/grad_norm_pre_clip': 18.456022262573242, 'timing/data': 0.0005453289486467838, 'timing/model': 1.4170834849355742, 'learning_rate/qwen_vl_interface': 8.988254522023364e-06, 'learning_rate/action_model': 8.988254522023363e-05, 'epoch': 6.91}) +[2026-06-20 15:56:37,437][starVLA.training.train_starvla][INFO] - Step 3738, Loss: {'train/loss': 0.39266955852508545, 'train/grad_norm_pre_clip': 13.57422924041748, 'timing/data': 0.0005484588909894228, 'timing/model': 1.1780793480575085, 'learning_rate/qwen_vl_interface': 8.986374875633441e-06, 'learning_rate/action_model': 8.98637487563344e-05, 'epoch': 6.91}) +[2026-06-20 15:56:42,349][starVLA.training.train_starvla][INFO] - Step 3739, Loss: {'train/loss': 0.026322532445192337, 'train/grad_norm_pre_clip': 2.041954517364502, 'timing/data': 0.000544599024578929, 'timing/model': 1.1782183859031647, 'learning_rate/qwen_vl_interface': 8.984493693360944e-06, 'learning_rate/action_model': 8.984493693360943e-05, 'epoch': 6.91}) +[2026-06-20 15:56:47,409][starVLA.training.train_starvla][INFO] - Step 3740, Loss: {'train/loss': 0.302552193403244, 'train/grad_norm_pre_clip': 16.62765884399414, 'timing/data': 0.0005421090172603726, 'timing/model': 1.2386724320240319, 'learning_rate/qwen_vl_interface': 8.982610975979153e-06, 'learning_rate/action_model': 8.982610975979152e-05, 'epoch': 6.92}) +[2026-06-20 15:56:52,367][starVLA.training.train_starvla][INFO] - Step 3741, Loss: {'train/loss': 0.0835593119263649, 'train/grad_norm_pre_clip': 8.78563117980957, 'timing/data': 0.000530109042301774, 'timing/model': 1.2272194719407707, 'learning_rate/qwen_vl_interface': 8.98072672426198e-06, 'learning_rate/action_model': 8.980726724261981e-05, 'epoch': 6.92}) +[2026-06-20 15:56:57,316][starVLA.training.train_starvla][INFO] - Step 3742, Loss: {'train/loss': 0.10867497324943542, 'train/grad_norm_pre_clip': 9.155552864074707, 'timing/data': 0.0005231690593063831, 'timing/model': 1.465822676080279, 'learning_rate/qwen_vl_interface': 8.978840938983977e-06, 'learning_rate/action_model': 8.978840938983977e-05, 'epoch': 6.92}) +[2026-06-20 15:57:02,226][starVLA.training.train_starvla][INFO] - Step 3743, Loss: {'train/loss': 0.12066663056612015, 'train/grad_norm_pre_clip': 11.085162162780762, 'timing/data': 0.0005416589556261897, 'timing/model': 1.3820852410281077, 'learning_rate/qwen_vl_interface': 8.976953620920313e-06, 'learning_rate/action_model': 8.976953620920312e-05, 'epoch': 6.92}) +[2026-06-20 15:57:06,892][starVLA.training.train_starvla][INFO] - Step 3744, Loss: {'train/loss': 0.31216177344322205, 'train/grad_norm_pre_clip': 12.397551536560059, 'timing/data': 0.0005538789555430412, 'timing/model': 1.1810569450026378, 'learning_rate/qwen_vl_interface': 8.975064770846794e-06, 'learning_rate/action_model': 8.975064770846794e-05, 'epoch': 6.92}) +[2026-06-20 15:57:11,784][starVLA.training.train_starvla][INFO] - Step 3745, Loss: {'train/loss': 0.12365231662988663, 'train/grad_norm_pre_clip': 6.807906150817871, 'timing/data': 0.0005475090583786368, 'timing/model': 1.2287635339889675, 'learning_rate/qwen_vl_interface': 8.973174389539856e-06, 'learning_rate/action_model': 8.973174389539855e-05, 'epoch': 6.93}) +[2026-06-20 15:57:16,634][starVLA.training.train_starvla][INFO] - Step 3746, Loss: {'train/loss': 0.015174439176917076, 'train/grad_norm_pre_clip': 1.8546115159988403, 'timing/data': 0.0005319589981809258, 'timing/model': 1.175830160966143, 'learning_rate/qwen_vl_interface': 8.971282477776563e-06, 'learning_rate/action_model': 8.971282477776563e-05, 'epoch': 6.93}) +[2026-06-20 15:57:21,487][starVLA.training.train_starvla][INFO] - Step 3747, Loss: {'train/loss': 0.12143751978874207, 'train/grad_norm_pre_clip': 3.645167827606201, 'timing/data': 0.0005472189513966441, 'timing/model': 1.180720061995089, 'learning_rate/qwen_vl_interface': 8.969389036334609e-06, 'learning_rate/action_model': 8.969389036334609e-05, 'epoch': 6.93}) +[2026-06-20 15:57:26,401][starVLA.training.train_starvla][INFO] - Step 3748, Loss: {'train/loss': 0.2870912253856659, 'train/grad_norm_pre_clip': 14.254606246948242, 'timing/data': 0.0005205890629440546, 'timing/model': 1.4789796809200197, 'learning_rate/qwen_vl_interface': 8.967494065992315e-06, 'learning_rate/action_model': 8.967494065992315e-05, 'epoch': 6.93}) +[2026-06-20 15:57:31,228][starVLA.training.train_starvla][INFO] - Step 3749, Loss: {'train/loss': 0.15919651091098785, 'train/grad_norm_pre_clip': 15.456480979919434, 'timing/data': 0.0005410090088844299, 'timing/model': 1.3895395669387653, 'learning_rate/qwen_vl_interface': 8.965597567528634e-06, 'learning_rate/action_model': 8.965597567528633e-05, 'epoch': 6.93}) +[2026-06-20 15:57:35,940][starVLA.training.train_starvla][INFO] - Step 3750, Loss: {'train/loss': 0.04927779734134674, 'train/grad_norm_pre_clip': 4.611067771911621, 'timing/data': 0.0005429789889603853, 'timing/model': 1.228615327971056, 'learning_rate/qwen_vl_interface': 8.963699541723143e-06, 'learning_rate/action_model': 8.963699541723142e-05, 'epoch': 6.94}) +[2026-06-20 15:57:40,818][starVLA.training.train_starvla][INFO] - Step 3751, Loss: {'train/loss': 0.20474353432655334, 'train/grad_norm_pre_clip': 16.204601287841797, 'timing/data': 0.0005287489620968699, 'timing/model': 1.178653920069337, 'learning_rate/qwen_vl_interface': 8.961799989356048e-06, 'learning_rate/action_model': 8.961799989356048e-05, 'epoch': 6.94}) +[2026-06-20 15:57:45,680][starVLA.training.train_starvla][INFO] - Step 3752, Loss: {'train/loss': 0.07155774533748627, 'train/grad_norm_pre_clip': 9.375732421875, 'timing/data': 0.0005443389527499676, 'timing/model': 1.2303952419897541, 'learning_rate/qwen_vl_interface': 8.959898911208185e-06, 'learning_rate/action_model': 8.959898911208185e-05, 'epoch': 6.94}) +[2026-06-20 15:57:50,590][starVLA.training.train_starvla][INFO] - Step 3753, Loss: {'train/loss': 0.17509205639362335, 'train/grad_norm_pre_clip': 12.218804359436035, 'timing/data': 0.0005400889785960317, 'timing/model': 1.2287814130540937, 'learning_rate/qwen_vl_interface': 8.957996308061015e-06, 'learning_rate/action_model': 8.957996308061014e-05, 'epoch': 6.94}) +[2026-06-20 15:57:55,511][starVLA.training.train_starvla][INFO] - Step 3754, Loss: {'train/loss': 0.05481497198343277, 'train/grad_norm_pre_clip': 6.180869102478027, 'timing/data': 0.0005520989652723074, 'timing/model': 1.4207947419490665, 'learning_rate/qwen_vl_interface': 8.956092180696626e-06, 'learning_rate/action_model': 8.956092180696625e-05, 'epoch': 6.94}) +[2026-06-20 15:58:00,124][starVLA.training.train_starvla][INFO] - Step 3755, Loss: {'train/loss': 0.12980878353118896, 'train/grad_norm_pre_clip': 14.985153198242188, 'timing/data': 0.0005403879331424832, 'timing/model': 1.1791798940394074, 'learning_rate/qwen_vl_interface': 8.95418652989773e-06, 'learning_rate/action_model': 8.95418652989773e-05, 'epoch': 6.94}) +[2026-06-20 15:58:04,977][starVLA.training.train_starvla][INFO] - Step 3756, Loss: {'train/loss': 0.25718313455581665, 'train/grad_norm_pre_clip': 10.023138999938965, 'timing/data': 0.0005343889351934195, 'timing/model': 1.244975250097923, 'learning_rate/qwen_vl_interface': 8.952279356447675e-06, 'learning_rate/action_model': 8.952279356447675e-05, 'epoch': 6.95}) +[2026-06-20 15:58:09,881][starVLA.training.train_starvla][INFO] - Step 3757, Loss: {'train/loss': 0.16678282618522644, 'train/grad_norm_pre_clip': 10.818889617919922, 'timing/data': 0.0005299689946696162, 'timing/model': 1.2262995839118958, 'learning_rate/qwen_vl_interface': 8.950370661130421e-06, 'learning_rate/action_model': 8.950370661130422e-05, 'epoch': 6.95}) +[2026-06-20 15:58:14,715][starVLA.training.train_starvla][INFO] - Step 3758, Loss: {'train/loss': 0.2174832969903946, 'train/grad_norm_pre_clip': 19.398880004882812, 'timing/data': 0.0005311089335009456, 'timing/model': 1.1795273500028998, 'learning_rate/qwen_vl_interface': 8.948460444730566e-06, 'learning_rate/action_model': 8.948460444730566e-05, 'epoch': 6.95}) +[2026-06-20 15:58:19,657][starVLA.training.train_starvla][INFO] - Step 3759, Loss: {'train/loss': 0.23905158042907715, 'train/grad_norm_pre_clip': 13.56818675994873, 'timing/data': 0.0005389680154621601, 'timing/model': 1.22621354192961, 'learning_rate/qwen_vl_interface': 8.946548708033324e-06, 'learning_rate/action_model': 8.946548708033323e-05, 'epoch': 6.95}) +[2026-06-20 15:58:24,565][starVLA.training.train_starvla][INFO] - Step 3760, Loss: {'train/loss': 0.11747682839632034, 'train/grad_norm_pre_clip': 5.653897762298584, 'timing/data': 0.0005319589981809258, 'timing/model': 1.46579955809284, 'learning_rate/qwen_vl_interface': 8.944635451824538e-06, 'learning_rate/action_model': 8.944635451824538e-05, 'epoch': 6.95}) +[2026-06-20 15:58:29,292][starVLA.training.train_starvla][INFO] - Step 3761, Loss: {'train/loss': 0.2945578992366791, 'train/grad_norm_pre_clip': 9.41720199584961, 'timing/data': 0.0005299089243635535, 'timing/model': 1.2452807830413803, 'learning_rate/qwen_vl_interface': 8.942720676890677e-06, 'learning_rate/action_model': 8.942720676890677e-05, 'epoch': 6.96}) +[2026-06-20 15:58:34,257][starVLA.training.train_starvla][INFO] - Step 3762, Loss: {'train/loss': 0.1036035418510437, 'train/grad_norm_pre_clip': 7.727360248565674, 'timing/data': 0.0005331899737939239, 'timing/model': 1.2435813850024715, 'learning_rate/qwen_vl_interface': 8.940804384018833e-06, 'learning_rate/action_model': 8.940804384018832e-05, 'epoch': 6.96}) +[2026-06-20 15:58:39,011][starVLA.training.train_starvla][INFO] - Step 3763, Loss: {'train/loss': 0.0643535777926445, 'train/grad_norm_pre_clip': 5.986156463623047, 'timing/data': 0.0005417790962383151, 'timing/model': 1.1796233049826697, 'learning_rate/qwen_vl_interface': 8.93888657399672e-06, 'learning_rate/action_model': 8.93888657399672e-05, 'epoch': 6.96}) +[2026-06-20 15:58:43,901][starVLA.training.train_starvla][INFO] - Step 3764, Loss: {'train/loss': 0.19373789429664612, 'train/grad_norm_pre_clip': 10.961156845092773, 'timing/data': 0.0005559389246627688, 'timing/model': 1.1795770649332553, 'learning_rate/qwen_vl_interface': 8.936967247612678e-06, 'learning_rate/action_model': 8.936967247612678e-05, 'epoch': 6.96}) +[2026-06-20 15:58:48,852][starVLA.training.train_starvla][INFO] - Step 3765, Loss: {'train/loss': 0.11283376812934875, 'train/grad_norm_pre_clip': 13.416556358337402, 'timing/data': 0.0005278190365061164, 'timing/model': 1.228710770024918, 'learning_rate/qwen_vl_interface': 8.93504640565567e-06, 'learning_rate/action_model': 8.93504640565567e-05, 'epoch': 6.96}) +[2026-06-20 15:58:53,881][starVLA.training.train_starvla][INFO] - Step 3766, Loss: {'train/loss': 0.03615832328796387, 'train/grad_norm_pre_clip': 4.587575912475586, 'timing/data': 0.0005317990435287356, 'timing/model': 1.4911780529655516, 'learning_rate/qwen_vl_interface': 8.933124048915278e-06, 'learning_rate/action_model': 8.933124048915278e-05, 'epoch': 6.96}) +[2026-06-20 15:58:58,764][starVLA.training.train_starvla][INFO] - Step 3767, Loss: {'train/loss': 0.07105039060115814, 'train/grad_norm_pre_clip': 4.383957862854004, 'timing/data': 0.000542348949238658, 'timing/model': 1.3668216769583523, 'learning_rate/qwen_vl_interface': 8.931200178181715e-06, 'learning_rate/action_model': 8.931200178181715e-05, 'epoch': 6.97}) +[2026-06-20 15:59:03,431][starVLA.training.train_starvla][INFO] - Step 3768, Loss: {'train/loss': 0.06389785557985306, 'train/grad_norm_pre_clip': 5.413763046264648, 'timing/data': 0.0005360590294003487, 'timing/model': 1.2456738640321419, 'learning_rate/qwen_vl_interface': 8.929274794245812e-06, 'learning_rate/action_model': 8.929274794245812e-05, 'epoch': 6.97}) +[2026-06-20 15:59:08,391][starVLA.training.train_starvla][INFO] - Step 3769, Loss: {'train/loss': 0.2501820921897888, 'train/grad_norm_pre_clip': 14.7168550491333, 'timing/data': 0.0005432290490716696, 'timing/model': 1.2452690640930086, 'learning_rate/qwen_vl_interface': 8.927347897899018e-06, 'learning_rate/action_model': 8.927347897899016e-05, 'epoch': 6.97}) +[2026-06-20 15:59:13,351][starVLA.training.train_starvla][INFO] - Step 3770, Loss: {'train/loss': 0.34203463792800903, 'train/grad_norm_pre_clip': 14.595051765441895, 'timing/data': 0.0005302089266479015, 'timing/model': 1.2454691330203786, 'learning_rate/qwen_vl_interface': 8.925419489933407e-06, 'learning_rate/action_model': 8.925419489933408e-05, 'epoch': 6.97}) +[2026-06-20 15:59:18,271][starVLA.training.train_starvla][INFO] - Step 3771, Loss: {'train/loss': 0.22271409630775452, 'train/grad_norm_pre_clip': 12.18865966796875, 'timing/data': 0.0005448090378195047, 'timing/model': 1.1809254640247673, 'learning_rate/qwen_vl_interface': 8.923489571141679e-06, 'learning_rate/action_model': 8.923489571141679e-05, 'epoch': 6.97}) +[2026-06-20 15:59:23,122][starVLA.training.train_starvla][INFO] - Step 3772, Loss: {'train/loss': 0.12848323583602905, 'train/grad_norm_pre_clip': 9.929839134216309, 'timing/data': 0.0005331289721652865, 'timing/model': 1.4166615159483626, 'learning_rate/qwen_vl_interface': 8.921558142317147e-06, 'learning_rate/action_model': 8.921558142317147e-05, 'epoch': 6.98}) +[2026-06-20 15:59:28,079][starVLA.training.train_starvla][INFO] - Step 3773, Loss: {'train/loss': 0.10695898532867432, 'train/grad_norm_pre_clip': 7.371830463409424, 'timing/data': 0.0005271189147606492, 'timing/model': 1.4673598799854517, 'learning_rate/qwen_vl_interface': 8.91962520425375e-06, 'learning_rate/action_model': 8.91962520425375e-05, 'epoch': 6.98}) +[2026-06-20 15:59:32,757][starVLA.training.train_starvla][INFO] - Step 3774, Loss: {'train/loss': 0.08138471841812134, 'train/grad_norm_pre_clip': 12.727742195129395, 'timing/data': 0.0005175989354029298, 'timing/model': 1.2298938080202788, 'learning_rate/qwen_vl_interface': 8.917690757746044e-06, 'learning_rate/action_model': 8.917690757746044e-05, 'epoch': 6.98}) +[2026-06-20 15:59:37,756][starVLA.training.train_starvla][INFO] - Step 3775, Loss: {'train/loss': 0.05009763315320015, 'train/grad_norm_pre_clip': 5.358070373535156, 'timing/data': 0.000531268073245883, 'timing/model': 1.229271047981456, 'learning_rate/qwen_vl_interface': 8.915754803589208e-06, 'learning_rate/action_model': 8.915754803589209e-05, 'epoch': 6.98}) +[2026-06-20 15:59:42,549][starVLA.training.train_starvla][INFO] - Step 3776, Loss: {'train/loss': 0.05860466882586479, 'train/grad_norm_pre_clip': 16.590726852416992, 'timing/data': 0.0005428679287433624, 'timing/model': 1.180255500948988, 'learning_rate/qwen_vl_interface': 8.913817342579042e-06, 'learning_rate/action_model': 8.913817342579042e-05, 'epoch': 6.98}) +[2026-06-20 15:59:47,520][starVLA.training.train_starvla][INFO] - Step 3777, Loss: {'train/loss': 0.11013989895582199, 'train/grad_norm_pre_clip': 7.7192606925964355, 'timing/data': 0.0005332289729267359, 'timing/model': 1.2306526240427047, 'learning_rate/qwen_vl_interface': 8.91187837551196e-06, 'learning_rate/action_model': 8.911878375511961e-05, 'epoch': 6.99}) +[2026-06-20 15:59:52,460][starVLA.training.train_starvla][INFO] - Step 3778, Loss: {'train/loss': 0.48695647716522217, 'train/grad_norm_pre_clip': 18.39293098449707, 'timing/data': 0.0005344289820641279, 'timing/model': 1.4244533539749682, 'learning_rate/qwen_vl_interface': 8.909937903185e-06, 'learning_rate/action_model': 8.909937903185e-05, 'epoch': 6.99}) +[2026-06-20 15:59:57,173][starVLA.training.train_starvla][INFO] - Step 3779, Loss: {'train/loss': 0.03911077603697777, 'train/grad_norm_pre_clip': 4.231695175170898, 'timing/data': 0.0005383699899539351, 'timing/model': 1.2288876449456438, 'learning_rate/qwen_vl_interface': 8.907995926395814e-06, 'learning_rate/action_model': 8.907995926395815e-05, 'epoch': 6.99}) +[2026-06-20 16:00:02,033][starVLA.training.train_starvla][INFO] - Step 3780, Loss: {'train/loss': 0.13916167616844177, 'train/grad_norm_pre_clip': 14.544340133666992, 'timing/data': 0.0005302090430632234, 'timing/model': 1.2268561579985544, 'learning_rate/qwen_vl_interface': 8.90605244594268e-06, 'learning_rate/action_model': 8.906052445942679e-05, 'epoch': 6.99}) +[2026-06-20 16:00:06,908][starVLA.training.train_starvla][INFO] - Step 3781, Loss: {'train/loss': 0.1978933960199356, 'train/grad_norm_pre_clip': 10.746026992797852, 'timing/data': 0.0005328590050339699, 'timing/model': 1.182055352954194, 'learning_rate/qwen_vl_interface': 8.904107462624484e-06, 'learning_rate/action_model': 8.904107462624483e-05, 'epoch': 6.99}) +[2026-06-20 16:00:11,766][starVLA.training.train_starvla][INFO] - Step 3782, Loss: {'train/loss': 0.3408978581428528, 'train/grad_norm_pre_clip': 15.738669395446777, 'timing/data': 0.000530179007910192, 'timing/model': 1.2425077760126442, 'learning_rate/qwen_vl_interface': 8.902160977240739e-06, 'learning_rate/action_model': 8.902160977240739e-05, 'epoch': 6.99}) +[2026-06-20 16:00:16,572][starVLA.training.train_starvla][INFO] - Step 3783, Loss: {'train/loss': 0.043723300099372864, 'train/grad_norm_pre_clip': 10.745508193969727, 'timing/data': 0.0005334389861673117, 'timing/model': 1.1812970219179988, 'learning_rate/qwen_vl_interface': 8.90021299059157e-06, 'learning_rate/action_model': 8.90021299059157e-05, 'epoch': 7.0}) +[2026-06-20 16:00:21,400][starVLA.training.train_starvla][INFO] - Step 3784, Loss: {'train/loss': 0.2627495527267456, 'train/grad_norm_pre_clip': 10.985688209533691, 'timing/data': 0.0005191189702600241, 'timing/model': 1.383703044964932, 'learning_rate/qwen_vl_interface': 8.898263503477722e-06, 'learning_rate/action_model': 8.898263503477721e-05, 'epoch': 7.0}) +[2026-06-20 16:00:25,959][starVLA.training.train_starvla][INFO] - Step 3785, Loss: {'train/loss': 0.3675940930843353, 'train/grad_norm_pre_clip': 18.219032287597656, 'timing/data': 0.000533908954821527, 'timing/model': 1.1779486669693142, 'learning_rate/qwen_vl_interface': 8.896312516700553e-06, 'learning_rate/action_model': 8.896312516700554e-05, 'epoch': 7.0}) +[2026-06-20 16:00:30,783][starVLA.training.train_starvla][INFO] - Step 3786, Loss: {'train/loss': 0.05814962834119797, 'train/grad_norm_pre_clip': 6.497032165527344, 'timing/data': 0.0005272090202197433, 'timing/model': 1.1793621230171993, 'learning_rate/qwen_vl_interface': 8.894360031062043e-06, 'learning_rate/action_model': 8.894360031062043e-05, 'epoch': 7.0}) +[2026-06-20 16:00:35,592][starVLA.training.train_starvla][INFO] - Step 3787, Loss: {'train/loss': 0.0383661612868309, 'train/grad_norm_pre_clip': 11.762007713317871, 'timing/data': 0.0005299389595165849, 'timing/model': 1.2010723300045356, 'learning_rate/qwen_vl_interface': 8.892406047364783e-06, 'learning_rate/action_model': 8.892406047364784e-05, 'epoch': 7.0}) +[2026-06-20 16:00:40,404][starVLA.training.train_starvla][INFO] - Step 3788, Loss: {'train/loss': 0.2787225544452667, 'train/grad_norm_pre_clip': 7.481578350067139, 'timing/data': 0.0005181389860808849, 'timing/model': 1.1817279370734468, 'learning_rate/qwen_vl_interface': 8.890450566411984e-06, 'learning_rate/action_model': 8.890450566411984e-05, 'epoch': 7.01}) +[2026-06-20 16:00:45,236][starVLA.training.train_starvla][INFO] - Step 3789, Loss: {'train/loss': 0.2279702126979828, 'train/grad_norm_pre_clip': 7.33696174621582, 'timing/data': 0.000529948971234262, 'timing/model': 1.1796542899683118, 'learning_rate/qwen_vl_interface': 8.888493589007467e-06, 'learning_rate/action_model': 8.888493589007468e-05, 'epoch': 7.01}) +[2026-06-20 16:00:50,029][starVLA.training.train_starvla][INFO] - Step 3790, Loss: {'train/loss': 0.06797685474157333, 'train/grad_norm_pre_clip': 4.6771979331970215, 'timing/data': 0.0005400989903137088, 'timing/model': 1.3895070690196007, 'learning_rate/qwen_vl_interface': 8.886535115955676e-06, 'learning_rate/action_model': 8.886535115955675e-05, 'epoch': 7.01}) +[2026-06-20 16:00:54,607][starVLA.training.train_starvla][INFO] - Step 3791, Loss: {'train/loss': 0.43008339405059814, 'train/grad_norm_pre_clip': 16.300670623779297, 'timing/data': 0.0005406889831647277, 'timing/model': 1.195023439009674, 'learning_rate/qwen_vl_interface': 8.884575148061662e-06, 'learning_rate/action_model': 8.88457514806166e-05, 'epoch': 7.01}) +[2026-06-20 16:00:59,401][starVLA.training.train_starvla][INFO] - Step 3792, Loss: {'train/loss': 0.12064675241708755, 'train/grad_norm_pre_clip': 9.55388355255127, 'timing/data': 0.0005297890165820718, 'timing/model': 1.1775085020344704, 'learning_rate/qwen_vl_interface': 8.882613686131095e-06, 'learning_rate/action_model': 8.882613686131093e-05, 'epoch': 7.01}) +[2026-06-20 16:01:04,202][starVLA.training.train_starvla][INFO] - Step 3793, Loss: {'train/loss': 0.27331268787384033, 'train/grad_norm_pre_clip': 12.137188911437988, 'timing/data': 0.0005263390485197306, 'timing/model': 1.1788630479713902, 'learning_rate/qwen_vl_interface': 8.880650730970254e-06, 'learning_rate/action_model': 8.880650730970254e-05, 'epoch': 7.01}) +[2026-06-20 16:01:09,020][starVLA.training.train_starvla][INFO] - Step 3794, Loss: {'train/loss': 0.24157162010669708, 'train/grad_norm_pre_clip': 13.08489990234375, 'timing/data': 0.0005240789614617825, 'timing/model': 1.1960103450110182, 'learning_rate/qwen_vl_interface': 8.878686283386042e-06, 'learning_rate/action_model': 8.878686283386042e-05, 'epoch': 7.02}) +[2026-06-20 16:01:13,816][starVLA.training.train_starvla][INFO] - Step 3795, Loss: {'train/loss': 0.19895488023757935, 'train/grad_norm_pre_clip': 9.906373023986816, 'timing/data': 0.0005366189870983362, 'timing/model': 1.4172916919924319, 'learning_rate/qwen_vl_interface': 8.876720344185967e-06, 'learning_rate/action_model': 8.876720344185967e-05, 'epoch': 7.02}) +[2026-06-20 16:01:18,404][starVLA.training.train_starvla][INFO] - Step 3796, Loss: {'train/loss': 0.15912586450576782, 'train/grad_norm_pre_clip': 10.73906135559082, 'timing/data': 0.0005308089312165976, 'timing/model': 1.1825588690117002, 'learning_rate/qwen_vl_interface': 8.874752914178149e-06, 'learning_rate/action_model': 8.87475291417815e-05, 'epoch': 7.02}) +[2026-06-20 16:01:23,189][starVLA.training.train_starvla][INFO] - Step 3797, Loss: {'train/loss': 0.17315632104873657, 'train/grad_norm_pre_clip': 7.9073872566223145, 'timing/data': 0.0005257989978417754, 'timing/model': 1.178444825927727, 'learning_rate/qwen_vl_interface': 8.87278399417133e-06, 'learning_rate/action_model': 8.872783994171329e-05, 'epoch': 7.02}) +[2026-06-20 16:01:27,993][starVLA.training.train_starvla][INFO] - Step 3798, Loss: {'train/loss': 0.2300395369529724, 'train/grad_norm_pre_clip': 10.634993553161621, 'timing/data': 0.0005296589806675911, 'timing/model': 1.1970957990270108, 'learning_rate/qwen_vl_interface': 8.870813584974852e-06, 'learning_rate/action_model': 8.870813584974852e-05, 'epoch': 7.02}) +[2026-06-20 16:01:32,748][starVLA.training.train_starvla][INFO] - Step 3799, Loss: {'train/loss': 0.16093844175338745, 'train/grad_norm_pre_clip': 8.287372589111328, 'timing/data': 0.000533769023604691, 'timing/model': 1.1968440290074795, 'learning_rate/qwen_vl_interface': 8.868841687398683e-06, 'learning_rate/action_model': 8.868841687398683e-05, 'epoch': 7.03}) +[2026-06-20 16:01:37,546][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 3800 +[2026-06-20 16:08:06,963][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 3800 +[2026-06-20 16:08:06,964][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 3800 +[2026-06-20 16:15:06,524][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 3800 +[2026-06-20 16:15:06,526][starVLA.training.train_starvla][INFO] - Step 3800, Loss: {'train/loss': 0.18718892335891724, 'train/grad_norm_pre_clip': 8.62295150756836, 'eval/loss': 0.49417972564697266, 'eval/latency_0/loss': 0.5929424166679382, 'eval/latency_1/loss': 0.7113783359527588, 'eval/latency_2/loss': 0.7219423055648804, 'eval/latency_3/loss': 0.31958112120628357, 'eval/latency_4/loss': 0.10083045065402985, 'eval/flappy/loss': 0.49417972564697266, 'eval/flappy/latency_0/loss': 0.5929424166679382, 'eval/flappy/latency_1/loss': 0.7113783359527588, 'eval/flappy/latency_2/loss': 0.7219423055648804, 'eval/flappy/latency_3/loss': 0.31958112120628357, 'eval/flappy/latency_4/loss': 0.10083045065402985, 'eval/flap_precision': 0.1497326203208556, 'eval/flap_recall': 0.10181818181818182, 'eval/flap_f1': 0.12121212121212122, 'eval/flap_rate_model': 0.051944444444444446, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.12121212121212122, 'eval/latency_0/flap_precision': 0.1497326203208556, 'eval/latency_0/flap_recall': 0.10181818181818182, 'eval/latency_0/flap_f1': 0.12121212121212122, 'eval/latency_0/flap_rate_model': 0.051944444444444446, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.12121212121212122, 'timing/data': 0.0005365129327401519, 'timing/model': 1.1969886790029705, 'learning_rate/qwen_vl_interface': 8.86686830225339e-06, 'learning_rate/action_model': 8.866868302253389e-05, 'epoch': 7.03}) +[2026-06-20 16:15:06,526][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_3800_state +[2026-06-20 16:15:06,530][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-20 16:15:15,094][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_3800_state/model.safetensors +[2026-06-20 16:15:29,013][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_3800_state/optimizer.bin +[2026-06-20 16:15:29,014][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_3800_state/sampler.bin +[2026-06-20 16:15:29,014][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_3800_state/sampler_1.bin +[2026-06-20 16:15:29,015][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_3800_state/random_states_0.pkl +[2026-06-20 16:15:31,580][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3800 +[2026-06-20 16:15:31,580][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-20 16:15:31,639][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-20 16:15:36,464][starVLA.training.train_starvla][INFO] - Step 3801, Loss: {'train/loss': 0.11723300069570541, 'train/grad_norm_pre_clip': 7.231532096862793, 'timing/data': 0.0005341079086065292, 'timing/model': 1.2264092520345002, 'learning_rate/qwen_vl_interface': 8.864893430350159e-06, 'learning_rate/action_model': 8.864893430350159e-05, 'epoch': 7.03}) +[2026-06-20 16:15:41,302][starVLA.training.train_starvla][INFO] - Step 3802, Loss: {'train/loss': 0.13555696606636047, 'train/grad_norm_pre_clip': 7.763533592224121, 'timing/data': 0.0005334779853001237, 'timing/model': 1.4202883340185508, 'learning_rate/qwen_vl_interface': 8.862917072500785e-06, 'learning_rate/action_model': 8.862917072500784e-05, 'epoch': 7.03}) +[2026-06-20 16:15:45,806][starVLA.training.train_starvla][INFO] - Step 3803, Loss: {'train/loss': 0.044608913362026215, 'train/grad_norm_pre_clip': 3.0684497356414795, 'timing/data': 0.0005382569506764412, 'timing/model': 1.1670726599404588, 'learning_rate/qwen_vl_interface': 8.860939229517674e-06, 'learning_rate/action_model': 8.860939229517674e-05, 'epoch': 7.03}) +[2026-06-20 16:15:50,561][starVLA.training.train_starvla][INFO] - Step 3804, Loss: {'train/loss': 0.12678329646587372, 'train/grad_norm_pre_clip': 6.783209800720215, 'timing/data': 0.0005487669259309769, 'timing/model': 1.1696264679776505, 'learning_rate/qwen_vl_interface': 8.858959902213844e-06, 'learning_rate/action_model': 8.858959902213845e-05, 'epoch': 7.03}) +[2026-06-20 16:15:55,377][starVLA.training.train_starvla][INFO] - Step 3805, Loss: {'train/loss': 0.10031846910715103, 'train/grad_norm_pre_clip': 6.792888641357422, 'timing/data': 0.0008520749397575855, 'timing/model': 1.2074981580954045, 'learning_rate/qwen_vl_interface': 8.856979091402921e-06, 'learning_rate/action_model': 8.85697909140292e-05, 'epoch': 7.04}) +[2026-06-20 16:16:00,301][starVLA.training.train_starvla][INFO] - Step 3806, Loss: {'train/loss': 0.20462100207805634, 'train/grad_norm_pre_clip': 13.960338592529297, 'timing/data': 0.0005275779403746128, 'timing/model': 1.2369265040615574, 'learning_rate/qwen_vl_interface': 8.854996797899139e-06, 'learning_rate/action_model': 8.854996797899139e-05, 'epoch': 7.04}) +[2026-06-20 16:16:05,154][starVLA.training.train_starvla][INFO] - Step 3807, Loss: {'train/loss': 0.09886635094881058, 'train/grad_norm_pre_clip': 8.067983627319336, 'timing/data': 0.0005295879673212767, 'timing/model': 1.1708926489809528, 'learning_rate/qwen_vl_interface': 8.853013022517348e-06, 'learning_rate/action_model': 8.853013022517348e-05, 'epoch': 7.04}) +[2026-06-20 16:16:10,033][starVLA.training.train_starvla][INFO] - Step 3808, Loss: {'train/loss': 0.4633252024650574, 'train/grad_norm_pre_clip': 16.150981903076172, 'timing/data': 0.0005332380533218384, 'timing/model': 1.462546337977983, 'learning_rate/qwen_vl_interface': 8.851027766073002e-06, 'learning_rate/action_model': 8.851027766073001e-05, 'epoch': 7.04}) +[2026-06-20 16:16:14,679][starVLA.training.train_starvla][INFO] - Step 3809, Loss: {'train/loss': 0.07156559079885483, 'train/grad_norm_pre_clip': 7.482799053192139, 'timing/data': 0.0005310369888320565, 'timing/model': 1.2397654100786895, 'learning_rate/qwen_vl_interface': 8.84904102938216e-06, 'learning_rate/action_model': 8.84904102938216e-05, 'epoch': 7.04}) +[2026-06-20 16:16:19,508][starVLA.training.train_starvla][INFO] - Step 3810, Loss: {'train/loss': 0.09188289940357208, 'train/grad_norm_pre_clip': 8.718546867370605, 'timing/data': 0.0005349280545488, 'timing/model': 1.1779070519842207, 'learning_rate/qwen_vl_interface': 8.8470528132615e-06, 'learning_rate/action_model': 8.8470528132615e-05, 'epoch': 7.05}) +[2026-06-20 16:16:24,385][starVLA.training.train_starvla][INFO] - Step 3811, Loss: {'train/loss': 0.22889220714569092, 'train/grad_norm_pre_clip': 9.538288116455078, 'timing/data': 0.0005273170536383986, 'timing/model': 1.2133030500262976, 'learning_rate/qwen_vl_interface': 8.845063118528303e-06, 'learning_rate/action_model': 8.845063118528304e-05, 'epoch': 7.05}) +[2026-06-20 16:16:29,318][starVLA.training.train_starvla][INFO] - Step 3812, Loss: {'train/loss': 0.06362663954496384, 'train/grad_norm_pre_clip': 7.126495361328125, 'timing/data': 0.0005364770768210292, 'timing/model': 1.2434889330761507, 'learning_rate/qwen_vl_interface': 8.843071946000455e-06, 'learning_rate/action_model': 8.843071946000456e-05, 'epoch': 7.05}) +[2026-06-20 16:16:34,153][starVLA.training.train_starvla][INFO] - Step 3813, Loss: {'train/loss': 0.06145704537630081, 'train/grad_norm_pre_clip': 4.588691234588623, 'timing/data': 0.0005398779176175594, 'timing/model': 1.194045540993102, 'learning_rate/qwen_vl_interface': 8.841079296496451e-06, 'learning_rate/action_model': 8.84107929649645e-05, 'epoch': 7.05}) +[2026-06-20 16:16:38,952][starVLA.training.train_starvla][INFO] - Step 3814, Loss: {'train/loss': 0.09814008325338364, 'train/grad_norm_pre_clip': 7.981178283691406, 'timing/data': 0.000642877072095871, 'timing/model': 1.422458309913054, 'learning_rate/qwen_vl_interface': 8.839085170835396e-06, 'learning_rate/action_model': 8.839085170835395e-05, 'epoch': 7.05}) +[2026-06-20 16:16:43,520][starVLA.training.train_starvla][INFO] - Step 3815, Loss: {'train/loss': 0.09504349529743195, 'train/grad_norm_pre_clip': 7.123714447021484, 'timing/data': 0.0005452469922602177, 'timing/model': 1.2021632069954649, 'learning_rate/qwen_vl_interface': 8.837089569837e-06, 'learning_rate/action_model': 8.837089569836998e-05, 'epoch': 7.06}) +[2026-06-20 16:16:48,340][starVLA.training.train_starvla][INFO] - Step 3816, Loss: {'train/loss': 0.06957801431417465, 'train/grad_norm_pre_clip': 8.869232177734375, 'timing/data': 0.0005635070847347379, 'timing/model': 1.2002225749893114, 'learning_rate/qwen_vl_interface': 8.835092494321575e-06, 'learning_rate/action_model': 8.835092494321575e-05, 'epoch': 7.06}) +[2026-06-20 16:16:53,132][starVLA.training.train_starvla][INFO] - Step 3817, Loss: {'train/loss': 0.09532012045383453, 'train/grad_norm_pre_clip': 14.216869354248047, 'timing/data': 0.0005391179583966732, 'timing/model': 1.1800645589828491, 'learning_rate/qwen_vl_interface': 8.833093945110048e-06, 'learning_rate/action_model': 8.833093945110047e-05, 'epoch': 7.06}) +[2026-06-20 16:16:57,928][starVLA.training.train_starvla][INFO] - Step 3818, Loss: {'train/loss': 0.10081449151039124, 'train/grad_norm_pre_clip': 9.267830848693848, 'timing/data': 0.0005380680086091161, 'timing/model': 1.1781547350110486, 'learning_rate/qwen_vl_interface': 8.831093923023946e-06, 'learning_rate/action_model': 8.831093923023945e-05, 'epoch': 7.06}) +[2026-06-20 16:17:02,761][starVLA.training.train_starvla][INFO] - Step 3819, Loss: {'train/loss': 0.04243763908743858, 'train/grad_norm_pre_clip': 5.764682769775391, 'timing/data': 0.0005350880092009902, 'timing/model': 1.1787324199685827, 'learning_rate/qwen_vl_interface': 8.8290924288854e-06, 'learning_rate/action_model': 8.8290924288854e-05, 'epoch': 7.06}) +[2026-06-20 16:17:07,556][starVLA.training.train_starvla][INFO] - Step 3820, Loss: {'train/loss': 0.10571342706680298, 'train/grad_norm_pre_clip': 10.633724212646484, 'timing/data': 0.0005398578941822052, 'timing/model': 1.422770454082638, 'learning_rate/qwen_vl_interface': 8.827089463517153e-06, 'learning_rate/action_model': 8.827089463517153e-05, 'epoch': 7.06}) +[2026-06-20 16:17:12,121][starVLA.training.train_starvla][INFO] - Step 3821, Loss: {'train/loss': 0.14183823764324188, 'train/grad_norm_pre_clip': 13.377519607543945, 'timing/data': 0.0006553269922733307, 'timing/model': 1.176533145015128, 'learning_rate/qwen_vl_interface': 8.825085027742545e-06, 'learning_rate/action_model': 8.825085027742544e-05, 'epoch': 7.07}) +[2026-06-20 16:17:16,909][starVLA.training.train_starvla][INFO] - Step 3822, Loss: {'train/loss': 0.1035914495587349, 'train/grad_norm_pre_clip': 19.621061325073242, 'timing/data': 0.0005345470272004604, 'timing/model': 1.1783763320418075, 'learning_rate/qwen_vl_interface': 8.823079122385523e-06, 'learning_rate/action_model': 8.823079122385523e-05, 'epoch': 7.07}) +[2026-06-20 16:17:21,661][starVLA.training.train_starvla][INFO] - Step 3823, Loss: {'train/loss': 0.12223872542381287, 'train/grad_norm_pre_clip': 14.162003517150879, 'timing/data': 0.0007497159531340003, 'timing/model': 1.1755863219732419, 'learning_rate/qwen_vl_interface': 8.821071748270644e-06, 'learning_rate/action_model': 8.821071748270644e-05, 'epoch': 7.07}) +[2026-06-20 16:17:26,500][starVLA.training.train_starvla][INFO] - Step 3824, Loss: {'train/loss': 0.11331558972597122, 'train/grad_norm_pre_clip': 23.79057502746582, 'timing/data': 0.0005395569605752826, 'timing/model': 1.2063056050101295, 'learning_rate/qwen_vl_interface': 8.81906290622306e-06, 'learning_rate/action_model': 8.81906290622306e-05, 'epoch': 7.07}) +[2026-06-20 16:17:31,322][starVLA.training.train_starvla][INFO] - Step 3825, Loss: {'train/loss': 0.10085084289312363, 'train/grad_norm_pre_clip': 11.302167892456055, 'timing/data': 0.0005343070952221751, 'timing/model': 1.1765158800408244, 'learning_rate/qwen_vl_interface': 8.81705259706853e-06, 'learning_rate/action_model': 8.817052597068529e-05, 'epoch': 7.07}) +[2026-06-20 16:17:36,127][starVLA.training.train_starvla][INFO] - Step 3826, Loss: {'train/loss': 0.08727482706308365, 'train/grad_norm_pre_clip': 8.2254638671875, 'timing/data': 0.0005407480057328939, 'timing/model': 1.4242534571094438, 'learning_rate/qwen_vl_interface': 8.815040821633416e-06, 'learning_rate/action_model': 8.815040821633416e-05, 'epoch': 7.08}) +[2026-06-20 16:17:40,925][starVLA.training.train_starvla][INFO] - Step 3827, Loss: {'train/loss': 0.09840438514947891, 'train/grad_norm_pre_clip': 9.1461181640625, 'timing/data': 0.0005512069910764694, 'timing/model': 1.417273357976228, 'learning_rate/qwen_vl_interface': 8.813027580744685e-06, 'learning_rate/action_model': 8.813027580744684e-05, 'epoch': 7.08}) +[2026-06-20 16:17:45,516][starVLA.training.train_starvla][INFO] - Step 3828, Loss: {'train/loss': 0.11290293186903, 'train/grad_norm_pre_clip': 15.5360746383667, 'timing/data': 0.0005314269801601768, 'timing/model': 1.1967137950705364, 'learning_rate/qwen_vl_interface': 8.8110128752299e-06, 'learning_rate/action_model': 8.8110128752299e-05, 'epoch': 7.08}) +[2026-06-20 16:17:50,307][starVLA.training.train_starvla][INFO] - Step 3829, Loss: {'train/loss': 0.1906229853630066, 'train/grad_norm_pre_clip': 11.337715148925781, 'timing/data': 0.0005573580274358392, 'timing/model': 1.187354882946238, 'learning_rate/qwen_vl_interface': 8.808996705917235e-06, 'learning_rate/action_model': 8.808996705917235e-05, 'epoch': 7.08}) +[2026-06-20 16:17:55,344][starVLA.training.train_starvla][INFO] - Step 3830, Loss: {'train/loss': 0.3148139715194702, 'train/grad_norm_pre_clip': 21.297286987304688, 'timing/data': 0.0006364569999277592, 'timing/model': 1.186707641929388, 'learning_rate/qwen_vl_interface': 8.80697907363546e-06, 'learning_rate/action_model': 8.80697907363546e-05, 'epoch': 7.08}) +[2026-06-20 16:18:00,199][starVLA.training.train_starvla][INFO] - Step 3831, Loss: {'train/loss': 0.20995889604091644, 'train/grad_norm_pre_clip': 14.012346267700195, 'timing/data': 0.0005393780302256346, 'timing/model': 1.1798647500108927, 'learning_rate/qwen_vl_interface': 8.804959979213945e-06, 'learning_rate/action_model': 8.804959979213945e-05, 'epoch': 7.08}) +[2026-06-20 16:18:05,032][starVLA.training.train_starvla][INFO] - Step 3832, Loss: {'train/loss': 0.13651932775974274, 'train/grad_norm_pre_clip': 15.844935417175293, 'timing/data': 0.0005473169730976224, 'timing/model': 1.4490752479759976, 'learning_rate/qwen_vl_interface': 8.802939423482667e-06, 'learning_rate/action_model': 8.802939423482666e-05, 'epoch': 7.09}) +[2026-06-20 16:18:09,632][starVLA.training.train_starvla][INFO] - Step 3833, Loss: {'train/loss': 0.37431591749191284, 'train/grad_norm_pre_clip': 9.138609886169434, 'timing/data': 0.000542517052963376, 'timing/model': 1.2018996089464054, 'learning_rate/qwen_vl_interface': 8.800917407272198e-06, 'learning_rate/action_model': 8.800917407272198e-05, 'epoch': 7.09}) +[2026-06-20 16:18:14,437][starVLA.training.train_starvla][INFO] - Step 3834, Loss: {'train/loss': 0.01379117276519537, 'train/grad_norm_pre_clip': 1.2950527667999268, 'timing/data': 0.0005319879855960608, 'timing/model': 1.1824263330781832, 'learning_rate/qwen_vl_interface': 8.798893931413711e-06, 'learning_rate/action_model': 8.79889393141371e-05, 'epoch': 7.09}) +[2026-06-20 16:18:19,285][starVLA.training.train_starvla][INFO] - Step 3835, Loss: {'train/loss': 0.014934496022760868, 'train/grad_norm_pre_clip': 2.669520854949951, 'timing/data': 0.0005337969632819295, 'timing/model': 1.1785625569755211, 'learning_rate/qwen_vl_interface': 8.796868996738985e-06, 'learning_rate/action_model': 8.796868996738984e-05, 'epoch': 7.09}) +[2026-06-20 16:18:24,088][starVLA.training.train_starvla][INFO] - Step 3836, Loss: {'train/loss': 0.0030910528730601072, 'train/grad_norm_pre_clip': 0.2262052744626999, 'timing/data': 0.0005348579725250602, 'timing/model': 1.1769262680318207, 'learning_rate/qwen_vl_interface': 8.794842604080389e-06, 'learning_rate/action_model': 8.794842604080389e-05, 'epoch': 7.09}) +[2026-06-20 16:18:28,892][starVLA.training.train_starvla][INFO] - Step 3837, Loss: {'train/loss': 0.13275808095932007, 'train/grad_norm_pre_clip': 12.727540969848633, 'timing/data': 0.0006286679999902844, 'timing/model': 1.1748347729444504, 'learning_rate/qwen_vl_interface': 8.7928147542709e-06, 'learning_rate/action_model': 8.7928147542709e-05, 'epoch': 7.1}) +[2026-06-20 16:18:33,658][starVLA.training.train_starvla][INFO] - Step 3838, Loss: {'train/loss': 0.07554031908512115, 'train/grad_norm_pre_clip': 13.572251319885254, 'timing/data': 0.0005475770449265838, 'timing/model': 1.387619503075257, 'learning_rate/qwen_vl_interface': 8.790785448144088e-06, 'learning_rate/action_model': 8.790785448144088e-05, 'epoch': 7.1}) +[2026-06-20 16:18:38,231][starVLA.training.train_starvla][INFO] - Step 3839, Loss: {'train/loss': 0.024641696363687515, 'train/grad_norm_pre_clip': 11.163559913635254, 'timing/data': 0.0005522369174286723, 'timing/model': 1.201154383015819, 'learning_rate/qwen_vl_interface': 8.788754686534126e-06, 'learning_rate/action_model': 8.788754686534127e-05, 'epoch': 7.1}) +[2026-06-20 16:18:43,050][starVLA.training.train_starvla][INFO] - Step 3840, Loss: {'train/loss': 0.01361594907939434, 'train/grad_norm_pre_clip': 1.2394154071807861, 'timing/data': 0.0005433779442682862, 'timing/model': 1.1907721729949117, 'learning_rate/qwen_vl_interface': 8.786722470275784e-06, 'learning_rate/action_model': 8.786722470275784e-05, 'epoch': 7.1}) +[2026-06-20 16:18:47,917][starVLA.training.train_starvla][INFO] - Step 3841, Loss: {'train/loss': 0.305525541305542, 'train/grad_norm_pre_clip': 17.333032608032227, 'timing/data': 0.0005422279937192798, 'timing/model': 1.177328846999444, 'learning_rate/qwen_vl_interface': 8.784688800204427e-06, 'learning_rate/action_model': 8.784688800204426e-05, 'epoch': 7.1}) +[2026-06-20 16:18:52,752][starVLA.training.train_starvla][INFO] - Step 3842, Loss: {'train/loss': 0.16441410779953003, 'train/grad_norm_pre_clip': 11.920466423034668, 'timing/data': 0.0005267879460006952, 'timing/model': 1.1766149739269167, 'learning_rate/qwen_vl_interface': 8.782653677156022e-06, 'learning_rate/action_model': 8.782653677156022e-05, 'epoch': 7.11}) +[2026-06-20 16:18:57,565][starVLA.training.train_starvla][INFO] - Step 3843, Loss: {'train/loss': 0.010452117770910263, 'train/grad_norm_pre_clip': 1.5443683862686157, 'timing/data': 0.0005294380243867636, 'timing/model': 1.1781347920186818, 'learning_rate/qwen_vl_interface': 8.780617101967129e-06, 'learning_rate/action_model': 8.780617101967129e-05, 'epoch': 7.11}) +[2026-06-20 16:19:02,408][starVLA.training.train_starvla][INFO] - Step 3844, Loss: {'train/loss': 0.12047276645898819, 'train/grad_norm_pre_clip': 13.411423683166504, 'timing/data': 0.0005545479943975806, 'timing/model': 1.412043854012154, 'learning_rate/qwen_vl_interface': 8.77857907547491e-06, 'learning_rate/action_model': 8.77857907547491e-05, 'epoch': 7.11}) +[2026-06-20 16:19:07,000][starVLA.training.train_starvla][INFO] - Step 3845, Loss: {'train/loss': 0.37955671548843384, 'train/grad_norm_pre_clip': 24.264944076538086, 'timing/data': 0.000562497996725142, 'timing/model': 1.1819371019955724, 'learning_rate/qwen_vl_interface': 8.776539598517119e-06, 'learning_rate/action_model': 8.776539598517118e-05, 'epoch': 7.11}) +[2026-06-20 16:19:11,834][starVLA.training.train_starvla][INFO] - Step 3846, Loss: {'train/loss': 0.4357941150665283, 'train/grad_norm_pre_clip': 16.77243423461914, 'timing/data': 0.0005417380016297102, 'timing/model': 1.1958195840707049, 'learning_rate/qwen_vl_interface': 8.774498671932108e-06, 'learning_rate/action_model': 8.774498671932107e-05, 'epoch': 7.11}) +[2026-06-20 16:19:16,687][starVLA.training.train_starvla][INFO] - Step 3847, Loss: {'train/loss': 0.13223913311958313, 'train/grad_norm_pre_clip': 29.1549015045166, 'timing/data': 0.0006824869196861982, 'timing/model': 1.1909520289627835, 'learning_rate/qwen_vl_interface': 8.772456296558824e-06, 'learning_rate/action_model': 8.772456296558824e-05, 'epoch': 7.11}) +[2026-06-20 16:19:21,509][starVLA.training.train_starvla][INFO] - Step 3848, Loss: {'train/loss': 0.10665375739336014, 'train/grad_norm_pre_clip': 13.919341087341309, 'timing/data': 0.00052931800018996, 'timing/model': 1.1812492549652234, 'learning_rate/qwen_vl_interface': 8.77041247323681e-06, 'learning_rate/action_model': 8.770412473236811e-05, 'epoch': 7.12}) +[2026-06-20 16:19:26,331][starVLA.training.train_starvla][INFO] - Step 3849, Loss: {'train/loss': 0.13478019833564758, 'train/grad_norm_pre_clip': 16.617115020751953, 'timing/data': 0.0005422969115898013, 'timing/model': 1.4448782210238278, 'learning_rate/qwen_vl_interface': 8.76836720280621e-06, 'learning_rate/action_model': 8.768367202806209e-05, 'epoch': 7.12}) +[2026-06-20 16:19:30,913][starVLA.training.train_starvla][INFO] - Step 3850, Loss: {'train/loss': 0.26505744457244873, 'train/grad_norm_pre_clip': 7.977733612060547, 'timing/data': 0.0005287469830363989, 'timing/model': 1.1782300540944561, 'learning_rate/qwen_vl_interface': 8.766320486107749e-06, 'learning_rate/action_model': 8.766320486107749e-05, 'epoch': 7.12}) +[2026-06-20 16:19:35,733][starVLA.training.train_starvla][INFO] - Step 3851, Loss: {'train/loss': 0.4483093321323395, 'train/grad_norm_pre_clip': 18.636247634887695, 'timing/data': 0.0005301979836076498, 'timing/model': 1.1780620669014752, 'learning_rate/qwen_vl_interface': 8.76427232398276e-06, 'learning_rate/action_model': 8.76427232398276e-05, 'epoch': 7.12}) +[2026-06-20 16:19:40,585][starVLA.training.train_starvla][INFO] - Step 3852, Loss: {'train/loss': 0.137991264462471, 'train/grad_norm_pre_clip': 15.645593643188477, 'timing/data': 0.000529618002474308, 'timing/model': 1.1914954929379746, 'learning_rate/qwen_vl_interface': 8.762222717273167e-06, 'learning_rate/action_model': 8.762222717273167e-05, 'epoch': 7.12}) +[2026-06-20 16:19:45,359][starVLA.training.train_starvla][INFO] - Step 3853, Loss: {'train/loss': 0.005089310929179192, 'train/grad_norm_pre_clip': 0.5236185789108276, 'timing/data': 0.0005336269969120622, 'timing/model': 1.178938459022902, 'learning_rate/qwen_vl_interface': 8.76017166682148e-06, 'learning_rate/action_model': 8.760171666821481e-05, 'epoch': 7.13}) +[2026-06-20 16:19:50,134][starVLA.training.train_starvla][INFO] - Step 3854, Loss: {'train/loss': 0.08524161577224731, 'train/grad_norm_pre_clip': 13.25581169128418, 'timing/data': 0.0005101580172777176, 'timing/model': 1.1772162590641528, 'learning_rate/qwen_vl_interface': 8.758119173470815e-06, 'learning_rate/action_model': 8.758119173470815e-05, 'epoch': 7.13}) +[2026-06-20 16:19:54,944][starVLA.training.train_starvla][INFO] - Step 3855, Loss: {'train/loss': 0.12775921821594238, 'train/grad_norm_pre_clip': 19.86550521850586, 'timing/data': 0.0005129080964252353, 'timing/model': 1.423213675036095, 'learning_rate/qwen_vl_interface': 8.75606523806487e-06, 'learning_rate/action_model': 8.756065238064868e-05, 'epoch': 7.13}) +[2026-06-20 16:19:59,504][starVLA.training.train_starvla][INFO] - Step 3856, Loss: {'train/loss': 0.06821974366903305, 'train/grad_norm_pre_clip': 10.347794532775879, 'timing/data': 0.000523227034136653, 'timing/model': 1.1803161399438977, 'learning_rate/qwen_vl_interface': 8.75400986144794e-06, 'learning_rate/action_model': 8.75400986144794e-05, 'epoch': 7.13}) +[2026-06-20 16:20:04,319][starVLA.training.train_starvla][INFO] - Step 3857, Loss: {'train/loss': 0.3520865738391876, 'train/grad_norm_pre_clip': 11.688603401184082, 'timing/data': 0.0005461979890242219, 'timing/model': 1.176282348926179, 'learning_rate/qwen_vl_interface': 8.751953044464913e-06, 'learning_rate/action_model': 8.751953044464913e-05, 'epoch': 7.13}) +[2026-06-20 16:20:09,212][starVLA.training.train_starvla][INFO] - Step 3858, Loss: {'train/loss': 0.03703286498785019, 'train/grad_norm_pre_clip': 6.227647304534912, 'timing/data': 0.0005320180207490921, 'timing/model': 1.19049259100575, 'learning_rate/qwen_vl_interface': 8.749894787961272e-06, 'learning_rate/action_model': 8.749894787961271e-05, 'epoch': 7.13}) +[2026-06-20 16:20:14,028][starVLA.training.train_starvla][INFO] - Step 3859, Loss: {'train/loss': 0.005336257629096508, 'train/grad_norm_pre_clip': 0.6221637725830078, 'timing/data': 0.0005184180336073041, 'timing/model': 1.1801364870043471, 'learning_rate/qwen_vl_interface': 8.747835092783086e-06, 'learning_rate/action_model': 8.747835092783085e-05, 'epoch': 7.14}) +[2026-06-20 16:20:18,788][starVLA.training.train_starvla][INFO] - Step 3860, Loss: {'train/loss': 0.11574944108724594, 'train/grad_norm_pre_clip': 14.812060356140137, 'timing/data': 0.0005207679932937026, 'timing/model': 1.1750730889616534, 'learning_rate/qwen_vl_interface': 8.745773959777018e-06, 'learning_rate/action_model': 8.745773959777017e-05, 'epoch': 7.14}) +[2026-06-20 16:20:23,762][starVLA.training.train_starvla][INFO] - Step 3861, Loss: {'train/loss': 0.007995665073394775, 'train/grad_norm_pre_clip': 0.7121520042419434, 'timing/data': 0.0005285879597067833, 'timing/model': 1.4761640189681202, 'learning_rate/qwen_vl_interface': 8.743711389790321e-06, 'learning_rate/action_model': 8.74371138979032e-05, 'epoch': 7.14}) +[2026-06-20 16:20:28,416][starVLA.training.train_starvla][INFO] - Step 3862, Loss: {'train/loss': 0.04659929871559143, 'train/grad_norm_pre_clip': 7.96162748336792, 'timing/data': 0.0005214270204305649, 'timing/model': 1.2254575421102345, 'learning_rate/qwen_vl_interface': 8.741647383670839e-06, 'learning_rate/action_model': 8.741647383670839e-05, 'epoch': 7.14}) +[2026-06-20 16:20:33,277][starVLA.training.train_starvla][INFO] - Step 3863, Loss: {'train/loss': 0.027655838057398796, 'train/grad_norm_pre_clip': 2.6509599685668945, 'timing/data': 0.0005386279663071036, 'timing/model': 1.1760686910711229, 'learning_rate/qwen_vl_interface': 8.739581942267011e-06, 'learning_rate/action_model': 8.739581942267011e-05, 'epoch': 7.14}) +[2026-06-20 16:20:38,169][starVLA.training.train_starvla][INFO] - Step 3864, Loss: {'train/loss': 0.2918452024459839, 'train/grad_norm_pre_clip': 19.41143226623535, 'timing/data': 0.0006848969496786594, 'timing/model': 1.1772573369089514, 'learning_rate/qwen_vl_interface': 8.737515066427861e-06, 'learning_rate/action_model': 8.73751506642786e-05, 'epoch': 7.15}) +[2026-06-20 16:20:42,984][starVLA.training.train_starvla][INFO] - Step 3865, Loss: {'train/loss': 0.11780697107315063, 'train/grad_norm_pre_clip': 16.94306755065918, 'timing/data': 0.0005239680176600814, 'timing/model': 1.1729923560051247, 'learning_rate/qwen_vl_interface': 8.735446757003e-06, 'learning_rate/action_model': 8.735446757003e-05, 'epoch': 7.15}) +[2026-06-20 16:20:47,761][starVLA.training.train_starvla][INFO] - Step 3866, Loss: {'train/loss': 0.4922691881656647, 'train/grad_norm_pre_clip': 29.05099105834961, 'timing/data': 0.0006198569899424911, 'timing/model': 1.4115521690109745, 'learning_rate/qwen_vl_interface': 8.733377014842635e-06, 'learning_rate/action_model': 8.733377014842635e-05, 'epoch': 7.15}) +[2026-06-20 16:20:52,484][starVLA.training.train_starvla][INFO] - Step 3867, Loss: {'train/loss': 0.11850101500749588, 'train/grad_norm_pre_clip': 4.617904186248779, 'timing/data': 0.00060760707128793, 'timing/model': 1.3297104539815336, 'learning_rate/qwen_vl_interface': 8.731305840797557e-06, 'learning_rate/action_model': 8.731305840797558e-05, 'epoch': 7.15}) +[2026-06-20 16:20:57,055][starVLA.training.train_starvla][INFO] - Step 3868, Loss: {'train/loss': 0.26763761043548584, 'train/grad_norm_pre_clip': 20.483230590820312, 'timing/data': 0.0006198280025273561, 'timing/model': 1.1824793990235776, 'learning_rate/qwen_vl_interface': 8.72923323571915e-06, 'learning_rate/action_model': 8.72923323571915e-05, 'epoch': 7.15}) +[2026-06-20 16:21:01,996][starVLA.training.train_starvla][INFO] - Step 3869, Loss: {'train/loss': 0.3141132891178131, 'train/grad_norm_pre_clip': 13.384329795837402, 'timing/data': 0.0005362380761653185, 'timing/model': 1.1733432080363855, 'learning_rate/qwen_vl_interface': 8.727159200459381e-06, 'learning_rate/action_model': 8.727159200459381e-05, 'epoch': 7.16}) +[2026-06-20 16:21:07,013][starVLA.training.train_starvla][INFO] - Step 3870, Loss: {'train/loss': 0.05573878437280655, 'train/grad_norm_pre_clip': 9.408677101135254, 'timing/data': 0.0005197679856792092, 'timing/model': 1.232484873966314, 'learning_rate/qwen_vl_interface': 8.725083735870807e-06, 'learning_rate/action_model': 8.725083735870807e-05, 'epoch': 7.16}) +[2026-06-20 16:21:11,865][starVLA.training.train_starvla][INFO] - Step 3871, Loss: {'train/loss': 0.3670291006565094, 'train/grad_norm_pre_clip': 14.778054237365723, 'timing/data': 0.0005339080234989524, 'timing/model': 1.238875049049966, 'learning_rate/qwen_vl_interface': 8.723006842806575e-06, 'learning_rate/action_model': 8.723006842806576e-05, 'epoch': 7.16}) +[2026-06-20 16:21:16,711][starVLA.training.train_starvla][INFO] - Step 3872, Loss: {'train/loss': 0.248349130153656, 'train/grad_norm_pre_clip': 97.0756607055664, 'timing/data': 0.0005271780537441373, 'timing/model': 1.405272213043645, 'learning_rate/qwen_vl_interface': 8.720928522120419e-06, 'learning_rate/action_model': 8.720928522120418e-05, 'epoch': 7.16}) +[2026-06-20 16:21:21,409][starVLA.training.train_starvla][INFO] - Step 3873, Loss: {'train/loss': 0.20431400835514069, 'train/grad_norm_pre_clip': 16.5415096282959, 'timing/data': 0.0005038679810240865, 'timing/model': 1.3449097250122577, 'learning_rate/qwen_vl_interface': 8.718848774666655e-06, 'learning_rate/action_model': 8.718848774666655e-05, 'epoch': 7.16}) +[2026-06-20 16:21:25,936][starVLA.training.train_starvla][INFO] - Step 3874, Loss: {'train/loss': 0.05768904089927673, 'train/grad_norm_pre_clip': 8.145981788635254, 'timing/data': 0.000508237979374826, 'timing/model': 1.176702696015127, 'learning_rate/qwen_vl_interface': 8.716767601300189e-06, 'learning_rate/action_model': 8.716767601300188e-05, 'epoch': 7.16}) +[2026-06-20 16:21:52,789][starVLA.training.train_starvla][INFO] - Step 3875, Loss: {'train/loss': 0.22527994215488434, 'train/grad_norm_pre_clip': 18.322507858276367, 'timing/data': 19.838451171992347, 'timing/model': 1.2490199590101838, 'learning_rate/qwen_vl_interface': 8.714685002876514e-06, 'learning_rate/action_model': 8.714685002876513e-05, 'epoch': 7.17}) +[2026-06-20 16:21:57,652][starVLA.training.train_starvla][INFO] - Step 3876, Loss: {'train/loss': 0.12631778419017792, 'train/grad_norm_pre_clip': 12.09196949005127, 'timing/data': 0.007781057967804372, 'timing/model': 1.2177905059652403, 'learning_rate/qwen_vl_interface': 8.712600980251706e-06, 'learning_rate/action_model': 8.712600980251705e-05, 'epoch': 7.17}) +[2026-06-20 16:22:02,496][starVLA.training.train_starvla][INFO] - Step 3877, Loss: {'train/loss': 0.03129153698682785, 'train/grad_norm_pre_clip': 7.42265510559082, 'timing/data': 0.0076514590764418244, 'timing/model': 1.2176408569794148, 'learning_rate/qwen_vl_interface': 8.710515534282428e-06, 'learning_rate/action_model': 8.710515534282428e-05, 'epoch': 7.17}) +[2026-06-20 16:22:07,380][starVLA.training.train_starvla][INFO] - Step 3878, Loss: {'train/loss': 0.14767563343048096, 'train/grad_norm_pre_clip': 10.630120277404785, 'timing/data': 0.006391964037902653, 'timing/model': 1.4593320220010355, 'learning_rate/qwen_vl_interface': 8.708428665825931e-06, 'learning_rate/action_model': 8.708428665825931e-05, 'epoch': 7.17}) +[2026-06-20 16:22:11,948][starVLA.training.train_starvla][INFO] - Step 3879, Loss: {'train/loss': 0.2551431953907013, 'train/grad_norm_pre_clip': 19.651649475097656, 'timing/data': 0.006371543975546956, 'timing/model': 1.1766251110238954, 'learning_rate/qwen_vl_interface': 8.706340375740045e-06, 'learning_rate/action_model': 8.706340375740045e-05, 'epoch': 7.17}) +[2026-06-20 16:22:16,753][starVLA.training.train_starvla][INFO] - Step 3880, Loss: {'train/loss': 0.1708628088235855, 'train/grad_norm_pre_clip': 9.454571723937988, 'timing/data': 0.006426414009183645, 'timing/model': 1.1825305769452825, 'learning_rate/qwen_vl_interface': 8.704250664883188e-06, 'learning_rate/action_model': 8.704250664883188e-05, 'epoch': 7.18}) +[2026-06-20 16:22:21,526][starVLA.training.train_starvla][INFO] - Step 3881, Loss: {'train/loss': 0.061591170728206635, 'train/grad_norm_pre_clip': 5.638495922088623, 'timing/data': 0.006498283939436078, 'timing/model': 1.1727173359831795, 'learning_rate/qwen_vl_interface': 8.70215953411436e-06, 'learning_rate/action_model': 8.70215953411436e-05, 'epoch': 7.18}) +[2026-06-20 16:22:26,347][starVLA.training.train_starvla][INFO] - Step 3882, Loss: {'train/loss': 0.16252368688583374, 'train/grad_norm_pre_clip': 13.397238731384277, 'timing/data': 0.006442774087190628, 'timing/model': 1.1868360179942101, 'learning_rate/qwen_vl_interface': 8.70006698429315e-06, 'learning_rate/action_model': 8.700066984293149e-05, 'epoch': 7.18}) +[2026-06-20 16:22:31,192][starVLA.training.train_starvla][INFO] - Step 3883, Loss: {'train/loss': 0.16339848935604095, 'train/grad_norm_pre_clip': 15.210033416748047, 'timing/data': 0.006520393071696162, 'timing/model': 1.1757682519964874, 'learning_rate/qwen_vl_interface': 8.697973016279722e-06, 'learning_rate/action_model': 8.697973016279722e-05, 'epoch': 7.18}) +[2026-06-20 16:22:35,994][starVLA.training.train_starvla][INFO] - Step 3884, Loss: {'train/loss': 0.018269073218107224, 'train/grad_norm_pre_clip': 1.0303322076797485, 'timing/data': 0.006280855042859912, 'timing/model': 1.4254848940763623, 'learning_rate/qwen_vl_interface': 8.695877630934827e-06, 'learning_rate/action_model': 8.695877630934826e-05, 'epoch': 7.18}) +[2026-06-20 16:22:40,536][starVLA.training.train_starvla][INFO] - Step 3885, Loss: {'train/loss': 0.10499155521392822, 'train/grad_norm_pre_clip': 13.92600154876709, 'timing/data': 0.00645508395973593, 'timing/model': 1.1840235670097172, 'learning_rate/qwen_vl_interface': 8.693780829119801e-06, 'learning_rate/action_model': 8.693780829119801e-05, 'epoch': 7.18}) +[2026-06-20 16:22:45,395][starVLA.training.train_starvla][INFO] - Step 3886, Loss: {'train/loss': 0.06933347880840302, 'train/grad_norm_pre_clip': 7.387646675109863, 'timing/data': 0.00688507198356092, 'timing/model': 1.1887425769818947, 'learning_rate/qwen_vl_interface': 8.69168261169656e-06, 'learning_rate/action_model': 8.691682611696559e-05, 'epoch': 7.19}) +[2026-06-20 16:22:50,204][starVLA.training.train_starvla][INFO] - Step 3887, Loss: {'train/loss': 0.17159141600131989, 'train/grad_norm_pre_clip': 10.806201934814453, 'timing/data': 0.007035022019408643, 'timing/model': 1.1760935951024294, 'learning_rate/qwen_vl_interface': 8.689582979527599e-06, 'learning_rate/action_model': 8.689582979527599e-05, 'epoch': 7.19}) +[2026-06-20 16:22:55,057][starVLA.training.train_starvla][INFO] - Step 3888, Loss: {'train/loss': 0.20225219428539276, 'train/grad_norm_pre_clip': 10.052841186523438, 'timing/data': 0.006021416047587991, 'timing/model': 1.2013230549637228, 'learning_rate/qwen_vl_interface': 8.687481933476e-06, 'learning_rate/action_model': 8.687481933476e-05, 'epoch': 7.19}) +[2026-06-20 16:22:59,985][starVLA.training.train_starvla][INFO] - Step 3889, Loss: {'train/loss': 0.04851694777607918, 'train/grad_norm_pre_clip': 6.447434902191162, 'timing/data': 0.006431134999729693, 'timing/model': 1.4433337970403954, 'learning_rate/qwen_vl_interface': 8.685379474405423e-06, 'learning_rate/action_model': 8.685379474405423e-05, 'epoch': 7.19}) +[2026-06-20 16:23:04,884][starVLA.training.train_starvla][INFO] - Step 3890, Loss: {'train/loss': 0.24553826451301575, 'train/grad_norm_pre_clip': 7.797040939331055, 'timing/data': 0.00760393007658422, 'timing/model': 1.4126799680525437, 'learning_rate/qwen_vl_interface': 8.68327560318011e-06, 'learning_rate/action_model': 8.68327560318011e-05, 'epoch': 7.19}) +[2026-06-20 16:23:09,621][starVLA.training.train_starvla][INFO] - Step 3891, Loss: {'train/loss': 0.2546999156475067, 'train/grad_norm_pre_clip': 11.163639068603516, 'timing/data': 0.007628818973898888, 'timing/model': 1.1800567649770528, 'learning_rate/qwen_vl_interface': 8.681170320664883e-06, 'learning_rate/action_model': 8.681170320664883e-05, 'epoch': 7.2}) +[2026-06-20 16:23:14,548][starVLA.training.train_starvla][INFO] - Step 3892, Loss: {'train/loss': 0.10444764792919159, 'train/grad_norm_pre_clip': 19.24266242980957, 'timing/data': 0.006475504953414202, 'timing/model': 1.174786094110459, 'learning_rate/qwen_vl_interface': 8.679063627725143e-06, 'learning_rate/action_model': 8.679063627725144e-05, 'epoch': 7.2}) +[2026-06-20 16:23:19,344][starVLA.training.train_starvla][INFO] - Step 3893, Loss: {'train/loss': 0.0780283510684967, 'train/grad_norm_pre_clip': 5.6052470207214355, 'timing/data': 0.00830740702804178, 'timing/model': 1.180127820931375, 'learning_rate/qwen_vl_interface': 8.676955525226876e-06, 'learning_rate/action_model': 8.676955525226875e-05, 'epoch': 7.2}) +[2026-06-20 16:23:24,274][starVLA.training.train_starvla][INFO] - Step 3894, Loss: {'train/loss': 0.10826420038938522, 'train/grad_norm_pre_clip': 10.151005744934082, 'timing/data': 0.00763725908473134, 'timing/model': 1.1881034990074113, 'learning_rate/qwen_vl_interface': 8.674846014036636e-06, 'learning_rate/action_model': 8.674846014036636e-05, 'epoch': 7.2}) +[2026-06-20 16:23:29,713][starVLA.training.train_starvla][INFO] - Step 3895, Loss: {'train/loss': 0.04773435741662979, 'train/grad_norm_pre_clip': 9.686914443969727, 'timing/data': 0.009863901999779046, 'timing/model': 1.6483495290158316, 'learning_rate/qwen_vl_interface': 8.672735095021569e-06, 'learning_rate/action_model': 8.672735095021569e-05, 'epoch': 7.2}) +[2026-06-20 16:23:35,195][starVLA.training.train_starvla][INFO] - Step 3896, Loss: {'train/loss': 0.06140662729740143, 'train/grad_norm_pre_clip': 7.681384563446045, 'timing/data': 0.01047432899940759, 'timing/model': 1.6286745590623468, 'learning_rate/qwen_vl_interface': 8.670622769049393e-06, 'learning_rate/action_model': 8.670622769049392e-05, 'epoch': 7.21}) +[2026-06-20 16:23:40,418][starVLA.training.train_starvla][INFO] - Step 3897, Loss: {'train/loss': 0.15875482559204102, 'train/grad_norm_pre_clip': 13.517497062683105, 'timing/data': 0.010360950022004545, 'timing/model': 1.3486235910095274, 'learning_rate/qwen_vl_interface': 8.668509036988406e-06, 'learning_rate/action_model': 8.668509036988405e-05, 'epoch': 7.21}) +[2026-06-20 16:23:45,913][starVLA.training.train_starvla][INFO] - Step 3898, Loss: {'train/loss': 0.1436036080121994, 'train/grad_norm_pre_clip': 12.967978477478027, 'timing/data': 0.010258489986881614, 'timing/model': 1.3479297129670158, 'learning_rate/qwen_vl_interface': 8.66639389970748e-06, 'learning_rate/action_model': 8.666393899707479e-05, 'epoch': 7.21}) +[2026-06-20 16:23:51,420][starVLA.training.train_starvla][INFO] - Step 3899, Loss: {'train/loss': 0.22614514827728271, 'train/grad_norm_pre_clip': 14.946306228637695, 'timing/data': 0.010260991053655744, 'timing/model': 1.348273611976765, 'learning_rate/qwen_vl_interface': 8.66427735807607e-06, 'learning_rate/action_model': 8.664277358076069e-05, 'epoch': 7.21}) +[2026-06-20 16:23:56,947][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 3900 +[2026-06-20 16:30:30,412][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 3900 +[2026-06-20 16:30:30,413][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 3900 +[2026-06-20 16:37:29,378][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 3900 +[2026-06-20 16:37:29,380][starVLA.training.train_starvla][INFO] - Step 3900, Loss: {'train/loss': 0.14822235703468323, 'train/grad_norm_pre_clip': 9.466261863708496, 'eval/loss': 0.6252211928367615, 'eval/latency_0/loss': 0.6673551797866821, 'eval/latency_1/loss': 0.8857616186141968, 'eval/latency_2/loss': 0.9705256819725037, 'eval/latency_3/loss': 0.460584819316864, 'eval/latency_4/loss': 0.10762688517570496, 'eval/flappy/loss': 0.6252211928367615, 'eval/flappy/latency_0/loss': 0.6673551797866821, 'eval/flappy/latency_1/loss': 0.8857616186141968, 'eval/flappy/latency_2/loss': 0.9705256819725037, 'eval/flappy/latency_3/loss': 0.460584819316864, 'eval/flappy/latency_4/loss': 0.10762688517570496, 'eval/flap_precision': 0.3068181818181818, 'eval/flap_recall': 0.09818181818181818, 'eval/flap_f1': 0.14876033057851237, 'eval/flap_rate_model': 0.024444444444444446, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.14876033057851237, 'eval/latency_0/flap_precision': 0.3068181818181818, 'eval/latency_0/flap_recall': 0.09818181818181818, 'eval/latency_0/flap_f1': 0.14876033057851237, 'eval/latency_0/flap_rate_model': 0.024444444444444446, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.14876033057851237, 'timing/data': 0.010315090999938548, 'timing/model': 1.3448664359748363, 'learning_rate/qwen_vl_interface': 8.662159412964208e-06, 'learning_rate/action_model': 8.662159412964207e-05, 'epoch': 7.21}) +[2026-06-20 16:37:29,380][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_3900_state +[2026-06-20 16:37:29,384][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-20 16:37:38,009][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_3900_state/model.safetensors +[2026-06-20 16:37:51,019][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_3900_state/optimizer.bin +[2026-06-20 16:37:51,020][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_3900_state/sampler.bin +[2026-06-20 16:37:51,020][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_3900_state/sampler_1.bin +[2026-06-20 16:37:51,021][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_3900_state/random_states_0.pkl +[2026-06-20 16:37:53,630][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3900 +[2026-06-20 16:37:53,630][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-20 16:37:53,645][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-20 16:37:58,635][starVLA.training.train_starvla][INFO] - Step 3901, Loss: {'train/loss': 0.26537445187568665, 'train/grad_norm_pre_clip': 8.640508651733398, 'timing/data': 0.007256988901644945, 'timing/model': 1.229047448025085, 'learning_rate/qwen_vl_interface': 8.660040065242499e-06, 'learning_rate/action_model': 8.660040065242498e-05, 'epoch': 7.21}) +[2026-06-20 16:38:03,635][starVLA.training.train_starvla][INFO] - Step 3902, Loss: {'train/loss': 0.0713270902633667, 'train/grad_norm_pre_clip': 9.175127029418945, 'timing/data': 0.006422638078220189, 'timing/model': 1.5234702880261466, 'learning_rate/qwen_vl_interface': 8.657919315782127e-06, 'learning_rate/action_model': 8.657919315782127e-05, 'epoch': 7.22}) +[2026-06-20 16:38:08,671][starVLA.training.train_starvla][INFO] - Step 3903, Loss: {'train/loss': 0.05451072379946709, 'train/grad_norm_pre_clip': 4.510594367980957, 'timing/data': 0.0065451470436528325, 'timing/model': 1.2993000019341707, 'learning_rate/qwen_vl_interface': 8.655797165454853e-06, 'learning_rate/action_model': 8.655797165454853e-05, 'epoch': 7.22}) +[2026-06-20 16:38:13,924][starVLA.training.train_starvla][INFO] - Step 3904, Loss: {'train/loss': 0.05653596669435501, 'train/grad_norm_pre_clip': 7.617049694061279, 'timing/data': 0.006292990990914404, 'timing/model': 1.2887640510452911, 'learning_rate/qwen_vl_interface': 8.653673615133013e-06, 'learning_rate/action_model': 8.653673615133013e-05, 'epoch': 7.22}) +[2026-06-20 16:38:19,104][starVLA.training.train_starvla][INFO] - Step 3905, Loss: {'train/loss': 0.07853177189826965, 'train/grad_norm_pre_clip': 6.168482780456543, 'timing/data': 0.0065123289823532104, 'timing/model': 1.2961348519893363, 'learning_rate/qwen_vl_interface': 8.651548665689519e-06, 'learning_rate/action_model': 8.651548665689518e-05, 'epoch': 7.22}) +[2026-06-20 16:38:24,366][starVLA.training.train_starvla][INFO] - Step 3906, Loss: {'train/loss': 0.07419203966856003, 'train/grad_norm_pre_clip': 6.464465141296387, 'timing/data': 0.006489339051768184, 'timing/model': 1.2905568679561839, 'learning_rate/qwen_vl_interface': 8.649422317997855e-06, 'learning_rate/action_model': 8.649422317997856e-05, 'epoch': 7.22}) +[2026-06-20 16:38:29,244][starVLA.training.train_starvla][INFO] - Step 3907, Loss: {'train/loss': 0.13670536875724792, 'train/grad_norm_pre_clip': 12.876104354858398, 'timing/data': 0.005840766942128539, 'timing/model': 1.16949672694318, 'learning_rate/qwen_vl_interface': 8.647294572932082e-06, 'learning_rate/action_model': 8.647294572932082e-05, 'epoch': 7.23}) +[2026-06-20 16:38:34,006][starVLA.training.train_starvla][INFO] - Step 3908, Loss: {'train/loss': 0.028908202424645424, 'train/grad_norm_pre_clip': 9.008575439453125, 'timing/data': 0.005868306965567172, 'timing/model': 1.4067071590106934, 'learning_rate/qwen_vl_interface': 8.64516543136684e-06, 'learning_rate/action_model': 8.64516543136684e-05, 'epoch': 7.23}) +[2026-06-20 16:38:38,755][starVLA.training.train_starvla][INFO] - Step 3909, Loss: {'train/loss': 0.32330983877182007, 'train/grad_norm_pre_clip': 12.391347885131836, 'timing/data': 0.00597536691930145, 'timing/model': 1.4076674790121615, 'learning_rate/qwen_vl_interface': 8.643034894177335e-06, 'learning_rate/action_model': 8.643034894177335e-05, 'epoch': 7.23}) +[2026-06-20 16:38:43,275][starVLA.training.train_starvla][INFO] - Step 3910, Loss: {'train/loss': 0.04097365960478783, 'train/grad_norm_pre_clip': 2.584197998046875, 'timing/data': 0.005853677052073181, 'timing/model': 1.1731444710167125, 'learning_rate/qwen_vl_interface': 8.64090296223935e-06, 'learning_rate/action_model': 8.64090296223935e-05, 'epoch': 7.23}) +[2026-06-20 16:38:48,172][starVLA.training.train_starvla][INFO] - Step 3911, Loss: {'train/loss': 0.06419782340526581, 'train/grad_norm_pre_clip': 6.108124732971191, 'timing/data': 0.006459722062572837, 'timing/model': 1.3047887859866023, 'learning_rate/qwen_vl_interface': 8.638769636429245e-06, 'learning_rate/action_model': 8.638769636429245e-05, 'epoch': 7.23}) +[2026-06-20 16:38:53,483][starVLA.training.train_starvla][INFO] - Step 3912, Loss: {'train/loss': 0.21705473959445953, 'train/grad_norm_pre_clip': 11.585634231567383, 'timing/data': 0.006412282004021108, 'timing/model': 1.3039318759692833, 'learning_rate/qwen_vl_interface': 8.636634917623946e-06, 'learning_rate/action_model': 8.636634917623944e-05, 'epoch': 7.23}) +[2026-06-20 16:38:58,783][starVLA.training.train_starvla][INFO] - Step 3913, Loss: {'train/loss': 0.19595138728618622, 'train/grad_norm_pre_clip': 10.4304838180542, 'timing/data': 0.006570011028088629, 'timing/model': 1.3038244739873335, 'learning_rate/qwen_vl_interface': 8.634498806700955e-06, 'learning_rate/action_model': 8.634498806700954e-05, 'epoch': 7.24}) +[2026-06-20 16:39:04,075][starVLA.training.train_starvla][INFO] - Step 3914, Loss: {'train/loss': 0.07063501328229904, 'train/grad_norm_pre_clip': 7.974082946777344, 'timing/data': 0.00650051306001842, 'timing/model': 1.5414891350083053, 'learning_rate/qwen_vl_interface': 8.63236130453835e-06, 'learning_rate/action_model': 8.63236130453835e-05, 'epoch': 7.24}) +[2026-06-20 16:39:09,135][starVLA.training.train_starvla][INFO] - Step 3915, Loss: {'train/loss': 0.21466505527496338, 'train/grad_norm_pre_clip': 11.985673904418945, 'timing/data': 0.006312784040346742, 'timing/model': 1.3067116140155122, 'learning_rate/qwen_vl_interface': 8.630222412014777e-06, 'learning_rate/action_model': 8.630222412014776e-05, 'epoch': 7.24}) +[2026-06-20 16:39:14,071][starVLA.training.train_starvla][INFO] - Step 3916, Loss: {'train/loss': 0.040957558900117874, 'train/grad_norm_pre_clip': 3.007096290588379, 'timing/data': 0.00595540797803551, 'timing/model': 1.1734038529684767, 'learning_rate/qwen_vl_interface': 8.628082130009451e-06, 'learning_rate/action_model': 8.628082130009451e-05, 'epoch': 7.24}) +[2026-06-20 16:39:18,960][starVLA.training.train_starvla][INFO] - Step 3917, Loss: {'train/loss': 0.061446573585271835, 'train/grad_norm_pre_clip': 7.968784809112549, 'timing/data': 0.006428593071177602, 'timing/model': 1.3038836129708216, 'learning_rate/qwen_vl_interface': 8.62594045940216e-06, 'learning_rate/action_model': 8.62594045940216e-05, 'epoch': 7.24}) +[2026-06-20 16:39:23,719][starVLA.training.train_starvla][INFO] - Step 3918, Loss: {'train/loss': 0.21803823113441467, 'train/grad_norm_pre_clip': 8.415757179260254, 'timing/data': 0.006171587039716542, 'timing/model': 1.1756903530331329, 'learning_rate/qwen_vl_interface': 8.623797401073272e-06, 'learning_rate/action_model': 8.623797401073271e-05, 'epoch': 7.25}) +[2026-06-20 16:39:28,447][starVLA.training.train_starvla][INFO] - Step 3919, Loss: {'train/loss': 0.08084853738546371, 'train/grad_norm_pre_clip': 6.549256324768066, 'timing/data': 0.006210056017152965, 'timing/model': 1.1759470190154389, 'learning_rate/qwen_vl_interface': 8.62165295590371e-06, 'learning_rate/action_model': 8.621652955903709e-05, 'epoch': 7.25}) +[2026-06-20 16:39:33,198][starVLA.training.train_starvla][INFO] - Step 3920, Loss: {'train/loss': 0.13251912593841553, 'train/grad_norm_pre_clip': 9.850356101989746, 'timing/data': 0.006389115937054157, 'timing/model': 1.3915666020475328, 'learning_rate/qwen_vl_interface': 8.619507124774975e-06, 'learning_rate/action_model': 8.619507124774975e-05, 'epoch': 7.25}) +[2026-06-20 16:39:37,728][starVLA.training.train_starvla][INFO] - Step 3921, Loss: {'train/loss': 0.1318517029285431, 'train/grad_norm_pre_clip': 10.793120384216309, 'timing/data': 0.00572975305840373, 'timing/model': 1.1735972280148417, 'learning_rate/qwen_vl_interface': 8.61735990856914e-06, 'learning_rate/action_model': 8.617359908569139e-05, 'epoch': 7.25}) +[2026-06-20 16:39:42,502][starVLA.training.train_starvla][INFO] - Step 3922, Loss: {'train/loss': 0.07241743057966232, 'train/grad_norm_pre_clip': 5.879271030426025, 'timing/data': 0.006264956900849938, 'timing/model': 1.1763692180393264, 'learning_rate/qwen_vl_interface': 8.615211308168842e-06, 'learning_rate/action_model': 8.61521130816884e-05, 'epoch': 7.25}) +[2026-06-20 16:39:47,615][starVLA.training.train_starvla][INFO] - Step 3923, Loss: {'train/loss': 0.0486626997590065, 'train/grad_norm_pre_clip': 7.200122833251953, 'timing/data': 0.006421176018193364, 'timing/model': 1.3081742740469053, 'learning_rate/qwen_vl_interface': 8.613061324457289e-06, 'learning_rate/action_model': 8.613061324457288e-05, 'epoch': 7.26}) +[2026-06-20 16:39:52,931][starVLA.training.train_starvla][INFO] - Step 3924, Loss: {'train/loss': 0.02991061843931675, 'train/grad_norm_pre_clip': 3.7121617794036865, 'timing/data': 0.006410075933672488, 'timing/model': 1.3075134499231353, 'learning_rate/qwen_vl_interface': 8.610909958318262e-06, 'learning_rate/action_model': 8.610909958318261e-05, 'epoch': 7.26}) +[2026-06-20 16:39:58,195][starVLA.training.train_starvla][INFO] - Step 3925, Loss: {'train/loss': 0.22421294450759888, 'train/grad_norm_pre_clip': 12.045308113098145, 'timing/data': 0.006358037004247308, 'timing/model': 1.3082439630525187, 'learning_rate/qwen_vl_interface': 8.608757210636101e-06, 'learning_rate/action_model': 8.608757210636101e-05, 'epoch': 7.26}) +[2026-06-20 16:40:03,515][starVLA.training.train_starvla][INFO] - Step 3926, Loss: {'train/loss': 0.5681746006011963, 'train/grad_norm_pre_clip': 13.859222412109375, 'timing/data': 0.006661123945377767, 'timing/model': 1.564922641031444, 'learning_rate/qwen_vl_interface': 8.606603082295724e-06, 'learning_rate/action_model': 8.606603082295724e-05, 'epoch': 7.26}) +[2026-06-20 16:40:08,571][starVLA.training.train_starvla][INFO] - Step 3927, Loss: {'train/loss': 0.24202591180801392, 'train/grad_norm_pre_clip': 11.828729629516602, 'timing/data': 0.006473926012404263, 'timing/model': 1.2988396299770102, 'learning_rate/qwen_vl_interface': 8.60444757418261e-06, 'learning_rate/action_model': 8.604447574182609e-05, 'epoch': 7.26}) +[2026-06-20 16:40:13,904][starVLA.training.train_starvla][INFO] - Step 3928, Loss: {'train/loss': 0.010632387362420559, 'train/grad_norm_pre_clip': 0.6690820455551147, 'timing/data': 0.006487878039479256, 'timing/model': 1.3056839889613912, 'learning_rate/qwen_vl_interface': 8.602290687182805e-06, 'learning_rate/action_model': 8.602290687182806e-05, 'epoch': 7.26}) +[2026-06-20 16:40:19,229][starVLA.training.train_starvla][INFO] - Step 3929, Loss: {'train/loss': 0.04938696697354317, 'train/grad_norm_pre_clip': 7.762557506561279, 'timing/data': 0.006294079008512199, 'timing/model': 1.3053912610048428, 'learning_rate/qwen_vl_interface': 8.600132422182929e-06, 'learning_rate/action_model': 8.600132422182928e-05, 'epoch': 7.27}) +[2026-06-20 16:40:24,615][starVLA.training.train_starvla][INFO] - Step 3930, Loss: {'train/loss': 0.1588289886713028, 'train/grad_norm_pre_clip': 10.229914665222168, 'timing/data': 0.0065136769553646445, 'timing/model': 1.306791924056597, 'learning_rate/qwen_vl_interface': 8.597972780070156e-06, 'learning_rate/action_model': 8.597972780070155e-05, 'epoch': 7.27}) +[2026-06-20 16:40:29,351][starVLA.training.train_starvla][INFO] - Step 3931, Loss: {'train/loss': 0.08343039453029633, 'train/grad_norm_pre_clip': 8.556652069091797, 'timing/data': 0.007761385990306735, 'timing/model': 1.1711451510200277, 'learning_rate/qwen_vl_interface': 8.595811761732239e-06, 'learning_rate/action_model': 8.595811761732238e-05, 'epoch': 7.27}) +[2026-06-20 16:40:34,389][starVLA.training.train_starvla][INFO] - Step 3932, Loss: {'train/loss': 0.03630288317799568, 'train/grad_norm_pre_clip': 5.629669189453125, 'timing/data': 0.006443258957006037, 'timing/model': 1.5411098019685596, 'learning_rate/qwen_vl_interface': 8.593649368057489e-06, 'learning_rate/action_model': 8.59364936805749e-05, 'epoch': 7.27}) +[2026-06-20 16:40:39,443][starVLA.training.train_starvla][INFO] - Step 3933, Loss: {'train/loss': 0.1386515498161316, 'train/grad_norm_pre_clip': 5.926061153411865, 'timing/data': 0.0064396189991384745, 'timing/model': 1.3073584811063483, 'learning_rate/qwen_vl_interface': 8.591485599934786e-06, 'learning_rate/action_model': 8.591485599934785e-05, 'epoch': 7.27}) +[2026-06-20 16:40:44,759][starVLA.training.train_starvla][INFO] - Step 3934, Loss: {'train/loss': 0.20470993220806122, 'train/grad_norm_pre_clip': 11.686031341552734, 'timing/data': 0.0064182489877566695, 'timing/model': 1.3066266409587115, 'learning_rate/qwen_vl_interface': 8.58932045825357e-06, 'learning_rate/action_model': 8.589320458253569e-05, 'epoch': 7.28}) +[2026-06-20 16:40:50,064][starVLA.training.train_starvla][INFO] - Step 3935, Loss: {'train/loss': 0.19532087445259094, 'train/grad_norm_pre_clip': 13.083147048950195, 'timing/data': 0.006554009043611586, 'timing/model': 1.3053741150069982, 'learning_rate/qwen_vl_interface': 8.58715394390385e-06, 'learning_rate/action_model': 8.58715394390385e-05, 'epoch': 7.28}) +[2026-06-20 16:40:55,396][starVLA.training.train_starvla][INFO] - Step 3936, Loss: {'train/loss': 0.07269041240215302, 'train/grad_norm_pre_clip': 5.806523323059082, 'timing/data': 0.006397419958375394, 'timing/model': 1.3123731929808855, 'learning_rate/qwen_vl_interface': 8.5849860577762e-06, 'learning_rate/action_model': 8.5849860577762e-05, 'epoch': 7.28}) +[2026-06-20 16:41:00,685][starVLA.training.train_starvla][INFO] - Step 3937, Loss: {'train/loss': 0.2862491309642792, 'train/grad_norm_pre_clip': 20.639286041259766, 'timing/data': 0.006455990020185709, 'timing/model': 1.5410256000468507, 'learning_rate/qwen_vl_interface': 8.582816800761756e-06, 'learning_rate/action_model': 8.582816800761755e-05, 'epoch': 7.28}) +[2026-06-20 16:41:05,882][starVLA.training.train_starvla][INFO] - Step 3938, Loss: {'train/loss': 0.21395090222358704, 'train/grad_norm_pre_clip': 11.72790813446045, 'timing/data': 0.006530500017106533, 'timing/model': 1.4502899390645325, 'learning_rate/qwen_vl_interface': 8.580646173752214e-06, 'learning_rate/action_model': 8.580646173752213e-05, 'epoch': 7.28}) +[2026-06-20 16:41:10,436][starVLA.training.train_starvla][INFO] - Step 3939, Loss: {'train/loss': 0.2658386826515198, 'train/grad_norm_pre_clip': 19.48857307434082, 'timing/data': 0.006330772070214152, 'timing/model': 1.1832432959927246, 'learning_rate/qwen_vl_interface': 8.578474177639841e-06, 'learning_rate/action_model': 8.578474177639842e-05, 'epoch': 7.28}) +[2026-06-20 16:41:15,268][starVLA.training.train_starvla][INFO] - Step 3940, Loss: {'train/loss': 0.17106421291828156, 'train/grad_norm_pre_clip': 5.897196292877197, 'timing/data': 0.006725398008711636, 'timing/model': 1.1804354419000447, 'learning_rate/qwen_vl_interface': 8.576300813317463e-06, 'learning_rate/action_model': 8.576300813317462e-05, 'epoch': 7.29}) +[2026-06-20 16:41:20,089][starVLA.training.train_starvla][INFO] - Step 3941, Loss: {'train/loss': 0.30225086212158203, 'train/grad_norm_pre_clip': 19.080577850341797, 'timing/data': 0.0066241699969395995, 'timing/model': 1.177593725034967, 'learning_rate/qwen_vl_interface': 8.574126081678463e-06, 'learning_rate/action_model': 8.574126081678462e-05, 'epoch': 7.29}) +[2026-06-20 16:41:24,904][starVLA.training.train_starvla][INFO] - Step 3942, Loss: {'train/loss': 0.13942286372184753, 'train/grad_norm_pre_clip': 5.343921661376953, 'timing/data': 0.0063487529987469316, 'timing/model': 1.1745378409978002, 'learning_rate/qwen_vl_interface': 8.571949983616796e-06, 'learning_rate/action_model': 8.571949983616795e-05, 'epoch': 7.29}) +[2026-06-20 16:41:29,751][starVLA.training.train_starvla][INFO] - Step 3943, Loss: {'train/loss': 0.25601866841316223, 'train/grad_norm_pre_clip': 8.887523651123047, 'timing/data': 0.006480591022409499, 'timing/model': 1.4632340490352362, 'learning_rate/qwen_vl_interface': 8.569772520026971e-06, 'learning_rate/action_model': 8.569772520026971e-05, 'epoch': 7.29}) +[2026-06-20 16:41:35,184][starVLA.training.train_starvla][INFO] - Step 3944, Loss: {'train/loss': 0.22139212489128113, 'train/grad_norm_pre_clip': 19.4548282623291, 'timing/data': 0.0062822329346090555, 'timing/model': 1.5241600869921967, 'learning_rate/qwen_vl_interface': 8.567593691804061e-06, 'learning_rate/action_model': 8.567593691804061e-05, 'epoch': 7.29}) +[2026-06-20 16:41:40,300][starVLA.training.train_starvla][INFO] - Step 3945, Loss: {'train/loss': 0.19765247404575348, 'train/grad_norm_pre_clip': 11.079898834228516, 'timing/data': 0.006594590027816594, 'timing/model': 1.3234348790720105, 'learning_rate/qwen_vl_interface': 8.565413499843703e-06, 'learning_rate/action_model': 8.565413499843703e-05, 'epoch': 7.3}) +[2026-06-20 16:41:45,597][starVLA.training.train_starvla][INFO] - Step 3946, Loss: {'train/loss': 0.0767555832862854, 'train/grad_norm_pre_clip': 6.8357110023498535, 'timing/data': 0.006587011041119695, 'timing/model': 1.3199515669839457, 'learning_rate/qwen_vl_interface': 8.563231945042087e-06, 'learning_rate/action_model': 8.563231945042088e-05, 'epoch': 7.3}) +[2026-06-20 16:41:50,911][starVLA.training.train_starvla][INFO] - Step 3947, Loss: {'train/loss': 0.032036494463682175, 'train/grad_norm_pre_clip': 2.867082357406616, 'timing/data': 0.00642716302536428, 'timing/model': 1.3259785390691832, 'learning_rate/qwen_vl_interface': 8.56104902829597e-06, 'learning_rate/action_model': 8.56104902829597e-05, 'epoch': 7.3}) +[2026-06-20 16:41:55,958][starVLA.training.train_starvla][INFO] - Step 3948, Loss: {'train/loss': 0.05246325209736824, 'train/grad_norm_pre_clip': 6.362946510314941, 'timing/data': 0.00643058295827359, 'timing/model': 1.182641597930342, 'learning_rate/qwen_vl_interface': 8.558864750502669e-06, 'learning_rate/action_model': 8.558864750502668e-05, 'epoch': 7.3}) +[2026-06-20 16:42:00,819][starVLA.training.train_starvla][INFO] - Step 3949, Loss: {'train/loss': 0.151148721575737, 'train/grad_norm_pre_clip': 14.792548179626465, 'timing/data': 0.0065885219955816865, 'timing/model': 1.419169276021421, 'learning_rate/qwen_vl_interface': 8.556679112560055e-06, 'learning_rate/action_model': 8.556679112560054e-05, 'epoch': 7.3}) +[2026-06-20 16:42:05,774][starVLA.training.train_starvla][INFO] - Step 3950, Loss: {'train/loss': 0.05577264353632927, 'train/grad_norm_pre_clip': 4.404594421386719, 'timing/data': 0.0063966530142351985, 'timing/model': 1.303935587988235, 'learning_rate/qwen_vl_interface': 8.554492115366562e-06, 'learning_rate/action_model': 8.554492115366561e-05, 'epoch': 7.3}) +[2026-06-20 16:42:11,092][starVLA.training.train_starvla][INFO] - Step 3951, Loss: {'train/loss': 0.18936485052108765, 'train/grad_norm_pre_clip': 12.320775985717773, 'timing/data': 0.006452834000810981, 'timing/model': 1.3085176430176944, 'learning_rate/qwen_vl_interface': 8.55230375982118e-06, 'learning_rate/action_model': 8.55230375982118e-05, 'epoch': 7.31}) +[2026-06-20 16:42:16,471][starVLA.training.train_starvla][INFO] - Step 3952, Loss: {'train/loss': 0.1579822599887848, 'train/grad_norm_pre_clip': 10.267755508422852, 'timing/data': 0.006177856936119497, 'timing/model': 1.2998506920412183, 'learning_rate/qwen_vl_interface': 8.550114046823463e-06, 'learning_rate/action_model': 8.550114046823463e-05, 'epoch': 7.31}) +[2026-06-20 16:42:21,267][starVLA.training.train_starvla][INFO] - Step 3953, Loss: {'train/loss': 0.31125006079673767, 'train/grad_norm_pre_clip': 14.137824058532715, 'timing/data': 0.006519082933664322, 'timing/model': 1.1875205190153793, 'learning_rate/qwen_vl_interface': 8.547922977273518e-06, 'learning_rate/action_model': 8.547922977273518e-05, 'epoch': 7.31}) +[2026-06-20 16:42:26,103][starVLA.training.train_starvla][INFO] - Step 3954, Loss: {'train/loss': 0.13928380608558655, 'train/grad_norm_pre_clip': 11.756728172302246, 'timing/data': 0.006420954945497215, 'timing/model': 1.1997301270021126, 'learning_rate/qwen_vl_interface': 8.54573055207201e-06, 'learning_rate/action_model': 8.545730552072008e-05, 'epoch': 7.31}) +[2026-06-20 16:42:31,119][starVLA.training.train_starvla][INFO] - Step 3955, Loss: {'train/loss': 0.15145418047904968, 'train/grad_norm_pre_clip': 10.15217399597168, 'timing/data': 0.006561825051903725, 'timing/model': 1.4235776549903676, 'learning_rate/qwen_vl_interface': 8.543536772120161e-06, 'learning_rate/action_model': 8.54353677212016e-05, 'epoch': 7.31}) +[2026-06-20 16:42:35,708][starVLA.training.train_starvla][INFO] - Step 3956, Loss: {'train/loss': 0.157161682844162, 'train/grad_norm_pre_clip': 6.682643413543701, 'timing/data': 0.006669854046776891, 'timing/model': 1.1800093600759283, 'learning_rate/qwen_vl_interface': 8.541341638319754e-06, 'learning_rate/action_model': 8.541341638319754e-05, 'epoch': 7.32}) +[2026-06-20 16:42:40,531][starVLA.training.train_starvla][INFO] - Step 3957, Loss: {'train/loss': 0.09184898436069489, 'train/grad_norm_pre_clip': 7.952485084533691, 'timing/data': 0.007633714936673641, 'timing/model': 1.1928998429793864, 'learning_rate/qwen_vl_interface': 8.539145151573126e-06, 'learning_rate/action_model': 8.539145151573125e-05, 'epoch': 7.32}) +[2026-06-20 16:42:45,333][starVLA.training.train_starvla][INFO] - Step 3958, Loss: {'train/loss': 0.04965267330408096, 'train/grad_norm_pre_clip': 4.354351997375488, 'timing/data': 0.006603653891943395, 'timing/model': 1.177773382049054, 'learning_rate/qwen_vl_interface': 8.536947312783167e-06, 'learning_rate/action_model': 8.536947312783166e-05, 'epoch': 7.32}) +[2026-06-20 16:42:50,200][starVLA.training.train_starvla][INFO] - Step 3959, Loss: {'train/loss': 0.1543150544166565, 'train/grad_norm_pre_clip': 6.317890167236328, 'timing/data': 0.0065690838964655995, 'timing/model': 1.1787665160372853, 'learning_rate/qwen_vl_interface': 8.534748122853326e-06, 'learning_rate/action_model': 8.534748122853325e-05, 'epoch': 7.32}) +[2026-06-20 16:42:55,077][starVLA.training.train_starvla][INFO] - Step 3960, Loss: {'train/loss': 0.030499571934342384, 'train/grad_norm_pre_clip': 1.58784818649292, 'timing/data': 0.006555475061759353, 'timing/model': 1.2003377750515938, 'learning_rate/qwen_vl_interface': 8.532547582687609e-06, 'learning_rate/action_model': 8.532547582687608e-05, 'epoch': 7.32}) +[2026-06-20 16:42:59,945][starVLA.training.train_starvla][INFO] - Step 3961, Loss: {'train/loss': 0.23712080717086792, 'train/grad_norm_pre_clip': 9.695960998535156, 'timing/data': 0.008264800999313593, 'timing/model': 1.4355349929537624, 'learning_rate/qwen_vl_interface': 8.530345693190573e-06, 'learning_rate/action_model': 8.530345693190572e-05, 'epoch': 7.33}) +[2026-06-20 16:43:04,726][starVLA.training.train_starvla][INFO] - Step 3962, Loss: {'train/loss': 0.0517902635037899, 'train/grad_norm_pre_clip': 6.843301773071289, 'timing/data': 0.006435037008486688, 'timing/model': 1.361271557980217, 'learning_rate/qwen_vl_interface': 8.528142455267328e-06, 'learning_rate/action_model': 8.528142455267329e-05, 'epoch': 7.33}) +[2026-06-20 16:43:09,304][starVLA.training.train_starvla][INFO] - Step 3963, Loss: {'train/loss': 0.07555132359266281, 'train/grad_norm_pre_clip': 10.619686126708984, 'timing/data': 0.006183387944474816, 'timing/model': 1.1779761200305074, 'learning_rate/qwen_vl_interface': 8.52593786982355e-06, 'learning_rate/action_model': 8.525937869823551e-05, 'epoch': 7.33}) +[2026-06-20 16:43:14,083][starVLA.training.train_starvla][INFO] - Step 3964, Loss: {'train/loss': 0.12038775533437729, 'train/grad_norm_pre_clip': 8.56635856628418, 'timing/data': 0.006543945986777544, 'timing/model': 1.1864922679960728, 'learning_rate/qwen_vl_interface': 8.523731937765456e-06, 'learning_rate/action_model': 8.523731937765456e-05, 'epoch': 7.33}) +[2026-06-20 16:43:18,893][starVLA.training.train_starvla][INFO] - Step 3965, Loss: {'train/loss': 0.4184081256389618, 'train/grad_norm_pre_clip': 16.311767578125, 'timing/data': 0.006435046903789043, 'timing/model': 1.1855225049657747, 'learning_rate/qwen_vl_interface': 8.521524659999823e-06, 'learning_rate/action_model': 8.521524659999822e-05, 'epoch': 7.33}) +[2026-06-20 16:43:23,699][starVLA.training.train_starvla][INFO] - Step 3966, Loss: {'train/loss': 0.09319048374891281, 'train/grad_norm_pre_clip': 8.642664909362793, 'timing/data': 0.0063827779376879334, 'timing/model': 1.187651906046085, 'learning_rate/qwen_vl_interface': 8.519316037433976e-06, 'learning_rate/action_model': 8.519316037433977e-05, 'epoch': 7.33}) +[2026-06-20 16:43:28,500][starVLA.training.train_starvla][INFO] - Step 3967, Loss: {'train/loss': 0.05225905403494835, 'train/grad_norm_pre_clip': 12.307891845703125, 'timing/data': 0.006335869082249701, 'timing/model': 1.432612476986833, 'learning_rate/qwen_vl_interface': 8.517106070975801e-06, 'learning_rate/action_model': 8.5171060709758e-05, 'epoch': 7.34}) +[2026-06-20 16:43:33,301][starVLA.training.train_starvla][INFO] - Step 3968, Loss: {'train/loss': 0.11407116800546646, 'train/grad_norm_pre_clip': 10.15778636932373, 'timing/data': 0.00632934900932014, 'timing/model': 1.4184072479838505, 'learning_rate/qwen_vl_interface': 8.514894761533729e-06, 'learning_rate/action_model': 8.514894761533728e-05, 'epoch': 7.34}) +[2026-06-20 16:43:37,848][starVLA.training.train_starvla][INFO] - Step 3969, Loss: {'train/loss': 0.26550036668777466, 'train/grad_norm_pre_clip': 17.714582443237305, 'timing/data': 0.006347039015963674, 'timing/model': 1.1799737990368158, 'learning_rate/qwen_vl_interface': 8.512682110016747e-06, 'learning_rate/action_model': 8.512682110016747e-05, 'epoch': 7.34}) +[2026-06-20 16:43:42,621][starVLA.training.train_starvla][INFO] - Step 3970, Loss: {'train/loss': 0.04621986672282219, 'train/grad_norm_pre_clip': 3.8788695335388184, 'timing/data': 0.006476747919805348, 'timing/model': 1.1806162600405514, 'learning_rate/qwen_vl_interface': 8.51046811733439e-06, 'learning_rate/action_model': 8.51046811733439e-05, 'epoch': 7.34}) +[2026-06-20 16:43:47,411][starVLA.training.train_starvla][INFO] - Step 3971, Loss: {'train/loss': 0.07747910171747208, 'train/grad_norm_pre_clip': 16.35066032409668, 'timing/data': 0.006394219002686441, 'timing/model': 1.1777858190471306, 'learning_rate/qwen_vl_interface': 8.50825278439675e-06, 'learning_rate/action_model': 8.508252784396749e-05, 'epoch': 7.34}) +[2026-06-20 16:43:52,207][starVLA.training.train_starvla][INFO] - Step 3972, Loss: {'train/loss': 0.21654976904392242, 'train/grad_norm_pre_clip': 7.243947982788086, 'timing/data': 0.00619253097102046, 'timing/model': 1.1788821660447866, 'learning_rate/qwen_vl_interface': 8.506036112114465e-06, 'learning_rate/action_model': 8.506036112114464e-05, 'epoch': 7.35}) +[2026-06-20 16:43:56,992][starVLA.training.train_starvla][INFO] - Step 3973, Loss: {'train/loss': 0.006297662388533354, 'train/grad_norm_pre_clip': 0.5185820460319519, 'timing/data': 0.006418138975277543, 'timing/model': 1.4189960081130266, 'learning_rate/qwen_vl_interface': 8.503818101398725e-06, 'learning_rate/action_model': 8.503818101398725e-05, 'epoch': 7.35}) +[2026-06-20 16:44:01,540][starVLA.training.train_starvla][INFO] - Step 3974, Loss: {'train/loss': 0.20164678990840912, 'train/grad_norm_pre_clip': 12.075296401977539, 'timing/data': 0.006471499102190137, 'timing/model': 1.1797325679799542, 'learning_rate/qwen_vl_interface': 8.501598753161273e-06, 'learning_rate/action_model': 8.501598753161272e-05, 'epoch': 7.35}) +[2026-06-20 16:44:06,332][starVLA.training.train_starvla][INFO] - Step 3975, Loss: {'train/loss': 0.3671044409275055, 'train/grad_norm_pre_clip': 6.731740951538086, 'timing/data': 0.0064627398969605565, 'timing/model': 1.1757977950619534, 'learning_rate/qwen_vl_interface': 8.499378068314395e-06, 'learning_rate/action_model': 8.499378068314395e-05, 'epoch': 7.35}) +[2026-06-20 16:44:11,118][starVLA.training.train_starvla][INFO] - Step 3976, Loss: {'train/loss': 0.043885961174964905, 'train/grad_norm_pre_clip': 8.190953254699707, 'timing/data': 0.006577517953701317, 'timing/model': 1.177149558905512, 'learning_rate/qwen_vl_interface': 8.497156047770932e-06, 'learning_rate/action_model': 8.497156047770933e-05, 'epoch': 7.35}) +[2026-06-20 16:44:15,892][starVLA.training.train_starvla][INFO] - Step 3977, Loss: {'train/loss': 0.15780892968177795, 'train/grad_norm_pre_clip': 15.172314643859863, 'timing/data': 0.008872120059095323, 'timing/model': 1.1765696379588917, 'learning_rate/qwen_vl_interface': 8.494932692444276e-06, 'learning_rate/action_model': 8.494932692444276e-05, 'epoch': 7.35}) +[2026-06-20 16:44:20,623][starVLA.training.train_starvla][INFO] - Step 3978, Loss: {'train/loss': 0.06277075409889221, 'train/grad_norm_pre_clip': 7.579594135284424, 'timing/data': 0.006453639012761414, 'timing/model': 1.1898654169635847, 'learning_rate/qwen_vl_interface': 8.492708003248362e-06, 'learning_rate/action_model': 8.492708003248362e-05, 'epoch': 7.36}) +[2026-06-20 16:44:25,367][starVLA.training.train_starvla][INFO] - Step 3979, Loss: {'train/loss': 0.059403352439403534, 'train/grad_norm_pre_clip': 6.398072242736816, 'timing/data': 0.007105294964276254, 'timing/model': 1.3893285570666194, 'learning_rate/qwen_vl_interface': 8.490481981097674e-06, 'learning_rate/action_model': 8.490481981097674e-05, 'epoch': 7.36}) +[2026-06-20 16:44:29,912][starVLA.training.train_starvla][INFO] - Step 3980, Loss: {'train/loss': 0.0814116895198822, 'train/grad_norm_pre_clip': 11.251434326171875, 'timing/data': 0.007155484054237604, 'timing/model': 1.181717345956713, 'learning_rate/qwen_vl_interface': 8.488254626907249e-06, 'learning_rate/action_model': 8.488254626907249e-05, 'epoch': 7.36}) +[2026-06-20 16:44:34,695][starVLA.training.train_starvla][INFO] - Step 3981, Loss: {'train/loss': 0.04895593598484993, 'train/grad_norm_pre_clip': 6.271825790405273, 'timing/data': 0.007070216001011431, 'timing/model': 1.1782633860129863, 'learning_rate/qwen_vl_interface': 8.486025941592664e-06, 'learning_rate/action_model': 8.486025941592664e-05, 'epoch': 7.36}) +[2026-06-20 16:44:39,475][starVLA.training.train_starvla][INFO] - Step 3982, Loss: {'train/loss': 0.05087171494960785, 'train/grad_norm_pre_clip': 14.82819938659668, 'timing/data': 0.006412070943042636, 'timing/model': 1.1794888679869473, 'learning_rate/qwen_vl_interface': 8.483795926070054e-06, 'learning_rate/action_model': 8.483795926070052e-05, 'epoch': 7.36}) +[2026-06-20 16:44:44,268][starVLA.training.train_starvla][INFO] - Step 3983, Loss: {'train/loss': 0.12886448204517365, 'train/grad_norm_pre_clip': 4.260970115661621, 'timing/data': 0.006708458065986633, 'timing/model': 1.1822934600058943, 'learning_rate/qwen_vl_interface': 8.481564581256087e-06, 'learning_rate/action_model': 8.481564581256087e-05, 'epoch': 7.37}) +[2026-06-20 16:44:49,051][starVLA.training.train_starvla][INFO] - Step 3984, Loss: {'train/loss': 0.20748074352741241, 'train/grad_norm_pre_clip': 16.90386199951172, 'timing/data': 0.00652253907173872, 'timing/model': 1.179276154958643, 'learning_rate/qwen_vl_interface': 8.479331908067987e-06, 'learning_rate/action_model': 8.479331908067987e-05, 'epoch': 7.37}) +[2026-06-20 16:44:53,821][starVLA.training.train_starvla][INFO] - Step 3985, Loss: {'train/loss': 0.05987470597028732, 'train/grad_norm_pre_clip': 7.9929423332214355, 'timing/data': 0.0065397199941799045, 'timing/model': 1.4033487669657916, 'learning_rate/qwen_vl_interface': 8.477097907423524e-06, 'learning_rate/action_model': 8.477097907423524e-05, 'epoch': 7.37}) +[2026-06-20 16:44:58,377][starVLA.training.train_starvla][INFO] - Step 3986, Loss: {'train/loss': 0.15964418649673462, 'train/grad_norm_pre_clip': 12.127264022827148, 'timing/data': 0.006647908943705261, 'timing/model': 1.1867907610721886, 'learning_rate/qwen_vl_interface': 8.474862580241006e-06, 'learning_rate/action_model': 8.474862580241006e-05, 'epoch': 7.37}) +[2026-06-20 16:45:03,204][starVLA.training.train_starvla][INFO] - Step 3987, Loss: {'train/loss': 0.15656530857086182, 'train/grad_norm_pre_clip': 5.116086006164551, 'timing/data': 0.0065868510864675045, 'timing/model': 1.1873133180197328, 'learning_rate/qwen_vl_interface': 8.472625927439299e-06, 'learning_rate/action_model': 8.472625927439299e-05, 'epoch': 7.37}) +[2026-06-20 16:45:08,031][starVLA.training.train_starvla][INFO] - Step 3988, Loss: {'train/loss': 0.023546621203422546, 'train/grad_norm_pre_clip': 1.9561867713928223, 'timing/data': 0.006702080019749701, 'timing/model': 1.1820070890244097, 'learning_rate/qwen_vl_interface': 8.4703879499378e-06, 'learning_rate/action_model': 8.4703879499378e-05, 'epoch': 7.38}) +[2026-06-20 16:45:12,841][starVLA.training.train_starvla][INFO] - Step 3989, Loss: {'train/loss': 0.0826394110918045, 'train/grad_norm_pre_clip': 12.846480369567871, 'timing/data': 0.007384223979897797, 'timing/model': 1.1965086810523644, 'learning_rate/qwen_vl_interface': 8.46814864865646e-06, 'learning_rate/action_model': 8.46814864865646e-05, 'epoch': 7.38}) +[2026-06-20 16:45:17,668][starVLA.training.train_starvla][INFO] - Step 3990, Loss: {'train/loss': 0.08135330677032471, 'train/grad_norm_pre_clip': 7.338834762573242, 'timing/data': 0.006340513005852699, 'timing/model': 1.4189298839773983, 'learning_rate/qwen_vl_interface': 8.465908024515771e-06, 'learning_rate/action_model': 8.46590802451577e-05, 'epoch': 7.38}) +[2026-06-20 16:45:22,232][starVLA.training.train_starvla][INFO] - Step 3991, Loss: {'train/loss': 0.1490255445241928, 'train/grad_norm_pre_clip': 7.625192642211914, 'timing/data': 0.006598990992642939, 'timing/model': 1.1907377740135416, 'learning_rate/qwen_vl_interface': 8.463666078436769e-06, 'learning_rate/action_model': 8.463666078436769e-05, 'epoch': 7.38}) +[2026-06-20 16:45:27,046][starVLA.training.train_starvla][INFO] - Step 3992, Loss: {'train/loss': 0.2797013223171234, 'train/grad_norm_pre_clip': 11.79251480102539, 'timing/data': 0.006599221029318869, 'timing/model': 1.1895360930357128, 'learning_rate/qwen_vl_interface': 8.46142281134103e-06, 'learning_rate/action_model': 8.461422811341031e-05, 'epoch': 7.38}) +[2026-06-20 16:45:31,820][starVLA.training.train_starvla][INFO] - Step 3993, Loss: {'train/loss': 0.37748628854751587, 'train/grad_norm_pre_clip': 14.071632385253906, 'timing/data': 0.006466340972110629, 'timing/model': 1.1787029230035841, 'learning_rate/qwen_vl_interface': 8.459178224150683e-06, 'learning_rate/action_model': 8.459178224150683e-05, 'epoch': 7.38}) +[2026-06-20 16:45:36,544][starVLA.training.train_starvla][INFO] - Step 3994, Loss: {'train/loss': 0.06216420233249664, 'train/grad_norm_pre_clip': 7.218990325927734, 'timing/data': 0.006509172031655908, 'timing/model': 1.178712422028184, 'learning_rate/qwen_vl_interface': 8.456932317788389e-06, 'learning_rate/action_model': 8.456932317788389e-05, 'epoch': 7.39}) +[2026-06-20 16:45:41,291][starVLA.training.train_starvla][INFO] - Step 3995, Loss: {'train/loss': 0.09640581160783768, 'train/grad_norm_pre_clip': 10.129117965698242, 'timing/data': 0.006385412998497486, 'timing/model': 1.1781118559883907, 'learning_rate/qwen_vl_interface': 8.454685093177356e-06, 'learning_rate/action_model': 8.454685093177354e-05, 'epoch': 7.39}) +[2026-06-20 16:45:46,073][starVLA.training.train_starvla][INFO] - Step 3996, Loss: {'train/loss': 0.1722428798675537, 'train/grad_norm_pre_clip': 7.564482688903809, 'timing/data': 0.006489921943284571, 'timing/model': 1.4167539889458567, 'learning_rate/qwen_vl_interface': 8.452436551241328e-06, 'learning_rate/action_model': 8.452436551241328e-05, 'epoch': 7.39}) +[2026-06-20 16:45:50,616][starVLA.training.train_starvla][INFO] - Step 3997, Loss: {'train/loss': 0.033807381987571716, 'train/grad_norm_pre_clip': 4.569620132446289, 'timing/data': 0.006377254030667245, 'timing/model': 1.1794758010655642, 'learning_rate/qwen_vl_interface': 8.450186692904605e-06, 'learning_rate/action_model': 8.450186692904606e-05, 'epoch': 7.39}) +[2026-06-20 16:45:55,417][starVLA.training.train_starvla][INFO] - Step 3998, Loss: {'train/loss': 0.07275640964508057, 'train/grad_norm_pre_clip': 7.358210563659668, 'timing/data': 0.006187084945850074, 'timing/model': 1.1802321539726108, 'learning_rate/qwen_vl_interface': 8.447935519092014e-06, 'learning_rate/action_model': 8.447935519092014e-05, 'epoch': 7.39}) +[2026-06-20 16:46:00,194][starVLA.training.train_starvla][INFO] - Step 3999, Loss: {'train/loss': 0.09465324878692627, 'train/grad_norm_pre_clip': 11.095671653747559, 'timing/data': 0.006669210968539119, 'timing/model': 1.1764515599934384, 'learning_rate/qwen_vl_interface': 8.445683030728925e-06, 'learning_rate/action_model': 8.445683030728925e-05, 'epoch': 7.4}) +[2026-06-20 16:46:04,988][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 4000 +[2026-06-20 16:52:34,613][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 4000 +[2026-06-20 16:52:34,614][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 4000 +[2026-06-20 16:59:33,523][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 4000 +[2026-06-20 16:59:33,525][starVLA.training.train_starvla][INFO] - Step 4000, Loss: {'train/loss': 0.05931940674781799, 'train/grad_norm_pre_clip': 7.259359836578369, 'eval/loss': 0.6993016004562378, 'eval/latency_0/loss': 0.8522630333900452, 'eval/latency_1/loss': 1.066237211227417, 'eval/latency_2/loss': 1.0002543926239014, 'eval/latency_3/loss': 0.46796369552612305, 'eval/latency_4/loss': 0.07829348742961884, 'eval/flappy/loss': 0.6993016004562378, 'eval/flappy/latency_0/loss': 0.8522630333900452, 'eval/flappy/latency_1/loss': 1.066237211227417, 'eval/flappy/latency_2/loss': 1.0002543926239014, 'eval/flappy/latency_3/loss': 0.46796369552612305, 'eval/flappy/latency_4/loss': 0.07829348742961884, 'eval/flap_precision': 0.11619718309859155, 'eval/flap_recall': 0.12, 'eval/flap_f1': 0.1180679785330948, 'eval/flap_rate_model': 0.07888888888888888, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.1180679785330948, 'eval/latency_0/flap_precision': 0.11619718309859155, 'eval/latency_0/flap_recall': 0.12, 'eval/latency_0/flap_f1': 0.1180679785330948, 'eval/latency_0/flap_rate_model': 0.07888888888888888, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.1180679785330948, 'timing/data': 0.006648872047662735, 'timing/model': 1.1787344299955294, 'learning_rate/qwen_vl_interface': 8.443429228741256e-06, 'learning_rate/action_model': 8.443429228741255e-05, 'epoch': 7.4}) +[2026-06-20 16:59:33,525][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_4000_state +[2026-06-20 16:59:33,529][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-20 16:59:41,459][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_4000_state/model.safetensors +[2026-06-20 17:00:01,269][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_4000_state/optimizer.bin +[2026-06-20 17:00:01,269][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_4000_state/sampler.bin +[2026-06-20 17:00:01,270][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_4000_state/sampler_1.bin +[2026-06-20 17:00:01,270][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_4000_state/random_states_0.pkl +[2026-06-20 17:00:03,843][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4000 +[2026-06-20 17:00:03,844][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-20 17:00:03,857][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-20 17:00:08,580][starVLA.training.train_starvla][INFO] - Step 4001, Loss: {'train/loss': 0.03306141495704651, 'train/grad_norm_pre_clip': 3.5665457248687744, 'timing/data': 0.006424166960641742, 'timing/model': 1.1602351069450378, 'learning_rate/qwen_vl_interface': 8.441174114055457e-06, 'learning_rate/action_model': 8.441174114055456e-05, 'epoch': 7.4}) +[2026-06-20 17:00:13,345][starVLA.training.train_starvla][INFO] - Step 4002, Loss: {'train/loss': 0.14883017539978027, 'train/grad_norm_pre_clip': 8.45953369140625, 'timing/data': 0.007700139074586332, 'timing/model': 1.1908317219931632, 'learning_rate/qwen_vl_interface': 8.43891768759852e-06, 'learning_rate/action_model': 8.43891768759852e-05, 'epoch': 7.4}) +[2026-06-20 17:00:18,116][starVLA.training.train_starvla][INFO] - Step 4003, Loss: {'train/loss': 0.10281649231910706, 'train/grad_norm_pre_clip': 12.59666919708252, 'timing/data': 0.006458015996031463, 'timing/model': 1.419645912013948, 'learning_rate/qwen_vl_interface': 8.436659950297978e-06, 'learning_rate/action_model': 8.436659950297977e-05, 'epoch': 7.4}) +[2026-06-20 17:00:22,741][starVLA.training.train_starvla][INFO] - Step 4004, Loss: {'train/loss': 0.44902393221855164, 'train/grad_norm_pre_clip': 18.812158584594727, 'timing/data': 0.00650356593541801, 'timing/model': 1.1836952900048345, 'learning_rate/qwen_vl_interface': 8.4344009030819e-06, 'learning_rate/action_model': 8.4344009030819e-05, 'epoch': 7.4}) +[2026-06-20 17:00:27,539][starVLA.training.train_starvla][INFO] - Step 4005, Loss: {'train/loss': 0.14061211049556732, 'train/grad_norm_pre_clip': 8.20908260345459, 'timing/data': 0.008144887979142368, 'timing/model': 1.2012143039610237, 'learning_rate/qwen_vl_interface': 8.432140546878899e-06, 'learning_rate/action_model': 8.432140546878898e-05, 'epoch': 7.41}) +[2026-06-20 17:00:32,291][starVLA.training.train_starvla][INFO] - Step 4006, Loss: {'train/loss': 0.042668815702199936, 'train/grad_norm_pre_clip': 6.20155143737793, 'timing/data': 0.0072575119556859136, 'timing/model': 1.1832451929803938, 'learning_rate/qwen_vl_interface': 8.429878882618116e-06, 'learning_rate/action_model': 8.429878882618116e-05, 'epoch': 7.41}) +[2026-06-20 17:00:37,024][starVLA.training.train_starvla][INFO] - Step 4007, Loss: {'train/loss': 0.04062124341726303, 'train/grad_norm_pre_clip': 4.658181190490723, 'timing/data': 0.006589996046386659, 'timing/model': 1.1765364730963483, 'learning_rate/qwen_vl_interface': 8.42761591122924e-06, 'learning_rate/action_model': 8.427615911229241e-05, 'epoch': 7.41}) +[2026-06-20 17:00:41,791][starVLA.training.train_starvla][INFO] - Step 4008, Loss: {'train/loss': 0.1623411327600479, 'train/grad_norm_pre_clip': 10.597492218017578, 'timing/data': 0.006395366042852402, 'timing/model': 1.171960951993242, 'learning_rate/qwen_vl_interface': 8.425351633642491e-06, 'learning_rate/action_model': 8.425351633642492e-05, 'epoch': 7.41}) +[2026-06-20 17:00:46,565][starVLA.training.train_starvla][INFO] - Step 4009, Loss: {'train/loss': 0.06938400864601135, 'train/grad_norm_pre_clip': 7.548334121704102, 'timing/data': 0.006387615925632417, 'timing/model': 1.4219713270431384, 'learning_rate/qwen_vl_interface': 8.423086050788632e-06, 'learning_rate/action_model': 8.423086050788631e-05, 'epoch': 7.41}) +[2026-06-20 17:00:51,075][starVLA.training.train_starvla][INFO] - Step 4010, Loss: {'train/loss': 0.24645909667015076, 'train/grad_norm_pre_clip': 13.221894264221191, 'timing/data': 0.006288807024247944, 'timing/model': 1.1698873119894415, 'learning_rate/qwen_vl_interface': 8.420819163598953e-06, 'learning_rate/action_model': 8.420819163598952e-05, 'epoch': 7.42}) +[2026-06-20 17:00:55,858][starVLA.training.train_starvla][INFO] - Step 4011, Loss: {'train/loss': 0.31468135118484497, 'train/grad_norm_pre_clip': 11.899955749511719, 'timing/data': 0.006852604099549353, 'timing/model': 1.1822392019676045, 'learning_rate/qwen_vl_interface': 8.418550973005289e-06, 'learning_rate/action_model': 8.418550973005289e-05, 'epoch': 7.42}) +[2026-06-20 17:01:00,628][starVLA.training.train_starvla][INFO] - Step 4012, Loss: {'train/loss': 0.12343760579824448, 'train/grad_norm_pre_clip': 9.767794609069824, 'timing/data': 0.0064102859469130635, 'timing/model': 1.1848652239423245, 'learning_rate/qwen_vl_interface': 8.416281479940005e-06, 'learning_rate/action_model': 8.416281479940005e-05, 'epoch': 7.42}) +[2026-06-20 17:01:05,476][starVLA.training.train_starvla][INFO] - Step 4013, Loss: {'train/loss': 0.04759477823972702, 'train/grad_norm_pre_clip': 8.917710304260254, 'timing/data': 0.006288758013397455, 'timing/model': 1.1916871829889715, 'learning_rate/qwen_vl_interface': 8.414010685336008e-06, 'learning_rate/action_model': 8.414010685336008e-05, 'epoch': 7.42}) +[2026-06-20 17:01:10,268][starVLA.training.train_starvla][INFO] - Step 4014, Loss: {'train/loss': 0.14472301304340363, 'train/grad_norm_pre_clip': 18.584938049316406, 'timing/data': 0.006453337031416595, 'timing/model': 1.1802667370066047, 'learning_rate/qwen_vl_interface': 8.411738590126733e-06, 'learning_rate/action_model': 8.411738590126732e-05, 'epoch': 7.42}) +[2026-06-20 17:01:15,067][starVLA.training.train_starvla][INFO] - Step 4015, Loss: {'train/loss': 0.08199229836463928, 'train/grad_norm_pre_clip': 6.755870342254639, 'timing/data': 0.006836024927906692, 'timing/model': 1.4335128009552136, 'learning_rate/qwen_vl_interface': 8.409465195246152e-06, 'learning_rate/action_model': 8.409465195246152e-05, 'epoch': 7.43}) +[2026-06-20 17:01:19,591][starVLA.training.train_starvla][INFO] - Step 4016, Loss: {'train/loss': 0.10518280416727066, 'train/grad_norm_pre_clip': 4.484273433685303, 'timing/data': 0.006060368032194674, 'timing/model': 1.174247278016992, 'learning_rate/qwen_vl_interface': 8.407190501628772e-06, 'learning_rate/action_model': 8.407190501628772e-05, 'epoch': 7.43}) +[2026-06-20 17:01:24,373][starVLA.training.train_starvla][INFO] - Step 4017, Loss: {'train/loss': 0.009156303480267525, 'train/grad_norm_pre_clip': 0.9300744533538818, 'timing/data': 0.006461116019636393, 'timing/model': 1.1861045010155067, 'learning_rate/qwen_vl_interface': 8.404914510209635e-06, 'learning_rate/action_model': 8.404914510209635e-05, 'epoch': 7.43}) +[2026-06-20 17:01:29,260][starVLA.training.train_starvla][INFO] - Step 4018, Loss: {'train/loss': 0.22246918082237244, 'train/grad_norm_pre_clip': 20.096424102783203, 'timing/data': 0.007933509070426226, 'timing/model': 1.2011728669749573, 'learning_rate/qwen_vl_interface': 8.402637221924315e-06, 'learning_rate/action_model': 8.402637221924314e-05, 'epoch': 7.43}) +[2026-06-20 17:01:34,056][starVLA.training.train_starvla][INFO] - Step 4019, Loss: {'train/loss': 0.22693389654159546, 'train/grad_norm_pre_clip': 5.868028163909912, 'timing/data': 0.006625586072914302, 'timing/model': 1.1808366959448904, 'learning_rate/qwen_vl_interface': 8.400358637708919e-06, 'learning_rate/action_model': 8.400358637708918e-05, 'epoch': 7.43}) +[2026-06-20 17:01:39,029][starVLA.training.train_starvla][INFO] - Step 4020, Loss: {'train/loss': 0.04131398722529411, 'train/grad_norm_pre_clip': 6.869772911071777, 'timing/data': 0.0065506360260769725, 'timing/model': 1.242788402014412, 'learning_rate/qwen_vl_interface': 8.398078758500086e-06, 'learning_rate/action_model': 8.398078758500086e-05, 'epoch': 7.43}) +[2026-06-20 17:01:43,922][starVLA.training.train_starvla][INFO] - Step 4021, Loss: {'train/loss': 0.03712368384003639, 'train/grad_norm_pre_clip': 3.6672887802124023, 'timing/data': 0.006452806992456317, 'timing/model': 1.4061896349303424, 'learning_rate/qwen_vl_interface': 8.395797585234987e-06, 'learning_rate/action_model': 8.395797585234988e-05, 'epoch': 7.44}) +[2026-06-20 17:01:48,674][starVLA.training.train_starvla][INFO] - Step 4022, Loss: {'train/loss': 0.18918141722679138, 'train/grad_norm_pre_clip': 9.11828327178955, 'timing/data': 0.006654454977251589, 'timing/model': 1.3594677810324356, 'learning_rate/qwen_vl_interface': 8.39351511885133e-06, 'learning_rate/action_model': 8.393515118851329e-05, 'epoch': 7.44}) +[2026-06-20 17:01:53,207][starVLA.training.train_starvla][INFO] - Step 4023, Loss: {'train/loss': 0.14084874093532562, 'train/grad_norm_pre_clip': 6.948893070220947, 'timing/data': 0.0068235950311645865, 'timing/model': 1.175548821920529, 'learning_rate/qwen_vl_interface': 8.391231360287348e-06, 'learning_rate/action_model': 8.391231360287347e-05, 'epoch': 7.44}) +[2026-06-20 17:01:57,962][starVLA.training.train_starvla][INFO] - Step 4024, Loss: {'train/loss': 0.17574524879455566, 'train/grad_norm_pre_clip': 14.036256790161133, 'timing/data': 0.006803925964049995, 'timing/model': 1.1936687430134043, 'learning_rate/qwen_vl_interface': 8.388946310481807e-06, 'learning_rate/action_model': 8.388946310481807e-05, 'epoch': 7.44}) +[2026-06-20 17:02:02,749][starVLA.training.train_starvla][INFO] - Step 4025, Loss: {'train/loss': 0.19057337939739227, 'train/grad_norm_pre_clip': 12.843626022338867, 'timing/data': 0.0067625450901687145, 'timing/model': 1.1890304409898818, 'learning_rate/qwen_vl_interface': 8.386659970374007e-06, 'learning_rate/action_model': 8.386659970374006e-05, 'epoch': 7.44}) +[2026-06-20 17:02:07,529][starVLA.training.train_starvla][INFO] - Step 4026, Loss: {'train/loss': 0.04964825138449669, 'train/grad_norm_pre_clip': 4.430471897125244, 'timing/data': 0.005970739992335439, 'timing/model': 1.18995555001311, 'learning_rate/qwen_vl_interface': 8.384372340903775e-06, 'learning_rate/action_model': 8.384372340903775e-05, 'epoch': 7.45}) +[2026-06-20 17:02:12,362][starVLA.training.train_starvla][INFO] - Step 4027, Loss: {'train/loss': 0.27762508392333984, 'train/grad_norm_pre_clip': 9.74268913269043, 'timing/data': 0.006744785001501441, 'timing/model': 1.4661020680796355, 'learning_rate/qwen_vl_interface': 8.382083423011472e-06, 'learning_rate/action_model': 8.382083423011471e-05, 'epoch': 7.45}) +[2026-06-20 17:02:17,140][starVLA.training.train_starvla][INFO] - Step 4028, Loss: {'train/loss': 0.1698145568370819, 'train/grad_norm_pre_clip': 6.419591903686523, 'timing/data': 0.00632630696054548, 'timing/model': 1.4259268189780414, 'learning_rate/qwen_vl_interface': 8.379793217637979e-06, 'learning_rate/action_model': 8.379793217637979e-05, 'epoch': 7.45}) +[2026-06-20 17:02:21,679][starVLA.training.train_starvla][INFO] - Step 4029, Loss: {'train/loss': 0.12311761826276779, 'train/grad_norm_pre_clip': 3.4587252140045166, 'timing/data': 0.0077472199918702245, 'timing/model': 1.179782125982456, 'learning_rate/qwen_vl_interface': 8.377501725724719e-06, 'learning_rate/action_model': 8.377501725724719e-05, 'epoch': 7.45}) +[2026-06-20 17:02:26,444][starVLA.training.train_starvla][INFO] - Step 4030, Loss: {'train/loss': 0.08433762937784195, 'train/grad_norm_pre_clip': 9.208511352539062, 'timing/data': 0.006857314030639827, 'timing/model': 1.175716730998829, 'learning_rate/qwen_vl_interface': 8.375208948213636e-06, 'learning_rate/action_model': 8.375208948213635e-05, 'epoch': 7.45}) +[2026-06-20 17:02:31,269][starVLA.training.train_starvla][INFO] - Step 4031, Loss: {'train/loss': 0.12045083194971085, 'train/grad_norm_pre_clip': 9.19404125213623, 'timing/data': 0.006532836938276887, 'timing/model': 1.1717010549036786, 'learning_rate/qwen_vl_interface': 8.372914886047203e-06, 'learning_rate/action_model': 8.372914886047203e-05, 'epoch': 7.45}) +[2026-06-20 17:02:36,104][starVLA.training.train_starvla][INFO] - Step 4032, Loss: {'train/loss': 0.01300343219190836, 'train/grad_norm_pre_clip': 1.5518590211868286, 'timing/data': 0.020780132967047393, 'timing/model': 1.191622746991925, 'learning_rate/qwen_vl_interface': 8.370619540168422e-06, 'learning_rate/action_model': 8.370619540168423e-05, 'epoch': 7.46}) +[2026-06-20 17:02:40,894][starVLA.training.train_starvla][INFO] - Step 4033, Loss: {'train/loss': 0.11640358716249466, 'train/grad_norm_pre_clip': 9.527348518371582, 'timing/data': 0.012340966030023992, 'timing/model': 1.4196161759318784, 'learning_rate/qwen_vl_interface': 8.368322911520826e-06, 'learning_rate/action_model': 8.368322911520826e-05, 'epoch': 7.46}) +[2026-06-20 17:02:45,432][starVLA.training.train_starvla][INFO] - Step 4034, Loss: {'train/loss': 0.021899253129959106, 'train/grad_norm_pre_clip': 1.1910996437072754, 'timing/data': 0.01820127700921148, 'timing/model': 1.1860656830249354, 'learning_rate/qwen_vl_interface': 8.36602500104847e-06, 'learning_rate/action_model': 8.36602500104847e-05, 'epoch': 7.46}) +[2026-06-20 17:02:50,269][starVLA.training.train_starvla][INFO] - Step 4035, Loss: {'train/loss': 0.06399144977331161, 'train/grad_norm_pre_clip': 6.5648698806762695, 'timing/data': 0.012003829004243016, 'timing/model': 1.1767801940441132, 'learning_rate/qwen_vl_interface': 8.36372580969594e-06, 'learning_rate/action_model': 8.36372580969594e-05, 'epoch': 7.46}) +[2026-06-20 17:02:55,043][starVLA.training.train_starvla][INFO] - Step 4036, Loss: {'train/loss': 0.1508886069059372, 'train/grad_norm_pre_clip': 8.73337173461914, 'timing/data': 0.011969998944550753, 'timing/model': 1.1878913909895346, 'learning_rate/qwen_vl_interface': 8.361425338408344e-06, 'learning_rate/action_model': 8.361425338408343e-05, 'epoch': 7.46}) +[2026-06-20 17:02:59,816][starVLA.training.train_starvla][INFO] - Step 4037, Loss: {'train/loss': 0.21690386533737183, 'train/grad_norm_pre_clip': 10.333794593811035, 'timing/data': 0.011658760020509362, 'timing/model': 1.1798278060741723, 'learning_rate/qwen_vl_interface': 8.359123588131319e-06, 'learning_rate/action_model': 8.359123588131319e-05, 'epoch': 7.47}) +[2026-06-20 17:03:04,531][starVLA.training.train_starvla][INFO] - Step 4038, Loss: {'train/loss': 0.13828474283218384, 'train/grad_norm_pre_clip': 10.480151176452637, 'timing/data': 0.011872028931975365, 'timing/model': 1.1882305570179597, 'learning_rate/qwen_vl_interface': 8.35682055981103e-06, 'learning_rate/action_model': 8.35682055981103e-05, 'epoch': 7.47}) +[2026-06-20 17:03:09,323][starVLA.training.train_starvla][INFO] - Step 4039, Loss: {'train/loss': 0.14946718513965607, 'train/grad_norm_pre_clip': 11.593086242675781, 'timing/data': 0.011759640066884458, 'timing/model': 1.4277810009662062, 'learning_rate/qwen_vl_interface': 8.354516254394163e-06, 'learning_rate/action_model': 8.354516254394162e-05, 'epoch': 7.47}) +[2026-06-20 17:03:13,857][starVLA.training.train_starvla][INFO] - Step 4040, Loss: {'train/loss': 0.0717463418841362, 'train/grad_norm_pre_clip': 6.7319655418396, 'timing/data': 0.01175150997005403, 'timing/model': 1.197956525022164, 'learning_rate/qwen_vl_interface': 8.352210672827932e-06, 'learning_rate/action_model': 8.352210672827931e-05, 'epoch': 7.47}) +[2026-06-20 17:03:18,652][starVLA.training.train_starvla][INFO] - Step 4041, Loss: {'train/loss': 0.1800709217786789, 'train/grad_norm_pre_clip': 10.249783515930176, 'timing/data': 0.012127508060075343, 'timing/model': 1.1990126229356974, 'learning_rate/qwen_vl_interface': 8.349903816060074e-06, 'learning_rate/action_model': 8.349903816060075e-05, 'epoch': 7.47}) +[2026-06-20 17:03:23,466][starVLA.training.train_starvla][INFO] - Step 4042, Loss: {'train/loss': 0.07131655514240265, 'train/grad_norm_pre_clip': 8.893157958984375, 'timing/data': 0.017170242965221405, 'timing/model': 1.2239171399269253, 'learning_rate/qwen_vl_interface': 8.34759568503885e-06, 'learning_rate/action_model': 8.347595685038851e-05, 'epoch': 7.48}) +[2026-06-20 17:03:28,268][starVLA.training.train_starvla][INFO] - Step 4043, Loss: {'train/loss': 0.23489488661289215, 'train/grad_norm_pre_clip': 14.660059928894043, 'timing/data': 0.018196498043835163, 'timing/model': 1.2049432899802923, 'learning_rate/qwen_vl_interface': 8.345286280713049e-06, 'learning_rate/action_model': 8.345286280713048e-05, 'epoch': 7.48}) +[2026-06-20 17:03:33,036][starVLA.training.train_starvla][INFO] - Step 4044, Loss: {'train/loss': 0.15511398017406464, 'train/grad_norm_pre_clip': 10.47150707244873, 'timing/data': 0.011572811054065824, 'timing/model': 1.1927955050487071, 'learning_rate/qwen_vl_interface': 8.342975604031974e-06, 'learning_rate/action_model': 8.342975604031974e-05, 'epoch': 7.48}) +[2026-06-20 17:03:37,802][starVLA.training.train_starvla][INFO] - Step 4045, Loss: {'train/loss': 0.059008095413446426, 'train/grad_norm_pre_clip': 4.995687961578369, 'timing/data': 0.011706560966558754, 'timing/model': 1.4069072629790753, 'learning_rate/qwen_vl_interface': 8.340663655945462e-06, 'learning_rate/action_model': 8.340663655945462e-05, 'epoch': 7.48}) +[2026-06-20 17:03:42,363][starVLA.training.train_starvla][INFO] - Step 4046, Loss: {'train/loss': 0.1591094732284546, 'train/grad_norm_pre_clip': 8.183257102966309, 'timing/data': 0.011700219940394163, 'timing/model': 1.1841823969734833, 'learning_rate/qwen_vl_interface': 8.338350437403866e-06, 'learning_rate/action_model': 8.338350437403865e-05, 'epoch': 7.48}) +[2026-06-20 17:03:47,133][starVLA.training.train_starvla][INFO] - Step 4047, Loss: {'train/loss': 0.15574470162391663, 'train/grad_norm_pre_clip': 15.825408935546875, 'timing/data': 0.011708681005984545, 'timing/model': 1.183484912966378, 'learning_rate/qwen_vl_interface': 8.336035949358063e-06, 'learning_rate/action_model': 8.336035949358062e-05, 'epoch': 7.48}) +[2026-06-20 17:03:51,912][starVLA.training.train_starvla][INFO] - Step 4048, Loss: {'train/loss': 0.24426092207431793, 'train/grad_norm_pre_clip': 13.109094619750977, 'timing/data': 0.0144974070135504, 'timing/model': 1.1921162731014192, 'learning_rate/qwen_vl_interface': 8.333720192759448e-06, 'learning_rate/action_model': 8.333720192759448e-05, 'epoch': 7.49}) +[2026-06-20 17:03:56,688][starVLA.training.train_starvla][INFO] - Step 4049, Loss: {'train/loss': 0.11142770200967789, 'train/grad_norm_pre_clip': 13.441117286682129, 'timing/data': 0.01190249901264906, 'timing/model': 1.17685497389175, 'learning_rate/qwen_vl_interface': 8.331403168559946e-06, 'learning_rate/action_model': 8.331403168559945e-05, 'epoch': 7.49}) +[2026-06-20 17:04:01,480][starVLA.training.train_starvla][INFO] - Step 4050, Loss: {'train/loss': 0.26496562361717224, 'train/grad_norm_pre_clip': 10.288960456848145, 'timing/data': 0.011907489039003849, 'timing/model': 1.4493924379348755, 'learning_rate/qwen_vl_interface': 8.329084877711995e-06, 'learning_rate/action_model': 8.329084877711996e-05, 'epoch': 7.49}) +[2026-06-20 17:04:06,002][starVLA.training.train_starvla][INFO] - Step 4051, Loss: {'train/loss': 0.11007390171289444, 'train/grad_norm_pre_clip': 7.618995666503906, 'timing/data': 0.012049418990500271, 'timing/model': 1.1770051789935678, 'learning_rate/qwen_vl_interface': 8.32676532116856e-06, 'learning_rate/action_model': 8.326765321168559e-05, 'epoch': 7.49}) +[2026-06-20 17:04:10,816][starVLA.training.train_starvla][INFO] - Step 4052, Loss: {'train/loss': 0.13131994009017944, 'train/grad_norm_pre_clip': 9.602042198181152, 'timing/data': 0.01302305399440229, 'timing/model': 1.1977590080350637, 'learning_rate/qwen_vl_interface': 8.324444499883119e-06, 'learning_rate/action_model': 8.324444499883119e-05, 'epoch': 7.49}) +[2026-06-20 17:04:15,562][starVLA.training.train_starvla][INFO] - Step 4053, Loss: {'train/loss': 0.011641905643045902, 'train/grad_norm_pre_clip': 1.377172589302063, 'timing/data': 0.011839480022899806, 'timing/model': 1.191810051095672, 'learning_rate/qwen_vl_interface': 8.322122414809678e-06, 'learning_rate/action_model': 8.322122414809677e-05, 'epoch': 7.5}) +[2026-06-20 17:04:20,297][starVLA.training.train_starvla][INFO] - Step 4054, Loss: {'train/loss': 0.00923094805330038, 'train/grad_norm_pre_clip': 1.5522944927215576, 'timing/data': 0.011915159993804991, 'timing/model': 1.17759684601333, 'learning_rate/qwen_vl_interface': 8.319799066902755e-06, 'learning_rate/action_model': 8.319799066902755e-05, 'epoch': 7.5}) +[2026-06-20 17:04:25,060][starVLA.training.train_starvla][INFO] - Step 4055, Loss: {'train/loss': 0.15794380009174347, 'train/grad_norm_pre_clip': 11.692296981811523, 'timing/data': 0.013222642941400409, 'timing/model': 1.192961851018481, 'learning_rate/qwen_vl_interface': 8.317474457117393e-06, 'learning_rate/action_model': 8.317474457117392e-05, 'epoch': 7.5}) +[2026-06-20 17:04:29,908][starVLA.training.train_starvla][INFO] - Step 4056, Loss: {'train/loss': 0.02803519181907177, 'train/grad_norm_pre_clip': 4.932525634765625, 'timing/data': 0.013578730984590948, 'timing/model': 1.4563783770427108, 'learning_rate/qwen_vl_interface': 8.315148586409148e-06, 'learning_rate/action_model': 8.315148586409148e-05, 'epoch': 7.5}) +[2026-06-20 17:04:34,473][starVLA.training.train_starvla][INFO] - Step 4057, Loss: {'train/loss': 0.21950341761112213, 'train/grad_norm_pre_clip': 20.946086883544922, 'timing/data': 0.01208074891474098, 'timing/model': 1.1877228950615972, 'learning_rate/qwen_vl_interface': 8.312821455734104e-06, 'learning_rate/action_model': 8.312821455734103e-05, 'epoch': 7.5}) +[2026-06-20 17:04:39,295][starVLA.training.train_starvla][INFO] - Step 4058, Loss: {'train/loss': 0.2884998321533203, 'train/grad_norm_pre_clip': 13.79544448852539, 'timing/data': 0.011314023053273559, 'timing/model': 1.1842342159943655, 'learning_rate/qwen_vl_interface': 8.31049306604885e-06, 'learning_rate/action_model': 8.310493066048849e-05, 'epoch': 7.5}) +[2026-06-20 17:04:44,588][starVLA.training.train_starvla][INFO] - Step 4059, Loss: {'train/loss': 0.26570942997932434, 'train/grad_norm_pre_clip': 10.314530372619629, 'timing/data': 0.012030599056743085, 'timing/model': 1.3024599240161479, 'learning_rate/qwen_vl_interface': 8.3081634183105e-06, 'learning_rate/action_model': 8.308163418310501e-05, 'epoch': 7.51}) +[2026-06-20 17:04:49,969][starVLA.training.train_starvla][INFO] - Step 4060, Loss: {'train/loss': 0.40600132942199707, 'train/grad_norm_pre_clip': 18.101539611816406, 'timing/data': 0.011595521005801857, 'timing/model': 1.3118883810238913, 'learning_rate/qwen_vl_interface': 8.305832513476689e-06, 'learning_rate/action_model': 8.305832513476688e-05, 'epoch': 7.51}) +[2026-06-20 17:04:55,244][starVLA.training.train_starvla][INFO] - Step 4061, Loss: {'train/loss': 0.030708231031894684, 'train/grad_norm_pre_clip': 5.197605133056641, 'timing/data': 0.01647959800902754, 'timing/model': 1.3174735460197553, 'learning_rate/qwen_vl_interface': 8.303500352505557e-06, 'learning_rate/action_model': 8.303500352505557e-05, 'epoch': 7.51}) +[2026-06-20 17:05:00,540][starVLA.training.train_starvla][INFO] - Step 4062, Loss: {'train/loss': 0.11669386923313141, 'train/grad_norm_pre_clip': 6.121054649353027, 'timing/data': 0.011650550994090736, 'timing/model': 1.5303691099397838, 'learning_rate/qwen_vl_interface': 8.301166936355773e-06, 'learning_rate/action_model': 8.301166936355773e-05, 'epoch': 7.51}) +[2026-06-20 17:05:05,572][starVLA.training.train_starvla][INFO] - Step 4063, Loss: {'train/loss': 0.021967163309454918, 'train/grad_norm_pre_clip': 3.8812010288238525, 'timing/data': 0.013024804065935314, 'timing/model': 1.3111511450260878, 'learning_rate/qwen_vl_interface': 8.298832265986514e-06, 'learning_rate/action_model': 8.298832265986513e-05, 'epoch': 7.51}) +[2026-06-20 17:05:10,765][starVLA.training.train_starvla][INFO] - Step 4064, Loss: {'train/loss': 0.08905863016843796, 'train/grad_norm_pre_clip': 13.44704818725586, 'timing/data': 0.011867430992424488, 'timing/model': 1.1690674920100719, 'learning_rate/qwen_vl_interface': 8.296496342357475e-06, 'learning_rate/action_model': 8.296496342357474e-05, 'epoch': 7.52}) +[2026-06-20 17:05:15,539][starVLA.training.train_starvla][INFO] - Step 4065, Loss: {'train/loss': 0.10403610020875931, 'train/grad_norm_pre_clip': 10.474898338317871, 'timing/data': 0.011772989993914962, 'timing/model': 1.1708349459804595, 'learning_rate/qwen_vl_interface': 8.294159166428864e-06, 'learning_rate/action_model': 8.294159166428864e-05, 'epoch': 7.52}) +[2026-06-20 17:05:20,319][starVLA.training.train_starvla][INFO] - Step 4066, Loss: {'train/loss': 0.16410313546657562, 'train/grad_norm_pre_clip': 11.831546783447266, 'timing/data': 0.011495551909320056, 'timing/model': 1.1823427309282124, 'learning_rate/qwen_vl_interface': 8.29182073916141e-06, 'learning_rate/action_model': 8.29182073916141e-05, 'epoch': 7.52}) +[2026-06-20 17:05:25,352][starVLA.training.train_starvla][INFO] - Step 4067, Loss: {'train/loss': 0.11196496337652206, 'train/grad_norm_pre_clip': 15.067787170410156, 'timing/data': 0.013121622963808477, 'timing/model': 1.5381089710863307, 'learning_rate/qwen_vl_interface': 8.28948106151635e-06, 'learning_rate/action_model': 8.28948106151635e-05, 'epoch': 7.52}) +[2026-06-20 17:05:30,548][starVLA.training.train_starvla][INFO] - Step 4068, Loss: {'train/loss': 0.29829394817352295, 'train/grad_norm_pre_clip': 14.268664360046387, 'timing/data': 0.012147660017944872, 'timing/model': 1.4523305939510465, 'learning_rate/qwen_vl_interface': 8.287140134455435e-06, 'learning_rate/action_model': 8.287140134455435e-05, 'epoch': 7.52}) +[2026-06-20 17:05:35,578][starVLA.training.train_starvla][INFO] - Step 4069, Loss: {'train/loss': 0.2162473499774933, 'train/grad_norm_pre_clip': 11.518280029296875, 'timing/data': 0.012228569015860558, 'timing/model': 1.293074386077933, 'learning_rate/qwen_vl_interface': 8.284797958940934e-06, 'learning_rate/action_model': 8.284797958940935e-05, 'epoch': 7.53}) +[2026-06-20 17:05:40,848][starVLA.training.train_starvla][INFO] - Step 4070, Loss: {'train/loss': 0.25945785641670227, 'train/grad_norm_pre_clip': 17.65839385986328, 'timing/data': 0.013774150982499123, 'timing/model': 1.3108840100467205, 'learning_rate/qwen_vl_interface': 8.282454535935626e-06, 'learning_rate/action_model': 8.282454535935626e-05, 'epoch': 7.53}) +[2026-06-20 17:05:46,120][starVLA.training.train_starvla][INFO] - Step 4071, Loss: {'train/loss': 0.20447222888469696, 'train/grad_norm_pre_clip': 18.08990478515625, 'timing/data': 0.013957310933619738, 'timing/model': 1.311821099021472, 'learning_rate/qwen_vl_interface': 8.280109866402805e-06, 'learning_rate/action_model': 8.280109866402806e-05, 'epoch': 7.53}) +[2026-06-20 17:05:51,427][starVLA.training.train_starvla][INFO] - Step 4072, Loss: {'train/loss': 0.10435072332620621, 'train/grad_norm_pre_clip': 11.907607078552246, 'timing/data': 0.013963590958155692, 'timing/model': 1.3119489119853824, 'learning_rate/qwen_vl_interface': 8.277763951306277e-06, 'learning_rate/action_model': 8.277763951306277e-05, 'epoch': 7.53}) +[2026-06-20 17:05:56,743][starVLA.training.train_starvla][INFO] - Step 4073, Loss: {'train/loss': 0.15774337947368622, 'train/grad_norm_pre_clip': 12.041311264038086, 'timing/data': 0.015024265041574836, 'timing/model': 1.5540034850127995, 'learning_rate/qwen_vl_interface': 8.275416791610356e-06, 'learning_rate/action_model': 8.275416791610355e-05, 'epoch': 7.53}) +[2026-06-20 17:06:01,804][starVLA.training.train_starvla][INFO] - Step 4074, Loss: {'train/loss': 0.03598468005657196, 'train/grad_norm_pre_clip': 7.148662090301514, 'timing/data': 0.017369683017022908, 'timing/model': 1.3183003059821203, 'learning_rate/qwen_vl_interface': 8.27306838827987e-06, 'learning_rate/action_model': 8.273068388279871e-05, 'epoch': 7.53}) +[2026-06-20 17:06:07,076][starVLA.training.train_starvla][INFO] - Step 4075, Loss: {'train/loss': 0.22128210961818695, 'train/grad_norm_pre_clip': 12.209038734436035, 'timing/data': 0.013203223934397101, 'timing/model': 1.312822075909935, 'learning_rate/qwen_vl_interface': 8.270718742280162e-06, 'learning_rate/action_model': 8.270718742280162e-05, 'epoch': 7.54}) +[2026-06-20 17:06:12,343][starVLA.training.train_starvla][INFO] - Step 4076, Loss: {'train/loss': 0.017001081258058548, 'train/grad_norm_pre_clip': 1.7465541362762451, 'timing/data': 0.011742840986698866, 'timing/model': 1.2987444000318646, 'learning_rate/qwen_vl_interface': 8.268367854577084e-06, 'learning_rate/action_model': 8.268367854577082e-05, 'epoch': 7.54}) +[2026-06-20 17:06:17,607][starVLA.training.train_starvla][INFO] - Step 4077, Loss: {'train/loss': 0.048301585018634796, 'train/grad_norm_pre_clip': 3.269723892211914, 'timing/data': 0.014088019961491227, 'timing/model': 1.3114451100118458, 'learning_rate/qwen_vl_interface': 8.266015726136991e-06, 'learning_rate/action_model': 8.266015726136991e-05, 'epoch': 7.54}) +[2026-06-20 17:06:22,896][starVLA.training.train_starvla][INFO] - Step 4078, Loss: {'train/loss': 0.027475204318761826, 'train/grad_norm_pre_clip': 2.6906445026397705, 'timing/data': 0.013897591037675738, 'timing/model': 1.3100399199174717, 'learning_rate/qwen_vl_interface': 8.26366235792676e-06, 'learning_rate/action_model': 8.26366235792676e-05, 'epoch': 7.54}) +[2026-06-20 17:06:28,187][starVLA.training.train_starvla][INFO] - Step 4079, Loss: {'train/loss': 0.0464746430516243, 'train/grad_norm_pre_clip': 4.018911838531494, 'timing/data': 0.01164429192431271, 'timing/model': 1.534686223953031, 'learning_rate/qwen_vl_interface': 8.261307750913769e-06, 'learning_rate/action_model': 8.261307750913769e-05, 'epoch': 7.54}) +[2026-06-20 17:06:33,240][starVLA.training.train_starvla][INFO] - Step 4080, Loss: {'train/loss': 0.1751510500907898, 'train/grad_norm_pre_clip': 11.637558937072754, 'timing/data': 0.011760432040318847, 'timing/model': 1.2990912100067362, 'learning_rate/qwen_vl_interface': 8.258951906065908e-06, 'learning_rate/action_model': 8.258951906065908e-05, 'epoch': 7.55}) +[2026-06-20 17:06:38,783][starVLA.training.train_starvla][INFO] - Step 4081, Loss: {'train/loss': 0.16169531643390656, 'train/grad_norm_pre_clip': 8.661203384399414, 'timing/data': 0.014313378022052348, 'timing/model': 1.2909651250811294, 'learning_rate/qwen_vl_interface': 8.256594824351577e-06, 'learning_rate/action_model': 8.256594824351576e-05, 'epoch': 7.55}) +[2026-06-20 17:06:43,579][starVLA.training.train_starvla][INFO] - Step 4082, Loss: {'train/loss': 0.16524147987365723, 'train/grad_norm_pre_clip': 11.444462776184082, 'timing/data': 0.011770891956984997, 'timing/model': 1.1743483460741118, 'learning_rate/qwen_vl_interface': 8.254236506739683e-06, 'learning_rate/action_model': 8.254236506739681e-05, 'epoch': 7.55}) +[2026-06-20 17:06:48,350][starVLA.training.train_starvla][INFO] - Step 4083, Loss: {'train/loss': 0.13237731158733368, 'train/grad_norm_pre_clip': 11.090035438537598, 'timing/data': 0.01295189606025815, 'timing/model': 1.1883745299419388, 'learning_rate/qwen_vl_interface': 8.25187695419964e-06, 'learning_rate/action_model': 8.25187695419964e-05, 'epoch': 7.55}) +[2026-06-20 17:06:53,147][starVLA.training.train_starvla][INFO] - Step 4084, Loss: {'train/loss': 0.1274874359369278, 'train/grad_norm_pre_clip': 9.786977767944336, 'timing/data': 0.013352283975109458, 'timing/model': 1.186712131020613, 'learning_rate/qwen_vl_interface': 8.249516167701373e-06, 'learning_rate/action_model': 8.249516167701374e-05, 'epoch': 7.55}) +[2026-06-20 17:06:57,955][starVLA.training.train_starvla][INFO] - Step 4085, Loss: {'train/loss': 0.04026167467236519, 'train/grad_norm_pre_clip': 5.554011344909668, 'timing/data': 0.011767360963858664, 'timing/model': 1.4184966009343043, 'learning_rate/qwen_vl_interface': 8.247154148215311e-06, 'learning_rate/action_model': 8.24715414821531e-05, 'epoch': 7.55}) +[2026-06-20 17:07:02,644][starVLA.training.train_starvla][INFO] - Step 4086, Loss: {'train/loss': 0.02224554866552353, 'train/grad_norm_pre_clip': 7.730814456939697, 'timing/data': 0.011907570995390415, 'timing/model': 1.1752575819846243, 'learning_rate/qwen_vl_interface': 8.24479089671239e-06, 'learning_rate/action_model': 8.24479089671239e-05, 'epoch': 7.56}) +[2026-06-20 17:07:07,463][starVLA.training.train_starvla][INFO] - Step 4087, Loss: {'train/loss': 0.19070205092430115, 'train/grad_norm_pre_clip': 14.474501609802246, 'timing/data': 0.011673692963086069, 'timing/model': 1.1749832449713722, 'learning_rate/qwen_vl_interface': 8.242426414164056e-06, 'learning_rate/action_model': 8.242426414164056e-05, 'epoch': 7.56}) +[2026-06-20 17:07:12,198][starVLA.training.train_starvla][INFO] - Step 4088, Loss: {'train/loss': 0.1539350003004074, 'train/grad_norm_pre_clip': 15.299812316894531, 'timing/data': 0.0115849319845438, 'timing/model': 1.1772465170361102, 'learning_rate/qwen_vl_interface': 8.240060701542257e-06, 'learning_rate/action_model': 8.240060701542257e-05, 'epoch': 7.56}) +[2026-06-20 17:07:17,267][starVLA.training.train_starvla][INFO] - Step 4089, Loss: {'train/loss': 0.06033588945865631, 'train/grad_norm_pre_clip': 7.9924492835998535, 'timing/data': 0.013204304035753012, 'timing/model': 1.32444125902839, 'learning_rate/qwen_vl_interface': 8.237693759819447e-06, 'learning_rate/action_model': 8.237693759819448e-05, 'epoch': 7.56}) +[2026-06-20 17:07:22,584][starVLA.training.train_starvla][INFO] - Step 4090, Loss: {'train/loss': 0.11621753871440887, 'train/grad_norm_pre_clip': 9.591184616088867, 'timing/data': 0.011732362094335258, 'timing/model': 1.3065308299846947, 'learning_rate/qwen_vl_interface': 8.235325589968587e-06, 'learning_rate/action_model': 8.235325589968587e-05, 'epoch': 7.56}) +[2026-06-20 17:07:27,891][starVLA.training.train_starvla][INFO] - Step 4091, Loss: {'train/loss': 0.022869905456900597, 'train/grad_norm_pre_clip': 2.8521888256073, 'timing/data': 0.011873671086505055, 'timing/model': 1.5511611310066655, 'learning_rate/qwen_vl_interface': 8.232956192963143e-06, 'learning_rate/action_model': 8.232956192963144e-05, 'epoch': 7.57}) +[2026-06-20 17:07:33,180][starVLA.training.train_starvla][INFO] - Step 4092, Loss: {'train/loss': 0.18618786334991455, 'train/grad_norm_pre_clip': 11.842621803283691, 'timing/data': 0.011752552003599703, 'timing/model': 1.4844974930165336, 'learning_rate/qwen_vl_interface': 8.230585569777086e-06, 'learning_rate/action_model': 8.230585569777085e-05, 'epoch': 7.57}) +[2026-06-20 17:07:38,299][starVLA.training.train_starvla][INFO] - Step 4093, Loss: {'train/loss': 0.018420385196805, 'train/grad_norm_pre_clip': 1.384913682937622, 'timing/data': 0.018239498953334987, 'timing/model': 1.323232444934547, 'learning_rate/qwen_vl_interface': 8.228213721384885e-06, 'learning_rate/action_model': 8.228213721384885e-05, 'epoch': 7.57}) +[2026-06-20 17:07:43,374][starVLA.training.train_starvla][INFO] - Step 4094, Loss: {'train/loss': 0.01390785165131092, 'train/grad_norm_pre_clip': 1.5776700973510742, 'timing/data': 0.011860101018100977, 'timing/model': 1.175688249990344, 'learning_rate/qwen_vl_interface': 8.225840648761523e-06, 'learning_rate/action_model': 8.225840648761523e-05, 'epoch': 7.57}) +[2026-06-20 17:07:48,587][starVLA.training.train_starvla][INFO] - Step 4095, Loss: {'train/loss': 0.12417936325073242, 'train/grad_norm_pre_clip': 14.32513427734375, 'timing/data': 0.011801311047747731, 'timing/model': 1.308514715055935, 'learning_rate/qwen_vl_interface': 8.223466352882476e-06, 'learning_rate/action_model': 8.223466352882475e-05, 'epoch': 7.57}) +[2026-06-20 17:07:53,648][starVLA.training.train_starvla][INFO] - Step 4096, Loss: {'train/loss': 0.2583601772785187, 'train/grad_norm_pre_clip': 19.508926391601562, 'timing/data': 0.011854161974042654, 'timing/model': 1.3192883939482272, 'learning_rate/qwen_vl_interface': 8.22109083472373e-06, 'learning_rate/action_model': 8.22109083472373e-05, 'epoch': 7.57}) +[2026-06-20 17:07:58,435][starVLA.training.train_starvla][INFO] - Step 4097, Loss: {'train/loss': 0.38728979229927063, 'train/grad_norm_pre_clip': 10.503179550170898, 'timing/data': 0.011790272081270814, 'timing/model': 1.418848072993569, 'learning_rate/qwen_vl_interface': 8.218714095261769e-06, 'learning_rate/action_model': 8.218714095261768e-05, 'epoch': 7.58}) +[2026-06-20 17:08:03,780][starVLA.training.train_starvla][INFO] - Step 4098, Loss: {'train/loss': 0.131336972117424, 'train/grad_norm_pre_clip': 8.865121841430664, 'timing/data': 0.011969489976763725, 'timing/model': 1.6308601689524949, 'learning_rate/qwen_vl_interface': 8.21633613547358e-06, 'learning_rate/action_model': 8.21633613547358e-05, 'epoch': 7.58}) +[2026-06-20 17:08:08,881][starVLA.training.train_starvla][INFO] - Step 4099, Loss: {'train/loss': 0.22081473469734192, 'train/grad_norm_pre_clip': 14.057196617126465, 'timing/data': 0.011937931063584983, 'timing/model': 1.315473800059408, 'learning_rate/qwen_vl_interface': 8.213956956336656e-06, 'learning_rate/action_model': 8.213956956336656e-05, 'epoch': 7.58}) +[2026-06-20 17:08:14,185][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 4100 +[2026-06-20 17:14:34,697][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 4100 +[2026-06-20 17:14:34,698][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 4100 +[2026-06-20 17:21:25,462][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 4100 +[2026-06-20 17:21:25,463][starVLA.training.train_starvla][INFO] - Step 4100, Loss: {'train/loss': 0.36539986729621887, 'train/grad_norm_pre_clip': 17.173484802246094, 'eval/loss': 0.6842801570892334, 'eval/latency_0/loss': 0.7632033228874207, 'eval/latency_1/loss': 1.0034936666488647, 'eval/latency_2/loss': 1.0471612215042114, 'eval/latency_3/loss': 0.4738469421863556, 'eval/latency_4/loss': 0.09614787250757217, 'eval/flappy/loss': 0.6842801570892334, 'eval/flappy/latency_0/loss': 0.7632033228874207, 'eval/flappy/latency_1/loss': 1.0034936666488647, 'eval/flappy/latency_2/loss': 1.0471612215042114, 'eval/flappy/latency_3/loss': 0.4738469421863556, 'eval/flappy/latency_4/loss': 0.09614787250757217, 'eval/flap_precision': 0.17333333333333334, 'eval/flap_recall': 0.09454545454545454, 'eval/flap_f1': 0.12235294117647058, 'eval/flap_rate_model': 0.041666666666666664, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.12235294117647058, 'eval/latency_0/flap_precision': 0.17333333333333334, 'eval/latency_0/flap_recall': 0.09454545454545454, 'eval/latency_0/flap_f1': 0.12235294117647058, 'eval/latency_0/flap_rate_model': 0.041666666666666664, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.12235294117647058, 'timing/data': 0.011741681955754757, 'timing/model': 1.3077556709758937, 'learning_rate/qwen_vl_interface': 8.211576558828985e-06, 'learning_rate/action_model': 8.211576558828985e-05, 'epoch': 7.58}) +[2026-06-20 17:21:25,464][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_4100_state +[2026-06-20 17:21:25,468][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-20 17:21:33,912][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_4100_state/model.safetensors +[2026-06-20 17:21:54,057][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_4100_state/optimizer.bin +[2026-06-20 17:21:54,058][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_4100_state/sampler.bin +[2026-06-20 17:21:54,058][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_4100_state/sampler_1.bin +[2026-06-20 17:21:54,059][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_4100_state/random_states_0.pkl +[2026-06-20 17:21:56,705][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4100 +[2026-06-20 17:21:56,705][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-20 17:21:56,718][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-20 17:22:02,013][starVLA.training.train_starvla][INFO] - Step 4101, Loss: {'train/loss': 0.13068875670433044, 'train/grad_norm_pre_clip': 12.86175537109375, 'timing/data': 0.012236083042807877, 'timing/model': 1.291734875063412, 'learning_rate/qwen_vl_interface': 8.209194943929059e-06, 'learning_rate/action_model': 8.20919494392906e-05, 'epoch': 7.58}) +[2026-06-20 17:22:06,916][starVLA.training.train_starvla][INFO] - Step 4102, Loss: {'train/loss': 0.2480337917804718, 'train/grad_norm_pre_clip': 13.33292007446289, 'timing/data': 0.01315745001193136, 'timing/model': 1.2058118570130318, 'learning_rate/qwen_vl_interface': 8.20681211261587e-06, 'learning_rate/action_model': 8.20681211261587e-05, 'epoch': 7.59}) +[2026-06-20 17:22:11,815][starVLA.training.train_starvla][INFO] - Step 4103, Loss: {'train/loss': 0.10789377987384796, 'train/grad_norm_pre_clip': 7.4668192863464355, 'timing/data': 0.012052393984049559, 'timing/model': 1.409252212033607, 'learning_rate/qwen_vl_interface': 8.204428065868911e-06, 'learning_rate/action_model': 8.204428065868911e-05, 'epoch': 7.59}) +[2026-06-20 17:22:16,521][starVLA.training.train_starvla][INFO] - Step 4104, Loss: {'train/loss': 0.018571941182017326, 'train/grad_norm_pre_clip': 1.792054295539856, 'timing/data': 0.0119401840493083, 'timing/model': 1.170434219064191, 'learning_rate/qwen_vl_interface': 8.202042804668173e-06, 'learning_rate/action_model': 8.202042804668172e-05, 'epoch': 7.59}) +[2026-06-20 17:22:21,445][starVLA.training.train_starvla][INFO] - Step 4105, Loss: {'train/loss': 0.2236533761024475, 'train/grad_norm_pre_clip': 11.1359224319458, 'timing/data': 0.017608847003430128, 'timing/model': 1.2260967760812491, 'learning_rate/qwen_vl_interface': 8.199656329994145e-06, 'learning_rate/action_model': 8.199656329994145e-05, 'epoch': 7.59}) +[2026-06-20 17:22:26,421][starVLA.training.train_starvla][INFO] - Step 4106, Loss: {'train/loss': 0.12190273404121399, 'train/grad_norm_pre_clip': 88.19735717773438, 'timing/data': 0.015620321966707706, 'timing/model': 1.2180497290100902, 'learning_rate/qwen_vl_interface': 8.19726864282782e-06, 'learning_rate/action_model': 8.197268642827818e-05, 'epoch': 7.59}) +[2026-06-20 17:22:31,296][starVLA.training.train_starvla][INFO] - Step 4107, Loss: {'train/loss': 0.09783697873353958, 'train/grad_norm_pre_clip': 6.5418243408203125, 'timing/data': 0.016718908911570907, 'timing/model': 1.214842814952135, 'learning_rate/qwen_vl_interface': 8.194879744150682e-06, 'learning_rate/action_model': 8.194879744150681e-05, 'epoch': 7.6}) +[2026-06-20 17:22:36,209][starVLA.training.train_starvla][INFO] - Step 4108, Loss: {'train/loss': 0.12177176028490067, 'train/grad_norm_pre_clip': 13.480743408203125, 'timing/data': 0.012082513072527945, 'timing/model': 1.1800049289595336, 'learning_rate/qwen_vl_interface': 8.192489634944718e-06, 'learning_rate/action_model': 8.192489634944718e-05, 'epoch': 7.6}) +[2026-06-20 17:22:41,025][starVLA.training.train_starvla][INFO] - Step 4109, Loss: {'train/loss': 0.453934907913208, 'train/grad_norm_pre_clip': 11.640548706054688, 'timing/data': 0.01815978391095996, 'timing/model': 1.44624289998319, 'learning_rate/qwen_vl_interface': 8.190098316192416e-06, 'learning_rate/action_model': 8.190098316192416e-05, 'epoch': 7.6}) +[2026-06-20 17:22:45,585][starVLA.training.train_starvla][INFO] - Step 4110, Loss: {'train/loss': 0.1880491077899933, 'train/grad_norm_pre_clip': 11.703152656555176, 'timing/data': 0.012055223109200597, 'timing/model': 1.1785059500252828, 'learning_rate/qwen_vl_interface': 8.18770578887675e-06, 'learning_rate/action_model': 8.18770578887675e-05, 'epoch': 7.6}) +[2026-06-20 17:22:50,376][starVLA.training.train_starvla][INFO] - Step 4111, Loss: {'train/loss': 0.16941462457180023, 'train/grad_norm_pre_clip': 3.85776424407959, 'timing/data': 0.012168252025730908, 'timing/model': 1.189831943018362, 'learning_rate/qwen_vl_interface': 8.185312053981204e-06, 'learning_rate/action_model': 8.185312053981204e-05, 'epoch': 7.6}) +[2026-06-20 17:22:55,347][starVLA.training.train_starvla][INFO] - Step 4112, Loss: {'train/loss': 0.14121964573860168, 'train/grad_norm_pre_clip': 9.903600692749023, 'timing/data': 0.01487573399208486, 'timing/model': 1.2405901639722288, 'learning_rate/qwen_vl_interface': 8.182917112489747e-06, 'learning_rate/action_model': 8.182917112489746e-05, 'epoch': 7.6}) +[2026-06-20 17:23:00,308][starVLA.training.train_starvla][INFO] - Step 4113, Loss: {'train/loss': 0.14101172983646393, 'train/grad_norm_pre_clip': 9.825368881225586, 'timing/data': 0.017499365960247815, 'timing/model': 1.197373146074824, 'learning_rate/qwen_vl_interface': 8.18052096538685e-06, 'learning_rate/action_model': 8.18052096538685e-05, 'epoch': 7.61}) +[2026-06-20 17:23:04,999][starVLA.training.train_starvla][INFO] - Step 4114, Loss: {'train/loss': 0.20986244082450867, 'train/grad_norm_pre_clip': 12.596661567687988, 'timing/data': 0.013578868005424738, 'timing/model': 1.1902250769780949, 'learning_rate/qwen_vl_interface': 8.178123613657483e-06, 'learning_rate/action_model': 8.178123613657482e-05, 'epoch': 7.61}) +[2026-06-20 17:23:09,808][starVLA.training.train_starvla][INFO] - Step 4115, Loss: {'train/loss': 0.03473372012376785, 'train/grad_norm_pre_clip': 6.247644901275635, 'timing/data': 0.011994922999292612, 'timing/model': 1.4446704989532009, 'learning_rate/qwen_vl_interface': 8.175725058287102e-06, 'learning_rate/action_model': 8.175725058287102e-05, 'epoch': 7.61}) +[2026-06-20 17:23:14,352][starVLA.training.train_starvla][INFO] - Step 4116, Loss: {'train/loss': 0.04882949963212013, 'train/grad_norm_pre_clip': 4.06574821472168, 'timing/data': 0.012663970002904534, 'timing/model': 1.1776399540249258, 'learning_rate/qwen_vl_interface': 8.173325300261664e-06, 'learning_rate/action_model': 8.173325300261663e-05, 'epoch': 7.61}) +[2026-06-20 17:23:19,136][starVLA.training.train_starvla][INFO] - Step 4117, Loss: {'train/loss': 0.18755602836608887, 'train/grad_norm_pre_clip': 9.968408584594727, 'timing/data': 0.013329648063518107, 'timing/model': 1.1763851770665497, 'learning_rate/qwen_vl_interface': 8.170924340567617e-06, 'learning_rate/action_model': 8.170924340567616e-05, 'epoch': 7.61}) +[2026-06-20 17:23:23,908][starVLA.training.train_starvla][INFO] - Step 4118, Loss: {'train/loss': 0.10272523760795593, 'train/grad_norm_pre_clip': 7.675938606262207, 'timing/data': 0.011867332039400935, 'timing/model': 1.1762781559955329, 'learning_rate/qwen_vl_interface': 8.168522180191909e-06, 'learning_rate/action_model': 8.168522180191909e-05, 'epoch': 7.62}) +[2026-06-20 17:23:28,629][starVLA.training.train_starvla][INFO] - Step 4119, Loss: {'train/loss': 0.17497016489505768, 'train/grad_norm_pre_clip': 8.220508575439453, 'timing/data': 0.012059352011419833, 'timing/model': 1.18501529796049, 'learning_rate/qwen_vl_interface': 8.166118820121972e-06, 'learning_rate/action_model': 8.166118820121973e-05, 'epoch': 7.62}) +[2026-06-20 17:23:33,419][starVLA.training.train_starvla][INFO] - Step 4120, Loss: {'train/loss': 0.1982383280992508, 'train/grad_norm_pre_clip': 8.539358139038086, 'timing/data': 0.01155442395247519, 'timing/model': 1.1735983330290765, 'learning_rate/qwen_vl_interface': 8.163714261345744e-06, 'learning_rate/action_model': 8.163714261345742e-05, 'epoch': 7.62}) +[2026-06-20 17:23:38,399][starVLA.training.train_starvla][INFO] - Step 4121, Loss: {'train/loss': 0.03679359331727028, 'train/grad_norm_pre_clip': 6.761135578155518, 'timing/data': 0.012349069933407009, 'timing/model': 1.4197769890306517, 'learning_rate/qwen_vl_interface': 8.16130850485164e-06, 'learning_rate/action_model': 8.161308504851639e-05, 'epoch': 7.62}) +[2026-06-20 17:23:43,009][starVLA.training.train_starvla][INFO] - Step 4122, Loss: {'train/loss': 0.15692861378192902, 'train/grad_norm_pre_clip': 10.998889923095703, 'timing/data': 0.012005220982246101, 'timing/model': 1.1796569819562137, 'learning_rate/qwen_vl_interface': 8.158901551628579e-06, 'learning_rate/action_model': 8.158901551628579e-05, 'epoch': 7.62}) +[2026-06-20 17:23:47,957][starVLA.training.train_starvla][INFO] - Step 4123, Loss: {'train/loss': 0.3716721832752228, 'train/grad_norm_pre_clip': 8.765899658203125, 'timing/data': 0.012580640031956136, 'timing/model': 1.1807600290048867, 'learning_rate/qwen_vl_interface': 8.156493402665971e-06, 'learning_rate/action_model': 8.15649340266597e-05, 'epoch': 7.62}) +[2026-06-20 17:23:52,892][starVLA.training.train_starvla][INFO] - Step 4124, Loss: {'train/loss': 0.0667625218629837, 'train/grad_norm_pre_clip': 5.139589309692383, 'timing/data': 0.013150467886589468, 'timing/model': 1.224074808997102, 'learning_rate/qwen_vl_interface': 8.154084058953713e-06, 'learning_rate/action_model': 8.154084058953713e-05, 'epoch': 7.63}) +[2026-06-20 17:23:57,776][starVLA.training.train_starvla][INFO] - Step 4125, Loss: {'train/loss': 0.10160025209188461, 'train/grad_norm_pre_clip': 8.444483757019043, 'timing/data': 0.012474599992856383, 'timing/model': 1.1774081079056486, 'learning_rate/qwen_vl_interface': 8.151673521482193e-06, 'learning_rate/action_model': 8.151673521482194e-05, 'epoch': 7.63}) +[2026-06-20 17:24:02,635][starVLA.training.train_starvla][INFO] - Step 4126, Loss: {'train/loss': 0.13371047377586365, 'train/grad_norm_pre_clip': 8.444832801818848, 'timing/data': 0.012029460980556905, 'timing/model': 1.1779451860347763, 'learning_rate/qwen_vl_interface': 8.149261791242297e-06, 'learning_rate/action_model': 8.149261791242297e-05, 'epoch': 7.63}) +[2026-06-20 17:24:07,584][starVLA.training.train_starvla][INFO] - Step 4127, Loss: {'train/loss': 0.14763134717941284, 'train/grad_norm_pre_clip': 6.901650428771973, 'timing/data': 0.014367664000019431, 'timing/model': 1.460282011074014, 'learning_rate/qwen_vl_interface': 8.146848869225391e-06, 'learning_rate/action_model': 8.146848869225391e-05, 'epoch': 7.63}) +[2026-06-20 17:24:12,248][starVLA.training.train_starvla][INFO] - Step 4128, Loss: {'train/loss': 0.08312734961509705, 'train/grad_norm_pre_clip': 4.104393482208252, 'timing/data': 0.012100009946152568, 'timing/model': 1.1812209960771725, 'learning_rate/qwen_vl_interface': 8.14443475642334e-06, 'learning_rate/action_model': 8.144434756423339e-05, 'epoch': 7.63}) +[2026-06-20 17:24:17,270][starVLA.training.train_starvla][INFO] - Step 4129, Loss: {'train/loss': 0.13908877968788147, 'train/grad_norm_pre_clip': 19.641469955444336, 'timing/data': 0.017354982905089855, 'timing/model': 1.2366941740037873, 'learning_rate/qwen_vl_interface': 8.142019453828493e-06, 'learning_rate/action_model': 8.142019453828493e-05, 'epoch': 7.64}) +[2026-06-20 17:24:22,283][starVLA.training.train_starvla][INFO] - Step 4130, Loss: {'train/loss': 0.06634697318077087, 'train/grad_norm_pre_clip': 5.242541790008545, 'timing/data': 0.013811875018291175, 'timing/model': 1.2189395610475913, 'learning_rate/qwen_vl_interface': 8.139602962433688e-06, 'learning_rate/action_model': 8.139602962433688e-05, 'epoch': 7.64}) +[2026-06-20 17:24:27,056][starVLA.training.train_starvla][INFO] - Step 4131, Loss: {'train/loss': 0.1174774020910263, 'train/grad_norm_pre_clip': 6.276352882385254, 'timing/data': 0.016158158075995743, 'timing/model': 1.182055331999436, 'learning_rate/qwen_vl_interface': 8.137185283232255e-06, 'learning_rate/action_model': 8.137185283232256e-05, 'epoch': 7.64}) +[2026-06-20 17:24:31,884][starVLA.training.train_starvla][INFO] - Step 4132, Loss: {'train/loss': 0.060440320521593094, 'train/grad_norm_pre_clip': 3.099126100540161, 'timing/data': 0.012194689945317805, 'timing/model': 1.178216425003484, 'learning_rate/qwen_vl_interface': 8.134766417218014e-06, 'learning_rate/action_model': 8.134766417218012e-05, 'epoch': 7.64}) +[2026-06-20 17:24:36,789][starVLA.training.train_starvla][INFO] - Step 4133, Loss: {'train/loss': 0.05544033274054527, 'train/grad_norm_pre_clip': 5.282411575317383, 'timing/data': 0.012236089911311865, 'timing/model': 1.4140788089716807, 'learning_rate/qwen_vl_interface': 8.132346365385262e-06, 'learning_rate/action_model': 8.132346365385261e-05, 'epoch': 7.64}) +[2026-06-20 17:24:41,424][starVLA.training.train_starvla][INFO] - Step 4134, Loss: {'train/loss': 0.061192724853754044, 'train/grad_norm_pre_clip': 4.2625732421875, 'timing/data': 0.012117060017772019, 'timing/model': 1.1802974690217525, 'learning_rate/qwen_vl_interface': 8.129925128728796e-06, 'learning_rate/action_model': 8.129925128728796e-05, 'epoch': 7.65}) +[2026-06-20 17:24:46,235][starVLA.training.train_starvla][INFO] - Step 4135, Loss: {'train/loss': 0.04008897766470909, 'train/grad_norm_pre_clip': 2.6989240646362305, 'timing/data': 0.011851271032355726, 'timing/model': 1.1809181270655245, 'learning_rate/qwen_vl_interface': 8.127502708243895e-06, 'learning_rate/action_model': 8.127502708243895e-05, 'epoch': 7.65}) +[2026-06-20 17:24:50,994][starVLA.training.train_starvla][INFO] - Step 4136, Loss: {'train/loss': 0.11491448432207108, 'train/grad_norm_pre_clip': 6.915287971496582, 'timing/data': 0.012215389986522496, 'timing/model': 1.1797305330401286, 'learning_rate/qwen_vl_interface': 8.125079104926324e-06, 'learning_rate/action_model': 8.125079104926323e-05, 'epoch': 7.65}) +[2026-06-20 17:24:55,812][starVLA.training.train_starvla][INFO] - Step 4137, Loss: {'train/loss': 0.04942530021071434, 'train/grad_norm_pre_clip': 8.210463523864746, 'timing/data': 0.011972280917689204, 'timing/model': 1.181348957004957, 'learning_rate/qwen_vl_interface': 8.122654319772332e-06, 'learning_rate/action_model': 8.122654319772332e-05, 'epoch': 7.65}) +[2026-06-20 17:25:00,636][starVLA.training.train_starvla][INFO] - Step 4138, Loss: {'train/loss': 0.05445723980665207, 'train/grad_norm_pre_clip': 4.127713680267334, 'timing/data': 0.012628756929188967, 'timing/model': 1.1933289589360356, 'learning_rate/qwen_vl_interface': 8.120228353778661e-06, 'learning_rate/action_model': 8.12022835377866e-05, 'epoch': 7.65}) +[2026-06-20 17:25:05,435][starVLA.training.train_starvla][INFO] - Step 4139, Loss: {'train/loss': 0.20568859577178955, 'train/grad_norm_pre_clip': 10.00017261505127, 'timing/data': 0.011902719968929887, 'timing/model': 1.4155818390427157, 'learning_rate/qwen_vl_interface': 8.117801207942528e-06, 'learning_rate/action_model': 8.117801207942528e-05, 'epoch': 7.65}) +[2026-06-20 17:25:10,188][starVLA.training.train_starvla][INFO] - Step 4140, Loss: {'train/loss': 0.05471884831786156, 'train/grad_norm_pre_clip': 5.781394004821777, 'timing/data': 0.011859431047923863, 'timing/model': 1.361290468950756, 'learning_rate/qwen_vl_interface': 8.115372883261648e-06, 'learning_rate/action_model': 8.115372883261647e-05, 'epoch': 7.66}) +[2026-06-20 17:25:14,741][starVLA.training.train_starvla][INFO] - Step 4141, Loss: {'train/loss': 0.1322294920682907, 'train/grad_norm_pre_clip': 10.781190872192383, 'timing/data': 0.011903380043804646, 'timing/model': 1.1793953989399597, 'learning_rate/qwen_vl_interface': 8.11294338073421e-06, 'learning_rate/action_model': 8.11294338073421e-05, 'epoch': 7.66}) +[2026-06-20 17:25:19,506][starVLA.training.train_starvla][INFO] - Step 4142, Loss: {'train/loss': 0.055545441806316376, 'train/grad_norm_pre_clip': 5.040861129760742, 'timing/data': 0.011874969932250679, 'timing/model': 1.1808762539876625, 'learning_rate/qwen_vl_interface': 8.110512701358892e-06, 'learning_rate/action_model': 8.110512701358891e-05, 'epoch': 7.66}) +[2026-06-20 17:25:24,667][starVLA.training.train_starvla][INFO] - Step 4143, Loss: {'train/loss': 0.5003852844238281, 'train/grad_norm_pre_clip': 14.91757583618164, 'timing/data': 0.019942143000662327, 'timing/model': 1.357417021994479, 'learning_rate/qwen_vl_interface': 8.108080846134852e-06, 'learning_rate/action_model': 8.108080846134852e-05, 'epoch': 7.66}) +[2026-06-20 17:25:30,246][starVLA.training.train_starvla][INFO] - Step 4144, Loss: {'train/loss': 0.16166429221630096, 'train/grad_norm_pre_clip': 8.020523071289062, 'timing/data': 0.020437101018615067, 'timing/model': 1.3619792949175462, 'learning_rate/qwen_vl_interface': 8.105647816061734e-06, 'learning_rate/action_model': 8.105647816061735e-05, 'epoch': 7.66}) +[2026-06-20 17:25:35,813][starVLA.training.train_starvla][INFO] - Step 4145, Loss: {'train/loss': 0.08656074851751328, 'train/grad_norm_pre_clip': 8.88486099243164, 'timing/data': 0.017882758984342217, 'timing/model': 1.6740022390149534, 'learning_rate/qwen_vl_interface': 8.103213612139667e-06, 'learning_rate/action_model': 8.103213612139667e-05, 'epoch': 7.67}) +[2026-06-20 17:25:41,293][starVLA.training.train_starvla][INFO] - Step 4146, Loss: {'train/loss': 0.13698992133140564, 'train/grad_norm_pre_clip': 12.875292778015137, 'timing/data': 0.01793542900122702, 'timing/model': 1.5958848100854084, 'learning_rate/qwen_vl_interface': 8.100778235369262e-06, 'learning_rate/action_model': 8.10077823536926e-05, 'epoch': 7.67}) +[2026-06-20 17:25:46,547][starVLA.training.train_starvla][INFO] - Step 4147, Loss: {'train/loss': 0.21006827056407928, 'train/grad_norm_pre_clip': 12.313643455505371, 'timing/data': 0.018242997932247818, 'timing/model': 1.3640756849199533, 'learning_rate/qwen_vl_interface': 8.098341686751605e-06, 'learning_rate/action_model': 8.098341686751605e-05, 'epoch': 7.67}) +[2026-06-20 17:25:52,111][starVLA.training.train_starvla][INFO] - Step 4148, Loss: {'train/loss': 0.08283507823944092, 'train/grad_norm_pre_clip': 16.005565643310547, 'timing/data': 0.01882352598477155, 'timing/model': 1.361916390946135, 'learning_rate/qwen_vl_interface': 8.095903967288273e-06, 'learning_rate/action_model': 8.095903967288272e-05, 'epoch': 7.67}) +[2026-06-20 17:25:57,674][starVLA.training.train_starvla][INFO] - Step 4149, Loss: {'train/loss': 0.01014044787734747, 'train/grad_norm_pre_clip': 1.333616852760315, 'timing/data': 0.017988178064115345, 'timing/model': 1.3552312640240416, 'learning_rate/qwen_vl_interface': 8.093465077981318e-06, 'learning_rate/action_model': 8.093465077981318e-05, 'epoch': 7.67}) +[2026-06-20 17:26:03,074][starVLA.training.train_starvla][INFO] - Step 4150, Loss: {'train/loss': 0.14129126071929932, 'train/grad_norm_pre_clip': 8.027488708496094, 'timing/data': 0.018510916037485003, 'timing/model': 1.190405016997829, 'learning_rate/qwen_vl_interface': 8.091025019833276e-06, 'learning_rate/action_model': 8.091025019833276e-05, 'epoch': 7.67}) +[2026-06-20 17:26:07,872][starVLA.training.train_starvla][INFO] - Step 4151, Loss: {'train/loss': 0.02462298609316349, 'train/grad_norm_pre_clip': 5.735464096069336, 'timing/data': 0.0119972990360111, 'timing/model': 1.4263154700165614, 'learning_rate/qwen_vl_interface': 8.088583793847165e-06, 'learning_rate/action_model': 8.088583793847164e-05, 'epoch': 7.68}) +[2026-06-20 17:26:12,436][starVLA.training.train_starvla][INFO] - Step 4152, Loss: {'train/loss': 0.1709427386522293, 'train/grad_norm_pre_clip': 23.564697265625, 'timing/data': 0.01196257898118347, 'timing/model': 1.1789011399960145, 'learning_rate/qwen_vl_interface': 8.086141401026478e-06, 'learning_rate/action_model': 8.086141401026478e-05, 'epoch': 7.68}) +[2026-06-20 17:26:17,232][starVLA.training.train_starvla][INFO] - Step 4153, Loss: {'train/loss': 0.07409193366765976, 'train/grad_norm_pre_clip': 11.239871978759766, 'timing/data': 0.0122315080370754, 'timing/model': 1.17732337792404, 'learning_rate/qwen_vl_interface': 8.083697842375193e-06, 'learning_rate/action_model': 8.083697842375193e-05, 'epoch': 7.68}) +[2026-06-20 17:26:22,023][starVLA.training.train_starvla][INFO] - Step 4154, Loss: {'train/loss': 0.00871679000556469, 'train/grad_norm_pre_clip': 1.343447208404541, 'timing/data': 0.011909408960491419, 'timing/model': 1.185442692018114, 'learning_rate/qwen_vl_interface': 8.081253118897762e-06, 'learning_rate/action_model': 8.081253118897761e-05, 'epoch': 7.68}) +[2026-06-20 17:26:26,831][starVLA.training.train_starvla][INFO] - Step 4155, Loss: {'train/loss': 0.01023709774017334, 'train/grad_norm_pre_clip': 0.8598842024803162, 'timing/data': 0.011827428941614926, 'timing/model': 1.1929709980031475, 'learning_rate/qwen_vl_interface': 8.07880723159912e-06, 'learning_rate/action_model': 8.07880723159912e-05, 'epoch': 7.68}) +[2026-06-20 17:26:31,571][starVLA.training.train_starvla][INFO] - Step 4156, Loss: {'train/loss': 0.13320612907409668, 'train/grad_norm_pre_clip': 16.731231689453125, 'timing/data': 0.011850678944028914, 'timing/model': 1.1908606280339882, 'learning_rate/qwen_vl_interface': 8.076360181484676e-06, 'learning_rate/action_model': 8.076360181484676e-05, 'epoch': 7.69}) +[2026-06-20 17:26:36,584][starVLA.training.train_starvla][INFO] - Step 4157, Loss: {'train/loss': 0.16287925839424133, 'train/grad_norm_pre_clip': 10.12519359588623, 'timing/data': 0.011862128973007202, 'timing/model': 1.5051023710984737, 'learning_rate/qwen_vl_interface': 8.073911969560326e-06, 'learning_rate/action_model': 8.073911969560326e-05, 'epoch': 7.69}) +[2026-06-20 17:26:41,654][starVLA.training.train_starvla][INFO] - Step 4158, Loss: {'train/loss': 0.21007205545902252, 'train/grad_norm_pre_clip': 18.860549926757812, 'timing/data': 0.012051988043822348, 'timing/model': 1.3084383939858526, 'learning_rate/qwen_vl_interface': 8.071462596832432e-06, 'learning_rate/action_model': 8.071462596832432e-05, 'epoch': 7.69}) +[2026-06-20 17:26:46,987][starVLA.training.train_starvla][INFO] - Step 4159, Loss: {'train/loss': 0.17699038982391357, 'train/grad_norm_pre_clip': 14.406793594360352, 'timing/data': 0.01191321900114417, 'timing/model': 1.301082529942505, 'learning_rate/qwen_vl_interface': 8.06901206430784e-06, 'learning_rate/action_model': 8.06901206430784e-05, 'epoch': 7.69}) +[2026-06-20 17:26:52,335][starVLA.training.train_starvla][INFO] - Step 4160, Loss: {'train/loss': 0.013205934315919876, 'train/grad_norm_pre_clip': 2.6180338859558105, 'timing/data': 0.013638313044793904, 'timing/model': 1.3216930880444124, 'learning_rate/qwen_vl_interface': 8.066560372993873e-06, 'learning_rate/action_model': 8.066560372993874e-05, 'epoch': 7.69}) +[2026-06-20 17:26:57,695][starVLA.training.train_starvla][INFO] - Step 4161, Loss: {'train/loss': 0.08438654989004135, 'train/grad_norm_pre_clip': 9.592870712280273, 'timing/data': 0.013408903032541275, 'timing/model': 1.3214435609988868, 'learning_rate/qwen_vl_interface': 8.064107523898327e-06, 'learning_rate/action_model': 8.064107523898326e-05, 'epoch': 7.7}) +[2026-06-20 17:27:03,086][starVLA.training.train_starvla][INFO] - Step 4162, Loss: {'train/loss': 0.004997638054192066, 'train/grad_norm_pre_clip': 0.35865235328674316, 'timing/data': 0.011969347950071096, 'timing/model': 1.3139516579685733, 'learning_rate/qwen_vl_interface': 8.061653518029476e-06, 'learning_rate/action_model': 8.061653518029476e-05, 'epoch': 7.7}) +[2026-06-20 17:27:08,344][starVLA.training.train_starvla][INFO] - Step 4163, Loss: {'train/loss': 0.172647625207901, 'train/grad_norm_pre_clip': 23.978893280029297, 'timing/data': 0.011945739039219916, 'timing/model': 1.4598101070150733, 'learning_rate/qwen_vl_interface': 8.059198356396069e-06, 'learning_rate/action_model': 8.059198356396068e-05, 'epoch': 7.7}) +[2026-06-20 17:27:12,899][starVLA.training.train_starvla][INFO] - Step 4164, Loss: {'train/loss': 0.1496727615594864, 'train/grad_norm_pre_clip': 6.583407402038574, 'timing/data': 0.011928358115255833, 'timing/model': 1.1750236100051552, 'learning_rate/qwen_vl_interface': 8.05674204000733e-06, 'learning_rate/action_model': 8.05674204000733e-05, 'epoch': 7.7}) +[2026-06-20 17:27:17,688][starVLA.training.train_starvla][INFO] - Step 4165, Loss: {'train/loss': 0.23099233210086823, 'train/grad_norm_pre_clip': 29.212303161621094, 'timing/data': 0.012285937089473009, 'timing/model': 1.1865359229268506, 'learning_rate/qwen_vl_interface': 8.05428456987296e-06, 'learning_rate/action_model': 8.05428456987296e-05, 'epoch': 7.7}) +[2026-06-20 17:27:22,484][starVLA.training.train_starvla][INFO] - Step 4166, Loss: {'train/loss': 0.09588120132684708, 'train/grad_norm_pre_clip': 14.921058654785156, 'timing/data': 0.012138567049987614, 'timing/model': 1.1768178500933573, 'learning_rate/qwen_vl_interface': 8.05182594700313e-06, 'learning_rate/action_model': 8.05182594700313e-05, 'epoch': 7.7}) +[2026-06-20 18:46:54,936][starVLA.training.train_starvla][INFO] - VLA Training :: Warming Up +[2026-06-20 18:46:54,949][starVLA.training.train_starvla][INFO] - ✅ Configuration wrapped for access tracking +[2026-06-20 18:46:56,607][starVLA.training.train_starvla][INFO] - Creating VLA Dataset with Mixture `flappy_mixed_latency_train__bridge` +[2026-06-20 18:46:57,280][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Behavior (data_config): ['BEHAVIOR_challenge'] +[2026-06-20 18:46:57,280][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from DOMINO: ['robotwin'] +[2026-06-20 18:46:57,280][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from DOMINO (data_config): [] +[2026-06-20 18:46:57,280][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from DOMINO (data_config): ['domino', 'domino_clean', 'domino_random', 'domino_cotrain'] +[2026-06-20 18:46:57,280][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Franka: ['custom_robot_config', 'demo_sim_franka_delta_joints', 'SO101'] +[2026-06-20 18:46:57,280][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Franka (data_config): [] +[2026-06-20 18:46:57,280][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Franka (data_config): ['custom_dataset', 'custom_dataset_2', 'demo_sim_pick_place', 'SO101_pick'] +[2026-06-20 18:46:57,280][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from LIBERO: ['libero_franka'] +[2026-06-20 18:46:57,280][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from LIBERO (data_config): [] +[2026-06-20 18:46:57,280][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from LIBERO (data_config): ['libero_all', 'libero_goal', 'multi_robot'] +[2026-06-20 18:46:57,281][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from RoboChallenge_table30v2: ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-20 18:46:57,281][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from RoboChallenge_table30v2 (data_config): ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-20 18:46:57,281][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from RoboChallenge_table30v2 (data_config): ['robochallenge_table30v2_shred_paper', 'robochallenge_table30v2_ur5_all', 'robochallenge_table30v2_arx5_all', 'robochallenge_table30v2_dosw1_all'] +[2026-06-20 18:46:57,281][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_365: ['panda_omron_robocasa365'] +[2026-06-20 18:46:57,281][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_365 (data_config): [] +[2026-06-20 18:46:57,281][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_365 (data_config): ['robocasa365_open_drawer_target_human', 'robocasa365_atomic_target_human_all', 'robocasa365_composite_target_human_all', 'robocasa365_target_human_all'] +[2026-06-20 18:46:57,281][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_tabletop: ['fourier_gr1_arms_waist'] +[2026-06-20 18:46:57,282][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_tabletop (data_config): [] +[2026-06-20 18:46:57,282][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_tabletop (data_config): ['fourier_gr1_unified_1000'] +[2026-06-20 18:46:57,282][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robotwin: ['robotwin', 'robotwin50', 'arx_x5'] +[2026-06-20 18:46:57,282][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robotwin (data_config): [] +[2026-06-20 18:46:57,282][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robotwin (data_config): ['robotwin_all', 'robotwin_all_50', 'robotwin', 'robotwin_task1', 'robotwin_task2', 'arx_x5'] +[2026-06-20 18:46:57,282][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from SimplerEnv: ['oxe_droid', 'oxe_bridge', 'oxe_rt1'] +[2026-06-20 18:46:57,282][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from SimplerEnv (data_config): [] +[2026-06-20 18:46:57,282][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from SimplerEnv (data_config): ['bridge', 'bridge_rt_1'] +[2026-06-20 18:46:57,283][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from VLA-Arena: ['vla_arena_franka'] +[2026-06-20 18:46:57,283][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from VLA-Arena (data_config): [] +[2026-06-20 18:46:57,283][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from VLA-Arena (data_config): ['vla_arena_L0_S', 'vla_arena_L0_M', 'vla_arena_L0_L'] +[2026-06-20 18:46:57,283][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from rl_games: ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-20 18:46:57,283][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from rl_games (data_config): ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-20 18:46:57,283][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from rl_games (data_config): ['flappy_train', 'flappy_train__bridge', 'flappy_mixed_latency_train', 'flappy_mixed_latency_train__bridge', 'demon_attack_train', 'demon_attack_train__bridge', 'demon_attack_mixed_latency_train', 'demon_attack_mixed_latency_train__bridge', 'deadly_corridor_train', 'deadly_corridor_train__bridge', 'deadly_corridor_mixed_latency_train', 'deadly_corridor_mixed_latency_train__bridge'] +[2026-06-20 18:46:57,355][starVLA.training.train_starvla][INFO] - Creating VLA Eval Dataset with Mixture `flappy_mixed_latency_train__bridge__val` +[2026-06-20 18:46:57,412][starVLA.training.train_starvla][INFO] - 📝 Full config saved at /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/config.full.yaml +[2026-06-20 18:46:57,419][starVLA.training.train_starvla][INFO] - 📊 Accessed config snapshot saved at /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/config.yaml +[2026-06-20 18:46:57,419][starVLA.training.train_starvla][INFO] - Will resume full training state from checkpoint: /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_4100_state, steps: 4100 +[2026-06-20 18:47:01,071][accelerate.accelerator][INFO] - Loading states from /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_4100_state +[2026-06-20 18:47:02,767][accelerate.checkpointing][INFO] - All model weights loaded successfully +[2026-06-20 18:47:11,672][accelerate.checkpointing][INFO] - All optimizer states loaded successfully +[2026-06-20 18:47:11,672][accelerate.checkpointing][INFO] - All scheduler states loaded successfully +[2026-06-20 18:47:11,672][accelerate.checkpointing][INFO] - All dataloader sampler states loaded successfully +[2026-06-20 18:47:11,675][accelerate.checkpointing][INFO] - All random states loaded successfully +[2026-06-20 18:47:12,464][accelerate.accelerator][INFO] - Loading in 0 custom states +[2026-06-20 18:47:14,154][starVLA.training.train_starvla][INFO] - ***** Training Configuration ***** +[2026-06-20 18:47:14,154][starVLA.training.train_starvla][INFO] - Total optimization steps = 5000 +[2026-06-20 18:47:14,155][starVLA.training.train_starvla][INFO] - Per device batch size = 32 +[2026-06-20 18:47:14,155][starVLA.training.train_starvla][INFO] - Gradient accumulation steps = 4 +[2026-06-20 18:47:14,155][starVLA.training.train_starvla][INFO] - Total batch size = 128 +[2026-06-20 18:47:14,883][starVLA.training.train_starvla][INFO] - Latency curriculum phase 4 at step 4100: active_latencies=[4], dataset_size=69213 +[2026-06-20 18:47:37,215][starVLA.training.train_starvla][INFO] - Step 4101, Loss: {'train/loss': 0.042002491652965546, 'train/grad_norm_pre_clip': 4.852111339569092, 'timing/data': 0.007238966994918883, 'timing/model': 0.911050351918675, 'learning_rate/qwen_vl_interface': 1.0000000000000001e-07, 'learning_rate/action_model': 1.0000000000000002e-06, 'epoch': 7.58}) +[2026-06-20 18:47:40,600][starVLA.training.train_starvla][INFO] - Step 4102, Loss: {'train/loss': 0.024595675989985466, 'train/grad_norm_pre_clip': 4.69415283203125, 'timing/data': 0.006365709938108921, 'timing/model': 0.8876961739733815, 'learning_rate/qwen_vl_interface': 2.0000000000000002e-07, 'learning_rate/action_model': 2.0000000000000003e-06, 'epoch': 7.59}) +[2026-06-20 18:47:43,992][starVLA.training.train_starvla][INFO] - Step 4103, Loss: {'train/loss': 0.06378281861543655, 'train/grad_norm_pre_clip': 13.65320110321045, 'timing/data': 0.005767991999164224, 'timing/model': 0.8886117800138891, 'learning_rate/qwen_vl_interface': 3.0000000000000004e-07, 'learning_rate/action_model': 3e-06, 'epoch': 7.59}) +[2026-06-20 18:47:47,405][starVLA.training.train_starvla][INFO] - Step 4104, Loss: {'train/loss': 0.44315871596336365, 'train/grad_norm_pre_clip': 33.7631721496582, 'timing/data': 0.0073224169900640845, 'timing/model': 0.9026189050637186, 'learning_rate/qwen_vl_interface': 4.0000000000000003e-07, 'learning_rate/action_model': 4.000000000000001e-06, 'epoch': 7.59}) +[2026-06-20 18:47:50,806][starVLA.training.train_starvla][INFO] - Step 4105, Loss: {'train/loss': 0.06133782118558884, 'train/grad_norm_pre_clip': 8.898704528808594, 'timing/data': 0.005630452069453895, 'timing/model': 0.8914447099668905, 'learning_rate/qwen_vl_interface': 5.000000000000001e-07, 'learning_rate/action_model': 5e-06, 'epoch': 7.59}) +[2026-06-20 18:47:54,211][starVLA.training.train_starvla][INFO] - Step 4106, Loss: {'train/loss': 0.23095273971557617, 'train/grad_norm_pre_clip': 12.241000175476074, 'timing/data': 0.005726201925426722, 'timing/model': 0.8918271770235151, 'learning_rate/qwen_vl_interface': 6.000000000000001e-07, 'learning_rate/action_model': 6e-06, 'epoch': 7.59}) +[2026-06-20 18:47:57,632][starVLA.training.train_starvla][INFO] - Step 4107, Loss: {'train/loss': 0.31703704595565796, 'train/grad_norm_pre_clip': 16.880983352661133, 'timing/data': 0.0055462130112573504, 'timing/model': 0.8924072349909693, 'learning_rate/qwen_vl_interface': 7.000000000000001e-07, 'learning_rate/action_model': 7.000000000000001e-06, 'epoch': 7.6}) +[2026-06-20 18:48:01,043][starVLA.training.train_starvla][INFO] - Step 4108, Loss: {'train/loss': 0.048230599611997604, 'train/grad_norm_pre_clip': 8.32318115234375, 'timing/data': 0.005398403969593346, 'timing/model': 0.8961920610163361, 'learning_rate/qwen_vl_interface': 8.000000000000001e-07, 'learning_rate/action_model': 8.000000000000001e-06, 'epoch': 7.6}) +[2026-06-20 18:48:04,615][starVLA.training.train_starvla][INFO] - Step 4109, Loss: {'train/loss': 0.1815768927335739, 'train/grad_norm_pre_clip': 14.667669296264648, 'timing/data': 0.005601472919806838, 'timing/model': 0.8940077479928732, 'learning_rate/qwen_vl_interface': 9.000000000000001e-07, 'learning_rate/action_model': 9e-06, 'epoch': 7.6}) +[2026-06-20 18:48:08,043][starVLA.training.train_starvla][INFO] - Step 4110, Loss: {'train/loss': 0.03080054372549057, 'train/grad_norm_pre_clip': 4.268087863922119, 'timing/data': 0.005769571987912059, 'timing/model': 0.894186355988495, 'learning_rate/qwen_vl_interface': 1.0000000000000002e-06, 'learning_rate/action_model': 1e-05, 'epoch': 7.6}) +[2026-06-20 18:48:11,460][starVLA.training.train_starvla][INFO] - Step 4111, Loss: {'train/loss': 0.010333732701838017, 'train/grad_norm_pre_clip': 0.8350149989128113, 'timing/data': 0.005652921972796321, 'timing/model': 0.8949560329783708, 'learning_rate/qwen_vl_interface': 1.1e-06, 'learning_rate/action_model': 1.1000000000000001e-05, 'epoch': 7.6}) +[2026-06-20 18:48:14,879][starVLA.training.train_starvla][INFO] - Step 4112, Loss: {'train/loss': 0.0809863954782486, 'train/grad_norm_pre_clip': 21.41415023803711, 'timing/data': 0.006573049002327025, 'timing/model': 0.896367356996052, 'learning_rate/qwen_vl_interface': 1.2000000000000002e-06, 'learning_rate/action_model': 1.2e-05, 'epoch': 7.6}) +[2026-06-20 18:48:18,305][starVLA.training.train_starvla][INFO] - Step 4113, Loss: {'train/loss': 0.06578502058982849, 'train/grad_norm_pre_clip': 8.015823364257812, 'timing/data': 0.005646912031807005, 'timing/model': 0.9009804719826207, 'learning_rate/qwen_vl_interface': 1.3e-06, 'learning_rate/action_model': 1.3000000000000001e-05, 'epoch': 7.61}) +[2026-06-20 18:48:21,729][starVLA.training.train_starvla][INFO] - Step 4114, Loss: {'train/loss': 0.13169194757938385, 'train/grad_norm_pre_clip': 11.90884017944336, 'timing/data': 0.0054856520146131516, 'timing/model': 0.8972298020962626, 'learning_rate/qwen_vl_interface': 1.4000000000000001e-06, 'learning_rate/action_model': 1.4000000000000001e-05, 'epoch': 7.61}) +[2026-06-20 18:48:25,144][starVLA.training.train_starvla][INFO] - Step 4115, Loss: {'train/loss': 0.15505149960517883, 'train/grad_norm_pre_clip': 8.084786415100098, 'timing/data': 0.005662171985022724, 'timing/model': 0.8943065409548581, 'learning_rate/qwen_vl_interface': 1.5e-06, 'learning_rate/action_model': 1.5e-05, 'epoch': 7.61}) +[2026-06-20 18:48:28,582][starVLA.training.train_starvla][INFO] - Step 4116, Loss: {'train/loss': 0.15994104743003845, 'train/grad_norm_pre_clip': 11.471158027648926, 'timing/data': 0.005510512972250581, 'timing/model': 0.9129798110807315, 'learning_rate/qwen_vl_interface': 1.6000000000000001e-06, 'learning_rate/action_model': 1.6000000000000003e-05, 'epoch': 7.61}) +[2026-06-20 18:48:32,031][starVLA.training.train_starvla][INFO] - Step 4117, Loss: {'train/loss': 0.016557427123188972, 'train/grad_norm_pre_clip': 1.8579190969467163, 'timing/data': 0.005889790947549045, 'timing/model': 0.9015744370408356, 'learning_rate/qwen_vl_interface': 1.7000000000000002e-06, 'learning_rate/action_model': 1.7000000000000003e-05, 'epoch': 7.61}) +[2026-06-20 18:48:35,575][starVLA.training.train_starvla][INFO] - Step 4118, Loss: {'train/loss': 0.008199024014174938, 'train/grad_norm_pre_clip': 0.8683929443359375, 'timing/data': 0.006932417978532612, 'timing/model': 0.9447099110111594, 'learning_rate/qwen_vl_interface': 1.8000000000000001e-06, 'learning_rate/action_model': 1.8e-05, 'epoch': 7.62}) +[2026-06-20 18:48:39,126][starVLA.training.train_starvla][INFO] - Step 4119, Loss: {'train/loss': 0.045159682631492615, 'train/grad_norm_pre_clip': 6.5505290031433105, 'timing/data': 0.005939672002568841, 'timing/model': 0.9013548460789025, 'learning_rate/qwen_vl_interface': 1.9000000000000002e-06, 'learning_rate/action_model': 1.9e-05, 'epoch': 7.62}) +[2026-06-20 18:48:42,606][starVLA.training.train_starvla][INFO] - Step 4120, Loss: {'train/loss': 0.04040538892149925, 'train/grad_norm_pre_clip': 4.249044895172119, 'timing/data': 0.0055259529035538435, 'timing/model': 0.8970066679175943, 'learning_rate/qwen_vl_interface': 2.0000000000000003e-06, 'learning_rate/action_model': 2e-05, 'epoch': 7.62}) +[2026-06-20 18:48:46,086][starVLA.training.train_starvla][INFO] - Step 4121, Loss: {'train/loss': 0.15375691652297974, 'train/grad_norm_pre_clip': 13.761232376098633, 'timing/data': 0.005334873916581273, 'timing/model': 0.915392930037342, 'learning_rate/qwen_vl_interface': 2.1000000000000002e-06, 'learning_rate/action_model': 2.1e-05, 'epoch': 7.62}) +[2026-06-20 18:48:49,524][starVLA.training.train_starvla][INFO] - Step 4122, Loss: {'train/loss': 0.06722486764192581, 'train/grad_norm_pre_clip': 6.734430313110352, 'timing/data': 0.00559897196944803, 'timing/model': 0.9000431470340118, 'learning_rate/qwen_vl_interface': 2.2e-06, 'learning_rate/action_model': 2.2000000000000003e-05, 'epoch': 7.62}) +[2026-06-20 18:48:52,964][starVLA.training.train_starvla][INFO] - Step 4123, Loss: {'train/loss': 0.006412267219275236, 'train/grad_norm_pre_clip': 0.4297330379486084, 'timing/data': 0.005725672002881765, 'timing/model': 0.9008371540112421, 'learning_rate/qwen_vl_interface': 2.3000000000000004e-06, 'learning_rate/action_model': 2.3000000000000003e-05, 'epoch': 7.62}) +[2026-06-20 18:48:56,406][starVLA.training.train_starvla][INFO] - Step 4124, Loss: {'train/loss': 0.06792163848876953, 'train/grad_norm_pre_clip': 6.578848361968994, 'timing/data': 0.00574453198350966, 'timing/model': 0.90096849296242, 'learning_rate/qwen_vl_interface': 2.4000000000000003e-06, 'learning_rate/action_model': 2.4e-05, 'epoch': 7.63}) +[2026-06-20 18:48:59,854][starVLA.training.train_starvla][INFO] - Step 4125, Loss: {'train/loss': 0.030794013291597366, 'train/grad_norm_pre_clip': 8.598726272583008, 'timing/data': 0.005522512947209179, 'timing/model': 0.9015955099603161, 'learning_rate/qwen_vl_interface': 2.5e-06, 'learning_rate/action_model': 2.5e-05, 'epoch': 7.63}) +[2026-06-20 18:49:03,491][starVLA.training.train_starvla][INFO] - Step 4126, Loss: {'train/loss': 0.11646687239408493, 'train/grad_norm_pre_clip': 10.759255409240723, 'timing/data': 0.005995861021801829, 'timing/model': 0.9018492980394512, 'learning_rate/qwen_vl_interface': 2.6e-06, 'learning_rate/action_model': 2.6000000000000002e-05, 'epoch': 7.63}) +[2026-06-20 18:49:07,118][starVLA.training.train_starvla][INFO] - Step 4127, Loss: {'train/loss': 0.016062144190073013, 'train/grad_norm_pre_clip': 1.5395570993423462, 'timing/data': 0.005727902054786682, 'timing/model': 0.9701905628899112, 'learning_rate/qwen_vl_interface': 2.7000000000000004e-06, 'learning_rate/action_model': 2.7000000000000002e-05, 'epoch': 7.63}) +[2026-06-20 18:49:10,679][starVLA.training.train_starvla][INFO] - Step 4128, Loss: {'train/loss': 0.03716624155640602, 'train/grad_norm_pre_clip': 8.826593399047852, 'timing/data': 0.007450627046637237, 'timing/model': 0.968571987003088, 'learning_rate/qwen_vl_interface': 2.8000000000000003e-06, 'learning_rate/action_model': 2.8000000000000003e-05, 'epoch': 7.63}) +[2026-06-20 18:49:14,250][starVLA.training.train_starvla][INFO] - Step 4129, Loss: {'train/loss': 0.23649191856384277, 'train/grad_norm_pre_clip': 14.749635696411133, 'timing/data': 0.007826735032722354, 'timing/model': 0.9704413500148803, 'learning_rate/qwen_vl_interface': 2.9e-06, 'learning_rate/action_model': 2.9e-05, 'epoch': 7.64}) +[2026-06-20 18:49:17,810][starVLA.training.train_starvla][INFO] - Step 4130, Loss: {'train/loss': 0.08061852306127548, 'train/grad_norm_pre_clip': 10.274094581604004, 'timing/data': 0.009173161000944674, 'timing/model': 0.949177645961754, 'learning_rate/qwen_vl_interface': 3e-06, 'learning_rate/action_model': 3e-05, 'epoch': 7.64}) +[2026-06-20 18:49:21,562][starVLA.training.train_starvla][INFO] - Step 4131, Loss: {'train/loss': 0.09648538380861282, 'train/grad_norm_pre_clip': 11.924222946166992, 'timing/data': 0.008417754084803164, 'timing/model': 1.010475772083737, 'learning_rate/qwen_vl_interface': 3.1000000000000004e-06, 'learning_rate/action_model': 3.1e-05, 'epoch': 7.64}) +[2026-06-20 18:49:25,434][starVLA.training.train_starvla][INFO] - Step 4132, Loss: {'train/loss': 0.054217059165239334, 'train/grad_norm_pre_clip': 7.921416759490967, 'timing/data': 0.00891528103966266, 'timing/model': 1.0147920689778402, 'learning_rate/qwen_vl_interface': 3.2000000000000003e-06, 'learning_rate/action_model': 3.2000000000000005e-05, 'epoch': 7.64}) +[2026-06-20 18:49:29,285][starVLA.training.train_starvla][INFO] - Step 4133, Loss: {'train/loss': 0.20069393515586853, 'train/grad_norm_pre_clip': 16.217182159423828, 'timing/data': 0.009030701941810548, 'timing/model': 0.9949871290009469, 'learning_rate/qwen_vl_interface': 3.3000000000000006e-06, 'learning_rate/action_model': 3.3e-05, 'epoch': 7.64}) +[2026-06-20 18:49:32,747][starVLA.training.train_starvla][INFO] - Step 4134, Loss: {'train/loss': 0.23034578561782837, 'train/grad_norm_pre_clip': 14.555193901062012, 'timing/data': 0.0056332029635086656, 'timing/model': 0.9037567749619484, 'learning_rate/qwen_vl_interface': 3.4000000000000005e-06, 'learning_rate/action_model': 3.4000000000000007e-05, 'epoch': 7.65}) +[2026-06-20 18:49:36,491][starVLA.training.train_starvla][INFO] - Step 4135, Loss: {'train/loss': 0.11091618239879608, 'train/grad_norm_pre_clip': 13.173249244689941, 'timing/data': 0.008577492088079453, 'timing/model': 0.9380455770296976, 'learning_rate/qwen_vl_interface': 3.5e-06, 'learning_rate/action_model': 3.5e-05, 'epoch': 7.65}) +[2026-06-20 18:49:39,941][starVLA.training.train_starvla][INFO] - Step 4136, Loss: {'train/loss': 0.044672805815935135, 'train/grad_norm_pre_clip': 8.907252311706543, 'timing/data': 0.005565083003602922, 'timing/model': 0.9044506020145491, 'learning_rate/qwen_vl_interface': 3.6000000000000003e-06, 'learning_rate/action_model': 3.6e-05, 'epoch': 7.65}) +[2026-06-20 18:49:43,389][starVLA.training.train_starvla][INFO] - Step 4137, Loss: {'train/loss': 0.13183243572711945, 'train/grad_norm_pre_clip': 14.29029655456543, 'timing/data': 0.005551442038267851, 'timing/model': 0.9059220569906756, 'learning_rate/qwen_vl_interface': 3.7e-06, 'learning_rate/action_model': 3.7e-05, 'epoch': 7.65}) +[2026-06-20 18:49:47,100][starVLA.training.train_starvla][INFO] - Step 4138, Loss: {'train/loss': 0.32322102785110474, 'train/grad_norm_pre_clip': 14.44497013092041, 'timing/data': 0.006865308969281614, 'timing/model': 0.9536038449732587, 'learning_rate/qwen_vl_interface': 3.8000000000000005e-06, 'learning_rate/action_model': 3.8e-05, 'epoch': 7.65}) +[2026-06-20 18:49:50,723][starVLA.training.train_starvla][INFO] - Step 4139, Loss: {'train/loss': 0.011714160442352295, 'train/grad_norm_pre_clip': 1.3554472923278809, 'timing/data': 0.006645179004408419, 'timing/model': 0.9696314539760351, 'learning_rate/qwen_vl_interface': 3.900000000000001e-06, 'learning_rate/action_model': 3.9000000000000006e-05, 'epoch': 7.65}) +[2026-06-20 18:49:54,327][starVLA.training.train_starvla][INFO] - Step 4140, Loss: {'train/loss': 0.043789323419332504, 'train/grad_norm_pre_clip': 8.104541778564453, 'timing/data': 0.006093161064200103, 'timing/model': 0.9671556510729715, 'learning_rate/qwen_vl_interface': 4.000000000000001e-06, 'learning_rate/action_model': 4e-05, 'epoch': 7.66}) +[2026-06-20 18:49:57,967][starVLA.training.train_starvla][INFO] - Step 4141, Loss: {'train/loss': 0.0629868283867836, 'train/grad_norm_pre_clip': 8.822264671325684, 'timing/data': 0.0066723289201036096, 'timing/model': 0.9669197910698131, 'learning_rate/qwen_vl_interface': 4.1e-06, 'learning_rate/action_model': 4.1e-05, 'epoch': 7.66}) +[2026-06-20 18:50:01,473][starVLA.training.train_starvla][INFO] - Step 4142, Loss: {'train/loss': 0.14364540576934814, 'train/grad_norm_pre_clip': 8.835834503173828, 'timing/data': 0.0075169759802520275, 'timing/model': 0.9671707290690392, 'learning_rate/qwen_vl_interface': 4.2000000000000004e-06, 'learning_rate/action_model': 4.2e-05, 'epoch': 7.66}) +[2026-06-20 18:50:05,110][starVLA.training.train_starvla][INFO] - Step 4143, Loss: {'train/loss': 0.16614660620689392, 'train/grad_norm_pre_clip': 15.489913940429688, 'timing/data': 0.007862944970838726, 'timing/model': 0.9665845209965482, 'learning_rate/qwen_vl_interface': 4.3e-06, 'learning_rate/action_model': 4.3e-05, 'epoch': 7.66}) +[2026-06-20 18:50:08,582][starVLA.training.train_starvla][INFO] - Step 4144, Loss: {'train/loss': 0.09221582114696503, 'train/grad_norm_pre_clip': 8.906941413879395, 'timing/data': 0.0069709079107269645, 'timing/model': 0.9169467060128227, 'learning_rate/qwen_vl_interface': 4.4e-06, 'learning_rate/action_model': 4.4000000000000006e-05, 'epoch': 7.66}) +[2026-06-20 18:50:12,039][starVLA.training.train_starvla][INFO] - Step 4145, Loss: {'train/loss': 0.1080770194530487, 'train/grad_norm_pre_clip': 12.682392120361328, 'timing/data': 0.005701772985048592, 'timing/model': 0.9039543860126287, 'learning_rate/qwen_vl_interface': 4.5e-06, 'learning_rate/action_model': 4.5e-05, 'epoch': 7.67}) +[2026-06-20 18:50:15,489][starVLA.training.train_starvla][INFO] - Step 4146, Loss: {'train/loss': 0.14531037211418152, 'train/grad_norm_pre_clip': 6.808502197265625, 'timing/data': 0.005629622028209269, 'timing/model': 0.9043398250360042, 'learning_rate/qwen_vl_interface': 4.600000000000001e-06, 'learning_rate/action_model': 4.600000000000001e-05, 'epoch': 7.67}) +[2026-06-20 18:50:18,945][starVLA.training.train_starvla][INFO] - Step 4147, Loss: {'train/loss': 0.34438011050224304, 'train/grad_norm_pre_clip': 9.170127868652344, 'timing/data': 0.0056988210417330265, 'timing/model': 0.904202553909272, 'learning_rate/qwen_vl_interface': 4.7e-06, 'learning_rate/action_model': 4.7e-05, 'epoch': 7.67}) +[2026-06-20 18:50:22,492][starVLA.training.train_starvla][INFO] - Step 4148, Loss: {'train/loss': 0.026312844827771187, 'train/grad_norm_pre_clip': 8.620192527770996, 'timing/data': 0.007530107046477497, 'timing/model': 0.9669991140253842, 'learning_rate/qwen_vl_interface': 4.800000000000001e-06, 'learning_rate/action_model': 4.8e-05, 'epoch': 7.67}) +[2026-06-20 18:50:26,180][starVLA.training.train_starvla][INFO] - Step 4149, Loss: {'train/loss': 0.20473043620586395, 'train/grad_norm_pre_clip': 10.267717361450195, 'timing/data': 0.0064641800709068775, 'timing/model': 0.9967615600908175, 'learning_rate/qwen_vl_interface': 4.9000000000000005e-06, 'learning_rate/action_model': 4.9e-05, 'epoch': 7.67}) +[2026-06-20 18:50:29,823][starVLA.training.train_starvla][INFO] - Step 4150, Loss: {'train/loss': 0.022501716390252113, 'train/grad_norm_pre_clip': 3.7598421573638916, 'timing/data': 0.007442875998094678, 'timing/model': 0.9681702699745074, 'learning_rate/qwen_vl_interface': 5e-06, 'learning_rate/action_model': 5e-05, 'epoch': 7.67}) +[2026-06-20 18:50:33,360][starVLA.training.train_starvla][INFO] - Step 4151, Loss: {'train/loss': 0.08676803112030029, 'train/grad_norm_pre_clip': 8.119111061096191, 'timing/data': 0.005786301917396486, 'timing/model': 0.9489054000005126, 'learning_rate/qwen_vl_interface': 5.1e-06, 'learning_rate/action_model': 5.1000000000000006e-05, 'epoch': 7.68}) +[2026-06-20 18:50:36,967][starVLA.training.train_starvla][INFO] - Step 4152, Loss: {'train/loss': 0.16328659653663635, 'train/grad_norm_pre_clip': 7.523524284362793, 'timing/data': 0.00680810899939388, 'timing/model': 0.9501476649893448, 'learning_rate/qwen_vl_interface': 5.2e-06, 'learning_rate/action_model': 5.2000000000000004e-05, 'epoch': 7.68}) +[2026-06-20 18:50:40,527][starVLA.training.train_starvla][INFO] - Step 4153, Loss: {'train/loss': 0.07912076264619827, 'train/grad_norm_pre_clip': 13.170618057250977, 'timing/data': 0.005710261990316212, 'timing/model': 0.9025393649935722, 'learning_rate/qwen_vl_interface': 5.300000000000001e-06, 'learning_rate/action_model': 5.300000000000001e-05, 'epoch': 7.68}) +[2026-06-20 18:50:44,135][starVLA.training.train_starvla][INFO] - Step 4154, Loss: {'train/loss': 0.05030500888824463, 'train/grad_norm_pre_clip': 6.935454845428467, 'timing/data': 0.006441189092583954, 'timing/model': 0.9490832779556513, 'learning_rate/qwen_vl_interface': 5.400000000000001e-06, 'learning_rate/action_model': 5.4000000000000005e-05, 'epoch': 7.68}) +[2026-06-20 18:50:47,687][starVLA.training.train_starvla][INFO] - Step 4155, Loss: {'train/loss': 0.1413540095090866, 'train/grad_norm_pre_clip': 11.35657024383545, 'timing/data': 0.006639568950049579, 'timing/model': 0.9498091940768063, 'learning_rate/qwen_vl_interface': 5.500000000000001e-06, 'learning_rate/action_model': 5.500000000000001e-05, 'epoch': 7.68}) +[2026-06-20 18:50:51,251][starVLA.training.train_starvla][INFO] - Step 4156, Loss: {'train/loss': 0.19987858831882477, 'train/grad_norm_pre_clip': 5.502852439880371, 'timing/data': 0.006195281050167978, 'timing/model': 0.9500958930002525, 'learning_rate/qwen_vl_interface': 5.600000000000001e-06, 'learning_rate/action_model': 5.6000000000000006e-05, 'epoch': 7.69}) +[2026-06-20 18:50:54,810][starVLA.training.train_starvla][INFO] - Step 4157, Loss: {'train/loss': 0.08151187747716904, 'train/grad_norm_pre_clip': 6.795620441436768, 'timing/data': 0.005832092021591961, 'timing/model': 0.9020728030009195, 'learning_rate/qwen_vl_interface': 5.7e-06, 'learning_rate/action_model': 5.6999999999999996e-05, 'epoch': 7.69}) +[2026-06-20 18:50:58,370][starVLA.training.train_starvla][INFO] - Step 4158, Loss: {'train/loss': 0.07672020047903061, 'train/grad_norm_pre_clip': 5.215551853179932, 'timing/data': 0.005681262002326548, 'timing/model': 0.9023837619461119, 'learning_rate/qwen_vl_interface': 5.8e-06, 'learning_rate/action_model': 5.8e-05, 'epoch': 7.69}) +[2026-06-20 18:51:02,020][starVLA.training.train_starvla][INFO] - Step 4159, Loss: {'train/loss': 0.07382047176361084, 'train/grad_norm_pre_clip': 7.054515361785889, 'timing/data': 0.005870731081813574, 'timing/model': 0.9028108809143305, 'learning_rate/qwen_vl_interface': 5.9e-06, 'learning_rate/action_model': 5.9e-05, 'epoch': 7.69}) +[2026-06-20 18:51:05,579][starVLA.training.train_starvla][INFO] - Step 4160, Loss: {'train/loss': 0.14171431958675385, 'train/grad_norm_pre_clip': 10.453862190246582, 'timing/data': 0.006809409009292722, 'timing/model': 0.9492824419867247, 'learning_rate/qwen_vl_interface': 6e-06, 'learning_rate/action_model': 6e-05, 'epoch': 7.69}) +[2026-06-20 18:51:09,140][starVLA.training.train_starvla][INFO] - Step 4161, Loss: {'train/loss': 0.02884075976908207, 'train/grad_norm_pre_clip': 2.5855872631073, 'timing/data': 0.005608701962046325, 'timing/model': 0.9026514689903706, 'learning_rate/qwen_vl_interface': 6.1e-06, 'learning_rate/action_model': 6.1e-05, 'epoch': 7.7}) +[2026-06-20 18:51:13,013][starVLA.training.train_starvla][INFO] - Step 4162, Loss: {'train/loss': 0.031007805839180946, 'train/grad_norm_pre_clip': 4.473602294921875, 'timing/data': 0.007566767046228051, 'timing/model': 0.9662018269300461, 'learning_rate/qwen_vl_interface': 6.200000000000001e-06, 'learning_rate/action_model': 6.2e-05, 'epoch': 7.7}) +[2026-06-20 18:51:16,642][starVLA.training.train_starvla][INFO] - Step 4163, Loss: {'train/loss': 0.08077438920736313, 'train/grad_norm_pre_clip': 11.068790435791016, 'timing/data': 0.005707541946321726, 'timing/model': 0.9036834739381447, 'learning_rate/qwen_vl_interface': 6.300000000000001e-06, 'learning_rate/action_model': 6.3e-05, 'epoch': 7.7}) +[2026-06-20 18:51:20,241][starVLA.training.train_starvla][INFO] - Step 4164, Loss: {'train/loss': 0.16635572910308838, 'train/grad_norm_pre_clip': 12.505882263183594, 'timing/data': 0.005710612051188946, 'timing/model': 0.9036453039152548, 'learning_rate/qwen_vl_interface': 6.4000000000000006e-06, 'learning_rate/action_model': 6.400000000000001e-05, 'epoch': 7.7}) +[2026-06-20 18:51:23,887][starVLA.training.train_starvla][INFO] - Step 4165, Loss: {'train/loss': 0.15990306437015533, 'train/grad_norm_pre_clip': 14.439608573913574, 'timing/data': 0.008339473977684975, 'timing/model': 0.9755872059613466, 'learning_rate/qwen_vl_interface': 6.5000000000000004e-06, 'learning_rate/action_model': 6.500000000000001e-05, 'epoch': 7.7}) +[2026-06-20 18:51:27,555][starVLA.training.train_starvla][INFO] - Step 4166, Loss: {'train/loss': 0.08708523958921432, 'train/grad_norm_pre_clip': 12.968476295471191, 'timing/data': 0.006670788978226483, 'timing/model': 0.9531885960604995, 'learning_rate/qwen_vl_interface': 6.600000000000001e-06, 'learning_rate/action_model': 6.6e-05, 'epoch': 7.7}) +[2026-06-20 18:51:31,180][starVLA.training.train_starvla][INFO] - Step 4167, Loss: {'train/loss': 0.008035248145461082, 'train/grad_norm_pre_clip': 1.469359040260315, 'timing/data': 0.006911079050041735, 'timing/model': 0.952859906014055, 'learning_rate/qwen_vl_interface': 6.700000000000001e-06, 'learning_rate/action_model': 6.7e-05, 'epoch': 7.71}) +[2026-06-20 18:51:34,699][starVLA.training.train_starvla][INFO] - Step 4168, Loss: {'train/loss': 0.07406871765851974, 'train/grad_norm_pre_clip': 8.630583763122559, 'timing/data': 0.005586431943811476, 'timing/model': 0.9017820579465479, 'learning_rate/qwen_vl_interface': 6.800000000000001e-06, 'learning_rate/action_model': 6.800000000000001e-05, 'epoch': 7.71}) +[2026-06-20 18:51:38,254][starVLA.training.train_starvla][INFO] - Step 4169, Loss: {'train/loss': 0.006402935832738876, 'train/grad_norm_pre_clip': 0.42502808570861816, 'timing/data': 0.005578452022746205, 'timing/model': 0.9024837639881298, 'learning_rate/qwen_vl_interface': 6.9e-06, 'learning_rate/action_model': 6.9e-05, 'epoch': 7.71}) +[2026-06-20 18:51:41,860][starVLA.training.train_starvla][INFO] - Step 4170, Loss: {'train/loss': 0.023036407306790352, 'train/grad_norm_pre_clip': 2.7914931774139404, 'timing/data': 0.007104858057573438, 'timing/model': 0.9496781240450218, 'learning_rate/qwen_vl_interface': 7e-06, 'learning_rate/action_model': 7e-05, 'epoch': 7.71}) +[2026-06-20 18:51:45,361][starVLA.training.train_starvla][INFO] - Step 4171, Loss: {'train/loss': 0.09926943480968475, 'train/grad_norm_pre_clip': 10.618969917297363, 'timing/data': 0.00596521096304059, 'timing/model': 0.9051632949849591, 'learning_rate/qwen_vl_interface': 7.100000000000001e-06, 'learning_rate/action_model': 7.1e-05, 'epoch': 7.71}) +[2026-06-20 18:51:48,840][starVLA.training.train_starvla][INFO] - Step 4172, Loss: {'train/loss': 0.13990941643714905, 'train/grad_norm_pre_clip': 10.770100593566895, 'timing/data': 0.005592311965301633, 'timing/model': 0.9027294520055875, 'learning_rate/qwen_vl_interface': 7.2000000000000005e-06, 'learning_rate/action_model': 7.2e-05, 'epoch': 7.72}) +[2026-06-20 18:51:52,344][starVLA.training.train_starvla][INFO] - Step 4173, Loss: {'train/loss': 0.022430511191487312, 'train/grad_norm_pre_clip': 2.6600193977355957, 'timing/data': 0.00562243303284049, 'timing/model': 0.9193532379576936, 'learning_rate/qwen_vl_interface': 7.3e-06, 'learning_rate/action_model': 7.3e-05, 'epoch': 7.72}) +[2026-06-20 18:51:55,801][starVLA.training.train_starvla][INFO] - Step 4174, Loss: {'train/loss': 0.12184436619281769, 'train/grad_norm_pre_clip': 14.084668159484863, 'timing/data': 0.005701432004570961, 'timing/model': 0.9020361930597574, 'learning_rate/qwen_vl_interface': 7.4e-06, 'learning_rate/action_model': 7.4e-05, 'epoch': 7.72}) +[2026-06-20 18:51:59,298][starVLA.training.train_starvla][INFO] - Step 4175, Loss: {'train/loss': 0.021387942135334015, 'train/grad_norm_pre_clip': 6.202520847320557, 'timing/data': 0.006635149009525776, 'timing/model': 0.9202441739616916, 'learning_rate/qwen_vl_interface': 7.500000000000001e-06, 'learning_rate/action_model': 7.500000000000001e-05, 'epoch': 7.72}) +[2026-06-20 18:52:02,769][starVLA.training.train_starvla][INFO] - Step 4176, Loss: {'train/loss': 0.06315641105175018, 'train/grad_norm_pre_clip': 5.328614711761475, 'timing/data': 0.005474762991070747, 'timing/model': 0.9026242790278047, 'learning_rate/qwen_vl_interface': 7.600000000000001e-06, 'learning_rate/action_model': 7.6e-05, 'epoch': 7.72}) +[2026-06-20 18:52:06,264][starVLA.training.train_starvla][INFO] - Step 4177, Loss: {'train/loss': 0.010292247869074345, 'train/grad_norm_pre_clip': 4.626927852630615, 'timing/data': 0.00546018301974982, 'timing/model': 0.9029599779751152, 'learning_rate/qwen_vl_interface': 7.7e-06, 'learning_rate/action_model': 7.7e-05, 'epoch': 7.72}) +[2026-06-20 18:52:09,743][starVLA.training.train_starvla][INFO] - Step 4178, Loss: {'train/loss': 0.12330672889947891, 'train/grad_norm_pre_clip': 6.185184478759766, 'timing/data': 0.005559151992201805, 'timing/model': 0.9035943859489635, 'learning_rate/qwen_vl_interface': 7.800000000000002e-06, 'learning_rate/action_model': 7.800000000000001e-05, 'epoch': 7.73}) +[2026-06-20 18:52:13,277][starVLA.training.train_starvla][INFO] - Step 4179, Loss: {'train/loss': 0.03917239233851433, 'train/grad_norm_pre_clip': 6.844799995422363, 'timing/data': 0.006147261010482907, 'timing/model': 0.9349010550649837, 'learning_rate/qwen_vl_interface': 7.9e-06, 'learning_rate/action_model': 7.900000000000001e-05, 'epoch': 7.73}) +[2026-06-20 18:52:16,776][starVLA.training.train_starvla][INFO] - Step 4180, Loss: {'train/loss': 0.15576201677322388, 'train/grad_norm_pre_clip': 17.844905853271484, 'timing/data': 0.005504408967681229, 'timing/model': 0.9072876379359514, 'learning_rate/qwen_vl_interface': 8.000000000000001e-06, 'learning_rate/action_model': 8e-05, 'epoch': 7.73}) +[2026-06-20 18:52:20,240][starVLA.training.train_starvla][INFO] - Step 4181, Loss: {'train/loss': 0.18108470737934113, 'train/grad_norm_pre_clip': 14.562912940979004, 'timing/data': 0.005722354049794376, 'timing/model': 0.9052498560631648, 'learning_rate/qwen_vl_interface': 8.1e-06, 'learning_rate/action_model': 8.1e-05, 'epoch': 7.73}) +[2026-06-20 18:52:23,736][starVLA.training.train_starvla][INFO] - Step 4182, Loss: {'train/loss': 0.019204389303922653, 'train/grad_norm_pre_clip': 5.053343296051025, 'timing/data': 0.005597874987870455, 'timing/model': 0.9043048090534285, 'learning_rate/qwen_vl_interface': 8.2e-06, 'learning_rate/action_model': 8.2e-05, 'epoch': 7.73}) +[2026-06-20 18:52:27,382][starVLA.training.train_starvla][INFO] - Step 4183, Loss: {'train/loss': 0.15879425406455994, 'train/grad_norm_pre_clip': 14.617199897766113, 'timing/data': 0.006400499027222395, 'timing/model': 0.9136968589155003, 'learning_rate/qwen_vl_interface': 8.3e-06, 'learning_rate/action_model': 8.3e-05, 'epoch': 7.74}) +[2026-06-20 18:52:30,851][starVLA.training.train_starvla][INFO] - Step 4184, Loss: {'train/loss': 0.0504610612988472, 'train/grad_norm_pre_clip': 7.747836589813232, 'timing/data': 0.006180742057040334, 'timing/model': 0.9112368730129674, 'learning_rate/qwen_vl_interface': 8.400000000000001e-06, 'learning_rate/action_model': 8.4e-05, 'epoch': 7.74}) +[2026-06-20 18:52:34,322][starVLA.training.train_starvla][INFO] - Step 4185, Loss: {'train/loss': 0.37452632188796997, 'train/grad_norm_pre_clip': 22.004562377929688, 'timing/data': 0.005800573970191181, 'timing/model': 0.9153602740261704, 'learning_rate/qwen_vl_interface': 8.5e-06, 'learning_rate/action_model': 8.5e-05, 'epoch': 7.74}) +[2026-06-20 18:52:37,787][starVLA.training.train_starvla][INFO] - Step 4186, Loss: {'train/loss': 0.013497211039066315, 'train/grad_norm_pre_clip': 2.0883607864379883, 'timing/data': 0.005676086060702801, 'timing/model': 0.9070095040369779, 'learning_rate/qwen_vl_interface': 8.6e-06, 'learning_rate/action_model': 8.6e-05, 'epoch': 7.74}) +[2026-06-20 18:52:41,260][starVLA.training.train_starvla][INFO] - Step 4187, Loss: {'train/loss': 0.31371474266052246, 'train/grad_norm_pre_clip': 22.93962860107422, 'timing/data': 0.0056845450308173895, 'timing/model': 0.9132648679660633, 'learning_rate/qwen_vl_interface': 8.700000000000001e-06, 'learning_rate/action_model': 8.7e-05, 'epoch': 7.74}) +[2026-06-20 18:52:44,738][starVLA.training.train_starvla][INFO] - Step 4188, Loss: {'train/loss': 0.040390461683273315, 'train/grad_norm_pre_clip': 4.570314407348633, 'timing/data': 0.007550409995019436, 'timing/model': 0.9131065539550036, 'learning_rate/qwen_vl_interface': 8.8e-06, 'learning_rate/action_model': 8.800000000000001e-05, 'epoch': 7.75}) +[2026-06-20 18:52:48,347][starVLA.training.train_starvla][INFO] - Step 4189, Loss: {'train/loss': 0.05083953216671944, 'train/grad_norm_pre_clip': 8.385075569152832, 'timing/data': 0.007391802035272121, 'timing/model': 0.9339079349301755, 'learning_rate/qwen_vl_interface': 8.900000000000001e-06, 'learning_rate/action_model': 8.900000000000001e-05, 'epoch': 7.75}) +[2026-06-20 18:52:51,996][starVLA.training.train_starvla][INFO] - Step 4190, Loss: {'train/loss': 0.09199250489473343, 'train/grad_norm_pre_clip': 17.66351318359375, 'timing/data': 0.005648977006785572, 'timing/model': 0.9092708909884095, 'learning_rate/qwen_vl_interface': 9e-06, 'learning_rate/action_model': 9e-05, 'epoch': 7.75}) +[2026-06-20 18:52:55,590][starVLA.training.train_starvla][INFO] - Step 4191, Loss: {'train/loss': 0.04317507520318031, 'train/grad_norm_pre_clip': 8.351311683654785, 'timing/data': 0.009701163973659277, 'timing/model': 1.0295840239850804, 'learning_rate/qwen_vl_interface': 9.100000000000001e-06, 'learning_rate/action_model': 9.1e-05, 'epoch': 7.75}) +[2026-06-20 18:52:59,050][starVLA.training.train_starvla][INFO] - Step 4192, Loss: {'train/loss': 0.2998020350933075, 'train/grad_norm_pre_clip': 13.020573616027832, 'timing/data': 0.004613435012288392, 'timing/model': 0.904059918015264, 'learning_rate/qwen_vl_interface': 9.200000000000002e-06, 'learning_rate/action_model': 9.200000000000001e-05, 'epoch': 7.75}) +[2026-06-20 18:53:02,518][starVLA.training.train_starvla][INFO] - Step 4193, Loss: {'train/loss': 0.052164144814014435, 'train/grad_norm_pre_clip': 9.350491523742676, 'timing/data': 0.005603117053397, 'timing/model': 0.9072942760540172, 'learning_rate/qwen_vl_interface': 9.3e-06, 'learning_rate/action_model': 9.300000000000001e-05, 'epoch': 7.75}) +[2026-06-20 18:53:06,023][starVLA.training.train_starvla][INFO] - Step 4194, Loss: {'train/loss': 0.07241038233041763, 'train/grad_norm_pre_clip': 8.271814346313477, 'timing/data': 0.006070454022847116, 'timing/model': 0.9098846080014482, 'learning_rate/qwen_vl_interface': 9.4e-06, 'learning_rate/action_model': 9.4e-05, 'epoch': 7.76}) +[2026-06-20 18:53:09,506][starVLA.training.train_starvla][INFO] - Step 4195, Loss: {'train/loss': 0.010162020102143288, 'train/grad_norm_pre_clip': 1.0549877882003784, 'timing/data': 0.006464280071668327, 'timing/model': 0.9264462450519204, 'learning_rate/qwen_vl_interface': 9.5e-06, 'learning_rate/action_model': 9.5e-05, 'epoch': 7.76}) +[2026-06-20 18:53:13,020][starVLA.training.train_starvla][INFO] - Step 4196, Loss: {'train/loss': 0.012572117149829865, 'train/grad_norm_pre_clip': 4.55206298828125, 'timing/data': 0.0056357369758188725, 'timing/model': 0.9151190029224381, 'learning_rate/qwen_vl_interface': 9.600000000000001e-06, 'learning_rate/action_model': 9.6e-05, 'epoch': 7.76}) +[2026-06-20 18:53:16,477][starVLA.training.train_starvla][INFO] - Step 4197, Loss: {'train/loss': 0.03739141300320625, 'train/grad_norm_pre_clip': 10.362455368041992, 'timing/data': 0.005418209009803832, 'timing/model': 0.9061885440023616, 'learning_rate/qwen_vl_interface': 9.7e-06, 'learning_rate/action_model': 9.7e-05, 'epoch': 7.76}) +[2026-06-20 18:53:19,972][starVLA.training.train_starvla][INFO] - Step 4198, Loss: {'train/loss': 0.13449935615062714, 'train/grad_norm_pre_clip': 15.083688735961914, 'timing/data': 0.006753328954800963, 'timing/model': 0.9232521870872006, 'learning_rate/qwen_vl_interface': 9.800000000000001e-06, 'learning_rate/action_model': 9.8e-05, 'epoch': 7.76}) +[2026-06-20 18:53:23,446][starVLA.training.train_starvla][INFO] - Step 4199, Loss: {'train/loss': 0.39670032262802124, 'train/grad_norm_pre_clip': 21.38627052307129, 'timing/data': 0.005679667112417519, 'timing/model': 0.9042954520555213, 'learning_rate/qwen_vl_interface': 9.9e-06, 'learning_rate/action_model': 9.900000000000001e-05, 'epoch': 7.77}) +[2026-06-20 18:53:26,938][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 4200 +[2026-06-20 19:00:39,606][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 4200 +[2026-06-20 19:00:39,607][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 4200 +[2026-06-20 19:00:40,269][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_openvla/include -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_openvla/include -fPIC -c /tmp/tmpqfj4rbd7/test.c -o /tmp/tmpqfj4rbd7/test.o +[2026-06-20 19:00:40,312][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat /tmp/tmpqfj4rbd7/test.o -laio -o /tmp/tmpqfj4rbd7/a.out +[2026-06-20 19:00:40,708][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_openvla/include -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_openvla/include -fPIC -c /tmp/tmp_3ynxlo7/test.c -o /tmp/tmp_3ynxlo7/test.o +[2026-06-20 19:00:40,729][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat /tmp/tmp_3ynxlo7/test.o -L/usr/local/cuda -L/usr/local/cuda/lib64 -lcufile -o /tmp/tmp_3ynxlo7/a.out +[2026-06-20 19:00:40,775][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_openvla/include -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_openvla/include -fPIC -c /tmp/tmpyaurl1at/test.c -o /tmp/tmpyaurl1at/test.o +[2026-06-20 19:00:40,808][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat /tmp/tmpyaurl1at/test.o -laio -o /tmp/tmpyaurl1at/a.out +[2026-06-20 19:07:45,782][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 4200 +[2026-06-20 19:07:45,784][starVLA.training.train_starvla][INFO] - Step 4200, Loss: {'train/loss': 0.2101549655199051, 'train/grad_norm_pre_clip': 8.903695106506348, 'eval/loss': 0.8160579800605774, 'eval/latency_0/loss': 1.013348937034607, 'eval/latency_1/loss': 1.2714256048202515, 'eval/latency_2/loss': 1.1375675201416016, 'eval/latency_3/loss': 0.5241535305976868, 'eval/latency_4/loss': 0.09765734523534775, 'eval/flappy/loss': 0.8160579800605774, 'eval/flappy/latency_0/loss': 1.013348937034607, 'eval/flappy/latency_1/loss': 1.2714256048202515, 'eval/flappy/latency_2/loss': 1.1375675201416016, 'eval/flappy/latency_3/loss': 0.5241535305976868, 'eval/flappy/latency_4/loss': 0.09765734523534775, 'eval/flap_precision': 0.13513513513513514, 'eval/flap_recall': 0.10909090909090909, 'eval/flap_f1': 0.12072434607645877, 'eval/flap_rate_model': 0.06166666666666667, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.12072434607645875, 'eval/latency_0/flap_precision': 0.13513513513513514, 'eval/latency_0/flap_recall': 0.10909090909090909, 'eval/latency_0/flap_f1': 0.12072434607645877, 'eval/latency_0/flap_rate_model': 0.06166666666666667, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.12072434607645875, 'timing/data': 0.006127294036559761, 'timing/model': 0.923125421977602, 'learning_rate/qwen_vl_interface': 1e-05, 'learning_rate/action_model': 0.0001, 'epoch': 7.77}) +[2026-06-20 19:07:45,784][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_4200_state +[2026-06-20 19:07:45,788][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-20 19:07:53,807][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_4200_state/model.safetensors +[2026-06-20 19:08:08,154][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_4200_state/optimizer.bin +[2026-06-20 19:08:08,155][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_4200_state/sampler.bin +[2026-06-20 19:08:08,155][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_4200_state/sampler_1.bin +[2026-06-20 19:08:08,156][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_4200_state/random_states_0.pkl +[2026-06-20 19:08:10,492][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4200 +[2026-06-20 19:08:10,492][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-20 19:08:10,506][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-20 19:08:15,075][starVLA.training.train_starvla][INFO] - Step 4201, Loss: {'train/loss': 0.1214730367064476, 'train/grad_norm_pre_clip': 8.428349494934082, 'timing/data': 0.006463599973358214, 'timing/model': 1.3631545649841428, 'learning_rate/qwen_vl_interface': 9.999999023727212e-06, 'learning_rate/action_model': 9.99999902372721e-05, 'epoch': 7.77}) +[2026-06-20 19:08:19,777][starVLA.training.train_starvla][INFO] - Step 4202, Loss: {'train/loss': 0.23290826380252838, 'train/grad_norm_pre_clip': 15.49677848815918, 'timing/data': 0.007387346005998552, 'timing/model': 1.411437739036046, 'learning_rate/qwen_vl_interface': 9.999996094909247e-06, 'learning_rate/action_model': 9.999996094909248e-05, 'epoch': 7.77}) +[2026-06-20 19:08:24,307][starVLA.training.train_starvla][INFO] - Step 4203, Loss: {'train/loss': 0.17340964078903198, 'train/grad_norm_pre_clip': 14.94037914276123, 'timing/data': 0.006323771085590124, 'timing/model': 1.1285043460084125, 'learning_rate/qwen_vl_interface': 9.999991213547314e-06, 'learning_rate/action_model': 9.999991213547313e-05, 'epoch': 7.77}) +[2026-06-20 19:08:29,112][starVLA.training.train_starvla][INFO] - Step 4204, Loss: {'train/loss': 0.010635606944561005, 'train/grad_norm_pre_clip': 1.387198567390442, 'timing/data': 0.006375261000357568, 'timing/model': 1.201513524982147, 'learning_rate/qwen_vl_interface': 9.999984379643413e-06, 'learning_rate/action_model': 9.999984379643412e-05, 'epoch': 7.77}) +[2026-06-20 19:08:33,952][starVLA.training.train_starvla][INFO] - Step 4205, Loss: {'train/loss': 0.039103906601667404, 'train/grad_norm_pre_clip': 6.0308709144592285, 'timing/data': 0.0064191799610853195, 'timing/model': 1.1324816909618676, 'learning_rate/qwen_vl_interface': 9.999975593200357e-06, 'learning_rate/action_model': 9.999975593200357e-05, 'epoch': 7.78}) +[2026-06-20 19:08:38,683][starVLA.training.train_starvla][INFO] - Step 4206, Loss: {'train/loss': 0.14561258256435394, 'train/grad_norm_pre_clip': 12.153726577758789, 'timing/data': 0.006469379994086921, 'timing/model': 1.1368446070700884, 'learning_rate/qwen_vl_interface': 9.999964854221756e-06, 'learning_rate/action_model': 9.999964854221754e-05, 'epoch': 7.78}) +[2026-06-20 19:08:43,466][starVLA.training.train_starvla][INFO] - Step 4207, Loss: {'train/loss': 0.02017974480986595, 'train/grad_norm_pre_clip': 2.997077226638794, 'timing/data': 0.006394071038812399, 'timing/model': 1.4212325429543853, 'learning_rate/qwen_vl_interface': 9.999952162712027e-06, 'learning_rate/action_model': 9.999952162712026e-05, 'epoch': 7.78}) +[2026-06-20 19:08:48,000][starVLA.training.train_starvla][INFO] - Step 4208, Loss: {'train/loss': 0.13139685988426208, 'train/grad_norm_pre_clip': 14.33481216430664, 'timing/data': 0.007856454001739621, 'timing/model': 1.2377815439831465, 'learning_rate/qwen_vl_interface': 9.999937518676382e-06, 'learning_rate/action_model': 9.999937518676382e-05, 'epoch': 7.78}) +[2026-06-20 19:08:52,791][starVLA.training.train_starvla][INFO] - Step 4209, Loss: {'train/loss': 0.16599951684474945, 'train/grad_norm_pre_clip': 15.895955085754395, 'timing/data': 0.00626924098469317, 'timing/model': 1.1394245829433203, 'learning_rate/qwen_vl_interface': 9.999920922120846e-06, 'learning_rate/action_model': 9.999920922120846e-05, 'epoch': 7.78}) +[2026-06-20 19:08:57,631][starVLA.training.train_starvla][INFO] - Step 4210, Loss: {'train/loss': 0.13913069665431976, 'train/grad_norm_pre_clip': 8.753669738769531, 'timing/data': 0.006269801058806479, 'timing/model': 1.1411912209587172, 'learning_rate/qwen_vl_interface': 9.999902373052239e-06, 'learning_rate/action_model': 9.999902373052239e-05, 'epoch': 7.79}) +[2026-06-20 19:09:02,468][starVLA.training.train_starvla][INFO] - Step 4211, Loss: {'train/loss': 0.08443782478570938, 'train/grad_norm_pre_clip': 11.325142860412598, 'timing/data': 0.010324342991225421, 'timing/model': 1.2097367219394073, 'learning_rate/qwen_vl_interface': 9.999881871478188e-06, 'learning_rate/action_model': 9.999881871478187e-05, 'epoch': 7.79}) +[2026-06-20 19:09:07,244][starVLA.training.train_starvla][INFO] - Step 4212, Loss: {'train/loss': 0.014825735241174698, 'train/grad_norm_pre_clip': 2.5452187061309814, 'timing/data': 0.006290021003223956, 'timing/model': 1.1406681960215792, 'learning_rate/qwen_vl_interface': 9.999859417407115e-06, 'learning_rate/action_model': 9.999859417407115e-05, 'epoch': 7.79}) +[2026-06-20 19:09:11,993][starVLA.training.train_starvla][INFO] - Step 4213, Loss: {'train/loss': 0.01454242318868637, 'train/grad_norm_pre_clip': 1.722485899925232, 'timing/data': 0.006460270029492676, 'timing/model': 1.3811513709370047, 'learning_rate/qwen_vl_interface': 9.999835010848254e-06, 'learning_rate/action_model': 9.999835010848254e-05, 'epoch': 7.79}) +[2026-06-20 19:09:16,564][starVLA.training.train_starvla][INFO] - Step 4214, Loss: {'train/loss': 0.06160092353820801, 'train/grad_norm_pre_clip': 13.58051872253418, 'timing/data': 0.00647423998452723, 'timing/model': 1.1440765709849074, 'learning_rate/qwen_vl_interface': 9.999808651811639e-06, 'learning_rate/action_model': 9.999808651811639e-05, 'epoch': 7.79}) +[2026-06-20 19:09:21,324][starVLA.training.train_starvla][INFO] - Step 4215, Loss: {'train/loss': 0.19012582302093506, 'train/grad_norm_pre_clip': 17.176340103149414, 'timing/data': 0.0072258670115843415, 'timing/model': 1.2149304630002007, 'learning_rate/qwen_vl_interface': 9.999780340308101e-06, 'learning_rate/action_model': 9.999780340308101e-05, 'epoch': 7.8}) +[2026-06-20 19:09:26,224][starVLA.training.train_starvla][INFO] - Step 4216, Loss: {'train/loss': 0.018846018239855766, 'train/grad_norm_pre_clip': 2.3544111251831055, 'timing/data': 0.0063410509610548615, 'timing/model': 1.2115952849853784, 'learning_rate/qwen_vl_interface': 9.999750076349282e-06, 'learning_rate/action_model': 9.999750076349282e-05, 'epoch': 7.8}) +[2026-06-20 19:09:31,037][starVLA.training.train_starvla][INFO] - Step 4217, Loss: {'train/loss': 0.10404430329799652, 'train/grad_norm_pre_clip': 13.752593040466309, 'timing/data': 0.007527895970270038, 'timing/model': 1.19781481300015, 'learning_rate/qwen_vl_interface': 9.999717859947618e-06, 'learning_rate/action_model': 9.999717859947618e-05, 'epoch': 7.8}) +[2026-06-20 19:09:35,776][starVLA.training.train_starvla][INFO] - Step 4218, Loss: {'train/loss': 0.28329119086265564, 'train/grad_norm_pre_clip': 15.515826225280762, 'timing/data': 0.006375400931574404, 'timing/model': 1.1451501089613885, 'learning_rate/qwen_vl_interface': 9.999683691116356e-06, 'learning_rate/action_model': 9.999683691116355e-05, 'epoch': 7.8}) +[2026-06-20 19:09:40,997][starVLA.training.train_starvla][INFO] - Step 4219, Loss: {'train/loss': 0.09668341279029846, 'train/grad_norm_pre_clip': 14.601890563964844, 'timing/data': 0.006306240917183459, 'timing/model': 1.3741516039008275, 'learning_rate/qwen_vl_interface': 9.999647569869539e-06, 'learning_rate/action_model': 9.99964756986954e-05, 'epoch': 7.8}) +[2026-06-20 19:09:45,661][starVLA.training.train_starvla][INFO] - Step 4220, Loss: {'train/loss': 0.005998712498694658, 'train/grad_norm_pre_clip': 1.5544308423995972, 'timing/data': 0.007626206032000482, 'timing/model': 1.1979920910671353, 'learning_rate/qwen_vl_interface': 9.999609496222017e-06, 'learning_rate/action_model': 9.999609496222017e-05, 'epoch': 7.8}) +[2026-06-20 19:09:50,432][starVLA.training.train_starvla][INFO] - Step 4221, Loss: {'train/loss': 0.04145257547497749, 'train/grad_norm_pre_clip': 24.453819274902344, 'timing/data': 0.006831538979895413, 'timing/model': 1.143988901982084, 'learning_rate/qwen_vl_interface': 9.99956947018944e-06, 'learning_rate/action_model': 9.999569470189439e-05, 'epoch': 7.81}) +[2026-06-20 19:09:55,273][starVLA.training.train_starvla][INFO] - Step 4222, Loss: {'train/loss': 0.17659282684326172, 'train/grad_norm_pre_clip': 20.60319709777832, 'timing/data': 0.006197681068442762, 'timing/model': 1.1435880189528689, 'learning_rate/qwen_vl_interface': 9.99952749178826e-06, 'learning_rate/action_model': 9.999527491788259e-05, 'epoch': 7.81}) +[2026-06-20 19:10:00,157][starVLA.training.train_starvla][INFO] - Step 4223, Loss: {'train/loss': 0.28487756848335266, 'train/grad_norm_pre_clip': 33.140228271484375, 'timing/data': 0.006301352055743337, 'timing/model': 1.144894749042578, 'learning_rate/qwen_vl_interface': 9.999483561035732e-06, 'learning_rate/action_model': 9.999483561035732e-05, 'epoch': 7.81}) +[2026-06-20 19:10:04,999][starVLA.training.train_starvla][INFO] - Step 4224, Loss: {'train/loss': 0.16198541224002838, 'train/grad_norm_pre_clip': 13.425725936889648, 'timing/data': 0.006286040996201336, 'timing/model': 1.372694174060598, 'learning_rate/qwen_vl_interface': 9.999437677949917e-06, 'learning_rate/action_model': 9.999437677949917e-05, 'epoch': 7.81}) +[2026-06-20 19:10:09,707][starVLA.training.train_starvla][INFO] - Step 4225, Loss: {'train/loss': 0.24781669676303864, 'train/grad_norm_pre_clip': 17.936525344848633, 'timing/data': 0.006314381957054138, 'timing/model': 1.2935563169885427, 'learning_rate/qwen_vl_interface': 9.999389842549674e-06, 'learning_rate/action_model': 9.999389842549673e-05, 'epoch': 7.81}) +[2026-06-20 19:10:14,219][starVLA.training.train_starvla][INFO] - Step 4226, Loss: {'train/loss': 0.12737572193145752, 'train/grad_norm_pre_clip': 11.887471199035645, 'timing/data': 0.006689200992695987, 'timing/model': 1.1949174279579893, 'learning_rate/qwen_vl_interface': 9.999340054854669e-06, 'learning_rate/action_model': 9.999340054854667e-05, 'epoch': 7.82}) +[2026-06-20 19:10:19,062][starVLA.training.train_starvla][INFO] - Step 4227, Loss: {'train/loss': 0.12232277542352676, 'train/grad_norm_pre_clip': 17.816028594970703, 'timing/data': 0.00631214096210897, 'timing/model': 1.2130409919191152, 'learning_rate/qwen_vl_interface': 9.999288314885362e-06, 'learning_rate/action_model': 9.999288314885362e-05, 'epoch': 7.82}) +[2026-06-20 19:10:23,780][starVLA.training.train_starvla][INFO] - Step 4228, Loss: {'train/loss': 0.13536955416202545, 'train/grad_norm_pre_clip': 12.796392440795898, 'timing/data': 0.006492440006695688, 'timing/model': 1.1444913659943268, 'learning_rate/qwen_vl_interface': 9.999234622663027e-06, 'learning_rate/action_model': 9.999234622663028e-05, 'epoch': 7.82}) +[2026-06-20 19:10:28,603][starVLA.training.train_starvla][INFO] - Step 4229, Loss: {'train/loss': 0.18667645752429962, 'train/grad_norm_pre_clip': 4.8491034507751465, 'timing/data': 0.008747200015932322, 'timing/model': 1.2135763010010123, 'learning_rate/qwen_vl_interface': 9.999178978209732e-06, 'learning_rate/action_model': 9.999178978209732e-05, 'epoch': 7.82}) +[2026-06-20 19:10:33,464][starVLA.training.train_starvla][INFO] - Step 4230, Loss: {'train/loss': 0.2519165873527527, 'train/grad_norm_pre_clip': 17.173877716064453, 'timing/data': 0.007583555998280644, 'timing/model': 1.4542185040190816, 'learning_rate/qwen_vl_interface': 9.999121381548354e-06, 'learning_rate/action_model': 9.999121381548353e-05, 'epoch': 7.82}) +[2026-06-20 19:10:38,252][starVLA.training.train_starvla][INFO] - Step 4231, Loss: {'train/loss': 0.1493670493364334, 'train/grad_norm_pre_clip': 9.268229484558105, 'timing/data': 0.006359581951983273, 'timing/model': 1.3765889289788902, 'learning_rate/qwen_vl_interface': 9.999061832702563e-06, 'learning_rate/action_model': 9.999061832702563e-05, 'epoch': 7.82}) +[2026-06-20 19:10:42,839][starVLA.training.train_starvla][INFO] - Step 4232, Loss: {'train/loss': 0.2289835810661316, 'train/grad_norm_pre_clip': 18.453693389892578, 'timing/data': 0.006793268956243992, 'timing/model': 1.1991745210252702, 'learning_rate/qwen_vl_interface': 9.999000331696842e-06, 'learning_rate/action_model': 9.999000331696841e-05, 'epoch': 7.83}) +[2026-06-20 19:10:47,584][starVLA.training.train_starvla][INFO] - Step 4233, Loss: {'train/loss': 0.037169042974710464, 'train/grad_norm_pre_clip': 5.228936195373535, 'timing/data': 0.006468050996772945, 'timing/model': 1.144573800964281, 'learning_rate/qwen_vl_interface': 9.99893687855647e-06, 'learning_rate/action_model': 9.99893687855647e-05, 'epoch': 7.83}) +[2026-06-20 19:10:52,435][starVLA.training.train_starvla][INFO] - Step 4234, Loss: {'train/loss': 0.2486913502216339, 'train/grad_norm_pre_clip': 11.376269340515137, 'timing/data': 0.007677846006117761, 'timing/model': 1.1967120420886204, 'learning_rate/qwen_vl_interface': 9.998871473307529e-06, 'learning_rate/action_model': 9.99887147330753e-05, 'epoch': 7.83}) +[2026-06-20 19:10:57,143][starVLA.training.train_starvla][INFO] - Step 4235, Loss: {'train/loss': 0.03375314176082611, 'train/grad_norm_pre_clip': 10.233063697814941, 'timing/data': 0.006438391050323844, 'timing/model': 1.1433919260744005, 'learning_rate/qwen_vl_interface': 9.998804115976907e-06, 'learning_rate/action_model': 9.998804115976908e-05, 'epoch': 7.83}) +[2026-06-20 19:11:01,867][starVLA.training.train_starvla][INFO] - Step 4236, Loss: {'train/loss': 0.11282578110694885, 'train/grad_norm_pre_clip': 11.435510635375977, 'timing/data': 0.00640948093496263, 'timing/model': 1.3816547859460115, 'learning_rate/qwen_vl_interface': 9.99873480659229e-06, 'learning_rate/action_model': 9.99873480659229e-05, 'epoch': 7.83}) +[2026-06-20 19:11:06,519][starVLA.training.train_starvla][INFO] - Step 4237, Loss: {'train/loss': 0.10018806904554367, 'train/grad_norm_pre_clip': 9.02066707611084, 'timing/data': 0.007642826065421104, 'timing/model': 1.2001418309519067, 'learning_rate/qwen_vl_interface': 9.99866354518217e-06, 'learning_rate/action_model': 9.99866354518217e-05, 'epoch': 7.84}) +[2026-06-20 19:11:11,307][starVLA.training.train_starvla][INFO] - Step 4238, Loss: {'train/loss': 0.1250045895576477, 'train/grad_norm_pre_clip': 14.453478813171387, 'timing/data': 0.0067697500344365835, 'timing/model': 1.1473472019424662, 'learning_rate/qwen_vl_interface': 9.99859033177584e-06, 'learning_rate/action_model': 9.99859033177584e-05, 'epoch': 7.84}) +[2026-06-20 19:11:16,183][starVLA.training.train_starvla][INFO] - Step 4239, Loss: {'train/loss': 0.1865188628435135, 'train/grad_norm_pre_clip': 12.20556640625, 'timing/data': 0.006942679057829082, 'timing/model': 1.1452719159424305, 'learning_rate/qwen_vl_interface': 9.998515166403394e-06, 'learning_rate/action_model': 9.998515166403394e-05, 'epoch': 7.84}) +[2026-06-20 19:11:21,000][starVLA.training.train_starvla][INFO] - Step 4240, Loss: {'train/loss': 0.09977440536022186, 'train/grad_norm_pre_clip': 5.449709415435791, 'timing/data': 0.0069927789736539125, 'timing/model': 1.143664747942239, 'learning_rate/qwen_vl_interface': 9.99843804909573e-06, 'learning_rate/action_model': 9.99843804909573e-05, 'epoch': 7.84}) +[2026-06-20 19:11:25,779][starVLA.training.train_starvla][INFO] - Step 4241, Loss: {'train/loss': 0.08585669100284576, 'train/grad_norm_pre_clip': 7.600797653198242, 'timing/data': 0.0067269300343468785, 'timing/model': 1.1440408809576184, 'learning_rate/qwen_vl_interface': 9.998358979884548e-06, 'learning_rate/action_model': 9.998358979884547e-05, 'epoch': 7.84}) +[2026-06-20 19:11:30,624][starVLA.training.train_starvla][INFO] - Step 4242, Loss: {'train/loss': 0.044263679534196854, 'train/grad_norm_pre_clip': 5.120937824249268, 'timing/data': 0.006809330079704523, 'timing/model': 1.4500303650274873, 'learning_rate/qwen_vl_interface': 9.998277958802352e-06, 'learning_rate/action_model': 9.998277958802351e-05, 'epoch': 7.85}) +[2026-06-20 19:11:35,207][starVLA.training.train_starvla][INFO] - Step 4243, Loss: {'train/loss': 0.10256985574960709, 'train/grad_norm_pre_clip': 7.724034786224365, 'timing/data': 0.006263442104682326, 'timing/model': 1.1447686259634793, 'learning_rate/qwen_vl_interface': 9.998194985882442e-06, 'learning_rate/action_model': 9.998194985882442e-05, 'epoch': 7.85}) +[2026-06-20 19:11:39,992][starVLA.training.train_starvla][INFO] - Step 4244, Loss: {'train/loss': 0.2222275584936142, 'train/grad_norm_pre_clip': 19.40257453918457, 'timing/data': 0.006277901935391128, 'timing/model': 1.1444539329968393, 'learning_rate/qwen_vl_interface': 9.99811006115893e-06, 'learning_rate/action_model': 9.998110061158931e-05, 'epoch': 7.85}) +[2026-06-20 19:11:44,778][starVLA.training.train_starvla][INFO] - Step 4245, Loss: {'train/loss': 0.1875147670507431, 'train/grad_norm_pre_clip': 11.428284645080566, 'timing/data': 0.006124922074377537, 'timing/model': 1.1463027080753818, 'learning_rate/qwen_vl_interface': 9.998023184666726e-06, 'learning_rate/action_model': 9.998023184666726e-05, 'epoch': 7.85}) +[2026-06-20 19:11:49,555][starVLA.training.train_starvla][INFO] - Step 4246, Loss: {'train/loss': 0.06363534927368164, 'train/grad_norm_pre_clip': 5.051809787750244, 'timing/data': 0.006461841054260731, 'timing/model': 1.2168085599550977, 'learning_rate/qwen_vl_interface': 9.997934356441538e-06, 'learning_rate/action_model': 9.997934356441537e-05, 'epoch': 7.85}) +[2026-06-20 19:11:54,374][starVLA.training.train_starvla][INFO] - Step 4247, Loss: {'train/loss': 0.31001123785972595, 'train/grad_norm_pre_clip': 16.76366424560547, 'timing/data': 0.006425431929528713, 'timing/model': 1.2150015330407768, 'learning_rate/qwen_vl_interface': 9.997843576519879e-06, 'learning_rate/action_model': 9.997843576519878e-05, 'epoch': 7.85}) +[2026-06-20 19:11:59,212][starVLA.training.train_starvla][INFO] - Step 4248, Loss: {'train/loss': 0.1725444346666336, 'train/grad_norm_pre_clip': 12.22127628326416, 'timing/data': 0.00749690702650696, 'timing/model': 1.426125384052284, 'learning_rate/qwen_vl_interface': 9.99775084493907e-06, 'learning_rate/action_model': 9.997750844939069e-05, 'epoch': 7.86}) +[2026-06-20 19:12:03,948][starVLA.training.train_starvla][INFO] - Step 4249, Loss: {'train/loss': 0.12966492772102356, 'train/grad_norm_pre_clip': 7.284249782562256, 'timing/data': 0.006559300934895873, 'timing/model': 1.3220807379111648, 'learning_rate/qwen_vl_interface': 9.997656161737227e-06, 'learning_rate/action_model': 9.997656161737226e-05, 'epoch': 7.86}) +[2026-06-20 19:12:08,539][starVLA.training.train_starvla][INFO] - Step 4250, Loss: {'train/loss': 0.18241377174854279, 'train/grad_norm_pre_clip': 9.227348327636719, 'timing/data': 0.006267742021009326, 'timing/model': 1.2159020929830149, 'learning_rate/qwen_vl_interface': 9.99755952695327e-06, 'learning_rate/action_model': 9.997559526953269e-05, 'epoch': 7.86}) +[2026-06-20 19:12:13,335][starVLA.training.train_starvla][INFO] - Step 4251, Loss: {'train/loss': 0.06971435993909836, 'train/grad_norm_pre_clip': 8.190794944763184, 'timing/data': 0.007888155058026314, 'timing/model': 1.1988223619991913, 'learning_rate/qwen_vl_interface': 9.99746094062692e-06, 'learning_rate/action_model': 9.99746094062692e-05, 'epoch': 7.86}) +[2026-06-20 19:12:18,194][starVLA.training.train_starvla][INFO] - Step 4252, Loss: {'train/loss': 0.1412528157234192, 'train/grad_norm_pre_clip': 6.6381330490112305, 'timing/data': 0.006490750936791301, 'timing/model': 1.2153321440564469, 'learning_rate/qwen_vl_interface': 9.997360402798706e-06, 'learning_rate/action_model': 9.997360402798705e-05, 'epoch': 7.86}) +[2026-06-20 19:12:22,958][starVLA.training.train_starvla][INFO] - Step 4253, Loss: {'train/loss': 0.1331930309534073, 'train/grad_norm_pre_clip': 5.793239593505859, 'timing/data': 0.006623429944738746, 'timing/model': 1.143718654057011, 'learning_rate/qwen_vl_interface': 9.997257913509955e-06, 'learning_rate/action_model': 9.997257913509955e-05, 'epoch': 7.87}) +[2026-06-20 19:12:27,735][starVLA.training.train_starvla][INFO] - Step 4254, Loss: {'train/loss': 0.1397993564605713, 'train/grad_norm_pre_clip': 8.236945152282715, 'timing/data': 0.006318081985227764, 'timing/model': 1.383190723019652, 'learning_rate/qwen_vl_interface': 9.997153472802794e-06, 'learning_rate/action_model': 9.997153472802795e-05, 'epoch': 7.87}) +[2026-06-20 19:12:32,587][starVLA.training.train_starvla][INFO] - Step 4255, Loss: {'train/loss': 0.05171199515461922, 'train/grad_norm_pre_clip': 3.8085484504699707, 'timing/data': 0.007606846047565341, 'timing/model': 1.4400826080236584, 'learning_rate/qwen_vl_interface': 9.997047080720157e-06, 'learning_rate/action_model': 9.997047080720158e-05, 'epoch': 7.87}) +[2026-06-20 19:12:37,252][starVLA.training.train_starvla][INFO] - Step 4256, Loss: {'train/loss': 0.17498880624771118, 'train/grad_norm_pre_clip': 11.306768417358398, 'timing/data': 0.008050094009377062, 'timing/model': 1.2152005909010768, 'learning_rate/qwen_vl_interface': 9.996938737305777e-06, 'learning_rate/action_model': 9.996938737305776e-05, 'epoch': 7.87}) +[2026-06-20 19:12:42,093][starVLA.training.train_starvla][INFO] - Step 4257, Loss: {'train/loss': 0.2676865756511688, 'train/grad_norm_pre_clip': 9.988024711608887, 'timing/data': 0.00894543994218111, 'timing/model': 1.2153129450744018, 'learning_rate/qwen_vl_interface': 9.996828442604189e-06, 'learning_rate/action_model': 9.99682844260419e-05, 'epoch': 7.87}) +[2026-06-20 19:12:46,951][starVLA.training.train_starvla][INFO] - Step 4258, Loss: {'train/loss': 0.05480984225869179, 'train/grad_norm_pre_clip': 3.3154947757720947, 'timing/data': 0.006427751039154828, 'timing/model': 1.218398105003871, 'learning_rate/qwen_vl_interface': 9.996716196660733e-06, 'learning_rate/action_model': 9.996716196660731e-05, 'epoch': 7.87}) +[2026-06-20 19:12:51,790][starVLA.training.train_starvla][INFO] - Step 4259, Loss: {'train/loss': 0.15849119424819946, 'train/grad_norm_pre_clip': 7.886526584625244, 'timing/data': 0.006459672003984451, 'timing/model': 1.1460715259891003, 'learning_rate/qwen_vl_interface': 9.996601999521546e-06, 'learning_rate/action_model': 9.996601999521546e-05, 'epoch': 7.88}) +[2026-06-20 19:12:56,579][starVLA.training.train_starvla][INFO] - Step 4260, Loss: {'train/loss': 0.0894368514418602, 'train/grad_norm_pre_clip': 7.650279521942139, 'timing/data': 0.00647933199070394, 'timing/model': 1.3882202510721982, 'learning_rate/qwen_vl_interface': 9.996485851233572e-06, 'learning_rate/action_model': 9.996485851233572e-05, 'epoch': 7.88}) +[2026-06-20 19:13:01,047][starVLA.training.train_starvla][INFO] - Step 4261, Loss: {'train/loss': 0.08148973435163498, 'train/grad_norm_pre_clip': 4.736465930938721, 'timing/data': 0.005983683979138732, 'timing/model': 1.1459420350147411, 'learning_rate/qwen_vl_interface': 9.996367751844556e-06, 'learning_rate/action_model': 9.996367751844555e-05, 'epoch': 7.88}) +[2026-06-20 19:13:05,957][starVLA.training.train_starvla][INFO] - Step 4262, Loss: {'train/loss': 0.045119140297174454, 'train/grad_norm_pre_clip': 4.997040748596191, 'timing/data': 0.005855205003172159, 'timing/model': 1.1461206370731816, 'learning_rate/qwen_vl_interface': 9.996247701403043e-06, 'learning_rate/action_model': 9.996247701403042e-05, 'epoch': 7.88}) +[2026-06-20 19:13:10,814][starVLA.training.train_starvla][INFO] - Step 4263, Loss: {'train/loss': 0.1554449498653412, 'train/grad_norm_pre_clip': 5.740055561065674, 'timing/data': 0.006607061950489879, 'timing/model': 1.1471926979720592, 'learning_rate/qwen_vl_interface': 9.99612569995838e-06, 'learning_rate/action_model': 9.99612569995838e-05, 'epoch': 7.88}) +[2026-06-20 19:13:15,544][starVLA.training.train_starvla][INFO] - Step 4264, Loss: {'train/loss': 0.2125139683485031, 'train/grad_norm_pre_clip': 9.984844207763672, 'timing/data': 0.0058228750713169575, 'timing/model': 1.1475257689598948, 'learning_rate/qwen_vl_interface': 9.99600174756072e-06, 'learning_rate/action_model': 9.99600174756072e-05, 'epoch': 7.89}) +[2026-06-20 19:13:20,331][starVLA.training.train_starvla][INFO] - Step 4265, Loss: {'train/loss': 0.13292638957500458, 'train/grad_norm_pre_clip': 13.30655574798584, 'timing/data': 0.0069525010185316205, 'timing/model': 1.1999088729498908, 'learning_rate/qwen_vl_interface': 9.995875844261012e-06, 'learning_rate/action_model': 9.995875844261013e-05, 'epoch': 7.89}) +[2026-06-20 19:13:25,079][starVLA.training.train_starvla][INFO] - Step 4266, Loss: {'train/loss': 0.13362190127372742, 'train/grad_norm_pre_clip': 7.39555549621582, 'timing/data': 0.005892353947274387, 'timing/model': 1.3506728599313647, 'learning_rate/qwen_vl_interface': 9.995747990111015e-06, 'learning_rate/action_model': 9.995747990111013e-05, 'epoch': 7.89}) +[2026-06-20 19:13:29,622][starVLA.training.train_starvla][INFO] - Step 4267, Loss: {'train/loss': 0.11896422505378723, 'train/grad_norm_pre_clip': 12.020231246948242, 'timing/data': 0.00802420498803258, 'timing/model': 1.2184697799384594, 'learning_rate/qwen_vl_interface': 9.99561818516328e-06, 'learning_rate/action_model': 9.99561818516328e-05, 'epoch': 7.89}) +[2026-06-20 19:13:34,468][starVLA.training.train_starvla][INFO] - Step 4268, Loss: {'train/loss': 0.23944756388664246, 'train/grad_norm_pre_clip': 10.253332138061523, 'timing/data': 0.006023364025168121, 'timing/model': 1.1462053690338507, 'learning_rate/qwen_vl_interface': 9.995486429471166e-06, 'learning_rate/action_model': 9.995486429471166e-05, 'epoch': 7.89}) +[2026-06-20 19:13:39,306][starVLA.training.train_starvla][INFO] - Step 4269, Loss: {'train/loss': 0.20120607316493988, 'train/grad_norm_pre_clip': 18.043720245361328, 'timing/data': 0.007052398985251784, 'timing/model': 1.201513092033565, 'learning_rate/qwen_vl_interface': 9.995352723088833e-06, 'learning_rate/action_model': 9.995352723088834e-05, 'epoch': 7.89}) +[2026-06-20 19:13:44,033][starVLA.training.train_starvla][INFO] - Step 4270, Loss: {'train/loss': 0.4058670103549957, 'train/grad_norm_pre_clip': 10.08877944946289, 'timing/data': 0.006443041958846152, 'timing/model': 1.1459131069714203, 'learning_rate/qwen_vl_interface': 9.995217066071245e-06, 'learning_rate/action_model': 9.995217066071246e-05, 'epoch': 7.9}) +[2026-06-20 19:13:48,896][starVLA.training.train_starvla][INFO] - Step 4271, Loss: {'train/loss': 0.1932152360677719, 'train/grad_norm_pre_clip': 10.080914497375488, 'timing/data': 0.0059579540975391865, 'timing/model': 1.145439003012143, 'learning_rate/qwen_vl_interface': 9.995079458474163e-06, 'learning_rate/action_model': 9.995079458474164e-05, 'epoch': 7.9}) +[2026-06-20 19:13:53,707][starVLA.training.train_starvla][INFO] - Step 4272, Loss: {'train/loss': 0.049416426569223404, 'train/grad_norm_pre_clip': 4.104848384857178, 'timing/data': 0.007917374954558909, 'timing/model': 1.4104814960155636, 'learning_rate/qwen_vl_interface': 9.994939900354154e-06, 'learning_rate/action_model': 9.994939900354154e-05, 'epoch': 7.9}) +[2026-06-20 19:13:58,263][starVLA.training.train_starvla][INFO] - Step 4273, Loss: {'train/loss': 0.13397815823554993, 'train/grad_norm_pre_clip': 10.647139549255371, 'timing/data': 0.006528800935484469, 'timing/model': 1.1467897839611396, 'learning_rate/qwen_vl_interface': 9.994798391768582e-06, 'learning_rate/action_model': 9.994798391768582e-05, 'epoch': 7.9}) +[2026-06-20 19:14:03,082][starVLA.training.train_starvla][INFO] - Step 4274, Loss: {'train/loss': 0.14770223200321198, 'train/grad_norm_pre_clip': 5.828213214874268, 'timing/data': 0.006203872966580093, 'timing/model': 1.2192156930686906, 'learning_rate/qwen_vl_interface': 9.99465493277562e-06, 'learning_rate/action_model': 9.994654932775618e-05, 'epoch': 7.9}) +[2026-06-20 19:14:07,934][starVLA.training.train_starvla][INFO] - Step 4275, Loss: {'train/loss': 0.06197278946638107, 'train/grad_norm_pre_clip': 6.184751510620117, 'timing/data': 0.006265213014557958, 'timing/model': 1.1474717779783532, 'learning_rate/qwen_vl_interface': 9.994509523434235e-06, 'learning_rate/action_model': 9.994509523434235e-05, 'epoch': 7.91}) +[2026-06-20 19:14:12,736][starVLA.training.train_starvla][INFO] - Step 4276, Loss: {'train/loss': 0.16507385671138763, 'train/grad_norm_pre_clip': 9.114276885986328, 'timing/data': 0.005797913996502757, 'timing/model': 1.1449847019976005, 'learning_rate/qwen_vl_interface': 9.994362163804202e-06, 'learning_rate/action_model': 9.9943621638042e-05, 'epoch': 7.91}) +[2026-06-20 19:14:17,591][starVLA.training.train_starvla][INFO] - Step 4277, Loss: {'train/loss': 0.09414287656545639, 'train/grad_norm_pre_clip': 5.77775764465332, 'timing/data': 0.005961354007013142, 'timing/model': 1.391596027999185, 'learning_rate/qwen_vl_interface': 9.994212853946092e-06, 'learning_rate/action_model': 9.994212853946092e-05, 'epoch': 7.91}) +[2026-06-20 19:14:22,136][starVLA.training.train_starvla][INFO] - Step 4278, Loss: {'train/loss': 0.046426255255937576, 'train/grad_norm_pre_clip': 3.6179208755493164, 'timing/data': 0.006386901950463653, 'timing/model': 1.152629804913886, 'learning_rate/qwen_vl_interface': 9.994061593921283e-06, 'learning_rate/action_model': 9.994061593921282e-05, 'epoch': 7.91}) +[2026-06-20 19:14:26,964][starVLA.training.train_starvla][INFO] - Step 4279, Loss: {'train/loss': 0.022276680916547775, 'train/grad_norm_pre_clip': 1.5730713605880737, 'timing/data': 0.006632631062529981, 'timing/model': 1.1501191690331325, 'learning_rate/qwen_vl_interface': 9.993908383791951e-06, 'learning_rate/action_model': 9.993908383791951e-05, 'epoch': 7.91}) +[2026-06-20 19:14:31,642][starVLA.training.train_starvla][INFO] - Step 4280, Loss: {'train/loss': 0.1318938434123993, 'train/grad_norm_pre_clip': 9.41196346282959, 'timing/data': 0.009459818946197629, 'timing/model': 1.175528471940197, 'learning_rate/qwen_vl_interface': 9.993753223621076e-06, 'learning_rate/action_model': 9.993753223621077e-05, 'epoch': 7.92}) +[2026-06-20 19:14:36,278][starVLA.training.train_starvla][INFO] - Step 4281, Loss: {'train/loss': 0.052247993648052216, 'train/grad_norm_pre_clip': 3.262216329574585, 'timing/data': 0.006171423010528088, 'timing/model': 1.1534290310228243, 'learning_rate/qwen_vl_interface': 9.99359611347244e-06, 'learning_rate/action_model': 9.993596113472439e-05, 'epoch': 7.92}) +[2026-06-20 19:14:40,942][starVLA.training.train_starvla][INFO] - Step 4282, Loss: {'train/loss': 0.09678557515144348, 'train/grad_norm_pre_clip': 8.48363208770752, 'timing/data': 0.006282293004915118, 'timing/model': 1.1518719200976193, 'learning_rate/qwen_vl_interface': 9.99343705341062e-06, 'learning_rate/action_model': 9.99343705341062e-05, 'epoch': 7.92}) +[2026-06-20 19:14:45,640][starVLA.training.train_starvla][INFO] - Step 4283, Loss: {'train/loss': 0.02676788531243801, 'train/grad_norm_pre_clip': 3.0033137798309326, 'timing/data': 0.006495852023363113, 'timing/model': 1.3961357300868258, 'learning_rate/qwen_vl_interface': 9.993276043501004e-06, 'learning_rate/action_model': 9.993276043501004e-05, 'epoch': 7.92}) +[2026-06-20 19:14:50,119][starVLA.training.train_starvla][INFO] - Step 4284, Loss: {'train/loss': 0.18961173295974731, 'train/grad_norm_pre_clip': 11.991595268249512, 'timing/data': 0.006377083016559482, 'timing/model': 1.1476946539478377, 'learning_rate/qwen_vl_interface': 9.993113083809776e-06, 'learning_rate/action_model': 9.993113083809775e-05, 'epoch': 7.92}) +[2026-06-20 19:14:54,804][starVLA.training.train_starvla][INFO] - Step 4285, Loss: {'train/loss': 0.07979757338762283, 'train/grad_norm_pre_clip': 8.013201713562012, 'timing/data': 0.006281764013692737, 'timing/model': 1.1479848569724709, 'learning_rate/qwen_vl_interface': 9.992948174403921e-06, 'learning_rate/action_model': 9.992948174403922e-05, 'epoch': 7.92}) +[2026-06-20 19:14:59,489][starVLA.training.train_starvla][INFO] - Step 4286, Loss: {'train/loss': 0.16981841623783112, 'train/grad_norm_pre_clip': 12.42790699005127, 'timing/data': 0.006437851930968463, 'timing/model': 1.1491593039827421, 'learning_rate/qwen_vl_interface': 9.99278131535123e-06, 'learning_rate/action_model': 9.992781315351229e-05, 'epoch': 7.93}) +[2026-06-20 19:15:04,176][starVLA.training.train_starvla][INFO] - Step 4287, Loss: {'train/loss': 0.04593096673488617, 'train/grad_norm_pre_clip': 7.511281967163086, 'timing/data': 0.00639092200435698, 'timing/model': 1.1482209509704262, 'learning_rate/qwen_vl_interface': 9.992612506720289e-06, 'learning_rate/action_model': 9.99261250672029e-05, 'epoch': 7.93}) +[2026-06-20 19:15:08,889][starVLA.training.train_starvla][INFO] - Step 4288, Loss: {'train/loss': 0.17871588468551636, 'train/grad_norm_pre_clip': 16.228361129760742, 'timing/data': 0.008746281964704394, 'timing/model': 1.1745187209453434, 'learning_rate/qwen_vl_interface': 9.992441748580492e-06, 'learning_rate/action_model': 9.99244174858049e-05, 'epoch': 7.93}) +[2026-06-20 19:15:13,553][starVLA.training.train_starvla][INFO] - Step 4289, Loss: {'train/loss': 0.006842765025794506, 'train/grad_norm_pre_clip': 1.2620587348937988, 'timing/data': 0.006357623031362891, 'timing/model': 1.393176565063186, 'learning_rate/qwen_vl_interface': 9.992269041002028e-06, 'learning_rate/action_model': 9.992269041002027e-05, 'epoch': 7.93}) +[2026-06-20 19:15:17,989][starVLA.training.train_starvla][INFO] - Step 4290, Loss: {'train/loss': 0.12348032742738724, 'train/grad_norm_pre_clip': 16.55510902404785, 'timing/data': 0.006450801971368492, 'timing/model': 1.1446339150425047, 'learning_rate/qwen_vl_interface': 9.992094384055894e-06, 'learning_rate/action_model': 9.992094384055895e-05, 'epoch': 7.93}) +[2026-06-20 19:15:22,640][starVLA.training.train_starvla][INFO] - Step 4291, Loss: {'train/loss': 0.3059884011745453, 'train/grad_norm_pre_clip': 17.3972110748291, 'timing/data': 0.006414233008399606, 'timing/model': 1.1442963290028274, 'learning_rate/qwen_vl_interface': 9.991917777813884e-06, 'learning_rate/action_model': 9.991917777813884e-05, 'epoch': 7.94}) +[2026-06-20 19:15:27,327][starVLA.training.train_starvla][INFO] - Step 4292, Loss: {'train/loss': 0.1769251674413681, 'train/grad_norm_pre_clip': 14.451071739196777, 'timing/data': 0.006415931973606348, 'timing/model': 1.1430981369921938, 'learning_rate/qwen_vl_interface': 9.991739222348593e-06, 'learning_rate/action_model': 9.991739222348592e-05, 'epoch': 7.94}) +[2026-06-20 19:15:32,021][starVLA.training.train_starvla][INFO] - Step 4293, Loss: {'train/loss': 0.14608317613601685, 'train/grad_norm_pre_clip': 19.82793426513672, 'timing/data': 0.006338233011774719, 'timing/model': 1.1631377349840477, 'learning_rate/qwen_vl_interface': 9.991558717733418e-06, 'learning_rate/action_model': 9.991558717733417e-05, 'epoch': 7.94}) +[2026-06-20 19:15:36,747][starVLA.training.train_starvla][INFO] - Step 4294, Loss: {'train/loss': 0.2737833261489868, 'train/grad_norm_pre_clip': 18.938627243041992, 'timing/data': 0.007625806960277259, 'timing/model': 1.4188749609747902, 'learning_rate/qwen_vl_interface': 9.99137626404256e-06, 'learning_rate/action_model': 9.991376264042559e-05, 'epoch': 7.94}) +[2026-06-20 19:15:41,360][starVLA.training.train_starvla][INFO] - Step 4295, Loss: {'train/loss': 0.15457071363925934, 'train/grad_norm_pre_clip': 5.770223617553711, 'timing/data': 0.006381963030435145, 'timing/model': 1.324529608944431, 'learning_rate/qwen_vl_interface': 9.991191861351015e-06, 'learning_rate/action_model': 9.991191861351014e-05, 'epoch': 7.94}) +[2026-06-20 19:15:45,784][starVLA.training.train_starvla][INFO] - Step 4296, Loss: {'train/loss': 0.12520171701908112, 'train/grad_norm_pre_clip': 20.910160064697266, 'timing/data': 0.006454552058130503, 'timing/model': 1.1629245929652825, 'learning_rate/qwen_vl_interface': 9.991005509734587e-06, 'learning_rate/action_model': 9.991005509734587e-05, 'epoch': 7.94}) +[2026-06-20 19:15:50,635][starVLA.training.train_starvla][INFO] - Step 4297, Loss: {'train/loss': 0.05177530273795128, 'train/grad_norm_pre_clip': 10.771109580993652, 'timing/data': 0.007857256103307009, 'timing/model': 1.216998154995963, 'learning_rate/qwen_vl_interface': 9.990817209269878e-06, 'learning_rate/action_model': 9.990817209269878e-05, 'epoch': 7.95}) +[2026-06-20 19:15:55,283][starVLA.training.train_starvla][INFO] - Step 4298, Loss: {'train/loss': 0.08622601628303528, 'train/grad_norm_pre_clip': 10.286998748779297, 'timing/data': 0.0064264730317518115, 'timing/model': 1.143163513042964, 'learning_rate/qwen_vl_interface': 9.990626960034291e-06, 'learning_rate/action_model': 9.990626960034291e-05, 'epoch': 7.95}) +[2026-06-20 19:16:00,121][starVLA.training.train_starvla][INFO] - Step 4299, Loss: {'train/loss': 0.38290050625801086, 'train/grad_norm_pre_clip': 16.391666412353516, 'timing/data': 0.00639084295835346, 'timing/model': 1.2152580369729549, 'learning_rate/qwen_vl_interface': 9.990434762106029e-06, 'learning_rate/action_model': 9.990434762106028e-05, 'epoch': 7.95}) +[2026-06-20 19:16:04,988][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 4300 +[2026-06-20 19:22:34,880][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 4300 +[2026-06-20 19:22:34,881][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 4300 +[2026-06-20 19:29:31,005][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 4300 +[2026-06-20 19:29:31,007][starVLA.training.train_starvla][INFO] - Step 4300, Loss: {'train/loss': 0.011568406596779823, 'train/grad_norm_pre_clip': 1.6989192962646484, 'eval/loss': 0.7286357879638672, 'eval/latency_0/loss': 0.7959065437316895, 'eval/latency_1/loss': 1.0733486413955688, 'eval/latency_2/loss': 1.1094505786895752, 'eval/latency_3/loss': 0.4893278479576111, 'eval/latency_4/loss': 0.13287736475467682, 'eval/flappy/loss': 0.7286357879638672, 'eval/flappy/latency_0/loss': 0.7959065437316895, 'eval/flappy/latency_1/loss': 1.0733486413955688, 'eval/flappy/latency_2/loss': 1.1094505786895752, 'eval/flappy/latency_3/loss': 0.4893278479576111, 'eval/flappy/latency_4/loss': 0.13287736475467682, 'eval/flap_precision': 0.16071428571428573, 'eval/flap_recall': 0.13090909090909092, 'eval/flap_f1': 0.1442885771543086, 'eval/flap_rate_model': 0.06222222222222222, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.1442885771543086, 'eval/latency_0/flap_precision': 0.16071428571428573, 'eval/latency_0/flap_recall': 0.13090909090909092, 'eval/latency_0/flap_f1': 0.1442885771543086, 'eval/latency_0/flap_rate_model': 0.06222222222222222, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.1442885771543086, 'timing/data': 0.006526621989905834, 'timing/model': 1.3914692780235782, 'learning_rate/qwen_vl_interface': 9.990240615564099e-06, 'learning_rate/action_model': 9.990240615564099e-05, 'epoch': 7.95}) +[2026-06-20 19:29:31,008][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_4300_state +[2026-06-20 19:29:31,015][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-20 19:29:39,422][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_4300_state/model.safetensors +[2026-06-20 19:29:53,054][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_4300_state/optimizer.bin +[2026-06-20 19:29:53,054][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_4300_state/sampler.bin +[2026-06-20 19:29:53,054][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_4300_state/sampler_1.bin +[2026-06-20 19:29:53,055][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_4300_state/random_states_0.pkl +[2026-06-20 19:29:55,663][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4300 +[2026-06-20 19:29:55,664][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-20 19:29:55,677][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-20 19:30:00,317][starVLA.training.train_starvla][INFO] - Step 4301, Loss: {'train/loss': 0.08826126158237457, 'train/grad_norm_pre_clip': 10.450840950012207, 'timing/data': 0.0066186729818582535, 'timing/model': 1.1283949080388993, 'learning_rate/qwen_vl_interface': 9.990044520488305e-06, 'learning_rate/action_model': 9.990044520488305e-05, 'epoch': 7.95}) +[2026-06-20 19:30:04,948][starVLA.training.train_starvla][INFO] - Step 4302, Loss: {'train/loss': 0.07838702201843262, 'train/grad_norm_pre_clip': 14.210023880004883, 'timing/data': 0.005797120044007897, 'timing/model': 1.3696451620198786, 'learning_rate/qwen_vl_interface': 9.989846476959258e-06, 'learning_rate/action_model': 9.989846476959257e-05, 'epoch': 7.96}) +[2026-06-20 19:30:09,487][starVLA.training.train_starvla][INFO] - Step 4303, Loss: {'train/loss': 0.1510300189256668, 'train/grad_norm_pre_clip': 6.3203253746032715, 'timing/data': 0.0069891540333628654, 'timing/model': 1.2850018759490922, 'learning_rate/qwen_vl_interface': 9.989646485058361e-06, 'learning_rate/action_model': 9.989646485058361e-05, 'epoch': 7.96}) +[2026-06-20 19:30:13,916][starVLA.training.train_starvla][INFO] - Step 4304, Loss: {'train/loss': 0.22579587996006012, 'train/grad_norm_pre_clip': 15.105668067932129, 'timing/data': 0.006464162026531994, 'timing/model': 1.1313978120451793, 'learning_rate/qwen_vl_interface': 9.98944454486783e-06, 'learning_rate/action_model': 9.989444544867829e-05, 'epoch': 7.96}) +[2026-06-20 19:30:18,635][starVLA.training.train_starvla][INFO] - Step 4305, Loss: {'train/loss': 0.17468398809432983, 'train/grad_norm_pre_clip': 9.627598762512207, 'timing/data': 0.006328960997052491, 'timing/model': 1.1363833140349016, 'learning_rate/qwen_vl_interface': 9.989240656470667e-06, 'learning_rate/action_model': 9.989240656470666e-05, 'epoch': 7.96}) +[2026-06-20 19:30:23,287][starVLA.training.train_starvla][INFO] - Step 4306, Loss: {'train/loss': 0.19349047541618347, 'train/grad_norm_pre_clip': 11.865586280822754, 'timing/data': 0.0063982210122048855, 'timing/model': 1.1538817969849333, 'learning_rate/qwen_vl_interface': 9.98903481995069e-06, 'learning_rate/action_model': 9.989034819950689e-05, 'epoch': 7.96}) +[2026-06-20 19:30:27,921][starVLA.training.train_starvla][INFO] - Step 4307, Loss: {'train/loss': 0.19441908597946167, 'train/grad_norm_pre_clip': 12.165345191955566, 'timing/data': 0.006385450949892402, 'timing/model': 1.1353677359875292, 'learning_rate/qwen_vl_interface': 9.988827035392506e-06, 'learning_rate/action_model': 9.988827035392507e-05, 'epoch': 7.97}) +[2026-06-20 19:30:32,561][starVLA.training.train_starvla][INFO] - Step 4308, Loss: {'train/loss': 0.03393552452325821, 'train/grad_norm_pre_clip': 4.131683826446533, 'timing/data': 0.00640499999281019, 'timing/model': 1.1368552889907733, 'learning_rate/qwen_vl_interface': 9.988617302881529e-06, 'learning_rate/action_model': 9.988617302881529e-05, 'epoch': 7.97}) +[2026-06-20 19:30:37,263][starVLA.training.train_starvla][INFO] - Step 4309, Loss: {'train/loss': 0.15507949888706207, 'train/grad_norm_pre_clip': 8.908339500427246, 'timing/data': 0.009450589073821902, 'timing/model': 1.3850860319798812, 'learning_rate/qwen_vl_interface': 9.988405622503972e-06, 'learning_rate/action_model': 9.988405622503971e-05, 'epoch': 7.97}) +[2026-06-20 19:30:41,706][starVLA.training.train_starvla][INFO] - Step 4310, Loss: {'train/loss': 0.08036041259765625, 'train/grad_norm_pre_clip': 7.462681293487549, 'timing/data': 0.006307910080067813, 'timing/model': 1.1402862990507856, 'learning_rate/qwen_vl_interface': 9.988191994346848e-06, 'learning_rate/action_model': 9.988191994346848e-05, 'epoch': 7.97}) +[2026-06-20 19:30:46,399][starVLA.training.train_starvla][INFO] - Step 4311, Loss: {'train/loss': 0.12632356584072113, 'train/grad_norm_pre_clip': 8.460562705993652, 'timing/data': 0.006183809018693864, 'timing/model': 1.1407848390517756, 'learning_rate/qwen_vl_interface': 9.987976418497974e-06, 'learning_rate/action_model': 9.987976418497972e-05, 'epoch': 7.97}) +[2026-06-20 19:30:51,092][starVLA.training.train_starvla][INFO] - Step 4312, Loss: {'train/loss': 0.11342348158359528, 'train/grad_norm_pre_clip': 9.639033317565918, 'timing/data': 0.006534259067848325, 'timing/model': 1.1660352740436792, 'learning_rate/qwen_vl_interface': 9.987758895045964e-06, 'learning_rate/action_model': 9.987758895045963e-05, 'epoch': 7.97}) +[2026-06-20 19:30:55,743][starVLA.training.train_starvla][INFO] - Step 4313, Loss: {'train/loss': 0.2441176474094391, 'train/grad_norm_pre_clip': 10.926501274108887, 'timing/data': 0.006506138946861029, 'timing/model': 1.1416335829999298, 'learning_rate/qwen_vl_interface': 9.98753942408023e-06, 'learning_rate/action_model': 9.98753942408023e-05, 'epoch': 7.98}) +[2026-06-20 19:31:00,383][starVLA.training.train_starvla][INFO] - Step 4314, Loss: {'train/loss': 0.08410128206014633, 'train/grad_norm_pre_clip': 6.459635257720947, 'timing/data': 0.008596724946983159, 'timing/model': 1.3765966349747032, 'learning_rate/qwen_vl_interface': 9.987318005690992e-06, 'learning_rate/action_model': 9.987318005690992e-05, 'epoch': 7.98}) +[2026-06-20 19:31:05,051][starVLA.training.train_starvla][INFO] - Step 4315, Loss: {'train/loss': 0.16846732795238495, 'train/grad_norm_pre_clip': 9.734726905822754, 'timing/data': 0.006342118023894727, 'timing/model': 1.3744379020063207, 'learning_rate/qwen_vl_interface': 9.987094639969268e-06, 'learning_rate/action_model': 9.987094639969268e-05, 'epoch': 7.98}) +[2026-06-20 19:31:09,493][starVLA.training.train_starvla][INFO] - Step 4316, Loss: {'train/loss': 0.051981788128614426, 'train/grad_norm_pre_clip': 4.718685150146484, 'timing/data': 0.006296427920460701, 'timing/model': 1.1402123059378937, 'learning_rate/qwen_vl_interface': 9.986869327006872e-06, 'learning_rate/action_model': 9.986869327006873e-05, 'epoch': 7.98}) +[2026-06-20 19:31:14,341][starVLA.training.train_starvla][INFO] - Step 4317, Loss: {'train/loss': 0.009563781321048737, 'train/grad_norm_pre_clip': 0.9335771799087524, 'timing/data': 0.006490687956102192, 'timing/model': 1.2153275680029765, 'learning_rate/qwen_vl_interface': 9.986642066896424e-06, 'learning_rate/action_model': 9.986642066896423e-05, 'epoch': 7.98}) +[2026-06-20 19:31:19,125][starVLA.training.train_starvla][INFO] - Step 4318, Loss: {'train/loss': 0.12757764756679535, 'train/grad_norm_pre_clip': 8.672800064086914, 'timing/data': 0.006378917023539543, 'timing/model': 1.1410516200121492, 'learning_rate/qwen_vl_interface': 9.986412859731341e-06, 'learning_rate/action_model': 9.98641285973134e-05, 'epoch': 7.99}) +[2026-06-20 19:31:23,823][starVLA.training.train_starvla][INFO] - Step 4319, Loss: {'train/loss': 0.11168034374713898, 'train/grad_norm_pre_clip': 9.693168640136719, 'timing/data': 0.006349076051265001, 'timing/model': 1.1384281780337915, 'learning_rate/qwen_vl_interface': 9.986181705605839e-06, 'learning_rate/action_model': 9.986181705605839e-05, 'epoch': 7.99}) +[2026-06-20 19:31:28,476][starVLA.training.train_starvla][INFO] - Step 4320, Loss: {'train/loss': 0.02152678556740284, 'train/grad_norm_pre_clip': 2.023366928100586, 'timing/data': 0.006452476023696363, 'timing/model': 1.408007180900313, 'learning_rate/qwen_vl_interface': 9.98594860461494e-06, 'learning_rate/action_model': 9.98594860461494e-05, 'epoch': 7.99}) +[2026-06-20 19:31:33,005][starVLA.training.train_starvla][INFO] - Step 4321, Loss: {'train/loss': 0.06665456295013428, 'train/grad_norm_pre_clip': 5.127556324005127, 'timing/data': 0.006329506053589284, 'timing/model': 1.2843731610337272, 'learning_rate/qwen_vl_interface': 9.985713556854461e-06, 'learning_rate/action_model': 9.98571355685446e-05, 'epoch': 7.99}) +[2026-06-20 19:31:37,389][starVLA.training.train_starvla][INFO] - Step 4322, Loss: {'train/loss': 0.03048766404390335, 'train/grad_norm_pre_clip': 4.894237995147705, 'timing/data': 0.006483935983851552, 'timing/model': 1.1386656580725685, 'learning_rate/qwen_vl_interface': 9.985476562421024e-06, 'learning_rate/action_model': 9.985476562421024e-05, 'epoch': 7.99}) +[2026-06-20 19:31:42,034][starVLA.training.train_starvla][INFO] - Step 4323, Loss: {'train/loss': 0.0302719883620739, 'train/grad_norm_pre_clip': 4.399001121520996, 'timing/data': 0.00650687504094094, 'timing/model': 1.1384309640852734, 'learning_rate/qwen_vl_interface': 9.985237621412047e-06, 'learning_rate/action_model': 9.985237621412046e-05, 'epoch': 7.99}) +[2026-06-20 19:31:46,658][starVLA.training.train_starvla][INFO] - Step 4324, Loss: {'train/loss': 0.021947283297777176, 'train/grad_norm_pre_clip': 3.6742985248565674, 'timing/data': 0.006670856033451855, 'timing/model': 1.1392676020041108, 'learning_rate/qwen_vl_interface': 9.984996733925749e-06, 'learning_rate/action_model': 9.984996733925748e-05, 'epoch': 8.0}) +[2026-06-20 19:31:51,313][starVLA.training.train_starvla][INFO] - Step 4325, Loss: {'train/loss': 0.011839773505926132, 'train/grad_norm_pre_clip': 2.0365827083587646, 'timing/data': 0.006617674953304231, 'timing/model': 1.1718655629083514, 'learning_rate/qwen_vl_interface': 9.98475390006115e-06, 'learning_rate/action_model': 9.98475390006115e-05, 'epoch': 8.0}) +[2026-06-20 19:31:55,932][starVLA.training.train_starvla][INFO] - Step 4326, Loss: {'train/loss': 0.5236960053443909, 'train/grad_norm_pre_clip': 36.18669128417969, 'timing/data': 0.006420094985514879, 'timing/model': 1.373414661968127, 'learning_rate/qwen_vl_interface': 9.98450911991807e-06, 'learning_rate/action_model': 9.98450911991807e-05, 'epoch': 8.0}) +[2026-06-20 19:32:00,559][starVLA.training.train_starvla][INFO] - Step 4327, Loss: {'train/loss': 0.5724462866783142, 'train/grad_norm_pre_clip': 21.103504180908203, 'timing/data': 0.006393584073521197, 'timing/model': 1.2973434220766649, 'learning_rate/qwen_vl_interface': 9.98426239359713e-06, 'learning_rate/action_model': 9.984262393597129e-05, 'epoch': 8.0}) +[2026-06-20 19:32:05,016][starVLA.training.train_starvla][INFO] - Step 4328, Loss: {'train/loss': 0.07892882078886032, 'train/grad_norm_pre_clip': 10.648148536682129, 'timing/data': 0.006300554028712213, 'timing/model': 1.1394618949852884, 'learning_rate/qwen_vl_interface': 9.984013721199747e-06, 'learning_rate/action_model': 9.984013721199747e-05, 'epoch': 8.0}) +[2026-06-20 19:32:09,634][starVLA.training.train_starvla][INFO] - Step 4329, Loss: {'train/loss': 0.2594021260738373, 'train/grad_norm_pre_clip': 9.356605529785156, 'timing/data': 0.008642817963846028, 'timing/model': 1.1601927189622074, 'learning_rate/qwen_vl_interface': 9.983763102828143e-06, 'learning_rate/action_model': 9.983763102828143e-05, 'epoch': 8.01}) +[2026-06-20 19:32:14,253][starVLA.training.train_starvla][INFO] - Step 4330, Loss: {'train/loss': 0.04859687760472298, 'train/grad_norm_pre_clip': 9.099649429321289, 'timing/data': 0.006361974054016173, 'timing/model': 1.1462178779765964, 'learning_rate/qwen_vl_interface': 9.983510538585336e-06, 'learning_rate/action_model': 9.983510538585337e-05, 'epoch': 8.01}) +[2026-06-20 19:32:19,055][starVLA.training.train_starvla][INFO] - Step 4331, Loss: {'train/loss': 0.08987099677324295, 'train/grad_norm_pre_clip': 11.901288986206055, 'timing/data': 0.00729000405408442, 'timing/model': 1.1962163569405675, 'learning_rate/qwen_vl_interface': 9.98325602857515e-06, 'learning_rate/action_model': 9.98325602857515e-05, 'epoch': 8.01}) +[2026-06-20 19:32:23,808][starVLA.training.train_starvla][INFO] - Step 4332, Loss: {'train/loss': 0.09146305173635483, 'train/grad_norm_pre_clip': 13.962274551391602, 'timing/data': 0.006703603081405163, 'timing/model': 1.3786850139731541, 'learning_rate/qwen_vl_interface': 9.9829995729022e-06, 'learning_rate/action_model': 9.982999572902199e-05, 'epoch': 8.01}) +[2026-06-20 19:32:28,235][starVLA.training.train_starvla][INFO] - Step 4333, Loss: {'train/loss': 0.0016499925404787064, 'train/grad_norm_pre_clip': 0.18773625791072845, 'timing/data': 0.006346223060972989, 'timing/model': 1.1420802499633282, 'learning_rate/qwen_vl_interface': 9.982741171671905e-06, 'learning_rate/action_model': 9.982741171671906e-05, 'epoch': 8.01}) +[2026-06-20 19:32:32,919][starVLA.training.train_starvla][INFO] - Step 4334, Loss: {'train/loss': 0.0061897821724414825, 'train/grad_norm_pre_clip': 1.2979214191436768, 'timing/data': 0.006577302003279328, 'timing/model': 1.1499604060081765, 'learning_rate/qwen_vl_interface': 9.982480824990489e-06, 'learning_rate/action_model': 9.982480824990488e-05, 'epoch': 8.02}) +[2026-06-20 19:32:37,590][starVLA.training.train_starvla][INFO] - Step 4335, Loss: {'train/loss': 0.2010958343744278, 'train/grad_norm_pre_clip': 13.289092063903809, 'timing/data': 0.006503013079054654, 'timing/model': 1.1492682249518111, 'learning_rate/qwen_vl_interface': 9.982218532964963e-06, 'learning_rate/action_model': 9.982218532964964e-05, 'epoch': 8.02}) +[2026-06-20 19:32:42,240][starVLA.training.train_starvla][INFO] - Step 4336, Loss: {'train/loss': 0.20972922444343567, 'train/grad_norm_pre_clip': 21.977680206298828, 'timing/data': 0.006517310976050794, 'timing/model': 1.1621428590733558, 'learning_rate/qwen_vl_interface': 9.981954295703152e-06, 'learning_rate/action_model': 9.981954295703152e-05, 'epoch': 8.02}) +[2026-06-20 19:32:46,947][starVLA.training.train_starvla][INFO] - Step 4337, Loss: {'train/loss': 0.10371804982423782, 'train/grad_norm_pre_clip': 16.017192840576172, 'timing/data': 0.006413711933419108, 'timing/model': 1.1415944750187919, 'learning_rate/qwen_vl_interface': 9.981688113313672e-06, 'learning_rate/action_model': 9.981688113313672e-05, 'epoch': 8.02}) +[2026-06-20 19:32:51,600][starVLA.training.train_starvla][INFO] - Step 4338, Loss: {'train/loss': 0.48085951805114746, 'train/grad_norm_pre_clip': 14.547931671142578, 'timing/data': 0.00790415401570499, 'timing/model': 1.3879846410127357, 'learning_rate/qwen_vl_interface': 9.98141998590594e-06, 'learning_rate/action_model': 9.981419985905939e-05, 'epoch': 8.02}) +[2026-06-20 19:32:56,055][starVLA.training.train_starvla][INFO] - Step 4339, Loss: {'train/loss': 0.24865783751010895, 'train/grad_norm_pre_clip': 9.019501686096191, 'timing/data': 0.0061841499991714954, 'timing/model': 1.1556559060700238, 'learning_rate/qwen_vl_interface': 9.981149913590171e-06, 'learning_rate/action_model': 9.981149913590172e-05, 'epoch': 8.02}) +[2026-06-20 19:33:00,735][starVLA.training.train_starvla][INFO] - Step 4340, Loss: {'train/loss': 0.2956482172012329, 'train/grad_norm_pre_clip': 14.753483772277832, 'timing/data': 0.007637083064764738, 'timing/model': 1.150752171059139, 'learning_rate/qwen_vl_interface': 9.980877896477386e-06, 'learning_rate/action_model': 9.980877896477386e-05, 'epoch': 8.03}) +[2026-06-20 19:33:05,387][starVLA.training.train_starvla][INFO] - Step 4341, Loss: {'train/loss': 0.015283222310245037, 'train/grad_norm_pre_clip': 3.053272008895874, 'timing/data': 0.005732019082643092, 'timing/model': 1.1519458059919998, 'learning_rate/qwen_vl_interface': 9.980603934679397e-06, 'learning_rate/action_model': 9.980603934679398e-05, 'epoch': 8.03}) +[2026-06-20 19:33:10,050][starVLA.training.train_starvla][INFO] - Step 4342, Loss: {'train/loss': 0.5457831025123596, 'train/grad_norm_pre_clip': 30.23748207092285, 'timing/data': 0.006158760050311685, 'timing/model': 1.1502854869468138, 'learning_rate/qwen_vl_interface': 9.980328028308824e-06, 'learning_rate/action_model': 9.980328028308824e-05, 'epoch': 8.03}) +[2026-06-20 19:33:14,744][starVLA.training.train_starvla][INFO] - Step 4343, Loss: {'train/loss': 0.0510367713868618, 'train/grad_norm_pre_clip': 10.51482105255127, 'timing/data': 0.007734862039797008, 'timing/model': 1.159046835033223, 'learning_rate/qwen_vl_interface': 9.980050177479075e-06, 'learning_rate/action_model': 9.980050177479076e-05, 'epoch': 8.03}) +[2026-06-20 19:33:19,442][starVLA.training.train_starvla][INFO] - Step 4344, Loss: {'train/loss': 0.17460250854492188, 'train/grad_norm_pre_clip': 18.97797203063965, 'timing/data': 0.005674957996234298, 'timing/model': 1.3713057420682162, 'learning_rate/qwen_vl_interface': 9.97977038230437e-06, 'learning_rate/action_model': 9.97977038230437e-05, 'epoch': 8.03}) +[2026-06-20 19:33:24,407][starVLA.training.train_starvla][INFO] - Step 4345, Loss: {'train/loss': 0.4931924045085907, 'train/grad_norm_pre_clip': 20.819210052490234, 'timing/data': 0.005645148106850684, 'timing/model': 1.3171936420258135, 'learning_rate/qwen_vl_interface': 9.97948864289972e-06, 'learning_rate/action_model': 9.97948864289972e-05, 'epoch': 8.04}) +[2026-06-20 19:33:29,124][starVLA.training.train_starvla][INFO] - Step 4346, Loss: {'train/loss': 0.20502161979675293, 'train/grad_norm_pre_clip': 17.416492462158203, 'timing/data': 0.006437259027734399, 'timing/model': 1.1420653309905902, 'learning_rate/qwen_vl_interface': 9.979204959380936e-06, 'learning_rate/action_model': 9.979204959380936e-05, 'epoch': 8.04}) +[2026-06-20 19:33:34,007][starVLA.training.train_starvla][INFO] - Step 4347, Loss: {'train/loss': 0.06263545155525208, 'train/grad_norm_pre_clip': 13.979689598083496, 'timing/data': 0.00861369096674025, 'timing/model': 1.2144540249137208, 'learning_rate/qwen_vl_interface': 9.978919331864631e-06, 'learning_rate/action_model': 9.97891933186463e-05, 'epoch': 8.04}) +[2026-06-20 19:33:38,671][starVLA.training.train_starvla][INFO] - Step 4348, Loss: {'train/loss': 0.019303344190120697, 'train/grad_norm_pre_clip': 2.660562515258789, 'timing/data': 0.00986625300720334, 'timing/model': 1.1641815518960357, 'learning_rate/qwen_vl_interface': 9.978631760468216e-06, 'learning_rate/action_model': 9.978631760468216e-05, 'epoch': 8.04}) +[2026-06-20 19:33:43,307][starVLA.training.train_starvla][INFO] - Step 4349, Loss: {'train/loss': 0.10839267075061798, 'train/grad_norm_pre_clip': 8.899578094482422, 'timing/data': 0.006247468059882522, 'timing/model': 1.1433630810352042, 'learning_rate/qwen_vl_interface': 9.9783422453099e-06, 'learning_rate/action_model': 9.9783422453099e-05, 'epoch': 8.04}) +[2026-06-20 19:33:47,943][starVLA.training.train_starvla][INFO] - Step 4350, Loss: {'train/loss': 0.2560398280620575, 'train/grad_norm_pre_clip': 10.152599334716797, 'timing/data': 0.006351977935992181, 'timing/model': 1.3751098091015592, 'learning_rate/qwen_vl_interface': 9.978050786508693e-06, 'learning_rate/action_model': 9.978050786508692e-05, 'epoch': 8.04}) +[2026-06-20 19:33:52,634][starVLA.training.train_starvla][INFO] - Step 4351, Loss: {'train/loss': 0.06373874843120575, 'train/grad_norm_pre_clip': 6.252761363983154, 'timing/data': 0.006377157988026738, 'timing/model': 1.3801861229585484, 'learning_rate/qwen_vl_interface': 9.9777573841844e-06, 'learning_rate/action_model': 9.977757384184401e-05, 'epoch': 8.05}) +[2026-06-20 19:33:57,041][starVLA.training.train_starvla][INFO] - Step 4352, Loss: {'train/loss': 0.049075208604335785, 'train/grad_norm_pre_clip': 4.06651496887207, 'timing/data': 0.006601657019928098, 'timing/model': 1.1418855270603672, 'learning_rate/qwen_vl_interface': 9.97746203845763e-06, 'learning_rate/action_model': 9.97746203845763e-05, 'epoch': 8.05}) +[2026-06-20 19:34:01,694][starVLA.training.train_starvla][INFO] - Step 4353, Loss: {'train/loss': 0.05496586486697197, 'train/grad_norm_pre_clip': 4.342860221862793, 'timing/data': 0.006072676973417401, 'timing/model': 1.1452922689495608, 'learning_rate/qwen_vl_interface': 9.97716474944979e-06, 'learning_rate/action_model': 9.977164749449788e-05, 'epoch': 8.05}) +[2026-06-20 19:34:06,387][starVLA.training.train_starvla][INFO] - Step 4354, Loss: {'train/loss': 0.05602854862809181, 'train/grad_norm_pre_clip': 5.69429349899292, 'timing/data': 0.00614678708370775, 'timing/model': 1.1470147290965542, 'learning_rate/qwen_vl_interface': 9.976865517283079e-06, 'learning_rate/action_model': 9.976865517283078e-05, 'epoch': 8.05}) +[2026-06-20 19:34:11,036][starVLA.training.train_starvla][INFO] - Step 4355, Loss: {'train/loss': 0.03850453346967697, 'train/grad_norm_pre_clip': 3.5547568798065186, 'timing/data': 0.0061892259400337934, 'timing/model': 1.1457397679332644, 'learning_rate/qwen_vl_interface': 9.976564342080505e-06, 'learning_rate/action_model': 9.976564342080504e-05, 'epoch': 8.05}) +[2026-06-20 19:34:15,722][starVLA.training.train_starvla][INFO] - Step 4356, Loss: {'train/loss': 0.11361601948738098, 'train/grad_norm_pre_clip': 9.317627906799316, 'timing/data': 0.006255766027607024, 'timing/model': 1.3902717089513317, 'learning_rate/qwen_vl_interface': 9.976261223965867e-06, 'learning_rate/action_model': 9.976261223965865e-05, 'epoch': 8.06}) +[2026-06-20 19:34:20,164][starVLA.training.train_starvla][INFO] - Step 4357, Loss: {'train/loss': 0.05396665632724762, 'train/grad_norm_pre_clip': 6.109353065490723, 'timing/data': 0.00641538598574698, 'timing/model': 1.1452506670029834, 'learning_rate/qwen_vl_interface': 9.975956163063767e-06, 'learning_rate/action_model': 9.975956163063767e-05, 'epoch': 8.06}) +[2026-06-20 19:34:24,871][starVLA.training.train_starvla][INFO] - Step 4358, Loss: {'train/loss': 0.1345222443342209, 'train/grad_norm_pre_clip': 7.730863571166992, 'timing/data': 0.0061362560372799635, 'timing/model': 1.1465035800356418, 'learning_rate/qwen_vl_interface': 9.975649159499605e-06, 'learning_rate/action_model': 9.975649159499603e-05, 'epoch': 8.06}) +[2026-06-20 19:34:29,556][starVLA.training.train_starvla][INFO] - Step 4359, Loss: {'train/loss': 0.030117182061076164, 'train/grad_norm_pre_clip': 3.371103525161743, 'timing/data': 0.00605584611184895, 'timing/model': 1.1463519900571555, 'learning_rate/qwen_vl_interface': 9.975340213399573e-06, 'learning_rate/action_model': 9.975340213399573e-05, 'epoch': 8.06}) +[2026-06-20 19:34:34,254][starVLA.training.train_starvla][INFO] - Step 4360, Loss: {'train/loss': 0.16321180760860443, 'train/grad_norm_pre_clip': 10.972719192504883, 'timing/data': 0.0064947259379550815, 'timing/model': 1.1423823470249772, 'learning_rate/qwen_vl_interface': 9.975029324890674e-06, 'learning_rate/action_model': 9.975029324890674e-05, 'epoch': 8.06}) +[2026-06-20 19:34:38,911][starVLA.training.train_starvla][INFO] - Step 4361, Loss: {'train/loss': 0.1700916737318039, 'train/grad_norm_pre_clip': 15.117446899414062, 'timing/data': 0.006320385029539466, 'timing/model': 1.1642484880285338, 'learning_rate/qwen_vl_interface': 9.974716494100698e-06, 'learning_rate/action_model': 9.974716494100697e-05, 'epoch': 8.07}) +[2026-06-20 19:34:43,556][starVLA.training.train_starvla][INFO] - Step 4362, Loss: {'train/loss': 0.3793851435184479, 'train/grad_norm_pre_clip': 13.350830078125, 'timing/data': 0.006246184930205345, 'timing/model': 1.3500821019988507, 'learning_rate/qwen_vl_interface': 9.97440172115824e-06, 'learning_rate/action_model': 9.974401721158239e-05, 'epoch': 8.07}) +[2026-06-20 19:34:48,015][starVLA.training.train_starvla][INFO] - Step 4363, Loss: {'train/loss': 0.11074017733335495, 'train/grad_norm_pre_clip': 8.919941902160645, 'timing/data': 0.006361954961903393, 'timing/model': 1.1472778089810163, 'learning_rate/qwen_vl_interface': 9.97408500619269e-06, 'learning_rate/action_model': 9.97408500619269e-05, 'epoch': 8.07}) +[2026-06-20 19:34:52,731][starVLA.training.train_starvla][INFO] - Step 4364, Loss: {'train/loss': 0.14151005446910858, 'train/grad_norm_pre_clip': 6.083970069885254, 'timing/data': 0.006221345043741167, 'timing/model': 1.1642536240397021, 'learning_rate/qwen_vl_interface': 9.973766349334239e-06, 'learning_rate/action_model': 9.973766349334239e-05, 'epoch': 8.07}) +[2026-06-20 19:34:57,399][starVLA.training.train_starvla][INFO] - Step 4365, Loss: {'train/loss': 0.05934492126107216, 'train/grad_norm_pre_clip': 10.916203498840332, 'timing/data': 0.006281225010752678, 'timing/model': 1.1474106319947168, 'learning_rate/qwen_vl_interface': 9.973445750713875e-06, 'learning_rate/action_model': 9.973445750713875e-05, 'epoch': 8.07}) +[2026-06-20 19:35:02,086][starVLA.training.train_starvla][INFO] - Step 4366, Loss: {'train/loss': 0.06779897212982178, 'train/grad_norm_pre_clip': 7.456961631774902, 'timing/data': 0.006218774011358619, 'timing/model': 1.1465743250446394, 'learning_rate/qwen_vl_interface': 9.97312321046338e-06, 'learning_rate/action_model': 9.97312321046338e-05, 'epoch': 8.07}) +[2026-06-20 19:35:06,748][starVLA.training.train_starvla][INFO] - Step 4367, Loss: {'train/loss': 0.16799722611904144, 'train/grad_norm_pre_clip': 14.21325397491455, 'timing/data': 0.006206133984960616, 'timing/model': 1.1453093079617247, 'learning_rate/qwen_vl_interface': 9.972798728715343e-06, 'learning_rate/action_model': 9.972798728715344e-05, 'epoch': 8.08}) +[2026-06-20 19:35:11,504][starVLA.training.train_starvla][INFO] - Step 4368, Loss: {'train/loss': 0.18011456727981567, 'train/grad_norm_pre_clip': 14.485188484191895, 'timing/data': 0.0064800039399415255, 'timing/model': 1.3527432520641014, 'learning_rate/qwen_vl_interface': 9.972472305603145e-06, 'learning_rate/action_model': 9.972472305603144e-05, 'epoch': 8.08}) +[2026-06-20 19:35:16,075][starVLA.training.train_starvla][INFO] - Step 4369, Loss: {'train/loss': 0.152383491396904, 'train/grad_norm_pre_clip': 5.294262886047363, 'timing/data': 0.006298064021393657, 'timing/model': 1.1470719769131392, 'learning_rate/qwen_vl_interface': 9.972143941260965e-06, 'learning_rate/action_model': 9.972143941260965e-05, 'epoch': 8.08}) +[2026-06-20 19:35:20,912][starVLA.training.train_starvla][INFO] - Step 4370, Loss: {'train/loss': 0.10025378316640854, 'train/grad_norm_pre_clip': 13.1835355758667, 'timing/data': 0.006367323105223477, 'timing/model': 1.1837396209593862, 'learning_rate/qwen_vl_interface': 9.971813635823781e-06, 'learning_rate/action_model': 9.971813635823781e-05, 'epoch': 8.08}) +[2026-06-20 19:35:25,612][starVLA.training.train_starvla][INFO] - Step 4371, Loss: {'train/loss': 0.08179760724306107, 'train/grad_norm_pre_clip': 6.761263370513916, 'timing/data': 0.006460343021899462, 'timing/model': 1.146222084062174, 'learning_rate/qwen_vl_interface': 9.971481389427372e-06, 'learning_rate/action_model': 9.971481389427371e-05, 'epoch': 8.08}) +[2026-06-20 19:35:30,352][starVLA.training.train_starvla][INFO] - Step 4372, Loss: {'train/loss': 0.024288447573781013, 'train/grad_norm_pre_clip': 2.076237201690674, 'timing/data': 0.006407544016838074, 'timing/model': 1.1530959820374846, 'learning_rate/qwen_vl_interface': 9.971147202208308e-06, 'learning_rate/action_model': 9.971147202208308e-05, 'epoch': 8.09}) +[2026-06-20 19:35:35,043][starVLA.training.train_starvla][INFO] - Step 4373, Loss: {'train/loss': 0.03434281051158905, 'train/grad_norm_pre_clip': 4.084981441497803, 'timing/data': 0.006409222958609462, 'timing/model': 1.4190111649222672, 'learning_rate/qwen_vl_interface': 9.970811074303962e-06, 'learning_rate/action_model': 9.970811074303962e-05, 'epoch': 8.09}) +[2026-06-20 19:35:39,599][starVLA.training.train_starvla][INFO] - Step 4374, Loss: {'train/loss': 0.11593394726514816, 'train/grad_norm_pre_clip': 3.534200668334961, 'timing/data': 0.006518763955682516, 'timing/model': 1.1502795809647068, 'learning_rate/qwen_vl_interface': 9.970473005852505e-06, 'learning_rate/action_model': 9.970473005852505e-05, 'epoch': 8.09}) +[2026-06-20 19:35:44,412][starVLA.training.train_starvla][INFO] - Step 4375, Loss: {'train/loss': 0.27505800127983093, 'train/grad_norm_pre_clip': 8.073278427124023, 'timing/data': 0.008271403959952295, 'timing/model': 1.1695842039771378, 'learning_rate/qwen_vl_interface': 9.970132996992904e-06, 'learning_rate/action_model': 9.970132996992904e-05, 'epoch': 8.09}) +[2026-06-20 19:35:49,128][starVLA.training.train_starvla][INFO] - Step 4376, Loss: {'train/loss': 0.1170865073800087, 'train/grad_norm_pre_clip': 11.56924057006836, 'timing/data': 0.006534553016535938, 'timing/model': 1.1464560909662396, 'learning_rate/qwen_vl_interface': 9.969791047864921e-06, 'learning_rate/action_model': 9.969791047864921e-05, 'epoch': 8.09}) +[2026-06-20 19:35:53,864][starVLA.training.train_starvla][INFO] - Step 4377, Loss: {'train/loss': 0.06884219497442245, 'train/grad_norm_pre_clip': 8.842732429504395, 'timing/data': 0.007582201971672475, 'timing/model': 1.2179152939934283, 'learning_rate/qwen_vl_interface': 9.969447158609122e-06, 'learning_rate/action_model': 9.96944715860912e-05, 'epoch': 8.09}) +[2026-06-20 19:35:58,684][starVLA.training.train_starvla][INFO] - Step 4378, Loss: {'train/loss': 0.3584221303462982, 'train/grad_norm_pre_clip': 20.959693908691406, 'timing/data': 0.0066346320090815425, 'timing/model': 1.1993239010917023, 'learning_rate/qwen_vl_interface': 9.969101329366864e-06, 'learning_rate/action_model': 9.969101329366863e-05, 'epoch': 8.1}) +[2026-06-20 19:36:03,462][starVLA.training.train_starvla][INFO] - Step 4379, Loss: {'train/loss': 0.02951692044734955, 'train/grad_norm_pre_clip': 3.4038400650024414, 'timing/data': 0.006601462024264038, 'timing/model': 1.384937901981175, 'learning_rate/qwen_vl_interface': 9.968753560280305e-06, 'learning_rate/action_model': 9.968753560280304e-05, 'epoch': 8.1}) +[2026-06-20 19:36:08,025][starVLA.training.train_starvla][INFO] - Step 4380, Loss: {'train/loss': 0.02475089021027088, 'train/grad_norm_pre_clip': 1.493574857711792, 'timing/data': 0.006506451987661421, 'timing/model': 1.144396607996896, 'learning_rate/qwen_vl_interface': 9.968403851492402e-06, 'learning_rate/action_model': 9.968403851492401e-05, 'epoch': 8.1}) +[2026-06-20 19:36:12,735][starVLA.training.train_starvla][INFO] - Step 4381, Loss: {'train/loss': 0.1529947966337204, 'train/grad_norm_pre_clip': 6.738986968994141, 'timing/data': 0.006493820925243199, 'timing/model': 1.1463406860129908, 'learning_rate/qwen_vl_interface': 9.968052203146904e-06, 'learning_rate/action_model': 9.968052203146903e-05, 'epoch': 8.1}) +[2026-06-20 19:36:17,592][starVLA.training.train_starvla][INFO] - Step 4382, Loss: {'train/loss': 0.022529030218720436, 'train/grad_norm_pre_clip': 1.4948912858963013, 'timing/data': 0.006437842035666108, 'timing/model': 1.2182467579841614, 'learning_rate/qwen_vl_interface': 9.967698615388363e-06, 'learning_rate/action_model': 9.967698615388362e-05, 'epoch': 8.1}) +[2026-06-20 19:36:22,391][starVLA.training.train_starvla][INFO] - Step 4383, Loss: {'train/loss': 0.0864134207367897, 'train/grad_norm_pre_clip': 7.805360794067383, 'timing/data': 0.006591081037186086, 'timing/model': 1.145090409903787, 'learning_rate/qwen_vl_interface': 9.967343088362123e-06, 'learning_rate/action_model': 9.967343088362122e-05, 'epoch': 8.11}) +[2026-06-20 19:36:27,257][starVLA.training.train_starvla][INFO] - Step 4384, Loss: {'train/loss': 0.03441431373357773, 'train/grad_norm_pre_clip': 5.204789161682129, 'timing/data': 0.007781471009366214, 'timing/model': 1.1984170150244609, 'learning_rate/qwen_vl_interface': 9.966985622214328e-06, 'learning_rate/action_model': 9.966985622214329e-05, 'epoch': 8.11}) +[2026-06-20 19:36:32,028][starVLA.training.train_starvla][INFO] - Step 4385, Loss: {'train/loss': 0.22034431993961334, 'train/grad_norm_pre_clip': 9.962394714355469, 'timing/data': 0.006290180957876146, 'timing/model': 1.3780886770691723, 'learning_rate/qwen_vl_interface': 9.966626217091922e-06, 'learning_rate/action_model': 9.966626217091922e-05, 'epoch': 8.11}) +[2026-06-20 19:36:36,601][starVLA.training.train_starvla][INFO] - Step 4386, Loss: {'train/loss': 0.02633712999522686, 'train/grad_norm_pre_clip': 4.395434379577637, 'timing/data': 0.0067163510248064995, 'timing/model': 1.1972286109812558, 'learning_rate/qwen_vl_interface': 9.966264873142641e-06, 'learning_rate/action_model': 9.96626487314264e-05, 'epoch': 8.11}) +[2026-06-20 19:36:41,454][starVLA.training.train_starvla][INFO] - Step 4387, Loss: {'train/loss': 0.020994799211621284, 'train/grad_norm_pre_clip': 3.182386875152588, 'timing/data': 0.00820923107676208, 'timing/model': 1.2143245589686558, 'learning_rate/qwen_vl_interface': 9.965901590515018e-06, 'learning_rate/action_model': 9.965901590515018e-05, 'epoch': 8.11}) +[2026-06-20 19:36:46,240][starVLA.training.train_starvla][INFO] - Step 4388, Loss: {'train/loss': 0.2531397044658661, 'train/grad_norm_pre_clip': 11.136160850524902, 'timing/data': 0.006369719980284572, 'timing/model': 1.22036210692022, 'learning_rate/qwen_vl_interface': 9.965536369358386e-06, 'learning_rate/action_model': 9.965536369358386e-05, 'epoch': 8.12}) +[2026-06-20 19:36:51,064][starVLA.training.train_starvla][INFO] - Step 4389, Loss: {'train/loss': 0.06455373764038086, 'train/grad_norm_pre_clip': 9.556753158569336, 'timing/data': 0.006390800001099706, 'timing/model': 1.1441786929499358, 'learning_rate/qwen_vl_interface': 9.965169209822874e-06, 'learning_rate/action_model': 9.965169209822874e-05, 'epoch': 8.12}) +[2026-06-20 19:36:55,964][starVLA.training.train_starvla][INFO] - Step 4390, Loss: {'train/loss': 0.11247032135725021, 'train/grad_norm_pre_clip': 11.010516166687012, 'timing/data': 0.007427140022628009, 'timing/model': 1.4292873119702563, 'learning_rate/qwen_vl_interface': 9.96480011205941e-06, 'learning_rate/action_model': 9.964800112059409e-05, 'epoch': 8.12}) +[2026-06-20 19:37:00,608][starVLA.training.train_starvla][INFO] - Step 4391, Loss: {'train/loss': 0.4524703919887543, 'train/grad_norm_pre_clip': 13.400038719177246, 'timing/data': 0.005729358992539346, 'timing/model': 1.2970142170088366, 'learning_rate/qwen_vl_interface': 9.964429076219711e-06, 'learning_rate/action_model': 9.964429076219711e-05, 'epoch': 8.12}) +[2026-06-20 19:37:05,218][starVLA.training.train_starvla][INFO] - Step 4392, Loss: {'train/loss': 0.00777288805693388, 'train/grad_norm_pre_clip': 1.1061501502990723, 'timing/data': 0.008168009924702346, 'timing/model': 1.2203376189572737, 'learning_rate/qwen_vl_interface': 9.964056102456299e-06, 'learning_rate/action_model': 9.964056102456299e-05, 'epoch': 8.12}) +[2026-06-20 19:37:09,962][starVLA.training.train_starvla][INFO] - Step 4393, Loss: {'train/loss': 0.3413389027118683, 'train/grad_norm_pre_clip': 15.525343894958496, 'timing/data': 0.006402719998732209, 'timing/model': 1.1455365249421448, 'learning_rate/qwen_vl_interface': 9.963681190922489e-06, 'learning_rate/action_model': 9.963681190922488e-05, 'epoch': 8.12}) +[2026-06-20 19:37:14,787][starVLA.training.train_starvla][INFO] - Step 4394, Loss: {'train/loss': 0.2933536469936371, 'train/grad_norm_pre_clip': 21.015180587768555, 'timing/data': 0.007304438971914351, 'timing/model': 1.2016215299954638, 'learning_rate/qwen_vl_interface': 9.963304341772391e-06, 'learning_rate/action_model': 9.963304341772391e-05, 'epoch': 8.13}) +[2026-06-20 19:37:19,563][starVLA.training.train_starvla][INFO] - Step 4395, Loss: {'train/loss': 0.11146827042102814, 'train/grad_norm_pre_clip': 9.221053123474121, 'timing/data': 0.0065116999903693795, 'timing/model': 1.1470733160385862, 'learning_rate/qwen_vl_interface': 9.962925555160919e-06, 'learning_rate/action_model': 9.962925555160918e-05, 'epoch': 8.13}) +[2026-06-20 19:37:24,412][starVLA.training.train_starvla][INFO] - Step 4396, Loss: {'train/loss': 0.01686951145529747, 'train/grad_norm_pre_clip': 1.9541116952896118, 'timing/data': 0.006380119011737406, 'timing/model': 1.3911981889978051, 'learning_rate/qwen_vl_interface': 9.96254483124377e-06, 'learning_rate/action_model': 9.96254483124377e-05, 'epoch': 8.13}) +[2026-06-20 19:37:29,273][starVLA.training.train_starvla][INFO] - Step 4397, Loss: {'train/loss': 0.0915181040763855, 'train/grad_norm_pre_clip': 12.822543144226074, 'timing/data': 0.009081238065846264, 'timing/model': 1.3828466149279848, 'learning_rate/qwen_vl_interface': 9.96216217017745e-06, 'learning_rate/action_model': 9.962162170177451e-05, 'epoch': 8.13}) +[2026-06-20 19:37:33,899][starVLA.training.train_starvla][INFO] - Step 4398, Loss: {'train/loss': 0.04619342461228371, 'train/grad_norm_pre_clip': 8.671952247619629, 'timing/data': 0.0069881490198895335, 'timing/model': 1.2167555170599371, 'learning_rate/qwen_vl_interface': 9.961777572119258e-06, 'learning_rate/action_model': 9.961777572119258e-05, 'epoch': 8.13}) +[2026-06-20 19:37:38,720][starVLA.training.train_starvla][INFO] - Step 4399, Loss: {'train/loss': 0.14195211231708527, 'train/grad_norm_pre_clip': 6.410172462463379, 'timing/data': 0.007663037977181375, 'timing/model': 1.2032521800138056, 'learning_rate/qwen_vl_interface': 9.961391037227286e-06, 'learning_rate/action_model': 9.961391037227284e-05, 'epoch': 8.14}) +[2026-06-20 19:37:43,467][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 4400 +[2026-06-20 19:44:18,117][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 4400 +[2026-06-20 19:44:18,118][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 4400 +[2026-06-20 19:51:14,301][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 4400 +[2026-06-20 19:51:14,303][starVLA.training.train_starvla][INFO] - Step 4400, Loss: {'train/loss': 0.023611510172486305, 'train/grad_norm_pre_clip': 3.981018543243408, 'eval/loss': 0.7011892199516296, 'eval/latency_0/loss': 0.8029714822769165, 'eval/latency_1/loss': 0.9650476574897766, 'eval/latency_2/loss': 1.0696173906326294, 'eval/latency_3/loss': 0.564661979675293, 'eval/latency_4/loss': 0.0747406929731369, 'eval/flappy/loss': 0.7011892199516296, 'eval/flappy/latency_0/loss': 0.8029714822769165, 'eval/flappy/latency_1/loss': 0.9650476574897766, 'eval/flappy/latency_2/loss': 1.0696173906326294, 'eval/flappy/latency_3/loss': 0.564661979675293, 'eval/flappy/latency_4/loss': 0.0747406929731369, 'eval/flap_precision': 0.1115702479338843, 'eval/flap_recall': 0.09818181818181818, 'eval/flap_f1': 0.1044487427466151, 'eval/flap_rate_model': 0.06722222222222222, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.1044487427466151, 'eval/latency_0/flap_precision': 0.1115702479338843, 'eval/latency_0/flap_recall': 0.09818181818181818, 'eval/latency_0/flap_f1': 0.1044487427466151, 'eval/latency_0/flap_rate_model': 0.06722222222222222, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.1044487427466151, 'timing/data': 0.00628433795645833, 'timing/model': 1.1470261649228632, 'learning_rate/qwen_vl_interface': 9.961002565660419e-06, 'learning_rate/action_model': 9.961002565660419e-05, 'epoch': 8.14}) +[2026-06-20 19:51:14,304][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_4400_state +[2026-06-20 19:51:14,308][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-20 19:51:22,928][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_4400_state/model.safetensors +[2026-06-20 19:51:36,455][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_4400_state/optimizer.bin +[2026-06-20 19:51:36,455][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_4400_state/sampler.bin +[2026-06-20 19:51:36,455][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_4400_state/sampler_1.bin +[2026-06-20 19:51:36,456][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_4400_state/random_states_0.pkl +[2026-06-20 19:51:39,106][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4400 +[2026-06-20 19:51:39,107][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-20 19:51:39,120][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-20 19:51:43,736][starVLA.training.train_starvla][INFO] - Step 4401, Loss: {'train/loss': 0.05372216925024986, 'train/grad_norm_pre_clip': 8.642046928405762, 'timing/data': 0.006428134976886213, 'timing/model': 1.1252884160494432, 'learning_rate/qwen_vl_interface': 9.96061215757835e-06, 'learning_rate/action_model': 9.960612157578349e-05, 'epoch': 8.14}) +[2026-06-20 19:51:48,436][starVLA.training.train_starvla][INFO] - Step 4402, Loss: {'train/loss': 0.036426790058612823, 'train/grad_norm_pre_clip': 4.726864337921143, 'timing/data': 0.007489271927624941, 'timing/model': 1.200450395932421, 'learning_rate/qwen_vl_interface': 9.96021981314156e-06, 'learning_rate/action_model': 9.96021981314156e-05, 'epoch': 8.14}) +[2026-06-20 19:51:53,093][starVLA.training.train_starvla][INFO] - Step 4403, Loss: {'train/loss': 0.1112494245171547, 'train/grad_norm_pre_clip': 9.398208618164062, 'timing/data': 0.006286885007284582, 'timing/model': 1.3763003709027544, 'learning_rate/qwen_vl_interface': 9.959825532511322e-06, 'learning_rate/action_model': 9.959825532511322e-05, 'epoch': 8.14}) +[2026-06-20 19:51:57,468][starVLA.training.train_starvla][INFO] - Step 4404, Loss: {'train/loss': 0.1697874665260315, 'train/grad_norm_pre_clip': 12.166671752929688, 'timing/data': 0.006505644996650517, 'timing/model': 1.1510611409321427, 'learning_rate/qwen_vl_interface': 9.959429315849715e-06, 'learning_rate/action_model': 9.959429315849715e-05, 'epoch': 8.14}) +[2026-06-20 19:52:02,104][starVLA.training.train_starvla][INFO] - Step 4405, Loss: {'train/loss': 0.16441954672336578, 'train/grad_norm_pre_clip': 13.631743431091309, 'timing/data': 0.006202835007570684, 'timing/model': 1.1518689439399168, 'learning_rate/qwen_vl_interface': 9.959031163319606e-06, 'learning_rate/action_model': 9.959031163319607e-05, 'epoch': 8.15}) +[2026-06-20 19:52:06,768][starVLA.training.train_starvla][INFO] - Step 4406, Loss: {'train/loss': 0.04513031989336014, 'train/grad_norm_pre_clip': 9.879146575927734, 'timing/data': 0.009740486973896623, 'timing/model': 1.1594386319629848, 'learning_rate/qwen_vl_interface': 9.958631075084664e-06, 'learning_rate/action_model': 9.958631075084664e-05, 'epoch': 8.15}) +[2026-06-20 19:52:11,487][starVLA.training.train_starvla][INFO] - Step 4407, Loss: {'train/loss': 0.07049227505922318, 'train/grad_norm_pre_clip': 13.845250129699707, 'timing/data': 0.008339100982993841, 'timing/model': 1.1740742509718984, 'learning_rate/qwen_vl_interface': 9.958229051309346e-06, 'learning_rate/action_model': 9.958229051309346e-05, 'epoch': 8.15}) +[2026-06-20 19:52:16,169][starVLA.training.train_starvla][INFO] - Step 4408, Loss: {'train/loss': 0.13336314260959625, 'train/grad_norm_pre_clip': 14.852300643920898, 'timing/data': 0.006379804923199117, 'timing/model': 1.4196851099841297, 'learning_rate/qwen_vl_interface': 9.95782509215891e-06, 'learning_rate/action_model': 9.95782509215891e-05, 'epoch': 8.15}) +[2026-06-20 19:52:20,590][starVLA.training.train_starvla][INFO] - Step 4409, Loss: {'train/loss': 0.3314436972141266, 'train/grad_norm_pre_clip': 14.430749893188477, 'timing/data': 0.006467364029958844, 'timing/model': 1.1538309609750286, 'learning_rate/qwen_vl_interface': 9.95741919779941e-06, 'learning_rate/action_model': 9.95741919779941e-05, 'epoch': 8.15}) +[2026-06-20 19:52:25,211][starVLA.training.train_starvla][INFO] - Step 4410, Loss: {'train/loss': 0.015320617705583572, 'train/grad_norm_pre_clip': 2.5886218547821045, 'timing/data': 0.006645963992923498, 'timing/model': 1.1391512900590897, 'learning_rate/qwen_vl_interface': 9.957011368397692e-06, 'learning_rate/action_model': 9.95701136839769e-05, 'epoch': 8.16}) +[2026-06-20 19:52:29,785][starVLA.training.train_starvla][INFO] - Step 4411, Loss: {'train/loss': 0.06634396314620972, 'train/grad_norm_pre_clip': 20.089187622070312, 'timing/data': 0.006462384015321732, 'timing/model': 1.1560256159864366, 'learning_rate/qwen_vl_interface': 9.956601604121398e-06, 'learning_rate/action_model': 9.956601604121397e-05, 'epoch': 8.16}) +[2026-06-20 19:52:34,449][starVLA.training.train_starvla][INFO] - Step 4412, Loss: {'train/loss': 0.17991243302822113, 'train/grad_norm_pre_clip': 13.234308242797852, 'timing/data': 0.006478543975390494, 'timing/model': 1.140066039049998, 'learning_rate/qwen_vl_interface': 9.95618990513897e-06, 'learning_rate/action_model': 9.95618990513897e-05, 'epoch': 8.16}) +[2026-06-20 19:52:39,097][starVLA.training.train_starvla][INFO] - Step 4413, Loss: {'train/loss': 0.41962286829948425, 'train/grad_norm_pre_clip': 12.587553024291992, 'timing/data': 0.006440875004045665, 'timing/model': 1.140399752999656, 'learning_rate/qwen_vl_interface': 9.95577627161964e-06, 'learning_rate/action_model': 9.95577627161964e-05, 'epoch': 8.16}) +[2026-06-20 19:52:43,741][starVLA.training.train_starvla][INFO] - Step 4414, Loss: {'train/loss': 0.10339724272489548, 'train/grad_norm_pre_clip': 19.589500427246094, 'timing/data': 0.006623834953643382, 'timing/model': 1.3862585040042177, 'learning_rate/qwen_vl_interface': 9.955360703733438e-06, 'learning_rate/action_model': 9.955360703733437e-05, 'epoch': 8.16}) +[2026-06-20 19:52:48,148][starVLA.training.train_starvla][INFO] - Step 4415, Loss: {'train/loss': 0.10209932178258896, 'train/grad_norm_pre_clip': 7.848212242126465, 'timing/data': 0.006446294020861387, 'timing/model': 1.149905162048526, 'learning_rate/qwen_vl_interface': 9.954943201651187e-06, 'learning_rate/action_model': 9.954943201651187e-05, 'epoch': 8.16}) +[2026-06-20 19:52:52,783][starVLA.training.train_starvla][INFO] - Step 4416, Loss: {'train/loss': 0.16987986862659454, 'train/grad_norm_pre_clip': 12.522212028503418, 'timing/data': 0.006275474093854427, 'timing/model': 1.139882370014675, 'learning_rate/qwen_vl_interface': 9.954523765544507e-06, 'learning_rate/action_model': 9.954523765544508e-05, 'epoch': 8.17}) +[2026-06-20 19:52:57,435][starVLA.training.train_starvla][INFO] - Step 4417, Loss: {'train/loss': 0.029010459780693054, 'train/grad_norm_pre_clip': 1.783809781074524, 'timing/data': 0.006198955001309514, 'timing/model': 1.1410923829535022, 'learning_rate/qwen_vl_interface': 9.954102395585814e-06, 'learning_rate/action_model': 9.954102395585813e-05, 'epoch': 8.17}) +[2026-06-20 19:53:02,098][starVLA.training.train_starvla][INFO] - Step 4418, Loss: {'train/loss': 0.02320871129631996, 'train/grad_norm_pre_clip': 2.976545810699463, 'timing/data': 0.006287735071964562, 'timing/model': 1.1425862850155681, 'learning_rate/qwen_vl_interface': 9.953679091948313e-06, 'learning_rate/action_model': 9.953679091948314e-05, 'epoch': 8.17}) +[2026-06-20 19:53:06,707][starVLA.training.train_starvla][INFO] - Step 4419, Loss: {'train/loss': 0.07080702483654022, 'train/grad_norm_pre_clip': 14.150575637817383, 'timing/data': 0.006304965005256236, 'timing/model': 1.1400549169629812, 'learning_rate/qwen_vl_interface': 9.953253854806013e-06, 'learning_rate/action_model': 9.953253854806012e-05, 'epoch': 8.17}) +[2026-06-20 19:53:11,364][starVLA.training.train_starvla][INFO] - Step 4420, Loss: {'train/loss': 0.21995829045772552, 'train/grad_norm_pre_clip': 9.953967094421387, 'timing/data': 0.009170747944153845, 'timing/model': 1.377449483028613, 'learning_rate/qwen_vl_interface': 9.95282668433371e-06, 'learning_rate/action_model': 9.952826684333709e-05, 'epoch': 8.17}) +[2026-06-20 19:53:15,803][starVLA.training.train_starvla][INFO] - Step 4421, Loss: {'train/loss': 0.007608562707901001, 'train/grad_norm_pre_clip': 1.2365813255310059, 'timing/data': 0.006416875054128468, 'timing/model': 1.1428070020629093, 'learning_rate/qwen_vl_interface': 9.952397580706998e-06, 'learning_rate/action_model': 9.952397580706998e-05, 'epoch': 8.18}) +[2026-06-20 19:53:20,495][starVLA.training.train_starvla][INFO] - Step 4422, Loss: {'train/loss': 0.09471944719552994, 'train/grad_norm_pre_clip': 8.887947082519531, 'timing/data': 0.006442434038035572, 'timing/model': 1.14036423410289, 'learning_rate/qwen_vl_interface': 9.951966544102267e-06, 'learning_rate/action_model': 9.951966544102266e-05, 'epoch': 8.18}) +[2026-06-20 19:53:25,208][starVLA.training.train_starvla][INFO] - Step 4423, Loss: {'train/loss': 0.10245125740766525, 'train/grad_norm_pre_clip': 7.800970554351807, 'timing/data': 0.006536273984238505, 'timing/model': 1.2049012830248103, 'learning_rate/qwen_vl_interface': 9.951533574696696e-06, 'learning_rate/action_model': 9.951533574696697e-05, 'epoch': 8.18}) +[2026-06-20 19:53:29,854][starVLA.training.train_starvla][INFO] - Step 4424, Loss: {'train/loss': 0.43928948044776917, 'train/grad_norm_pre_clip': 27.468955993652344, 'timing/data': 0.006443983991630375, 'timing/model': 1.1408460839884356, 'learning_rate/qwen_vl_interface': 9.951098672668266e-06, 'learning_rate/action_model': 9.951098672668265e-05, 'epoch': 8.18}) +[2026-06-20 19:53:34,542][starVLA.training.train_starvla][INFO] - Step 4425, Loss: {'train/loss': 0.0826035887002945, 'train/grad_norm_pre_clip': 13.216670989990234, 'timing/data': 0.006425403989851475, 'timing/model': 1.3845204809913412, 'learning_rate/qwen_vl_interface': 9.95066183819575e-06, 'learning_rate/action_model': 9.950661838195749e-05, 'epoch': 8.18}) +[2026-06-20 19:53:38,945][starVLA.training.train_starvla][INFO] - Step 4426, Loss: {'train/loss': 0.17701201140880585, 'train/grad_norm_pre_clip': 15.308146476745605, 'timing/data': 0.006544764037244022, 'timing/model': 1.1411164960591123, 'learning_rate/qwen_vl_interface': 9.95022307145871e-06, 'learning_rate/action_model': 9.950223071458708e-05, 'epoch': 8.19}) +[2026-06-20 19:53:43,580][starVLA.training.train_starvla][INFO] - Step 4427, Loss: {'train/loss': 0.050026245415210724, 'train/grad_norm_pre_clip': 7.5285797119140625, 'timing/data': 0.0063431450398638844, 'timing/model': 1.1415701790247113, 'learning_rate/qwen_vl_interface': 9.949782372637507e-06, 'learning_rate/action_model': 9.949782372637506e-05, 'epoch': 8.19}) +[2026-06-20 19:53:48,259][starVLA.training.train_starvla][INFO] - Step 4428, Loss: {'train/loss': 0.08637801557779312, 'train/grad_norm_pre_clip': 8.808393478393555, 'timing/data': 0.006488824030384421, 'timing/model': 1.1720820319605991, 'learning_rate/qwen_vl_interface': 9.949339741913297e-06, 'learning_rate/action_model': 9.949339741913297e-05, 'epoch': 8.19}) +[2026-06-20 19:53:52,945][starVLA.training.train_starvla][INFO] - Step 4429, Loss: {'train/loss': 0.08970550447702408, 'train/grad_norm_pre_clip': 10.265466690063477, 'timing/data': 0.006381685030646622, 'timing/model': 1.1592602789169177, 'learning_rate/qwen_vl_interface': 9.948895179468032e-06, 'learning_rate/action_model': 9.948895179468031e-05, 'epoch': 8.19}) +[2026-06-20 19:53:57,607][starVLA.training.train_starvla][INFO] - Step 4430, Loss: {'train/loss': 0.07160241156816483, 'train/grad_norm_pre_clip': 6.230748653411865, 'timing/data': 0.008606928982771933, 'timing/model': 1.1936804309953004, 'learning_rate/qwen_vl_interface': 9.948448685484448e-06, 'learning_rate/action_model': 9.948448685484447e-05, 'epoch': 8.19}) +[2026-06-20 19:54:02,389][starVLA.training.train_starvla][INFO] - Step 4431, Loss: {'train/loss': 0.24022990465164185, 'train/grad_norm_pre_clip': 9.995461463928223, 'timing/data': 0.00651885406114161, 'timing/model': 1.3846899010241032, 'learning_rate/qwen_vl_interface': 9.948000260146087e-06, 'learning_rate/action_model': 9.948000260146087e-05, 'epoch': 8.19}) +[2026-06-20 19:54:06,922][starVLA.training.train_starvla][INFO] - Step 4432, Loss: {'train/loss': 0.016817891970276833, 'train/grad_norm_pre_clip': 1.7816206216812134, 'timing/data': 0.0064083850011229515, 'timing/model': 1.142076549003832, 'learning_rate/qwen_vl_interface': 9.947549903637277e-06, 'learning_rate/action_model': 9.947549903637278e-05, 'epoch': 8.2}) +[2026-06-20 19:54:11,633][starVLA.training.train_starvla][INFO] - Step 4433, Loss: {'train/loss': 0.015096766874194145, 'train/grad_norm_pre_clip': 2.3070011138916016, 'timing/data': 0.0064215639140456915, 'timing/model': 1.1404611890902743, 'learning_rate/qwen_vl_interface': 9.947097616143145e-06, 'learning_rate/action_model': 9.947097616143144e-05, 'epoch': 8.2}) +[2026-06-20 19:54:16,329][starVLA.training.train_starvla][INFO] - Step 4434, Loss: {'train/loss': 0.16756314039230347, 'train/grad_norm_pre_clip': 10.016250610351562, 'timing/data': 0.007458881940692663, 'timing/model': 1.1966648480156437, 'learning_rate/qwen_vl_interface': 9.946643397849607e-06, 'learning_rate/action_model': 9.946643397849607e-05, 'epoch': 8.2}) +[2026-06-20 19:54:21,030][starVLA.training.train_starvla][INFO] - Step 4435, Loss: {'train/loss': 0.19412313401699066, 'train/grad_norm_pre_clip': 17.080158233642578, 'timing/data': 0.007712100981734693, 'timing/model': 1.1949066870147362, 'learning_rate/qwen_vl_interface': 9.946187248943375e-06, 'learning_rate/action_model': 9.946187248943374e-05, 'epoch': 8.2}) +[2026-06-20 19:54:25,826][starVLA.training.train_starvla][INFO] - Step 4436, Loss: {'train/loss': 0.06598611176013947, 'train/grad_norm_pre_clip': 13.45307731628418, 'timing/data': 0.006454013986513019, 'timing/model': 1.1412274959729984, 'learning_rate/qwen_vl_interface': 9.945729169611957e-06, 'learning_rate/action_model': 9.945729169611958e-05, 'epoch': 8.2}) +[2026-06-20 19:54:30,606][starVLA.training.train_starvla][INFO] - Step 4437, Loss: {'train/loss': 0.06247352436184883, 'train/grad_norm_pre_clip': 7.2753472328186035, 'timing/data': 0.00643989397212863, 'timing/model': 1.3784056779695675, 'learning_rate/qwen_vl_interface': 9.94526916004365e-06, 'learning_rate/action_model': 9.94526916004365e-05, 'epoch': 8.21}) +[2026-06-20 19:54:35,080][starVLA.training.train_starvla][INFO] - Step 4438, Loss: {'train/loss': 0.3365871012210846, 'train/grad_norm_pre_clip': 13.947773933410645, 'timing/data': 0.006367483991198242, 'timing/model': 1.1415723479585722, 'learning_rate/qwen_vl_interface': 9.944807220427549e-06, 'learning_rate/action_model': 9.944807220427548e-05, 'epoch': 8.21}) +[2026-06-20 19:54:39,724][starVLA.training.train_starvla][INFO] - Step 4439, Loss: {'train/loss': 0.26107537746429443, 'train/grad_norm_pre_clip': 20.74243927001953, 'timing/data': 0.006424274994060397, 'timing/model': 1.1414003539830446, 'learning_rate/qwen_vl_interface': 9.944343350953534e-06, 'learning_rate/action_model': 9.944343350953533e-05, 'epoch': 8.21}) +[2026-06-20 19:54:44,370][starVLA.training.train_starvla][INFO] - Step 4440, Loss: {'train/loss': 0.08952180296182632, 'train/grad_norm_pre_clip': 11.229055404663086, 'timing/data': 0.006412504008039832, 'timing/model': 1.141815340030007, 'learning_rate/qwen_vl_interface': 9.94387755181229e-06, 'learning_rate/action_model': 9.94387755181229e-05, 'epoch': 8.21}) +[2026-06-20 19:54:49,066][starVLA.training.train_starvla][INFO] - Step 4441, Loss: {'train/loss': 0.03659231588244438, 'train/grad_norm_pre_clip': 5.26188850402832, 'timing/data': 0.006468574050813913, 'timing/model': 1.1411493569612503, 'learning_rate/qwen_vl_interface': 9.943409823195289e-06, 'learning_rate/action_model': 9.943409823195288e-05, 'epoch': 8.21}) +[2026-06-20 19:54:53,740][starVLA.training.train_starvla][INFO] - Step 4442, Loss: {'train/loss': 0.08094097673892975, 'train/grad_norm_pre_clip': 11.020915031433105, 'timing/data': 0.006594123900867999, 'timing/model': 1.4179504279745743, 'learning_rate/qwen_vl_interface': 9.942940165294794e-06, 'learning_rate/action_model': 9.942940165294793e-05, 'epoch': 8.21}) +[2026-06-20 19:54:58,291][starVLA.training.train_starvla][INFO] - Step 4443, Loss: {'train/loss': 0.14228329062461853, 'train/grad_norm_pre_clip': 30.416282653808594, 'timing/data': 0.006660272949375212, 'timing/model': 1.292659503989853, 'learning_rate/qwen_vl_interface': 9.942468578303865e-06, 'learning_rate/action_model': 9.942468578303865e-05, 'epoch': 8.22}) +[2026-06-20 19:55:02,745][starVLA.training.train_starvla][INFO] - Step 4444, Loss: {'train/loss': 0.09847688674926758, 'train/grad_norm_pre_clip': 22.091651916503906, 'timing/data': 0.00641153403557837, 'timing/model': 1.1423097050283104, 'learning_rate/qwen_vl_interface': 9.941995062416354e-06, 'learning_rate/action_model': 9.941995062416354e-05, 'epoch': 8.22}) +[2026-06-20 19:55:07,475][starVLA.training.train_starvla][INFO] - Step 4445, Loss: {'train/loss': 0.18634143471717834, 'train/grad_norm_pre_clip': 16.621782302856445, 'timing/data': 0.005786765948869288, 'timing/model': 1.144597084959969, 'learning_rate/qwen_vl_interface': 9.941519617826906e-06, 'learning_rate/action_model': 9.941519617826906e-05, 'epoch': 8.22}) +[2026-06-20 19:55:12,132][starVLA.training.train_starvla][INFO] - Step 4446, Loss: {'train/loss': 0.08287541568279266, 'train/grad_norm_pre_clip': 12.496335983276367, 'timing/data': 0.005925415083765984, 'timing/model': 1.1436888539465144, 'learning_rate/qwen_vl_interface': 9.941042244730954e-06, 'learning_rate/action_model': 9.941042244730955e-05, 'epoch': 8.22}) +[2026-06-20 19:55:16,931][starVLA.training.train_starvla][INFO] - Step 4447, Loss: {'train/loss': 0.22743503749370575, 'train/grad_norm_pre_clip': 10.366501808166504, 'timing/data': 0.005764255998656154, 'timing/model': 1.2153933519730344, 'learning_rate/qwen_vl_interface': 9.940562943324736e-06, 'learning_rate/action_model': 9.940562943324735e-05, 'epoch': 8.22}) +[2026-06-20 19:55:21,681][starVLA.training.train_starvla][INFO] - Step 4448, Loss: {'train/loss': 0.1895667463541031, 'train/grad_norm_pre_clip': 13.10064697265625, 'timing/data': 0.0058281749952584505, 'timing/model': 1.3813955039950088, 'learning_rate/qwen_vl_interface': 9.940081713805266e-06, 'learning_rate/action_model': 9.940081713805266e-05, 'epoch': 8.23}) +[2026-06-20 19:55:26,294][starVLA.training.train_starvla][INFO] - Step 4449, Loss: {'train/loss': 0.23711268603801727, 'train/grad_norm_pre_clip': 7.069504261016846, 'timing/data': 0.0059729949571192265, 'timing/model': 1.3034810139797628, 'learning_rate/qwen_vl_interface': 9.939598556370366e-06, 'learning_rate/action_model': 9.939598556370365e-05, 'epoch': 8.23}) +[2026-06-20 19:55:30,827][starVLA.training.train_starvla][INFO] - Step 4450, Loss: {'train/loss': 0.3472776412963867, 'train/grad_norm_pre_clip': 13.951237678527832, 'timing/data': 0.0058435860555619, 'timing/model': 1.145299636060372, 'learning_rate/qwen_vl_interface': 9.93911347121864e-06, 'learning_rate/action_model': 9.939113471218639e-05, 'epoch': 8.23}) +[2026-06-20 19:55:35,519][starVLA.training.train_starvla][INFO] - Step 4451, Loss: {'train/loss': 0.3707866072654724, 'train/grad_norm_pre_clip': 17.031347274780273, 'timing/data': 0.00838431902229786, 'timing/model': 1.2140327000524849, 'learning_rate/qwen_vl_interface': 9.938626458549488e-06, 'learning_rate/action_model': 9.938626458549488e-05, 'epoch': 8.23}) +[2026-06-20 19:55:40,316][starVLA.training.train_starvla][INFO] - Step 4452, Loss: {'train/loss': 0.2914641797542572, 'train/grad_norm_pre_clip': 18.08636474609375, 'timing/data': 0.00740391097497195, 'timing/model': 1.1969943899894133, 'learning_rate/qwen_vl_interface': 9.938137518563106e-06, 'learning_rate/action_model': 9.938137518563106e-05, 'epoch': 8.23}) +[2026-06-20 19:55:45,142][starVLA.training.train_starvla][INFO] - Step 4453, Loss: {'train/loss': 0.08424364030361176, 'train/grad_norm_pre_clip': 7.752590179443359, 'timing/data': 0.006235143984667957, 'timing/model': 1.1442711590789258, 'learning_rate/qwen_vl_interface': 9.937646651460475e-06, 'learning_rate/action_model': 9.937646651460474e-05, 'epoch': 8.24}) +[2026-06-20 19:55:49,935][starVLA.training.train_starvla][INFO] - Step 4454, Loss: {'train/loss': 0.15262751281261444, 'train/grad_norm_pre_clip': 23.14274024963379, 'timing/data': 0.006432253983803093, 'timing/model': 1.3849374200217426, 'learning_rate/qwen_vl_interface': 9.937153857443375e-06, 'learning_rate/action_model': 9.937153857443374e-05, 'epoch': 8.24}) +[2026-06-20 19:55:54,532][starVLA.training.train_starvla][INFO] - Step 4455, Loss: {'train/loss': 0.2609495520591736, 'train/grad_norm_pre_clip': 17.026283264160156, 'timing/data': 0.006596704013645649, 'timing/model': 1.195302173960954, 'learning_rate/qwen_vl_interface': 9.936659136714372e-06, 'learning_rate/action_model': 9.936659136714371e-05, 'epoch': 8.24}) +[2026-06-20 19:55:59,363][starVLA.training.train_starvla][INFO] - Step 4456, Loss: {'train/loss': 0.221427321434021, 'train/grad_norm_pre_clip': 11.617563247680664, 'timing/data': 0.0063183329766616225, 'timing/model': 1.1443081389879808, 'learning_rate/qwen_vl_interface': 9.936162489476827e-06, 'learning_rate/action_model': 9.936162489476828e-05, 'epoch': 8.24}) +[2026-06-20 19:56:04,226][starVLA.training.train_starvla][INFO] - Step 4457, Loss: {'train/loss': 0.14710435271263123, 'train/grad_norm_pre_clip': 7.738301753997803, 'timing/data': 0.008905167924240232, 'timing/model': 1.2167284429306164, 'learning_rate/qwen_vl_interface': 9.935663915934899e-06, 'learning_rate/action_model': 9.935663915934898e-05, 'epoch': 8.24}) +[2026-06-20 19:56:09,037][starVLA.training.train_starvla][INFO] - Step 4458, Loss: {'train/loss': 0.12624314427375793, 'train/grad_norm_pre_clip': 7.396556854248047, 'timing/data': 0.006408783956430852, 'timing/model': 1.2107940249843523, 'learning_rate/qwen_vl_interface': 9.935163416293525e-06, 'learning_rate/action_model': 9.935163416293525e-05, 'epoch': 8.24}) +[2026-06-20 19:56:13,791][starVLA.training.train_starvla][INFO] - Step 4459, Loss: {'train/loss': 0.0708857849240303, 'train/grad_norm_pre_clip': 3.08019757270813, 'timing/data': 0.008127339999191463, 'timing/model': 1.1971878649201244, 'learning_rate/qwen_vl_interface': 9.934660990758445e-06, 'learning_rate/action_model': 9.934660990758445e-05, 'epoch': 8.25}) +[2026-06-20 19:56:18,588][starVLA.training.train_starvla][INFO] - Step 4460, Loss: {'train/loss': 0.019057132303714752, 'train/grad_norm_pre_clip': 1.3775498867034912, 'timing/data': 0.009688055026344955, 'timing/model': 1.4552531499648467, 'learning_rate/qwen_vl_interface': 9.934156639536186e-06, 'learning_rate/action_model': 9.934156639536187e-05, 'epoch': 8.25}) +[2026-06-20 19:56:23,167][starVLA.training.train_starvla][INFO] - Step 4461, Loss: {'train/loss': 0.05365155637264252, 'train/grad_norm_pre_clip': 3.590907335281372, 'timing/data': 0.008115608943626285, 'timing/model': 1.196919810026884, 'learning_rate/qwen_vl_interface': 9.93365036283407e-06, 'learning_rate/action_model': 9.93365036283407e-05, 'epoch': 8.25}) +[2026-06-20 19:56:27,863][starVLA.training.train_starvla][INFO] - Step 4462, Loss: {'train/loss': 0.08060527592897415, 'train/grad_norm_pre_clip': 7.865070819854736, 'timing/data': 0.006355843972414732, 'timing/model': 1.1434655430493876, 'learning_rate/qwen_vl_interface': 9.933142160860208e-06, 'learning_rate/action_model': 9.933142160860208e-05, 'epoch': 8.25}) +[2026-06-20 19:56:32,602][starVLA.training.train_starvla][INFO] - Step 4463, Loss: {'train/loss': 0.22697800397872925, 'train/grad_norm_pre_clip': 6.611399173736572, 'timing/data': 0.006400955026037991, 'timing/model': 1.2144163090270013, 'learning_rate/qwen_vl_interface': 9.932632033823499e-06, 'learning_rate/action_model': 9.932632033823499e-05, 'epoch': 8.25}) +[2026-06-20 19:56:37,524][starVLA.training.train_starvla][INFO] - Step 4464, Loss: {'train/loss': 0.12907898426055908, 'train/grad_norm_pre_clip': 10.416419982910156, 'timing/data': 0.009166167001239955, 'timing/model': 1.2143110160250217, 'learning_rate/qwen_vl_interface': 9.93211998193364e-06, 'learning_rate/action_model': 9.932119981933639e-05, 'epoch': 8.26}) +[2026-06-20 19:56:42,380][starVLA.training.train_starvla][INFO] - Step 4465, Loss: {'train/loss': 0.06116260588169098, 'train/grad_norm_pre_clip': 5.855795860290527, 'timing/data': 0.00747183000203222, 'timing/model': 1.2139373240061104, 'learning_rate/qwen_vl_interface': 9.931606005401117e-06, 'learning_rate/action_model': 9.931606005401116e-05, 'epoch': 8.26}) +[2026-06-20 19:56:47,123][starVLA.training.train_starvla][INFO] - Step 4466, Loss: {'train/loss': 0.17774276435375214, 'train/grad_norm_pre_clip': 9.287177085876465, 'timing/data': 0.006275244057178497, 'timing/model': 1.4062772609759122, 'learning_rate/qwen_vl_interface': 9.931090104437202e-06, 'learning_rate/action_model': 9.931090104437202e-05, 'epoch': 8.26}) +[2026-06-20 19:56:51,719][starVLA.training.train_starvla][INFO] - Step 4467, Loss: {'train/loss': 0.19554933905601501, 'train/grad_norm_pre_clip': 17.242353439331055, 'timing/data': 0.00888287799898535, 'timing/model': 1.3368737440323457, 'learning_rate/qwen_vl_interface': 9.930572279253967e-06, 'learning_rate/action_model': 9.930572279253967e-05, 'epoch': 8.26}) +[2026-06-20 19:56:56,148][starVLA.training.train_starvla][INFO] - Step 4468, Loss: {'train/loss': 0.2504594922065735, 'train/grad_norm_pre_clip': 8.818706512451172, 'timing/data': 0.00622956408187747, 'timing/model': 1.1437987349927425, 'learning_rate/qwen_vl_interface': 9.930052530064269e-06, 'learning_rate/action_model': 9.93005253006427e-05, 'epoch': 8.26}) +[2026-06-20 19:57:00,781][starVLA.training.train_starvla][INFO] - Step 4469, Loss: {'train/loss': 0.10283271223306656, 'train/grad_norm_pre_clip': 6.255605220794678, 'timing/data': 0.006288503995165229, 'timing/model': 1.1444563789991662, 'learning_rate/qwen_vl_interface': 9.929530857081757e-06, 'learning_rate/action_model': 9.929530857081757e-05, 'epoch': 8.26}) +[2026-06-20 19:57:05,437][starVLA.training.train_starvla][INFO] - Step 4470, Loss: {'train/loss': 0.013581987470388412, 'train/grad_norm_pre_clip': 1.233988881111145, 'timing/data': 0.006299293949268758, 'timing/model': 1.1434138690819964, 'learning_rate/qwen_vl_interface': 9.92900726052087e-06, 'learning_rate/action_model': 9.929007260520869e-05, 'epoch': 8.27}) +[2026-06-20 19:57:10,492][starVLA.training.train_starvla][INFO] - Step 4471, Loss: {'train/loss': 0.2800150513648987, 'train/grad_norm_pre_clip': 14.61103630065918, 'timing/data': 0.020802726969122887, 'timing/model': 1.1438016950851306, 'learning_rate/qwen_vl_interface': 9.92848174059684e-06, 'learning_rate/action_model': 9.92848174059684e-05, 'epoch': 8.27}) +[2026-06-20 19:57:15,153][starVLA.training.train_starvla][INFO] - Step 4472, Loss: {'train/loss': 0.38766446709632874, 'train/grad_norm_pre_clip': 20.728172302246094, 'timing/data': 0.007970159058459103, 'timing/model': 1.382204532972537, 'learning_rate/qwen_vl_interface': 9.92795429752569e-06, 'learning_rate/action_model': 9.927954297525689e-05, 'epoch': 8.27}) +[2026-06-20 19:57:19,794][starVLA.training.train_starvla][INFO] - Step 4473, Loss: {'train/loss': 0.17439553141593933, 'train/grad_norm_pre_clip': 11.01959228515625, 'timing/data': 0.006428303895518184, 'timing/model': 1.3808193129952997, 'learning_rate/qwen_vl_interface': 9.927424931524228e-06, 'learning_rate/action_model': 9.927424931524228e-05, 'epoch': 8.27}) +[2026-06-20 19:57:24,220][starVLA.training.train_starvla][INFO] - Step 4474, Loss: {'train/loss': 0.07681537419557571, 'train/grad_norm_pre_clip': 9.2222318649292, 'timing/data': 0.006225932971574366, 'timing/model': 1.1429855589522049, 'learning_rate/qwen_vl_interface': 9.926893642810058e-06, 'learning_rate/action_model': 9.926893642810059e-05, 'epoch': 8.27}) +[2026-06-20 19:57:28,881][starVLA.training.train_starvla][INFO] - Step 4475, Loss: {'train/loss': 0.01783282496035099, 'train/grad_norm_pre_clip': 2.2641522884368896, 'timing/data': 0.006330893957056105, 'timing/model': 1.1433488059556112, 'learning_rate/qwen_vl_interface': 9.926360431601575e-06, 'learning_rate/action_model': 9.926360431601574e-05, 'epoch': 8.28}) +[2026-06-20 19:57:33,520][starVLA.training.train_starvla][INFO] - Step 4476, Loss: {'train/loss': 0.015538460575044155, 'train/grad_norm_pre_clip': 1.013608694076538, 'timing/data': 0.006657933001406491, 'timing/model': 1.1419728469336405, 'learning_rate/qwen_vl_interface': 9.92582529811796e-06, 'learning_rate/action_model': 9.925825298117959e-05, 'epoch': 8.28}) +[2026-06-20 19:57:38,159][starVLA.training.train_starvla][INFO] - Step 4477, Loss: {'train/loss': 0.2897934317588806, 'train/grad_norm_pre_clip': 24.746679306030273, 'timing/data': 0.006431672954931855, 'timing/model': 1.1488042270066217, 'learning_rate/qwen_vl_interface': 9.925288242579184e-06, 'learning_rate/action_model': 9.925288242579184e-05, 'epoch': 8.28}) +[2026-06-20 19:57:42,872][starVLA.training.train_starvla][INFO] - Step 4478, Loss: {'train/loss': 0.13799530267715454, 'train/grad_norm_pre_clip': 9.78588581085205, 'timing/data': 0.008992067072540522, 'timing/model': 1.4216522320639342, 'learning_rate/qwen_vl_interface': 9.924749265206015e-06, 'learning_rate/action_model': 9.924749265206015e-05, 'epoch': 8.28}) +[2026-06-20 19:57:47,411][starVLA.training.train_starvla][INFO] - Step 4479, Loss: {'train/loss': 0.1956256479024887, 'train/grad_norm_pre_clip': 12.063312530517578, 'timing/data': 0.007063311990350485, 'timing/model': 1.1454479100648314, 'learning_rate/qwen_vl_interface': 9.924208366220003e-06, 'learning_rate/action_model': 9.924208366220002e-05, 'epoch': 8.28}) +[2026-06-20 19:57:52,142][starVLA.training.train_starvla][INFO] - Step 4480, Loss: {'train/loss': 0.01929706335067749, 'train/grad_norm_pre_clip': 1.7698712348937988, 'timing/data': 0.006151992944069207, 'timing/model': 1.1477743409341201, 'learning_rate/qwen_vl_interface': 9.923665545843492e-06, 'learning_rate/action_model': 9.923665545843491e-05, 'epoch': 8.29}) +[2026-06-20 19:57:56,848][starVLA.training.train_starvla][INFO] - Step 4481, Loss: {'train/loss': 0.0488191656768322, 'train/grad_norm_pre_clip': 6.307709693908691, 'timing/data': 0.009294035960920155, 'timing/model': 1.1737524119671434, 'learning_rate/qwen_vl_interface': 9.923120804299615e-06, 'learning_rate/action_model': 9.923120804299613e-05, 'epoch': 8.29}) +[2026-06-20 19:58:01,512][starVLA.training.train_starvla][INFO] - Step 4482, Loss: {'train/loss': 0.181366428732872, 'train/grad_norm_pre_clip': 13.1669282913208, 'timing/data': 0.00631420302670449, 'timing/model': 1.1465451599797234, 'learning_rate/qwen_vl_interface': 9.922574141812295e-06, 'learning_rate/action_model': 9.922574141812294e-05, 'epoch': 8.29}) +[2026-06-20 19:58:06,215][starVLA.training.train_starvla][INFO] - Step 4483, Loss: {'train/loss': 0.10451271384954453, 'train/grad_norm_pre_clip': 6.930380344390869, 'timing/data': 0.007788219954818487, 'timing/model': 1.187761110952124, 'learning_rate/qwen_vl_interface': 9.922025558606242e-06, 'learning_rate/action_model': 9.922025558606242e-05, 'epoch': 8.29}) +[2026-06-20 19:58:10,889][starVLA.training.train_starvla][INFO] - Step 4484, Loss: {'train/loss': 0.18810681998729706, 'train/grad_norm_pre_clip': 26.657594680786133, 'timing/data': 0.006692343042232096, 'timing/model': 1.3534058110089973, 'learning_rate/qwen_vl_interface': 9.92147505490696e-06, 'learning_rate/action_model': 9.92147505490696e-05, 'epoch': 8.29}) +[2026-06-20 19:58:15,342][starVLA.training.train_starvla][INFO] - Step 4485, Loss: {'train/loss': 0.17699015140533447, 'train/grad_norm_pre_clip': 11.95331859588623, 'timing/data': 0.006347364047542214, 'timing/model': 1.1481170490151271, 'learning_rate/qwen_vl_interface': 9.920922630940741e-06, 'learning_rate/action_model': 9.920922630940742e-05, 'epoch': 8.29}) +[2026-06-20 19:58:20,012][starVLA.training.train_starvla][INFO] - Step 4486, Loss: {'train/loss': 0.007216441910713911, 'train/grad_norm_pre_clip': 1.4175342321395874, 'timing/data': 0.006574313039891422, 'timing/model': 1.144298356026411, 'learning_rate/qwen_vl_interface': 9.920368286934664e-06, 'learning_rate/action_model': 9.920368286934663e-05, 'epoch': 8.3}) +[2026-06-20 19:58:24,707][starVLA.training.train_starvla][INFO] - Step 4487, Loss: {'train/loss': 0.2595216929912567, 'train/grad_norm_pre_clip': 18.631301879882812, 'timing/data': 0.006461863988079131, 'timing/model': 1.1437846950720996, 'learning_rate/qwen_vl_interface': 9.9198120231166e-06, 'learning_rate/action_model': 9.9198120231166e-05, 'epoch': 8.3}) +[2026-06-20 19:58:29,472][starVLA.training.train_starvla][INFO] - Step 4488, Loss: {'train/loss': 0.008034935221076012, 'train/grad_norm_pre_clip': 0.859483003616333, 'timing/data': 0.006896972074173391, 'timing/model': 1.1955148379784077, 'learning_rate/qwen_vl_interface': 9.919253839715208e-06, 'learning_rate/action_model': 9.919253839715208e-05, 'epoch': 8.3}) +[2026-06-20 19:58:34,232][starVLA.training.train_starvla][INFO] - Step 4489, Loss: {'train/loss': 0.04376722127199173, 'train/grad_norm_pre_clip': 14.80107307434082, 'timing/data': 0.0066996830282732844, 'timing/model': 1.1404655289370567, 'learning_rate/qwen_vl_interface': 9.918693736959935e-06, 'learning_rate/action_model': 9.918693736959934e-05, 'epoch': 8.3}) +[2026-06-20 19:58:38,991][starVLA.training.train_starvla][INFO] - Step 4490, Loss: {'train/loss': 0.13371127843856812, 'train/grad_norm_pre_clip': 11.562657356262207, 'timing/data': 0.006356794037856162, 'timing/model': 1.342959069996141, 'learning_rate/qwen_vl_interface': 9.918131715081019e-06, 'learning_rate/action_model': 9.918131715081018e-05, 'epoch': 8.3}) +[2026-06-20 19:58:43,679][starVLA.training.train_starvla][INFO] - Step 4491, Loss: {'train/loss': 0.1368614286184311, 'train/grad_norm_pre_clip': 6.850366592407227, 'timing/data': 0.00676435197237879, 'timing/model': 1.1952938119648024, 'learning_rate/qwen_vl_interface': 9.917567774309486e-06, 'learning_rate/action_model': 9.917567774309485e-05, 'epoch': 8.31}) +[2026-06-20 19:58:48,463][starVLA.training.train_starvla][INFO] - Step 4492, Loss: {'train/loss': 0.11432967334985733, 'train/grad_norm_pre_clip': 16.538362503051758, 'timing/data': 0.006624083034694195, 'timing/model': 1.1452615089947358, 'learning_rate/qwen_vl_interface': 9.91700191487715e-06, 'learning_rate/action_model': 9.91700191487715e-05, 'epoch': 8.31}) +[2026-06-20 19:58:53,300][starVLA.training.train_starvla][INFO] - Step 4493, Loss: {'train/loss': 0.2820991277694702, 'train/grad_norm_pre_clip': 18.491697311401367, 'timing/data': 0.006419743993319571, 'timing/model': 1.1476792010944337, 'learning_rate/qwen_vl_interface': 9.916434137016614e-06, 'learning_rate/action_model': 9.916434137016613e-05, 'epoch': 8.31}) +[2026-06-20 19:58:58,014][starVLA.training.train_starvla][INFO] - Step 4494, Loss: {'train/loss': 0.23875932395458221, 'train/grad_norm_pre_clip': 10.860528945922852, 'timing/data': 0.006641501910053194, 'timing/model': 1.1435805689543486, 'learning_rate/qwen_vl_interface': 9.915864440961272e-06, 'learning_rate/action_model': 9.915864440961272e-05, 'epoch': 8.31}) +[2026-06-20 19:59:02,719][starVLA.training.train_starvla][INFO] - Step 4495, Loss: {'train/loss': 0.10345591604709625, 'train/grad_norm_pre_clip': 11.952046394348145, 'timing/data': 0.006240474060177803, 'timing/model': 1.380546798929572, 'learning_rate/qwen_vl_interface': 9.915292826945303e-06, 'learning_rate/action_model': 9.915292826945303e-05, 'epoch': 8.31}) +[2026-06-20 19:59:07,232][starVLA.training.train_starvla][INFO] - Step 4496, Loss: {'train/loss': 0.15166756510734558, 'train/grad_norm_pre_clip': 14.772612571716309, 'timing/data': 0.009047795902006328, 'timing/model': 1.1739717950113118, 'learning_rate/qwen_vl_interface': 9.914719295203676e-06, 'learning_rate/action_model': 9.914719295203676e-05, 'epoch': 8.31}) +[2026-06-20 19:59:11,870][starVLA.training.train_starvla][INFO] - Step 4497, Loss: {'train/loss': 0.3472711443901062, 'train/grad_norm_pre_clip': 17.282150268554688, 'timing/data': 0.006259732996113598, 'timing/model': 1.1427807649597526, 'learning_rate/qwen_vl_interface': 9.91414384597215e-06, 'learning_rate/action_model': 9.91414384597215e-05, 'epoch': 8.32}) +[2026-06-20 19:59:16,519][starVLA.training.train_starvla][INFO] - Step 4498, Loss: {'train/loss': 0.07779914885759354, 'train/grad_norm_pre_clip': 7.896472454071045, 'timing/data': 0.006478813011199236, 'timing/model': 1.14343897101935, 'learning_rate/qwen_vl_interface': 9.913566479487267e-06, 'learning_rate/action_model': 9.913566479487267e-05, 'epoch': 8.32}) +[2026-06-20 19:59:21,100][starVLA.training.train_starvla][INFO] - Step 4499, Loss: {'train/loss': 0.07186542451381683, 'train/grad_norm_pre_clip': 7.614990234375, 'timing/data': 0.006269064033403993, 'timing/model': 1.141773832961917, 'learning_rate/qwen_vl_interface': 9.912987195986366e-06, 'learning_rate/action_model': 9.912987195986365e-05, 'epoch': 8.32}) +[2026-06-20 19:59:25,703][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 4500 +[2026-06-20 20:05:53,495][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 4500 +[2026-06-20 20:05:53,496][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 4500 +[2026-06-20 20:12:48,526][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 4500 +[2026-06-20 20:12:48,527][starVLA.training.train_starvla][INFO] - Step 4500, Loss: {'train/loss': 0.2881883680820465, 'train/grad_norm_pre_clip': 16.5257625579834, 'eval/loss': 0.6666690111160278, 'eval/latency_0/loss': 0.713348925113678, 'eval/latency_1/loss': 0.978222668170929, 'eval/latency_2/loss': 1.0025044679641724, 'eval/latency_3/loss': 0.49184027314186096, 'eval/latency_4/loss': 0.11246431618928909, 'eval/flappy/loss': 0.6666690111160278, 'eval/flappy/latency_0/loss': 0.713348925113678, 'eval/flappy/latency_1/loss': 0.978222668170929, 'eval/flappy/latency_2/loss': 1.0025044679641724, 'eval/flappy/latency_3/loss': 0.49184027314186096, 'eval/flappy/latency_4/loss': 0.11246431618928909, 'eval/flap_precision': 0.2689075630252101, 'eval/flap_recall': 0.11636363636363636, 'eval/flap_f1': 0.16243654822335024, 'eval/flap_rate_model': 0.03305555555555555, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.16243654822335024, 'eval/latency_0/flap_precision': 0.2689075630252101, 'eval/latency_0/flap_recall': 0.11636363636363636, 'eval/latency_0/flap_f1': 0.16243654822335024, 'eval/latency_0/flap_rate_model': 0.03305555555555555, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.16243654822335024, 'timing/data': 0.0063179240096360445, 'timing/model': 1.1425393290119246, 'learning_rate/qwen_vl_interface': 9.912405995707562e-06, 'learning_rate/action_model': 9.912405995707561e-05, 'epoch': 8.32}) +[2026-06-20 20:12:48,528][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_4500_state +[2026-06-20 20:12:48,532][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-20 20:12:57,063][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_4500_state/model.safetensors +[2026-06-20 20:13:10,437][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_4500_state/optimizer.bin +[2026-06-20 20:13:10,437][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_4500_state/sampler.bin +[2026-06-20 20:13:10,437][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_4500_state/sampler_1.bin +[2026-06-20 20:13:10,438][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_4500_state/random_states_0.pkl +[2026-06-20 20:13:13,098][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4500 +[2026-06-20 20:13:13,098][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-20 20:13:13,111][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-20 20:13:17,716][starVLA.training.train_starvla][INFO] - Step 4501, Loss: {'train/loss': 0.06288328021764755, 'train/grad_norm_pre_clip': 5.552126884460449, 'timing/data': 0.0063672029646113515, 'timing/model': 1.3959216820076108, 'learning_rate/qwen_vl_interface': 9.911822878889767e-06, 'learning_rate/action_model': 9.911822878889767e-05, 'epoch': 8.32}) +[2026-06-20 20:13:22,284][starVLA.training.train_starvla][INFO] - Step 4502, Loss: {'train/loss': 0.05896353721618652, 'train/grad_norm_pre_clip': 5.566479206085205, 'timing/data': 0.006345152971334755, 'timing/model': 1.3299125430639833, 'learning_rate/qwen_vl_interface': 9.911237845772681e-06, 'learning_rate/action_model': 9.911237845772681e-05, 'epoch': 8.33}) +[2026-06-20 20:13:26,657][starVLA.training.train_starvla][INFO] - Step 4503, Loss: {'train/loss': 0.060588035732507706, 'train/grad_norm_pre_clip': 6.92112922668457, 'timing/data': 0.008318262989632785, 'timing/model': 1.1432910490548238, 'learning_rate/qwen_vl_interface': 9.910650896596787e-06, 'learning_rate/action_model': 9.910650896596786e-05, 'epoch': 8.33}) +[2026-06-20 20:13:31,400][starVLA.training.train_starvla][INFO] - Step 4504, Loss: {'train/loss': 0.206874281167984, 'train/grad_norm_pre_clip': 8.217426300048828, 'timing/data': 0.006474343012087047, 'timing/model': 1.2066683979937807, 'learning_rate/qwen_vl_interface': 9.910062031603356e-06, 'learning_rate/action_model': 9.910062031603356e-05, 'epoch': 8.33}) +[2026-06-20 20:13:36,162][starVLA.training.train_starvla][INFO] - Step 4505, Loss: {'train/loss': 0.020532485097646713, 'train/grad_norm_pre_clip': 4.0116801261901855, 'timing/data': 0.0062669349135831, 'timing/model': 1.1825596949784085, 'learning_rate/qwen_vl_interface': 9.909471251034451e-06, 'learning_rate/action_model': 9.90947125103445e-05, 'epoch': 8.33}) +[2026-06-20 20:13:40,872][starVLA.training.train_starvla][INFO] - Step 4506, Loss: {'train/loss': 0.3338252604007721, 'train/grad_norm_pre_clip': 7.47515344619751, 'timing/data': 0.006386094028130174, 'timing/model': 1.1744563999818638, 'learning_rate/qwen_vl_interface': 9.908878555132918e-06, 'learning_rate/action_model': 9.908878555132918e-05, 'epoch': 8.33}) +[2026-06-20 20:13:45,559][starVLA.training.train_starvla][INFO] - Step 4507, Loss: {'train/loss': 0.335976779460907, 'train/grad_norm_pre_clip': 18.513357162475586, 'timing/data': 0.00639311398845166, 'timing/model': 1.3651918340474367, 'learning_rate/qwen_vl_interface': 9.908283944142391e-06, 'learning_rate/action_model': 9.908283944142391e-05, 'epoch': 8.34}) +[2026-06-20 20:13:50,133][starVLA.training.train_starvla][INFO] - Step 4508, Loss: {'train/loss': 0.07812883704900742, 'train/grad_norm_pre_clip': 9.661880493164062, 'timing/data': 0.006255273940041661, 'timing/model': 1.2972305670846254, 'learning_rate/qwen_vl_interface': 9.907687418307297e-06, 'learning_rate/action_model': 9.907687418307296e-05, 'epoch': 8.34}) +[2026-06-20 20:13:54,636][starVLA.training.train_starvla][INFO] - Step 4509, Loss: {'train/loss': 0.06820755451917648, 'train/grad_norm_pre_clip': 7.419887542724609, 'timing/data': 0.007389168022200465, 'timing/model': 1.1797119870316237, 'learning_rate/qwen_vl_interface': 9.907088977872837e-06, 'learning_rate/action_model': 9.907088977872837e-05, 'epoch': 8.34}) +[2026-06-20 20:13:59,311][starVLA.training.train_starvla][INFO] - Step 4510, Loss: {'train/loss': 0.3254198133945465, 'train/grad_norm_pre_clip': 15.656509399414062, 'timing/data': 0.006518773036077619, 'timing/model': 1.1470962379826233, 'learning_rate/qwen_vl_interface': 9.906488623085015e-06, 'learning_rate/action_model': 9.906488623085015e-05, 'epoch': 8.34}) +[2026-06-20 20:14:04,041][starVLA.training.train_starvla][INFO] - Step 4511, Loss: {'train/loss': 0.12592542171478271, 'train/grad_norm_pre_clip': 6.581493854522705, 'timing/data': 0.006780641968362033, 'timing/model': 1.1415372668998316, 'learning_rate/qwen_vl_interface': 9.905886354190611e-06, 'learning_rate/action_model': 9.90588635419061e-05, 'epoch': 8.34}) +[2026-06-20 20:14:08,761][starVLA.training.train_starvla][INFO] - Step 4512, Loss: {'train/loss': 0.07393021136522293, 'train/grad_norm_pre_clip': 5.417125701904297, 'timing/data': 0.009079218958504498, 'timing/model': 1.1383828830439597, 'learning_rate/qwen_vl_interface': 9.905282171437192e-06, 'learning_rate/action_model': 9.905282171437192e-05, 'epoch': 8.34}) +[2026-06-20 20:14:13,439][starVLA.training.train_starvla][INFO] - Step 4513, Loss: {'train/loss': 0.07193247973918915, 'train/grad_norm_pre_clip': 7.722016334533691, 'timing/data': 0.006552623934112489, 'timing/model': 1.386632083915174, 'learning_rate/qwen_vl_interface': 9.904676075073121e-06, 'learning_rate/action_model': 9.904676075073121e-05, 'epoch': 8.35}) +[2026-06-20 20:14:17,881][starVLA.training.train_starvla][INFO] - Step 4514, Loss: {'train/loss': 0.10032204538583755, 'train/grad_norm_pre_clip': 5.091470718383789, 'timing/data': 0.006378805031999946, 'timing/model': 1.1440524749923497, 'learning_rate/qwen_vl_interface': 9.904068065347536e-06, 'learning_rate/action_model': 9.904068065347536e-05, 'epoch': 8.35}) +[2026-06-20 20:14:22,552][starVLA.training.train_starvla][INFO] - Step 4515, Loss: {'train/loss': 0.022972729057073593, 'train/grad_norm_pre_clip': 2.2307043075561523, 'timing/data': 0.006498003960587084, 'timing/model': 1.1456925929524004, 'learning_rate/qwen_vl_interface': 9.90345814251037e-06, 'learning_rate/action_model': 9.90345814251037e-05, 'epoch': 8.35}) +[2026-06-20 20:14:27,231][starVLA.training.train_starvla][INFO] - Step 4516, Loss: {'train/loss': 0.06334897875785828, 'train/grad_norm_pre_clip': 5.6734089851379395, 'timing/data': 0.006824663025327027, 'timing/model': 1.164718855987303, 'learning_rate/qwen_vl_interface': 9.902846306812339e-06, 'learning_rate/action_model': 9.902846306812338e-05, 'epoch': 8.35}) +[2026-06-20 20:14:31,882][starVLA.training.train_starvla][INFO] - Step 4517, Loss: {'train/loss': 0.0662308782339096, 'train/grad_norm_pre_clip': 6.721391201019287, 'timing/data': 0.006405763910152018, 'timing/model': 1.1623475360684097, 'learning_rate/qwen_vl_interface': 9.902232558504942e-06, 'learning_rate/action_model': 9.902232558504941e-05, 'epoch': 8.35}) +[2026-06-20 20:14:36,556][starVLA.training.train_starvla][INFO] - Step 4518, Loss: {'train/loss': 0.1168956384062767, 'train/grad_norm_pre_clip': 6.561938285827637, 'timing/data': 0.006371124996803701, 'timing/model': 1.1668552689952776, 'learning_rate/qwen_vl_interface': 9.901616897840471e-06, 'learning_rate/action_model': 9.90161689784047e-05, 'epoch': 8.36}) +[2026-06-20 20:14:41,217][starVLA.training.train_starvla][INFO] - Step 4519, Loss: {'train/loss': 0.17312540113925934, 'train/grad_norm_pre_clip': 11.238387107849121, 'timing/data': 0.006505594938062131, 'timing/model': 1.3935400489717722, 'learning_rate/qwen_vl_interface': 9.900999325072e-06, 'learning_rate/action_model': 9.900999325072001e-05, 'epoch': 8.36}) +[2026-06-20 20:14:45,685][starVLA.training.train_starvla][INFO] - Step 4520, Loss: {'train/loss': 0.18245477974414825, 'train/grad_norm_pre_clip': 12.373251914978027, 'timing/data': 0.007748428033664823, 'timing/model': 1.158990975935012, 'learning_rate/qwen_vl_interface': 9.900379840453392e-06, 'learning_rate/action_model': 9.900379840453392e-05, 'epoch': 8.36}) +[2026-06-20 20:14:50,344][starVLA.training.train_starvla][INFO] - Step 4521, Loss: {'train/loss': 0.06965027004480362, 'train/grad_norm_pre_clip': 5.6859130859375, 'timing/data': 0.0065312039805576205, 'timing/model': 1.1434362860163674, 'learning_rate/qwen_vl_interface': 9.89975844423929e-06, 'learning_rate/action_model': 9.899758444239289e-05, 'epoch': 8.36}) +[2026-06-20 20:14:55,001][starVLA.training.train_starvla][INFO] - Step 4522, Loss: {'train/loss': 0.22291482985019684, 'train/grad_norm_pre_clip': 13.439338684082031, 'timing/data': 0.006636874051764607, 'timing/model': 1.1428804650204256, 'learning_rate/qwen_vl_interface': 9.899135136685131e-06, 'learning_rate/action_model': 9.89913513668513e-05, 'epoch': 8.36}) +[2026-06-20 20:14:59,665][starVLA.training.train_starvla][INFO] - Step 4523, Loss: {'train/loss': 0.15408936142921448, 'train/grad_norm_pre_clip': 5.2815117835998535, 'timing/data': 0.006314906058833003, 'timing/model': 1.1480574130546302, 'learning_rate/qwen_vl_interface': 9.898509918047127e-06, 'learning_rate/action_model': 9.898509918047126e-05, 'epoch': 8.36}) +[2026-06-20 20:15:04,322][starVLA.training.train_starvla][INFO] - Step 4524, Loss: {'train/loss': 0.03986092284321785, 'train/grad_norm_pre_clip': 3.676910877227783, 'timing/data': 0.006237806985154748, 'timing/model': 1.1466754059074447, 'learning_rate/qwen_vl_interface': 9.897882788582286e-06, 'learning_rate/action_model': 9.897882788582287e-05, 'epoch': 8.37}) +[2026-06-20 20:15:08,974][starVLA.training.train_starvla][INFO] - Step 4525, Loss: {'train/loss': 0.052138131111860275, 'train/grad_norm_pre_clip': 5.671413421630859, 'timing/data': 0.006398114957846701, 'timing/model': 1.3884601140161976, 'learning_rate/qwen_vl_interface': 9.897253748548398e-06, 'learning_rate/action_model': 9.897253748548399e-05, 'epoch': 8.37}) +[2026-06-20 20:15:13,615][starVLA.training.train_starvla][INFO] - Step 4526, Loss: {'train/loss': 0.05888333544135094, 'train/grad_norm_pre_clip': 3.50424861907959, 'timing/data': 0.006214647088199854, 'timing/model': 1.324076427030377, 'learning_rate/qwen_vl_interface': 9.896622798204035e-06, 'learning_rate/action_model': 9.896622798204035e-05, 'epoch': 8.37}) +[2026-06-20 20:15:18,018][starVLA.training.train_starvla][INFO] - Step 4527, Loss: {'train/loss': 0.0380764976143837, 'train/grad_norm_pre_clip': 7.175990104675293, 'timing/data': 0.006249036057852209, 'timing/model': 1.144780763075687, 'learning_rate/qwen_vl_interface': 9.89598993780856e-06, 'learning_rate/action_model': 9.89598993780856e-05, 'epoch': 8.37}) +[2026-06-20 20:15:22,631][starVLA.training.train_starvla][INFO] - Step 4528, Loss: {'train/loss': 0.07213860005140305, 'train/grad_norm_pre_clip': 5.585449695587158, 'timing/data': 0.006615525111556053, 'timing/model': 1.1431556369643658, 'learning_rate/qwen_vl_interface': 9.895355167622113e-06, 'learning_rate/action_model': 9.895355167622113e-05, 'epoch': 8.37}) +[2026-06-20 20:15:27,452][starVLA.training.train_starvla][INFO] - Step 4529, Loss: {'train/loss': 0.16217361390590668, 'train/grad_norm_pre_clip': 11.247841835021973, 'timing/data': 0.0072640819707885385, 'timing/model': 1.1971451350254938, 'learning_rate/qwen_vl_interface': 9.894718487905629e-06, 'learning_rate/action_model': 9.894718487905629e-05, 'epoch': 8.38}) +[2026-06-20 20:15:32,219][starVLA.training.train_starvla][INFO] - Step 4530, Loss: {'train/loss': 0.04074512794613838, 'train/grad_norm_pre_clip': 3.7003557682037354, 'timing/data': 0.006231527077034116, 'timing/model': 1.1412968679796904, 'learning_rate/qwen_vl_interface': 9.89407989892082e-06, 'learning_rate/action_model': 9.894079898920821e-05, 'epoch': 8.38}) +[2026-06-20 20:15:37,036][starVLA.training.train_starvla][INFO] - Step 4531, Loss: {'train/loss': 0.03572922199964523, 'train/grad_norm_pre_clip': 3.3678860664367676, 'timing/data': 0.007519201026298106, 'timing/model': 1.4271989489207044, 'learning_rate/qwen_vl_interface': 9.893439400930187e-06, 'learning_rate/action_model': 9.893439400930187e-05, 'epoch': 8.38}) +[2026-06-20 20:15:41,886][starVLA.training.train_starvla][INFO] - Step 4532, Loss: {'train/loss': 0.23855379223823547, 'train/grad_norm_pre_clip': 15.756875991821289, 'timing/data': 0.006159038050100207, 'timing/model': 1.3827601439552382, 'learning_rate/qwen_vl_interface': 9.892796994197014e-06, 'learning_rate/action_model': 9.892796994197014e-05, 'epoch': 8.38}) +[2026-06-20 20:15:46,396][starVLA.training.train_starvla][INFO] - Step 4533, Loss: {'train/loss': 0.013625485822558403, 'train/grad_norm_pre_clip': 2.75626802444458, 'timing/data': 0.006302066962234676, 'timing/model': 1.1441007470712066, 'learning_rate/qwen_vl_interface': 9.892152678985372e-06, 'learning_rate/action_model': 9.892152678985372e-05, 'epoch': 8.38}) +[2026-06-20 20:15:51,088][starVLA.training.train_starvla][INFO] - Step 4534, Loss: {'train/loss': 0.23522664606571198, 'train/grad_norm_pre_clip': 8.878366470336914, 'timing/data': 0.0059818190056830645, 'timing/model': 1.2186870999867097, 'learning_rate/qwen_vl_interface': 9.891506455560111e-06, 'learning_rate/action_model': 9.891506455560111e-05, 'epoch': 8.39}) +[2026-06-20 20:15:55,979][starVLA.training.train_starvla][INFO] - Step 4535, Loss: {'train/loss': 0.06442026793956757, 'train/grad_norm_pre_clip': 10.288127899169922, 'timing/data': 0.006123897968791425, 'timing/model': 1.221773380995728, 'learning_rate/qwen_vl_interface': 9.890858324186872e-06, 'learning_rate/action_model': 9.890858324186872e-05, 'epoch': 8.39}) +[2026-06-20 20:16:00,684][starVLA.training.train_starvla][INFO] - Step 4536, Loss: {'train/loss': 0.18630927801132202, 'train/grad_norm_pre_clip': 10.413300514221191, 'timing/data': 0.006010549026541412, 'timing/model': 1.1424278900958598, 'learning_rate/qwen_vl_interface': 9.890208285132077e-06, 'learning_rate/action_model': 9.890208285132077e-05, 'epoch': 8.39}) +[2026-06-20 20:16:05,457][starVLA.training.train_starvla][INFO] - Step 4537, Loss: {'train/loss': 0.3778156340122223, 'train/grad_norm_pre_clip': 10.9894380569458, 'timing/data': 0.006270137964747846, 'timing/model': 1.3860111749963835, 'learning_rate/qwen_vl_interface': 9.889556338662933e-06, 'learning_rate/action_model': 9.889556338662932e-05, 'epoch': 8.39}) +[2026-06-20 20:16:09,942][starVLA.training.train_starvla][INFO] - Step 4538, Loss: {'train/loss': 0.01895718090236187, 'train/grad_norm_pre_clip': 3.5591604709625244, 'timing/data': 0.006524795899167657, 'timing/model': 1.1614294990431517, 'learning_rate/qwen_vl_interface': 9.888902485047427e-06, 'learning_rate/action_model': 9.888902485047427e-05, 'epoch': 8.39}) +[2026-06-20 20:16:14,657][starVLA.training.train_starvla][INFO] - Step 4539, Loss: {'train/loss': 0.49809157848358154, 'train/grad_norm_pre_clip': 16.719762802124023, 'timing/data': 0.0063349169213324785, 'timing/model': 1.1602767700096592, 'learning_rate/qwen_vl_interface': 9.888246724554337e-06, 'learning_rate/action_model': 9.888246724554337e-05, 'epoch': 8.39}) +[2026-06-20 20:16:19,363][starVLA.training.train_starvla][INFO] - Step 4540, Loss: {'train/loss': 0.08665294200181961, 'train/grad_norm_pre_clip': 8.890105247497559, 'timing/data': 0.008274017018266022, 'timing/model': 1.1507283330429345, 'learning_rate/qwen_vl_interface': 9.887589057453222e-06, 'learning_rate/action_model': 9.88758905745322e-05, 'epoch': 8.4}) +[2026-06-20 20:16:24,078][starVLA.training.train_starvla][INFO] - Step 4541, Loss: {'train/loss': 0.21837550401687622, 'train/grad_norm_pre_clip': 17.356660842895508, 'timing/data': 0.006400256999768317, 'timing/model': 1.1534437630325556, 'learning_rate/qwen_vl_interface': 9.88692948401442e-06, 'learning_rate/action_model': 9.88692948401442e-05, 'epoch': 8.4}) +[2026-06-20 20:16:28,695][starVLA.training.train_starvla][INFO] - Step 4542, Loss: {'train/loss': 0.05865991488099098, 'train/grad_norm_pre_clip': 7.946161270141602, 'timing/data': 0.006386087043210864, 'timing/model': 1.1530925780534744, 'learning_rate/qwen_vl_interface': 9.88626800450906e-06, 'learning_rate/action_model': 9.88626800450906e-05, 'epoch': 8.4}) +[2026-06-20 20:16:33,325][starVLA.training.train_starvla][INFO] - Step 4543, Loss: {'train/loss': 0.007070666179060936, 'train/grad_norm_pre_clip': 0.4972067177295685, 'timing/data': 0.006388867041096091, 'timing/model': 1.340966706047766, 'learning_rate/qwen_vl_interface': 9.885604619209052e-06, 'learning_rate/action_model': 9.885604619209051e-05, 'epoch': 8.4}) +[2026-06-20 20:16:37,732][starVLA.training.train_starvla][INFO] - Step 4544, Loss: {'train/loss': 0.2961697578430176, 'train/grad_norm_pre_clip': 13.854056358337402, 'timing/data': 0.006406208034604788, 'timing/model': 1.1445330779533833, 'learning_rate/qwen_vl_interface': 9.884939328387085e-06, 'learning_rate/action_model': 9.884939328387084e-05, 'epoch': 8.4}) +[2026-06-20 20:16:42,387][starVLA.training.train_starvla][INFO] - Step 4545, Loss: {'train/loss': 0.1600535809993744, 'train/grad_norm_pre_clip': 7.7765421867370605, 'timing/data': 0.006309688091278076, 'timing/model': 1.1445092010544613, 'learning_rate/qwen_vl_interface': 9.884272132316638e-06, 'learning_rate/action_model': 9.884272132316637e-05, 'epoch': 8.41}) +[2026-06-20 20:16:47,060][starVLA.training.train_starvla][INFO] - Step 4546, Loss: {'train/loss': 0.029850728809833527, 'train/grad_norm_pre_clip': 4.27396821975708, 'timing/data': 0.0063309569377452135, 'timing/model': 1.144060107995756, 'learning_rate/qwen_vl_interface': 9.883603031271968e-06, 'learning_rate/action_model': 9.883603031271969e-05, 'epoch': 8.41}) +[2026-06-20 20:16:51,716][starVLA.training.train_starvla][INFO] - Step 4547, Loss: {'train/loss': 0.03242317959666252, 'train/grad_norm_pre_clip': 3.7751753330230713, 'timing/data': 0.006364276981912553, 'timing/model': 1.1435523430118337, 'learning_rate/qwen_vl_interface': 9.882932025528122e-06, 'learning_rate/action_model': 9.882932025528121e-05, 'epoch': 8.41}) +[2026-06-20 20:16:56,381][starVLA.training.train_starvla][INFO] - Step 4548, Loss: {'train/loss': 0.04413092881441116, 'train/grad_norm_pre_clip': 5.4057512283325195, 'timing/data': 0.00812238804064691, 'timing/model': 1.165873043006286, 'learning_rate/qwen_vl_interface': 9.882259115360918e-06, 'learning_rate/action_model': 9.882259115360917e-05, 'epoch': 8.41}) +[2026-06-20 20:17:01,023][starVLA.training.train_starvla][INFO] - Step 4549, Loss: {'train/loss': 0.06728952378034592, 'train/grad_norm_pre_clip': 3.743588447570801, 'timing/data': 0.006453036912716925, 'timing/model': 1.3521513589657843, 'learning_rate/qwen_vl_interface': 9.881584301046968e-06, 'learning_rate/action_model': 9.881584301046968e-05, 'epoch': 8.41}) +[2026-06-20 20:17:05,464][starVLA.training.train_starvla][INFO] - Step 4550, Loss: {'train/loss': 0.17236638069152832, 'train/grad_norm_pre_clip': 15.140710830688477, 'timing/data': 0.006400798098184168, 'timing/model': 1.1442858759546652, 'learning_rate/qwen_vl_interface': 9.880907582863662e-06, 'learning_rate/action_model': 9.880907582863662e-05, 'epoch': 8.41}) +[2026-06-20 20:17:10,112][starVLA.training.train_starvla][INFO] - Step 4551, Loss: {'train/loss': 0.049612801522016525, 'train/grad_norm_pre_clip': 11.425899505615234, 'timing/data': 0.006506316945888102, 'timing/model': 1.1448494549840689, 'learning_rate/qwen_vl_interface': 9.880228961089175e-06, 'learning_rate/action_model': 9.880228961089175e-05, 'epoch': 8.42}) +[2026-06-20 20:17:14,764][starVLA.training.train_starvla][INFO] - Step 4552, Loss: {'train/loss': 0.3066417872905731, 'train/grad_norm_pre_clip': 22.30399513244629, 'timing/data': 0.006316428072750568, 'timing/model': 1.1434000260196626, 'learning_rate/qwen_vl_interface': 9.87954843600246e-06, 'learning_rate/action_model': 9.879548436002459e-05, 'epoch': 8.42}) +[2026-06-20 20:17:19,464][starVLA.training.train_starvla][INFO] - Step 4553, Loss: {'train/loss': 0.13919256627559662, 'train/grad_norm_pre_clip': 15.009058952331543, 'timing/data': 0.006399097037501633, 'timing/model': 1.1602868429617956, 'learning_rate/qwen_vl_interface': 9.878866007883256e-06, 'learning_rate/action_model': 9.878866007883257e-05, 'epoch': 8.42}) +[2026-06-20 20:17:24,145][starVLA.training.train_starvla][INFO] - Step 4554, Loss: {'train/loss': 0.2016931176185608, 'train/grad_norm_pre_clip': 9.753997802734375, 'timing/data': 0.006259928923100233, 'timing/model': 1.3807075179647654, 'learning_rate/qwen_vl_interface': 9.878181677012084e-06, 'learning_rate/action_model': 9.878181677012085e-05, 'epoch': 8.42}) +[2026-06-20 20:17:28,554][starVLA.training.train_starvla][INFO] - Step 4555, Loss: {'train/loss': 0.07295457273721695, 'train/grad_norm_pre_clip': 13.558444023132324, 'timing/data': 0.006408286979421973, 'timing/model': 1.1452080130111426, 'learning_rate/qwen_vl_interface': 9.877495443670247e-06, 'learning_rate/action_model': 9.877495443670246e-05, 'epoch': 8.42}) +[2026-06-20 20:17:33,234][starVLA.training.train_starvla][INFO] - Step 4556, Loss: {'train/loss': 0.26745638251304626, 'train/grad_norm_pre_clip': 13.924095153808594, 'timing/data': 0.006835476029664278, 'timing/model': 1.1446585659869015, 'learning_rate/qwen_vl_interface': 9.876807308139828e-06, 'learning_rate/action_model': 9.876807308139829e-05, 'epoch': 8.43}) +[2026-06-20 20:17:37,836][starVLA.training.train_starvla][INFO] - Step 4557, Loss: {'train/loss': 0.13533273339271545, 'train/grad_norm_pre_clip': 12.27547836303711, 'timing/data': 0.0064114079577848315, 'timing/model': 1.1446436190744862, 'learning_rate/qwen_vl_interface': 9.876117270703695e-06, 'learning_rate/action_model': 9.876117270703694e-05, 'epoch': 8.43}) +[2026-06-20 20:17:42,443][starVLA.training.train_starvla][INFO] - Step 4558, Loss: {'train/loss': 0.09860917925834656, 'train/grad_norm_pre_clip': 12.771561622619629, 'timing/data': 0.009059735108166933, 'timing/model': 1.1446232600137591, 'learning_rate/qwen_vl_interface': 9.875425331645495e-06, 'learning_rate/action_model': 9.875425331645495e-05, 'epoch': 8.43}) +[2026-06-20 20:17:47,122][starVLA.training.train_starvla][INFO] - Step 4559, Loss: {'train/loss': 0.17313721776008606, 'train/grad_norm_pre_clip': 22.97819709777832, 'timing/data': 0.009216904058121145, 'timing/model': 1.1595178989227861, 'learning_rate/qwen_vl_interface': 9.874731491249658e-06, 'learning_rate/action_model': 9.874731491249657e-05, 'epoch': 8.43}) +[2026-06-20 20:17:51,776][starVLA.training.train_starvla][INFO] - Step 4560, Loss: {'train/loss': 0.12256795167922974, 'train/grad_norm_pre_clip': 11.238384246826172, 'timing/data': 0.006476796930655837, 'timing/model': 1.3882263149134815, 'learning_rate/qwen_vl_interface': 9.874035749801398e-06, 'learning_rate/action_model': 9.874035749801399e-05, 'epoch': 8.43}) +[2026-06-20 20:17:56,211][starVLA.training.train_starvla][INFO] - Step 4561, Loss: {'train/loss': 0.19342143833637238, 'train/grad_norm_pre_clip': 8.838448524475098, 'timing/data': 0.006435798015445471, 'timing/model': 1.1443500459427014, 'learning_rate/qwen_vl_interface': 9.873338107586704e-06, 'learning_rate/action_model': 9.873338107586705e-05, 'epoch': 8.43}) +[2026-06-20 20:18:00,884][starVLA.training.train_starvla][INFO] - Step 4562, Loss: {'train/loss': 0.05067289620637894, 'train/grad_norm_pre_clip': 4.398348331451416, 'timing/data': 0.006363617954775691, 'timing/model': 1.1469743339112028, 'learning_rate/qwen_vl_interface': 9.872638564892358e-06, 'learning_rate/action_model': 9.872638564892357e-05, 'epoch': 8.44}) +[2026-06-20 20:18:05,524][starVLA.training.train_starvla][INFO] - Step 4563, Loss: {'train/loss': 0.03425103798508644, 'train/grad_norm_pre_clip': 3.891578435897827, 'timing/data': 0.0063759880140423775, 'timing/model': 1.1453484839294106, 'learning_rate/qwen_vl_interface': 9.871937122005905e-06, 'learning_rate/action_model': 9.871937122005905e-05, 'epoch': 8.44}) +[2026-06-20 20:18:10,205][starVLA.training.train_starvla][INFO] - Step 4564, Loss: {'train/loss': 0.24258777499198914, 'train/grad_norm_pre_clip': 16.834672927856445, 'timing/data': 0.008578606997616589, 'timing/model': 1.1594631749903783, 'learning_rate/qwen_vl_interface': 9.87123377921569e-06, 'learning_rate/action_model': 9.871233779215689e-05, 'epoch': 8.44}) +[2026-06-20 20:18:14,867][starVLA.training.train_starvla][INFO] - Step 4565, Loss: {'train/loss': 0.09801968187093735, 'train/grad_norm_pre_clip': 12.031579971313477, 'timing/data': 0.006304018897935748, 'timing/model': 1.1627620400395244, 'learning_rate/qwen_vl_interface': 9.870528536810825e-06, 'learning_rate/action_model': 9.870528536810825e-05, 'epoch': 8.44}) +[2026-06-20 20:18:19,548][starVLA.training.train_starvla][INFO] - Step 4566, Loss: {'train/loss': 0.2634917199611664, 'train/grad_norm_pre_clip': 9.671074867248535, 'timing/data': 0.006547037977725267, 'timing/model': 1.3798592169769108, 'learning_rate/qwen_vl_interface': 9.869821395081214e-06, 'learning_rate/action_model': 9.869821395081214e-05, 'epoch': 8.44}) +[2026-06-20 20:18:23,983][starVLA.training.train_starvla][INFO] - Step 4567, Loss: {'train/loss': 0.20610716938972473, 'train/grad_norm_pre_clip': 8.14005184173584, 'timing/data': 0.006563637056387961, 'timing/model': 1.1435686049517244, 'learning_rate/qwen_vl_interface': 9.869112354317532e-06, 'learning_rate/action_model': 9.869112354317532e-05, 'epoch': 8.45}) +[2026-06-20 20:18:28,662][starVLA.training.train_starvla][INFO] - Step 4568, Loss: {'train/loss': 0.1556144505739212, 'train/grad_norm_pre_clip': 11.530308723449707, 'timing/data': 0.006502278964035213, 'timing/model': 1.1455424360465258, 'learning_rate/qwen_vl_interface': 9.86840141481124e-06, 'learning_rate/action_model': 9.868401414811239e-05, 'epoch': 8.45}) +[2026-06-20 20:18:33,329][starVLA.training.train_starvla][INFO] - Step 4569, Loss: {'train/loss': 0.43862035870552063, 'train/grad_norm_pre_clip': 11.063151359558105, 'timing/data': 0.006239688955247402, 'timing/model': 1.1439866150030866, 'learning_rate/qwen_vl_interface': 9.867688576854578e-06, 'learning_rate/action_model': 9.867688576854578e-05, 'epoch': 8.45}) +[2026-06-20 20:18:37,989][starVLA.training.train_starvla][INFO] - Step 4570, Loss: {'train/loss': 0.060998160392045975, 'train/grad_norm_pre_clip': 6.089873313903809, 'timing/data': 0.0064696589251980186, 'timing/model': 1.143249579006806, 'learning_rate/qwen_vl_interface': 9.86697384074057e-06, 'learning_rate/action_model': 9.866973840740568e-05, 'epoch': 8.45}) +[2026-06-20 20:18:42,635][starVLA.training.train_starvla][INFO] - Step 4571, Loss: {'train/loss': 0.13377399742603302, 'train/grad_norm_pre_clip': 8.646881103515625, 'timing/data': 0.006433759001083672, 'timing/model': 1.3763199470704421, 'learning_rate/qwen_vl_interface': 9.86625720676301e-06, 'learning_rate/action_model': 9.86625720676301e-05, 'epoch': 8.45}) +[2026-06-20 20:18:47,192][starVLA.training.train_starvla][INFO] - Step 4572, Loss: {'train/loss': 0.039277080446481705, 'train/grad_norm_pre_clip': 3.099700927734375, 'timing/data': 0.0065111080184578896, 'timing/model': 1.2908530340064317, 'learning_rate/qwen_vl_interface': 9.865538675216488e-06, 'learning_rate/action_model': 9.865538675216486e-05, 'epoch': 8.46}) +[2026-06-20 20:18:51,605][starVLA.training.train_starvla][INFO] - Step 4573, Loss: {'train/loss': 0.10583114624023438, 'train/grad_norm_pre_clip': 6.8215012550354, 'timing/data': 0.00624034006614238, 'timing/model': 1.1442483440041542, 'learning_rate/qwen_vl_interface': 9.864818246396358e-06, 'learning_rate/action_model': 9.864818246396358e-05, 'epoch': 8.46}) +[2026-06-20 20:18:56,384][starVLA.training.train_starvla][INFO] - Step 4574, Loss: {'train/loss': 0.19967424869537354, 'train/grad_norm_pre_clip': 8.411149024963379, 'timing/data': 0.006584088085219264, 'timing/model': 1.1469637710833922, 'learning_rate/qwen_vl_interface': 9.864095920598764e-06, 'learning_rate/action_model': 9.864095920598763e-05, 'epoch': 8.46}) +[2026-06-20 20:19:01,171][starVLA.training.train_starvla][INFO] - Step 4575, Loss: {'train/loss': 0.12058506906032562, 'train/grad_norm_pre_clip': 5.5175042152404785, 'timing/data': 0.006343869026750326, 'timing/model': 1.1467178330058232, 'learning_rate/qwen_vl_interface': 9.86337169812063e-06, 'learning_rate/action_model': 9.863371698120629e-05, 'epoch': 8.46}) +[2026-06-20 20:19:05,976][starVLA.training.train_starvla][INFO] - Step 4576, Loss: {'train/loss': 0.20961779356002808, 'train/grad_norm_pre_clip': 7.086739540100098, 'timing/data': 0.0076484230812639, 'timing/model': 1.2192154140211642, 'learning_rate/qwen_vl_interface': 9.862645579259648e-06, 'learning_rate/action_model': 9.862645579259648e-05, 'epoch': 8.46}) +[2026-06-20 20:19:10,764][starVLA.training.train_starvla][INFO] - Step 4577, Loss: {'train/loss': 0.09104747325181961, 'train/grad_norm_pre_clip': 5.935173988342285, 'timing/data': 0.00877824705094099, 'timing/model': 1.4376379249151796, 'learning_rate/qwen_vl_interface': 9.861917564314308e-06, 'learning_rate/action_model': 9.861917564314308e-05, 'epoch': 8.46}) +[2026-06-20 20:19:15,564][starVLA.training.train_starvla][INFO] - Step 4578, Loss: {'train/loss': 0.14194545149803162, 'train/grad_norm_pre_clip': 8.809276580810547, 'timing/data': 0.006447240011766553, 'timing/model': 1.319020746042952, 'learning_rate/qwen_vl_interface': 9.861187653583862e-06, 'learning_rate/action_model': 9.861187653583862e-05, 'epoch': 8.47}) +[2026-06-20 20:19:19,987][starVLA.training.train_starvla][INFO] - Step 4579, Loss: {'train/loss': 0.08059465885162354, 'train/grad_norm_pre_clip': 5.899817943572998, 'timing/data': 0.006326399045065045, 'timing/model': 1.151163280941546, 'learning_rate/qwen_vl_interface': 9.860455847368352e-06, 'learning_rate/action_model': 9.860455847368351e-05, 'epoch': 8.47}) +[2026-06-20 20:19:24,795][starVLA.training.train_starvla][INFO] - Step 4580, Loss: {'train/loss': 0.1353731006383896, 'train/grad_norm_pre_clip': 7.665670394897461, 'timing/data': 0.006331759970635176, 'timing/model': 1.2216450439300388, 'learning_rate/qwen_vl_interface': 9.859722145968595e-06, 'learning_rate/action_model': 9.859722145968595e-05, 'epoch': 8.47}) +[2026-06-20 20:19:29,577][starVLA.training.train_starvla][INFO] - Step 4581, Loss: {'train/loss': 0.3441682457923889, 'train/grad_norm_pre_clip': 10.916888236999512, 'timing/data': 0.0068683159770444036, 'timing/model': 1.2171899259556085, 'learning_rate/qwen_vl_interface': 9.858986549686188e-06, 'learning_rate/action_model': 9.858986549686188e-05, 'epoch': 8.47}) +[2026-06-20 20:19:34,448][starVLA.training.train_starvla][INFO] - Step 4582, Loss: {'train/loss': 0.13664017617702484, 'train/grad_norm_pre_clip': 5.191305637359619, 'timing/data': 0.006986445980146527, 'timing/model': 1.1469958110246807, 'learning_rate/qwen_vl_interface': 9.85824905882351e-06, 'learning_rate/action_model': 9.858249058823509e-05, 'epoch': 8.47}) +[2026-06-20 20:19:39,241][starVLA.training.train_starvla][INFO] - Step 4583, Loss: {'train/loss': 0.05562007799744606, 'train/grad_norm_pre_clip': 4.516536235809326, 'timing/data': 0.006796088069677353, 'timing/model': 1.39172196702566, 'learning_rate/qwen_vl_interface': 9.857509673683709e-06, 'learning_rate/action_model': 9.857509673683708e-05, 'epoch': 8.48}) +[2026-06-20 20:19:43,817][starVLA.training.train_starvla][INFO] - Step 4584, Loss: {'train/loss': 0.1727687418460846, 'train/grad_norm_pre_clip': 5.816850185394287, 'timing/data': 0.007910473039373755, 'timing/model': 1.1995936590246856, 'learning_rate/qwen_vl_interface': 9.856768394570725e-06, 'learning_rate/action_model': 9.856768394570723e-05, 'epoch': 8.48}) +[2026-06-20 20:19:48,693][starVLA.training.train_starvla][INFO] - Step 4585, Loss: {'train/loss': 0.19285932183265686, 'train/grad_norm_pre_clip': 5.689450263977051, 'timing/data': 0.006640748935751617, 'timing/model': 1.1459209639579058, 'learning_rate/qwen_vl_interface': 9.856025221789264e-06, 'learning_rate/action_model': 9.856025221789264e-05, 'epoch': 8.48}) +[2026-06-20 20:19:53,533][starVLA.training.train_starvla][INFO] - Step 4586, Loss: {'train/loss': 0.13372337818145752, 'train/grad_norm_pre_clip': 9.102676391601562, 'timing/data': 0.0077255439246073365, 'timing/model': 1.198938032030128, 'learning_rate/qwen_vl_interface': 9.85528015564482e-06, 'learning_rate/action_model': 9.85528015564482e-05, 'epoch': 8.48}) +[2026-06-20 20:19:58,229][starVLA.training.train_starvla][INFO] - Step 4587, Loss: {'train/loss': 0.08561492711305618, 'train/grad_norm_pre_clip': 4.712133884429932, 'timing/data': 0.00655767903663218, 'timing/model': 1.148741299053654, 'learning_rate/qwen_vl_interface': 9.854533196443661e-06, 'learning_rate/action_model': 9.854533196443661e-05, 'epoch': 8.48}) +[2026-06-20 20:20:03,170][starVLA.training.train_starvla][INFO] - Step 4588, Loss: {'train/loss': 0.11125976592302322, 'train/grad_norm_pre_clip': 6.081856727600098, 'timing/data': 0.006953687989152968, 'timing/model': 1.1979120369069278, 'learning_rate/qwen_vl_interface': 9.853784344492833e-06, 'learning_rate/action_model': 9.853784344492833e-05, 'epoch': 8.48}) +[2026-06-20 20:20:07,872][starVLA.training.train_starvla][INFO] - Step 4589, Loss: {'train/loss': 0.19855912029743195, 'train/grad_norm_pre_clip': 8.965517044067383, 'timing/data': 0.006503198994323611, 'timing/model': 1.3845345081062987, 'learning_rate/qwen_vl_interface': 9.853033600100162e-06, 'learning_rate/action_model': 9.853033600100162e-05, 'epoch': 8.49}) +[2026-06-20 20:20:12,393][starVLA.training.train_starvla][INFO] - Step 4590, Loss: {'train/loss': 0.3602587878704071, 'train/grad_norm_pre_clip': 12.519957542419434, 'timing/data': 0.006520479917526245, 'timing/model': 1.1451211230596527, 'learning_rate/qwen_vl_interface': 9.85228096357425e-06, 'learning_rate/action_model': 9.852280963574249e-05, 'epoch': 8.49}) +[2026-06-20 20:20:17,112][starVLA.training.train_starvla][INFO] - Step 4591, Loss: {'train/loss': 0.25807303190231323, 'train/grad_norm_pre_clip': 8.529919624328613, 'timing/data': 0.006549899000674486, 'timing/model': 1.1463059559464455, 'learning_rate/qwen_vl_interface': 9.851526435224475e-06, 'learning_rate/action_model': 9.851526435224474e-05, 'epoch': 8.49}) +[2026-06-20 20:20:21,861][starVLA.training.train_starvla][INFO] - Step 4592, Loss: {'train/loss': 0.11378798633813858, 'train/grad_norm_pre_clip': 5.886353492736816, 'timing/data': 0.006338399951346219, 'timing/model': 1.1446910530794412, 'learning_rate/qwen_vl_interface': 9.850770015360999e-06, 'learning_rate/action_model': 9.850770015360999e-05, 'epoch': 8.49}) +[2026-06-20 20:20:26,655][starVLA.training.train_starvla][INFO] - Step 4593, Loss: {'train/loss': 0.1723051369190216, 'train/grad_norm_pre_clip': 9.270614624023438, 'timing/data': 0.006526949000544846, 'timing/model': 1.145875354995951, 'learning_rate/qwen_vl_interface': 9.850011704294755e-06, 'learning_rate/action_model': 9.850011704294755e-05, 'epoch': 8.49}) +[2026-06-20 20:20:31,507][starVLA.training.train_starvla][INFO] - Step 4594, Loss: {'train/loss': 0.10365529358386993, 'train/grad_norm_pre_clip': 6.901144981384277, 'timing/data': 0.006474809953942895, 'timing/model': 1.1464065610198304, 'learning_rate/qwen_vl_interface': 9.849251502337458e-06, 'learning_rate/action_model': 9.849251502337458e-05, 'epoch': 8.5}) +[2026-06-20 20:20:36,299][starVLA.training.train_starvla][INFO] - Step 4595, Loss: {'train/loss': 0.18806838989257812, 'train/grad_norm_pre_clip': 4.892978668212891, 'timing/data': 0.010240852017886937, 'timing/model': 1.4517303080065176, 'learning_rate/qwen_vl_interface': 9.848489409801598e-06, 'learning_rate/action_model': 9.848489409801598e-05, 'epoch': 8.5}) +[2026-06-20 20:20:40,904][starVLA.training.train_starvla][INFO] - Step 4596, Loss: {'train/loss': 0.13195331394672394, 'train/grad_norm_pre_clip': 7.850947856903076, 'timing/data': 0.006453251000493765, 'timing/model': 1.3330294699408114, 'learning_rate/qwen_vl_interface': 9.84772542700044e-06, 'learning_rate/action_model': 9.84772542700044e-05, 'epoch': 8.5}) +[2026-06-20 20:20:45,908][starVLA.training.train_starvla][INFO] - Step 4597, Loss: {'train/loss': 0.2708904445171356, 'train/grad_norm_pre_clip': 6.529291152954102, 'timing/data': 0.020780054037459195, 'timing/model': 1.213318796013482, 'learning_rate/qwen_vl_interface': 9.846959554248032e-06, 'learning_rate/action_model': 9.84695955424803e-05, 'epoch': 8.5}) +[2026-06-20 20:20:50,704][starVLA.training.train_starvla][INFO] - Step 4598, Loss: {'train/loss': 0.15246187150478363, 'train/grad_norm_pre_clip': 6.098788261413574, 'timing/data': 0.006911647971719503, 'timing/model': 1.201525259995833, 'learning_rate/qwen_vl_interface': 9.846191791859192e-06, 'learning_rate/action_model': 9.846191791859192e-05, 'epoch': 8.5}) +[2026-06-20 20:20:55,572][starVLA.training.train_starvla][INFO] - Step 4599, Loss: {'train/loss': 0.06995377689599991, 'train/grad_norm_pre_clip': 3.8969430923461914, 'timing/data': 0.006642388994805515, 'timing/model': 1.147046381025575, 'learning_rate/qwen_vl_interface': 9.845422140149518e-06, 'learning_rate/action_model': 9.845422140149519e-05, 'epoch': 8.51}) +[2026-06-20 20:21:00,536][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 4600 +[2026-06-20 20:27:30,243][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 4600 +[2026-06-20 20:27:30,243][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 4600 +[2026-06-20 20:34:25,657][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 4600 +[2026-06-20 20:34:25,659][starVLA.training.train_starvla][INFO] - Step 4600, Loss: {'train/loss': 0.09838841110467911, 'train/grad_norm_pre_clip': 7.314492702484131, 'eval/loss': 0.48220351338386536, 'eval/latency_0/loss': 0.5561911463737488, 'eval/latency_1/loss': 0.7019485831260681, 'eval/latency_2/loss': 0.6922750473022461, 'eval/latency_3/loss': 0.3548836410045624, 'eval/latency_4/loss': 0.08500908315181732, 'eval/flappy/loss': 0.48220351338386536, 'eval/flappy/latency_0/loss': 0.5561911463737488, 'eval/flappy/latency_1/loss': 0.7019485831260681, 'eval/flappy/latency_2/loss': 0.6922750473022461, 'eval/flappy/latency_3/loss': 0.3548836410045624, 'eval/flappy/latency_4/loss': 0.08500908315181732, 'eval/flap_precision': 0.1935483870967742, 'eval/flap_recall': 0.08727272727272728, 'eval/flap_f1': 0.12030075187969924, 'eval/flap_rate_model': 0.034444444444444444, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.12030075187969924, 'eval/latency_0/flap_precision': 0.1935483870967742, 'eval/latency_0/flap_recall': 0.08727272727272728, 'eval/latency_0/flap_f1': 0.12030075187969924, 'eval/latency_0/flap_rate_model': 0.034444444444444444, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.12030075187969924, 'timing/data': 0.006336610997095704, 'timing/model': 1.2192012059967965, 'learning_rate/qwen_vl_interface': 9.84465059943539e-06, 'learning_rate/action_model': 9.84465059943539e-05, 'epoch': 8.51}) +[2026-06-20 20:34:25,660][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_4600_state +[2026-06-20 20:34:25,664][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-20 20:34:34,232][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_4600_state/model.safetensors +[2026-06-20 20:34:47,664][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_4600_state/optimizer.bin +[2026-06-20 20:34:47,665][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_4600_state/sampler.bin +[2026-06-20 20:34:47,665][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_4600_state/sampler_1.bin +[2026-06-20 20:34:47,666][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_4600_state/random_states_0.pkl +[2026-06-20 20:34:50,278][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4600 +[2026-06-20 20:34:50,278][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-20 20:34:50,293][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-20 20:34:54,907][starVLA.training.train_starvla][INFO] - Step 4601, Loss: {'train/loss': 0.0947924554347992, 'train/grad_norm_pre_clip': 4.255918502807617, 'timing/data': 0.006434611044824123, 'timing/model': 1.1267258540028706, 'learning_rate/qwen_vl_interface': 9.843877170033956e-06, 'learning_rate/action_model': 9.843877170033954e-05, 'epoch': 8.51}) +[2026-06-20 20:34:59,627][starVLA.training.train_starvla][INFO] - Step 4602, Loss: {'train/loss': 0.12338393181562424, 'train/grad_norm_pre_clip': 7.926476001739502, 'timing/data': 0.00636217906139791, 'timing/model': 1.356076505035162, 'learning_rate/qwen_vl_interface': 9.84310185226314e-06, 'learning_rate/action_model': 9.84310185226314e-05, 'epoch': 8.51}) +[2026-06-20 20:35:04,201][starVLA.training.train_starvla][INFO] - Step 4603, Loss: {'train/loss': 0.11949563771486282, 'train/grad_norm_pre_clip': 6.4909892082214355, 'timing/data': 0.008546625031158328, 'timing/model': 1.216986265964806, 'learning_rate/qwen_vl_interface': 9.842324646441652e-06, 'learning_rate/action_model': 9.842324646441652e-05, 'epoch': 8.51}) +[2026-06-20 20:35:09,036][starVLA.training.train_starvla][INFO] - Step 4604, Loss: {'train/loss': 0.05220365151762962, 'train/grad_norm_pre_clip': 2.6415822505950928, 'timing/data': 0.010410649003461003, 'timing/model': 1.2182979830540717, 'learning_rate/qwen_vl_interface': 9.841545552888968e-06, 'learning_rate/action_model': 9.841545552888967e-05, 'epoch': 8.51}) +[2026-06-20 20:35:13,835][starVLA.training.train_starvla][INFO] - Step 4605, Loss: {'train/loss': 0.198777973651886, 'train/grad_norm_pre_clip': 6.443330764770508, 'timing/data': 0.006484808051027358, 'timing/model': 1.1502669269684702, 'learning_rate/qwen_vl_interface': 9.840764571925346e-06, 'learning_rate/action_model': 9.840764571925345e-05, 'epoch': 8.52}) +[2026-06-20 20:35:18,576][starVLA.training.train_starvla][INFO] - Step 4606, Loss: {'train/loss': 0.09250523149967194, 'train/grad_norm_pre_clip': 7.294497966766357, 'timing/data': 0.007417483953759074, 'timing/model': 1.1976707669673488, 'learning_rate/qwen_vl_interface': 9.839981703871815e-06, 'learning_rate/action_model': 9.839981703871814e-05, 'epoch': 8.52}) +[2026-06-20 20:35:23,291][starVLA.training.train_starvla][INFO] - Step 4607, Loss: {'train/loss': 0.16397003829479218, 'train/grad_norm_pre_clip': 7.92633056640625, 'timing/data': 0.009287177002988756, 'timing/model': 1.4569135669153184, 'learning_rate/qwen_vl_interface': 9.839196949050185e-06, 'learning_rate/action_model': 9.839196949050185e-05, 'epoch': 8.52}) +[2026-06-20 20:35:27,813][starVLA.training.train_starvla][INFO] - Step 4608, Loss: {'train/loss': 0.053517378866672516, 'train/grad_norm_pre_clip': 4.67366886138916, 'timing/data': 0.008392070070840418, 'timing/model': 1.206013065064326, 'learning_rate/qwen_vl_interface': 9.838410307783038e-06, 'learning_rate/action_model': 9.838410307783038e-05, 'epoch': 8.52}) +[2026-06-20 20:35:32,651][starVLA.training.train_starvla][INFO] - Step 4609, Loss: {'train/loss': 0.11622630059719086, 'train/grad_norm_pre_clip': 7.170516490936279, 'timing/data': 0.008009565994143486, 'timing/model': 1.214576985919848, 'learning_rate/qwen_vl_interface': 9.837621780393734e-06, 'learning_rate/action_model': 9.837621780393733e-05, 'epoch': 8.52}) +[2026-06-20 20:35:37,513][starVLA.training.train_starvla][INFO] - Step 4610, Loss: {'train/loss': 0.032572027295827866, 'train/grad_norm_pre_clip': 2.831958055496216, 'timing/data': 0.00647926505189389, 'timing/model': 1.2093871800461784, 'learning_rate/qwen_vl_interface': 9.836831367206406e-06, 'learning_rate/action_model': 9.836831367206406e-05, 'epoch': 8.53}) +[2026-06-20 20:35:42,172][starVLA.training.train_starvla][INFO] - Step 4611, Loss: {'train/loss': 0.032985322177410126, 'train/grad_norm_pre_clip': 3.7569127082824707, 'timing/data': 0.00669714598916471, 'timing/model': 1.1383550880709663, 'learning_rate/qwen_vl_interface': 9.836039068545962e-06, 'learning_rate/action_model': 9.836039068545961e-05, 'epoch': 8.53}) +[2026-06-20 20:35:46,918][starVLA.training.train_starvla][INFO] - Step 4612, Loss: {'train/loss': 0.04052605852484703, 'train/grad_norm_pre_clip': 5.19370698928833, 'timing/data': 0.006532794097438455, 'timing/model': 1.13769602694083, 'learning_rate/qwen_vl_interface': 9.835244884738086e-06, 'learning_rate/action_model': 9.835244884738085e-05, 'epoch': 8.53}) +[2026-06-20 20:35:51,761][starVLA.training.train_starvla][INFO] - Step 4613, Loss: {'train/loss': 0.21823741495609283, 'train/grad_norm_pre_clip': 7.872747898101807, 'timing/data': 0.008849038975313306, 'timing/model': 1.4454711019061506, 'learning_rate/qwen_vl_interface': 9.83444881610924e-06, 'learning_rate/action_model': 9.834448816109238e-05, 'epoch': 8.53}) +[2026-06-20 20:35:56,352][starVLA.training.train_starvla][INFO] - Step 4614, Loss: {'train/loss': 0.17944280803203583, 'train/grad_norm_pre_clip': 7.848377704620361, 'timing/data': 0.006361840991303325, 'timing/model': 1.21188308193814, 'learning_rate/qwen_vl_interface': 9.833650862986652e-06, 'learning_rate/action_model': 9.833650862986653e-05, 'epoch': 8.53}) +[2026-06-20 20:36:01,204][starVLA.training.train_starvla][INFO] - Step 4615, Loss: {'train/loss': 0.3422473669052124, 'train/grad_norm_pre_clip': 18.374448776245117, 'timing/data': 0.007571008987724781, 'timing/model': 1.2131655750563368, 'learning_rate/qwen_vl_interface': 9.832851025698335e-06, 'learning_rate/action_model': 9.832851025698335e-05, 'epoch': 8.53}) +[2026-06-20 20:36:06,011][starVLA.training.train_starvla][INFO] - Step 4616, Loss: {'train/loss': 0.06553579121828079, 'train/grad_norm_pre_clip': 7.0983171463012695, 'timing/data': 0.006287859985604882, 'timing/model': 1.2662340570241213, 'learning_rate/qwen_vl_interface': 9.832049304573072e-06, 'learning_rate/action_model': 9.832049304573072e-05, 'epoch': 8.54}) +[2026-06-20 20:36:10,777][starVLA.training.train_starvla][INFO] - Step 4617, Loss: {'train/loss': 0.3220835030078888, 'train/grad_norm_pre_clip': 22.991365432739258, 'timing/data': 0.006370010902173817, 'timing/model': 1.211609233985655, 'learning_rate/qwen_vl_interface': 9.831245699940419e-06, 'learning_rate/action_model': 9.831245699940418e-05, 'epoch': 8.54}) +[2026-06-20 20:36:15,565][starVLA.training.train_starvla][INFO] - Step 4618, Loss: {'train/loss': 0.09045954048633575, 'train/grad_norm_pre_clip': 7.4308953285217285, 'timing/data': 0.006506461068056524, 'timing/model': 1.2380161980399862, 'learning_rate/qwen_vl_interface': 9.830440212130707e-06, 'learning_rate/action_model': 9.830440212130708e-05, 'epoch': 8.54}) +[2026-06-20 20:36:20,448][starVLA.training.train_starvla][INFO] - Step 4619, Loss: {'train/loss': 0.13650688529014587, 'train/grad_norm_pre_clip': 7.795812606811523, 'timing/data': 0.007398896035738289, 'timing/model': 1.4220705500338227, 'learning_rate/qwen_vl_interface': 9.829632841475041e-06, 'learning_rate/action_model': 9.829632841475041e-05, 'epoch': 8.54}) +[2026-06-20 20:36:24,969][starVLA.training.train_starvla][INFO] - Step 4620, Loss: {'train/loss': 0.03888269513845444, 'train/grad_norm_pre_clip': 3.010963201522827, 'timing/data': 0.007904367987066507, 'timing/model': 1.197592254029587, 'learning_rate/qwen_vl_interface': 9.828823588305305e-06, 'learning_rate/action_model': 9.828823588305303e-05, 'epoch': 8.54}) +[2026-06-20 20:36:29,806][starVLA.training.train_starvla][INFO] - Step 4621, Loss: {'train/loss': 0.25278159976005554, 'train/grad_norm_pre_clip': 5.655937671661377, 'timing/data': 0.0064082680037245154, 'timing/model': 1.211139648919925, 'learning_rate/qwen_vl_interface': 9.828012452954146e-06, 'learning_rate/action_model': 9.828012452954145e-05, 'epoch': 8.55}) +[2026-06-20 20:36:34,691][starVLA.training.train_starvla][INFO] - Step 4622, Loss: {'train/loss': 0.09317149966955185, 'train/grad_norm_pre_clip': 9.793572425842285, 'timing/data': 0.007889917003922164, 'timing/model': 1.1977459701010957, 'learning_rate/qwen_vl_interface': 9.827199435754995e-06, 'learning_rate/action_model': 9.827199435754995e-05, 'epoch': 8.55}) +[2026-06-20 20:36:39,535][starVLA.training.train_starvla][INFO] - Step 4623, Loss: {'train/loss': 0.06721669435501099, 'train/grad_norm_pre_clip': 5.9189887046813965, 'timing/data': 0.006425997940823436, 'timing/model': 1.1429728030925617, 'learning_rate/qwen_vl_interface': 9.826384537042052e-06, 'learning_rate/action_model': 9.826384537042052e-05, 'epoch': 8.55}) +[2026-06-20 20:36:44,364][starVLA.training.train_starvla][INFO] - Step 4624, Loss: {'train/loss': 0.11635260283946991, 'train/grad_norm_pre_clip': 11.918551445007324, 'timing/data': 0.006350056966766715, 'timing/model': 1.3766244520666078, 'learning_rate/qwen_vl_interface': 9.825567757150291e-06, 'learning_rate/action_model': 9.82556775715029e-05, 'epoch': 8.55}) +[2026-06-20 20:36:48,996][starVLA.training.train_starvla][INFO] - Step 4625, Loss: {'train/loss': 0.14030443131923676, 'train/grad_norm_pre_clip': 8.680733680725098, 'timing/data': 0.006340235937386751, 'timing/model': 1.294160602032207, 'learning_rate/qwen_vl_interface': 9.82474909641546e-06, 'learning_rate/action_model': 9.82474909641546e-05, 'epoch': 8.55}) +[2026-06-20 20:36:53,530][starVLA.training.train_starvla][INFO] - Step 4626, Loss: {'train/loss': 0.3825821578502655, 'train/grad_norm_pre_clip': 18.18052101135254, 'timing/data': 0.006367155932821333, 'timing/model': 1.144515193067491, 'learning_rate/qwen_vl_interface': 9.823928555174078e-06, 'learning_rate/action_model': 9.823928555174077e-05, 'epoch': 8.56}) +[2026-06-20 20:36:58,304][starVLA.training.train_starvla][INFO] - Step 4627, Loss: {'train/loss': 0.15125823020935059, 'train/grad_norm_pre_clip': 5.680703639984131, 'timing/data': 0.0065775769762694836, 'timing/model': 1.1436487509636208, 'learning_rate/qwen_vl_interface': 9.82310613376344e-06, 'learning_rate/action_model': 9.82310613376344e-05, 'epoch': 8.56}) +[2026-06-20 20:37:03,007][starVLA.training.train_starvla][INFO] - Step 4628, Loss: {'train/loss': 0.04744004085659981, 'train/grad_norm_pre_clip': 5.607616901397705, 'timing/data': 0.006418965058401227, 'timing/model': 1.1436007079901174, 'learning_rate/qwen_vl_interface': 9.822281832521613e-06, 'learning_rate/action_model': 9.822281832521613e-05, 'epoch': 8.56}) +[2026-06-20 20:37:07,808][starVLA.training.train_starvla][INFO] - Step 4629, Loss: {'train/loss': 0.22313843667507172, 'train/grad_norm_pre_clip': 11.985174179077148, 'timing/data': 0.0063317540334537625, 'timing/model': 1.1439413140760735, 'learning_rate/qwen_vl_interface': 9.821455651787432e-06, 'learning_rate/action_model': 9.821455651787432e-05, 'epoch': 8.56}) +[2026-06-20 20:37:12,517][starVLA.training.train_starvla][INFO] - Step 4630, Loss: {'train/loss': 0.055237267166376114, 'train/grad_norm_pre_clip': 4.333652496337891, 'timing/data': 0.006442454061470926, 'timing/model': 1.3744067780207843, 'learning_rate/qwen_vl_interface': 9.820627591900514e-06, 'learning_rate/action_model': 9.820627591900513e-05, 'epoch': 8.56}) +[2026-06-20 20:37:17,208][starVLA.training.train_starvla][INFO] - Step 4631, Loss: {'train/loss': 0.18527625501155853, 'train/grad_norm_pre_clip': 9.852057456970215, 'timing/data': 0.0063938540406525135, 'timing/model': 1.296783820958808, 'learning_rate/qwen_vl_interface': 9.819797653201238e-06, 'learning_rate/action_model': 9.819797653201238e-05, 'epoch': 8.56}) +[2026-06-20 20:37:21,760][starVLA.training.train_starvla][INFO] - Step 4632, Loss: {'train/loss': 0.10252414643764496, 'train/grad_norm_pre_clip': 10.53042221069336, 'timing/data': 0.006285742972977459, 'timing/model': 1.1437195379985496, 'learning_rate/qwen_vl_interface': 9.818965836030765e-06, 'learning_rate/action_model': 9.818965836030764e-05, 'epoch': 8.57}) +[2026-06-20 20:37:26,538][starVLA.training.train_starvla][INFO] - Step 4633, Loss: {'train/loss': 0.04115613177418709, 'train/grad_norm_pre_clip': 3.341693162918091, 'timing/data': 0.007613069959916174, 'timing/model': 1.1990574379451573, 'learning_rate/qwen_vl_interface': 9.81813214073102e-06, 'learning_rate/action_model': 9.81813214073102e-05, 'epoch': 8.57}) +[2026-06-20 20:37:31,342][starVLA.training.train_starvla][INFO] - Step 4634, Loss: {'train/loss': 0.06522821635007858, 'train/grad_norm_pre_clip': 6.139864444732666, 'timing/data': 0.00621708098333329, 'timing/model': 1.2169126420048997, 'learning_rate/qwen_vl_interface': 9.817296567644709e-06, 'learning_rate/action_model': 9.817296567644709e-05, 'epoch': 8.57}) +[2026-06-20 20:37:36,168][starVLA.training.train_starvla][INFO] - Step 4635, Loss: {'train/loss': 0.055465538054704666, 'train/grad_norm_pre_clip': 10.470789909362793, 'timing/data': 0.007652887026779354, 'timing/model': 1.196563804987818, 'learning_rate/qwen_vl_interface': 9.816459117115296e-06, 'learning_rate/action_model': 9.816459117115297e-05, 'epoch': 8.57}) +[2026-06-20 20:37:40,867][starVLA.training.train_starvla][INFO] - Step 4636, Loss: {'train/loss': 0.12487345933914185, 'train/grad_norm_pre_clip': 10.417159080505371, 'timing/data': 0.00628355098888278, 'timing/model': 1.3759944640332833, 'learning_rate/qwen_vl_interface': 9.815619789487035e-06, 'learning_rate/action_model': 9.815619789487034e-05, 'epoch': 8.57}) +[2026-06-20 20:37:45,351][starVLA.training.train_starvla][INFO] - Step 4637, Loss: {'train/loss': 0.022228654474020004, 'train/grad_norm_pre_clip': 2.929100275039673, 'timing/data': 0.006447961088269949, 'timing/model': 1.1459557600319386, 'learning_rate/qwen_vl_interface': 9.814778585104933e-06, 'learning_rate/action_model': 9.814778585104933e-05, 'epoch': 8.58}) +[2026-06-20 20:37:50,175][starVLA.training.train_starvla][INFO] - Step 4638, Loss: {'train/loss': 0.060577183961868286, 'train/grad_norm_pre_clip': 6.062981605529785, 'timing/data': 0.006442751036956906, 'timing/model': 1.1449790350161493, 'learning_rate/qwen_vl_interface': 9.813935504314784e-06, 'learning_rate/action_model': 9.813935504314784e-05, 'epoch': 8.58}) +[2026-06-20 20:37:54,974][starVLA.training.train_starvla][INFO] - Step 4639, Loss: {'train/loss': 0.031734008342027664, 'train/grad_norm_pre_clip': 7.452304363250732, 'timing/data': 0.007628206047229469, 'timing/model': 1.2093086519744247, 'learning_rate/qwen_vl_interface': 9.813090547463144e-06, 'learning_rate/action_model': 9.813090547463143e-05, 'epoch': 8.58}) +[2026-06-20 20:37:59,652][starVLA.training.train_starvla][INFO] - Step 4640, Loss: {'train/loss': 0.09344377368688583, 'train/grad_norm_pre_clip': 10.834386825561523, 'timing/data': 0.006068247952498496, 'timing/model': 1.1429118720116094, 'learning_rate/qwen_vl_interface': 9.812243714897342e-06, 'learning_rate/action_model': 9.812243714897343e-05, 'epoch': 8.58}) +[2026-06-20 20:38:23,844][starVLA.training.train_starvla][INFO] - Step 4641, Loss: {'train/loss': 0.061983752995729446, 'train/grad_norm_pre_clip': 8.711831092834473, 'timing/data': 18.601578275905922, 'timing/model': 1.2099200040102005, 'learning_rate/qwen_vl_interface': 9.811395006965481e-06, 'learning_rate/action_model': 9.81139500696548e-05, 'epoch': 8.58}) +[2026-06-20 20:38:28,568][starVLA.training.train_starvla][INFO] - Step 4642, Loss: {'train/loss': 0.01458531804382801, 'train/grad_norm_pre_clip': 1.2215020656585693, 'timing/data': 0.006196595961228013, 'timing/model': 1.3646855941042304, 'learning_rate/qwen_vl_interface': 9.81054442401643e-06, 'learning_rate/action_model': 9.81054442401643e-05, 'epoch': 8.58}) +[2026-06-20 20:38:33,051][starVLA.training.train_starvla][INFO] - Step 4643, Loss: {'train/loss': 0.0515250489115715, 'train/grad_norm_pre_clip': 6.137875080108643, 'timing/data': 0.013544226996600628, 'timing/model': 1.1906334049999714, 'learning_rate/qwen_vl_interface': 9.809691966399834e-06, 'learning_rate/action_model': 9.809691966399835e-05, 'epoch': 8.59}) +[2026-06-20 20:38:37,884][starVLA.training.train_starvla][INFO] - Step 4644, Loss: {'train/loss': 0.17575204372406006, 'train/grad_norm_pre_clip': 8.180325508117676, 'timing/data': 0.01599669607821852, 'timing/model': 1.2025386788882315, 'learning_rate/qwen_vl_interface': 9.808837634466104e-06, 'learning_rate/action_model': 9.808837634466104e-05, 'epoch': 8.59}) +[2026-06-20 20:38:42,657][starVLA.training.train_starvla][INFO] - Step 4645, Loss: {'train/loss': 0.08339156210422516, 'train/grad_norm_pre_clip': 12.54965877532959, 'timing/data': 0.011990058002993464, 'timing/model': 1.137914621969685, 'learning_rate/qwen_vl_interface': 9.807981428566428e-06, 'learning_rate/action_model': 9.807981428566427e-05, 'epoch': 8.59}) +[2026-06-20 20:38:47,407][starVLA.training.train_starvla][INFO] - Step 4646, Loss: {'train/loss': 0.07343993335962296, 'train/grad_norm_pre_clip': 6.9047698974609375, 'timing/data': 0.012065747985616326, 'timing/model': 1.1421893290244043, 'learning_rate/qwen_vl_interface': 9.807123349052754e-06, 'learning_rate/action_model': 9.807123349052753e-05, 'epoch': 8.59}) +[2026-06-20 20:38:52,255][starVLA.training.train_starvla][INFO] - Step 4647, Loss: {'train/loss': 0.08547016978263855, 'train/grad_norm_pre_clip': 10.338814735412598, 'timing/data': 0.016489275032654405, 'timing/model': 1.212318824022077, 'learning_rate/qwen_vl_interface': 9.80626339627781e-06, 'learning_rate/action_model': 9.806263396277809e-05, 'epoch': 8.59}) +[2026-06-20 20:38:57,087][starVLA.training.train_starvla][INFO] - Step 4648, Loss: {'train/loss': 0.10160853713750839, 'train/grad_norm_pre_clip': 10.666268348693848, 'timing/data': 0.011900934972800314, 'timing/model': 1.3841927100438625, 'learning_rate/qwen_vl_interface': 9.805401570595087e-06, 'learning_rate/action_model': 9.805401570595087e-05, 'epoch': 8.6}) +[2026-06-20 20:39:01,628][starVLA.training.train_starvla][INFO] - Step 4649, Loss: {'train/loss': 0.17272689938545227, 'train/grad_norm_pre_clip': 11.678282737731934, 'timing/data': 0.013372862013056874, 'timing/model': 1.161124787060544, 'learning_rate/qwen_vl_interface': 9.804537872358853e-06, 'learning_rate/action_model': 9.804537872358853e-05, 'epoch': 8.6}) +[2026-06-20 20:39:06,295][starVLA.training.train_starvla][INFO] - Step 4650, Loss: {'train/loss': 0.11452215909957886, 'train/grad_norm_pre_clip': 9.329565048217773, 'timing/data': 0.012264576042070985, 'timing/model': 1.1456255209632218, 'learning_rate/qwen_vl_interface': 9.80367230192414e-06, 'learning_rate/action_model': 9.803672301924138e-05, 'epoch': 8.6}) +[2026-06-20 20:39:10,970][starVLA.training.train_starvla][INFO] - Step 4651, Loss: {'train/loss': 0.11090919375419617, 'train/grad_norm_pre_clip': 11.407147407531738, 'timing/data': 0.012028504977934062, 'timing/model': 1.1442322420189157, 'learning_rate/qwen_vl_interface': 9.80280485964675e-06, 'learning_rate/action_model': 9.80280485964675e-05, 'epoch': 8.6}) +[2026-06-20 20:39:15,636][starVLA.training.train_starvla][INFO] - Step 4652, Loss: {'train/loss': 0.01019253395497799, 'train/grad_norm_pre_clip': 2.083456516265869, 'timing/data': 0.012266804929822683, 'timing/model': 1.1448105479357764, 'learning_rate/qwen_vl_interface': 9.801935545883255e-06, 'learning_rate/action_model': 9.801935545883255e-05, 'epoch': 8.6}) +[2026-06-20 20:39:20,344][starVLA.training.train_starvla][INFO] - Step 4653, Loss: {'train/loss': 0.00785788893699646, 'train/grad_norm_pre_clip': 0.8738143444061279, 'timing/data': 0.01789112389087677, 'timing/model': 1.1614442729623988, 'learning_rate/qwen_vl_interface': 9.801064360991e-06, 'learning_rate/action_model': 9.801064360991e-05, 'epoch': 8.61}) +[2026-06-20 20:39:25,007][starVLA.training.train_starvla][INFO] - Step 4654, Loss: {'train/loss': 0.04781300947070122, 'train/grad_norm_pre_clip': 7.46157693862915, 'timing/data': 0.013220587046816945, 'timing/model': 1.3954478980740532, 'learning_rate/qwen_vl_interface': 9.800191305328095e-06, 'learning_rate/action_model': 9.800191305328094e-05, 'epoch': 8.61}) +[2026-06-20 20:39:29,477][starVLA.training.train_starvla][INFO] - Step 4655, Loss: {'train/loss': 0.003560108132660389, 'train/grad_norm_pre_clip': 0.5006805062294006, 'timing/data': 0.013374915928579867, 'timing/model': 1.1586448970483616, 'learning_rate/qwen_vl_interface': 9.79931637925342e-06, 'learning_rate/action_model': 9.79931637925342e-05, 'epoch': 8.61}) +[2026-06-20 20:39:34,177][starVLA.training.train_starvla][INFO] - Step 4656, Loss: {'train/loss': 0.17754259705543518, 'train/grad_norm_pre_clip': 18.348609924316406, 'timing/data': 0.013766897027380764, 'timing/model': 1.1592519549885765, 'learning_rate/qwen_vl_interface': 9.798439583126625e-06, 'learning_rate/action_model': 9.798439583126625e-05, 'epoch': 8.61}) +[2026-06-20 20:39:38,824][starVLA.training.train_starvla][INFO] - Step 4657, Loss: {'train/loss': 0.02909587137401104, 'train/grad_norm_pre_clip': 5.838430881500244, 'timing/data': 0.012249312014319003, 'timing/model': 1.160996128921397, 'learning_rate/qwen_vl_interface': 9.797560917308126e-06, 'learning_rate/action_model': 9.797560917308126e-05, 'epoch': 8.61}) +[2026-06-20 20:39:43,528][starVLA.training.train_starvla][INFO] - Step 4658, Loss: {'train/loss': 0.05676913633942604, 'train/grad_norm_pre_clip': 6.214657783508301, 'timing/data': 0.012498890981078148, 'timing/model': 1.1488347869599238, 'learning_rate/qwen_vl_interface': 9.796680382159112e-06, 'learning_rate/action_model': 9.79668038215911e-05, 'epoch': 8.61}) +[2026-06-20 20:39:48,208][starVLA.training.train_starvla][INFO] - Step 4659, Loss: {'train/loss': 0.37716519832611084, 'train/grad_norm_pre_clip': 42.47854232788086, 'timing/data': 0.012288999976590276, 'timing/model': 1.1426820650231093, 'learning_rate/qwen_vl_interface': 9.795797978041533e-06, 'learning_rate/action_model': 9.795797978041533e-05, 'epoch': 8.62}) +[2026-06-20 20:39:53,109][starVLA.training.train_starvla][INFO] - Step 4660, Loss: {'train/loss': 0.012676847167313099, 'train/grad_norm_pre_clip': 2.651240587234497, 'timing/data': 0.011258075945079327, 'timing/model': 1.4436263300012797, 'learning_rate/qwen_vl_interface': 9.794913705318117e-06, 'learning_rate/action_model': 9.794913705318118e-05, 'epoch': 8.62}) +[2026-06-20 20:39:57,975][starVLA.training.train_starvla][INFO] - Step 4661, Loss: {'train/loss': 0.24418072402477264, 'train/grad_norm_pre_clip': 12.083596229553223, 'timing/data': 0.012565869954414666, 'timing/model': 1.3894020300358534, 'learning_rate/qwen_vl_interface': 9.794027564352352e-06, 'learning_rate/action_model': 9.794027564352352e-05, 'epoch': 8.62}) +[2026-06-20 20:40:02,504][starVLA.training.train_starvla][INFO] - Step 4662, Loss: {'train/loss': 0.18820562958717346, 'train/grad_norm_pre_clip': 16.83891487121582, 'timing/data': 0.012531199026852846, 'timing/model': 1.1948899539420381, 'learning_rate/qwen_vl_interface': 9.793139555508501e-06, 'learning_rate/action_model': 9.7931395555085e-05, 'epoch': 8.62}) +[2026-06-20 20:40:07,391][starVLA.training.train_starvla][INFO] - Step 4663, Loss: {'train/loss': 0.021735504269599915, 'train/grad_norm_pre_clip': 6.037020683288574, 'timing/data': 0.013569111935794353, 'timing/model': 1.2296466859988868, 'learning_rate/qwen_vl_interface': 9.792249679151585e-06, 'learning_rate/action_model': 9.792249679151585e-05, 'epoch': 8.62}) +[2026-06-20 20:40:12,108][starVLA.training.train_starvla][INFO] - Step 4664, Loss: {'train/loss': 0.03722725063562393, 'train/grad_norm_pre_clip': 5.591226100921631, 'timing/data': 0.011452754959464073, 'timing/model': 1.1551002589985728, 'learning_rate/qwen_vl_interface': 9.791357935647403e-06, 'learning_rate/action_model': 9.791357935647403e-05, 'epoch': 8.63}) +[2026-06-20 20:40:16,978][starVLA.training.train_starvla][INFO] - Step 4665, Loss: {'train/loss': 0.08433545380830765, 'train/grad_norm_pre_clip': 10.18603229522705, 'timing/data': 0.012044594972394407, 'timing/model': 1.232304197968915, 'learning_rate/qwen_vl_interface': 9.790464325362516e-06, 'learning_rate/action_model': 9.790464325362516e-05, 'epoch': 8.63}) +[2026-06-20 20:40:21,767][starVLA.training.train_starvla][INFO] - Step 4666, Loss: {'train/loss': 0.047326140105724335, 'train/grad_norm_pre_clip': 6.209599494934082, 'timing/data': 0.011130332946777344, 'timing/model': 1.3820724380202591, 'learning_rate/qwen_vl_interface': 9.789568848664254e-06, 'learning_rate/action_model': 9.789568848664253e-05, 'epoch': 8.63}) +[2026-06-20 20:40:26,575][starVLA.training.train_starvla][INFO] - Step 4667, Loss: {'train/loss': 0.1704997718334198, 'train/grad_norm_pre_clip': 18.600801467895508, 'timing/data': 0.012943946989253163, 'timing/model': 1.4305643449770287, 'learning_rate/qwen_vl_interface': 9.788671505920712e-06, 'learning_rate/action_model': 9.788671505920711e-05, 'epoch': 8.63}) +[2026-06-20 20:40:31,142][starVLA.training.train_starvla][INFO] - Step 4668, Loss: {'train/loss': 0.04203750938177109, 'train/grad_norm_pre_clip': 8.391733169555664, 'timing/data': 0.012153684045188129, 'timing/model': 1.144146905047819, 'learning_rate/qwen_vl_interface': 9.787772297500754e-06, 'learning_rate/action_model': 9.787772297500754e-05, 'epoch': 8.63}) +[2026-06-20 20:40:35,892][starVLA.training.train_starvla][INFO] - Step 4669, Loss: {'train/loss': 0.28914350271224976, 'train/grad_norm_pre_clip': 7.546278953552246, 'timing/data': 0.012158863013610244, 'timing/model': 1.1463650950463489, 'learning_rate/qwen_vl_interface': 9.78687122377401e-06, 'learning_rate/action_model': 9.78687122377401e-05, 'epoch': 8.63}) +[2026-06-20 20:40:40,592][starVLA.training.train_starvla][INFO] - Step 4670, Loss: {'train/loss': 0.19519862532615662, 'train/grad_norm_pre_clip': 19.1460018157959, 'timing/data': 0.012414074037224054, 'timing/model': 1.1433647410012782, 'learning_rate/qwen_vl_interface': 9.78596828511088e-06, 'learning_rate/action_model': 9.78596828511088e-05, 'epoch': 8.64}) +[2026-06-20 20:40:45,312][starVLA.training.train_starvla][INFO] - Step 4671, Loss: {'train/loss': 0.09694703668355942, 'train/grad_norm_pre_clip': 21.047847747802734, 'timing/data': 0.016255574068054557, 'timing/model': 1.1461622039787471, 'learning_rate/qwen_vl_interface': 9.785063481882527e-06, 'learning_rate/action_model': 9.785063481882527e-05, 'epoch': 8.64}) +[2026-06-20 20:40:50,131][starVLA.training.train_starvla][INFO] - Step 4672, Loss: {'train/loss': 0.16646362841129303, 'train/grad_norm_pre_clip': 12.544600486755371, 'timing/data': 0.014191507012583315, 'timing/model': 1.4353787919972092, 'learning_rate/qwen_vl_interface': 9.78415681446088e-06, 'learning_rate/action_model': 9.784156814460879e-05, 'epoch': 8.64}) +[2026-06-20 20:40:54,724][starVLA.training.train_starvla][INFO] - Step 4673, Loss: {'train/loss': 0.08816106617450714, 'train/grad_norm_pre_clip': 14.21809196472168, 'timing/data': 0.01201188110280782, 'timing/model': 1.1431597680784762, 'learning_rate/qwen_vl_interface': 9.783248283218638e-06, 'learning_rate/action_model': 9.783248283218638e-05, 'epoch': 8.64}) +[2026-06-20 20:40:59,555][starVLA.training.train_starvla][INFO] - Step 4674, Loss: {'train/loss': 0.14423759281635284, 'train/grad_norm_pre_clip': 25.9805850982666, 'timing/data': 0.01703716197516769, 'timing/model': 1.209277329966426, 'learning_rate/qwen_vl_interface': 9.782337888529261e-06, 'learning_rate/action_model': 9.782337888529261e-05, 'epoch': 8.64}) +[2026-06-20 20:41:04,367][starVLA.training.train_starvla][INFO] - Step 4675, Loss: {'train/loss': 0.1813950091600418, 'train/grad_norm_pre_clip': 10.602185249328613, 'timing/data': 0.011539808008819818, 'timing/model': 1.1425047379452735, 'learning_rate/qwen_vl_interface': 9.781425630766978e-06, 'learning_rate/action_model': 9.781425630766979e-05, 'epoch': 8.65}) +[2026-06-20 20:41:09,071][starVLA.training.train_starvla][INFO] - Step 4676, Loss: {'train/loss': 0.1523234248161316, 'train/grad_norm_pre_clip': 14.802128791809082, 'timing/data': 0.011477658059448004, 'timing/model': 1.1421984740300104, 'learning_rate/qwen_vl_interface': 9.780511510306789e-06, 'learning_rate/action_model': 9.780511510306788e-05, 'epoch': 8.65}) +[2026-06-20 20:41:13,728][starVLA.training.train_starvla][INFO] - Step 4677, Loss: {'train/loss': 0.031323838979005814, 'train/grad_norm_pre_clip': 5.3316755294799805, 'timing/data': 0.01153678703121841, 'timing/model': 1.152215447044, 'learning_rate/qwen_vl_interface': 9.779595527524447e-06, 'learning_rate/action_model': 9.779595527524447e-05, 'epoch': 8.65}) +[2026-06-20 20:41:18,463][starVLA.training.train_starvla][INFO] - Step 4678, Loss: {'train/loss': 0.04307403042912483, 'train/grad_norm_pre_clip': 12.313010215759277, 'timing/data': 0.011487536947242916, 'timing/model': 1.3506753859110177, 'learning_rate/qwen_vl_interface': 9.778677682796483e-06, 'learning_rate/action_model': 9.778677682796483e-05, 'epoch': 8.65}) +[2026-06-20 20:41:23,005][starVLA.training.train_starvla][INFO] - Step 4679, Loss: {'train/loss': 0.06050018221139908, 'train/grad_norm_pre_clip': 5.893087387084961, 'timing/data': 0.013628251035697758, 'timing/model': 1.2110652909614146, 'learning_rate/qwen_vl_interface': 9.777757976500185e-06, 'learning_rate/action_model': 9.777757976500186e-05, 'epoch': 8.65}) +[2026-06-20 20:41:27,703][starVLA.training.train_starvla][INFO] - Step 4680, Loss: {'train/loss': 0.033577997237443924, 'train/grad_norm_pre_clip': 5.94647741317749, 'timing/data': 0.011134244967252016, 'timing/model': 1.155845653032884, 'learning_rate/qwen_vl_interface': 9.776836409013614e-06, 'learning_rate/action_model': 9.776836409013615e-05, 'epoch': 8.66}) +[2026-06-20 20:41:32,521][starVLA.training.train_starvla][INFO] - Step 4681, Loss: {'train/loss': 0.10509766638278961, 'train/grad_norm_pre_clip': 12.919227600097656, 'timing/data': 0.01350774907041341, 'timing/model': 1.2008731999667361, 'learning_rate/qwen_vl_interface': 9.775912980715589e-06, 'learning_rate/action_model': 9.775912980715588e-05, 'epoch': 8.66}) +[2026-06-20 20:41:37,297][starVLA.training.train_starvla][INFO] - Step 4682, Loss: {'train/loss': 0.09315692633390427, 'train/grad_norm_pre_clip': 9.591574668884277, 'timing/data': 0.016070015029981732, 'timing/model': 1.2144678749609739, 'learning_rate/qwen_vl_interface': 9.774987691985694e-06, 'learning_rate/action_model': 9.774987691985694e-05, 'epoch': 8.66}) +[2026-06-20 20:41:42,142][starVLA.training.train_starvla][INFO] - Step 4683, Loss: {'train/loss': 0.07308963686227798, 'train/grad_norm_pre_clip': 19.640026092529297, 'timing/data': 0.013648879015818238, 'timing/model': 1.197720126947388, 'learning_rate/qwen_vl_interface': 9.774060543204283e-06, 'learning_rate/action_model': 9.774060543204283e-05, 'epoch': 8.66}) +[2026-06-20 20:41:46,897][starVLA.training.train_starvla][INFO] - Step 4684, Loss: {'train/loss': 0.010149423032999039, 'train/grad_norm_pre_clip': 1.4079701900482178, 'timing/data': 0.012185965082608163, 'timing/model': 1.3592249799985439, 'learning_rate/qwen_vl_interface': 9.773131534752472e-06, 'learning_rate/action_model': 9.773131534752472e-05, 'epoch': 8.66}) +[2026-06-20 20:41:51,445][starVLA.training.train_starvla][INFO] - Step 4685, Loss: {'train/loss': 0.07283641397953033, 'train/grad_norm_pre_clip': 7.533339023590088, 'timing/data': 0.012114173965528607, 'timing/model': 1.1473563460167497, 'learning_rate/qwen_vl_interface': 9.772200667012142e-06, 'learning_rate/action_model': 9.772200667012142e-05, 'epoch': 8.66}) +[2026-06-20 20:41:56,285][starVLA.training.train_starvla][INFO] - Step 4686, Loss: {'train/loss': 0.03448653593659401, 'train/grad_norm_pre_clip': 11.058063507080078, 'timing/data': 0.017959266086108983, 'timing/model': 1.2151741489069536, 'learning_rate/qwen_vl_interface': 9.771267940365934e-06, 'learning_rate/action_model': 9.771267940365935e-05, 'epoch': 8.67}) +[2026-06-20 20:42:01,133][starVLA.training.train_starvla][INFO] - Step 4687, Loss: {'train/loss': 0.024861954152584076, 'train/grad_norm_pre_clip': 6.482889175415039, 'timing/data': 0.014122367952950299, 'timing/model': 1.2159963899757713, 'learning_rate/qwen_vl_interface': 9.770333355197262e-06, 'learning_rate/action_model': 9.77033335519726e-05, 'epoch': 8.67}) +[2026-06-20 20:42:05,839][starVLA.training.train_starvla][INFO] - Step 4688, Loss: {'train/loss': 0.10820851475000381, 'train/grad_norm_pre_clip': 12.5034818649292, 'timing/data': 0.011687892023473978, 'timing/model': 1.1472542600240558, 'learning_rate/qwen_vl_interface': 9.769396911890291e-06, 'learning_rate/action_model': 9.769396911890291e-05, 'epoch': 8.67}) +[2026-06-20 20:42:10,717][starVLA.training.train_starvla][INFO] - Step 4689, Loss: {'train/loss': 0.0686640664935112, 'train/grad_norm_pre_clip': 7.567819595336914, 'timing/data': 0.01384638505987823, 'timing/model': 1.4482558900490403, 'learning_rate/qwen_vl_interface': 9.768458610829965e-06, 'learning_rate/action_model': 9.768458610829965e-05, 'epoch': 8.67}) +[2026-06-20 20:42:15,316][starVLA.training.train_starvla][INFO] - Step 4690, Loss: {'train/loss': 0.23166392743587494, 'train/grad_norm_pre_clip': 16.17386245727539, 'timing/data': 0.015803648973815143, 'timing/model': 1.2011000839993358, 'learning_rate/qwen_vl_interface': 9.767518452401981e-06, 'learning_rate/action_model': 9.76751845240198e-05, 'epoch': 8.67}) +[2026-06-20 20:42:20,133][starVLA.training.train_starvla][INFO] - Step 4691, Loss: {'train/loss': 0.039496950805187225, 'train/grad_norm_pre_clip': 3.6555163860321045, 'timing/data': 0.011922292062081397, 'timing/model': 1.2147119790315628, 'learning_rate/qwen_vl_interface': 9.766576436992802e-06, 'learning_rate/action_model': 9.766576436992801e-05, 'epoch': 8.68}) +[2026-06-20 20:42:24,783][starVLA.training.train_starvla][INFO] - Step 4692, Loss: {'train/loss': 0.07183040678501129, 'train/grad_norm_pre_clip': 11.486672401428223, 'timing/data': 0.011945970938540995, 'timing/model': 1.1458860830171034, 'learning_rate/qwen_vl_interface': 9.765632564989657e-06, 'learning_rate/action_model': 9.765632564989656e-05, 'epoch': 8.68}) +[2026-06-20 20:42:29,388][starVLA.training.train_starvla][INFO] - Step 4693, Loss: {'train/loss': 0.22340868413448334, 'train/grad_norm_pre_clip': 9.856498718261719, 'timing/data': 0.01215531094931066, 'timing/model': 1.1545176829677075, 'learning_rate/qwen_vl_interface': 9.764686836780533e-06, 'learning_rate/action_model': 9.764686836780532e-05, 'epoch': 8.68}) +[2026-06-20 20:42:34,031][starVLA.training.train_starvla][INFO] - Step 4694, Loss: {'train/loss': 0.0957353413105011, 'train/grad_norm_pre_clip': 12.381257057189941, 'timing/data': 0.011963550001382828, 'timing/model': 1.1447946929838508, 'learning_rate/qwen_vl_interface': 9.763739252754187e-06, 'learning_rate/action_model': 9.763739252754186e-05, 'epoch': 8.68}) +[2026-06-20 20:42:38,719][starVLA.training.train_starvla][INFO] - Step 4695, Loss: {'train/loss': 0.2551678419113159, 'train/grad_norm_pre_clip': 11.201407432556152, 'timing/data': 0.01317403104621917, 'timing/model': 1.4003941339906305, 'learning_rate/qwen_vl_interface': 9.76278981330013e-06, 'learning_rate/action_model': 9.762789813300131e-05, 'epoch': 8.68}) +[2026-06-20 20:42:43,215][starVLA.training.train_starvla][INFO] - Step 4696, Loss: {'train/loss': 0.008283701725304127, 'train/grad_norm_pre_clip': 1.856908917427063, 'timing/data': 0.015018983976915479, 'timing/model': 1.1754829259589314, 'learning_rate/qwen_vl_interface': 9.761838518808647e-06, 'learning_rate/action_model': 9.761838518808646e-05, 'epoch': 8.68}) +[2026-06-20 20:42:48,085][starVLA.training.train_starvla][INFO] - Step 4697, Loss: {'train/loss': 0.09758896380662918, 'train/grad_norm_pre_clip': 7.7203145027160645, 'timing/data': 0.013782221009023488, 'timing/model': 1.2046928459312767, 'learning_rate/qwen_vl_interface': 9.760885369670776e-06, 'learning_rate/action_model': 9.760885369670776e-05, 'epoch': 8.69}) +[2026-06-20 20:42:53,064][starVLA.training.train_starvla][INFO] - Step 4698, Loss: {'train/loss': 0.04316159710288048, 'train/grad_norm_pre_clip': 7.289357662200928, 'timing/data': 0.013647800078615546, 'timing/model': 1.1905316269258037, 'learning_rate/qwen_vl_interface': 9.759930366278319e-06, 'learning_rate/action_model': 9.759930366278318e-05, 'epoch': 8.69}) +[2026-06-20 20:42:57,873][starVLA.training.train_starvla][INFO] - Step 4699, Loss: {'train/loss': 0.1701512336730957, 'train/grad_norm_pre_clip': 15.096993446350098, 'timing/data': 0.013783750007860363, 'timing/model': 1.1892476490465924, 'learning_rate/qwen_vl_interface': 9.758973509023842e-06, 'learning_rate/action_model': 9.758973509023843e-05, 'epoch': 8.69}) +[2026-06-20 20:43:02,721][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 4700 +[2026-06-20 20:49:29,169][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 4700 +[2026-06-20 20:49:29,170][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 4700 +[2026-06-20 20:56:24,473][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 4700 +[2026-06-20 20:56:24,474][starVLA.training.train_starvla][INFO] - Step 4700, Loss: {'train/loss': 0.06596605479717255, 'train/grad_norm_pre_clip': 4.1804351806640625, 'eval/loss': 0.7553841471672058, 'eval/latency_0/loss': 0.9312015175819397, 'eval/latency_1/loss': 1.1491633653640747, 'eval/latency_2/loss': 1.0716277360916138, 'eval/latency_3/loss': 0.5074697732925415, 'eval/latency_4/loss': 0.08486610651016235, 'eval/flappy/loss': 0.7553841471672058, 'eval/flappy/latency_0/loss': 0.9312015175819397, 'eval/flappy/latency_1/loss': 1.1491633653640747, 'eval/flappy/latency_2/loss': 1.0716277360916138, 'eval/flappy/latency_3/loss': 0.5074697732925415, 'eval/flappy/latency_4/loss': 0.08486610651016235, 'eval/flap_precision': 0.12286689419795221, 'eval/flap_recall': 0.13090909090909092, 'eval/flap_f1': 0.1267605633802817, 'eval/flap_rate_model': 0.08138888888888889, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.1267605633802817, 'eval/latency_0/flap_precision': 0.12286689419795221, 'eval/latency_0/flap_recall': 0.13090909090909092, 'eval/latency_0/flap_f1': 0.1267605633802817, 'eval/latency_0/flap_rate_model': 0.08138888888888889, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.1267605633802817, 'timing/data': 0.013235940015874803, 'timing/model': 1.2092564930208027, 'learning_rate/qwen_vl_interface': 9.758014798300677e-06, 'learning_rate/action_model': 9.758014798300676e-05, 'epoch': 8.69}) +[2026-06-20 20:56:24,475][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_4700_state +[2026-06-20 20:56:24,482][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-20 20:56:32,011][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_4700_state/model.safetensors +[2026-06-20 20:56:46,060][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_4700_state/optimizer.bin +[2026-06-20 20:56:46,061][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_4700_state/sampler.bin +[2026-06-20 20:56:46,061][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_4700_state/sampler_1.bin +[2026-06-20 20:56:46,062][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_4700_state/random_states_0.pkl +[2026-06-20 20:56:48,709][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4700 +[2026-06-20 20:56:48,709][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-20 20:56:48,722][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-20 20:56:53,487][starVLA.training.train_starvla][INFO] - Step 4701, Loss: {'train/loss': 0.31944841146469116, 'train/grad_norm_pre_clip': 22.123764038085938, 'timing/data': 0.013599035912193358, 'timing/model': 1.1785125460010022, 'learning_rate/qwen_vl_interface': 9.757054234502909e-06, 'learning_rate/action_model': 9.757054234502909e-05, 'epoch': 8.69}) +[2026-06-20 20:56:58,315][starVLA.training.train_starvla][INFO] - Step 4702, Loss: {'train/loss': 0.03364593908190727, 'train/grad_norm_pre_clip': 9.434032440185547, 'timing/data': 0.01564989297185093, 'timing/model': 1.4235864550573751, 'learning_rate/qwen_vl_interface': 9.756091818025394e-06, 'learning_rate/action_model': 9.756091818025394e-05, 'epoch': 8.7}) +[2026-06-20 20:57:03,129][starVLA.training.train_starvla][INFO] - Step 4703, Loss: {'train/loss': 0.049438443034887314, 'train/grad_norm_pre_clip': 17.506263732910156, 'timing/data': 0.019669165019877255, 'timing/model': 1.3797248729970306, 'learning_rate/qwen_vl_interface': 9.75512754926374e-06, 'learning_rate/action_model': 9.755127549263741e-05, 'epoch': 8.7}) +[2026-06-20 20:57:07,741][starVLA.training.train_starvla][INFO] - Step 4704, Loss: {'train/loss': 0.080977663397789, 'train/grad_norm_pre_clip': 9.569624900817871, 'timing/data': 0.014621284091845155, 'timing/model': 1.1997401340631768, 'learning_rate/qwen_vl_interface': 9.754161428614325e-06, 'learning_rate/action_model': 9.754161428614324e-05, 'epoch': 8.7}) +[2026-06-20 20:57:12,600][starVLA.training.train_starvla][INFO] - Step 4705, Loss: {'train/loss': 0.13442499935626984, 'train/grad_norm_pre_clip': 10.630335807800293, 'timing/data': 0.012638787971809506, 'timing/model': 1.2095839969115332, 'learning_rate/qwen_vl_interface': 9.753193456474284e-06, 'learning_rate/action_model': 9.753193456474284e-05, 'epoch': 8.7}) +[2026-06-20 20:57:17,373][starVLA.training.train_starvla][INFO] - Step 4706, Loss: {'train/loss': 0.2469007521867752, 'train/grad_norm_pre_clip': 7.411187648773193, 'timing/data': 0.011345589999109507, 'timing/model': 1.13472161302343, 'learning_rate/qwen_vl_interface': 9.752223633241514e-06, 'learning_rate/action_model': 9.752223633241514e-05, 'epoch': 8.7}) +[2026-06-20 20:57:22,268][starVLA.training.train_starvla][INFO] - Step 4707, Loss: {'train/loss': 0.22993190586566925, 'train/grad_norm_pre_clip': 20.31631088256836, 'timing/data': 0.016203321050852537, 'timing/model': 1.2038392599206418, 'learning_rate/qwen_vl_interface': 9.751251959314671e-06, 'learning_rate/action_model': 9.751251959314672e-05, 'epoch': 8.7}) +[2026-06-20 20:57:27,124][starVLA.training.train_starvla][INFO] - Step 4708, Loss: {'train/loss': 0.03160163015127182, 'train/grad_norm_pre_clip': 3.1907663345336914, 'timing/data': 0.012497856980189681, 'timing/model': 1.4129434539936483, 'learning_rate/qwen_vl_interface': 9.750278435093176e-06, 'learning_rate/action_model': 9.750278435093177e-05, 'epoch': 8.71}) +[2026-06-20 20:57:32,079][starVLA.training.train_starvla][INFO] - Step 4709, Loss: {'train/loss': 0.11199330538511276, 'train/grad_norm_pre_clip': 7.428267478942871, 'timing/data': 0.014692222932353616, 'timing/model': 1.4300755430012941, 'learning_rate/qwen_vl_interface': 9.749303060977205e-06, 'learning_rate/action_model': 9.749303060977205e-05, 'epoch': 8.71}) +[2026-06-20 20:57:36,737][starVLA.training.train_starvla][INFO] - Step 4710, Loss: {'train/loss': 0.15386462211608887, 'train/grad_norm_pre_clip': 12.001133918762207, 'timing/data': 0.013164675910957158, 'timing/model': 1.1892876119818538, 'learning_rate/qwen_vl_interface': 9.748325837367701e-06, 'learning_rate/action_model': 9.748325837367702e-05, 'epoch': 8.71}) +[2026-06-20 20:57:41,581][starVLA.training.train_starvla][INFO] - Step 4711, Loss: {'train/loss': 0.08903378248214722, 'train/grad_norm_pre_clip': 5.887892246246338, 'timing/data': 0.01359938399400562, 'timing/model': 1.18998383206781, 'learning_rate/qwen_vl_interface': 9.747346764666362e-06, 'learning_rate/action_model': 9.747346764666361e-05, 'epoch': 8.71}) +[2026-06-20 20:57:46,892][starVLA.training.train_starvla][INFO] - Step 4712, Loss: {'train/loss': 0.13298362493515015, 'train/grad_norm_pre_clip': 4.814705848693848, 'timing/data': 0.4846629899693653, 'timing/model': 1.21926700102631, 'learning_rate/qwen_vl_interface': 9.746365843275648e-06, 'learning_rate/action_model': 9.746365843275646e-05, 'epoch': 8.71}) +[2026-06-20 20:57:51,716][starVLA.training.train_starvla][INFO] - Step 4713, Loss: {'train/loss': 0.1193104237318039, 'train/grad_norm_pre_clip': 7.653040409088135, 'timing/data': 0.01936576410662383, 'timing/model': 1.2075387339573354, 'learning_rate/qwen_vl_interface': 9.745383073598776e-06, 'learning_rate/action_model': 9.745383073598775e-05, 'epoch': 8.72}) +[2026-06-20 20:57:56,503][starVLA.training.train_starvla][INFO] - Step 4714, Loss: {'train/loss': 0.015870792791247368, 'train/grad_norm_pre_clip': 1.2927420139312744, 'timing/data': 0.013083496014587581, 'timing/model': 1.390559887047857, 'learning_rate/qwen_vl_interface': 9.744398456039729e-06, 'learning_rate/action_model': 9.744398456039729e-05, 'epoch': 8.72}) +[2026-06-20 20:58:00,903][starVLA.training.train_starvla][INFO] - Step 4715, Loss: {'train/loss': 0.19141682982444763, 'train/grad_norm_pre_clip': 13.489100456237793, 'timing/data': 0.012707386049441993, 'timing/model': 1.1461395929800346, 'learning_rate/qwen_vl_interface': 9.743411991003246e-06, 'learning_rate/action_model': 9.743411991003247e-05, 'epoch': 8.72}) +[2026-06-20 20:58:05,600][starVLA.training.train_starvla][INFO] - Step 4716, Loss: {'train/loss': 0.18100668489933014, 'train/grad_norm_pre_clip': 9.308660507202148, 'timing/data': 0.01257036707829684, 'timing/model': 1.1533653530059382, 'learning_rate/qwen_vl_interface': 9.742423678894823e-06, 'learning_rate/action_model': 9.742423678894823e-05, 'epoch': 8.72}) +[2026-06-20 20:58:10,267][starVLA.training.train_starvla][INFO] - Step 4717, Loss: {'train/loss': 0.031216157600283623, 'train/grad_norm_pre_clip': 4.782825469970703, 'timing/data': 0.011858747922815382, 'timing/model': 1.156557149020955, 'learning_rate/qwen_vl_interface': 9.741433520120722e-06, 'learning_rate/action_model': 9.74143352012072e-05, 'epoch': 8.72}) +[2026-06-20 20:58:14,995][starVLA.training.train_starvla][INFO] - Step 4718, Loss: {'train/loss': 0.1120789647102356, 'train/grad_norm_pre_clip': 5.889337539672852, 'timing/data': 0.01197619701270014, 'timing/model': 1.1920504650333896, 'learning_rate/qwen_vl_interface': 9.740441515087953e-06, 'learning_rate/action_model': 9.740441515087953e-05, 'epoch': 8.73}) +[2026-06-20 20:58:19,809][starVLA.training.train_starvla][INFO] - Step 4719, Loss: {'train/loss': 0.2262994647026062, 'train/grad_norm_pre_clip': 17.589231491088867, 'timing/data': 0.011956527945585549, 'timing/model': 1.2071901599410921, 'learning_rate/qwen_vl_interface': 9.739447664204298e-06, 'learning_rate/action_model': 9.739447664204297e-05, 'epoch': 8.73}) +[2026-06-20 20:58:24,530][starVLA.training.train_starvla][INFO] - Step 4720, Loss: {'train/loss': 0.13168513774871826, 'train/grad_norm_pre_clip': 8.374711036682129, 'timing/data': 0.012044608010910451, 'timing/model': 1.3443014939548448, 'learning_rate/qwen_vl_interface': 9.73845196787829e-06, 'learning_rate/action_model': 9.738451967878289e-05, 'epoch': 8.73}) +[2026-06-20 20:58:29,111][starVLA.training.train_starvla][INFO] - Step 4721, Loss: {'train/loss': 0.09833141416311264, 'train/grad_norm_pre_clip': 5.363005638122559, 'timing/data': 0.012006008066236973, 'timing/model': 1.207977193989791, 'learning_rate/qwen_vl_interface': 9.737454426519221e-06, 'learning_rate/action_model': 9.737454426519221e-05, 'epoch': 8.73}) +[2026-06-20 20:58:33,915][starVLA.training.train_starvla][INFO] - Step 4722, Loss: {'train/loss': 0.21917413175106049, 'train/grad_norm_pre_clip': 7.988512992858887, 'timing/data': 0.012087216950021684, 'timing/model': 1.1410204730927944, 'learning_rate/qwen_vl_interface': 9.736455040537145e-06, 'learning_rate/action_model': 9.736455040537144e-05, 'epoch': 8.73}) +[2026-06-20 20:58:38,794][starVLA.training.train_starvla][INFO] - Step 4723, Loss: {'train/loss': 0.11584428697824478, 'train/grad_norm_pre_clip': 8.970174789428711, 'timing/data': 0.0155532609205693, 'timing/model': 1.1942667650291696, 'learning_rate/qwen_vl_interface': 9.735453810342868e-06, 'learning_rate/action_model': 9.735453810342867e-05, 'epoch': 8.73}) +[2026-06-20 20:58:43,660][starVLA.training.train_starvla][INFO] - Step 4724, Loss: {'train/loss': 0.22992345690727234, 'train/grad_norm_pre_clip': 9.321419715881348, 'timing/data': 0.012101457105018198, 'timing/model': 1.2082715010037646, 'learning_rate/qwen_vl_interface': 9.734450736347958e-06, 'learning_rate/action_model': 9.73445073634796e-05, 'epoch': 8.74}) +[2026-06-20 20:58:48,426][starVLA.training.train_starvla][INFO] - Step 4725, Loss: {'train/loss': 0.15656618773937225, 'train/grad_norm_pre_clip': 12.548330307006836, 'timing/data': 0.012037236941978335, 'timing/model': 1.209101282991469, 'learning_rate/qwen_vl_interface': 9.733445818964746e-06, 'learning_rate/action_model': 9.733445818964746e-05, 'epoch': 8.74}) +[2026-06-20 20:58:53,134][starVLA.training.train_starvla][INFO] - Step 4726, Loss: {'train/loss': 0.2322218120098114, 'train/grad_norm_pre_clip': 8.784051895141602, 'timing/data': 0.011844216962344944, 'timing/model': 1.3990439110202715, 'learning_rate/qwen_vl_interface': 9.732439058606311e-06, 'learning_rate/action_model': 9.732439058606312e-05, 'epoch': 8.74}) +[2026-06-20 20:58:57,653][starVLA.training.train_starvla][INFO] - Step 4727, Loss: {'train/loss': 0.07634637504816055, 'train/grad_norm_pre_clip': 7.339631080627441, 'timing/data': 0.011790037038736045, 'timing/model': 1.14213670801837, 'learning_rate/qwen_vl_interface': 9.731430455686499e-06, 'learning_rate/action_model': 9.731430455686498e-05, 'epoch': 8.74}) +[2026-06-20 20:59:02,440][starVLA.training.train_starvla][INFO] - Step 4728, Loss: {'train/loss': 0.23894406855106354, 'train/grad_norm_pre_clip': 10.842040061950684, 'timing/data': 0.012150326045230031, 'timing/model': 1.1427264199592173, 'learning_rate/qwen_vl_interface': 9.730420010619901e-06, 'learning_rate/action_model': 9.7304200106199e-05, 'epoch': 8.74}) +[2026-06-20 20:59:07,197][starVLA.training.train_starvla][INFO] - Step 4729, Loss: {'train/loss': 0.2590082287788391, 'train/grad_norm_pre_clip': 11.672858238220215, 'timing/data': 0.01228926598560065, 'timing/model': 1.1424765750998631, 'learning_rate/qwen_vl_interface': 9.72940772382188e-06, 'learning_rate/action_model': 9.729407723821878e-05, 'epoch': 8.75}) +[2026-06-20 20:59:12,029][starVLA.training.train_starvla][INFO] - Step 4730, Loss: {'train/loss': 0.1745210736989975, 'train/grad_norm_pre_clip': 10.52350902557373, 'timing/data': 0.011998075991868973, 'timing/model': 1.1931713999947533, 'learning_rate/qwen_vl_interface': 9.728393595708542e-06, 'learning_rate/action_model': 9.728393595708542e-05, 'epoch': 8.75}) +[2026-06-20 20:59:16,798][starVLA.training.train_starvla][INFO] - Step 4731, Loss: {'train/loss': 0.11474142968654633, 'train/grad_norm_pre_clip': 4.064024925231934, 'timing/data': 0.011956776026636362, 'timing/model': 1.4520260089775547, 'learning_rate/qwen_vl_interface': 9.727377626696762e-06, 'learning_rate/action_model': 9.727377626696762e-05, 'epoch': 8.75}) +[2026-06-20 20:59:21,320][starVLA.training.train_starvla][INFO] - Step 4732, Loss: {'train/loss': 0.15562692284584045, 'train/grad_norm_pre_clip': 4.818297863006592, 'timing/data': 0.011849067057482898, 'timing/model': 1.1944698740262538, 'learning_rate/qwen_vl_interface': 9.726359817204165e-06, 'learning_rate/action_model': 9.726359817204166e-05, 'epoch': 8.75}) +[2026-06-20 20:59:26,032][starVLA.training.train_starvla][INFO] - Step 4733, Loss: {'train/loss': 0.09152701497077942, 'train/grad_norm_pre_clip': 4.94537878036499, 'timing/data': 0.01185800600796938, 'timing/model': 1.2116823539836332, 'learning_rate/qwen_vl_interface': 9.725340167649136e-06, 'learning_rate/action_model': 9.725340167649136e-05, 'epoch': 8.75}) +[2026-06-20 20:59:30,860][starVLA.training.train_starvla][INFO] - Step 4734, Loss: {'train/loss': 0.16581840813159943, 'train/grad_norm_pre_clip': 6.858012676239014, 'timing/data': 0.014223012025468051, 'timing/model': 1.2086744740372524, 'learning_rate/qwen_vl_interface': 9.72431867845081e-06, 'learning_rate/action_model': 9.72431867845081e-05, 'epoch': 8.75}) +[2026-06-20 20:59:35,484][starVLA.training.train_starvla][INFO] - Step 4735, Loss: {'train/loss': 0.06107285991311073, 'train/grad_norm_pre_clip': 3.169236421585083, 'timing/data': 0.012875903979875147, 'timing/model': 1.1536823159549385, 'learning_rate/qwen_vl_interface': 9.723295350029086e-06, 'learning_rate/action_model': 9.723295350029086e-05, 'epoch': 8.76}) +[2026-06-20 20:59:40,128][starVLA.training.train_starvla][INFO] - Step 4736, Loss: {'train/loss': 0.028345469385385513, 'train/grad_norm_pre_clip': 1.8919708728790283, 'timing/data': 0.011821227031759918, 'timing/model': 1.143456480000168, 'learning_rate/qwen_vl_interface': 9.722270182804616e-06, 'learning_rate/action_model': 9.722270182804614e-05, 'epoch': 8.76}) +[2026-06-20 20:59:44,800][starVLA.training.train_starvla][INFO] - Step 4737, Loss: {'train/loss': 0.05529886484146118, 'train/grad_norm_pre_clip': 3.283076524734497, 'timing/data': 0.01183557603508234, 'timing/model': 1.3907945540267974, 'learning_rate/qwen_vl_interface': 9.721243177198805e-06, 'learning_rate/action_model': 9.721243177198804e-05, 'epoch': 8.76}) +[2026-06-20 20:59:49,252][starVLA.training.train_starvla][INFO] - Step 4738, Loss: {'train/loss': 0.10069002956151962, 'train/grad_norm_pre_clip': 16.048341751098633, 'timing/data': 0.01380686298944056, 'timing/model': 1.161120412987657, 'learning_rate/qwen_vl_interface': 9.720214333633817e-06, 'learning_rate/action_model': 9.720214333633817e-05, 'epoch': 8.76}) +[2026-06-20 20:59:53,952][starVLA.training.train_starvla][INFO] - Step 4739, Loss: {'train/loss': 0.15582875907421112, 'train/grad_norm_pre_clip': 4.853575229644775, 'timing/data': 0.016835546935908496, 'timing/model': 1.16426047089044, 'learning_rate/qwen_vl_interface': 9.719183652532572e-06, 'learning_rate/action_model': 9.719183652532572e-05, 'epoch': 8.76}) +[2026-06-20 20:59:58,756][starVLA.training.train_starvla][INFO] - Step 4740, Loss: {'train/loss': 0.21472635865211487, 'train/grad_norm_pre_clip': 9.555545806884766, 'timing/data': 0.012980893952772021, 'timing/model': 1.188952696043998, 'learning_rate/qwen_vl_interface': 9.718151134318743e-06, 'learning_rate/action_model': 9.718151134318743e-05, 'epoch': 8.77}) +[2026-06-20 21:00:03,442][starVLA.training.train_starvla][INFO] - Step 4741, Loss: {'train/loss': 0.11726892739534378, 'train/grad_norm_pre_clip': 10.030316352844238, 'timing/data': 0.014209171989932656, 'timing/model': 1.1519557730061933, 'learning_rate/qwen_vl_interface': 9.71711677941676e-06, 'learning_rate/action_model': 9.717116779416758e-05, 'epoch': 8.77}) +[2026-06-20 21:00:08,182][starVLA.training.train_starvla][INFO] - Step 4742, Loss: {'train/loss': 0.09389850497245789, 'train/grad_norm_pre_clip': 5.998143672943115, 'timing/data': 0.014152202056720853, 'timing/model': 1.1955405119806528, 'learning_rate/qwen_vl_interface': 9.716080588251805e-06, 'learning_rate/action_model': 9.716080588251805e-05, 'epoch': 8.77}) +[2026-06-20 21:00:13,064][starVLA.training.train_starvla][INFO] - Step 4743, Loss: {'train/loss': 0.09411389380693436, 'train/grad_norm_pre_clip': 7.798515796661377, 'timing/data': 0.015196799067780375, 'timing/model': 1.439066014951095, 'learning_rate/qwen_vl_interface': 9.71504256124982e-06, 'learning_rate/action_model': 9.71504256124982e-05, 'epoch': 8.77}) +[2026-06-20 21:00:17,700][starVLA.training.train_starvla][INFO] - Step 4744, Loss: {'train/loss': 0.5045012831687927, 'train/grad_norm_pre_clip': 16.406538009643555, 'timing/data': 0.01617053698282689, 'timing/model': 1.2108873979886994, 'learning_rate/qwen_vl_interface': 9.714002698837499e-06, 'learning_rate/action_model': 9.714002698837498e-05, 'epoch': 8.77}) +[2026-06-20 21:00:22,454][starVLA.training.train_starvla][INFO] - Step 4745, Loss: {'train/loss': 0.15617622435092926, 'train/grad_norm_pre_clip': 10.899507522583008, 'timing/data': 0.013383973040618002, 'timing/model': 1.1964949009707198, 'learning_rate/qwen_vl_interface': 9.712961001442285e-06, 'learning_rate/action_model': 9.712961001442286e-05, 'epoch': 8.78}) +[2026-06-20 21:00:27,353][starVLA.training.train_starvla][INFO] - Step 4746, Loss: {'train/loss': 0.5390089154243469, 'train/grad_norm_pre_clip': 22.895536422729492, 'timing/data': 0.01628849795088172, 'timing/model': 1.210774715989828, 'learning_rate/qwen_vl_interface': 9.711917469492388e-06, 'learning_rate/action_model': 9.711917469492388e-05, 'epoch': 8.78}) +[2026-06-20 21:00:32,267][starVLA.training.train_starvla][INFO] - Step 4747, Loss: {'train/loss': 0.15729670226573944, 'train/grad_norm_pre_clip': 10.68597412109375, 'timing/data': 0.014380801003426313, 'timing/model': 1.2141379240201786, 'learning_rate/qwen_vl_interface': 9.71087210341676e-06, 'learning_rate/action_model': 9.71087210341676e-05, 'epoch': 8.78}) +[2026-06-20 21:00:36,868][starVLA.training.train_starvla][INFO] - Step 4748, Loss: {'train/loss': 0.06354019790887833, 'train/grad_norm_pre_clip': 8.263477325439453, 'timing/data': 0.012212934903800488, 'timing/model': 1.143037413014099, 'learning_rate/qwen_vl_interface': 9.70982490364511e-06, 'learning_rate/action_model': 9.70982490364511e-05, 'epoch': 8.78}) +[2026-06-20 21:00:41,514][starVLA.training.train_starvla][INFO] - Step 4749, Loss: {'train/loss': 0.03704235330224037, 'train/grad_norm_pre_clip': 3.5140380859375, 'timing/data': 0.012107835034839809, 'timing/model': 1.3838836170034483, 'learning_rate/qwen_vl_interface': 9.708775870607905e-06, 'learning_rate/action_model': 9.708775870607904e-05, 'epoch': 8.78}) +[2026-06-20 21:00:45,921][starVLA.training.train_starvla][INFO] - Step 4750, Loss: {'train/loss': 0.028599221259355545, 'train/grad_norm_pre_clip': 2.339012861251831, 'timing/data': 0.011927556013688445, 'timing/model': 1.1464165460783988, 'learning_rate/qwen_vl_interface': 9.707725004736363e-06, 'learning_rate/action_model': 9.707725004736363e-05, 'epoch': 8.78}) +[2026-06-20 21:00:50,572][starVLA.training.train_starvla][INFO] - Step 4751, Loss: {'train/loss': 0.06877885013818741, 'train/grad_norm_pre_clip': 7.212845325469971, 'timing/data': 0.011906305910088122, 'timing/model': 1.1459660020191222, 'learning_rate/qwen_vl_interface': 9.706672306462454e-06, 'learning_rate/action_model': 9.706672306462454e-05, 'epoch': 8.79}) +[2026-06-20 21:00:55,228][starVLA.training.train_starvla][INFO] - Step 4752, Loss: {'train/loss': 0.20820465683937073, 'train/grad_norm_pre_clip': 11.741083145141602, 'timing/data': 0.011926445993594825, 'timing/model': 1.1457990769995376, 'learning_rate/qwen_vl_interface': 9.705617776218905e-06, 'learning_rate/action_model': 9.705617776218904e-05, 'epoch': 8.79}) +[2026-06-20 21:00:59,845][starVLA.training.train_starvla][INFO] - Step 4753, Loss: {'train/loss': 0.22807620465755463, 'train/grad_norm_pre_clip': 12.299105644226074, 'timing/data': 0.018983601010404527, 'timing/model': 1.1634174849605188, 'learning_rate/qwen_vl_interface': 9.70456141443919e-06, 'learning_rate/action_model': 9.704561414439188e-05, 'epoch': 8.79}) +[2026-06-20 21:01:04,536][starVLA.training.train_starvla][INFO] - Step 4754, Loss: {'train/loss': 0.1621778905391693, 'train/grad_norm_pre_clip': 8.964632034301758, 'timing/data': 0.01342068298254162, 'timing/model': 1.1608574860729277, 'learning_rate/qwen_vl_interface': 9.703503221557538e-06, 'learning_rate/action_model': 9.703503221557538e-05, 'epoch': 8.79}) +[2026-06-20 21:01:09,245][starVLA.training.train_starvla][INFO] - Step 4755, Loss: {'train/loss': 0.15412911772727966, 'train/grad_norm_pre_clip': 13.285289764404297, 'timing/data': 0.01949937001336366, 'timing/model': 1.4069159020436928, 'learning_rate/qwen_vl_interface': 9.70244319800894e-06, 'learning_rate/action_model': 9.70244319800894e-05, 'epoch': 8.79}) +[2026-06-20 21:01:13,661][starVLA.training.train_starvla][INFO] - Step 4756, Loss: {'train/loss': 0.1024356335401535, 'train/grad_norm_pre_clip': 6.162930011749268, 'timing/data': 0.012586124008521438, 'timing/model': 1.1436593920225278, 'learning_rate/qwen_vl_interface': 9.701381344229124e-06, 'learning_rate/action_model': 9.701381344229123e-05, 'epoch': 8.8}) +[2026-06-20 21:01:18,364][starVLA.training.train_starvla][INFO] - Step 4757, Loss: {'train/loss': 0.07080496102571487, 'train/grad_norm_pre_clip': 7.989371299743652, 'timing/data': 0.011954655055887997, 'timing/model': 1.1448987440671772, 'learning_rate/qwen_vl_interface': 9.700317660654581e-06, 'learning_rate/action_model': 9.700317660654581e-05, 'epoch': 8.8}) +[2026-06-20 21:01:23,151][starVLA.training.train_starvla][INFO] - Step 4758, Loss: {'train/loss': 0.12791338562965393, 'train/grad_norm_pre_clip': 8.268903732299805, 'timing/data': 0.013522911933250725, 'timing/model': 1.2117417700355873, 'learning_rate/qwen_vl_interface': 9.699252147722554e-06, 'learning_rate/action_model': 9.699252147722553e-05, 'epoch': 8.8}) +[2026-06-20 21:01:28,001][starVLA.training.train_starvla][INFO] - Step 4759, Loss: {'train/loss': 0.09289294481277466, 'train/grad_norm_pre_clip': 7.777802467346191, 'timing/data': 0.011947244987823069, 'timing/model': 1.1430973570095375, 'learning_rate/qwen_vl_interface': 9.698184805871028e-06, 'learning_rate/action_model': 9.698184805871028e-05, 'epoch': 8.8}) +[2026-06-20 21:01:32,880][starVLA.training.train_starvla][INFO] - Step 4760, Loss: {'train/loss': 0.24624097347259521, 'train/grad_norm_pre_clip': 8.044636726379395, 'timing/data': 0.013718321919441223, 'timing/model': 1.1967589210253209, 'learning_rate/qwen_vl_interface': 9.697115635538753e-06, 'learning_rate/action_model': 9.697115635538753e-05, 'epoch': 8.8}) +[2026-06-20 21:01:37,721][starVLA.training.train_starvla][INFO] - Step 4761, Loss: {'train/loss': 0.09069642424583435, 'train/grad_norm_pre_clip': 6.647658348083496, 'timing/data': 0.012031605001538992, 'timing/model': 1.3955199999036267, 'learning_rate/qwen_vl_interface': 9.69604463716522e-06, 'learning_rate/action_model': 9.69604463716522e-05, 'epoch': 8.8}) +[2026-06-20 21:01:42,329][starVLA.training.train_starvla][INFO] - Step 4762, Loss: {'train/loss': 0.02740069292485714, 'train/grad_norm_pre_clip': 2.258639335632324, 'timing/data': 0.014896299107931554, 'timing/model': 1.199148635030724, 'learning_rate/qwen_vl_interface': 9.69497181119068e-06, 'learning_rate/action_model': 9.694971811190681e-05, 'epoch': 8.81}) +[2026-06-20 21:01:47,213][starVLA.training.train_starvla][INFO] - Step 4763, Loss: {'train/loss': 0.08539401739835739, 'train/grad_norm_pre_clip': 5.471604824066162, 'timing/data': 0.014174690004438162, 'timing/model': 1.212969962041825, 'learning_rate/qwen_vl_interface': 9.693897158056132e-06, 'learning_rate/action_model': 9.693897158056132e-05, 'epoch': 8.81}) +[2026-06-20 21:01:51,988][starVLA.training.train_starvla][INFO] - Step 4764, Loss: {'train/loss': 0.047922927886247635, 'train/grad_norm_pre_clip': 3.050935983657837, 'timing/data': 0.01194997492711991, 'timing/model': 1.145360748982057, 'learning_rate/qwen_vl_interface': 9.692820678203322e-06, 'learning_rate/action_model': 9.692820678203322e-05, 'epoch': 8.81}) +[2026-06-20 21:01:56,831][starVLA.training.train_starvla][INFO] - Step 4765, Loss: {'train/loss': 0.08548315614461899, 'train/grad_norm_pre_clip': 3.540753126144409, 'timing/data': 0.016947404015809298, 'timing/model': 1.2158865749370307, 'learning_rate/qwen_vl_interface': 9.691742372074749e-06, 'learning_rate/action_model': 9.691742372074749e-05, 'epoch': 8.81}) +[2026-06-20 21:02:01,682][starVLA.training.train_starvla][INFO] - Step 4766, Loss: {'train/loss': 0.10472957789897919, 'train/grad_norm_pre_clip': 12.096044540405273, 'timing/data': 0.013728942023590207, 'timing/model': 1.1986844870261848, 'learning_rate/qwen_vl_interface': 9.690662240113669e-06, 'learning_rate/action_model': 9.690662240113668e-05, 'epoch': 8.81}) +[2026-06-20 21:02:06,575][starVLA.training.train_starvla][INFO] - Step 4767, Loss: {'train/loss': 0.01816113106906414, 'train/grad_norm_pre_clip': 1.786742925643921, 'timing/data': 0.014637438929639757, 'timing/model': 1.4342406740179285, 'learning_rate/qwen_vl_interface': 9.689580282764077e-06, 'learning_rate/action_model': 9.689580282764077e-05, 'epoch': 8.82}) +[2026-06-20 21:02:11,179][starVLA.training.train_starvla][INFO] - Step 4768, Loss: {'train/loss': 0.2299017608165741, 'train/grad_norm_pre_clip': 10.486200332641602, 'timing/data': 0.015053478069603443, 'timing/model': 1.2050137739861384, 'learning_rate/qwen_vl_interface': 9.688496500470731e-06, 'learning_rate/action_model': 9.68849650047073e-05, 'epoch': 8.82}) +[2026-06-20 21:02:15,825][starVLA.training.train_starvla][INFO] - Step 4769, Loss: {'train/loss': 0.22117610275745392, 'train/grad_norm_pre_clip': 14.467951774597168, 'timing/data': 0.011861764010973275, 'timing/model': 1.143793600029312, 'learning_rate/qwen_vl_interface': 9.68741089367913e-06, 'learning_rate/action_model': 9.687410893679129e-05, 'epoch': 8.82}) +[2026-06-20 21:02:20,451][starVLA.training.train_starvla][INFO] - Step 4770, Loss: {'train/loss': 0.011536858044564724, 'train/grad_norm_pre_clip': 1.0402430295944214, 'timing/data': 0.011850946000777185, 'timing/model': 1.1461837600218132, 'learning_rate/qwen_vl_interface': 9.686323462835527e-06, 'learning_rate/action_model': 9.686323462835526e-05, 'epoch': 8.82}) +[2026-06-20 21:02:25,147][starVLA.training.train_starvla][INFO] - Step 4771, Loss: {'train/loss': 0.17481805384159088, 'train/grad_norm_pre_clip': 6.547276020050049, 'timing/data': 0.01447465899400413, 'timing/model': 1.16259572096169, 'learning_rate/qwen_vl_interface': 9.685234208386922e-06, 'learning_rate/action_model': 9.685234208386921e-05, 'epoch': 8.82}) +[2026-06-20 21:02:29,805][starVLA.training.train_starvla][INFO] - Step 4772, Loss: {'train/loss': 0.02721928246319294, 'train/grad_norm_pre_clip': 2.0139691829681396, 'timing/data': 0.012814101995900273, 'timing/model': 1.1445812559686601, 'learning_rate/qwen_vl_interface': 9.684143130781066e-06, 'learning_rate/action_model': 9.684143130781066e-05, 'epoch': 8.83}) +[2026-06-20 21:02:34,505][starVLA.training.train_starvla][INFO] - Step 4773, Loss: {'train/loss': 0.09790662676095963, 'train/grad_norm_pre_clip': 7.203593730926514, 'timing/data': 0.0117710349150002, 'timing/model': 1.3841492879437283, 'learning_rate/qwen_vl_interface': 9.683050230466462e-06, 'learning_rate/action_model': 9.683050230466462e-05, 'epoch': 8.83}) +[2026-06-20 21:02:39,105][starVLA.training.train_starvla][INFO] - Step 4774, Loss: {'train/loss': 0.1577461212873459, 'train/grad_norm_pre_clip': 25.153839111328125, 'timing/data': 0.012070523109287024, 'timing/model': 1.3346868100343272, 'learning_rate/qwen_vl_interface': 9.681955507892361e-06, 'learning_rate/action_model': 9.68195550789236e-05, 'epoch': 8.83}) +[2026-06-20 21:02:43,555][starVLA.training.train_starvla][INFO] - Step 4775, Loss: {'train/loss': 0.5696563720703125, 'train/grad_norm_pre_clip': 22.74602508544922, 'timing/data': 0.01222008396871388, 'timing/model': 1.1455432609654963, 'learning_rate/qwen_vl_interface': 9.680858963508759e-06, 'learning_rate/action_model': 9.68085896350876e-05, 'epoch': 8.83}) +[2026-06-20 21:02:48,208][starVLA.training.train_starvla][INFO] - Step 4776, Loss: {'train/loss': 0.024550624191761017, 'train/grad_norm_pre_clip': 3.4526476860046387, 'timing/data': 0.012421032995916903, 'timing/model': 1.1456705369055271, 'learning_rate/qwen_vl_interface': 9.679760597766405e-06, 'learning_rate/action_model': 9.679760597766405e-05, 'epoch': 8.83}) +[2026-06-20 21:02:52,909][starVLA.training.train_starvla][INFO] - Step 4777, Loss: {'train/loss': 0.1303018480539322, 'train/grad_norm_pre_clip': 11.642335891723633, 'timing/data': 0.01186156494077295, 'timing/model': 1.146912039956078, 'learning_rate/qwen_vl_interface': 9.678660411116796e-06, 'learning_rate/action_model': 9.678660411116796e-05, 'epoch': 8.83}) +[2026-06-20 21:02:57,579][starVLA.training.train_starvla][INFO] - Step 4778, Loss: {'train/loss': 0.22247572243213654, 'train/grad_norm_pre_clip': 12.85212516784668, 'timing/data': 0.01199588505551219, 'timing/model': 1.1454553080257028, 'learning_rate/qwen_vl_interface': 9.677558404012178e-06, 'learning_rate/action_model': 9.677558404012177e-05, 'epoch': 8.84}) +[2026-06-20 21:03:02,270][starVLA.training.train_starvla][INFO] - Step 4779, Loss: {'train/loss': 0.3067547380924225, 'train/grad_norm_pre_clip': 32.71630096435547, 'timing/data': 0.011999754002317786, 'timing/model': 1.3888665890553966, 'learning_rate/qwen_vl_interface': 9.676454576905542e-06, 'learning_rate/action_model': 9.676454576905542e-05, 'epoch': 8.84}) +[2026-06-20 21:03:06,937][starVLA.training.train_starvla][INFO] - Step 4780, Loss: {'train/loss': 0.04611188545823097, 'train/grad_norm_pre_clip': 4.8602190017700195, 'timing/data': 0.011866645072586834, 'timing/model': 1.3831322659971192, 'learning_rate/qwen_vl_interface': 9.675348930250633e-06, 'learning_rate/action_model': 9.675348930250632e-05, 'epoch': 8.84}) +[2026-06-20 21:03:11,388][starVLA.training.train_starvla][INFO] - Step 4781, Loss: {'train/loss': 0.025367356836795807, 'train/grad_norm_pre_clip': 4.050662517547607, 'timing/data': 0.011979614035226405, 'timing/model': 1.1452690379228443, 'learning_rate/qwen_vl_interface': 9.674241464501938e-06, 'learning_rate/action_model': 9.674241464501937e-05, 'epoch': 8.84}) +[2026-06-20 21:03:16,059][starVLA.training.train_starvla][INFO] - Step 4782, Loss: {'train/loss': 0.058385219424963, 'train/grad_norm_pre_clip': 7.948500633239746, 'timing/data': 0.012069532996974885, 'timing/model': 1.1446673239115626, 'learning_rate/qwen_vl_interface': 9.673132180114695e-06, 'learning_rate/action_model': 9.673132180114695e-05, 'epoch': 8.84}) +[2026-06-20 21:03:20,797][starVLA.training.train_starvla][INFO] - Step 4783, Loss: {'train/loss': 0.09304514527320862, 'train/grad_norm_pre_clip': 7.478287220001221, 'timing/data': 0.012884610914625227, 'timing/model': 1.161892813979648, 'learning_rate/qwen_vl_interface': 9.67202107754489e-06, 'learning_rate/action_model': 9.672021077544889e-05, 'epoch': 8.85}) +[2026-06-20 21:03:25,467][starVLA.training.train_starvla][INFO] - Step 4784, Loss: {'train/loss': 0.14280778169631958, 'train/grad_norm_pre_clip': 10.470538139343262, 'timing/data': 0.012009173980914056, 'timing/model': 1.1452698960201815, 'learning_rate/qwen_vl_interface': 9.670908157249254e-06, 'learning_rate/action_model': 9.670908157249253e-05, 'epoch': 8.85}) +[2026-06-20 21:03:30,163][starVLA.training.train_starvla][INFO] - Step 4785, Loss: {'train/loss': 0.27373552322387695, 'train/grad_norm_pre_clip': 19.13711166381836, 'timing/data': 0.013808620045892894, 'timing/model': 1.410790322930552, 'learning_rate/qwen_vl_interface': 9.669793419685267e-06, 'learning_rate/action_model': 9.669793419685266e-05, 'epoch': 8.85}) +[2026-06-20 21:03:34,615][starVLA.training.train_starvla][INFO] - Step 4786, Loss: {'train/loss': 0.2092866152524948, 'train/grad_norm_pre_clip': 17.164154052734375, 'timing/data': 0.011377425049431622, 'timing/model': 1.1498579879989848, 'learning_rate/qwen_vl_interface': 9.668676865311155e-06, 'learning_rate/action_model': 9.668676865311154e-05, 'epoch': 8.85}) +[2026-06-20 21:03:39,278][starVLA.training.train_starvla][INFO] - Step 4787, Loss: {'train/loss': 0.028112148866057396, 'train/grad_norm_pre_clip': 4.603456020355225, 'timing/data': 0.012161624035798013, 'timing/model': 1.152424039086327, 'learning_rate/qwen_vl_interface': 9.667558494585891e-06, 'learning_rate/action_model': 9.66755849458589e-05, 'epoch': 8.85}) +[2026-06-20 21:03:43,972][starVLA.training.train_starvla][INFO] - Step 4788, Loss: {'train/loss': 0.2663753926753998, 'train/grad_norm_pre_clip': 11.686529159545898, 'timing/data': 0.012304213014431298, 'timing/model': 1.155971827916801, 'learning_rate/qwen_vl_interface': 9.666438307969194e-06, 'learning_rate/action_model': 9.666438307969195e-05, 'epoch': 8.85}) +[2026-06-20 21:03:48,699][starVLA.training.train_starvla][INFO] - Step 4789, Loss: {'train/loss': 0.14912579953670502, 'train/grad_norm_pre_clip': 4.9465813636779785, 'timing/data': 0.012236202950589359, 'timing/model': 1.160876383073628, 'learning_rate/qwen_vl_interface': 9.665316305921535e-06, 'learning_rate/action_model': 9.665316305921535e-05, 'epoch': 8.86}) +[2026-06-20 21:03:53,345][starVLA.training.train_starvla][INFO] - Step 4790, Loss: {'train/loss': 0.13176469504833221, 'train/grad_norm_pre_clip': 13.147634506225586, 'timing/data': 0.012169123045168817, 'timing/model': 1.1551902419887483, 'learning_rate/qwen_vl_interface': 9.664192488904123e-06, 'learning_rate/action_model': 9.664192488904122e-05, 'epoch': 8.86}) +[2026-06-20 21:03:57,981][starVLA.training.train_starvla][INFO] - Step 4791, Loss: {'train/loss': 0.11837224662303925, 'train/grad_norm_pre_clip': 8.89298152923584, 'timing/data': 0.011953654000535607, 'timing/model': 1.3632619509007782, 'learning_rate/qwen_vl_interface': 9.663066857378917e-06, 'learning_rate/action_model': 9.663066857378917e-05, 'epoch': 8.86}) +[2026-06-20 21:04:02,421][starVLA.training.train_starvla][INFO] - Step 4792, Loss: {'train/loss': 0.19636990129947662, 'train/grad_norm_pre_clip': 10.841236114501953, 'timing/data': 0.012039472931064665, 'timing/model': 1.1478637220570818, 'learning_rate/qwen_vl_interface': 9.661939411808623e-06, 'learning_rate/action_model': 9.661939411808622e-05, 'epoch': 8.86}) +[2026-06-20 21:04:07,087][starVLA.training.train_starvla][INFO] - Step 4793, Loss: {'train/loss': 0.20839443802833557, 'train/grad_norm_pre_clip': 13.246251106262207, 'timing/data': 0.011887024040333927, 'timing/model': 1.1447588739683852, 'learning_rate/qwen_vl_interface': 9.66081015265669e-06, 'learning_rate/action_model': 9.660810152656689e-05, 'epoch': 8.86}) +[2026-06-20 21:04:11,772][starVLA.training.train_starvla][INFO] - Step 4794, Loss: {'train/loss': 0.30956926941871643, 'train/grad_norm_pre_clip': 15.741671562194824, 'timing/data': 0.011162574985064566, 'timing/model': 1.1456866490188986, 'learning_rate/qwen_vl_interface': 9.659679080387314e-06, 'learning_rate/action_model': 9.659679080387313e-05, 'epoch': 8.87}) +[2026-06-20 21:04:16,441][starVLA.training.train_starvla][INFO] - Step 4795, Loss: {'train/loss': 0.1945720762014389, 'train/grad_norm_pre_clip': 12.071974754333496, 'timing/data': 0.011803593952208757, 'timing/model': 1.1465432540280744, 'learning_rate/qwen_vl_interface': 9.658546195465436e-06, 'learning_rate/action_model': 9.658546195465435e-05, 'epoch': 8.87}) +[2026-06-20 21:04:21,097][starVLA.training.train_starvla][INFO] - Step 4796, Loss: {'train/loss': 0.09267237037420273, 'train/grad_norm_pre_clip': 5.850778102874756, 'timing/data': 0.012019822956062853, 'timing/model': 1.144159734947607, 'learning_rate/qwen_vl_interface': 9.65741149835674e-06, 'learning_rate/action_model': 9.657411498356741e-05, 'epoch': 8.87}) +[2026-06-20 21:04:25,804][starVLA.training.train_starvla][INFO] - Step 4797, Loss: {'train/loss': 0.15664511919021606, 'train/grad_norm_pre_clip': 9.443787574768066, 'timing/data': 0.013898398960009217, 'timing/model': 1.376730925985612, 'learning_rate/qwen_vl_interface': 9.656274989527663e-06, 'learning_rate/action_model': 9.656274989527664e-05, 'epoch': 8.87}) +[2026-06-20 21:04:30,253][starVLA.training.train_starvla][INFO] - Step 4798, Loss: {'train/loss': 0.2706720530986786, 'train/grad_norm_pre_clip': 10.96954345703125, 'timing/data': 0.011843804037198424, 'timing/model': 1.1520957810571417, 'learning_rate/qwen_vl_interface': 9.655136669445378e-06, 'learning_rate/action_model': 9.655136669445378e-05, 'epoch': 8.87}) +[2026-06-20 21:04:34,913][starVLA.training.train_starvla][INFO] - Step 4799, Loss: {'train/loss': 0.22927600145339966, 'train/grad_norm_pre_clip': 7.248827934265137, 'timing/data': 0.01214045297820121, 'timing/model': 1.146151271997951, 'learning_rate/qwen_vl_interface': 9.653996538577802e-06, 'learning_rate/action_model': 9.653996538577802e-05, 'epoch': 8.88}) +[2026-06-20 21:04:39,611][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 4800 +[2026-06-20 21:11:05,880][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 4800 +[2026-06-20 21:11:05,881][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 4800 +[2026-06-20 21:17:59,040][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 4800 +[2026-06-20 21:17:59,042][starVLA.training.train_starvla][INFO] - Step 4800, Loss: {'train/loss': 0.08608651161193848, 'train/grad_norm_pre_clip': 3.3465523719787598, 'eval/loss': 0.45437005162239075, 'eval/latency_0/loss': 0.5165876150131226, 'eval/latency_1/loss': 0.6292045712471008, 'eval/latency_2/loss': 0.6917465329170227, 'eval/latency_3/loss': 0.3320351839065552, 'eval/latency_4/loss': 0.08024834841489792, 'eval/flappy/loss': 0.45437005162239075, 'eval/flappy/latency_0/loss': 0.5165876150131226, 'eval/flappy/latency_1/loss': 0.6292045712471008, 'eval/flappy/latency_2/loss': 0.6917465329170227, 'eval/flappy/latency_3/loss': 0.3320351839065552, 'eval/flappy/latency_4/loss': 0.08024834841489792, 'eval/flap_precision': 0.13405797101449277, 'eval/flap_recall': 0.13454545454545455, 'eval/flap_f1': 0.13430127041742287, 'eval/flap_rate_model': 0.07666666666666666, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.13430127041742287, 'eval/latency_0/flap_precision': 0.13405797101449277, 'eval/latency_0/flap_recall': 0.13454545454545455, 'eval/latency_0/flap_f1': 0.13430127041742287, 'eval/latency_0/flap_rate_model': 0.07666666666666666, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.13430127041742287, 'timing/data': 0.011969823040999472, 'timing/model': 1.1461099970620126, 'learning_rate/qwen_vl_interface': 9.652854597393604e-06, 'learning_rate/action_model': 9.652854597393603e-05, 'epoch': 8.88}) +[2026-06-20 21:17:59,042][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_4800_state +[2026-06-20 21:17:59,046][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-20 21:18:07,586][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_4800_state/model.safetensors +[2026-06-20 21:18:21,639][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_4800_state/optimizer.bin +[2026-06-20 21:18:21,640][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_4800_state/sampler.bin +[2026-06-20 21:18:21,640][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_4800_state/sampler_1.bin +[2026-06-20 21:18:21,641][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_4800_state/random_states_0.pkl +[2026-06-20 21:18:24,283][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4800 +[2026-06-20 21:18:24,284][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-20 21:18:24,316][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-20 21:18:28,885][starVLA.training.train_starvla][INFO] - Step 4801, Loss: {'train/loss': 0.16421960294246674, 'train/grad_norm_pre_clip': 6.356255531311035, 'timing/data': 0.011840887018479407, 'timing/model': 1.127111512934789, 'learning_rate/qwen_vl_interface': 9.651710846362192e-06, 'learning_rate/action_model': 9.651710846362192e-05, 'epoch': 8.88}) +[2026-06-20 21:18:33,557][starVLA.training.train_starvla][INFO] - Step 4802, Loss: {'train/loss': 0.12999112904071808, 'train/grad_norm_pre_clip': 5.752913475036621, 'timing/data': 0.011973515036515892, 'timing/model': 1.130692987004295, 'learning_rate/qwen_vl_interface': 9.650565285953716e-06, 'learning_rate/action_model': 9.650565285953715e-05, 'epoch': 8.88}) +[2026-06-20 21:18:38,233][starVLA.training.train_starvla][INFO] - Step 4803, Loss: {'train/loss': 0.05223117023706436, 'train/grad_norm_pre_clip': 2.6591620445251465, 'timing/data': 0.011865156004205346, 'timing/model': 1.3715204509207979, 'learning_rate/qwen_vl_interface': 9.649417916639077e-06, 'learning_rate/action_model': 9.649417916639076e-05, 'epoch': 8.88}) +[2026-06-20 21:18:42,792][starVLA.training.train_starvla][INFO] - Step 4804, Loss: {'train/loss': 0.04768265783786774, 'train/grad_norm_pre_clip': 2.1005868911743164, 'timing/data': 0.014799147029407322, 'timing/model': 1.1859724660171196, 'learning_rate/qwen_vl_interface': 9.648268738889911e-06, 'learning_rate/action_model': 9.648268738889911e-05, 'epoch': 8.88}) +[2026-06-20 21:18:47,422][starVLA.training.train_starvla][INFO] - Step 4805, Loss: {'train/loss': 0.2498028725385666, 'train/grad_norm_pre_clip': 9.219935417175293, 'timing/data': 0.011986095923930407, 'timing/model': 1.131558867986314, 'learning_rate/qwen_vl_interface': 9.647117753178604e-06, 'learning_rate/action_model': 9.647117753178604e-05, 'epoch': 8.89}) +[2026-06-20 21:18:52,056][starVLA.training.train_starvla][INFO] - Step 4806, Loss: {'train/loss': 0.040840137749910355, 'train/grad_norm_pre_clip': 2.145023822784424, 'timing/data': 0.012035537045449018, 'timing/model': 1.1338560540461913, 'learning_rate/qwen_vl_interface': 9.645964959978284e-06, 'learning_rate/action_model': 9.645964959978284e-05, 'epoch': 8.89}) +[2026-06-20 21:18:56,751][starVLA.training.train_starvla][INFO] - Step 4807, Loss: {'train/loss': 0.08052510768175125, 'train/grad_norm_pre_clip': 5.037839889526367, 'timing/data': 0.014963436988182366, 'timing/model': 1.1626520600402728, 'learning_rate/qwen_vl_interface': 9.644810359762818e-06, 'learning_rate/action_model': 9.644810359762817e-05, 'epoch': 8.89}) +[2026-06-20 21:19:01,346][starVLA.training.train_starvla][INFO] - Step 4808, Loss: {'train/loss': 0.12226443737745285, 'train/grad_norm_pre_clip': 7.490569591522217, 'timing/data': 0.012074557016603649, 'timing/model': 1.1377801820635796, 'learning_rate/qwen_vl_interface': 9.64365395300682e-06, 'learning_rate/action_model': 9.643653953006819e-05, 'epoch': 8.89}) +[2026-06-20 21:19:05,987][starVLA.training.train_starvla][INFO] - Step 4809, Loss: {'train/loss': 0.03355377912521362, 'train/grad_norm_pre_clip': 2.0206093788146973, 'timing/data': 0.012175757088698447, 'timing/model': 1.3683497769525275, 'learning_rate/qwen_vl_interface': 9.642495740185644e-06, 'learning_rate/action_model': 9.642495740185643e-05, 'epoch': 8.89}) +[2026-06-20 21:19:10,408][starVLA.training.train_starvla][INFO] - Step 4810, Loss: {'train/loss': 0.08187014609575272, 'train/grad_norm_pre_clip': 7.270593166351318, 'timing/data': 0.011982380063273013, 'timing/model': 1.1517679549288005, 'learning_rate/qwen_vl_interface': 9.641335721775388e-06, 'learning_rate/action_model': 9.641335721775387e-05, 'epoch': 8.9}) +[2026-06-20 21:19:15,060][starVLA.training.train_starvla][INFO] - Step 4811, Loss: {'train/loss': 0.13705766201019287, 'train/grad_norm_pre_clip': 10.110124588012695, 'timing/data': 0.013599063036963344, 'timing/model': 1.1517137930495664, 'learning_rate/qwen_vl_interface': 9.640173898252893e-06, 'learning_rate/action_model': 9.640173898252892e-05, 'epoch': 8.9}) +[2026-06-20 21:19:19,763][starVLA.training.train_starvla][INFO] - Step 4812, Loss: {'train/loss': 0.19618840515613556, 'train/grad_norm_pre_clip': 11.785310745239258, 'timing/data': 0.011999520007520914, 'timing/model': 1.1389325089985505, 'learning_rate/qwen_vl_interface': 9.639010270095738e-06, 'learning_rate/action_model': 9.639010270095739e-05, 'epoch': 8.9}) +[2026-06-20 21:19:24,435][starVLA.training.train_starvla][INFO] - Step 4813, Loss: {'train/loss': 0.024776509031653404, 'train/grad_norm_pre_clip': 3.948085069656372, 'timing/data': 0.011839453014545143, 'timing/model': 1.1380881340010092, 'learning_rate/qwen_vl_interface': 9.637844837782251e-06, 'learning_rate/action_model': 9.63784483778225e-05, 'epoch': 8.9}) +[2026-06-20 21:19:29,072][starVLA.training.train_starvla][INFO] - Step 4814, Loss: {'train/loss': 0.029397141188383102, 'train/grad_norm_pre_clip': 2.7801778316497803, 'timing/data': 0.012662143912166357, 'timing/model': 1.3852152520557866, 'learning_rate/qwen_vl_interface': 9.636677601791494e-06, 'learning_rate/action_model': 9.636677601791493e-05, 'epoch': 8.9}) +[2026-06-20 21:19:33,655][starVLA.training.train_starvla][INFO] - Step 4815, Loss: {'train/loss': 0.16405951976776123, 'train/grad_norm_pre_clip': 13.060733795166016, 'timing/data': 0.013374938047491014, 'timing/model': 1.3070802389411256, 'learning_rate/qwen_vl_interface': 9.635508562603273e-06, 'learning_rate/action_model': 9.635508562603272e-05, 'epoch': 8.9}) +[2026-06-20 21:19:38,072][starVLA.training.train_starvla][INFO] - Step 4816, Loss: {'train/loss': 0.019675372168421745, 'train/grad_norm_pre_clip': 1.84042227268219, 'timing/data': 0.011735344072803855, 'timing/model': 1.1408147029578686, 'learning_rate/qwen_vl_interface': 9.634337720698139e-06, 'learning_rate/action_model': 9.634337720698137e-05, 'epoch': 8.91}) +[2026-06-20 21:19:42,755][starVLA.training.train_starvla][INFO] - Step 4817, Loss: {'train/loss': 0.17827025055885315, 'train/grad_norm_pre_clip': 20.754709243774414, 'timing/data': 0.018341080052778125, 'timing/model': 1.1558840100187808, 'learning_rate/qwen_vl_interface': 9.633165076557376e-06, 'learning_rate/action_model': 9.633165076557375e-05, 'epoch': 8.91}) +[2026-06-20 21:19:47,395][starVLA.training.train_starvla][INFO] - Step 4818, Loss: {'train/loss': 0.023315107449889183, 'train/grad_norm_pre_clip': 5.6547651290893555, 'timing/data': 0.011929453001357615, 'timing/model': 1.1410648999735713, 'learning_rate/qwen_vl_interface': 9.631990630663019e-06, 'learning_rate/action_model': 9.631990630663018e-05, 'epoch': 8.91}) +[2026-06-20 21:19:52,058][starVLA.training.train_starvla][INFO] - Step 4819, Loss: {'train/loss': 0.008856629021465778, 'train/grad_norm_pre_clip': 1.808509349822998, 'timing/data': 0.013385210069827735, 'timing/model': 1.1540876280050725, 'learning_rate/qwen_vl_interface': 9.630814383497836e-06, 'learning_rate/action_model': 9.630814383497835e-05, 'epoch': 8.91}) +[2026-06-20 21:19:56,740][starVLA.training.train_starvla][INFO] - Step 4820, Loss: {'train/loss': 0.16171380877494812, 'train/grad_norm_pre_clip': 15.382583618164062, 'timing/data': 0.012652616947889328, 'timing/model': 1.3893320269417018, 'learning_rate/qwen_vl_interface': 9.629636335545336e-06, 'learning_rate/action_model': 9.629636335545335e-05, 'epoch': 8.91}) +[2026-06-20 21:20:01,305][starVLA.training.train_starvla][INFO] - Step 4821, Loss: {'train/loss': 0.32369163632392883, 'train/grad_norm_pre_clip': 14.01529598236084, 'timing/data': 0.011630277032963932, 'timing/model': 1.2948779999278486, 'learning_rate/qwen_vl_interface': 9.628456487289773e-06, 'learning_rate/action_model': 9.628456487289773e-05, 'epoch': 8.92}) +[2026-06-20 21:20:05,740][starVLA.training.train_starvla][INFO] - Step 4822, Loss: {'train/loss': 0.24994076788425446, 'train/grad_norm_pre_clip': 11.983474731445312, 'timing/data': 0.013123763957992196, 'timing/model': 1.1537296100286767, 'learning_rate/qwen_vl_interface': 9.627274839216135e-06, 'learning_rate/action_model': 9.627274839216136e-05, 'epoch': 8.92}) +[2026-06-20 21:20:10,367][starVLA.training.train_starvla][INFO] - Step 4823, Loss: {'train/loss': 0.07598811388015747, 'train/grad_norm_pre_clip': 9.24508285522461, 'timing/data': 0.013644990045577288, 'timing/model': 1.1555703160120174, 'learning_rate/qwen_vl_interface': 9.626091391810158e-06, 'learning_rate/action_model': 9.626091391810157e-05, 'epoch': 8.92}) +[2026-06-20 21:20:14,991][starVLA.training.train_starvla][INFO] - Step 4824, Loss: {'train/loss': 0.18662802875041962, 'train/grad_norm_pre_clip': 8.846577644348145, 'timing/data': 0.012079813983291388, 'timing/model': 1.1439579389989376, 'learning_rate/qwen_vl_interface': 9.624906145558309e-06, 'learning_rate/action_model': 9.624906145558307e-05, 'epoch': 8.92}) +[2026-06-20 21:20:19,656][starVLA.training.train_starvla][INFO] - Step 4825, Loss: {'train/loss': 0.1594470739364624, 'train/grad_norm_pre_clip': 6.654037952423096, 'timing/data': 0.011824037996120751, 'timing/model': 1.1414009659783915, 'learning_rate/qwen_vl_interface': 9.623719100947795e-06, 'learning_rate/action_model': 9.623719100947795e-05, 'epoch': 8.92}) +[2026-06-20 21:20:24,333][starVLA.training.train_starvla][INFO] - Step 4826, Loss: {'train/loss': 0.0042788260616362095, 'train/grad_norm_pre_clip': 1.5782066583633423, 'timing/data': 0.012653509969823062, 'timing/model': 1.389061728026718, 'learning_rate/qwen_vl_interface': 9.622530258466571e-06, 'learning_rate/action_model': 9.62253025846657e-05, 'epoch': 8.93}) +[2026-06-20 21:20:28,776][starVLA.training.train_starvla][INFO] - Step 4827, Loss: {'train/loss': 0.10286092013120651, 'train/grad_norm_pre_clip': 18.676485061645508, 'timing/data': 0.013611392001621425, 'timing/model': 1.1561272889375687, 'learning_rate/qwen_vl_interface': 9.62133961860332e-06, 'learning_rate/action_model': 9.62133961860332e-05, 'epoch': 8.93}) +[2026-06-20 21:20:33,453][starVLA.training.train_starvla][INFO] - Step 4828, Loss: {'train/loss': 0.1790207028388977, 'train/grad_norm_pre_clip': 14.739694595336914, 'timing/data': 0.017140169977210462, 'timing/model': 1.1614151609828696, 'learning_rate/qwen_vl_interface': 9.620147181847474e-06, 'learning_rate/action_model': 9.620147181847475e-05, 'epoch': 8.93}) +[2026-06-20 21:20:38,089][starVLA.training.train_starvla][INFO] - Step 4829, Loss: {'train/loss': 0.18392319977283478, 'train/grad_norm_pre_clip': 19.0474910736084, 'timing/data': 0.012065987102687359, 'timing/model': 1.1421700719511136, 'learning_rate/qwen_vl_interface': 9.618952948689196e-06, 'learning_rate/action_model': 9.618952948689196e-05, 'epoch': 8.93}) +[2026-06-20 21:20:42,724][starVLA.training.train_starvla][INFO] - Step 4830, Loss: {'train/loss': 0.1394316554069519, 'train/grad_norm_pre_clip': 10.487235069274902, 'timing/data': 0.015094739967025816, 'timing/model': 1.155910104047507, 'learning_rate/qwen_vl_interface': 9.617756919619393e-06, 'learning_rate/action_model': 9.617756919619392e-05, 'epoch': 8.93}) +[2026-06-20 21:20:47,577][starVLA.training.train_starvla][INFO] - Step 4831, Loss: {'train/loss': 0.009820069186389446, 'train/grad_norm_pre_clip': 3.755669116973877, 'timing/data': 0.019012692966498435, 'timing/model': 1.3131363180000335, 'learning_rate/qwen_vl_interface': 9.616559095129701e-06, 'learning_rate/action_model': 9.6165590951297e-05, 'epoch': 8.93}) +[2026-06-20 21:20:52,940][starVLA.training.train_starvla][INFO] - Step 4832, Loss: {'train/loss': 0.23922327160835266, 'train/grad_norm_pre_clip': 11.54450798034668, 'timing/data': 0.02004349499475211, 'timing/model': 1.6283334089675918, 'learning_rate/qwen_vl_interface': 9.615359475712506e-06, 'learning_rate/action_model': 9.615359475712506e-05, 'epoch': 8.94}) +[2026-06-20 21:20:57,992][starVLA.training.train_starvla][INFO] - Step 4833, Loss: {'train/loss': 0.04930117353796959, 'train/grad_norm_pre_clip': 8.375186920166016, 'timing/data': 0.01892028597649187, 'timing/model': 1.3108631490031257, 'learning_rate/qwen_vl_interface': 9.614158061860924e-06, 'learning_rate/action_model': 9.614158061860924e-05, 'epoch': 8.94}) +[2026-06-20 21:21:03,351][starVLA.training.train_starvla][INFO] - Step 4834, Loss: {'train/loss': 0.014349306002259254, 'train/grad_norm_pre_clip': 2.6664562225341797, 'timing/data': 0.018769338028505445, 'timing/model': 1.311507805949077, 'learning_rate/qwen_vl_interface': 9.612954854068813e-06, 'learning_rate/action_model': 9.612954854068812e-05, 'epoch': 8.94}) +[2026-06-20 21:21:08,704][starVLA.training.train_starvla][INFO] - Step 4835, Loss: {'train/loss': 0.16271980106830597, 'train/grad_norm_pre_clip': 9.73891544342041, 'timing/data': 0.018350392929278314, 'timing/model': 1.3140538720181212, 'learning_rate/qwen_vl_interface': 9.611749852830763e-06, 'learning_rate/action_model': 9.611749852830763e-05, 'epoch': 8.94}) +[2026-06-20 21:21:14,074][starVLA.training.train_starvla][INFO] - Step 4836, Loss: {'train/loss': 0.019045842811465263, 'train/grad_norm_pre_clip': 5.157435417175293, 'timing/data': 0.018914327956736088, 'timing/model': 1.310427705058828, 'learning_rate/qwen_vl_interface': 9.610543058642107e-06, 'learning_rate/action_model': 9.610543058642107e-05, 'epoch': 8.94}) +[2026-06-20 21:21:19,440][starVLA.training.train_starvla][INFO] - Step 4837, Loss: {'train/loss': 0.018084008246660233, 'train/grad_norm_pre_clip': 6.496412754058838, 'timing/data': 0.01884277001954615, 'timing/model': 1.3119911419926211, 'learning_rate/qwen_vl_interface': 9.609334471998911e-06, 'learning_rate/action_model': 9.609334471998912e-05, 'epoch': 8.95}) +[2026-06-20 21:21:25,189][starVLA.training.train_starvla][INFO] - Step 4838, Loss: {'train/loss': 0.01694369502365589, 'train/grad_norm_pre_clip': 1.8817826509475708, 'timing/data': 0.0696954820305109, 'timing/model': 1.6457314980216324, 'learning_rate/qwen_vl_interface': 9.60812409339798e-06, 'learning_rate/action_model': 9.60812409339798e-05, 'epoch': 8.95}) +[2026-06-20 21:21:30,527][starVLA.training.train_starvla][INFO] - Step 4839, Loss: {'train/loss': 0.12386634200811386, 'train/grad_norm_pre_clip': 23.28973388671875, 'timing/data': 0.018561913049779832, 'timing/model': 1.5993324659066275, 'learning_rate/qwen_vl_interface': 9.606911923336856e-06, 'learning_rate/action_model': 9.606911923336856e-05, 'epoch': 8.95}) +[2026-06-20 21:21:35,573][starVLA.training.train_starvla][INFO] - Step 4840, Loss: {'train/loss': 0.05111177638173103, 'train/grad_norm_pre_clip': 4.981270790100098, 'timing/data': 0.01893357199151069, 'timing/model': 1.3095775080146268, 'learning_rate/qwen_vl_interface': 9.605697962313817e-06, 'learning_rate/action_model': 9.605697962313816e-05, 'epoch': 8.95}) +[2026-06-20 21:21:40,925][starVLA.training.train_starvla][INFO] - Step 4841, Loss: {'train/loss': 0.15764111280441284, 'train/grad_norm_pre_clip': 11.401561737060547, 'timing/data': 0.018719583051279187, 'timing/model': 1.3105197479017079, 'learning_rate/qwen_vl_interface': 9.604482210827871e-06, 'learning_rate/action_model': 9.604482210827871e-05, 'epoch': 8.95}) +[2026-06-20 21:21:46,292][starVLA.training.train_starvla][INFO] - Step 4842, Loss: {'train/loss': 0.08183251321315765, 'train/grad_norm_pre_clip': 11.474786758422852, 'timing/data': 0.018819123040884733, 'timing/model': 1.3085745519492775, 'learning_rate/qwen_vl_interface': 9.603264669378772e-06, 'learning_rate/action_model': 9.603264669378772e-05, 'epoch': 8.95}) +[2026-06-20 21:21:51,660][starVLA.training.train_starvla][INFO] - Step 4843, Loss: {'train/loss': 0.0683804526925087, 'train/grad_norm_pre_clip': 9.943599700927734, 'timing/data': 0.019117682008072734, 'timing/model': 1.3097197979222983, 'learning_rate/qwen_vl_interface': 9.602045338467007e-06, 'learning_rate/action_model': 9.602045338467006e-05, 'epoch': 8.96}) +[2026-06-20 21:21:57,027][starVLA.training.train_starvla][INFO] - Step 4844, Loss: {'train/loss': 0.07616155594587326, 'train/grad_norm_pre_clip': 9.29743766784668, 'timing/data': 0.020068823941983283, 'timing/model': 1.6263420910108835, 'learning_rate/qwen_vl_interface': 9.600824218593791e-06, 'learning_rate/action_model': 9.600824218593791e-05, 'epoch': 8.96}) +[2026-06-20 21:22:02,328][starVLA.training.train_starvla][INFO] - Step 4845, Loss: {'train/loss': 0.08824314922094345, 'train/grad_norm_pre_clip': 13.464515686035156, 'timing/data': 0.019176362082362175, 'timing/model': 1.5626222949940711, 'learning_rate/qwen_vl_interface': 9.599601310261087e-06, 'learning_rate/action_model': 9.599601310261085e-05, 'epoch': 8.96}) +[2026-06-20 21:22:06,845][starVLA.training.train_starvla][INFO] - Step 4846, Loss: {'train/loss': 0.12321598082780838, 'train/grad_norm_pre_clip': 6.562045097351074, 'timing/data': 0.01493172999471426, 'timing/model': 1.1962324370397255, 'learning_rate/qwen_vl_interface': 9.598376613971582e-06, 'learning_rate/action_model': 9.598376613971582e-05, 'epoch': 8.96}) +[2026-06-20 21:22:11,537][starVLA.training.train_starvla][INFO] - Step 4847, Loss: {'train/loss': 0.21783997118473053, 'train/grad_norm_pre_clip': 15.421219825744629, 'timing/data': 0.011942195938900113, 'timing/model': 1.143800419056788, 'learning_rate/qwen_vl_interface': 9.597150130228703e-06, 'learning_rate/action_model': 9.597150130228703e-05, 'epoch': 8.96}) +[2026-06-20 21:22:16,375][starVLA.training.train_starvla][INFO] - Step 4848, Loss: {'train/loss': 0.12253287434577942, 'train/grad_norm_pre_clip': 13.985600471496582, 'timing/data': 0.01225037407130003, 'timing/model': 1.1426241050940007, 'learning_rate/qwen_vl_interface': 9.595921859536612e-06, 'learning_rate/action_model': 9.595921859536611e-05, 'epoch': 8.97}) +[2026-06-20 21:22:21,249][starVLA.training.train_starvla][INFO] - Step 4849, Loss: {'train/loss': 0.2623482346534729, 'train/grad_norm_pre_clip': 18.385717391967773, 'timing/data': 0.014687752933241427, 'timing/model': 1.1951877860119566, 'learning_rate/qwen_vl_interface': 9.594691802400203e-06, 'learning_rate/action_model': 9.594691802400202e-05, 'epoch': 8.97}) +[2026-06-20 21:22:26,120][starVLA.training.train_starvla][INFO] - Step 4850, Loss: {'train/loss': 0.26016420125961304, 'train/grad_norm_pre_clip': 18.758394241333008, 'timing/data': 0.015342148020863533, 'timing/model': 1.433739678002894, 'learning_rate/qwen_vl_interface': 9.59345995932511e-06, 'learning_rate/action_model': 9.59345995932511e-05, 'epoch': 8.97}) +[2026-06-20 21:22:30,848][starVLA.training.train_starvla][INFO] - Step 4851, Loss: {'train/loss': 0.21289218962192535, 'train/grad_norm_pre_clip': 10.191171646118164, 'timing/data': 0.011995336972177029, 'timing/model': 1.380037251044996, 'learning_rate/qwen_vl_interface': 9.592226330817693e-06, 'learning_rate/action_model': 9.592226330817693e-05, 'epoch': 8.97}) +[2026-06-20 21:22:35,400][starVLA.training.train_starvla][INFO] - Step 4852, Loss: {'train/loss': 0.06586231291294098, 'train/grad_norm_pre_clip': 6.906148433685303, 'timing/data': 0.012983708991669118, 'timing/model': 1.2176032569259405, 'learning_rate/qwen_vl_interface': 9.590990917385052e-06, 'learning_rate/action_model': 9.590990917385052e-05, 'epoch': 8.97}) +[2026-06-20 21:22:40,280][starVLA.training.train_starvla][INFO] - Step 4853, Loss: {'train/loss': 0.1563374251127243, 'train/grad_norm_pre_clip': 15.29586410522461, 'timing/data': 0.016342650982551277, 'timing/model': 1.2102616759948432, 'learning_rate/qwen_vl_interface': 9.58975371953502e-06, 'learning_rate/action_model': 9.589753719535018e-05, 'epoch': 8.97}) +[2026-06-20 21:22:45,131][starVLA.training.train_starvla][INFO] - Step 4854, Loss: {'train/loss': 0.06460165977478027, 'train/grad_norm_pre_clip': 7.691572189331055, 'timing/data': 0.016364691080525517, 'timing/model': 1.2126946600619704, 'learning_rate/qwen_vl_interface': 9.588514737776158e-06, 'learning_rate/action_model': 9.588514737776157e-05, 'epoch': 8.98}) +[2026-06-20 21:22:50,093][starVLA.training.train_starvla][INFO] - Step 4855, Loss: {'train/loss': 0.1457282304763794, 'train/grad_norm_pre_clip': 10.753011703491211, 'timing/data': 0.013629995053634048, 'timing/model': 1.1938890549354255, 'learning_rate/qwen_vl_interface': 9.587273972617769e-06, 'learning_rate/action_model': 9.587273972617768e-05, 'epoch': 8.98}) +[2026-06-20 21:22:54,929][starVLA.training.train_starvla][INFO] - Step 4856, Loss: {'train/loss': 0.26833048462867737, 'train/grad_norm_pre_clip': 20.492019653320312, 'timing/data': 0.015314780990593135, 'timing/model': 1.4431663659634069, 'learning_rate/qwen_vl_interface': 9.586031424569882e-06, 'learning_rate/action_model': 9.58603142456988e-05, 'epoch': 8.98}) +[2026-06-20 21:22:59,511][starVLA.training.train_starvla][INFO] - Step 4857, Loss: {'train/loss': 0.01543367374688387, 'train/grad_norm_pre_clip': 2.1192450523376465, 'timing/data': 0.013699925038963556, 'timing/model': 1.196677788044326, 'learning_rate/qwen_vl_interface': 9.584787094143265e-06, 'learning_rate/action_model': 9.584787094143264e-05, 'epoch': 8.98}) +[2026-06-20 21:23:04,339][starVLA.training.train_starvla][INFO] - Step 4858, Loss: {'train/loss': 0.14531759917736053, 'train/grad_norm_pre_clip': 6.4428019523620605, 'timing/data': 0.012133557931520045, 'timing/model': 1.144387008040212, 'learning_rate/qwen_vl_interface': 9.58354098184941e-06, 'learning_rate/action_model': 9.583540981849409e-05, 'epoch': 8.98}) +[2026-06-20 21:23:08,985][starVLA.training.train_starvla][INFO] - Step 4859, Loss: {'train/loss': 0.039180926978588104, 'train/grad_norm_pre_clip': 7.255809783935547, 'timing/data': 0.011990108992904425, 'timing/model': 1.142786552081816, 'learning_rate/qwen_vl_interface': 9.582293088200549e-06, 'learning_rate/action_model': 9.582293088200549e-05, 'epoch': 8.99}) +[2026-06-20 21:23:13,670][starVLA.training.train_starvla][INFO] - Step 4860, Loss: {'train/loss': 0.04627028852701187, 'train/grad_norm_pre_clip': 4.793503761291504, 'timing/data': 0.013795965001918375, 'timing/model': 1.158514080918394, 'learning_rate/qwen_vl_interface': 9.581043413709646e-06, 'learning_rate/action_model': 9.581043413709646e-05, 'epoch': 8.99}) +[2026-06-20 21:23:18,255][starVLA.training.train_starvla][INFO] - Step 4861, Loss: {'train/loss': 0.10130302608013153, 'train/grad_norm_pre_clip': 10.242633819580078, 'timing/data': 0.01175116200465709, 'timing/model': 1.1440322420094162, 'learning_rate/qwen_vl_interface': 9.579791958890393e-06, 'learning_rate/action_model': 9.579791958890394e-05, 'epoch': 8.99}) +[2026-06-20 21:23:22,908][starVLA.training.train_starvla][INFO] - Step 4862, Loss: {'train/loss': 0.2677629888057709, 'train/grad_norm_pre_clip': 6.101298809051514, 'timing/data': 0.011967530939728022, 'timing/model': 1.366323191090487, 'learning_rate/qwen_vl_interface': 9.578538724257218e-06, 'learning_rate/action_model': 9.578538724257216e-05, 'epoch': 8.99}) +[2026-06-20 21:23:27,312][starVLA.training.train_starvla][INFO] - Step 4863, Loss: {'train/loss': 0.07487958669662476, 'train/grad_norm_pre_clip': 5.1786723136901855, 'timing/data': 0.011812573065981269, 'timing/model': 1.1448112919460982, 'learning_rate/qwen_vl_interface': 9.577283710325274e-06, 'learning_rate/action_model': 9.577283710325274e-05, 'epoch': 8.99}) +[2026-06-20 21:23:32,034][starVLA.training.train_starvla][INFO] - Step 4864, Loss: {'train/loss': 0.026916639879345894, 'train/grad_norm_pre_clip': 2.0660581588745117, 'timing/data': 0.011961540905758739, 'timing/model': 1.142518141074106, 'learning_rate/qwen_vl_interface': 9.576026917610455e-06, 'learning_rate/action_model': 9.576026917610455e-05, 'epoch': 9.0}) +[2026-06-20 21:23:36,864][starVLA.training.train_starvla][INFO] - Step 4865, Loss: {'train/loss': 0.13629475235939026, 'train/grad_norm_pre_clip': 8.897512435913086, 'timing/data': 0.013600289006717503, 'timing/model': 1.1964753470383584, 'learning_rate/qwen_vl_interface': 9.574768346629376e-06, 'learning_rate/action_model': 9.574768346629376e-05, 'epoch': 9.0}) +[2026-06-20 21:23:41,688][starVLA.training.train_starvla][INFO] - Step 4866, Loss: {'train/loss': 0.03442825376987457, 'train/grad_norm_pre_clip': 3.4695498943328857, 'timing/data': 0.01198601198848337, 'timing/model': 1.1448781189974397, 'learning_rate/qwen_vl_interface': 9.573507997899391e-06, 'learning_rate/action_model': 9.573507997899392e-05, 'epoch': 9.0}) +[2026-06-20 21:23:46,447][starVLA.training.train_starvla][INFO] - Step 4867, Loss: {'train/loss': 0.19112646579742432, 'train/grad_norm_pre_clip': 17.167634963989258, 'timing/data': 0.0117531749419868, 'timing/model': 1.1410514790331945, 'learning_rate/qwen_vl_interface': 9.572245871938582e-06, 'learning_rate/action_model': 9.572245871938582e-05, 'epoch': 9.0}) +[2026-06-20 21:23:51,263][starVLA.training.train_starvla][INFO] - Step 4868, Loss: {'train/loss': 0.15046750009059906, 'train/grad_norm_pre_clip': 17.686939239501953, 'timing/data': 0.014427072019316256, 'timing/model': 1.4235804240452126, 'learning_rate/qwen_vl_interface': 9.570981969265761e-06, 'learning_rate/action_model': 9.570981969265761e-05, 'epoch': 9.0}) +[2026-06-20 21:23:55,834][starVLA.training.train_starvla][INFO] - Step 4869, Loss: {'train/loss': 0.13657501339912415, 'train/grad_norm_pre_clip': 7.008749008178711, 'timing/data': 0.013745938893407583, 'timing/model': 1.1950230080401525, 'learning_rate/qwen_vl_interface': 9.56971629040047e-06, 'learning_rate/action_model': 9.56971629040047e-05, 'epoch': 9.0}) +[2026-06-20 21:24:00,664][starVLA.training.train_starvla][INFO] - Step 4870, Loss: {'train/loss': 0.07790637761354446, 'train/grad_norm_pre_clip': 8.113351821899414, 'timing/data': 0.014108235016465187, 'timing/model': 1.2148336879909039, 'learning_rate/qwen_vl_interface': 9.56844883586298e-06, 'learning_rate/action_model': 9.568448835862981e-05, 'epoch': 9.01}) +[2026-06-20 21:24:05,601][starVLA.training.train_starvla][INFO] - Step 4871, Loss: {'train/loss': 0.15356135368347168, 'train/grad_norm_pre_clip': 10.687359809875488, 'timing/data': 0.014770030975341797, 'timing/model': 1.2121928400592878, 'learning_rate/qwen_vl_interface': 9.567179606174299e-06, 'learning_rate/action_model': 9.567179606174299e-05, 'epoch': 9.01}) +[2026-06-20 21:24:10,431][starVLA.training.train_starvla][INFO] - Step 4872, Loss: {'train/loss': 0.20469540357589722, 'train/grad_norm_pre_clip': 5.7729082107543945, 'timing/data': 0.011826884932816029, 'timing/model': 1.1451158670242876, 'learning_rate/qwen_vl_interface': 9.565908601856155e-06, 'learning_rate/action_model': 9.565908601856154e-05, 'epoch': 9.01}) +[2026-06-20 21:24:15,322][starVLA.training.train_starvla][INFO] - Step 4873, Loss: {'train/loss': 0.03146732598543167, 'train/grad_norm_pre_clip': 7.33065128326416, 'timing/data': 0.013791149947792292, 'timing/model': 1.4424131480045617, 'learning_rate/qwen_vl_interface': 9.56463582343101e-06, 'learning_rate/action_model': 9.56463582343101e-05, 'epoch': 9.01}) +[2026-06-20 21:24:19,857][starVLA.training.train_starvla][INFO] - Step 4874, Loss: {'train/loss': 0.09847535938024521, 'train/grad_norm_pre_clip': 3.6649630069732666, 'timing/data': 0.015573136042803526, 'timing/model': 1.2149999630637467, 'learning_rate/qwen_vl_interface': 9.563361271422056e-06, 'learning_rate/action_model': 9.563361271422056e-05, 'epoch': 9.01}) +[2026-06-20 21:24:24,684][starVLA.training.train_starvla][INFO] - Step 4875, Loss: {'train/loss': 0.053845107555389404, 'train/grad_norm_pre_clip': 5.66217565536499, 'timing/data': 0.014779143035411835, 'timing/model': 1.2144242270151153, 'learning_rate/qwen_vl_interface': 9.562084946353212e-06, 'learning_rate/action_model': 9.562084946353213e-05, 'epoch': 9.02}) +[2026-06-20 21:24:29,436][starVLA.training.train_starvla][INFO] - Step 4876, Loss: {'train/loss': 0.029652368277311325, 'train/grad_norm_pre_clip': 3.5922043323516846, 'timing/data': 0.015409477986395359, 'timing/model': 1.1954220969928429, 'learning_rate/qwen_vl_interface': 9.560806848749132e-06, 'learning_rate/action_model': 9.560806848749132e-05, 'epoch': 9.02}) +[2026-06-20 21:24:34,215][starVLA.training.train_starvla][INFO] - Step 4877, Loss: {'train/loss': 0.01935369148850441, 'train/grad_norm_pre_clip': 1.584431767463684, 'timing/data': 0.014625063980929554, 'timing/model': 1.212784867035225, 'learning_rate/qwen_vl_interface': 9.559526979135186e-06, 'learning_rate/action_model': 9.559526979135187e-05, 'epoch': 9.02}) +[2026-06-20 21:24:38,957][starVLA.training.train_starvla][INFO] - Step 4878, Loss: {'train/loss': 0.16573801636695862, 'train/grad_norm_pre_clip': 16.532629013061523, 'timing/data': 0.017020765924826264, 'timing/model': 1.2134393780725077, 'learning_rate/qwen_vl_interface': 9.558245338037486e-06, 'learning_rate/action_model': 9.558245338037486e-05, 'epoch': 9.02}) +[2026-06-20 21:24:43,789][starVLA.training.train_starvla][INFO] - Step 4879, Loss: {'train/loss': 0.07599761337041855, 'train/grad_norm_pre_clip': 5.709966659545898, 'timing/data': 0.013146156095899642, 'timing/model': 1.438294888008386, 'learning_rate/qwen_vl_interface': 9.556961925982864e-06, 'learning_rate/action_model': 9.556961925982863e-05, 'epoch': 9.02}) +[2026-06-20 21:24:48,381][starVLA.training.train_starvla][INFO] - Step 4880, Loss: {'train/loss': 0.0709909126162529, 'train/grad_norm_pre_clip': 6.8782734870910645, 'timing/data': 0.013900711084716022, 'timing/model': 1.1969183819601312, 'learning_rate/qwen_vl_interface': 9.555676743498882e-06, 'learning_rate/action_model': 9.555676743498881e-05, 'epoch': 9.02}) +[2026-06-20 21:24:53,307][starVLA.training.train_starvla][INFO] - Step 4881, Loss: {'train/loss': 0.048556361347436905, 'train/grad_norm_pre_clip': 3.219863176345825, 'timing/data': 0.01481337402947247, 'timing/model': 1.2165420859819278, 'learning_rate/qwen_vl_interface': 9.55438979111383e-06, 'learning_rate/action_model': 9.554389791113831e-05, 'epoch': 9.03}) +[2026-06-20 21:24:58,172][starVLA.training.train_starvla][INFO] - Step 4882, Loss: {'train/loss': 0.050833504647016525, 'train/grad_norm_pre_clip': 6.419135570526123, 'timing/data': 0.012061576009728014, 'timing/model': 1.1451804890530184, 'learning_rate/qwen_vl_interface': 9.553101069356728e-06, 'learning_rate/action_model': 9.553101069356727e-05, 'epoch': 9.03}) +[2026-06-20 21:25:02,893][starVLA.training.train_starvla][INFO] - Step 4883, Loss: {'train/loss': 0.10810967534780502, 'train/grad_norm_pre_clip': 8.055824279785156, 'timing/data': 0.012138505931943655, 'timing/model': 1.1456635989015922, 'learning_rate/qwen_vl_interface': 9.551810578757316e-06, 'learning_rate/action_model': 9.551810578757315e-05, 'epoch': 9.03}) +[2026-06-20 21:25:07,709][starVLA.training.train_starvla][INFO] - Step 4884, Loss: {'train/loss': 0.26349541544914246, 'train/grad_norm_pre_clip': 24.230995178222656, 'timing/data': 0.013433046988211572, 'timing/model': 1.1975219229934737, 'learning_rate/qwen_vl_interface': 9.550518319846069e-06, 'learning_rate/action_model': 9.550518319846069e-05, 'epoch': 9.03}) +[2026-06-20 21:25:12,512][starVLA.training.train_starvla][INFO] - Step 4885, Loss: {'train/loss': 0.07348901033401489, 'train/grad_norm_pre_clip': 10.85372257232666, 'timing/data': 0.012001557974144816, 'timing/model': 1.3893681350164115, 'learning_rate/qwen_vl_interface': 9.549224293154187e-06, 'learning_rate/action_model': 9.549224293154187e-05, 'epoch': 9.03}) +[2026-06-20 21:25:16,936][starVLA.training.train_starvla][INFO] - Step 4886, Loss: {'train/loss': 0.06285785138607025, 'train/grad_norm_pre_clip': 10.615877151489258, 'timing/data': 0.015993306995369494, 'timing/model': 1.156567316967994, 'learning_rate/qwen_vl_interface': 9.547928499213593e-06, 'learning_rate/action_model': 9.547928499213593e-05, 'epoch': 9.04}) +[2026-06-20 21:25:21,593][starVLA.training.train_starvla][INFO] - Step 4887, Loss: {'train/loss': 0.06903837621212006, 'train/grad_norm_pre_clip': 11.578563690185547, 'timing/data': 0.012027828022837639, 'timing/model': 1.1430157959694043, 'learning_rate/qwen_vl_interface': 9.54663093855694e-06, 'learning_rate/action_model': 9.54663093855694e-05, 'epoch': 9.04}) +[2026-06-20 21:25:26,257][starVLA.training.train_starvla][INFO] - Step 4888, Loss: {'train/loss': 0.008711818605661392, 'train/grad_norm_pre_clip': 1.1045551300048828, 'timing/data': 0.017491335980594158, 'timing/model': 1.1610343719366938, 'learning_rate/qwen_vl_interface': 9.545331611717607e-06, 'learning_rate/action_model': 9.545331611717608e-05, 'epoch': 9.04}) +[2026-06-20 21:25:30,939][starVLA.training.train_starvla][INFO] - Step 4889, Loss: {'train/loss': 0.12193816155195236, 'train/grad_norm_pre_clip': 7.002227306365967, 'timing/data': 0.011577021912671626, 'timing/model': 1.1422416269779205, 'learning_rate/qwen_vl_interface': 9.544030519229697e-06, 'learning_rate/action_model': 9.544030519229697e-05, 'epoch': 9.04}) +[2026-06-20 21:25:35,512][starVLA.training.train_starvla][INFO] - Step 4890, Loss: {'train/loss': 0.1485319584608078, 'train/grad_norm_pre_clip': 9.68707275390625, 'timing/data': 0.017909773043356836, 'timing/model': 1.1615111330756918, 'learning_rate/qwen_vl_interface': 9.54272766162804e-06, 'learning_rate/action_model': 9.54272766162804e-05, 'epoch': 9.04}) +[2026-06-20 21:25:40,173][starVLA.training.train_starvla][INFO] - Step 4891, Loss: {'train/loss': 0.009073955938220024, 'train/grad_norm_pre_clip': 1.2431713342666626, 'timing/data': 0.011990298982709646, 'timing/model': 1.3956785739865154, 'learning_rate/qwen_vl_interface': 9.54142303944819e-06, 'learning_rate/action_model': 9.541423039448191e-05, 'epoch': 9.05}) +[2026-06-20 21:25:44,603][starVLA.training.train_starvla][INFO] - Step 4892, Loss: {'train/loss': 0.22135187685489655, 'train/grad_norm_pre_clip': 9.186854362487793, 'timing/data': 0.011810400988906622, 'timing/model': 1.1443552969722077, 'learning_rate/qwen_vl_interface': 9.540116653226433e-06, 'learning_rate/action_model': 9.540116653226431e-05, 'epoch': 9.05}) +[2026-06-20 21:25:49,287][starVLA.training.train_starvla][INFO] - Step 4893, Loss: {'train/loss': 0.1781197339296341, 'train/grad_norm_pre_clip': 3.5028576850891113, 'timing/data': 0.012005068943835795, 'timing/model': 1.170625041006133, 'learning_rate/qwen_vl_interface': 9.53880850349977e-06, 'learning_rate/action_model': 9.538808503499768e-05, 'epoch': 9.05}) +[2026-06-20 21:25:53,986][starVLA.training.train_starvla][INFO] - Step 4894, Loss: {'train/loss': 0.023504741489887238, 'train/grad_norm_pre_clip': 8.464847564697266, 'timing/data': 0.012024959083646536, 'timing/model': 1.155104301054962, 'learning_rate/qwen_vl_interface': 9.537498590805934e-06, 'learning_rate/action_model': 9.537498590805933e-05, 'epoch': 9.05}) +[2026-06-20 21:25:58,552][starVLA.training.train_starvla][INFO] - Step 4895, Loss: {'train/loss': 0.2308540940284729, 'train/grad_norm_pre_clip': 7.0072712898254395, 'timing/data': 0.012050998979248106, 'timing/model': 1.1472413619048893, 'learning_rate/qwen_vl_interface': 9.536186915683379e-06, 'learning_rate/action_model': 9.53618691568338e-05, 'epoch': 9.05}) +[2026-06-20 21:26:03,208][starVLA.training.train_starvla][INFO] - Step 4896, Loss: {'train/loss': 0.016435693949460983, 'train/grad_norm_pre_clip': 2.027686357498169, 'timing/data': 0.011772001977078617, 'timing/model': 1.1499245229642838, 'learning_rate/qwen_vl_interface': 9.534873478671289e-06, 'learning_rate/action_model': 9.534873478671288e-05, 'epoch': 9.05}) +[2026-06-20 21:26:07,922][starVLA.training.train_starvla][INFO] - Step 4897, Loss: {'train/loss': 0.02311333827674389, 'train/grad_norm_pre_clip': 4.387554168701172, 'timing/data': 0.012001789989881217, 'timing/model': 1.4184368220157921, 'learning_rate/qwen_vl_interface': 9.533558280309563e-06, 'learning_rate/action_model': 9.533558280309563e-05, 'epoch': 9.06}) +[2026-06-20 21:26:12,338][starVLA.training.train_starvla][INFO] - Step 4898, Loss: {'train/loss': 0.32875677943229675, 'train/grad_norm_pre_clip': 8.182947158813477, 'timing/data': 0.011819000937975943, 'timing/model': 1.1548750480869785, 'learning_rate/qwen_vl_interface': 9.532241321138832e-06, 'learning_rate/action_model': 9.532241321138833e-05, 'epoch': 9.06}) +[2026-06-20 21:26:16,997][starVLA.training.train_starvla][INFO] - Step 4899, Loss: {'train/loss': 0.01602371409535408, 'train/grad_norm_pre_clip': 5.835625171661377, 'timing/data': 0.011894171009771526, 'timing/model': 1.151967871002853, 'learning_rate/qwen_vl_interface': 9.530922601700451e-06, 'learning_rate/action_model': 9.530922601700452e-05, 'epoch': 9.06}) +[2026-06-20 21:26:21,655][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 4900 +[2026-06-20 21:32:51,679][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 4900 +[2026-06-20 21:32:51,680][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 4900 +[2026-06-20 21:39:46,781][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 4900 +[2026-06-20 21:39:46,783][starVLA.training.train_starvla][INFO] - Step 4900, Loss: {'train/loss': 0.20764979720115662, 'train/grad_norm_pre_clip': 11.821213722229004, 'eval/loss': 0.6986148953437805, 'eval/latency_0/loss': 0.8319660425186157, 'eval/latency_1/loss': 1.1078693866729736, 'eval/latency_2/loss': 0.9617787599563599, 'eval/latency_3/loss': 0.47032713890075684, 'eval/latency_4/loss': 0.09012756496667862, 'eval/flappy/loss': 0.6986148953437805, 'eval/flappy/latency_0/loss': 0.8319660425186157, 'eval/flappy/latency_1/loss': 1.1078693866729736, 'eval/flappy/latency_2/loss': 0.9617787599563599, 'eval/flappy/latency_3/loss': 0.47032713890075684, 'eval/flappy/latency_4/loss': 0.09012756496667862, 'eval/flap_precision': 0.2589928057553957, 'eval/flap_recall': 0.13090909090909092, 'eval/flap_f1': 0.17391304347826086, 'eval/flap_rate_model': 0.03861111111111111, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.17391304347826086, 'eval/latency_0/flap_precision': 0.2589928057553957, 'eval/latency_0/flap_recall': 0.13090909090909092, 'eval/latency_0/flap_f1': 0.17391304347826086, 'eval/latency_0/flap_rate_model': 0.03861111111111111, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.17391304347826086, 'timing/data': 0.011969971004873514, 'timing/model': 1.1574517410481349, 'learning_rate/qwen_vl_interface': 9.52960212253649e-06, 'learning_rate/action_model': 9.529602122536491e-05, 'epoch': 9.06}) +[2026-06-20 21:39:46,783][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_4900_state +[2026-06-20 21:39:46,787][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-20 21:39:55,464][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_4900_state/model.safetensors +[2026-06-20 21:40:08,926][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_4900_state/optimizer.bin +[2026-06-20 21:40:08,927][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_4900_state/sampler.bin +[2026-06-20 21:40:08,927][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_4900_state/sampler_1.bin +[2026-06-20 21:40:08,928][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_4900_state/random_states_0.pkl +[2026-06-20 21:40:11,577][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4900 +[2026-06-20 21:40:11,577][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-20 21:40:11,590][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-20 21:40:16,275][starVLA.training.train_starvla][INFO] - Step 4901, Loss: {'train/loss': 0.25275319814682007, 'train/grad_norm_pre_clip': 17.214887619018555, 'timing/data': 0.014619985013268888, 'timing/model': 1.4145848699845374, 'learning_rate/qwen_vl_interface': 9.528279884189754e-06, 'learning_rate/action_model': 9.528279884189754e-05, 'epoch': 9.06}) +[2026-06-20 21:40:21,052][starVLA.training.train_starvla][INFO] - Step 4902, Loss: {'train/loss': 0.10401874035596848, 'train/grad_norm_pre_clip': 9.678116798400879, 'timing/data': 0.012492966023273766, 'timing/model': 1.3697694360744208, 'learning_rate/qwen_vl_interface': 9.526955887203762e-06, 'learning_rate/action_model': 9.526955887203761e-05, 'epoch': 9.07}) +[2026-06-20 21:40:25,599][starVLA.training.train_starvla][INFO] - Step 4903, Loss: {'train/loss': 0.07883360236883163, 'train/grad_norm_pre_clip': 18.776046752929688, 'timing/data': 0.01426656695548445, 'timing/model': 1.187895291019231, 'learning_rate/qwen_vl_interface': 9.525630132122758e-06, 'learning_rate/action_model': 9.525630132122758e-05, 'epoch': 9.07}) +[2026-06-20 21:40:30,320][starVLA.training.train_starvla][INFO] - Step 4904, Loss: {'train/loss': 0.037240490317344666, 'train/grad_norm_pre_clip': 6.747281074523926, 'timing/data': 0.017646285938099027, 'timing/model': 1.1333840939914808, 'learning_rate/qwen_vl_interface': 9.524302619491712e-06, 'learning_rate/action_model': 9.524302619491712e-05, 'epoch': 9.07}) +[2026-06-20 21:40:35,079][starVLA.training.train_starvla][INFO] - Step 4905, Loss: {'train/loss': 0.0572487972676754, 'train/grad_norm_pre_clip': 8.963889122009277, 'timing/data': 0.02219240996055305, 'timing/model': 1.2031489009968936, 'learning_rate/qwen_vl_interface': 9.522973349856315e-06, 'learning_rate/action_model': 9.522973349856315e-05, 'epoch': 9.07}) +[2026-06-20 21:40:39,772][starVLA.training.train_starvla][INFO] - Step 4906, Loss: {'train/loss': 0.20273542404174805, 'train/grad_norm_pre_clip': 24.444246292114258, 'timing/data': 0.020541230915114284, 'timing/model': 1.20197051204741, 'learning_rate/qwen_vl_interface': 9.521642323762976e-06, 'learning_rate/action_model': 9.521642323762976e-05, 'epoch': 9.07}) +[2026-06-20 21:40:44,459][starVLA.training.train_starvla][INFO] - Step 4907, Loss: {'train/loss': 0.07101397961378098, 'train/grad_norm_pre_clip': 7.018948554992676, 'timing/data': 0.017469919053837657, 'timing/model': 1.3684774959692731, 'learning_rate/qwen_vl_interface': 9.520309541758833e-06, 'learning_rate/action_model': 9.520309541758833e-05, 'epoch': 9.07}) +[2026-06-20 21:40:49,220][starVLA.training.train_starvla][INFO] - Step 4908, Loss: {'train/loss': 0.22567905485630035, 'train/grad_norm_pre_clip': 13.181926727294922, 'timing/data': 0.025335702928714454, 'timing/model': 1.419589903904125, 'learning_rate/qwen_vl_interface': 9.51897500439174e-06, 'learning_rate/action_model': 9.51897500439174e-05, 'epoch': 9.08}) +[2026-06-20 21:40:53,725][starVLA.training.train_starvla][INFO] - Step 4909, Loss: {'train/loss': 0.2746242880821228, 'train/grad_norm_pre_clip': 16.173294067382812, 'timing/data': 0.02266409806907177, 'timing/model': 1.1917283129878342, 'learning_rate/qwen_vl_interface': 9.517638712210277e-06, 'learning_rate/action_model': 9.517638712210276e-05, 'epoch': 9.08}) +[2026-06-20 21:40:58,561][starVLA.training.train_starvla][INFO] - Step 4910, Loss: {'train/loss': 0.12100665271282196, 'train/grad_norm_pre_clip': 13.902637481689453, 'timing/data': 0.02209211199078709, 'timing/model': 1.193451686995104, 'learning_rate/qwen_vl_interface': 9.51630066576374e-06, 'learning_rate/action_model': 9.51630066576374e-05, 'epoch': 9.08}) +[2026-06-20 21:41:03,412][starVLA.training.train_starvla][INFO] - Step 4911, Loss: {'train/loss': 0.006562582682818174, 'train/grad_norm_pre_clip': 0.4079788029193878, 'timing/data': 0.025421843049116433, 'timing/model': 1.2076951380586252, 'learning_rate/qwen_vl_interface': 9.514960865602154e-06, 'learning_rate/action_model': 9.514960865602153e-05, 'epoch': 9.08}) +[2026-06-20 21:41:08,175][starVLA.training.train_starvla][INFO] - Step 4912, Loss: {'train/loss': 0.09879015386104584, 'train/grad_norm_pre_clip': 10.826096534729004, 'timing/data': 0.023087245994247496, 'timing/model': 1.2098775999620557, 'learning_rate/qwen_vl_interface': 9.513619312276259e-06, 'learning_rate/action_model': 9.513619312276258e-05, 'epoch': 9.08}) +[2026-06-20 21:41:13,088][starVLA.training.train_starvla][INFO] - Step 4913, Loss: {'train/loss': 0.21034090220928192, 'train/grad_norm_pre_clip': 12.194928169250488, 'timing/data': 0.023926961002871394, 'timing/model': 1.209316757041961, 'learning_rate/qwen_vl_interface': 9.512276006337514e-06, 'learning_rate/action_model': 9.512276006337513e-05, 'epoch': 9.09}) +[2026-06-20 21:41:17,817][starVLA.training.train_starvla][INFO] - Step 4914, Loss: {'train/loss': 0.23449811339378357, 'train/grad_norm_pre_clip': 11.46422290802002, 'timing/data': 0.017867177026346326, 'timing/model': 1.352745920070447, 'learning_rate/qwen_vl_interface': 9.510930948338102e-06, 'learning_rate/action_model': 9.510930948338102e-05, 'epoch': 9.09}) +[2026-06-20 21:41:22,407][starVLA.training.train_starvla][INFO] - Step 4915, Loss: {'train/loss': 0.22913053631782532, 'train/grad_norm_pre_clip': 19.058002471923828, 'timing/data': 0.017331709968857467, 'timing/model': 1.1433699269546196, 'learning_rate/qwen_vl_interface': 9.509584138830929e-06, 'learning_rate/action_model': 9.509584138830928e-05, 'epoch': 9.09}) +[2026-06-20 21:41:27,177][starVLA.training.train_starvla][INFO] - Step 4916, Loss: {'train/loss': 0.22490064799785614, 'train/grad_norm_pre_clip': 9.699660301208496, 'timing/data': 0.018749510985799134, 'timing/model': 1.1948969929944724, 'learning_rate/qwen_vl_interface': 9.508235578369613e-06, 'learning_rate/action_model': 9.508235578369612e-05, 'epoch': 9.09}) +[2026-06-20 21:41:32,027][starVLA.training.train_starvla][INFO] - Step 4917, Loss: {'train/loss': 0.14461416006088257, 'train/grad_norm_pre_clip': 11.19832992553711, 'timing/data': 0.02406024094671011, 'timing/model': 1.21214868593961, 'learning_rate/qwen_vl_interface': 9.506885267508498e-06, 'learning_rate/action_model': 9.506885267508498e-05, 'epoch': 9.09}) +[2026-06-20 21:41:36,925][starVLA.training.train_starvla][INFO] - Step 4918, Loss: {'train/loss': 0.07682657241821289, 'train/grad_norm_pre_clip': 6.31728982925415, 'timing/data': 0.020576100097969174, 'timing/model': 1.1945124429184943, 'learning_rate/qwen_vl_interface': 9.505533206802648e-06, 'learning_rate/action_model': 9.505533206802647e-05, 'epoch': 9.1}) +[2026-06-20 21:41:41,699][starVLA.training.train_starvla][INFO] - Step 4919, Loss: {'train/loss': 0.16634346544742584, 'train/grad_norm_pre_clip': 8.58231258392334, 'timing/data': 0.01969150605145842, 'timing/model': 1.450763123924844, 'learning_rate/qwen_vl_interface': 9.50417939680784e-06, 'learning_rate/action_model': 9.504179396807839e-05, 'epoch': 9.1}) +[2026-06-20 21:41:46,301][starVLA.training.train_starvla][INFO] - Step 4920, Loss: {'train/loss': 0.14007623493671417, 'train/grad_norm_pre_clip': 7.938348770141602, 'timing/data': 0.017495658015832305, 'timing/model': 1.1444056989857927, 'learning_rate/qwen_vl_interface': 9.502823838080578e-06, 'learning_rate/action_model': 9.502823838080578e-05, 'epoch': 9.1}) +[2026-06-20 21:41:51,044][starVLA.training.train_starvla][INFO] - Step 4921, Loss: {'train/loss': 0.09537290781736374, 'train/grad_norm_pre_clip': 7.743836879730225, 'timing/data': 0.017514458973892033, 'timing/model': 1.1444330819649622, 'learning_rate/qwen_vl_interface': 9.501466531178079e-06, 'learning_rate/action_model': 9.501466531178079e-05, 'epoch': 9.1}) +[2026-06-20 21:41:55,684][starVLA.training.train_starvla][INFO] - Step 4922, Loss: {'train/loss': 0.045175306499004364, 'train/grad_norm_pre_clip': 4.349863052368164, 'timing/data': 0.017767647048458457, 'timing/model': 1.1436022699344903, 'learning_rate/qwen_vl_interface': 9.50010747665828e-06, 'learning_rate/action_model': 9.500107476658281e-05, 'epoch': 9.1}) +[2026-06-20 21:42:00,435][starVLA.training.train_starvla][INFO] - Step 4923, Loss: {'train/loss': 0.16496215760707855, 'train/grad_norm_pre_clip': 7.8850860595703125, 'timing/data': 0.02211592299863696, 'timing/model': 1.2138150780228898, 'learning_rate/qwen_vl_interface': 9.498746675079842e-06, 'learning_rate/action_model': 9.49874667507984e-05, 'epoch': 9.1}) +[2026-06-20 21:42:05,237][starVLA.training.train_starvla][INFO] - Step 4924, Loss: {'train/loss': 0.05109724774956703, 'train/grad_norm_pre_clip': 4.075944423675537, 'timing/data': 0.021030789939686656, 'timing/model': 1.196187873953022, 'learning_rate/qwen_vl_interface': 9.49738412700213e-06, 'learning_rate/action_model': 9.49738412700213e-05, 'epoch': 9.11}) +[2026-06-20 21:42:10,075][starVLA.training.train_starvla][INFO] - Step 4925, Loss: {'train/loss': 0.04768526926636696, 'train/grad_norm_pre_clip': 5.777768611907959, 'timing/data': 0.02053854102268815, 'timing/model': 1.435995835927315, 'learning_rate/qwen_vl_interface': 9.496019832985243e-06, 'learning_rate/action_model': 9.496019832985243e-05, 'epoch': 9.11}) +[2026-06-20 21:42:14,619][starVLA.training.train_starvla][INFO] - Step 4926, Loss: {'train/loss': 0.041491784155368805, 'train/grad_norm_pre_clip': 4.1391167640686035, 'timing/data': 0.017539708991535008, 'timing/model': 1.1468787949997932, 'learning_rate/qwen_vl_interface': 9.49465379358999e-06, 'learning_rate/action_model': 9.49465379358999e-05, 'epoch': 9.11}) +[2026-06-20 21:42:19,443][starVLA.training.train_starvla][INFO] - Step 4927, Loss: {'train/loss': 0.030766014009714127, 'train/grad_norm_pre_clip': 3.3253023624420166, 'timing/data': 0.017427109996788204, 'timing/model': 1.1916285599581897, 'learning_rate/qwen_vl_interface': 9.493286009377896e-06, 'learning_rate/action_model': 9.493286009377896e-05, 'epoch': 9.11}) +[2026-06-20 21:42:24,239][starVLA.training.train_starvla][INFO] - Step 4928, Loss: {'train/loss': 0.06651397049427032, 'train/grad_norm_pre_clip': 4.870034217834473, 'timing/data': 0.025794351007789373, 'timing/model': 1.213067800970748, 'learning_rate/qwen_vl_interface': 9.491916480911208e-06, 'learning_rate/action_model': 9.491916480911207e-05, 'epoch': 9.11}) +[2026-06-20 21:42:29,081][starVLA.training.train_starvla][INFO] - Step 4929, Loss: {'train/loss': 0.06982254981994629, 'train/grad_norm_pre_clip': 7.2362961769104, 'timing/data': 0.019669436966069043, 'timing/model': 1.195581594016403, 'learning_rate/qwen_vl_interface': 9.490545208752884e-06, 'learning_rate/action_model': 9.490545208752883e-05, 'epoch': 9.12}) +[2026-06-20 21:42:33,899][starVLA.training.train_starvla][INFO] - Step 4930, Loss: {'train/loss': 0.02492651715874672, 'train/grad_norm_pre_clip': 1.8584659099578857, 'timing/data': 0.017376929987221956, 'timing/model': 1.1435042370576411, 'learning_rate/qwen_vl_interface': 9.489172193466603e-06, 'learning_rate/action_model': 9.489172193466603e-05, 'epoch': 9.12}) +[2026-06-20 21:42:38,759][starVLA.training.train_starvla][INFO] - Step 4931, Loss: {'train/loss': 0.2578098475933075, 'train/grad_norm_pre_clip': 16.594039916992188, 'timing/data': 0.02216205291915685, 'timing/model': 1.4481079899705946, 'learning_rate/qwen_vl_interface': 9.487797435616762e-06, 'learning_rate/action_model': 9.487797435616761e-05, 'epoch': 9.12}) +[2026-06-20 21:42:43,291][starVLA.training.train_starvla][INFO] - Step 4932, Loss: {'train/loss': 0.030571220442652702, 'train/grad_norm_pre_clip': 2.5223312377929688, 'timing/data': 0.019575868034735322, 'timing/model': 1.1977727019693702, 'learning_rate/qwen_vl_interface': 9.486420935768472e-06, 'learning_rate/action_model': 9.486420935768472e-05, 'epoch': 9.12}) +[2026-06-20 21:42:48,023][starVLA.training.train_starvla][INFO] - Step 4933, Loss: {'train/loss': 0.49586236476898193, 'train/grad_norm_pre_clip': 18.445898056030273, 'timing/data': 0.017476059962064028, 'timing/model': 1.1444876899477094, 'learning_rate/qwen_vl_interface': 9.485042694487558e-06, 'learning_rate/action_model': 9.485042694487559e-05, 'epoch': 9.12}) +[2026-06-20 21:42:52,855][starVLA.training.train_starvla][INFO] - Step 4934, Loss: {'train/loss': 0.041730087250471115, 'train/grad_norm_pre_clip': 7.229304790496826, 'timing/data': 0.01845707301981747, 'timing/model': 1.1945018579717726, 'learning_rate/qwen_vl_interface': 9.483662712340565e-06, 'learning_rate/action_model': 9.483662712340564e-05, 'epoch': 9.12}) +[2026-06-20 21:42:57,680][starVLA.training.train_starvla][INFO] - Step 4935, Loss: {'train/loss': 0.1480282098054886, 'train/grad_norm_pre_clip': 10.4312162399292, 'timing/data': 0.02234641695395112, 'timing/model': 1.1962887790286914, 'learning_rate/qwen_vl_interface': 9.482280989894748e-06, 'learning_rate/action_model': 9.482280989894748e-05, 'epoch': 9.13}) +[2026-06-20 21:43:02,363][starVLA.training.train_starvla][INFO] - Step 4936, Loss: {'train/loss': 0.06083156168460846, 'train/grad_norm_pre_clip': 12.640536308288574, 'timing/data': 0.021817470085807145, 'timing/model': 1.2103683010209352, 'learning_rate/qwen_vl_interface': 9.480897527718083e-06, 'learning_rate/action_model': 9.480897527718084e-05, 'epoch': 9.13}) +[2026-06-20 21:43:07,134][starVLA.training.train_starvla][INFO] - Step 4937, Loss: {'train/loss': 0.06544288992881775, 'train/grad_norm_pre_clip': 6.0782270431518555, 'timing/data': 0.019547587959095836, 'timing/model': 1.4359821929829195, 'learning_rate/qwen_vl_interface': 9.479512326379263e-06, 'learning_rate/action_model': 9.479512326379264e-05, 'epoch': 9.13}) +[2026-06-20 21:43:11,740][starVLA.training.train_starvla][INFO] - Step 4938, Loss: {'train/loss': 0.16597411036491394, 'train/grad_norm_pre_clip': 15.513293266296387, 'timing/data': 0.02135542395990342, 'timing/model': 1.196204319014214, 'learning_rate/qwen_vl_interface': 9.478125386447686e-06, 'learning_rate/action_model': 9.478125386447686e-05, 'epoch': 9.13}) +[2026-06-20 21:43:16,515][starVLA.training.train_starvla][INFO] - Step 4939, Loss: {'train/loss': 0.04356786236166954, 'train/grad_norm_pre_clip': 9.532014846801758, 'timing/data': 0.02079141209833324, 'timing/model': 1.2011525540146977, 'learning_rate/qwen_vl_interface': 9.476736708493472e-06, 'learning_rate/action_model': 9.476736708493472e-05, 'epoch': 9.13}) +[2026-06-20 21:43:21,415][starVLA.training.train_starvla][INFO] - Step 4940, Loss: {'train/loss': 0.12497919052839279, 'train/grad_norm_pre_clip': 11.881783485412598, 'timing/data': 0.0185000819619745, 'timing/model': 1.1987301389453933, 'learning_rate/qwen_vl_interface': 9.475346293087457e-06, 'learning_rate/action_model': 9.475346293087458e-05, 'epoch': 9.14}) +[2026-06-20 21:43:26,099][starVLA.training.train_starvla][INFO] - Step 4941, Loss: {'train/loss': 0.22859056293964386, 'train/grad_norm_pre_clip': 14.595248222351074, 'timing/data': 0.022530760965310037, 'timing/model': 1.1974110350711271, 'learning_rate/qwen_vl_interface': 9.473954140801184e-06, 'learning_rate/action_model': 9.473954140801185e-05, 'epoch': 9.14}) +[2026-06-20 21:43:30,960][starVLA.training.train_starvla][INFO] - Step 4942, Loss: {'train/loss': 0.031187262386083603, 'train/grad_norm_pre_clip': 4.497241973876953, 'timing/data': 0.02373401599470526, 'timing/model': 1.2132597410818562, 'learning_rate/qwen_vl_interface': 9.472560252206918e-06, 'learning_rate/action_model': 9.472560252206918e-05, 'epoch': 9.14}) +[2026-06-20 21:43:35,804][starVLA.training.train_starvla][INFO] - Step 4943, Loss: {'train/loss': 0.17584830522537231, 'train/grad_norm_pre_clip': 22.3007755279541, 'timing/data': 0.019225833006203175, 'timing/model': 1.4400701959384605, 'learning_rate/qwen_vl_interface': 9.471164627877633e-06, 'learning_rate/action_model': 9.471164627877633e-05, 'epoch': 9.14}) +[2026-06-20 21:43:40,485][starVLA.training.train_starvla][INFO] - Step 4944, Loss: {'train/loss': 0.2578026056289673, 'train/grad_norm_pre_clip': 6.783394813537598, 'timing/data': 0.023240520036779344, 'timing/model': 1.1985772649059072, 'learning_rate/qwen_vl_interface': 9.469767268387016e-06, 'learning_rate/action_model': 9.469767268387016e-05, 'epoch': 9.14}) +[2026-06-20 21:43:45,310][starVLA.training.train_starvla][INFO] - Step 4945, Loss: {'train/loss': 0.3545364439487457, 'train/grad_norm_pre_clip': 12.990211486816406, 'timing/data': 0.025223065982572734, 'timing/model': 1.2179846490034834, 'learning_rate/qwen_vl_interface': 9.468368174309468e-06, 'learning_rate/action_model': 9.468368174309469e-05, 'epoch': 9.15}) +[2026-06-20 21:43:50,092][starVLA.training.train_starvla][INFO] - Step 4946, Loss: {'train/loss': 0.2296377718448639, 'train/grad_norm_pre_clip': 17.78034210205078, 'timing/data': 0.01746519806329161, 'timing/model': 1.1468881510663778, 'learning_rate/qwen_vl_interface': 9.466967346220107e-06, 'learning_rate/action_model': 9.466967346220107e-05, 'epoch': 9.15}) +[2026-06-20 21:43:54,925][starVLA.training.train_starvla][INFO] - Step 4947, Loss: {'train/loss': 0.04930214211344719, 'train/grad_norm_pre_clip': 9.094322204589844, 'timing/data': 0.022271095076575875, 'timing/model': 1.2129094840493053, 'learning_rate/qwen_vl_interface': 9.465564784694759e-06, 'learning_rate/action_model': 9.465564784694759e-05, 'epoch': 9.15}) +[2026-06-20 21:43:59,736][starVLA.training.train_starvla][INFO] - Step 4948, Loss: {'train/loss': 0.45761266350746155, 'train/grad_norm_pre_clip': 19.32611656188965, 'timing/data': 0.017172417021356523, 'timing/model': 1.1479539789725095, 'learning_rate/qwen_vl_interface': 9.464160490309962e-06, 'learning_rate/action_model': 9.464160490309962e-05, 'epoch': 9.15}) +[2026-06-20 21:44:04,613][starVLA.training.train_starvla][INFO] - Step 4949, Loss: {'train/loss': 0.2330751121044159, 'train/grad_norm_pre_clip': 12.97913932800293, 'timing/data': 0.023070888011716306, 'timing/model': 1.448996192892082, 'learning_rate/qwen_vl_interface': 9.462754463642972e-06, 'learning_rate/action_model': 9.46275446364297e-05, 'epoch': 9.15}) +[2026-06-20 21:44:09,304][starVLA.training.train_starvla][INFO] - Step 4950, Loss: {'train/loss': 0.0696113333106041, 'train/grad_norm_pre_clip': 9.532296180725098, 'timing/data': 0.02680875698570162, 'timing/model': 1.2116799330106005, 'learning_rate/qwen_vl_interface': 9.461346705271753e-06, 'learning_rate/action_model': 9.461346705271752e-05, 'epoch': 9.15}) +[2026-06-20 21:44:14,194][starVLA.training.train_starvla][INFO] - Step 4951, Loss: {'train/loss': 0.047814782708883286, 'train/grad_norm_pre_clip': 4.243082046508789, 'timing/data': 0.023287334945052862, 'timing/model': 1.2117536080768332, 'learning_rate/qwen_vl_interface': 9.459937215774976e-06, 'learning_rate/action_model': 9.459937215774976e-05, 'epoch': 9.16}) +[2026-06-20 21:44:18,953][starVLA.training.train_starvla][INFO] - Step 4952, Loss: {'train/loss': 0.03519246354699135, 'train/grad_norm_pre_clip': 4.331364631652832, 'timing/data': 0.02274429309181869, 'timing/model': 1.195836492930539, 'learning_rate/qwen_vl_interface': 9.458525995732035e-06, 'learning_rate/action_model': 9.458525995732035e-05, 'epoch': 9.16}) +[2026-06-20 21:44:23,733][starVLA.training.train_starvla][INFO] - Step 4953, Loss: {'train/loss': 0.09911585599184036, 'train/grad_norm_pre_clip': 10.74989128112793, 'timing/data': 0.02200666896533221, 'timing/model': 1.2132949529914185, 'learning_rate/qwen_vl_interface': 9.457113045723027e-06, 'learning_rate/action_model': 9.457113045723026e-05, 'epoch': 9.16}) +[2026-06-20 21:44:28,568][starVLA.training.train_starvla][INFO] - Step 4954, Loss: {'train/loss': 0.1277141273021698, 'train/grad_norm_pre_clip': 7.36927604675293, 'timing/data': 0.02425137592945248, 'timing/model': 1.2141263030935079, 'learning_rate/qwen_vl_interface': 9.455698366328762e-06, 'learning_rate/action_model': 9.455698366328762e-05, 'epoch': 9.16}) +[2026-06-20 21:44:33,398][starVLA.training.train_starvla][INFO] - Step 4955, Loss: {'train/loss': 0.0841609314084053, 'train/grad_norm_pre_clip': 8.671911239624023, 'timing/data': 0.02391049894504249, 'timing/model': 1.4481168439378962, 'learning_rate/qwen_vl_interface': 9.454281958130761e-06, 'learning_rate/action_model': 9.454281958130761e-05, 'epoch': 9.16}) +[2026-06-20 21:44:37,995][starVLA.training.train_starvla][INFO] - Step 4956, Loss: {'train/loss': 0.031741540879011154, 'train/grad_norm_pre_clip': 5.107644557952881, 'timing/data': 0.024449630989693105, 'timing/model': 1.1952726839808747, 'learning_rate/qwen_vl_interface': 9.452863821711257e-06, 'learning_rate/action_model': 9.452863821711258e-05, 'epoch': 9.17}) +[2026-06-20 21:44:42,840][starVLA.training.train_starvla][INFO] - Step 4957, Loss: {'train/loss': 0.20690283179283142, 'train/grad_norm_pre_clip': 12.766609191894531, 'timing/data': 0.02170003193896264, 'timing/model': 1.1986032349523157, 'learning_rate/qwen_vl_interface': 9.451443957653195e-06, 'learning_rate/action_model': 9.451443957653195e-05, 'epoch': 9.17}) +[2026-06-20 21:44:47,657][starVLA.training.train_starvla][INFO] - Step 4958, Loss: {'train/loss': 0.140370711684227, 'train/grad_norm_pre_clip': 10.123114585876465, 'timing/data': 0.02201050298754126, 'timing/model': 1.213615366956219, 'learning_rate/qwen_vl_interface': 9.450022366540221e-06, 'learning_rate/action_model': 9.45002236654022e-05, 'epoch': 9.17}) +[2026-06-20 21:44:52,455][starVLA.training.train_starvla][INFO] - Step 4959, Loss: {'train/loss': 0.208296000957489, 'train/grad_norm_pre_clip': 12.455336570739746, 'timing/data': 0.023324246052652597, 'timing/model': 1.2141766829881817, 'learning_rate/qwen_vl_interface': 9.448599048956702e-06, 'learning_rate/action_model': 9.448599048956702e-05, 'epoch': 9.17}) +[2026-06-20 21:44:57,275][starVLA.training.train_starvla][INFO] - Step 4960, Loss: {'train/loss': 0.19081899523735046, 'train/grad_norm_pre_clip': 13.007303237915039, 'timing/data': 0.021162745892070234, 'timing/model': 1.1959249339997768, 'learning_rate/qwen_vl_interface': 9.447174005487708e-06, 'learning_rate/action_model': 9.447174005487708e-05, 'epoch': 9.17}) +[2026-06-20 21:45:02,097][starVLA.training.train_starvla][INFO] - Step 4961, Loss: {'train/loss': 0.09046648442745209, 'train/grad_norm_pre_clip': 8.835418701171875, 'timing/data': 0.021548517979681492, 'timing/model': 1.4244853009004146, 'learning_rate/qwen_vl_interface': 9.445747236719022e-06, 'learning_rate/action_model': 9.445747236719022e-05, 'epoch': 9.17}) +[2026-06-20 21:45:06,760][starVLA.training.train_starvla][INFO] - Step 4962, Loss: {'train/loss': 0.08453700691461563, 'train/grad_norm_pre_clip': 6.896298885345459, 'timing/data': 0.017295561963692307, 'timing/model': 1.3251540920464322, 'learning_rate/qwen_vl_interface': 9.444318743237135e-06, 'learning_rate/action_model': 9.444318743237134e-05, 'epoch': 9.18}) +[2026-06-20 21:45:11,718][starVLA.training.train_starvla][INFO] - Step 4963, Loss: {'train/loss': 0.02896112948656082, 'train/grad_norm_pre_clip': 2.8475420475006104, 'timing/data': 0.021936105913482606, 'timing/model': 1.1968883420340717, 'learning_rate/qwen_vl_interface': 9.442888525629244e-06, 'learning_rate/action_model': 9.442888525629245e-05, 'epoch': 9.18}) +[2026-06-20 21:45:16,488][starVLA.training.train_starvla][INFO] - Step 4964, Loss: {'train/loss': 0.056204382330179214, 'train/grad_norm_pre_clip': 5.229748725891113, 'timing/data': 0.025376664008945227, 'timing/model': 1.1972027369774878, 'learning_rate/qwen_vl_interface': 9.441456584483261e-06, 'learning_rate/action_model': 9.441456584483261e-05, 'epoch': 9.18}) +[2026-06-20 21:45:21,314][starVLA.training.train_starvla][INFO] - Step 4965, Loss: {'train/loss': 0.03677979111671448, 'train/grad_norm_pre_clip': 3.757239818572998, 'timing/data': 0.021530624013394117, 'timing/model': 1.193483673967421, 'learning_rate/qwen_vl_interface': 9.440022920387801e-06, 'learning_rate/action_model': 9.440022920387801e-05, 'epoch': 9.18}) +[2026-06-20 21:45:26,160][starVLA.training.train_starvla][INFO] - Step 4966, Loss: {'train/loss': 0.16214072704315186, 'train/grad_norm_pre_clip': 10.785687446594238, 'timing/data': 0.018614977947436273, 'timing/model': 1.1466150879859924, 'learning_rate/qwen_vl_interface': 9.438587533932191e-06, 'learning_rate/action_model': 9.43858753393219e-05, 'epoch': 9.18}) +[2026-06-20 21:45:31,033][starVLA.training.train_starvla][INFO] - Step 4967, Loss: {'train/loss': 0.14396759867668152, 'train/grad_norm_pre_clip': 6.660931587219238, 'timing/data': 0.02629457402508706, 'timing/model': 1.4230447199661285, 'learning_rate/qwen_vl_interface': 9.43715042570646e-06, 'learning_rate/action_model': 9.43715042570646e-05, 'epoch': 9.19}) +[2026-06-20 21:45:35,947][starVLA.training.train_starvla][INFO] - Step 4968, Loss: {'train/loss': 0.11375456303358078, 'train/grad_norm_pre_clip': 13.34097957611084, 'timing/data': 0.022337756003253162, 'timing/model': 1.4504645650740713, 'learning_rate/qwen_vl_interface': 9.435711596301353e-06, 'learning_rate/action_model': 9.435711596301352e-05, 'epoch': 9.19}) +[2026-06-20 21:45:40,583][starVLA.training.train_starvla][INFO] - Step 4969, Loss: {'train/loss': 0.00618917727842927, 'train/grad_norm_pre_clip': 0.36207494139671326, 'timing/data': 0.023243071977049112, 'timing/model': 1.2152160860132426, 'learning_rate/qwen_vl_interface': 9.434271046308317e-06, 'learning_rate/action_model': 9.434271046308317e-05, 'epoch': 9.19}) +[2026-06-20 21:45:45,296][starVLA.training.train_starvla][INFO] - Step 4970, Loss: {'train/loss': 0.14920823276042938, 'train/grad_norm_pre_clip': 9.898972511291504, 'timing/data': 0.018742799991741776, 'timing/model': 1.1526275780051947, 'learning_rate/qwen_vl_interface': 9.432828776319507e-06, 'learning_rate/action_model': 9.432828776319505e-05, 'epoch': 9.19}) +[2026-06-20 21:45:50,161][starVLA.training.train_starvla][INFO] - Step 4971, Loss: {'train/loss': 0.02247687056660652, 'train/grad_norm_pre_clip': 2.0368998050689697, 'timing/data': 0.018794717965647578, 'timing/model': 1.1483290260657668, 'learning_rate/qwen_vl_interface': 9.431384786927785e-06, 'learning_rate/action_model': 9.431384786927786e-05, 'epoch': 9.19}) +[2026-06-20 21:45:55,028][starVLA.training.train_starvla][INFO] - Step 4972, Loss: {'train/loss': 0.119841568171978, 'train/grad_norm_pre_clip': 8.107101440429688, 'timing/data': 0.021465764031745493, 'timing/model': 1.2139315020758659, 'learning_rate/qwen_vl_interface': 9.429939078726723e-06, 'learning_rate/action_model': 9.429939078726722e-05, 'epoch': 9.2}) +[2026-06-20 21:45:59,904][starVLA.training.train_starvla][INFO] - Step 4973, Loss: {'train/loss': 0.024500805884599686, 'train/grad_norm_pre_clip': 4.23268985748291, 'timing/data': 0.021812874008901417, 'timing/model': 1.455219198949635, 'learning_rate/qwen_vl_interface': 9.428491652310598e-06, 'learning_rate/action_model': 9.428491652310597e-05, 'epoch': 9.2}) +[2026-06-20 21:46:04,375][starVLA.training.train_starvla][INFO] - Step 4974, Loss: {'train/loss': 0.15662871301174164, 'train/grad_norm_pre_clip': 8.583647727966309, 'timing/data': 0.018533195951022208, 'timing/model': 1.1473378560040146, 'learning_rate/qwen_vl_interface': 9.427042508274386e-06, 'learning_rate/action_model': 9.427042508274386e-05, 'epoch': 9.2}) +[2026-06-20 21:46:09,225][starVLA.training.train_starvla][INFO] - Step 4975, Loss: {'train/loss': 0.04614228755235672, 'train/grad_norm_pre_clip': 6.905299663543701, 'timing/data': 0.022744318936020136, 'timing/model': 1.2130024069920182, 'learning_rate/qwen_vl_interface': 9.425591647213783e-06, 'learning_rate/action_model': 9.425591647213782e-05, 'epoch': 9.2}) +[2026-06-20 21:46:14,004][starVLA.training.train_starvla][INFO] - Step 4976, Loss: {'train/loss': 0.02777407318353653, 'train/grad_norm_pre_clip': 6.982367992401123, 'timing/data': 0.021306321024894714, 'timing/model': 1.2147105570184067, 'learning_rate/qwen_vl_interface': 9.424139069725177e-06, 'learning_rate/action_model': 9.424139069725178e-05, 'epoch': 9.2}) +[2026-06-20 21:46:18,911][starVLA.training.train_starvla][INFO] - Step 4977, Loss: {'train/loss': 0.007174189668148756, 'train/grad_norm_pre_clip': 0.8135793209075928, 'timing/data': 0.02113028708845377, 'timing/model': 1.2151106050005183, 'learning_rate/qwen_vl_interface': 9.422684776405673e-06, 'learning_rate/action_model': 9.422684776405672e-05, 'epoch': 9.2}) +[2026-06-20 21:46:23,693][starVLA.training.train_starvla][INFO] - Step 4978, Loss: {'train/loss': 0.0951167568564415, 'train/grad_norm_pre_clip': 16.69366455078125, 'timing/data': 0.02289976191241294, 'timing/model': 1.2147194910794497, 'learning_rate/qwen_vl_interface': 9.421228767853072e-06, 'learning_rate/action_model': 9.421228767853072e-05, 'epoch': 9.21}) +[2026-06-20 21:46:28,440][starVLA.training.train_starvla][INFO] - Step 4979, Loss: {'train/loss': 0.014743022620677948, 'train/grad_norm_pre_clip': 3.7252180576324463, 'timing/data': 0.018774577067233622, 'timing/model': 1.3495138708967716, 'learning_rate/qwen_vl_interface': 9.419771044665887e-06, 'learning_rate/action_model': 9.419771044665887e-05, 'epoch': 9.21}) +[2026-06-20 21:46:32,925][starVLA.training.train_starvla][INFO] - Step 4980, Loss: {'train/loss': 0.15082789957523346, 'train/grad_norm_pre_clip': 16.378236770629883, 'timing/data': 0.02087919705081731, 'timing/model': 1.2135613650316373, 'learning_rate/qwen_vl_interface': 9.41831160744333e-06, 'learning_rate/action_model': 9.41831160744333e-05, 'epoch': 9.21}) +[2026-06-20 21:46:37,703][starVLA.training.train_starvla][INFO] - Step 4981, Loss: {'train/loss': 0.011577649973332882, 'train/grad_norm_pre_clip': 1.7273406982421875, 'timing/data': 0.02224602794740349, 'timing/model': 1.2137164699379355, 'learning_rate/qwen_vl_interface': 9.416850456785322e-06, 'learning_rate/action_model': 9.416850456785321e-05, 'epoch': 9.21}) +[2026-06-20 21:46:42,624][starVLA.training.train_starvla][INFO] - Step 4982, Loss: {'train/loss': 0.09674942493438721, 'train/grad_norm_pre_clip': 15.151167869567871, 'timing/data': 0.022980371955782175, 'timing/model': 1.2134222090244293, 'learning_rate/qwen_vl_interface': 9.415387593292489e-06, 'learning_rate/action_model': 9.415387593292488e-05, 'epoch': 9.21}) +[2026-06-20 21:46:47,487][starVLA.training.train_starvla][INFO] - Step 4983, Loss: {'train/loss': 0.013241183944046497, 'train/grad_norm_pre_clip': 3.147282361984253, 'timing/data': 0.018103803042322397, 'timing/model': 1.1468983190134168, 'learning_rate/qwen_vl_interface': 9.413923017566155e-06, 'learning_rate/action_model': 9.413923017566155e-05, 'epoch': 9.22}) +[2026-06-20 21:46:52,347][starVLA.training.train_starvla][INFO] - Step 4984, Loss: {'train/loss': 0.08412361890077591, 'train/grad_norm_pre_clip': 5.0902018547058105, 'timing/data': 0.0180424100253731, 'timing/model': 1.1435136310756207, 'learning_rate/qwen_vl_interface': 9.412456730208352e-06, 'learning_rate/action_model': 9.412456730208352e-05, 'epoch': 9.22}) +[2026-06-20 21:46:57,296][starVLA.training.train_starvla][INFO] - Step 4985, Loss: {'train/loss': 0.2576352655887604, 'train/grad_norm_pre_clip': 12.18785572052002, 'timing/data': 0.023998982971534133, 'timing/model': 1.4529978529317304, 'learning_rate/qwen_vl_interface': 9.41098873182182e-06, 'learning_rate/action_model': 9.41098873182182e-05, 'epoch': 9.22}) +[2026-06-20 21:47:01,844][starVLA.training.train_starvla][INFO] - Step 4986, Loss: {'train/loss': 0.14527520537376404, 'train/grad_norm_pre_clip': 15.089524269104004, 'timing/data': 0.022139244014397264, 'timing/model': 1.1964659380028024, 'learning_rate/qwen_vl_interface': 9.409519023009993e-06, 'learning_rate/action_model': 9.409519023009993e-05, 'epoch': 9.22}) +[2026-06-20 21:47:06,735][starVLA.training.train_starvla][INFO] - Step 4987, Loss: {'train/loss': 0.4921729266643524, 'train/grad_norm_pre_clip': 13.13376522064209, 'timing/data': 0.02341670193709433, 'timing/model': 1.2141322799725458, 'learning_rate/qwen_vl_interface': 9.408047604377014e-06, 'learning_rate/action_model': 9.408047604377014e-05, 'epoch': 9.22}) +[2026-06-20 21:47:11,537][starVLA.training.train_starvla][INFO] - Step 4988, Loss: {'train/loss': 0.20695838332176208, 'train/grad_norm_pre_clip': 11.999797821044922, 'timing/data': 0.027164764003828168, 'timing/model': 1.2111399039858952, 'learning_rate/qwen_vl_interface': 9.40657447652773e-06, 'learning_rate/action_model': 9.406574476527731e-05, 'epoch': 9.22}) +[2026-06-20 21:47:16,284][starVLA.training.train_starvla][INFO] - Step 4989, Loss: {'train/loss': 0.0038063405081629753, 'train/grad_norm_pre_clip': 0.4357530176639557, 'timing/data': 0.017782707000151277, 'timing/model': 1.1446544119389728, 'learning_rate/qwen_vl_interface': 9.405099640067685e-06, 'learning_rate/action_model': 9.405099640067685e-05, 'epoch': 9.23}) +[2026-06-20 21:47:21,021][starVLA.training.train_starvla][INFO] - Step 4990, Loss: {'train/loss': 0.0811876505613327, 'train/grad_norm_pre_clip': 7.89748477935791, 'timing/data': 0.017667743959464133, 'timing/model': 1.143545379047282, 'learning_rate/qwen_vl_interface': 9.40362309560313e-06, 'learning_rate/action_model': 9.40362309560313e-05, 'epoch': 9.23}) +[2026-06-20 21:47:25,840][starVLA.training.train_starvla][INFO] - Step 4991, Loss: {'train/loss': 0.06563974171876907, 'train/grad_norm_pre_clip': 11.206818580627441, 'timing/data': 0.01766212296206504, 'timing/model': 1.387129407026805, 'learning_rate/qwen_vl_interface': 9.402144843741016e-06, 'learning_rate/action_model': 9.402144843741017e-05, 'epoch': 9.23}) +[2026-06-20 21:47:30,484][starVLA.training.train_starvla][INFO] - Step 4992, Loss: {'train/loss': 0.10784568637609482, 'train/grad_norm_pre_clip': 5.378344535827637, 'timing/data': 0.028067008941434324, 'timing/model': 1.2118570749880746, 'learning_rate/qwen_vl_interface': 9.400664885088999e-06, 'learning_rate/action_model': 9.400664885088999e-05, 'epoch': 9.23}) +[2026-06-20 21:47:35,245][starVLA.training.train_starvla][INFO] - Step 4993, Loss: {'train/loss': 0.07816216349601746, 'train/grad_norm_pre_clip': 15.27008056640625, 'timing/data': 0.01934340293519199, 'timing/model': 1.186321689048782, 'learning_rate/qwen_vl_interface': 9.39918322025543e-06, 'learning_rate/action_model': 9.39918322025543e-05, 'epoch': 9.23}) +[2026-06-20 21:47:40,113][starVLA.training.train_starvla][INFO] - Step 4994, Loss: {'train/loss': 0.013418310321867466, 'train/grad_norm_pre_clip': 1.9325741529464722, 'timing/data': 0.021500439033843577, 'timing/model': 1.1967024080222473, 'learning_rate/qwen_vl_interface': 9.397699849849368e-06, 'learning_rate/action_model': 9.397699849849367e-05, 'epoch': 9.24}) +[2026-06-20 21:47:44,866][starVLA.training.train_starvla][INFO] - Step 4995, Loss: {'train/loss': 0.10110145807266235, 'train/grad_norm_pre_clip': 6.25537109375, 'timing/data': 0.01900417695287615, 'timing/model': 1.1966909150360152, 'learning_rate/qwen_vl_interface': 9.39621477448057e-06, 'learning_rate/action_model': 9.39621477448057e-05, 'epoch': 9.24}) +[2026-06-20 21:47:49,743][starVLA.training.train_starvla][INFO] - Step 4996, Loss: {'train/loss': 0.08269762247800827, 'train/grad_norm_pre_clip': 4.3516340255737305, 'timing/data': 0.02019300393294543, 'timing/model': 1.197473724023439, 'learning_rate/qwen_vl_interface': 9.394727994759495e-06, 'learning_rate/action_model': 9.394727994759495e-05, 'epoch': 9.24}) +[2026-06-20 21:47:54,593][starVLA.training.train_starvla][INFO] - Step 4997, Loss: {'train/loss': 0.02039555087685585, 'train/grad_norm_pre_clip': 5.744691371917725, 'timing/data': 0.023561088950373232, 'timing/model': 1.4504373689414933, 'learning_rate/qwen_vl_interface': 9.3932395112973e-06, 'learning_rate/action_model': 9.3932395112973e-05, 'epoch': 9.24}) +[2026-06-20 21:47:59,143][starVLA.training.train_starvla][INFO] - Step 4998, Loss: {'train/loss': 0.3233996629714966, 'train/grad_norm_pre_clip': 15.359376907348633, 'timing/data': 0.017563728033564985, 'timing/model': 1.1469495819183066, 'learning_rate/qwen_vl_interface': 9.391749324705846e-06, 'learning_rate/action_model': 9.391749324705846e-05, 'epoch': 9.24}) +[2026-06-20 21:48:03,988][starVLA.training.train_starvla][INFO] - Step 4999, Loss: {'train/loss': 0.06416626274585724, 'train/grad_norm_pre_clip': 11.74998664855957, 'timing/data': 0.022604166064411402, 'timing/model': 1.211778491968289, 'learning_rate/qwen_vl_interface': 9.390257435597692e-06, 'learning_rate/action_model': 9.390257435597692e-05, 'epoch': 9.24}) +[2026-06-20 21:48:08,840][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 5000 +[2026-06-20 21:54:38,980][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 5000 +[2026-06-20 21:54:38,981][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 5000 +[2026-06-20 22:01:33,365][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 5000 +[2026-06-20 22:01:33,367][starVLA.training.train_starvla][INFO] - Step 5000, Loss: {'train/loss': 0.0223684199154377, 'train/grad_norm_pre_clip': 2.3423521518707275, 'eval/loss': 0.7652456760406494, 'eval/latency_0/loss': 0.8967652320861816, 'eval/latency_1/loss': 1.1681649684906006, 'eval/latency_2/loss': 1.0782390832901, 'eval/latency_3/loss': 0.561813235282898, 'eval/latency_4/loss': 0.08969016373157501, 'eval/flappy/loss': 0.7652456760406494, 'eval/flappy/latency_0/loss': 0.8967652320861816, 'eval/flappy/latency_1/loss': 1.1681649684906006, 'eval/flappy/latency_2/loss': 1.0782390832901, 'eval/flappy/latency_3/loss': 0.561813235282898, 'eval/flappy/latency_4/loss': 0.08969016373157501, 'eval/flap_precision': 0.1787709497206704, 'eval/flap_recall': 0.11636363636363636, 'eval/flap_f1': 0.1409691629955947, 'eval/flap_rate_model': 0.04972222222222222, 'eval/flap_rate_teacher': 0.0763888888888889, 'eval/flap_cc_f1': 0.1409691629955947, 'eval/latency_0/flap_precision': 0.1787709497206704, 'eval/latency_0/flap_recall': 0.11636363636363636, 'eval/latency_0/flap_f1': 0.1409691629955947, 'eval/latency_0/flap_rate_model': 0.04972222222222222, 'eval/latency_0/flap_rate_teacher': 0.0763888888888889, 'eval/latency_0/flap_cc_f1': 0.1409691629955947, 'timing/data': 0.022611313033849, 'timing/model': 1.1961900279857218, 'learning_rate/qwen_vl_interface': 9.3887638445861e-06, 'learning_rate/action_model': 9.3887638445861e-05, 'epoch': 9.25}) +[2026-06-20 22:01:33,367][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_5000_state +[2026-06-20 22:01:33,371][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-20 22:01:42,011][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_5000_state/model.safetensors +[2026-06-20 22:01:57,411][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_5000_state/optimizer.bin +[2026-06-20 22:01:57,412][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_5000_state/sampler.bin +[2026-06-20 22:01:57,412][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_5000_state/sampler_1.bin +[2026-06-20 22:01:57,413][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints/steps_5000_state/random_states_0.pkl +[2026-06-20 22:02:00,040][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 5000 +[2026-06-20 22:02:00,041][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-20 22:02:00,055][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-20 22:02:00,055][starVLA.training.train_starvla][INFO] - Training complete. Final training state is saved under /workspace/starVLA/results/Checkpoints/flappy_curriculum_exclusive_40ep_per_latency/checkpoints +[2026-06-20 22:02:02,142][starVLA.training.train_starvla][INFO] - ... and that's all, folks! diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_curriculum_exclusive_40ep_per_latency/summary.jsonl b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_curriculum_exclusive_40ep_per_latency/summary.jsonl new file mode 100644 index 0000000000000000000000000000000000000000..ffa1ed16fa7cd08f0c83739d659e9d2a8bd4e827 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_curriculum_exclusive_40ep_per_latency/summary.jsonl @@ -0,0 +1,50 @@ +{"steps": 100} +{"steps": 200} +{"steps": 300} +{"steps": 400} +{"steps": 500} +{"steps": 600} +{"steps": 700} +{"steps": 800} +{"steps": 900} +{"steps": 1000} +{"steps": 1100} +{"steps": 1200} +{"steps": 1300} +{"steps": 1400} +{"steps": 1500} +{"steps": 1600} +{"steps": 1700} +{"steps": 1800} +{"steps": 1900} +{"steps": 2000} +{"steps": 2100} +{"steps": 2200} +{"steps": 2300} +{"steps": 2400} +{"steps": 2500} +{"steps": 2600} +{"steps": 2700} +{"steps": 2800} +{"steps": 2900} +{"steps": 3000} +{"steps": 3100} +{"steps": 3200} +{"steps": 3300} +{"steps": 3400} +{"steps": 3500} +{"steps": 3600} +{"steps": 3700} +{"steps": 3800} +{"steps": 3900} +{"steps": 4000} +{"steps": 4100} +{"steps": 4200} +{"steps": 4300} +{"steps": 4400} +{"steps": 4500} +{"steps": 4600} +{"steps": 4700} +{"steps": 4800} +{"steps": 4900} +{"steps": 5000} diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15/checkpoints/steps_5000_model.safetensors b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15/checkpoints/steps_5000_model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..37135bb93d1d4d0a8003ae533d71bf4d37d3883d --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15/checkpoints/steps_5000_model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:91931822871b9c3af55f58a7254963b512e478bda6d7514dd2a194ac27311cd5 +size 9138230516 diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15/config.full.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15/config.full.yaml new file mode 100644 index 0000000000000000000000000000000000000000..ddb8ec58094229108b0d4786dcb4d23dc25ad121 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15/config.full.yaml @@ -0,0 +1,254 @@ +framework: + name: QwenOFT + qwenvl: + base_vlm: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/playground/Pretrained_models/Qwen3-VL-4B-Instruct + attn_implementation: flash_attention_2 + enable_gradient_checkpointing: true + action_model: + action_model_type: MLP + action_dim: 7 + action_hidden_dim: 2560 + future_action_window_size: 0 + past_action_window_size: 0 + loss_type: discrete_ce + state_dim: 7 + action_horizon: 1 + action_env_dim: 2 + kv_memory: + enabled: false + window: 4 + rollout_len: 8 + packed_train: false + rebased_sink: true +datasets: + vla_data: + dataset_py: lerobot_datasets + include_state: true + data_root_dir: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/playground/Datasets/rl_games + data_mix: flappy_train__bridge + eval_data_mix: flappy_train__bridge__val + custom_mixtures_path: null + action_type: discrete + sequential_step_sampling: true + eval_sequential_step_sampling: null + num_workers: 8 + eval_num_workers: 8 + prefetch_factor: 4 + persistent_workers: true + pin_memory: true + shuffle: true + action_balance: + enabled: false + strategy: balanced_epoch + action_key: action_id + target_flap_fraction: 0.3 + noop_id: 0 + flap_id: 1 + latency_curriculum: + enabled: false + strategy: exclusive + latencies: null + phase_steps: null + phase_distributions: null + new_latency_passes: 1.0 + replay_passes: 0.25 + target_total_passes: 2.0 + final_equalization: true + step_budget_mode: auto + eval_at_phase_end: false + save_at_phase_end: false + computed_plan: null + per_device_batch_size: 8 + load_all_data_for_training: true + num_obs_frames: 1 + image_mode: single + prompt_mode: raw + stitch_grid: + - 2 + - 2 + obs_image_size: null + video_backend: torchvision_av +dataset: + source_hf: latency-sensitive-bench/flappy_fixed_latency_3_200ep_7k2steps_ghost15 + config_name: null + source_subdir: null + converted_name: flappy_train + single_source_hf: '' + mixed_source_hf: '' + single_converted_name: flappy_train + mixed_converted_name: flappy_mixed_latency_train + single_latency_filter: null + mixed_latency_filter: null + force_download: false + setup_force: false + skip_verification: false + target_latency_unit: observation_steps + verify_rows: 200 + max_episodes: null + episodes_per_latency: 200 + latency_filter: + - 3 + debug_subset: + enabled: false + max_episodes: 5 + suffix: debug +base_model: + repo_id: Qwen/Qwen3-VL-4B-Instruct +initialization: + checkpoint_local_dir: playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_hf_repo_id: StarVLA/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_filename: checkpoints/steps_5000_pytorch_model.pt +trainer: + max_train_steps: 5000 + num_warmup_steps: 0 + save_interval: 100 + eval_interval: 5000 + eval_num_batches: 100 + per_latency_eval_num_batches: 5 + eval_action_classification: false + eval_action_classification_interval: null + cc_f1_tolerance: 1 + learning_rate: + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + action_model: 0.0001 + lr_scheduler_type: cosine_with_min_lr + scheduler_specific_kwargs: + min_lr: 1.0e-06 + freeze_modules: '' + freeze_vit: false + freeze_tied_embedding: false + freeze_llm_layers: [] + loss_scale: + vla: 1.0 + vlm: 0.1 + max_grad_norm: 1.0 + weight_decay: 0.0 + logging_frequency: 1 + profile_timing: + enabled: false + log_interval: 10 + gradient_clipping: 1.0 + gradient_accumulation_steps: 4 + distributed_backend: none + is_resume: true + pretrained_checkpoint: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/results/Checkpoints/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15/checkpoints/steps_5000_state + resume_step: 5000 + reload_modules: null + optimizer: + name: AdamW + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + weight_decay: 1.0e-08 + fused: true + save_format: pt +workspace_dir: /lustre/fsw/portfolios/nvr/users/zihwang/latency-sensitive-bench +run_root_dir: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/results/Checkpoints +seed: 42 +wandb_entity: zihanwang-ai-northwestern-university +wandb_project: starVLA_rl_games +auth: + env_file: null + hf_token_env: HF_TOKEN + wandb_api_key_env: WANDB_API_KEY +paths: + run_root_dir: results/Checkpoints + dataset_local_dir: playground/Datasets/rl_games + dataset_cache_dir: null + base_model_dir: playground/Pretrained_models/Qwen3-VL-4B-Instruct + accelerate_config: starVLA/config/deepseeds/deepspeed_zero2.yaml +launch: + use_accelerate: true + gpus: null + num_processes: 1 + dry_run: false +conda: + enabled: true + env_name: null +rl_games: + model_alias: openvla + env_eval: + image_size: 224 + frameskip: 1 + image_transform: flappy_ghost_trail + prompt_mode: raw + ghost_trail: + history_frames: 15 + gamma: 1.3 + min_alpha: 35 + scroll_px_per_step: 4.0 + ground_fraction: 0.22 + seed: 42 + fixed_episode_seeds: true + latency_seed_stride: 0 + task_seed_stride: 0 + task_description: '' + eval_parallel_envs: 5 + action_chunk_execution: + enabled: false + chunk_size: null + enabled: true + eval_backend: eval_core + distributed_mode: none + vectorized: + enabled: false + batch_size: 1 + latency: + prompt_map_path: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/playground/Datasets/rl_games/flappy_train__bridge/latency_prompt_map.json + mode: single + values: + - 0 + mid_train: + enabled: false + interval_steps: 100 + latencies: + - 0 + num_episodes: 5 + max_steps_per_episode: 3600 + post_train: + enabled: true + latencies: + - 3 + num_episodes: 20 + max_steps_per_episode: 3600 + task: flappy + initialization_mode: bridge + action_carrier: bridge +model: openvla +env: flappy +init: bridge +bridge_base_model: + repo_id: + openvla: Qwen/Qwen3-VL-4B-Instruct + pi0: StarVLA/Qwen2.5-VL-3B-Instruct-Action + pi05: Qwen/Qwen3-VL-4B-Instruct + gr00t: Qwen/Qwen3-VL-4B-Instruct + local_dir: + openvla: playground/Pretrained_models/Qwen3-VL-4B-Instruct + pi0: playground/Pretrained_models/Qwen2.5-VL-3B-Instruct-Action + pi05: playground/Pretrained_models/Qwen3-VL-4B-Instruct + gr00t: playground/Pretrained_models/Qwen3-VL-4B-Instruct +mode: single +checkpoint: + load: auto + hf_repo_id: talha15032/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15 + save_best_model: false + save_final_model: true + save_pt_file: false + save_training_state: true + save_safetensors_file: false + local: + keep_last_n: 1 + sync: + enabled: true + repo_id: talha15032/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15 + keep_last_n: 0 + sync_every_n_checkpoints: 1 + resume_policy: local_latest +run_id: openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15 +output_dir: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/results/Checkpoints/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15 +config_yaml: null +is_debug: false +version_id: '0.21' diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15/config.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15/config.yaml new file mode 100644 index 0000000000000000000000000000000000000000..9437f53708ea9aebd198616b4cb579f66e7d4f99 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15/config.yaml @@ -0,0 +1,73 @@ +checkpoint: + local: + keep_last_n: 1 + save_best_model: false + save_final_model: true + save_pt_file: false + save_safetensors_file: false + save_training_state: true + sync: + enabled: true + keep_last_n: 0 + repo_id: talha15032/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15 +datasets: + vla_data: + data_mix: flappy_train__bridge + dataset_py: lerobot_datasets + eval_data_mix: flappy_train__bridge__val + latency_curriculum: + enabled: false + per_device_batch_size: 8 +framework: + action_model: + action_dim: 7 + action_env_dim: 2 + action_hidden_dim: 2560 + action_horizon: 1 + action_model_type: MLP + loss_type: discrete_ce + kv_memory: + enabled: false + packed_train: false + rebased_sink: true + rollout_len: 8 + window: 4 + name: QwenOFT + qwenvl: + attn_implementation: flash_attention_2 + base_vlm: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/playground/Pretrained_models/Qwen3-VL-4B-Instruct + enable_gradient_checkpointing: true +output_dir: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/results/Checkpoints/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15 +rl_games: + env_eval: + enabled: true + eval_backend: eval_core + prompt_mode: raw +run_id: openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15 +run_root_dir: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/results/Checkpoints +seed: 42 +trainer: + distributed_backend: none + freeze_llm_layers: [] + freeze_modules: '' + freeze_tied_embedding: false + freeze_vit: false + gradient_accumulation_steps: 4 + is_resume: true + learning_rate: + action_model: 0.0001 + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + lr_scheduler_type: cosine_with_min_lr + max_train_steps: 5000 + num_warmup_steps: 0 + optimizer: + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + fused: true + weight_decay: 1.0e-08 + pretrained_checkpoint: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/results/Checkpoints/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15/checkpoints/steps_5000_state + scheduler_specific_kwargs: + min_lr: 1.0e-06 diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15/dataset_statistics.json b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15/dataset_statistics.json new file mode 100644 index 0000000000000000000000000000000000000000..17bca16d4540e82f0909055ba72980988769eae1 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15/dataset_statistics.json @@ -0,0 +1,127 @@ +{ + "new_embodiment": { + "action": { + "mean": [ + 0.8039520978927612, + 0.19604788720607758, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.3938750624656677, + 0.3938750624656677, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "mask": [ + true, + true, + true, + true, + true, + true, + true + ] + }, + "state": { + "mean": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ] + }, + "num_transitions": 1396215, + "num_trajectories": 198 + } +} \ No newline at end of file diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15/dataset_statistics_eval.json b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15/dataset_statistics_eval.json new file mode 100644 index 0000000000000000000000000000000000000000..cebd0fc5f320c5ffca2a9ed1621e9a91e9d9db3f --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15/dataset_statistics_eval.json @@ -0,0 +1,127 @@ +{ + "new_embodiment": { + "action": { + "mean": [ + 0.8033333420753479, + 0.1966666728258133, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.39748135209083557, + 0.39748135209083557, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "mask": [ + true, + true, + true, + true, + true, + true, + true + ] + }, + "state": { + "mean": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ] + }, + "num_transitions": 14400, + "num_trajectories": 2 + } +} \ No newline at end of file diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15/summary.jsonl b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15/summary.jsonl new file mode 100644 index 0000000000000000000000000000000000000000..ffa1ed16fa7cd08f0c83739d659e9d2a8bd4e827 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15/summary.jsonl @@ -0,0 +1,50 @@ +{"steps": 100} +{"steps": 200} +{"steps": 300} +{"steps": 400} +{"steps": 500} +{"steps": 600} +{"steps": 700} +{"steps": 800} +{"steps": 900} +{"steps": 1000} +{"steps": 1100} +{"steps": 1200} +{"steps": 1300} +{"steps": 1400} +{"steps": 1500} +{"steps": 1600} +{"steps": 1700} +{"steps": 1800} +{"steps": 1900} +{"steps": 2000} +{"steps": 2100} +{"steps": 2200} +{"steps": 2300} +{"steps": 2400} +{"steps": 2500} +{"steps": 2600} +{"steps": 2700} +{"steps": 2800} +{"steps": 2900} +{"steps": 3000} +{"steps": 3100} +{"steps": 3200} +{"steps": 3300} +{"steps": 3400} +{"steps": 3500} +{"steps": 3600} +{"steps": 3700} +{"steps": 3800} +{"steps": 3900} +{"steps": 4000} +{"steps": 4100} +{"steps": 4200} +{"steps": 4300} +{"steps": 4400} +{"steps": 4500} +{"steps": 4600} +{"steps": 4700} +{"steps": 4800} +{"steps": 4900} +{"steps": 5000} diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15_exp2/checkpoints/steps_5000_model.safetensors b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15_exp2/checkpoints/steps_5000_model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..37135bb93d1d4d0a8003ae533d71bf4d37d3883d --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15_exp2/checkpoints/steps_5000_model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:91931822871b9c3af55f58a7254963b512e478bda6d7514dd2a194ac27311cd5 +size 9138230516 diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15_exp2/config.full.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15_exp2/config.full.yaml new file mode 100644 index 0000000000000000000000000000000000000000..a1dd33c878dca951aaa453f8fc3a0123c8e2665f --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15_exp2/config.full.yaml @@ -0,0 +1,254 @@ +framework: + name: QwenOFT + qwenvl: + base_vlm: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/playground/Pretrained_models/Qwen3-VL-4B-Instruct + attn_implementation: flash_attention_2 + enable_gradient_checkpointing: true + action_model: + action_model_type: MLP + action_dim: 7 + action_hidden_dim: 2560 + future_action_window_size: 0 + past_action_window_size: 0 + loss_type: discrete_ce + state_dim: 7 + action_horizon: 1 + action_env_dim: 2 + kv_memory: + enabled: false + window: 4 + rollout_len: 8 + packed_train: false + rebased_sink: true +datasets: + vla_data: + dataset_py: lerobot_datasets + include_state: true + data_root_dir: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/playground/Datasets/rl_games + data_mix: flappy_train__bridge + eval_data_mix: flappy_train__bridge__val + custom_mixtures_path: null + action_type: discrete + sequential_step_sampling: true + eval_sequential_step_sampling: null + num_workers: 8 + eval_num_workers: 8 + prefetch_factor: 4 + persistent_workers: true + pin_memory: true + shuffle: true + action_balance: + enabled: false + strategy: balanced_epoch + action_key: action_id + target_flap_fraction: 0.3 + noop_id: 0 + flap_id: 1 + latency_curriculum: + enabled: false + strategy: exclusive + latencies: null + phase_steps: null + phase_distributions: null + new_latency_passes: 1.0 + replay_passes: 0.25 + target_total_passes: 2.0 + final_equalization: true + step_budget_mode: auto + eval_at_phase_end: false + save_at_phase_end: false + computed_plan: null + per_device_batch_size: 8 + load_all_data_for_training: true + num_obs_frames: 1 + image_mode: single + prompt_mode: raw + stitch_grid: + - 2 + - 2 + obs_image_size: null + video_backend: torchvision_av +dataset: + source_hf: latency-sensitive-bench/flappy_fixed_latency_3_200ep_7k2steps_ghost15 + config_name: null + source_subdir: null + converted_name: flappy_train + single_source_hf: '' + mixed_source_hf: '' + single_converted_name: flappy_train + mixed_converted_name: flappy_mixed_latency_train + single_latency_filter: null + mixed_latency_filter: null + force_download: false + setup_force: false + skip_verification: false + target_latency_unit: observation_steps + verify_rows: 200 + max_episodes: null + episodes_per_latency: 200 + latency_filter: + - 3 + debug_subset: + enabled: false + max_episodes: 5 + suffix: debug +base_model: + repo_id: Qwen/Qwen3-VL-4B-Instruct +initialization: + checkpoint_local_dir: playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_hf_repo_id: StarVLA/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_filename: checkpoints/steps_5000_pytorch_model.pt +trainer: + max_train_steps: 5000 + num_warmup_steps: 0 + save_interval: 500 + eval_interval: 5000 + eval_num_batches: 100 + per_latency_eval_num_batches: 5 + eval_action_classification: false + eval_action_classification_interval: null + cc_f1_tolerance: 1 + learning_rate: + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + action_model: 0.0001 + lr_scheduler_type: cosine_with_min_lr + scheduler_specific_kwargs: + min_lr: 1.0e-06 + freeze_modules: '' + freeze_vit: false + freeze_tied_embedding: false + freeze_llm_layers: [] + loss_scale: + vla: 1.0 + vlm: 0.1 + max_grad_norm: 1.0 + weight_decay: 0.0 + logging_frequency: 1 + profile_timing: + enabled: false + log_interval: 10 + gradient_clipping: 1.0 + gradient_accumulation_steps: 4 + distributed_backend: none + is_resume: true + pretrained_checkpoint: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/results/Checkpoints/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15_exp2/checkpoints/steps_5000_state + resume_step: 5000 + reload_modules: null + optimizer: + name: AdamW + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + weight_decay: 1.0e-08 + fused: true + save_format: pt +workspace_dir: /lustre/fsw/portfolios/nvr/users/zihwang/latency-sensitive-bench +run_root_dir: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/results/Checkpoints +seed: 42 +wandb_entity: zihanwang-ai-northwestern-university +wandb_project: starVLA_rl_games +auth: + env_file: null + hf_token_env: HF_TOKEN + wandb_api_key_env: WANDB_API_KEY +paths: + run_root_dir: results/Checkpoints + dataset_local_dir: playground/Datasets/rl_games + dataset_cache_dir: null + base_model_dir: playground/Pretrained_models/Qwen3-VL-4B-Instruct + accelerate_config: starVLA/config/deepseeds/deepspeed_zero2.yaml +launch: + use_accelerate: true + gpus: null + num_processes: 1 + dry_run: false +conda: + enabled: true + env_name: null +rl_games: + model_alias: openvla + env_eval: + image_size: 224 + frameskip: 1 + image_transform: flappy_ghost_trail + prompt_mode: raw + ghost_trail: + history_frames: 15 + gamma: 1.3 + min_alpha: 35 + scroll_px_per_step: 4.0 + ground_fraction: 0.22 + seed: 42 + fixed_episode_seeds: true + latency_seed_stride: 0 + task_seed_stride: 0 + task_description: '' + eval_parallel_envs: 5 + action_chunk_execution: + enabled: false + chunk_size: null + enabled: true + eval_backend: eval_core + distributed_mode: none + vectorized: + enabled: false + batch_size: 1 + latency: + prompt_map_path: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/playground/Datasets/rl_games/flappy_train__bridge/latency_prompt_map.json + mode: single + values: + - 0 + mid_train: + enabled: false + interval_steps: 100 + latencies: + - 0 + num_episodes: 5 + max_steps_per_episode: 3600 + post_train: + enabled: true + latencies: + - 3 + num_episodes: 20 + max_steps_per_episode: 3600 + task: flappy + initialization_mode: bridge + action_carrier: bridge +model: openvla +env: flappy +init: bridge +bridge_base_model: + repo_id: + openvla: Qwen/Qwen3-VL-4B-Instruct + pi0: StarVLA/Qwen2.5-VL-3B-Instruct-Action + pi05: Qwen/Qwen3-VL-4B-Instruct + gr00t: Qwen/Qwen3-VL-4B-Instruct + local_dir: + openvla: playground/Pretrained_models/Qwen3-VL-4B-Instruct + pi0: playground/Pretrained_models/Qwen2.5-VL-3B-Instruct-Action + pi05: playground/Pretrained_models/Qwen3-VL-4B-Instruct + gr00t: playground/Pretrained_models/Qwen3-VL-4B-Instruct +mode: single +checkpoint: + load: auto + hf_repo_id: talha15032/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15_exp2 + save_best_model: false + save_final_model: true + save_pt_file: false + save_training_state: true + save_safetensors_file: false + local: + keep_last_n: 1 + sync: + enabled: true + repo_id: talha15032/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15_exp2 + keep_last_n: 0 + sync_every_n_checkpoints: 1 + resume_policy: local_latest +run_id: openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15_exp2 +output_dir: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/results/Checkpoints/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15_exp2 +config_yaml: null +is_debug: false +version_id: '0.21' diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15_exp2/config.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15_exp2/config.yaml new file mode 100644 index 0000000000000000000000000000000000000000..48f2d956234049533c3a66adb829bdee7293aa88 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15_exp2/config.yaml @@ -0,0 +1,73 @@ +checkpoint: + local: + keep_last_n: 1 + save_best_model: false + save_final_model: true + save_pt_file: false + save_safetensors_file: false + save_training_state: true + sync: + enabled: true + keep_last_n: 0 + repo_id: talha15032/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15_exp2 +datasets: + vla_data: + data_mix: flappy_train__bridge + dataset_py: lerobot_datasets + eval_data_mix: flappy_train__bridge__val + latency_curriculum: + enabled: false + per_device_batch_size: 8 +framework: + action_model: + action_dim: 7 + action_env_dim: 2 + action_hidden_dim: 2560 + action_horizon: 1 + action_model_type: MLP + loss_type: discrete_ce + kv_memory: + enabled: false + packed_train: false + rebased_sink: true + rollout_len: 8 + window: 4 + name: QwenOFT + qwenvl: + attn_implementation: flash_attention_2 + base_vlm: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/playground/Pretrained_models/Qwen3-VL-4B-Instruct + enable_gradient_checkpointing: true +output_dir: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/results/Checkpoints/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15_exp2 +rl_games: + env_eval: + enabled: true + eval_backend: eval_core + prompt_mode: raw +run_id: openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15_exp2 +run_root_dir: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/results/Checkpoints +seed: 42 +trainer: + distributed_backend: none + freeze_llm_layers: [] + freeze_modules: '' + freeze_tied_embedding: false + freeze_vit: false + gradient_accumulation_steps: 4 + is_resume: true + learning_rate: + action_model: 0.0001 + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + lr_scheduler_type: cosine_with_min_lr + max_train_steps: 5000 + num_warmup_steps: 0 + optimizer: + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + fused: true + weight_decay: 1.0e-08 + pretrained_checkpoint: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/results/Checkpoints/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15_exp2/checkpoints/steps_5000_state + scheduler_specific_kwargs: + min_lr: 1.0e-06 diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15_exp2/dataset_statistics.json b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15_exp2/dataset_statistics.json new file mode 100644 index 0000000000000000000000000000000000000000..17bca16d4540e82f0909055ba72980988769eae1 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15_exp2/dataset_statistics.json @@ -0,0 +1,127 @@ +{ + "new_embodiment": { + "action": { + "mean": [ + 0.8039520978927612, + 0.19604788720607758, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.3938750624656677, + 0.3938750624656677, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "mask": [ + true, + true, + true, + true, + true, + true, + true + ] + }, + "state": { + "mean": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ] + }, + "num_transitions": 1396215, + "num_trajectories": 198 + } +} \ No newline at end of file diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15_exp2/dataset_statistics_eval.json b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15_exp2/dataset_statistics_eval.json new file mode 100644 index 0000000000000000000000000000000000000000..cebd0fc5f320c5ffca2a9ed1621e9a91e9d9db3f --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15_exp2/dataset_statistics_eval.json @@ -0,0 +1,127 @@ +{ + "new_embodiment": { + "action": { + "mean": [ + 0.8033333420753479, + 0.1966666728258133, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.39748135209083557, + 0.39748135209083557, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "mask": [ + true, + true, + true, + true, + true, + true, + true + ] + }, + "state": { + "mean": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ] + }, + "num_transitions": 14400, + "num_trajectories": 2 + } +} \ No newline at end of file diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15_exp2/summary.jsonl b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15_exp2/summary.jsonl new file mode 100644 index 0000000000000000000000000000000000000000..a0e21985e6b7ed5b88fddcc5f74f6ec45063b603 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15_exp2/summary.jsonl @@ -0,0 +1,10 @@ +{"steps": 500} +{"steps": 1000} +{"steps": 1500} +{"steps": 2000} +{"steps": 2500} +{"steps": 3000} +{"steps": 3500} +{"steps": 4000} +{"steps": 4500} +{"steps": 5000} diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15_seed1/checkpoints/steps_5000_model.safetensors b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15_seed1/checkpoints/steps_5000_model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..37135bb93d1d4d0a8003ae533d71bf4d37d3883d --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15_seed1/checkpoints/steps_5000_model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:91931822871b9c3af55f58a7254963b512e478bda6d7514dd2a194ac27311cd5 +size 9138230516 diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15_seed1/config.full.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15_seed1/config.full.yaml new file mode 100644 index 0000000000000000000000000000000000000000..baafabf85fc93a0a0917ffc16db21c5f440db691 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15_seed1/config.full.yaml @@ -0,0 +1,254 @@ +framework: + name: QwenOFT + qwenvl: + base_vlm: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/playground/Pretrained_models/Qwen3-VL-4B-Instruct + attn_implementation: flash_attention_2 + enable_gradient_checkpointing: true + action_model: + action_model_type: MLP + action_dim: 7 + action_hidden_dim: 2560 + future_action_window_size: 0 + past_action_window_size: 0 + loss_type: discrete_ce + state_dim: 7 + action_horizon: 1 + action_env_dim: 2 + kv_memory: + enabled: false + window: 4 + rollout_len: 8 + packed_train: false + rebased_sink: true +datasets: + vla_data: + dataset_py: lerobot_datasets + include_state: true + data_root_dir: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/playground/Datasets/rl_games + data_mix: flappy_train__bridge + eval_data_mix: flappy_train__bridge__val + custom_mixtures_path: null + action_type: discrete + sequential_step_sampling: true + eval_sequential_step_sampling: null + num_workers: 8 + eval_num_workers: 8 + prefetch_factor: 4 + persistent_workers: true + pin_memory: true + shuffle: true + action_balance: + enabled: false + strategy: balanced_epoch + action_key: action_id + target_flap_fraction: 0.3 + noop_id: 0 + flap_id: 1 + latency_curriculum: + enabled: false + strategy: exclusive + latencies: null + phase_steps: null + phase_distributions: null + new_latency_passes: 1.0 + replay_passes: 0.25 + target_total_passes: 2.0 + final_equalization: true + step_budget_mode: auto + eval_at_phase_end: false + save_at_phase_end: false + computed_plan: null + per_device_batch_size: 8 + load_all_data_for_training: true + num_obs_frames: 1 + image_mode: single + prompt_mode: raw + stitch_grid: + - 2 + - 2 + obs_image_size: null + video_backend: torchvision_av +dataset: + source_hf: latency-sensitive-bench/flappy_fixed_latency_3_200ep_7k2steps_ghost15 + config_name: null + source_subdir: null + converted_name: flappy_train + single_source_hf: '' + mixed_source_hf: '' + single_converted_name: flappy_train + mixed_converted_name: flappy_mixed_latency_train + single_latency_filter: null + mixed_latency_filter: null + force_download: false + setup_force: false + skip_verification: false + target_latency_unit: observation_steps + verify_rows: 200 + max_episodes: null + episodes_per_latency: 200 + latency_filter: + - 3 + debug_subset: + enabled: false + max_episodes: 5 + suffix: debug +base_model: + repo_id: Qwen/Qwen3-VL-4B-Instruct +initialization: + checkpoint_local_dir: playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_hf_repo_id: StarVLA/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_filename: checkpoints/steps_5000_pytorch_model.pt +trainer: + max_train_steps: 5000 + num_warmup_steps: 0 + save_interval: 500 + eval_interval: 5000 + eval_num_batches: 100 + per_latency_eval_num_batches: 5 + eval_action_classification: false + eval_action_classification_interval: null + cc_f1_tolerance: 1 + learning_rate: + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + action_model: 0.0001 + lr_scheduler_type: cosine_with_min_lr + scheduler_specific_kwargs: + min_lr: 1.0e-06 + freeze_modules: '' + freeze_vit: false + freeze_tied_embedding: false + freeze_llm_layers: [] + loss_scale: + vla: 1.0 + vlm: 0.1 + max_grad_norm: 1.0 + weight_decay: 0.0 + logging_frequency: 1 + profile_timing: + enabled: false + log_interval: 10 + gradient_clipping: 1.0 + gradient_accumulation_steps: 4 + distributed_backend: none + is_resume: true + pretrained_checkpoint: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/results/Checkpoints/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15_seed1/checkpoints/steps_5000_state + resume_step: 5000 + reload_modules: null + optimizer: + name: AdamW + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + weight_decay: 1.0e-08 + fused: true + save_format: pt +workspace_dir: /lustre/fsw/portfolios/nvr/users/zihwang/latency-sensitive-bench +run_root_dir: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/results/Checkpoints +seed: 1 +wandb_entity: zihanwang-ai-northwestern-university +wandb_project: starVLA_rl_games +auth: + env_file: null + hf_token_env: HF_TOKEN + wandb_api_key_env: WANDB_API_KEY +paths: + run_root_dir: results/Checkpoints + dataset_local_dir: playground/Datasets/rl_games + dataset_cache_dir: null + base_model_dir: playground/Pretrained_models/Qwen3-VL-4B-Instruct + accelerate_config: starVLA/config/deepseeds/deepspeed_zero2.yaml +launch: + use_accelerate: true + gpus: null + num_processes: 1 + dry_run: false +conda: + enabled: true + env_name: null +rl_games: + model_alias: openvla + env_eval: + image_size: 224 + frameskip: 1 + image_transform: flappy_ghost_trail + prompt_mode: raw + ghost_trail: + history_frames: 15 + gamma: 1.3 + min_alpha: 35 + scroll_px_per_step: 4.0 + ground_fraction: 0.22 + seed: 1 + fixed_episode_seeds: true + latency_seed_stride: 0 + task_seed_stride: 0 + task_description: '' + eval_parallel_envs: 5 + action_chunk_execution: + enabled: false + chunk_size: null + enabled: true + eval_backend: eval_core + distributed_mode: none + vectorized: + enabled: false + batch_size: 1 + latency: + prompt_map_path: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/playground/Datasets/rl_games/flappy_train__bridge/latency_prompt_map.json + mode: single + values: + - 0 + mid_train: + enabled: false + interval_steps: 100 + latencies: + - 0 + num_episodes: 5 + max_steps_per_episode: 3600 + post_train: + enabled: true + latencies: + - 3 + num_episodes: 20 + max_steps_per_episode: 3600 + task: flappy + initialization_mode: bridge + action_carrier: bridge +model: openvla +env: flappy +init: bridge +bridge_base_model: + repo_id: + openvla: Qwen/Qwen3-VL-4B-Instruct + pi0: StarVLA/Qwen2.5-VL-3B-Instruct-Action + pi05: Qwen/Qwen3-VL-4B-Instruct + gr00t: Qwen/Qwen3-VL-4B-Instruct + local_dir: + openvla: playground/Pretrained_models/Qwen3-VL-4B-Instruct + pi0: playground/Pretrained_models/Qwen2.5-VL-3B-Instruct-Action + pi05: playground/Pretrained_models/Qwen3-VL-4B-Instruct + gr00t: playground/Pretrained_models/Qwen3-VL-4B-Instruct +mode: single +checkpoint: + load: auto + hf_repo_id: talha15032/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15_seed1 + save_best_model: false + save_final_model: true + save_pt_file: false + save_training_state: true + save_safetensors_file: false + local: + keep_last_n: 1 + sync: + enabled: false + repo_id: talha15032/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15_seed1 + keep_last_n: 0 + sync_every_n_checkpoints: 1 + resume_policy: local_latest +run_id: openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15_seed1 +output_dir: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/results/Checkpoints/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15_seed1 +config_yaml: null +is_debug: false +version_id: '0.21' diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15_seed1/config.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15_seed1/config.yaml new file mode 100644 index 0000000000000000000000000000000000000000..895316a83f3cc2cb6d655651f742ab6b39bd54ee --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15_seed1/config.yaml @@ -0,0 +1,73 @@ +checkpoint: + local: + keep_last_n: 1 + save_best_model: false + save_final_model: true + save_pt_file: false + save_safetensors_file: false + save_training_state: true + sync: + enabled: false + keep_last_n: 0 + repo_id: talha15032/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15_seed1 +datasets: + vla_data: + data_mix: flappy_train__bridge + dataset_py: lerobot_datasets + eval_data_mix: flappy_train__bridge__val + latency_curriculum: + enabled: false + per_device_batch_size: 8 +framework: + action_model: + action_dim: 7 + action_env_dim: 2 + action_hidden_dim: 2560 + action_horizon: 1 + action_model_type: MLP + loss_type: discrete_ce + kv_memory: + enabled: false + packed_train: false + rebased_sink: true + rollout_len: 8 + window: 4 + name: QwenOFT + qwenvl: + attn_implementation: flash_attention_2 + base_vlm: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/playground/Pretrained_models/Qwen3-VL-4B-Instruct + enable_gradient_checkpointing: true +output_dir: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/results/Checkpoints/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15_seed1 +rl_games: + env_eval: + enabled: true + eval_backend: eval_core + prompt_mode: raw +run_id: openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15_seed1 +run_root_dir: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/results/Checkpoints +seed: 1 +trainer: + distributed_backend: none + freeze_llm_layers: [] + freeze_modules: '' + freeze_tied_embedding: false + freeze_vit: false + gradient_accumulation_steps: 4 + is_resume: true + learning_rate: + action_model: 0.0001 + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + lr_scheduler_type: cosine_with_min_lr + max_train_steps: 5000 + num_warmup_steps: 0 + optimizer: + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + fused: true + weight_decay: 1.0e-08 + pretrained_checkpoint: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/results/Checkpoints/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15_seed1/checkpoints/steps_5000_state + scheduler_specific_kwargs: + min_lr: 1.0e-06 diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15_seed1/dataset_statistics.json b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15_seed1/dataset_statistics.json new file mode 100644 index 0000000000000000000000000000000000000000..17bca16d4540e82f0909055ba72980988769eae1 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15_seed1/dataset_statistics.json @@ -0,0 +1,127 @@ +{ + "new_embodiment": { + "action": { + "mean": [ + 0.8039520978927612, + 0.19604788720607758, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.3938750624656677, + 0.3938750624656677, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "mask": [ + true, + true, + true, + true, + true, + true, + true + ] + }, + "state": { + "mean": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ] + }, + "num_transitions": 1396215, + "num_trajectories": 198 + } +} \ No newline at end of file diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15_seed1/dataset_statistics_eval.json b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15_seed1/dataset_statistics_eval.json new file mode 100644 index 0000000000000000000000000000000000000000..cebd0fc5f320c5ffca2a9ed1621e9a91e9d9db3f --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15_seed1/dataset_statistics_eval.json @@ -0,0 +1,127 @@ +{ + "new_embodiment": { + "action": { + "mean": [ + 0.8033333420753479, + 0.1966666728258133, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.39748135209083557, + 0.39748135209083557, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "mask": [ + true, + true, + true, + true, + true, + true, + true + ] + }, + "state": { + "mean": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ] + }, + "num_transitions": 14400, + "num_trajectories": 2 + } +} \ No newline at end of file diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15_seed1/summary.jsonl b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15_seed1/summary.jsonl new file mode 100644 index 0000000000000000000000000000000000000000..a0e21985e6b7ed5b88fddcc5f74f6ec45063b603 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost15_seed1/summary.jsonl @@ -0,0 +1,10 @@ +{"steps": 500} +{"steps": 1000} +{"steps": 1500} +{"steps": 2000} +{"steps": 2500} +{"steps": 3000} +{"steps": 3500} +{"steps": 4000} +{"steps": 4500} +{"steps": 5000} diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost7_exp3/checkpoints/steps_4000_model.safetensors b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost7_exp3/checkpoints/steps_4000_model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..60e22beab76b1320fc118c7ac6b0509e1968535b --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost7_exp3/checkpoints/steps_4000_model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:553d47011aa16cef4609610ea44342630fb4b0b7aabc65da5fe58d133915f40a +size 9784896958 diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost7_exp3/config.full.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost7_exp3/config.full.yaml new file mode 100644 index 0000000000000000000000000000000000000000..02ff1207840da2c240db3ed0351142bab7d7952b --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost7_exp3/config.full.yaml @@ -0,0 +1,254 @@ +framework: + name: QwenOFT + qwenvl: + base_vlm: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/playground/Pretrained_models/Qwen3-VL-4B-Instruct + attn_implementation: flash_attention_2 + enable_gradient_checkpointing: true + action_model: + action_model_type: MLP + action_dim: 7 + action_hidden_dim: 2560 + future_action_window_size: 0 + past_action_window_size: 0 + loss_type: discrete_ce + state_dim: 7 + action_horizon: 1 + action_env_dim: 2 + kv_memory: + enabled: false + window: 4 + rollout_len: 8 + packed_train: false + rebased_sink: true +datasets: + vla_data: + dataset_py: lerobot_datasets + include_state: true + data_root_dir: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/playground/Datasets/rl_games + data_mix: flappy_train__bridge + eval_data_mix: flappy_train__bridge__val + custom_mixtures_path: null + action_type: discrete + sequential_step_sampling: true + eval_sequential_step_sampling: null + num_workers: 8 + eval_num_workers: 0 + prefetch_factor: 4 + persistent_workers: true + pin_memory: true + shuffle: true + action_balance: + enabled: false + strategy: balanced_epoch + action_key: action_id + target_flap_fraction: 0.3 + noop_id: 0 + flap_id: 1 + latency_curriculum: + enabled: false + strategy: exclusive + latencies: null + phase_steps: null + phase_distributions: null + new_latency_passes: 1.0 + replay_passes: 0.25 + target_total_passes: 2.0 + final_equalization: true + step_budget_mode: auto + eval_at_phase_end: false + save_at_phase_end: false + computed_plan: null + per_device_batch_size: 8 + load_all_data_for_training: true + num_obs_frames: 1 + image_mode: single + prompt_mode: raw + stitch_grid: + - 2 + - 2 + obs_image_size: null + video_backend: torchvision_av +dataset: + source_hf: latency-sensitive-bench/memory-rollouts + config_name: null + source_subdir: flappy_fixed_latency_3_200ep_7k2steps_ghost7 + converted_name: flappy_train + single_source_hf: '' + mixed_source_hf: '' + single_converted_name: flappy_train + mixed_converted_name: flappy_mixed_latency_train + single_latency_filter: null + mixed_latency_filter: null + force_download: false + setup_force: false + skip_verification: false + target_latency_unit: observation_steps + verify_rows: 200 + max_episodes: null + episodes_per_latency: 200 + latency_filter: + - 3 + debug_subset: + enabled: false + max_episodes: 5 + suffix: debug +base_model: + repo_id: Qwen/Qwen3-VL-4B-Instruct +initialization: + checkpoint_local_dir: playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_hf_repo_id: StarVLA/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_filename: checkpoints/steps_5000_pytorch_model.pt +trainer: + max_train_steps: 4000 + num_warmup_steps: 0 + save_interval: 500 + eval_interval: 4000 + eval_num_batches: 100 + per_latency_eval_num_batches: 5 + eval_action_classification: false + eval_action_classification_interval: null + cc_f1_tolerance: 1 + learning_rate: + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + action_model: 0.0001 + lr_scheduler_type: cosine_with_min_lr + scheduler_specific_kwargs: + min_lr: 1.0e-06 + freeze_modules: '' + freeze_vit: false + freeze_tied_embedding: false + freeze_llm_layers: [] + loss_scale: + vla: 1.0 + vlm: 0.1 + max_grad_norm: 1.0 + weight_decay: 0.0 + logging_frequency: 1 + profile_timing: + enabled: false + log_interval: 10 + gradient_clipping: 1.0 + gradient_accumulation_steps: 16 + distributed_backend: deepspeed + is_resume: true + pretrained_checkpoint: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/results/Checkpoints/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost7_exp3/checkpoints/steps_4000_state + resume_step: 4000 + reload_modules: null + optimizer: + name: AdamW + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + weight_decay: 1.0e-08 + fused: true + save_format: pt +workspace_dir: /lustre/fsw/portfolios/nvr/users/zihwang/latency-sensitive-bench +run_root_dir: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/results/Checkpoints +seed: 42 +wandb_entity: zihanwang-ai-northwestern-university +wandb_project: starVLA_rl_games +auth: + env_file: null + hf_token_env: HF_TOKEN + wandb_api_key_env: WANDB_API_KEY +paths: + run_root_dir: results/Checkpoints + dataset_local_dir: playground/Datasets/rl_games + dataset_cache_dir: null + base_model_dir: playground/Pretrained_models/Qwen3-VL-4B-Instruct + accelerate_config: starVLA/config/deepseeds/deepspeed_zero2.yaml +launch: + use_accelerate: true + gpus: null + num_processes: 1 + dry_run: false +conda: + enabled: true + env_name: null +rl_games: + model_alias: openvla + env_eval: + image_size: 224 + frameskip: 1 + image_transform: flappy_ghost_trail + prompt_mode: raw + ghost_trail: + history_frames: 7 + gamma: 1.3 + min_alpha: 35 + scroll_px_per_step: 4.0 + ground_fraction: 0.22 + seed: 42 + fixed_episode_seeds: true + latency_seed_stride: 0 + task_seed_stride: 0 + task_description: '' + eval_parallel_envs: 5 + action_chunk_execution: + enabled: false + chunk_size: null + enabled: true + eval_backend: latency_bench + distributed_mode: rank_sharded + vectorized: + enabled: false + batch_size: 1 + latency: + prompt_map_path: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/playground/Datasets/rl_games/flappy_train__bridge/latency_prompt_map.json + mode: single + values: + - 0 + mid_train: + enabled: false + interval_steps: 100 + latencies: + - 0 + num_episodes: 5 + max_steps_per_episode: 3600 + post_train: + enabled: true + latencies: + - 3 + num_episodes: 100 + max_steps_per_episode: 3600 + task: flappy + initialization_mode: bridge + action_carrier: bridge +model: openvla +env: flappy +init: bridge +bridge_base_model: + repo_id: + openvla: Qwen/Qwen3-VL-4B-Instruct + pi0: StarVLA/Qwen2.5-VL-3B-Instruct-Action + pi05: Qwen/Qwen3-VL-4B-Instruct + gr00t: Qwen/Qwen3-VL-4B-Instruct + local_dir: + openvla: playground/Pretrained_models/Qwen3-VL-4B-Instruct + pi0: playground/Pretrained_models/Qwen2.5-VL-3B-Instruct-Action + pi05: playground/Pretrained_models/Qwen3-VL-4B-Instruct + gr00t: playground/Pretrained_models/Qwen3-VL-4B-Instruct +mode: single +checkpoint: + load: auto + hf_repo_id: talha15032/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost7_exp3 + save_best_model: false + save_final_model: true + save_pt_file: false + save_training_state: true + save_safetensors_file: false + local: + keep_last_n: 1 + sync: + enabled: false + repo_id: talha15032/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost7_exp3 + keep_last_n: 0 + sync_every_n_checkpoints: 1 + resume_policy: local_latest +run_id: openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost7_exp3 +output_dir: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/results/Checkpoints/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost7_exp3 +config_yaml: null +is_debug: false +version_id: '0.21' diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost7_exp3/config.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost7_exp3/config.yaml new file mode 100644 index 0000000000000000000000000000000000000000..e5f12921fa691770bc26f4dcd300fc3dfb290bd9 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost7_exp3/config.yaml @@ -0,0 +1,73 @@ +checkpoint: + local: + keep_last_n: 1 + save_best_model: false + save_final_model: true + save_pt_file: false + save_safetensors_file: false + save_training_state: true + sync: + enabled: false + keep_last_n: 0 + repo_id: talha15032/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost7_exp3 +datasets: + vla_data: + data_mix: flappy_train__bridge + dataset_py: lerobot_datasets + eval_data_mix: flappy_train__bridge__val + latency_curriculum: + enabled: false + per_device_batch_size: 8 +framework: + action_model: + action_dim: 7 + action_env_dim: 2 + action_hidden_dim: 2560 + action_horizon: 1 + action_model_type: MLP + loss_type: discrete_ce + kv_memory: + enabled: false + packed_train: false + rebased_sink: true + rollout_len: 8 + window: 4 + name: QwenOFT + qwenvl: + attn_implementation: flash_attention_2 + base_vlm: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/playground/Pretrained_models/Qwen3-VL-4B-Instruct + enable_gradient_checkpointing: true +output_dir: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/results/Checkpoints/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost7_exp3 +rl_games: + env_eval: + enabled: true + eval_backend: latency_bench + prompt_mode: raw +run_id: openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost7_exp3 +run_root_dir: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/results/Checkpoints +seed: 42 +trainer: + distributed_backend: deepspeed + freeze_llm_layers: [] + freeze_modules: '' + freeze_tied_embedding: false + freeze_vit: false + gradient_accumulation_steps: 16 + is_resume: true + learning_rate: + action_model: 0.0001 + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + lr_scheduler_type: cosine_with_min_lr + max_train_steps: 4000 + num_warmup_steps: 0 + optimizer: + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + fused: true + weight_decay: 1.0e-08 + pretrained_checkpoint: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/results/Checkpoints/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost7_exp3/checkpoints/steps_4000_state + scheduler_specific_kwargs: + min_lr: 1.0e-06 diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost7_exp3/dataset_statistics.json b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost7_exp3/dataset_statistics.json new file mode 100644 index 0000000000000000000000000000000000000000..e5a582890ec84b8c80351f23a5e0839f9300a755 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost7_exp3/dataset_statistics.json @@ -0,0 +1,127 @@ +{ + "new_embodiment": { + "action": { + "mean": [ + 0.8235060572624207, + 0.17649392783641815, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.3809901773929596, + 0.3809901773929596, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "mask": [ + true, + true, + true, + true, + true, + true, + true + ] + }, + "state": { + "mean": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ] + }, + "num_transitions": 1393742, + "num_trajectories": 198 + } +} \ No newline at end of file diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost7_exp3/dataset_statistics_eval.json b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost7_exp3/dataset_statistics_eval.json new file mode 100644 index 0000000000000000000000000000000000000000..10b466b71241f0381f1789828d788caffc5a0386 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost7_exp3/dataset_statistics_eval.json @@ -0,0 +1,127 @@ +{ + "new_embodiment": { + "action": { + "mean": [ + 0.8237500190734863, + 0.17624999582767487, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.3810042142868042, + 0.3810042142868042, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "mask": [ + true, + true, + true, + true, + true, + true, + true + ] + }, + "state": { + "mean": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ] + }, + "num_transitions": 14400, + "num_trajectories": 2 + } +} \ No newline at end of file diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost7_exp3/summary.jsonl b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost7_exp3/summary.jsonl new file mode 100644 index 0000000000000000000000000000000000000000..ce7dad56d7abf845dd57d6f02d76df86c6434c49 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_bridge_flappy_fixed_latency_3_200ep_7k2steps_ghost7_exp3/summary.jsonl @@ -0,0 +1,8 @@ +{"steps": 500} +{"steps": 1000} +{"steps": 1500} +{"steps": 2000} +{"steps": 2500} +{"steps": 3000} +{"steps": 3500} +{"steps": 4000} diff --git a/latency-aware/flappy/vla/starvla-wanoft-h8/wan_oft_flappy_fix_latency_3_context5_standard_sft_2000_effbs128_224_currentce/checkpoints/steps_2000_pytorch_model.pt b/latency-aware/flappy/vla/starvla-wanoft-h8/wan_oft_flappy_fix_latency_3_context5_standard_sft_2000_effbs128_224_currentce/checkpoints/steps_2000_pytorch_model.pt new file mode 100644 index 0000000000000000000000000000000000000000..9c1d22f25f129fdb81262ae2c9c3734c3a4fa254 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-wanoft-h8/wan_oft_flappy_fix_latency_3_context5_standard_sft_2000_effbs128_224_currentce/checkpoints/steps_2000_pytorch_model.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:898e0b3e19c6d07987be834b503b315688dda724ffa25e5642c16c1cb316c660 +size 25211531190 diff --git a/latency-aware/flappy/vla/starvla-wanoft-h8/wan_oft_flappy_fix_latency_3_context5_standard_sft_2000_effbs128_224_currentce/config.full.yaml b/latency-aware/flappy/vla/starvla-wanoft-h8/wan_oft_flappy_fix_latency_3_context5_standard_sft_2000_effbs128_224_currentce/config.full.yaml new file mode 100644 index 0000000000000000000000000000000000000000..778c88f4812563c011863c2f8f56202d873c3703 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-wanoft-h8/wan_oft_flappy_fix_latency_3_context5_standard_sft_2000_effbs128_224_currentce/config.full.yaml @@ -0,0 +1,277 @@ +framework: + name: WanOFT + world_model: + base_wm: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/wanoft_bench/third_party/starVLA/playground/Pretrained_models/Wan-AI/Wan2.2-TI2V-5B-Diffusers + extract_layers: + - -1 + num_frames: 5 + qwenvl: + base_vlm: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/wanoft_bench/third_party/starVLA/playground/Pretrained_models/Wan-AI/Wan2.2-TI2V-5B-Diffusers + attn_implementation: sdpa + flex_backend: triton + enable_gradient_checkpointing: true + action_model: + action_model_type: MLP + action_dim: 7 + action_hidden_dim: 3072 + action_horizon: 8 + future_action_window_size: 7 + past_action_window_size: 0 + loss_type: current_discrete_ce + class_weights: null + future_loss_weight: null + action_query_source: mean + action_query_num_heads: 24 + state_dim: 7 + action_env_dim: 2 + kv_memory: + enabled: false + window: 4 + rollout_len: 8 + packed_train: false + rebased_sink: true +datasets: + vla_data: + dataset_py: lerobot_datasets + include_state: true + data_root_dir: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/wanoft_bench/third_party/starVLA/data/flappy_fix_latency_3_200ep_context5 + data_mix: flappy_train__bridge + eval_data_mix: null + custom_mixtures_path: null + action_type: discrete + sequential_step_sampling: false + eval_sequential_step_sampling: null + num_workers: 8 + eval_num_workers: 8 + prefetch_factor: 4 + persistent_workers: true + pin_memory: true + shuffle: true + action_balance: + enabled: false + strategy: balanced_epoch + action_key: action_id + target_flap_fraction: 0.3 + noop_id: 0 + flap_id: 1 + latency_curriculum: + enabled: false + strategy: exclusive + latencies: null + phase_steps: null + phase_distributions: null + new_latency_passes: 1.0 + replay_passes: 0.25 + target_total_passes: 2.0 + final_equalization: true + step_budget_mode: auto + eval_at_phase_end: false + save_at_phase_end: false + computed_plan: null + per_device_batch_size: 1 + load_all_data_for_training: true + num_obs_frames: 1 + image_mode: single + prompt_mode: raw + stitch_grid: + - 2 + - 2 + obs_image_size: + - 224 + - 224 + video_backend: torchvision_av + observation_indices: + - -4 + - -3 + - -2 + - -1 + - 0 + language_indices: + - 0 + state_indices: + - 0 + action_indices: + - 0 + - 1 + - 2 + - 3 + - 4 + - 5 + - 6 + - 7 + pack_image_sequence: true + image_sequence_length: 5 + context_images_column: observation.context_images +dataset: + source_hf: latency-sensitive-bench/memory-data + config_name: flappy_fixed_latency_3_200ep_7k2steps + source_subdir: null + converted_name: flappy_train__bridge + single_source_hf: '' + mixed_source_hf: '' + single_converted_name: flappy_train__bridge + mixed_converted_name: flappy_mixed_latency_train__bridge + single_latency_filter: null + mixed_latency_filter: null + force_download: false + setup_force: false + skip_verification: false + target_latency_unit: observation_steps + verify_rows: 200 + max_episodes: 200 + episodes_per_latency: null + latency_filter: null + debug_subset: + enabled: false + max_episodes: 5 + suffix: debug +base_model: + repo_id: Wan-AI/Wan2.2-TI2V-5B-Diffusers +initialization: + checkpoint_local_dir: playground/Pretrained_models/WM4A-Wan2d2-OFT-LIBERO-4in1 + checkpoint_hf_repo_id: StarVLA/WM4A-Wan2d2-OFT-LIBERO-4in1 + checkpoint_filename: checkpoints/steps_60000_pytorch_model.pt +trainer: + max_train_steps: 2000 + num_warmup_steps: 0 + save_interval: 500 + eval_interval: 100 + eval_num_batches: 100 + per_latency_eval_num_batches: null + eval_action_classification: false + eval_action_classification_interval: null + cc_f1_tolerance: 1 + learning_rate: + base: 2.5e-05 + qwen_vl_interface: 1.0e-05 + action_model: 0.0001 + action_query_proj: 0.0001 + lr_scheduler_type: cosine_with_min_lr + scheduler_specific_kwargs: + min_lr: 1.0e-06 + freeze_modules: '' + freeze_vit: false + freeze_tied_embedding: false + freeze_llm_layers: [] + loss_scale: + vla: 1.0 + vlm: 0.1 + max_grad_norm: 1.0 + weight_decay: 0.0 + logging_frequency: 1 + profile_timing: + enabled: false + log_interval: 10 + gradient_clipping: 1.0 + gradient_accumulation_steps: 16 + distributed_backend: deepspeed + is_resume: false + pretrained_checkpoint: null + resume_step: 0 + reload_modules: null + optimizer: + name: AdamW + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + weight_decay: 1.0e-08 + fused: true + save_format: pt +workspace_dir: WORKSPACE_DIR +run_root_dir: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/wanoft_bench/third_party/starVLA/results/Checkpoints +seed: 42 +wandb_entity: zihanwang-ai-northwestern-university +wandb_project: starVLA_rl_games +auth: + env_file: null + hf_token_env: HF_TOKEN + wandb_api_key_env: WANDB_API_KEY +paths: + run_root_dir: results/Checkpoints + dataset_local_dir: data/flappy_fix_latency_3_200ep_context5 + dataset_cache_dir: null + base_model_dir: playground/Pretrained_models/Wan-AI/Wan2.2-TI2V-5B-Diffusers + accelerate_config: starVLA/config/deepseeds/deepspeed_zero2.yaml +launch: + use_accelerate: true + gpus: null + num_processes: 1 + dry_run: false +conda: + enabled: true + env_name: null +rl_games: + model_alias: wan_oft + env_eval: + image_size: 224 + frameskip: 1 + image_transform: raw_rgb + prompt_mode: raw + ghost_trail: + history_frames: 5 + gamma: 1.3 + min_alpha: 35 + scroll_px_per_step: 4.0 + ground_fraction: 0.22 + seed: 42 + fixed_episode_seeds: true + latency_seed_stride: 0 + task_seed_stride: 0 + task_description: '' + eval_parallel_envs: 5 + action_chunk_execution: + enabled: false + chunk_size: null + enabled: true + eval_backend: eval_core + distributed_mode: rank_sharded + vectorized: + enabled: false + batch_size: 1 + latency: + prompt_map_path: data/flappy_fix_latency_3_200ep_context5/flappy_train__bridge/latency_prompt_map.json + mode: single + values: + - 0 + mid_train: + enabled: false + interval_steps: 100 + latencies: + - 0 + num_episodes: 5 + max_steps_per_episode: 3600 + post_train: + enabled: true + latencies: + - 3 + num_episodes: 20 + max_steps_per_episode: 3600 + task: flappy + initialization_mode: bridge + action_carrier: bridge +model: wan_oft +env: flappy +init: wan_oft_libero +mode: single +checkpoint: + load: auto + hf_repo_id: null + save_best_model: false + save_final_model: true + save_pt_file: true + save_training_state: true + save_safetensors_file: false + local: + keep_last_n: 1 + sync: + enabled: false + repo_id: null + keep_last_n: 0 + sync_every_n_checkpoints: 1 + resume_policy: local_latest +run_id: wan_oft_flappy_fix_latency_3_context5_standard_sft_2000_effbs128_224_currentce +output_dir: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/wanoft_bench/third_party/starVLA/results/Checkpoints/wan_oft_flappy_fix_latency_3_context5_standard_sft_2000_effbs128_224_currentce +config_yaml: null +is_debug: false +version_id: '0.21' diff --git a/latency-aware/flappy/vla/starvla-wanoft-h8/wan_oft_flappy_fix_latency_3_context5_standard_sft_2000_effbs128_224_currentce/config.yaml b/latency-aware/flappy/vla/starvla-wanoft-h8/wan_oft_flappy_fix_latency_3_context5_standard_sft_2000_effbs128_224_currentce/config.yaml new file mode 100644 index 0000000000000000000000000000000000000000..1ba08acfcf5c22f73ad8f39ac63a58886b5a6e3b --- /dev/null +++ b/latency-aware/flappy/vla/starvla-wanoft-h8/wan_oft_flappy_fix_latency_3_context5_standard_sft_2000_effbs128_224_currentce/config.yaml @@ -0,0 +1,89 @@ +checkpoint: + local: + keep_last_n: 1 + save_best_model: false + save_final_model: true + save_pt_file: true + save_safetensors_file: false + save_training_state: true + sync: + enabled: false + keep_last_n: 0 + repo_id: null +datasets: + vla_data: + data_mix: flappy_train__bridge + dataset_py: lerobot_datasets + eval_data_mix: null + latency_curriculum: + enabled: false + obs_image_size: + - 224 + - 224 + per_device_batch_size: 1 +framework: + action_model: + action_dim: 7 + action_env_dim: 2 + action_hidden_dim: 3072 + action_horizon: 8 + action_model_type: MLP + action_query_source: mean + class_weights: null + loss_type: current_discrete_ce + name: WanOFT + qwenvl: + base_vlm: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/wanoft_bench/third_party/starVLA/playground/Pretrained_models/Wan-AI/Wan2.2-TI2V-5B-Diffusers + world_model: + base_wm: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/wanoft_bench/third_party/starVLA/playground/Pretrained_models/Wan-AI/Wan2.2-TI2V-5B-Diffusers + extract_layers: + - -1 + num_frames: 5 +output_dir: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/wanoft_bench/third_party/starVLA/results/Checkpoints/wan_oft_flappy_fix_latency_3_context5_standard_sft_2000_effbs128_224_currentce +rl_games: + env_eval: + enabled: true + eval_backend: eval_core + mid_train: + enabled: false + interval_steps: 100 + prompt_mode: raw +run_id: wan_oft_flappy_fix_latency_3_context5_standard_sft_2000_effbs128_224_currentce +run_root_dir: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/wanoft_bench/third_party/starVLA/results/Checkpoints +seed: 42 +trainer: + distributed_backend: deepspeed + eval_action_classification: false + eval_action_classification_interval: null + eval_interval: 100 + freeze_llm_layers: [] + freeze_modules: '' + freeze_tied_embedding: false + freeze_vit: false + gradient_accumulation_steps: 16 + is_resume: false + learning_rate: + action_model: 0.0001 + action_query_proj: 0.0001 + base: 2.5e-05 + qwen_vl_interface: 1.0e-05 + logging_frequency: 1 + lr_scheduler_type: cosine_with_min_lr + max_train_steps: 2000 + num_warmup_steps: 0 + optimizer: + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + fused: true + weight_decay: 1.0e-08 + pretrained_checkpoint: null + profile_timing: + enabled: false + reload_modules: null + save_interval: 500 + scheduler_specific_kwargs: + min_lr: 1.0e-06 +wandb_entity: zihanwang-ai-northwestern-university +wandb_project: starVLA_rl_games diff --git a/latency-aware/flappy/vla/starvla-wanoft-h8/wan_oft_flappy_fix_latency_3_context5_standard_sft_2000_effbs128_224_currentce/dataset_statistics.json b/latency-aware/flappy/vla/starvla-wanoft-h8/wan_oft_flappy_fix_latency_3_context5_standard_sft_2000_effbs128_224_currentce/dataset_statistics.json new file mode 100644 index 0000000000000000000000000000000000000000..e5a582890ec84b8c80351f23a5e0839f9300a755 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-wanoft-h8/wan_oft_flappy_fix_latency_3_context5_standard_sft_2000_effbs128_224_currentce/dataset_statistics.json @@ -0,0 +1,127 @@ +{ + "new_embodiment": { + "action": { + "mean": [ + 0.8235060572624207, + 0.17649392783641815, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.3809901773929596, + 0.3809901773929596, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "mask": [ + true, + true, + true, + true, + true, + true, + true + ] + }, + "state": { + "mean": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ] + }, + "num_transitions": 1393742, + "num_trajectories": 198 + } +} \ No newline at end of file diff --git a/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2/checkpoints/steps_5000_pytorch_model.pt b/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2/checkpoints/steps_5000_pytorch_model.pt new file mode 100644 index 0000000000000000000000000000000000000000..de7b3132e15b8aa88946b9f49c6df3b03c877840 --- /dev/null +++ b/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2/checkpoints/steps_5000_pytorch_model.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:28b6ca56206555b081f481247649ac06762debcb7cf74776b2ca240338d14974 +size 9138435024 diff --git a/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2/config.full.yaml b/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2/config.full.yaml new file mode 100644 index 0000000000000000000000000000000000000000..43096cec882e79f1776b99d8c1792082685bb20b --- /dev/null +++ b/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2/config.full.yaml @@ -0,0 +1,338 @@ +framework: + name: QwenOFT + qwenvl: + base_vlm: /home/ubuntu/talha/playground/Pretrained_models/Qwen3-VL-4B-Instruct + attn_implementation: flash_attention_2 + enable_gradient_checkpointing: true + action_model: + action_model_type: MLP + action_dim: 5 + state_dim: 4 + action_hidden_dim: 2560 + future_action_window_size: 0 + past_action_window_size: 0 + loss_type: discrete_ce + state_encoding: discretized_text + action_horizon: 1 + action_env_dim: 5 + kv_memory: + enabled: false + window: 4 + rollout_len: 8 + packed_train: false + rebased_sink: true +datasets: + vla_data: + dataset_py: lerobot_datasets + include_state: true + data_root_dir: /home/ubuntu/talha/playground/Datasets/rl_games + data_mix: openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2 + eval_data_mix: openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2__val + custom_mixtures_path: /home/ubuntu/talha/playground/Datasets/rl_games/_generated_mixtures/single_77a677bfbf8e73b8.json + action_type: discrete + sequential_step_sampling: false + eval_sequential_step_sampling: null + num_workers: 8 + eval_num_workers: 8 + prefetch_factor: 4 + persistent_workers: true + pin_memory: true + shuffle: true + action_balance: + enabled: false + strategy: balanced_epoch + action_key: action_id + target_flap_fraction: 0.3 + noop_id: 0 + flap_id: 1 + latency_curriculum: + enabled: false + strategy: exclusive + latencies: null + phase_steps: null + phase_distributions: null + new_latency_passes: 1.0 + replay_passes: 0.25 + target_total_passes: 2.0 + final_equalization: true + step_budget_mode: auto + eval_at_phase_end: false + save_at_phase_end: false + computed_plan: null + per_device_batch_size: 8 + load_all_data_for_training: true + num_obs_frames: 1 + image_mode: single + prompt_mode: raw + stitch_grid: + - 2 + - 2 + obs_image_size: null + video_backend: torchvision_av + gymnasium_task_contract: + task_name: inverted_pendulum_rgb_state + env_id: LatencyBench/InvertedPendulumDiscrete-v0 + registration_imports: + - latency_bench.envs.gymnasium_mujoco + env_fps: 25.0 + obs_fps: 25.0 + frame_stack: 1 + base_prompt: 'Keep the pole upright by applying horizontal cart forces. Choose + exactly one action from: push_left_hard, push_left_soft, zero_force, push_right_soft, + push_right_hard.' + action_labels: + - push_left_hard + - push_left_soft + - zero_force + - push_right_soft + - push_right_hard + action_values: + - 0 + - 1 + - 2 + - 3 + - 4 + noop_action_id: 2 + make_kwargs: + base_env_id: InvertedPendulum-v4 + render_mode: rgb_array + force_values: + - -3.0 + - -1.5 + - 0.0 + - 1.5 + - 3.0 + base_make_kwargs: {} + state_space: + labels: + - cart_position + - pole_angle + - cart_velocity + - pole_angular_velocity + active_action_dim: 5 +dataset: + source_hf: latency-sensitive-bench/inverted_pendulum_200ep_2 + config_name: null + source_subdir: inverted_pendulum_fixed_latency_0_100ep_1ksteps_og,inverted_pendulum_fixed_latency_2_1000ep_7k2steps,inverted_pendulum_fixed_latency_4_1000ep_7k2steps + converted_name: openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2 + single_source_hf: '' + mixed_source_hf: '' + single_converted_name: flappy_train + mixed_converted_name: flappy_train + single_latency_filter: null + mixed_latency_filter: null + force_download: false + setup_force: true + skip_verification: true + target_latency_unit: observation_steps + verify_rows: 200 + max_episodes: null + max_steps_per_episode: 1000 + episodes_per_latency: null + episodes_per_latency_by_latency: 0:90,2:1000,4:1000 + latency_filter: + - 0 + - 2 + - 4 + debug_subset: + enabled: false + max_episodes: 5 + suffix: debug +base_model: + repo_id: Qwen/Qwen3-VL-4B-Instruct +initialization: + checkpoint_local_dir: null + checkpoint_hf_repo_id: null + checkpoint_filename: null +trainer: + max_train_steps: 5000 + num_warmup_steps: 100 + save_interval: 5000 + eval_interval: 5000 + eval_num_batches: 200 + per_latency_eval_num_batches: 5 + eval_action_classification: true + eval_action_classification_interval: null + cc_f1_tolerance: 1 + learning_rate: + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + action_model: 0.0001 + lr_scheduler_type: cosine_with_min_lr + scheduler_specific_kwargs: + min_lr: 1.0e-06 + freeze_modules: '' + freeze_vit: false + freeze_tied_embedding: false + freeze_llm_layers: [] + loss_scale: + vla: 1.0 + vlm: 0.1 + max_grad_norm: 1.0 + weight_decay: 0.0 + logging_frequency: 1 + profile_timing: + enabled: false + log_interval: 10 + gradient_clipping: 1.0 + gradient_accumulation_steps: 2 + distributed_backend: none + is_resume: false + pretrained_checkpoint: null + resume_step: 0 + reload_modules: null + optimizer: + name: AdamW + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + weight_decay: 1.0e-08 + fused: true + save_format: pt +workspace_dir: /home/ubuntu/talha +run_root_dir: /home/ubuntu/talha/results/Checkpoints +seed: 42 +wandb_entity: talha1503 +wandb_project: latency-sensitive-bench +wandb_group: openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2 +wandb_tags: [] +training_latency_condition: unspecified +auth: + env_file: null + hf_token_env: HF_TOKEN + wandb_api_key_env: WANDB_API_KEY +paths: + run_root_dir: results/Checkpoints + dataset_local_dir: playground/Datasets/rl_games + dataset_cache_dir: null + base_model_dir: /home/ubuntu/talha/playground/Pretrained_models/Qwen3-VL-4B-Instruct + accelerate_config: starVLA/config/deepseeds/deepspeed_zero2.yaml +launch: + use_accelerate: true + gpus: null + num_processes: 1 + dry_run: false +conda: + enabled: true + env_name: null +rl_games: + model_alias: openvla + env_eval: + image_size: 224 + frameskip: 1 + image_transform: raw_rgb + prompt_mode: raw + ghost_trail: + history_frames: 5 + gamma: 1.3 + min_alpha: 35 + scroll_px_per_step: 4.0 + ground_fraction: 0.22 + seed: 42 + fixed_episode_seeds: true + latency_seed_stride: 0 + task_seed_stride: 0 + task_description: '' + eval_parallel_envs: 5 + action_chunk_execution: + enabled: false + chunk_size: null + enabled: true + gymnasium: + make_kwargs: {} + mid_train: + enabled: false + interval_steps: 100 + latencies: + - 0 + - 2 + - 4 + num_episodes: 5 + max_steps_per_episode: 3600 + post_train: + enabled: true + latencies: + - 0 + latencies_from_prompt_map: false + num_episodes: 20 + max_steps_per_episode: 3600 + eval_backend: latency_bench + distributed_mode: none + vectorized: + enabled: false + batch_size: 1 + latency: + prompt_map_path: /home/ubuntu/talha/playground/Datasets/rl_games/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2/latency_prompt_map.json + mode: mixed + values: + - 0 + task: gymnasium + gymnasium: + task_contract: + task_name: inverted_pendulum_rgb_state + env_id: LatencyBench/InvertedPendulumDiscrete-v0 + registration_imports: + - latency_bench.envs.gymnasium_mujoco + env_fps: 25.0 + obs_fps: 25.0 + frame_stack: 1 + base_prompt: 'Keep the pole upright by applying horizontal cart forces. Choose + exactly one action from: push_left_hard, push_left_soft, zero_force, push_right_soft, + push_right_hard.' + action_labels: + - push_left_hard + - push_left_soft + - zero_force + - push_right_soft + - push_right_hard + action_values: + - 0 + - 1 + - 2 + - 3 + - 4 + noop_action_id: 2 + make_kwargs: + base_env_id: InvertedPendulum-v4 + render_mode: rgb_array + force_values: + - -3.0 + - -1.5 + - 0.0 + - 1.5 + - 3.0 + state_space: + labels: + - cart_position + - pole_angle + - cart_velocity + - pole_angular_velocity + task_name: inverted_pendulum_rgb_state + initialization_mode: scratch + action_carrier: native +model: openvla +env: gymnasium +init: scratch +mode: mixed_latency +checkpoint: + load: none + hf_repo_id: talha15032/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2 + save_best_model: false + save_final_model: true + save_pt_file: true + save_training_state: false + save_safetensors_file: false + local: + keep_last_n: 1 + sync: + enabled: true + repo_id: talha15032/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2 + keep_last_n: 0 + sync_every_n_checkpoints: 1 + resume_policy: local_latest +run_id: openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2 +output_dir: /home/ubuntu/talha/results/Checkpoints/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2 +config_yaml: null +is_debug: false +version_id: '0.21' diff --git a/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2/config.yaml b/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2/config.yaml new file mode 100644 index 0000000000000000000000000000000000000000..13e583cfe91ebd79127b94983752b6cda0551dd1 --- /dev/null +++ b/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2/config.yaml @@ -0,0 +1,93 @@ +checkpoint: + local: + keep_last_n: 1 + save_best_model: false + save_final_model: true + save_pt_file: true + save_safetensors_file: false + save_training_state: false + sync: + enabled: true + keep_last_n: 0 + repo_id: talha15032/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2 +datasets: + vla_data: + data_mix: openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2 + dataset_py: lerobot_datasets + eval_data_mix: openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2__val + latency_curriculum: + enabled: false + per_device_batch_size: 8 +framework: + action_model: + action_dim: 5 + action_env_dim: 5 + action_hidden_dim: 2560 + action_horizon: 1 + action_model_type: MLP + loss_type: discrete_ce + state_encoding: discretized_text + kv_memory: + enabled: false + packed_train: false + rebased_sink: true + rollout_len: 8 + window: 4 + name: QwenOFT + qwenvl: + attn_implementation: flash_attention_2 + base_vlm: /home/ubuntu/talha/playground/Pretrained_models/Qwen3-VL-4B-Instruct + enable_gradient_checkpointing: true +output_dir: /home/ubuntu/talha/results/Checkpoints/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2 +rl_games: + env_eval: + enabled: true + eval_backend: latency_bench + mid_train: + enabled: false + interval_steps: 100 + prompt_mode: raw + task: gymnasium +run_id: openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2 +run_root_dir: /home/ubuntu/talha/results/Checkpoints +seed: 42 +trainer: + distributed_backend: none + eval_action_classification: true + eval_action_classification_interval: null + eval_interval: 5000 + eval_num_batches: 200 + freeze_llm_layers: [] + freeze_modules: '' + freeze_tied_embedding: false + freeze_vit: false + gradient_accumulation_steps: 2 + gradient_clipping: 1.0 + is_resume: false + learning_rate: + action_model: 0.0001 + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + logging_frequency: 1 + lr_scheduler_type: cosine_with_min_lr + max_train_steps: 5000 + num_warmup_steps: 100 + optimizer: + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + fused: true + weight_decay: 1.0e-08 + per_latency_eval_num_batches: 5 + pretrained_checkpoint: null + profile_timing: + enabled: false + save_interval: 5000 + scheduler_specific_kwargs: + min_lr: 1.0e-06 +training_latency_condition: unspecified +wandb_entity: talha1503 +wandb_group: openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2 +wandb_project: latency-sensitive-bench +wandb_tags: [] diff --git a/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2/dataset_statistics.json b/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2/dataset_statistics.json new file mode 100644 index 0000000000000000000000000000000000000000..48310ea5d5254896f48d5e8d123c1528088202a0 --- /dev/null +++ b/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2/dataset_statistics.json @@ -0,0 +1,95 @@ +{ + "new_embodiment": { + "action": { + "mean": [ + 0.027996724471449852, + 0.1508392095565796, + 0.6481470465660095, + 0.13600370287895203, + 0.03701331838965416 + ], + "std": [ + 0.16490621864795685, + 0.3581511974334717, + 0.4778982102870941, + 0.3424498736858368, + 0.18899837136268616 + ], + "max": [ + 1.0, + 1.0, + 1.0, + 1.0, + 1.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 1.0, + 1.0, + 1.0, + 1.0, + 1.0 + ], + "mask": [ + true, + true, + true, + true, + true + ] + }, + "state": { + "mean": [ + 0.143598273396492, + 0.004461213015019894, + 0.11432920396327972, + -0.006049089599400759 + ], + "std": [ + 0.31354930996894836, + 0.06130174919962883, + 0.4503016769886017, + 0.8575727343559265 + ], + "max": [ + 1.031674861907959, + 0.19994768500328064, + 2.3079874515533447, + 3.7387092113494873 + ], + "min": [ + -1.0259008407592773, + -0.1999799758195877, + -2.2849948406219482, + -4.038050174713135 + ], + "q01": [ + -0.3022584795951843, + -0.15527626156806945, + -0.9133486723899841, + -2.359431056976318 + ], + "q99": [ + 1.0026830291748048, + 0.16135655224323286, + 1.3828536510467528, + 2.107583684921267 + ] + }, + "num_transitions": 218597, + "num_trajectories": 2070 + } +} \ No newline at end of file diff --git a/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2/dataset_statistics_eval.json b/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2/dataset_statistics_eval.json new file mode 100644 index 0000000000000000000000000000000000000000..d8b9a06ce9d7f472c01600f4e742aa2411c89047 --- /dev/null +++ b/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2/dataset_statistics_eval.json @@ -0,0 +1,95 @@ +{ + "new_embodiment": { + "action": { + "mean": [ + 0.035599857568740845, + 0.1577073633670807, + 0.6541473865509033, + 0.07760769128799438, + 0.07493770122528076 + ], + "std": [ + 0.1853053718805313, + 0.3644678294658661, + 0.4756529927253723, + 0.26755642890930176, + 0.26328903436660767 + ], + "max": [ + 1.0, + 1.0, + 1.0, + 1.0, + 1.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 1.0, + 1.0, + 1.0, + 1.0, + 1.0 + ], + "mask": [ + true, + true, + true, + true, + true + ] + }, + "state": { + "mean": [ + -0.006056365091353655, + -0.001078602741472423, + 0.021814774721860886, + -0.004301219247281551 + ], + "std": [ + 0.15048037469387054, + 0.0324825718998909, + 0.33326858282089233, + 0.713808000087738 + ], + "max": [ + 1.0273027420043945, + 0.198622465133667, + 2.3032615184783936, + 3.4846131801605225 + ], + "min": [ + -0.41009554266929626, + -0.19869747757911682, + -1.3806304931640625, + -3.9696083068847656 + ], + "q01": [ + -0.08295003287494182, + -0.1170046754181385, + -0.977174437046051, + -2.3071367263793947 + ], + "q99": [ + 0.9016301155090326, + 0.11161688491702067, + 1.2677143096923813, + 2.1670001268386816 + ] + }, + "num_transitions": 11236, + "num_trajectories": 30 + } +} \ No newline at end of file diff --git a/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2/hydra/.hydra/config.yaml b/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2/hydra/.hydra/config.yaml new file mode 100644 index 0000000000000000000000000000000000000000..516ab34e4b6ae4de778bbfa2b871746a30cb6ed1 --- /dev/null +++ b/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2/hydra/.hydra/config.yaml @@ -0,0 +1,338 @@ +framework: + qwenvl: + base_vlm: /home/ubuntu/talha/playground/Pretrained_models/Qwen3-VL-4B-Instruct + attn_implementation: flash_attention_2 + enable_gradient_checkpointing: true + action_model: + state_dim: 4 + loss_type: discrete_ce + action_horizon: 1 + future_action_window_size: 0 + past_action_window_size: 0 + action_dim: 5 + action_env_dim: 5 + state_encoding: discretized_text + action_model_type: MLP + action_hidden_dim: 2560 + kv_memory: + enabled: false + window: 4 + rollout_len: 8 + packed_train: false + rebased_sink: true + name: QwenOFT +datasets: + vla_data: + dataset_py: lerobot_datasets + include_state: true + data_root_dir: /home/ubuntu/talha/playground/Datasets/rl_games + data_mix: openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2 + eval_data_mix: openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2__val + custom_mixtures_path: /home/ubuntu/talha/playground/Datasets/rl_games/_generated_mixtures/single_77a677bfbf8e73b8.json + action_type: discrete + sequential_step_sampling: false + eval_sequential_step_sampling: null + num_workers: 8 + eval_num_workers: 8 + prefetch_factor: 4 + persistent_workers: true + pin_memory: true + shuffle: true + action_balance: + enabled: false + strategy: balanced_epoch + action_key: action_id + target_flap_fraction: 0.3 + noop_id: 0 + flap_id: 1 + latency_curriculum: + enabled: false + strategy: exclusive + latencies: null + phase_steps: null + phase_distributions: null + new_latency_passes: 1.0 + replay_passes: 0.25 + target_total_passes: 2.0 + final_equalization: true + step_budget_mode: auto + eval_at_phase_end: false + save_at_phase_end: false + computed_plan: null + per_device_batch_size: 8 + load_all_data_for_training: true + num_obs_frames: 1 + image_mode: single + prompt_mode: raw + stitch_grid: + - 2 + - 2 + obs_image_size: null + video_backend: torchvision_av + gymnasium_task_contract: + task_name: inverted_pendulum_rgb_state + env_id: LatencyBench/InvertedPendulumDiscrete-v0 + registration_imports: + - latency_bench.envs.gymnasium_mujoco + env_fps: 25.0 + obs_fps: 25.0 + frame_stack: 1 + base_prompt: 'Keep the pole upright by applying horizontal cart forces. Choose + exactly one action from: push_left_hard, push_left_soft, zero_force, push_right_soft, + push_right_hard.' + action_labels: + - push_left_hard + - push_left_soft + - zero_force + - push_right_soft + - push_right_hard + action_values: + - 0 + - 1 + - 2 + - 3 + - 4 + noop_action_id: 2 + make_kwargs: + base_env_id: InvertedPendulum-v4 + render_mode: rgb_array + force_values: + - -3.0 + - -1.5 + - 0.0 + - 1.5 + - 3.0 + base_make_kwargs: {} + state_space: + labels: + - cart_position + - pole_angle + - cart_velocity + - pole_angular_velocity + active_action_dim: 5 +dataset: + source_hf: latency-sensitive-bench/inverted_pendulum_200ep_2 + config_name: null + source_subdir: inverted_pendulum_fixed_latency_0_100ep_1ksteps_og,inverted_pendulum_fixed_latency_2_1000ep_7k2steps,inverted_pendulum_fixed_latency_4_1000ep_7k2steps + converted_name: openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2 + single_source_hf: '' + mixed_source_hf: '' + single_converted_name: flappy_train + mixed_converted_name: flappy_train + single_latency_filter: null + mixed_latency_filter: null + force_download: false + setup_force: true + skip_verification: true + target_latency_unit: observation_steps + verify_rows: 200 + max_episodes: null + max_steps_per_episode: 1000 + episodes_per_latency: null + episodes_per_latency_by_latency: 0:90,2:1000,4:1000 + latency_filter: + - 0 + - 2 + - 4 + debug_subset: + enabled: false + max_episodes: 5 + suffix: debug +base_model: + repo_id: Qwen/Qwen3-VL-4B-Instruct +initialization: + checkpoint_local_dir: null + checkpoint_hf_repo_id: null + checkpoint_filename: null +trainer: + max_train_steps: 5000 + num_warmup_steps: 100 + save_interval: 5000 + eval_interval: 5000 + eval_num_batches: 200 + per_latency_eval_num_batches: 5 + eval_action_classification: true + eval_action_classification_interval: null + cc_f1_tolerance: 1 + learning_rate: + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + action_model: 0.0001 + lr_scheduler_type: cosine_with_min_lr + scheduler_specific_kwargs: + min_lr: 1.0e-06 + freeze_modules: '' + freeze_vit: false + freeze_tied_embedding: false + freeze_llm_layers: [] + loss_scale: + vla: 1.0 + vlm: 0.1 + max_grad_norm: 1.0 + weight_decay: 0.0 + logging_frequency: 1 + profile_timing: + enabled: false + log_interval: 10 + gradient_clipping: 1.0 + gradient_accumulation_steps: 2 + distributed_backend: none + is_resume: false + pretrained_checkpoint: null + resume_step: 0 + reload_modules: null + optimizer: + name: AdamW + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + weight_decay: 1.0e-08 + fused: true + save_format: pt +workspace_dir: /home/ubuntu/talha +run_root_dir: /home/ubuntu/talha/results/Checkpoints +seed: 42 +wandb_entity: ${oc.env:WANDB_ENTITY} +wandb_project: ${oc.env:WANDB_PROJECT,starVLA_rl_games} +wandb_group: openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2 +wandb_tags: [] +training_latency_condition: unspecified +auth: + env_file: null + hf_token_env: HF_TOKEN + wandb_api_key_env: WANDB_API_KEY +paths: + run_root_dir: results/Checkpoints + dataset_local_dir: playground/Datasets/rl_games + dataset_cache_dir: null + base_model_dir: /home/ubuntu/talha/playground/Pretrained_models/Qwen3-VL-4B-Instruct + accelerate_config: starVLA/config/deepseeds/deepspeed_zero2.yaml +launch: + use_accelerate: true + gpus: null + num_processes: 1 + dry_run: false +conda: + enabled: true + env_name: null +rl_games: + model_alias: openvla + env_eval: + image_size: 224 + frameskip: 1 + image_transform: raw_rgb + prompt_mode: raw + ghost_trail: + history_frames: 5 + gamma: 1.3 + min_alpha: 35 + scroll_px_per_step: 4.0 + ground_fraction: 0.22 + seed: 42 + fixed_episode_seeds: true + latency_seed_stride: 0 + task_seed_stride: 0 + task_description: '' + eval_parallel_envs: 5 + action_chunk_execution: + enabled: false + chunk_size: null + enabled: true + gymnasium: + make_kwargs: {} + mid_train: + enabled: false + interval_steps: 100 + latencies: + - 0 + - 2 + - 4 + num_episodes: 5 + max_steps_per_episode: 3600 + post_train: + enabled: true + latencies: + - 0 + latencies_from_prompt_map: false + num_episodes: 20 + max_steps_per_episode: 3600 + eval_backend: latency_bench + distributed_mode: none + vectorized: + enabled: false + batch_size: 1 + latency: + prompt_map_path: /home/ubuntu/talha/playground/Datasets/rl_games/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2/latency_prompt_map.json + mode: mixed + values: + - 0 + task: gymnasium + gymnasium: + task_contract: + task_name: inverted_pendulum_rgb_state + env_id: LatencyBench/InvertedPendulumDiscrete-v0 + registration_imports: + - latency_bench.envs.gymnasium_mujoco + env_fps: 25.0 + obs_fps: 25.0 + frame_stack: 1 + base_prompt: 'Keep the pole upright by applying horizontal cart forces. Choose + exactly one action from: push_left_hard, push_left_soft, zero_force, push_right_soft, + push_right_hard.' + action_labels: + - push_left_hard + - push_left_soft + - zero_force + - push_right_soft + - push_right_hard + action_values: + - 0 + - 1 + - 2 + - 3 + - 4 + noop_action_id: 2 + make_kwargs: + base_env_id: InvertedPendulum-v4 + render_mode: rgb_array + force_values: + - -3.0 + - -1.5 + - 0.0 + - 1.5 + - 3.0 + state_space: + labels: + - cart_position + - pole_angle + - cart_velocity + - pole_angular_velocity + task_name: inverted_pendulum_rgb_state + initialization_mode: scratch + action_carrier: native +model: openvla +env: gymnasium +init: scratch +mode: mixed_latency +checkpoint: + load: none + hf_repo_id: talha15032/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2 + save_best_model: false + save_final_model: true + save_pt_file: true + save_training_state: false + save_safetensors_file: false + local: + keep_last_n: 1 + sync: + enabled: true + repo_id: talha15032/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2 + keep_last_n: 0 + sync_every_n_checkpoints: 1 + resume_policy: local_latest +run_id: openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2 +output_dir: null +config_yaml: null +is_debug: false +version_id: 0.21 diff --git a/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2/hydra/.hydra/hydra.yaml b/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2/hydra/.hydra/hydra.yaml new file mode 100644 index 0000000000000000000000000000000000000000..c44e0229c2da6f99a959fe88c4f85cc529f4f4dc --- /dev/null +++ b/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2/hydra/.hydra/hydra.yaml @@ -0,0 +1,394 @@ +hydra: + run: + dir: ${run_root_dir}/${run_id}/hydra + sweep: + dir: multirun/${now:%Y-%m-%d}/${now:%H-%M-%S} + subdir: ${hydra.job.num} + launcher: + _target_: hydra._internal.core_plugins.basic_launcher.BasicLauncher + sweeper: + _target_: hydra._internal.core_plugins.basic_sweeper.BasicSweeper + max_batch_size: null + params: null + help: + app_name: ${hydra.job.name} + header: '${hydra.help.app_name} is powered by Hydra. + + ' + footer: 'Powered by Hydra (https://hydra.cc) + + Use --hydra-help to view Hydra specific help + + ' + template: '${hydra.help.header} + + == Configuration groups == + + Compose your configuration from those groups (group=option) + + + $APP_CONFIG_GROUPS + + + == Config == + + Override anything in the config (foo.bar=value) + + + $CONFIG + + + ${hydra.help.footer} + + ' + hydra_help: + template: 'Hydra (${hydra.runtime.version}) + + See https://hydra.cc for more info. + + + == Flags == + + $FLAGS_HELP + + + == Configuration groups == + + Compose your configuration from those groups (For example, append hydra/job_logging=disabled + to command line) + + + $HYDRA_CONFIG_GROUPS + + + Use ''--cfg hydra'' to Show the Hydra config. + + ' + hydra_help: ??? + hydra_logging: + version: 1 + formatters: + simple: + format: '[%(asctime)s][HYDRA] %(message)s' + handlers: + console: + class: logging.StreamHandler + formatter: simple + stream: ext://sys.stdout + root: + level: INFO + handlers: + - console + loggers: + logging_example: + level: DEBUG + disable_existing_loggers: false + job_logging: + version: 1 + formatters: + simple: + format: '[%(asctime)s][%(name)s][%(levelname)s] - %(message)s' + handlers: + console: + class: logging.StreamHandler + formatter: simple + stream: ext://sys.stdout + file: + class: logging.FileHandler + formatter: simple + filename: ${hydra.runtime.output_dir}/${hydra.job.name}.log + root: + level: INFO + handlers: + - console + - file + disable_existing_loggers: false + env: {} + mode: RUN + searchpath: [] + callbacks: {} + output_subdir: .hydra + overrides: + hydra: + - hydra.mode=RUN + task: + - model=openvla + - env=gymnasium + - init=scratch + - mode=mixed_latency + - ++framework.qwenvl.base_vlm=/home/ubuntu/talha/playground/Pretrained_models/Qwen3-VL-4B-Instruct + - ++framework.qwenvl.attn_implementation=flash_attention_2 + - ++framework.qwenvl.enable_gradient_checkpointing=true + - ++framework.action_model.state_dim=4 + - ++framework.action_model.loss_type=discrete_ce + - ++framework.action_model.action_horizon=1 + - ++framework.action_model.future_action_window_size=0 + - ++framework.action_model.past_action_window_size=0 + - ++framework.action_model.action_dim=5 + - ++framework.action_model.action_env_dim=5 + - ++framework.action_model.state_encoding=discretized_text + - ++framework.action_model.action_model_type=MLP + - ++framework.action_model.action_hidden_dim=2560 + - ++framework.kv_memory.enabled=false + - ++framework.kv_memory.window=4 + - ++framework.kv_memory.rollout_len=8 + - ++framework.kv_memory.packed_train=false + - ++framework.kv_memory.rebased_sink=true + - ++framework.name=QwenOFT + - ++datasets.vla_data.dataset_py=lerobot_datasets + - ++datasets.vla_data.include_state=true + - ++datasets.vla_data.data_root_dir=playground/Datasets/rl_games + - ++datasets.vla_data.data_mix=openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2 + - ++datasets.vla_data.eval_data_mix=openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2__val + - ++datasets.vla_data.custom_mixtures_path=null + - ++datasets.vla_data.action_type=discrete + - ++datasets.vla_data.sequential_step_sampling=false + - ++datasets.vla_data.eval_sequential_step_sampling=null + - ++datasets.vla_data.num_workers=8 + - ++datasets.vla_data.eval_num_workers=8 + - ++datasets.vla_data.prefetch_factor=4 + - ++datasets.vla_data.persistent_workers=true + - ++datasets.vla_data.pin_memory=true + - ++datasets.vla_data.shuffle=true + - ++datasets.vla_data.action_balance.enabled=false + - ++datasets.vla_data.action_balance.strategy=balanced_epoch + - ++datasets.vla_data.action_balance.action_key=action_id + - ++datasets.vla_data.action_balance.target_flap_fraction=0.3 + - ++datasets.vla_data.action_balance.noop_id=0 + - ++datasets.vla_data.action_balance.flap_id=1 + - ++datasets.vla_data.latency_curriculum.enabled=false + - ++datasets.vla_data.latency_curriculum.strategy=exclusive + - ++datasets.vla_data.latency_curriculum.latencies=null + - ++datasets.vla_data.latency_curriculum.phase_steps=null + - ++datasets.vla_data.latency_curriculum.phase_distributions=null + - ++datasets.vla_data.latency_curriculum.new_latency_passes=1.0 + - ++datasets.vla_data.latency_curriculum.replay_passes=0.25 + - ++datasets.vla_data.latency_curriculum.target_total_passes=2.0 + - ++datasets.vla_data.latency_curriculum.final_equalization=true + - ++datasets.vla_data.latency_curriculum.step_budget_mode=auto + - ++datasets.vla_data.latency_curriculum.eval_at_phase_end=false + - ++datasets.vla_data.latency_curriculum.save_at_phase_end=false + - ++datasets.vla_data.latency_curriculum.computed_plan=null + - ++datasets.vla_data.per_device_batch_size=8 + - ++datasets.vla_data.load_all_data_for_training=true + - ++datasets.vla_data.num_obs_frames=1 + - ++datasets.vla_data.image_mode=single + - ++datasets.vla_data.prompt_mode=raw + - ++datasets.vla_data.stitch_grid=[2,2] + - ++datasets.vla_data.obs_image_size=null + - ++datasets.vla_data.video_backend=torchvision_av + - '++datasets.vla_data.gymnasium_task_contract={task_name:inverted_pendulum_rgb_state,env_id:LatencyBench/InvertedPendulumDiscrete-v0,registration_imports:[latency_bench.envs.gymnasium_mujoco],env_fps:25.0,obs_fps:25.0,frame_stack:1,base_prompt:''Keep + the pole upright by applying horizontal cart forces. Choose exactly one action + from: push_left_hard, push_left_soft, zero_force, push_right_soft, push_right_hard.'',action_labels:[push_left_hard,push_left_soft,zero_force,push_right_soft,push_right_hard],action_values:[0,1,2,3,4],noop_action_id:2,make_kwargs:{base_env_id:InvertedPendulum-v4,render_mode:rgb_array,force_values:[-3.0,-1.5,0.0,1.5,3.0],base_make_kwargs:{}},state_space:{labels:[cart_position,pole_angle,cart_velocity,pole_angular_velocity]}}' + - ++datasets.vla_data.active_action_dim=5 + - ++dataset.source_hf=latency-sensitive-bench/inverted_pendulum_200ep_2 + - ++dataset.config_name=null + - ++dataset.source_subdir='inverted_pendulum_fixed_latency_0_100ep_1ksteps_og,inverted_pendulum_fixed_latency_2_1000ep_7k2steps,inverted_pendulum_fixed_latency_4_1000ep_7k2steps' + - ++dataset.converted_name=openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2 + - ++dataset.single_source_hf= + - ++dataset.mixed_source_hf= + - ++dataset.single_converted_name=flappy_train + - ++dataset.mixed_converted_name=flappy_train + - ++dataset.single_latency_filter=null + - ++dataset.mixed_latency_filter=null + - ++dataset.force_download=false + - ++dataset.setup_force=true + - ++dataset.skip_verification=true + - ++dataset.target_latency_unit=observation_steps + - ++dataset.verify_rows=200 + - ++dataset.max_episodes=null + - ++dataset.max_steps_per_episode=1000 + - ++dataset.episodes_per_latency=null + - ++dataset.episodes_per_latency_by_latency='0:90,2:1000,4:1000' + - ++dataset.latency_filter=[0,2,4] + - ++dataset.debug_subset.enabled=false + - ++dataset.debug_subset.max_episodes=5 + - ++dataset.debug_subset.suffix=debug + - ++base_model.repo_id=Qwen/Qwen3-VL-4B-Instruct + - ++initialization.checkpoint_local_dir=null + - ++initialization.checkpoint_hf_repo_id=null + - ++initialization.checkpoint_filename=null + - ++trainer.max_train_steps=5000 + - ++trainer.num_warmup_steps=100 + - ++trainer.save_interval=5000 + - ++trainer.eval_interval=5000 + - ++trainer.eval_num_batches=200 + - ++trainer.per_latency_eval_num_batches=5 + - ++trainer.eval_action_classification=true + - ++trainer.eval_action_classification_interval=null + - ++trainer.cc_f1_tolerance=1 + - ++trainer.learning_rate.base=2e-05 + - ++trainer.learning_rate.qwen_vl_interface=1e-05 + - ++trainer.learning_rate.action_model=0.0001 + - ++trainer.lr_scheduler_type=cosine_with_min_lr + - ++trainer.scheduler_specific_kwargs.min_lr=1e-06 + - ++trainer.freeze_modules= + - ++trainer.freeze_vit=false + - ++trainer.freeze_tied_embedding=false + - ++trainer.freeze_llm_layers=[] + - ++trainer.loss_scale.vla=1.0 + - ++trainer.loss_scale.vlm=0.1 + - ++trainer.max_grad_norm=1.0 + - ++trainer.weight_decay=0.0 + - ++trainer.logging_frequency=1 + - ++trainer.profile_timing.enabled=false + - ++trainer.profile_timing.log_interval=10 + - ++trainer.gradient_clipping=1.0 + - ++trainer.gradient_accumulation_steps=2 + - ++trainer.distributed_backend=none + - ++trainer.is_resume=false + - ++trainer.pretrained_checkpoint=null + - ++trainer.resume_step=0 + - ++trainer.reload_modules=null + - ++trainer.optimizer.name=AdamW + - ++trainer.optimizer.betas=[0.9,0.95] + - ++trainer.optimizer.eps=1e-08 + - ++trainer.optimizer.weight_decay=1e-08 + - ++trainer.optimizer.fused=true + - ++trainer.save_format=pt + - ++workspace_dir=/home/ubuntu/talha + - ++run_root_dir=results/Checkpoints + - ++seed=42 + - ++wandb_group=openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2 + - ++wandb_tags=[] + - ++training_latency_condition=unspecified + - ++auth.env_file=null + - ++auth.hf_token_env=HF_TOKEN + - ++auth.wandb_api_key_env=WANDB_API_KEY + - ++paths.run_root_dir=results/Checkpoints + - ++paths.dataset_local_dir=playground/Datasets/rl_games + - ++paths.dataset_cache_dir=null + - ++paths.base_model_dir=/home/ubuntu/talha/playground/Pretrained_models/Qwen3-VL-4B-Instruct + - ++paths.accelerate_config=starVLA/config/deepseeds/deepspeed_zero2.yaml + - ++rl_games.model_alias=openvla + - ++rl_games.env_eval.image_size=224 + - ++rl_games.env_eval.frameskip=1 + - ++rl_games.env_eval.image_transform=raw_rgb + - ++rl_games.env_eval.prompt_mode=raw + - ++rl_games.env_eval.ghost_trail.history_frames=5 + - ++rl_games.env_eval.ghost_trail.gamma=1.3 + - ++rl_games.env_eval.ghost_trail.min_alpha=35 + - ++rl_games.env_eval.ghost_trail.scroll_px_per_step=4.0 + - ++rl_games.env_eval.ghost_trail.ground_fraction=0.22 + - ++rl_games.env_eval.seed=42 + - ++rl_games.env_eval.fixed_episode_seeds=true + - ++rl_games.env_eval.latency_seed_stride=0 + - ++rl_games.env_eval.task_seed_stride=0 + - ++rl_games.env_eval.task_description= + - ++rl_games.env_eval.eval_parallel_envs=5 + - ++rl_games.env_eval.action_chunk_execution.enabled=false + - ++rl_games.env_eval.action_chunk_execution.chunk_size=null + - ++rl_games.env_eval.enabled=true + - ++rl_games.env_eval.mid_train.enabled=false + - ++rl_games.env_eval.mid_train.interval_steps=100 + - ++rl_games.env_eval.mid_train.latencies=[0,2,4] + - ++rl_games.env_eval.mid_train.num_episodes=5 + - ++rl_games.env_eval.mid_train.max_steps_per_episode=3600 + - ++rl_games.env_eval.post_train.enabled=true + - ++rl_games.env_eval.post_train.latencies=[0] + - ++rl_games.env_eval.post_train.latencies_from_prompt_map=false + - ++rl_games.env_eval.post_train.num_episodes=20 + - ++rl_games.env_eval.post_train.max_steps_per_episode=3600 + - ++rl_games.env_eval.eval_backend=latency_bench + - ++rl_games.env_eval.distributed_mode=none + - ++rl_games.env_eval.vectorized.enabled=false + - ++rl_games.env_eval.vectorized.batch_size=1 + - ++rl_games.env_eval.latency.prompt_map_path=null + - ++rl_games.env_eval.latency.mode=mixed + - ++rl_games.env_eval.latency.values=[0] + - ++rl_games.task=gymnasium + - ++rl_games.gymnasium.task_contract.task_name=inverted_pendulum_rgb_state + - ++rl_games.gymnasium.task_contract.env_id=LatencyBench/InvertedPendulumDiscrete-v0 + - ++rl_games.gymnasium.task_contract.registration_imports=[latency_bench.envs.gymnasium_mujoco] + - ++rl_games.gymnasium.task_contract.env_fps=25.0 + - ++rl_games.gymnasium.task_contract.obs_fps=25.0 + - ++rl_games.gymnasium.task_contract.frame_stack=1 + - '++rl_games.gymnasium.task_contract.base_prompt=''Keep the pole upright by applying + horizontal cart forces. Choose exactly one action from: push_left_hard, push_left_soft, + zero_force, push_right_soft, push_right_hard.''' + - ++rl_games.gymnasium.task_contract.action_labels=[push_left_hard,push_left_soft,zero_force,push_right_soft,push_right_hard] + - ++rl_games.gymnasium.task_contract.action_values=[0,1,2,3,4] + - ++rl_games.gymnasium.task_contract.noop_action_id=2 + - ++rl_games.gymnasium.task_contract.make_kwargs.base_env_id=InvertedPendulum-v4 + - ++rl_games.gymnasium.task_contract.make_kwargs.render_mode=rgb_array + - ++rl_games.gymnasium.task_contract.make_kwargs.force_values=[-3.0,-1.5,0.0,1.5,3.0] + - ++rl_games.gymnasium.task_contract.state_space.labels=[cart_position,pole_angle,cart_velocity,pole_angular_velocity] + - ++rl_games.gymnasium.task_name=inverted_pendulum_rgb_state + - ++rl_games.initialization_mode=scratch + - ++rl_games.action_carrier=native + - ++checkpoint.load=none + - ++checkpoint.hf_repo_id=talha15032/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2 + - ++checkpoint.save_best_model=false + - ++checkpoint.save_final_model=true + - ++checkpoint.save_pt_file=true + - ++checkpoint.save_training_state=false + - ++checkpoint.save_safetensors_file=false + - ++checkpoint.local.keep_last_n=1 + - ++checkpoint.sync.enabled=true + - ++checkpoint.sync.repo_id=talha15032/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2 + - ++checkpoint.sync.keep_last_n=0 + - ++checkpoint.sync.sync_every_n_checkpoints=1 + - ++checkpoint.sync.resume_policy=local_latest + - ++run_id=openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2 + - ++output_dir=null + - ++config_yaml=null + - ++is_debug=false + - ++version_id=0.21 + - ++run_root_dir=/home/ubuntu/talha/results/Checkpoints + - ++trainer.is_resume=false + - ++datasets.vla_data.data_root_dir=/home/ubuntu/talha/playground/Datasets/rl_games + - ++datasets.vla_data.data_mix=openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2 + - ++datasets.vla_data.eval_data_mix=openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2__val + - ++datasets.vla_data.custom_mixtures_path=/home/ubuntu/talha/playground/Datasets/rl_games/_generated_mixtures/single_77a677bfbf8e73b8.json + - ++framework.qwenvl.base_vlm=/home/ubuntu/talha/playground/Pretrained_models/Qwen3-VL-4B-Instruct + - ++rl_games.env_eval.latency.prompt_map_path=/home/ubuntu/talha/playground/Datasets/rl_games/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2/latency_prompt_map.json + job: + name: train_starvla_hydra + chdir: false + override_dirname: '++auth.env_file=null,++auth.hf_token_env=HF_TOKEN,++auth.wandb_api_key_env=WANDB_API_KEY,++base_model.repo_id=Qwen/Qwen3-VL-4B-Instruct,++checkpoint.hf_repo_id=talha15032/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2,++checkpoint.load=none,++checkpoint.local.keep_last_n=1,++checkpoint.save_best_model=false,++checkpoint.save_final_model=true,++checkpoint.save_pt_file=true,++checkpoint.save_safetensors_file=false,++checkpoint.save_training_state=false,++checkpoint.sync.enabled=true,++checkpoint.sync.keep_last_n=0,++checkpoint.sync.repo_id=talha15032/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2,++checkpoint.sync.resume_policy=local_latest,++checkpoint.sync.sync_every_n_checkpoints=1,++config_yaml=null,++dataset.config_name=null,++dataset.converted_name=openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2,++dataset.debug_subset.enabled=false,++dataset.debug_subset.max_episodes=5,++dataset.debug_subset.suffix=debug,++dataset.episodes_per_latency=null,++dataset.episodes_per_latency_by_latency=''0:90,2:1000,4:1000'',++dataset.force_download=false,++dataset.latency_filter=[0,2,4],++dataset.max_episodes=null,++dataset.max_steps_per_episode=1000,++dataset.mixed_converted_name=flappy_train,++dataset.mixed_latency_filter=null,++dataset.mixed_source_hf=,++dataset.setup_force=true,++dataset.single_converted_name=flappy_train,++dataset.single_latency_filter=null,++dataset.single_source_hf=,++dataset.skip_verification=true,++dataset.source_hf=latency-sensitive-bench/inverted_pendulum_200ep_2,++dataset.source_subdir=''inverted_pendulum_fixed_latency_0_100ep_1ksteps_og,inverted_pendulum_fixed_latency_2_1000ep_7k2steps,inverted_pendulum_fixed_latency_4_1000ep_7k2steps'',++dataset.target_latency_unit=observation_steps,++dataset.verify_rows=200,++datasets.vla_data.action_balance.action_key=action_id,++datasets.vla_data.action_balance.enabled=false,++datasets.vla_data.action_balance.flap_id=1,++datasets.vla_data.action_balance.noop_id=0,++datasets.vla_data.action_balance.strategy=balanced_epoch,++datasets.vla_data.action_balance.target_flap_fraction=0.3,++datasets.vla_data.action_type=discrete,++datasets.vla_data.active_action_dim=5,++datasets.vla_data.custom_mixtures_path=/home/ubuntu/talha/playground/Datasets/rl_games/_generated_mixtures/single_77a677bfbf8e73b8.json,++datasets.vla_data.custom_mixtures_path=null,++datasets.vla_data.data_mix=openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2,++datasets.vla_data.data_mix=openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2,++datasets.vla_data.data_root_dir=/home/ubuntu/talha/playground/Datasets/rl_games,++datasets.vla_data.data_root_dir=playground/Datasets/rl_games,++datasets.vla_data.dataset_py=lerobot_datasets,++datasets.vla_data.eval_data_mix=openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2__val,++datasets.vla_data.eval_data_mix=openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2__val,++datasets.vla_data.eval_num_workers=8,++datasets.vla_data.eval_sequential_step_sampling=null,++datasets.vla_data.gymnasium_task_contract={task_name:inverted_pendulum_rgb_state,env_id:LatencyBench/InvertedPendulumDiscrete-v0,registration_imports:[latency_bench.envs.gymnasium_mujoco],env_fps:25.0,obs_fps:25.0,frame_stack:1,base_prompt:''Keep + the pole upright by applying horizontal cart forces. Choose exactly one action + from: push_left_hard, push_left_soft, zero_force, push_right_soft, push_right_hard.'',action_labels:[push_left_hard,push_left_soft,zero_force,push_right_soft,push_right_hard],action_values:[0,1,2,3,4],noop_action_id:2,make_kwargs:{base_env_id:InvertedPendulum-v4,render_mode:rgb_array,force_values:[-3.0,-1.5,0.0,1.5,3.0],base_make_kwargs:{}},state_space:{labels:[cart_position,pole_angle,cart_velocity,pole_angular_velocity]}},++datasets.vla_data.image_mode=single,++datasets.vla_data.include_state=true,++datasets.vla_data.latency_curriculum.computed_plan=null,++datasets.vla_data.latency_curriculum.enabled=false,++datasets.vla_data.latency_curriculum.eval_at_phase_end=false,++datasets.vla_data.latency_curriculum.final_equalization=true,++datasets.vla_data.latency_curriculum.latencies=null,++datasets.vla_data.latency_curriculum.new_latency_passes=1.0,++datasets.vla_data.latency_curriculum.phase_distributions=null,++datasets.vla_data.latency_curriculum.phase_steps=null,++datasets.vla_data.latency_curriculum.replay_passes=0.25,++datasets.vla_data.latency_curriculum.save_at_phase_end=false,++datasets.vla_data.latency_curriculum.step_budget_mode=auto,++datasets.vla_data.latency_curriculum.strategy=exclusive,++datasets.vla_data.latency_curriculum.target_total_passes=2.0,++datasets.vla_data.load_all_data_for_training=true,++datasets.vla_data.num_obs_frames=1,++datasets.vla_data.num_workers=8,++datasets.vla_data.obs_image_size=null,++datasets.vla_data.per_device_batch_size=8,++datasets.vla_data.persistent_workers=true,++datasets.vla_data.pin_memory=true,++datasets.vla_data.prefetch_factor=4,++datasets.vla_data.prompt_mode=raw,++datasets.vla_data.sequential_step_sampling=false,++datasets.vla_data.shuffle=true,++datasets.vla_data.stitch_grid=[2,2],++datasets.vla_data.video_backend=torchvision_av,++framework.action_model.action_dim=5,++framework.action_model.action_env_dim=5,++framework.action_model.action_hidden_dim=2560,++framework.action_model.action_horizon=1,++framework.action_model.action_model_type=MLP,++framework.action_model.future_action_window_size=0,++framework.action_model.loss_type=discrete_ce,++framework.action_model.past_action_window_size=0,++framework.action_model.state_dim=4,++framework.action_model.state_encoding=discretized_text,++framework.kv_memory.enabled=false,++framework.kv_memory.packed_train=false,++framework.kv_memory.rebased_sink=true,++framework.kv_memory.rollout_len=8,++framework.kv_memory.window=4,++framework.name=QwenOFT,++framework.qwenvl.attn_implementation=flash_attention_2,++framework.qwenvl.base_vlm=/home/ubuntu/talha/playground/Pretrained_models/Qwen3-VL-4B-Instruct,++framework.qwenvl.base_vlm=/home/ubuntu/talha/playground/Pretrained_models/Qwen3-VL-4B-Instruct,++framework.qwenvl.enable_gradient_checkpointing=true,++initialization.checkpoint_filename=null,++initialization.checkpoint_hf_repo_id=null,++initialization.checkpoint_local_dir=null,++is_debug=false,++output_dir=null,++paths.accelerate_config=starVLA/config/deepseeds/deepspeed_zero2.yaml,++paths.base_model_dir=/home/ubuntu/talha/playground/Pretrained_models/Qwen3-VL-4B-Instruct,++paths.dataset_cache_dir=null,++paths.dataset_local_dir=playground/Datasets/rl_games,++paths.run_root_dir=results/Checkpoints,++rl_games.action_carrier=native,++rl_games.env_eval.action_chunk_execution.chunk_size=null,++rl_games.env_eval.action_chunk_execution.enabled=false,++rl_games.env_eval.distributed_mode=none,++rl_games.env_eval.enabled=true,++rl_games.env_eval.eval_backend=latency_bench,++rl_games.env_eval.eval_parallel_envs=5,++rl_games.env_eval.fixed_episode_seeds=true,++rl_games.env_eval.frameskip=1,++rl_games.env_eval.ghost_trail.gamma=1.3,++rl_games.env_eval.ghost_trail.ground_fraction=0.22,++rl_games.env_eval.ghost_trail.history_frames=5,++rl_games.env_eval.ghost_trail.min_alpha=35,++rl_games.env_eval.ghost_trail.scroll_px_per_step=4.0,++rl_games.env_eval.image_size=224,++rl_games.env_eval.image_transform=raw_rgb,++rl_games.env_eval.latency.mode=mixed,++rl_games.env_eval.latency.prompt_map_path=/home/ubuntu/talha/playground/Datasets/rl_games/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2/latency_prompt_map.json,++rl_games.env_eval.latency.prompt_map_path=null,++rl_games.env_eval.latency.values=[0],++rl_games.env_eval.latency_seed_stride=0,++rl_games.env_eval.mid_train.enabled=false,++rl_games.env_eval.mid_train.interval_steps=100,++rl_games.env_eval.mid_train.latencies=[0,2,4],++rl_games.env_eval.mid_train.max_steps_per_episode=3600,++rl_games.env_eval.mid_train.num_episodes=5,++rl_games.env_eval.post_train.enabled=true,++rl_games.env_eval.post_train.latencies=[0],++rl_games.env_eval.post_train.latencies_from_prompt_map=false,++rl_games.env_eval.post_train.max_steps_per_episode=3600,++rl_games.env_eval.post_train.num_episodes=20,++rl_games.env_eval.prompt_mode=raw,++rl_games.env_eval.seed=42,++rl_games.env_eval.task_description=,++rl_games.env_eval.task_seed_stride=0,++rl_games.env_eval.vectorized.batch_size=1,++rl_games.env_eval.vectorized.enabled=false,++rl_games.gymnasium.task_contract.action_labels=[push_left_hard,push_left_soft,zero_force,push_right_soft,push_right_hard],++rl_games.gymnasium.task_contract.action_values=[0,1,2,3,4],++rl_games.gymnasium.task_contract.base_prompt=''Keep + the pole upright by applying horizontal cart forces. Choose exactly one action + from: push_left_hard, push_left_soft, zero_force, push_right_soft, push_right_hard.'',++rl_games.gymnasium.task_contract.env_fps=25.0,++rl_games.gymnasium.task_contract.env_id=LatencyBench/InvertedPendulumDiscrete-v0,++rl_games.gymnasium.task_contract.frame_stack=1,++rl_games.gymnasium.task_contract.make_kwargs.base_env_id=InvertedPendulum-v4,++rl_games.gymnasium.task_contract.make_kwargs.force_values=[-3.0,-1.5,0.0,1.5,3.0],++rl_games.gymnasium.task_contract.make_kwargs.render_mode=rgb_array,++rl_games.gymnasium.task_contract.noop_action_id=2,++rl_games.gymnasium.task_contract.obs_fps=25.0,++rl_games.gymnasium.task_contract.registration_imports=[latency_bench.envs.gymnasium_mujoco],++rl_games.gymnasium.task_contract.state_space.labels=[cart_position,pole_angle,cart_velocity,pole_angular_velocity],++rl_games.gymnasium.task_contract.task_name=inverted_pendulum_rgb_state,++rl_games.gymnasium.task_name=inverted_pendulum_rgb_state,++rl_games.initialization_mode=scratch,++rl_games.model_alias=openvla,++rl_games.task=gymnasium,++run_id=openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2,++run_root_dir=/home/ubuntu/talha/results/Checkpoints,++run_root_dir=results/Checkpoints,++seed=42,++trainer.cc_f1_tolerance=1,++trainer.distributed_backend=none,++trainer.eval_action_classification=true,++trainer.eval_action_classification_interval=null,++trainer.eval_interval=5000,++trainer.eval_num_batches=200,++trainer.freeze_llm_layers=[],++trainer.freeze_modules=,++trainer.freeze_tied_embedding=false,++trainer.freeze_vit=false,++trainer.gradient_accumulation_steps=2,++trainer.gradient_clipping=1.0,++trainer.is_resume=false,++trainer.is_resume=false,++trainer.learning_rate.action_model=0.0001,++trainer.learning_rate.base=2e-05,++trainer.learning_rate.qwen_vl_interface=1e-05,++trainer.logging_frequency=1,++trainer.loss_scale.vla=1.0,++trainer.loss_scale.vlm=0.1,++trainer.lr_scheduler_type=cosine_with_min_lr,++trainer.max_grad_norm=1.0,++trainer.max_train_steps=5000,++trainer.num_warmup_steps=100,++trainer.optimizer.betas=[0.9,0.95],++trainer.optimizer.eps=1e-08,++trainer.optimizer.fused=true,++trainer.optimizer.name=AdamW,++trainer.optimizer.weight_decay=1e-08,++trainer.per_latency_eval_num_batches=5,++trainer.pretrained_checkpoint=null,++trainer.profile_timing.enabled=false,++trainer.profile_timing.log_interval=10,++trainer.reload_modules=null,++trainer.resume_step=0,++trainer.save_format=pt,++trainer.save_interval=5000,++trainer.scheduler_specific_kwargs.min_lr=1e-06,++trainer.weight_decay=0.0,++training_latency_condition=unspecified,++version_id=0.21,++wandb_group=openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2,++wandb_tags=[],++workspace_dir=/home/ubuntu/talha,env=gymnasium,init=scratch,mode=mixed_latency,model=openvla' + id: ??? + num: ??? + config_name: train + env_set: {} + env_copy: [] + config: + override_dirname: + kv_sep: '=' + item_sep: ',' + exclude_keys: [] + runtime: + version: 1.3.7 + version_base: '1.1' + cwd: /home/ubuntu/talha/starVLA + config_sources: + - path: hydra.conf + schema: pkg + provider: hydra + - path: /home/ubuntu/talha/starVLA/examples/rl_games/config + schema: file + provider: main + - path: '' + schema: structured + provider: schema + output_dir: /home/ubuntu/talha/results/Checkpoints/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2/hydra + choices: + cross_task_setup: null + checkpoint: default + mode: mixed_latency + init: scratch + env: gymnasium + model: openvla + hydra/env: default + hydra/callbacks: null + hydra/job_logging: default + hydra/hydra_logging: default + hydra/hydra_help: default + hydra/help: default + hydra/sweeper: basic + hydra/launcher: basic + hydra/output: default + verbose: false diff --git a/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2/hydra/.hydra/overrides.yaml b/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2/hydra/.hydra/overrides.yaml new file mode 100644 index 0000000000000000000000000000000000000000..b4d85f27ccf551b8423331246d7760e05f4b827b --- /dev/null +++ b/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2/hydra/.hydra/overrides.yaml @@ -0,0 +1,230 @@ +- model=openvla +- env=gymnasium +- init=scratch +- mode=mixed_latency +- ++framework.qwenvl.base_vlm=/home/ubuntu/talha/playground/Pretrained_models/Qwen3-VL-4B-Instruct +- ++framework.qwenvl.attn_implementation=flash_attention_2 +- ++framework.qwenvl.enable_gradient_checkpointing=true +- ++framework.action_model.state_dim=4 +- ++framework.action_model.loss_type=discrete_ce +- ++framework.action_model.action_horizon=1 +- ++framework.action_model.future_action_window_size=0 +- ++framework.action_model.past_action_window_size=0 +- ++framework.action_model.action_dim=5 +- ++framework.action_model.action_env_dim=5 +- ++framework.action_model.state_encoding=discretized_text +- ++framework.action_model.action_model_type=MLP +- ++framework.action_model.action_hidden_dim=2560 +- ++framework.kv_memory.enabled=false +- ++framework.kv_memory.window=4 +- ++framework.kv_memory.rollout_len=8 +- ++framework.kv_memory.packed_train=false +- ++framework.kv_memory.rebased_sink=true +- ++framework.name=QwenOFT +- ++datasets.vla_data.dataset_py=lerobot_datasets +- ++datasets.vla_data.include_state=true +- ++datasets.vla_data.data_root_dir=playground/Datasets/rl_games +- ++datasets.vla_data.data_mix=openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2 +- ++datasets.vla_data.eval_data_mix=openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2__val +- ++datasets.vla_data.custom_mixtures_path=null +- ++datasets.vla_data.action_type=discrete +- ++datasets.vla_data.sequential_step_sampling=false +- ++datasets.vla_data.eval_sequential_step_sampling=null +- ++datasets.vla_data.num_workers=8 +- ++datasets.vla_data.eval_num_workers=8 +- ++datasets.vla_data.prefetch_factor=4 +- ++datasets.vla_data.persistent_workers=true +- ++datasets.vla_data.pin_memory=true +- ++datasets.vla_data.shuffle=true +- ++datasets.vla_data.action_balance.enabled=false +- ++datasets.vla_data.action_balance.strategy=balanced_epoch +- ++datasets.vla_data.action_balance.action_key=action_id +- ++datasets.vla_data.action_balance.target_flap_fraction=0.3 +- ++datasets.vla_data.action_balance.noop_id=0 +- ++datasets.vla_data.action_balance.flap_id=1 +- ++datasets.vla_data.latency_curriculum.enabled=false +- ++datasets.vla_data.latency_curriculum.strategy=exclusive +- ++datasets.vla_data.latency_curriculum.latencies=null +- ++datasets.vla_data.latency_curriculum.phase_steps=null +- ++datasets.vla_data.latency_curriculum.phase_distributions=null +- ++datasets.vla_data.latency_curriculum.new_latency_passes=1.0 +- ++datasets.vla_data.latency_curriculum.replay_passes=0.25 +- ++datasets.vla_data.latency_curriculum.target_total_passes=2.0 +- ++datasets.vla_data.latency_curriculum.final_equalization=true +- ++datasets.vla_data.latency_curriculum.step_budget_mode=auto +- ++datasets.vla_data.latency_curriculum.eval_at_phase_end=false +- ++datasets.vla_data.latency_curriculum.save_at_phase_end=false +- ++datasets.vla_data.latency_curriculum.computed_plan=null +- ++datasets.vla_data.per_device_batch_size=8 +- ++datasets.vla_data.load_all_data_for_training=true +- ++datasets.vla_data.num_obs_frames=1 +- ++datasets.vla_data.image_mode=single +- ++datasets.vla_data.prompt_mode=raw +- ++datasets.vla_data.stitch_grid=[2,2] +- ++datasets.vla_data.obs_image_size=null +- ++datasets.vla_data.video_backend=torchvision_av +- '++datasets.vla_data.gymnasium_task_contract={task_name:inverted_pendulum_rgb_state,env_id:LatencyBench/InvertedPendulumDiscrete-v0,registration_imports:[latency_bench.envs.gymnasium_mujoco],env_fps:25.0,obs_fps:25.0,frame_stack:1,base_prompt:''Keep + the pole upright by applying horizontal cart forces. Choose exactly one action from: + push_left_hard, push_left_soft, zero_force, push_right_soft, push_right_hard.'',action_labels:[push_left_hard,push_left_soft,zero_force,push_right_soft,push_right_hard],action_values:[0,1,2,3,4],noop_action_id:2,make_kwargs:{base_env_id:InvertedPendulum-v4,render_mode:rgb_array,force_values:[-3.0,-1.5,0.0,1.5,3.0],base_make_kwargs:{}},state_space:{labels:[cart_position,pole_angle,cart_velocity,pole_angular_velocity]}}' +- ++datasets.vla_data.active_action_dim=5 +- ++dataset.source_hf=latency-sensitive-bench/inverted_pendulum_200ep_2 +- ++dataset.config_name=null +- ++dataset.source_subdir='inverted_pendulum_fixed_latency_0_100ep_1ksteps_og,inverted_pendulum_fixed_latency_2_1000ep_7k2steps,inverted_pendulum_fixed_latency_4_1000ep_7k2steps' +- ++dataset.converted_name=openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2 +- ++dataset.single_source_hf= +- ++dataset.mixed_source_hf= +- ++dataset.single_converted_name=flappy_train +- ++dataset.mixed_converted_name=flappy_train +- ++dataset.single_latency_filter=null +- ++dataset.mixed_latency_filter=null +- ++dataset.force_download=false +- ++dataset.setup_force=true +- ++dataset.skip_verification=true +- ++dataset.target_latency_unit=observation_steps +- ++dataset.verify_rows=200 +- ++dataset.max_episodes=null +- ++dataset.max_steps_per_episode=1000 +- ++dataset.episodes_per_latency=null +- ++dataset.episodes_per_latency_by_latency='0:90,2:1000,4:1000' +- ++dataset.latency_filter=[0,2,4] +- ++dataset.debug_subset.enabled=false +- ++dataset.debug_subset.max_episodes=5 +- ++dataset.debug_subset.suffix=debug +- ++base_model.repo_id=Qwen/Qwen3-VL-4B-Instruct +- ++initialization.checkpoint_local_dir=null +- ++initialization.checkpoint_hf_repo_id=null +- ++initialization.checkpoint_filename=null +- ++trainer.max_train_steps=5000 +- ++trainer.num_warmup_steps=100 +- ++trainer.save_interval=5000 +- ++trainer.eval_interval=5000 +- ++trainer.eval_num_batches=200 +- ++trainer.per_latency_eval_num_batches=5 +- ++trainer.eval_action_classification=true +- ++trainer.eval_action_classification_interval=null +- ++trainer.cc_f1_tolerance=1 +- ++trainer.learning_rate.base=2e-05 +- ++trainer.learning_rate.qwen_vl_interface=1e-05 +- ++trainer.learning_rate.action_model=0.0001 +- ++trainer.lr_scheduler_type=cosine_with_min_lr +- ++trainer.scheduler_specific_kwargs.min_lr=1e-06 +- ++trainer.freeze_modules= +- ++trainer.freeze_vit=false +- ++trainer.freeze_tied_embedding=false +- ++trainer.freeze_llm_layers=[] +- ++trainer.loss_scale.vla=1.0 +- ++trainer.loss_scale.vlm=0.1 +- ++trainer.max_grad_norm=1.0 +- ++trainer.weight_decay=0.0 +- ++trainer.logging_frequency=1 +- ++trainer.profile_timing.enabled=false +- ++trainer.profile_timing.log_interval=10 +- ++trainer.gradient_clipping=1.0 +- ++trainer.gradient_accumulation_steps=2 +- ++trainer.distributed_backend=none +- ++trainer.is_resume=false +- ++trainer.pretrained_checkpoint=null +- ++trainer.resume_step=0 +- ++trainer.reload_modules=null +- ++trainer.optimizer.name=AdamW +- ++trainer.optimizer.betas=[0.9,0.95] +- ++trainer.optimizer.eps=1e-08 +- ++trainer.optimizer.weight_decay=1e-08 +- ++trainer.optimizer.fused=true +- ++trainer.save_format=pt +- ++workspace_dir=/home/ubuntu/talha +- ++run_root_dir=results/Checkpoints +- ++seed=42 +- ++wandb_group=openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2 +- ++wandb_tags=[] +- ++training_latency_condition=unspecified +- ++auth.env_file=null +- ++auth.hf_token_env=HF_TOKEN +- ++auth.wandb_api_key_env=WANDB_API_KEY +- ++paths.run_root_dir=results/Checkpoints +- ++paths.dataset_local_dir=playground/Datasets/rl_games +- ++paths.dataset_cache_dir=null +- ++paths.base_model_dir=/home/ubuntu/talha/playground/Pretrained_models/Qwen3-VL-4B-Instruct +- ++paths.accelerate_config=starVLA/config/deepseeds/deepspeed_zero2.yaml +- ++rl_games.model_alias=openvla +- ++rl_games.env_eval.image_size=224 +- ++rl_games.env_eval.frameskip=1 +- ++rl_games.env_eval.image_transform=raw_rgb +- ++rl_games.env_eval.prompt_mode=raw +- ++rl_games.env_eval.ghost_trail.history_frames=5 +- ++rl_games.env_eval.ghost_trail.gamma=1.3 +- ++rl_games.env_eval.ghost_trail.min_alpha=35 +- ++rl_games.env_eval.ghost_trail.scroll_px_per_step=4.0 +- ++rl_games.env_eval.ghost_trail.ground_fraction=0.22 +- ++rl_games.env_eval.seed=42 +- ++rl_games.env_eval.fixed_episode_seeds=true +- ++rl_games.env_eval.latency_seed_stride=0 +- ++rl_games.env_eval.task_seed_stride=0 +- ++rl_games.env_eval.task_description= +- ++rl_games.env_eval.eval_parallel_envs=5 +- ++rl_games.env_eval.action_chunk_execution.enabled=false +- ++rl_games.env_eval.action_chunk_execution.chunk_size=null +- ++rl_games.env_eval.enabled=true +- ++rl_games.env_eval.mid_train.enabled=false +- ++rl_games.env_eval.mid_train.interval_steps=100 +- ++rl_games.env_eval.mid_train.latencies=[0,2,4] +- ++rl_games.env_eval.mid_train.num_episodes=5 +- ++rl_games.env_eval.mid_train.max_steps_per_episode=3600 +- ++rl_games.env_eval.post_train.enabled=true +- ++rl_games.env_eval.post_train.latencies=[0] +- ++rl_games.env_eval.post_train.latencies_from_prompt_map=false +- ++rl_games.env_eval.post_train.num_episodes=20 +- ++rl_games.env_eval.post_train.max_steps_per_episode=3600 +- ++rl_games.env_eval.eval_backend=latency_bench +- ++rl_games.env_eval.distributed_mode=none +- ++rl_games.env_eval.vectorized.enabled=false +- ++rl_games.env_eval.vectorized.batch_size=1 +- ++rl_games.env_eval.latency.prompt_map_path=null +- ++rl_games.env_eval.latency.mode=mixed +- ++rl_games.env_eval.latency.values=[0] +- ++rl_games.task=gymnasium +- ++rl_games.gymnasium.task_contract.task_name=inverted_pendulum_rgb_state +- ++rl_games.gymnasium.task_contract.env_id=LatencyBench/InvertedPendulumDiscrete-v0 +- ++rl_games.gymnasium.task_contract.registration_imports=[latency_bench.envs.gymnasium_mujoco] +- ++rl_games.gymnasium.task_contract.env_fps=25.0 +- ++rl_games.gymnasium.task_contract.obs_fps=25.0 +- ++rl_games.gymnasium.task_contract.frame_stack=1 +- '++rl_games.gymnasium.task_contract.base_prompt=''Keep the pole upright by applying + horizontal cart forces. Choose exactly one action from: push_left_hard, push_left_soft, + zero_force, push_right_soft, push_right_hard.''' +- ++rl_games.gymnasium.task_contract.action_labels=[push_left_hard,push_left_soft,zero_force,push_right_soft,push_right_hard] +- ++rl_games.gymnasium.task_contract.action_values=[0,1,2,3,4] +- ++rl_games.gymnasium.task_contract.noop_action_id=2 +- ++rl_games.gymnasium.task_contract.make_kwargs.base_env_id=InvertedPendulum-v4 +- ++rl_games.gymnasium.task_contract.make_kwargs.render_mode=rgb_array +- ++rl_games.gymnasium.task_contract.make_kwargs.force_values=[-3.0,-1.5,0.0,1.5,3.0] +- ++rl_games.gymnasium.task_contract.state_space.labels=[cart_position,pole_angle,cart_velocity,pole_angular_velocity] +- ++rl_games.gymnasium.task_name=inverted_pendulum_rgb_state +- ++rl_games.initialization_mode=scratch +- ++rl_games.action_carrier=native +- ++checkpoint.load=none +- ++checkpoint.hf_repo_id=talha15032/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2 +- ++checkpoint.save_best_model=false +- ++checkpoint.save_final_model=true +- ++checkpoint.save_pt_file=true +- ++checkpoint.save_training_state=false +- ++checkpoint.save_safetensors_file=false +- ++checkpoint.local.keep_last_n=1 +- ++checkpoint.sync.enabled=true +- ++checkpoint.sync.repo_id=talha15032/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2 +- ++checkpoint.sync.keep_last_n=0 +- ++checkpoint.sync.sync_every_n_checkpoints=1 +- ++checkpoint.sync.resume_policy=local_latest +- ++run_id=openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2 +- ++output_dir=null +- ++config_yaml=null +- ++is_debug=false +- ++version_id=0.21 +- ++run_root_dir=/home/ubuntu/talha/results/Checkpoints +- ++trainer.is_resume=false +- ++datasets.vla_data.data_root_dir=/home/ubuntu/talha/playground/Datasets/rl_games +- ++datasets.vla_data.data_mix=openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2 +- ++datasets.vla_data.eval_data_mix=openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2__val +- ++datasets.vla_data.custom_mixtures_path=/home/ubuntu/talha/playground/Datasets/rl_games/_generated_mixtures/single_77a677bfbf8e73b8.json +- ++framework.qwenvl.base_vlm=/home/ubuntu/talha/playground/Pretrained_models/Qwen3-VL-4B-Instruct +- ++rl_games.env_eval.latency.prompt_map_path=/home/ubuntu/talha/playground/Datasets/rl_games/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2/latency_prompt_map.json diff --git a/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2/hydra/train_starvla_hydra.log b/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2/hydra/train_starvla_hydra.log new file mode 100644 index 0000000000000000000000000000000000000000..a651759cc3480799333355dcd0aaf756d6b3c1e8 --- /dev/null +++ b/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2/hydra/train_starvla_hydra.log @@ -0,0 +1,5065 @@ +[2026-10-03 13:00:02,801][starVLA.model.framework.share_tools][INFO] - [*] [apply_config_compat] normalised config from version_id=0.21 to '0.21' +[2026-10-03 13:00:02,801][starVLA.training.train_starvla][INFO] - VLA Training :: Warming Up +[2026-10-03 13:00:02,815][starVLA.training.train_starvla][INFO] - ✅ Configuration wrapped for access tracking +[2026-10-03 13:00:03,750][starVLA.training.train_starvla][INFO] - Creating VLA Dataset with Mixture `openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2` +[2026-10-03 13:00:04,126][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Behavior (data_config): ['BEHAVIOR_challenge'] +[2026-10-03 13:00:04,127][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from DOMINO: ['robotwin'] +[2026-10-03 13:00:04,127][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from DOMINO (data_config): [] +[2026-10-03 13:00:04,127][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from DOMINO (data_config): ['domino', 'domino_clean', 'domino_random', 'domino_cotrain'] +[2026-10-03 13:00:04,127][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Franka: ['custom_robot_config', 'demo_sim_franka_delta_joints', 'SO101'] +[2026-10-03 13:00:04,127][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Franka (data_config): [] +[2026-10-03 13:00:04,127][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Franka (data_config): ['custom_dataset', 'custom_dataset_2', 'demo_sim_pick_place', 'SO101_pick'] +[2026-10-03 13:00:04,127][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from LIBERO: ['libero_franka'] +[2026-10-03 13:00:04,127][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from LIBERO (data_config): [] +[2026-10-03 13:00:04,127][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from LIBERO (data_config): ['libero_all', 'libero_goal', 'multi_robot'] +[2026-10-03 13:00:04,127][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from MIKASA: ['mikasa_franka_h1'] +[2026-10-03 13:00:04,127][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from MIKASA (data_config): ['mikasa_franka_h1'] +[2026-10-03 13:00:04,127][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from MIKASA (data_config): ['local/intercept_grab_fast_vla_v0_h1_train'] +[2026-10-03 13:00:04,127][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from RoboChallenge_table30v2: ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-10-03 13:00:04,127][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from RoboChallenge_table30v2 (data_config): ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-10-03 13:00:04,127][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from RoboChallenge_table30v2 (data_config): ['robochallenge_table30v2_shred_paper', 'robochallenge_table30v2_ur5_all', 'robochallenge_table30v2_arx5_all', 'robochallenge_table30v2_dosw1_all'] +[2026-10-03 13:00:04,127][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_365: ['panda_omron_robocasa365'] +[2026-10-03 13:00:04,127][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_365 (data_config): [] +[2026-10-03 13:00:04,128][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_365 (data_config): ['robocasa365_open_drawer_target_human', 'robocasa365_atomic_target_human_all', 'robocasa365_composite_target_human_all', 'robocasa365_target_human_all'] +[2026-10-03 13:00:04,128][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_tabletop: ['fourier_gr1_arms_waist'] +[2026-10-03 13:00:04,128][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_tabletop (data_config): [] +[2026-10-03 13:00:04,128][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_tabletop (data_config): ['fourier_gr1_unified_1000'] +[2026-10-03 13:00:04,128][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robotwin: ['robotwin', 'robotwin50', 'arx_x5'] +[2026-10-03 13:00:04,128][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robotwin (data_config): [] +[2026-10-03 13:00:04,128][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robotwin (data_config): ['robotwin_all', 'robotwin_all_50', 'robotwin', 'robotwin_task1', 'robotwin_task2', 'arx_x5'] +[2026-10-03 13:00:04,128][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from SimplerEnv: ['oxe_droid', 'oxe_bridge', 'oxe_rt1'] +[2026-10-03 13:00:04,128][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from SimplerEnv (data_config): [] +[2026-10-03 13:00:04,128][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from SimplerEnv (data_config): ['bridge', 'bridge_rt_1'] +[2026-10-03 13:00:04,128][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from VLA-Arena: ['vla_arena_franka'] +[2026-10-03 13:00:04,128][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from VLA-Arena (data_config): [] +[2026-10-03 13:00:04,128][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from VLA-Arena (data_config): ['vla_arena_L0_S', 'vla_arena_L0_M', 'vla_arena_L0_L'] +[2026-10-03 13:00:04,128][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from rl_games: ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_defend_the_line', 'rl_games_deadly_corridor', 'rl_games_asterix', 'rl_games_atlantis', 'rl_games_gymnasium', 'rl_games_gymnasium_discrete', 'rl_games_gymnasium_native'] +[2026-10-03 13:00:04,128][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from rl_games (data_config): ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_defend_the_line', 'rl_games_deadly_corridor', 'rl_games_asterix', 'rl_games_atlantis', 'rl_games_gymnasium', 'rl_games_gymnasium_discrete', 'rl_games_gymnasium_native'] +[2026-10-03 13:00:04,128][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from rl_games (data_config): ['flappy_train', 'flappy_train__bridge', 'flappy_mixed_latency_train', 'flappy_mixed_latency_train__bridge', 'demon_attack_train', 'demon_attack_train__bridge', 'demon_attack_mixed_latency_train', 'demon_attack_mixed_latency_train__bridge', 'defend_the_line_train', 'defend_the_line_train__bridge', 'defend_the_line_mixed_latency_train', 'defend_the_line_mixed_latency_train__bridge', 'deadly_corridor_train', 'deadly_corridor_train__bridge', 'deadly_corridor_mixed_latency_train', 'deadly_corridor_mixed_latency_train__bridge', 'asterix_train', 'asterix_train__bridge', 'asterix_mixed_latency_train', 'asterix_mixed_latency_train__bridge', 'atlantis_train', 'atlantis_train__bridge', 'atlantis_mixed_latency_train', 'atlantis_mixed_latency_train__bridge', 'inverted_pendulum_mixed_latency_train', 'inverted_pendulum_mixed_latency_train__bridge', 'inverted_pendulum_mixed_latency_train_no_latency_prompt', 'inverted_pendulum_mixed_latency_train_no_latency_prompt__bridge', 'h1hand_balance_hard'] +[2026-10-03 13:00:04,136][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded custom mixtures from /home/ubuntu/talha/playground/Datasets/rl_games/_generated_mixtures/single_77a677bfbf8e73b8.json: ['openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2', 'openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2__val'] +[2026-10-03 13:00:04,156][starVLA.dataloader][INFO] - Created train VLA dataloader with num_workers=8, persistent_workers=True, cpu_only_workers=True +[2026-10-03 13:00:04,156][starVLA.training.train_starvla][INFO] - Creating VLA Eval Dataset with Mixture `openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2__val` +[2026-10-03 13:00:04,163][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded custom mixtures from /home/ubuntu/talha/playground/Datasets/rl_games/_generated_mixtures/single_77a677bfbf8e73b8.json: ['openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2', 'openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2__val'] +[2026-10-03 13:00:04,167][starVLA.dataloader][INFO] - Created eval VLA dataloader with num_workers=8, persistent_workers=True, cpu_only_workers=True +[2026-10-03 13:00:04,221][starVLA.training.train_starvla][INFO] - 📝 Full config saved at /home/ubuntu/talha/results/Checkpoints/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2/config.full.yaml +[2026-10-03 13:00:04,231][starVLA.training.train_starvla][INFO] - 📊 Accessed config snapshot saved at /home/ubuntu/talha/results/Checkpoints/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2/config.yaml +[2026-10-03 13:00:04,231][starVLA.training.train_starvla][INFO] - No pretrained checkpoint provided. Starting training from scratch. +[2026-10-03 13:00:05,987][starVLA.training.train_starvla][INFO] - ***** Training Configuration ***** +[2026-10-03 13:00:05,988][starVLA.training.train_starvla][INFO] - Total optimization steps = 5000 +[2026-10-03 13:00:05,988][starVLA.training.train_starvla][INFO] - Per device batch size = 8 +[2026-10-03 13:00:05,988][starVLA.training.train_starvla][INFO] - Gradient accumulation steps = 2 +[2026-10-03 13:00:05,988][starVLA.training.train_starvla][INFO] - Total batch size = 16 +[2026-10-03 13:00:31,222][starVLA.training.train_starvla][INFO] - Step 1, Loss: {'train/loss': 1.7027273178100586, 'train/grad_norm_pre_clip': 120.72057342529297, 'timing/data': 0.00044099999649915844, 'timing/model': 0.48191684699850157, 'learning_rate/qwen_vl_interface': 1.0000000000000001e-07, 'learning_rate/action_model': 1.0000000000000002e-06, 'epoch': 0.0}) +[2026-10-03 13:00:31,938][starVLA.training.train_starvla][INFO] - Step 2, Loss: {'train/loss': 1.6730077266693115, 'train/grad_norm_pre_clip': 176.24241638183594, 'timing/data': 0.0002845699928002432, 'timing/model': 0.397512994997669, 'learning_rate/qwen_vl_interface': 2.0000000000000002e-07, 'learning_rate/action_model': 2.0000000000000003e-06, 'epoch': 0.0}) +[2026-10-03 13:00:32,647][starVLA.training.train_starvla][INFO] - Step 3, Loss: {'train/loss': 1.487200140953064, 'train/grad_norm_pre_clip': 102.98754119873047, 'timing/data': 0.0003516200085869059, 'timing/model': 0.3958668559935177, 'learning_rate/qwen_vl_interface': 3.0000000000000004e-07, 'learning_rate/action_model': 3e-06, 'epoch': 0.0}) +[2026-10-03 13:00:33,352][starVLA.training.train_starvla][INFO] - Step 4, Loss: {'train/loss': 1.4558849334716797, 'train/grad_norm_pre_clip': 74.97704315185547, 'timing/data': 0.0002341899962630123, 'timing/model': 0.3945402660028776, 'learning_rate/qwen_vl_interface': 4.0000000000000003e-07, 'learning_rate/action_model': 4.000000000000001e-06, 'epoch': 0.0}) +[2026-10-03 13:00:34,060][starVLA.training.train_starvla][INFO] - Step 5, Loss: {'train/loss': 1.1396434307098389, 'train/grad_norm_pre_clip': 35.84294891357422, 'timing/data': 0.00020103999122511595, 'timing/model': 0.396847665993846, 'learning_rate/qwen_vl_interface': 5.000000000000001e-07, 'learning_rate/action_model': 5e-06, 'epoch': 0.0}) +[2026-10-03 13:00:34,954][starVLA.training.train_starvla][INFO] - Step 6, Loss: {'train/loss': 1.0874924659729004, 'train/grad_norm_pre_clip': 35.19933319091797, 'timing/data': 0.00024670000129844993, 'timing/model': 0.5816438049951103, 'learning_rate/qwen_vl_interface': 6.000000000000001e-07, 'learning_rate/action_model': 6e-06, 'epoch': 0.0}) +[2026-10-03 13:00:35,662][starVLA.training.train_starvla][INFO] - Step 7, Loss: {'train/loss': 1.4500700235366821, 'train/grad_norm_pre_clip': 61.82386016845703, 'timing/data': 0.00024553999537602067, 'timing/model': 0.3952748659939971, 'learning_rate/qwen_vl_interface': 7.000000000000001e-07, 'learning_rate/action_model': 7.000000000000001e-06, 'epoch': 0.0}) +[2026-10-03 13:00:36,370][starVLA.training.train_starvla][INFO] - Step 8, Loss: {'train/loss': 0.6299459338188171, 'train/grad_norm_pre_clip': 12.166075706481934, 'timing/data': 0.0003738599916687235, 'timing/model': 0.3955336760118371, 'learning_rate/qwen_vl_interface': 8.000000000000001e-07, 'learning_rate/action_model': 8.000000000000001e-06, 'epoch': 0.0}) +[2026-10-03 13:00:37,077][starVLA.training.train_starvla][INFO] - Step 9, Loss: {'train/loss': 1.0842801332473755, 'train/grad_norm_pre_clip': 50.908775329589844, 'timing/data': 0.00022047999664209783, 'timing/model': 0.39625231600075494, 'learning_rate/qwen_vl_interface': 9.000000000000001e-07, 'learning_rate/action_model': 9e-06, 'epoch': 0.0}) +[2026-10-03 13:00:37,779][starVLA.training.train_starvla][INFO] - Step 10, Loss: {'train/loss': 1.1383869647979736, 'train/grad_norm_pre_clip': 32.750579833984375, 'timing/data': 0.000444909994257614, 'timing/model': 0.39327043599041644, 'learning_rate/qwen_vl_interface': 1.0000000000000002e-06, 'learning_rate/action_model': 1e-05, 'epoch': 0.0}) +[2026-10-03 13:00:38,678][starVLA.training.train_starvla][INFO] - Step 11, Loss: {'train/loss': 0.8533065915107727, 'train/grad_norm_pre_clip': 29.886192321777344, 'timing/data': 0.0003914499975508079, 'timing/model': 0.5879957940051099, 'learning_rate/qwen_vl_interface': 1.1e-06, 'learning_rate/action_model': 1.1000000000000001e-05, 'epoch': 0.0}) +[2026-10-03 13:00:39,382][starVLA.training.train_starvla][INFO] - Step 12, Loss: {'train/loss': 0.6908367872238159, 'train/grad_norm_pre_clip': 46.449337005615234, 'timing/data': 0.00019561000226531178, 'timing/model': 0.39269446600519586, 'learning_rate/qwen_vl_interface': 1.2000000000000002e-06, 'learning_rate/action_model': 1.2e-05, 'epoch': 0.0}) +[2026-10-03 13:00:40,087][starVLA.training.train_starvla][INFO] - Step 13, Loss: {'train/loss': 1.1050795316696167, 'train/grad_norm_pre_clip': 47.162559509277344, 'timing/data': 0.0003197690093657002, 'timing/model': 0.39407596600358374, 'learning_rate/qwen_vl_interface': 1.3e-06, 'learning_rate/action_model': 1.3000000000000001e-05, 'epoch': 0.0}) +[2026-10-03 13:00:40,788][starVLA.training.train_starvla][INFO] - Step 14, Loss: {'train/loss': 0.6362172365188599, 'train/grad_norm_pre_clip': 44.13529586791992, 'timing/data': 0.00032233999809250236, 'timing/model': 0.392102605997934, 'learning_rate/qwen_vl_interface': 1.4000000000000001e-06, 'learning_rate/action_model': 1.4000000000000001e-05, 'epoch': 0.0}) +[2026-10-03 13:00:41,491][starVLA.training.train_starvla][INFO] - Step 15, Loss: {'train/loss': 1.3478400707244873, 'train/grad_norm_pre_clip': 67.4061050415039, 'timing/data': 0.00019038999744225293, 'timing/model': 0.3932254359970102, 'learning_rate/qwen_vl_interface': 1.5e-06, 'learning_rate/action_model': 1.5e-05, 'epoch': 0.0}) +[2026-10-03 13:00:42,194][starVLA.training.train_starvla][INFO] - Step 16, Loss: {'train/loss': 0.8154697418212891, 'train/grad_norm_pre_clip': 29.925201416015625, 'timing/data': 0.00023449001309927553, 'timing/model': 0.3927876060042763, 'learning_rate/qwen_vl_interface': 1.6000000000000001e-06, 'learning_rate/action_model': 1.6000000000000003e-05, 'epoch': 0.0}) +[2026-10-03 13:00:43,089][starVLA.training.train_starvla][INFO] - Step 17, Loss: {'train/loss': 1.4463181495666504, 'train/grad_norm_pre_clip': 80.39584350585938, 'timing/data': 0.00021856000239495188, 'timing/model': 0.3941218359977938, 'learning_rate/qwen_vl_interface': 1.7000000000000002e-06, 'learning_rate/action_model': 1.7000000000000003e-05, 'epoch': 0.0}) +[2026-10-03 13:00:43,792][starVLA.training.train_starvla][INFO] - Step 18, Loss: {'train/loss': 1.1740593910217285, 'train/grad_norm_pre_clip': 113.10293579101562, 'timing/data': 0.0002416500065010041, 'timing/model': 0.39356917700206395, 'learning_rate/qwen_vl_interface': 1.8000000000000001e-06, 'learning_rate/action_model': 1.8e-05, 'epoch': 0.0}) +[2026-10-03 13:00:44,498][starVLA.training.train_starvla][INFO] - Step 19, Loss: {'train/loss': 1.252253770828247, 'train/grad_norm_pre_clip': 115.10820770263672, 'timing/data': 0.00027898000553250313, 'timing/model': 0.39352218600106426, 'learning_rate/qwen_vl_interface': 1.9000000000000002e-06, 'learning_rate/action_model': 1.9e-05, 'epoch': 0.0}) +[2026-10-03 13:00:45,201][starVLA.training.train_starvla][INFO] - Step 20, Loss: {'train/loss': 0.9487059712409973, 'train/grad_norm_pre_clip': 116.11084747314453, 'timing/data': 0.00029861999792046845, 'timing/model': 0.3920199370040791, 'learning_rate/qwen_vl_interface': 2.0000000000000003e-06, 'learning_rate/action_model': 2e-05, 'epoch': 0.0}) +[2026-10-03 13:00:45,904][starVLA.training.train_starvla][INFO] - Step 21, Loss: {'train/loss': 0.7594298124313354, 'train/grad_norm_pre_clip': 57.035980224609375, 'timing/data': 0.00026560999685898423, 'timing/model': 0.3926013259915635, 'learning_rate/qwen_vl_interface': 2.1000000000000002e-06, 'learning_rate/action_model': 2.1e-05, 'epoch': 0.0}) +[2026-10-03 13:00:46,800][starVLA.training.train_starvla][INFO] - Step 22, Loss: {'train/loss': 1.4672693014144897, 'train/grad_norm_pre_clip': 122.2135238647461, 'timing/data': 0.00023871999292168766, 'timing/model': 0.5862795650027692, 'learning_rate/qwen_vl_interface': 2.2e-06, 'learning_rate/action_model': 2.2000000000000003e-05, 'epoch': 0.0}) +[2026-10-03 13:00:47,506][starVLA.training.train_starvla][INFO] - Step 23, Loss: {'train/loss': 1.1046638488769531, 'train/grad_norm_pre_clip': 92.81817626953125, 'timing/data': 0.0003598500043153763, 'timing/model': 0.3930375060008373, 'learning_rate/qwen_vl_interface': 2.3000000000000004e-06, 'learning_rate/action_model': 2.3000000000000003e-05, 'epoch': 0.0}) +[2026-10-03 13:00:48,202][starVLA.training.train_starvla][INFO] - Step 24, Loss: {'train/loss': 1.4500499963760376, 'train/grad_norm_pre_clip': 128.88494873046875, 'timing/data': 0.00019813000108115375, 'timing/model': 0.38823544699698687, 'learning_rate/qwen_vl_interface': 2.4000000000000003e-06, 'learning_rate/action_model': 2.4e-05, 'epoch': 0.0}) +[2026-10-03 13:00:48,898][starVLA.training.train_starvla][INFO] - Step 25, Loss: {'train/loss': 1.0666602849960327, 'train/grad_norm_pre_clip': 47.202728271484375, 'timing/data': 0.0002638099977048114, 'timing/model': 0.3883091669995338, 'learning_rate/qwen_vl_interface': 2.5e-06, 'learning_rate/action_model': 2.5e-05, 'epoch': 0.0}) +[2026-10-03 13:00:49,592][starVLA.training.train_starvla][INFO] - Step 26, Loss: {'train/loss': 2.312594413757324, 'train/grad_norm_pre_clip': 47.80818557739258, 'timing/data': 0.0002910900075221434, 'timing/model': 0.38810359600756783, 'learning_rate/qwen_vl_interface': 2.6e-06, 'learning_rate/action_model': 2.6000000000000002e-05, 'epoch': 0.0}) +[2026-10-03 13:00:50,474][starVLA.training.train_starvla][INFO] - Step 27, Loss: {'train/loss': 1.182059645652771, 'train/grad_norm_pre_clip': 77.99227905273438, 'timing/data': 0.00028021000616718084, 'timing/model': 0.5766612659936072, 'learning_rate/qwen_vl_interface': 2.7000000000000004e-06, 'learning_rate/action_model': 2.7000000000000002e-05, 'epoch': 0.0}) +[2026-10-03 13:00:51,171][starVLA.training.train_starvla][INFO] - Step 28, Loss: {'train/loss': 1.1124004125595093, 'train/grad_norm_pre_clip': 79.77297973632812, 'timing/data': 0.00020763000065926462, 'timing/model': 0.38830017700092867, 'learning_rate/qwen_vl_interface': 2.8000000000000003e-06, 'learning_rate/action_model': 2.8000000000000003e-05, 'epoch': 0.0}) +[2026-10-03 13:00:51,869][starVLA.training.train_starvla][INFO] - Step 29, Loss: {'train/loss': 1.0034446716308594, 'train/grad_norm_pre_clip': 105.2850112915039, 'timing/data': 0.0002568800118751824, 'timing/model': 0.3880314670095686, 'learning_rate/qwen_vl_interface': 2.9e-06, 'learning_rate/action_model': 2.9e-05, 'epoch': 0.0}) +[2026-10-03 13:00:52,561][starVLA.training.train_starvla][INFO] - Step 30, Loss: {'train/loss': 0.38774555921554565, 'train/grad_norm_pre_clip': 103.01605224609375, 'timing/data': 0.0004187199956504628, 'timing/model': 0.387083938010619, 'learning_rate/qwen_vl_interface': 3e-06, 'learning_rate/action_model': 3e-05, 'epoch': 0.0}) +[2026-10-03 13:00:53,254][starVLA.training.train_starvla][INFO] - Step 31, Loss: {'train/loss': 0.9010929465293884, 'train/grad_norm_pre_clip': 53.48701858520508, 'timing/data': 0.0002523700095480308, 'timing/model': 0.38838655600557104, 'learning_rate/qwen_vl_interface': 3.1000000000000004e-06, 'learning_rate/action_model': 3.1e-05, 'epoch': 0.0}) +[2026-10-03 13:00:53,951][starVLA.training.train_starvla][INFO] - Step 32, Loss: {'train/loss': 2.0056676864624023, 'train/grad_norm_pre_clip': 167.4644775390625, 'timing/data': 0.00020867999410256743, 'timing/model': 0.3878426469891565, 'learning_rate/qwen_vl_interface': 3.2000000000000003e-06, 'learning_rate/action_model': 3.2000000000000005e-05, 'epoch': 0.0}) +[2026-10-03 13:00:54,836][starVLA.training.train_starvla][INFO] - Step 33, Loss: {'train/loss': 2.1864607334136963, 'train/grad_norm_pre_clip': 218.43577575683594, 'timing/data': 0.0003603200020734221, 'timing/model': 0.3876216759963427, 'learning_rate/qwen_vl_interface': 3.3000000000000006e-06, 'learning_rate/action_model': 3.3e-05, 'epoch': 0.0}) +[2026-10-03 13:00:55,529][starVLA.training.train_starvla][INFO] - Step 34, Loss: {'train/loss': 1.343637228012085, 'train/grad_norm_pre_clip': 131.1472930908203, 'timing/data': 0.0002930100017692894, 'timing/model': 0.38758794699970167, 'learning_rate/qwen_vl_interface': 3.4000000000000005e-06, 'learning_rate/action_model': 3.4000000000000007e-05, 'epoch': 0.0}) +[2026-10-03 13:00:56,222][starVLA.training.train_starvla][INFO] - Step 35, Loss: {'train/loss': 1.3005155324935913, 'train/grad_norm_pre_clip': 58.286895751953125, 'timing/data': 0.00030597001023124903, 'timing/model': 0.3875521669979207, 'learning_rate/qwen_vl_interface': 3.5e-06, 'learning_rate/action_model': 3.5e-05, 'epoch': 0.0}) +[2026-10-03 13:00:56,919][starVLA.training.train_starvla][INFO] - Step 36, Loss: {'train/loss': 0.7799490690231323, 'train/grad_norm_pre_clip': 57.843109130859375, 'timing/data': 0.0003111300029559061, 'timing/model': 0.38765991700347513, 'learning_rate/qwen_vl_interface': 3.6000000000000003e-06, 'learning_rate/action_model': 3.6e-05, 'epoch': 0.0}) +[2026-10-03 13:00:57,613][starVLA.training.train_starvla][INFO] - Step 37, Loss: {'train/loss': 0.9729896783828735, 'train/grad_norm_pre_clip': 115.36251831054688, 'timing/data': 0.0002436700015095994, 'timing/model': 0.3878367870056536, 'learning_rate/qwen_vl_interface': 3.7e-06, 'learning_rate/action_model': 3.7e-05, 'epoch': 0.0}) +[2026-10-03 13:00:58,499][starVLA.training.train_starvla][INFO] - Step 38, Loss: {'train/loss': 1.8260565996170044, 'train/grad_norm_pre_clip': 155.01011657714844, 'timing/data': 0.00026315001014154404, 'timing/model': 0.3873774860112462, 'learning_rate/qwen_vl_interface': 3.8000000000000005e-06, 'learning_rate/action_model': 3.8e-05, 'epoch': 0.0}) +[2026-10-03 13:00:59,192][starVLA.training.train_starvla][INFO] - Step 39, Loss: {'train/loss': 1.036700963973999, 'train/grad_norm_pre_clip': 86.01849365234375, 'timing/data': 0.00030502000299748033, 'timing/model': 0.3876744670124026, 'learning_rate/qwen_vl_interface': 3.900000000000001e-06, 'learning_rate/action_model': 3.9000000000000006e-05, 'epoch': 0.0}) +[2026-10-03 13:00:59,888][starVLA.training.train_starvla][INFO] - Step 40, Loss: {'train/loss': 1.3172252178192139, 'train/grad_norm_pre_clip': 72.07511901855469, 'timing/data': 0.00024680999922566116, 'timing/model': 0.3877098569937516, 'learning_rate/qwen_vl_interface': 4.000000000000001e-06, 'learning_rate/action_model': 4e-05, 'epoch': 0.0}) +[2026-10-03 13:01:00,581][starVLA.training.train_starvla][INFO] - Step 41, Loss: {'train/loss': 1.073251485824585, 'train/grad_norm_pre_clip': 62.856788635253906, 'timing/data': 0.0003033300017705187, 'timing/model': 0.38809902699722443, 'learning_rate/qwen_vl_interface': 4.1e-06, 'learning_rate/action_model': 4.1e-05, 'epoch': 0.0}) +[2026-10-03 13:01:01,273][starVLA.training.train_starvla][INFO] - Step 42, Loss: {'train/loss': 0.9822144508361816, 'train/grad_norm_pre_clip': 47.28632736206055, 'timing/data': 0.0003351600025780499, 'timing/model': 0.38673632699646987, 'learning_rate/qwen_vl_interface': 4.2000000000000004e-06, 'learning_rate/action_model': 4.2e-05, 'epoch': 0.0}) +[2026-10-03 13:01:02,157][starVLA.training.train_starvla][INFO] - Step 43, Loss: {'train/loss': 1.1221439838409424, 'train/grad_norm_pre_clip': 35.90151596069336, 'timing/data': 0.00019560998771339655, 'timing/model': 0.5803145359968767, 'learning_rate/qwen_vl_interface': 4.3e-06, 'learning_rate/action_model': 4.3e-05, 'epoch': 0.0}) +[2026-10-03 13:01:02,852][starVLA.training.train_starvla][INFO] - Step 44, Loss: {'train/loss': 0.964195966720581, 'train/grad_norm_pre_clip': 83.84492492675781, 'timing/data': 0.0003323499986436218, 'timing/model': 0.3870074959995691, 'learning_rate/qwen_vl_interface': 4.4e-06, 'learning_rate/action_model': 4.4000000000000006e-05, 'epoch': 0.0}) +[2026-10-03 13:01:03,545][starVLA.training.train_starvla][INFO] - Step 45, Loss: {'train/loss': 1.720271110534668, 'train/grad_norm_pre_clip': 207.1060333251953, 'timing/data': 0.00028138999186921865, 'timing/model': 0.3873211970058037, 'learning_rate/qwen_vl_interface': 4.5e-06, 'learning_rate/action_model': 4.5e-05, 'epoch': 0.0}) +[2026-10-03 13:01:04,237][starVLA.training.train_starvla][INFO] - Step 46, Loss: {'train/loss': 1.2527072429656982, 'train/grad_norm_pre_clip': 195.1995849609375, 'timing/data': 0.0003616199974203482, 'timing/model': 0.38719480799045414, 'learning_rate/qwen_vl_interface': 4.600000000000001e-06, 'learning_rate/action_model': 4.600000000000001e-05, 'epoch': 0.0}) +[2026-10-03 13:01:04,928][starVLA.training.train_starvla][INFO] - Step 47, Loss: {'train/loss': 1.291862964630127, 'train/grad_norm_pre_clip': 40.47221755981445, 'timing/data': 0.0003447600029176101, 'timing/model': 0.38692890699894633, 'learning_rate/qwen_vl_interface': 4.7e-06, 'learning_rate/action_model': 4.7e-05, 'epoch': 0.0}) +[2026-10-03 13:01:05,816][starVLA.training.train_starvla][INFO] - Step 48, Loss: {'train/loss': 1.2033772468566895, 'train/grad_norm_pre_clip': 129.98387145996094, 'timing/data': 0.00029888900462538004, 'timing/model': 0.5798298560112016, 'learning_rate/qwen_vl_interface': 4.800000000000001e-06, 'learning_rate/action_model': 4.8e-05, 'epoch': 0.0}) +[2026-10-03 13:01:06,509][starVLA.training.train_starvla][INFO] - Step 49, Loss: {'train/loss': 0.9924026727676392, 'train/grad_norm_pre_clip': 100.95817565917969, 'timing/data': 0.00036021000414621085, 'timing/model': 0.38722432700160425, 'learning_rate/qwen_vl_interface': 4.9000000000000005e-06, 'learning_rate/action_model': 4.9e-05, 'epoch': 0.0}) +[2026-10-03 13:01:07,200][starVLA.training.train_starvla][INFO] - Step 50, Loss: {'train/loss': 0.9816746711730957, 'train/grad_norm_pre_clip': 109.51363372802734, 'timing/data': 0.00020010999287478626, 'timing/model': 0.38643353698716965, 'learning_rate/qwen_vl_interface': 5e-06, 'learning_rate/action_model': 5e-05, 'epoch': 0.0}) +[2026-10-03 13:01:07,892][starVLA.training.train_starvla][INFO] - Step 51, Loss: {'train/loss': 1.1348971128463745, 'train/grad_norm_pre_clip': 87.77255249023438, 'timing/data': 0.0003379099944140762, 'timing/model': 0.38735076699231286, 'learning_rate/qwen_vl_interface': 5.1e-06, 'learning_rate/action_model': 5.1000000000000006e-05, 'epoch': 0.0}) +[2026-10-03 13:01:08,587][starVLA.training.train_starvla][INFO] - Step 52, Loss: {'train/loss': 0.6112332344055176, 'train/grad_norm_pre_clip': 103.14403533935547, 'timing/data': 0.0003895900008501485, 'timing/model': 0.3870428969967179, 'learning_rate/qwen_vl_interface': 5.2e-06, 'learning_rate/action_model': 5.2000000000000004e-05, 'epoch': 0.0}) +[2026-10-03 13:01:09,279][starVLA.training.train_starvla][INFO] - Step 53, Loss: {'train/loss': 1.8003791570663452, 'train/grad_norm_pre_clip': 201.07493591308594, 'timing/data': 0.00023327999224420637, 'timing/model': 0.38684442700468935, 'learning_rate/qwen_vl_interface': 5.300000000000001e-06, 'learning_rate/action_model': 5.300000000000001e-05, 'epoch': 0.0}) +[2026-10-03 13:01:10,160][starVLA.training.train_starvla][INFO] - Step 54, Loss: {'train/loss': 1.3538705110549927, 'train/grad_norm_pre_clip': 84.42237091064453, 'timing/data': 0.00034651999885682017, 'timing/model': 0.3868029469886096, 'learning_rate/qwen_vl_interface': 5.400000000000001e-06, 'learning_rate/action_model': 5.4000000000000005e-05, 'epoch': 0.0}) +[2026-10-03 13:01:10,852][starVLA.training.train_starvla][INFO] - Step 55, Loss: {'train/loss': 1.5317308902740479, 'train/grad_norm_pre_clip': 213.7039337158203, 'timing/data': 0.00019746999896597117, 'timing/model': 0.3875569080119021, 'learning_rate/qwen_vl_interface': 5.500000000000001e-06, 'learning_rate/action_model': 5.500000000000001e-05, 'epoch': 0.0}) +[2026-10-03 13:01:11,548][starVLA.training.train_starvla][INFO] - Step 56, Loss: {'train/loss': 1.4878836870193481, 'train/grad_norm_pre_clip': 328.500732421875, 'timing/data': 0.00031809900247026235, 'timing/model': 0.3878535170078976, 'learning_rate/qwen_vl_interface': 5.600000000000001e-06, 'learning_rate/action_model': 5.6000000000000006e-05, 'epoch': 0.0}) +[2026-10-03 13:01:12,241][starVLA.training.train_starvla][INFO] - Step 57, Loss: {'train/loss': 0.6877805590629578, 'train/grad_norm_pre_clip': 255.7489471435547, 'timing/data': 0.00039644999196752906, 'timing/model': 0.38746438699308783, 'learning_rate/qwen_vl_interface': 5.7e-06, 'learning_rate/action_model': 5.6999999999999996e-05, 'epoch': 0.0}) +[2026-10-03 13:01:12,931][starVLA.training.train_starvla][INFO] - Step 58, Loss: {'train/loss': 1.3794902563095093, 'train/grad_norm_pre_clip': 59.46781921386719, 'timing/data': 0.00034837999555747956, 'timing/model': 0.38608833801117726, 'learning_rate/qwen_vl_interface': 5.8e-06, 'learning_rate/action_model': 5.8e-05, 'epoch': 0.0}) +[2026-10-03 13:01:13,815][starVLA.training.train_starvla][INFO] - Step 59, Loss: {'train/loss': 0.39530831575393677, 'train/grad_norm_pre_clip': 102.29320526123047, 'timing/data': 0.0003736500075319782, 'timing/model': 0.38697736799076665, 'learning_rate/qwen_vl_interface': 5.9e-06, 'learning_rate/action_model': 5.9e-05, 'epoch': 0.0}) +[2026-10-03 13:01:14,510][starVLA.training.train_starvla][INFO] - Step 60, Loss: {'train/loss': 0.7919625639915466, 'train/grad_norm_pre_clip': 128.7531280517578, 'timing/data': 0.00040347999311052263, 'timing/model': 0.38750151699059643, 'learning_rate/qwen_vl_interface': 6e-06, 'learning_rate/action_model': 6e-05, 'epoch': 0.0}) +[2026-10-03 13:01:15,201][starVLA.training.train_starvla][INFO] - Step 61, Loss: {'train/loss': 1.1071667671203613, 'train/grad_norm_pre_clip': 96.34098815917969, 'timing/data': 0.0002867400035029277, 'timing/model': 0.38718735698785167, 'learning_rate/qwen_vl_interface': 6.1e-06, 'learning_rate/action_model': 6.1e-05, 'epoch': 0.0}) +[2026-10-03 13:01:15,892][starVLA.training.train_starvla][INFO] - Step 62, Loss: {'train/loss': 1.174344778060913, 'train/grad_norm_pre_clip': 61.62997817993164, 'timing/data': 0.00030433999199885875, 'timing/model': 0.38701282799593173, 'learning_rate/qwen_vl_interface': 6.200000000000001e-06, 'learning_rate/action_model': 6.2e-05, 'epoch': 0.0}) +[2026-10-03 13:01:16,585][starVLA.training.train_starvla][INFO] - Step 63, Loss: {'train/loss': 1.3004295825958252, 'train/grad_norm_pre_clip': 381.2536926269531, 'timing/data': 0.00030451999919023365, 'timing/model': 0.38769840700842906, 'learning_rate/qwen_vl_interface': 6.300000000000001e-06, 'learning_rate/action_model': 6.3e-05, 'epoch': 0.0}) +[2026-10-03 13:01:17,473][starVLA.training.train_starvla][INFO] - Step 64, Loss: {'train/loss': 0.4580705463886261, 'train/grad_norm_pre_clip': 68.6540298461914, 'timing/data': 0.00026666000485420227, 'timing/model': 0.5800815560069168, 'learning_rate/qwen_vl_interface': 6.4000000000000006e-06, 'learning_rate/action_model': 6.400000000000001e-05, 'epoch': 0.0}) +[2026-10-03 13:01:18,168][starVLA.training.train_starvla][INFO] - Step 65, Loss: {'train/loss': 0.034121885895729065, 'train/grad_norm_pre_clip': 32.45881271362305, 'timing/data': 0.00024075999681372195, 'timing/model': 0.38841507700271904, 'learning_rate/qwen_vl_interface': 6.5000000000000004e-06, 'learning_rate/action_model': 6.500000000000001e-05, 'epoch': 0.0}) +[2026-10-03 13:01:18,860][starVLA.training.train_starvla][INFO] - Step 66, Loss: {'train/loss': 1.110194444656372, 'train/grad_norm_pre_clip': 220.13218688964844, 'timing/data': 0.00021123001351952553, 'timing/model': 0.38637080701300874, 'learning_rate/qwen_vl_interface': 6.600000000000001e-06, 'learning_rate/action_model': 6.6e-05, 'epoch': 0.0}) +[2026-10-03 13:01:19,554][starVLA.training.train_starvla][INFO] - Step 67, Loss: {'train/loss': 0.7100881338119507, 'train/grad_norm_pre_clip': 194.41647338867188, 'timing/data': 0.0001983499969355762, 'timing/model': 0.3879632970056264, 'learning_rate/qwen_vl_interface': 6.700000000000001e-06, 'learning_rate/action_model': 6.7e-05, 'epoch': 0.0}) +[2026-10-03 13:01:20,251][starVLA.training.train_starvla][INFO] - Step 68, Loss: {'train/loss': 0.8618963956832886, 'train/grad_norm_pre_clip': 116.97868347167969, 'timing/data': 0.000256139988778159, 'timing/model': 0.3880857659969479, 'learning_rate/qwen_vl_interface': 6.800000000000001e-06, 'learning_rate/action_model': 6.800000000000001e-05, 'epoch': 0.0}) +[2026-10-03 13:01:21,141][starVLA.training.train_starvla][INFO] - Step 69, Loss: {'train/loss': 0.8007127642631531, 'train/grad_norm_pre_clip': 375.22454833984375, 'timing/data': 0.0002839899971149862, 'timing/model': 0.583229295007186, 'learning_rate/qwen_vl_interface': 6.9e-06, 'learning_rate/action_model': 6.9e-05, 'epoch': 0.01}) +[2026-10-03 13:01:21,834][starVLA.training.train_starvla][INFO] - Step 70, Loss: {'train/loss': 1.2642977237701416, 'train/grad_norm_pre_clip': 224.3206024169922, 'timing/data': 0.0002529199991840869, 'timing/model': 0.3874397070030682, 'learning_rate/qwen_vl_interface': 7e-06, 'learning_rate/action_model': 7e-05, 'epoch': 0.01}) +[2026-10-03 13:01:22,528][starVLA.training.train_starvla][INFO] - Step 71, Loss: {'train/loss': 1.0494699478149414, 'train/grad_norm_pre_clip': 250.6042938232422, 'timing/data': 0.00023997999960556626, 'timing/model': 0.38866374699864537, 'learning_rate/qwen_vl_interface': 7.100000000000001e-06, 'learning_rate/action_model': 7.1e-05, 'epoch': 0.01}) +[2026-10-03 13:01:23,225][starVLA.training.train_starvla][INFO] - Step 72, Loss: {'train/loss': 0.736369252204895, 'train/grad_norm_pre_clip': 113.80880737304688, 'timing/data': 0.00020246999338269234, 'timing/model': 0.38828016800107434, 'learning_rate/qwen_vl_interface': 7.2000000000000005e-06, 'learning_rate/action_model': 7.2e-05, 'epoch': 0.01}) +[2026-10-03 13:01:23,919][starVLA.training.train_starvla][INFO] - Step 73, Loss: {'train/loss': 2.1353650093078613, 'train/grad_norm_pre_clip': 189.55624389648438, 'timing/data': 0.00020388000120874494, 'timing/model': 0.3882682060066145, 'learning_rate/qwen_vl_interface': 7.3e-06, 'learning_rate/action_model': 7.3e-05, 'epoch': 0.01}) +[2026-10-03 13:01:24,610][starVLA.training.train_starvla][INFO] - Step 74, Loss: {'train/loss': 0.579296350479126, 'train/grad_norm_pre_clip': 100.09652709960938, 'timing/data': 0.00034319001133553684, 'timing/model': 0.38720369699876755, 'learning_rate/qwen_vl_interface': 7.4e-06, 'learning_rate/action_model': 7.4e-05, 'epoch': 0.01}) +[2026-10-03 13:01:25,496][starVLA.training.train_starvla][INFO] - Step 75, Loss: {'train/loss': 1.254788875579834, 'train/grad_norm_pre_clip': 173.53375244140625, 'timing/data': 0.00030641000194009393, 'timing/model': 0.38770841699442826, 'learning_rate/qwen_vl_interface': 7.500000000000001e-06, 'learning_rate/action_model': 7.500000000000001e-05, 'epoch': 0.01}) +[2026-10-03 13:01:26,192][starVLA.training.train_starvla][INFO] - Step 76, Loss: {'train/loss': 1.7431390285491943, 'train/grad_norm_pre_clip': 160.0484161376953, 'timing/data': 0.0002400499943178147, 'timing/model': 0.3877670570072951, 'learning_rate/qwen_vl_interface': 7.600000000000001e-06, 'learning_rate/action_model': 7.6e-05, 'epoch': 0.01}) +[2026-10-03 13:01:26,885][starVLA.training.train_starvla][INFO] - Step 77, Loss: {'train/loss': 0.9758698344230652, 'train/grad_norm_pre_clip': 46.0667724609375, 'timing/data': 0.0002926500019384548, 'timing/model': 0.3875856670056237, 'learning_rate/qwen_vl_interface': 7.7e-06, 'learning_rate/action_model': 7.7e-05, 'epoch': 0.01}) +[2026-10-03 13:01:27,577][starVLA.training.train_starvla][INFO] - Step 78, Loss: {'train/loss': 0.43845582008361816, 'train/grad_norm_pre_clip': 74.23665618896484, 'timing/data': 0.00019597000209614635, 'timing/model': 0.38752611698873807, 'learning_rate/qwen_vl_interface': 7.800000000000002e-06, 'learning_rate/action_model': 7.800000000000001e-05, 'epoch': 0.01}) +[2026-10-03 13:01:28,270][starVLA.training.train_starvla][INFO] - Step 79, Loss: {'train/loss': 1.2094988822937012, 'train/grad_norm_pre_clip': 125.81075286865234, 'timing/data': 0.00028918001044075936, 'timing/model': 0.3877890669973567, 'learning_rate/qwen_vl_interface': 7.9e-06, 'learning_rate/action_model': 7.900000000000001e-05, 'epoch': 0.01}) +[2026-10-03 13:01:29,157][starVLA.training.train_starvla][INFO] - Step 80, Loss: {'train/loss': 0.42604389786720276, 'train/grad_norm_pre_clip': 89.56310272216797, 'timing/data': 0.00021574999846052378, 'timing/model': 0.5780490160104819, 'learning_rate/qwen_vl_interface': 8.000000000000001e-06, 'learning_rate/action_model': 8e-05, 'epoch': 0.01}) +[2026-10-03 13:01:29,850][starVLA.training.train_starvla][INFO] - Step 81, Loss: {'train/loss': 1.4919838905334473, 'train/grad_norm_pre_clip': 52.272438049316406, 'timing/data': 0.000279119994957, 'timing/model': 0.3876932770072017, 'learning_rate/qwen_vl_interface': 8.1e-06, 'learning_rate/action_model': 8.1e-05, 'epoch': 0.01}) +[2026-10-03 13:01:30,543][starVLA.training.train_starvla][INFO] - Step 82, Loss: {'train/loss': 1.0534080266952515, 'train/grad_norm_pre_clip': 160.25131225585938, 'timing/data': 0.00023697999131400138, 'timing/model': 0.3873811670055147, 'learning_rate/qwen_vl_interface': 8.2e-06, 'learning_rate/action_model': 8.2e-05, 'epoch': 0.01}) +[2026-10-03 13:01:31,234][starVLA.training.train_starvla][INFO] - Step 83, Loss: {'train/loss': 1.304744839668274, 'train/grad_norm_pre_clip': 186.44180297851562, 'timing/data': 0.00033260000054724514, 'timing/model': 0.3868800670024939, 'learning_rate/qwen_vl_interface': 8.3e-06, 'learning_rate/action_model': 8.3e-05, 'epoch': 0.01}) +[2026-10-03 13:01:31,930][starVLA.training.train_starvla][INFO] - Step 84, Loss: {'train/loss': 0.44579261541366577, 'train/grad_norm_pre_clip': 26.24103355407715, 'timing/data': 0.00029922999965492636, 'timing/model': 0.3877737669972703, 'learning_rate/qwen_vl_interface': 8.400000000000001e-06, 'learning_rate/action_model': 8.4e-05, 'epoch': 0.01}) +[2026-10-03 13:01:32,817][starVLA.training.train_starvla][INFO] - Step 85, Loss: {'train/loss': 0.6578391790390015, 'train/grad_norm_pre_clip': 21.85904312133789, 'timing/data': 0.0002905600122176111, 'timing/model': 0.5811767150007654, 'learning_rate/qwen_vl_interface': 8.5e-06, 'learning_rate/action_model': 8.5e-05, 'epoch': 0.01}) +[2026-10-03 13:01:33,510][starVLA.training.train_starvla][INFO] - Step 86, Loss: {'train/loss': 1.0907318592071533, 'train/grad_norm_pre_clip': 154.62437438964844, 'timing/data': 0.0003632000007200986, 'timing/model': 0.3875896660028957, 'learning_rate/qwen_vl_interface': 8.6e-06, 'learning_rate/action_model': 8.6e-05, 'epoch': 0.01}) +[2026-10-03 13:01:34,203][starVLA.training.train_starvla][INFO] - Step 87, Loss: {'train/loss': 1.0570696592330933, 'train/grad_norm_pre_clip': 276.3758239746094, 'timing/data': 0.000351310009136796, 'timing/model': 0.3880758279992733, 'learning_rate/qwen_vl_interface': 8.700000000000001e-06, 'learning_rate/action_model': 8.7e-05, 'epoch': 0.01}) +[2026-10-03 13:01:34,898][starVLA.training.train_starvla][INFO] - Step 88, Loss: {'train/loss': 1.4992122650146484, 'train/grad_norm_pre_clip': 117.40604400634766, 'timing/data': 0.00039365999691653997, 'timing/model': 0.38600475699058734, 'learning_rate/qwen_vl_interface': 8.8e-06, 'learning_rate/action_model': 8.800000000000001e-05, 'epoch': 0.01}) +[2026-10-03 13:01:35,592][starVLA.training.train_starvla][INFO] - Step 89, Loss: {'train/loss': 0.910429060459137, 'train/grad_norm_pre_clip': 86.80088806152344, 'timing/data': 0.000285360001726076, 'timing/model': 0.38877005799440667, 'learning_rate/qwen_vl_interface': 8.900000000000001e-06, 'learning_rate/action_model': 8.900000000000001e-05, 'epoch': 0.01}) +[2026-10-03 13:01:36,287][starVLA.training.train_starvla][INFO] - Step 90, Loss: {'train/loss': 0.9423068761825562, 'train/grad_norm_pre_clip': 119.06575775146484, 'timing/data': 0.0001961699890671298, 'timing/model': 0.3891553660068894, 'learning_rate/qwen_vl_interface': 9e-06, 'learning_rate/action_model': 9e-05, 'epoch': 0.01}) +[2026-10-03 13:01:37,174][starVLA.training.train_starvla][INFO] - Step 91, Loss: {'train/loss': 0.7517255544662476, 'train/grad_norm_pre_clip': 67.88011169433594, 'timing/data': 0.00039604000630788505, 'timing/model': 0.3884423170093214, 'learning_rate/qwen_vl_interface': 9.100000000000001e-06, 'learning_rate/action_model': 9.1e-05, 'epoch': 0.01}) +[2026-10-03 13:01:37,872][starVLA.training.train_starvla][INFO] - Step 92, Loss: {'train/loss': 1.0185348987579346, 'train/grad_norm_pre_clip': 50.60332107543945, 'timing/data': 0.00020461999520193785, 'timing/model': 0.3887862179981312, 'learning_rate/qwen_vl_interface': 9.200000000000002e-06, 'learning_rate/action_model': 9.200000000000001e-05, 'epoch': 0.01}) +[2026-10-03 13:01:38,567][starVLA.training.train_starvla][INFO] - Step 93, Loss: {'train/loss': 1.0660483837127686, 'train/grad_norm_pre_clip': 129.65756225585938, 'timing/data': 0.0003018500137841329, 'timing/model': 0.38816219600266777, 'learning_rate/qwen_vl_interface': 9.3e-06, 'learning_rate/action_model': 9.300000000000001e-05, 'epoch': 0.01}) +[2026-10-03 13:01:39,261][starVLA.training.train_starvla][INFO] - Step 94, Loss: {'train/loss': 1.404861569404602, 'train/grad_norm_pre_clip': 111.68168640136719, 'timing/data': 0.0003315700014354661, 'timing/model': 0.3883201479911804, 'learning_rate/qwen_vl_interface': 9.4e-06, 'learning_rate/action_model': 9.4e-05, 'epoch': 0.01}) +[2026-10-03 13:01:39,955][starVLA.training.train_starvla][INFO] - Step 95, Loss: {'train/loss': 0.943350076675415, 'train/grad_norm_pre_clip': 54.57770538330078, 'timing/data': 0.0002996300027007237, 'timing/model': 0.38866180600598454, 'learning_rate/qwen_vl_interface': 9.5e-06, 'learning_rate/action_model': 9.5e-05, 'epoch': 0.01}) +[2026-10-03 13:01:40,842][starVLA.training.train_starvla][INFO] - Step 96, Loss: {'train/loss': 1.0224320888519287, 'train/grad_norm_pre_clip': 208.19094848632812, 'timing/data': 0.00024839999969117343, 'timing/model': 0.38835717699839734, 'learning_rate/qwen_vl_interface': 9.600000000000001e-06, 'learning_rate/action_model': 9.6e-05, 'epoch': 0.01}) +[2026-10-03 13:01:41,535][starVLA.training.train_starvla][INFO] - Step 97, Loss: {'train/loss': 1.6876850128173828, 'train/grad_norm_pre_clip': 212.1693878173828, 'timing/data': 0.00036751000152435154, 'timing/model': 0.3879303369903937, 'learning_rate/qwen_vl_interface': 9.7e-06, 'learning_rate/action_model': 9.7e-05, 'epoch': 0.01}) +[2026-10-03 13:01:42,229][starVLA.training.train_starvla][INFO] - Step 98, Loss: {'train/loss': 0.9484137296676636, 'train/grad_norm_pre_clip': 170.80784606933594, 'timing/data': 0.0003446299961069599, 'timing/model': 0.38768272699962836, 'learning_rate/qwen_vl_interface': 9.800000000000001e-06, 'learning_rate/action_model': 9.8e-05, 'epoch': 0.01}) +[2026-10-03 13:01:42,924][starVLA.training.train_starvla][INFO] - Step 99, Loss: {'train/loss': 1.1263405084609985, 'train/grad_norm_pre_clip': 102.73956298828125, 'timing/data': 0.0001971400051843375, 'timing/model': 0.3890253970021149, 'learning_rate/qwen_vl_interface': 9.9e-06, 'learning_rate/action_model': 9.900000000000001e-05, 'epoch': 0.01}) +[2026-10-03 13:01:43,622][starVLA.training.train_starvla][INFO] - Step 100, Loss: {'train/loss': 1.1139719486236572, 'train/grad_norm_pre_clip': 185.6865692138672, 'timing/data': 0.00023748999228700995, 'timing/model': 0.3887769959983416, 'learning_rate/qwen_vl_interface': 1e-05, 'learning_rate/action_model': 0.0001, 'epoch': 0.01}) +[2026-10-03 13:01:44,506][starVLA.training.train_starvla][INFO] - Step 101, Loss: {'train/loss': 2.193148612976074, 'train/grad_norm_pre_clip': 345.7972412109375, 'timing/data': 0.00031092000426724553, 'timing/model': 0.5761871660070028, 'learning_rate/qwen_vl_interface': 9.999999023727212e-06, 'learning_rate/action_model': 9.99999902372721e-05, 'epoch': 0.01}) +[2026-10-03 13:01:45,200][starVLA.training.train_starvla][INFO] - Step 102, Loss: {'train/loss': 1.2743638753890991, 'train/grad_norm_pre_clip': 328.9493103027344, 'timing/data': 0.00031015000422485173, 'timing/model': 0.3880894360045204, 'learning_rate/qwen_vl_interface': 9.999996094909247e-06, 'learning_rate/action_model': 9.999996094909248e-05, 'epoch': 0.01}) +[2026-10-03 13:01:45,893][starVLA.training.train_starvla][INFO] - Step 103, Loss: {'train/loss': 0.6652551293373108, 'train/grad_norm_pre_clip': 66.818603515625, 'timing/data': 0.00024795001081656665, 'timing/model': 0.3883413760049734, 'learning_rate/qwen_vl_interface': 9.999991213547314e-06, 'learning_rate/action_model': 9.999991213547313e-05, 'epoch': 0.01}) +[2026-10-03 13:01:46,592][starVLA.training.train_starvla][INFO] - Step 104, Loss: {'train/loss': 0.8929495811462402, 'train/grad_norm_pre_clip': 47.72836685180664, 'timing/data': 0.00031782999576535076, 'timing/model': 0.3883570970065193, 'learning_rate/qwen_vl_interface': 9.999984379643413e-06, 'learning_rate/action_model': 9.999984379643412e-05, 'epoch': 0.01}) +[2026-10-03 13:01:47,287][starVLA.training.train_starvla][INFO] - Step 105, Loss: {'train/loss': 0.9932169914245605, 'train/grad_norm_pre_clip': 148.4534149169922, 'timing/data': 0.00034221999521832913, 'timing/model': 0.38790776699897833, 'learning_rate/qwen_vl_interface': 9.999975593200357e-06, 'learning_rate/action_model': 9.999975593200357e-05, 'epoch': 0.01}) +[2026-10-03 13:01:48,173][starVLA.training.train_starvla][INFO] - Step 106, Loss: {'train/loss': 1.869853138923645, 'train/grad_norm_pre_clip': 164.67001342773438, 'timing/data': 0.00034761999268084764, 'timing/model': 0.5805309459974524, 'learning_rate/qwen_vl_interface': 9.999964854221756e-06, 'learning_rate/action_model': 9.999964854221754e-05, 'epoch': 0.01}) +[2026-10-03 13:01:48,867][starVLA.training.train_starvla][INFO] - Step 107, Loss: {'train/loss': 0.8378232717514038, 'train/grad_norm_pre_clip': 107.68096923828125, 'timing/data': 0.00032171000202652067, 'timing/model': 0.38874742700136267, 'learning_rate/qwen_vl_interface': 9.999952162712027e-06, 'learning_rate/action_model': 9.999952162712026e-05, 'epoch': 0.01}) +[2026-10-03 13:01:49,564][starVLA.training.train_starvla][INFO] - Step 108, Loss: {'train/loss': 0.4766554832458496, 'train/grad_norm_pre_clip': 30.55817222595215, 'timing/data': 0.0003169999981764704, 'timing/model': 0.3884357469942188, 'learning_rate/qwen_vl_interface': 9.999937518676382e-06, 'learning_rate/action_model': 9.999937518676382e-05, 'epoch': 0.01}) +[2026-10-03 13:01:50,260][starVLA.training.train_starvla][INFO] - Step 109, Loss: {'train/loss': 0.7735843658447266, 'train/grad_norm_pre_clip': 24.97170639038086, 'timing/data': 0.0002005600108532235, 'timing/model': 0.38870084699010476, 'learning_rate/qwen_vl_interface': 9.999920922120846e-06, 'learning_rate/action_model': 9.999920922120846e-05, 'epoch': 0.01}) +[2026-10-03 13:01:50,953][starVLA.training.train_starvla][INFO] - Step 110, Loss: {'train/loss': 0.9386069178581238, 'train/grad_norm_pre_clip': 50.06280517578125, 'timing/data': 0.0003447799972491339, 'timing/model': 0.3882838969875593, 'learning_rate/qwen_vl_interface': 9.999902373052239e-06, 'learning_rate/action_model': 9.999902373052239e-05, 'epoch': 0.01}) +[2026-10-03 13:01:51,839][starVLA.training.train_starvla][INFO] - Step 111, Loss: {'train/loss': 0.6395965218544006, 'train/grad_norm_pre_clip': 55.03965759277344, 'timing/data': 0.00033304998942185193, 'timing/model': 0.5787660460046027, 'learning_rate/qwen_vl_interface': 9.999881871478188e-06, 'learning_rate/action_model': 9.999881871478187e-05, 'epoch': 0.01}) +[2026-10-03 13:01:52,536][starVLA.training.train_starvla][INFO] - Step 112, Loss: {'train/loss': 0.7886514663696289, 'train/grad_norm_pre_clip': 41.1170654296875, 'timing/data': 0.0002578999992692843, 'timing/model': 0.387982706000912, 'learning_rate/qwen_vl_interface': 9.999859417407115e-06, 'learning_rate/action_model': 9.999859417407115e-05, 'epoch': 0.01}) +[2026-10-03 13:01:53,231][starVLA.training.train_starvla][INFO] - Step 113, Loss: {'train/loss': 0.6120100021362305, 'train/grad_norm_pre_clip': 57.40627670288086, 'timing/data': 0.00024063000455498695, 'timing/model': 0.38832988700596616, 'learning_rate/qwen_vl_interface': 9.999835010848254e-06, 'learning_rate/action_model': 9.999835010848254e-05, 'epoch': 0.01}) +[2026-10-03 13:01:53,925][starVLA.training.train_starvla][INFO] - Step 114, Loss: {'train/loss': 0.7940768003463745, 'train/grad_norm_pre_clip': 42.10763931274414, 'timing/data': 0.00024126000062096864, 'timing/model': 0.3881386669963831, 'learning_rate/qwen_vl_interface': 9.999808651811639e-06, 'learning_rate/action_model': 9.999808651811639e-05, 'epoch': 0.01}) +[2026-10-03 13:01:54,619][starVLA.training.train_starvla][INFO] - Step 115, Loss: {'train/loss': 1.225208044052124, 'train/grad_norm_pre_clip': 142.8638153076172, 'timing/data': 0.00033364999399054796, 'timing/model': 0.3883018170017749, 'learning_rate/qwen_vl_interface': 9.999780340308101e-06, 'learning_rate/action_model': 9.999780340308101e-05, 'epoch': 0.01}) +[2026-10-03 13:01:55,316][starVLA.training.train_starvla][INFO] - Step 116, Loss: {'train/loss': 1.074184536933899, 'train/grad_norm_pre_clip': 65.99114990234375, 'timing/data': 0.0003140699991490692, 'timing/model': 0.3879070280090673, 'learning_rate/qwen_vl_interface': 9.999750076349282e-06, 'learning_rate/action_model': 9.999750076349282e-05, 'epoch': 0.01}) +[2026-10-03 13:01:56,205][starVLA.training.train_starvla][INFO] - Step 117, Loss: {'train/loss': 0.8109111785888672, 'train/grad_norm_pre_clip': 36.228416442871094, 'timing/data': 0.0002862000110326335, 'timing/model': 0.3884191370016197, 'learning_rate/qwen_vl_interface': 9.999717859947618e-06, 'learning_rate/action_model': 9.999717859947618e-05, 'epoch': 0.01}) +[2026-10-03 13:01:56,898][starVLA.training.train_starvla][INFO] - Step 118, Loss: {'train/loss': 0.8571575880050659, 'train/grad_norm_pre_clip': 40.569580078125, 'timing/data': 0.0002006699942285195, 'timing/model': 0.38823937700362876, 'learning_rate/qwen_vl_interface': 9.999683691116356e-06, 'learning_rate/action_model': 9.999683691116355e-05, 'epoch': 0.01}) +[2026-10-03 13:01:57,593][starVLA.training.train_starvla][INFO] - Step 119, Loss: {'train/loss': 0.722567081451416, 'train/grad_norm_pre_clip': 108.90629577636719, 'timing/data': 0.0002783499949146062, 'timing/model': 0.3883685969922226, 'learning_rate/qwen_vl_interface': 9.999647569869539e-06, 'learning_rate/action_model': 9.99964756986954e-05, 'epoch': 0.01}) +[2026-10-03 13:01:58,290][starVLA.training.train_starvla][INFO] - Step 120, Loss: {'train/loss': 0.9185184240341187, 'train/grad_norm_pre_clip': 31.506027221679688, 'timing/data': 0.0003262899990659207, 'timing/model': 0.3887707469984889, 'learning_rate/qwen_vl_interface': 9.999609496222017e-06, 'learning_rate/action_model': 9.999609496222017e-05, 'epoch': 0.01}) +[2026-10-03 13:01:58,985][starVLA.training.train_starvla][INFO] - Step 121, Loss: {'train/loss': 0.6379657983779907, 'train/grad_norm_pre_clip': 22.305883407592773, 'timing/data': 0.00028434999694582075, 'timing/model': 0.38823891601350624, 'learning_rate/qwen_vl_interface': 9.99956947018944e-06, 'learning_rate/action_model': 9.999569470189439e-05, 'epoch': 0.01}) +[2026-10-03 13:01:59,871][starVLA.training.train_starvla][INFO] - Step 122, Loss: {'train/loss': 0.5541465878486633, 'train/grad_norm_pre_clip': 31.119722366333008, 'timing/data': 0.0003642999945441261, 'timing/model': 0.5795524860004662, 'learning_rate/qwen_vl_interface': 9.99952749178826e-06, 'learning_rate/action_model': 9.999527491788259e-05, 'epoch': 0.01}) +[2026-10-03 13:02:00,565][starVLA.training.train_starvla][INFO] - Step 123, Loss: {'train/loss': 1.1857789754867554, 'train/grad_norm_pre_clip': 112.33816528320312, 'timing/data': 0.00024345000565517694, 'timing/model': 0.3882839160069125, 'learning_rate/qwen_vl_interface': 9.999483561035732e-06, 'learning_rate/action_model': 9.999483561035732e-05, 'epoch': 0.01}) +[2026-10-03 13:02:01,261][starVLA.training.train_starvla][INFO] - Step 124, Loss: {'train/loss': 0.814720869064331, 'train/grad_norm_pre_clip': 47.515323638916016, 'timing/data': 0.0002039300015894696, 'timing/model': 0.38831787700473797, 'learning_rate/qwen_vl_interface': 9.999437677949917e-06, 'learning_rate/action_model': 9.999437677949917e-05, 'epoch': 0.01}) +[2026-10-03 13:02:01,956][starVLA.training.train_starvla][INFO] - Step 125, Loss: {'train/loss': 1.0451405048370361, 'train/grad_norm_pre_clip': 44.4427490234375, 'timing/data': 0.00020210999355185777, 'timing/model': 0.3881490269995993, 'learning_rate/qwen_vl_interface': 9.999389842549674e-06, 'learning_rate/action_model': 9.999389842549673e-05, 'epoch': 0.01}) +[2026-10-03 13:02:02,648][starVLA.training.train_starvla][INFO] - Step 126, Loss: {'train/loss': 1.0466128587722778, 'train/grad_norm_pre_clip': 124.54217529296875, 'timing/data': 0.00023975000658538193, 'timing/model': 0.38752420799573883, 'learning_rate/qwen_vl_interface': 9.999340054854669e-06, 'learning_rate/action_model': 9.999340054854667e-05, 'epoch': 0.01}) +[2026-10-03 13:02:03,342][starVLA.training.train_starvla][INFO] - Step 127, Loss: {'train/loss': 1.2833354473114014, 'train/grad_norm_pre_clip': 99.83876037597656, 'timing/data': 0.00019642000552266836, 'timing/model': 0.387872936000349, 'learning_rate/qwen_vl_interface': 9.999288314885362e-06, 'learning_rate/action_model': 9.999288314885362e-05, 'epoch': 0.01}) +[2026-10-03 13:02:04,229][starVLA.training.train_starvla][INFO] - Step 128, Loss: {'train/loss': 0.9772636890411377, 'train/grad_norm_pre_clip': 33.89665222167969, 'timing/data': 0.00029368000105023384, 'timing/model': 0.3882159269996919, 'learning_rate/qwen_vl_interface': 9.999234622663027e-06, 'learning_rate/action_model': 9.999234622663028e-05, 'epoch': 0.01}) +[2026-10-03 13:02:04,924][starVLA.training.train_starvla][INFO] - Step 129, Loss: {'train/loss': 0.4985797703266144, 'train/grad_norm_pre_clip': 28.890304565429688, 'timing/data': 0.00032654999813530594, 'timing/model': 0.38817379700776655, 'learning_rate/qwen_vl_interface': 9.999178978209732e-06, 'learning_rate/action_model': 9.999178978209732e-05, 'epoch': 0.01}) +[2026-10-03 13:02:05,618][starVLA.training.train_starvla][INFO] - Step 130, Loss: {'train/loss': 0.687678337097168, 'train/grad_norm_pre_clip': 16.698659896850586, 'timing/data': 0.00024742000096011907, 'timing/model': 0.38847229699604213, 'learning_rate/qwen_vl_interface': 9.999121381548354e-06, 'learning_rate/action_model': 9.999121381548353e-05, 'epoch': 0.01}) +[2026-10-03 13:02:06,314][starVLA.training.train_starvla][INFO] - Step 131, Loss: {'train/loss': 1.364220142364502, 'train/grad_norm_pre_clip': 49.48310470581055, 'timing/data': 0.00034128999686799943, 'timing/model': 0.38936957700934727, 'learning_rate/qwen_vl_interface': 9.999061832702563e-06, 'learning_rate/action_model': 9.999061832702563e-05, 'epoch': 0.01}) +[2026-10-03 13:02:07,012][starVLA.training.train_starvla][INFO] - Step 132, Loss: {'train/loss': 0.972115695476532, 'train/grad_norm_pre_clip': 34.88639831542969, 'timing/data': 0.0003241099911974743, 'timing/model': 0.38865791700663976, 'learning_rate/qwen_vl_interface': 9.999000331696842e-06, 'learning_rate/action_model': 9.999000331696841e-05, 'epoch': 0.01}) +[2026-10-03 13:02:07,899][starVLA.training.train_starvla][INFO] - Step 133, Loss: {'train/loss': 0.4607832729816437, 'train/grad_norm_pre_clip': 25.933612823486328, 'timing/data': 0.000269329990260303, 'timing/model': 0.38789192600233946, 'learning_rate/qwen_vl_interface': 9.99893687855647e-06, 'learning_rate/action_model': 9.99893687855647e-05, 'epoch': 0.01}) +[2026-10-03 13:02:08,593][starVLA.training.train_starvla][INFO] - Step 134, Loss: {'train/loss': 0.6443790197372437, 'train/grad_norm_pre_clip': 36.439090728759766, 'timing/data': 0.0004064899985678494, 'timing/model': 0.38842888800718356, 'learning_rate/qwen_vl_interface': 9.998871473307529e-06, 'learning_rate/action_model': 9.99887147330753e-05, 'epoch': 0.01}) +[2026-10-03 13:02:09,288][starVLA.training.train_starvla][INFO] - Step 135, Loss: {'train/loss': 0.7736939191818237, 'train/grad_norm_pre_clip': 23.645597457885742, 'timing/data': 0.0002996489929500967, 'timing/model': 0.3887996569974348, 'learning_rate/qwen_vl_interface': 9.998804115976907e-06, 'learning_rate/action_model': 9.998804115976908e-05, 'epoch': 0.01}) +[2026-10-03 13:02:09,985][starVLA.training.train_starvla][INFO] - Step 136, Loss: {'train/loss': 0.46707338094711304, 'train/grad_norm_pre_clip': 22.279680252075195, 'timing/data': 0.0003365699958521873, 'timing/model': 0.38830285699805245, 'learning_rate/qwen_vl_interface': 9.99873480659229e-06, 'learning_rate/action_model': 9.99873480659229e-05, 'epoch': 0.01}) +[2026-10-03 13:02:10,680][starVLA.training.train_starvla][INFO] - Step 137, Loss: {'train/loss': 0.9944355487823486, 'train/grad_norm_pre_clip': 38.04800796508789, 'timing/data': 0.0003990400000475347, 'timing/model': 0.38840502699895296, 'learning_rate/qwen_vl_interface': 9.99866354518217e-06, 'learning_rate/action_model': 9.99866354518217e-05, 'epoch': 0.01}) +[2026-10-03 13:02:11,565][starVLA.training.train_starvla][INFO] - Step 138, Loss: {'train/loss': 0.41032874584198, 'train/grad_norm_pre_clip': 37.37366485595703, 'timing/data': 0.00034200999652966857, 'timing/model': 0.5788188860024093, 'learning_rate/qwen_vl_interface': 9.99859033177584e-06, 'learning_rate/action_model': 9.99859033177584e-05, 'epoch': 0.01}) +[2026-10-03 13:02:12,261][starVLA.training.train_starvla][INFO] - Step 139, Loss: {'train/loss': 0.9816302061080933, 'train/grad_norm_pre_clip': 67.54622650146484, 'timing/data': 0.00030306998814921826, 'timing/model': 0.3886124170094263, 'learning_rate/qwen_vl_interface': 9.998515166403394e-06, 'learning_rate/action_model': 9.998515166403394e-05, 'epoch': 0.01}) +[2026-10-03 13:02:12,957][starVLA.training.train_starvla][INFO] - Step 140, Loss: {'train/loss': 1.4194951057434082, 'train/grad_norm_pre_clip': 106.24201965332031, 'timing/data': 0.0003325300058349967, 'timing/model': 0.3884204670030158, 'learning_rate/qwen_vl_interface': 9.99843804909573e-06, 'learning_rate/action_model': 9.99843804909573e-05, 'epoch': 0.01}) +[2026-10-03 13:02:13,652][starVLA.training.train_starvla][INFO] - Step 141, Loss: {'train/loss': 0.7379077672958374, 'train/grad_norm_pre_clip': 186.7466583251953, 'timing/data': 0.0002528199984226376, 'timing/model': 0.3884297070035245, 'learning_rate/qwen_vl_interface': 9.998358979884548e-06, 'learning_rate/action_model': 9.998358979884547e-05, 'epoch': 0.01}) +[2026-10-03 13:02:14,347][starVLA.training.train_starvla][INFO] - Step 142, Loss: {'train/loss': 0.9242198467254639, 'train/grad_norm_pre_clip': 53.50649642944336, 'timing/data': 0.000262079993262887, 'timing/model': 0.38831244700122625, 'learning_rate/qwen_vl_interface': 9.998277958802352e-06, 'learning_rate/action_model': 9.998277958802351e-05, 'epoch': 0.01}) +[2026-10-03 13:02:15,041][starVLA.training.train_starvla][INFO] - Step 143, Loss: {'train/loss': 1.3338916301727295, 'train/grad_norm_pre_clip': 129.68179321289062, 'timing/data': 0.00029663000896107405, 'timing/model': 0.388468256991473, 'learning_rate/qwen_vl_interface': 9.998194985882442e-06, 'learning_rate/action_model': 9.998194985882442e-05, 'epoch': 0.01}) +[2026-10-03 13:02:15,929][starVLA.training.train_starvla][INFO] - Step 144, Loss: {'train/loss': 1.4417117834091187, 'train/grad_norm_pre_clip': 161.3857421875, 'timing/data': 0.0002386999985901639, 'timing/model': 0.38859182699525263, 'learning_rate/qwen_vl_interface': 9.99811006115893e-06, 'learning_rate/action_model': 9.998110061158931e-05, 'epoch': 0.01}) +[2026-10-03 13:02:16,624][starVLA.training.train_starvla][INFO] - Step 145, Loss: {'train/loss': 1.2935717105865479, 'train/grad_norm_pre_clip': 58.97739791870117, 'timing/data': 0.00024033000227063894, 'timing/model': 0.3881784660043195, 'learning_rate/qwen_vl_interface': 9.998023184666726e-06, 'learning_rate/action_model': 9.998023184666726e-05, 'epoch': 0.01}) +[2026-10-03 13:02:17,318][starVLA.training.train_starvla][INFO] - Step 146, Loss: {'train/loss': 0.9358127117156982, 'train/grad_norm_pre_clip': 128.86016845703125, 'timing/data': 0.00024670000129844993, 'timing/model': 0.3882184079993749, 'learning_rate/qwen_vl_interface': 9.997934356441538e-06, 'learning_rate/action_model': 9.997934356441537e-05, 'epoch': 0.01}) +[2026-10-03 13:02:18,013][starVLA.training.train_starvla][INFO] - Step 147, Loss: {'train/loss': 1.1944007873535156, 'train/grad_norm_pre_clip': 154.5673370361328, 'timing/data': 0.0003094100102316588, 'timing/model': 0.38844646599318366, 'learning_rate/qwen_vl_interface': 9.997843576519879e-06, 'learning_rate/action_model': 9.997843576519878e-05, 'epoch': 0.01}) +[2026-10-03 13:02:18,710][starVLA.training.train_starvla][INFO] - Step 148, Loss: {'train/loss': 0.9515427947044373, 'train/grad_norm_pre_clip': 82.77271270751953, 'timing/data': 0.00026915999478660524, 'timing/model': 0.3885555160086369, 'learning_rate/qwen_vl_interface': 9.99775084493907e-06, 'learning_rate/action_model': 9.997750844939069e-05, 'epoch': 0.01}) +[2026-10-03 13:02:19,597][starVLA.training.train_starvla][INFO] - Step 149, Loss: {'train/loss': 1.029445767402649, 'train/grad_norm_pre_clip': 63.07467269897461, 'timing/data': 0.00023611000506207347, 'timing/model': 0.38790298700041603, 'learning_rate/qwen_vl_interface': 9.997656161737227e-06, 'learning_rate/action_model': 9.997656161737226e-05, 'epoch': 0.01}) +[2026-10-03 13:02:20,291][starVLA.training.train_starvla][INFO] - Step 150, Loss: {'train/loss': 0.44357946515083313, 'train/grad_norm_pre_clip': 13.055256843566895, 'timing/data': 0.00030370999593287706, 'timing/model': 0.3880391370039433, 'learning_rate/qwen_vl_interface': 9.99755952695327e-06, 'learning_rate/action_model': 9.997559526953269e-05, 'epoch': 0.01}) +[2026-10-03 13:02:20,986][starVLA.training.train_starvla][INFO] - Step 151, Loss: {'train/loss': 0.6376262903213501, 'train/grad_norm_pre_clip': 79.17427062988281, 'timing/data': 0.00029557000380009413, 'timing/model': 0.3894782270072028, 'learning_rate/qwen_vl_interface': 9.99746094062692e-06, 'learning_rate/action_model': 9.99746094062692e-05, 'epoch': 0.01}) +[2026-10-03 13:02:21,683][starVLA.training.train_starvla][INFO] - Step 152, Loss: {'train/loss': 0.6679190993309021, 'train/grad_norm_pre_clip': 61.06837463378906, 'timing/data': 0.00033672999416012317, 'timing/model': 0.38872324700059835, 'learning_rate/qwen_vl_interface': 9.997360402798706e-06, 'learning_rate/action_model': 9.997360402798705e-05, 'epoch': 0.01}) +[2026-10-03 13:02:22,378][starVLA.training.train_starvla][INFO] - Step 153, Loss: {'train/loss': 0.7897936105728149, 'train/grad_norm_pre_clip': 78.99576568603516, 'timing/data': 0.00020029000006616116, 'timing/model': 0.38863754700287245, 'learning_rate/qwen_vl_interface': 9.997257913509955e-06, 'learning_rate/action_model': 9.997257913509955e-05, 'epoch': 0.01}) +[2026-10-03 13:02:23,262][starVLA.training.train_starvla][INFO] - Step 154, Loss: {'train/loss': 0.7727910280227661, 'train/grad_norm_pre_clip': 73.67662811279297, 'timing/data': 0.00020809999841731042, 'timing/model': 0.5788225159922149, 'learning_rate/qwen_vl_interface': 9.997153472802794e-06, 'learning_rate/action_model': 9.997153472802795e-05, 'epoch': 0.01}) +[2026-10-03 13:02:23,957][starVLA.training.train_starvla][INFO] - Step 155, Loss: {'train/loss': 0.6069188117980957, 'train/grad_norm_pre_clip': 123.77772521972656, 'timing/data': 0.00028469000244513154, 'timing/model': 0.388916136987973, 'learning_rate/qwen_vl_interface': 9.997047080720157e-06, 'learning_rate/action_model': 9.997047080720158e-05, 'epoch': 0.01}) +[2026-10-03 13:02:24,653][starVLA.training.train_starvla][INFO] - Step 156, Loss: {'train/loss': 0.6939113140106201, 'train/grad_norm_pre_clip': 33.11494445800781, 'timing/data': 0.0002011600008700043, 'timing/model': 0.38811580699984916, 'learning_rate/qwen_vl_interface': 9.996938737305777e-06, 'learning_rate/action_model': 9.996938737305776e-05, 'epoch': 0.01}) +[2026-10-03 13:02:25,348][starVLA.training.train_starvla][INFO] - Step 157, Loss: {'train/loss': 0.2936363220214844, 'train/grad_norm_pre_clip': 86.5754623413086, 'timing/data': 0.00020452000899240375, 'timing/model': 0.38832393700431567, 'learning_rate/qwen_vl_interface': 9.996828442604189e-06, 'learning_rate/action_model': 9.99682844260419e-05, 'epoch': 0.01}) +[2026-10-03 13:02:26,042][starVLA.training.train_starvla][INFO] - Step 158, Loss: {'train/loss': 0.6498686075210571, 'train/grad_norm_pre_clip': 64.4959945678711, 'timing/data': 0.00021933999960310757, 'timing/model': 0.38811583699134644, 'learning_rate/qwen_vl_interface': 9.996716196660733e-06, 'learning_rate/action_model': 9.996716196660731e-05, 'epoch': 0.01}) +[2026-10-03 13:02:26,925][starVLA.training.train_starvla][INFO] - Step 159, Loss: {'train/loss': 1.0362468957901, 'train/grad_norm_pre_clip': 246.1795196533203, 'timing/data': 0.0003030500083696097, 'timing/model': 0.577960106995306, 'learning_rate/qwen_vl_interface': 9.996601999521546e-06, 'learning_rate/action_model': 9.996601999521546e-05, 'epoch': 0.01}) +[2026-10-03 13:02:27,622][starVLA.training.train_starvla][INFO] - Step 160, Loss: {'train/loss': 1.431508183479309, 'train/grad_norm_pre_clip': 198.86318969726562, 'timing/data': 0.0003753199998755008, 'timing/model': 0.3887548469938338, 'learning_rate/qwen_vl_interface': 9.996485851233572e-06, 'learning_rate/action_model': 9.996485851233572e-05, 'epoch': 0.01}) +[2026-10-03 13:02:28,317][starVLA.training.train_starvla][INFO] - Step 161, Loss: {'train/loss': 0.2991754412651062, 'train/grad_norm_pre_clip': 39.52566909790039, 'timing/data': 0.00026185000024270266, 'timing/model': 0.38824114699673373, 'learning_rate/qwen_vl_interface': 9.996367751844556e-06, 'learning_rate/action_model': 9.996367751844555e-05, 'epoch': 0.01}) +[2026-10-03 13:02:29,012][starVLA.training.train_starvla][INFO] - Step 162, Loss: {'train/loss': 1.1908650398254395, 'train/grad_norm_pre_clip': 124.80000305175781, 'timing/data': 0.00021102000027894974, 'timing/model': 0.38844115698884707, 'learning_rate/qwen_vl_interface': 9.996247701403043e-06, 'learning_rate/action_model': 9.996247701403042e-05, 'epoch': 0.01}) +[2026-10-03 13:02:29,706][starVLA.training.train_starvla][INFO] - Step 163, Loss: {'train/loss': 0.07822898030281067, 'train/grad_norm_pre_clip': 40.63560485839844, 'timing/data': 0.0003566000086721033, 'timing/model': 0.3888486270006979, 'learning_rate/qwen_vl_interface': 9.99612569995838e-06, 'learning_rate/action_model': 9.99612569995838e-05, 'epoch': 0.01}) +[2026-10-03 13:02:30,596][starVLA.training.train_starvla][INFO] - Step 164, Loss: {'train/loss': 1.142660140991211, 'train/grad_norm_pre_clip': 375.8200988769531, 'timing/data': 0.00027554000553209335, 'timing/model': 0.5803180060029263, 'learning_rate/qwen_vl_interface': 9.99600174756072e-06, 'learning_rate/action_model': 9.99600174756072e-05, 'epoch': 0.01}) +[2026-10-03 13:02:31,290][starVLA.training.train_starvla][INFO] - Step 165, Loss: {'train/loss': 0.30247563123703003, 'train/grad_norm_pre_clip': 59.5789680480957, 'timing/data': 0.0002011299948208034, 'timing/model': 0.3882941070041852, 'learning_rate/qwen_vl_interface': 9.995875844261012e-06, 'learning_rate/action_model': 9.995875844261013e-05, 'epoch': 0.01}) +[2026-10-03 13:02:31,982][starVLA.training.train_starvla][INFO] - Step 166, Loss: {'train/loss': 0.5836406350135803, 'train/grad_norm_pre_clip': 149.66818237304688, 'timing/data': 0.00032666999322827905, 'timing/model': 0.3883993769995868, 'learning_rate/qwen_vl_interface': 9.995747990111015e-06, 'learning_rate/action_model': 9.995747990111013e-05, 'epoch': 0.01}) +[2026-10-03 13:02:32,675][starVLA.training.train_starvla][INFO] - Step 167, Loss: {'train/loss': 0.6558711528778076, 'train/grad_norm_pre_clip': 80.61624145507812, 'timing/data': 0.00025427000946365297, 'timing/model': 0.3873938269971404, 'learning_rate/qwen_vl_interface': 9.99561818516328e-06, 'learning_rate/action_model': 9.99561818516328e-05, 'epoch': 0.01}) +[2026-10-03 13:02:33,373][starVLA.training.train_starvla][INFO] - Step 168, Loss: {'train/loss': 1.0931808948516846, 'train/grad_norm_pre_clip': 120.73788452148438, 'timing/data': 0.00026761999470181763, 'timing/model': 0.38876458699814975, 'learning_rate/qwen_vl_interface': 9.995486429471166e-06, 'learning_rate/action_model': 9.995486429471166e-05, 'epoch': 0.01}) +[2026-10-03 13:02:34,068][starVLA.training.train_starvla][INFO] - Step 169, Loss: {'train/loss': 1.4179797172546387, 'train/grad_norm_pre_clip': 350.51580810546875, 'timing/data': 0.0002015000063693151, 'timing/model': 0.38833295699441805, 'learning_rate/qwen_vl_interface': 9.995352723088833e-06, 'learning_rate/action_model': 9.995352723088834e-05, 'epoch': 0.01}) +[2026-10-03 13:02:34,957][starVLA.training.train_starvla][INFO] - Step 170, Loss: {'train/loss': 0.8582022190093994, 'train/grad_norm_pre_clip': 157.86289978027344, 'timing/data': 0.0002075200027320534, 'timing/model': 0.38792895698861685, 'learning_rate/qwen_vl_interface': 9.995217066071245e-06, 'learning_rate/action_model': 9.995217066071246e-05, 'epoch': 0.01}) +[2026-10-03 13:02:35,652][starVLA.training.train_starvla][INFO] - Step 171, Loss: {'train/loss': 0.5273915529251099, 'train/grad_norm_pre_clip': 79.89546966552734, 'timing/data': 0.00029764999635517597, 'timing/model': 0.388725106997299, 'learning_rate/qwen_vl_interface': 9.995079458474163e-06, 'learning_rate/action_model': 9.995079458474164e-05, 'epoch': 0.01}) +[2026-10-03 13:02:36,352][starVLA.training.train_starvla][INFO] - Step 172, Loss: {'train/loss': 0.44828447699546814, 'train/grad_norm_pre_clip': 37.31499099731445, 'timing/data': 0.00036407999868970364, 'timing/model': 0.390845335990889, 'learning_rate/qwen_vl_interface': 9.994939900354154e-06, 'learning_rate/action_model': 9.994939900354154e-05, 'epoch': 0.01}) +[2026-10-03 13:02:37,047][starVLA.training.train_starvla][INFO] - Step 173, Loss: {'train/loss': 0.25596728920936584, 'train/grad_norm_pre_clip': 29.847829818725586, 'timing/data': 0.00020107999444007874, 'timing/model': 0.38873654599592555, 'learning_rate/qwen_vl_interface': 9.994798391768582e-06, 'learning_rate/action_model': 9.994798391768582e-05, 'epoch': 0.01}) +[2026-10-03 13:02:37,740][starVLA.training.train_starvla][INFO] - Step 174, Loss: {'train/loss': 0.3510880768299103, 'train/grad_norm_pre_clip': 15.02023696899414, 'timing/data': 0.0003104999923380092, 'timing/model': 0.38808349799364805, 'learning_rate/qwen_vl_interface': 9.99465493277562e-06, 'learning_rate/action_model': 9.994654932775618e-05, 'epoch': 0.01}) +[2026-10-03 13:02:38,629][starVLA.training.train_starvla][INFO] - Step 175, Loss: {'train/loss': 0.3253973126411438, 'train/grad_norm_pre_clip': 58.103179931640625, 'timing/data': 0.0001973899925360456, 'timing/model': 0.5822108849970391, 'learning_rate/qwen_vl_interface': 9.994509523434235e-06, 'learning_rate/action_model': 9.994509523434235e-05, 'epoch': 0.01}) +[2026-10-03 13:02:39,326][starVLA.training.train_starvla][INFO] - Step 176, Loss: {'train/loss': 0.886133074760437, 'train/grad_norm_pre_clip': 65.48111724853516, 'timing/data': 0.0004390499962028116, 'timing/model': 0.3883412169961957, 'learning_rate/qwen_vl_interface': 9.994362163804202e-06, 'learning_rate/action_model': 9.9943621638042e-05, 'epoch': 0.01}) +[2026-10-03 13:02:40,021][starVLA.training.train_starvla][INFO] - Step 177, Loss: {'train/loss': 1.9490811824798584, 'train/grad_norm_pre_clip': 289.07220458984375, 'timing/data': 0.0004196399968350306, 'timing/model': 0.3883058370120125, 'learning_rate/qwen_vl_interface': 9.994212853946092e-06, 'learning_rate/action_model': 9.994212853946092e-05, 'epoch': 0.01}) +[2026-10-03 13:02:40,716][starVLA.training.train_starvla][INFO] - Step 178, Loss: {'train/loss': 0.412814199924469, 'train/grad_norm_pre_clip': 37.959983825683594, 'timing/data': 0.0002603399916552007, 'timing/model': 0.38858093800081406, 'learning_rate/qwen_vl_interface': 9.994061593921283e-06, 'learning_rate/action_model': 9.994061593921282e-05, 'epoch': 0.01}) +[2026-10-03 13:02:41,410][starVLA.training.train_starvla][INFO] - Step 179, Loss: {'train/loss': 1.1500310897827148, 'train/grad_norm_pre_clip': 106.85261535644531, 'timing/data': 0.00031066899828147143, 'timing/model': 0.38854222701047547, 'learning_rate/qwen_vl_interface': 9.993908383791951e-06, 'learning_rate/action_model': 9.993908383791951e-05, 'epoch': 0.01}) +[2026-10-03 13:02:42,108][starVLA.training.train_starvla][INFO] - Step 180, Loss: {'train/loss': 0.42651063203811646, 'train/grad_norm_pre_clip': 25.039722442626953, 'timing/data': 0.00020291999680921435, 'timing/model': 0.38873809699725825, 'learning_rate/qwen_vl_interface': 9.993753223621076e-06, 'learning_rate/action_model': 9.993753223621077e-05, 'epoch': 0.01}) +[2026-10-03 13:02:42,994][starVLA.training.train_starvla][INFO] - Step 181, Loss: {'train/loss': 0.6499403715133667, 'train/grad_norm_pre_clip': 84.5020980834961, 'timing/data': 0.0002577300037955865, 'timing/model': 0.38820557699364144, 'learning_rate/qwen_vl_interface': 9.99359611347244e-06, 'learning_rate/action_model': 9.993596113472439e-05, 'epoch': 0.01}) +[2026-10-03 13:02:43,689][starVLA.training.train_starvla][INFO] - Step 182, Loss: {'train/loss': 0.8243908882141113, 'train/grad_norm_pre_clip': 87.06819152832031, 'timing/data': 0.0003234699979657307, 'timing/model': 0.38969035699847154, 'learning_rate/qwen_vl_interface': 9.99343705341062e-06, 'learning_rate/action_model': 9.99343705341062e-05, 'epoch': 0.01}) +[2026-10-03 13:02:44,384][starVLA.training.train_starvla][INFO] - Step 183, Loss: {'train/loss': 0.9446181058883667, 'train/grad_norm_pre_clip': 71.01475524902344, 'timing/data': 0.0002912700001616031, 'timing/model': 0.38825234699470457, 'learning_rate/qwen_vl_interface': 9.993276043501004e-06, 'learning_rate/action_model': 9.993276043501004e-05, 'epoch': 0.01}) +[2026-10-03 13:02:45,081][starVLA.training.train_starvla][INFO] - Step 184, Loss: {'train/loss': 0.4749097228050232, 'train/grad_norm_pre_clip': 117.4572525024414, 'timing/data': 0.000305499997921288, 'timing/model': 0.3886114670021925, 'learning_rate/qwen_vl_interface': 9.993113083809776e-06, 'learning_rate/action_model': 9.993113083809775e-05, 'epoch': 0.01}) +[2026-10-03 13:02:45,776][starVLA.training.train_starvla][INFO] - Step 185, Loss: {'train/loss': 1.1124507188796997, 'train/grad_norm_pre_clip': 99.12108612060547, 'timing/data': 0.00023515999782830477, 'timing/model': 0.3887592169921845, 'learning_rate/qwen_vl_interface': 9.992948174403921e-06, 'learning_rate/action_model': 9.992948174403922e-05, 'epoch': 0.01}) +[2026-10-03 13:02:46,664][starVLA.training.train_starvla][INFO] - Step 186, Loss: {'train/loss': 0.8546670079231262, 'train/grad_norm_pre_clip': 28.204326629638672, 'timing/data': 0.00020769999537151307, 'timing/model': 0.38833813701057807, 'learning_rate/qwen_vl_interface': 9.99278131535123e-06, 'learning_rate/action_model': 9.992781315351229e-05, 'epoch': 0.01}) +[2026-10-03 13:02:47,359][starVLA.training.train_starvla][INFO] - Step 187, Loss: {'train/loss': 0.7708798050880432, 'train/grad_norm_pre_clip': 50.7370719909668, 'timing/data': 0.0002501900016795844, 'timing/model': 0.38869038700067904, 'learning_rate/qwen_vl_interface': 9.992612506720289e-06, 'learning_rate/action_model': 9.99261250672029e-05, 'epoch': 0.01}) +[2026-10-03 13:02:48,057][starVLA.training.train_starvla][INFO] - Step 188, Loss: {'train/loss': 0.859708309173584, 'train/grad_norm_pre_clip': 20.789913177490234, 'timing/data': 0.00033005999284796417, 'timing/model': 0.38882288700551726, 'learning_rate/qwen_vl_interface': 9.992441748580492e-06, 'learning_rate/action_model': 9.99244174858049e-05, 'epoch': 0.01}) +[2026-10-03 13:02:48,752][starVLA.training.train_starvla][INFO] - Step 189, Loss: {'train/loss': 0.5800954103469849, 'train/grad_norm_pre_clip': 62.858638763427734, 'timing/data': 0.00032637998810969293, 'timing/model': 0.3881446059967857, 'learning_rate/qwen_vl_interface': 9.992269041002028e-06, 'learning_rate/action_model': 9.992269041002027e-05, 'epoch': 0.01}) +[2026-10-03 13:02:49,447][starVLA.training.train_starvla][INFO] - Step 190, Loss: {'train/loss': 0.3852205276489258, 'train/grad_norm_pre_clip': 22.291208267211914, 'timing/data': 0.0002010200114455074, 'timing/model': 0.38886362699849997, 'learning_rate/qwen_vl_interface': 9.992094384055894e-06, 'learning_rate/action_model': 9.992094384055895e-05, 'epoch': 0.01}) +[2026-10-03 13:02:50,332][starVLA.training.train_starvla][INFO] - Step 191, Loss: {'train/loss': 1.5948293209075928, 'train/grad_norm_pre_clip': 116.52348327636719, 'timing/data': 0.00020090999896638095, 'timing/model': 0.5794295959931333, 'learning_rate/qwen_vl_interface': 9.991917777813884e-06, 'learning_rate/action_model': 9.991917777813884e-05, 'epoch': 0.01}) +[2026-10-03 13:02:51,030][starVLA.training.train_starvla][INFO] - Step 192, Loss: {'train/loss': 1.3351330757141113, 'train/grad_norm_pre_clip': 177.1553497314453, 'timing/data': 0.0003370800113771111, 'timing/model': 0.3889843469951302, 'learning_rate/qwen_vl_interface': 9.991739222348593e-06, 'learning_rate/action_model': 9.991739222348592e-05, 'epoch': 0.01}) +[2026-10-03 13:02:51,729][starVLA.training.train_starvla][INFO] - Step 193, Loss: {'train/loss': 0.5951238870620728, 'train/grad_norm_pre_clip': 45.46243667602539, 'timing/data': 0.000281530010397546, 'timing/model': 0.3874791670095874, 'learning_rate/qwen_vl_interface': 9.991558717733418e-06, 'learning_rate/action_model': 9.991558717733417e-05, 'epoch': 0.01}) +[2026-10-03 13:02:52,421][starVLA.training.train_starvla][INFO] - Step 194, Loss: {'train/loss': 0.8981634378433228, 'train/grad_norm_pre_clip': 42.49319839477539, 'timing/data': 0.00033160999009851366, 'timing/model': 0.38738252699840814, 'learning_rate/qwen_vl_interface': 9.99137626404256e-06, 'learning_rate/action_model': 9.991376264042559e-05, 'epoch': 0.01}) +[2026-10-03 13:02:53,115][starVLA.training.train_starvla][INFO] - Step 195, Loss: {'train/loss': 0.7533630728721619, 'train/grad_norm_pre_clip': 36.69656753540039, 'timing/data': 0.00026463999529369175, 'timing/model': 0.38817476700933184, 'learning_rate/qwen_vl_interface': 9.991191861351015e-06, 'learning_rate/action_model': 9.991191861351014e-05, 'epoch': 0.01}) +[2026-10-03 13:02:53,811][starVLA.training.train_starvla][INFO] - Step 196, Loss: {'train/loss': 0.6578251719474792, 'train/grad_norm_pre_clip': 38.131126403808594, 'timing/data': 0.0003643000090960413, 'timing/model': 0.388360607001232, 'learning_rate/qwen_vl_interface': 9.991005509734587e-06, 'learning_rate/action_model': 9.991005509734587e-05, 'epoch': 0.01}) +[2026-10-03 13:02:54,695][starVLA.training.train_starvla][INFO] - Step 197, Loss: {'train/loss': 0.7668672800064087, 'train/grad_norm_pre_clip': 50.19002151489258, 'timing/data': 0.0002898099919548258, 'timing/model': 0.38829480699496344, 'learning_rate/qwen_vl_interface': 9.990817209269878e-06, 'learning_rate/action_model': 9.990817209269878e-05, 'epoch': 0.01}) +[2026-10-03 13:02:55,389][starVLA.training.train_starvla][INFO] - Step 198, Loss: {'train/loss': 0.29264432191848755, 'train/grad_norm_pre_clip': 19.05173110961914, 'timing/data': 0.0003571400011423975, 'timing/model': 0.3879656469944166, 'learning_rate/qwen_vl_interface': 9.990626960034291e-06, 'learning_rate/action_model': 9.990626960034291e-05, 'epoch': 0.01}) +[2026-10-03 13:02:56,082][starVLA.training.train_starvla][INFO] - Step 199, Loss: {'train/loss': 1.1054306030273438, 'train/grad_norm_pre_clip': 56.939002990722656, 'timing/data': 0.0003062799951294437, 'timing/model': 0.38788971801113803, 'learning_rate/qwen_vl_interface': 9.990434762106029e-06, 'learning_rate/action_model': 9.990434762106028e-05, 'epoch': 0.01}) +[2026-10-03 13:02:56,780][starVLA.training.train_starvla][INFO] - Step 200, Loss: {'train/loss': 0.8170733451843262, 'train/grad_norm_pre_clip': 35.306827545166016, 'timing/data': 0.00031627999851480126, 'timing/model': 0.3884263079962693, 'learning_rate/qwen_vl_interface': 9.990240615564099e-06, 'learning_rate/action_model': 9.990240615564099e-05, 'epoch': 0.01}) +[2026-10-03 13:02:57,474][starVLA.training.train_starvla][INFO] - Step 201, Loss: {'train/loss': 0.7038578987121582, 'train/grad_norm_pre_clip': 35.334774017333984, 'timing/data': 0.000235440005781129, 'timing/model': 0.3882254480122356, 'learning_rate/qwen_vl_interface': 9.990044520488305e-06, 'learning_rate/action_model': 9.990044520488305e-05, 'epoch': 0.01}) +[2026-10-03 13:02:58,363][starVLA.training.train_starvla][INFO] - Step 202, Loss: {'train/loss': 0.4965749979019165, 'train/grad_norm_pre_clip': 20.093271255493164, 'timing/data': 0.0003067299985559657, 'timing/model': 0.38855952701123897, 'learning_rate/qwen_vl_interface': 9.989846476959258e-06, 'learning_rate/action_model': 9.989846476959257e-05, 'epoch': 0.01}) +[2026-10-03 13:02:59,057][starVLA.training.train_starvla][INFO] - Step 203, Loss: {'train/loss': 0.40718019008636475, 'train/grad_norm_pre_clip': 27.831588745117188, 'timing/data': 0.00020243000471964478, 'timing/model': 0.38863671700528357, 'learning_rate/qwen_vl_interface': 9.989646485058361e-06, 'learning_rate/action_model': 9.989646485058361e-05, 'epoch': 0.01}) +[2026-10-03 13:02:59,754][starVLA.training.train_starvla][INFO] - Step 204, Loss: {'train/loss': 0.7235510945320129, 'train/grad_norm_pre_clip': 75.7652816772461, 'timing/data': 0.0003626399993663654, 'timing/model': 0.3884268169931602, 'learning_rate/qwen_vl_interface': 9.98944454486783e-06, 'learning_rate/action_model': 9.989444544867829e-05, 'epoch': 0.01}) +[2026-10-03 13:03:00,448][starVLA.training.train_starvla][INFO] - Step 205, Loss: {'train/loss': 0.8104597330093384, 'train/grad_norm_pre_clip': 72.176513671875, 'timing/data': 0.00036808000004384667, 'timing/model': 0.38834986799338367, 'learning_rate/qwen_vl_interface': 9.989240656470667e-06, 'learning_rate/action_model': 9.989240656470666e-05, 'epoch': 0.02}) +[2026-10-03 13:03:01,143][starVLA.training.train_starvla][INFO] - Step 206, Loss: {'train/loss': 0.6757882237434387, 'train/grad_norm_pre_clip': 50.34861373901367, 'timing/data': 0.000391419991501607, 'timing/model': 0.3880898770003114, 'learning_rate/qwen_vl_interface': 9.98903481995069e-06, 'learning_rate/action_model': 9.989034819950689e-05, 'epoch': 0.02}) +[2026-10-03 13:03:02,026][starVLA.training.train_starvla][INFO] - Step 207, Loss: {'train/loss': 0.31256648898124695, 'train/grad_norm_pre_clip': 14.56369686126709, 'timing/data': 0.00029679000726900995, 'timing/model': 0.5766757460078225, 'learning_rate/qwen_vl_interface': 9.988827035392506e-06, 'learning_rate/action_model': 9.988827035392507e-05, 'epoch': 0.02}) +[2026-10-03 13:03:02,723][starVLA.training.train_starvla][INFO] - Step 208, Loss: {'train/loss': 0.419201523065567, 'train/grad_norm_pre_clip': 23.056354522705078, 'timing/data': 0.00039164999907370657, 'timing/model': 0.3886140769900521, 'learning_rate/qwen_vl_interface': 9.988617302881529e-06, 'learning_rate/action_model': 9.988617302881529e-05, 'epoch': 0.02}) +[2026-10-03 13:03:03,417][starVLA.training.train_starvla][INFO] - Step 209, Loss: {'train/loss': 0.742796003818512, 'train/grad_norm_pre_clip': 119.9793930053711, 'timing/data': 0.00030951001099310815, 'timing/model': 0.3879073770076502, 'learning_rate/qwen_vl_interface': 9.988405622503972e-06, 'learning_rate/action_model': 9.988405622503971e-05, 'epoch': 0.02}) +[2026-10-03 13:03:04,111][starVLA.training.train_starvla][INFO] - Step 210, Loss: {'train/loss': 0.45126986503601074, 'train/grad_norm_pre_clip': 46.52666473388672, 'timing/data': 0.00035276000562589616, 'timing/model': 0.3881959580030525, 'learning_rate/qwen_vl_interface': 9.988191994346848e-06, 'learning_rate/action_model': 9.988191994346848e-05, 'epoch': 0.02}) +[2026-10-03 13:03:04,805][starVLA.training.train_starvla][INFO] - Step 211, Loss: {'train/loss': 0.398368239402771, 'train/grad_norm_pre_clip': 27.907604217529297, 'timing/data': 0.0002602800086606294, 'timing/model': 0.3879009960073745, 'learning_rate/qwen_vl_interface': 9.987976418497974e-06, 'learning_rate/action_model': 9.987976418497972e-05, 'epoch': 0.02}) +[2026-10-03 13:03:05,695][starVLA.training.train_starvla][INFO] - Step 212, Loss: {'train/loss': 0.28015008568763733, 'train/grad_norm_pre_clip': 19.164119720458984, 'timing/data': 0.00032796000596135855, 'timing/model': 0.5810766350041376, 'learning_rate/qwen_vl_interface': 9.987758895045964e-06, 'learning_rate/action_model': 9.987758895045963e-05, 'epoch': 0.02}) +[2026-10-03 13:03:06,391][starVLA.training.train_starvla][INFO] - Step 213, Loss: {'train/loss': 0.49754685163497925, 'train/grad_norm_pre_clip': 29.985118865966797, 'timing/data': 0.0002983800077345222, 'timing/model': 0.3885062469926197, 'learning_rate/qwen_vl_interface': 9.98753942408023e-06, 'learning_rate/action_model': 9.98753942408023e-05, 'epoch': 0.02}) +[2026-10-03 13:03:07,085][starVLA.training.train_starvla][INFO] - Step 214, Loss: {'train/loss': 1.2551500797271729, 'train/grad_norm_pre_clip': 60.158939361572266, 'timing/data': 0.00026652000087779015, 'timing/model': 0.3884356370108435, 'learning_rate/qwen_vl_interface': 9.987318005690992e-06, 'learning_rate/action_model': 9.987318005690992e-05, 'epoch': 0.02}) +[2026-10-03 13:03:07,777][starVLA.training.train_starvla][INFO] - Step 215, Loss: {'train/loss': 0.49186018109321594, 'train/grad_norm_pre_clip': 20.993541717529297, 'timing/data': 0.00030638999305665493, 'timing/model': 0.38668508699629456, 'learning_rate/qwen_vl_interface': 9.987094639969268e-06, 'learning_rate/action_model': 9.987094639969268e-05, 'epoch': 0.02}) +[2026-10-03 13:03:08,473][starVLA.training.train_starvla][INFO] - Step 216, Loss: {'train/loss': 0.3139975666999817, 'train/grad_norm_pre_clip': 14.995645523071289, 'timing/data': 0.0003103700000792742, 'timing/model': 0.38781261700205505, 'learning_rate/qwen_vl_interface': 9.986869327006872e-06, 'learning_rate/action_model': 9.986869327006873e-05, 'epoch': 0.02}) +[2026-10-03 13:03:09,360][starVLA.training.train_starvla][INFO] - Step 217, Loss: {'train/loss': 0.4990745782852173, 'train/grad_norm_pre_clip': 41.092159271240234, 'timing/data': 0.00030872999923303723, 'timing/model': 0.5819343559996923, 'learning_rate/qwen_vl_interface': 9.986642066896424e-06, 'learning_rate/action_model': 9.986642066896423e-05, 'epoch': 0.02}) +[2026-10-03 13:03:10,053][starVLA.training.train_starvla][INFO] - Step 218, Loss: {'train/loss': 0.8877319693565369, 'train/grad_norm_pre_clip': 21.29277801513672, 'timing/data': 0.0002021400141529739, 'timing/model': 0.3877161170094041, 'learning_rate/qwen_vl_interface': 9.986412859731341e-06, 'learning_rate/action_model': 9.98641285973134e-05, 'epoch': 0.02}) +[2026-10-03 13:03:10,747][starVLA.training.train_starvla][INFO] - Step 219, Loss: {'train/loss': 0.49012142419815063, 'train/grad_norm_pre_clip': 25.142446517944336, 'timing/data': 0.00026821999927051365, 'timing/model': 0.38802381699497346, 'learning_rate/qwen_vl_interface': 9.986181705605839e-06, 'learning_rate/action_model': 9.986181705605839e-05, 'epoch': 0.02}) +[2026-10-03 13:03:11,443][starVLA.training.train_starvla][INFO] - Step 220, Loss: {'train/loss': 0.4013317823410034, 'train/grad_norm_pre_clip': 69.7081298828125, 'timing/data': 0.0002991899964399636, 'timing/model': 0.3881890670018038, 'learning_rate/qwen_vl_interface': 9.98594860461494e-06, 'learning_rate/action_model': 9.98594860461494e-05, 'epoch': 0.02}) +[2026-10-03 13:03:12,137][starVLA.training.train_starvla][INFO] - Step 221, Loss: {'train/loss': 0.2382410615682602, 'train/grad_norm_pre_clip': 28.2508487701416, 'timing/data': 0.0002557599946158007, 'timing/model': 0.3878749669966055, 'learning_rate/qwen_vl_interface': 9.985713556854461e-06, 'learning_rate/action_model': 9.98571355685446e-05, 'epoch': 0.02}) +[2026-10-03 13:03:12,830][starVLA.training.train_starvla][INFO] - Step 222, Loss: {'train/loss': 0.9051055908203125, 'train/grad_norm_pre_clip': 40.998931884765625, 'timing/data': 0.00025918999745044857, 'timing/model': 0.38774213699798565, 'learning_rate/qwen_vl_interface': 9.985476562421024e-06, 'learning_rate/action_model': 9.985476562421024e-05, 'epoch': 0.02}) +[2026-10-03 13:03:13,718][starVLA.training.train_starvla][INFO] - Step 223, Loss: {'train/loss': 0.37479084730148315, 'train/grad_norm_pre_clip': 30.483802795410156, 'timing/data': 0.0003379600093467161, 'timing/model': 0.3892036769975675, 'learning_rate/qwen_vl_interface': 9.985237621412047e-06, 'learning_rate/action_model': 9.985237621412046e-05, 'epoch': 0.02}) +[2026-10-03 13:03:14,416][starVLA.training.train_starvla][INFO] - Step 224, Loss: {'train/loss': 0.6518675088882446, 'train/grad_norm_pre_clip': 114.10431671142578, 'timing/data': 0.00023954000789672136, 'timing/model': 0.38894392699876335, 'learning_rate/qwen_vl_interface': 9.984996733925749e-06, 'learning_rate/action_model': 9.984996733925748e-05, 'epoch': 0.02}) +[2026-10-03 13:03:15,113][starVLA.training.train_starvla][INFO] - Step 225, Loss: {'train/loss': 0.400718629360199, 'train/grad_norm_pre_clip': 26.19940757751465, 'timing/data': 0.0003906700003426522, 'timing/model': 0.3893819880031515, 'learning_rate/qwen_vl_interface': 9.98475390006115e-06, 'learning_rate/action_model': 9.98475390006115e-05, 'epoch': 0.02}) +[2026-10-03 13:03:15,809][starVLA.training.train_starvla][INFO] - Step 226, Loss: {'train/loss': 0.660377562046051, 'train/grad_norm_pre_clip': 39.992122650146484, 'timing/data': 0.0002916699886554852, 'timing/model': 0.3892633080104133, 'learning_rate/qwen_vl_interface': 9.98450911991807e-06, 'learning_rate/action_model': 9.98450911991807e-05, 'epoch': 0.02}) +[2026-10-03 13:03:16,505][starVLA.training.train_starvla][INFO] - Step 227, Loss: {'train/loss': 0.34467387199401855, 'train/grad_norm_pre_clip': 63.276859283447266, 'timing/data': 0.00023565000446978956, 'timing/model': 0.38925906700023916, 'learning_rate/qwen_vl_interface': 9.98426239359713e-06, 'learning_rate/action_model': 9.984262393597129e-05, 'epoch': 0.02}) +[2026-10-03 13:03:17,396][starVLA.training.train_starvla][INFO] - Step 228, Loss: {'train/loss': 0.38264748454093933, 'train/grad_norm_pre_clip': 37.948326110839844, 'timing/data': 0.0002925099979620427, 'timing/model': 0.5813881760113873, 'learning_rate/qwen_vl_interface': 9.984013721199747e-06, 'learning_rate/action_model': 9.984013721199747e-05, 'epoch': 0.02}) +[2026-10-03 13:03:18,092][starVLA.training.train_starvla][INFO] - Step 229, Loss: {'train/loss': 0.3379967510700226, 'train/grad_norm_pre_clip': 41.20539474487305, 'timing/data': 0.0002575100079411641, 'timing/model': 0.38896921699051745, 'learning_rate/qwen_vl_interface': 9.983763102828143e-06, 'learning_rate/action_model': 9.983763102828143e-05, 'epoch': 0.02}) +[2026-10-03 13:03:18,788][starVLA.training.train_starvla][INFO] - Step 230, Loss: {'train/loss': 0.27704837918281555, 'train/grad_norm_pre_clip': 50.214908599853516, 'timing/data': 0.0003888600040227175, 'timing/model': 0.3887438569945516, 'learning_rate/qwen_vl_interface': 9.983510538585336e-06, 'learning_rate/action_model': 9.983510538585337e-05, 'epoch': 0.02}) +[2026-10-03 13:03:19,484][starVLA.training.train_starvla][INFO] - Step 231, Loss: {'train/loss': 0.5175377726554871, 'train/grad_norm_pre_clip': 72.28834533691406, 'timing/data': 0.00020074000349268317, 'timing/model': 0.38943214599567, 'learning_rate/qwen_vl_interface': 9.98325602857515e-06, 'learning_rate/action_model': 9.98325602857515e-05, 'epoch': 0.02}) +[2026-10-03 13:03:20,181][starVLA.training.train_starvla][INFO] - Step 232, Loss: {'train/loss': 0.583541750907898, 'train/grad_norm_pre_clip': 110.58375549316406, 'timing/data': 0.00037174001045059413, 'timing/model': 0.3888082360062981, 'learning_rate/qwen_vl_interface': 9.9829995729022e-06, 'learning_rate/action_model': 9.982999572902199e-05, 'epoch': 0.02}) +[2026-10-03 13:03:20,877][starVLA.training.train_starvla][INFO] - Step 233, Loss: {'train/loss': 0.8408575057983398, 'train/grad_norm_pre_clip': 127.25196075439453, 'timing/data': 0.0002858400112017989, 'timing/model': 0.38888704699638765, 'learning_rate/qwen_vl_interface': 9.982741171671905e-06, 'learning_rate/action_model': 9.982741171671906e-05, 'epoch': 0.02}) +[2026-10-03 13:03:21,769][starVLA.training.train_starvla][INFO] - Step 234, Loss: {'train/loss': 0.42655372619628906, 'train/grad_norm_pre_clip': 45.25144577026367, 'timing/data': 0.0002755800087470561, 'timing/model': 0.3892895570024848, 'learning_rate/qwen_vl_interface': 9.982480824990489e-06, 'learning_rate/action_model': 9.982480824990488e-05, 'epoch': 0.02}) +[2026-10-03 13:03:22,464][starVLA.training.train_starvla][INFO] - Step 235, Loss: {'train/loss': 0.39883604645729065, 'train/grad_norm_pre_clip': 41.19125747680664, 'timing/data': 0.0002903400018112734, 'timing/model': 0.3892742169991834, 'learning_rate/qwen_vl_interface': 9.982218532964963e-06, 'learning_rate/action_model': 9.982218532964964e-05, 'epoch': 0.02}) +[2026-10-03 13:03:23,162][starVLA.training.train_starvla][INFO] - Step 236, Loss: {'train/loss': 0.5059908032417297, 'train/grad_norm_pre_clip': 170.02438354492188, 'timing/data': 0.00028105999808758497, 'timing/model': 0.38949659699574113, 'learning_rate/qwen_vl_interface': 9.981954295703152e-06, 'learning_rate/action_model': 9.981954295703152e-05, 'epoch': 0.02}) +[2026-10-03 13:03:23,859][starVLA.training.train_starvla][INFO] - Step 237, Loss: {'train/loss': 0.12000759690999985, 'train/grad_norm_pre_clip': 16.132183074951172, 'timing/data': 0.0003259700024500489, 'timing/model': 0.38941214700753335, 'learning_rate/qwen_vl_interface': 9.981688113313672e-06, 'learning_rate/action_model': 9.981688113313672e-05, 'epoch': 0.02}) +[2026-10-03 13:03:24,556][starVLA.training.train_starvla][INFO] - Step 238, Loss: {'train/loss': 0.8727268576622009, 'train/grad_norm_pre_clip': 62.63997268676758, 'timing/data': 0.00023841999063733965, 'timing/model': 0.38975189700431656, 'learning_rate/qwen_vl_interface': 9.98141998590594e-06, 'learning_rate/action_model': 9.981419985905939e-05, 'epoch': 0.02}) +[2026-10-03 13:03:25,448][starVLA.training.train_starvla][INFO] - Step 239, Loss: {'train/loss': 0.675549328327179, 'train/grad_norm_pre_clip': 130.13905334472656, 'timing/data': 0.0003748699964489788, 'timing/model': 0.3892678470001556, 'learning_rate/qwen_vl_interface': 9.981149913590171e-06, 'learning_rate/action_model': 9.981149913590172e-05, 'epoch': 0.02}) +[2026-10-03 13:03:26,148][starVLA.training.train_starvla][INFO] - Step 240, Loss: {'train/loss': 0.5196146965026855, 'train/grad_norm_pre_clip': 231.67897033691406, 'timing/data': 0.00038536000647582114, 'timing/model': 0.3903881360020023, 'learning_rate/qwen_vl_interface': 9.980877896477386e-06, 'learning_rate/action_model': 9.980877896477386e-05, 'epoch': 0.02}) +[2026-10-03 13:03:26,844][starVLA.training.train_starvla][INFO] - Step 241, Loss: {'train/loss': 0.39299964904785156, 'train/grad_norm_pre_clip': 65.18387603759766, 'timing/data': 0.0002914000069722533, 'timing/model': 0.38926447700941935, 'learning_rate/qwen_vl_interface': 9.980603934679397e-06, 'learning_rate/action_model': 9.980603934679398e-05, 'epoch': 0.02}) +[2026-10-03 13:03:27,540][starVLA.training.train_starvla][INFO] - Step 242, Loss: {'train/loss': 0.4368375539779663, 'train/grad_norm_pre_clip': 60.53272247314453, 'timing/data': 0.0002020499960053712, 'timing/model': 0.3893190379894804, 'learning_rate/qwen_vl_interface': 9.980328028308824e-06, 'learning_rate/action_model': 9.980328028308824e-05, 'epoch': 0.02}) +[2026-10-03 13:03:28,236][starVLA.training.train_starvla][INFO] - Step 243, Loss: {'train/loss': 0.4573981463909149, 'train/grad_norm_pre_clip': 117.05458068847656, 'timing/data': 0.0003514099953463301, 'timing/model': 0.38918821699917316, 'learning_rate/qwen_vl_interface': 9.980050177479075e-06, 'learning_rate/action_model': 9.980050177479076e-05, 'epoch': 0.02}) +[2026-10-03 13:03:29,128][starVLA.training.train_starvla][INFO] - Step 244, Loss: {'train/loss': 0.3742430806159973, 'train/grad_norm_pre_clip': 42.98100280761719, 'timing/data': 0.00043026999628636986, 'timing/model': 0.5824558360036463, 'learning_rate/qwen_vl_interface': 9.97977038230437e-06, 'learning_rate/action_model': 9.97977038230437e-05, 'epoch': 0.02}) +[2026-10-03 13:03:29,825][starVLA.training.train_starvla][INFO] - Step 245, Loss: {'train/loss': 1.1502714157104492, 'train/grad_norm_pre_clip': 91.75204467773438, 'timing/data': 0.00020357000175863504, 'timing/model': 0.38905301700287964, 'learning_rate/qwen_vl_interface': 9.97948864289972e-06, 'learning_rate/action_model': 9.97948864289972e-05, 'epoch': 0.02}) +[2026-10-03 13:03:30,521][starVLA.training.train_starvla][INFO] - Step 246, Loss: {'train/loss': 0.7939271926879883, 'train/grad_norm_pre_clip': 50.47679901123047, 'timing/data': 0.0003181999927619472, 'timing/model': 0.389534107001964, 'learning_rate/qwen_vl_interface': 9.979204959380936e-06, 'learning_rate/action_model': 9.979204959380936e-05, 'epoch': 0.02}) +[2026-10-03 13:03:31,216][starVLA.training.train_starvla][INFO] - Step 247, Loss: {'train/loss': 0.5054328441619873, 'train/grad_norm_pre_clip': 28.62282943725586, 'timing/data': 0.00020148999465163797, 'timing/model': 0.3890695779991802, 'learning_rate/qwen_vl_interface': 9.978919331864631e-06, 'learning_rate/action_model': 9.97891933186463e-05, 'epoch': 0.02}) +[2026-10-03 13:03:31,913][starVLA.training.train_starvla][INFO] - Step 248, Loss: {'train/loss': 0.4130461812019348, 'train/grad_norm_pre_clip': 41.53019332885742, 'timing/data': 0.0002602699969429523, 'timing/model': 0.3890132579981582, 'learning_rate/qwen_vl_interface': 9.978631760468216e-06, 'learning_rate/action_model': 9.978631760468216e-05, 'epoch': 0.02}) +[2026-10-03 13:03:32,609][starVLA.training.train_starvla][INFO] - Step 249, Loss: {'train/loss': 0.816216766834259, 'train/grad_norm_pre_clip': 51.30451965332031, 'timing/data': 0.0003897300048265606, 'timing/model': 0.3887489269982325, 'learning_rate/qwen_vl_interface': 9.9783422453099e-06, 'learning_rate/action_model': 9.9783422453099e-05, 'epoch': 0.02}) +[2026-10-03 13:03:33,493][starVLA.training.train_starvla][INFO] - Step 250, Loss: {'train/loss': 0.6291978359222412, 'train/grad_norm_pre_clip': 58.47056579589844, 'timing/data': 0.00020248000510036945, 'timing/model': 0.389315587002784, 'learning_rate/qwen_vl_interface': 9.978050786508693e-06, 'learning_rate/action_model': 9.978050786508692e-05, 'epoch': 0.02}) +[2026-10-03 13:03:34,189][starVLA.training.train_starvla][INFO] - Step 251, Loss: {'train/loss': 0.5169476270675659, 'train/grad_norm_pre_clip': 41.879234313964844, 'timing/data': 0.00029928000003565103, 'timing/model': 0.38937159700435586, 'learning_rate/qwen_vl_interface': 9.9777573841844e-06, 'learning_rate/action_model': 9.977757384184401e-05, 'epoch': 0.02}) +[2026-10-03 13:03:34,887][starVLA.training.train_starvla][INFO] - Step 252, Loss: {'train/loss': 0.5072237849235535, 'train/grad_norm_pre_clip': 34.095916748046875, 'timing/data': 0.00020207000488881022, 'timing/model': 0.38967444699665066, 'learning_rate/qwen_vl_interface': 9.97746203845763e-06, 'learning_rate/action_model': 9.97746203845763e-05, 'epoch': 0.02}) +[2026-10-03 13:03:35,584][starVLA.training.train_starvla][INFO] - Step 253, Loss: {'train/loss': 0.7571622133255005, 'train/grad_norm_pre_clip': 112.48686218261719, 'timing/data': 0.0002652100083651021, 'timing/model': 0.38932800698967185, 'learning_rate/qwen_vl_interface': 9.97716474944979e-06, 'learning_rate/action_model': 9.977164749449788e-05, 'epoch': 0.02}) +[2026-10-03 13:03:36,281][starVLA.training.train_starvla][INFO] - Step 254, Loss: {'train/loss': 0.27752894163131714, 'train/grad_norm_pre_clip': 16.279085159301758, 'timing/data': 0.00025828999059740454, 'timing/model': 0.39004330699390266, 'learning_rate/qwen_vl_interface': 9.976865517283079e-06, 'learning_rate/action_model': 9.976865517283078e-05, 'epoch': 0.02}) +[2026-10-03 13:03:37,171][starVLA.training.train_starvla][INFO] - Step 255, Loss: {'train/loss': 0.009899081662297249, 'train/grad_norm_pre_clip': 1.7459843158721924, 'timing/data': 0.0003114200080744922, 'timing/model': 0.389849787010462, 'learning_rate/qwen_vl_interface': 9.976564342080505e-06, 'learning_rate/action_model': 9.976564342080504e-05, 'epoch': 0.02}) +[2026-10-03 13:03:37,870][starVLA.training.train_starvla][INFO] - Step 256, Loss: {'train/loss': 0.9290031790733337, 'train/grad_norm_pre_clip': 53.889892578125, 'timing/data': 0.0002990300126839429, 'timing/model': 0.38947157700022217, 'learning_rate/qwen_vl_interface': 9.976261223965867e-06, 'learning_rate/action_model': 9.976261223965865e-05, 'epoch': 0.02}) +[2026-10-03 13:03:38,566][starVLA.training.train_starvla][INFO] - Step 257, Loss: {'train/loss': 0.630180299282074, 'train/grad_norm_pre_clip': 50.90949249267578, 'timing/data': 0.0003144999936921522, 'timing/model': 0.3892344769992633, 'learning_rate/qwen_vl_interface': 9.975956163063767e-06, 'learning_rate/action_model': 9.975956163063767e-05, 'epoch': 0.02}) +[2026-10-03 13:03:39,263][starVLA.training.train_starvla][INFO] - Step 258, Loss: {'train/loss': 0.7512518763542175, 'train/grad_norm_pre_clip': 54.876556396484375, 'timing/data': 0.0003273600013926625, 'timing/model': 0.3894184769887943, 'learning_rate/qwen_vl_interface': 9.975649159499605e-06, 'learning_rate/action_model': 9.975649159499603e-05, 'epoch': 0.02}) +[2026-10-03 13:03:39,958][starVLA.training.train_starvla][INFO] - Step 259, Loss: {'train/loss': 0.8248144388198853, 'train/grad_norm_pre_clip': 76.46233367919922, 'timing/data': 0.00020324000797700137, 'timing/model': 0.3892714370012982, 'learning_rate/qwen_vl_interface': 9.975340213399573e-06, 'learning_rate/action_model': 9.975340213399573e-05, 'epoch': 0.02}) +[2026-10-03 13:03:40,848][starVLA.training.train_starvla][INFO] - Step 260, Loss: {'train/loss': 0.8301659822463989, 'train/grad_norm_pre_clip': 22.373830795288086, 'timing/data': 0.0003341900010127574, 'timing/model': 0.5800651559984544, 'learning_rate/qwen_vl_interface': 9.975029324890674e-06, 'learning_rate/action_model': 9.975029324890674e-05, 'epoch': 0.02}) +[2026-10-03 13:03:41,545][starVLA.training.train_starvla][INFO] - Step 261, Loss: {'train/loss': 0.2740466892719269, 'train/grad_norm_pre_clip': 21.361454010009766, 'timing/data': 0.00031119000050239265, 'timing/model': 0.38915287698910106, 'learning_rate/qwen_vl_interface': 9.974716494100698e-06, 'learning_rate/action_model': 9.974716494100697e-05, 'epoch': 0.02}) +[2026-10-03 13:03:42,240][starVLA.training.train_starvla][INFO] - Step 262, Loss: {'train/loss': 0.6001954674720764, 'train/grad_norm_pre_clip': 42.66990280151367, 'timing/data': 0.00034876000427175313, 'timing/model': 0.38876083699869923, 'learning_rate/qwen_vl_interface': 9.97440172115824e-06, 'learning_rate/action_model': 9.974401721158239e-05, 'epoch': 0.02}) +[2026-10-03 13:03:42,936][starVLA.training.train_starvla][INFO] - Step 263, Loss: {'train/loss': 0.6411148309707642, 'train/grad_norm_pre_clip': 48.113487243652344, 'timing/data': 0.00019526999676600099, 'timing/model': 0.38968277799722273, 'learning_rate/qwen_vl_interface': 9.97408500619269e-06, 'learning_rate/action_model': 9.97408500619269e-05, 'epoch': 0.02}) +[2026-10-03 13:03:43,635][starVLA.training.train_starvla][INFO] - Step 264, Loss: {'train/loss': 0.704001247882843, 'train/grad_norm_pre_clip': 28.433002471923828, 'timing/data': 0.00019957000040449202, 'timing/model': 0.38941310699738096, 'learning_rate/qwen_vl_interface': 9.973766349334239e-06, 'learning_rate/action_model': 9.973766349334239e-05, 'epoch': 0.02}) +[2026-10-03 13:03:44,522][starVLA.training.train_starvla][INFO] - Step 265, Loss: {'train/loss': 0.7132889032363892, 'train/grad_norm_pre_clip': 32.39619827270508, 'timing/data': 0.00028685000143013895, 'timing/model': 0.5800866059871623, 'learning_rate/qwen_vl_interface': 9.973445750713875e-06, 'learning_rate/action_model': 9.973445750713875e-05, 'epoch': 0.02}) +[2026-10-03 13:03:45,217][starVLA.training.train_starvla][INFO] - Step 266, Loss: {'train/loss': 0.47616639733314514, 'train/grad_norm_pre_clip': 53.2452507019043, 'timing/data': 0.00025392998941242695, 'timing/model': 0.3883354680001503, 'learning_rate/qwen_vl_interface': 9.97312321046338e-06, 'learning_rate/action_model': 9.97312321046338e-05, 'epoch': 0.02}) +[2026-10-03 13:03:45,910][starVLA.training.train_starvla][INFO] - Step 267, Loss: {'train/loss': 0.1865888386964798, 'train/grad_norm_pre_clip': 14.236608505249023, 'timing/data': 0.0002346499968552962, 'timing/model': 0.38797797699226066, 'learning_rate/qwen_vl_interface': 9.972798728715343e-06, 'learning_rate/action_model': 9.972798728715344e-05, 'epoch': 0.02}) +[2026-10-03 13:03:46,607][starVLA.training.train_starvla][INFO] - Step 268, Loss: {'train/loss': 0.19510601460933685, 'train/grad_norm_pre_clip': 44.02802658081055, 'timing/data': 0.0003034800029126927, 'timing/model': 0.38845510699320585, 'learning_rate/qwen_vl_interface': 9.972472305603145e-06, 'learning_rate/action_model': 9.972472305603144e-05, 'epoch': 0.02}) +[2026-10-03 13:03:47,300][starVLA.training.train_starvla][INFO] - Step 269, Loss: {'train/loss': 0.6281734108924866, 'train/grad_norm_pre_clip': 55.10805892944336, 'timing/data': 0.00020015001064166427, 'timing/model': 0.38775836800050456, 'learning_rate/qwen_vl_interface': 9.972143941260965e-06, 'learning_rate/action_model': 9.972143941260965e-05, 'epoch': 0.02}) +[2026-10-03 13:03:48,184][starVLA.training.train_starvla][INFO] - Step 270, Loss: {'train/loss': 0.5485571026802063, 'train/grad_norm_pre_clip': 38.17549133300781, 'timing/data': 0.00020356000459287316, 'timing/model': 0.5779446460073814, 'learning_rate/qwen_vl_interface': 9.971813635823781e-06, 'learning_rate/action_model': 9.971813635823781e-05, 'epoch': 0.02}) +[2026-10-03 13:03:48,877][starVLA.training.train_starvla][INFO] - Step 271, Loss: {'train/loss': 0.253661572933197, 'train/grad_norm_pre_clip': 17.004737854003906, 'timing/data': 0.00032066000858321786, 'timing/model': 0.38794172699272167, 'learning_rate/qwen_vl_interface': 9.971481389427372e-06, 'learning_rate/action_model': 9.971481389427371e-05, 'epoch': 0.02}) +[2026-10-03 13:03:49,574][starVLA.training.train_starvla][INFO] - Step 272, Loss: {'train/loss': 0.29744139313697815, 'train/grad_norm_pre_clip': 17.370534896850586, 'timing/data': 0.0002479800023138523, 'timing/model': 0.3885237380018225, 'learning_rate/qwen_vl_interface': 9.971147202208308e-06, 'learning_rate/action_model': 9.971147202208308e-05, 'epoch': 0.02}) +[2026-10-03 13:03:50,268][starVLA.training.train_starvla][INFO] - Step 273, Loss: {'train/loss': 0.09205614030361176, 'train/grad_norm_pre_clip': 42.975929260253906, 'timing/data': 0.0002391800080658868, 'timing/model': 0.38826000699191354, 'learning_rate/qwen_vl_interface': 9.970811074303962e-06, 'learning_rate/action_model': 9.970811074303962e-05, 'epoch': 0.02}) +[2026-10-03 13:03:50,963][starVLA.training.train_starvla][INFO] - Step 274, Loss: {'train/loss': 1.1101195812225342, 'train/grad_norm_pre_clip': 146.82363891601562, 'timing/data': 0.0001926300028571859, 'timing/model': 0.3887876570079243, 'learning_rate/qwen_vl_interface': 9.970473005852505e-06, 'learning_rate/action_model': 9.970473005852505e-05, 'epoch': 0.02}) +[2026-10-03 13:03:51,657][starVLA.training.train_starvla][INFO] - Step 275, Loss: {'train/loss': 1.2288477420806885, 'train/grad_norm_pre_clip': 82.61697387695312, 'timing/data': 0.0002899599930969998, 'timing/model': 0.3882675369968638, 'learning_rate/qwen_vl_interface': 9.970132996992904e-06, 'learning_rate/action_model': 9.970132996992904e-05, 'epoch': 0.02}) +[2026-10-03 13:03:52,548][starVLA.training.train_starvla][INFO] - Step 276, Loss: {'train/loss': 0.8717162609100342, 'train/grad_norm_pre_clip': 69.35176849365234, 'timing/data': 0.0002075500087812543, 'timing/model': 0.3885464960039826, 'learning_rate/qwen_vl_interface': 9.969791047864921e-06, 'learning_rate/action_model': 9.969791047864921e-05, 'epoch': 0.02}) +[2026-10-03 13:03:53,243][starVLA.training.train_starvla][INFO] - Step 277, Loss: {'train/loss': 1.1168231964111328, 'train/grad_norm_pre_clip': 68.62320709228516, 'timing/data': 0.00020416000916156918, 'timing/model': 0.38836234700283967, 'learning_rate/qwen_vl_interface': 9.969447158609122e-06, 'learning_rate/action_model': 9.96944715860912e-05, 'epoch': 0.02}) +[2026-10-03 13:03:53,938][starVLA.training.train_starvla][INFO] - Step 278, Loss: {'train/loss': 0.397094190120697, 'train/grad_norm_pre_clip': 18.184661865234375, 'timing/data': 0.0002912900090450421, 'timing/model': 0.3888650580047397, 'learning_rate/qwen_vl_interface': 9.969101329366864e-06, 'learning_rate/action_model': 9.969101329366863e-05, 'epoch': 0.02}) +[2026-10-03 13:03:54,632][starVLA.training.train_starvla][INFO] - Step 279, Loss: {'train/loss': 0.9632859826087952, 'train/grad_norm_pre_clip': 109.4477767944336, 'timing/data': 0.0002857200015569106, 'timing/model': 0.388422736999928, 'learning_rate/qwen_vl_interface': 9.968753560280305e-06, 'learning_rate/action_model': 9.968753560280304e-05, 'epoch': 0.02}) +[2026-10-03 13:03:55,328][starVLA.training.train_starvla][INFO] - Step 280, Loss: {'train/loss': 0.3590211868286133, 'train/grad_norm_pre_clip': 44.34245300292969, 'timing/data': 0.000308899994706735, 'timing/model': 0.3880405870004324, 'learning_rate/qwen_vl_interface': 9.968403851492402e-06, 'learning_rate/action_model': 9.968403851492401e-05, 'epoch': 0.02}) +[2026-10-03 13:03:56,214][starVLA.training.train_starvla][INFO] - Step 281, Loss: {'train/loss': 0.30750933289527893, 'train/grad_norm_pre_clip': 29.969263076782227, 'timing/data': 0.00033432000782340765, 'timing/model': 0.5801187460019719, 'learning_rate/qwen_vl_interface': 9.968052203146904e-06, 'learning_rate/action_model': 9.968052203146903e-05, 'epoch': 0.02}) +[2026-10-03 13:03:56,909][starVLA.training.train_starvla][INFO] - Step 282, Loss: {'train/loss': 0.3996012806892395, 'train/grad_norm_pre_clip': 39.29807662963867, 'timing/data': 0.0002472100022714585, 'timing/model': 0.3882141770009184, 'learning_rate/qwen_vl_interface': 9.967698615388363e-06, 'learning_rate/action_model': 9.967698615388362e-05, 'epoch': 0.02}) +[2026-10-03 13:03:57,602][starVLA.training.train_starvla][INFO] - Step 283, Loss: {'train/loss': 0.7073307037353516, 'train/grad_norm_pre_clip': 44.95335006713867, 'timing/data': 0.00020121999841649085, 'timing/model': 0.38860177699825726, 'learning_rate/qwen_vl_interface': 9.967343088362123e-06, 'learning_rate/action_model': 9.967343088362122e-05, 'epoch': 0.02}) +[2026-10-03 13:03:58,298][starVLA.training.train_starvla][INFO] - Step 284, Loss: {'train/loss': 0.27679356932640076, 'train/grad_norm_pre_clip': 26.38968276977539, 'timing/data': 0.0003311199980089441, 'timing/model': 0.38796546700177714, 'learning_rate/qwen_vl_interface': 9.966985622214328e-06, 'learning_rate/action_model': 9.966985622214329e-05, 'epoch': 0.02}) +[2026-10-03 13:03:58,992][starVLA.training.train_starvla][INFO] - Step 285, Loss: {'train/loss': 0.7493072748184204, 'train/grad_norm_pre_clip': 126.52009582519531, 'timing/data': 0.00030901000718586147, 'timing/model': 0.3879453069966985, 'learning_rate/qwen_vl_interface': 9.966626217091922e-06, 'learning_rate/action_model': 9.966626217091922e-05, 'epoch': 0.02}) +[2026-10-03 13:03:59,687][starVLA.training.train_starvla][INFO] - Step 286, Loss: {'train/loss': 1.0718473196029663, 'train/grad_norm_pre_clip': 53.836181640625, 'timing/data': 0.00020061999384779483, 'timing/model': 0.3886904780083569, 'learning_rate/qwen_vl_interface': 9.966264873142641e-06, 'learning_rate/action_model': 9.96626487314264e-05, 'epoch': 0.02}) +[2026-10-03 13:04:00,573][starVLA.training.train_starvla][INFO] - Step 287, Loss: {'train/loss': 1.0897495746612549, 'train/grad_norm_pre_clip': 120.96903991699219, 'timing/data': 0.00035925999691244215, 'timing/model': 0.3889206369931344, 'learning_rate/qwen_vl_interface': 9.965901590515018e-06, 'learning_rate/action_model': 9.965901590515018e-05, 'epoch': 0.02}) +[2026-10-03 13:04:01,271][starVLA.training.train_starvla][INFO] - Step 288, Loss: {'train/loss': 0.5338553190231323, 'train/grad_norm_pre_clip': 92.40043640136719, 'timing/data': 0.0003758100065169856, 'timing/model': 0.388653478003107, 'learning_rate/qwen_vl_interface': 9.965536369358386e-06, 'learning_rate/action_model': 9.965536369358386e-05, 'epoch': 0.02}) +[2026-10-03 13:04:01,965][starVLA.training.train_starvla][INFO] - Step 289, Loss: {'train/loss': 0.5448951125144958, 'train/grad_norm_pre_clip': 28.769113540649414, 'timing/data': 0.0003204700042260811, 'timing/model': 0.3879314669902669, 'learning_rate/qwen_vl_interface': 9.965169209822874e-06, 'learning_rate/action_model': 9.965169209822874e-05, 'epoch': 0.02}) +[2026-10-03 13:04:02,659][starVLA.training.train_starvla][INFO] - Step 290, Loss: {'train/loss': 0.35331881046295166, 'train/grad_norm_pre_clip': 21.04442024230957, 'timing/data': 0.0003428899944992736, 'timing/model': 0.38811113699921407, 'learning_rate/qwen_vl_interface': 9.96480011205941e-06, 'learning_rate/action_model': 9.964800112059409e-05, 'epoch': 0.02}) +[2026-10-03 13:04:03,354][starVLA.training.train_starvla][INFO] - Step 291, Loss: {'train/loss': 0.5508142709732056, 'train/grad_norm_pre_clip': 112.11153411865234, 'timing/data': 0.0002501700073480606, 'timing/model': 0.38890296799945645, 'learning_rate/qwen_vl_interface': 9.964429076219711e-06, 'learning_rate/action_model': 9.964429076219711e-05, 'epoch': 0.02}) +[2026-10-03 13:04:04,246][starVLA.training.train_starvla][INFO] - Step 292, Loss: {'train/loss': 0.37548506259918213, 'train/grad_norm_pre_clip': 72.4249267578125, 'timing/data': 0.00030984000477474183, 'timing/model': 0.38832490699132904, 'learning_rate/qwen_vl_interface': 9.964056102456299e-06, 'learning_rate/action_model': 9.964056102456299e-05, 'epoch': 0.02}) +[2026-10-03 13:04:04,942][starVLA.training.train_starvla][INFO] - Step 293, Loss: {'train/loss': 0.3713981509208679, 'train/grad_norm_pre_clip': 28.616657257080078, 'timing/data': 0.00031504999788012356, 'timing/model': 0.38993334700353444, 'learning_rate/qwen_vl_interface': 9.963681190922489e-06, 'learning_rate/action_model': 9.963681190922488e-05, 'epoch': 0.02}) +[2026-10-03 13:04:05,638][starVLA.training.train_starvla][INFO] - Step 294, Loss: {'train/loss': 0.3188568949699402, 'train/grad_norm_pre_clip': 12.57133960723877, 'timing/data': 0.0003418300038902089, 'timing/model': 0.38985441699333023, 'learning_rate/qwen_vl_interface': 9.963304341772391e-06, 'learning_rate/action_model': 9.963304341772391e-05, 'epoch': 0.02}) +[2026-10-03 13:04:06,336][starVLA.training.train_starvla][INFO] - Step 295, Loss: {'train/loss': 0.4500734508037567, 'train/grad_norm_pre_clip': 21.976869583129883, 'timing/data': 0.00035709000076167285, 'timing/model': 0.39110789699770976, 'learning_rate/qwen_vl_interface': 9.962925555160919e-06, 'learning_rate/action_model': 9.962925555160918e-05, 'epoch': 0.02}) +[2026-10-03 13:04:07,036][starVLA.training.train_starvla][INFO] - Step 296, Loss: {'train/loss': 0.3995783030986786, 'train/grad_norm_pre_clip': 25.554733276367188, 'timing/data': 0.00019963999511674047, 'timing/model': 0.39069558700430207, 'learning_rate/qwen_vl_interface': 9.96254483124377e-06, 'learning_rate/action_model': 9.96254483124377e-05, 'epoch': 0.02}) +[2026-10-03 13:04:07,925][starVLA.training.train_starvla][INFO] - Step 297, Loss: {'train/loss': 0.23509511351585388, 'train/grad_norm_pre_clip': 6.3435282707214355, 'timing/data': 0.0002756000030785799, 'timing/model': 0.5812605859973701, 'learning_rate/qwen_vl_interface': 9.96216217017745e-06, 'learning_rate/action_model': 9.962162170177451e-05, 'epoch': 0.02}) +[2026-10-03 13:04:08,624][starVLA.training.train_starvla][INFO] - Step 298, Loss: {'train/loss': 0.11193720996379852, 'train/grad_norm_pre_clip': 23.43863868713379, 'timing/data': 0.000252609999733977, 'timing/model': 0.3900648170092609, 'learning_rate/qwen_vl_interface': 9.961777572119258e-06, 'learning_rate/action_model': 9.961777572119258e-05, 'epoch': 0.02}) +[2026-10-03 13:04:09,321][starVLA.training.train_starvla][INFO] - Step 299, Loss: {'train/loss': 0.5100449323654175, 'train/grad_norm_pre_clip': 29.136878967285156, 'timing/data': 0.00019703000725712627, 'timing/model': 0.3900091569958022, 'learning_rate/qwen_vl_interface': 9.961391037227286e-06, 'learning_rate/action_model': 9.961391037227284e-05, 'epoch': 0.02}) +[2026-10-03 13:04:10,020][starVLA.training.train_starvla][INFO] - Step 300, Loss: {'train/loss': 0.5494346618652344, 'train/grad_norm_pre_clip': 32.644561767578125, 'timing/data': 0.00033352000173181295, 'timing/model': 0.3894034769909922, 'learning_rate/qwen_vl_interface': 9.961002565660419e-06, 'learning_rate/action_model': 9.961002565660419e-05, 'epoch': 0.02}) +[2026-10-03 13:04:10,719][starVLA.training.train_starvla][INFO] - Step 301, Loss: {'train/loss': 0.5760757923126221, 'train/grad_norm_pre_clip': 94.20355224609375, 'timing/data': 0.00020614999812096357, 'timing/model': 0.3901904469967121, 'learning_rate/qwen_vl_interface': 9.96061215757835e-06, 'learning_rate/action_model': 9.960612157578349e-05, 'epoch': 0.02}) +[2026-10-03 13:04:11,416][starVLA.training.train_starvla][INFO] - Step 302, Loss: {'train/loss': 0.22233566641807556, 'train/grad_norm_pre_clip': 12.561111450195312, 'timing/data': 0.0004350499948486686, 'timing/model': 0.3893296970054507, 'learning_rate/qwen_vl_interface': 9.96021981314156e-06, 'learning_rate/action_model': 9.96021981314156e-05, 'epoch': 0.02}) +[2026-10-03 13:04:12,303][starVLA.training.train_starvla][INFO] - Step 303, Loss: {'train/loss': 0.6302800178527832, 'train/grad_norm_pre_clip': 240.53738403320312, 'timing/data': 0.00020445999689400196, 'timing/model': 0.39038590699783526, 'learning_rate/qwen_vl_interface': 9.959825532511322e-06, 'learning_rate/action_model': 9.959825532511322e-05, 'epoch': 0.02}) +[2026-10-03 13:04:13,003][starVLA.training.train_starvla][INFO] - Step 304, Loss: {'train/loss': 0.15532086789608002, 'train/grad_norm_pre_clip': 7.7934722900390625, 'timing/data': 0.00037134999001864344, 'timing/model': 0.38984875798632856, 'learning_rate/qwen_vl_interface': 9.959429315849715e-06, 'learning_rate/action_model': 9.959429315849715e-05, 'epoch': 0.02}) +[2026-10-03 13:04:13,700][starVLA.training.train_starvla][INFO] - Step 305, Loss: {'train/loss': 0.794026792049408, 'train/grad_norm_pre_clip': 59.56501388549805, 'timing/data': 0.00019955000607296824, 'timing/model': 0.3894007579947356, 'learning_rate/qwen_vl_interface': 9.959031163319606e-06, 'learning_rate/action_model': 9.959031163319607e-05, 'epoch': 0.02}) +[2026-10-03 13:04:14,397][starVLA.training.train_starvla][INFO] - Step 306, Loss: {'train/loss': 0.7553873658180237, 'train/grad_norm_pre_clip': 49.69718933105469, 'timing/data': 0.00037041999166831374, 'timing/model': 0.38943560799816623, 'learning_rate/qwen_vl_interface': 9.958631075084664e-06, 'learning_rate/action_model': 9.958631075084664e-05, 'epoch': 0.02}) +[2026-10-03 13:04:15,094][starVLA.training.train_starvla][INFO] - Step 307, Loss: {'train/loss': 0.4597814679145813, 'train/grad_norm_pre_clip': 97.16187286376953, 'timing/data': 0.00020296999718993902, 'timing/model': 0.3902898170053959, 'learning_rate/qwen_vl_interface': 9.958229051309346e-06, 'learning_rate/action_model': 9.958229051309346e-05, 'epoch': 0.02}) +[2026-10-03 13:04:15,988][starVLA.training.train_starvla][INFO] - Step 308, Loss: {'train/loss': 0.3519844114780426, 'train/grad_norm_pre_clip': 67.46695709228516, 'timing/data': 0.00030577000870835036, 'timing/model': 0.3897328269958962, 'learning_rate/qwen_vl_interface': 9.95782509215891e-06, 'learning_rate/action_model': 9.95782509215891e-05, 'epoch': 0.02}) +[2026-10-03 13:04:16,687][starVLA.training.train_starvla][INFO] - Step 309, Loss: {'train/loss': 0.6653497219085693, 'train/grad_norm_pre_clip': 178.5581512451172, 'timing/data': 0.0003080000024056062, 'timing/model': 0.38992272700124886, 'learning_rate/qwen_vl_interface': 9.95741919779941e-06, 'learning_rate/action_model': 9.95741919779941e-05, 'epoch': 0.02}) +[2026-10-03 13:04:17,384][starVLA.training.train_starvla][INFO] - Step 310, Loss: {'train/loss': 0.32003054022789, 'train/grad_norm_pre_clip': 96.4621353149414, 'timing/data': 0.00028217000362928957, 'timing/model': 0.38963387699914165, 'learning_rate/qwen_vl_interface': 9.957011368397692e-06, 'learning_rate/action_model': 9.95701136839769e-05, 'epoch': 0.02}) +[2026-10-03 13:04:18,083][starVLA.training.train_starvla][INFO] - Step 311, Loss: {'train/loss': 0.43974220752716064, 'train/grad_norm_pre_clip': 97.55731964111328, 'timing/data': 0.00035072999889962375, 'timing/model': 0.39024956700450275, 'learning_rate/qwen_vl_interface': 9.956601604121398e-06, 'learning_rate/action_model': 9.956601604121397e-05, 'epoch': 0.02}) +[2026-10-03 13:04:18,782][starVLA.training.train_starvla][INFO] - Step 312, Loss: {'train/loss': 0.36399045586586, 'train/grad_norm_pre_clip': 60.210086822509766, 'timing/data': 0.00022897000599186867, 'timing/model': 0.39015979699615855, 'learning_rate/qwen_vl_interface': 9.95618990513897e-06, 'learning_rate/action_model': 9.95618990513897e-05, 'epoch': 0.02}) +[2026-10-03 13:04:19,671][starVLA.training.train_starvla][INFO] - Step 313, Loss: {'train/loss': 0.41488713026046753, 'train/grad_norm_pre_clip': 60.84130859375, 'timing/data': 0.0003050189989153296, 'timing/model': 0.580959917002474, 'learning_rate/qwen_vl_interface': 9.95577627161964e-06, 'learning_rate/action_model': 9.95577627161964e-05, 'epoch': 0.02}) +[2026-10-03 13:04:20,368][starVLA.training.train_starvla][INFO] - Step 314, Loss: {'train/loss': 0.4096761643886566, 'train/grad_norm_pre_clip': 65.90653991699219, 'timing/data': 0.00028030999237671494, 'timing/model': 0.3899263269995572, 'learning_rate/qwen_vl_interface': 9.955360703733438e-06, 'learning_rate/action_model': 9.955360703733437e-05, 'epoch': 0.02}) +[2026-10-03 13:04:21,066][starVLA.training.train_starvla][INFO] - Step 315, Loss: {'train/loss': 0.39170610904693604, 'train/grad_norm_pre_clip': 62.79758071899414, 'timing/data': 0.00034389900974929333, 'timing/model': 0.3900692680035718, 'learning_rate/qwen_vl_interface': 9.954943201651187e-06, 'learning_rate/action_model': 9.954943201651187e-05, 'epoch': 0.02}) +[2026-10-03 13:04:21,766][starVLA.training.train_starvla][INFO] - Step 316, Loss: {'train/loss': 0.9273951649665833, 'train/grad_norm_pre_clip': 158.6305389404297, 'timing/data': 0.00024880999990273267, 'timing/model': 0.3901653270004317, 'learning_rate/qwen_vl_interface': 9.954523765544507e-06, 'learning_rate/action_model': 9.954523765544508e-05, 'epoch': 0.02}) +[2026-10-03 13:04:22,464][starVLA.training.train_starvla][INFO] - Step 317, Loss: {'train/loss': 0.3524443507194519, 'train/grad_norm_pre_clip': 117.85710906982422, 'timing/data': 0.00031335999665316194, 'timing/model': 0.38950531599402893, 'learning_rate/qwen_vl_interface': 9.954102395585814e-06, 'learning_rate/action_model': 9.954102395585813e-05, 'epoch': 0.02}) +[2026-10-03 13:04:23,353][starVLA.training.train_starvla][INFO] - Step 318, Loss: {'train/loss': 1.2654719352722168, 'train/grad_norm_pre_clip': 164.40940856933594, 'timing/data': 0.00019864000205416232, 'timing/model': 0.5815862149902387, 'learning_rate/qwen_vl_interface': 9.953679091948313e-06, 'learning_rate/action_model': 9.953679091948314e-05, 'epoch': 0.02}) +[2026-10-03 13:04:24,050][starVLA.training.train_starvla][INFO] - Step 319, Loss: {'train/loss': 0.4571837782859802, 'train/grad_norm_pre_clip': 166.1129913330078, 'timing/data': 0.00020170000789221376, 'timing/model': 0.3906219259952195, 'learning_rate/qwen_vl_interface': 9.953253854806013e-06, 'learning_rate/action_model': 9.953253854806012e-05, 'epoch': 0.02}) +[2026-10-03 13:04:24,750][starVLA.training.train_starvla][INFO] - Step 320, Loss: {'train/loss': 0.17658880352973938, 'train/grad_norm_pre_clip': 38.54481506347656, 'timing/data': 0.00026325999351684004, 'timing/model': 0.3898768770013703, 'learning_rate/qwen_vl_interface': 9.95282668433371e-06, 'learning_rate/action_model': 9.952826684333709e-05, 'epoch': 0.02}) +[2026-10-03 13:04:25,448][starVLA.training.train_starvla][INFO] - Step 321, Loss: {'train/loss': 0.4412427246570587, 'train/grad_norm_pre_clip': 30.003000259399414, 'timing/data': 0.0003185899986419827, 'timing/model': 0.3908514970098622, 'learning_rate/qwen_vl_interface': 9.952397580706998e-06, 'learning_rate/action_model': 9.952397580706998e-05, 'epoch': 0.02}) +[2026-10-03 13:04:26,147][starVLA.training.train_starvla][INFO] - Step 322, Loss: {'train/loss': 0.4992448687553406, 'train/grad_norm_pre_clip': 79.02119445800781, 'timing/data': 0.00020980001136194915, 'timing/model': 0.3911323770007584, 'learning_rate/qwen_vl_interface': 9.951966544102267e-06, 'learning_rate/action_model': 9.951966544102266e-05, 'epoch': 0.02}) +[2026-10-03 13:04:27,035][starVLA.training.train_starvla][INFO] - Step 323, Loss: {'train/loss': 0.37754371762275696, 'train/grad_norm_pre_clip': 62.69466018676758, 'timing/data': 0.0001988300064112991, 'timing/model': 0.5816214859951288, 'learning_rate/qwen_vl_interface': 9.951533574696696e-06, 'learning_rate/action_model': 9.951533574696697e-05, 'epoch': 0.02}) +[2026-10-03 13:04:27,735][starVLA.training.train_starvla][INFO] - Step 324, Loss: {'train/loss': 0.7954932451248169, 'train/grad_norm_pre_clip': 62.13127899169922, 'timing/data': 0.0003310790052637458, 'timing/model': 0.3900691180024296, 'learning_rate/qwen_vl_interface': 9.951098672668266e-06, 'learning_rate/action_model': 9.951098672668265e-05, 'epoch': 0.02}) +[2026-10-03 13:04:28,433][starVLA.training.train_starvla][INFO] - Step 325, Loss: {'train/loss': 0.3331257700920105, 'train/grad_norm_pre_clip': 60.913482666015625, 'timing/data': 0.0003084299969486892, 'timing/model': 0.38931565699749626, 'learning_rate/qwen_vl_interface': 9.95066183819575e-06, 'learning_rate/action_model': 9.950661838195749e-05, 'epoch': 0.02}) +[2026-10-03 13:04:29,130][starVLA.training.train_starvla][INFO] - Step 326, Loss: {'train/loss': 0.5170526504516602, 'train/grad_norm_pre_clip': 41.6641845703125, 'timing/data': 0.000278219988103956, 'timing/model': 0.3899656569992658, 'learning_rate/qwen_vl_interface': 9.95022307145871e-06, 'learning_rate/action_model': 9.950223071458708e-05, 'epoch': 0.02}) +[2026-10-03 13:04:29,827][starVLA.training.train_starvla][INFO] - Step 327, Loss: {'train/loss': 0.17233459651470184, 'train/grad_norm_pre_clip': 20.72939109802246, 'timing/data': 0.00024326999846380204, 'timing/model': 0.3900860870053293, 'learning_rate/qwen_vl_interface': 9.949782372637507e-06, 'learning_rate/action_model': 9.949782372637506e-05, 'epoch': 0.02}) +[2026-10-03 13:04:30,526][starVLA.training.train_starvla][INFO] - Step 328, Loss: {'train/loss': 0.2830682098865509, 'train/grad_norm_pre_clip': 43.740684509277344, 'timing/data': 0.0002663199993548915, 'timing/model': 0.38999914699525107, 'learning_rate/qwen_vl_interface': 9.949339741913297e-06, 'learning_rate/action_model': 9.949339741913297e-05, 'epoch': 0.02}) +[2026-10-03 13:04:31,415][starVLA.training.train_starvla][INFO] - Step 329, Loss: {'train/loss': 0.05307602882385254, 'train/grad_norm_pre_clip': 15.229310035705566, 'timing/data': 0.0002314200100954622, 'timing/model': 0.38987635700323153, 'learning_rate/qwen_vl_interface': 9.948895179468032e-06, 'learning_rate/action_model': 9.948895179468031e-05, 'epoch': 0.02}) +[2026-10-03 13:04:32,112][starVLA.training.train_starvla][INFO] - Step 330, Loss: {'train/loss': 0.40626445412635803, 'train/grad_norm_pre_clip': 147.02561950683594, 'timing/data': 0.000331150004058145, 'timing/model': 0.3895194469951093, 'learning_rate/qwen_vl_interface': 9.948448685484448e-06, 'learning_rate/action_model': 9.948448685484447e-05, 'epoch': 0.02}) +[2026-10-03 13:04:32,810][starVLA.training.train_starvla][INFO] - Step 331, Loss: {'train/loss': 0.34255215525627136, 'train/grad_norm_pre_clip': 10.679673194885254, 'timing/data': 0.00020113000937271863, 'timing/model': 0.39036314700206276, 'learning_rate/qwen_vl_interface': 9.948000260146087e-06, 'learning_rate/action_model': 9.948000260146087e-05, 'epoch': 0.02}) +[2026-10-03 13:04:33,510][starVLA.training.train_starvla][INFO] - Step 332, Loss: {'train/loss': 0.7009655833244324, 'train/grad_norm_pre_clip': 163.8758544921875, 'timing/data': 0.0003240999940317124, 'timing/model': 0.3899800770013826, 'learning_rate/qwen_vl_interface': 9.947549903637277e-06, 'learning_rate/action_model': 9.947549903637278e-05, 'epoch': 0.02}) +[2026-10-03 13:04:34,208][starVLA.training.train_starvla][INFO] - Step 333, Loss: {'train/loss': 0.3058628737926483, 'train/grad_norm_pre_clip': 197.02191162109375, 'timing/data': 0.0002533899969421327, 'timing/model': 0.38969183700100984, 'learning_rate/qwen_vl_interface': 9.947097616143145e-06, 'learning_rate/action_model': 9.947097616143144e-05, 'epoch': 0.02}) +[2026-10-03 13:04:35,098][starVLA.training.train_starvla][INFO] - Step 334, Loss: {'train/loss': 0.16103394329547882, 'train/grad_norm_pre_clip': 71.9176254272461, 'timing/data': 0.00035289999505039304, 'timing/model': 0.5817703960055951, 'learning_rate/qwen_vl_interface': 9.946643397849607e-06, 'learning_rate/action_model': 9.946643397849607e-05, 'epoch': 0.02}) +[2026-10-03 13:04:35,796][starVLA.training.train_starvla][INFO] - Step 335, Loss: {'train/loss': 0.5256850719451904, 'train/grad_norm_pre_clip': 44.106868743896484, 'timing/data': 0.000215859996387735, 'timing/model': 0.39007370799663477, 'learning_rate/qwen_vl_interface': 9.946187248943375e-06, 'learning_rate/action_model': 9.946187248943374e-05, 'epoch': 0.02}) +[2026-10-03 13:04:36,496][starVLA.training.train_starvla][INFO] - Step 336, Loss: {'train/loss': 0.5476107597351074, 'train/grad_norm_pre_clip': 126.64775085449219, 'timing/data': 0.00020290000247769058, 'timing/model': 0.39048113700118847, 'learning_rate/qwen_vl_interface': 9.945729169611957e-06, 'learning_rate/action_model': 9.945729169611958e-05, 'epoch': 0.02}) +[2026-10-03 13:04:37,194][starVLA.training.train_starvla][INFO] - Step 337, Loss: {'train/loss': 0.6396777033805847, 'train/grad_norm_pre_clip': 59.5650749206543, 'timing/data': 0.0002126299950759858, 'timing/model': 0.3897392859944375, 'learning_rate/qwen_vl_interface': 9.94526916004365e-06, 'learning_rate/action_model': 9.94526916004365e-05, 'epoch': 0.02}) +[2026-10-03 13:04:37,891][starVLA.training.train_starvla][INFO] - Step 338, Loss: {'train/loss': 0.3843896985054016, 'train/grad_norm_pre_clip': 34.028324127197266, 'timing/data': 0.00034250000317115337, 'timing/model': 0.38996073699672706, 'learning_rate/qwen_vl_interface': 9.944807220427549e-06, 'learning_rate/action_model': 9.944807220427548e-05, 'epoch': 0.02}) +[2026-10-03 13:04:38,589][starVLA.training.train_starvla][INFO] - Step 339, Loss: {'train/loss': 0.3686351180076599, 'train/grad_norm_pre_clip': 64.6622085571289, 'timing/data': 0.0003376300010131672, 'timing/model': 0.3898035869933665, 'learning_rate/qwen_vl_interface': 9.944343350953534e-06, 'learning_rate/action_model': 9.944343350953533e-05, 'epoch': 0.02}) +[2026-10-03 13:04:39,478][starVLA.training.train_starvla][INFO] - Step 340, Loss: {'train/loss': 0.25494247674942017, 'train/grad_norm_pre_clip': 39.58266830444336, 'timing/data': 0.00023163000878412277, 'timing/model': 0.38990517699858174, 'learning_rate/qwen_vl_interface': 9.94387755181229e-06, 'learning_rate/action_model': 9.94387755181229e-05, 'epoch': 0.02}) +[2026-10-03 13:04:40,175][starVLA.training.train_starvla][INFO] - Step 341, Loss: {'train/loss': 0.1053684651851654, 'train/grad_norm_pre_clip': 8.608999252319336, 'timing/data': 0.00025601001107133925, 'timing/model': 0.3883050870063016, 'learning_rate/qwen_vl_interface': 9.943409823195289e-06, 'learning_rate/action_model': 9.943409823195288e-05, 'epoch': 0.02}) +[2026-10-03 13:04:40,872][starVLA.training.train_starvla][INFO] - Step 342, Loss: {'train/loss': 0.19487930834293365, 'train/grad_norm_pre_clip': 53.056724548339844, 'timing/data': 0.00022573999012820423, 'timing/model': 0.38995612699363846, 'learning_rate/qwen_vl_interface': 9.942940165294794e-06, 'learning_rate/action_model': 9.942940165294793e-05, 'epoch': 0.03}) +[2026-10-03 13:04:41,569][starVLA.training.train_starvla][INFO] - Step 343, Loss: {'train/loss': 0.29089590907096863, 'train/grad_norm_pre_clip': 24.182933807373047, 'timing/data': 0.00020265000057406723, 'timing/model': 0.38983666700369213, 'learning_rate/qwen_vl_interface': 9.942468578303865e-06, 'learning_rate/action_model': 9.942468578303865e-05, 'epoch': 0.03}) +[2026-10-03 13:04:42,268][starVLA.training.train_starvla][INFO] - Step 344, Loss: {'train/loss': 0.6198808550834656, 'train/grad_norm_pre_clip': 45.56611251831055, 'timing/data': 0.00019811998936347663, 'timing/model': 0.3897873170062667, 'learning_rate/qwen_vl_interface': 9.941995062416354e-06, 'learning_rate/action_model': 9.941995062416354e-05, 'epoch': 0.03}) +[2026-10-03 13:04:43,161][starVLA.training.train_starvla][INFO] - Step 345, Loss: {'train/loss': 0.2689805030822754, 'train/grad_norm_pre_clip': 148.91798400878906, 'timing/data': 0.0002547599870013073, 'timing/model': 0.3891508770029759, 'learning_rate/qwen_vl_interface': 9.941519617826906e-06, 'learning_rate/action_model': 9.941519617826906e-05, 'epoch': 0.03}) +[2026-10-03 13:04:43,860][starVLA.training.train_starvla][INFO] - Step 346, Loss: {'train/loss': 0.31467369198799133, 'train/grad_norm_pre_clip': 71.45891571044922, 'timing/data': 0.00024780000967439264, 'timing/model': 0.391862767006387, 'learning_rate/qwen_vl_interface': 9.941042244730954e-06, 'learning_rate/action_model': 9.941042244730955e-05, 'epoch': 0.03}) +[2026-10-03 13:04:44,557][starVLA.training.train_starvla][INFO] - Step 347, Loss: {'train/loss': 0.5878052711486816, 'train/grad_norm_pre_clip': 48.77159118652344, 'timing/data': 0.00028634999762289226, 'timing/model': 0.3899066869926173, 'learning_rate/qwen_vl_interface': 9.940562943324736e-06, 'learning_rate/action_model': 9.940562943324735e-05, 'epoch': 0.03}) +[2026-10-03 13:04:45,256][starVLA.training.train_starvla][INFO] - Step 348, Loss: {'train/loss': 0.18066461384296417, 'train/grad_norm_pre_clip': 19.63542938232422, 'timing/data': 0.0003503399930195883, 'timing/model': 0.38941667799372226, 'learning_rate/qwen_vl_interface': 9.940081713805266e-06, 'learning_rate/action_model': 9.940081713805266e-05, 'epoch': 0.03}) +[2026-10-03 13:04:45,953][starVLA.training.train_starvla][INFO] - Step 349, Loss: {'train/loss': 0.18640673160552979, 'train/grad_norm_pre_clip': 38.12699890136719, 'timing/data': 0.00028950000705663115, 'timing/model': 0.3895943880052073, 'learning_rate/qwen_vl_interface': 9.939598556370366e-06, 'learning_rate/action_model': 9.939598556370365e-05, 'epoch': 0.03}) +[2026-10-03 13:04:46,844][starVLA.training.train_starvla][INFO] - Step 350, Loss: {'train/loss': 0.2665930986404419, 'train/grad_norm_pre_clip': 31.84152603149414, 'timing/data': 0.00036738000926561654, 'timing/model': 0.5826989860070171, 'learning_rate/qwen_vl_interface': 9.93911347121864e-06, 'learning_rate/action_model': 9.939113471218639e-05, 'epoch': 0.03}) +[2026-10-03 13:04:47,541][starVLA.training.train_starvla][INFO] - Step 351, Loss: {'train/loss': 0.18588922917842865, 'train/grad_norm_pre_clip': 16.62693214416504, 'timing/data': 0.0002637399884406477, 'timing/model': 0.3898747970088152, 'learning_rate/qwen_vl_interface': 9.938626458549488e-06, 'learning_rate/action_model': 9.938626458549488e-05, 'epoch': 0.03}) +[2026-10-03 13:04:48,240][starVLA.training.train_starvla][INFO] - Step 352, Loss: {'train/loss': 0.3831331133842468, 'train/grad_norm_pre_clip': 108.07221984863281, 'timing/data': 0.0003072700055781752, 'timing/model': 0.3894672770111356, 'learning_rate/qwen_vl_interface': 9.938137518563106e-06, 'learning_rate/action_model': 9.938137518563106e-05, 'epoch': 0.03}) +[2026-10-03 13:04:48,933][starVLA.training.train_starvla][INFO] - Step 353, Loss: {'train/loss': 0.5631594657897949, 'train/grad_norm_pre_clip': 18.378320693969727, 'timing/data': 0.0003314300120109692, 'timing/model': 0.38741208700230345, 'learning_rate/qwen_vl_interface': 9.937646651460475e-06, 'learning_rate/action_model': 9.937646651460474e-05, 'epoch': 0.03}) +[2026-10-03 13:04:49,626][starVLA.training.train_starvla][INFO] - Step 354, Loss: {'train/loss': 0.28081750869750977, 'train/grad_norm_pre_clip': 16.285634994506836, 'timing/data': 0.00030778000655118376, 'timing/model': 0.3877178469992941, 'learning_rate/qwen_vl_interface': 9.937153857443375e-06, 'learning_rate/action_model': 9.937153857443374e-05, 'epoch': 0.03}) +[2026-10-03 13:04:50,320][starVLA.training.train_starvla][INFO] - Step 355, Loss: {'train/loss': 0.16669535636901855, 'train/grad_norm_pre_clip': 13.913971900939941, 'timing/data': 0.0002415999915683642, 'timing/model': 0.38818349699431565, 'learning_rate/qwen_vl_interface': 9.936659136714372e-06, 'learning_rate/action_model': 9.936659136714371e-05, 'epoch': 0.03}) +[2026-10-03 13:04:51,231][starVLA.training.train_starvla][INFO] - Step 356, Loss: {'train/loss': 0.20692653954029083, 'train/grad_norm_pre_clip': 79.80577850341797, 'timing/data': 0.0002770089922705665, 'timing/model': 0.3936792470049113, 'learning_rate/qwen_vl_interface': 9.936162489476827e-06, 'learning_rate/action_model': 9.936162489476828e-05, 'epoch': 0.03}) +[2026-10-03 13:04:51,937][starVLA.training.train_starvla][INFO] - Step 357, Loss: {'train/loss': 0.6117833852767944, 'train/grad_norm_pre_clip': 49.806663513183594, 'timing/data': 0.0003992300044046715, 'timing/model': 0.3938718670106027, 'learning_rate/qwen_vl_interface': 9.935663915934899e-06, 'learning_rate/action_model': 9.935663915934898e-05, 'epoch': 0.03}) +[2026-10-03 13:04:52,642][starVLA.training.train_starvla][INFO] - Step 358, Loss: {'train/loss': 0.05706878751516342, 'train/grad_norm_pre_clip': 9.424589157104492, 'timing/data': 0.0002055000077234581, 'timing/model': 0.39436096700956114, 'learning_rate/qwen_vl_interface': 9.935163416293525e-06, 'learning_rate/action_model': 9.935163416293525e-05, 'epoch': 0.03}) +[2026-10-03 13:04:53,349][starVLA.training.train_starvla][INFO] - Step 359, Loss: {'train/loss': 0.622308611869812, 'train/grad_norm_pre_clip': 42.98893356323242, 'timing/data': 0.0002277899911860004, 'timing/model': 0.39450313699489925, 'learning_rate/qwen_vl_interface': 9.934660990758445e-06, 'learning_rate/action_model': 9.934660990758445e-05, 'epoch': 0.03}) +[2026-10-03 13:04:54,054][starVLA.training.train_starvla][INFO] - Step 360, Loss: {'train/loss': 0.3099015951156616, 'train/grad_norm_pre_clip': 27.846559524536133, 'timing/data': 0.00021530999219976366, 'timing/model': 0.39342476599267684, 'learning_rate/qwen_vl_interface': 9.934156639536186e-06, 'learning_rate/action_model': 9.934156639536187e-05, 'epoch': 0.03}) +[2026-10-03 13:04:54,954][starVLA.training.train_starvla][INFO] - Step 361, Loss: {'train/loss': 0.004602584056556225, 'train/grad_norm_pre_clip': 2.046074151992798, 'timing/data': 0.00022463999630417675, 'timing/model': 0.39416869600245263, 'learning_rate/qwen_vl_interface': 9.93365036283407e-06, 'learning_rate/action_model': 9.93365036283407e-05, 'epoch': 0.03}) +[2026-10-03 13:04:55,664][starVLA.training.train_starvla][INFO] - Step 362, Loss: {'train/loss': 1.2344977855682373, 'train/grad_norm_pre_clip': 52.28483963012695, 'timing/data': 0.00020758000027853996, 'timing/model': 0.39387788699241355, 'learning_rate/qwen_vl_interface': 9.933142160860208e-06, 'learning_rate/action_model': 9.933142160860208e-05, 'epoch': 0.03}) +[2026-10-03 13:04:56,371][starVLA.training.train_starvla][INFO] - Step 363, Loss: {'train/loss': 0.007309365086257458, 'train/grad_norm_pre_clip': 1.3900072574615479, 'timing/data': 0.00022560000070370734, 'timing/model': 0.3949234869942302, 'learning_rate/qwen_vl_interface': 9.932632033823499e-06, 'learning_rate/action_model': 9.932632033823499e-05, 'epoch': 0.03}) +[2026-10-03 13:04:57,076][starVLA.training.train_starvla][INFO] - Step 364, Loss: {'train/loss': 0.5445713400840759, 'train/grad_norm_pre_clip': 40.545204162597656, 'timing/data': 0.000463610005681403, 'timing/model': 0.3935852760041598, 'learning_rate/qwen_vl_interface': 9.93211998193364e-06, 'learning_rate/action_model': 9.932119981933639e-05, 'epoch': 0.03}) +[2026-10-03 13:04:57,780][starVLA.training.train_starvla][INFO] - Step 365, Loss: {'train/loss': 0.8153870701789856, 'train/grad_norm_pre_clip': 38.1181526184082, 'timing/data': 0.0002065900043817237, 'timing/model': 0.3938998759986134, 'learning_rate/qwen_vl_interface': 9.931606005401117e-06, 'learning_rate/action_model': 9.931606005401116e-05, 'epoch': 0.03}) +[2026-10-03 13:04:58,677][starVLA.training.train_starvla][INFO] - Step 366, Loss: {'train/loss': 0.13541536033153534, 'train/grad_norm_pre_clip': 26.955413818359375, 'timing/data': 0.0002106199972331524, 'timing/model': 0.5867312549962662, 'learning_rate/qwen_vl_interface': 9.931090104437202e-06, 'learning_rate/action_model': 9.931090104437202e-05, 'epoch': 0.03}) +[2026-10-03 13:04:59,384][starVLA.training.train_starvla][INFO] - Step 367, Loss: {'train/loss': 0.49825477600097656, 'train/grad_norm_pre_clip': 90.88475799560547, 'timing/data': 0.000347719993442297, 'timing/model': 0.3940344360016752, 'learning_rate/qwen_vl_interface': 9.930572279253967e-06, 'learning_rate/action_model': 9.930572279253967e-05, 'epoch': 0.03}) +[2026-10-03 13:05:00,089][starVLA.training.train_starvla][INFO] - Step 368, Loss: {'train/loss': 0.09611210227012634, 'train/grad_norm_pre_clip': 11.95350170135498, 'timing/data': 0.00021470000501722097, 'timing/model': 0.39306911699532066, 'learning_rate/qwen_vl_interface': 9.930052530064269e-06, 'learning_rate/action_model': 9.93005253006427e-05, 'epoch': 0.03}) +[2026-10-03 13:05:00,795][starVLA.training.train_starvla][INFO] - Step 369, Loss: {'train/loss': 0.11232462525367737, 'train/grad_norm_pre_clip': 45.662628173828125, 'timing/data': 0.00032110000029206276, 'timing/model': 0.394669265995617, 'learning_rate/qwen_vl_interface': 9.929530857081757e-06, 'learning_rate/action_model': 9.929530857081757e-05, 'epoch': 0.03}) +[2026-10-03 13:05:01,500][starVLA.training.train_starvla][INFO] - Step 370, Loss: {'train/loss': 0.1241808831691742, 'train/grad_norm_pre_clip': 59.19544219970703, 'timing/data': 0.0002680899924598634, 'timing/model': 0.39441053599875886, 'learning_rate/qwen_vl_interface': 9.92900726052087e-06, 'learning_rate/action_model': 9.929007260520869e-05, 'epoch': 0.03}) +[2026-10-03 13:05:02,401][starVLA.training.train_starvla][INFO] - Step 371, Loss: {'train/loss': 0.6624370813369751, 'train/grad_norm_pre_clip': 228.0922088623047, 'timing/data': 0.00045741000212728977, 'timing/model': 0.5880062349897344, 'learning_rate/qwen_vl_interface': 9.92848174059684e-06, 'learning_rate/action_model': 9.92848174059684e-05, 'epoch': 0.03}) +[2026-10-03 13:05:03,105][starVLA.training.train_starvla][INFO] - Step 372, Loss: {'train/loss': 0.5078380107879639, 'train/grad_norm_pre_clip': 39.208961486816406, 'timing/data': 0.0002619300066726282, 'timing/model': 0.39286235699546523, 'learning_rate/qwen_vl_interface': 9.92795429752569e-06, 'learning_rate/action_model': 9.927954297525689e-05, 'epoch': 0.03}) +[2026-10-03 13:05:03,808][starVLA.training.train_starvla][INFO] - Step 373, Loss: {'train/loss': 0.5490267872810364, 'train/grad_norm_pre_clip': 40.936622619628906, 'timing/data': 0.0002580500004114583, 'timing/model': 0.3937058770097792, 'learning_rate/qwen_vl_interface': 9.927424931524228e-06, 'learning_rate/action_model': 9.927424931524228e-05, 'epoch': 0.03}) +[2026-10-03 13:05:04,513][starVLA.training.train_starvla][INFO] - Step 374, Loss: {'train/loss': 0.41571635007858276, 'train/grad_norm_pre_clip': 74.30199432373047, 'timing/data': 0.00020908999431412667, 'timing/model': 0.3941208370088134, 'learning_rate/qwen_vl_interface': 9.926893642810058e-06, 'learning_rate/action_model': 9.926893642810059e-05, 'epoch': 0.03}) +[2026-10-03 13:05:05,218][starVLA.training.train_starvla][INFO] - Step 375, Loss: {'train/loss': 0.22648823261260986, 'train/grad_norm_pre_clip': 29.39369773864746, 'timing/data': 0.0003327799931867048, 'timing/model': 0.39296245700097643, 'learning_rate/qwen_vl_interface': 9.926360431601575e-06, 'learning_rate/action_model': 9.926360431601574e-05, 'epoch': 0.03}) +[2026-10-03 13:05:06,115][starVLA.training.train_starvla][INFO] - Step 376, Loss: {'train/loss': 0.6488723158836365, 'train/grad_norm_pre_clip': 25.172677993774414, 'timing/data': 0.00033922999864444137, 'timing/model': 0.586179905993049, 'learning_rate/qwen_vl_interface': 9.92582529811796e-06, 'learning_rate/action_model': 9.925825298117959e-05, 'epoch': 0.03}) +[2026-10-03 13:05:06,818][starVLA.training.train_starvla][INFO] - Step 377, Loss: {'train/loss': 0.052882444113492966, 'train/grad_norm_pre_clip': 9.353187561035156, 'timing/data': 0.00041926000267267227, 'timing/model': 0.3925832859968068, 'learning_rate/qwen_vl_interface': 9.925288242579184e-06, 'learning_rate/action_model': 9.925288242579184e-05, 'epoch': 0.03}) +[2026-10-03 13:05:07,522][starVLA.training.train_starvla][INFO] - Step 378, Loss: {'train/loss': 0.4738372266292572, 'train/grad_norm_pre_clip': 65.77987670898438, 'timing/data': 0.00020669000514317304, 'timing/model': 0.3937075569992885, 'learning_rate/qwen_vl_interface': 9.924749265206015e-06, 'learning_rate/action_model': 9.924749265206015e-05, 'epoch': 0.03}) +[2026-10-03 13:05:08,228][starVLA.training.train_starvla][INFO] - Step 379, Loss: {'train/loss': 0.0833083838224411, 'train/grad_norm_pre_clip': 10.728711128234863, 'timing/data': 0.00022200999956112355, 'timing/model': 0.3943621170037659, 'learning_rate/qwen_vl_interface': 9.924208366220003e-06, 'learning_rate/action_model': 9.924208366220002e-05, 'epoch': 0.03}) +[2026-10-03 13:05:08,931][starVLA.training.train_starvla][INFO] - Step 380, Loss: {'train/loss': 0.3582541346549988, 'train/grad_norm_pre_clip': 37.0549201965332, 'timing/data': 0.00020698999287560582, 'timing/model': 0.3915483460004907, 'learning_rate/qwen_vl_interface': 9.923665545843492e-06, 'learning_rate/action_model': 9.923665545843491e-05, 'epoch': 0.03}) +[2026-10-03 13:05:09,634][starVLA.training.train_starvla][INFO] - Step 381, Loss: {'train/loss': 0.1606827676296234, 'train/grad_norm_pre_clip': 35.88969802856445, 'timing/data': 0.0002109100023517385, 'timing/model': 0.39303022698732093, 'learning_rate/qwen_vl_interface': 9.923120804299615e-06, 'learning_rate/action_model': 9.923120804299613e-05, 'epoch': 0.03}) +[2026-10-03 13:05:10,535][starVLA.training.train_starvla][INFO] - Step 382, Loss: {'train/loss': 0.04729820042848587, 'train/grad_norm_pre_clip': 39.09552764892578, 'timing/data': 0.00039544000173918903, 'timing/model': 0.3941762570029823, 'learning_rate/qwen_vl_interface': 9.922574141812295e-06, 'learning_rate/action_model': 9.922574141812294e-05, 'epoch': 0.03}) +[2026-10-03 13:05:11,241][starVLA.training.train_starvla][INFO] - Step 383, Loss: {'train/loss': 0.43561679124832153, 'train/grad_norm_pre_clip': 84.5645523071289, 'timing/data': 0.0003532900009304285, 'timing/model': 0.39442761600366794, 'learning_rate/qwen_vl_interface': 9.922025558606242e-06, 'learning_rate/action_model': 9.922025558606242e-05, 'epoch': 0.03}) +[2026-10-03 13:05:11,946][starVLA.training.train_starvla][INFO] - Step 384, Loss: {'train/loss': 0.2669205069541931, 'train/grad_norm_pre_clip': 43.16476058959961, 'timing/data': 0.0002138499985449016, 'timing/model': 0.3933535060059512, 'learning_rate/qwen_vl_interface': 9.92147505490696e-06, 'learning_rate/action_model': 9.92147505490696e-05, 'epoch': 0.03}) +[2026-10-03 13:05:12,649][starVLA.training.train_starvla][INFO] - Step 385, Loss: {'train/loss': 0.04859205335378647, 'train/grad_norm_pre_clip': 15.064098358154297, 'timing/data': 0.0002064600121229887, 'timing/model': 0.3932324469933519, 'learning_rate/qwen_vl_interface': 9.920922630940741e-06, 'learning_rate/action_model': 9.920922630940742e-05, 'epoch': 0.03}) +[2026-10-03 13:05:13,353][starVLA.training.train_starvla][INFO] - Step 386, Loss: {'train/loss': 0.390952467918396, 'train/grad_norm_pre_clip': 86.70011901855469, 'timing/data': 0.00021238000772427768, 'timing/model': 0.39400289600598626, 'learning_rate/qwen_vl_interface': 9.920368286934664e-06, 'learning_rate/action_model': 9.920368286934663e-05, 'epoch': 0.03}) +[2026-10-03 13:05:14,251][starVLA.training.train_starvla][INFO] - Step 387, Loss: {'train/loss': 0.09704011678695679, 'train/grad_norm_pre_clip': 6.815078258514404, 'timing/data': 0.0003358200046932325, 'timing/model': 0.5850810349948006, 'learning_rate/qwen_vl_interface': 9.9198120231166e-06, 'learning_rate/action_model': 9.9198120231166e-05, 'epoch': 0.03}) +[2026-10-03 13:05:14,955][starVLA.training.train_starvla][INFO] - Step 388, Loss: {'train/loss': 0.08675879240036011, 'train/grad_norm_pre_clip': 74.45022583007812, 'timing/data': 0.00020559001131914556, 'timing/model': 0.3930237470049178, 'learning_rate/qwen_vl_interface': 9.919253839715208e-06, 'learning_rate/action_model': 9.919253839715208e-05, 'epoch': 0.03}) +[2026-10-03 13:05:15,658][starVLA.training.train_starvla][INFO] - Step 389, Loss: {'train/loss': 0.13417397439479828, 'train/grad_norm_pre_clip': 50.47015380859375, 'timing/data': 0.0003020899894181639, 'timing/model': 0.39287249599874485, 'learning_rate/qwen_vl_interface': 9.918693736959935e-06, 'learning_rate/action_model': 9.918693736959934e-05, 'epoch': 0.03}) +[2026-10-03 13:05:16,364][starVLA.training.train_starvla][INFO] - Step 390, Loss: {'train/loss': 0.8149962425231934, 'train/grad_norm_pre_clip': 39.9884033203125, 'timing/data': 0.00020942999981343746, 'timing/model': 0.3948543169972254, 'learning_rate/qwen_vl_interface': 9.918131715081019e-06, 'learning_rate/action_model': 9.918131715081018e-05, 'epoch': 0.03}) +[2026-10-03 13:05:17,069][starVLA.training.train_starvla][INFO] - Step 391, Loss: {'train/loss': 0.010745679959654808, 'train/grad_norm_pre_clip': 1.9528027772903442, 'timing/data': 0.00038035000034142286, 'timing/model': 0.39338631700957194, 'learning_rate/qwen_vl_interface': 9.917567774309486e-06, 'learning_rate/action_model': 9.917567774309485e-05, 'epoch': 0.03}) +[2026-10-03 13:05:17,754][starVLA.training.train_starvla][INFO] - Step 392, Loss: {'train/loss': 0.01289825513958931, 'train/grad_norm_pre_clip': 5.479889392852783, 'timing/data': 0.0003042900061700493, 'timing/model': 0.39678847600589506, 'learning_rate/qwen_vl_interface': 9.91700191487715e-06, 'learning_rate/action_model': 9.91700191487715e-05, 'epoch': 0.03}) +[2026-10-03 13:05:18,653][starVLA.training.train_starvla][INFO] - Step 393, Loss: {'train/loss': 0.1297338902950287, 'train/grad_norm_pre_clip': 69.16441345214844, 'timing/data': 0.00037369999336078763, 'timing/model': 0.3941102359967772, 'learning_rate/qwen_vl_interface': 9.916434137016614e-06, 'learning_rate/action_model': 9.916434137016613e-05, 'epoch': 0.03}) +[2026-10-03 13:05:19,358][starVLA.training.train_starvla][INFO] - Step 394, Loss: {'train/loss': 0.12202852964401245, 'train/grad_norm_pre_clip': 53.72609329223633, 'timing/data': 0.00025873999402392656, 'timing/model': 0.39459501700184774, 'learning_rate/qwen_vl_interface': 9.915864440961272e-06, 'learning_rate/action_model': 9.915864440961272e-05, 'epoch': 0.03}) +[2026-10-03 13:05:20,065][starVLA.training.train_starvla][INFO] - Step 395, Loss: {'train/loss': 0.14115387201309204, 'train/grad_norm_pre_clip': 153.47726440429688, 'timing/data': 0.0002765300014289096, 'timing/model': 0.3947748560021864, 'learning_rate/qwen_vl_interface': 9.915292826945303e-06, 'learning_rate/action_model': 9.915292826945303e-05, 'epoch': 0.03}) +[2026-10-03 13:05:20,770][starVLA.training.train_starvla][INFO] - Step 396, Loss: {'train/loss': 0.5054312348365784, 'train/grad_norm_pre_clip': 40.22553634643555, 'timing/data': 0.00040240000816993415, 'timing/model': 0.393034147011349, 'learning_rate/qwen_vl_interface': 9.914719295203676e-06, 'learning_rate/action_model': 9.914719295203676e-05, 'epoch': 0.03}) +[2026-10-03 13:05:21,475][starVLA.training.train_starvla][INFO] - Step 397, Loss: {'train/loss': 0.9178268313407898, 'train/grad_norm_pre_clip': 132.75701904296875, 'timing/data': 0.00021164999634493142, 'timing/model': 0.39440615600324236, 'learning_rate/qwen_vl_interface': 9.91414384597215e-06, 'learning_rate/action_model': 9.91414384597215e-05, 'epoch': 0.03}) +[2026-10-03 13:05:22,373][starVLA.training.train_starvla][INFO] - Step 398, Loss: {'train/loss': 0.4625246226787567, 'train/grad_norm_pre_clip': 31.8630428314209, 'timing/data': 0.00022076000459492207, 'timing/model': 0.39417936699464917, 'learning_rate/qwen_vl_interface': 9.913566479487267e-06, 'learning_rate/action_model': 9.913566479487267e-05, 'epoch': 0.03}) +[2026-10-03 13:05:23,080][starVLA.training.train_starvla][INFO] - Step 399, Loss: {'train/loss': 0.47584399580955505, 'train/grad_norm_pre_clip': 72.7071762084961, 'timing/data': 0.0002735599991865456, 'timing/model': 0.3938362269982463, 'learning_rate/qwen_vl_interface': 9.912987195986366e-06, 'learning_rate/action_model': 9.912987195986365e-05, 'epoch': 0.03}) +[2026-10-03 13:05:23,786][starVLA.training.train_starvla][INFO] - Step 400, Loss: {'train/loss': 0.014129227958619595, 'train/grad_norm_pre_clip': 7.743015289306641, 'timing/data': 0.0002412899921182543, 'timing/model': 0.3938358169980347, 'learning_rate/qwen_vl_interface': 9.912405995707562e-06, 'learning_rate/action_model': 9.912405995707561e-05, 'epoch': 0.03}) +[2026-10-03 13:05:24,490][starVLA.training.train_starvla][INFO] - Step 401, Loss: {'train/loss': 0.7346423864364624, 'train/grad_norm_pre_clip': 169.6487579345703, 'timing/data': 0.00021782999101560563, 'timing/model': 0.3937262360122986, 'learning_rate/qwen_vl_interface': 9.911822878889767e-06, 'learning_rate/action_model': 9.911822878889767e-05, 'epoch': 0.03}) +[2026-10-03 13:05:25,195][starVLA.training.train_starvla][INFO] - Step 402, Loss: {'train/loss': 0.005815952084958553, 'train/grad_norm_pre_clip': 1.4068325757980347, 'timing/data': 0.00033429000177420676, 'timing/model': 0.3944355769926915, 'learning_rate/qwen_vl_interface': 9.911237845772681e-06, 'learning_rate/action_model': 9.911237845772681e-05, 'epoch': 0.03}) +[2026-10-03 13:05:26,094][starVLA.training.train_starvla][INFO] - Step 403, Loss: {'train/loss': 0.32553812861442566, 'train/grad_norm_pre_clip': 28.38947105407715, 'timing/data': 0.00022818999423179775, 'timing/model': 0.5863206849899143, 'learning_rate/qwen_vl_interface': 9.910650896596787e-06, 'learning_rate/action_model': 9.910650896596786e-05, 'epoch': 0.03}) +[2026-10-03 13:05:26,798][starVLA.training.train_starvla][INFO] - Step 404, Loss: {'train/loss': 0.244196355342865, 'train/grad_norm_pre_clip': 53.387088775634766, 'timing/data': 0.0002104599989252165, 'timing/model': 0.39277260699600447, 'learning_rate/qwen_vl_interface': 9.910062031603356e-06, 'learning_rate/action_model': 9.910062031603356e-05, 'epoch': 0.03}) +[2026-10-03 13:05:27,503][starVLA.training.train_starvla][INFO] - Step 405, Loss: {'train/loss': 0.613690972328186, 'train/grad_norm_pre_clip': 81.24044036865234, 'timing/data': 0.000275849990430288, 'timing/model': 0.39445585598878097, 'learning_rate/qwen_vl_interface': 9.909471251034451e-06, 'learning_rate/action_model': 9.90947125103445e-05, 'epoch': 0.03}) +[2026-10-03 13:05:28,208][starVLA.training.train_starvla][INFO] - Step 406, Loss: {'train/loss': 0.011833545751869678, 'train/grad_norm_pre_clip': 10.513652801513672, 'timing/data': 0.0003235700132790953, 'timing/model': 0.3940759969991632, 'learning_rate/qwen_vl_interface': 9.908878555132918e-06, 'learning_rate/action_model': 9.908878555132918e-05, 'epoch': 0.03}) +[2026-10-03 13:05:28,915][starVLA.training.train_starvla][INFO] - Step 407, Loss: {'train/loss': 0.3817627429962158, 'train/grad_norm_pre_clip': 171.485107421875, 'timing/data': 0.0003306400030851364, 'timing/model': 0.3943895670090569, 'learning_rate/qwen_vl_interface': 9.908283944142391e-06, 'learning_rate/action_model': 9.908283944142391e-05, 'epoch': 0.03}) +[2026-10-03 13:05:29,814][starVLA.training.train_starvla][INFO] - Step 408, Loss: {'train/loss': 0.0043899305164813995, 'train/grad_norm_pre_clip': 1.7386815547943115, 'timing/data': 0.0003342099953442812, 'timing/model': 0.5865921049989993, 'learning_rate/qwen_vl_interface': 9.907687418307297e-06, 'learning_rate/action_model': 9.907687418307296e-05, 'epoch': 0.03}) +[2026-10-03 13:05:30,518][starVLA.training.train_starvla][INFO] - Step 409, Loss: {'train/loss': 0.2669447064399719, 'train/grad_norm_pre_clip': 30.917097091674805, 'timing/data': 0.0002075500087812543, 'timing/model': 0.3939000259997556, 'learning_rate/qwen_vl_interface': 9.907088977872837e-06, 'learning_rate/action_model': 9.907088977872837e-05, 'epoch': 0.03}) +[2026-10-03 13:05:31,222][starVLA.training.train_starvla][INFO] - Step 410, Loss: {'train/loss': 0.31299251317977905, 'train/grad_norm_pre_clip': 219.76971435546875, 'timing/data': 0.0002169700019294396, 'timing/model': 0.3932310370000778, 'learning_rate/qwen_vl_interface': 9.906488623085015e-06, 'learning_rate/action_model': 9.906488623085015e-05, 'epoch': 0.03}) +[2026-10-03 13:05:31,929][starVLA.training.train_starvla][INFO] - Step 411, Loss: {'train/loss': 0.0527750626206398, 'train/grad_norm_pre_clip': 52.217376708984375, 'timing/data': 0.00022899000032339245, 'timing/model': 0.3948860559903551, 'learning_rate/qwen_vl_interface': 9.905886354190611e-06, 'learning_rate/action_model': 9.90588635419061e-05, 'epoch': 0.03}) +[2026-10-03 13:05:32,633][starVLA.training.train_starvla][INFO] - Step 412, Loss: {'train/loss': 0.2005539834499359, 'train/grad_norm_pre_clip': 80.08490753173828, 'timing/data': 0.0002078200050164014, 'timing/model': 0.39314151600410696, 'learning_rate/qwen_vl_interface': 9.905282171437192e-06, 'learning_rate/action_model': 9.905282171437192e-05, 'epoch': 0.03}) +[2026-10-03 13:05:33,338][starVLA.training.train_starvla][INFO] - Step 413, Loss: {'train/loss': 0.013320974074304104, 'train/grad_norm_pre_clip': 4.448812961578369, 'timing/data': 0.00020501000108197331, 'timing/model': 0.3945408859872259, 'learning_rate/qwen_vl_interface': 9.904676075073121e-06, 'learning_rate/action_model': 9.904676075073121e-05, 'epoch': 0.03}) +[2026-10-03 13:05:34,239][starVLA.training.train_starvla][INFO] - Step 414, Loss: {'train/loss': 0.12266907095909119, 'train/grad_norm_pre_clip': 89.52172088623047, 'timing/data': 0.0003269700100645423, 'timing/model': 0.39458513699355535, 'learning_rate/qwen_vl_interface': 9.904068065347536e-06, 'learning_rate/action_model': 9.904068065347536e-05, 'epoch': 0.03}) +[2026-10-03 13:05:34,946][starVLA.training.train_starvla][INFO] - Step 415, Loss: {'train/loss': 0.6130566596984863, 'train/grad_norm_pre_clip': 39.14857864379883, 'timing/data': 0.00022904000070411712, 'timing/model': 0.39468666599714197, 'learning_rate/qwen_vl_interface': 9.90345814251037e-06, 'learning_rate/action_model': 9.90345814251037e-05, 'epoch': 0.03}) +[2026-10-03 13:05:35,651][starVLA.training.train_starvla][INFO] - Step 416, Loss: {'train/loss': 0.30302780866622925, 'train/grad_norm_pre_clip': 13.644816398620605, 'timing/data': 0.00033805899147409946, 'timing/model': 0.39351198700023815, 'learning_rate/qwen_vl_interface': 9.902846306812339e-06, 'learning_rate/action_model': 9.902846306812338e-05, 'epoch': 0.03}) +[2026-10-03 13:05:36,433][starVLA.training.train_starvla][INFO] - Step 417, Loss: {'train/loss': 0.07970233261585236, 'train/grad_norm_pre_clip': 81.40532684326172, 'timing/data': 0.00029152000206522644, 'timing/model': 0.3950895460002357, 'learning_rate/qwen_vl_interface': 9.902232558504942e-06, 'learning_rate/action_model': 9.902232558504941e-05, 'epoch': 0.03}) +[2026-10-03 13:05:37,138][starVLA.training.train_starvla][INFO] - Step 418, Loss: {'train/loss': 0.019154643639922142, 'train/grad_norm_pre_clip': 14.457991600036621, 'timing/data': 0.0002648499939823523, 'timing/model': 0.3944892760046059, 'learning_rate/qwen_vl_interface': 9.901616897840471e-06, 'learning_rate/action_model': 9.90161689784047e-05, 'epoch': 0.03}) +[2026-10-03 13:05:38,039][starVLA.training.train_starvla][INFO] - Step 419, Loss: {'train/loss': 0.3105584681034088, 'train/grad_norm_pre_clip': 49.24394226074219, 'timing/data': 0.00027555000269785523, 'timing/model': 0.5881995349918725, 'learning_rate/qwen_vl_interface': 9.900999325072e-06, 'learning_rate/action_model': 9.900999325072001e-05, 'epoch': 0.03}) +[2026-10-03 13:05:38,744][starVLA.training.train_starvla][INFO] - Step 420, Loss: {'train/loss': 0.46640586853027344, 'train/grad_norm_pre_clip': 71.57508087158203, 'timing/data': 0.00020534000941552222, 'timing/model': 0.39367906699771993, 'learning_rate/qwen_vl_interface': 9.900379840453392e-06, 'learning_rate/action_model': 9.900379840453392e-05, 'epoch': 0.03}) +[2026-10-03 13:05:39,449][starVLA.training.train_starvla][INFO] - Step 421, Loss: {'train/loss': 0.6398158669471741, 'train/grad_norm_pre_clip': 115.97634887695312, 'timing/data': 0.00021748000290244818, 'timing/model': 0.39405261700449046, 'learning_rate/qwen_vl_interface': 9.89975844423929e-06, 'learning_rate/action_model': 9.899758444239289e-05, 'epoch': 0.03}) +[2026-10-03 13:05:40,154][starVLA.training.train_starvla][INFO] - Step 422, Loss: {'train/loss': 0.6413442492485046, 'train/grad_norm_pre_clip': 62.02195358276367, 'timing/data': 0.0002608899958431721, 'timing/model': 0.39419643599831033, 'learning_rate/qwen_vl_interface': 9.899135136685131e-06, 'learning_rate/action_model': 9.89913513668513e-05, 'epoch': 0.03}) +[2026-10-03 13:05:40,860][starVLA.training.train_starvla][INFO] - Step 423, Loss: {'train/loss': 0.17135648429393768, 'train/grad_norm_pre_clip': 35.68677520751953, 'timing/data': 0.00028070001280866563, 'timing/model': 0.39526650599145796, 'learning_rate/qwen_vl_interface': 9.898509918047127e-06, 'learning_rate/action_model': 9.898509918047126e-05, 'epoch': 0.03}) +[2026-10-03 13:05:41,757][starVLA.training.train_starvla][INFO] - Step 424, Loss: {'train/loss': 0.15748144686222076, 'train/grad_norm_pre_clip': 101.0964126586914, 'timing/data': 0.00027400000544730574, 'timing/model': 0.5847299849992851, 'learning_rate/qwen_vl_interface': 9.897882788582286e-06, 'learning_rate/action_model': 9.897882788582287e-05, 'epoch': 0.03}) +[2026-10-03 13:05:42,461][starVLA.training.train_starvla][INFO] - Step 425, Loss: {'train/loss': 0.5572509765625, 'train/grad_norm_pre_clip': 15.400650978088379, 'timing/data': 0.00031334999948740005, 'timing/model': 0.39368195600400213, 'learning_rate/qwen_vl_interface': 9.897253748548398e-06, 'learning_rate/action_model': 9.897253748548399e-05, 'epoch': 0.03}) +[2026-10-03 13:05:43,165][starVLA.training.train_starvla][INFO] - Step 426, Loss: {'train/loss': 0.09721314907073975, 'train/grad_norm_pre_clip': 30.737897872924805, 'timing/data': 0.0003602700016926974, 'timing/model': 0.39362131600501016, 'learning_rate/qwen_vl_interface': 9.896622798204035e-06, 'learning_rate/action_model': 9.896622798204035e-05, 'epoch': 0.03}) +[2026-10-03 13:05:43,886][starVLA.training.train_starvla][INFO] - Step 427, Loss: {'train/loss': 0.0440024733543396, 'train/grad_norm_pre_clip': 42.23427200317383, 'timing/data': 0.000360909994924441, 'timing/model': 0.3937287059961818, 'learning_rate/qwen_vl_interface': 9.89598993780856e-06, 'learning_rate/action_model': 9.89598993780856e-05, 'epoch': 0.03}) +[2026-10-03 13:05:44,589][starVLA.training.train_starvla][INFO] - Step 428, Loss: {'train/loss': 0.31057968735694885, 'train/grad_norm_pre_clip': 126.30192565917969, 'timing/data': 0.0003603200020734221, 'timing/model': 0.3923483670077985, 'learning_rate/qwen_vl_interface': 9.895355167622113e-06, 'learning_rate/action_model': 9.895355167622113e-05, 'epoch': 0.03}) +[2026-10-03 13:05:45,292][starVLA.training.train_starvla][INFO] - Step 429, Loss: {'train/loss': 0.011704882606863976, 'train/grad_norm_pre_clip': 2.130995035171509, 'timing/data': 0.0002620699960971251, 'timing/model': 0.39272139700187836, 'learning_rate/qwen_vl_interface': 9.894718487905629e-06, 'learning_rate/action_model': 9.894718487905629e-05, 'epoch': 0.03}) +[2026-10-03 13:05:46,201][starVLA.training.train_starvla][INFO] - Step 430, Loss: {'train/loss': 0.7158466577529907, 'train/grad_norm_pre_clip': 41.44898986816406, 'timing/data': 0.0004422100028023124, 'timing/model': 0.406639555003494, 'learning_rate/qwen_vl_interface': 9.89407989892082e-06, 'learning_rate/action_model': 9.894079898920821e-05, 'epoch': 0.03}) +[2026-10-03 13:05:46,919][starVLA.training.train_starvla][INFO] - Step 431, Loss: {'train/loss': 0.03806207329034805, 'train/grad_norm_pre_clip': 41.250064849853516, 'timing/data': 0.00030641000194009393, 'timing/model': 0.39355250699736644, 'learning_rate/qwen_vl_interface': 9.893439400930187e-06, 'learning_rate/action_model': 9.893439400930187e-05, 'epoch': 0.03}) +[2026-10-03 13:05:47,622][starVLA.training.train_starvla][INFO] - Step 432, Loss: {'train/loss': 0.19443972408771515, 'train/grad_norm_pre_clip': 98.15501403808594, 'timing/data': 0.00026096000510733575, 'timing/model': 0.3922523170040222, 'learning_rate/qwen_vl_interface': 9.892796994197014e-06, 'learning_rate/action_model': 9.892796994197014e-05, 'epoch': 0.03}) +[2026-10-03 13:05:48,325][starVLA.training.train_starvla][INFO] - Step 433, Loss: {'train/loss': 0.031405434012413025, 'train/grad_norm_pre_clip': 35.92038345336914, 'timing/data': 0.00024622000637464225, 'timing/model': 0.3929954969935352, 'learning_rate/qwen_vl_interface': 9.892152678985372e-06, 'learning_rate/action_model': 9.892152678985372e-05, 'epoch': 0.03}) +[2026-10-03 13:05:49,030][starVLA.training.train_starvla][INFO] - Step 434, Loss: {'train/loss': 0.599877119064331, 'train/grad_norm_pre_clip': 167.27781677246094, 'timing/data': 0.0003459900035522878, 'timing/model': 0.39412446701317094, 'learning_rate/qwen_vl_interface': 9.891506455560111e-06, 'learning_rate/action_model': 9.891506455560111e-05, 'epoch': 0.03}) +[2026-10-03 13:05:49,933][starVLA.training.train_starvla][INFO] - Step 435, Loss: {'train/loss': 0.4366973042488098, 'train/grad_norm_pre_clip': 57.867801666259766, 'timing/data': 0.0003585099912015721, 'timing/model': 0.39377589599462226, 'learning_rate/qwen_vl_interface': 9.890858324186872e-06, 'learning_rate/action_model': 9.890858324186872e-05, 'epoch': 0.03}) +[2026-10-03 13:05:50,638][starVLA.training.train_starvla][INFO] - Step 436, Loss: {'train/loss': 0.4547298848628998, 'train/grad_norm_pre_clip': 176.11514282226562, 'timing/data': 0.00025842999457381666, 'timing/model': 0.3926586069865152, 'learning_rate/qwen_vl_interface': 9.890208285132077e-06, 'learning_rate/action_model': 9.890208285132077e-05, 'epoch': 0.03}) +[2026-10-03 13:05:51,341][starVLA.training.train_starvla][INFO] - Step 437, Loss: {'train/loss': 0.2841499447822571, 'train/grad_norm_pre_clip': 53.7216682434082, 'timing/data': 0.0003195489989593625, 'timing/model': 0.3929756970028393, 'learning_rate/qwen_vl_interface': 9.889556338662933e-06, 'learning_rate/action_model': 9.889556338662932e-05, 'epoch': 0.03}) +[2026-10-03 13:05:52,047][starVLA.training.train_starvla][INFO] - Step 438, Loss: {'train/loss': 0.023828424513339996, 'train/grad_norm_pre_clip': 17.165451049804688, 'timing/data': 0.0003712399920914322, 'timing/model': 0.3943831670039799, 'learning_rate/qwen_vl_interface': 9.888902485047427e-06, 'learning_rate/action_model': 9.888902485047427e-05, 'epoch': 0.03}) +[2026-10-03 13:05:52,753][starVLA.training.train_starvla][INFO] - Step 439, Loss: {'train/loss': 0.43371155858039856, 'train/grad_norm_pre_clip': 31.33132553100586, 'timing/data': 0.00033621001057326794, 'timing/model': 0.3942841270036297, 'learning_rate/qwen_vl_interface': 9.888246724554337e-06, 'learning_rate/action_model': 9.888246724554337e-05, 'epoch': 0.03}) +[2026-10-03 13:05:53,648][starVLA.training.train_starvla][INFO] - Step 440, Loss: {'train/loss': 0.6963781118392944, 'train/grad_norm_pre_clip': 109.27781677246094, 'timing/data': 0.00021057001140434295, 'timing/model': 0.5841384759987704, 'learning_rate/qwen_vl_interface': 9.887589057453222e-06, 'learning_rate/action_model': 9.88758905745322e-05, 'epoch': 0.03}) +[2026-10-03 13:05:54,351][starVLA.training.train_starvla][INFO] - Step 441, Loss: {'train/loss': 0.500565767288208, 'train/grad_norm_pre_clip': 83.4852523803711, 'timing/data': 0.00024634999863337725, 'timing/model': 0.39313282701186836, 'learning_rate/qwen_vl_interface': 9.88692948401442e-06, 'learning_rate/action_model': 9.88692948401442e-05, 'epoch': 0.03}) +[2026-10-03 13:05:55,055][starVLA.training.train_starvla][INFO] - Step 442, Loss: {'train/loss': 0.015227829106152058, 'train/grad_norm_pre_clip': 2.7234246730804443, 'timing/data': 0.00023369000700768083, 'timing/model': 0.39357050698890816, 'learning_rate/qwen_vl_interface': 9.88626800450906e-06, 'learning_rate/action_model': 9.88626800450906e-05, 'epoch': 0.03}) +[2026-10-03 13:05:55,761][starVLA.training.train_starvla][INFO] - Step 443, Loss: {'train/loss': 0.3241005837917328, 'train/grad_norm_pre_clip': 33.06398010253906, 'timing/data': 0.00026478999643586576, 'timing/model': 0.3944298270071158, 'learning_rate/qwen_vl_interface': 9.885604619209052e-06, 'learning_rate/action_model': 9.885604619209051e-05, 'epoch': 0.03}) +[2026-10-03 13:05:56,466][starVLA.training.train_starvla][INFO] - Step 444, Loss: {'train/loss': 0.16602009534835815, 'train/grad_norm_pre_clip': 129.18190002441406, 'timing/data': 0.00020984999719075859, 'timing/model': 0.39320179700735025, 'learning_rate/qwen_vl_interface': 9.884939328387085e-06, 'learning_rate/action_model': 9.884939328387084e-05, 'epoch': 0.03}) +[2026-10-03 13:05:57,368][starVLA.training.train_starvla][INFO] - Step 445, Loss: {'train/loss': 0.6097298860549927, 'train/grad_norm_pre_clip': 128.91781616210938, 'timing/data': 0.00020806999236810952, 'timing/model': 0.5912354449974373, 'learning_rate/qwen_vl_interface': 9.884272132316638e-06, 'learning_rate/action_model': 9.884272132316637e-05, 'epoch': 0.03}) +[2026-10-03 13:05:58,073][starVLA.training.train_starvla][INFO] - Step 446, Loss: {'train/loss': 0.14483961462974548, 'train/grad_norm_pre_clip': 202.99586486816406, 'timing/data': 0.00021369000023696572, 'timing/model': 0.39429448699229397, 'learning_rate/qwen_vl_interface': 9.883603031271968e-06, 'learning_rate/action_model': 9.883603031271969e-05, 'epoch': 0.03}) +[2026-10-03 13:05:58,779][starVLA.training.train_starvla][INFO] - Step 447, Loss: {'train/loss': 0.48988744616508484, 'train/grad_norm_pre_clip': 20.550708770751953, 'timing/data': 0.00022591000015381724, 'timing/model': 0.39448481699218974, 'learning_rate/qwen_vl_interface': 9.882932025528122e-06, 'learning_rate/action_model': 9.882932025528121e-05, 'epoch': 0.03}) +[2026-10-03 13:05:59,484][starVLA.training.train_starvla][INFO] - Step 448, Loss: {'train/loss': 0.3829047381877899, 'train/grad_norm_pre_clip': 66.63382720947266, 'timing/data': 0.000259280001046136, 'timing/model': 0.3932512959872838, 'learning_rate/qwen_vl_interface': 9.882259115360918e-06, 'learning_rate/action_model': 9.882259115360917e-05, 'epoch': 0.03}) +[2026-10-03 13:06:00,188][starVLA.training.train_starvla][INFO] - Step 449, Loss: {'train/loss': 0.2651079595088959, 'train/grad_norm_pre_clip': 144.7869873046875, 'timing/data': 0.00021256000036373734, 'timing/model': 0.3929436869948404, 'learning_rate/qwen_vl_interface': 9.881584301046968e-06, 'learning_rate/action_model': 9.881584301046968e-05, 'epoch': 0.03}) +[2026-10-03 13:06:01,101][starVLA.training.train_starvla][INFO] - Step 450, Loss: {'train/loss': 0.2695786654949188, 'train/grad_norm_pre_clip': 55.9854736328125, 'timing/data': 0.00021549001394305378, 'timing/model': 0.6013722339994274, 'learning_rate/qwen_vl_interface': 9.880907582863662e-06, 'learning_rate/action_model': 9.880907582863662e-05, 'epoch': 0.03}) +[2026-10-03 13:06:01,808][starVLA.training.train_starvla][INFO] - Step 451, Loss: {'train/loss': 0.11678348481655121, 'train/grad_norm_pre_clip': 24.755769729614258, 'timing/data': 0.0002768800040939823, 'timing/model': 0.3945575660036411, 'learning_rate/qwen_vl_interface': 9.880228961089175e-06, 'learning_rate/action_model': 9.880228961089175e-05, 'epoch': 0.03}) +[2026-10-03 13:06:02,512][starVLA.training.train_starvla][INFO] - Step 452, Loss: {'train/loss': 0.6087786555290222, 'train/grad_norm_pre_clip': 54.952247619628906, 'timing/data': 0.00035539000236894935, 'timing/model': 0.39258952700765803, 'learning_rate/qwen_vl_interface': 9.87954843600246e-06, 'learning_rate/action_model': 9.879548436002459e-05, 'epoch': 0.03}) +[2026-10-03 13:06:03,217][starVLA.training.train_starvla][INFO] - Step 453, Loss: {'train/loss': 0.8066025376319885, 'train/grad_norm_pre_clip': 54.70520782470703, 'timing/data': 0.0003452499950071797, 'timing/model': 0.3933528159977868, 'learning_rate/qwen_vl_interface': 9.878866007883256e-06, 'learning_rate/action_model': 9.878866007883257e-05, 'epoch': 0.03}) +[2026-10-03 13:06:03,923][starVLA.training.train_starvla][INFO] - Step 454, Loss: {'train/loss': 0.15975232422351837, 'train/grad_norm_pre_clip': 32.82905578613281, 'timing/data': 0.0003206999972462654, 'timing/model': 0.39491353699122556, 'learning_rate/qwen_vl_interface': 9.878181677012084e-06, 'learning_rate/action_model': 9.878181677012085e-05, 'epoch': 0.03}) +[2026-10-03 13:06:04,628][starVLA.training.train_starvla][INFO] - Step 455, Loss: {'train/loss': 0.09646490216255188, 'train/grad_norm_pre_clip': 23.29245948791504, 'timing/data': 0.0002752799919107929, 'timing/model': 0.393028926991974, 'learning_rate/qwen_vl_interface': 9.877495443670247e-06, 'learning_rate/action_model': 9.877495443670246e-05, 'epoch': 0.03}) +[2026-10-03 13:06:05,534][starVLA.training.train_starvla][INFO] - Step 456, Loss: {'train/loss': 0.20268778502941132, 'train/grad_norm_pre_clip': 46.90290069580078, 'timing/data': 0.00042084000597242266, 'timing/model': 0.3939673270069761, 'learning_rate/qwen_vl_interface': 9.876807308139828e-06, 'learning_rate/action_model': 9.876807308139829e-05, 'epoch': 0.03}) +[2026-10-03 13:06:06,238][starVLA.training.train_starvla][INFO] - Step 457, Loss: {'train/loss': 0.024326564744114876, 'train/grad_norm_pre_clip': 5.15966796875, 'timing/data': 0.00030575999699067324, 'timing/model': 0.3932733469991945, 'learning_rate/qwen_vl_interface': 9.876117270703695e-06, 'learning_rate/action_model': 9.876117270703694e-05, 'epoch': 0.03}) +[2026-10-03 13:06:06,944][starVLA.training.train_starvla][INFO] - Step 458, Loss: {'train/loss': 0.10116955637931824, 'train/grad_norm_pre_clip': 38.28451156616211, 'timing/data': 0.00023179998970590532, 'timing/model': 0.3949523459887132, 'learning_rate/qwen_vl_interface': 9.875425331645495e-06, 'learning_rate/action_model': 9.875425331645495e-05, 'epoch': 0.03}) +[2026-10-03 13:06:07,651][starVLA.training.train_starvla][INFO] - Step 459, Loss: {'train/loss': 0.13964715600013733, 'train/grad_norm_pre_clip': 20.975431442260742, 'timing/data': 0.0003568599931895733, 'timing/model': 0.3946039759903215, 'learning_rate/qwen_vl_interface': 9.874731491249658e-06, 'learning_rate/action_model': 9.874731491249657e-05, 'epoch': 0.03}) +[2026-10-03 13:06:08,356][starVLA.training.train_starvla][INFO] - Step 460, Loss: {'train/loss': 0.07933321595191956, 'train/grad_norm_pre_clip': 44.126651763916016, 'timing/data': 0.00022801000159233809, 'timing/model': 0.3933092770021176, 'learning_rate/qwen_vl_interface': 9.874035749801398e-06, 'learning_rate/action_model': 9.874035749801399e-05, 'epoch': 0.03}) +[2026-10-03 13:06:09,261][starVLA.training.train_starvla][INFO] - Step 461, Loss: {'train/loss': 0.5019424557685852, 'train/grad_norm_pre_clip': 17.559371948242188, 'timing/data': 0.00022128999989945441, 'timing/model': 0.592571824992774, 'learning_rate/qwen_vl_interface': 9.873338107586704e-06, 'learning_rate/action_model': 9.873338107586705e-05, 'epoch': 0.03}) +[2026-10-03 13:06:09,967][starVLA.training.train_starvla][INFO] - Step 462, Loss: {'train/loss': 1.0753310918807983, 'train/grad_norm_pre_clip': 134.4999237060547, 'timing/data': 0.00022001999604981393, 'timing/model': 0.39527884699054994, 'learning_rate/qwen_vl_interface': 9.872638564892358e-06, 'learning_rate/action_model': 9.872638564892357e-05, 'epoch': 0.03}) +[2026-10-03 13:06:10,673][starVLA.training.train_starvla][INFO] - Step 463, Loss: {'train/loss': 0.5606558918952942, 'train/grad_norm_pre_clip': 84.20901489257812, 'timing/data': 0.00036345999978948385, 'timing/model': 0.3942305969976587, 'learning_rate/qwen_vl_interface': 9.871937122005905e-06, 'learning_rate/action_model': 9.871937122005905e-05, 'epoch': 0.03}) +[2026-10-03 13:06:11,379][starVLA.training.train_starvla][INFO] - Step 464, Loss: {'train/loss': 0.04739472270011902, 'train/grad_norm_pre_clip': 41.88857650756836, 'timing/data': 0.0003143900103168562, 'timing/model': 0.39331194599799346, 'learning_rate/qwen_vl_interface': 9.87123377921569e-06, 'learning_rate/action_model': 9.871233779215689e-05, 'epoch': 0.03}) +[2026-10-03 13:06:12,083][starVLA.training.train_starvla][INFO] - Step 465, Loss: {'train/loss': 0.5265758633613586, 'train/grad_norm_pre_clip': 65.20764923095703, 'timing/data': 0.00040304000140167773, 'timing/model': 0.3936040669941576, 'learning_rate/qwen_vl_interface': 9.870528536810825e-06, 'learning_rate/action_model': 9.870528536810825e-05, 'epoch': 0.03}) +[2026-10-03 13:06:12,787][starVLA.training.train_starvla][INFO] - Step 466, Loss: {'train/loss': 0.05503789335489273, 'train/grad_norm_pre_clip': 7.565854549407959, 'timing/data': 0.00022040998737793416, 'timing/model': 0.39398357601021416, 'learning_rate/qwen_vl_interface': 9.869821395081214e-06, 'learning_rate/action_model': 9.869821395081214e-05, 'epoch': 0.03}) +[2026-10-03 13:06:13,690][starVLA.training.train_starvla][INFO] - Step 467, Loss: {'train/loss': 0.30307161808013916, 'train/grad_norm_pre_clip': 64.43927764892578, 'timing/data': 0.0002267500094603747, 'timing/model': 0.39454992600076366, 'learning_rate/qwen_vl_interface': 9.869112354317532e-06, 'learning_rate/action_model': 9.869112354317532e-05, 'epoch': 0.03}) +[2026-10-03 13:06:14,396][starVLA.training.train_starvla][INFO] - Step 468, Loss: {'train/loss': 0.5741220712661743, 'train/grad_norm_pre_clip': 76.54029846191406, 'timing/data': 0.00021150000975467265, 'timing/model': 0.39432102598948404, 'learning_rate/qwen_vl_interface': 9.86840141481124e-06, 'learning_rate/action_model': 9.868401414811239e-05, 'epoch': 0.03}) +[2026-10-03 13:06:15,102][starVLA.training.train_starvla][INFO] - Step 469, Loss: {'train/loss': 0.05662798136472702, 'train/grad_norm_pre_clip': 45.626991271972656, 'timing/data': 0.00036576999991666526, 'timing/model': 0.39376777700090315, 'learning_rate/qwen_vl_interface': 9.867688576854578e-06, 'learning_rate/action_model': 9.867688576854578e-05, 'epoch': 0.03}) +[2026-10-03 13:06:15,808][starVLA.training.train_starvla][INFO] - Step 470, Loss: {'train/loss': 0.6785878539085388, 'train/grad_norm_pre_clip': 156.6057586669922, 'timing/data': 0.00022187001013662666, 'timing/model': 0.3943774270010181, 'learning_rate/qwen_vl_interface': 9.86697384074057e-06, 'learning_rate/action_model': 9.866973840740568e-05, 'epoch': 0.03}) +[2026-10-03 13:06:16,515][starVLA.training.train_starvla][INFO] - Step 471, Loss: {'train/loss': 0.6318638324737549, 'train/grad_norm_pre_clip': 81.83875274658203, 'timing/data': 0.00022616999922320247, 'timing/model': 0.3946084460039856, 'learning_rate/qwen_vl_interface': 9.86625720676301e-06, 'learning_rate/action_model': 9.86625720676301e-05, 'epoch': 0.03}) +[2026-10-03 13:06:17,419][starVLA.training.train_starvla][INFO] - Step 472, Loss: {'train/loss': 0.014136350713670254, 'train/grad_norm_pre_clip': 1.270194411277771, 'timing/data': 0.0003083600022364408, 'timing/model': 0.3933746569964569, 'learning_rate/qwen_vl_interface': 9.865538675216488e-06, 'learning_rate/action_model': 9.865538675216486e-05, 'epoch': 0.03}) +[2026-10-03 13:06:18,121][starVLA.training.train_starvla][INFO] - Step 473, Loss: {'train/loss': 0.20781973004341125, 'train/grad_norm_pre_clip': 61.48839569091797, 'timing/data': 0.0003686600102810189, 'timing/model': 0.391551916996832, 'learning_rate/qwen_vl_interface': 9.864818246396358e-06, 'learning_rate/action_model': 9.864818246396358e-05, 'epoch': 0.03}) +[2026-10-03 13:06:18,824][starVLA.training.train_starvla][INFO] - Step 474, Loss: {'train/loss': 0.7289613485336304, 'train/grad_norm_pre_clip': 176.0904998779297, 'timing/data': 0.00021189999824855477, 'timing/model': 0.3931613469903823, 'learning_rate/qwen_vl_interface': 9.864095920598764e-06, 'learning_rate/action_model': 9.864095920598763e-05, 'epoch': 0.03}) +[2026-10-03 13:06:19,531][starVLA.training.train_starvla][INFO] - Step 475, Loss: {'train/loss': 0.007450812496244907, 'train/grad_norm_pre_clip': 1.651737093925476, 'timing/data': 0.00038225000025704503, 'timing/model': 0.3945473070052685, 'learning_rate/qwen_vl_interface': 9.86337169812063e-06, 'learning_rate/action_model': 9.863371698120629e-05, 'epoch': 0.03}) +[2026-10-03 13:06:20,235][starVLA.training.train_starvla][INFO] - Step 476, Loss: {'train/loss': 0.021677304059267044, 'train/grad_norm_pre_clip': 7.300516128540039, 'timing/data': 0.00041861999488901347, 'timing/model': 0.3924916269897949, 'learning_rate/qwen_vl_interface': 9.862645579259648e-06, 'learning_rate/action_model': 9.862645579259648e-05, 'epoch': 0.03}) +[2026-10-03 13:06:21,134][starVLA.training.train_starvla][INFO] - Step 477, Loss: {'train/loss': 0.10813175141811371, 'train/grad_norm_pre_clip': 163.40879821777344, 'timing/data': 0.00031955000304151326, 'timing/model': 0.5883206949947635, 'learning_rate/qwen_vl_interface': 9.861917564314308e-06, 'learning_rate/action_model': 9.861917564314308e-05, 'epoch': 0.03}) +[2026-10-03 13:06:21,841][starVLA.training.train_starvla][INFO] - Step 478, Loss: {'train/loss': 0.27976277470588684, 'train/grad_norm_pre_clip': 54.399330139160156, 'timing/data': 0.00020368999685160816, 'timing/model': 0.3965334870008519, 'learning_rate/qwen_vl_interface': 9.861187653583862e-06, 'learning_rate/action_model': 9.861187653583862e-05, 'epoch': 0.03}) +[2026-10-03 13:06:22,547][starVLA.training.train_starvla][INFO] - Step 479, Loss: {'train/loss': 0.018708089366555214, 'train/grad_norm_pre_clip': 37.41008377075195, 'timing/data': 0.0003478700091363862, 'timing/model': 0.39339307699992787, 'learning_rate/qwen_vl_interface': 9.860455847368352e-06, 'learning_rate/action_model': 9.860455847368351e-05, 'epoch': 0.04}) +[2026-10-03 13:06:23,251][starVLA.training.train_starvla][INFO] - Step 480, Loss: {'train/loss': 0.11166365444660187, 'train/grad_norm_pre_clip': 132.07186889648438, 'timing/data': 0.00030076999973971397, 'timing/model': 0.3932073460018728, 'learning_rate/qwen_vl_interface': 9.859722145968595e-06, 'learning_rate/action_model': 9.859722145968595e-05, 'epoch': 0.04}) +[2026-10-03 13:06:23,959][starVLA.training.train_starvla][INFO] - Step 481, Loss: {'train/loss': 0.0828484445810318, 'train/grad_norm_pre_clip': 130.03640747070312, 'timing/data': 0.00041156000224873424, 'timing/model': 0.3967665859963745, 'learning_rate/qwen_vl_interface': 9.858986549686188e-06, 'learning_rate/action_model': 9.858986549686188e-05, 'epoch': 0.04}) +[2026-10-03 13:06:24,665][starVLA.training.train_starvla][INFO] - Step 482, Loss: {'train/loss': 0.5588463544845581, 'train/grad_norm_pre_clip': 57.91215896606445, 'timing/data': 0.00030432999483309686, 'timing/model': 0.39499090699246153, 'learning_rate/qwen_vl_interface': 9.85824905882351e-06, 'learning_rate/action_model': 9.858249058823509e-05, 'epoch': 0.04}) +[2026-10-03 13:06:25,592][starVLA.training.train_starvla][INFO] - Step 483, Loss: {'train/loss': 0.5010392069816589, 'train/grad_norm_pre_clip': 25.49525260925293, 'timing/data': 0.0003800999984377995, 'timing/model': 0.3951504059950821, 'learning_rate/qwen_vl_interface': 9.857509673683709e-06, 'learning_rate/action_model': 9.857509673683708e-05, 'epoch': 0.04}) +[2026-10-03 13:06:26,300][starVLA.training.train_starvla][INFO] - Step 484, Loss: {'train/loss': 0.06864628195762634, 'train/grad_norm_pre_clip': 40.0887565612793, 'timing/data': 0.0005025400023441762, 'timing/model': 0.39459574699867517, 'learning_rate/qwen_vl_interface': 9.856768394570725e-06, 'learning_rate/action_model': 9.856768394570723e-05, 'epoch': 0.04}) +[2026-10-03 13:06:27,020][starVLA.training.train_starvla][INFO] - Step 485, Loss: {'train/loss': 0.00039177382132038474, 'train/grad_norm_pre_clip': 0.0617140457034111, 'timing/data': 0.0002793199964798987, 'timing/model': 0.39336186699802056, 'learning_rate/qwen_vl_interface': 9.856025221789264e-06, 'learning_rate/action_model': 9.856025221789264e-05, 'epoch': 0.04}) +[2026-10-03 13:06:27,724][starVLA.training.train_starvla][INFO] - Step 486, Loss: {'train/loss': 0.0052537983283400536, 'train/grad_norm_pre_clip': 4.538281440734863, 'timing/data': 0.0003009000065503642, 'timing/model': 0.39346737699816003, 'learning_rate/qwen_vl_interface': 9.85528015564482e-06, 'learning_rate/action_model': 9.85528015564482e-05, 'epoch': 0.04}) +[2026-10-03 13:06:28,428][starVLA.training.train_starvla][INFO] - Step 487, Loss: {'train/loss': 0.4471999406814575, 'train/grad_norm_pre_clip': 45.96299362182617, 'timing/data': 0.0002609100047266111, 'timing/model': 0.3932192870124709, 'learning_rate/qwen_vl_interface': 9.854533196443661e-06, 'learning_rate/action_model': 9.854533196443661e-05, 'epoch': 0.04}) +[2026-10-03 13:06:29,324][starVLA.training.train_starvla][INFO] - Step 488, Loss: {'train/loss': 0.001388351432979107, 'train/grad_norm_pre_clip': 0.2688133418560028, 'timing/data': 0.0003475199919193983, 'timing/model': 0.3926798169995891, 'learning_rate/qwen_vl_interface': 9.853784344492833e-06, 'learning_rate/action_model': 9.853784344492833e-05, 'epoch': 0.04}) +[2026-10-03 13:06:30,044][starVLA.training.train_starvla][INFO] - Step 489, Loss: {'train/loss': 0.017988262698054314, 'train/grad_norm_pre_clip': 56.46906661987305, 'timing/data': 0.0004413590067997575, 'timing/model': 0.3940413670061389, 'learning_rate/qwen_vl_interface': 9.853033600100162e-06, 'learning_rate/action_model': 9.853033600100162e-05, 'epoch': 0.04}) +[2026-10-03 13:06:30,748][starVLA.training.train_starvla][INFO] - Step 490, Loss: {'train/loss': 0.0004729857901111245, 'train/grad_norm_pre_clip': 0.12768475711345673, 'timing/data': 0.00020973999926354736, 'timing/model': 0.3938543969998136, 'learning_rate/qwen_vl_interface': 9.85228096357425e-06, 'learning_rate/action_model': 9.852280963574249e-05, 'epoch': 0.04}) +[2026-10-03 13:06:31,453][starVLA.training.train_starvla][INFO] - Step 491, Loss: {'train/loss': 0.37458640336990356, 'train/grad_norm_pre_clip': 24.941434860229492, 'timing/data': 0.0003567200037650764, 'timing/model': 0.39309643699380103, 'learning_rate/qwen_vl_interface': 9.851526435224475e-06, 'learning_rate/action_model': 9.851526435224474e-05, 'epoch': 0.04}) +[2026-10-03 13:06:32,157][starVLA.training.train_starvla][INFO] - Step 492, Loss: {'train/loss': 0.003673089435324073, 'train/grad_norm_pre_clip': 3.016308069229126, 'timing/data': 0.0003937599976779893, 'timing/model': 0.39298017699911725, 'learning_rate/qwen_vl_interface': 9.850770015360999e-06, 'learning_rate/action_model': 9.850770015360999e-05, 'epoch': 0.04}) +[2026-10-03 13:06:33,049][starVLA.training.train_starvla][INFO] - Step 493, Loss: {'train/loss': 0.8083474636077881, 'train/grad_norm_pre_clip': 112.97796630859375, 'timing/data': 0.000358539997250773, 'timing/model': 0.579466795999906, 'learning_rate/qwen_vl_interface': 9.850011704294755e-06, 'learning_rate/action_model': 9.850011704294755e-05, 'epoch': 0.04}) +[2026-10-03 13:06:33,753][starVLA.training.train_starvla][INFO] - Step 494, Loss: {'train/loss': 0.46631020307540894, 'train/grad_norm_pre_clip': 52.88467788696289, 'timing/data': 0.00025644000561442226, 'timing/model': 0.3931729159958195, 'learning_rate/qwen_vl_interface': 9.849251502337458e-06, 'learning_rate/action_model': 9.849251502337458e-05, 'epoch': 0.04}) +[2026-10-03 13:06:34,456][starVLA.training.train_starvla][INFO] - Step 495, Loss: {'train/loss': 0.393793523311615, 'train/grad_norm_pre_clip': 70.54055786132812, 'timing/data': 0.00023587000032421201, 'timing/model': 0.39281640699482523, 'learning_rate/qwen_vl_interface': 9.848489409801598e-06, 'learning_rate/action_model': 9.848489409801598e-05, 'epoch': 0.04}) +[2026-10-03 13:06:35,161][starVLA.training.train_starvla][INFO] - Step 496, Loss: {'train/loss': 0.01691003143787384, 'train/grad_norm_pre_clip': 20.67452621459961, 'timing/data': 0.00031335999665316194, 'timing/model': 0.39253922599891666, 'learning_rate/qwen_vl_interface': 9.84772542700044e-06, 'learning_rate/action_model': 9.84772542700044e-05, 'epoch': 0.04}) +[2026-10-03 13:06:35,867][starVLA.training.train_starvla][INFO] - Step 497, Loss: {'train/loss': 0.013726182281970978, 'train/grad_norm_pre_clip': 108.82518768310547, 'timing/data': 0.0003365399898029864, 'timing/model': 0.3940939270105446, 'learning_rate/qwen_vl_interface': 9.846959554248032e-06, 'learning_rate/action_model': 9.84695955424803e-05, 'epoch': 0.04}) +[2026-10-03 13:06:36,761][starVLA.training.train_starvla][INFO] - Step 498, Loss: {'train/loss': 1.556279182434082, 'train/grad_norm_pre_clip': 86.58345794677734, 'timing/data': 0.00021491000370588154, 'timing/model': 0.5815499259915669, 'learning_rate/qwen_vl_interface': 9.846191791859192e-06, 'learning_rate/action_model': 9.846191791859192e-05, 'epoch': 0.04}) +[2026-10-03 13:06:37,464][starVLA.training.train_starvla][INFO] - Step 499, Loss: {'train/loss': 0.7822428345680237, 'train/grad_norm_pre_clip': 36.6915283203125, 'timing/data': 0.0003493599942885339, 'timing/model': 0.392735786997946, 'learning_rate/qwen_vl_interface': 9.845422140149518e-06, 'learning_rate/action_model': 9.845422140149519e-05, 'epoch': 0.04}) +[2026-10-03 13:06:38,168][starVLA.training.train_starvla][INFO] - Step 500, Loss: {'train/loss': 0.013257667422294617, 'train/grad_norm_pre_clip': 2.818913221359253, 'timing/data': 0.0003193000011378899, 'timing/model': 0.3925303870055359, 'learning_rate/qwen_vl_interface': 9.84465059943539e-06, 'learning_rate/action_model': 9.84465059943539e-05, 'epoch': 0.04}) +[2026-10-03 13:06:38,875][starVLA.training.train_starvla][INFO] - Step 501, Loss: {'train/loss': 0.8696218729019165, 'train/grad_norm_pre_clip': 100.79869842529297, 'timing/data': 0.0003694600018206984, 'timing/model': 0.39397072698920965, 'learning_rate/qwen_vl_interface': 9.843877170033956e-06, 'learning_rate/action_model': 9.843877170033954e-05, 'epoch': 0.04}) +[2026-10-03 13:06:39,580][starVLA.training.train_starvla][INFO] - Step 502, Loss: {'train/loss': 0.5695862770080566, 'train/grad_norm_pre_clip': 173.9149627685547, 'timing/data': 0.00038660000427626073, 'timing/model': 0.39297328700195067, 'learning_rate/qwen_vl_interface': 9.84310185226314e-06, 'learning_rate/action_model': 9.84310185226314e-05, 'epoch': 0.04}) +[2026-10-03 13:06:40,472][starVLA.training.train_starvla][INFO] - Step 503, Loss: {'train/loss': 0.7232733964920044, 'train/grad_norm_pre_clip': 239.79774475097656, 'timing/data': 0.00021564000053331256, 'timing/model': 0.5806075159925967, 'learning_rate/qwen_vl_interface': 9.842324646441652e-06, 'learning_rate/action_model': 9.842324646441652e-05, 'epoch': 0.04}) +[2026-10-03 13:06:41,176][starVLA.training.train_starvla][INFO] - Step 504, Loss: {'train/loss': 0.6856140494346619, 'train/grad_norm_pre_clip': 88.17383575439453, 'timing/data': 0.00021179999748710543, 'timing/model': 0.3927376569918124, 'learning_rate/qwen_vl_interface': 9.841545552888968e-06, 'learning_rate/action_model': 9.841545552888967e-05, 'epoch': 0.04}) +[2026-10-03 13:06:41,884][starVLA.training.train_starvla][INFO] - Step 505, Loss: {'train/loss': 0.20423832535743713, 'train/grad_norm_pre_clip': 122.6716079711914, 'timing/data': 0.000232299993513152, 'timing/model': 0.39435315600712784, 'learning_rate/qwen_vl_interface': 9.840764571925346e-06, 'learning_rate/action_model': 9.840764571925345e-05, 'epoch': 0.04}) +[2026-10-03 13:06:42,590][starVLA.training.train_starvla][INFO] - Step 506, Loss: {'train/loss': 0.12663596868515015, 'train/grad_norm_pre_clip': 34.89771270751953, 'timing/data': 0.0003519099991535768, 'timing/model': 0.39401376601017546, 'learning_rate/qwen_vl_interface': 9.839981703871815e-06, 'learning_rate/action_model': 9.839981703871814e-05, 'epoch': 0.04}) +[2026-10-03 13:06:43,295][starVLA.training.train_starvla][INFO] - Step 507, Loss: {'train/loss': 1.7012434005737305, 'train/grad_norm_pre_clip': 288.4346618652344, 'timing/data': 0.00033722000080160797, 'timing/model': 0.3935479460051283, 'learning_rate/qwen_vl_interface': 9.839196949050185e-06, 'learning_rate/action_model': 9.839196949050185e-05, 'epoch': 0.04}) +[2026-10-03 13:06:43,999][starVLA.training.train_starvla][INFO] - Step 508, Loss: {'train/loss': 0.4304264783859253, 'train/grad_norm_pre_clip': 160.66795349121094, 'timing/data': 0.00020634999964386225, 'timing/model': 0.39337483599956613, 'learning_rate/qwen_vl_interface': 9.838410307783038e-06, 'learning_rate/action_model': 9.838410307783038e-05, 'epoch': 0.04}) +[2026-10-03 13:06:44,901][starVLA.training.train_starvla][INFO] - Step 509, Loss: {'train/loss': 0.21502648293972015, 'train/grad_norm_pre_clip': 25.297048568725586, 'timing/data': 0.0002796199987642467, 'timing/model': 0.3955114160053199, 'learning_rate/qwen_vl_interface': 9.837621780393734e-06, 'learning_rate/action_model': 9.837621780393733e-05, 'epoch': 0.04}) +[2026-10-03 13:06:45,606][starVLA.training.train_starvla][INFO] - Step 510, Loss: {'train/loss': 0.8855682611465454, 'train/grad_norm_pre_clip': 96.62155151367188, 'timing/data': 0.00020948999735992402, 'timing/model': 0.39398203599557746, 'learning_rate/qwen_vl_interface': 9.836831367206406e-06, 'learning_rate/action_model': 9.836831367206406e-05, 'epoch': 0.04}) +[2026-10-03 13:06:46,310][starVLA.training.train_starvla][INFO] - Step 511, Loss: {'train/loss': 0.3063933253288269, 'train/grad_norm_pre_clip': 54.710723876953125, 'timing/data': 0.000321080005960539, 'timing/model': 0.39355883600364905, 'learning_rate/qwen_vl_interface': 9.836039068545962e-06, 'learning_rate/action_model': 9.836039068545961e-05, 'epoch': 0.04}) +[2026-10-03 13:06:47,014][starVLA.training.train_starvla][INFO] - Step 512, Loss: {'train/loss': 0.5438172817230225, 'train/grad_norm_pre_clip': 95.13263702392578, 'timing/data': 0.00020556000526994467, 'timing/model': 0.3927182570041623, 'learning_rate/qwen_vl_interface': 9.835244884738086e-06, 'learning_rate/action_model': 9.835244884738085e-05, 'epoch': 0.04}) +[2026-10-03 13:06:47,721][starVLA.training.train_starvla][INFO] - Step 513, Loss: {'train/loss': 0.3362962007522583, 'train/grad_norm_pre_clip': 30.283390045166016, 'timing/data': 0.0003469299990683794, 'timing/model': 0.3939795970072737, 'learning_rate/qwen_vl_interface': 9.83444881610924e-06, 'learning_rate/action_model': 9.834448816109238e-05, 'epoch': 0.04}) +[2026-10-03 13:06:48,613][starVLA.training.train_starvla][INFO] - Step 514, Loss: {'train/loss': 0.09157972037792206, 'train/grad_norm_pre_clip': 23.175861358642578, 'timing/data': 0.00036163000913802534, 'timing/model': 0.5808249159890693, 'learning_rate/qwen_vl_interface': 9.833650862986652e-06, 'learning_rate/action_model': 9.833650862986653e-05, 'epoch': 0.04}) +[2026-10-03 13:06:49,316][starVLA.training.train_starvla][INFO] - Step 515, Loss: {'train/loss': 0.20094071328639984, 'train/grad_norm_pre_clip': 29.72591209411621, 'timing/data': 0.00020568000036291778, 'timing/model': 0.39297627699852455, 'learning_rate/qwen_vl_interface': 9.832851025698335e-06, 'learning_rate/action_model': 9.832851025698335e-05, 'epoch': 0.04}) +[2026-10-03 13:06:50,020][starVLA.training.train_starvla][INFO] - Step 516, Loss: {'train/loss': 0.12870925664901733, 'train/grad_norm_pre_clip': 16.764842987060547, 'timing/data': 0.0003244399995310232, 'timing/model': 0.39226460599456914, 'learning_rate/qwen_vl_interface': 9.832049304573072e-06, 'learning_rate/action_model': 9.832049304573072e-05, 'epoch': 0.04}) +[2026-10-03 13:06:50,727][starVLA.training.train_starvla][INFO] - Step 517, Loss: {'train/loss': 0.144764244556427, 'train/grad_norm_pre_clip': 38.548057556152344, 'timing/data': 0.00026590999914333224, 'timing/model': 0.39428051699360367, 'learning_rate/qwen_vl_interface': 9.831245699940419e-06, 'learning_rate/action_model': 9.831245699940418e-05, 'epoch': 0.04}) +[2026-10-03 13:06:51,431][starVLA.training.train_starvla][INFO] - Step 518, Loss: {'train/loss': 0.30163058638572693, 'train/grad_norm_pre_clip': 29.036792755126953, 'timing/data': 0.0002708500105654821, 'timing/model': 0.39399371699255425, 'learning_rate/qwen_vl_interface': 9.830440212130707e-06, 'learning_rate/action_model': 9.830440212130708e-05, 'epoch': 0.04}) +[2026-10-03 13:06:52,135][starVLA.training.train_starvla][INFO] - Step 519, Loss: {'train/loss': 0.19084954261779785, 'train/grad_norm_pre_clip': 30.656373977661133, 'timing/data': 0.0002394299954175949, 'timing/model': 0.39303155700326897, 'learning_rate/qwen_vl_interface': 9.829632841475041e-06, 'learning_rate/action_model': 9.829632841475041e-05, 'epoch': 0.04}) +[2026-10-03 13:06:53,025][starVLA.training.train_starvla][INFO] - Step 520, Loss: {'train/loss': 0.5632092952728271, 'train/grad_norm_pre_clip': 89.8542251586914, 'timing/data': 0.000351290000253357, 'timing/model': 0.3927384160051588, 'learning_rate/qwen_vl_interface': 9.828823588305305e-06, 'learning_rate/action_model': 9.828823588305303e-05, 'epoch': 0.04}) +[2026-10-03 13:06:53,732][starVLA.training.train_starvla][INFO] - Step 521, Loss: {'train/loss': 0.43892768025398254, 'train/grad_norm_pre_clip': 180.10617065429688, 'timing/data': 0.00022131000878289342, 'timing/model': 0.3944488970009843, 'learning_rate/qwen_vl_interface': 9.828012452954146e-06, 'learning_rate/action_model': 9.828012452954145e-05, 'epoch': 0.04}) +[2026-10-03 13:06:54,437][starVLA.training.train_starvla][INFO] - Step 522, Loss: {'train/loss': 0.0015477088745683432, 'train/grad_norm_pre_clip': 0.4201977849006653, 'timing/data': 0.00028347001352813095, 'timing/model': 0.3935034669993911, 'learning_rate/qwen_vl_interface': 9.827199435754995e-06, 'learning_rate/action_model': 9.827199435754995e-05, 'epoch': 0.04}) +[2026-10-03 13:06:55,141][starVLA.training.train_starvla][INFO] - Step 523, Loss: {'train/loss': 0.8295159935951233, 'train/grad_norm_pre_clip': 144.56668090820312, 'timing/data': 0.0003154700098093599, 'timing/model': 0.39250726699538063, 'learning_rate/qwen_vl_interface': 9.826384537042052e-06, 'learning_rate/action_model': 9.826384537042052e-05, 'epoch': 0.04}) +[2026-10-03 13:06:55,847][starVLA.training.train_starvla][INFO] - Step 524, Loss: {'train/loss': 0.09025020152330399, 'train/grad_norm_pre_clip': 36.1204948425293, 'timing/data': 0.00028747999749612063, 'timing/model': 0.3931583570083603, 'learning_rate/qwen_vl_interface': 9.825567757150291e-06, 'learning_rate/action_model': 9.82556775715029e-05, 'epoch': 0.04}) +[2026-10-03 13:06:56,744][starVLA.training.train_starvla][INFO] - Step 525, Loss: {'train/loss': 0.10766616463661194, 'train/grad_norm_pre_clip': 63.67181396484375, 'timing/data': 0.0003254100010963157, 'timing/model': 0.3939198070002021, 'learning_rate/qwen_vl_interface': 9.82474909641546e-06, 'learning_rate/action_model': 9.82474909641546e-05, 'epoch': 0.04}) +[2026-10-03 13:06:57,447][starVLA.training.train_starvla][INFO] - Step 526, Loss: {'train/loss': 0.5341516733169556, 'train/grad_norm_pre_clip': 154.55807495117188, 'timing/data': 0.0003676799969980493, 'timing/model': 0.39282847600406967, 'learning_rate/qwen_vl_interface': 9.823928555174078e-06, 'learning_rate/action_model': 9.823928555174077e-05, 'epoch': 0.04}) +[2026-10-03 13:06:58,150][starVLA.training.train_starvla][INFO] - Step 527, Loss: {'train/loss': 0.027602680027484894, 'train/grad_norm_pre_clip': 17.934429168701172, 'timing/data': 0.00032546000147704035, 'timing/model': 0.39244090700231027, 'learning_rate/qwen_vl_interface': 9.82310613376344e-06, 'learning_rate/action_model': 9.82310613376344e-05, 'epoch': 0.04}) +[2026-10-03 13:06:58,852][starVLA.training.train_starvla][INFO] - Step 528, Loss: {'train/loss': 1.3120007514953613, 'train/grad_norm_pre_clip': 49.442142486572266, 'timing/data': 0.00029584999720100313, 'timing/model': 0.3918017070100177, 'learning_rate/qwen_vl_interface': 9.822281832521613e-06, 'learning_rate/action_model': 9.822281832521613e-05, 'epoch': 0.04}) +[2026-10-03 13:06:59,557][starVLA.training.train_starvla][INFO] - Step 529, Loss: {'train/loss': 0.01089910976588726, 'train/grad_norm_pre_clip': 1.868462324142456, 'timing/data': 0.0002920400002039969, 'timing/model': 0.39317882699833717, 'learning_rate/qwen_vl_interface': 9.821455651787432e-06, 'learning_rate/action_model': 9.821455651787432e-05, 'epoch': 0.04}) +[2026-10-03 13:07:00,451][starVLA.training.train_starvla][INFO] - Step 530, Loss: {'train/loss': 0.1107674166560173, 'train/grad_norm_pre_clip': 48.617286682128906, 'timing/data': 0.00031215000490192324, 'timing/model': 0.5836892060033279, 'learning_rate/qwen_vl_interface': 9.820627591900514e-06, 'learning_rate/action_model': 9.820627591900513e-05, 'epoch': 0.04}) +[2026-10-03 13:07:01,154][starVLA.training.train_starvla][INFO] - Step 531, Loss: {'train/loss': 0.4252040982246399, 'train/grad_norm_pre_clip': 49.283809661865234, 'timing/data': 0.00029615999665111303, 'timing/model': 0.39267732700682245, 'learning_rate/qwen_vl_interface': 9.819797653201238e-06, 'learning_rate/action_model': 9.819797653201238e-05, 'epoch': 0.04}) +[2026-10-03 13:07:01,857][starVLA.training.train_starvla][INFO] - Step 532, Loss: {'train/loss': 0.10386376082897186, 'train/grad_norm_pre_clip': 29.824188232421875, 'timing/data': 0.0003294600028311834, 'timing/model': 0.39175789699947927, 'learning_rate/qwen_vl_interface': 9.818965836030765e-06, 'learning_rate/action_model': 9.818965836030764e-05, 'epoch': 0.04}) +[2026-10-03 13:07:02,562][starVLA.training.train_starvla][INFO] - Step 533, Loss: {'train/loss': 0.17892301082611084, 'train/grad_norm_pre_clip': 33.50458526611328, 'timing/data': 0.0002866199938580394, 'timing/model': 0.39335506700444967, 'learning_rate/qwen_vl_interface': 9.81813214073102e-06, 'learning_rate/action_model': 9.81813214073102e-05, 'epoch': 0.04}) +[2026-10-03 13:07:03,264][starVLA.training.train_starvla][INFO] - Step 534, Loss: {'train/loss': 0.026733815670013428, 'train/grad_norm_pre_clip': 13.595301628112793, 'timing/data': 0.0002519200061215088, 'timing/model': 0.3924781069945311, 'learning_rate/qwen_vl_interface': 9.817296567644709e-06, 'learning_rate/action_model': 9.817296567644709e-05, 'epoch': 0.04}) +[2026-10-03 13:07:03,944][starVLA.training.train_starvla][INFO] - Step 535, Loss: {'train/loss': 0.039229441434144974, 'train/grad_norm_pre_clip': 11.9014892578125, 'timing/data': 0.0003825699968729168, 'timing/model': 0.3682891599892173, 'learning_rate/qwen_vl_interface': 9.816459117115296e-06, 'learning_rate/action_model': 9.816459117115297e-05, 'epoch': 0.04}) +[2026-10-03 13:07:04,835][starVLA.training.train_starvla][INFO] - Step 536, Loss: {'train/loss': 0.017276868224143982, 'train/grad_norm_pre_clip': 4.568387985229492, 'timing/data': 0.00021513999672606587, 'timing/model': 0.3923448069981532, 'learning_rate/qwen_vl_interface': 9.815619789487035e-06, 'learning_rate/action_model': 9.815619789487034e-05, 'epoch': 0.04}) +[2026-10-03 13:07:05,540][starVLA.training.train_starvla][INFO] - Step 537, Loss: {'train/loss': 0.020625758916139603, 'train/grad_norm_pre_clip': 18.275527954101562, 'timing/data': 0.0003426000039326027, 'timing/model': 0.3929454769968288, 'learning_rate/qwen_vl_interface': 9.814778585104933e-06, 'learning_rate/action_model': 9.814778585104933e-05, 'epoch': 0.04}) +[2026-10-03 13:07:06,243][starVLA.training.train_starvla][INFO] - Step 538, Loss: {'train/loss': 0.006313296500593424, 'train/grad_norm_pre_clip': 3.155371904373169, 'timing/data': 0.0003800999984377995, 'timing/model': 0.3931513869902119, 'learning_rate/qwen_vl_interface': 9.813935504314784e-06, 'learning_rate/action_model': 9.813935504314784e-05, 'epoch': 0.04}) +[2026-10-03 13:07:06,945][starVLA.training.train_starvla][INFO] - Step 539, Loss: {'train/loss': 0.018379466608166695, 'train/grad_norm_pre_clip': 3.8533730506896973, 'timing/data': 0.00022722900030203164, 'timing/model': 0.3922581169899786, 'learning_rate/qwen_vl_interface': 9.813090547463144e-06, 'learning_rate/action_model': 9.813090547463143e-05, 'epoch': 0.04}) +[2026-10-03 13:07:07,649][starVLA.training.train_starvla][INFO] - Step 540, Loss: {'train/loss': 0.7426248788833618, 'train/grad_norm_pre_clip': 104.16535186767578, 'timing/data': 0.0003253799950471148, 'timing/model': 0.39236533700022846, 'learning_rate/qwen_vl_interface': 9.812243714897342e-06, 'learning_rate/action_model': 9.812243714897343e-05, 'epoch': 0.04}) +[2026-10-03 13:07:08,545][starVLA.training.train_starvla][INFO] - Step 541, Loss: {'train/loss': 0.22788889706134796, 'train/grad_norm_pre_clip': 153.21853637695312, 'timing/data': 0.00030285000684671104, 'timing/model': 0.3938910169963492, 'learning_rate/qwen_vl_interface': 9.811395006965481e-06, 'learning_rate/action_model': 9.81139500696548e-05, 'epoch': 0.04}) +[2026-10-03 13:07:09,249][starVLA.training.train_starvla][INFO] - Step 542, Loss: {'train/loss': 0.4415631592273712, 'train/grad_norm_pre_clip': 100.59896850585938, 'timing/data': 0.0003768199967453256, 'timing/model': 0.39301286700356286, 'learning_rate/qwen_vl_interface': 9.81054442401643e-06, 'learning_rate/action_model': 9.81054442401643e-05, 'epoch': 0.04}) +[2026-10-03 13:07:09,951][starVLA.training.train_starvla][INFO] - Step 543, Loss: {'train/loss': 0.06060408800840378, 'train/grad_norm_pre_clip': 32.83003616333008, 'timing/data': 0.00034501000482123345, 'timing/model': 0.392274907004321, 'learning_rate/qwen_vl_interface': 9.809691966399834e-06, 'learning_rate/action_model': 9.809691966399835e-05, 'epoch': 0.04}) +[2026-10-03 13:07:10,653][starVLA.training.train_starvla][INFO] - Step 544, Loss: {'train/loss': 0.025839857757091522, 'train/grad_norm_pre_clip': 24.498659133911133, 'timing/data': 0.0003438400017330423, 'timing/model': 0.39157660699856933, 'learning_rate/qwen_vl_interface': 9.808837634466104e-06, 'learning_rate/action_model': 9.808837634466104e-05, 'epoch': 0.04}) +[2026-10-03 13:07:11,358][starVLA.training.train_starvla][INFO] - Step 545, Loss: {'train/loss': 0.36876916885375977, 'train/grad_norm_pre_clip': 126.2420883178711, 'timing/data': 0.00022562000958714634, 'timing/model': 0.3944263659941498, 'learning_rate/qwen_vl_interface': 9.807981428566428e-06, 'learning_rate/action_model': 9.807981428566427e-05, 'epoch': 0.04}) +[2026-10-03 13:07:12,248][starVLA.training.train_starvla][INFO] - Step 546, Loss: {'train/loss': 0.5654619336128235, 'train/grad_norm_pre_clip': 289.15118408203125, 'timing/data': 0.000258100000792183, 'timing/model': 0.5781307769939303, 'learning_rate/qwen_vl_interface': 9.807123349052754e-06, 'learning_rate/action_model': 9.807123349052753e-05, 'epoch': 0.04}) +[2026-10-03 13:07:12,950][starVLA.training.train_starvla][INFO] - Step 547, Loss: {'train/loss': 0.11044538021087646, 'train/grad_norm_pre_clip': 348.2940979003906, 'timing/data': 0.00020426000992301852, 'timing/model': 0.3923731270042481, 'learning_rate/qwen_vl_interface': 9.80626339627781e-06, 'learning_rate/action_model': 9.806263396277809e-05, 'epoch': 0.04}) +[2026-10-03 13:07:13,653][starVLA.training.train_starvla][INFO] - Step 548, Loss: {'train/loss': 0.9262509346008301, 'train/grad_norm_pre_clip': 219.4634552001953, 'timing/data': 0.0002545000024838373, 'timing/model': 0.39219684699492063, 'learning_rate/qwen_vl_interface': 9.805401570595087e-06, 'learning_rate/action_model': 9.805401570595087e-05, 'epoch': 0.04}) +[2026-10-03 13:07:14,357][starVLA.training.train_starvla][INFO] - Step 549, Loss: {'train/loss': 0.06781405210494995, 'train/grad_norm_pre_clip': 47.1602897644043, 'timing/data': 0.0002642299950821325, 'timing/model': 0.39334087699535303, 'learning_rate/qwen_vl_interface': 9.804537872358853e-06, 'learning_rate/action_model': 9.804537872358853e-05, 'epoch': 0.04}) +[2026-10-03 13:07:15,061][starVLA.training.train_starvla][INFO] - Step 550, Loss: {'train/loss': 0.16630534827709198, 'train/grad_norm_pre_clip': 83.97505187988281, 'timing/data': 0.00035710999509319663, 'timing/model': 0.39267174698761664, 'learning_rate/qwen_vl_interface': 9.80367230192414e-06, 'learning_rate/action_model': 9.803672301924138e-05, 'epoch': 0.04}) +[2026-10-03 13:07:15,951][starVLA.training.train_starvla][INFO] - Step 551, Loss: {'train/loss': 0.32900965213775635, 'train/grad_norm_pre_clip': 25.92246437072754, 'timing/data': 0.00030158000299707055, 'timing/model': 0.5794194469926879, 'learning_rate/qwen_vl_interface': 9.80280485964675e-06, 'learning_rate/action_model': 9.80280485964675e-05, 'epoch': 0.04}) +[2026-10-03 13:07:16,653][starVLA.training.train_starvla][INFO] - Step 552, Loss: {'train/loss': 0.02053792215883732, 'train/grad_norm_pre_clip': 18.33466148376465, 'timing/data': 0.0002554399979999289, 'timing/model': 0.3918690470018191, 'learning_rate/qwen_vl_interface': 9.801935545883255e-06, 'learning_rate/action_model': 9.801935545883255e-05, 'epoch': 0.04}) +[2026-10-03 13:07:17,358][starVLA.training.train_starvla][INFO] - Step 553, Loss: {'train/loss': 0.6254810690879822, 'train/grad_norm_pre_clip': 30.650474548339844, 'timing/data': 0.00026663999597076327, 'timing/model': 0.3937095059955027, 'learning_rate/qwen_vl_interface': 9.801064360991e-06, 'learning_rate/action_model': 9.801064360991e-05, 'epoch': 0.04}) +[2026-10-03 13:07:18,061][starVLA.training.train_starvla][INFO] - Step 554, Loss: {'train/loss': 0.2545887231826782, 'train/grad_norm_pre_clip': 217.14236450195312, 'timing/data': 0.0002495499938959256, 'timing/model': 0.39260300699970685, 'learning_rate/qwen_vl_interface': 9.800191305328095e-06, 'learning_rate/action_model': 9.800191305328094e-05, 'epoch': 0.04}) +[2026-10-03 13:07:18,764][starVLA.training.train_starvla][INFO] - Step 555, Loss: {'train/loss': 0.5599719882011414, 'train/grad_norm_pre_clip': 54.23257827758789, 'timing/data': 0.00033082999289035797, 'timing/model': 0.3922113470034674, 'learning_rate/qwen_vl_interface': 9.79931637925342e-06, 'learning_rate/action_model': 9.79931637925342e-05, 'epoch': 0.04}) +[2026-10-03 13:07:19,467][starVLA.training.train_starvla][INFO] - Step 556, Loss: {'train/loss': 0.06864618510007858, 'train/grad_norm_pre_clip': 110.79734802246094, 'timing/data': 0.00025313999503850937, 'timing/model': 0.39192699700652156, 'learning_rate/qwen_vl_interface': 9.798439583126625e-06, 'learning_rate/action_model': 9.798439583126625e-05, 'epoch': 0.04}) +[2026-10-03 13:07:20,360][starVLA.training.train_starvla][INFO] - Step 557, Loss: {'train/loss': 0.38719263672828674, 'train/grad_norm_pre_clip': 45.45492172241211, 'timing/data': 0.0002827400021487847, 'timing/model': 0.39351866699871607, 'learning_rate/qwen_vl_interface': 9.797560917308126e-06, 'learning_rate/action_model': 9.797560917308126e-05, 'epoch': 0.04}) +[2026-10-03 13:07:21,063][starVLA.training.train_starvla][INFO] - Step 558, Loss: {'train/loss': 0.830389142036438, 'train/grad_norm_pre_clip': 54.42416763305664, 'timing/data': 0.00032993000058922917, 'timing/model': 0.39278459700290114, 'learning_rate/qwen_vl_interface': 9.796680382159112e-06, 'learning_rate/action_model': 9.79668038215911e-05, 'epoch': 0.04}) +[2026-10-03 13:07:21,765][starVLA.training.train_starvla][INFO] - Step 559, Loss: {'train/loss': 0.08601593971252441, 'train/grad_norm_pre_clip': 37.431617736816406, 'timing/data': 0.0003030799998668954, 'timing/model': 0.3921249870036263, 'learning_rate/qwen_vl_interface': 9.795797978041533e-06, 'learning_rate/action_model': 9.795797978041533e-05, 'epoch': 0.04}) +[2026-10-03 13:07:22,468][starVLA.training.train_starvla][INFO] - Step 560, Loss: {'train/loss': 1.0350419282913208, 'train/grad_norm_pre_clip': 105.55497741699219, 'timing/data': 0.00037274000351317227, 'timing/model': 0.391824357007863, 'learning_rate/qwen_vl_interface': 9.794913705318117e-06, 'learning_rate/action_model': 9.794913705318118e-05, 'epoch': 0.04}) +[2026-10-03 13:07:23,173][starVLA.training.train_starvla][INFO] - Step 561, Loss: {'train/loss': 0.9932858943939209, 'train/grad_norm_pre_clip': 68.91801452636719, 'timing/data': 0.0003554299910319969, 'timing/model': 0.39358892700693104, 'learning_rate/qwen_vl_interface': 9.794027564352352e-06, 'learning_rate/action_model': 9.794027564352352e-05, 'epoch': 0.04}) +[2026-10-03 13:07:24,068][starVLA.training.train_starvla][INFO] - Step 562, Loss: {'train/loss': 0.02307017520070076, 'train/grad_norm_pre_clip': 12.21191692352295, 'timing/data': 0.00021284999093040824, 'timing/model': 0.394026166992262, 'learning_rate/qwen_vl_interface': 9.793139555508501e-06, 'learning_rate/action_model': 9.7931395555085e-05, 'epoch': 0.04}) +[2026-10-03 13:07:24,772][starVLA.training.train_starvla][INFO] - Step 563, Loss: {'train/loss': 0.09208305180072784, 'train/grad_norm_pre_clip': 56.0066032409668, 'timing/data': 0.00020700000459328294, 'timing/model': 0.39293158700456843, 'learning_rate/qwen_vl_interface': 9.792249679151585e-06, 'learning_rate/action_model': 9.792249679151585e-05, 'epoch': 0.04}) +[2026-10-03 13:07:25,474][starVLA.training.train_starvla][INFO] - Step 564, Loss: {'train/loss': 0.21353915333747864, 'train/grad_norm_pre_clip': 44.55348587036133, 'timing/data': 0.00029826999525539577, 'timing/model': 0.39201445700018667, 'learning_rate/qwen_vl_interface': 9.791357935647403e-06, 'learning_rate/action_model': 9.791357935647403e-05, 'epoch': 0.04}) +[2026-10-03 13:07:26,180][starVLA.training.train_starvla][INFO] - Step 565, Loss: {'train/loss': 0.015396523289382458, 'train/grad_norm_pre_clip': 2.970761775970459, 'timing/data': 0.00045680999755859375, 'timing/model': 0.39362937699479517, 'learning_rate/qwen_vl_interface': 9.790464325362516e-06, 'learning_rate/action_model': 9.790464325362516e-05, 'epoch': 0.04}) +[2026-10-03 13:07:26,883][starVLA.training.train_starvla][INFO] - Step 566, Loss: {'train/loss': 0.022428803145885468, 'train/grad_norm_pre_clip': 6.045121669769287, 'timing/data': 0.0002893900091294199, 'timing/model': 0.3927393070043763, 'learning_rate/qwen_vl_interface': 9.789568848664254e-06, 'learning_rate/action_model': 9.789568848664253e-05, 'epoch': 0.04}) +[2026-10-03 13:07:27,773][starVLA.training.train_starvla][INFO] - Step 567, Loss: {'train/loss': 0.03507333993911743, 'train/grad_norm_pre_clip': 12.02509880065918, 'timing/data': 0.0002665499923750758, 'timing/model': 0.5800761759892339, 'learning_rate/qwen_vl_interface': 9.788671505920712e-06, 'learning_rate/action_model': 9.788671505920711e-05, 'epoch': 0.04}) +[2026-10-03 13:07:28,478][starVLA.training.train_starvla][INFO] - Step 568, Loss: {'train/loss': 0.4560390114784241, 'train/grad_norm_pre_clip': 61.32966613769531, 'timing/data': 0.00021476999972946942, 'timing/model': 0.39292898699932266, 'learning_rate/qwen_vl_interface': 9.787772297500754e-06, 'learning_rate/action_model': 9.787772297500754e-05, 'epoch': 0.04}) +[2026-10-03 13:07:29,183][starVLA.training.train_starvla][INFO] - Step 569, Loss: {'train/loss': 0.05195958539843559, 'train/grad_norm_pre_clip': 42.26007843017578, 'timing/data': 0.0002875399950426072, 'timing/model': 0.39359174598939717, 'learning_rate/qwen_vl_interface': 9.78687122377401e-06, 'learning_rate/action_model': 9.78687122377401e-05, 'epoch': 0.04}) +[2026-10-03 13:07:29,888][starVLA.training.train_starvla][INFO] - Step 570, Loss: {'train/loss': 0.21632817387580872, 'train/grad_norm_pre_clip': 127.54051208496094, 'timing/data': 0.00022231999901123345, 'timing/model': 0.39439621700148564, 'learning_rate/qwen_vl_interface': 9.78596828511088e-06, 'learning_rate/action_model': 9.78596828511088e-05, 'epoch': 0.04}) +[2026-10-03 13:07:30,592][starVLA.training.train_starvla][INFO] - Step 571, Loss: {'train/loss': 0.012283165007829666, 'train/grad_norm_pre_clip': 2.4702718257904053, 'timing/data': 0.0003006799961440265, 'timing/model': 0.393167286994867, 'learning_rate/qwen_vl_interface': 9.785063481882527e-06, 'learning_rate/action_model': 9.785063481882527e-05, 'epoch': 0.04}) +[2026-10-03 13:07:31,486][starVLA.training.train_starvla][INFO] - Step 572, Loss: {'train/loss': 0.23719225823879242, 'train/grad_norm_pre_clip': 61.50644302368164, 'timing/data': 0.0002559300046414137, 'timing/model': 0.5820940970006632, 'learning_rate/qwen_vl_interface': 9.78415681446088e-06, 'learning_rate/action_model': 9.784156814460879e-05, 'epoch': 0.04}) +[2026-10-03 13:07:32,191][starVLA.training.train_starvla][INFO] - Step 573, Loss: {'train/loss': 0.07836070656776428, 'train/grad_norm_pre_clip': 58.42654800415039, 'timing/data': 0.0002647500077728182, 'timing/model': 0.39429927700257394, 'learning_rate/qwen_vl_interface': 9.783248283218638e-06, 'learning_rate/action_model': 9.783248283218638e-05, 'epoch': 0.04}) +[2026-10-03 13:07:32,896][starVLA.training.train_starvla][INFO] - Step 574, Loss: {'train/loss': 0.28782719373703003, 'train/grad_norm_pre_clip': 29.385040283203125, 'timing/data': 0.0003900500014424324, 'timing/model': 0.3931693469930906, 'learning_rate/qwen_vl_interface': 9.782337888529261e-06, 'learning_rate/action_model': 9.782337888529261e-05, 'epoch': 0.04}) +[2026-10-03 13:07:33,599][starVLA.training.train_starvla][INFO] - Step 575, Loss: {'train/loss': 0.003965276293456554, 'train/grad_norm_pre_clip': 1.54120671749115, 'timing/data': 0.0003533799899742007, 'timing/model': 0.39298389699251857, 'learning_rate/qwen_vl_interface': 9.781425630766978e-06, 'learning_rate/action_model': 9.781425630766979e-05, 'epoch': 0.04}) +[2026-10-03 13:07:34,305][starVLA.training.train_starvla][INFO] - Step 576, Loss: {'train/loss': 0.018647141754627228, 'train/grad_norm_pre_clip': 12.2394380569458, 'timing/data': 0.00021399000252131373, 'timing/model': 0.393251487010275, 'learning_rate/qwen_vl_interface': 9.780511510306789e-06, 'learning_rate/action_model': 9.780511510306788e-05, 'epoch': 0.04}) +[2026-10-03 13:07:35,204][starVLA.training.train_starvla][INFO] - Step 577, Loss: {'train/loss': 0.36156928539276123, 'train/grad_norm_pre_clip': 10.008089065551758, 'timing/data': 0.00026111000624950975, 'timing/model': 0.5872180650039809, 'learning_rate/qwen_vl_interface': 9.779595527524447e-06, 'learning_rate/action_model': 9.779595527524447e-05, 'epoch': 0.04}) +[2026-10-03 13:07:35,911][starVLA.training.train_starvla][INFO] - Step 578, Loss: {'train/loss': 0.00552596477791667, 'train/grad_norm_pre_clip': 2.9259793758392334, 'timing/data': 0.0003530100075295195, 'timing/model': 0.395252136993804, 'learning_rate/qwen_vl_interface': 9.778677682796483e-06, 'learning_rate/action_model': 9.778677682796483e-05, 'epoch': 0.04}) +[2026-10-03 13:07:36,616][starVLA.training.train_starvla][INFO] - Step 579, Loss: {'train/loss': 0.4922024607658386, 'train/grad_norm_pre_clip': 104.61969757080078, 'timing/data': 0.00027771999884862453, 'timing/model': 0.3933913959917845, 'learning_rate/qwen_vl_interface': 9.777757976500185e-06, 'learning_rate/action_model': 9.777757976500186e-05, 'epoch': 0.04}) +[2026-10-03 13:07:37,321][starVLA.training.train_starvla][INFO] - Step 580, Loss: {'train/loss': 0.1448131799697876, 'train/grad_norm_pre_clip': 51.125858306884766, 'timing/data': 0.0003053899999940768, 'timing/model': 0.3927659769979073, 'learning_rate/qwen_vl_interface': 9.776836409013614e-06, 'learning_rate/action_model': 9.776836409013615e-05, 'epoch': 0.04}) +[2026-10-03 13:07:38,027][starVLA.training.train_starvla][INFO] - Step 581, Loss: {'train/loss': 0.30607789754867554, 'train/grad_norm_pre_clip': 75.95701599121094, 'timing/data': 0.00030361000972334296, 'timing/model': 0.39442924699687865, 'learning_rate/qwen_vl_interface': 9.775912980715589e-06, 'learning_rate/action_model': 9.775912980715588e-05, 'epoch': 0.04}) +[2026-10-03 13:07:38,732][starVLA.training.train_starvla][INFO] - Step 582, Loss: {'train/loss': 0.0033500168938189745, 'train/grad_norm_pre_clip': 0.37074199318885803, 'timing/data': 0.0003557900054147467, 'timing/model': 0.3935465969989309, 'learning_rate/qwen_vl_interface': 9.774987691985694e-06, 'learning_rate/action_model': 9.774987691985694e-05, 'epoch': 0.04}) +[2026-10-03 13:07:39,628][starVLA.training.train_starvla][INFO] - Step 583, Loss: {'train/loss': 0.24512067437171936, 'train/grad_norm_pre_clip': 61.37358856201172, 'timing/data': 0.0003353999927639961, 'timing/model': 0.39144557699910365, 'learning_rate/qwen_vl_interface': 9.774060543204283e-06, 'learning_rate/action_model': 9.774060543204283e-05, 'epoch': 0.04}) +[2026-10-03 13:07:40,332][starVLA.training.train_starvla][INFO] - Step 584, Loss: {'train/loss': 0.004347153007984161, 'train/grad_norm_pre_clip': 0.5462140440940857, 'timing/data': 0.00020759999461006373, 'timing/model': 0.3926637469994603, 'learning_rate/qwen_vl_interface': 9.773131534752472e-06, 'learning_rate/action_model': 9.773131534752472e-05, 'epoch': 0.04}) +[2026-10-03 13:07:41,038][starVLA.training.train_starvla][INFO] - Step 585, Loss: {'train/loss': 0.19119255244731903, 'train/grad_norm_pre_clip': 33.0659065246582, 'timing/data': 0.00027176999719813466, 'timing/model': 0.39394175600318704, 'learning_rate/qwen_vl_interface': 9.772200667012142e-06, 'learning_rate/action_model': 9.772200667012142e-05, 'epoch': 0.04}) +[2026-10-03 13:07:41,743][starVLA.training.train_starvla][INFO] - Step 586, Loss: {'train/loss': 0.0038058259524405003, 'train/grad_norm_pre_clip': 0.706351637840271, 'timing/data': 0.00036019999242853373, 'timing/model': 0.39390912700037006, 'learning_rate/qwen_vl_interface': 9.771267940365934e-06, 'learning_rate/action_model': 9.771267940365935e-05, 'epoch': 0.04}) +[2026-10-03 13:07:42,446][starVLA.training.train_starvla][INFO] - Step 587, Loss: {'train/loss': 0.5915639996528625, 'train/grad_norm_pre_clip': 38.27401351928711, 'timing/data': 0.0002933200012193993, 'timing/model': 0.392279306994169, 'learning_rate/qwen_vl_interface': 9.770333355197262e-06, 'learning_rate/action_model': 9.77033335519726e-05, 'epoch': 0.04}) +[2026-10-03 13:07:43,339][starVLA.training.train_starvla][INFO] - Step 588, Loss: {'train/loss': 1.2423802614212036, 'train/grad_norm_pre_clip': 126.34078216552734, 'timing/data': 0.0002622600004542619, 'timing/model': 0.5821663060050923, 'learning_rate/qwen_vl_interface': 9.769396911890291e-06, 'learning_rate/action_model': 9.769396911890291e-05, 'epoch': 0.04}) +[2026-10-03 13:07:44,046][starVLA.training.train_starvla][INFO] - Step 589, Loss: {'train/loss': 0.0009666912956163287, 'train/grad_norm_pre_clip': 0.09377897530794144, 'timing/data': 0.0003202700027031824, 'timing/model': 0.3940991770068649, 'learning_rate/qwen_vl_interface': 9.768458610829965e-06, 'learning_rate/action_model': 9.768458610829965e-05, 'epoch': 0.04}) +[2026-10-03 13:07:44,752][starVLA.training.train_starvla][INFO] - Step 590, Loss: {'train/loss': 0.38087326288223267, 'train/grad_norm_pre_clip': 38.917171478271484, 'timing/data': 0.00026061000244226307, 'timing/model': 0.39400219699018635, 'learning_rate/qwen_vl_interface': 9.767518452401981e-06, 'learning_rate/action_model': 9.76751845240198e-05, 'epoch': 0.04}) +[2026-10-03 13:07:45,458][starVLA.training.train_starvla][INFO] - Step 591, Loss: {'train/loss': 0.46891167759895325, 'train/grad_norm_pre_clip': 109.20100402832031, 'timing/data': 0.0002432999899610877, 'timing/model': 0.3933424959977856, 'learning_rate/qwen_vl_interface': 9.766576436992802e-06, 'learning_rate/action_model': 9.766576436992801e-05, 'epoch': 0.04}) +[2026-10-03 13:07:46,167][starVLA.training.train_starvla][INFO] - Step 592, Loss: {'train/loss': 0.6347594261169434, 'train/grad_norm_pre_clip': 36.133243560791016, 'timing/data': 0.0002392799942754209, 'timing/model': 0.3930462160060415, 'learning_rate/qwen_vl_interface': 9.765632564989657e-06, 'learning_rate/action_model': 9.765632564989656e-05, 'epoch': 0.04}) +[2026-10-03 13:07:46,874][starVLA.training.train_starvla][INFO] - Step 593, Loss: {'train/loss': 0.23248329758644104, 'train/grad_norm_pre_clip': 55.587223052978516, 'timing/data': 0.0003286700084572658, 'timing/model': 0.39420486699964385, 'learning_rate/qwen_vl_interface': 9.764686836780533e-06, 'learning_rate/action_model': 9.764686836780532e-05, 'epoch': 0.04}) +[2026-10-03 13:07:47,768][starVLA.training.train_starvla][INFO] - Step 594, Loss: {'train/loss': 0.37260180711746216, 'train/grad_norm_pre_clip': 78.59229278564453, 'timing/data': 0.00022031000116840005, 'timing/model': 0.3935750460077543, 'learning_rate/qwen_vl_interface': 9.763739252754187e-06, 'learning_rate/action_model': 9.763739252754186e-05, 'epoch': 0.04}) +[2026-10-03 13:07:48,472][starVLA.training.train_starvla][INFO] - Step 595, Loss: {'train/loss': 0.48259100317955017, 'train/grad_norm_pre_clip': 75.39496612548828, 'timing/data': 0.00027840999246109277, 'timing/model': 0.39311076699232217, 'learning_rate/qwen_vl_interface': 9.76278981330013e-06, 'learning_rate/action_model': 9.762789813300131e-05, 'epoch': 0.04}) +[2026-10-03 13:07:49,176][starVLA.training.train_starvla][INFO] - Step 596, Loss: {'train/loss': 0.09172284603118896, 'train/grad_norm_pre_clip': 45.34976577758789, 'timing/data': 0.00038645000313408673, 'timing/model': 0.3924028570036171, 'learning_rate/qwen_vl_interface': 9.761838518808647e-06, 'learning_rate/action_model': 9.761838518808646e-05, 'epoch': 0.04}) +[2026-10-03 13:07:49,883][starVLA.training.train_starvla][INFO] - Step 597, Loss: {'train/loss': 0.005561657715588808, 'train/grad_norm_pre_clip': 2.0829484462738037, 'timing/data': 0.00029568999889306724, 'timing/model': 0.3940373869991163, 'learning_rate/qwen_vl_interface': 9.760885369670776e-06, 'learning_rate/action_model': 9.760885369670776e-05, 'epoch': 0.04}) +[2026-10-03 13:07:50,587][starVLA.training.train_starvla][INFO] - Step 598, Loss: {'train/loss': 0.10169044882059097, 'train/grad_norm_pre_clip': 53.34950637817383, 'timing/data': 0.0003350899933138862, 'timing/model': 0.39334101699932944, 'learning_rate/qwen_vl_interface': 9.759930366278319e-06, 'learning_rate/action_model': 9.759930366278318e-05, 'epoch': 0.04}) +[2026-10-03 13:07:51,485][starVLA.training.train_starvla][INFO] - Step 599, Loss: {'train/loss': 0.5864434242248535, 'train/grad_norm_pre_clip': 59.9364013671875, 'timing/data': 0.00028033999842591584, 'timing/model': 0.39447560599364806, 'learning_rate/qwen_vl_interface': 9.758973509023842e-06, 'learning_rate/action_model': 9.758973509023843e-05, 'epoch': 0.04}) +[2026-10-03 13:07:52,190][starVLA.training.train_starvla][INFO] - Step 600, Loss: {'train/loss': 0.3191325068473816, 'train/grad_norm_pre_clip': 50.57755661010742, 'timing/data': 0.0003267199936090037, 'timing/model': 0.39310437699896283, 'learning_rate/qwen_vl_interface': 9.758014798300677e-06, 'learning_rate/action_model': 9.758014798300676e-05, 'epoch': 0.04}) +[2026-10-03 13:07:52,896][starVLA.training.train_starvla][INFO] - Step 601, Loss: {'train/loss': 0.11725948750972748, 'train/grad_norm_pre_clip': 42.36493682861328, 'timing/data': 0.0003254100010963157, 'timing/model': 0.39413356700970326, 'learning_rate/qwen_vl_interface': 9.757054234502909e-06, 'learning_rate/action_model': 9.757054234502909e-05, 'epoch': 0.04}) +[2026-10-03 13:07:53,601][starVLA.training.train_starvla][INFO] - Step 602, Loss: {'train/loss': 0.02884187363088131, 'train/grad_norm_pre_clip': 15.838210105895996, 'timing/data': 0.00021543999901041389, 'timing/model': 0.39375791599741206, 'learning_rate/qwen_vl_interface': 9.756091818025394e-06, 'learning_rate/action_model': 9.756091818025394e-05, 'epoch': 0.04}) +[2026-10-03 13:07:54,304][starVLA.training.train_starvla][INFO] - Step 603, Loss: {'train/loss': 0.06288744509220123, 'train/grad_norm_pre_clip': 34.0449333190918, 'timing/data': 0.0003067500074394047, 'timing/model': 0.39268896699650213, 'learning_rate/qwen_vl_interface': 9.75512754926374e-06, 'learning_rate/action_model': 9.755127549263741e-05, 'epoch': 0.04}) +[2026-10-03 13:07:55,198][starVLA.training.train_starvla][INFO] - Step 604, Loss: {'train/loss': 0.09008714556694031, 'train/grad_norm_pre_clip': 55.40277862548828, 'timing/data': 0.0002110799978254363, 'timing/model': 0.5825689459888963, 'learning_rate/qwen_vl_interface': 9.754161428614325e-06, 'learning_rate/action_model': 9.754161428614324e-05, 'epoch': 0.04}) +[2026-10-03 13:07:55,904][starVLA.training.train_starvla][INFO] - Step 605, Loss: {'train/loss': 0.01166885532438755, 'train/grad_norm_pre_clip': 3.9396398067474365, 'timing/data': 0.00025321899738628417, 'timing/model': 0.39403306700114626, 'learning_rate/qwen_vl_interface': 9.753193456474284e-06, 'learning_rate/action_model': 9.753193456474284e-05, 'epoch': 0.04}) +[2026-10-03 13:07:56,610][starVLA.training.train_starvla][INFO] - Step 606, Loss: {'train/loss': 0.3867531716823578, 'train/grad_norm_pre_clip': 69.14207458496094, 'timing/data': 0.00033506000181660056, 'timing/model': 0.3940991870040307, 'learning_rate/qwen_vl_interface': 9.752223633241514e-06, 'learning_rate/action_model': 9.752223633241514e-05, 'epoch': 0.04}) +[2026-10-03 13:07:57,314][starVLA.training.train_starvla][INFO] - Step 607, Loss: {'train/loss': 0.045241039246320724, 'train/grad_norm_pre_clip': 22.44830894470215, 'timing/data': 0.0002923499996541068, 'timing/model': 0.3933936669927789, 'learning_rate/qwen_vl_interface': 9.751251959314671e-06, 'learning_rate/action_model': 9.751251959314672e-05, 'epoch': 0.04}) +[2026-10-03 13:07:58,019][starVLA.training.train_starvla][INFO] - Step 608, Loss: {'train/loss': 0.1571197211742401, 'train/grad_norm_pre_clip': 56.12080383300781, 'timing/data': 0.0003555799921741709, 'timing/model': 0.3928594969911501, 'learning_rate/qwen_vl_interface': 9.750278435093176e-06, 'learning_rate/action_model': 9.750278435093177e-05, 'epoch': 0.04}) +[2026-10-03 13:07:58,726][starVLA.training.train_starvla][INFO] - Step 609, Loss: {'train/loss': 0.10093677788972855, 'train/grad_norm_pre_clip': 64.5009765625, 'timing/data': 0.00031181999656837434, 'timing/model': 0.39500733598833904, 'learning_rate/qwen_vl_interface': 9.749303060977205e-06, 'learning_rate/action_model': 9.749303060977205e-05, 'epoch': 0.04}) +[2026-10-03 13:07:59,622][starVLA.training.train_starvla][INFO] - Step 610, Loss: {'train/loss': 0.13478726148605347, 'train/grad_norm_pre_clip': 127.55206298828125, 'timing/data': 0.0003292600013082847, 'timing/model': 0.3941560769890202, 'learning_rate/qwen_vl_interface': 9.748325837367701e-06, 'learning_rate/action_model': 9.748325837367702e-05, 'epoch': 0.04}) +[2026-10-03 13:08:00,327][starVLA.training.train_starvla][INFO] - Step 611, Loss: {'train/loss': 0.0028852936811745167, 'train/grad_norm_pre_clip': 0.7228174209594727, 'timing/data': 0.0003156000020680949, 'timing/model': 0.39310872700298205, 'learning_rate/qwen_vl_interface': 9.747346764666362e-06, 'learning_rate/action_model': 9.747346764666361e-05, 'epoch': 0.04}) +[2026-10-03 13:08:01,032][starVLA.training.train_starvla][INFO] - Step 612, Loss: {'train/loss': 0.008977913297712803, 'train/grad_norm_pre_clip': 3.6580698490142822, 'timing/data': 0.0003477100108284503, 'timing/model': 0.39294570600031875, 'learning_rate/qwen_vl_interface': 9.746365843275648e-06, 'learning_rate/action_model': 9.746365843275646e-05, 'epoch': 0.04}) +[2026-10-03 13:08:01,740][starVLA.training.train_starvla][INFO] - Step 613, Loss: {'train/loss': 0.1807975023984909, 'train/grad_norm_pre_clip': 58.317054748535156, 'timing/data': 0.00025286000163760036, 'timing/model': 0.3947704369929852, 'learning_rate/qwen_vl_interface': 9.745383073598776e-06, 'learning_rate/action_model': 9.745383073598775e-05, 'epoch': 0.04}) +[2026-10-03 13:08:02,446][starVLA.training.train_starvla][INFO] - Step 614, Loss: {'train/loss': 0.001038515125401318, 'train/grad_norm_pre_clip': 0.35336461663246155, 'timing/data': 0.00027211999986320734, 'timing/model': 0.39438529800099786, 'learning_rate/qwen_vl_interface': 9.744398456039729e-06, 'learning_rate/action_model': 9.744398456039729e-05, 'epoch': 0.04}) +[2026-10-03 13:08:03,342][starVLA.training.train_starvla][INFO] - Step 615, Loss: {'train/loss': 0.5921127796173096, 'train/grad_norm_pre_clip': 90.30807495117188, 'timing/data': 0.00030355999479070306, 'timing/model': 0.3931238770019263, 'learning_rate/qwen_vl_interface': 9.743411991003246e-06, 'learning_rate/action_model': 9.743411991003247e-05, 'epoch': 0.05}) +[2026-10-03 13:08:04,046][starVLA.training.train_starvla][INFO] - Step 616, Loss: {'train/loss': 0.3064078986644745, 'train/grad_norm_pre_clip': 152.9904327392578, 'timing/data': 0.000347869994584471, 'timing/model': 0.39268920599715784, 'learning_rate/qwen_vl_interface': 9.742423678894823e-06, 'learning_rate/action_model': 9.742423678894823e-05, 'epoch': 0.05}) +[2026-10-03 13:08:04,753][starVLA.training.train_starvla][INFO] - Step 617, Loss: {'train/loss': 0.01063590683043003, 'train/grad_norm_pre_clip': 11.598325729370117, 'timing/data': 0.0003632299922173843, 'timing/model': 0.3942226170038339, 'learning_rate/qwen_vl_interface': 9.741433520120722e-06, 'learning_rate/action_model': 9.74143352012072e-05, 'epoch': 0.05}) +[2026-10-03 13:08:05,459][starVLA.training.train_starvla][INFO] - Step 618, Loss: {'train/loss': 0.14908714592456818, 'train/grad_norm_pre_clip': 71.29385375976562, 'timing/data': 0.0002980299905175343, 'timing/model': 0.3938330570090329, 'learning_rate/qwen_vl_interface': 9.740441515087953e-06, 'learning_rate/action_model': 9.740441515087953e-05, 'epoch': 0.05}) +[2026-10-03 13:08:06,164][starVLA.training.train_starvla][INFO] - Step 619, Loss: {'train/loss': 0.06758076697587967, 'train/grad_norm_pre_clip': 75.9315185546875, 'timing/data': 0.00031821000447962433, 'timing/model': 0.39332449599169195, 'learning_rate/qwen_vl_interface': 9.739447664204298e-06, 'learning_rate/action_model': 9.739447664204297e-05, 'epoch': 0.05}) +[2026-10-03 13:08:07,057][starVLA.training.train_starvla][INFO] - Step 620, Loss: {'train/loss': 0.6828480958938599, 'train/grad_norm_pre_clip': 155.43472290039062, 'timing/data': 0.0003697700012708083, 'timing/model': 0.5816369060048601, 'learning_rate/qwen_vl_interface': 9.73845196787829e-06, 'learning_rate/action_model': 9.738451967878289e-05, 'epoch': 0.05}) +[2026-10-03 13:08:07,764][starVLA.training.train_starvla][INFO] - Step 621, Loss: {'train/loss': 0.16719238460063934, 'train/grad_norm_pre_clip': 63.850399017333984, 'timing/data': 0.0003330100007588044, 'timing/model': 0.3944221060082782, 'learning_rate/qwen_vl_interface': 9.737454426519221e-06, 'learning_rate/action_model': 9.737454426519221e-05, 'epoch': 0.05}) +[2026-10-03 13:08:08,469][starVLA.training.train_starvla][INFO] - Step 622, Loss: {'train/loss': 0.11395099759101868, 'train/grad_norm_pre_clip': 54.82609558105469, 'timing/data': 0.00021109999215696007, 'timing/model': 0.393923267009086, 'learning_rate/qwen_vl_interface': 9.736455040537145e-06, 'learning_rate/action_model': 9.736455040537144e-05, 'epoch': 0.05}) +[2026-10-03 13:08:09,174][starVLA.training.train_starvla][INFO] - Step 623, Loss: {'train/loss': 0.3243142366409302, 'train/grad_norm_pre_clip': 75.73466491699219, 'timing/data': 0.00022463999630417675, 'timing/model': 0.3931442769971909, 'learning_rate/qwen_vl_interface': 9.735453810342868e-06, 'learning_rate/action_model': 9.735453810342867e-05, 'epoch': 0.05}) +[2026-10-03 13:08:09,878][starVLA.training.train_starvla][INFO] - Step 624, Loss: {'train/loss': 0.003622031304985285, 'train/grad_norm_pre_clip': 1.0101613998413086, 'timing/data': 0.0003190300049027428, 'timing/model': 0.3929957869986538, 'learning_rate/qwen_vl_interface': 9.734450736347958e-06, 'learning_rate/action_model': 9.73445073634796e-05, 'epoch': 0.05}) +[2026-10-03 13:08:10,776][starVLA.training.train_starvla][INFO] - Step 625, Loss: {'train/loss': 0.30444610118865967, 'train/grad_norm_pre_clip': 105.3454818725586, 'timing/data': 0.0003031700034625828, 'timing/model': 0.5848671659914544, 'learning_rate/qwen_vl_interface': 9.733445818964746e-06, 'learning_rate/action_model': 9.733445818964746e-05, 'epoch': 0.05}) +[2026-10-03 13:08:11,482][starVLA.training.train_starvla][INFO] - Step 626, Loss: {'train/loss': 0.6033874750137329, 'train/grad_norm_pre_clip': 73.58028411865234, 'timing/data': 0.00039716000901535153, 'timing/model': 0.3944441880012164, 'learning_rate/qwen_vl_interface': 9.732439058606311e-06, 'learning_rate/action_model': 9.732439058606312e-05, 'epoch': 0.05}) +[2026-10-03 13:08:12,187][starVLA.training.train_starvla][INFO] - Step 627, Loss: {'train/loss': 0.0030965390615165234, 'train/grad_norm_pre_clip': 3.3069241046905518, 'timing/data': 0.0003264099941588938, 'timing/model': 0.39312571700429544, 'learning_rate/qwen_vl_interface': 9.731430455686499e-06, 'learning_rate/action_model': 9.731430455686498e-05, 'epoch': 0.05}) +[2026-10-03 13:08:12,892][starVLA.training.train_starvla][INFO] - Step 628, Loss: {'train/loss': 0.556195080280304, 'train/grad_norm_pre_clip': 105.23709869384766, 'timing/data': 0.0003219799982616678, 'timing/model': 0.3930790969898226, 'learning_rate/qwen_vl_interface': 9.730420010619901e-06, 'learning_rate/action_model': 9.7304200106199e-05, 'epoch': 0.05}) +[2026-10-03 13:08:13,599][starVLA.training.train_starvla][INFO] - Step 629, Loss: {'train/loss': 0.009100046008825302, 'train/grad_norm_pre_clip': 8.91602611541748, 'timing/data': 0.0002688100066734478, 'timing/model': 0.3946546569932252, 'learning_rate/qwen_vl_interface': 9.72940772382188e-06, 'learning_rate/action_model': 9.729407723821878e-05, 'epoch': 0.05}) +[2026-10-03 13:08:14,498][starVLA.training.train_starvla][INFO] - Step 630, Loss: {'train/loss': 0.06638409197330475, 'train/grad_norm_pre_clip': 110.42228698730469, 'timing/data': 0.0003718000079970807, 'timing/model': 0.587230195000302, 'learning_rate/qwen_vl_interface': 9.728393595708542e-06, 'learning_rate/action_model': 9.728393595708542e-05, 'epoch': 0.05}) +[2026-10-03 13:08:15,202][starVLA.training.train_starvla][INFO] - Step 631, Loss: {'train/loss': 0.3071024715900421, 'train/grad_norm_pre_clip': 53.46237564086914, 'timing/data': 0.0002149700012523681, 'timing/model': 0.3933464560104767, 'learning_rate/qwen_vl_interface': 9.727377626696762e-06, 'learning_rate/action_model': 9.727377626696762e-05, 'epoch': 0.05}) +[2026-10-03 13:08:15,906][starVLA.training.train_starvla][INFO] - Step 632, Loss: {'train/loss': 0.0006254848558455706, 'train/grad_norm_pre_clip': 0.18536968529224396, 'timing/data': 0.000245080009335652, 'timing/model': 0.3925250870088348, 'learning_rate/qwen_vl_interface': 9.726359817204165e-06, 'learning_rate/action_model': 9.726359817204166e-05, 'epoch': 0.05}) +[2026-10-03 13:08:16,613][starVLA.training.train_starvla][INFO] - Step 633, Loss: {'train/loss': 0.006854054052382708, 'train/grad_norm_pre_clip': 7.499763011932373, 'timing/data': 0.0003464390028966591, 'timing/model': 0.3949085280037252, 'learning_rate/qwen_vl_interface': 9.725340167649136e-06, 'learning_rate/action_model': 9.725340167649136e-05, 'epoch': 0.05}) +[2026-10-03 13:08:17,319][starVLA.training.train_starvla][INFO] - Step 634, Loss: {'train/loss': 0.00047360791359096766, 'train/grad_norm_pre_clip': 0.14169317483901978, 'timing/data': 0.0003348400059621781, 'timing/model': 0.39417698699980974, 'learning_rate/qwen_vl_interface': 9.72431867845081e-06, 'learning_rate/action_model': 9.72431867845081e-05, 'epoch': 0.05}) +[2026-10-03 13:08:18,024][starVLA.training.train_starvla][INFO] - Step 635, Loss: {'train/loss': 0.00033469669870100915, 'train/grad_norm_pre_clip': 0.1899067610502243, 'timing/data': 0.00032763999479357153, 'timing/model': 0.39333477600303013, 'learning_rate/qwen_vl_interface': 9.723295350029086e-06, 'learning_rate/action_model': 9.723295350029086e-05, 'epoch': 0.05}) +[2026-10-03 13:08:18,924][starVLA.training.train_starvla][INFO] - Step 636, Loss: {'train/loss': 0.7217212915420532, 'train/grad_norm_pre_clip': 22.20235824584961, 'timing/data': 0.0003712600009748712, 'timing/model': 0.3933305270038545, 'learning_rate/qwen_vl_interface': 9.722270182804616e-06, 'learning_rate/action_model': 9.722270182804614e-05, 'epoch': 0.05}) +[2026-10-03 13:08:19,631][starVLA.training.train_starvla][INFO] - Step 637, Loss: {'train/loss': 0.19590692222118378, 'train/grad_norm_pre_clip': 76.37783813476562, 'timing/data': 0.00025112999719567597, 'timing/model': 0.39507007600332145, 'learning_rate/qwen_vl_interface': 9.721243177198805e-06, 'learning_rate/action_model': 9.721243177198804e-05, 'epoch': 0.05}) +[2026-10-03 13:08:20,338][starVLA.training.train_starvla][INFO] - Step 638, Loss: {'train/loss': 0.19767554104328156, 'train/grad_norm_pre_clip': 151.72467041015625, 'timing/data': 0.0002963299921248108, 'timing/model': 0.394985427003121, 'learning_rate/qwen_vl_interface': 9.720214333633817e-06, 'learning_rate/action_model': 9.720214333633817e-05, 'epoch': 0.05}) +[2026-10-03 13:08:21,042][starVLA.training.train_starvla][INFO] - Step 639, Loss: {'train/loss': 0.7612869143486023, 'train/grad_norm_pre_clip': 118.71478271484375, 'timing/data': 0.0003003699966939166, 'timing/model': 0.3935515769990161, 'learning_rate/qwen_vl_interface': 9.719183652532572e-06, 'learning_rate/action_model': 9.719183652532572e-05, 'epoch': 0.05}) +[2026-10-03 13:08:21,748][starVLA.training.train_starvla][INFO] - Step 640, Loss: {'train/loss': 0.3062225580215454, 'train/grad_norm_pre_clip': 65.56021118164062, 'timing/data': 0.00026852999872062355, 'timing/model': 0.39291843699174933, 'learning_rate/qwen_vl_interface': 9.718151134318743e-06, 'learning_rate/action_model': 9.718151134318743e-05, 'epoch': 0.05}) +[2026-10-03 13:08:22,645][starVLA.training.train_starvla][INFO] - Step 641, Loss: {'train/loss': 0.011559400707483292, 'train/grad_norm_pre_clip': 21.98436164855957, 'timing/data': 0.0002535399980843067, 'timing/model': 0.5841161370044574, 'learning_rate/qwen_vl_interface': 9.71711677941676e-06, 'learning_rate/action_model': 9.717116779416758e-05, 'epoch': 0.05}) +[2026-10-03 13:08:23,351][starVLA.training.train_starvla][INFO] - Step 642, Loss: {'train/loss': 0.004823905415832996, 'train/grad_norm_pre_clip': 1.2181416749954224, 'timing/data': 0.00032716999703552574, 'timing/model': 0.3934950469993055, 'learning_rate/qwen_vl_interface': 9.716080588251805e-06, 'learning_rate/action_model': 9.716080588251805e-05, 'epoch': 0.05}) +[2026-10-03 13:08:24,055][starVLA.training.train_starvla][INFO] - Step 643, Loss: {'train/loss': 1.2223453521728516, 'train/grad_norm_pre_clip': 108.38115692138672, 'timing/data': 0.00037220001104287803, 'timing/model': 0.3930746870028088, 'learning_rate/qwen_vl_interface': 9.71504256124982e-06, 'learning_rate/action_model': 9.71504256124982e-05, 'epoch': 0.05}) +[2026-10-03 13:08:24,759][starVLA.training.train_starvla][INFO] - Step 644, Loss: {'train/loss': 0.2724258005619049, 'train/grad_norm_pre_clip': 96.48778533935547, 'timing/data': 0.0002976300020236522, 'timing/model': 0.3922474959981628, 'learning_rate/qwen_vl_interface': 9.714002698837499e-06, 'learning_rate/action_model': 9.714002698837498e-05, 'epoch': 0.05}) +[2026-10-03 13:08:25,465][starVLA.training.train_starvla][INFO] - Step 645, Loss: {'train/loss': 0.016653237864375114, 'train/grad_norm_pre_clip': 28.9355525970459, 'timing/data': 0.0002796199987642467, 'timing/model': 0.3935513169999467, 'learning_rate/qwen_vl_interface': 9.712961001442285e-06, 'learning_rate/action_model': 9.712961001442286e-05, 'epoch': 0.05}) +[2026-10-03 13:08:26,169][starVLA.training.train_starvla][INFO] - Step 646, Loss: {'train/loss': 0.0008709619287401438, 'train/grad_norm_pre_clip': 0.1863042265176773, 'timing/data': 0.00029399000050034374, 'timing/model': 0.39306854699680116, 'learning_rate/qwen_vl_interface': 9.711917469492388e-06, 'learning_rate/action_model': 9.711917469492388e-05, 'epoch': 0.05}) +[2026-10-03 13:08:27,064][starVLA.training.train_starvla][INFO] - Step 647, Loss: {'train/loss': 0.5921816229820251, 'train/grad_norm_pre_clip': 13.34832763671875, 'timing/data': 0.0003235400072298944, 'timing/model': 0.39304413698846474, 'learning_rate/qwen_vl_interface': 9.71087210341676e-06, 'learning_rate/action_model': 9.71087210341676e-05, 'epoch': 0.05}) +[2026-10-03 13:08:27,768][starVLA.training.train_starvla][INFO] - Step 648, Loss: {'train/loss': 0.007105010561645031, 'train/grad_norm_pre_clip': 17.937788009643555, 'timing/data': 0.00031582999508827925, 'timing/model': 0.39257422799710184, 'learning_rate/qwen_vl_interface': 9.70982490364511e-06, 'learning_rate/action_model': 9.70982490364511e-05, 'epoch': 0.05}) +[2026-10-03 13:08:28,474][starVLA.training.train_starvla][INFO] - Step 649, Loss: {'train/loss': 0.1292305737733841, 'train/grad_norm_pre_clip': 130.901611328125, 'timing/data': 0.0003346900048200041, 'timing/model': 0.3939853369956836, 'learning_rate/qwen_vl_interface': 9.708775870607905e-06, 'learning_rate/action_model': 9.708775870607904e-05, 'epoch': 0.05}) +[2026-10-03 13:08:29,179][starVLA.training.train_starvla][INFO] - Step 650, Loss: {'train/loss': 0.1371644139289856, 'train/grad_norm_pre_clip': 120.54649353027344, 'timing/data': 0.0004075699980603531, 'timing/model': 0.3934633059980115, 'learning_rate/qwen_vl_interface': 9.707725004736363e-06, 'learning_rate/action_model': 9.707725004736363e-05, 'epoch': 0.05}) +[2026-10-03 13:08:29,882][starVLA.training.train_starvla][INFO] - Step 651, Loss: {'train/loss': 1.322298288345337, 'train/grad_norm_pre_clip': 144.5841827392578, 'timing/data': 0.0002543400041759014, 'timing/model': 0.39343122599530034, 'learning_rate/qwen_vl_interface': 9.706672306462454e-06, 'learning_rate/action_model': 9.706672306462454e-05, 'epoch': 0.05}) +[2026-10-03 13:08:30,779][starVLA.training.train_starvla][INFO] - Step 652, Loss: {'train/loss': 0.08740396797657013, 'train/grad_norm_pre_clip': 115.84551239013672, 'timing/data': 0.0004629800096154213, 'timing/model': 0.3929754069977207, 'learning_rate/qwen_vl_interface': 9.705617776218905e-06, 'learning_rate/action_model': 9.705617776218904e-05, 'epoch': 0.05}) +[2026-10-03 13:08:31,486][starVLA.training.train_starvla][INFO] - Step 653, Loss: {'train/loss': 0.024566443637013435, 'train/grad_norm_pre_clip': 44.072540283203125, 'timing/data': 0.0003650500002549961, 'timing/model': 0.39439991700055543, 'learning_rate/qwen_vl_interface': 9.70456141443919e-06, 'learning_rate/action_model': 9.704561414439188e-05, 'epoch': 0.05}) +[2026-10-03 13:08:32,192][starVLA.training.train_starvla][INFO] - Step 654, Loss: {'train/loss': 0.3221985101699829, 'train/grad_norm_pre_clip': 85.2832260131836, 'timing/data': 0.0002682699996512383, 'timing/model': 0.3944604769931175, 'learning_rate/qwen_vl_interface': 9.703503221557538e-06, 'learning_rate/action_model': 9.703503221557538e-05, 'epoch': 0.05}) +[2026-10-03 13:08:32,897][starVLA.training.train_starvla][INFO] - Step 655, Loss: {'train/loss': 0.15844687819480896, 'train/grad_norm_pre_clip': 39.44158935546875, 'timing/data': 0.00020614999812096357, 'timing/model': 0.3935677069966914, 'learning_rate/qwen_vl_interface': 9.70244319800894e-06, 'learning_rate/action_model': 9.70244319800894e-05, 'epoch': 0.05}) +[2026-10-03 13:08:33,602][starVLA.training.train_starvla][INFO] - Step 656, Loss: {'train/loss': 0.0787850022315979, 'train/grad_norm_pre_clip': 80.3233871459961, 'timing/data': 0.00020444999972824007, 'timing/model': 0.39342213700001594, 'learning_rate/qwen_vl_interface': 9.701381344229124e-06, 'learning_rate/action_model': 9.701381344229123e-05, 'epoch': 0.05}) +[2026-10-03 13:08:34,501][starVLA.training.train_starvla][INFO] - Step 657, Loss: {'train/loss': 0.017584960907697678, 'train/grad_norm_pre_clip': 5.63934850692749, 'timing/data': 0.00027393999334890395, 'timing/model': 0.5868466060055653, 'learning_rate/qwen_vl_interface': 9.700317660654581e-06, 'learning_rate/action_model': 9.700317660654581e-05, 'epoch': 0.05}) +[2026-10-03 13:08:35,208][starVLA.training.train_starvla][INFO] - Step 658, Loss: {'train/loss': 0.3624937832355499, 'train/grad_norm_pre_clip': 78.46356964111328, 'timing/data': 0.00020917000074405223, 'timing/model': 0.39521181798772886, 'learning_rate/qwen_vl_interface': 9.699252147722554e-06, 'learning_rate/action_model': 9.699252147722553e-05, 'epoch': 0.05}) +[2026-10-03 13:08:35,912][starVLA.training.train_starvla][INFO] - Step 659, Loss: {'train/loss': 0.48983290791511536, 'train/grad_norm_pre_clip': 15.221967697143555, 'timing/data': 0.000341070001013577, 'timing/model': 0.3936306769901421, 'learning_rate/qwen_vl_interface': 9.698184805871028e-06, 'learning_rate/action_model': 9.698184805871028e-05, 'epoch': 0.05}) +[2026-10-03 13:08:36,617][starVLA.training.train_starvla][INFO] - Step 660, Loss: {'train/loss': 0.33662062883377075, 'train/grad_norm_pre_clip': 73.27803802490234, 'timing/data': 0.00030667000100947917, 'timing/model': 0.39369518600869924, 'learning_rate/qwen_vl_interface': 9.697115635538753e-06, 'learning_rate/action_model': 9.697115635538753e-05, 'epoch': 0.05}) +[2026-10-03 13:08:37,324][starVLA.training.train_starvla][INFO] - Step 661, Loss: {'train/loss': 0.8436379432678223, 'train/grad_norm_pre_clip': 41.730247497558594, 'timing/data': 0.0004605200083460659, 'timing/model': 0.3945076270028949, 'learning_rate/qwen_vl_interface': 9.69604463716522e-06, 'learning_rate/action_model': 9.69604463716522e-05, 'epoch': 0.05}) +[2026-10-03 13:08:38,030][starVLA.training.train_starvla][INFO] - Step 662, Loss: {'train/loss': 0.8730348348617554, 'train/grad_norm_pre_clip': 137.39816284179688, 'timing/data': 0.00027057999977841973, 'timing/model': 0.394132246990921, 'learning_rate/qwen_vl_interface': 9.69497181119068e-06, 'learning_rate/action_model': 9.694971811190681e-05, 'epoch': 0.05}) +[2026-10-03 13:08:38,923][starVLA.training.train_starvla][INFO] - Step 663, Loss: {'train/loss': 0.007899153046309948, 'train/grad_norm_pre_clip': 2.163588285446167, 'timing/data': 0.0003566800005501136, 'timing/model': 0.39378849700733554, 'learning_rate/qwen_vl_interface': 9.693897158056132e-06, 'learning_rate/action_model': 9.693897158056132e-05, 'epoch': 0.05}) +[2026-10-03 13:08:39,628][starVLA.training.train_starvla][INFO] - Step 664, Loss: {'train/loss': 0.47983646392822266, 'train/grad_norm_pre_clip': 23.26576805114746, 'timing/data': 0.0002728300023591146, 'timing/model': 0.3933021260017995, 'learning_rate/qwen_vl_interface': 9.692820678203322e-06, 'learning_rate/action_model': 9.692820678203322e-05, 'epoch': 0.05}) +[2026-10-03 13:08:40,335][starVLA.training.train_starvla][INFO] - Step 665, Loss: {'train/loss': 0.02967856638133526, 'train/grad_norm_pre_clip': 22.735925674438477, 'timing/data': 0.00034679999225772917, 'timing/model': 0.3944876369932899, 'learning_rate/qwen_vl_interface': 9.691742372074749e-06, 'learning_rate/action_model': 9.691742372074749e-05, 'epoch': 0.05}) +[2026-10-03 13:08:41,040][starVLA.training.train_starvla][INFO] - Step 666, Loss: {'train/loss': 0.16304415464401245, 'train/grad_norm_pre_clip': 95.4809341430664, 'timing/data': 0.0003352499916218221, 'timing/model': 0.3938513769971905, 'learning_rate/qwen_vl_interface': 9.690662240113669e-06, 'learning_rate/action_model': 9.690662240113668e-05, 'epoch': 0.05}) +[2026-10-03 13:08:41,744][starVLA.training.train_starvla][INFO] - Step 667, Loss: {'train/loss': 0.018762651830911636, 'train/grad_norm_pre_clip': 2.3745689392089844, 'timing/data': 0.0002509099867893383, 'timing/model': 0.39326110600086395, 'learning_rate/qwen_vl_interface': 9.689580282764077e-06, 'learning_rate/action_model': 9.689580282764077e-05, 'epoch': 0.05}) +[2026-10-03 13:08:42,641][starVLA.training.train_starvla][INFO] - Step 668, Loss: {'train/loss': 0.07217275351285934, 'train/grad_norm_pre_clip': 60.84415054321289, 'timing/data': 0.00021782000840175897, 'timing/model': 0.3937115669978084, 'learning_rate/qwen_vl_interface': 9.688496500470731e-06, 'learning_rate/action_model': 9.68849650047073e-05, 'epoch': 0.05}) +[2026-10-03 13:08:43,349][starVLA.training.train_starvla][INFO] - Step 669, Loss: {'train/loss': 0.24525687098503113, 'train/grad_norm_pre_clip': 56.01758575439453, 'timing/data': 0.0003514300042297691, 'timing/model': 0.394531707992428, 'learning_rate/qwen_vl_interface': 9.68741089367913e-06, 'learning_rate/action_model': 9.687410893679129e-05, 'epoch': 0.05}) +[2026-10-03 13:08:44,055][starVLA.training.train_starvla][INFO] - Step 670, Loss: {'train/loss': 0.07383603602647781, 'train/grad_norm_pre_clip': 48.96723937988281, 'timing/data': 0.00036702999204862863, 'timing/model': 0.3953643970016856, 'learning_rate/qwen_vl_interface': 9.686323462835527e-06, 'learning_rate/action_model': 9.686323462835526e-05, 'epoch': 0.05}) +[2026-10-03 13:08:44,760][starVLA.training.train_starvla][INFO] - Step 671, Loss: {'train/loss': 0.22509756684303284, 'train/grad_norm_pre_clip': 171.9269256591797, 'timing/data': 0.00020147999748587608, 'timing/model': 0.3933832159964368, 'learning_rate/qwen_vl_interface': 9.685234208386922e-06, 'learning_rate/action_model': 9.685234208386921e-05, 'epoch': 0.05}) +[2026-10-03 13:08:45,464][starVLA.training.train_starvla][INFO] - Step 672, Loss: {'train/loss': 0.4213966131210327, 'train/grad_norm_pre_clip': 9.936859130859375, 'timing/data': 0.00028248000307939947, 'timing/model': 0.39317083699279465, 'learning_rate/qwen_vl_interface': 9.684143130781066e-06, 'learning_rate/action_model': 9.684143130781066e-05, 'epoch': 0.05}) +[2026-10-03 13:08:46,359][starVLA.training.train_starvla][INFO] - Step 673, Loss: {'train/loss': 0.011459993198513985, 'train/grad_norm_pre_clip': 3.3152101039886475, 'timing/data': 0.0002804399991873652, 'timing/model': 0.5837835689890198, 'learning_rate/qwen_vl_interface': 9.683050230466462e-06, 'learning_rate/action_model': 9.683050230466462e-05, 'epoch': 0.05}) +[2026-10-03 13:08:47,065][starVLA.training.train_starvla][INFO] - Step 674, Loss: {'train/loss': 0.5180074572563171, 'train/grad_norm_pre_clip': 199.36654663085938, 'timing/data': 0.0004209500038996339, 'timing/model': 0.3941212199861184, 'learning_rate/qwen_vl_interface': 9.681955507892361e-06, 'learning_rate/action_model': 9.68195550789236e-05, 'epoch': 0.05}) +[2026-10-03 13:08:47,769][starVLA.training.train_starvla][INFO] - Step 675, Loss: {'train/loss': 0.6835999488830566, 'train/grad_norm_pre_clip': 143.29635620117188, 'timing/data': 0.0002776299952529371, 'timing/model': 0.39306800899794325, 'learning_rate/qwen_vl_interface': 9.680858963508759e-06, 'learning_rate/action_model': 9.68085896350876e-05, 'epoch': 0.05}) +[2026-10-03 13:08:48,473][starVLA.training.train_starvla][INFO] - Step 676, Loss: {'train/loss': 0.012345482595264912, 'train/grad_norm_pre_clip': 1.540844202041626, 'timing/data': 0.0002760499919531867, 'timing/model': 0.3929179500119062, 'learning_rate/qwen_vl_interface': 9.679760597766405e-06, 'learning_rate/action_model': 9.679760597766405e-05, 'epoch': 0.05}) +[2026-10-03 13:08:49,180][starVLA.training.train_starvla][INFO] - Step 677, Loss: {'train/loss': 0.6282202005386353, 'train/grad_norm_pre_clip': 38.89385986328125, 'timing/data': 0.0003467799979262054, 'timing/model': 0.39495132000593003, 'learning_rate/qwen_vl_interface': 9.678660411116796e-06, 'learning_rate/action_model': 9.678660411116796e-05, 'epoch': 0.05}) +[2026-10-03 13:08:50,077][starVLA.training.train_starvla][INFO] - Step 678, Loss: {'train/loss': 0.005748945288360119, 'train/grad_norm_pre_clip': 1.6041133403778076, 'timing/data': 0.00021780999668408185, 'timing/model': 0.5854127250058809, 'learning_rate/qwen_vl_interface': 9.677558404012178e-06, 'learning_rate/action_model': 9.677558404012177e-05, 'epoch': 0.05}) +[2026-10-03 13:08:50,782][starVLA.training.train_starvla][INFO] - Step 679, Loss: {'train/loss': 0.24701444804668427, 'train/grad_norm_pre_clip': 11.710511207580566, 'timing/data': 0.0003227300039725378, 'timing/model': 0.39380946999881417, 'learning_rate/qwen_vl_interface': 9.676454576905542e-06, 'learning_rate/action_model': 9.676454576905542e-05, 'epoch': 0.05}) +[2026-10-03 13:08:51,488][starVLA.training.train_starvla][INFO] - Step 680, Loss: {'train/loss': 0.2538842558860779, 'train/grad_norm_pre_clip': 32.0847282409668, 'timing/data': 0.00021661999926436692, 'timing/model': 0.39437404900672846, 'learning_rate/qwen_vl_interface': 9.675348930250633e-06, 'learning_rate/action_model': 9.675348930250632e-05, 'epoch': 0.05}) +[2026-10-03 13:08:52,196][starVLA.training.train_starvla][INFO] - Step 681, Loss: {'train/loss': 0.298467218875885, 'train/grad_norm_pre_clip': 55.8826904296875, 'timing/data': 0.0003546700027072802, 'timing/model': 0.3948938600078691, 'learning_rate/qwen_vl_interface': 9.674241464501938e-06, 'learning_rate/action_model': 9.674241464501937e-05, 'epoch': 0.05}) +[2026-10-03 13:08:52,903][starVLA.training.train_starvla][INFO] - Step 682, Loss: {'train/loss': 0.8932445645332336, 'train/grad_norm_pre_clip': 90.30036163330078, 'timing/data': 0.0003496589924907312, 'timing/model': 0.39449842998874374, 'learning_rate/qwen_vl_interface': 9.673132180114695e-06, 'learning_rate/action_model': 9.673132180114695e-05, 'epoch': 0.05}) +[2026-10-03 13:08:53,798][starVLA.training.train_starvla][INFO] - Step 683, Loss: {'train/loss': 0.3249208629131317, 'train/grad_norm_pre_clip': 48.61623764038086, 'timing/data': 0.00030267999682109803, 'timing/model': 0.5840329859929625, 'learning_rate/qwen_vl_interface': 9.67202107754489e-06, 'learning_rate/action_model': 9.672021077544889e-05, 'epoch': 0.05}) +[2026-10-03 13:08:54,503][starVLA.training.train_starvla][INFO] - Step 684, Loss: {'train/loss': 0.016142891719937325, 'train/grad_norm_pre_clip': 2.82206130027771, 'timing/data': 0.00029167899629101157, 'timing/model': 0.39307510999788065, 'learning_rate/qwen_vl_interface': 9.670908157249254e-06, 'learning_rate/action_model': 9.670908157249253e-05, 'epoch': 0.05}) +[2026-10-03 13:08:55,211][starVLA.training.train_starvla][INFO] - Step 685, Loss: {'train/loss': 0.047056395560503006, 'train/grad_norm_pre_clip': 39.893062591552734, 'timing/data': 0.00027132000832352787, 'timing/model': 0.3949627099937061, 'learning_rate/qwen_vl_interface': 9.669793419685267e-06, 'learning_rate/action_model': 9.669793419685266e-05, 'epoch': 0.05}) +[2026-10-03 13:08:55,915][starVLA.training.train_starvla][INFO] - Step 686, Loss: {'train/loss': 0.017741389572620392, 'train/grad_norm_pre_clip': 13.323071479797363, 'timing/data': 0.0002666500076884404, 'timing/model': 0.3939547890040558, 'learning_rate/qwen_vl_interface': 9.668676865311155e-06, 'learning_rate/action_model': 9.668676865311154e-05, 'epoch': 0.05}) +[2026-10-03 13:08:56,620][starVLA.training.train_starvla][INFO] - Step 687, Loss: {'train/loss': 0.021305030211806297, 'train/grad_norm_pre_clip': 7.517963409423828, 'timing/data': 0.0003428999916650355, 'timing/model': 0.3932525300042471, 'learning_rate/qwen_vl_interface': 9.667558494585891e-06, 'learning_rate/action_model': 9.66755849458589e-05, 'epoch': 0.05}) +[2026-10-03 13:08:57,325][starVLA.training.train_starvla][INFO] - Step 688, Loss: {'train/loss': 0.00263569806702435, 'train/grad_norm_pre_clip': 0.786736011505127, 'timing/data': 0.0003365199954714626, 'timing/model': 0.3931202800013125, 'learning_rate/qwen_vl_interface': 9.666438307969194e-06, 'learning_rate/action_model': 9.666438307969195e-05, 'epoch': 0.05}) +[2026-10-03 13:08:58,228][starVLA.training.train_starvla][INFO] - Step 689, Loss: {'train/loss': 0.005630088038742542, 'train/grad_norm_pre_clip': 4.852416038513184, 'timing/data': 0.0003590599953895435, 'timing/model': 0.39383750899287406, 'learning_rate/qwen_vl_interface': 9.665316305921535e-06, 'learning_rate/action_model': 9.665316305921535e-05, 'epoch': 0.05}) +[2026-10-03 13:08:58,933][starVLA.training.train_starvla][INFO] - Step 690, Loss: {'train/loss': 0.2513105869293213, 'train/grad_norm_pre_clip': 13.665311813354492, 'timing/data': 0.0003693100006785244, 'timing/model': 0.39439778999076225, 'learning_rate/qwen_vl_interface': 9.664192488904123e-06, 'learning_rate/action_model': 9.664192488904122e-05, 'epoch': 0.05}) +[2026-10-03 13:08:59,637][starVLA.training.train_starvla][INFO] - Step 691, Loss: {'train/loss': 0.0060425312258303165, 'train/grad_norm_pre_clip': 1.2014503479003906, 'timing/data': 0.0003141499910270795, 'timing/model': 0.3929048100108048, 'learning_rate/qwen_vl_interface': 9.663066857378917e-06, 'learning_rate/action_model': 9.663066857378917e-05, 'epoch': 0.05}) +[2026-10-03 13:09:00,340][starVLA.training.train_starvla][INFO] - Step 692, Loss: {'train/loss': 0.4237932562828064, 'train/grad_norm_pre_clip': 19.752193450927734, 'timing/data': 0.0002115300012519583, 'timing/model': 0.3930684099905193, 'learning_rate/qwen_vl_interface': 9.661939411808623e-06, 'learning_rate/action_model': 9.661939411808622e-05, 'epoch': 0.05}) +[2026-10-03 13:09:01,047][starVLA.training.train_starvla][INFO] - Step 693, Loss: {'train/loss': 0.0009233254240825772, 'train/grad_norm_pre_clip': 0.13978345692157745, 'timing/data': 0.00028684000426437706, 'timing/model': 0.39430831000208855, 'learning_rate/qwen_vl_interface': 9.66081015265669e-06, 'learning_rate/action_model': 9.660810152656689e-05, 'epoch': 0.05}) +[2026-10-03 13:09:01,943][starVLA.training.train_starvla][INFO] - Step 694, Loss: {'train/loss': 0.5508108735084534, 'train/grad_norm_pre_clip': 189.5029754638672, 'timing/data': 0.000341520004440099, 'timing/model': 0.5843072150018997, 'learning_rate/qwen_vl_interface': 9.659679080387314e-06, 'learning_rate/action_model': 9.659679080387313e-05, 'epoch': 0.05}) +[2026-10-03 13:09:02,646][starVLA.training.train_starvla][INFO] - Step 695, Loss: {'train/loss': 0.706579864025116, 'train/grad_norm_pre_clip': 57.7471923828125, 'timing/data': 0.0003382599970791489, 'timing/model': 0.39238181100517977, 'learning_rate/qwen_vl_interface': 9.658546195465436e-06, 'learning_rate/action_model': 9.658546195465435e-05, 'epoch': 0.05}) +[2026-10-03 13:09:03,350][starVLA.training.train_starvla][INFO] - Step 696, Loss: {'train/loss': 0.011471924372017384, 'train/grad_norm_pre_clip': 5.214977741241455, 'timing/data': 0.00025611999444663525, 'timing/model': 0.3925739100086503, 'learning_rate/qwen_vl_interface': 9.65741149835674e-06, 'learning_rate/action_model': 9.657411498356741e-05, 'epoch': 0.05}) +[2026-10-03 13:09:04,056][starVLA.training.train_starvla][INFO] - Step 697, Loss: {'train/loss': 0.00027596589643508196, 'train/grad_norm_pre_clip': 0.06234738230705261, 'timing/data': 0.00030329999572131783, 'timing/model': 0.39417333000164945, 'learning_rate/qwen_vl_interface': 9.656274989527663e-06, 'learning_rate/action_model': 9.656274989527664e-05, 'epoch': 0.05}) +[2026-10-03 13:09:04,762][starVLA.training.train_starvla][INFO] - Step 698, Loss: {'train/loss': 0.003063156269490719, 'train/grad_norm_pre_clip': 0.6922208666801453, 'timing/data': 0.0002734500012593344, 'timing/model': 0.39361034000467043, 'learning_rate/qwen_vl_interface': 9.655136669445378e-06, 'learning_rate/action_model': 9.655136669445378e-05, 'epoch': 0.05}) +[2026-10-03 13:09:05,466][starVLA.training.train_starvla][INFO] - Step 699, Loss: {'train/loss': 0.0028662304393947124, 'train/grad_norm_pre_clip': 0.5789371728897095, 'timing/data': 0.00037403000169433653, 'timing/model': 0.3927562799945008, 'learning_rate/qwen_vl_interface': 9.653996538577802e-06, 'learning_rate/action_model': 9.653996538577802e-05, 'epoch': 0.05}) +[2026-10-03 13:09:06,360][starVLA.training.train_starvla][INFO] - Step 700, Loss: {'train/loss': 0.004695852752774954, 'train/grad_norm_pre_clip': 8.877835273742676, 'timing/data': 0.00036963999446015805, 'timing/model': 0.3933194999990519, 'learning_rate/qwen_vl_interface': 9.652854597393604e-06, 'learning_rate/action_model': 9.652854597393603e-05, 'epoch': 0.05}) +[2026-10-03 13:09:07,066][starVLA.training.train_starvla][INFO] - Step 701, Loss: {'train/loss': 0.0001538258948130533, 'train/grad_norm_pre_clip': 0.019826989620923996, 'timing/data': 0.00032962999830488116, 'timing/model': 0.3943119700124953, 'learning_rate/qwen_vl_interface': 9.651710846362192e-06, 'learning_rate/action_model': 9.651710846362192e-05, 'epoch': 0.05}) +[2026-10-03 13:09:07,771][starVLA.training.train_starvla][INFO] - Step 702, Loss: {'train/loss': 0.23912911117076874, 'train/grad_norm_pre_clip': 89.41155242919922, 'timing/data': 0.00021363999985624105, 'timing/model': 0.3936830400052713, 'learning_rate/qwen_vl_interface': 9.650565285953716e-06, 'learning_rate/action_model': 9.650565285953715e-05, 'epoch': 0.05}) +[2026-10-03 13:09:08,473][starVLA.training.train_starvla][INFO] - Step 703, Loss: {'train/loss': 0.040898397564888, 'train/grad_norm_pre_clip': 111.27510070800781, 'timing/data': 0.00029897000058554113, 'timing/model': 0.3922415400011232, 'learning_rate/qwen_vl_interface': 9.649417916639077e-06, 'learning_rate/action_model': 9.649417916639076e-05, 'epoch': 0.05}) +[2026-10-03 13:09:09,176][starVLA.training.train_starvla][INFO] - Step 704, Loss: {'train/loss': 0.6594634056091309, 'train/grad_norm_pre_clip': 36.70830154418945, 'timing/data': 0.0002110700006596744, 'timing/model': 0.3923664199974155, 'learning_rate/qwen_vl_interface': 9.648268738889911e-06, 'learning_rate/action_model': 9.648268738889911e-05, 'epoch': 0.05}) +[2026-10-03 13:09:10,076][starVLA.training.train_starvla][INFO] - Step 705, Loss: {'train/loss': 0.5242299437522888, 'train/grad_norm_pre_clip': 166.21090698242188, 'timing/data': 0.00027939000574406236, 'timing/model': 0.39382686000317335, 'learning_rate/qwen_vl_interface': 9.647117753178604e-06, 'learning_rate/action_model': 9.647117753178604e-05, 'epoch': 0.05}) +[2026-10-03 13:09:10,781][starVLA.training.train_starvla][INFO] - Step 706, Loss: {'train/loss': 0.12525387108325958, 'train/grad_norm_pre_clip': 187.463134765625, 'timing/data': 0.00027491000946611166, 'timing/model': 0.39355691900709644, 'learning_rate/qwen_vl_interface': 9.645964959978284e-06, 'learning_rate/action_model': 9.645964959978284e-05, 'epoch': 0.05}) +[2026-10-03 13:09:11,484][starVLA.training.train_starvla][INFO] - Step 707, Loss: {'train/loss': 0.22888034582138062, 'train/grad_norm_pre_clip': 107.9213638305664, 'timing/data': 0.00026188000629190356, 'timing/model': 0.3923472009919351, 'learning_rate/qwen_vl_interface': 9.644810359762818e-06, 'learning_rate/action_model': 9.644810359762817e-05, 'epoch': 0.05}) +[2026-10-03 13:09:12,188][starVLA.training.train_starvla][INFO] - Step 708, Loss: {'train/loss': 0.780585527420044, 'train/grad_norm_pre_clip': 76.55833435058594, 'timing/data': 0.00021451999782584608, 'timing/model': 0.3926211300131399, 'learning_rate/qwen_vl_interface': 9.64365395300682e-06, 'learning_rate/action_model': 9.643653953006819e-05, 'epoch': 0.05}) +[2026-10-03 13:09:12,893][starVLA.training.train_starvla][INFO] - Step 709, Loss: {'train/loss': 0.6821243166923523, 'train/grad_norm_pre_clip': 239.7623748779297, 'timing/data': 0.0003339590039104223, 'timing/model': 0.3941834610013757, 'learning_rate/qwen_vl_interface': 9.642495740185644e-06, 'learning_rate/action_model': 9.642495740185643e-05, 'epoch': 0.05}) +[2026-10-03 13:09:13,788][starVLA.training.train_starvla][INFO] - Step 710, Loss: {'train/loss': 1.5265628099441528, 'train/grad_norm_pre_clip': 59.541255950927734, 'timing/data': 0.00027737999334931374, 'timing/model': 0.5844096049986547, 'learning_rate/qwen_vl_interface': 9.641335721775388e-06, 'learning_rate/action_model': 9.641335721775387e-05, 'epoch': 0.05}) +[2026-10-03 13:09:14,491][starVLA.training.train_starvla][INFO] - Step 711, Loss: {'train/loss': 0.3203858435153961, 'train/grad_norm_pre_clip': 24.957355499267578, 'timing/data': 0.00034189999860245734, 'timing/model': 0.39291275999858044, 'learning_rate/qwen_vl_interface': 9.640173898252893e-06, 'learning_rate/action_model': 9.640173898252892e-05, 'epoch': 0.05}) +[2026-10-03 13:09:15,195][starVLA.training.train_starvla][INFO] - Step 712, Loss: {'train/loss': 0.0006814088555984199, 'train/grad_norm_pre_clip': 0.6214829683303833, 'timing/data': 0.0003347100137034431, 'timing/model': 0.39286001000436954, 'learning_rate/qwen_vl_interface': 9.639010270095738e-06, 'learning_rate/action_model': 9.639010270095739e-05, 'epoch': 0.05}) +[2026-10-03 13:09:15,902][starVLA.training.train_starvla][INFO] - Step 713, Loss: {'train/loss': 0.28926634788513184, 'train/grad_norm_pre_clip': 48.590152740478516, 'timing/data': 0.00027936999686062336, 'timing/model': 0.39436345000285655, 'learning_rate/qwen_vl_interface': 9.637844837782251e-06, 'learning_rate/action_model': 9.63784483778225e-05, 'epoch': 0.05}) +[2026-10-03 13:09:16,606][starVLA.training.train_starvla][INFO] - Step 714, Loss: {'train/loss': 0.6140905022621155, 'train/grad_norm_pre_clip': 171.44760131835938, 'timing/data': 0.000387310006772168, 'timing/model': 0.3919291509955656, 'learning_rate/qwen_vl_interface': 9.636677601791494e-06, 'learning_rate/action_model': 9.636677601791493e-05, 'epoch': 0.05}) +[2026-10-03 13:09:17,309][starVLA.training.train_starvla][INFO] - Step 715, Loss: {'train/loss': 0.836858332157135, 'train/grad_norm_pre_clip': 67.42716217041016, 'timing/data': 0.00034372998925391585, 'timing/model': 0.3923061899986351, 'learning_rate/qwen_vl_interface': 9.635508562603273e-06, 'learning_rate/action_model': 9.635508562603272e-05, 'epoch': 0.05}) +[2026-10-03 13:09:18,202][starVLA.training.train_starvla][INFO] - Step 716, Loss: {'train/loss': 0.23821324110031128, 'train/grad_norm_pre_clip': 77.55957794189453, 'timing/data': 0.0003397899999981746, 'timing/model': 0.3922100009949645, 'learning_rate/qwen_vl_interface': 9.634337720698139e-06, 'learning_rate/action_model': 9.634337720698137e-05, 'epoch': 0.05}) +[2026-10-03 13:09:18,910][starVLA.training.train_starvla][INFO] - Step 717, Loss: {'train/loss': 0.23933741450309753, 'train/grad_norm_pre_clip': 23.489437103271484, 'timing/data': 0.00031586999830324203, 'timing/model': 0.39487145000020973, 'learning_rate/qwen_vl_interface': 9.633165076557376e-06, 'learning_rate/action_model': 9.633165076557375e-05, 'epoch': 0.05}) +[2026-10-03 13:09:19,614][starVLA.training.train_starvla][INFO] - Step 718, Loss: {'train/loss': 0.09614516794681549, 'train/grad_norm_pre_clip': 53.336341857910156, 'timing/data': 0.00028341999859549105, 'timing/model': 0.39382231001218315, 'learning_rate/qwen_vl_interface': 9.631990630663019e-06, 'learning_rate/action_model': 9.631990630663018e-05, 'epoch': 0.05}) +[2026-10-03 13:09:20,319][starVLA.training.train_starvla][INFO] - Step 719, Loss: {'train/loss': 0.15155819058418274, 'train/grad_norm_pre_clip': 47.656890869140625, 'timing/data': 0.0002364999963901937, 'timing/model': 0.39336667099269107, 'learning_rate/qwen_vl_interface': 9.630814383497836e-06, 'learning_rate/action_model': 9.630814383497835e-05, 'epoch': 0.05}) +[2026-10-03 13:09:21,023][starVLA.training.train_starvla][INFO] - Step 720, Loss: {'train/loss': 0.006085851229727268, 'train/grad_norm_pre_clip': 4.339569091796875, 'timing/data': 0.00034579999919515103, 'timing/model': 0.39300527100567706, 'learning_rate/qwen_vl_interface': 9.629636335545336e-06, 'learning_rate/action_model': 9.629636335545335e-05, 'epoch': 0.05}) +[2026-10-03 13:09:21,927][starVLA.training.train_starvla][INFO] - Step 721, Loss: {'train/loss': 0.055728763341903687, 'train/grad_norm_pre_clip': 9.368812561035156, 'timing/data': 0.00029176998941693455, 'timing/model': 0.3951582099980442, 'learning_rate/qwen_vl_interface': 9.628456487289773e-06, 'learning_rate/action_model': 9.628456487289773e-05, 'epoch': 0.05}) +[2026-10-03 13:09:22,635][starVLA.training.train_starvla][INFO] - Step 722, Loss: {'train/loss': 0.0462738536298275, 'train/grad_norm_pre_clip': 9.417731285095215, 'timing/data': 0.0003052700049011037, 'timing/model': 0.3949074800038943, 'learning_rate/qwen_vl_interface': 9.627274839216135e-06, 'learning_rate/action_model': 9.627274839216136e-05, 'epoch': 0.05}) +[2026-10-03 13:09:23,343][starVLA.training.train_starvla][INFO] - Step 723, Loss: {'train/loss': 0.6679099798202515, 'train/grad_norm_pre_clip': 107.13892364501953, 'timing/data': 0.0003030299994861707, 'timing/model': 0.3956235499936156, 'learning_rate/qwen_vl_interface': 9.626091391810158e-06, 'learning_rate/action_model': 9.626091391810157e-05, 'epoch': 0.05}) +[2026-10-03 13:09:24,050][starVLA.training.train_starvla][INFO] - Step 724, Loss: {'train/loss': 0.4061465859413147, 'train/grad_norm_pre_clip': 20.194202423095703, 'timing/data': 0.00027801000396721065, 'timing/model': 0.39403214999765623, 'learning_rate/qwen_vl_interface': 9.624906145558309e-06, 'learning_rate/action_model': 9.624906145558307e-05, 'epoch': 0.05}) +[2026-10-03 13:09:24,758][starVLA.training.train_starvla][INFO] - Step 725, Loss: {'train/loss': 0.3422665596008301, 'train/grad_norm_pre_clip': 39.342262268066406, 'timing/data': 0.00025504000950604677, 'timing/model': 0.3950837500015041, 'learning_rate/qwen_vl_interface': 9.623719100947795e-06, 'learning_rate/action_model': 9.623719100947795e-05, 'epoch': 0.05}) +[2026-10-03 13:09:25,652][starVLA.training.train_starvla][INFO] - Step 726, Loss: {'train/loss': 0.4374716281890869, 'train/grad_norm_pre_clip': 24.351398468017578, 'timing/data': 0.0002560599969001487, 'timing/model': 0.5826660360035021, 'learning_rate/qwen_vl_interface': 9.622530258466571e-06, 'learning_rate/action_model': 9.62253025846657e-05, 'epoch': 0.05}) +[2026-10-03 13:09:26,358][starVLA.training.train_starvla][INFO] - Step 727, Loss: {'train/loss': 0.26446935534477234, 'train/grad_norm_pre_clip': 101.6302490234375, 'timing/data': 0.00020568999752867967, 'timing/model': 0.39420749999408145, 'learning_rate/qwen_vl_interface': 9.62133961860332e-06, 'learning_rate/action_model': 9.62133961860332e-05, 'epoch': 0.05}) +[2026-10-03 13:09:27,063][starVLA.training.train_starvla][INFO] - Step 728, Loss: {'train/loss': 0.05562338978052139, 'train/grad_norm_pre_clip': 26.66193962097168, 'timing/data': 0.00026951001200359315, 'timing/model': 0.39285452099284157, 'learning_rate/qwen_vl_interface': 9.620147181847474e-06, 'learning_rate/action_model': 9.620147181847475e-05, 'epoch': 0.05}) +[2026-10-03 13:09:27,770][starVLA.training.train_starvla][INFO] - Step 729, Loss: {'train/loss': 0.013674397021532059, 'train/grad_norm_pre_clip': 1.4878333806991577, 'timing/data': 0.0003183799999533221, 'timing/model': 0.3946790000045439, 'learning_rate/qwen_vl_interface': 9.618952948689196e-06, 'learning_rate/action_model': 9.618952948689196e-05, 'epoch': 0.05}) +[2026-10-03 13:09:28,476][starVLA.training.train_starvla][INFO] - Step 730, Loss: {'train/loss': 0.3341711163520813, 'train/grad_norm_pre_clip': 54.271034240722656, 'timing/data': 0.00025601001107133925, 'timing/model': 0.39474216000235174, 'learning_rate/qwen_vl_interface': 9.617756919619393e-06, 'learning_rate/action_model': 9.617756919619392e-05, 'epoch': 0.05}) +[2026-10-03 13:09:29,371][starVLA.training.train_starvla][INFO] - Step 731, Loss: {'train/loss': 0.11191579699516296, 'train/grad_norm_pre_clip': 39.60124206542969, 'timing/data': 0.0003656000044429675, 'timing/model': 0.5820672260015272, 'learning_rate/qwen_vl_interface': 9.616559095129701e-06, 'learning_rate/action_model': 9.6165590951297e-05, 'epoch': 0.05}) +[2026-10-03 13:09:30,076][starVLA.training.train_starvla][INFO] - Step 732, Loss: {'train/loss': 0.16728541254997253, 'train/grad_norm_pre_clip': 60.129756927490234, 'timing/data': 0.0002109100023517385, 'timing/model': 0.39361900999210775, 'learning_rate/qwen_vl_interface': 9.615359475712506e-06, 'learning_rate/action_model': 9.615359475712506e-05, 'epoch': 0.05}) +[2026-10-03 13:09:30,783][starVLA.training.train_starvla][INFO] - Step 733, Loss: {'train/loss': 0.44652342796325684, 'train/grad_norm_pre_clip': 98.90068817138672, 'timing/data': 0.0003632800071500242, 'timing/model': 0.3943823500012513, 'learning_rate/qwen_vl_interface': 9.614158061860924e-06, 'learning_rate/action_model': 9.614158061860924e-05, 'epoch': 0.05}) +[2026-10-03 13:09:31,490][starVLA.training.train_starvla][INFO] - Step 734, Loss: {'train/loss': 0.6178373694419861, 'train/grad_norm_pre_clip': 58.253726959228516, 'timing/data': 0.00028020000900141895, 'timing/model': 0.39551235000544693, 'learning_rate/qwen_vl_interface': 9.612954854068813e-06, 'learning_rate/action_model': 9.612954854068812e-05, 'epoch': 0.05}) +[2026-10-03 13:09:32,197][starVLA.training.train_starvla][INFO] - Step 735, Loss: {'train/loss': 0.1192731186747551, 'train/grad_norm_pre_clip': 138.9243621826172, 'timing/data': 0.0003301699907751754, 'timing/model': 0.3935402310016798, 'learning_rate/qwen_vl_interface': 9.611749852830763e-06, 'learning_rate/action_model': 9.611749852830763e-05, 'epoch': 0.05}) +[2026-10-03 13:09:33,093][starVLA.training.train_starvla][INFO] - Step 736, Loss: {'train/loss': 0.019352464005351067, 'train/grad_norm_pre_clip': 26.596818923950195, 'timing/data': 0.00030652999703306705, 'timing/model': 0.5836687370028812, 'learning_rate/qwen_vl_interface': 9.610543058642107e-06, 'learning_rate/action_model': 9.610543058642107e-05, 'epoch': 0.05}) +[2026-10-03 13:09:33,800][starVLA.training.train_starvla][INFO] - Step 737, Loss: {'train/loss': 0.4682610332965851, 'train/grad_norm_pre_clip': 48.4334716796875, 'timing/data': 0.00036114999966230243, 'timing/model': 0.3942722100036917, 'learning_rate/qwen_vl_interface': 9.609334471998911e-06, 'learning_rate/action_model': 9.609334471998912e-05, 'epoch': 0.05}) +[2026-10-03 13:09:34,506][starVLA.training.train_starvla][INFO] - Step 738, Loss: {'train/loss': 0.049788981676101685, 'train/grad_norm_pre_clip': 35.97298812866211, 'timing/data': 0.0003293599875178188, 'timing/model': 0.394652410002891, 'learning_rate/qwen_vl_interface': 9.60812409339798e-06, 'learning_rate/action_model': 9.60812409339798e-05, 'epoch': 0.05}) +[2026-10-03 13:09:35,212][starVLA.training.train_starvla][INFO] - Step 739, Loss: {'train/loss': 0.5795807838439941, 'train/grad_norm_pre_clip': 90.79000091552734, 'timing/data': 0.00032885999826248735, 'timing/model': 0.39427854999667034, 'learning_rate/qwen_vl_interface': 9.606911923336856e-06, 'learning_rate/action_model': 9.606911923336856e-05, 'epoch': 0.05}) +[2026-10-03 13:09:35,917][starVLA.training.train_starvla][INFO] - Step 740, Loss: {'train/loss': 0.3119453489780426, 'train/grad_norm_pre_clip': 35.67422103881836, 'timing/data': 0.0002338699996471405, 'timing/model': 0.3931830800138414, 'learning_rate/qwen_vl_interface': 9.605697962313817e-06, 'learning_rate/action_model': 9.605697962313816e-05, 'epoch': 0.05}) +[2026-10-03 13:09:36,624][starVLA.training.train_starvla][INFO] - Step 741, Loss: {'train/loss': 0.04180660471320152, 'train/grad_norm_pre_clip': 6.785839080810547, 'timing/data': 0.00028423000185284764, 'timing/model': 0.39439558000594843, 'learning_rate/qwen_vl_interface': 9.604482210827871e-06, 'learning_rate/action_model': 9.604482210827871e-05, 'epoch': 0.05}) +[2026-10-03 13:09:37,529][starVLA.training.train_starvla][INFO] - Step 742, Loss: {'train/loss': 0.5411168336868286, 'train/grad_norm_pre_clip': 87.76624298095703, 'timing/data': 0.0002250099933007732, 'timing/model': 0.3953301110013854, 'learning_rate/qwen_vl_interface': 9.603264669378772e-06, 'learning_rate/action_model': 9.603264669378772e-05, 'epoch': 0.05}) +[2026-10-03 13:09:38,234][starVLA.training.train_starvla][INFO] - Step 743, Loss: {'train/loss': 0.001209949143230915, 'train/grad_norm_pre_clip': 0.30559518933296204, 'timing/data': 0.00032310000096913427, 'timing/model': 0.39351649100717623, 'learning_rate/qwen_vl_interface': 9.602045338467007e-06, 'learning_rate/action_model': 9.602045338467006e-05, 'epoch': 0.05}) +[2026-10-03 13:09:38,939][starVLA.training.train_starvla][INFO] - Step 744, Loss: {'train/loss': 0.09124954789876938, 'train/grad_norm_pre_clip': 46.5946159362793, 'timing/data': 0.00030803999106865376, 'timing/model': 0.3929809699911857, 'learning_rate/qwen_vl_interface': 9.600824218593791e-06, 'learning_rate/action_model': 9.600824218593791e-05, 'epoch': 0.05}) +[2026-10-03 13:09:39,647][starVLA.training.train_starvla][INFO] - Step 745, Loss: {'train/loss': 0.18227936327457428, 'train/grad_norm_pre_clip': 36.966121673583984, 'timing/data': 0.0002522300055716187, 'timing/model': 0.3945799300126964, 'learning_rate/qwen_vl_interface': 9.599601310261087e-06, 'learning_rate/action_model': 9.599601310261085e-05, 'epoch': 0.05}) +[2026-10-03 13:09:40,353][starVLA.training.train_starvla][INFO] - Step 746, Loss: {'train/loss': 0.2498008757829666, 'train/grad_norm_pre_clip': 40.55952072143555, 'timing/data': 0.0003227400011382997, 'timing/model': 0.3945336999895517, 'learning_rate/qwen_vl_interface': 9.598376613971582e-06, 'learning_rate/action_model': 9.598376613971582e-05, 'epoch': 0.05}) +[2026-10-03 13:09:41,249][starVLA.training.train_starvla][INFO] - Step 747, Loss: {'train/loss': 0.6554428935050964, 'train/grad_norm_pre_clip': 46.96144485473633, 'timing/data': 0.00033429000177420676, 'timing/model': 0.5840617260109866, 'learning_rate/qwen_vl_interface': 9.597150130228703e-06, 'learning_rate/action_model': 9.597150130228703e-05, 'epoch': 0.05}) +[2026-10-03 13:09:41,955][starVLA.training.train_starvla][INFO] - Step 748, Loss: {'train/loss': 0.015092424117028713, 'train/grad_norm_pre_clip': 3.588593006134033, 'timing/data': 0.00037434000114444643, 'timing/model': 0.39330850100668613, 'learning_rate/qwen_vl_interface': 9.595921859536612e-06, 'learning_rate/action_model': 9.595921859536611e-05, 'epoch': 0.05}) +[2026-10-03 13:09:42,662][starVLA.training.train_starvla][INFO] - Step 749, Loss: {'train/loss': 0.006722829304635525, 'train/grad_norm_pre_clip': 1.1493607759475708, 'timing/data': 0.00028355998801998794, 'timing/model': 0.39476793000358157, 'learning_rate/qwen_vl_interface': 9.594691802400203e-06, 'learning_rate/action_model': 9.594691802400202e-05, 'epoch': 0.05}) +[2026-10-03 13:09:43,369][starVLA.training.train_starvla][INFO] - Step 750, Loss: {'train/loss': 0.5114864110946655, 'train/grad_norm_pre_clip': 73.33602905273438, 'timing/data': 0.0003599199990276247, 'timing/model': 0.3948352099978365, 'learning_rate/qwen_vl_interface': 9.59345995932511e-06, 'learning_rate/action_model': 9.59345995932511e-05, 'epoch': 0.05}) +[2026-10-03 13:09:44,075][starVLA.training.train_starvla][INFO] - Step 751, Loss: {'train/loss': 0.03405172377824783, 'train/grad_norm_pre_clip': 14.858550071716309, 'timing/data': 0.00030872999923303723, 'timing/model': 0.3937914199923398, 'learning_rate/qwen_vl_interface': 9.592226330817693e-06, 'learning_rate/action_model': 9.592226330817693e-05, 'epoch': 0.05}) +[2026-10-03 13:09:44,779][starVLA.training.train_starvla][INFO] - Step 752, Loss: {'train/loss': 0.9801026582717896, 'train/grad_norm_pre_clip': 75.83021545410156, 'timing/data': 0.00033460999839007854, 'timing/model': 0.3932801710034255, 'learning_rate/qwen_vl_interface': 9.590990917385052e-06, 'learning_rate/action_model': 9.590990917385052e-05, 'epoch': 0.06}) +[2026-10-03 13:09:45,682][starVLA.training.train_starvla][INFO] - Step 753, Loss: {'train/loss': 0.8511709570884705, 'train/grad_norm_pre_clip': 18.216121673583984, 'timing/data': 0.0002134500100510195, 'timing/model': 0.3946036810084479, 'learning_rate/qwen_vl_interface': 9.58975371953502e-06, 'learning_rate/action_model': 9.589753719535018e-05, 'epoch': 0.06}) +[2026-10-03 13:09:46,389][starVLA.training.train_starvla][INFO] - Step 754, Loss: {'train/loss': 0.3636306822299957, 'train/grad_norm_pre_clip': 12.416455268859863, 'timing/data': 0.00026201999571640044, 'timing/model': 0.39477099099894986, 'learning_rate/qwen_vl_interface': 9.588514737776158e-06, 'learning_rate/action_model': 9.588514737776157e-05, 'epoch': 0.06}) +[2026-10-03 13:09:47,094][starVLA.training.train_starvla][INFO] - Step 755, Loss: {'train/loss': 0.0027795173227787018, 'train/grad_norm_pre_clip': 0.5601881742477417, 'timing/data': 0.0002833399921655655, 'timing/model': 0.3937441899906844, 'learning_rate/qwen_vl_interface': 9.587273972617769e-06, 'learning_rate/action_model': 9.587273972617768e-05, 'epoch': 0.06}) +[2026-10-03 13:09:47,799][starVLA.training.train_starvla][INFO] - Step 756, Loss: {'train/loss': 0.011825408786535263, 'train/grad_norm_pre_clip': 3.7500367164611816, 'timing/data': 0.0002544599992688745, 'timing/model': 0.39344631000130903, 'learning_rate/qwen_vl_interface': 9.586031424569882e-06, 'learning_rate/action_model': 9.58603142456988e-05, 'epoch': 0.06}) +[2026-10-03 13:09:48,505][starVLA.training.train_starvla][INFO] - Step 757, Loss: {'train/loss': 0.021152794361114502, 'train/grad_norm_pre_clip': 17.628461837768555, 'timing/data': 0.0002865999995265156, 'timing/model': 0.39415609098796267, 'learning_rate/qwen_vl_interface': 9.584787094143265e-06, 'learning_rate/action_model': 9.584787094143264e-05, 'epoch': 0.06}) +[2026-10-03 13:09:49,406][starVLA.training.train_starvla][INFO] - Step 758, Loss: {'train/loss': 0.008331254124641418, 'train/grad_norm_pre_clip': 4.466497898101807, 'timing/data': 0.00027925999893341213, 'timing/model': 0.3949767099984456, 'learning_rate/qwen_vl_interface': 9.58354098184941e-06, 'learning_rate/action_model': 9.583540981849409e-05, 'epoch': 0.06}) +[2026-10-03 13:09:50,112][starVLA.training.train_starvla][INFO] - Step 759, Loss: {'train/loss': 0.005745799746364355, 'train/grad_norm_pre_clip': 3.091681957244873, 'timing/data': 0.00020791999122593552, 'timing/model': 0.3940406610054197, 'learning_rate/qwen_vl_interface': 9.582293088200549e-06, 'learning_rate/action_model': 9.582293088200549e-05, 'epoch': 0.06}) +[2026-10-03 13:09:50,817][starVLA.training.train_starvla][INFO] - Step 760, Loss: {'train/loss': 0.1781018227338791, 'train/grad_norm_pre_clip': 23.923507690429688, 'timing/data': 0.0002114000089932233, 'timing/model': 0.39358151001215447, 'learning_rate/qwen_vl_interface': 9.581043413709646e-06, 'learning_rate/action_model': 9.581043413709646e-05, 'epoch': 0.06}) +[2026-10-03 13:09:51,524][starVLA.training.train_starvla][INFO] - Step 761, Loss: {'train/loss': 0.35143527388572693, 'train/grad_norm_pre_clip': 44.50621032714844, 'timing/data': 0.00031218001095112413, 'timing/model': 0.39458895100688096, 'learning_rate/qwen_vl_interface': 9.579791958890393e-06, 'learning_rate/action_model': 9.579791958890394e-05, 'epoch': 0.06}) +[2026-10-03 13:09:52,231][starVLA.training.train_starvla][INFO] - Step 762, Loss: {'train/loss': 0.09070345759391785, 'train/grad_norm_pre_clip': 193.1263427734375, 'timing/data': 0.00021833000937476754, 'timing/model': 0.3946134799916763, 'learning_rate/qwen_vl_interface': 9.578538724257218e-06, 'learning_rate/action_model': 9.578538724257216e-05, 'epoch': 0.06}) +[2026-10-03 13:09:53,129][starVLA.training.train_starvla][INFO] - Step 763, Loss: {'train/loss': 0.012649010866880417, 'train/grad_norm_pre_clip': 2.361287832260132, 'timing/data': 0.00030906000756658614, 'timing/model': 0.5866448559972923, 'learning_rate/qwen_vl_interface': 9.577283710325274e-06, 'learning_rate/action_model': 9.577283710325274e-05, 'epoch': 0.06}) +[2026-10-03 13:09:53,835][starVLA.training.train_starvla][INFO] - Step 764, Loss: {'train/loss': 0.8291674852371216, 'train/grad_norm_pre_clip': 144.75927734375, 'timing/data': 0.0002668899978743866, 'timing/model': 0.39351774100214243, 'learning_rate/qwen_vl_interface': 9.576026917610455e-06, 'learning_rate/action_model': 9.576026917610455e-05, 'epoch': 0.06}) +[2026-10-03 13:09:54,541][starVLA.training.train_starvla][INFO] - Step 765, Loss: {'train/loss': 0.4729265868663788, 'train/grad_norm_pre_clip': 37.41115951538086, 'timing/data': 0.0002878200029954314, 'timing/model': 0.394051559997024, 'learning_rate/qwen_vl_interface': 9.574768346629376e-06, 'learning_rate/action_model': 9.574768346629376e-05, 'epoch': 0.06}) +[2026-10-03 13:09:55,248][starVLA.training.train_starvla][INFO] - Step 766, Loss: {'train/loss': 0.002020995132625103, 'train/grad_norm_pre_clip': 0.4138321578502655, 'timing/data': 0.00024463000590913, 'timing/model': 0.3950650300103007, 'learning_rate/qwen_vl_interface': 9.573507997899391e-06, 'learning_rate/action_model': 9.573507997899392e-05, 'epoch': 0.06}) +[2026-10-03 13:09:55,949][starVLA.training.train_starvla][INFO] - Step 767, Loss: {'train/loss': 0.043686602264642715, 'train/grad_norm_pre_clip': 37.57912826538086, 'timing/data': 0.000209210003959015, 'timing/model': 0.3913597720093094, 'learning_rate/qwen_vl_interface': 9.572245871938582e-06, 'learning_rate/action_model': 9.572245871938582e-05, 'epoch': 0.06}) +[2026-10-03 13:09:56,646][starVLA.training.train_starvla][INFO] - Step 768, Loss: {'train/loss': 0.32250261306762695, 'train/grad_norm_pre_clip': 64.60799407958984, 'timing/data': 0.0003729399904841557, 'timing/model': 0.3894871009979397, 'learning_rate/qwen_vl_interface': 9.570981969265761e-06, 'learning_rate/action_model': 9.570981969265761e-05, 'epoch': 0.06}) +[2026-10-03 13:09:57,532][starVLA.training.train_starvla][INFO] - Step 769, Loss: {'train/loss': 0.16595807671546936, 'train/grad_norm_pre_clip': 152.03981018066406, 'timing/data': 0.00025659000675659627, 'timing/model': 0.3895302309974795, 'learning_rate/qwen_vl_interface': 9.56971629040047e-06, 'learning_rate/action_model': 9.56971629040047e-05, 'epoch': 0.06}) +[2026-10-03 13:09:58,230][starVLA.training.train_starvla][INFO] - Step 770, Loss: {'train/loss': 0.22980906069278717, 'train/grad_norm_pre_clip': 30.572750091552734, 'timing/data': 0.00030112999957054853, 'timing/model': 0.39103137199708726, 'learning_rate/qwen_vl_interface': 9.56844883586298e-06, 'learning_rate/action_model': 9.568448835862981e-05, 'epoch': 0.06}) +[2026-10-03 13:09:58,930][starVLA.training.train_starvla][INFO] - Step 771, Loss: {'train/loss': 0.0025938008911907673, 'train/grad_norm_pre_clip': 0.38850507140159607, 'timing/data': 0.0002564699971117079, 'timing/model': 0.3906159009929979, 'learning_rate/qwen_vl_interface': 9.567179606174299e-06, 'learning_rate/action_model': 9.567179606174299e-05, 'epoch': 0.06}) +[2026-10-03 13:09:59,626][starVLA.training.train_starvla][INFO] - Step 772, Loss: {'train/loss': 0.0063203005120158195, 'train/grad_norm_pre_clip': 3.7277300357818604, 'timing/data': 0.00024669000413268805, 'timing/model': 0.38963948099990375, 'learning_rate/qwen_vl_interface': 9.565908601856155e-06, 'learning_rate/action_model': 9.565908601856154e-05, 'epoch': 0.06}) +[2026-10-03 13:10:00,322][starVLA.training.train_starvla][INFO] - Step 773, Loss: {'train/loss': 0.008824024349451065, 'train/grad_norm_pre_clip': 4.272238731384277, 'timing/data': 0.0002290599950356409, 'timing/model': 0.3890797209896846, 'learning_rate/qwen_vl_interface': 9.56463582343101e-06, 'learning_rate/action_model': 9.56463582343101e-05, 'epoch': 0.06}) +[2026-10-03 13:10:01,216][starVLA.training.train_starvla][INFO] - Step 774, Loss: {'train/loss': 0.11462701112031937, 'train/grad_norm_pre_clip': 25.017044067382812, 'timing/data': 0.00019796000560745597, 'timing/model': 0.3910131409938913, 'learning_rate/qwen_vl_interface': 9.563361271422056e-06, 'learning_rate/action_model': 9.563361271422056e-05, 'epoch': 0.06}) +[2026-10-03 13:10:01,917][starVLA.training.train_starvla][INFO] - Step 775, Loss: {'train/loss': 0.4256116449832916, 'train/grad_norm_pre_clip': 14.69959831237793, 'timing/data': 0.0002967489999718964, 'timing/model': 0.3918240920029348, 'learning_rate/qwen_vl_interface': 9.562084946353212e-06, 'learning_rate/action_model': 9.562084946353213e-05, 'epoch': 0.06}) +[2026-10-03 13:10:02,614][starVLA.training.train_starvla][INFO] - Step 776, Loss: {'train/loss': 0.007041840348392725, 'train/grad_norm_pre_clip': 3.4617748260498047, 'timing/data': 0.0001929499994730577, 'timing/model': 0.38975130100152455, 'learning_rate/qwen_vl_interface': 9.560806848749132e-06, 'learning_rate/action_model': 9.560806848749132e-05, 'epoch': 0.06}) +[2026-10-03 13:10:03,308][starVLA.training.train_starvla][INFO] - Step 777, Loss: {'train/loss': 0.0010173748014494777, 'train/grad_norm_pre_clip': 0.5010452270507812, 'timing/data': 0.0003571400011423975, 'timing/model': 0.38891581099596806, 'learning_rate/qwen_vl_interface': 9.559526979135186e-06, 'learning_rate/action_model': 9.559526979135187e-05, 'epoch': 0.06}) +[2026-10-03 13:10:04,005][starVLA.training.train_starvla][INFO] - Step 778, Loss: {'train/loss': 0.002039851387962699, 'train/grad_norm_pre_clip': 1.2766140699386597, 'timing/data': 0.00024632998974993825, 'timing/model': 0.3907340209989343, 'learning_rate/qwen_vl_interface': 9.558245338037486e-06, 'learning_rate/action_model': 9.558245338037486e-05, 'epoch': 0.06}) +[2026-10-03 13:10:04,896][starVLA.training.train_starvla][INFO] - Step 779, Loss: {'train/loss': 0.9189116954803467, 'train/grad_norm_pre_clip': 42.008941650390625, 'timing/data': 0.00027239999326411635, 'timing/model': 0.5828261069982545, 'learning_rate/qwen_vl_interface': 9.556961925982864e-06, 'learning_rate/action_model': 9.556961925982863e-05, 'epoch': 0.06}) +[2026-10-03 13:10:05,591][starVLA.training.train_starvla][INFO] - Step 780, Loss: {'train/loss': 0.22902631759643555, 'train/grad_norm_pre_clip': 64.23523712158203, 'timing/data': 0.00031382999441120774, 'timing/model': 0.38872330100275576, 'learning_rate/qwen_vl_interface': 9.555676743498882e-06, 'learning_rate/action_model': 9.555676743498881e-05, 'epoch': 0.06}) +[2026-10-03 13:10:06,285][starVLA.training.train_starvla][INFO] - Step 781, Loss: {'train/loss': 0.0024088078644126654, 'train/grad_norm_pre_clip': 3.3383326530456543, 'timing/data': 0.00031359998683910817, 'timing/model': 0.38854219199856743, 'learning_rate/qwen_vl_interface': 9.55438979111383e-06, 'learning_rate/action_model': 9.554389791113831e-05, 'epoch': 0.06}) +[2026-10-03 13:10:06,984][starVLA.training.train_starvla][INFO] - Step 782, Loss: {'train/loss': 0.0005721450434066355, 'train/grad_norm_pre_clip': 0.09820510447025299, 'timing/data': 0.000284630004898645, 'timing/model': 0.3913004509959137, 'learning_rate/qwen_vl_interface': 9.553101069356728e-06, 'learning_rate/action_model': 9.553101069356727e-05, 'epoch': 0.06}) +[2026-10-03 13:10:07,684][starVLA.training.train_starvla][INFO] - Step 783, Loss: {'train/loss': 0.006096866447478533, 'train/grad_norm_pre_clip': 0.998725414276123, 'timing/data': 0.00019841999164782465, 'timing/model': 0.3914240919984877, 'learning_rate/qwen_vl_interface': 9.551810578757316e-06, 'learning_rate/action_model': 9.551810578757315e-05, 'epoch': 0.06}) +[2026-10-03 13:10:08,573][starVLA.training.train_starvla][INFO] - Step 784, Loss: {'train/loss': 1.1809868812561035, 'train/grad_norm_pre_clip': 201.69508361816406, 'timing/data': 0.0002611499949125573, 'timing/model': 0.581957076996332, 'learning_rate/qwen_vl_interface': 9.550518319846069e-06, 'learning_rate/action_model': 9.550518319846069e-05, 'epoch': 0.06}) +[2026-10-03 13:10:09,267][starVLA.training.train_starvla][INFO] - Step 785, Loss: {'train/loss': 0.0068010250106453896, 'train/grad_norm_pre_clip': 1.1903265714645386, 'timing/data': 0.00024494998797308654, 'timing/model': 0.3886140520044137, 'learning_rate/qwen_vl_interface': 9.549224293154187e-06, 'learning_rate/action_model': 9.549224293154187e-05, 'epoch': 0.06}) +[2026-10-03 13:10:09,964][starVLA.training.train_starvla][INFO] - Step 786, Loss: {'train/loss': 0.42562636733055115, 'train/grad_norm_pre_clip': 84.07252502441406, 'timing/data': 0.00027129999944008887, 'timing/model': 0.3904324210016057, 'learning_rate/qwen_vl_interface': 9.547928499213593e-06, 'learning_rate/action_model': 9.547928499213593e-05, 'epoch': 0.06}) +[2026-10-03 13:10:10,664][starVLA.training.train_starvla][INFO] - Step 787, Loss: {'train/loss': 1.1332836151123047, 'train/grad_norm_pre_clip': 177.68777465820312, 'timing/data': 0.00028080999618396163, 'timing/model': 0.3915730809967499, 'learning_rate/qwen_vl_interface': 9.54663093855694e-06, 'learning_rate/action_model': 9.54663093855694e-05, 'epoch': 0.06}) +[2026-10-03 13:10:11,358][starVLA.training.train_starvla][INFO] - Step 788, Loss: {'train/loss': 0.720883846282959, 'train/grad_norm_pre_clip': 100.49279022216797, 'timing/data': 0.0003638499911176041, 'timing/model': 0.38809566100826487, 'learning_rate/qwen_vl_interface': 9.545331611717607e-06, 'learning_rate/action_model': 9.545331611717608e-05, 'epoch': 0.06}) +[2026-10-03 13:10:12,250][starVLA.training.train_starvla][INFO] - Step 789, Loss: {'train/loss': 0.6072556376457214, 'train/grad_norm_pre_clip': 41.7254524230957, 'timing/data': 0.00028099000337533653, 'timing/model': 0.5858708359883167, 'learning_rate/qwen_vl_interface': 9.544030519229697e-06, 'learning_rate/action_model': 9.544030519229697e-05, 'epoch': 0.06}) +[2026-10-03 13:10:12,949][starVLA.training.train_starvla][INFO] - Step 790, Loss: {'train/loss': 0.005247069057077169, 'train/grad_norm_pre_clip': 11.357043266296387, 'timing/data': 0.00031802000012248755, 'timing/model': 0.3908395619946532, 'learning_rate/qwen_vl_interface': 9.54272766162804e-06, 'learning_rate/action_model': 9.54272766162804e-05, 'epoch': 0.06}) +[2026-10-03 13:10:13,651][starVLA.training.train_starvla][INFO] - Step 791, Loss: {'train/loss': 0.34969061613082886, 'train/grad_norm_pre_clip': 48.229888916015625, 'timing/data': 0.00023524000425823033, 'timing/model': 0.3905684010096593, 'learning_rate/qwen_vl_interface': 9.54142303944819e-06, 'learning_rate/action_model': 9.541423039448191e-05, 'epoch': 0.06}) +[2026-10-03 13:10:14,345][starVLA.training.train_starvla][INFO] - Step 792, Loss: {'train/loss': 0.21372553706169128, 'train/grad_norm_pre_clip': 132.72006225585938, 'timing/data': 0.00022430000535678118, 'timing/model': 0.38839440199080855, 'learning_rate/qwen_vl_interface': 9.540116653226433e-06, 'learning_rate/action_model': 9.540116653226431e-05, 'epoch': 0.06}) +[2026-10-03 13:10:15,040][starVLA.training.train_starvla][INFO] - Step 793, Loss: {'train/loss': 0.32703715562820435, 'train/grad_norm_pre_clip': 34.84861755371094, 'timing/data': 0.00019914998847525567, 'timing/model': 0.38867294100055005, 'learning_rate/qwen_vl_interface': 9.53880850349977e-06, 'learning_rate/action_model': 9.538808503499768e-05, 'epoch': 0.06}) +[2026-10-03 13:10:15,737][starVLA.training.train_starvla][INFO] - Step 794, Loss: {'train/loss': 0.27914464473724365, 'train/grad_norm_pre_clip': 34.86708450317383, 'timing/data': 0.00025938000180758536, 'timing/model': 0.3907816110004205, 'learning_rate/qwen_vl_interface': 9.537498590805934e-06, 'learning_rate/action_model': 9.537498590805933e-05, 'epoch': 0.06}) +[2026-10-03 13:10:16,630][starVLA.training.train_starvla][INFO] - Step 795, Loss: {'train/loss': 0.4522332549095154, 'train/grad_norm_pre_clip': 10.875783920288086, 'timing/data': 0.00033167999936267734, 'timing/model': 0.38910953199956566, 'learning_rate/qwen_vl_interface': 9.536186915683379e-06, 'learning_rate/action_model': 9.53618691568338e-05, 'epoch': 0.06}) +[2026-10-03 13:10:17,324][starVLA.training.train_starvla][INFO] - Step 796, Loss: {'train/loss': 0.0029867684934288263, 'train/grad_norm_pre_clip': 1.0372350215911865, 'timing/data': 0.00035705999471247196, 'timing/model': 0.38787987201067153, 'learning_rate/qwen_vl_interface': 9.534873478671289e-06, 'learning_rate/action_model': 9.534873478671288e-05, 'epoch': 0.06}) +[2026-10-03 13:10:18,018][starVLA.training.train_starvla][INFO] - Step 797, Loss: {'train/loss': 0.005429328419268131, 'train/grad_norm_pre_clip': 2.1406826972961426, 'timing/data': 0.0002490699989721179, 'timing/model': 0.38862153100490104, 'learning_rate/qwen_vl_interface': 9.533558280309563e-06, 'learning_rate/action_model': 9.533558280309563e-05, 'epoch': 0.06}) +[2026-10-03 13:10:18,716][starVLA.training.train_starvla][INFO] - Step 798, Loss: {'train/loss': 0.0010123655665665865, 'train/grad_norm_pre_clip': 0.19777636229991913, 'timing/data': 0.00033221999183297157, 'timing/model': 0.39112446100625675, 'learning_rate/qwen_vl_interface': 9.532241321138832e-06, 'learning_rate/action_model': 9.532241321138833e-05, 'epoch': 0.06}) +[2026-10-03 13:10:19,415][starVLA.training.train_starvla][INFO] - Step 799, Loss: {'train/loss': 0.0023140141274780035, 'train/grad_norm_pre_clip': 0.2642248570919037, 'timing/data': 0.00020193999807816, 'timing/model': 0.3909018310077954, 'learning_rate/qwen_vl_interface': 9.530922601700451e-06, 'learning_rate/action_model': 9.530922601700452e-05, 'epoch': 0.06}) +[2026-10-03 13:10:20,301][starVLA.training.train_starvla][INFO] - Step 800, Loss: {'train/loss': 0.02004915662109852, 'train/grad_norm_pre_clip': 40.63461685180664, 'timing/data': 0.000424469995778054, 'timing/model': 0.5797095669986447, 'learning_rate/qwen_vl_interface': 9.52960212253649e-06, 'learning_rate/action_model': 9.529602122536491e-05, 'epoch': 0.06}) +[2026-10-03 13:10:20,996][starVLA.training.train_starvla][INFO] - Step 801, Loss: {'train/loss': 0.0026997611857950687, 'train/grad_norm_pre_clip': 0.31584662199020386, 'timing/data': 0.00031553000735584646, 'timing/model': 0.388437702000374, 'learning_rate/qwen_vl_interface': 9.528279884189754e-06, 'learning_rate/action_model': 9.528279884189754e-05, 'epoch': 0.06}) +[2026-10-03 13:10:21,694][starVLA.training.train_starvla][INFO] - Step 802, Loss: {'train/loss': 0.0018915042746812105, 'train/grad_norm_pre_clip': 0.27271807193756104, 'timing/data': 0.0003304999991087243, 'timing/model': 0.391039150999859, 'learning_rate/qwen_vl_interface': 9.526955887203762e-06, 'learning_rate/action_model': 9.526955887203761e-05, 'epoch': 0.06}) +[2026-10-03 13:10:22,394][starVLA.training.train_starvla][INFO] - Step 803, Loss: {'train/loss': 0.4270608127117157, 'train/grad_norm_pre_clip': 37.89044952392578, 'timing/data': 0.0002629200025694445, 'timing/model': 0.3911989420012105, 'learning_rate/qwen_vl_interface': 9.525630132122758e-06, 'learning_rate/action_model': 9.525630132122758e-05, 'epoch': 0.06}) +[2026-10-03 13:10:23,089][starVLA.training.train_starvla][INFO] - Step 804, Loss: {'train/loss': 0.25738680362701416, 'train/grad_norm_pre_clip': 42.01462936401367, 'timing/data': 0.00019900999905075878, 'timing/model': 0.3891402519948315, 'learning_rate/qwen_vl_interface': 9.524302619491712e-06, 'learning_rate/action_model': 9.524302619491712e-05, 'epoch': 0.06}) +[2026-10-03 13:10:23,786][starVLA.training.train_starvla][INFO] - Step 805, Loss: {'train/loss': 0.23472854495048523, 'train/grad_norm_pre_clip': 90.6847152709961, 'timing/data': 0.00020090999896638095, 'timing/model': 0.39079523099644575, 'learning_rate/qwen_vl_interface': 9.522973349856315e-06, 'learning_rate/action_model': 9.522973349856315e-05, 'epoch': 0.06}) +[2026-10-03 13:10:24,678][starVLA.training.train_starvla][INFO] - Step 806, Loss: {'train/loss': 0.0001700732682365924, 'train/grad_norm_pre_clip': 0.028070814907550812, 'timing/data': 0.00034683999547269195, 'timing/model': 0.3912997410079697, 'learning_rate/qwen_vl_interface': 9.521642323762976e-06, 'learning_rate/action_model': 9.521642323762976e-05, 'epoch': 0.06}) +[2026-10-03 13:10:25,376][starVLA.training.train_starvla][INFO] - Step 807, Loss: {'train/loss': 0.9428095817565918, 'train/grad_norm_pre_clip': 114.1877212524414, 'timing/data': 0.00020502900588326156, 'timing/model': 0.3912305519916117, 'learning_rate/qwen_vl_interface': 9.520309541758833e-06, 'learning_rate/action_model': 9.520309541758833e-05, 'epoch': 0.06}) +[2026-10-03 13:10:26,071][starVLA.training.train_starvla][INFO] - Step 808, Loss: {'train/loss': 0.0009987839730456471, 'train/grad_norm_pre_clip': 0.35027316212654114, 'timing/data': 0.000394110000343062, 'timing/model': 0.38874692098761443, 'learning_rate/qwen_vl_interface': 9.51897500439174e-06, 'learning_rate/action_model': 9.51897500439174e-05, 'epoch': 0.06}) +[2026-10-03 13:10:26,767][starVLA.training.train_starvla][INFO] - Step 809, Loss: {'train/loss': 0.09746120870113373, 'train/grad_norm_pre_clip': 93.3114242553711, 'timing/data': 0.00019689000328071415, 'timing/model': 0.389578841990442, 'learning_rate/qwen_vl_interface': 9.517638712210277e-06, 'learning_rate/action_model': 9.517638712210276e-05, 'epoch': 0.06}) +[2026-10-03 13:10:27,465][starVLA.training.train_starvla][INFO] - Step 810, Loss: {'train/loss': 0.40605607628822327, 'train/grad_norm_pre_clip': 19.017736434936523, 'timing/data': 0.00021028000628575683, 'timing/model': 0.39111843099817634, 'learning_rate/qwen_vl_interface': 9.51630066576374e-06, 'learning_rate/action_model': 9.51630066576374e-05, 'epoch': 0.06}) +[2026-10-03 13:10:28,360][starVLA.training.train_starvla][INFO] - Step 811, Loss: {'train/loss': 0.5248190760612488, 'train/grad_norm_pre_clip': 10.680839538574219, 'timing/data': 0.00037426999188028276, 'timing/model': 0.3916905910009518, 'learning_rate/qwen_vl_interface': 9.514960865602154e-06, 'learning_rate/action_model': 9.514960865602153e-05, 'epoch': 0.06}) +[2026-10-03 13:10:29,056][starVLA.training.train_starvla][INFO] - Step 812, Loss: {'train/loss': 0.4271201193332672, 'train/grad_norm_pre_clip': 56.758636474609375, 'timing/data': 0.0002973200025735423, 'timing/model': 0.3891791609930806, 'learning_rate/qwen_vl_interface': 9.513619312276259e-06, 'learning_rate/action_model': 9.513619312276258e-05, 'epoch': 0.06}) +[2026-10-03 13:10:29,752][starVLA.training.train_starvla][INFO] - Step 813, Loss: {'train/loss': 0.0006451222579926252, 'train/grad_norm_pre_clip': 0.13784530758857727, 'timing/data': 0.00030948000494390726, 'timing/model': 0.38897954201092944, 'learning_rate/qwen_vl_interface': 9.512276006337514e-06, 'learning_rate/action_model': 9.512276006337513e-05, 'epoch': 0.06}) +[2026-10-03 13:10:30,449][starVLA.training.train_starvla][INFO] - Step 814, Loss: {'train/loss': 0.4542083740234375, 'train/grad_norm_pre_clip': 15.354511260986328, 'timing/data': 0.0002444199926685542, 'timing/model': 0.3913153410103405, 'learning_rate/qwen_vl_interface': 9.510930948338102e-06, 'learning_rate/action_model': 9.510930948338102e-05, 'epoch': 0.06}) +[2026-10-03 13:10:31,149][starVLA.training.train_starvla][INFO] - Step 815, Loss: {'train/loss': 0.0270075723528862, 'train/grad_norm_pre_clip': 11.79384708404541, 'timing/data': 0.00027694000164046884, 'timing/model': 0.3906391209893627, 'learning_rate/qwen_vl_interface': 9.509584138830929e-06, 'learning_rate/action_model': 9.509584138830928e-05, 'epoch': 0.06}) +[2026-10-03 13:10:32,037][starVLA.training.train_starvla][INFO] - Step 816, Loss: {'train/loss': 0.0055481502786278725, 'train/grad_norm_pre_clip': 1.4708302021026611, 'timing/data': 0.00029819000337738544, 'timing/model': 0.5810349279927323, 'learning_rate/qwen_vl_interface': 9.508235578369613e-06, 'learning_rate/action_model': 9.508235578369612e-05, 'epoch': 0.06}) +[2026-10-03 13:10:32,732][starVLA.training.train_starvla][INFO] - Step 817, Loss: {'train/loss': 0.35738006234169006, 'train/grad_norm_pre_clip': 13.767190933227539, 'timing/data': 0.0001951099984580651, 'timing/model': 0.3899477420054609, 'learning_rate/qwen_vl_interface': 9.506885267508498e-06, 'learning_rate/action_model': 9.506885267508498e-05, 'epoch': 0.06}) +[2026-10-03 13:10:33,429][starVLA.training.train_starvla][INFO] - Step 818, Loss: {'train/loss': 0.0015768927987664938, 'train/grad_norm_pre_clip': 0.3559389114379883, 'timing/data': 0.00025287999596912414, 'timing/model': 0.39084451099915896, 'learning_rate/qwen_vl_interface': 9.505533206802648e-06, 'learning_rate/action_model': 9.505533206802647e-05, 'epoch': 0.06}) +[2026-10-03 13:10:34,130][starVLA.training.train_starvla][INFO] - Step 819, Loss: {'train/loss': 0.3818821907043457, 'train/grad_norm_pre_clip': 54.332969665527344, 'timing/data': 0.00019884998619090766, 'timing/model': 0.39128456100297626, 'learning_rate/qwen_vl_interface': 9.50417939680784e-06, 'learning_rate/action_model': 9.504179396807839e-05, 'epoch': 0.06}) +[2026-10-03 13:10:34,826][starVLA.training.train_starvla][INFO] - Step 820, Loss: {'train/loss': 0.035800255835056305, 'train/grad_norm_pre_clip': 27.9229736328125, 'timing/data': 0.00028603999817278236, 'timing/model': 0.38910368099459447, 'learning_rate/qwen_vl_interface': 9.502823838080578e-06, 'learning_rate/action_model': 9.502823838080578e-05, 'epoch': 0.06}) +[2026-10-03 13:10:35,522][starVLA.training.train_starvla][INFO] - Step 821, Loss: {'train/loss': 0.002812336664646864, 'train/grad_norm_pre_clip': 1.278968095779419, 'timing/data': 0.00019833000260405242, 'timing/model': 0.38959397200960666, 'learning_rate/qwen_vl_interface': 9.501466531178079e-06, 'learning_rate/action_model': 9.501466531178079e-05, 'epoch': 0.06}) +[2026-10-03 13:10:36,412][starVLA.training.train_starvla][INFO] - Step 822, Loss: {'train/loss': 0.5374369621276855, 'train/grad_norm_pre_clip': 33.518165588378906, 'timing/data': 0.0003527200024109334, 'timing/model': 0.3919431709946366, 'learning_rate/qwen_vl_interface': 9.50010747665828e-06, 'learning_rate/action_model': 9.500107476658281e-05, 'epoch': 0.06}) +[2026-10-03 13:10:37,112][starVLA.training.train_starvla][INFO] - Step 823, Loss: {'train/loss': 0.0030740262009203434, 'train/grad_norm_pre_clip': 0.7214831113815308, 'timing/data': 0.0002585100010037422, 'timing/model': 0.39077930100029334, 'learning_rate/qwen_vl_interface': 9.498746675079842e-06, 'learning_rate/action_model': 9.49874667507984e-05, 'epoch': 0.06}) +[2026-10-03 13:10:37,807][starVLA.training.train_starvla][INFO] - Step 824, Loss: {'train/loss': 0.014907272532582283, 'train/grad_norm_pre_clip': 7.326902866363525, 'timing/data': 0.0003171999996993691, 'timing/model': 0.38860544199997094, 'learning_rate/qwen_vl_interface': 9.49738412700213e-06, 'learning_rate/action_model': 9.49738412700213e-05, 'epoch': 0.06}) +[2026-10-03 13:10:38,504][starVLA.training.train_starvla][INFO] - Step 825, Loss: {'train/loss': 0.0024604196660220623, 'train/grad_norm_pre_clip': 0.4355475604534149, 'timing/data': 0.0002839399967342615, 'timing/model': 0.38949488100479357, 'learning_rate/qwen_vl_interface': 9.496019832985243e-06, 'learning_rate/action_model': 9.496019832985243e-05, 'epoch': 0.06}) +[2026-10-03 13:10:39,202][starVLA.training.train_starvla][INFO] - Step 826, Loss: {'train/loss': 0.011516627855598927, 'train/grad_norm_pre_clip': 5.160035133361816, 'timing/data': 0.00033881000126712024, 'timing/model': 0.391280370997265, 'learning_rate/qwen_vl_interface': 9.49465379358999e-06, 'learning_rate/action_model': 9.49465379358999e-05, 'epoch': 0.06}) +[2026-10-03 13:10:40,091][starVLA.training.train_starvla][INFO] - Step 827, Loss: {'train/loss': 0.13969765603542328, 'train/grad_norm_pre_clip': 25.637741088867188, 'timing/data': 0.000198840003577061, 'timing/model': 0.3903196010069223, 'learning_rate/qwen_vl_interface': 9.493286009377896e-06, 'learning_rate/action_model': 9.493286009377896e-05, 'epoch': 0.06}) +[2026-10-03 13:10:40,787][starVLA.training.train_starvla][INFO] - Step 828, Loss: {'train/loss': 0.002659029792994261, 'train/grad_norm_pre_clip': 0.956358015537262, 'timing/data': 0.0004784799966728315, 'timing/model': 0.3889582420088118, 'learning_rate/qwen_vl_interface': 9.491916480911208e-06, 'learning_rate/action_model': 9.491916480911207e-05, 'epoch': 0.06}) +[2026-10-03 13:10:41,482][starVLA.training.train_starvla][INFO] - Step 829, Loss: {'train/loss': 0.054170526564121246, 'train/grad_norm_pre_clip': 48.57915496826172, 'timing/data': 0.0001965300034498796, 'timing/model': 0.3888187210104661, 'learning_rate/qwen_vl_interface': 9.490545208752884e-06, 'learning_rate/action_model': 9.490545208752883e-05, 'epoch': 0.06}) +[2026-10-03 13:10:42,180][starVLA.training.train_starvla][INFO] - Step 830, Loss: {'train/loss': 0.4699117839336395, 'train/grad_norm_pre_clip': 70.28243255615234, 'timing/data': 0.0002745200035860762, 'timing/model': 0.39092359200003557, 'learning_rate/qwen_vl_interface': 9.489172193466603e-06, 'learning_rate/action_model': 9.489172193466603e-05, 'epoch': 0.06}) +[2026-10-03 13:10:42,880][starVLA.training.train_starvla][INFO] - Step 831, Loss: {'train/loss': 0.0021581612527370453, 'train/grad_norm_pre_clip': 0.40200167894363403, 'timing/data': 0.000254809987382032, 'timing/model': 0.3913018919993192, 'learning_rate/qwen_vl_interface': 9.487797435616762e-06, 'learning_rate/action_model': 9.487797435616761e-05, 'epoch': 0.06}) +[2026-10-03 13:10:43,767][starVLA.training.train_starvla][INFO] - Step 832, Loss: {'train/loss': 0.2154284566640854, 'train/grad_norm_pre_clip': 161.70777893066406, 'timing/data': 0.0002510600024834275, 'timing/model': 0.5794744779996108, 'learning_rate/qwen_vl_interface': 9.486420935768472e-06, 'learning_rate/action_model': 9.486420935768472e-05, 'epoch': 0.06}) +[2026-10-03 13:10:44,463][starVLA.training.train_starvla][INFO] - Step 833, Loss: {'train/loss': 0.09802499413490295, 'train/grad_norm_pre_clip': 82.24412536621094, 'timing/data': 0.00019670999608933926, 'timing/model': 0.3893037810048554, 'learning_rate/qwen_vl_interface': 9.485042694487558e-06, 'learning_rate/action_model': 9.485042694487559e-05, 'epoch': 0.06}) +[2026-10-03 13:10:45,160][starVLA.training.train_starvla][INFO] - Step 834, Loss: {'train/loss': 0.4050537347793579, 'train/grad_norm_pre_clip': 52.33375930786133, 'timing/data': 0.00026200999855063856, 'timing/model': 0.390774372004671, 'learning_rate/qwen_vl_interface': 9.483662712340565e-06, 'learning_rate/action_model': 9.483662712340564e-05, 'epoch': 0.06}) +[2026-10-03 13:10:45,861][starVLA.training.train_starvla][INFO] - Step 835, Loss: {'train/loss': 0.001998115796595812, 'train/grad_norm_pre_clip': 0.26270636916160583, 'timing/data': 0.00020700000459328294, 'timing/model': 0.39135639200685546, 'learning_rate/qwen_vl_interface': 9.482280989894748e-06, 'learning_rate/action_model': 9.482280989894748e-05, 'epoch': 0.06}) +[2026-10-03 13:10:46,558][starVLA.training.train_starvla][INFO] - Step 836, Loss: {'train/loss': 0.47610607743263245, 'train/grad_norm_pre_clip': 104.60572052001953, 'timing/data': 0.00025533999723847955, 'timing/model': 0.3892480520007666, 'learning_rate/qwen_vl_interface': 9.480897527718083e-06, 'learning_rate/action_model': 9.480897527718084e-05, 'epoch': 0.06}) +[2026-10-03 13:10:47,446][starVLA.training.train_starvla][INFO] - Step 837, Loss: {'train/loss': 0.0014345725066959858, 'train/grad_norm_pre_clip': 0.40712398290634155, 'timing/data': 0.0002790200087474659, 'timing/model': 0.5817809069994837, 'learning_rate/qwen_vl_interface': 9.479512326379263e-06, 'learning_rate/action_model': 9.479512326379264e-05, 'epoch': 0.06}) +[2026-10-03 13:10:48,145][starVLA.training.train_starvla][INFO] - Step 838, Loss: {'train/loss': 0.0016323430463671684, 'train/grad_norm_pre_clip': 0.2471165657043457, 'timing/data': 0.00026482000248506665, 'timing/model': 0.39099420100683346, 'learning_rate/qwen_vl_interface': 9.478125386447686e-06, 'learning_rate/action_model': 9.478125386447686e-05, 'epoch': 0.06}) +[2026-10-03 13:10:48,844][starVLA.training.train_starvla][INFO] - Step 839, Loss: {'train/loss': 0.7273821234703064, 'train/grad_norm_pre_clip': 147.2137451171875, 'timing/data': 0.00020373000006657094, 'timing/model': 0.39095142199948896, 'learning_rate/qwen_vl_interface': 9.476736708493472e-06, 'learning_rate/action_model': 9.476736708493472e-05, 'epoch': 0.06}) +[2026-10-03 13:10:49,540][starVLA.training.train_starvla][INFO] - Step 840, Loss: {'train/loss': 0.0005283025093376637, 'train/grad_norm_pre_clip': 0.0827229917049408, 'timing/data': 0.00019829999655485153, 'timing/model': 0.38915910200739745, 'learning_rate/qwen_vl_interface': 9.475346293087457e-06, 'learning_rate/action_model': 9.475346293087458e-05, 'epoch': 0.06}) +[2026-10-03 13:10:50,236][starVLA.training.train_starvla][INFO] - Step 841, Loss: {'train/loss': 0.04040155187249184, 'train/grad_norm_pre_clip': 67.186279296875, 'timing/data': 0.00025549999554641545, 'timing/model': 0.38918848100001924, 'learning_rate/qwen_vl_interface': 9.473954140801184e-06, 'learning_rate/action_model': 9.473954140801185e-05, 'epoch': 0.06}) +[2026-10-03 13:10:51,127][starVLA.training.train_starvla][INFO] - Step 842, Loss: {'train/loss': 0.6362903714179993, 'train/grad_norm_pre_clip': 55.05371856689453, 'timing/data': 0.0002627200010465458, 'timing/model': 0.5845421980047831, 'learning_rate/qwen_vl_interface': 9.472560252206918e-06, 'learning_rate/action_model': 9.472560252206918e-05, 'epoch': 0.06}) +[2026-10-03 13:10:51,825][starVLA.training.train_starvla][INFO] - Step 843, Loss: {'train/loss': 0.4096137285232544, 'train/grad_norm_pre_clip': 106.82939147949219, 'timing/data': 0.0002581799926701933, 'timing/model': 0.3904739509889623, 'learning_rate/qwen_vl_interface': 9.471164627877633e-06, 'learning_rate/action_model': 9.471164627877633e-05, 'epoch': 0.06}) +[2026-10-03 13:10:52,520][starVLA.training.train_starvla][INFO] - Step 844, Loss: {'train/loss': 0.35211920738220215, 'train/grad_norm_pre_clip': 284.48773193359375, 'timing/data': 0.00023508000595029444, 'timing/model': 0.38886018299672287, 'learning_rate/qwen_vl_interface': 9.469767268387016e-06, 'learning_rate/action_model': 9.469767268387016e-05, 'epoch': 0.06}) +[2026-10-03 13:10:53,217][starVLA.training.train_starvla][INFO] - Step 845, Loss: {'train/loss': 0.09478048980236053, 'train/grad_norm_pre_clip': 91.19285583496094, 'timing/data': 0.00023237999994307756, 'timing/model': 0.3903018519922625, 'learning_rate/qwen_vl_interface': 9.468368174309468e-06, 'learning_rate/action_model': 9.468368174309469e-05, 'epoch': 0.06}) +[2026-10-03 13:10:53,914][starVLA.training.train_starvla][INFO] - Step 846, Loss: {'train/loss': 0.716667652130127, 'train/grad_norm_pre_clip': 75.48866271972656, 'timing/data': 0.00030621000041719526, 'timing/model': 0.39071851200424135, 'learning_rate/qwen_vl_interface': 9.466967346220107e-06, 'learning_rate/action_model': 9.466967346220107e-05, 'epoch': 0.06}) +[2026-10-03 13:10:54,613][starVLA.training.train_starvla][INFO] - Step 847, Loss: {'train/loss': 0.39332345128059387, 'train/grad_norm_pre_clip': 78.24971771240234, 'timing/data': 0.0002588400093372911, 'timing/model': 0.39087249200383667, 'learning_rate/qwen_vl_interface': 9.465564784694759e-06, 'learning_rate/action_model': 9.465564784694759e-05, 'epoch': 0.06}) +[2026-10-03 13:10:55,503][starVLA.training.train_starvla][INFO] - Step 848, Loss: {'train/loss': 0.00043215660843998194, 'train/grad_norm_pre_clip': 0.05855896323919296, 'timing/data': 0.00034360999416094273, 'timing/model': 0.38885738198587205, 'learning_rate/qwen_vl_interface': 9.464160490309962e-06, 'learning_rate/action_model': 9.464160490309962e-05, 'epoch': 0.06}) +[2026-10-03 13:10:56,198][starVLA.training.train_starvla][INFO] - Step 849, Loss: {'train/loss': 0.7014650106430054, 'train/grad_norm_pre_clip': 143.4590301513672, 'timing/data': 0.00027057000261265785, 'timing/model': 0.3886641819990473, 'learning_rate/qwen_vl_interface': 9.462754463642972e-06, 'learning_rate/action_model': 9.46275446364297e-05, 'epoch': 0.06}) +[2026-10-03 13:10:56,896][starVLA.training.train_starvla][INFO] - Step 850, Loss: {'train/loss': 0.028429480269551277, 'train/grad_norm_pre_clip': 39.6861572265625, 'timing/data': 0.0002858600055333227, 'timing/model': 0.3910109620046569, 'learning_rate/qwen_vl_interface': 9.461346705271753e-06, 'learning_rate/action_model': 9.461346705271752e-05, 'epoch': 0.06}) +[2026-10-03 13:10:57,595][starVLA.training.train_starvla][INFO] - Step 851, Loss: {'train/loss': 0.06048961356282234, 'train/grad_norm_pre_clip': 49.996559143066406, 'timing/data': 0.00019977999909315258, 'timing/model': 0.39116243099852, 'learning_rate/qwen_vl_interface': 9.459937215774976e-06, 'learning_rate/action_model': 9.459937215774976e-05, 'epoch': 0.06}) +[2026-10-03 13:10:58,289][starVLA.training.train_starvla][INFO] - Step 852, Loss: {'train/loss': 0.4217367470264435, 'train/grad_norm_pre_clip': 105.12957763671875, 'timing/data': 0.0003732000041054562, 'timing/model': 0.3883409819973167, 'learning_rate/qwen_vl_interface': 9.458525995732035e-06, 'learning_rate/action_model': 9.458525995732035e-05, 'epoch': 0.06}) +[2026-10-03 13:10:59,175][starVLA.training.train_starvla][INFO] - Step 853, Loss: {'train/loss': 0.4838533401489258, 'train/grad_norm_pre_clip': 93.99999237060547, 'timing/data': 0.0002497300010873005, 'timing/model': 0.578970628004754, 'learning_rate/qwen_vl_interface': 9.457113045723027e-06, 'learning_rate/action_model': 9.457113045723026e-05, 'epoch': 0.06}) +[2026-10-03 13:10:59,871][starVLA.training.train_starvla][INFO] - Step 854, Loss: {'train/loss': 0.25284597277641296, 'train/grad_norm_pre_clip': 45.422332763671875, 'timing/data': 0.00027147999207954854, 'timing/model': 0.3901859610050451, 'learning_rate/qwen_vl_interface': 9.455698366328762e-06, 'learning_rate/action_model': 9.455698366328762e-05, 'epoch': 0.06}) +[2026-10-03 13:11:00,570][starVLA.training.train_starvla][INFO] - Step 855, Loss: {'train/loss': 0.004954512696713209, 'train/grad_norm_pre_clip': 6.391554832458496, 'timing/data': 0.0002581499866209924, 'timing/model': 0.3914736619917676, 'learning_rate/qwen_vl_interface': 9.454281958130761e-06, 'learning_rate/action_model': 9.454281958130761e-05, 'epoch': 0.06}) +[2026-10-03 13:11:01,265][starVLA.training.train_starvla][INFO] - Step 856, Loss: {'train/loss': 0.0018617197638377547, 'train/grad_norm_pre_clip': 0.4585999548435211, 'timing/data': 0.0003278299991507083, 'timing/model': 0.38811617200553883, 'learning_rate/qwen_vl_interface': 9.452863821711257e-06, 'learning_rate/action_model': 9.452863821711258e-05, 'epoch': 0.06}) +[2026-10-03 13:11:01,960][starVLA.training.train_starvla][INFO] - Step 857, Loss: {'train/loss': 0.38656777143478394, 'train/grad_norm_pre_clip': 248.39402770996094, 'timing/data': 0.00019950000569224358, 'timing/model': 0.38898779200098943, 'learning_rate/qwen_vl_interface': 9.451443957653195e-06, 'learning_rate/action_model': 9.451443957653195e-05, 'epoch': 0.06}) +[2026-10-03 13:11:02,658][starVLA.training.train_starvla][INFO] - Step 858, Loss: {'train/loss': 0.0035695258993655443, 'train/grad_norm_pre_clip': 1.573033332824707, 'timing/data': 0.00027445000887382776, 'timing/model': 0.3915024819871178, 'learning_rate/qwen_vl_interface': 9.450022366540221e-06, 'learning_rate/action_model': 9.45002236654022e-05, 'epoch': 0.06}) +[2026-10-03 13:11:03,549][starVLA.training.train_starvla][INFO] - Step 859, Loss: {'train/loss': 0.001545105827972293, 'train/grad_norm_pre_clip': 1.2071149349212646, 'timing/data': 0.00028592000307980925, 'timing/model': 0.3910793019895209, 'learning_rate/qwen_vl_interface': 9.448599048956702e-06, 'learning_rate/action_model': 9.448599048956702e-05, 'epoch': 0.06}) +[2026-10-03 13:11:04,245][starVLA.training.train_starvla][INFO] - Step 860, Loss: {'train/loss': 0.0012227436527609825, 'train/grad_norm_pre_clip': 0.8616220951080322, 'timing/data': 0.00019647000590339303, 'timing/model': 0.3890611109964084, 'learning_rate/qwen_vl_interface': 9.447174005487708e-06, 'learning_rate/action_model': 9.447174005487708e-05, 'epoch': 0.06}) +[2026-10-03 13:11:04,940][starVLA.training.train_starvla][INFO] - Step 861, Loss: {'train/loss': 0.0005368433194234967, 'train/grad_norm_pre_clip': 0.07718588411808014, 'timing/data': 0.0002555999963078648, 'timing/model': 0.38875179100432433, 'learning_rate/qwen_vl_interface': 9.445747236719022e-06, 'learning_rate/action_model': 9.445747236719022e-05, 'epoch': 0.06}) +[2026-10-03 13:11:05,637][starVLA.training.train_starvla][INFO] - Step 862, Loss: {'train/loss': 0.026113372296094894, 'train/grad_norm_pre_clip': 38.22627258300781, 'timing/data': 0.00029711000388488173, 'timing/model': 0.3913931910065003, 'learning_rate/qwen_vl_interface': 9.444318743237135e-06, 'learning_rate/action_model': 9.444318743237134e-05, 'epoch': 0.06}) +[2026-10-03 13:11:06,337][starVLA.training.train_starvla][INFO] - Step 863, Loss: {'train/loss': 0.5394770503044128, 'train/grad_norm_pre_clip': 90.04412841796875, 'timing/data': 0.00025826999626588076, 'timing/model': 0.3911182320007356, 'learning_rate/qwen_vl_interface': 9.442888525629244e-06, 'learning_rate/action_model': 9.442888525629245e-05, 'epoch': 0.06}) +[2026-10-03 13:11:07,228][starVLA.training.train_starvla][INFO] - Step 864, Loss: {'train/loss': 0.28678908944129944, 'train/grad_norm_pre_clip': 46.3591423034668, 'timing/data': 0.0003573200083337724, 'timing/model': 0.3879589819989633, 'learning_rate/qwen_vl_interface': 9.441456584483261e-06, 'learning_rate/action_model': 9.441456584483261e-05, 'epoch': 0.06}) +[2026-10-03 13:11:07,922][starVLA.training.train_starvla][INFO] - Step 865, Loss: {'train/loss': 1.4860711097717285, 'train/grad_norm_pre_clip': 344.02105712890625, 'timing/data': 0.0002196300047216937, 'timing/model': 0.3888231719902251, 'learning_rate/qwen_vl_interface': 9.440022920387801e-06, 'learning_rate/action_model': 9.440022920387801e-05, 'epoch': 0.06}) +[2026-10-03 13:11:08,619][starVLA.training.train_starvla][INFO] - Step 866, Loss: {'train/loss': 0.21731524169445038, 'train/grad_norm_pre_clip': 271.67156982421875, 'timing/data': 0.00027412999770604074, 'timing/model': 0.3912000909913331, 'learning_rate/qwen_vl_interface': 9.438587533932191e-06, 'learning_rate/action_model': 9.43858753393219e-05, 'epoch': 0.06}) +[2026-10-03 13:11:09,318][starVLA.training.train_starvla][INFO] - Step 867, Loss: {'train/loss': 0.008485318161547184, 'train/grad_norm_pre_clip': 6.998815536499023, 'timing/data': 0.0002027099981205538, 'timing/model': 0.39051049199770205, 'learning_rate/qwen_vl_interface': 9.43715042570646e-06, 'learning_rate/action_model': 9.43715042570646e-05, 'epoch': 0.06}) +[2026-10-03 13:11:10,014][starVLA.training.train_starvla][INFO] - Step 868, Loss: {'train/loss': 0.004786635749042034, 'train/grad_norm_pre_clip': 3.234785795211792, 'timing/data': 0.00019393999536987394, 'timing/model': 0.3891152019932633, 'learning_rate/qwen_vl_interface': 9.435711596301353e-06, 'learning_rate/action_model': 9.435711596301352e-05, 'epoch': 0.06}) +[2026-10-03 13:11:10,903][starVLA.training.train_starvla][INFO] - Step 869, Loss: {'train/loss': 8.491907647112384e-05, 'train/grad_norm_pre_clip': 0.02205282263457775, 'timing/data': 0.00020683000911958516, 'timing/model': 0.5839217379980255, 'learning_rate/qwen_vl_interface': 9.434271046308317e-06, 'learning_rate/action_model': 9.434271046308317e-05, 'epoch': 0.06}) +[2026-10-03 13:11:11,601][starVLA.training.train_starvla][INFO] - Step 870, Loss: {'train/loss': 0.13436542451381683, 'train/grad_norm_pre_clip': 49.251583099365234, 'timing/data': 0.00019947999680880457, 'timing/model': 0.39151178099564277, 'learning_rate/qwen_vl_interface': 9.432828776319507e-06, 'learning_rate/action_model': 9.432828776319505e-05, 'epoch': 0.06}) +[2026-10-03 13:11:12,299][starVLA.training.train_starvla][INFO] - Step 871, Loss: {'train/loss': 0.030804375186562538, 'train/grad_norm_pre_clip': 32.46541976928711, 'timing/data': 0.0002680100005818531, 'timing/model': 0.39016235200688243, 'learning_rate/qwen_vl_interface': 9.431384786927785e-06, 'learning_rate/action_model': 9.431384786927786e-05, 'epoch': 0.06}) +[2026-10-03 13:11:12,996][starVLA.training.train_starvla][INFO] - Step 872, Loss: {'train/loss': 0.00035649945493787527, 'train/grad_norm_pre_clip': 0.35413721203804016, 'timing/data': 0.0002966600004583597, 'timing/model': 0.38944283200544305, 'learning_rate/qwen_vl_interface': 9.429939078726723e-06, 'learning_rate/action_model': 9.429939078726722e-05, 'epoch': 0.06}) +[2026-10-03 13:11:13,692][starVLA.training.train_starvla][INFO] - Step 873, Loss: {'train/loss': 0.4488893151283264, 'train/grad_norm_pre_clip': 124.87867736816406, 'timing/data': 0.00030683999648317695, 'timing/model': 0.38942929200129583, 'learning_rate/qwen_vl_interface': 9.428491652310598e-06, 'learning_rate/action_model': 9.428491652310597e-05, 'epoch': 0.06}) +[2026-10-03 13:11:14,390][starVLA.training.train_starvla][INFO] - Step 874, Loss: {'train/loss': 0.8696175813674927, 'train/grad_norm_pre_clip': 89.58625030517578, 'timing/data': 0.00019376000273041427, 'timing/model': 0.39104218198917806, 'learning_rate/qwen_vl_interface': 9.427042508274386e-06, 'learning_rate/action_model': 9.427042508274386e-05, 'epoch': 0.06}) +[2026-10-03 13:11:15,284][starVLA.training.train_starvla][INFO] - Step 875, Loss: {'train/loss': 0.0017530807526782155, 'train/grad_norm_pre_clip': 1.434027075767517, 'timing/data': 0.0003451899974606931, 'timing/model': 0.3905196609994164, 'learning_rate/qwen_vl_interface': 9.425591647213783e-06, 'learning_rate/action_model': 9.425591647213782e-05, 'epoch': 0.06}) +[2026-10-03 13:11:15,980][starVLA.training.train_starvla][INFO] - Step 876, Loss: {'train/loss': 0.47313207387924194, 'train/grad_norm_pre_clip': 12.547382354736328, 'timing/data': 0.0002341499930480495, 'timing/model': 0.38927837200753856, 'learning_rate/qwen_vl_interface': 9.424139069725177e-06, 'learning_rate/action_model': 9.424139069725178e-05, 'epoch': 0.06}) +[2026-10-03 13:11:16,675][starVLA.training.train_starvla][INFO] - Step 877, Loss: {'train/loss': 0.007603678852319717, 'train/grad_norm_pre_clip': 9.1014986038208, 'timing/data': 0.0004410299879964441, 'timing/model': 0.38862630201037973, 'learning_rate/qwen_vl_interface': 9.422684776405673e-06, 'learning_rate/action_model': 9.422684776405672e-05, 'epoch': 0.06}) +[2026-10-03 13:11:17,371][starVLA.training.train_starvla][INFO] - Step 878, Loss: {'train/loss': 0.6601059436798096, 'train/grad_norm_pre_clip': 91.78118896484375, 'timing/data': 0.0002103000006172806, 'timing/model': 0.3893007519945968, 'learning_rate/qwen_vl_interface': 9.421228767853072e-06, 'learning_rate/action_model': 9.421228767853072e-05, 'epoch': 0.06}) +[2026-10-03 13:11:18,070][starVLA.training.train_starvla][INFO] - Step 879, Loss: {'train/loss': 0.3430830240249634, 'train/grad_norm_pre_clip': 163.60643005371094, 'timing/data': 0.0002829799923347309, 'timing/model': 0.3907005119981477, 'learning_rate/qwen_vl_interface': 9.419771044665887e-06, 'learning_rate/action_model': 9.419771044665887e-05, 'epoch': 0.06}) +[2026-10-03 13:11:18,962][starVLA.training.train_starvla][INFO] - Step 880, Loss: {'train/loss': 0.0015583470230922103, 'train/grad_norm_pre_clip': 0.6072074770927429, 'timing/data': 0.0002506099990569055, 'timing/model': 0.38950924199889414, 'learning_rate/qwen_vl_interface': 9.41831160744333e-06, 'learning_rate/action_model': 9.41831160744333e-05, 'epoch': 0.06}) +[2026-10-03 13:11:19,658][starVLA.training.train_starvla][INFO] - Step 881, Loss: {'train/loss': 0.2508804500102997, 'train/grad_norm_pre_clip': 133.44776916503906, 'timing/data': 0.00028279000252950937, 'timing/model': 0.38907730201026425, 'learning_rate/qwen_vl_interface': 9.416850456785322e-06, 'learning_rate/action_model': 9.416850456785321e-05, 'epoch': 0.06}) +[2026-10-03 13:11:20,355][starVLA.training.train_starvla][INFO] - Step 882, Loss: {'train/loss': 0.2058764100074768, 'train/grad_norm_pre_clip': 95.79219818115234, 'timing/data': 0.0002297999890288338, 'timing/model': 0.3906407619942911, 'learning_rate/qwen_vl_interface': 9.415387593292489e-06, 'learning_rate/action_model': 9.415387593292488e-05, 'epoch': 0.06}) +[2026-10-03 13:11:21,056][starVLA.training.train_starvla][INFO] - Step 883, Loss: {'train/loss': 0.3676103949546814, 'train/grad_norm_pre_clip': 54.48974609375, 'timing/data': 0.00025570001162122935, 'timing/model': 0.391145581001183, 'learning_rate/qwen_vl_interface': 9.413923017566155e-06, 'learning_rate/action_model': 9.413923017566155e-05, 'epoch': 0.06}) +[2026-10-03 13:11:21,752][starVLA.training.train_starvla][INFO] - Step 884, Loss: {'train/loss': 0.5293987393379211, 'train/grad_norm_pre_clip': 23.371824264526367, 'timing/data': 0.0002983390004374087, 'timing/model': 0.38992960199539084, 'learning_rate/qwen_vl_interface': 9.412456730208352e-06, 'learning_rate/action_model': 9.412456730208352e-05, 'epoch': 0.06}) +[2026-10-03 13:11:22,644][starVLA.training.train_starvla][INFO] - Step 885, Loss: {'train/loss': 0.002045747358351946, 'train/grad_norm_pre_clip': 1.2800005674362183, 'timing/data': 0.00019967999833170325, 'timing/model': 0.3893700229964452, 'learning_rate/qwen_vl_interface': 9.41098873182182e-06, 'learning_rate/action_model': 9.41098873182182e-05, 'epoch': 0.06}) +[2026-10-03 13:11:23,341][starVLA.training.train_starvla][INFO] - Step 886, Loss: {'train/loss': 0.24176526069641113, 'train/grad_norm_pre_clip': 10.904666900634766, 'timing/data': 0.00019916000019293278, 'timing/model': 0.39068251199205406, 'learning_rate/qwen_vl_interface': 9.409519023009993e-06, 'learning_rate/action_model': 9.409519023009993e-05, 'epoch': 0.06}) +[2026-10-03 13:11:24,042][starVLA.training.train_starvla][INFO] - Step 887, Loss: {'train/loss': 0.010385220870375633, 'train/grad_norm_pre_clip': 4.805645942687988, 'timing/data': 0.00033476999669801444, 'timing/model': 0.3917852720041992, 'learning_rate/qwen_vl_interface': 9.408047604377014e-06, 'learning_rate/action_model': 9.408047604377014e-05, 'epoch': 0.06}) +[2026-10-03 13:11:24,737][starVLA.training.train_starvla][INFO] - Step 888, Loss: {'train/loss': 0.04841914772987366, 'train/grad_norm_pre_clip': 29.22125816345215, 'timing/data': 0.00025456000003032386, 'timing/model': 0.38875151200045366, 'learning_rate/qwen_vl_interface': 9.40657447652773e-06, 'learning_rate/action_model': 9.406574476527731e-05, 'epoch': 0.06}) +[2026-10-03 13:11:25,432][starVLA.training.train_starvla][INFO] - Step 889, Loss: {'train/loss': 0.019758880138397217, 'train/grad_norm_pre_clip': 5.775009632110596, 'timing/data': 0.0003302099939901382, 'timing/model': 0.38966736200382, 'learning_rate/qwen_vl_interface': 9.405099640067685e-06, 'learning_rate/action_model': 9.405099640067685e-05, 'epoch': 0.07}) +[2026-10-03 13:11:26,321][starVLA.training.train_starvla][INFO] - Step 890, Loss: {'train/loss': 0.02532120607793331, 'train/grad_norm_pre_clip': 5.684986114501953, 'timing/data': 0.00025707000168040395, 'timing/model': 0.5827901279990328, 'learning_rate/qwen_vl_interface': 9.40362309560313e-06, 'learning_rate/action_model': 9.40362309560313e-05, 'epoch': 0.07}) +[2026-10-03 13:11:27,019][starVLA.training.train_starvla][INFO] - Step 891, Loss: {'train/loss': 0.04514393210411072, 'train/grad_norm_pre_clip': 46.9452018737793, 'timing/data': 0.0002808600111166015, 'timing/model': 0.39110393199371174, 'learning_rate/qwen_vl_interface': 9.402144843741016e-06, 'learning_rate/action_model': 9.402144843741017e-05, 'epoch': 0.07}) +[2026-10-03 13:11:27,714][starVLA.training.train_starvla][INFO] - Step 892, Loss: {'train/loss': 0.005474551115185022, 'train/grad_norm_pre_clip': 0.8406881093978882, 'timing/data': 0.00019767000048886985, 'timing/model': 0.38888724200660363, 'learning_rate/qwen_vl_interface': 9.400664885088999e-06, 'learning_rate/action_model': 9.400664885088999e-05, 'epoch': 0.07}) +[2026-10-03 13:11:28,408][starVLA.training.train_starvla][INFO] - Step 893, Loss: {'train/loss': 0.6647540330886841, 'train/grad_norm_pre_clip': 44.440879821777344, 'timing/data': 0.000348319998010993, 'timing/model': 0.3882649820006918, 'learning_rate/qwen_vl_interface': 9.39918322025543e-06, 'learning_rate/action_model': 9.39918322025543e-05, 'epoch': 0.07}) +[2026-10-03 13:11:29,105][starVLA.training.train_starvla][INFO] - Step 894, Loss: {'train/loss': 0.0027444595471024513, 'train/grad_norm_pre_clip': 0.791733980178833, 'timing/data': 0.0003434000100241974, 'timing/model': 0.39031812200846616, 'learning_rate/qwen_vl_interface': 9.397699849849368e-06, 'learning_rate/action_model': 9.397699849849367e-05, 'epoch': 0.07}) +[2026-10-03 13:11:29,995][starVLA.training.train_starvla][INFO] - Step 895, Loss: {'train/loss': 0.2495560348033905, 'train/grad_norm_pre_clip': 10.384142875671387, 'timing/data': 0.0002725100057432428, 'timing/model': 0.5824451779917581, 'learning_rate/qwen_vl_interface': 9.39621477448057e-06, 'learning_rate/action_model': 9.39621477448057e-05, 'epoch': 0.07}) +[2026-10-03 13:11:30,691][starVLA.training.train_starvla][INFO] - Step 896, Loss: {'train/loss': 0.5859500169754028, 'train/grad_norm_pre_clip': 26.17540740966797, 'timing/data': 0.00027847000455949455, 'timing/model': 0.38940834200184327, 'learning_rate/qwen_vl_interface': 9.394727994759495e-06, 'learning_rate/action_model': 9.394727994759495e-05, 'epoch': 0.07}) +[2026-10-03 13:11:31,386][starVLA.training.train_starvla][INFO] - Step 897, Loss: {'train/loss': 0.0033820776734501123, 'train/grad_norm_pre_clip': 0.3583381772041321, 'timing/data': 0.0002793199964798987, 'timing/model': 0.38811257199267857, 'learning_rate/qwen_vl_interface': 9.3932395112973e-06, 'learning_rate/action_model': 9.3932395112973e-05, 'epoch': 0.07}) +[2026-10-03 13:11:32,084][starVLA.training.train_starvla][INFO] - Step 898, Loss: {'train/loss': 0.0032566580921411514, 'train/grad_norm_pre_clip': 3.882582664489746, 'timing/data': 0.0002549500059103593, 'timing/model': 0.3909398810064886, 'learning_rate/qwen_vl_interface': 9.391749324705846e-06, 'learning_rate/action_model': 9.391749324705846e-05, 'epoch': 0.07}) +[2026-10-03 13:11:32,782][starVLA.training.train_starvla][INFO] - Step 899, Loss: {'train/loss': 0.7536072134971619, 'train/grad_norm_pre_clip': 73.82252502441406, 'timing/data': 0.0002585100010037422, 'timing/model': 0.3899862229882274, 'learning_rate/qwen_vl_interface': 9.390257435597692e-06, 'learning_rate/action_model': 9.390257435597692e-05, 'epoch': 0.07}) +[2026-10-03 13:11:33,476][starVLA.training.train_starvla][INFO] - Step 900, Loss: {'train/loss': 0.0006503799231722951, 'train/grad_norm_pre_clip': 0.20602814853191376, 'timing/data': 0.00019613000040408224, 'timing/model': 0.388542681990657, 'learning_rate/qwen_vl_interface': 9.3887638445861e-06, 'learning_rate/action_model': 9.3887638445861e-05, 'epoch': 0.07}) +[2026-10-03 13:11:34,365][starVLA.training.train_starvla][INFO] - Step 901, Loss: {'train/loss': 0.33653244376182556, 'train/grad_norm_pre_clip': 113.7125473022461, 'timing/data': 0.00033590001112315804, 'timing/model': 0.38806470201234333, 'learning_rate/qwen_vl_interface': 9.387268552285024e-06, 'learning_rate/action_model': 9.387268552285025e-05, 'epoch': 0.07}) +[2026-10-03 13:11:35,062][starVLA.training.train_starvla][INFO] - Step 902, Loss: {'train/loss': 0.006669067312031984, 'train/grad_norm_pre_clip': 0.8758928775787354, 'timing/data': 0.00024943999596871436, 'timing/model': 0.390243471993017, 'learning_rate/qwen_vl_interface': 9.385771559309125e-06, 'learning_rate/action_model': 9.385771559309125e-05, 'epoch': 0.07}) +[2026-10-03 13:11:35,760][starVLA.training.train_starvla][INFO] - Step 903, Loss: {'train/loss': 0.04779810830950737, 'train/grad_norm_pre_clip': 44.72685241699219, 'timing/data': 0.0002765100070973858, 'timing/model': 0.3911127720057266, 'learning_rate/qwen_vl_interface': 9.38427286627376e-06, 'learning_rate/action_model': 9.38427286627376e-05, 'epoch': 0.07}) +[2026-10-03 13:11:36,455][starVLA.training.train_starvla][INFO] - Step 904, Loss: {'train/loss': 0.0001009276820695959, 'train/grad_norm_pre_clip': 0.021594051271677017, 'timing/data': 0.00025800000003073364, 'timing/model': 0.3881053530058125, 'learning_rate/qwen_vl_interface': 9.382772473794985e-06, 'learning_rate/action_model': 9.382772473794985e-05, 'epoch': 0.07}) +[2026-10-03 13:11:37,148][starVLA.training.train_starvla][INFO] - Step 905, Loss: {'train/loss': 0.009202668443322182, 'train/grad_norm_pre_clip': 7.51676082611084, 'timing/data': 0.00032577999809291214, 'timing/model': 0.38862530200276524, 'learning_rate/qwen_vl_interface': 9.381270382489554e-06, 'learning_rate/action_model': 9.381270382489554e-05, 'epoch': 0.07}) +[2026-10-03 13:11:38,036][starVLA.training.train_starvla][INFO] - Step 906, Loss: {'train/loss': 0.005179139319807291, 'train/grad_norm_pre_clip': 3.2898693084716797, 'timing/data': 0.0002486200100975111, 'timing/model': 0.5819328579964349, 'learning_rate/qwen_vl_interface': 9.37976659297492e-06, 'learning_rate/action_model': 9.379766592974921e-05, 'epoch': 0.07}) +[2026-10-03 13:11:38,734][starVLA.training.train_starvla][INFO] - Step 907, Loss: {'train/loss': 0.8175337910652161, 'train/grad_norm_pre_clip': 150.48489379882812, 'timing/data': 0.0003151900018565357, 'timing/model': 0.39040386100532487, 'learning_rate/qwen_vl_interface': 9.378261105869237e-06, 'learning_rate/action_model': 9.378261105869237e-05, 'epoch': 0.07}) +[2026-10-03 13:11:39,429][starVLA.training.train_starvla][INFO] - Step 908, Loss: {'train/loss': 0.00027489938656799495, 'train/grad_norm_pre_clip': 0.1527438908815384, 'timing/data': 0.00038748999941162765, 'timing/model': 0.38813690199458506, 'learning_rate/qwen_vl_interface': 9.376753921791348e-06, 'learning_rate/action_model': 9.376753921791348e-05, 'epoch': 0.07}) +[2026-10-03 13:11:40,124][starVLA.training.train_starvla][INFO] - Step 909, Loss: {'train/loss': 0.28403177857398987, 'train/grad_norm_pre_clip': 80.4526138305664, 'timing/data': 0.0002509399928385392, 'timing/model': 0.38835120300063863, 'learning_rate/qwen_vl_interface': 9.375245041360805e-06, 'learning_rate/action_model': 9.375245041360804e-05, 'epoch': 0.07}) +[2026-10-03 13:11:40,822][starVLA.training.train_starvla][INFO] - Step 910, Loss: {'train/loss': 0.4009877145290375, 'train/grad_norm_pre_clip': 63.403018951416016, 'timing/data': 0.00031304000003729016, 'timing/model': 0.3912295019981684, 'learning_rate/qwen_vl_interface': 9.37373446519785e-06, 'learning_rate/action_model': 9.373734465197849e-05, 'epoch': 0.07}) +[2026-10-03 13:11:41,521][starVLA.training.train_starvla][INFO] - Step 911, Loss: {'train/loss': 0.00022594361507799476, 'train/grad_norm_pre_clip': 0.13453340530395508, 'timing/data': 0.0003123900096397847, 'timing/model': 0.390753080995637, 'learning_rate/qwen_vl_interface': 9.372222193923421e-06, 'learning_rate/action_model': 9.372222193923421e-05, 'epoch': 0.07}) +[2026-10-03 13:11:42,408][starVLA.training.train_starvla][INFO] - Step 912, Loss: {'train/loss': 0.3566809892654419, 'train/grad_norm_pre_clip': 56.716896057128906, 'timing/data': 0.00036255001032259315, 'timing/model': 0.38865031200111844, 'learning_rate/qwen_vl_interface': 9.370708228159158e-06, 'learning_rate/action_model': 9.370708228159159e-05, 'epoch': 0.07}) +[2026-10-03 13:11:43,103][starVLA.training.train_starvla][INFO] - Step 913, Loss: {'train/loss': 0.21525728702545166, 'train/grad_norm_pre_clip': 16.411972045898438, 'timing/data': 0.0002632300020195544, 'timing/model': 0.38795269299589563, 'learning_rate/qwen_vl_interface': 9.369192568527398e-06, 'learning_rate/action_model': 9.369192568527398e-05, 'epoch': 0.07}) +[2026-10-03 13:11:43,799][starVLA.training.train_starvla][INFO] - Step 914, Loss: {'train/loss': 0.06401678174734116, 'train/grad_norm_pre_clip': 22.078166961669922, 'timing/data': 0.0002633800031617284, 'timing/model': 0.3902449819870526, 'learning_rate/qwen_vl_interface': 9.367675215651168e-06, 'learning_rate/action_model': 9.367675215651168e-05, 'epoch': 0.07}) +[2026-10-03 13:11:44,497][starVLA.training.train_starvla][INFO] - Step 915, Loss: {'train/loss': 0.4169984459877014, 'train/grad_norm_pre_clip': 79.26268768310547, 'timing/data': 0.0002571099903434515, 'timing/model': 0.3902137120021507, 'learning_rate/qwen_vl_interface': 9.366156170154194e-06, 'learning_rate/action_model': 9.366156170154194e-05, 'epoch': 0.07}) +[2026-10-03 13:11:45,191][starVLA.training.train_starvla][INFO] - Step 916, Loss: {'train/loss': 0.004569300916045904, 'train/grad_norm_pre_clip': 0.8395262956619263, 'timing/data': 0.0003136000013910234, 'timing/model': 0.38811243200325407, 'learning_rate/qwen_vl_interface': 9.3646354326609e-06, 'learning_rate/action_model': 9.364635432660899e-05, 'epoch': 0.07}) +[2026-10-03 13:11:46,080][starVLA.training.train_starvla][INFO] - Step 917, Loss: {'train/loss': 0.054737962782382965, 'train/grad_norm_pre_clip': 48.42845153808594, 'timing/data': 0.0003229600115446374, 'timing/model': 0.3879877220024355, 'learning_rate/qwen_vl_interface': 9.3631130037964e-06, 'learning_rate/action_model': 9.363113003796401e-05, 'epoch': 0.07}) +[2026-10-03 13:11:46,775][starVLA.training.train_starvla][INFO] - Step 918, Loss: {'train/loss': 0.007068915292620659, 'train/grad_norm_pre_clip': 1.1551685333251953, 'timing/data': 0.0002765300014289096, 'timing/model': 0.3892031819996191, 'learning_rate/qwen_vl_interface': 9.361588884186513e-06, 'learning_rate/action_model': 9.361588884186513e-05, 'epoch': 0.07}) +[2026-10-03 13:11:47,473][starVLA.training.train_starvla][INFO] - Step 919, Loss: {'train/loss': 1.01678466796875, 'train/grad_norm_pre_clip': 31.27790069580078, 'timing/data': 0.00030401999538298696, 'timing/model': 0.3898724409955321, 'learning_rate/qwen_vl_interface': 9.360063074457744e-06, 'learning_rate/action_model': 9.360063074457743e-05, 'epoch': 0.07}) +[2026-10-03 13:11:48,167][starVLA.training.train_starvla][INFO] - Step 920, Loss: {'train/loss': 0.0027574358973652124, 'train/grad_norm_pre_clip': 0.7635694146156311, 'timing/data': 0.0003207300032954663, 'timing/model': 0.38788603199645877, 'learning_rate/qwen_vl_interface': 9.358535575237296e-06, 'learning_rate/action_model': 9.358535575237295e-05, 'epoch': 0.07}) +[2026-10-03 13:11:48,861][starVLA.training.train_starvla][INFO] - Step 921, Loss: {'train/loss': 0.7285361289978027, 'train/grad_norm_pre_clip': 54.564353942871094, 'timing/data': 0.00031914000282995403, 'timing/model': 0.38806173199554905, 'learning_rate/qwen_vl_interface': 9.357006387153065e-06, 'learning_rate/action_model': 9.357006387153065e-05, 'epoch': 0.07}) +[2026-10-03 13:11:49,751][starVLA.training.train_starvla][INFO] - Step 922, Loss: {'train/loss': 0.6053905487060547, 'train/grad_norm_pre_clip': 66.3341293334961, 'timing/data': 0.00024528999347239733, 'timing/model': 0.584812057990348, 'learning_rate/qwen_vl_interface': 9.355475510833643e-06, 'learning_rate/action_model': 9.355475510833643e-05, 'epoch': 0.07}) +[2026-10-03 13:11:50,449][starVLA.training.train_starvla][INFO] - Step 923, Loss: {'train/loss': 0.004785060882568359, 'train/grad_norm_pre_clip': 4.083305358886719, 'timing/data': 0.0002582199958851561, 'timing/model': 0.39061442199454177, 'learning_rate/qwen_vl_interface': 9.353942946908318e-06, 'learning_rate/action_model': 9.353942946908317e-05, 'epoch': 0.07}) +[2026-10-03 13:11:51,145][starVLA.training.train_starvla][INFO] - Step 924, Loss: {'train/loss': 0.0024375207722187042, 'train/grad_norm_pre_clip': 1.0196774005889893, 'timing/data': 0.0003696200001286343, 'timing/model': 0.38842163200024515, 'learning_rate/qwen_vl_interface': 9.352408696007065e-06, 'learning_rate/action_model': 9.352408696007065e-05, 'epoch': 0.07}) +[2026-10-03 13:11:51,841][starVLA.training.train_starvla][INFO] - Step 925, Loss: {'train/loss': 0.0020884096156805754, 'train/grad_norm_pre_clip': 1.1596333980560303, 'timing/data': 0.0002513799990992993, 'timing/model': 0.3896239419991616, 'learning_rate/qwen_vl_interface': 9.350872758760558e-06, 'learning_rate/action_model': 9.350872758760559e-05, 'epoch': 0.07}) +[2026-10-03 13:11:52,538][starVLA.training.train_starvla][INFO] - Step 926, Loss: {'train/loss': 0.576449990272522, 'train/grad_norm_pre_clip': 88.19144439697266, 'timing/data': 0.00031643999682273716, 'timing/model': 0.3907741820003139, 'learning_rate/qwen_vl_interface': 9.349335135800164e-06, 'learning_rate/action_model': 9.349335135800163e-05, 'epoch': 0.07}) +[2026-10-03 13:11:53,236][starVLA.training.train_starvla][INFO] - Step 927, Loss: {'train/loss': 0.5755404829978943, 'train/grad_norm_pre_clip': 50.60051345825195, 'timing/data': 0.0002738499897532165, 'timing/model': 0.39025796299392823, 'learning_rate/qwen_vl_interface': 9.347795827757942e-06, 'learning_rate/action_model': 9.347795827757941e-05, 'epoch': 0.07}) +[2026-10-03 13:11:54,122][starVLA.training.train_starvla][INFO] - Step 928, Loss: {'train/loss': 1.140298843383789, 'train/grad_norm_pre_clip': 62.29750061035156, 'timing/data': 0.00026538998645264655, 'timing/model': 0.3884081229916774, 'learning_rate/qwen_vl_interface': 9.346254835266643e-06, 'learning_rate/action_model': 9.346254835266642e-05, 'epoch': 0.07}) +[2026-10-03 13:11:54,818][starVLA.training.train_starvla][INFO] - Step 929, Loss: {'train/loss': 0.0007753060781396925, 'train/grad_norm_pre_clip': 0.23107261955738068, 'timing/data': 0.0002872600016416982, 'timing/model': 0.3888836420082953, 'learning_rate/qwen_vl_interface': 9.344712158959709e-06, 'learning_rate/action_model': 9.344712158959709e-05, 'epoch': 0.07}) +[2026-10-03 13:11:55,515][starVLA.training.train_starvla][INFO] - Step 930, Loss: {'train/loss': 0.13454005122184753, 'train/grad_norm_pre_clip': 51.60409927368164, 'timing/data': 0.00023444999533239752, 'timing/model': 0.3906347820011433, 'learning_rate/qwen_vl_interface': 9.343167799471278e-06, 'learning_rate/action_model': 9.343167799471277e-05, 'epoch': 0.07}) +[2026-10-03 13:11:56,215][starVLA.training.train_starvla][INFO] - Step 931, Loss: {'train/loss': 0.011308836750686169, 'train/grad_norm_pre_clip': 15.65034008026123, 'timing/data': 0.0003123700007563457, 'timing/model': 0.3912362519913586, 'learning_rate/qwen_vl_interface': 9.341621757436177e-06, 'learning_rate/action_model': 9.341621757436177e-05, 'epoch': 0.07}) +[2026-10-03 13:11:56,909][starVLA.training.train_starvla][INFO] - Step 932, Loss: {'train/loss': 0.0018667076947167516, 'train/grad_norm_pre_clip': 0.34474417567253113, 'timing/data': 0.0003740800020750612, 'timing/model': 0.3883902119996492, 'learning_rate/qwen_vl_interface': 9.340074033489924e-06, 'learning_rate/action_model': 9.340074033489923e-05, 'epoch': 0.07}) +[2026-10-03 13:11:57,798][starVLA.training.train_starvla][INFO] - Step 933, Loss: {'train/loss': 0.002074922202154994, 'train/grad_norm_pre_clip': 0.9145950078964233, 'timing/data': 0.00028655001369770616, 'timing/model': 0.38834314300038386, 'learning_rate/qwen_vl_interface': 9.338524628268734e-06, 'learning_rate/action_model': 9.338524628268733e-05, 'epoch': 0.07}) +[2026-10-03 13:11:58,493][starVLA.training.train_starvla][INFO] - Step 934, Loss: {'train/loss': 0.004085138440132141, 'train/grad_norm_pre_clip': 2.173861503601074, 'timing/data': 0.00025921998894773424, 'timing/model': 0.3896982619917253, 'learning_rate/qwen_vl_interface': 9.336973542409505e-06, 'learning_rate/action_model': 9.336973542409504e-05, 'epoch': 0.07}) +[2026-10-03 13:11:59,190][starVLA.training.train_starvla][INFO] - Step 935, Loss: {'train/loss': 0.02862602099776268, 'train/grad_norm_pre_clip': 25.449941635131836, 'timing/data': 0.0002768099948298186, 'timing/model': 0.38982370200392324, 'learning_rate/qwen_vl_interface': 9.33542077654983e-06, 'learning_rate/action_model': 9.33542077654983e-05, 'epoch': 0.07}) +[2026-10-03 13:11:59,882][starVLA.training.train_starvla][INFO] - Step 936, Loss: {'train/loss': 0.004181908909231424, 'train/grad_norm_pre_clip': 1.580273985862732, 'timing/data': 0.0002555699902586639, 'timing/model': 0.3877920030063251, 'learning_rate/qwen_vl_interface': 9.333866331327995e-06, 'learning_rate/action_model': 9.333866331327995e-05, 'epoch': 0.07}) +[2026-10-03 13:12:00,576][starVLA.training.train_starvla][INFO] - Step 937, Loss: {'train/loss': 0.008426355198025703, 'train/grad_norm_pre_clip': 12.984621047973633, 'timing/data': 0.0002363600069656968, 'timing/model': 0.3886303330073133, 'learning_rate/qwen_vl_interface': 9.332310207382972e-06, 'learning_rate/action_model': 9.332310207382972e-05, 'epoch': 0.07}) +[2026-10-03 13:12:01,464][starVLA.training.train_starvla][INFO] - Step 938, Loss: {'train/loss': 0.30796512961387634, 'train/grad_norm_pre_clip': 119.37511444091797, 'timing/data': 0.0001967999996850267, 'timing/model': 0.5822543579997728, 'learning_rate/qwen_vl_interface': 9.330752405354426e-06, 'learning_rate/action_model': 9.330752405354426e-05, 'epoch': 0.07}) +[2026-10-03 13:12:02,162][starVLA.training.train_starvla][INFO] - Step 939, Loss: {'train/loss': 0.8126460313796997, 'train/grad_norm_pre_clip': 155.85562133789062, 'timing/data': 0.00029379999614320695, 'timing/model': 0.3899998730048537, 'learning_rate/qwen_vl_interface': 9.32919292588271e-06, 'learning_rate/action_model': 9.32919292588271e-05, 'epoch': 0.07}) +[2026-10-03 13:12:02,855][starVLA.training.train_starvla][INFO] - Step 940, Loss: {'train/loss': 0.0007855934672988951, 'train/grad_norm_pre_clip': 0.15983137488365173, 'timing/data': 0.00023206000332720578, 'timing/model': 0.3878089820063906, 'learning_rate/qwen_vl_interface': 9.327631769608864e-06, 'learning_rate/action_model': 9.327631769608864e-05, 'epoch': 0.07}) +[2026-10-03 13:12:03,549][starVLA.training.train_starvla][INFO] - Step 941, Loss: {'train/loss': 0.020356852561235428, 'train/grad_norm_pre_clip': 58.64100646972656, 'timing/data': 0.00028850999660789967, 'timing/model': 0.38832429199828766, 'learning_rate/qwen_vl_interface': 9.326068937174627e-06, 'learning_rate/action_model': 9.326068937174626e-05, 'epoch': 0.07}) +[2026-10-03 13:12:04,243][starVLA.training.train_starvla][INFO] - Step 942, Loss: {'train/loss': 0.0006200245115906, 'train/grad_norm_pre_clip': 0.6715466976165771, 'timing/data': 0.0002289300027769059, 'timing/model': 0.3886859829945024, 'learning_rate/qwen_vl_interface': 9.324504429222412e-06, 'learning_rate/action_model': 9.324504429222412e-05, 'epoch': 0.07}) +[2026-10-03 13:12:05,144][starVLA.training.train_starvla][INFO] - Step 943, Loss: {'train/loss': 0.38666069507598877, 'train/grad_norm_pre_clip': 87.47832489013672, 'timing/data': 0.00025224999990314245, 'timing/model': 0.5933221770101227, 'learning_rate/qwen_vl_interface': 9.322938246395336e-06, 'learning_rate/action_model': 9.322938246395337e-05, 'epoch': 0.07}) +[2026-10-03 13:12:05,841][starVLA.training.train_starvla][INFO] - Step 944, Loss: {'train/loss': 4.653095675166696e-05, 'train/grad_norm_pre_clip': 0.016415048390626907, 'timing/data': 0.0003728000010596588, 'timing/model': 0.38866449298802763, 'learning_rate/qwen_vl_interface': 9.321370389337196e-06, 'learning_rate/action_model': 9.321370389337196e-05, 'epoch': 0.07}) +[2026-10-03 13:12:06,537][starVLA.training.train_starvla][INFO] - Step 945, Loss: {'train/loss': 0.002617067424580455, 'train/grad_norm_pre_clip': 0.9112798571586609, 'timing/data': 0.00020238000433892012, 'timing/model': 0.38903351200860925, 'learning_rate/qwen_vl_interface': 9.319800858692478e-06, 'learning_rate/action_model': 9.319800858692476e-05, 'epoch': 0.07}) +[2026-10-03 13:12:07,236][starVLA.training.train_starvla][INFO] - Step 946, Loss: {'train/loss': 0.3414098918437958, 'train/grad_norm_pre_clip': 44.197879791259766, 'timing/data': 0.00031315999513026327, 'timing/model': 0.3914029220031807, 'learning_rate/qwen_vl_interface': 9.318229655106357e-06, 'learning_rate/action_model': 9.318229655106356e-05, 'epoch': 0.07}) +[2026-10-03 13:12:07,937][starVLA.training.train_starvla][INFO] - Step 947, Loss: {'train/loss': 0.00025155951152555645, 'train/grad_norm_pre_clip': 0.045456044375896454, 'timing/data': 0.00037804999738000333, 'timing/model': 0.39132451299519744, 'learning_rate/qwen_vl_interface': 9.316656779224693e-06, 'learning_rate/action_model': 9.316656779224693e-05, 'epoch': 0.07}) +[2026-10-03 13:12:08,827][starVLA.training.train_starvla][INFO] - Step 948, Loss: {'train/loss': 0.00010181344259763137, 'train/grad_norm_pre_clip': 0.030315883457660675, 'timing/data': 0.00023902999237179756, 'timing/model': 0.582686799010844, 'learning_rate/qwen_vl_interface': 9.31508223169404e-06, 'learning_rate/action_model': 9.31508223169404e-05, 'epoch': 0.07}) +[2026-10-03 13:12:09,522][starVLA.training.train_starvla][INFO] - Step 949, Loss: {'train/loss': 0.0006564853247255087, 'train/grad_norm_pre_clip': 0.3037755787372589, 'timing/data': 0.00028601998928934336, 'timing/model': 0.3893004630081123, 'learning_rate/qwen_vl_interface': 9.313506013161633e-06, 'learning_rate/action_model': 9.313506013161633e-05, 'epoch': 0.07}) +[2026-10-03 13:12:10,220][starVLA.training.train_starvla][INFO] - Step 950, Loss: {'train/loss': 0.08338741213083267, 'train/grad_norm_pre_clip': 106.04986572265625, 'timing/data': 0.00034783899900503457, 'timing/model': 0.3907746129989391, 'learning_rate/qwen_vl_interface': 9.311928124275396e-06, 'learning_rate/action_model': 9.311928124275395e-05, 'epoch': 0.07}) +[2026-10-03 13:12:10,920][starVLA.training.train_starvla][INFO] - Step 951, Loss: {'train/loss': 0.0003715890343300998, 'train/grad_norm_pre_clip': 0.4036674201488495, 'timing/data': 0.0003573300054995343, 'timing/model': 0.3909429920022376, 'learning_rate/qwen_vl_interface': 9.310348565683938e-06, 'learning_rate/action_model': 9.310348565683939e-05, 'epoch': 0.07}) +[2026-10-03 13:12:11,616][starVLA.training.train_starvla][INFO] - Step 952, Loss: {'train/loss': 0.7250608801841736, 'train/grad_norm_pre_clip': 120.34762573242188, 'timing/data': 0.00031165999826043844, 'timing/model': 0.38933498298865743, 'learning_rate/qwen_vl_interface': 9.30876733803656e-06, 'learning_rate/action_model': 9.308767338036559e-05, 'epoch': 0.07}) +[2026-10-03 13:12:12,311][starVLA.training.train_starvla][INFO] - Step 953, Loss: {'train/loss': 0.4169948101043701, 'train/grad_norm_pre_clip': 27.853199005126953, 'timing/data': 0.0004353800031822175, 'timing/model': 0.38914567200117745, 'learning_rate/qwen_vl_interface': 9.30718444198324e-06, 'learning_rate/action_model': 9.307184441983239e-05, 'epoch': 0.07}) +[2026-10-03 13:12:13,203][starVLA.training.train_starvla][INFO] - Step 954, Loss: {'train/loss': 0.41099533438682556, 'train/grad_norm_pre_clip': 106.7654800415039, 'timing/data': 0.0002532399957999587, 'timing/model': 0.3909932129899971, 'learning_rate/qwen_vl_interface': 9.30559987817465e-06, 'learning_rate/action_model': 9.30559987817465e-05, 'epoch': 0.07}) +[2026-10-03 13:12:13,902][starVLA.training.train_starvla][INFO] - Step 955, Loss: {'train/loss': 0.0009200540371239185, 'train/grad_norm_pre_clip': 1.7270456552505493, 'timing/data': 0.00029207998886704445, 'timing/model': 0.39066108300175983, 'learning_rate/qwen_vl_interface': 9.304013647262143e-06, 'learning_rate/action_model': 9.304013647262143e-05, 'epoch': 0.07}) +[2026-10-03 13:12:14,598][starVLA.training.train_starvla][INFO] - Step 956, Loss: {'train/loss': 0.006020680069923401, 'train/grad_norm_pre_clip': 10.242986679077148, 'timing/data': 0.000351590002537705, 'timing/model': 0.38921087299240753, 'learning_rate/qwen_vl_interface': 9.302425749897761e-06, 'learning_rate/action_model': 9.302425749897761e-05, 'epoch': 0.07}) +[2026-10-03 13:12:15,295][starVLA.training.train_starvla][INFO] - Step 957, Loss: {'train/loss': 0.43887224793434143, 'train/grad_norm_pre_clip': 21.05337142944336, 'timing/data': 0.000255279999691993, 'timing/model': 0.3901953430031426, 'learning_rate/qwen_vl_interface': 9.300836186734225e-06, 'learning_rate/action_model': 9.300836186734226e-05, 'epoch': 0.07}) +[2026-10-03 13:12:15,993][starVLA.training.train_starvla][INFO] - Step 958, Loss: {'train/loss': 0.5804297924041748, 'train/grad_norm_pre_clip': 29.597248077392578, 'timing/data': 0.00022844999330118299, 'timing/model': 0.3909317610086873, 'learning_rate/qwen_vl_interface': 9.299244958424947e-06, 'learning_rate/action_model': 9.299244958424946e-05, 'epoch': 0.07}) +[2026-10-03 13:12:16,885][starVLA.training.train_starvla][INFO] - Step 959, Loss: {'train/loss': 0.0004703174054156989, 'train/grad_norm_pre_clip': 0.1468537151813507, 'timing/data': 0.00020303999190218747, 'timing/model': 0.5842057089903392, 'learning_rate/qwen_vl_interface': 9.29765206562402e-06, 'learning_rate/action_model': 9.297652065624018e-05, 'epoch': 0.07}) +[2026-10-03 13:12:17,580][starVLA.training.train_starvla][INFO] - Step 960, Loss: {'train/loss': 0.05988357961177826, 'train/grad_norm_pre_clip': 132.96450805664062, 'timing/data': 0.0002567390038166195, 'timing/model': 0.3893531229987275, 'learning_rate/qwen_vl_interface': 9.296057508986221e-06, 'learning_rate/action_model': 9.296057508986221e-05, 'epoch': 0.07}) +[2026-10-03 13:12:18,274][starVLA.training.train_starvla][INFO] - Step 961, Loss: {'train/loss': 0.0001915050088427961, 'train/grad_norm_pre_clip': 0.23299317061901093, 'timing/data': 0.0003650500002549961, 'timing/model': 0.3879486030054977, 'learning_rate/qwen_vl_interface': 9.294461289167013e-06, 'learning_rate/action_model': 9.294461289167014e-05, 'epoch': 0.07}) +[2026-10-03 13:12:18,968][starVLA.training.train_starvla][INFO] - Step 962, Loss: {'train/loss': 0.20070749521255493, 'train/grad_norm_pre_clip': 151.02188110351562, 'timing/data': 0.00019858000450767577, 'timing/model': 0.3888335520023247, 'learning_rate/qwen_vl_interface': 9.292863406822543e-06, 'learning_rate/action_model': 9.292863406822542e-05, 'epoch': 0.07}) +[2026-10-03 13:12:19,668][starVLA.training.train_starvla][INFO] - Step 963, Loss: {'train/loss': 0.9382970333099365, 'train/grad_norm_pre_clip': 30.86210060119629, 'timing/data': 0.0002536699903430417, 'timing/model': 0.3916001230099937, 'learning_rate/qwen_vl_interface': 9.291263862609638e-06, 'learning_rate/action_model': 9.291263862609638e-05, 'epoch': 0.07}) +[2026-10-03 13:12:20,364][starVLA.training.train_starvla][INFO] - Step 964, Loss: {'train/loss': 0.12382449954748154, 'train/grad_norm_pre_clip': 123.69644927978516, 'timing/data': 0.000275869999313727, 'timing/model': 0.38952462299494073, 'learning_rate/qwen_vl_interface': 9.28966265718581e-06, 'learning_rate/action_model': 9.28966265718581e-05, 'epoch': 0.07}) +[2026-10-03 13:12:21,257][starVLA.training.train_starvla][INFO] - Step 965, Loss: {'train/loss': 0.008494753390550613, 'train/grad_norm_pre_clip': 2.073284149169922, 'timing/data': 0.00022759998682886362, 'timing/model': 0.39036979200318456, 'learning_rate/qwen_vl_interface': 9.288059791209256e-06, 'learning_rate/action_model': 9.288059791209256e-05, 'epoch': 0.07}) +[2026-10-03 13:12:21,957][starVLA.training.train_starvla][INFO] - Step 966, Loss: {'train/loss': 0.5611034631729126, 'train/grad_norm_pre_clip': 39.145843505859375, 'timing/data': 0.0003059799928450957, 'timing/model': 0.391723622000427, 'learning_rate/qwen_vl_interface': 9.286455265338853e-06, 'learning_rate/action_model': 9.286455265338853e-05, 'epoch': 0.07}) +[2026-10-03 13:12:22,658][starVLA.training.train_starvla][INFO] - Step 967, Loss: {'train/loss': 0.0016289388295263052, 'train/grad_norm_pre_clip': 0.36988407373428345, 'timing/data': 0.0002296799939358607, 'timing/model': 0.3915988719963934, 'learning_rate/qwen_vl_interface': 9.284849080234161e-06, 'learning_rate/action_model': 9.28484908023416e-05, 'epoch': 0.07}) +[2026-10-03 13:12:23,353][starVLA.training.train_starvla][INFO] - Step 968, Loss: {'train/loss': 0.0005381428636610508, 'train/grad_norm_pre_clip': 0.12948855757713318, 'timing/data': 0.0003917700087185949, 'timing/model': 0.38813706299697515, 'learning_rate/qwen_vl_interface': 9.283241236555421e-06, 'learning_rate/action_model': 9.283241236555421e-05, 'epoch': 0.07}) +[2026-10-03 13:12:24,051][starVLA.training.train_starvla][INFO] - Step 969, Loss: {'train/loss': 0.2086687684059143, 'train/grad_norm_pre_clip': 11.4512300491333, 'timing/data': 0.0003520600002957508, 'timing/model': 0.3912588530074572, 'learning_rate/qwen_vl_interface': 9.281631734963557e-06, 'learning_rate/action_model': 9.281631734963557e-05, 'epoch': 0.07}) +[2026-10-03 13:12:24,945][starVLA.training.train_starvla][INFO] - Step 970, Loss: {'train/loss': 0.0038775107823312283, 'train/grad_norm_pre_clip': 0.6891558766365051, 'timing/data': 0.0003159689949825406, 'timing/model': 0.3909860229905462, 'learning_rate/qwen_vl_interface': 9.280020576120176e-06, 'learning_rate/action_model': 9.280020576120176e-05, 'epoch': 0.07}) +[2026-10-03 13:12:25,646][starVLA.training.train_starvla][INFO] - Step 971, Loss: {'train/loss': 0.37510809302330017, 'train/grad_norm_pre_clip': 237.4692840576172, 'timing/data': 0.0003841700090561062, 'timing/model': 0.3911405129911145, 'learning_rate/qwen_vl_interface': 9.278407760687562e-06, 'learning_rate/action_model': 9.278407760687562e-05, 'epoch': 0.07}) +[2026-10-03 13:12:26,339][starVLA.training.train_starvla][INFO] - Step 972, Loss: {'train/loss': 0.0034501361660659313, 'train/grad_norm_pre_clip': 1.4590169191360474, 'timing/data': 0.0003019800060428679, 'timing/model': 0.3876779520069249, 'learning_rate/qwen_vl_interface': 9.276793289328686e-06, 'learning_rate/action_model': 9.276793289328685e-05, 'epoch': 0.07}) +[2026-10-03 13:12:27,034][starVLA.training.train_starvla][INFO] - Step 973, Loss: {'train/loss': 0.0045051416382193565, 'train/grad_norm_pre_clip': 2.973646879196167, 'timing/data': 0.0002519799891160801, 'timing/model': 0.3889746730128536, 'learning_rate/qwen_vl_interface': 9.27517716270719e-06, 'learning_rate/action_model': 9.275177162707191e-05, 'epoch': 0.07}) +[2026-10-03 13:12:27,731][starVLA.training.train_starvla][INFO] - Step 974, Loss: {'train/loss': 0.0769876092672348, 'train/grad_norm_pre_clip': 21.87969207763672, 'timing/data': 0.0002736000024015084, 'timing/model': 0.3910839720047079, 'learning_rate/qwen_vl_interface': 9.273559381487409e-06, 'learning_rate/action_model': 9.273559381487409e-05, 'epoch': 0.07}) +[2026-10-03 13:12:28,621][starVLA.training.train_starvla][INFO] - Step 975, Loss: {'train/loss': 6.245353142730892e-05, 'train/grad_norm_pre_clip': 0.01454156357795, 'timing/data': 0.0002747899998212233, 'timing/model': 0.5824140089971479, 'learning_rate/qwen_vl_interface': 9.27193994633435e-06, 'learning_rate/action_model': 9.27193994633435e-05, 'epoch': 0.07}) +[2026-10-03 13:12:29,315][starVLA.training.train_starvla][INFO] - Step 976, Loss: {'train/loss': 0.000185057389899157, 'train/grad_norm_pre_clip': 0.09543938934803009, 'timing/data': 0.0003636699984781444, 'timing/model': 0.38849822198972106, 'learning_rate/qwen_vl_interface': 9.2703188579137e-06, 'learning_rate/action_model': 9.2703188579137e-05, 'epoch': 0.07}) +[2026-10-03 13:12:30,009][starVLA.training.train_starvla][INFO] - Step 977, Loss: {'train/loss': 0.00216250610537827, 'train/grad_norm_pre_clip': 0.33283060789108276, 'timing/data': 0.00034687000152189285, 'timing/model': 0.38804598299611825, 'learning_rate/qwen_vl_interface': 9.26869611689183e-06, 'learning_rate/action_model': 9.26869611689183e-05, 'epoch': 0.07}) +[2026-10-03 13:12:30,706][starVLA.training.train_starvla][INFO] - Step 978, Loss: {'train/loss': 8.966803579824045e-05, 'train/grad_norm_pre_clip': 0.02304750308394432, 'timing/data': 0.0002960600104415789, 'timing/model': 0.3906907820055494, 'learning_rate/qwen_vl_interface': 9.267071723935784e-06, 'learning_rate/action_model': 9.267071723935784e-05, 'epoch': 0.07}) +[2026-10-03 13:12:31,406][starVLA.training.train_starvla][INFO] - Step 979, Loss: {'train/loss': 1.3894721632823348e-05, 'train/grad_norm_pre_clip': 0.002465939847752452, 'timing/data': 0.0002673700073501095, 'timing/model': 0.39053402200806886, 'learning_rate/qwen_vl_interface': 9.26544567971329e-06, 'learning_rate/action_model': 9.26544567971329e-05, 'epoch': 0.07}) +[2026-10-03 13:12:32,100][starVLA.training.train_starvla][INFO] - Step 980, Loss: {'train/loss': 0.0003122369234915823, 'train/grad_norm_pre_clip': 0.1847679615020752, 'timing/data': 0.00023830999271012843, 'timing/model': 0.38871291300165467, 'learning_rate/qwen_vl_interface': 9.263817984892756e-06, 'learning_rate/action_model': 9.263817984892757e-05, 'epoch': 0.07}) +[2026-10-03 13:12:32,988][starVLA.training.train_starvla][INFO] - Step 981, Loss: {'train/loss': 0.22441016137599945, 'train/grad_norm_pre_clip': 210.33026123046875, 'timing/data': 0.00034315000812057406, 'timing/model': 0.3885736719967099, 'learning_rate/qwen_vl_interface': 9.262188640143262e-06, 'learning_rate/action_model': 9.262188640143262e-05, 'epoch': 0.07}) +[2026-10-03 13:12:33,689][starVLA.training.train_starvla][INFO] - Step 982, Loss: {'train/loss': 0.2356363981962204, 'train/grad_norm_pre_clip': 10.381514549255371, 'timing/data': 0.0002474100037943572, 'timing/model': 0.3937681820098078, 'learning_rate/qwen_vl_interface': 9.260557646134572e-06, 'learning_rate/action_model': 9.260557646134572e-05, 'epoch': 0.07}) +[2026-10-03 13:12:34,391][starVLA.training.train_starvla][INFO] - Step 983, Loss: {'train/loss': 0.574492871761322, 'train/grad_norm_pre_clip': 128.6433563232422, 'timing/data': 0.0002979400014737621, 'timing/model': 0.3914576919924002, 'learning_rate/qwen_vl_interface': 9.258925003537127e-06, 'learning_rate/action_model': 9.258925003537126e-05, 'epoch': 0.07}) +[2026-10-03 13:12:35,085][starVLA.training.train_starvla][INFO] - Step 984, Loss: {'train/loss': 1.5064649232954253e-05, 'train/grad_norm_pre_clip': 0.005811233073472977, 'timing/data': 0.0003129399992758408, 'timing/model': 0.3885566429962637, 'learning_rate/qwen_vl_interface': 9.257290713022041e-06, 'learning_rate/action_model': 9.257290713022041e-05, 'epoch': 0.07}) +[2026-10-03 13:12:35,782][starVLA.training.train_starvla][INFO] - Step 985, Loss: {'train/loss': 0.2969368100166321, 'train/grad_norm_pre_clip': 252.9886474609375, 'timing/data': 0.0002293399884365499, 'timing/model': 0.39051861299958546, 'learning_rate/qwen_vl_interface': 9.255654775261112e-06, 'learning_rate/action_model': 9.255654775261112e-05, 'epoch': 0.07}) +[2026-10-03 13:12:36,674][starVLA.training.train_starvla][INFO] - Step 986, Loss: {'train/loss': 0.6857985854148865, 'train/grad_norm_pre_clip': 346.4725036621094, 'timing/data': 0.0003075499989790842, 'timing/model': 0.3910261730052298, 'learning_rate/qwen_vl_interface': 9.254017190926813e-06, 'learning_rate/action_model': 9.254017190926813e-05, 'epoch': 0.07}) +[2026-10-03 13:12:37,373][starVLA.training.train_starvla][INFO] - Step 987, Loss: {'train/loss': 5.379935464588925e-05, 'train/grad_norm_pre_clip': 0.032622773200273514, 'timing/data': 0.00023299999884329736, 'timing/model': 0.39057861299079377, 'learning_rate/qwen_vl_interface': 9.25237796069229e-06, 'learning_rate/action_model': 9.25237796069229e-05, 'epoch': 0.07}) +[2026-10-03 13:12:38,067][starVLA.training.train_starvla][INFO] - Step 988, Loss: {'train/loss': 1.3060199307801668e-05, 'train/grad_norm_pre_clip': 0.00634712865576148, 'timing/data': 0.00032293998810928315, 'timing/model': 0.3876950130070327, 'learning_rate/qwen_vl_interface': 9.250737085231371e-06, 'learning_rate/action_model': 9.250737085231372e-05, 'epoch': 0.07}) +[2026-10-03 13:12:38,761][starVLA.training.train_starvla][INFO] - Step 989, Loss: {'train/loss': 0.7296770811080933, 'train/grad_norm_pre_clip': 162.39913940429688, 'timing/data': 0.0002509399928385392, 'timing/model': 0.3887779430078808, 'learning_rate/qwen_vl_interface': 9.24909456521856e-06, 'learning_rate/action_model': 9.249094565218559e-05, 'epoch': 0.07}) +[2026-10-03 13:12:39,461][starVLA.training.train_starvla][INFO] - Step 990, Loss: {'train/loss': 0.3141939043998718, 'train/grad_norm_pre_clip': 129.06654357910156, 'timing/data': 0.0003090599930146709, 'timing/model': 0.39004493301035836, 'learning_rate/qwen_vl_interface': 9.247450401329027e-06, 'learning_rate/action_model': 9.247450401329027e-05, 'epoch': 0.07}) +[2026-10-03 13:12:40,352][starVLA.training.train_starvla][INFO] - Step 991, Loss: {'train/loss': 0.0003168615512549877, 'train/grad_norm_pre_clip': 0.574241042137146, 'timing/data': 0.0003298899973742664, 'timing/model': 0.5831143190007424, 'learning_rate/qwen_vl_interface': 9.245804594238637e-06, 'learning_rate/action_model': 9.245804594238636e-05, 'epoch': 0.07}) +[2026-10-03 13:12:41,046][starVLA.training.train_starvla][INFO] - Step 992, Loss: {'train/loss': 0.4654637575149536, 'train/grad_norm_pre_clip': 71.32713317871094, 'timing/data': 0.0001988099975278601, 'timing/model': 0.38815700299164746, 'learning_rate/qwen_vl_interface': 9.244157144623909e-06, 'learning_rate/action_model': 9.244157144623909e-05, 'epoch': 0.07}) +[2026-10-03 13:12:41,740][starVLA.training.train_starvla][INFO] - Step 993, Loss: {'train/loss': 0.12642589211463928, 'train/grad_norm_pre_clip': 209.560302734375, 'timing/data': 0.00019516999600455165, 'timing/model': 0.3887910430057673, 'learning_rate/qwen_vl_interface': 9.242508053162051e-06, 'learning_rate/action_model': 9.242508053162051e-05, 'epoch': 0.07}) +[2026-10-03 13:12:42,435][starVLA.training.train_starvla][INFO] - Step 994, Loss: {'train/loss': 0.03667796403169632, 'train/grad_norm_pre_clip': 5.0678300857543945, 'timing/data': 0.00025288001052103937, 'timing/model': 0.3899597220006399, 'learning_rate/qwen_vl_interface': 9.240857320530944e-06, 'learning_rate/action_model': 9.240857320530943e-05, 'epoch': 0.07}) +[2026-10-03 13:12:43,133][starVLA.training.train_starvla][INFO] - Step 995, Loss: {'train/loss': 0.00136966269928962, 'train/grad_norm_pre_clip': 4.325809001922607, 'timing/data': 0.0002906599984271452, 'timing/model': 0.38989331299671903, 'learning_rate/qwen_vl_interface': 9.23920494740914e-06, 'learning_rate/action_model': 9.23920494740914e-05, 'epoch': 0.07}) +[2026-10-03 13:12:44,021][starVLA.training.train_starvla][INFO] - Step 996, Loss: {'train/loss': 0.0009590562549419701, 'train/grad_norm_pre_clip': 1.678108811378479, 'timing/data': 0.00019334000535309315, 'timing/model': 0.582032150006853, 'learning_rate/qwen_vl_interface': 9.237550934475868e-06, 'learning_rate/action_model': 9.237550934475867e-05, 'epoch': 0.07}) +[2026-10-03 13:12:44,716][starVLA.training.train_starvla][INFO] - Step 997, Loss: {'train/loss': 0.13480912148952484, 'train/grad_norm_pre_clip': 57.739044189453125, 'timing/data': 0.00019245001021772623, 'timing/model': 0.3893660719913896, 'learning_rate/qwen_vl_interface': 9.235895282411029e-06, 'learning_rate/action_model': 9.235895282411029e-05, 'epoch': 0.07}) +[2026-10-03 13:12:45,412][starVLA.training.train_starvla][INFO] - Step 998, Loss: {'train/loss': 4.73108229925856e-06, 'train/grad_norm_pre_clip': 0.0013095592148602009, 'timing/data': 0.0002135499962605536, 'timing/model': 0.3903469830111135, 'learning_rate/qwen_vl_interface': 9.2342379918952e-06, 'learning_rate/action_model': 9.2342379918952e-05, 'epoch': 0.07}) +[2026-10-03 13:12:46,112][starVLA.training.train_starvla][INFO] - Step 999, Loss: {'train/loss': 0.47216537594795227, 'train/grad_norm_pre_clip': 172.25001525878906, 'timing/data': 0.00020963999850209802, 'timing/model': 0.3910465020017, 'learning_rate/qwen_vl_interface': 9.232579063609631e-06, 'learning_rate/action_model': 9.232579063609631e-05, 'epoch': 0.07}) +[2026-10-03 13:12:46,807][starVLA.training.train_starvla][INFO] - Step 1000, Loss: {'train/loss': 0.5788588523864746, 'train/grad_norm_pre_clip': 128.71510314941406, 'timing/data': 0.00024947000201791525, 'timing/model': 0.38846746300987434, 'learning_rate/qwen_vl_interface': 9.230918498236243e-06, 'learning_rate/action_model': 9.230918498236244e-05, 'epoch': 0.07}) +[2026-10-03 13:12:47,692][starVLA.training.train_starvla][INFO] - Step 1001, Loss: {'train/loss': 0.11503791064023972, 'train/grad_norm_pre_clip': 95.20259857177734, 'timing/data': 0.00029552000341936946, 'timing/model': 0.579618979987572, 'learning_rate/qwen_vl_interface': 9.229256296457635e-06, 'learning_rate/action_model': 9.229256296457634e-05, 'epoch': 0.07}) +[2026-10-03 13:12:48,389][starVLA.training.train_starvla][INFO] - Step 1002, Loss: {'train/loss': 0.0033508646301925182, 'train/grad_norm_pre_clip': 0.5592294335365295, 'timing/data': 0.00034887999936472625, 'timing/model': 0.39057469299586955, 'learning_rate/qwen_vl_interface': 9.22759245895707e-06, 'learning_rate/action_model': 9.227592458957071e-05, 'epoch': 0.07}) +[2026-10-03 13:12:49,086][starVLA.training.train_starvla][INFO] - Step 1003, Loss: {'train/loss': 0.672398567199707, 'train/grad_norm_pre_clip': 74.09178161621094, 'timing/data': 0.0003391799982637167, 'timing/model': 0.3896280030021444, 'learning_rate/qwen_vl_interface': 9.225926986418494e-06, 'learning_rate/action_model': 9.225926986418494e-05, 'epoch': 0.07}) +[2026-10-03 13:12:49,781][starVLA.training.train_starvla][INFO] - Step 1004, Loss: {'train/loss': 0.006133557762950659, 'train/grad_norm_pre_clip': 10.759671211242676, 'timing/data': 0.0002043699932983145, 'timing/model': 0.38936273200670257, 'learning_rate/qwen_vl_interface': 9.224259879526519e-06, 'learning_rate/action_model': 9.224259879526519e-05, 'epoch': 0.07}) +[2026-10-03 13:12:50,475][starVLA.training.train_starvla][INFO] - Step 1005, Loss: {'train/loss': 0.002716951770707965, 'train/grad_norm_pre_clip': 1.8575725555419922, 'timing/data': 0.00030975999834481627, 'timing/model': 0.3881288930133451, 'learning_rate/qwen_vl_interface': 9.222591138966427e-06, 'learning_rate/action_model': 9.222591138966426e-05, 'epoch': 0.07}) +[2026-10-03 13:12:51,173][starVLA.training.train_starvla][INFO] - Step 1006, Loss: {'train/loss': 8.460241224383935e-05, 'train/grad_norm_pre_clip': 0.12285488843917847, 'timing/data': 0.00029363000066950917, 'timing/model': 0.3913412330002757, 'learning_rate/qwen_vl_interface': 9.220920765424175e-06, 'learning_rate/action_model': 9.220920765424174e-05, 'epoch': 0.07}) +[2026-10-03 13:12:52,078][starVLA.training.train_starvla][INFO] - Step 1007, Loss: {'train/loss': 1.127211332321167, 'train/grad_norm_pre_clip': 87.6671142578125, 'timing/data': 0.000387160005629994, 'timing/model': 0.4030159110116074, 'learning_rate/qwen_vl_interface': 9.219248759586392e-06, 'learning_rate/action_model': 9.219248759586391e-05, 'epoch': 0.07}) +[2026-10-03 13:12:52,771][starVLA.training.train_starvla][INFO] - Step 1008, Loss: {'train/loss': 0.5593775510787964, 'train/grad_norm_pre_clip': 119.15593719482422, 'timing/data': 0.00034871000389102846, 'timing/model': 0.38779767300002277, 'learning_rate/qwen_vl_interface': 9.217575122140375e-06, 'learning_rate/action_model': 9.217575122140375e-05, 'epoch': 0.07}) +[2026-10-03 13:12:53,464][starVLA.training.train_starvla][INFO] - Step 1009, Loss: {'train/loss': 0.022260421887040138, 'train/grad_norm_pre_clip': 52.73462677001953, 'timing/data': 0.0002915499935625121, 'timing/model': 0.38821516300959047, 'learning_rate/qwen_vl_interface': 9.215899853774094e-06, 'learning_rate/action_model': 9.215899853774093e-05, 'epoch': 0.07}) +[2026-10-03 13:12:54,161][starVLA.training.train_starvla][INFO] - Step 1010, Loss: {'train/loss': 0.314371258020401, 'train/grad_norm_pre_clip': 46.62065124511719, 'timing/data': 0.0003042299940716475, 'timing/model': 0.39024487300775945, 'learning_rate/qwen_vl_interface': 9.214222955176189e-06, 'learning_rate/action_model': 9.214222955176189e-05, 'epoch': 0.07}) +[2026-10-03 13:12:54,872][starVLA.training.train_starvla][INFO] - Step 1011, Loss: {'train/loss': 0.0001595066860318184, 'train/grad_norm_pre_clip': 0.07427255809307098, 'timing/data': 0.00022231000184547156, 'timing/model': 0.40276169100252446, 'learning_rate/qwen_vl_interface': 9.212544427035967e-06, 'learning_rate/action_model': 9.212544427035967e-05, 'epoch': 0.07}) +[2026-10-03 13:12:55,759][starVLA.training.train_starvla][INFO] - Step 1012, Loss: {'train/loss': 0.003252386348322034, 'train/grad_norm_pre_clip': 4.526767730712891, 'timing/data': 0.0003100000030826777, 'timing/model': 0.5814430589962285, 'learning_rate/qwen_vl_interface': 9.21086427004341e-06, 'learning_rate/action_model': 9.210864270043411e-05, 'epoch': 0.07}) +[2026-10-03 13:12:56,432][starVLA.training.train_starvla][INFO] - Step 1013, Loss: {'train/loss': 0.0021466214675456285, 'train/grad_norm_pre_clip': 1.3747888803482056, 'timing/data': 0.00026652999804355204, 'timing/model': 0.3883130039903335, 'learning_rate/qwen_vl_interface': 9.209182484889168e-06, 'learning_rate/action_model': 9.209182484889167e-05, 'epoch': 0.07}) +[2026-10-03 13:12:57,127][starVLA.training.train_starvla][INFO] - Step 1014, Loss: {'train/loss': 0.4223034381866455, 'train/grad_norm_pre_clip': 69.02690887451172, 'timing/data': 0.0002971600042656064, 'timing/model': 0.39023459200689103, 'learning_rate/qwen_vl_interface': 9.207499072264557e-06, 'learning_rate/action_model': 9.207499072264558e-05, 'epoch': 0.07}) +[2026-10-03 13:12:57,827][starVLA.training.train_starvla][INFO] - Step 1015, Loss: {'train/loss': 0.2521601617336273, 'train/grad_norm_pre_clip': 22.25909996032715, 'timing/data': 0.0002858499938156456, 'timing/model': 0.391694991994882, 'learning_rate/qwen_vl_interface': 9.205814032861566e-06, 'learning_rate/action_model': 9.205814032861566e-05, 'epoch': 0.07}) +[2026-10-03 13:12:58,524][starVLA.training.train_starvla][INFO] - Step 1016, Loss: {'train/loss': 0.00041431363206356764, 'train/grad_norm_pre_clip': 0.3117096722126007, 'timing/data': 0.0003557700110832229, 'timing/model': 0.38928673199552577, 'learning_rate/qwen_vl_interface': 9.204127367372849e-06, 'learning_rate/action_model': 9.204127367372849e-05, 'epoch': 0.07}) +[2026-10-03 13:12:59,218][starVLA.training.train_starvla][INFO] - Step 1017, Loss: {'train/loss': 0.13538946211338043, 'train/grad_norm_pre_clip': 97.94097900390625, 'timing/data': 0.00026654000976122916, 'timing/model': 0.38831140199908987, 'learning_rate/qwen_vl_interface': 9.202439076491732e-06, 'learning_rate/action_model': 9.202439076491732e-05, 'epoch': 0.07}) +[2026-10-03 13:13:00,109][starVLA.training.train_starvla][INFO] - Step 1018, Loss: {'train/loss': 0.05035576596856117, 'train/grad_norm_pre_clip': 43.579246520996094, 'timing/data': 0.00020049000158905983, 'timing/model': 0.39044779300456867, 'learning_rate/qwen_vl_interface': 9.200749160912208e-06, 'learning_rate/action_model': 9.200749160912208e-05, 'epoch': 0.07}) +[2026-10-03 13:13:00,806][starVLA.training.train_starvla][INFO] - Step 1019, Loss: {'train/loss': 0.0013851462863385677, 'train/grad_norm_pre_clip': 0.2843337059020996, 'timing/data': 0.00019950000569224358, 'timing/model': 0.3902286520024063, 'learning_rate/qwen_vl_interface': 9.199057621328937e-06, 'learning_rate/action_model': 9.199057621328935e-05, 'epoch': 0.07}) +[2026-10-03 13:13:01,500][starVLA.training.train_starvla][INFO] - Step 1020, Loss: {'train/loss': 0.12988416850566864, 'train/grad_norm_pre_clip': 63.67813491821289, 'timing/data': 0.00034127999970223755, 'timing/model': 0.3879245139978593, 'learning_rate/qwen_vl_interface': 9.197364458437244e-06, 'learning_rate/action_model': 9.197364458437245e-05, 'epoch': 0.07}) +[2026-10-03 13:13:02,195][starVLA.training.train_starvla][INFO] - Step 1021, Loss: {'train/loss': 0.01028767041862011, 'train/grad_norm_pre_clip': 7.028936862945557, 'timing/data': 0.00019428000086918473, 'timing/model': 0.38826018299732823, 'learning_rate/qwen_vl_interface': 9.19566967293313e-06, 'learning_rate/action_model': 9.19566967293313e-05, 'epoch': 0.07}) +[2026-10-03 13:13:02,892][starVLA.training.train_starvla][INFO] - Step 1022, Loss: {'train/loss': 0.002225463045760989, 'train/grad_norm_pre_clip': 2.2030913829803467, 'timing/data': 0.0002570200012996793, 'timing/model': 0.3907935520110186, 'learning_rate/qwen_vl_interface': 9.193973265513253e-06, 'learning_rate/action_model': 9.193973265513253e-05, 'epoch': 0.07}) +[2026-10-03 13:13:03,783][starVLA.training.train_starvla][INFO] - Step 1023, Loss: {'train/loss': 0.3236125409603119, 'train/grad_norm_pre_clip': 21.427244186401367, 'timing/data': 0.0002726600068854168, 'timing/model': 0.38887700300256256, 'learning_rate/qwen_vl_interface': 9.192275236874946e-06, 'learning_rate/action_model': 9.192275236874946e-05, 'epoch': 0.07}) +[2026-10-03 13:13:04,477][starVLA.training.train_starvla][INFO] - Step 1024, Loss: {'train/loss': 0.0022316407412290573, 'train/grad_norm_pre_clip': 2.130505084991455, 'timing/data': 0.0002508000034140423, 'timing/model': 0.38806851200934034, 'learning_rate/qwen_vl_interface': 9.190575587716199e-06, 'learning_rate/action_model': 9.190575587716199e-05, 'epoch': 0.07}) +[2026-10-03 13:13:05,171][starVLA.training.train_starvla][INFO] - Step 1025, Loss: {'train/loss': 0.002269129501655698, 'train/grad_norm_pre_clip': 0.7724575400352478, 'timing/data': 0.00029573000210803, 'timing/model': 0.3885094240104081, 'learning_rate/qwen_vl_interface': 9.188874318735677e-06, 'learning_rate/action_model': 9.188874318735676e-05, 'epoch': 0.08}) +[2026-10-03 13:13:05,867][starVLA.training.train_starvla][INFO] - Step 1026, Loss: {'train/loss': 0.0007502884254790843, 'train/grad_norm_pre_clip': 0.3748706877231598, 'timing/data': 0.0002459599927533418, 'timing/model': 0.3903691029991023, 'learning_rate/qwen_vl_interface': 9.187171430632707e-06, 'learning_rate/action_model': 9.187171430632707e-05, 'epoch': 0.08}) +[2026-10-03 13:13:06,567][starVLA.training.train_starvla][INFO] - Step 1027, Loss: {'train/loss': 0.0003756906953640282, 'train/grad_norm_pre_clip': 0.1528717279434204, 'timing/data': 0.0002612899988889694, 'timing/model': 0.3915630230039824, 'learning_rate/qwen_vl_interface': 9.185466924107283e-06, 'learning_rate/action_model': 9.185466924107282e-05, 'epoch': 0.08}) +[2026-10-03 13:13:07,456][starVLA.training.train_starvla][INFO] - Step 1028, Loss: {'train/loss': 0.29236164689064026, 'train/grad_norm_pre_clip': 101.83973693847656, 'timing/data': 0.00033601000905036926, 'timing/model': 0.5828373089898378, 'learning_rate/qwen_vl_interface': 9.18376079986006e-06, 'learning_rate/action_model': 9.183760799860061e-05, 'epoch': 0.08}) +[2026-10-03 13:13:08,150][starVLA.training.train_starvla][INFO] - Step 1029, Loss: {'train/loss': 0.00011691908002831042, 'train/grad_norm_pre_clip': 0.04712122306227684, 'timing/data': 0.0002500300033716485, 'timing/model': 0.3883620039996458, 'learning_rate/qwen_vl_interface': 9.182053058592365e-06, 'learning_rate/action_model': 9.182053058592365e-05, 'epoch': 0.08}) +[2026-10-03 13:13:08,847][starVLA.training.train_starvla][INFO] - Step 1030, Loss: {'train/loss': 0.0004752931999973953, 'train/grad_norm_pre_clip': 0.8400028944015503, 'timing/data': 0.00029908999567851424, 'timing/model': 0.390514902988798, 'learning_rate/qwen_vl_interface': 9.180343701006183e-06, 'learning_rate/action_model': 9.180343701006183e-05, 'epoch': 0.08}) +[2026-10-03 13:13:09,545][starVLA.training.train_starvla][INFO] - Step 1031, Loss: {'train/loss': 5.732281715609133e-05, 'train/grad_norm_pre_clip': 0.018639715388417244, 'timing/data': 0.0003336099907755852, 'timing/model': 0.39042683300795034, 'learning_rate/qwen_vl_interface': 9.17863272780417e-06, 'learning_rate/action_model': 9.178632727804168e-05, 'epoch': 0.08}) +[2026-10-03 13:13:10,239][starVLA.training.train_starvla][INFO] - Step 1032, Loss: {'train/loss': 0.0013988784048706293, 'train/grad_norm_pre_clip': 0.4427121579647064, 'timing/data': 0.00034656999923754483, 'timing/model': 0.3878648130048532, 'learning_rate/qwen_vl_interface': 9.176920139689638e-06, 'learning_rate/action_model': 9.176920139689637e-05, 'epoch': 0.08}) +[2026-10-03 13:13:11,134][starVLA.training.train_starvla][INFO] - Step 1033, Loss: {'train/loss': 0.20679064095020294, 'train/grad_norm_pre_clip': 212.1358642578125, 'timing/data': 0.0002749800041783601, 'timing/model': 0.5880999590008287, 'learning_rate/qwen_vl_interface': 9.175205937366568e-06, 'learning_rate/action_model': 9.175205937366569e-05, 'epoch': 0.08}) +[2026-10-03 13:13:11,831][starVLA.training.train_starvla][INFO] - Step 1034, Loss: {'train/loss': 1.8543876649346203e-05, 'train/grad_norm_pre_clip': 0.0054160961881279945, 'timing/data': 0.00029579999682027847, 'timing/model': 0.3905567530018743, 'learning_rate/qwen_vl_interface': 9.173490121539608e-06, 'learning_rate/action_model': 9.173490121539608e-05, 'epoch': 0.08}) +[2026-10-03 13:13:12,527][starVLA.training.train_starvla][INFO] - Step 1035, Loss: {'train/loss': 0.40328696370124817, 'train/grad_norm_pre_clip': 123.70682525634766, 'timing/data': 0.00031820000731386244, 'timing/model': 0.3882468230003724, 'learning_rate/qwen_vl_interface': 9.171772692914062e-06, 'learning_rate/action_model': 9.171772692914061e-05, 'epoch': 0.08}) +[2026-10-03 13:13:13,221][starVLA.training.train_starvla][INFO] - Step 1036, Loss: {'train/loss': 1.22484088933561e-05, 'train/grad_norm_pre_clip': 0.007010912522673607, 'timing/data': 0.00033052999060600996, 'timing/model': 0.38846029300475493, 'learning_rate/qwen_vl_interface': 9.1700536521959e-06, 'learning_rate/action_model': 9.170053652195899e-05, 'epoch': 0.08}) +[2026-10-03 13:13:13,916][starVLA.training.train_starvla][INFO] - Step 1037, Loss: {'train/loss': 0.0891796201467514, 'train/grad_norm_pre_clip': 48.40202331542969, 'timing/data': 0.0003032000095117837, 'timing/model': 0.3887050729972543, 'learning_rate/qwen_vl_interface': 9.168333000091754e-06, 'learning_rate/action_model': 9.168333000091754e-05, 'epoch': 0.08}) +[2026-10-03 13:13:14,612][starVLA.training.train_starvla][INFO] - Step 1038, Loss: {'train/loss': 0.04508573189377785, 'train/grad_norm_pre_clip': 40.71430206298828, 'timing/data': 0.0003799499972956255, 'timing/model': 0.3907476130116265, 'learning_rate/qwen_vl_interface': 9.16661073730892e-06, 'learning_rate/action_model': 9.166610737308919e-05, 'epoch': 0.08}) +[2026-10-03 13:13:15,506][starVLA.training.train_starvla][INFO] - Step 1039, Loss: {'train/loss': 0.0003265285340603441, 'train/grad_norm_pre_clip': 0.5498809814453125, 'timing/data': 0.0001999400119530037, 'timing/model': 0.39091667300090194, 'learning_rate/qwen_vl_interface': 9.164886864555355e-06, 'learning_rate/action_model': 9.164886864555355e-05, 'epoch': 0.08}) +[2026-10-03 13:13:16,203][starVLA.training.train_starvla][INFO] - Step 1040, Loss: {'train/loss': 0.0346423014998436, 'train/grad_norm_pre_clip': 120.62203979492188, 'timing/data': 0.0003720700042322278, 'timing/model': 0.3893047729943646, 'learning_rate/qwen_vl_interface': 9.163161382539678e-06, 'learning_rate/action_model': 9.163161382539678e-05, 'epoch': 0.08}) +[2026-10-03 13:13:16,900][starVLA.training.train_starvla][INFO] - Step 1041, Loss: {'train/loss': 0.39583563804626465, 'train/grad_norm_pre_clip': 40.58137893676758, 'timing/data': 0.0003929899976355955, 'timing/model': 0.38884377300564665, 'learning_rate/qwen_vl_interface': 9.16143429197117e-06, 'learning_rate/action_model': 9.16143429197117e-05, 'epoch': 0.08}) +[2026-10-03 13:13:17,597][starVLA.training.train_starvla][INFO] - Step 1042, Loss: {'train/loss': 0.00017133846995420754, 'train/grad_norm_pre_clip': 0.0621127113699913, 'timing/data': 0.0002462700067553669, 'timing/model': 0.3908761220081942, 'learning_rate/qwen_vl_interface': 9.159705593559773e-06, 'learning_rate/action_model': 9.159705593559773e-05, 'epoch': 0.08}) +[2026-10-03 13:13:18,296][starVLA.training.train_starvla][INFO] - Step 1043, Loss: {'train/loss': 0.889258623123169, 'train/grad_norm_pre_clip': 185.9079132080078, 'timing/data': 0.00032021000515669584, 'timing/model': 0.3906203320075292, 'learning_rate/qwen_vl_interface': 9.157975288016089e-06, 'learning_rate/action_model': 9.157975288016088e-05, 'epoch': 0.08}) +[2026-10-03 13:13:19,184][starVLA.training.train_starvla][INFO] - Step 1044, Loss: {'train/loss': 0.001472417963668704, 'train/grad_norm_pre_clip': 0.9881846904754639, 'timing/data': 0.0003017000126419589, 'timing/model': 0.5815064099879237, 'learning_rate/qwen_vl_interface': 9.156243376051381e-06, 'learning_rate/action_model': 9.156243376051381e-05, 'epoch': 0.08}) +[2026-10-03 13:13:19,881][starVLA.training.train_starvla][INFO] - Step 1045, Loss: {'train/loss': 0.0005979050183668733, 'train/grad_norm_pre_clip': 0.33681023120880127, 'timing/data': 0.0003347400052007288, 'timing/model': 0.3891449829970952, 'learning_rate/qwen_vl_interface': 9.154509858377574e-06, 'learning_rate/action_model': 9.154509858377574e-05, 'epoch': 0.08}) +[2026-10-03 13:13:20,581][starVLA.training.train_starvla][INFO] - Step 1046, Loss: {'train/loss': 0.34102967381477356, 'train/grad_norm_pre_clip': 217.6781768798828, 'timing/data': 0.000199249989236705, 'timing/model': 0.39238634299545083, 'learning_rate/qwen_vl_interface': 9.152774735707248e-06, 'learning_rate/action_model': 9.152774735707247e-05, 'epoch': 0.08}) +[2026-10-03 13:13:21,280][starVLA.training.train_starvla][INFO] - Step 1047, Loss: {'train/loss': 0.1084502711892128, 'train/grad_norm_pre_clip': 149.74066162109375, 'timing/data': 0.00031184000545181334, 'timing/model': 0.3910764320025919, 'learning_rate/qwen_vl_interface': 9.15103800875365e-06, 'learning_rate/action_model': 9.15103800875365e-05, 'epoch': 0.08}) +[2026-10-03 13:13:21,976][starVLA.training.train_starvla][INFO] - Step 1048, Loss: {'train/loss': 1.9378514289855957, 'train/grad_norm_pre_clip': 984.9697265625, 'timing/data': 0.00020447999122552574, 'timing/model': 0.38891355200030375, 'learning_rate/qwen_vl_interface': 9.149299678230681e-06, 'learning_rate/action_model': 9.149299678230681e-05, 'epoch': 0.08}) +[2026-10-03 13:13:22,872][starVLA.training.train_starvla][INFO] - Step 1049, Loss: {'train/loss': 0.41899386048316956, 'train/grad_norm_pre_clip': 24.06073760986328, 'timing/data': 0.0003106499934801832, 'timing/model': 0.5893042579991743, 'learning_rate/qwen_vl_interface': 9.147559744852905e-06, 'learning_rate/action_model': 9.147559744852905e-05, 'epoch': 0.08}) +[2026-10-03 13:13:23,570][starVLA.training.train_starvla][INFO] - Step 1050, Loss: {'train/loss': 0.48698097467422485, 'train/grad_norm_pre_clip': 197.80331420898438, 'timing/data': 0.0002676599979167804, 'timing/model': 0.39226156201038975, 'learning_rate/qwen_vl_interface': 9.14581820933554e-06, 'learning_rate/action_model': 9.14581820933554e-05, 'epoch': 0.08}) +[2026-10-03 13:13:24,271][starVLA.training.train_starvla][INFO] - Step 1051, Loss: {'train/loss': 7.154663762776181e-05, 'train/grad_norm_pre_clip': 0.02088320627808571, 'timing/data': 0.0002654599957168102, 'timing/model': 0.39058767299866304, 'learning_rate/qwen_vl_interface': 9.14407507239447e-06, 'learning_rate/action_model': 9.14407507239447e-05, 'epoch': 0.08}) +[2026-10-03 13:13:24,967][starVLA.training.train_starvla][INFO] - Step 1052, Loss: {'train/loss': 0.00021922352607361972, 'train/grad_norm_pre_clip': 0.08611306548118591, 'timing/data': 0.00020696999854408205, 'timing/model': 0.3890764430107083, 'learning_rate/qwen_vl_interface': 9.142330334746227e-06, 'learning_rate/action_model': 9.142330334746227e-05, 'epoch': 0.08}) +[2026-10-03 13:13:25,664][starVLA.training.train_starvla][INFO] - Step 1053, Loss: {'train/loss': 0.6449202299118042, 'train/grad_norm_pre_clip': 58.616363525390625, 'timing/data': 0.0002680700126802549, 'timing/model': 0.39003532301285304, 'learning_rate/qwen_vl_interface': 9.14058399710801e-06, 'learning_rate/action_model': 9.14058399710801e-05, 'epoch': 0.08}) +[2026-10-03 13:13:26,361][starVLA.training.train_starvla][INFO] - Step 1054, Loss: {'train/loss': 0.23480185866355896, 'train/grad_norm_pre_clip': 114.85700988769531, 'timing/data': 0.00021216999448370188, 'timing/model': 0.3910691330092959, 'learning_rate/qwen_vl_interface': 9.138836060197673e-06, 'learning_rate/action_model': 9.138836060197672e-05, 'epoch': 0.08}) +[2026-10-03 13:13:27,259][starVLA.training.train_starvla][INFO] - Step 1055, Loss: {'train/loss': 0.08519021421670914, 'train/grad_norm_pre_clip': 47.924293518066406, 'timing/data': 0.0002067500026896596, 'timing/model': 0.3910678530082805, 'learning_rate/qwen_vl_interface': 9.137086524733725e-06, 'learning_rate/action_model': 9.137086524733724e-05, 'epoch': 0.08}) +[2026-10-03 13:13:27,956][starVLA.training.train_starvla][INFO] - Step 1056, Loss: {'train/loss': 0.6420445442199707, 'train/grad_norm_pre_clip': 75.32588195800781, 'timing/data': 0.0003125100047327578, 'timing/model': 0.3893095930106938, 'learning_rate/qwen_vl_interface': 9.135335391435336e-06, 'learning_rate/action_model': 9.135335391435335e-05, 'epoch': 0.08}) +[2026-10-03 13:13:28,651][starVLA.training.train_starvla][INFO] - Step 1057, Loss: {'train/loss': 0.0418737567961216, 'train/grad_norm_pre_clip': 27.647586822509766, 'timing/data': 0.0002503200084902346, 'timing/model': 0.3893137929990189, 'learning_rate/qwen_vl_interface': 9.13358266102233e-06, 'learning_rate/action_model': 9.133582661022328e-05, 'epoch': 0.08}) +[2026-10-03 13:13:29,349][starVLA.training.train_starvla][INFO] - Step 1058, Loss: {'train/loss': 0.016243955120444298, 'train/grad_norm_pre_clip': 11.29576301574707, 'timing/data': 0.00023713000700809062, 'timing/model': 0.3915595129947178, 'learning_rate/qwen_vl_interface': 9.131828334215185e-06, 'learning_rate/action_model': 9.131828334215185e-05, 'epoch': 0.08}) +[2026-10-03 13:13:30,050][starVLA.training.train_starvla][INFO] - Step 1059, Loss: {'train/loss': 0.0006290787132456899, 'train/grad_norm_pre_clip': 0.20159178972244263, 'timing/data': 0.00034436999703757465, 'timing/model': 0.39096532300754916, 'learning_rate/qwen_vl_interface': 9.130072411735043e-06, 'learning_rate/action_model': 9.130072411735043e-05, 'epoch': 0.08}) +[2026-10-03 13:13:30,944][starVLA.training.train_starvla][INFO] - Step 1060, Loss: {'train/loss': 0.05360355228185654, 'train/grad_norm_pre_clip': 29.014339447021484, 'timing/data': 0.00020866999693680555, 'timing/model': 0.3890729429986095, 'learning_rate/qwen_vl_interface': 9.128314894303693e-06, 'learning_rate/action_model': 9.128314894303693e-05, 'epoch': 0.08}) +[2026-10-03 13:13:31,641][starVLA.training.train_starvla][INFO] - Step 1061, Loss: {'train/loss': 0.004765850957483053, 'train/grad_norm_pre_clip': 3.97153377532959, 'timing/data': 0.0001980700035346672, 'timing/model': 0.39001636300235987, 'learning_rate/qwen_vl_interface': 9.126555782643592e-06, 'learning_rate/action_model': 9.126555782643591e-05, 'epoch': 0.08}) +[2026-10-03 13:13:32,338][starVLA.training.train_starvla][INFO] - Step 1062, Loss: {'train/loss': 0.34071969985961914, 'train/grad_norm_pre_clip': 39.257354736328125, 'timing/data': 0.0002650600072229281, 'timing/model': 0.3908334029983962, 'learning_rate/qwen_vl_interface': 9.124795077477835e-06, 'learning_rate/action_model': 9.124795077477834e-05, 'epoch': 0.08}) +[2026-10-03 13:13:33,038][starVLA.training.train_starvla][INFO] - Step 1063, Loss: {'train/loss': 0.003979150205850601, 'train/grad_norm_pre_clip': 5.284463882446289, 'timing/data': 0.00025431001267861575, 'timing/model': 0.39108491300430615, 'learning_rate/qwen_vl_interface': 9.123032779530186e-06, 'learning_rate/action_model': 9.123032779530187e-05, 'epoch': 0.08}) +[2026-10-03 13:13:33,732][starVLA.training.train_starvla][INFO] - Step 1064, Loss: {'train/loss': 0.00137920246925205, 'train/grad_norm_pre_clip': 1.3289096355438232, 'timing/data': 0.0004022700013592839, 'timing/model': 0.3879724139987957, 'learning_rate/qwen_vl_interface': 9.12126888952506e-06, 'learning_rate/action_model': 9.12126888952506e-05, 'epoch': 0.08}) +[2026-10-03 13:13:34,622][starVLA.training.train_starvla][INFO] - Step 1065, Loss: {'train/loss': 0.001344868796877563, 'train/grad_norm_pre_clip': 0.9860429167747498, 'timing/data': 0.0003252199967391789, 'timing/model': 0.38868472298781853, 'learning_rate/qwen_vl_interface': 9.119503408187521e-06, 'learning_rate/action_model': 9.119503408187522e-05, 'epoch': 0.08}) +[2026-10-03 13:13:35,318][starVLA.training.train_starvla][INFO] - Step 1066, Loss: {'train/loss': 0.05603731423616409, 'train/grad_norm_pre_clip': 46.23384475708008, 'timing/data': 0.0002417799987597391, 'timing/model': 0.3909647219988983, 'learning_rate/qwen_vl_interface': 9.117736336243299e-06, 'learning_rate/action_model': 9.117736336243298e-05, 'epoch': 0.08}) +[2026-10-03 13:13:36,016][starVLA.training.train_starvla][INFO] - Step 1067, Loss: {'train/loss': 0.5399509072303772, 'train/grad_norm_pre_clip': 80.03966522216797, 'timing/data': 0.0003240200021537021, 'timing/model': 0.39069732300413307, 'learning_rate/qwen_vl_interface': 9.115967674418765e-06, 'learning_rate/action_model': 9.115967674418765e-05, 'epoch': 0.08}) +[2026-10-03 13:13:36,712][starVLA.training.train_starvla][INFO] - Step 1068, Loss: {'train/loss': 0.000202607610845007, 'train/grad_norm_pre_clip': 0.055620864033699036, 'timing/data': 0.00019547999545466155, 'timing/model': 0.3896880329994019, 'learning_rate/qwen_vl_interface': 9.114197423440949e-06, 'learning_rate/action_model': 9.11419742344095e-05, 'epoch': 0.08}) +[2026-10-03 13:13:37,409][starVLA.training.train_starvla][INFO] - Step 1069, Loss: {'train/loss': 0.5321542620658875, 'train/grad_norm_pre_clip': 30.851848602294922, 'timing/data': 0.00032297000871039927, 'timing/model': 0.38858937300392427, 'learning_rate/qwen_vl_interface': 9.112425584037536e-06, 'learning_rate/action_model': 9.112425584037536e-05, 'epoch': 0.08}) +[2026-10-03 13:13:38,302][starVLA.training.train_starvla][INFO] - Step 1070, Loss: {'train/loss': 0.0006815785309299827, 'train/grad_norm_pre_clip': 0.3382323384284973, 'timing/data': 0.00020336000306997448, 'timing/model': 0.5864745500002755, 'learning_rate/qwen_vl_interface': 9.110652156936863e-06, 'learning_rate/action_model': 9.110652156936863e-05, 'epoch': 0.08}) +[2026-10-03 13:13:39,002][starVLA.training.train_starvla][INFO] - Step 1071, Loss: {'train/loss': 0.06678397208452225, 'train/grad_norm_pre_clip': 70.54686737060547, 'timing/data': 0.0003216500044800341, 'timing/model': 0.3904664129950106, 'learning_rate/qwen_vl_interface': 9.108877142867918e-06, 'learning_rate/action_model': 9.108877142867918e-05, 'epoch': 0.08}) +[2026-10-03 13:13:39,696][starVLA.training.train_starvla][INFO] - Step 1072, Loss: {'train/loss': 0.003536170581355691, 'train/grad_norm_pre_clip': 3.947437286376953, 'timing/data': 0.0003324999997857958, 'timing/model': 0.3883567830052925, 'learning_rate/qwen_vl_interface': 9.107100542560342e-06, 'learning_rate/action_model': 9.107100542560342e-05, 'epoch': 0.08}) +[2026-10-03 13:13:40,392][starVLA.training.train_starvla][INFO] - Step 1073, Loss: {'train/loss': 0.2936129570007324, 'train/grad_norm_pre_clip': 57.279869079589844, 'timing/data': 0.0001952800084836781, 'timing/model': 0.3890110329957679, 'learning_rate/qwen_vl_interface': 9.105322356744429e-06, 'learning_rate/action_model': 9.105322356744429e-05, 'epoch': 0.08}) +[2026-10-03 13:13:41,089][starVLA.training.train_starvla][INFO] - Step 1074, Loss: {'train/loss': 0.4996131658554077, 'train/grad_norm_pre_clip': 14.46975326538086, 'timing/data': 0.0002473600034136325, 'timing/model': 0.390691271997639, 'learning_rate/qwen_vl_interface': 9.103542586151121e-06, 'learning_rate/action_model': 9.10354258615112e-05, 'epoch': 0.08}) +[2026-10-03 13:13:41,984][starVLA.training.train_starvla][INFO] - Step 1075, Loss: {'train/loss': 0.001003142329864204, 'train/grad_norm_pre_clip': 1.3985282182693481, 'timing/data': 0.0002485200093360618, 'timing/model': 0.5875489289901452, 'learning_rate/qwen_vl_interface': 9.101761231512017e-06, 'learning_rate/action_model': 9.101761231512017e-05, 'epoch': 0.08}) +[2026-10-03 13:13:42,679][starVLA.training.train_starvla][INFO] - Step 1076, Loss: {'train/loss': 0.4599681496620178, 'train/grad_norm_pre_clip': 22.58501434326172, 'timing/data': 0.0001951099984580651, 'timing/model': 0.38876926300872583, 'learning_rate/qwen_vl_interface': 9.099978293559365e-06, 'learning_rate/action_model': 9.099978293559366e-05, 'epoch': 0.08}) +[2026-10-03 13:13:43,376][starVLA.training.train_starvla][INFO] - Step 1077, Loss: {'train/loss': 0.012055186554789543, 'train/grad_norm_pre_clip': 16.005342483520508, 'timing/data': 0.00019698998949024826, 'timing/model': 0.38935014299931936, 'learning_rate/qwen_vl_interface': 9.098193773026065e-06, 'learning_rate/action_model': 9.098193773026064e-05, 'epoch': 0.08}) +[2026-10-03 13:13:44,073][starVLA.training.train_starvla][INFO] - Step 1078, Loss: {'train/loss': 0.05435119569301605, 'train/grad_norm_pre_clip': 23.146099090576172, 'timing/data': 0.0003373299987288192, 'timing/model': 0.39064014299947303, 'learning_rate/qwen_vl_interface': 9.096407670645663e-06, 'learning_rate/action_model': 9.096407670645662e-05, 'epoch': 0.08}) +[2026-10-03 13:13:44,773][starVLA.training.train_starvla][INFO] - Step 1079, Loss: {'train/loss': 0.46029117703437805, 'train/grad_norm_pre_clip': 94.26973724365234, 'timing/data': 0.0003317200025776401, 'timing/model': 0.39127096299489494, 'learning_rate/qwen_vl_interface': 9.094619987152357e-06, 'learning_rate/action_model': 9.094619987152358e-05, 'epoch': 0.08}) +[2026-10-03 13:13:45,469][starVLA.training.train_starvla][INFO] - Step 1080, Loss: {'train/loss': 0.2610054910182953, 'train/grad_norm_pre_clip': 68.46839141845703, 'timing/data': 0.0001994500053115189, 'timing/model': 0.38958707300480455, 'learning_rate/qwen_vl_interface': 9.092830723281e-06, 'learning_rate/action_model': 9.092830723281e-05, 'epoch': 0.08}) +[2026-10-03 13:13:46,359][starVLA.training.train_starvla][INFO] - Step 1081, Loss: {'train/loss': 0.001262062694877386, 'train/grad_norm_pre_clip': 0.5534369945526123, 'timing/data': 0.00037877999420743436, 'timing/model': 0.3891680929955328, 'learning_rate/qwen_vl_interface': 9.09103987976709e-06, 'learning_rate/action_model': 9.09103987976709e-05, 'epoch': 0.08}) +[2026-10-03 13:13:47,058][starVLA.training.train_starvla][INFO] - Step 1082, Loss: {'train/loss': 0.032058291137218475, 'train/grad_norm_pre_clip': 47.891448974609375, 'timing/data': 0.0001948100107256323, 'timing/model': 0.39194081199821085, 'learning_rate/qwen_vl_interface': 9.089247457346774e-06, 'learning_rate/action_model': 9.089247457346773e-05, 'epoch': 0.08}) +[2026-10-03 13:13:47,760][starVLA.training.train_starvla][INFO] - Step 1083, Loss: {'train/loss': 0.03234047442674637, 'train/grad_norm_pre_clip': 41.06493377685547, 'timing/data': 0.00032109000312630087, 'timing/model': 0.3920038529904559, 'learning_rate/qwen_vl_interface': 9.087453456756849e-06, 'learning_rate/action_model': 9.087453456756849e-05, 'epoch': 0.08}) +[2026-10-03 13:13:48,456][starVLA.training.train_starvla][INFO] - Step 1084, Loss: {'train/loss': 0.000752763997297734, 'train/grad_norm_pre_clip': 0.5605228543281555, 'timing/data': 0.00019258000247646123, 'timing/model': 0.38937540400365833, 'learning_rate/qwen_vl_interface': 9.085657878734759e-06, 'learning_rate/action_model': 9.085657878734758e-05, 'epoch': 0.08}) +[2026-10-03 13:13:49,153][starVLA.training.train_starvla][INFO] - Step 1085, Loss: {'train/loss': 0.007714098319411278, 'train/grad_norm_pre_clip': 2.6084837913513184, 'timing/data': 0.00026434999017510563, 'timing/model': 0.38966427400009707, 'learning_rate/qwen_vl_interface': 9.083860724018603e-06, 'learning_rate/action_model': 9.083860724018602e-05, 'epoch': 0.08}) +[2026-10-03 13:13:50,045][starVLA.training.train_starvla][INFO] - Step 1086, Loss: {'train/loss': 0.0013987756101414561, 'train/grad_norm_pre_clip': 0.36749783158302307, 'timing/data': 0.00032691001251805574, 'timing/model': 0.5839184799988288, 'learning_rate/qwen_vl_interface': 9.08206199334712e-06, 'learning_rate/action_model': 9.082061993347119e-05, 'epoch': 0.08}) +[2026-10-03 13:13:50,746][starVLA.training.train_starvla][INFO] - Step 1087, Loss: {'train/loss': 1.333604097366333, 'train/grad_norm_pre_clip': 177.94064331054688, 'timing/data': 0.000202369992621243, 'timing/model': 0.39195830300741363, 'learning_rate/qwen_vl_interface': 9.0802616874597e-06, 'learning_rate/action_model': 9.0802616874597e-05, 'epoch': 0.08}) +[2026-10-03 13:13:51,445][starVLA.training.train_starvla][INFO] - Step 1088, Loss: {'train/loss': 0.0005821421509608626, 'train/grad_norm_pre_clip': 0.10963132232427597, 'timing/data': 0.0002662600018084049, 'timing/model': 0.38964052399387583, 'learning_rate/qwen_vl_interface': 9.078459807096385e-06, 'learning_rate/action_model': 9.078459807096385e-05, 'epoch': 0.08}) +[2026-10-03 13:13:52,143][starVLA.training.train_starvla][INFO] - Step 1089, Loss: {'train/loss': 0.84975266456604, 'train/grad_norm_pre_clip': 35.37498092651367, 'timing/data': 0.00019972000154666603, 'timing/model': 0.38860895299876574, 'learning_rate/qwen_vl_interface': 9.076656352997854e-06, 'learning_rate/action_model': 9.076656352997853e-05, 'epoch': 0.08}) +[2026-10-03 13:13:52,840][starVLA.training.train_starvla][INFO] - Step 1090, Loss: {'train/loss': 0.05925966799259186, 'train/grad_norm_pre_clip': 57.46464920043945, 'timing/data': 0.00019645999418571591, 'timing/model': 0.39096683300158475, 'learning_rate/qwen_vl_interface': 9.074851325905443e-06, 'learning_rate/action_model': 9.074851325905442e-05, 'epoch': 0.08}) +[2026-10-03 13:13:53,538][starVLA.training.train_starvla][INFO] - Step 1091, Loss: {'train/loss': 0.0010616716463118792, 'train/grad_norm_pre_clip': 0.3889617621898651, 'timing/data': 0.00022521999198943377, 'timing/model': 0.3910044330114033, 'learning_rate/qwen_vl_interface': 9.073044726561128e-06, 'learning_rate/action_model': 9.073044726561129e-05, 'epoch': 0.08}) +[2026-10-03 13:13:54,425][starVLA.training.train_starvla][INFO] - Step 1092, Loss: {'train/loss': 0.6729836463928223, 'train/grad_norm_pre_clip': 43.73897171020508, 'timing/data': 0.000198969995835796, 'timing/model': 0.38833296300435904, 'learning_rate/qwen_vl_interface': 9.071236555707536e-06, 'learning_rate/action_model': 9.071236555707537e-05, 'epoch': 0.08}) +[2026-10-03 13:13:55,120][starVLA.training.train_starvla][INFO] - Step 1093, Loss: {'train/loss': 0.12713126838207245, 'train/grad_norm_pre_clip': 57.102813720703125, 'timing/data': 0.00019490999693516642, 'timing/model': 0.3883608329924755, 'learning_rate/qwen_vl_interface': 9.069426814087937e-06, 'learning_rate/action_model': 9.069426814087937e-05, 'epoch': 0.08}) +[2026-10-03 13:13:55,816][starVLA.training.train_starvla][INFO] - Step 1094, Loss: {'train/loss': 0.11649417132139206, 'train/grad_norm_pre_clip': 49.97214126586914, 'timing/data': 0.0002763900120044127, 'timing/model': 0.39088061299116816, 'learning_rate/qwen_vl_interface': 9.067615502446247e-06, 'learning_rate/action_model': 9.067615502446246e-05, 'epoch': 0.08}) +[2026-10-03 13:13:56,515][starVLA.training.train_starvla][INFO] - Step 1095, Loss: {'train/loss': 0.2117827981710434, 'train/grad_norm_pre_clip': 29.117136001586914, 'timing/data': 0.000261950001004152, 'timing/model': 0.39106081299541984, 'learning_rate/qwen_vl_interface': 9.06580262152703e-06, 'learning_rate/action_model': 9.065802621527028e-05, 'epoch': 0.08}) +[2026-10-03 13:13:57,210][starVLA.training.train_starvla][INFO] - Step 1096, Loss: {'train/loss': 0.0077056605368852615, 'train/grad_norm_pre_clip': 6.9878106117248535, 'timing/data': 0.00019397999858483672, 'timing/model': 0.38838983299501706, 'learning_rate/qwen_vl_interface': 9.063988172075486e-06, 'learning_rate/action_model': 9.063988172075486e-05, 'epoch': 0.08}) +[2026-10-03 13:13:58,104][starVLA.training.train_starvla][INFO] - Step 1097, Loss: {'train/loss': 0.000350229034665972, 'train/grad_norm_pre_clip': 0.18397514522075653, 'timing/data': 0.00029986999288666993, 'timing/model': 0.38821624299453106, 'learning_rate/qwen_vl_interface': 9.062172154837475e-06, 'learning_rate/action_model': 9.062172154837475e-05, 'epoch': 0.08}) +[2026-10-03 13:13:58,800][starVLA.training.train_starvla][INFO] - Step 1098, Loss: {'train/loss': 0.12367378175258636, 'train/grad_norm_pre_clip': 57.86158752441406, 'timing/data': 0.00021183000353630632, 'timing/model': 0.3909054230025504, 'learning_rate/qwen_vl_interface': 9.060354570559487e-06, 'learning_rate/action_model': 9.060354570559487e-05, 'epoch': 0.08}) +[2026-10-03 13:13:59,499][starVLA.training.train_starvla][INFO] - Step 1099, Loss: {'train/loss': 0.48478052020072937, 'train/grad_norm_pre_clip': 110.00724029541016, 'timing/data': 0.00027816000510938466, 'timing/model': 0.3903397829999449, 'learning_rate/qwen_vl_interface': 9.058535419988667e-06, 'learning_rate/action_model': 9.058535419988667e-05, 'epoch': 0.08}) +[2026-10-03 13:14:00,193][starVLA.training.train_starvla][INFO] - Step 1100, Loss: {'train/loss': 0.4806268513202667, 'train/grad_norm_pre_clip': 24.112356185913086, 'timing/data': 0.00036608999653253704, 'timing/model': 0.3879759129922604, 'learning_rate/qwen_vl_interface': 9.056714703872794e-06, 'learning_rate/action_model': 9.056714703872794e-05, 'epoch': 0.08}) +[2026-10-03 13:14:00,889][starVLA.training.train_starvla][INFO] - Step 1101, Loss: {'train/loss': 0.3561057150363922, 'train/grad_norm_pre_clip': 73.06011962890625, 'timing/data': 0.00020169999334029853, 'timing/model': 0.38801673300622497, 'learning_rate/qwen_vl_interface': 9.0548924229603e-06, 'learning_rate/action_model': 9.054892422960301e-05, 'epoch': 0.08}) +[2026-10-03 13:14:01,778][starVLA.training.train_starvla][INFO] - Step 1102, Loss: {'train/loss': 0.4570772051811218, 'train/grad_norm_pre_clip': 21.214468002319336, 'timing/data': 0.00025220999668817967, 'timing/model': 0.5835152999934508, 'learning_rate/qwen_vl_interface': 9.053068578000256e-06, 'learning_rate/action_model': 9.053068578000256e-05, 'epoch': 0.08}) +[2026-10-03 13:14:02,477][starVLA.training.train_starvla][INFO] - Step 1103, Loss: {'train/loss': 0.012159409001469612, 'train/grad_norm_pre_clip': 4.602131366729736, 'timing/data': 0.00030394000350497663, 'timing/model': 0.39098475300124846, 'learning_rate/qwen_vl_interface': 9.051243169742372e-06, 'learning_rate/action_model': 9.051243169742372e-05, 'epoch': 0.08}) +[2026-10-03 13:14:03,172][starVLA.training.train_starvla][INFO] - Step 1104, Loss: {'train/loss': 0.014439240097999573, 'train/grad_norm_pre_clip': 4.573663711547852, 'timing/data': 0.0003042699972866103, 'timing/model': 0.38837431398860645, 'learning_rate/qwen_vl_interface': 9.049416198937008e-06, 'learning_rate/action_model': 9.049416198937007e-05, 'epoch': 0.08}) +[2026-10-03 13:14:03,866][starVLA.training.train_starvla][INFO] - Step 1105, Loss: {'train/loss': 0.12175553292036057, 'train/grad_norm_pre_clip': 23.102550506591797, 'timing/data': 0.00031288999889511615, 'timing/model': 0.38682321400847286, 'learning_rate/qwen_vl_interface': 9.047587666335159e-06, 'learning_rate/action_model': 9.047587666335159e-05, 'epoch': 0.08}) +[2026-10-03 13:14:04,564][starVLA.training.train_starvla][INFO] - Step 1106, Loss: {'train/loss': 0.24691809713840485, 'train/grad_norm_pre_clip': 17.74854278564453, 'timing/data': 0.0002568500058259815, 'timing/model': 0.39175639300083276, 'learning_rate/qwen_vl_interface': 9.045757572688469e-06, 'learning_rate/action_model': 9.04575757268847e-05, 'epoch': 0.08}) +[2026-10-03 13:14:05,263][starVLA.training.train_starvla][INFO] - Step 1107, Loss: {'train/loss': 0.02874397113919258, 'train/grad_norm_pre_clip': 14.29858112335205, 'timing/data': 0.0002301700005773455, 'timing/model': 0.3908256630093092, 'learning_rate/qwen_vl_interface': 9.043925918749218e-06, 'learning_rate/action_model': 9.043925918749218e-05, 'epoch': 0.08}) +[2026-10-03 13:14:06,152][starVLA.training.train_starvla][INFO] - Step 1108, Loss: {'train/loss': 0.45766371488571167, 'train/grad_norm_pre_clip': 67.17255401611328, 'timing/data': 0.0002036399964708835, 'timing/model': 0.38895650299673434, 'learning_rate/qwen_vl_interface': 9.042092705270333e-06, 'learning_rate/action_model': 9.042092705270333e-05, 'epoch': 0.08}) +[2026-10-03 13:14:06,848][starVLA.training.train_starvla][INFO] - Step 1109, Loss: {'train/loss': 0.5559245944023132, 'train/grad_norm_pre_clip': 141.24838256835938, 'timing/data': 0.00028867999208159745, 'timing/model': 0.3897098529996583, 'learning_rate/qwen_vl_interface': 9.040257933005375e-06, 'learning_rate/action_model': 9.040257933005374e-05, 'epoch': 0.08}) +[2026-10-03 13:14:07,546][starVLA.training.train_starvla][INFO] - Step 1110, Loss: {'train/loss': 0.592710018157959, 'train/grad_norm_pre_clip': 59.740684509277344, 'timing/data': 0.000265370006673038, 'timing/model': 0.3913121130026411, 'learning_rate/qwen_vl_interface': 9.038421602708552e-06, 'learning_rate/action_model': 9.038421602708552e-05, 'epoch': 0.08}) +[2026-10-03 13:14:08,244][starVLA.training.train_starvla][INFO] - Step 1111, Loss: {'train/loss': 0.4293295741081238, 'train/grad_norm_pre_clip': 51.14620590209961, 'timing/data': 0.00030497999978251755, 'timing/model': 0.3904823330085492, 'learning_rate/qwen_vl_interface': 9.036583715134709e-06, 'learning_rate/action_model': 9.036583715134708e-05, 'epoch': 0.08}) +[2026-10-03 13:14:08,939][starVLA.training.train_starvla][INFO] - Step 1112, Loss: {'train/loss': 0.0023750746622681618, 'train/grad_norm_pre_clip': 0.6446624994277954, 'timing/data': 0.00032720000308472663, 'timing/model': 0.38835789400036447, 'learning_rate/qwen_vl_interface': 9.034744271039332e-06, 'learning_rate/action_model': 9.034744271039332e-05, 'epoch': 0.08}) +[2026-10-03 13:14:09,826][starVLA.training.train_starvla][INFO] - Step 1113, Loss: {'train/loss': 0.00357536762021482, 'train/grad_norm_pre_clip': 3.7156946659088135, 'timing/data': 0.00020237000717315823, 'timing/model': 0.3883087229914963, 'learning_rate/qwen_vl_interface': 9.03290327117855e-06, 'learning_rate/action_model': 9.032903271178549e-05, 'epoch': 0.08}) +[2026-10-03 13:14:10,523][starVLA.training.train_starvla][INFO] - Step 1114, Loss: {'train/loss': 0.0016042309580370784, 'train/grad_norm_pre_clip': 0.6504185199737549, 'timing/data': 0.00023234999389387667, 'timing/model': 0.391023062999011, 'learning_rate/qwen_vl_interface': 9.031060716309122e-06, 'learning_rate/action_model': 9.031060716309121e-05, 'epoch': 0.08}) +[2026-10-03 13:14:11,221][starVLA.training.train_starvla][INFO] - Step 1115, Loss: {'train/loss': 0.11465582251548767, 'train/grad_norm_pre_clip': 26.047332763671875, 'timing/data': 0.00023174998932518065, 'timing/model': 0.39002541299851146, 'learning_rate/qwen_vl_interface': 9.02921660718846e-06, 'learning_rate/action_model': 9.029216607188459e-05, 'epoch': 0.08}) +[2026-10-03 13:14:11,917][starVLA.training.train_starvla][INFO] - Step 1116, Loss: {'train/loss': 0.03818870335817337, 'train/grad_norm_pre_clip': 7.879293918609619, 'timing/data': 0.00032624999585095793, 'timing/model': 0.3889014629967278, 'learning_rate/qwen_vl_interface': 9.027370944574604e-06, 'learning_rate/action_model': 9.027370944574604e-05, 'epoch': 0.08}) +[2026-10-03 13:14:12,612][starVLA.training.train_starvla][INFO] - Step 1117, Loss: {'train/loss': 0.012950534000992775, 'train/grad_norm_pre_clip': 8.295599937438965, 'timing/data': 0.00037292001070454717, 'timing/model': 0.38878636399749666, 'learning_rate/qwen_vl_interface': 9.025523729226236e-06, 'learning_rate/action_model': 9.025523729226236e-05, 'epoch': 0.08}) +[2026-10-03 13:14:13,503][starVLA.training.train_starvla][INFO] - Step 1118, Loss: {'train/loss': 0.1868400275707245, 'train/grad_norm_pre_clip': 24.260339736938477, 'timing/data': 0.0001952599996002391, 'timing/model': 0.5846013010013849, 'learning_rate/qwen_vl_interface': 9.023674961902677e-06, 'learning_rate/action_model': 9.023674961902676e-05, 'epoch': 0.08}) +[2026-10-03 13:14:14,202][starVLA.training.train_starvla][INFO] - Step 1119, Loss: {'train/loss': 0.005146597977727652, 'train/grad_norm_pre_clip': 1.0535935163497925, 'timing/data': 0.0002775999892037362, 'timing/model': 0.39070660299330484, 'learning_rate/qwen_vl_interface': 9.021824643363883e-06, 'learning_rate/action_model': 9.021824643363883e-05, 'epoch': 0.08}) +[2026-10-03 13:14:14,898][starVLA.training.train_starvla][INFO] - Step 1120, Loss: {'train/loss': 0.000815697421785444, 'train/grad_norm_pre_clip': 0.1786031424999237, 'timing/data': 0.0002641800092533231, 'timing/model': 0.38879857400024775, 'learning_rate/qwen_vl_interface': 9.019972774370456e-06, 'learning_rate/action_model': 9.019972774370455e-05, 'epoch': 0.08}) +[2026-10-03 13:14:15,593][starVLA.training.train_starvla][INFO] - Step 1121, Loss: {'train/loss': 0.4320201575756073, 'train/grad_norm_pre_clip': 375.5049743652344, 'timing/data': 0.0002661399921635166, 'timing/model': 0.38915167399682105, 'learning_rate/qwen_vl_interface': 9.018119355683623e-06, 'learning_rate/action_model': 9.018119355683623e-05, 'epoch': 0.08}) +[2026-10-03 13:14:16,291][starVLA.training.train_starvla][INFO] - Step 1122, Loss: {'train/loss': 0.00944439321756363, 'train/grad_norm_pre_clip': 3.803325653076172, 'timing/data': 0.00026030000299215317, 'timing/model': 0.39109541299694683, 'learning_rate/qwen_vl_interface': 9.016264388065258e-06, 'learning_rate/action_model': 9.016264388065258e-05, 'epoch': 0.08}) +[2026-10-03 13:14:17,183][starVLA.training.train_starvla][INFO] - Step 1123, Loss: {'train/loss': 0.03033634088933468, 'train/grad_norm_pre_clip': 41.12312316894531, 'timing/data': 0.00027322000823915005, 'timing/model': 0.5845730300061405, 'learning_rate/qwen_vl_interface': 9.014407872277869e-06, 'learning_rate/action_model': 9.014407872277869e-05, 'epoch': 0.08}) +[2026-10-03 13:14:17,879][starVLA.training.train_starvla][INFO] - Step 1124, Loss: {'train/loss': 0.00045323348604142666, 'train/grad_norm_pre_clip': 0.1243601068854332, 'timing/data': 0.000304938992485404, 'timing/model': 0.38850983399606775, 'learning_rate/qwen_vl_interface': 9.012549809084595e-06, 'learning_rate/action_model': 9.012549809084595e-05, 'epoch': 0.08}) +[2026-10-03 13:14:18,575][starVLA.training.train_starvla][INFO] - Step 1125, Loss: {'train/loss': 0.37163594365119934, 'train/grad_norm_pre_clip': 38.65801239013672, 'timing/data': 0.0002002399996854365, 'timing/model': 0.3894435929978499, 'learning_rate/qwen_vl_interface': 9.01069019924922e-06, 'learning_rate/action_model': 9.010690199249219e-05, 'epoch': 0.08}) +[2026-10-03 13:14:19,270][starVLA.training.train_starvla][INFO] - Step 1126, Loss: {'train/loss': 0.0010454758303239942, 'train/grad_norm_pre_clip': 0.5558006763458252, 'timing/data': 0.0003315999929327518, 'timing/model': 0.388148772995919, 'learning_rate/qwen_vl_interface': 9.008829043536157e-06, 'learning_rate/action_model': 9.008829043536156e-05, 'epoch': 0.08}) +[2026-10-03 13:14:19,970][starVLA.training.train_starvla][INFO] - Step 1127, Loss: {'train/loss': 0.2846899926662445, 'train/grad_norm_pre_clip': 31.92317008972168, 'timing/data': 0.0004418500029714778, 'timing/model': 0.39041709300363436, 'learning_rate/qwen_vl_interface': 9.006966342710458e-06, 'learning_rate/action_model': 9.006966342710457e-05, 'epoch': 0.08}) +[2026-10-03 13:14:20,860][starVLA.training.train_starvla][INFO] - Step 1128, Loss: {'train/loss': 0.00017059891251847148, 'train/grad_norm_pre_clip': 0.04306741803884506, 'timing/data': 0.00020153999503236264, 'timing/model': 0.5835839200008195, 'learning_rate/qwen_vl_interface': 9.005102097537809e-06, 'learning_rate/action_model': 9.005102097537809e-05, 'epoch': 0.08}) +[2026-10-03 13:14:21,558][starVLA.training.train_starvla][INFO] - Step 1129, Loss: {'train/loss': 0.19300171732902527, 'train/grad_norm_pre_clip': 82.83148956298828, 'timing/data': 0.00020092001068405807, 'timing/model': 0.3906848939950578, 'learning_rate/qwen_vl_interface': 9.00323630878453e-06, 'learning_rate/action_model': 9.00323630878453e-05, 'epoch': 0.08}) +[2026-10-03 13:14:22,255][starVLA.training.train_starvla][INFO] - Step 1130, Loss: {'train/loss': 0.006850077770650387, 'train/grad_norm_pre_clip': 9.045388221740723, 'timing/data': 0.00032114000350702554, 'timing/model': 0.39032709300227, 'learning_rate/qwen_vl_interface': 9.001368977217577e-06, 'learning_rate/action_model': 9.001368977217577e-05, 'epoch': 0.08}) +[2026-10-03 13:14:22,960][starVLA.training.train_starvla][INFO] - Step 1131, Loss: {'train/loss': 0.27027562260627747, 'train/grad_norm_pre_clip': 37.418067932128906, 'timing/data': 0.0003078599984291941, 'timing/model': 0.3959756129916059, 'learning_rate/qwen_vl_interface': 8.999500103604538e-06, 'learning_rate/action_model': 8.999500103604536e-05, 'epoch': 0.08}) +[2026-10-03 13:14:23,666][starVLA.training.train_starvla][INFO] - Step 1132, Loss: {'train/loss': 8.487536251777783e-05, 'train/grad_norm_pre_clip': 0.013516482897102833, 'timing/data': 0.00041043000237550586, 'timing/model': 0.39396320199011825, 'learning_rate/qwen_vl_interface': 8.997629688713639e-06, 'learning_rate/action_model': 8.997629688713638e-05, 'epoch': 0.08}) +[2026-10-03 13:14:24,371][starVLA.training.train_starvla][INFO] - Step 1133, Loss: {'train/loss': 0.000793107901699841, 'train/grad_norm_pre_clip': 0.18147766590118408, 'timing/data': 0.00026296000578440726, 'timing/model': 0.3930059629929019, 'learning_rate/qwen_vl_interface': 8.995757733313733e-06, 'learning_rate/action_model': 8.995757733313732e-05, 'epoch': 0.08}) +[2026-10-03 13:14:25,272][starVLA.training.train_starvla][INFO] - Step 1134, Loss: {'train/loss': 0.0006751271430402994, 'train/grad_norm_pre_clip': 0.31757852435112, 'timing/data': 0.0003624599921749905, 'timing/model': 0.3941327329957858, 'learning_rate/qwen_vl_interface': 8.993884238174314e-06, 'learning_rate/action_model': 8.993884238174314e-05, 'epoch': 0.08}) +[2026-10-03 13:14:25,978][starVLA.training.train_starvla][INFO] - Step 1135, Loss: {'train/loss': 0.00024339040101040155, 'train/grad_norm_pre_clip': 0.045923445373773575, 'timing/data': 0.00031493999995291233, 'timing/model': 0.3941996619978454, 'learning_rate/qwen_vl_interface': 8.992009204065505e-06, 'learning_rate/action_model': 8.992009204065505e-05, 'epoch': 0.08}) +[2026-10-03 13:14:26,686][starVLA.training.train_starvla][INFO] - Step 1136, Loss: {'train/loss': 0.0011099458206444979, 'train/grad_norm_pre_clip': 0.2954387068748474, 'timing/data': 0.0003187300026183948, 'timing/model': 0.394526363001205, 'learning_rate/qwen_vl_interface': 8.990132631758058e-06, 'learning_rate/action_model': 8.990132631758059e-05, 'epoch': 0.08}) +[2026-10-03 13:14:27,388][starVLA.training.train_starvla][INFO] - Step 1137, Loss: {'train/loss': 0.5359101891517639, 'train/grad_norm_pre_clip': 50.99361801147461, 'timing/data': 0.00029495000489987433, 'timing/model': 0.39239035300852265, 'learning_rate/qwen_vl_interface': 8.988254522023364e-06, 'learning_rate/action_model': 8.988254522023363e-05, 'epoch': 0.08}) +[2026-10-03 13:14:28,092][starVLA.training.train_starvla][INFO] - Step 1138, Loss: {'train/loss': 0.0015960745513439178, 'train/grad_norm_pre_clip': 1.0869956016540527, 'timing/data': 0.00032046998967416584, 'timing/model': 0.39298851300554816, 'learning_rate/qwen_vl_interface': 8.986374875633441e-06, 'learning_rate/action_model': 8.98637487563344e-05, 'epoch': 0.08}) +[2026-10-03 13:14:28,986][starVLA.training.train_starvla][INFO] - Step 1139, Loss: {'train/loss': 0.43601226806640625, 'train/grad_norm_pre_clip': 188.07357788085938, 'timing/data': 0.00026713999977800995, 'timing/model': 0.5833861200080719, 'learning_rate/qwen_vl_interface': 8.984493693360944e-06, 'learning_rate/action_model': 8.984493693360943e-05, 'epoch': 0.08}) +[2026-10-03 13:14:29,690][starVLA.training.train_starvla][INFO] - Step 1140, Loss: {'train/loss': 0.0013602669350802898, 'train/grad_norm_pre_clip': 2.2429680824279785, 'timing/data': 0.00035604000731837004, 'timing/model': 0.3924708140111761, 'learning_rate/qwen_vl_interface': 8.982610975979153e-06, 'learning_rate/action_model': 8.982610975979152e-05, 'epoch': 0.08}) +[2026-10-03 13:14:30,391][starVLA.training.train_starvla][INFO] - Step 1141, Loss: {'train/loss': 0.1162492111325264, 'train/grad_norm_pre_clip': 116.0257339477539, 'timing/data': 0.00031894000130705535, 'timing/model': 0.39140443300129846, 'learning_rate/qwen_vl_interface': 8.98072672426198e-06, 'learning_rate/action_model': 8.980726724261981e-05, 'epoch': 0.08}) +[2026-10-03 13:14:31,093][starVLA.training.train_starvla][INFO] - Step 1142, Loss: {'train/loss': 0.0004626433365046978, 'train/grad_norm_pre_clip': 0.26331013441085815, 'timing/data': 0.0002532299986341968, 'timing/model': 0.39254564300063066, 'learning_rate/qwen_vl_interface': 8.978840938983977e-06, 'learning_rate/action_model': 8.978840938983977e-05, 'epoch': 0.08}) +[2026-10-03 13:14:31,794][starVLA.training.train_starvla][INFO] - Step 1143, Loss: {'train/loss': 0.002605883404612541, 'train/grad_norm_pre_clip': 4.8387346267700195, 'timing/data': 0.00023312999110203236, 'timing/model': 0.39179637300549075, 'learning_rate/qwen_vl_interface': 8.976953620920313e-06, 'learning_rate/action_model': 8.976953620920312e-05, 'epoch': 0.08}) +[2026-10-03 13:14:32,497][starVLA.training.train_starvla][INFO] - Step 1144, Loss: {'train/loss': 0.5143089890480042, 'train/grad_norm_pre_clip': 151.0567169189453, 'timing/data': 0.00033590001112315804, 'timing/model': 0.3925471029942855, 'learning_rate/qwen_vl_interface': 8.975064770846794e-06, 'learning_rate/action_model': 8.975064770846794e-05, 'epoch': 0.08}) +[2026-10-03 13:14:33,391][starVLA.training.train_starvla][INFO] - Step 1145, Loss: {'train/loss': 0.00044145973515696824, 'train/grad_norm_pre_clip': 0.3030187785625458, 'timing/data': 0.0003186999965691939, 'timing/model': 0.3917480539967073, 'learning_rate/qwen_vl_interface': 8.973174389539856e-06, 'learning_rate/action_model': 8.973174389539855e-05, 'epoch': 0.08}) +[2026-10-03 13:14:34,094][starVLA.training.train_starvla][INFO] - Step 1146, Loss: {'train/loss': 0.42683175206184387, 'train/grad_norm_pre_clip': 84.28558349609375, 'timing/data': 0.0003171599964844063, 'timing/model': 0.3930067129986128, 'learning_rate/qwen_vl_interface': 8.971282477776563e-06, 'learning_rate/action_model': 8.971282477776563e-05, 'epoch': 0.08}) +[2026-10-03 13:14:34,797][starVLA.training.train_starvla][INFO] - Step 1147, Loss: {'train/loss': 0.19436977803707123, 'train/grad_norm_pre_clip': 95.47303771972656, 'timing/data': 0.000312019998091273, 'timing/model': 0.3923769030079711, 'learning_rate/qwen_vl_interface': 8.969389036334609e-06, 'learning_rate/action_model': 8.969389036334609e-05, 'epoch': 0.08}) +[2026-10-03 13:14:35,500][starVLA.training.train_starvla][INFO] - Step 1148, Loss: {'train/loss': 0.3438607454299927, 'train/grad_norm_pre_clip': 305.09564208984375, 'timing/data': 0.0002658300072653219, 'timing/model': 0.39229927300766576, 'learning_rate/qwen_vl_interface': 8.967494065992315e-06, 'learning_rate/action_model': 8.967494065992315e-05, 'epoch': 0.08}) +[2026-10-03 13:14:36,203][starVLA.training.train_starvla][INFO] - Step 1149, Loss: {'train/loss': 0.00628339359536767, 'train/grad_norm_pre_clip': 7.142284870147705, 'timing/data': 0.00025630000163801014, 'timing/model': 0.39266150299226865, 'learning_rate/qwen_vl_interface': 8.965597567528634e-06, 'learning_rate/action_model': 8.965597567528633e-05, 'epoch': 0.08}) +[2026-10-03 13:14:37,099][starVLA.training.train_starvla][INFO] - Step 1150, Loss: {'train/loss': 0.0005130781792104244, 'train/grad_norm_pre_clip': 0.226796954870224, 'timing/data': 0.0002465699944877997, 'timing/model': 0.3929844719968969, 'learning_rate/qwen_vl_interface': 8.963699541723143e-06, 'learning_rate/action_model': 8.963699541723142e-05, 'epoch': 0.08}) +[2026-10-03 13:14:37,801][starVLA.training.train_starvla][INFO] - Step 1151, Loss: {'train/loss': 0.00013016117736697197, 'train/grad_norm_pre_clip': 0.0540737546980381, 'timing/data': 0.0003543490020092577, 'timing/model': 0.39244075400347356, 'learning_rate/qwen_vl_interface': 8.961799989356048e-06, 'learning_rate/action_model': 8.961799989356048e-05, 'epoch': 0.08}) +[2026-10-03 13:14:38,504][starVLA.training.train_starvla][INFO] - Step 1152, Loss: {'train/loss': 0.6993690133094788, 'train/grad_norm_pre_clip': 179.25917053222656, 'timing/data': 0.00035428999399300665, 'timing/model': 0.3925445830100216, 'learning_rate/qwen_vl_interface': 8.959898911208185e-06, 'learning_rate/action_model': 8.959898911208185e-05, 'epoch': 0.08}) +[2026-10-03 13:14:39,206][starVLA.training.train_starvla][INFO] - Step 1153, Loss: {'train/loss': 0.00028638759977184236, 'train/grad_norm_pre_clip': 0.19976770877838135, 'timing/data': 0.00031589000718668103, 'timing/model': 0.3921733440074604, 'learning_rate/qwen_vl_interface': 8.957996308061015e-06, 'learning_rate/action_model': 8.957996308061014e-05, 'epoch': 0.08}) +[2026-10-03 13:14:39,908][starVLA.training.train_starvla][INFO] - Step 1154, Loss: {'train/loss': 0.5402050018310547, 'train/grad_norm_pre_clip': 35.78115463256836, 'timing/data': 0.0003239999932702631, 'timing/model': 0.39218814299965743, 'learning_rate/qwen_vl_interface': 8.956092180696626e-06, 'learning_rate/action_model': 8.956092180696625e-05, 'epoch': 0.08}) +[2026-10-03 13:14:40,803][starVLA.training.train_starvla][INFO] - Step 1155, Loss: {'train/loss': 0.00016204467101488262, 'train/grad_norm_pre_clip': 0.08483779430389404, 'timing/data': 0.00023258999863173813, 'timing/model': 0.5846467599913012, 'learning_rate/qwen_vl_interface': 8.95418652989773e-06, 'learning_rate/action_model': 8.95418652989773e-05, 'epoch': 0.08}) +[2026-10-03 13:14:41,506][starVLA.training.train_starvla][INFO] - Step 1156, Loss: {'train/loss': 0.15852795541286469, 'train/grad_norm_pre_clip': 210.33499145507812, 'timing/data': 0.00034340999263804406, 'timing/model': 0.39261760299268644, 'learning_rate/qwen_vl_interface': 8.952279356447675e-06, 'learning_rate/action_model': 8.952279356447675e-05, 'epoch': 0.08}) +[2026-10-03 13:14:42,208][starVLA.training.train_starvla][INFO] - Step 1157, Loss: {'train/loss': 0.000289247022010386, 'train/grad_norm_pre_clip': 0.24206797778606415, 'timing/data': 0.00030487000185530633, 'timing/model': 0.3924590740061831, 'learning_rate/qwen_vl_interface': 8.950370661130421e-06, 'learning_rate/action_model': 8.950370661130422e-05, 'epoch': 0.08}) +[2026-10-03 13:14:42,911][starVLA.training.train_starvla][INFO] - Step 1158, Loss: {'train/loss': 0.00015847002214286476, 'train/grad_norm_pre_clip': 0.07242776453495026, 'timing/data': 0.00031408999348059297, 'timing/model': 0.392551443001139, 'learning_rate/qwen_vl_interface': 8.948460444730566e-06, 'learning_rate/action_model': 8.948460444730566e-05, 'epoch': 0.08}) +[2026-10-03 13:14:43,613][starVLA.training.train_starvla][INFO] - Step 1159, Loss: {'train/loss': 2.9018396162427962e-05, 'train/grad_norm_pre_clip': 0.006875724531710148, 'timing/data': 0.00033341000380460173, 'timing/model': 0.39230263300123625, 'learning_rate/qwen_vl_interface': 8.946548708033324e-06, 'learning_rate/action_model': 8.946548708033323e-05, 'epoch': 0.08}) +[2026-10-03 13:14:44,319][starVLA.training.train_starvla][INFO] - Step 1160, Loss: {'train/loss': 8.860310481395572e-05, 'train/grad_norm_pre_clip': 0.031500883400440216, 'timing/data': 0.0003522299957694486, 'timing/model': 0.39437413300038315, 'learning_rate/qwen_vl_interface': 8.944635451824538e-06, 'learning_rate/action_model': 8.944635451824538e-05, 'epoch': 0.08}) +[2026-10-03 13:14:45,211][starVLA.training.train_starvla][INFO] - Step 1161, Loss: {'train/loss': 0.01484499592334032, 'train/grad_norm_pre_clip': 26.349197387695312, 'timing/data': 0.00025649000599514693, 'timing/model': 0.39218486299796496, 'learning_rate/qwen_vl_interface': 8.942720676890677e-06, 'learning_rate/action_model': 8.942720676890677e-05, 'epoch': 0.08}) +[2026-10-03 13:14:45,914][starVLA.training.train_starvla][INFO] - Step 1162, Loss: {'train/loss': 0.4977881610393524, 'train/grad_norm_pre_clip': 70.67951965332031, 'timing/data': 0.00023068999871611595, 'timing/model': 0.3929997730010655, 'learning_rate/qwen_vl_interface': 8.940804384018833e-06, 'learning_rate/action_model': 8.940804384018832e-05, 'epoch': 0.09}) +[2026-10-03 13:14:46,618][starVLA.training.train_starvla][INFO] - Step 1163, Loss: {'train/loss': 9.933435649145395e-05, 'train/grad_norm_pre_clip': 0.027913277968764305, 'timing/data': 0.0002616599958855659, 'timing/model': 0.39302315399982035, 'learning_rate/qwen_vl_interface': 8.93888657399672e-06, 'learning_rate/action_model': 8.93888657399672e-05, 'epoch': 0.09}) +[2026-10-03 13:14:47,322][starVLA.training.train_starvla][INFO] - Step 1164, Loss: {'train/loss': 0.618147611618042, 'train/grad_norm_pre_clip': 28.363933563232422, 'timing/data': 0.0003616199974203482, 'timing/model': 0.3927208429959137, 'learning_rate/qwen_vl_interface': 8.936967247612678e-06, 'learning_rate/action_model': 8.936967247612678e-05, 'epoch': 0.09}) +[2026-10-03 13:14:48,026][starVLA.training.train_starvla][INFO] - Step 1165, Loss: {'train/loss': 0.4006901979446411, 'train/grad_norm_pre_clip': 75.4352798461914, 'timing/data': 0.00031055998988449574, 'timing/model': 0.3926267039933009, 'learning_rate/qwen_vl_interface': 8.93504640565567e-06, 'learning_rate/action_model': 8.93504640565567e-05, 'epoch': 0.09}) +[2026-10-03 13:14:48,923][starVLA.training.train_starvla][INFO] - Step 1166, Loss: {'train/loss': 9.595610026735812e-05, 'train/grad_norm_pre_clip': 0.021902242675423622, 'timing/data': 0.0004493790038395673, 'timing/model': 0.3924430840124842, 'learning_rate/qwen_vl_interface': 8.933124048915278e-06, 'learning_rate/action_model': 8.933124048915278e-05, 'epoch': 0.09}) +[2026-10-03 13:14:49,626][starVLA.training.train_starvla][INFO] - Step 1167, Loss: {'train/loss': 0.00017388784908689559, 'train/grad_norm_pre_clip': 0.05688416212797165, 'timing/data': 0.00021112000104039907, 'timing/model': 0.3925793829985196, 'learning_rate/qwen_vl_interface': 8.931200178181715e-06, 'learning_rate/action_model': 8.931200178181715e-05, 'epoch': 0.09}) +[2026-10-03 13:14:50,330][starVLA.training.train_starvla][INFO] - Step 1168, Loss: {'train/loss': 0.0003851205692626536, 'train/grad_norm_pre_clip': 0.08053788542747498, 'timing/data': 0.00037868000799790025, 'timing/model': 0.39282750300480984, 'learning_rate/qwen_vl_interface': 8.929274794245812e-06, 'learning_rate/action_model': 8.929274794245812e-05, 'epoch': 0.09}) +[2026-10-03 13:14:51,033][starVLA.training.train_starvla][INFO] - Step 1169, Loss: {'train/loss': 0.0009433858213014901, 'train/grad_norm_pre_clip': 0.6934914588928223, 'timing/data': 0.0003055800043512136, 'timing/model': 0.3924887640023371, 'learning_rate/qwen_vl_interface': 8.927347897899018e-06, 'learning_rate/action_model': 8.927347897899016e-05, 'epoch': 0.09}) +[2026-10-03 13:14:51,738][starVLA.training.train_starvla][INFO] - Step 1170, Loss: {'train/loss': 6.639947241637856e-05, 'train/grad_norm_pre_clip': 0.014021195471286774, 'timing/data': 0.00034567000693641603, 'timing/model': 0.39362096300465055, 'learning_rate/qwen_vl_interface': 8.925419489933407e-06, 'learning_rate/action_model': 8.925419489933408e-05, 'epoch': 0.09}) +[2026-10-03 13:14:52,633][starVLA.training.train_starvla][INFO] - Step 1171, Loss: {'train/loss': 0.34776535630226135, 'train/grad_norm_pre_clip': 30.55243682861328, 'timing/data': 0.00024411000777035952, 'timing/model': 0.3926882130035665, 'learning_rate/qwen_vl_interface': 8.923489571141679e-06, 'learning_rate/action_model': 8.923489571141679e-05, 'epoch': 0.09}) +[2026-10-03 13:14:53,338][starVLA.training.train_starvla][INFO] - Step 1172, Loss: {'train/loss': 0.6070191860198975, 'train/grad_norm_pre_clip': 50.520057678222656, 'timing/data': 0.00024028999905567616, 'timing/model': 0.3933042130083777, 'learning_rate/qwen_vl_interface': 8.921558142317147e-06, 'learning_rate/action_model': 8.921558142317147e-05, 'epoch': 0.09}) +[2026-10-03 13:14:54,040][starVLA.training.train_starvla][INFO] - Step 1173, Loss: {'train/loss': 1.0248353481292725, 'train/grad_norm_pre_clip': 156.020263671875, 'timing/data': 0.0002757699985522777, 'timing/model': 0.39228928300144617, 'learning_rate/qwen_vl_interface': 8.91962520425375e-06, 'learning_rate/action_model': 8.91962520425375e-05, 'epoch': 0.09}) +[2026-10-03 13:14:54,744][starVLA.training.train_starvla][INFO] - Step 1174, Loss: {'train/loss': 0.0006314708152785897, 'train/grad_norm_pre_clip': 0.14941570162773132, 'timing/data': 0.0002110700006596744, 'timing/model': 0.3933090530044865, 'learning_rate/qwen_vl_interface': 8.917690757746044e-06, 'learning_rate/action_model': 8.917690757746044e-05, 'epoch': 0.09}) +[2026-10-03 13:14:55,447][starVLA.training.train_starvla][INFO] - Step 1175, Loss: {'train/loss': 0.001131368917413056, 'train/grad_norm_pre_clip': 0.3983406722545624, 'timing/data': 0.0004389300011098385, 'timing/model': 0.39282706400263123, 'learning_rate/qwen_vl_interface': 8.915754803589208e-06, 'learning_rate/action_model': 8.915754803589209e-05, 'epoch': 0.09}) +[2026-10-03 13:14:56,346][starVLA.training.train_starvla][INFO] - Step 1176, Loss: {'train/loss': 0.6739031076431274, 'train/grad_norm_pre_clip': 28.25450897216797, 'timing/data': 0.000281489992630668, 'timing/model': 0.5866718600009335, 'learning_rate/qwen_vl_interface': 8.913817342579042e-06, 'learning_rate/action_model': 8.913817342579042e-05, 'epoch': 0.09}) +[2026-10-03 13:14:57,049][starVLA.training.train_starvla][INFO] - Step 1177, Loss: {'train/loss': 0.34421664476394653, 'train/grad_norm_pre_clip': 9.967782974243164, 'timing/data': 0.00025068999093491584, 'timing/model': 0.3930470930063166, 'learning_rate/qwen_vl_interface': 8.91187837551196e-06, 'learning_rate/action_model': 8.911878375511961e-05, 'epoch': 0.09}) +[2026-10-03 13:14:57,753][starVLA.training.train_starvla][INFO] - Step 1178, Loss: {'train/loss': 0.0011064715217798948, 'train/grad_norm_pre_clip': 0.24668489396572113, 'timing/data': 0.00038004999805707484, 'timing/model': 0.39317016300628893, 'learning_rate/qwen_vl_interface': 8.909937903185e-06, 'learning_rate/action_model': 8.909937903185e-05, 'epoch': 0.09}) +[2026-10-03 13:14:58,456][starVLA.training.train_starvla][INFO] - Step 1179, Loss: {'train/loss': 0.051843881607055664, 'train/grad_norm_pre_clip': 34.05794143676758, 'timing/data': 0.00028902001213282347, 'timing/model': 0.3927324529940961, 'learning_rate/qwen_vl_interface': 8.907995926395814e-06, 'learning_rate/action_model': 8.907995926395815e-05, 'epoch': 0.09}) +[2026-10-03 13:14:59,161][starVLA.training.train_starvla][INFO] - Step 1180, Loss: {'train/loss': 0.1904974728822708, 'train/grad_norm_pre_clip': 29.765592575073242, 'timing/data': 0.00028315000236034393, 'timing/model': 0.39300975399964955, 'learning_rate/qwen_vl_interface': 8.90605244594268e-06, 'learning_rate/action_model': 8.906052445942679e-05, 'epoch': 0.09}) +[2026-10-03 13:15:00,056][starVLA.training.train_starvla][INFO] - Step 1181, Loss: {'train/loss': 0.04706208407878876, 'train/grad_norm_pre_clip': 22.789134979248047, 'timing/data': 0.00025220999668817967, 'timing/model': 0.5848839700047392, 'learning_rate/qwen_vl_interface': 8.904107462624484e-06, 'learning_rate/action_model': 8.904107462624483e-05, 'epoch': 0.09}) +[2026-10-03 13:15:00,760][starVLA.training.train_starvla][INFO] - Step 1182, Loss: {'train/loss': 0.012437744066119194, 'train/grad_norm_pre_clip': 3.2137887477874756, 'timing/data': 0.00020534000941552222, 'timing/model': 0.393009343999438, 'learning_rate/qwen_vl_interface': 8.902160977240739e-06, 'learning_rate/action_model': 8.902160977240739e-05, 'epoch': 0.09}) +[2026-10-03 13:15:01,462][starVLA.training.train_starvla][INFO] - Step 1183, Loss: {'train/loss': 0.3653962016105652, 'train/grad_norm_pre_clip': 30.876720428466797, 'timing/data': 0.0003292799956398085, 'timing/model': 0.3927121629967587, 'learning_rate/qwen_vl_interface': 8.90021299059157e-06, 'learning_rate/action_model': 8.90021299059157e-05, 'epoch': 0.09}) +[2026-10-03 13:15:02,167][starVLA.training.train_starvla][INFO] - Step 1184, Loss: {'train/loss': 0.014625666663050652, 'train/grad_norm_pre_clip': 7.813931465148926, 'timing/data': 0.00029101000109221786, 'timing/model': 0.393255943999975, 'learning_rate/qwen_vl_interface': 8.898263503477722e-06, 'learning_rate/action_model': 8.898263503477721e-05, 'epoch': 0.09}) +[2026-10-03 13:15:02,870][starVLA.training.train_starvla][INFO] - Step 1185, Loss: {'train/loss': 0.034294113516807556, 'train/grad_norm_pre_clip': 17.353662490844727, 'timing/data': 0.00035185000160709023, 'timing/model': 0.39193115300440695, 'learning_rate/qwen_vl_interface': 8.896312516700553e-06, 'learning_rate/action_model': 8.896312516700554e-05, 'epoch': 0.09}) +[2026-10-03 13:15:03,573][starVLA.training.train_starvla][INFO] - Step 1186, Loss: {'train/loss': 0.04280854016542435, 'train/grad_norm_pre_clip': 34.116546630859375, 'timing/data': 0.00021065000328235328, 'timing/model': 0.3936545039905468, 'learning_rate/qwen_vl_interface': 8.894360031062043e-06, 'learning_rate/action_model': 8.894360031062043e-05, 'epoch': 0.09}) +[2026-10-03 13:15:04,470][starVLA.training.train_starvla][INFO] - Step 1187, Loss: {'train/loss': 0.45538556575775146, 'train/grad_norm_pre_clip': 16.664613723754883, 'timing/data': 0.0003369999903952703, 'timing/model': 0.3926288530055899, 'learning_rate/qwen_vl_interface': 8.892406047364783e-06, 'learning_rate/action_model': 8.892406047364784e-05, 'epoch': 0.09}) +[2026-10-03 13:15:05,174][starVLA.training.train_starvla][INFO] - Step 1188, Loss: {'train/loss': 0.008165116421878338, 'train/grad_norm_pre_clip': 3.0629525184631348, 'timing/data': 0.00029582998831756413, 'timing/model': 0.3928886740031885, 'learning_rate/qwen_vl_interface': 8.890450566411984e-06, 'learning_rate/action_model': 8.890450566411984e-05, 'epoch': 0.09}) +[2026-10-03 13:15:05,878][starVLA.training.train_starvla][INFO] - Step 1189, Loss: {'train/loss': 0.007784952409565449, 'train/grad_norm_pre_clip': 1.3942335844039917, 'timing/data': 0.00023567999596707523, 'timing/model': 0.3926480629888829, 'learning_rate/qwen_vl_interface': 8.888493589007467e-06, 'learning_rate/action_model': 8.888493589007468e-05, 'epoch': 0.09}) +[2026-10-03 13:15:06,586][starVLA.training.train_starvla][INFO] - Step 1190, Loss: {'train/loss': 0.012183964252471924, 'train/grad_norm_pre_clip': 5.769583702087402, 'timing/data': 0.00037280899414326996, 'timing/model': 0.39388917401083745, 'learning_rate/qwen_vl_interface': 8.886535115955676e-06, 'learning_rate/action_model': 8.886535115955675e-05, 'epoch': 0.09}) +[2026-10-03 13:15:07,289][starVLA.training.train_starvla][INFO] - Step 1191, Loss: {'train/loss': 0.0007682928116992116, 'train/grad_norm_pre_clip': 0.15527307987213135, 'timing/data': 0.0003324999997857958, 'timing/model': 0.3928349430061644, 'learning_rate/qwen_vl_interface': 8.884575148061662e-06, 'learning_rate/action_model': 8.88457514806166e-05, 'epoch': 0.09}) +[2026-10-03 13:15:08,185][starVLA.training.train_starvla][INFO] - Step 1192, Loss: {'train/loss': 0.002724583027884364, 'train/grad_norm_pre_clip': 0.39313188195228577, 'timing/data': 0.0003414800012251362, 'timing/model': 0.5843445509963203, 'learning_rate/qwen_vl_interface': 8.882613686131095e-06, 'learning_rate/action_model': 8.882613686131093e-05, 'epoch': 0.09}) +[2026-10-03 13:15:08,887][starVLA.training.train_starvla][INFO] - Step 1193, Loss: {'train/loss': 0.00793258287012577, 'train/grad_norm_pre_clip': 1.8274682760238647, 'timing/data': 0.0003453399986028671, 'timing/model': 0.392281082997215, 'learning_rate/qwen_vl_interface': 8.880650730970254e-06, 'learning_rate/action_model': 8.880650730970254e-05, 'epoch': 0.09}) +[2026-10-03 13:15:09,591][starVLA.training.train_starvla][INFO] - Step 1194, Loss: {'train/loss': 0.8370212912559509, 'train/grad_norm_pre_clip': 59.630123138427734, 'timing/data': 0.00034686000435613096, 'timing/model': 0.3932438540068688, 'learning_rate/qwen_vl_interface': 8.878686283386042e-06, 'learning_rate/action_model': 8.878686283386042e-05, 'epoch': 0.09}) +[2026-10-03 13:15:10,294][starVLA.training.train_starvla][INFO] - Step 1195, Loss: {'train/loss': 0.003474349854514003, 'train/grad_norm_pre_clip': 0.889309287071228, 'timing/data': 0.0003281899989815429, 'timing/model': 0.3927217229938833, 'learning_rate/qwen_vl_interface': 8.876720344185967e-06, 'learning_rate/action_model': 8.876720344185967e-05, 'epoch': 0.09}) +[2026-10-03 13:15:10,998][starVLA.training.train_starvla][INFO] - Step 1196, Loss: {'train/loss': 0.0006756834918633103, 'train/grad_norm_pre_clip': 0.1937507838010788, 'timing/data': 0.00036094999813940376, 'timing/model': 0.392731833999278, 'learning_rate/qwen_vl_interface': 8.874752914178149e-06, 'learning_rate/action_model': 8.87475291417815e-05, 'epoch': 0.09}) +[2026-10-03 13:15:11,700][starVLA.training.train_starvla][INFO] - Step 1197, Loss: {'train/loss': 0.8903107047080994, 'train/grad_norm_pre_clip': 46.2001953125, 'timing/data': 0.00031123000371735543, 'timing/model': 0.39238856300653424, 'learning_rate/qwen_vl_interface': 8.87278399417133e-06, 'learning_rate/action_model': 8.872783994171329e-05, 'epoch': 0.09}) +[2026-10-03 13:15:12,594][starVLA.training.train_starvla][INFO] - Step 1198, Loss: {'train/loss': 0.4734327495098114, 'train/grad_norm_pre_clip': 31.461559295654297, 'timing/data': 0.00021795999782625586, 'timing/model': 0.3934431039961055, 'learning_rate/qwen_vl_interface': 8.870813584974852e-06, 'learning_rate/action_model': 8.870813584974852e-05, 'epoch': 0.09}) +[2026-10-03 13:15:13,297][starVLA.training.train_starvla][INFO] - Step 1199, Loss: {'train/loss': 0.01617436110973358, 'train/grad_norm_pre_clip': 13.063718795776367, 'timing/data': 0.00031924998620525, 'timing/model': 0.392776012988179, 'learning_rate/qwen_vl_interface': 8.868841687398683e-06, 'learning_rate/action_model': 8.868841687398683e-05, 'epoch': 0.09}) +[2026-10-03 13:15:14,003][starVLA.training.train_starvla][INFO] - Step 1200, Loss: {'train/loss': 0.0012456441763788462, 'train/grad_norm_pre_clip': 1.1757553815841675, 'timing/data': 0.0003146600065520033, 'timing/model': 0.3930620730097871, 'learning_rate/qwen_vl_interface': 8.86686830225339e-06, 'learning_rate/action_model': 8.866868302253389e-05, 'epoch': 0.09}) +[2026-10-03 13:15:14,706][starVLA.training.train_starvla][INFO] - Step 1201, Loss: {'train/loss': 0.00022059000912122428, 'train/grad_norm_pre_clip': 0.07115612924098969, 'timing/data': 0.00029521000396925956, 'timing/model': 0.3924644639919279, 'learning_rate/qwen_vl_interface': 8.864893430350159e-06, 'learning_rate/action_model': 8.864893430350159e-05, 'epoch': 0.09}) +[2026-10-03 13:15:15,410][starVLA.training.train_starvla][INFO] - Step 1202, Loss: {'train/loss': 0.2658388912677765, 'train/grad_norm_pre_clip': 15.207603454589844, 'timing/data': 0.0003756900114240125, 'timing/model': 0.39310189300158527, 'learning_rate/qwen_vl_interface': 8.862917072500785e-06, 'learning_rate/action_model': 8.862917072500784e-05, 'epoch': 0.09}) +[2026-10-03 13:15:16,306][starVLA.training.train_starvla][INFO] - Step 1203, Loss: {'train/loss': 0.0041132778860628605, 'train/grad_norm_pre_clip': 8.121430397033691, 'timing/data': 0.00033216000883840024, 'timing/model': 0.39280601299833506, 'learning_rate/qwen_vl_interface': 8.860939229517674e-06, 'learning_rate/action_model': 8.860939229517674e-05, 'epoch': 0.09}) +[2026-10-03 13:15:17,010][starVLA.training.train_starvla][INFO] - Step 1204, Loss: {'train/loss': 0.003591547254472971, 'train/grad_norm_pre_clip': 1.0860975980758667, 'timing/data': 0.0003472199896350503, 'timing/model': 0.3927431140036788, 'learning_rate/qwen_vl_interface': 8.858959902213844e-06, 'learning_rate/action_model': 8.858959902213845e-05, 'epoch': 0.09}) +[2026-10-03 13:15:17,712][starVLA.training.train_starvla][INFO] - Step 1205, Loss: {'train/loss': 0.000301994412438944, 'train/grad_norm_pre_clip': 0.07146468013525009, 'timing/data': 0.0003135800070594996, 'timing/model': 0.39230243299971335, 'learning_rate/qwen_vl_interface': 8.856979091402921e-06, 'learning_rate/action_model': 8.85697909140292e-05, 'epoch': 0.09}) +[2026-10-03 13:15:18,415][starVLA.training.train_starvla][INFO] - Step 1206, Loss: {'train/loss': 0.5264424681663513, 'train/grad_norm_pre_clip': 35.76013946533203, 'timing/data': 0.00021040999854449183, 'timing/model': 0.392560444000992, 'learning_rate/qwen_vl_interface': 8.854996797899139e-06, 'learning_rate/action_model': 8.854996797899139e-05, 'epoch': 0.09}) +[2026-10-03 13:15:19,118][starVLA.training.train_starvla][INFO] - Step 1207, Loss: {'train/loss': 0.00018497784913051873, 'train/grad_norm_pre_clip': 0.05696927756071091, 'timing/data': 0.000324309992720373, 'timing/model': 0.39339619199745357, 'learning_rate/qwen_vl_interface': 8.853013022517348e-06, 'learning_rate/action_model': 8.853013022517348e-05, 'epoch': 0.09}) +[2026-10-03 13:15:20,016][starVLA.training.train_starvla][INFO] - Step 1208, Loss: {'train/loss': 0.009621983394026756, 'train/grad_norm_pre_clip': 7.967092990875244, 'timing/data': 0.00047441000060644, 'timing/model': 0.5859529199951794, 'learning_rate/qwen_vl_interface': 8.851027766073002e-06, 'learning_rate/action_model': 8.851027766073001e-05, 'epoch': 0.09}) +[2026-10-03 13:15:20,717][starVLA.training.train_starvla][INFO] - Step 1209, Loss: {'train/loss': 0.0014834983740001917, 'train/grad_norm_pre_clip': 0.3326500952243805, 'timing/data': 0.00020788000256288797, 'timing/model': 0.3919045239890693, 'learning_rate/qwen_vl_interface': 8.84904102938216e-06, 'learning_rate/action_model': 8.84904102938216e-05, 'epoch': 0.09}) +[2026-10-03 13:15:21,420][starVLA.training.train_starvla][INFO] - Step 1210, Loss: {'train/loss': 0.36255529522895813, 'train/grad_norm_pre_clip': 42.2320671081543, 'timing/data': 0.0004030599957332015, 'timing/model': 0.3933102930022869, 'learning_rate/qwen_vl_interface': 8.8470528132615e-06, 'learning_rate/action_model': 8.8470528132615e-05, 'epoch': 0.09}) +[2026-10-03 13:15:22,122][starVLA.training.train_starvla][INFO] - Step 1211, Loss: {'train/loss': 0.0002446684520691633, 'train/grad_norm_pre_clip': 0.08853346854448318, 'timing/data': 0.00034767000761348754, 'timing/model': 0.3922223640111042, 'learning_rate/qwen_vl_interface': 8.845063118528303e-06, 'learning_rate/action_model': 8.845063118528304e-05, 'epoch': 0.09}) +[2026-10-03 13:15:22,826][starVLA.training.train_starvla][INFO] - Step 1212, Loss: {'train/loss': 5.29656172147952e-05, 'train/grad_norm_pre_clip': 0.019524332135915756, 'timing/data': 0.0002605699992273003, 'timing/model': 0.3926994830108015, 'learning_rate/qwen_vl_interface': 8.843071946000455e-06, 'learning_rate/action_model': 8.843071946000456e-05, 'epoch': 0.09}) +[2026-10-03 13:15:23,528][starVLA.training.train_starvla][INFO] - Step 1213, Loss: {'train/loss': 0.00029065069975331426, 'train/grad_norm_pre_clip': 0.16562248766422272, 'timing/data': 0.000272019999101758, 'timing/model': 0.3921177140000509, 'learning_rate/qwen_vl_interface': 8.841079296496451e-06, 'learning_rate/action_model': 8.84107929649645e-05, 'epoch': 0.09}) +[2026-10-03 13:15:24,422][starVLA.training.train_starvla][INFO] - Step 1214, Loss: {'train/loss': 0.0008987970068119466, 'train/grad_norm_pre_clip': 0.1495443880558014, 'timing/data': 0.00036235999141354114, 'timing/model': 0.3923175829986576, 'learning_rate/qwen_vl_interface': 8.839085170835396e-06, 'learning_rate/action_model': 8.839085170835395e-05, 'epoch': 0.09}) +[2026-10-03 13:15:25,124][starVLA.training.train_starvla][INFO] - Step 1215, Loss: {'train/loss': 0.001791802467778325, 'train/grad_norm_pre_clip': 0.36853528022766113, 'timing/data': 0.0002825800038408488, 'timing/model': 0.3925124739907915, 'learning_rate/qwen_vl_interface': 8.837089569837e-06, 'learning_rate/action_model': 8.837089569836998e-05, 'epoch': 0.09}) +[2026-10-03 13:15:25,827][starVLA.training.train_starvla][INFO] - Step 1216, Loss: {'train/loss': 0.0005486963782459497, 'train/grad_norm_pre_clip': 0.38478514552116394, 'timing/data': 0.0002624600019771606, 'timing/model': 0.3920860829966841, 'learning_rate/qwen_vl_interface': 8.835092494321575e-06, 'learning_rate/action_model': 8.835092494321575e-05, 'epoch': 0.09}) +[2026-10-03 13:15:26,530][starVLA.training.train_starvla][INFO] - Step 1217, Loss: {'train/loss': 0.3606188893318176, 'train/grad_norm_pre_clip': 54.113258361816406, 'timing/data': 0.00032500000088475645, 'timing/model': 0.3926722530013649, 'learning_rate/qwen_vl_interface': 8.833093945110048e-06, 'learning_rate/action_model': 8.833093945110047e-05, 'epoch': 0.09}) +[2026-10-03 13:15:27,233][starVLA.training.train_starvla][INFO] - Step 1218, Loss: {'train/loss': 0.00044483368401415646, 'train/grad_norm_pre_clip': 0.09358152747154236, 'timing/data': 0.00033055999665521085, 'timing/model': 0.39265845400223043, 'learning_rate/qwen_vl_interface': 8.831093923023946e-06, 'learning_rate/action_model': 8.831093923023945e-05, 'epoch': 0.09}) +[2026-10-03 13:15:28,126][starVLA.training.train_starvla][INFO] - Step 1219, Loss: {'train/loss': 0.00022106108372099698, 'train/grad_norm_pre_clip': 0.04371337965130806, 'timing/data': 0.00022159000218380243, 'timing/model': 0.3926581430132501, 'learning_rate/qwen_vl_interface': 8.8290924288854e-06, 'learning_rate/action_model': 8.8290924288854e-05, 'epoch': 0.09}) +[2026-10-03 13:15:28,831][starVLA.training.train_starvla][INFO] - Step 1220, Loss: {'train/loss': 0.4502938985824585, 'train/grad_norm_pre_clip': 105.08512878417969, 'timing/data': 0.0003173300065100193, 'timing/model': 0.3932532140024705, 'learning_rate/qwen_vl_interface': 8.827089463517153e-06, 'learning_rate/action_model': 8.827089463517153e-05, 'epoch': 0.09}) +[2026-10-03 13:15:29,533][starVLA.training.train_starvla][INFO] - Step 1221, Loss: {'train/loss': 0.004007645882666111, 'train/grad_norm_pre_clip': 5.921148300170898, 'timing/data': 0.0002490799961378798, 'timing/model': 0.3917970539914677, 'learning_rate/qwen_vl_interface': 8.825085027742545e-06, 'learning_rate/action_model': 8.825085027742544e-05, 'epoch': 0.09}) +[2026-10-03 13:15:30,235][starVLA.training.train_starvla][INFO] - Step 1222, Loss: {'train/loss': 0.40062376856803894, 'train/grad_norm_pre_clip': 32.82592010498047, 'timing/data': 0.00035147999005857855, 'timing/model': 0.3921967640053481, 'learning_rate/qwen_vl_interface': 8.823079122385523e-06, 'learning_rate/action_model': 8.823079122385523e-05, 'epoch': 0.09}) +[2026-10-03 13:15:30,935][starVLA.training.train_starvla][INFO] - Step 1223, Loss: {'train/loss': 0.0005087417084723711, 'train/grad_norm_pre_clip': 0.08858432620763779, 'timing/data': 0.00033911000355146825, 'timing/model': 0.3911601729923859, 'learning_rate/qwen_vl_interface': 8.821071748270644e-06, 'learning_rate/action_model': 8.821071748270644e-05, 'epoch': 0.09}) +[2026-10-03 13:15:31,828][starVLA.training.train_starvla][INFO] - Step 1224, Loss: {'train/loss': 0.00016178848454728723, 'train/grad_norm_pre_clip': 0.03518809750676155, 'timing/data': 0.00021186000958550721, 'timing/model': 0.5825111710000783, 'learning_rate/qwen_vl_interface': 8.81906290622306e-06, 'learning_rate/action_model': 8.81906290622306e-05, 'epoch': 0.09}) +[2026-10-03 13:15:32,529][starVLA.training.train_starvla][INFO] - Step 1225, Loss: {'train/loss': 0.001518087345175445, 'train/grad_norm_pre_clip': 2.0822980403900146, 'timing/data': 0.00028866999491583556, 'timing/model': 0.39148298400687054, 'learning_rate/qwen_vl_interface': 8.81705259706853e-06, 'learning_rate/action_model': 8.817052597068529e-05, 'epoch': 0.09}) +[2026-10-03 13:15:33,231][starVLA.training.train_starvla][INFO] - Step 1226, Loss: {'train/loss': 0.6824912428855896, 'train/grad_norm_pre_clip': 17.634138107299805, 'timing/data': 0.000291709991870448, 'timing/model': 0.3924633839924354, 'learning_rate/qwen_vl_interface': 8.815040821633416e-06, 'learning_rate/action_model': 8.815040821633416e-05, 'epoch': 0.09}) +[2026-10-03 13:15:33,934][starVLA.training.train_starvla][INFO] - Step 1227, Loss: {'train/loss': 0.0009296464850194752, 'train/grad_norm_pre_clip': 1.3703184127807617, 'timing/data': 0.0003271899913670495, 'timing/model': 0.39263015300093684, 'learning_rate/qwen_vl_interface': 8.813027580744685e-06, 'learning_rate/action_model': 8.813027580744684e-05, 'epoch': 0.09}) +[2026-10-03 13:15:34,637][starVLA.training.train_starvla][INFO] - Step 1228, Loss: {'train/loss': 0.6008687615394592, 'train/grad_norm_pre_clip': 94.66975402832031, 'timing/data': 0.00025191999156959355, 'timing/model': 0.3926024030079134, 'learning_rate/qwen_vl_interface': 8.8110128752299e-06, 'learning_rate/action_model': 8.8110128752299e-05, 'epoch': 0.09}) +[2026-10-03 13:15:35,532][starVLA.training.train_starvla][INFO] - Step 1229, Loss: {'train/loss': 0.00040222046663984656, 'train/grad_norm_pre_clip': 0.1446761041879654, 'timing/data': 0.00034897901059594005, 'timing/model': 0.5849879820016213, 'learning_rate/qwen_vl_interface': 8.808996705917235e-06, 'learning_rate/action_model': 8.808996705917235e-05, 'epoch': 0.09}) +[2026-10-03 13:15:36,237][starVLA.training.train_starvla][INFO] - Step 1230, Loss: {'train/loss': 0.0009302626713179052, 'train/grad_norm_pre_clip': 0.19519802927970886, 'timing/data': 0.0002573590027168393, 'timing/model': 0.3940826839971123, 'learning_rate/qwen_vl_interface': 8.80697907363546e-06, 'learning_rate/action_model': 8.80697907363546e-05, 'epoch': 0.09}) +[2026-10-03 13:15:36,939][starVLA.training.train_starvla][INFO] - Step 1231, Loss: {'train/loss': 0.3626059591770172, 'train/grad_norm_pre_clip': 37.00752639770508, 'timing/data': 0.00025749999622348696, 'timing/model': 0.3924090829968918, 'learning_rate/qwen_vl_interface': 8.804959979213945e-06, 'learning_rate/action_model': 8.804959979213945e-05, 'epoch': 0.09}) +[2026-10-03 13:15:37,643][starVLA.training.train_starvla][INFO] - Step 1232, Loss: {'train/loss': 0.4716564118862152, 'train/grad_norm_pre_clip': 47.018653869628906, 'timing/data': 0.00030973998946137726, 'timing/model': 0.39263463400129694, 'learning_rate/qwen_vl_interface': 8.802939423482667e-06, 'learning_rate/action_model': 8.802939423482666e-05, 'epoch': 0.09}) +[2026-10-03 13:15:38,345][starVLA.training.train_starvla][INFO] - Step 1233, Loss: {'train/loss': 0.6260011196136475, 'train/grad_norm_pre_clip': 134.91822814941406, 'timing/data': 0.00023253000108525157, 'timing/model': 0.3915358340018429, 'learning_rate/qwen_vl_interface': 8.800917407272198e-06, 'learning_rate/action_model': 8.800917407272198e-05, 'epoch': 0.09}) +[2026-10-03 13:15:39,239][starVLA.training.train_starvla][INFO] - Step 1234, Loss: {'train/loss': 0.5473546981811523, 'train/grad_norm_pre_clip': 23.070716857910156, 'timing/data': 0.00033349999284837395, 'timing/model': 0.583786090996, 'learning_rate/qwen_vl_interface': 8.798893931413711e-06, 'learning_rate/action_model': 8.79889393141371e-05, 'epoch': 0.09}) +[2026-10-03 13:15:39,941][starVLA.training.train_starvla][INFO] - Step 1235, Loss: {'train/loss': 0.262294739484787, 'train/grad_norm_pre_clip': 192.0498046875, 'timing/data': 0.0002582000015536323, 'timing/model': 0.3924832540069474, 'learning_rate/qwen_vl_interface': 8.796868996738985e-06, 'learning_rate/action_model': 8.796868996738984e-05, 'epoch': 0.09}) +[2026-10-03 13:15:40,644][starVLA.training.train_starvla][INFO] - Step 1236, Loss: {'train/loss': 0.0004631757619790733, 'train/grad_norm_pre_clip': 0.12185857445001602, 'timing/data': 0.0003698900109156966, 'timing/model': 0.39213835400005337, 'learning_rate/qwen_vl_interface': 8.794842604080389e-06, 'learning_rate/action_model': 8.794842604080389e-05, 'epoch': 0.09}) +[2026-10-03 13:15:41,346][starVLA.training.train_starvla][INFO] - Step 1237, Loss: {'train/loss': 0.00038265364128164947, 'train/grad_norm_pre_clip': 0.16035255789756775, 'timing/data': 0.0002683999919099733, 'timing/model': 0.3920171039935667, 'learning_rate/qwen_vl_interface': 8.7928147542709e-06, 'learning_rate/action_model': 8.7928147542709e-05, 'epoch': 0.09}) +[2026-10-03 13:15:42,049][starVLA.training.train_starvla][INFO] - Step 1238, Loss: {'train/loss': 0.004391918424516916, 'train/grad_norm_pre_clip': 3.2069995403289795, 'timing/data': 0.0003104700008407235, 'timing/model': 0.39250238401291426, 'learning_rate/qwen_vl_interface': 8.790785448144088e-06, 'learning_rate/action_model': 8.790785448144088e-05, 'epoch': 0.09}) +[2026-10-03 13:15:42,751][starVLA.training.train_starvla][INFO] - Step 1239, Loss: {'train/loss': 0.0008094539516605437, 'train/grad_norm_pre_clip': 0.19079884886741638, 'timing/data': 0.0002109689958160743, 'timing/model': 0.39304637401073705, 'learning_rate/qwen_vl_interface': 8.788754686534126e-06, 'learning_rate/action_model': 8.788754686534127e-05, 'epoch': 0.09}) +[2026-10-03 13:15:43,650][starVLA.training.train_starvla][INFO] - Step 1240, Loss: {'train/loss': 0.4694726765155792, 'train/grad_norm_pre_clip': 44.84965133666992, 'timing/data': 0.0004348299989942461, 'timing/model': 0.39296841400209814, 'learning_rate/qwen_vl_interface': 8.786722470275784e-06, 'learning_rate/action_model': 8.786722470275784e-05, 'epoch': 0.09}) +[2026-10-03 13:15:44,351][starVLA.training.train_starvla][INFO] - Step 1241, Loss: {'train/loss': 0.002439648611471057, 'train/grad_norm_pre_clip': 0.30543777346611023, 'timing/data': 0.00026853999588638544, 'timing/model': 0.390890042996034, 'learning_rate/qwen_vl_interface': 8.784688800204427e-06, 'learning_rate/action_model': 8.784688800204426e-05, 'epoch': 0.09}) +[2026-10-03 13:15:45,055][starVLA.training.train_starvla][INFO] - Step 1242, Loss: {'train/loss': 0.0047284821048378944, 'train/grad_norm_pre_clip': 0.6796542406082153, 'timing/data': 0.00021247999393381178, 'timing/model': 0.39330330400844105, 'learning_rate/qwen_vl_interface': 8.782653677156022e-06, 'learning_rate/action_model': 8.782653677156022e-05, 'epoch': 0.09}) +[2026-10-03 13:15:45,759][starVLA.training.train_starvla][INFO] - Step 1243, Loss: {'train/loss': 0.0008925521979108453, 'train/grad_norm_pre_clip': 0.28224220871925354, 'timing/data': 0.00025241999537684023, 'timing/model': 0.3932951840106398, 'learning_rate/qwen_vl_interface': 8.780617101967129e-06, 'learning_rate/action_model': 8.780617101967129e-05, 'epoch': 0.09}) +[2026-10-03 13:15:46,463][starVLA.training.train_starvla][INFO] - Step 1244, Loss: {'train/loss': 0.0021883719600737095, 'train/grad_norm_pre_clip': 0.6171892881393433, 'timing/data': 0.0003416899999137968, 'timing/model': 0.39304734399775043, 'learning_rate/qwen_vl_interface': 8.77857907547491e-06, 'learning_rate/action_model': 8.77857907547491e-05, 'epoch': 0.09}) +[2026-10-03 13:15:47,358][starVLA.training.train_starvla][INFO] - Step 1245, Loss: {'train/loss': 0.38855668902397156, 'train/grad_norm_pre_clip': 29.543907165527344, 'timing/data': 0.0003362500137882307, 'timing/model': 0.583581981001771, 'learning_rate/qwen_vl_interface': 8.776539598517119e-06, 'learning_rate/action_model': 8.776539598517118e-05, 'epoch': 0.09}) +[2026-10-03 13:15:48,062][starVLA.training.train_starvla][INFO] - Step 1246, Loss: {'train/loss': 0.006563901901245117, 'train/grad_norm_pre_clip': 2.469325065612793, 'timing/data': 0.00038205999589990824, 'timing/model': 0.3927178230078425, 'learning_rate/qwen_vl_interface': 8.774498671932108e-06, 'learning_rate/action_model': 8.774498671932107e-05, 'epoch': 0.09}) +[2026-10-03 13:15:48,765][starVLA.training.train_starvla][INFO] - Step 1247, Loss: {'train/loss': 0.0013010280672460794, 'train/grad_norm_pre_clip': 0.18197302520275116, 'timing/data': 0.00032372999703511596, 'timing/model': 0.3927607929945225, 'learning_rate/qwen_vl_interface': 8.772456296558824e-06, 'learning_rate/action_model': 8.772456296558824e-05, 'epoch': 0.09}) +[2026-10-03 13:15:49,469][starVLA.training.train_starvla][INFO] - Step 1248, Loss: {'train/loss': 1.0969215631484985, 'train/grad_norm_pre_clip': 30.806503295898438, 'timing/data': 0.00040282998816110194, 'timing/model': 0.3931236040079966, 'learning_rate/qwen_vl_interface': 8.77041247323681e-06, 'learning_rate/action_model': 8.770412473236811e-05, 'epoch': 0.09}) +[2026-10-03 13:15:50,172][starVLA.training.train_starvla][INFO] - Step 1249, Loss: {'train/loss': 0.27465224266052246, 'train/grad_norm_pre_clip': 82.35762786865234, 'timing/data': 0.0003464800101937726, 'timing/model': 0.39207104399974924, 'learning_rate/qwen_vl_interface': 8.76836720280621e-06, 'learning_rate/action_model': 8.768367202806209e-05, 'epoch': 0.09}) +[2026-10-03 13:15:50,875][starVLA.training.train_starvla][INFO] - Step 1250, Loss: {'train/loss': 0.028472470119595528, 'train/grad_norm_pre_clip': 35.084228515625, 'timing/data': 0.00027859999681822956, 'timing/model': 0.39320762299757916, 'learning_rate/qwen_vl_interface': 8.766320486107749e-06, 'learning_rate/action_model': 8.766320486107749e-05, 'epoch': 0.09}) +[2026-10-03 13:15:51,771][starVLA.training.train_starvla][INFO] - Step 1251, Loss: {'train/loss': 0.4044998288154602, 'train/grad_norm_pre_clip': 26.950143814086914, 'timing/data': 0.0002710800035856664, 'timing/model': 0.3936811429884983, 'learning_rate/qwen_vl_interface': 8.76427232398276e-06, 'learning_rate/action_model': 8.76427232398276e-05, 'epoch': 0.09}) +[2026-10-03 13:15:52,476][starVLA.training.train_starvla][INFO] - Step 1252, Loss: {'train/loss': 0.25961342453956604, 'train/grad_norm_pre_clip': 106.95429229736328, 'timing/data': 0.0002271999983349815, 'timing/model': 0.3929458729980979, 'learning_rate/qwen_vl_interface': 8.762222717273167e-06, 'learning_rate/action_model': 8.762222717273167e-05, 'epoch': 0.09}) +[2026-10-03 13:15:53,179][starVLA.training.train_starvla][INFO] - Step 1253, Loss: {'train/loss': 0.007424307055771351, 'train/grad_norm_pre_clip': 4.9999775886535645, 'timing/data': 0.0003395300009287894, 'timing/model': 0.39281198399839923, 'learning_rate/qwen_vl_interface': 8.76017166682148e-06, 'learning_rate/action_model': 8.760171666821481e-05, 'epoch': 0.09}) +[2026-10-03 13:15:53,884][starVLA.training.train_starvla][INFO] - Step 1254, Loss: {'train/loss': 0.03882395848631859, 'train/grad_norm_pre_clip': 42.807716369628906, 'timing/data': 0.0003796800010604784, 'timing/model': 0.39314475399442017, 'learning_rate/qwen_vl_interface': 8.758119173470815e-06, 'learning_rate/action_model': 8.758119173470815e-05, 'epoch': 0.09}) +[2026-10-03 13:15:54,587][starVLA.training.train_starvla][INFO] - Step 1255, Loss: {'train/loss': 0.0004295053076930344, 'train/grad_norm_pre_clip': 0.05450107529759407, 'timing/data': 0.00033617000735830516, 'timing/model': 0.39285596400441136, 'learning_rate/qwen_vl_interface': 8.75606523806487e-06, 'learning_rate/action_model': 8.756065238064868e-05, 'epoch': 0.09}) +[2026-10-03 13:15:55,487][starVLA.training.train_starvla][INFO] - Step 1256, Loss: {'train/loss': 0.006800975650548935, 'train/grad_norm_pre_clip': 8.786816596984863, 'timing/data': 0.00040131001151166856, 'timing/model': 0.393030973005807, 'learning_rate/qwen_vl_interface': 8.75400986144794e-06, 'learning_rate/action_model': 8.75400986144794e-05, 'epoch': 0.09}) +[2026-10-03 13:15:56,189][starVLA.training.train_starvla][INFO] - Step 1257, Loss: {'train/loss': 0.3918692469596863, 'train/grad_norm_pre_clip': 77.88101959228516, 'timing/data': 0.00026254000840708613, 'timing/model': 0.3921950539952377, 'learning_rate/qwen_vl_interface': 8.751953044464913e-06, 'learning_rate/action_model': 8.751953044464913e-05, 'epoch': 0.09}) +[2026-10-03 13:15:56,893][starVLA.training.train_starvla][INFO] - Step 1258, Loss: {'train/loss': 0.0008279953035525978, 'train/grad_norm_pre_clip': 0.20153400301933289, 'timing/data': 0.00020840000070165843, 'timing/model': 0.3931751039926894, 'learning_rate/qwen_vl_interface': 8.749894787961272e-06, 'learning_rate/action_model': 8.749894787961271e-05, 'epoch': 0.09}) +[2026-10-03 13:15:57,595][starVLA.training.train_starvla][INFO] - Step 1259, Loss: {'train/loss': 0.4694710373878479, 'train/grad_norm_pre_clip': 67.13388061523438, 'timing/data': 0.00034335000964347273, 'timing/model': 0.3929723839974031, 'learning_rate/qwen_vl_interface': 8.747835092783086e-06, 'learning_rate/action_model': 8.747835092783085e-05, 'epoch': 0.09}) +[2026-10-03 13:15:58,299][starVLA.training.train_starvla][INFO] - Step 1260, Loss: {'train/loss': 0.3223302662372589, 'train/grad_norm_pre_clip': 92.39148712158203, 'timing/data': 0.0003709500015247613, 'timing/model': 0.39246100300806575, 'learning_rate/qwen_vl_interface': 8.745773959777018e-06, 'learning_rate/action_model': 8.745773959777017e-05, 'epoch': 0.09}) +[2026-10-03 13:15:59,196][starVLA.training.train_starvla][INFO] - Step 1261, Loss: {'train/loss': 0.0003090329119004309, 'train/grad_norm_pre_clip': 0.03139469400048256, 'timing/data': 0.000304599991068244, 'timing/model': 0.5852985609963071, 'learning_rate/qwen_vl_interface': 8.743711389790321e-06, 'learning_rate/action_model': 8.74371138979032e-05, 'epoch': 0.09}) +[2026-10-03 13:15:59,900][starVLA.training.train_starvla][INFO] - Step 1262, Loss: {'train/loss': 0.4423699676990509, 'train/grad_norm_pre_clip': 43.11650466918945, 'timing/data': 0.0003076399880228564, 'timing/model': 0.39374033399508335, 'learning_rate/qwen_vl_interface': 8.741647383670839e-06, 'learning_rate/action_model': 8.741647383670839e-05, 'epoch': 0.09}) +[2026-10-03 13:16:00,604][starVLA.training.train_starvla][INFO] - Step 1263, Loss: {'train/loss': 0.17734920978546143, 'train/grad_norm_pre_clip': 62.52064514160156, 'timing/data': 0.0002674100105650723, 'timing/model': 0.39352215301187243, 'learning_rate/qwen_vl_interface': 8.739581942267011e-06, 'learning_rate/action_model': 8.739581942267011e-05, 'epoch': 0.09}) +[2026-10-03 13:16:01,308][starVLA.training.train_starvla][INFO] - Step 1264, Loss: {'train/loss': 0.0028969519771635532, 'train/grad_norm_pre_clip': 1.9472862482070923, 'timing/data': 0.00026326000806875527, 'timing/model': 0.3927381230023457, 'learning_rate/qwen_vl_interface': 8.737515066427861e-06, 'learning_rate/action_model': 8.73751506642786e-05, 'epoch': 0.09}) +[2026-10-03 13:16:02,011][starVLA.training.train_starvla][INFO] - Step 1265, Loss: {'train/loss': 0.0008205934427678585, 'train/grad_norm_pre_clip': 0.3220267593860626, 'timing/data': 0.0003013399982592091, 'timing/model': 0.39244989400322083, 'learning_rate/qwen_vl_interface': 8.735446757003e-06, 'learning_rate/action_model': 8.735446757003e-05, 'epoch': 0.09}) +[2026-10-03 13:16:02,716][starVLA.training.train_starvla][INFO] - Step 1266, Loss: {'train/loss': 0.001459346734918654, 'train/grad_norm_pre_clip': 0.9144517779350281, 'timing/data': 0.0003737199876923114, 'timing/model': 0.3935917530034203, 'learning_rate/qwen_vl_interface': 8.733377014842635e-06, 'learning_rate/action_model': 8.733377014842635e-05, 'epoch': 0.09}) +[2026-10-03 13:16:03,609][starVLA.training.train_starvla][INFO] - Step 1267, Loss: {'train/loss': 0.0005766602698713541, 'train/grad_norm_pre_clip': 0.11094032973051071, 'timing/data': 0.0003320599935250357, 'timing/model': 0.3938109549926594, 'learning_rate/qwen_vl_interface': 8.731305840797557e-06, 'learning_rate/action_model': 8.731305840797558e-05, 'epoch': 0.09}) +[2026-10-03 13:16:04,313][starVLA.training.train_starvla][INFO] - Step 1268, Loss: {'train/loss': 0.5253961682319641, 'train/grad_norm_pre_clip': 21.778074264526367, 'timing/data': 0.00021178000315558165, 'timing/model': 0.3932344540080521, 'learning_rate/qwen_vl_interface': 8.72923323571915e-06, 'learning_rate/action_model': 8.72923323571915e-05, 'epoch': 0.09}) +[2026-10-03 13:16:05,016][starVLA.training.train_starvla][INFO] - Step 1269, Loss: {'train/loss': 0.0022152583114802837, 'train/grad_norm_pre_clip': 2.195866823196411, 'timing/data': 0.00021102000027894974, 'timing/model': 0.3925475540017942, 'learning_rate/qwen_vl_interface': 8.727159200459381e-06, 'learning_rate/action_model': 8.727159200459381e-05, 'epoch': 0.09}) +[2026-10-03 13:16:05,720][starVLA.training.train_starvla][INFO] - Step 1270, Loss: {'train/loss': 0.32788121700286865, 'train/grad_norm_pre_clip': 68.31491088867188, 'timing/data': 0.00021335000928957015, 'timing/model': 0.39293231299961917, 'learning_rate/qwen_vl_interface': 8.725083735870807e-06, 'learning_rate/action_model': 8.725083735870807e-05, 'epoch': 0.09}) +[2026-10-03 13:16:06,424][starVLA.training.train_starvla][INFO] - Step 1271, Loss: {'train/loss': 0.009916584938764572, 'train/grad_norm_pre_clip': 13.84009075164795, 'timing/data': 0.00021209000260569155, 'timing/model': 0.39381870299985167, 'learning_rate/qwen_vl_interface': 8.723006842806575e-06, 'learning_rate/action_model': 8.723006842806576e-05, 'epoch': 0.09}) +[2026-10-03 13:16:07,326][starVLA.training.train_starvla][INFO] - Step 1272, Loss: {'train/loss': 0.4107261598110199, 'train/grad_norm_pre_clip': 115.3294448852539, 'timing/data': 0.00034443900221958756, 'timing/model': 0.3925568340055179, 'learning_rate/qwen_vl_interface': 8.720928522120419e-06, 'learning_rate/action_model': 8.720928522120418e-05, 'epoch': 0.09}) +[2026-10-03 13:16:08,030][starVLA.training.train_starvla][INFO] - Step 1273, Loss: {'train/loss': 0.0011643588077276945, 'train/grad_norm_pre_clip': 0.23404523730278015, 'timing/data': 0.0002129800122929737, 'timing/model': 0.39252500400471035, 'learning_rate/qwen_vl_interface': 8.718848774666655e-06, 'learning_rate/action_model': 8.718848774666655e-05, 'epoch': 0.09}) +[2026-10-03 13:16:08,734][starVLA.training.train_starvla][INFO] - Step 1274, Loss: {'train/loss': 0.02887067198753357, 'train/grad_norm_pre_clip': 46.75885772705078, 'timing/data': 0.00032376000308431685, 'timing/model': 0.39313007399323396, 'learning_rate/qwen_vl_interface': 8.716767601300189e-06, 'learning_rate/action_model': 8.716767601300188e-05, 'epoch': 0.09}) +[2026-10-03 13:16:09,436][starVLA.training.train_starvla][INFO] - Step 1275, Loss: {'train/loss': 0.001878336421214044, 'train/grad_norm_pre_clip': 0.2979506552219391, 'timing/data': 0.00033151000388897955, 'timing/model': 0.3922828749928158, 'learning_rate/qwen_vl_interface': 8.714685002876514e-06, 'learning_rate/action_model': 8.714685002876513e-05, 'epoch': 0.09}) +[2026-10-03 13:16:10,140][starVLA.training.train_starvla][INFO] - Step 1276, Loss: {'train/loss': 0.48216503858566284, 'train/grad_norm_pre_clip': 30.597007751464844, 'timing/data': 0.0002845900016836822, 'timing/model': 0.3930656050069956, 'learning_rate/qwen_vl_interface': 8.712600980251706e-06, 'learning_rate/action_model': 8.712600980251705e-05, 'epoch': 0.09}) +[2026-10-03 13:16:11,032][starVLA.training.train_starvla][INFO] - Step 1277, Loss: {'train/loss': 0.32594192028045654, 'train/grad_norm_pre_clip': 19.969268798828125, 'timing/data': 0.00026498000079300255, 'timing/model': 0.5824773509957595, 'learning_rate/qwen_vl_interface': 8.710515534282428e-06, 'learning_rate/action_model': 8.710515534282428e-05, 'epoch': 0.09}) +[2026-10-03 13:16:11,733][starVLA.training.train_starvla][INFO] - Step 1278, Loss: {'train/loss': 0.001030593877658248, 'train/grad_norm_pre_clip': 1.434469223022461, 'timing/data': 0.0002475700021022931, 'timing/model': 0.3909053529932862, 'learning_rate/qwen_vl_interface': 8.708428665825931e-06, 'learning_rate/action_model': 8.708428665825931e-05, 'epoch': 0.09}) +[2026-10-03 13:16:12,436][starVLA.training.train_starvla][INFO] - Step 1279, Loss: {'train/loss': 0.3957192301750183, 'train/grad_norm_pre_clip': 18.78542137145996, 'timing/data': 0.0002632200048537925, 'timing/model': 0.39229946400155313, 'learning_rate/qwen_vl_interface': 8.706340375740045e-06, 'learning_rate/action_model': 8.706340375740045e-05, 'epoch': 0.09}) +[2026-10-03 13:16:13,139][starVLA.training.train_starvla][INFO] - Step 1280, Loss: {'train/loss': 0.3554474711418152, 'train/grad_norm_pre_clip': 25.652280807495117, 'timing/data': 0.00026725999487098306, 'timing/model': 0.39228461400489323, 'learning_rate/qwen_vl_interface': 8.704250664883188e-06, 'learning_rate/action_model': 8.704250664883188e-05, 'epoch': 0.09}) +[2026-10-03 13:16:13,840][starVLA.training.train_starvla][INFO] - Step 1281, Loss: {'train/loss': 0.001946641132235527, 'train/grad_norm_pre_clip': 0.6222652196884155, 'timing/data': 0.00025383899628650397, 'timing/model': 0.39191323399427347, 'learning_rate/qwen_vl_interface': 8.70215953411436e-06, 'learning_rate/action_model': 8.70215953411436e-05, 'epoch': 0.09}) +[2026-10-03 13:16:14,735][starVLA.training.train_starvla][INFO] - Step 1282, Loss: {'train/loss': 0.40326637029647827, 'train/grad_norm_pre_clip': 14.114612579345703, 'timing/data': 0.00021068999194540083, 'timing/model': 0.585269030998461, 'learning_rate/qwen_vl_interface': 8.70006698429315e-06, 'learning_rate/action_model': 8.700066984293149e-05, 'epoch': 0.09}) +[2026-10-03 13:16:15,438][starVLA.training.train_starvla][INFO] - Step 1283, Loss: {'train/loss': 0.0006481001037172973, 'train/grad_norm_pre_clip': 0.16472026705741882, 'timing/data': 0.00024697999469935894, 'timing/model': 0.3925607729906915, 'learning_rate/qwen_vl_interface': 8.697973016279722e-06, 'learning_rate/action_model': 8.697973016279722e-05, 'epoch': 0.09}) +[2026-10-03 13:16:16,141][starVLA.training.train_starvla][INFO] - Step 1284, Loss: {'train/loss': 0.0023565415758639574, 'train/grad_norm_pre_clip': 2.021928548812866, 'timing/data': 0.00046552999992854893, 'timing/model': 0.39230192301329225, 'learning_rate/qwen_vl_interface': 8.695877630934827e-06, 'learning_rate/action_model': 8.695877630934826e-05, 'epoch': 0.09}) +[2026-10-03 13:16:16,842][starVLA.training.train_starvla][INFO] - Step 1285, Loss: {'train/loss': 0.0004068751004524529, 'train/grad_norm_pre_clip': 0.1302584856748581, 'timing/data': 0.0002595200057839975, 'timing/model': 0.39168783399509266, 'learning_rate/qwen_vl_interface': 8.693780829119801e-06, 'learning_rate/action_model': 8.693780829119801e-05, 'epoch': 0.09}) +[2026-10-03 13:16:17,545][starVLA.training.train_starvla][INFO] - Step 1286, Loss: {'train/loss': 0.0037084612995386124, 'train/grad_norm_pre_clip': 1.4618844985961914, 'timing/data': 0.0002047199959633872, 'timing/model': 0.39180866499373224, 'learning_rate/qwen_vl_interface': 8.69168261169656e-06, 'learning_rate/action_model': 8.691682611696559e-05, 'epoch': 0.09}) +[2026-10-03 13:16:18,439][starVLA.training.train_starvla][INFO] - Step 1287, Loss: {'train/loss': 0.609808623790741, 'train/grad_norm_pre_clip': 36.77290725708008, 'timing/data': 0.0002078600082313642, 'timing/model': 0.5851210510008968, 'learning_rate/qwen_vl_interface': 8.689582979527599e-06, 'learning_rate/action_model': 8.689582979527599e-05, 'epoch': 0.09}) +[2026-10-03 13:16:19,142][starVLA.training.train_starvla][INFO] - Step 1288, Loss: {'train/loss': 0.0011042087571695447, 'train/grad_norm_pre_clip': 0.9168655872344971, 'timing/data': 0.0003576099989004433, 'timing/model': 0.3922668939922005, 'learning_rate/qwen_vl_interface': 8.687481933476e-06, 'learning_rate/action_model': 8.687481933476e-05, 'epoch': 0.09}) +[2026-10-03 13:16:19,843][starVLA.training.train_starvla][INFO] - Step 1289, Loss: {'train/loss': 0.16718870401382446, 'train/grad_norm_pre_clip': 121.38794708251953, 'timing/data': 0.00022602999524679035, 'timing/model': 0.3917464939877391, 'learning_rate/qwen_vl_interface': 8.685379474405423e-06, 'learning_rate/action_model': 8.685379474405423e-05, 'epoch': 0.09}) +[2026-10-03 13:16:20,546][starVLA.training.train_starvla][INFO] - Step 1290, Loss: {'train/loss': 0.00012785880244337022, 'train/grad_norm_pre_clip': 0.045859333127737045, 'timing/data': 0.0002208700025221333, 'timing/model': 0.39245003399264533, 'learning_rate/qwen_vl_interface': 8.68327560318011e-06, 'learning_rate/action_model': 8.68327560318011e-05, 'epoch': 0.09}) +[2026-10-03 13:16:21,246][starVLA.training.train_starvla][INFO] - Step 1291, Loss: {'train/loss': 0.37458837032318115, 'train/grad_norm_pre_clip': 85.26292419433594, 'timing/data': 0.00021088001085445285, 'timing/model': 0.3919251740007894, 'learning_rate/qwen_vl_interface': 8.681170320664883e-06, 'learning_rate/action_model': 8.681170320664883e-05, 'epoch': 0.09}) +[2026-10-03 13:16:21,949][starVLA.training.train_starvla][INFO] - Step 1292, Loss: {'train/loss': 0.07690384238958359, 'train/grad_norm_pre_clip': 61.87124252319336, 'timing/data': 0.00026775000151246786, 'timing/model': 0.3919628439907683, 'learning_rate/qwen_vl_interface': 8.679063627725143e-06, 'learning_rate/action_model': 8.679063627725144e-05, 'epoch': 0.09}) +[2026-10-03 13:16:22,847][starVLA.training.train_starvla][INFO] - Step 1293, Loss: {'train/loss': 0.005233958829194307, 'train/grad_norm_pre_clip': 2.188636541366577, 'timing/data': 0.00032612000359222293, 'timing/model': 0.39177352300612256, 'learning_rate/qwen_vl_interface': 8.676955525226876e-06, 'learning_rate/action_model': 8.676955525226875e-05, 'epoch': 0.09}) +[2026-10-03 13:16:23,549][starVLA.training.train_starvla][INFO] - Step 1294, Loss: {'train/loss': 0.06926341354846954, 'train/grad_norm_pre_clip': 12.240887641906738, 'timing/data': 0.0003055600100196898, 'timing/model': 0.39177220400597434, 'learning_rate/qwen_vl_interface': 8.674846014036636e-06, 'learning_rate/action_model': 8.674846014036636e-05, 'epoch': 0.09}) +[2026-10-03 13:16:24,252][starVLA.training.train_starvla][INFO] - Step 1295, Loss: {'train/loss': 0.0036501085851341486, 'train/grad_norm_pre_clip': 0.8733525276184082, 'timing/data': 0.0003898800059687346, 'timing/model': 0.39230950399360154, 'learning_rate/qwen_vl_interface': 8.672735095021569e-06, 'learning_rate/action_model': 8.672735095021569e-05, 'epoch': 0.09}) +[2026-10-03 13:16:24,957][starVLA.training.train_starvla][INFO] - Step 1296, Loss: {'train/loss': 0.35313329100608826, 'train/grad_norm_pre_clip': 197.02755737304688, 'timing/data': 0.0003657100023701787, 'timing/model': 0.39407545400899835, 'learning_rate/qwen_vl_interface': 8.670622769049393e-06, 'learning_rate/action_model': 8.670622769049392e-05, 'epoch': 0.09}) +[2026-10-03 13:16:25,659][starVLA.training.train_starvla][INFO] - Step 1297, Loss: {'train/loss': 0.002881102729588747, 'train/grad_norm_pre_clip': 3.4682559967041016, 'timing/data': 0.00021166000806260854, 'timing/model': 0.39220438500342425, 'learning_rate/qwen_vl_interface': 8.668509036988406e-06, 'learning_rate/action_model': 8.668509036988405e-05, 'epoch': 0.09}) +[2026-10-03 13:16:26,556][starVLA.training.train_starvla][INFO] - Step 1298, Loss: {'train/loss': 0.2687471807003021, 'train/grad_norm_pre_clip': 58.51926040649414, 'timing/data': 0.00022522998915519565, 'timing/model': 0.5860868210002081, 'learning_rate/qwen_vl_interface': 8.66639389970748e-06, 'learning_rate/action_model': 8.666393899707479e-05, 'epoch': 0.1}) +[2026-10-03 13:16:27,257][starVLA.training.train_starvla][INFO] - Step 1299, Loss: {'train/loss': 0.41540321707725525, 'train/grad_norm_pre_clip': 53.65528869628906, 'timing/data': 0.00036869999894406646, 'timing/model': 0.39215331300511025, 'learning_rate/qwen_vl_interface': 8.66427735807607e-06, 'learning_rate/action_model': 8.664277358076069e-05, 'epoch': 0.1}) +[2026-10-03 13:16:27,961][starVLA.training.train_starvla][INFO] - Step 1300, Loss: {'train/loss': 0.18040604889392853, 'train/grad_norm_pre_clip': 105.10444641113281, 'timing/data': 0.0003373100043972954, 'timing/model': 0.3921433240029728, 'learning_rate/qwen_vl_interface': 8.662159412964208e-06, 'learning_rate/action_model': 8.662159412964207e-05, 'epoch': 0.1}) +[2026-10-03 13:16:28,661][starVLA.training.train_starvla][INFO] - Step 1301, Loss: {'train/loss': 0.026289325207471848, 'train/grad_norm_pre_clip': 9.953083992004395, 'timing/data': 0.0002169399958802387, 'timing/model': 0.3912808040040545, 'learning_rate/qwen_vl_interface': 8.660040065242499e-06, 'learning_rate/action_model': 8.660040065242498e-05, 'epoch': 0.1}) +[2026-10-03 13:16:29,363][starVLA.training.train_starvla][INFO] - Step 1302, Loss: {'train/loss': 0.06260412186384201, 'train/grad_norm_pre_clip': 66.81812286376953, 'timing/data': 0.0003357899986440316, 'timing/model': 0.39189227399765514, 'learning_rate/qwen_vl_interface': 8.657919315782127e-06, 'learning_rate/action_model': 8.657919315782127e-05, 'epoch': 0.1}) +[2026-10-03 13:16:30,064][starVLA.training.train_starvla][INFO] - Step 1303, Loss: {'train/loss': 0.006788972765207291, 'train/grad_norm_pre_clip': 4.081323146820068, 'timing/data': 0.00031676000799052417, 'timing/model': 0.3921536130073946, 'learning_rate/qwen_vl_interface': 8.655797165454853e-06, 'learning_rate/action_model': 8.655797165454853e-05, 'epoch': 0.1}) +[2026-10-03 13:16:30,959][starVLA.training.train_starvla][INFO] - Step 1304, Loss: {'train/loss': 0.003178776940330863, 'train/grad_norm_pre_clip': 2.338045597076416, 'timing/data': 0.00037874998815823346, 'timing/model': 0.3929540730023291, 'learning_rate/qwen_vl_interface': 8.653673615133013e-06, 'learning_rate/action_model': 8.653673615133013e-05, 'epoch': 0.1}) +[2026-10-03 13:16:31,661][starVLA.training.train_starvla][INFO] - Step 1305, Loss: {'train/loss': 0.2586589455604553, 'train/grad_norm_pre_clip': 74.55522918701172, 'timing/data': 0.0003484799963189289, 'timing/model': 0.3913185429992154, 'learning_rate/qwen_vl_interface': 8.651548665689519e-06, 'learning_rate/action_model': 8.651548665689518e-05, 'epoch': 0.1}) +[2026-10-03 13:16:32,364][starVLA.training.train_starvla][INFO] - Step 1306, Loss: {'train/loss': 0.595967173576355, 'train/grad_norm_pre_clip': 42.423946380615234, 'timing/data': 0.000251450008363463, 'timing/model': 0.3915545740019297, 'learning_rate/qwen_vl_interface': 8.649422317997855e-06, 'learning_rate/action_model': 8.649422317997856e-05, 'epoch': 0.1}) +[2026-10-03 13:16:33,065][starVLA.training.train_starvla][INFO] - Step 1307, Loss: {'train/loss': 0.5779249668121338, 'train/grad_norm_pre_clip': 274.20465087890625, 'timing/data': 0.00031709000177215785, 'timing/model': 0.3915119239973137, 'learning_rate/qwen_vl_interface': 8.647294572932082e-06, 'learning_rate/action_model': 8.647294572932082e-05, 'epoch': 0.1}) +[2026-10-03 13:16:33,768][starVLA.training.train_starvla][INFO] - Step 1308, Loss: {'train/loss': 0.0032788310199975967, 'train/grad_norm_pre_clip': 1.7629075050354004, 'timing/data': 0.0002136999974027276, 'timing/model': 0.39249205398664344, 'learning_rate/qwen_vl_interface': 8.64516543136684e-06, 'learning_rate/action_model': 8.64516543136684e-05, 'epoch': 0.1}) +[2026-10-03 13:16:34,664][starVLA.training.train_starvla][INFO] - Step 1309, Loss: {'train/loss': 0.22082944214344025, 'train/grad_norm_pre_clip': 239.0540008544922, 'timing/data': 0.0003284199920017272, 'timing/model': 0.39212330398731865, 'learning_rate/qwen_vl_interface': 8.643034894177335e-06, 'learning_rate/action_model': 8.643034894177335e-05, 'epoch': 0.1}) +[2026-10-03 13:16:35,365][starVLA.training.train_starvla][INFO] - Step 1310, Loss: {'train/loss': 0.00135463394690305, 'train/grad_norm_pre_clip': 0.6399286985397339, 'timing/data': 0.0003476900019450113, 'timing/model': 0.39167248499870766, 'learning_rate/qwen_vl_interface': 8.64090296223935e-06, 'learning_rate/action_model': 8.64090296223935e-05, 'epoch': 0.1}) +[2026-10-03 13:16:36,067][starVLA.training.train_starvla][INFO] - Step 1311, Loss: {'train/loss': 0.00026007206179201603, 'train/grad_norm_pre_clip': 0.06587563455104828, 'timing/data': 0.0003410200006328523, 'timing/model': 0.39192454400472343, 'learning_rate/qwen_vl_interface': 8.638769636429245e-06, 'learning_rate/action_model': 8.638769636429245e-05, 'epoch': 0.1}) +[2026-10-03 13:16:36,771][starVLA.training.train_starvla][INFO] - Step 1312, Loss: {'train/loss': 0.15425032377243042, 'train/grad_norm_pre_clip': 65.72464752197266, 'timing/data': 0.0003650500002549961, 'timing/model': 0.39223893400048837, 'learning_rate/qwen_vl_interface': 8.636634917623946e-06, 'learning_rate/action_model': 8.636634917623944e-05, 'epoch': 0.1}) +[2026-10-03 13:16:37,473][starVLA.training.train_starvla][INFO] - Step 1313, Loss: {'train/loss': 0.009394564665853977, 'train/grad_norm_pre_clip': 10.837850570678711, 'timing/data': 0.00034204999974463135, 'timing/model': 0.39153131398779806, 'learning_rate/qwen_vl_interface': 8.634498806700955e-06, 'learning_rate/action_model': 8.634498806700954e-05, 'epoch': 0.1}) +[2026-10-03 13:16:38,366][starVLA.training.train_starvla][INFO] - Step 1314, Loss: {'train/loss': 0.84101402759552, 'train/grad_norm_pre_clip': 84.57544708251953, 'timing/data': 0.00032387999817728996, 'timing/model': 0.5821739219973097, 'learning_rate/qwen_vl_interface': 8.63236130453835e-06, 'learning_rate/action_model': 8.63236130453835e-05, 'epoch': 0.1}) +[2026-10-03 13:16:39,067][starVLA.training.train_starvla][INFO] - Step 1315, Loss: {'train/loss': 6.575982843060046e-05, 'train/grad_norm_pre_clip': 0.02060365118086338, 'timing/data': 0.0002478300011716783, 'timing/model': 0.39192825500504114, 'learning_rate/qwen_vl_interface': 8.630222412014777e-06, 'learning_rate/action_model': 8.630222412014776e-05, 'epoch': 0.1}) +[2026-10-03 13:16:39,771][starVLA.training.train_starvla][INFO] - Step 1316, Loss: {'train/loss': 2.0368912373669446e-05, 'train/grad_norm_pre_clip': 0.011120932176709175, 'timing/data': 0.0003181899955961853, 'timing/model': 0.392871884003398, 'learning_rate/qwen_vl_interface': 8.628082130009451e-06, 'learning_rate/action_model': 8.628082130009451e-05, 'epoch': 0.1}) +[2026-10-03 13:16:40,473][starVLA.training.train_starvla][INFO] - Step 1317, Loss: {'train/loss': 0.11566253751516342, 'train/grad_norm_pre_clip': 99.78955078125, 'timing/data': 0.0003021489974344149, 'timing/model': 0.3917367039975943, 'learning_rate/qwen_vl_interface': 8.62594045940216e-06, 'learning_rate/action_model': 8.62594045940216e-05, 'epoch': 0.1}) +[2026-10-03 13:16:41,175][starVLA.training.train_starvla][INFO] - Step 1318, Loss: {'train/loss': 0.20470307767391205, 'train/grad_norm_pre_clip': 154.5815887451172, 'timing/data': 0.0002521900023566559, 'timing/model': 0.3921897239924874, 'learning_rate/qwen_vl_interface': 8.623797401073272e-06, 'learning_rate/action_model': 8.623797401073271e-05, 'epoch': 0.1}) +[2026-10-03 13:16:41,877][starVLA.training.train_starvla][INFO] - Step 1319, Loss: {'train/loss': 0.00011668886145344004, 'train/grad_norm_pre_clip': 0.26431015133857727, 'timing/data': 0.00025243000709451735, 'timing/model': 0.39202152300276794, 'learning_rate/qwen_vl_interface': 8.62165295590371e-06, 'learning_rate/action_model': 8.621652955903709e-05, 'epoch': 0.1}) +[2026-10-03 13:16:42,769][starVLA.training.train_starvla][INFO] - Step 1320, Loss: {'train/loss': 0.002785828895866871, 'train/grad_norm_pre_clip': 0.7742308378219604, 'timing/data': 0.0002564000023994595, 'timing/model': 0.39197424399026204, 'learning_rate/qwen_vl_interface': 8.619507124774975e-06, 'learning_rate/action_model': 8.619507124774975e-05, 'epoch': 0.1}) +[2026-10-03 13:16:43,471][starVLA.training.train_starvla][INFO] - Step 1321, Loss: {'train/loss': 0.34903767704963684, 'train/grad_norm_pre_clip': 161.5297088623047, 'timing/data': 0.00027812999906018376, 'timing/model': 0.39170660299714655, 'learning_rate/qwen_vl_interface': 8.61735990856914e-06, 'learning_rate/action_model': 8.617359908569139e-05, 'epoch': 0.1}) +[2026-10-03 13:16:44,172][starVLA.training.train_starvla][INFO] - Step 1322, Loss: {'train/loss': 4.955434633302502e-05, 'train/grad_norm_pre_clip': 0.025560587644577026, 'timing/data': 0.0002102699945680797, 'timing/model': 0.3920095039939042, 'learning_rate/qwen_vl_interface': 8.615211308168842e-06, 'learning_rate/action_model': 8.61521130816884e-05, 'epoch': 0.1}) +[2026-10-03 13:16:44,874][starVLA.training.train_starvla][INFO] - Step 1323, Loss: {'train/loss': 0.00764109194278717, 'train/grad_norm_pre_clip': 8.477151870727539, 'timing/data': 0.00020176998805254698, 'timing/model': 0.392088294000132, 'learning_rate/qwen_vl_interface': 8.613061324457289e-06, 'learning_rate/action_model': 8.613061324457288e-05, 'epoch': 0.1}) +[2026-10-03 13:16:45,577][starVLA.training.train_starvla][INFO] - Step 1324, Loss: {'train/loss': 0.5644601583480835, 'train/grad_norm_pre_clip': 20.846351623535156, 'timing/data': 0.0003219700010959059, 'timing/model': 0.392266605005716, 'learning_rate/qwen_vl_interface': 8.610909958318262e-06, 'learning_rate/action_model': 8.610909958318261e-05, 'epoch': 0.1}) +[2026-10-03 13:16:46,471][starVLA.training.train_starvla][INFO] - Step 1325, Loss: {'train/loss': 0.45415782928466797, 'train/grad_norm_pre_clip': 32.255218505859375, 'timing/data': 0.00032345000363420695, 'timing/model': 0.39256133399612736, 'learning_rate/qwen_vl_interface': 8.608757210636101e-06, 'learning_rate/action_model': 8.608757210636101e-05, 'epoch': 0.1}) +[2026-10-03 13:16:47,174][starVLA.training.train_starvla][INFO] - Step 1326, Loss: {'train/loss': 0.5010493397712708, 'train/grad_norm_pre_clip': 92.10652160644531, 'timing/data': 0.0002538900007493794, 'timing/model': 0.392297113998211, 'learning_rate/qwen_vl_interface': 8.606603082295724e-06, 'learning_rate/action_model': 8.606603082295724e-05, 'epoch': 0.1}) +[2026-10-03 13:16:47,875][starVLA.training.train_starvla][INFO] - Step 1327, Loss: {'train/loss': 0.0004736444097943604, 'train/grad_norm_pre_clip': 0.07013814151287079, 'timing/data': 0.00031180999940261245, 'timing/model': 0.3917976539960364, 'learning_rate/qwen_vl_interface': 8.60444757418261e-06, 'learning_rate/action_model': 8.604447574182609e-05, 'epoch': 0.1}) +[2026-10-03 13:16:48,579][starVLA.training.train_starvla][INFO] - Step 1328, Loss: {'train/loss': 5.766646154370392e-06, 'train/grad_norm_pre_clip': 0.0016143099637702107, 'timing/data': 0.000288759998511523, 'timing/model': 0.39269171400519554, 'learning_rate/qwen_vl_interface': 8.602290687182805e-06, 'learning_rate/action_model': 8.602290687182806e-05, 'epoch': 0.1}) +[2026-10-03 13:16:49,281][starVLA.training.train_starvla][INFO] - Step 1329, Loss: {'train/loss': 7.103555981302634e-05, 'train/grad_norm_pre_clip': 0.050008371472358704, 'timing/data': 0.0002668800007086247, 'timing/model': 0.39169037400279194, 'learning_rate/qwen_vl_interface': 8.600132422182929e-06, 'learning_rate/action_model': 8.600132422182928e-05, 'epoch': 0.1}) +[2026-10-03 13:16:50,176][starVLA.training.train_starvla][INFO] - Step 1330, Loss: {'train/loss': 0.36553090810775757, 'train/grad_norm_pre_clip': 45.728084564208984, 'timing/data': 0.00039973999082576483, 'timing/model': 0.39247899400652386, 'learning_rate/qwen_vl_interface': 8.597972780070156e-06, 'learning_rate/action_model': 8.597972780070155e-05, 'epoch': 0.1}) +[2026-10-03 13:16:50,877][starVLA.training.train_starvla][INFO] - Step 1331, Loss: {'train/loss': 9.544444037601352e-05, 'train/grad_norm_pre_clip': 0.05340724065899849, 'timing/data': 0.0003486700006760657, 'timing/model': 0.39172227299422957, 'learning_rate/qwen_vl_interface': 8.595811761732239e-06, 'learning_rate/action_model': 8.595811761732238e-05, 'epoch': 0.1}) +[2026-10-03 13:16:51,582][starVLA.training.train_starvla][INFO] - Step 1332, Loss: {'train/loss': 0.0006552701815962791, 'train/grad_norm_pre_clip': 0.1893993616104126, 'timing/data': 0.0002905300061684102, 'timing/model': 0.39347760399687104, 'learning_rate/qwen_vl_interface': 8.593649368057489e-06, 'learning_rate/action_model': 8.59364936805749e-05, 'epoch': 0.1}) +[2026-10-03 13:16:52,283][starVLA.training.train_starvla][INFO] - Step 1333, Loss: {'train/loss': 0.00030933727975934744, 'train/grad_norm_pre_clip': 0.07137920707464218, 'timing/data': 0.00026603999140206724, 'timing/model': 0.39171826299570967, 'learning_rate/qwen_vl_interface': 8.591485599934786e-06, 'learning_rate/action_model': 8.591485599934785e-05, 'epoch': 0.1}) +[2026-10-03 13:16:52,985][starVLA.training.train_starvla][INFO] - Step 1334, Loss: {'train/loss': 0.00026784063084051013, 'train/grad_norm_pre_clip': 0.16814479231834412, 'timing/data': 0.00034396001137793064, 'timing/model': 0.3922492540004896, 'learning_rate/qwen_vl_interface': 8.58932045825357e-06, 'learning_rate/action_model': 8.589320458253569e-05, 'epoch': 0.1}) +[2026-10-03 13:16:53,880][starVLA.training.train_starvla][INFO] - Step 1335, Loss: {'train/loss': 8.721624908503145e-05, 'train/grad_norm_pre_clip': 0.03324420750141144, 'timing/data': 0.0003014800022356212, 'timing/model': 0.5844972610066179, 'learning_rate/qwen_vl_interface': 8.58715394390385e-06, 'learning_rate/action_model': 8.58715394390385e-05, 'epoch': 0.1}) +[2026-10-03 13:16:54,581][starVLA.training.train_starvla][INFO] - Step 1336, Loss: {'train/loss': 0.000491759623400867, 'train/grad_norm_pre_clip': 0.2950231432914734, 'timing/data': 0.00024028999905567616, 'timing/model': 0.39111802401021123, 'learning_rate/qwen_vl_interface': 8.5849860577762e-06, 'learning_rate/action_model': 8.5849860577762e-05, 'epoch': 0.1}) +[2026-10-03 13:16:55,282][starVLA.training.train_starvla][INFO] - Step 1337, Loss: {'train/loss': 0.001066702650859952, 'train/grad_norm_pre_clip': 0.3171515166759491, 'timing/data': 0.0002443200064590201, 'timing/model': 0.3916045440128073, 'learning_rate/qwen_vl_interface': 8.582816800761756e-06, 'learning_rate/action_model': 8.582816800761755e-05, 'epoch': 0.1}) +[2026-10-03 13:16:55,985][starVLA.training.train_starvla][INFO] - Step 1338, Loss: {'train/loss': 1.0107733011245728, 'train/grad_norm_pre_clip': 118.97986602783203, 'timing/data': 0.00025535999157000333, 'timing/model': 0.392375023002387, 'learning_rate/qwen_vl_interface': 8.580646173752214e-06, 'learning_rate/action_model': 8.580646173752213e-05, 'epoch': 0.1}) +[2026-10-03 13:16:56,686][starVLA.training.train_starvla][INFO] - Step 1339, Loss: {'train/loss': 0.0004916326142847538, 'train/grad_norm_pre_clip': 0.09232667833566666, 'timing/data': 0.0003245799889555201, 'timing/model': 0.391338283996447, 'learning_rate/qwen_vl_interface': 8.578474177639841e-06, 'learning_rate/action_model': 8.578474177639842e-05, 'epoch': 0.1}) +[2026-10-03 13:16:57,583][starVLA.training.train_starvla][INFO] - Step 1340, Loss: {'train/loss': 0.00035539886448532343, 'train/grad_norm_pre_clip': 0.1565750688314438, 'timing/data': 0.00031324000156018883, 'timing/model': 0.5859712509991368, 'learning_rate/qwen_vl_interface': 8.576300813317463e-06, 'learning_rate/action_model': 8.576300813317462e-05, 'epoch': 0.1}) +[2026-10-03 13:16:58,284][starVLA.training.train_starvla][INFO] - Step 1341, Loss: {'train/loss': 0.0005357047193683684, 'train/grad_norm_pre_clip': 0.17355522513389587, 'timing/data': 0.0002963500010082498, 'timing/model': 0.3913707540050382, 'learning_rate/qwen_vl_interface': 8.574126081678463e-06, 'learning_rate/action_model': 8.574126081678462e-05, 'epoch': 0.1}) +[2026-10-03 13:16:58,986][starVLA.training.train_starvla][INFO] - Step 1342, Loss: {'train/loss': 0.001661019749008119, 'train/grad_norm_pre_clip': 2.1857950687408447, 'timing/data': 0.00020326000230852515, 'timing/model': 0.39199259399902076, 'learning_rate/qwen_vl_interface': 8.571949983616796e-06, 'learning_rate/action_model': 8.571949983616795e-05, 'epoch': 0.1}) +[2026-10-03 13:16:59,687][starVLA.training.train_starvla][INFO] - Step 1343, Loss: {'train/loss': 0.14586946368217468, 'train/grad_norm_pre_clip': 62.277671813964844, 'timing/data': 0.0002642499894136563, 'timing/model': 0.3919654939963948, 'learning_rate/qwen_vl_interface': 8.569772520026971e-06, 'learning_rate/action_model': 8.569772520026971e-05, 'epoch': 0.1}) +[2026-10-03 13:17:00,390][starVLA.training.train_starvla][INFO] - Step 1344, Loss: {'train/loss': 0.13809579610824585, 'train/grad_norm_pre_clip': 71.19452667236328, 'timing/data': 0.0003446299961069599, 'timing/model': 0.3925227139989147, 'learning_rate/qwen_vl_interface': 8.567593691804061e-06, 'learning_rate/action_model': 8.567593691804061e-05, 'epoch': 0.1}) +[2026-10-03 13:17:01,090][starVLA.training.train_starvla][INFO] - Step 1345, Loss: {'train/loss': 0.4899899959564209, 'train/grad_norm_pre_clip': 21.940940856933594, 'timing/data': 0.0002805199910653755, 'timing/model': 0.3903606740059331, 'learning_rate/qwen_vl_interface': 8.565413499843703e-06, 'learning_rate/action_model': 8.565413499843703e-05, 'epoch': 0.1}) +[2026-10-03 13:17:01,985][starVLA.training.train_starvla][INFO] - Step 1346, Loss: {'train/loss': 0.16949336230754852, 'train/grad_norm_pre_clip': 10.738825798034668, 'timing/data': 0.0003843299928121269, 'timing/model': 0.391927435004618, 'learning_rate/qwen_vl_interface': 8.563231945042087e-06, 'learning_rate/action_model': 8.563231945042088e-05, 'epoch': 0.1}) +[2026-10-03 13:17:02,686][starVLA.training.train_starvla][INFO] - Step 1347, Loss: {'train/loss': 0.011722286231815815, 'train/grad_norm_pre_clip': 37.27244186401367, 'timing/data': 0.00031459999445360154, 'timing/model': 0.3913956739997957, 'learning_rate/qwen_vl_interface': 8.56104902829597e-06, 'learning_rate/action_model': 8.56104902829597e-05, 'epoch': 0.1}) +[2026-10-03 13:17:03,389][starVLA.training.train_starvla][INFO] - Step 1348, Loss: {'train/loss': 6.177707109600306e-05, 'train/grad_norm_pre_clip': 0.028944771736860275, 'timing/data': 0.00031957999453879893, 'timing/model': 0.3919843339972431, 'learning_rate/qwen_vl_interface': 8.558864750502669e-06, 'learning_rate/action_model': 8.558864750502668e-05, 'epoch': 0.1}) +[2026-10-03 13:17:04,090][starVLA.training.train_starvla][INFO] - Step 1349, Loss: {'train/loss': 0.008041349239647388, 'train/grad_norm_pre_clip': 17.506032943725586, 'timing/data': 0.0003116099978797138, 'timing/model': 0.3918150639947271, 'learning_rate/qwen_vl_interface': 8.556679112560055e-06, 'learning_rate/action_model': 8.556679112560054e-05, 'epoch': 0.1}) +[2026-10-03 13:17:04,793][starVLA.training.train_starvla][INFO] - Step 1350, Loss: {'train/loss': 0.08805925399065018, 'train/grad_norm_pre_clip': 37.961299896240234, 'timing/data': 0.0002512990031391382, 'timing/model': 0.39243482500023674, 'learning_rate/qwen_vl_interface': 8.554492115366562e-06, 'learning_rate/action_model': 8.554492115366561e-05, 'epoch': 0.1}) +[2026-10-03 13:17:05,684][starVLA.training.train_starvla][INFO] - Step 1351, Loss: {'train/loss': 0.2520669400691986, 'train/grad_norm_pre_clip': 103.39445495605469, 'timing/data': 0.00020530000620055944, 'timing/model': 0.5821315119974315, 'learning_rate/qwen_vl_interface': 8.55230375982118e-06, 'learning_rate/action_model': 8.55230375982118e-05, 'epoch': 0.1}) +[2026-10-03 13:17:06,387][starVLA.training.train_starvla][INFO] - Step 1352, Loss: {'train/loss': 0.0011812779121100903, 'train/grad_norm_pre_clip': 0.7346799373626709, 'timing/data': 0.0003723999980138615, 'timing/model': 0.3926122039993061, 'learning_rate/qwen_vl_interface': 8.550114046823463e-06, 'learning_rate/action_model': 8.550114046823463e-05, 'epoch': 0.1}) +[2026-10-03 13:17:07,089][starVLA.training.train_starvla][INFO] - Step 1353, Loss: {'train/loss': 0.00010972032032441348, 'train/grad_norm_pre_clip': 0.051321517676115036, 'timing/data': 0.00025908999668899924, 'timing/model': 0.3917581449932186, 'learning_rate/qwen_vl_interface': 8.547922977273518e-06, 'learning_rate/action_model': 8.547922977273518e-05, 'epoch': 0.1}) +[2026-10-03 13:17:07,791][starVLA.training.train_starvla][INFO] - Step 1354, Loss: {'train/loss': 0.001311896601691842, 'train/grad_norm_pre_clip': 0.27635645866394043, 'timing/data': 0.00025437999283894897, 'timing/model': 0.3922139549977146, 'learning_rate/qwen_vl_interface': 8.54573055207201e-06, 'learning_rate/action_model': 8.545730552072008e-05, 'epoch': 0.1}) +[2026-10-03 13:17:08,491][starVLA.training.train_starvla][INFO] - Step 1355, Loss: {'train/loss': 0.002261203480884433, 'train/grad_norm_pre_clip': 0.36997175216674805, 'timing/data': 0.00026073999470099807, 'timing/model': 0.3918596349976724, 'learning_rate/qwen_vl_interface': 8.543536772120161e-06, 'learning_rate/action_model': 8.54353677212016e-05, 'epoch': 0.1}) +[2026-10-03 13:17:09,195][starVLA.training.train_starvla][INFO] - Step 1356, Loss: {'train/loss': 0.21308395266532898, 'train/grad_norm_pre_clip': 124.0122299194336, 'timing/data': 0.0002442699915263802, 'timing/model': 0.3929358540044632, 'learning_rate/qwen_vl_interface': 8.541341638319754e-06, 'learning_rate/action_model': 8.541341638319754e-05, 'epoch': 0.1}) +[2026-10-03 13:17:10,089][starVLA.training.train_starvla][INFO] - Step 1357, Loss: {'train/loss': 0.00024591758847236633, 'train/grad_norm_pre_clip': 0.1442725658416748, 'timing/data': 0.00032901999657042325, 'timing/model': 0.3919667939917417, 'learning_rate/qwen_vl_interface': 8.539145151573126e-06, 'learning_rate/action_model': 8.539145151573125e-05, 'epoch': 0.1}) +[2026-10-03 13:17:10,791][starVLA.training.train_starvla][INFO] - Step 1358, Loss: {'train/loss': 0.010164852254092693, 'train/grad_norm_pre_clip': 1.3386987447738647, 'timing/data': 0.0002560400025686249, 'timing/model': 0.3921965940098744, 'learning_rate/qwen_vl_interface': 8.536947312783167e-06, 'learning_rate/action_model': 8.536947312783166e-05, 'epoch': 0.1}) +[2026-10-03 13:17:11,492][starVLA.training.train_starvla][INFO] - Step 1359, Loss: {'train/loss': 0.00018241655197925866, 'train/grad_norm_pre_clip': 0.09978420287370682, 'timing/data': 0.0003207600093446672, 'timing/model': 0.39117514400277287, 'learning_rate/qwen_vl_interface': 8.534748122853326e-06, 'learning_rate/action_model': 8.534748122853325e-05, 'epoch': 0.1}) +[2026-10-03 13:17:12,195][starVLA.training.train_starvla][INFO] - Step 1360, Loss: {'train/loss': 0.11825630813837051, 'train/grad_norm_pre_clip': 10.665681838989258, 'timing/data': 0.00036375998752191663, 'timing/model': 0.392010413997923, 'learning_rate/qwen_vl_interface': 8.532547582687609e-06, 'learning_rate/action_model': 8.532547582687608e-05, 'epoch': 0.1}) +[2026-10-03 13:17:12,896][starVLA.training.train_starvla][INFO] - Step 1361, Loss: {'train/loss': 0.007262994069606066, 'train/grad_norm_pre_clip': 9.413670539855957, 'timing/data': 0.0003022199962288141, 'timing/model': 0.391453434000141, 'learning_rate/qwen_vl_interface': 8.530345693190573e-06, 'learning_rate/action_model': 8.530345693190572e-05, 'epoch': 0.1}) +[2026-10-03 13:17:13,792][starVLA.training.train_starvla][INFO] - Step 1362, Loss: {'train/loss': 0.00025103252846747637, 'train/grad_norm_pre_clip': 0.04983718693256378, 'timing/data': 0.0003442599991103634, 'timing/model': 0.39241216400114354, 'learning_rate/qwen_vl_interface': 8.528142455267328e-06, 'learning_rate/action_model': 8.528142455267329e-05, 'epoch': 0.1}) +[2026-10-03 13:17:14,493][starVLA.training.train_starvla][INFO] - Step 1363, Loss: {'train/loss': 0.00012827709724660963, 'train/grad_norm_pre_clip': 0.028743352741003036, 'timing/data': 0.0003582799981813878, 'timing/model': 0.3920010539877694, 'learning_rate/qwen_vl_interface': 8.52593786982355e-06, 'learning_rate/action_model': 8.525937869823551e-05, 'epoch': 0.1}) +[2026-10-03 13:17:15,197][starVLA.training.train_starvla][INFO] - Step 1364, Loss: {'train/loss': 0.001824384555220604, 'train/grad_norm_pre_clip': 1.3396782875061035, 'timing/data': 0.0003887000057147816, 'timing/model': 0.39241790400410537, 'learning_rate/qwen_vl_interface': 8.523731937765456e-06, 'learning_rate/action_model': 8.523731937765456e-05, 'epoch': 0.1}) +[2026-10-03 13:17:15,898][starVLA.training.train_starvla][INFO] - Step 1365, Loss: {'train/loss': 0.00010845654469449073, 'train/grad_norm_pre_clip': 0.028938082978129387, 'timing/data': 0.00033758000063244253, 'timing/model': 0.39147684500494506, 'learning_rate/qwen_vl_interface': 8.521524659999823e-06, 'learning_rate/action_model': 8.521524659999822e-05, 'epoch': 0.1}) +[2026-10-03 13:17:16,601][starVLA.training.train_starvla][INFO] - Step 1366, Loss: {'train/loss': 0.030721358954906464, 'train/grad_norm_pre_clip': 70.56792449951172, 'timing/data': 0.0003473499964457005, 'timing/model': 0.3925422540050931, 'learning_rate/qwen_vl_interface': 8.519316037433976e-06, 'learning_rate/action_model': 8.519316037433977e-05, 'epoch': 0.1}) +[2026-10-03 13:17:17,495][starVLA.training.train_starvla][INFO] - Step 1367, Loss: {'train/loss': 0.09365332126617432, 'train/grad_norm_pre_clip': 35.78300476074219, 'timing/data': 0.0002599400031613186, 'timing/model': 0.5847552219929639, 'learning_rate/qwen_vl_interface': 8.517106070975801e-06, 'learning_rate/action_model': 8.5171060709758e-05, 'epoch': 0.1}) +[2026-10-03 13:17:18,198][starVLA.training.train_starvla][INFO] - Step 1368, Loss: {'train/loss': 4.099482612218708e-05, 'train/grad_norm_pre_clip': 0.024242211133241653, 'timing/data': 0.000314899007207714, 'timing/model': 0.3925495849980507, 'learning_rate/qwen_vl_interface': 8.514894761533729e-06, 'learning_rate/action_model': 8.514894761533728e-05, 'epoch': 0.1}) +[2026-10-03 13:17:18,900][starVLA.training.train_starvla][INFO] - Step 1369, Loss: {'train/loss': 0.0292070135474205, 'train/grad_norm_pre_clip': 13.583141326904297, 'timing/data': 0.0002686399966478348, 'timing/model': 0.3917758750030771, 'learning_rate/qwen_vl_interface': 8.512682110016747e-06, 'learning_rate/action_model': 8.512682110016747e-05, 'epoch': 0.1}) +[2026-10-03 13:17:19,602][starVLA.training.train_starvla][INFO] - Step 1370, Loss: {'train/loss': 0.285879909992218, 'train/grad_norm_pre_clip': 85.67958068847656, 'timing/data': 0.0003432399971643463, 'timing/model': 0.3918805249995785, 'learning_rate/qwen_vl_interface': 8.51046811733439e-06, 'learning_rate/action_model': 8.51046811733439e-05, 'epoch': 0.1}) +[2026-10-03 13:17:20,302][starVLA.training.train_starvla][INFO] - Step 1371, Loss: {'train/loss': 0.3077056109905243, 'train/grad_norm_pre_clip': 41.68992233276367, 'timing/data': 0.00025196000933647156, 'timing/model': 0.39113316500151996, 'learning_rate/qwen_vl_interface': 8.50825278439675e-06, 'learning_rate/action_model': 8.508252784396749e-05, 'epoch': 0.1}) +[2026-10-03 13:17:21,004][starVLA.training.train_starvla][INFO] - Step 1372, Loss: {'train/loss': 0.6881695985794067, 'train/grad_norm_pre_clip': 156.5398712158203, 'timing/data': 0.00036851999175269157, 'timing/model': 0.3916924149962142, 'learning_rate/qwen_vl_interface': 8.506036112114465e-06, 'learning_rate/action_model': 8.506036112114464e-05, 'epoch': 0.1}) +[2026-10-03 13:17:21,894][starVLA.training.train_starvla][INFO] - Step 1373, Loss: {'train/loss': 7.028285472188145e-05, 'train/grad_norm_pre_clip': 0.017958518117666245, 'timing/data': 0.00029035999614279717, 'timing/model': 0.3915788650047034, 'learning_rate/qwen_vl_interface': 8.503818101398725e-06, 'learning_rate/action_model': 8.503818101398725e-05, 'epoch': 0.1}) +[2026-10-03 13:17:22,594][starVLA.training.train_starvla][INFO] - Step 1374, Loss: {'train/loss': 0.00011226569768041372, 'train/grad_norm_pre_clip': 0.0693567767739296, 'timing/data': 0.0002484299911884591, 'timing/model': 0.39158283500000834, 'learning_rate/qwen_vl_interface': 8.501598753161273e-06, 'learning_rate/action_model': 8.501598753161272e-05, 'epoch': 0.1}) +[2026-10-03 13:17:23,294][starVLA.training.train_starvla][INFO] - Step 1375, Loss: {'train/loss': 4.4404660002328455e-06, 'train/grad_norm_pre_clip': 0.0010748617351055145, 'timing/data': 0.0002028599992627278, 'timing/model': 0.391095294005936, 'learning_rate/qwen_vl_interface': 8.499378068314395e-06, 'learning_rate/action_model': 8.499378068314395e-05, 'epoch': 0.1}) +[2026-10-03 13:17:23,997][starVLA.training.train_starvla][INFO] - Step 1376, Loss: {'train/loss': 0.2882811427116394, 'train/grad_norm_pre_clip': 128.4702606201172, 'timing/data': 0.0003194800083292648, 'timing/model': 0.392450753992307, 'learning_rate/qwen_vl_interface': 8.497156047770932e-06, 'learning_rate/action_model': 8.497156047770933e-05, 'epoch': 0.1}) +[2026-10-03 13:17:24,698][starVLA.training.train_starvla][INFO] - Step 1377, Loss: {'train/loss': 3.1664235393691342e-06, 'train/grad_norm_pre_clip': 0.0008429118315689266, 'timing/data': 0.0002782100054901093, 'timing/model': 0.39156577399990056, 'learning_rate/qwen_vl_interface': 8.494932692444276e-06, 'learning_rate/action_model': 8.494932692444276e-05, 'epoch': 0.1}) +[2026-10-03 13:17:25,592][starVLA.training.train_starvla][INFO] - Step 1378, Loss: {'train/loss': 0.5033087730407715, 'train/grad_norm_pre_clip': 36.056312561035156, 'timing/data': 0.0004199399991193786, 'timing/model': 0.3920569940091809, 'learning_rate/qwen_vl_interface': 8.492708003248362e-06, 'learning_rate/action_model': 8.492708003248362e-05, 'epoch': 0.1}) +[2026-10-03 13:17:26,294][starVLA.training.train_starvla][INFO] - Step 1379, Loss: {'train/loss': 0.6607801914215088, 'train/grad_norm_pre_clip': 49.23595428466797, 'timing/data': 0.00020267000945750624, 'timing/model': 0.3928774140076712, 'learning_rate/qwen_vl_interface': 8.490481981097674e-06, 'learning_rate/action_model': 8.490481981097674e-05, 'epoch': 0.1}) +[2026-10-03 13:17:26,997][starVLA.training.train_starvla][INFO] - Step 1380, Loss: {'train/loss': 8.705552318133414e-05, 'train/grad_norm_pre_clip': 0.024775352329015732, 'timing/data': 0.00034455000422894955, 'timing/model': 0.3923856540059205, 'learning_rate/qwen_vl_interface': 8.488254626907249e-06, 'learning_rate/action_model': 8.488254626907249e-05, 'epoch': 0.1}) +[2026-10-03 13:17:27,698][starVLA.training.train_starvla][INFO] - Step 1381, Loss: {'train/loss': 0.5354253053665161, 'train/grad_norm_pre_clip': 82.1066665649414, 'timing/data': 0.0003345800068927929, 'timing/model': 0.3915874439990148, 'learning_rate/qwen_vl_interface': 8.486025941592664e-06, 'learning_rate/action_model': 8.486025941592664e-05, 'epoch': 0.1}) +[2026-10-03 13:17:28,401][starVLA.training.train_starvla][INFO] - Step 1382, Loss: {'train/loss': 0.0003797906683757901, 'train/grad_norm_pre_clip': 0.23207291960716248, 'timing/data': 0.0003013200039276853, 'timing/model': 0.3924578439909965, 'learning_rate/qwen_vl_interface': 8.483795926070054e-06, 'learning_rate/action_model': 8.483795926070052e-05, 'epoch': 0.1}) +[2026-10-03 13:17:29,296][starVLA.training.train_starvla][INFO] - Step 1383, Loss: {'train/loss': 0.01447487436234951, 'train/grad_norm_pre_clip': 34.8787841796875, 'timing/data': 0.00021365999418776482, 'timing/model': 0.392485863994807, 'learning_rate/qwen_vl_interface': 8.481564581256087e-06, 'learning_rate/action_model': 8.481564581256087e-05, 'epoch': 0.1}) +[2026-10-03 13:17:29,998][starVLA.training.train_starvla][INFO] - Step 1384, Loss: {'train/loss': 0.0026274926494807005, 'train/grad_norm_pre_clip': 1.8684468269348145, 'timing/data': 0.00028198999643791467, 'timing/model': 0.3913696340023307, 'learning_rate/qwen_vl_interface': 8.479331908067987e-06, 'learning_rate/action_model': 8.479331908067987e-05, 'epoch': 0.1}) +[2026-10-03 13:17:30,699][starVLA.training.train_starvla][INFO] - Step 1385, Loss: {'train/loss': 1.4602728697354905e-05, 'train/grad_norm_pre_clip': 0.0029737590812146664, 'timing/data': 0.00029785999504383653, 'timing/model': 0.39139670300937723, 'learning_rate/qwen_vl_interface': 8.477097907423524e-06, 'learning_rate/action_model': 8.477097907423524e-05, 'epoch': 0.1}) +[2026-10-03 13:17:31,400][starVLA.training.train_starvla][INFO] - Step 1386, Loss: {'train/loss': 0.2944267690181732, 'train/grad_norm_pre_clip': 53.52702331542969, 'timing/data': 0.00034185999538749456, 'timing/model': 0.39145862399891485, 'learning_rate/qwen_vl_interface': 8.474862580241006e-06, 'learning_rate/action_model': 8.474862580241006e-05, 'epoch': 0.1}) +[2026-10-03 13:17:32,101][starVLA.training.train_starvla][INFO] - Step 1387, Loss: {'train/loss': 0.00017925829160958529, 'train/grad_norm_pre_clip': 0.08020897954702377, 'timing/data': 0.00023766000231262296, 'timing/model': 0.3922437939909287, 'learning_rate/qwen_vl_interface': 8.472625927439299e-06, 'learning_rate/action_model': 8.472625927439299e-05, 'epoch': 0.1}) +[2026-10-03 13:17:32,997][starVLA.training.train_starvla][INFO] - Step 1388, Loss: {'train/loss': 0.48135116696357727, 'train/grad_norm_pre_clip': 114.53825378417969, 'timing/data': 0.0002965199964819476, 'timing/model': 0.584353172002011, 'learning_rate/qwen_vl_interface': 8.4703879499378e-06, 'learning_rate/action_model': 8.4703879499378e-05, 'epoch': 0.1}) +[2026-10-03 13:17:33,703][starVLA.training.train_starvla][INFO] - Step 1389, Loss: {'train/loss': 0.004101024940609932, 'train/grad_norm_pre_clip': 3.285775661468506, 'timing/data': 0.00021569999807979912, 'timing/model': 0.3925175939948531, 'learning_rate/qwen_vl_interface': 8.46814864865646e-06, 'learning_rate/action_model': 8.46814864865646e-05, 'epoch': 0.1}) +[2026-10-03 13:17:34,405][starVLA.training.train_starvla][INFO] - Step 1390, Loss: {'train/loss': 8.630873344372958e-05, 'train/grad_norm_pre_clip': 0.06809995323419571, 'timing/data': 0.00020148900512140244, 'timing/model': 0.39189237500249874, 'learning_rate/qwen_vl_interface': 8.465908024515771e-06, 'learning_rate/action_model': 8.46590802451577e-05, 'epoch': 0.1}) +[2026-10-03 13:17:35,105][starVLA.training.train_starvla][INFO] - Step 1391, Loss: {'train/loss': 0.019069714471697807, 'train/grad_norm_pre_clip': 10.18858814239502, 'timing/data': 0.0001990899909287691, 'timing/model': 0.3915877340041334, 'learning_rate/qwen_vl_interface': 8.463666078436769e-06, 'learning_rate/action_model': 8.463666078436769e-05, 'epoch': 0.1}) +[2026-10-03 13:17:35,809][starVLA.training.train_starvla][INFO] - Step 1392, Loss: {'train/loss': 3.805374581133947e-05, 'train/grad_norm_pre_clip': 0.02948549948632717, 'timing/data': 0.0003524300118442625, 'timing/model': 0.39228773400827777, 'learning_rate/qwen_vl_interface': 8.46142281134103e-06, 'learning_rate/action_model': 8.461422811341031e-05, 'epoch': 0.1}) +[2026-10-03 13:17:36,704][starVLA.training.train_starvla][INFO] - Step 1393, Loss: {'train/loss': 0.007262512110173702, 'train/grad_norm_pre_clip': 5.709064960479736, 'timing/data': 0.000314990000333637, 'timing/model': 0.5849019419983961, 'learning_rate/qwen_vl_interface': 8.459178224150683e-06, 'learning_rate/action_model': 8.459178224150683e-05, 'epoch': 0.1}) +[2026-10-03 13:17:37,405][starVLA.training.train_starvla][INFO] - Step 1394, Loss: {'train/loss': 5.613000030280091e-05, 'train/grad_norm_pre_clip': 0.05291423201560974, 'timing/data': 0.00023910000163596123, 'timing/model': 0.3916180449887179, 'learning_rate/qwen_vl_interface': 8.456932317788389e-06, 'learning_rate/action_model': 8.456932317788389e-05, 'epoch': 0.1}) +[2026-10-03 13:17:38,105][starVLA.training.train_starvla][INFO] - Step 1395, Loss: {'train/loss': 0.0008652035030536354, 'train/grad_norm_pre_clip': 0.9270473718643188, 'timing/data': 0.00020571000641211867, 'timing/model': 0.39108300500083715, 'learning_rate/qwen_vl_interface': 8.454685093177356e-06, 'learning_rate/action_model': 8.454685093177354e-05, 'epoch': 0.1}) +[2026-10-03 13:17:38,806][starVLA.training.train_starvla][INFO] - Step 1396, Loss: {'train/loss': 0.00022725264716427773, 'train/grad_norm_pre_clip': 0.057511717081069946, 'timing/data': 0.0002437799994368106, 'timing/model': 0.3921996939898236, 'learning_rate/qwen_vl_interface': 8.452436551241328e-06, 'learning_rate/action_model': 8.452436551241328e-05, 'epoch': 0.1}) +[2026-10-03 13:17:39,506][starVLA.training.train_starvla][INFO] - Step 1397, Loss: {'train/loss': 0.0014584532473236322, 'train/grad_norm_pre_clip': 0.75532066822052, 'timing/data': 0.00033178999728988856, 'timing/model': 0.3909622150094947, 'learning_rate/qwen_vl_interface': 8.450186692904605e-06, 'learning_rate/action_model': 8.450186692904606e-05, 'epoch': 0.1}) +[2026-10-03 13:17:40,208][starVLA.training.train_starvla][INFO] - Step 1398, Loss: {'train/loss': 9.25301865208894e-05, 'train/grad_norm_pre_clip': 0.04216057434678078, 'timing/data': 0.00037077000888530165, 'timing/model': 0.3917282150068786, 'learning_rate/qwen_vl_interface': 8.447935519092014e-06, 'learning_rate/action_model': 8.447935519092014e-05, 'epoch': 0.1}) +[2026-10-03 13:17:41,100][starVLA.training.train_starvla][INFO] - Step 1399, Loss: {'train/loss': 3.598617695388384e-05, 'train/grad_norm_pre_clip': 0.06097126752138138, 'timing/data': 0.00021194000146351755, 'timing/model': 0.3913223240087973, 'learning_rate/qwen_vl_interface': 8.445683030728925e-06, 'learning_rate/action_model': 8.445683030728925e-05, 'epoch': 0.1}) +[2026-10-03 13:17:41,803][starVLA.training.train_starvla][INFO] - Step 1400, Loss: {'train/loss': 5.729380791308358e-05, 'train/grad_norm_pre_clip': 0.09855690598487854, 'timing/data': 0.0002391500020166859, 'timing/model': 0.3923369139956776, 'learning_rate/qwen_vl_interface': 8.443429228741256e-06, 'learning_rate/action_model': 8.443429228741255e-05, 'epoch': 0.1}) +[2026-10-03 13:17:42,503][starVLA.training.train_starvla][INFO] - Step 1401, Loss: {'train/loss': 0.04809315502643585, 'train/grad_norm_pre_clip': 69.17569732666016, 'timing/data': 0.0003278299991507083, 'timing/model': 0.3911591439973563, 'learning_rate/qwen_vl_interface': 8.441174114055457e-06, 'learning_rate/action_model': 8.441174114055456e-05, 'epoch': 0.1}) +[2026-10-03 13:17:43,204][starVLA.training.train_starvla][INFO] - Step 1402, Loss: {'train/loss': 0.018008992075920105, 'train/grad_norm_pre_clip': 54.37968826293945, 'timing/data': 0.0002545099996495992, 'timing/model': 0.39220178400864825, 'learning_rate/qwen_vl_interface': 8.43891768759852e-06, 'learning_rate/action_model': 8.43891768759852e-05, 'epoch': 0.1}) +[2026-10-03 13:17:43,905][starVLA.training.train_starvla][INFO] - Step 1403, Loss: {'train/loss': 0.00014317578461486846, 'train/grad_norm_pre_clip': 0.04756760969758034, 'timing/data': 0.0002278100000694394, 'timing/model': 0.39154492400120944, 'learning_rate/qwen_vl_interface': 8.436659950297978e-06, 'learning_rate/action_model': 8.436659950297977e-05, 'epoch': 0.1}) +[2026-10-03 13:17:44,798][starVLA.training.train_starvla][INFO] - Step 1404, Loss: {'train/loss': 3.780367478611879e-05, 'train/grad_norm_pre_clip': 0.010642698965966702, 'timing/data': 0.00036029999318998307, 'timing/model': 0.5824524019990349, 'learning_rate/qwen_vl_interface': 8.4344009030819e-06, 'learning_rate/action_model': 8.4344009030819e-05, 'epoch': 0.1}) +[2026-10-03 13:17:45,499][starVLA.training.train_starvla][INFO] - Step 1405, Loss: {'train/loss': 0.0005578413838520646, 'train/grad_norm_pre_clip': 2.1476244926452637, 'timing/data': 0.00022792001254856586, 'timing/model': 0.3915748339932179, 'learning_rate/qwen_vl_interface': 8.432140546878899e-06, 'learning_rate/action_model': 8.432140546878898e-05, 'epoch': 0.1}) +[2026-10-03 13:17:46,200][starVLA.training.train_starvla][INFO] - Step 1406, Loss: {'train/loss': 1.5764806448714808e-05, 'train/grad_norm_pre_clip': 0.004043439868837595, 'timing/data': 0.0002465299912728369, 'timing/model': 0.3919901039917022, 'learning_rate/qwen_vl_interface': 8.429878882618116e-06, 'learning_rate/action_model': 8.429878882618116e-05, 'epoch': 0.1}) +[2026-10-03 13:17:46,901][starVLA.training.train_starvla][INFO] - Step 1407, Loss: {'train/loss': 0.08466333150863647, 'train/grad_norm_pre_clip': 37.276512145996094, 'timing/data': 0.00029028000426478684, 'timing/model': 0.39187456500076223, 'learning_rate/qwen_vl_interface': 8.42761591122924e-06, 'learning_rate/action_model': 8.427615911229241e-05, 'epoch': 0.1}) +[2026-10-03 13:17:47,603][starVLA.training.train_starvla][INFO] - Step 1408, Loss: {'train/loss': 3.119293978670612e-05, 'train/grad_norm_pre_clip': 0.01719794049859047, 'timing/data': 0.00032469000143464655, 'timing/model': 0.39147515399963595, 'learning_rate/qwen_vl_interface': 8.425351633642491e-06, 'learning_rate/action_model': 8.425351633642492e-05, 'epoch': 0.1}) +[2026-10-03 13:17:48,304][starVLA.training.train_starvla][INFO] - Step 1409, Loss: {'train/loss': 0.2086033672094345, 'train/grad_norm_pre_clip': 63.12241744995117, 'timing/data': 0.00029221999284345657, 'timing/model': 0.39159428399580065, 'learning_rate/qwen_vl_interface': 8.423086050788632e-06, 'learning_rate/action_model': 8.423086050788631e-05, 'epoch': 0.1}) +[2026-10-03 13:17:49,196][starVLA.training.train_starvla][INFO] - Step 1410, Loss: {'train/loss': 0.05683687701821327, 'train/grad_norm_pre_clip': 206.09347534179688, 'timing/data': 0.0003538699966156855, 'timing/model': 0.3922129440034041, 'learning_rate/qwen_vl_interface': 8.420819163598953e-06, 'learning_rate/action_model': 8.420819163598952e-05, 'epoch': 0.1}) +[2026-10-03 13:17:49,896][starVLA.training.train_starvla][INFO] - Step 1411, Loss: {'train/loss': 0.001037060865201056, 'train/grad_norm_pre_clip': 2.7908432483673096, 'timing/data': 0.0003236899938201532, 'timing/model': 0.39129065499582794, 'learning_rate/qwen_vl_interface': 8.418550973005289e-06, 'learning_rate/action_model': 8.418550973005289e-05, 'epoch': 0.1}) +[2026-10-03 13:17:50,600][starVLA.training.train_starvla][INFO] - Step 1412, Loss: {'train/loss': 7.681281203986146e-06, 'train/grad_norm_pre_clip': 0.0019951462745666504, 'timing/data': 0.00029630999779328704, 'timing/model': 0.3925719639955787, 'learning_rate/qwen_vl_interface': 8.416281479940005e-06, 'learning_rate/action_model': 8.416281479940005e-05, 'epoch': 0.1}) +[2026-10-03 13:17:51,301][starVLA.training.train_starvla][INFO] - Step 1413, Loss: {'train/loss': 4.2044710426125675e-05, 'train/grad_norm_pre_clip': 0.014995268546044827, 'timing/data': 0.0003564000071492046, 'timing/model': 0.39183258400589693, 'learning_rate/qwen_vl_interface': 8.414010685336008e-06, 'learning_rate/action_model': 8.414010685336008e-05, 'epoch': 0.1}) +[2026-10-03 13:17:52,004][starVLA.training.train_starvla][INFO] - Step 1414, Loss: {'train/loss': 0.000943138322327286, 'train/grad_norm_pre_clip': 1.4774514436721802, 'timing/data': 0.00020080000103916973, 'timing/model': 0.39289699398796074, 'learning_rate/qwen_vl_interface': 8.411738590126733e-06, 'learning_rate/action_model': 8.411738590126732e-05, 'epoch': 0.1}) +[2026-10-03 13:17:52,896][starVLA.training.train_starvla][INFO] - Step 1415, Loss: {'train/loss': 2.328156006115023e-05, 'train/grad_norm_pre_clip': 0.0070259906351566315, 'timing/data': 0.0003400200075702742, 'timing/model': 0.3920168349868618, 'learning_rate/qwen_vl_interface': 8.409465195246152e-06, 'learning_rate/action_model': 8.409465195246152e-05, 'epoch': 0.1}) +[2026-10-03 13:17:53,599][starVLA.training.train_starvla][INFO] - Step 1416, Loss: {'train/loss': 0.11326642334461212, 'train/grad_norm_pre_clip': 115.40811920166016, 'timing/data': 0.00027238999609835446, 'timing/model': 0.3926710740051931, 'learning_rate/qwen_vl_interface': 8.407190501628772e-06, 'learning_rate/action_model': 8.407190501628772e-05, 'epoch': 0.1}) +[2026-10-03 13:17:54,300][starVLA.training.train_starvla][INFO] - Step 1417, Loss: {'train/loss': 0.37670356035232544, 'train/grad_norm_pre_clip': 137.06434631347656, 'timing/data': 0.00032857000769581646, 'timing/model': 0.3915218940092018, 'learning_rate/qwen_vl_interface': 8.404914510209635e-06, 'learning_rate/action_model': 8.404914510209635e-05, 'epoch': 0.1}) +[2026-10-03 13:17:55,002][starVLA.training.train_starvla][INFO] - Step 1418, Loss: {'train/loss': 0.025922447443008423, 'train/grad_norm_pre_clip': 18.773927688598633, 'timing/data': 0.00031735999800730497, 'timing/model': 0.39224788499996066, 'learning_rate/qwen_vl_interface': 8.402637221924315e-06, 'learning_rate/action_model': 8.402637221924314e-05, 'epoch': 0.1}) +[2026-10-03 13:17:55,703][starVLA.training.train_starvla][INFO] - Step 1419, Loss: {'train/loss': 0.48769307136535645, 'train/grad_norm_pre_clip': 37.723304748535156, 'timing/data': 0.0003119300090475008, 'timing/model': 0.39111054499517195, 'learning_rate/qwen_vl_interface': 8.400358637708919e-06, 'learning_rate/action_model': 8.400358637708918e-05, 'epoch': 0.1}) +[2026-10-03 13:17:56,599][starVLA.training.train_starvla][INFO] - Step 1420, Loss: {'train/loss': 1.3603712432086468e-05, 'train/grad_norm_pre_clip': 0.006280659697949886, 'timing/data': 0.00024187000235542655, 'timing/model': 0.5860116319963709, 'learning_rate/qwen_vl_interface': 8.398078758500086e-06, 'learning_rate/action_model': 8.398078758500086e-05, 'epoch': 0.1}) +[2026-10-03 13:17:57,301][starVLA.training.train_starvla][INFO] - Step 1421, Loss: {'train/loss': 0.3736809194087982, 'train/grad_norm_pre_clip': 217.8249969482422, 'timing/data': 0.0003076599969062954, 'timing/model': 0.39183051500003785, 'learning_rate/qwen_vl_interface': 8.395797585234987e-06, 'learning_rate/action_model': 8.395797585234988e-05, 'epoch': 0.1}) +[2026-10-03 13:17:58,003][starVLA.training.train_starvla][INFO] - Step 1422, Loss: {'train/loss': 7.003527571214363e-07, 'train/grad_norm_pre_clip': 0.00016357790445908904, 'timing/data': 0.0002453500055707991, 'timing/model': 0.3927249749976909, 'learning_rate/qwen_vl_interface': 8.39351511885133e-06, 'learning_rate/action_model': 8.393515118851329e-05, 'epoch': 0.1}) +[2026-10-03 13:17:58,704][starVLA.training.train_starvla][INFO] - Step 1423, Loss: {'train/loss': 0.5502289533615112, 'train/grad_norm_pre_clip': 109.32015228271484, 'timing/data': 0.0002498199901310727, 'timing/model': 0.39186484400124755, 'learning_rate/qwen_vl_interface': 8.391231360287348e-06, 'learning_rate/action_model': 8.391231360287347e-05, 'epoch': 0.1}) +[2026-10-03 13:17:59,407][starVLA.training.train_starvla][INFO] - Step 1424, Loss: {'train/loss': 0.4618004858493805, 'train/grad_norm_pre_clip': 423.283935546875, 'timing/data': 0.00030239000625442713, 'timing/model': 0.3926678839925444, 'learning_rate/qwen_vl_interface': 8.388946310481807e-06, 'learning_rate/action_model': 8.388946310481807e-05, 'epoch': 0.1}) +[2026-10-03 13:18:00,109][starVLA.training.train_starvla][INFO] - Step 1425, Loss: {'train/loss': 6.35514243185753e-06, 'train/grad_norm_pre_clip': 0.008018757216632366, 'timing/data': 0.0002474799985066056, 'timing/model': 0.39156126500165556, 'learning_rate/qwen_vl_interface': 8.386659970374007e-06, 'learning_rate/action_model': 8.386659970374006e-05, 'epoch': 0.1}) +[2026-10-03 13:18:00,999][starVLA.training.train_starvla][INFO] - Step 1426, Loss: {'train/loss': 0.7966787219047546, 'train/grad_norm_pre_clip': 48.0847282409668, 'timing/data': 0.0002878300001611933, 'timing/model': 0.392555935002747, 'learning_rate/qwen_vl_interface': 8.384372340903775e-06, 'learning_rate/action_model': 8.384372340903775e-05, 'epoch': 0.1}) +[2026-10-03 13:18:01,700][starVLA.training.train_starvla][INFO] - Step 1427, Loss: {'train/loss': 0.2859005331993103, 'train/grad_norm_pre_clip': 188.21336364746094, 'timing/data': 0.00025037999148480594, 'timing/model': 0.39168556399818044, 'learning_rate/qwen_vl_interface': 8.382083423011472e-06, 'learning_rate/action_model': 8.382083423011471e-05, 'epoch': 0.1}) +[2026-10-03 13:18:02,402][starVLA.training.train_starvla][INFO] - Step 1428, Loss: {'train/loss': 0.00035150747862644494, 'train/grad_norm_pre_clip': 0.15094879269599915, 'timing/data': 0.00037911000254098326, 'timing/model': 0.3915933340031188, 'learning_rate/qwen_vl_interface': 8.379793217637979e-06, 'learning_rate/action_model': 8.379793217637979e-05, 'epoch': 0.1}) +[2026-10-03 13:18:03,103][starVLA.training.train_starvla][INFO] - Step 1429, Loss: {'train/loss': 0.0007678434485569596, 'train/grad_norm_pre_clip': 3.023541212081909, 'timing/data': 0.00030028900073375553, 'timing/model': 0.39155022500199266, 'learning_rate/qwen_vl_interface': 8.377501725724719e-06, 'learning_rate/action_model': 8.377501725724719e-05, 'epoch': 0.1}) +[2026-10-03 13:18:03,806][starVLA.training.train_starvla][INFO] - Step 1430, Loss: {'train/loss': 0.3750838339328766, 'train/grad_norm_pre_clip': 722.8585815429688, 'timing/data': 0.00020279000455047935, 'timing/model': 0.3919318550033495, 'learning_rate/qwen_vl_interface': 8.375208948213636e-06, 'learning_rate/action_model': 8.375208948213635e-05, 'epoch': 0.1}) +[2026-10-03 13:18:04,701][starVLA.training.train_starvla][INFO] - Step 1431, Loss: {'train/loss': 0.6765955090522766, 'train/grad_norm_pre_clip': 280.2657165527344, 'timing/data': 0.00022318901028484106, 'timing/model': 0.3927810449968092, 'learning_rate/qwen_vl_interface': 8.372914886047203e-06, 'learning_rate/action_model': 8.372914886047203e-05, 'epoch': 0.1}) +[2026-10-03 13:18:05,404][starVLA.training.train_starvla][INFO] - Step 1432, Loss: {'train/loss': 0.6861194372177124, 'train/grad_norm_pre_clip': 23.804672241210938, 'timing/data': 0.00033557000278960913, 'timing/model': 0.3918008039909182, 'learning_rate/qwen_vl_interface': 8.370619540168422e-06, 'learning_rate/action_model': 8.370619540168423e-05, 'epoch': 0.1}) +[2026-10-03 13:18:06,105][starVLA.training.train_starvla][INFO] - Step 1433, Loss: {'train/loss': 0.3587442636489868, 'train/grad_norm_pre_clip': 213.349365234375, 'timing/data': 0.0002036200021393597, 'timing/model': 0.3913390539964894, 'learning_rate/qwen_vl_interface': 8.368322911520826e-06, 'learning_rate/action_model': 8.368322911520826e-05, 'epoch': 0.1}) +[2026-10-03 13:18:06,809][starVLA.training.train_starvla][INFO] - Step 1434, Loss: {'train/loss': 0.0017151004867628217, 'train/grad_norm_pre_clip': 2.719309091567993, 'timing/data': 0.00031106000824365765, 'timing/model': 0.3927440350089455, 'learning_rate/qwen_vl_interface': 8.36602500104847e-06, 'learning_rate/action_model': 8.36602500104847e-05, 'epoch': 0.1}) +[2026-10-03 13:18:07,513][starVLA.training.train_starvla][INFO] - Step 1435, Loss: {'train/loss': 0.13559143245220184, 'train/grad_norm_pre_clip': 127.5053482055664, 'timing/data': 0.00025411999376956373, 'timing/model': 0.39371927399770357, 'learning_rate/qwen_vl_interface': 8.36372580969594e-06, 'learning_rate/action_model': 8.36372580969594e-05, 'epoch': 0.11}) +[2026-10-03 13:18:08,413][starVLA.training.train_starvla][INFO] - Step 1436, Loss: {'train/loss': 0.0005020816461183131, 'train/grad_norm_pre_clip': 0.22695456445217133, 'timing/data': 0.00033104000613093376, 'timing/model': 0.39377567400515545, 'learning_rate/qwen_vl_interface': 8.361425338408344e-06, 'learning_rate/action_model': 8.361425338408343e-05, 'epoch': 0.11}) +[2026-10-03 13:18:09,117][starVLA.training.train_starvla][INFO] - Step 1437, Loss: {'train/loss': 0.21241767704486847, 'train/grad_norm_pre_clip': 96.29763793945312, 'timing/data': 0.00024608000239823014, 'timing/model': 0.39293673500651494, 'learning_rate/qwen_vl_interface': 8.359123588131319e-06, 'learning_rate/action_model': 8.359123588131319e-05, 'epoch': 0.11}) +[2026-10-03 13:18:09,822][starVLA.training.train_starvla][INFO] - Step 1438, Loss: {'train/loss': 0.35537880659103394, 'train/grad_norm_pre_clip': 194.32354736328125, 'timing/data': 0.0002071899943985045, 'timing/model': 0.39416815400181804, 'learning_rate/qwen_vl_interface': 8.35682055981103e-06, 'learning_rate/action_model': 8.35682055981103e-05, 'epoch': 0.11}) +[2026-10-03 13:18:10,526][starVLA.training.train_starvla][INFO] - Step 1439, Loss: {'train/loss': 0.0009735296480357647, 'train/grad_norm_pre_clip': 0.3920409381389618, 'timing/data': 0.0003156700113322586, 'timing/model': 0.3931163439992815, 'learning_rate/qwen_vl_interface': 8.354516254394163e-06, 'learning_rate/action_model': 8.354516254394162e-05, 'epoch': 0.11}) +[2026-10-03 13:18:11,233][starVLA.training.train_starvla][INFO] - Step 1440, Loss: {'train/loss': 0.0002832838799804449, 'train/grad_norm_pre_clip': 0.07488734275102615, 'timing/data': 0.00029643999005202204, 'timing/model': 0.3943533040001057, 'learning_rate/qwen_vl_interface': 8.352210672827932e-06, 'learning_rate/action_model': 8.352210672827931e-05, 'epoch': 0.11}) +[2026-10-03 13:18:12,128][starVLA.training.train_starvla][INFO] - Step 1441, Loss: {'train/loss': 0.0004646811285056174, 'train/grad_norm_pre_clip': 0.09636425971984863, 'timing/data': 0.00021189000108279288, 'timing/model': 0.5849925609945785, 'learning_rate/qwen_vl_interface': 8.349903816060074e-06, 'learning_rate/action_model': 8.349903816060075e-05, 'epoch': 0.11}) +[2026-10-03 13:18:12,834][starVLA.training.train_starvla][INFO] - Step 1442, Loss: {'train/loss': 0.00026542512932792306, 'train/grad_norm_pre_clip': 0.07167832553386688, 'timing/data': 0.00020228000357747078, 'timing/model': 0.3943624339881353, 'learning_rate/qwen_vl_interface': 8.34759568503885e-06, 'learning_rate/action_model': 8.347595685038851e-05, 'epoch': 0.11}) +[2026-10-03 13:18:13,537][starVLA.training.train_starvla][INFO] - Step 1443, Loss: {'train/loss': 0.3501797318458557, 'train/grad_norm_pre_clip': 380.9775390625, 'timing/data': 0.00020138001127634197, 'timing/model': 0.3930840940010967, 'learning_rate/qwen_vl_interface': 8.345286280713049e-06, 'learning_rate/action_model': 8.345286280713048e-05, 'epoch': 0.11}) +[2026-10-03 13:18:14,243][starVLA.training.train_starvla][INFO] - Step 1444, Loss: {'train/loss': 0.0008346894173882902, 'train/grad_norm_pre_clip': 0.14842955768108368, 'timing/data': 0.00021713999740313739, 'timing/model': 0.39364499400835484, 'learning_rate/qwen_vl_interface': 8.342975604031974e-06, 'learning_rate/action_model': 8.342975604031974e-05, 'epoch': 0.11}) +[2026-10-03 13:18:14,947][starVLA.training.train_starvla][INFO] - Step 1445, Loss: {'train/loss': 0.10140763968229294, 'train/grad_norm_pre_clip': 38.603904724121094, 'timing/data': 0.00027501999284140766, 'timing/model': 0.39281775400741026, 'learning_rate/qwen_vl_interface': 8.340663655945462e-06, 'learning_rate/action_model': 8.340663655945462e-05, 'epoch': 0.11}) +[2026-10-03 13:18:15,845][starVLA.training.train_starvla][INFO] - Step 1446, Loss: {'train/loss': 0.00040198443457484245, 'train/grad_norm_pre_clip': 0.08774656057357788, 'timing/data': 0.00020450000010896474, 'timing/model': 0.5863467419985682, 'learning_rate/qwen_vl_interface': 8.338350437403866e-06, 'learning_rate/action_model': 8.338350437403865e-05, 'epoch': 0.11}) +[2026-10-03 13:18:16,550][starVLA.training.train_starvla][INFO] - Step 1447, Loss: {'train/loss': 0.6719214916229248, 'train/grad_norm_pre_clip': 97.7151107788086, 'timing/data': 0.0003009100037161261, 'timing/model': 0.393536514005973, 'learning_rate/qwen_vl_interface': 8.336035949358063e-06, 'learning_rate/action_model': 8.336035949358062e-05, 'epoch': 0.11}) +[2026-10-03 13:18:17,256][starVLA.training.train_starvla][INFO] - Step 1448, Loss: {'train/loss': 0.6292456984519958, 'train/grad_norm_pre_clip': 68.6937484741211, 'timing/data': 0.0003427699994063005, 'timing/model': 0.3934254140040139, 'learning_rate/qwen_vl_interface': 8.333720192759448e-06, 'learning_rate/action_model': 8.333720192759448e-05, 'epoch': 0.11}) +[2026-10-03 13:18:17,959][starVLA.training.train_starvla][INFO] - Step 1449, Loss: {'train/loss': 0.2593013346195221, 'train/grad_norm_pre_clip': 58.94321060180664, 'timing/data': 0.00029605001327581704, 'timing/model': 0.39275137399090454, 'learning_rate/qwen_vl_interface': 8.331403168559946e-06, 'learning_rate/action_model': 8.331403168559945e-05, 'epoch': 0.11}) +[2026-10-03 13:18:18,664][starVLA.training.train_starvla][INFO] - Step 1450, Loss: {'train/loss': 0.0005673552514053881, 'train/grad_norm_pre_clip': 0.19045232236385345, 'timing/data': 0.00030030999914743006, 'timing/model': 0.3934531139966566, 'learning_rate/qwen_vl_interface': 8.329084877711995e-06, 'learning_rate/action_model': 8.329084877711996e-05, 'epoch': 0.11}) +[2026-10-03 13:18:19,368][starVLA.training.train_starvla][INFO] - Step 1451, Loss: {'train/loss': 0.0049320706166327, 'train/grad_norm_pre_clip': 4.590897083282471, 'timing/data': 0.00033618000452406704, 'timing/model': 0.3928123250079807, 'learning_rate/qwen_vl_interface': 8.32676532116856e-06, 'learning_rate/action_model': 8.326765321168559e-05, 'epoch': 0.11}) +[2026-10-03 13:18:20,268][starVLA.training.train_starvla][INFO] - Step 1452, Loss: {'train/loss': 0.003328395076096058, 'train/grad_norm_pre_clip': 5.491415023803711, 'timing/data': 0.0002975490060634911, 'timing/model': 0.39396538499568123, 'learning_rate/qwen_vl_interface': 8.324444499883119e-06, 'learning_rate/action_model': 8.324444499883119e-05, 'epoch': 0.11}) +[2026-10-03 13:18:20,971][starVLA.training.train_starvla][INFO] - Step 1453, Loss: {'train/loss': 0.24251919984817505, 'train/grad_norm_pre_clip': 36.46765899658203, 'timing/data': 0.00030075998802203685, 'timing/model': 0.3926768739911495, 'learning_rate/qwen_vl_interface': 8.322122414809678e-06, 'learning_rate/action_model': 8.322122414809677e-05, 'epoch': 0.11}) +[2026-10-03 13:18:21,677][starVLA.training.train_starvla][INFO] - Step 1454, Loss: {'train/loss': 0.014043956995010376, 'train/grad_norm_pre_clip': 18.35487174987793, 'timing/data': 0.000208910001674667, 'timing/model': 0.3941318940051133, 'learning_rate/qwen_vl_interface': 8.319799066902755e-06, 'learning_rate/action_model': 8.319799066902755e-05, 'epoch': 0.11}) +[2026-10-03 13:18:22,382][starVLA.training.train_starvla][INFO] - Step 1455, Loss: {'train/loss': 0.10039687156677246, 'train/grad_norm_pre_clip': 38.89104080200195, 'timing/data': 0.0002062900020973757, 'timing/model': 0.39426230400567874, 'learning_rate/qwen_vl_interface': 8.317474457117393e-06, 'learning_rate/action_model': 8.317474457117392e-05, 'epoch': 0.11}) +[2026-10-03 13:18:23,088][starVLA.training.train_starvla][INFO] - Step 1456, Loss: {'train/loss': 0.03539228066802025, 'train/grad_norm_pre_clip': 17.59757423400879, 'timing/data': 0.0003503600019030273, 'timing/model': 0.3940155440068338, 'learning_rate/qwen_vl_interface': 8.315148586409148e-06, 'learning_rate/action_model': 8.315148586409148e-05, 'epoch': 0.11}) +[2026-10-03 13:18:23,984][starVLA.training.train_starvla][INFO] - Step 1457, Loss: {'train/loss': 0.00044178011012263596, 'train/grad_norm_pre_clip': 0.10751347243785858, 'timing/data': 0.00020202000450808555, 'timing/model': 0.5847760319884401, 'learning_rate/qwen_vl_interface': 8.312821455734104e-06, 'learning_rate/action_model': 8.312821455734103e-05, 'epoch': 0.11}) +[2026-10-03 13:18:24,689][starVLA.training.train_starvla][INFO] - Step 1458, Loss: {'train/loss': 0.5850048065185547, 'train/grad_norm_pre_clip': 37.069881439208984, 'timing/data': 0.00021923000167589635, 'timing/model': 0.3941880540078273, 'learning_rate/qwen_vl_interface': 8.31049306604885e-06, 'learning_rate/action_model': 8.310493066048849e-05, 'epoch': 0.11}) +[2026-10-03 13:18:25,368][starVLA.training.train_starvla][INFO] - Step 1459, Loss: {'train/loss': 0.0007129247533157468, 'train/grad_norm_pre_clip': 0.2128453552722931, 'timing/data': 0.000315010009217076, 'timing/model': 0.39326420400175266, 'learning_rate/qwen_vl_interface': 8.3081634183105e-06, 'learning_rate/action_model': 8.308163418310501e-05, 'epoch': 0.11}) +[2026-10-03 13:18:26,074][starVLA.training.train_starvla][INFO] - Step 1460, Loss: {'train/loss': 0.40052521228790283, 'train/grad_norm_pre_clip': 154.8791046142578, 'timing/data': 0.0003865400067297742, 'timing/model': 0.39358319400344044, 'learning_rate/qwen_vl_interface': 8.305832513476689e-06, 'learning_rate/action_model': 8.305832513476688e-05, 'epoch': 0.11}) +[2026-10-03 13:18:26,780][starVLA.training.train_starvla][INFO] - Step 1461, Loss: {'train/loss': 0.2773980498313904, 'train/grad_norm_pre_clip': 130.3068084716797, 'timing/data': 0.00021470000501722097, 'timing/model': 0.393506353997509, 'learning_rate/qwen_vl_interface': 8.303500352505557e-06, 'learning_rate/action_model': 8.303500352505557e-05, 'epoch': 0.11}) +[2026-10-03 13:18:27,485][starVLA.training.train_starvla][INFO] - Step 1462, Loss: {'train/loss': 0.0005864548729732633, 'train/grad_norm_pre_clip': 0.384532630443573, 'timing/data': 0.0002051699993899092, 'timing/model': 0.3937133139988873, 'learning_rate/qwen_vl_interface': 8.301166936355773e-06, 'learning_rate/action_model': 8.301166936355773e-05, 'epoch': 0.11}) +[2026-10-03 13:18:28,380][starVLA.training.train_starvla][INFO] - Step 1463, Loss: {'train/loss': 0.6820936799049377, 'train/grad_norm_pre_clip': 176.83518981933594, 'timing/data': 0.0003764099965337664, 'timing/model': 0.393769763992168, 'learning_rate/qwen_vl_interface': 8.298832265986514e-06, 'learning_rate/action_model': 8.298832265986513e-05, 'epoch': 0.11}) +[2026-10-03 13:18:29,086][starVLA.training.train_starvla][INFO] - Step 1464, Loss: {'train/loss': 0.3542272448539734, 'train/grad_norm_pre_clip': 79.2870864868164, 'timing/data': 0.00021425000159069896, 'timing/model': 0.393683164002141, 'learning_rate/qwen_vl_interface': 8.296496342357475e-06, 'learning_rate/action_model': 8.296496342357474e-05, 'epoch': 0.11}) +[2026-10-03 13:18:29,790][starVLA.training.train_starvla][INFO] - Step 1465, Loss: {'train/loss': 0.004715072922408581, 'train/grad_norm_pre_clip': 11.900320053100586, 'timing/data': 0.00020781000785063952, 'timing/model': 0.3927380349923624, 'learning_rate/qwen_vl_interface': 8.294159166428864e-06, 'learning_rate/action_model': 8.294159166428864e-05, 'epoch': 0.11}) +[2026-10-03 13:18:30,496][starVLA.training.train_starvla][INFO] - Step 1466, Loss: {'train/loss': 0.10054955631494522, 'train/grad_norm_pre_clip': 51.22090530395508, 'timing/data': 0.0003036599955521524, 'timing/model': 0.3939823639957467, 'learning_rate/qwen_vl_interface': 8.29182073916141e-06, 'learning_rate/action_model': 8.29182073916141e-05, 'epoch': 0.11}) +[2026-10-03 13:18:31,200][starVLA.training.train_starvla][INFO] - Step 1467, Loss: {'train/loss': 0.00010367995855631307, 'train/grad_norm_pre_clip': 0.022790230810642242, 'timing/data': 0.00030309999419841915, 'timing/model': 0.39321949500299525, 'learning_rate/qwen_vl_interface': 8.28948106151635e-06, 'learning_rate/action_model': 8.28948106151635e-05, 'epoch': 0.11}) +[2026-10-03 13:18:32,094][starVLA.training.train_starvla][INFO] - Step 1468, Loss: {'train/loss': 0.051902614533901215, 'train/grad_norm_pre_clip': 42.47019577026367, 'timing/data': 0.00031477000447921455, 'timing/model': 0.3939036849915283, 'learning_rate/qwen_vl_interface': 8.287140134455435e-06, 'learning_rate/action_model': 8.287140134455435e-05, 'epoch': 0.11}) +[2026-10-03 13:18:32,798][starVLA.training.train_starvla][INFO] - Step 1469, Loss: {'train/loss': 0.00029561776318587363, 'train/grad_norm_pre_clip': 0.08384872227907181, 'timing/data': 0.00024446999304927886, 'timing/model': 0.393072034989018, 'learning_rate/qwen_vl_interface': 8.284797958940934e-06, 'learning_rate/action_model': 8.284797958940935e-05, 'epoch': 0.11}) +[2026-10-03 13:18:33,502][starVLA.training.train_starvla][INFO] - Step 1470, Loss: {'train/loss': 0.12232096493244171, 'train/grad_norm_pre_clip': 109.56193542480469, 'timing/data': 0.0003368199977558106, 'timing/model': 0.39381337400118355, 'learning_rate/qwen_vl_interface': 8.282454535935626e-06, 'learning_rate/action_model': 8.282454535935626e-05, 'epoch': 0.11}) +[2026-10-03 13:18:34,208][starVLA.training.train_starvla][INFO] - Step 1471, Loss: {'train/loss': 0.0008452791953459382, 'train/grad_norm_pre_clip': 0.28379538655281067, 'timing/data': 0.00032435999310109764, 'timing/model': 0.3934263740084134, 'learning_rate/qwen_vl_interface': 8.280109866402805e-06, 'learning_rate/action_model': 8.280109866402806e-05, 'epoch': 0.11}) +[2026-10-03 13:18:34,914][starVLA.training.train_starvla][INFO] - Step 1472, Loss: {'train/loss': 0.3580363988876343, 'train/grad_norm_pre_clip': 67.4383316040039, 'timing/data': 0.00026092999905813485, 'timing/model': 0.39391742400766816, 'learning_rate/qwen_vl_interface': 8.277763951306277e-06, 'learning_rate/action_model': 8.277763951306277e-05, 'epoch': 0.11}) +[2026-10-03 13:18:35,809][starVLA.training.train_starvla][INFO] - Step 1473, Loss: {'train/loss': 0.002616968471556902, 'train/grad_norm_pre_clip': 5.723601341247559, 'timing/data': 0.00022935900778975338, 'timing/model': 0.582968582995818, 'learning_rate/qwen_vl_interface': 8.275416791610356e-06, 'learning_rate/action_model': 8.275416791610355e-05, 'epoch': 0.11}) +[2026-10-03 13:18:36,515][starVLA.training.train_starvla][INFO] - Step 1474, Loss: {'train/loss': 0.4505371153354645, 'train/grad_norm_pre_clip': 22.965774536132812, 'timing/data': 0.00028634999762289226, 'timing/model': 0.39437727400218137, 'learning_rate/qwen_vl_interface': 8.27306838827987e-06, 'learning_rate/action_model': 8.273068388279871e-05, 'epoch': 0.11}) +[2026-10-03 13:18:37,219][starVLA.training.train_starvla][INFO] - Step 1475, Loss: {'train/loss': 0.2457071989774704, 'train/grad_norm_pre_clip': 54.18601989746094, 'timing/data': 0.00020954999490641057, 'timing/model': 0.3940471240057377, 'learning_rate/qwen_vl_interface': 8.270718742280162e-06, 'learning_rate/action_model': 8.270718742280162e-05, 'epoch': 0.11}) +[2026-10-03 13:18:37,925][starVLA.training.train_starvla][INFO] - Step 1476, Loss: {'train/loss': 8.093357610050589e-05, 'train/grad_norm_pre_clip': 0.032052356749773026, 'timing/data': 0.0002689300017664209, 'timing/model': 0.3940328739990946, 'learning_rate/qwen_vl_interface': 8.268367854577084e-06, 'learning_rate/action_model': 8.268367854577082e-05, 'epoch': 0.11}) +[2026-10-03 13:18:38,629][starVLA.training.train_starvla][INFO] - Step 1477, Loss: {'train/loss': 0.522364616394043, 'train/grad_norm_pre_clip': 38.021766662597656, 'timing/data': 0.00030166999204084277, 'timing/model': 0.3930667640088359, 'learning_rate/qwen_vl_interface': 8.266015726136991e-06, 'learning_rate/action_model': 8.266015726136991e-05, 'epoch': 0.11}) +[2026-10-03 13:18:39,508][starVLA.training.train_starvla][INFO] - Step 1478, Loss: {'train/loss': 0.21088500320911407, 'train/grad_norm_pre_clip': 101.68738555908203, 'timing/data': 0.0003097400040132925, 'timing/model': 0.5903105509933084, 'learning_rate/qwen_vl_interface': 8.26366235792676e-06, 'learning_rate/action_model': 8.26366235792676e-05, 'epoch': 0.11}) +[2026-10-03 13:18:40,212][starVLA.training.train_starvla][INFO] - Step 1479, Loss: {'train/loss': 0.00019371465896256268, 'train/grad_norm_pre_clip': 0.1494010090827942, 'timing/data': 0.0002042599953711033, 'timing/model': 0.3933960849972209, 'learning_rate/qwen_vl_interface': 8.261307750913769e-06, 'learning_rate/action_model': 8.261307750913769e-05, 'epoch': 0.11}) +[2026-10-03 13:18:40,916][starVLA.training.train_starvla][INFO] - Step 1480, Loss: {'train/loss': 0.009034913964569569, 'train/grad_norm_pre_clip': 11.32417106628418, 'timing/data': 0.00020916000357829034, 'timing/model': 0.39189891400747, 'learning_rate/qwen_vl_interface': 8.258951906065908e-06, 'learning_rate/action_model': 8.258951906065908e-05, 'epoch': 0.11}) +[2026-10-03 13:18:41,619][starVLA.training.train_starvla][INFO] - Step 1481, Loss: {'train/loss': 0.5247483253479004, 'train/grad_norm_pre_clip': 192.62294006347656, 'timing/data': 0.000242219990468584, 'timing/model': 0.3925026240031002, 'learning_rate/qwen_vl_interface': 8.256594824351577e-06, 'learning_rate/action_model': 8.256594824351576e-05, 'epoch': 0.11}) +[2026-10-03 13:18:42,325][starVLA.training.train_starvla][INFO] - Step 1482, Loss: {'train/loss': 0.02822823077440262, 'train/grad_norm_pre_clip': 52.255027770996094, 'timing/data': 0.00031536999449599534, 'timing/model': 0.39378514399868436, 'learning_rate/qwen_vl_interface': 8.254236506739683e-06, 'learning_rate/action_model': 8.254236506739681e-05, 'epoch': 0.11}) +[2026-10-03 13:18:43,030][starVLA.training.train_starvla][INFO] - Step 1483, Loss: {'train/loss': 0.0023564102593809366, 'train/grad_norm_pre_clip': 1.0630924701690674, 'timing/data': 0.0003134599974146113, 'timing/model': 0.39362166399951093, 'learning_rate/qwen_vl_interface': 8.25187695419964e-06, 'learning_rate/action_model': 8.25187695419964e-05, 'epoch': 0.11}) +[2026-10-03 13:18:43,952][starVLA.training.train_starvla][INFO] - Step 1484, Loss: {'train/loss': 0.27262282371520996, 'train/grad_norm_pre_clip': 39.871253967285156, 'timing/data': 0.00043069999082945287, 'timing/model': 0.3934642639942467, 'learning_rate/qwen_vl_interface': 8.249516167701373e-06, 'learning_rate/action_model': 8.249516167701374e-05, 'epoch': 0.11}) +[2026-10-03 13:18:44,655][starVLA.training.train_starvla][INFO] - Step 1485, Loss: {'train/loss': 0.018926195800304413, 'train/grad_norm_pre_clip': 34.08028030395508, 'timing/data': 0.0002633800031617284, 'timing/model': 0.3930672440037597, 'learning_rate/qwen_vl_interface': 8.247154148215311e-06, 'learning_rate/action_model': 8.24715414821531e-05, 'epoch': 0.11}) +[2026-10-03 13:18:45,362][starVLA.training.train_starvla][INFO] - Step 1486, Loss: {'train/loss': 4.507131961872801e-05, 'train/grad_norm_pre_clip': 0.00861463975161314, 'timing/data': 0.0002268699900014326, 'timing/model': 0.39500187399971765, 'learning_rate/qwen_vl_interface': 8.24479089671239e-06, 'learning_rate/action_model': 8.24479089671239e-05, 'epoch': 0.11}) +[2026-10-03 13:18:46,070][starVLA.training.train_starvla][INFO] - Step 1487, Loss: {'train/loss': 0.0014983094297349453, 'train/grad_norm_pre_clip': 0.5003827810287476, 'timing/data': 0.00028716999804601073, 'timing/model': 0.3948568339983467, 'learning_rate/qwen_vl_interface': 8.242426414164056e-06, 'learning_rate/action_model': 8.242426414164056e-05, 'epoch': 0.11}) +[2026-10-03 13:18:46,775][starVLA.training.train_starvla][INFO] - Step 1488, Loss: {'train/loss': 0.00040130072738975286, 'train/grad_norm_pre_clip': 0.2573765218257904, 'timing/data': 0.0002462900010868907, 'timing/model': 0.3936938539991388, 'learning_rate/qwen_vl_interface': 8.240060701542257e-06, 'learning_rate/action_model': 8.240060701542257e-05, 'epoch': 0.11}) +[2026-10-03 13:18:47,673][starVLA.training.train_starvla][INFO] - Step 1489, Loss: {'train/loss': 0.08416129648685455, 'train/grad_norm_pre_clip': 21.02941131591797, 'timing/data': 0.0002866700087906793, 'timing/model': 0.3935861840000143, 'learning_rate/qwen_vl_interface': 8.237693759819447e-06, 'learning_rate/action_model': 8.237693759819448e-05, 'epoch': 0.11}) +[2026-10-03 13:18:48,379][starVLA.training.train_starvla][INFO] - Step 1490, Loss: {'train/loss': 0.00024257382028736174, 'train/grad_norm_pre_clip': 0.07751330733299255, 'timing/data': 0.00034625000262167305, 'timing/model': 0.39450417400803417, 'learning_rate/qwen_vl_interface': 8.235325589968587e-06, 'learning_rate/action_model': 8.235325589968587e-05, 'epoch': 0.11}) +[2026-10-03 13:18:49,084][starVLA.training.train_starvla][INFO] - Step 1491, Loss: {'train/loss': 0.003007725812494755, 'train/grad_norm_pre_clip': 1.3622400760650635, 'timing/data': 0.00042953000229317695, 'timing/model': 0.3926801439956762, 'learning_rate/qwen_vl_interface': 8.232956192963143e-06, 'learning_rate/action_model': 8.232956192963144e-05, 'epoch': 0.11}) +[2026-10-03 13:18:49,789][starVLA.training.train_starvla][INFO] - Step 1492, Loss: {'train/loss': 1.0289057172485627e-05, 'train/grad_norm_pre_clip': 0.0028290918562561274, 'timing/data': 0.0003837200056295842, 'timing/model': 0.3931064049975248, 'learning_rate/qwen_vl_interface': 8.230585569777086e-06, 'learning_rate/action_model': 8.230585569777085e-05, 'epoch': 0.11}) +[2026-10-03 13:18:50,491][starVLA.training.train_starvla][INFO] - Step 1493, Loss: {'train/loss': 0.0008806255646049976, 'train/grad_norm_pre_clip': 0.5086928009986877, 'timing/data': 0.00030638001044280827, 'timing/model': 0.39220815400767606, 'learning_rate/qwen_vl_interface': 8.228213721384885e-06, 'learning_rate/action_model': 8.228213721384885e-05, 'epoch': 0.11}) +[2026-10-03 13:18:51,389][starVLA.training.train_starvla][INFO] - Step 1494, Loss: {'train/loss': 0.436465859413147, 'train/grad_norm_pre_clip': 35.766143798828125, 'timing/data': 0.00038304000918287784, 'timing/model': 0.5867226429982111, 'learning_rate/qwen_vl_interface': 8.225840648761523e-06, 'learning_rate/action_model': 8.225840648761523e-05, 'epoch': 0.11}) +[2026-10-03 13:18:52,092][starVLA.training.train_starvla][INFO] - Step 1495, Loss: {'train/loss': 0.013889061287045479, 'train/grad_norm_pre_clip': 52.60615158081055, 'timing/data': 0.0002488699974492192, 'timing/model': 0.3924738950008759, 'learning_rate/qwen_vl_interface': 8.223466352882476e-06, 'learning_rate/action_model': 8.223466352882475e-05, 'epoch': 0.11}) +[2026-10-03 13:18:52,798][starVLA.training.train_starvla][INFO] - Step 1496, Loss: {'train/loss': 0.00035163230495527387, 'train/grad_norm_pre_clip': 0.1677340567111969, 'timing/data': 0.00038853999285493046, 'timing/model': 0.39362280500063207, 'learning_rate/qwen_vl_interface': 8.22109083472373e-06, 'learning_rate/action_model': 8.22109083472373e-05, 'epoch': 0.11}) +[2026-10-03 13:18:53,501][starVLA.training.train_starvla][INFO] - Step 1497, Loss: {'train/loss': 2.296142884006258e-05, 'train/grad_norm_pre_clip': 0.01068657636642456, 'timing/data': 0.00023127999156713486, 'timing/model': 0.3930659749894403, 'learning_rate/qwen_vl_interface': 8.218714095261769e-06, 'learning_rate/action_model': 8.218714095261768e-05, 'epoch': 0.11}) +[2026-10-03 13:18:54,205][starVLA.training.train_starvla][INFO] - Step 1498, Loss: {'train/loss': 0.002186054363846779, 'train/grad_norm_pre_clip': 2.0117788314819336, 'timing/data': 0.00025567000557202846, 'timing/model': 0.39297953400819097, 'learning_rate/qwen_vl_interface': 8.21633613547358e-06, 'learning_rate/action_model': 8.21633613547358e-05, 'epoch': 0.11}) +[2026-10-03 13:18:54,908][starVLA.training.train_starvla][INFO] - Step 1499, Loss: {'train/loss': 0.00027833087369799614, 'train/grad_norm_pre_clip': 0.07666227221488953, 'timing/data': 0.0003484800108708441, 'timing/model': 0.3926945139974123, 'learning_rate/qwen_vl_interface': 8.213956956336656e-06, 'learning_rate/action_model': 8.213956956336656e-05, 'epoch': 0.11}) +[2026-10-03 13:18:55,804][starVLA.training.train_starvla][INFO] - Step 1500, Loss: {'train/loss': 0.3065466284751892, 'train/grad_norm_pre_clip': 32.38987350463867, 'timing/data': 0.00032123998971655965, 'timing/model': 0.3929672840022249, 'learning_rate/qwen_vl_interface': 8.211576558828985e-06, 'learning_rate/action_model': 8.211576558828985e-05, 'epoch': 0.11}) +[2026-10-03 13:18:56,507][starVLA.training.train_starvla][INFO] - Step 1501, Loss: {'train/loss': 0.00029711370007134974, 'train/grad_norm_pre_clip': 0.09844855964183807, 'timing/data': 0.00032423999800812453, 'timing/model': 0.392590604009456, 'learning_rate/qwen_vl_interface': 8.209194943929059e-06, 'learning_rate/action_model': 8.20919494392906e-05, 'epoch': 0.11}) +[2026-10-03 13:18:57,211][starVLA.training.train_starvla][INFO] - Step 1502, Loss: {'train/loss': 0.0437263660132885, 'train/grad_norm_pre_clip': 92.00250244140625, 'timing/data': 0.00037990001146681607, 'timing/model': 0.39266807500098366, 'learning_rate/qwen_vl_interface': 8.20681211261587e-06, 'learning_rate/action_model': 8.20681211261587e-05, 'epoch': 0.11}) +[2026-10-03 13:18:57,889][starVLA.training.train_starvla][INFO] - Step 1503, Loss: {'train/loss': 0.7192971110343933, 'train/grad_norm_pre_clip': 69.11612701416016, 'timing/data': 0.0002584400062914938, 'timing/model': 0.39269492399762385, 'learning_rate/qwen_vl_interface': 8.204428065868911e-06, 'learning_rate/action_model': 8.204428065868911e-05, 'epoch': 0.11}) +[2026-10-03 13:18:58,594][starVLA.training.train_starvla][INFO] - Step 1504, Loss: {'train/loss': 0.6761669516563416, 'train/grad_norm_pre_clip': 84.4157485961914, 'timing/data': 0.0002081400016322732, 'timing/model': 0.3931600040086778, 'learning_rate/qwen_vl_interface': 8.202042804668173e-06, 'learning_rate/action_model': 8.202042804668172e-05, 'epoch': 0.11}) +[2026-10-03 13:18:59,487][starVLA.training.train_starvla][INFO] - Step 1505, Loss: {'train/loss': 6.634049350395799e-05, 'train/grad_norm_pre_clip': 0.018276549875736237, 'timing/data': 0.00025146000552922487, 'timing/model': 0.5835202620073687, 'learning_rate/qwen_vl_interface': 8.199656329994145e-06, 'learning_rate/action_model': 8.199656329994145e-05, 'epoch': 0.11}) +[2026-10-03 13:19:00,191][starVLA.training.train_starvla][INFO] - Step 1506, Loss: {'train/loss': 0.1257598102092743, 'train/grad_norm_pre_clip': 87.48290252685547, 'timing/data': 0.00025024000206030905, 'timing/model': 0.39290608499140944, 'learning_rate/qwen_vl_interface': 8.19726864282782e-06, 'learning_rate/action_model': 8.197268642827818e-05, 'epoch': 0.11}) +[2026-10-03 13:19:00,895][starVLA.training.train_starvla][INFO] - Step 1507, Loss: {'train/loss': 0.07833472639322281, 'train/grad_norm_pre_clip': 88.36614990234375, 'timing/data': 0.00031854001281317323, 'timing/model': 0.39252447499893606, 'learning_rate/qwen_vl_interface': 8.194879744150682e-06, 'learning_rate/action_model': 8.194879744150681e-05, 'epoch': 0.11}) +[2026-10-03 13:19:01,600][starVLA.training.train_starvla][INFO] - Step 1508, Loss: {'train/loss': 0.3891012668609619, 'train/grad_norm_pre_clip': 101.88569641113281, 'timing/data': 0.00029718999576289207, 'timing/model': 0.3932379839970963, 'learning_rate/qwen_vl_interface': 8.192489634944718e-06, 'learning_rate/action_model': 8.192489634944718e-05, 'epoch': 0.11}) +[2026-10-03 13:19:02,303][starVLA.training.train_starvla][INFO] - Step 1509, Loss: {'train/loss': 0.002341249957680702, 'train/grad_norm_pre_clip': 1.497391700744629, 'timing/data': 0.0002526400057831779, 'timing/model': 0.3928246750001563, 'learning_rate/qwen_vl_interface': 8.190098316192416e-06, 'learning_rate/action_model': 8.190098316192416e-05, 'epoch': 0.11}) +[2026-10-03 13:19:03,198][starVLA.training.train_starvla][INFO] - Step 1510, Loss: {'train/loss': 0.12146655470132828, 'train/grad_norm_pre_clip': 172.7506561279297, 'timing/data': 0.0004045999958179891, 'timing/model': 0.5831304919993272, 'learning_rate/qwen_vl_interface': 8.18770578887675e-06, 'learning_rate/action_model': 8.18770578887675e-05, 'epoch': 0.11}) +[2026-10-03 13:19:03,901][starVLA.training.train_starvla][INFO] - Step 1511, Loss: {'train/loss': 0.0007532802410423756, 'train/grad_norm_pre_clip': 0.3643597960472107, 'timing/data': 0.00023682000755798072, 'timing/model': 0.3928036149882246, 'learning_rate/qwen_vl_interface': 8.185312053981204e-06, 'learning_rate/action_model': 8.185312053981204e-05, 'epoch': 0.11}) +[2026-10-03 13:19:04,604][starVLA.training.train_starvla][INFO] - Step 1512, Loss: {'train/loss': 0.3473328948020935, 'train/grad_norm_pre_clip': 33.29402542114258, 'timing/data': 0.00035327000659890473, 'timing/model': 0.39126707399555016, 'learning_rate/qwen_vl_interface': 8.182917112489747e-06, 'learning_rate/action_model': 8.182917112489746e-05, 'epoch': 0.11}) +[2026-10-03 13:19:05,307][starVLA.training.train_starvla][INFO] - Step 1513, Loss: {'train/loss': 0.00011487895244499668, 'train/grad_norm_pre_clip': 0.0675862729549408, 'timing/data': 0.0002513999934308231, 'timing/model': 0.3923830249987077, 'learning_rate/qwen_vl_interface': 8.18052096538685e-06, 'learning_rate/action_model': 8.18052096538685e-05, 'epoch': 0.11}) +[2026-10-03 13:19:06,010][starVLA.training.train_starvla][INFO] - Step 1514, Loss: {'train/loss': 5.326505197444931e-05, 'train/grad_norm_pre_clip': 0.014986081048846245, 'timing/data': 0.00030528000206686556, 'timing/model': 0.39322618499863893, 'learning_rate/qwen_vl_interface': 8.178123613657483e-06, 'learning_rate/action_model': 8.178123613657482e-05, 'epoch': 0.11}) +[2026-10-03 13:19:06,715][starVLA.training.train_starvla][INFO] - Step 1515, Loss: {'train/loss': 0.0009542300249449909, 'train/grad_norm_pre_clip': 0.25103744864463806, 'timing/data': 0.0003190199931850657, 'timing/model': 0.39354285500303376, 'learning_rate/qwen_vl_interface': 8.175725058287102e-06, 'learning_rate/action_model': 8.175725058287102e-05, 'epoch': 0.11}) +[2026-10-03 13:19:07,610][starVLA.training.train_starvla][INFO] - Step 1516, Loss: {'train/loss': 0.38952934741973877, 'train/grad_norm_pre_clip': 106.52017974853516, 'timing/data': 0.00026513000193517655, 'timing/model': 0.3927276640024502, 'learning_rate/qwen_vl_interface': 8.173325300261664e-06, 'learning_rate/action_model': 8.173325300261663e-05, 'epoch': 0.11}) +[2026-10-03 13:19:08,313][starVLA.training.train_starvla][INFO] - Step 1517, Loss: {'train/loss': 0.000297272956231609, 'train/grad_norm_pre_clip': 0.1163320317864418, 'timing/data': 0.00026674001128412783, 'timing/model': 0.3923146739980439, 'learning_rate/qwen_vl_interface': 8.170924340567617e-06, 'learning_rate/action_model': 8.170924340567616e-05, 'epoch': 0.11}) +[2026-10-03 13:19:09,018][starVLA.training.train_starvla][INFO] - Step 1518, Loss: {'train/loss': 0.001544550177641213, 'train/grad_norm_pre_clip': 0.7902273535728455, 'timing/data': 0.00034192000748589635, 'timing/model': 0.39303897399804555, 'learning_rate/qwen_vl_interface': 8.168522180191909e-06, 'learning_rate/action_model': 8.168522180191909e-05, 'epoch': 0.11}) +[2026-10-03 13:19:09,721][starVLA.training.train_starvla][INFO] - Step 1519, Loss: {'train/loss': 0.2590979337692261, 'train/grad_norm_pre_clip': 87.90304565429688, 'timing/data': 0.00022852999973110855, 'timing/model': 0.39303931400354486, 'learning_rate/qwen_vl_interface': 8.166118820121972e-06, 'learning_rate/action_model': 8.166118820121973e-05, 'epoch': 0.11}) +[2026-10-03 13:19:10,425][starVLA.training.train_starvla][INFO] - Step 1520, Loss: {'train/loss': 0.010098092257976532, 'train/grad_norm_pre_clip': 20.413930892944336, 'timing/data': 0.00036501999420579523, 'timing/model': 0.39272110399906524, 'learning_rate/qwen_vl_interface': 8.163714261345744e-06, 'learning_rate/action_model': 8.163714261345742e-05, 'epoch': 0.11}) +[2026-10-03 13:19:11,315][starVLA.training.train_starvla][INFO] - Step 1521, Loss: {'train/loss': 5.447395233204588e-05, 'train/grad_norm_pre_clip': 0.01130911335349083, 'timing/data': 0.0003221700026188046, 'timing/model': 0.5782727040059399, 'learning_rate/qwen_vl_interface': 8.16130850485164e-06, 'learning_rate/action_model': 8.161308504851639e-05, 'epoch': 0.11}) +[2026-10-03 13:19:12,018][starVLA.training.train_starvla][INFO] - Step 1522, Loss: {'train/loss': 0.00019635001081041992, 'train/grad_norm_pre_clip': 0.08001291006803513, 'timing/data': 0.00026006001280620694, 'timing/model': 0.392471724000643, 'learning_rate/qwen_vl_interface': 8.158901551628579e-06, 'learning_rate/action_model': 8.158901551628579e-05, 'epoch': 0.11}) +[2026-10-03 13:19:12,721][starVLA.training.train_starvla][INFO] - Step 1523, Loss: {'train/loss': 1.0523756742477417, 'train/grad_norm_pre_clip': 183.41062927246094, 'timing/data': 0.00023226000485010445, 'timing/model': 0.39313568398938514, 'learning_rate/qwen_vl_interface': 8.156493402665971e-06, 'learning_rate/action_model': 8.15649340266597e-05, 'epoch': 0.11}) +[2026-10-03 13:19:13,427][starVLA.training.train_starvla][INFO] - Step 1524, Loss: {'train/loss': 0.0030050333589315414, 'train/grad_norm_pre_clip': 0.8762182593345642, 'timing/data': 0.0003875600086757913, 'timing/model': 0.39337068499298766, 'learning_rate/qwen_vl_interface': 8.154084058953713e-06, 'learning_rate/action_model': 8.154084058953713e-05, 'epoch': 0.11}) +[2026-10-03 13:19:14,129][starVLA.training.train_starvla][INFO] - Step 1525, Loss: {'train/loss': 0.00018468123744241893, 'train/grad_norm_pre_clip': 0.11230301856994629, 'timing/data': 0.00027285999385640025, 'timing/model': 0.3924432950007031, 'learning_rate/qwen_vl_interface': 8.151673521482193e-06, 'learning_rate/action_model': 8.151673521482194e-05, 'epoch': 0.11}) +[2026-10-03 13:19:15,023][starVLA.training.train_starvla][INFO] - Step 1526, Loss: {'train/loss': 0.000714777794200927, 'train/grad_norm_pre_clip': 0.32704436779022217, 'timing/data': 0.00033139000879600644, 'timing/model': 0.5819869529950665, 'learning_rate/qwen_vl_interface': 8.149261791242297e-06, 'learning_rate/action_model': 8.149261791242297e-05, 'epoch': 0.11}) +[2026-10-03 13:19:15,726][starVLA.training.train_starvla][INFO] - Step 1527, Loss: {'train/loss': 0.0005015922943130136, 'train/grad_norm_pre_clip': 0.4655454754829407, 'timing/data': 0.00020029999723192304, 'timing/model': 0.39289563400961924, 'learning_rate/qwen_vl_interface': 8.146848869225391e-06, 'learning_rate/action_model': 8.146848869225391e-05, 'epoch': 0.11}) +[2026-10-03 13:19:16,430][starVLA.training.train_starvla][INFO] - Step 1528, Loss: {'train/loss': 0.0007720834692008793, 'train/grad_norm_pre_clip': 0.413638710975647, 'timing/data': 0.00030526999034918845, 'timing/model': 0.3930792840110371, 'learning_rate/qwen_vl_interface': 8.14443475642334e-06, 'learning_rate/action_model': 8.144434756423339e-05, 'epoch': 0.11}) +[2026-10-03 13:19:17,133][starVLA.training.train_starvla][INFO] - Step 1529, Loss: {'train/loss': 0.7013216018676758, 'train/grad_norm_pre_clip': 18.156766891479492, 'timing/data': 0.0003174299927195534, 'timing/model': 0.3923768640088383, 'learning_rate/qwen_vl_interface': 8.142019453828493e-06, 'learning_rate/action_model': 8.142019453828493e-05, 'epoch': 0.11}) +[2026-10-03 13:19:17,837][starVLA.training.train_starvla][INFO] - Step 1530, Loss: {'train/loss': 0.005148435942828655, 'train/grad_norm_pre_clip': 9.139955520629883, 'timing/data': 0.0003393600054550916, 'timing/model': 0.39289679400098976, 'learning_rate/qwen_vl_interface': 8.139602962433688e-06, 'learning_rate/action_model': 8.139602962433688e-05, 'epoch': 0.11}) +[2026-10-03 13:19:18,540][starVLA.training.train_starvla][INFO] - Step 1531, Loss: {'train/loss': 0.0015265537658706307, 'train/grad_norm_pre_clip': 1.6034588813781738, 'timing/data': 0.000251429999480024, 'timing/model': 0.3929836239985889, 'learning_rate/qwen_vl_interface': 8.137185283232255e-06, 'learning_rate/action_model': 8.137185283232256e-05, 'epoch': 0.11}) +[2026-10-03 13:19:19,434][starVLA.training.train_starvla][INFO] - Step 1532, Loss: {'train/loss': 0.00015348714077845216, 'train/grad_norm_pre_clip': 0.06252913922071457, 'timing/data': 0.00036111999361310154, 'timing/model': 0.3927175640128553, 'learning_rate/qwen_vl_interface': 8.134766417218014e-06, 'learning_rate/action_model': 8.134766417218012e-05, 'epoch': 0.11}) +[2026-10-03 13:19:20,137][starVLA.training.train_starvla][INFO] - Step 1533, Loss: {'train/loss': 0.47955968976020813, 'train/grad_norm_pre_clip': 32.19344711303711, 'timing/data': 0.0003139100008411333, 'timing/model': 0.3924614950083196, 'learning_rate/qwen_vl_interface': 8.132346365385262e-06, 'learning_rate/action_model': 8.132346365385261e-05, 'epoch': 0.11}) +[2026-10-03 13:19:20,841][starVLA.training.train_starvla][INFO] - Step 1534, Loss: {'train/loss': 0.18012842535972595, 'train/grad_norm_pre_clip': 88.39814758300781, 'timing/data': 0.0002053299976978451, 'timing/model': 0.39314727400778793, 'learning_rate/qwen_vl_interface': 8.129925128728796e-06, 'learning_rate/action_model': 8.129925128728796e-05, 'epoch': 0.11}) +[2026-10-03 13:19:21,545][starVLA.training.train_starvla][INFO] - Step 1535, Loss: {'train/loss': 0.0011935560032725334, 'train/grad_norm_pre_clip': 1.6790400743484497, 'timing/data': 0.0003024600009666756, 'timing/model': 0.39325405399722513, 'learning_rate/qwen_vl_interface': 8.127502708243895e-06, 'learning_rate/action_model': 8.127502708243895e-05, 'epoch': 0.11}) +[2026-10-03 13:19:22,250][starVLA.training.train_starvla][INFO] - Step 1536, Loss: {'train/loss': 0.10926613956689835, 'train/grad_norm_pre_clip': 147.51229858398438, 'timing/data': 0.00040409000939689577, 'timing/model': 0.3930702450015815, 'learning_rate/qwen_vl_interface': 8.125079104926324e-06, 'learning_rate/action_model': 8.125079104926323e-05, 'epoch': 0.11}) +[2026-10-03 13:19:23,145][starVLA.training.train_starvla][INFO] - Step 1537, Loss: {'train/loss': 0.06412195414304733, 'train/grad_norm_pre_clip': 158.0498809814453, 'timing/data': 0.00033129999064840376, 'timing/model': 0.392502593997051, 'learning_rate/qwen_vl_interface': 8.122654319772332e-06, 'learning_rate/action_model': 8.122654319772332e-05, 'epoch': 0.11}) +[2026-10-03 13:19:23,849][starVLA.training.train_starvla][INFO] - Step 1538, Loss: {'train/loss': 0.3436141908168793, 'train/grad_norm_pre_clip': 87.16561126708984, 'timing/data': 0.00028657000802922994, 'timing/model': 0.3929894150060136, 'learning_rate/qwen_vl_interface': 8.120228353778661e-06, 'learning_rate/action_model': 8.12022835377866e-05, 'epoch': 0.11}) +[2026-10-03 13:19:24,551][starVLA.training.train_starvla][INFO] - Step 1539, Loss: {'train/loss': 0.13945043087005615, 'train/grad_norm_pre_clip': 57.53856658935547, 'timing/data': 0.00029952998738735914, 'timing/model': 0.39249692400335334, 'learning_rate/qwen_vl_interface': 8.117801207942528e-06, 'learning_rate/action_model': 8.117801207942528e-05, 'epoch': 0.11}) +[2026-10-03 13:19:25,256][starVLA.training.train_starvla][INFO] - Step 1540, Loss: {'train/loss': 0.0013898346805945039, 'train/grad_norm_pre_clip': 0.2960958182811737, 'timing/data': 0.00039835000643506646, 'timing/model': 0.39330018499458674, 'learning_rate/qwen_vl_interface': 8.115372883261648e-06, 'learning_rate/action_model': 8.115372883261647e-05, 'epoch': 0.11}) +[2026-10-03 13:19:25,958][starVLA.training.train_starvla][INFO] - Step 1541, Loss: {'train/loss': 0.49413442611694336, 'train/grad_norm_pre_clip': 34.11149978637695, 'timing/data': 0.00024434000079054385, 'timing/model': 0.39231341500999406, 'learning_rate/qwen_vl_interface': 8.11294338073421e-06, 'learning_rate/action_model': 8.11294338073421e-05, 'epoch': 0.11}) +[2026-10-03 13:19:26,866][starVLA.training.train_starvla][INFO] - Step 1542, Loss: {'train/loss': 0.006679557263851166, 'train/grad_norm_pre_clip': 1.2528133392333984, 'timing/data': 0.0003199700004188344, 'timing/model': 0.5969192010088591, 'learning_rate/qwen_vl_interface': 8.110512701358892e-06, 'learning_rate/action_model': 8.110512701358891e-05, 'epoch': 0.11}) +[2026-10-03 13:19:27,576][starVLA.training.train_starvla][INFO] - Step 1543, Loss: {'train/loss': 0.0034410192165523767, 'train/grad_norm_pre_clip': 1.131466269493103, 'timing/data': 0.0004010299890069291, 'timing/model': 0.3965000140015036, 'learning_rate/qwen_vl_interface': 8.108080846134852e-06, 'learning_rate/action_model': 8.108080846134852e-05, 'epoch': 0.11}) +[2026-10-03 13:19:28,284][starVLA.training.train_starvla][INFO] - Step 1544, Loss: {'train/loss': 0.1841733753681183, 'train/grad_norm_pre_clip': 81.12317657470703, 'timing/data': 0.00021230000129435211, 'timing/model': 0.3953037739993306, 'learning_rate/qwen_vl_interface': 8.105647816061734e-06, 'learning_rate/action_model': 8.105647816061735e-05, 'epoch': 0.11}) +[2026-10-03 13:19:28,991][starVLA.training.train_starvla][INFO] - Step 1545, Loss: {'train/loss': 0.0017047192668542266, 'train/grad_norm_pre_clip': 0.31221726536750793, 'timing/data': 0.0002016100042965263, 'timing/model': 0.3946516639989568, 'learning_rate/qwen_vl_interface': 8.103213612139667e-06, 'learning_rate/action_model': 8.103213612139667e-05, 'epoch': 0.11}) +[2026-10-03 13:19:29,701][starVLA.training.train_starvla][INFO] - Step 1546, Loss: {'train/loss': 0.006402337457984686, 'train/grad_norm_pre_clip': 2.190764904022217, 'timing/data': 0.00020993000362068415, 'timing/model': 0.3965807950007729, 'learning_rate/qwen_vl_interface': 8.100778235369262e-06, 'learning_rate/action_model': 8.10077823536926e-05, 'epoch': 0.11}) +[2026-10-03 13:19:30,408][starVLA.training.train_starvla][INFO] - Step 1547, Loss: {'train/loss': 0.00021524590556509793, 'train/grad_norm_pre_clip': 0.038469284772872925, 'timing/data': 0.00027805000718217343, 'timing/model': 0.39435638500435743, 'learning_rate/qwen_vl_interface': 8.098341686751605e-06, 'learning_rate/action_model': 8.098341686751605e-05, 'epoch': 0.11}) +[2026-10-03 13:19:31,305][starVLA.training.train_starvla][INFO] - Step 1548, Loss: {'train/loss': 4.934157186653465e-05, 'train/grad_norm_pre_clip': 0.00938389077782631, 'timing/data': 0.00033445999724790454, 'timing/model': 0.3932647750043543, 'learning_rate/qwen_vl_interface': 8.095903967288273e-06, 'learning_rate/action_model': 8.095903967288272e-05, 'epoch': 0.11}) +[2026-10-03 13:19:32,008][starVLA.training.train_starvla][INFO] - Step 1549, Loss: {'train/loss': 0.001028604106977582, 'train/grad_norm_pre_clip': 1.6402581930160522, 'timing/data': 0.00025179999647662044, 'timing/model': 0.3929460239887703, 'learning_rate/qwen_vl_interface': 8.093465077981318e-06, 'learning_rate/action_model': 8.093465077981318e-05, 'epoch': 0.11}) +[2026-10-03 13:19:32,715][starVLA.training.train_starvla][INFO] - Step 1550, Loss: {'train/loss': 1.6420581340789795, 'train/grad_norm_pre_clip': 102.1865005493164, 'timing/data': 0.0002074490039376542, 'timing/model': 0.39513802499277517, 'learning_rate/qwen_vl_interface': 8.091025019833276e-06, 'learning_rate/action_model': 8.091025019833276e-05, 'epoch': 0.11}) +[2026-10-03 13:19:33,423][starVLA.training.train_starvla][INFO] - Step 1551, Loss: {'train/loss': 0.7574222683906555, 'train/grad_norm_pre_clip': 129.694580078125, 'timing/data': 0.0002222800103481859, 'timing/model': 0.39539150400378276, 'learning_rate/qwen_vl_interface': 8.088583793847165e-06, 'learning_rate/action_model': 8.088583793847164e-05, 'epoch': 0.11}) +[2026-10-03 13:19:34,128][starVLA.training.train_starvla][INFO] - Step 1552, Loss: {'train/loss': 0.10380950570106506, 'train/grad_norm_pre_clip': 41.7977180480957, 'timing/data': 0.0002653300034580752, 'timing/model': 0.3933257449971279, 'learning_rate/qwen_vl_interface': 8.086141401026478e-06, 'learning_rate/action_model': 8.086141401026478e-05, 'epoch': 0.11}) +[2026-10-03 13:19:35,024][starVLA.training.train_starvla][INFO] - Step 1553, Loss: {'train/loss': 1.6807785868877545e-05, 'train/grad_norm_pre_clip': 0.005690538324415684, 'timing/data': 0.0003464500041445717, 'timing/model': 0.39345853400300257, 'learning_rate/qwen_vl_interface': 8.083697842375193e-06, 'learning_rate/action_model': 8.083697842375193e-05, 'epoch': 0.11}) +[2026-10-03 13:19:35,732][starVLA.training.train_starvla][INFO] - Step 1554, Loss: {'train/loss': 3.1349507480626926e-05, 'train/grad_norm_pre_clip': 0.005604481790214777, 'timing/data': 0.00028833998658228666, 'timing/model': 0.3956260539998766, 'learning_rate/qwen_vl_interface': 8.081253118897762e-06, 'learning_rate/action_model': 8.081253118897761e-05, 'epoch': 0.11}) +[2026-10-03 13:19:36,440][starVLA.training.train_starvla][INFO] - Step 1555, Loss: {'train/loss': 0.21662181615829468, 'train/grad_norm_pre_clip': 151.3668975830078, 'timing/data': 0.00022137998894322664, 'timing/model': 0.3952177749888506, 'learning_rate/qwen_vl_interface': 8.07880723159912e-06, 'learning_rate/action_model': 8.07880723159912e-05, 'epoch': 0.11}) +[2026-10-03 13:19:37,145][starVLA.training.train_starvla][INFO] - Step 1556, Loss: {'train/loss': 0.00010403118358226493, 'train/grad_norm_pre_clip': 0.0823553279042244, 'timing/data': 0.0002682000049389899, 'timing/model': 0.39331061400298495, 'learning_rate/qwen_vl_interface': 8.076360181484676e-06, 'learning_rate/action_model': 8.076360181484676e-05, 'epoch': 0.11}) +[2026-10-03 13:19:37,848][starVLA.training.train_starvla][INFO] - Step 1557, Loss: {'train/loss': 0.0003421037981752306, 'train/grad_norm_pre_clip': 0.15335384011268616, 'timing/data': 0.0002925099979620427, 'timing/model': 0.39293533499585465, 'learning_rate/qwen_vl_interface': 8.073911969560326e-06, 'learning_rate/action_model': 8.073911969560326e-05, 'epoch': 0.11}) +[2026-10-03 13:19:38,749][starVLA.training.train_starvla][INFO] - Step 1558, Loss: {'train/loss': 0.00023390166461467743, 'train/grad_norm_pre_clip': 0.04695522040128708, 'timing/data': 0.00033303999225609004, 'timing/model': 0.5885484619939234, 'learning_rate/qwen_vl_interface': 8.071462596832432e-06, 'learning_rate/action_model': 8.071462596832432e-05, 'epoch': 0.11}) +[2026-10-03 13:19:39,455][starVLA.training.train_starvla][INFO] - Step 1559, Loss: {'train/loss': 1.2901355028152466, 'train/grad_norm_pre_clip': 79.40576934814453, 'timing/data': 0.0002738900075200945, 'timing/model': 0.3945389939908637, 'learning_rate/qwen_vl_interface': 8.06901206430784e-06, 'learning_rate/action_model': 8.06901206430784e-05, 'epoch': 0.11}) +[2026-10-03 13:19:40,160][starVLA.training.train_starvla][INFO] - Step 1560, Loss: {'train/loss': 4.005747177870944e-05, 'train/grad_norm_pre_clip': 0.009001895785331726, 'timing/data': 0.0002037899976130575, 'timing/model': 0.3936074649973307, 'learning_rate/qwen_vl_interface': 8.066560372993873e-06, 'learning_rate/action_model': 8.066560372993874e-05, 'epoch': 0.11}) +[2026-10-03 13:19:40,863][starVLA.training.train_starvla][INFO] - Step 1561, Loss: {'train/loss': 0.8297903537750244, 'train/grad_norm_pre_clip': 112.5093994140625, 'timing/data': 0.00031362999288830906, 'timing/model': 0.39296199499221984, 'learning_rate/qwen_vl_interface': 8.064107523898327e-06, 'learning_rate/action_model': 8.064107523898326e-05, 'epoch': 0.11}) +[2026-10-03 13:19:41,570][starVLA.training.train_starvla][INFO] - Step 1562, Loss: {'train/loss': 0.036191198974847794, 'train/grad_norm_pre_clip': 97.95747375488281, 'timing/data': 0.0003224500105716288, 'timing/model': 0.39475021400721744, 'learning_rate/qwen_vl_interface': 8.061653518029476e-06, 'learning_rate/action_model': 8.061653518029476e-05, 'epoch': 0.11}) +[2026-10-03 13:19:42,471][starVLA.training.train_starvla][INFO] - Step 1563, Loss: {'train/loss': 0.0009915306000038981, 'train/grad_norm_pre_clip': 0.2751910090446472, 'timing/data': 0.00036080999416299164, 'timing/model': 0.5887787120009307, 'learning_rate/qwen_vl_interface': 8.059198356396069e-06, 'learning_rate/action_model': 8.059198356396068e-05, 'epoch': 0.11}) +[2026-10-03 13:19:43,175][starVLA.training.train_starvla][INFO] - Step 1564, Loss: {'train/loss': 0.046754997223615646, 'train/grad_norm_pre_clip': 88.15065002441406, 'timing/data': 0.000307460009935312, 'timing/model': 0.3931661739916308, 'learning_rate/qwen_vl_interface': 8.05674204000733e-06, 'learning_rate/action_model': 8.05674204000733e-05, 'epoch': 0.11}) +[2026-10-03 13:19:43,878][starVLA.training.train_starvla][INFO] - Step 1565, Loss: {'train/loss': 0.27392998337745667, 'train/grad_norm_pre_clip': 34.820098876953125, 'timing/data': 0.00036202999763190746, 'timing/model': 0.3927510139910737, 'learning_rate/qwen_vl_interface': 8.05428456987296e-06, 'learning_rate/action_model': 8.05428456987296e-05, 'epoch': 0.11}) +[2026-10-03 13:19:44,584][starVLA.training.train_starvla][INFO] - Step 1566, Loss: {'train/loss': 0.009582950733602047, 'train/grad_norm_pre_clip': 17.76435089111328, 'timing/data': 0.0002994499955093488, 'timing/model': 0.39505802499479614, 'learning_rate/qwen_vl_interface': 8.05182594700313e-06, 'learning_rate/action_model': 8.05182594700313e-05, 'epoch': 0.11}) +[2026-10-03 13:19:45,292][starVLA.training.train_starvla][INFO] - Step 1567, Loss: {'train/loss': 0.24680979549884796, 'train/grad_norm_pre_clip': 27.132728576660156, 'timing/data': 0.0002559600106906146, 'timing/model': 0.3948174149991246, 'learning_rate/qwen_vl_interface': 8.049366172408484e-06, 'learning_rate/action_model': 8.049366172408485e-05, 'epoch': 0.11}) +[2026-10-03 13:19:46,192][starVLA.training.train_starvla][INFO] - Step 1568, Loss: {'train/loss': 0.43717458844184875, 'train/grad_norm_pre_clip': 52.81089401245117, 'timing/data': 0.00034528000105638057, 'timing/model': 0.5890519919921644, 'learning_rate/qwen_vl_interface': 8.04690524710015e-06, 'learning_rate/action_model': 8.046905247100151e-05, 'epoch': 0.11}) +[2026-10-03 13:19:46,897][starVLA.training.train_starvla][INFO] - Step 1569, Loss: {'train/loss': 0.0041943020187318325, 'train/grad_norm_pre_clip': 1.5030299425125122, 'timing/data': 0.000335150005412288, 'timing/model': 0.39376904400705826, 'learning_rate/qwen_vl_interface': 8.044443172089715e-06, 'learning_rate/action_model': 8.044443172089714e-05, 'epoch': 0.11}) +[2026-10-03 13:19:47,603][starVLA.training.train_starvla][INFO] - Step 1570, Loss: {'train/loss': 4.261632057023235e-06, 'train/grad_norm_pre_clip': 0.0041778720915317535, 'timing/data': 0.00030359000083990395, 'timing/model': 0.39474332399549894, 'learning_rate/qwen_vl_interface': 8.041979948389248e-06, 'learning_rate/action_model': 8.041979948389248e-05, 'epoch': 0.11}) +[2026-10-03 13:19:48,311][starVLA.training.train_starvla][INFO] - Step 1571, Loss: {'train/loss': 0.02014138363301754, 'train/grad_norm_pre_clip': 23.542009353637695, 'timing/data': 0.00031596999906469136, 'timing/model': 0.3951091040071333, 'learning_rate/qwen_vl_interface': 8.039515577011287e-06, 'learning_rate/action_model': 8.039515577011288e-05, 'epoch': 0.11}) +[2026-10-03 13:19:49,018][starVLA.training.train_starvla][INFO] - Step 1572, Loss: {'train/loss': 0.03694121167063713, 'train/grad_norm_pre_clip': 17.728199005126953, 'timing/data': 0.000331000002915971, 'timing/model': 0.3941482549998909, 'learning_rate/qwen_vl_interface': 8.037050058968841e-06, 'learning_rate/action_model': 8.03705005896884e-05, 'epoch': 0.12}) +[2026-10-03 13:19:49,722][starVLA.training.train_starvla][INFO] - Step 1573, Loss: {'train/loss': 0.013539677485823631, 'train/grad_norm_pre_clip': 5.667154788970947, 'timing/data': 0.00026890001026913524, 'timing/model': 0.39338167500682175, 'learning_rate/qwen_vl_interface': 8.034583395275392e-06, 'learning_rate/action_model': 8.034583395275392e-05, 'epoch': 0.12}) +[2026-10-03 13:19:50,622][starVLA.training.train_starvla][INFO] - Step 1574, Loss: {'train/loss': 1.1883835792541504, 'train/grad_norm_pre_clip': 135.89891052246094, 'timing/data': 0.00027368000883143395, 'timing/model': 0.3951292050041957, 'learning_rate/qwen_vl_interface': 8.032115586944893e-06, 'learning_rate/action_model': 8.032115586944892e-05, 'epoch': 0.12}) +[2026-10-03 13:19:51,329][starVLA.training.train_starvla][INFO] - Step 1575, Loss: {'train/loss': 4.2468275296414504e-07, 'train/grad_norm_pre_clip': 0.00017696863505989313, 'timing/data': 0.0003280600067228079, 'timing/model': 0.3949185540113831, 'learning_rate/qwen_vl_interface': 8.029646634991762e-06, 'learning_rate/action_model': 8.029646634991762e-05, 'epoch': 0.12}) +[2026-10-03 13:19:52,034][starVLA.training.train_starvla][INFO] - Step 1576, Loss: {'train/loss': 0.11219222098588943, 'train/grad_norm_pre_clip': 36.78520202636719, 'timing/data': 0.00045140000293031335, 'timing/model': 0.39344133499253076, 'learning_rate/qwen_vl_interface': 8.0271765404309e-06, 'learning_rate/action_model': 8.0271765404309e-05, 'epoch': 0.12}) +[2026-10-03 13:19:52,738][starVLA.training.train_starvla][INFO] - Step 1577, Loss: {'train/loss': 0.23381200432777405, 'train/grad_norm_pre_clip': 62.71059799194336, 'timing/data': 0.00033411999174859375, 'timing/model': 0.39361774499411695, 'learning_rate/qwen_vl_interface': 8.024705304277664e-06, 'learning_rate/action_model': 8.024705304277664e-05, 'epoch': 0.12}) +[2026-10-03 13:19:53,445][starVLA.training.train_starvla][INFO] - Step 1578, Loss: {'train/loss': 0.4728052020072937, 'train/grad_norm_pre_clip': 63.26976013183594, 'timing/data': 0.0003234999894630164, 'timing/model': 0.3956557040073676, 'learning_rate/qwen_vl_interface': 8.022232927547888e-06, 'learning_rate/action_model': 8.022232927547886e-05, 'epoch': 0.12}) +[2026-10-03 13:19:54,346][starVLA.training.train_starvla][INFO] - Step 1579, Loss: {'train/loss': 0.01973259076476097, 'train/grad_norm_pre_clip': 11.897125244140625, 'timing/data': 0.00027804001001641154, 'timing/model': 0.5874914820014965, 'learning_rate/qwen_vl_interface': 8.01975941125787e-06, 'learning_rate/action_model': 8.01975941125787e-05, 'epoch': 0.12}) +[2026-10-03 13:19:55,053][starVLA.training.train_starvla][INFO] - Step 1580, Loss: {'train/loss': 0.014498747885227203, 'train/grad_norm_pre_clip': 7.817039489746094, 'timing/data': 0.00033249899570364505, 'timing/model': 0.395196423996822, 'learning_rate/qwen_vl_interface': 8.017284756424381e-06, 'learning_rate/action_model': 8.017284756424381e-05, 'epoch': 0.12}) +[2026-10-03 13:19:55,759][starVLA.training.train_starvla][INFO] - Step 1581, Loss: {'train/loss': 0.0018741488456726074, 'train/grad_norm_pre_clip': 0.5735549926757812, 'timing/data': 0.00026200999855063856, 'timing/model': 0.39482520399906207, 'learning_rate/qwen_vl_interface': 8.01480896406466e-06, 'learning_rate/action_model': 8.01480896406466e-05, 'epoch': 0.12}) +[2026-10-03 13:19:56,469][starVLA.training.train_starvla][INFO] - Step 1582, Loss: {'train/loss': 0.0009662087541073561, 'train/grad_norm_pre_clip': 0.21609219908714294, 'timing/data': 0.00030169999809004366, 'timing/model': 0.39584584499243647, 'learning_rate/qwen_vl_interface': 8.01233203519641e-06, 'learning_rate/action_model': 8.012332035196409e-05, 'epoch': 0.12}) +[2026-10-03 13:19:57,179][starVLA.training.train_starvla][INFO] - Step 1583, Loss: {'train/loss': 0.2796686291694641, 'train/grad_norm_pre_clip': 53.433433532714844, 'timing/data': 0.00030947999039199203, 'timing/model': 0.39595394500065595, 'learning_rate/qwen_vl_interface': 8.009853970837801e-06, 'learning_rate/action_model': 8.009853970837801e-05, 'epoch': 0.12}) +[2026-10-03 13:19:57,885][starVLA.training.train_starvla][INFO] - Step 1584, Loss: {'train/loss': 0.00040196083136834204, 'train/grad_norm_pre_clip': 0.07796625792980194, 'timing/data': 0.0002708000101847574, 'timing/model': 0.3947482340008719, 'learning_rate/qwen_vl_interface': 8.007374772007474e-06, 'learning_rate/action_model': 8.007374772007474e-05, 'epoch': 0.12}) +[2026-10-03 13:19:58,787][starVLA.training.train_starvla][INFO] - Step 1585, Loss: {'train/loss': 1.005064677883638e-05, 'train/grad_norm_pre_clip': 0.0061065759509801865, 'timing/data': 0.0003568399988580495, 'timing/model': 0.3949403339938726, 'learning_rate/qwen_vl_interface': 8.004894439724533e-06, 'learning_rate/action_model': 8.004894439724533e-05, 'epoch': 0.12}) +[2026-10-03 13:19:59,497][starVLA.training.train_starvla][INFO] - Step 1586, Loss: {'train/loss': 0.0003705275012180209, 'train/grad_norm_pre_clip': 0.09184551239013672, 'timing/data': 0.0003701899986481294, 'timing/model': 0.39621688399347477, 'learning_rate/qwen_vl_interface': 8.002412975008552e-06, 'learning_rate/action_model': 8.002412975008552e-05, 'epoch': 0.12}) +[2026-10-03 13:20:00,208][starVLA.training.train_starvla][INFO] - Step 1587, Loss: {'train/loss': 0.06767433881759644, 'train/grad_norm_pre_clip': 48.39119338989258, 'timing/data': 0.00032844000088516623, 'timing/model': 0.39622435500496067, 'learning_rate/qwen_vl_interface': 7.999930378879562e-06, 'learning_rate/action_model': 7.999930378879561e-05, 'epoch': 0.12}) +[2026-10-03 13:20:00,918][starVLA.training.train_starvla][INFO] - Step 1588, Loss: {'train/loss': 0.05008352920413017, 'train/grad_norm_pre_clip': 28.015676498413086, 'timing/data': 0.00025973000447265804, 'timing/model': 0.3947679650009377, 'learning_rate/qwen_vl_interface': 7.997446652358067e-06, 'learning_rate/action_model': 7.997446652358068e-05, 'epoch': 0.12}) +[2026-10-03 13:20:01,625][starVLA.training.train_starvla][INFO] - Step 1589, Loss: {'train/loss': 0.00041741091990843415, 'train/grad_norm_pre_clip': 0.12265953421592712, 'timing/data': 0.000294400000711903, 'timing/model': 0.3947162440017564, 'learning_rate/qwen_vl_interface': 7.994961796465035e-06, 'learning_rate/action_model': 7.994961796465035e-05, 'epoch': 0.12}) +[2026-10-03 13:20:02,529][starVLA.training.train_starvla][INFO] - Step 1590, Loss: {'train/loss': 3.8425965612987056e-05, 'train/grad_norm_pre_clip': 0.015275509096682072, 'timing/data': 0.00030789901211392134, 'timing/model': 0.3957461149984738, 'learning_rate/qwen_vl_interface': 7.992475812221894e-06, 'learning_rate/action_model': 7.992475812221894e-05, 'epoch': 0.12}) +[2026-10-03 13:20:03,239][starVLA.training.train_starvla][INFO] - Step 1591, Loss: {'train/loss': 0.47641777992248535, 'train/grad_norm_pre_clip': 149.25137329101562, 'timing/data': 0.0003415900137042627, 'timing/model': 0.39637554400542285, 'learning_rate/qwen_vl_interface': 7.989988700650538e-06, 'learning_rate/action_model': 7.989988700650538e-05, 'epoch': 0.12}) +[2026-10-03 13:20:03,944][starVLA.training.train_starvla][INFO] - Step 1592, Loss: {'train/loss': 0.2805927097797394, 'train/grad_norm_pre_clip': 71.26456451416016, 'timing/data': 0.0002912500058300793, 'timing/model': 0.3930900749983266, 'learning_rate/qwen_vl_interface': 7.987500462773326e-06, 'learning_rate/action_model': 7.987500462773326e-05, 'epoch': 0.12}) +[2026-10-03 13:20:04,647][starVLA.training.train_starvla][INFO] - Step 1593, Loss: {'train/loss': 5.590432556346059e-05, 'train/grad_norm_pre_clip': 0.024566320702433586, 'timing/data': 0.00030609000532422215, 'timing/model': 0.3925889149977593, 'learning_rate/qwen_vl_interface': 7.985011099613078e-06, 'learning_rate/action_model': 7.985011099613077e-05, 'epoch': 0.12}) +[2026-10-03 13:20:05,353][starVLA.training.train_starvla][INFO] - Step 1594, Loss: {'train/loss': 0.0009549550013616681, 'train/grad_norm_pre_clip': 0.27867886424064636, 'timing/data': 0.0002748700062511489, 'timing/model': 0.3948106239986373, 'learning_rate/qwen_vl_interface': 7.98252061219308e-06, 'learning_rate/action_model': 7.982520612193078e-05, 'epoch': 0.12}) +[2026-10-03 13:20:06,255][starVLA.training.train_starvla][INFO] - Step 1595, Loss: {'train/loss': 0.004578378051519394, 'train/grad_norm_pre_clip': 4.0417585372924805, 'timing/data': 0.00021995000133756548, 'timing/model': 0.591576870996505, 'learning_rate/qwen_vl_interface': 7.980029001537073e-06, 'learning_rate/action_model': 7.980029001537072e-05, 'epoch': 0.12}) +[2026-10-03 13:20:06,960][starVLA.training.train_starvla][INFO] - Step 1596, Loss: {'train/loss': 0.00013667347957380116, 'train/grad_norm_pre_clip': 0.1754770129919052, 'timing/data': 0.00029804000223521143, 'timing/model': 0.3935199950064998, 'learning_rate/qwen_vl_interface': 7.977536268669264e-06, 'learning_rate/action_model': 7.977536268669264e-05, 'epoch': 0.12}) +[2026-10-03 13:20:07,664][starVLA.training.train_starvla][INFO] - Step 1597, Loss: {'train/loss': 0.00013514337479136884, 'train/grad_norm_pre_clip': 0.0813080295920372, 'timing/data': 0.00025935999292414635, 'timing/model': 0.3935207150061615, 'learning_rate/qwen_vl_interface': 7.975042414614324e-06, 'learning_rate/action_model': 7.975042414614324e-05, 'epoch': 0.12}) +[2026-10-03 13:20:08,384][starVLA.training.train_starvla][INFO] - Step 1598, Loss: {'train/loss': 0.005718662869185209, 'train/grad_norm_pre_clip': 10.24888801574707, 'timing/data': 0.0003008600033354014, 'timing/model': 0.40827658300986513, 'learning_rate/qwen_vl_interface': 7.972547440397382e-06, 'learning_rate/action_model': 7.972547440397383e-05, 'epoch': 0.12}) +[2026-10-03 13:20:09,092][starVLA.training.train_starvla][INFO] - Step 1599, Loss: {'train/loss': 0.6980732083320618, 'train/grad_norm_pre_clip': 76.06818389892578, 'timing/data': 0.0003577399911591783, 'timing/model': 0.3951262949994998, 'learning_rate/qwen_vl_interface': 7.970051347044026e-06, 'learning_rate/action_model': 7.970051347044025e-05, 'epoch': 0.12}) +[2026-10-03 13:20:09,799][starVLA.training.train_starvla][INFO] - Step 1600, Loss: {'train/loss': 0.12994897365570068, 'train/grad_norm_pre_clip': 53.90629196166992, 'timing/data': 0.0003320800024084747, 'timing/model': 0.3937019040022278, 'learning_rate/qwen_vl_interface': 7.967554135580307e-06, 'learning_rate/action_model': 7.967554135580307e-05, 'epoch': 0.12}) +[2026-10-03 13:20:10,699][starVLA.training.train_starvla][INFO] - Step 1601, Loss: {'train/loss': 0.13518208265304565, 'train/grad_norm_pre_clip': 47.329978942871094, 'timing/data': 0.00032300000020768493, 'timing/model': 0.39465809499961324, 'learning_rate/qwen_vl_interface': 7.965055807032734e-06, 'learning_rate/action_model': 7.965055807032735e-05, 'epoch': 0.12}) +[2026-10-03 13:20:11,418][starVLA.training.train_starvla][INFO] - Step 1602, Loss: {'train/loss': 0.49236705899238586, 'train/grad_norm_pre_clip': 72.66218566894531, 'timing/data': 0.00022930999693926424, 'timing/model': 0.4086881330003962, 'learning_rate/qwen_vl_interface': 7.962556362428275e-06, 'learning_rate/action_model': 7.962556362428274e-05, 'epoch': 0.12}) +[2026-10-03 13:20:12,125][starVLA.training.train_starvla][INFO] - Step 1603, Loss: {'train/loss': 0.0002285814261995256, 'train/grad_norm_pre_clip': 0.0654912069439888, 'timing/data': 0.0002180500014219433, 'timing/model': 0.3946988940006122, 'learning_rate/qwen_vl_interface': 7.960055802794358e-06, 'learning_rate/action_model': 7.960055802794357e-05, 'epoch': 0.12}) +[2026-10-03 13:20:12,830][starVLA.training.train_starvla][INFO] - Step 1604, Loss: {'train/loss': 0.04660932719707489, 'train/grad_norm_pre_clip': 75.48347473144531, 'timing/data': 0.00032839000050444156, 'timing/model': 0.3934284549904987, 'learning_rate/qwen_vl_interface': 7.957554129158866e-06, 'learning_rate/action_model': 7.957554129158866e-05, 'epoch': 0.12}) +[2026-10-03 13:20:13,535][starVLA.training.train_starvla][INFO] - Step 1605, Loss: {'train/loss': 0.27427223324775696, 'train/grad_norm_pre_clip': 108.18374633789062, 'timing/data': 0.00030478999542538077, 'timing/model': 0.39368369498697575, 'learning_rate/qwen_vl_interface': 7.955051342550148e-06, 'learning_rate/action_model': 7.955051342550148e-05, 'epoch': 0.12}) +[2026-10-03 13:20:14,438][starVLA.training.train_starvla][INFO] - Step 1606, Loss: {'train/loss': 0.6439652442932129, 'train/grad_norm_pre_clip': 33.88417053222656, 'timing/data': 0.000268470001174137, 'timing/model': 0.3935431650024839, 'learning_rate/qwen_vl_interface': 7.952547443997e-06, 'learning_rate/action_model': 7.952547443997e-05, 'epoch': 0.12}) +[2026-10-03 13:20:15,146][starVLA.training.train_starvla][INFO] - Step 1607, Loss: {'train/loss': 0.00021343831031117588, 'train/grad_norm_pre_clip': 0.05280322954058647, 'timing/data': 0.00035755999851971865, 'timing/model': 0.39497304399264976, 'learning_rate/qwen_vl_interface': 7.950042434528681e-06, 'learning_rate/action_model': 7.950042434528682e-05, 'epoch': 0.12}) +[2026-10-03 13:20:15,851][starVLA.training.train_starvla][INFO] - Step 1608, Loss: {'train/loss': 0.004395033698529005, 'train/grad_norm_pre_clip': 9.889177322387695, 'timing/data': 0.0003605700039770454, 'timing/model': 0.3937371840002015, 'learning_rate/qwen_vl_interface': 7.94753631517491e-06, 'learning_rate/action_model': 7.947536315174909e-05, 'epoch': 0.12}) +[2026-10-03 13:20:16,556][starVLA.training.train_starvla][INFO] - Step 1609, Loss: {'train/loss': 0.3272605240345001, 'train/grad_norm_pre_clip': 38.444908142089844, 'timing/data': 0.000208480007131584, 'timing/model': 0.3935047450067941, 'learning_rate/qwen_vl_interface': 7.945029086965854e-06, 'learning_rate/action_model': 7.945029086965853e-05, 'epoch': 0.12}) +[2026-10-03 13:20:17,262][starVLA.training.train_starvla][INFO] - Step 1610, Loss: {'train/loss': 0.2385566234588623, 'train/grad_norm_pre_clip': 124.30773162841797, 'timing/data': 0.00020642999152187258, 'timing/model': 0.3949060840095626, 'learning_rate/qwen_vl_interface': 7.942520750932138e-06, 'learning_rate/action_model': 7.942520750932138e-05, 'epoch': 0.12}) +[2026-10-03 13:20:18,169][starVLA.training.train_starvla][INFO] - Step 1611, Loss: {'train/loss': 0.00045179110020399094, 'train/grad_norm_pre_clip': 1.0140703916549683, 'timing/data': 0.000320930004818365, 'timing/model': 0.39620700500381645, 'learning_rate/qwen_vl_interface': 7.940011308104848e-06, 'learning_rate/action_model': 7.940011308104848e-05, 'epoch': 0.12}) +[2026-10-03 13:20:18,877][starVLA.training.train_starvla][INFO] - Step 1612, Loss: {'train/loss': 0.6575640439987183, 'train/grad_norm_pre_clip': 64.47486877441406, 'timing/data': 0.0003755490033654496, 'timing/model': 0.39500586499343626, 'learning_rate/qwen_vl_interface': 7.93750075951552e-06, 'learning_rate/action_model': 7.93750075951552e-05, 'epoch': 0.12}) +[2026-10-03 13:20:19,583][starVLA.training.train_starvla][INFO] - Step 1613, Loss: {'train/loss': 0.003885007929056883, 'train/grad_norm_pre_clip': 4.1022419929504395, 'timing/data': 0.00031319999834522605, 'timing/model': 0.3946317140071187, 'learning_rate/qwen_vl_interface': 7.934989106196145e-06, 'learning_rate/action_model': 7.934989106196146e-05, 'epoch': 0.12}) +[2026-10-03 13:20:20,292][starVLA.training.train_starvla][INFO] - Step 1614, Loss: {'train/loss': 0.0020037514623254538, 'train/grad_norm_pre_clip': 2.471398115158081, 'timing/data': 0.0002445399877615273, 'timing/model': 0.39596224400156643, 'learning_rate/qwen_vl_interface': 7.93247634917917e-06, 'learning_rate/action_model': 7.93247634917917e-05, 'epoch': 0.12}) +[2026-10-03 13:20:20,999][starVLA.training.train_starvla][INFO] - Step 1615, Loss: {'train/loss': 0.07457978278398514, 'train/grad_norm_pre_clip': 39.44145965576172, 'timing/data': 0.0003507100045681, 'timing/model': 0.3947652439965168, 'learning_rate/qwen_vl_interface': 7.929962489497494e-06, 'learning_rate/action_model': 7.929962489497493e-05, 'epoch': 0.12}) +[2026-10-03 13:20:21,902][starVLA.training.train_starvla][INFO] - Step 1616, Loss: {'train/loss': 0.0005755784222856164, 'train/grad_norm_pre_clip': 0.3394296169281006, 'timing/data': 0.00020838000637013465, 'timing/model': 0.5931892819935456, 'learning_rate/qwen_vl_interface': 7.927447528184469e-06, 'learning_rate/action_model': 7.927447528184468e-05, 'epoch': 0.12}) +[2026-10-03 13:20:22,606][starVLA.training.train_starvla][INFO] - Step 1617, Loss: {'train/loss': 1.3541338443756104, 'train/grad_norm_pre_clip': 57.28192901611328, 'timing/data': 0.00027106000925414264, 'timing/model': 0.39354430499952286, 'learning_rate/qwen_vl_interface': 7.924931466273901e-06, 'learning_rate/action_model': 7.924931466273902e-05, 'epoch': 0.12}) +[2026-10-03 13:20:23,312][starVLA.training.train_starvla][INFO] - Step 1618, Loss: {'train/loss': 0.005600559525191784, 'train/grad_norm_pre_clip': 4.658136367797852, 'timing/data': 0.00033231000998057425, 'timing/model': 0.3945645640051225, 'learning_rate/qwen_vl_interface': 7.92241430480005e-06, 'learning_rate/action_model': 7.922414304800049e-05, 'epoch': 0.12}) +[2026-10-03 13:20:24,020][starVLA.training.train_starvla][INFO] - Step 1619, Loss: {'train/loss': 0.00014433404430747032, 'train/grad_norm_pre_clip': 0.06336092203855515, 'timing/data': 0.0003621400101110339, 'timing/model': 0.39586932500242256, 'learning_rate/qwen_vl_interface': 7.919896044797622e-06, 'learning_rate/action_model': 7.91989604479762e-05, 'epoch': 0.12}) +[2026-10-03 13:20:24,725][starVLA.training.train_starvla][INFO] - Step 1620, Loss: {'train/loss': 0.009681561030447483, 'train/grad_norm_pre_clip': 27.362009048461914, 'timing/data': 0.00027171999681741, 'timing/model': 0.39360907500667963, 'learning_rate/qwen_vl_interface': 7.91737668730178e-06, 'learning_rate/action_model': 7.91737668730178e-05, 'epoch': 0.12}) +[2026-10-03 13:20:25,626][starVLA.training.train_starvla][INFO] - Step 1621, Loss: {'train/loss': 0.38616034388542175, 'train/grad_norm_pre_clip': 41.47603988647461, 'timing/data': 0.0003805699961958453, 'timing/model': 0.5894046320026973, 'learning_rate/qwen_vl_interface': 7.91485623334814e-06, 'learning_rate/action_model': 7.91485623334814e-05, 'epoch': 0.12}) +[2026-10-03 13:20:26,333][starVLA.training.train_starvla][INFO] - Step 1622, Loss: {'train/loss': 0.0009688622085377574, 'train/grad_norm_pre_clip': 0.1404123455286026, 'timing/data': 0.00028510999982245266, 'timing/model': 0.39472370500152465, 'learning_rate/qwen_vl_interface': 7.91233468397276e-06, 'learning_rate/action_model': 7.912334683972761e-05, 'epoch': 0.12}) +[2026-10-03 13:20:27,040][starVLA.training.train_starvla][INFO] - Step 1623, Loss: {'train/loss': 7.104407268343493e-05, 'train/grad_norm_pre_clip': 0.019246041774749756, 'timing/data': 0.0003400999994482845, 'timing/model': 0.39497001400741283, 'learning_rate/qwen_vl_interface': 7.90981204021216e-06, 'learning_rate/action_model': 7.90981204021216e-05, 'epoch': 0.12}) +[2026-10-03 13:20:27,744][starVLA.training.train_starvla][INFO] - Step 1624, Loss: {'train/loss': 0.002562347101047635, 'train/grad_norm_pre_clip': 0.6374725103378296, 'timing/data': 0.0003273900074418634, 'timing/model': 0.3933829149900703, 'learning_rate/qwen_vl_interface': 7.907288303103298e-06, 'learning_rate/action_model': 7.907288303103298e-05, 'epoch': 0.12}) +[2026-10-03 13:20:28,450][starVLA.training.train_starvla][INFO] - Step 1625, Loss: {'train/loss': 0.0022309478372335434, 'train/grad_norm_pre_clip': 0.49793457984924316, 'timing/data': 0.00022891000844538212, 'timing/model': 0.3941410949919373, 'learning_rate/qwen_vl_interface': 7.90476347368359e-06, 'learning_rate/action_model': 7.90476347368359e-05, 'epoch': 0.12}) +[2026-10-03 13:20:29,157][starVLA.training.train_starvla][INFO] - Step 1626, Loss: {'train/loss': 0.10185037553310394, 'train/grad_norm_pre_clip': 19.524091720581055, 'timing/data': 0.0003984400100307539, 'timing/model': 0.3951275839935988, 'learning_rate/qwen_vl_interface': 7.902237552990897e-06, 'learning_rate/action_model': 7.902237552990897e-05, 'epoch': 0.12}) +[2026-10-03 13:20:30,055][starVLA.training.train_starvla][INFO] - Step 1627, Loss: {'train/loss': 0.010187669657170773, 'train/grad_norm_pre_clip': 13.84652042388916, 'timing/data': 0.00022854999406263232, 'timing/model': 0.3946310750034172, 'learning_rate/qwen_vl_interface': 7.89971054206353e-06, 'learning_rate/action_model': 7.89971054206353e-05, 'epoch': 0.12}) +[2026-10-03 13:20:30,760][starVLA.training.train_starvla][INFO] - Step 1628, Loss: {'train/loss': 0.42638012766838074, 'train/grad_norm_pre_clip': 15.232739448547363, 'timing/data': 0.0002619699953356758, 'timing/model': 0.39404944398847874, 'learning_rate/qwen_vl_interface': 7.897182441940247e-06, 'learning_rate/action_model': 7.897182441940247e-05, 'epoch': 0.12}) +[2026-10-03 13:20:31,465][starVLA.training.train_starvla][INFO] - Step 1629, Loss: {'train/loss': 0.0025307941250503063, 'train/grad_norm_pre_clip': 1.3476142883300781, 'timing/data': 0.0002116599935106933, 'timing/model': 0.39396778399532195, 'learning_rate/qwen_vl_interface': 7.894653253660256e-06, 'learning_rate/action_model': 7.894653253660256e-05, 'epoch': 0.12}) +[2026-10-03 13:20:32,172][starVLA.training.train_starvla][INFO] - Step 1630, Loss: {'train/loss': 0.0010160522069782019, 'train/grad_norm_pre_clip': 0.2599278390407562, 'timing/data': 0.00030933000380173326, 'timing/model': 0.39533842499076854, 'learning_rate/qwen_vl_interface': 7.89212297826321e-06, 'learning_rate/action_model': 7.892122978263209e-05, 'epoch': 0.12}) +[2026-10-03 13:20:32,879][starVLA.training.train_starvla][INFO] - Step 1631, Loss: {'train/loss': 0.0016040924238041043, 'train/grad_norm_pre_clip': 0.7282419204711914, 'timing/data': 0.0002208600053563714, 'timing/model': 0.39487171500513796, 'learning_rate/qwen_vl_interface': 7.889591616789206e-06, 'learning_rate/action_model': 7.889591616789206e-05, 'epoch': 0.12}) +[2026-10-03 13:20:33,776][starVLA.training.train_starvla][INFO] - Step 1632, Loss: {'train/loss': 0.43452632427215576, 'train/grad_norm_pre_clip': 11.437461853027344, 'timing/data': 0.00030061999859753996, 'timing/model': 0.5853871630097274, 'learning_rate/qwen_vl_interface': 7.887059170278799e-06, 'learning_rate/action_model': 7.887059170278797e-05, 'epoch': 0.12}) +[2026-10-03 13:20:34,480][starVLA.training.train_starvla][INFO] - Step 1633, Loss: {'train/loss': 0.35822561383247375, 'train/grad_norm_pre_clip': 88.37175750732422, 'timing/data': 0.0002630600065458566, 'timing/model': 0.3931146950053517, 'learning_rate/qwen_vl_interface': 7.884525639772974e-06, 'learning_rate/action_model': 7.884525639772974e-05, 'epoch': 0.12}) +[2026-10-03 13:20:35,186][starVLA.training.train_starvla][INFO] - Step 1634, Loss: {'train/loss': 0.0025445991195738316, 'train/grad_norm_pre_clip': 0.9344243407249451, 'timing/data': 0.0002570300130173564, 'timing/model': 0.39446684400900267, 'learning_rate/qwen_vl_interface': 7.881991026313176e-06, 'learning_rate/action_model': 7.881991026313175e-05, 'epoch': 0.12}) +[2026-10-03 13:20:35,892][starVLA.training.train_starvla][INFO] - Step 1635, Loss: {'train/loss': 0.027759362012147903, 'train/grad_norm_pre_clip': 47.0914421081543, 'timing/data': 0.00022965999960433692, 'timing/model': 0.3946332239866024, 'learning_rate/qwen_vl_interface': 7.879455330941283e-06, 'learning_rate/action_model': 7.879455330941282e-05, 'epoch': 0.12}) +[2026-10-03 13:20:36,596][starVLA.training.train_starvla][INFO] - Step 1636, Loss: {'train/loss': 0.0006384610896930099, 'train/grad_norm_pre_clip': 0.1852876842021942, 'timing/data': 0.0003284199920017272, 'timing/model': 0.39313353500619996, 'learning_rate/qwen_vl_interface': 7.876918554699627e-06, 'learning_rate/action_model': 7.876918554699627e-05, 'epoch': 0.12}) +[2026-10-03 13:20:37,300][starVLA.training.train_starvla][INFO] - Step 1637, Loss: {'train/loss': 0.2970905303955078, 'train/grad_norm_pre_clip': 64.52677154541016, 'timing/data': 0.00020874000620096922, 'timing/model': 0.3932214749947889, 'learning_rate/qwen_vl_interface': 7.874380698630982e-06, 'learning_rate/action_model': 7.874380698630982e-05, 'epoch': 0.12}) +[2026-10-03 13:20:38,200][starVLA.training.train_starvla][INFO] - Step 1638, Loss: {'train/loss': 0.009277409873902798, 'train/grad_norm_pre_clip': 3.474036455154419, 'timing/data': 0.00032051000744104385, 'timing/model': 0.3950895950110862, 'learning_rate/qwen_vl_interface': 7.871841763778562e-06, 'learning_rate/action_model': 7.871841763778563e-05, 'epoch': 0.12}) +[2026-10-03 13:20:38,907][starVLA.training.train_starvla][INFO] - Step 1639, Loss: {'train/loss': 0.33603283762931824, 'train/grad_norm_pre_clip': 19.672290802001953, 'timing/data': 0.00022231000184547156, 'timing/model': 0.3947801949980203, 'learning_rate/qwen_vl_interface': 7.86930175118603e-06, 'learning_rate/action_model': 7.86930175118603e-05, 'epoch': 0.12}) +[2026-10-03 13:20:39,612][starVLA.training.train_starvla][INFO] - Step 1640, Loss: {'train/loss': 0.20375026762485504, 'train/grad_norm_pre_clip': 126.73294067382812, 'timing/data': 0.00021074000687804073, 'timing/model': 0.39333859499311075, 'learning_rate/qwen_vl_interface': 7.866760661897487e-06, 'learning_rate/action_model': 7.866760661897487e-05, 'epoch': 0.12}) +[2026-10-03 13:20:40,315][starVLA.training.train_starvla][INFO] - Step 1641, Loss: {'train/loss': 0.001089174416847527, 'train/grad_norm_pre_clip': 0.3051794469356537, 'timing/data': 0.00028046999068465084, 'timing/model': 0.39269824500661343, 'learning_rate/qwen_vl_interface': 7.86421849695748e-06, 'learning_rate/action_model': 7.864218496957481e-05, 'epoch': 0.12}) +[2026-10-03 13:20:41,022][starVLA.training.train_starvla][INFO] - Step 1642, Loss: {'train/loss': 0.32755210995674133, 'train/grad_norm_pre_clip': 46.16535568237305, 'timing/data': 0.00032495999766979367, 'timing/model': 0.39487047400325537, 'learning_rate/qwen_vl_interface': 7.861675257410998e-06, 'learning_rate/action_model': 7.861675257410997e-05, 'epoch': 0.12}) +[2026-10-03 13:20:41,927][starVLA.training.train_starvla][INFO] - Step 1643, Loss: {'train/loss': 0.047082118690013885, 'train/grad_norm_pre_clip': 115.55555725097656, 'timing/data': 0.0003765599976759404, 'timing/model': 0.3942553740052972, 'learning_rate/qwen_vl_interface': 7.859130944303468e-06, 'learning_rate/action_model': 7.859130944303467e-05, 'epoch': 0.12}) +[2026-10-03 13:20:42,630][starVLA.training.train_starvla][INFO] - Step 1644, Loss: {'train/loss': 0.49660736322402954, 'train/grad_norm_pre_clip': 19.542367935180664, 'timing/data': 0.00027011000202037394, 'timing/model': 0.39290791501116473, 'learning_rate/qwen_vl_interface': 7.856585558680761e-06, 'learning_rate/action_model': 7.856585558680761e-05, 'epoch': 0.12}) +[2026-10-03 13:20:43,334][starVLA.training.train_starvla][INFO] - Step 1645, Loss: {'train/loss': 0.0025353103410452604, 'train/grad_norm_pre_clip': 6.225148677825928, 'timing/data': 0.0003002099983859807, 'timing/model': 0.3930187649966683, 'learning_rate/qwen_vl_interface': 7.854039101589191e-06, 'learning_rate/action_model': 7.854039101589191e-05, 'epoch': 0.12}) +[2026-10-03 13:20:44,040][starVLA.training.train_starvla][INFO] - Step 1646, Loss: {'train/loss': 0.00025607418501749635, 'train/grad_norm_pre_clip': 0.06949058175086975, 'timing/data': 0.0002687699889065698, 'timing/model': 0.39475764498638455, 'learning_rate/qwen_vl_interface': 7.851491574075509e-06, 'learning_rate/action_model': 7.851491574075509e-05, 'epoch': 0.12}) +[2026-10-03 13:20:44,748][starVLA.training.train_starvla][INFO] - Step 1647, Loss: {'train/loss': 0.31600508093833923, 'train/grad_norm_pre_clip': 164.97323608398438, 'timing/data': 0.0003009000065503642, 'timing/model': 0.39564557498670183, 'learning_rate/qwen_vl_interface': 7.848942977186909e-06, 'learning_rate/action_model': 7.848942977186908e-05, 'epoch': 0.12}) +[2026-10-03 13:20:45,648][starVLA.training.train_starvla][INFO] - Step 1648, Loss: {'train/loss': 0.22807571291923523, 'train/grad_norm_pre_clip': 44.14242172241211, 'timing/data': 0.00030559000151697546, 'timing/model': 0.5887377720064251, 'learning_rate/qwen_vl_interface': 7.84639331197102e-06, 'learning_rate/action_model': 7.84639331197102e-05, 'epoch': 0.12}) +[2026-10-03 13:20:46,351][starVLA.training.train_starvla][INFO] - Step 1649, Loss: {'train/loss': 0.46627873182296753, 'train/grad_norm_pre_clip': 105.52727508544922, 'timing/data': 0.00025653900229372084, 'timing/model': 0.39269849499396514, 'learning_rate/qwen_vl_interface': 7.843842579475918e-06, 'learning_rate/action_model': 7.843842579475918e-05, 'epoch': 0.12}) +[2026-10-03 13:20:47,058][starVLA.training.train_starvla][INFO] - Step 1650, Loss: {'train/loss': 0.19579601287841797, 'train/grad_norm_pre_clip': 50.957122802734375, 'timing/data': 0.0003209799906471744, 'timing/model': 0.39538628500304185, 'learning_rate/qwen_vl_interface': 7.841290780750107e-06, 'learning_rate/action_model': 7.841290780750107e-05, 'epoch': 0.12}) +[2026-10-03 13:20:47,765][starVLA.training.train_starvla][INFO] - Step 1651, Loss: {'train/loss': 0.0009343340643681586, 'train/grad_norm_pre_clip': 0.1659696400165558, 'timing/data': 0.0002619800070533529, 'timing/model': 0.39469755500613246, 'learning_rate/qwen_vl_interface': 7.838737916842538e-06, 'learning_rate/action_model': 7.838737916842538e-05, 'epoch': 0.12}) +[2026-10-03 13:20:48,470][starVLA.training.train_starvla][INFO] - Step 1652, Loss: {'train/loss': 0.0050804587081074715, 'train/grad_norm_pre_clip': 21.640987396240234, 'timing/data': 0.00035848999687004834, 'timing/model': 0.39315416499448474, 'learning_rate/qwen_vl_interface': 7.836183988802596e-06, 'learning_rate/action_model': 7.836183988802595e-05, 'epoch': 0.12}) +[2026-10-03 13:20:49,175][starVLA.training.train_starvla][INFO] - Step 1653, Loss: {'train/loss': 0.001039846451021731, 'train/grad_norm_pre_clip': 0.2152450829744339, 'timing/data': 0.0003489700029604137, 'timing/model': 0.39304783500847407, 'learning_rate/qwen_vl_interface': 7.833628997680107e-06, 'learning_rate/action_model': 7.833628997680107e-05, 'epoch': 0.12}) +[2026-10-03 13:20:50,072][starVLA.training.train_starvla][INFO] - Step 1654, Loss: {'train/loss': 0.11490218341350555, 'train/grad_norm_pre_clip': 58.89448547363281, 'timing/data': 0.000219130000914447, 'timing/model': 0.3958087350038113, 'learning_rate/qwen_vl_interface': 7.831072944525329e-06, 'learning_rate/action_model': 7.831072944525328e-05, 'epoch': 0.12}) +[2026-10-03 13:20:50,779][starVLA.training.train_starvla][INFO] - Step 1655, Loss: {'train/loss': 0.1347331702709198, 'train/grad_norm_pre_clip': 16.79132652282715, 'timing/data': 0.0003281399986008182, 'timing/model': 0.39485032499942463, 'learning_rate/qwen_vl_interface': 7.828515830388956e-06, 'learning_rate/action_model': 7.828515830388956e-05, 'epoch': 0.12}) +[2026-10-03 13:20:51,479][starVLA.training.train_starvla][INFO] - Step 1656, Loss: {'train/loss': 0.5317989587783813, 'train/grad_norm_pre_clip': 255.40945434570312, 'timing/data': 0.00037336000241339207, 'timing/model': 0.3890252950077411, 'learning_rate/qwen_vl_interface': 7.825957656322129e-06, 'learning_rate/action_model': 7.825957656322128e-05, 'epoch': 0.12}) +[2026-10-03 13:20:52,174][starVLA.training.train_starvla][INFO] - Step 1657, Loss: {'train/loss': 0.0011506166774779558, 'train/grad_norm_pre_clip': 1.3089359998703003, 'timing/data': 0.00027539000438991934, 'timing/model': 0.3883601150009781, 'learning_rate/qwen_vl_interface': 7.823398423376406e-06, 'learning_rate/action_model': 7.823398423376406e-05, 'epoch': 0.12}) +[2026-10-03 13:20:52,873][starVLA.training.train_starvla][INFO] - Step 1658, Loss: {'train/loss': 0.004247335717082024, 'train/grad_norm_pre_clip': 0.8036672472953796, 'timing/data': 0.00023937999503687024, 'timing/model': 0.389711175012053, 'learning_rate/qwen_vl_interface': 7.820838132603801e-06, 'learning_rate/action_model': 7.8208381326038e-05, 'epoch': 0.12}) +[2026-10-03 13:20:53,758][starVLA.training.train_starvla][INFO] - Step 1659, Loss: {'train/loss': 0.04847578704357147, 'train/grad_norm_pre_clip': 13.19006061553955, 'timing/data': 0.00025221999385394156, 'timing/model': 0.3894163049990311, 'learning_rate/qwen_vl_interface': 7.818276785056748e-06, 'learning_rate/action_model': 7.818276785056749e-05, 'epoch': 0.12}) +[2026-10-03 13:20:54,454][starVLA.training.train_starvla][INFO] - Step 1660, Loss: {'train/loss': 0.023158375173807144, 'train/grad_norm_pre_clip': 42.753108978271484, 'timing/data': 0.00020700000459328294, 'timing/model': 0.3882460250024451, 'learning_rate/qwen_vl_interface': 7.815714381788123e-06, 'learning_rate/action_model': 7.815714381788122e-05, 'epoch': 0.12}) +[2026-10-03 13:20:55,151][starVLA.training.train_starvla][INFO] - Step 1661, Loss: {'train/loss': 0.00025664010900072753, 'train/grad_norm_pre_clip': 0.1344393938779831, 'timing/data': 0.00019612000323832035, 'timing/model': 0.3897338059905451, 'learning_rate/qwen_vl_interface': 7.813150923851227e-06, 'learning_rate/action_model': 7.813150923851227e-05, 'epoch': 0.12}) +[2026-10-03 13:20:55,851][starVLA.training.train_starvla][INFO] - Step 1662, Loss: {'train/loss': 0.0022435092832893133, 'train/grad_norm_pre_clip': 0.9716149568557739, 'timing/data': 0.00034502999915275723, 'timing/model': 0.39055691500834655, 'learning_rate/qwen_vl_interface': 7.810586412299807e-06, 'learning_rate/action_model': 7.810586412299806e-05, 'epoch': 0.12}) +[2026-10-03 13:20:56,546][starVLA.training.train_starvla][INFO] - Step 1663, Loss: {'train/loss': 0.3120568096637726, 'train/grad_norm_pre_clip': 10.630032539367676, 'timing/data': 0.0001963600079761818, 'timing/model': 0.3903774250065908, 'learning_rate/qwen_vl_interface': 7.808020848188034e-06, 'learning_rate/action_model': 7.808020848188034e-05, 'epoch': 0.12}) +[2026-10-03 13:20:57,437][starVLA.training.train_starvla][INFO] - Step 1664, Loss: {'train/loss': 0.6308048367500305, 'train/grad_norm_pre_clip': 33.61302185058594, 'timing/data': 0.000285210000583902, 'timing/model': 0.38943208499404136, 'learning_rate/qwen_vl_interface': 7.805454232570517e-06, 'learning_rate/action_model': 7.805454232570517e-05, 'epoch': 0.12}) +[2026-10-03 13:20:58,133][starVLA.training.train_starvla][INFO] - Step 1665, Loss: {'train/loss': 0.4282285273075104, 'train/grad_norm_pre_clip': 18.352354049682617, 'timing/data': 0.00019571000302676111, 'timing/model': 0.3895189049944747, 'learning_rate/qwen_vl_interface': 7.80288656650229e-06, 'learning_rate/action_model': 7.80288656650229e-05, 'epoch': 0.12}) +[2026-10-03 13:20:58,832][starVLA.training.train_starvla][INFO] - Step 1666, Loss: {'train/loss': 0.2458074688911438, 'train/grad_norm_pre_clip': 94.19573974609375, 'timing/data': 0.000344749991199933, 'timing/model': 0.3901506149995839, 'learning_rate/qwen_vl_interface': 7.800317851038829e-06, 'learning_rate/action_model': 7.800317851038828e-05, 'epoch': 0.12}) +[2026-10-03 13:20:59,529][starVLA.training.train_starvla][INFO] - Step 1667, Loss: {'train/loss': 0.007751902565360069, 'train/grad_norm_pre_clip': 6.732556343078613, 'timing/data': 0.00018567900406196713, 'timing/model': 0.3898240260023158, 'learning_rate/qwen_vl_interface': 7.797748087236033e-06, 'learning_rate/action_model': 7.797748087236033e-05, 'epoch': 0.12}) +[2026-10-03 13:21:00,226][starVLA.training.train_starvla][INFO] - Step 1668, Loss: {'train/loss': 7.53225467633456e-05, 'train/grad_norm_pre_clip': 0.022479569539427757, 'timing/data': 0.0002850500022759661, 'timing/model': 0.38943853600358125, 'learning_rate/qwen_vl_interface': 7.795177276150234e-06, 'learning_rate/action_model': 7.795177276150234e-05, 'epoch': 0.12}) +[2026-10-03 13:21:01,110][starVLA.training.train_starvla][INFO] - Step 1669, Loss: {'train/loss': 0.30442655086517334, 'train/grad_norm_pre_clip': 10.80139446258545, 'timing/data': 0.00022496000747196376, 'timing/model': 0.5766690730088158, 'learning_rate/qwen_vl_interface': 7.792605418838196e-06, 'learning_rate/action_model': 7.792605418838197e-05, 'epoch': 0.12}) +[2026-10-03 13:21:01,808][starVLA.training.train_starvla][INFO] - Step 1670, Loss: {'train/loss': 8.674745913594961e-05, 'train/grad_norm_pre_clip': 0.03592722490429878, 'timing/data': 0.00023746999795548618, 'timing/model': 0.3894302850094391, 'learning_rate/qwen_vl_interface': 7.790032516357116e-06, 'learning_rate/action_model': 7.790032516357117e-05, 'epoch': 0.12}) +[2026-10-03 13:21:02,504][starVLA.training.train_starvla][INFO] - Step 1671, Loss: {'train/loss': 0.0035644115414470434, 'train/grad_norm_pre_clip': 2.6759326457977295, 'timing/data': 0.00025194999761879444, 'timing/model': 0.38941128499573097, 'learning_rate/qwen_vl_interface': 7.787458569764613e-06, 'learning_rate/action_model': 7.787458569764612e-05, 'epoch': 0.12}) +[2026-10-03 13:21:03,201][starVLA.training.train_starvla][INFO] - Step 1672, Loss: {'train/loss': 0.3375399112701416, 'train/grad_norm_pre_clip': 69.66239929199219, 'timing/data': 0.00029181000718381256, 'timing/model': 0.38964948600914795, 'learning_rate/qwen_vl_interface': 7.784883580118738e-06, 'learning_rate/action_model': 7.784883580118738e-05, 'epoch': 0.12}) +[2026-10-03 13:21:03,898][starVLA.training.train_starvla][INFO] - Step 1673, Loss: {'train/loss': 0.004410688299685717, 'train/grad_norm_pre_clip': 3.40342378616333, 'timing/data': 0.00022787999478168786, 'timing/model': 0.38978664499882143, 'learning_rate/qwen_vl_interface': 7.782307548477975e-06, 'learning_rate/action_model': 7.782307548477975e-05, 'epoch': 0.12}) +[2026-10-03 13:21:04,786][starVLA.training.train_starvla][INFO] - Step 1674, Loss: {'train/loss': 0.07479426264762878, 'train/grad_norm_pre_clip': 100.66488647460938, 'timing/data': 0.0003204700042260811, 'timing/model': 0.5784812140045688, 'learning_rate/qwen_vl_interface': 7.779730475901233e-06, 'learning_rate/action_model': 7.779730475901232e-05, 'epoch': 0.12}) +[2026-10-03 13:21:05,482][starVLA.training.train_starvla][INFO] - Step 1675, Loss: {'train/loss': 0.20182695984840393, 'train/grad_norm_pre_clip': 10.5061674118042, 'timing/data': 0.0003144799993606284, 'timing/model': 0.3894766260054894, 'learning_rate/qwen_vl_interface': 7.777152363447849e-06, 'learning_rate/action_model': 7.77715236344785e-05, 'epoch': 0.12}) +[2026-10-03 13:21:06,179][starVLA.training.train_starvla][INFO] - Step 1676, Loss: {'train/loss': 0.00011196720879524946, 'train/grad_norm_pre_clip': 0.02762436494231224, 'timing/data': 0.00028861001192126423, 'timing/model': 0.3899364849930862, 'learning_rate/qwen_vl_interface': 7.774573212177587e-06, 'learning_rate/action_model': 7.774573212177588e-05, 'epoch': 0.12}) +[2026-10-03 13:21:06,876][starVLA.training.train_starvla][INFO] - Step 1677, Loss: {'train/loss': 0.006325739435851574, 'train/grad_norm_pre_clip': 0.970984935760498, 'timing/data': 0.0002800390066113323, 'timing/model': 0.3898080859944457, 'learning_rate/qwen_vl_interface': 7.77199302315064e-06, 'learning_rate/action_model': 7.77199302315064e-05, 'epoch': 0.12}) +[2026-10-03 13:21:07,576][starVLA.training.train_starvla][INFO] - Step 1678, Loss: {'train/loss': 0.16139709949493408, 'train/grad_norm_pre_clip': 64.05956268310547, 'timing/data': 0.00029116000223439187, 'timing/model': 0.39022979501169175, 'learning_rate/qwen_vl_interface': 7.769411797427622e-06, 'learning_rate/action_model': 7.769411797427623e-05, 'epoch': 0.12}) +[2026-10-03 13:21:08,272][starVLA.training.train_starvla][INFO] - Step 1679, Loss: {'train/loss': 0.02509290724992752, 'train/grad_norm_pre_clip': 14.520492553710938, 'timing/data': 0.00021282999659888446, 'timing/model': 0.3896354250027798, 'learning_rate/qwen_vl_interface': 7.766829536069584e-06, 'learning_rate/action_model': 7.766829536069584e-05, 'epoch': 0.12}) +[2026-10-03 13:21:09,160][starVLA.training.train_starvla][INFO] - Step 1680, Loss: {'train/loss': 0.3280239403247833, 'train/grad_norm_pre_clip': 52.746952056884766, 'timing/data': 0.00036716999602504075, 'timing/model': 0.3894782850111369, 'learning_rate/qwen_vl_interface': 7.764246240137991e-06, 'learning_rate/action_model': 7.764246240137991e-05, 'epoch': 0.12}) +[2026-10-03 13:21:09,858][starVLA.training.train_starvla][INFO] - Step 1681, Loss: {'train/loss': 0.00036632874980568886, 'train/grad_norm_pre_clip': 0.1754639893770218, 'timing/data': 0.0003096000000368804, 'timing/model': 0.38949355599470437, 'learning_rate/qwen_vl_interface': 7.761661910694741e-06, 'learning_rate/action_model': 7.761661910694741e-05, 'epoch': 0.12}) +[2026-10-03 13:21:10,557][starVLA.training.train_starvla][INFO] - Step 1682, Loss: {'train/loss': 0.2756904065608978, 'train/grad_norm_pre_clip': 14.400474548339844, 'timing/data': 0.00030110000807326287, 'timing/model': 0.3903979949973291, 'learning_rate/qwen_vl_interface': 7.759076548802153e-06, 'learning_rate/action_model': 7.759076548802152e-05, 'epoch': 0.12}) +[2026-10-03 13:21:11,252][starVLA.training.train_starvla][INFO] - Step 1683, Loss: {'train/loss': 0.01283502671867609, 'train/grad_norm_pre_clip': 10.905327796936035, 'timing/data': 0.00024541901075281203, 'timing/model': 0.3897991260018898, 'learning_rate/qwen_vl_interface': 7.75649015552297e-06, 'learning_rate/action_model': 7.75649015552297e-05, 'epoch': 0.12}) +[2026-10-03 13:21:11,949][starVLA.training.train_starvla][INFO] - Step 1684, Loss: {'train/loss': 0.01007635984569788, 'train/grad_norm_pre_clip': 1.6587318181991577, 'timing/data': 0.00019244001305196434, 'timing/model': 0.38960708600643557, 'learning_rate/qwen_vl_interface': 7.753902731920364e-06, 'learning_rate/action_model': 7.753902731920364e-05, 'epoch': 0.12}) +[2026-10-03 13:21:12,832][starVLA.training.train_starvla][INFO] - Step 1685, Loss: {'train/loss': 0.00030929362401366234, 'train/grad_norm_pre_clip': 0.21193760633468628, 'timing/data': 0.0001941700029419735, 'timing/model': 0.5762661839980865, 'learning_rate/qwen_vl_interface': 7.751314279057925e-06, 'learning_rate/action_model': 7.751314279057923e-05, 'epoch': 0.12}) +[2026-10-03 13:21:13,530][starVLA.training.train_starvla][INFO] - Step 1686, Loss: {'train/loss': 0.013138561509549618, 'train/grad_norm_pre_clip': 1.8571505546569824, 'timing/data': 0.00039018000825308263, 'timing/model': 0.38990765500057023, 'learning_rate/qwen_vl_interface': 7.74872479799967e-06, 'learning_rate/action_model': 7.74872479799967e-05, 'epoch': 0.12}) +[2026-10-03 13:21:14,227][starVLA.training.train_starvla][INFO] - Step 1687, Loss: {'train/loss': 7.112399180186912e-05, 'train/grad_norm_pre_clip': 0.017787566408514977, 'timing/data': 0.0003757399972528219, 'timing/model': 0.38920340500772, 'learning_rate/qwen_vl_interface': 7.746134289810034e-06, 'learning_rate/action_model': 7.746134289810033e-05, 'epoch': 0.12}) +[2026-10-03 13:21:14,924][starVLA.training.train_starvla][INFO] - Step 1688, Loss: {'train/loss': 9.881556616164744e-05, 'train/grad_norm_pre_clip': 0.053782347589731216, 'timing/data': 0.0002522590075386688, 'timing/model': 0.3897588960098801, 'learning_rate/qwen_vl_interface': 7.74354275555388e-06, 'learning_rate/action_model': 7.74354275555388e-05, 'epoch': 0.12}) +[2026-10-03 13:21:15,621][starVLA.training.train_starvla][INFO] - Step 1689, Loss: {'train/loss': 0.043582089245319366, 'train/grad_norm_pre_clip': 29.54462432861328, 'timing/data': 0.00024405999283771962, 'timing/model': 0.3893906559969764, 'learning_rate/qwen_vl_interface': 7.740950196296488e-06, 'learning_rate/action_model': 7.740950196296488e-05, 'epoch': 0.12}) +[2026-10-03 13:21:16,319][starVLA.training.train_starvla][INFO] - Step 1690, Loss: {'train/loss': 0.015766821801662445, 'train/grad_norm_pre_clip': 16.512378692626953, 'timing/data': 0.00020138999389018863, 'timing/model': 0.39023767500475515, 'learning_rate/qwen_vl_interface': 7.738356613103562e-06, 'learning_rate/action_model': 7.73835661310356e-05, 'epoch': 0.12}) +[2026-10-03 13:21:17,206][starVLA.training.train_starvla][INFO] - Step 1691, Loss: {'train/loss': 0.00018002127762883902, 'train/grad_norm_pre_clip': 0.11208683252334595, 'timing/data': 0.000238389999140054, 'timing/model': 0.3894168849947164, 'learning_rate/qwen_vl_interface': 7.735762007041226e-06, 'learning_rate/action_model': 7.735762007041226e-05, 'epoch': 0.12}) +[2026-10-03 13:21:17,903][starVLA.training.train_starvla][INFO] - Step 1692, Loss: {'train/loss': 0.15668551623821259, 'train/grad_norm_pre_clip': 91.79537200927734, 'timing/data': 0.00020663999021053314, 'timing/model': 0.389665225011413, 'learning_rate/qwen_vl_interface': 7.733166379176023e-06, 'learning_rate/action_model': 7.733166379176023e-05, 'epoch': 0.12}) +[2026-10-03 13:21:18,600][starVLA.training.train_starvla][INFO] - Step 1693, Loss: {'train/loss': 0.7211406230926514, 'train/grad_norm_pre_clip': 139.3927001953125, 'timing/data': 0.0003075700078625232, 'timing/model': 0.3895094949984923, 'learning_rate/qwen_vl_interface': 7.730569730574921e-06, 'learning_rate/action_model': 7.73056973057492e-05, 'epoch': 0.12}) +[2026-10-03 13:21:19,299][starVLA.training.train_starvla][INFO] - Step 1694, Loss: {'train/loss': 0.006214236840605736, 'train/grad_norm_pre_clip': 1.9088408946990967, 'timing/data': 0.00019736999820452183, 'timing/model': 0.3901971949962899, 'learning_rate/qwen_vl_interface': 7.727972062305299e-06, 'learning_rate/action_model': 7.727972062305299e-05, 'epoch': 0.12}) +[2026-10-03 13:21:19,996][starVLA.training.train_starvla][INFO] - Step 1695, Loss: {'train/loss': 2.1112866306793876e-05, 'train/grad_norm_pre_clip': 0.005534532014280558, 'timing/data': 0.00031006999779492617, 'timing/model': 0.3899862149992259, 'learning_rate/qwen_vl_interface': 7.725373375434965e-06, 'learning_rate/action_model': 7.725373375434965e-05, 'epoch': 0.12}) +[2026-10-03 13:21:20,885][starVLA.training.train_starvla][INFO] - Step 1696, Loss: {'train/loss': 0.0016447675880044699, 'train/grad_norm_pre_clip': 0.2909933626651764, 'timing/data': 0.00020190000941511244, 'timing/model': 0.3902970049966825, 'learning_rate/qwen_vl_interface': 7.722773671032141e-06, 'learning_rate/action_model': 7.722773671032142e-05, 'epoch': 0.12}) +[2026-10-03 13:21:21,582][starVLA.training.train_starvla][INFO] - Step 1697, Loss: {'train/loss': 0.0010289610363543034, 'train/grad_norm_pre_clip': 0.2665691077709198, 'timing/data': 0.00027933999081142247, 'timing/model': 0.3900264659896493, 'learning_rate/qwen_vl_interface': 7.720172950165466e-06, 'learning_rate/action_model': 7.720172950165465e-05, 'epoch': 0.12}) +[2026-10-03 13:21:22,281][starVLA.training.train_starvla][INFO] - Step 1698, Loss: {'train/loss': 0.00017240640590898693, 'train/grad_norm_pre_clip': 0.2794218063354492, 'timing/data': 0.0002540599962230772, 'timing/model': 0.39000065499567427, 'learning_rate/qwen_vl_interface': 7.717571213903996e-06, 'learning_rate/action_model': 7.717571213903996e-05, 'epoch': 0.12}) +[2026-10-03 13:21:22,977][starVLA.training.train_starvla][INFO] - Step 1699, Loss: {'train/loss': 0.0029422463849186897, 'train/grad_norm_pre_clip': 0.5413148403167725, 'timing/data': 0.00028358999406918883, 'timing/model': 0.3895095460029552, 'learning_rate/qwen_vl_interface': 7.714968463317207e-06, 'learning_rate/action_model': 7.714968463317207e-05, 'epoch': 0.12}) +[2026-10-03 13:21:23,675][starVLA.training.train_starvla][INFO] - Step 1700, Loss: {'train/loss': 0.9140982627868652, 'train/grad_norm_pre_clip': 117.14300537109375, 'timing/data': 0.0002929699985543266, 'timing/model': 0.39019339499645866, 'learning_rate/qwen_vl_interface': 7.712364699474996e-06, 'learning_rate/action_model': 7.712364699474996e-05, 'epoch': 0.12}) +[2026-10-03 13:21:24,564][starVLA.training.train_starvla][INFO] - Step 1701, Loss: {'train/loss': 5.206941568758339e-05, 'train/grad_norm_pre_clip': 0.1363363265991211, 'timing/data': 0.0002503299911040813, 'timing/model': 0.5796607739903266, 'learning_rate/qwen_vl_interface': 7.709759923447667e-06, 'learning_rate/action_model': 7.709759923447665e-05, 'epoch': 0.12}) +[2026-10-03 13:21:25,263][starVLA.training.train_starvla][INFO] - Step 1702, Loss: {'train/loss': 0.18249334394931793, 'train/grad_norm_pre_clip': 273.7012634277344, 'timing/data': 0.0003000799915753305, 'timing/model': 0.39072626500274055, 'learning_rate/qwen_vl_interface': 7.707154136305945e-06, 'learning_rate/action_model': 7.707154136305945e-05, 'epoch': 0.12}) +[2026-10-03 13:21:25,960][starVLA.training.train_starvla][INFO] - Step 1703, Loss: {'train/loss': 0.0007001726771704853, 'train/grad_norm_pre_clip': 1.9129077196121216, 'timing/data': 0.0002841500099748373, 'timing/model': 0.3894523159979144, 'learning_rate/qwen_vl_interface': 7.704547339120974e-06, 'learning_rate/action_model': 7.704547339120974e-05, 'epoch': 0.12}) +[2026-10-03 13:21:26,657][starVLA.training.train_starvla][INFO] - Step 1704, Loss: {'train/loss': 0.006271978374570608, 'train/grad_norm_pre_clip': 14.343999862670898, 'timing/data': 0.00019444999634288251, 'timing/model': 0.3905697950103786, 'learning_rate/qwen_vl_interface': 7.70193953296431e-06, 'learning_rate/action_model': 7.70193953296431e-05, 'epoch': 0.12}) +[2026-10-03 13:21:27,354][starVLA.training.train_starvla][INFO] - Step 1705, Loss: {'train/loss': 7.380438910331577e-05, 'train/grad_norm_pre_clip': 0.05915191397070885, 'timing/data': 0.0002776200126390904, 'timing/model': 0.3893852559995139, 'learning_rate/qwen_vl_interface': 7.69933071890792e-06, 'learning_rate/action_model': 7.699330718907919e-05, 'epoch': 0.12}) +[2026-10-03 13:21:28,054][starVLA.training.train_starvla][INFO] - Step 1706, Loss: {'train/loss': 0.04986637085676193, 'train/grad_norm_pre_clip': 97.44761657714844, 'timing/data': 0.00022137998894322664, 'timing/model': 0.3907921150093898, 'learning_rate/qwen_vl_interface': 7.696720898024188e-06, 'learning_rate/action_model': 7.69672089802419e-05, 'epoch': 0.12}) +[2026-10-03 13:21:28,938][starVLA.training.train_starvla][INFO] - Step 1707, Loss: {'train/loss': 0.2433754950761795, 'train/grad_norm_pre_clip': 284.8796081542969, 'timing/data': 0.0002858299994841218, 'timing/model': 0.39000714500434697, 'learning_rate/qwen_vl_interface': 7.69411007138592e-06, 'learning_rate/action_model': 7.694110071385919e-05, 'epoch': 0.12}) +[2026-10-03 13:21:29,635][starVLA.training.train_starvla][INFO] - Step 1708, Loss: {'train/loss': 0.35306626558303833, 'train/grad_norm_pre_clip': 29.09952735900879, 'timing/data': 0.00025614000333007425, 'timing/model': 0.390107685001567, 'learning_rate/qwen_vl_interface': 7.691498240066322e-06, 'learning_rate/action_model': 7.691498240066322e-05, 'epoch': 0.13}) +[2026-10-03 13:21:30,332][starVLA.training.train_starvla][INFO] - Step 1709, Loss: {'train/loss': 0.4309786260128021, 'train/grad_norm_pre_clip': 20.238906860351562, 'timing/data': 0.00030934999813325703, 'timing/model': 0.3896177959977649, 'learning_rate/qwen_vl_interface': 7.688885405139022e-06, 'learning_rate/action_model': 7.688885405139021e-05, 'epoch': 0.13}) +[2026-10-03 13:21:31,032][starVLA.training.train_starvla][INFO] - Step 1710, Loss: {'train/loss': 0.0005468110321089625, 'train/grad_norm_pre_clip': 0.36890825629234314, 'timing/data': 0.00019767999765463173, 'timing/model': 0.39077124500181526, 'learning_rate/qwen_vl_interface': 7.686271567678056e-06, 'learning_rate/action_model': 7.686271567678054e-05, 'epoch': 0.13}) +[2026-10-03 13:21:31,725][starVLA.training.train_starvla][INFO] - Step 1711, Loss: {'train/loss': 0.0026140580885112286, 'train/grad_norm_pre_clip': 2.975717544555664, 'timing/data': 0.00018961000023409724, 'timing/model': 0.3879532770079095, 'learning_rate/qwen_vl_interface': 7.683656728757873e-06, 'learning_rate/action_model': 7.683656728757872e-05, 'epoch': 0.13}) +[2026-10-03 13:21:32,610][starVLA.training.train_starvla][INFO] - Step 1712, Loss: {'train/loss': 0.28278234601020813, 'train/grad_norm_pre_clip': 60.43179702758789, 'timing/data': 0.00019901999621652067, 'timing/model': 0.3880453949968796, 'learning_rate/qwen_vl_interface': 7.681040889453337e-06, 'learning_rate/action_model': 7.681040889453336e-05, 'epoch': 0.13}) +[2026-10-03 13:21:33,304][starVLA.training.train_starvla][INFO] - Step 1713, Loss: {'train/loss': 0.0003334169741719961, 'train/grad_norm_pre_clip': 0.07931148260831833, 'timing/data': 0.0004094600008102134, 'timing/model': 0.3879193650063826, 'learning_rate/qwen_vl_interface': 7.678424050839722e-06, 'learning_rate/action_model': 7.67842405083972e-05, 'epoch': 0.13}) +[2026-10-03 13:21:34,000][starVLA.training.train_starvla][INFO] - Step 1714, Loss: {'train/loss': 4.3319931137375534e-05, 'train/grad_norm_pre_clip': 0.0066901445388793945, 'timing/data': 0.0002706200029933825, 'timing/model': 0.3889331660029711, 'learning_rate/qwen_vl_interface': 7.675806213992706e-06, 'learning_rate/action_model': 7.675806213992707e-05, 'epoch': 0.13}) +[2026-10-03 13:21:34,696][starVLA.training.train_starvla][INFO] - Step 1715, Loss: {'train/loss': 0.00011510458716657013, 'train/grad_norm_pre_clip': 0.031568221747875214, 'timing/data': 0.0002514999941922724, 'timing/model': 0.3901142959948629, 'learning_rate/qwen_vl_interface': 7.673187379988388e-06, 'learning_rate/action_model': 7.673187379988388e-05, 'epoch': 0.13}) +[2026-10-03 13:21:35,392][starVLA.training.train_starvla][INFO] - Step 1716, Loss: {'train/loss': 7.561855454696342e-05, 'train/grad_norm_pre_clip': 0.01818915642797947, 'timing/data': 0.00025409000227227807, 'timing/model': 0.3891357149987016, 'learning_rate/qwen_vl_interface': 7.67056754990327e-06, 'learning_rate/action_model': 7.670567549903269e-05, 'epoch': 0.13}) +[2026-10-03 13:21:36,280][starVLA.training.train_starvla][INFO] - Step 1717, Loss: {'train/loss': 0.36916446685791016, 'train/grad_norm_pre_clip': 21.060043334960938, 'timing/data': 0.00031899000168778, 'timing/model': 0.3899709060060559, 'learning_rate/qwen_vl_interface': 7.667946724814264e-06, 'learning_rate/action_model': 7.667946724814265e-05, 'epoch': 0.13}) +[2026-10-03 13:21:36,980][starVLA.training.train_starvla][INFO] - Step 1718, Loss: {'train/loss': 0.7367280125617981, 'train/grad_norm_pre_clip': 24.62372398376465, 'timing/data': 0.0001977299980353564, 'timing/model': 0.3904039049957646, 'learning_rate/qwen_vl_interface': 7.665324905798694e-06, 'learning_rate/action_model': 7.665324905798695e-05, 'epoch': 0.13}) +[2026-10-03 13:21:37,676][starVLA.training.train_starvla][INFO] - Step 1719, Loss: {'train/loss': 0.0006900955340825021, 'train/grad_norm_pre_clip': 0.19924716651439667, 'timing/data': 0.00019767999765463173, 'timing/model': 0.38968842499889433, 'learning_rate/qwen_vl_interface': 7.66270209393429e-06, 'learning_rate/action_model': 7.66270209393429e-05, 'epoch': 0.13}) +[2026-10-03 13:21:38,372][starVLA.training.train_starvla][INFO] - Step 1720, Loss: {'train/loss': 0.0002794352767523378, 'train/grad_norm_pre_clip': 0.07696962356567383, 'timing/data': 0.0003149599942844361, 'timing/model': 0.3891426360059995, 'learning_rate/qwen_vl_interface': 7.66007829029919e-06, 'learning_rate/action_model': 7.660078290299191e-05, 'epoch': 0.13}) +[2026-10-03 13:21:39,070][starVLA.training.train_starvla][INFO] - Step 1721, Loss: {'train/loss': 0.0010585106210783124, 'train/grad_norm_pre_clip': 0.7918686866760254, 'timing/data': 0.0003037999995285645, 'timing/model': 0.38976629498938564, 'learning_rate/qwen_vl_interface': 7.657453495971942e-06, 'learning_rate/action_model': 7.657453495971943e-05, 'epoch': 0.13}) +[2026-10-03 13:21:39,958][starVLA.training.train_starvla][INFO] - Step 1722, Loss: {'train/loss': 0.0006633921293541789, 'train/grad_norm_pre_clip': 0.15215712785720825, 'timing/data': 0.00025037999148480594, 'timing/model': 0.5791922229982447, 'learning_rate/qwen_vl_interface': 7.6548277120315e-06, 'learning_rate/action_model': 7.6548277120315e-05, 'epoch': 0.13}) +[2026-10-03 13:21:40,654][starVLA.training.train_starvla][INFO] - Step 1723, Loss: {'train/loss': 0.0011716856388375163, 'train/grad_norm_pre_clip': 0.36511629819869995, 'timing/data': 0.00023741000040899962, 'timing/model': 0.3895515750045888, 'learning_rate/qwen_vl_interface': 7.652200939557225e-06, 'learning_rate/action_model': 7.652200939557225e-05, 'epoch': 0.13}) +[2026-10-03 13:21:41,352][starVLA.training.train_starvla][INFO] - Step 1724, Loss: {'train/loss': 0.03069387376308441, 'train/grad_norm_pre_clip': 11.859554290771484, 'timing/data': 0.0002917200035881251, 'timing/model': 0.39033225500315893, 'learning_rate/qwen_vl_interface': 7.649573179628881e-06, 'learning_rate/action_model': 7.649573179628881e-05, 'epoch': 0.13}) +[2026-10-03 13:21:42,050][starVLA.training.train_starvla][INFO] - Step 1725, Loss: {'train/loss': 3.9023998397169635e-05, 'train/grad_norm_pre_clip': 0.01630079187452793, 'timing/data': 0.00027645000955089927, 'timing/model': 0.3901995049964171, 'learning_rate/qwen_vl_interface': 7.646944433326645e-06, 'learning_rate/action_model': 7.646944433326643e-05, 'epoch': 0.13}) +[2026-10-03 13:21:42,749][starVLA.training.train_starvla][INFO] - Step 1726, Loss: {'train/loss': 0.33954373002052307, 'train/grad_norm_pre_clip': 148.31451416015625, 'timing/data': 0.00024669000413268805, 'timing/model': 0.39023474499117583, 'learning_rate/qwen_vl_interface': 7.64431470173109e-06, 'learning_rate/action_model': 7.64431470173109e-05, 'epoch': 0.13}) +[2026-10-03 13:21:43,636][starVLA.training.train_starvla][INFO] - Step 1727, Loss: {'train/loss': 0.11074024438858032, 'train/grad_norm_pre_clip': 54.871192932128906, 'timing/data': 0.0002357600023970008, 'timing/model': 0.5799637829913991, 'learning_rate/qwen_vl_interface': 7.641683985923203e-06, 'learning_rate/action_model': 7.641683985923203e-05, 'epoch': 0.13}) +[2026-10-03 13:21:44,334][starVLA.training.train_starvla][INFO] - Step 1728, Loss: {'train/loss': 0.004615819547325373, 'train/grad_norm_pre_clip': 1.3534067869186401, 'timing/data': 0.0002916899975389242, 'timing/model': 0.3902209350053454, 'learning_rate/qwen_vl_interface': 7.63905228698437e-06, 'learning_rate/action_model': 7.63905228698437e-05, 'epoch': 0.13}) +[2026-10-03 13:21:45,030][starVLA.training.train_starvla][INFO] - Step 1729, Loss: {'train/loss': 0.5174597501754761, 'train/grad_norm_pre_clip': 20.337087631225586, 'timing/data': 0.000235440005781129, 'timing/model': 0.38956596500065643, 'learning_rate/qwen_vl_interface': 7.636419605996385e-06, 'learning_rate/action_model': 7.636419605996384e-05, 'epoch': 0.13}) +[2026-10-03 13:21:45,729][starVLA.training.train_starvla][INFO] - Step 1730, Loss: {'train/loss': 0.0008083686698228121, 'train/grad_norm_pre_clip': 0.17578639090061188, 'timing/data': 0.0002449000021442771, 'timing/model': 0.3902138059929712, 'learning_rate/qwen_vl_interface': 7.633785944041444e-06, 'learning_rate/action_model': 7.633785944041444e-05, 'epoch': 0.13}) +[2026-10-03 13:21:46,427][starVLA.training.train_starvla][INFO] - Step 1731, Loss: {'train/loss': 0.7333862781524658, 'train/grad_norm_pre_clip': 32.07790756225586, 'timing/data': 0.00023993999639060348, 'timing/model': 0.38963524499558844, 'learning_rate/qwen_vl_interface': 7.631151302202145e-06, 'learning_rate/action_model': 7.631151302202144e-05, 'epoch': 0.13}) +[2026-10-03 13:21:47,124][starVLA.training.train_starvla][INFO] - Step 1732, Loss: {'train/loss': 0.2950221598148346, 'train/grad_norm_pre_clip': 46.1894645690918, 'timing/data': 0.0001979100052267313, 'timing/model': 0.39001468500646297, 'learning_rate/qwen_vl_interface': 7.628515681561488e-06, 'learning_rate/action_model': 7.628515681561487e-05, 'epoch': 0.13}) +[2026-10-03 13:21:48,012][starVLA.training.train_starvla][INFO] - Step 1733, Loss: {'train/loss': 0.2981588542461395, 'train/grad_norm_pre_clip': 34.572330474853516, 'timing/data': 0.00031772001239005476, 'timing/model': 0.3896219950111117, 'learning_rate/qwen_vl_interface': 7.62587908320288e-06, 'learning_rate/action_model': 7.62587908320288e-05, 'epoch': 0.13}) +[2026-10-03 13:21:48,712][starVLA.training.train_starvla][INFO] - Step 1734, Loss: {'train/loss': 0.0005911333719268441, 'train/grad_norm_pre_clip': 0.15878266096115112, 'timing/data': 0.00036222999915480614, 'timing/model': 0.3901349359948654, 'learning_rate/qwen_vl_interface': 7.6232415082101255e-06, 'learning_rate/action_model': 7.623241508210124e-05, 'epoch': 0.13}) +[2026-10-03 13:21:49,408][starVLA.training.train_starvla][INFO] - Step 1735, Loss: {'train/loss': 0.0005742664798162878, 'train/grad_norm_pre_clip': 0.23962101340293884, 'timing/data': 0.00023801000497769564, 'timing/model': 0.3892372849950334, 'learning_rate/qwen_vl_interface': 7.620602957667432e-06, 'learning_rate/action_model': 7.620602957667432e-05, 'epoch': 0.13}) +[2026-10-03 13:21:50,106][starVLA.training.train_starvla][INFO] - Step 1736, Loss: {'train/loss': 0.1309054046869278, 'train/grad_norm_pre_clip': 14.717636108398438, 'timing/data': 0.00019797901040874422, 'timing/model': 0.3898832160048187, 'learning_rate/qwen_vl_interface': 7.617963432659408e-06, 'learning_rate/action_model': 7.617963432659408e-05, 'epoch': 0.13}) +[2026-10-03 13:21:50,804][starVLA.training.train_starvla][INFO] - Step 1737, Loss: {'train/loss': 0.25867947936058044, 'train/grad_norm_pre_clip': 27.605430603027344, 'timing/data': 0.0002584799949545413, 'timing/model': 0.39002581499516964, 'learning_rate/qwen_vl_interface': 7.615322934271063e-06, 'learning_rate/action_model': 7.615322934271064e-05, 'epoch': 0.13}) +[2026-10-03 13:21:51,690][starVLA.training.train_starvla][INFO] - Step 1738, Loss: {'train/loss': 0.001288929139263928, 'train/grad_norm_pre_clip': 0.3854072391986847, 'timing/data': 0.0002784999960567802, 'timing/model': 0.5769970339897554, 'learning_rate/qwen_vl_interface': 7.6126814635878055e-06, 'learning_rate/action_model': 7.612681463587805e-05, 'epoch': 0.13}) +[2026-10-03 13:21:52,386][starVLA.training.train_starvla][INFO] - Step 1739, Loss: {'train/loss': 0.0025377138517796993, 'train/grad_norm_pre_clip': 0.39478299021720886, 'timing/data': 0.0002681500045582652, 'timing/model': 0.389372755002114, 'learning_rate/qwen_vl_interface': 7.6100390216954445e-06, 'learning_rate/action_model': 7.610039021695445e-05, 'epoch': 0.13}) +[2026-10-03 13:21:53,082][starVLA.training.train_starvla][INFO] - Step 1740, Loss: {'train/loss': 0.00014572296640835702, 'train/grad_norm_pre_clip': 0.027395347133278847, 'timing/data': 0.00023686999338679016, 'timing/model': 0.3893511350033805, 'learning_rate/qwen_vl_interface': 7.607395609680189e-06, 'learning_rate/action_model': 7.60739560968019e-05, 'epoch': 0.13}) +[2026-10-03 13:21:53,779][starVLA.training.train_starvla][INFO] - Step 1741, Loss: {'train/loss': 0.02911173552274704, 'train/grad_norm_pre_clip': 8.030923843383789, 'timing/data': 0.0002850900054909289, 'timing/model': 0.39000600499275606, 'learning_rate/qwen_vl_interface': 7.604751228628644e-06, 'learning_rate/action_model': 7.604751228628644e-05, 'epoch': 0.13}) +[2026-10-03 13:21:54,478][starVLA.training.train_starvla][INFO] - Step 1742, Loss: {'train/loss': 0.3204149305820465, 'train/grad_norm_pre_clip': 21.64404296875, 'timing/data': 0.00031099999614525586, 'timing/model': 0.39005997600906994, 'learning_rate/qwen_vl_interface': 7.602105879627814e-06, 'learning_rate/action_model': 7.602105879627814e-05, 'epoch': 0.13}) +[2026-10-03 13:21:55,174][starVLA.training.train_starvla][INFO] - Step 1743, Loss: {'train/loss': 0.011412781663239002, 'train/grad_norm_pre_clip': 5.361827850341797, 'timing/data': 0.00032789999386295676, 'timing/model': 0.38979035599913914, 'learning_rate/qwen_vl_interface': 7.599459563765107e-06, 'learning_rate/action_model': 7.599459563765106e-05, 'epoch': 0.13}) +[2026-10-03 13:21:56,057][starVLA.training.train_starvla][INFO] - Step 1744, Loss: {'train/loss': 0.030846377834677696, 'train/grad_norm_pre_clip': 27.610349655151367, 'timing/data': 0.00022431000252254307, 'timing/model': 0.38927455501107033, 'learning_rate/qwen_vl_interface': 7.596812282128319e-06, 'learning_rate/action_model': 7.596812282128319e-05, 'epoch': 0.13}) +[2026-10-03 13:21:56,753][starVLA.training.train_starvla][INFO] - Step 1745, Loss: {'train/loss': 0.0040720608085393906, 'train/grad_norm_pre_clip': 0.8076092004776001, 'timing/data': 0.00023252000391948968, 'timing/model': 0.38916036499722395, 'learning_rate/qwen_vl_interface': 7.594164035805648e-06, 'learning_rate/action_model': 7.594164035805648e-05, 'epoch': 0.13}) +[2026-10-03 13:21:57,453][starVLA.training.train_starvla][INFO] - Step 1746, Loss: {'train/loss': 0.03733492270112038, 'train/grad_norm_pre_clip': 50.955936431884766, 'timing/data': 0.00031104001391213387, 'timing/model': 0.38987585600989405, 'learning_rate/qwen_vl_interface': 7.59151482588569e-06, 'learning_rate/action_model': 7.59151482588569e-05, 'epoch': 0.13}) +[2026-10-03 13:21:58,149][starVLA.training.train_starvla][INFO] - Step 1747, Loss: {'train/loss': 0.025420870631933212, 'train/grad_norm_pre_clip': 61.74797439575195, 'timing/data': 0.0003304100100649521, 'timing/model': 0.38909574499120936, 'learning_rate/qwen_vl_interface': 7.588864653457433e-06, 'learning_rate/action_model': 7.588864653457432e-05, 'epoch': 0.13}) +[2026-10-03 13:21:58,846][starVLA.training.train_starvla][INFO] - Step 1748, Loss: {'train/loss': 0.004673958756029606, 'train/grad_norm_pre_clip': 0.8193669319152832, 'timing/data': 0.00019039999460801482, 'timing/model': 0.3897560449986486, 'learning_rate/qwen_vl_interface': 7.5862135196102634e-06, 'learning_rate/action_model': 7.586213519610263e-05, 'epoch': 0.13}) +[2026-10-03 13:21:59,731][starVLA.training.train_starvla][INFO] - Step 1749, Loss: {'train/loss': 0.004209127742797136, 'train/grad_norm_pre_clip': 0.7515049576759338, 'timing/data': 0.0002349299902562052, 'timing/model': 0.3883599859982496, 'learning_rate/qwen_vl_interface': 7.583561425433962e-06, 'learning_rate/action_model': 7.583561425433962e-05, 'epoch': 0.13}) +[2026-10-03 13:22:00,429][starVLA.training.train_starvla][INFO] - Step 1750, Loss: {'train/loss': 0.16424548625946045, 'train/grad_norm_pre_clip': 52.175376892089844, 'timing/data': 0.00032284000189974904, 'timing/model': 0.38980730599723756, 'learning_rate/qwen_vl_interface': 7.580908372018706e-06, 'learning_rate/action_model': 7.580908372018705e-05, 'epoch': 0.13}) +[2026-10-03 13:22:01,124][starVLA.training.train_starvla][INFO] - Step 1751, Loss: {'train/loss': 0.00012624448572751135, 'train/grad_norm_pre_clip': 0.07975827902555466, 'timing/data': 0.00019623999833129346, 'timing/model': 0.38892649600165896, 'learning_rate/qwen_vl_interface': 7.5782543604550655e-06, 'learning_rate/action_model': 7.578254360455064e-05, 'epoch': 0.13}) +[2026-10-03 13:22:01,820][starVLA.training.train_starvla][INFO] - Step 1752, Loss: {'train/loss': 0.00012678517668973655, 'train/grad_norm_pre_clip': 0.09320521354675293, 'timing/data': 0.00019193999469280243, 'timing/model': 0.3893693450081628, 'learning_rate/qwen_vl_interface': 7.575599391834e-06, 'learning_rate/action_model': 7.575599391834e-05, 'epoch': 0.13}) +[2026-10-03 13:22:02,516][starVLA.training.train_starvla][INFO] - Step 1753, Loss: {'train/loss': 0.0009810416959226131, 'train/grad_norm_pre_clip': 0.4045712649822235, 'timing/data': 0.0002550890058046207, 'timing/model': 0.3889083559915889, 'learning_rate/qwen_vl_interface': 7.572943467246871e-06, 'learning_rate/action_model': 7.57294346724687e-05, 'epoch': 0.13}) +[2026-10-03 13:22:03,402][starVLA.training.train_starvla][INFO] - Step 1754, Loss: {'train/loss': 0.16799309849739075, 'train/grad_norm_pre_clip': 64.34734344482422, 'timing/data': 0.00019949999114032835, 'timing/model': 0.5793325539998477, 'learning_rate/qwen_vl_interface': 7.570286587785428e-06, 'learning_rate/action_model': 7.570286587785428e-05, 'epoch': 0.13}) +[2026-10-03 13:22:04,094][starVLA.training.train_starvla][INFO] - Step 1755, Loss: {'train/loss': 0.12151428312063217, 'train/grad_norm_pre_clip': 132.84295654296875, 'timing/data': 0.000321680010529235, 'timing/model': 0.38733238499844447, 'learning_rate/qwen_vl_interface': 7.567628754541815e-06, 'learning_rate/action_model': 7.567628754541814e-05, 'epoch': 0.13}) +[2026-10-03 13:22:04,786][starVLA.training.train_starvla][INFO] - Step 1756, Loss: {'train/loss': 0.0013923445949330926, 'train/grad_norm_pre_clip': 2.0882930755615234, 'timing/data': 0.00023929000599309802, 'timing/model': 0.387361975008389, 'learning_rate/qwen_vl_interface': 7.564969968608565e-06, 'learning_rate/action_model': 7.564969968608565e-05, 'epoch': 0.13}) +[2026-10-03 13:22:05,480][starVLA.training.train_starvla][INFO] - Step 1757, Loss: {'train/loss': 0.27391543984413147, 'train/grad_norm_pre_clip': 46.82093048095703, 'timing/data': 0.00024334900081157684, 'timing/model': 0.3878229460096918, 'learning_rate/qwen_vl_interface': 7.562310231078605e-06, 'learning_rate/action_model': 7.562310231078604e-05, 'epoch': 0.13}) +[2026-10-03 13:22:06,176][starVLA.training.train_starvla][INFO] - Step 1758, Loss: {'train/loss': 0.00043371220817789435, 'train/grad_norm_pre_clip': 0.31634047627449036, 'timing/data': 0.0002022700064117089, 'timing/model': 0.3883396250021178, 'learning_rate/qwen_vl_interface': 7.559649543045252e-06, 'learning_rate/action_model': 7.559649543045252e-05, 'epoch': 0.13}) +[2026-10-03 13:22:06,869][starVLA.training.train_starvla][INFO] - Step 1759, Loss: {'train/loss': 2.1700147044612095e-05, 'train/grad_norm_pre_clip': 0.006261200178414583, 'timing/data': 0.00021590999676845968, 'timing/model': 0.3881010250042891, 'learning_rate/qwen_vl_interface': 7.556987905602215e-06, 'learning_rate/action_model': 7.556987905602215e-05, 'epoch': 0.13}) +[2026-10-03 13:22:07,750][starVLA.training.train_starvla][INFO] - Step 1760, Loss: {'train/loss': 1.557862742629368e-05, 'train/grad_norm_pre_clip': 0.004169847816228867, 'timing/data': 0.00021196001034695655, 'timing/model': 0.3877535549981985, 'learning_rate/qwen_vl_interface': 7.554325319843593e-06, 'learning_rate/action_model': 7.554325319843592e-05, 'epoch': 0.13}) +[2026-10-03 13:22:08,442][starVLA.training.train_starvla][INFO] - Step 1761, Loss: {'train/loss': 0.00022149921278469265, 'train/grad_norm_pre_clip': 0.04939493536949158, 'timing/data': 0.00031524000223726034, 'timing/model': 0.38727449500584044, 'learning_rate/qwen_vl_interface': 7.551661786863874e-06, 'learning_rate/action_model': 7.551661786863874e-05, 'epoch': 0.13}) +[2026-10-03 13:22:09,137][starVLA.training.train_starvla][INFO] - Step 1762, Loss: {'train/loss': 0.003825938096269965, 'train/grad_norm_pre_clip': 4.675686359405518, 'timing/data': 0.00033183000050485134, 'timing/model': 0.38756466600170825, 'learning_rate/qwen_vl_interface': 7.548997307757936e-06, 'learning_rate/action_model': 7.548997307757935e-05, 'epoch': 0.13}) +[2026-10-03 13:22:09,829][starVLA.training.train_starvla][INFO] - Step 1763, Loss: {'train/loss': 0.21776846051216125, 'train/grad_norm_pre_clip': 81.55426788330078, 'timing/data': 0.000403770012781024, 'timing/model': 0.3875009659968782, 'learning_rate/qwen_vl_interface': 7.546331883621044e-06, 'learning_rate/action_model': 7.546331883621044e-05, 'epoch': 0.13}) +[2026-10-03 13:22:10,521][starVLA.training.train_starvla][INFO] - Step 1764, Loss: {'train/loss': 0.338926762342453, 'train/grad_norm_pre_clip': 63.67527389526367, 'timing/data': 0.00032577000092715025, 'timing/model': 0.38720700598787516, 'learning_rate/qwen_vl_interface': 7.543665515548857e-06, 'learning_rate/action_model': 7.543665515548856e-05, 'epoch': 0.13}) +[2026-10-03 13:22:11,407][starVLA.training.train_starvla][INFO] - Step 1765, Loss: {'train/loss': 0.00012611890269909054, 'train/grad_norm_pre_clip': 0.10416680574417114, 'timing/data': 0.00027119999867863953, 'timing/model': 0.3877688560023671, 'learning_rate/qwen_vl_interface': 7.540998204637413e-06, 'learning_rate/action_model': 7.540998204637413e-05, 'epoch': 0.13}) +[2026-10-03 13:22:12,103][starVLA.training.train_starvla][INFO] - Step 1766, Loss: {'train/loss': 0.0020644906908273697, 'train/grad_norm_pre_clip': 0.872958242893219, 'timing/data': 0.00020095999934710562, 'timing/model': 0.38816399598727, 'learning_rate/qwen_vl_interface': 7.538329951983148e-06, 'learning_rate/action_model': 7.538329951983147e-05, 'epoch': 0.13}) +[2026-10-03 13:22:12,797][starVLA.training.train_starvla][INFO] - Step 1767, Loss: {'train/loss': 4.770703162648715e-05, 'train/grad_norm_pre_clip': 0.008430157788097858, 'timing/data': 0.0001936599874170497, 'timing/model': 0.38863757500075735, 'learning_rate/qwen_vl_interface': 7.535660758682877e-06, 'learning_rate/action_model': 7.535660758682876e-05, 'epoch': 0.13}) +[2026-10-03 13:22:13,491][starVLA.training.train_starvla][INFO] - Step 1768, Loss: {'train/loss': 0.7959335446357727, 'train/grad_norm_pre_clip': 140.28599548339844, 'timing/data': 0.0003119899920420721, 'timing/model': 0.3879061749903485, 'learning_rate/qwen_vl_interface': 7.532990625833806e-06, 'learning_rate/action_model': 7.532990625833805e-05, 'epoch': 0.13}) +[2026-10-03 13:22:14,184][starVLA.training.train_starvla][INFO] - Step 1769, Loss: {'train/loss': 0.057450417429208755, 'train/grad_norm_pre_clip': 91.49513244628906, 'timing/data': 0.000212459999602288, 'timing/model': 0.3873825149930781, 'learning_rate/qwen_vl_interface': 7.5303195545335225e-06, 'learning_rate/action_model': 7.530319554533522e-05, 'epoch': 0.13}) +[2026-10-03 13:22:15,069][starVLA.training.train_starvla][INFO] - Step 1770, Loss: {'train/loss': 9.233810851583257e-05, 'train/grad_norm_pre_clip': 0.01864507794380188, 'timing/data': 0.0002324599918210879, 'timing/model': 0.3877260550070787, 'learning_rate/qwen_vl_interface': 7.527647545880007e-06, 'learning_rate/action_model': 7.527647545880007e-05, 'epoch': 0.13}) +[2026-10-03 13:22:15,762][starVLA.training.train_starvla][INFO] - Step 1771, Loss: {'train/loss': 0.00813392922282219, 'train/grad_norm_pre_clip': 9.737525939941406, 'timing/data': 0.0001892099971882999, 'timing/model': 0.38781617500353605, 'learning_rate/qwen_vl_interface': 7.52497460097162e-06, 'learning_rate/action_model': 7.52497460097162e-05, 'epoch': 0.13}) +[2026-10-03 13:22:16,454][starVLA.training.train_starvla][INFO] - Step 1772, Loss: {'train/loss': 0.007780758664011955, 'train/grad_norm_pre_clip': 25.87958335876465, 'timing/data': 0.00023660999431740493, 'timing/model': 0.38759408600162715, 'learning_rate/qwen_vl_interface': 7.522300720907105e-06, 'learning_rate/action_model': 7.522300720907105e-05, 'epoch': 0.13}) +[2026-10-03 13:22:17,148][starVLA.training.train_starvla][INFO] - Step 1773, Loss: {'train/loss': 2.4211680283769965e-05, 'train/grad_norm_pre_clip': 0.011757870204746723, 'timing/data': 0.00029417000769171864, 'timing/model': 0.38777851598570123, 'learning_rate/qwen_vl_interface': 7.519625906785601e-06, 'learning_rate/action_model': 7.5196259067856e-05, 'epoch': 0.13}) +[2026-10-03 13:22:17,844][starVLA.training.train_starvla][INFO] - Step 1774, Loss: {'train/loss': 7.763296707707923e-06, 'train/grad_norm_pre_clip': 0.006705339532345533, 'timing/data': 0.0003808699984801933, 'timing/model': 0.3886867959954543, 'learning_rate/qwen_vl_interface': 7.516950159706614e-06, 'learning_rate/action_model': 7.516950159706614e-05, 'epoch': 0.13}) +[2026-10-03 13:22:18,729][starVLA.training.train_starvla][INFO] - Step 1775, Loss: {'train/loss': 3.092444967478514e-05, 'train/grad_norm_pre_clip': 0.015876762568950653, 'timing/data': 0.00019357999553903937, 'timing/model': 0.5797372240049299, 'learning_rate/qwen_vl_interface': 7.514273480770047e-06, 'learning_rate/action_model': 7.514273480770047e-05, 'epoch': 0.13}) +[2026-10-03 13:22:19,422][starVLA.training.train_starvla][INFO] - Step 1776, Loss: {'train/loss': 0.27958032488822937, 'train/grad_norm_pre_clip': 61.73969268798828, 'timing/data': 0.00025431001267861575, 'timing/model': 0.38760859599278774, 'learning_rate/qwen_vl_interface': 7.511595871076181e-06, 'learning_rate/action_model': 7.511595871076181e-05, 'epoch': 0.13}) +[2026-10-03 13:22:20,117][starVLA.training.train_starvla][INFO] - Step 1777, Loss: {'train/loss': 2.7864939511346165e-06, 'train/grad_norm_pre_clip': 0.0009392325882799923, 'timing/data': 0.00019527999393176287, 'timing/model': 0.3882141249923734, 'learning_rate/qwen_vl_interface': 7.50891733172568e-06, 'learning_rate/action_model': 7.50891733172568e-05, 'epoch': 0.13}) +[2026-10-03 13:22:20,812][starVLA.training.train_starvla][INFO] - Step 1778, Loss: {'train/loss': 2.4959290385595523e-06, 'train/grad_norm_pre_clip': 0.0011152341030538082, 'timing/data': 0.00019725000311154872, 'timing/model': 0.38867578499775846, 'learning_rate/qwen_vl_interface': 7.506237863819593e-06, 'learning_rate/action_model': 7.506237863819593e-05, 'epoch': 0.13}) +[2026-10-03 13:22:21,507][starVLA.training.train_starvla][INFO] - Step 1779, Loss: {'train/loss': 0.6213408708572388, 'train/grad_norm_pre_clip': 168.9976348876953, 'timing/data': 0.0003039800067199394, 'timing/model': 0.388273096003104, 'learning_rate/qwen_vl_interface': 7.503557468459345e-06, 'learning_rate/action_model': 7.503557468459344e-05, 'epoch': 0.13}) +[2026-10-03 13:22:22,393][starVLA.training.train_starvla][INFO] - Step 1780, Loss: {'train/loss': 3.4272370612598024e-06, 'train/grad_norm_pre_clip': 0.00318434601649642, 'timing/data': 0.000287729999399744, 'timing/model': 0.5809476340073161, 'learning_rate/qwen_vl_interface': 7.5008761467467435e-06, 'learning_rate/action_model': 7.500876146746743e-05, 'epoch': 0.13}) +[2026-10-03 13:22:23,087][starVLA.training.train_starvla][INFO] - Step 1781, Loss: {'train/loss': 0.004188417457044125, 'train/grad_norm_pre_clip': 4.482509136199951, 'timing/data': 0.0001897999900393188, 'timing/model': 0.3883157160016708, 'learning_rate/qwen_vl_interface': 7.498193899783983e-06, 'learning_rate/action_model': 7.498193899783983e-05, 'epoch': 0.13}) +[2026-10-03 13:22:23,784][starVLA.training.train_starvla][INFO] - Step 1782, Loss: {'train/loss': 0.0012131419498473406, 'train/grad_norm_pre_clip': 1.206787347793579, 'timing/data': 0.0002535700041335076, 'timing/model': 0.3895407350064488, 'learning_rate/qwen_vl_interface': 7.495510728673631e-06, 'learning_rate/action_model': 7.495510728673632e-05, 'epoch': 0.13}) +[2026-10-03 13:22:24,480][starVLA.training.train_starvla][INFO] - Step 1783, Loss: {'train/loss': 0.0001783025509212166, 'train/grad_norm_pre_clip': 0.03864738345146179, 'timing/data': 0.00019639999663922936, 'timing/model': 0.38969013500900473, 'learning_rate/qwen_vl_interface': 7.492826634518642e-06, 'learning_rate/action_model': 7.492826634518641e-05, 'epoch': 0.13}) +[2026-10-03 13:22:25,175][starVLA.training.train_starvla][INFO] - Step 1784, Loss: {'train/loss': 0.00010607727017486468, 'train/grad_norm_pre_clip': 0.03591948747634888, 'timing/data': 0.00033418000384699553, 'timing/model': 0.38824666499567684, 'learning_rate/qwen_vl_interface': 7.4901416184223404e-06, 'learning_rate/action_model': 7.49014161842234e-05, 'epoch': 0.13}) +[2026-10-03 13:22:25,869][starVLA.training.train_starvla][INFO] - Step 1785, Loss: {'train/loss': 1.5124635410757037e-06, 'train/grad_norm_pre_clip': 0.00041112149483524263, 'timing/data': 0.0002392799942754209, 'timing/model': 0.3881822159892181, 'learning_rate/qwen_vl_interface': 7.487455681488437e-06, 'learning_rate/action_model': 7.487455681488436e-05, 'epoch': 0.13}) +[2026-10-03 13:22:26,756][starVLA.training.train_starvla][INFO] - Step 1786, Loss: {'train/loss': 1.356002712782356e-06, 'train/grad_norm_pre_clip': 0.0004228928592056036, 'timing/data': 0.000281360000371933, 'timing/model': 0.3882975660089869, 'learning_rate/qwen_vl_interface': 7.484768824821018e-06, 'learning_rate/action_model': 7.484768824821018e-05, 'epoch': 0.13}) +[2026-10-03 13:22:27,453][starVLA.training.train_starvla][INFO] - Step 1787, Loss: {'train/loss': 0.00028465050854720175, 'train/grad_norm_pre_clip': 0.8172360062599182, 'timing/data': 0.00021354001364670694, 'timing/model': 0.3874633359955624, 'learning_rate/qwen_vl_interface': 7.482081049524551e-06, 'learning_rate/action_model': 7.482081049524551e-05, 'epoch': 0.13}) +[2026-10-03 13:22:28,148][starVLA.training.train_starvla][INFO] - Step 1788, Loss: {'train/loss': 4.5798875362379476e-05, 'train/grad_norm_pre_clip': 0.011363724246621132, 'timing/data': 0.0002438400115352124, 'timing/model': 0.38839917599398177, 'learning_rate/qwen_vl_interface': 7.479392356703878e-06, 'learning_rate/action_model': 7.479392356703878e-05, 'epoch': 0.13}) +[2026-10-03 13:22:28,841][starVLA.training.train_starvla][INFO] - Step 1789, Loss: {'train/loss': 4.47028332928312e-06, 'train/grad_norm_pre_clip': 0.0011392151936888695, 'timing/data': 0.00019693000649567693, 'timing/model': 0.3875110059889266, 'learning_rate/qwen_vl_interface': 7.476702747464219e-06, 'learning_rate/action_model': 7.476702747464219e-05, 'epoch': 0.13}) +[2026-10-03 13:22:29,537][starVLA.training.train_starvla][INFO] - Step 1790, Loss: {'train/loss': 4.6267732614069246e-06, 'train/grad_norm_pre_clip': 0.0026801733765751123, 'timing/data': 0.0003260999947087839, 'timing/model': 0.38872715599427465, 'learning_rate/qwen_vl_interface': 7.474012222911169e-06, 'learning_rate/action_model': 7.474012222911169e-05, 'epoch': 0.13}) +[2026-10-03 13:22:30,420][starVLA.training.train_starvla][INFO] - Step 1791, Loss: {'train/loss': 0.9128040671348572, 'train/grad_norm_pre_clip': 134.8260498046875, 'timing/data': 0.00019128000712953508, 'timing/model': 0.5774277849995997, 'learning_rate/qwen_vl_interface': 7.471320784150703e-06, 'learning_rate/action_model': 7.471320784150703e-05, 'epoch': 0.13}) +[2026-10-03 13:22:31,114][starVLA.training.train_starvla][INFO] - Step 1792, Loss: {'train/loss': 5.0243539590155706e-05, 'train/grad_norm_pre_clip': 0.009775628335773945, 'timing/data': 0.00019196000357624143, 'timing/model': 0.38866681599756703, 'learning_rate/qwen_vl_interface': 7.468628432289169e-06, 'learning_rate/action_model': 7.468628432289168e-05, 'epoch': 0.13}) +[2026-10-03 13:22:31,808][starVLA.training.train_starvla][INFO] - Step 1793, Loss: {'train/loss': 4.325167174101807e-05, 'train/grad_norm_pre_clip': 0.008565167896449566, 'timing/data': 0.00031596999906469136, 'timing/model': 0.3877244460018119, 'learning_rate/qwen_vl_interface': 7.465935168433292e-06, 'learning_rate/action_model': 7.465935168433292e-05, 'epoch': 0.13}) +[2026-10-03 13:22:32,504][starVLA.training.train_starvla][INFO] - Step 1794, Loss: {'train/loss': 0.000156545196659863, 'train/grad_norm_pre_clip': 0.4661094844341278, 'timing/data': 0.0003839299897663295, 'timing/model': 0.38894422599696554, 'learning_rate/qwen_vl_interface': 7.463240993690168e-06, 'learning_rate/action_model': 7.463240993690168e-05, 'epoch': 0.13}) +[2026-10-03 13:22:33,197][starVLA.training.train_starvla][INFO] - Step 1795, Loss: {'train/loss': 0.00019433864508755505, 'train/grad_norm_pre_clip': 0.1867617964744568, 'timing/data': 0.00032716999703552574, 'timing/model': 0.38795184600166976, 'learning_rate/qwen_vl_interface': 7.460545909167276e-06, 'learning_rate/action_model': 7.460545909167275e-05, 'epoch': 0.13}) +[2026-10-03 13:22:33,890][starVLA.training.train_starvla][INFO] - Step 1796, Loss: {'train/loss': 0.00018884113524109125, 'train/grad_norm_pre_clip': 0.10991819202899933, 'timing/data': 0.0001930599974002689, 'timing/model': 0.3883826059900457, 'learning_rate/qwen_vl_interface': 7.457849915972459e-06, 'learning_rate/action_model': 7.457849915972458e-05, 'epoch': 0.13}) +[2026-10-03 13:22:34,775][starVLA.training.train_starvla][INFO] - Step 1797, Loss: {'train/loss': 2.8684430617431644e-06, 'train/grad_norm_pre_clip': 0.002392613561823964, 'timing/data': 0.00032373999420087785, 'timing/model': 0.38790559698827565, 'learning_rate/qwen_vl_interface': 7.455153015213941e-06, 'learning_rate/action_model': 7.45515301521394e-05, 'epoch': 0.13}) +[2026-10-03 13:22:35,472][starVLA.training.train_starvla][INFO] - Step 1798, Loss: {'train/loss': 3.179917621309869e-05, 'train/grad_norm_pre_clip': 0.007485790178179741, 'timing/data': 0.00036253000143915415, 'timing/model': 0.38850018600351177, 'learning_rate/qwen_vl_interface': 7.452455208000311e-06, 'learning_rate/action_model': 7.452455208000311e-05, 'epoch': 0.13}) +[2026-10-03 13:22:36,166][starVLA.training.train_starvla][INFO] - Step 1799, Loss: {'train/loss': 0.7290663719177246, 'train/grad_norm_pre_clip': 35.1847038269043, 'timing/data': 0.00019781998707912862, 'timing/model': 0.38833328599866945, 'learning_rate/qwen_vl_interface': 7.449756495440541e-06, 'learning_rate/action_model': 7.44975649544054e-05, 'epoch': 0.13}) +[2026-10-03 13:22:36,861][starVLA.training.train_starvla][INFO] - Step 1800, Loss: {'train/loss': 0.2564159035682678, 'train/grad_norm_pre_clip': 202.74647521972656, 'timing/data': 0.00020949001191183925, 'timing/model': 0.38888542499626055, 'learning_rate/qwen_vl_interface': 7.447056878643969e-06, 'learning_rate/action_model': 7.447056878643968e-05, 'epoch': 0.13}) +[2026-10-03 13:22:37,555][starVLA.training.train_starvla][INFO] - Step 1801, Loss: {'train/loss': 7.710701174801216e-05, 'train/grad_norm_pre_clip': 0.0393088161945343, 'timing/data': 0.00033026999153662473, 'timing/model': 0.38796375499805436, 'learning_rate/qwen_vl_interface': 7.444356358720303e-06, 'learning_rate/action_model': 7.444356358720302e-05, 'epoch': 0.13}) +[2026-10-03 13:22:38,441][starVLA.training.train_starvla][INFO] - Step 1802, Loss: {'train/loss': 0.4239009618759155, 'train/grad_norm_pre_clip': 67.61490631103516, 'timing/data': 0.00022767001064494252, 'timing/model': 0.3882532860006904, 'learning_rate/qwen_vl_interface': 7.4416549367796255e-06, 'learning_rate/action_model': 7.441654936779625e-05, 'epoch': 0.13}) +[2026-10-03 13:22:39,135][starVLA.training.train_starvla][INFO] - Step 1803, Loss: {'train/loss': 0.0002634126867633313, 'train/grad_norm_pre_clip': 0.4053841233253479, 'timing/data': 0.00024067000776994973, 'timing/model': 0.3887389349984005, 'learning_rate/qwen_vl_interface': 7.4389526139323895e-06, 'learning_rate/action_model': 7.438952613932389e-05, 'epoch': 0.13}) +[2026-10-03 13:22:39,829][starVLA.training.train_starvla][INFO] - Step 1804, Loss: {'train/loss': 0.20125974714756012, 'train/grad_norm_pre_clip': 194.84059143066406, 'timing/data': 0.00024300000222865492, 'timing/model': 0.3879831559897866, 'learning_rate/qwen_vl_interface': 7.436249391289419e-06, 'learning_rate/action_model': 7.436249391289418e-05, 'epoch': 0.13}) +[2026-10-03 13:22:40,522][starVLA.training.train_starvla][INFO] - Step 1805, Loss: {'train/loss': 2.3333237550104968e-05, 'train/grad_norm_pre_clip': 0.0038446763064712286, 'timing/data': 0.0002386999985901639, 'timing/model': 0.38771998599986546, 'learning_rate/qwen_vl_interface': 7.433545269961906e-06, 'learning_rate/action_model': 7.433545269961906e-05, 'epoch': 0.13}) +[2026-10-03 13:22:41,218][starVLA.training.train_starvla][INFO] - Step 1806, Loss: {'train/loss': 0.0005311415297910571, 'train/grad_norm_pre_clip': 0.6477837562561035, 'timing/data': 0.0004009399999631569, 'timing/model': 0.3880533350020414, 'learning_rate/qwen_vl_interface': 7.430840251061411e-06, 'learning_rate/action_model': 7.430840251061411e-05, 'epoch': 0.13}) +[2026-10-03 13:22:42,104][starVLA.training.train_starvla][INFO] - Step 1807, Loss: {'train/loss': 0.00037910364335402846, 'train/grad_norm_pre_clip': 0.26024261116981506, 'timing/data': 0.0002994899987243116, 'timing/model': 0.5814771129953442, 'learning_rate/qwen_vl_interface': 7.428134335699866e-06, 'learning_rate/action_model': 7.428134335699865e-05, 'epoch': 0.13}) +[2026-10-03 13:22:42,798][starVLA.training.train_starvla][INFO] - Step 1808, Loss: {'train/loss': 0.00016335328109562397, 'train/grad_norm_pre_clip': 0.08135276287794113, 'timing/data': 0.00019257000531069934, 'timing/model': 0.3883118350058794, 'learning_rate/qwen_vl_interface': 7.425427524989572e-06, 'learning_rate/action_model': 7.425427524989572e-05, 'epoch': 0.13}) +[2026-10-03 13:22:43,493][starVLA.training.train_starvla][INFO] - Step 1809, Loss: {'train/loss': 7.256550452439114e-05, 'train/grad_norm_pre_clip': 0.011118721216917038, 'timing/data': 0.00032557999657001346, 'timing/model': 0.3880980260000797, 'learning_rate/qwen_vl_interface': 7.422719820043194e-06, 'learning_rate/action_model': 7.422719820043193e-05, 'epoch': 0.13}) +[2026-10-03 13:22:44,190][starVLA.training.train_starvla][INFO] - Step 1810, Loss: {'train/loss': 0.22683994472026825, 'train/grad_norm_pre_clip': 65.38739776611328, 'timing/data': 0.00020392000442370772, 'timing/model': 0.3890620960009983, 'learning_rate/qwen_vl_interface': 7.420011221973771e-06, 'learning_rate/action_model': 7.420011221973771e-05, 'epoch': 0.13}) +[2026-10-03 13:22:44,885][starVLA.training.train_starvla][INFO] - Step 1811, Loss: {'train/loss': 3.697911961353384e-05, 'train/grad_norm_pre_clip': 0.009132212027907372, 'timing/data': 0.00024018000112846494, 'timing/model': 0.38878463600121904, 'learning_rate/qwen_vl_interface': 7.417301731894702e-06, 'learning_rate/action_model': 7.417301731894701e-05, 'epoch': 0.13}) +[2026-10-03 13:22:45,580][starVLA.training.train_starvla][INFO] - Step 1812, Loss: {'train/loss': 6.146440864540637e-05, 'train/grad_norm_pre_clip': 0.0170504841953516, 'timing/data': 0.0002803899988066405, 'timing/model': 0.3887629359960556, 'learning_rate/qwen_vl_interface': 7.414591350919756e-06, 'learning_rate/action_model': 7.414591350919756e-05, 'epoch': 0.13}) +[2026-10-03 13:22:46,464][starVLA.training.train_starvla][INFO] - Step 1813, Loss: {'train/loss': 0.012507151812314987, 'train/grad_norm_pre_clip': 38.831539154052734, 'timing/data': 0.00025502000062260777, 'timing/model': 0.38820718599890824, 'learning_rate/qwen_vl_interface': 7.4118800801630715e-06, 'learning_rate/action_model': 7.411880080163072e-05, 'epoch': 0.13}) +[2026-10-03 13:22:47,161][starVLA.training.train_starvla][INFO] - Step 1814, Loss: {'train/loss': 0.16607844829559326, 'train/grad_norm_pre_clip': 101.45872497558594, 'timing/data': 0.00021034000383224338, 'timing/model': 0.38883695600088686, 'learning_rate/qwen_vl_interface': 7.409167920739146e-06, 'learning_rate/action_model': 7.409167920739147e-05, 'epoch': 0.13}) +[2026-10-03 13:22:47,855][starVLA.training.train_starvla][INFO] - Step 1815, Loss: {'train/loss': 0.0047195968218147755, 'train/grad_norm_pre_clip': 4.519010066986084, 'timing/data': 0.0002394600014667958, 'timing/model': 0.38819214499380905, 'learning_rate/qwen_vl_interface': 7.406454873762848e-06, 'learning_rate/action_model': 7.406454873762848e-05, 'epoch': 0.13}) +[2026-10-03 13:22:48,550][starVLA.training.train_starvla][INFO] - Step 1816, Loss: {'train/loss': 3.243557148380205e-05, 'train/grad_norm_pre_clip': 0.033214837312698364, 'timing/data': 0.0002295099984621629, 'timing/model': 0.3885757949901745, 'learning_rate/qwen_vl_interface': 7.403740940349408e-06, 'learning_rate/action_model': 7.403740940349408e-05, 'epoch': 0.13}) +[2026-10-03 13:22:49,245][starVLA.training.train_starvla][INFO] - Step 1817, Loss: {'train/loss': 0.00014988855400588363, 'train/grad_norm_pre_clip': 0.05772889032959938, 'timing/data': 0.0002896000078180805, 'timing/model': 0.3880634259985527, 'learning_rate/qwen_vl_interface': 7.401026121614421e-06, 'learning_rate/action_model': 7.40102612161442e-05, 'epoch': 0.13}) +[2026-10-03 13:22:50,132][starVLA.training.train_starvla][INFO] - Step 1818, Loss: {'train/loss': 6.144505459815264e-05, 'train/grad_norm_pre_clip': 0.015305954962968826, 'timing/data': 0.0002017700026044622, 'timing/model': 0.38836762600112706, 'learning_rate/qwen_vl_interface': 7.398310418673846e-06, 'learning_rate/action_model': 7.398310418673846e-05, 'epoch': 0.13}) +[2026-10-03 13:22:50,824][starVLA.training.train_starvla][INFO] - Step 1819, Loss: {'train/loss': 8.42273875605315e-05, 'train/grad_norm_pre_clip': 0.2968139350414276, 'timing/data': 0.00026602999423630536, 'timing/model': 0.3881095760007156, 'learning_rate/qwen_vl_interface': 7.395593832644007e-06, 'learning_rate/action_model': 7.395593832644007e-05, 'epoch': 0.13}) +[2026-10-03 13:22:51,520][starVLA.training.train_starvla][INFO] - Step 1820, Loss: {'train/loss': 0.5953274965286255, 'train/grad_norm_pre_clip': 15.355280876159668, 'timing/data': 0.00032811000710353255, 'timing/model': 0.38832969599752687, 'learning_rate/qwen_vl_interface': 7.39287636464159e-06, 'learning_rate/action_model': 7.39287636464159e-05, 'epoch': 0.13}) +[2026-10-03 13:22:52,214][starVLA.training.train_starvla][INFO] - Step 1821, Loss: {'train/loss': 0.427580326795578, 'train/grad_norm_pre_clip': 76.65052032470703, 'timing/data': 0.0002476600056979805, 'timing/model': 0.3882374050008366, 'learning_rate/qwen_vl_interface': 7.3901580157836425e-06, 'learning_rate/action_model': 7.390158015783643e-05, 'epoch': 0.13}) +[2026-10-03 13:22:52,911][starVLA.training.train_starvla][INFO] - Step 1822, Loss: {'train/loss': 3.884901525452733e-05, 'train/grad_norm_pre_clip': 0.009582890197634697, 'timing/data': 0.0003376300010131672, 'timing/model': 0.38876768499903847, 'learning_rate/qwen_vl_interface': 7.387438787187576e-06, 'learning_rate/action_model': 7.387438787187576e-05, 'epoch': 0.13}) +[2026-10-03 13:22:53,795][starVLA.training.train_starvla][INFO] - Step 1823, Loss: {'train/loss': 4.8169760702876374e-05, 'train/grad_norm_pre_clip': 0.039852432906627655, 'timing/data': 0.00025522000214550644, 'timing/model': 0.38908345499658026, 'learning_rate/qwen_vl_interface': 7.3847186799711625e-06, 'learning_rate/action_model': 7.384718679971162e-05, 'epoch': 0.13}) +[2026-10-03 13:22:54,489][starVLA.training.train_starvla][INFO] - Step 1824, Loss: {'train/loss': 0.11713207513093948, 'train/grad_norm_pre_clip': 202.03225708007812, 'timing/data': 0.00024606000806670636, 'timing/model': 0.38812262599822134, 'learning_rate/qwen_vl_interface': 7.381997695252535e-06, 'learning_rate/action_model': 7.381997695252535e-05, 'epoch': 0.13}) +[2026-10-03 13:22:55,184][starVLA.training.train_starvla][INFO] - Step 1825, Loss: {'train/loss': 0.06912758946418762, 'train/grad_norm_pre_clip': 86.3486328125, 'timing/data': 0.0002492900093784556, 'timing/model': 0.3880403560033301, 'learning_rate/qwen_vl_interface': 7.379275834150186e-06, 'learning_rate/action_model': 7.379275834150186e-05, 'epoch': 0.13}) +[2026-10-03 13:22:55,880][starVLA.training.train_starvla][INFO] - Step 1826, Loss: {'train/loss': 0.3834999203681946, 'train/grad_norm_pre_clip': 141.84486389160156, 'timing/data': 0.00019839999731630087, 'timing/model': 0.38844484600122087, 'learning_rate/qwen_vl_interface': 7.376553097782972e-06, 'learning_rate/action_model': 7.376553097782972e-05, 'epoch': 0.13}) +[2026-10-03 13:22:56,572][starVLA.training.train_starvla][INFO] - Step 1827, Loss: {'train/loss': 0.48553624749183655, 'train/grad_norm_pre_clip': 16.6792049407959, 'timing/data': 0.00023841999063733965, 'timing/model': 0.388585755994427, 'learning_rate/qwen_vl_interface': 7.373829487270109e-06, 'learning_rate/action_model': 7.373829487270109e-05, 'epoch': 0.13}) +[2026-10-03 13:22:57,457][starVLA.training.train_starvla][INFO] - Step 1828, Loss: {'train/loss': 5.117101318319328e-05, 'train/grad_norm_pre_clip': 0.05047931149601936, 'timing/data': 0.0002725300000747666, 'timing/model': 0.578975743992487, 'learning_rate/qwen_vl_interface': 7.371105003731165e-06, 'learning_rate/action_model': 7.371105003731164e-05, 'epoch': 0.13}) +[2026-10-03 13:22:58,151][starVLA.training.train_starvla][INFO] - Step 1829, Loss: {'train/loss': 0.00011675851419568062, 'train/grad_norm_pre_clip': 0.051881685853004456, 'timing/data': 0.0001969800068764016, 'timing/model': 0.3880484859982971, 'learning_rate/qwen_vl_interface': 7.3683796482860755e-06, 'learning_rate/action_model': 7.368379648286075e-05, 'epoch': 0.13}) +[2026-10-03 13:22:58,848][starVLA.training.train_starvla][INFO] - Step 1830, Loss: {'train/loss': 0.058437932282686234, 'train/grad_norm_pre_clip': 69.4460678100586, 'timing/data': 0.0003122300113318488, 'timing/model': 0.3887090159987565, 'learning_rate/qwen_vl_interface': 7.365653422055131e-06, 'learning_rate/action_model': 7.365653422055131e-05, 'epoch': 0.13}) +[2026-10-03 13:22:59,543][starVLA.training.train_starvla][INFO] - Step 1831, Loss: {'train/loss': 0.020941531285643578, 'train/grad_norm_pre_clip': 36.64176559448242, 'timing/data': 0.00019345000328030437, 'timing/model': 0.3884946059988579, 'learning_rate/qwen_vl_interface': 7.3629263261589816e-06, 'learning_rate/action_model': 7.362926326158981e-05, 'epoch': 0.13}) +[2026-10-03 13:23:00,238][starVLA.training.train_starvla][INFO] - Step 1832, Loss: {'train/loss': 0.736370861530304, 'train/grad_norm_pre_clip': 223.52601623535156, 'timing/data': 0.0003157500032102689, 'timing/model': 0.3884339850046672, 'learning_rate/qwen_vl_interface': 7.36019836171863e-06, 'learning_rate/action_model': 7.360198361718629e-05, 'epoch': 0.13}) +[2026-10-03 13:23:01,123][starVLA.training.train_starvla][INFO] - Step 1833, Loss: {'train/loss': 0.0017233282560482621, 'train/grad_norm_pre_clip': 1.7585177421569824, 'timing/data': 0.0003491399984341115, 'timing/model': 0.5793630440020934, 'learning_rate/qwen_vl_interface': 7.357469529855441e-06, 'learning_rate/action_model': 7.35746952985544e-05, 'epoch': 0.13}) +[2026-10-03 13:23:01,820][starVLA.training.train_starvla][INFO] - Step 1834, Loss: {'train/loss': 0.11307784914970398, 'train/grad_norm_pre_clip': 109.14543914794922, 'timing/data': 0.00022671000624541193, 'timing/model': 0.38864258499233983, 'learning_rate/qwen_vl_interface': 7.354739831691134e-06, 'learning_rate/action_model': 7.354739831691135e-05, 'epoch': 0.13}) +[2026-10-03 13:23:02,514][starVLA.training.train_starvla][INFO] - Step 1835, Loss: {'train/loss': 0.005045885220170021, 'train/grad_norm_pre_clip': 13.406224250793457, 'timing/data': 0.00019654000061564147, 'timing/model': 0.38846761600871105, 'learning_rate/qwen_vl_interface': 7.352009268347786e-06, 'learning_rate/action_model': 7.352009268347786e-05, 'epoch': 0.13}) +[2026-10-03 13:23:03,209][starVLA.training.train_starvla][INFO] - Step 1836, Loss: {'train/loss': 0.003605070523917675, 'train/grad_norm_pre_clip': 5.76358699798584, 'timing/data': 0.00019542999507393688, 'timing/model': 0.38838398600637447, 'learning_rate/qwen_vl_interface': 7.349277840947828e-06, 'learning_rate/action_model': 7.349277840947828e-05, 'epoch': 0.13}) +[2026-10-03 13:23:03,904][starVLA.training.train_starvla][INFO] - Step 1837, Loss: {'train/loss': 0.002485377714037895, 'train/grad_norm_pre_clip': 0.6386113166809082, 'timing/data': 0.0002484099968569353, 'timing/model': 0.3883921350061428, 'learning_rate/qwen_vl_interface': 7.346545550614046e-06, 'learning_rate/action_model': 7.346545550614046e-05, 'epoch': 0.13}) +[2026-10-03 13:23:04,602][starVLA.training.train_starvla][INFO] - Step 1838, Loss: {'train/loss': 0.6145715713500977, 'train/grad_norm_pre_clip': 74.60003662109375, 'timing/data': 0.0003301699907751754, 'timing/model': 0.3885752559872344, 'learning_rate/qwen_vl_interface': 7.343812398469582e-06, 'learning_rate/action_model': 7.343812398469581e-05, 'epoch': 0.13}) +[2026-10-03 13:23:05,488][starVLA.training.train_starvla][INFO] - Step 1839, Loss: {'train/loss': 0.032384198158979416, 'train/grad_norm_pre_clip': 11.618398666381836, 'timing/data': 0.0002523500006645918, 'timing/model': 0.3887533460074337, 'learning_rate/qwen_vl_interface': 7.34107838563793e-06, 'learning_rate/action_model': 7.341078385637929e-05, 'epoch': 0.13}) +[2026-10-03 13:23:06,183][starVLA.training.train_starvla][INFO] - Step 1840, Loss: {'train/loss': 0.6395041942596436, 'train/grad_norm_pre_clip': 129.39505004882812, 'timing/data': 0.000321789993904531, 'timing/model': 0.38862874600454234, 'learning_rate/qwen_vl_interface': 7.338343513242941e-06, 'learning_rate/action_model': 7.338343513242941e-05, 'epoch': 0.13}) +[2026-10-03 13:23:06,879][starVLA.training.train_starvla][INFO] - Step 1841, Loss: {'train/loss': 0.4963776171207428, 'train/grad_norm_pre_clip': 125.97660827636719, 'timing/data': 0.00018945999909192324, 'timing/model': 0.3889484549872577, 'learning_rate/qwen_vl_interface': 7.335607782408818e-06, 'learning_rate/action_model': 7.335607782408818e-05, 'epoch': 0.13}) +[2026-10-03 13:23:07,575][starVLA.training.train_starvla][INFO] - Step 1842, Loss: {'train/loss': 2.0562754798447713e-05, 'train/grad_norm_pre_clip': 0.018810128793120384, 'timing/data': 0.00019566000264603645, 'timing/model': 0.38879925599030685, 'learning_rate/qwen_vl_interface': 7.3328711942601145e-06, 'learning_rate/action_model': 7.332871194260115e-05, 'epoch': 0.13}) +[2026-10-03 13:23:08,270][starVLA.training.train_starvla][INFO] - Step 1843, Loss: {'train/loss': 0.0005347257829271257, 'train/grad_norm_pre_clip': 0.5191948413848877, 'timing/data': 0.000245789997279644, 'timing/model': 0.38869565600180067, 'learning_rate/qwen_vl_interface': 7.33013374992174e-06, 'learning_rate/action_model': 7.33013374992174e-05, 'epoch': 0.13}) +[2026-10-03 13:23:09,152][starVLA.training.train_starvla][INFO] - Step 1844, Loss: {'train/loss': 0.02229352295398712, 'train/grad_norm_pre_clip': 34.16439437866211, 'timing/data': 0.00025260000256821513, 'timing/model': 0.5773593350022566, 'learning_rate/qwen_vl_interface': 7.327395450518956e-06, 'learning_rate/action_model': 7.327395450518956e-05, 'epoch': 0.13}) +[2026-10-03 13:23:09,847][starVLA.training.train_starvla][INFO] - Step 1845, Loss: {'train/loss': 1.4386032489710487e-05, 'train/grad_norm_pre_clip': 0.0033370747696608305, 'timing/data': 0.0002545000024838373, 'timing/model': 0.38825219499995, 'learning_rate/qwen_vl_interface': 7.324656297177372e-06, 'learning_rate/action_model': 7.324656297177371e-05, 'epoch': 0.14}) +[2026-10-03 13:23:10,544][starVLA.training.train_starvla][INFO] - Step 1846, Loss: {'train/loss': 0.000432029424700886, 'train/grad_norm_pre_clip': 0.6815980672836304, 'timing/data': 0.0003202090010745451, 'timing/model': 0.3889821460034, 'learning_rate/qwen_vl_interface': 7.32191629102295e-06, 'learning_rate/action_model': 7.32191629102295e-05, 'epoch': 0.14}) +[2026-10-03 13:23:11,239][starVLA.training.train_starvla][INFO] - Step 1847, Loss: {'train/loss': 3.886564081767574e-05, 'train/grad_norm_pre_clip': 0.01579987071454525, 'timing/data': 0.00023997999960556626, 'timing/model': 0.3888765260053333, 'learning_rate/qwen_vl_interface': 7.319175433182004e-06, 'learning_rate/action_model': 7.319175433182004e-05, 'epoch': 0.14}) +[2026-10-03 13:23:11,933][starVLA.training.train_starvla][INFO] - Step 1848, Loss: {'train/loss': 0.12565173208713531, 'train/grad_norm_pre_clip': 41.00144577026367, 'timing/data': 0.00019176999921910465, 'timing/model': 0.38836475599964615, 'learning_rate/qwen_vl_interface': 7.316433724781198e-06, 'learning_rate/action_model': 7.316433724781197e-05, 'epoch': 0.14}) +[2026-10-03 13:23:12,628][starVLA.training.train_starvla][INFO] - Step 1849, Loss: {'train/loss': 7.51992265577428e-05, 'train/grad_norm_pre_clip': 0.01472723949700594, 'timing/data': 0.00024480999854858965, 'timing/model': 0.38813289599784184, 'learning_rate/qwen_vl_interface': 7.313691166947543e-06, 'learning_rate/action_model': 7.313691166947543e-05, 'epoch': 0.14}) +[2026-10-03 13:23:13,515][starVLA.training.train_starvla][INFO] - Step 1850, Loss: {'train/loss': 0.00019369280198588967, 'train/grad_norm_pre_clip': 0.10026265680789948, 'timing/data': 0.00020496999786701053, 'timing/model': 0.3887326359981671, 'learning_rate/qwen_vl_interface': 7.310947760808401e-06, 'learning_rate/action_model': 7.3109477608084e-05, 'epoch': 0.14}) +[2026-10-03 13:23:14,208][starVLA.training.train_starvla][INFO] - Step 1851, Loss: {'train/loss': 0.0010497068287804723, 'train/grad_norm_pre_clip': 0.2435528188943863, 'timing/data': 0.00024142999609466642, 'timing/model': 0.38783743599196896, 'learning_rate/qwen_vl_interface': 7.308203507491485e-06, 'learning_rate/action_model': 7.308203507491484e-05, 'epoch': 0.14}) +[2026-10-03 13:23:14,902][starVLA.training.train_starvla][INFO] - Step 1852, Loss: {'train/loss': 3.407648910069838e-05, 'train/grad_norm_pre_clip': 0.006798040587455034, 'timing/data': 0.0003256399941165, 'timing/model': 0.3881506449979497, 'learning_rate/qwen_vl_interface': 7.3054584081248515e-06, 'learning_rate/action_model': 7.305458408124852e-05, 'epoch': 0.14}) +[2026-10-03 13:23:15,596][starVLA.training.train_starvla][INFO] - Step 1853, Loss: {'train/loss': 3.390396886970848e-05, 'train/grad_norm_pre_clip': 0.00680083641782403, 'timing/data': 0.0002999900025315583, 'timing/model': 0.38805180600320455, 'learning_rate/qwen_vl_interface': 7.302712463836907e-06, 'learning_rate/action_model': 7.302712463836907e-05, 'epoch': 0.14}) +[2026-10-03 13:23:16,291][starVLA.training.train_starvla][INFO] - Step 1854, Loss: {'train/loss': 0.002260009292513132, 'train/grad_norm_pre_clip': 0.8136414289474487, 'timing/data': 0.0003178899933118373, 'timing/model': 0.38813041600224096, 'learning_rate/qwen_vl_interface': 7.2999656757564086e-06, 'learning_rate/action_model': 7.299965675756409e-05, 'epoch': 0.14}) +[2026-10-03 13:23:17,177][starVLA.training.train_starvla][INFO] - Step 1855, Loss: {'train/loss': 0.28414782881736755, 'train/grad_norm_pre_clip': 31.35318946838379, 'timing/data': 0.0003109599929302931, 'timing/model': 0.38778773600643035, 'learning_rate/qwen_vl_interface': 7.297218045012453e-06, 'learning_rate/action_model': 7.297218045012453e-05, 'epoch': 0.14}) +[2026-10-03 13:23:17,871][starVLA.training.train_starvla][INFO] - Step 1856, Loss: {'train/loss': 0.0002507723111193627, 'train/grad_norm_pre_clip': 0.2501507103443146, 'timing/data': 0.0002952700015157461, 'timing/model': 0.38838725599634927, 'learning_rate/qwen_vl_interface': 7.29446957273449e-06, 'learning_rate/action_model': 7.294469572734489e-05, 'epoch': 0.14}) +[2026-10-03 13:23:18,565][starVLA.training.train_starvla][INFO] - Step 1857, Loss: {'train/loss': 6.172891153255478e-05, 'train/grad_norm_pre_clip': 0.016291167587041855, 'timing/data': 0.0003208800044376403, 'timing/model': 0.3879164960089838, 'learning_rate/qwen_vl_interface': 7.2917202600523104e-06, 'learning_rate/action_model': 7.29172026005231e-05, 'epoch': 0.14}) +[2026-10-03 13:23:19,260][starVLA.training.train_starvla][INFO] - Step 1858, Loss: {'train/loss': 0.23962226510047913, 'train/grad_norm_pre_clip': 40.325199127197266, 'timing/data': 0.00032115000067278743, 'timing/model': 0.38810961600393057, 'learning_rate/qwen_vl_interface': 7.2889701080960565e-06, 'learning_rate/action_model': 7.288970108096056e-05, 'epoch': 0.14}) +[2026-10-03 13:23:19,954][starVLA.training.train_starvla][INFO] - Step 1859, Loss: {'train/loss': 0.49694982171058655, 'train/grad_norm_pre_clip': 129.01405334472656, 'timing/data': 0.00019348999194335192, 'timing/model': 0.3885540760093136, 'learning_rate/qwen_vl_interface': 7.286219117996209e-06, 'learning_rate/action_model': 7.286219117996207e-05, 'epoch': 0.14}) +[2026-10-03 13:23:20,838][starVLA.training.train_starvla][INFO] - Step 1860, Loss: {'train/loss': 0.013780138455331326, 'train/grad_norm_pre_clip': 12.232650756835938, 'timing/data': 0.000198819994693622, 'timing/model': 0.5772459049912868, 'learning_rate/qwen_vl_interface': 7.283467290883598e-06, 'learning_rate/action_model': 7.283467290883598e-05, 'epoch': 0.14}) +[2026-10-03 13:23:21,532][starVLA.training.train_starvla][INFO] - Step 1861, Loss: {'train/loss': 0.0011041064281016588, 'train/grad_norm_pre_clip': 1.2393710613250732, 'timing/data': 0.0003326699952594936, 'timing/model': 0.38759184601076413, 'learning_rate/qwen_vl_interface': 7.2807146278893925e-06, 'learning_rate/action_model': 7.280714627889393e-05, 'epoch': 0.14}) +[2026-10-03 13:23:22,205][starVLA.training.train_starvla][INFO] - Step 1862, Loss: {'train/loss': 1.5421768694068305e-05, 'train/grad_norm_pre_clip': 0.004145585000514984, 'timing/data': 0.00019924000662285835, 'timing/model': 0.3653354690031847, 'learning_rate/qwen_vl_interface': 7.277961130145111e-06, 'learning_rate/action_model': 7.277961130145111e-05, 'epoch': 0.14}) +[2026-10-03 13:23:22,899][starVLA.training.train_starvla][INFO] - Step 1863, Loss: {'train/loss': 1.2053923606872559, 'train/grad_norm_pre_clip': 156.5216522216797, 'timing/data': 0.00019743999291677028, 'timing/model': 0.3881157559953863, 'learning_rate/qwen_vl_interface': 7.2752067987826116e-06, 'learning_rate/action_model': 7.275206798782611e-05, 'epoch': 0.14}) +[2026-10-03 13:23:23,592][starVLA.training.train_starvla][INFO] - Step 1864, Loss: {'train/loss': 0.00017791750724427402, 'train/grad_norm_pre_clip': 0.06609421223402023, 'timing/data': 0.0001944800023920834, 'timing/model': 0.3885211659944616, 'learning_rate/qwen_vl_interface': 7.272451634934096e-06, 'learning_rate/action_model': 7.272451634934096e-05, 'epoch': 0.14}) +[2026-10-03 13:23:24,287][starVLA.training.train_starvla][INFO] - Step 1865, Loss: {'train/loss': 0.4043402671813965, 'train/grad_norm_pre_clip': 122.16799926757812, 'timing/data': 0.00020033000328112394, 'timing/model': 0.3879252059996361, 'learning_rate/qwen_vl_interface': 7.269695639732108e-06, 'learning_rate/action_model': 7.269695639732108e-05, 'epoch': 0.14}) +[2026-10-03 13:23:25,172][starVLA.training.train_starvla][INFO] - Step 1866, Loss: {'train/loss': 0.0017249148804694414, 'train/grad_norm_pre_clip': 1.2626343965530396, 'timing/data': 0.00020193999807816, 'timing/model': 0.38871452599414624, 'learning_rate/qwen_vl_interface': 7.266938814309533e-06, 'learning_rate/action_model': 7.266938814309533e-05, 'epoch': 0.14}) +[2026-10-03 13:23:25,865][starVLA.training.train_starvla][INFO] - Step 1867, Loss: {'train/loss': 0.0015167040983214974, 'train/grad_norm_pre_clip': 1.1073442697525024, 'timing/data': 0.0001895199966384098, 'timing/model': 0.3880689159996109, 'learning_rate/qwen_vl_interface': 7.264181159799598e-06, 'learning_rate/action_model': 7.264181159799597e-05, 'epoch': 0.14}) +[2026-10-03 13:23:26,557][starVLA.training.train_starvla][INFO] - Step 1868, Loss: {'train/loss': 0.00039254894363693893, 'train/grad_norm_pre_clip': 0.5028665065765381, 'timing/data': 0.00021083999308757484, 'timing/model': 0.3882751359924441, 'learning_rate/qwen_vl_interface': 7.261422677335871e-06, 'learning_rate/action_model': 7.261422677335871e-05, 'epoch': 0.14}) +[2026-10-03 13:23:27,251][starVLA.training.train_starvla][INFO] - Step 1869, Loss: {'train/loss': 0.0026893464382737875, 'train/grad_norm_pre_clip': 2.3846020698547363, 'timing/data': 0.00019813000108115375, 'timing/model': 0.38791466600378044, 'learning_rate/qwen_vl_interface': 7.258663368052259e-06, 'learning_rate/action_model': 7.258663368052259e-05, 'epoch': 0.14}) +[2026-10-03 13:23:27,947][starVLA.training.train_starvla][INFO] - Step 1870, Loss: {'train/loss': 0.0028173986356705427, 'train/grad_norm_pre_clip': 1.797938585281372, 'timing/data': 0.00020370999118313193, 'timing/model': 0.38899327599210665, 'learning_rate/qwen_vl_interface': 7.2559032330830105e-06, 'learning_rate/action_model': 7.25590323308301e-05, 'epoch': 0.14}) +[2026-10-03 13:23:28,845][starVLA.training.train_starvla][INFO] - Step 1871, Loss: {'train/loss': 1.8044498574454337e-05, 'train/grad_norm_pre_clip': 0.004718394484370947, 'timing/data': 0.00026635000540409237, 'timing/model': 0.5915949710033601, 'learning_rate/qwen_vl_interface': 7.253142273562712e-06, 'learning_rate/action_model': 7.253142273562712e-05, 'epoch': 0.14}) +[2026-10-03 13:23:29,551][starVLA.training.train_starvla][INFO] - Step 1872, Loss: {'train/loss': 0.001512225833721459, 'train/grad_norm_pre_clip': 2.420822858810425, 'timing/data': 0.00020403999951668084, 'timing/model': 0.39424441500159446, 'learning_rate/qwen_vl_interface': 7.25038049062629e-06, 'learning_rate/action_model': 7.250380490626289e-05, 'epoch': 0.14}) +[2026-10-03 13:23:30,257][starVLA.training.train_starvla][INFO] - Step 1873, Loss: {'train/loss': 0.0007109709549695253, 'train/grad_norm_pre_clip': 2.971778154373169, 'timing/data': 0.0002794699976220727, 'timing/model': 0.3936298149928916, 'learning_rate/qwen_vl_interface': 7.24761788540901e-06, 'learning_rate/action_model': 7.24761788540901e-05, 'epoch': 0.14}) +[2026-10-03 13:23:30,964][starVLA.training.train_starvla][INFO] - Step 1874, Loss: {'train/loss': 5.450027310871519e-05, 'train/grad_norm_pre_clip': 0.05553978681564331, 'timing/data': 0.00020130000484641641, 'timing/model': 0.39648167400446255, 'learning_rate/qwen_vl_interface': 7.244854459046473e-06, 'learning_rate/action_model': 7.244854459046473e-05, 'epoch': 0.14}) +[2026-10-03 13:23:31,683][starVLA.training.train_starvla][INFO] - Step 1875, Loss: {'train/loss': 0.052758701145648956, 'train/grad_norm_pre_clip': 25.135040283203125, 'timing/data': 0.0002716799936024472, 'timing/model': 0.3947026560053928, 'learning_rate/qwen_vl_interface': 7.24209021267462e-06, 'learning_rate/action_model': 7.24209021267462e-05, 'epoch': 0.14}) +[2026-10-03 13:23:32,579][starVLA.training.train_starvla][INFO] - Step 1876, Loss: {'train/loss': 1.1116038876934908e-05, 'train/grad_norm_pre_clip': 0.0020681049209088087, 'timing/data': 0.0002997599949594587, 'timing/model': 0.5854312039882643, 'learning_rate/qwen_vl_interface': 7.23932514742973e-06, 'learning_rate/action_model': 7.23932514742973e-05, 'epoch': 0.14}) +[2026-10-03 13:23:33,282][starVLA.training.train_starvla][INFO] - Step 1877, Loss: {'train/loss': 0.00022696686210110784, 'train/grad_norm_pre_clip': 0.034072961658239365, 'timing/data': 0.0003228399873478338, 'timing/model': 0.3926016350014834, 'learning_rate/qwen_vl_interface': 7.236559264448413e-06, 'learning_rate/action_model': 7.236559264448412e-05, 'epoch': 0.14}) +[2026-10-03 13:23:33,988][starVLA.training.train_starvla][INFO] - Step 1878, Loss: {'train/loss': 1.1011515198333655e-05, 'train/grad_norm_pre_clip': 0.005852686706930399, 'timing/data': 0.0002511400089133531, 'timing/model': 0.3945269349933369, 'learning_rate/qwen_vl_interface': 7.233792564867623e-06, 'learning_rate/action_model': 7.233792564867623e-05, 'epoch': 0.14}) +[2026-10-03 13:23:34,692][starVLA.training.train_starvla][INFO] - Step 1879, Loss: {'train/loss': 0.025007139891386032, 'train/grad_norm_pre_clip': 43.6835823059082, 'timing/data': 0.00025036999431904405, 'timing/model': 0.39456109500315506, 'learning_rate/qwen_vl_interface': 7.231025049824644e-06, 'learning_rate/action_model': 7.231025049824643e-05, 'epoch': 0.14}) +[2026-10-03 13:23:35,396][starVLA.training.train_starvla][INFO] - Step 1880, Loss: {'train/loss': 1.1204646825790405, 'train/grad_norm_pre_clip': 99.3111572265625, 'timing/data': 0.000372009992133826, 'timing/model': 0.3929004859965062, 'learning_rate/qwen_vl_interface': 7.228256720457097e-06, 'learning_rate/action_model': 7.228256720457096e-05, 'epoch': 0.14}) +[2026-10-03 13:23:36,100][starVLA.training.train_starvla][INFO] - Step 1881, Loss: {'train/loss': 9.618184776627459e-06, 'train/grad_norm_pre_clip': 0.007699683308601379, 'timing/data': 0.00032588999602012336, 'timing/model': 0.3935213739896426, 'learning_rate/qwen_vl_interface': 7.225487577902937e-06, 'learning_rate/action_model': 7.225487577902936e-05, 'epoch': 0.14}) +[2026-10-03 13:23:37,003][starVLA.training.train_starvla][INFO] - Step 1882, Loss: {'train/loss': 0.2693409323692322, 'train/grad_norm_pre_clip': 28.65151023864746, 'timing/data': 0.0002177800051867962, 'timing/model': 0.39538190499297343, 'learning_rate/qwen_vl_interface': 7.222717623300455e-06, 'learning_rate/action_model': 7.222717623300455e-05, 'epoch': 0.14}) +[2026-10-03 13:23:37,724][starVLA.training.train_starvla][INFO] - Step 1883, Loss: {'train/loss': 0.09350117295980453, 'train/grad_norm_pre_clip': 146.64706420898438, 'timing/data': 0.0002712300047278404, 'timing/model': 0.39529708500776906, 'learning_rate/qwen_vl_interface': 7.219946857788273e-06, 'learning_rate/action_model': 7.219946857788272e-05, 'epoch': 0.14}) +[2026-10-03 13:23:38,428][starVLA.training.train_starvla][INFO] - Step 1884, Loss: {'train/loss': 0.3467056155204773, 'train/grad_norm_pre_clip': 107.64591217041016, 'timing/data': 0.00021020999702159315, 'timing/model': 0.39318074499897193, 'learning_rate/qwen_vl_interface': 7.21717528250535e-06, 'learning_rate/action_model': 7.21717528250535e-05, 'epoch': 0.14}) +[2026-10-03 13:23:39,132][starVLA.training.train_starvla][INFO] - Step 1885, Loss: {'train/loss': 0.27723872661590576, 'train/grad_norm_pre_clip': 124.13240051269531, 'timing/data': 0.0003974699939135462, 'timing/model': 0.39301232500292826, 'learning_rate/qwen_vl_interface': 7.214402898590974e-06, 'learning_rate/action_model': 7.214402898590973e-05, 'epoch': 0.14}) +[2026-10-03 13:23:39,840][starVLA.training.train_starvla][INFO] - Step 1886, Loss: {'train/loss': 0.3371681272983551, 'train/grad_norm_pre_clip': 97.95899200439453, 'timing/data': 0.0002101199934259057, 'timing/model': 0.3955578649911331, 'learning_rate/qwen_vl_interface': 7.211629707184768e-06, 'learning_rate/action_model': 7.211629707184768e-05, 'epoch': 0.14}) +[2026-10-03 13:23:40,772][starVLA.training.train_starvla][INFO] - Step 1887, Loss: {'train/loss': 0.00018054284737445414, 'train/grad_norm_pre_clip': 0.031011367216706276, 'timing/data': 0.00021899999410379678, 'timing/model': 0.3937853550014552, 'learning_rate/qwen_vl_interface': 7.2088557094266885e-06, 'learning_rate/action_model': 7.208855709426688e-05, 'epoch': 0.14}) +[2026-10-03 13:23:41,491][starVLA.training.train_starvla][INFO] - Step 1888, Loss: {'train/loss': 0.6336430311203003, 'train/grad_norm_pre_clip': 31.894479751586914, 'timing/data': 0.00028649999876506627, 'timing/model': 0.39379356399876997, 'learning_rate/qwen_vl_interface': 7.2060809064570195e-06, 'learning_rate/action_model': 7.20608090645702e-05, 'epoch': 0.14}) +[2026-10-03 13:23:42,195][starVLA.training.train_starvla][INFO] - Step 1889, Loss: {'train/loss': 0.0001567222352605313, 'train/grad_norm_pre_clip': 0.0939583033323288, 'timing/data': 0.00034306899760849774, 'timing/model': 0.39321583599667065, 'learning_rate/qwen_vl_interface': 7.203305299416375e-06, 'learning_rate/action_model': 7.203305299416375e-05, 'epoch': 0.14}) +[2026-10-03 13:23:42,899][starVLA.training.train_starvla][INFO] - Step 1890, Loss: {'train/loss': 0.022563688457012177, 'train/grad_norm_pre_clip': 8.892401695251465, 'timing/data': 0.00033755999174900353, 'timing/model': 0.3933090850041481, 'learning_rate/qwen_vl_interface': 7.200528889445707e-06, 'learning_rate/action_model': 7.200528889445706e-05, 'epoch': 0.14}) +[2026-10-03 13:23:43,604][starVLA.training.train_starvla][INFO] - Step 1891, Loss: {'train/loss': 0.04634808003902435, 'train/grad_norm_pre_clip': 7.479516983032227, 'timing/data': 0.0002383700048085302, 'timing/model': 0.39336337499844376, 'learning_rate/qwen_vl_interface': 7.19775167768629e-06, 'learning_rate/action_model': 7.19775167768629e-05, 'epoch': 0.14}) +[2026-10-03 13:23:44,514][starVLA.training.train_starvla][INFO] - Step 1892, Loss: {'train/loss': 0.05845940485596657, 'train/grad_norm_pre_clip': 9.439340591430664, 'timing/data': 0.000383890001103282, 'timing/model': 0.5860260420013219, 'learning_rate/qwen_vl_interface': 7.194973665279731e-06, 'learning_rate/action_model': 7.19497366527973e-05, 'epoch': 0.14}) +[2026-10-03 13:23:45,219][starVLA.training.train_starvla][INFO] - Step 1893, Loss: {'train/loss': 0.029883023351430893, 'train/grad_norm_pre_clip': 50.200531005859375, 'timing/data': 0.00034374000097159296, 'timing/model': 0.39308328500192147, 'learning_rate/qwen_vl_interface': 7.1921948533679666e-06, 'learning_rate/action_model': 7.192194853367967e-05, 'epoch': 0.14}) +[2026-10-03 13:23:45,924][starVLA.training.train_starvla][INFO] - Step 1894, Loss: {'train/loss': 0.023132165893912315, 'train/grad_norm_pre_clip': 4.217427730560303, 'timing/data': 0.0002736799942795187, 'timing/model': 0.39374132500961423, 'learning_rate/qwen_vl_interface': 7.1894152430932605e-06, 'learning_rate/action_model': 7.189415243093261e-05, 'epoch': 0.14}) +[2026-10-03 13:23:46,629][starVLA.training.train_starvla][INFO] - Step 1895, Loss: {'train/loss': 0.20455071330070496, 'train/grad_norm_pre_clip': 138.4001007080078, 'timing/data': 0.00028608000138774514, 'timing/model': 0.3935561460093595, 'learning_rate/qwen_vl_interface': 7.186634835598205e-06, 'learning_rate/action_model': 7.186634835598205e-05, 'epoch': 0.14}) +[2026-10-03 13:23:47,348][starVLA.training.train_starvla][INFO] - Step 1896, Loss: {'train/loss': 0.00025334610836580396, 'train/grad_norm_pre_clip': 0.16631010174751282, 'timing/data': 0.00033466001332271844, 'timing/model': 0.3941187849995913, 'learning_rate/qwen_vl_interface': 7.183853632025724e-06, 'learning_rate/action_model': 7.183853632025723e-05, 'epoch': 0.14}) +[2026-10-03 13:23:48,242][starVLA.training.train_starvla][INFO] - Step 1897, Loss: {'train/loss': 0.7948055863380432, 'train/grad_norm_pre_clip': 134.9123992919922, 'timing/data': 0.00033987000642810017, 'timing/model': 0.5830474730028072, 'learning_rate/qwen_vl_interface': 7.181071633519062e-06, 'learning_rate/action_model': 7.181071633519062e-05, 'epoch': 0.14}) +[2026-10-03 13:23:48,948][starVLA.training.train_starvla][INFO] - Step 1898, Loss: {'train/loss': 2.894428325816989e-05, 'train/grad_norm_pre_clip': 0.009211357682943344, 'timing/data': 0.00022359000286087394, 'timing/model': 0.39440544600074645, 'learning_rate/qwen_vl_interface': 7.178288841221795e-06, 'learning_rate/action_model': 7.178288841221794e-05, 'epoch': 0.14}) +[2026-10-03 13:23:49,652][starVLA.training.train_starvla][INFO] - Step 1899, Loss: {'train/loss': 0.004017319064587355, 'train/grad_norm_pre_clip': 1.2316964864730835, 'timing/data': 0.0003473899996606633, 'timing/model': 0.39327441599743906, 'learning_rate/qwen_vl_interface': 7.1755052562778215e-06, 'learning_rate/action_model': 7.175505256277822e-05, 'epoch': 0.14}) +[2026-10-03 13:23:50,371][starVLA.training.train_starvla][INFO] - Step 1900, Loss: {'train/loss': 0.0007842822815291584, 'train/grad_norm_pre_clip': 0.4407441020011902, 'timing/data': 0.00035645999014377594, 'timing/model': 0.3934226849960396, 'learning_rate/qwen_vl_interface': 7.1727208798313704e-06, 'learning_rate/action_model': 7.172720879831371e-05, 'epoch': 0.14}) +[2026-10-03 13:23:51,075][starVLA.training.train_starvla][INFO] - Step 1901, Loss: {'train/loss': 0.005592693109065294, 'train/grad_norm_pre_clip': 22.231374740600586, 'timing/data': 0.00026455000624991953, 'timing/model': 0.39344435499515384, 'learning_rate/qwen_vl_interface': 7.169935713026994e-06, 'learning_rate/action_model': 7.169935713026994e-05, 'epoch': 0.14}) +[2026-10-03 13:23:51,780][starVLA.training.train_starvla][INFO] - Step 1902, Loss: {'train/loss': 3.706311326823197e-05, 'train/grad_norm_pre_clip': 0.012116076424717903, 'timing/data': 0.00026197999250143766, 'timing/model': 0.39359844599675853, 'learning_rate/qwen_vl_interface': 7.1671497570095684e-06, 'learning_rate/action_model': 7.167149757009568e-05, 'epoch': 0.14}) +[2026-10-03 13:23:52,675][starVLA.training.train_starvla][INFO] - Step 1903, Loss: {'train/loss': 0.011929179541766644, 'train/grad_norm_pre_clip': 38.985504150390625, 'timing/data': 0.00035573000786826015, 'timing/model': 0.393190874994616, 'learning_rate/qwen_vl_interface': 7.164363012924295e-06, 'learning_rate/action_model': 7.164363012924294e-05, 'epoch': 0.14}) +[2026-10-03 13:23:53,392][starVLA.training.train_starvla][INFO] - Step 1904, Loss: {'train/loss': 0.002972749527543783, 'train/grad_norm_pre_clip': 6.141314506530762, 'timing/data': 0.00026282999897375703, 'timing/model': 0.3932417850010097, 'learning_rate/qwen_vl_interface': 7.161575481916699e-06, 'learning_rate/action_model': 7.1615754819167e-05, 'epoch': 0.14}) +[2026-10-03 13:23:54,097][starVLA.training.train_starvla][INFO] - Step 1905, Loss: {'train/loss': 7.432397251250222e-05, 'train/grad_norm_pre_clip': 0.04413454979658127, 'timing/data': 0.00021337000362109393, 'timing/model': 0.39349323500937317, 'learning_rate/qwen_vl_interface': 7.158787165132627e-06, 'learning_rate/action_model': 7.158787165132626e-05, 'epoch': 0.14}) +[2026-10-03 13:23:54,802][starVLA.training.train_starvla][INFO] - Step 1906, Loss: {'train/loss': 1.7313557691522874e-05, 'train/grad_norm_pre_clip': 0.007987325079739094, 'timing/data': 0.00034690000757109374, 'timing/model': 0.3931609059945913, 'learning_rate/qwen_vl_interface': 7.1559980637182535e-06, 'learning_rate/action_model': 7.155998063718254e-05, 'epoch': 0.14}) +[2026-10-03 13:23:55,504][starVLA.training.train_starvla][INFO] - Step 1907, Loss: {'train/loss': 0.0001088757926481776, 'train/grad_norm_pre_clip': 0.058303095400333405, 'timing/data': 0.0002940800040960312, 'timing/model': 0.39272598500247113, 'learning_rate/qwen_vl_interface': 7.153208178820071e-06, 'learning_rate/action_model': 7.153208178820072e-05, 'epoch': 0.14}) +[2026-10-03 13:23:56,400][starVLA.training.train_starvla][INFO] - Step 1908, Loss: {'train/loss': 1.3597000361187384e-05, 'train/grad_norm_pre_clip': 0.003172324039041996, 'timing/data': 0.000289060000795871, 'timing/model': 0.39296356499835383, 'learning_rate/qwen_vl_interface': 7.150417511584898e-06, 'learning_rate/action_model': 7.150417511584898e-05, 'epoch': 0.14}) +[2026-10-03 13:23:57,104][starVLA.training.train_starvla][INFO] - Step 1909, Loss: {'train/loss': 0.34261107444763184, 'train/grad_norm_pre_clip': 51.3163948059082, 'timing/data': 0.00025976001052185893, 'timing/model': 0.393312034997507, 'learning_rate/qwen_vl_interface': 7.147626063159872e-06, 'learning_rate/action_model': 7.147626063159872e-05, 'epoch': 0.14}) +[2026-10-03 13:23:57,808][starVLA.training.train_starvla][INFO] - Step 1910, Loss: {'train/loss': 4.410708697832888e-06, 'train/grad_norm_pre_clip': 0.0017538962420076132, 'timing/data': 0.0002098199911415577, 'timing/model': 0.39360570500139147, 'learning_rate/qwen_vl_interface': 7.14483383469245e-06, 'learning_rate/action_model': 7.144833834692449e-05, 'epoch': 0.14}) +[2026-10-03 13:23:58,512][starVLA.training.train_starvla][INFO] - Step 1911, Loss: {'train/loss': 2.65723756456282e-05, 'train/grad_norm_pre_clip': 0.1007029190659523, 'timing/data': 0.00025705000734888017, 'timing/model': 0.3934338749968447, 'learning_rate/qwen_vl_interface': 7.142040827330411e-06, 'learning_rate/action_model': 7.142040827330411e-05, 'epoch': 0.14}) +[2026-10-03 13:23:59,217][starVLA.training.train_starvla][INFO] - Step 1912, Loss: {'train/loss': 7.241855655593099e-06, 'train/grad_norm_pre_clip': 0.0016954385209828615, 'timing/data': 0.0002849899901775643, 'timing/model': 0.39352524500282016, 'learning_rate/qwen_vl_interface': 7.139247042221859e-06, 'learning_rate/action_model': 7.139247042221858e-05, 'epoch': 0.14}) +[2026-10-03 13:24:00,112][starVLA.training.train_starvla][INFO] - Step 1913, Loss: {'train/loss': 6.137972377473488e-05, 'train/grad_norm_pre_clip': 0.03546816110610962, 'timing/data': 0.00029830000130459666, 'timing/model': 0.5838326329976553, 'learning_rate/qwen_vl_interface': 7.136452480515208e-06, 'learning_rate/action_model': 7.136452480515208e-05, 'epoch': 0.14}) +[2026-10-03 13:24:00,817][starVLA.training.train_starvla][INFO] - Step 1914, Loss: {'train/loss': 4.298939074942609e-06, 'train/grad_norm_pre_clip': 0.0008798739872872829, 'timing/data': 0.0002056099910987541, 'timing/model': 0.39330037600302603, 'learning_rate/qwen_vl_interface': 7.133657143359203e-06, 'learning_rate/action_model': 7.133657143359202e-05, 'epoch': 0.14}) +[2026-10-03 13:24:01,521][starVLA.training.train_starvla][INFO] - Step 1915, Loss: {'train/loss': 0.0947924554347992, 'train/grad_norm_pre_clip': 53.40882873535156, 'timing/data': 0.00031752999348100275, 'timing/model': 0.39326754500507377, 'learning_rate/qwen_vl_interface': 7.130861031902897e-06, 'learning_rate/action_model': 7.130861031902897e-05, 'epoch': 0.14}) +[2026-10-03 13:24:02,226][starVLA.training.train_starvla][INFO] - Step 1916, Loss: {'train/loss': 7.5249740802973974e-06, 'train/grad_norm_pre_clip': 0.0017136692767962813, 'timing/data': 0.0002997499977936968, 'timing/model': 0.3936592349928105, 'learning_rate/qwen_vl_interface': 7.128064147295667e-06, 'learning_rate/action_model': 7.128064147295667e-05, 'epoch': 0.14}) +[2026-10-03 13:24:02,930][starVLA.training.train_starvla][INFO] - Step 1917, Loss: {'train/loss': 0.057517800480127335, 'train/grad_norm_pre_clip': 25.594257354736328, 'timing/data': 0.0003168299881508574, 'timing/model': 0.39247892500134185, 'learning_rate/qwen_vl_interface': 7.125266490687206e-06, 'learning_rate/action_model': 7.125266490687206e-05, 'epoch': 0.14}) +[2026-10-03 13:24:03,824][starVLA.training.train_starvla][INFO] - Step 1918, Loss: {'train/loss': 3.6582068787538446e-06, 'train/grad_norm_pre_clip': 0.0009225856047123671, 'timing/data': 0.00034158000198658556, 'timing/model': 0.5840751740033738, 'learning_rate/qwen_vl_interface': 7.1224680632275265e-06, 'learning_rate/action_model': 7.122468063227527e-05, 'epoch': 0.14}) +[2026-10-03 13:24:04,527][starVLA.training.train_starvla][INFO] - Step 1919, Loss: {'train/loss': 1.7732290871208534e-06, 'train/grad_norm_pre_clip': 0.0002889624156523496, 'timing/data': 0.00020012000459246337, 'timing/model': 0.3923939149972284, 'learning_rate/qwen_vl_interface': 7.1196688660669555e-06, 'learning_rate/action_model': 7.119668866066955e-05, 'epoch': 0.14}) +[2026-10-03 13:24:05,232][starVLA.training.train_starvla][INFO] - Step 1920, Loss: {'train/loss': 0.00010041167843155563, 'train/grad_norm_pre_clip': 0.027296166867017746, 'timing/data': 0.00033491999784018844, 'timing/model': 0.3933848149899859, 'learning_rate/qwen_vl_interface': 7.116868900356138e-06, 'learning_rate/action_model': 7.116868900356138e-05, 'epoch': 0.14}) +[2026-10-03 13:24:05,934][starVLA.training.train_starvla][INFO] - Step 1921, Loss: {'train/loss': 1.8431050193612464e-05, 'train/grad_norm_pre_clip': 0.014972427859902382, 'timing/data': 0.00020507999579422176, 'timing/model': 0.39264864500728436, 'learning_rate/qwen_vl_interface': 7.114068167246033e-06, 'learning_rate/action_model': 7.114068167246033e-05, 'epoch': 0.14}) +[2026-10-03 13:24:06,638][starVLA.training.train_starvla][INFO] - Step 1922, Loss: {'train/loss': 0.004605842288583517, 'train/grad_norm_pre_clip': 3.3679192066192627, 'timing/data': 0.00036980000732000917, 'timing/model': 0.39289515500422567, 'learning_rate/qwen_vl_interface': 7.111266667887916e-06, 'learning_rate/action_model': 7.111266667887915e-05, 'epoch': 0.14}) +[2026-10-03 13:24:07,343][starVLA.training.train_starvla][INFO] - Step 1923, Loss: {'train/loss': 0.4667504131793976, 'train/grad_norm_pre_clip': 71.24752807617188, 'timing/data': 0.00026298000011593103, 'timing/model': 0.3938071150041651, 'learning_rate/qwen_vl_interface': 7.108464403433378e-06, 'learning_rate/action_model': 7.108464403433378e-05, 'epoch': 0.14}) +[2026-10-03 13:24:08,244][starVLA.training.train_starvla][INFO] - Step 1924, Loss: {'train/loss': 0.19394411146640778, 'train/grad_norm_pre_clip': 49.691490173339844, 'timing/data': 0.00039074999222066253, 'timing/model': 0.39410684500762727, 'learning_rate/qwen_vl_interface': 7.105661375034324e-06, 'learning_rate/action_model': 7.105661375034324e-05, 'epoch': 0.14}) +[2026-10-03 13:24:08,949][starVLA.training.train_starvla][INFO] - Step 1925, Loss: {'train/loss': 0.0001034290689858608, 'train/grad_norm_pre_clip': 0.020879508927464485, 'timing/data': 0.0003009000065503642, 'timing/model': 0.393975604994921, 'learning_rate/qwen_vl_interface': 7.102857583842973e-06, 'learning_rate/action_model': 7.102857583842974e-05, 'epoch': 0.14}) +[2026-10-03 13:24:09,655][starVLA.training.train_starvla][INFO] - Step 1926, Loss: {'train/loss': 6.843721348559484e-05, 'train/grad_norm_pre_clip': 0.02661876380443573, 'timing/data': 0.0003517800068948418, 'timing/model': 0.3937788950133836, 'learning_rate/qwen_vl_interface': 7.100053031011859e-06, 'learning_rate/action_model': 7.100053031011859e-05, 'epoch': 0.14}) +[2026-10-03 13:24:10,360][starVLA.training.train_starvla][INFO] - Step 1927, Loss: {'train/loss': 0.4988897442817688, 'train/grad_norm_pre_clip': 265.4586181640625, 'timing/data': 0.00025178000214509666, 'timing/model': 0.3937868659995729, 'learning_rate/qwen_vl_interface': 7.097247717693827e-06, 'learning_rate/action_model': 7.097247717693827e-05, 'epoch': 0.14}) +[2026-10-03 13:24:11,068][starVLA.training.train_starvla][INFO] - Step 1928, Loss: {'train/loss': 2.7870140911545604e-05, 'train/grad_norm_pre_clip': 0.010487666353583336, 'timing/data': 0.0003036400012206286, 'timing/model': 0.3945098149997648, 'learning_rate/qwen_vl_interface': 7.094441645042033e-06, 'learning_rate/action_model': 7.094441645042032e-05, 'epoch': 0.14}) +[2026-10-03 13:24:11,958][starVLA.training.train_starvla][INFO] - Step 1929, Loss: {'train/loss': 0.23213431239128113, 'train/grad_norm_pre_clip': 274.46728515625, 'timing/data': 0.0002730900014284998, 'timing/model': 0.5796336039929884, 'learning_rate/qwen_vl_interface': 7.09163481420995e-06, 'learning_rate/action_model': 7.09163481420995e-05, 'epoch': 0.14}) +[2026-10-03 13:24:12,664][starVLA.training.train_starvla][INFO] - Step 1930, Loss: {'train/loss': 0.4524837136268616, 'train/grad_norm_pre_clip': 69.79578399658203, 'timing/data': 0.00033629000245127827, 'timing/model': 0.3935784149944084, 'learning_rate/qwen_vl_interface': 7.0888272263513595e-06, 'learning_rate/action_model': 7.088827226351359e-05, 'epoch': 0.14}) +[2026-10-03 13:24:13,369][starVLA.training.train_starvla][INFO] - Step 1931, Loss: {'train/loss': 0.00018724967958405614, 'train/grad_norm_pre_clip': 0.11491818726062775, 'timing/data': 0.00039726999239064753, 'timing/model': 0.3935890249995282, 'learning_rate/qwen_vl_interface': 7.086018882620354e-06, 'learning_rate/action_model': 7.086018882620354e-05, 'epoch': 0.14}) +[2026-10-03 13:24:14,075][starVLA.training.train_starvla][INFO] - Step 1932, Loss: {'train/loss': 3.085075150011107e-05, 'train/grad_norm_pre_clip': 0.012875979766249657, 'timing/data': 0.00032934000773821026, 'timing/model': 0.39414316600596067, 'learning_rate/qwen_vl_interface': 7.083209784171341e-06, 'learning_rate/action_model': 7.083209784171341e-05, 'epoch': 0.14}) +[2026-10-03 13:24:14,779][starVLA.training.train_starvla][INFO] - Step 1933, Loss: {'train/loss': 0.000367151340469718, 'train/grad_norm_pre_clip': 0.2236335426568985, 'timing/data': 0.0002464599965605885, 'timing/model': 0.39222076600708533, 'learning_rate/qwen_vl_interface': 7.080399932159029e-06, 'learning_rate/action_model': 7.080399932159029e-05, 'epoch': 0.14}) +[2026-10-03 13:24:15,671][starVLA.training.train_starvla][INFO] - Step 1934, Loss: {'train/loss': 0.0002327959518879652, 'train/grad_norm_pre_clip': 0.2323485165834427, 'timing/data': 0.0002989999920828268, 'timing/model': 0.5822871039999882, 'learning_rate/qwen_vl_interface': 7.077589327738448e-06, 'learning_rate/action_model': 7.077589327738447e-05, 'epoch': 0.14}) +[2026-10-03 13:24:16,374][starVLA.training.train_starvla][INFO] - Step 1935, Loss: {'train/loss': 0.4108278453350067, 'train/grad_norm_pre_clip': 242.18545532226562, 'timing/data': 0.0003025799960596487, 'timing/model': 0.3932116060022963, 'learning_rate/qwen_vl_interface': 7.0747779720649265e-06, 'learning_rate/action_model': 7.074777972064927e-05, 'epoch': 0.14}) +[2026-10-03 13:24:17,079][starVLA.training.train_starvla][INFO] - Step 1936, Loss: {'train/loss': 8.847701246850193e-05, 'train/grad_norm_pre_clip': 0.06377402693033218, 'timing/data': 0.00037215001066215336, 'timing/model': 0.3931224659900181, 'learning_rate/qwen_vl_interface': 7.071965866294108e-06, 'learning_rate/action_model': 7.071965866294108e-05, 'epoch': 0.14}) +[2026-10-03 13:24:17,780][starVLA.training.train_starvla][INFO] - Step 1937, Loss: {'train/loss': 3.0507884730468504e-05, 'train/grad_norm_pre_clip': 0.009106085635721684, 'timing/data': 0.00024401000700891018, 'timing/model': 0.39248259500891436, 'learning_rate/qwen_vl_interface': 7.0691530115819454e-06, 'learning_rate/action_model': 7.069153011581945e-05, 'epoch': 0.14}) +[2026-10-03 13:24:18,482][starVLA.training.train_starvla][INFO] - Step 1938, Loss: {'train/loss': 0.33277803659439087, 'train/grad_norm_pre_clip': 61.158111572265625, 'timing/data': 0.00024718001077417284, 'timing/model': 0.3924713750020601, 'learning_rate/qwen_vl_interface': 7.066339409084694e-06, 'learning_rate/action_model': 7.066339409084694e-05, 'epoch': 0.14}) +[2026-10-03 13:24:19,185][starVLA.training.train_starvla][INFO] - Step 1939, Loss: {'train/loss': 6.072656105970964e-05, 'train/grad_norm_pre_clip': 0.024572143331170082, 'timing/data': 0.00031165999826043844, 'timing/model': 0.3926986059959745, 'learning_rate/qwen_vl_interface': 7.0635250599589215e-06, 'learning_rate/action_model': 7.06352505995892e-05, 'epoch': 0.14}) +[2026-10-03 13:24:20,078][starVLA.training.train_starvla][INFO] - Step 1940, Loss: {'train/loss': 1.3537361155613326e-05, 'train/grad_norm_pre_clip': 0.002610709983855486, 'timing/data': 0.0002903800050262362, 'timing/model': 0.39257719500164967, 'learning_rate/qwen_vl_interface': 7.060709965361499e-06, 'learning_rate/action_model': 7.060709965361499e-05, 'epoch': 0.14}) +[2026-10-03 13:24:20,779][starVLA.training.train_starvla][INFO] - Step 1941, Loss: {'train/loss': 0.3406582474708557, 'train/grad_norm_pre_clip': 74.653564453125, 'timing/data': 0.0001999200030695647, 'timing/model': 0.3910927060060203, 'learning_rate/qwen_vl_interface': 7.057894126449606e-06, 'learning_rate/action_model': 7.057894126449606e-05, 'epoch': 0.14}) +[2026-10-03 13:24:21,483][starVLA.training.train_starvla][INFO] - Step 1942, Loss: {'train/loss': 0.5203350186347961, 'train/grad_norm_pre_clip': 40.03333282470703, 'timing/data': 0.0003157500032102689, 'timing/model': 0.39330561499809846, 'learning_rate/qwen_vl_interface': 7.055077544380728e-06, 'learning_rate/action_model': 7.055077544380728e-05, 'epoch': 0.14}) +[2026-10-03 13:24:22,185][starVLA.training.train_starvla][INFO] - Step 1943, Loss: {'train/loss': 0.00023937539663165808, 'train/grad_norm_pre_clip': 0.03802148252725601, 'timing/data': 0.00032048999855760485, 'timing/model': 0.39296452498820145, 'learning_rate/qwen_vl_interface': 7.0522602203126554e-06, 'learning_rate/action_model': 7.052260220312656e-05, 'epoch': 0.14}) +[2026-10-03 13:24:22,889][starVLA.training.train_starvla][INFO] - Step 1944, Loss: {'train/loss': 0.01882510632276535, 'train/grad_norm_pre_clip': 34.93529510498047, 'timing/data': 0.0003334900102345273, 'timing/model': 0.39324732500244863, 'learning_rate/qwen_vl_interface': 7.0494421554034845e-06, 'learning_rate/action_model': 7.049442155403484e-05, 'epoch': 0.14}) +[2026-10-03 13:24:23,786][starVLA.training.train_starvla][INFO] - Step 1945, Loss: {'train/loss': 0.5833690762519836, 'train/grad_norm_pre_clip': 50.797691345214844, 'timing/data': 0.00023257000430021435, 'timing/model': 0.3926658259879332, 'learning_rate/qwen_vl_interface': 7.046623350811612e-06, 'learning_rate/action_model': 7.046623350811612e-05, 'epoch': 0.14}) +[2026-10-03 13:24:24,491][starVLA.training.train_starvla][INFO] - Step 1946, Loss: {'train/loss': 0.006618613842874765, 'train/grad_norm_pre_clip': 15.082881927490234, 'timing/data': 0.0003703099937411025, 'timing/model': 0.393124185007764, 'learning_rate/qwen_vl_interface': 7.043803807695747e-06, 'learning_rate/action_model': 7.043803807695747e-05, 'epoch': 0.14}) +[2026-10-03 13:24:25,194][starVLA.training.train_starvla][INFO] - Step 1947, Loss: {'train/loss': 0.0002246332005597651, 'train/grad_norm_pre_clip': 0.10672356933355331, 'timing/data': 0.0003403000009711832, 'timing/model': 0.3930146960046841, 'learning_rate/qwen_vl_interface': 7.040983527214893e-06, 'learning_rate/action_model': 7.040983527214893e-05, 'epoch': 0.14}) +[2026-10-03 13:24:25,899][starVLA.training.train_starvla][INFO] - Step 1948, Loss: {'train/loss': 0.14095760881900787, 'train/grad_norm_pre_clip': 106.0020751953125, 'timing/data': 0.0003759489918593317, 'timing/model': 0.3932966559950728, 'learning_rate/qwen_vl_interface': 7.038162510528363e-06, 'learning_rate/action_model': 7.038162510528362e-05, 'epoch': 0.14}) +[2026-10-03 13:24:26,604][starVLA.training.train_starvla][INFO] - Step 1949, Loss: {'train/loss': 0.000154179084347561, 'train/grad_norm_pre_clip': 0.0360393151640892, 'timing/data': 0.00027122000756207854, 'timing/model': 0.3933657250017859, 'learning_rate/qwen_vl_interface': 7.03534075879577e-06, 'learning_rate/action_model': 7.03534075879577e-05, 'epoch': 0.14}) +[2026-10-03 13:24:27,498][starVLA.training.train_starvla][INFO] - Step 1950, Loss: {'train/loss': 8.682695624884218e-05, 'train/grad_norm_pre_clip': 0.023862572386860847, 'timing/data': 0.00029644000460393727, 'timing/model': 0.3938143150007818, 'learning_rate/qwen_vl_interface': 7.032518273177028e-06, 'learning_rate/action_model': 7.032518273177028e-05, 'epoch': 0.14}) +[2026-10-03 13:24:28,201][starVLA.training.train_starvla][INFO] - Step 1951, Loss: {'train/loss': 0.2284715473651886, 'train/grad_norm_pre_clip': 121.52767944335938, 'timing/data': 0.0002529900084482506, 'timing/model': 0.3930845059949206, 'learning_rate/qwen_vl_interface': 7.029695054832357e-06, 'learning_rate/action_model': 7.029695054832357e-05, 'epoch': 0.14}) +[2026-10-03 13:24:28,906][starVLA.training.train_starvla][INFO] - Step 1952, Loss: {'train/loss': 0.0023812910076230764, 'train/grad_norm_pre_clip': 4.219796180725098, 'timing/data': 0.0003509600064717233, 'timing/model': 0.39310797600774094, 'learning_rate/qwen_vl_interface': 7.026871104922273e-06, 'learning_rate/action_model': 7.026871104922273e-05, 'epoch': 0.14}) +[2026-10-03 13:24:29,609][starVLA.training.train_starvla][INFO] - Step 1953, Loss: {'train/loss': 0.029830018058419228, 'train/grad_norm_pre_clip': 37.92295455932617, 'timing/data': 0.00022667000303044915, 'timing/model': 0.39257295499555767, 'learning_rate/qwen_vl_interface': 7.024046424607596e-06, 'learning_rate/action_model': 7.024046424607595e-05, 'epoch': 0.14}) +[2026-10-03 13:24:30,312][starVLA.training.train_starvla][INFO] - Step 1954, Loss: {'train/loss': 0.00010066917457152158, 'train/grad_norm_pre_clip': 0.03917158767580986, 'timing/data': 0.0002968299959320575, 'timing/model': 0.39310970601218287, 'learning_rate/qwen_vl_interface': 7.021221015049445e-06, 'learning_rate/action_model': 7.021221015049446e-05, 'epoch': 0.14}) +[2026-10-03 13:24:31,207][starVLA.training.train_starvla][INFO] - Step 1955, Loss: {'train/loss': 0.00016112675075419247, 'train/grad_norm_pre_clip': 0.036261096596717834, 'timing/data': 0.00025221999385394156, 'timing/model': 0.5838310340041062, 'learning_rate/qwen_vl_interface': 7.01839487740924e-06, 'learning_rate/action_model': 7.01839487740924e-05, 'epoch': 0.14}) +[2026-10-03 13:24:31,912][starVLA.training.train_starvla][INFO] - Step 1956, Loss: {'train/loss': 0.06034957244992256, 'train/grad_norm_pre_clip': 52.92219161987305, 'timing/data': 0.0003192000003764406, 'timing/model': 0.3933721659996081, 'learning_rate/qwen_vl_interface': 7.015568012848699e-06, 'learning_rate/action_model': 7.015568012848699e-05, 'epoch': 0.14}) +[2026-10-03 13:24:32,614][starVLA.training.train_starvla][INFO] - Step 1957, Loss: {'train/loss': 0.00015068052744027227, 'train/grad_norm_pre_clip': 0.03829994052648544, 'timing/data': 0.0003155799931846559, 'timing/model': 0.39232750500377733, 'learning_rate/qwen_vl_interface': 7.01274042252984e-06, 'learning_rate/action_model': 7.01274042252984e-05, 'epoch': 0.14}) +[2026-10-03 13:24:33,319][starVLA.training.train_starvla][INFO] - Step 1958, Loss: {'train/loss': 0.007262186612933874, 'train/grad_norm_pre_clip': 6.267294406890869, 'timing/data': 0.0003562199999578297, 'timing/model': 0.3933083349984372, 'learning_rate/qwen_vl_interface': 7.009912107614978e-06, 'learning_rate/action_model': 7.009912107614978e-05, 'epoch': 0.14}) +[2026-10-03 13:24:34,022][starVLA.training.train_starvla][INFO] - Step 1959, Loss: {'train/loss': 0.00014164112508296967, 'train/grad_norm_pre_clip': 0.13667044043540955, 'timing/data': 0.00026667000201996416, 'timing/model': 0.39260356599697843, 'learning_rate/qwen_vl_interface': 7.0070830692667244e-06, 'learning_rate/action_model': 7.007083069266724e-05, 'epoch': 0.14}) +[2026-10-03 13:24:34,919][starVLA.training.train_starvla][INFO] - Step 1960, Loss: {'train/loss': 0.055543381720781326, 'train/grad_norm_pre_clip': 50.171871185302734, 'timing/data': 0.00035334999847691506, 'timing/model': 0.5843900140025653, 'learning_rate/qwen_vl_interface': 7.004253308647993e-06, 'learning_rate/action_model': 7.004253308647994e-05, 'epoch': 0.14}) +[2026-10-03 13:24:35,622][starVLA.training.train_starvla][INFO] - Step 1961, Loss: {'train/loss': 0.0004330791998654604, 'train/grad_norm_pre_clip': 0.2254786491394043, 'timing/data': 0.0003094300045631826, 'timing/model': 0.39263975599897094, 'learning_rate/qwen_vl_interface': 7.001422826921991e-06, 'learning_rate/action_model': 7.00142282692199e-05, 'epoch': 0.14}) +[2026-10-03 13:24:36,326][starVLA.training.train_starvla][INFO] - Step 1962, Loss: {'train/loss': 0.2731000781059265, 'train/grad_norm_pre_clip': 102.89053344726562, 'timing/data': 0.0002312099968548864, 'timing/model': 0.393176265002694, 'learning_rate/qwen_vl_interface': 6.9985916252522225e-06, 'learning_rate/action_model': 6.998591625252222e-05, 'epoch': 0.14}) +[2026-10-03 13:24:37,029][starVLA.training.train_starvla][INFO] - Step 1963, Loss: {'train/loss': 0.0008357465267181396, 'train/grad_norm_pre_clip': 0.5841527581214905, 'timing/data': 0.00020200999279040843, 'timing/model': 0.3931607359991176, 'learning_rate/qwen_vl_interface': 6.995759704802485e-06, 'learning_rate/action_model': 6.995759704802485e-05, 'epoch': 0.14}) +[2026-10-03 13:24:37,734][starVLA.training.train_starvla][INFO] - Step 1964, Loss: {'train/loss': 9.596185554983094e-05, 'train/grad_norm_pre_clip': 0.01502927951514721, 'timing/data': 0.00024121999740600586, 'timing/model': 0.3937189760035835, 'learning_rate/qwen_vl_interface': 6.992927066736876e-06, 'learning_rate/action_model': 6.992927066736875e-05, 'epoch': 0.14}) +[2026-10-03 13:24:38,437][starVLA.training.train_starvla][INFO] - Step 1965, Loss: {'train/loss': 0.00010448378452565521, 'train/grad_norm_pre_clip': 0.06018949672579765, 'timing/data': 0.0003143700014334172, 'timing/model': 0.3925313539948547, 'learning_rate/qwen_vl_interface': 6.990093712219787e-06, 'learning_rate/action_model': 6.990093712219787e-05, 'epoch': 0.14}) +[2026-10-03 13:24:39,332][starVLA.training.train_starvla][INFO] - Step 1966, Loss: {'train/loss': 0.023337045684456825, 'train/grad_norm_pre_clip': 32.96746826171875, 'timing/data': 0.0003353100037202239, 'timing/model': 0.39321903500240296, 'learning_rate/qwen_vl_interface': 6.9872596424159e-06, 'learning_rate/action_model': 6.9872596424159e-05, 'epoch': 0.14}) +[2026-10-03 13:24:40,035][starVLA.training.train_starvla][INFO] - Step 1967, Loss: {'train/loss': 9.700408554635942e-06, 'train/grad_norm_pre_clip': 0.0016236366936936975, 'timing/data': 0.00022783999156672508, 'timing/model': 0.392632606002735, 'learning_rate/qwen_vl_interface': 6.9844248584902e-06, 'learning_rate/action_model': 6.984424858490199e-05, 'epoch': 0.14}) +[2026-10-03 13:24:40,740][starVLA.training.train_starvla][INFO] - Step 1968, Loss: {'train/loss': 0.006772379856556654, 'train/grad_norm_pre_clip': 13.290533065795898, 'timing/data': 0.00023705999774392694, 'timing/model': 0.39322277599421795, 'learning_rate/qwen_vl_interface': 6.981589361607952e-06, 'learning_rate/action_model': 6.981589361607953e-05, 'epoch': 0.14}) +[2026-10-03 13:24:41,444][starVLA.training.train_starvla][INFO] - Step 1969, Loss: {'train/loss': 1.8640741473063827e-05, 'train/grad_norm_pre_clip': 0.0049528563395142555, 'timing/data': 0.0002017599908867851, 'timing/model': 0.393125365008018, 'learning_rate/qwen_vl_interface': 6.9787531529347274e-06, 'learning_rate/action_model': 6.978753152934727e-05, 'epoch': 0.14}) +[2026-10-03 13:24:42,148][starVLA.training.train_starvla][INFO] - Step 1970, Loss: {'train/loss': 5.165960465092212e-05, 'train/grad_norm_pre_clip': 0.008483587764203548, 'timing/data': 0.0003224600077373907, 'timing/model': 0.39352774499275256, 'learning_rate/qwen_vl_interface': 6.97591623363638e-06, 'learning_rate/action_model': 6.97591623363638e-05, 'epoch': 0.14}) +[2026-10-03 13:24:43,039][starVLA.training.train_starvla][INFO] - Step 1971, Loss: {'train/loss': 2.4779699742794037e-05, 'train/grad_norm_pre_clip': 0.004679690580815077, 'timing/data': 0.0003156200109515339, 'timing/model': 0.5801494740007911, 'learning_rate/qwen_vl_interface': 6.9730786048790644e-06, 'learning_rate/action_model': 6.973078604879064e-05, 'epoch': 0.14}) +[2026-10-03 13:24:43,744][starVLA.training.train_starvla][INFO] - Step 1972, Loss: {'train/loss': 0.00015735405031591654, 'train/grad_norm_pre_clip': 0.19459135830402374, 'timing/data': 0.00026011999580077827, 'timing/model': 0.39340140600688756, 'learning_rate/qwen_vl_interface': 6.970240267829221e-06, 'learning_rate/action_model': 6.970240267829222e-05, 'epoch': 0.14}) +[2026-10-03 13:24:44,448][starVLA.training.train_starvla][INFO] - Step 1973, Loss: {'train/loss': 1.928886558744125e-05, 'train/grad_norm_pre_clip': 0.013732453808188438, 'timing/data': 0.0003385199961485341, 'timing/model': 0.39265527499082964, 'learning_rate/qwen_vl_interface': 6.967401223653582e-06, 'learning_rate/action_model': 6.967401223653581e-05, 'epoch': 0.14}) +[2026-10-03 13:24:45,154][starVLA.training.train_starvla][INFO] - Step 1974, Loss: {'train/loss': 9.662044612923637e-05, 'train/grad_norm_pre_clip': 0.10139032453298569, 'timing/data': 0.0003623500087996945, 'timing/model': 0.39340181500301696, 'learning_rate/qwen_vl_interface': 6.9645614735191724e-06, 'learning_rate/action_model': 6.964561473519172e-05, 'epoch': 0.14}) +[2026-10-03 13:24:45,858][starVLA.training.train_starvla][INFO] - Step 1975, Loss: {'train/loss': 0.0001421127381036058, 'train/grad_norm_pre_clip': 0.03553533926606178, 'timing/data': 0.00031694999779574573, 'timing/model': 0.3929299150040606, 'learning_rate/qwen_vl_interface': 6.961721018593305e-06, 'learning_rate/action_model': 6.961721018593305e-05, 'epoch': 0.14}) +[2026-10-03 13:24:46,563][starVLA.training.train_starvla][INFO] - Step 1976, Loss: {'train/loss': 9.759828571986873e-06, 'train/grad_norm_pre_clip': 0.005506701301783323, 'timing/data': 0.00038020999636501074, 'timing/model': 0.39342471600684803, 'learning_rate/qwen_vl_interface': 6.958879860043585e-06, 'learning_rate/action_model': 6.958879860043585e-05, 'epoch': 0.14}) +[2026-10-03 13:24:47,458][starVLA.training.train_starvla][INFO] - Step 1977, Loss: {'train/loss': 0.5390195846557617, 'train/grad_norm_pre_clip': 80.1893539428711, 'timing/data': 0.00023689000227022916, 'timing/model': 0.39339860499603674, 'learning_rate/qwen_vl_interface': 6.9560379990379025e-06, 'learning_rate/action_model': 6.956037999037902e-05, 'epoch': 0.14}) +[2026-10-03 13:24:48,162][starVLA.training.train_starvla][INFO] - Step 1978, Loss: {'train/loss': 0.18218448758125305, 'train/grad_norm_pre_clip': 42.41885757446289, 'timing/data': 0.00028234999626874924, 'timing/model': 0.392999335002969, 'learning_rate/qwen_vl_interface': 6.953195436744442e-06, 'learning_rate/action_model': 6.953195436744442e-05, 'epoch': 0.14}) +[2026-10-03 13:24:48,866][starVLA.training.train_starvla][INFO] - Step 1979, Loss: {'train/loss': 0.0011177759151905775, 'train/grad_norm_pre_clip': 0.5102604031562805, 'timing/data': 0.0002443999983370304, 'timing/model': 0.3926511860045139, 'learning_rate/qwen_vl_interface': 6.950352174331676e-06, 'learning_rate/action_model': 6.950352174331675e-05, 'epoch': 0.14}) +[2026-10-03 13:24:49,571][starVLA.training.train_starvla][INFO] - Step 1980, Loss: {'train/loss': 0.0006898796418681741, 'train/grad_norm_pre_clip': 0.26511237025260925, 'timing/data': 0.00025334999372716993, 'timing/model': 0.3931478349986719, 'learning_rate/qwen_vl_interface': 6.947508212968356e-06, 'learning_rate/action_model': 6.947508212968356e-05, 'epoch': 0.14}) +[2026-10-03 13:24:50,271][starVLA.training.train_starvla][INFO] - Step 1981, Loss: {'train/loss': 9.648281775298528e-06, 'train/grad_norm_pre_clip': 0.0088350186124444, 'timing/data': 0.0002796199987642467, 'timing/model': 0.39072678600496147, 'learning_rate/qwen_vl_interface': 6.9446635538235295e-06, 'learning_rate/action_model': 6.944663553823529e-05, 'epoch': 0.14}) +[2026-10-03 13:24:51,167][starVLA.training.train_starvla][INFO] - Step 1982, Loss: {'train/loss': 5.793589662062004e-05, 'train/grad_norm_pre_clip': 0.04089251905679703, 'timing/data': 0.00026369999977760017, 'timing/model': 0.39304198599711526, 'learning_rate/qwen_vl_interface': 6.94181819806653e-06, 'learning_rate/action_model': 6.94181819806653e-05, 'epoch': 0.15}) +[2026-10-03 13:24:51,871][starVLA.training.train_starvla][INFO] - Step 1983, Loss: {'train/loss': 2.7661004423862323e-05, 'train/grad_norm_pre_clip': 0.03486521169543266, 'timing/data': 0.0003297100047348067, 'timing/model': 0.3931402759917546, 'learning_rate/qwen_vl_interface': 6.938972146866975e-06, 'learning_rate/action_model': 6.938972146866974e-05, 'epoch': 0.15}) +[2026-10-03 13:24:52,579][starVLA.training.train_starvla][INFO] - Step 1984, Loss: {'train/loss': 7.0928163040662184e-06, 'train/grad_norm_pre_clip': 0.0017977317329496145, 'timing/data': 0.00024111999664455652, 'timing/model': 0.3943867859925376, 'learning_rate/qwen_vl_interface': 6.936125401394769e-06, 'learning_rate/action_model': 6.936125401394768e-05, 'epoch': 0.15}) +[2026-10-03 13:24:53,282][starVLA.training.train_starvla][INFO] - Step 1985, Loss: {'train/loss': 3.18971615342889e-05, 'train/grad_norm_pre_clip': 0.01077578030526638, 'timing/data': 0.00031392999517265707, 'timing/model': 0.39272256499680225, 'learning_rate/qwen_vl_interface': 6.933277962820101e-06, 'learning_rate/action_model': 6.9332779628201e-05, 'epoch': 0.15}) +[2026-10-03 13:24:53,987][starVLA.training.train_starvla][INFO] - Step 1986, Loss: {'train/loss': 2.536706415412482e-05, 'train/grad_norm_pre_clip': 0.009075731039047241, 'timing/data': 0.00023501999385189265, 'timing/model': 0.3931133150035748, 'learning_rate/qwen_vl_interface': 6.930429832313444e-06, 'learning_rate/action_model': 6.930429832313444e-05, 'epoch': 0.15}) +[2026-10-03 13:24:54,883][starVLA.training.train_starvla][INFO] - Step 1987, Loss: {'train/loss': 2.2036925656720996e-05, 'train/grad_norm_pre_clip': 0.014058123342692852, 'timing/data': 0.0002760399947874248, 'timing/model': 0.582101774009061, 'learning_rate/qwen_vl_interface': 6.92758101104556e-06, 'learning_rate/action_model': 6.92758101104556e-05, 'epoch': 0.15}) +[2026-10-03 13:24:55,588][starVLA.training.train_starvla][INFO] - Step 1988, Loss: {'train/loss': 5.538752884604037e-05, 'train/grad_norm_pre_clip': 0.01944015361368656, 'timing/data': 0.0002721190103329718, 'timing/model': 0.39350338600343093, 'learning_rate/qwen_vl_interface': 6.9247315001874885e-06, 'learning_rate/action_model': 6.924731500187488e-05, 'epoch': 0.15}) +[2026-10-03 13:24:56,291][starVLA.training.train_starvla][INFO] - Step 1989, Loss: {'train/loss': 0.03440311923623085, 'train/grad_norm_pre_clip': 17.302085876464844, 'timing/data': 0.00034224000410176814, 'timing/model': 0.3928309049952077, 'learning_rate/qwen_vl_interface': 6.9218813009105605e-06, 'learning_rate/action_model': 6.92188130091056e-05, 'epoch': 0.15}) +[2026-10-03 13:24:56,996][starVLA.training.train_starvla][INFO] - Step 1990, Loss: {'train/loss': 1.862643159711297e-07, 'train/grad_norm_pre_clip': 4.872029967373237e-05, 'timing/data': 0.0002474000066285953, 'timing/model': 0.3927591450046748, 'learning_rate/qwen_vl_interface': 6.919030414386378e-06, 'learning_rate/action_model': 6.919030414386379e-05, 'epoch': 0.15}) +[2026-10-03 13:24:57,699][starVLA.training.train_starvla][INFO] - Step 1991, Loss: {'train/loss': 0.00014397116319742054, 'train/grad_norm_pre_clip': 0.06689989566802979, 'timing/data': 0.0002584100002422929, 'timing/model': 0.39251455599151086, 'learning_rate/qwen_vl_interface': 6.916178841786839e-06, 'learning_rate/action_model': 6.916178841786838e-05, 'epoch': 0.15}) +[2026-10-03 13:24:58,404][starVLA.training.train_starvla][INFO] - Step 1992, Loss: {'train/loss': 0.5286636352539062, 'train/grad_norm_pre_clip': 76.77133178710938, 'timing/data': 0.0003108200035057962, 'timing/model': 0.39289138599997386, 'learning_rate/qwen_vl_interface': 6.913326584284115e-06, 'learning_rate/action_model': 6.913326584284115e-05, 'epoch': 0.15}) +[2026-10-03 13:24:59,295][starVLA.training.train_starvla][INFO] - Step 1993, Loss: {'train/loss': 7.474891754100099e-05, 'train/grad_norm_pre_clip': 0.052265435457229614, 'timing/data': 0.0002315599995199591, 'timing/model': 0.39306147399474867, 'learning_rate/qwen_vl_interface': 6.910473643050662e-06, 'learning_rate/action_model': 6.910473643050662e-05, 'epoch': 0.15}) +[2026-10-03 13:25:00,000][starVLA.training.train_starvla][INFO] - Step 1994, Loss: {'train/loss': 0.005287305451929569, 'train/grad_norm_pre_clip': 3.43814754486084, 'timing/data': 0.0003574500005925074, 'timing/model': 0.3932232860097429, 'learning_rate/qwen_vl_interface': 6.907620019259215e-06, 'learning_rate/action_model': 6.907620019259216e-05, 'epoch': 0.15}) +[2026-10-03 13:25:00,702][starVLA.training.train_starvla][INFO] - Step 1995, Loss: {'train/loss': 1.4698754057462793e-05, 'train/grad_norm_pre_clip': 0.01822730340063572, 'timing/data': 0.00037652000901289284, 'timing/model': 0.39292682500672527, 'learning_rate/qwen_vl_interface': 6.9047657140827935e-06, 'learning_rate/action_model': 6.904765714082794e-05, 'epoch': 0.15}) +[2026-10-03 13:25:01,408][starVLA.training.train_starvla][INFO] - Step 1996, Loss: {'train/loss': 3.7550203160208184e-06, 'train/grad_norm_pre_clip': 0.00407030712813139, 'timing/data': 0.000311999989207834, 'timing/model': 0.39346879499498755, 'learning_rate/qwen_vl_interface': 6.901910728694694e-06, 'learning_rate/action_model': 6.901910728694694e-05, 'epoch': 0.15}) +[2026-10-03 13:25:02,111][starVLA.training.train_starvla][INFO] - Step 1997, Loss: {'train/loss': 0.006205007433891296, 'train/grad_norm_pre_clip': 17.932416915893555, 'timing/data': 0.0003152299905195832, 'timing/model': 0.3923446559929289, 'learning_rate/qwen_vl_interface': 6.899055064268489e-06, 'learning_rate/action_model': 6.899055064268489e-05, 'epoch': 0.15}) +[2026-10-03 13:25:03,007][starVLA.training.train_starvla][INFO] - Step 1998, Loss: {'train/loss': 0.5550695061683655, 'train/grad_norm_pre_clip': 49.456138610839844, 'timing/data': 0.00032172999635804445, 'timing/model': 0.3936718449986074, 'learning_rate/qwen_vl_interface': 6.896198721978038e-06, 'learning_rate/action_model': 6.896198721978038e-05, 'epoch': 0.15}) +[2026-10-03 13:25:03,710][starVLA.training.train_starvla][INFO] - Step 1999, Loss: {'train/loss': 2.6633771994966082e-05, 'train/grad_norm_pre_clip': 0.02096572518348694, 'timing/data': 0.00020788000256288797, 'timing/model': 0.3928819049906451, 'learning_rate/qwen_vl_interface': 6.893341702997475e-06, 'learning_rate/action_model': 6.893341702997475e-05, 'epoch': 0.15}) +[2026-10-03 13:25:04,415][starVLA.training.train_starvla][INFO] - Step 2000, Loss: {'train/loss': 0.12274599075317383, 'train/grad_norm_pre_clip': 43.536834716796875, 'timing/data': 0.0002115399984177202, 'timing/model': 0.3931057959998725, 'learning_rate/qwen_vl_interface': 6.890484008501211e-06, 'learning_rate/action_model': 6.890484008501211e-05, 'epoch': 0.15}) +[2026-10-03 13:25:05,118][starVLA.training.train_starvla][INFO] - Step 2001, Loss: {'train/loss': 0.4078577160835266, 'train/grad_norm_pre_clip': 64.021728515625, 'timing/data': 0.00033193999843206257, 'timing/model': 0.39237619600316975, 'learning_rate/qwen_vl_interface': 6.887625639663941e-06, 'learning_rate/action_model': 6.88762563966394e-05, 'epoch': 0.15}) +[2026-10-03 13:25:05,822][starVLA.training.train_starvla][INFO] - Step 2002, Loss: {'train/loss': 8.701819751877338e-06, 'train/grad_norm_pre_clip': 0.00451078824698925, 'timing/data': 0.0002526199968997389, 'timing/model': 0.39342363498872146, 'learning_rate/qwen_vl_interface': 6.884766597660626e-06, 'learning_rate/action_model': 6.884766597660626e-05, 'epoch': 0.15}) +[2026-10-03 13:25:06,719][starVLA.training.train_starvla][INFO] - Step 2003, Loss: {'train/loss': 1.616766439838102e-06, 'train/grad_norm_pre_clip': 0.00034783349838107824, 'timing/data': 0.00036267000541556627, 'timing/model': 0.39363534600124694, 'learning_rate/qwen_vl_interface': 6.881906883666515e-06, 'learning_rate/action_model': 6.881906883666515e-05, 'epoch': 0.15}) +[2026-10-03 13:25:07,424][starVLA.training.train_starvla][INFO] - Step 2004, Loss: {'train/loss': 2.4100654627545737e-05, 'train/grad_norm_pre_clip': 0.013661650009453297, 'timing/data': 0.00026717000582721084, 'timing/model': 0.3928123749938095, 'learning_rate/qwen_vl_interface': 6.879046498857125e-06, 'learning_rate/action_model': 6.879046498857124e-05, 'epoch': 0.15}) +[2026-10-03 13:25:08,127][starVLA.training.train_starvla][INFO] - Step 2005, Loss: {'train/loss': 0.0011972435750067234, 'train/grad_norm_pre_clip': 1.1662921905517578, 'timing/data': 0.00031319999834522605, 'timing/model': 0.3925915360014187, 'learning_rate/qwen_vl_interface': 6.876185444408255e-06, 'learning_rate/action_model': 6.876185444408255e-05, 'epoch': 0.15}) +[2026-10-03 13:25:08,831][starVLA.training.train_starvla][INFO] - Step 2006, Loss: {'train/loss': 8.241059549618512e-05, 'train/grad_norm_pre_clip': 0.07376113533973694, 'timing/data': 0.0003659899957710877, 'timing/model': 0.39290062501095235, 'learning_rate/qwen_vl_interface': 6.873323721495976e-06, 'learning_rate/action_model': 6.873323721495976e-05, 'epoch': 0.15}) +[2026-10-03 13:25:09,535][starVLA.training.train_starvla][INFO] - Step 2007, Loss: {'train/loss': 0.4453548192977905, 'train/grad_norm_pre_clip': 52.920528411865234, 'timing/data': 0.00020604999735951424, 'timing/model': 0.3934749250038294, 'learning_rate/qwen_vl_interface': 6.870461331296632e-06, 'learning_rate/action_model': 6.870461331296632e-05, 'epoch': 0.15}) +[2026-10-03 13:25:10,432][starVLA.training.train_starvla][INFO] - Step 2008, Loss: {'train/loss': 0.6894568204879761, 'train/grad_norm_pre_clip': 37.41624450683594, 'timing/data': 0.00020978999964427203, 'timing/model': 0.5852360440039774, 'learning_rate/qwen_vl_interface': 6.8675982749868455e-06, 'learning_rate/action_model': 6.867598274986845e-05, 'epoch': 0.15}) +[2026-10-03 13:25:11,136][starVLA.training.train_starvla][INFO] - Step 2009, Loss: {'train/loss': 2.649939233378973e-05, 'train/grad_norm_pre_clip': 0.00930820219218731, 'timing/data': 0.0003187400143360719, 'timing/model': 0.39292122499318793, 'learning_rate/qwen_vl_interface': 6.864734553743509e-06, 'learning_rate/action_model': 6.86473455374351e-05, 'epoch': 0.15}) +[2026-10-03 13:25:11,840][starVLA.training.train_starvla][INFO] - Step 2010, Loss: {'train/loss': 2.2961208742344752e-05, 'train/grad_norm_pre_clip': 0.03491688892245293, 'timing/data': 0.0003486700006760657, 'timing/model': 0.3930112649977673, 'learning_rate/qwen_vl_interface': 6.861870168743793e-06, 'learning_rate/action_model': 6.861870168743793e-05, 'epoch': 0.15}) +[2026-10-03 13:25:12,542][starVLA.training.train_starvla][INFO] - Step 2011, Loss: {'train/loss': 0.25477302074432373, 'train/grad_norm_pre_clip': 44.006134033203125, 'timing/data': 0.00028078000468667597, 'timing/model': 0.39265462598996237, 'learning_rate/qwen_vl_interface': 6.859005121165135e-06, 'learning_rate/action_model': 6.859005121165135e-05, 'epoch': 0.15}) +[2026-10-03 13:25:13,248][starVLA.training.train_starvla][INFO] - Step 2012, Loss: {'train/loss': 2.020471220021136e-05, 'train/grad_norm_pre_clip': 0.014214043505489826, 'timing/data': 0.00027351999597158283, 'timing/model': 0.39335793600184843, 'learning_rate/qwen_vl_interface': 6.85613941218525e-06, 'learning_rate/action_model': 6.85613941218525e-05, 'epoch': 0.15}) +[2026-10-03 13:25:14,142][starVLA.training.train_starvla][INFO] - Step 2013, Loss: {'train/loss': 0.0006362695130519569, 'train/grad_norm_pre_clip': 0.35804876685142517, 'timing/data': 0.00020692999532911927, 'timing/model': 0.5832504040008644, 'learning_rate/qwen_vl_interface': 6.8532730429821215e-06, 'learning_rate/action_model': 6.853273042982122e-05, 'epoch': 0.15}) +[2026-10-03 13:25:14,846][starVLA.training.train_starvla][INFO] - Step 2014, Loss: {'train/loss': 0.0009808404138311744, 'train/grad_norm_pre_clip': 0.6319897174835205, 'timing/data': 0.0002544799936003983, 'timing/model': 0.39300491599715315, 'learning_rate/qwen_vl_interface': 6.850406014734005e-06, 'learning_rate/action_model': 6.850406014734005e-05, 'epoch': 0.15}) +[2026-10-03 13:25:15,548][starVLA.training.train_starvla][INFO] - Step 2015, Loss: {'train/loss': 0.046208713203668594, 'train/grad_norm_pre_clip': 31.157814025878906, 'timing/data': 0.00021000001288484782, 'timing/model': 0.39260656500118785, 'learning_rate/qwen_vl_interface': 6.8475383286194295e-06, 'learning_rate/action_model': 6.84753832861943e-05, 'epoch': 0.15}) +[2026-10-03 13:25:16,253][starVLA.training.train_starvla][INFO] - Step 2016, Loss: {'train/loss': 0.11331503093242645, 'train/grad_norm_pre_clip': 38.15183639526367, 'timing/data': 0.0003117300075246021, 'timing/model': 0.3929443450033432, 'learning_rate/qwen_vl_interface': 6.844669985817192e-06, 'learning_rate/action_model': 6.844669985817192e-05, 'epoch': 0.15}) +[2026-10-03 13:25:16,957][starVLA.training.train_starvla][INFO] - Step 2017, Loss: {'train/loss': 0.00011732285202015191, 'train/grad_norm_pre_clip': 0.03683294355869293, 'timing/data': 0.0005774199962615967, 'timing/model': 0.3929282249882817, 'learning_rate/qwen_vl_interface': 6.8418009875063594e-06, 'learning_rate/action_model': 6.84180098750636e-05, 'epoch': 0.15}) +[2026-10-03 13:25:17,662][starVLA.training.train_starvla][INFO] - Step 2018, Loss: {'train/loss': 0.41546082496643066, 'train/grad_norm_pre_clip': 101.6612777709961, 'timing/data': 0.00036189999082125723, 'timing/model': 0.39295094599947333, 'learning_rate/qwen_vl_interface': 6.838931334866267e-06, 'learning_rate/action_model': 6.838931334866266e-05, 'epoch': 0.15}) +[2026-10-03 13:25:18,556][starVLA.training.train_starvla][INFO] - Step 2019, Loss: {'train/loss': 3.848614142043516e-05, 'train/grad_norm_pre_clip': 0.008717416785657406, 'timing/data': 0.00021321000531315804, 'timing/model': 0.39216585499525536, 'learning_rate/qwen_vl_interface': 6.836061029076523e-06, 'learning_rate/action_model': 6.836061029076522e-05, 'epoch': 0.15}) +[2026-10-03 13:25:19,259][starVLA.training.train_starvla][INFO] - Step 2020, Loss: {'train/loss': 5.205674096941948e-05, 'train/grad_norm_pre_clip': 0.010849570855498314, 'timing/data': 0.00035627999750431627, 'timing/model': 0.39237940601015, 'learning_rate/qwen_vl_interface': 6.833190071317e-06, 'learning_rate/action_model': 6.833190071316999e-05, 'epoch': 0.15}) +[2026-10-03 13:25:19,962][starVLA.training.train_starvla][INFO] - Step 2021, Loss: {'train/loss': 0.27609410881996155, 'train/grad_norm_pre_clip': 96.0260009765625, 'timing/data': 0.0002509800106054172, 'timing/model': 0.39241977500205394, 'learning_rate/qwen_vl_interface': 6.830318462767841e-06, 'learning_rate/action_model': 6.830318462767841e-05, 'epoch': 0.15}) +[2026-10-03 13:25:20,665][starVLA.training.train_starvla][INFO] - Step 2022, Loss: {'train/loss': 5.5311640608124435e-05, 'train/grad_norm_pre_clip': 0.012775546871125698, 'timing/data': 0.0003279800002928823, 'timing/model': 0.39293196600920055, 'learning_rate/qwen_vl_interface': 6.827446204609456e-06, 'learning_rate/action_model': 6.827446204609456e-05, 'epoch': 0.15}) +[2026-10-03 13:25:21,371][starVLA.training.train_starvla][INFO] - Step 2023, Loss: {'train/loss': 0.0005348356207832694, 'train/grad_norm_pre_clip': 0.5536049008369446, 'timing/data': 0.000241660003666766, 'timing/model': 0.39549649599939585, 'learning_rate/qwen_vl_interface': 6.824573298022522e-06, 'learning_rate/action_model': 6.824573298022522e-05, 'epoch': 0.15}) +[2026-10-03 13:25:22,264][starVLA.training.train_starvla][INFO] - Step 2024, Loss: {'train/loss': 0.0006240099901333451, 'train/grad_norm_pre_clip': 0.3580821752548218, 'timing/data': 0.00028322999423835427, 'timing/model': 0.5806572139990749, 'learning_rate/qwen_vl_interface': 6.8216997441879805e-06, 'learning_rate/action_model': 6.821699744187981e-05, 'epoch': 0.15}) +[2026-10-03 13:25:22,973][starVLA.training.train_starvla][INFO] - Step 2025, Loss: {'train/loss': 0.00032490951707586646, 'train/grad_norm_pre_clip': 0.10662990063428879, 'timing/data': 0.00034270000469405204, 'timing/model': 0.3960094749927521, 'learning_rate/qwen_vl_interface': 6.818825544287045e-06, 'learning_rate/action_model': 6.818825544287045e-05, 'epoch': 0.15}) +[2026-10-03 13:25:23,680][starVLA.training.train_starvla][INFO] - Step 2026, Loss: {'train/loss': 6.479713192675263e-05, 'train/grad_norm_pre_clip': 0.011905113235116005, 'timing/data': 0.0002749400009633973, 'timing/model': 0.3939680450130254, 'learning_rate/qwen_vl_interface': 6.815950699501186e-06, 'learning_rate/action_model': 6.815950699501186e-05, 'epoch': 0.15}) +[2026-10-03 13:25:24,387][starVLA.training.train_starvla][INFO] - Step 2027, Loss: {'train/loss': 0.003271522233262658, 'train/grad_norm_pre_clip': 9.672107696533203, 'timing/data': 0.0002609999937703833, 'timing/model': 0.3941026749962475, 'learning_rate/qwen_vl_interface': 6.813075211012148e-06, 'learning_rate/action_model': 6.813075211012147e-05, 'epoch': 0.15}) +[2026-10-03 13:25:25,091][starVLA.training.train_starvla][INFO] - Step 2028, Loss: {'train/loss': 8.444827108178288e-05, 'train/grad_norm_pre_clip': 0.03009391948580742, 'timing/data': 0.00020638000569306314, 'timing/model': 0.39340887501020916, 'learning_rate/qwen_vl_interface': 6.81019908000193e-06, 'learning_rate/action_model': 6.81019908000193e-05, 'epoch': 0.15}) +[2026-10-03 13:25:25,799][starVLA.training.train_starvla][INFO] - Step 2029, Loss: {'train/loss': 0.06852534413337708, 'train/grad_norm_pre_clip': 30.534378051757812, 'timing/data': 0.00021486000332515687, 'timing/model': 0.39468870499695186, 'learning_rate/qwen_vl_interface': 6.807322307652806e-06, 'learning_rate/action_model': 6.807322307652806e-05, 'epoch': 0.15}) +[2026-10-03 13:25:26,691][starVLA.training.train_starvla][INFO] - Step 2030, Loss: {'train/loss': 0.0006563824135810137, 'train/grad_norm_pre_clip': 0.29056820273399353, 'timing/data': 0.000225629992200993, 'timing/model': 0.39450673600367736, 'learning_rate/qwen_vl_interface': 6.804444895147306e-06, 'learning_rate/action_model': 6.804444895147306e-05, 'epoch': 0.15}) +[2026-10-03 13:25:27,395][starVLA.training.train_starvla][INFO] - Step 2031, Loss: {'train/loss': 0.004025644157081842, 'train/grad_norm_pre_clip': 5.0279059410095215, 'timing/data': 0.00020650000078603625, 'timing/model': 0.3935371050029062, 'learning_rate/qwen_vl_interface': 6.8015668436682255e-06, 'learning_rate/action_model': 6.801566843668226e-05, 'epoch': 0.15}) +[2026-10-03 13:25:28,101][starVLA.training.train_starvla][INFO] - Step 2032, Loss: {'train/loss': 4.227687531965785e-05, 'train/grad_norm_pre_clip': 0.01630134880542755, 'timing/data': 0.0003173400036757812, 'timing/model': 0.3935021850047633, 'learning_rate/qwen_vl_interface': 6.7986881543986245e-06, 'learning_rate/action_model': 6.798688154398624e-05, 'epoch': 0.15}) +[2026-10-03 13:25:28,808][starVLA.training.train_starvla][INFO] - Step 2033, Loss: {'train/loss': 7.969158468768e-05, 'train/grad_norm_pre_clip': 0.017943471670150757, 'timing/data': 0.00029457999335136265, 'timing/model': 0.3941849849943537, 'learning_rate/qwen_vl_interface': 6.795808828521822e-06, 'learning_rate/action_model': 6.795808828521821e-05, 'epoch': 0.15}) +[2026-10-03 13:25:29,516][starVLA.training.train_starvla][INFO] - Step 2034, Loss: {'train/loss': 0.201351597905159, 'train/grad_norm_pre_clip': 54.301963806152344, 'timing/data': 0.0002513899962650612, 'timing/model': 0.3948967350006569, 'learning_rate/qwen_vl_interface': 6.792928867221401e-06, 'learning_rate/action_model': 6.792928867221401e-05, 'epoch': 0.15}) +[2026-10-03 13:25:30,411][starVLA.training.train_starvla][INFO] - Step 2035, Loss: {'train/loss': 0.0023925912100821733, 'train/grad_norm_pre_clip': 4.244659423828125, 'timing/data': 0.0003343599964864552, 'timing/model': 0.39383533599902876, 'learning_rate/qwen_vl_interface': 6.790048271681202e-06, 'learning_rate/action_model': 6.790048271681202e-05, 'epoch': 0.15}) +[2026-10-03 13:25:31,116][starVLA.training.train_starvla][INFO] - Step 2036, Loss: {'train/loss': 5.939544644206762e-05, 'train/grad_norm_pre_clip': 0.01682792231440544, 'timing/data': 0.0002070600021397695, 'timing/model': 0.3940577360044699, 'learning_rate/qwen_vl_interface': 6.7871670430853335e-06, 'learning_rate/action_model': 6.787167043085333e-05, 'epoch': 0.15}) +[2026-10-03 13:25:31,823][starVLA.training.train_starvla][INFO] - Step 2037, Loss: {'train/loss': 0.16997253894805908, 'train/grad_norm_pre_clip': 59.10149002075195, 'timing/data': 0.00030856000375933945, 'timing/model': 0.39409630598674994, 'learning_rate/qwen_vl_interface': 6.784285182618156e-06, 'learning_rate/action_model': 6.784285182618156e-05, 'epoch': 0.15}) +[2026-10-03 13:25:32,529][starVLA.training.train_starvla][INFO] - Step 2038, Loss: {'train/loss': 0.00450178375467658, 'train/grad_norm_pre_clip': 1.3901581764221191, 'timing/data': 0.0002534600062062964, 'timing/model': 0.3943311959883431, 'learning_rate/qwen_vl_interface': 6.781402691464297e-06, 'learning_rate/action_model': 6.781402691464297e-05, 'epoch': 0.15}) +[2026-10-03 13:25:33,233][starVLA.training.train_starvla][INFO] - Step 2039, Loss: {'train/loss': 0.009156656451523304, 'train/grad_norm_pre_clip': 29.080078125, 'timing/data': 0.00031334000232163817, 'timing/model': 0.39320147600665223, 'learning_rate/qwen_vl_interface': 6.778519570808636e-06, 'learning_rate/action_model': 6.778519570808636e-05, 'epoch': 0.15}) +[2026-10-03 13:25:34,124][starVLA.training.train_starvla][INFO] - Step 2040, Loss: {'train/loss': 0.5341706275939941, 'train/grad_norm_pre_clip': 21.775758743286133, 'timing/data': 0.00035262000164948404, 'timing/model': 0.5796148549998179, 'learning_rate/qwen_vl_interface': 6.775635821836317e-06, 'learning_rate/action_model': 6.775635821836317e-05, 'epoch': 0.15}) +[2026-10-03 13:25:34,830][starVLA.training.train_starvla][INFO] - Step 2041, Loss: {'train/loss': 5.866861101821996e-05, 'train/grad_norm_pre_clip': 0.035605911165475845, 'timing/data': 0.0002470699982950464, 'timing/model': 0.39416945600532927, 'learning_rate/qwen_vl_interface': 6.77275144573274e-06, 'learning_rate/action_model': 6.77275144573274e-05, 'epoch': 0.15}) +[2026-10-03 13:25:35,538][starVLA.training.train_starvla][INFO] - Step 2042, Loss: {'train/loss': 0.00011721826012944803, 'train/grad_norm_pre_clip': 0.045487795025110245, 'timing/data': 0.000351569993654266, 'timing/model': 0.39483724599995185, 'learning_rate/qwen_vl_interface': 6.769866443683564e-06, 'learning_rate/action_model': 6.769866443683563e-05, 'epoch': 0.15}) +[2026-10-03 13:25:36,243][starVLA.training.train_starvla][INFO] - Step 2043, Loss: {'train/loss': 0.2206307202577591, 'train/grad_norm_pre_clip': 214.3587188720703, 'timing/data': 0.00033368000003974885, 'timing/model': 0.3939389159932034, 'learning_rate/qwen_vl_interface': 6.766980816874701e-06, 'learning_rate/action_model': 6.7669808168747e-05, 'epoch': 0.15}) +[2026-10-03 13:25:36,949][starVLA.training.train_starvla][INFO] - Step 2044, Loss: {'train/loss': 0.09218727797269821, 'train/grad_norm_pre_clip': 56.3374137878418, 'timing/data': 0.00020619000133592635, 'timing/model': 0.39490214599936735, 'learning_rate/qwen_vl_interface': 6.764094566492327e-06, 'learning_rate/action_model': 6.764094566492328e-05, 'epoch': 0.15}) +[2026-10-03 13:25:37,656][starVLA.training.train_starvla][INFO] - Step 2045, Loss: {'train/loss': 0.001140295877121389, 'train/grad_norm_pre_clip': 4.6841888427734375, 'timing/data': 0.0002663999912329018, 'timing/model': 0.39413663500454277, 'learning_rate/qwen_vl_interface': 6.7612076937228685e-06, 'learning_rate/action_model': 6.761207693722869e-05, 'epoch': 0.15}) +[2026-10-03 13:25:38,547][starVLA.training.train_starvla][INFO] - Step 2046, Loss: {'train/loss': 0.00020275065617170185, 'train/grad_norm_pre_clip': 0.06703772395849228, 'timing/data': 0.00040202999662142247, 'timing/model': 0.3947024650115054, 'learning_rate/qwen_vl_interface': 6.7583201997530126e-06, 'learning_rate/action_model': 6.758320199753012e-05, 'epoch': 0.15}) +[2026-10-03 13:25:39,252][starVLA.training.train_starvla][INFO] - Step 2047, Loss: {'train/loss': 0.000271306314971298, 'train/grad_norm_pre_clip': 0.24118280410766602, 'timing/data': 0.0002509100013412535, 'timing/model': 0.3938591960031772, 'learning_rate/qwen_vl_interface': 6.755432085769693e-06, 'learning_rate/action_model': 6.755432085769694e-05, 'epoch': 0.15}) +[2026-10-03 13:25:39,958][starVLA.training.train_starvla][INFO] - Step 2048, Loss: {'train/loss': 0.4962039291858673, 'train/grad_norm_pre_clip': 101.2297134399414, 'timing/data': 0.0003405499883228913, 'timing/model': 0.39405727600387763, 'learning_rate/qwen_vl_interface': 6.75254335296011e-06, 'learning_rate/action_model': 6.75254335296011e-05, 'epoch': 0.15}) +[2026-10-03 13:25:40,665][starVLA.training.train_starvla][INFO] - Step 2049, Loss: {'train/loss': 5.1461691327858716e-05, 'train/grad_norm_pre_clip': 0.024152161553502083, 'timing/data': 0.0003448100032983348, 'timing/model': 0.3941340860037599, 'learning_rate/qwen_vl_interface': 6.749654002511712e-06, 'learning_rate/action_model': 6.749654002511712e-05, 'epoch': 0.15}) +[2026-10-03 13:25:41,372][starVLA.training.train_starvla][INFO] - Step 2050, Loss: {'train/loss': 0.0008642080938443542, 'train/grad_norm_pre_clip': 0.3882312774658203, 'timing/data': 0.0003408900083741173, 'timing/model': 0.3947266250033863, 'learning_rate/qwen_vl_interface': 6.7467640356122e-06, 'learning_rate/action_model': 6.746764035612199e-05, 'epoch': 0.15}) +[2026-10-03 13:25:42,264][starVLA.training.train_starvla][INFO] - Step 2051, Loss: {'train/loss': 0.00020310611580498517, 'train/grad_norm_pre_clip': 0.12053506076335907, 'timing/data': 0.0004424499929882586, 'timing/model': 0.3934772549982881, 'learning_rate/qwen_vl_interface': 6.74387345344953e-06, 'learning_rate/action_model': 6.74387345344953e-05, 'epoch': 0.15}) +[2026-10-03 13:25:42,970][starVLA.training.train_starvla][INFO] - Step 2052, Loss: {'train/loss': 0.0005941904382780194, 'train/grad_norm_pre_clip': 0.33529382944107056, 'timing/data': 0.00020655000116676092, 'timing/model': 0.39360382499580737, 'learning_rate/qwen_vl_interface': 6.740982257211911e-06, 'learning_rate/action_model': 6.740982257211911e-05, 'epoch': 0.15}) +[2026-10-03 13:25:43,676][starVLA.training.train_starvla][INFO] - Step 2053, Loss: {'train/loss': 0.013736426830291748, 'train/grad_norm_pre_clip': 14.832477569580078, 'timing/data': 0.00028898999153170735, 'timing/model': 0.3942764649982564, 'learning_rate/qwen_vl_interface': 6.738090448087806e-06, 'learning_rate/action_model': 6.738090448087806e-05, 'epoch': 0.15}) +[2026-10-03 13:25:44,383][starVLA.training.train_starvla][INFO] - Step 2054, Loss: {'train/loss': 3.853574889944866e-05, 'train/grad_norm_pre_clip': 0.020434701815247536, 'timing/data': 0.00021165001089684665, 'timing/model': 0.39460269499977585, 'learning_rate/qwen_vl_interface': 6.735198027265928e-06, 'learning_rate/action_model': 6.735198027265928e-05, 'epoch': 0.15}) +[2026-10-03 13:25:45,089][starVLA.training.train_starvla][INFO] - Step 2055, Loss: {'train/loss': 0.35952523350715637, 'train/grad_norm_pre_clip': 53.85994338989258, 'timing/data': 0.00020323999342508614, 'timing/model': 0.3950479049963178, 'learning_rate/qwen_vl_interface': 6.7323049959352424e-06, 'learning_rate/action_model': 6.732304995935243e-05, 'epoch': 0.15}) +[2026-10-03 13:25:45,984][starVLA.training.train_starvla][INFO] - Step 2056, Loss: {'train/loss': 7.547002041974338e-06, 'train/grad_norm_pre_clip': 0.0025774510577321053, 'timing/data': 0.00029541000549215823, 'timing/model': 0.39435535500524566, 'learning_rate/qwen_vl_interface': 6.729411355284962e-06, 'learning_rate/action_model': 6.729411355284963e-05, 'epoch': 0.15}) +[2026-10-03 13:25:46,690][starVLA.training.train_starvla][INFO] - Step 2057, Loss: {'train/loss': 0.00239034672267735, 'train/grad_norm_pre_clip': 5.218847751617432, 'timing/data': 0.0002995999966515228, 'timing/model': 0.39445775499916635, 'learning_rate/qwen_vl_interface': 6.726517106504559e-06, 'learning_rate/action_model': 6.726517106504558e-05, 'epoch': 0.15}) +[2026-10-03 13:25:47,396][starVLA.training.train_starvla][INFO] - Step 2058, Loss: {'train/loss': 0.00022486006491817534, 'train/grad_norm_pre_clip': 0.18456780910491943, 'timing/data': 0.00024557000142522156, 'timing/model': 0.3937599260098068, 'learning_rate/qwen_vl_interface': 6.723622250783745e-06, 'learning_rate/action_model': 6.723622250783744e-05, 'epoch': 0.15}) +[2026-10-03 13:25:48,100][starVLA.training.train_starvla][INFO] - Step 2059, Loss: {'train/loss': 0.040567249059677124, 'train/grad_norm_pre_clip': 37.567909240722656, 'timing/data': 0.00032144000579137355, 'timing/model': 0.3934917559963651, 'learning_rate/qwen_vl_interface': 6.720726789312487e-06, 'learning_rate/action_model': 6.720726789312487e-05, 'epoch': 0.15}) +[2026-10-03 13:25:48,806][starVLA.training.train_starvla][INFO] - Step 2060, Loss: {'train/loss': 8.672321200720035e-06, 'train/grad_norm_pre_clip': 0.0019055139273405075, 'timing/data': 0.00029121899569872767, 'timing/model': 0.39366607600823045, 'learning_rate/qwen_vl_interface': 6.7178307232810015e-06, 'learning_rate/action_model': 6.717830723281e-05, 'epoch': 0.15}) +[2026-10-03 13:25:49,700][starVLA.training.train_starvla][INFO] - Step 2061, Loss: {'train/loss': 9.888986824080348e-05, 'train/grad_norm_pre_clip': 0.11968231946229935, 'timing/data': 0.00034779000270646065, 'timing/model': 0.5816388840030413, 'learning_rate/qwen_vl_interface': 6.714934053879749e-06, 'learning_rate/action_model': 6.71493405387975e-05, 'epoch': 0.15}) +[2026-10-03 13:25:50,407][starVLA.training.train_starvla][INFO] - Step 2062, Loss: {'train/loss': 0.15721014142036438, 'train/grad_norm_pre_clip': 47.814640045166016, 'timing/data': 0.00025089000700972974, 'timing/model': 0.39473553499556147, 'learning_rate/qwen_vl_interface': 6.712036782299442e-06, 'learning_rate/action_model': 6.712036782299443e-05, 'epoch': 0.15}) +[2026-10-03 13:25:51,112][starVLA.training.train_starvla][INFO] - Step 2063, Loss: {'train/loss': 4.90302627440542e-05, 'train/grad_norm_pre_clip': 0.013890553265810013, 'timing/data': 0.0002503199939383194, 'timing/model': 0.39334960500127636, 'learning_rate/qwen_vl_interface': 6.709138909731041e-06, 'learning_rate/action_model': 6.709138909731041e-05, 'epoch': 0.15}) +[2026-10-03 13:25:51,817][starVLA.training.train_starvla][INFO] - Step 2064, Loss: {'train/loss': 0.0002649036468937993, 'train/grad_norm_pre_clip': 0.18257345259189606, 'timing/data': 0.0002630200033308938, 'timing/model': 0.39345960499485955, 'learning_rate/qwen_vl_interface': 6.706240437365751e-06, 'learning_rate/action_model': 6.70624043736575e-05, 'epoch': 0.15}) +[2026-10-03 13:25:52,523][starVLA.training.train_starvla][INFO] - Step 2065, Loss: {'train/loss': 0.0008424544939771295, 'train/grad_norm_pre_clip': 1.6588401794433594, 'timing/data': 0.00034317000245209783, 'timing/model': 0.3938857549946988, 'learning_rate/qwen_vl_interface': 6.703341366395023e-06, 'learning_rate/action_model': 6.703341366395023e-05, 'epoch': 0.15}) +[2026-10-03 13:25:53,416][starVLA.training.train_starvla][INFO] - Step 2066, Loss: {'train/loss': 3.6432945762499003e-06, 'train/grad_norm_pre_clip': 0.0009320076205767691, 'timing/data': 0.0002436399954603985, 'timing/model': 0.5808281139907194, 'learning_rate/qwen_vl_interface': 6.700441698010558e-06, 'learning_rate/action_model': 6.700441698010558e-05, 'epoch': 0.15}) +[2026-10-03 13:25:54,121][starVLA.training.train_starvla][INFO] - Step 2067, Loss: {'train/loss': 0.002503950847312808, 'train/grad_norm_pre_clip': 7.443230628967285, 'timing/data': 0.00019859999883919954, 'timing/model': 0.3941375450085616, 'learning_rate/qwen_vl_interface': 6.697541433404299e-06, 'learning_rate/action_model': 6.697541433404299e-05, 'epoch': 0.15}) +[2026-10-03 13:25:54,826][starVLA.training.train_starvla][INFO] - Step 2068, Loss: {'train/loss': 0.14499787986278534, 'train/grad_norm_pre_clip': 161.18185424804688, 'timing/data': 0.0003203899977961555, 'timing/model': 0.3935028049891116, 'learning_rate/qwen_vl_interface': 6.6946405737684325e-06, 'learning_rate/action_model': 6.694640573768433e-05, 'epoch': 0.15}) +[2026-10-03 13:25:55,534][starVLA.training.train_starvla][INFO] - Step 2069, Loss: {'train/loss': 4.1052207961911336e-06, 'train/grad_norm_pre_clip': 0.0015321302926167846, 'timing/data': 0.00038682999729644507, 'timing/model': 0.39409615499607753, 'learning_rate/qwen_vl_interface': 6.691739120295397e-06, 'learning_rate/action_model': 6.691739120295397e-05, 'epoch': 0.15}) +[2026-10-03 13:25:56,239][starVLA.training.train_starvla][INFO] - Step 2070, Loss: {'train/loss': 0.04865478724241257, 'train/grad_norm_pre_clip': 67.49102783203125, 'timing/data': 0.00040290001197718084, 'timing/model': 0.3938503649987979, 'learning_rate/qwen_vl_interface': 6.688837074177868e-06, 'learning_rate/action_model': 6.688837074177868e-05, 'epoch': 0.15}) +[2026-10-03 13:25:56,945][starVLA.training.train_starvla][INFO] - Step 2071, Loss: {'train/loss': 0.2988024055957794, 'train/grad_norm_pre_clip': 149.12399291992188, 'timing/data': 0.0002456299989717081, 'timing/model': 0.39476445500622503, 'learning_rate/qwen_vl_interface': 6.68593443660877e-06, 'learning_rate/action_model': 6.685934436608769e-05, 'epoch': 0.15}) +[2026-10-03 13:25:57,841][starVLA.training.train_starvla][INFO] - Step 2072, Loss: {'train/loss': 1.9743952179851476e-06, 'train/grad_norm_pre_clip': 0.00039828204899095, 'timing/data': 0.0003175000019837171, 'timing/model': 0.39328460600518156, 'learning_rate/qwen_vl_interface': 6.6830312087812625e-06, 'learning_rate/action_model': 6.683031208781262e-05, 'epoch': 0.15}) +[2026-10-03 13:25:58,548][starVLA.training.train_starvla][INFO] - Step 2073, Loss: {'train/loss': 0.42361125349998474, 'train/grad_norm_pre_clip': 114.79837799072266, 'timing/data': 0.0003654000029200688, 'timing/model': 0.3944440350023797, 'learning_rate/qwen_vl_interface': 6.680127391888754e-06, 'learning_rate/action_model': 6.680127391888754e-05, 'epoch': 0.15}) +[2026-10-03 13:25:59,254][starVLA.training.train_starvla][INFO] - Step 2074, Loss: {'train/loss': 6.826710159657523e-05, 'train/grad_norm_pre_clip': 0.01960308849811554, 'timing/data': 0.0002859899977920577, 'timing/model': 0.39422777500294615, 'learning_rate/qwen_vl_interface': 6.677222987124897e-06, 'learning_rate/action_model': 6.677222987124897e-05, 'epoch': 0.15}) +[2026-10-03 13:25:59,959][starVLA.training.train_starvla][INFO] - Step 2075, Loss: {'train/loss': 5.538589175557718e-05, 'train/grad_norm_pre_clip': 0.02206055074930191, 'timing/data': 0.00032295999699272215, 'timing/model': 0.39400646499416325, 'learning_rate/qwen_vl_interface': 6.67431799568358e-06, 'learning_rate/action_model': 6.67431799568358e-05, 'epoch': 0.15}) +[2026-10-03 13:26:00,664][starVLA.training.train_starvla][INFO] - Step 2076, Loss: {'train/loss': 6.057794962543994e-05, 'train/grad_norm_pre_clip': 0.022585708647966385, 'timing/data': 0.0003111099940724671, 'timing/model': 0.3936738860065816, 'learning_rate/qwen_vl_interface': 6.671412418758939e-06, 'learning_rate/action_model': 6.671412418758938e-05, 'epoch': 0.15}) +[2026-10-03 13:26:01,557][starVLA.training.train_starvla][INFO] - Step 2077, Loss: {'train/loss': 6.3774823502171785e-06, 'train/grad_norm_pre_clip': 0.009464308619499207, 'timing/data': 0.0003402299917070195, 'timing/model': 0.5799268239934463, 'learning_rate/qwen_vl_interface': 6.668506257545341e-06, 'learning_rate/action_model': 6.668506257545341e-05, 'epoch': 0.15}) +[2026-10-03 13:26:02,264][starVLA.training.train_starvla][INFO] - Step 2078, Loss: {'train/loss': 9.68574113358045e-07, 'train/grad_norm_pre_clip': 0.00016659566608723253, 'timing/data': 0.0003556099982233718, 'timing/model': 0.39464022500033025, 'learning_rate/qwen_vl_interface': 6.665599513237403e-06, 'learning_rate/action_model': 6.665599513237402e-05, 'epoch': 0.15}) +[2026-10-03 13:26:02,970][starVLA.training.train_starvla][INFO] - Step 2079, Loss: {'train/loss': 1.9167626305716112e-05, 'train/grad_norm_pre_clip': 0.020338278263807297, 'timing/data': 0.0002560400025686249, 'timing/model': 0.3938776850118302, 'learning_rate/qwen_vl_interface': 6.662692187029978e-06, 'learning_rate/action_model': 6.662692187029978e-05, 'epoch': 0.15}) +[2026-10-03 13:26:03,676][starVLA.training.train_starvla][INFO] - Step 2080, Loss: {'train/loss': 1.6710249838070013e-05, 'train/grad_norm_pre_clip': 0.017795855179429054, 'timing/data': 0.00021050000214017928, 'timing/model': 0.39441816598991863, 'learning_rate/qwen_vl_interface': 6.659784280118156e-06, 'learning_rate/action_model': 6.659784280118156e-05, 'epoch': 0.15}) +[2026-10-03 13:26:04,385][starVLA.training.train_starvla][INFO] - Step 2081, Loss: {'train/loss': 0.4341849386692047, 'train/grad_norm_pre_clip': 22.03072166442871, 'timing/data': 0.0003097300068475306, 'timing/model': 0.3944953550089849, 'learning_rate/qwen_vl_interface': 6.65687579369727e-06, 'learning_rate/action_model': 6.656875793697271e-05, 'epoch': 0.15}) +[2026-10-03 13:26:05,091][starVLA.training.train_starvla][INFO] - Step 2082, Loss: {'train/loss': 0.023593872785568237, 'train/grad_norm_pre_clip': 23.32062339782715, 'timing/data': 0.00044720999721903354, 'timing/model': 0.39389464499254245, 'learning_rate/qwen_vl_interface': 6.653966728962885e-06, 'learning_rate/action_model': 6.653966728962885e-05, 'epoch': 0.15}) +[2026-10-03 13:26:05,983][starVLA.training.train_starvla][INFO] - Step 2083, Loss: {'train/loss': 0.03760059177875519, 'train/grad_norm_pre_clip': 52.64051055908203, 'timing/data': 0.0003427699994063005, 'timing/model': 0.3948245060018962, 'learning_rate/qwen_vl_interface': 6.651057087110813e-06, 'learning_rate/action_model': 6.651057087110813e-05, 'epoch': 0.15}) +[2026-10-03 13:26:06,690][starVLA.training.train_starvla][INFO] - Step 2084, Loss: {'train/loss': 0.0002687321975827217, 'train/grad_norm_pre_clip': 0.08724769949913025, 'timing/data': 0.00021507999917957932, 'timing/model': 0.3946633049927186, 'learning_rate/qwen_vl_interface': 6.6481468693370924e-06, 'learning_rate/action_model': 6.648146869337091e-05, 'epoch': 0.15}) +[2026-10-03 13:26:07,398][starVLA.training.train_starvla][INFO] - Step 2085, Loss: {'train/loss': 0.0009892850648611784, 'train/grad_norm_pre_clip': 2.558519124984741, 'timing/data': 0.0002987500047311187, 'timing/model': 0.3943027649947908, 'learning_rate/qwen_vl_interface': 6.645236076838009e-06, 'learning_rate/action_model': 6.645236076838009e-05, 'epoch': 0.15}) +[2026-10-03 13:26:08,105][starVLA.training.train_starvla][INFO] - Step 2086, Loss: {'train/loss': 0.6591039896011353, 'train/grad_norm_pre_clip': 35.22016143798828, 'timing/data': 0.0002558190026320517, 'timing/model': 0.3948374059982598, 'learning_rate/qwen_vl_interface': 6.642324710810075e-06, 'learning_rate/action_model': 6.642324710810076e-05, 'epoch': 0.15}) +[2026-10-03 13:26:08,811][starVLA.training.train_starvla][INFO] - Step 2087, Loss: {'train/loss': 1.475210410717409e-06, 'train/grad_norm_pre_clip': 0.0006303998525254428, 'timing/data': 0.00033214999712072313, 'timing/model': 0.39405904499290045, 'learning_rate/qwen_vl_interface': 6.639412772450047e-06, 'learning_rate/action_model': 6.639412772450047e-05, 'epoch': 0.15}) +[2026-10-03 13:26:09,704][starVLA.training.train_starvla][INFO] - Step 2088, Loss: {'train/loss': 7.525076171077671e-07, 'train/grad_norm_pre_clip': 0.0004110542358830571, 'timing/data': 0.00031823001336306334, 'timing/model': 0.3944142649997957, 'learning_rate/qwen_vl_interface': 6.6365002629549104e-06, 'learning_rate/action_model': 6.63650026295491e-05, 'epoch': 0.15}) +[2026-10-03 13:26:10,411][starVLA.training.train_starvla][INFO] - Step 2089, Loss: {'train/loss': 0.39435023069381714, 'train/grad_norm_pre_clip': 130.46652221679688, 'timing/data': 0.00034498999593779445, 'timing/model': 0.3943376649986021, 'learning_rate/qwen_vl_interface': 6.633587183521887e-06, 'learning_rate/action_model': 6.633587183521887e-05, 'epoch': 0.15}) +[2026-10-03 13:26:11,119][starVLA.training.train_starvla][INFO] - Step 2090, Loss: {'train/loss': 0.0002631284005474299, 'train/grad_norm_pre_clip': 0.07787461578845978, 'timing/data': 0.00033814999915193766, 'timing/model': 0.39493472600588575, 'learning_rate/qwen_vl_interface': 6.6306735353484365e-06, 'learning_rate/action_model': 6.630673535348436e-05, 'epoch': 0.15}) +[2026-10-03 13:26:11,825][starVLA.training.train_starvla][INFO] - Step 2091, Loss: {'train/loss': 3.384875526535325e-05, 'train/grad_norm_pre_clip': 0.03609035909175873, 'timing/data': 0.0002738699986366555, 'timing/model': 0.3946586259989999, 'learning_rate/qwen_vl_interface': 6.627759319632246e-06, 'learning_rate/action_model': 6.627759319632247e-05, 'epoch': 0.15}) +[2026-10-03 13:26:12,531][starVLA.training.train_starvla][INFO] - Step 2092, Loss: {'train/loss': 7.896000170148909e-05, 'train/grad_norm_pre_clip': 0.04810318350791931, 'timing/data': 0.0002043000131379813, 'timing/model': 0.39430523599730805, 'learning_rate/qwen_vl_interface': 6.624844537571245e-06, 'learning_rate/action_model': 6.624844537571245e-05, 'epoch': 0.15}) +[2026-10-03 13:26:13,427][starVLA.training.train_starvla][INFO] - Step 2093, Loss: {'train/loss': 0.012915803119540215, 'train/grad_norm_pre_clip': 15.366039276123047, 'timing/data': 0.00031600000511389226, 'timing/model': 0.5826176139962627, 'learning_rate/qwen_vl_interface': 6.621929190363585e-06, 'learning_rate/action_model': 6.621929190363585e-05, 'epoch': 0.15}) +[2026-10-03 13:26:14,134][starVLA.training.train_starvla][INFO] - Step 2094, Loss: {'train/loss': 0.0002826181298587471, 'train/grad_norm_pre_clip': 0.06508854031562805, 'timing/data': 0.00029213000379968435, 'timing/model': 0.3949631550058257, 'learning_rate/qwen_vl_interface': 6.619013279207657e-06, 'learning_rate/action_model': 6.619013279207656e-05, 'epoch': 0.15}) +[2026-10-03 13:26:14,841][starVLA.training.train_starvla][INFO] - Step 2095, Loss: {'train/loss': 0.28861716389656067, 'train/grad_norm_pre_clip': 12.084933280944824, 'timing/data': 0.00032053000177256763, 'timing/model': 0.39475186499475967, 'learning_rate/qwen_vl_interface': 6.616096805302083e-06, 'learning_rate/action_model': 6.616096805302082e-05, 'epoch': 0.15}) +[2026-10-03 13:26:15,547][starVLA.training.train_starvla][INFO] - Step 2096, Loss: {'train/loss': 0.0007446073577739298, 'train/grad_norm_pre_clip': 0.15304766595363617, 'timing/data': 0.00020155998936388642, 'timing/model': 0.39472812600433826, 'learning_rate/qwen_vl_interface': 6.6131797698457135e-06, 'learning_rate/action_model': 6.613179769845713e-05, 'epoch': 0.15}) +[2026-10-03 13:26:16,255][starVLA.training.train_starvla][INFO] - Step 2097, Loss: {'train/loss': 0.41439634561538696, 'train/grad_norm_pre_clip': 74.49394989013672, 'timing/data': 0.0003305200079921633, 'timing/model': 0.39469718499458395, 'learning_rate/qwen_vl_interface': 6.610262174037633e-06, 'learning_rate/action_model': 6.610262174037633e-05, 'epoch': 0.15}) +[2026-10-03 13:26:16,964][starVLA.training.train_starvla][INFO] - Step 2098, Loss: {'train/loss': 9.142734052147716e-05, 'train/grad_norm_pre_clip': 0.038457438349723816, 'timing/data': 0.0002764899982139468, 'timing/model': 0.3944504660030361, 'learning_rate/qwen_vl_interface': 6.607344019077155e-06, 'learning_rate/action_model': 6.607344019077156e-05, 'epoch': 0.15}) +[2026-10-03 13:26:17,856][starVLA.training.train_starvla][INFO] - Step 2099, Loss: {'train/loss': 0.00011972880020039156, 'train/grad_norm_pre_clip': 0.0626194030046463, 'timing/data': 0.00022973999148234725, 'timing/model': 0.39448024600278586, 'learning_rate/qwen_vl_interface': 6.604425306163822e-06, 'learning_rate/action_model': 6.604425306163822e-05, 'epoch': 0.15}) +[2026-10-03 13:26:18,562][starVLA.training.train_starvla][INFO] - Step 2100, Loss: {'train/loss': 0.0035894857719540596, 'train/grad_norm_pre_clip': 3.626579523086548, 'timing/data': 0.0002923199936049059, 'timing/model': 0.39480944599199574, 'learning_rate/qwen_vl_interface': 6.6015060364974045e-06, 'learning_rate/action_model': 6.601506036497404e-05, 'epoch': 0.15}) +[2026-10-03 13:26:19,270][starVLA.training.train_starvla][INFO] - Step 2101, Loss: {'train/loss': 3.126687806798145e-05, 'train/grad_norm_pre_clip': 0.013671111315488815, 'timing/data': 0.00034918999881483614, 'timing/model': 0.39453353600401897, 'learning_rate/qwen_vl_interface': 6.598586211277907e-06, 'learning_rate/action_model': 6.598586211277907e-05, 'epoch': 0.15}) +[2026-10-03 13:26:19,977][starVLA.training.train_starvla][INFO] - Step 2102, Loss: {'train/loss': 0.5186801552772522, 'train/grad_norm_pre_clip': 35.433414459228516, 'timing/data': 0.0003221700026188046, 'timing/model': 0.3946535849972861, 'learning_rate/qwen_vl_interface': 6.59566583170556e-06, 'learning_rate/action_model': 6.59566583170556e-05, 'epoch': 0.15}) +[2026-10-03 13:26:20,683][starVLA.training.train_starvla][INFO] - Step 2103, Loss: {'train/loss': 0.003939823247492313, 'train/grad_norm_pre_clip': 0.7360131740570068, 'timing/data': 0.00032187000033445656, 'timing/model': 0.3946209850109881, 'learning_rate/qwen_vl_interface': 6.59274489898082e-06, 'learning_rate/action_model': 6.592744898980819e-05, 'epoch': 0.15}) +[2026-10-03 13:26:21,577][starVLA.training.train_starvla][INFO] - Step 2104, Loss: {'train/loss': 5.4538882977794856e-05, 'train/grad_norm_pre_clip': 0.02210891619324684, 'timing/data': 0.00027516999398358166, 'timing/model': 0.39479308600130025, 'learning_rate/qwen_vl_interface': 6.5898234143043716e-06, 'learning_rate/action_model': 6.589823414304371e-05, 'epoch': 0.15}) +[2026-10-03 13:26:22,285][starVLA.training.train_starvla][INFO] - Step 2105, Loss: {'train/loss': 3.4830914955819026e-05, 'train/grad_norm_pre_clip': 0.01804734580218792, 'timing/data': 0.00030559999868273735, 'timing/model': 0.3944950649893144, 'learning_rate/qwen_vl_interface': 6.5869013788771264e-06, 'learning_rate/action_model': 6.586901378877126e-05, 'epoch': 0.15}) +[2026-10-03 13:26:22,993][starVLA.training.train_starvla][INFO] - Step 2106, Loss: {'train/loss': 0.0014578173868358135, 'train/grad_norm_pre_clip': 0.2743335962295532, 'timing/data': 0.0003259599907323718, 'timing/model': 0.3949051659874385, 'learning_rate/qwen_vl_interface': 6.583978793900223e-06, 'learning_rate/action_model': 6.583978793900223e-05, 'epoch': 0.15}) +[2026-10-03 13:26:23,700][starVLA.training.train_starvla][INFO] - Step 2107, Loss: {'train/loss': 3.3527078358019935e-06, 'train/grad_norm_pre_clip': 0.0006226251134648919, 'timing/data': 0.00021065000328235328, 'timing/model': 0.39433947500947397, 'learning_rate/qwen_vl_interface': 6.581055660575028e-06, 'learning_rate/action_model': 6.581055660575027e-05, 'epoch': 0.15}) +[2026-10-03 13:26:24,406][starVLA.training.train_starvla][INFO] - Step 2108, Loss: {'train/loss': 0.3596899211406708, 'train/grad_norm_pre_clip': 107.86911010742188, 'timing/data': 0.00032321900653187186, 'timing/model': 0.3941568959999131, 'learning_rate/qwen_vl_interface': 6.578131980103128e-06, 'learning_rate/action_model': 6.578131980103128e-05, 'epoch': 0.15}) +[2026-10-03 13:26:25,298][starVLA.training.train_starvla][INFO] - Step 2109, Loss: {'train/loss': 0.007163866423070431, 'train/grad_norm_pre_clip': 10.876372337341309, 'timing/data': 0.0003242899983888492, 'timing/model': 0.5792774440051289, 'learning_rate/qwen_vl_interface': 6.5752077536863404e-06, 'learning_rate/action_model': 6.57520775368634e-05, 'epoch': 0.15}) +[2026-10-03 13:26:26,006][starVLA.training.train_starvla][INFO] - Step 2110, Loss: {'train/loss': 0.000963847094681114, 'train/grad_norm_pre_clip': 0.1685263216495514, 'timing/data': 0.0003791999915847555, 'timing/model': 0.3947755950066494, 'learning_rate/qwen_vl_interface': 6.572282982526702e-06, 'learning_rate/action_model': 6.572282982526702e-05, 'epoch': 0.15}) +[2026-10-03 13:26:26,711][starVLA.training.train_starvla][INFO] - Step 2111, Loss: {'train/loss': 0.19798313081264496, 'train/grad_norm_pre_clip': 125.40933227539062, 'timing/data': 0.00028105999808758497, 'timing/model': 0.3937066060025245, 'learning_rate/qwen_vl_interface': 6.5693576678264755e-06, 'learning_rate/action_model': 6.569357667826476e-05, 'epoch': 0.15}) +[2026-10-03 13:26:27,417][starVLA.training.train_starvla][INFO] - Step 2112, Loss: {'train/loss': 0.0006621037609875202, 'train/grad_norm_pre_clip': 0.12926365435123444, 'timing/data': 0.00020527999731712043, 'timing/model': 0.39401195499522146, 'learning_rate/qwen_vl_interface': 6.566431810788148e-06, 'learning_rate/action_model': 6.566431810788148e-05, 'epoch': 0.15}) +[2026-10-03 13:26:28,125][starVLA.training.train_starvla][INFO] - Step 2113, Loss: {'train/loss': 8.620007065474056e-06, 'train/grad_norm_pre_clip': 0.002242211252450943, 'timing/data': 0.00025877000007312745, 'timing/model': 0.39512832499167416, 'learning_rate/qwen_vl_interface': 6.5635054126144306e-06, 'learning_rate/action_model': 6.56350541261443e-05, 'epoch': 0.15}) +[2026-10-03 13:26:29,019][starVLA.training.train_starvla][INFO] - Step 2114, Loss: {'train/loss': 0.4285929799079895, 'train/grad_norm_pre_clip': 75.97382354736328, 'timing/data': 0.00022035000438336283, 'timing/model': 0.5807028140116017, 'learning_rate/qwen_vl_interface': 6.560578474508254e-06, 'learning_rate/action_model': 6.560578474508253e-05, 'epoch': 0.15}) +[2026-10-03 13:26:29,726][starVLA.training.train_starvla][INFO] - Step 2115, Loss: {'train/loss': 0.011813201941549778, 'train/grad_norm_pre_clip': 13.805290222167969, 'timing/data': 0.0002450500032864511, 'timing/model': 0.39490497500810307, 'learning_rate/qwen_vl_interface': 6.5576509976727685e-06, 'learning_rate/action_model': 6.557650997672769e-05, 'epoch': 0.15}) +[2026-10-03 13:26:30,431][starVLA.training.train_starvla][INFO] - Step 2116, Loss: {'train/loss': 0.11705470830202103, 'train/grad_norm_pre_clip': 54.3087158203125, 'timing/data': 0.000338829995598644, 'timing/model': 0.3942128960043192, 'learning_rate/qwen_vl_interface': 6.554722983311354e-06, 'learning_rate/action_model': 6.554722983311355e-05, 'epoch': 0.15}) +[2026-10-03 13:26:31,137][starVLA.training.train_starvla][INFO] - Step 2117, Loss: {'train/loss': 0.00011238112347200513, 'train/grad_norm_pre_clip': 0.0956820622086525, 'timing/data': 0.00027019999106414616, 'timing/model': 0.3927301249932498, 'learning_rate/qwen_vl_interface': 6.551794432627605e-06, 'learning_rate/action_model': 6.551794432627605e-05, 'epoch': 0.15}) +[2026-10-03 13:26:31,843][starVLA.training.train_starvla][INFO] - Step 2118, Loss: {'train/loss': 0.00015732485917396843, 'train/grad_norm_pre_clip': 0.13487140834331512, 'timing/data': 0.0003362600109539926, 'timing/model': 0.39334354500169866, 'learning_rate/qwen_vl_interface': 6.548865346825338e-06, 'learning_rate/action_model': 6.548865346825338e-05, 'epoch': 0.16}) +[2026-10-03 13:26:32,733][starVLA.training.train_starvla][INFO] - Step 2119, Loss: {'train/loss': 0.0026560663245618343, 'train/grad_norm_pre_clip': 3.590411901473999, 'timing/data': 0.0002736999886110425, 'timing/model': 0.5796292040031403, 'learning_rate/qwen_vl_interface': 6.545935727108589e-06, 'learning_rate/action_model': 6.545935727108588e-05, 'epoch': 0.16}) +[2026-10-03 13:26:33,435][starVLA.training.train_starvla][INFO] - Step 2120, Loss: {'train/loss': 0.5408194661140442, 'train/grad_norm_pre_clip': 46.44168472290039, 'timing/data': 0.0003081600007135421, 'timing/model': 0.39241002600465436, 'learning_rate/qwen_vl_interface': 6.543005574681613e-06, 'learning_rate/action_model': 6.543005574681613e-05, 'epoch': 0.16}) +[2026-10-03 13:26:34,140][starVLA.training.train_starvla][INFO] - Step 2121, Loss: {'train/loss': 3.138002648483962e-05, 'train/grad_norm_pre_clip': 0.008186942897737026, 'timing/data': 0.000257949999650009, 'timing/model': 0.39269311500538606, 'learning_rate/qwen_vl_interface': 6.5400748907488886e-06, 'learning_rate/action_model': 6.540074890748888e-05, 'epoch': 0.16}) +[2026-10-03 13:26:34,845][starVLA.training.train_starvla][INFO] - Step 2122, Loss: {'train/loss': 1.1317260941723362e-05, 'train/grad_norm_pre_clip': 0.00244026817381382, 'timing/data': 0.00021035000099800527, 'timing/model': 0.3935701959999278, 'learning_rate/qwen_vl_interface': 6.537143676515105e-06, 'learning_rate/action_model': 6.537143676515105e-05, 'epoch': 0.16}) +[2026-10-03 13:26:35,547][starVLA.training.train_starvla][INFO] - Step 2123, Loss: {'train/loss': 0.0072953770868480206, 'train/grad_norm_pre_clip': 8.721991539001465, 'timing/data': 0.0002055200020549819, 'timing/model': 0.3925251069886144, 'learning_rate/qwen_vl_interface': 6.534211933185176e-06, 'learning_rate/action_model': 6.534211933185176e-05, 'epoch': 0.16}) +[2026-10-03 13:26:36,250][starVLA.training.train_starvla][INFO] - Step 2124, Loss: {'train/loss': 8.35192531667417e-06, 'train/grad_norm_pre_clip': 0.0022133777383714914, 'timing/data': 0.0002439700037939474, 'timing/model': 0.39290324501052964, 'learning_rate/qwen_vl_interface': 6.531279661964231e-06, 'learning_rate/action_model': 6.531279661964231e-05, 'epoch': 0.16}) +[2026-10-03 13:26:37,147][starVLA.training.train_starvla][INFO] - Step 2125, Loss: {'train/loss': 5.811443770653568e-07, 'train/grad_norm_pre_clip': 0.00014121409913059324, 'timing/data': 0.0002880100073525682, 'timing/model': 0.39319554599933326, 'learning_rate/qwen_vl_interface': 6.528346864057616e-06, 'learning_rate/action_model': 6.528346864057615e-05, 'epoch': 0.16}) +[2026-10-03 13:26:37,851][starVLA.training.train_starvla][INFO] - Step 2126, Loss: {'train/loss': 9.462210073252209e-07, 'train/grad_norm_pre_clip': 0.00033216355950571597, 'timing/data': 0.0002692599955480546, 'timing/model': 0.39290182500553783, 'learning_rate/qwen_vl_interface': 6.5254135406708906e-06, 'learning_rate/action_model': 6.52541354067089e-05, 'epoch': 0.16}) +[2026-10-03 13:26:38,553][starVLA.training.train_starvla][INFO] - Step 2127, Loss: {'train/loss': 0.03215275704860687, 'train/grad_norm_pre_clip': 60.35499954223633, 'timing/data': 0.0002520400012144819, 'timing/model': 0.3923527769948123, 'learning_rate/qwen_vl_interface': 6.522479693009836e-06, 'learning_rate/action_model': 6.522479693009837e-05, 'epoch': 0.16}) +[2026-10-03 13:26:39,256][starVLA.training.train_starvla][INFO] - Step 2128, Loss: {'train/loss': 0.00220962380990386, 'train/grad_norm_pre_clip': 2.2011401653289795, 'timing/data': 0.0002010300086112693, 'timing/model': 0.39261218599858694, 'learning_rate/qwen_vl_interface': 6.519545322280446e-06, 'learning_rate/action_model': 6.519545322280446e-05, 'epoch': 0.16}) +[2026-10-03 13:26:39,960][starVLA.training.train_starvla][INFO] - Step 2129, Loss: {'train/loss': 0.48309415578842163, 'train/grad_norm_pre_clip': 240.64907836914062, 'timing/data': 0.0002602199965622276, 'timing/model': 0.3934367749898229, 'learning_rate/qwen_vl_interface': 6.5166104296889275e-06, 'learning_rate/action_model': 6.516610429688927e-05, 'epoch': 0.16}) +[2026-10-03 13:26:40,852][starVLA.training.train_starvla][INFO] - Step 2130, Loss: {'train/loss': 0.0006913512479513884, 'train/grad_norm_pre_clip': 0.137502983212471, 'timing/data': 0.00031378999119624496, 'timing/model': 0.5805225739895832, 'learning_rate/qwen_vl_interface': 6.513675016441707e-06, 'learning_rate/action_model': 6.513675016441707e-05, 'epoch': 0.16}) +[2026-10-03 13:26:41,555][starVLA.training.train_starvla][INFO] - Step 2131, Loss: {'train/loss': 4.393782728584483e-05, 'train/grad_norm_pre_clip': 0.15266627073287964, 'timing/data': 0.0002865999995265156, 'timing/model': 0.39249758499499876, 'learning_rate/qwen_vl_interface': 6.510739083745422e-06, 'learning_rate/action_model': 6.510739083745422e-05, 'epoch': 0.16}) +[2026-10-03 13:26:42,258][starVLA.training.train_starvla][INFO] - Step 2132, Loss: {'train/loss': 0.0006451787776313722, 'train/grad_norm_pre_clip': 0.12610653042793274, 'timing/data': 0.0003198000049451366, 'timing/model': 0.39176425500772893, 'learning_rate/qwen_vl_interface': 6.507802632806919e-06, 'learning_rate/action_model': 6.507802632806919e-05, 'epoch': 0.16}) +[2026-10-03 13:26:42,963][starVLA.training.train_starvla][INFO] - Step 2133, Loss: {'train/loss': 0.4537467658519745, 'train/grad_norm_pre_clip': 28.878482818603516, 'timing/data': 0.0003261300007579848, 'timing/model': 0.39327987500291783, 'learning_rate/qwen_vl_interface': 6.5048656648332646e-06, 'learning_rate/action_model': 6.504865664833264e-05, 'epoch': 0.16}) +[2026-10-03 13:26:43,667][starVLA.training.train_starvla][INFO] - Step 2134, Loss: {'train/loss': 0.012799003161489964, 'train/grad_norm_pre_clip': 15.328780174255371, 'timing/data': 0.00024391000624746084, 'timing/model': 0.3931053050037008, 'learning_rate/qwen_vl_interface': 6.501928181031737e-06, 'learning_rate/action_model': 6.501928181031737e-05, 'epoch': 0.16}) +[2026-10-03 13:26:44,369][starVLA.training.train_starvla][INFO] - Step 2135, Loss: {'train/loss': 0.06765288859605789, 'train/grad_norm_pre_clip': 232.78646850585938, 'timing/data': 0.0003062499890802428, 'timing/model': 0.3926091460016323, 'learning_rate/qwen_vl_interface': 6.498990182609823e-06, 'learning_rate/action_model': 6.498990182609823e-05, 'epoch': 0.16}) +[2026-10-03 13:26:45,258][starVLA.training.train_starvla][INFO] - Step 2136, Loss: {'train/loss': 0.008893696591258049, 'train/grad_norm_pre_clip': 6.650341033935547, 'timing/data': 0.0002430500026093796, 'timing/model': 0.3924575860000914, 'learning_rate/qwen_vl_interface': 6.4960516707752255e-06, 'learning_rate/action_model': 6.496051670775226e-05, 'epoch': 0.16}) +[2026-10-03 13:26:45,962][starVLA.training.train_starvla][INFO] - Step 2137, Loss: {'train/loss': 7.092619398463285e-06, 'train/grad_norm_pre_clip': 0.004288572818040848, 'timing/data': 0.0003540699981385842, 'timing/model': 0.3933165950002149, 'learning_rate/qwen_vl_interface': 6.493112646735851e-06, 'learning_rate/action_model': 6.493112646735851e-05, 'epoch': 0.16}) +[2026-10-03 13:26:46,668][starVLA.training.train_starvla][INFO] - Step 2138, Loss: {'train/loss': 1.631670784263406e-06, 'train/grad_norm_pre_clip': 0.0003525668871589005, 'timing/data': 0.00022015000286046416, 'timing/model': 0.393569206004031, 'learning_rate/qwen_vl_interface': 6.490173111699824e-06, 'learning_rate/action_model': 6.490173111699825e-05, 'epoch': 0.16}) +[2026-10-03 13:26:47,370][starVLA.training.train_starvla][INFO] - Step 2139, Loss: {'train/loss': 0.6054060459136963, 'train/grad_norm_pre_clip': 28.56821632385254, 'timing/data': 0.00028467000811360776, 'timing/model': 0.39241675499943085, 'learning_rate/qwen_vl_interface': 6.487233066875479e-06, 'learning_rate/action_model': 6.487233066875479e-05, 'epoch': 0.16}) +[2026-10-03 13:26:48,073][starVLA.training.train_starvla][INFO] - Step 2140, Loss: {'train/loss': 0.00033012323547154665, 'train/grad_norm_pre_clip': 0.08524913340806961, 'timing/data': 0.0002826999989338219, 'timing/model': 0.3921118450089125, 'learning_rate/qwen_vl_interface': 6.484292513471352e-06, 'learning_rate/action_model': 6.484292513471352e-05, 'epoch': 0.16}) +[2026-10-03 13:26:48,967][starVLA.training.train_starvla][INFO] - Step 2141, Loss: {'train/loss': 1.62713295139838e-05, 'train/grad_norm_pre_clip': 0.006706279702484608, 'timing/data': 0.0002164189936593175, 'timing/model': 0.39285935599764343, 'learning_rate/qwen_vl_interface': 6.4813514526961975e-06, 'learning_rate/action_model': 6.481351452696198e-05, 'epoch': 0.16}) +[2026-10-03 13:26:49,671][starVLA.training.train_starvla][INFO] - Step 2142, Loss: {'train/loss': 0.6026186347007751, 'train/grad_norm_pre_clip': 45.03169250488281, 'timing/data': 0.0002406900021014735, 'timing/model': 0.3929366759984987, 'learning_rate/qwen_vl_interface': 6.478409885758972e-06, 'learning_rate/action_model': 6.478409885758971e-05, 'epoch': 0.16}) +[2026-10-03 13:26:50,374][starVLA.training.train_starvla][INFO] - Step 2143, Loss: {'train/loss': 8.962798347056378e-06, 'train/grad_norm_pre_clip': 0.0024600608740001917, 'timing/data': 0.00034281000262126327, 'timing/model': 0.3925467750086682, 'learning_rate/qwen_vl_interface': 6.475467813868844e-06, 'learning_rate/action_model': 6.475467813868844e-05, 'epoch': 0.16}) +[2026-10-03 13:26:51,076][starVLA.training.train_starvla][INFO] - Step 2144, Loss: {'train/loss': 0.00022641551913693547, 'train/grad_norm_pre_clip': 0.061375692486763, 'timing/data': 0.00032619999547023326, 'timing/model': 0.3920602350117406, 'learning_rate/qwen_vl_interface': 6.472525238235187e-06, 'learning_rate/action_model': 6.472525238235186e-05, 'epoch': 0.16}) +[2026-10-03 13:26:51,781][starVLA.training.train_starvla][INFO] - Step 2145, Loss: {'train/loss': 0.36576923727989197, 'train/grad_norm_pre_clip': 56.70651626586914, 'timing/data': 0.0002915099903475493, 'timing/model': 0.39351404599437956, 'learning_rate/qwen_vl_interface': 6.4695821600675834e-06, 'learning_rate/action_model': 6.469582160067583e-05, 'epoch': 0.16}) +[2026-10-03 13:26:52,677][starVLA.training.train_starvla][INFO] - Step 2146, Loss: {'train/loss': 5.0588432713993825e-06, 'train/grad_norm_pre_clip': 0.0013118407223373652, 'timing/data': 0.0003261800011387095, 'timing/model': 0.5849013740080409, 'learning_rate/qwen_vl_interface': 6.466638580575824e-06, 'learning_rate/action_model': 6.466638580575824e-05, 'epoch': 0.16}) +[2026-10-03 13:26:53,380][starVLA.training.train_starvla][INFO] - Step 2147, Loss: {'train/loss': 0.00012394515215419233, 'train/grad_norm_pre_clip': 0.03669247776269913, 'timing/data': 0.00028777000261470675, 'timing/model': 0.392547454990563, 'learning_rate/qwen_vl_interface': 6.463694500969898e-06, 'learning_rate/action_model': 6.463694500969898e-05, 'epoch': 0.16}) +[2026-10-03 13:26:54,082][starVLA.training.train_starvla][INFO] - Step 2148, Loss: {'train/loss': 1.624122160137631e-05, 'train/grad_norm_pre_clip': 0.002968632150441408, 'timing/data': 0.00023774999135639518, 'timing/model': 0.39266105499700643, 'learning_rate/qwen_vl_interface': 6.460749922460008e-06, 'learning_rate/action_model': 6.460749922460008e-05, 'epoch': 0.16}) +[2026-10-03 13:26:54,786][starVLA.training.train_starvla][INFO] - Step 2149, Loss: {'train/loss': 0.00027084106113761663, 'train/grad_norm_pre_clip': 0.06534861028194427, 'timing/data': 0.00026050000451505184, 'timing/model': 0.3929528959997697, 'learning_rate/qwen_vl_interface': 6.457804846256559e-06, 'learning_rate/action_model': 6.457804846256559e-05, 'epoch': 0.16}) +[2026-10-03 13:26:55,491][starVLA.training.train_starvla][INFO] - Step 2150, Loss: {'train/loss': 9.114505519391969e-05, 'train/grad_norm_pre_clip': 0.17058375477790833, 'timing/data': 0.00025625999842304736, 'timing/model': 0.393401774999802, 'learning_rate/qwen_vl_interface': 6.45485927357016e-06, 'learning_rate/action_model': 6.454859273570159e-05, 'epoch': 0.16}) +[2026-10-03 13:26:56,195][starVLA.training.train_starvla][INFO] - Step 2151, Loss: {'train/loss': 0.0020883558318018913, 'train/grad_norm_pre_clip': 3.433926820755005, 'timing/data': 0.00024639000184834003, 'timing/model': 0.39371100600692444, 'learning_rate/qwen_vl_interface': 6.451913205611626e-06, 'learning_rate/action_model': 6.451913205611626e-05, 'epoch': 0.16}) +[2026-10-03 13:26:57,093][starVLA.training.train_starvla][INFO] - Step 2152, Loss: {'train/loss': 0.9097406268119812, 'train/grad_norm_pre_clip': 198.29222106933594, 'timing/data': 0.00029752000409644097, 'timing/model': 0.39433044599718414, 'learning_rate/qwen_vl_interface': 6.448966643591974e-06, 'learning_rate/action_model': 6.448966643591974e-05, 'epoch': 0.16}) +[2026-10-03 13:26:57,797][starVLA.training.train_starvla][INFO] - Step 2153, Loss: {'train/loss': 0.0020956830121576786, 'train/grad_norm_pre_clip': 1.0034120082855225, 'timing/data': 0.00020804999803658575, 'timing/model': 0.3935878459888045, 'learning_rate/qwen_vl_interface': 6.446019588722423e-06, 'learning_rate/action_model': 6.446019588722422e-05, 'epoch': 0.16}) +[2026-10-03 13:26:58,504][starVLA.training.train_starvla][INFO] - Step 2154, Loss: {'train/loss': 0.0011634170077741146, 'train/grad_norm_pre_clip': 0.6817736029624939, 'timing/data': 0.0002537200052756816, 'timing/model': 0.3951045249996241, 'learning_rate/qwen_vl_interface': 6.443072042214398e-06, 'learning_rate/action_model': 6.443072042214397e-05, 'epoch': 0.16}) +[2026-10-03 13:26:59,206][starVLA.training.train_starvla][INFO] - Step 2155, Loss: {'train/loss': 0.7645074129104614, 'train/grad_norm_pre_clip': 98.77681732177734, 'timing/data': 0.00020179999410174787, 'timing/model': 0.3926040669903159, 'learning_rate/qwen_vl_interface': 6.440124005279521e-06, 'learning_rate/action_model': 6.440124005279521e-05, 'epoch': 0.16}) +[2026-10-03 13:26:59,910][starVLA.training.train_starvla][INFO] - Step 2156, Loss: {'train/loss': 0.1410810500383377, 'train/grad_norm_pre_clip': 94.36449432373047, 'timing/data': 0.00029772000561933964, 'timing/model': 0.39252232600119896, 'learning_rate/qwen_vl_interface': 6.43717547912962e-06, 'learning_rate/action_model': 6.437175479129621e-05, 'epoch': 0.16}) +[2026-10-03 13:27:00,806][starVLA.training.train_starvla][INFO] - Step 2157, Loss: {'train/loss': 0.009137281216681004, 'train/grad_norm_pre_clip': 6.462032794952393, 'timing/data': 0.0002653300034580752, 'timing/model': 0.3932859750057105, 'learning_rate/qwen_vl_interface': 6.434226464976725e-06, 'learning_rate/action_model': 6.434226464976724e-05, 'epoch': 0.16}) +[2026-10-03 13:27:01,512][starVLA.training.train_starvla][INFO] - Step 2158, Loss: {'train/loss': 0.00011051108594983816, 'train/grad_norm_pre_clip': 0.037320174276828766, 'timing/data': 0.00021006999304518104, 'timing/model': 0.39325251600530464, 'learning_rate/qwen_vl_interface': 6.431276964033062e-06, 'learning_rate/action_model': 6.431276964033062e-05, 'epoch': 0.16}) +[2026-10-03 13:27:02,214][starVLA.training.train_starvla][INFO] - Step 2159, Loss: {'train/loss': 9.913671965477988e-05, 'train/grad_norm_pre_clip': 0.025549139827489853, 'timing/data': 0.00021353000192902982, 'timing/model': 0.39194440499704797, 'learning_rate/qwen_vl_interface': 6.428326977511059e-06, 'learning_rate/action_model': 6.428326977511058e-05, 'epoch': 0.16}) +[2026-10-03 13:27:02,917][starVLA.training.train_starvla][INFO] - Step 2160, Loss: {'train/loss': 0.0006056425627321005, 'train/grad_norm_pre_clip': 0.18960842490196228, 'timing/data': 0.000348040004610084, 'timing/model': 0.39243440599238966, 'learning_rate/qwen_vl_interface': 6.425376506623344e-06, 'learning_rate/action_model': 6.425376506623344e-05, 'epoch': 0.16}) +[2026-10-03 13:27:03,622][starVLA.training.train_starvla][INFO] - Step 2161, Loss: {'train/loss': 0.0001294909161515534, 'train/grad_norm_pre_clip': 0.1356048285961151, 'timing/data': 0.000294529992970638, 'timing/model': 0.3937522650085157, 'learning_rate/qwen_vl_interface': 6.422425552582747e-06, 'learning_rate/action_model': 6.422425552582747e-05, 'epoch': 0.16}) +[2026-10-03 13:27:04,519][starVLA.training.train_starvla][INFO] - Step 2162, Loss: {'train/loss': 0.3202752470970154, 'train/grad_norm_pre_clip': 167.03578186035156, 'timing/data': 0.00031091000710148364, 'timing/model': 0.3930562150053447, 'learning_rate/qwen_vl_interface': 6.41947411660229e-06, 'learning_rate/action_model': 6.41947411660229e-05, 'epoch': 0.16}) +[2026-10-03 13:27:05,222][starVLA.training.train_starvla][INFO] - Step 2163, Loss: {'train/loss': 0.00018889583589043468, 'train/grad_norm_pre_clip': 0.04341355338692665, 'timing/data': 0.00020203999883960932, 'timing/model': 0.39277160599885974, 'learning_rate/qwen_vl_interface': 6.4165221998951995e-06, 'learning_rate/action_model': 6.4165221998952e-05, 'epoch': 0.16}) +[2026-10-03 13:27:05,924][starVLA.training.train_starvla][INFO] - Step 2164, Loss: {'train/loss': 0.0001750838418956846, 'train/grad_norm_pre_clip': 0.08024635165929794, 'timing/data': 0.000328650014125742, 'timing/model': 0.3921049360069446, 'learning_rate/qwen_vl_interface': 6.413569803674896e-06, 'learning_rate/action_model': 6.413569803674896e-05, 'epoch': 0.16}) +[2026-10-03 13:27:06,628][starVLA.training.train_starvla][INFO] - Step 2165, Loss: {'train/loss': 0.28688615560531616, 'train/grad_norm_pre_clip': 25.524425506591797, 'timing/data': 0.00030999000591691583, 'timing/model': 0.39291233599942643, 'learning_rate/qwen_vl_interface': 6.410616929154997e-06, 'learning_rate/action_model': 6.410616929154996e-05, 'epoch': 0.16}) +[2026-10-03 13:27:07,346][starVLA.training.train_starvla][INFO] - Step 2166, Loss: {'train/loss': 3.3664000511635095e-05, 'train/grad_norm_pre_clip': 0.007606683764606714, 'timing/data': 0.0003108300006715581, 'timing/model': 0.406911443991703, 'learning_rate/qwen_vl_interface': 6.40766357754932e-06, 'learning_rate/action_model': 6.407663577549319e-05, 'epoch': 0.16}) +[2026-10-03 13:27:08,237][starVLA.training.train_starvla][INFO] - Step 2167, Loss: {'train/loss': 0.00035849513369612396, 'train/grad_norm_pre_clip': 0.0989813432097435, 'timing/data': 0.0001998899970203638, 'timing/model': 0.5823799039935693, 'learning_rate/qwen_vl_interface': 6.404709750071875e-06, 'learning_rate/action_model': 6.404709750071874e-05, 'epoch': 0.16}) +[2026-10-03 13:27:08,940][starVLA.training.train_starvla][INFO] - Step 2168, Loss: {'train/loss': 0.0022805258631706238, 'train/grad_norm_pre_clip': 7.406753063201904, 'timing/data': 0.0003163600049447268, 'timing/model': 0.3925667260045884, 'learning_rate/qwen_vl_interface': 6.401755447936869e-06, 'learning_rate/action_model': 6.401755447936869e-05, 'epoch': 0.16}) +[2026-10-03 13:27:09,644][starVLA.training.train_starvla][INFO] - Step 2169, Loss: {'train/loss': 2.2351733264258655e-07, 'train/grad_norm_pre_clip': 9.846586181083694e-05, 'timing/data': 0.00036123998870607466, 'timing/model': 0.3931544160004705, 'learning_rate/qwen_vl_interface': 6.3988006723587036e-06, 'learning_rate/action_model': 6.398800672358703e-05, 'epoch': 0.16}) +[2026-10-03 13:27:10,348][starVLA.training.train_starvla][INFO] - Step 2170, Loss: {'train/loss': 0.00017599607235752046, 'train/grad_norm_pre_clip': 0.05274539068341255, 'timing/data': 0.00020730899996124208, 'timing/model': 0.3934768970066216, 'learning_rate/qwen_vl_interface': 6.395845424551975e-06, 'learning_rate/action_model': 6.395845424551975e-05, 'epoch': 0.16}) +[2026-10-03 13:27:11,051][starVLA.training.train_starvla][INFO] - Step 2171, Loss: {'train/loss': 0.0001001728160190396, 'train/grad_norm_pre_clip': 0.029291721060872078, 'timing/data': 0.0002921799896284938, 'timing/model': 0.3921214759902796, 'learning_rate/qwen_vl_interface': 6.392889705731475e-06, 'learning_rate/action_model': 6.392889705731476e-05, 'epoch': 0.16}) +[2026-10-03 13:27:11,944][starVLA.training.train_starvla][INFO] - Step 2172, Loss: {'train/loss': 0.02844252623617649, 'train/grad_norm_pre_clip': 38.410926818847656, 'timing/data': 0.00033178999728988856, 'timing/model': 0.5831265839951811, 'learning_rate/qwen_vl_interface': 6.389933517112188e-06, 'learning_rate/action_model': 6.389933517112187e-05, 'epoch': 0.16}) +[2026-10-03 13:27:12,649][starVLA.training.train_starvla][INFO] - Step 2173, Loss: {'train/loss': 0.012981511652469635, 'train/grad_norm_pre_clip': 15.225656509399414, 'timing/data': 0.0002099900011671707, 'timing/model': 0.3935075760091422, 'learning_rate/qwen_vl_interface': 6.386976859909293e-06, 'learning_rate/action_model': 6.386976859909292e-05, 'epoch': 0.16}) +[2026-10-03 13:27:13,353][starVLA.training.train_starvla][INFO] - Step 2174, Loss: {'train/loss': 0.0007112419698387384, 'train/grad_norm_pre_clip': 0.25935277342796326, 'timing/data': 0.0002668599918251857, 'timing/model': 0.39261005600565113, 'learning_rate/qwen_vl_interface': 6.384019735338156e-06, 'learning_rate/action_model': 6.384019735338156e-05, 'epoch': 0.16}) +[2026-10-03 13:27:14,056][starVLA.training.train_starvla][INFO] - Step 2175, Loss: {'train/loss': 0.0003109777462668717, 'train/grad_norm_pre_clip': 0.0822952538728714, 'timing/data': 0.00030651000270154327, 'timing/model': 0.3927391969918972, 'learning_rate/qwen_vl_interface': 6.381062144614342e-06, 'learning_rate/action_model': 6.381062144614342e-05, 'epoch': 0.16}) +[2026-10-03 13:27:14,758][starVLA.training.train_starvla][INFO] - Step 2176, Loss: {'train/loss': 0.5314547419548035, 'train/grad_norm_pre_clip': 40.120357513427734, 'timing/data': 0.0003420890134293586, 'timing/model': 0.392263196001295, 'learning_rate/qwen_vl_interface': 6.378104088953605e-06, 'learning_rate/action_model': 6.378104088953605e-05, 'epoch': 0.16}) +[2026-10-03 13:27:15,462][starVLA.training.train_starvla][INFO] - Step 2177, Loss: {'train/loss': 0.00017767168174032122, 'train/grad_norm_pre_clip': 0.07421249896287918, 'timing/data': 0.0002485399891156703, 'timing/model': 0.39321486600965727, 'learning_rate/qwen_vl_interface': 6.375145569571889e-06, 'learning_rate/action_model': 6.375145569571889e-05, 'epoch': 0.16}) +[2026-10-03 13:27:16,359][starVLA.training.train_starvla][INFO] - Step 2178, Loss: {'train/loss': 0.00023450510343536735, 'train/grad_norm_pre_clip': 0.34315913915634155, 'timing/data': 0.00040797000110615045, 'timing/model': 0.39323463599430397, 'learning_rate/qwen_vl_interface': 6.372186587685329e-06, 'learning_rate/action_model': 6.372186587685329e-05, 'epoch': 0.16}) +[2026-10-03 13:27:17,062][starVLA.training.train_starvla][INFO] - Step 2179, Loss: {'train/loss': 0.00027356267673894763, 'train/grad_norm_pre_clip': 0.06041822209954262, 'timing/data': 0.00029117999656591564, 'timing/model': 0.39258027500181925, 'learning_rate/qwen_vl_interface': 6.369227144510252e-06, 'learning_rate/action_model': 6.369227144510252e-05, 'epoch': 0.16}) +[2026-10-03 13:27:17,764][starVLA.training.train_starvla][INFO] - Step 2180, Loss: {'train/loss': 0.1727910041809082, 'train/grad_norm_pre_clip': 145.64060974121094, 'timing/data': 0.0002497099922038615, 'timing/model': 0.39212855699588545, 'learning_rate/qwen_vl_interface': 6.3662672412631736e-06, 'learning_rate/action_model': 6.366267241263173e-05, 'epoch': 0.16}) +[2026-10-03 13:27:18,468][starVLA.training.train_starvla][INFO] - Step 2181, Loss: {'train/loss': 0.0002982321893796325, 'train/grad_norm_pre_clip': 0.08787285536527634, 'timing/data': 0.0003496200079098344, 'timing/model': 0.393079935995047, 'learning_rate/qwen_vl_interface': 6.3633068791607954e-06, 'learning_rate/action_model': 6.363306879160795e-05, 'epoch': 0.16}) +[2026-10-03 13:27:19,172][starVLA.training.train_starvla][INFO] - Step 2182, Loss: {'train/loss': 3.52001006831415e-05, 'train/grad_norm_pre_clip': 0.008675429970026016, 'timing/data': 0.00020809000125154853, 'timing/model': 0.39278856600867584, 'learning_rate/qwen_vl_interface': 6.360346059420014e-06, 'learning_rate/action_model': 6.360346059420014e-05, 'epoch': 0.16}) +[2026-10-03 13:27:20,064][starVLA.training.train_starvla][INFO] - Step 2183, Loss: {'train/loss': 0.0001061901930370368, 'train/grad_norm_pre_clip': 0.029629599303007126, 'timing/data': 0.00031971999851521105, 'timing/model': 0.5812507639930118, 'learning_rate/qwen_vl_interface': 6.357384783257909e-06, 'learning_rate/action_model': 6.357384783257908e-05, 'epoch': 0.16}) +[2026-10-03 13:27:20,766][starVLA.training.train_starvla][INFO] - Step 2184, Loss: {'train/loss': 4.988806904293597e-05, 'train/grad_norm_pre_clip': 0.012384557165205479, 'timing/data': 0.0002449400053592399, 'timing/model': 0.3922321860009106, 'learning_rate/qwen_vl_interface': 6.354423051891748e-06, 'learning_rate/action_model': 6.354423051891749e-05, 'epoch': 0.16}) +[2026-10-03 13:27:21,472][starVLA.training.train_starvla][INFO] - Step 2185, Loss: {'train/loss': 0.019487563520669937, 'train/grad_norm_pre_clip': 61.262489318847656, 'timing/data': 0.00035522000689525157, 'timing/model': 0.39327555599447805, 'learning_rate/qwen_vl_interface': 6.35146086653899e-06, 'learning_rate/action_model': 6.35146086653899e-05, 'epoch': 0.16}) +[2026-10-03 13:27:22,196][starVLA.training.train_starvla][INFO] - Step 2186, Loss: {'train/loss': 4.222877396387048e-05, 'train/grad_norm_pre_clip': 0.010096491314470768, 'timing/data': 0.0002238099987152964, 'timing/model': 0.4081806839967612, 'learning_rate/qwen_vl_interface': 6.348498228417273e-06, 'learning_rate/action_model': 6.348498228417273e-05, 'epoch': 0.16}) +[2026-10-03 13:27:22,902][starVLA.training.train_starvla][INFO] - Step 2187, Loss: {'train/loss': 0.22881214320659637, 'train/grad_norm_pre_clip': 33.88502502441406, 'timing/data': 0.00025212000764440745, 'timing/model': 0.39442276500631124, 'learning_rate/qwen_vl_interface': 6.34553513874443e-06, 'learning_rate/action_model': 6.34553513874443e-05, 'epoch': 0.16}) +[2026-10-03 13:27:23,608][starVLA.training.train_starvla][INFO] - Step 2188, Loss: {'train/loss': 0.00020425707043614239, 'train/grad_norm_pre_clip': 0.08943193405866623, 'timing/data': 0.0003954599960707128, 'timing/model': 0.3943502150068525, 'learning_rate/qwen_vl_interface': 6.342571598738473e-06, 'learning_rate/action_model': 6.342571598738472e-05, 'epoch': 0.16}) +[2026-10-03 13:27:24,522][starVLA.training.train_starvla][INFO] - Step 2189, Loss: {'train/loss': 0.000322947307722643, 'train/grad_norm_pre_clip': 0.07636482268571854, 'timing/data': 0.0003086100041400641, 'timing/model': 0.3948673859995324, 'learning_rate/qwen_vl_interface': 6.339607609617603e-06, 'learning_rate/action_model': 6.339607609617603e-05, 'epoch': 0.16}) +[2026-10-03 13:27:25,232][starVLA.training.train_starvla][INFO] - Step 2190, Loss: {'train/loss': 3.100053800153546e-05, 'train/grad_norm_pre_clip': 0.005563409999012947, 'timing/data': 0.00030406999576371163, 'timing/model': 0.3958433650113875, 'learning_rate/qwen_vl_interface': 6.336643172600202e-06, 'learning_rate/action_model': 6.336643172600202e-05, 'epoch': 0.16}) +[2026-10-03 13:27:25,939][starVLA.training.train_starvla][INFO] - Step 2191, Loss: {'train/loss': 0.06651989370584488, 'train/grad_norm_pre_clip': 114.83338165283203, 'timing/data': 0.00022236000222619623, 'timing/model': 0.39449041499756277, 'learning_rate/qwen_vl_interface': 6.333678288904838e-06, 'learning_rate/action_model': 6.333678288904837e-05, 'epoch': 0.16}) +[2026-10-03 13:27:26,645][starVLA.training.train_starvla][INFO] - Step 2192, Loss: {'train/loss': 1.862644296579674e-07, 'train/grad_norm_pre_clip': 8.853094914229587e-05, 'timing/data': 0.00026467000134289265, 'timing/model': 0.3951815660111606, 'learning_rate/qwen_vl_interface': 6.330712959750264e-06, 'learning_rate/action_model': 6.330712959750263e-05, 'epoch': 0.16}) +[2026-10-03 13:27:27,354][starVLA.training.train_starvla][INFO] - Step 2193, Loss: {'train/loss': 1.7115039825439453, 'train/grad_norm_pre_clip': 21.18262481689453, 'timing/data': 0.00035432999720796943, 'timing/model': 0.39497356599895284, 'learning_rate/qwen_vl_interface': 6.3277471863554145e-06, 'learning_rate/action_model': 6.327747186355414e-05, 'epoch': 0.16}) +[2026-10-03 13:27:28,260][starVLA.training.train_starvla][INFO] - Step 2194, Loss: {'train/loss': 0.00012771598994731903, 'train/grad_norm_pre_clip': 0.04179314523935318, 'timing/data': 0.00029953000193927437, 'timing/model': 0.39588367500982713, 'learning_rate/qwen_vl_interface': 6.324780969939406e-06, 'learning_rate/action_model': 6.324780969939405e-05, 'epoch': 0.16}) +[2026-10-03 13:27:28,966][starVLA.training.train_starvla][INFO] - Step 2195, Loss: {'train/loss': 0.0006588913965970278, 'train/grad_norm_pre_clip': 0.33840110898017883, 'timing/data': 0.00028951000422239304, 'timing/model': 0.3947391859983327, 'learning_rate/qwen_vl_interface': 6.32181431172154e-06, 'learning_rate/action_model': 6.32181431172154e-05, 'epoch': 0.16}) +[2026-10-03 13:27:29,673][starVLA.training.train_starvla][INFO] - Step 2196, Loss: {'train/loss': 0.00010394638229627162, 'train/grad_norm_pre_clip': 0.017626356333494186, 'timing/data': 0.00020183999731671065, 'timing/model': 0.39442151501134504, 'learning_rate/qwen_vl_interface': 6.318847212921295e-06, 'learning_rate/action_model': 6.318847212921295e-05, 'epoch': 0.16}) +[2026-10-03 13:27:30,381][starVLA.training.train_starvla][INFO] - Step 2197, Loss: {'train/loss': 8.231351239373907e-05, 'train/grad_norm_pre_clip': 0.021712910383939743, 'timing/data': 0.0003801099956035614, 'timing/model': 0.3946835649985587, 'learning_rate/qwen_vl_interface': 6.3158796747583364e-06, 'learning_rate/action_model': 6.315879674758336e-05, 'epoch': 0.16}) +[2026-10-03 13:27:31,091][starVLA.training.train_starvla][INFO] - Step 2198, Loss: {'train/loss': 0.003365475684404373, 'train/grad_norm_pre_clip': 2.1205508708953857, 'timing/data': 0.0002929600013885647, 'timing/model': 0.396439375996124, 'learning_rate/qwen_vl_interface': 6.312911698452503e-06, 'learning_rate/action_model': 6.312911698452502e-05, 'epoch': 0.16}) +[2026-10-03 13:27:31,992][starVLA.training.train_starvla][INFO] - Step 2199, Loss: {'train/loss': 0.00010686452151276171, 'train/grad_norm_pre_clip': 0.03565200790762901, 'timing/data': 0.0003262599930167198, 'timing/model': 0.5887619740096852, 'learning_rate/qwen_vl_interface': 6.309943285223824e-06, 'learning_rate/action_model': 6.309943285223824e-05, 'epoch': 0.16}) +[2026-10-03 13:27:32,698][starVLA.training.train_starvla][INFO] - Step 2200, Loss: {'train/loss': 0.001317015034146607, 'train/grad_norm_pre_clip': 1.2908508777618408, 'timing/data': 0.00020942999981343746, 'timing/model': 0.39480432600248605, 'learning_rate/qwen_vl_interface': 6.306974436292493e-06, 'learning_rate/action_model': 6.306974436292494e-05, 'epoch': 0.16}) +[2026-10-03 13:27:33,407][starVLA.training.train_starvla][INFO] - Step 2201, Loss: {'train/loss': 0.0004602441913448274, 'train/grad_norm_pre_clip': 0.2586609721183777, 'timing/data': 0.00034780999703798443, 'timing/model': 0.39503577499999665, 'learning_rate/qwen_vl_interface': 6.304005152878902e-06, 'learning_rate/action_model': 6.304005152878901e-05, 'epoch': 0.16}) +[2026-10-03 13:27:34,116][starVLA.training.train_starvla][INFO] - Step 2202, Loss: {'train/loss': 0.2921507954597473, 'train/grad_norm_pre_clip': 299.87615966796875, 'timing/data': 0.00030414998764172196, 'timing/model': 0.39644628499809187, 'learning_rate/qwen_vl_interface': 6.301035436203602e-06, 'learning_rate/action_model': 6.301035436203602e-05, 'epoch': 0.16}) +[2026-10-03 13:27:34,825][starVLA.training.train_starvla][INFO] - Step 2203, Loss: {'train/loss': 0.34177741408348083, 'train/grad_norm_pre_clip': 59.841548919677734, 'timing/data': 0.00029815999732818455, 'timing/model': 0.3950586649880279, 'learning_rate/qwen_vl_interface': 6.298065287487337e-06, 'learning_rate/action_model': 6.298065287487337e-05, 'epoch': 0.16}) +[2026-10-03 13:27:35,532][starVLA.training.train_starvla][INFO] - Step 2204, Loss: {'train/loss': 0.0037266146391630173, 'train/grad_norm_pre_clip': 12.722701072692871, 'timing/data': 0.00020348999532870948, 'timing/model': 0.3947029660048429, 'learning_rate/qwen_vl_interface': 6.295094707951019e-06, 'learning_rate/action_model': 6.295094707951019e-05, 'epoch': 0.16}) +[2026-10-03 13:27:36,435][starVLA.training.train_starvla][INFO] - Step 2205, Loss: {'train/loss': 0.076483353972435, 'train/grad_norm_pre_clip': 137.18212890625, 'timing/data': 0.0002194700064137578, 'timing/model': 0.39525800599949434, 'learning_rate/qwen_vl_interface': 6.292123698815745e-06, 'learning_rate/action_model': 6.292123698815745e-05, 'epoch': 0.16}) +[2026-10-03 13:27:37,144][starVLA.training.train_starvla][INFO] - Step 2206, Loss: {'train/loss': 1.861649434431456e-05, 'train/grad_norm_pre_clip': 0.006152177229523659, 'timing/data': 0.00020870000298600644, 'timing/model': 0.3966263550100848, 'learning_rate/qwen_vl_interface': 6.289152261302783e-06, 'learning_rate/action_model': 6.289152261302784e-05, 'epoch': 0.16}) +[2026-10-03 13:27:37,848][starVLA.training.train_starvla][INFO] - Step 2207, Loss: {'train/loss': 0.6408892273902893, 'train/grad_norm_pre_clip': 82.4492416381836, 'timing/data': 0.00020154999219812453, 'timing/model': 0.39409311600320507, 'learning_rate/qwen_vl_interface': 6.286180396633579e-06, 'learning_rate/action_model': 6.286180396633579e-05, 'epoch': 0.16}) +[2026-10-03 13:27:38,555][starVLA.training.train_starvla][INFO] - Step 2208, Loss: {'train/loss': 0.0021671298891305923, 'train/grad_norm_pre_clip': 3.3675615787506104, 'timing/data': 0.00031347999174613506, 'timing/model': 0.395133245998295, 'learning_rate/qwen_vl_interface': 6.283208106029753e-06, 'learning_rate/action_model': 6.283208106029752e-05, 'epoch': 0.16}) +[2026-10-03 13:27:39,264][starVLA.training.train_starvla][INFO] - Step 2209, Loss: {'train/loss': 0.5874370336532593, 'train/grad_norm_pre_clip': 26.34820556640625, 'timing/data': 0.0003385800082469359, 'timing/model': 0.3950379960006103, 'learning_rate/qwen_vl_interface': 6.280235390713102e-06, 'learning_rate/action_model': 6.280235390713102e-05, 'epoch': 0.16}) +[2026-10-03 13:27:40,169][starVLA.training.train_starvla][INFO] - Step 2210, Loss: {'train/loss': 0.00027565108030103147, 'train/grad_norm_pre_clip': 0.21306701004505157, 'timing/data': 0.00024684000527486205, 'timing/model': 0.39560636499663815, 'learning_rate/qwen_vl_interface': 6.277262251905597e-06, 'learning_rate/action_model': 6.277262251905597e-05, 'epoch': 0.16}) +[2026-10-03 13:27:40,876][starVLA.training.train_starvla][INFO] - Step 2211, Loss: {'train/loss': 0.015730271115899086, 'train/grad_norm_pre_clip': 6.6220293045043945, 'timing/data': 0.0002953999937744811, 'timing/model': 0.39418888599902857, 'learning_rate/qwen_vl_interface': 6.274288690829385e-06, 'learning_rate/action_model': 6.274288690829384e-05, 'epoch': 0.16}) +[2026-10-03 13:27:41,583][starVLA.training.train_starvla][INFO] - Step 2212, Loss: {'train/loss': 0.30196985602378845, 'train/grad_norm_pre_clip': 37.57975387573242, 'timing/data': 0.0002115900133503601, 'timing/model': 0.39514450599381234, 'learning_rate/qwen_vl_interface': 6.271314708706781e-06, 'learning_rate/action_model': 6.271314708706782e-05, 'epoch': 0.16}) +[2026-10-03 13:27:42,292][starVLA.training.train_starvla][INFO] - Step 2213, Loss: {'train/loss': 0.31377315521240234, 'train/grad_norm_pre_clip': 252.95980834960938, 'timing/data': 0.000397639989387244, 'timing/model': 0.3950876659946516, 'learning_rate/qwen_vl_interface': 6.26834030676028e-06, 'learning_rate/action_model': 6.26834030676028e-05, 'epoch': 0.16}) +[2026-10-03 13:27:43,000][starVLA.training.train_starvla][INFO] - Step 2214, Loss: {'train/loss': 0.498099684715271, 'train/grad_norm_pre_clip': 18.488853454589844, 'timing/data': 0.0003097199951298535, 'timing/model': 0.396120974997757, 'learning_rate/qwen_vl_interface': 6.265365486212545e-06, 'learning_rate/action_model': 6.265365486212545e-05, 'epoch': 0.16}) +[2026-10-03 13:27:43,900][starVLA.training.train_starvla][INFO] - Step 2215, Loss: {'train/loss': 0.00018783917766995728, 'train/grad_norm_pre_clip': 0.06396634876728058, 'timing/data': 0.00033845999860204756, 'timing/model': 0.3950679949921323, 'learning_rate/qwen_vl_interface': 6.26239024828641e-06, 'learning_rate/action_model': 6.26239024828641e-05, 'epoch': 0.16}) +[2026-10-03 13:27:44,608][starVLA.training.train_starvla][INFO] - Step 2216, Loss: {'train/loss': 3.872346132993698e-05, 'train/grad_norm_pre_clip': 0.012500370852649212, 'timing/data': 0.0003402799920877442, 'timing/model': 0.3953239759866847, 'learning_rate/qwen_vl_interface': 6.259414594204888e-06, 'learning_rate/action_model': 6.259414594204888e-05, 'epoch': 0.16}) +[2026-10-03 13:27:45,317][starVLA.training.train_starvla][INFO] - Step 2217, Loss: {'train/loss': 0.00015968848310876638, 'train/grad_norm_pre_clip': 0.08555018901824951, 'timing/data': 0.00038096999924164265, 'timing/model': 0.39552904599986505, 'learning_rate/qwen_vl_interface': 6.256438525191155e-06, 'learning_rate/action_model': 6.256438525191156e-05, 'epoch': 0.16}) +[2026-10-03 13:27:46,027][starVLA.training.train_starvla][INFO] - Step 2218, Loss: {'train/loss': 0.006966888438910246, 'train/grad_norm_pre_clip': 11.10684585571289, 'timing/data': 0.00036286000977270305, 'timing/model': 0.39615175599465147, 'learning_rate/qwen_vl_interface': 6.253462042468561e-06, 'learning_rate/action_model': 6.25346204246856e-05, 'epoch': 0.16}) +[2026-10-03 13:27:46,735][starVLA.training.train_starvla][INFO] - Step 2219, Loss: {'train/loss': 0.00013623513223137707, 'train/grad_norm_pre_clip': 0.03312402218580246, 'timing/data': 0.00029888999415561557, 'timing/model': 0.3952036050031893, 'learning_rate/qwen_vl_interface': 6.2504851472606246e-06, 'learning_rate/action_model': 6.250485147260625e-05, 'epoch': 0.16}) +[2026-10-03 13:27:47,632][starVLA.training.train_starvla][INFO] - Step 2220, Loss: {'train/loss': 0.0018881798023357987, 'train/grad_norm_pre_clip': 3.4114906787872314, 'timing/data': 0.00025458000891376287, 'timing/model': 0.5851833039923804, 'learning_rate/qwen_vl_interface': 6.247507840791035e-06, 'learning_rate/action_model': 6.247507840791035e-05, 'epoch': 0.16}) +[2026-10-03 13:27:48,342][starVLA.training.train_starvla][INFO] - Step 2221, Loss: {'train/loss': 9.398241672897711e-05, 'train/grad_norm_pre_clip': 0.04397639259696007, 'timing/data': 0.0003541999903973192, 'timing/model': 0.39590426499489695, 'learning_rate/qwen_vl_interface': 6.244530124283651e-06, 'learning_rate/action_model': 6.24453012428365e-05, 'epoch': 0.16}) +[2026-10-03 13:27:49,050][starVLA.training.train_starvla][INFO] - Step 2222, Loss: {'train/loss': 8.744003571337089e-05, 'train/grad_norm_pre_clip': 0.03919572755694389, 'timing/data': 0.000305479989037849, 'timing/model': 0.39611122499627527, 'learning_rate/qwen_vl_interface': 6.2415519989625016e-06, 'learning_rate/action_model': 6.241551998962501e-05, 'epoch': 0.16}) +[2026-10-03 13:27:49,757][starVLA.training.train_starvla][INFO] - Step 2223, Loss: {'train/loss': 0.4891662299633026, 'train/grad_norm_pre_clip': 147.3284454345703, 'timing/data': 0.00022429999080486596, 'timing/model': 0.39473962600459345, 'learning_rate/qwen_vl_interface': 6.238573466051781e-06, 'learning_rate/action_model': 6.23857346605178e-05, 'epoch': 0.16}) +[2026-10-03 13:27:50,465][starVLA.training.train_starvla][INFO] - Step 2224, Loss: {'train/loss': 0.016963379457592964, 'train/grad_norm_pre_clip': 18.065534591674805, 'timing/data': 0.00026972999330610037, 'timing/model': 0.3953200460091466, 'learning_rate/qwen_vl_interface': 6.2355945267758465e-06, 'learning_rate/action_model': 6.235594526775846e-05, 'epoch': 0.16}) +[2026-10-03 13:27:51,367][starVLA.training.train_starvla][INFO] - Step 2225, Loss: {'train/loss': 0.23562419414520264, 'train/grad_norm_pre_clip': 95.54269409179688, 'timing/data': 0.00026647999766282737, 'timing/model': 0.5882043640012853, 'learning_rate/qwen_vl_interface': 6.232615182359231e-06, 'learning_rate/action_model': 6.232615182359231e-05, 'epoch': 0.16}) +[2026-10-03 13:27:52,076][starVLA.training.train_starvla][INFO] - Step 2226, Loss: {'train/loss': 0.0022252846974879503, 'train/grad_norm_pre_clip': 1.576737642288208, 'timing/data': 0.00020696000137832016, 'timing/model': 0.39629883599991444, 'learning_rate/qwen_vl_interface': 6.229635434026634e-06, 'learning_rate/action_model': 6.229635434026634e-05, 'epoch': 0.16}) +[2026-10-03 13:27:52,782][starVLA.training.train_starvla][INFO] - Step 2227, Loss: {'train/loss': 0.3968203365802765, 'train/grad_norm_pre_clip': 103.85211944580078, 'timing/data': 0.00020433000463526696, 'timing/model': 0.394399565004278, 'learning_rate/qwen_vl_interface': 6.226655283002912e-06, 'learning_rate/action_model': 6.226655283002911e-05, 'epoch': 0.16}) +[2026-10-03 13:27:53,490][starVLA.training.train_starvla][INFO] - Step 2228, Loss: {'train/loss': 0.00652357516810298, 'train/grad_norm_pre_clip': 7.181267261505127, 'timing/data': 0.0003134199941996485, 'timing/model': 0.3951388660061639, 'learning_rate/qwen_vl_interface': 6.223674730513096e-06, 'learning_rate/action_model': 6.223674730513096e-05, 'epoch': 0.16}) +[2026-10-03 13:27:54,199][starVLA.training.train_starvla][INFO] - Step 2229, Loss: {'train/loss': 0.0004630057665053755, 'train/grad_norm_pre_clip': 0.14511464536190033, 'timing/data': 0.00031180000223685056, 'timing/model': 0.3952175860031275, 'learning_rate/qwen_vl_interface': 6.220693777782375e-06, 'learning_rate/action_model': 6.220693777782375e-05, 'epoch': 0.16}) +[2026-10-03 13:27:54,907][starVLA.training.train_starvla][INFO] - Step 2230, Loss: {'train/loss': 0.00037110032280907035, 'train/grad_norm_pre_clip': 0.07408041507005692, 'timing/data': 0.0003127000090898946, 'timing/model': 0.3960157049878035, 'learning_rate/qwen_vl_interface': 6.21771242603611e-06, 'learning_rate/action_model': 6.21771242603611e-05, 'epoch': 0.16}) +[2026-10-03 13:27:55,810][starVLA.training.train_starvla][INFO] - Step 2231, Loss: {'train/loss': 2.4801640392979607e-05, 'train/grad_norm_pre_clip': 0.005242232233285904, 'timing/data': 0.00025505000667180866, 'timing/model': 0.3940213160094572, 'learning_rate/qwen_vl_interface': 6.21473067649982e-06, 'learning_rate/action_model': 6.21473067649982e-05, 'epoch': 0.16}) +[2026-10-03 13:27:56,518][starVLA.training.train_starvla][INFO] - Step 2232, Loss: {'train/loss': 0.3949296176433563, 'train/grad_norm_pre_clip': 34.175506591796875, 'timing/data': 0.00022841000463813543, 'timing/model': 0.3951789160055341, 'learning_rate/qwen_vl_interface': 6.211748530399189e-06, 'learning_rate/action_model': 6.211748530399189e-05, 'epoch': 0.16}) +[2026-10-03 13:27:57,228][starVLA.training.train_starvla][INFO] - Step 2233, Loss: {'train/loss': 0.0016439611790701747, 'train/grad_norm_pre_clip': 0.8506012558937073, 'timing/data': 0.000271140001132153, 'timing/model': 0.3953722649894189, 'learning_rate/qwen_vl_interface': 6.208765988960067e-06, 'learning_rate/action_model': 6.208765988960066e-05, 'epoch': 0.16}) +[2026-10-03 13:27:57,936][starVLA.training.train_starvla][INFO] - Step 2234, Loss: {'train/loss': 0.5688493847846985, 'train/grad_norm_pre_clip': 64.01770782470703, 'timing/data': 0.00030521900043822825, 'timing/model': 0.3964895060053095, 'learning_rate/qwen_vl_interface': 6.2057830534084615e-06, 'learning_rate/action_model': 6.205783053408461e-05, 'epoch': 0.16}) +[2026-10-03 13:27:58,643][starVLA.training.train_starvla][INFO] - Step 2235, Loss: {'train/loss': 0.0004365279455669224, 'train/grad_norm_pre_clip': 0.13579516112804413, 'timing/data': 0.00031102998764254153, 'timing/model': 0.39486009601387195, 'learning_rate/qwen_vl_interface': 6.202799724970546e-06, 'learning_rate/action_model': 6.202799724970546e-05, 'epoch': 0.16}) +[2026-10-03 13:27:59,540][starVLA.training.train_starvla][INFO] - Step 2236, Loss: {'train/loss': 0.02115079201757908, 'train/grad_norm_pre_clip': 15.154031753540039, 'timing/data': 0.0002369699941482395, 'timing/model': 0.5836141850013519, 'learning_rate/qwen_vl_interface': 6.199816004872655e-06, 'learning_rate/action_model': 6.199816004872655e-05, 'epoch': 0.16}) +[2026-10-03 13:28:00,248][starVLA.training.train_starvla][INFO] - Step 2237, Loss: {'train/loss': 0.0006420080899260938, 'train/grad_norm_pre_clip': 0.9052807688713074, 'timing/data': 0.0002795890031848103, 'timing/model': 0.3951946860033786, 'learning_rate/qwen_vl_interface': 6.196831894341281e-06, 'learning_rate/action_model': 6.19683189434128e-05, 'epoch': 0.16}) +[2026-10-03 13:28:00,957][starVLA.training.train_starvla][INFO] - Step 2238, Loss: {'train/loss': 0.1521754115819931, 'train/grad_norm_pre_clip': 23.98128890991211, 'timing/data': 0.0002676299918675795, 'timing/model': 0.3963076350046322, 'learning_rate/qwen_vl_interface': 6.1938473946030795e-06, 'learning_rate/action_model': 6.193847394603079e-05, 'epoch': 0.16}) +[2026-10-03 13:28:01,664][starVLA.training.train_starvla][INFO] - Step 2239, Loss: {'train/loss': 0.0052282377146184444, 'train/grad_norm_pre_clip': 2.920489549636841, 'timing/data': 0.0003097299922956154, 'timing/model': 0.3946099559980212, 'learning_rate/qwen_vl_interface': 6.1908625068848695e-06, 'learning_rate/action_model': 6.19086250688487e-05, 'epoch': 0.16}) +[2026-10-03 13:28:02,371][starVLA.training.train_starvla][INFO] - Step 2240, Loss: {'train/loss': 0.0007520313374698162, 'train/grad_norm_pre_clip': 0.23828160762786865, 'timing/data': 0.0003269400040153414, 'timing/model': 0.39482404600130394, 'learning_rate/qwen_vl_interface': 6.187877232413619e-06, 'learning_rate/action_model': 6.18787723241362e-05, 'epoch': 0.16}) +[2026-10-03 13:28:03,078][starVLA.training.train_starvla][INFO] - Step 2241, Loss: {'train/loss': 0.001191123272292316, 'train/grad_norm_pre_clip': 0.3063180148601532, 'timing/data': 0.00037372000224422663, 'timing/model': 0.3934487059887033, 'learning_rate/qwen_vl_interface': 6.184891572416467e-06, 'learning_rate/action_model': 6.184891572416466e-05, 'epoch': 0.16}) +[2026-10-03 13:28:03,976][starVLA.training.train_starvla][INFO] - Step 2242, Loss: {'train/loss': 0.041975267231464386, 'train/grad_norm_pre_clip': 62.838321685791016, 'timing/data': 0.0002895799989346415, 'timing/model': 0.39642874599667266, 'learning_rate/qwen_vl_interface': 6.181905528120703e-06, 'learning_rate/action_model': 6.181905528120703e-05, 'epoch': 0.16}) +[2026-10-03 13:28:04,683][starVLA.training.train_starvla][INFO] - Step 2243, Loss: {'train/loss': 0.0013074394082650542, 'train/grad_norm_pre_clip': 0.6162928342819214, 'timing/data': 0.00030193000566214323, 'timing/model': 0.39449677600350697, 'learning_rate/qwen_vl_interface': 6.178919100753778e-06, 'learning_rate/action_model': 6.178919100753778e-05, 'epoch': 0.16}) +[2026-10-03 13:28:05,391][starVLA.training.train_starvla][INFO] - Step 2244, Loss: {'train/loss': 4.00662756874226e-05, 'train/grad_norm_pre_clip': 0.008692936040461063, 'timing/data': 0.0002528100012568757, 'timing/model': 0.39487077500962187, 'learning_rate/qwen_vl_interface': 6.175932291543299e-06, 'learning_rate/action_model': 6.175932291543298e-05, 'epoch': 0.16}) +[2026-10-03 13:28:06,100][starVLA.training.train_starvla][INFO] - Step 2245, Loss: {'train/loss': 0.00010608257434796542, 'train/grad_norm_pre_clip': 0.0658329427242279, 'timing/data': 0.00026993000938091427, 'timing/model': 0.39560119500674773, 'learning_rate/qwen_vl_interface': 6.172945101717029e-06, 'learning_rate/action_model': 6.17294510171703e-05, 'epoch': 0.16}) +[2026-10-03 13:28:06,810][starVLA.training.train_starvla][INFO] - Step 2246, Loss: {'train/loss': 0.0008349369745701551, 'train/grad_norm_pre_clip': 0.5254420638084412, 'timing/data': 0.00023243000032380223, 'timing/model': 0.39655100599338766, 'learning_rate/qwen_vl_interface': 6.169957532502892e-06, 'learning_rate/action_model': 6.169957532502892e-05, 'epoch': 0.16}) +[2026-10-03 13:28:07,711][starVLA.training.train_starvla][INFO] - Step 2247, Loss: {'train/loss': 0.0005183401517570019, 'train/grad_norm_pre_clip': 0.2407417595386505, 'timing/data': 0.00025159999495372176, 'timing/model': 0.3949803459981922, 'learning_rate/qwen_vl_interface': 6.166969585128962e-06, 'learning_rate/action_model': 6.166969585128962e-05, 'epoch': 0.16}) +[2026-10-03 13:28:08,418][starVLA.training.train_starvla][INFO] - Step 2248, Loss: {'train/loss': 4.554953193292022e-05, 'train/grad_norm_pre_clip': 0.06683448702096939, 'timing/data': 0.00023063999833539128, 'timing/model': 0.39486071599822026, 'learning_rate/qwen_vl_interface': 6.163981260823473e-06, 'learning_rate/action_model': 6.163981260823473e-05, 'epoch': 0.16}) +[2026-10-03 13:28:09,126][starVLA.training.train_starvla][INFO] - Step 2249, Loss: {'train/loss': 0.11945635825395584, 'train/grad_norm_pre_clip': 84.10358428955078, 'timing/data': 0.0003241399972466752, 'timing/model': 0.39465014600136783, 'learning_rate/qwen_vl_interface': 6.16099256081481e-06, 'learning_rate/action_model': 6.16099256081481e-05, 'epoch': 0.16}) +[2026-10-03 13:28:09,835][starVLA.training.train_starvla][INFO] - Step 2250, Loss: {'train/loss': 1.577226066729054e-05, 'train/grad_norm_pre_clip': 0.004216201137751341, 'timing/data': 0.00029121000261511654, 'timing/model': 0.3971209450101014, 'learning_rate/qwen_vl_interface': 6.15800348633152e-06, 'learning_rate/action_model': 6.15800348633152e-05, 'epoch': 0.16}) +[2026-10-03 13:28:10,542][starVLA.training.train_starvla][INFO] - Step 2251, Loss: {'train/loss': 0.0034768283367156982, 'train/grad_norm_pre_clip': 4.778720855712891, 'timing/data': 0.0002059599937638268, 'timing/model': 0.39527097599057015, 'learning_rate/qwen_vl_interface': 6.155014038602293e-06, 'learning_rate/action_model': 6.155014038602292e-05, 'epoch': 0.16}) +[2026-10-03 13:28:11,441][starVLA.training.train_starvla][INFO] - Step 2252, Loss: {'train/loss': 0.00028091634158045053, 'train/grad_norm_pre_clip': 0.1660662591457367, 'timing/data': 0.00020896000205539167, 'timing/model': 0.5861702039983356, 'learning_rate/qwen_vl_interface': 6.152024218855979e-06, 'learning_rate/action_model': 6.15202421885598e-05, 'epoch': 0.16}) +[2026-10-03 13:28:12,150][starVLA.training.train_starvla][INFO] - Step 2253, Loss: {'train/loss': 0.00027844548458233476, 'train/grad_norm_pre_clip': 0.08843550086021423, 'timing/data': 0.0003584299993235618, 'timing/model': 0.3952928359940415, 'learning_rate/qwen_vl_interface': 6.149034028321583e-06, 'learning_rate/action_model': 6.149034028321583e-05, 'epoch': 0.16}) +[2026-10-03 13:28:12,859][starVLA.training.train_starvla][INFO] - Step 2254, Loss: {'train/loss': 0.00027670973213389516, 'train/grad_norm_pre_clip': 0.11183086782693863, 'timing/data': 0.00028717899112962186, 'timing/model': 0.39674281601037364, 'learning_rate/qwen_vl_interface': 6.1460434682282566e-06, 'learning_rate/action_model': 6.146043468228256e-05, 'epoch': 0.16}) +[2026-10-03 13:28:13,566][starVLA.training.train_starvla][INFO] - Step 2255, Loss: {'train/loss': 0.027997275814414024, 'train/grad_norm_pre_clip': 48.16401290893555, 'timing/data': 0.000231740006711334, 'timing/model': 0.3945963950100122, 'learning_rate/qwen_vl_interface': 6.143052539805308e-06, 'learning_rate/action_model': 6.143052539805308e-05, 'epoch': 0.17}) +[2026-10-03 13:28:14,273][starVLA.training.train_starvla][INFO] - Step 2256, Loss: {'train/loss': 5.131920988787897e-05, 'train/grad_norm_pre_clip': 0.018318988382816315, 'timing/data': 0.00023015000624582171, 'timing/model': 0.3946902659954503, 'learning_rate/qwen_vl_interface': 6.140061244282194e-06, 'learning_rate/action_model': 6.140061244282194e-05, 'epoch': 0.17}) +[2026-10-03 13:28:14,983][starVLA.training.train_starvla][INFO] - Step 2257, Loss: {'train/loss': 0.0001562294492032379, 'train/grad_norm_pre_clip': 0.036413438618183136, 'timing/data': 0.00034435000270605087, 'timing/model': 0.39514198500546627, 'learning_rate/qwen_vl_interface': 6.137069582888521e-06, 'learning_rate/action_model': 6.137069582888521e-05, 'epoch': 0.17}) +[2026-10-03 13:28:15,880][starVLA.training.train_starvla][INFO] - Step 2258, Loss: {'train/loss': 0.416620671749115, 'train/grad_norm_pre_clip': 42.86699676513672, 'timing/data': 0.0003730099997483194, 'timing/model': 0.396414664995973, 'learning_rate/qwen_vl_interface': 6.13407755685405e-06, 'learning_rate/action_model': 6.13407755685405e-05, 'epoch': 0.17}) +[2026-10-03 13:28:16,587][starVLA.training.train_starvla][INFO] - Step 2259, Loss: {'train/loss': 0.00015850945783313364, 'train/grad_norm_pre_clip': 0.09970665723085403, 'timing/data': 0.00032364000799134374, 'timing/model': 0.39483727600600105, 'learning_rate/qwen_vl_interface': 6.131085167408692e-06, 'learning_rate/action_model': 6.131085167408692e-05, 'epoch': 0.17}) +[2026-10-03 13:28:17,295][starVLA.training.train_starvla][INFO] - Step 2260, Loss: {'train/loss': 0.01508363801985979, 'train/grad_norm_pre_clip': 19.228580474853516, 'timing/data': 0.00032001000363379717, 'timing/model': 0.3947734660032438, 'learning_rate/qwen_vl_interface': 6.128092415782502e-06, 'learning_rate/action_model': 6.128092415782501e-05, 'epoch': 0.17}) +[2026-10-03 13:28:18,003][starVLA.training.train_starvla][INFO] - Step 2261, Loss: {'train/loss': 8.143074410327245e-06, 'train/grad_norm_pre_clip': 0.002848448231816292, 'timing/data': 0.000314689998049289, 'timing/model': 0.39476203500817064, 'learning_rate/qwen_vl_interface': 6.125099303205688e-06, 'learning_rate/action_model': 6.125099303205688e-05, 'epoch': 0.17}) +[2026-10-03 13:28:18,713][starVLA.training.train_starvla][INFO] - Step 2262, Loss: {'train/loss': 0.3108433485031128, 'train/grad_norm_pre_clip': 198.24978637695312, 'timing/data': 0.00031542000942863524, 'timing/model': 0.39688308599579614, 'learning_rate/qwen_vl_interface': 6.122105830908607e-06, 'learning_rate/action_model': 6.122105830908606e-05, 'epoch': 0.17}) +[2026-10-03 13:28:19,613][starVLA.training.train_starvla][INFO] - Step 2263, Loss: {'train/loss': 1.5183786672423594e-05, 'train/grad_norm_pre_clip': 0.005558453034609556, 'timing/data': 0.00031455900170840323, 'timing/model': 0.39439608600514475, 'learning_rate/qwen_vl_interface': 6.11911200012176e-06, 'learning_rate/action_model': 6.11911200012176e-05, 'epoch': 0.17}) +[2026-10-03 13:28:20,319][starVLA.training.train_starvla][INFO] - Step 2264, Loss: {'train/loss': 0.00011113745131297037, 'train/grad_norm_pre_clip': 0.021161383017897606, 'timing/data': 0.00030528000206686556, 'timing/model': 0.3931678350054426, 'learning_rate/qwen_vl_interface': 6.116117812075799e-06, 'learning_rate/action_model': 6.116117812075799e-05, 'epoch': 0.17}) +[2026-10-03 13:28:21,026][starVLA.training.train_starvla][INFO] - Step 2265, Loss: {'train/loss': 0.730959951877594, 'train/grad_norm_pre_clip': 98.3194808959961, 'timing/data': 0.0003473899996606633, 'timing/model': 0.3950244949955959, 'learning_rate/qwen_vl_interface': 6.113123268001524e-06, 'learning_rate/action_model': 6.113123268001524e-05, 'epoch': 0.17}) +[2026-10-03 13:28:21,736][starVLA.training.train_starvla][INFO] - Step 2266, Loss: {'train/loss': 6.980945727264043e-06, 'train/grad_norm_pre_clip': 0.0025518147740513086, 'timing/data': 0.0002575599937699735, 'timing/model': 0.3974296850065002, 'learning_rate/qwen_vl_interface': 6.110128369129873e-06, 'learning_rate/action_model': 6.110128369129873e-05, 'epoch': 0.17}) +[2026-10-03 13:28:22,443][starVLA.training.train_starvla][INFO] - Step 2267, Loss: {'train/loss': 0.0001984480331884697, 'train/grad_norm_pre_clip': 0.4310334324836731, 'timing/data': 0.0002566799958003685, 'timing/model': 0.39449298499675933, 'learning_rate/qwen_vl_interface': 6.107133116691943e-06, 'learning_rate/action_model': 6.107133116691942e-05, 'epoch': 0.17}) +[2026-10-03 13:28:23,344][starVLA.training.train_starvla][INFO] - Step 2268, Loss: {'train/loss': 0.43904635310173035, 'train/grad_norm_pre_clip': 22.954404830932617, 'timing/data': 0.00022092000290285796, 'timing/model': 0.39569838500756305, 'learning_rate/qwen_vl_interface': 6.104137511918963e-06, 'learning_rate/action_model': 6.104137511918963e-05, 'epoch': 0.17}) +[2026-10-03 13:28:24,053][starVLA.training.train_starvla][INFO] - Step 2269, Loss: {'train/loss': 0.00013448302343022078, 'train/grad_norm_pre_clip': 0.028631998226046562, 'timing/data': 0.0002231800026493147, 'timing/model': 0.39534152499982156, 'learning_rate/qwen_vl_interface': 6.1011415560423155e-06, 'learning_rate/action_model': 6.101141556042316e-05, 'epoch': 0.17}) +[2026-10-03 13:28:24,762][starVLA.training.train_starvla][INFO] - Step 2270, Loss: {'train/loss': 1.0795502021210268e-05, 'train/grad_norm_pre_clip': 0.003394202096387744, 'timing/data': 0.0002997490082634613, 'timing/model': 0.39658521600358654, 'learning_rate/qwen_vl_interface': 6.098145250293525e-06, 'learning_rate/action_model': 6.0981452502935246e-05, 'epoch': 0.17}) +[2026-10-03 13:28:25,469][starVLA.training.train_starvla][INFO] - Step 2271, Loss: {'train/loss': 0.7299292683601379, 'train/grad_norm_pre_clip': 63.492610931396484, 'timing/data': 0.00023053999757394195, 'timing/model': 0.39446476599550806, 'learning_rate/qwen_vl_interface': 6.09514859590426e-06, 'learning_rate/action_model': 6.09514859590426e-05, 'epoch': 0.17}) +[2026-10-03 13:28:26,176][starVLA.training.train_starvla][INFO] - Step 2272, Loss: {'train/loss': 3.5852775909006596e-05, 'train/grad_norm_pre_clip': 0.04246656596660614, 'timing/data': 0.00020842900266870856, 'timing/model': 0.3953224959986983, 'learning_rate/qwen_vl_interface': 6.09215159410633e-06, 'learning_rate/action_model': 6.09215159410633e-05, 'epoch': 0.17}) +[2026-10-03 13:28:27,078][starVLA.training.train_starvla][INFO] - Step 2273, Loss: {'train/loss': 0.0002025035792030394, 'train/grad_norm_pre_clip': 0.06615373492240906, 'timing/data': 0.0002402300015091896, 'timing/model': 0.5878526040032739, 'learning_rate/qwen_vl_interface': 6.089154246131692e-06, 'learning_rate/action_model': 6.089154246131692e-05, 'epoch': 0.17}) +[2026-10-03 13:28:27,787][starVLA.training.train_starvla][INFO] - Step 2274, Loss: {'train/loss': 0.10214526206254959, 'train/grad_norm_pre_clip': 48.43454360961914, 'timing/data': 0.000254550002864562, 'timing/model': 0.3966884350084001, 'learning_rate/qwen_vl_interface': 6.08615655321244e-06, 'learning_rate/action_model': 6.08615655321244e-05, 'epoch': 0.17}) +[2026-10-03 13:28:28,494][starVLA.training.train_starvla][INFO] - Step 2275, Loss: {'train/loss': 0.00026580371195450425, 'train/grad_norm_pre_clip': 0.05591582879424095, 'timing/data': 0.00028406000637914985, 'timing/model': 0.3947810749959899, 'learning_rate/qwen_vl_interface': 6.083158516580814e-06, 'learning_rate/action_model': 6.083158516580814e-05, 'epoch': 0.17}) +[2026-10-03 13:28:29,201][starVLA.training.train_starvla][INFO] - Step 2276, Loss: {'train/loss': 0.27763092517852783, 'train/grad_norm_pre_clip': 26.858917236328125, 'timing/data': 0.00026234000688418746, 'timing/model': 0.39531271600571927, 'learning_rate/qwen_vl_interface': 6.080160137469192e-06, 'learning_rate/action_model': 6.0801601374691916e-05, 'epoch': 0.17}) +[2026-10-03 13:28:29,911][starVLA.training.train_starvla][INFO] - Step 2277, Loss: {'train/loss': 0.28820160031318665, 'train/grad_norm_pre_clip': 19.096580505371094, 'timing/data': 0.00032441000803373754, 'timing/model': 0.39524688498931937, 'learning_rate/qwen_vl_interface': 6.077161417110097e-06, 'learning_rate/action_model': 6.0771614171100965e-05, 'epoch': 0.17}) +[2026-10-03 13:28:30,813][starVLA.training.train_starvla][INFO] - Step 2278, Loss: {'train/loss': 4.481741052586585e-05, 'train/grad_norm_pre_clip': 0.01787031814455986, 'timing/data': 0.0002597299899207428, 'timing/model': 0.5895234829949914, 'learning_rate/qwen_vl_interface': 6.0741623567361864e-06, 'learning_rate/action_model': 6.074162356736186e-05, 'epoch': 0.17}) +[2026-10-03 13:28:31,520][starVLA.training.train_starvla][INFO] - Step 2279, Loss: {'train/loss': 0.000588757568039, 'train/grad_norm_pre_clip': 0.10705018043518066, 'timing/data': 0.0002042700070887804, 'timing/model': 0.3951044150016969, 'learning_rate/qwen_vl_interface': 6.071162957580262e-06, 'learning_rate/action_model': 6.071162957580262e-05, 'epoch': 0.17}) +[2026-10-03 13:28:32,227][starVLA.training.train_starvla][INFO] - Step 2280, Loss: {'train/loss': 9.749761375132948e-05, 'train/grad_norm_pre_clip': 0.039492640644311905, 'timing/data': 0.00022620000527240336, 'timing/model': 0.39480165600252803, 'learning_rate/qwen_vl_interface': 6.068163220875266e-06, 'learning_rate/action_model': 6.068163220875266e-05, 'epoch': 0.17}) +[2026-10-03 13:28:32,936][starVLA.training.train_starvla][INFO] - Step 2281, Loss: {'train/loss': 0.0002848103758879006, 'train/grad_norm_pre_clip': 0.0803714245557785, 'timing/data': 0.00035139000101480633, 'timing/model': 0.3950083050003741, 'learning_rate/qwen_vl_interface': 6.065163147854273e-06, 'learning_rate/action_model': 6.065163147854273e-05, 'epoch': 0.17}) +[2026-10-03 13:28:33,644][starVLA.training.train_starvla][INFO] - Step 2282, Loss: {'train/loss': 0.003614649176597595, 'train/grad_norm_pre_clip': 0.9416059851646423, 'timing/data': 0.0002036600053543225, 'timing/model': 0.3962685949954903, 'learning_rate/qwen_vl_interface': 6.062162739750502e-06, 'learning_rate/action_model': 6.062162739750502e-05, 'epoch': 0.17}) +[2026-10-03 13:28:34,350][starVLA.training.train_starvla][INFO] - Step 2283, Loss: {'train/loss': 0.0015706561971455812, 'train/grad_norm_pre_clip': 0.34899115562438965, 'timing/data': 0.00020512999617494643, 'timing/model': 0.39423430600436404, 'learning_rate/qwen_vl_interface': 6.059161997797306e-06, 'learning_rate/action_model': 6.0591619977973056e-05, 'epoch': 0.17}) +[2026-10-03 13:28:35,254][starVLA.training.train_starvla][INFO] - Step 2284, Loss: {'train/loss': 0.2692696154117584, 'train/grad_norm_pre_clip': 144.7913818359375, 'timing/data': 0.0002773000014713034, 'timing/model': 0.39494349600863643, 'learning_rate/qwen_vl_interface': 6.056160923228178e-06, 'learning_rate/action_model': 6.056160923228178e-05, 'epoch': 0.17}) +[2026-10-03 13:28:35,963][starVLA.training.train_starvla][INFO] - Step 2285, Loss: {'train/loss': 0.32222121953964233, 'train/grad_norm_pre_clip': 102.45386505126953, 'timing/data': 0.00029302999610081315, 'timing/model': 0.3953664459986612, 'learning_rate/qwen_vl_interface': 6.0531595172767455e-06, 'learning_rate/action_model': 6.0531595172767455e-05, 'epoch': 0.17}) +[2026-10-03 13:28:36,673][starVLA.training.train_starvla][INFO] - Step 2286, Loss: {'train/loss': 0.0002102758880937472, 'train/grad_norm_pre_clip': 0.09715934097766876, 'timing/data': 0.00020898999355267733, 'timing/model': 0.39705866599979345, 'learning_rate/qwen_vl_interface': 6.050157781176774e-06, 'learning_rate/action_model': 6.050157781176774e-05, 'epoch': 0.17}) +[2026-10-03 13:28:37,380][starVLA.training.train_starvla][INFO] - Step 2287, Loss: {'train/loss': 0.3631986081600189, 'train/grad_norm_pre_clip': 38.41615295410156, 'timing/data': 0.00020842900266870856, 'timing/model': 0.39384343700658064, 'learning_rate/qwen_vl_interface': 6.047155716162164e-06, 'learning_rate/action_model': 6.047155716162164e-05, 'epoch': 0.17}) +[2026-10-03 13:28:38,086][starVLA.training.train_starvla][INFO] - Step 2288, Loss: {'train/loss': 0.007837917655706406, 'train/grad_norm_pre_clip': 8.221776008605957, 'timing/data': 0.0002441600081510842, 'timing/model': 0.39431820598838385, 'learning_rate/qwen_vl_interface': 6.044153323466948e-06, 'learning_rate/action_model': 6.044153323466948e-05, 'epoch': 0.17}) +[2026-10-03 13:28:38,982][starVLA.training.train_starvla][INFO] - Step 2289, Loss: {'train/loss': 0.6059836745262146, 'train/grad_norm_pre_clip': 39.304805755615234, 'timing/data': 0.0002952499926323071, 'timing/model': 0.5835987440077588, 'learning_rate/qwen_vl_interface': 6.0411506043253e-06, 'learning_rate/action_model': 6.0411506043253e-05, 'epoch': 0.17}) +[2026-10-03 13:28:39,689][starVLA.training.train_starvla][INFO] - Step 2290, Loss: {'train/loss': 0.000874240358825773, 'train/grad_norm_pre_clip': 0.15895840525627136, 'timing/data': 0.0002478500100551173, 'timing/model': 0.395743995992234, 'learning_rate/qwen_vl_interface': 6.038147559971523e-06, 'learning_rate/action_model': 6.038147559971523e-05, 'epoch': 0.17}) +[2026-10-03 13:28:40,395][starVLA.training.train_starvla][INFO] - Step 2291, Loss: {'train/loss': 0.30534037947654724, 'train/grad_norm_pre_clip': 45.989871978759766, 'timing/data': 0.0002102799917338416, 'timing/model': 0.3936158360011177, 'learning_rate/qwen_vl_interface': 6.035144191640055e-06, 'learning_rate/action_model': 6.035144191640055e-05, 'epoch': 0.17}) +[2026-10-03 13:28:41,100][starVLA.training.train_starvla][INFO] - Step 2292, Loss: {'train/loss': 0.0015432446962222457, 'train/grad_norm_pre_clip': 0.5534924268722534, 'timing/data': 0.00029594999796245247, 'timing/model': 0.39448247601103503, 'learning_rate/qwen_vl_interface': 6.032140500565467e-06, 'learning_rate/action_model': 6.0321405005654665e-05, 'epoch': 0.17}) +[2026-10-03 13:28:41,808][starVLA.training.train_starvla][INFO] - Step 2293, Loss: {'train/loss': 7.229112088680267e-05, 'train/grad_norm_pre_clip': 0.036828216165304184, 'timing/data': 0.00036143000761512667, 'timing/model': 0.39467530499678105, 'learning_rate/qwen_vl_interface': 6.029136487982463e-06, 'learning_rate/action_model': 6.029136487982463e-05, 'epoch': 0.17}) +[2026-10-03 13:28:42,516][starVLA.training.train_starvla][INFO] - Step 2294, Loss: {'train/loss': 0.0017078997334465384, 'train/grad_norm_pre_clip': 0.8127660751342773, 'timing/data': 0.00033149999217130244, 'timing/model': 0.3960602050065063, 'learning_rate/qwen_vl_interface': 6.02613215512588e-06, 'learning_rate/action_model': 6.0261321551258796e-05, 'epoch': 0.17}) +[2026-10-03 13:28:43,412][starVLA.training.train_starvla][INFO] - Step 2295, Loss: {'train/loss': 0.0005017195944674313, 'train/grad_norm_pre_clip': 0.3075543940067291, 'timing/data': 0.000331150004058145, 'timing/model': 0.3940820660063764, 'learning_rate/qwen_vl_interface': 6.023127503230684e-06, 'learning_rate/action_model': 6.023127503230683e-05, 'epoch': 0.17}) +[2026-10-03 13:28:44,118][starVLA.training.train_starvla][INFO] - Step 2296, Loss: {'train/loss': 3.554396971594542e-05, 'train/grad_norm_pre_clip': 0.009645315818488598, 'timing/data': 0.00020854000467807055, 'timing/model': 0.3946871359949, 'learning_rate/qwen_vl_interface': 6.020122533531976e-06, 'learning_rate/action_model': 6.0201225335319755e-05, 'epoch': 0.17}) +[2026-10-03 13:28:44,825][starVLA.training.train_starvla][INFO] - Step 2297, Loss: {'train/loss': 0.0010691529605537653, 'train/grad_norm_pre_clip': 0.6198508143424988, 'timing/data': 0.0003405600000405684, 'timing/model': 0.39461672499601264, 'learning_rate/qwen_vl_interface': 6.017117247264983e-06, 'learning_rate/action_model': 6.017117247264983e-05, 'epoch': 0.17}) +[2026-10-03 13:28:45,534][starVLA.training.train_starvla][INFO] - Step 2298, Loss: {'train/loss': 0.0005746848182752728, 'train/grad_norm_pre_clip': 0.29885008931159973, 'timing/data': 0.0002537300024414435, 'timing/model': 0.39609166599984746, 'learning_rate/qwen_vl_interface': 6.014111645665069e-06, 'learning_rate/action_model': 6.0141116456650684e-05, 'epoch': 0.17}) +[2026-10-03 13:28:46,239][starVLA.training.train_starvla][INFO] - Step 2299, Loss: {'train/loss': 0.32820290327072144, 'train/grad_norm_pre_clip': 126.53621673583984, 'timing/data': 0.0002090699999826029, 'timing/model': 0.39460969500942156, 'learning_rate/qwen_vl_interface': 6.011105729967719e-06, 'learning_rate/action_model': 6.0111057299677184e-05, 'epoch': 0.17}) +[2026-10-03 13:28:47,140][starVLA.training.train_starvla][INFO] - Step 2300, Loss: {'train/loss': 0.012549222446978092, 'train/grad_norm_pre_clip': 20.0839786529541, 'timing/data': 0.00025239000387955457, 'timing/model': 0.3944099560030736, 'learning_rate/qwen_vl_interface': 6.008099501408552e-06, 'learning_rate/action_model': 6.008099501408552e-05, 'epoch': 0.17}) +[2026-10-03 13:28:47,847][starVLA.training.train_starvla][INFO] - Step 2301, Loss: {'train/loss': 0.0043924227356910706, 'train/grad_norm_pre_clip': 5.288192272186279, 'timing/data': 0.00042424000275786966, 'timing/model': 0.39472426498832647, 'learning_rate/qwen_vl_interface': 6.0050929612233165e-06, 'learning_rate/action_model': 6.005092961223316e-05, 'epoch': 0.17}) +[2026-10-03 13:28:48,554][starVLA.training.train_starvla][INFO] - Step 2302, Loss: {'train/loss': 0.3331063687801361, 'train/grad_norm_pre_clip': 54.99674987792969, 'timing/data': 0.0002077999961329624, 'timing/model': 0.39526259500416927, 'learning_rate/qwen_vl_interface': 6.002086110647888e-06, 'learning_rate/action_model': 6.002086110647888e-05, 'epoch': 0.17}) +[2026-10-03 13:28:49,259][starVLA.training.train_starvla][INFO] - Step 2303, Loss: {'train/loss': 0.0010813260450959206, 'train/grad_norm_pre_clip': 0.23865899443626404, 'timing/data': 0.00027030899946112186, 'timing/model': 0.39391867599624675, 'learning_rate/qwen_vl_interface': 5.999078950918267e-06, 'learning_rate/action_model': 5.999078950918267e-05, 'epoch': 0.17}) +[2026-10-03 13:28:49,965][starVLA.training.train_starvla][INFO] - Step 2304, Loss: {'train/loss': 4.098899808013812e-05, 'train/grad_norm_pre_clip': 0.01799529604613781, 'timing/data': 0.00029228000494185835, 'timing/model': 0.39408173599804286, 'learning_rate/qwen_vl_interface': 5.996071483270585e-06, 'learning_rate/action_model': 5.996071483270585e-05, 'epoch': 0.17}) +[2026-10-03 13:28:50,863][starVLA.training.train_starvla][INFO] - Step 2305, Loss: {'train/loss': 0.0002325605455553159, 'train/grad_norm_pre_clip': 0.12184233963489532, 'timing/data': 0.00033340000663883984, 'timing/model': 0.5851689340051962, 'learning_rate/qwen_vl_interface': 5.993063708941098e-06, 'learning_rate/action_model': 5.993063708941098e-05, 'epoch': 0.17}) +[2026-10-03 13:28:51,570][starVLA.training.train_starvla][INFO] - Step 2306, Loss: {'train/loss': 0.0022523975931108, 'train/grad_norm_pre_clip': 3.522324562072754, 'timing/data': 0.00021015999664086848, 'timing/model': 0.39559063500200864, 'learning_rate/qwen_vl_interface': 5.990055629166187e-06, 'learning_rate/action_model': 5.990055629166187e-05, 'epoch': 0.17}) +[2026-10-03 13:28:52,276][starVLA.training.train_starvla][INFO] - Step 2307, Loss: {'train/loss': 9.315981878899038e-05, 'train/grad_norm_pre_clip': 0.04518714174628258, 'timing/data': 0.00020542999845929444, 'timing/model': 0.39443429600214586, 'learning_rate/qwen_vl_interface': 5.987047245182363e-06, 'learning_rate/action_model': 5.987047245182362e-05, 'epoch': 0.17}) +[2026-10-03 13:28:52,983][starVLA.training.train_starvla][INFO] - Step 2308, Loss: {'train/loss': 0.09635882079601288, 'train/grad_norm_pre_clip': 121.68364715576172, 'timing/data': 0.00021047999325674027, 'timing/model': 0.39450222499726806, 'learning_rate/qwen_vl_interface': 5.984038558226255e-06, 'learning_rate/action_model': 5.9840385582262546e-05, 'epoch': 0.17}) +[2026-10-03 13:28:53,690][starVLA.training.train_starvla][INFO] - Step 2309, Loss: {'train/loss': 0.41646137833595276, 'train/grad_norm_pre_clip': 64.3371810913086, 'timing/data': 0.00026277999859303236, 'timing/model': 0.39430721499957144, 'learning_rate/qwen_vl_interface': 5.981029569534625e-06, 'learning_rate/action_model': 5.981029569534625e-05, 'epoch': 0.17}) +[2026-10-03 13:28:54,397][starVLA.training.train_starvla][INFO] - Step 2310, Loss: {'train/loss': 5.059666000306606e-05, 'train/grad_norm_pre_clip': 0.039947062730789185, 'timing/data': 0.00025397000717930496, 'timing/model': 0.39522762600972783, 'learning_rate/qwen_vl_interface': 5.978020280344351e-06, 'learning_rate/action_model': 5.9780202803443506e-05, 'epoch': 0.17}) +[2026-10-03 13:28:55,291][starVLA.training.train_starvla][INFO] - Step 2311, Loss: {'train/loss': 0.734407365322113, 'train/grad_norm_pre_clip': 83.55371856689453, 'timing/data': 0.0003110799880232662, 'timing/model': 0.39403658600349445, 'learning_rate/qwen_vl_interface': 5.97501069189244e-06, 'learning_rate/action_model': 5.97501069189244e-05, 'epoch': 0.17}) +[2026-10-03 13:28:55,997][starVLA.training.train_starvla][INFO] - Step 2312, Loss: {'train/loss': 0.0007351043750531971, 'train/grad_norm_pre_clip': 0.38753795623779297, 'timing/data': 0.0002070600021397695, 'timing/model': 0.3945024559943704, 'learning_rate/qwen_vl_interface': 5.97200080541602e-06, 'learning_rate/action_model': 5.97200080541602e-05, 'epoch': 0.17}) +[2026-10-03 13:28:56,705][starVLA.training.train_starvla][INFO] - Step 2313, Loss: {'train/loss': 0.0009816167876124382, 'train/grad_norm_pre_clip': 2.4253320693969727, 'timing/data': 0.00030937000701669604, 'timing/model': 0.3945006559952162, 'learning_rate/qwen_vl_interface': 5.968990622152341e-06, 'learning_rate/action_model': 5.9689906221523404e-05, 'epoch': 0.17}) +[2026-10-03 13:28:57,413][starVLA.training.train_starvla][INFO] - Step 2314, Loss: {'train/loss': 0.24425509572029114, 'train/grad_norm_pre_clip': 84.196044921875, 'timing/data': 0.00020446999405976385, 'timing/model': 0.39610956600517966, 'learning_rate/qwen_vl_interface': 5.965980143338774e-06, 'learning_rate/action_model': 5.965980143338774e-05, 'epoch': 0.17}) +[2026-10-03 13:28:58,119][starVLA.training.train_starvla][INFO] - Step 2315, Loss: {'train/loss': 1.451297794119455e-05, 'train/grad_norm_pre_clip': 0.006032963749021292, 'timing/data': 0.0002434899943182245, 'timing/model': 0.39371133501117583, 'learning_rate/qwen_vl_interface': 5.962969370212819e-06, 'learning_rate/action_model': 5.9629693702128185e-05, 'epoch': 0.17}) +[2026-10-03 13:28:59,019][starVLA.training.train_starvla][INFO] - Step 2316, Loss: {'train/loss': 3.383585135452449e-05, 'train/grad_norm_pre_clip': 0.02010604552924633, 'timing/data': 0.0003517300065141171, 'timing/model': 0.39461155601020437, 'learning_rate/qwen_vl_interface': 5.959958304012085e-06, 'learning_rate/action_model': 5.9599583040120845e-05, 'epoch': 0.17}) +[2026-10-03 13:28:59,725][starVLA.training.train_starvla][INFO] - Step 2317, Loss: {'train/loss': 0.00010047088289866224, 'train/grad_norm_pre_clip': 0.03300873190164566, 'timing/data': 0.0003462799941189587, 'timing/model': 0.39426204600022174, 'learning_rate/qwen_vl_interface': 5.956946945974309e-06, 'learning_rate/action_model': 5.9569469459743085e-05, 'epoch': 0.17}) +[2026-10-03 13:29:00,432][starVLA.training.train_starvla][INFO] - Step 2318, Loss: {'train/loss': 2.883974593714811e-05, 'train/grad_norm_pre_clip': 0.01019109133630991, 'timing/data': 0.0002420999953756109, 'timing/model': 0.3953533059975598, 'learning_rate/qwen_vl_interface': 5.953935297337347e-06, 'learning_rate/action_model': 5.953935297337347e-05, 'epoch': 0.17}) +[2026-10-03 13:29:01,137][starVLA.training.train_starvla][INFO] - Step 2319, Loss: {'train/loss': 0.0020016669295728207, 'train/grad_norm_pre_clip': 2.410571813583374, 'timing/data': 0.000305370005662553, 'timing/model': 0.3943722860130947, 'learning_rate/qwen_vl_interface': 5.950923359339175e-06, 'learning_rate/action_model': 5.950923359339174e-05, 'epoch': 0.17}) +[2026-10-03 13:29:01,844][starVLA.training.train_starvla][INFO] - Step 2320, Loss: {'train/loss': 5.7708140957402065e-05, 'train/grad_norm_pre_clip': 0.025464797392487526, 'timing/data': 0.00030958998831920326, 'timing/model': 0.39484187599737197, 'learning_rate/qwen_vl_interface': 5.947911133217886e-06, 'learning_rate/action_model': 5.947911133217886e-05, 'epoch': 0.17}) +[2026-10-03 13:29:02,747][starVLA.training.train_starvla][INFO] - Step 2321, Loss: {'train/loss': 0.009068860672414303, 'train/grad_norm_pre_clip': 6.154633045196533, 'timing/data': 0.000215859996387735, 'timing/model': 0.39463810600864235, 'learning_rate/qwen_vl_interface': 5.944898620211693e-06, 'learning_rate/action_model': 5.944898620211693e-05, 'epoch': 0.17}) +[2026-10-03 13:29:03,454][starVLA.training.train_starvla][INFO] - Step 2322, Loss: {'train/loss': 1.2434446034603752e-05, 'train/grad_norm_pre_clip': 0.005109282676130533, 'timing/data': 0.0002513199870008975, 'timing/model': 0.3953089759888826, 'learning_rate/qwen_vl_interface': 5.941885821558925e-06, 'learning_rate/action_model': 5.9418858215589245e-05, 'epoch': 0.17}) +[2026-10-03 13:29:04,160][starVLA.training.train_starvla][INFO] - Step 2323, Loss: {'train/loss': 5.4821386584080756e-05, 'train/grad_norm_pre_clip': 0.023505590856075287, 'timing/data': 0.0002671200054464862, 'timing/model': 0.39395611599320546, 'learning_rate/qwen_vl_interface': 5.938872738498031e-06, 'learning_rate/action_model': 5.93887273849803e-05, 'epoch': 0.17}) +[2026-10-03 13:29:04,866][starVLA.training.train_starvla][INFO] - Step 2324, Loss: {'train/loss': 6.094512173149269e-06, 'train/grad_norm_pre_clip': 0.0011790768476203084, 'timing/data': 0.00032094000198412687, 'timing/model': 0.39454967599886004, 'learning_rate/qwen_vl_interface': 5.935859372267575e-06, 'learning_rate/action_model': 5.935859372267575e-05, 'epoch': 0.17}) +[2026-10-03 13:29:05,574][starVLA.training.train_starvla][INFO] - Step 2325, Loss: {'train/loss': 0.0004046998219564557, 'train/grad_norm_pre_clip': 0.9561139345169067, 'timing/data': 0.000297970007522963, 'timing/model': 0.39453248599602375, 'learning_rate/qwen_vl_interface': 5.9328457241062396e-06, 'learning_rate/action_model': 5.9328457241062394e-05, 'epoch': 0.17}) +[2026-10-03 13:29:06,474][starVLA.training.train_starvla][INFO] - Step 2326, Loss: {'train/loss': 0.0187226515263319, 'train/grad_norm_pre_clip': 28.569616317749023, 'timing/data': 0.0002475299988873303, 'timing/model': 0.5888455940003041, 'learning_rate/qwen_vl_interface': 5.929831795252819e-06, 'learning_rate/action_model': 5.9298317952528184e-05, 'epoch': 0.17}) +[2026-10-03 13:29:07,182][starVLA.training.train_starvla][INFO] - Step 2327, Loss: {'train/loss': 1.862637418525992e-06, 'train/grad_norm_pre_clip': 0.00038778147427365184, 'timing/data': 0.0003065499913645908, 'timing/model': 0.3946962949994486, 'learning_rate/qwen_vl_interface': 5.926817586946227e-06, 'learning_rate/action_model': 5.926817586946227e-05, 'epoch': 0.17}) +[2026-10-03 13:29:07,888][starVLA.training.train_starvla][INFO] - Step 2328, Loss: {'train/loss': 0.002055673860013485, 'train/grad_norm_pre_clip': 3.5579264163970947, 'timing/data': 0.0003619099879870191, 'timing/model': 0.39432178500283044, 'learning_rate/qwen_vl_interface': 5.923803100425491e-06, 'learning_rate/action_model': 5.92380310042549e-05, 'epoch': 0.17}) +[2026-10-03 13:29:08,595][starVLA.training.train_starvla][INFO] - Step 2329, Loss: {'train/loss': 0.21547460556030273, 'train/grad_norm_pre_clip': 64.94847106933594, 'timing/data': 0.0002802300004987046, 'timing/model': 0.3942895759973908, 'learning_rate/qwen_vl_interface': 5.92078833692975e-06, 'learning_rate/action_model': 5.92078833692975e-05, 'epoch': 0.17}) +[2026-10-03 13:29:09,303][starVLA.training.train_starvla][INFO] - Step 2330, Loss: {'train/loss': 0.00020855672482866794, 'train/grad_norm_pre_clip': 0.3866022825241089, 'timing/data': 0.00031936999585013837, 'timing/model': 0.3959142950043315, 'learning_rate/qwen_vl_interface': 5.917773297698264e-06, 'learning_rate/action_model': 5.9177732976982636e-05, 'epoch': 0.17}) +[2026-10-03 13:29:10,199][starVLA.training.train_starvla][INFO] - Step 2331, Loss: {'train/loss': 8.610953227616847e-05, 'train/grad_norm_pre_clip': 0.028709614649415016, 'timing/data': 0.00020456001220736653, 'timing/model': 0.5844371050043264, 'learning_rate/qwen_vl_interface': 5.9147579839704e-06, 'learning_rate/action_model': 5.9147579839704e-05, 'epoch': 0.17}) +[2026-10-03 13:29:10,905][starVLA.training.train_starvla][INFO] - Step 2332, Loss: {'train/loss': 0.058654073625802994, 'train/grad_norm_pre_clip': 265.8257141113281, 'timing/data': 0.0003227400011382997, 'timing/model': 0.3943359160039108, 'learning_rate/qwen_vl_interface': 5.911742396985637e-06, 'learning_rate/action_model': 5.9117423969856364e-05, 'epoch': 0.17}) +[2026-10-03 13:29:11,613][starVLA.training.train_starvla][INFO] - Step 2333, Loss: {'train/loss': 0.0002834312035702169, 'train/grad_norm_pre_clip': 0.189486563205719, 'timing/data': 0.00036068999907001853, 'timing/model': 0.39408674600417726, 'learning_rate/qwen_vl_interface': 5.908726537983571e-06, 'learning_rate/action_model': 5.908726537983571e-05, 'epoch': 0.17}) +[2026-10-03 13:29:12,321][starVLA.training.train_starvla][INFO] - Step 2334, Loss: {'train/loss': 0.995122492313385, 'train/grad_norm_pre_clip': 129.46640014648438, 'timing/data': 0.0002558100095484406, 'timing/model': 0.3959928849944845, 'learning_rate/qwen_vl_interface': 5.9057104082039085e-06, 'learning_rate/action_model': 5.905710408203908e-05, 'epoch': 0.17}) +[2026-10-03 13:29:13,026][starVLA.training.train_starvla][INFO] - Step 2335, Loss: {'train/loss': 0.0030380436219274998, 'train/grad_norm_pre_clip': 1.6393336057662964, 'timing/data': 0.0003055400011362508, 'timing/model': 0.39416678600537125, 'learning_rate/qwen_vl_interface': 5.902694008886466e-06, 'learning_rate/action_model': 5.902694008886466e-05, 'epoch': 0.17}) +[2026-10-03 13:29:13,732][starVLA.training.train_starvla][INFO] - Step 2336, Loss: {'train/loss': 0.0008255925495177507, 'train/grad_norm_pre_clip': 0.5848257541656494, 'timing/data': 0.00032731000101193786, 'timing/model': 0.39434410599642433, 'learning_rate/qwen_vl_interface': 5.8996773412711694e-06, 'learning_rate/action_model': 5.899677341271169e-05, 'epoch': 0.17}) +[2026-10-03 13:29:14,634][starVLA.training.train_starvla][INFO] - Step 2337, Loss: {'train/loss': 0.3574138879776001, 'train/grad_norm_pre_clip': 53.90187454223633, 'timing/data': 0.00033275000168941915, 'timing/model': 0.3939249050017679, 'learning_rate/qwen_vl_interface': 5.896660406598059e-06, 'learning_rate/action_model': 5.896660406598059e-05, 'epoch': 0.17}) +[2026-10-03 13:29:15,341][starVLA.training.train_starvla][INFO] - Step 2338, Loss: {'train/loss': 7.495114914490841e-06, 'train/grad_norm_pre_clip': 0.003319463925436139, 'timing/data': 0.0003131199919153005, 'timing/model': 0.39550865499768406, 'learning_rate/qwen_vl_interface': 5.893643206107283e-06, 'learning_rate/action_model': 5.893643206107282e-05, 'epoch': 0.17}) +[2026-10-03 13:29:16,047][starVLA.training.train_starvla][INFO] - Step 2339, Loss: {'train/loss': 0.5654422044754028, 'train/grad_norm_pre_clip': 24.38084602355957, 'timing/data': 0.00020410999422892928, 'timing/model': 0.3944994960038457, 'learning_rate/qwen_vl_interface': 5.890625741039096e-06, 'learning_rate/action_model': 5.890625741039095e-05, 'epoch': 0.17}) +[2026-10-03 13:29:16,754][starVLA.training.train_starvla][INFO] - Step 2340, Loss: {'train/loss': 0.0008408462745137513, 'train/grad_norm_pre_clip': 1.2200818061828613, 'timing/data': 0.0003132999991066754, 'timing/model': 0.3952042049932061, 'learning_rate/qwen_vl_interface': 5.887608012633864e-06, 'learning_rate/action_model': 5.887608012633864e-05, 'epoch': 0.17}) +[2026-10-03 13:29:17,461][starVLA.training.train_starvla][INFO] - Step 2341, Loss: {'train/loss': 3.016282971657347e-05, 'train/grad_norm_pre_clip': 0.019500277936458588, 'timing/data': 0.00035130999458488077, 'timing/model': 0.39402969600632787, 'learning_rate/qwen_vl_interface': 5.884590022132064e-06, 'learning_rate/action_model': 5.884590022132064e-05, 'epoch': 0.17}) +[2026-10-03 13:29:18,364][starVLA.training.train_starvla][INFO] - Step 2342, Loss: {'train/loss': 0.0027706834953278303, 'train/grad_norm_pre_clip': 6.474810600280762, 'timing/data': 0.00028917999588884413, 'timing/model': 0.5900943340093363, 'learning_rate/qwen_vl_interface': 5.881571770774278e-06, 'learning_rate/action_model': 5.881571770774278e-05, 'epoch': 0.17}) +[2026-10-03 13:29:19,069][starVLA.training.train_starvla][INFO] - Step 2343, Loss: {'train/loss': 0.0006394495721906424, 'train/grad_norm_pre_clip': 0.44097328186035156, 'timing/data': 0.0002439700037939474, 'timing/model': 0.3940920260065468, 'learning_rate/qwen_vl_interface': 5.878553259801192e-06, 'learning_rate/action_model': 5.878553259801192e-05, 'epoch': 0.17}) +[2026-10-03 13:29:19,775][starVLA.training.train_starvla][INFO] - Step 2344, Loss: {'train/loss': 0.30531275272369385, 'train/grad_norm_pre_clip': 135.51641845703125, 'timing/data': 0.00031231000320985913, 'timing/model': 0.3943617559998529, 'learning_rate/qwen_vl_interface': 5.8755344904536025e-06, 'learning_rate/action_model': 5.8755344904536025e-05, 'epoch': 0.17}) +[2026-10-03 13:29:20,482][starVLA.training.train_starvla][INFO] - Step 2345, Loss: {'train/loss': 0.00014397597988136113, 'train/grad_norm_pre_clip': 0.0792950987815857, 'timing/data': 0.0003497599973343313, 'timing/model': 0.3947988160070963, 'learning_rate/qwen_vl_interface': 5.872515463972415e-06, 'learning_rate/action_model': 5.8725154639724146e-05, 'epoch': 0.17}) +[2026-10-03 13:29:21,189][starVLA.training.train_starvla][INFO] - Step 2346, Loss: {'train/loss': 0.000110406763269566, 'train/grad_norm_pre_clip': 0.07525637000799179, 'timing/data': 0.00023474000045098364, 'timing/model': 0.39522370501072146, 'learning_rate/qwen_vl_interface': 5.869496181598633e-06, 'learning_rate/action_model': 5.869496181598633e-05, 'epoch': 0.17}) +[2026-10-03 13:29:21,895][starVLA.training.train_starvla][INFO] - Step 2347, Loss: {'train/loss': 0.00012046733900206164, 'train/grad_norm_pre_clip': 0.045617230236530304, 'timing/data': 0.000252459998591803, 'timing/model': 0.39390618600009475, 'learning_rate/qwen_vl_interface': 5.866476644573371e-06, 'learning_rate/action_model': 5.866476644573371e-05, 'epoch': 0.17}) +[2026-10-03 13:29:22,794][starVLA.training.train_starvla][INFO] - Step 2348, Loss: {'train/loss': 1.8625556549523026e-05, 'train/grad_norm_pre_clip': 0.005747123621404171, 'timing/data': 0.00030066999897826463, 'timing/model': 0.3944342649920145, 'learning_rate/qwen_vl_interface': 5.863456854137848e-06, 'learning_rate/action_model': 5.8634568541378474e-05, 'epoch': 0.17}) +[2026-10-03 13:29:23,501][starVLA.training.train_starvla][INFO] - Step 2349, Loss: {'train/loss': 0.0004433294234331697, 'train/grad_norm_pre_clip': 1.2428226470947266, 'timing/data': 0.0002683800121303648, 'timing/model': 0.39416865599923767, 'learning_rate/qwen_vl_interface': 5.860436811533383e-06, 'learning_rate/action_model': 5.860436811533383e-05, 'epoch': 0.17}) +[2026-10-03 13:29:24,207][starVLA.training.train_starvla][INFO] - Step 2350, Loss: {'train/loss': 0.41756460070610046, 'train/grad_norm_pre_clip': 46.40022277832031, 'timing/data': 0.00029463000828400254, 'timing/model': 0.3952559250028571, 'learning_rate/qwen_vl_interface': 5.857416518001404e-06, 'learning_rate/action_model': 5.857416518001404e-05, 'epoch': 0.17}) +[2026-10-03 13:29:24,913][starVLA.training.train_starvla][INFO] - Step 2351, Loss: {'train/loss': 2.2274132788879797e-05, 'train/grad_norm_pre_clip': 0.007265444844961166, 'timing/data': 0.00032650999492034316, 'timing/model': 0.39401406499382574, 'learning_rate/qwen_vl_interface': 5.854395974783439e-06, 'learning_rate/action_model': 5.8543959747834386e-05, 'epoch': 0.17}) +[2026-10-03 13:29:25,620][starVLA.training.train_starvla][INFO] - Step 2352, Loss: {'train/loss': 5.3631669288733974e-05, 'train/grad_norm_pre_clip': 0.012613820843398571, 'timing/data': 0.0003040300071006641, 'timing/model': 0.39555600500898436, 'learning_rate/qwen_vl_interface': 5.851375183121119e-06, 'learning_rate/action_model': 5.8513751831211186e-05, 'epoch': 0.17}) +[2026-10-03 13:29:26,524][starVLA.training.train_starvla][INFO] - Step 2353, Loss: {'train/loss': 0.4875929057598114, 'train/grad_norm_pre_clip': 32.52286148071289, 'timing/data': 0.0003408899938222021, 'timing/model': 0.394130036002025, 'learning_rate/qwen_vl_interface': 5.848354144256175e-06, 'learning_rate/action_model': 5.8483541442561754e-05, 'epoch': 0.17}) +[2026-10-03 13:29:27,231][starVLA.training.train_starvla][INFO] - Step 2354, Loss: {'train/loss': 5.774062174168648e-06, 'train/grad_norm_pre_clip': 0.0030063423328101635, 'timing/data': 0.0003068999940296635, 'timing/model': 0.39570732599531766, 'learning_rate/qwen_vl_interface': 5.845332859430446e-06, 'learning_rate/action_model': 5.845332859430446e-05, 'epoch': 0.17}) +[2026-10-03 13:29:27,937][starVLA.training.train_starvla][INFO] - Step 2355, Loss: {'train/loss': 0.00024531775852665305, 'train/grad_norm_pre_clip': 0.05013422295451164, 'timing/data': 0.000225479991058819, 'timing/model': 0.3939815159974387, 'learning_rate/qwen_vl_interface': 5.842311329885866e-06, 'learning_rate/action_model': 5.842311329885866e-05, 'epoch': 0.17}) +[2026-10-03 13:29:28,643][starVLA.training.train_starvla][INFO] - Step 2356, Loss: {'train/loss': 0.0051978337578475475, 'train/grad_norm_pre_clip': 3.525519371032715, 'timing/data': 0.0003027399943675846, 'timing/model': 0.39514187499298714, 'learning_rate/qwen_vl_interface': 5.839289556864472e-06, 'learning_rate/action_model': 5.8392895568644714e-05, 'epoch': 0.17}) +[2026-10-03 13:29:29,350][starVLA.training.train_starvla][INFO] - Step 2357, Loss: {'train/loss': 2.9290036763995886e-05, 'train/grad_norm_pre_clip': 0.006640711333602667, 'timing/data': 0.0003156299935653806, 'timing/model': 0.39463265499216504, 'learning_rate/qwen_vl_interface': 5.836267541608398e-06, 'learning_rate/action_model': 5.836267541608398e-05, 'epoch': 0.17}) +[2026-10-03 13:29:30,251][starVLA.training.train_starvla][INFO] - Step 2358, Loss: {'train/loss': 0.007950603030622005, 'train/grad_norm_pre_clip': 5.302975654602051, 'timing/data': 0.000338099998771213, 'timing/model': 0.5896457339986227, 'learning_rate/qwen_vl_interface': 5.833245285359887e-06, 'learning_rate/action_model': 5.833245285359886e-05, 'epoch': 0.17}) +[2026-10-03 13:29:30,957][starVLA.training.train_starvla][INFO] - Step 2359, Loss: {'train/loss': 0.0018938672728836536, 'train/grad_norm_pre_clip': 1.3404114246368408, 'timing/data': 0.0003055400011362508, 'timing/model': 0.3941411359992344, 'learning_rate/qwen_vl_interface': 5.830222789361268e-06, 'learning_rate/action_model': 5.8302227893612684e-05, 'epoch': 0.17}) +[2026-10-03 13:29:31,665][starVLA.training.train_starvla][INFO] - Step 2360, Loss: {'train/loss': 2.4605285943835042e-05, 'train/grad_norm_pre_clip': 0.005876774434000254, 'timing/data': 0.00029802000790368766, 'timing/model': 0.39591977499367204, 'learning_rate/qwen_vl_interface': 5.8272000548549776e-06, 'learning_rate/action_model': 5.827200054854978e-05, 'epoch': 0.17}) +[2026-10-03 13:29:32,372][starVLA.training.train_starvla][INFO] - Step 2361, Loss: {'train/loss': 8.418921424890868e-06, 'train/grad_norm_pre_clip': 0.0026675143744796515, 'timing/data': 0.0002738799958024174, 'timing/model': 0.3943491559912218, 'learning_rate/qwen_vl_interface': 5.824177083083547e-06, 'learning_rate/action_model': 5.8241770830835475e-05, 'epoch': 0.17}) +[2026-10-03 13:29:33,080][starVLA.training.train_starvla][INFO] - Step 2362, Loss: {'train/loss': 0.31469616293907166, 'train/grad_norm_pre_clip': 28.430221557617188, 'timing/data': 0.0003329600003780797, 'timing/model': 0.39571386500028893, 'learning_rate/qwen_vl_interface': 5.8211538752896056e-06, 'learning_rate/action_model': 5.821153875289606e-05, 'epoch': 0.17}) +[2026-10-03 13:29:33,787][starVLA.training.train_starvla][INFO] - Step 2363, Loss: {'train/loss': 0.19271418452262878, 'train/grad_norm_pre_clip': 218.56210327148438, 'timing/data': 0.00032423999800812453, 'timing/model': 0.39393269599531777, 'learning_rate/qwen_vl_interface': 5.8181304327158814e-06, 'learning_rate/action_model': 5.8181304327158814e-05, 'epoch': 0.17}) +[2026-10-03 13:29:34,685][starVLA.training.train_starvla][INFO] - Step 2364, Loss: {'train/loss': 5.495543882716447e-05, 'train/grad_norm_pre_clip': 0.013798605650663376, 'timing/data': 0.00026794000586960465, 'timing/model': 0.3952317649964243, 'learning_rate/qwen_vl_interface': 5.8151067566051965e-06, 'learning_rate/action_model': 5.815106756605196e-05, 'epoch': 0.17}) +[2026-10-03 13:29:35,392][starVLA.training.train_starvla][INFO] - Step 2365, Loss: {'train/loss': 0.37108781933784485, 'train/grad_norm_pre_clip': 31.861007690429688, 'timing/data': 0.00026280000747647136, 'timing/model': 0.3942985059984494, 'learning_rate/qwen_vl_interface': 5.81208284820047e-06, 'learning_rate/action_model': 5.81208284820047e-05, 'epoch': 0.17}) +[2026-10-03 13:29:36,099][starVLA.training.train_starvla][INFO] - Step 2366, Loss: {'train/loss': 2.391526140854694e-05, 'train/grad_norm_pre_clip': 0.011596724390983582, 'timing/data': 0.00027253999724052846, 'timing/model': 0.395788096007891, 'learning_rate/qwen_vl_interface': 5.809058708744716e-06, 'learning_rate/action_model': 5.809058708744716e-05, 'epoch': 0.17}) +[2026-10-03 13:29:36,804][starVLA.training.train_starvla][INFO] - Step 2367, Loss: {'train/loss': 0.000605221139267087, 'train/grad_norm_pre_clip': 0.3215336799621582, 'timing/data': 0.0002805989934131503, 'timing/model': 0.3940211669978453, 'learning_rate/qwen_vl_interface': 5.8060343394810445e-06, 'learning_rate/action_model': 5.806034339481045e-05, 'epoch': 0.17}) +[2026-10-03 13:29:37,513][starVLA.training.train_starvla][INFO] - Step 2368, Loss: {'train/loss': 0.14068417251110077, 'train/grad_norm_pre_clip': 145.6468505859375, 'timing/data': 0.00021732000459451228, 'timing/model': 0.3958598559984239, 'learning_rate/qwen_vl_interface': 5.803009741652659e-06, 'learning_rate/action_model': 5.8030097416526585e-05, 'epoch': 0.17}) +[2026-10-03 13:29:38,412][starVLA.training.train_starvla][INFO] - Step 2369, Loss: {'train/loss': 6.386423774529248e-05, 'train/grad_norm_pre_clip': 0.015169537626206875, 'timing/data': 0.00027470999339129776, 'timing/model': 0.3941520660009701, 'learning_rate/qwen_vl_interface': 5.799984916502859e-06, 'learning_rate/action_model': 5.799984916502859e-05, 'epoch': 0.17}) +[2026-10-03 13:29:39,121][starVLA.training.train_starvla][INFO] - Step 2370, Loss: {'train/loss': 0.029425807297229767, 'train/grad_norm_pre_clip': 81.47953796386719, 'timing/data': 0.00027458000113256276, 'timing/model': 0.3959028549870709, 'learning_rate/qwen_vl_interface': 5.7969598652750334e-06, 'learning_rate/action_model': 5.796959865275033e-05, 'epoch': 0.17}) +[2026-10-03 13:29:39,827][starVLA.training.train_starvla][INFO] - Step 2371, Loss: {'train/loss': 0.11206021904945374, 'train/grad_norm_pre_clip': 66.9619369506836, 'timing/data': 0.0003105000068899244, 'timing/model': 0.39425760599260684, 'learning_rate/qwen_vl_interface': 5.793934589212666e-06, 'learning_rate/action_model': 5.793934589212666e-05, 'epoch': 0.17}) +[2026-10-03 13:29:40,534][starVLA.training.train_starvla][INFO] - Step 2372, Loss: {'train/loss': 0.912380039691925, 'train/grad_norm_pre_clip': 27.03040313720703, 'timing/data': 0.00032605999149382114, 'timing/model': 0.39430067599460017, 'learning_rate/qwen_vl_interface': 5.7909090895593375e-06, 'learning_rate/action_model': 5.790909089559338e-05, 'epoch': 0.17}) +[2026-10-03 13:29:41,242][starVLA.training.train_starvla][INFO] - Step 2373, Loss: {'train/loss': 0.008239670656621456, 'train/grad_norm_pre_clip': 32.89150619506836, 'timing/data': 0.0002658599987626076, 'timing/model': 0.3945750849961769, 'learning_rate/qwen_vl_interface': 5.7878833675587135e-06, 'learning_rate/action_model': 5.787883367558713e-05, 'epoch': 0.17}) +[2026-10-03 13:29:42,144][starVLA.training.train_starvla][INFO] - Step 2374, Loss: {'train/loss': 0.0004481622891034931, 'train/grad_norm_pre_clip': 0.11948135495185852, 'timing/data': 0.0002617499994812533, 'timing/model': 0.3954418649955187, 'learning_rate/qwen_vl_interface': 5.784857424454556e-06, 'learning_rate/action_model': 5.7848574244545563e-05, 'epoch': 0.17}) +[2026-10-03 13:29:42,850][starVLA.training.train_starvla][INFO] - Step 2375, Loss: {'train/loss': 0.00013254286022856832, 'train/grad_norm_pre_clip': 0.035990696400403976, 'timing/data': 0.00020836001203861088, 'timing/model': 0.3944873159925919, 'learning_rate/qwen_vl_interface': 5.7818312614907125e-06, 'learning_rate/action_model': 5.781831261490712e-05, 'epoch': 0.17}) +[2026-10-03 13:29:43,559][starVLA.training.train_starvla][INFO] - Step 2376, Loss: {'train/loss': 0.0006150726694613695, 'train/grad_norm_pre_clip': 0.2778273820877075, 'timing/data': 0.00026228999195154756, 'timing/model': 0.39638250500138383, 'learning_rate/qwen_vl_interface': 5.778804879911129e-06, 'learning_rate/action_model': 5.778804879911128e-05, 'epoch': 0.17}) +[2026-10-03 13:29:44,266][starVLA.training.train_starvla][INFO] - Step 2377, Loss: {'train/loss': 0.00040007702773436904, 'train/grad_norm_pre_clip': 0.14044597744941711, 'timing/data': 0.0003387400065548718, 'timing/model': 0.3945106959872646, 'learning_rate/qwen_vl_interface': 5.775778280959832e-06, 'learning_rate/action_model': 5.7757782809598315e-05, 'epoch': 0.17}) +[2026-10-03 13:29:44,974][starVLA.training.train_starvla][INFO] - Step 2378, Loss: {'train/loss': 0.016563061624765396, 'train/grad_norm_pre_clip': 31.657899856567383, 'timing/data': 0.0002978900010930374, 'timing/model': 0.395373995997943, 'learning_rate/qwen_vl_interface': 5.7727514658809435e-06, 'learning_rate/action_model': 5.772751465880944e-05, 'epoch': 0.17}) +[2026-10-03 13:29:45,873][starVLA.training.train_starvla][INFO] - Step 2379, Loss: {'train/loss': 0.00035450421273708344, 'train/grad_norm_pre_clip': 0.11056862771511078, 'timing/data': 0.000274819991318509, 'timing/model': 0.5878327550017275, 'learning_rate/qwen_vl_interface': 5.769724435918676e-06, 'learning_rate/action_model': 5.7697244359186755e-05, 'epoch': 0.17}) +[2026-10-03 13:29:46,582][starVLA.training.train_starvla][INFO] - Step 2380, Loss: {'train/loss': 0.0007852771668694913, 'train/grad_norm_pre_clip': 0.32781824469566345, 'timing/data': 0.00038904999382793903, 'timing/model': 0.3964650350098964, 'learning_rate/qwen_vl_interface': 5.766697192317323e-06, 'learning_rate/action_model': 5.766697192317323e-05, 'epoch': 0.17}) +[2026-10-03 13:29:47,290][starVLA.training.train_starvla][INFO] - Step 2381, Loss: {'train/loss': 0.00016271669301204383, 'train/grad_norm_pre_clip': 0.06627733260393143, 'timing/data': 0.0003368600009707734, 'timing/model': 0.3943680049997056, 'learning_rate/qwen_vl_interface': 5.763669736321271e-06, 'learning_rate/action_model': 5.763669736321271e-05, 'epoch': 0.17}) +[2026-10-03 13:29:47,998][starVLA.training.train_starvla][INFO] - Step 2382, Loss: {'train/loss': 0.0004948053392581642, 'train/grad_norm_pre_clip': 0.17059089243412018, 'timing/data': 0.00024942999880295247, 'timing/model': 0.3955058050050866, 'learning_rate/qwen_vl_interface': 5.760642069174994e-06, 'learning_rate/action_model': 5.760642069174994e-05, 'epoch': 0.17}) +[2026-10-03 13:29:48,703][starVLA.training.train_starvla][INFO] - Step 2383, Loss: {'train/loss': 0.0003409793716855347, 'train/grad_norm_pre_clip': 0.12078351527452469, 'timing/data': 0.00025097999605350196, 'timing/model': 0.39397486600500997, 'learning_rate/qwen_vl_interface': 5.757614192123051e-06, 'learning_rate/action_model': 5.757614192123051e-05, 'epoch': 0.17}) +[2026-10-03 13:29:49,608][starVLA.training.train_starvla][INFO] - Step 2384, Loss: {'train/loss': 0.01344558596611023, 'train/grad_norm_pre_clip': 6.333672046661377, 'timing/data': 0.00031040998874232173, 'timing/model': 0.5918805230030557, 'learning_rate/qwen_vl_interface': 5.754586106410089e-06, 'learning_rate/action_model': 5.754586106410088e-05, 'epoch': 0.17}) +[2026-10-03 13:29:50,319][starVLA.training.train_starvla][INFO] - Step 2385, Loss: {'train/loss': 0.8051385879516602, 'train/grad_norm_pre_clip': 41.877410888671875, 'timing/data': 0.000268449992290698, 'timing/model': 0.39424729500024114, 'learning_rate/qwen_vl_interface': 5.751557813280838e-06, 'learning_rate/action_model': 5.751557813280838e-05, 'epoch': 0.17}) +[2026-10-03 13:29:51,026][starVLA.training.train_starvla][INFO] - Step 2386, Loss: {'train/loss': 0.04105418175458908, 'train/grad_norm_pre_clip': 75.6963119506836, 'timing/data': 0.0002101700083585456, 'timing/model': 0.3953809250087943, 'learning_rate/qwen_vl_interface': 5.7485293139801146e-06, 'learning_rate/action_model': 5.7485293139801146e-05, 'epoch': 0.17}) +[2026-10-03 13:29:51,732][starVLA.training.train_starvla][INFO] - Step 2387, Loss: {'train/loss': 0.00010409468086436391, 'train/grad_norm_pre_clip': 0.03978850692510605, 'timing/data': 0.0002048800088232383, 'timing/model': 0.39441326599626336, 'learning_rate/qwen_vl_interface': 5.745500609752821e-06, 'learning_rate/action_model': 5.745500609752821e-05, 'epoch': 0.17}) +[2026-10-03 13:29:52,441][starVLA.training.train_starvla][INFO] - Step 2388, Loss: {'train/loss': 3.476733036222868e-05, 'train/grad_norm_pre_clip': 0.010075167752802372, 'timing/data': 0.00038394999864976853, 'timing/model': 0.3952251759910723, 'learning_rate/qwen_vl_interface': 5.742471701843943e-06, 'learning_rate/action_model': 5.742471701843943e-05, 'epoch': 0.17}) +[2026-10-03 13:29:53,148][starVLA.training.train_starvla][INFO] - Step 2389, Loss: {'train/loss': 0.43035224080085754, 'train/grad_norm_pre_clip': 140.86317443847656, 'timing/data': 0.00030151999089866877, 'timing/model': 0.39462430599087384, 'learning_rate/qwen_vl_interface': 5.739442591498553e-06, 'learning_rate/action_model': 5.739442591498553e-05, 'epoch': 0.17}) +[2026-10-03 13:29:54,050][starVLA.training.train_starvla][INFO] - Step 2390, Loss: {'train/loss': 0.0002621267340146005, 'train/grad_norm_pre_clip': 0.4560375213623047, 'timing/data': 0.00021563000336755067, 'timing/model': 0.3956065149977803, 'learning_rate/qwen_vl_interface': 5.7364132799617985e-06, 'learning_rate/action_model': 5.7364132799617984e-05, 'epoch': 0.17}) +[2026-10-03 13:29:54,755][starVLA.training.train_starvla][INFO] - Step 2391, Loss: {'train/loss': 0.004870562348514795, 'train/grad_norm_pre_clip': 10.711790084838867, 'timing/data': 0.00028137999470345676, 'timing/model': 0.3938430959969992, 'learning_rate/qwen_vl_interface': 5.733383768478923e-06, 'learning_rate/action_model': 5.7333837684789225e-05, 'epoch': 0.18}) +[2026-10-03 13:29:55,463][starVLA.training.train_starvla][INFO] - Step 2392, Loss: {'train/loss': 0.7212172746658325, 'train/grad_norm_pre_clip': 50.12214660644531, 'timing/data': 0.00025980999635066837, 'timing/model': 0.395377295993967, 'learning_rate/qwen_vl_interface': 5.730354058295234e-06, 'learning_rate/action_model': 5.730354058295234e-05, 'epoch': 0.18}) +[2026-10-03 13:29:56,168][starVLA.training.train_starvla][INFO] - Step 2393, Loss: {'train/loss': 9.187609248328954e-05, 'train/grad_norm_pre_clip': 0.030937381088733673, 'timing/data': 0.0003498899895930663, 'timing/model': 0.3930842069967184, 'learning_rate/qwen_vl_interface': 5.72732415065614e-06, 'learning_rate/action_model': 5.727324150656139e-05, 'epoch': 0.18}) +[2026-10-03 13:29:56,871][starVLA.training.train_starvla][INFO] - Step 2394, Loss: {'train/loss': 0.3090255856513977, 'train/grad_norm_pre_clip': 33.49295425415039, 'timing/data': 0.00020480000239331275, 'timing/model': 0.39370416599558666, 'learning_rate/qwen_vl_interface': 5.724294046807115e-06, 'learning_rate/action_model': 5.724294046807115e-05, 'epoch': 0.18}) +[2026-10-03 13:29:57,761][starVLA.training.train_starvla][INFO] - Step 2395, Loss: {'train/loss': 2.3080450773704797e-05, 'train/grad_norm_pre_clip': 0.011515629477798939, 'timing/data': 0.0002931000053649768, 'timing/model': 0.58015015500132, 'learning_rate/qwen_vl_interface': 5.7212637479937234e-06, 'learning_rate/action_model': 5.721263747993724e-05, 'epoch': 0.18}) +[2026-10-03 13:29:58,466][starVLA.training.train_starvla][INFO] - Step 2396, Loss: {'train/loss': 0.00041364418575540185, 'train/grad_norm_pre_clip': 0.9457152485847473, 'timing/data': 0.0002633499971125275, 'timing/model': 0.39375196599576157, 'learning_rate/qwen_vl_interface': 5.718233255461607e-06, 'learning_rate/action_model': 5.718233255461607e-05, 'epoch': 0.18}) +[2026-10-03 13:29:59,169][starVLA.training.train_starvla][INFO] - Step 2397, Loss: {'train/loss': 3.974287392338738e-05, 'train/grad_norm_pre_clip': 0.014350948855280876, 'timing/data': 0.00030967000930104405, 'timing/model': 0.3924893160001375, 'learning_rate/qwen_vl_interface': 5.715202570456486e-06, 'learning_rate/action_model': 5.715202570456486e-05, 'epoch': 0.18}) +[2026-10-03 13:29:59,873][starVLA.training.train_starvla][INFO] - Step 2398, Loss: {'train/loss': 0.6137251853942871, 'train/grad_norm_pre_clip': 34.490570068359375, 'timing/data': 0.00020911000319756567, 'timing/model': 0.3937363059958443, 'learning_rate/qwen_vl_interface': 5.7121716942241615e-06, 'learning_rate/action_model': 5.712171694224161e-05, 'epoch': 0.18}) +[2026-10-03 13:30:00,575][starVLA.training.train_starvla][INFO] - Step 2399, Loss: {'train/loss': 9.351945482194424e-05, 'train/grad_norm_pre_clip': 0.02326989360153675, 'timing/data': 0.00020527000015135854, 'timing/model': 0.39228796600946225, 'learning_rate/qwen_vl_interface': 5.709140628010509e-06, 'learning_rate/action_model': 5.7091406280105085e-05, 'epoch': 0.18}) +[2026-10-03 13:30:01,280][starVLA.training.train_starvla][INFO] - Step 2400, Loss: {'train/loss': 0.00031553354347124696, 'train/grad_norm_pre_clip': 0.28916171193122864, 'timing/data': 0.0003245800035074353, 'timing/model': 0.3939455759973498, 'learning_rate/qwen_vl_interface': 5.706109373061489e-06, 'learning_rate/action_model': 5.706109373061489e-05, 'epoch': 0.18}) +[2026-10-03 13:30:02,173][starVLA.training.train_starvla][INFO] - Step 2401, Loss: {'train/loss': 0.00016934046288952231, 'train/grad_norm_pre_clip': 0.03958833962678909, 'timing/data': 0.00029865000396966934, 'timing/model': 0.39208191601210274, 'learning_rate/qwen_vl_interface': 5.703077930623134e-06, 'learning_rate/action_model': 5.7030779306231344e-05, 'epoch': 0.18}) +[2026-10-03 13:30:02,877][starVLA.training.train_starvla][INFO] - Step 2402, Loss: {'train/loss': 0.6309820413589478, 'train/grad_norm_pre_clip': 33.326866149902344, 'timing/data': 0.0003083600022364408, 'timing/model': 0.3934629460127326, 'learning_rate/qwen_vl_interface': 5.7000463019415565e-06, 'learning_rate/action_model': 5.7000463019415563e-05, 'epoch': 0.18}) +[2026-10-03 13:30:03,579][starVLA.training.train_starvla][INFO] - Step 2403, Loss: {'train/loss': 1.6137553757289425e-05, 'train/grad_norm_pre_clip': 0.0025372712407261133, 'timing/data': 0.00029379999614320695, 'timing/model': 0.39211827599501703, 'learning_rate/qwen_vl_interface': 5.697014488262945e-06, 'learning_rate/action_model': 5.697014488262945e-05, 'epoch': 0.18}) +[2026-10-03 13:30:04,285][starVLA.training.train_starvla][INFO] - Step 2404, Loss: {'train/loss': 3.748903327505104e-05, 'train/grad_norm_pre_clip': 0.0096668666228652, 'timing/data': 0.0003265699924668297, 'timing/model': 0.3941665359889157, 'learning_rate/qwen_vl_interface': 5.693982490833561e-06, 'learning_rate/action_model': 5.693982490833561e-05, 'epoch': 0.18}) +[2026-10-03 13:30:04,989][starVLA.training.train_starvla][INFO] - Step 2405, Loss: {'train/loss': 0.00020576591487042606, 'train/grad_norm_pre_clip': 0.10408459603786469, 'timing/data': 0.00035625998862087727, 'timing/model': 0.3930933060037205, 'learning_rate/qwen_vl_interface': 5.690950310899745e-06, 'learning_rate/action_model': 5.6909503108997456e-05, 'epoch': 0.18}) +[2026-10-03 13:30:05,886][starVLA.training.train_starvla][INFO] - Step 2406, Loss: {'train/loss': 0.0007963438401930034, 'train/grad_norm_pre_clip': 0.5570516586303711, 'timing/data': 0.00031115999445319176, 'timing/model': 0.393700735992752, 'learning_rate/qwen_vl_interface': 5.687917949707913e-06, 'learning_rate/action_model': 5.687917949707913e-05, 'epoch': 0.18}) +[2026-10-03 13:30:06,588][starVLA.training.train_starvla][INFO] - Step 2407, Loss: {'train/loss': 0.17264051735401154, 'train/grad_norm_pre_clip': 31.625343322753906, 'timing/data': 0.00021313999604899436, 'timing/model': 0.3927051849896088, 'learning_rate/qwen_vl_interface': 5.684885408504555e-06, 'learning_rate/action_model': 5.684885408504554e-05, 'epoch': 0.18}) +[2026-10-03 13:30:07,294][starVLA.training.train_starvla][INFO] - Step 2408, Loss: {'train/loss': 0.019002068787813187, 'train/grad_norm_pre_clip': 18.81060028076172, 'timing/data': 0.0002976399991894141, 'timing/model': 0.39380550499481615, 'learning_rate/qwen_vl_interface': 5.681852688536231e-06, 'learning_rate/action_model': 5.681852688536231e-05, 'epoch': 0.18}) +[2026-10-03 13:30:08,000][starVLA.training.train_starvla][INFO] - Step 2409, Loss: {'train/loss': 0.4856719970703125, 'train/grad_norm_pre_clip': 112.13182067871094, 'timing/data': 0.0003081600007135421, 'timing/model': 0.39338735600176733, 'learning_rate/qwen_vl_interface': 5.67881979104958e-06, 'learning_rate/action_model': 5.67881979104958e-05, 'epoch': 0.18}) +[2026-10-03 13:30:08,706][starVLA.training.train_starvla][INFO] - Step 2410, Loss: {'train/loss': 2.845285416697152e-05, 'train/grad_norm_pre_clip': 0.006459810305386782, 'timing/data': 0.00024806999135762453, 'timing/model': 0.3942669259995455, 'learning_rate/qwen_vl_interface': 5.67578671729131e-06, 'learning_rate/action_model': 5.675786717291309e-05, 'epoch': 0.18}) +[2026-10-03 13:30:09,605][starVLA.training.train_starvla][INFO] - Step 2411, Loss: {'train/loss': 4.7025583626236767e-05, 'train/grad_norm_pre_clip': 0.011469998396933079, 'timing/data': 0.00024623000354040414, 'timing/model': 0.5874340139998822, 'learning_rate/qwen_vl_interface': 5.672753468508204e-06, 'learning_rate/action_model': 5.672753468508204e-05, 'epoch': 0.18}) +[2026-10-03 13:30:10,311][starVLA.training.train_starvla][INFO] - Step 2412, Loss: {'train/loss': 0.3880324959754944, 'train/grad_norm_pre_clip': 73.69316101074219, 'timing/data': 0.0002636300050653517, 'timing/model': 0.39511516600032337, 'learning_rate/qwen_vl_interface': 5.6697200459471165e-06, 'learning_rate/action_model': 5.669720045947117e-05, 'epoch': 0.18}) +[2026-10-03 13:30:11,017][starVLA.training.train_starvla][INFO] - Step 2413, Loss: {'train/loss': 0.0002962687867693603, 'train/grad_norm_pre_clip': 0.17340728640556335, 'timing/data': 0.00030298999627120793, 'timing/model': 0.3935315559938317, 'learning_rate/qwen_vl_interface': 5.666686450854973e-06, 'learning_rate/action_model': 5.666686450854973e-05, 'epoch': 0.18}) +[2026-10-03 13:30:11,722][starVLA.training.train_starvla][INFO] - Step 2414, Loss: {'train/loss': 0.0002529500052332878, 'train/grad_norm_pre_clip': 0.11020725965499878, 'timing/data': 0.00025154001195915043, 'timing/model': 0.3944146760040894, 'learning_rate/qwen_vl_interface': 5.663652684478768e-06, 'learning_rate/action_model': 5.6636526844787676e-05, 'epoch': 0.18}) +[2026-10-03 13:30:12,427][starVLA.training.train_starvla][INFO] - Step 2415, Loss: {'train/loss': 0.0002155504043912515, 'train/grad_norm_pre_clip': 0.057763487100601196, 'timing/data': 0.00019921000057365745, 'timing/model': 0.3937714860076085, 'learning_rate/qwen_vl_interface': 5.660618748065569e-06, 'learning_rate/action_model': 5.6606187480655695e-05, 'epoch': 0.18}) +[2026-10-03 13:30:13,134][starVLA.training.train_starvla][INFO] - Step 2416, Loss: {'train/loss': 0.003110595978796482, 'train/grad_norm_pre_clip': 8.3826322555542, 'timing/data': 0.0002619900042191148, 'timing/model': 0.3946015750116203, 'learning_rate/qwen_vl_interface': 5.657584642862514e-06, 'learning_rate/action_model': 5.657584642862513e-05, 'epoch': 0.18}) +[2026-10-03 13:30:14,031][starVLA.training.train_starvla][INFO] - Step 2417, Loss: {'train/loss': 7.428413664456457e-05, 'train/grad_norm_pre_clip': 0.01501571387052536, 'timing/data': 0.0002631999959703535, 'timing/model': 0.3934907860093517, 'learning_rate/qwen_vl_interface': 5.654550370116808e-06, 'learning_rate/action_model': 5.654550370116808e-05, 'epoch': 0.18}) +[2026-10-03 13:30:14,737][starVLA.training.train_starvla][INFO] - Step 2418, Loss: {'train/loss': 0.000181303737917915, 'train/grad_norm_pre_clip': 0.05640784651041031, 'timing/data': 0.00020492999465204775, 'timing/model': 0.39506273600272834, 'learning_rate/qwen_vl_interface': 5.651515931075726e-06, 'learning_rate/action_model': 5.651515931075726e-05, 'epoch': 0.18}) +[2026-10-03 13:30:15,441][starVLA.training.train_starvla][INFO] - Step 2419, Loss: {'train/loss': 9.115000284509733e-05, 'train/grad_norm_pre_clip': 0.01984252966940403, 'timing/data': 0.00027165999927092344, 'timing/model': 0.3930605750065297, 'learning_rate/qwen_vl_interface': 5.6484813269866126e-06, 'learning_rate/action_model': 5.648481326986612e-05, 'epoch': 0.18}) +[2026-10-03 13:30:16,148][starVLA.training.train_starvla][INFO] - Step 2420, Loss: {'train/loss': 0.00017233517428394407, 'train/grad_norm_pre_clip': 0.043975990265607834, 'timing/data': 0.00030432999483309686, 'timing/model': 0.39478989600320347, 'learning_rate/qwen_vl_interface': 5.645446559096875e-06, 'learning_rate/action_model': 5.645446559096875e-05, 'epoch': 0.18}) +[2026-10-03 13:30:16,854][starVLA.training.train_starvla][INFO] - Step 2421, Loss: {'train/loss': 3.266169369453564e-05, 'train/grad_norm_pre_clip': 0.007420726120471954, 'timing/data': 0.00030061999859753996, 'timing/model': 0.3941986959980568, 'learning_rate/qwen_vl_interface': 5.6424116286539965e-06, 'learning_rate/action_model': 5.642411628653996e-05, 'epoch': 0.18}) +[2026-10-03 13:30:17,751][starVLA.training.train_starvla][INFO] - Step 2422, Loss: {'train/loss': 0.004969967529177666, 'train/grad_norm_pre_clip': 5.915243148803711, 'timing/data': 0.0003108999953838065, 'timing/model': 0.39463246599189006, 'learning_rate/qwen_vl_interface': 5.639376536905517e-06, 'learning_rate/action_model': 5.6393765369055175e-05, 'epoch': 0.18}) +[2026-10-03 13:30:18,456][starVLA.training.train_starvla][INFO] - Step 2423, Loss: {'train/loss': 0.019677525386214256, 'train/grad_norm_pre_clip': 5.855908393859863, 'timing/data': 0.00033696000173222274, 'timing/model': 0.3933688159886515, 'learning_rate/qwen_vl_interface': 5.636341285099055e-06, 'learning_rate/action_model': 5.636341285099055e-05, 'epoch': 0.18}) +[2026-10-03 13:30:19,163][starVLA.training.train_starvla][INFO] - Step 2424, Loss: {'train/loss': 0.00023484055418521166, 'train/grad_norm_pre_clip': 0.0817941203713417, 'timing/data': 0.0002904799912357703, 'timing/model': 0.3949283149995608, 'learning_rate/qwen_vl_interface': 5.6333058744822835e-06, 'learning_rate/action_model': 5.633305874482283e-05, 'epoch': 0.18}) +[2026-10-03 13:30:19,870][starVLA.training.train_starvla][INFO] - Step 2425, Loss: {'train/loss': 0.0003048972284886986, 'train/grad_norm_pre_clip': 0.0699484571814537, 'timing/data': 0.000321830011671409, 'timing/model': 0.3937006259948248, 'learning_rate/qwen_vl_interface': 5.630270306302944e-06, 'learning_rate/action_model': 5.6302703063029436e-05, 'epoch': 0.18}) +[2026-10-03 13:30:20,575][starVLA.training.train_starvla][INFO] - Step 2426, Loss: {'train/loss': 0.0010366920614615083, 'train/grad_norm_pre_clip': 1.1766690015792847, 'timing/data': 0.0003077000001212582, 'timing/model': 0.3945337650075089, 'learning_rate/qwen_vl_interface': 5.627234581808848e-06, 'learning_rate/action_model': 5.627234581808848e-05, 'epoch': 0.18}) +[2026-10-03 13:30:21,474][starVLA.training.train_starvla][INFO] - Step 2427, Loss: {'train/loss': 7.396966975647956e-05, 'train/grad_norm_pre_clip': 0.015200418420135975, 'timing/data': 0.00021360999380704015, 'timing/model': 0.5862132140027825, 'learning_rate/qwen_vl_interface': 5.624198702247863e-06, 'learning_rate/action_model': 5.624198702247863e-05, 'epoch': 0.18}) +[2026-10-03 13:30:22,181][starVLA.training.train_starvla][INFO] - Step 2428, Loss: {'train/loss': 0.05077129602432251, 'train/grad_norm_pre_clip': 37.35247802734375, 'timing/data': 0.00031939000473357737, 'timing/model': 0.3948116149986163, 'learning_rate/qwen_vl_interface': 5.621162668867925e-06, 'learning_rate/action_model': 5.621162668867925e-05, 'epoch': 0.18}) +[2026-10-03 13:30:22,886][starVLA.training.train_starvla][INFO] - Step 2429, Loss: {'train/loss': 1.576513022882864e-05, 'train/grad_norm_pre_clip': 0.0030558514408767223, 'timing/data': 0.00031242999830283225, 'timing/model': 0.39357234600174706, 'learning_rate/qwen_vl_interface': 5.618126482917037e-06, 'learning_rate/action_model': 5.618126482917036e-05, 'epoch': 0.18}) +[2026-10-03 13:30:23,593][starVLA.training.train_starvla][INFO] - Step 2430, Loss: {'train/loss': 0.0003341874689795077, 'train/grad_norm_pre_clip': 0.24738042056560516, 'timing/data': 0.00020684000628534704, 'timing/model': 0.3957626859919401, 'learning_rate/qwen_vl_interface': 5.6150901456432546e-06, 'learning_rate/action_model': 5.615090145643255e-05, 'epoch': 0.18}) +[2026-10-03 13:30:24,300][starVLA.training.train_starvla][INFO] - Step 2431, Loss: {'train/loss': 0.443037211894989, 'train/grad_norm_pre_clip': 92.4229965209961, 'timing/data': 0.00033044000156223774, 'timing/model': 0.3947199549875222, 'learning_rate/qwen_vl_interface': 5.612053658294705e-06, 'learning_rate/action_model': 5.6120536582947045e-05, 'epoch': 0.18}) +[2026-10-03 13:30:25,203][starVLA.training.train_starvla][INFO] - Step 2432, Loss: {'train/loss': 0.14601552486419678, 'train/grad_norm_pre_clip': 118.62751007080078, 'timing/data': 0.00039556999399792403, 'timing/model': 0.5907021040038671, 'learning_rate/qwen_vl_interface': 5.609017022119572e-06, 'learning_rate/action_model': 5.609017022119573e-05, 'epoch': 0.18}) +[2026-10-03 13:30:25,910][starVLA.training.train_starvla][INFO] - Step 2433, Loss: {'train/loss': 1.929626705532428e-05, 'train/grad_norm_pre_clip': 0.0059264167211949825, 'timing/data': 0.000372159993276, 'timing/model': 0.3942307860124856, 'learning_rate/qwen_vl_interface': 5.605980238366104e-06, 'learning_rate/action_model': 5.605980238366103e-05, 'epoch': 0.18}) +[2026-10-03 13:30:26,616][starVLA.training.train_starvla][INFO] - Step 2434, Loss: {'train/loss': 0.01416916586458683, 'train/grad_norm_pre_clip': 11.57528305053711, 'timing/data': 0.00029899000946898013, 'timing/model': 0.3948737660102779, 'learning_rate/qwen_vl_interface': 5.602943308282606e-06, 'learning_rate/action_model': 5.602943308282605e-05, 'epoch': 0.18}) +[2026-10-03 13:30:27,322][starVLA.training.train_starvla][INFO] - Step 2435, Loss: {'train/loss': 0.00011086705490015447, 'train/grad_norm_pre_clip': 0.040239181369543076, 'timing/data': 0.00028113000735174865, 'timing/model': 0.393775464995997, 'learning_rate/qwen_vl_interface': 5.599906233117445e-06, 'learning_rate/action_model': 5.5999062331174446e-05, 'epoch': 0.18}) +[2026-10-03 13:30:28,030][starVLA.training.train_starvla][INFO] - Step 2436, Loss: {'train/loss': 1.786602479114663e-05, 'train/grad_norm_pre_clip': 0.004658609628677368, 'timing/data': 0.00035581999691203237, 'timing/model': 0.3947211759950733, 'learning_rate/qwen_vl_interface': 5.596869014119048e-06, 'learning_rate/action_model': 5.596869014119048e-05, 'epoch': 0.18}) +[2026-10-03 13:30:28,930][starVLA.training.train_starvla][INFO] - Step 2437, Loss: {'train/loss': 0.0005484624416567385, 'train/grad_norm_pre_clip': 0.324957013130188, 'timing/data': 0.00033290000283159316, 'timing/model': 0.5877605440036859, 'learning_rate/qwen_vl_interface': 5.593831652535902e-06, 'learning_rate/action_model': 5.593831652535902e-05, 'epoch': 0.18}) +[2026-10-03 13:30:29,635][starVLA.training.train_starvla][INFO] - Step 2438, Loss: {'train/loss': 0.0004081499937456101, 'train/grad_norm_pre_clip': 0.08900387585163116, 'timing/data': 0.0002828900032909587, 'timing/model': 0.3948601150041213, 'learning_rate/qwen_vl_interface': 5.590794149616552e-06, 'learning_rate/action_model': 5.590794149616552e-05, 'epoch': 0.18}) +[2026-10-03 13:30:30,340][starVLA.training.train_starvla][INFO] - Step 2439, Loss: {'train/loss': 0.0016223916318267584, 'train/grad_norm_pre_clip': 1.2286914587020874, 'timing/data': 0.00029210999491624534, 'timing/model': 0.3938137560035102, 'learning_rate/qwen_vl_interface': 5.587756506609601e-06, 'learning_rate/action_model': 5.5877565066096006e-05, 'epoch': 0.18}) +[2026-10-03 13:30:31,049][starVLA.training.train_starvla][INFO] - Step 2440, Loss: {'train/loss': 1.960922236321494e-05, 'train/grad_norm_pre_clip': 0.005503651686012745, 'timing/data': 0.0003492300020297989, 'timing/model': 0.3957966559974011, 'learning_rate/qwen_vl_interface': 5.584718724763707e-06, 'learning_rate/action_model': 5.584718724763707e-05, 'epoch': 0.18}) +[2026-10-03 13:30:31,755][starVLA.training.train_starvla][INFO] - Step 2441, Loss: {'train/loss': 0.0013484330847859383, 'train/grad_norm_pre_clip': 1.790023922920227, 'timing/data': 0.00033348999568261206, 'timing/model': 0.3941374659916619, 'learning_rate/qwen_vl_interface': 5.581680805327588e-06, 'learning_rate/action_model': 5.5816808053275885e-05, 'epoch': 0.18}) +[2026-10-03 13:30:32,460][starVLA.training.train_starvla][INFO] - Step 2442, Loss: {'train/loss': 0.004239615984261036, 'train/grad_norm_pre_clip': 6.775060176849365, 'timing/data': 0.00025533000007271767, 'timing/model': 0.39530188600474503, 'learning_rate/qwen_vl_interface': 5.578642749550019e-06, 'learning_rate/action_model': 5.578642749550019e-05, 'epoch': 0.18}) +[2026-10-03 13:30:33,332][starVLA.training.train_starvla][INFO] - Step 2443, Loss: {'train/loss': 0.0001795005373423919, 'train/grad_norm_pre_clip': 0.07479532808065414, 'timing/data': 0.00029843999072909355, 'timing/model': 0.58386207500007, 'learning_rate/qwen_vl_interface': 5.575604558679831e-06, 'learning_rate/action_model': 5.57560455867983e-05, 'epoch': 0.18}) +[2026-10-03 13:30:34,040][starVLA.training.train_starvla][INFO] - Step 2444, Loss: {'train/loss': 0.004983231890946627, 'train/grad_norm_pre_clip': 5.139513969421387, 'timing/data': 0.0003188299888279289, 'timing/model': 0.39504715500515886, 'learning_rate/qwen_vl_interface': 5.5725662339659085e-06, 'learning_rate/action_model': 5.572566233965908e-05, 'epoch': 0.18}) +[2026-10-03 13:30:34,746][starVLA.training.train_starvla][INFO] - Step 2445, Loss: {'train/loss': 0.00010423991625430062, 'train/grad_norm_pre_clip': 0.03643861785531044, 'timing/data': 0.00027256000612396747, 'timing/model': 0.39402426600281615, 'learning_rate/qwen_vl_interface': 5.56952777665719e-06, 'learning_rate/action_model': 5.56952777665719e-05, 'epoch': 0.18}) +[2026-10-03 13:30:35,452][starVLA.training.train_starvla][INFO] - Step 2446, Loss: {'train/loss': 1.6122783563332632e-05, 'train/grad_norm_pre_clip': 0.0042958795092999935, 'timing/data': 0.0002569499920355156, 'timing/model': 0.394842455003527, 'learning_rate/qwen_vl_interface': 5.566489188002671e-06, 'learning_rate/action_model': 5.56648918800267e-05, 'epoch': 0.18}) +[2026-10-03 13:30:36,158][starVLA.training.train_starvla][INFO] - Step 2447, Loss: {'train/loss': 0.3777152895927429, 'train/grad_norm_pre_clip': 67.91587829589844, 'timing/data': 0.00021235000167507678, 'timing/model': 0.39394698600517586, 'learning_rate/qwen_vl_interface': 5.563450469251403e-06, 'learning_rate/action_model': 5.563450469251402e-05, 'epoch': 0.18}) +[2026-10-03 13:30:37,058][starVLA.training.train_starvla][INFO] - Step 2448, Loss: {'train/loss': 5.734445949201472e-05, 'train/grad_norm_pre_clip': 0.028642795979976654, 'timing/data': 0.0003652989980764687, 'timing/model': 0.5884951139887562, 'learning_rate/qwen_vl_interface': 5.560411621652488e-06, 'learning_rate/action_model': 5.560411621652488e-05, 'epoch': 0.18}) +[2026-10-03 13:30:37,764][starVLA.training.train_starvla][INFO] - Step 2449, Loss: {'train/loss': 7.08753868821077e-05, 'train/grad_norm_pre_clip': 0.07583951950073242, 'timing/data': 0.0003501499886624515, 'timing/model': 0.3941057159972843, 'learning_rate/qwen_vl_interface': 5.557372646455081e-06, 'learning_rate/action_model': 5.55737264645508e-05, 'epoch': 0.18}) +[2026-10-03 13:30:38,471][starVLA.training.train_starvla][INFO] - Step 2450, Loss: {'train/loss': 0.6439752578735352, 'train/grad_norm_pre_clip': 241.4910888671875, 'timing/data': 0.0002568700001575053, 'timing/model': 0.3950217150122626, 'learning_rate/qwen_vl_interface': 5.554333544908388e-06, 'learning_rate/action_model': 5.5543335449083875e-05, 'epoch': 0.18}) +[2026-10-03 13:30:39,176][starVLA.training.train_starvla][INFO] - Step 2451, Loss: {'train/loss': 0.5332499146461487, 'train/grad_norm_pre_clip': 14.530192375183105, 'timing/data': 0.0002199400041718036, 'timing/model': 0.39369745600561146, 'learning_rate/qwen_vl_interface': 5.551294318261671e-06, 'learning_rate/action_model': 5.5512943182616705e-05, 'epoch': 0.18}) +[2026-10-03 13:30:39,884][starVLA.training.train_starvla][INFO] - Step 2452, Loss: {'train/loss': 0.00016551064618397504, 'train/grad_norm_pre_clip': 0.18634025752544403, 'timing/data': 0.0002197599969804287, 'timing/model': 0.3950613560009515, 'learning_rate/qwen_vl_interface': 5.548254967764239e-06, 'learning_rate/action_model': 5.5482549677642394e-05, 'epoch': 0.18}) +[2026-10-03 13:30:40,781][starVLA.training.train_starvla][INFO] - Step 2453, Loss: {'train/loss': 0.0001066041222657077, 'train/grad_norm_pre_clip': 0.026453088968992233, 'timing/data': 0.00031974000739865005, 'timing/model': 0.5850320840108907, 'learning_rate/qwen_vl_interface': 5.545215494665458e-06, 'learning_rate/action_model': 5.545215494665458e-05, 'epoch': 0.18}) +[2026-10-03 13:30:41,489][starVLA.training.train_starvla][INFO] - Step 2454, Loss: {'train/loss': 0.8447692394256592, 'train/grad_norm_pre_clip': 206.1134796142578, 'timing/data': 0.0002615500125102699, 'timing/model': 0.3951029760064557, 'learning_rate/qwen_vl_interface': 5.5421759002147376e-06, 'learning_rate/action_model': 5.542175900214738e-05, 'epoch': 0.18}) +[2026-10-03 13:30:42,194][starVLA.training.train_starvla][INFO] - Step 2455, Loss: {'train/loss': 0.0012959323357790709, 'train/grad_norm_pre_clip': 8.377554893493652, 'timing/data': 0.0002955000090878457, 'timing/model': 0.3939679960021749, 'learning_rate/qwen_vl_interface': 5.53913618566154e-06, 'learning_rate/action_model': 5.53913618566154e-05, 'epoch': 0.18}) +[2026-10-03 13:30:42,902][starVLA.training.train_starvla][INFO] - Step 2456, Loss: {'train/loss': 5.185538611840457e-06, 'train/grad_norm_pre_clip': 0.0018108357908204198, 'timing/data': 0.0003674100007629022, 'timing/model': 0.39523532601015177, 'learning_rate/qwen_vl_interface': 5.536096352255381e-06, 'learning_rate/action_model': 5.536096352255381e-05, 'epoch': 0.18}) +[2026-10-03 13:30:43,608][starVLA.training.train_starvla][INFO] - Step 2457, Loss: {'train/loss': 0.0001367047370877117, 'train/grad_norm_pre_clip': 0.031829532235860825, 'timing/data': 0.0002708500105654821, 'timing/model': 0.39387699599319603, 'learning_rate/qwen_vl_interface': 5.533056401245817e-06, 'learning_rate/action_model': 5.5330564012458164e-05, 'epoch': 0.18}) +[2026-10-03 13:30:44,315][starVLA.training.train_starvla][INFO] - Step 2458, Loss: {'train/loss': 0.48065343499183655, 'train/grad_norm_pre_clip': 43.61784744262695, 'timing/data': 0.000259280001046136, 'timing/model': 0.3956282760045724, 'learning_rate/qwen_vl_interface': 5.530016333882456e-06, 'learning_rate/action_model': 5.5300163338824565e-05, 'epoch': 0.18}) +[2026-10-03 13:30:45,214][starVLA.training.train_starvla][INFO] - Step 2459, Loss: {'train/loss': 0.0002652195980772376, 'train/grad_norm_pre_clip': 0.21688586473464966, 'timing/data': 0.000257949999650009, 'timing/model': 0.39415480601019226, 'learning_rate/qwen_vl_interface': 5.52697615141496e-06, 'learning_rate/action_model': 5.52697615141496e-05, 'epoch': 0.18}) +[2026-10-03 13:30:45,921][starVLA.training.train_starvla][INFO] - Step 2460, Loss: {'train/loss': 4.271924262866378e-05, 'train/grad_norm_pre_clip': 0.023900505155324936, 'timing/data': 0.00031693000346422195, 'timing/model': 0.3952026759943692, 'learning_rate/qwen_vl_interface': 5.52393585509303e-06, 'learning_rate/action_model': 5.52393585509303e-05, 'epoch': 0.18}) +[2026-10-03 13:30:46,627][starVLA.training.train_starvla][INFO] - Step 2461, Loss: {'train/loss': 0.4577306807041168, 'train/grad_norm_pre_clip': 115.72562408447266, 'timing/data': 0.0002931100025307387, 'timing/model': 0.39385590598976705, 'learning_rate/qwen_vl_interface': 5.520895446166418e-06, 'learning_rate/action_model': 5.5208954461664175e-05, 'epoch': 0.18}) +[2026-10-03 13:30:47,333][starVLA.training.train_starvla][INFO] - Step 2462, Loss: {'train/loss': 1.877448994491715e-05, 'train/grad_norm_pre_clip': 0.006961664650589228, 'timing/data': 0.00034749999758787453, 'timing/model': 0.394805705989711, 'learning_rate/qwen_vl_interface': 5.517854925884919e-06, 'learning_rate/action_model': 5.517854925884919e-05, 'epoch': 0.18}) +[2026-10-03 13:30:48,040][starVLA.training.train_starvla][INFO] - Step 2463, Loss: {'train/loss': 0.0002936887904070318, 'train/grad_norm_pre_clip': 0.5468276739120483, 'timing/data': 0.00025040000036824495, 'timing/model': 0.39367309599765576, 'learning_rate/qwen_vl_interface': 5.514814295498379e-06, 'learning_rate/action_model': 5.514814295498378e-05, 'epoch': 0.18}) +[2026-10-03 13:30:48,940][starVLA.training.train_starvla][INFO] - Step 2464, Loss: {'train/loss': 0.055388953536748886, 'train/grad_norm_pre_clip': 319.2904357910156, 'timing/data': 0.0003452300006756559, 'timing/model': 0.39484822600206826, 'learning_rate/qwen_vl_interface': 5.5117735562566846e-06, 'learning_rate/action_model': 5.5117735562566844e-05, 'epoch': 0.18}) +[2026-10-03 13:30:49,645][starVLA.training.train_starvla][INFO] - Step 2465, Loss: {'train/loss': 0.0004299189895391464, 'train/grad_norm_pre_clip': 0.16906186938285828, 'timing/data': 0.00036825999268330634, 'timing/model': 0.39385880599729717, 'learning_rate/qwen_vl_interface': 5.5087327094097685e-06, 'learning_rate/action_model': 5.508732709409769e-05, 'epoch': 0.18}) +[2026-10-03 13:30:50,349][starVLA.training.train_starvla][INFO] - Step 2466, Loss: {'train/loss': 0.00378593010827899, 'train/grad_norm_pre_clip': 2.093719959259033, 'timing/data': 0.00026760000037029386, 'timing/model': 0.3930337260098895, 'learning_rate/qwen_vl_interface': 5.505691756207609e-06, 'learning_rate/action_model': 5.505691756207609e-05, 'epoch': 0.18}) +[2026-10-03 13:30:51,055][starVLA.training.train_starvla][INFO] - Step 2467, Loss: {'train/loss': 0.0003403364971745759, 'train/grad_norm_pre_clip': 0.18446239829063416, 'timing/data': 0.0003331400075694546, 'timing/model': 0.3939563060121145, 'learning_rate/qwen_vl_interface': 5.502650697900228e-06, 'learning_rate/action_model': 5.502650697900228e-05, 'epoch': 0.18}) +[2026-10-03 13:30:51,763][starVLA.training.train_starvla][INFO] - Step 2468, Loss: {'train/loss': 5.6325889090658166e-06, 'train/grad_norm_pre_clip': 0.002146864077076316, 'timing/data': 0.0003334700013510883, 'timing/model': 0.39460023598803673, 'learning_rate/qwen_vl_interface': 5.499609535737688e-06, 'learning_rate/action_model': 5.4996095357376875e-05, 'epoch': 0.18}) +[2026-10-03 13:30:52,664][starVLA.training.train_starvla][INFO] - Step 2469, Loss: {'train/loss': 0.04055982828140259, 'train/grad_norm_pre_clip': 26.412017822265625, 'timing/data': 0.0003651300066849217, 'timing/model': 0.5878705740033183, 'learning_rate/qwen_vl_interface': 5.496568270970096e-06, 'learning_rate/action_model': 5.496568270970096e-05, 'epoch': 0.18}) +[2026-10-03 13:30:53,370][starVLA.training.train_starvla][INFO] - Step 2470, Loss: {'train/loss': 0.00396573543548584, 'train/grad_norm_pre_clip': 3.0201919078826904, 'timing/data': 0.0002969999914057553, 'timing/model': 0.39531161601189524, 'learning_rate/qwen_vl_interface': 5.4935269048476015e-06, 'learning_rate/action_model': 5.493526904847601e-05, 'epoch': 0.18}) +[2026-10-03 13:30:54,076][starVLA.training.train_starvla][INFO] - Step 2471, Loss: {'train/loss': 5.341464202501811e-05, 'train/grad_norm_pre_clip': 0.05234311521053314, 'timing/data': 0.0004088499990757555, 'timing/model': 0.39411492500221357, 'learning_rate/qwen_vl_interface': 5.490485438620397e-06, 'learning_rate/action_model': 5.4904854386203965e-05, 'epoch': 0.18}) +[2026-10-03 13:30:54,781][starVLA.training.train_starvla][INFO] - Step 2472, Loss: {'train/loss': 0.00010750476940302178, 'train/grad_norm_pre_clip': 0.02374301292002201, 'timing/data': 0.0003150300035485998, 'timing/model': 0.39460856599907856, 'learning_rate/qwen_vl_interface': 5.487443873538712e-06, 'learning_rate/action_model': 5.4874438735387115e-05, 'epoch': 0.18}) +[2026-10-03 13:30:55,488][starVLA.training.train_starvla][INFO] - Step 2473, Loss: {'train/loss': 0.630943775177002, 'train/grad_norm_pre_clip': 269.5760803222656, 'timing/data': 0.0003481700114207342, 'timing/model': 0.39421333601057995, 'learning_rate/qwen_vl_interface': 5.484402210852819e-06, 'learning_rate/action_model': 5.484402210852818e-05, 'epoch': 0.18}) +[2026-10-03 13:30:56,389][starVLA.training.train_starvla][INFO] - Step 2474, Loss: {'train/loss': 1.470705319661647e-05, 'train/grad_norm_pre_clip': 0.004527710378170013, 'timing/data': 0.00025795999681577086, 'timing/model': 0.5893408940028166, 'learning_rate/qwen_vl_interface': 5.48136045181303e-06, 'learning_rate/action_model': 5.48136045181303e-05, 'epoch': 0.18}) +[2026-10-03 13:30:57,095][starVLA.training.train_starvla][INFO] - Step 2475, Loss: {'train/loss': 8.590351171733346e-06, 'train/grad_norm_pre_clip': 0.0021266101393848658, 'timing/data': 0.00040608999552205205, 'timing/model': 0.39449487600359134, 'learning_rate/qwen_vl_interface': 5.478318597669699e-06, 'learning_rate/action_model': 5.4783185976696995e-05, 'epoch': 0.18}) +[2026-10-03 13:30:57,801][starVLA.training.train_starvla][INFO] - Step 2476, Loss: {'train/loss': 4.827916200156324e-06, 'train/grad_norm_pre_clip': 0.0010386788053438067, 'timing/data': 0.0003339099930599332, 'timing/model': 0.3945428949955385, 'learning_rate/qwen_vl_interface': 5.475276649673218e-06, 'learning_rate/action_model': 5.4752766496732174e-05, 'epoch': 0.18}) +[2026-10-03 13:30:58,506][starVLA.training.train_starvla][INFO] - Step 2477, Loss: {'train/loss': 0.0021305435802787542, 'train/grad_norm_pre_clip': 11.558753967285156, 'timing/data': 0.0003477200079942122, 'timing/model': 0.39288937600213103, 'learning_rate/qwen_vl_interface': 5.472234609074014e-06, 'learning_rate/action_model': 5.472234609074014e-05, 'epoch': 0.18}) +[2026-10-03 13:30:59,211][starVLA.training.train_starvla][INFO] - Step 2478, Loss: {'train/loss': 0.0003045511257369071, 'train/grad_norm_pre_clip': 0.23907233774662018, 'timing/data': 0.00029845999961253256, 'timing/model': 0.3946460860024672, 'learning_rate/qwen_vl_interface': 5.46919247712256e-06, 'learning_rate/action_model': 5.46919247712256e-05, 'epoch': 0.18}) +[2026-10-03 13:30:59,916][starVLA.training.train_starvla][INFO] - Step 2479, Loss: {'train/loss': 0.004257573746144772, 'train/grad_norm_pre_clip': 14.322413444519043, 'timing/data': 0.00020534900249913335, 'timing/model': 0.3933392860053573, 'learning_rate/qwen_vl_interface': 5.466150255069356e-06, 'learning_rate/action_model': 5.4661502550693555e-05, 'epoch': 0.18}) +[2026-10-03 13:31:00,811][starVLA.training.train_starvla][INFO] - Step 2480, Loss: {'train/loss': 0.00015229221025947481, 'train/grad_norm_pre_clip': 0.04059606045484543, 'timing/data': 0.00021429000480566174, 'timing/model': 0.39577541600738186, 'learning_rate/qwen_vl_interface': 5.463107944164947e-06, 'learning_rate/action_model': 5.4631079441649466e-05, 'epoch': 0.18}) +[2026-10-03 13:31:01,518][starVLA.training.train_starvla][INFO] - Step 2481, Loss: {'train/loss': 0.001464603003114462, 'train/grad_norm_pre_clip': 4.146410942077637, 'timing/data': 0.00035288999788463116, 'timing/model': 0.3937486950017046, 'learning_rate/qwen_vl_interface': 5.46006554565991e-06, 'learning_rate/action_model': 5.4600655456599105e-05, 'epoch': 0.18}) +[2026-10-03 13:31:02,224][starVLA.training.train_starvla][INFO] - Step 2482, Loss: {'train/loss': 0.00024333986220881343, 'train/grad_norm_pre_clip': 0.05758360028266907, 'timing/data': 0.0003129899996565655, 'timing/model': 0.39533405599649996, 'learning_rate/qwen_vl_interface': 5.457023060804865e-06, 'learning_rate/action_model': 5.457023060804864e-05, 'epoch': 0.18}) +[2026-10-03 13:31:02,925][starVLA.training.train_starvla][INFO] - Step 2483, Loss: {'train/loss': 0.0016553116729483008, 'train/grad_norm_pre_clip': 1.125687837600708, 'timing/data': 0.0002402399986749515, 'timing/model': 0.38838235700677615, 'learning_rate/qwen_vl_interface': 5.453980490850459e-06, 'learning_rate/action_model': 5.453980490850459e-05, 'epoch': 0.18}) +[2026-10-03 13:31:03,624][starVLA.training.train_starvla][INFO] - Step 2484, Loss: {'train/loss': 0.0002905103610828519, 'train/grad_norm_pre_clip': 0.05631159991025925, 'timing/data': 0.00031215000490192324, 'timing/model': 0.3893774649914121, 'learning_rate/qwen_vl_interface': 5.4509378370473785e-06, 'learning_rate/action_model': 5.450937837047378e-05, 'epoch': 0.18}) +[2026-10-03 13:31:04,513][starVLA.training.train_starvla][INFO] - Step 2485, Loss: {'train/loss': 1.5355140931205824e-05, 'train/grad_norm_pre_clip': 0.004405651241540909, 'timing/data': 0.00020520998805295676, 'timing/model': 0.3881795269990107, 'learning_rate/qwen_vl_interface': 5.4478951006463424e-06, 'learning_rate/action_model': 5.447895100646342e-05, 'epoch': 0.18}) +[2026-10-03 13:31:05,208][starVLA.training.train_starvla][INFO] - Step 2486, Loss: {'train/loss': 1.0013052815338597e-05, 'train/grad_norm_pre_clip': 0.005702740512788296, 'timing/data': 0.0002647399960551411, 'timing/model': 0.3906322069960879, 'learning_rate/qwen_vl_interface': 5.444852282898106e-06, 'learning_rate/action_model': 5.444852282898106e-05, 'epoch': 0.18}) +[2026-10-03 13:31:05,903][starVLA.training.train_starvla][INFO] - Step 2487, Loss: {'train/loss': 4.8726319619163405e-06, 'train/grad_norm_pre_clip': 0.0008299470064230263, 'timing/data': 0.0002406900021014735, 'timing/model': 0.38859387701086234, 'learning_rate/qwen_vl_interface': 5.441809385053456e-06, 'learning_rate/action_model': 5.4418093850534555e-05, 'epoch': 0.18}) +[2026-10-03 13:31:06,600][starVLA.training.train_starvla][INFO] - Step 2488, Loss: {'train/loss': 0.47502249479293823, 'train/grad_norm_pre_clip': 43.78334045410156, 'timing/data': 0.00021862999710720032, 'timing/model': 0.3901798059960129, 'learning_rate/qwen_vl_interface': 5.438766408363214e-06, 'learning_rate/action_model': 5.438766408363214e-05, 'epoch': 0.18}) +[2026-10-03 13:31:07,296][starVLA.training.train_starvla][INFO] - Step 2489, Loss: {'train/loss': 0.0009590377449057996, 'train/grad_norm_pre_clip': 3.162534236907959, 'timing/data': 0.00021770999592263252, 'timing/model': 0.3884652770066168, 'learning_rate/qwen_vl_interface': 5.43572335407823e-06, 'learning_rate/action_model': 5.43572335407823e-05, 'epoch': 0.18}) +[2026-10-03 13:31:08,185][starVLA.training.train_starvla][INFO] - Step 2490, Loss: {'train/loss': 9.422327275387943e-05, 'train/grad_norm_pre_clip': 0.02731637842953205, 'timing/data': 0.0002928999892901629, 'timing/model': 0.5824655549949966, 'learning_rate/qwen_vl_interface': 5.432680223449392e-06, 'learning_rate/action_model': 5.4326802234493924e-05, 'epoch': 0.18}) +[2026-10-03 13:31:08,878][starVLA.training.train_starvla][INFO] - Step 2491, Loss: {'train/loss': 0.000739423674531281, 'train/grad_norm_pre_clip': 1.195228934288025, 'timing/data': 0.00019289999909233302, 'timing/model': 0.38809787698846776, 'learning_rate/qwen_vl_interface': 5.429637017727614e-06, 'learning_rate/action_model': 5.4296370177276144e-05, 'epoch': 0.18}) +[2026-10-03 13:31:09,576][starVLA.training.train_starvla][INFO] - Step 2492, Loss: {'train/loss': 0.00010859331086976454, 'train/grad_norm_pre_clip': 0.07966972887516022, 'timing/data': 0.0002740299969445914, 'timing/model': 0.3905682059994433, 'learning_rate/qwen_vl_interface': 5.426593738163845e-06, 'learning_rate/action_model': 5.4265937381638446e-05, 'epoch': 0.18}) +[2026-10-03 13:31:10,272][starVLA.training.train_starvla][INFO] - Step 2493, Loss: {'train/loss': 0.001090149162337184, 'train/grad_norm_pre_clip': 1.0867935419082642, 'timing/data': 0.00023126999440137297, 'timing/model': 0.3884464559960179, 'learning_rate/qwen_vl_interface': 5.42355038600906e-06, 'learning_rate/action_model': 5.42355038600906e-05, 'epoch': 0.18}) +[2026-10-03 13:31:10,969][starVLA.training.train_starvla][INFO] - Step 2494, Loss: {'train/loss': 0.05405062809586525, 'train/grad_norm_pre_clip': 82.07672119140625, 'timing/data': 0.0003300499956822023, 'timing/model': 0.3911853460012935, 'learning_rate/qwen_vl_interface': 5.420506962514267e-06, 'learning_rate/action_model': 5.420506962514267e-05, 'epoch': 0.18}) +[2026-10-03 13:31:11,859][starVLA.training.train_starvla][INFO] - Step 2495, Loss: {'train/loss': 3.49428364643245e-06, 'train/grad_norm_pre_clip': 0.0009622682700864971, 'timing/data': 0.00024588999804109335, 'timing/model': 0.5832969939947361, 'learning_rate/qwen_vl_interface': 5.4174634689305015e-06, 'learning_rate/action_model': 5.417463468930501e-05, 'epoch': 0.18}) +[2026-10-03 13:31:12,557][starVLA.training.train_starvla][INFO] - Step 2496, Loss: {'train/loss': 1.244182203663513e-05, 'train/grad_norm_pre_clip': 0.006172639783471823, 'timing/data': 0.00034790999779943377, 'timing/model': 0.39056409600016195, 'learning_rate/qwen_vl_interface': 5.41441990650883e-06, 'learning_rate/action_model': 5.4144199065088295e-05, 'epoch': 0.18}) +[2026-10-03 13:31:13,254][starVLA.training.train_starvla][INFO] - Step 2497, Loss: {'train/loss': 0.0011021388927474618, 'train/grad_norm_pre_clip': 1.7169461250305176, 'timing/data': 0.0003313489869469777, 'timing/model': 0.3889012669969816, 'learning_rate/qwen_vl_interface': 5.411376276500345e-06, 'learning_rate/action_model': 5.411376276500345e-05, 'epoch': 0.18}) +[2026-10-03 13:31:13,952][starVLA.training.train_starvla][INFO] - Step 2498, Loss: {'train/loss': 0.004110672511160374, 'train/grad_norm_pre_clip': 2.1985244750976562, 'timing/data': 0.00030934000096749514, 'timing/model': 0.3914329160033958, 'learning_rate/qwen_vl_interface': 5.408332580156168e-06, 'learning_rate/action_model': 5.408332580156168e-05, 'epoch': 0.18}) +[2026-10-03 13:31:14,647][starVLA.training.train_starvla][INFO] - Step 2499, Loss: {'train/loss': 2.816308096953435e-06, 'train/grad_norm_pre_clip': 0.0007636796217411757, 'timing/data': 0.0002594899997347966, 'timing/model': 0.38867058599134907, 'learning_rate/qwen_vl_interface': 5.40528881872745e-06, 'learning_rate/action_model': 5.40528881872745e-05, 'epoch': 0.18}) +[2026-10-03 13:31:15,344][starVLA.training.train_starvla][INFO] - Step 2500, Loss: {'train/loss': 0.0001051898580044508, 'train/grad_norm_pre_clip': 0.032740507274866104, 'timing/data': 0.0002475200017215684, 'timing/model': 0.389159237005515, 'learning_rate/qwen_vl_interface': 5.402244993465364e-06, 'learning_rate/action_model': 5.402244993465364e-05, 'epoch': 0.18}) +[2026-10-03 13:31:16,233][starVLA.training.train_starvla][INFO] - Step 2501, Loss: {'train/loss': 0.0005732675199396908, 'train/grad_norm_pre_clip': 0.2254677712917328, 'timing/data': 0.0002596099948277697, 'timing/model': 0.38872759700461756, 'learning_rate/qwen_vl_interface': 5.399201105621111e-06, 'learning_rate/action_model': 5.39920110562111e-05, 'epoch': 0.18}) +[2026-10-03 13:31:16,931][starVLA.training.train_starvla][INFO] - Step 2502, Loss: {'train/loss': 8.835836524667684e-06, 'train/grad_norm_pre_clip': 0.005099009722471237, 'timing/data': 0.00019842000619973987, 'timing/model': 0.3910441850021016, 'learning_rate/qwen_vl_interface': 5.396157156445918e-06, 'learning_rate/action_model': 5.396157156445918e-05, 'epoch': 0.18}) +[2026-10-03 13:31:17,625][starVLA.training.train_starvla][INFO] - Step 2503, Loss: {'train/loss': 0.00011918204836547375, 'train/grad_norm_pre_clip': 0.023153843358159065, 'timing/data': 0.00019396000425331295, 'timing/model': 0.3887869869940914, 'learning_rate/qwen_vl_interface': 5.39311314719104e-06, 'learning_rate/action_model': 5.39311314719104e-05, 'epoch': 0.18}) +[2026-10-03 13:31:18,323][starVLA.training.train_starvla][INFO] - Step 2504, Loss: {'train/loss': 0.00010849427053472027, 'train/grad_norm_pre_clip': 0.030215909704566002, 'timing/data': 0.00028326000028755516, 'timing/model': 0.3903671060106717, 'learning_rate/qwen_vl_interface': 5.390069079107755e-06, 'learning_rate/action_model': 5.390069079107755e-05, 'epoch': 0.18}) +[2026-10-03 13:31:19,020][starVLA.training.train_starvla][INFO] - Step 2505, Loss: {'train/loss': 0.3662812411785126, 'train/grad_norm_pre_clip': 23.60112953186035, 'timing/data': 0.0002855299971997738, 'timing/model': 0.38890569600334857, 'learning_rate/qwen_vl_interface': 5.387024953447361e-06, 'learning_rate/action_model': 5.387024953447361e-05, 'epoch': 0.18}) +[2026-10-03 13:31:19,910][starVLA.training.train_starvla][INFO] - Step 2506, Loss: {'train/loss': 0.07803112268447876, 'train/grad_norm_pre_clip': 127.47139739990234, 'timing/data': 0.0002804300020216033, 'timing/model': 0.5834908649994759, 'learning_rate/qwen_vl_interface': 5.383980771461184e-06, 'learning_rate/action_model': 5.3839807714611846e-05, 'epoch': 0.18}) +[2026-10-03 13:31:20,605][starVLA.training.train_starvla][INFO] - Step 2507, Loss: {'train/loss': 0.08535023778676987, 'train/grad_norm_pre_clip': 603.2588500976562, 'timing/data': 0.0003117099986411631, 'timing/model': 0.38889883599767927, 'learning_rate/qwen_vl_interface': 5.380936534400572e-06, 'learning_rate/action_model': 5.3809365344005726e-05, 'epoch': 0.18}) +[2026-10-03 13:31:21,304][starVLA.training.train_starvla][INFO] - Step 2508, Loss: {'train/loss': 2.5778799681575038e-06, 'train/grad_norm_pre_clip': 0.000584998691920191, 'timing/data': 0.00025769999774638563, 'timing/model': 0.39044147700769827, 'learning_rate/qwen_vl_interface': 5.377892243516899e-06, 'learning_rate/action_model': 5.3778922435168985e-05, 'epoch': 0.18}) +[2026-10-03 13:31:21,999][starVLA.training.train_starvla][INFO] - Step 2509, Loss: {'train/loss': 5.3205785661702976e-05, 'train/grad_norm_pre_clip': 0.01475805789232254, 'timing/data': 0.0003380900016054511, 'timing/model': 0.388561836007284, 'learning_rate/qwen_vl_interface': 5.3748479000615525e-06, 'learning_rate/action_model': 5.374847900061553e-05, 'epoch': 0.18}) +[2026-10-03 13:31:22,695][starVLA.training.train_starvla][INFO] - Step 2510, Loss: {'train/loss': 0.0004479825438465923, 'train/grad_norm_pre_clip': 0.999531626701355, 'timing/data': 0.00023057000362314284, 'timing/model': 0.3906340760004241, 'learning_rate/qwen_vl_interface': 5.371803505285954e-06, 'learning_rate/action_model': 5.371803505285954e-05, 'epoch': 0.18}) +[2026-10-03 13:31:23,583][starVLA.training.train_starvla][INFO] - Step 2511, Loss: {'train/loss': 2.1159487459954107e-06, 'train/grad_norm_pre_clip': 0.0007535942713730037, 'timing/data': 0.00022159999934956431, 'timing/model': 0.5818380459968466, 'learning_rate/qwen_vl_interface': 5.368759060441534e-06, 'learning_rate/action_model': 5.368759060441534e-05, 'epoch': 0.18}) +[2026-10-03 13:31:24,282][starVLA.training.train_starvla][INFO] - Step 2512, Loss: {'train/loss': 4.6668752474943176e-05, 'train/grad_norm_pre_clip': 0.013823792338371277, 'timing/data': 0.00029433000599965453, 'timing/model': 0.3907187760050874, 'learning_rate/qwen_vl_interface': 5.365714566779749e-06, 'learning_rate/action_model': 5.3657145667797484e-05, 'epoch': 0.18}) +[2026-10-03 13:31:24,979][starVLA.training.train_starvla][INFO] - Step 2513, Loss: {'train/loss': 0.0018557138973847032, 'train/grad_norm_pre_clip': 3.706073045730591, 'timing/data': 0.0003043100005015731, 'timing/model': 0.3888507660012692, 'learning_rate/qwen_vl_interface': 5.362670025552077e-06, 'learning_rate/action_model': 5.362670025552077e-05, 'epoch': 0.18}) +[2026-10-03 13:31:25,676][starVLA.training.train_starvla][INFO] - Step 2514, Loss: {'train/loss': 0.5672958493232727, 'train/grad_norm_pre_clip': 103.78414916992188, 'timing/data': 0.00023477000650018454, 'timing/model': 0.3906226360122673, 'learning_rate/qwen_vl_interface': 5.359625438010013e-06, 'learning_rate/action_model': 5.359625438010013e-05, 'epoch': 0.18}) +[2026-10-03 13:31:26,371][starVLA.training.train_starvla][INFO] - Step 2515, Loss: {'train/loss': 0.2558160424232483, 'train/grad_norm_pre_clip': 127.87513732910156, 'timing/data': 0.0002599400031613186, 'timing/model': 0.3884796160127735, 'learning_rate/qwen_vl_interface': 5.356580805405075e-06, 'learning_rate/action_model': 5.356580805405075e-05, 'epoch': 0.18}) +[2026-10-03 13:31:27,071][starVLA.training.train_starvla][INFO] - Step 2516, Loss: {'train/loss': 1.3932545925854356e-06, 'train/grad_norm_pre_clip': 0.0007892783032730222, 'timing/data': 0.0003033399989362806, 'timing/model': 0.3911286059883423, 'learning_rate/qwen_vl_interface': 5.353536128988791e-06, 'learning_rate/action_model': 5.353536128988791e-05, 'epoch': 0.18}) +[2026-10-03 13:31:27,960][starVLA.training.train_starvla][INFO] - Step 2517, Loss: {'train/loss': 1.866139609774109e-05, 'train/grad_norm_pre_clip': 0.017584824934601784, 'timing/data': 0.0003140089975204319, 'timing/model': 0.3893392760073766, 'learning_rate/qwen_vl_interface': 5.350491410012717e-06, 'learning_rate/action_model': 5.350491410012717e-05, 'epoch': 0.18}) +[2026-10-03 13:31:28,658][starVLA.training.train_starvla][INFO] - Step 2518, Loss: {'train/loss': 0.0002277345338370651, 'train/grad_norm_pre_clip': 0.0788852795958519, 'timing/data': 0.00020062999101355672, 'timing/model': 0.3918123170005856, 'learning_rate/qwen_vl_interface': 5.347446649728419e-06, 'learning_rate/action_model': 5.347446649728418e-05, 'epoch': 0.18}) +[2026-10-03 13:31:29,354][starVLA.training.train_starvla][INFO] - Step 2519, Loss: {'train/loss': 0.2206392139196396, 'train/grad_norm_pre_clip': 25.483726501464844, 'timing/data': 0.0003289700107416138, 'timing/model': 0.3890882660052739, 'learning_rate/qwen_vl_interface': 5.344401849387484e-06, 'learning_rate/action_model': 5.344401849387484e-05, 'epoch': 0.18}) +[2026-10-03 13:31:30,054][starVLA.training.train_starvla][INFO] - Step 2520, Loss: {'train/loss': 5.3052688599564135e-05, 'train/grad_norm_pre_clip': 0.011326748877763748, 'timing/data': 0.00020417998894117773, 'timing/model': 0.39155084600497503, 'learning_rate/qwen_vl_interface': 5.3413570102415144e-06, 'learning_rate/action_model': 5.3413570102415144e-05, 'epoch': 0.18}) +[2026-10-03 13:31:30,753][starVLA.training.train_starvla][INFO] - Step 2521, Loss: {'train/loss': 7.942979573272169e-05, 'train/grad_norm_pre_clip': 0.018369438126683235, 'timing/data': 0.00025442999321967363, 'timing/model': 0.3896667770022759, 'learning_rate/qwen_vl_interface': 5.338312133542131e-06, 'learning_rate/action_model': 5.3383121335421305e-05, 'epoch': 0.18}) +[2026-10-03 13:31:31,643][starVLA.training.train_starvla][INFO] - Step 2522, Loss: {'train/loss': 3.874268713843776e-06, 'train/grad_norm_pre_clip': 0.002226817887276411, 'timing/data': 0.00030123000033199787, 'timing/model': 0.3910530459979782, 'learning_rate/qwen_vl_interface': 5.335267220540963e-06, 'learning_rate/action_model': 5.335267220540963e-05, 'epoch': 0.18}) +[2026-10-03 13:31:32,338][starVLA.training.train_starvla][INFO] - Step 2523, Loss: {'train/loss': 1.0579781246633502e-06, 'train/grad_norm_pre_clip': 0.00041012425208464265, 'timing/data': 0.0002627500070957467, 'timing/model': 0.38906804600264877, 'learning_rate/qwen_vl_interface': 5.332222272489662e-06, 'learning_rate/action_model': 5.332222272489662e-05, 'epoch': 0.18}) +[2026-10-03 13:31:33,037][starVLA.training.train_starvla][INFO] - Step 2524, Loss: {'train/loss': 0.0045190006494522095, 'train/grad_norm_pre_clip': 10.67831039428711, 'timing/data': 0.0002453299966873601, 'timing/model': 0.3906126859947108, 'learning_rate/qwen_vl_interface': 5.3291772906398895e-06, 'learning_rate/action_model': 5.32917729063989e-05, 'epoch': 0.18}) +[2026-10-03 13:31:33,736][starVLA.training.train_starvla][INFO] - Step 2525, Loss: {'train/loss': 1.1622867077676347e-06, 'train/grad_norm_pre_clip': 0.0006601798813790083, 'timing/data': 0.0002921900013461709, 'timing/model': 0.38981244699971285, 'learning_rate/qwen_vl_interface': 5.326132276243327e-06, 'learning_rate/action_model': 5.3261322762433266e-05, 'epoch': 0.18}) +[2026-10-03 13:31:34,435][starVLA.training.train_starvla][INFO] - Step 2526, Loss: {'train/loss': 1.4752074548596283e-06, 'train/grad_norm_pre_clip': 0.0007800621096976101, 'timing/data': 0.0002870500029530376, 'timing/model': 0.3913862650078954, 'learning_rate/qwen_vl_interface': 5.323087230551659e-06, 'learning_rate/action_model': 5.323087230551659e-05, 'epoch': 0.18}) +[2026-10-03 13:31:35,325][starVLA.training.train_starvla][INFO] - Step 2527, Loss: {'train/loss': 0.000294280267553404, 'train/grad_norm_pre_clip': 0.14169250428676605, 'timing/data': 0.0002312989963684231, 'timing/model': 0.389863417003653, 'learning_rate/qwen_vl_interface': 5.3200421548165916e-06, 'learning_rate/action_model': 5.3200421548165914e-05, 'epoch': 0.18}) +[2026-10-03 13:31:36,025][starVLA.training.train_starvla][INFO] - Step 2528, Loss: {'train/loss': 0.6347832679748535, 'train/grad_norm_pre_clip': 67.63912200927734, 'timing/data': 0.00028761000430677086, 'timing/model': 0.3917040759988595, 'learning_rate/qwen_vl_interface': 5.3169970502898405e-06, 'learning_rate/action_model': 5.31699705028984e-05, 'epoch': 0.19}) +[2026-10-03 13:31:36,724][starVLA.training.train_starvla][INFO] - Step 2529, Loss: {'train/loss': 0.0012143683852627873, 'train/grad_norm_pre_clip': 1.0855886936187744, 'timing/data': 0.0002934399963123724, 'timing/model': 0.39011995600594673, 'learning_rate/qwen_vl_interface': 5.31395191822313e-06, 'learning_rate/action_model': 5.31395191822313e-05, 'epoch': 0.19}) +[2026-10-03 13:31:37,423][starVLA.training.train_starvla][INFO] - Step 2530, Loss: {'train/loss': 5.0886546887340955e-06, 'train/grad_norm_pre_clip': 0.0023118213284760714, 'timing/data': 0.00026513000193517655, 'timing/model': 0.3913234060019022, 'learning_rate/qwen_vl_interface': 5.310906759868205e-06, 'learning_rate/action_model': 5.310906759868205e-05, 'epoch': 0.19}) +[2026-10-03 13:31:38,120][starVLA.training.train_starvla][INFO] - Step 2531, Loss: {'train/loss': 8.775788592174649e-05, 'train/grad_norm_pre_clip': 0.01864757388830185, 'timing/data': 0.00030218999017961323, 'timing/model': 0.3897184959932929, 'learning_rate/qwen_vl_interface': 5.30786157647681e-06, 'learning_rate/action_model': 5.30786157647681e-05, 'epoch': 0.19}) +[2026-10-03 13:31:39,014][starVLA.training.train_starvla][INFO] - Step 2532, Loss: {'train/loss': 0.07290933281183243, 'train/grad_norm_pre_clip': 227.87254333496094, 'timing/data': 0.00024976000713650137, 'timing/model': 0.5857473439973546, 'learning_rate/qwen_vl_interface': 5.304816369300709e-06, 'learning_rate/action_model': 5.304816369300709e-05, 'epoch': 0.19}) +[2026-10-03 13:31:39,712][starVLA.training.train_starvla][INFO] - Step 2533, Loss: {'train/loss': 2.8748641852871515e-05, 'train/grad_norm_pre_clip': 0.02941933088004589, 'timing/data': 0.0002031600015470758, 'timing/model': 0.3896033269993495, 'learning_rate/qwen_vl_interface': 5.301771139591666e-06, 'learning_rate/action_model': 5.301771139591666e-05, 'epoch': 0.19}) +[2026-10-03 13:31:40,410][starVLA.training.train_starvla][INFO] - Step 2534, Loss: {'train/loss': 0.011961028911173344, 'train/grad_norm_pre_clip': 32.56419372558594, 'timing/data': 0.00027212999702896923, 'timing/model': 0.39101091599150095, 'learning_rate/qwen_vl_interface': 5.298725888601465e-06, 'learning_rate/action_model': 5.298725888601465e-05, 'epoch': 0.19}) +[2026-10-03 13:31:41,106][starVLA.training.train_starvla][INFO] - Step 2535, Loss: {'train/loss': 0.003987136762589216, 'train/grad_norm_pre_clip': 5.628676414489746, 'timing/data': 0.00023316999431699514, 'timing/model': 0.3889975469937781, 'learning_rate/qwen_vl_interface': 5.295680617581891e-06, 'learning_rate/action_model': 5.2956806175818904e-05, 'epoch': 0.19}) +[2026-10-03 13:31:41,806][starVLA.training.train_starvla][INFO] - Step 2536, Loss: {'train/loss': 2.4731160010560416e-05, 'train/grad_norm_pre_clip': 0.017013072967529297, 'timing/data': 0.0003188800037605688, 'timing/model': 0.39152071499847807, 'learning_rate/qwen_vl_interface': 5.2926353277847395e-06, 'learning_rate/action_model': 5.292635327784739e-05, 'epoch': 0.19}) +[2026-10-03 13:31:42,698][starVLA.training.train_starvla][INFO] - Step 2537, Loss: {'train/loss': 5.304760634317063e-06, 'train/grad_norm_pre_clip': 0.0017286394722759724, 'timing/data': 0.0003479000006336719, 'timing/model': 0.582801174998167, 'learning_rate/qwen_vl_interface': 5.2895900204618175e-06, 'learning_rate/action_model': 5.289590020461818e-05, 'epoch': 0.19}) +[2026-10-03 13:31:43,397][starVLA.training.train_starvla][INFO] - Step 2538, Loss: {'train/loss': 0.6770556569099426, 'train/grad_norm_pre_clip': 53.069305419921875, 'timing/data': 0.00027689000125974417, 'timing/model': 0.39184740600467194, 'learning_rate/qwen_vl_interface': 5.286544696864933e-06, 'learning_rate/action_model': 5.2865446968649324e-05, 'epoch': 0.19}) +[2026-10-03 13:31:44,093][starVLA.training.train_starvla][INFO] - Step 2539, Loss: {'train/loss': 2.2445079594035633e-05, 'train/grad_norm_pre_clip': 0.005766434594988823, 'timing/data': 0.00035246000334154814, 'timing/model': 0.38885512699198443, 'learning_rate/qwen_vl_interface': 5.283499358245904e-06, 'learning_rate/action_model': 5.2834993582459034e-05, 'epoch': 0.19}) +[2026-10-03 13:31:44,792][starVLA.training.train_starvla][INFO] - Step 2540, Loss: {'train/loss': 0.3497990369796753, 'train/grad_norm_pre_clip': 116.99715423583984, 'timing/data': 0.00023503000556956977, 'timing/model': 0.39154977700673044, 'learning_rate/qwen_vl_interface': 5.280454005856553e-06, 'learning_rate/action_model': 5.2804540058565534e-05, 'epoch': 0.19}) +[2026-10-03 13:31:45,492][starVLA.training.train_starvla][INFO] - Step 2541, Loss: {'train/loss': 0.0007887305691838264, 'train/grad_norm_pre_clip': 0.38428252935409546, 'timing/data': 0.00021872999786864966, 'timing/model': 0.3904160960082663, 'learning_rate/qwen_vl_interface': 5.277408640948711e-06, 'learning_rate/action_model': 5.277408640948711e-05, 'epoch': 0.19}) +[2026-10-03 13:31:46,190][starVLA.training.train_starvla][INFO] - Step 2542, Loss: {'train/loss': 1.8699320207815617e-05, 'train/grad_norm_pre_clip': 0.011893795803189278, 'timing/data': 0.00027678000333253294, 'timing/model': 0.39127129600092303, 'learning_rate/qwen_vl_interface': 5.274363264774213e-06, 'learning_rate/action_model': 5.274363264774212e-05, 'epoch': 0.19}) +[2026-10-03 13:31:47,079][starVLA.training.train_starvla][INFO] - Step 2543, Loss: {'train/loss': 0.0002900107065215707, 'train/grad_norm_pre_clip': 0.24776838719844818, 'timing/data': 0.0002948100009234622, 'timing/model': 0.3900429459899897, 'learning_rate/qwen_vl_interface': 5.271317878584897e-06, 'learning_rate/action_model': 5.2713178785848974e-05, 'epoch': 0.19}) +[2026-10-03 13:31:47,778][starVLA.training.train_starvla][INFO] - Step 2544, Loss: {'train/loss': 2.8785409085685387e-05, 'train/grad_norm_pre_clip': 0.017640136182308197, 'timing/data': 0.00020407000556588173, 'timing/model': 0.39118635600607377, 'learning_rate/qwen_vl_interface': 5.268272483632606e-06, 'learning_rate/action_model': 5.268272483632606e-05, 'epoch': 0.19}) +[2026-10-03 13:31:48,476][starVLA.training.train_starvla][INFO] - Step 2545, Loss: {'train/loss': 0.01146889477968216, 'train/grad_norm_pre_clip': 13.358266830444336, 'timing/data': 0.00028064000071026385, 'timing/model': 0.39016794599592686, 'learning_rate/qwen_vl_interface': 5.265227081169187e-06, 'learning_rate/action_model': 5.265227081169186e-05, 'epoch': 0.19}) +[2026-10-03 13:31:49,175][starVLA.training.train_starvla][INFO] - Step 2546, Loss: {'train/loss': 4.214211003272794e-05, 'train/grad_norm_pre_clip': 0.00847471784800291, 'timing/data': 0.000244759998167865, 'timing/model': 0.3919578060013009, 'learning_rate/qwen_vl_interface': 5.2621816724464895e-06, 'learning_rate/action_model': 5.262181672446489e-05, 'epoch': 0.19}) +[2026-10-03 13:31:49,872][starVLA.training.train_starvla][INFO] - Step 2547, Loss: {'train/loss': 3.1978779588826e-05, 'train/grad_norm_pre_clip': 0.013271072879433632, 'timing/data': 0.0002741700009210035, 'timing/model': 0.3899086869932944, 'learning_rate/qwen_vl_interface': 5.259136258716367e-06, 'learning_rate/action_model': 5.2591362587163675e-05, 'epoch': 0.19}) +[2026-10-03 13:31:50,762][starVLA.training.train_starvla][INFO] - Step 2548, Loss: {'train/loss': 0.00012469223293010145, 'train/grad_norm_pre_clip': 0.04233231022953987, 'timing/data': 0.00020210999355185777, 'timing/model': 0.5819676940009231, 'learning_rate/qwen_vl_interface': 5.256090841230675e-06, 'learning_rate/action_model': 5.256090841230674e-05, 'epoch': 0.19}) +[2026-10-03 13:31:51,460][starVLA.training.train_starvla][INFO] - Step 2549, Loss: {'train/loss': 0.49100473523139954, 'train/grad_norm_pre_clip': 31.482568740844727, 'timing/data': 0.00023628999770153314, 'timing/model': 0.3898267559998203, 'learning_rate/qwen_vl_interface': 5.253045421241266e-06, 'learning_rate/action_model': 5.253045421241266e-05, 'epoch': 0.19}) +[2026-10-03 13:31:52,160][starVLA.training.train_starvla][INFO] - Step 2550, Loss: {'train/loss': 1.945092844835017e-05, 'train/grad_norm_pre_clip': 0.014952769502997398, 'timing/data': 0.00026824000815395266, 'timing/model': 0.3916947160032578, 'learning_rate/qwen_vl_interface': 5.2500000000000006e-06, 'learning_rate/action_model': 5.25e-05, 'epoch': 0.19}) +[2026-10-03 13:31:52,858][starVLA.training.train_starvla][INFO] - Step 2551, Loss: {'train/loss': 0.7013881802558899, 'train/grad_norm_pre_clip': 78.26538848876953, 'timing/data': 0.00024176000442821532, 'timing/model': 0.39021730600507, 'learning_rate/qwen_vl_interface': 5.246954578758735e-06, 'learning_rate/action_model': 5.246954578758735e-05, 'epoch': 0.19}) +[2026-10-03 13:31:53,557][starVLA.training.train_starvla][INFO] - Step 2552, Loss: {'train/loss': 0.0001773027324816212, 'train/grad_norm_pre_clip': 0.0805213674902916, 'timing/data': 0.0002478300011716783, 'timing/model': 0.3916430469980696, 'learning_rate/qwen_vl_interface': 5.243909158769327e-06, 'learning_rate/action_model': 5.2439091587693264e-05, 'epoch': 0.19}) +[2026-10-03 13:31:54,255][starVLA.training.train_starvla][INFO] - Step 2553, Loss: {'train/loss': 0.06452928483486176, 'train/grad_norm_pre_clip': 27.588302612304688, 'timing/data': 0.0002684099890757352, 'timing/model': 0.3896957959950669, 'learning_rate/qwen_vl_interface': 5.240863741283634e-06, 'learning_rate/action_model': 5.2408637412836336e-05, 'epoch': 0.19}) +[2026-10-03 13:31:55,145][starVLA.training.train_starvla][INFO] - Step 2554, Loss: {'train/loss': 0.0003395517123863101, 'train/grad_norm_pre_clip': 0.20412155985832214, 'timing/data': 0.0002700400073081255, 'timing/model': 0.39089731700369157, 'learning_rate/qwen_vl_interface': 5.2378183275535125e-06, 'learning_rate/action_model': 5.237818327553512e-05, 'epoch': 0.19}) +[2026-10-03 13:31:55,842][starVLA.training.train_starvla][INFO] - Step 2555, Loss: {'train/loss': 0.00020712385594379157, 'train/grad_norm_pre_clip': 0.12723642587661743, 'timing/data': 0.0003497400030028075, 'timing/model': 0.38938199600670487, 'learning_rate/qwen_vl_interface': 5.234772918830815e-06, 'learning_rate/action_model': 5.234772918830815e-05, 'epoch': 0.19}) +[2026-10-03 13:31:56,541][starVLA.training.train_starvla][INFO] - Step 2556, Loss: {'train/loss': 5.2918767323717475e-05, 'train/grad_norm_pre_clip': 0.024611901491880417, 'timing/data': 0.00023451999004464597, 'timing/model': 0.3911403469974175, 'learning_rate/qwen_vl_interface': 5.231727516367395e-06, 'learning_rate/action_model': 5.231727516367395e-05, 'epoch': 0.19}) +[2026-10-03 13:31:57,239][starVLA.training.train_starvla][INFO] - Step 2557, Loss: {'train/loss': 4.343360342318192e-05, 'train/grad_norm_pre_clip': 0.01568933017551899, 'timing/data': 0.00029429999995045364, 'timing/model': 0.3890882470004726, 'learning_rate/qwen_vl_interface': 5.228682121415104e-06, 'learning_rate/action_model': 5.228682121415104e-05, 'epoch': 0.19}) +[2026-10-03 13:31:57,937][starVLA.training.train_starvla][INFO] - Step 2558, Loss: {'train/loss': 7.092730811564252e-06, 'train/grad_norm_pre_clip': 0.0013409702805802226, 'timing/data': 0.0002759100025286898, 'timing/model': 0.39169921599386726, 'learning_rate/qwen_vl_interface': 5.225636735225789e-06, 'learning_rate/action_model': 5.225636735225788e-05, 'epoch': 0.19}) +[2026-10-03 13:31:58,826][starVLA.training.train_starvla][INFO] - Step 2559, Loss: {'train/loss': 8.922778943087906e-05, 'train/grad_norm_pre_clip': 0.05658341944217682, 'timing/data': 0.000281360000371933, 'timing/model': 0.3891802369907964, 'learning_rate/qwen_vl_interface': 5.22259135905129e-06, 'learning_rate/action_model': 5.2225913590512895e-05, 'epoch': 0.19}) +[2026-10-03 13:31:59,525][starVLA.training.train_starvla][INFO] - Step 2560, Loss: {'train/loss': 6.258474627429678e-07, 'train/grad_norm_pre_clip': 0.00020785166998393834, 'timing/data': 0.00023328998940996826, 'timing/model': 0.39118389498617034, 'learning_rate/qwen_vl_interface': 5.219545994143448e-06, 'learning_rate/action_model': 5.219545994143448e-05, 'epoch': 0.19}) +[2026-10-03 13:32:00,224][starVLA.training.train_starvla][INFO] - Step 2561, Loss: {'train/loss': 0.0038657160475850105, 'train/grad_norm_pre_clip': 2.5366556644439697, 'timing/data': 0.00027709999994840473, 'timing/model': 0.3894471070088912, 'learning_rate/qwen_vl_interface': 5.216500641754097e-06, 'learning_rate/action_model': 5.216500641754097e-05, 'epoch': 0.19}) +[2026-10-03 13:32:00,923][starVLA.training.train_starvla][INFO] - Step 2562, Loss: {'train/loss': 0.00022558023920282722, 'train/grad_norm_pre_clip': 0.12831036746501923, 'timing/data': 0.00022962999355513602, 'timing/model': 0.3921871250058757, 'learning_rate/qwen_vl_interface': 5.213455303135067e-06, 'learning_rate/action_model': 5.2134553031350673e-05, 'epoch': 0.19}) +[2026-10-03 13:32:01,620][starVLA.training.train_starvla][INFO] - Step 2563, Loss: {'train/loss': 0.04726003482937813, 'train/grad_norm_pre_clip': 101.19113159179688, 'timing/data': 0.000232470003538765, 'timing/model': 0.38936174599803053, 'learning_rate/qwen_vl_interface': 5.2104099795381836e-06, 'learning_rate/action_model': 5.210409979538183e-05, 'epoch': 0.19}) +[2026-10-03 13:32:02,513][starVLA.training.train_starvla][INFO] - Step 2564, Loss: {'train/loss': 1.966105992323719e-05, 'train/grad_norm_pre_clip': 0.0045094662345945835, 'timing/data': 0.00027892999059986323, 'timing/model': 0.5850209049967816, 'learning_rate/qwen_vl_interface': 5.207364672215261e-06, 'learning_rate/action_model': 5.2073646722152604e-05, 'epoch': 0.19}) +[2026-10-03 13:32:03,211][starVLA.training.train_starvla][INFO] - Step 2565, Loss: {'train/loss': 1.0460205885465257e-05, 'train/grad_norm_pre_clip': 0.0018171640112996101, 'timing/data': 0.00037475999852176756, 'timing/model': 0.38906552600383293, 'learning_rate/qwen_vl_interface': 5.2043193824181095e-06, 'learning_rate/action_model': 5.2043193824181094e-05, 'epoch': 0.19}) +[2026-10-03 13:32:03,909][starVLA.training.train_starvla][INFO] - Step 2566, Loss: {'train/loss': 0.0001936050393851474, 'train/grad_norm_pre_clip': 0.3193819522857666, 'timing/data': 0.00022448999516200274, 'timing/model': 0.3918166259973077, 'learning_rate/qwen_vl_interface': 5.201274111398535e-06, 'learning_rate/action_model': 5.2012741113985356e-05, 'epoch': 0.19}) +[2026-10-03 13:32:04,606][starVLA.training.train_starvla][INFO] - Step 2567, Loss: {'train/loss': 6.545137148350477e-05, 'train/grad_norm_pre_clip': 0.024412736296653748, 'timing/data': 0.00027987000066787004, 'timing/model': 0.3896170259977225, 'learning_rate/qwen_vl_interface': 5.198228860408334e-06, 'learning_rate/action_model': 5.198228860408334e-05, 'epoch': 0.19}) +[2026-10-03 13:32:05,307][starVLA.training.train_starvla][INFO] - Step 2568, Loss: {'train/loss': 0.07229596376419067, 'train/grad_norm_pre_clip': 31.408620834350586, 'timing/data': 0.00023802999930921942, 'timing/model': 0.3913961360085523, 'learning_rate/qwen_vl_interface': 5.195183630699292e-06, 'learning_rate/action_model': 5.195183630699292e-05, 'epoch': 0.19}) +[2026-10-03 13:32:06,005][starVLA.training.train_starvla][INFO] - Step 2569, Loss: {'train/loss': 9.455891995457932e-05, 'train/grad_norm_pre_clip': 0.21854999661445618, 'timing/data': 0.00023623999732080847, 'timing/model': 0.39005047699902207, 'learning_rate/qwen_vl_interface': 5.192138423523189e-06, 'learning_rate/action_model': 5.1921384235231895e-05, 'epoch': 0.19}) +[2026-10-03 13:32:06,896][starVLA.training.train_starvla][INFO] - Step 2570, Loss: {'train/loss': 3.1903582566883415e-05, 'train/grad_norm_pre_clip': 0.01386373583227396, 'timing/data': 0.00025270000332966447, 'timing/model': 0.39133679600490723, 'learning_rate/qwen_vl_interface': 5.189093240131795e-06, 'learning_rate/action_model': 5.189093240131795e-05, 'epoch': 0.19}) +[2026-10-03 13:32:07,591][starVLA.training.train_starvla][INFO] - Step 2571, Loss: {'train/loss': 6.756874063285068e-05, 'train/grad_norm_pre_clip': 0.026637142524123192, 'timing/data': 0.0002033299970207736, 'timing/model': 0.38789996699779294, 'learning_rate/qwen_vl_interface': 5.186048081776869e-06, 'learning_rate/action_model': 5.1860480817768685e-05, 'epoch': 0.19}) +[2026-10-03 13:32:08,290][starVLA.training.train_starvla][INFO] - Step 2572, Loss: {'train/loss': 0.003128797747194767, 'train/grad_norm_pre_clip': 2.598419189453125, 'timing/data': 0.00023909000447019935, 'timing/model': 0.39097945600224193, 'learning_rate/qwen_vl_interface': 5.183002949710161e-06, 'learning_rate/action_model': 5.18300294971016e-05, 'epoch': 0.19}) +[2026-10-03 13:32:08,989][starVLA.training.train_starvla][INFO] - Step 2573, Loss: {'train/loss': 2.287312327098334e-06, 'train/grad_norm_pre_clip': 0.0005252309492789209, 'timing/data': 0.00022924999939277768, 'timing/model': 0.38935975600907113, 'learning_rate/qwen_vl_interface': 5.179957845183411e-06, 'learning_rate/action_model': 5.179957845183411e-05, 'epoch': 0.19}) +[2026-10-03 13:32:09,687][starVLA.training.train_starvla][INFO] - Step 2574, Loss: {'train/loss': 1.0884907245635986, 'train/grad_norm_pre_clip': 84.95133209228516, 'timing/data': 0.0002634799893712625, 'timing/model': 0.3910590759915067, 'learning_rate/qwen_vl_interface': 5.176912769448343e-06, 'learning_rate/action_model': 5.176912769448343e-05, 'epoch': 0.19}) +[2026-10-03 13:32:10,574][starVLA.training.train_starvla][INFO] - Step 2575, Loss: {'train/loss': 8.338447514688596e-05, 'train/grad_norm_pre_clip': 0.04400474578142166, 'timing/data': 0.0002715599985094741, 'timing/model': 0.389539407013217, 'learning_rate/qwen_vl_interface': 5.173867723756676e-06, 'learning_rate/action_model': 5.173867723756676e-05, 'epoch': 0.19}) +[2026-10-03 13:32:11,275][starVLA.training.train_starvla][INFO] - Step 2576, Loss: {'train/loss': 1.2234091758728027, 'train/grad_norm_pre_clip': 234.91818237304688, 'timing/data': 0.00022928000544197857, 'timing/model': 0.3910987470007967, 'learning_rate/qwen_vl_interface': 5.170822709360112e-06, 'learning_rate/action_model': 5.170822709360111e-05, 'epoch': 0.19}) +[2026-10-03 13:32:11,972][starVLA.training.train_starvla][INFO] - Step 2577, Loss: {'train/loss': 2.0322911950643174e-05, 'train/grad_norm_pre_clip': 0.01262756623327732, 'timing/data': 0.00030978998984210193, 'timing/model': 0.38912244599487167, 'learning_rate/qwen_vl_interface': 5.16777772751034e-06, 'learning_rate/action_model': 5.16777772751034e-05, 'epoch': 0.19}) +[2026-10-03 13:32:12,670][starVLA.training.train_starvla][INFO] - Step 2578, Loss: {'train/loss': 1.3434357643127441, 'train/grad_norm_pre_clip': 57.947391510009766, 'timing/data': 0.0002492500061634928, 'timing/model': 0.39128045600955375, 'learning_rate/qwen_vl_interface': 5.16473277945904e-06, 'learning_rate/action_model': 5.1647327794590396e-05, 'epoch': 0.19}) +[2026-10-03 13:32:13,365][starVLA.training.train_starvla][INFO] - Step 2579, Loss: {'train/loss': 0.7238630652427673, 'train/grad_norm_pre_clip': 121.62373352050781, 'timing/data': 0.0002453199995215982, 'timing/model': 0.38856842600216623, 'learning_rate/qwen_vl_interface': 5.161687866457872e-06, 'learning_rate/action_model': 5.161687866457871e-05, 'epoch': 0.19}) +[2026-10-03 13:32:14,254][starVLA.training.train_starvla][INFO] - Step 2580, Loss: {'train/loss': 4.012931458419189e-05, 'train/grad_norm_pre_clip': 0.019739242270588875, 'timing/data': 0.0003064699994865805, 'timing/model': 0.5808530149952276, 'learning_rate/qwen_vl_interface': 5.158642989758487e-06, 'learning_rate/action_model': 5.158642989758487e-05, 'epoch': 0.19}) +[2026-10-03 13:32:14,951][starVLA.training.train_starvla][INFO] - Step 2581, Loss: {'train/loss': 1.5183544746832922e-05, 'train/grad_norm_pre_clip': 0.010167360305786133, 'timing/data': 0.000327879999531433, 'timing/model': 0.38908805599203333, 'learning_rate/qwen_vl_interface': 5.155598150612519e-06, 'learning_rate/action_model': 5.1555981506125185e-05, 'epoch': 0.19}) +[2026-10-03 13:32:15,648][starVLA.training.train_starvla][INFO] - Step 2582, Loss: {'train/loss': 4.334904224378988e-05, 'train/grad_norm_pre_clip': 0.01287376880645752, 'timing/data': 0.0002611199888633564, 'timing/model': 0.39079914600006305, 'learning_rate/qwen_vl_interface': 5.152553350271584e-06, 'learning_rate/action_model': 5.152553350271584e-05, 'epoch': 0.19}) +[2026-10-03 13:32:16,344][starVLA.training.train_starvla][INFO] - Step 2583, Loss: {'train/loss': 0.4356396496295929, 'train/grad_norm_pre_clip': 45.17696762084961, 'timing/data': 0.0002348300040466711, 'timing/model': 0.3886164059949806, 'learning_rate/qwen_vl_interface': 5.149508589987286e-06, 'learning_rate/action_model': 5.1495085899872855e-05, 'epoch': 0.19}) +[2026-10-03 13:32:17,041][starVLA.training.train_starvla][INFO] - Step 2584, Loss: {'train/loss': 0.8826473951339722, 'train/grad_norm_pre_clip': 84.94635772705078, 'timing/data': 0.00031548998958896846, 'timing/model': 0.38934594600868877, 'learning_rate/qwen_vl_interface': 5.14646387101121e-06, 'learning_rate/action_model': 5.14646387101121e-05, 'epoch': 0.19}) +[2026-10-03 13:32:17,931][starVLA.training.train_starvla][INFO] - Step 2585, Loss: {'train/loss': 0.00019361531303729862, 'train/grad_norm_pre_clip': 0.2599470913410187, 'timing/data': 0.00028802998713217676, 'timing/model': 0.5790488749917131, 'learning_rate/qwen_vl_interface': 5.143419194594928e-06, 'learning_rate/action_model': 5.143419194594927e-05, 'epoch': 0.19}) +[2026-10-03 13:32:18,629][starVLA.training.train_starvla][INFO] - Step 2586, Loss: {'train/loss': 1.7224723706021905e-05, 'train/grad_norm_pre_clip': 0.006262119393795729, 'timing/data': 0.0002144799946108833, 'timing/model': 0.3913619149971055, 'learning_rate/qwen_vl_interface': 5.140374561989988e-06, 'learning_rate/action_model': 5.140374561989988e-05, 'epoch': 0.19}) +[2026-10-03 13:32:19,324][starVLA.training.train_starvla][INFO] - Step 2587, Loss: {'train/loss': 0.0004882226639892906, 'train/grad_norm_pre_clip': 0.42166274785995483, 'timing/data': 0.00019732999498955905, 'timing/model': 0.388877567005693, 'learning_rate/qwen_vl_interface': 5.137329974447924e-06, 'learning_rate/action_model': 5.137329974447924e-05, 'epoch': 0.19}) +[2026-10-03 13:32:20,024][starVLA.training.train_starvla][INFO] - Step 2588, Loss: {'train/loss': 0.00011325984087307006, 'train/grad_norm_pre_clip': 0.0422833114862442, 'timing/data': 0.00020212000526953489, 'timing/model': 0.3915954849944683, 'learning_rate/qwen_vl_interface': 5.1342854332202516e-06, 'learning_rate/action_model': 5.1342854332202514e-05, 'epoch': 0.19}) +[2026-10-03 13:32:20,722][starVLA.training.train_starvla][INFO] - Step 2589, Loss: {'train/loss': 0.00029218269628472626, 'train/grad_norm_pre_clip': 0.11261284351348877, 'timing/data': 0.0002682699996512383, 'timing/model': 0.38988061700365506, 'learning_rate/qwen_vl_interface': 5.131240939558467e-06, 'learning_rate/action_model': 5.131240939558467e-05, 'epoch': 0.19}) +[2026-10-03 13:32:21,614][starVLA.training.train_starvla][INFO] - Step 2590, Loss: {'train/loss': 5.1538761908886954e-05, 'train/grad_norm_pre_clip': 0.017909176647663116, 'timing/data': 0.00036148000799585134, 'timing/model': 0.5843534750019899, 'learning_rate/qwen_vl_interface': 5.128196494714048e-06, 'learning_rate/action_model': 5.128196494714047e-05, 'epoch': 0.19}) +[2026-10-03 13:32:22,311][starVLA.training.train_starvla][INFO] - Step 2591, Loss: {'train/loss': 4.3803825974464417e-05, 'train/grad_norm_pre_clip': 0.02611502632498741, 'timing/data': 0.0002877000079024583, 'timing/model': 0.38963855599286035, 'learning_rate/qwen_vl_interface': 5.125152099938448e-06, 'learning_rate/action_model': 5.125152099938447e-05, 'epoch': 0.19}) +[2026-10-03 13:32:23,011][starVLA.training.train_starvla][INFO] - Step 2592, Loss: {'train/loss': 7.696104148635641e-05, 'train/grad_norm_pre_clip': 0.012032038532197475, 'timing/data': 0.00023350000265054405, 'timing/model': 0.39153030700981617, 'learning_rate/qwen_vl_interface': 5.122107756483104e-06, 'learning_rate/action_model': 5.122107756483103e-05, 'epoch': 0.19}) +[2026-10-03 13:32:23,709][starVLA.training.train_starvla][INFO] - Step 2593, Loss: {'train/loss': 1.0236914022243582e-05, 'train/grad_norm_pre_clip': 0.0036112647503614426, 'timing/data': 0.0003367199969943613, 'timing/model': 0.3898321769956965, 'learning_rate/qwen_vl_interface': 5.119063465599429e-06, 'learning_rate/action_model': 5.1190634655994285e-05, 'epoch': 0.19}) +[2026-10-03 13:32:24,409][starVLA.training.train_starvla][INFO] - Step 2594, Loss: {'train/loss': 0.00022164019173942506, 'train/grad_norm_pre_clip': 0.06149685010313988, 'timing/data': 0.00025707999884616584, 'timing/model': 0.39152446600201074, 'learning_rate/qwen_vl_interface': 5.116019228538818e-06, 'learning_rate/action_model': 5.116019228538818e-05, 'epoch': 0.19}) +[2026-10-03 13:32:25,105][starVLA.training.train_starvla][INFO] - Step 2595, Loss: {'train/loss': 0.5811487436294556, 'train/grad_norm_pre_clip': 47.65107345581055, 'timing/data': 0.00025826999626588076, 'timing/model': 0.3892300369916484, 'learning_rate/qwen_vl_interface': 5.112975046552641e-06, 'learning_rate/action_model': 5.1129750465526404e-05, 'epoch': 0.19}) +[2026-10-03 13:32:25,996][starVLA.training.train_starvla][INFO] - Step 2596, Loss: {'train/loss': 7.380267197731882e-05, 'train/grad_norm_pre_clip': 0.032447438687086105, 'timing/data': 0.00032221000583376735, 'timing/model': 0.390984676007065, 'learning_rate/qwen_vl_interface': 5.109930920892246e-06, 'learning_rate/action_model': 5.1099309208922464e-05, 'epoch': 0.19}) +[2026-10-03 13:32:26,694][starVLA.training.train_starvla][INFO] - Step 2597, Loss: {'train/loss': 0.3254348635673523, 'train/grad_norm_pre_clip': 76.53990173339844, 'timing/data': 0.00029424999956972897, 'timing/model': 0.3892189569887705, 'learning_rate/qwen_vl_interface': 5.10688685280896e-06, 'learning_rate/action_model': 5.10688685280896e-05, 'epoch': 0.19}) +[2026-10-03 13:32:27,392][starVLA.training.train_starvla][INFO] - Step 2598, Loss: {'train/loss': 0.017357124015688896, 'train/grad_norm_pre_clip': 88.63565826416016, 'timing/data': 0.00028829899383708835, 'timing/model': 0.3911712770059239, 'learning_rate/qwen_vl_interface': 5.103842843554082e-06, 'learning_rate/action_model': 5.103842843554082e-05, 'epoch': 0.19}) +[2026-10-03 13:32:28,088][starVLA.training.train_starvla][INFO] - Step 2599, Loss: {'train/loss': 0.007938560098409653, 'train/grad_norm_pre_clip': 19.264318466186523, 'timing/data': 0.00026982001145370305, 'timing/model': 0.38888556700840127, 'learning_rate/qwen_vl_interface': 5.100798894378891e-06, 'learning_rate/action_model': 5.100798894378891e-05, 'epoch': 0.19}) +[2026-10-03 13:32:28,788][starVLA.training.train_starvla][INFO] - Step 2600, Loss: {'train/loss': 4.611906115314923e-05, 'train/grad_norm_pre_clip': 0.014291596598923206, 'timing/data': 0.00026361001073382795, 'timing/model': 0.39138963699224405, 'learning_rate/qwen_vl_interface': 5.097755006534638e-06, 'learning_rate/action_model': 5.097755006534638e-05, 'epoch': 0.19}) +[2026-10-03 13:32:29,678][starVLA.training.train_starvla][INFO] - Step 2601, Loss: {'train/loss': 0.5627183318138123, 'train/grad_norm_pre_clip': 75.73068237304688, 'timing/data': 0.00030352998874150217, 'timing/model': 0.5809415949916001, 'learning_rate/qwen_vl_interface': 5.094711181272551e-06, 'learning_rate/action_model': 5.094711181272551e-05, 'epoch': 0.19}) +[2026-10-03 13:32:30,375][starVLA.training.train_starvla][INFO] - Step 2602, Loss: {'train/loss': 0.018757201731204987, 'train/grad_norm_pre_clip': 61.64883804321289, 'timing/data': 0.00027997999859508127, 'timing/model': 0.39110279601300135, 'learning_rate/qwen_vl_interface': 5.091667419843833e-06, 'learning_rate/action_model': 5.091667419843833e-05, 'epoch': 0.19}) +[2026-10-03 13:32:31,071][starVLA.training.train_starvla][INFO] - Step 2603, Loss: {'train/loss': 0.00026455463375896215, 'train/grad_norm_pre_clip': 0.07534467428922653, 'timing/data': 0.00024044999736361206, 'timing/model': 0.3894214759930037, 'learning_rate/qwen_vl_interface': 5.088623723499657e-06, 'learning_rate/action_model': 5.088623723499657e-05, 'epoch': 0.19}) +[2026-10-03 13:32:31,772][starVLA.training.train_starvla][INFO] - Step 2604, Loss: {'train/loss': 0.0006264664698392153, 'train/grad_norm_pre_clip': 0.1253833919763565, 'timing/data': 0.000295000005280599, 'timing/model': 0.3913006770017091, 'learning_rate/qwen_vl_interface': 5.085580093491171e-06, 'learning_rate/action_model': 5.085580093491171e-05, 'epoch': 0.19}) +[2026-10-03 13:32:32,470][starVLA.training.train_starvla][INFO] - Step 2605, Loss: {'train/loss': 0.0013226439477875829, 'train/grad_norm_pre_clip': 1.534213662147522, 'timing/data': 0.00023499999952036887, 'timing/model': 0.38918099699367303, 'learning_rate/qwen_vl_interface': 5.0825365310695e-06, 'learning_rate/action_model': 5.082536531069499e-05, 'epoch': 0.19}) +[2026-10-03 13:32:33,168][starVLA.training.train_starvla][INFO] - Step 2606, Loss: {'train/loss': 0.0006841717404313385, 'train/grad_norm_pre_clip': 0.11993616819381714, 'timing/data': 0.00023438000062014908, 'timing/model': 0.3913448349921964, 'learning_rate/qwen_vl_interface': 5.079493037485735e-06, 'learning_rate/action_model': 5.0794930374857344e-05, 'epoch': 0.19}) +[2026-10-03 13:32:34,053][starVLA.training.train_starvla][INFO] - Step 2607, Loss: {'train/loss': 0.04677338898181915, 'train/grad_norm_pre_clip': 27.12662696838379, 'timing/data': 0.0003613299923017621, 'timing/model': 0.38868015600019135, 'learning_rate/qwen_vl_interface': 5.076449613990941e-06, 'learning_rate/action_model': 5.076449613990941e-05, 'epoch': 0.19}) +[2026-10-03 13:32:34,753][starVLA.training.train_starvla][INFO] - Step 2608, Loss: {'train/loss': 8.048392191994935e-05, 'train/grad_norm_pre_clip': 0.01906590349972248, 'timing/data': 0.0002551399957155809, 'timing/model': 0.39084064700000454, 'learning_rate/qwen_vl_interface': 5.073406261836157e-06, 'learning_rate/action_model': 5.0734062618361565e-05, 'epoch': 0.19}) +[2026-10-03 13:32:35,451][starVLA.training.train_starvla][INFO] - Step 2609, Loss: {'train/loss': 0.01123987790197134, 'train/grad_norm_pre_clip': 18.80831527709961, 'timing/data': 0.0003325400030007586, 'timing/model': 0.3892066669941414, 'learning_rate/qwen_vl_interface': 5.070362982272386e-06, 'learning_rate/action_model': 5.070362982272385e-05, 'epoch': 0.19}) +[2026-10-03 13:32:36,149][starVLA.training.train_starvla][INFO] - Step 2610, Loss: {'train/loss': 0.0007702101720497012, 'train/grad_norm_pre_clip': 0.404786616563797, 'timing/data': 0.00021191999258007854, 'timing/model': 0.3915161360055208, 'learning_rate/qwen_vl_interface': 5.067319776550607e-06, 'learning_rate/action_model': 5.067319776550607e-05, 'epoch': 0.19}) +[2026-10-03 13:32:36,847][starVLA.training.train_starvla][INFO] - Step 2611, Loss: {'train/loss': 0.3863600194454193, 'train/grad_norm_pre_clip': 20.590660095214844, 'timing/data': 0.0003245699917897582, 'timing/model': 0.39067066700954456, 'learning_rate/qwen_vl_interface': 5.064276645921768e-06, 'learning_rate/action_model': 5.064276645921768e-05, 'epoch': 0.19}) +[2026-10-03 13:32:37,740][starVLA.training.train_starvla][INFO] - Step 2612, Loss: {'train/loss': 0.00018347168224863708, 'train/grad_norm_pre_clip': 0.06554870307445526, 'timing/data': 0.0002049200120382011, 'timing/model': 0.39136411699291784, 'learning_rate/qwen_vl_interface': 5.061233591636787e-06, 'learning_rate/action_model': 5.061233591636787e-05, 'epoch': 0.19}) +[2026-10-03 13:32:38,437][starVLA.training.train_starvla][INFO] - Step 2613, Loss: {'train/loss': 0.0004168688028585166, 'train/grad_norm_pre_clip': 0.09974837303161621, 'timing/data': 0.0003090799873461947, 'timing/model': 0.38903313600167166, 'learning_rate/qwen_vl_interface': 5.058190614946544e-06, 'learning_rate/action_model': 5.058190614946544e-05, 'epoch': 0.19}) +[2026-10-03 13:32:39,134][starVLA.training.train_starvla][INFO] - Step 2614, Loss: {'train/loss': 0.04933596029877663, 'train/grad_norm_pre_clip': 35.73961639404297, 'timing/data': 0.00023657998826820403, 'timing/model': 0.3909143570053857, 'learning_rate/qwen_vl_interface': 5.055147717101895e-06, 'learning_rate/action_model': 5.055147717101894e-05, 'epoch': 0.19}) +[2026-10-03 13:32:39,830][starVLA.training.train_starvla][INFO] - Step 2615, Loss: {'train/loss': 0.0002814926265273243, 'train/grad_norm_pre_clip': 0.11441600322723389, 'timing/data': 0.00026467000134289265, 'timing/model': 0.38905989700288046, 'learning_rate/qwen_vl_interface': 5.052104899353658e-06, 'learning_rate/action_model': 5.0521048993536577e-05, 'epoch': 0.19}) +[2026-10-03 13:32:40,530][starVLA.training.train_starvla][INFO] - Step 2616, Loss: {'train/loss': 0.6787511706352234, 'train/grad_norm_pre_clip': 58.75506591796875, 'timing/data': 0.0002823599934345111, 'timing/model': 0.3916605770064052, 'learning_rate/qwen_vl_interface': 5.049062162952622e-06, 'learning_rate/action_model': 5.0490621629526214e-05, 'epoch': 0.19}) +[2026-10-03 13:32:41,422][starVLA.training.train_starvla][INFO] - Step 2617, Loss: {'train/loss': 0.00031373760430142283, 'train/grad_norm_pre_clip': 0.08945727348327637, 'timing/data': 0.00023766999947838485, 'timing/model': 0.5827281149977352, 'learning_rate/qwen_vl_interface': 5.046019509149541e-06, 'learning_rate/action_model': 5.046019509149541e-05, 'epoch': 0.19}) +[2026-10-03 13:32:42,118][starVLA.training.train_starvla][INFO] - Step 2618, Loss: {'train/loss': 8.804010576568544e-05, 'train/grad_norm_pre_clip': 0.039290592074394226, 'timing/data': 0.00027862000570166856, 'timing/model': 0.39066373600508086, 'learning_rate/qwen_vl_interface': 5.042976939195135e-06, 'learning_rate/action_model': 5.0429769391951355e-05, 'epoch': 0.19}) +[2026-10-03 13:32:42,813][starVLA.training.train_starvla][INFO] - Step 2619, Loss: {'train/loss': 0.2623036503791809, 'train/grad_norm_pre_clip': 98.97972106933594, 'timing/data': 0.00025421999453101307, 'timing/model': 0.3883928260038374, 'learning_rate/qwen_vl_interface': 5.039934454340091e-06, 'learning_rate/action_model': 5.0399344543400907e-05, 'epoch': 0.19}) +[2026-10-03 13:32:43,511][starVLA.training.train_starvla][INFO] - Step 2620, Loss: {'train/loss': 0.00019548610725905746, 'train/grad_norm_pre_clip': 0.046553835272789, 'timing/data': 0.00020485000277403742, 'timing/model': 0.3908507370069856, 'learning_rate/qwen_vl_interface': 5.036892055835056e-06, 'learning_rate/action_model': 5.0368920558350566e-05, 'epoch': 0.19}) +[2026-10-03 13:32:44,209][starVLA.training.train_starvla][INFO] - Step 2621, Loss: {'train/loss': 0.0017898764926940203, 'train/grad_norm_pre_clip': 2.5158190727233887, 'timing/data': 0.0002434799971524626, 'timing/model': 0.3889430359995458, 'learning_rate/qwen_vl_interface': 5.033849744930647e-06, 'learning_rate/action_model': 5.033849744930646e-05, 'epoch': 0.19}) +[2026-10-03 13:32:44,907][starVLA.training.train_starvla][INFO] - Step 2622, Loss: {'train/loss': 0.0010512133594602346, 'train/grad_norm_pre_clip': 1.0607659816741943, 'timing/data': 0.00022154999896883965, 'timing/model': 0.39103567600250244, 'learning_rate/qwen_vl_interface': 5.030807522877443e-06, 'learning_rate/action_model': 5.030807522877443e-05, 'epoch': 0.19}) +[2026-10-03 13:32:45,791][starVLA.training.train_starvla][INFO] - Step 2623, Loss: {'train/loss': 0.4801042675971985, 'train/grad_norm_pre_clip': 13.972132682800293, 'timing/data': 0.00028107999241910875, 'timing/model': 0.3896703969949158, 'learning_rate/qwen_vl_interface': 5.027765390925987e-06, 'learning_rate/action_model': 5.027765390925987e-05, 'epoch': 0.19}) +[2026-10-03 13:32:46,492][starVLA.training.train_starvla][INFO] - Step 2624, Loss: {'train/loss': 0.0003767735615838319, 'train/grad_norm_pre_clip': 0.2427082061767578, 'timing/data': 0.0002866400027414784, 'timing/model': 0.39163813600316644, 'learning_rate/qwen_vl_interface': 5.024723350326783e-06, 'learning_rate/action_model': 5.024723350326783e-05, 'epoch': 0.19}) +[2026-10-03 13:32:47,190][starVLA.training.train_starvla][INFO] - Step 2625, Loss: {'train/loss': 1.0148941278457642, 'train/grad_norm_pre_clip': 97.6351089477539, 'timing/data': 0.0003199499915353954, 'timing/model': 0.38946988599491306, 'learning_rate/qwen_vl_interface': 5.021681402330303e-06, 'learning_rate/action_model': 5.021681402330303e-05, 'epoch': 0.19}) +[2026-10-03 13:32:47,888][starVLA.training.train_starvla][INFO] - Step 2626, Loss: {'train/loss': 0.00016942656657192856, 'train/grad_norm_pre_clip': 0.03528362140059471, 'timing/data': 0.0002717000024858862, 'timing/model': 0.3910926259995904, 'learning_rate/qwen_vl_interface': 5.018639548186972e-06, 'learning_rate/action_model': 5.0186395481869716e-05, 'epoch': 0.19}) +[2026-10-03 13:32:48,582][starVLA.training.train_starvla][INFO] - Step 2627, Loss: {'train/loss': 0.05213550850749016, 'train/grad_norm_pre_clip': 65.51143646240234, 'timing/data': 0.000347440000041388, 'timing/model': 0.3888364360027481, 'learning_rate/qwen_vl_interface': 5.015597789147184e-06, 'learning_rate/action_model': 5.015597789147184e-05, 'epoch': 0.19}) +[2026-10-03 13:32:49,468][starVLA.training.train_starvla][INFO] - Step 2628, Loss: {'train/loss': 0.18171219527721405, 'train/grad_norm_pre_clip': 32.281822204589844, 'timing/data': 0.0002529199991840869, 'timing/model': 0.3885077260056278, 'learning_rate/qwen_vl_interface': 5.012556126461291e-06, 'learning_rate/action_model': 5.012556126461291e-05, 'epoch': 0.19}) +[2026-10-03 13:32:50,165][starVLA.training.train_starvla][INFO] - Step 2629, Loss: {'train/loss': 0.002976837567985058, 'train/grad_norm_pre_clip': 5.860576152801514, 'timing/data': 0.0003250500012654811, 'timing/model': 0.3891219170036493, 'learning_rate/qwen_vl_interface': 5.0095145613796044e-06, 'learning_rate/action_model': 5.0095145613796046e-05, 'epoch': 0.19}) +[2026-10-03 13:32:50,862][starVLA.training.train_starvla][INFO] - Step 2630, Loss: {'train/loss': 0.00011881002137670293, 'train/grad_norm_pre_clip': 0.042437776923179626, 'timing/data': 0.0003261899983044714, 'timing/model': 0.3907037360040704, 'learning_rate/qwen_vl_interface': 5.0064730951524e-06, 'learning_rate/action_model': 5.0064730951524e-05, 'epoch': 0.19}) +[2026-10-03 13:32:51,558][starVLA.training.train_starvla][INFO] - Step 2631, Loss: {'train/loss': 0.00010445024236105382, 'train/grad_norm_pre_clip': 0.019770575687289238, 'timing/data': 0.0003126700030406937, 'timing/model': 0.3892186570010381, 'learning_rate/qwen_vl_interface': 5.003431729029905e-06, 'learning_rate/action_model': 5.003431729029905e-05, 'epoch': 0.19}) +[2026-10-03 13:32:52,256][starVLA.training.train_starvla][INFO] - Step 2632, Loss: {'train/loss': 0.29598504304885864, 'train/grad_norm_pre_clip': 57.22650146484375, 'timing/data': 0.00022922000789549202, 'timing/model': 0.39033038599882275, 'learning_rate/qwen_vl_interface': 5.000390464262314e-06, 'learning_rate/action_model': 5.0003904642623136e-05, 'epoch': 0.19}) +[2026-10-03 13:32:53,144][starVLA.training.train_starvla][INFO] - Step 2633, Loss: {'train/loss': 6.347634189296514e-05, 'train/grad_norm_pre_clip': 0.050446704030036926, 'timing/data': 0.0002629899972816929, 'timing/model': 0.3885722369886935, 'learning_rate/qwen_vl_interface': 4.997349302099774e-06, 'learning_rate/action_model': 4.997349302099773e-05, 'epoch': 0.19}) +[2026-10-03 13:32:53,841][starVLA.training.train_starvla][INFO] - Step 2634, Loss: {'train/loss': 0.0006893934914842248, 'train/grad_norm_pre_clip': 0.576082170009613, 'timing/data': 0.00019686999439727515, 'timing/model': 0.39084860699949786, 'learning_rate/qwen_vl_interface': 4.994308243792391e-06, 'learning_rate/action_model': 4.994308243792391e-05, 'epoch': 0.19}) +[2026-10-03 13:32:54,536][starVLA.training.train_starvla][INFO] - Step 2635, Loss: {'train/loss': 0.10402479022741318, 'train/grad_norm_pre_clip': 80.13067626953125, 'timing/data': 0.0002455500070936978, 'timing/model': 0.38863679600763135, 'learning_rate/qwen_vl_interface': 4.991267290590233e-06, 'learning_rate/action_model': 4.991267290590233e-05, 'epoch': 0.19}) +[2026-10-03 13:32:55,233][starVLA.training.train_starvla][INFO] - Step 2636, Loss: {'train/loss': 5.5024447647156194e-05, 'train/grad_norm_pre_clip': 0.014165845699608326, 'timing/data': 0.0002848699950845912, 'timing/model': 0.39027457700285595, 'learning_rate/qwen_vl_interface': 4.988226443743317e-06, 'learning_rate/action_model': 4.9882264437433174e-05, 'epoch': 0.19}) +[2026-10-03 13:32:55,930][starVLA.training.train_starvla][INFO] - Step 2637, Loss: {'train/loss': 0.0005238883313722908, 'train/grad_norm_pre_clip': 0.26007965207099915, 'timing/data': 0.0002414799964753911, 'timing/model': 0.38855567600694485, 'learning_rate/qwen_vl_interface': 4.985185704501623e-06, 'learning_rate/action_model': 4.985185704501623e-05, 'epoch': 0.19}) +[2026-10-03 13:32:56,817][starVLA.training.train_starvla][INFO] - Step 2638, Loss: {'train/loss': 0.0013667980674654245, 'train/grad_norm_pre_clip': 2.795837879180908, 'timing/data': 0.00023936999787110835, 'timing/model': 0.5812184050009819, 'learning_rate/qwen_vl_interface': 4.982145074115082e-06, 'learning_rate/action_model': 4.982145074115082e-05, 'epoch': 0.19}) +[2026-10-03 13:32:57,511][starVLA.training.train_starvla][INFO] - Step 2639, Loss: {'train/loss': 4.735376933240332e-05, 'train/grad_norm_pre_clip': 0.01948375068604946, 'timing/data': 0.00022821000311523676, 'timing/model': 0.38803070700669196, 'learning_rate/qwen_vl_interface': 4.979104553833584e-06, 'learning_rate/action_model': 4.979104553833584e-05, 'epoch': 0.19}) +[2026-10-03 13:32:58,210][starVLA.training.train_starvla][INFO] - Step 2640, Loss: {'train/loss': 7.809490489307791e-05, 'train/grad_norm_pre_clip': 0.02635015733540058, 'timing/data': 0.00032099000236485153, 'timing/model': 0.3902596959960647, 'learning_rate/qwen_vl_interface': 4.9760641449069724e-06, 'learning_rate/action_model': 4.976064144906972e-05, 'epoch': 0.19}) +[2026-10-03 13:32:58,908][starVLA.training.train_starvla][INFO] - Step 2641, Loss: {'train/loss': 0.10548419505357742, 'train/grad_norm_pre_clip': 334.0157165527344, 'timing/data': 0.0003415600076550618, 'timing/model': 0.38892009601113386, 'learning_rate/qwen_vl_interface': 4.973023848585042e-06, 'learning_rate/action_model': 4.973023848585042e-05, 'epoch': 0.19}) +[2026-10-03 13:32:59,605][starVLA.training.train_starvla][INFO] - Step 2642, Loss: {'train/loss': 3.126889714621939e-05, 'train/grad_norm_pre_clip': 0.007852185517549515, 'timing/data': 0.00028558999474626034, 'timing/model': 0.39049026599968784, 'learning_rate/qwen_vl_interface': 4.969983666117545e-06, 'learning_rate/action_model': 4.9699836661175446e-05, 'epoch': 0.19}) +[2026-10-03 13:33:00,492][starVLA.training.train_starvla][INFO] - Step 2643, Loss: {'train/loss': 3.18027196044568e-05, 'train/grad_norm_pre_clip': 0.014920656569302082, 'timing/data': 0.0002433600020594895, 'timing/model': 0.5802352449973114, 'learning_rate/qwen_vl_interface': 4.966943598754185e-06, 'learning_rate/action_model': 4.9669435987541854e-05, 'epoch': 0.19}) +[2026-10-03 13:33:01,192][starVLA.training.train_starvla][INFO] - Step 2644, Loss: {'train/loss': 0.0004683908191509545, 'train/grad_norm_pre_clip': 0.11109214276075363, 'timing/data': 0.00019882999185938388, 'timing/model': 0.3911251359968446, 'learning_rate/qwen_vl_interface': 4.963903647744621e-06, 'learning_rate/action_model': 4.96390364774462e-05, 'epoch': 0.19}) +[2026-10-03 13:33:01,888][starVLA.training.train_starvla][INFO] - Step 2645, Loss: {'train/loss': 1.9765651813941076e-05, 'train/grad_norm_pre_clip': 0.005829538218677044, 'timing/data': 0.0003151499986415729, 'timing/model': 0.3885644459951436, 'learning_rate/qwen_vl_interface': 4.9608638143384604e-06, 'learning_rate/action_model': 4.960863814338461e-05, 'epoch': 0.19}) +[2026-10-03 13:33:02,585][starVLA.training.train_starvla][INFO] - Step 2646, Loss: {'train/loss': 0.03674734756350517, 'train/grad_norm_pre_clip': 77.6036605834961, 'timing/data': 0.0003027200000360608, 'timing/model': 0.3904892860009568, 'learning_rate/qwen_vl_interface': 4.9578240997852636e-06, 'learning_rate/action_model': 4.957824099785264e-05, 'epoch': 0.19}) +[2026-10-03 13:33:03,280][starVLA.training.train_starvla][INFO] - Step 2647, Loss: {'train/loss': 0.0005450170137919486, 'train/grad_norm_pre_clip': 0.3329305052757263, 'timing/data': 0.00018971000099554658, 'timing/model': 0.38893666699004825, 'learning_rate/qwen_vl_interface': 4.954784505334544e-06, 'learning_rate/action_model': 4.9547845053345434e-05, 'epoch': 0.19}) +[2026-10-03 13:33:03,978][starVLA.training.train_starvla][INFO] - Step 2648, Loss: {'train/loss': 0.07634582370519638, 'train/grad_norm_pre_clip': 91.955322265625, 'timing/data': 0.0003189399867551401, 'timing/model': 0.39030721499875654, 'learning_rate/qwen_vl_interface': 4.951745032235762e-06, 'learning_rate/action_model': 4.951745032235762e-05, 'epoch': 0.19}) +[2026-10-03 13:33:04,867][starVLA.training.train_starvla][INFO] - Step 2649, Loss: {'train/loss': 9.130944818025455e-05, 'train/grad_norm_pre_clip': 0.04436253011226654, 'timing/data': 0.0002922099956776947, 'timing/model': 0.38835776700580027, 'learning_rate/qwen_vl_interface': 4.94870568173833e-06, 'learning_rate/action_model': 4.9487056817383307e-05, 'epoch': 0.19}) +[2026-10-03 13:33:05,564][starVLA.training.train_starvla][INFO] - Step 2650, Loss: {'train/loss': 0.001104990253224969, 'train/grad_norm_pre_clip': 1.1507630348205566, 'timing/data': 0.00029106000147294253, 'timing/model': 0.39104118599789217, 'learning_rate/qwen_vl_interface': 4.945666455091614e-06, 'learning_rate/action_model': 4.9456664550916136e-05, 'epoch': 0.19}) +[2026-10-03 13:33:06,260][starVLA.training.train_starvla][INFO] - Step 2651, Loss: {'train/loss': 0.017940638586878777, 'train/grad_norm_pre_clip': 39.285240173339844, 'timing/data': 0.0002166699996450916, 'timing/model': 0.38929452700540423, 'learning_rate/qwen_vl_interface': 4.942627353544921e-06, 'learning_rate/action_model': 4.942627353544921e-05, 'epoch': 0.19}) +[2026-10-03 13:33:06,958][starVLA.training.train_starvla][INFO] - Step 2652, Loss: {'train/loss': 0.10578098893165588, 'train/grad_norm_pre_clip': 66.10047149658203, 'timing/data': 0.00019828000222332776, 'timing/model': 0.39062426499731373, 'learning_rate/qwen_vl_interface': 4.939588378347513e-06, 'learning_rate/action_model': 4.9395883783475125e-05, 'epoch': 0.19}) +[2026-10-03 13:33:07,655][starVLA.training.train_starvla][INFO] - Step 2653, Loss: {'train/loss': 4.139916200074367e-05, 'train/grad_norm_pre_clip': 0.014518337324261665, 'timing/data': 0.00024031000793911517, 'timing/model': 0.3890459170070244, 'learning_rate/qwen_vl_interface': 4.936549530748597e-06, 'learning_rate/action_model': 4.9365495307485963e-05, 'epoch': 0.19}) +[2026-10-03 13:33:08,541][starVLA.training.train_starvla][INFO] - Step 2654, Loss: {'train/loss': 3.3756477932911366e-05, 'train/grad_norm_pre_clip': 0.013057717122137547, 'timing/data': 0.00025986999389715493, 'timing/model': 0.5782916150055826, 'learning_rate/qwen_vl_interface': 4.9335108119973286e-06, 'learning_rate/action_model': 4.933510811997328e-05, 'epoch': 0.19}) +[2026-10-03 13:33:09,235][starVLA.training.train_starvla][INFO] - Step 2655, Loss: {'train/loss': 0.0006293477490544319, 'train/grad_norm_pre_clip': 0.2605171799659729, 'timing/data': 0.0002642699982970953, 'timing/model': 0.38840009600971825, 'learning_rate/qwen_vl_interface': 4.930472223342812e-06, 'learning_rate/action_model': 4.930472223342812e-05, 'epoch': 0.19}) +[2026-10-03 13:33:09,934][starVLA.training.train_starvla][INFO] - Step 2656, Loss: {'train/loss': 3.414232924114913e-05, 'train/grad_norm_pre_clip': 0.015706658363342285, 'timing/data': 0.00033008999889716506, 'timing/model': 0.3903926070051966, 'learning_rate/qwen_vl_interface': 4.927433766034093e-06, 'learning_rate/action_model': 4.9274337660340934e-05, 'epoch': 0.19}) +[2026-10-03 13:33:10,632][starVLA.training.train_starvla][INFO] - Step 2657, Loss: {'train/loss': 0.0005054106586612761, 'train/grad_norm_pre_clip': 0.19090808928012848, 'timing/data': 0.00045433000195771456, 'timing/model': 0.38911473599728197, 'learning_rate/qwen_vl_interface': 4.92439544132017e-06, 'learning_rate/action_model': 4.92439544132017e-05, 'epoch': 0.19}) +[2026-10-03 13:33:11,330][starVLA.training.train_starvla][INFO] - Step 2658, Loss: {'train/loss': 0.04065653309226036, 'train/grad_norm_pre_clip': 44.4018440246582, 'timing/data': 0.0003482100000837818, 'timing/model': 0.3908935960062081, 'learning_rate/qwen_vl_interface': 4.92135725044998e-06, 'learning_rate/action_model': 4.92135725044998e-05, 'epoch': 0.19}) +[2026-10-03 13:33:12,025][starVLA.training.train_starvla][INFO] - Step 2659, Loss: {'train/loss': 7.579737575724721e-05, 'train/grad_norm_pre_clip': 0.024015307426452637, 'timing/data': 0.00019171999883837998, 'timing/model': 0.3890752959996462, 'learning_rate/qwen_vl_interface': 4.918319194672412e-06, 'learning_rate/action_model': 4.918319194672412e-05, 'epoch': 0.19}) +[2026-10-03 13:33:12,912][starVLA.training.train_starvla][INFO] - Step 2660, Loss: {'train/loss': 0.02934976853430271, 'train/grad_norm_pre_clip': 94.31880950927734, 'timing/data': 0.0002344599924981594, 'timing/model': 0.3908667359937681, 'learning_rate/qwen_vl_interface': 4.9152812752362935e-06, 'learning_rate/action_model': 4.915281275236293e-05, 'epoch': 0.19}) +[2026-10-03 13:33:13,610][starVLA.training.train_starvla][INFO] - Step 2661, Loss: {'train/loss': 0.005424535367637873, 'train/grad_norm_pre_clip': 4.8158979415893555, 'timing/data': 0.0002413500042166561, 'timing/model': 0.38911551600904204, 'learning_rate/qwen_vl_interface': 4.9122434933904e-06, 'learning_rate/action_model': 4.9122434933904e-05, 'epoch': 0.19}) +[2026-10-03 13:33:14,307][starVLA.training.train_starvla][INFO] - Step 2662, Loss: {'train/loss': 0.08083085715770721, 'train/grad_norm_pre_clip': 161.55055236816406, 'timing/data': 0.0003238200006308034, 'timing/model': 0.39061811700230464, 'learning_rate/qwen_vl_interface': 4.909205850383449e-06, 'learning_rate/action_model': 4.9092058503834485e-05, 'epoch': 0.19}) +[2026-10-03 13:33:15,003][starVLA.training.train_starvla][INFO] - Step 2663, Loss: {'train/loss': 0.00014256819849833846, 'train/grad_norm_pre_clip': 0.04005945846438408, 'timing/data': 0.00019406998762860894, 'timing/model': 0.38937412700033747, 'learning_rate/qwen_vl_interface': 4.906168347464098e-06, 'learning_rate/action_model': 4.906168347464098e-05, 'epoch': 0.19}) +[2026-10-03 13:33:15,702][starVLA.training.train_starvla][INFO] - Step 2664, Loss: {'train/loss': 0.28608158230781555, 'train/grad_norm_pre_clip': 117.27680206298828, 'timing/data': 0.00024468000628985465, 'timing/model': 0.39131658700353, 'learning_rate/qwen_vl_interface': 4.903130985880955e-06, 'learning_rate/action_model': 4.903130985880954e-05, 'epoch': 0.19}) +[2026-10-03 13:33:16,591][starVLA.training.train_starvla][INFO] - Step 2665, Loss: {'train/loss': 0.7591610550880432, 'train/grad_norm_pre_clip': 82.41912078857422, 'timing/data': 0.00030606999644078314, 'timing/model': 0.3894021860032808, 'learning_rate/qwen_vl_interface': 4.900093766882558e-06, 'learning_rate/action_model': 4.900093766882558e-05, 'epoch': 0.2}) +[2026-10-03 13:33:17,289][starVLA.training.train_starvla][INFO] - Step 2666, Loss: {'train/loss': 0.18118640780448914, 'train/grad_norm_pre_clip': 46.94929885864258, 'timing/data': 0.00029676000121980906, 'timing/model': 0.3906744569976581, 'learning_rate/qwen_vl_interface': 4.897056691717397e-06, 'learning_rate/action_model': 4.8970566917173966e-05, 'epoch': 0.2}) +[2026-10-03 13:33:17,985][starVLA.training.train_starvla][INFO] - Step 2667, Loss: {'train/loss': 0.2437426894903183, 'train/grad_norm_pre_clip': 101.71200561523438, 'timing/data': 0.0001983399997698143, 'timing/model': 0.38953569599834736, 'learning_rate/qwen_vl_interface': 4.894019761633898e-06, 'learning_rate/action_model': 4.894019761633898e-05, 'epoch': 0.2}) +[2026-10-03 13:33:18,684][starVLA.training.train_starvla][INFO] - Step 2668, Loss: {'train/loss': 8.201556192943826e-05, 'train/grad_norm_pre_clip': 0.022149186581373215, 'timing/data': 0.0002450199972372502, 'timing/model': 0.39102258600178175, 'learning_rate/qwen_vl_interface': 4.8909829778804304e-06, 'learning_rate/action_model': 4.89098297788043e-05, 'epoch': 0.2}) +[2026-10-03 13:33:19,383][starVLA.training.train_starvla][INFO] - Step 2669, Loss: {'train/loss': 2.7200181648368016e-05, 'train/grad_norm_pre_clip': 0.012768874876201153, 'timing/data': 0.00028449000092223287, 'timing/model': 0.3902409370057285, 'learning_rate/qwen_vl_interface': 4.887946341705298e-06, 'learning_rate/action_model': 4.887946341705297e-05, 'epoch': 0.2}) +[2026-10-03 13:33:20,270][starVLA.training.train_starvla][INFO] - Step 2670, Loss: {'train/loss': 0.5701671838760376, 'train/grad_norm_pre_clip': 15.316258430480957, 'timing/data': 0.0001932700106408447, 'timing/model': 0.5807195649977075, 'learning_rate/qwen_vl_interface': 4.8849098543567465e-06, 'learning_rate/action_model': 4.884909854356747e-05, 'epoch': 0.2}) +[2026-10-03 13:33:20,965][starVLA.training.train_starvla][INFO] - Step 2671, Loss: {'train/loss': 7.128441211534664e-05, 'train/grad_norm_pre_clip': 0.032826222479343414, 'timing/data': 0.0002729099942371249, 'timing/model': 0.38926309598900843, 'learning_rate/qwen_vl_interface': 4.881873517082965e-06, 'learning_rate/action_model': 4.8818735170829655e-05, 'epoch': 0.2}) +[2026-10-03 13:33:21,665][starVLA.training.train_starvla][INFO] - Step 2672, Loss: {'train/loss': 0.0001824954233597964, 'train/grad_norm_pre_clip': 0.1364946812391281, 'timing/data': 0.0003433300007600337, 'timing/model': 0.39072439599840436, 'learning_rate/qwen_vl_interface': 4.878837331132076e-06, 'learning_rate/action_model': 4.8788373311320754e-05, 'epoch': 0.2}) +[2026-10-03 13:33:22,363][starVLA.training.train_starvla][INFO] - Step 2673, Loss: {'train/loss': 3.125162038486451e-05, 'train/grad_norm_pre_clip': 0.01521850936114788, 'timing/data': 0.0002404100087005645, 'timing/model': 0.3895970669982489, 'learning_rate/qwen_vl_interface': 4.87580129775214e-06, 'learning_rate/action_model': 4.87580129775214e-05, 'epoch': 0.2}) +[2026-10-03 13:33:23,060][starVLA.training.train_starvla][INFO] - Step 2674, Loss: {'train/loss': 5.0934238970512524e-05, 'train/grad_norm_pre_clip': 0.012054737657308578, 'timing/data': 0.00025985999673139304, 'timing/model': 0.3912930660007987, 'learning_rate/qwen_vl_interface': 4.872765418191156e-06, 'learning_rate/action_model': 4.872765418191155e-05, 'epoch': 0.2}) +[2026-10-03 13:33:23,757][starVLA.training.train_starvla][INFO] - Step 2675, Loss: {'train/loss': 0.5203354954719543, 'train/grad_norm_pre_clip': 71.47965240478516, 'timing/data': 0.00027181999757885933, 'timing/model': 0.3897512769908644, 'learning_rate/qwen_vl_interface': 4.8697296936970575e-06, 'learning_rate/action_model': 4.869729693697057e-05, 'epoch': 0.2}) +[2026-10-03 13:33:24,643][starVLA.training.train_starvla][INFO] - Step 2676, Loss: {'train/loss': 0.015989670529961586, 'train/grad_norm_pre_clip': 57.315650939941406, 'timing/data': 0.00026575000083539635, 'timing/model': 0.39112238699453883, 'learning_rate/qwen_vl_interface': 4.8666941255177185e-06, 'learning_rate/action_model': 4.8666941255177185e-05, 'epoch': 0.2}) +[2026-10-03 13:33:25,341][starVLA.training.train_starvla][INFO] - Step 2677, Loss: {'train/loss': 3.440900400164537e-05, 'train/grad_norm_pre_clip': 0.010805006138980389, 'timing/data': 0.0003437300038058311, 'timing/model': 0.3897569270047825, 'learning_rate/qwen_vl_interface': 4.863658714900945e-06, 'learning_rate/action_model': 4.863658714900945e-05, 'epoch': 0.2}) +[2026-10-03 13:33:26,039][starVLA.training.train_starvla][INFO] - Step 2678, Loss: {'train/loss': 1.935595355462283e-05, 'train/grad_norm_pre_clip': 0.004993116483092308, 'timing/data': 0.00020848998974543065, 'timing/model': 0.39085687599435914, 'learning_rate/qwen_vl_interface': 4.860623463094483e-06, 'learning_rate/action_model': 4.860623463094483e-05, 'epoch': 0.2}) +[2026-10-03 13:33:26,733][starVLA.training.train_starvla][INFO] - Step 2679, Loss: {'train/loss': 0.00020199020218569785, 'train/grad_norm_pre_clip': 0.03344389796257019, 'timing/data': 0.00019500900816638023, 'timing/model': 0.38920503699046094, 'learning_rate/qwen_vl_interface': 4.857588371346006e-06, 'learning_rate/action_model': 4.8575883713460057e-05, 'epoch': 0.2}) +[2026-10-03 13:33:27,433][starVLA.training.train_starvla][INFO] - Step 2680, Loss: {'train/loss': 0.1917124092578888, 'train/grad_norm_pre_clip': 111.84486389160156, 'timing/data': 0.0002839899971149862, 'timing/model': 0.3905357759940671, 'learning_rate/qwen_vl_interface': 4.854553440903127e-06, 'learning_rate/action_model': 4.854553440903126e-05, 'epoch': 0.2}) +[2026-10-03 13:33:28,322][starVLA.training.train_starvla][INFO] - Step 2681, Loss: {'train/loss': 3.797230601776391e-05, 'train/grad_norm_pre_clip': 0.013440987095236778, 'timing/data': 0.00020201000734232366, 'timing/model': 0.38963942699774634, 'learning_rate/qwen_vl_interface': 4.85151867301339e-06, 'learning_rate/action_model': 4.85151867301339e-05, 'epoch': 0.2}) +[2026-10-03 13:33:29,020][starVLA.training.train_starvla][INFO] - Step 2682, Loss: {'train/loss': 0.17973493039608002, 'train/grad_norm_pre_clip': 51.327850341796875, 'timing/data': 0.0002772599982563406, 'timing/model': 0.3910597750073066, 'learning_rate/qwen_vl_interface': 4.848484068924274e-06, 'learning_rate/action_model': 4.848484068924275e-05, 'epoch': 0.2}) +[2026-10-03 13:33:29,716][starVLA.training.train_starvla][INFO] - Step 2683, Loss: {'train/loss': 0.04487857595086098, 'train/grad_norm_pre_clip': 18.649410247802734, 'timing/data': 0.00029551000625360757, 'timing/model': 0.3892292070086114, 'learning_rate/qwen_vl_interface': 4.845449629883193e-06, 'learning_rate/action_model': 4.8454496298831934e-05, 'epoch': 0.2}) +[2026-10-03 13:33:30,415][starVLA.training.train_starvla][INFO] - Step 2684, Loss: {'train/loss': 0.3142026364803314, 'train/grad_norm_pre_clip': 65.79435729980469, 'timing/data': 0.0002482599957147613, 'timing/model': 0.3910839360032696, 'learning_rate/qwen_vl_interface': 4.842415357137487e-06, 'learning_rate/action_model': 4.842415357137487e-05, 'epoch': 0.2}) +[2026-10-03 13:33:31,112][starVLA.training.train_starvla][INFO] - Step 2685, Loss: {'train/loss': 0.00035321441828273237, 'train/grad_norm_pre_clip': 0.05167645588517189, 'timing/data': 0.0002515500091249123, 'timing/model': 0.38922101599746384, 'learning_rate/qwen_vl_interface': 4.839381251934432e-06, 'learning_rate/action_model': 4.8393812519344316e-05, 'epoch': 0.2}) +[2026-10-03 13:33:32,000][starVLA.training.train_starvla][INFO] - Step 2686, Loss: {'train/loss': 0.0001902607618831098, 'train/grad_norm_pre_clip': 0.028253603726625443, 'timing/data': 0.00020202000450808555, 'timing/model': 0.3912798359961016, 'learning_rate/qwen_vl_interface': 4.836347315521233e-06, 'learning_rate/action_model': 4.836347315521233e-05, 'epoch': 0.2}) +[2026-10-03 13:33:32,696][starVLA.training.train_starvla][INFO] - Step 2687, Loss: {'train/loss': 0.008545317687094212, 'train/grad_norm_pre_clip': 16.184301376342773, 'timing/data': 0.00030030999914743006, 'timing/model': 0.3889604069991037, 'learning_rate/qwen_vl_interface': 4.833313549145029e-06, 'learning_rate/action_model': 4.833313549145029e-05, 'epoch': 0.2}) +[2026-10-03 13:33:33,395][starVLA.training.train_starvla][INFO] - Step 2688, Loss: {'train/loss': 7.4803424467972945e-06, 'train/grad_norm_pre_clip': 0.0019328509224578738, 'timing/data': 0.000334700001985766, 'timing/model': 0.39088191599876154, 'learning_rate/qwen_vl_interface': 4.830279954052884e-06, 'learning_rate/action_model': 4.8302799540528836e-05, 'epoch': 0.2}) +[2026-10-03 13:33:34,091][starVLA.training.train_starvla][INFO] - Step 2689, Loss: {'train/loss': 0.00020217431301716715, 'train/grad_norm_pre_clip': 0.132570281624794, 'timing/data': 0.0002509500045562163, 'timing/model': 0.38765303599939216, 'learning_rate/qwen_vl_interface': 4.827246531491797e-06, 'learning_rate/action_model': 4.827246531491797e-05, 'epoch': 0.2}) +[2026-10-03 13:33:34,790][starVLA.training.train_starvla][INFO] - Step 2690, Loss: {'train/loss': 0.008608430624008179, 'train/grad_norm_pre_clip': 12.101802825927734, 'timing/data': 0.0001969299919437617, 'timing/model': 0.3913922560022911, 'learning_rate/qwen_vl_interface': 4.824213282708691e-06, 'learning_rate/action_model': 4.824213282708691e-05, 'epoch': 0.2}) +[2026-10-03 13:33:35,673][starVLA.training.train_starvla][INFO] - Step 2691, Loss: {'train/loss': 0.4794062077999115, 'train/grad_norm_pre_clip': 111.70548248291016, 'timing/data': 0.00019175000488758087, 'timing/model': 0.5773355350102065, 'learning_rate/qwen_vl_interface': 4.821180208950422e-06, 'learning_rate/action_model': 4.821180208950421e-05, 'epoch': 0.2}) +[2026-10-03 13:33:36,372][starVLA.training.train_starvla][INFO] - Step 2692, Loss: {'train/loss': 0.017986251041293144, 'train/grad_norm_pre_clip': 9.670065879821777, 'timing/data': 0.0002904900029534474, 'timing/model': 0.39073933599865995, 'learning_rate/qwen_vl_interface': 4.81814731146377e-06, 'learning_rate/action_model': 4.8181473114637695e-05, 'epoch': 0.2}) +[2026-10-03 13:33:37,070][starVLA.training.train_starvla][INFO] - Step 2693, Loss: {'train/loss': 0.00733567401766777, 'train/grad_norm_pre_clip': 7.796245098114014, 'timing/data': 0.0003365099983057007, 'timing/model': 0.38933383699622937, 'learning_rate/qwen_vl_interface': 4.8151145914954465e-06, 'learning_rate/action_model': 4.815114591495446e-05, 'epoch': 0.2}) +[2026-10-03 13:33:37,768][starVLA.training.train_starvla][INFO] - Step 2694, Loss: {'train/loss': 2.0115992811042815e-05, 'train/grad_norm_pre_clip': 0.00559275783598423, 'timing/data': 0.0001951000012923032, 'timing/model': 0.3914234660041984, 'learning_rate/qwen_vl_interface': 4.812082050292088e-06, 'learning_rate/action_model': 4.8120820502920876e-05, 'epoch': 0.2}) +[2026-10-03 13:33:38,463][starVLA.training.train_starvla][INFO] - Step 2695, Loss: {'train/loss': 0.01777571067214012, 'train/grad_norm_pre_clip': 17.137378692626953, 'timing/data': 0.00029488999280147254, 'timing/model': 0.38872432700009085, 'learning_rate/qwen_vl_interface': 4.809049689100256e-06, 'learning_rate/action_model': 4.8090496891002555e-05, 'epoch': 0.2}) +[2026-10-03 13:33:39,354][starVLA.training.train_starvla][INFO] - Step 2696, Loss: {'train/loss': 0.0001131084281951189, 'train/grad_norm_pre_clip': 0.052078817039728165, 'timing/data': 0.0003618900082074106, 'timing/model': 0.5821790349873481, 'learning_rate/qwen_vl_interface': 4.80601750916644e-06, 'learning_rate/action_model': 4.8060175091664404e-05, 'epoch': 0.2}) +[2026-10-03 13:33:40,053][starVLA.training.train_starvla][INFO] - Step 2697, Loss: {'train/loss': 2.6151490146730794e-06, 'train/grad_norm_pre_clip': 0.0007213271455839276, 'timing/data': 0.0003127800009679049, 'timing/model': 0.38940928599913605, 'learning_rate/qwen_vl_interface': 4.802985511737057e-06, 'learning_rate/action_model': 4.802985511737057e-05, 'epoch': 0.2}) +[2026-10-03 13:33:40,749][starVLA.training.train_starvla][INFO] - Step 2698, Loss: {'train/loss': 0.0006668604910373688, 'train/grad_norm_pre_clip': 0.19224074482917786, 'timing/data': 0.00033660999906715006, 'timing/model': 0.3894510569953127, 'learning_rate/qwen_vl_interface': 4.799953698058444e-06, 'learning_rate/action_model': 4.7999536980584434e-05, 'epoch': 0.2}) +[2026-10-03 13:33:41,446][starVLA.training.train_starvla][INFO] - Step 2699, Loss: {'train/loss': 0.44505763053894043, 'train/grad_norm_pre_clip': 80.32991027832031, 'timing/data': 0.00026173000514972955, 'timing/model': 0.3895886359969154, 'learning_rate/qwen_vl_interface': 4.796922069376867e-06, 'learning_rate/action_model': 4.796922069376867e-05, 'epoch': 0.2}) +[2026-10-03 13:33:42,145][starVLA.training.train_starvla][INFO] - Step 2700, Loss: {'train/loss': 0.02710793726146221, 'train/grad_norm_pre_clip': 39.798702239990234, 'timing/data': 0.0002577800041763112, 'timing/model': 0.3908766769891372, 'learning_rate/qwen_vl_interface': 4.793890626938511e-06, 'learning_rate/action_model': 4.793890626938511e-05, 'epoch': 0.2}) +[2026-10-03 13:33:42,842][starVLA.training.train_starvla][INFO] - Step 2701, Loss: {'train/loss': 0.0022700424306094646, 'train/grad_norm_pre_clip': 2.742307662963867, 'timing/data': 0.0002042800042545423, 'timing/model': 0.3892979170050239, 'learning_rate/qwen_vl_interface': 4.7908593719894914e-06, 'learning_rate/action_model': 4.790859371989491e-05, 'epoch': 0.2}) +[2026-10-03 13:33:43,732][starVLA.training.train_starvla][INFO] - Step 2702, Loss: {'train/loss': 0.000248576863668859, 'train/grad_norm_pre_clip': 0.22470098733901978, 'timing/data': 0.00027407999732531607, 'timing/model': 0.39076429601118434, 'learning_rate/qwen_vl_interface': 4.78782830577584e-06, 'learning_rate/action_model': 4.7878283057758396e-05, 'epoch': 0.2}) +[2026-10-03 13:33:44,427][starVLA.training.train_starvla][INFO] - Step 2703, Loss: {'train/loss': 0.00038595442310906947, 'train/grad_norm_pre_clip': 0.19418014585971832, 'timing/data': 0.00031945000228006393, 'timing/model': 0.38756322700646706, 'learning_rate/qwen_vl_interface': 4.784797429543513e-06, 'learning_rate/action_model': 4.784797429543513e-05, 'epoch': 0.2}) +[2026-10-03 13:33:45,125][starVLA.training.train_starvla][INFO] - Step 2704, Loss: {'train/loss': 0.000748760940041393, 'train/grad_norm_pre_clip': 0.17066989839076996, 'timing/data': 0.00026746001094579697, 'timing/model': 0.3896404370025266, 'learning_rate/qwen_vl_interface': 4.781766744538393e-06, 'learning_rate/action_model': 4.781766744538393e-05, 'epoch': 0.2}) +[2026-10-03 13:33:45,822][starVLA.training.train_starvla][INFO] - Step 2705, Loss: {'train/loss': 1.5422321666846983e-05, 'train/grad_norm_pre_clip': 0.0037891664542257786, 'timing/data': 0.0002593300014268607, 'timing/model': 0.3891557070019189, 'learning_rate/qwen_vl_interface': 4.778736252006277e-06, 'learning_rate/action_model': 4.7787362520062765e-05, 'epoch': 0.2}) +[2026-10-03 13:33:46,520][starVLA.training.train_starvla][INFO] - Step 2706, Loss: {'train/loss': 2.054064862022642e-05, 'train/grad_norm_pre_clip': 0.004430660977959633, 'timing/data': 0.00019397000141907483, 'timing/model': 0.391403846006142, 'learning_rate/qwen_vl_interface': 4.775705953192885e-06, 'learning_rate/action_model': 4.7757059531928856e-05, 'epoch': 0.2}) +[2026-10-03 13:33:47,405][starVLA.training.train_starvla][INFO] - Step 2707, Loss: {'train/loss': 4.542845272226259e-05, 'train/grad_norm_pre_clip': 0.02396433614194393, 'timing/data': 0.00023004999093245715, 'timing/model': 0.5781195450108498, 'learning_rate/qwen_vl_interface': 4.7726758493438615e-06, 'learning_rate/action_model': 4.7726758493438613e-05, 'epoch': 0.2}) +[2026-10-03 13:33:48,105][starVLA.training.train_starvla][INFO] - Step 2708, Loss: {'train/loss': 0.0005561340949498117, 'train/grad_norm_pre_clip': 0.6211398243904114, 'timing/data': 0.0002381999947829172, 'timing/model': 0.39097458600008395, 'learning_rate/qwen_vl_interface': 4.769645941704766e-06, 'learning_rate/action_model': 4.769645941704766e-05, 'epoch': 0.2}) +[2026-10-03 13:33:48,802][starVLA.training.train_starvla][INFO] - Step 2709, Loss: {'train/loss': 3.460481821093708e-05, 'train/grad_norm_pre_clip': 0.011038674041628838, 'timing/data': 0.00023826000688131899, 'timing/model': 0.38908328699471895, 'learning_rate/qwen_vl_interface': 4.766616231521081e-06, 'learning_rate/action_model': 4.7666162315210806e-05, 'epoch': 0.2}) +[2026-10-03 13:33:49,500][starVLA.training.train_starvla][INFO] - Step 2710, Loss: {'train/loss': 5.111067821417237e-06, 'train/grad_norm_pre_clip': 0.0010058408370241523, 'timing/data': 0.000282260007224977, 'timing/model': 0.3912697359919548, 'learning_rate/qwen_vl_interface': 4.763586720038202e-06, 'learning_rate/action_model': 4.7635867200382014e-05, 'epoch': 0.2}) +[2026-10-03 13:33:50,196][starVLA.training.train_starvla][INFO] - Step 2711, Loss: {'train/loss': 0.005738651845604181, 'train/grad_norm_pre_clip': 9.957094192504883, 'timing/data': 0.0002701299963518977, 'timing/model': 0.3888738670066232, 'learning_rate/qwen_vl_interface': 4.76055740850145e-06, 'learning_rate/action_model': 4.76055740850145e-05, 'epoch': 0.2}) +[2026-10-03 13:33:50,895][starVLA.training.train_starvla][INFO] - Step 2712, Loss: {'train/loss': 0.00017184963508043438, 'train/grad_norm_pre_clip': 0.04582512751221657, 'timing/data': 0.0002870499884011224, 'timing/model': 0.3907957460032776, 'learning_rate/qwen_vl_interface': 4.757528298156058e-06, 'learning_rate/action_model': 4.757528298156058e-05, 'epoch': 0.2}) +[2026-10-03 13:33:51,778][starVLA.training.train_starvla][INFO] - Step 2713, Loss: {'train/loss': 0.0002151051303371787, 'train/grad_norm_pre_clip': 0.0326942503452301, 'timing/data': 0.00019858998712152243, 'timing/model': 0.38879707700107247, 'learning_rate/qwen_vl_interface': 4.754499390247181e-06, 'learning_rate/action_model': 4.754499390247181e-05, 'epoch': 0.2}) +[2026-10-03 13:33:52,475][starVLA.training.train_starvla][INFO] - Step 2714, Loss: {'train/loss': 0.00017919160018209368, 'train/grad_norm_pre_clip': 0.023207135498523712, 'timing/data': 0.0002488799946149811, 'timing/model': 0.3905242069886299, 'learning_rate/qwen_vl_interface': 4.751470686019887e-06, 'learning_rate/action_model': 4.751470686019888e-05, 'epoch': 0.2}) +[2026-10-03 13:33:53,169][starVLA.training.train_starvla][INFO] - Step 2715, Loss: {'train/loss': 0.0006098903249949217, 'train/grad_norm_pre_clip': 0.6531382203102112, 'timing/data': 0.00019272000645287335, 'timing/model': 0.3882963159994688, 'learning_rate/qwen_vl_interface': 4.7484421867191645e-06, 'learning_rate/action_model': 4.748442186719165e-05, 'epoch': 0.2}) +[2026-10-03 13:33:53,867][starVLA.training.train_starvla][INFO] - Step 2716, Loss: {'train/loss': 0.2265302836894989, 'train/grad_norm_pre_clip': 736.7713012695312, 'timing/data': 0.00019918999169021845, 'timing/model': 0.39051955699687824, 'learning_rate/qwen_vl_interface': 4.745413893589913e-06, 'learning_rate/action_model': 4.7454138935899125e-05, 'epoch': 0.2}) +[2026-10-03 13:33:54,564][starVLA.training.train_starvla][INFO] - Step 2717, Loss: {'train/loss': 0.0003679986111819744, 'train/grad_norm_pre_clip': 0.8238988518714905, 'timing/data': 0.00029233899840619415, 'timing/model': 0.38891733600758016, 'learning_rate/qwen_vl_interface': 4.7423858078769506e-06, 'learning_rate/action_model': 4.742385807876951e-05, 'epoch': 0.2}) +[2026-10-03 13:33:55,450][starVLA.training.train_starvla][INFO] - Step 2718, Loss: {'train/loss': 0.0007854202995076776, 'train/grad_norm_pre_clip': 1.0003039836883545, 'timing/data': 0.0002611499949125573, 'timing/model': 0.3900402059953194, 'learning_rate/qwen_vl_interface': 4.739357930825008e-06, 'learning_rate/action_model': 4.739357930825008e-05, 'epoch': 0.2}) +[2026-10-03 13:33:56,144][starVLA.training.train_starvla][INFO] - Step 2719, Loss: {'train/loss': 0.00010027170355897397, 'train/grad_norm_pre_clip': 0.014744744636118412, 'timing/data': 0.00029984000138938427, 'timing/model': 0.3882348670013016, 'learning_rate/qwen_vl_interface': 4.736330263678731e-06, 'learning_rate/action_model': 4.736330263678731e-05, 'epoch': 0.2}) +[2026-10-03 13:33:56,843][starVLA.training.train_starvla][INFO] - Step 2720, Loss: {'train/loss': 0.0028692036867141724, 'train/grad_norm_pre_clip': 6.478521823883057, 'timing/data': 0.000365039988537319, 'timing/model': 0.39171193599759135, 'learning_rate/qwen_vl_interface': 4.733302807682679e-06, 'learning_rate/action_model': 4.733302807682679e-05, 'epoch': 0.2}) +[2026-10-03 13:33:57,540][starVLA.training.train_starvla][INFO] - Step 2721, Loss: {'train/loss': 2.2252694179769605e-05, 'train/grad_norm_pre_clip': 0.014288017526268959, 'timing/data': 0.0004600099928211421, 'timing/model': 0.38879466700018384, 'learning_rate/qwen_vl_interface': 4.7302755640813255e-06, 'learning_rate/action_model': 4.7302755640813256e-05, 'epoch': 0.2}) +[2026-10-03 13:33:58,237][starVLA.training.train_starvla][INFO] - Step 2722, Loss: {'train/loss': 0.00016356847481802106, 'train/grad_norm_pre_clip': 0.023467164486646652, 'timing/data': 0.0001986599963856861, 'timing/model': 0.39070894599717576, 'learning_rate/qwen_vl_interface': 4.727248534119058e-06, 'learning_rate/action_model': 4.7272485341190574e-05, 'epoch': 0.2}) +[2026-10-03 13:33:59,123][starVLA.training.train_starvla][INFO] - Step 2723, Loss: {'train/loss': 0.0005106855533085763, 'train/grad_norm_pre_clip': 0.7860918641090393, 'timing/data': 0.00019449999672360718, 'timing/model': 0.5801906560081989, 'learning_rate/qwen_vl_interface': 4.72422171904017e-06, 'learning_rate/action_model': 4.7242217190401697e-05, 'epoch': 0.2}) +[2026-10-03 13:33:59,821][starVLA.training.train_starvla][INFO] - Step 2724, Loss: {'train/loss': 4.2020978980872314e-06, 'train/grad_norm_pre_clip': 0.001264590653590858, 'timing/data': 0.00030734000029042363, 'timing/model': 0.390434255998116, 'learning_rate/qwen_vl_interface': 4.721195120088873e-06, 'learning_rate/action_model': 4.721195120088873e-05, 'epoch': 0.2}) +[2026-10-03 13:34:00,518][starVLA.training.train_starvla][INFO] - Step 2725, Loss: {'train/loss': 0.00019158337090630084, 'train/grad_norm_pre_clip': 0.02801414765417576, 'timing/data': 0.00019828999938908964, 'timing/model': 0.3893248370004585, 'learning_rate/qwen_vl_interface': 4.718168738509288e-06, 'learning_rate/action_model': 4.718168738509288e-05, 'epoch': 0.2}) +[2026-10-03 13:34:01,215][starVLA.training.train_starvla][INFO] - Step 2726, Loss: {'train/loss': 3.658193236333318e-06, 'train/grad_norm_pre_clip': 0.0011190318036824465, 'timing/data': 0.00029289000667631626, 'timing/model': 0.3903864460007753, 'learning_rate/qwen_vl_interface': 4.715142575545445e-06, 'learning_rate/action_model': 4.7151425755454454e-05, 'epoch': 0.2}) +[2026-10-03 13:34:01,910][starVLA.training.train_starvla][INFO] - Step 2727, Loss: {'train/loss': 0.00011548704060260206, 'train/grad_norm_pre_clip': 0.03241647034883499, 'timing/data': 0.00019617000361904502, 'timing/model': 0.3884067170001799, 'learning_rate/qwen_vl_interface': 4.712116632441287e-06, 'learning_rate/action_model': 4.7121166324412866e-05, 'epoch': 0.2}) +[2026-10-03 13:34:02,609][starVLA.training.train_starvla][INFO] - Step 2728, Loss: {'train/loss': 0.00041956081986427307, 'train/grad_norm_pre_clip': 0.2441207319498062, 'timing/data': 0.0004326499911257997, 'timing/model': 0.39033060599467717, 'learning_rate/qwen_vl_interface': 4.709090910440664e-06, 'learning_rate/action_model': 4.7090909104406634e-05, 'epoch': 0.2}) +[2026-10-03 13:34:03,493][starVLA.training.train_starvla][INFO] - Step 2729, Loss: {'train/loss': 4.184016870567575e-05, 'train/grad_norm_pre_clip': 0.009841653518378735, 'timing/data': 0.0002462800039211288, 'timing/model': 0.3887552770029288, 'learning_rate/qwen_vl_interface': 4.7060654107873335e-06, 'learning_rate/action_model': 4.7060654107873335e-05, 'epoch': 0.2}) +[2026-10-03 13:34:04,190][starVLA.training.train_starvla][INFO] - Step 2730, Loss: {'train/loss': 0.4401351511478424, 'train/grad_norm_pre_clip': 153.876220703125, 'timing/data': 0.00018854999507311732, 'timing/model': 0.3908931869955268, 'learning_rate/qwen_vl_interface': 4.7030401347249685e-06, 'learning_rate/action_model': 4.703040134724968e-05, 'epoch': 0.2}) +[2026-10-03 13:34:04,886][starVLA.training.train_starvla][INFO] - Step 2731, Loss: {'train/loss': 1.1920897122763563e-06, 'train/grad_norm_pre_clip': 0.00043243696563877165, 'timing/data': 0.00030696998874191195, 'timing/model': 0.3886261360021308, 'learning_rate/qwen_vl_interface': 4.700015083497143e-06, 'learning_rate/action_model': 4.700015083497142e-05, 'epoch': 0.2}) +[2026-10-03 13:34:05,584][starVLA.training.train_starvla][INFO] - Step 2732, Loss: {'train/loss': 1.7732323840391473e-06, 'train/grad_norm_pre_clip': 0.0005505512235686183, 'timing/data': 0.00022042999626137316, 'timing/model': 0.39070957600779366, 'learning_rate/qwen_vl_interface': 4.696990258347342e-06, 'learning_rate/action_model': 4.696990258347342e-05, 'epoch': 0.2}) +[2026-10-03 13:34:06,282][starVLA.training.train_starvla][INFO] - Step 2733, Loss: {'train/loss': 0.0012687151320278645, 'train/grad_norm_pre_clip': 4.700294494628906, 'timing/data': 0.0002536100073484704, 'timing/model': 0.38958122600161005, 'learning_rate/qwen_vl_interface': 4.693965660518957e-06, 'learning_rate/action_model': 4.693965660518957e-05, 'epoch': 0.2}) +[2026-10-03 13:34:07,170][starVLA.training.train_starvla][INFO] - Step 2734, Loss: {'train/loss': 0.00013534053869079798, 'train/grad_norm_pre_clip': 0.06739367544651031, 'timing/data': 0.0002575800026534125, 'timing/model': 0.3900595559971407, 'learning_rate/qwen_vl_interface': 4.690941291255285e-06, 'learning_rate/action_model': 4.6909412912552844e-05, 'epoch': 0.2}) +[2026-10-03 13:34:07,864][starVLA.training.train_starvla][INFO] - Step 2735, Loss: {'train/loss': 6.847495387773961e-05, 'train/grad_norm_pre_clip': 0.02464304305613041, 'timing/data': 0.0002676900039659813, 'timing/model': 0.388775565996184, 'learning_rate/qwen_vl_interface': 4.68791715179953e-06, 'learning_rate/action_model': 4.68791715179953e-05, 'epoch': 0.2}) +[2026-10-03 13:34:08,562][starVLA.training.train_starvla][INFO] - Step 2736, Loss: {'train/loss': 0.3186805844306946, 'train/grad_norm_pre_clip': 181.05935668945312, 'timing/data': 0.00024172999837901443, 'timing/model': 0.3906618269975297, 'learning_rate/qwen_vl_interface': 4.684893243394804e-06, 'learning_rate/action_model': 4.6848932433948036e-05, 'epoch': 0.2}) +[2026-10-03 13:34:09,259][starVLA.training.train_starvla][INFO] - Step 2737, Loss: {'train/loss': 9.685745681053959e-07, 'train/grad_norm_pre_clip': 0.00015785134746693075, 'timing/data': 0.00033811999310273677, 'timing/model': 0.3885509269894101, 'learning_rate/qwen_vl_interface': 4.68186956728412e-06, 'learning_rate/action_model': 4.68186956728412e-05, 'epoch': 0.2}) +[2026-10-03 13:34:09,955][starVLA.training.train_starvla][INFO] - Step 2738, Loss: {'train/loss': 2.210422644566279e-05, 'train/grad_norm_pre_clip': 0.008013801649212837, 'timing/data': 0.00018930000078398734, 'timing/model': 0.39045825599168893, 'learning_rate/qwen_vl_interface': 4.678846124710395e-06, 'learning_rate/action_model': 4.678846124710394e-05, 'epoch': 0.2}) +[2026-10-03 13:34:10,844][starVLA.training.train_starvla][INFO] - Step 2739, Loss: {'train/loss': 0.48591116070747375, 'train/grad_norm_pre_clip': 89.11042785644531, 'timing/data': 0.00026784998772200197, 'timing/model': 0.3890711359999841, 'learning_rate/qwen_vl_interface': 4.675822916916454e-06, 'learning_rate/action_model': 4.6758229169164536e-05, 'epoch': 0.2}) +[2026-10-03 13:34:11,543][starVLA.training.train_starvla][INFO] - Step 2740, Loss: {'train/loss': 5.923125627305126e-06, 'train/grad_norm_pre_clip': 0.003009672509506345, 'timing/data': 0.0004355600103735924, 'timing/model': 0.39097797599970363, 'learning_rate/qwen_vl_interface': 4.6727999451450236e-06, 'learning_rate/action_model': 4.672799945145023e-05, 'epoch': 0.2}) +[2026-10-03 13:34:12,237][starVLA.training.train_starvla][INFO] - Step 2741, Loss: {'train/loss': 1.2367947874736274e-06, 'train/grad_norm_pre_clip': 0.00021650028065778315, 'timing/data': 0.0004015500016976148, 'timing/model': 0.3883225760073401, 'learning_rate/qwen_vl_interface': 4.669777210638733e-06, 'learning_rate/action_model': 4.669777210638733e-05, 'epoch': 0.2}) +[2026-10-03 13:34:12,934][starVLA.training.train_starvla][INFO] - Step 2742, Loss: {'train/loss': 3.2632947295496706e-06, 'train/grad_norm_pre_clip': 0.0017680157907307148, 'timing/data': 0.00026440000510774553, 'timing/model': 0.3909396969975205, 'learning_rate/qwen_vl_interface': 4.666754714640115e-06, 'learning_rate/action_model': 4.666754714640115e-05, 'epoch': 0.2}) +[2026-10-03 13:34:13,629][starVLA.training.train_starvla][INFO] - Step 2743, Loss: {'train/loss': 0.0007470107520930469, 'train/grad_norm_pre_clip': 3.3871829509735107, 'timing/data': 0.0002708100073505193, 'timing/model': 0.3885469759989064, 'learning_rate/qwen_vl_interface': 4.663732458391602e-06, 'learning_rate/action_model': 4.663732458391601e-05, 'epoch': 0.2}) +[2026-10-03 13:34:14,516][starVLA.training.train_starvla][INFO] - Step 2744, Loss: {'train/loss': 0.0002975641982629895, 'train/grad_norm_pre_clip': 0.3391232490539551, 'timing/data': 0.000308920003590174, 'timing/model': 0.5815573649888393, 'learning_rate/qwen_vl_interface': 4.6607104431355295e-06, 'learning_rate/action_model': 4.66071044313553e-05, 'epoch': 0.2}) +[2026-10-03 13:34:15,212][starVLA.training.train_starvla][INFO] - Step 2745, Loss: {'train/loss': 0.18753446638584137, 'train/grad_norm_pre_clip': 148.6925048828125, 'timing/data': 0.00033382000401616096, 'timing/model': 0.3879213170002913, 'learning_rate/qwen_vl_interface': 4.657688670114135e-06, 'learning_rate/action_model': 4.657688670114135e-05, 'epoch': 0.2}) +[2026-10-03 13:34:15,908][starVLA.training.train_starvla][INFO] - Step 2746, Loss: {'train/loss': 6.559467146871611e-05, 'train/grad_norm_pre_clip': 0.01756799779832363, 'timing/data': 0.0003603599907364696, 'timing/model': 0.39015725599892903, 'learning_rate/qwen_vl_interface': 4.654667140569555e-06, 'learning_rate/action_model': 4.6546671405695544e-05, 'epoch': 0.2}) +[2026-10-03 13:34:16,603][starVLA.training.train_starvla][INFO] - Step 2747, Loss: {'train/loss': 8.195158443413675e-06, 'train/grad_norm_pre_clip': 0.007033518515527248, 'timing/data': 0.00023262000468093902, 'timing/model': 0.3885220369993476, 'learning_rate/qwen_vl_interface': 4.651645855743825e-06, 'learning_rate/action_model': 4.651645855743825e-05, 'epoch': 0.2}) +[2026-10-03 13:34:17,300][starVLA.training.train_starvla][INFO] - Step 2748, Loss: {'train/loss': 0.46124985814094543, 'train/grad_norm_pre_clip': 12.0166597366333, 'timing/data': 0.0002562500012572855, 'timing/model': 0.3901738260028651, 'learning_rate/qwen_vl_interface': 4.648624816878883e-06, 'learning_rate/action_model': 4.6486248168788825e-05, 'epoch': 0.2}) +[2026-10-03 13:34:18,188][starVLA.training.train_starvla][INFO] - Step 2749, Loss: {'train/loss': 0.3936499357223511, 'train/grad_norm_pre_clip': 168.71839904785156, 'timing/data': 0.00026292999973520637, 'timing/model': 0.5798412949952763, 'learning_rate/qwen_vl_interface': 4.6456040252165615e-06, 'learning_rate/action_model': 4.645604025216562e-05, 'epoch': 0.2}) +[2026-10-03 13:34:18,886][starVLA.training.train_starvla][INFO] - Step 2750, Loss: {'train/loss': 0.6095779538154602, 'train/grad_norm_pre_clip': 53.14384460449219, 'timing/data': 0.00030159000016283244, 'timing/model': 0.39093759600655176, 'learning_rate/qwen_vl_interface': 4.642583481998596e-06, 'learning_rate/action_model': 4.642583481998596e-05, 'epoch': 0.2}) +[2026-10-03 13:34:19,579][starVLA.training.train_starvla][INFO] - Step 2751, Loss: {'train/loss': 0.0033455269876867533, 'train/grad_norm_pre_clip': 16.685325622558594, 'timing/data': 0.00023624999448657036, 'timing/model': 0.3880613969959086, 'learning_rate/qwen_vl_interface': 4.639563188466616e-06, 'learning_rate/action_model': 4.6395631884666165e-05, 'epoch': 0.2}) +[2026-10-03 13:34:20,277][starVLA.training.train_starvla][INFO] - Step 2752, Loss: {'train/loss': 0.5913069844245911, 'train/grad_norm_pre_clip': 20.360750198364258, 'timing/data': 0.00023233999672811478, 'timing/model': 0.3904822260083165, 'learning_rate/qwen_vl_interface': 4.636543145862152e-06, 'learning_rate/action_model': 4.6365431458621523e-05, 'epoch': 0.2}) +[2026-10-03 13:34:20,973][starVLA.training.train_starvla][INFO] - Step 2753, Loss: {'train/loss': 0.26656341552734375, 'train/grad_norm_pre_clip': 27.124441146850586, 'timing/data': 0.000261929992120713, 'timing/model': 0.38810030699823983, 'learning_rate/qwen_vl_interface': 4.633523355426629e-06, 'learning_rate/action_model': 4.6335233554266295e-05, 'epoch': 0.2}) +[2026-10-03 13:34:21,669][starVLA.training.train_starvla][INFO] - Step 2754, Loss: {'train/loss': 0.07234068214893341, 'train/grad_norm_pre_clip': 235.71360778808594, 'timing/data': 0.00019671999325510114, 'timing/model': 0.3903055270056939, 'learning_rate/qwen_vl_interface': 4.6305038184013675e-06, 'learning_rate/action_model': 4.6305038184013675e-05, 'epoch': 0.2}) +[2026-10-03 13:34:22,553][starVLA.training.train_starvla][INFO] - Step 2755, Loss: {'train/loss': 0.09741563349962234, 'train/grad_norm_pre_clip': 121.24609375, 'timing/data': 0.00026810000417754054, 'timing/model': 0.3872105069895042, 'learning_rate/qwen_vl_interface': 4.627484536027589e-06, 'learning_rate/action_model': 4.627484536027588e-05, 'epoch': 0.2}) +[2026-10-03 13:34:23,252][starVLA.training.train_starvla][INFO] - Step 2756, Loss: {'train/loss': 0.0029455905314534903, 'train/grad_norm_pre_clip': 4.098919868469238, 'timing/data': 0.0002741899952525273, 'timing/model': 0.390123646997381, 'learning_rate/qwen_vl_interface': 4.624465509546399e-06, 'learning_rate/action_model': 4.6244655095463986e-05, 'epoch': 0.2}) +[2026-10-03 13:34:23,948][starVLA.training.train_starvla][INFO] - Step 2757, Loss: {'train/loss': 5.3823951020604e-05, 'train/grad_norm_pre_clip': 0.010784031823277473, 'timing/data': 0.0003961100010201335, 'timing/model': 0.3896267959935358, 'learning_rate/qwen_vl_interface': 4.6214467401988106e-06, 'learning_rate/action_model': 4.621446740198811e-05, 'epoch': 0.2}) +[2026-10-03 13:34:24,645][starVLA.training.train_starvla][INFO] - Step 2758, Loss: {'train/loss': 0.040694717317819595, 'train/grad_norm_pre_clip': 130.62017822265625, 'timing/data': 0.00025645999994594604, 'timing/model': 0.3904320770088816, 'learning_rate/qwen_vl_interface': 4.6184282292257245e-06, 'learning_rate/action_model': 4.618428229225725e-05, 'epoch': 0.2}) +[2026-10-03 13:34:25,340][starVLA.training.train_starvla][INFO] - Step 2759, Loss: {'train/loss': 0.0001407949603162706, 'train/grad_norm_pre_clip': 0.030418772250413895, 'timing/data': 0.0002440799871692434, 'timing/model': 0.3886593770002946, 'learning_rate/qwen_vl_interface': 4.615409977867938e-06, 'learning_rate/action_model': 4.615409977867937e-05, 'epoch': 0.2}) +[2026-10-03 13:34:26,231][starVLA.training.train_starvla][INFO] - Step 2760, Loss: {'train/loss': 0.0001332058454863727, 'train/grad_norm_pre_clip': 0.039000362157821655, 'timing/data': 0.00020557999960146844, 'timing/model': 0.5823705359944142, 'learning_rate/qwen_vl_interface': 4.612391987366138e-06, 'learning_rate/action_model': 4.612391987366138e-05, 'epoch': 0.2}) +[2026-10-03 13:34:26,927][starVLA.training.train_starvla][INFO] - Step 2761, Loss: {'train/loss': 0.0002251110563520342, 'train/grad_norm_pre_clip': 0.05167326703667641, 'timing/data': 0.0003461400046944618, 'timing/model': 0.3884779759973753, 'learning_rate/qwen_vl_interface': 4.609374258960908e-06, 'learning_rate/action_model': 4.6093742589609074e-05, 'epoch': 0.2}) +[2026-10-03 13:34:27,623][starVLA.training.train_starvla][INFO] - Step 2762, Loss: {'train/loss': 0.031402621418237686, 'train/grad_norm_pre_clip': 42.43561553955078, 'timing/data': 0.0003359690017532557, 'timing/model': 0.39031164700281806, 'learning_rate/qwen_vl_interface': 4.60635679389272e-06, 'learning_rate/action_model': 4.60635679389272e-05, 'epoch': 0.2}) +[2026-10-03 13:34:28,317][starVLA.training.train_starvla][INFO] - Step 2763, Loss: {'train/loss': 0.00022397379507310688, 'train/grad_norm_pre_clip': 0.055778272449970245, 'timing/data': 0.00022602999524679035, 'timing/model': 0.38718309700198006, 'learning_rate/qwen_vl_interface': 4.603339593401943e-06, 'learning_rate/action_model': 4.603339593401942e-05, 'epoch': 0.2}) +[2026-10-03 13:34:29,015][starVLA.training.train_starvla][INFO] - Step 2764, Loss: {'train/loss': 0.5864010453224182, 'train/grad_norm_pre_clip': 82.98474884033203, 'timing/data': 0.00027830999169964343, 'timing/model': 0.3900111169932643, 'learning_rate/qwen_vl_interface': 4.600322658728832e-06, 'learning_rate/action_model': 4.600322658728832e-05, 'epoch': 0.2}) +[2026-10-03 13:34:29,711][starVLA.training.train_starvla][INFO] - Step 2765, Loss: {'train/loss': 4.744745456264354e-05, 'train/grad_norm_pre_clip': 0.0336277112364769, 'timing/data': 0.0003205199900548905, 'timing/model': 0.3886364159989171, 'learning_rate/qwen_vl_interface': 4.597305991113537e-06, 'learning_rate/action_model': 4.597305991113537e-05, 'epoch': 0.2}) +[2026-10-03 13:34:30,596][starVLA.training.train_starvla][INFO] - Step 2766, Loss: {'train/loss': 0.00010446133092045784, 'train/grad_norm_pre_clip': 0.10549023747444153, 'timing/data': 0.0002864299895009026, 'timing/model': 0.3900204359961208, 'learning_rate/qwen_vl_interface': 4.5942895917960935e-06, 'learning_rate/action_model': 4.5942895917960935e-05, 'epoch': 0.2}) +[2026-10-03 13:34:31,290][starVLA.training.train_starvla][INFO] - Step 2767, Loss: {'train/loss': 8.575191895943135e-05, 'train/grad_norm_pre_clip': 0.07214199006557465, 'timing/data': 0.00034266000147908926, 'timing/model': 0.3882424170005834, 'learning_rate/qwen_vl_interface': 4.59127346201643e-06, 'learning_rate/action_model': 4.59127346201643e-05, 'epoch': 0.2}) +[2026-10-03 13:34:31,987][starVLA.training.train_starvla][INFO] - Step 2768, Loss: {'train/loss': 0.0007402157643809915, 'train/grad_norm_pre_clip': 1.4541634321212769, 'timing/data': 0.00027333000616636127, 'timing/model': 0.38986570700944867, 'learning_rate/qwen_vl_interface': 4.588257603014364e-06, 'learning_rate/action_model': 4.588257603014364e-05, 'epoch': 0.2}) +[2026-10-03 13:34:32,685][starVLA.training.train_starvla][INFO] - Step 2769, Loss: {'train/loss': 0.00021353569172788411, 'train/grad_norm_pre_clip': 0.0511811263859272, 'timing/data': 0.0002431799948681146, 'timing/model': 0.3891229760047281, 'learning_rate/qwen_vl_interface': 4.585242016029602e-06, 'learning_rate/action_model': 4.5852420160296015e-05, 'epoch': 0.2}) +[2026-10-03 13:34:33,381][starVLA.training.train_starvla][INFO] - Step 2770, Loss: {'train/loss': 0.0028774039819836617, 'train/grad_norm_pre_clip': 8.67080307006836, 'timing/data': 0.0003139300097245723, 'timing/model': 0.3907866059889784, 'learning_rate/qwen_vl_interface': 4.582226702301737e-06, 'learning_rate/action_model': 4.5822267023017375e-05, 'epoch': 0.2}) +[2026-10-03 13:34:34,270][starVLA.training.train_starvla][INFO] - Step 2771, Loss: {'train/loss': 5.4892552725505084e-05, 'train/grad_norm_pre_clip': 0.06577222049236298, 'timing/data': 0.00020299998868722469, 'timing/model': 0.3887797259958461, 'learning_rate/qwen_vl_interface': 4.57921166307025e-06, 'learning_rate/action_model': 4.57921166307025e-05, 'epoch': 0.2}) +[2026-10-03 13:34:34,968][starVLA.training.train_starvla][INFO] - Step 2772, Loss: {'train/loss': 0.0008588623604737222, 'train/grad_norm_pre_clip': 0.3466508984565735, 'timing/data': 0.0002003599947784096, 'timing/model': 0.3906955359998392, 'learning_rate/qwen_vl_interface': 4.576196899574511e-06, 'learning_rate/action_model': 4.576196899574511e-05, 'epoch': 0.2}) +[2026-10-03 13:34:35,665][starVLA.training.train_starvla][INFO] - Step 2773, Loss: {'train/loss': 0.0004391574766486883, 'train/grad_norm_pre_clip': 0.5922437906265259, 'timing/data': 0.00024036000831983984, 'timing/model': 0.3890144270117162, 'learning_rate/qwen_vl_interface': 4.573182413053775e-06, 'learning_rate/action_model': 4.5731824130537746e-05, 'epoch': 0.2}) +[2026-10-03 13:34:36,364][starVLA.training.train_starvla][INFO] - Step 2774, Loss: {'train/loss': 0.010855947621166706, 'train/grad_norm_pre_clip': 6.9389214515686035, 'timing/data': 0.00024290000146720558, 'timing/model': 0.3916683660063427, 'learning_rate/qwen_vl_interface': 4.570168204747183e-06, 'learning_rate/action_model': 4.570168204747183e-05, 'epoch': 0.2}) +[2026-10-03 13:34:37,060][starVLA.training.train_starvla][INFO] - Step 2775, Loss: {'train/loss': 0.00017906595894601196, 'train/grad_norm_pre_clip': 0.13318414986133575, 'timing/data': 0.0002495700027793646, 'timing/model': 0.3894181070063496, 'learning_rate/qwen_vl_interface': 4.567154275893762e-06, 'learning_rate/action_model': 4.5671542758937624e-05, 'epoch': 0.2}) +[2026-10-03 13:34:37,953][starVLA.training.train_starvla][INFO] - Step 2776, Loss: {'train/loss': 0.7523853182792664, 'train/grad_norm_pre_clip': 43.11817169189453, 'timing/data': 0.0002508100005798042, 'timing/model': 0.5859816540032625, 'learning_rate/qwen_vl_interface': 4.564140627732426e-06, 'learning_rate/action_model': 4.564140627732426e-05, 'epoch': 0.2}) +[2026-10-03 13:34:38,651][starVLA.training.train_starvla][INFO] - Step 2777, Loss: {'train/loss': 0.0015426194295287132, 'train/grad_norm_pre_clip': 6.5988898277282715, 'timing/data': 0.0002013400080613792, 'timing/model': 0.3891318259993568, 'learning_rate/qwen_vl_interface': 4.56112726150197e-06, 'learning_rate/action_model': 4.56112726150197e-05, 'epoch': 0.2}) +[2026-10-03 13:34:39,348][starVLA.training.train_starvla][INFO] - Step 2778, Loss: {'train/loss': 4.082898158230819e-06, 'train/grad_norm_pre_clip': 0.0013305661268532276, 'timing/data': 0.0002516999957151711, 'timing/model': 0.3905595960095525, 'learning_rate/qwen_vl_interface': 4.558114178441076e-06, 'learning_rate/action_model': 4.558114178441076e-05, 'epoch': 0.2}) +[2026-10-03 13:34:40,044][starVLA.training.train_starvla][INFO] - Step 2779, Loss: {'train/loss': 0.13959167897701263, 'train/grad_norm_pre_clip': 77.3280258178711, 'timing/data': 0.0003296999930171296, 'timing/model': 0.3889013969892403, 'learning_rate/qwen_vl_interface': 4.555101379788308e-06, 'learning_rate/action_model': 4.555101379788308e-05, 'epoch': 0.2}) +[2026-10-03 13:34:40,743][starVLA.training.train_starvla][INFO] - Step 2780, Loss: {'train/loss': 0.03377939760684967, 'train/grad_norm_pre_clip': 46.39324188232422, 'timing/data': 0.00030772999161854386, 'timing/model': 0.39110335599980317, 'learning_rate/qwen_vl_interface': 4.552088866782115e-06, 'learning_rate/action_model': 4.552088866782115e-05, 'epoch': 0.2}) +[2026-10-03 13:34:41,441][starVLA.training.train_starvla][INFO] - Step 2781, Loss: {'train/loss': 0.02734021469950676, 'train/grad_norm_pre_clip': 37.85939407348633, 'timing/data': 0.000248030002694577, 'timing/model': 0.38954967600875534, 'learning_rate/qwen_vl_interface': 4.5490766406608265e-06, 'learning_rate/action_model': 4.549076640660827e-05, 'epoch': 0.2}) +[2026-10-03 13:34:42,328][starVLA.training.train_starvla][INFO] - Step 2782, Loss: {'train/loss': 0.27758169174194336, 'train/grad_norm_pre_clip': 32.3878173828125, 'timing/data': 0.00023330999829340726, 'timing/model': 0.39107024599798024, 'learning_rate/qwen_vl_interface': 4.546064702662655e-06, 'learning_rate/action_model': 4.546064702662655e-05, 'epoch': 0.2}) +[2026-10-03 13:34:43,025][starVLA.training.train_starvla][INFO] - Step 2783, Loss: {'train/loss': 0.39789748191833496, 'train/grad_norm_pre_clip': 100.53952026367188, 'timing/data': 0.00025214000197593123, 'timing/model': 0.3890957370022079, 'learning_rate/qwen_vl_interface': 4.543053054025693e-06, 'learning_rate/action_model': 4.5430530540256926e-05, 'epoch': 0.2}) +[2026-10-03 13:34:43,727][starVLA.training.train_starvla][INFO] - Step 2784, Loss: {'train/loss': 5.215370038058609e-06, 'train/grad_norm_pre_clip': 0.001658254535868764, 'timing/data': 0.00019844000053126365, 'timing/model': 0.3908553560031578, 'learning_rate/qwen_vl_interface': 4.540041695987917e-06, 'learning_rate/action_model': 4.540041695987917e-05, 'epoch': 0.2}) +[2026-10-03 13:34:44,425][starVLA.training.train_starvla][INFO] - Step 2785, Loss: {'train/loss': 1.7389187632943504e-05, 'train/grad_norm_pre_clip': 0.00481494190171361, 'timing/data': 0.00033040999551303685, 'timing/model': 0.3892286970076384, 'learning_rate/qwen_vl_interface': 4.537030629787182e-06, 'learning_rate/action_model': 4.5370306297871826e-05, 'epoch': 0.2}) +[2026-10-03 13:34:45,123][starVLA.training.train_starvla][INFO] - Step 2786, Loss: {'train/loss': 0.0003959052264690399, 'train/grad_norm_pre_clip': 0.6783069968223572, 'timing/data': 0.0002905200090026483, 'timing/model': 0.3914686560019618, 'learning_rate/qwen_vl_interface': 4.534019856661226e-06, 'learning_rate/action_model': 4.534019856661226e-05, 'epoch': 0.2}) +[2026-10-03 13:34:46,015][starVLA.training.train_starvla][INFO] - Step 2787, Loss: {'train/loss': 0.00047944000107236207, 'train/grad_norm_pre_clip': 0.2905036211013794, 'timing/data': 0.00032294999982696027, 'timing/model': 0.38928768699406646, 'learning_rate/qwen_vl_interface': 4.531009377847661e-06, 'learning_rate/action_model': 4.531009377847661e-05, 'epoch': 0.2}) +[2026-10-03 13:34:46,718][starVLA.training.train_starvla][INFO] - Step 2788, Loss: {'train/loss': 4.8138823331100866e-05, 'train/grad_norm_pre_clip': 0.024209504947066307, 'timing/data': 0.0002750699932221323, 'timing/model': 0.39256082600331865, 'learning_rate/qwen_vl_interface': 4.527999194583982e-06, 'learning_rate/action_model': 4.527999194583981e-05, 'epoch': 0.2}) +[2026-10-03 13:34:47,418][starVLA.training.train_starvla][INFO] - Step 2789, Loss: {'train/loss': 0.00014203129103407264, 'train/grad_norm_pre_clip': 0.10954412072896957, 'timing/data': 0.00025040999753400683, 'timing/model': 0.3905055259965593, 'learning_rate/qwen_vl_interface': 4.52498930810756e-06, 'learning_rate/action_model': 4.52498930810756e-05, 'epoch': 0.2}) +[2026-10-03 13:34:48,119][starVLA.training.train_starvla][INFO] - Step 2790, Loss: {'train/loss': 7.028569234535098e-05, 'train/grad_norm_pre_clip': 0.01756129413843155, 'timing/data': 0.000225650001084432, 'timing/model': 0.39278042600199115, 'learning_rate/qwen_vl_interface': 4.521979719655648e-06, 'learning_rate/action_model': 4.5219797196556484e-05, 'epoch': 0.2}) +[2026-10-03 13:34:48,818][starVLA.training.train_starvla][INFO] - Step 2791, Loss: {'train/loss': 0.2869836091995239, 'train/grad_norm_pre_clip': 98.51200866699219, 'timing/data': 0.000275719998171553, 'timing/model': 0.3904945160029456, 'learning_rate/qwen_vl_interface': 4.518970430465377e-06, 'learning_rate/action_model': 4.518970430465376e-05, 'epoch': 0.2}) +[2026-10-03 13:34:49,713][starVLA.training.train_starvla][INFO] - Step 2792, Loss: {'train/loss': 2.3311491531785578e-05, 'train/grad_norm_pre_clip': 0.009190227836370468, 'timing/data': 0.0002477100060787052, 'timing/model': 0.3923982059932314, 'learning_rate/qwen_vl_interface': 4.515961441773745e-06, 'learning_rate/action_model': 4.515961441773745e-05, 'epoch': 0.2}) +[2026-10-03 13:34:50,413][starVLA.training.train_starvla][INFO] - Step 2793, Loss: {'train/loss': 0.0011800043284893036, 'train/grad_norm_pre_clip': 1.365297794342041, 'timing/data': 0.0002588800125522539, 'timing/model': 0.3899984760064399, 'learning_rate/qwen_vl_interface': 4.5129527548176385e-06, 'learning_rate/action_model': 4.512952754817638e-05, 'epoch': 0.2}) +[2026-10-03 13:34:51,114][starVLA.training.train_starvla][INFO] - Step 2794, Loss: {'train/loss': 0.0001342192990705371, 'train/grad_norm_pre_clip': 0.10260089486837387, 'timing/data': 0.0002452199987601489, 'timing/model': 0.3930072060029488, 'learning_rate/qwen_vl_interface': 4.509944370833813e-06, 'learning_rate/action_model': 4.509944370833813e-05, 'epoch': 0.2}) +[2026-10-03 13:34:51,814][starVLA.training.train_starvla][INFO] - Step 2795, Loss: {'train/loss': 0.0011243750341236591, 'train/grad_norm_pre_clip': 0.3837131857872009, 'timing/data': 0.00030586999491788447, 'timing/model': 0.3914508460002253, 'learning_rate/qwen_vl_interface': 4.506936291058902e-06, 'learning_rate/action_model': 4.506936291058902e-05, 'epoch': 0.2}) +[2026-10-03 13:34:52,516][starVLA.training.train_starvla][INFO] - Step 2796, Loss: {'train/loss': 0.2218109518289566, 'train/grad_norm_pre_clip': 361.16644287109375, 'timing/data': 0.00024287999258376658, 'timing/model': 0.39217346599616576, 'learning_rate/qwen_vl_interface': 4.503928516729416e-06, 'learning_rate/action_model': 4.503928516729416e-05, 'epoch': 0.2}) +[2026-10-03 13:34:53,408][starVLA.training.train_starvla][INFO] - Step 2797, Loss: {'train/loss': 3.0060762583161704e-05, 'train/grad_norm_pre_clip': 0.009900416247546673, 'timing/data': 0.00023692000831943005, 'timing/model': 0.58231044501008, 'learning_rate/qwen_vl_interface': 4.500921049081734e-06, 'learning_rate/action_model': 4.500921049081734e-05, 'epoch': 0.2}) +[2026-10-03 13:34:54,108][starVLA.training.train_starvla][INFO] - Step 2798, Loss: {'train/loss': 0.029245823621749878, 'train/grad_norm_pre_clip': 58.16651916503906, 'timing/data': 0.00021765999554190785, 'timing/model': 0.3926291859970661, 'learning_rate/qwen_vl_interface': 4.497913889352113e-06, 'learning_rate/action_model': 4.4979138893521126e-05, 'epoch': 0.2}) +[2026-10-03 13:34:54,807][starVLA.training.train_starvla][INFO] - Step 2799, Loss: {'train/loss': 2.595580372144468e-05, 'train/grad_norm_pre_clip': 0.009838012047111988, 'timing/data': 0.0001924800017150119, 'timing/model': 0.39104747600504197, 'learning_rate/qwen_vl_interface': 4.494907038776684e-06, 'learning_rate/action_model': 4.4949070387766836e-05, 'epoch': 0.2}) +[2026-10-03 13:34:55,510][starVLA.training.train_starvla][INFO] - Step 2800, Loss: {'train/loss': 0.0011645080521702766, 'train/grad_norm_pre_clip': 3.69720196723938, 'timing/data': 0.00028704000578727573, 'timing/model': 0.3925933860009536, 'learning_rate/qwen_vl_interface': 4.49190049859145e-06, 'learning_rate/action_model': 4.49190049859145e-05, 'epoch': 0.2}) +[2026-10-03 13:34:56,210][starVLA.training.train_starvla][INFO] - Step 2801, Loss: {'train/loss': 8.27751136966981e-06, 'train/grad_norm_pre_clip': 0.0022679127287119627, 'timing/data': 0.00028457000735215843, 'timing/model': 0.39091318599821534, 'learning_rate/qwen_vl_interface': 4.488894270032284e-06, 'learning_rate/action_model': 4.4888942700322834e-05, 'epoch': 0.21}) +[2026-10-03 13:34:57,107][starVLA.training.train_starvla][INFO] - Step 2802, Loss: {'train/loss': 3.78488221031148e-06, 'train/grad_norm_pre_clip': 0.0012091412208974361, 'timing/data': 0.0002284600050188601, 'timing/model': 0.5877301540022017, 'learning_rate/qwen_vl_interface': 4.485888354334933e-06, 'learning_rate/action_model': 4.4858883543349334e-05, 'epoch': 0.21}) +[2026-10-03 13:34:57,805][starVLA.training.train_starvla][INFO] - Step 2803, Loss: {'train/loss': 0.008730296976864338, 'train/grad_norm_pre_clip': 47.347293853759766, 'timing/data': 0.0002947100001620129, 'timing/model': 0.3904837860027328, 'learning_rate/qwen_vl_interface': 4.482882752735017e-06, 'learning_rate/action_model': 4.482882752735017e-05, 'epoch': 0.21}) +[2026-10-03 13:34:58,507][starVLA.training.train_starvla][INFO] - Step 2804, Loss: {'train/loss': 0.09585276991128922, 'train/grad_norm_pre_clip': 84.79475402832031, 'timing/data': 0.00023953999334480613, 'timing/model': 0.3923380260093836, 'learning_rate/qwen_vl_interface': 4.479877466468027e-06, 'learning_rate/action_model': 4.479877466468026e-05, 'epoch': 0.21}) +[2026-10-03 13:34:59,207][starVLA.training.train_starvla][INFO] - Step 2805, Loss: {'train/loss': 0.0012927863281220198, 'train/grad_norm_pre_clip': 0.3245543837547302, 'timing/data': 0.0003101600013906136, 'timing/model': 0.39092792599694803, 'learning_rate/qwen_vl_interface': 4.476872496769318e-06, 'learning_rate/action_model': 4.476872496769318e-05, 'epoch': 0.21}) +[2026-10-03 13:34:59,908][starVLA.training.train_starvla][INFO] - Step 2806, Loss: {'train/loss': 7.100297352735652e-06, 'train/grad_norm_pre_clip': 0.0017990368651226163, 'timing/data': 0.0002166399935958907, 'timing/model': 0.3928416659909999, 'learning_rate/qwen_vl_interface': 4.473867844874122e-06, 'learning_rate/action_model': 4.473867844874122e-05, 'epoch': 0.21}) +[2026-10-03 13:35:00,606][starVLA.training.train_starvla][INFO] - Step 2807, Loss: {'train/loss': 0.0005497190868481994, 'train/grad_norm_pre_clip': 0.1757875233888626, 'timing/data': 0.0002312199940206483, 'timing/model': 0.3905199760047253, 'learning_rate/qwen_vl_interface': 4.470863512017539e-06, 'learning_rate/action_model': 4.470863512017539e-05, 'epoch': 0.21}) +[2026-10-03 13:35:01,500][starVLA.training.train_starvla][INFO] - Step 2808, Loss: {'train/loss': 0.0010042290668934584, 'train/grad_norm_pre_clip': 0.25049951672554016, 'timing/data': 0.00029486999846994877, 'timing/model': 0.392471386003308, 'learning_rate/qwen_vl_interface': 4.4678594994345344e-06, 'learning_rate/action_model': 4.467859499434534e-05, 'epoch': 0.21}) +[2026-10-03 13:35:02,201][starVLA.training.train_starvla][INFO] - Step 2809, Loss: {'train/loss': 4.313864337746054e-06, 'train/grad_norm_pre_clip': 0.0011743427021428943, 'timing/data': 0.0002528100012568757, 'timing/model': 0.39047031600784976, 'learning_rate/qwen_vl_interface': 4.4648558083599475e-06, 'learning_rate/action_model': 4.464855808359947e-05, 'epoch': 0.21}) +[2026-10-03 13:35:02,902][starVLA.training.train_starvla][INFO] - Step 2810, Loss: {'train/loss': 0.4835573136806488, 'train/grad_norm_pre_clip': 314.4572448730469, 'timing/data': 0.00021549999655690044, 'timing/model': 0.39300434599863365, 'learning_rate/qwen_vl_interface': 4.461852440028479e-06, 'learning_rate/action_model': 4.461852440028479e-05, 'epoch': 0.21}) +[2026-10-03 13:35:03,600][starVLA.training.train_starvla][INFO] - Step 2811, Loss: {'train/loss': 6.467035291279899e-06, 'train/grad_norm_pre_clip': 0.001759798382408917, 'timing/data': 0.0002577400009613484, 'timing/model': 0.39046553600928746, 'learning_rate/qwen_vl_interface': 4.458849395674702e-06, 'learning_rate/action_model': 4.458849395674701e-05, 'epoch': 0.21}) +[2026-10-03 13:35:04,303][starVLA.training.train_starvla][INFO] - Step 2812, Loss: {'train/loss': 0.0005537681863643229, 'train/grad_norm_pre_clip': 0.2655833065509796, 'timing/data': 0.0003191499999957159, 'timing/model': 0.3926202560105594, 'learning_rate/qwen_vl_interface': 4.455846676533052e-06, 'learning_rate/action_model': 4.4558466765330525e-05, 'epoch': 0.21}) +[2026-10-03 13:35:05,194][starVLA.training.train_starvla][INFO] - Step 2813, Loss: {'train/loss': 0.02196846343576908, 'train/grad_norm_pre_clip': 68.62926483154297, 'timing/data': 0.0002122100122505799, 'timing/model': 0.5808986549964175, 'learning_rate/qwen_vl_interface': 4.452844283837838e-06, 'learning_rate/action_model': 4.452844283837838e-05, 'epoch': 0.21}) +[2026-10-03 13:35:05,894][starVLA.training.train_starvla][INFO] - Step 2814, Loss: {'train/loss': 2.644011692609638e-05, 'train/grad_norm_pre_clip': 0.006828491110354662, 'timing/data': 0.0002377299970248714, 'timing/model': 0.39264778699725866, 'learning_rate/qwen_vl_interface': 4.449842218823228e-06, 'learning_rate/action_model': 4.449842218823228e-05, 'epoch': 0.21}) +[2026-10-03 13:35:06,593][starVLA.training.train_starvla][INFO] - Step 2815, Loss: {'train/loss': 2.8535621368064312e-06, 'train/grad_norm_pre_clip': 0.0007021215278655291, 'timing/data': 0.00026795000303536654, 'timing/model': 0.3908186969929375, 'learning_rate/qwen_vl_interface': 4.446840482723256e-06, 'learning_rate/action_model': 4.446840482723255e-05, 'epoch': 0.21}) +[2026-10-03 13:35:07,296][starVLA.training.train_starvla][INFO] - Step 2816, Loss: {'train/loss': 0.00011990093480562791, 'train/grad_norm_pre_clip': 0.06633885204792023, 'timing/data': 0.00023799999326001853, 'timing/model': 0.3924616159929428, 'learning_rate/qwen_vl_interface': 4.443839076771823e-06, 'learning_rate/action_model': 4.443839076771823e-05, 'epoch': 0.21}) +[2026-10-03 13:35:07,996][starVLA.training.train_starvla][INFO] - Step 2817, Loss: {'train/loss': 3.047273821721319e-06, 'train/grad_norm_pre_clip': 0.0006346849841065705, 'timing/data': 0.00022768999042455107, 'timing/model': 0.39097154600312933, 'learning_rate/qwen_vl_interface': 4.4408380022026945e-06, 'learning_rate/action_model': 4.440838002202694e-05, 'epoch': 0.21}) +[2026-10-03 13:35:08,696][starVLA.training.train_starvla][INFO] - Step 2818, Loss: {'train/loss': 0.00046819818089716136, 'train/grad_norm_pre_clip': 2.496105194091797, 'timing/data': 0.00019645999418571591, 'timing/model': 0.3925436759891454, 'learning_rate/qwen_vl_interface': 4.437837260249499e-06, 'learning_rate/action_model': 4.437837260249499e-05, 'epoch': 0.21}) +[2026-10-03 13:35:09,583][starVLA.training.train_starvla][INFO] - Step 2819, Loss: {'train/loss': 0.5983178019523621, 'train/grad_norm_pre_clip': 50.1571044921875, 'timing/data': 0.000202410010388121, 'timing/model': 0.3907704559969716, 'learning_rate/qwen_vl_interface': 4.434836852145728e-06, 'learning_rate/action_model': 4.434836852145728e-05, 'epoch': 0.21}) +[2026-10-03 13:35:10,286][starVLA.training.train_starvla][INFO] - Step 2820, Loss: {'train/loss': 0.029256301000714302, 'train/grad_norm_pre_clip': 97.7633285522461, 'timing/data': 0.000195720000192523, 'timing/model': 0.3925010859966278, 'learning_rate/qwen_vl_interface': 4.431836779124736e-06, 'learning_rate/action_model': 4.431836779124735e-05, 'epoch': 0.21}) +[2026-10-03 13:35:10,986][starVLA.training.train_starvla][INFO] - Step 2821, Loss: {'train/loss': 0.3264601528644562, 'train/grad_norm_pre_clip': 248.01304626464844, 'timing/data': 0.0002834799961419776, 'timing/model': 0.39087696600472555, 'learning_rate/qwen_vl_interface': 4.428837042419738e-06, 'learning_rate/action_model': 4.428837042419738e-05, 'epoch': 0.21}) +[2026-10-03 13:35:11,687][starVLA.training.train_starvla][INFO] - Step 2822, Loss: {'train/loss': 0.0005299696931615472, 'train/grad_norm_pre_clip': 1.8475782871246338, 'timing/data': 0.00020009999570902437, 'timing/model': 0.39263047599524725, 'learning_rate/qwen_vl_interface': 4.425837643263815e-06, 'learning_rate/action_model': 4.425837643263814e-05, 'epoch': 0.21}) +[2026-10-03 13:35:12,385][starVLA.training.train_starvla][INFO] - Step 2823, Loss: {'train/loss': 0.00010799663868965581, 'train/grad_norm_pre_clip': 0.0908648744225502, 'timing/data': 0.0002965599996969104, 'timing/model': 0.39035075600259006, 'learning_rate/qwen_vl_interface': 4.4228385828899045e-06, 'learning_rate/action_model': 4.4228385828899046e-05, 'epoch': 0.21}) +[2026-10-03 13:35:13,279][starVLA.training.train_starvla][INFO] - Step 2824, Loss: {'train/loss': 0.0009431940270587802, 'train/grad_norm_pre_clip': 1.0034818649291992, 'timing/data': 0.00031619000947102904, 'timing/model': 0.39206935599213466, 'learning_rate/qwen_vl_interface': 4.419839862530809e-06, 'learning_rate/action_model': 4.419839862530809e-05, 'epoch': 0.21}) +[2026-10-03 13:35:13,981][starVLA.training.train_starvla][INFO] - Step 2825, Loss: {'train/loss': 0.00012631944264285266, 'train/grad_norm_pre_clip': 0.05041293799877167, 'timing/data': 0.00040391000220552087, 'timing/model': 0.3912054060056107, 'learning_rate/qwen_vl_interface': 4.416841483419188e-06, 'learning_rate/action_model': 4.416841483419188e-05, 'epoch': 0.21}) +[2026-10-03 13:35:14,682][starVLA.training.train_starvla][INFO] - Step 2826, Loss: {'train/loss': 5.252606570138596e-06, 'train/grad_norm_pre_clip': 0.0011807900154963136, 'timing/data': 0.00019885000074282289, 'timing/model': 0.39312404599331785, 'learning_rate/qwen_vl_interface': 4.413843446787561e-06, 'learning_rate/action_model': 4.413843446787561e-05, 'epoch': 0.21}) +[2026-10-03 13:35:15,380][starVLA.training.train_starvla][INFO] - Step 2827, Loss: {'train/loss': 0.0006845354801043868, 'train/grad_norm_pre_clip': 0.4548339545726776, 'timing/data': 0.00025394999829586595, 'timing/model': 0.3899626860074932, 'learning_rate/qwen_vl_interface': 4.41084575386831e-06, 'learning_rate/action_model': 4.4108457538683093e-05, 'epoch': 0.21}) +[2026-10-03 13:35:16,082][starVLA.training.train_starvla][INFO] - Step 2828, Loss: {'train/loss': 0.5370197296142578, 'train/grad_norm_pre_clip': 168.9512939453125, 'timing/data': 0.0003063100011786446, 'timing/model': 0.39218082699517254, 'learning_rate/qwen_vl_interface': 4.40784840589367e-06, 'learning_rate/action_model': 4.4078484058936706e-05, 'epoch': 0.21}) +[2026-10-03 13:35:16,976][starVLA.training.train_starvla][INFO] - Step 2829, Loss: {'train/loss': 8.390406583203003e-05, 'train/grad_norm_pre_clip': 0.02351967990398407, 'timing/data': 0.00024151999969035387, 'timing/model': 0.5839673650043551, 'learning_rate/qwen_vl_interface': 4.404851404095742e-06, 'learning_rate/action_model': 4.404851404095742e-05, 'epoch': 0.21}) +[2026-10-03 13:35:17,677][starVLA.training.train_starvla][INFO] - Step 2830, Loss: {'train/loss': 0.01661386527121067, 'train/grad_norm_pre_clip': 106.86155700683594, 'timing/data': 0.0002270700060762465, 'timing/model': 0.3930248560063774, 'learning_rate/qwen_vl_interface': 4.4018547497064765e-06, 'learning_rate/action_model': 4.401854749706476e-05, 'epoch': 0.21}) +[2026-10-03 13:35:18,375][starVLA.training.train_starvla][INFO] - Step 2831, Loss: {'train/loss': 0.0011918069794774055, 'train/grad_norm_pre_clip': 1.566145896911621, 'timing/data': 0.00023522000992670655, 'timing/model': 0.39030096599890385, 'learning_rate/qwen_vl_interface': 4.398858443957686e-06, 'learning_rate/action_model': 4.398858443957686e-05, 'epoch': 0.21}) +[2026-10-03 13:35:19,077][starVLA.training.train_starvla][INFO] - Step 2832, Loss: {'train/loss': 0.00014126842143014073, 'train/grad_norm_pre_clip': 0.04198504984378815, 'timing/data': 0.00020090000180061907, 'timing/model': 0.3924694969900884, 'learning_rate/qwen_vl_interface': 4.395862488081038e-06, 'learning_rate/action_model': 4.395862488081038e-05, 'epoch': 0.21}) +[2026-10-03 13:35:19,777][starVLA.training.train_starvla][INFO] - Step 2833, Loss: {'train/loss': 0.5524618625640869, 'train/grad_norm_pre_clip': 21.099365234375, 'timing/data': 0.0001994299964280799, 'timing/model': 0.39054859601310454, 'learning_rate/qwen_vl_interface': 4.392866883308059e-06, 'learning_rate/action_model': 4.392866883308059e-05, 'epoch': 0.21}) +[2026-10-03 13:35:20,478][starVLA.training.train_starvla][INFO] - Step 2834, Loss: {'train/loss': 0.044150084257125854, 'train/grad_norm_pre_clip': 80.5535888671875, 'timing/data': 0.00020936000510118902, 'timing/model': 0.39267861600092147, 'learning_rate/qwen_vl_interface': 4.389871630870127e-06, 'learning_rate/action_model': 4.389871630870127e-05, 'epoch': 0.21}) +[2026-10-03 13:35:21,368][starVLA.training.train_starvla][INFO] - Step 2835, Loss: {'train/loss': 7.914157322375104e-05, 'train/grad_norm_pre_clip': 0.01690066047012806, 'timing/data': 0.00020471001334954053, 'timing/model': 0.3915010960045038, 'learning_rate/qwen_vl_interface': 4.386876731998478e-06, 'learning_rate/action_model': 4.386876731998478e-05, 'epoch': 0.21}) +[2026-10-03 13:35:22,071][starVLA.training.train_starvla][INFO] - Step 2836, Loss: {'train/loss': 1.095222160074627e-05, 'train/grad_norm_pre_clip': 0.0024104234762489796, 'timing/data': 0.00020299000607337803, 'timing/model': 0.3935845870000776, 'learning_rate/qwen_vl_interface': 4.3838821879242006e-06, 'learning_rate/action_model': 4.3838821879242e-05, 'epoch': 0.21}) +[2026-10-03 13:35:22,771][starVLA.training.train_starvla][INFO] - Step 2837, Loss: {'train/loss': 1.6189766029128805e-05, 'train/grad_norm_pre_clip': 0.004597225226461887, 'timing/data': 0.00028202999965287745, 'timing/model': 0.3903270459995838, 'learning_rate/qwen_vl_interface': 4.38088799987824e-06, 'learning_rate/action_model': 4.38088799987824e-05, 'epoch': 0.21}) +[2026-10-03 13:35:23,472][starVLA.training.train_starvla][INFO] - Step 2838, Loss: {'train/loss': 0.15088047087192535, 'train/grad_norm_pre_clip': 257.0780944824219, 'timing/data': 0.0002796900080284104, 'timing/model': 0.3923793459980516, 'learning_rate/qwen_vl_interface': 4.3778941690913935e-06, 'learning_rate/action_model': 4.3778941690913934e-05, 'epoch': 0.21}) +[2026-10-03 13:35:24,173][starVLA.training.train_starvla][INFO] - Step 2839, Loss: {'train/loss': 0.0002020936517510563, 'train/grad_norm_pre_clip': 0.1532817780971527, 'timing/data': 0.00033849000465124846, 'timing/model': 0.39086270600091666, 'learning_rate/qwen_vl_interface': 4.374900696794311e-06, 'learning_rate/action_model': 4.374900696794311e-05, 'epoch': 0.21}) +[2026-10-03 13:35:25,068][starVLA.training.train_starvla][INFO] - Step 2840, Loss: {'train/loss': 0.0001800601021386683, 'train/grad_norm_pre_clip': 0.04206365346908569, 'timing/data': 0.0003199700004188344, 'timing/model': 0.39247627700387966, 'learning_rate/qwen_vl_interface': 4.371907584217499e-06, 'learning_rate/action_model': 4.371907584217498e-05, 'epoch': 0.21}) +[2026-10-03 13:35:25,768][starVLA.training.train_starvla][INFO] - Step 2841, Loss: {'train/loss': 1.6554589819861576e-05, 'train/grad_norm_pre_clip': 0.004529683385044336, 'timing/data': 0.0002541499998187646, 'timing/model': 0.3907217060041148, 'learning_rate/qwen_vl_interface': 4.368914832591309e-06, 'learning_rate/action_model': 4.368914832591309e-05, 'epoch': 0.21}) +[2026-10-03 13:35:26,469][starVLA.training.train_starvla][INFO] - Step 2842, Loss: {'train/loss': 0.08406354486942291, 'train/grad_norm_pre_clip': 148.59823608398438, 'timing/data': 0.00025587000709492713, 'timing/model': 0.3918246560060652, 'learning_rate/qwen_vl_interface': 4.365922443145949e-06, 'learning_rate/action_model': 4.365922443145949e-05, 'epoch': 0.21}) +[2026-10-03 13:35:27,168][starVLA.training.train_starvla][INFO] - Step 2843, Loss: {'train/loss': 0.0005848966538906097, 'train/grad_norm_pre_clip': 0.8445174694061279, 'timing/data': 0.00019323000742588192, 'timing/model': 0.3916153860045597, 'learning_rate/qwen_vl_interface': 4.362930417111479e-06, 'learning_rate/action_model': 4.362930417111479e-05, 'epoch': 0.21}) +[2026-10-03 13:35:27,870][starVLA.training.train_starvla][INFO] - Step 2844, Loss: {'train/loss': 1.7351663700537756e-05, 'train/grad_norm_pre_clip': 0.006741099059581757, 'timing/data': 0.0003207400004612282, 'timing/model': 0.3925057259912137, 'learning_rate/qwen_vl_interface': 4.359938755717807e-06, 'learning_rate/action_model': 4.359938755717807e-05, 'epoch': 0.21}) +[2026-10-03 13:35:28,762][starVLA.training.train_starvla][INFO] - Step 2845, Loss: {'train/loss': 2.685828621906694e-05, 'train/grad_norm_pre_clip': 0.00760538037866354, 'timing/data': 0.00020645000040531158, 'timing/model': 0.39037263600039296, 'learning_rate/qwen_vl_interface': 4.3569474601946934e-06, 'learning_rate/action_model': 4.356947460194694e-05, 'epoch': 0.21}) +[2026-10-03 13:35:29,462][starVLA.training.train_starvla][INFO] - Step 2846, Loss: {'train/loss': 3.291470784461126e-05, 'train/grad_norm_pre_clip': 0.01934031955897808, 'timing/data': 0.0002590500080259517, 'timing/model': 0.39214359599282034, 'learning_rate/qwen_vl_interface': 4.3539565317717446e-06, 'learning_rate/action_model': 4.353956531771745e-05, 'epoch': 0.21}) +[2026-10-03 13:35:30,160][starVLA.training.train_starvla][INFO] - Step 2847, Loss: {'train/loss': 0.038707368075847626, 'train/grad_norm_pre_clip': 26.017410278320312, 'timing/data': 0.0003714800113812089, 'timing/model': 0.3903142960043624, 'learning_rate/qwen_vl_interface': 4.350965971678419e-06, 'learning_rate/action_model': 4.3509659716784185e-05, 'epoch': 0.21}) +[2026-10-03 13:35:30,862][starVLA.training.train_starvla][INFO] - Step 2848, Loss: {'train/loss': 1.2196440366096795e-05, 'train/grad_norm_pre_clip': 0.0039177522994577885, 'timing/data': 0.0003184000088367611, 'timing/model': 0.39216365599713754, 'learning_rate/qwen_vl_interface': 4.3479757811440225e-06, 'learning_rate/action_model': 4.347975781144022e-05, 'epoch': 0.21}) +[2026-10-03 13:35:31,563][starVLA.training.train_starvla][INFO] - Step 2849, Loss: {'train/loss': 0.001098834560252726, 'train/grad_norm_pre_clip': 1.6031559705734253, 'timing/data': 0.0003117500018561259, 'timing/model': 0.3906025460019009, 'learning_rate/qwen_vl_interface': 4.34498596139771e-06, 'learning_rate/action_model': 4.34498596139771e-05, 'epoch': 0.21}) +[2026-10-03 13:35:32,455][starVLA.training.train_starvla][INFO] - Step 2850, Loss: {'train/loss': 4.880094820691738e-06, 'train/grad_norm_pre_clip': 0.0013452101266011596, 'timing/data': 0.0003404000017326325, 'timing/model': 0.5841557239909889, 'learning_rate/qwen_vl_interface': 4.341996513668483e-06, 'learning_rate/action_model': 4.3419965136684824e-05, 'epoch': 0.21}) +[2026-10-03 13:35:33,153][starVLA.training.train_starvla][INFO] - Step 2851, Loss: {'train/loss': 2.8115668101236224e-05, 'train/grad_norm_pre_clip': 0.01110717374831438, 'timing/data': 0.00033464000443927944, 'timing/model': 0.3896670760004781, 'learning_rate/qwen_vl_interface': 4.33900743918519e-06, 'learning_rate/action_model': 4.33900743918519e-05, 'epoch': 0.21}) +[2026-10-03 13:35:33,856][starVLA.training.train_starvla][INFO] - Step 2852, Loss: {'train/loss': 0.00010545145778451115, 'train/grad_norm_pre_clip': 0.07899324595928192, 'timing/data': 0.00022910999541636556, 'timing/model': 0.39245318599569146, 'learning_rate/qwen_vl_interface': 4.336018739176529e-06, 'learning_rate/action_model': 4.33601873917653e-05, 'epoch': 0.21}) +[2026-10-03 13:35:34,556][starVLA.training.train_starvla][INFO] - Step 2853, Loss: {'train/loss': 3.813038347288966e-05, 'train/grad_norm_pre_clip': 0.012191982008516788, 'timing/data': 0.00032691001251805574, 'timing/model': 0.39086293600848876, 'learning_rate/qwen_vl_interface': 4.333030414871041e-06, 'learning_rate/action_model': 4.33303041487104e-05, 'epoch': 0.21}) +[2026-10-03 13:35:35,257][starVLA.training.train_starvla][INFO] - Step 2854, Loss: {'train/loss': 1.6562375094508752e-05, 'train/grad_norm_pre_clip': 0.007161459419876337, 'timing/data': 0.00019096999312750995, 'timing/model': 0.3928458860027604, 'learning_rate/qwen_vl_interface': 4.330042467497111e-06, 'learning_rate/action_model': 4.330042467497111e-05, 'epoch': 0.21}) +[2026-10-03 13:35:36,148][starVLA.training.train_starvla][INFO] - Step 2855, Loss: {'train/loss': 2.4950684746727347e-05, 'train/grad_norm_pre_clip': 0.01588216796517372, 'timing/data': 0.0002936200035037473, 'timing/model': 0.5827627039980143, 'learning_rate/qwen_vl_interface': 4.327054898282973e-06, 'learning_rate/action_model': 4.3270548982829727e-05, 'epoch': 0.21}) +[2026-10-03 13:35:36,850][starVLA.training.train_starvla][INFO] - Step 2856, Loss: {'train/loss': 1.892380169010721e-05, 'train/grad_norm_pre_clip': 0.010457138530910015, 'timing/data': 0.000315569996018894, 'timing/model': 0.39210432600521017, 'learning_rate/qwen_vl_interface': 4.324067708456703e-06, 'learning_rate/action_model': 4.3240677084567036e-05, 'epoch': 0.21}) +[2026-10-03 13:35:37,551][starVLA.training.train_starvla][INFO] - Step 2857, Loss: {'train/loss': 0.00023699522716924548, 'train/grad_norm_pre_clip': 0.05131081864237785, 'timing/data': 0.000375300005543977, 'timing/model': 0.39077316600014456, 'learning_rate/qwen_vl_interface': 4.321080899246225e-06, 'learning_rate/action_model': 4.321080899246225e-05, 'epoch': 0.21}) +[2026-10-03 13:35:38,252][starVLA.training.train_starvla][INFO] - Step 2858, Loss: {'train/loss': 9.059747753781267e-06, 'train/grad_norm_pre_clip': 0.002950612222775817, 'timing/data': 0.0002073700015898794, 'timing/model': 0.39265278598759323, 'learning_rate/qwen_vl_interface': 4.318094471879299e-06, 'learning_rate/action_model': 4.3180944718792985e-05, 'epoch': 0.21}) +[2026-10-03 13:35:38,951][starVLA.training.train_starvla][INFO] - Step 2859, Loss: {'train/loss': 0.005800766870379448, 'train/grad_norm_pre_clip': 39.27548599243164, 'timing/data': 0.00023582999710924923, 'timing/model': 0.39090473600663245, 'learning_rate/qwen_vl_interface': 4.315108427583535e-06, 'learning_rate/action_model': 4.315108427583535e-05, 'epoch': 0.21}) +[2026-10-03 13:35:39,661][starVLA.training.train_starvla][INFO] - Step 2860, Loss: {'train/loss': 0.3477441072463989, 'train/grad_norm_pre_clip': 371.5636291503906, 'timing/data': 0.00021769999875687063, 'timing/model': 0.39651615600450896, 'learning_rate/qwen_vl_interface': 4.312122767586381e-06, 'learning_rate/action_model': 4.3121227675863815e-05, 'epoch': 0.21}) +[2026-10-03 13:35:40,565][starVLA.training.train_starvla][INFO] - Step 2861, Loss: {'train/loss': 0.45239490270614624, 'train/grad_norm_pre_clip': 278.1979675292969, 'timing/data': 0.00034919999598059803, 'timing/model': 0.39551538600062486, 'learning_rate/qwen_vl_interface': 4.3091374931151324e-06, 'learning_rate/action_model': 4.309137493115132e-05, 'epoch': 0.21}) +[2026-10-03 13:35:41,274][starVLA.training.train_starvla][INFO] - Step 2862, Loss: {'train/loss': 0.0014831613516435027, 'train/grad_norm_pre_clip': 4.874266147613525, 'timing/data': 0.00033055999665521085, 'timing/model': 0.396153656009119, 'learning_rate/qwen_vl_interface': 4.306152605396921e-06, 'learning_rate/action_model': 4.306152605396921e-05, 'epoch': 0.21}) +[2026-10-03 13:35:41,982][starVLA.training.train_starvla][INFO] - Step 2863, Loss: {'train/loss': 7.189755251602037e-06, 'train/grad_norm_pre_clip': 0.0024266676045954227, 'timing/data': 0.0003556199953891337, 'timing/model': 0.3952297360083321, 'learning_rate/qwen_vl_interface': 4.30316810565872e-06, 'learning_rate/action_model': 4.3031681056587206e-05, 'epoch': 0.21}) +[2026-10-03 13:35:42,691][starVLA.training.train_starvla][INFO] - Step 2864, Loss: {'train/loss': 4.880103006144054e-06, 'train/grad_norm_pre_clip': 0.0021418025717139244, 'timing/data': 0.00036238000029698014, 'timing/model': 0.3956322259909939, 'learning_rate/qwen_vl_interface': 4.300183995127347e-06, 'learning_rate/action_model': 4.300183995127346e-05, 'epoch': 0.21}) +[2026-10-03 13:35:43,394][starVLA.training.train_starvla][INFO] - Step 2865, Loss: {'train/loss': 0.8169206380844116, 'train/grad_norm_pre_clip': 90.82977294921875, 'timing/data': 0.00034440000308677554, 'timing/model': 0.39253404599730857, 'learning_rate/qwen_vl_interface': 4.297200275029454e-06, 'learning_rate/action_model': 4.297200275029454e-05, 'epoch': 0.21}) +[2026-10-03 13:35:44,287][starVLA.training.train_starvla][INFO] - Step 2866, Loss: {'train/loss': 0.09718196839094162, 'train/grad_norm_pre_clip': 116.58385467529297, 'timing/data': 0.0003580799966584891, 'timing/model': 0.5815915349958232, 'learning_rate/qwen_vl_interface': 4.29421694659154e-06, 'learning_rate/action_model': 4.294216946591539e-05, 'epoch': 0.21}) +[2026-10-03 13:35:44,990][starVLA.training.train_starvla][INFO] - Step 2867, Loss: {'train/loss': 7.91593556641601e-05, 'train/grad_norm_pre_clip': 0.013149612583220005, 'timing/data': 0.0004013699945062399, 'timing/model': 0.3928534970036708, 'learning_rate/qwen_vl_interface': 4.2912340110399345e-06, 'learning_rate/action_model': 4.291234011039934e-05, 'epoch': 0.21}) +[2026-10-03 13:35:45,695][starVLA.training.train_starvla][INFO] - Step 2868, Loss: {'train/loss': 1.5839410480111837e-05, 'train/grad_norm_pre_clip': 0.005811384413391352, 'timing/data': 0.0003058399888686836, 'timing/model': 0.3931152059958549, 'learning_rate/qwen_vl_interface': 4.2882514696008125e-06, 'learning_rate/action_model': 4.2882514696008125e-05, 'epoch': 0.21}) +[2026-10-03 13:35:46,398][starVLA.training.train_starvla][INFO] - Step 2869, Loss: {'train/loss': 5.9979778598062694e-05, 'train/grad_norm_pre_clip': 0.009814473800361156, 'timing/data': 0.0002794800093397498, 'timing/model': 0.393018665999989, 'learning_rate/qwen_vl_interface': 4.2852693235001815e-06, 'learning_rate/action_model': 4.2852693235001815e-05, 'epoch': 0.21}) +[2026-10-03 13:35:47,106][starVLA.training.train_starvla][INFO] - Step 2870, Loss: {'train/loss': 0.004328953102231026, 'train/grad_norm_pre_clip': 30.57897186279297, 'timing/data': 0.00033183000050485134, 'timing/model': 0.39512203600315843, 'learning_rate/qwen_vl_interface': 4.2822875739638915e-06, 'learning_rate/action_model': 4.282287573963891e-05, 'epoch': 0.21}) +[2026-10-03 13:35:47,809][starVLA.training.train_starvla][INFO] - Step 2871, Loss: {'train/loss': 6.0091693740105256e-05, 'train/grad_norm_pre_clip': 0.01578551158308983, 'timing/data': 0.00027224999212194234, 'timing/model': 0.3931722559937043, 'learning_rate/qwen_vl_interface': 4.279306222217625e-06, 'learning_rate/action_model': 4.279306222217626e-05, 'epoch': 0.21}) +[2026-10-03 13:35:48,703][starVLA.training.train_starvla][INFO] - Step 2872, Loss: {'train/loss': 2.8591348382178694e-05, 'train/grad_norm_pre_clip': 0.011715894564986229, 'timing/data': 0.00027622899506241083, 'timing/model': 0.3943732170009753, 'learning_rate/qwen_vl_interface': 4.276325269486906e-06, 'learning_rate/action_model': 4.276325269486906e-05, 'epoch': 0.21}) +[2026-10-03 13:35:49,409][starVLA.training.train_starvla][INFO] - Step 2873, Loss: {'train/loss': 0.00015995677676983178, 'train/grad_norm_pre_clip': 0.15324722230434418, 'timing/data': 0.00030235000303946435, 'timing/model': 0.3939054560032673, 'learning_rate/qwen_vl_interface': 4.273344716997089e-06, 'learning_rate/action_model': 4.273344716997089e-05, 'epoch': 0.21}) +[2026-10-03 13:35:50,116][starVLA.training.train_starvla][INFO] - Step 2874, Loss: {'train/loss': 9.78968819254078e-06, 'train/grad_norm_pre_clip': 0.0048780012875795364, 'timing/data': 0.0002618099970277399, 'timing/model': 0.3949983859929489, 'learning_rate/qwen_vl_interface': 4.270364565973368e-06, 'learning_rate/action_model': 4.270364565973367e-05, 'epoch': 0.21}) +[2026-10-03 13:35:50,820][starVLA.training.train_starvla][INFO] - Step 2875, Loss: {'train/loss': 2.1612011551042087e-05, 'train/grad_norm_pre_clip': 0.008161856792867184, 'timing/data': 0.0002808400022331625, 'timing/model': 0.3931755770026939, 'learning_rate/qwen_vl_interface': 4.267384817640768e-06, 'learning_rate/action_model': 4.2673848176407685e-05, 'epoch': 0.21}) +[2026-10-03 13:35:51,527][starVLA.training.train_starvla][INFO] - Step 2876, Loss: {'train/loss': 0.00941415037959814, 'train/grad_norm_pre_clip': 38.05426788330078, 'timing/data': 0.00026822999643627554, 'timing/model': 0.3952792159980163, 'learning_rate/qwen_vl_interface': 4.264405473224155e-06, 'learning_rate/action_model': 4.2644054732241546e-05, 'epoch': 0.21}) +[2026-10-03 13:35:52,425][starVLA.training.train_starvla][INFO] - Step 2877, Loss: {'train/loss': 3.5037115594604984e-05, 'train/grad_norm_pre_clip': 0.015093921683728695, 'timing/data': 0.00030170999525580555, 'timing/model': 0.39405325600819197, 'learning_rate/qwen_vl_interface': 4.261426533948221e-06, 'learning_rate/action_model': 4.2614265339482213e-05, 'epoch': 0.21}) +[2026-10-03 13:35:53,130][starVLA.training.train_starvla][INFO] - Step 2878, Loss: {'train/loss': 5.275238072499633e-05, 'train/grad_norm_pre_clip': 0.008626345545053482, 'timing/data': 0.0002661500038811937, 'timing/model': 0.39467860500735696, 'learning_rate/qwen_vl_interface': 4.258448001037499e-06, 'learning_rate/action_model': 4.2584480010374985e-05, 'epoch': 0.21}) +[2026-10-03 13:35:53,834][starVLA.training.train_starvla][INFO] - Step 2879, Loss: {'train/loss': 1.7590184143045917e-05, 'train/grad_norm_pre_clip': 0.006584830116480589, 'timing/data': 0.00030762000824324787, 'timing/model': 0.39312393600994255, 'learning_rate/qwen_vl_interface': 4.255469875716348e-06, 'learning_rate/action_model': 4.255469875716348e-05, 'epoch': 0.21}) +[2026-10-03 13:35:54,540][starVLA.training.train_starvla][INFO] - Step 2880, Loss: {'train/loss': 1.0415504220873117e-05, 'train/grad_norm_pre_clip': 0.003073540050536394, 'timing/data': 0.0002597800048533827, 'timing/model': 0.3946763259882573, 'learning_rate/qwen_vl_interface': 4.252492159208966e-06, 'learning_rate/action_model': 4.252492159208966e-05, 'epoch': 0.21}) +[2026-10-03 13:35:55,246][starVLA.training.train_starvla][INFO] - Step 2881, Loss: {'train/loss': 2.749255827438901e-06, 'train/grad_norm_pre_clip': 0.0008080420666374266, 'timing/data': 0.00026497000362724066, 'timing/model': 0.3946698459913023, 'learning_rate/qwen_vl_interface': 4.2495148527393766e-06, 'learning_rate/action_model': 4.2495148527393766e-05, 'epoch': 0.21}) +[2026-10-03 13:35:56,141][starVLA.training.train_starvla][INFO] - Step 2882, Loss: {'train/loss': 1.2993566997465678e-05, 'train/grad_norm_pre_clip': 0.0037885617930442095, 'timing/data': 0.0002406999992672354, 'timing/model': 0.5836043760064058, 'learning_rate/qwen_vl_interface': 4.246537957531441e-06, 'learning_rate/action_model': 4.2465379575314404e-05, 'epoch': 0.21}) +[2026-10-03 13:35:56,845][starVLA.training.train_starvla][INFO] - Step 2883, Loss: {'train/loss': 0.7568861842155457, 'train/grad_norm_pre_clip': 334.23345947265625, 'timing/data': 0.00020622000738512725, 'timing/model': 0.3934940960025415, 'learning_rate/qwen_vl_interface': 4.243561474808846e-06, 'learning_rate/action_model': 4.2435614748088456e-05, 'epoch': 0.21}) +[2026-10-03 13:35:57,552][starVLA.training.train_starvla][INFO] - Step 2884, Loss: {'train/loss': 1.3216895240475424e-05, 'train/grad_norm_pre_clip': 0.00442046532407403, 'timing/data': 0.0003302099939901382, 'timing/model': 0.39448739599902183, 'learning_rate/qwen_vl_interface': 4.240585405795112e-06, 'learning_rate/action_model': 4.240585405795112e-05, 'epoch': 0.21}) +[2026-10-03 13:35:58,258][starVLA.training.train_starvla][INFO] - Step 2885, Loss: {'train/loss': 1.2858970876550302e-05, 'train/grad_norm_pre_clip': 0.005215812008827925, 'timing/data': 0.0002902999985963106, 'timing/model': 0.3940785960003268, 'learning_rate/qwen_vl_interface': 4.237609751713589e-06, 'learning_rate/action_model': 4.2376097517135896e-05, 'epoch': 0.21}) +[2026-10-03 13:35:58,963][starVLA.training.train_starvla][INFO] - Step 2886, Loss: {'train/loss': 0.00015409763727802783, 'train/grad_norm_pre_clip': 0.05166597291827202, 'timing/data': 0.0002838200016412884, 'timing/model': 0.3944992159958929, 'learning_rate/qwen_vl_interface': 4.234634513787455e-06, 'learning_rate/action_model': 4.2346345137874556e-05, 'epoch': 0.21}) +[2026-10-03 13:35:59,668][starVLA.training.train_starvla][INFO] - Step 2887, Loss: {'train/loss': 1.2501675882958807e-05, 'train/grad_norm_pre_clip': 0.004480949603021145, 'timing/data': 0.0002826200070558116, 'timing/model': 0.3931764959997963, 'learning_rate/qwen_vl_interface': 4.2316596932397195e-06, 'learning_rate/action_model': 4.2316596932397195e-05, 'epoch': 0.21}) +[2026-10-03 13:36:00,563][starVLA.training.train_starvla][INFO] - Step 2888, Loss: {'train/loss': 1.4699293387820944e-05, 'train/grad_norm_pre_clip': 0.004995663184672594, 'timing/data': 0.00024337999639101326, 'timing/model': 0.39471996600332204, 'learning_rate/qwen_vl_interface': 4.22868529129322e-06, 'learning_rate/action_model': 4.22868529129322e-05, 'epoch': 0.21}) +[2026-10-03 13:36:01,282][starVLA.training.train_starvla][INFO] - Step 2889, Loss: {'train/loss': 9.238713118975284e-07, 'train/grad_norm_pre_clip': 0.000190119375474751, 'timing/data': 0.00032083000405691564, 'timing/model': 0.3938976960052969, 'learning_rate/qwen_vl_interface': 4.2257113091706165e-06, 'learning_rate/action_model': 4.2257113091706165e-05, 'epoch': 0.21}) +[2026-10-03 13:36:01,985][starVLA.training.train_starvla][INFO] - Step 2890, Loss: {'train/loss': 0.00038419870543293655, 'train/grad_norm_pre_clip': 1.1486423015594482, 'timing/data': 0.0002895799989346415, 'timing/model': 0.39268658599758055, 'learning_rate/qwen_vl_interface': 4.222737748094404e-06, 'learning_rate/action_model': 4.222737748094404e-05, 'epoch': 0.21}) +[2026-10-03 13:36:02,687][starVLA.training.train_starvla][INFO] - Step 2891, Loss: {'train/loss': 4.2018778913188726e-05, 'train/grad_norm_pre_clip': 0.008979639038443565, 'timing/data': 0.0002872100012609735, 'timing/model': 0.39284170599421486, 'learning_rate/qwen_vl_interface': 4.219764609286901e-06, 'learning_rate/action_model': 4.219764609286901e-05, 'epoch': 0.21}) +[2026-10-03 13:36:03,393][starVLA.training.train_starvla][INFO] - Step 2892, Loss: {'train/loss': 3.44212276104372e-06, 'train/grad_norm_pre_clip': 0.001080513233318925, 'timing/data': 0.00030975999834481627, 'timing/model': 0.3936511259962572, 'learning_rate/qwen_vl_interface': 4.21679189397025e-06, 'learning_rate/action_model': 4.21679189397025e-05, 'epoch': 0.21}) +[2026-10-03 13:36:04,300][starVLA.training.train_starvla][INFO] - Step 2893, Loss: {'train/loss': 0.00011681925388984382, 'train/grad_norm_pre_clip': 0.1084553673863411, 'timing/data': 0.0002946699969470501, 'timing/model': 0.3933246560045518, 'learning_rate/qwen_vl_interface': 4.213819603366424e-06, 'learning_rate/action_model': 4.213819603366424e-05, 'epoch': 0.21}) +[2026-10-03 13:36:05,005][starVLA.training.train_starvla][INFO] - Step 2894, Loss: {'train/loss': 0.3578605353832245, 'train/grad_norm_pre_clip': 133.55038452148438, 'timing/data': 0.0002361899969400838, 'timing/model': 0.393788945992128, 'learning_rate/qwen_vl_interface': 4.210847738697219e-06, 'learning_rate/action_model': 4.210847738697218e-05, 'epoch': 0.21}) +[2026-10-03 13:36:05,708][starVLA.training.train_starvla][INFO] - Step 2895, Loss: {'train/loss': 0.27678364515304565, 'train/grad_norm_pre_clip': 178.8301544189453, 'timing/data': 0.00019777999841608107, 'timing/model': 0.3928839760046685, 'learning_rate/qwen_vl_interface': 4.207876301184256e-06, 'learning_rate/action_model': 4.207876301184256e-05, 'epoch': 0.21}) +[2026-10-03 13:36:06,414][starVLA.training.train_starvla][INFO] - Step 2896, Loss: {'train/loss': 2.2848116714158095e-05, 'train/grad_norm_pre_clip': 0.0037357527762651443, 'timing/data': 0.0002583599998615682, 'timing/model': 0.3942883760028053, 'learning_rate/qwen_vl_interface': 4.2049052920489825e-06, 'learning_rate/action_model': 4.2049052920489825e-05, 'epoch': 0.21}) +[2026-10-03 13:36:07,135][starVLA.training.train_starvla][INFO] - Step 2897, Loss: {'train/loss': 4.895745951216668e-05, 'train/grad_norm_pre_clip': 0.022855529561638832, 'timing/data': 0.0003218199999537319, 'timing/model': 0.39543990600213874, 'learning_rate/qwen_vl_interface': 4.201934712512666e-06, 'learning_rate/action_model': 4.2019347125126656e-05, 'epoch': 0.21}) +[2026-10-03 13:36:08,035][starVLA.training.train_starvla][INFO] - Step 2898, Loss: {'train/loss': 4.850148252444342e-05, 'train/grad_norm_pre_clip': 0.009399596601724625, 'timing/data': 0.0003539700119290501, 'timing/model': 0.3957663259934634, 'learning_rate/qwen_vl_interface': 4.1989645637964e-06, 'learning_rate/action_model': 4.1989645637964e-05, 'epoch': 0.21}) +[2026-10-03 13:36:08,741][starVLA.training.train_starvla][INFO] - Step 2899, Loss: {'train/loss': 0.5752686858177185, 'train/grad_norm_pre_clip': 79.67044830322266, 'timing/data': 0.0002569399948697537, 'timing/model': 0.3938065559923416, 'learning_rate/qwen_vl_interface': 4.195994847121101e-06, 'learning_rate/action_model': 4.195994847121101e-05, 'epoch': 0.21}) +[2026-10-03 13:36:09,449][starVLA.training.train_starvla][INFO] - Step 2900, Loss: {'train/loss': 3.8601530832238495e-05, 'train/grad_norm_pre_clip': 0.02098861150443554, 'timing/data': 0.0003025900077773258, 'timing/model': 0.39560295500268694, 'learning_rate/qwen_vl_interface': 4.193025563707507e-06, 'learning_rate/action_model': 4.1930255637075066e-05, 'epoch': 0.21}) +[2026-10-03 13:36:10,157][starVLA.training.train_starvla][INFO] - Step 2901, Loss: {'train/loss': 0.08941544592380524, 'train/grad_norm_pre_clip': 488.3879699707031, 'timing/data': 0.00034032999246846884, 'timing/model': 0.3944134860066697, 'learning_rate/qwen_vl_interface': 4.19005671477618e-06, 'learning_rate/action_model': 4.1900567147761794e-05, 'epoch': 0.21}) +[2026-10-03 13:36:10,865][starVLA.training.train_starvla][INFO] - Step 2902, Loss: {'train/loss': 0.00011826565605588257, 'train/grad_norm_pre_clip': 0.07737074792385101, 'timing/data': 0.00022182999236974865, 'timing/model': 0.3962762660084991, 'learning_rate/qwen_vl_interface': 4.1870883015474975e-06, 'learning_rate/action_model': 4.187088301547497e-05, 'epoch': 0.21}) +[2026-10-03 13:36:11,761][starVLA.training.train_starvla][INFO] - Step 2903, Loss: {'train/loss': 1.1101353720732732e-06, 'train/grad_norm_pre_clip': 0.00045640862663276494, 'timing/data': 0.00020778999896720052, 'timing/model': 0.5841948540037265, 'learning_rate/qwen_vl_interface': 4.1841203252416655e-06, 'learning_rate/action_model': 4.1841203252416655e-05, 'epoch': 0.21}) +[2026-10-03 13:36:12,471][starVLA.training.train_starvla][INFO] - Step 2904, Loss: {'train/loss': 3.934423148166388e-05, 'train/grad_norm_pre_clip': 0.048451878130435944, 'timing/data': 0.00035242999729234725, 'timing/model': 0.39702465500158723, 'learning_rate/qwen_vl_interface': 4.181152787078705e-06, 'learning_rate/action_model': 4.1811527870787056e-05, 'epoch': 0.21}) +[2026-10-03 13:36:13,178][starVLA.training.train_starvla][INFO] - Step 2905, Loss: {'train/loss': 0.00019523946684785187, 'train/grad_norm_pre_clip': 0.14502167701721191, 'timing/data': 0.00031694999779574573, 'timing/model': 0.39424229599535465, 'learning_rate/qwen_vl_interface': 4.178185688278462e-06, 'learning_rate/action_model': 4.1781856882784616e-05, 'epoch': 0.21}) +[2026-10-03 13:36:13,885][starVLA.training.train_starvla][INFO] - Step 2906, Loss: {'train/loss': 0.30704569816589355, 'train/grad_norm_pre_clip': 147.6388702392578, 'timing/data': 0.00028433000261429697, 'timing/model': 0.39586043600866105, 'learning_rate/qwen_vl_interface': 4.175219030060595e-06, 'learning_rate/action_model': 4.175219030060595e-05, 'epoch': 0.21}) +[2026-10-03 13:36:14,589][starVLA.training.train_starvla][INFO] - Step 2907, Loss: {'train/loss': 0.02835986763238907, 'train/grad_norm_pre_clip': 325.0999755859375, 'timing/data': 0.00020478999067563564, 'timing/model': 0.39325691599515267, 'learning_rate/qwen_vl_interface': 4.1722528136445875e-06, 'learning_rate/action_model': 4.172252813644588e-05, 'epoch': 0.21}) +[2026-10-03 13:36:15,482][starVLA.training.train_starvla][INFO] - Step 2908, Loss: {'train/loss': 0.0003413965168874711, 'train/grad_norm_pre_clip': 0.7274419069290161, 'timing/data': 0.00022045000514481217, 'timing/model': 0.5820057650125818, 'learning_rate/qwen_vl_interface': 4.169287040249738e-06, 'learning_rate/action_model': 4.169287040249738e-05, 'epoch': 0.21}) +[2026-10-03 13:36:16,186][starVLA.training.train_starvla][INFO] - Step 2909, Loss: {'train/loss': 2.9374197765719146e-05, 'train/grad_norm_pre_clip': 0.007002606987953186, 'timing/data': 0.00020509000751189888, 'timing/model': 0.39304006700695027, 'learning_rate/qwen_vl_interface': 4.166321711095164e-06, 'learning_rate/action_model': 4.166321711095164e-05, 'epoch': 0.21}) +[2026-10-03 13:36:16,893][starVLA.training.train_starvla][INFO] - Step 2910, Loss: {'train/loss': 2.7981879611616023e-05, 'train/grad_norm_pre_clip': 0.014790026471018791, 'timing/data': 0.00035586000012699515, 'timing/model': 0.3949545060022501, 'learning_rate/qwen_vl_interface': 4.1633568273998e-06, 'learning_rate/action_model': 4.1633568273998e-05, 'epoch': 0.21}) +[2026-10-03 13:36:17,596][starVLA.training.train_starvla][INFO] - Step 2911, Loss: {'train/loss': 1.2061675988661591e-05, 'train/grad_norm_pre_clip': 0.0019838670268654823, 'timing/data': 0.00026342000637669116, 'timing/model': 0.3925896360015031, 'learning_rate/qwen_vl_interface': 4.1603923903823995e-06, 'learning_rate/action_model': 4.160392390382399e-05, 'epoch': 0.21}) +[2026-10-03 13:36:18,300][starVLA.training.train_starvla][INFO] - Step 2912, Loss: {'train/loss': 0.00615866482257843, 'train/grad_norm_pre_clip': 15.391891479492188, 'timing/data': 0.00029229000210762024, 'timing/model': 0.39324130599561613, 'learning_rate/qwen_vl_interface': 4.1574284012615276e-06, 'learning_rate/action_model': 4.157428401261528e-05, 'epoch': 0.21}) +[2026-10-03 13:36:19,003][starVLA.training.train_starvla][INFO] - Step 2913, Loss: {'train/loss': 3.6879796425637323e-06, 'train/grad_norm_pre_clip': 0.0020758386235684156, 'timing/data': 0.00027034999220632017, 'timing/model': 0.39306256599957123, 'learning_rate/qwen_vl_interface': 4.154464861255571e-06, 'learning_rate/action_model': 4.154464861255571e-05, 'epoch': 0.21}) +[2026-10-03 13:36:19,906][starVLA.training.train_starvla][INFO] - Step 2914, Loss: {'train/loss': 1.8073244063998573e-05, 'train/grad_norm_pre_clip': 0.010509314015507698, 'timing/data': 0.0003167300019413233, 'timing/model': 0.3938809960091021, 'learning_rate/qwen_vl_interface': 4.151501771582728e-06, 'learning_rate/action_model': 4.151501771582728e-05, 'epoch': 0.21}) +[2026-10-03 13:36:20,610][starVLA.training.train_starvla][INFO] - Step 2915, Loss: {'train/loss': 0.4507249593734741, 'train/grad_norm_pre_clip': 58.216800689697266, 'timing/data': 0.0003089099918724969, 'timing/model': 0.39329094599816017, 'learning_rate/qwen_vl_interface': 4.148539133461012e-06, 'learning_rate/action_model': 4.1485391334610116e-05, 'epoch': 0.21}) +[2026-10-03 13:36:21,313][starVLA.training.train_starvla][INFO] - Step 2916, Loss: {'train/loss': 5.3718267736257985e-06, 'train/grad_norm_pre_clip': 0.0014340380439534783, 'timing/data': 0.00033989000075962394, 'timing/model': 0.39298021599825006, 'learning_rate/qwen_vl_interface': 4.145576948108254e-06, 'learning_rate/action_model': 4.145576948108253e-05, 'epoch': 0.21}) +[2026-10-03 13:36:22,017][starVLA.training.train_starvla][INFO] - Step 2917, Loss: {'train/loss': 4.663559593609534e-05, 'train/grad_norm_pre_clip': 0.012921257875859737, 'timing/data': 0.00020763899374287575, 'timing/model': 0.3933484869921813, 'learning_rate/qwen_vl_interface': 4.142615216742093e-06, 'learning_rate/action_model': 4.142615216742093e-05, 'epoch': 0.21}) +[2026-10-03 13:36:22,723][starVLA.training.train_starvla][INFO] - Step 2918, Loss: {'train/loss': 0.0047548264265060425, 'train/grad_norm_pre_clip': 14.38846492767334, 'timing/data': 0.0004017000028397888, 'timing/model': 0.3938682459993288, 'learning_rate/qwen_vl_interface': 4.139653940579988e-06, 'learning_rate/action_model': 4.1396539405799874e-05, 'epoch': 0.21}) +[2026-10-03 13:36:23,613][starVLA.training.train_starvla][INFO] - Step 2919, Loss: {'train/loss': 4.6192876652639825e-06, 'train/grad_norm_pre_clip': 0.0015001081628724933, 'timing/data': 0.0002013890043599531, 'timing/model': 0.5799079049902502, 'learning_rate/qwen_vl_interface': 4.136693120839206e-06, 'learning_rate/action_model': 4.136693120839205e-05, 'epoch': 0.21}) +[2026-10-03 13:36:24,317][starVLA.training.train_starvla][INFO] - Step 2920, Loss: {'train/loss': 4.062289372086525e-05, 'train/grad_norm_pre_clip': 0.014619036577641964, 'timing/data': 0.0002112299989676103, 'timing/model': 0.39342684599978384, 'learning_rate/qwen_vl_interface': 4.133732758736828e-06, 'learning_rate/action_model': 4.133732758736828e-05, 'epoch': 0.21}) +[2026-10-03 13:36:25,020][starVLA.training.train_starvla][INFO] - Step 2921, Loss: {'train/loss': 1.069103745976463e-05, 'train/grad_norm_pre_clip': 0.008700272999703884, 'timing/data': 0.00025402900064364076, 'timing/model': 0.3928119570045965, 'learning_rate/qwen_vl_interface': 4.13077285548975e-06, 'learning_rate/action_model': 4.13077285548975e-05, 'epoch': 0.21}) +[2026-10-03 13:36:25,726][starVLA.training.train_starvla][INFO] - Step 2922, Loss: {'train/loss': 2.3554519430035725e-05, 'train/grad_norm_pre_clip': 0.0047157839871943, 'timing/data': 0.0003145099908579141, 'timing/model': 0.39410893600143027, 'learning_rate/qwen_vl_interface': 4.127813412314672e-06, 'learning_rate/action_model': 4.1278134123146725e-05, 'epoch': 0.21}) +[2026-10-03 13:36:26,429][starVLA.training.train_starvla][INFO] - Step 2923, Loss: {'train/loss': 0.27633821964263916, 'train/grad_norm_pre_clip': 46.741966247558594, 'timing/data': 0.0002487799938535318, 'timing/model': 0.39280943700578064, 'learning_rate/qwen_vl_interface': 4.124854430428113e-06, 'learning_rate/action_model': 4.1248544304281126e-05, 'epoch': 0.21}) +[2026-10-03 13:36:27,132][starVLA.training.train_starvla][INFO] - Step 2924, Loss: {'train/loss': 2.828482502081897e-05, 'train/grad_norm_pre_clip': 0.010627375915646553, 'timing/data': 0.0002728399995248765, 'timing/model': 0.3931537059979746, 'learning_rate/qwen_vl_interface': 4.121895911046397e-06, 'learning_rate/action_model': 4.121895911046396e-05, 'epoch': 0.21}) +[2026-10-03 13:36:28,022][starVLA.training.train_starvla][INFO] - Step 2925, Loss: {'train/loss': 0.00020948139717802405, 'train/grad_norm_pre_clip': 0.24922527372837067, 'timing/data': 0.0003279100055806339, 'timing/model': 0.3928765460004797, 'learning_rate/qwen_vl_interface': 4.118937855385659e-06, 'learning_rate/action_model': 4.1189378553856584e-05, 'epoch': 0.21}) +[2026-10-03 13:36:28,727][starVLA.training.train_starvla][INFO] - Step 2926, Loss: {'train/loss': 0.000279843108728528, 'train/grad_norm_pre_clip': 0.17050983011722565, 'timing/data': 0.00021926000772509724, 'timing/model': 0.3929332560073817, 'learning_rate/qwen_vl_interface': 4.115980264661844e-06, 'learning_rate/action_model': 4.115980264661844e-05, 'epoch': 0.21}) +[2026-10-03 13:36:29,431][starVLA.training.train_starvla][INFO] - Step 2927, Loss: {'train/loss': 4.741067823488265e-05, 'train/grad_norm_pre_clip': 0.008525530807673931, 'timing/data': 0.00025329999334644526, 'timing/model': 0.39261710600112565, 'learning_rate/qwen_vl_interface': 4.113023140090709e-06, 'learning_rate/action_model': 4.113023140090709e-05, 'epoch': 0.21}) +[2026-10-03 13:36:30,134][starVLA.training.train_starvla][INFO] - Step 2928, Loss: {'train/loss': 0.005172360222786665, 'train/grad_norm_pre_clip': 5.924326419830322, 'timing/data': 0.0002059800026472658, 'timing/model': 0.39241125600528903, 'learning_rate/qwen_vl_interface': 4.110066482887812e-06, 'learning_rate/action_model': 4.110066482887812e-05, 'epoch': 0.21}) +[2026-10-03 13:36:30,836][starVLA.training.train_starvla][INFO] - Step 2929, Loss: {'train/loss': 0.5949842929840088, 'train/grad_norm_pre_clip': 36.099281311035156, 'timing/data': 0.00020353999570943415, 'timing/model': 0.39261440600967035, 'learning_rate/qwen_vl_interface': 4.107110294268525e-06, 'learning_rate/action_model': 4.107110294268525e-05, 'epoch': 0.21}) +[2026-10-03 13:36:31,731][starVLA.training.train_starvla][INFO] - Step 2930, Loss: {'train/loss': 5.5059167607396375e-06, 'train/grad_norm_pre_clip': 0.0015771050238981843, 'timing/data': 0.0003746900038095191, 'timing/model': 0.393332505991566, 'learning_rate/qwen_vl_interface': 4.104154575448026e-06, 'learning_rate/action_model': 4.104154575448025e-05, 'epoch': 0.21}) +[2026-10-03 13:36:32,432][starVLA.training.train_starvla][INFO] - Step 2931, Loss: {'train/loss': 4.906315734842792e-05, 'train/grad_norm_pre_clip': 0.013250559568405151, 'timing/data': 0.00023893899924587458, 'timing/model': 0.3922565469983965, 'learning_rate/qwen_vl_interface': 4.101199327641297e-06, 'learning_rate/action_model': 4.101199327641297e-05, 'epoch': 0.21}) +[2026-10-03 13:36:33,135][starVLA.training.train_starvla][INFO] - Step 2932, Loss: {'train/loss': 0.4294629991054535, 'train/grad_norm_pre_clip': 53.35688400268555, 'timing/data': 0.0002469500032020733, 'timing/model': 0.39212731600855477, 'learning_rate/qwen_vl_interface': 4.098244552063133e-06, 'learning_rate/action_model': 4.098244552063132e-05, 'epoch': 0.21}) +[2026-10-03 13:36:33,838][starVLA.training.train_starvla][INFO] - Step 2933, Loss: {'train/loss': 0.5229276418685913, 'train/grad_norm_pre_clip': 20.626420974731445, 'timing/data': 0.0003135299921268597, 'timing/model': 0.39244243600114714, 'learning_rate/qwen_vl_interface': 4.0952902499281254e-06, 'learning_rate/action_model': 4.095290249928125e-05, 'epoch': 0.21}) +[2026-10-03 13:36:34,542][starVLA.training.train_starvla][INFO] - Step 2934, Loss: {'train/loss': 0.034802865236997604, 'train/grad_norm_pre_clip': 22.041139602661133, 'timing/data': 0.0002820000081555918, 'timing/model': 0.39357476599980146, 'learning_rate/qwen_vl_interface': 4.092336422450681e-06, 'learning_rate/action_model': 4.0923364224506804e-05, 'epoch': 0.21}) +[2026-10-03 13:36:35,432][starVLA.training.train_starvla][INFO] - Step 2935, Loss: {'train/loss': 0.15105384588241577, 'train/grad_norm_pre_clip': 113.33267211914062, 'timing/data': 0.00020284998754505068, 'timing/model': 0.5800591059960425, 'learning_rate/qwen_vl_interface': 4.089383070845003e-06, 'learning_rate/action_model': 4.089383070845003e-05, 'epoch': 0.21}) +[2026-10-03 13:36:36,134][starVLA.training.train_starvla][INFO] - Step 2936, Loss: {'train/loss': 1.3768281860393472e-05, 'train/grad_norm_pre_clip': 0.004332018084824085, 'timing/data': 0.0001994500053115189, 'timing/model': 0.3927372070029378, 'learning_rate/qwen_vl_interface': 4.086430196325106e-06, 'learning_rate/action_model': 4.086430196325106e-05, 'epoch': 0.21}) +[2026-10-03 13:36:36,837][starVLA.training.train_starvla][INFO] - Step 2937, Loss: {'train/loss': 0.7493686676025391, 'train/grad_norm_pre_clip': 272.4417724609375, 'timing/data': 0.0003240899968659505, 'timing/model': 0.3923755370051367, 'learning_rate/qwen_vl_interface': 4.083477800104802e-06, 'learning_rate/action_model': 4.083477800104802e-05, 'epoch': 0.21}) +[2026-10-03 13:36:37,542][starVLA.training.train_starvla][INFO] - Step 2938, Loss: {'train/loss': 0.7171089053153992, 'train/grad_norm_pre_clip': 78.59822082519531, 'timing/data': 0.000354260002495721, 'timing/model': 0.39321226699394174, 'learning_rate/qwen_vl_interface': 4.080525883397711e-06, 'learning_rate/action_model': 4.080525883397711e-05, 'epoch': 0.22}) +[2026-10-03 13:36:38,245][starVLA.training.train_starvla][INFO] - Step 2939, Loss: {'train/loss': 7.022463978501037e-05, 'train/grad_norm_pre_clip': 0.014303474687039852, 'timing/data': 0.000248309996095486, 'timing/model': 0.39243409599293955, 'learning_rate/qwen_vl_interface': 4.077574447417256e-06, 'learning_rate/action_model': 4.077574447417256e-05, 'epoch': 0.22}) +[2026-10-03 13:36:38,948][starVLA.training.train_starvla][INFO] - Step 2940, Loss: {'train/loss': 6.742367258993909e-05, 'train/grad_norm_pre_clip': 0.020699413493275642, 'timing/data': 0.00020550999033730477, 'timing/model': 0.39294347699615173, 'learning_rate/qwen_vl_interface': 4.074623493376657e-06, 'learning_rate/action_model': 4.074623493376657e-05, 'epoch': 0.22}) +[2026-10-03 13:36:39,838][starVLA.training.train_starvla][INFO] - Step 2941, Loss: {'train/loss': 1.665814852458425e-05, 'train/grad_norm_pre_clip': 0.009407538920640945, 'timing/data': 0.00021683999511878937, 'timing/model': 0.39262534699810203, 'learning_rate/qwen_vl_interface': 4.071673022488944e-06, 'learning_rate/action_model': 4.071673022488944e-05, 'epoch': 0.22}) +[2026-10-03 13:36:40,542][starVLA.training.train_starvla][INFO] - Step 2942, Loss: {'train/loss': 0.3256920874118805, 'train/grad_norm_pre_clip': 80.85184478759766, 'timing/data': 0.0002750699932221323, 'timing/model': 0.3935929959989153, 'learning_rate/qwen_vl_interface': 4.06872303596694e-06, 'learning_rate/action_model': 4.06872303596694e-05, 'epoch': 0.22}) +[2026-10-03 13:36:41,245][starVLA.training.train_starvla][INFO] - Step 2943, Loss: {'train/loss': 0.0010022108908742666, 'train/grad_norm_pre_clip': 2.729546070098877, 'timing/data': 0.00027362999389879405, 'timing/model': 0.39276659599272534, 'learning_rate/qwen_vl_interface': 4.065773535023277e-06, 'learning_rate/action_model': 4.0657735350232766e-05, 'epoch': 0.22}) +[2026-10-03 13:36:41,948][starVLA.training.train_starvla][INFO] - Step 2944, Loss: {'train/loss': 9.99715703073889e-05, 'train/grad_norm_pre_clip': 0.05433148145675659, 'timing/data': 0.000305649999063462, 'timing/model': 0.39272482600063086, 'learning_rate/qwen_vl_interface': 4.062824520870382e-06, 'learning_rate/action_model': 4.062824520870381e-05, 'epoch': 0.22}) +[2026-10-03 13:36:42,652][starVLA.training.train_starvla][INFO] - Step 2945, Loss: {'train/loss': 0.00011250277748331428, 'train/grad_norm_pre_clip': 0.024720996618270874, 'timing/data': 0.00025999998615588993, 'timing/model': 0.39306284599297214, 'learning_rate/qwen_vl_interface': 4.059875994720481e-06, 'learning_rate/action_model': 4.059875994720481e-05, 'epoch': 0.22}) +[2026-10-03 13:36:43,544][starVLA.training.train_starvla][INFO] - Step 2946, Loss: {'train/loss': 0.0001740663283271715, 'train/grad_norm_pre_clip': 0.046495016664266586, 'timing/data': 0.00033788000291679054, 'timing/model': 0.3934106160013471, 'learning_rate/qwen_vl_interface': 4.056927957785605e-06, 'learning_rate/action_model': 4.056927957785605e-05, 'epoch': 0.22}) +[2026-10-03 13:36:44,248][starVLA.training.train_starvla][INFO] - Step 2947, Loss: {'train/loss': 0.00012490234803408384, 'train/grad_norm_pre_clip': 0.1872372329235077, 'timing/data': 0.00020836999465245754, 'timing/model': 0.3932532659964636, 'learning_rate/qwen_vl_interface': 4.0539804112775785e-06, 'learning_rate/action_model': 4.053980411277578e-05, 'epoch': 0.22}) +[2026-10-03 13:36:44,952][starVLA.training.train_starvla][INFO] - Step 2948, Loss: {'train/loss': 1.6196943761315197e-05, 'train/grad_norm_pre_clip': 0.0060775321908295155, 'timing/data': 0.00028546999965328723, 'timing/model': 0.39283286599675193, 'learning_rate/qwen_vl_interface': 4.051033356408027e-06, 'learning_rate/action_model': 4.0510333564080264e-05, 'epoch': 0.22}) +[2026-10-03 13:36:45,656][starVLA.training.train_starvla][INFO] - Step 2949, Loss: {'train/loss': 0.00021937943529337645, 'train/grad_norm_pre_clip': 1.2397949695587158, 'timing/data': 0.0002559200074756518, 'timing/model': 0.3926953470072476, 'learning_rate/qwen_vl_interface': 4.048086794388375e-06, 'learning_rate/action_model': 4.048086794388375e-05, 'epoch': 0.22}) +[2026-10-03 13:36:46,362][starVLA.training.train_starvla][INFO] - Step 2950, Loss: {'train/loss': 0.0014433705946430564, 'train/grad_norm_pre_clip': 3.502108097076416, 'timing/data': 0.00032603999716229737, 'timing/model': 0.3939839660015423, 'learning_rate/qwen_vl_interface': 4.045140726429841e-06, 'learning_rate/action_model': 4.0451407264298406e-05, 'epoch': 0.22}) +[2026-10-03 13:36:47,252][starVLA.training.train_starvla][INFO] - Step 2951, Loss: {'train/loss': 0.30630892515182495, 'train/grad_norm_pre_clip': 25.97100257873535, 'timing/data': 0.00025289000768680125, 'timing/model': 0.580670763985836, 'learning_rate/qwen_vl_interface': 4.042195153743443e-06, 'learning_rate/action_model': 4.042195153743443e-05, 'epoch': 0.22}) +[2026-10-03 13:36:47,955][starVLA.training.train_starvla][INFO] - Step 2952, Loss: {'train/loss': 0.00010795413982123137, 'train/grad_norm_pre_clip': 0.022605298087000847, 'timing/data': 0.0003228699933970347, 'timing/model': 0.39262396599224303, 'learning_rate/qwen_vl_interface': 4.039250077539993e-06, 'learning_rate/action_model': 4.039250077539993e-05, 'epoch': 0.22}) +[2026-10-03 13:36:48,658][starVLA.training.train_starvla][INFO] - Step 2953, Loss: {'train/loss': 0.00032592882052995265, 'train/grad_norm_pre_clip': 0.08892209827899933, 'timing/data': 0.0003168300027027726, 'timing/model': 0.3923150160117075, 'learning_rate/qwen_vl_interface': 4.036305499030104e-06, 'learning_rate/action_model': 4.0363054990301036e-05, 'epoch': 0.22}) +[2026-10-03 13:36:49,363][starVLA.training.train_starvla][INFO] - Step 2954, Loss: {'train/loss': 1.2241000149515457e-05, 'train/grad_norm_pre_clip': 0.002684313803911209, 'timing/data': 0.00021912000374868512, 'timing/model': 0.3934473460103618, 'learning_rate/qwen_vl_interface': 4.033361419424178e-06, 'learning_rate/action_model': 4.033361419424178e-05, 'epoch': 0.22}) +[2026-10-03 13:36:50,066][starVLA.training.train_starvla][INFO] - Step 2955, Loss: {'train/loss': 1.049762977345381e-05, 'train/grad_norm_pre_clip': 0.002893542405217886, 'timing/data': 0.0003817100077867508, 'timing/model': 0.39274298600503244, 'learning_rate/qwen_vl_interface': 4.030417839932418e-06, 'learning_rate/action_model': 4.030417839932417e-05, 'epoch': 0.22}) +[2026-10-03 13:36:50,967][starVLA.training.train_starvla][INFO] - Step 2956, Loss: {'train/loss': 0.469858855009079, 'train/grad_norm_pre_clip': 68.86813354492188, 'timing/data': 0.0003179699997417629, 'timing/model': 0.5896843250084203, 'learning_rate/qwen_vl_interface': 4.027474761764815e-06, 'learning_rate/action_model': 4.027474761764814e-05, 'epoch': 0.22}) +[2026-10-03 13:36:51,670][starVLA.training.train_starvla][INFO] - Step 2957, Loss: {'train/loss': 0.33496883511543274, 'train/grad_norm_pre_clip': 49.042518615722656, 'timing/data': 0.00026976999652106315, 'timing/model': 0.39275646599708125, 'learning_rate/qwen_vl_interface': 4.024532186131158e-06, 'learning_rate/action_model': 4.0245321861311575e-05, 'epoch': 0.22}) +[2026-10-03 13:36:52,377][starVLA.training.train_starvla][INFO] - Step 2958, Loss: {'train/loss': 2.6790921765496023e-05, 'train/grad_norm_pre_clip': 0.009931678883731365, 'timing/data': 0.0003069399972446263, 'timing/model': 0.3938338659936562, 'learning_rate/qwen_vl_interface': 4.021590114241029e-06, 'learning_rate/action_model': 4.021590114241029e-05, 'epoch': 0.22}) +[2026-10-03 13:36:53,081][starVLA.training.train_starvla][INFO] - Step 2959, Loss: {'train/loss': 0.00033054285449907184, 'train/grad_norm_pre_clip': 0.10818295180797577, 'timing/data': 0.00021005999587941915, 'timing/model': 0.39353507600026205, 'learning_rate/qwen_vl_interface': 4.018648547303804e-06, 'learning_rate/action_model': 4.0186485473038036e-05, 'epoch': 0.22}) +[2026-10-03 13:36:53,785][starVLA.training.train_starvla][INFO] - Step 2960, Loss: {'train/loss': 2.1716879928135313e-05, 'train/grad_norm_pre_clip': 0.008927213959395885, 'timing/data': 0.00024705998657736927, 'timing/model': 0.39293119599460624, 'learning_rate/qwen_vl_interface': 4.015707486528649e-06, 'learning_rate/action_model': 4.015707486528649e-05, 'epoch': 0.22}) +[2026-10-03 13:36:54,678][starVLA.training.train_starvla][INFO] - Step 2961, Loss: {'train/loss': 2.1731579181505367e-05, 'train/grad_norm_pre_clip': 0.006944421213120222, 'timing/data': 0.00021894001110922545, 'timing/model': 0.5823078050016193, 'learning_rate/qwen_vl_interface': 4.012766933124523e-06, 'learning_rate/action_model': 4.012766933124523e-05, 'epoch': 0.22}) +[2026-10-03 13:36:55,383][starVLA.training.train_starvla][INFO] - Step 2962, Loss: {'train/loss': 0.001820308156311512, 'train/grad_norm_pre_clip': 1.1927330493927002, 'timing/data': 0.00026440000510774553, 'timing/model': 0.393686595998588, 'learning_rate/qwen_vl_interface': 4.009826888300176e-06, 'learning_rate/action_model': 4.0098268883001755e-05, 'epoch': 0.22}) +[2026-10-03 13:36:56,086][starVLA.training.train_starvla][INFO] - Step 2963, Loss: {'train/loss': 7.181763794505969e-05, 'train/grad_norm_pre_clip': 0.03694944828748703, 'timing/data': 0.0002719799958867952, 'timing/model': 0.3931921559997136, 'learning_rate/qwen_vl_interface': 4.00688735326415e-06, 'learning_rate/action_model': 4.0068873532641495e-05, 'epoch': 0.22}) +[2026-10-03 13:36:56,789][starVLA.training.train_starvla][INFO] - Step 2964, Loss: {'train/loss': 1.147375314758392e-05, 'train/grad_norm_pre_clip': 0.0022306174505501986, 'timing/data': 0.00034043999039568007, 'timing/model': 0.39280924599734135, 'learning_rate/qwen_vl_interface': 4.0039483292247765e-06, 'learning_rate/action_model': 4.003948329224776e-05, 'epoch': 0.22}) +[2026-10-03 13:36:57,493][starVLA.training.train_starvla][INFO] - Step 2965, Loss: {'train/loss': 0.0011975757079198956, 'train/grad_norm_pre_clip': 0.7418995499610901, 'timing/data': 0.00026764000358525664, 'timing/model': 0.39258544699987397, 'learning_rate/qwen_vl_interface': 4.001009817390178e-06, 'learning_rate/action_model': 4.001009817390178e-05, 'epoch': 0.22}) +[2026-10-03 13:36:58,199][starVLA.training.train_starvla][INFO] - Step 2966, Loss: {'train/loss': 0.00018271204316988587, 'train/grad_norm_pre_clip': 0.09191156923770905, 'timing/data': 0.00032952000037766993, 'timing/model': 0.39336817599541973, 'learning_rate/qwen_vl_interface': 3.998071818968264e-06, 'learning_rate/action_model': 3.998071818968264e-05, 'epoch': 0.22}) +[2026-10-03 13:36:59,088][starVLA.training.train_starvla][INFO] - Step 2967, Loss: {'train/loss': 4.1294246329925954e-05, 'train/grad_norm_pre_clip': 0.01261389534920454, 'timing/data': 0.0003019699943251908, 'timing/model': 0.39327262599545065, 'learning_rate/qwen_vl_interface': 3.995134335166736e-06, 'learning_rate/action_model': 3.995134335166736e-05, 'epoch': 0.22}) +[2026-10-03 13:36:59,791][starVLA.training.train_starvla][INFO] - Step 2968, Loss: {'train/loss': 0.0009817725513130426, 'train/grad_norm_pre_clip': 0.8008115291595459, 'timing/data': 0.00020237000717315823, 'timing/model': 0.3928182569943601, 'learning_rate/qwen_vl_interface': 3.992197367193083e-06, 'learning_rate/action_model': 3.992197367193082e-05, 'epoch': 0.22}) +[2026-10-03 13:37:00,494][starVLA.training.train_starvla][INFO] - Step 2969, Loss: {'train/loss': 0.20960205793380737, 'train/grad_norm_pre_clip': 47.29664611816406, 'timing/data': 0.00022604998957831413, 'timing/model': 0.3926308960071765, 'learning_rate/qwen_vl_interface': 3.98926091625458e-06, 'learning_rate/action_model': 3.9892609162545795e-05, 'epoch': 0.22}) +[2026-10-03 13:37:01,199][starVLA.training.train_starvla][INFO] - Step 2970, Loss: {'train/loss': 0.15463221073150635, 'train/grad_norm_pre_clip': 24.799007415771484, 'timing/data': 0.00023450001026503742, 'timing/model': 0.39378946599026676, 'learning_rate/qwen_vl_interface': 3.986324983558294e-06, 'learning_rate/action_model': 3.9863249835582934e-05, 'epoch': 0.22}) +[2026-10-03 13:37:01,903][starVLA.training.train_starvla][INFO] - Step 2971, Loss: {'train/loss': 1.1831272786366753e-05, 'train/grad_norm_pre_clip': 0.0034357071854174137, 'timing/data': 0.0002582199958851561, 'timing/model': 0.39282740600174293, 'learning_rate/qwen_vl_interface': 3.983389570311073e-06, 'learning_rate/action_model': 3.9833895703110724e-05, 'epoch': 0.22}) +[2026-10-03 13:37:02,797][starVLA.training.train_starvla][INFO] - Step 2972, Loss: {'train/loss': 0.00017573365767020732, 'train/grad_norm_pre_clip': 0.14296089112758636, 'timing/data': 0.0003524999920045957, 'timing/model': 0.39262532599968836, 'learning_rate/qwen_vl_interface': 3.9804546777195546e-06, 'learning_rate/action_model': 3.980454677719554e-05, 'epoch': 0.22}) +[2026-10-03 13:37:03,500][starVLA.training.train_starvla][INFO] - Step 2973, Loss: {'train/loss': 0.3939208686351776, 'train/grad_norm_pre_clip': 130.5708465576172, 'timing/data': 0.0002538700064178556, 'timing/model': 0.39244990701263305, 'learning_rate/qwen_vl_interface': 3.977520306990164e-06, 'learning_rate/action_model': 3.977520306990163e-05, 'epoch': 0.22}) +[2026-10-03 13:37:04,205][starVLA.training.train_starvla][INFO] - Step 2974, Loss: {'train/loss': 5.84120334679028e-06, 'train/grad_norm_pre_clip': 0.0014443506952375174, 'timing/data': 0.0003911700041498989, 'timing/model': 0.393462616004399, 'learning_rate/qwen_vl_interface': 3.974586459329109e-06, 'learning_rate/action_model': 3.9745864593291085e-05, 'epoch': 0.22}) +[2026-10-03 13:37:04,909][starVLA.training.train_starvla][INFO] - Step 2975, Loss: {'train/loss': 8.046163566177711e-05, 'train/grad_norm_pre_clip': 0.03608591854572296, 'timing/data': 0.0002687300002435222, 'timing/model': 0.3930265560047701, 'learning_rate/qwen_vl_interface': 3.9716531359423855e-06, 'learning_rate/action_model': 3.971653135942385e-05, 'epoch': 0.22}) +[2026-10-03 13:37:05,612][starVLA.training.train_starvla][INFO] - Step 2976, Loss: {'train/loss': 1.2539127055788413e-05, 'train/grad_norm_pre_clip': 0.002545770490542054, 'timing/data': 0.0001996200007852167, 'timing/model': 0.3932382159982808, 'learning_rate/qwen_vl_interface': 3.968720338035769e-06, 'learning_rate/action_model': 3.968720338035769e-05, 'epoch': 0.22}) +[2026-10-03 13:37:06,506][starVLA.training.train_starvla][INFO] - Step 2977, Loss: {'train/loss': 2.451842010486871e-05, 'train/grad_norm_pre_clip': 0.008862044662237167, 'timing/data': 0.00022000999888405204, 'timing/model': 0.39326505600183737, 'learning_rate/qwen_vl_interface': 3.965788066814824e-06, 'learning_rate/action_model': 3.965788066814823e-05, 'epoch': 0.22}) +[2026-10-03 13:37:07,212][starVLA.training.train_starvla][INFO] - Step 2978, Loss: {'train/loss': 3.2018404453992844e-05, 'train/grad_norm_pre_clip': 0.018605953082442284, 'timing/data': 0.00033939999411813915, 'timing/model': 0.3938288959907368, 'learning_rate/qwen_vl_interface': 3.962856323484895e-06, 'learning_rate/action_model': 3.9628563234848945e-05, 'epoch': 0.22}) +[2026-10-03 13:37:07,915][starVLA.training.train_starvla][INFO] - Step 2979, Loss: {'train/loss': 0.0004033817385789007, 'train/grad_norm_pre_clip': 0.24350643157958984, 'timing/data': 0.0002846799907274544, 'timing/model': 0.39261472699581645, 'learning_rate/qwen_vl_interface': 3.959925109251112e-06, 'learning_rate/action_model': 3.959925109251112e-05, 'epoch': 0.22}) +[2026-10-03 13:37:08,617][starVLA.training.train_starvla][INFO] - Step 2980, Loss: {'train/loss': 0.00036925883614458144, 'train/grad_norm_pre_clip': 0.1390550434589386, 'timing/data': 0.0003072200051974505, 'timing/model': 0.3915559369925177, 'learning_rate/qwen_vl_interface': 3.956994425318387e-06, 'learning_rate/action_model': 3.956994425318387e-05, 'epoch': 0.22}) +[2026-10-03 13:37:09,321][starVLA.training.train_starvla][INFO] - Step 2981, Loss: {'train/loss': 0.0036121439188718796, 'train/grad_norm_pre_clip': 7.289494514465332, 'timing/data': 0.00020772000425495207, 'timing/model': 0.39303897599165794, 'learning_rate/qwen_vl_interface': 3.9540642728914115e-06, 'learning_rate/action_model': 3.9540642728914115e-05, 'epoch': 0.22}) +[2026-10-03 13:37:10,221][starVLA.training.train_starvla][INFO] - Step 2982, Loss: {'train/loss': 3.480844679870643e-05, 'train/grad_norm_pre_clip': 0.023720910772681236, 'timing/data': 0.0002514400111977011, 'timing/model': 0.5867357839888427, 'learning_rate/qwen_vl_interface': 3.951134653174664e-06, 'learning_rate/action_model': 3.951134653174663e-05, 'epoch': 0.22}) +[2026-10-03 13:37:10,927][starVLA.training.train_starvla][INFO] - Step 2983, Loss: {'train/loss': 1.2106143003620673e-05, 'train/grad_norm_pre_clip': 0.008740170858800411, 'timing/data': 0.00031866000790614635, 'timing/model': 0.3925919059984153, 'learning_rate/qwen_vl_interface': 3.948205567372397e-06, 'learning_rate/action_model': 3.948205567372397e-05, 'epoch': 0.22}) +[2026-10-03 13:37:11,630][starVLA.training.train_starvla][INFO] - Step 2984, Loss: {'train/loss': 0.002437136834487319, 'train/grad_norm_pre_clip': 3.6361424922943115, 'timing/data': 0.0002963500010082498, 'timing/model': 0.39263923600083217, 'learning_rate/qwen_vl_interface': 3.945277016688647e-06, 'learning_rate/action_model': 3.945277016688647e-05, 'epoch': 0.22}) +[2026-10-03 13:37:12,332][starVLA.training.train_starvla][INFO] - Step 2985, Loss: {'train/loss': 0.023242680355906487, 'train/grad_norm_pre_clip': 21.54215431213379, 'timing/data': 0.00027146999491378665, 'timing/model': 0.3923015070031397, 'learning_rate/qwen_vl_interface': 3.9423490023272326e-06, 'learning_rate/action_model': 3.9423490023272324e-05, 'epoch': 0.22}) +[2026-10-03 13:37:13,037][starVLA.training.train_starvla][INFO] - Step 2986, Loss: {'train/loss': 0.0026990193873643875, 'train/grad_norm_pre_clip': 3.506316900253296, 'timing/data': 0.0002675599971553311, 'timing/model': 0.39274699600355234, 'learning_rate/qwen_vl_interface': 3.939421525491748e-06, 'learning_rate/action_model': 3.9394215254917483e-05, 'epoch': 0.22}) +[2026-10-03 13:37:13,928][starVLA.training.train_starvla][INFO] - Step 2987, Loss: {'train/loss': 4.858147804043256e-05, 'train/grad_norm_pre_clip': 0.029407039284706116, 'timing/data': 0.00020021999080199748, 'timing/model': 0.5817255750007462, 'learning_rate/qwen_vl_interface': 3.9364945873855705e-06, 'learning_rate/action_model': 3.93649458738557e-05, 'epoch': 0.22}) +[2026-10-03 13:37:14,631][starVLA.training.train_starvla][INFO] - Step 2988, Loss: {'train/loss': 0.1481465846300125, 'train/grad_norm_pre_clip': 94.93719482421875, 'timing/data': 0.00029965999419800937, 'timing/model': 0.39235726700280793, 'learning_rate/qwen_vl_interface': 3.9335681892118536e-06, 'learning_rate/action_model': 3.933568189211854e-05, 'epoch': 0.22}) +[2026-10-03 13:37:15,333][starVLA.training.train_starvla][INFO] - Step 2989, Loss: {'train/loss': 0.0053016385063529015, 'train/grad_norm_pre_clip': 7.402312755584717, 'timing/data': 0.00022358000569511205, 'timing/model': 0.39148864599701483, 'learning_rate/qwen_vl_interface': 3.930642332173527e-06, 'learning_rate/action_model': 3.930642332173527e-05, 'epoch': 0.22}) +[2026-10-03 13:37:16,039][starVLA.training.train_starvla][INFO] - Step 2990, Loss: {'train/loss': 0.44158685207366943, 'train/grad_norm_pre_clip': 108.89787292480469, 'timing/data': 0.0002919699909398332, 'timing/model': 0.3941758260043571, 'learning_rate/qwen_vl_interface': 3.927717017473301e-06, 'learning_rate/action_model': 3.9277170174733006e-05, 'epoch': 0.22}) +[2026-10-03 13:37:16,743][starVLA.training.train_starvla][INFO] - Step 2991, Loss: {'train/loss': 0.00014855309564154595, 'train/grad_norm_pre_clip': 0.15816347301006317, 'timing/data': 0.00020294000569265336, 'timing/model': 0.3929136459919391, 'learning_rate/qwen_vl_interface': 3.9247922463136615e-06, 'learning_rate/action_model': 3.924792246313662e-05, 'epoch': 0.22}) +[2026-10-03 13:37:17,446][starVLA.training.train_starvla][INFO] - Step 2992, Loss: {'train/loss': 0.003041812451556325, 'train/grad_norm_pre_clip': 2.6250720024108887, 'timing/data': 0.00022301000717561692, 'timing/model': 0.3926252560049761, 'learning_rate/qwen_vl_interface': 3.921868019896873e-06, 'learning_rate/action_model': 3.921868019896873e-05, 'epoch': 0.22}) +[2026-10-03 13:37:18,345][starVLA.training.train_starvla][INFO] - Step 2993, Loss: {'train/loss': 0.2687297463417053, 'train/grad_norm_pre_clip': 425.82757568359375, 'timing/data': 0.00021908999769948423, 'timing/model': 0.393231845999253, 'learning_rate/qwen_vl_interface': 3.918944339424974e-06, 'learning_rate/action_model': 3.9189443394249736e-05, 'epoch': 0.22}) +[2026-10-03 13:37:19,049][starVLA.training.train_starvla][INFO] - Step 2994, Loss: {'train/loss': 8.731250272830948e-05, 'train/grad_norm_pre_clip': 0.07495617121458054, 'timing/data': 0.0002920400002039969, 'timing/model': 0.3925214360060636, 'learning_rate/qwen_vl_interface': 3.9160212060997774e-06, 'learning_rate/action_model': 3.9160212060997774e-05, 'epoch': 0.22}) +[2026-10-03 13:37:19,752][starVLA.training.train_starvla][INFO] - Step 2995, Loss: {'train/loss': 4.470572457648814e-05, 'train/grad_norm_pre_clip': 0.008932927623391151, 'timing/data': 0.00020616999245248735, 'timing/model': 0.3932352759875357, 'learning_rate/qwen_vl_interface': 3.913098621122875e-06, 'learning_rate/action_model': 3.913098621122875e-05, 'epoch': 0.22}) +[2026-10-03 13:37:20,456][starVLA.training.train_starvla][INFO] - Step 2996, Loss: {'train/loss': 0.31986817717552185, 'train/grad_norm_pre_clip': 34.71487045288086, 'timing/data': 0.00021680000645574182, 'timing/model': 0.3932137059891829, 'learning_rate/qwen_vl_interface': 3.91017658569563e-06, 'learning_rate/action_model': 3.91017658569563e-05, 'epoch': 0.22}) +[2026-10-03 13:37:21,159][starVLA.training.train_starvla][INFO] - Step 2997, Loss: {'train/loss': 0.007993962615728378, 'train/grad_norm_pre_clip': 24.831655502319336, 'timing/data': 0.00020343999494798481, 'timing/model': 0.3925679170060903, 'learning_rate/qwen_vl_interface': 3.907255101019181e-06, 'learning_rate/action_model': 3.907255101019181e-05, 'epoch': 0.22}) +[2026-10-03 13:37:22,056][starVLA.training.train_starvla][INFO] - Step 2998, Loss: {'train/loss': 0.45969313383102417, 'train/grad_norm_pre_clip': 60.699466705322266, 'timing/data': 0.00032292999094352126, 'timing/model': 0.5847392050054623, 'learning_rate/qwen_vl_interface': 3.904334168294442e-06, 'learning_rate/action_model': 3.9043341682944415e-05, 'epoch': 0.22}) +[2026-10-03 13:37:22,759][starVLA.training.train_starvla][INFO] - Step 2999, Loss: {'train/loss': 4.5835666242055595e-05, 'train/grad_norm_pre_clip': 0.01372725609689951, 'timing/data': 0.000281510001514107, 'timing/model': 0.3929420069907792, 'learning_rate/qwen_vl_interface': 3.901413788722093e-06, 'learning_rate/action_model': 3.901413788722093e-05, 'epoch': 0.22}) +[2026-10-03 13:37:23,463][starVLA.training.train_starvla][INFO] - Step 3000, Loss: {'train/loss': 9.73632704699412e-05, 'train/grad_norm_pre_clip': 0.03843601793050766, 'timing/data': 0.0002655499993124977, 'timing/model': 0.3928272359917173, 'learning_rate/qwen_vl_interface': 3.8984939635025966e-06, 'learning_rate/action_model': 3.898493963502596e-05, 'epoch': 0.22}) +[2026-10-03 13:37:24,167][starVLA.training.train_starvla][INFO] - Step 3001, Loss: {'train/loss': 9.232333104591817e-05, 'train/grad_norm_pre_clip': 0.03992446884512901, 'timing/data': 0.0003644899989012629, 'timing/model': 0.39318451601138804, 'learning_rate/qwen_vl_interface': 3.895574693836181e-06, 'learning_rate/action_model': 3.8955746938361804e-05, 'epoch': 0.22}) +[2026-10-03 13:37:24,872][starVLA.training.train_starvla][INFO] - Step 3002, Loss: {'train/loss': 0.0023537734523415565, 'train/grad_norm_pre_clip': 1.9209061861038208, 'timing/data': 0.0002698799944482744, 'timing/model': 0.39372488600201905, 'learning_rate/qwen_vl_interface': 3.892655980922847e-06, 'learning_rate/action_model': 3.8926559809228466e-05, 'epoch': 0.22}) +[2026-10-03 13:37:25,576][starVLA.training.train_starvla][INFO] - Step 3003, Loss: {'train/loss': 0.011705871671438217, 'train/grad_norm_pre_clip': 90.86689758300781, 'timing/data': 0.00029832001018803567, 'timing/model': 0.3929885760007892, 'learning_rate/qwen_vl_interface': 3.889737825962368e-06, 'learning_rate/action_model': 3.889737825962368e-05, 'epoch': 0.22}) +[2026-10-03 13:37:26,466][starVLA.training.train_starvla][INFO] - Step 3004, Loss: {'train/loss': 2.6909536245511845e-05, 'train/grad_norm_pre_clip': 0.009040835313498974, 'timing/data': 0.00031601000227965415, 'timing/model': 0.3929530960012926, 'learning_rate/qwen_vl_interface': 3.886820230154288e-06, 'learning_rate/action_model': 3.886820230154287e-05, 'epoch': 0.22}) +[2026-10-03 13:37:27,168][starVLA.training.train_starvla][INFO] - Step 3005, Loss: {'train/loss': 0.00012840877752751112, 'train/grad_norm_pre_clip': 0.09005393087863922, 'timing/data': 0.00022513000294566154, 'timing/model': 0.3926103569974657, 'learning_rate/qwen_vl_interface': 3.8839031946979184e-06, 'learning_rate/action_model': 3.883903194697918e-05, 'epoch': 0.22}) +[2026-10-03 13:37:27,875][starVLA.training.train_starvla][INFO] - Step 3006, Loss: {'train/loss': 4.7275323595386e-05, 'train/grad_norm_pre_clip': 0.08438177406787872, 'timing/data': 0.00033745000837370753, 'timing/model': 0.39405464701121673, 'learning_rate/qwen_vl_interface': 3.880986720792344e-06, 'learning_rate/action_model': 3.880986720792344e-05, 'epoch': 0.22}) +[2026-10-03 13:37:28,578][starVLA.training.train_starvla][INFO] - Step 3007, Loss: {'train/loss': 0.28050151467323303, 'train/grad_norm_pre_clip': 30.93916130065918, 'timing/data': 0.0002912399941124022, 'timing/model': 0.39251864700054284, 'learning_rate/qwen_vl_interface': 3.878070809636416e-06, 'learning_rate/action_model': 3.878070809636416e-05, 'epoch': 0.22}) +[2026-10-03 13:37:29,281][starVLA.training.train_starvla][INFO] - Step 3008, Loss: {'train/loss': 6.18360863882117e-05, 'train/grad_norm_pre_clip': 0.013934623450040817, 'timing/data': 0.00024777000362519175, 'timing/model': 0.3928361459984444, 'learning_rate/qwen_vl_interface': 3.875155462428758e-06, 'learning_rate/action_model': 3.875155462428758e-05, 'epoch': 0.22}) +[2026-10-03 13:37:30,177][starVLA.training.train_starvla][INFO] - Step 3009, Loss: {'train/loss': 2.2028898456483148e-05, 'train/grad_norm_pre_clip': 0.0036476608365774155, 'timing/data': 0.0003202399966539815, 'timing/model': 0.39297976599482354, 'learning_rate/qwen_vl_interface': 3.872240680367754e-06, 'learning_rate/action_model': 3.872240680367754e-05, 'epoch': 0.22}) +[2026-10-03 13:37:30,882][starVLA.training.train_starvla][INFO] - Step 3010, Loss: {'train/loss': 0.6018322110176086, 'train/grad_norm_pre_clip': 49.82899856567383, 'timing/data': 0.0003258700016885996, 'timing/model': 0.39365960599388927, 'learning_rate/qwen_vl_interface': 3.869326464651566e-06, 'learning_rate/action_model': 3.869326464651566e-05, 'epoch': 0.22}) +[2026-10-03 13:37:31,586][starVLA.training.train_starvla][INFO] - Step 3011, Loss: {'train/loss': 0.00011679393355734646, 'train/grad_norm_pre_clip': 0.048343539237976074, 'timing/data': 0.00020227998902555555, 'timing/model': 0.39288089600449894, 'learning_rate/qwen_vl_interface': 3.866412816478114e-06, 'learning_rate/action_model': 3.8664128164781144e-05, 'epoch': 0.22}) +[2026-10-03 13:37:32,290][starVLA.training.train_starvla][INFO] - Step 3012, Loss: {'train/loss': 9.87580424407497e-05, 'train/grad_norm_pre_clip': 0.03799242526292801, 'timing/data': 0.0003431799996178597, 'timing/model': 0.39303772599669173, 'learning_rate/qwen_vl_interface': 3.8634997370450915e-06, 'learning_rate/action_model': 3.8634997370450914e-05, 'epoch': 0.22}) +[2026-10-03 13:37:32,994][starVLA.training.train_starvla][INFO] - Step 3013, Loss: {'train/loss': 0.23258070647716522, 'train/grad_norm_pre_clip': 94.51273345947266, 'timing/data': 0.00032358999305870384, 'timing/model': 0.3937159170018276, 'learning_rate/qwen_vl_interface': 3.860587227549956e-06, 'learning_rate/action_model': 3.860587227549955e-05, 'epoch': 0.22}) +[2026-10-03 13:37:33,889][starVLA.training.train_starvla][INFO] - Step 3014, Loss: {'train/loss': 6.307850708253682e-05, 'train/grad_norm_pre_clip': 0.0158827006816864, 'timing/data': 0.0003270900051575154, 'timing/model': 0.583219794993056, 'learning_rate/qwen_vl_interface': 3.857675289189926e-06, 'learning_rate/action_model': 3.857675289189926e-05, 'epoch': 0.22}) +[2026-10-03 13:37:34,593][starVLA.training.train_starvla][INFO] - Step 3015, Loss: {'train/loss': 2.345948996662628e-05, 'train/grad_norm_pre_clip': 0.005026009399443865, 'timing/data': 0.0002603600005386397, 'timing/model': 0.39308008599618915, 'learning_rate/qwen_vl_interface': 3.854763923161993e-06, 'learning_rate/action_model': 3.854763923161992e-05, 'epoch': 0.22}) +[2026-10-03 13:37:35,295][starVLA.training.train_starvla][INFO] - Step 3016, Loss: {'train/loss': 9.126784789259546e-06, 'train/grad_norm_pre_clip': 0.0016759986756369472, 'timing/data': 0.0001944699906744063, 'timing/model': 0.3925277460075449, 'learning_rate/qwen_vl_interface': 3.851853130662908e-06, 'learning_rate/action_model': 3.8518531306629077e-05, 'epoch': 0.22}) +[2026-10-03 13:37:35,996][starVLA.training.train_starvla][INFO] - Step 3017, Loss: {'train/loss': 0.0024234014563262463, 'train/grad_norm_pre_clip': 5.74993896484375, 'timing/data': 0.00030720001086592674, 'timing/model': 0.39220357600424904, 'learning_rate/qwen_vl_interface': 3.8489429128891884e-06, 'learning_rate/action_model': 3.848942912889188e-05, 'epoch': 0.22}) +[2026-10-03 13:37:36,702][starVLA.training.train_starvla][INFO] - Step 3018, Loss: {'train/loss': 1.1972655556746759e-05, 'train/grad_norm_pre_clip': 0.0023312976118177176, 'timing/data': 0.00033193999843206257, 'timing/model': 0.39375633701274637, 'learning_rate/qwen_vl_interface': 3.846033271037114e-06, 'learning_rate/action_model': 3.8460332710371136e-05, 'epoch': 0.22}) +[2026-10-03 13:37:37,405][starVLA.training.train_starvla][INFO] - Step 3019, Loss: {'train/loss': 1.2442142178770155e-05, 'train/grad_norm_pre_clip': 0.004517918452620506, 'timing/data': 0.00028350998763926327, 'timing/model': 0.392657476986642, 'learning_rate/qwen_vl_interface': 3.843124206302732e-06, 'learning_rate/action_model': 3.8431242063027316e-05, 'epoch': 0.22}) +[2026-10-03 13:37:38,296][starVLA.training.train_starvla][INFO] - Step 3020, Loss: {'train/loss': 1.208456524182111e-05, 'train/grad_norm_pre_clip': 0.0029615454841405153, 'timing/data': 0.00025892999838106334, 'timing/model': 0.3928037160076201, 'learning_rate/qwen_vl_interface': 3.840215719881844e-06, 'learning_rate/action_model': 3.8402157198818445e-05, 'epoch': 0.22}) +[2026-10-03 13:37:38,998][starVLA.training.train_starvla][INFO] - Step 3021, Loss: {'train/loss': 7.046705286484212e-05, 'train/grad_norm_pre_clip': 0.03255946934223175, 'timing/data': 0.0002439800009597093, 'timing/model': 0.3923671860102331, 'learning_rate/qwen_vl_interface': 3.837307812970022e-06, 'learning_rate/action_model': 3.837307812970022e-05, 'epoch': 0.22}) +[2026-10-03 13:37:39,703][starVLA.training.train_starvla][INFO] - Step 3022, Loss: {'train/loss': 2.1427087631309405e-05, 'train/grad_norm_pre_clip': 0.0056265564635396, 'timing/data': 0.000345049993484281, 'timing/model': 0.39346838599885814, 'learning_rate/qwen_vl_interface': 3.834400486762596e-06, 'learning_rate/action_model': 3.834400486762596e-05, 'epoch': 0.22}) +[2026-10-03 13:37:40,406][starVLA.training.train_starvla][INFO] - Step 3023, Loss: {'train/loss': 0.5357306599617004, 'train/grad_norm_pre_clip': 14.241849899291992, 'timing/data': 0.0002833299949998036, 'timing/model': 0.39273541599686723, 'learning_rate/qwen_vl_interface': 3.8314937424546585e-06, 'learning_rate/action_model': 3.8314937424546584e-05, 'epoch': 0.22}) +[2026-10-03 13:37:41,109][starVLA.training.train_starvla][INFO] - Step 3024, Loss: {'train/loss': 0.39418840408325195, 'train/grad_norm_pre_clip': 174.0801239013672, 'timing/data': 0.00029615999665111303, 'timing/model': 0.3923382470093202, 'learning_rate/qwen_vl_interface': 3.828587581241064e-06, 'learning_rate/action_model': 3.8285875812410635e-05, 'epoch': 0.22}) +[2026-10-03 13:37:42,004][starVLA.training.train_starvla][INFO] - Step 3025, Loss: {'train/loss': 0.00011397944763302803, 'train/grad_norm_pre_clip': 0.019780905917286873, 'timing/data': 0.00034605999826453626, 'timing/model': 0.3929914159962209, 'learning_rate/qwen_vl_interface': 3.825682004316419e-06, 'learning_rate/action_model': 3.825682004316419e-05, 'epoch': 0.22}) +[2026-10-03 13:37:42,709][starVLA.training.train_starvla][INFO] - Step 3026, Loss: {'train/loss': 6.187443796079606e-05, 'train/grad_norm_pre_clip': 0.017904354259371758, 'timing/data': 0.0003285290003987029, 'timing/model': 0.3929400670022005, 'learning_rate/qwen_vl_interface': 3.822777012875104e-06, 'learning_rate/action_model': 3.822777012875103e-05, 'epoch': 0.22}) +[2026-10-03 13:37:43,413][starVLA.training.train_starvla][INFO] - Step 3027, Loss: {'train/loss': 3.235238546039909e-05, 'train/grad_norm_pre_clip': 0.017234938219189644, 'timing/data': 0.0002963699953397736, 'timing/model': 0.3936463659920264, 'learning_rate/qwen_vl_interface': 3.819872608111248e-06, 'learning_rate/action_model': 3.819872608111248e-05, 'epoch': 0.22}) +[2026-10-03 13:37:44,116][starVLA.training.train_starvla][INFO] - Step 3028, Loss: {'train/loss': 4.725215694634244e-05, 'train/grad_norm_pre_clip': 0.02431260235607624, 'timing/data': 0.0002993399975821376, 'timing/model': 0.3928051660041092, 'learning_rate/qwen_vl_interface': 3.816968791218741e-06, 'learning_rate/action_model': 3.816968791218741e-05, 'epoch': 0.22}) +[2026-10-03 13:37:44,819][starVLA.training.train_starvla][INFO] - Step 3029, Loss: {'train/loss': 1.9028244423680007e-05, 'train/grad_norm_pre_clip': 0.0031173659954220057, 'timing/data': 0.0002760600036708638, 'timing/model': 0.39252199699694756, 'learning_rate/qwen_vl_interface': 3.8140655633912337e-06, 'learning_rate/action_model': 3.814065563391234e-05, 'epoch': 0.22}) +[2026-10-03 13:37:45,716][starVLA.training.train_starvla][INFO] - Step 3030, Loss: {'train/loss': 0.0008695434080436826, 'train/grad_norm_pre_clip': 0.5040735006332397, 'timing/data': 0.00031879000016488135, 'timing/model': 0.3935350960091455, 'learning_rate/qwen_vl_interface': 3.811162925822133e-06, 'learning_rate/action_model': 3.811162925822132e-05, 'epoch': 0.22}) +[2026-10-03 13:37:46,419][starVLA.training.train_starvla][INFO] - Step 3031, Loss: {'train/loss': 1.99595742742531e-05, 'train/grad_norm_pre_clip': 0.0035419827327132225, 'timing/data': 0.000228899996727705, 'timing/model': 0.3924669570114929, 'learning_rate/qwen_vl_interface': 3.8082608797046045e-06, 'learning_rate/action_model': 3.8082608797046045e-05, 'epoch': 0.22}) +[2026-10-03 13:37:47,121][starVLA.training.train_starvla][INFO] - Step 3032, Loss: {'train/loss': 0.06239819899201393, 'train/grad_norm_pre_clip': 319.140625, 'timing/data': 0.0002776199980871752, 'timing/model': 0.39216768598998897, 'learning_rate/qwen_vl_interface': 3.8053594262315687e-06, 'learning_rate/action_model': 3.805359426231568e-05, 'epoch': 0.22}) +[2026-10-03 13:37:47,823][starVLA.training.train_starvla][INFO] - Step 3033, Loss: {'train/loss': 0.00013047085667494684, 'train/grad_norm_pre_clip': 0.023167695850133896, 'timing/data': 0.0002005099959205836, 'timing/model': 0.392295835990808, 'learning_rate/qwen_vl_interface': 3.802458566595704e-06, 'learning_rate/action_model': 3.802458566595704e-05, 'epoch': 0.22}) +[2026-10-03 13:37:48,528][starVLA.training.train_starvla][INFO] - Step 3034, Loss: {'train/loss': 1.4453567928285338e-05, 'train/grad_norm_pre_clip': 0.005566779058426619, 'timing/data': 0.00026645899924915284, 'timing/model': 0.3935247470071772, 'learning_rate/qwen_vl_interface': 3.7995583019894444e-06, 'learning_rate/action_model': 3.7995583019894445e-05, 'epoch': 0.22}) +[2026-10-03 13:37:49,420][starVLA.training.train_starvla][INFO] - Step 3035, Loss: {'train/loss': 0.00013046353706158698, 'train/grad_norm_pre_clip': 0.026059255003929138, 'timing/data': 0.00020160998974461108, 'timing/model': 0.5839024849992711, 'learning_rate/qwen_vl_interface': 3.7966586336049782e-06, 'learning_rate/action_model': 3.796658633604978e-05, 'epoch': 0.22}) +[2026-10-03 13:37:50,123][starVLA.training.train_starvla][INFO] - Step 3036, Loss: {'train/loss': 1.0736005606304388e-05, 'train/grad_norm_pre_clip': 0.0034060548059642315, 'timing/data': 0.00020058000518474728, 'timing/model': 0.39315025600080844, 'learning_rate/qwen_vl_interface': 3.7937595626342516e-06, 'learning_rate/action_model': 3.793759562634252e-05, 'epoch': 0.22}) +[2026-10-03 13:37:50,825][starVLA.training.train_starvla][INFO] - Step 3037, Loss: {'train/loss': 0.35603463649749756, 'train/grad_norm_pre_clip': 79.59754943847656, 'timing/data': 0.00025810999795794487, 'timing/model': 0.3914127059979364, 'learning_rate/qwen_vl_interface': 3.79086109026896e-06, 'learning_rate/action_model': 3.7908610902689604e-05, 'epoch': 0.22}) +[2026-10-03 13:37:51,531][starVLA.training.train_starvla][INFO] - Step 3038, Loss: {'train/loss': 0.0003339706745464355, 'train/grad_norm_pre_clip': 0.49032965302467346, 'timing/data': 0.00034043999039568007, 'timing/model': 0.39339247600582894, 'learning_rate/qwen_vl_interface': 3.7879632177005584e-06, 'learning_rate/action_model': 3.787963217700558e-05, 'epoch': 0.22}) +[2026-10-03 13:37:52,235][starVLA.training.train_starvla][INFO] - Step 3039, Loss: {'train/loss': 0.0003360370174050331, 'train/grad_norm_pre_clip': 0.14851053059101105, 'timing/data': 0.00028449000092223287, 'timing/model': 0.39267056599783245, 'learning_rate/qwen_vl_interface': 3.7850659461202523e-06, 'learning_rate/action_model': 3.785065946120252e-05, 'epoch': 0.22}) +[2026-10-03 13:37:53,127][starVLA.training.train_starvla][INFO] - Step 3040, Loss: {'train/loss': 2.548575503169559e-05, 'train/grad_norm_pre_clip': 0.015016268938779831, 'timing/data': 0.0002459899988025427, 'timing/model': 0.5816863660002127, 'learning_rate/qwen_vl_interface': 3.782169276719e-06, 'learning_rate/action_model': 3.782169276719e-05, 'epoch': 0.22}) +[2026-10-03 13:37:53,830][starVLA.training.train_starvla][INFO] - Step 3041, Loss: {'train/loss': 6.877764826640487e-05, 'train/grad_norm_pre_clip': 0.015896758064627647, 'timing/data': 0.00030605000210925937, 'timing/model': 0.39282358600758016, 'learning_rate/qwen_vl_interface': 3.779273210687514e-06, 'learning_rate/action_model': 3.779273210687514e-05, 'epoch': 0.22}) +[2026-10-03 13:37:54,535][starVLA.training.train_starvla][INFO] - Step 3042, Loss: {'train/loss': 0.00017656898126006126, 'train/grad_norm_pre_clip': 0.16045626997947693, 'timing/data': 0.00022435000573750585, 'timing/model': 0.393383205999271, 'learning_rate/qwen_vl_interface': 3.776377749216257e-06, 'learning_rate/action_model': 3.776377749216257e-05, 'epoch': 0.22}) +[2026-10-03 13:37:55,238][starVLA.training.train_starvla][INFO] - Step 3043, Loss: {'train/loss': 0.00011566698958631605, 'train/grad_norm_pre_clip': 0.07269454002380371, 'timing/data': 0.000258829997619614, 'timing/model': 0.39290941700164694, 'learning_rate/qwen_vl_interface': 3.773482893495442e-06, 'learning_rate/action_model': 3.773482893495442e-05, 'epoch': 0.22}) +[2026-10-03 13:37:55,942][starVLA.training.train_starvla][INFO] - Step 3044, Loss: {'train/loss': 1.7805634342948906e-05, 'train/grad_norm_pre_clip': 0.005414173938333988, 'timing/data': 0.00020959001267328858, 'timing/model': 0.39296965599351097, 'learning_rate/qwen_vl_interface': 3.7705886447150375e-06, 'learning_rate/action_model': 3.7705886447150376e-05, 'epoch': 0.22}) +[2026-10-03 13:37:56,645][starVLA.training.train_starvla][INFO] - Step 3045, Loss: {'train/loss': 0.0021107937209308147, 'train/grad_norm_pre_clip': 2.1013295650482178, 'timing/data': 0.00021229000412859023, 'timing/model': 0.39302818599389866, 'learning_rate/qwen_vl_interface': 3.767695004064759e-06, 'learning_rate/action_model': 3.767695004064759e-05, 'epoch': 0.22}) +[2026-10-03 13:37:57,544][starVLA.training.train_starvla][INFO] - Step 3046, Loss: {'train/loss': 0.00023630858049727976, 'train/grad_norm_pre_clip': 0.08629868179559708, 'timing/data': 0.00023130000045057386, 'timing/model': 0.39367947600840125, 'learning_rate/qwen_vl_interface': 3.7648019727340725e-06, 'learning_rate/action_model': 3.764801972734072e-05, 'epoch': 0.22}) +[2026-10-03 13:37:58,248][starVLA.training.train_starvla][INFO] - Step 3047, Loss: {'train/loss': 0.659102737903595, 'train/grad_norm_pre_clip': 44.74073791503906, 'timing/data': 0.00021319999359548092, 'timing/model': 0.3931465159985237, 'learning_rate/qwen_vl_interface': 3.761909551912196e-06, 'learning_rate/action_model': 3.761909551912196e-05, 'epoch': 0.22}) +[2026-10-03 13:37:58,951][starVLA.training.train_starvla][INFO] - Step 3048, Loss: {'train/loss': 0.00028442085022106767, 'train/grad_norm_pre_clip': 0.21587052941322327, 'timing/data': 0.0001991800090763718, 'timing/model': 0.39237184599915054, 'learning_rate/qwen_vl_interface': 3.75901774278809e-06, 'learning_rate/action_model': 3.7590177427880904e-05, 'epoch': 0.22}) +[2026-10-03 13:37:59,654][starVLA.training.train_starvla][INFO] - Step 3049, Loss: {'train/loss': 0.00014648954675067216, 'train/grad_norm_pre_clip': 0.05455894023180008, 'timing/data': 0.0002861600078176707, 'timing/model': 0.3930521659931401, 'learning_rate/qwen_vl_interface': 3.756126546550472e-06, 'learning_rate/action_model': 3.756126546550472e-05, 'epoch': 0.22}) +[2026-10-03 13:38:00,358][starVLA.training.train_starvla][INFO] - Step 3050, Loss: {'train/loss': 7.277166150743142e-05, 'train/grad_norm_pre_clip': 0.013870196416974068, 'timing/data': 0.0003359600086696446, 'timing/model': 0.39304305700352415, 'learning_rate/qwen_vl_interface': 3.753235964387802e-06, 'learning_rate/action_model': 3.753235964387802e-05, 'epoch': 0.22}) +[2026-10-03 13:38:01,249][starVLA.training.train_starvla][INFO] - Step 3051, Loss: {'train/loss': 7.558708603028208e-05, 'train/grad_norm_pre_clip': 0.023485247045755386, 'timing/data': 0.0002081400016322732, 'timing/model': 0.5814361039956566, 'learning_rate/qwen_vl_interface': 3.750345997488289e-06, 'learning_rate/action_model': 3.7503459974882886e-05, 'epoch': 0.22}) +[2026-10-03 13:38:01,951][starVLA.training.train_starvla][INFO] - Step 3052, Loss: {'train/loss': 5.086835517431609e-05, 'train/grad_norm_pre_clip': 0.013986615464091301, 'timing/data': 0.00020429000142030418, 'timing/model': 0.39220533600018825, 'learning_rate/qwen_vl_interface': 3.7474566470398903e-06, 'learning_rate/action_model': 3.74745664703989e-05, 'epoch': 0.22}) +[2026-10-03 13:38:02,652][starVLA.training.train_starvla][INFO] - Step 3053, Loss: {'train/loss': 2.5159592041745782e-05, 'train/grad_norm_pre_clip': 0.0063773589208722115, 'timing/data': 0.00020737999875564128, 'timing/model': 0.39223723601025995, 'learning_rate/qwen_vl_interface': 3.744567914230307e-06, 'learning_rate/action_model': 3.744567914230307e-05, 'epoch': 0.22}) +[2026-10-03 13:38:03,348][starVLA.training.train_starvla][INFO] - Step 3054, Loss: {'train/loss': 3.519746314850636e-05, 'train/grad_norm_pre_clip': 0.014876111410558224, 'timing/data': 0.00019998999778181314, 'timing/model': 0.38922159699723125, 'learning_rate/qwen_vl_interface': 3.7416798002469902e-06, 'learning_rate/action_model': 3.74167980024699e-05, 'epoch': 0.22}) +[2026-10-03 13:38:04,044][starVLA.training.train_starvla][INFO] - Step 3055, Loss: {'train/loss': 0.38710829615592957, 'train/grad_norm_pre_clip': 48.032501220703125, 'timing/data': 0.0003291000030003488, 'timing/model': 0.388804998001433, 'learning_rate/qwen_vl_interface': 3.7387923062771313e-06, 'learning_rate/action_model': 3.738792306277131e-05, 'epoch': 0.22}) +[2026-10-03 13:38:04,740][starVLA.training.train_starvla][INFO] - Step 3056, Loss: {'train/loss': 0.000328603433445096, 'train/grad_norm_pre_clip': 0.5743613839149475, 'timing/data': 0.0002644600026542321, 'timing/model': 0.3891160959901754, 'learning_rate/qwen_vl_interface': 3.7359054335076732e-06, 'learning_rate/action_model': 3.7359054335076734e-05, 'epoch': 0.22}) +[2026-10-03 13:38:05,626][starVLA.training.train_starvla][INFO] - Step 3057, Loss: {'train/loss': 6.839582056272775e-05, 'train/grad_norm_pre_clip': 0.01107640191912651, 'timing/data': 0.00034706899896264076, 'timing/model': 0.3884126769989962, 'learning_rate/qwen_vl_interface': 3.7330191831253006e-06, 'learning_rate/action_model': 3.7330191831253e-05, 'epoch': 0.22}) +[2026-10-03 13:38:06,322][starVLA.training.train_starvla][INFO] - Step 3058, Loss: {'train/loss': 0.00011547462054295465, 'train/grad_norm_pre_clip': 0.02030264586210251, 'timing/data': 0.00019669000175781548, 'timing/model': 0.38913478700851556, 'learning_rate/qwen_vl_interface': 3.7301335563164382e-06, 'learning_rate/action_model': 3.730133556316438e-05, 'epoch': 0.22}) +[2026-10-03 13:38:07,017][starVLA.training.train_starvla][INFO] - Step 3059, Loss: {'train/loss': 0.00018510672089178115, 'train/grad_norm_pre_clip': 0.14992991089820862, 'timing/data': 0.00026974998763762414, 'timing/model': 0.3893793659954099, 'learning_rate/qwen_vl_interface': 3.727248554267261e-06, 'learning_rate/action_model': 3.727248554267261e-05, 'epoch': 0.22}) +[2026-10-03 13:38:07,712][starVLA.training.train_starvla][INFO] - Step 3060, Loss: {'train/loss': 3.3080375487770652e-06, 'train/grad_norm_pre_clip': 0.0010295496322214603, 'timing/data': 0.00031954998848959804, 'timing/model': 0.3884542470041197, 'learning_rate/qwen_vl_interface': 3.7243641781636838e-06, 'learning_rate/action_model': 3.7243641781636836e-05, 'epoch': 0.22}) +[2026-10-03 13:38:08,407][starVLA.training.train_starvla][INFO] - Step 3061, Loss: {'train/loss': 1.613710082892794e-05, 'train/grad_norm_pre_clip': 0.004276888445019722, 'timing/data': 0.00027363999106455594, 'timing/model': 0.38898221698764246, 'learning_rate/qwen_vl_interface': 3.721480429191364e-06, 'learning_rate/action_model': 3.721480429191364e-05, 'epoch': 0.22}) +[2026-10-03 13:38:09,295][starVLA.training.train_starvla][INFO] - Step 3062, Loss: {'train/loss': 9.954218694474548e-05, 'train/grad_norm_pre_clip': 0.02134704403579235, 'timing/data': 0.0003101699985563755, 'timing/model': 0.3883816869929433, 'learning_rate/qwen_vl_interface': 3.7185973085357045e-06, 'learning_rate/action_model': 3.718597308535705e-05, 'epoch': 0.22}) +[2026-10-03 13:38:09,991][starVLA.training.train_starvla][INFO] - Step 3063, Loss: {'train/loss': 0.29804956912994385, 'train/grad_norm_pre_clip': 65.17343139648438, 'timing/data': 0.00020642999152187258, 'timing/model': 0.38914599599957, 'learning_rate/qwen_vl_interface': 3.7157148173818443e-06, 'learning_rate/action_model': 3.715714817381844e-05, 'epoch': 0.22}) +[2026-10-03 13:38:10,686][starVLA.training.train_starvla][INFO] - Step 3064, Loss: {'train/loss': 7.472824108845089e-06, 'train/grad_norm_pre_clip': 0.0013421766925603151, 'timing/data': 0.0002897100057452917, 'timing/model': 0.38709922700945754, 'learning_rate/qwen_vl_interface': 3.7128329569146672e-06, 'learning_rate/action_model': 3.7128329569146666e-05, 'epoch': 0.22}) +[2026-10-03 13:38:11,380][starVLA.training.train_starvla][INFO] - Step 3065, Loss: {'train/loss': 8.43270099721849e-05, 'train/grad_norm_pre_clip': 0.03471752628684044, 'timing/data': 0.00025409999943803996, 'timing/model': 0.3879913369892165, 'learning_rate/qwen_vl_interface': 3.7099517283187977e-06, 'learning_rate/action_model': 3.7099517283187976e-05, 'epoch': 0.22}) +[2026-10-03 13:38:12,076][starVLA.training.train_starvla][INFO] - Step 3066, Loss: {'train/loss': 1.0341031156713143e-05, 'train/grad_norm_pre_clip': 0.004199125338345766, 'timing/data': 0.00026015000184997916, 'timing/model': 0.38895676798711065, 'learning_rate/qwen_vl_interface': 3.7070711327786e-06, 'learning_rate/action_model': 3.7070711327785997e-05, 'epoch': 0.22}) +[2026-10-03 13:38:12,964][starVLA.training.train_starvla][INFO] - Step 3067, Loss: {'train/loss': 1.3731142644246574e-05, 'train/grad_norm_pre_clip': 0.002684781327843666, 'timing/data': 0.00019614999473560601, 'timing/model': 0.5813929750001989, 'learning_rate/qwen_vl_interface': 3.704191171478179e-06, 'learning_rate/action_model': 3.704191171478179e-05, 'epoch': 0.22}) +[2026-10-03 13:38:13,659][starVLA.training.train_starvla][INFO] - Step 3068, Loss: {'train/loss': 0.340282678604126, 'train/grad_norm_pre_clip': 50.60813522338867, 'timing/data': 0.00028417998692020774, 'timing/model': 0.3884642569901189, 'learning_rate/qwen_vl_interface': 3.701311845601376e-06, 'learning_rate/action_model': 3.701311845601376e-05, 'epoch': 0.22}) +[2026-10-03 13:38:14,353][starVLA.training.train_starvla][INFO] - Step 3069, Loss: {'train/loss': 0.5809326171875, 'train/grad_norm_pre_clip': 47.661861419677734, 'timing/data': 0.0003417000116314739, 'timing/model': 0.3881032769859303, 'learning_rate/qwen_vl_interface': 3.6984331563317744e-06, 'learning_rate/action_model': 3.698433156331774e-05, 'epoch': 0.22}) +[2026-10-03 13:38:15,049][starVLA.training.train_starvla][INFO] - Step 3070, Loss: {'train/loss': 3.061140887439251e-05, 'train/grad_norm_pre_clip': 0.010662935674190521, 'timing/data': 0.0003091099933953956, 'timing/model': 0.3889921880036127, 'learning_rate/qwen_vl_interface': 3.695555104852695e-06, 'learning_rate/action_model': 3.6955551048526946e-05, 'epoch': 0.22}) +[2026-10-03 13:38:15,744][starVLA.training.train_starvla][INFO] - Step 3071, Loss: {'train/loss': 8.774837624514475e-05, 'train/grad_norm_pre_clip': 0.01745469495654106, 'timing/data': 0.00019425000937189907, 'timing/model': 0.3886561370018171, 'learning_rate/qwen_vl_interface': 3.6926776923471948e-06, 'learning_rate/action_model': 3.6926776923471946e-05, 'epoch': 0.22}) +[2026-10-03 13:38:16,440][starVLA.training.train_starvla][INFO] - Step 3072, Loss: {'train/loss': 1.4356949577631894e-05, 'train/grad_norm_pre_clip': 0.002682679332792759, 'timing/data': 0.00024963000032585114, 'timing/model': 0.38879722700221464, 'learning_rate/qwen_vl_interface': 3.6898009199980715e-06, 'learning_rate/action_model': 3.6898009199980714e-05, 'epoch': 0.22}) +[2026-10-03 13:38:17,321][starVLA.training.train_starvla][INFO] - Step 3073, Loss: {'train/loss': 0.03069651871919632, 'train/grad_norm_pre_clip': 9.781091690063477, 'timing/data': 0.0003813800140051171, 'timing/model': 0.3892684559978079, 'learning_rate/qwen_vl_interface': 3.686924788987855e-06, 'learning_rate/action_model': 3.686924788987855e-05, 'epoch': 0.22}) +[2026-10-03 13:38:18,015][starVLA.training.train_starvla][INFO] - Step 3074, Loss: {'train/loss': 0.00012921669986099005, 'train/grad_norm_pre_clip': 0.0459429956972599, 'timing/data': 0.00030901998979970813, 'timing/model': 0.3885378970007878, 'learning_rate/qwen_vl_interface': 3.6840493004988155e-06, 'learning_rate/action_model': 3.684049300498815e-05, 'epoch': 0.22}) +[2026-10-03 13:38:18,710][starVLA.training.train_starvla][INFO] - Step 3075, Loss: {'train/loss': 0.004834107588976622, 'train/grad_norm_pre_clip': 13.905994415283203, 'timing/data': 0.00027721999504137784, 'timing/model': 0.38852339700679295, 'learning_rate/qwen_vl_interface': 3.681174455712958e-06, 'learning_rate/action_model': 3.681174455712958e-05, 'epoch': 0.23}) +[2026-10-03 13:38:19,405][starVLA.training.train_starvla][INFO] - Step 3076, Loss: {'train/loss': 4.3710646423278376e-05, 'train/grad_norm_pre_clip': 0.019037315621972084, 'timing/data': 0.00031195000337902457, 'timing/model': 0.3885665069974493, 'learning_rate/qwen_vl_interface': 3.6783002558120206e-06, 'learning_rate/action_model': 3.67830025581202e-05, 'epoch': 0.23}) +[2026-10-03 13:38:20,099][starVLA.training.train_starvla][INFO] - Step 3077, Loss: {'train/loss': 3.9290538552450016e-05, 'train/grad_norm_pre_clip': 0.012676158919930458, 'timing/data': 0.00030278999474830925, 'timing/model': 0.38815343599708285, 'learning_rate/qwen_vl_interface': 3.6754267019774804e-06, 'learning_rate/action_model': 3.67542670197748e-05, 'epoch': 0.23}) +[2026-10-03 13:38:20,985][starVLA.training.train_starvla][INFO] - Step 3078, Loss: {'train/loss': 0.00030460156267508864, 'train/grad_norm_pre_clip': 0.2609146237373352, 'timing/data': 0.00019656999211292714, 'timing/model': 0.3884826160065131, 'learning_rate/qwen_vl_interface': 3.6725537953905453e-06, 'learning_rate/action_model': 3.672553795390545e-05, 'epoch': 0.23}) +[2026-10-03 13:38:21,678][starVLA.training.train_starvla][INFO] - Step 3079, Loss: {'train/loss': 0.10702367126941681, 'train/grad_norm_pre_clip': 264.6025085449219, 'timing/data': 0.00030575999699067324, 'timing/model': 0.3880591079941951, 'learning_rate/qwen_vl_interface': 3.6696815372321603e-06, 'learning_rate/action_model': 3.6696815372321604e-05, 'epoch': 0.23}) +[2026-10-03 13:38:22,373][starVLA.training.train_starvla][INFO] - Step 3080, Loss: {'train/loss': 0.5248585939407349, 'train/grad_norm_pre_clip': 423.2803955078125, 'timing/data': 0.00022618999355472624, 'timing/model': 0.38783161700121127, 'learning_rate/qwen_vl_interface': 3.6668099286830023e-06, 'learning_rate/action_model': 3.666809928683002e-05, 'epoch': 0.23}) +[2026-10-03 13:38:23,067][starVLA.training.train_starvla][INFO] - Step 3081, Loss: {'train/loss': 5.27493921254063e-06, 'train/grad_norm_pre_clip': 0.001194006996229291, 'timing/data': 0.0002398999931756407, 'timing/model': 0.3882588169944938, 'learning_rate/qwen_vl_interface': 3.66393897092348e-06, 'learning_rate/action_model': 3.66393897092348e-05, 'epoch': 0.23}) +[2026-10-03 13:38:23,762][starVLA.training.train_starvla][INFO] - Step 3082, Loss: {'train/loss': 0.0001682784059084952, 'train/grad_norm_pre_clip': 0.07616061717271805, 'timing/data': 0.0003036700072698295, 'timing/model': 0.38878345700504724, 'learning_rate/qwen_vl_interface': 3.661068665133735e-06, 'learning_rate/action_model': 3.661068665133735e-05, 'epoch': 0.23}) +[2026-10-03 13:38:24,645][starVLA.training.train_starvla][INFO] - Step 3083, Loss: {'train/loss': 2.426440732961055e-05, 'train/grad_norm_pre_clip': 0.010887788608670235, 'timing/data': 0.0003683499962789938, 'timing/model': 0.5781461350125028, 'learning_rate/qwen_vl_interface': 3.6581990124936434e-06, 'learning_rate/action_model': 3.6581990124936434e-05, 'epoch': 0.23}) +[2026-10-03 13:38:25,340][starVLA.training.train_starvla][INFO] - Step 3084, Loss: {'train/loss': 2.0316439986345358e-05, 'train/grad_norm_pre_clip': 0.006855475716292858, 'timing/data': 0.00019120000069960952, 'timing/model': 0.3885051069955807, 'learning_rate/qwen_vl_interface': 3.65533001418281e-06, 'learning_rate/action_model': 3.655330014182809e-05, 'epoch': 0.23}) +[2026-10-03 13:38:26,034][starVLA.training.train_starvla][INFO] - Step 3085, Loss: {'train/loss': 1.1190593795618042e-05, 'train/grad_norm_pre_clip': 0.002897165948525071, 'timing/data': 0.00019666999287437648, 'timing/model': 0.3884339770011138, 'learning_rate/qwen_vl_interface': 3.652461671380572e-06, 'learning_rate/action_model': 3.652461671380572e-05, 'epoch': 0.23}) +[2026-10-03 13:38:26,728][starVLA.training.train_starvla][INFO] - Step 3086, Loss: {'train/loss': 0.27140843868255615, 'train/grad_norm_pre_clip': 48.51138687133789, 'timing/data': 0.0002994599926751107, 'timing/model': 0.3884355860063806, 'learning_rate/qwen_vl_interface': 3.6495939852659966e-06, 'learning_rate/action_model': 3.649593985265997e-05, 'epoch': 0.23}) +[2026-10-03 13:38:27,422][starVLA.training.train_starvla][INFO] - Step 3087, Loss: {'train/loss': 1.3581944585894234e-05, 'train/grad_norm_pre_clip': 0.0048174806870520115, 'timing/data': 0.00022869999520480633, 'timing/model': 0.38844465599686373, 'learning_rate/qwen_vl_interface': 3.646726957017881e-06, 'learning_rate/action_model': 3.646726957017881e-05, 'epoch': 0.23}) +[2026-10-03 13:38:28,307][starVLA.training.train_starvla][INFO] - Step 3088, Loss: {'train/loss': 0.0002636366989463568, 'train/grad_norm_pre_clip': 0.2806587219238281, 'timing/data': 0.00028141000075265765, 'timing/model': 0.5781335259962361, 'learning_rate/qwen_vl_interface': 3.643860587814752e-06, 'learning_rate/action_model': 3.643860587814752e-05, 'epoch': 0.23}) +[2026-10-03 13:38:29,001][starVLA.training.train_starvla][INFO] - Step 3089, Loss: {'train/loss': 0.000517812732141465, 'train/grad_norm_pre_clip': 0.5419803857803345, 'timing/data': 0.00030533000244759023, 'timing/model': 0.38797478799824603, 'learning_rate/qwen_vl_interface': 3.640994878834866e-06, 'learning_rate/action_model': 3.640994878834866e-05, 'epoch': 0.23}) +[2026-10-03 13:38:29,695][starVLA.training.train_starvla][INFO] - Step 3090, Loss: {'train/loss': 8.825893746688962e-05, 'train/grad_norm_pre_clip': 0.019110044464468956, 'timing/data': 0.00025585999537725, 'timing/model': 0.3884409460006282, 'learning_rate/qwen_vl_interface': 3.6381298312562095e-06, 'learning_rate/action_model': 3.6381298312562094e-05, 'epoch': 0.23}) +[2026-10-03 13:38:30,389][starVLA.training.train_starvla][INFO] - Step 3091, Loss: {'train/loss': 0.00014812999870628119, 'train/grad_norm_pre_clip': 0.025258472189307213, 'timing/data': 0.0002680100005818531, 'timing/model': 0.3879792270017788, 'learning_rate/qwen_vl_interface': 3.635265446256493e-06, 'learning_rate/action_model': 3.6352654462564926e-05, 'epoch': 0.23}) +[2026-10-03 13:38:31,083][starVLA.training.train_starvla][INFO] - Step 3092, Loss: {'train/loss': 0.00016706823953427374, 'train/grad_norm_pre_clip': 0.035118382424116135, 'timing/data': 0.0001880000054370612, 'timing/model': 0.38837210700148717, 'learning_rate/qwen_vl_interface': 3.6324017250131565e-06, 'learning_rate/action_model': 3.6324017250131567e-05, 'epoch': 0.23}) +[2026-10-03 13:38:31,969][starVLA.training.train_starvla][INFO] - Step 3093, Loss: {'train/loss': 7.592076144646853e-06, 'train/grad_norm_pre_clip': 0.001314285909757018, 'timing/data': 0.00025151000590994954, 'timing/model': 0.5797764849994564, 'learning_rate/qwen_vl_interface': 3.6295386687033696e-06, 'learning_rate/action_model': 3.6295386687033695e-05, 'epoch': 0.23}) +[2026-10-03 13:38:32,663][starVLA.training.train_starvla][INFO] - Step 3094, Loss: {'train/loss': 7.027111132629216e-05, 'train/grad_norm_pre_clip': 0.029863161966204643, 'timing/data': 0.00039406001451425254, 'timing/model': 0.3886814269935712, 'learning_rate/qwen_vl_interface': 3.6266762785040253e-06, 'learning_rate/action_model': 3.626676278504025e-05, 'epoch': 0.23}) +[2026-10-03 13:38:33,356][starVLA.training.train_starvla][INFO] - Step 3095, Loss: {'train/loss': 6.942393520148471e-05, 'train/grad_norm_pre_clip': 0.035469502210617065, 'timing/data': 0.0003539299941621721, 'timing/model': 0.3879173070017714, 'learning_rate/qwen_vl_interface': 3.6238145555917464e-06, 'learning_rate/action_model': 3.6238145555917465e-05, 'epoch': 0.23}) +[2026-10-03 13:38:34,051][starVLA.training.train_starvla][INFO] - Step 3096, Loss: {'train/loss': 7.681472197873518e-06, 'train/grad_norm_pre_clip': 0.0016062876675277948, 'timing/data': 0.00019227000302635133, 'timing/model': 0.3884956260008039, 'learning_rate/qwen_vl_interface': 3.6209535011428757e-06, 'learning_rate/action_model': 3.620953501142876e-05, 'epoch': 0.23}) +[2026-10-03 13:38:34,745][starVLA.training.train_starvla][INFO] - Step 3097, Loss: {'train/loss': 0.0001608807360753417, 'train/grad_norm_pre_clip': 0.04212237894535065, 'timing/data': 0.00019539000641088933, 'timing/model': 0.3881136269919807, 'learning_rate/qwen_vl_interface': 3.6180931163334863e-06, 'learning_rate/action_model': 3.6180931163334863e-05, 'epoch': 0.23}) +[2026-10-03 13:38:35,439][starVLA.training.train_starvla][INFO] - Step 3098, Loss: {'train/loss': 0.06942415237426758, 'train/grad_norm_pre_clip': 131.61700439453125, 'timing/data': 0.00019356999837327749, 'timing/model': 0.3883955369965406, 'learning_rate/qwen_vl_interface': 3.6152334023393747e-06, 'learning_rate/action_model': 3.6152334023393745e-05, 'epoch': 0.23}) +[2026-10-03 13:38:36,327][starVLA.training.train_starvla][INFO] - Step 3099, Loss: {'train/loss': 0.0006158089963719249, 'train/grad_norm_pre_clip': 1.1568853855133057, 'timing/data': 0.00029237999115139246, 'timing/model': 0.38847677598823793, 'learning_rate/qwen_vl_interface': 3.6123743603360613e-06, 'learning_rate/action_model': 3.612374360336061e-05, 'epoch': 0.23}) +[2026-10-03 13:38:37,022][starVLA.training.train_starvla][INFO] - Step 3100, Loss: {'train/loss': 0.016945701092481613, 'train/grad_norm_pre_clip': 87.19874572753906, 'timing/data': 0.000377670003217645, 'timing/model': 0.3888830869982485, 'learning_rate/qwen_vl_interface': 3.6095159914987893e-06, 'learning_rate/action_model': 3.6095159914987894e-05, 'epoch': 0.23}) +[2026-10-03 13:38:37,716][starVLA.training.train_starvla][INFO] - Step 3101, Loss: {'train/loss': 0.09588608145713806, 'train/grad_norm_pre_clip': 226.6333465576172, 'timing/data': 0.0001967000134754926, 'timing/model': 0.38806730599026196, 'learning_rate/qwen_vl_interface': 3.6066582970025266e-06, 'learning_rate/action_model': 3.6066582970025264e-05, 'epoch': 0.23}) +[2026-10-03 13:38:38,411][starVLA.training.train_starvla][INFO] - Step 3102, Loss: {'train/loss': 0.035117119550704956, 'train/grad_norm_pre_clip': 191.12449645996094, 'timing/data': 0.00022622999676968902, 'timing/model': 0.38883517700014636, 'learning_rate/qwen_vl_interface': 3.603801278021964e-06, 'learning_rate/action_model': 3.6038012780219635e-05, 'epoch': 0.23}) +[2026-10-03 13:38:39,106][starVLA.training.train_starvla][INFO] - Step 3103, Loss: {'train/loss': 8.075522782746702e-05, 'train/grad_norm_pre_clip': 0.09994398057460785, 'timing/data': 0.0002295400045113638, 'timing/model': 0.38855264701123815, 'learning_rate/qwen_vl_interface': 3.6009449357315122e-06, 'learning_rate/action_model': 3.6009449357315125e-05, 'epoch': 0.23}) +[2026-10-03 13:38:39,994][starVLA.training.train_starvla][INFO] - Step 3104, Loss: {'train/loss': 1.3120103176333942e-05, 'train/grad_norm_pre_clip': 0.004553842358291149, 'timing/data': 0.0003030599909834564, 'timing/model': 0.5816712950036163, 'learning_rate/qwen_vl_interface': 3.5980892713053084e-06, 'learning_rate/action_model': 3.598089271305308e-05, 'epoch': 0.23}) +[2026-10-03 13:38:40,688][starVLA.training.train_starvla][INFO] - Step 3105, Loss: {'train/loss': 0.5649951100349426, 'train/grad_norm_pre_clip': 55.46571731567383, 'timing/data': 0.000251429999480024, 'timing/model': 0.38841294700978324, 'learning_rate/qwen_vl_interface': 3.5952342859172063e-06, 'learning_rate/action_model': 3.595234285917206e-05, 'epoch': 0.23}) +[2026-10-03 13:38:41,383][starVLA.training.train_starvla][INFO] - Step 3106, Loss: {'train/loss': 0.578480064868927, 'train/grad_norm_pre_clip': 80.83992004394531, 'timing/data': 0.00028059999749530107, 'timing/model': 0.38883157599775586, 'learning_rate/qwen_vl_interface': 3.592379980740785e-06, 'learning_rate/action_model': 3.5923799807407845e-05, 'epoch': 0.23}) +[2026-10-03 13:38:42,078][starVLA.training.train_starvla][INFO] - Step 3107, Loss: {'train/loss': 8.297005842905492e-05, 'train/grad_norm_pre_clip': 0.02879401296377182, 'timing/data': 0.0002497300010873005, 'timing/model': 0.3887218169984408, 'learning_rate/qwen_vl_interface': 3.5895263569493384e-06, 'learning_rate/action_model': 3.589526356949338e-05, 'epoch': 0.23}) +[2026-10-03 13:38:42,772][starVLA.training.train_starvla][INFO] - Step 3108, Loss: {'train/loss': 2.2171980162966065e-05, 'train/grad_norm_pre_clip': 0.007376156747341156, 'timing/data': 0.00027956999838352203, 'timing/model': 0.38808178700855933, 'learning_rate/qwen_vl_interface': 3.586673415715885e-06, 'learning_rate/action_model': 3.586673415715885e-05, 'epoch': 0.23}) +[2026-10-03 13:38:43,466][starVLA.training.train_starvla][INFO] - Step 3109, Loss: {'train/loss': 9.774989848665427e-06, 'train/grad_norm_pre_clip': 0.004450177773833275, 'timing/data': 0.00044334000267554075, 'timing/model': 0.387823228011257, 'learning_rate/qwen_vl_interface': 3.5838211582131603e-06, 'learning_rate/action_model': 3.58382115821316e-05, 'epoch': 0.23}) +[2026-10-03 13:38:44,348][starVLA.training.train_starvla][INFO] - Step 3110, Loss: {'train/loss': 4.124765837332234e-05, 'train/grad_norm_pre_clip': 0.013129434548318386, 'timing/data': 0.0002397700009169057, 'timing/model': 0.38858008600072935, 'learning_rate/qwen_vl_interface': 3.580969585613622e-06, 'learning_rate/action_model': 3.580969585613622e-05, 'epoch': 0.23}) +[2026-10-03 13:38:45,041][starVLA.training.train_starvla][INFO] - Step 3111, Loss: {'train/loss': 6.033451791154221e-05, 'train/grad_norm_pre_clip': 0.019909942522644997, 'timing/data': 0.00019377999706193805, 'timing/model': 0.3867564370011678, 'learning_rate/qwen_vl_interface': 3.5781186990894423e-06, 'learning_rate/action_model': 3.578118699089442e-05, 'epoch': 0.23}) +[2026-10-03 13:38:45,735][starVLA.training.train_starvla][INFO] - Step 3112, Loss: {'train/loss': 0.00015194671868812293, 'train/grad_norm_pre_clip': 0.03255457803606987, 'timing/data': 0.0004191799962427467, 'timing/model': 0.38772900699405, 'learning_rate/qwen_vl_interface': 3.5752684998125113e-06, 'learning_rate/action_model': 3.5752684998125116e-05, 'epoch': 0.23}) +[2026-10-03 13:38:46,428][starVLA.training.train_starvla][INFO] - Step 3113, Loss: {'train/loss': 0.0018268540734425187, 'train/grad_norm_pre_clip': 4.110230922698975, 'timing/data': 0.00024731000303290784, 'timing/model': 0.3876573970046593, 'learning_rate/qwen_vl_interface': 3.5724189889544407e-06, 'learning_rate/action_model': 3.572418988954441e-05, 'epoch': 0.23}) +[2026-10-03 13:38:47,123][starVLA.training.train_starvla][INFO] - Step 3114, Loss: {'train/loss': 4.444429214345291e-05, 'train/grad_norm_pre_clip': 0.026623887941241264, 'timing/data': 0.0002064400032395497, 'timing/model': 0.3892187369929161, 'learning_rate/qwen_vl_interface': 3.569570167686556e-06, 'learning_rate/action_model': 3.569570167686556e-05, 'epoch': 0.23}) +[2026-10-03 13:38:48,007][starVLA.training.train_starvla][INFO] - Step 3115, Loss: {'train/loss': 0.00017268025840166956, 'train/grad_norm_pre_clip': 0.11289431154727936, 'timing/data': 0.00033363999682478607, 'timing/model': 0.38872848699975293, 'learning_rate/qwen_vl_interface': 3.5667220371798995e-06, 'learning_rate/action_model': 3.566722037179899e-05, 'epoch': 0.23}) +[2026-10-03 13:38:48,703][starVLA.training.train_starvla][INFO] - Step 3116, Loss: {'train/loss': 3.674135950859636e-05, 'train/grad_norm_pre_clip': 0.01751112751662731, 'timing/data': 0.000272019999101758, 'timing/model': 0.3891923369956203, 'learning_rate/qwen_vl_interface': 3.563874598605232e-06, 'learning_rate/action_model': 3.5638745986052316e-05, 'epoch': 0.23}) +[2026-10-03 13:38:49,397][starVLA.training.train_starvla][INFO] - Step 3117, Loss: {'train/loss': 2.323739681742154e-05, 'train/grad_norm_pre_clip': 0.009062306955456734, 'timing/data': 0.0002355799952056259, 'timing/model': 0.38842275699425954, 'learning_rate/qwen_vl_interface': 3.561027853133025e-06, 'learning_rate/action_model': 3.561027853133025e-05, 'epoch': 0.23}) +[2026-10-03 13:38:50,091][starVLA.training.train_starvla][INFO] - Step 3118, Loss: {'train/loss': 0.0026411942671984434, 'train/grad_norm_pre_clip': 1.1581305265426636, 'timing/data': 0.0002970500063383952, 'timing/model': 0.38865651600644924, 'learning_rate/qwen_vl_interface': 3.558181801933472e-06, 'learning_rate/action_model': 3.5581818019334714e-05, 'epoch': 0.23}) +[2026-10-03 13:38:50,786][starVLA.training.train_starvla][INFO] - Step 3119, Loss: {'train/loss': 0.0003699523804243654, 'train/grad_norm_pre_clip': 0.12600918114185333, 'timing/data': 0.00026797001191880554, 'timing/model': 0.3890948769985698, 'learning_rate/qwen_vl_interface': 3.555336446176473e-06, 'learning_rate/action_model': 3.5553364461764726e-05, 'epoch': 0.23}) +[2026-10-03 13:38:51,675][starVLA.training.train_starvla][INFO] - Step 3120, Loss: {'train/loss': 0.0004488107515498996, 'train/grad_norm_pre_clip': 0.6877889633178711, 'timing/data': 0.00026548000460024923, 'timing/model': 0.5810735150007531, 'learning_rate/qwen_vl_interface': 3.552491787031647e-06, 'learning_rate/action_model': 3.552491787031647e-05, 'epoch': 0.23}) +[2026-10-03 13:38:52,369][starVLA.training.train_starvla][INFO] - Step 3121, Loss: {'train/loss': 0.00023461636737920344, 'train/grad_norm_pre_clip': 0.04917679727077484, 'timing/data': 0.00028330000350251794, 'timing/model': 0.38858467699901666, 'learning_rate/qwen_vl_interface': 3.5496478256683277e-06, 'learning_rate/action_model': 3.549647825668327e-05, 'epoch': 0.23}) +[2026-10-03 13:38:53,064][starVLA.training.train_starvla][INFO] - Step 3122, Loss: {'train/loss': 0.0016344485338777304, 'train/grad_norm_pre_clip': 3.9118030071258545, 'timing/data': 0.00023443000100087374, 'timing/model': 0.38858359699952416, 'learning_rate/qwen_vl_interface': 3.5468045632555585e-06, 'learning_rate/action_model': 3.546804563255559e-05, 'epoch': 0.23}) +[2026-10-03 13:38:53,758][starVLA.training.train_starvla][INFO] - Step 3123, Loss: {'train/loss': 9.44451239774935e-05, 'train/grad_norm_pre_clip': 0.044267311692237854, 'timing/data': 0.00031211999885272235, 'timing/model': 0.3889088659925619, 'learning_rate/qwen_vl_interface': 3.5439620009620995e-06, 'learning_rate/action_model': 3.543962000962099e-05, 'epoch': 0.23}) +[2026-10-03 13:38:54,455][starVLA.training.train_starvla][INFO] - Step 3124, Loss: {'train/loss': 2.716307972150389e-05, 'train/grad_norm_pre_clip': 0.007604160811752081, 'timing/data': 0.00019133000751025975, 'timing/model': 0.38946022698655725, 'learning_rate/qwen_vl_interface': 3.541120139956419e-06, 'learning_rate/action_model': 3.5411201399564184e-05, 'epoch': 0.23}) +[2026-10-03 13:38:55,150][starVLA.training.train_starvla][INFO] - Step 3125, Loss: {'train/loss': 0.24645887315273285, 'train/grad_norm_pre_clip': 34.70756912231445, 'timing/data': 0.00019859999883919954, 'timing/model': 0.38866319600492716, 'learning_rate/qwen_vl_interface': 3.538278981406697e-06, 'learning_rate/action_model': 3.5382789814066966e-05, 'epoch': 0.23}) +[2026-10-03 13:38:56,032][starVLA.training.train_starvla][INFO] - Step 3126, Loss: {'train/loss': 3.360125265317038e-05, 'train/grad_norm_pre_clip': 0.013735828921198845, 'timing/data': 0.00029729900415986776, 'timing/model': 0.38871135699446313, 'learning_rate/qwen_vl_interface': 3.5354385264808295e-06, 'learning_rate/action_model': 3.535438526480829e-05, 'epoch': 0.23}) +[2026-10-03 13:38:56,728][starVLA.training.train_starvla][INFO] - Step 3127, Loss: {'train/loss': 0.0001371016987832263, 'train/grad_norm_pre_clip': 0.17256152629852295, 'timing/data': 0.00027628999669104815, 'timing/model': 0.38943900700542144, 'learning_rate/qwen_vl_interface': 3.5325987763464196e-06, 'learning_rate/action_model': 3.5325987763464194e-05, 'epoch': 0.23}) +[2026-10-03 13:38:57,424][starVLA.training.train_starvla][INFO] - Step 3128, Loss: {'train/loss': 0.0002035298093687743, 'train/grad_norm_pre_clip': 0.04884572699666023, 'timing/data': 0.0002568789932411164, 'timing/model': 0.3883933269971749, 'learning_rate/qwen_vl_interface': 3.529759732170781e-06, 'learning_rate/action_model': 3.529759732170781e-05, 'epoch': 0.23}) +[2026-10-03 13:38:58,119][starVLA.training.train_starvla][INFO] - Step 3129, Loss: {'train/loss': 8.959288970800117e-05, 'train/grad_norm_pre_clip': 0.028969403356313705, 'timing/data': 0.00019341999723110348, 'timing/model': 0.38887799700023606, 'learning_rate/qwen_vl_interface': 3.5269213951209375e-06, 'learning_rate/action_model': 3.5269213951209374e-05, 'epoch': 0.23}) +[2026-10-03 13:38:58,813][starVLA.training.train_starvla][INFO] - Step 3130, Loss: {'train/loss': 9.99489493551664e-05, 'train/grad_norm_pre_clip': 0.03781045600771904, 'timing/data': 0.00030107999918982387, 'timing/model': 0.3884997170098359, 'learning_rate/qwen_vl_interface': 3.5240837663636207e-06, 'learning_rate/action_model': 3.5240837663636204e-05, 'epoch': 0.23}) +[2026-10-03 13:38:59,697][starVLA.training.train_starvla][INFO] - Step 3131, Loss: {'train/loss': 0.015627402812242508, 'train/grad_norm_pre_clip': 27.901195526123047, 'timing/data': 0.00019768999482039362, 'timing/model': 0.38856031600153074, 'learning_rate/qwen_vl_interface': 3.5212468470652745e-06, 'learning_rate/action_model': 3.521246847065274e-05, 'epoch': 0.23}) +[2026-10-03 13:39:00,392][starVLA.training.train_starvla][INFO] - Step 3132, Loss: {'train/loss': 0.0003962252812925726, 'train/grad_norm_pre_clip': 0.14936916530132294, 'timing/data': 0.00027704999956768006, 'timing/model': 0.38840087699645665, 'learning_rate/qwen_vl_interface': 3.5184106383920487e-06, 'learning_rate/action_model': 3.5184106383920486e-05, 'epoch': 0.23}) +[2026-10-03 13:39:01,085][starVLA.training.train_starvla][INFO] - Step 3133, Loss: {'train/loss': 0.0010232900967821479, 'train/grad_norm_pre_clip': 0.4434419870376587, 'timing/data': 0.00034516000596340746, 'timing/model': 0.387471367008402, 'learning_rate/qwen_vl_interface': 3.5155751415098022e-06, 'learning_rate/action_model': 3.515575141509802e-05, 'epoch': 0.23}) +[2026-10-03 13:39:01,780][starVLA.training.train_starvla][INFO] - Step 3134, Loss: {'train/loss': 0.00017731203115545213, 'train/grad_norm_pre_clip': 0.04509403929114342, 'timing/data': 0.00030571999377571046, 'timing/model': 0.3887161170132458, 'learning_rate/qwen_vl_interface': 3.5127403575841e-06, 'learning_rate/action_model': 3.5127403575841e-05, 'epoch': 0.23}) +[2026-10-03 13:39:02,474][starVLA.training.train_starvla][INFO] - Step 3135, Loss: {'train/loss': 1.1622387319221161e-05, 'train/grad_norm_pre_clip': 0.009523133747279644, 'timing/data': 0.0002681600017240271, 'timing/model': 0.38831153699720744, 'learning_rate/qwen_vl_interface': 3.509906287780215e-06, 'learning_rate/action_model': 3.5099062877802144e-05, 'epoch': 0.23}) +[2026-10-03 13:39:03,355][starVLA.training.train_starvla][INFO] - Step 3136, Loss: {'train/loss': 1.396178049617447e-05, 'train/grad_norm_pre_clip': 0.005662340670824051, 'timing/data': 0.0002743200020631775, 'timing/model': 0.5750696970062563, 'learning_rate/qwen_vl_interface': 3.5070729332631247e-06, 'learning_rate/action_model': 3.507072933263125e-05, 'epoch': 0.23}) +[2026-10-03 13:39:04,049][starVLA.training.train_starvla][INFO] - Step 3137, Loss: {'train/loss': 6.579376349691302e-05, 'train/grad_norm_pre_clip': 0.03788386285305023, 'timing/data': 0.0002310800045961514, 'timing/model': 0.38826391600014176, 'learning_rate/qwen_vl_interface': 3.5042402951975167e-06, 'learning_rate/action_model': 3.5042402951975167e-05, 'epoch': 0.23}) +[2026-10-03 13:39:04,744][starVLA.training.train_starvla][INFO] - Step 3138, Loss: {'train/loss': 0.0007584282429888844, 'train/grad_norm_pre_clip': 0.542120635509491, 'timing/data': 0.0003060300077777356, 'timing/model': 0.38891988700197544, 'learning_rate/qwen_vl_interface': 3.50140837474778e-06, 'learning_rate/action_model': 3.50140837474778e-05, 'epoch': 0.23}) +[2026-10-03 13:39:05,438][starVLA.training.train_starvla][INFO] - Step 3139, Loss: {'train/loss': 2.071254584734561e-06, 'train/grad_norm_pre_clip': 0.0006927911890670657, 'timing/data': 0.0002711500128498301, 'timing/model': 0.3883032470039325, 'learning_rate/qwen_vl_interface': 3.498577173078011e-06, 'learning_rate/action_model': 3.498577173078011e-05, 'epoch': 0.23}) +[2026-10-03 13:39:06,133][starVLA.training.train_starvla][INFO] - Step 3140, Loss: {'train/loss': 0.00015058113785926253, 'train/grad_norm_pre_clip': 0.10583779960870743, 'timing/data': 0.0002899999963119626, 'timing/model': 0.38840589601022657, 'learning_rate/qwen_vl_interface': 3.495746691352008e-06, 'learning_rate/action_model': 3.495746691352008e-05, 'epoch': 0.23}) +[2026-10-03 13:39:07,018][starVLA.training.train_starvla][INFO] - Step 3141, Loss: {'train/loss': 0.0002014355850405991, 'train/grad_norm_pre_clip': 0.0476822704076767, 'timing/data': 0.000192170002264902, 'timing/model': 0.579325215992867, 'learning_rate/qwen_vl_interface': 3.4929169307332763e-06, 'learning_rate/action_model': 3.492916930733276e-05, 'epoch': 0.23}) +[2026-10-03 13:39:07,712][starVLA.training.train_starvla][INFO] - Step 3142, Loss: {'train/loss': 0.008820162154734135, 'train/grad_norm_pre_clip': 24.76600456237793, 'timing/data': 0.0003050399973290041, 'timing/model': 0.38839814699895214, 'learning_rate/qwen_vl_interface': 3.4900878923850234e-06, 'learning_rate/action_model': 3.490087892385023e-05, 'epoch': 0.23}) +[2026-10-03 13:39:08,407][starVLA.training.train_starvla][INFO] - Step 3143, Loss: {'train/loss': 4.705533501692116e-05, 'train/grad_norm_pre_clip': 0.011769226752221584, 'timing/data': 0.0002511000056983903, 'timing/model': 0.38837488599529024, 'learning_rate/qwen_vl_interface': 3.4872595774701607e-06, 'learning_rate/action_model': 3.487259577470161e-05, 'epoch': 0.23}) +[2026-10-03 13:39:09,102][starVLA.training.train_starvla][INFO] - Step 3144, Loss: {'train/loss': 9.417624823981896e-05, 'train/grad_norm_pre_clip': 0.059651631861925125, 'timing/data': 0.00019557999621611089, 'timing/model': 0.38857366699085105, 'learning_rate/qwen_vl_interface': 3.4844319871513014e-06, 'learning_rate/action_model': 3.484431987151301e-05, 'epoch': 0.23}) +[2026-10-03 13:39:09,796][starVLA.training.train_starvla][INFO] - Step 3145, Loss: {'train/loss': 0.22019678354263306, 'train/grad_norm_pre_clip': 222.266357421875, 'timing/data': 0.000238690001424402, 'timing/model': 0.3869764260016382, 'learning_rate/qwen_vl_interface': 3.4816051225907606e-06, 'learning_rate/action_model': 3.48160512259076e-05, 'epoch': 0.23}) +[2026-10-03 13:39:10,685][starVLA.training.train_starvla][INFO] - Step 3146, Loss: {'train/loss': 3.6507613003777806e-06, 'train/grad_norm_pre_clip': 0.0009284047991968691, 'timing/data': 0.00020349999249447137, 'timing/model': 0.5824818160035647, 'learning_rate/qwen_vl_interface': 3.478778984950556e-06, 'learning_rate/action_model': 3.478778984950556e-05, 'epoch': 0.23}) +[2026-10-03 13:39:11,380][starVLA.training.train_starvla][INFO] - Step 3147, Loss: {'train/loss': 5.215309556660941e-06, 'train/grad_norm_pre_clip': 0.002394148614257574, 'timing/data': 0.0002995000104419887, 'timing/model': 0.3889819960022578, 'learning_rate/qwen_vl_interface': 3.475953575392405e-06, 'learning_rate/action_model': 3.475953575392405e-05, 'epoch': 0.23}) +[2026-10-03 13:39:12,075][starVLA.training.train_starvla][INFO] - Step 3148, Loss: {'train/loss': 7.388963422272354e-05, 'train/grad_norm_pre_clip': 0.034513313323259354, 'timing/data': 0.00025348000053782016, 'timing/model': 0.3883814870059723, 'learning_rate/qwen_vl_interface': 3.473128895077728e-06, 'learning_rate/action_model': 3.473128895077728e-05, 'epoch': 0.23}) +[2026-10-03 13:39:12,770][starVLA.training.train_starvla][INFO] - Step 3149, Loss: {'train/loss': 7.644049219379667e-06, 'train/grad_norm_pre_clip': 0.011588484048843384, 'timing/data': 0.0002691300032893196, 'timing/model': 0.38895999699889217, 'learning_rate/qwen_vl_interface': 3.4703049451676453e-06, 'learning_rate/action_model': 3.4703049451676454e-05, 'epoch': 0.23}) +[2026-10-03 13:39:13,466][starVLA.training.train_starvla][INFO] - Step 3150, Loss: {'train/loss': 3.8098907680250704e-05, 'train/grad_norm_pre_clip': 0.011296345852315426, 'timing/data': 0.00023128000611905009, 'timing/model': 0.3893567970080767, 'learning_rate/qwen_vl_interface': 3.4674817268229732e-06, 'learning_rate/action_model': 3.467481726822973e-05, 'epoch': 0.23}) +[2026-10-03 13:39:14,161][starVLA.training.train_starvla][INFO] - Step 3151, Loss: {'train/loss': 7.25669178791577e-06, 'train/grad_norm_pre_clip': 0.0036754736211150885, 'timing/data': 0.00023516999499406666, 'timing/model': 0.38892234599916264, 'learning_rate/qwen_vl_interface': 3.464659241204232e-06, 'learning_rate/action_model': 3.4646592412042314e-05, 'epoch': 0.23}) +[2026-10-03 13:39:15,050][starVLA.training.train_starvla][INFO] - Step 3152, Loss: {'train/loss': 0.00017773767467588186, 'train/grad_norm_pre_clip': 0.16711458563804626, 'timing/data': 0.0002911100018536672, 'timing/model': 0.3882209160074126, 'learning_rate/qwen_vl_interface': 3.4618374894716377e-06, 'learning_rate/action_model': 3.461837489471638e-05, 'epoch': 0.23}) +[2026-10-03 13:39:15,745][starVLA.training.train_starvla][INFO] - Step 3153, Loss: {'train/loss': 1.5273647022695513e-06, 'train/grad_norm_pre_clip': 0.00046829605707898736, 'timing/data': 0.00025190999440383166, 'timing/model': 0.3885514569992665, 'learning_rate/qwen_vl_interface': 3.459016472785107e-06, 'learning_rate/action_model': 3.459016472785107e-05, 'epoch': 0.23}) +[2026-10-03 13:39:16,441][starVLA.training.train_starvla][INFO] - Step 3154, Loss: {'train/loss': 6.482000571850222e-07, 'train/grad_norm_pre_clip': 0.00013655000657308847, 'timing/data': 0.00031501999183092266, 'timing/model': 0.3892159060051199, 'learning_rate/qwen_vl_interface': 3.456196192304254e-06, 'learning_rate/action_model': 3.456196192304254e-05, 'epoch': 0.23}) +[2026-10-03 13:39:17,136][starVLA.training.train_starvla][INFO] - Step 3155, Loss: {'train/loss': 0.001880924217402935, 'train/grad_norm_pre_clip': 5.850612640380859, 'timing/data': 0.00034240000240970403, 'timing/model': 0.3885974069999065, 'learning_rate/qwen_vl_interface': 3.4533766491883887e-06, 'learning_rate/action_model': 3.453376649188388e-05, 'epoch': 0.23}) +[2026-10-03 13:39:17,831][starVLA.training.train_starvla][INFO] - Step 3156, Loss: {'train/loss': 0.6689468026161194, 'train/grad_norm_pre_clip': 374.9750061035156, 'timing/data': 0.0002510899939807132, 'timing/model': 0.3889917170017725, 'learning_rate/qwen_vl_interface': 3.450557844596517e-06, 'learning_rate/action_model': 3.450557844596517e-05, 'epoch': 0.23}) +[2026-10-03 13:39:18,716][starVLA.training.train_starvla][INFO] - Step 3157, Loss: {'train/loss': 2.2649671791441506e-06, 'train/grad_norm_pre_clip': 0.0006493359105661511, 'timing/data': 0.00021221999486442655, 'timing/model': 0.5786094749928452, 'learning_rate/qwen_vl_interface': 3.447739779687345e-06, 'learning_rate/action_model': 3.4477397796873446e-05, 'epoch': 0.23}) +[2026-10-03 13:39:19,411][starVLA.training.train_starvla][INFO] - Step 3158, Loss: {'train/loss': 0.5101380944252014, 'train/grad_norm_pre_clip': 125.38969421386719, 'timing/data': 0.0002463200071360916, 'timing/model': 0.38875951699446887, 'learning_rate/qwen_vl_interface': 3.4449224556192722e-06, 'learning_rate/action_model': 3.444922455619272e-05, 'epoch': 0.23}) +[2026-10-03 13:39:20,106][starVLA.training.train_starvla][INFO] - Step 3159, Loss: {'train/loss': 0.04544873163104057, 'train/grad_norm_pre_clip': 25.42481803894043, 'timing/data': 0.00022857000294607133, 'timing/model': 0.3885198659991147, 'learning_rate/qwen_vl_interface': 3.442105873550395e-06, 'learning_rate/action_model': 3.442105873550395e-05, 'epoch': 0.23}) +[2026-10-03 13:39:20,801][starVLA.training.train_starvla][INFO] - Step 3160, Loss: {'train/loss': 4.138206713832915e-05, 'train/grad_norm_pre_clip': 0.017292387783527374, 'timing/data': 0.00030115000845398754, 'timing/model': 0.3884570380032528, 'learning_rate/qwen_vl_interface': 3.4392900346385022e-06, 'learning_rate/action_model': 3.4392900346385024e-05, 'epoch': 0.23}) +[2026-10-03 13:39:21,496][starVLA.training.train_starvla][INFO] - Step 3161, Loss: {'train/loss': 1.3485500858223531e-06, 'train/grad_norm_pre_clip': 0.0006799295078963041, 'timing/data': 0.0003025799960596487, 'timing/model': 0.3889405659865588, 'learning_rate/qwen_vl_interface': 3.4364749400410796e-06, 'learning_rate/action_model': 3.43647494004108e-05, 'epoch': 0.23}) +[2026-10-03 13:39:22,194][starVLA.training.train_starvla][INFO] - Step 3162, Loss: {'train/loss': 1.0430795782667701e-06, 'train/grad_norm_pre_clip': 0.00029609640478156507, 'timing/data': 0.0003346500016050413, 'timing/model': 0.39018268599465955, 'learning_rate/qwen_vl_interface': 3.4336605909153066e-06, 'learning_rate/action_model': 3.433660590915306e-05, 'epoch': 0.23}) +[2026-10-03 13:39:23,080][starVLA.training.train_starvla][INFO] - Step 3163, Loss: {'train/loss': 5.736943649026216e-07, 'train/grad_norm_pre_clip': 0.00021732803725171834, 'timing/data': 0.00040109999827109277, 'timing/model': 0.39025135700649116, 'learning_rate/qwen_vl_interface': 3.430846988418057e-06, 'learning_rate/action_model': 3.430846988418057e-05, 'epoch': 0.23}) +[2026-10-03 13:39:23,779][starVLA.training.train_starvla][INFO] - Step 3164, Loss: {'train/loss': 0.10216569900512695, 'train/grad_norm_pre_clip': 92.12044525146484, 'timing/data': 0.00029618000553455204, 'timing/model': 0.3900764559948584, 'learning_rate/qwen_vl_interface': 3.428034133705893e-06, 'learning_rate/action_model': 3.428034133705893e-05, 'epoch': 0.23}) +[2026-10-03 13:39:24,476][starVLA.training.train_starvla][INFO] - Step 3165, Loss: {'train/loss': 4.589483069139533e-06, 'train/grad_norm_pre_clip': 0.0020049407612532377, 'timing/data': 0.00028731000202242285, 'timing/model': 0.3903694570035441, 'learning_rate/qwen_vl_interface': 3.4252220279350763e-06, 'learning_rate/action_model': 3.425222027935076e-05, 'epoch': 0.23}) +[2026-10-03 13:39:25,175][starVLA.training.train_starvla][INFO] - Step 3166, Loss: {'train/loss': 0.44432905316352844, 'train/grad_norm_pre_clip': 124.49032592773438, 'timing/data': 0.00029198999982327223, 'timing/model': 0.3901500570063945, 'learning_rate/qwen_vl_interface': 3.4224106722615547e-06, 'learning_rate/action_model': 3.422410672261554e-05, 'epoch': 0.23}) +[2026-10-03 13:39:25,872][starVLA.training.train_starvla][INFO] - Step 3167, Loss: {'train/loss': 0.05493087321519852, 'train/grad_norm_pre_clip': 20.769941329956055, 'timing/data': 0.000248780008405447, 'timing/model': 0.39023887598887086, 'learning_rate/qwen_vl_interface': 3.419600067840973e-06, 'learning_rate/action_model': 3.419600067840973e-05, 'epoch': 0.23}) +[2026-10-03 13:39:26,760][starVLA.training.train_starvla][INFO] - Step 3168, Loss: {'train/loss': 2.491769555490464e-05, 'train/grad_norm_pre_clip': 0.010077131912112236, 'timing/data': 0.000285230009467341, 'timing/model': 0.39026736600499135, 'learning_rate/qwen_vl_interface': 3.4167902158286627e-06, 'learning_rate/action_model': 3.4167902158286626e-05, 'epoch': 0.23}) +[2026-10-03 13:39:27,458][starVLA.training.train_starvla][INFO] - Step 3169, Loss: {'train/loss': 3.610246130847372e-05, 'train/grad_norm_pre_clip': 0.022603658959269524, 'timing/data': 0.0003176200116286054, 'timing/model': 0.39026117698813323, 'learning_rate/qwen_vl_interface': 3.4139811173796468e-06, 'learning_rate/action_model': 3.413981117379647e-05, 'epoch': 0.23}) +[2026-10-03 13:39:28,156][starVLA.training.train_starvla][INFO] - Step 3170, Loss: {'train/loss': 0.0005520934355445206, 'train/grad_norm_pre_clip': 0.6907826066017151, 'timing/data': 0.0003517299919622019, 'timing/model': 0.3904956069891341, 'learning_rate/qwen_vl_interface': 3.411172773648642e-06, 'learning_rate/action_model': 3.411172773648642e-05, 'epoch': 0.23}) +[2026-10-03 13:39:28,854][starVLA.training.train_starvla][INFO] - Step 3171, Loss: {'train/loss': 0.00010977635974995792, 'train/grad_norm_pre_clip': 0.042833734303712845, 'timing/data': 0.00031570999999530613, 'timing/model': 0.3901903469959507, 'learning_rate/qwen_vl_interface': 3.408365185790052e-06, 'learning_rate/action_model': 3.4083651857900515e-05, 'epoch': 0.23}) +[2026-10-03 13:39:29,553][starVLA.training.train_starvla][INFO] - Step 3172, Loss: {'train/loss': 0.003605199744924903, 'train/grad_norm_pre_clip': 6.607896327972412, 'timing/data': 0.00026991999766323715, 'timing/model': 0.39060100600181613, 'learning_rate/qwen_vl_interface': 3.40555835495797e-06, 'learning_rate/action_model': 3.40555835495797e-05, 'epoch': 0.23}) +[2026-10-03 13:39:30,439][starVLA.training.train_starvla][INFO] - Step 3173, Loss: {'train/loss': 0.061158690601587296, 'train/grad_norm_pre_clip': 220.3317413330078, 'timing/data': 0.00023694999981671572, 'timing/model': 0.5779156060016248, 'learning_rate/qwen_vl_interface': 3.4027522823061763e-06, 'learning_rate/action_model': 3.4027522823061765e-05, 'epoch': 0.23}) +[2026-10-03 13:39:31,138][starVLA.training.train_starvla][INFO] - Step 3174, Loss: {'train/loss': 4.656478722608881e-06, 'train/grad_norm_pre_clip': 0.00456848181784153, 'timing/data': 0.00026066000282298774, 'timing/model': 0.3909138669987442, 'learning_rate/qwen_vl_interface': 3.399946968988143e-06, 'learning_rate/action_model': 3.399946968988143e-05, 'epoch': 0.23}) +[2026-10-03 13:39:31,836][starVLA.training.train_starvla][INFO] - Step 3175, Loss: {'train/loss': 0.0001502819504821673, 'train/grad_norm_pre_clip': 0.7440465688705444, 'timing/data': 0.00019557999621611089, 'timing/model': 0.3904127769928891, 'learning_rate/qwen_vl_interface': 3.397142416157028e-06, 'learning_rate/action_model': 3.397142416157028e-05, 'epoch': 0.23}) +[2026-10-03 13:39:32,535][starVLA.training.train_starvla][INFO] - Step 3176, Loss: {'train/loss': 3.300586740806466e-06, 'train/grad_norm_pre_clip': 0.001275195274502039, 'timing/data': 0.0003018099960172549, 'timing/model': 0.39012342599744443, 'learning_rate/qwen_vl_interface': 3.3943386249656766e-06, 'learning_rate/action_model': 3.394338624965677e-05, 'epoch': 0.23}) +[2026-10-03 13:39:33,233][starVLA.training.train_starvla][INFO] - Step 3177, Loss: {'train/loss': 0.00016419953317381442, 'train/grad_norm_pre_clip': 0.09371872246265411, 'timing/data': 0.0002808999997796491, 'timing/model': 0.39027842599898577, 'learning_rate/qwen_vl_interface': 3.391535596566624e-06, 'learning_rate/action_model': 3.391535596566624e-05, 'epoch': 0.23}) +[2026-10-03 13:39:33,931][starVLA.training.train_starvla][INFO] - Step 3178, Loss: {'train/loss': 1.0236436537525151e-05, 'train/grad_norm_pre_clip': 0.0036223563365638256, 'timing/data': 0.00019356999837327749, 'timing/model': 0.3904695959936362, 'learning_rate/qwen_vl_interface': 3.388733332112086e-06, 'learning_rate/action_model': 3.3887333321120856e-05, 'epoch': 0.23}) +[2026-10-03 13:39:34,816][starVLA.training.train_starvla][INFO] - Step 3179, Loss: {'train/loss': 0.0001406716473866254, 'train/grad_norm_pre_clip': 0.4954056441783905, 'timing/data': 0.0002013199991779402, 'timing/model': 0.39060711600177456, 'learning_rate/qwen_vl_interface': 3.3859318327539693e-06, 'learning_rate/action_model': 3.385931832753969e-05, 'epoch': 0.23}) +[2026-10-03 13:39:35,514][starVLA.training.train_starvla][INFO] - Step 3180, Loss: {'train/loss': 0.5453150272369385, 'train/grad_norm_pre_clip': 135.56216430664062, 'timing/data': 0.00019428000086918473, 'timing/model': 0.39013593600247987, 'learning_rate/qwen_vl_interface': 3.3831310996438633e-06, 'learning_rate/action_model': 3.383131099643863e-05, 'epoch': 0.23}) +[2026-10-03 13:39:36,212][starVLA.training.train_starvla][INFO] - Step 3181, Loss: {'train/loss': 2.3564096409245394e-05, 'train/grad_norm_pre_clip': 0.01101988647133112, 'timing/data': 0.00034442001197021455, 'timing/model': 0.3906318469962571, 'learning_rate/qwen_vl_interface': 3.3803311339330447e-06, 'learning_rate/action_model': 3.3803311339330444e-05, 'epoch': 0.23}) +[2026-10-03 13:39:36,916][starVLA.training.train_starvla][INFO] - Step 3182, Loss: {'train/loss': 3.168280454701744e-05, 'train/grad_norm_pre_clip': 0.044172465801239014, 'timing/data': 0.0001976100029423833, 'timing/model': 0.3958210760029033, 'learning_rate/qwen_vl_interface': 3.3775319367724746e-06, 'learning_rate/action_model': 3.3775319367724745e-05, 'epoch': 0.23}) +[2026-10-03 13:39:37,614][starVLA.training.train_starvla][INFO] - Step 3183, Loss: {'train/loss': 1.9371443613636075e-06, 'train/grad_norm_pre_clip': 0.0009270616574212909, 'timing/data': 0.00019841999164782465, 'timing/model': 0.3898925359972054, 'learning_rate/qwen_vl_interface': 3.374733509312795e-06, 'learning_rate/action_model': 3.374733509312795e-05, 'epoch': 0.23}) +[2026-10-03 13:39:38,502][starVLA.training.train_starvla][INFO] - Step 3184, Loss: {'train/loss': 1.6092297300929204e-05, 'train/grad_norm_pre_clip': 0.01265647541731596, 'timing/data': 0.0002780600043479353, 'timing/model': 0.3898978960060049, 'learning_rate/qwen_vl_interface': 3.3719358527043345e-06, 'learning_rate/action_model': 3.371935852704334e-05, 'epoch': 0.23}) +[2026-10-03 13:39:39,200][starVLA.training.train_starvla][INFO] - Step 3185, Loss: {'train/loss': 0.00017878753715194762, 'train/grad_norm_pre_clip': 0.3047860264778137, 'timing/data': 0.0002962200087495148, 'timing/model': 0.3902698970050551, 'learning_rate/qwen_vl_interface': 3.3691389680971047e-06, 'learning_rate/action_model': 3.3691389680971044e-05, 'epoch': 0.23}) +[2026-10-03 13:39:39,899][starVLA.training.train_starvla][INFO] - Step 3186, Loss: {'train/loss': 0.005482954904437065, 'train/grad_norm_pre_clip': 6.620855331420898, 'timing/data': 0.0001996100036194548, 'timing/model': 0.3907624069979647, 'learning_rate/qwen_vl_interface': 3.3663428566407985e-06, 'learning_rate/action_model': 3.366342856640798e-05, 'epoch': 0.23}) +[2026-10-03 13:39:40,598][starVLA.training.train_starvla][INFO] - Step 3187, Loss: {'train/loss': 8.07607557362644e-06, 'train/grad_norm_pre_clip': 0.005635980516672134, 'timing/data': 0.00030343000253196806, 'timing/model': 0.39063320599962026, 'learning_rate/qwen_vl_interface': 3.3635475194847927e-06, 'learning_rate/action_model': 3.363547519484793e-05, 'epoch': 0.23}) +[2026-10-03 13:39:41,296][starVLA.training.train_starvla][INFO] - Step 3188, Loss: {'train/loss': 1.2963962490175618e-06, 'train/grad_norm_pre_clip': 0.000563626701477915, 'timing/data': 0.0001931000006152317, 'timing/model': 0.39046915600192733, 'learning_rate/qwen_vl_interface': 3.3607529577781433e-06, 'learning_rate/action_model': 3.360752957778143e-05, 'epoch': 0.23}) +[2026-10-03 13:39:42,182][starVLA.training.train_starvla][INFO] - Step 3189, Loss: {'train/loss': 1.8998927089342033e-06, 'train/grad_norm_pre_clip': 0.0005746897659264505, 'timing/data': 0.0002785499964375049, 'timing/model': 0.577899735988467, 'learning_rate/qwen_vl_interface': 3.3579591726695906e-06, 'learning_rate/action_model': 3.35795917266959e-05, 'epoch': 0.23}) +[2026-10-03 13:39:42,880][starVLA.training.train_starvla][INFO] - Step 3190, Loss: {'train/loss': 0.531839907169342, 'train/grad_norm_pre_clip': 458.1065673828125, 'timing/data': 0.00031921999470796436, 'timing/model': 0.3907457259920193, 'learning_rate/qwen_vl_interface': 3.3551661653075522e-06, 'learning_rate/action_model': 3.355166165307552e-05, 'epoch': 0.23}) +[2026-10-03 13:39:43,578][starVLA.training.train_starvla][INFO] - Step 3191, Loss: {'train/loss': 2.3584216251038015e-05, 'train/grad_norm_pre_clip': 0.016987185925245285, 'timing/data': 0.00024071999359875917, 'timing/model': 0.3901745770126581, 'learning_rate/qwen_vl_interface': 3.35237393684013e-06, 'learning_rate/action_model': 3.35237393684013e-05, 'epoch': 0.23}) +[2026-10-03 13:39:44,276][starVLA.training.train_starvla][INFO] - Step 3192, Loss: {'train/loss': 0.0032856157049536705, 'train/grad_norm_pre_clip': 25.687894821166992, 'timing/data': 0.00034627001150511205, 'timing/model': 0.3898996859934414, 'learning_rate/qwen_vl_interface': 3.349582488415103e-06, 'learning_rate/action_model': 3.3495824884151026e-05, 'epoch': 0.23}) +[2026-10-03 13:39:44,974][starVLA.training.train_starvla][INFO] - Step 3193, Loss: {'train/loss': 1.2591455060828594e-06, 'train/grad_norm_pre_clip': 0.0003363514260854572, 'timing/data': 0.00023084999702405185, 'timing/model': 0.39033761600148864, 'learning_rate/qwen_vl_interface': 3.3467918211799294e-06, 'learning_rate/action_model': 3.3467918211799295e-05, 'epoch': 0.23}) +[2026-10-03 13:39:45,862][starVLA.training.train_starvla][INFO] - Step 3194, Loss: {'train/loss': 9.76025035015482e-07, 'train/grad_norm_pre_clip': 0.0002087102475343272, 'timing/data': 0.0002615100092953071, 'timing/model': 0.5784825259906938, 'learning_rate/qwen_vl_interface': 3.3440019362817476e-06, 'learning_rate/action_model': 3.344001936281747e-05, 'epoch': 0.23}) +[2026-10-03 13:39:46,559][starVLA.training.train_starvla][INFO] - Step 3195, Loss: {'train/loss': 2.6747395622805925e-06, 'train/grad_norm_pre_clip': 0.0007784321205690503, 'timing/data': 0.0003009899955941364, 'timing/model': 0.3907585859997198, 'learning_rate/qwen_vl_interface': 3.3412128348673745e-06, 'learning_rate/action_model': 3.341212834867374e-05, 'epoch': 0.23}) +[2026-10-03 13:39:47,258][starVLA.training.train_starvla][INFO] - Step 3196, Loss: {'train/loss': 9.834750471782172e-07, 'train/grad_norm_pre_clip': 0.00027507590129971504, 'timing/data': 0.00029320000612642616, 'timing/model': 0.3902401759987697, 'learning_rate/qwen_vl_interface': 3.338424518083303e-06, 'learning_rate/action_model': 3.338424518083303e-05, 'epoch': 0.23}) +[2026-10-03 13:39:47,953][starVLA.training.train_starvla][INFO] - Step 3197, Loss: {'train/loss': 0.7166443467140198, 'train/grad_norm_pre_clip': 35.004371643066406, 'timing/data': 0.00023335000150837004, 'timing/model': 0.3885183770034928, 'learning_rate/qwen_vl_interface': 3.3356369870757056e-06, 'learning_rate/action_model': 3.335636987075705e-05, 'epoch': 0.23}) +[2026-10-03 13:39:48,649][starVLA.training.train_starvla][INFO] - Step 3198, Loss: {'train/loss': 9.164201060229971e-07, 'train/grad_norm_pre_clip': 0.0002271642442792654, 'timing/data': 0.00020416000916156918, 'timing/model': 0.389052297003218, 'learning_rate/qwen_vl_interface': 3.3328502429904327e-06, 'learning_rate/action_model': 3.3328502429904325e-05, 'epoch': 0.23}) +[2026-10-03 13:39:49,535][starVLA.training.train_starvla][INFO] - Step 3199, Loss: {'train/loss': 5.677209628629498e-06, 'train/grad_norm_pre_clip': 0.004824080970138311, 'timing/data': 0.0002946700114989653, 'timing/model': 0.580499215007876, 'learning_rate/qwen_vl_interface': 3.3300642869730066e-06, 'learning_rate/action_model': 3.330064286973006e-05, 'epoch': 0.23}) +[2026-10-03 13:39:50,230][starVLA.training.train_starvla][INFO] - Step 3200, Loss: {'train/loss': 0.6631847023963928, 'train/grad_norm_pre_clip': 125.82442474365234, 'timing/data': 0.00020739999308716506, 'timing/model': 0.3883364770008484, 'learning_rate/qwen_vl_interface': 3.3272791201686294e-06, 'learning_rate/action_model': 3.327279120168629e-05, 'epoch': 0.23}) +[2026-10-03 13:39:50,925][starVLA.training.train_starvla][INFO] - Step 3201, Loss: {'train/loss': 1.0787816790980287e-05, 'train/grad_norm_pre_clip': 0.004374940879642963, 'timing/data': 0.0003366699966136366, 'timing/model': 0.3884350970038213, 'learning_rate/qwen_vl_interface': 3.324494743722179e-06, 'learning_rate/action_model': 3.3244947437221785e-05, 'epoch': 0.23}) +[2026-10-03 13:39:51,620][starVLA.training.train_starvla][INFO] - Step 3202, Loss: {'train/loss': 0.29371312260627747, 'train/grad_norm_pre_clip': 158.07688903808594, 'timing/data': 0.00031833999673835933, 'timing/model': 0.38887064599839505, 'learning_rate/qwen_vl_interface': 3.321711158778205e-06, 'learning_rate/action_model': 3.321711158778205e-05, 'epoch': 0.23}) +[2026-10-03 13:39:52,315][starVLA.training.train_starvla][INFO] - Step 3203, Loss: {'train/loss': 2.510434751457069e-05, 'train/grad_norm_pre_clip': 0.027535755187273026, 'timing/data': 0.0003445799957262352, 'timing/model': 0.38808760700339917, 'learning_rate/qwen_vl_interface': 3.3189283664809393e-06, 'learning_rate/action_model': 3.318928366480939e-05, 'epoch': 0.23}) +[2026-10-03 13:39:53,010][starVLA.training.train_starvla][INFO] - Step 3204, Loss: {'train/loss': 2.0786967525054934e-06, 'train/grad_norm_pre_clip': 0.0008155895047821105, 'timing/data': 0.000211880003917031, 'timing/model': 0.3884320659999503, 'learning_rate/qwen_vl_interface': 3.316146367974276e-06, 'learning_rate/action_model': 3.316146367974276e-05, 'epoch': 0.23}) +[2026-10-03 13:39:53,902][starVLA.training.train_starvla][INFO] - Step 3205, Loss: {'train/loss': 1.8856242604670115e-05, 'train/grad_norm_pre_clip': 0.006107796914875507, 'timing/data': 0.00033024999720510095, 'timing/model': 0.38860819600813556, 'learning_rate/qwen_vl_interface': 3.3133651644017945e-06, 'learning_rate/action_model': 3.3133651644017944e-05, 'epoch': 0.23}) +[2026-10-03 13:39:54,597][starVLA.training.train_starvla][INFO] - Step 3206, Loss: {'train/loss': 0.6319699883460999, 'train/grad_norm_pre_clip': 26.99652099609375, 'timing/data': 0.00019620999228209257, 'timing/model': 0.38870093699370045, 'learning_rate/qwen_vl_interface': 3.31058475690674e-06, 'learning_rate/action_model': 3.31058475690674e-05, 'epoch': 0.23}) +[2026-10-03 13:39:55,292][starVLA.training.train_starvla][INFO] - Step 3207, Loss: {'train/loss': 1.1562940017029177e-05, 'train/grad_norm_pre_clip': 0.0035557106602936983, 'timing/data': 0.00019919000624213368, 'timing/model': 0.3893583960016258, 'learning_rate/qwen_vl_interface': 3.3078051466320337e-06, 'learning_rate/action_model': 3.307805146632034e-05, 'epoch': 0.23}) +[2026-10-03 13:39:55,988][starVLA.training.train_starvla][INFO] - Step 3208, Loss: {'train/loss': 0.27622053027153015, 'train/grad_norm_pre_clip': 258.1841735839844, 'timing/data': 0.00027303000388201326, 'timing/model': 0.3885298780078301, 'learning_rate/qwen_vl_interface': 3.3050263347202712e-06, 'learning_rate/action_model': 3.305026334720271e-05, 'epoch': 0.23}) +[2026-10-03 13:39:56,682][starVLA.training.train_starvla][INFO] - Step 3209, Loss: {'train/loss': 1.344730299024377e-05, 'train/grad_norm_pre_clip': 0.00970419030636549, 'timing/data': 0.00033241999335587025, 'timing/model': 0.3878047859907383, 'learning_rate/qwen_vl_interface': 3.302248322313712e-06, 'learning_rate/action_model': 3.302248322313712e-05, 'epoch': 0.23}) +[2026-10-03 13:39:57,565][starVLA.training.train_starvla][INFO] - Step 3210, Loss: {'train/loss': 4.4107150642958e-06, 'train/grad_norm_pre_clip': 0.0009652599110268056, 'timing/data': 0.00019238999811932445, 'timing/model': 0.5774039960087975, 'learning_rate/qwen_vl_interface': 3.299471110554295e-06, 'learning_rate/action_model': 3.299471110554295e-05, 'epoch': 0.23}) +[2026-10-03 13:39:58,259][starVLA.training.train_starvla][INFO] - Step 3211, Loss: {'train/loss': 4.246069147484377e-05, 'train/grad_norm_pre_clip': 0.018460946157574654, 'timing/data': 0.00019875999714713544, 'timing/model': 0.38877707699430175, 'learning_rate/qwen_vl_interface': 3.296694700583626e-06, 'learning_rate/action_model': 3.296694700583626e-05, 'epoch': 0.24}) +[2026-10-03 13:39:58,953][starVLA.training.train_starvla][INFO] - Step 3212, Loss: {'train/loss': 0.5635993480682373, 'train/grad_norm_pre_clip': 424.2119140625, 'timing/data': 0.0002033199998550117, 'timing/model': 0.38809480700001586, 'learning_rate/qwen_vl_interface': 3.2939190935429833e-06, 'learning_rate/action_model': 3.293919093542983e-05, 'epoch': 0.24}) +[2026-10-03 13:39:59,648][starVLA.training.train_starvla][INFO] - Step 3213, Loss: {'train/loss': 4.947132310917368e-06, 'train/grad_norm_pre_clip': 0.0015201467322185636, 'timing/data': 0.00029178999830037355, 'timing/model': 0.3885312159982277, 'learning_rate/qwen_vl_interface': 3.2911442905733126e-06, 'learning_rate/action_model': 3.291144290573312e-05, 'epoch': 0.24}) +[2026-10-03 13:40:00,343][starVLA.training.train_starvla][INFO] - Step 3214, Loss: {'train/loss': 1.0549612852628343e-05, 'train/grad_norm_pre_clip': 0.0032617233227938414, 'timing/data': 0.00023061000683810562, 'timing/model': 0.3885258869995596, 'learning_rate/qwen_vl_interface': 3.2883702928152323e-06, 'learning_rate/action_model': 3.288370292815232e-05, 'epoch': 0.24}) +[2026-10-03 13:40:01,036][starVLA.training.train_starvla][INFO] - Step 3215, Loss: {'train/loss': 1.3858052625437267e-06, 'train/grad_norm_pre_clip': 0.00031079386826604605, 'timing/data': 0.00019424001220613718, 'timing/model': 0.3882512570125982, 'learning_rate/qwen_vl_interface': 3.2855971014090275e-06, 'learning_rate/action_model': 3.285597101409028e-05, 'epoch': 0.24}) +[2026-10-03 13:40:01,921][starVLA.training.train_starvla][INFO] - Step 3216, Loss: {'train/loss': 8.567954864702187e-06, 'train/grad_norm_pre_clip': 0.004025437403470278, 'timing/data': 0.0001994700141949579, 'timing/model': 0.388588975998573, 'learning_rate/qwen_vl_interface': 3.2828247174946528e-06, 'learning_rate/action_model': 3.2828247174946524e-05, 'epoch': 0.24}) +[2026-10-03 13:40:02,616][starVLA.training.train_starvla][INFO] - Step 3217, Loss: {'train/loss': 2.0041961761307903e-06, 'train/grad_norm_pre_clip': 0.000561519933398813, 'timing/data': 0.0002653900010045618, 'timing/model': 0.38797343798796646, 'learning_rate/qwen_vl_interface': 3.280053142211729e-06, 'learning_rate/action_model': 3.280053142211729e-05, 'epoch': 0.24}) +[2026-10-03 13:40:03,309][starVLA.training.train_starvla][INFO] - Step 3218, Loss: {'train/loss': 3.660623406176455e-05, 'train/grad_norm_pre_clip': 0.021467246115207672, 'timing/data': 0.00022509999689646065, 'timing/model': 0.3871177359978901, 'learning_rate/qwen_vl_interface': 3.2772823766995475e-06, 'learning_rate/action_model': 3.277282376699548e-05, 'epoch': 0.24}) +[2026-10-03 13:40:04,003][starVLA.training.train_starvla][INFO] - Step 3219, Loss: {'train/loss': 3.12788033625111e-05, 'train/grad_norm_pre_clip': 0.062123481184244156, 'timing/data': 0.0003020900039700791, 'timing/model': 0.38859696699364576, 'learning_rate/qwen_vl_interface': 3.2745124220970646e-06, 'learning_rate/action_model': 3.274512422097065e-05, 'epoch': 0.24}) +[2026-10-03 13:40:04,698][starVLA.training.train_starvla][INFO] - Step 3220, Loss: {'train/loss': 4.433052708918694e-06, 'train/grad_norm_pre_clip': 0.001144133391790092, 'timing/data': 0.000202669994905591, 'timing/model': 0.38856993700028397, 'learning_rate/qwen_vl_interface': 3.2717432795429057e-06, 'learning_rate/action_model': 3.271743279542905e-05, 'epoch': 0.24}) +[2026-10-03 13:40:05,584][starVLA.training.train_starvla][INFO] - Step 3221, Loss: {'train/loss': 5.816660268465057e-05, 'train/grad_norm_pre_clip': 0.060758285224437714, 'timing/data': 0.00020146901078987867, 'timing/model': 0.38836061699839775, 'learning_rate/qwen_vl_interface': 3.2689749501753578e-06, 'learning_rate/action_model': 3.2689749501753574e-05, 'epoch': 0.24}) +[2026-10-03 13:40:06,278][starVLA.training.train_starvla][INFO] - Step 3222, Loss: {'train/loss': 6.77250091030146e-06, 'train/grad_norm_pre_clip': 0.0020486165303736925, 'timing/data': 0.00032540998654440045, 'timing/model': 0.38765583699569106, 'learning_rate/qwen_vl_interface': 3.2662074351323774e-06, 'learning_rate/action_model': 3.266207435132378e-05, 'epoch': 0.24}) +[2026-10-03 13:40:06,972][starVLA.training.train_starvla][INFO] - Step 3223, Loss: {'train/loss': 0.23803158104419708, 'train/grad_norm_pre_clip': 463.8139343261719, 'timing/data': 0.00033541000448167324, 'timing/model': 0.388510827004211, 'learning_rate/qwen_vl_interface': 3.2634407355515875e-06, 'learning_rate/action_model': 3.2634407355515875e-05, 'epoch': 0.24}) +[2026-10-03 13:40:07,667][starVLA.training.train_starvla][INFO] - Step 3224, Loss: {'train/loss': 2.491801933501847e-05, 'train/grad_norm_pre_clip': 0.007211968302726746, 'timing/data': 0.00030590000096708536, 'timing/model': 0.3884469869954046, 'learning_rate/qwen_vl_interface': 3.2606748525702715e-06, 'learning_rate/action_model': 3.260674852570271e-05, 'epoch': 0.24}) +[2026-10-03 13:40:08,361][starVLA.training.train_starvla][INFO] - Step 3225, Loss: {'train/loss': 2.8302005375735462e-05, 'train/grad_norm_pre_clip': 0.013230782002210617, 'timing/data': 0.0003133799909846857, 'timing/model': 0.3881568159995368, 'learning_rate/qwen_vl_interface': 3.2579097873253803e-06, 'learning_rate/action_model': 3.25790978732538e-05, 'epoch': 0.24}) +[2026-10-03 13:40:09,246][starVLA.training.train_starvla][INFO] - Step 3226, Loss: {'train/loss': 0.008313058875501156, 'train/grad_norm_pre_clip': 12.370270729064941, 'timing/data': 0.00024869000480975956, 'timing/model': 0.5802419950050535, 'learning_rate/qwen_vl_interface': 3.2551455409535274e-06, 'learning_rate/action_model': 3.255145540953528e-05, 'epoch': 0.24}) +[2026-10-03 13:40:09,941][starVLA.training.train_starvla][INFO] - Step 3227, Loss: {'train/loss': 0.0030577436555176973, 'train/grad_norm_pre_clip': 2.37933087348938, 'timing/data': 0.000248630007263273, 'timing/model': 0.38884815700293984, 'learning_rate/qwen_vl_interface': 3.2523821145909907e-06, 'learning_rate/action_model': 3.252382114590991e-05, 'epoch': 0.24}) +[2026-10-03 13:40:10,635][starVLA.training.train_starvla][INFO] - Step 3228, Loss: {'train/loss': 3.204724634997547e-05, 'train/grad_norm_pre_clip': 0.00983165018260479, 'timing/data': 0.00029232000815682113, 'timing/model': 0.38772752600198146, 'learning_rate/qwen_vl_interface': 3.249619509373712e-06, 'learning_rate/action_model': 3.249619509373712e-05, 'epoch': 0.24}) +[2026-10-03 13:40:11,329][starVLA.training.train_starvla][INFO] - Step 3229, Loss: {'train/loss': 0.2644316256046295, 'train/grad_norm_pre_clip': 57.12314987182617, 'timing/data': 0.0003227599954698235, 'timing/model': 0.38824316699174233, 'learning_rate/qwen_vl_interface': 3.2468577264372905e-06, 'learning_rate/action_model': 3.24685772643729e-05, 'epoch': 0.24}) +[2026-10-03 13:40:12,024][starVLA.training.train_starvla][INFO] - Step 3230, Loss: {'train/loss': 1.100921869277954, 'train/grad_norm_pre_clip': 107.35562896728516, 'timing/data': 0.0002495399967301637, 'timing/model': 0.38888269699236844, 'learning_rate/qwen_vl_interface': 3.244096766916992e-06, 'learning_rate/action_model': 3.244096766916992e-05, 'epoch': 0.24}) +[2026-10-03 13:40:12,720][starVLA.training.train_starvla][INFO] - Step 3231, Loss: {'train/loss': 3.645036485977471e-05, 'train/grad_norm_pre_clip': 0.007575578056275845, 'timing/data': 0.00032357999589294195, 'timing/model': 0.3887742470105877, 'learning_rate/qwen_vl_interface': 3.241336631947743e-06, 'learning_rate/action_model': 3.2413366319477427e-05, 'epoch': 0.24}) +[2026-10-03 13:40:13,600][starVLA.training.train_starvla][INFO] - Step 3232, Loss: {'train/loss': 5.594587855739519e-05, 'train/grad_norm_pre_clip': 0.020421411842107773, 'timing/data': 0.00019739000708796084, 'timing/model': 0.38879857599386014, 'learning_rate/qwen_vl_interface': 3.238577322664131e-06, 'learning_rate/action_model': 3.238577322664131e-05, 'epoch': 0.24}) +[2026-10-03 13:40:14,294][starVLA.training.train_starvla][INFO] - Step 3233, Loss: {'train/loss': 0.00013018521713092923, 'train/grad_norm_pre_clip': 0.05187435448169708, 'timing/data': 0.00025101999926846474, 'timing/model': 0.38806132799072657, 'learning_rate/qwen_vl_interface': 3.2358188402004026e-06, 'learning_rate/action_model': 3.2358188402004024e-05, 'epoch': 0.24}) +[2026-10-03 13:40:14,990][starVLA.training.train_starvla][INFO] - Step 3234, Loss: {'train/loss': 0.00016738369595259428, 'train/grad_norm_pre_clip': 0.2559733986854553, 'timing/data': 0.00024651999410707504, 'timing/model': 0.38865924600395374, 'learning_rate/qwen_vl_interface': 3.2330611856904675e-06, 'learning_rate/action_model': 3.233061185690468e-05, 'epoch': 0.24}) +[2026-10-03 13:40:15,684][starVLA.training.train_starvla][INFO] - Step 3235, Loss: {'train/loss': 0.0001968878204934299, 'train/grad_norm_pre_clip': 0.7405672669410706, 'timing/data': 0.0003030099906027317, 'timing/model': 0.38819869700819254, 'learning_rate/qwen_vl_interface': 3.2303043602678928e-06, 'learning_rate/action_model': 3.230304360267893e-05, 'epoch': 0.24}) +[2026-10-03 13:40:16,378][starVLA.training.train_starvla][INFO] - Step 3236, Loss: {'train/loss': 0.04010703042149544, 'train/grad_norm_pre_clip': 79.21953582763672, 'timing/data': 0.00021618000755552202, 'timing/model': 0.38836516698938794, 'learning_rate/qwen_vl_interface': 3.227548365065905e-06, 'learning_rate/action_model': 3.227548365065905e-05, 'epoch': 0.24}) +[2026-10-03 13:40:17,264][starVLA.training.train_starvla][INFO] - Step 3237, Loss: {'train/loss': 0.00023691462411079556, 'train/grad_norm_pre_clip': 0.09349151700735092, 'timing/data': 0.00033429999893996865, 'timing/model': 0.38875942600134294, 'learning_rate/qwen_vl_interface': 3.2247932012173887e-06, 'learning_rate/action_model': 3.2247932012173884e-05, 'epoch': 0.24}) +[2026-10-03 13:40:17,959][starVLA.training.train_starvla][INFO] - Step 3238, Loss: {'train/loss': 5.479743049363606e-05, 'train/grad_norm_pre_clip': 0.012308266013860703, 'timing/data': 0.0003226000117138028, 'timing/model': 0.38906228699488565, 'learning_rate/qwen_vl_interface': 3.2220388698548897e-06, 'learning_rate/action_model': 3.22203886985489e-05, 'epoch': 0.24}) +[2026-10-03 13:40:18,654][starVLA.training.train_starvla][INFO] - Step 3239, Loss: {'train/loss': 1.1987593097728677e-05, 'train/grad_norm_pre_clip': 0.0036190389655530453, 'timing/data': 0.00024280000070575625, 'timing/model': 0.3882831060036551, 'learning_rate/qwen_vl_interface': 3.21928537211061e-06, 'learning_rate/action_model': 3.21928537211061e-05, 'epoch': 0.24}) +[2026-10-03 13:40:19,349][starVLA.training.train_starvla][INFO] - Step 3240, Loss: {'train/loss': 0.003815081436187029, 'train/grad_norm_pre_clip': 2.671470880508423, 'timing/data': 0.00024543999461457133, 'timing/model': 0.38889178801036905, 'learning_rate/qwen_vl_interface': 3.216532709116405e-06, 'learning_rate/action_model': 3.216532709116405e-05, 'epoch': 0.24}) +[2026-10-03 13:40:20,044][starVLA.training.train_starvla][INFO] - Step 3241, Loss: {'train/loss': 0.43415382504463196, 'train/grad_norm_pre_clip': 267.436767578125, 'timing/data': 0.0001963399990927428, 'timing/model': 0.38895583599514794, 'learning_rate/qwen_vl_interface': 3.2137808820037925e-06, 'learning_rate/action_model': 3.2137808820037924e-05, 'epoch': 0.24}) +[2026-10-03 13:40:20,930][starVLA.training.train_starvla][INFO] - Step 3242, Loss: {'train/loss': 7.606963663420174e-06, 'train/grad_norm_pre_clip': 0.001798194949515164, 'timing/data': 0.00033071999496314675, 'timing/model': 0.3897237469936954, 'learning_rate/qwen_vl_interface': 3.211029891903945e-06, 'learning_rate/action_model': 3.211029891903945e-05, 'epoch': 0.24}) +[2026-10-03 13:40:21,626][starVLA.training.train_starvla][INFO] - Step 3243, Loss: {'train/loss': 8.210369742300827e-06, 'train/grad_norm_pre_clip': 0.0013317647390067577, 'timing/data': 0.00032716999703552574, 'timing/model': 0.38936589600052685, 'learning_rate/qwen_vl_interface': 3.2082797399476894e-06, 'learning_rate/action_model': 3.2082797399476895e-05, 'epoch': 0.24}) +[2026-10-03 13:40:22,321][starVLA.training.train_starvla][INFO] - Step 3244, Loss: {'train/loss': 0.0017407594714313745, 'train/grad_norm_pre_clip': 3.076969861984253, 'timing/data': 0.00024080000002868474, 'timing/model': 0.3887320270005148, 'learning_rate/qwen_vl_interface': 3.2055304272655114e-06, 'learning_rate/action_model': 3.205530427265511e-05, 'epoch': 0.24}) +[2026-10-03 13:40:23,017][starVLA.training.train_starvla][INFO] - Step 3245, Loss: {'train/loss': 1.45205958688166e-05, 'train/grad_norm_pre_clip': 0.008017976768314838, 'timing/data': 0.0003511300019454211, 'timing/model': 0.3885925360082183, 'learning_rate/qwen_vl_interface': 3.2027819549875477e-06, 'learning_rate/action_model': 3.202781954987548e-05, 'epoch': 0.24}) +[2026-10-03 13:40:23,713][starVLA.training.train_starvla][INFO] - Step 3246, Loss: {'train/loss': 7.02712350175716e-05, 'train/grad_norm_pre_clip': 0.019293611869215965, 'timing/data': 0.00019907999376300722, 'timing/model': 0.3897453170066001, 'learning_rate/qwen_vl_interface': 3.2000343242435926e-06, 'learning_rate/action_model': 3.2000343242435924e-05, 'epoch': 0.24}) +[2026-10-03 13:40:24,598][starVLA.training.train_starvla][INFO] - Step 3247, Loss: {'train/loss': 0.0007400306058116257, 'train/grad_norm_pre_clip': 0.5478565096855164, 'timing/data': 0.00024916000256780535, 'timing/model': 0.5789913150074426, 'learning_rate/qwen_vl_interface': 3.1972875361630923e-06, 'learning_rate/action_model': 3.1972875361630923e-05, 'epoch': 0.24}) +[2026-10-03 13:40:25,292][starVLA.training.train_starvla][INFO] - Step 3248, Loss: {'train/loss': 0.0006344994180835783, 'train/grad_norm_pre_clip': 0.6165914535522461, 'timing/data': 0.00020796900207642466, 'timing/model': 0.3886608280008659, 'learning_rate/qwen_vl_interface': 3.1945415918751484e-06, 'learning_rate/action_model': 3.194541591875148e-05, 'epoch': 0.24}) +[2026-10-03 13:40:25,987][starVLA.training.train_starvla][INFO] - Step 3249, Loss: {'train/loss': 0.02124645933508873, 'train/grad_norm_pre_clip': 54.009422302246094, 'timing/data': 0.00031850000959821045, 'timing/model': 0.38876574599999003, 'learning_rate/qwen_vl_interface': 3.191796492508516e-06, 'learning_rate/action_model': 3.191796492508516e-05, 'epoch': 0.24}) +[2026-10-03 13:40:26,683][starVLA.training.train_starvla][INFO] - Step 3250, Loss: {'train/loss': 2.40777571889339e-05, 'train/grad_norm_pre_clip': 0.0112552335485816, 'timing/data': 0.0003201500076102093, 'timing/model': 0.3892607269954169, 'learning_rate/qwen_vl_interface': 3.189052239191599e-06, 'learning_rate/action_model': 3.189052239191599e-05, 'epoch': 0.24}) +[2026-10-03 13:40:27,378][starVLA.training.train_starvla][INFO] - Step 3251, Loss: {'train/loss': 5.5580621847184375e-06, 'train/grad_norm_pre_clip': 0.0015785022405907512, 'timing/data': 0.0002414900081930682, 'timing/model': 0.3887608660006663, 'learning_rate/qwen_vl_interface': 3.1863088330524577e-06, 'learning_rate/action_model': 3.186308833052458e-05, 'epoch': 0.24}) +[2026-10-03 13:40:28,263][starVLA.training.train_starvla][INFO] - Step 3252, Loss: {'train/loss': 0.000828671851195395, 'train/grad_norm_pre_clip': 0.49188557267189026, 'timing/data': 0.00025117000041063875, 'timing/model': 0.5782802159956191, 'learning_rate/qwen_vl_interface': 3.183566275218803e-06, 'learning_rate/action_model': 3.1835662752188026e-05, 'epoch': 0.24}) +[2026-10-03 13:40:28,958][starVLA.training.train_starvla][INFO] - Step 3253, Loss: {'train/loss': 5.371764473238727e-06, 'train/grad_norm_pre_clip': 0.0028428451623767614, 'timing/data': 0.0002470299950800836, 'timing/model': 0.38869350698951166, 'learning_rate/qwen_vl_interface': 3.180824566817996e-06, 'learning_rate/action_model': 3.1808245668179956e-05, 'epoch': 0.24}) +[2026-10-03 13:40:29,653][starVLA.training.train_starvla][INFO] - Step 3254, Loss: {'train/loss': 0.000924629857763648, 'train/grad_norm_pre_clip': 2.401310682296753, 'timing/data': 0.00029205999453552067, 'timing/model': 0.388462207003613, 'learning_rate/qwen_vl_interface': 3.1780837089770524e-06, 'learning_rate/action_model': 3.1780837089770524e-05, 'epoch': 0.24}) +[2026-10-03 13:40:30,349][starVLA.training.train_starvla][INFO] - Step 3255, Loss: {'train/loss': 0.03163965046405792, 'train/grad_norm_pre_clip': 28.819034576416016, 'timing/data': 0.00030375999631360173, 'timing/model': 0.3890045059961267, 'learning_rate/qwen_vl_interface': 3.17534370282263e-06, 'learning_rate/action_model': 3.1753437028226303e-05, 'epoch': 0.24}) +[2026-10-03 13:40:31,045][starVLA.training.train_starvla][INFO] - Step 3256, Loss: {'train/loss': 2.5704380277602468e-06, 'train/grad_norm_pre_clip': 0.0008407188579440117, 'timing/data': 0.00019916999735869467, 'timing/model': 0.38887057700776495, 'learning_rate/qwen_vl_interface': 3.172604549481046e-06, 'learning_rate/action_model': 3.1726045494810455e-05, 'epoch': 0.24}) +[2026-10-03 13:40:31,741][starVLA.training.train_starvla][INFO] - Step 3257, Loss: {'train/loss': 0.00040272960904985666, 'train/grad_norm_pre_clip': 1.392539143562317, 'timing/data': 0.00030556999263353646, 'timing/model': 0.3886907260020962, 'learning_rate/qwen_vl_interface': 3.169866250078261e-06, 'learning_rate/action_model': 3.169866250078261e-05, 'epoch': 0.24}) +[2026-10-03 13:40:32,633][starVLA.training.train_starvla][INFO] - Step 3258, Loss: {'train/loss': 0.7850803136825562, 'train/grad_norm_pre_clip': 64.80877685546875, 'timing/data': 0.00020364001102279872, 'timing/model': 0.38840606599114835, 'learning_rate/qwen_vl_interface': 3.167128805739887e-06, 'learning_rate/action_model': 3.1671288057398866e-05, 'epoch': 0.24}) +[2026-10-03 13:40:33,328][starVLA.training.train_starvla][INFO] - Step 3259, Loss: {'train/loss': 2.4063021555775777e-05, 'train/grad_norm_pre_clip': 0.005459498148411512, 'timing/data': 0.000281530010397546, 'timing/model': 0.38869274700118694, 'learning_rate/qwen_vl_interface': 3.1643922175911847e-06, 'learning_rate/action_model': 3.164392217591185e-05, 'epoch': 0.24}) +[2026-10-03 13:40:34,024][starVLA.training.train_starvla][INFO] - Step 3260, Loss: {'train/loss': 3.042342359549366e-05, 'train/grad_norm_pre_clip': 0.02261352725327015, 'timing/data': 0.0003100100002484396, 'timing/model': 0.3892476770124631, 'learning_rate/qwen_vl_interface': 3.1616564867570603e-06, 'learning_rate/action_model': 3.16165648675706e-05, 'epoch': 0.24}) +[2026-10-03 13:40:34,719][starVLA.training.train_starvla][INFO] - Step 3261, Loss: {'train/loss': 0.36097559332847595, 'train/grad_norm_pre_clip': 36.18474578857422, 'timing/data': 0.0003290200111223385, 'timing/model': 0.38875333601026796, 'learning_rate/qwen_vl_interface': 3.1589216143620733e-06, 'learning_rate/action_model': 3.1589216143620734e-05, 'epoch': 0.24}) +[2026-10-03 13:40:35,415][starVLA.training.train_starvla][INFO] - Step 3262, Loss: {'train/loss': 0.00016599567607045174, 'train/grad_norm_pre_clip': 0.22572767734527588, 'timing/data': 0.00035009998828172684, 'timing/model': 0.38885737699456513, 'learning_rate/qwen_vl_interface': 3.1561876015304215e-06, 'learning_rate/action_model': 3.156187601530421e-05, 'epoch': 0.24}) +[2026-10-03 13:40:36,299][starVLA.training.train_starvla][INFO] - Step 3263, Loss: {'train/loss': 5.1969283958896995e-05, 'train/grad_norm_pre_clip': 0.017349371686577797, 'timing/data': 0.00025379999715369195, 'timing/model': 0.5781655159953516, 'learning_rate/qwen_vl_interface': 3.1534544493859563e-06, 'learning_rate/action_model': 3.1534544493859566e-05, 'epoch': 0.24}) +[2026-10-03 13:40:36,995][starVLA.training.train_starvla][INFO] - Step 3264, Loss: {'train/loss': 0.39668893814086914, 'train/grad_norm_pre_clip': 56.91200256347656, 'timing/data': 0.0002000399981625378, 'timing/model': 0.3888745259901043, 'learning_rate/qwen_vl_interface': 3.150722159052174e-06, 'learning_rate/action_model': 3.150722159052174e-05, 'epoch': 0.24}) +[2026-10-03 13:40:37,690][starVLA.training.train_starvla][INFO] - Step 3265, Loss: {'train/loss': 0.0009214591700583696, 'train/grad_norm_pre_clip': 2.3007266521453857, 'timing/data': 0.0001963399990927428, 'timing/model': 0.3887741369981086, 'learning_rate/qwen_vl_interface': 3.147990731652215e-06, 'learning_rate/action_model': 3.147990731652215e-05, 'epoch': 0.24}) +[2026-10-03 13:40:38,386][starVLA.training.train_starvla][INFO] - Step 3266, Loss: {'train/loss': 0.04045438393950462, 'train/grad_norm_pre_clip': 200.34910583496094, 'timing/data': 0.00030914999661035836, 'timing/model': 0.38951030699536204, 'learning_rate/qwen_vl_interface': 3.1452601683088673e-06, 'learning_rate/action_model': 3.145260168308867e-05, 'epoch': 0.24}) +[2026-10-03 13:40:39,081][starVLA.training.train_starvla][INFO] - Step 3267, Loss: {'train/loss': 3.165256202919409e-05, 'train/grad_norm_pre_clip': 0.03228361904621124, 'timing/data': 0.00026529999740887433, 'timing/model': 0.3888208979915362, 'learning_rate/qwen_vl_interface': 3.1425304701445608e-06, 'learning_rate/action_model': 3.1425304701445605e-05, 'epoch': 0.24}) +[2026-10-03 13:40:39,776][starVLA.training.train_starvla][INFO] - Step 3268, Loss: {'train/loss': 0.0002891167241614312, 'train/grad_norm_pre_clip': 0.8020750880241394, 'timing/data': 0.0001946499978657812, 'timing/model': 0.3892609359900234, 'learning_rate/qwen_vl_interface': 3.139801638281372e-06, 'learning_rate/action_model': 3.139801638281372e-05, 'epoch': 0.24}) +[2026-10-03 13:40:40,659][starVLA.training.train_starvla][INFO] - Step 3269, Loss: {'train/loss': 6.556390417244984e-06, 'train/grad_norm_pre_clip': 0.0014413715107366443, 'timing/data': 0.00019885900837834924, 'timing/model': 0.38895784701162484, 'learning_rate/qwen_vl_interface': 3.13707367384102e-06, 'learning_rate/action_model': 3.13707367384102e-05, 'epoch': 0.24}) +[2026-10-03 13:40:41,354][starVLA.training.train_starvla][INFO] - Step 3270, Loss: {'train/loss': 3.874276444548741e-06, 'train/grad_norm_pre_clip': 0.0008239540620706975, 'timing/data': 0.000311160009005107, 'timing/model': 0.3890333970048232, 'learning_rate/qwen_vl_interface': 3.134346577944869e-06, 'learning_rate/action_model': 3.134346577944869e-05, 'epoch': 0.24}) +[2026-10-03 13:40:42,050][starVLA.training.train_starvla][INFO] - Step 3271, Loss: {'train/loss': 1.6391238659707597e-06, 'train/grad_norm_pre_clip': 0.001491972478106618, 'timing/data': 0.000292459997581318, 'timing/model': 0.38930987600178923, 'learning_rate/qwen_vl_interface': 3.1316203517139243e-06, 'learning_rate/action_model': 3.131620351713924e-05, 'epoch': 0.24}) +[2026-10-03 13:40:42,745][starVLA.training.train_starvla][INFO] - Step 3272, Loss: {'train/loss': 0.01545715518295765, 'train/grad_norm_pre_clip': 31.72858428955078, 'timing/data': 0.00033363999682478607, 'timing/model': 0.38901594700291753, 'learning_rate/qwen_vl_interface': 3.128894996268837e-06, 'learning_rate/action_model': 3.128894996268837e-05, 'epoch': 0.24}) +[2026-10-03 13:40:43,440][starVLA.training.train_starvla][INFO] - Step 3273, Loss: {'train/loss': 0.4462033212184906, 'train/grad_norm_pre_clip': 96.1016845703125, 'timing/data': 0.000337520003085956, 'timing/model': 0.38858940699719824, 'learning_rate/qwen_vl_interface': 3.1261705127298934e-06, 'learning_rate/action_model': 3.1261705127298935e-05, 'epoch': 0.24}) +[2026-10-03 13:40:44,325][starVLA.training.train_starvla][INFO] - Step 3274, Loss: {'train/loss': 8.627535862615332e-05, 'train/grad_norm_pre_clip': 0.6134414076805115, 'timing/data': 0.00020268000662326813, 'timing/model': 0.3888666479906533, 'learning_rate/qwen_vl_interface': 3.123446902217028e-06, 'learning_rate/action_model': 3.1234469022170283e-05, 'epoch': 0.24}) +[2026-10-03 13:40:45,021][starVLA.training.train_starvla][INFO] - Step 3275, Loss: {'train/loss': 2.197902176703792e-06, 'train/grad_norm_pre_clip': 0.0007074174354784191, 'timing/data': 0.00019425000937189907, 'timing/model': 0.38980136600730475, 'learning_rate/qwen_vl_interface': 3.1207241658498145e-06, 'learning_rate/action_model': 3.1207241658498145e-05, 'epoch': 0.24}) +[2026-10-03 13:40:45,717][starVLA.training.train_starvla][INFO] - Step 3276, Loss: {'train/loss': 3.873452078551054e-05, 'train/grad_norm_pre_clip': 0.011844308115541935, 'timing/data': 0.00020126999879721552, 'timing/model': 0.3892299669969361, 'learning_rate/qwen_vl_interface': 3.1180023047474667e-06, 'learning_rate/action_model': 3.1180023047474664e-05, 'epoch': 0.24}) +[2026-10-03 13:40:46,412][starVLA.training.train_starvla][INFO] - Step 3277, Loss: {'train/loss': 0.0002517801185604185, 'train/grad_norm_pre_clip': 0.17857511341571808, 'timing/data': 0.00019295999663881958, 'timing/model': 0.38854387699393556, 'learning_rate/qwen_vl_interface': 3.1152813200288386e-06, 'learning_rate/action_model': 3.115281320028838e-05, 'epoch': 0.24}) +[2026-10-03 13:40:47,109][starVLA.training.train_starvla][INFO] - Step 3278, Loss: {'train/loss': 1.8074173567583784e-05, 'train/grad_norm_pre_clip': 0.006476504262536764, 'timing/data': 0.00031449001107830554, 'timing/model': 0.38906772699556313, 'learning_rate/qwen_vl_interface': 3.1125612128124244e-06, 'learning_rate/action_model': 3.112561212812424e-05, 'epoch': 0.24}) +[2026-10-03 13:40:47,995][starVLA.training.train_starvla][INFO] - Step 3279, Loss: {'train/loss': 0.17301194369792938, 'train/grad_norm_pre_clip': 451.40142822265625, 'timing/data': 0.0002544999879319221, 'timing/model': 0.5802878250106005, 'learning_rate/qwen_vl_interface': 3.1098419842163573e-06, 'learning_rate/action_model': 3.109841984216357e-05, 'epoch': 0.24}) +[2026-10-03 13:40:48,690][starVLA.training.train_starvla][INFO] - Step 3280, Loss: {'train/loss': 9.648087143432349e-06, 'train/grad_norm_pre_clip': 0.005636734887957573, 'timing/data': 0.00031724000291433185, 'timing/model': 0.3888282469997648, 'learning_rate/qwen_vl_interface': 3.10712363535841e-06, 'learning_rate/action_model': 3.10712363535841e-05, 'epoch': 0.24}) +[2026-10-03 13:40:49,384][starVLA.training.train_starvla][INFO] - Step 3281, Loss: {'train/loss': 1.1562307008716743e-05, 'train/grad_norm_pre_clip': 0.010027715004980564, 'timing/data': 0.00019347001216374338, 'timing/model': 0.38860943699546624, 'learning_rate/qwen_vl_interface': 3.1044061673559926e-06, 'learning_rate/action_model': 3.1044061673559925e-05, 'epoch': 0.24}) +[2026-10-03 13:40:50,080][starVLA.training.train_starvla][INFO] - Step 3282, Loss: {'train/loss': 3.524103249219479e-06, 'train/grad_norm_pre_clip': 0.0007447390235029161, 'timing/data': 0.00024061001022346318, 'timing/model': 0.389286125995568, 'learning_rate/qwen_vl_interface': 3.1016895813261556e-06, 'learning_rate/action_model': 3.101689581326155e-05, 'epoch': 0.24}) +[2026-10-03 13:40:50,774][starVLA.training.train_starvla][INFO] - Step 3283, Loss: {'train/loss': 0.00033826479921117425, 'train/grad_norm_pre_clip': 0.10732749104499817, 'timing/data': 0.00021034899691585451, 'timing/model': 0.3885712170012994, 'learning_rate/qwen_vl_interface': 3.098973878385581e-06, 'learning_rate/action_model': 3.098973878385581e-05, 'epoch': 0.24}) +[2026-10-03 13:40:51,470][starVLA.training.train_starvla][INFO] - Step 3284, Loss: {'train/loss': 0.30846574902534485, 'train/grad_norm_pre_clip': 384.11126708984375, 'timing/data': 0.0003038200084120035, 'timing/model': 0.38858243699360173, 'learning_rate/qwen_vl_interface': 3.096259059650594e-06, 'learning_rate/action_model': 3.096259059650594e-05, 'epoch': 0.24}) +[2026-10-03 13:40:52,352][starVLA.training.train_starvla][INFO] - Step 3285, Loss: {'train/loss': 4.470280146051664e-06, 'train/grad_norm_pre_clip': 0.0013103893725201488, 'timing/data': 0.0002005199930863455, 'timing/model': 0.38832302700029686, 'learning_rate/qwen_vl_interface': 3.093545126237153e-06, 'learning_rate/action_model': 3.0935451262371527e-05, 'epoch': 0.24}) +[2026-10-03 13:40:53,048][starVLA.training.train_starvla][INFO] - Step 3286, Loss: {'train/loss': 4.9216901970794424e-05, 'train/grad_norm_pre_clip': 0.017034271731972694, 'timing/data': 0.0002450500032864511, 'timing/model': 0.38905643700854853, 'learning_rate/qwen_vl_interface': 3.090832079260855e-06, 'learning_rate/action_model': 3.0908320792608544e-05, 'epoch': 0.24}) +[2026-10-03 13:40:53,743][starVLA.training.train_starvla][INFO] - Step 3287, Loss: {'train/loss': 0.7452077865600586, 'train/grad_norm_pre_clip': 524.2167358398438, 'timing/data': 0.00028627000574488193, 'timing/model': 0.3887149760121247, 'learning_rate/qwen_vl_interface': 3.088119919836929e-06, 'learning_rate/action_model': 3.088119919836929e-05, 'epoch': 0.24}) +[2026-10-03 13:40:54,438][starVLA.training.train_starvla][INFO] - Step 3288, Loss: {'train/loss': 5.832665556226857e-05, 'train/grad_norm_pre_clip': 0.020133502781391144, 'timing/data': 0.00019612000323832035, 'timing/model': 0.3893049470061669, 'learning_rate/qwen_vl_interface': 3.0854086490802433e-06, 'learning_rate/action_model': 3.085408649080243e-05, 'epoch': 0.24}) +[2026-10-03 13:40:55,134][starVLA.training.train_starvla][INFO] - Step 3289, Loss: {'train/loss': 2.376711108809104e-06, 'train/grad_norm_pre_clip': 0.00046428959467448294, 'timing/data': 0.00035558000672608614, 'timing/model': 0.38910111600125674, 'learning_rate/qwen_vl_interface': 3.082698268105299e-06, 'learning_rate/action_model': 3.082698268105299e-05, 'epoch': 0.24}) +[2026-10-03 13:40:56,020][starVLA.training.train_starvla][INFO] - Step 3290, Loss: {'train/loss': 2.6523925953370053e-06, 'train/grad_norm_pre_clip': 0.0009807535680010915, 'timing/data': 0.00019811000674962997, 'timing/model': 0.3888532869896153, 'learning_rate/qwen_vl_interface': 3.0799887780262295e-06, 'learning_rate/action_model': 3.0799887780262296e-05, 'epoch': 0.24}) +[2026-10-03 13:40:56,715][starVLA.training.train_starvla][INFO] - Step 3291, Loss: {'train/loss': 0.03800874203443527, 'train/grad_norm_pre_clip': 214.24807739257812, 'timing/data': 0.0001980499946512282, 'timing/model': 0.38906649699492846, 'learning_rate/qwen_vl_interface': 3.0772801799568054e-06, 'learning_rate/action_model': 3.0772801799568054e-05, 'epoch': 0.24}) +[2026-10-03 13:40:57,411][starVLA.training.train_starvla][INFO] - Step 3292, Loss: {'train/loss': 0.002456250600516796, 'train/grad_norm_pre_clip': 2.5222887992858887, 'timing/data': 0.00027614001010078937, 'timing/model': 0.38849113800097257, 'learning_rate/qwen_vl_interface': 3.07457247501043e-06, 'learning_rate/action_model': 3.0745724750104296e-05, 'epoch': 0.24}) +[2026-10-03 13:40:58,106][starVLA.training.train_starvla][INFO] - Step 3293, Loss: {'train/loss': 0.4214020073413849, 'train/grad_norm_pre_clip': 45.743900299072266, 'timing/data': 0.00030944999889470637, 'timing/model': 0.38847208600782324, 'learning_rate/qwen_vl_interface': 3.0718656643001353e-06, 'learning_rate/action_model': 3.071865664300135e-05, 'epoch': 0.24}) +[2026-10-03 13:40:58,802][starVLA.training.train_starvla][INFO] - Step 3294, Loss: {'train/loss': 2.6037227144115604e-05, 'train/grad_norm_pre_clip': 0.012695006094872952, 'timing/data': 0.0002245499927084893, 'timing/model': 0.3893348870042246, 'learning_rate/qwen_vl_interface': 3.069159748938591e-06, 'learning_rate/action_model': 3.069159748938591e-05, 'epoch': 0.24}) +[2026-10-03 13:40:59,687][starVLA.training.train_starvla][INFO] - Step 3295, Loss: {'train/loss': 8.553070983907674e-06, 'train/grad_norm_pre_clip': 0.0026948319282382727, 'timing/data': 0.0002343399974051863, 'timing/model': 0.3893409559968859, 'learning_rate/qwen_vl_interface': 3.066454730038096e-06, 'learning_rate/action_model': 3.0664547300380956e-05, 'epoch': 0.24}) +[2026-10-03 13:41:00,382][starVLA.training.train_starvla][INFO] - Step 3296, Loss: {'train/loss': 0.008954639546573162, 'train/grad_norm_pre_clip': 91.9032974243164, 'timing/data': 0.00023267998767551035, 'timing/model': 0.38879133699811064, 'learning_rate/qwen_vl_interface': 3.0637506087105816e-06, 'learning_rate/action_model': 3.0637506087105815e-05, 'epoch': 0.24}) +[2026-10-03 13:41:01,077][starVLA.training.train_starvla][INFO] - Step 3297, Loss: {'train/loss': 2.7567066354095004e-06, 'train/grad_norm_pre_clip': 0.0010287107434123755, 'timing/data': 0.0003068000078201294, 'timing/model': 0.3885554669977864, 'learning_rate/qwen_vl_interface': 3.0610473860676103e-06, 'learning_rate/action_model': 3.06104738606761e-05, 'epoch': 0.24}) +[2026-10-03 13:41:01,773][starVLA.training.train_starvla][INFO] - Step 3298, Loss: {'train/loss': 0.7465090751647949, 'train/grad_norm_pre_clip': 80.57170867919922, 'timing/data': 0.00032544998975936323, 'timing/model': 0.38899041700642556, 'learning_rate/qwen_vl_interface': 3.0583450632203748e-06, 'learning_rate/action_model': 3.058345063220375e-05, 'epoch': 0.24}) +[2026-10-03 13:41:02,468][starVLA.training.train_starvla][INFO] - Step 3299, Loss: {'train/loss': 1.876658461696934e-05, 'train/grad_norm_pre_clip': 0.006893550045788288, 'timing/data': 0.0002182000025641173, 'timing/model': 0.3885899759916356, 'learning_rate/qwen_vl_interface': 3.055643641279699e-06, 'learning_rate/action_model': 3.055643641279699e-05, 'epoch': 0.24}) +[2026-10-03 13:41:03,354][starVLA.training.train_starvla][INFO] - Step 3300, Loss: {'train/loss': 7.331276265176712e-06, 'train/grad_norm_pre_clip': 0.0022000051103532314, 'timing/data': 0.00020197000412736088, 'timing/model': 0.5794375359982951, 'learning_rate/qwen_vl_interface': 3.0529431213560334e-06, 'learning_rate/action_model': 3.0529431213560334e-05, 'epoch': 0.24}) +[2026-10-03 13:41:04,049][starVLA.training.train_starvla][INFO] - Step 3301, Loss: {'train/loss': 6.347827820718521e-06, 'train/grad_norm_pre_clip': 0.0014469402376562357, 'timing/data': 0.000307869995594956, 'timing/model': 0.388781535992166, 'learning_rate/qwen_vl_interface': 3.0502435045594596e-06, 'learning_rate/action_model': 3.0502435045594595e-05, 'epoch': 0.24}) +[2026-10-03 13:41:04,745][starVLA.training.train_starvla][INFO] - Step 3302, Loss: {'train/loss': 0.00010192902118433267, 'train/grad_norm_pre_clip': 0.030207360163331032, 'timing/data': 0.0003127800009679049, 'timing/model': 0.3892142469994724, 'learning_rate/qwen_vl_interface': 3.047544791999689e-06, 'learning_rate/action_model': 3.0475447919996892e-05, 'epoch': 0.24}) +[2026-10-03 13:41:05,440][starVLA.training.train_starvla][INFO] - Step 3303, Loss: {'train/loss': 0.01786446012556553, 'train/grad_norm_pre_clip': 45.29563903808594, 'timing/data': 0.00019566999981179833, 'timing/model': 0.38940412701049354, 'learning_rate/qwen_vl_interface': 3.044846984786061e-06, 'learning_rate/action_model': 3.044846984786061e-05, 'epoch': 0.24}) +[2026-10-03 13:41:06,136][starVLA.training.train_starvla][INFO] - Step 3304, Loss: {'train/loss': 0.0781300961971283, 'train/grad_norm_pre_clip': 158.6849822998047, 'timing/data': 0.00019619999511633068, 'timing/model': 0.3892032069998095, 'learning_rate/qwen_vl_interface': 3.0421500840275413e-06, 'learning_rate/action_model': 3.0421500840275413e-05, 'epoch': 0.24}) +[2026-10-03 13:41:07,022][starVLA.training.train_starvla][INFO] - Step 3305, Loss: {'train/loss': 0.00029400456696748734, 'train/grad_norm_pre_clip': 0.0655793622136116, 'timing/data': 0.00028750000637955964, 'timing/model': 0.579482184999506, 'learning_rate/qwen_vl_interface': 3.039454090832727e-06, 'learning_rate/action_model': 3.0394540908327268e-05, 'epoch': 0.24}) +[2026-10-03 13:41:07,722][starVLA.training.train_starvla][INFO] - Step 3306, Loss: {'train/loss': 2.6400444767205045e-05, 'train/grad_norm_pre_clip': 0.011667738668620586, 'timing/data': 0.0003548800013959408, 'timing/model': 0.389835485999356, 'learning_rate/qwen_vl_interface': 3.0367590063098333e-06, 'learning_rate/action_model': 3.0367590063098333e-05, 'epoch': 0.24}) +[2026-10-03 13:41:08,417][starVLA.training.train_starvla][INFO] - Step 3307, Loss: {'train/loss': 2.4712018785066903e-05, 'train/grad_norm_pre_clip': 0.008221499621868134, 'timing/data': 0.00019689000328071415, 'timing/model': 0.38916281699493993, 'learning_rate/qwen_vl_interface': 3.0340648315667115e-06, 'learning_rate/action_model': 3.0340648315667113e-05, 'epoch': 0.24}) +[2026-10-03 13:41:09,113][starVLA.training.train_starvla][INFO] - Step 3308, Loss: {'train/loss': 1.2814579349651467e-05, 'train/grad_norm_pre_clip': 0.012427475303411484, 'timing/data': 0.00034487000084482133, 'timing/model': 0.3891788170003565, 'learning_rate/qwen_vl_interface': 3.0313715677108336e-06, 'learning_rate/action_model': 3.0313715677108335e-05, 'epoch': 0.24}) +[2026-10-03 13:41:09,809][starVLA.training.train_starvla][INFO] - Step 3309, Loss: {'train/loss': 0.0003521092003211379, 'train/grad_norm_pre_clip': 0.14792948961257935, 'timing/data': 0.00019714998779818416, 'timing/model': 0.38882471599208657, 'learning_rate/qwen_vl_interface': 3.0286792158492996e-06, 'learning_rate/action_model': 3.0286792158492994e-05, 'epoch': 0.24}) +[2026-10-03 13:41:10,504][starVLA.training.train_starvla][INFO] - Step 3310, Loss: {'train/loss': 0.0001087108685169369, 'train/grad_norm_pre_clip': 0.03297553211450577, 'timing/data': 0.0002523599978303537, 'timing/model': 0.38880646700272337, 'learning_rate/qwen_vl_interface': 3.0259877770888334e-06, 'learning_rate/action_model': 3.0259877770888333e-05, 'epoch': 0.24}) +[2026-10-03 13:41:11,393][starVLA.training.train_starvla][INFO] - Step 3311, Loss: {'train/loss': 0.5460476875305176, 'train/grad_norm_pre_clip': 56.941959381103516, 'timing/data': 0.00033860999974422157, 'timing/model': 0.3883591359917773, 'learning_rate/qwen_vl_interface': 3.0232972525357836e-06, 'learning_rate/action_model': 3.0232972525357834e-05, 'epoch': 0.24}) +[2026-10-03 13:41:12,088][starVLA.training.train_starvla][INFO] - Step 3312, Loss: {'train/loss': 0.0014290307881310582, 'train/grad_norm_pre_clip': 12.611278533935547, 'timing/data': 0.00019548999262042344, 'timing/model': 0.3892703470046399, 'learning_rate/qwen_vl_interface': 3.020607643296123e-06, 'learning_rate/action_model': 3.0206076432961226e-05, 'epoch': 0.24}) +[2026-10-03 13:41:12,784][starVLA.training.train_starvla][INFO] - Step 3313, Loss: {'train/loss': 4.149653250351548e-05, 'train/grad_norm_pre_clip': 0.019984625279903412, 'timing/data': 0.00019678000535350293, 'timing/model': 0.38966565701412037, 'learning_rate/qwen_vl_interface': 3.017918950475449e-06, 'learning_rate/action_model': 3.017918950475449e-05, 'epoch': 0.24}) +[2026-10-03 13:41:13,479][starVLA.training.train_starvla][INFO] - Step 3314, Loss: {'train/loss': 8.344460184162017e-06, 'train/grad_norm_pre_clip': 0.0030755861662328243, 'timing/data': 0.00034502999915275723, 'timing/model': 0.38912865601014346, 'learning_rate/qwen_vl_interface': 3.0152311751789815e-06, 'learning_rate/action_model': 3.0152311751789814e-05, 'epoch': 0.24}) +[2026-10-03 13:41:14,175][starVLA.training.train_starvla][INFO] - Step 3315, Loss: {'train/loss': 8.806657569948584e-05, 'train/grad_norm_pre_clip': 0.02283487096428871, 'timing/data': 0.00022124999668449163, 'timing/model': 0.3889624869916588, 'learning_rate/qwen_vl_interface': 3.0125443185115655e-06, 'learning_rate/action_model': 3.0125443185115655e-05, 'epoch': 0.24}) +[2026-10-03 13:41:15,054][starVLA.training.train_starvla][INFO] - Step 3316, Loss: {'train/loss': 0.33940625190734863, 'train/grad_norm_pre_clip': 19.574260711669922, 'timing/data': 0.0002925200096797198, 'timing/model': 0.5727251659991452, 'learning_rate/qwen_vl_interface': 3.0098583815776615e-06, 'learning_rate/action_model': 3.0098583815776615e-05, 'epoch': 0.24}) +[2026-10-03 13:41:15,749][starVLA.training.train_starvla][INFO] - Step 3317, Loss: {'train/loss': 9.506804417469539e-06, 'train/grad_norm_pre_clip': 0.0026746902149170637, 'timing/data': 0.00034625999978743494, 'timing/model': 0.38858078699558973, 'learning_rate/qwen_vl_interface': 3.0071733654813595e-06, 'learning_rate/action_model': 3.0071733654813595e-05, 'epoch': 0.24}) +[2026-10-03 13:41:16,445][starVLA.training.train_starvla][INFO] - Step 3318, Loss: {'train/loss': 0.18155942857265472, 'train/grad_norm_pre_clip': 68.0337142944336, 'timing/data': 0.00033583999902475625, 'timing/model': 0.38893918599933386, 'learning_rate/qwen_vl_interface': 3.0044892713263685e-06, 'learning_rate/action_model': 3.0044892713263682e-05, 'epoch': 0.24}) +[2026-10-03 13:41:17,140][starVLA.training.train_starvla][INFO] - Step 3319, Loss: {'train/loss': 3.9504393498646095e-05, 'train/grad_norm_pre_clip': 0.019087890163064003, 'timing/data': 0.00019426998915150762, 'timing/model': 0.38958013799856417, 'learning_rate/qwen_vl_interface': 3.001806100216017e-06, 'learning_rate/action_model': 3.001806100216017e-05, 'epoch': 0.24}) +[2026-10-03 13:41:17,835][starVLA.training.train_starvla][INFO] - Step 3320, Loss: {'train/loss': 0.005725635681301355, 'train/grad_norm_pre_clip': 3.9646284580230713, 'timing/data': 0.00019746999896597117, 'timing/model': 0.38882924600329716, 'learning_rate/qwen_vl_interface': 2.9991238532532567e-06, 'learning_rate/action_model': 2.9991238532532567e-05, 'epoch': 0.24}) +[2026-10-03 13:41:18,530][starVLA.training.train_starvla][INFO] - Step 3321, Loss: {'train/loss': 0.0007197708473540843, 'train/grad_norm_pre_clip': 0.2920178771018982, 'timing/data': 0.00022088999685365707, 'timing/model': 0.3891435569967143, 'learning_rate/qwen_vl_interface': 2.996442531540657e-06, 'learning_rate/action_model': 2.9964425315406568e-05, 'epoch': 0.24}) +[2026-10-03 13:41:19,414][starVLA.training.train_starvla][INFO] - Step 3322, Loss: {'train/loss': 0.00022391515085473657, 'train/grad_norm_pre_clip': 0.06414026021957397, 'timing/data': 0.00019817000429611653, 'timing/model': 0.3892645769956289, 'learning_rate/qwen_vl_interface': 2.993762136180408e-06, 'learning_rate/action_model': 2.993762136180408e-05, 'epoch': 0.24}) +[2026-10-03 13:41:20,109][starVLA.training.train_starvla][INFO] - Step 3323, Loss: {'train/loss': 0.28835001587867737, 'train/grad_norm_pre_clip': 992.2229614257812, 'timing/data': 0.0001981999957934022, 'timing/model': 0.3891927459917497, 'learning_rate/qwen_vl_interface': 2.9910826682743195e-06, 'learning_rate/action_model': 2.9910826682743192e-05, 'epoch': 0.24}) +[2026-10-03 13:41:20,804][starVLA.training.train_starvla][INFO] - Step 3324, Loss: {'train/loss': 0.0006093591800890863, 'train/grad_norm_pre_clip': 0.5689883232116699, 'timing/data': 0.0002532499929657206, 'timing/model': 0.3886392069980502, 'learning_rate/qwen_vl_interface': 2.9884041289238192e-06, 'learning_rate/action_model': 2.988404128923819e-05, 'epoch': 0.24}) +[2026-10-03 13:41:21,499][starVLA.training.train_starvla][INFO] - Step 3325, Loss: {'train/loss': 0.00025280355475842953, 'train/grad_norm_pre_clip': 0.06808146834373474, 'timing/data': 0.00019767999765463173, 'timing/model': 0.3888253769982839, 'learning_rate/qwen_vl_interface': 2.9857265192299557e-06, 'learning_rate/action_model': 2.9857265192299556e-05, 'epoch': 0.24}) +[2026-10-03 13:41:22,195][starVLA.training.train_starvla][INFO] - Step 3326, Loss: {'train/loss': 4.765370249515399e-05, 'train/grad_norm_pre_clip': 0.01904706470668316, 'timing/data': 0.00035168000613339245, 'timing/model': 0.38915275600447785, 'learning_rate/qwen_vl_interface': 2.9830498402933894e-06, 'learning_rate/action_model': 2.9830498402933894e-05, 'epoch': 0.24}) +[2026-10-03 13:41:23,079][starVLA.training.train_starvla][INFO] - Step 3327, Loss: {'train/loss': 1.716553197184112e-05, 'train/grad_norm_pre_clip': 0.005700208712369204, 'timing/data': 0.00019879000319633633, 'timing/model': 0.3886538460064912, 'learning_rate/qwen_vl_interface': 2.9803740932144025e-06, 'learning_rate/action_model': 2.9803740932144024e-05, 'epoch': 0.24}) +[2026-10-03 13:41:23,775][starVLA.training.train_starvla][INFO] - Step 3328, Loss: {'train/loss': 5.5655441428825725e-06, 'train/grad_norm_pre_clip': 0.001197870122268796, 'timing/data': 0.00019847000658046454, 'timing/model': 0.3894055870041484, 'learning_rate/qwen_vl_interface': 2.9776992790928953e-06, 'learning_rate/action_model': 2.9776992790928952e-05, 'epoch': 0.24}) +[2026-10-03 13:41:24,470][starVLA.training.train_starvla][INFO] - Step 3329, Loss: {'train/loss': 1.724730827845633e-05, 'train/grad_norm_pre_clip': 0.006983416620641947, 'timing/data': 0.00024163999478332698, 'timing/model': 0.3885929469979601, 'learning_rate/qwen_vl_interface': 2.975025399028382e-06, 'learning_rate/action_model': 2.975025399028382e-05, 'epoch': 0.24}) +[2026-10-03 13:41:25,167][starVLA.training.train_starvla][INFO] - Step 3330, Loss: {'train/loss': 1.555633934913203e-05, 'train/grad_norm_pre_clip': 0.006951596587896347, 'timing/data': 0.0001986499992199242, 'timing/model': 0.38961715600453317, 'learning_rate/qwen_vl_interface': 2.9723524541199943e-06, 'learning_rate/action_model': 2.972352454119994e-05, 'epoch': 0.24}) +[2026-10-03 13:41:25,862][starVLA.training.train_starvla][INFO] - Step 3331, Loss: {'train/loss': 0.00013471556303557009, 'train/grad_norm_pre_clip': 0.03885835036635399, 'timing/data': 0.0001933899911819026, 'timing/model': 0.3887210969987791, 'learning_rate/qwen_vl_interface': 2.969680445466478e-06, 'learning_rate/action_model': 2.969680445466478e-05, 'epoch': 0.24}) +[2026-10-03 13:41:26,750][starVLA.training.train_starvla][INFO] - Step 3332, Loss: {'train/loss': 9.729040175443515e-05, 'train/grad_norm_pre_clip': 0.027408599853515625, 'timing/data': 0.0002580900036264211, 'timing/model': 0.5813106450077612, 'learning_rate/qwen_vl_interface': 2.9670093741661955e-06, 'learning_rate/action_model': 2.9670093741661954e-05, 'epoch': 0.24}) +[2026-10-03 13:41:27,445][starVLA.training.train_starvla][INFO] - Step 3333, Loss: {'train/loss': 0.00021386196021921933, 'train/grad_norm_pre_clip': 0.17631648480892181, 'timing/data': 0.000304640008835122, 'timing/model': 0.388770546996966, 'learning_rate/qwen_vl_interface': 2.964339241317123e-06, 'learning_rate/action_model': 2.964339241317123e-05, 'epoch': 0.24}) +[2026-10-03 13:41:28,141][starVLA.training.train_starvla][INFO] - Step 3334, Loss: {'train/loss': 5.657197834807448e-05, 'train/grad_norm_pre_clip': 0.032110389322042465, 'timing/data': 0.000371429996448569, 'timing/model': 0.38947556698985863, 'learning_rate/qwen_vl_interface': 2.961670048016852e-06, 'learning_rate/action_model': 2.9616700480168518e-05, 'epoch': 0.24}) +[2026-10-03 13:41:28,834][starVLA.training.train_starvla][INFO] - Step 3335, Loss: {'train/loss': 0.4755347967147827, 'train/grad_norm_pre_clip': 97.73873138427734, 'timing/data': 0.00031596999906469136, 'timing/model': 0.3889892669976689, 'learning_rate/qwen_vl_interface': 2.9590017953625854e-06, 'learning_rate/action_model': 2.9590017953625853e-05, 'epoch': 0.24}) +[2026-10-03 13:41:29,530][starVLA.training.train_starvla][INFO] - Step 3336, Loss: {'train/loss': 2.1329240553313866e-05, 'train/grad_norm_pre_clip': 0.012539512477815151, 'timing/data': 0.00024993000261019915, 'timing/model': 0.389064425995457, 'learning_rate/qwen_vl_interface': 2.9563344844511455e-06, 'learning_rate/action_model': 2.9563344844511455e-05, 'epoch': 0.24}) +[2026-10-03 13:41:30,224][starVLA.training.train_starvla][INFO] - Step 3337, Loss: {'train/loss': 1.1250021998421289e-05, 'train/grad_norm_pre_clip': 0.005284728016704321, 'timing/data': 0.0002546699979575351, 'timing/model': 0.38699453699518926, 'learning_rate/qwen_vl_interface': 2.9536681163789568e-06, 'learning_rate/action_model': 2.9536681163789564e-05, 'epoch': 0.24}) +[2026-10-03 13:41:31,105][starVLA.training.train_starvla][INFO] - Step 3338, Loss: {'train/loss': 0.0003027784696314484, 'train/grad_norm_pre_clip': 0.22396399080753326, 'timing/data': 0.0001995599886868149, 'timing/model': 0.38903187699906994, 'learning_rate/qwen_vl_interface': 2.9510026922420653e-06, 'learning_rate/action_model': 2.9510026922420653e-05, 'epoch': 0.24}) +[2026-10-03 13:41:31,800][starVLA.training.train_starvla][INFO] - Step 3339, Loss: {'train/loss': 5.423959009931423e-06, 'train/grad_norm_pre_clip': 0.0018100016750395298, 'timing/data': 0.00030973998946137726, 'timing/model': 0.38851061699097045, 'learning_rate/qwen_vl_interface': 2.9483382131361273e-06, 'learning_rate/action_model': 2.948338213136127e-05, 'epoch': 0.24}) +[2026-10-03 13:41:32,494][starVLA.training.train_starvla][INFO] - Step 3340, Loss: {'train/loss': 0.003501131432130933, 'train/grad_norm_pre_clip': 4.914242744445801, 'timing/data': 0.00033398000232409686, 'timing/model': 0.3886582270060899, 'learning_rate/qwen_vl_interface': 2.9456746801564072e-06, 'learning_rate/action_model': 2.945674680156407e-05, 'epoch': 0.24}) +[2026-10-03 13:41:33,188][starVLA.training.train_starvla][INFO] - Step 3341, Loss: {'train/loss': 0.0021978248842060566, 'train/grad_norm_pre_clip': 2.1448380947113037, 'timing/data': 0.0003618600021582097, 'timing/model': 0.3881767470011255, 'learning_rate/qwen_vl_interface': 2.9430120943977854e-06, 'learning_rate/action_model': 2.9430120943977853e-05, 'epoch': 0.24}) +[2026-10-03 13:41:33,883][starVLA.training.train_starvla][INFO] - Step 3342, Loss: {'train/loss': 0.024759754538536072, 'train/grad_norm_pre_clip': 37.95520782470703, 'timing/data': 0.00018875999376177788, 'timing/model': 0.3887249560066266, 'learning_rate/qwen_vl_interface': 2.940350456954748e-06, 'learning_rate/action_model': 2.940350456954748e-05, 'epoch': 0.24}) +[2026-10-03 13:41:34,766][starVLA.training.train_starvla][INFO] - Step 3343, Loss: {'train/loss': 4.485215868044179e-06, 'train/grad_norm_pre_clip': 0.0012066980125382543, 'timing/data': 0.0002735599991865456, 'timing/model': 0.3881277170003159, 'learning_rate/qwen_vl_interface': 2.9376897689213952e-06, 'learning_rate/action_model': 2.937689768921395e-05, 'epoch': 0.24}) +[2026-10-03 13:41:35,461][starVLA.training.train_starvla][INFO] - Step 3344, Loss: {'train/loss': 0.0003505547356326133, 'train/grad_norm_pre_clip': 0.40268030762672424, 'timing/data': 0.0002838800137396902, 'timing/model': 0.3884290670102928, 'learning_rate/qwen_vl_interface': 2.9350300313914364e-06, 'learning_rate/action_model': 2.9350300313914363e-05, 'epoch': 0.24}) +[2026-10-03 13:41:36,154][starVLA.training.train_starvla][INFO] - Step 3345, Loss: {'train/loss': 0.4530702829360962, 'train/grad_norm_pre_clip': 56.708702087402344, 'timing/data': 0.00019247000454925, 'timing/model': 0.38732271701155696, 'learning_rate/qwen_vl_interface': 2.9323712454581867e-06, 'learning_rate/action_model': 2.9323712454581866e-05, 'epoch': 0.24}) +[2026-10-03 13:41:36,849][starVLA.training.train_starvla][INFO] - Step 3346, Loss: {'train/loss': 0.00020707250223495066, 'train/grad_norm_pre_clip': 0.1187642440199852, 'timing/data': 0.0003826100000878796, 'timing/model': 0.38882900700264145, 'learning_rate/qwen_vl_interface': 2.929713412214572e-06, 'learning_rate/action_model': 2.929713412214572e-05, 'epoch': 0.24}) +[2026-10-03 13:41:37,544][starVLA.training.train_starvla][INFO] - Step 3347, Loss: {'train/loss': 7.100245511537651e-06, 'train/grad_norm_pre_clip': 0.003339093178510666, 'timing/data': 0.00028546999965328723, 'timing/model': 0.3886638160038274, 'learning_rate/qwen_vl_interface': 2.9270565327531294e-06, 'learning_rate/action_model': 2.927056532753129e-05, 'epoch': 0.24}) +[2026-10-03 13:41:38,431][starVLA.training.train_starvla][INFO] - Step 3348, Loss: {'train/loss': 4.026812530355528e-05, 'train/grad_norm_pre_clip': 0.01057493407279253, 'timing/data': 0.0002598999999463558, 'timing/model': 0.3884512869990431, 'learning_rate/qwen_vl_interface': 2.924400608166003e-06, 'learning_rate/action_model': 2.924400608166003e-05, 'epoch': 0.25}) +[2026-10-03 13:41:39,126][starVLA.training.train_starvla][INFO] - Step 3349, Loss: {'train/loss': 2.980220870085759e-06, 'train/grad_norm_pre_clip': 0.0007728260825388134, 'timing/data': 0.00033803000405896455, 'timing/model': 0.38828984599967953, 'learning_rate/qwen_vl_interface': 2.921745639544939e-06, 'learning_rate/action_model': 2.921745639544939e-05, 'epoch': 0.25}) +[2026-10-03 13:41:39,821][starVLA.training.train_starvla][INFO] - Step 3350, Loss: {'train/loss': 3.9061775169102475e-05, 'train/grad_norm_pre_clip': 0.018214451149106026, 'timing/data': 0.00019688000611495227, 'timing/model': 0.388940648001153, 'learning_rate/qwen_vl_interface': 2.919091627981297e-06, 'learning_rate/action_model': 2.919091627981297e-05, 'epoch': 0.25}) +[2026-10-03 13:41:40,516][starVLA.training.train_starvla][INFO] - Step 3351, Loss: {'train/loss': 0.0003347571473568678, 'train/grad_norm_pre_clip': 0.3102226257324219, 'timing/data': 0.00019140999938827008, 'timing/model': 0.3887236759910593, 'learning_rate/qwen_vl_interface': 2.9164385745660396e-06, 'learning_rate/action_model': 2.9164385745660394e-05, 'epoch': 0.25}) +[2026-10-03 13:41:41,211][starVLA.training.train_starvla][INFO] - Step 3352, Loss: {'train/loss': 3.6283804547565524e-06, 'train/grad_norm_pre_clip': 0.0040957024320960045, 'timing/data': 0.0001963499962585047, 'timing/model': 0.38883090700255707, 'learning_rate/qwen_vl_interface': 2.9137864803897385e-06, 'learning_rate/action_model': 2.9137864803897387e-05, 'epoch': 0.25}) +[2026-10-03 13:41:42,093][starVLA.training.train_starvla][INFO] - Step 3353, Loss: {'train/loss': 5.801154475193471e-05, 'train/grad_norm_pre_clip': 0.013595791533589363, 'timing/data': 0.00020454000332392752, 'timing/model': 0.5768422359979013, 'learning_rate/qwen_vl_interface': 2.9111353465425696e-06, 'learning_rate/action_model': 2.9111353465425694e-05, 'epoch': 0.25}) +[2026-10-03 13:41:42,786][starVLA.training.train_starvla][INFO] - Step 3354, Loss: {'train/loss': 0.04555739462375641, 'train/grad_norm_pre_clip': 379.0248107910156, 'timing/data': 0.0002428599982522428, 'timing/model': 0.38713389699114487, 'learning_rate/qwen_vl_interface': 2.908485174114312e-06, 'learning_rate/action_model': 2.908485174114312e-05, 'epoch': 0.25}) +[2026-10-03 13:41:43,479][starVLA.training.train_starvla][INFO] - Step 3355, Loss: {'train/loss': 9.871688234852627e-06, 'train/grad_norm_pre_clip': 0.00399212259799242, 'timing/data': 0.00020083000708837062, 'timing/model': 0.38833264699496794, 'learning_rate/qwen_vl_interface': 2.9058359641943523e-06, 'learning_rate/action_model': 2.9058359641943523e-05, 'epoch': 0.25}) +[2026-10-03 13:41:44,172][starVLA.training.train_starvla][INFO] - Step 3356, Loss: {'train/loss': 0.00011415449989726767, 'train/grad_norm_pre_clip': 0.12750102579593658, 'timing/data': 0.00019376999989617616, 'timing/model': 0.38880814699223265, 'learning_rate/qwen_vl_interface': 2.9031877178716823e-06, 'learning_rate/action_model': 2.903187717871682e-05, 'epoch': 0.25}) +[2026-10-03 13:41:44,867][starVLA.training.train_starvla][INFO] - Step 3357, Loss: {'train/loss': 6.273306553339353e-06, 'train/grad_norm_pre_clip': 0.001636514556594193, 'timing/data': 0.0003577600000426173, 'timing/model': 0.3885611860023346, 'learning_rate/qwen_vl_interface': 2.900540436234894e-06, 'learning_rate/action_model': 2.9005404362348937e-05, 'epoch': 0.25}) +[2026-10-03 13:41:45,754][starVLA.training.train_starvla][INFO] - Step 3358, Loss: {'train/loss': 1.1108382750535384e-05, 'train/grad_norm_pre_clip': 0.0051717860624194145, 'timing/data': 0.0004235700034769252, 'timing/model': 0.5803248860029271, 'learning_rate/qwen_vl_interface': 2.897894120372185e-06, 'learning_rate/action_model': 2.897894120372185e-05, 'epoch': 0.25}) +[2026-10-03 13:41:46,448][starVLA.training.train_starvla][INFO] - Step 3359, Loss: {'train/loss': 3.8966281863395125e-06, 'train/grad_norm_pre_clip': 0.0009479313739575446, 'timing/data': 0.0002875399950426072, 'timing/model': 0.3884949769999366, 'learning_rate/qwen_vl_interface': 2.8952487713713585e-06, 'learning_rate/action_model': 2.8952487713713582e-05, 'epoch': 0.25}) +[2026-10-03 13:41:47,143][starVLA.training.train_starvla][INFO] - Step 3360, Loss: {'train/loss': 0.09463184326887131, 'train/grad_norm_pre_clip': 164.23941040039062, 'timing/data': 0.00019576000340748578, 'timing/model': 0.38849676700192504, 'learning_rate/qwen_vl_interface': 2.8926043903198135e-06, 'learning_rate/action_model': 2.8926043903198135e-05, 'epoch': 0.25}) +[2026-10-03 13:41:47,837][starVLA.training.train_starvla][INFO] - Step 3361, Loss: {'train/loss': 2.840630622813478e-05, 'train/grad_norm_pre_clip': 0.011056660674512386, 'timing/data': 0.000317939993692562, 'timing/model': 0.38865274700219743, 'learning_rate/qwen_vl_interface': 2.8899609783045574e-06, 'learning_rate/action_model': 2.8899609783045573e-05, 'epoch': 0.25}) +[2026-10-03 13:41:48,536][starVLA.training.train_starvla][INFO] - Step 3362, Loss: {'train/loss': 6.876788575027604e-06, 'train/grad_norm_pre_clip': 0.002415397437289357, 'timing/data': 0.0002008400042541325, 'timing/model': 0.3895669469929999, 'learning_rate/qwen_vl_interface': 2.8873185364121965e-06, 'learning_rate/action_model': 2.8873185364121963e-05, 'epoch': 0.25}) +[2026-10-03 13:41:49,231][starVLA.training.train_starvla][INFO] - Step 3363, Loss: {'train/loss': 0.33476722240448, 'train/grad_norm_pre_clip': 45.88447570800781, 'timing/data': 0.0001944599935086444, 'timing/model': 0.38899838700308464, 'learning_rate/qwen_vl_interface': 2.8846770657289386e-06, 'learning_rate/action_model': 2.8846770657289386e-05, 'epoch': 0.25}) +[2026-10-03 13:41:50,121][starVLA.training.train_starvla][INFO] - Step 3364, Loss: {'train/loss': 8.493420864397194e-06, 'train/grad_norm_pre_clip': 0.0033968081697821617, 'timing/data': 0.0002979400014737621, 'timing/model': 0.3881269169942243, 'learning_rate/qwen_vl_interface': 2.882036567340593e-06, 'learning_rate/action_model': 2.8820365673405932e-05, 'epoch': 0.25}) +[2026-10-03 13:41:50,816][starVLA.training.train_starvla][INFO] - Step 3365, Loss: {'train/loss': 3.084518311879947e-06, 'train/grad_norm_pre_clip': 0.0013612157199531794, 'timing/data': 0.0001930599974002689, 'timing/model': 0.3885911569959717, 'learning_rate/qwen_vl_interface': 2.8793970423325684e-06, 'learning_rate/action_model': 2.8793970423325685e-05, 'epoch': 0.25}) +[2026-10-03 13:41:51,513][starVLA.training.train_starvla][INFO] - Step 3366, Loss: {'train/loss': 6.101680264691822e-05, 'train/grad_norm_pre_clip': 0.0171684380620718, 'timing/data': 0.00019687000894919038, 'timing/model': 0.390039286998217, 'learning_rate/qwen_vl_interface': 2.8767584917898756e-06, 'learning_rate/action_model': 2.8767584917898754e-05, 'epoch': 0.25}) +[2026-10-03 13:41:52,209][starVLA.training.train_starvla][INFO] - Step 3367, Loss: {'train/loss': 6.32543014944531e-06, 'train/grad_norm_pre_clip': 0.0029173363000154495, 'timing/data': 0.0003323800046928227, 'timing/model': 0.3886274760006927, 'learning_rate/qwen_vl_interface': 2.874120916797121e-06, 'learning_rate/action_model': 2.8741209167971207e-05, 'epoch': 0.25}) +[2026-10-03 13:41:52,904][starVLA.training.train_starvla][INFO] - Step 3368, Loss: {'train/loss': 6.191325155668892e-06, 'train/grad_norm_pre_clip': 0.00256749102845788, 'timing/data': 0.00029020999500062317, 'timing/model': 0.3885289979953086, 'learning_rate/qwen_vl_interface': 2.8714843184385124e-06, 'learning_rate/action_model': 2.871484318438512e-05, 'epoch': 0.25}) +[2026-10-03 13:41:53,787][starVLA.training.train_starvla][INFO] - Step 3369, Loss: {'train/loss': 4.276587787899189e-06, 'train/grad_norm_pre_clip': 0.0014325276715680957, 'timing/data': 0.00031632999889552593, 'timing/model': 0.5757148470001994, 'learning_rate/qwen_vl_interface': 2.8688486977978575e-06, 'learning_rate/action_model': 2.8688486977978572e-05, 'epoch': 0.25}) +[2026-10-03 13:41:54,483][starVLA.training.train_starvla][INFO] - Step 3370, Loss: {'train/loss': 5.465512367663905e-05, 'train/grad_norm_pre_clip': 0.01771414466202259, 'timing/data': 0.00027590899844653904, 'timing/model': 0.3894411170040257, 'learning_rate/qwen_vl_interface': 2.8662140559585584e-06, 'learning_rate/action_model': 2.866214055958558e-05, 'epoch': 0.25}) +[2026-10-03 13:41:55,178][starVLA.training.train_starvla][INFO] - Step 3371, Loss: {'train/loss': 2.847912401193753e-05, 'train/grad_norm_pre_clip': 0.007392294239252806, 'timing/data': 0.00020805999520234764, 'timing/model': 0.3887387869908707, 'learning_rate/qwen_vl_interface': 2.8635803940036162e-06, 'learning_rate/action_model': 2.863580394003616e-05, 'epoch': 0.25}) +[2026-10-03 13:41:55,875][starVLA.training.train_starvla][INFO] - Step 3372, Loss: {'train/loss': 8.582946975366212e-06, 'train/grad_norm_pre_clip': 0.004874010104686022, 'timing/data': 0.0003232199960621074, 'timing/model': 0.3892791069956729, 'learning_rate/qwen_vl_interface': 2.8609477130156317e-06, 'learning_rate/action_model': 2.8609477130156314e-05, 'epoch': 0.25}) +[2026-10-03 13:41:56,570][starVLA.training.train_starvla][INFO] - Step 3373, Loss: {'train/loss': 2.469484752509743e-05, 'train/grad_norm_pre_clip': 0.039576511830091476, 'timing/data': 0.0003206100082024932, 'timing/model': 0.3887269569968339, 'learning_rate/qwen_vl_interface': 2.8583160140767997e-06, 'learning_rate/action_model': 2.8583160140767994e-05, 'epoch': 0.25}) +[2026-10-03 13:41:57,266][starVLA.training.train_starvla][INFO] - Step 3374, Loss: {'train/loss': 1.8253834923598333e-06, 'train/grad_norm_pre_clip': 0.00044292956590652466, 'timing/data': 0.00033922999864444137, 'timing/model': 0.388809506999678, 'learning_rate/qwen_vl_interface': 2.8556852982689125e-06, 'learning_rate/action_model': 2.8556852982689126e-05, 'epoch': 0.25}) +[2026-10-03 13:41:58,149][starVLA.training.train_starvla][INFO] - Step 3375, Loss: {'train/loss': 0.00014050351455807686, 'train/grad_norm_pre_clip': 0.1465008407831192, 'timing/data': 0.0002039499959209934, 'timing/model': 0.3893464870052412, 'learning_rate/qwen_vl_interface': 2.8530555666733584e-06, 'learning_rate/action_model': 2.8530555666733583e-05, 'epoch': 0.25}) +[2026-10-03 13:41:58,845][starVLA.training.train_starvla][INFO] - Step 3376, Loss: {'train/loss': 0.07919038832187653, 'train/grad_norm_pre_clip': 54.070465087890625, 'timing/data': 0.0001982200046768412, 'timing/model': 0.3891872560052434, 'learning_rate/qwen_vl_interface': 2.85042682037112e-06, 'learning_rate/action_model': 2.85042682037112e-05, 'epoch': 0.25}) +[2026-10-03 13:41:59,539][starVLA.training.train_starvla][INFO] - Step 3377, Loss: {'train/loss': 5.2759889513254166e-05, 'train/grad_norm_pre_clip': 0.031095024198293686, 'timing/data': 0.00020013000175822526, 'timing/model': 0.3885931169934338, 'learning_rate/qwen_vl_interface': 2.8477990604427764e-06, 'learning_rate/action_model': 2.8477990604427763e-05, 'epoch': 0.25}) +[2026-10-03 13:42:00,235][starVLA.training.train_starvla][INFO] - Step 3378, Loss: {'train/loss': 0.01842116191983223, 'train/grad_norm_pre_clip': 119.89105224609375, 'timing/data': 0.0004018400068162009, 'timing/model': 0.3893007070000749, 'learning_rate/qwen_vl_interface': 2.8451722879684994e-06, 'learning_rate/action_model': 2.8451722879684994e-05, 'epoch': 0.25}) +[2026-10-03 13:42:00,931][starVLA.training.train_starvla][INFO] - Step 3379, Loss: {'train/loss': 3.06624460790772e-05, 'train/grad_norm_pre_clip': 0.007275515701621771, 'timing/data': 0.0002486500015947968, 'timing/model': 0.3889083960093558, 'learning_rate/qwen_vl_interface': 2.842546504028059e-06, 'learning_rate/action_model': 2.8425465040280586e-05, 'epoch': 0.25}) +[2026-10-03 13:42:01,816][starVLA.training.train_starvla][INFO] - Step 3380, Loss: {'train/loss': 0.00010690641647670418, 'train/grad_norm_pre_clip': 0.06674601137638092, 'timing/data': 0.00043820000428240746, 'timing/model': 0.3882562070066342, 'learning_rate/qwen_vl_interface': 2.83992170970081e-06, 'learning_rate/action_model': 2.8399217097008102e-05, 'epoch': 0.25}) +[2026-10-03 13:42:02,512][starVLA.training.train_starvla][INFO] - Step 3381, Loss: {'train/loss': 1.5793764760019258e-05, 'train/grad_norm_pre_clip': 0.03652437403798103, 'timing/data': 0.00024002000282052904, 'timing/model': 0.38851611700374633, 'learning_rate/qwen_vl_interface': 2.837297906065711e-06, 'learning_rate/action_model': 2.8372979060657107e-05, 'epoch': 0.25}) +[2026-10-03 13:42:03,211][starVLA.training.train_starvla][INFO] - Step 3382, Loss: {'train/loss': 0.49920642375946045, 'train/grad_norm_pre_clip': 64.22492218017578, 'timing/data': 0.00020436001068446785, 'timing/model': 0.389134887009277, 'learning_rate/qwen_vl_interface': 2.834675094201306e-06, 'learning_rate/action_model': 2.8346750942013063e-05, 'epoch': 0.25}) +[2026-10-03 13:42:03,906][starVLA.training.train_starvla][INFO] - Step 3383, Loss: {'train/loss': 7.078889029799029e-05, 'train/grad_norm_pre_clip': 0.028640564531087875, 'timing/data': 0.000342120009008795, 'timing/model': 0.38865714699204545, 'learning_rate/qwen_vl_interface': 2.8320532751857365e-06, 'learning_rate/action_model': 2.8320532751857366e-05, 'epoch': 0.25}) +[2026-10-03 13:42:04,602][starVLA.training.train_starvla][INFO] - Step 3384, Loss: {'train/loss': 0.06124601885676384, 'train/grad_norm_pre_clip': 122.66841888427734, 'timing/data': 0.0002994600072270259, 'timing/model': 0.38881490600761026, 'learning_rate/qwen_vl_interface': 2.829432450096731e-06, 'learning_rate/action_model': 2.8294324500967306e-05, 'epoch': 0.25}) +[2026-10-03 13:42:05,486][starVLA.training.train_starvla][INFO] - Step 3385, Loss: {'train/loss': 1.4177487173583359e-05, 'train/grad_norm_pre_clip': 0.005973936524242163, 'timing/data': 0.00019654999778140336, 'timing/model': 0.5780118959955871, 'learning_rate/qwen_vl_interface': 2.8268126200116118e-06, 'learning_rate/action_model': 2.8268126200116117e-05, 'epoch': 0.25}) +[2026-10-03 13:42:06,180][starVLA.training.train_starvla][INFO] - Step 3386, Loss: {'train/loss': 6.364748696796596e-05, 'train/grad_norm_pre_clip': 0.020847994834184647, 'timing/data': 0.00025411999376956373, 'timing/model': 0.38798754700110294, 'learning_rate/qwen_vl_interface': 2.824193786007294e-06, 'learning_rate/action_model': 2.8241937860072937e-05, 'epoch': 0.25}) +[2026-10-03 13:42:06,876][starVLA.training.train_starvla][INFO] - Step 3387, Loss: {'train/loss': 8.538183465134352e-06, 'train/grad_norm_pre_clip': 0.005315935704857111, 'timing/data': 0.0003470999945420772, 'timing/model': 0.3889720270090038, 'learning_rate/qwen_vl_interface': 2.821575949160279e-06, 'learning_rate/action_model': 2.8215759491602787e-05, 'epoch': 0.25}) +[2026-10-03 13:42:07,572][starVLA.training.train_starvla][INFO] - Step 3388, Loss: {'train/loss': 4.656583769246936e-06, 'train/grad_norm_pre_clip': 0.0009118902962654829, 'timing/data': 0.00019729998894035816, 'timing/model': 0.3893941959977383, 'learning_rate/qwen_vl_interface': 2.8189591105466625e-06, 'learning_rate/action_model': 2.8189591105466624e-05, 'epoch': 0.25}) +[2026-10-03 13:42:08,267][starVLA.training.train_starvla][INFO] - Step 3389, Loss: {'train/loss': 3.4868508009822108e-06, 'train/grad_norm_pre_clip': 0.0010402356274425983, 'timing/data': 0.00019721999706234783, 'timing/model': 0.3891475570126204, 'learning_rate/qwen_vl_interface': 2.816343271242126e-06, 'learning_rate/action_model': 2.816343271242126e-05, 'epoch': 0.25}) +[2026-10-03 13:42:08,964][starVLA.training.train_starvla][INFO] - Step 3390, Loss: {'train/loss': 5.7280209148302674e-05, 'train/grad_norm_pre_clip': 0.021353116258978844, 'timing/data': 0.0001990000018849969, 'timing/model': 0.38956448700628243, 'learning_rate/qwen_vl_interface': 2.8137284323219467e-06, 'learning_rate/action_model': 2.8137284323219464e-05, 'epoch': 0.25}) +[2026-10-03 13:42:09,846][starVLA.training.train_starvla][INFO] - Step 3391, Loss: {'train/loss': 0.00012136962322983891, 'train/grad_norm_pre_clip': 0.040806643664836884, 'timing/data': 0.000248180003836751, 'timing/model': 0.38913705699087586, 'learning_rate/qwen_vl_interface': 2.81111459486098e-06, 'learning_rate/action_model': 2.81111459486098e-05, 'epoch': 0.25}) +[2026-10-03 13:42:10,542][starVLA.training.train_starvla][INFO] - Step 3392, Loss: {'train/loss': 0.00024629916879348457, 'train/grad_norm_pre_clip': 0.6150878667831421, 'timing/data': 0.0002045299916062504, 'timing/model': 0.388859266997315, 'learning_rate/qwen_vl_interface': 2.8085017599336806e-06, 'learning_rate/action_model': 2.8085017599336806e-05, 'epoch': 0.25}) +[2026-10-03 13:42:11,237][starVLA.training.train_starvla][INFO] - Step 3393, Loss: {'train/loss': 2.5629888114053756e-06, 'train/grad_norm_pre_clip': 0.0008025583229027689, 'timing/data': 0.00028048999956808984, 'timing/model': 0.3889222160069039, 'learning_rate/qwen_vl_interface': 2.8058899286140828e-06, 'learning_rate/action_model': 2.805889928614083e-05, 'epoch': 0.25}) +[2026-10-03 13:42:11,932][starVLA.training.train_starvla][INFO] - Step 3394, Loss: {'train/loss': 9.947621583705768e-05, 'train/grad_norm_pre_clip': 0.18984538316726685, 'timing/data': 0.00019536000036168844, 'timing/model': 0.3891178470075829, 'learning_rate/qwen_vl_interface': 2.8032791019758127e-06, 'learning_rate/action_model': 2.8032791019758127e-05, 'epoch': 0.25}) +[2026-10-03 13:42:12,628][starVLA.training.train_starvla][INFO] - Step 3395, Loss: {'train/loss': 7.822216866770759e-05, 'train/grad_norm_pre_clip': 0.026179878041148186, 'timing/data': 0.00023797999892849475, 'timing/model': 0.3890442370029632, 'learning_rate/qwen_vl_interface': 2.800669281092084e-06, 'learning_rate/action_model': 2.8006692810920838e-05, 'epoch': 0.25}) +[2026-10-03 13:42:13,514][starVLA.training.train_starvla][INFO] - Step 3396, Loss: {'train/loss': 4.270621866453439e-05, 'train/grad_norm_pre_clip': 0.013623221777379513, 'timing/data': 0.00020013999892398715, 'timing/model': 0.3887092969962396, 'learning_rate/qwen_vl_interface': 2.7980604670356938e-06, 'learning_rate/action_model': 2.7980604670356937e-05, 'epoch': 0.25}) +[2026-10-03 13:42:14,208][starVLA.training.train_starvla][INFO] - Step 3397, Loss: {'train/loss': 0.2264367640018463, 'train/grad_norm_pre_clip': 409.3098449707031, 'timing/data': 0.00019427000370342284, 'timing/model': 0.3886417769972468, 'learning_rate/qwen_vl_interface': 2.795452660879027e-06, 'learning_rate/action_model': 2.7954526608790273e-05, 'epoch': 0.25}) +[2026-10-03 13:42:14,904][starVLA.training.train_starvla][INFO] - Step 3398, Loss: {'train/loss': 2.542523361626081e-05, 'train/grad_norm_pre_clip': 0.025215089321136475, 'timing/data': 0.00025769000058062375, 'timing/model': 0.38891328699537553, 'learning_rate/qwen_vl_interface': 2.7928458636940565e-06, 'learning_rate/action_model': 2.7928458636940564e-05, 'epoch': 0.25}) +[2026-10-03 13:42:15,599][starVLA.training.train_starvla][INFO] - Step 3399, Loss: {'train/loss': 5.47613399248803e-06, 'train/grad_norm_pre_clip': 0.0010354651603847742, 'timing/data': 0.00034295998921152204, 'timing/model': 0.38923206701292656, 'learning_rate/qwen_vl_interface': 2.790240076552336e-06, 'learning_rate/action_model': 2.7902400765523355e-05, 'epoch': 0.25}) +[2026-10-03 13:42:16,296][starVLA.training.train_starvla][INFO] - Step 3400, Loss: {'train/loss': 8.283193892566487e-05, 'train/grad_norm_pre_clip': 0.09319379180669785, 'timing/data': 0.0003085000062128529, 'timing/model': 0.38927715600584634, 'learning_rate/qwen_vl_interface': 2.787635300525006e-06, 'learning_rate/action_model': 2.787635300525006e-05, 'epoch': 0.25}) +[2026-10-03 13:42:17,176][starVLA.training.train_starvla][INFO] - Step 3401, Loss: {'train/loss': 9.071792737813666e-05, 'train/grad_norm_pre_clip': 0.04260648414492607, 'timing/data': 0.00019859000167343765, 'timing/model': 0.5745586859993637, 'learning_rate/qwen_vl_interface': 2.7850315366827933e-06, 'learning_rate/action_model': 2.7850315366827934e-05, 'epoch': 0.25}) +[2026-10-03 13:42:17,872][starVLA.training.train_starvla][INFO] - Step 3402, Loss: {'train/loss': 1.8477370531400084e-06, 'train/grad_norm_pre_clip': 0.0006114773568697274, 'timing/data': 0.0003526099899318069, 'timing/model': 0.38881989699439146, 'learning_rate/qwen_vl_interface': 2.7824287860960076e-06, 'learning_rate/action_model': 2.7824287860960075e-05, 'epoch': 0.25}) +[2026-10-03 13:42:18,567][starVLA.training.train_starvla][INFO] - Step 3403, Loss: {'train/loss': 8.464259008178487e-05, 'train/grad_norm_pre_clip': 0.04072146490216255, 'timing/data': 0.00023862000671215355, 'timing/model': 0.3889575560024241, 'learning_rate/qwen_vl_interface': 2.7798270498345377e-06, 'learning_rate/action_model': 2.7798270498345376e-05, 'epoch': 0.25}) +[2026-10-03 13:42:19,263][starVLA.training.train_starvla][INFO] - Step 3404, Loss: {'train/loss': 3.340806870255619e-05, 'train/grad_norm_pre_clip': 0.009954888373613358, 'timing/data': 0.0003041000018129125, 'timing/model': 0.3887279379996471, 'learning_rate/qwen_vl_interface': 2.777226328967861e-06, 'learning_rate/action_model': 2.7772263289678607e-05, 'epoch': 0.25}) +[2026-10-03 13:42:19,958][starVLA.training.train_starvla][INFO] - Step 3405, Loss: {'train/loss': 3.904071490978822e-06, 'train/grad_norm_pre_clip': 0.0013978466158732772, 'timing/data': 0.00040092899871524423, 'timing/model': 0.38885773699439596, 'learning_rate/qwen_vl_interface': 2.774626624565035e-06, 'learning_rate/action_model': 2.774626624565035e-05, 'epoch': 0.25}) +[2026-10-03 13:42:20,842][starVLA.training.train_starvla][INFO] - Step 3406, Loss: {'train/loss': 0.00010464661318110302, 'train/grad_norm_pre_clip': 0.1064925342798233, 'timing/data': 0.00028356000257190317, 'timing/model': 0.5767478159978054, 'learning_rate/qwen_vl_interface': 2.772027937694701e-06, 'learning_rate/action_model': 2.772027937694701e-05, 'epoch': 0.25}) +[2026-10-03 13:42:21,537][starVLA.training.train_starvla][INFO] - Step 3407, Loss: {'train/loss': 0.00010295797255821526, 'train/grad_norm_pre_clip': 0.10583220422267914, 'timing/data': 0.00025134001043625176, 'timing/model': 0.3888363160076551, 'learning_rate/qwen_vl_interface': 2.7694302694250806e-06, 'learning_rate/action_model': 2.7694302694250808e-05, 'epoch': 0.25}) +[2026-10-03 13:42:22,233][starVLA.training.train_starvla][INFO] - Step 3408, Loss: {'train/loss': 0.0013394528068602085, 'train/grad_norm_pre_clip': 8.276944160461426, 'timing/data': 0.0002905099972849712, 'timing/model': 0.388534717989387, 'learning_rate/qwen_vl_interface': 2.7668336208239775e-06, 'learning_rate/action_model': 2.7668336208239776e-05, 'epoch': 0.25}) +[2026-10-03 13:42:22,928][starVLA.training.train_starvla][INFO] - Step 3409, Loss: {'train/loss': 0.0028280611149966717, 'train/grad_norm_pre_clip': 2.6801304817199707, 'timing/data': 0.000192170002264902, 'timing/model': 0.38872812599583995, 'learning_rate/qwen_vl_interface': 2.7642379929587755e-06, 'learning_rate/action_model': 2.7642379929587754e-05, 'epoch': 0.25}) +[2026-10-03 13:42:23,624][starVLA.training.train_starvla][INFO] - Step 3410, Loss: {'train/loss': 3.5613634281617124e-06, 'train/grad_norm_pre_clip': 0.0006479246076196432, 'timing/data': 0.00029429000278469175, 'timing/model': 0.38902436700300314, 'learning_rate/qwen_vl_interface': 2.7616433868964387e-06, 'learning_rate/action_model': 2.7616433868964388e-05, 'epoch': 0.25}) +[2026-10-03 13:42:24,508][starVLA.training.train_starvla][INFO] - Step 3411, Loss: {'train/loss': 5.990196768834721e-06, 'train/grad_norm_pre_clip': 0.0016456227749586105, 'timing/data': 0.00019844999769702554, 'timing/model': 0.5773335059930105, 'learning_rate/qwen_vl_interface': 2.7590498037035122e-06, 'learning_rate/action_model': 2.7590498037035124e-05, 'epoch': 0.25}) +[2026-10-03 13:42:25,201][starVLA.training.train_starvla][INFO] - Step 3412, Loss: {'train/loss': 5.810454968013801e-05, 'train/grad_norm_pre_clip': 0.026807120069861412, 'timing/data': 0.00019310999778099358, 'timing/model': 0.38706484700378496, 'learning_rate/qwen_vl_interface': 2.7564572444461218e-06, 'learning_rate/action_model': 2.7564572444461217e-05, 'epoch': 0.25}) +[2026-10-03 13:42:25,895][starVLA.training.train_starvla][INFO] - Step 3413, Loss: {'train/loss': 2.913150865424541e-06, 'train/grad_norm_pre_clip': 0.0006231290171854198, 'timing/data': 0.0003726699942490086, 'timing/model': 0.38821243699931074, 'learning_rate/qwen_vl_interface': 2.7538657101899674e-06, 'learning_rate/action_model': 2.7538657101899673e-05, 'epoch': 0.25}) +[2026-10-03 13:42:26,591][starVLA.training.train_starvla][INFO] - Step 3414, Loss: {'train/loss': 0.7044335007667542, 'train/grad_norm_pre_clip': 215.51165771484375, 'timing/data': 0.00021456999820657074, 'timing/model': 0.3894584460067563, 'learning_rate/qwen_vl_interface': 2.751275202000332e-06, 'learning_rate/action_model': 2.7512752020003317e-05, 'epoch': 0.25}) +[2026-10-03 13:42:27,287][starVLA.training.train_starvla][INFO] - Step 3415, Loss: {'train/loss': 6.064676199457608e-06, 'train/grad_norm_pre_clip': 0.0014489003224298358, 'timing/data': 0.0003007299965247512, 'timing/model': 0.3891073669947218, 'learning_rate/qwen_vl_interface': 2.7486857209420763e-06, 'learning_rate/action_model': 2.7486857209420763e-05, 'epoch': 0.25}) +[2026-10-03 13:42:27,981][starVLA.training.train_starvla][INFO] - Step 3416, Loss: {'train/loss': 7.614853529958054e-05, 'train/grad_norm_pre_clip': 0.03477299213409424, 'timing/data': 0.00028762999863829464, 'timing/model': 0.3882632670138264, 'learning_rate/qwen_vl_interface': 2.7460972680796377e-06, 'learning_rate/action_model': 2.7460972680796375e-05, 'epoch': 0.25}) +[2026-10-03 13:42:28,870][starVLA.training.train_starvla][INFO] - Step 3417, Loss: {'train/loss': 0.00017930165631696582, 'train/grad_norm_pre_clip': 0.1253303587436676, 'timing/data': 0.0004042599903186783, 'timing/model': 0.3878471369971521, 'learning_rate/qwen_vl_interface': 2.7435098444770314e-06, 'learning_rate/action_model': 2.7435098444770315e-05, 'epoch': 0.25}) +[2026-10-03 13:42:29,565][starVLA.training.train_starvla][INFO] - Step 3418, Loss: {'train/loss': 1.2643018635571934e-05, 'train/grad_norm_pre_clip': 0.009594141505658627, 'timing/data': 0.00036518000706564635, 'timing/model': 0.3888089070096612, 'learning_rate/qwen_vl_interface': 2.7409234511978495e-06, 'learning_rate/action_model': 2.7409234511978492e-05, 'epoch': 0.25}) +[2026-10-03 13:42:30,259][starVLA.training.train_starvla][INFO] - Step 3419, Loss: {'train/loss': 0.021950537338852882, 'train/grad_norm_pre_clip': 32.701820373535156, 'timing/data': 0.00031516001035925, 'timing/model': 0.3885195559996646, 'learning_rate/qwen_vl_interface': 2.7383380893052606e-06, 'learning_rate/action_model': 2.7383380893052608e-05, 'epoch': 0.25}) +[2026-10-03 13:42:30,953][starVLA.training.train_starvla][INFO] - Step 3420, Loss: {'train/loss': 4.1723222921064007e-07, 'train/grad_norm_pre_clip': 0.00024015220697037876, 'timing/data': 0.00019285999587737024, 'timing/model': 0.3884203180059558, 'learning_rate/qwen_vl_interface': 2.73575375986201e-06, 'learning_rate/action_model': 2.7357537598620096e-05, 'epoch': 0.25}) +[2026-10-03 13:42:31,647][starVLA.training.train_starvla][INFO] - Step 3421, Loss: {'train/loss': 2.047302041319199e-05, 'train/grad_norm_pre_clip': 0.0057671405375003815, 'timing/data': 0.00022898000315763056, 'timing/model': 0.3882194559992058, 'learning_rate/qwen_vl_interface': 2.733170463930417e-06, 'learning_rate/action_model': 2.7331704639304167e-05, 'epoch': 0.25}) +[2026-10-03 13:42:32,528][starVLA.training.train_starvla][INFO] - Step 3422, Loss: {'train/loss': 0.00015449435159098357, 'train/grad_norm_pre_clip': 0.18865127861499786, 'timing/data': 0.00021440000273287296, 'timing/model': 0.5741676470061066, 'learning_rate/qwen_vl_interface': 2.7305882025723774e-06, 'learning_rate/action_model': 2.7305882025723774e-05, 'epoch': 0.25}) +[2026-10-03 13:42:33,223][starVLA.training.train_starvla][INFO] - Step 3423, Loss: {'train/loss': 3.6507290133158676e-06, 'train/grad_norm_pre_clip': 0.0020460179075598717, 'timing/data': 0.00019659999816212803, 'timing/model': 0.389632795995567, 'learning_rate/qwen_vl_interface': 2.7280069768493626e-06, 'learning_rate/action_model': 2.7280069768493628e-05, 'epoch': 0.25}) +[2026-10-03 13:42:33,918][starVLA.training.train_starvla][INFO] - Step 3424, Loss: {'train/loss': 6.330259202513844e-05, 'train/grad_norm_pre_clip': 0.022391699254512787, 'timing/data': 0.00035370000114198774, 'timing/model': 0.3882688770099776, 'learning_rate/qwen_vl_interface': 2.7254267878224144e-06, 'learning_rate/action_model': 2.7254267878224143e-05, 'epoch': 0.25}) +[2026-10-03 13:42:34,612][starVLA.training.train_starvla][INFO] - Step 3425, Loss: {'train/loss': 7.301558753169957e-07, 'train/grad_norm_pre_clip': 0.0001353885163553059, 'timing/data': 0.00039513000228907913, 'timing/model': 0.38819212700764183, 'learning_rate/qwen_vl_interface': 2.7228476365521524e-06, 'learning_rate/action_model': 2.722847636552152e-05, 'epoch': 0.25}) +[2026-10-03 13:42:35,307][starVLA.training.train_starvla][INFO] - Step 3426, Loss: {'train/loss': 2.566971306805499e-05, 'train/grad_norm_pre_clip': 0.026818405836820602, 'timing/data': 0.00036887999158352613, 'timing/model': 0.38855796700227074, 'learning_rate/qwen_vl_interface': 2.7202695240987674e-06, 'learning_rate/action_model': 2.7202695240987676e-05, 'epoch': 0.25}) +[2026-10-03 13:42:36,002][starVLA.training.train_starvla][INFO] - Step 3427, Loss: {'train/loss': 7.435170846292749e-05, 'train/grad_norm_pre_clip': 0.0261031873524189, 'timing/data': 0.0002548999909777194, 'timing/model': 0.3886612759961281, 'learning_rate/qwen_vl_interface': 2.717692451522025e-06, 'learning_rate/action_model': 2.7176924515220248e-05, 'epoch': 0.25}) +[2026-10-03 13:42:36,887][starVLA.training.train_starvla][INFO] - Step 3428, Loss: {'train/loss': 7.770286902086809e-05, 'train/grad_norm_pre_clip': 0.05739954486489296, 'timing/data': 0.00025273000937886536, 'timing/model': 0.38969778700266033, 'learning_rate/qwen_vl_interface': 2.7151164198812624e-06, 'learning_rate/action_model': 2.7151164198812622e-05, 'epoch': 0.25}) +[2026-10-03 13:42:37,580][starVLA.training.train_starvla][INFO] - Step 3429, Loss: {'train/loss': 1.8289101717527956e-05, 'train/grad_norm_pre_clip': 0.014698042534291744, 'timing/data': 0.00021297000057529658, 'timing/model': 0.3866191470006015, 'learning_rate/qwen_vl_interface': 2.712541430235388e-06, 'learning_rate/action_model': 2.712541430235388e-05, 'epoch': 0.25}) +[2026-10-03 13:42:38,276][starVLA.training.train_starvla][INFO] - Step 3430, Loss: {'train/loss': 9.983755262510385e-07, 'train/grad_norm_pre_clip': 0.000344281637808308, 'timing/data': 0.0003528600063873455, 'timing/model': 0.38932784700591583, 'learning_rate/qwen_vl_interface': 2.7099674836428842e-06, 'learning_rate/action_model': 2.709967483642884e-05, 'epoch': 0.25}) +[2026-10-03 13:42:38,973][starVLA.training.train_starvla][INFO] - Step 3431, Loss: {'train/loss': 9.162278001895174e-05, 'train/grad_norm_pre_clip': 0.1339099258184433, 'timing/data': 0.00019828999938908964, 'timing/model': 0.3894061359897023, 'learning_rate/qwen_vl_interface': 2.7073945811618018e-06, 'learning_rate/action_model': 2.7073945811618017e-05, 'epoch': 0.25}) +[2026-10-03 13:42:39,670][starVLA.training.train_starvla][INFO] - Step 3432, Loss: {'train/loss': 4.596923190547386e-06, 'train/grad_norm_pre_clip': 0.0038367505185306072, 'timing/data': 0.0001960399968083948, 'timing/model': 0.3901853870047489, 'learning_rate/qwen_vl_interface': 2.704822723849765e-06, 'learning_rate/action_model': 2.7048227238497648e-05, 'epoch': 0.25}) +[2026-10-03 13:42:40,560][starVLA.training.train_starvla][INFO] - Step 3433, Loss: {'train/loss': 1.4139730410533957e-05, 'train/grad_norm_pre_clip': 0.006826462224125862, 'timing/data': 0.0003121399931842461, 'timing/model': 0.3896044169960078, 'learning_rate/qwen_vl_interface': 2.7022519127639687e-06, 'learning_rate/action_model': 2.7022519127639682e-05, 'epoch': 0.25}) +[2026-10-03 13:42:41,256][starVLA.training.train_starvla][INFO] - Step 3434, Loss: {'train/loss': 1.1622281817835756e-05, 'train/grad_norm_pre_clip': 0.0033835831563919783, 'timing/data': 0.0002855200000340119, 'timing/model': 0.3886193170037586, 'learning_rate/qwen_vl_interface': 2.6996821489611713e-06, 'learning_rate/action_model': 2.699682148961171e-05, 'epoch': 0.25}) +[2026-10-03 13:42:41,954][starVLA.training.train_starvla][INFO] - Step 3435, Loss: {'train/loss': 2.4810233298921958e-06, 'train/grad_norm_pre_clip': 0.0017303806962445378, 'timing/data': 0.00019614999473560601, 'timing/model': 0.39053623599465936, 'learning_rate/qwen_vl_interface': 2.697113433497712e-06, 'learning_rate/action_model': 2.6971134334977115e-05, 'epoch': 0.25}) +[2026-10-03 13:42:42,650][starVLA.training.train_starvla][INFO] - Step 3436, Loss: {'train/loss': 1.8551802440924803e-06, 'train/grad_norm_pre_clip': 0.0008585173636674881, 'timing/data': 0.00027686999237630516, 'timing/model': 0.3899813070020173, 'learning_rate/qwen_vl_interface': 2.6945457674294863e-06, 'learning_rate/action_model': 2.694545767429486e-05, 'epoch': 0.25}) +[2026-10-03 13:42:43,348][starVLA.training.train_starvla][INFO] - Step 3437, Loss: {'train/loss': 0.3430640995502472, 'train/grad_norm_pre_clip': 230.5013427734375, 'timing/data': 0.0001951300073415041, 'timing/model': 0.3899414560000878, 'learning_rate/qwen_vl_interface': 2.6919791518119676e-06, 'learning_rate/action_model': 2.6919791518119674e-05, 'epoch': 0.25}) +[2026-10-03 13:42:44,239][starVLA.training.train_starvla][INFO] - Step 3438, Loss: {'train/loss': 1.1041749715805054, 'train/grad_norm_pre_clip': 105.10079193115234, 'timing/data': 0.00020172999938949943, 'timing/model': 0.5835793450096389, 'learning_rate/qwen_vl_interface': 2.6894135877001946e-06, 'learning_rate/action_model': 2.689413587700195e-05, 'epoch': 0.25}) +[2026-10-03 13:42:44,936][starVLA.training.train_starvla][INFO] - Step 3439, Loss: {'train/loss': 0.015335489064455032, 'train/grad_norm_pre_clip': 63.014984130859375, 'timing/data': 0.00028435999411158264, 'timing/model': 0.3901887969987001, 'learning_rate/qwen_vl_interface': 2.686849076148775e-06, 'learning_rate/action_model': 2.6868490761487747e-05, 'epoch': 0.25}) +[2026-10-03 13:42:45,610][starVLA.training.train_starvla][INFO] - Step 3440, Loss: {'train/loss': 2.5580287910997868e-05, 'train/grad_norm_pre_clip': 0.012106752023100853, 'timing/data': 0.00023810999118722975, 'timing/model': 0.3904359770094743, 'learning_rate/qwen_vl_interface': 2.68428561821188e-06, 'learning_rate/action_model': 2.6842856182118802e-05, 'epoch': 0.25}) +[2026-10-03 13:42:46,309][starVLA.training.train_starvla][INFO] - Step 3441, Loss: {'train/loss': 4.514926786214346e-06, 'train/grad_norm_pre_clip': 0.004478760529309511, 'timing/data': 0.0003503800107864663, 'timing/model': 0.39027954699122347, 'learning_rate/qwen_vl_interface': 2.681723214943253e-06, 'learning_rate/action_model': 2.681723214943253e-05, 'epoch': 0.25}) +[2026-10-03 13:42:47,008][starVLA.training.train_starvla][INFO] - Step 3442, Loss: {'train/loss': 2.071249582513701e-06, 'train/grad_norm_pre_clip': 0.0006963909836485982, 'timing/data': 0.00019987999985460192, 'timing/model': 0.39121459699526895, 'learning_rate/qwen_vl_interface': 2.6791618673961996e-06, 'learning_rate/action_model': 2.6791618673961994e-05, 'epoch': 0.25}) +[2026-10-03 13:42:47,705][starVLA.training.train_starvla][INFO] - Step 3443, Loss: {'train/loss': 4.9992240747087635e-06, 'train/grad_norm_pre_clip': 0.001565137761645019, 'timing/data': 0.00020435001351870596, 'timing/model': 0.3894814279919956, 'learning_rate/qwen_vl_interface': 2.676601576623594e-06, 'learning_rate/action_model': 2.6766015766235937e-05, 'epoch': 0.25}) +[2026-10-03 13:42:48,598][starVLA.training.train_starvla][INFO] - Step 3444, Loss: {'train/loss': 0.5305347442626953, 'train/grad_norm_pre_clip': 187.45973205566406, 'timing/data': 0.00020311000116635114, 'timing/model': 0.3904344670008868, 'learning_rate/qwen_vl_interface': 2.674042343677874e-06, 'learning_rate/action_model': 2.6740423436778738e-05, 'epoch': 0.25}) +[2026-10-03 13:42:49,297][starVLA.training.train_starvla][INFO] - Step 3445, Loss: {'train/loss': 0.011495876125991344, 'train/grad_norm_pre_clip': 32.99528121948242, 'timing/data': 0.000327859990647994, 'timing/model': 0.39020673699269537, 'learning_rate/qwen_vl_interface': 2.6714841696110456e-06, 'learning_rate/action_model': 2.6714841696110456e-05, 'epoch': 0.25}) +[2026-10-03 13:42:49,995][starVLA.training.train_starvla][INFO] - Step 3446, Loss: {'train/loss': 5.997623702569399e-06, 'train/grad_norm_pre_clip': 0.0015445379540324211, 'timing/data': 0.00020056999346707016, 'timing/model': 0.3904115069890395, 'learning_rate/qwen_vl_interface': 2.6689270554746743e-06, 'learning_rate/action_model': 2.6689270554746743e-05, 'epoch': 0.25}) +[2026-10-03 13:42:50,692][starVLA.training.train_starvla][INFO] - Step 3447, Loss: {'train/loss': 1.9234605133533478e-05, 'train/grad_norm_pre_clip': 0.039830129593610764, 'timing/data': 0.00023710000095888972, 'timing/model': 0.38988807699934114, 'learning_rate/qwen_vl_interface': 2.6663710023198956e-06, 'learning_rate/action_model': 2.6663710023198956e-05, 'epoch': 0.25}) +[2026-10-03 13:42:51,390][starVLA.training.train_starvla][INFO] - Step 3448, Loss: {'train/loss': 3.270771685492946e-06, 'train/grad_norm_pre_clip': 0.0007215967052616179, 'timing/data': 0.0002971200010506436, 'timing/model': 0.3902772860019468, 'learning_rate/qwen_vl_interface': 2.663816011197405e-06, 'learning_rate/action_model': 2.6638160111974048e-05, 'epoch': 0.25}) +[2026-10-03 13:42:52,280][starVLA.training.train_starvla][INFO] - Step 3449, Loss: {'train/loss': 3.5581095289671794e-05, 'train/grad_norm_pre_clip': 0.03549805283546448, 'timing/data': 0.0002150699874619022, 'timing/model': 0.5823854750051396, 'learning_rate/qwen_vl_interface': 2.661262083157463e-06, 'learning_rate/action_model': 2.661262083157463e-05, 'epoch': 0.25}) +[2026-10-03 13:42:52,978][starVLA.training.train_starvla][INFO] - Step 3450, Loss: {'train/loss': 2.5257245397369843e-06, 'train/grad_norm_pre_clip': 0.0005938507383689284, 'timing/data': 0.0003267599968239665, 'timing/model': 0.39050104600028135, 'learning_rate/qwen_vl_interface': 2.6587092192498943e-06, 'learning_rate/action_model': 2.658709219249894e-05, 'epoch': 0.25}) +[2026-10-03 13:42:53,676][starVLA.training.train_starvla][INFO] - Step 3451, Loss: {'train/loss': 0.5801193714141846, 'train/grad_norm_pre_clip': 94.84906005859375, 'timing/data': 0.00028016000578645617, 'timing/model': 0.3902245770004811, 'learning_rate/qwen_vl_interface': 2.656157420524084e-06, 'learning_rate/action_model': 2.6561574205240842e-05, 'epoch': 0.25}) +[2026-10-03 13:42:54,374][starVLA.training.train_starvla][INFO] - Step 3452, Loss: {'train/loss': 0.3218877613544464, 'train/grad_norm_pre_clip': 52.32668685913086, 'timing/data': 0.0002520199923310429, 'timing/model': 0.3898397919983836, 'learning_rate/qwen_vl_interface': 2.6536066880289795e-06, 'learning_rate/action_model': 2.6536066880289794e-05, 'epoch': 0.25}) +[2026-10-03 13:42:55,071][starVLA.training.train_starvla][INFO] - Step 3453, Loss: {'train/loss': 8.717154855730769e-07, 'train/grad_norm_pre_clip': 0.00015555752906948328, 'timing/data': 0.00019724000594578683, 'timing/model': 0.3902163750026375, 'learning_rate/qwen_vl_interface': 2.6510570228130916e-06, 'learning_rate/action_model': 2.6510570228130914e-05, 'epoch': 0.25}) +[2026-10-03 13:42:55,957][starVLA.training.train_starvla][INFO] - Step 3454, Loss: {'train/loss': 1.3038477391091874e-06, 'train/grad_norm_pre_clip': 0.000250972923822701, 'timing/data': 0.0002601300075184554, 'timing/model': 0.5781163270003162, 'learning_rate/qwen_vl_interface': 2.648508425924491e-06, 'learning_rate/action_model': 2.6485084259244907e-05, 'epoch': 0.25}) +[2026-10-03 13:42:56,655][starVLA.training.train_starvla][INFO] - Step 3455, Loss: {'train/loss': 2.973750451928936e-05, 'train/grad_norm_pre_clip': 0.049546949565410614, 'timing/data': 0.0001960900117410347, 'timing/model': 0.39023698699020315, 'learning_rate/qwen_vl_interface': 2.6459608984108093e-06, 'learning_rate/action_model': 2.645960898410809e-05, 'epoch': 0.25}) +[2026-10-03 13:42:57,353][starVLA.training.train_starvla][INFO] - Step 3456, Loss: {'train/loss': 2.250065108455601e-06, 'train/grad_norm_pre_clip': 0.0004176034708507359, 'timing/data': 0.00020275000133551657, 'timing/model': 0.38932171699707396, 'learning_rate/qwen_vl_interface': 2.6434144413192414e-06, 'learning_rate/action_model': 2.6434144413192414e-05, 'epoch': 0.25}) +[2026-10-03 13:42:58,050][starVLA.training.train_starvla][INFO] - Step 3457, Loss: {'train/loss': 0.0009618340991437435, 'train/grad_norm_pre_clip': 4.117501735687256, 'timing/data': 0.0002511899947421625, 'timing/model': 0.3896923659922322, 'learning_rate/qwen_vl_interface': 2.6408690556965347e-06, 'learning_rate/action_model': 2.640869055696535e-05, 'epoch': 0.25}) +[2026-10-03 13:42:58,748][starVLA.training.train_starvla][INFO] - Step 3458, Loss: {'train/loss': 2.495922217349289e-06, 'train/grad_norm_pre_clip': 0.0016853116685524583, 'timing/data': 0.00028456001018639654, 'timing/model': 0.38996678599505685, 'learning_rate/qwen_vl_interface': 2.6383247425890047e-06, 'learning_rate/action_model': 2.6383247425890047e-05, 'epoch': 0.25}) +[2026-10-03 13:42:59,445][starVLA.training.train_starvla][INFO] - Step 3459, Loss: {'train/loss': 5.699647317669587e-06, 'train/grad_norm_pre_clip': 0.0011785555398091674, 'timing/data': 0.00024913999368436635, 'timing/model': 0.39023866699426435, 'learning_rate/qwen_vl_interface': 2.6357815030425204e-06, 'learning_rate/action_model': 2.6357815030425205e-05, 'epoch': 0.25}) +[2026-10-03 13:43:00,337][starVLA.training.train_starvla][INFO] - Step 3460, Loss: {'train/loss': 6.243497409741394e-06, 'train/grad_norm_pre_clip': 0.0013284825254231691, 'timing/data': 0.00020156000391580164, 'timing/model': 0.3905203159956727, 'learning_rate/qwen_vl_interface': 2.6332393381025132e-06, 'learning_rate/action_model': 2.6332393381025133e-05, 'epoch': 0.25}) +[2026-10-03 13:43:01,034][starVLA.training.train_starvla][INFO] - Step 3461, Loss: {'train/loss': 5.8474055549595505e-05, 'train/grad_norm_pre_clip': 0.0330844484269619, 'timing/data': 0.0003815300005953759, 'timing/model': 0.3896164469915675, 'learning_rate/qwen_vl_interface': 2.6306982488139704e-06, 'learning_rate/action_model': 2.6306982488139702e-05, 'epoch': 0.25}) +[2026-10-03 13:43:01,732][starVLA.training.train_starvla][INFO] - Step 3462, Loss: {'train/loss': 1.6018118913052604e-05, 'train/grad_norm_pre_clip': 0.007752659730613232, 'timing/data': 0.00019445001089479774, 'timing/model': 0.39063580699439626, 'learning_rate/qwen_vl_interface': 2.628158236221438e-06, 'learning_rate/action_model': 2.628158236221438e-05, 'epoch': 0.25}) +[2026-10-03 13:43:02,429][starVLA.training.train_starvla][INFO] - Step 3463, Loss: {'train/loss': 0.448635995388031, 'train/grad_norm_pre_clip': 219.56787109375, 'timing/data': 0.00026546001026872545, 'timing/model': 0.38978020699869376, 'learning_rate/qwen_vl_interface': 2.6256193013690183e-06, 'learning_rate/action_model': 2.6256193013690182e-05, 'epoch': 0.25}) +[2026-10-03 13:43:03,126][starVLA.training.train_starvla][INFO] - Step 3464, Loss: {'train/loss': 3.322920747450553e-06, 'train/grad_norm_pre_clip': 0.0006087995134294033, 'timing/data': 0.00019785000768024474, 'timing/model': 0.3903749969904311, 'learning_rate/qwen_vl_interface': 2.623081445300373e-06, 'learning_rate/action_model': 2.6230814453003728e-05, 'epoch': 0.25}) +[2026-10-03 13:43:04,015][starVLA.training.train_starvla][INFO] - Step 3465, Loss: {'train/loss': 6.437232968892204e-06, 'train/grad_norm_pre_clip': 0.001769375172443688, 'timing/data': 0.0002028399903792888, 'timing/model': 0.38982920699345414, 'learning_rate/qwen_vl_interface': 2.6205446690587183e-06, 'learning_rate/action_model': 2.6205446690587182e-05, 'epoch': 0.25}) +[2026-10-03 13:43:04,713][starVLA.training.train_starvla][INFO] - Step 3466, Loss: {'train/loss': 7.681446732021868e-06, 'train/grad_norm_pre_clip': 0.003357921727001667, 'timing/data': 0.00023594999220222235, 'timing/model': 0.3903741669928422, 'learning_rate/qwen_vl_interface': 2.618008973686828e-06, 'learning_rate/action_model': 2.618008973686828e-05, 'epoch': 0.25}) +[2026-10-03 13:43:05,409][starVLA.training.train_starvla][INFO] - Step 3467, Loss: {'train/loss': 0.5482688546180725, 'train/grad_norm_pre_clip': 40.6384162902832, 'timing/data': 0.00028529000701382756, 'timing/model': 0.38974176700867247, 'learning_rate/qwen_vl_interface': 2.615474360227028e-06, 'learning_rate/action_model': 2.615474360227028e-05, 'epoch': 0.25}) +[2026-10-03 13:43:06,108][starVLA.training.train_starvla][INFO] - Step 3468, Loss: {'train/loss': 0.0030010228510946035, 'train/grad_norm_pre_clip': 23.161174774169922, 'timing/data': 0.0002859700034605339, 'timing/model': 0.3911455159977777, 'learning_rate/qwen_vl_interface': 2.612940829721204e-06, 'learning_rate/action_model': 2.612940829721204e-05, 'epoch': 0.25}) +[2026-10-03 13:43:06,806][starVLA.training.train_starvla][INFO] - Step 3469, Loss: {'train/loss': 8.552964573027566e-06, 'train/grad_norm_pre_clip': 0.0024802011903375387, 'timing/data': 0.0003318000090075657, 'timing/model': 0.3896634469856508, 'learning_rate/qwen_vl_interface': 2.6104083832107948e-06, 'learning_rate/action_model': 2.6104083832107945e-05, 'epoch': 0.25}) +[2026-10-03 13:43:07,696][starVLA.training.train_starvla][INFO] - Step 3470, Loss: {'train/loss': 3.8296660932246596e-05, 'train/grad_norm_pre_clip': 0.09274879842996597, 'timing/data': 0.0001956899941433221, 'timing/model': 0.5831401059986092, 'learning_rate/qwen_vl_interface': 2.607877021736793e-06, 'learning_rate/action_model': 2.6078770217367926e-05, 'epoch': 0.25}) +[2026-10-03 13:43:08,393][starVLA.training.train_starvla][INFO] - Step 3471, Loss: {'train/loss': 1.4058889064472169e-05, 'train/grad_norm_pre_clip': 0.0030063227750360966, 'timing/data': 0.0002529600023990497, 'timing/model': 0.38974208700528834, 'learning_rate/qwen_vl_interface': 2.605346746339745e-06, 'learning_rate/action_model': 2.605346746339745e-05, 'epoch': 0.25}) +[2026-10-03 13:43:09,091][starVLA.training.train_starvla][INFO] - Step 3472, Loss: {'train/loss': 0.011549712158739567, 'train/grad_norm_pre_clip': 24.357759475708008, 'timing/data': 0.0002577199920779094, 'timing/model': 0.3900823880103417, 'learning_rate/qwen_vl_interface': 2.602817558059753e-06, 'learning_rate/action_model': 2.6028175580597532e-05, 'epoch': 0.25}) +[2026-10-03 13:43:09,789][starVLA.training.train_starvla][INFO] - Step 3473, Loss: {'train/loss': 1.1197967978660017e-05, 'train/grad_norm_pre_clip': 0.003395630279555917, 'timing/data': 0.0003028899955097586, 'timing/model': 0.39002710800559726, 'learning_rate/qwen_vl_interface': 2.60028945793647e-06, 'learning_rate/action_model': 2.60028945793647e-05, 'epoch': 0.25}) +[2026-10-03 13:43:10,486][starVLA.training.train_starvla][INFO] - Step 3474, Loss: {'train/loss': 3.834959716186859e-05, 'train/grad_norm_pre_clip': 0.02896750718355179, 'timing/data': 0.00024434999795630574, 'timing/model': 0.3901006269879872, 'learning_rate/qwen_vl_interface': 2.5977624470091024e-06, 'learning_rate/action_model': 2.5977624470091023e-05, 'epoch': 0.25}) +[2026-10-03 13:43:11,376][starVLA.training.train_starvla][INFO] - Step 3475, Loss: {'train/loss': 0.0007784450426697731, 'train/grad_norm_pre_clip': 3.2202014923095703, 'timing/data': 0.00019615000928752124, 'timing/model': 0.5830814369983273, 'learning_rate/qwen_vl_interface': 2.59523652631641e-06, 'learning_rate/action_model': 2.5952365263164097e-05, 'epoch': 0.25}) +[2026-10-03 13:43:12,074][starVLA.training.train_starvla][INFO] - Step 3476, Loss: {'train/loss': 3.010012278537033e-06, 'train/grad_norm_pre_clip': 0.0009352653869427741, 'timing/data': 0.0002014600031543523, 'timing/model': 0.3899180270091165, 'learning_rate/qwen_vl_interface': 2.5927116968967024e-06, 'learning_rate/action_model': 2.5927116968967025e-05, 'epoch': 0.25}) +[2026-10-03 13:43:12,770][starVLA.training.train_starvla][INFO] - Step 3477, Loss: {'train/loss': 1.8037037079920992e-05, 'train/grad_norm_pre_clip': 0.008427564054727554, 'timing/data': 0.00024604999634902924, 'timing/model': 0.38976016800734214, 'learning_rate/qwen_vl_interface': 2.5901879597878417e-06, 'learning_rate/action_model': 2.5901879597878415e-05, 'epoch': 0.25}) +[2026-10-03 13:43:13,468][starVLA.training.train_starvla][INFO] - Step 3478, Loss: {'train/loss': 1.922237061080523e-06, 'train/grad_norm_pre_clip': 0.0009034028626047075, 'timing/data': 0.00024896999821066856, 'timing/model': 0.39029507800296415, 'learning_rate/qwen_vl_interface': 2.5876653160272395e-06, 'learning_rate/action_model': 2.5876653160272397e-05, 'epoch': 0.25}) +[2026-10-03 13:43:14,165][starVLA.training.train_starvla][INFO] - Step 3479, Loss: {'train/loss': 2.167200182157103e-05, 'train/grad_norm_pre_clip': 0.010725117288529873, 'timing/data': 0.00020136999955866486, 'timing/model': 0.39042188800522126, 'learning_rate/qwen_vl_interface': 2.5851437666518607e-06, 'learning_rate/action_model': 2.5851437666518607e-05, 'epoch': 0.25}) +[2026-10-03 13:43:14,862][starVLA.training.train_starvla][INFO] - Step 3480, Loss: {'train/loss': 4.863808862864971e-05, 'train/grad_norm_pre_clip': 0.047649092972278595, 'timing/data': 0.00024138999287970364, 'timing/model': 0.3898909670097055, 'learning_rate/qwen_vl_interface': 2.5826233126982198e-06, 'learning_rate/action_model': 2.5826233126982196e-05, 'epoch': 0.25}) +[2026-10-03 13:43:15,753][starVLA.training.train_starvla][INFO] - Step 3481, Loss: {'train/loss': 0.025454746559262276, 'train/grad_norm_pre_clip': 25.79853057861328, 'timing/data': 0.000232170001254417, 'timing/model': 0.38972215801186394, 'learning_rate/qwen_vl_interface': 2.580103955202381e-06, 'learning_rate/action_model': 2.580103955202381e-05, 'epoch': 0.25}) +[2026-10-03 13:43:16,451][starVLA.training.train_starvla][INFO] - Step 3482, Loss: {'train/loss': 0.2608270049095154, 'train/grad_norm_pre_clip': 187.3525390625, 'timing/data': 0.00031242000113707036, 'timing/model': 0.3900404079904547, 'learning_rate/qwen_vl_interface': 2.577585695199954e-06, 'learning_rate/action_model': 2.5775856951999538e-05, 'epoch': 0.25}) +[2026-10-03 13:43:17,145][starVLA.training.train_starvla][INFO] - Step 3483, Loss: {'train/loss': 3.476404526736587e-05, 'train/grad_norm_pre_clip': 0.052461639046669006, 'timing/data': 0.00043431000085547566, 'timing/model': 0.38833388799685054, 'learning_rate/qwen_vl_interface': 2.575068533726101e-06, 'learning_rate/action_model': 2.575068533726101e-05, 'epoch': 0.25}) +[2026-10-03 13:43:17,838][starVLA.training.train_starvla][INFO] - Step 3484, Loss: {'train/loss': 0.4322415590286255, 'train/grad_norm_pre_clip': 99.83816528320312, 'timing/data': 0.0002016200014622882, 'timing/model': 0.38849897900945507, 'learning_rate/qwen_vl_interface': 2.5725524718155333e-06, 'learning_rate/action_model': 2.572552471815533e-05, 'epoch': 0.26}) +[2026-10-03 13:43:18,533][starVLA.training.train_starvla][INFO] - Step 3485, Loss: {'train/loss': 0.2034723460674286, 'train/grad_norm_pre_clip': 162.02145385742188, 'timing/data': 0.00023160000273492187, 'timing/model': 0.3880050379957538, 'learning_rate/qwen_vl_interface': 2.570037510502508e-06, 'learning_rate/action_model': 2.5700375105025083e-05, 'epoch': 0.26}) +[2026-10-03 13:43:19,420][starVLA.training.train_starvla][INFO] - Step 3486, Loss: {'train/loss': 7.927273145469371e-06, 'train/grad_norm_pre_clip': 0.003357918933033943, 'timing/data': 0.0001980900124181062, 'timing/model': 0.38783272900036536, 'learning_rate/qwen_vl_interface': 2.5675236508208313e-06, 'learning_rate/action_model': 2.5675236508208313e-05, 'epoch': 0.26}) +[2026-10-03 13:43:20,114][starVLA.training.train_starvla][INFO] - Step 3487, Loss: {'train/loss': 1.4274513887357898e-05, 'train/grad_norm_pre_clip': 0.013350251130759716, 'timing/data': 0.0001983399997698143, 'timing/model': 0.3882678490044782, 'learning_rate/qwen_vl_interface': 2.565010893803856e-06, 'learning_rate/action_model': 2.565010893803856e-05, 'epoch': 0.26}) +[2026-10-03 13:43:20,808][starVLA.training.train_starvla][INFO] - Step 3488, Loss: {'train/loss': 1.8394735889160074e-05, 'train/grad_norm_pre_clip': 0.011411964893341064, 'timing/data': 0.00025953000294975936, 'timing/model': 0.38857162799104117, 'learning_rate/qwen_vl_interface': 2.562499240484482e-06, 'learning_rate/action_model': 2.5624992404844816e-05, 'epoch': 0.26}) +[2026-10-03 13:43:21,502][starVLA.training.train_starvla][INFO] - Step 3489, Loss: {'train/loss': 1.1145155440317467e-05, 'train/grad_norm_pre_clip': 0.0039817821234464645, 'timing/data': 0.000245489994995296, 'timing/model': 0.3880968080047751, 'learning_rate/qwen_vl_interface': 2.559988691895155e-06, 'learning_rate/action_model': 2.559988691895155e-05, 'epoch': 0.26}) +[2026-10-03 13:43:22,196][starVLA.training.train_starvla][INFO] - Step 3490, Loss: {'train/loss': 3.2409852792625315e-06, 'train/grad_norm_pre_clip': 0.0006558798486366868, 'timing/data': 0.00031442999897990376, 'timing/model': 0.38765933900140226, 'learning_rate/qwen_vl_interface': 2.5574792490678648e-06, 'learning_rate/action_model': 2.5574792490678644e-05, 'epoch': 0.26}) +[2026-10-03 13:43:23,081][starVLA.training.train_starvla][INFO] - Step 3491, Loss: {'train/loss': 0.5408268570899963, 'train/grad_norm_pre_clip': 62.561485290527344, 'timing/data': 0.00028311999631114304, 'timing/model': 0.5795085279969499, 'learning_rate/qwen_vl_interface': 2.5549709130341493e-06, 'learning_rate/action_model': 2.5549709130341493e-05, 'epoch': 0.26}) +[2026-10-03 13:43:23,775][starVLA.training.train_starvla][INFO] - Step 3492, Loss: {'train/loss': 3.1183724786387756e-05, 'train/grad_norm_pre_clip': 0.008859777823090553, 'timing/data': 0.00030481000430881977, 'timing/model': 0.38817461898725014, 'learning_rate/qwen_vl_interface': 2.5524636848250924e-06, 'learning_rate/action_model': 2.552463684825092e-05, 'epoch': 0.26}) +[2026-10-03 13:43:24,468][starVLA.training.train_starvla][INFO] - Step 3493, Loss: {'train/loss': 2.741794787652907e-06, 'train/grad_norm_pre_clip': 0.0004841267073061317, 'timing/data': 0.00029445000109262764, 'timing/model': 0.3881039489933755, 'learning_rate/qwen_vl_interface': 2.5499575654713186e-06, 'learning_rate/action_model': 2.5499575654713185e-05, 'epoch': 0.26}) +[2026-10-03 13:43:25,162][starVLA.training.train_starvla][INFO] - Step 3494, Loss: {'train/loss': 6.43713883619057e-06, 'train/grad_norm_pre_clip': 0.004027322866022587, 'timing/data': 0.0001969299919437617, 'timing/model': 0.38814395999361295, 'learning_rate/qwen_vl_interface': 2.547452556003001e-06, 'learning_rate/action_model': 2.547452556003001e-05, 'epoch': 0.26}) +[2026-10-03 13:43:25,857][starVLA.training.train_starvla][INFO] - Step 3495, Loss: {'train/loss': 1.8446389731252566e-05, 'train/grad_norm_pre_clip': 0.004876026418060064, 'timing/data': 0.0003092400002060458, 'timing/model': 0.3883536590001313, 'learning_rate/qwen_vl_interface': 2.5449486574498535e-06, 'learning_rate/action_model': 2.5449486574498534e-05, 'epoch': 0.26}) +[2026-10-03 13:43:26,746][starVLA.training.train_starvla][INFO] - Step 3496, Loss: {'train/loss': 0.0004917174810543656, 'train/grad_norm_pre_clip': 0.44741684198379517, 'timing/data': 0.00019978999625891447, 'timing/model': 0.5830838380061323, 'learning_rate/qwen_vl_interface': 2.5424458708411335e-06, 'learning_rate/action_model': 2.5424458708411335e-05, 'epoch': 0.26}) +[2026-10-03 13:43:27,439][starVLA.training.train_starvla][INFO] - Step 3497, Loss: {'train/loss': 4.865150913246907e-06, 'train/grad_norm_pre_clip': 0.0039892978966236115, 'timing/data': 0.0003179699997417629, 'timing/model': 0.38809446900268085, 'learning_rate/qwen_vl_interface': 2.5399441972056437e-06, 'learning_rate/action_model': 2.5399441972056438e-05, 'epoch': 0.26}) +[2026-10-03 13:43:28,134][starVLA.training.train_starvla][INFO] - Step 3498, Loss: {'train/loss': 0.07386307418346405, 'train/grad_norm_pre_clip': 157.16000366210938, 'timing/data': 0.00020209999638609588, 'timing/model': 0.3889777790027438, 'learning_rate/qwen_vl_interface': 2.537443637571726e-06, 'learning_rate/action_model': 2.537443637571726e-05, 'epoch': 0.26}) +[2026-10-03 13:43:28,830][starVLA.training.train_starvla][INFO] - Step 3499, Loss: {'train/loss': 8.991952199721709e-05, 'train/grad_norm_pre_clip': 0.07539689540863037, 'timing/data': 0.0003331789921503514, 'timing/model': 0.3899683789932169, 'learning_rate/qwen_vl_interface': 2.534944192967269e-06, 'learning_rate/action_model': 2.534944192967269e-05, 'epoch': 0.26}) +[2026-10-03 13:43:29,527][starVLA.training.train_starvla][INFO] - Step 3500, Loss: {'train/loss': 1.9675579096656293e-05, 'train/grad_norm_pre_clip': 0.005137850996106863, 'timing/data': 0.00021173000277485698, 'timing/model': 0.390318847988965, 'learning_rate/qwen_vl_interface': 2.532445864419695e-06, 'learning_rate/action_model': 2.5324458644196952e-05, 'epoch': 0.26}) +[2026-10-03 13:43:30,226][starVLA.training.train_starvla][INFO] - Step 3501, Loss: {'train/loss': 0.3228459358215332, 'train/grad_norm_pre_clip': 58.10103225708008, 'timing/data': 0.00035090000892523676, 'timing/model': 0.3907527389965253, 'learning_rate/qwen_vl_interface': 2.5299486529559758e-06, 'learning_rate/action_model': 2.5299486529559756e-05, 'epoch': 0.26}) +[2026-10-03 13:43:31,127][starVLA.training.train_starvla][INFO] - Step 3502, Loss: {'train/loss': 0.00030912228976376355, 'train/grad_norm_pre_clip': 0.13840167224407196, 'timing/data': 0.00023459899239242077, 'timing/model': 0.3903150789992651, 'learning_rate/qwen_vl_interface': 2.5274525596026202e-06, 'learning_rate/action_model': 2.52745255960262e-05, 'epoch': 0.26}) +[2026-10-03 13:43:31,826][starVLA.training.train_starvla][INFO] - Step 3503, Loss: {'train/loss': 0.3629273474216461, 'train/grad_norm_pre_clip': 371.58087158203125, 'timing/data': 0.0003512399998726323, 'timing/model': 0.3907076390023576, 'learning_rate/qwen_vl_interface': 2.524957585385677e-06, 'learning_rate/action_model': 2.524957585385677e-05, 'epoch': 0.26}) +[2026-10-03 13:43:32,524][starVLA.training.train_starvla][INFO] - Step 3504, Loss: {'train/loss': 4.7784895286895335e-05, 'train/grad_norm_pre_clip': 0.055016402155160904, 'timing/data': 0.0002535099920351058, 'timing/model': 0.39057039898762014, 'learning_rate/qwen_vl_interface': 2.522463731330738e-06, 'learning_rate/action_model': 2.5224637313307374e-05, 'epoch': 0.26}) +[2026-10-03 13:43:33,221][starVLA.training.train_starvla][INFO] - Step 3505, Loss: {'train/loss': 3.004438622156158e-05, 'train/grad_norm_pre_clip': 0.03707119822502136, 'timing/data': 0.0002651899994816631, 'timing/model': 0.389968098999816, 'learning_rate/qwen_vl_interface': 2.5199709984629294e-06, 'learning_rate/action_model': 2.5199709984629292e-05, 'epoch': 0.26}) +[2026-10-03 13:43:33,919][starVLA.training.train_starvla][INFO] - Step 3506, Loss: {'train/loss': 6.37842167634517e-05, 'train/grad_norm_pre_clip': 0.018700389191508293, 'timing/data': 0.00031892998958937824, 'timing/model': 0.39073664900206495, 'learning_rate/qwen_vl_interface': 2.517479387806922e-06, 'learning_rate/action_model': 2.517479387806922e-05, 'epoch': 0.26}) +[2026-10-03 13:43:34,819][starVLA.training.train_starvla][INFO] - Step 3507, Loss: {'train/loss': 0.2911563813686371, 'train/grad_norm_pre_clip': 201.1337890625, 'timing/data': 0.00035837000177707523, 'timing/model': 0.5910751670016907, 'learning_rate/qwen_vl_interface': 2.5149889003869217e-06, 'learning_rate/action_model': 2.5149889003869214e-05, 'epoch': 0.26}) +[2026-10-03 13:43:35,517][starVLA.training.train_starvla][INFO] - Step 3508, Loss: {'train/loss': 0.00016706250607967377, 'train/grad_norm_pre_clip': 0.09995283931493759, 'timing/data': 0.00033573000109754503, 'timing/model': 0.3903177789907204, 'learning_rate/qwen_vl_interface': 2.5124995372266736e-06, 'learning_rate/action_model': 2.5124995372266735e-05, 'epoch': 0.26}) +[2026-10-03 13:43:36,215][starVLA.training.train_starvla][INFO] - Step 3509, Loss: {'train/loss': 1.861753480625339e-05, 'train/grad_norm_pre_clip': 0.005657099653035402, 'timing/data': 0.00034712998603936285, 'timing/model': 0.39062063899473287, 'learning_rate/qwen_vl_interface': 2.5100112993494626e-06, 'learning_rate/action_model': 2.5100112993494624e-05, 'epoch': 0.26}) +[2026-10-03 13:43:36,912][starVLA.training.train_starvla][INFO] - Step 3510, Loss: {'train/loss': 6.294756894931197e-05, 'train/grad_norm_pre_clip': 0.017851868644356728, 'timing/data': 0.0001991989993257448, 'timing/model': 0.3898155889910413, 'learning_rate/qwen_vl_interface': 2.5075241877781085e-06, 'learning_rate/action_model': 2.5075241877781086e-05, 'epoch': 0.26}) +[2026-10-03 13:43:37,610][starVLA.training.train_starvla][INFO] - Step 3511, Loss: {'train/loss': 4.817696753889322e-05, 'train/grad_norm_pre_clip': 0.06938089430332184, 'timing/data': 0.00031401000160258263, 'timing/model': 0.39004185900557786, 'learning_rate/qwen_vl_interface': 2.5050382035349673e-06, 'learning_rate/action_model': 2.5050382035349674e-05, 'epoch': 0.26}) +[2026-10-03 13:43:38,510][starVLA.training.train_starvla][INFO] - Step 3512, Loss: {'train/loss': 0.27897509932518005, 'train/grad_norm_pre_clip': 150.53515625, 'timing/data': 0.0003383699950063601, 'timing/model': 0.5925784579885658, 'learning_rate/qwen_vl_interface': 2.5025533476419343e-06, 'learning_rate/action_model': 2.5025533476419343e-05, 'epoch': 0.26}) +[2026-10-03 13:43:39,207][starVLA.training.train_starvla][INFO] - Step 3513, Loss: {'train/loss': 0.06605000793933868, 'train/grad_norm_pre_clip': 92.61634063720703, 'timing/data': 0.0003216099867131561, 'timing/model': 0.39011009999376256, 'learning_rate/qwen_vl_interface': 2.5000696211204394e-06, 'learning_rate/action_model': 2.5000696211204394e-05, 'epoch': 0.26}) +[2026-10-03 13:43:39,905][starVLA.training.train_starvla][INFO] - Step 3514, Loss: {'train/loss': 0.0001248054759344086, 'train/grad_norm_pre_clip': 0.17524898052215576, 'timing/data': 0.00032503000693395734, 'timing/model': 0.3908283990022028, 'learning_rate/qwen_vl_interface': 2.4975870249914495e-06, 'learning_rate/action_model': 2.4975870249914497e-05, 'epoch': 0.26}) +[2026-10-03 13:43:40,603][starVLA.training.train_starvla][INFO] - Step 3515, Loss: {'train/loss': 0.0001773949625203386, 'train/grad_norm_pre_clip': 0.2844234108924866, 'timing/data': 0.00028310999914538115, 'timing/model': 0.3901686289900681, 'learning_rate/qwen_vl_interface': 2.4951055602754664e-06, 'learning_rate/action_model': 2.4951055602754665e-05, 'epoch': 0.26}) +[2026-10-03 13:43:41,302][starVLA.training.train_starvla][INFO] - Step 3516, Loss: {'train/loss': 0.00014223346079234034, 'train/grad_norm_pre_clip': 0.1894521266222, 'timing/data': 0.0002420800010440871, 'timing/model': 0.3911771089915419, 'learning_rate/qwen_vl_interface': 2.4926252279925255e-06, 'learning_rate/action_model': 2.4926252279925254e-05, 'epoch': 0.26}) +[2026-10-03 13:43:42,000][starVLA.training.train_starvla][INFO] - Step 3517, Loss: {'train/loss': 0.0009307941654697061, 'train/grad_norm_pre_clip': 2.16140079498291, 'timing/data': 0.0004291400109650567, 'timing/model': 0.3892644090083195, 'learning_rate/qwen_vl_interface': 2.490146029162199e-06, 'learning_rate/action_model': 2.4901460291621988e-05, 'epoch': 0.26}) +[2026-10-03 13:43:42,893][starVLA.training.train_starvla][INFO] - Step 3518, Loss: {'train/loss': 3.743592242244631e-05, 'train/grad_norm_pre_clip': 0.008214165456593037, 'timing/data': 0.00025036999431904405, 'timing/model': 0.3899588900094386, 'learning_rate/qwen_vl_interface': 2.4876679648035905e-06, 'learning_rate/action_model': 2.4876679648035903e-05, 'epoch': 0.26}) +[2026-10-03 13:43:43,588][starVLA.training.train_starvla][INFO] - Step 3519, Loss: {'train/loss': 1.2069920103385812e-06, 'train/grad_norm_pre_clip': 0.00028450778336264193, 'timing/data': 0.0002538599947001785, 'timing/model': 0.3884849199966993, 'learning_rate/qwen_vl_interface': 2.4851910359353392e-06, 'learning_rate/action_model': 2.485191035935339e-05, 'epoch': 0.26}) +[2026-10-03 13:43:44,281][starVLA.training.train_starvla][INFO] - Step 3520, Loss: {'train/loss': 4.201818592264317e-05, 'train/grad_norm_pre_clip': 0.01104161236435175, 'timing/data': 0.0002491100021870807, 'timing/model': 0.38772526000684593, 'learning_rate/qwen_vl_interface': 2.4827152435756185e-06, 'learning_rate/action_model': 2.4827152435756185e-05, 'epoch': 0.26}) +[2026-10-03 13:43:44,974][starVLA.training.train_starvla][INFO] - Step 3521, Loss: {'train/loss': 0.68997722864151, 'train/grad_norm_pre_clip': 52.31880569458008, 'timing/data': 0.00019661000987980515, 'timing/model': 0.38824856000428554, 'learning_rate/qwen_vl_interface': 2.4802405887421307e-06, 'learning_rate/action_model': 2.4802405887421306e-05, 'epoch': 0.26}) +[2026-10-03 13:43:45,669][starVLA.training.train_starvla][INFO] - Step 3522, Loss: {'train/loss': 0.0006212524604052305, 'train/grad_norm_pre_clip': 0.7663483619689941, 'timing/data': 0.00024499000573996454, 'timing/model': 0.3892023899970809, 'learning_rate/qwen_vl_interface': 2.4777670724521134e-06, 'learning_rate/action_model': 2.4777670724521134e-05, 'epoch': 0.26}) +[2026-10-03 13:43:46,557][starVLA.training.train_starvla][INFO] - Step 3523, Loss: {'train/loss': 4.060540959471837e-06, 'train/grad_norm_pre_clip': 0.001923333969898522, 'timing/data': 0.0003362499992363155, 'timing/model': 0.3880352300038794, 'learning_rate/qwen_vl_interface': 2.475294695722337e-06, 'learning_rate/action_model': 2.475294695722337e-05, 'epoch': 0.26}) +[2026-10-03 13:43:47,251][starVLA.training.train_starvla][INFO] - Step 3524, Loss: {'train/loss': 0.6056281328201294, 'train/grad_norm_pre_clip': 36.054649353027344, 'timing/data': 0.00023620000865776092, 'timing/model': 0.38842496000870597, 'learning_rate/qwen_vl_interface': 2.4728234595691006e-06, 'learning_rate/action_model': 2.4728234595691005e-05, 'epoch': 0.26}) +[2026-10-03 13:43:47,944][starVLA.training.train_starvla][INFO] - Step 3525, Loss: {'train/loss': 0.7555966973304749, 'train/grad_norm_pre_clip': 73.06220245361328, 'timing/data': 0.0002924699947470799, 'timing/model': 0.38768646999960765, 'learning_rate/qwen_vl_interface': 2.4703533650082368e-06, 'learning_rate/action_model': 2.4703533650082366e-05, 'epoch': 0.26}) +[2026-10-03 13:43:48,638][starVLA.training.train_starvla][INFO] - Step 3526, Loss: {'train/loss': 1.4669356460217386e-05, 'train/grad_norm_pre_clip': 0.003192687639966607, 'timing/data': 0.0001982200046768412, 'timing/model': 0.3879776209942065, 'learning_rate/qwen_vl_interface': 2.4678844130551103e-06, 'learning_rate/action_model': 2.4678844130551103e-05, 'epoch': 0.26}) +[2026-10-03 13:43:49,332][starVLA.training.train_starvla][INFO] - Step 3527, Loss: {'train/loss': 4.142488251090981e-06, 'train/grad_norm_pre_clip': 0.0009651329019106925, 'timing/data': 0.00024947000201791525, 'timing/model': 0.3880940500093857, 'learning_rate/qwen_vl_interface': 2.4654166047246102e-06, 'learning_rate/action_model': 2.46541660472461e-05, 'epoch': 0.26}) +[2026-10-03 13:43:50,222][starVLA.training.train_starvla][INFO] - Step 3528, Loss: {'train/loss': 1.9704999431269243e-05, 'train/grad_norm_pre_clip': 0.018295248970389366, 'timing/data': 0.00025782999000512064, 'timing/model': 0.38829788999282755, 'learning_rate/qwen_vl_interface': 2.4629499410311606e-06, 'learning_rate/action_model': 2.462949941031161e-05, 'epoch': 0.26}) +[2026-10-03 13:43:50,917][starVLA.training.train_starvla][INFO] - Step 3529, Loss: {'train/loss': 4.421298945089802e-05, 'train/grad_norm_pre_clip': 0.018130114302039146, 'timing/data': 0.00028573999588843435, 'timing/model': 0.38809232000494376, 'learning_rate/qwen_vl_interface': 2.460484422988715e-06, 'learning_rate/action_model': 2.4604844229887146e-05, 'epoch': 0.26}) +[2026-10-03 13:43:51,611][starVLA.training.train_starvla][INFO] - Step 3530, Loss: {'train/loss': 4.769442239194177e-05, 'train/grad_norm_pre_clip': 0.045672956854104996, 'timing/data': 0.00025028000527527183, 'timing/model': 0.38841616999707185, 'learning_rate/qwen_vl_interface': 2.458020051610753e-06, 'learning_rate/action_model': 2.4580200516107528e-05, 'epoch': 0.26}) +[2026-10-03 13:43:52,305][starVLA.training.train_starvla][INFO] - Step 3531, Loss: {'train/loss': 8.359432285942603e-06, 'train/grad_norm_pre_clip': 0.0015743935946375132, 'timing/data': 0.0003092300030402839, 'timing/model': 0.38868253100372385, 'learning_rate/qwen_vl_interface': 2.455556827910286e-06, 'learning_rate/action_model': 2.455556827910286e-05, 'epoch': 0.26}) +[2026-10-03 13:43:52,999][starVLA.training.train_starvla][INFO] - Step 3532, Loss: {'train/loss': 1.7299549654126167e-05, 'train/grad_norm_pre_clip': 0.007036449853330851, 'timing/data': 0.00029973000346217304, 'timing/model': 0.38825782100320794, 'learning_rate/qwen_vl_interface': 2.453094752899853e-06, 'learning_rate/action_model': 2.453094752899853e-05, 'epoch': 0.26}) +[2026-10-03 13:43:53,884][starVLA.training.train_starvla][INFO] - Step 3533, Loss: {'train/loss': 0.42767202854156494, 'train/grad_norm_pre_clip': 225.58538818359375, 'timing/data': 0.00031447000219486654, 'timing/model': 0.5791610209998908, 'learning_rate/qwen_vl_interface': 2.4506338275915173e-06, 'learning_rate/action_model': 2.4506338275915172e-05, 'epoch': 0.26}) +[2026-10-03 13:43:54,578][starVLA.training.train_starvla][INFO] - Step 3534, Loss: {'train/loss': 0.001722063054330647, 'train/grad_norm_pre_clip': 5.364284992218018, 'timing/data': 0.00028360000578686595, 'timing/model': 0.3881304700043984, 'learning_rate/qwen_vl_interface': 2.4481740529968735e-06, 'learning_rate/action_model': 2.4481740529968734e-05, 'epoch': 0.26}) +[2026-10-03 13:43:55,271][starVLA.training.train_starvla][INFO] - Step 3535, Loss: {'train/loss': 0.00017711662803776562, 'train/grad_norm_pre_clip': 0.04677915573120117, 'timing/data': 0.00023963999410625547, 'timing/model': 0.3884316899930127, 'learning_rate/qwen_vl_interface': 2.4457154301270427e-06, 'learning_rate/action_model': 2.4457154301270427e-05, 'epoch': 0.26}) +[2026-10-03 13:43:55,963][starVLA.training.train_starvla][INFO] - Step 3536, Loss: {'train/loss': 1.6830414097057655e-05, 'train/grad_norm_pre_clip': 0.006902844645082951, 'timing/data': 0.00029897000058554113, 'timing/model': 0.387695890996838, 'learning_rate/qwen_vl_interface': 2.443257959992671e-06, 'learning_rate/action_model': 2.443257959992671e-05, 'epoch': 0.26}) +[2026-10-03 13:43:56,656][starVLA.training.train_starvla][INFO] - Step 3537, Loss: {'train/loss': 0.0002598474093247205, 'train/grad_norm_pre_clip': 0.12092592567205429, 'timing/data': 0.0002807899873005226, 'timing/model': 0.3872538700088626, 'learning_rate/qwen_vl_interface': 2.440801643603933e-06, 'learning_rate/action_model': 2.440801643603933e-05, 'epoch': 0.26}) +[2026-10-03 13:43:57,542][starVLA.training.train_starvla][INFO] - Step 3538, Loss: {'train/loss': 0.0001400836481479928, 'train/grad_norm_pre_clip': 0.061705026775598526, 'timing/data': 0.0003072800027439371, 'timing/model': 0.5793024909944506, 'learning_rate/qwen_vl_interface': 2.4383464819705252e-06, 'learning_rate/action_model': 2.438346481970525e-05, 'epoch': 0.26}) +[2026-10-03 13:43:58,237][starVLA.training.train_starvla][INFO] - Step 3539, Loss: {'train/loss': 0.012246964499354362, 'train/grad_norm_pre_clip': 95.33257293701172, 'timing/data': 0.00019364000763744116, 'timing/model': 0.38986834000388626, 'learning_rate/qwen_vl_interface': 2.435892476101674e-06, 'learning_rate/action_model': 2.435892476101674e-05, 'epoch': 0.26}) +[2026-10-03 13:43:58,933][starVLA.training.train_starvla][INFO] - Step 3540, Loss: {'train/loss': 0.00038704072358086705, 'train/grad_norm_pre_clip': 0.3829628825187683, 'timing/data': 0.00019314000383019447, 'timing/model': 0.38966503999836277, 'learning_rate/qwen_vl_interface': 2.433439627006128e-06, 'learning_rate/action_model': 2.4334396270061277e-05, 'epoch': 0.26}) +[2026-10-03 13:43:59,629][starVLA.training.train_starvla][INFO] - Step 3541, Loss: {'train/loss': 5.120272180647589e-05, 'train/grad_norm_pre_clip': 0.016633056104183197, 'timing/data': 0.00019499000336509198, 'timing/model': 0.3898586299910676, 'learning_rate/qwen_vl_interface': 2.4309879356921602e-06, 'learning_rate/action_model': 2.43098793569216e-05, 'epoch': 0.26}) +[2026-10-03 13:44:00,326][starVLA.training.train_starvla][INFO] - Step 3542, Loss: {'train/loss': 2.7939588562730933e-06, 'train/grad_norm_pre_clip': 0.0005943274009041488, 'timing/data': 0.00021528999786823988, 'timing/model': 0.3900942199979909, 'learning_rate/qwen_vl_interface': 2.428537403167569e-06, 'learning_rate/action_model': 2.4285374031675687e-05, 'epoch': 0.26}) +[2026-10-03 13:44:01,022][starVLA.training.train_starvla][INFO] - Step 3543, Loss: {'train/loss': 6.715080962749198e-05, 'train/grad_norm_pre_clip': 0.053786009550094604, 'timing/data': 0.0001935500040417537, 'timing/model': 0.3899310900014825, 'learning_rate/qwen_vl_interface': 2.4260880304396765e-06, 'learning_rate/action_model': 2.4260880304396765e-05, 'epoch': 0.26}) +[2026-10-03 13:44:01,916][starVLA.training.train_starvla][INFO] - Step 3544, Loss: {'train/loss': 0.0003167852119076997, 'train/grad_norm_pre_clip': 0.28924357891082764, 'timing/data': 0.00022607999562751502, 'timing/model': 0.38931532100832555, 'learning_rate/qwen_vl_interface': 2.4236398185153245e-06, 'learning_rate/action_model': 2.4236398185153245e-05, 'epoch': 0.26}) +[2026-10-03 13:44:02,612][starVLA.training.train_starvla][INFO] - Step 3545, Loss: {'train/loss': 2.6982772396877408e-05, 'train/grad_norm_pre_clip': 0.03273671120405197, 'timing/data': 0.00019286999304313213, 'timing/model': 0.38948834101029206, 'learning_rate/qwen_vl_interface': 2.4211927684008823e-06, 'learning_rate/action_model': 2.4211927684008824e-05, 'epoch': 0.26}) +[2026-10-03 13:44:03,308][starVLA.training.train_starvla][INFO] - Step 3546, Loss: {'train/loss': 1.587652695889119e-05, 'train/grad_norm_pre_clip': 0.037895821034908295, 'timing/data': 0.00019598999642767012, 'timing/model': 0.38970179100579116, 'learning_rate/qwen_vl_interface': 2.4187468811022398e-06, 'learning_rate/action_model': 2.4187468811022396e-05, 'epoch': 0.26}) +[2026-10-03 13:44:04,004][starVLA.training.train_starvla][INFO] - Step 3547, Loss: {'train/loss': 0.00011976945825153962, 'train/grad_norm_pre_clip': 0.09973719716072083, 'timing/data': 0.00019888000679202378, 'timing/model': 0.3896199599985266, 'learning_rate/qwen_vl_interface': 2.416302157624809e-06, 'learning_rate/action_model': 2.4163021576248088e-05, 'epoch': 0.26}) +[2026-10-03 13:44:04,701][starVLA.training.train_starvla][INFO] - Step 3548, Loss: {'train/loss': 0.006013328209519386, 'train/grad_norm_pre_clip': 15.610191345214844, 'timing/data': 0.0001927900011651218, 'timing/model': 0.38997197098797187, 'learning_rate/qwen_vl_interface': 2.4138585989735227e-06, 'learning_rate/action_model': 2.4138585989735226e-05, 'epoch': 0.26}) +[2026-10-03 13:44:05,587][starVLA.training.train_starvla][INFO] - Step 3549, Loss: {'train/loss': 7.371003448497504e-05, 'train/grad_norm_pre_clip': 0.07737593352794647, 'timing/data': 0.00019709000480361283, 'timing/model': 0.5798620419955114, 'learning_rate/qwen_vl_interface': 2.4114162061528357e-06, 'learning_rate/action_model': 2.4114162061528354e-05, 'epoch': 0.26}) +[2026-10-03 13:44:06,285][starVLA.training.train_starvla][INFO] - Step 3550, Loss: {'train/loss': 4.164445635979064e-05, 'train/grad_norm_pre_clip': 0.014053897000849247, 'timing/data': 0.00019561000226531178, 'timing/model': 0.39040581999870483, 'learning_rate/qwen_vl_interface': 2.408974980166724e-06, 'learning_rate/action_model': 2.408974980166724e-05, 'epoch': 0.26}) +[2026-10-03 13:44:06,982][starVLA.training.train_starvla][INFO] - Step 3551, Loss: {'train/loss': 4.91733862872934e-06, 'train/grad_norm_pre_clip': 0.002420109463855624, 'timing/data': 0.00020034000044688582, 'timing/model': 0.3895499310019659, 'learning_rate/qwen_vl_interface': 2.406534922018683e-06, 'learning_rate/action_model': 2.4065349220186827e-05, 'epoch': 0.26}) +[2026-10-03 13:44:07,678][starVLA.training.train_starvla][INFO] - Step 3552, Loss: {'train/loss': 2.570441665739054e-06, 'train/grad_norm_pre_clip': 0.0008907146984711289, 'timing/data': 0.0003313999914098531, 'timing/model': 0.38973568099027034, 'learning_rate/qwen_vl_interface': 2.4040960327117287e-06, 'learning_rate/action_model': 2.4040960327117283e-05, 'epoch': 0.26}) +[2026-10-03 13:44:08,374][starVLA.training.train_starvla][INFO] - Step 3553, Loss: {'train/loss': 0.12324601411819458, 'train/grad_norm_pre_clip': 114.12667846679688, 'timing/data': 0.00026944999990519136, 'timing/model': 0.389331131009385, 'learning_rate/qwen_vl_interface': 2.4016583132483976e-06, 'learning_rate/action_model': 2.4016583132483976e-05, 'epoch': 0.26}) +[2026-10-03 13:44:09,071][starVLA.training.train_starvla][INFO] - Step 3554, Loss: {'train/loss': 0.0008699060417711735, 'train/grad_norm_pre_clip': 0.6504024267196655, 'timing/data': 0.00019631000759545714, 'timing/model': 0.39021739100280683, 'learning_rate/qwen_vl_interface': 2.399221764630741e-06, 'learning_rate/action_model': 2.3992217646307406e-05, 'epoch': 0.26}) +[2026-10-03 13:44:09,959][starVLA.training.train_starvla][INFO] - Step 3555, Loss: {'train/loss': 1.9147944385622395e-06, 'train/grad_norm_pre_clip': 0.0006563881179317832, 'timing/data': 0.00020136999955866486, 'timing/model': 0.38968202100659255, 'learning_rate/qwen_vl_interface': 2.3967863878603334e-06, 'learning_rate/action_model': 2.3967863878603332e-05, 'epoch': 0.26}) +[2026-10-03 13:44:10,655][starVLA.training.train_starvla][INFO] - Step 3556, Loss: {'train/loss': 4.363184416433796e-05, 'train/grad_norm_pre_clip': 0.05876116454601288, 'timing/data': 0.0001978799991775304, 'timing/model': 0.38971047100494616, 'learning_rate/qwen_vl_interface': 2.394352183938266e-06, 'learning_rate/action_model': 2.3943521839382664e-05, 'epoch': 0.26}) +[2026-10-03 13:44:11,352][starVLA.training.train_starvla][INFO] - Step 3557, Loss: {'train/loss': 2.7465452149044722e-05, 'train/grad_norm_pre_clip': 0.025834443047642708, 'timing/data': 0.0003137099993182346, 'timing/model': 0.38952616999449674, 'learning_rate/qwen_vl_interface': 2.39191915386515e-06, 'learning_rate/action_model': 2.39191915386515e-05, 'epoch': 0.26}) +[2026-10-03 13:44:12,049][starVLA.training.train_starvla][INFO] - Step 3558, Loss: {'train/loss': 0.00017965339065995067, 'train/grad_norm_pre_clip': 0.19938673079013824, 'timing/data': 0.00019492000865284353, 'timing/model': 0.3898966209962964, 'learning_rate/qwen_vl_interface': 2.3894872986411095e-06, 'learning_rate/action_model': 2.3894872986411094e-05, 'epoch': 0.26}) +[2026-10-03 13:44:12,745][starVLA.training.train_starvla][INFO] - Step 3559, Loss: {'train/loss': 6.962075713090599e-05, 'train/grad_norm_pre_clip': 0.019944878295063972, 'timing/data': 0.00032931999885477126, 'timing/model': 0.38982637200388126, 'learning_rate/qwen_vl_interface': 2.3870566192657903e-06, 'learning_rate/action_model': 2.38705661926579e-05, 'epoch': 0.26}) +[2026-10-03 13:44:13,636][starVLA.training.train_starvla][INFO] - Step 3560, Loss: {'train/loss': 0.5052710175514221, 'train/grad_norm_pre_clip': 32.183101654052734, 'timing/data': 0.00023119999968912452, 'timing/model': 0.38893472199561074, 'learning_rate/qwen_vl_interface': 2.3846271167383515e-06, 'learning_rate/action_model': 2.3846271167383515e-05, 'epoch': 0.26}) +[2026-10-03 13:44:14,332][starVLA.training.train_starvla][INFO] - Step 3561, Loss: {'train/loss': 0.00039988860953599215, 'train/grad_norm_pre_clip': 0.3260171115398407, 'timing/data': 0.00019366999913472682, 'timing/model': 0.38955541100585833, 'learning_rate/qwen_vl_interface': 2.382198792057471e-06, 'learning_rate/action_model': 2.3821987920574707e-05, 'epoch': 0.26}) +[2026-10-03 13:44:15,028][starVLA.training.train_starvla][INFO] - Step 3562, Loss: {'train/loss': 2.0018425857415423e-05, 'train/grad_norm_pre_clip': 0.006650798488408327, 'timing/data': 0.0001951099984580651, 'timing/model': 0.38985627199872397, 'learning_rate/qwen_vl_interface': 2.3797716462213402e-06, 'learning_rate/action_model': 2.3797716462213402e-05, 'epoch': 0.26}) +[2026-10-03 13:44:15,724][starVLA.training.train_starvla][INFO] - Step 3563, Loss: {'train/loss': 5.356893780117389e-06, 'train/grad_norm_pre_clip': 0.0013242056593298912, 'timing/data': 0.0002849000011337921, 'timing/model': 0.3894473910040688, 'learning_rate/qwen_vl_interface': 2.37734568022767e-06, 'learning_rate/action_model': 2.3773456802276697e-05, 'epoch': 0.26}) +[2026-10-03 13:44:16,420][starVLA.training.train_starvla][INFO] - Step 3564, Loss: {'train/loss': 1.3634526112582535e-06, 'train/grad_norm_pre_clip': 0.00023041761596687138, 'timing/data': 0.0003052199899684638, 'timing/model': 0.3893062910065055, 'learning_rate/qwen_vl_interface': 2.3749208950736783e-06, 'learning_rate/action_model': 2.374920895073678e-05, 'epoch': 0.26}) +[2026-10-03 13:44:17,309][starVLA.training.train_starvla][INFO] - Step 3565, Loss: {'train/loss': 0.42034217715263367, 'train/grad_norm_pre_clip': 73.0039291381836, 'timing/data': 0.00019582000095397234, 'timing/model': 0.5822530730074504, 'learning_rate/qwen_vl_interface': 2.372497291756106e-06, 'learning_rate/action_model': 2.3724972917561058e-05, 'epoch': 0.26}) +[2026-10-03 13:44:18,005][starVLA.training.train_starvla][INFO] - Step 3566, Loss: {'train/loss': 5.952952051302418e-06, 'train/grad_norm_pre_clip': 0.0014566117897629738, 'timing/data': 0.00031814999238122255, 'timing/model': 0.38932404200022575, 'learning_rate/qwen_vl_interface': 2.3700748712712043e-06, 'learning_rate/action_model': 2.370074871271204e-05, 'epoch': 0.26}) +[2026-10-03 13:44:18,699][starVLA.training.train_starvla][INFO] - Step 3567, Loss: {'train/loss': 1.6838286001075176e-06, 'train/grad_norm_pre_clip': 0.0005037327646277845, 'timing/data': 0.00031566999678034335, 'timing/model': 0.3888561310013756, 'learning_rate/qwen_vl_interface': 2.3676536346147388e-06, 'learning_rate/action_model': 2.3676536346147388e-05, 'epoch': 0.26}) +[2026-10-03 13:44:19,394][starVLA.training.train_starvla][INFO] - Step 3568, Loss: {'train/loss': 2.168110768252518e-06, 'train/grad_norm_pre_clip': 0.000628078356385231, 'timing/data': 0.00028534000739455223, 'timing/model': 0.3887956109974766, 'learning_rate/qwen_vl_interface': 2.365233582781988e-06, 'learning_rate/action_model': 2.365233582781988e-05, 'epoch': 0.26}) +[2026-10-03 13:44:20,088][starVLA.training.train_starvla][INFO] - Step 3569, Loss: {'train/loss': 3.740173042388051e-06, 'train/grad_norm_pre_clip': 0.0007528559071943164, 'timing/data': 0.00032091999310068786, 'timing/model': 0.38818582199746743, 'learning_rate/qwen_vl_interface': 2.3628147167677444e-06, 'learning_rate/action_model': 2.3628147167677443e-05, 'epoch': 0.26}) +[2026-10-03 13:44:20,784][starVLA.training.train_starvla][INFO] - Step 3570, Loss: {'train/loss': 0.0033070126082748175, 'train/grad_norm_pre_clip': 8.00737190246582, 'timing/data': 0.0003309700114186853, 'timing/model': 0.388600491991383, 'learning_rate/qwen_vl_interface': 2.3603970375663123e-06, 'learning_rate/action_model': 2.3603970375663125e-05, 'epoch': 0.26}) +[2026-10-03 13:44:21,669][starVLA.training.train_starvla][INFO] - Step 3571, Loss: {'train/loss': 1.583909397595562e-05, 'train/grad_norm_pre_clip': 0.0066800788044929504, 'timing/data': 0.00020358999609015882, 'timing/model': 0.3885479309974471, 'learning_rate/qwen_vl_interface': 2.35798054617151e-06, 'learning_rate/action_model': 2.3579805461715096e-05, 'epoch': 0.26}) +[2026-10-03 13:44:22,365][starVLA.training.train_starvla][INFO] - Step 3572, Loss: {'train/loss': 2.7659898478304967e-05, 'train/grad_norm_pre_clip': 0.049908459186553955, 'timing/data': 0.00020031999156344682, 'timing/model': 0.38891519200115, 'learning_rate/qwen_vl_interface': 2.355565243576662e-06, 'learning_rate/action_model': 2.3555652435766617e-05, 'epoch': 0.26}) +[2026-10-03 13:44:23,061][starVLA.training.train_starvla][INFO] - Step 3573, Loss: {'train/loss': 2.5762530640349723e-05, 'train/grad_norm_pre_clip': 0.007837312296032906, 'timing/data': 0.0003395100065972656, 'timing/model': 0.38916897200397216, 'learning_rate/qwen_vl_interface': 2.35315113077461e-06, 'learning_rate/action_model': 2.3531511307746098e-05, 'epoch': 0.26}) +[2026-10-03 13:44:23,759][starVLA.training.train_starvla][INFO] - Step 3574, Loss: {'train/loss': 3.874261892633513e-06, 'train/grad_norm_pre_clip': 0.0014096434460952878, 'timing/data': 0.00020735000725835562, 'timing/model': 0.3906574319989886, 'learning_rate/qwen_vl_interface': 2.3507382087577047e-06, 'learning_rate/action_model': 2.350738208757705e-05, 'epoch': 0.26}) +[2026-10-03 13:44:24,456][starVLA.training.train_starvla][INFO] - Step 3575, Loss: {'train/loss': 0.00023234140826389194, 'train/grad_norm_pre_clip': 0.1516723334789276, 'timing/data': 0.00032325999927707016, 'timing/model': 0.38971836199925747, 'learning_rate/qwen_vl_interface': 2.348326478517807e-06, 'learning_rate/action_model': 2.3483264785178072e-05, 'epoch': 0.26}) +[2026-10-03 13:44:25,357][starVLA.training.train_starvla][INFO] - Step 3576, Loss: {'train/loss': 0.5425989627838135, 'train/grad_norm_pre_clip': 93.64987182617188, 'timing/data': 0.00020275999850127846, 'timing/model': 0.38900106299843173, 'learning_rate/qwen_vl_interface': 2.3459159410462907e-06, 'learning_rate/action_model': 2.3459159410462905e-05, 'epoch': 0.26}) +[2026-10-03 13:44:26,056][starVLA.training.train_starvla][INFO] - Step 3577, Loss: {'train/loss': 1.1190498298674356e-05, 'train/grad_norm_pre_clip': 0.005147662479430437, 'timing/data': 0.0003387400065548718, 'timing/model': 0.391506311003468, 'learning_rate/qwen_vl_interface': 2.343506597334031e-06, 'learning_rate/action_model': 2.343506597334031e-05, 'epoch': 0.26}) +[2026-10-03 13:44:26,753][starVLA.training.train_starvla][INFO] - Step 3578, Loss: {'train/loss': 0.0022179451771080494, 'train/grad_norm_pre_clip': 6.364645004272461, 'timing/data': 0.0003400999994482845, 'timing/model': 0.38940920200548135, 'learning_rate/qwen_vl_interface': 2.341098448371423e-06, 'learning_rate/action_model': 2.3410984483714227e-05, 'epoch': 0.26}) +[2026-10-03 13:44:27,451][starVLA.training.train_starvla][INFO] - Step 3579, Loss: {'train/loss': 2.436329623378697e-06, 'train/grad_norm_pre_clip': 0.000748884049244225, 'timing/data': 0.00027081998996436596, 'timing/model': 0.39031919099215884, 'learning_rate/qwen_vl_interface': 2.338691495148363e-06, 'learning_rate/action_model': 2.3386914951483626e-05, 'epoch': 0.26}) +[2026-10-03 13:44:28,150][starVLA.training.train_starvla][INFO] - Step 3580, Loss: {'train/loss': 0.00039337947964668274, 'train/grad_norm_pre_clip': 0.6887006759643555, 'timing/data': 0.0003306599974166602, 'timing/model': 0.3892747520003468, 'learning_rate/qwen_vl_interface': 2.33628573865426e-06, 'learning_rate/action_model': 2.3362857386542596e-05, 'epoch': 0.26}) +[2026-10-03 13:44:29,062][starVLA.training.train_starvla][INFO] - Step 3581, Loss: {'train/loss': 1.7649668734520674e-05, 'train/grad_norm_pre_clip': 0.0034698396921157837, 'timing/data': 0.00021399999968707561, 'timing/model': 0.5994309899979271, 'learning_rate/qwen_vl_interface': 2.3338811798780282e-06, 'learning_rate/action_model': 2.333881179878028e-05, 'epoch': 0.26}) +[2026-10-03 13:44:29,757][starVLA.training.train_starvla][INFO] - Step 3582, Loss: {'train/loss': 0.2603326737880707, 'train/grad_norm_pre_clip': 38.575679779052734, 'timing/data': 0.00036143999022897333, 'timing/model': 0.3892924120009411, 'learning_rate/qwen_vl_interface': 2.3314778198080933e-06, 'learning_rate/action_model': 2.331477819808093e-05, 'epoch': 0.26}) +[2026-10-03 13:44:30,452][starVLA.training.train_starvla][INFO] - Step 3583, Loss: {'train/loss': 0.0001214483127114363, 'train/grad_norm_pre_clip': 0.10051009804010391, 'timing/data': 0.00019922999490518123, 'timing/model': 0.388956792987301, 'learning_rate/qwen_vl_interface': 2.3290756594323834e-06, 'learning_rate/action_model': 2.3290756594323836e-05, 'epoch': 0.26}) +[2026-10-03 13:44:31,148][starVLA.training.train_starvla][INFO] - Step 3584, Loss: {'train/loss': 0.00010829661187017336, 'train/grad_norm_pre_clip': 0.18002448976039886, 'timing/data': 0.0003026199992746115, 'timing/model': 0.38870054300059564, 'learning_rate/qwen_vl_interface': 2.3266746997383378e-06, 'learning_rate/action_model': 2.3266746997383377e-05, 'epoch': 0.26}) +[2026-10-03 13:44:31,842][starVLA.training.train_starvla][INFO] - Step 3585, Loss: {'train/loss': 0.12077400833368301, 'train/grad_norm_pre_clip': 222.4696807861328, 'timing/data': 0.0003264900005888194, 'timing/model': 0.3880300820019329, 'learning_rate/qwen_vl_interface': 2.324274941712899e-06, 'learning_rate/action_model': 2.324274941712899e-05, 'epoch': 0.26}) +[2026-10-03 13:44:32,731][starVLA.training.train_starvla][INFO] - Step 3586, Loss: {'train/loss': 0.002018213737756014, 'train/grad_norm_pre_clip': 13.535962104797363, 'timing/data': 0.00031820000731386244, 'timing/model': 0.5826561930007301, 'learning_rate/qwen_vl_interface': 2.321876386342519e-06, 'learning_rate/action_model': 2.3218763863425187e-05, 'epoch': 0.26}) +[2026-10-03 13:44:33,424][starVLA.training.train_starvla][INFO] - Step 3587, Loss: {'train/loss': 0.0010850998805835843, 'train/grad_norm_pre_clip': 4.2249755859375, 'timing/data': 0.00026875000912696123, 'timing/model': 0.38825087300210726, 'learning_rate/qwen_vl_interface': 2.31947903461315e-06, 'learning_rate/action_model': 2.3194790346131502e-05, 'epoch': 0.26}) +[2026-10-03 13:44:34,120][starVLA.training.train_starvla][INFO] - Step 3588, Loss: {'train/loss': 1.249397246283479e-05, 'train/grad_norm_pre_clip': 0.006455693393945694, 'timing/data': 0.00020330000552348793, 'timing/model': 0.389774982992094, 'learning_rate/qwen_vl_interface': 2.3170828875102544e-06, 'learning_rate/action_model': 2.3170828875102543e-05, 'epoch': 0.26}) +[2026-10-03 13:44:34,818][starVLA.training.train_starvla][INFO] - Step 3589, Loss: {'train/loss': 7.363813347183168e-05, 'train/grad_norm_pre_clip': 0.0405784510076046, 'timing/data': 0.00019733000954147428, 'timing/model': 0.3906809920008527, 'learning_rate/qwen_vl_interface': 2.3146879460187984e-06, 'learning_rate/action_model': 2.3146879460187982e-05, 'epoch': 0.26}) +[2026-10-03 13:44:35,517][starVLA.training.train_starvla][INFO] - Step 3590, Loss: {'train/loss': 4.113430986762978e-05, 'train/grad_norm_pre_clip': 0.040052033960819244, 'timing/data': 0.0002442499971948564, 'timing/model': 0.3903721829992719, 'learning_rate/qwen_vl_interface': 2.3122942111232506e-06, 'learning_rate/action_model': 2.3122942111232506e-05, 'epoch': 0.26}) +[2026-10-03 13:44:36,417][starVLA.training.train_starvla][INFO] - Step 3591, Loss: {'train/loss': 0.0035188915207982063, 'train/grad_norm_pre_clip': 5.0520782470703125, 'timing/data': 0.0003097700100624934, 'timing/model': 0.5927140819985652, 'learning_rate/qwen_vl_interface': 2.3099016838075857e-06, 'learning_rate/action_model': 2.3099016838075856e-05, 'epoch': 0.26}) +[2026-10-03 13:44:37,114][starVLA.training.train_starvla][INFO] - Step 3592, Loss: {'train/loss': 5.871010216651484e-06, 'train/grad_norm_pre_clip': 0.0017610711511224508, 'timing/data': 0.0003013399982592091, 'timing/model': 0.3909097220021067, 'learning_rate/qwen_vl_interface': 2.307510365055282e-06, 'learning_rate/action_model': 2.307510365055282e-05, 'epoch': 0.26}) +[2026-10-03 13:44:37,813][starVLA.training.train_starvla][INFO] - Step 3593, Loss: {'train/loss': 0.12462736666202545, 'train/grad_norm_pre_clip': 127.55301666259766, 'timing/data': 0.0002140300057362765, 'timing/model': 0.39060068198887166, 'learning_rate/qwen_vl_interface': 2.3051202558493196e-06, 'learning_rate/action_model': 2.3051202558493192e-05, 'epoch': 0.26}) +[2026-10-03 13:44:38,509][starVLA.training.train_starvla][INFO] - Step 3594, Loss: {'train/loss': 0.002628597430884838, 'train/grad_norm_pre_clip': 4.857202053070068, 'timing/data': 0.00031612000020686537, 'timing/model': 0.3890398830117192, 'learning_rate/qwen_vl_interface': 2.302731357172182e-06, 'learning_rate/action_model': 2.302731357172182e-05, 'epoch': 0.26}) +[2026-10-03 13:44:39,206][starVLA.training.train_starvla][INFO] - Step 3595, Loss: {'train/loss': 0.0011611698428168893, 'train/grad_norm_pre_clip': 0.7162578105926514, 'timing/data': 0.0003424900060053915, 'timing/model': 0.3892933129973244, 'learning_rate/qwen_vl_interface': 2.3003436700058556e-06, 'learning_rate/action_model': 2.3003436700058554e-05, 'epoch': 0.26}) +[2026-10-03 13:44:39,902][starVLA.training.train_starvla][INFO] - Step 3596, Loss: {'train/loss': 0.008218564093112946, 'train/grad_norm_pre_clip': 7.518558502197266, 'timing/data': 0.0002552700025262311, 'timing/model': 0.3895320229930803, 'learning_rate/qwen_vl_interface': 2.297957195331828e-06, 'learning_rate/action_model': 2.2979571953318276e-05, 'epoch': 0.26}) +[2026-10-03 13:44:40,818][starVLA.training.train_starvla][INFO] - Step 3597, Loss: {'train/loss': 2.1112666217959486e-05, 'train/grad_norm_pre_clip': 0.007408061530441046, 'timing/data': 0.0002528100012568757, 'timing/model': 0.3897913129912922, 'learning_rate/qwen_vl_interface': 2.2955719341310903e-06, 'learning_rate/action_model': 2.2955719341310903e-05, 'epoch': 0.26}) +[2026-10-03 13:44:41,515][starVLA.training.train_starvla][INFO] - Step 3598, Loss: {'train/loss': 5.848645287187537e-06, 'train/grad_norm_pre_clip': 0.0017396690091118217, 'timing/data': 0.0002878399973269552, 'timing/model': 0.38970666300156154, 'learning_rate/qwen_vl_interface': 2.2931878873841304e-06, 'learning_rate/action_model': 2.2931878873841302e-05, 'epoch': 0.26}) +[2026-10-03 13:44:42,211][starVLA.training.train_starvla][INFO] - Step 3599, Loss: {'train/loss': 4.8438276280649006e-05, 'train/grad_norm_pre_clip': 0.038725100457668304, 'timing/data': 0.00020188999769743532, 'timing/model': 0.3893590929947095, 'learning_rate/qwen_vl_interface': 2.290805056070942e-06, 'learning_rate/action_model': 2.290805056070942e-05, 'epoch': 0.26}) +[2026-10-03 13:44:42,911][starVLA.training.train_starvla][INFO] - Step 3600, Loss: {'train/loss': 6.9810957938898355e-06, 'train/grad_norm_pre_clip': 0.0028680970426648855, 'timing/data': 0.0002407300053164363, 'timing/model': 0.3919241730036447, 'learning_rate/qwen_vl_interface': 2.288423441171016e-06, 'learning_rate/action_model': 2.288423441171016e-05, 'epoch': 0.26}) +[2026-10-03 13:44:43,609][starVLA.training.train_starvla][INFO] - Step 3601, Loss: {'train/loss': 1.244245027010038e-06, 'train/grad_norm_pre_clip': 0.00037377001717686653, 'timing/data': 0.00020978999964427203, 'timing/model': 0.39029466299689375, 'learning_rate/qwen_vl_interface': 2.2860430436633447e-06, 'learning_rate/action_model': 2.2860430436633447e-05, 'epoch': 0.26}) +[2026-10-03 13:44:44,511][starVLA.training.train_starvla][INFO] - Step 3602, Loss: {'train/loss': 0.0011572480434551835, 'train/grad_norm_pre_clip': 0.45702528953552246, 'timing/data': 0.0003198500053258613, 'timing/model': 0.5947718919924228, 'learning_rate/qwen_vl_interface': 2.28366386452642e-06, 'learning_rate/action_model': 2.28366386452642e-05, 'epoch': 0.26}) +[2026-10-03 13:44:45,208][starVLA.training.train_starvla][INFO] - Step 3603, Loss: {'train/loss': 0.029067065566778183, 'train/grad_norm_pre_clip': 198.87091064453125, 'timing/data': 0.0003174699959345162, 'timing/model': 0.3910747730114963, 'learning_rate/qwen_vl_interface': 2.2812859047382326e-06, 'learning_rate/action_model': 2.2812859047382323e-05, 'epoch': 0.26}) +[2026-10-03 13:44:45,906][starVLA.training.train_starvla][INFO] - Step 3604, Loss: {'train/loss': 2.6374877961643506e-06, 'train/grad_norm_pre_clip': 0.0005616460111923516, 'timing/data': 0.0003601099888328463, 'timing/model': 0.3900214429886546, 'learning_rate/qwen_vl_interface': 2.2789091652762712e-06, 'learning_rate/action_model': 2.2789091652762712e-05, 'epoch': 0.26}) +[2026-10-03 13:44:46,603][starVLA.training.train_starvla][INFO] - Step 3605, Loss: {'train/loss': 0.09966161847114563, 'train/grad_norm_pre_clip': 600.6533203125, 'timing/data': 0.0003346399898873642, 'timing/model': 0.3892635529919062, 'learning_rate/qwen_vl_interface': 2.276533647117524e-06, 'learning_rate/action_model': 2.2765336471175242e-05, 'epoch': 0.26}) +[2026-10-03 13:44:47,301][starVLA.training.train_starvla][INFO] - Step 3606, Loss: {'train/loss': 1.2777263691532426e-05, 'train/grad_norm_pre_clip': 0.004647191613912582, 'timing/data': 0.0003566800005501136, 'timing/model': 0.39036799300811253, 'learning_rate/qwen_vl_interface': 2.2741593512384773e-06, 'learning_rate/action_model': 2.2741593512384772e-05, 'epoch': 0.26}) +[2026-10-03 13:44:48,000][starVLA.training.train_starvla][INFO] - Step 3607, Loss: {'train/loss': 1.3388220395427197e-05, 'train/grad_norm_pre_clip': 0.0024432437494397163, 'timing/data': 0.00032856999314390123, 'timing/model': 0.3918189620017074, 'learning_rate/qwen_vl_interface': 2.2717862786151153e-06, 'learning_rate/action_model': 2.271786278615115e-05, 'epoch': 0.26}) +[2026-10-03 13:44:48,904][starVLA.training.train_starvla][INFO] - Step 3608, Loss: {'train/loss': 0.5262681841850281, 'train/grad_norm_pre_clip': 66.19624328613281, 'timing/data': 0.00021340000967029482, 'timing/model': 0.38764569400518667, 'learning_rate/qwen_vl_interface': 2.2694144302229152e-06, 'learning_rate/action_model': 2.2694144302229152e-05, 'epoch': 0.26}) +[2026-10-03 13:44:49,601][starVLA.training.train_starvla][INFO] - Step 3609, Loss: {'train/loss': 1.2382446584524587e-05, 'train/grad_norm_pre_clip': 0.0030026102904230356, 'timing/data': 0.00021096999989822507, 'timing/model': 0.3895697029947769, 'learning_rate/qwen_vl_interface': 2.2670438070368568e-06, 'learning_rate/action_model': 2.2670438070368566e-05, 'epoch': 0.26}) +[2026-10-03 13:44:50,297][starVLA.training.train_starvla][INFO] - Step 3610, Loss: {'train/loss': 2.4512319214409217e-06, 'train/grad_norm_pre_clip': 0.0005995735991746187, 'timing/data': 0.00021660000493284315, 'timing/model': 0.3899015130009502, 'learning_rate/qwen_vl_interface': 2.264674410031413e-06, 'learning_rate/action_model': 2.2646744100314127e-05, 'epoch': 0.26}) +[2026-10-03 13:44:50,999][starVLA.training.train_starvla][INFO] - Step 3611, Loss: {'train/loss': 7.338695468206424e-06, 'train/grad_norm_pre_clip': 0.0022692654747515917, 'timing/data': 0.00024893999216146767, 'timing/model': 0.3922273829957703, 'learning_rate/qwen_vl_interface': 2.2623062401805537e-06, 'learning_rate/action_model': 2.2623062401805535e-05, 'epoch': 0.26}) +[2026-10-03 13:44:51,697][starVLA.training.train_starvla][INFO] - Step 3612, Loss: {'train/loss': 2.59843363892287e-05, 'train/grad_norm_pre_clip': 0.02195303700864315, 'timing/data': 0.0003328790044179186, 'timing/model': 0.38984299400181044, 'learning_rate/qwen_vl_interface': 2.2599392984577436e-06, 'learning_rate/action_model': 2.2599392984577438e-05, 'epoch': 0.26}) +[2026-10-03 13:44:52,599][starVLA.training.train_starvla][INFO] - Step 3613, Loss: {'train/loss': 1.1138179615954868e-05, 'train/grad_norm_pre_clip': 0.007917800918221474, 'timing/data': 0.00020589999621734023, 'timing/model': 0.38904068400734104, 'learning_rate/qwen_vl_interface': 2.2575735858359436e-06, 'learning_rate/action_model': 2.257573585835944e-05, 'epoch': 0.26}) +[2026-10-03 13:44:53,300][starVLA.training.train_starvla][INFO] - Step 3614, Loss: {'train/loss': 0.0012498932192102075, 'train/grad_norm_pre_clip': 2.9369940757751465, 'timing/data': 0.00035852000291924924, 'timing/model': 0.3920139030087739, 'learning_rate/qwen_vl_interface': 2.255209103287609e-06, 'learning_rate/action_model': 2.255209103287609e-05, 'epoch': 0.26}) +[2026-10-03 13:44:53,997][starVLA.training.train_starvla][INFO] - Step 3615, Loss: {'train/loss': 0.00040485538193024695, 'train/grad_norm_pre_clip': 3.792372226715088, 'timing/data': 0.00035196999670006335, 'timing/model': 0.38919526401150506, 'learning_rate/qwen_vl_interface': 2.2528458517846886e-06, 'learning_rate/action_model': 2.252845851784689e-05, 'epoch': 0.26}) +[2026-10-03 13:44:54,692][starVLA.training.train_starvla][INFO] - Step 3616, Loss: {'train/loss': 1.5816662198631093e-05, 'train/grad_norm_pre_clip': 0.004109700210392475, 'timing/data': 0.0003162200009683147, 'timing/model': 0.3891992639983073, 'learning_rate/qwen_vl_interface': 2.250483832298627e-06, 'learning_rate/action_model': 2.250483832298627e-05, 'epoch': 0.26}) +[2026-10-03 13:44:55,387][starVLA.training.train_starvla][INFO] - Step 3617, Loss: {'train/loss': 1.9147908005834324e-06, 'train/grad_norm_pre_clip': 0.0003237674536649138, 'timing/data': 0.00030128999787848443, 'timing/model': 0.3882719240064034, 'learning_rate/qwen_vl_interface': 2.2481230458003608e-06, 'learning_rate/action_model': 2.2481230458003606e-05, 'epoch': 0.26}) +[2026-10-03 13:44:56,300][starVLA.training.train_starvla][INFO] - Step 3618, Loss: {'train/loss': 0.00016836034774314612, 'train/grad_norm_pre_clip': 0.18846416473388672, 'timing/data': 0.00024355000641662627, 'timing/model': 0.6076089910056908, 'learning_rate/qwen_vl_interface': 2.2457634932603184e-06, 'learning_rate/action_model': 2.2457634932603183e-05, 'epoch': 0.26}) +[2026-10-03 13:44:57,001][starVLA.training.train_starvla][INFO] - Step 3619, Loss: {'train/loss': 1.1473865697553265e-06, 'train/grad_norm_pre_clip': 0.0002688198001123965, 'timing/data': 0.0002734099980443716, 'timing/model': 0.39133157300238963, 'learning_rate/qwen_vl_interface': 2.2434051756484255e-06, 'learning_rate/action_model': 2.2434051756484253e-05, 'epoch': 0.26}) +[2026-10-03 13:44:57,701][starVLA.training.train_starvla][INFO] - Step 3620, Loss: {'train/loss': 3.0385588615899906e-05, 'train/grad_norm_pre_clip': 0.035077132284641266, 'timing/data': 0.00020579001284204423, 'timing/model': 0.3914411639998434, 'learning_rate/qwen_vl_interface': 2.2410480939340945e-06, 'learning_rate/action_model': 2.2410480939340945e-05, 'epoch': 0.26}) +[2026-10-03 13:44:58,401][starVLA.training.train_starvla][INFO] - Step 3621, Loss: {'train/loss': 1.2255655747139826e-05, 'train/grad_norm_pre_clip': 0.008579356595873833, 'timing/data': 0.0003265699924668297, 'timing/model': 0.3911121639976045, 'learning_rate/qwen_vl_interface': 2.238692249086234e-06, 'learning_rate/action_model': 2.2386922490862336e-05, 'epoch': 0.27}) +[2026-10-03 13:44:59,105][starVLA.training.train_starvla][INFO] - Step 3622, Loss: {'train/loss': 3.3601882023504004e-06, 'train/grad_norm_pre_clip': 0.0008888221345841885, 'timing/data': 0.00035797999589703977, 'timing/model': 0.39312788301322144, 'learning_rate/qwen_vl_interface': 2.2363376420732426e-06, 'learning_rate/action_model': 2.2363376420732423e-05, 'epoch': 0.27}) +[2026-10-03 13:44:59,805][starVLA.training.train_starvla][INFO] - Step 3623, Loss: {'train/loss': 3.3154883567476645e-06, 'train/grad_norm_pre_clip': 0.00099339394364506, 'timing/data': 0.00028476999432314187, 'timing/model': 0.3918821929983096, 'learning_rate/qwen_vl_interface': 2.233984273863011e-06, 'learning_rate/action_model': 2.233984273863011e-05, 'epoch': 0.27}) +[2026-10-03 13:45:00,702][starVLA.training.train_starvla][INFO] - Step 3624, Loss: {'train/loss': 0.00016259515541605651, 'train/grad_norm_pre_clip': 0.2816007435321808, 'timing/data': 0.0003073800035053864, 'timing/model': 0.3913604539993685, 'learning_rate/qwen_vl_interface': 2.231632145422919e-06, 'learning_rate/action_model': 2.231632145422919e-05, 'epoch': 0.27}) +[2026-10-03 13:45:01,406][starVLA.training.train_starvla][INFO] - Step 3625, Loss: {'train/loss': 5.5804275689297356e-06, 'train/grad_norm_pre_clip': 0.0015398479299619794, 'timing/data': 0.00034732000494841486, 'timing/model': 0.3937436630076263, 'learning_rate/qwen_vl_interface': 2.2292812577198383e-06, 'learning_rate/action_model': 2.229281257719838e-05, 'epoch': 0.27}) +[2026-10-03 13:45:02,105][starVLA.training.train_starvla][INFO] - Step 3626, Loss: {'train/loss': 0.00017589391791261733, 'train/grad_norm_pre_clip': 0.3349476456642151, 'timing/data': 0.0002589189971331507, 'timing/model': 0.39058916400244925, 'learning_rate/qwen_vl_interface': 2.2269316117201307e-06, 'learning_rate/action_model': 2.2269316117201307e-05, 'epoch': 0.27}) +[2026-10-03 13:45:02,805][starVLA.training.train_starvla][INFO] - Step 3627, Loss: {'train/loss': 6.489300176326651e-06, 'train/grad_norm_pre_clip': 0.003170831361785531, 'timing/data': 0.00032214999373536557, 'timing/model': 0.3915385540021816, 'learning_rate/qwen_vl_interface': 2.224583208389646e-06, 'learning_rate/action_model': 2.224583208389646e-05, 'epoch': 0.27}) +[2026-10-03 13:45:03,503][starVLA.training.train_starvla][INFO] - Step 3628, Loss: {'train/loss': 0.36694467067718506, 'train/grad_norm_pre_clip': 115.89965057373047, 'timing/data': 0.00019906999659724534, 'timing/model': 0.39251440299267415, 'learning_rate/qwen_vl_interface': 2.2222360486937238e-06, 'learning_rate/action_model': 2.2222360486937235e-05, 'epoch': 0.27}) +[2026-10-03 13:45:04,411][starVLA.training.train_starvla][INFO] - Step 3629, Loss: {'train/loss': 2.9653110686922446e-06, 'train/grad_norm_pre_clip': 0.0007250849739648402, 'timing/data': 0.0003501699975458905, 'timing/model': 0.39246272399032023, 'learning_rate/qwen_vl_interface': 2.219890133597194e-06, 'learning_rate/action_model': 2.219890133597194e-05, 'epoch': 0.27}) +[2026-10-03 13:45:05,110][starVLA.training.train_starvla][INFO] - Step 3630, Loss: {'train/loss': 0.001205433625727892, 'train/grad_norm_pre_clip': 0.7353759407997131, 'timing/data': 0.0003278600051999092, 'timing/model': 0.3910615839995444, 'learning_rate/qwen_vl_interface': 2.217545464064374e-06, 'learning_rate/action_model': 2.217545464064374e-05, 'epoch': 0.27}) +[2026-10-03 13:45:05,810][starVLA.training.train_starvla][INFO] - Step 3631, Loss: {'train/loss': 3.599653427954763e-05, 'train/grad_norm_pre_clip': 0.017079602926969528, 'timing/data': 0.000329189992044121, 'timing/model': 0.39110741300100926, 'learning_rate/qwen_vl_interface': 2.215202041059067e-06, 'learning_rate/action_model': 2.215202041059067e-05, 'epoch': 0.27}) +[2026-10-03 13:45:06,511][starVLA.training.train_starvla][INFO] - Step 3632, Loss: {'train/loss': 4.872576937486883e-06, 'train/grad_norm_pre_clip': 0.0012193801812827587, 'timing/data': 0.00030544000037480146, 'timing/model': 0.3927242440113332, 'learning_rate/qwen_vl_interface': 2.212859865544566e-06, 'learning_rate/action_model': 2.212859865544566e-05, 'epoch': 0.27}) +[2026-10-03 13:45:07,214][starVLA.training.train_starvla][INFO] - Step 3633, Loss: {'train/loss': 5.632590273307869e-06, 'train/grad_norm_pre_clip': 0.002667965367436409, 'timing/data': 0.00023563999275211245, 'timing/model': 0.3922779940039618, 'learning_rate/qwen_vl_interface': 2.210518938483652e-06, 'learning_rate/action_model': 2.210518938483652e-05, 'epoch': 0.27}) +[2026-10-03 13:45:08,111][starVLA.training.train_starvla][INFO] - Step 3634, Loss: {'train/loss': 2.4512078198313247e-06, 'train/grad_norm_pre_clip': 0.0024589458480477333, 'timing/data': 0.00020098900131415576, 'timing/model': 0.3892203249997692, 'learning_rate/qwen_vl_interface': 2.20817926083859e-06, 'learning_rate/action_model': 2.20817926083859e-05, 'epoch': 0.27}) +[2026-10-03 13:45:08,807][starVLA.training.train_starvla][INFO] - Step 3635, Loss: {'train/loss': 4.351057214080356e-06, 'train/grad_norm_pre_clip': 0.004387492313981056, 'timing/data': 0.0003902199969161302, 'timing/model': 0.389255593996495, 'learning_rate/qwen_vl_interface': 2.205840833571135e-06, 'learning_rate/action_model': 2.205840833571135e-05, 'epoch': 0.27}) +[2026-10-03 13:45:09,507][starVLA.training.train_starvla][INFO] - Step 3636, Loss: {'train/loss': 3.97857183997985e-06, 'train/grad_norm_pre_clip': 0.000912271614652127, 'timing/data': 0.00034998000774066895, 'timing/model': 0.3905441440001596, 'learning_rate/qwen_vl_interface': 2.2035036576425264e-06, 'learning_rate/action_model': 2.203503657642526e-05, 'epoch': 0.27}) +[2026-10-03 13:45:10,204][starVLA.training.train_starvla][INFO] - Step 3637, Loss: {'train/loss': 2.436319391563302e-06, 'train/grad_norm_pre_clip': 0.0005690714460797608, 'timing/data': 0.0003255800111219287, 'timing/model': 0.38963668400538154, 'learning_rate/qwen_vl_interface': 2.2011677340134867e-06, 'learning_rate/action_model': 2.2011677340134867e-05, 'epoch': 0.27}) +[2026-10-03 13:45:10,901][starVLA.training.train_starvla][INFO] - Step 3638, Loss: {'train/loss': 1.7285287867707666e-06, 'train/grad_norm_pre_clip': 0.0005499293329194188, 'timing/data': 0.0002366800035815686, 'timing/model': 0.3902590139914537, 'learning_rate/qwen_vl_interface': 2.198833063644227e-06, 'learning_rate/action_model': 2.198833063644227e-05, 'epoch': 0.27}) +[2026-10-03 13:45:11,806][starVLA.training.train_starvla][INFO] - Step 3639, Loss: {'train/loss': 2.937358476629015e-05, 'train/grad_norm_pre_clip': 0.03054673597216606, 'timing/data': 0.0003386000025784597, 'timing/model': 0.5962909139925614, 'learning_rate/qwen_vl_interface': 2.196499647494443e-06, 'learning_rate/action_model': 2.196499647494443e-05, 'epoch': 0.27}) +[2026-10-03 13:45:12,506][starVLA.training.train_starvla][INFO] - Step 3640, Loss: {'train/loss': 3.49542242474854e-05, 'train/grad_norm_pre_clip': 0.027960672974586487, 'timing/data': 0.0002463300043018535, 'timing/model': 0.39152189399464987, 'learning_rate/qwen_vl_interface': 2.1941674865233144e-06, 'learning_rate/action_model': 2.1941674865233143e-05, 'epoch': 0.27}) +[2026-10-03 13:45:13,202][starVLA.training.train_starvla][INFO] - Step 3641, Loss: {'train/loss': 4.3836062104674056e-05, 'train/grad_norm_pre_clip': 0.038438573479652405, 'timing/data': 0.00029996999364811927, 'timing/model': 0.3890011140028946, 'learning_rate/qwen_vl_interface': 2.1918365816895013e-06, 'learning_rate/action_model': 2.191836581689501e-05, 'epoch': 0.27}) +[2026-10-03 13:45:13,899][starVLA.training.train_starvla][INFO] - Step 3642, Loss: {'train/loss': 0.18900372087955475, 'train/grad_norm_pre_clip': 272.4342041015625, 'timing/data': 0.000295000005280599, 'timing/model': 0.3893855850037653, 'learning_rate/qwen_vl_interface': 2.189506933951153e-06, 'learning_rate/action_model': 2.1895069339511525e-05, 'epoch': 0.27}) +[2026-10-03 13:45:14,596][starVLA.training.train_starvla][INFO] - Step 3643, Loss: {'train/loss': 4.185957004665397e-05, 'train/grad_norm_pre_clip': 0.043118808418512344, 'timing/data': 0.00031423999462276697, 'timing/model': 0.39068751499871723, 'learning_rate/qwen_vl_interface': 2.187178544265898e-06, 'learning_rate/action_model': 2.1871785442658983e-05, 'epoch': 0.27}) +[2026-10-03 13:45:15,493][starVLA.training.train_starvla][INFO] - Step 3644, Loss: {'train/loss': 0.6132376790046692, 'train/grad_norm_pre_clip': 50.57462692260742, 'timing/data': 0.0003098799934377894, 'timing/model': 0.5893043959949864, 'learning_rate/qwen_vl_interface': 2.184851413590852e-06, 'learning_rate/action_model': 2.1848514135908518e-05, 'epoch': 0.27}) +[2026-10-03 13:45:16,189][starVLA.training.train_starvla][INFO] - Step 3645, Loss: {'train/loss': 0.0011860596714541316, 'train/grad_norm_pre_clip': 0.8657553195953369, 'timing/data': 0.00023967000015545636, 'timing/model': 0.3896262540074531, 'learning_rate/qwen_vl_interface': 2.1825255428826087e-06, 'learning_rate/action_model': 2.182525542882609e-05, 'epoch': 0.27}) +[2026-10-03 13:45:16,886][starVLA.training.train_starvla][INFO] - Step 3646, Loss: {'train/loss': 0.00011574719246709719, 'train/grad_norm_pre_clip': 0.08031456172466278, 'timing/data': 0.0002062499988824129, 'timing/model': 0.38995706499554217, 'learning_rate/qwen_vl_interface': 2.180200933097246e-06, 'learning_rate/action_model': 2.180200933097246e-05, 'epoch': 0.27}) +[2026-10-03 13:45:17,584][starVLA.training.train_starvla][INFO] - Step 3647, Loss: {'train/loss': 2.6523948690737598e-06, 'train/grad_norm_pre_clip': 0.0006538340821862221, 'timing/data': 0.00031279001268558204, 'timing/model': 0.39065066499460954, 'learning_rate/qwen_vl_interface': 2.1778775851903235e-06, 'learning_rate/action_model': 2.177877585190323e-05, 'epoch': 0.27}) +[2026-10-03 13:45:18,282][starVLA.training.train_starvla][INFO] - Step 3648, Loss: {'train/loss': 0.505630612373352, 'train/grad_norm_pre_clip': 23.458703994750977, 'timing/data': 0.00037958999746479094, 'timing/model': 0.3900974949938245, 'learning_rate/qwen_vl_interface': 2.175555500116881e-06, 'learning_rate/action_model': 2.175555500116881e-05, 'epoch': 0.27}) +[2026-10-03 13:45:18,978][starVLA.training.train_starvla][INFO] - Step 3649, Loss: {'train/loss': 2.9355073820624966e-06, 'train/grad_norm_pre_clip': 0.0006718852091580629, 'timing/data': 0.0003744999994523823, 'timing/model': 0.3895725449983729, 'learning_rate/qwen_vl_interface': 2.1732346788314404e-06, 'learning_rate/action_model': 2.1732346788314402e-05, 'epoch': 0.27}) +[2026-10-03 13:45:19,879][starVLA.training.train_starvla][INFO] - Step 3650, Loss: {'train/loss': 4.7161602196865715e-06, 'train/grad_norm_pre_clip': 0.0015501222806051373, 'timing/data': 0.0002517200045986101, 'timing/model': 0.3901238149992423, 'learning_rate/qwen_vl_interface': 2.170915122288006e-06, 'learning_rate/action_model': 2.170915122288006e-05, 'epoch': 0.27}) +[2026-10-03 13:45:20,575][starVLA.training.train_starvla][INFO] - Step 3651, Loss: {'train/loss': 0.00021039342391304672, 'train/grad_norm_pre_clip': 0.3180779814720154, 'timing/data': 0.00030732000595889986, 'timing/model': 0.39007623499492183, 'learning_rate/qwen_vl_interface': 2.168596831440055e-06, 'learning_rate/action_model': 2.1685968314400552e-05, 'epoch': 0.27}) +[2026-10-03 13:45:21,273][starVLA.training.train_starvla][INFO] - Step 3652, Loss: {'train/loss': 3.2409657251264434e-06, 'train/grad_norm_pre_clip': 0.0012390222400426865, 'timing/data': 0.00024777000362519175, 'timing/model': 0.39067243500903714, 'learning_rate/qwen_vl_interface': 2.166279807240553e-06, 'learning_rate/action_model': 2.1662798072405528e-05, 'epoch': 0.27}) +[2026-10-03 13:45:21,969][starVLA.training.train_starvla][INFO] - Step 3653, Loss: {'train/loss': 2.7939508981944527e-06, 'train/grad_norm_pre_clip': 0.0007355904672294855, 'timing/data': 0.00019789999350905418, 'timing/model': 0.38924020499689505, 'learning_rate/qwen_vl_interface': 2.1639640506419394e-06, 'learning_rate/action_model': 2.163964050641939e-05, 'epoch': 0.27}) +[2026-10-03 13:45:22,667][starVLA.training.train_starvla][INFO] - Step 3654, Loss: {'train/loss': 0.4925224184989929, 'train/grad_norm_pre_clip': 45.31371307373047, 'timing/data': 0.00026826000248547643, 'timing/model': 0.39058929499879014, 'learning_rate/qwen_vl_interface': 2.161649562596135e-06, 'learning_rate/action_model': 2.161649562596135e-05, 'epoch': 0.27}) +[2026-10-03 13:45:23,565][starVLA.training.train_starvla][INFO] - Step 3655, Loss: {'train/loss': 3.0498809792334214e-05, 'train/grad_norm_pre_clip': 0.02479940839111805, 'timing/data': 0.00027080999279860407, 'timing/model': 0.5905502459936542, 'learning_rate/qwen_vl_interface': 2.1593363440545383e-06, 'learning_rate/action_model': 2.1593363440545384e-05, 'epoch': 0.27}) +[2026-10-03 13:45:24,263][starVLA.training.train_starvla][INFO] - Step 3656, Loss: {'train/loss': 0.17256349325180054, 'train/grad_norm_pre_clip': 97.6640396118164, 'timing/data': 0.00033044000156223774, 'timing/model': 0.3903997849993175, 'learning_rate/qwen_vl_interface': 2.157024395968025e-06, 'learning_rate/action_model': 2.1570243959680252e-05, 'epoch': 0.27}) +[2026-10-03 13:45:24,961][starVLA.training.train_starvla][INFO] - Step 3657, Loss: {'train/loss': 0.47794458270072937, 'train/grad_norm_pre_clip': 69.80245208740234, 'timing/data': 0.00020306000078562647, 'timing/model': 0.38981753498956095, 'learning_rate/qwen_vl_interface': 2.154713719286952e-06, 'learning_rate/action_model': 2.154713719286952e-05, 'epoch': 0.27}) +[2026-10-03 13:45:25,659][starVLA.training.train_starvla][INFO] - Step 3658, Loss: {'train/loss': 8.329563570441678e-06, 'train/grad_norm_pre_clip': 0.002813147846609354, 'timing/data': 0.00020435999613255262, 'timing/model': 0.3901248949987348, 'learning_rate/qwen_vl_interface': 2.152404314961149e-06, 'learning_rate/action_model': 2.152404314961149e-05, 'epoch': 0.27}) +[2026-10-03 13:45:26,356][starVLA.training.train_starvla][INFO] - Step 3659, Loss: {'train/loss': 1.8924415599030908e-06, 'train/grad_norm_pre_clip': 0.00036816662759520113, 'timing/data': 0.0002520199923310429, 'timing/model': 0.3900267860008171, 'learning_rate/qwen_vl_interface': 2.1500961839399264e-06, 'learning_rate/action_model': 2.1500961839399262e-05, 'epoch': 0.27}) +[2026-10-03 13:45:27,054][starVLA.training.train_starvla][INFO] - Step 3660, Loss: {'train/loss': 6.0951279010623693e-05, 'train/grad_norm_pre_clip': 0.03749678283929825, 'timing/data': 0.0003196500038029626, 'timing/model': 0.3902666460053297, 'learning_rate/qwen_vl_interface': 2.1477893271720683e-06, 'learning_rate/action_model': 2.1477893271720682e-05, 'epoch': 0.27}) +[2026-10-03 13:45:27,955][starVLA.training.train_starvla][INFO] - Step 3661, Loss: {'train/loss': 3.67265020031482e-05, 'train/grad_norm_pre_clip': 0.016611510887742043, 'timing/data': 0.0002451600012136623, 'timing/model': 0.3904338540014578, 'learning_rate/qwen_vl_interface': 2.1454837456058385e-06, 'learning_rate/action_model': 2.1454837456058386e-05, 'epoch': 0.27}) +[2026-10-03 13:45:28,653][starVLA.training.train_starvla][INFO] - Step 3662, Loss: {'train/loss': 4.723567599285161e-06, 'train/grad_norm_pre_clip': 0.004227781668305397, 'timing/data': 0.00024138999287970364, 'timing/model': 0.3900821160059422, 'learning_rate/qwen_vl_interface': 2.1431794401889722e-06, 'learning_rate/action_model': 2.143179440188972e-05, 'epoch': 0.27}) +[2026-10-03 13:45:29,354][starVLA.training.train_starvla][INFO] - Step 3663, Loss: {'train/loss': 2.108508397213882e-06, 'train/grad_norm_pre_clip': 0.0004979494260624051, 'timing/data': 0.0002487500023562461, 'timing/model': 0.39133944500645157, 'learning_rate/qwen_vl_interface': 2.1408764118686837e-06, 'learning_rate/action_model': 2.1408764118686838e-05, 'epoch': 0.27}) +[2026-10-03 13:45:30,054][starVLA.training.train_starvla][INFO] - Step 3664, Loss: {'train/loss': 1.3946548278909177e-05, 'train/grad_norm_pre_clip': 0.0030052519869059324, 'timing/data': 0.00021042900334578007, 'timing/model': 0.3916789760114625, 'learning_rate/qwen_vl_interface': 2.1385746615916598e-06, 'learning_rate/action_model': 2.1385746615916596e-05, 'epoch': 0.27}) +[2026-10-03 13:45:30,754][starVLA.training.train_starvla][INFO] - Step 3665, Loss: {'train/loss': 4.999265911465045e-06, 'train/grad_norm_pre_clip': 0.0016196905635297298, 'timing/data': 0.0002522200084058568, 'timing/model': 0.3916775149991736, 'learning_rate/qwen_vl_interface': 2.1362741903040636e-06, 'learning_rate/action_model': 2.1362741903040633e-05, 'epoch': 0.27}) +[2026-10-03 13:45:31,653][starVLA.training.train_starvla][INFO] - Step 3666, Loss: {'train/loss': 3.738734812941402e-05, 'train/grad_norm_pre_clip': 0.013545925728976727, 'timing/data': 0.0003246400010539219, 'timing/model': 0.38878905499586836, 'learning_rate/qwen_vl_interface': 2.1339749989515318e-06, 'learning_rate/action_model': 2.1339749989515316e-05, 'epoch': 0.27}) +[2026-10-03 13:45:32,350][starVLA.training.train_starvla][INFO] - Step 3667, Loss: {'train/loss': 2.779055421342491e-06, 'train/grad_norm_pre_clip': 0.0005985806346870959, 'timing/data': 0.0002045299916062504, 'timing/model': 0.38915953499963507, 'learning_rate/qwen_vl_interface': 2.131677088479175e-06, 'learning_rate/action_model': 2.1316770884791752e-05, 'epoch': 0.27}) +[2026-10-03 13:45:33,049][starVLA.training.train_starvla][INFO] - Step 3668, Loss: {'train/loss': 4.418159278429812e-06, 'train/grad_norm_pre_clip': 0.0016928109107539058, 'timing/data': 0.00021491000370588154, 'timing/model': 0.3911789849953493, 'learning_rate/qwen_vl_interface': 2.1293804598315787e-06, 'learning_rate/action_model': 2.1293804598315787e-05, 'epoch': 0.27}) +[2026-10-03 13:45:33,750][starVLA.training.train_starvla][INFO] - Step 3669, Loss: {'train/loss': 5.522264109458774e-05, 'train/grad_norm_pre_clip': 0.0553927943110466, 'timing/data': 0.0002445800055284053, 'timing/model': 0.39207257400266826, 'learning_rate/qwen_vl_interface': 2.1270851139527986e-06, 'learning_rate/action_model': 2.1270851139527983e-05, 'epoch': 0.27}) +[2026-10-03 13:45:34,449][starVLA.training.train_starvla][INFO] - Step 3670, Loss: {'train/loss': 0.00010037969332188368, 'train/grad_norm_pre_clip': 0.15280205011367798, 'timing/data': 0.0002450300089549273, 'timing/model': 0.3903558960009832, 'learning_rate/qwen_vl_interface': 2.1247910517863653e-06, 'learning_rate/action_model': 2.124791051786365e-05, 'epoch': 0.27}) +[2026-10-03 13:45:35,347][starVLA.training.train_starvla][INFO] - Step 3671, Loss: {'train/loss': 4.594229540089145e-05, 'train/grad_norm_pre_clip': 0.030931251123547554, 'timing/data': 0.00024465999740641564, 'timing/model': 0.5911852769932011, 'learning_rate/qwen_vl_interface': 2.1224982742752814e-06, 'learning_rate/action_model': 2.1224982742752812e-05, 'epoch': 0.27}) +[2026-10-03 13:45:36,048][starVLA.training.train_starvla][INFO] - Step 3672, Loss: {'train/loss': 4.872562385571655e-06, 'train/grad_norm_pre_clip': 0.0019362238235771656, 'timing/data': 0.00031598999339621514, 'timing/model': 0.3916370749939233, 'learning_rate/qwen_vl_interface': 2.1202067823620213e-06, 'learning_rate/action_model': 2.1202067823620212e-05, 'epoch': 0.27}) +[2026-10-03 13:45:36,747][starVLA.training.train_starvla][INFO] - Step 3673, Loss: {'train/loss': 4.4451444409787655e-05, 'train/grad_norm_pre_clip': 0.03702346980571747, 'timing/data': 0.000465049990452826, 'timing/model': 0.39068359599332325, 'learning_rate/qwen_vl_interface': 2.117916576988531e-06, 'learning_rate/action_model': 2.117916576988531e-05, 'epoch': 0.27}) +[2026-10-03 13:45:37,447][starVLA.training.train_starvla][INFO] - Step 3674, Loss: {'train/loss': 2.5670033210190013e-05, 'train/grad_norm_pre_clip': 0.010984588414430618, 'timing/data': 0.00020693999249488115, 'timing/model': 0.39179327500460204, 'learning_rate/qwen_vl_interface': 2.1156276590962257e-06, 'learning_rate/action_model': 2.1156276590962257e-05, 'epoch': 0.27}) +[2026-10-03 13:45:38,148][starVLA.training.train_starvla][INFO] - Step 3675, Loss: {'train/loss': 1.639759102545213e-05, 'train/grad_norm_pre_clip': 0.009413506835699081, 'timing/data': 0.000294420009595342, 'timing/model': 0.3927587950020097, 'learning_rate/qwen_vl_interface': 2.113340029625994e-06, 'learning_rate/action_model': 2.1133400296259938e-05, 'epoch': 0.27}) +[2026-10-03 13:45:38,851][starVLA.training.train_starvla][INFO] - Step 3676, Loss: {'train/loss': 3.0100088679319015e-06, 'train/grad_norm_pre_clip': 0.0008747220854274929, 'timing/data': 0.0002891199983423576, 'timing/model': 0.3920312349946471, 'learning_rate/qwen_vl_interface': 2.111053689518194e-06, 'learning_rate/action_model': 2.111053689518194e-05, 'epoch': 0.27}) +[2026-10-03 13:45:39,750][starVLA.training.train_starvla][INFO] - Step 3677, Loss: {'train/loss': 5.975281055725645e-06, 'train/grad_norm_pre_clip': 0.0015965297352522612, 'timing/data': 0.0004295100079616532, 'timing/model': 0.3920046759885736, 'learning_rate/qwen_vl_interface': 2.1087686397126544e-06, 'learning_rate/action_model': 2.1087686397126545e-05, 'epoch': 0.27}) +[2026-10-03 13:45:40,452][starVLA.training.train_starvla][INFO] - Step 3678, Loss: {'train/loss': 1.366393371426966e-05, 'train/grad_norm_pre_clip': 0.005073694046586752, 'timing/data': 0.0002492800122126937, 'timing/model': 0.39183554500050377, 'learning_rate/qwen_vl_interface': 2.106484881148672e-06, 'learning_rate/action_model': 2.1064848811486718e-05, 'epoch': 0.27}) +[2026-10-03 13:45:41,154][starVLA.training.train_starvla][INFO] - Step 3679, Loss: {'train/loss': 0.4641513526439667, 'train/grad_norm_pre_clip': 69.89048767089844, 'timing/data': 0.00020981000852771103, 'timing/model': 0.39302419500018004, 'learning_rate/qwen_vl_interface': 2.1042024147650137e-06, 'learning_rate/action_model': 2.1042024147650137e-05, 'epoch': 0.27}) +[2026-10-03 13:45:41,854][starVLA.training.train_starvla][INFO] - Step 3680, Loss: {'train/loss': 9.588742614141665e-06, 'train/grad_norm_pre_clip': 0.0042481799609959126, 'timing/data': 0.00020542999845929444, 'timing/model': 0.39147014598711394, 'learning_rate/qwen_vl_interface': 2.1019212414999163e-06, 'learning_rate/action_model': 2.101921241499916e-05, 'epoch': 0.27}) +[2026-10-03 13:45:42,554][starVLA.training.train_starvla][INFO] - Step 3681, Loss: {'train/loss': 0.00022274022921919823, 'train/grad_norm_pre_clip': 0.08608546853065491, 'timing/data': 0.00038327999936882406, 'timing/model': 0.3912354750063969, 'learning_rate/qwen_vl_interface': 2.099641362291082e-06, 'learning_rate/action_model': 2.099641362291082e-05, 'epoch': 0.27}) +[2026-10-03 13:45:43,453][starVLA.training.train_starvla][INFO] - Step 3682, Loss: {'train/loss': 2.2426197574532125e-06, 'train/grad_norm_pre_clip': 0.0004935553297400475, 'timing/data': 0.00023987999884411693, 'timing/model': 0.3929706850030925, 'learning_rate/qwen_vl_interface': 2.097362778075685e-06, 'learning_rate/action_model': 2.097362778075685e-05, 'epoch': 0.27}) +[2026-10-03 13:45:44,155][starVLA.training.train_starvla][INFO] - Step 3683, Loss: {'train/loss': 2.7830052204080857e-05, 'train/grad_norm_pre_clip': 0.00962267816066742, 'timing/data': 0.00021143000049050897, 'timing/model': 0.39245634600229096, 'learning_rate/qwen_vl_interface': 2.0950854897903663e-06, 'learning_rate/action_model': 2.095085489790366e-05, 'epoch': 0.27}) +[2026-10-03 13:45:44,854][starVLA.training.train_starvla][INFO] - Step 3684, Loss: {'train/loss': 2.1320433006621897e-05, 'train/grad_norm_pre_clip': 0.016547324135899544, 'timing/data': 0.00028500999906100333, 'timing/model': 0.3910038260073634, 'learning_rate/qwen_vl_interface': 2.092809498371229e-06, 'learning_rate/action_model': 2.0928094983712292e-05, 'epoch': 0.27}) +[2026-10-03 13:45:45,554][starVLA.training.train_starvla][INFO] - Step 3685, Loss: {'train/loss': 3.129233164145262e-06, 'train/grad_norm_pre_clip': 0.0005197857390157878, 'timing/data': 0.00024526999914087355, 'timing/model': 0.391310644990881, 'learning_rate/qwen_vl_interface': 2.0905348047538498e-06, 'learning_rate/action_model': 2.0905348047538497e-05, 'epoch': 0.27}) +[2026-10-03 13:45:46,258][starVLA.training.train_starvla][INFO] - Step 3686, Loss: {'train/loss': 0.6426002979278564, 'train/grad_norm_pre_clip': 42.11663818359375, 'timing/data': 0.0002942400024039671, 'timing/model': 0.3949607050017221, 'learning_rate/qwen_vl_interface': 2.0882614098732686e-06, 'learning_rate/action_model': 2.0882614098732685e-05, 'epoch': 0.27}) +[2026-10-03 13:45:47,174][starVLA.training.train_starvla][INFO] - Step 3687, Loss: {'train/loss': 0.3065599501132965, 'train/grad_norm_pre_clip': 263.52532958984375, 'timing/data': 0.00038152998604346067, 'timing/model': 0.3918121750029968, 'learning_rate/qwen_vl_interface': 2.085989314663993e-06, 'learning_rate/action_model': 2.0859893146639927e-05, 'epoch': 0.27}) +[2026-10-03 13:45:47,870][starVLA.training.train_starvla][INFO] - Step 3688, Loss: {'train/loss': 0.5791423320770264, 'train/grad_norm_pre_clip': 69.66854095458984, 'timing/data': 0.00031915899307932705, 'timing/model': 0.389286486999481, 'learning_rate/qwen_vl_interface': 2.083718520059995e-06, 'learning_rate/action_model': 2.0837185200599946e-05, 'epoch': 0.27}) +[2026-10-03 13:45:48,566][starVLA.training.train_starvla][INFO] - Step 3689, Loss: {'train/loss': 3.673118044389412e-06, 'train/grad_norm_pre_clip': 0.0012747214641422033, 'timing/data': 0.00029861999792046845, 'timing/model': 0.38882290599576663, 'learning_rate/qwen_vl_interface': 2.0814490269947117e-06, 'learning_rate/action_model': 2.0814490269947118e-05, 'epoch': 0.27}) +[2026-10-03 13:45:49,264][starVLA.training.train_starvla][INFO] - Step 3690, Loss: {'train/loss': 1.0385830137238372e-05, 'train/grad_norm_pre_clip': 0.007863888517022133, 'timing/data': 0.00033804000122472644, 'timing/model': 0.3910550659929868, 'learning_rate/qwen_vl_interface': 2.079180836401047e-06, 'learning_rate/action_model': 2.079180836401047e-05, 'epoch': 0.27}) +[2026-10-03 13:45:49,962][starVLA.training.train_starvla][INFO] - Step 3691, Loss: {'train/loss': 0.013988767750561237, 'train/grad_norm_pre_clip': 20.26090431213379, 'timing/data': 0.0003588999970816076, 'timing/model': 0.38984390599944163, 'learning_rate/qwen_vl_interface': 2.0769139492113687e-06, 'learning_rate/action_model': 2.076913949211369e-05, 'epoch': 0.27}) +[2026-10-03 13:45:50,859][starVLA.training.train_starvla][INFO] - Step 3692, Loss: {'train/loss': 0.42616400122642517, 'train/grad_norm_pre_clip': 73.89445495605469, 'timing/data': 0.0003093000123044476, 'timing/model': 0.5910545080114389, 'learning_rate/qwen_vl_interface': 2.0746483663575075e-06, 'learning_rate/action_model': 2.0746483663575076e-05, 'epoch': 0.27}) +[2026-10-03 13:45:51,556][starVLA.training.train_starvla][INFO] - Step 3693, Loss: {'train/loss': 1.0430538168293424e-05, 'train/grad_norm_pre_clip': 0.007744073402136564, 'timing/data': 0.0004060399951413274, 'timing/model': 0.3891448060021503, 'learning_rate/qwen_vl_interface': 2.0723840887707595e-06, 'learning_rate/action_model': 2.0723840887707594e-05, 'epoch': 0.27}) +[2026-10-03 13:45:52,256][starVLA.training.train_starvla][INFO] - Step 3694, Loss: {'train/loss': 0.00032255935366265476, 'train/grad_norm_pre_clip': 0.6481813192367554, 'timing/data': 0.000261629989836365, 'timing/model': 0.39132407600118313, 'learning_rate/qwen_vl_interface': 2.0701211173818854e-06, 'learning_rate/action_model': 2.0701211173818853e-05, 'epoch': 0.27}) +[2026-10-03 13:45:52,952][starVLA.training.train_starvla][INFO] - Step 3695, Loss: {'train/loss': 5.826276719744783e-06, 'train/grad_norm_pre_clip': 0.0017278887098655105, 'timing/data': 0.0002678500022739172, 'timing/model': 0.38958500600710977, 'learning_rate/qwen_vl_interface': 2.0678594531211036e-06, 'learning_rate/action_model': 2.0678594531211033e-05, 'epoch': 0.27}) +[2026-10-03 13:45:53,653][starVLA.training.train_starvla][INFO] - Step 3696, Loss: {'train/loss': 4.916541001875885e-05, 'train/grad_norm_pre_clip': 0.059323374181985855, 'timing/data': 0.0003129699907731265, 'timing/model': 0.39251342600618955, 'learning_rate/qwen_vl_interface': 2.0655990969181e-06, 'learning_rate/action_model': 2.0655990969181003e-05, 'epoch': 0.27}) +[2026-10-03 13:45:54,549][starVLA.training.train_starvla][INFO] - Step 3697, Loss: {'train/loss': 1.0631544682837557e-05, 'train/grad_norm_pre_clip': 0.004716624040156603, 'timing/data': 0.0002017799997702241, 'timing/model': 0.5879872189980233, 'learning_rate/qwen_vl_interface': 2.0633400497020236e-06, 'learning_rate/action_model': 2.0633400497020232e-05, 'epoch': 0.27}) +[2026-10-03 13:45:55,248][starVLA.training.train_starvla][INFO] - Step 3698, Loss: {'train/loss': 8.251483814092353e-05, 'train/grad_norm_pre_clip': 0.07257973402738571, 'timing/data': 0.00020482999389059842, 'timing/model': 0.3924459459958598, 'learning_rate/qwen_vl_interface': 2.061082312401481e-06, 'learning_rate/action_model': 2.061082312401481e-05, 'epoch': 0.27}) +[2026-10-03 13:45:55,945][starVLA.training.train_starvla][INFO] - Step 3699, Loss: {'train/loss': 4.1119856177829206e-05, 'train/grad_norm_pre_clip': 0.010180429555475712, 'timing/data': 0.00021004999871365726, 'timing/model': 0.38937556599557865, 'learning_rate/qwen_vl_interface': 2.058825885944544e-06, 'learning_rate/action_model': 2.0588258859445442e-05, 'epoch': 0.27}) +[2026-10-03 13:45:56,641][starVLA.training.train_starvla][INFO] - Step 3700, Loss: {'train/loss': 0.0010447090025991201, 'train/grad_norm_pre_clip': 1.1355781555175781, 'timing/data': 0.00029256999550852925, 'timing/model': 0.3896806559932884, 'learning_rate/qwen_vl_interface': 2.0565707712587447e-06, 'learning_rate/action_model': 2.0565707712587444e-05, 'epoch': 0.27}) +[2026-10-03 13:45:57,338][starVLA.training.train_starvla][INFO] - Step 3701, Loss: {'train/loss': 2.6225916371913627e-06, 'train/grad_norm_pre_clip': 0.0009007180924527347, 'timing/data': 0.0004020400083390996, 'timing/model': 0.3901671060011722, 'learning_rate/qwen_vl_interface': 2.0543169692710747e-06, 'learning_rate/action_model': 2.054316969271075e-05, 'epoch': 0.27}) +[2026-10-03 13:45:58,036][starVLA.training.train_starvla][INFO] - Step 3702, Loss: {'train/loss': 0.00030179228633642197, 'train/grad_norm_pre_clip': 1.899295687675476, 'timing/data': 0.0002104599989252165, 'timing/model': 0.39156555698718876, 'learning_rate/qwen_vl_interface': 2.0520644809079873e-06, 'learning_rate/action_model': 2.0520644809079873e-05, 'epoch': 0.27}) +[2026-10-03 13:45:58,937][starVLA.training.train_starvla][INFO] - Step 3703, Loss: {'train/loss': 8.292353413708042e-06, 'train/grad_norm_pre_clip': 0.0022334486711770296, 'timing/data': 0.0002089199988404289, 'timing/model': 0.3891974270081846, 'learning_rate/qwen_vl_interface': 2.049813307095395e-06, 'learning_rate/action_model': 2.0498133070953947e-05, 'epoch': 0.27}) +[2026-10-03 13:45:59,635][starVLA.training.train_starvla][INFO] - Step 3704, Loss: {'train/loss': 3.935575659852475e-05, 'train/grad_norm_pre_clip': 0.009565147571265697, 'timing/data': 0.00037970999255776405, 'timing/model': 0.3899396469932981, 'learning_rate/qwen_vl_interface': 2.047563448758672e-06, 'learning_rate/action_model': 2.0475634487586717e-05, 'epoch': 0.27}) +[2026-10-03 13:46:00,335][starVLA.training.train_starvla][INFO] - Step 3705, Loss: {'train/loss': 5.1442581025185063e-05, 'train/grad_norm_pre_clip': 0.013375196605920792, 'timing/data': 0.0002479899994796142, 'timing/model': 0.39146759700088296, 'learning_rate/qwen_vl_interface': 2.045314906822647e-06, 'learning_rate/action_model': 2.0453149068226467e-05, 'epoch': 0.27}) +[2026-10-03 13:46:01,033][starVLA.training.train_starvla][INFO] - Step 3706, Loss: {'train/loss': 0.00541484821587801, 'train/grad_norm_pre_clip': 11.332419395446777, 'timing/data': 0.0002107699983753264, 'timing/model': 0.38991296700260136, 'learning_rate/qwen_vl_interface': 2.0430676822116124e-06, 'learning_rate/action_model': 2.0430676822116123e-05, 'epoch': 0.27}) +[2026-10-03 13:46:01,729][starVLA.training.train_starvla][INFO] - Step 3707, Loss: {'train/loss': 1.797665208869148e-05, 'train/grad_norm_pre_clip': 0.008500497788190842, 'timing/data': 0.0002476399968145415, 'timing/model': 0.39023493700369727, 'learning_rate/qwen_vl_interface': 2.040821775849319e-06, 'learning_rate/action_model': 2.0408217758493188e-05, 'epoch': 0.27}) +[2026-10-03 13:46:02,622][starVLA.training.train_starvla][INFO] - Step 3708, Loss: {'train/loss': 4.641686246031895e-06, 'train/grad_norm_pre_clip': 0.0010651643387973309, 'timing/data': 0.0002943400031654164, 'timing/model': 0.5862507509882562, 'learning_rate/qwen_vl_interface': 2.03857718865897e-06, 'learning_rate/action_model': 2.0385771886589702e-05, 'epoch': 0.27}) +[2026-10-03 13:46:03,320][starVLA.training.train_starvla][INFO] - Step 3709, Loss: {'train/loss': 0.0005351860309019685, 'train/grad_norm_pre_clip': 1.2617870569229126, 'timing/data': 0.0002875299978768453, 'timing/model': 0.39061099699756596, 'learning_rate/qwen_vl_interface': 2.0363339215632337e-06, 'learning_rate/action_model': 2.0363339215632334e-05, 'epoch': 0.27}) +[2026-10-03 13:46:04,018][starVLA.training.train_starvla][INFO] - Step 3710, Loss: {'train/loss': 0.5767149329185486, 'train/grad_norm_pre_clip': 56.10459899902344, 'timing/data': 0.00035104999551549554, 'timing/model': 0.3898620270047104, 'learning_rate/qwen_vl_interface': 2.034091975484231e-06, 'learning_rate/action_model': 2.0340919754842307e-05, 'epoch': 0.27}) +[2026-10-03 13:46:04,713][starVLA.training.train_starvla][INFO] - Step 3711, Loss: {'train/loss': 3.1729206966701895e-05, 'train/grad_norm_pre_clip': 0.0061440556310117245, 'timing/data': 0.00019638000230770558, 'timing/model': 0.3890727369871456, 'learning_rate/qwen_vl_interface': 2.0318513513435415e-06, 'learning_rate/action_model': 2.0318513513435415e-05, 'epoch': 0.27}) +[2026-10-03 13:46:05,409][starVLA.training.train_starvla][INFO] - Step 3712, Loss: {'train/loss': 0.00011681314208544791, 'train/grad_norm_pre_clip': 0.04332498461008072, 'timing/data': 0.0003001700097229332, 'timing/model': 0.3890405669953907, 'learning_rate/qwen_vl_interface': 2.0296120500622018e-06, 'learning_rate/action_model': 2.0296120500622017e-05, 'epoch': 0.27}) +[2026-10-03 13:46:06,108][starVLA.training.train_starvla][INFO] - Step 3713, Loss: {'train/loss': 8.268166857305914e-05, 'train/grad_norm_pre_clip': 0.037916138768196106, 'timing/data': 0.0005130399949848652, 'timing/model': 0.39065664800000377, 'learning_rate/qwen_vl_interface': 2.0273740725607034e-06, 'learning_rate/action_model': 2.0273740725607032e-05, 'epoch': 0.27}) +[2026-10-03 13:46:07,026][starVLA.training.train_starvla][INFO] - Step 3714, Loss: {'train/loss': 3.4648695873329416e-05, 'train/grad_norm_pre_clip': 0.015460985712707043, 'timing/data': 0.0002203699987148866, 'timing/model': 0.3910075570020126, 'learning_rate/qwen_vl_interface': 2.0251374197589942e-06, 'learning_rate/action_model': 2.0251374197589943e-05, 'epoch': 0.27}) +[2026-10-03 13:46:07,724][starVLA.training.train_starvla][INFO] - Step 3715, Loss: {'train/loss': 1.541493293188978e-05, 'train/grad_norm_pre_clip': 0.007784674409776926, 'timing/data': 0.00020087999291718006, 'timing/model': 0.39038004699978046, 'learning_rate/qwen_vl_interface': 2.0229020925764785e-06, 'learning_rate/action_model': 2.0229020925764785e-05, 'epoch': 0.27}) +[2026-10-03 13:46:08,424][starVLA.training.train_starvla][INFO] - Step 3716, Loss: {'train/loss': 1.5765101124998182e-05, 'train/grad_norm_pre_clip': 0.005884331651031971, 'timing/data': 0.00022385999909602106, 'timing/model': 0.39149118699424434, 'learning_rate/qwen_vl_interface': 2.0206680919320145e-06, 'learning_rate/action_model': 2.020668091932014e-05, 'epoch': 0.27}) +[2026-10-03 13:46:09,124][starVLA.training.train_starvla][INFO] - Step 3717, Loss: {'train/loss': 1.5332938346546143e-05, 'train/grad_norm_pre_clip': 0.005531474482268095, 'timing/data': 0.0003344300057506189, 'timing/model': 0.3905757570028072, 'learning_rate/qwen_vl_interface': 2.0184354187439164e-06, 'learning_rate/action_model': 2.0184354187439162e-05, 'epoch': 0.27}) +[2026-10-03 13:46:09,821][starVLA.training.train_starvla][INFO] - Step 3718, Loss: {'train/loss': 3.985914736404084e-05, 'train/grad_norm_pre_clip': 0.015051648020744324, 'timing/data': 0.0003804099978879094, 'timing/model': 0.39018156701058615, 'learning_rate/qwen_vl_interface': 2.016204073929949e-06, 'learning_rate/action_model': 2.016204073929949e-05, 'epoch': 0.27}) +[2026-10-03 13:46:10,724][starVLA.training.train_starvla][INFO] - Step 3719, Loss: {'train/loss': 1.305370569229126, 'train/grad_norm_pre_clip': 256.9947814941406, 'timing/data': 0.0003623500087996945, 'timing/model': 0.389701607011375, 'learning_rate/qwen_vl_interface': 2.0139740584073363e-06, 'learning_rate/action_model': 2.0139740584073364e-05, 'epoch': 0.27}) +[2026-10-03 13:46:11,422][starVLA.training.train_starvla][INFO] - Step 3720, Loss: {'train/loss': 5.386231714510359e-05, 'train/grad_norm_pre_clip': 0.013589446432888508, 'timing/data': 0.0004886700044153258, 'timing/model': 0.3909602970088599, 'learning_rate/qwen_vl_interface': 2.011745373092753e-06, 'learning_rate/action_model': 2.011745373092753e-05, 'epoch': 0.27}) +[2026-10-03 13:46:12,118][starVLA.training.train_starvla][INFO] - Step 3721, Loss: {'train/loss': 0.00524948863312602, 'train/grad_norm_pre_clip': 14.109016418457031, 'timing/data': 0.00020302001212257892, 'timing/model': 0.3894068170047831, 'learning_rate/qwen_vl_interface': 2.0095180189023267e-06, 'learning_rate/action_model': 2.0095180189023265e-05, 'epoch': 0.27}) +[2026-10-03 13:46:12,815][starVLA.training.train_starvla][INFO] - Step 3722, Loss: {'train/loss': 1.5548881492577493e-05, 'train/grad_norm_pre_clip': 0.0038194512017071247, 'timing/data': 0.00027968999347649515, 'timing/model': 0.39090474699332844, 'learning_rate/qwen_vl_interface': 2.0072919967516393e-06, 'learning_rate/action_model': 2.0072919967516392e-05, 'epoch': 0.27}) +[2026-10-03 13:46:13,513][starVLA.training.train_starvla][INFO] - Step 3723, Loss: {'train/loss': 1.6495072486577556e-05, 'train/grad_norm_pre_clip': 0.005306340288370848, 'timing/data': 0.0002142000012099743, 'timing/model': 0.3890725270030089, 'learning_rate/qwen_vl_interface': 2.0050673075557245e-06, 'learning_rate/action_model': 2.0050673075557243e-05, 'epoch': 0.27}) +[2026-10-03 13:46:14,426][starVLA.training.train_starvla][INFO] - Step 3724, Loss: {'train/loss': 1.6316029359586537e-05, 'train/grad_norm_pre_clip': 0.006461856886744499, 'timing/data': 0.00021189999824855477, 'timing/model': 0.6052287979982793, 'learning_rate/qwen_vl_interface': 2.0028439522290677e-06, 'learning_rate/action_model': 2.0028439522290676e-05, 'epoch': 0.27}) +[2026-10-03 13:46:15,121][starVLA.training.train_starvla][INFO] - Step 3725, Loss: {'train/loss': 3.865432881866582e-05, 'train/grad_norm_pre_clip': 0.008075966499745846, 'timing/data': 0.00019794999388977885, 'timing/model': 0.3895187769958284, 'learning_rate/qwen_vl_interface': 2.000621931685606e-06, 'learning_rate/action_model': 2.000621931685606e-05, 'epoch': 0.27}) +[2026-10-03 13:46:15,817][starVLA.training.train_starvla][INFO] - Step 3726, Loss: {'train/loss': 9.94086149148643e-05, 'train/grad_norm_pre_clip': 0.23275701701641083, 'timing/data': 0.00020247000793460757, 'timing/model': 0.3895402979978826, 'learning_rate/qwen_vl_interface': 1.9984012468387285e-06, 'learning_rate/action_model': 1.9984012468387285e-05, 'epoch': 0.27}) +[2026-10-03 13:46:16,516][starVLA.training.train_starvla][INFO] - Step 3727, Loss: {'train/loss': 0.2500756084918976, 'train/grad_norm_pre_clip': 84.42997741699219, 'timing/data': 0.0002580700092948973, 'timing/model': 0.39154265700199176, 'learning_rate/qwen_vl_interface': 1.9961818986012767e-06, 'learning_rate/action_model': 1.9961818986012763e-05, 'epoch': 0.27}) +[2026-10-03 13:46:17,216][starVLA.training.train_starvla][INFO] - Step 3728, Loss: {'train/loss': 6.675553595414385e-05, 'train/grad_norm_pre_clip': 0.013224252499639988, 'timing/data': 0.0002641800092533231, 'timing/model': 0.3919466380029917, 'learning_rate/qwen_vl_interface': 1.9939638878855372e-06, 'learning_rate/action_model': 1.993963887885537e-05, 'epoch': 0.27}) +[2026-10-03 13:46:17,913][starVLA.training.train_starvla][INFO] - Step 3729, Loss: {'train/loss': 0.005285858176648617, 'train/grad_norm_pre_clip': 7.066102504730225, 'timing/data': 0.0003481800085864961, 'timing/model': 0.38907243800349534, 'learning_rate/qwen_vl_interface': 1.991747215603252e-06, 'learning_rate/action_model': 1.991747215603252e-05, 'epoch': 0.27}) +[2026-10-03 13:46:18,817][starVLA.training.train_starvla][INFO] - Step 3730, Loss: {'train/loss': 5.7741413002077024e-06, 'train/grad_norm_pre_clip': 0.0014593469677492976, 'timing/data': 0.00021791999461129308, 'timing/model': 0.39036741699965205, 'learning_rate/qwen_vl_interface': 1.9895318826656116e-06, 'learning_rate/action_model': 1.9895318826656114e-05, 'epoch': 0.27}) +[2026-10-03 13:46:19,517][starVLA.training.train_starvla][INFO] - Step 3731, Loss: {'train/loss': 0.028495758771896362, 'train/grad_norm_pre_clip': 70.3845443725586, 'timing/data': 0.00022878999880049378, 'timing/model': 0.39162909699371085, 'learning_rate/qwen_vl_interface': 1.9873178899832553e-06, 'learning_rate/action_model': 1.9873178899832554e-05, 'epoch': 0.27}) +[2026-10-03 13:46:20,215][starVLA.training.train_starvla][INFO] - Step 3732, Loss: {'train/loss': 4.292032463126816e-05, 'train/grad_norm_pre_clip': 0.09473453462123871, 'timing/data': 0.00020621001021936536, 'timing/model': 0.39112662700063083, 'learning_rate/qwen_vl_interface': 1.985105238466272e-06, 'learning_rate/action_model': 1.9851052384662722e-05, 'epoch': 0.27}) +[2026-10-03 13:46:20,912][starVLA.training.train_starvla][INFO] - Step 3733, Loss: {'train/loss': 0.00019784871255978942, 'train/grad_norm_pre_clip': 0.05533720180392265, 'timing/data': 0.0002016500075114891, 'timing/model': 0.389835456997389, 'learning_rate/qwen_vl_interface': 1.9828939290242e-06, 'learning_rate/action_model': 1.9828939290241998e-05, 'epoch': 0.27}) +[2026-10-03 13:46:21,609][starVLA.training.train_starvla][INFO] - Step 3734, Loss: {'train/loss': 0.522449254989624, 'train/grad_norm_pre_clip': 586.6047973632812, 'timing/data': 0.00036160000308882445, 'timing/model': 0.3895870380074484, 'learning_rate/qwen_vl_interface': 1.980683962566024e-06, 'learning_rate/action_model': 1.980683962566024e-05, 'epoch': 0.27}) +[2026-10-03 13:46:22,504][starVLA.training.train_starvla][INFO] - Step 3735, Loss: {'train/loss': 0.04998378828167915, 'train/grad_norm_pre_clip': 374.38922119140625, 'timing/data': 0.0003037000133190304, 'timing/model': 0.38873956799216103, 'learning_rate/qwen_vl_interface': 1.9784753400001784e-06, 'learning_rate/action_model': 1.978475340000178e-05, 'epoch': 0.27}) +[2026-10-03 13:46:23,201][starVLA.training.train_starvla][INFO] - Step 3736, Loss: {'train/loss': 0.4545831084251404, 'train/grad_norm_pre_clip': 123.58324432373047, 'timing/data': 0.0002504199946997687, 'timing/model': 0.39088335700216703, 'learning_rate/qwen_vl_interface': 1.9762680622345443e-06, 'learning_rate/action_model': 1.9762680622345442e-05, 'epoch': 0.27}) +[2026-10-03 13:46:23,901][starVLA.training.train_starvla][INFO] - Step 3737, Loss: {'train/loss': 0.0001995010970858857, 'train/grad_norm_pre_clip': 0.150528684258461, 'timing/data': 0.0003216799959773198, 'timing/model': 0.38982500799465925, 'learning_rate/qwen_vl_interface': 1.9740621301764513e-06, 'learning_rate/action_model': 1.9740621301764513e-05, 'epoch': 0.27}) +[2026-10-03 13:46:24,597][starVLA.training.train_starvla][INFO] - Step 3738, Loss: {'train/loss': 0.0013992543099448085, 'train/grad_norm_pre_clip': 4.101600170135498, 'timing/data': 0.00019727999460883439, 'timing/model': 0.3897069889935665, 'learning_rate/qwen_vl_interface': 1.971857544732672e-06, 'learning_rate/action_model': 1.971857544732672e-05, 'epoch': 0.27}) +[2026-10-03 13:46:25,294][starVLA.training.train_starvla][INFO] - Step 3739, Loss: {'train/loss': 6.87912106513977e-05, 'train/grad_norm_pre_clip': 0.02647673338651657, 'timing/data': 0.0003313300112495199, 'timing/model': 0.3909271980082849, 'learning_rate/qwen_vl_interface': 1.96965430680943e-06, 'learning_rate/action_model': 1.96965430680943e-05, 'epoch': 0.27}) +[2026-10-03 13:46:26,198][starVLA.training.train_starvla][INFO] - Step 3740, Loss: {'train/loss': 0.007261898368597031, 'train/grad_norm_pre_clip': 18.554359436035156, 'timing/data': 0.000251429999480024, 'timing/model': 0.5950491699913982, 'learning_rate/qwen_vl_interface': 1.967452417312393e-06, 'learning_rate/action_model': 1.9674524173123927e-05, 'epoch': 0.27}) +[2026-10-03 13:46:26,896][starVLA.training.train_starvla][INFO] - Step 3741, Loss: {'train/loss': 1.5541583707090467e-05, 'train/grad_norm_pre_clip': 0.006971007212996483, 'timing/data': 0.0002440899988869205, 'timing/model': 0.390621647995431, 'learning_rate/qwen_vl_interface': 1.9652518771466746e-06, 'learning_rate/action_model': 1.9652518771466747e-05, 'epoch': 0.27}) +[2026-10-03 13:46:27,596][starVLA.training.train_starvla][INFO] - Step 3742, Loss: {'train/loss': 2.1620762709062546e-05, 'train/grad_norm_pre_clip': 0.005097914952784777, 'timing/data': 0.00031610998848918825, 'timing/model': 0.3908615370019106, 'learning_rate/qwen_vl_interface': 1.963052687216834e-06, 'learning_rate/action_model': 1.963052687216834e-05, 'epoch': 0.27}) +[2026-10-03 13:46:28,294][starVLA.training.train_starvla][INFO] - Step 3743, Loss: {'train/loss': 0.0002110665664076805, 'train/grad_norm_pre_clip': 0.08366034924983978, 'timing/data': 0.00021101000311318785, 'timing/model': 0.3907175579952309, 'learning_rate/qwen_vl_interface': 1.960854848426874e-06, 'learning_rate/action_model': 1.960854848426874e-05, 'epoch': 0.27}) +[2026-10-03 13:46:28,994][starVLA.training.train_starvla][INFO] - Step 3744, Loss: {'train/loss': 0.0004391282272990793, 'train/grad_norm_pre_clip': 0.31990310549736023, 'timing/data': 0.0002579099964350462, 'timing/model': 0.39225062799232546, 'learning_rate/qwen_vl_interface': 1.9586583616802446e-06, 'learning_rate/action_model': 1.9586583616802446e-05, 'epoch': 0.27}) +[2026-10-03 13:46:29,897][starVLA.training.train_starvla][INFO] - Step 3745, Loss: {'train/loss': 0.19141839444637299, 'train/grad_norm_pre_clip': 172.3841552734375, 'timing/data': 0.0002746900136116892, 'timing/model': 0.5946174999990035, 'learning_rate/qwen_vl_interface': 1.956463227879838e-06, 'learning_rate/action_model': 1.9564632278798377e-05, 'epoch': 0.27}) +[2026-10-03 13:46:30,594][starVLA.training.train_starvla][INFO] - Step 3746, Loss: {'train/loss': 0.37852638959884644, 'train/grad_norm_pre_clip': 33.511234283447266, 'timing/data': 0.00026673001411836594, 'timing/model': 0.38964657799806446, 'learning_rate/qwen_vl_interface': 1.95426944792799e-06, 'learning_rate/action_model': 1.95426944792799e-05, 'epoch': 0.27}) +[2026-10-03 13:46:31,293][starVLA.training.train_starvla][INFO] - Step 3747, Loss: {'train/loss': 0.0001138008083216846, 'train/grad_norm_pre_clip': 0.05888790264725685, 'timing/data': 0.00027209999097976834, 'timing/model': 0.39023165800608695, 'learning_rate/qwen_vl_interface': 1.9520770227264817e-06, 'learning_rate/action_model': 1.9520770227264817e-05, 'epoch': 0.27}) +[2026-10-03 13:46:31,992][starVLA.training.train_starvla][INFO] - Step 3748, Loss: {'train/loss': 5.326019527274184e-05, 'train/grad_norm_pre_clip': 0.021910516545176506, 'timing/data': 0.0003438299900153652, 'timing/model': 0.39175287800026126, 'learning_rate/qwen_vl_interface': 1.9498859531765373e-06, 'learning_rate/action_model': 1.9498859531765372e-05, 'epoch': 0.27}) +[2026-10-03 13:46:32,691][starVLA.training.train_starvla][INFO] - Step 3749, Loss: {'train/loss': 0.46055465936660767, 'train/grad_norm_pre_clip': 99.4489974975586, 'timing/data': 0.00021108999499119818, 'timing/model': 0.39078840700676665, 'learning_rate/qwen_vl_interface': 1.9476962401788204e-06, 'learning_rate/action_model': 1.9476962401788204e-05, 'epoch': 0.27}) +[2026-10-03 13:46:33,596][starVLA.training.train_starvla][INFO] - Step 3750, Loss: {'train/loss': 9.183440124616027e-05, 'train/grad_norm_pre_clip': 0.02505672536790371, 'timing/data': 0.00027226000383961946, 'timing/model': 0.5964569299976574, 'learning_rate/qwen_vl_interface': 1.9455078846334397e-06, 'learning_rate/action_model': 1.9455078846334395e-05, 'epoch': 0.27}) +[2026-10-03 13:46:34,294][starVLA.training.train_starvla][INFO] - Step 3751, Loss: {'train/loss': 0.00013996503548696637, 'train/grad_norm_pre_clip': 0.032499078661203384, 'timing/data': 0.00023798999609425664, 'timing/model': 0.3906584680080414, 'learning_rate/qwen_vl_interface': 1.9433208874399465e-06, 'learning_rate/action_model': 1.9433208874399464e-05, 'epoch': 0.27}) +[2026-10-03 13:46:34,992][starVLA.training.train_starvla][INFO] - Step 3752, Loss: {'train/loss': 1.671857899054885e-05, 'train/grad_norm_pre_clip': 0.00454401969909668, 'timing/data': 0.00028870000096503645, 'timing/model': 0.3914497279911302, 'learning_rate/qwen_vl_interface': 1.9411352494973314e-06, 'learning_rate/action_model': 1.9411352494973315e-05, 'epoch': 0.27}) +[2026-10-03 13:46:35,689][starVLA.training.train_starvla][INFO] - Step 3753, Loss: {'train/loss': 0.0002658408193383366, 'train/grad_norm_pre_clip': 0.09923011064529419, 'timing/data': 0.00020699000742752105, 'timing/model': 0.3900433180097025, 'learning_rate/qwen_vl_interface': 1.938950971704031e-06, 'learning_rate/action_model': 1.9389509717040308e-05, 'epoch': 0.27}) +[2026-10-03 13:46:36,388][starVLA.training.train_starvla][INFO] - Step 3754, Loss: {'train/loss': 0.00018132766126655042, 'train/grad_norm_pre_clip': 0.040412697941064835, 'timing/data': 0.00035988999297842383, 'timing/model': 0.39045237899699714, 'learning_rate/qwen_vl_interface': 1.936768054957915e-06, 'learning_rate/action_model': 1.9367680549579147e-05, 'epoch': 0.27}) +[2026-10-03 13:46:37,087][starVLA.training.train_starvla][INFO] - Step 3755, Loss: {'train/loss': 0.0001883072982309386, 'train/grad_norm_pre_clip': 0.48956263065338135, 'timing/data': 0.00020850000146310776, 'timing/model': 0.39222989800327923, 'learning_rate/qwen_vl_interface': 1.9345865001562994e-06, 'learning_rate/action_model': 1.9345865001562994e-05, 'epoch': 0.27}) +[2026-10-03 13:46:37,995][starVLA.training.train_starvla][INFO] - Step 3756, Loss: {'train/loss': 6.392571231117472e-05, 'train/grad_norm_pre_clip': 0.015583956614136696, 'timing/data': 0.00021263000962790102, 'timing/model': 0.39197432799846865, 'learning_rate/qwen_vl_interface': 1.9324063081959394e-06, 'learning_rate/action_model': 1.9324063081959395e-05, 'epoch': 0.27}) +[2026-10-03 13:46:38,694][starVLA.training.train_starvla][INFO] - Step 3757, Loss: {'train/loss': 3.716935316333547e-05, 'train/grad_norm_pre_clip': 0.012690888717770576, 'timing/data': 0.00035898000351153314, 'timing/model': 0.39049748799880035, 'learning_rate/qwen_vl_interface': 1.9302274799730304e-06, 'learning_rate/action_model': 1.9302274799730306e-05, 'epoch': 0.27}) +[2026-10-03 13:46:39,394][starVLA.training.train_starvla][INFO] - Step 3758, Loss: {'train/loss': 9.444173338124529e-05, 'train/grad_norm_pre_clip': 0.031211281195282936, 'timing/data': 0.00037167000118643045, 'timing/model': 0.3920637980045285, 'learning_rate/qwen_vl_interface': 1.9280500163832055e-06, 'learning_rate/action_model': 1.9280500163832055e-05, 'epoch': 0.28}) +[2026-10-03 13:46:40,092][starVLA.training.train_starvla][INFO] - Step 3759, Loss: {'train/loss': 8.812879241304472e-05, 'train/grad_norm_pre_clip': 0.03141147270798683, 'timing/data': 0.00032508998992852867, 'timing/model': 0.3897720090026269, 'learning_rate/qwen_vl_interface': 1.925873918321537e-06, 'learning_rate/action_model': 1.925873918321537e-05, 'epoch': 0.28}) +[2026-10-03 13:46:40,790][starVLA.training.train_starvla][INFO] - Step 3760, Loss: {'train/loss': 0.00011053961497964337, 'train/grad_norm_pre_clip': 0.024762773886322975, 'timing/data': 0.00024866999592632055, 'timing/model': 0.3913589490111917, 'learning_rate/qwen_vl_interface': 1.92369918668254e-06, 'learning_rate/action_model': 1.92369918668254e-05, 'epoch': 0.28}) +[2026-10-03 13:46:41,682][starVLA.training.train_starvla][INFO] - Step 3761, Loss: {'train/loss': 0.00037796818651258945, 'train/grad_norm_pre_clip': 0.3251453936100006, 'timing/data': 0.00024564001068938524, 'timing/model': 0.5844485620036721, 'learning_rate/qwen_vl_interface': 1.9215258223601605e-06, 'learning_rate/action_model': 1.9215258223601605e-05, 'epoch': 0.28}) +[2026-10-03 13:46:42,377][starVLA.training.train_starvla][INFO] - Step 3762, Loss: {'train/loss': 0.0011153897503390908, 'train/grad_norm_pre_clip': 1.14961576461792, 'timing/data': 0.0002464400022290647, 'timing/model': 0.3896533490042202, 'learning_rate/qwen_vl_interface': 1.9193538262477874e-06, 'learning_rate/action_model': 1.9193538262477874e-05, 'epoch': 0.28}) +[2026-10-03 13:46:43,074][starVLA.training.train_starvla][INFO] - Step 3763, Loss: {'train/loss': 5.103659350425005e-05, 'train/grad_norm_pre_clip': 0.011485445313155651, 'timing/data': 0.0002346700057387352, 'timing/model': 0.38944625899603125, 'learning_rate/qwen_vl_interface': 1.917183199238247e-06, 'learning_rate/action_model': 1.917183199238247e-05, 'epoch': 0.28}) +[2026-10-03 13:46:43,775][starVLA.training.train_starvla][INFO] - Step 3764, Loss: {'train/loss': 0.006083684042096138, 'train/grad_norm_pre_clip': 23.662330627441406, 'timing/data': 0.00026080000679939985, 'timing/model': 0.3924077590054367, 'learning_rate/qwen_vl_interface': 1.9150139422238016e-06, 'learning_rate/action_model': 1.9150139422238018e-05, 'epoch': 0.28}) +[2026-10-03 13:46:44,474][starVLA.training.train_starvla][INFO] - Step 3765, Loss: {'train/loss': 7.936339534353465e-05, 'train/grad_norm_pre_clip': 0.036860544234514236, 'timing/data': 0.00024500000290572643, 'timing/model': 0.39023826901393477, 'learning_rate/qwen_vl_interface': 1.9128460560961513e-06, 'learning_rate/action_model': 1.9128460560961514e-05, 'epoch': 0.28}) +[2026-10-03 13:46:45,170][starVLA.training.train_starvla][INFO] - Step 3766, Loss: {'train/loss': 0.37370166182518005, 'train/grad_norm_pre_clip': 27.665321350097656, 'timing/data': 0.0002543900045566261, 'timing/model': 0.3890973980014678, 'learning_rate/qwen_vl_interface': 1.910679541746432e-06, 'learning_rate/action_model': 1.9106795417464323e-05, 'epoch': 0.28}) +[2026-10-03 13:46:46,061][starVLA.training.train_starvla][INFO] - Step 3767, Loss: {'train/loss': 1.823827187763527e-05, 'train/grad_norm_pre_clip': 0.004521776922047138, 'timing/data': 0.000348920002579689, 'timing/model': 0.38896329900308046, 'learning_rate/qwen_vl_interface': 1.9085144000652166e-06, 'learning_rate/action_model': 1.9085144000652166e-05, 'epoch': 0.28}) +[2026-10-03 13:46:46,766][starVLA.training.train_starvla][INFO] - Step 3768, Loss: {'train/loss': 2.219409179815557e-05, 'train/grad_norm_pre_clip': 0.008425825275480747, 'timing/data': 0.00036566000198945403, 'timing/model': 0.3940498279989697, 'learning_rate/qwen_vl_interface': 1.9063506319425116e-06, 'learning_rate/action_model': 1.9063506319425114e-05, 'epoch': 0.28}) +[2026-10-03 13:46:47,467][starVLA.training.train_starvla][INFO] - Step 3769, Loss: {'train/loss': 0.00039472454227507114, 'train/grad_norm_pre_clip': 0.5603799223899841, 'timing/data': 0.000345349995768629, 'timing/model': 0.39146236900705844, 'learning_rate/qwen_vl_interface': 1.9041882382677613e-06, 'learning_rate/action_model': 1.9041882382677612e-05, 'epoch': 0.28}) +[2026-10-03 13:46:48,166][starVLA.training.train_starvla][INFO] - Step 3770, Loss: {'train/loss': 8.274121501017362e-05, 'train/grad_norm_pre_clip': 0.03284596651792526, 'timing/data': 0.0003370500053279102, 'timing/model': 0.39174363799975254, 'learning_rate/qwen_vl_interface': 1.9020272199298452e-06, 'learning_rate/action_model': 1.902027219929845e-05, 'epoch': 0.28}) +[2026-10-03 13:46:48,866][starVLA.training.train_starvla][INFO] - Step 3771, Loss: {'train/loss': 0.06825314462184906, 'train/grad_norm_pre_clip': 177.4210968017578, 'timing/data': 0.00020882999524474144, 'timing/model': 0.39118736900854856, 'learning_rate/qwen_vl_interface': 1.8998675778170736e-06, 'learning_rate/action_model': 1.8998675778170734e-05, 'epoch': 0.28}) +[2026-10-03 13:46:49,777][starVLA.training.train_starvla][INFO] - Step 3772, Loss: {'train/loss': 5.737829997087829e-05, 'train/grad_norm_pre_clip': 0.054171059280633926, 'timing/data': 0.00023817000328563154, 'timing/model': 0.39401777800230775, 'learning_rate/qwen_vl_interface': 1.8977093128171955e-06, 'learning_rate/action_model': 1.8977093128171954e-05, 'epoch': 0.28}) +[2026-10-03 13:46:50,478][starVLA.training.train_starvla][INFO] - Step 3773, Loss: {'train/loss': 0.00697288615629077, 'train/grad_norm_pre_clip': 11.574164390563965, 'timing/data': 0.0003309299936518073, 'timing/model': 0.3916177979990607, 'learning_rate/qwen_vl_interface': 1.8955524258173919e-06, 'learning_rate/action_model': 1.895552425817392e-05, 'epoch': 0.28}) +[2026-10-03 13:46:51,178][starVLA.training.train_starvla][INFO] - Step 3774, Loss: {'train/loss': 6.543671042891219e-05, 'train/grad_norm_pre_clip': 0.01798064447939396, 'timing/data': 0.00026254000840708613, 'timing/model': 0.39191128899983596, 'learning_rate/qwen_vl_interface': 1.8933969177042769e-06, 'learning_rate/action_model': 1.8933969177042768e-05, 'epoch': 0.28}) +[2026-10-03 13:46:51,878][starVLA.training.train_starvla][INFO] - Step 3775, Loss: {'train/loss': 0.32063060998916626, 'train/grad_norm_pre_clip': 30.837547302246094, 'timing/data': 0.00031874999694991857, 'timing/model': 0.3914371280116029, 'learning_rate/qwen_vl_interface': 1.8912427893638996e-06, 'learning_rate/action_model': 1.8912427893638996e-05, 'epoch': 0.28}) +[2026-10-03 13:46:52,581][starVLA.training.train_starvla][INFO] - Step 3776, Loss: {'train/loss': 6.247262354008853e-05, 'train/grad_norm_pre_clip': 0.01269126683473587, 'timing/data': 0.00024337999639101326, 'timing/model': 0.39385877799941227, 'learning_rate/qwen_vl_interface': 1.8890900416817398e-06, 'learning_rate/action_model': 1.8890900416817397e-05, 'epoch': 0.28}) +[2026-10-03 13:46:53,487][starVLA.training.train_starvla][INFO] - Step 3777, Loss: {'train/loss': 0.00012347704614512622, 'train/grad_norm_pre_clip': 0.03931201994419098, 'timing/data': 0.00040842000453267246, 'timing/model': 0.5958424520067638, 'learning_rate/qwen_vl_interface': 1.8869386755427117e-06, 'learning_rate/action_model': 1.8869386755427114e-05, 'epoch': 0.28}) +[2026-10-03 13:46:54,186][starVLA.training.train_starvla][INFO] - Step 3778, Loss: {'train/loss': 3.8767975638620555e-05, 'train/grad_norm_pre_clip': 0.01861114427447319, 'timing/data': 0.00032950000604614615, 'timing/model': 0.39149395799904596, 'learning_rate/qwen_vl_interface': 1.8847886918311597e-06, 'learning_rate/action_model': 1.8847886918311598e-05, 'epoch': 0.28}) +[2026-10-03 13:46:54,889][starVLA.training.train_starvla][INFO] - Step 3779, Loss: {'train/loss': 0.001486544613726437, 'train/grad_norm_pre_clip': 2.246710777282715, 'timing/data': 0.0002462499978719279, 'timing/model': 0.39257975900545716, 'learning_rate/qwen_vl_interface': 1.8826400914308616e-06, 'learning_rate/action_model': 1.8826400914308613e-05, 'epoch': 0.28}) +[2026-10-03 13:46:55,596][starVLA.training.train_starvla][INFO] - Step 3780, Loss: {'train/loss': 0.4194088876247406, 'train/grad_norm_pre_clip': 59.12827682495117, 'timing/data': 0.00026308999804314226, 'timing/model': 0.39240165799856186, 'learning_rate/qwen_vl_interface': 1.880492875225025e-06, 'learning_rate/action_model': 1.8804928752250252e-05, 'epoch': 0.28}) +[2026-10-03 13:46:56,296][starVLA.training.train_starvla][INFO] - Step 3781, Loss: {'train/loss': 0.010345774702727795, 'train/grad_norm_pre_clip': 25.949462890625, 'timing/data': 0.0002106199972331524, 'timing/model': 0.3913873489946127, 'learning_rate/qwen_vl_interface': 1.8783470440962922e-06, 'learning_rate/action_model': 1.878347044096292e-05, 'epoch': 0.28}) +[2026-10-03 13:46:56,996][starVLA.training.train_starvla][INFO] - Step 3782, Loss: {'train/loss': 0.00020453965407796204, 'train/grad_norm_pre_clip': 0.074461430311203, 'timing/data': 0.0002134200040018186, 'timing/model': 0.39184361801017076, 'learning_rate/qwen_vl_interface': 1.8762025989267293e-06, 'learning_rate/action_model': 1.876202598926729e-05, 'epoch': 0.28}) +[2026-10-03 13:46:57,910][starVLA.training.train_starvla][INFO] - Step 3783, Loss: {'train/loss': 6.787419806641992e-06, 'train/grad_norm_pre_clip': 0.0033666500821709633, 'timing/data': 0.00022370000078808516, 'timing/model': 0.39164071800769307, 'learning_rate/qwen_vl_interface': 1.874059540597839e-06, 'learning_rate/action_model': 1.874059540597839e-05, 'epoch': 0.28}) +[2026-10-03 13:46:58,610][starVLA.training.train_starvla][INFO] - Step 3784, Loss: {'train/loss': 9.678202331997454e-06, 'train/grad_norm_pre_clip': 0.0033483055885881186, 'timing/data': 0.0002604100009193644, 'timing/model': 0.39239987899782136, 'learning_rate/qwen_vl_interface': 1.8719178699905508e-06, 'learning_rate/action_model': 1.8719178699905508e-05, 'epoch': 0.28}) +[2026-10-03 13:46:59,309][starVLA.training.train_starvla][INFO] - Step 3785, Loss: {'train/loss': 0.0005252004484646022, 'train/grad_norm_pre_clip': 0.4807358980178833, 'timing/data': 0.000285660004010424, 'timing/model': 0.3910060489870375, 'learning_rate/qwen_vl_interface': 1.8697775879852242e-06, 'learning_rate/action_model': 1.8697775879852242e-05, 'epoch': 0.28}) +[2026-10-03 13:47:00,011][starVLA.training.train_starvla][INFO] - Step 3786, Loss: {'train/loss': 1.2397507816785946e-05, 'train/grad_norm_pre_clip': 0.005818906705826521, 'timing/data': 0.00023173999215941876, 'timing/model': 0.39364707900676876, 'learning_rate/qwen_vl_interface': 1.86763869546165e-06, 'learning_rate/action_model': 1.86763869546165e-05, 'epoch': 0.28}) +[2026-10-03 13:47:00,713][starVLA.training.train_starvla][INFO] - Step 3787, Loss: {'train/loss': 0.0002121970901498571, 'train/grad_norm_pre_clip': 0.10981500893831253, 'timing/data': 0.00034705000871326774, 'timing/model': 0.3919782290031435, 'learning_rate/qwen_vl_interface': 1.865501193299045e-06, 'learning_rate/action_model': 1.8655011932990447e-05, 'epoch': 0.28}) +[2026-10-03 13:47:01,615][starVLA.training.train_starvla][INFO] - Step 3788, Loss: {'train/loss': 5.0814087444450706e-05, 'train/grad_norm_pre_clip': 0.030636165291070938, 'timing/data': 0.00032084000122267753, 'timing/model': 0.3919380900042597, 'learning_rate/qwen_vl_interface': 1.8633650823760555e-06, 'learning_rate/action_model': 1.8633650823760555e-05, 'epoch': 0.28}) +[2026-10-03 13:47:02,317][starVLA.training.train_starvla][INFO] - Step 3789, Loss: {'train/loss': 0.000542264140676707, 'train/grad_norm_pre_clip': 0.2991017699241638, 'timing/data': 0.0003556300071068108, 'timing/model': 0.39193843900284264, 'learning_rate/qwen_vl_interface': 1.8612303635707562e-06, 'learning_rate/action_model': 1.861230363570756e-05, 'epoch': 0.28}) +[2026-10-03 13:47:03,017][starVLA.training.train_starvla][INFO] - Step 3790, Loss: {'train/loss': 1.5787252777954563e-05, 'train/grad_norm_pre_clip': 0.004427797626703978, 'timing/data': 0.000355419993866235, 'timing/model': 0.3912407390016597, 'learning_rate/qwen_vl_interface': 1.8590970377606494e-06, 'learning_rate/action_model': 1.859097037760649e-05, 'epoch': 0.28}) +[2026-10-03 13:47:03,717][starVLA.training.train_starvla][INFO] - Step 3791, Loss: {'train/loss': 0.3683253228664398, 'train/grad_norm_pre_clip': 99.15464782714844, 'timing/data': 0.00019885000074282289, 'timing/model': 0.3912069399957545, 'learning_rate/qwen_vl_interface': 1.8569651058226667e-06, 'learning_rate/action_model': 1.8569651058226666e-05, 'epoch': 0.28}) +[2026-10-03 13:47:04,417][starVLA.training.train_starvla][INFO] - Step 3792, Loss: {'train/loss': 5.511574272532016e-05, 'train/grad_norm_pre_clip': 0.01847865991294384, 'timing/data': 0.00019607901049312204, 'timing/model': 0.3924831400072435, 'learning_rate/qwen_vl_interface': 1.854834568633161e-06, 'learning_rate/action_model': 1.854834568633161e-05, 'epoch': 0.28}) +[2026-10-03 13:47:05,330][starVLA.training.train_starvla][INFO] - Step 3793, Loss: {'train/loss': 0.3074936270713806, 'train/grad_norm_pre_clip': 80.11363220214844, 'timing/data': 0.00036360000376589596, 'timing/model': 0.603360031003831, 'learning_rate/qwen_vl_interface': 1.8527054270679174e-06, 'learning_rate/action_model': 1.8527054270679172e-05, 'epoch': 0.28}) +[2026-10-03 13:47:06,029][starVLA.training.train_starvla][INFO] - Step 3794, Loss: {'train/loss': 0.00030239581246860325, 'train/grad_norm_pre_clip': 0.32521897554397583, 'timing/data': 0.00037855899427086115, 'timing/model': 0.3911040800012415, 'learning_rate/qwen_vl_interface': 1.850577682002146e-06, 'learning_rate/action_model': 1.850577682002146e-05, 'epoch': 0.28}) +[2026-10-03 13:47:06,729][starVLA.training.train_starvla][INFO] - Step 3795, Loss: {'train/loss': 0.00019866577349603176, 'train/grad_norm_pre_clip': 0.040872570127248764, 'timing/data': 0.00020695998682640493, 'timing/model': 0.3908503099955851, 'learning_rate/qwen_vl_interface': 1.848451334310482e-06, 'learning_rate/action_model': 1.848451334310482e-05, 'epoch': 0.28}) +[2026-10-03 13:47:07,430][starVLA.training.train_starvla][INFO] - Step 3796, Loss: {'train/loss': 0.31823912262916565, 'train/grad_norm_pre_clip': 30.930883407592773, 'timing/data': 0.0002145200123777613, 'timing/model': 0.39244196900108363, 'learning_rate/qwen_vl_interface': 1.8463263848669865e-06, 'learning_rate/action_model': 1.8463263848669864e-05, 'epoch': 0.28}) +[2026-10-03 13:47:08,133][starVLA.training.train_starvla][INFO] - Step 3797, Loss: {'train/loss': 1.8618344256537966e-05, 'train/grad_norm_pre_clip': 0.005707944743335247, 'timing/data': 0.00023107999004423618, 'timing/model': 0.39233856899954844, 'learning_rate/qwen_vl_interface': 1.8442028345451468e-06, 'learning_rate/action_model': 1.8442028345451468e-05, 'epoch': 0.28}) +[2026-10-03 13:47:09,037][starVLA.training.train_starvla][INFO] - Step 3798, Loss: {'train/loss': 0.25241243839263916, 'train/grad_norm_pre_clip': 49.1567268371582, 'timing/data': 0.0004776499990839511, 'timing/model': 0.595199343006243, 'learning_rate/qwen_vl_interface': 1.8420806842178746e-06, 'learning_rate/action_model': 1.8420806842178745e-05, 'epoch': 0.28}) +[2026-10-03 13:47:09,734][starVLA.training.train_starvla][INFO] - Step 3799, Loss: {'train/loss': 0.1876004934310913, 'train/grad_norm_pre_clip': 441.7080078125, 'timing/data': 0.00023330999829340726, 'timing/model': 0.38969246999477036, 'learning_rate/qwen_vl_interface': 1.8399599347575034e-06, 'learning_rate/action_model': 1.8399599347575033e-05, 'epoch': 0.28}) +[2026-10-03 13:47:10,434][starVLA.training.train_starvla][INFO] - Step 3800, Loss: {'train/loss': 3.888278661179356e-05, 'train/grad_norm_pre_clip': 0.009258901700377464, 'timing/data': 0.00031906999356579036, 'timing/model': 0.3930901090061525, 'learning_rate/qwen_vl_interface': 1.8378405870357943e-06, 'learning_rate/action_model': 1.837840587035794e-05, 'epoch': 0.28}) +[2026-10-03 13:47:11,133][starVLA.training.train_starvla][INFO] - Step 3801, Loss: {'train/loss': 0.002238982357084751, 'train/grad_norm_pre_clip': 4.548300743103027, 'timing/data': 0.0003604100056691095, 'timing/model': 0.3898660389968427, 'learning_rate/qwen_vl_interface': 1.8357226419239316e-06, 'learning_rate/action_model': 1.8357226419239315e-05, 'epoch': 0.28}) +[2026-10-03 13:47:11,830][starVLA.training.train_starvla][INFO] - Step 3802, Loss: {'train/loss': 0.003305547870695591, 'train/grad_norm_pre_clip': 11.893266677856445, 'timing/data': 0.00020432999008335173, 'timing/model': 0.3900533499981975, 'learning_rate/qwen_vl_interface': 1.8336061002925224e-06, 'learning_rate/action_model': 1.8336061002925224e-05, 'epoch': 0.28}) +[2026-10-03 13:47:12,732][starVLA.training.train_starvla][INFO] - Step 3803, Loss: {'train/loss': 0.0008634806727059186, 'train/grad_norm_pre_clip': 0.6965911984443665, 'timing/data': 0.0003217799967387691, 'timing/model': 0.5948250929941423, 'learning_rate/qwen_vl_interface': 1.8314909630115985e-06, 'learning_rate/action_model': 1.8314909630115983e-05, 'epoch': 0.28}) +[2026-10-03 13:47:13,432][starVLA.training.train_starvla][INFO] - Step 3804, Loss: {'train/loss': 0.0002342334482818842, 'train/grad_norm_pre_clip': 0.12412033975124359, 'timing/data': 0.000327600006130524, 'timing/model': 0.3923734799900558, 'learning_rate/qwen_vl_interface': 1.8293772309506092e-06, 'learning_rate/action_model': 1.829377230950609e-05, 'epoch': 0.28}) +[2026-10-03 13:47:14,130][starVLA.training.train_starvla][INFO] - Step 3805, Loss: {'train/loss': 9.491780474490952e-06, 'train/grad_norm_pre_clip': 0.003920271992683411, 'timing/data': 0.00021267999545671046, 'timing/model': 0.39014628999575507, 'learning_rate/qwen_vl_interface': 1.8272649049784326e-06, 'learning_rate/action_model': 1.8272649049784323e-05, 'epoch': 0.28}) +[2026-10-03 13:47:14,827][starVLA.training.train_starvla][INFO] - Step 3806, Loss: {'train/loss': 0.00013690190098714083, 'train/grad_norm_pre_clip': 0.07282216101884842, 'timing/data': 0.00034032999246846884, 'timing/model': 0.38939393099281006, 'learning_rate/qwen_vl_interface': 1.8251539859633653e-06, 'learning_rate/action_model': 1.825153985963365e-05, 'epoch': 0.28}) +[2026-10-03 13:47:15,523][starVLA.training.train_starvla][INFO] - Step 3807, Loss: {'train/loss': 0.0028193097095936537, 'train/grad_norm_pre_clip': 15.332845687866211, 'timing/data': 0.00021297000057529658, 'timing/model': 0.39015925000421703, 'learning_rate/qwen_vl_interface': 1.8230444747731265e-06, 'learning_rate/action_model': 1.8230444747731264e-05, 'epoch': 0.28}) +[2026-10-03 13:47:16,224][starVLA.training.train_starvla][INFO] - Step 3808, Loss: {'train/loss': 0.2588685154914856, 'train/grad_norm_pre_clip': 143.7310333251953, 'timing/data': 0.00020997000683564693, 'timing/model': 0.3923502600082429, 'learning_rate/qwen_vl_interface': 1.820936372274857e-06, 'learning_rate/action_model': 1.820936372274857e-05, 'epoch': 0.28}) +[2026-10-03 13:47:17,130][starVLA.training.train_starvla][INFO] - Step 3809, Loss: {'train/loss': 9.149139259534422e-06, 'train/grad_norm_pre_clip': 0.0022243051789700985, 'timing/data': 0.0002032000047620386, 'timing/model': 0.3902758000040194, 'learning_rate/qwen_vl_interface': 1.8188296793351178e-06, 'learning_rate/action_model': 1.818829679335118e-05, 'epoch': 0.28}) +[2026-10-03 13:47:17,827][starVLA.training.train_starvla][INFO] - Step 3810, Loss: {'train/loss': 1.1965250450884923e-05, 'train/grad_norm_pre_clip': 0.00382316205650568, 'timing/data': 0.00023786000383552164, 'timing/model': 0.3895455499878153, 'learning_rate/qwen_vl_interface': 1.8167243968198905e-06, 'learning_rate/action_model': 1.8167243968198905e-05, 'epoch': 0.28}) +[2026-10-03 13:47:18,526][starVLA.training.train_starvla][INFO] - Step 3811, Loss: {'train/loss': 8.823165990179405e-05, 'train/grad_norm_pre_clip': 0.03144805133342743, 'timing/data': 0.0002430500026093796, 'timing/model': 0.39128949999576434, 'learning_rate/qwen_vl_interface': 1.8146205255945775e-06, 'learning_rate/action_model': 1.8146205255945777e-05, 'epoch': 0.28}) +[2026-10-03 13:47:19,225][starVLA.training.train_starvla][INFO] - Step 3812, Loss: {'train/loss': 4.306405116949463e-06, 'train/grad_norm_pre_clip': 0.0011329188710078597, 'timing/data': 0.000235719999182038, 'timing/model': 0.391468799993163, 'learning_rate/qwen_vl_interface': 1.8125180665240007e-06, 'learning_rate/action_model': 1.8125180665240008e-05, 'epoch': 0.28}) +[2026-10-03 13:47:19,920][starVLA.training.train_starvla][INFO] - Step 3813, Loss: {'train/loss': 3.833104346995242e-05, 'train/grad_norm_pre_clip': 0.010969756171107292, 'timing/data': 0.00020627000776585191, 'timing/model': 0.39001553000707645, 'learning_rate/qwen_vl_interface': 1.8104170204724025e-06, 'learning_rate/action_model': 1.8104170204724024e-05, 'epoch': 0.28}) +[2026-10-03 13:47:20,821][starVLA.training.train_starvla][INFO] - Step 3814, Loss: {'train/loss': 0.3746539056301117, 'train/grad_norm_pre_clip': 75.74987030029297, 'timing/data': 0.00034250000317115337, 'timing/model': 0.5938811240048381, 'learning_rate/qwen_vl_interface': 1.8083173883034435e-06, 'learning_rate/action_model': 1.8083173883034435e-05, 'epoch': 0.28}) +[2026-10-03 13:47:21,519][starVLA.training.train_starvla][INFO] - Step 3815, Loss: {'train/loss': 8.180627446563449e-06, 'train/grad_norm_pre_clip': 0.006118663586676121, 'timing/data': 0.00021347000438254327, 'timing/model': 0.3914158400002634, 'learning_rate/qwen_vl_interface': 1.8062191708802013e-06, 'learning_rate/action_model': 1.8062191708802012e-05, 'epoch': 0.28}) +[2026-10-03 13:47:22,217][starVLA.training.train_starvla][INFO] - Step 3816, Loss: {'train/loss': 0.00013665789447259158, 'train/grad_norm_pre_clip': 0.09912745654582977, 'timing/data': 0.0003353200008859858, 'timing/model': 0.39121344999875873, 'learning_rate/qwen_vl_interface': 1.804122369065175e-06, 'learning_rate/action_model': 1.8041223690651748e-05, 'epoch': 0.28}) +[2026-10-03 13:47:22,913][starVLA.training.train_starvla][INFO] - Step 3817, Loss: {'train/loss': 7.5771463343699e-06, 'train/grad_norm_pre_clip': 0.0017890722956508398, 'timing/data': 0.00030262999644037336, 'timing/model': 0.389685500995256, 'learning_rate/qwen_vl_interface': 1.8020269837202806e-06, 'learning_rate/action_model': 1.8020269837202804e-05, 'epoch': 0.28}) +[2026-10-03 13:47:23,608][starVLA.training.train_starvla][INFO] - Step 3818, Loss: {'train/loss': 0.00023010509903542697, 'train/grad_norm_pre_clip': 0.2268088012933731, 'timing/data': 0.00024569999368395656, 'timing/model': 0.3890488509932766, 'learning_rate/qwen_vl_interface': 1.7999330157068517e-06, 'learning_rate/action_model': 1.7999330157068515e-05, 'epoch': 0.28}) +[2026-10-03 13:47:24,309][starVLA.training.train_starvla][INFO] - Step 3819, Loss: {'train/loss': 0.0002519756671972573, 'train/grad_norm_pre_clip': 0.07006571441888809, 'timing/data': 0.0003681999951368198, 'timing/model': 0.39132333001180086, 'learning_rate/qwen_vl_interface': 1.7978404658856393e-06, 'learning_rate/action_model': 1.7978404658856394e-05, 'epoch': 0.28}) +[2026-10-03 13:47:25,214][starVLA.training.train_starvla][INFO] - Step 3820, Loss: {'train/loss': 0.002907697344198823, 'train/grad_norm_pre_clip': 5.082350730895996, 'timing/data': 0.0002800700021907687, 'timing/model': 0.39175089000491425, 'learning_rate/qwen_vl_interface': 1.7957493351168126e-06, 'learning_rate/action_model': 1.7957493351168126e-05, 'epoch': 0.28}) +[2026-10-03 13:47:25,913][starVLA.training.train_starvla][INFO] - Step 3821, Loss: {'train/loss': 5.0403581553837284e-05, 'train/grad_norm_pre_clip': 0.017060546204447746, 'timing/data': 0.00034564000088721514, 'timing/model': 0.39022931999352295, 'learning_rate/qwen_vl_interface': 1.7936596242599552e-06, 'learning_rate/action_model': 1.7936596242599553e-05, 'epoch': 0.28}) +[2026-10-03 13:47:26,611][starVLA.training.train_starvla][INFO] - Step 3822, Loss: {'train/loss': 3.426697730901651e-05, 'train/grad_norm_pre_clip': 0.01880459673702717, 'timing/data': 0.0001972699974430725, 'timing/model': 0.3917695609998191, 'learning_rate/qwen_vl_interface': 1.7915713341740692e-06, 'learning_rate/action_model': 1.791571334174069e-05, 'epoch': 0.28}) +[2026-10-03 13:47:27,312][starVLA.training.train_starvla][INFO] - Step 3823, Loss: {'train/loss': 0.0001386443618685007, 'train/grad_norm_pre_clip': 0.10086924582719803, 'timing/data': 0.0003294100024504587, 'timing/model': 0.39170603100501467, 'learning_rate/qwen_vl_interface': 1.7894844657175716e-06, 'learning_rate/action_model': 1.7894844657175717e-05, 'epoch': 0.28}) +[2026-10-03 13:47:28,013][starVLA.training.train_starvla][INFO] - Step 3824, Loss: {'train/loss': 0.0001969841541722417, 'train/grad_norm_pre_clip': 0.300903856754303, 'timing/data': 0.00026852999872062355, 'timing/model': 0.392664771003183, 'learning_rate/qwen_vl_interface': 1.7873990197482962e-06, 'learning_rate/action_model': 1.787399019748296e-05, 'epoch': 0.28}) +[2026-10-03 13:47:28,905][starVLA.training.train_starvla][INFO] - Step 3825, Loss: {'train/loss': 1.352986873826012e-05, 'train/grad_norm_pre_clip': 0.0037495195865631104, 'timing/data': 0.00023925000277813524, 'timing/model': 0.3883888510026736, 'learning_rate/qwen_vl_interface': 1.7853149971234883e-06, 'learning_rate/action_model': 1.785314997123488e-05, 'epoch': 0.28}) +[2026-10-03 13:47:29,601][starVLA.training.train_starvla][INFO] - Step 3826, Loss: {'train/loss': 0.11782759428024292, 'train/grad_norm_pre_clip': 113.70540618896484, 'timing/data': 0.0002036299993051216, 'timing/model': 0.3897333310014801, 'learning_rate/qwen_vl_interface': 1.7832323986998128e-06, 'learning_rate/action_model': 1.783232398699813e-05, 'epoch': 0.28}) +[2026-10-03 13:47:30,299][starVLA.training.train_starvla][INFO] - Step 3827, Loss: {'train/loss': 0.00017069422756321728, 'train/grad_norm_pre_clip': 0.12596799433231354, 'timing/data': 0.00020414999744389206, 'timing/model': 0.39100610099558253, 'learning_rate/qwen_vl_interface': 1.7811512253333468e-06, 'learning_rate/action_model': 1.7811512253333467e-05, 'epoch': 0.28}) +[2026-10-03 13:47:31,001][starVLA.training.train_starvla][INFO] - Step 3828, Loss: {'train/loss': 2.8241422114660963e-05, 'train/grad_norm_pre_clip': 0.008043724112212658, 'timing/data': 0.00030323999817483127, 'timing/model': 0.3933632189873606, 'learning_rate/qwen_vl_interface': 1.7790714778795823e-06, 'learning_rate/action_model': 1.779071477879582e-05, 'epoch': 0.28}) +[2026-10-03 13:47:31,702][starVLA.training.train_starvla][INFO] - Step 3829, Loss: {'train/loss': 6.0896436480106786e-05, 'train/grad_norm_pre_clip': 0.08420077711343765, 'timing/data': 0.0002609599905554205, 'timing/model': 0.391550731001189, 'learning_rate/qwen_vl_interface': 1.7769931571934246e-06, 'learning_rate/action_model': 1.7769931571934246e-05, 'epoch': 0.28}) +[2026-10-03 13:47:32,596][starVLA.training.train_starvla][INFO] - Step 3830, Loss: {'train/loss': 0.5039638876914978, 'train/grad_norm_pre_clip': 148.52452087402344, 'timing/data': 0.00021229000412859023, 'timing/model': 0.5874806960055139, 'learning_rate/qwen_vl_interface': 1.7749162641291933e-06, 'learning_rate/action_model': 1.7749162641291933e-05, 'epoch': 0.28}) +[2026-10-03 13:47:33,292][starVLA.training.train_starvla][INFO] - Step 3831, Loss: {'train/loss': 8.972641080617905e-05, 'train/grad_norm_pre_clip': 0.08974558115005493, 'timing/data': 0.00023286999203264713, 'timing/model': 0.38945590099319816, 'learning_rate/qwen_vl_interface': 1.7728407995406206e-06, 'learning_rate/action_model': 1.7728407995406205e-05, 'epoch': 0.28}) +[2026-10-03 13:47:33,993][starVLA.training.train_starvla][INFO] - Step 3832, Loss: {'train/loss': 3.5240932447777595e-06, 'train/grad_norm_pre_clip': 0.0012885876931250095, 'timing/data': 0.0002905099972849712, 'timing/model': 0.39181535999523476, 'learning_rate/qwen_vl_interface': 1.7707667642808513e-06, 'learning_rate/action_model': 1.7707667642808512e-05, 'epoch': 0.28}) +[2026-10-03 13:47:34,689][starVLA.training.train_starvla][INFO] - Step 3833, Loss: {'train/loss': 2.2297768737189472e-05, 'train/grad_norm_pre_clip': 0.015100005082786083, 'timing/data': 0.00020000999211333692, 'timing/model': 0.3894055010023294, 'learning_rate/qwen_vl_interface': 1.7686941592024423e-06, 'learning_rate/action_model': 1.7686941592024423e-05, 'epoch': 0.28}) +[2026-10-03 13:47:35,387][starVLA.training.train_starvla][INFO] - Step 3834, Loss: {'train/loss': 0.5140769481658936, 'train/grad_norm_pre_clip': 98.62762451171875, 'timing/data': 0.00019897001038771123, 'timing/model': 0.391033651001635, 'learning_rate/qwen_vl_interface': 1.7666229851573662e-06, 'learning_rate/action_model': 1.7666229851573664e-05, 'epoch': 0.28}) +[2026-10-03 13:47:36,086][starVLA.training.train_starvla][INFO] - Step 3835, Loss: {'train/loss': 0.8844582438468933, 'train/grad_norm_pre_clip': 82.77363586425781, 'timing/data': 0.000348490008036606, 'timing/model': 0.39086310101265553, 'learning_rate/qwen_vl_interface': 1.7645532429970006e-06, 'learning_rate/action_model': 1.7645532429970004e-05, 'epoch': 0.28}) +[2026-10-03 13:47:36,994][starVLA.training.train_starvla][INFO] - Step 3836, Loss: {'train/loss': 7.890089364082087e-06, 'train/grad_norm_pre_clip': 0.0017396537587046623, 'timing/data': 0.0002089000045089051, 'timing/model': 0.3937276600045152, 'learning_rate/qwen_vl_interface': 1.7624849335721402e-06, 'learning_rate/action_model': 1.7624849335721402e-05, 'epoch': 0.28}) +[2026-10-03 13:47:37,693][starVLA.training.train_starvla][INFO] - Step 3837, Loss: {'train/loss': 3.2829786505317315e-05, 'train/grad_norm_pre_clip': 0.0094381682574749, 'timing/data': 0.00031532999128103256, 'timing/model': 0.39095542101131286, 'learning_rate/qwen_vl_interface': 1.7604180577329883e-06, 'learning_rate/action_model': 1.7604180577329883e-05, 'epoch': 0.28}) +[2026-10-03 13:47:38,392][starVLA.training.train_starvla][INFO] - Step 3838, Loss: {'train/loss': 0.00010173647024203092, 'train/grad_norm_pre_clip': 0.09513554722070694, 'timing/data': 0.0002980590070364997, 'timing/model': 0.39018224099709187, 'learning_rate/qwen_vl_interface': 1.7583526163291604e-06, 'learning_rate/action_model': 1.7583526163291604e-05, 'epoch': 0.28}) +[2026-10-03 13:47:39,092][starVLA.training.train_starvla][INFO] - Step 3839, Loss: {'train/loss': 0.00010255105735268444, 'train/grad_norm_pre_clip': 0.05145310238003731, 'timing/data': 0.00023534000501967967, 'timing/model': 0.39179121000051964, 'learning_rate/qwen_vl_interface': 1.7562886102096807e-06, 'learning_rate/action_model': 1.7562886102096807e-05, 'epoch': 0.28}) +[2026-10-03 13:47:39,792][starVLA.training.train_starvla][INFO] - Step 3840, Loss: {'train/loss': 7.682903378736228e-05, 'train/grad_norm_pre_clip': 0.026896493509411812, 'timing/data': 0.00031100999331101775, 'timing/model': 0.3922680099931313, 'learning_rate/qwen_vl_interface': 1.7542260402229838e-06, 'learning_rate/action_model': 1.7542260402229838e-05, 'epoch': 0.28}) +[2026-10-03 13:47:40,695][starVLA.training.train_starvla][INFO] - Step 3841, Loss: {'train/loss': 2.724376281548757e-05, 'train/grad_norm_pre_clip': 0.005709703546017408, 'timing/data': 0.0002560900029493496, 'timing/model': 0.3896882309927605, 'learning_rate/qwen_vl_interface': 1.7521649072169152e-06, 'learning_rate/action_model': 1.752164907216915e-05, 'epoch': 0.28}) +[2026-10-03 13:47:41,392][starVLA.training.train_starvla][INFO] - Step 3842, Loss: {'train/loss': 0.31305426359176636, 'train/grad_norm_pre_clip': 145.11524963378906, 'timing/data': 0.00032088998705148697, 'timing/model': 0.3894486310018692, 'learning_rate/qwen_vl_interface': 1.7501052120387285e-06, 'learning_rate/action_model': 1.7501052120387285e-05, 'epoch': 0.28}) +[2026-10-03 13:47:42,090][starVLA.training.train_starvla][INFO] - Step 3843, Loss: {'train/loss': 5.617675014946144e-06, 'train/grad_norm_pre_clip': 0.0016865745419636369, 'timing/data': 0.00020838000637013465, 'timing/model': 0.3907103020028444, 'learning_rate/qwen_vl_interface': 1.7480469555350866e-06, 'learning_rate/action_model': 1.7480469555350865e-05, 'epoch': 0.28}) +[2026-10-03 13:47:42,790][starVLA.training.train_starvla][INFO] - Step 3844, Loss: {'train/loss': 1.9526552932802588e-05, 'train/grad_norm_pre_clip': 0.004521623719483614, 'timing/data': 0.00036732001171913, 'timing/model': 0.39247014999273233, 'learning_rate/qwen_vl_interface': 1.7459901385520619e-06, 'learning_rate/action_model': 1.745990138552062e-05, 'epoch': 0.28}) +[2026-10-03 13:47:43,491][starVLA.training.train_starvla][INFO] - Step 3845, Loss: {'train/loss': 6.742709501850186e-06, 'train/grad_norm_pre_clip': 0.0019002406625077128, 'timing/data': 0.00034016999416053295, 'timing/model': 0.3906474309915211, 'learning_rate/qwen_vl_interface': 1.7439347619351324e-06, 'learning_rate/action_model': 1.7439347619351324e-05, 'epoch': 0.28}) +[2026-10-03 13:47:44,400][starVLA.training.train_starvla][INFO] - Step 3846, Loss: {'train/loss': 4.485368845053017e-05, 'train/grad_norm_pre_clip': 0.014611602760851383, 'timing/data': 0.0003636399924289435, 'timing/model': 0.3914158010011306, 'learning_rate/qwen_vl_interface': 1.741880826529186e-06, 'learning_rate/action_model': 1.741880826529186e-05, 'epoch': 0.28}) +[2026-10-03 13:47:45,099][starVLA.training.train_starvla][INFO] - Step 3847, Loss: {'train/loss': 0.02284240536391735, 'train/grad_norm_pre_clip': 88.71332550048828, 'timing/data': 0.0002078300021821633, 'timing/model': 0.3903689909930108, 'learning_rate/qwen_vl_interface': 1.7398283331785205e-06, 'learning_rate/action_model': 1.7398283331785203e-05, 'epoch': 0.28}) +[2026-10-03 13:47:45,800][starVLA.training.train_starvla][INFO] - Step 3848, Loss: {'train/loss': 0.00021630636183544993, 'train/grad_norm_pre_clip': 0.3079407811164856, 'timing/data': 0.00020553999638650566, 'timing/model': 0.392646391002927, 'learning_rate/qwen_vl_interface': 1.737777282726835e-06, 'learning_rate/action_model': 1.737777282726835e-05, 'epoch': 0.28}) +[2026-10-03 13:47:46,497][starVLA.training.train_starvla][INFO] - Step 3849, Loss: {'train/loss': 6.839459274488036e-06, 'train/grad_norm_pre_clip': 0.0021734361071139574, 'timing/data': 0.00020494000636972487, 'timing/model': 0.38853626100171823, 'learning_rate/qwen_vl_interface': 1.7357276760172403e-06, 'learning_rate/action_model': 1.7357276760172404e-05, 'epoch': 0.28}) +[2026-10-03 13:47:47,196][starVLA.training.train_starvla][INFO] - Step 3850, Loss: {'train/loss': 1.30754515339504e-05, 'train/grad_norm_pre_clip': 0.004027821123600006, 'timing/data': 0.00021099898731335998, 'timing/model': 0.3912291419983376, 'learning_rate/qwen_vl_interface': 1.733679513892252e-06, 'learning_rate/action_model': 1.7336795138922517e-05, 'epoch': 0.28}) +[2026-10-03 13:47:48,100][starVLA.training.train_starvla][INFO] - Step 3851, Loss: {'train/loss': 0.5812147259712219, 'train/grad_norm_pre_clip': 209.63880920410156, 'timing/data': 0.00020093000784981996, 'timing/model': 0.5963144840061432, 'learning_rate/qwen_vl_interface': 1.7316327971937926e-06, 'learning_rate/action_model': 1.7316327971937926e-05, 'epoch': 0.28}) +[2026-10-03 13:47:48,801][starVLA.training.train_starvla][INFO] - Step 3852, Loss: {'train/loss': 5.949151091044769e-05, 'train/grad_norm_pre_clip': 0.07252620905637741, 'timing/data': 0.0002593399985926226, 'timing/model': 0.3919840309972642, 'learning_rate/qwen_vl_interface': 1.7295875267631904e-06, 'learning_rate/action_model': 1.7295875267631904e-05, 'epoch': 0.28}) +[2026-10-03 13:47:49,502][starVLA.training.train_starvla][INFO] - Step 3853, Loss: {'train/loss': 1.5190711565082893e-05, 'train/grad_norm_pre_clip': 0.006382564082741737, 'timing/data': 0.0003193099983036518, 'timing/model': 0.3923818009934621, 'learning_rate/qwen_vl_interface': 1.7275437034411778e-06, 'learning_rate/action_model': 1.7275437034411776e-05, 'epoch': 0.28}) +[2026-10-03 13:47:50,201][starVLA.training.train_starvla][INFO] - Step 3854, Loss: {'train/loss': 0.0010545600671321154, 'train/grad_norm_pre_clip': 0.41870975494384766, 'timing/data': 0.0003339399991091341, 'timing/model': 0.39108036100515164, 'learning_rate/qwen_vl_interface': 1.7255013280678941e-06, 'learning_rate/action_model': 1.725501328067894e-05, 'epoch': 0.28}) +[2026-10-03 13:47:50,900][starVLA.training.train_starvla][INFO] - Step 3855, Loss: {'train/loss': 4.888209514319897e-05, 'train/grad_norm_pre_clip': 0.03883783519268036, 'timing/data': 0.00020992000645492226, 'timing/model': 0.3912307110003894, 'learning_rate/qwen_vl_interface': 1.7234604014828824e-06, 'learning_rate/action_model': 1.7234604014828823e-05, 'epoch': 0.28}) +[2026-10-03 13:47:51,806][starVLA.training.train_starvla][INFO] - Step 3856, Loss: {'train/loss': 9.655760550231207e-06, 'train/grad_norm_pre_clip': 0.004362739622592926, 'timing/data': 0.00020675999985542148, 'timing/model': 0.5967325550009264, 'learning_rate/qwen_vl_interface': 1.7214209245250906e-06, 'learning_rate/action_model': 1.7214209245250905e-05, 'epoch': 0.28}) +[2026-10-03 13:47:52,507][starVLA.training.train_starvla][INFO] - Step 3857, Loss: {'train/loss': 3.645304241217673e-05, 'train/grad_norm_pre_clip': 0.03734660893678665, 'timing/data': 0.00022649001039098948, 'timing/model': 0.39144543099973816, 'learning_rate/qwen_vl_interface': 1.7193828980328723e-06, 'learning_rate/action_model': 1.719382898032872e-05, 'epoch': 0.28}) +[2026-10-03 13:47:53,206][starVLA.training.train_starvla][INFO] - Step 3858, Loss: {'train/loss': 9.988459351006895e-05, 'train/grad_norm_pre_clip': 0.0486258789896965, 'timing/data': 0.00033440999686717987, 'timing/model': 0.3910175210039597, 'learning_rate/qwen_vl_interface': 1.71734632284398e-06, 'learning_rate/action_model': 1.7173463228439798e-05, 'epoch': 0.28}) +[2026-10-03 13:47:53,904][starVLA.training.train_starvla][INFO] - Step 3859, Loss: {'train/loss': 6.377258250722662e-05, 'train/grad_norm_pre_clip': 0.04138823598623276, 'timing/data': 0.00020215999393258244, 'timing/model': 0.39044677100901026, 'learning_rate/qwen_vl_interface': 1.7153111997955744e-06, 'learning_rate/action_model': 1.7153111997955743e-05, 'epoch': 0.28}) +[2026-10-03 13:47:54,606][starVLA.training.train_starvla][INFO] - Step 3860, Loss: {'train/loss': 0.292636901140213, 'train/grad_norm_pre_clip': 110.98216247558594, 'timing/data': 0.0002980599965667352, 'timing/model': 0.39329151100537274, 'learning_rate/qwen_vl_interface': 1.7132775297242175e-06, 'learning_rate/action_model': 1.7132775297242172e-05, 'epoch': 0.28}) +[2026-10-03 13:47:55,304][starVLA.training.train_starvla][INFO] - Step 3861, Loss: {'train/loss': 1.8416958482703194e-05, 'train/grad_norm_pre_clip': 0.0072222016751766205, 'timing/data': 0.00020803000370506197, 'timing/model': 0.389784981001867, 'learning_rate/qwen_vl_interface': 1.7112453134658747e-06, 'learning_rate/action_model': 1.7112453134658747e-05, 'epoch': 0.28}) +[2026-10-03 13:47:56,207][starVLA.training.train_starvla][INFO] - Step 3862, Loss: {'train/loss': 0.5806064009666443, 'train/grad_norm_pre_clip': 52.66322708129883, 'timing/data': 0.00033368999720551074, 'timing/model': 0.38989846099866554, 'learning_rate/qwen_vl_interface': 1.7092145518559129e-06, 'learning_rate/action_model': 1.709214551855913e-05, 'epoch': 0.28}) +[2026-10-03 13:47:56,905][starVLA.training.train_starvla][INFO] - Step 3863, Loss: {'train/loss': 1.712837729428429e-05, 'train/grad_norm_pre_clip': 0.013198185712099075, 'timing/data': 0.00025569899298716336, 'timing/model': 0.39058706200739834, 'learning_rate/qwen_vl_interface': 1.7071852457291013e-06, 'learning_rate/action_model': 1.707185245729101e-05, 'epoch': 0.28}) +[2026-10-03 13:47:57,607][starVLA.training.train_starvla][INFO] - Step 3864, Loss: {'train/loss': 1.7976864910451695e-05, 'train/grad_norm_pre_clip': 0.012358984909951687, 'timing/data': 0.000348469999153167, 'timing/model': 0.3926777610031422, 'learning_rate/qwen_vl_interface': 1.7051573959196122e-06, 'learning_rate/action_model': 1.7051573959196122e-05, 'epoch': 0.28}) +[2026-10-03 13:47:58,303][starVLA.training.train_starvla][INFO] - Step 3865, Loss: {'train/loss': 0.2562417984008789, 'train/grad_norm_pre_clip': 187.77284240722656, 'timing/data': 0.00032845999521669, 'timing/model': 0.3897910219966434, 'learning_rate/qwen_vl_interface': 1.7031310032610166e-06, 'learning_rate/action_model': 1.7031310032610163e-05, 'epoch': 0.28}) +[2026-10-03 13:47:59,000][starVLA.training.train_starvla][INFO] - Step 3866, Loss: {'train/loss': 2.8503467547125183e-05, 'train/grad_norm_pre_clip': 0.011086824350059032, 'timing/data': 0.0002543500013416633, 'timing/model': 0.3898334419936873, 'learning_rate/qwen_vl_interface': 1.701106068586289e-06, 'learning_rate/action_model': 1.701106068586289e-05, 'epoch': 0.28}) +[2026-10-03 13:47:59,915][starVLA.training.train_starvla][INFO] - Step 3867, Loss: {'train/loss': 7.1066060627345e-05, 'train/grad_norm_pre_clip': 0.2353057712316513, 'timing/data': 0.00020063000556547195, 'timing/model': 0.6067798840085743, 'learning_rate/qwen_vl_interface': 1.6990825927278032e-06, 'learning_rate/action_model': 1.699082592727803e-05, 'epoch': 0.28}) +[2026-10-03 13:48:00,618][starVLA.training.train_starvla][INFO] - Step 3868, Loss: {'train/loss': 2.888844028348103e-05, 'train/grad_norm_pre_clip': 0.03321884572505951, 'timing/data': 0.0002615799894556403, 'timing/model': 0.39277148099790793, 'learning_rate/qwen_vl_interface': 1.6970605765173344e-06, 'learning_rate/action_model': 1.6970605765173346e-05, 'epoch': 0.28}) +[2026-10-03 13:48:01,316][starVLA.training.train_starvla][INFO] - Step 3869, Loss: {'train/loss': 0.0008697434095665812, 'train/grad_norm_pre_clip': 7.533980369567871, 'timing/data': 0.0003369899932295084, 'timing/model': 0.39090476201090496, 'learning_rate/qwen_vl_interface': 1.6950400207860557e-06, 'learning_rate/action_model': 1.6950400207860557e-05, 'epoch': 0.28}) +[2026-10-03 13:48:02,016][starVLA.training.train_starvla][INFO] - Step 3870, Loss: {'train/loss': 2.9374408768489957e-05, 'train/grad_norm_pre_clip': 0.008907146751880646, 'timing/data': 0.0004204600118100643, 'timing/model': 0.39136333200440276, 'learning_rate/qwen_vl_interface': 1.6930209263645412e-06, 'learning_rate/action_model': 1.6930209263645412e-05, 'epoch': 0.28}) +[2026-10-03 13:48:02,718][starVLA.training.train_starvla][INFO] - Step 3871, Loss: {'train/loss': 5.572939699050039e-06, 'train/grad_norm_pre_clip': 0.0016416197177022696, 'timing/data': 0.0002134990063495934, 'timing/model': 0.3919157920026919, 'learning_rate/qwen_vl_interface': 1.6910032940827656e-06, 'learning_rate/action_model': 1.6910032940827657e-05, 'epoch': 0.28}) +[2026-10-03 13:48:03,419][starVLA.training.train_starvla][INFO] - Step 3872, Loss: {'train/loss': 2.2117968910606578e-05, 'train/grad_norm_pre_clip': 0.005852223839610815, 'timing/data': 0.0003309700114186853, 'timing/model': 0.3921134519914631, 'learning_rate/qwen_vl_interface': 1.6889871247701003e-06, 'learning_rate/action_model': 1.6889871247701002e-05, 'epoch': 0.28}) +[2026-10-03 13:48:04,320][starVLA.training.train_starvla][INFO] - Step 3873, Loss: {'train/loss': 4.939681730320444e-06, 'train/grad_norm_pre_clip': 0.0014511335175484419, 'timing/data': 0.0003872700035572052, 'timing/model': 0.39069845199992415, 'learning_rate/qwen_vl_interface': 1.6869724192553168e-06, 'learning_rate/action_model': 1.686972419255317e-05, 'epoch': 0.28}) +[2026-10-03 13:48:05,022][starVLA.training.train_starvla][INFO] - Step 3874, Loss: {'train/loss': 8.572489605285227e-05, 'train/grad_norm_pre_clip': 0.02219974808394909, 'timing/data': 0.00020537999807856977, 'timing/model': 0.39323076099390164, 'learning_rate/qwen_vl_interface': 1.6849591783665847e-06, 'learning_rate/action_model': 1.6849591783665847e-05, 'epoch': 0.28}) +[2026-10-03 13:48:05,723][starVLA.training.train_starvla][INFO] - Step 3875, Loss: {'train/loss': 4.372137482278049e-05, 'train/grad_norm_pre_clip': 0.009545959532260895, 'timing/data': 0.00020562901045195758, 'timing/model': 0.3916921720083337, 'learning_rate/qwen_vl_interface': 1.6829474029314713e-06, 'learning_rate/action_model': 1.6829474029314713e-05, 'epoch': 0.28}) +[2026-10-03 13:48:06,424][starVLA.training.train_starvla][INFO] - Step 3876, Loss: {'train/loss': 2.8687813028227538e-05, 'train/grad_norm_pre_clip': 0.0073958043940365314, 'timing/data': 0.00021128999651409686, 'timing/model': 0.39283451199298725, 'learning_rate/qwen_vl_interface': 1.6809370937769408e-06, 'learning_rate/action_model': 1.680937093776941e-05, 'epoch': 0.28}) +[2026-10-03 13:48:07,124][starVLA.training.train_starvla][INFO] - Step 3877, Loss: {'train/loss': 2.625277375045698e-05, 'train/grad_norm_pre_clip': 0.03091462515294552, 'timing/data': 0.0003466100024525076, 'timing/model': 0.3907210220058914, 'learning_rate/qwen_vl_interface': 1.678928251729356e-06, 'learning_rate/action_model': 1.6789282517293557e-05, 'epoch': 0.28}) +[2026-10-03 13:48:08,022][starVLA.training.train_starvla][INFO] - Step 3878, Loss: {'train/loss': 6.0828377172583714e-05, 'train/grad_norm_pre_clip': 0.03736587241292, 'timing/data': 0.00021055000252090394, 'timing/model': 0.39267367200227454, 'learning_rate/qwen_vl_interface': 1.6769208776144767e-06, 'learning_rate/action_model': 1.6769208776144764e-05, 'epoch': 0.28}) +[2026-10-03 13:48:08,722][starVLA.training.train_starvla][INFO] - Step 3879, Loss: {'train/loss': 1.3202017726143822e-05, 'train/grad_norm_pre_clip': 0.004270625300705433, 'timing/data': 0.0003174699959345162, 'timing/model': 0.39106167199497577, 'learning_rate/qwen_vl_interface': 1.6749149722574564e-06, 'learning_rate/action_model': 1.6749149722574564e-05, 'epoch': 0.28}) +[2026-10-03 13:48:09,421][starVLA.training.train_starvla][INFO] - Step 3880, Loss: {'train/loss': 0.030472584068775177, 'train/grad_norm_pre_clip': 187.77638244628906, 'timing/data': 0.000208480007131584, 'timing/model': 0.3912041520088678, 'learning_rate/qwen_vl_interface': 1.6729105364828486e-06, 'learning_rate/action_model': 1.6729105364828484e-05, 'epoch': 0.28}) +[2026-10-03 13:48:10,119][starVLA.training.train_starvla][INFO] - Step 3881, Loss: {'train/loss': 6.198785740707535e-06, 'train/grad_norm_pre_clip': 0.0026180206332355738, 'timing/data': 0.00019986000552307814, 'timing/model': 0.3911843109963229, 'learning_rate/qwen_vl_interface': 1.6709075711146e-06, 'learning_rate/action_model': 1.6709075711146e-05, 'epoch': 0.28}) +[2026-10-03 13:48:10,819][starVLA.training.train_starvla][INFO] - Step 3882, Loss: {'train/loss': 8.597878149885219e-06, 'train/grad_norm_pre_clip': 0.0018321506213396788, 'timing/data': 0.00033907999750226736, 'timing/model': 0.39107653200335335, 'learning_rate/qwen_vl_interface': 1.6689060769760556e-06, 'learning_rate/action_model': 1.6689060769760557e-05, 'epoch': 0.28}) +[2026-10-03 13:48:11,719][starVLA.training.train_starvla][INFO] - Step 3883, Loss: {'train/loss': 1.605498073331546e-05, 'train/grad_norm_pre_clip': 0.0037776175886392593, 'timing/data': 0.00029950999305583537, 'timing/model': 0.5931825270090485, 'learning_rate/qwen_vl_interface': 1.6669060548899528e-06, 'learning_rate/action_model': 1.6669060548899525e-05, 'epoch': 0.28}) +[2026-10-03 13:48:12,418][starVLA.training.train_starvla][INFO] - Step 3884, Loss: {'train/loss': 3.8976162613835186e-05, 'train/grad_norm_pre_clip': 0.03126358613371849, 'timing/data': 0.00024731000303290784, 'timing/model': 0.3924640509940218, 'learning_rate/qwen_vl_interface': 1.6649075056784253e-06, 'learning_rate/action_model': 1.6649075056784253e-05, 'epoch': 0.28}) +[2026-10-03 13:48:13,117][starVLA.training.train_starvla][INFO] - Step 3885, Loss: {'train/loss': 0.634719967842102, 'train/grad_norm_pre_clip': 63.720001220703125, 'timing/data': 0.0003182299988111481, 'timing/model': 0.3909891119983513, 'learning_rate/qwen_vl_interface': 1.6629104301630019e-06, 'learning_rate/action_model': 1.662910430163002e-05, 'epoch': 0.28}) +[2026-10-03 13:48:13,815][starVLA.training.train_starvla][INFO] - Step 3886, Loss: {'train/loss': 0.1067066639661789, 'train/grad_norm_pre_clip': 144.1343231201172, 'timing/data': 0.00021343999833334237, 'timing/model': 0.38989853199745994, 'learning_rate/qwen_vl_interface': 1.6609148291646043e-06, 'learning_rate/action_model': 1.6609148291646042e-05, 'epoch': 0.28}) +[2026-10-03 13:48:14,512][starVLA.training.train_starvla][INFO] - Step 3887, Loss: {'train/loss': 1.1272277333773673e-05, 'train/grad_norm_pre_clip': 0.005674062296748161, 'timing/data': 0.00030308999703265727, 'timing/model': 0.39019848199677654, 'learning_rate/qwen_vl_interface': 1.658920703503549e-06, 'learning_rate/action_model': 1.6589207035035488e-05, 'epoch': 0.28}) +[2026-10-03 13:48:15,211][starVLA.training.train_starvla][INFO] - Step 3888, Loss: {'train/loss': 4.5671663428947795e-06, 'train/grad_norm_pre_clip': 0.0011018558871001005, 'timing/data': 0.0002000199892790988, 'timing/model': 0.39133038200088777, 'learning_rate/qwen_vl_interface': 1.6569280539995467e-06, 'learning_rate/action_model': 1.6569280539995467e-05, 'epoch': 0.28}) +[2026-10-03 13:48:16,115][starVLA.training.train_starvla][INFO] - Step 3889, Loss: {'train/loss': 0.001070646452717483, 'train/grad_norm_pre_clip': 1.694986343383789, 'timing/data': 0.00027413999487180263, 'timing/model': 0.39198246199521236, 'learning_rate/qwen_vl_interface': 1.6549368814716977e-06, 'learning_rate/action_model': 1.6549368814716977e-05, 'epoch': 0.28}) +[2026-10-03 13:48:16,813][starVLA.training.train_starvla][INFO] - Step 3890, Loss: {'train/loss': 1.498204073868692e-05, 'train/grad_norm_pre_clip': 0.006161503493785858, 'timing/data': 0.00020604999735951424, 'timing/model': 0.38982055299857166, 'learning_rate/qwen_vl_interface': 1.6529471867385009e-06, 'learning_rate/action_model': 1.6529471867385007e-05, 'epoch': 0.28}) +[2026-10-03 13:48:17,510][starVLA.training.train_starvla][INFO] - Step 3891, Loss: {'train/loss': 0.5125472545623779, 'train/grad_norm_pre_clip': 79.60406494140625, 'timing/data': 0.00020584999583661556, 'timing/model': 0.3897257820062805, 'learning_rate/qwen_vl_interface': 1.6509589706178417e-06, 'learning_rate/action_model': 1.6509589706178417e-05, 'epoch': 0.28}) +[2026-10-03 13:48:18,208][starVLA.training.train_starvla][INFO] - Step 3892, Loss: {'train/loss': 1.5452093066414818e-05, 'train/grad_norm_pre_clip': 0.003193238051608205, 'timing/data': 0.0003137099993182346, 'timing/model': 0.39052398200146854, 'learning_rate/qwen_vl_interface': 1.648972233927002e-06, 'learning_rate/action_model': 1.648972233927002e-05, 'epoch': 0.28}) +[2026-10-03 13:48:18,909][starVLA.training.train_starvla][INFO] - Step 3893, Loss: {'train/loss': 0.0001197253805003129, 'train/grad_norm_pre_clip': 0.08607421070337296, 'timing/data': 0.00024478998966515064, 'timing/model': 0.39198508199478965, 'learning_rate/qwen_vl_interface': 1.646986977482654e-06, 'learning_rate/action_model': 1.6469869774826538e-05, 'epoch': 0.28}) +[2026-10-03 13:48:19,814][starVLA.training.train_starvla][INFO] - Step 3894, Loss: {'train/loss': 9.775033504411113e-06, 'train/grad_norm_pre_clip': 0.002501495648175478, 'timing/data': 0.00034383998718112707, 'timing/model': 0.3897742319968529, 'learning_rate/qwen_vl_interface': 1.6450032021008618e-06, 'learning_rate/action_model': 1.6450032021008617e-05, 'epoch': 0.29}) +[2026-10-03 13:48:20,510][starVLA.training.train_starvla][INFO] - Step 3895, Loss: {'train/loss': 0.0001203493302455172, 'train/grad_norm_pre_clip': 0.1615811586380005, 'timing/data': 0.00031257000227924436, 'timing/model': 0.3892066620028345, 'learning_rate/qwen_vl_interface': 1.6430209085970812e-06, 'learning_rate/action_model': 1.643020908597081e-05, 'epoch': 0.29}) +[2026-10-03 13:48:21,210][starVLA.training.train_starvla][INFO] - Step 3896, Loss: {'train/loss': 0.0006044054171070457, 'train/grad_norm_pre_clip': 3.237173080444336, 'timing/data': 0.00031337000837083906, 'timing/model': 0.3929193719959585, 'learning_rate/qwen_vl_interface': 1.641040097786157e-06, 'learning_rate/action_model': 1.641040097786157e-05, 'epoch': 0.29}) +[2026-10-03 13:48:21,910][starVLA.training.train_starvla][INFO] - Step 3897, Loss: {'train/loss': 4.9319500249112025e-05, 'train/grad_norm_pre_clip': 0.011404300108551979, 'timing/data': 0.00035767900408245623, 'timing/model': 0.39117508300114423, 'learning_rate/qwen_vl_interface': 1.6390607704823267e-06, 'learning_rate/action_model': 1.6390607704823265e-05, 'epoch': 0.29}) +[2026-10-03 13:48:22,607][starVLA.training.train_starvla][INFO] - Step 3898, Loss: {'train/loss': 0.00024655094603076577, 'train/grad_norm_pre_clip': 0.35018277168273926, 'timing/data': 0.0002261299960082397, 'timing/model': 0.3894029219954973, 'learning_rate/qwen_vl_interface': 1.6370829274992162e-06, 'learning_rate/action_model': 1.6370829274992162e-05, 'epoch': 0.29}) +[2026-10-03 13:48:23,506][starVLA.training.train_starvla][INFO] - Step 3899, Loss: {'train/loss': 6.34092721156776e-05, 'train/grad_norm_pre_clip': 0.05050567910075188, 'timing/data': 0.00033863999124150723, 'timing/model': 0.38893106298928615, 'learning_rate/qwen_vl_interface': 1.6351065696498422e-06, 'learning_rate/action_model': 1.6351065696498422e-05, 'epoch': 0.29}) +[2026-10-03 13:48:24,207][starVLA.training.train_starvla][INFO] - Step 3900, Loss: {'train/loss': 6.265875526878517e-06, 'train/grad_norm_pre_clip': 0.0016963635571300983, 'timing/data': 0.00021100000594742596, 'timing/model': 0.3919172030000482, 'learning_rate/qwen_vl_interface': 1.6331316977466117e-06, 'learning_rate/action_model': 1.6331316977466115e-05, 'epoch': 0.29}) +[2026-10-03 13:48:24,905][starVLA.training.train_starvla][INFO] - Step 3901, Loss: {'train/loss': 6.276579370023683e-05, 'train/grad_norm_pre_clip': 0.038031093776226044, 'timing/data': 0.00021358999947551638, 'timing/model': 0.390091332999873, 'learning_rate/qwen_vl_interface': 1.6311583126013198e-06, 'learning_rate/action_model': 1.6311583126013198e-05, 'epoch': 0.29}) +[2026-10-03 13:48:25,601][starVLA.training.train_starvla][INFO] - Step 3902, Loss: {'train/loss': 1.4967390598030761e-05, 'train/grad_norm_pre_clip': 0.00694837374612689, 'timing/data': 0.00020327999664004892, 'timing/model': 0.38978945299459156, 'learning_rate/qwen_vl_interface': 1.6291864150251488e-06, 'learning_rate/action_model': 1.629186415025149e-05, 'epoch': 0.29}) +[2026-10-03 13:48:26,273][starVLA.training.train_starvla][INFO] - Step 3903, Loss: {'train/loss': 0.00038194318767637014, 'train/grad_norm_pre_clip': 0.3720620274543762, 'timing/data': 0.0003523600025800988, 'timing/model': 0.36434776699752547, 'learning_rate/qwen_vl_interface': 1.6272160058286727e-06, 'learning_rate/action_model': 1.6272160058286726e-05, 'epoch': 0.29}) +[2026-10-03 13:48:27,185][starVLA.training.train_starvla][INFO] - Step 3904, Loss: {'train/loss': 3.412743171793409e-05, 'train/grad_norm_pre_clip': 0.010194041766226292, 'timing/data': 0.0003110800025751814, 'timing/model': 0.6048632860038197, 'learning_rate/qwen_vl_interface': 1.6252470858218519e-06, 'learning_rate/action_model': 1.6252470858218517e-05, 'epoch': 0.29}) +[2026-10-03 13:48:27,885][starVLA.training.train_starvla][INFO] - Step 3905, Loss: {'train/loss': 9.365257938043214e-06, 'train/grad_norm_pre_clip': 0.0028492638375610113, 'timing/data': 0.00032029999420046806, 'timing/model': 0.39194252199376933, 'learning_rate/qwen_vl_interface': 1.6232796558140357e-06, 'learning_rate/action_model': 1.6232796558140357e-05, 'epoch': 0.29}) +[2026-10-03 13:48:28,584][starVLA.training.train_starvla][INFO] - Step 3906, Loss: {'train/loss': 1.5637973774573766e-05, 'train/grad_norm_pre_clip': 0.005894086789339781, 'timing/data': 0.000321209998219274, 'timing/model': 0.39126023299468216, 'learning_rate/qwen_vl_interface': 1.6213137166139592e-06, 'learning_rate/action_model': 1.621313716613959e-05, 'epoch': 0.29}) +[2026-10-03 13:48:29,285][starVLA.training.train_starvla][INFO] - Step 3907, Loss: {'train/loss': 0.00022767930931877345, 'train/grad_norm_pre_clip': 0.21768222749233246, 'timing/data': 0.0003275990020483732, 'timing/model': 0.39285716200538445, 'learning_rate/qwen_vl_interface': 1.6193492690297463e-06, 'learning_rate/action_model': 1.6193492690297464e-05, 'epoch': 0.29}) +[2026-10-03 13:48:29,988][starVLA.training.train_starvla][INFO] - Step 3908, Loss: {'train/loss': 0.3066832423210144, 'train/grad_norm_pre_clip': 57.92032241821289, 'timing/data': 0.0002118499978678301, 'timing/model': 0.39331550200586207, 'learning_rate/qwen_vl_interface': 1.6173863138689078e-06, 'learning_rate/action_model': 1.617386313868908e-05, 'epoch': 0.29}) +[2026-10-03 13:48:30,889][starVLA.training.train_starvla][INFO] - Step 3909, Loss: {'train/loss': 0.24676646292209625, 'train/grad_norm_pre_clip': 46.46767044067383, 'timing/data': 0.0003300499956822023, 'timing/model': 0.5923495680035558, 'learning_rate/qwen_vl_interface': 1.6154248519383396e-06, 'learning_rate/action_model': 1.6154248519383395e-05, 'epoch': 0.29}) +[2026-10-03 13:48:31,588][starVLA.training.train_starvla][INFO] - Step 3910, Loss: {'train/loss': 0.00019984522077720612, 'train/grad_norm_pre_clip': 0.42924436926841736, 'timing/data': 0.0003009100037161261, 'timing/model': 0.3909200929920189, 'learning_rate/qwen_vl_interface': 1.6134648840443248e-06, 'learning_rate/action_model': 1.613464884044325e-05, 'epoch': 0.29}) +[2026-10-03 13:48:32,291][starVLA.training.train_starvla][INFO] - Step 3911, Loss: {'train/loss': 1.7991311324294657e-05, 'train/grad_norm_pre_clip': 0.02136598899960518, 'timing/data': 0.00037545000668615103, 'timing/model': 0.3932911529991543, 'learning_rate/qwen_vl_interface': 1.6115064109925334e-06, 'learning_rate/action_model': 1.6115064109925336e-05, 'epoch': 0.29}) +[2026-10-03 13:48:32,994][starVLA.training.train_starvla][INFO] - Step 3912, Loss: {'train/loss': 1.5429663108079694e-05, 'train/grad_norm_pre_clip': 0.0037949220277369022, 'timing/data': 0.00031677000515628606, 'timing/model': 0.3928254820057191, 'learning_rate/qwen_vl_interface': 1.6095494335880173e-06, 'learning_rate/action_model': 1.6095494335880174e-05, 'epoch': 0.29}) +[2026-10-03 13:48:33,693][starVLA.training.train_starvla][INFO] - Step 3913, Loss: {'train/loss': 5.1664039347087964e-05, 'train/grad_norm_pre_clip': 0.017056407406926155, 'timing/data': 0.00028606000705622137, 'timing/model': 0.3909315820055781, 'learning_rate/qwen_vl_interface': 1.6075939526352175e-06, 'learning_rate/action_model': 1.6075939526352175e-05, 'epoch': 0.29}) +[2026-10-03 13:48:34,599][starVLA.training.train_starvla][INFO] - Step 3914, Loss: {'train/loss': 0.24836505949497223, 'train/grad_norm_pre_clip': 170.5284881591797, 'timing/data': 0.00023910999880172312, 'timing/model': 0.5972429969988298, 'learning_rate/qwen_vl_interface': 1.6056399689379576e-06, 'learning_rate/action_model': 1.6056399689379575e-05, 'epoch': 0.29}) +[2026-10-03 13:48:35,299][starVLA.training.train_starvla][INFO] - Step 3915, Loss: {'train/loss': 0.0005599851137958467, 'train/grad_norm_pre_clip': 0.6077242493629456, 'timing/data': 0.0003598500043153763, 'timing/model': 0.3905247530055931, 'learning_rate/qwen_vl_interface': 1.6036874832994473e-06, 'learning_rate/action_model': 1.603687483299447e-05, 'epoch': 0.29}) +[2026-10-03 13:48:35,999][starVLA.training.train_starvla][INFO] - Step 3916, Loss: {'train/loss': 2.252844933536835e-05, 'train/grad_norm_pre_clip': 0.006668131798505783, 'timing/data': 0.0003109599929302931, 'timing/model': 0.3919935029989574, 'learning_rate/qwen_vl_interface': 1.6017364965222796e-06, 'learning_rate/action_model': 1.6017364965222793e-05, 'epoch': 0.29}) +[2026-10-03 13:48:36,699][starVLA.training.train_starvla][INFO] - Step 3917, Loss: {'train/loss': 0.1909789890050888, 'train/grad_norm_pre_clip': 288.7239990234375, 'timing/data': 0.0003596199967432767, 'timing/model': 0.3907444220094476, 'learning_rate/qwen_vl_interface': 1.5997870094084305e-06, 'learning_rate/action_model': 1.5997870094084306e-05, 'epoch': 0.29}) +[2026-10-03 13:48:37,402][starVLA.training.train_starvla][INFO] - Step 3918, Loss: {'train/loss': 1.1220350643270649e-05, 'train/grad_norm_pre_clip': 0.0031655337661504745, 'timing/data': 0.000212009996175766, 'timing/model': 0.39325423300033435, 'learning_rate/qwen_vl_interface': 1.5978390227592616e-06, 'learning_rate/action_model': 1.5978390227592615e-05, 'epoch': 0.29}) +[2026-10-03 13:48:38,101][starVLA.training.train_starvla][INFO] - Step 3919, Loss: {'train/loss': 0.5710746049880981, 'train/grad_norm_pre_clip': 42.000648498535156, 'timing/data': 0.0002147799968952313, 'timing/model': 0.3903590029949555, 'learning_rate/qwen_vl_interface': 1.5958925373755164e-06, 'learning_rate/action_model': 1.5958925373755164e-05, 'epoch': 0.29}) +[2026-10-03 13:48:38,999][starVLA.training.train_starvla][INFO] - Step 3920, Loss: {'train/loss': 0.00016140032676048577, 'train/grad_norm_pre_clip': 0.12593714892864227, 'timing/data': 0.0004150099994149059, 'timing/model': 0.3910269430052722, 'learning_rate/qwen_vl_interface': 1.5939475540573213e-06, 'learning_rate/action_model': 1.5939475540573212e-05, 'epoch': 0.29}) +[2026-10-03 13:48:39,696][starVLA.training.train_starvla][INFO] - Step 3921, Loss: {'train/loss': 0.00013345708430279046, 'train/grad_norm_pre_clip': 0.0679893046617508, 'timing/data': 0.00019643000268843025, 'timing/model': 0.38946418299747165, 'learning_rate/qwen_vl_interface': 1.5920040736041867e-06, 'learning_rate/action_model': 1.5920040736041866e-05, 'epoch': 0.29}) +[2026-10-03 13:48:40,392][starVLA.training.train_starvla][INFO] - Step 3922, Loss: {'train/loss': 0.5448005199432373, 'train/grad_norm_pre_clip': 90.95320892333984, 'timing/data': 0.00031668900919612497, 'timing/model': 0.389436644007219, 'learning_rate/qwen_vl_interface': 1.5900620968150015e-06, 'learning_rate/action_model': 1.5900620968150015e-05, 'epoch': 0.29}) +[2026-10-03 13:48:41,087][starVLA.training.train_starvla][INFO] - Step 3923, Loss: {'train/loss': 4.431187335285358e-05, 'train/grad_norm_pre_clip': 0.01800389215350151, 'timing/data': 0.00031953999132383615, 'timing/model': 0.38894106399675366, 'learning_rate/qwen_vl_interface': 1.5881216244880407e-06, 'learning_rate/action_model': 1.5881216244880407e-05, 'epoch': 0.29}) +[2026-10-03 13:48:41,787][starVLA.training.train_starvla][INFO] - Step 3924, Loss: {'train/loss': 0.2453513890504837, 'train/grad_norm_pre_clip': 165.56301879882812, 'timing/data': 0.0003243999963160604, 'timing/model': 0.3924765830015531, 'learning_rate/qwen_vl_interface': 1.5861826574209587e-06, 'learning_rate/action_model': 1.5861826574209587e-05, 'epoch': 0.29}) +[2026-10-03 13:48:42,688][starVLA.training.train_starvla][INFO] - Step 3925, Loss: {'train/loss': 9.78253228822723e-06, 'train/grad_norm_pre_clip': 0.0024158665910363197, 'timing/data': 0.00022523000370711088, 'timing/model': 0.38845468300860375, 'learning_rate/qwen_vl_interface': 1.5842451964107916e-06, 'learning_rate/action_model': 1.5842451964107916e-05, 'epoch': 0.29}) +[2026-10-03 13:48:43,385][starVLA.training.train_starvla][INFO] - Step 3926, Loss: {'train/loss': 0.0002398226788500324, 'train/grad_norm_pre_clip': 0.7258056998252869, 'timing/data': 0.0003272699977969751, 'timing/model': 0.3905155230022501, 'learning_rate/qwen_vl_interface': 1.5823092422539567e-06, 'learning_rate/action_model': 1.582309242253957e-05, 'epoch': 0.29}) +[2026-10-03 13:48:44,083][starVLA.training.train_starvla][INFO] - Step 3927, Loss: {'train/loss': 4.57031128462404e-05, 'train/grad_norm_pre_clip': 0.03229780122637749, 'timing/data': 0.00020614999812096357, 'timing/model': 0.39074680299381725, 'learning_rate/qwen_vl_interface': 1.5803747957462514e-06, 'learning_rate/action_model': 1.5803747957462514e-05, 'epoch': 0.29}) +[2026-10-03 13:48:44,782][starVLA.training.train_starvla][INFO] - Step 3928, Loss: {'train/loss': 0.0008438251097686589, 'train/grad_norm_pre_clip': 2.8912243843078613, 'timing/data': 0.0002993300004163757, 'timing/model': 0.3915979830053402, 'learning_rate/qwen_vl_interface': 1.5784418576828541e-06, 'learning_rate/action_model': 1.5784418576828543e-05, 'epoch': 0.29}) +[2026-10-03 13:48:45,478][starVLA.training.train_starvla][INFO] - Step 3929, Loss: {'train/loss': 5.065826917416416e-05, 'train/grad_norm_pre_clip': 0.02351832389831543, 'timing/data': 0.0002095200034091249, 'timing/model': 0.38895086299453396, 'learning_rate/qwen_vl_interface': 1.576510428858322e-06, 'learning_rate/action_model': 1.5765104288583217e-05, 'epoch': 0.29}) +[2026-10-03 13:48:46,375][starVLA.training.train_starvla][INFO] - Step 3930, Loss: {'train/loss': 9.491906894254498e-06, 'train/grad_norm_pre_clip': 0.003242525737732649, 'timing/data': 0.00032391000422649086, 'timing/model': 0.5891441989952, 'learning_rate/qwen_vl_interface': 1.5745805100665922e-06, 'learning_rate/action_model': 1.5745805100665923e-05, 'epoch': 0.29}) +[2026-10-03 13:48:47,072][starVLA.training.train_starvla][INFO] - Step 3931, Loss: {'train/loss': 5.50592812942341e-06, 'train/grad_norm_pre_clip': 0.0009884891333058476, 'timing/data': 0.00024172999837901443, 'timing/model': 0.390785573006724, 'learning_rate/qwen_vl_interface': 1.572652102100983e-06, 'learning_rate/action_model': 1.572652102100983e-05, 'epoch': 0.29}) +[2026-10-03 13:48:47,770][starVLA.training.train_starvla][INFO] - Step 3932, Loss: {'train/loss': 0.21897931396961212, 'train/grad_norm_pre_clip': 293.41650390625, 'timing/data': 0.00020621001021936536, 'timing/model': 0.3912329930026317, 'learning_rate/qwen_vl_interface': 1.5707252057541898e-06, 'learning_rate/action_model': 1.5707252057541896e-05, 'epoch': 0.29}) +[2026-10-03 13:48:48,467][starVLA.training.train_starvla][INFO] - Step 3933, Loss: {'train/loss': 3.176520112901926e-05, 'train/grad_norm_pre_clip': 0.00862198043614626, 'timing/data': 0.00020180999126750976, 'timing/model': 0.39027848400291987, 'learning_rate/qwen_vl_interface': 1.5687998218182849e-06, 'learning_rate/action_model': 1.5687998218182847e-05, 'epoch': 0.29}) +[2026-10-03 13:48:49,165][starVLA.training.train_starvla][INFO] - Step 3934, Loss: {'train/loss': 1.989912925637327e-05, 'train/grad_norm_pre_clip': 0.007956567220389843, 'timing/data': 0.00023256998974829912, 'timing/model': 0.39007083399337716, 'learning_rate/qwen_vl_interface': 1.5668759510847235e-06, 'learning_rate/action_model': 1.5668759510847233e-05, 'epoch': 0.29}) +[2026-10-03 13:48:49,861][starVLA.training.train_starvla][INFO] - Step 3935, Loss: {'train/loss': 0.5921355485916138, 'train/grad_norm_pre_clip': 72.58985137939453, 'timing/data': 0.00020540000696200877, 'timing/model': 0.3890791039884789, 'learning_rate/qwen_vl_interface': 1.5649535943443331e-06, 'learning_rate/action_model': 1.564953594344333e-05, 'epoch': 0.29}) +[2026-10-03 13:48:50,762][starVLA.training.train_starvla][INFO] - Step 3936, Loss: {'train/loss': 5.120885180076584e-05, 'train/grad_norm_pre_clip': 0.009129935875535011, 'timing/data': 0.0002092299982905388, 'timing/model': 0.39097072299045976, 'learning_rate/qwen_vl_interface': 1.5630327523873244e-06, 'learning_rate/action_model': 1.563032752387324e-05, 'epoch': 0.29}) +[2026-10-03 13:48:51,461][starVLA.training.train_starvla][INFO] - Step 3937, Loss: {'train/loss': 0.00019948981935158372, 'train/grad_norm_pre_clip': 0.1945340782403946, 'timing/data': 0.00034174000029452145, 'timing/model': 0.39050036399567034, 'learning_rate/qwen_vl_interface': 1.561113426003282e-06, 'learning_rate/action_model': 1.561113426003282e-05, 'epoch': 0.29}) +[2026-10-03 13:48:52,158][starVLA.training.train_starvla][INFO] - Step 3938, Loss: {'train/loss': 0.0013982300879433751, 'train/grad_norm_pre_clip': 7.719736099243164, 'timing/data': 0.00034360999416094273, 'timing/model': 0.3905472530022962, 'learning_rate/qwen_vl_interface': 1.5591956159811686e-06, 'learning_rate/action_model': 1.5591956159811687e-05, 'epoch': 0.29}) +[2026-10-03 13:48:52,855][starVLA.training.train_starvla][INFO] - Step 3939, Loss: {'train/loss': 6.733942427672446e-05, 'train/grad_norm_pre_clip': 0.08523151278495789, 'timing/data': 0.0002579489955678582, 'timing/model': 0.38984039399656467, 'learning_rate/qwen_vl_interface': 1.5572793231093235e-06, 'learning_rate/action_model': 1.5572793231093236e-05, 'epoch': 0.29}) +[2026-10-03 13:48:53,553][starVLA.training.train_starvla][INFO] - Step 3940, Loss: {'train/loss': 1.7500609828857705e-05, 'train/grad_norm_pre_clip': 0.0046133240684866905, 'timing/data': 0.00031211999885272235, 'timing/model': 0.39135401300154626, 'learning_rate/qwen_vl_interface': 1.5553645481754626e-06, 'learning_rate/action_model': 1.5553645481754626e-05, 'epoch': 0.29}) +[2026-10-03 13:48:54,452][starVLA.training.train_starvla][INFO] - Step 3941, Loss: {'train/loss': 0.00027546737692318857, 'train/grad_norm_pre_clip': 0.5393579602241516, 'timing/data': 0.0002303789951838553, 'timing/model': 0.38938390399562195, 'learning_rate/qwen_vl_interface': 1.5534512919666775e-06, 'learning_rate/action_model': 1.5534512919666776e-05, 'epoch': 0.29}) +[2026-10-03 13:48:55,150][starVLA.training.train_starvla][INFO] - Step 3942, Loss: {'train/loss': 8.612807505414821e-06, 'train/grad_norm_pre_clip': 0.0018667583353817463, 'timing/data': 0.00020666999625973403, 'timing/model': 0.3907206430012593, 'learning_rate/qwen_vl_interface': 1.551539555269435e-06, 'learning_rate/action_model': 1.5515395552694348e-05, 'epoch': 0.29}) +[2026-10-03 13:48:55,846][starVLA.training.train_starvla][INFO] - Step 3943, Loss: {'train/loss': 2.7780635718954727e-05, 'train/grad_norm_pre_clip': 0.005883887875825167, 'timing/data': 0.00035242000012658536, 'timing/model': 0.38892854400910437, 'learning_rate/qwen_vl_interface': 1.5496293388695787e-06, 'learning_rate/action_model': 1.5496293388695786e-05, 'epoch': 0.29}) +[2026-10-03 13:48:56,545][starVLA.training.train_starvla][INFO] - Step 3944, Loss: {'train/loss': 0.0021867158357053995, 'train/grad_norm_pre_clip': 2.3627960681915283, 'timing/data': 0.0002464200078975409, 'timing/model': 0.3916573430033168, 'learning_rate/qwen_vl_interface': 1.547720643552327e-06, 'learning_rate/action_model': 1.547720643552327e-05, 'epoch': 0.29}) +[2026-10-03 13:48:57,244][starVLA.training.train_starvla][INFO] - Step 3945, Loss: {'train/loss': 0.53759765625, 'train/grad_norm_pre_clip': 38.91950988769531, 'timing/data': 0.00020364999363664538, 'timing/model': 0.39059316299972124, 'learning_rate/qwen_vl_interface': 1.5458134701022706e-06, 'learning_rate/action_model': 1.5458134701022705e-05, 'epoch': 0.29}) +[2026-10-03 13:48:58,148][starVLA.training.train_starvla][INFO] - Step 3946, Loss: {'train/loss': 0.00041434846934862435, 'train/grad_norm_pre_clip': 0.9524575471878052, 'timing/data': 0.00039834999188315123, 'timing/model': 0.5971326480066637, 'learning_rate/qwen_vl_interface': 1.5439078193033765e-06, 'learning_rate/action_model': 1.5439078193033766e-05, 'epoch': 0.29}) +[2026-10-03 13:48:58,846][starVLA.training.train_starvla][INFO] - Step 3947, Loss: {'train/loss': 1.637537934584543e-05, 'train/grad_norm_pre_clip': 0.008401818573474884, 'timing/data': 0.0003594900044845417, 'timing/model': 0.390096914008609, 'learning_rate/qwen_vl_interface': 1.5420036919389872e-06, 'learning_rate/action_model': 1.5420036919389873e-05, 'epoch': 0.29}) +[2026-10-03 13:48:59,545][starVLA.training.train_starvla][INFO] - Step 3948, Loss: {'train/loss': 0.00013026868691667914, 'train/grad_norm_pre_clip': 0.0791536346077919, 'timing/data': 0.00031424999178852886, 'timing/model': 0.39206317299976945, 'learning_rate/qwen_vl_interface': 1.540101088791817e-06, 'learning_rate/action_model': 1.5401010887918168e-05, 'epoch': 0.29}) +[2026-10-03 13:49:00,242][starVLA.training.train_starvla][INFO] - Step 3949, Loss: {'train/loss': 0.5084079504013062, 'train/grad_norm_pre_clip': 86.22981262207031, 'timing/data': 0.00032716999703552574, 'timing/model': 0.3899819829966873, 'learning_rate/qwen_vl_interface': 1.5382000106439532e-06, 'learning_rate/action_model': 1.5382000106439532e-05, 'epoch': 0.29}) +[2026-10-03 13:49:00,940][starVLA.training.train_starvla][INFO] - Step 3950, Loss: {'train/loss': 3.183362787240185e-05, 'train/grad_norm_pre_clip': 0.008805084973573685, 'timing/data': 0.0003481399908196181, 'timing/model': 0.3900534739950672, 'learning_rate/qwen_vl_interface': 1.5363004582768584e-06, 'learning_rate/action_model': 1.5363004582768584e-05, 'epoch': 0.29}) +[2026-10-03 13:49:01,835][starVLA.training.train_starvla][INFO] - Step 3951, Loss: {'train/loss': 5.127501572133042e-05, 'train/grad_norm_pre_clip': 0.009887060150504112, 'timing/data': 0.0003351699997438118, 'timing/model': 0.587554990008357, 'learning_rate/qwen_vl_interface': 1.534402432471367e-06, 'learning_rate/action_model': 1.534402432471367e-05, 'epoch': 0.29}) +[2026-10-03 13:49:02,537][starVLA.training.train_starvla][INFO] - Step 3952, Loss: {'train/loss': 1.3828062037646305e-05, 'train/grad_norm_pre_clip': 0.0034054520074278116, 'timing/data': 0.0002141900040442124, 'timing/model': 0.39371488300093915, 'learning_rate/qwen_vl_interface': 1.532505934007685e-06, 'learning_rate/action_model': 1.532505934007685e-05, 'epoch': 0.29}) +[2026-10-03 13:49:03,236][starVLA.training.train_starvla][INFO] - Step 3953, Loss: {'train/loss': 1.9958508346462622e-05, 'train/grad_norm_pre_clip': 0.005396158900111914, 'timing/data': 0.0004320500011090189, 'timing/model': 0.3911677339929156, 'learning_rate/qwen_vl_interface': 1.5306109636653915e-06, 'learning_rate/action_model': 1.5306109636653914e-05, 'epoch': 0.29}) +[2026-10-03 13:49:03,933][starVLA.training.train_starvla][INFO] - Step 3954, Loss: {'train/loss': 0.4483802616596222, 'train/grad_norm_pre_clip': 37.42830276489258, 'timing/data': 0.0001996000064536929, 'timing/model': 0.3895742640015669, 'learning_rate/qwen_vl_interface': 1.5287175222234383e-06, 'learning_rate/action_model': 1.528717522223438e-05, 'epoch': 0.29}) +[2026-10-03 13:49:04,631][starVLA.training.train_starvla][INFO] - Step 3955, Loss: {'train/loss': 8.085811714408919e-05, 'train/grad_norm_pre_clip': 0.07653770595788956, 'timing/data': 0.0002010999887716025, 'timing/model': 0.39049769401026424, 'learning_rate/qwen_vl_interface': 1.5268256104601457e-06, 'learning_rate/action_model': 1.5268256104601454e-05, 'epoch': 0.29}) +[2026-10-03 13:49:05,532][starVLA.training.train_starvla][INFO] - Step 3956, Loss: {'train/loss': 6.586272320419084e-06, 'train/grad_norm_pre_clip': 0.0013103247620165348, 'timing/data': 0.0002784300013445318, 'timing/model': 0.5921864799893228, 'learning_rate/qwen_vl_interface': 1.5249352291532075e-06, 'learning_rate/action_model': 1.5249352291532073e-05, 'epoch': 0.29}) +[2026-10-03 13:49:06,228][starVLA.training.train_starvla][INFO] - Step 3957, Loss: {'train/loss': 0.00036847853334620595, 'train/grad_norm_pre_clip': 0.4952841103076935, 'timing/data': 0.0003056000132346526, 'timing/model': 0.3897320340038277, 'learning_rate/qwen_vl_interface': 1.523046379079689e-06, 'learning_rate/action_model': 1.523046379079689e-05, 'epoch': 0.29}) +[2026-10-03 13:49:06,926][starVLA.training.train_starvla][INFO] - Step 3958, Loss: {'train/loss': 1.8156169971916825e-05, 'train/grad_norm_pre_clip': 0.011725147254765034, 'timing/data': 0.00033789000008255243, 'timing/model': 0.3904694539960474, 'learning_rate/qwen_vl_interface': 1.5211590610160244e-06, 'learning_rate/action_model': 1.5211590610160242e-05, 'epoch': 0.29}) +[2026-10-03 13:49:07,627][starVLA.training.train_starvla][INFO] - Step 3959, Loss: {'train/loss': 0.0006277650245465338, 'train/grad_norm_pre_clip': 2.611189603805542, 'timing/data': 0.0002118700067512691, 'timing/model': 0.3911948239983758, 'learning_rate/qwen_vl_interface': 1.5192732757380193e-06, 'learning_rate/action_model': 1.5192732757380194e-05, 'epoch': 0.29}) +[2026-10-03 13:49:08,324][starVLA.training.train_starvla][INFO] - Step 3960, Loss: {'train/loss': 9.394781045557465e-06, 'train/grad_norm_pre_clip': 0.002710673725232482, 'timing/data': 0.0002343600062886253, 'timing/model': 0.3909243430098286, 'learning_rate/qwen_vl_interface': 1.517389024020849e-06, 'learning_rate/action_model': 1.5173890240208488e-05, 'epoch': 0.29}) +[2026-10-03 13:49:09,025][starVLA.training.train_starvla][INFO] - Step 3961, Loss: {'train/loss': 3.128136449959129e-05, 'train/grad_norm_pre_clip': 0.013415946625173092, 'timing/data': 0.00031956000020727515, 'timing/model': 0.39173948399547953, 'learning_rate/qwen_vl_interface': 1.515506306639057e-06, 'learning_rate/action_model': 1.515506306639057e-05, 'epoch': 0.29}) +[2026-10-03 13:49:09,941][starVLA.training.train_starvla][INFO] - Step 3962, Loss: {'train/loss': 0.0005237691220827401, 'train/grad_norm_pre_clip': 0.5426999926567078, 'timing/data': 0.00021161000768188387, 'timing/model': 0.3892918249912327, 'learning_rate/qwen_vl_interface': 1.5136251243665589e-06, 'learning_rate/action_model': 1.5136251243665589e-05, 'epoch': 0.29}) +[2026-10-03 13:49:10,640][starVLA.training.train_starvla][INFO] - Step 3963, Loss: {'train/loss': 3.300598564237589e-06, 'train/grad_norm_pre_clip': 0.0008844592375680804, 'timing/data': 0.00021896000544074923, 'timing/model': 0.39109471399569884, 'learning_rate/qwen_vl_interface': 1.5117454779766372e-06, 'learning_rate/action_model': 1.5117454779766371e-05, 'epoch': 0.29}) +[2026-10-03 13:49:11,341][starVLA.training.train_starvla][INFO] - Step 3964, Loss: {'train/loss': 1.1369425010343548e-05, 'train/grad_norm_pre_clip': 0.0030223450157791376, 'timing/data': 0.00029884000832680613, 'timing/model': 0.3925576629990246, 'learning_rate/qwen_vl_interface': 1.5098673682419428e-06, 'learning_rate/action_model': 1.5098673682419429e-05, 'epoch': 0.29}) +[2026-10-03 13:49:12,036][starVLA.training.train_starvla][INFO] - Step 3965, Loss: {'train/loss': 0.32496023178100586, 'train/grad_norm_pre_clip': 306.178955078125, 'timing/data': 0.00019716999668162316, 'timing/model': 0.3889169539907016, 'learning_rate/qwen_vl_interface': 1.5079907959344977e-06, 'learning_rate/action_model': 1.5079907959344978e-05, 'epoch': 0.29}) +[2026-10-03 13:49:12,733][starVLA.training.train_starvla][INFO] - Step 3966, Loss: {'train/loss': 9.581317499396391e-06, 'train/grad_norm_pre_clip': 0.002914697863161564, 'timing/data': 0.000311740004690364, 'timing/model': 0.39062248400296085, 'learning_rate/qwen_vl_interface': 1.5061157618256873e-06, 'learning_rate/action_model': 1.5061157618256871e-05, 'epoch': 0.29}) +[2026-10-03 13:49:13,633][starVLA.training.train_starvla][INFO] - Step 3967, Loss: {'train/loss': 2.8778904379578307e-05, 'train/grad_norm_pre_clip': 0.009685847908258438, 'timing/data': 0.00027074001263827085, 'timing/model': 0.5908595109940507, 'learning_rate/qwen_vl_interface': 1.5042422666862675e-06, 'learning_rate/action_model': 1.5042422666862673e-05, 'epoch': 0.29}) +[2026-10-03 13:49:14,332][starVLA.training.train_starvla][INFO] - Step 3968, Loss: {'train/loss': 6.18816920905374e-05, 'train/grad_norm_pre_clip': 0.01856165938079357, 'timing/data': 0.0002629999944474548, 'timing/model': 0.3911526839947328, 'learning_rate/qwen_vl_interface': 1.5023703112863628e-06, 'learning_rate/action_model': 1.5023703112863626e-05, 'epoch': 0.29}) +[2026-10-03 13:49:15,028][starVLA.training.train_starvla][INFO] - Step 3969, Loss: {'train/loss': 1.9012792108696885e-05, 'train/grad_norm_pre_clip': 0.006924852728843689, 'timing/data': 0.00020008999854326248, 'timing/model': 0.3892366740037687, 'learning_rate/qwen_vl_interface': 1.5004998963954629e-06, 'learning_rate/action_model': 1.5004998963954628e-05, 'epoch': 0.29}) +[2026-10-03 13:49:15,730][starVLA.training.train_starvla][INFO] - Step 3970, Loss: {'train/loss': 0.00019372382666915655, 'train/grad_norm_pre_clip': 0.052935998886823654, 'timing/data': 0.0003632300067692995, 'timing/model': 0.39278785399801563, 'learning_rate/qwen_vl_interface': 1.498631022782424e-06, 'learning_rate/action_model': 1.4986310227824241e-05, 'epoch': 0.29}) +[2026-10-03 13:49:16,430][starVLA.training.train_starvla][INFO] - Step 3971, Loss: {'train/loss': 0.425624817609787, 'train/grad_norm_pre_clip': 64.6524887084961, 'timing/data': 0.00026155999512411654, 'timing/model': 0.3908191039954545, 'learning_rate/qwen_vl_interface': 1.4967636912154707e-06, 'learning_rate/action_model': 1.4967636912154706e-05, 'epoch': 0.29}) +[2026-10-03 13:49:17,324][starVLA.training.train_starvla][INFO] - Step 3972, Loss: {'train/loss': 0.0018684349488466978, 'train/grad_norm_pre_clip': 10.329658508300781, 'timing/data': 0.00020299000607337803, 'timing/model': 0.5865331120003248, 'learning_rate/qwen_vl_interface': 1.4948979024621915e-06, 'learning_rate/action_model': 1.4948979024621914e-05, 'epoch': 0.29}) +[2026-10-03 13:49:18,022][starVLA.training.train_starvla][INFO] - Step 3973, Loss: {'train/loss': 0.23405730724334717, 'train/grad_norm_pre_clip': 67.5390625, 'timing/data': 0.00020554999355226755, 'timing/model': 0.3902492049965076, 'learning_rate/qwen_vl_interface': 1.493033657289542e-06, 'learning_rate/action_model': 1.493033657289542e-05, 'epoch': 0.29}) +[2026-10-03 13:49:18,721][starVLA.training.train_starvla][INFO] - Step 3974, Loss: {'train/loss': 0.0053774891421198845, 'train/grad_norm_pre_clip': 5.380326271057129, 'timing/data': 0.00034209000295959413, 'timing/model': 0.3900566039956175, 'learning_rate/qwen_vl_interface': 1.4911709564638428e-06, 'learning_rate/action_model': 1.4911709564638427e-05, 'epoch': 0.29}) +[2026-10-03 13:49:19,419][starVLA.training.train_starvla][INFO] - Step 3975, Loss: {'train/loss': 0.0023195650428533554, 'train/grad_norm_pre_clip': 12.494050979614258, 'timing/data': 0.00020076999498996884, 'timing/model': 0.38962988401181065, 'learning_rate/qwen_vl_interface': 1.4893098007507817e-06, 'learning_rate/action_model': 1.4893098007507816e-05, 'epoch': 0.29}) +[2026-10-03 13:49:20,116][starVLA.training.train_starvla][INFO] - Step 3976, Loss: {'train/loss': 0.0019911392591893673, 'train/grad_norm_pre_clip': 3.517482280731201, 'timing/data': 0.0001954700128408149, 'timing/model': 0.39104462499381043, 'learning_rate/qwen_vl_interface': 1.487450190915406e-06, 'learning_rate/action_model': 1.487450190915406e-05, 'epoch': 0.29}) +[2026-10-03 13:49:20,814][starVLA.training.train_starvla][INFO] - Step 3977, Loss: {'train/loss': 0.0006035287515260279, 'train/grad_norm_pre_clip': 0.5927979946136475, 'timing/data': 0.0003115199942840263, 'timing/model': 0.3906777539959876, 'learning_rate/qwen_vl_interface': 1.4855921277221336e-06, 'learning_rate/action_model': 1.4855921277221336e-05, 'epoch': 0.29}) +[2026-10-03 13:49:21,720][starVLA.training.train_starvla][INFO] - Step 3978, Loss: {'train/loss': 0.00010032334103016183, 'train/grad_norm_pre_clip': 0.023326531052589417, 'timing/data': 0.0002092299982905388, 'timing/model': 0.3917948940070346, 'learning_rate/qwen_vl_interface': 1.4837356119347423e-06, 'learning_rate/action_model': 1.4837356119347423e-05, 'epoch': 0.29}) +[2026-10-03 13:49:22,426][starVLA.training.train_starvla][INFO] - Step 3979, Loss: {'train/loss': 0.0011562597937881947, 'train/grad_norm_pre_clip': 1.580595850944519, 'timing/data': 0.0003091099933953956, 'timing/model': 0.3960984329896746, 'learning_rate/qwen_vl_interface': 1.4818806443163774e-06, 'learning_rate/action_model': 1.4818806443163775e-05, 'epoch': 0.29}) +[2026-10-03 13:49:23,126][starVLA.training.train_starvla][INFO] - Step 3980, Loss: {'train/loss': 0.9672060608863831, 'train/grad_norm_pre_clip': 24.03008460998535, 'timing/data': 0.00028198999643791467, 'timing/model': 0.3940616039908491, 'learning_rate/qwen_vl_interface': 1.480027225629547e-06, 'learning_rate/action_model': 1.480027225629547e-05, 'epoch': 0.29}) +[2026-10-03 13:49:23,825][starVLA.training.train_starvla][INFO] - Step 3981, Loss: {'train/loss': 0.025237666442990303, 'train/grad_norm_pre_clip': 51.84603500366211, 'timing/data': 0.00034424000477883965, 'timing/model': 0.39027365400397684, 'learning_rate/qwen_vl_interface': 1.4781753566361182e-06, 'learning_rate/action_model': 1.4781753566361182e-05, 'epoch': 0.29}) +[2026-10-03 13:49:24,526][starVLA.training.train_starvla][INFO] - Step 3982, Loss: {'train/loss': 0.007479852996766567, 'train/grad_norm_pre_clip': 22.84720802307129, 'timing/data': 0.00033755999174900353, 'timing/model': 0.39220330500393175, 'learning_rate/qwen_vl_interface': 1.4763250380973255e-06, 'learning_rate/action_model': 1.4763250380973254e-05, 'epoch': 0.29}) +[2026-10-03 13:49:25,427][starVLA.training.train_starvla][INFO] - Step 3983, Loss: {'train/loss': 0.0001420290645910427, 'train/grad_norm_pre_clip': 0.14733858406543732, 'timing/data': 0.0003274000046076253, 'timing/model': 0.3907669949985575, 'learning_rate/qwen_vl_interface': 1.4744762707737662e-06, 'learning_rate/action_model': 1.4744762707737662e-05, 'epoch': 0.29}) +[2026-10-03 13:49:26,130][starVLA.training.train_starvla][INFO] - Step 3984, Loss: {'train/loss': 1.9586439520935528e-05, 'train/grad_norm_pre_clip': 0.012663427740335464, 'timing/data': 0.00024202000349760056, 'timing/model': 0.39394132400047965, 'learning_rate/qwen_vl_interface': 1.4726290554253977e-06, 'learning_rate/action_model': 1.4726290554253977e-05, 'epoch': 0.29}) +[2026-10-03 13:49:26,833][starVLA.training.train_starvla][INFO] - Step 3985, Loss: {'train/loss': 2.0681758542195894e-05, 'train/grad_norm_pre_clip': 0.006819781381636858, 'timing/data': 0.0002456899965181947, 'timing/model': 0.3928888939990429, 'learning_rate/qwen_vl_interface': 1.4707833928115406e-06, 'learning_rate/action_model': 1.4707833928115405e-05, 'epoch': 0.29}) +[2026-10-03 13:49:27,534][starVLA.training.train_starvla][INFO] - Step 3986, Loss: {'train/loss': 0.000524067843798548, 'train/grad_norm_pre_clip': 0.5046808123588562, 'timing/data': 0.00020919000962749124, 'timing/model': 0.3911744849901879, 'learning_rate/qwen_vl_interface': 1.4689392836908775e-06, 'learning_rate/action_model': 1.4689392836908774e-05, 'epoch': 0.29}) +[2026-10-03 13:49:28,237][starVLA.training.train_starvla][INFO] - Step 3987, Loss: {'train/loss': 3.155553713440895e-05, 'train/grad_norm_pre_clip': 0.01749235764145851, 'timing/data': 0.0003327300073578954, 'timing/model': 0.3939522530126851, 'learning_rate/qwen_vl_interface': 1.4670967288214514e-06, 'learning_rate/action_model': 1.4670967288214513e-05, 'epoch': 0.29}) +[2026-10-03 13:49:29,152][starVLA.training.train_starvla][INFO] - Step 3988, Loss: {'train/loss': 7.259180711116642e-05, 'train/grad_norm_pre_clip': 0.024733375757932663, 'timing/data': 0.00026231999800074846, 'timing/model': 0.6052410380070796, 'learning_rate/qwen_vl_interface': 1.4652557289606687e-06, 'learning_rate/action_model': 1.4652557289606685e-05, 'epoch': 0.29}) +[2026-10-03 13:49:29,852][starVLA.training.train_starvla][INFO] - Step 3989, Loss: {'train/loss': 5.029125077271601e-06, 'train/grad_norm_pre_clip': 0.001014619949273765, 'timing/data': 0.00021395899238996208, 'timing/model': 0.39188335499784444, 'learning_rate/qwen_vl_interface': 1.4634162848652918e-06, 'learning_rate/action_model': 1.4634162848652919e-05, 'epoch': 0.29}) +[2026-10-03 13:49:30,552][starVLA.training.train_starvla][INFO] - Step 3990, Loss: {'train/loss': 1.0132500392501242e-05, 'train/grad_norm_pre_clip': 0.003263295628130436, 'timing/data': 0.00023863000387791544, 'timing/model': 0.3922977339971112, 'learning_rate/qwen_vl_interface': 1.4615783972914494e-06, 'learning_rate/action_model': 1.4615783972914493e-05, 'epoch': 0.29}) +[2026-10-03 13:49:31,252][starVLA.training.train_starvla][INFO] - Step 3991, Loss: {'train/loss': 3.355958324391395e-05, 'train/grad_norm_pre_clip': 0.007698738016188145, 'timing/data': 0.0002530700003262609, 'timing/model': 0.3913417049916461, 'learning_rate/qwen_vl_interface': 1.459742066994626e-06, 'learning_rate/action_model': 1.4597420669946261e-05, 'epoch': 0.29}) +[2026-10-03 13:49:31,954][starVLA.training.train_starvla][INFO] - Step 3992, Loss: {'train/loss': 1.1950417501793709e-05, 'train/grad_norm_pre_clip': 0.0038960506208240986, 'timing/data': 0.0003230400034226477, 'timing/model': 0.39134647500759456, 'learning_rate/qwen_vl_interface': 1.4579072947296687e-06, 'learning_rate/action_model': 1.4579072947296687e-05, 'epoch': 0.29}) +[2026-10-03 13:49:32,864][starVLA.training.train_starvla][INFO] - Step 3993, Loss: {'train/loss': 0.006390931084752083, 'train/grad_norm_pre_clip': 46.143192291259766, 'timing/data': 0.00026790999982040375, 'timing/model': 0.6000845389935421, 'learning_rate/qwen_vl_interface': 1.4560740812507823e-06, 'learning_rate/action_model': 1.4560740812507822e-05, 'epoch': 0.29}) +[2026-10-03 13:49:33,562][starVLA.training.train_starvla][INFO] - Step 3994, Loss: {'train/loss': 1.6822730685817078e-05, 'train/grad_norm_pre_clip': 0.006819016300141811, 'timing/data': 0.00020136999955866486, 'timing/model': 0.39097285400202964, 'learning_rate/qwen_vl_interface': 1.4542424273115322e-06, 'learning_rate/action_model': 1.454242427311532e-05, 'epoch': 0.29}) +[2026-10-03 13:49:34,326][starVLA.training.train_starvla][INFO] - Step 3995, Loss: {'train/loss': 2.78698826150503e-05, 'train/grad_norm_pre_clip': 0.011513607576489449, 'timing/data': 0.00024260999634861946, 'timing/model': 0.3926720940071391, 'learning_rate/qwen_vl_interface': 1.4524123336648419e-06, 'learning_rate/action_model': 1.4524123336648418e-05, 'epoch': 0.29}) +[2026-10-03 13:49:35,030][starVLA.training.train_starvla][INFO] - Step 3996, Loss: {'train/loss': 3.9750222640577704e-05, 'train/grad_norm_pre_clip': 0.008430413901805878, 'timing/data': 0.0002854300109902397, 'timing/model': 0.3939049239997985, 'learning_rate/qwen_vl_interface': 1.4505838010629942e-06, 'learning_rate/action_model': 1.4505838010629941e-05, 'epoch': 0.29}) +[2026-10-03 13:49:35,729][starVLA.training.train_starvla][INFO] - Step 3997, Loss: {'train/loss': 1.681534922681749e-05, 'train/grad_norm_pre_clip': 0.007103653624653816, 'timing/data': 0.00025487999664619565, 'timing/model': 0.39066125500539783, 'learning_rate/qwen_vl_interface': 1.4487568302576284e-06, 'learning_rate/action_model': 1.4487568302576282e-05, 'epoch': 0.29}) +[2026-10-03 13:49:36,430][starVLA.training.train_starvla][INFO] - Step 3998, Loss: {'train/loss': 9.827086614677683e-06, 'train/grad_norm_pre_clip': 0.0027423338033258915, 'timing/data': 0.00030229998810682446, 'timing/model': 0.3917908650037134, 'learning_rate/qwen_vl_interface': 1.446931421999746e-06, 'learning_rate/action_model': 1.4469314219997459e-05, 'epoch': 0.29}) +[2026-10-03 13:49:37,345][starVLA.training.train_starvla][INFO] - Step 3999, Loss: {'train/loss': 1.863258148659952e-05, 'train/grad_norm_pre_clip': 0.008359427563846111, 'timing/data': 0.0002903400018112734, 'timing/model': 0.39252021499851253, 'learning_rate/qwen_vl_interface': 1.4451075770397004e-06, 'learning_rate/action_model': 1.4451075770397004e-05, 'epoch': 0.29}) +[2026-10-03 13:49:38,046][starVLA.training.train_starvla][INFO] - Step 4000, Loss: {'train/loss': 0.0006455049151554704, 'train/grad_norm_pre_clip': 0.5406057834625244, 'timing/data': 0.000315740006044507, 'timing/model': 0.3931186840054579, 'learning_rate/qwen_vl_interface': 1.4432852961272068e-06, 'learning_rate/action_model': 1.4432852961272066e-05, 'epoch': 0.29}) +[2026-10-03 13:49:38,744][starVLA.training.train_starvla][INFO] - Step 4001, Loss: {'train/loss': 3.8177113310666755e-05, 'train/grad_norm_pre_clip': 0.010009943507611752, 'timing/data': 0.0002824399998644367, 'timing/model': 0.38977152499137446, 'learning_rate/qwen_vl_interface': 1.4414645800113353e-06, 'learning_rate/action_model': 1.4414645800113352e-05, 'epoch': 0.29}) +[2026-10-03 13:49:39,446][starVLA.training.train_starvla][INFO] - Step 4002, Loss: {'train/loss': 0.0005930712213739753, 'train/grad_norm_pre_clip': 0.547273576259613, 'timing/data': 0.00028881999605800956, 'timing/model': 0.3919883549970109, 'learning_rate/qwen_vl_interface': 1.4396454294405138e-06, 'learning_rate/action_model': 1.4396454294405136e-05, 'epoch': 0.29}) +[2026-10-03 13:49:40,145][starVLA.training.train_starvla][INFO] - Step 4003, Loss: {'train/loss': 1.549595981487073e-05, 'train/grad_norm_pre_clip': 0.010884324088692665, 'timing/data': 0.0003199700004188344, 'timing/model': 0.391503004997503, 'learning_rate/qwen_vl_interface': 1.4378278451625271e-06, 'learning_rate/action_model': 1.4378278451625271e-05, 'epoch': 0.29}) +[2026-10-03 13:49:41,069][starVLA.training.train_starvla][INFO] - Step 4004, Loss: {'train/loss': 8.1657253758749e-06, 'train/grad_norm_pre_clip': 0.0028086616657674313, 'timing/data': 0.00026328000240027905, 'timing/model': 0.4049670719978167, 'learning_rate/qwen_vl_interface': 1.4360118279245143e-06, 'learning_rate/action_model': 1.4360118279245144e-05, 'epoch': 0.29}) +[2026-10-03 13:49:41,771][starVLA.training.train_starvla][INFO] - Step 4005, Loss: {'train/loss': 8.171671652235091e-05, 'train/grad_norm_pre_clip': 0.05155092105269432, 'timing/data': 0.00021764000121038407, 'timing/model': 0.39371950400527567, 'learning_rate/qwen_vl_interface': 1.434197378472973e-06, 'learning_rate/action_model': 1.4341973784729729e-05, 'epoch': 0.29}) +[2026-10-03 13:49:42,471][starVLA.training.train_starvla][INFO] - Step 4006, Loss: {'train/loss': 0.00014311437553260475, 'train/grad_norm_pre_clip': 0.11248265206813812, 'timing/data': 0.000314689998049289, 'timing/model': 0.3917188449995592, 'learning_rate/qwen_vl_interface': 1.4323844975537535e-06, 'learning_rate/action_model': 1.4323844975537534e-05, 'epoch': 0.29}) +[2026-10-03 13:49:43,169][starVLA.training.train_starvla][INFO] - Step 4007, Loss: {'train/loss': 8.873490514815785e-06, 'train/grad_norm_pre_clip': 0.003013916779309511, 'timing/data': 0.00026077000075019896, 'timing/model': 0.3909838350082282, 'learning_rate/qwen_vl_interface': 1.4305731859120628e-06, 'learning_rate/action_model': 1.4305731859120628e-05, 'epoch': 0.29}) +[2026-10-03 13:49:43,883][starVLA.training.train_starvla][INFO] - Step 4008, Loss: {'train/loss': 0.0008861215901561081, 'train/grad_norm_pre_clip': 0.8960134983062744, 'timing/data': 0.00021100998856127262, 'timing/model': 0.40527451199886855, 'learning_rate/qwen_vl_interface': 1.4287634442924647e-06, 'learning_rate/action_model': 1.4287634442924646e-05, 'epoch': 0.29}) +[2026-10-03 13:49:44,797][starVLA.training.train_starvla][INFO] - Step 4009, Loss: {'train/loss': 1.2203601727378555e-05, 'train/grad_norm_pre_clip': 0.004474650602787733, 'timing/data': 0.00021396999363787472, 'timing/model': 0.606276919002994, 'learning_rate/qwen_vl_interface': 1.4269552734388725e-06, 'learning_rate/action_model': 1.4269552734388724e-05, 'epoch': 0.29}) +[2026-10-03 13:49:45,495][starVLA.training.train_starvla][INFO] - Step 4010, Loss: {'train/loss': 0.00045256121666170657, 'train/grad_norm_pre_clip': 0.1874011754989624, 'timing/data': 0.0002473700005793944, 'timing/model': 0.3901641849952284, 'learning_rate/qwen_vl_interface': 1.4251486740945584e-06, 'learning_rate/action_model': 1.4251486740945583e-05, 'epoch': 0.29}) +[2026-10-03 13:49:46,192][starVLA.training.train_starvla][INFO] - Step 4011, Loss: {'train/loss': 0.5705947279930115, 'train/grad_norm_pre_clip': 90.59666442871094, 'timing/data': 0.00021983000624459237, 'timing/model': 0.39031193499977235, 'learning_rate/qwen_vl_interface': 1.423343647002147e-06, 'learning_rate/action_model': 1.423343647002147e-05, 'epoch': 0.29}) +[2026-10-03 13:49:46,902][starVLA.training.train_starvla][INFO] - Step 4012, Loss: {'train/loss': 3.889176696247887e-06, 'train/grad_norm_pre_clip': 0.0016331755323335528, 'timing/data': 0.00027571000100579113, 'timing/model': 0.4032923729973845, 'learning_rate/qwen_vl_interface': 1.421540192903617e-06, 'learning_rate/action_model': 1.421540192903617e-05, 'epoch': 0.29}) +[2026-10-03 13:49:47,599][starVLA.training.train_starvla][INFO] - Step 4013, Loss: {'train/loss': 2.2558164346264675e-05, 'train/grad_norm_pre_clip': 0.006157769355922937, 'timing/data': 0.0002376100019318983, 'timing/model': 0.390354994990048, 'learning_rate/qwen_vl_interface': 1.4197383125402997e-06, 'learning_rate/action_model': 1.4197383125402997e-05, 'epoch': 0.29}) +[2026-10-03 13:49:48,491][starVLA.training.train_starvla][INFO] - Step 4014, Loss: {'train/loss': 9.454510291106999e-06, 'train/grad_norm_pre_clip': 0.003932681865990162, 'timing/data': 0.0002701799967326224, 'timing/model': 0.5848219019972021, 'learning_rate/qwen_vl_interface': 1.4179380066528807e-06, 'learning_rate/action_model': 1.4179380066528807e-05, 'epoch': 0.29}) +[2026-10-03 13:49:49,187][starVLA.training.train_starvla][INFO] - Step 4015, Loss: {'train/loss': 0.03084113635122776, 'train/grad_norm_pre_clip': 76.00951385498047, 'timing/data': 0.0002499800029909238, 'timing/model': 0.38974493500427343, 'learning_rate/qwen_vl_interface': 1.4161392759813984e-06, 'learning_rate/action_model': 1.4161392759813983e-05, 'epoch': 0.29}) +[2026-10-03 13:49:49,898][starVLA.training.train_starvla][INFO] - Step 4016, Loss: {'train/loss': 0.0010834967251867056, 'train/grad_norm_pre_clip': 5.003212928771973, 'timing/data': 0.00028716000088024884, 'timing/model': 0.4041780830011703, 'learning_rate/qwen_vl_interface': 1.4143421212652411e-06, 'learning_rate/action_model': 1.414342121265241e-05, 'epoch': 0.29}) +[2026-10-03 13:49:50,597][starVLA.training.train_starvla][INFO] - Step 4017, Loss: {'train/loss': 0.04362308606505394, 'train/grad_norm_pre_clip': 101.46243286132812, 'timing/data': 0.00021209999977145344, 'timing/model': 0.3906610950070899, 'learning_rate/qwen_vl_interface': 1.4125465432431529e-06, 'learning_rate/action_model': 1.4125465432431529e-05, 'epoch': 0.29}) +[2026-10-03 13:49:51,297][starVLA.training.train_starvla][INFO] - Step 4018, Loss: {'train/loss': 1.3589038644568063e-05, 'train/grad_norm_pre_clip': 0.005602533929049969, 'timing/data': 0.00020967000455129892, 'timing/model': 0.3916259550023824, 'learning_rate/qwen_vl_interface': 1.4107525426532273e-06, 'learning_rate/action_model': 1.4107525426532272e-05, 'epoch': 0.29}) +[2026-10-03 13:49:51,995][starVLA.training.train_starvla][INFO] - Step 4019, Loss: {'train/loss': 0.0001792841067072004, 'train/grad_norm_pre_clip': 0.28305336833000183, 'timing/data': 0.0002065699954982847, 'timing/model': 0.39043230599781964, 'learning_rate/qwen_vl_interface': 1.4089601202329116e-06, 'learning_rate/action_model': 1.4089601202329114e-05, 'epoch': 0.29}) +[2026-10-03 13:49:52,916][starVLA.training.train_starvla][INFO] - Step 4020, Loss: {'train/loss': 2.1017382096033543e-05, 'train/grad_norm_pre_clip': 0.004146035760641098, 'timing/data': 0.00029009999707341194, 'timing/model': 0.4051069330016617, 'learning_rate/qwen_vl_interface': 1.407169276719e-06, 'learning_rate/action_model': 1.4071692767190001e-05, 'epoch': 0.29}) +[2026-10-03 13:49:53,613][starVLA.training.train_starvla][INFO] - Step 4021, Loss: {'train/loss': 5.543177394429222e-06, 'train/grad_norm_pre_clip': 0.0018558797892183065, 'timing/data': 0.00031681000837124884, 'timing/model': 0.3899192660028348, 'learning_rate/qwen_vl_interface': 1.4053800128476426e-06, 'learning_rate/action_model': 1.4053800128476427e-05, 'epoch': 0.29}) +[2026-10-03 13:49:54,311][starVLA.training.train_starvla][INFO] - Step 4022, Loss: {'train/loss': 0.00021493287931662053, 'train/grad_norm_pre_clip': 0.31174248456954956, 'timing/data': 0.00020619000133592635, 'timing/model': 0.3905553049989976, 'learning_rate/qwen_vl_interface': 1.403592329354338e-06, 'learning_rate/action_model': 1.403592329354338e-05, 'epoch': 0.29}) +[2026-10-03 13:49:55,009][starVLA.training.train_starvla][INFO] - Step 4023, Loss: {'train/loss': 7.170368189690635e-05, 'train/grad_norm_pre_clip': 0.08574333041906357, 'timing/data': 0.0002089000045089051, 'timing/model': 0.3902635660051601, 'learning_rate/qwen_vl_interface': 1.4018062269739354e-06, 'learning_rate/action_model': 1.4018062269739355e-05, 'epoch': 0.29}) +[2026-10-03 13:49:55,722][starVLA.training.train_starvla][INFO] - Step 4024, Loss: {'train/loss': 0.631813645362854, 'train/grad_norm_pre_clip': 206.47096252441406, 'timing/data': 0.0002611700037959963, 'timing/model': 0.4052871729945764, 'learning_rate/qwen_vl_interface': 1.4000217064406339e-06, 'learning_rate/action_model': 1.4000217064406337e-05, 'epoch': 0.29}) +[2026-10-03 13:49:56,624][starVLA.training.train_starvla][INFO] - Step 4025, Loss: {'train/loss': 1.6062656868598424e-05, 'train/grad_norm_pre_clip': 0.0037563536316156387, 'timing/data': 0.000258829997619614, 'timing/model': 0.5942769620014587, 'learning_rate/qwen_vl_interface': 1.3982387684879836e-06, 'learning_rate/action_model': 1.3982387684879836e-05, 'epoch': 0.29}) +[2026-10-03 13:49:57,324][starVLA.training.train_starvla][INFO] - Step 4026, Loss: {'train/loss': 0.4088781476020813, 'train/grad_norm_pre_clip': 76.19361114501953, 'timing/data': 0.00021448000916279852, 'timing/model': 0.3909056950069498, 'learning_rate/qwen_vl_interface': 1.3964574138488804e-06, 'learning_rate/action_model': 1.3964574138488803e-05, 'epoch': 0.29}) +[2026-10-03 13:49:58,021][starVLA.training.train_starvla][INFO] - Step 4027, Loss: {'train/loss': 4.142491889069788e-06, 'train/grad_norm_pre_clip': 0.0009699336369521916, 'timing/data': 0.00020701999892480671, 'timing/model': 0.39079257599951234, 'learning_rate/qwen_vl_interface': 1.3946776432555743e-06, 'learning_rate/action_model': 1.3946776432555741e-05, 'epoch': 0.29}) +[2026-10-03 13:49:58,734][starVLA.training.train_starvla][INFO] - Step 4028, Loss: {'train/loss': 9.327804036729503e-06, 'train/grad_norm_pre_clip': 0.002854488091543317, 'timing/data': 0.00027766999846789986, 'timing/model': 0.4048262130090734, 'learning_rate/qwen_vl_interface': 1.3928994574396595e-06, 'learning_rate/action_model': 1.3928994574396594e-05, 'epoch': 0.29}) +[2026-10-03 13:49:59,433][starVLA.training.train_starvla][INFO] - Step 4029, Loss: {'train/loss': 2.8626969651668333e-05, 'train/grad_norm_pre_clip': 0.014777271077036858, 'timing/data': 0.0002090900088660419, 'timing/model': 0.3910457750025671, 'learning_rate/qwen_vl_interface': 1.391122857132083e-06, 'learning_rate/action_model': 1.3911228571320831e-05, 'epoch': 0.29}) +[2026-10-03 13:50:00,338][starVLA.training.train_starvla][INFO] - Step 4030, Loss: {'train/loss': 0.0009428277262486517, 'train/grad_norm_pre_clip': 2.2004897594451904, 'timing/data': 0.00020747999951709062, 'timing/model': 0.5974799719988368, 'learning_rate/qwen_vl_interface': 1.3893478430631376e-06, 'learning_rate/action_model': 1.3893478430631376e-05, 'epoch': 0.29}) +[2026-10-03 13:50:01,036][starVLA.training.train_starvla][INFO] - Step 4031, Loss: {'train/loss': 7.822987754479982e-06, 'train/grad_norm_pre_clip': 0.005484886933118105, 'timing/data': 0.00026447999698575586, 'timing/model': 0.3910929859994212, 'learning_rate/qwen_vl_interface': 1.3875744159624652e-06, 'learning_rate/action_model': 1.3875744159624652e-05, 'epoch': 0.3}) +[2026-10-03 13:50:01,750][starVLA.training.train_starvla][INFO] - Step 4032, Loss: {'train/loss': 0.3049694895744324, 'train/grad_norm_pre_clip': 114.68531036376953, 'timing/data': 0.0003594699956011027, 'timing/model': 0.4056075230037095, 'learning_rate/qwen_vl_interface': 1.3858025765590525e-06, 'learning_rate/action_model': 1.3858025765590525e-05, 'epoch': 0.3}) +[2026-10-03 13:50:02,448][starVLA.training.train_starvla][INFO] - Step 4033, Loss: {'train/loss': 1.680016066529788e-05, 'train/grad_norm_pre_clip': 0.003724067471921444, 'timing/data': 0.00034587999107316136, 'timing/model': 0.39006114601215813, 'learning_rate/qwen_vl_interface': 1.384032325581237e-06, 'learning_rate/action_model': 1.3840323255812371e-05, 'epoch': 0.3}) +[2026-10-03 13:50:03,149][starVLA.training.train_starvla][INFO] - Step 4034, Loss: {'train/loss': 3.5567481972975656e-05, 'train/grad_norm_pre_clip': 0.011577825993299484, 'timing/data': 0.00020942999981343746, 'timing/model': 0.39223915500042494, 'learning_rate/qwen_vl_interface': 1.3822636637567025e-06, 'learning_rate/action_model': 1.3822636637567024e-05, 'epoch': 0.3}) +[2026-10-03 13:50:03,849][starVLA.training.train_starvla][INFO] - Step 4035, Loss: {'train/loss': 0.00012073949619662017, 'train/grad_norm_pre_clip': 0.16848096251487732, 'timing/data': 0.00021346998983062804, 'timing/model': 0.39087814498634543, 'learning_rate/qwen_vl_interface': 1.3804965918124788e-06, 'learning_rate/action_model': 1.3804965918124787e-05, 'epoch': 0.3}) +[2026-10-03 13:50:04,753][starVLA.training.train_starvla][INFO] - Step 4036, Loss: {'train/loss': 1.8542621546657756e-05, 'train/grad_norm_pre_clip': 0.0038606657180935144, 'timing/data': 0.0002826899872161448, 'timing/model': 0.392612966010347, 'learning_rate/qwen_vl_interface': 1.3787311104749418e-06, 'learning_rate/action_model': 1.3787311104749418e-05, 'epoch': 0.3}) +[2026-10-03 13:50:05,456][starVLA.training.train_starvla][INFO] - Step 4037, Loss: {'train/loss': 6.3925463109626435e-06, 'train/grad_norm_pre_clip': 0.0014095106162130833, 'timing/data': 0.0003008700005011633, 'timing/model': 0.39300128500326537, 'learning_rate/qwen_vl_interface': 1.3769672204698149e-06, 'learning_rate/action_model': 1.3769672204698147e-05, 'epoch': 0.3}) +[2026-10-03 13:50:06,156][starVLA.training.train_starvla][INFO] - Step 4038, Loss: {'train/loss': 7.897529940237291e-06, 'train/grad_norm_pre_clip': 0.0015688289422541857, 'timing/data': 0.0003367500030435622, 'timing/model': 0.3916274160001194, 'learning_rate/qwen_vl_interface': 1.3752049225221664e-06, 'learning_rate/action_model': 1.3752049225221663e-05, 'epoch': 0.3}) +[2026-10-03 13:50:06,856][starVLA.training.train_starvla][INFO] - Step 4039, Loss: {'train/loss': 3.356471279403195e-05, 'train/grad_norm_pre_clip': 0.012223908677697182, 'timing/data': 0.0003036499983863905, 'timing/model': 0.39107977600360755, 'learning_rate/qwen_vl_interface': 1.3734442173564103e-06, 'learning_rate/action_model': 1.3734442173564102e-05, 'epoch': 0.3}) +[2026-10-03 13:50:07,557][starVLA.training.train_starvla][INFO] - Step 4040, Loss: {'train/loss': 9.988552483264357e-05, 'train/grad_norm_pre_clip': 0.08341845124959946, 'timing/data': 0.0001974199985852465, 'timing/model': 0.3927368349977769, 'learning_rate/qwen_vl_interface': 1.3716851056963062e-06, 'learning_rate/action_model': 1.3716851056963062e-05, 'epoch': 0.3}) +[2026-10-03 13:50:08,474][starVLA.training.train_starvla][INFO] - Step 4041, Loss: {'train/loss': 3.65820960723795e-06, 'train/grad_norm_pre_clip': 0.0011151923099532723, 'timing/data': 0.00020559999393299222, 'timing/model': 0.39285504500730895, 'learning_rate/qwen_vl_interface': 1.3699275882649597e-06, 'learning_rate/action_model': 1.3699275882649596e-05, 'epoch': 0.3}) +[2026-10-03 13:50:09,178][starVLA.training.train_starvla][INFO] - Step 4042, Loss: {'train/loss': 0.0015958817675709724, 'train/grad_norm_pre_clip': 3.1543681621551514, 'timing/data': 0.00036396000359673053, 'timing/model': 0.3930484660086222, 'learning_rate/qwen_vl_interface': 1.368171665784817e-06, 'learning_rate/action_model': 1.3681716657848171e-05, 'epoch': 0.3}) +[2026-10-03 13:50:09,878][starVLA.training.train_starvla][INFO] - Step 4043, Loss: {'train/loss': 0.8981318473815918, 'train/grad_norm_pre_clip': 54.824310302734375, 'timing/data': 0.00033657999301794916, 'timing/model': 0.39043120499991346, 'learning_rate/qwen_vl_interface': 1.3664173389776735e-06, 'learning_rate/action_model': 1.3664173389776735e-05, 'epoch': 0.3}) +[2026-10-03 13:50:10,576][starVLA.training.train_starvla][INFO] - Step 4044, Loss: {'train/loss': 0.0006572044803760946, 'train/grad_norm_pre_clip': 4.540708065032959, 'timing/data': 0.0002678199962247163, 'timing/model': 0.3908199059951585, 'learning_rate/qwen_vl_interface': 1.3646646085646658e-06, 'learning_rate/action_model': 1.3646646085646659e-05, 'epoch': 0.3}) +[2026-10-03 13:50:11,279][starVLA.training.train_starvla][INFO] - Step 4045, Loss: {'train/loss': 3.355142689542845e-05, 'train/grad_norm_pre_clip': 0.05228456109762192, 'timing/data': 0.00029537999944295734, 'timing/model': 0.39342807499633636, 'learning_rate/qwen_vl_interface': 1.3629134752662756e-06, 'learning_rate/action_model': 1.3629134752662756e-05, 'epoch': 0.3}) +[2026-10-03 13:50:12,189][starVLA.training.train_starvla][INFO] - Step 4046, Loss: {'train/loss': 3.5837119867210276e-06, 'train/grad_norm_pre_clip': 0.000917587080039084, 'timing/data': 0.00023206998594105244, 'timing/model': 0.3905435960041359, 'learning_rate/qwen_vl_interface': 1.3611639398023282e-06, 'learning_rate/action_model': 1.3611639398023283e-05, 'epoch': 0.3}) +[2026-10-03 13:50:12,888][starVLA.training.train_starvla][INFO] - Step 4047, Loss: {'train/loss': 1.149583749793237e-05, 'train/grad_norm_pre_clip': 0.006396390963345766, 'timing/data': 0.0003402199945412576, 'timing/model': 0.3906898360000923, 'learning_rate/qwen_vl_interface': 1.35941600289199e-06, 'learning_rate/action_model': 1.35941600289199e-05, 'epoch': 0.3}) +[2026-10-03 13:50:13,589][starVLA.training.train_starvla][INFO] - Step 4048, Loss: {'train/loss': 4.4181724661029875e-06, 'train/grad_norm_pre_clip': 0.0019526057876646519, 'timing/data': 0.00019970000721514225, 'timing/model': 0.3925663859990891, 'learning_rate/qwen_vl_interface': 1.3576696652537735e-06, 'learning_rate/action_model': 1.3576696652537736e-05, 'epoch': 0.3}) +[2026-10-03 13:50:14,292][starVLA.training.train_starvla][INFO] - Step 4049, Loss: {'train/loss': 2.5121042199316435e-05, 'train/grad_norm_pre_clip': 0.028415022417902946, 'timing/data': 0.00021284000831656158, 'timing/model': 0.3925106159877032, 'learning_rate/qwen_vl_interface': 1.355924927605531e-06, 'learning_rate/action_model': 1.3559249276055311e-05, 'epoch': 0.3}) +[2026-10-03 13:50:14,994][starVLA.training.train_starvla][INFO] - Step 4050, Loss: {'train/loss': 1.0966486115648877e-05, 'train/grad_norm_pre_clip': 0.0036948423366993666, 'timing/data': 0.0003252599999541417, 'timing/model': 0.39435110600607004, 'learning_rate/qwen_vl_interface': 1.3541817906644591e-06, 'learning_rate/action_model': 1.3541817906644591e-05, 'epoch': 0.3}) +[2026-10-03 13:50:15,914][starVLA.training.train_starvla][INFO] - Step 4051, Loss: {'train/loss': 0.00908729713410139, 'train/grad_norm_pre_clip': 31.28524398803711, 'timing/data': 0.00036997999995946884, 'timing/model': 0.6090107600030024, 'learning_rate/qwen_vl_interface': 1.3524402551470955e-06, 'learning_rate/action_model': 1.3524402551470956e-05, 'epoch': 0.3}) +[2026-10-03 13:50:16,615][starVLA.training.train_starvla][INFO] - Step 4052, Loss: {'train/loss': 0.017116205766797066, 'train/grad_norm_pre_clip': 81.9926986694336, 'timing/data': 0.0002449000021442771, 'timing/model': 0.39373800600878894, 'learning_rate/qwen_vl_interface': 1.3507003217693196e-06, 'learning_rate/action_model': 1.3507003217693196e-05, 'epoch': 0.3}) +[2026-10-03 13:50:17,316][starVLA.training.train_starvla][INFO] - Step 4053, Loss: {'train/loss': 4.537364475254435e-06, 'train/grad_norm_pre_clip': 0.0014740143669769168, 'timing/data': 0.0003799800033448264, 'timing/model': 0.3918028859916376, 'learning_rate/qwen_vl_interface': 1.3489619912463515e-06, 'learning_rate/action_model': 1.3489619912463514e-05, 'epoch': 0.3}) +[2026-10-03 13:50:18,015][starVLA.training.train_starvla][INFO] - Step 4054, Loss: {'train/loss': 4.000931767222937e-06, 'train/grad_norm_pre_clip': 0.0013879382750019431, 'timing/data': 0.0002143500023521483, 'timing/model': 0.39097572599712294, 'learning_rate/qwen_vl_interface': 1.3472252642927534e-06, 'learning_rate/action_model': 1.3472252642927533e-05, 'epoch': 0.3}) +[2026-10-03 13:50:18,713][starVLA.training.train_starvla][INFO] - Step 4055, Loss: {'train/loss': 2.771606659734971e-06, 'train/grad_norm_pre_clip': 0.0007492860895581543, 'timing/data': 0.00033660000190138817, 'timing/model': 0.39049540599808097, 'learning_rate/qwen_vl_interface': 1.3454901416224284e-06, 'learning_rate/action_model': 1.3454901416224284e-05, 'epoch': 0.3}) +[2026-10-03 13:50:19,630][starVLA.training.train_starvla][INFO] - Step 4056, Loss: {'train/loss': 3.913149339496158e-05, 'train/grad_norm_pre_clip': 0.020600775256752968, 'timing/data': 0.0002776299952529371, 'timing/model': 0.6089148910104996, 'learning_rate/qwen_vl_interface': 1.34375662394862e-06, 'learning_rate/action_model': 1.34375662394862e-05, 'epoch': 0.3}) +[2026-10-03 13:50:20,330][starVLA.training.train_starvla][INFO] - Step 4057, Loss: {'train/loss': 7.40758259780705e-05, 'train/grad_norm_pre_clip': 0.04004795849323273, 'timing/data': 0.0003395990061108023, 'timing/model': 0.3910864569916157, 'learning_rate/qwen_vl_interface': 1.3420247119839116e-06, 'learning_rate/action_model': 1.3420247119839117e-05, 'epoch': 0.3}) +[2026-10-03 13:50:21,029][starVLA.training.train_starvla][INFO] - Step 4058, Loss: {'train/loss': 7.532462859671796e-06, 'train/grad_norm_pre_clip': 0.0015306606655940413, 'timing/data': 0.00024791000760160387, 'timing/model': 0.3907331560039893, 'learning_rate/qwen_vl_interface': 1.340294406440227e-06, 'learning_rate/action_model': 1.340294406440227e-05, 'epoch': 0.3}) +[2026-10-03 13:50:21,732][starVLA.training.train_starvla][INFO] - Step 4059, Loss: {'train/loss': 0.707656979560852, 'train/grad_norm_pre_clip': 265.0579833984375, 'timing/data': 0.0003556400042725727, 'timing/model': 0.39368748500419315, 'learning_rate/qwen_vl_interface': 1.3385657080288297e-06, 'learning_rate/action_model': 1.3385657080288297e-05, 'epoch': 0.3}) +[2026-10-03 13:50:22,438][starVLA.training.train_starvla][INFO] - Step 4060, Loss: {'train/loss': 3.762509322768892e-06, 'train/grad_norm_pre_clip': 0.0010269077029079199, 'timing/data': 0.0003295799979241565, 'timing/model': 0.3944589949969668, 'learning_rate/qwen_vl_interface': 1.336838617460322e-06, 'learning_rate/action_model': 1.336838617460322e-05, 'epoch': 0.3}) +[2026-10-03 13:50:23,139][starVLA.training.train_starvla][INFO] - Step 4061, Loss: {'train/loss': 2.8452113838284276e-05, 'train/grad_norm_pre_clip': 0.012897980399429798, 'timing/data': 0.0003612200089264661, 'timing/model': 0.39138889699825086, 'learning_rate/qwen_vl_interface': 1.335113135444646e-06, 'learning_rate/action_model': 1.335113135444646e-05, 'epoch': 0.3}) +[2026-10-03 13:50:24,037][starVLA.training.train_starvla][INFO] - Step 4062, Loss: {'train/loss': 7.025794275250519e-06, 'train/grad_norm_pre_clip': 0.0026220178697258234, 'timing/data': 0.0003202099906047806, 'timing/model': 0.39085523500398267, 'learning_rate/qwen_vl_interface': 1.3333892626910823e-06, 'learning_rate/action_model': 1.3333892626910823e-05, 'epoch': 0.3}) +[2026-10-03 13:50:24,738][starVLA.training.train_starvla][INFO] - Step 4063, Loss: {'train/loss': 8.441318641416728e-06, 'train/grad_norm_pre_clip': 0.002656012075021863, 'timing/data': 0.00024352999753318727, 'timing/model': 0.3934780760027934, 'learning_rate/qwen_vl_interface': 1.3316669999082486e-06, 'learning_rate/action_model': 1.3316669999082484e-05, 'epoch': 0.3}) +[2026-10-03 13:50:25,441][starVLA.training.train_starvla][INFO] - Step 4064, Loss: {'train/loss': 3.2409905088570667e-06, 'train/grad_norm_pre_clip': 0.0005630149971693754, 'timing/data': 0.00031088000105228275, 'timing/model': 0.3933499350096099, 'learning_rate/qwen_vl_interface': 1.3299463478041024e-06, 'learning_rate/action_model': 1.3299463478041024e-05, 'epoch': 0.3}) +[2026-10-03 13:50:26,146][starVLA.training.train_starvla][INFO] - Step 4065, Loss: {'train/loss': 0.00023532495833933353, 'train/grad_norm_pre_clip': 0.22544753551483154, 'timing/data': 0.0003062699979636818, 'timing/model': 0.39610953599913046, 'learning_rate/qwen_vl_interface': 1.3282273070859396e-06, 'learning_rate/action_model': 1.3282273070859396e-05, 'epoch': 0.3}) +[2026-10-03 13:50:26,961][starVLA.training.train_starvla][INFO] - Step 4066, Loss: {'train/loss': 4.611882104654796e-06, 'train/grad_norm_pre_clip': 0.0008799625793471932, 'timing/data': 0.00022178998915478587, 'timing/model': 0.39438571500068065, 'learning_rate/qwen_vl_interface': 1.3265098784603924e-06, 'learning_rate/action_model': 1.3265098784603924e-05, 'epoch': 0.3}) +[2026-10-03 13:50:27,880][starVLA.training.train_starvla][INFO] - Step 4067, Loss: {'train/loss': 1.4960145563236438e-05, 'train/grad_norm_pre_clip': 0.006460895761847496, 'timing/data': 0.0003413800004636869, 'timing/model': 0.6062623910111142, 'learning_rate/qwen_vl_interface': 1.324794062633431e-06, 'learning_rate/action_model': 1.324794062633431e-05, 'epoch': 0.3}) +[2026-10-03 13:50:28,588][starVLA.training.train_starvla][INFO] - Step 4068, Loss: {'train/loss': 0.00045198132283985615, 'train/grad_norm_pre_clip': 0.28297945857048035, 'timing/data': 0.0003637189947767183, 'timing/model': 0.39570565600297414, 'learning_rate/qwen_vl_interface': 1.323079860310363e-06, 'learning_rate/action_model': 1.3230798603103628e-05, 'epoch': 0.3}) +[2026-10-03 13:50:29,297][starVLA.training.train_starvla][INFO] - Step 4069, Loss: {'train/loss': 0.00026813976000994444, 'train/grad_norm_pre_clip': 0.2565551698207855, 'timing/data': 0.00025258000823669136, 'timing/model': 0.39545828600239474, 'learning_rate/qwen_vl_interface': 1.3213672721958309e-06, 'learning_rate/action_model': 1.321367272195831e-05, 'epoch': 0.3}) +[2026-10-03 13:50:30,001][starVLA.training.train_starvla][INFO] - Step 4070, Loss: {'train/loss': 0.00010732645750977099, 'train/grad_norm_pre_clip': 0.18692438304424286, 'timing/data': 0.00022063001233618706, 'timing/model': 0.3942478259996278, 'learning_rate/qwen_vl_interface': 1.3196562989938175e-06, 'learning_rate/action_model': 1.3196562989938174e-05, 'epoch': 0.3}) +[2026-10-03 13:50:30,707][starVLA.training.train_starvla][INFO] - Step 4071, Loss: {'train/loss': 1.4244878911995329e-05, 'train/grad_norm_pre_clip': 0.007998243905603886, 'timing/data': 0.00021194000146351755, 'timing/model': 0.39422787600778975, 'learning_rate/qwen_vl_interface': 1.3179469414076357e-06, 'learning_rate/action_model': 1.3179469414076356e-05, 'epoch': 0.3}) +[2026-10-03 13:50:31,414][starVLA.training.train_starvla][INFO] - Step 4072, Loss: {'train/loss': 5.10356449012761e-06, 'train/grad_norm_pre_clip': 0.0028587752021849155, 'timing/data': 0.0003528700035531074, 'timing/model': 0.3944336650019977, 'learning_rate/qwen_vl_interface': 1.3162392001399401e-06, 'learning_rate/action_model': 1.3162392001399401e-05, 'epoch': 0.3}) +[2026-10-03 13:50:32,323][starVLA.training.train_starvla][INFO] - Step 4073, Loss: {'train/loss': 0.00013666666927747428, 'train/grad_norm_pre_clip': 0.16233514249324799, 'timing/data': 0.00027076000696979463, 'timing/model': 0.39462202599679586, 'learning_rate/qwen_vl_interface': 1.3145330758927182e-06, 'learning_rate/action_model': 1.3145330758927182e-05, 'epoch': 0.3}) +[2026-10-03 13:50:33,027][starVLA.training.train_starvla][INFO] - Step 4074, Loss: {'train/loss': 0.00020341042545624077, 'train/grad_norm_pre_clip': 0.19957518577575684, 'timing/data': 0.00037137999606784433, 'timing/model': 0.39357527600077447, 'learning_rate/qwen_vl_interface': 1.3128285693672935e-06, 'learning_rate/action_model': 1.3128285693672934e-05, 'epoch': 0.3}) +[2026-10-03 13:50:33,735][starVLA.training.train_starvla][INFO] - Step 4075, Loss: {'train/loss': 0.4689151346683502, 'train/grad_norm_pre_clip': 79.85518646240234, 'timing/data': 0.00036720000207424164, 'timing/model': 0.39469239600293804, 'learning_rate/qwen_vl_interface': 1.311125681264325e-06, 'learning_rate/action_model': 1.311125681264325e-05, 'epoch': 0.3}) +[2026-10-03 13:50:34,440][starVLA.training.train_starvla][INFO] - Step 4076, Loss: {'train/loss': 0.0003000883443746716, 'train/grad_norm_pre_clip': 0.12592923641204834, 'timing/data': 0.00021923000167589635, 'timing/model': 0.3933417060034117, 'learning_rate/qwen_vl_interface': 1.309424412283803e-06, 'learning_rate/action_model': 1.3094244122838029e-05, 'epoch': 0.3}) +[2026-10-03 13:50:35,149][starVLA.training.train_starvla][INFO] - Step 4077, Loss: {'train/loss': 1.159252042270964e-05, 'train/grad_norm_pre_clip': 0.004766399506479502, 'timing/data': 0.00030137998692225665, 'timing/model': 0.39672083500772715, 'learning_rate/qwen_vl_interface': 1.307724763125057e-06, 'learning_rate/action_model': 1.307724763125057e-05, 'epoch': 0.3}) +[2026-10-03 13:50:36,081][starVLA.training.train_starvla][INFO] - Step 4078, Loss: {'train/loss': 0.002259073080495, 'train/grad_norm_pre_clip': 4.668489933013916, 'timing/data': 0.0003715900093084201, 'timing/model': 0.3938648260082118, 'learning_rate/qwen_vl_interface': 1.3060267344867478e-06, 'learning_rate/action_model': 1.3060267344867478e-05, 'epoch': 0.3}) +[2026-10-03 13:50:36,788][starVLA.training.train_starvla][INFO] - Step 4079, Loss: {'train/loss': 3.1421084713656455e-05, 'train/grad_norm_pre_clip': 0.014673381112515926, 'timing/data': 0.00037787000474054366, 'timing/model': 0.394636746001197, 'learning_rate/qwen_vl_interface': 1.3043303270668715e-06, 'learning_rate/action_model': 1.3043303270668713e-05, 'epoch': 0.3}) +[2026-10-03 13:50:37,493][starVLA.training.train_starvla][INFO] - Step 4080, Loss: {'train/loss': 4.595563223119825e-05, 'train/grad_norm_pre_clip': 0.05540088564157486, 'timing/data': 0.00022200999956112355, 'timing/model': 0.39408333599567413, 'learning_rate/qwen_vl_interface': 1.3026355415627565e-06, 'learning_rate/action_model': 1.3026355415627564e-05, 'epoch': 0.3}) +[2026-10-03 13:50:38,199][starVLA.training.train_starvla][INFO] - Step 4081, Loss: {'train/loss': 4.053077191201737e-06, 'train/grad_norm_pre_clip': 0.001226757769472897, 'timing/data': 0.000264208996668458, 'timing/model': 0.3945868160080863, 'learning_rate/qwen_vl_interface': 1.3009423786710653e-06, 'learning_rate/action_model': 1.3009423786710653e-05, 'epoch': 0.3}) +[2026-10-03 13:50:38,905][starVLA.training.train_starvla][INFO] - Step 4082, Loss: {'train/loss': 0.29157936573028564, 'train/grad_norm_pre_clip': 77.1829833984375, 'timing/data': 0.0003602399956434965, 'timing/model': 0.39426685600483324, 'learning_rate/qwen_vl_interface': 1.2992508390877933e-06, 'learning_rate/action_model': 1.2992508390877933e-05, 'epoch': 0.3}) +[2026-10-03 13:50:39,818][starVLA.training.train_starvla][INFO] - Step 4083, Loss: {'train/loss': 7.041684875730425e-05, 'train/grad_norm_pre_clip': 0.036647241562604904, 'timing/data': 0.00024563999613747, 'timing/model': 0.6011514929996338, 'learning_rate/qwen_vl_interface': 1.297560923508269e-06, 'learning_rate/action_model': 1.297560923508269e-05, 'epoch': 0.3}) +[2026-10-03 13:50:40,523][starVLA.training.train_starvla][INFO] - Step 4084, Loss: {'train/loss': 1.1868578440044075e-05, 'train/grad_norm_pre_clip': 0.004007179290056229, 'timing/data': 0.0003166499955113977, 'timing/model': 0.39371185599884484, 'learning_rate/qwen_vl_interface': 1.295872632627152e-06, 'learning_rate/action_model': 1.2958726326271519e-05, 'epoch': 0.3}) +[2026-10-03 13:50:41,230][starVLA.training.train_starvla][INFO] - Step 4085, Loss: {'train/loss': 4.790669208887266e-06, 'train/grad_norm_pre_clip': 0.001317871967330575, 'timing/data': 0.00031581000075675547, 'timing/model': 0.3955780459946254, 'learning_rate/qwen_vl_interface': 1.2941859671384364e-06, 'learning_rate/action_model': 1.2941859671384362e-05, 'epoch': 0.3}) +[2026-10-03 13:50:41,937][starVLA.training.train_starvla][INFO] - Step 4086, Loss: {'train/loss': 1.9354891264811158e-05, 'train/grad_norm_pre_clip': 0.01902112178504467, 'timing/data': 0.00037318999238777906, 'timing/model': 0.3949095260031754, 'learning_rate/qwen_vl_interface': 1.2925009277354444e-06, 'learning_rate/action_model': 1.2925009277354444e-05, 'epoch': 0.3}) +[2026-10-03 13:50:42,646][starVLA.training.train_starvla][INFO] - Step 4087, Loss: {'train/loss': 0.0007896878523752093, 'train/grad_norm_pre_clip': 0.8496909737586975, 'timing/data': 0.0003756499936571345, 'timing/model': 0.3952997560118092, 'learning_rate/qwen_vl_interface': 1.2908175151108335e-06, 'learning_rate/action_model': 1.2908175151108334e-05, 'epoch': 0.3}) +[2026-10-03 13:50:43,351][starVLA.training.train_starvla][INFO] - Step 4088, Loss: {'train/loss': 0.00029404828092083335, 'train/grad_norm_pre_clip': 0.3406449854373932, 'timing/data': 0.00033322999661322683, 'timing/model': 0.3941430460108677, 'learning_rate/qwen_vl_interface': 1.2891357299565899e-06, 'learning_rate/action_model': 1.2891357299565898e-05, 'epoch': 0.3}) +[2026-10-03 13:50:44,256][starVLA.training.train_starvla][INFO] - Step 4089, Loss: {'train/loss': 0.00019690360932145268, 'train/grad_norm_pre_clip': 0.19085170328617096, 'timing/data': 0.0002408199943602085, 'timing/model': 0.3948273560090456, 'learning_rate/qwen_vl_interface': 1.2874555729640333e-06, 'learning_rate/action_model': 1.2874555729640334e-05, 'epoch': 0.3}) +[2026-10-03 13:50:44,964][starVLA.training.train_starvla][INFO] - Step 4090, Loss: {'train/loss': 0.0009665561956353486, 'train/grad_norm_pre_clip': 1.359495997428894, 'timing/data': 0.00040194000757765025, 'timing/model': 0.39537505598855205, 'learning_rate/qwen_vl_interface': 1.2857770448238124e-06, 'learning_rate/action_model': 1.2857770448238123e-05, 'epoch': 0.3}) +[2026-10-03 13:50:45,669][starVLA.training.train_starvla][INFO] - Step 4091, Loss: {'train/loss': 0.3909403085708618, 'train/grad_norm_pre_clip': 62.74900436401367, 'timing/data': 0.0003207300032954663, 'timing/model': 0.39390285599802155, 'learning_rate/qwen_vl_interface': 1.2841001462259067e-06, 'learning_rate/action_model': 1.2841001462259067e-05, 'epoch': 0.3}) +[2026-10-03 13:50:46,372][starVLA.training.train_starvla][INFO] - Step 4092, Loss: {'train/loss': 3.608168844948523e-05, 'train/grad_norm_pre_clip': 0.07466140389442444, 'timing/data': 0.0003587499959394336, 'timing/model': 0.3919025170034729, 'learning_rate/qwen_vl_interface': 1.282424877859626e-06, 'learning_rate/action_model': 1.282424877859626e-05, 'epoch': 0.3}) +[2026-10-03 13:50:47,080][starVLA.training.train_starvla][INFO] - Step 4093, Loss: {'train/loss': 0.10493150353431702, 'train/grad_norm_pre_clip': 119.32913208007812, 'timing/data': 0.00026376001187600195, 'timing/model': 0.39565269599552266, 'learning_rate/qwen_vl_interface': 1.2807512404136088e-06, 'learning_rate/action_model': 1.2807512404136087e-05, 'epoch': 0.3}) +[2026-10-03 13:50:48,000][starVLA.training.train_starvla][INFO] - Step 4094, Loss: {'train/loss': 0.00026543339481577277, 'train/grad_norm_pre_clip': 0.5297307372093201, 'timing/data': 0.0002421200042590499, 'timing/model': 0.39352638699347153, 'learning_rate/qwen_vl_interface': 1.2790792345758255e-06, 'learning_rate/action_model': 1.2790792345758254e-05, 'epoch': 0.3}) +[2026-10-03 13:50:48,705][starVLA.training.train_starvla][INFO] - Step 4095, Loss: {'train/loss': 7.889833796070889e-06, 'train/grad_norm_pre_clip': 0.006539345718920231, 'timing/data': 0.0002064689906546846, 'timing/model': 0.39380919699033257, 'learning_rate/qwen_vl_interface': 1.2774088610335752e-06, 'learning_rate/action_model': 1.2774088610335752e-05, 'epoch': 0.3}) +[2026-10-03 13:50:49,413][starVLA.training.train_starvla][INFO] - Step 4096, Loss: {'train/loss': 4.8553021770203486e-05, 'train/grad_norm_pre_clip': 0.033328328281641006, 'timing/data': 0.0003706600109580904, 'timing/model': 0.3943575560115278, 'learning_rate/qwen_vl_interface': 1.2757401204734827e-06, 'learning_rate/action_model': 1.2757401204734826e-05, 'epoch': 0.3}) +[2026-10-03 13:50:50,118][starVLA.training.train_starvla][INFO] - Step 4097, Loss: {'train/loss': 0.0009048728388734162, 'train/grad_norm_pre_clip': 1.4904049634933472, 'timing/data': 0.0003269500011811033, 'timing/model': 0.39538410599925555, 'learning_rate/qwen_vl_interface': 1.2740730135815064e-06, 'learning_rate/action_model': 1.2740730135815062e-05, 'epoch': 0.3}) +[2026-10-03 13:50:50,825][starVLA.training.train_starvla][INFO] - Step 4098, Loss: {'train/loss': 2.749255827438901e-06, 'train/grad_norm_pre_clip': 0.00078546442091465, 'timing/data': 0.0002479700051480904, 'timing/model': 0.3941717160050757, 'learning_rate/qwen_vl_interface': 1.2724075410429298e-06, 'learning_rate/action_model': 1.2724075410429299e-05, 'epoch': 0.3}) +[2026-10-03 13:50:51,732][starVLA.training.train_starvla][INFO] - Step 4099, Loss: {'train/loss': 0.39551785588264465, 'train/grad_norm_pre_clip': 44.58920669555664, 'timing/data': 0.00022806999913882464, 'timing/model': 0.39329167699906975, 'learning_rate/qwen_vl_interface': 1.2707437035423664e-06, 'learning_rate/action_model': 1.2707437035423664e-05, 'epoch': 0.3}) +[2026-10-03 13:50:52,438][starVLA.training.train_starvla][INFO] - Step 4100, Loss: {'train/loss': 2.3368806068901904e-05, 'train/grad_norm_pre_clip': 0.030376899987459183, 'timing/data': 0.00033337000058963895, 'timing/model': 0.39352947699080687, 'learning_rate/qwen_vl_interface': 1.2690815017637573e-06, 'learning_rate/action_model': 1.2690815017637572e-05, 'epoch': 0.3}) +[2026-10-03 13:50:53,143][starVLA.training.train_starvla][INFO] - Step 4101, Loss: {'train/loss': 2.1374276911956258e-05, 'train/grad_norm_pre_clip': 0.004604777321219444, 'timing/data': 0.00035877998743671924, 'timing/model': 0.3944786159991054, 'learning_rate/qwen_vl_interface': 1.2674209363903694e-06, 'learning_rate/action_model': 1.2674209363903693e-05, 'epoch': 0.3}) +[2026-10-03 13:50:53,850][starVLA.training.train_starvla][INFO] - Step 4102, Loss: {'train/loss': 7.368470960500417e-06, 'train/grad_norm_pre_clip': 0.002389147412031889, 'timing/data': 0.0003671099984785542, 'timing/model': 0.3948407860007137, 'learning_rate/qwen_vl_interface': 1.2657620081048005e-06, 'learning_rate/action_model': 1.2657620081048006e-05, 'epoch': 0.3}) +[2026-10-03 13:50:54,558][starVLA.training.train_starvla][INFO] - Step 4103, Loss: {'train/loss': 1.4825196558376774e-05, 'train/grad_norm_pre_clip': 0.027849039062857628, 'timing/data': 0.00032473000464960933, 'timing/model': 0.39564943699224386, 'learning_rate/qwen_vl_interface': 1.2641047175889714e-06, 'learning_rate/action_model': 1.2641047175889714e-05, 'epoch': 0.3}) +[2026-10-03 13:50:55,474][starVLA.training.train_starvla][INFO] - Step 4104, Loss: {'train/loss': 8.994524250738323e-05, 'train/grad_norm_pre_clip': 0.059126995503902435, 'timing/data': 0.00041754999256227165, 'timing/model': 0.6034700230084127, 'learning_rate/qwen_vl_interface': 1.2624490655241332e-06, 'learning_rate/action_model': 1.262449065524133e-05, 'epoch': 0.3}) +[2026-10-03 13:50:56,181][starVLA.training.train_starvla][INFO] - Step 4105, Loss: {'train/loss': 3.9030652260407805e-05, 'train/grad_norm_pre_clip': 0.010157094337046146, 'timing/data': 0.0003168800030834973, 'timing/model': 0.3948273360001622, 'learning_rate/qwen_vl_interface': 1.2607950525908602e-06, 'learning_rate/action_model': 1.2607950525908601e-05, 'epoch': 0.3}) +[2026-10-03 13:50:56,886][starVLA.training.train_starvla][INFO] - Step 4106, Loss: {'train/loss': 0.0003743353590834886, 'train/grad_norm_pre_clip': 0.1616315096616745, 'timing/data': 0.00021887000184506178, 'timing/model': 0.39419770600216, 'learning_rate/qwen_vl_interface': 1.2591426794690575e-06, 'learning_rate/action_model': 1.2591426794690576e-05, 'epoch': 0.3}) +[2026-10-03 13:50:57,592][starVLA.training.train_starvla][INFO] - Step 4107, Loss: {'train/loss': 0.5926694273948669, 'train/grad_norm_pre_clip': 76.67359924316406, 'timing/data': 0.0002285599912283942, 'timing/model': 0.3949306769936811, 'learning_rate/qwen_vl_interface': 1.25749194683795e-06, 'learning_rate/action_model': 1.25749194683795e-05, 'epoch': 0.3}) +[2026-10-03 13:50:58,296][starVLA.training.train_starvla][INFO] - Step 4108, Loss: {'train/loss': 2.2855232600704767e-05, 'train/grad_norm_pre_clip': 0.008656768128275871, 'timing/data': 0.0003449000068940222, 'timing/model': 0.3928793569939444, 'learning_rate/qwen_vl_interface': 1.2558428553760933e-06, 'learning_rate/action_model': 1.2558428553760932e-05, 'epoch': 0.3}) +[2026-10-03 13:50:59,211][starVLA.training.train_starvla][INFO] - Step 4109, Loss: {'train/loss': 5.394123490987113e-06, 'train/grad_norm_pre_clip': 0.003406069241464138, 'timing/data': 0.00028604999533854425, 'timing/model': 0.6027394629927585, 'learning_rate/qwen_vl_interface': 1.2541954057613659e-06, 'learning_rate/action_model': 1.2541954057613658e-05, 'epoch': 0.3}) +[2026-10-03 13:50:59,916][starVLA.training.train_starvla][INFO] - Step 4110, Loss: {'train/loss': 0.00020093616330996156, 'train/grad_norm_pre_clip': 0.3028174340724945, 'timing/data': 0.00023456000781152397, 'timing/model': 0.3942864270065911, 'learning_rate/qwen_vl_interface': 1.2525495986709727e-06, 'learning_rate/action_model': 1.2525495986709726e-05, 'epoch': 0.3}) +[2026-10-03 13:51:00,622][starVLA.training.train_starvla][INFO] - Step 4111, Loss: {'train/loss': 0.32969483733177185, 'train/grad_norm_pre_clip': 94.45268249511719, 'timing/data': 0.00026841000362765044, 'timing/model': 0.3938967669964768, 'learning_rate/qwen_vl_interface': 1.2509054347814423e-06, 'learning_rate/action_model': 1.2509054347814423e-05, 'epoch': 0.3}) +[2026-10-03 13:51:01,327][starVLA.training.train_starvla][INFO] - Step 4112, Loss: {'train/loss': 1.4087985618971288e-05, 'train/grad_norm_pre_clip': 0.004169415682554245, 'timing/data': 0.000222079994273372, 'timing/model': 0.39419827600067947, 'learning_rate/qwen_vl_interface': 1.2492629147686288e-06, 'learning_rate/action_model': 1.2492629147686289e-05, 'epoch': 0.3}) +[2026-10-03 13:51:02,034][starVLA.training.train_starvla][INFO] - Step 4113, Loss: {'train/loss': 4.39386312791612e-05, 'train/grad_norm_pre_clip': 0.03937537968158722, 'timing/data': 0.00023133998911362141, 'timing/model': 0.3949804769945331, 'learning_rate/qwen_vl_interface': 1.2476220393077098e-06, 'learning_rate/action_model': 1.2476220393077098e-05, 'epoch': 0.3}) +[2026-10-03 13:51:02,737][starVLA.training.train_starvla][INFO] - Step 4114, Loss: {'train/loss': 2.332026951989974e-06, 'train/grad_norm_pre_clip': 0.0005982925067655742, 'timing/data': 0.00037075000000186265, 'timing/model': 0.39243151700065937, 'learning_rate/qwen_vl_interface': 1.2459828090731877e-06, 'learning_rate/action_model': 1.2459828090731876e-05, 'epoch': 0.3}) +[2026-10-03 13:51:03,645][starVLA.training.train_starvla][INFO] - Step 4115, Loss: {'train/loss': 8.716873708181083e-06, 'train/grad_norm_pre_clip': 0.0032592457719147205, 'timing/data': 0.000354860007064417, 'timing/model': 0.3928022469917778, 'learning_rate/qwen_vl_interface': 1.2443452247388876e-06, 'learning_rate/action_model': 1.2443452247388877e-05, 'epoch': 0.3}) +[2026-10-03 13:51:04,349][starVLA.training.train_starvla][INFO] - Step 4116, Loss: {'train/loss': 5.975225121801486e-06, 'train/grad_norm_pre_clip': 0.0020183189772069454, 'timing/data': 0.0003346900048200041, 'timing/model': 0.39359190699178725, 'learning_rate/qwen_vl_interface': 1.2427092869779603e-06, 'learning_rate/action_model': 1.2427092869779602e-05, 'epoch': 0.3}) +[2026-10-03 13:51:05,055][starVLA.training.train_starvla][INFO] - Step 4117, Loss: {'train/loss': 2.2827576685813256e-05, 'train/grad_norm_pre_clip': 0.007316080387681723, 'timing/data': 0.0003365199954714626, 'timing/model': 0.39461976700113155, 'learning_rate/qwen_vl_interface': 1.2410749964628751e-06, 'learning_rate/action_model': 1.2410749964628751e-05, 'epoch': 0.3}) +[2026-10-03 13:51:05,760][starVLA.training.train_starvla][INFO] - Step 4118, Loss: {'train/loss': 3.245125481043942e-05, 'train/grad_norm_pre_clip': 0.01654444821178913, 'timing/data': 0.00038627001049462706, 'timing/model': 0.3934808669873746, 'learning_rate/qwen_vl_interface': 1.2394423538654296e-06, 'learning_rate/action_model': 1.2394423538654296e-05, 'epoch': 0.3}) +[2026-10-03 13:51:06,466][starVLA.training.train_starvla][INFO] - Step 4119, Loss: {'train/loss': 9.522510663373396e-05, 'train/grad_norm_pre_clip': 0.07176990061998367, 'timing/data': 0.0003916590067092329, 'timing/model': 0.39442218700423837, 'learning_rate/qwen_vl_interface': 1.2378113598567394e-06, 'learning_rate/action_model': 1.2378113598567392e-05, 'epoch': 0.3}) +[2026-10-03 13:51:07,379][starVLA.training.train_starvla][INFO] - Step 4120, Loss: {'train/loss': 7.524911779910326e-06, 'train/grad_norm_pre_clip': 0.0025267715100198984, 'timing/data': 0.0002428800071356818, 'timing/model': 0.6015546039998299, 'learning_rate/qwen_vl_interface': 1.2361820151072454e-06, 'learning_rate/action_model': 1.2361820151072453e-05, 'epoch': 0.3}) +[2026-10-03 13:51:08,086][starVLA.training.train_starvla][INFO] - Step 4121, Loss: {'train/loss': 0.13158172369003296, 'train/grad_norm_pre_clip': 511.7640686035156, 'timing/data': 0.00037664000410586596, 'timing/model': 0.39510807700571604, 'learning_rate/qwen_vl_interface': 1.23455432028671e-06, 'learning_rate/action_model': 1.23455432028671e-05, 'epoch': 0.3}) +[2026-10-03 13:51:08,791][starVLA.training.train_starvla][INFO] - Step 4122, Loss: {'train/loss': 7.902256766101345e-05, 'train/grad_norm_pre_clip': 0.10158884525299072, 'timing/data': 0.00036051000643055886, 'timing/model': 0.3936316659965087, 'learning_rate/qwen_vl_interface': 1.2329282760642167e-06, 'learning_rate/action_model': 1.2329282760642166e-05, 'epoch': 0.3}) +[2026-10-03 13:51:09,496][starVLA.training.train_starvla][INFO] - Step 4123, Loss: {'train/loss': 2.4383463824051432e-05, 'train/grad_norm_pre_clip': 0.020838327705860138, 'timing/data': 0.0002779200003715232, 'timing/model': 0.39423781700315885, 'learning_rate/qwen_vl_interface': 1.2313038831081716e-06, 'learning_rate/action_model': 1.2313038831081717e-05, 'epoch': 0.3}) +[2026-10-03 13:51:10,201][starVLA.training.train_starvla][INFO] - Step 4124, Loss: {'train/loss': 1.6964000678854063e-05, 'train/grad_norm_pre_clip': 0.0042924946174025536, 'timing/data': 0.0003610199928516522, 'timing/model': 0.39367474599566776, 'learning_rate/qwen_vl_interface': 1.2296811420863002e-06, 'learning_rate/action_model': 1.2296811420863003e-05, 'epoch': 0.3}) +[2026-10-03 13:51:10,907][starVLA.training.train_starvla][INFO] - Step 4125, Loss: {'train/loss': 0.025589263066649437, 'train/grad_norm_pre_clip': 123.69499969482422, 'timing/data': 0.00022087999968789518, 'timing/model': 0.395114597005886, 'learning_rate/qwen_vl_interface': 1.2280600536656504e-06, 'learning_rate/action_model': 1.2280600536656503e-05, 'epoch': 0.3}) +[2026-10-03 13:51:11,815][starVLA.training.train_starvla][INFO] - Step 4126, Loss: {'train/loss': 3.234591713408008e-05, 'train/grad_norm_pre_clip': 0.027540752664208412, 'timing/data': 0.00023970000620465726, 'timing/model': 0.39450925700657535, 'learning_rate/qwen_vl_interface': 1.2264406185125907e-06, 'learning_rate/action_model': 1.2264406185125907e-05, 'epoch': 0.3}) +[2026-10-03 13:51:12,522][starVLA.training.train_starvla][INFO] - Step 4127, Loss: {'train/loss': 1.0221678166999482e-05, 'train/grad_norm_pre_clip': 0.0026195948012173176, 'timing/data': 0.00036125000042375177, 'timing/model': 0.39442401700944174, 'learning_rate/qwen_vl_interface': 1.2248228372928098e-06, 'learning_rate/action_model': 1.2248228372928097e-05, 'epoch': 0.3}) +[2026-10-03 13:51:13,229][starVLA.training.train_starvla][INFO] - Step 4128, Loss: {'train/loss': 0.0011651929235085845, 'train/grad_norm_pre_clip': 1.4166423082351685, 'timing/data': 0.00024623000354040414, 'timing/model': 0.39488988599623553, 'learning_rate/qwen_vl_interface': 1.2232067106713174e-06, 'learning_rate/action_model': 1.2232067106713174e-05, 'epoch': 0.3}) +[2026-10-03 13:51:13,936][starVLA.training.train_starvla][INFO] - Step 4129, Loss: {'train/loss': 3.659628055174835e-05, 'train/grad_norm_pre_clip': 0.03493890538811684, 'timing/data': 0.00023450999287888408, 'timing/model': 0.3949074459960684, 'learning_rate/qwen_vl_interface': 1.22159223931244e-06, 'learning_rate/action_model': 1.2215922393124399e-05, 'epoch': 0.3}) +[2026-10-03 13:51:14,642][starVLA.training.train_starvla][INFO] - Step 4130, Loss: {'train/loss': 0.00010589447629172355, 'train/grad_norm_pre_clip': 0.07218590378761292, 'timing/data': 0.00023484000121243298, 'timing/model': 0.3947095170005923, 'learning_rate/qwen_vl_interface': 1.2199794238798266e-06, 'learning_rate/action_model': 1.2199794238798264e-05, 'epoch': 0.3}) +[2026-10-03 13:51:15,555][starVLA.training.train_starvla][INFO] - Step 4131, Loss: {'train/loss': 3.213480158592574e-05, 'train/grad_norm_pre_clip': 0.01033538207411766, 'timing/data': 0.0003675699990708381, 'timing/model': 0.394344557003933, 'learning_rate/qwen_vl_interface': 1.2183682650364447e-06, 'learning_rate/action_model': 1.2183682650364447e-05, 'epoch': 0.3}) +[2026-10-03 13:51:16,261][starVLA.training.train_starvla][INFO] - Step 4132, Loss: {'train/loss': 1.9102284568361938e-05, 'train/grad_norm_pre_clip': 0.004125660285353661, 'timing/data': 0.00023171999782789499, 'timing/model': 0.3940049169905251, 'learning_rate/qwen_vl_interface': 1.216758763444581e-06, 'learning_rate/action_model': 1.2167587634445809e-05, 'epoch': 0.3}) +[2026-10-03 13:51:16,967][starVLA.training.train_starvla][INFO] - Step 4133, Loss: {'train/loss': 3.044059849344194e-05, 'train/grad_norm_pre_clip': 0.014529557898640633, 'timing/data': 0.00029350000841077417, 'timing/model': 0.395399097003974, 'learning_rate/qwen_vl_interface': 1.215150919765841e-06, 'learning_rate/action_model': 1.215150919765841e-05, 'epoch': 0.3}) +[2026-10-03 13:51:17,673][starVLA.training.train_starvla][INFO] - Step 4134, Loss: {'train/loss': 2.6006084226537496e-05, 'train/grad_norm_pre_clip': 0.009298302233219147, 'timing/data': 0.0002444400015519932, 'timing/model': 0.3943916169955628, 'learning_rate/qwen_vl_interface': 1.213544734661148e-06, 'learning_rate/action_model': 1.2135447346611478e-05, 'epoch': 0.3}) +[2026-10-03 13:51:18,379][starVLA.training.train_starvla][INFO] - Step 4135, Loss: {'train/loss': 9.029869033838622e-06, 'train/grad_norm_pre_clip': 0.0032855260651558638, 'timing/data': 0.00036838999949395657, 'timing/model': 0.39369849699141923, 'learning_rate/qwen_vl_interface': 1.2119402087907448e-06, 'learning_rate/action_model': 1.2119402087907448e-05, 'epoch': 0.3}) +[2026-10-03 13:51:19,290][starVLA.training.train_starvla][INFO] - Step 4136, Loss: {'train/loss': 3.7923218769719824e-06, 'train/grad_norm_pre_clip': 0.0009505103225819767, 'timing/data': 0.00023744998907204717, 'timing/model': 0.599926504000905, 'learning_rate/qwen_vl_interface': 1.2103373428141903e-06, 'learning_rate/action_model': 1.2103373428141903e-05, 'epoch': 0.3}) +[2026-10-03 13:51:19,996][starVLA.training.train_starvla][INFO] - Step 4137, Loss: {'train/loss': 4.224445092404494e-06, 'train/grad_norm_pre_clip': 0.0013155172346159816, 'timing/data': 0.0002373000024817884, 'timing/model': 0.3948048860038398, 'learning_rate/qwen_vl_interface': 1.2087361373903633e-06, 'learning_rate/action_model': 1.2087361373903633e-05, 'epoch': 0.3}) +[2026-10-03 13:51:20,702][starVLA.training.train_starvla][INFO] - Step 4138, Loss: {'train/loss': 2.3609209165442735e-05, 'train/grad_norm_pre_clip': 0.012401727959513664, 'timing/data': 0.0002470699982950464, 'timing/model': 0.3942800960066961, 'learning_rate/qwen_vl_interface': 1.2071365931774584e-06, 'learning_rate/action_model': 1.2071365931774584e-05, 'epoch': 0.3}) +[2026-10-03 13:51:21,406][starVLA.training.train_starvla][INFO] - Step 4139, Loss: {'train/loss': 3.5688049138116185e-06, 'train/grad_norm_pre_clip': 0.001146897324360907, 'timing/data': 0.00034019000304397196, 'timing/model': 0.3933545170002617, 'learning_rate/qwen_vl_interface': 1.2055387108329885e-06, 'learning_rate/action_model': 1.2055387108329886e-05, 'epoch': 0.3}) +[2026-10-03 13:51:22,113][starVLA.training.train_starvla][INFO] - Step 4140, Loss: {'train/loss': 0.0027006426826119423, 'train/grad_norm_pre_clip': 9.793963432312012, 'timing/data': 0.00038701000448781997, 'timing/model': 0.3947778779984219, 'learning_rate/qwen_vl_interface': 1.2039424910137807e-06, 'learning_rate/action_model': 1.2039424910137806e-05, 'epoch': 0.3}) +[2026-10-03 13:51:22,820][starVLA.training.train_starvla][INFO] - Step 4141, Loss: {'train/loss': 0.6093205213546753, 'train/grad_norm_pre_clip': 34.785823822021484, 'timing/data': 0.00023461000819224864, 'timing/model': 0.3951456470094854, 'learning_rate/qwen_vl_interface': 1.2023479343759822e-06, 'learning_rate/action_model': 1.2023479343759823e-05, 'epoch': 0.3}) +[2026-10-03 13:51:23,735][starVLA.training.train_starvla][INFO] - Step 4142, Loss: {'train/loss': 7.45776833355194e-06, 'train/grad_norm_pre_clip': 0.0019225151045247912, 'timing/data': 0.00024230999406427145, 'timing/model': 0.39530211700184736, 'learning_rate/qwen_vl_interface': 1.2007550415750546e-06, 'learning_rate/action_model': 1.2007550415750545e-05, 'epoch': 0.3}) +[2026-10-03 13:51:24,440][starVLA.training.train_starvla][INFO] - Step 4143, Loss: {'train/loss': 1.4305093145594583e-06, 'train/grad_norm_pre_clip': 0.0002669061650522053, 'timing/data': 0.0002481299889041111, 'timing/model': 0.39388501799840014, 'learning_rate/qwen_vl_interface': 1.1991638132657759e-06, 'learning_rate/action_model': 1.199163813265776e-05, 'epoch': 0.3}) +[2026-10-03 13:51:25,145][starVLA.training.train_starvla][INFO] - Step 4144, Loss: {'train/loss': 5.3910280257696286e-05, 'train/grad_norm_pre_clip': 0.0344577394425869, 'timing/data': 0.000288180002826266, 'timing/model': 0.39377986700856127, 'learning_rate/qwen_vl_interface': 1.1975742501022406e-06, 'learning_rate/action_model': 1.1975742501022405e-05, 'epoch': 0.3}) +[2026-10-03 13:51:25,851][starVLA.training.train_starvla][INFO] - Step 4145, Loss: {'train/loss': 0.5922747254371643, 'train/grad_norm_pre_clip': 42.310089111328125, 'timing/data': 0.00027841998962685466, 'timing/model': 0.3945900669932598, 'learning_rate/qwen_vl_interface': 1.1959863527378569e-06, 'learning_rate/action_model': 1.1959863527378567e-05, 'epoch': 0.3}) +[2026-10-03 13:51:26,555][starVLA.training.train_starvla][INFO] - Step 4146, Loss: {'train/loss': 3.0398202852666145e-06, 'train/grad_norm_pre_clip': 0.0010505638783797622, 'timing/data': 0.00036566000198945403, 'timing/model': 0.39312837799661793, 'learning_rate/qwen_vl_interface': 1.1944001218253505e-06, 'learning_rate/action_model': 1.1944001218253503e-05, 'epoch': 0.3}) +[2026-10-03 13:51:27,462][starVLA.training.train_starvla][INFO] - Step 4147, Loss: {'train/loss': 0.010015338659286499, 'train/grad_norm_pre_clip': 19.088895797729492, 'timing/data': 0.0002411399909760803, 'timing/model': 0.3940195470058825, 'learning_rate/qwen_vl_interface': 1.1928155580167608e-06, 'learning_rate/action_model': 1.1928155580167608e-05, 'epoch': 0.3}) +[2026-10-03 13:51:28,169][starVLA.training.train_starvla][INFO] - Step 4148, Loss: {'train/loss': 0.43360188603401184, 'train/grad_norm_pre_clip': 26.216482162475586, 'timing/data': 0.0003441199951339513, 'timing/model': 0.3947611569892615, 'learning_rate/qwen_vl_interface': 1.1912326619634412e-06, 'learning_rate/action_model': 1.1912326619634412e-05, 'epoch': 0.3}) +[2026-10-03 13:51:28,875][starVLA.training.train_starvla][INFO] - Step 4149, Loss: {'train/loss': 1.4766430467716418e-05, 'train/grad_norm_pre_clip': 0.006483457051217556, 'timing/data': 0.00028828899667132646, 'timing/model': 0.3945293079887051, 'learning_rate/qwen_vl_interface': 1.1896514343160632e-06, 'learning_rate/action_model': 1.189651434316063e-05, 'epoch': 0.3}) +[2026-10-03 13:51:29,580][starVLA.training.train_starvla][INFO] - Step 4150, Loss: {'train/loss': 2.0563556972774677e-06, 'train/grad_norm_pre_clip': 0.0004849074175581336, 'timing/data': 0.0003154099977109581, 'timing/model': 0.3937845879991073, 'learning_rate/qwen_vl_interface': 1.1880718757246062e-06, 'learning_rate/action_model': 1.1880718757246063e-05, 'epoch': 0.3}) +[2026-10-03 13:51:30,290][starVLA.training.train_starvla][INFO] - Step 4151, Loss: {'train/loss': 5.096164386486635e-06, 'train/grad_norm_pre_clip': 0.0018636825261637568, 'timing/data': 0.000254550002864562, 'timing/model': 0.3958157480083173, 'learning_rate/qwen_vl_interface': 1.186493986838369e-06, 'learning_rate/action_model': 1.186493986838369e-05, 'epoch': 0.3}) +[2026-10-03 13:51:31,213][starVLA.training.train_starvla][INFO] - Step 4152, Loss: {'train/loss': 7.681427632633131e-06, 'train/grad_norm_pre_clip': 0.0021641026251018047, 'timing/data': 0.0002935700031230226, 'timing/model': 0.3938969869923312, 'learning_rate/qwen_vl_interface': 1.1849177683059612e-06, 'learning_rate/action_model': 1.1849177683059612e-05, 'epoch': 0.3}) +[2026-10-03 13:51:31,919][starVLA.training.train_starvla][INFO] - Step 4153, Loss: {'train/loss': 4.462865035748109e-06, 'train/grad_norm_pre_clip': 0.002057042671367526, 'timing/data': 0.00021934999676886946, 'timing/model': 0.3948922369891079, 'learning_rate/qwen_vl_interface': 1.1833432207753075e-06, 'learning_rate/action_model': 1.1833432207753076e-05, 'epoch': 0.3}) +[2026-10-03 13:51:32,623][starVLA.training.train_starvla][INFO] - Step 4154, Loss: {'train/loss': 2.4339988158317283e-05, 'train/grad_norm_pre_clip': 0.011779122054576874, 'timing/data': 0.00037930000689812005, 'timing/model': 0.39277073700213805, 'learning_rate/qwen_vl_interface': 1.1817703448936448e-06, 'learning_rate/action_model': 1.181770344893645e-05, 'epoch': 0.3}) +[2026-10-03 13:51:33,328][starVLA.training.train_starvla][INFO] - Step 4155, Loss: {'train/loss': 7.2864609137468506e-06, 'train/grad_norm_pre_clip': 0.002647936111316085, 'timing/data': 0.00024171000404749066, 'timing/model': 0.39305130700813606, 'learning_rate/qwen_vl_interface': 1.1801991413075228e-06, 'learning_rate/action_model': 1.1801991413075227e-05, 'epoch': 0.3}) +[2026-10-03 13:51:34,032][starVLA.training.train_starvla][INFO] - Step 4156, Loss: {'train/loss': 0.00011418056965339929, 'train/grad_norm_pre_clip': 0.06493380665779114, 'timing/data': 0.00030453001090791076, 'timing/model': 0.39361259699217044, 'learning_rate/qwen_vl_interface': 1.1786296106628038e-06, 'learning_rate/action_model': 1.1786296106628036e-05, 'epoch': 0.3}) +[2026-10-03 13:51:34,954][starVLA.training.train_starvla][INFO] - Step 4157, Loss: {'train/loss': 0.40247395634651184, 'train/grad_norm_pre_clip': 223.43865966796875, 'timing/data': 0.00023125999723561108, 'timing/model': 0.6099196129944175, 'learning_rate/qwen_vl_interface': 1.1770617536046633e-06, 'learning_rate/action_model': 1.1770617536046632e-05, 'epoch': 0.3}) +[2026-10-03 13:51:35,662][starVLA.training.train_starvla][INFO] - Step 4158, Loss: {'train/loss': 9.136438893619925e-05, 'train/grad_norm_pre_clip': 0.048598162829875946, 'timing/data': 0.0002341499930480495, 'timing/model': 0.3963326180091826, 'learning_rate/qwen_vl_interface': 1.1754955707775873e-06, 'learning_rate/action_model': 1.1754955707775872e-05, 'epoch': 0.3}) +[2026-10-03 13:51:36,372][starVLA.training.train_starvla][INFO] - Step 4159, Loss: {'train/loss': 0.006214705295860767, 'train/grad_norm_pre_clip': 18.912561416625977, 'timing/data': 0.00038373999996110797, 'timing/model': 0.39541448700765613, 'learning_rate/qwen_vl_interface': 1.1739310628253753e-06, 'learning_rate/action_model': 1.1739310628253753e-05, 'epoch': 0.3}) +[2026-10-03 13:51:37,077][starVLA.training.train_starvla][INFO] - Step 4160, Loss: {'train/loss': 4.3560539779718965e-05, 'train/grad_norm_pre_clip': 0.025852948427200317, 'timing/data': 0.00033214999712072313, 'timing/model': 0.3935239080019528, 'learning_rate/qwen_vl_interface': 1.172368230391136e-06, 'learning_rate/action_model': 1.1723682303911359e-05, 'epoch': 0.3}) +[2026-10-03 13:51:37,785][starVLA.training.train_starvla][INFO] - Step 4161, Loss: {'train/loss': 8.383257227251306e-05, 'train/grad_norm_pre_clip': 0.029224487021565437, 'timing/data': 0.00025053998979274184, 'timing/model': 0.39616358700732235, 'learning_rate/qwen_vl_interface': 1.1708070741172928e-06, 'learning_rate/action_model': 1.1708070741172928e-05, 'epoch': 0.3}) +[2026-10-03 13:51:38,706][starVLA.training.train_starvla][INFO] - Step 4162, Loss: {'train/loss': 0.05335772782564163, 'train/grad_norm_pre_clip': 296.9712829589844, 'timing/data': 0.00021374000061769038, 'timing/model': 0.6099874830106273, 'learning_rate/qwen_vl_interface': 1.1692475946455755e-06, 'learning_rate/action_model': 1.1692475946455755e-05, 'epoch': 0.3}) +[2026-10-03 13:51:39,409][starVLA.training.train_starvla][INFO] - Step 4163, Loss: {'train/loss': 2.7027339456253685e-05, 'train/grad_norm_pre_clip': 0.025995735079050064, 'timing/data': 0.00026560999685898423, 'timing/model': 0.39230563800083473, 'learning_rate/qwen_vl_interface': 1.1676897926170289e-06, 'learning_rate/action_model': 1.1676897926170289e-05, 'epoch': 0.3}) +[2026-10-03 13:51:40,114][starVLA.training.train_starvla][INFO] - Step 4164, Loss: {'train/loss': 8.643587352707982e-05, 'train/grad_norm_pre_clip': 0.018903078511357307, 'timing/data': 0.0002810489968396723, 'timing/model': 0.393927948010969, 'learning_rate/qwen_vl_interface': 1.166133668672006e-06, 'learning_rate/action_model': 1.166133668672006e-05, 'epoch': 0.3}) +[2026-10-03 13:51:40,822][starVLA.training.train_starvla][INFO] - Step 4165, Loss: {'train/loss': 0.00044236559187993407, 'train/grad_norm_pre_clip': 0.24725358188152313, 'timing/data': 0.0002723500074353069, 'timing/model': 0.3961740770027973, 'learning_rate/qwen_vl_interface': 1.1645792234501706e-06, 'learning_rate/action_model': 1.1645792234501706e-05, 'epoch': 0.3}) +[2026-10-03 13:51:41,528][starVLA.training.train_starvla][INFO] - Step 4166, Loss: {'train/loss': 5.9326881455490366e-05, 'train/grad_norm_pre_clip': 0.028486592695116997, 'timing/data': 0.00028526000096462667, 'timing/model': 0.39251139799307566, 'learning_rate/qwen_vl_interface': 1.1630264575904972e-06, 'learning_rate/action_model': 1.1630264575904973e-05, 'epoch': 0.3}) +[2026-10-03 13:51:42,233][starVLA.training.train_starvla][INFO] - Step 4167, Loss: {'train/loss': 2.5008559532579966e-05, 'train/grad_norm_pre_clip': 0.009316555224359035, 'timing/data': 0.0002802700037136674, 'timing/model': 0.3944383280031616, 'learning_rate/qwen_vl_interface': 1.1614753717312683e-06, 'learning_rate/action_model': 1.1614753717312681e-05, 'epoch': 0.3}) +[2026-10-03 13:51:43,155][starVLA.training.train_starvla][INFO] - Step 4168, Loss: {'train/loss': 8.716796583030373e-05, 'train/grad_norm_pre_clip': 0.23850014805793762, 'timing/data': 0.0004040200001327321, 'timing/model': 0.3938021369976923, 'learning_rate/qwen_vl_interface': 1.1599259665100767e-06, 'learning_rate/action_model': 1.1599259665100767e-05, 'epoch': 0.31}) +[2026-10-03 13:51:43,864][starVLA.training.train_starvla][INFO] - Step 4169, Loss: {'train/loss': 5.168449933989905e-05, 'train/grad_norm_pre_clip': 0.01561778411269188, 'timing/data': 0.0003197900077793747, 'timing/model': 0.39565493699046783, 'learning_rate/qwen_vl_interface': 1.1583782425638252e-06, 'learning_rate/action_model': 1.158378242563825e-05, 'epoch': 0.31}) +[2026-10-03 13:51:44,573][starVLA.training.train_starvla][INFO] - Step 4170, Loss: {'train/loss': 8.95633056643419e-05, 'train/grad_norm_pre_clip': 0.047981709241867065, 'timing/data': 0.0002541699941502884, 'timing/model': 0.3932306180067826, 'learning_rate/qwen_vl_interface': 1.1568322005287237e-06, 'learning_rate/action_model': 1.1568322005287235e-05, 'epoch': 0.31}) +[2026-10-03 13:51:45,277][starVLA.training.train_starvla][INFO] - Step 4171, Loss: {'train/loss': 5.4033771448303014e-05, 'train/grad_norm_pre_clip': 0.02068471722304821, 'timing/data': 0.0004238099936628714, 'timing/model': 0.3933584279875504, 'learning_rate/qwen_vl_interface': 1.1552878410402923e-06, 'learning_rate/action_model': 1.1552878410402924e-05, 'epoch': 0.31}) +[2026-10-03 13:51:45,981][starVLA.training.train_starvla][INFO] - Step 4172, Loss: {'train/loss': 0.4652521014213562, 'train/grad_norm_pre_clip': 280.8780822753906, 'timing/data': 0.00021980000019539148, 'timing/model': 0.3933924769953592, 'learning_rate/qwen_vl_interface': 1.1537451647333594e-06, 'learning_rate/action_model': 1.1537451647333593e-05, 'epoch': 0.31}) +[2026-10-03 13:51:46,888][starVLA.training.train_starvla][INFO] - Step 4173, Loss: {'train/loss': 4.6863924580975436e-06, 'train/grad_norm_pre_clip': 0.001888066646642983, 'timing/data': 0.0002955799864139408, 'timing/model': 0.5953650860028574, 'learning_rate/qwen_vl_interface': 1.1522041722420594e-06, 'learning_rate/action_model': 1.1522041722420594e-05, 'epoch': 0.31}) +[2026-10-03 13:51:47,592][starVLA.training.train_starvla][INFO] - Step 4174, Loss: {'train/loss': 3.6889148759655654e-05, 'train/grad_norm_pre_clip': 0.009019189514219761, 'timing/data': 0.0003336200024932623, 'timing/model': 0.3931614580069436, 'learning_rate/qwen_vl_interface': 1.1506648641998365e-06, 'learning_rate/action_model': 1.1506648641998365e-05, 'epoch': 0.31}) +[2026-10-03 13:51:48,295][starVLA.training.train_starvla][INFO] - Step 4175, Loss: {'train/loss': 5.7741572163649835e-06, 'train/grad_norm_pre_clip': 0.0032149460166692734, 'timing/data': 0.0003143799985991791, 'timing/model': 0.3928354879899416, 'learning_rate/qwen_vl_interface': 1.1491272412394433e-06, 'learning_rate/action_model': 1.1491272412394433e-05, 'epoch': 0.31}) +[2026-10-03 13:51:49,001][starVLA.training.train_starvla][INFO] - Step 4176, Loss: {'train/loss': 0.0004594953206833452, 'train/grad_norm_pre_clip': 0.5429326295852661, 'timing/data': 0.00041059999784920365, 'timing/model': 0.3937616280018119, 'learning_rate/qwen_vl_interface': 1.1475913039929373e-06, 'learning_rate/action_model': 1.1475913039929372e-05, 'epoch': 0.31}) +[2026-10-03 13:51:49,705][starVLA.training.train_starvla][INFO] - Step 4177, Loss: {'train/loss': 7.91985439718701e-06, 'train/grad_norm_pre_clip': 0.012974407523870468, 'timing/data': 0.00023017999774310738, 'timing/model': 0.3940511270047864, 'learning_rate/qwen_vl_interface': 1.1460570530916844e-06, 'learning_rate/action_model': 1.1460570530916844e-05, 'epoch': 0.31}) +[2026-10-03 13:51:50,409][starVLA.training.train_starvla][INFO] - Step 4178, Loss: {'train/loss': 7.032475696178153e-05, 'train/grad_norm_pre_clip': 0.07189510017633438, 'timing/data': 0.00030169999809004366, 'timing/model': 0.39432807799312286, 'learning_rate/qwen_vl_interface': 1.1445244891663581e-06, 'learning_rate/action_model': 1.1445244891663581e-05, 'epoch': 0.31}) +[2026-10-03 13:51:51,315][starVLA.training.train_starvla][INFO] - Step 4179, Loss: {'train/loss': 3.6814657505601645e-05, 'train/grad_norm_pre_clip': 0.009793926030397415, 'timing/data': 0.00037367000186350197, 'timing/model': 0.39402454800438136, 'learning_rate/qwen_vl_interface': 1.1429936128469366e-06, 'learning_rate/action_model': 1.1429936128469366e-05, 'epoch': 0.31}) +[2026-10-03 13:51:52,019][starVLA.training.train_starvla][INFO] - Step 4180, Loss: {'train/loss': 0.00012453316594474018, 'train/grad_norm_pre_clip': 0.056899458169937134, 'timing/data': 0.0003090599930146709, 'timing/model': 0.3933853079943219, 'learning_rate/qwen_vl_interface': 1.1414644247627056e-06, 'learning_rate/action_model': 1.1414644247627058e-05, 'epoch': 0.31}) +[2026-10-03 13:51:52,726][starVLA.training.train_starvla][INFO] - Step 4181, Loss: {'train/loss': 6.68310713081155e-06, 'train/grad_norm_pre_clip': 0.002496489090844989, 'timing/data': 0.0002511000056983903, 'timing/model': 0.3945205880008871, 'learning_rate/qwen_vl_interface': 1.1399369255422568e-06, 'learning_rate/action_model': 1.1399369255422568e-05, 'epoch': 0.31}) +[2026-10-03 13:51:53,432][starVLA.training.train_starvla][INFO] - Step 4182, Loss: {'train/loss': 9.998990572057664e-05, 'train/grad_norm_pre_clip': 0.0758984237909317, 'timing/data': 0.00025378999998793006, 'timing/model': 0.39410276799753774, 'learning_rate/qwen_vl_interface': 1.138411115813488e-06, 'learning_rate/action_model': 1.138411115813488e-05, 'epoch': 0.31}) +[2026-10-03 13:51:54,139][starVLA.training.train_starvla][INFO] - Step 4183, Loss: {'train/loss': 0.4963708817958832, 'train/grad_norm_pre_clip': 57.74864959716797, 'timing/data': 0.00037667001015506685, 'timing/model': 0.39457283800584264, 'learning_rate/qwen_vl_interface': 1.1368869962036006e-06, 'learning_rate/action_model': 1.1368869962036005e-05, 'epoch': 0.31}) +[2026-10-03 13:51:55,049][starVLA.training.train_starvla][INFO] - Step 4184, Loss: {'train/loss': 1.1838633326988202e-05, 'train/grad_norm_pre_clip': 0.004217431880533695, 'timing/data': 0.00034097000025212765, 'timing/model': 0.3927056880056625, 'learning_rate/qwen_vl_interface': 1.1353645673391025e-06, 'learning_rate/action_model': 1.1353645673391026e-05, 'epoch': 0.31}) +[2026-10-03 13:51:55,754][starVLA.training.train_starvla][INFO] - Step 4185, Loss: {'train/loss': 0.00015726372657809407, 'train/grad_norm_pre_clip': 0.0754542127251625, 'timing/data': 0.0002597900020191446, 'timing/model': 0.3945562379958574, 'learning_rate/qwen_vl_interface': 1.1338438298458078e-06, 'learning_rate/action_model': 1.1338438298458078e-05, 'epoch': 0.31}) +[2026-10-03 13:51:56,459][starVLA.training.train_starvla][INFO] - Step 4186, Loss: {'train/loss': 0.35937798023223877, 'train/grad_norm_pre_clip': 164.31602478027344, 'timing/data': 0.0002619600127218291, 'timing/model': 0.3941161179973278, 'learning_rate/qwen_vl_interface': 1.1323247843488345e-06, 'learning_rate/action_model': 1.1323247843488344e-05, 'epoch': 0.31}) +[2026-10-03 13:51:57,162][starVLA.training.train_starvla][INFO] - Step 4187, Loss: {'train/loss': 0.4951121211051941, 'train/grad_norm_pre_clip': 100.60440063476562, 'timing/data': 0.000305368987028487, 'timing/model': 0.39294184900063556, 'learning_rate/qwen_vl_interface': 1.130807431472603e-06, 'learning_rate/action_model': 1.130807431472603e-05, 'epoch': 0.31}) +[2026-10-03 13:51:57,868][starVLA.training.train_starvla][INFO] - Step 4188, Loss: {'train/loss': 0.001207430032081902, 'train/grad_norm_pre_clip': 2.368699789047241, 'timing/data': 0.0002405400009592995, 'timing/model': 0.393892997992225, 'learning_rate/qwen_vl_interface': 1.1292917718408416e-06, 'learning_rate/action_model': 1.1292917718408416e-05, 'epoch': 0.31}) +[2026-10-03 13:51:58,783][starVLA.training.train_starvla][INFO] - Step 4189, Loss: {'train/loss': 7.600433309562504e-05, 'train/grad_norm_pre_clip': 0.028211543336510658, 'timing/data': 0.0003137500025331974, 'timing/model': 0.6044367050053552, 'learning_rate/qwen_vl_interface': 1.12777780607658e-06, 'learning_rate/action_model': 1.12777780607658e-05, 'epoch': 0.31}) +[2026-10-03 13:51:59,488][starVLA.training.train_starvla][INFO] - Step 4190, Loss: {'train/loss': 9.804150613490492e-05, 'train/grad_norm_pre_clip': 0.03992724418640137, 'timing/data': 0.0002999599964823574, 'timing/model': 0.39379533799365163, 'learning_rate/qwen_vl_interface': 1.1262655348021523e-06, 'learning_rate/action_model': 1.1262655348021523e-05, 'epoch': 0.31}) +[2026-10-03 13:52:00,194][starVLA.training.train_starvla][INFO] - Step 4191, Loss: {'train/loss': 1.4990020645200275e-05, 'train/grad_norm_pre_clip': 0.004641961306333542, 'timing/data': 0.0002676900039659813, 'timing/model': 0.39425430800474714, 'learning_rate/qwen_vl_interface': 1.1247549586391966e-06, 'learning_rate/action_model': 1.1247549586391964e-05, 'epoch': 0.31}) +[2026-10-03 13:52:00,900][starVLA.training.train_starvla][INFO] - Step 4192, Loss: {'train/loss': 9.348975436296314e-05, 'train/grad_norm_pre_clip': 0.11311681568622589, 'timing/data': 0.0003224300016881898, 'timing/model': 0.3940009179932531, 'learning_rate/qwen_vl_interface': 1.1232460782086534e-06, 'learning_rate/action_model': 1.1232460782086534e-05, 'epoch': 0.31}) +[2026-10-03 13:52:01,606][starVLA.training.train_starvla][INFO] - Step 4193, Loss: {'train/loss': 0.404598593711853, 'train/grad_norm_pre_clip': 79.26812744140625, 'timing/data': 0.00024992000544443727, 'timing/model': 0.39496804800000973, 'learning_rate/qwen_vl_interface': 1.1217388941307658e-06, 'learning_rate/action_model': 1.1217388941307658e-05, 'epoch': 0.31}) +[2026-10-03 13:52:02,313][starVLA.training.train_starvla][INFO] - Step 4194, Loss: {'train/loss': 7.249357167893322e-06, 'train/grad_norm_pre_clip': 0.0020604117307811975, 'timing/data': 0.00031392999517265707, 'timing/model': 0.394405518003623, 'learning_rate/qwen_vl_interface': 1.1202334070250807e-06, 'learning_rate/action_model': 1.1202334070250806e-05, 'epoch': 0.31}) +[2026-10-03 13:52:03,216][starVLA.training.train_starvla][INFO] - Step 4195, Loss: {'train/loss': 1.5258079656632617e-05, 'train/grad_norm_pre_clip': 0.0120400320738554, 'timing/data': 0.0003401700087124482, 'timing/model': 0.3940763979917392, 'learning_rate/qwen_vl_interface': 1.1187296175104472e-06, 'learning_rate/action_model': 1.1187296175104471e-05, 'epoch': 0.31}) +[2026-10-03 13:52:03,921][starVLA.training.train_starvla][INFO] - Step 4196, Loss: {'train/loss': 1.785703716450371e-05, 'train/grad_norm_pre_clip': 0.004762272350490093, 'timing/data': 0.0002920700062531978, 'timing/model': 0.39397410801029764, 'learning_rate/qwen_vl_interface': 1.1172275262050158e-06, 'learning_rate/action_model': 1.1172275262050159e-05, 'epoch': 0.31}) +[2026-10-03 13:52:04,627][starVLA.training.train_starvla][INFO] - Step 4197, Loss: {'train/loss': 6.318801752058789e-05, 'train/grad_norm_pre_clip': 0.03443321958184242, 'timing/data': 0.0003346500016050413, 'timing/model': 0.39483348801149987, 'learning_rate/qwen_vl_interface': 1.1157271337262401e-06, 'learning_rate/action_model': 1.11572713372624e-05, 'epoch': 0.31}) +[2026-10-03 13:52:05,332][starVLA.training.train_starvla][INFO] - Step 4198, Loss: {'train/loss': 6.160072371130809e-05, 'train/grad_norm_pre_clip': 0.021014023572206497, 'timing/data': 0.0003853999951388687, 'timing/model': 0.3934391679940745, 'learning_rate/qwen_vl_interface': 1.1142284406908751e-06, 'learning_rate/action_model': 1.114228440690875e-05, 'epoch': 0.31}) +[2026-10-03 13:52:06,037][starVLA.training.train_starvla][INFO] - Step 4199, Loss: {'train/loss': 0.0010287631303071976, 'train/grad_norm_pre_clip': 6.660747528076172, 'timing/data': 0.0002244400093331933, 'timing/model': 0.3934702980041038, 'learning_rate/qwen_vl_interface': 1.1127314477149758e-06, 'learning_rate/action_model': 1.1127314477149759e-05, 'epoch': 0.31}) +[2026-10-03 13:52:06,948][starVLA.training.train_starvla][INFO] - Step 4200, Loss: {'train/loss': 0.00011811578588094562, 'train/grad_norm_pre_clip': 0.04252253472805023, 'timing/data': 0.0003385900054126978, 'timing/model': 0.39332030799414497, 'learning_rate/qwen_vl_interface': 1.1112361554139003e-06, 'learning_rate/action_model': 1.1112361554139004e-05, 'epoch': 0.31}) +[2026-10-03 13:52:07,656][starVLA.training.train_starvla][INFO] - Step 4201, Loss: {'train/loss': 2.2065712983021513e-05, 'train/grad_norm_pre_clip': 0.005462080240249634, 'timing/data': 0.00032713999098632485, 'timing/model': 0.3943898180004908, 'learning_rate/qwen_vl_interface': 1.1097425644023065e-06, 'learning_rate/action_model': 1.1097425644023065e-05, 'epoch': 0.31}) +[2026-10-03 13:52:08,360][starVLA.training.train_starvla][INFO] - Step 4202, Loss: {'train/loss': 6.779899194953032e-06, 'train/grad_norm_pre_clip': 0.002682252088561654, 'timing/data': 0.00035197999386582524, 'timing/model': 0.3933046689926414, 'learning_rate/qwen_vl_interface': 1.1082506752941529e-06, 'learning_rate/action_model': 1.1082506752941528e-05, 'epoch': 0.31}) +[2026-10-03 13:52:09,065][starVLA.training.train_starvla][INFO] - Step 4203, Loss: {'train/loss': 0.00021163317433092743, 'train/grad_norm_pre_clip': 0.13850772380828857, 'timing/data': 0.0002745400124695152, 'timing/model': 0.39377238899760414, 'learning_rate/qwen_vl_interface': 1.1067604887027004e-06, 'learning_rate/action_model': 1.1067604887027004e-05, 'epoch': 0.31}) +[2026-10-03 13:52:09,772][starVLA.training.train_starvla][INFO] - Step 4204, Loss: {'train/loss': 0.000122017678222619, 'train/grad_norm_pre_clip': 0.07426881045103073, 'timing/data': 0.00047292999806813896, 'timing/model': 0.39522260800004005, 'learning_rate/qwen_vl_interface': 1.1052720052405066e-06, 'learning_rate/action_model': 1.1052720052405064e-05, 'epoch': 0.31}) +[2026-10-03 13:52:10,692][starVLA.training.train_starvla][INFO] - Step 4205, Loss: {'train/loss': 1.4103286048339214e-05, 'train/grad_norm_pre_clip': 0.004492634441703558, 'timing/data': 0.00031800000579096377, 'timing/model': 0.3961605380027322, 'learning_rate/qwen_vl_interface': 1.1037852255194309e-06, 'learning_rate/action_model': 1.1037852255194308e-05, 'epoch': 0.31}) +[2026-10-03 13:52:11,400][starVLA.training.train_starvla][INFO] - Step 4206, Loss: {'train/loss': 3.506912980810739e-05, 'train/grad_norm_pre_clip': 0.03905298933386803, 'timing/data': 0.0004668199981097132, 'timing/model': 0.3945091789937578, 'learning_rate/qwen_vl_interface': 1.1023001501506329e-06, 'learning_rate/action_model': 1.102300150150633e-05, 'epoch': 0.31}) +[2026-10-03 13:52:12,106][starVLA.training.train_starvla][INFO] - Step 4207, Loss: {'train/loss': 0.4419037103652954, 'train/grad_norm_pre_clip': 60.302127838134766, 'timing/data': 0.00023940000392030925, 'timing/model': 0.3942291190032847, 'learning_rate/qwen_vl_interface': 1.1008167797445717e-06, 'learning_rate/action_model': 1.1008167797445718e-05, 'epoch': 0.31}) +[2026-10-03 13:52:12,813][starVLA.training.train_starvla][INFO] - Step 4208, Loss: {'train/loss': 0.7075461149215698, 'train/grad_norm_pre_clip': 53.264888763427734, 'timing/data': 0.00026274999254383147, 'timing/model': 0.3946223479870241, 'learning_rate/qwen_vl_interface': 1.099335114911003e-06, 'learning_rate/action_model': 1.0993351149110031e-05, 'epoch': 0.31}) +[2026-10-03 13:52:13,520][starVLA.training.train_starvla][INFO] - Step 4209, Loss: {'train/loss': 6.006649346090853e-05, 'train/grad_norm_pre_clip': 0.054789334535598755, 'timing/data': 0.0002926499873865396, 'timing/model': 0.39574381901184097, 'learning_rate/qwen_vl_interface': 1.0978551562589846e-06, 'learning_rate/action_model': 1.0978551562589845e-05, 'epoch': 0.31}) +[2026-10-03 13:52:14,442][starVLA.training.train_starvla][INFO] - Step 4210, Loss: {'train/loss': 0.0016461991472169757, 'train/grad_norm_pre_clip': 7.5947065353393555, 'timing/data': 0.0004168390005361289, 'timing/model': 0.6088439449958969, 'learning_rate/qwen_vl_interface': 1.0963769043968708e-06, 'learning_rate/action_model': 1.0963769043968708e-05, 'epoch': 0.31}) +[2026-10-03 13:52:15,148][starVLA.training.train_starvla][INFO] - Step 4211, Loss: {'train/loss': 0.012787739746272564, 'train/grad_norm_pre_clip': 67.11380767822266, 'timing/data': 0.0002338399935979396, 'timing/model': 0.39491694899334107, 'learning_rate/qwen_vl_interface': 1.0949003599323158e-06, 'learning_rate/action_model': 1.0949003599323157e-05, 'epoch': 0.31}) +[2026-10-03 13:52:15,859][starVLA.training.train_starvla][INFO] - Step 4212, Loss: {'train/loss': 1.5742638424853794e-05, 'train/grad_norm_pre_clip': 0.004859357140958309, 'timing/data': 0.00034874000994022936, 'timing/model': 0.39859081800386775, 'learning_rate/qwen_vl_interface': 1.0934255234722708e-06, 'learning_rate/action_model': 1.0934255234722706e-05, 'epoch': 0.31}) +[2026-10-03 13:52:16,574][starVLA.training.train_starvla][INFO] - Step 4213, Loss: {'train/loss': 2.0874878828180954e-05, 'train/grad_norm_pre_clip': 0.005117144901305437, 'timing/data': 0.00029997999081388116, 'timing/model': 0.3984683680027956, 'learning_rate/qwen_vl_interface': 1.0919523956229858e-06, 'learning_rate/action_model': 1.0919523956229857e-05, 'epoch': 0.31}) +[2026-10-03 13:52:17,285][starVLA.training.train_starvla][INFO] - Step 4214, Loss: {'train/loss': 0.043391238898038864, 'train/grad_norm_pre_clip': 153.83074951171875, 'timing/data': 0.000367450003977865, 'timing/model': 0.3954900280077709, 'learning_rate/qwen_vl_interface': 1.0904809769900075e-06, 'learning_rate/action_model': 1.0904809769900073e-05, 'epoch': 0.31}) +[2026-10-03 13:52:18,207][starVLA.training.train_starvla][INFO] - Step 4215, Loss: {'train/loss': 4.3006089981645346e-05, 'train/grad_norm_pre_clip': 0.029241370037198067, 'timing/data': 0.00031673999910708517, 'timing/model': 0.6095200249983463, 'learning_rate/qwen_vl_interface': 1.0890112681781821e-06, 'learning_rate/action_model': 1.089011268178182e-05, 'epoch': 0.31}) +[2026-10-03 13:52:18,921][starVLA.training.train_starvla][INFO] - Step 4216, Loss: {'train/loss': 0.019778406247496605, 'train/grad_norm_pre_clip': 37.82005310058594, 'timing/data': 0.00036608999653253704, 'timing/model': 0.39764187799301, 'learning_rate/qwen_vl_interface': 1.0875432697916492e-06, 'learning_rate/action_model': 1.087543269791649e-05, 'epoch': 0.31}) +[2026-10-03 13:52:19,627][starVLA.training.train_starvla][INFO] - Step 4217, Loss: {'train/loss': 7.393136183964089e-05, 'train/grad_norm_pre_clip': 0.03178844973444939, 'timing/data': 0.00030728999990969896, 'timing/model': 0.39473926900245715, 'learning_rate/qwen_vl_interface': 1.0860769824338477e-06, 'learning_rate/action_model': 1.0860769824338478e-05, 'epoch': 0.31}) +[2026-10-03 13:52:20,335][starVLA.training.train_starvla][INFO] - Step 4218, Loss: {'train/loss': 0.2371959090232849, 'train/grad_norm_pre_clip': 215.612060546875, 'timing/data': 0.0002228399971500039, 'timing/model': 0.3959587779972935, 'learning_rate/qwen_vl_interface': 1.084612406707514e-06, 'learning_rate/action_model': 1.084612406707514e-05, 'epoch': 0.31}) +[2026-10-03 13:52:21,041][starVLA.training.train_starvla][INFO] - Step 4219, Loss: {'train/loss': 5.980980131425895e-05, 'train/grad_norm_pre_clip': 0.01698562689125538, 'timing/data': 0.0002676699950825423, 'timing/model': 0.39433707800344564, 'learning_rate/qwen_vl_interface': 1.0831495432146795e-06, 'learning_rate/action_model': 1.0831495432146795e-05, 'epoch': 0.31}) +[2026-10-03 13:52:21,749][starVLA.training.train_starvla][INFO] - Step 4220, Loss: {'train/loss': 0.41784483194351196, 'train/grad_norm_pre_clip': 92.34009552001953, 'timing/data': 0.00035427999682724476, 'timing/model': 0.3949585589871276, 'learning_rate/qwen_vl_interface': 1.0816883925566717e-06, 'learning_rate/action_model': 1.0816883925566716e-05, 'epoch': 0.31}) +[2026-10-03 13:52:22,659][starVLA.training.train_starvla][INFO] - Step 4221, Loss: {'train/loss': 0.676918625831604, 'train/grad_norm_pre_clip': 48.62813186645508, 'timing/data': 0.00030979000439401716, 'timing/model': 0.39486363899777643, 'learning_rate/qwen_vl_interface': 1.0802289553341153e-06, 'learning_rate/action_model': 1.0802289553341152e-05, 'epoch': 0.31}) +[2026-10-03 13:52:23,365][starVLA.training.train_starvla][INFO] - Step 4222, Loss: {'train/loss': 0.00015044413157738745, 'train/grad_norm_pre_clip': 0.04283948987722397, 'timing/data': 0.0004033900040667504, 'timing/model': 0.3931738090032013, 'learning_rate/qwen_vl_interface': 1.0787712321469289e-06, 'learning_rate/action_model': 1.0787712321469288e-05, 'epoch': 0.31}) +[2026-10-03 13:52:24,074][starVLA.training.train_starvla][INFO] - Step 4223, Loss: {'train/loss': 5.6922085605037864e-06, 'train/grad_norm_pre_clip': 0.0015552800614386797, 'timing/data': 0.0004476199974305928, 'timing/model': 0.3957554880034877, 'learning_rate/qwen_vl_interface': 1.077315223594328e-06, 'learning_rate/action_model': 1.077315223594328e-05, 'epoch': 0.31}) +[2026-10-03 13:52:24,780][starVLA.training.train_starvla][INFO] - Step 4224, Loss: {'train/loss': 0.23261642456054688, 'train/grad_norm_pre_clip': 271.7212829589844, 'timing/data': 0.00033704999077599496, 'timing/model': 0.3945392680034274, 'learning_rate/qwen_vl_interface': 1.0758609302748227e-06, 'learning_rate/action_model': 1.0758609302748227e-05, 'epoch': 0.31}) +[2026-10-03 13:52:25,487][starVLA.training.train_starvla][INFO] - Step 4225, Loss: {'train/loss': 0.0008458212832920253, 'train/grad_norm_pre_clip': 1.2490040063858032, 'timing/data': 0.0002744800003711134, 'timing/model': 0.39523415800067596, 'learning_rate/qwen_vl_interface': 1.0744083527862182e-06, 'learning_rate/action_model': 1.074408352786218e-05, 'epoch': 0.31}) +[2026-10-03 13:52:26,407][starVLA.training.train_starvla][INFO] - Step 4226, Loss: {'train/loss': 2.6566971428110264e-05, 'train/grad_norm_pre_clip': 0.021121978759765625, 'timing/data': 0.00035712000681087375, 'timing/model': 0.6081546749919653, 'learning_rate/qwen_vl_interface': 1.0729574917256144e-06, 'learning_rate/action_model': 1.0729574917256145e-05, 'epoch': 0.31}) +[2026-10-03 13:52:27,115][starVLA.training.train_starvla][INFO] - Step 4227, Loss: {'train/loss': 5.103620424051769e-06, 'train/grad_norm_pre_clip': 0.0015194510342553258, 'timing/data': 0.00027470900386106223, 'timing/model': 0.3959716590034077, 'learning_rate/qwen_vl_interface': 1.0715083476894047e-06, 'learning_rate/action_model': 1.0715083476894047e-05, 'epoch': 0.31}) +[2026-10-03 13:52:27,827][starVLA.training.train_starvla][INFO] - Step 4228, Loss: {'train/loss': 3.508289228193462e-05, 'train/grad_norm_pre_clip': 0.010935602709650993, 'timing/data': 0.0003215000033378601, 'timing/model': 0.39674690800893586, 'learning_rate/qwen_vl_interface': 1.0700609212732773e-06, 'learning_rate/action_model': 1.0700609212732774e-05, 'epoch': 0.31}) +[2026-10-03 13:52:28,541][starVLA.training.train_starvla][INFO] - Step 4229, Loss: {'train/loss': 0.00010390377428848296, 'train/grad_norm_pre_clip': 0.03526486083865166, 'timing/data': 0.00028020000900141895, 'timing/model': 0.3988913780049188, 'learning_rate/qwen_vl_interface': 1.0686152130722155e-06, 'learning_rate/action_model': 1.0686152130722154e-05, 'epoch': 0.31}) +[2026-10-03 13:52:29,252][starVLA.training.train_starvla][INFO] - Step 4230, Loss: {'train/loss': 5.237725417828187e-06, 'train/grad_norm_pre_clip': 0.0013169911690056324, 'timing/data': 0.00038147000304888934, 'timing/model': 0.3948959990084404, 'learning_rate/qwen_vl_interface': 1.0671712236804947e-06, 'learning_rate/action_model': 1.0671712236804946e-05, 'epoch': 0.31}) +[2026-10-03 13:52:29,960][starVLA.training.train_starvla][INFO] - Step 4231, Loss: {'train/loss': 8.255127795564476e-06, 'train/grad_norm_pre_clip': 0.002695907838642597, 'timing/data': 0.00040040000749286264, 'timing/model': 0.396248409000691, 'learning_rate/qwen_vl_interface': 1.0657289536916845e-06, 'learning_rate/action_model': 1.0657289536916845e-05, 'epoch': 0.31}) +[2026-10-03 13:52:30,877][starVLA.training.train_starvla][INFO] - Step 4232, Loss: {'train/loss': 0.0002487299789208919, 'train/grad_norm_pre_clip': 0.2617108225822449, 'timing/data': 0.00036071000795345753, 'timing/model': 0.39561321900691837, 'learning_rate/qwen_vl_interface': 1.0642884036986481e-06, 'learning_rate/action_model': 1.064288403698648e-05, 'epoch': 0.31}) +[2026-10-03 13:52:31,587][starVLA.training.train_starvla][INFO] - Step 4233, Loss: {'train/loss': 2.868467390726437e-06, 'train/grad_norm_pre_clip': 0.0007985626580193639, 'timing/data': 0.00044283000170253217, 'timing/model': 0.39780294900992885, 'learning_rate/qwen_vl_interface': 1.0628495742935404e-06, 'learning_rate/action_model': 1.0628495742935405e-05, 'epoch': 0.31}) +[2026-10-03 13:52:32,293][starVLA.training.train_starvla][INFO] - Step 4234, Loss: {'train/loss': 6.943846983631374e-06, 'train/grad_norm_pre_clip': 0.001467695808969438, 'timing/data': 0.00036736000038217753, 'timing/model': 0.3930960290017538, 'learning_rate/qwen_vl_interface': 1.0614124660678103e-06, 'learning_rate/action_model': 1.0614124660678102e-05, 'epoch': 0.31}) +[2026-10-03 13:52:33,000][starVLA.training.train_starvla][INFO] - Step 4235, Loss: {'train/loss': 0.00011208775686100125, 'train/grad_norm_pre_clip': 0.10519464313983917, 'timing/data': 0.0003505200002109632, 'timing/model': 0.39479726798890624, 'learning_rate/qwen_vl_interface': 1.0599770796121988e-06, 'learning_rate/action_model': 1.0599770796121987e-05, 'epoch': 0.31}) +[2026-10-03 13:52:33,705][starVLA.training.train_starvla][INFO] - Step 4236, Loss: {'train/loss': 0.00019876292208209634, 'train/grad_norm_pre_clip': 0.09862108528614044, 'timing/data': 0.0003463199973339215, 'timing/model': 0.3937641490047099, 'learning_rate/qwen_vl_interface': 1.0585434155167394e-06, 'learning_rate/action_model': 1.0585434155167393e-05, 'epoch': 0.31}) +[2026-10-03 13:52:34,629][starVLA.training.train_starvla][INFO] - Step 4237, Loss: {'train/loss': 0.36038026213645935, 'train/grad_norm_pre_clip': 65.43562316894531, 'timing/data': 0.00033707999682519585, 'timing/model': 0.3953523200034397, 'learning_rate/qwen_vl_interface': 1.0571114743707556e-06, 'learning_rate/action_model': 1.0571114743707555e-05, 'epoch': 0.31}) +[2026-10-03 13:52:35,333][starVLA.training.train_starvla][INFO] - Step 4238, Loss: {'train/loss': 0.44458428025245667, 'train/grad_norm_pre_clip': 30.31709861755371, 'timing/data': 0.00036535000253934413, 'timing/model': 0.39352936900104396, 'learning_rate/qwen_vl_interface': 1.0556812567628661e-06, 'learning_rate/action_model': 1.055681256762866e-05, 'epoch': 0.31}) +[2026-10-03 13:52:36,037][starVLA.training.train_starvla][INFO] - Step 4239, Loss: {'train/loss': 0.36770763993263245, 'train/grad_norm_pre_clip': 119.78721618652344, 'timing/data': 0.00024139999004546553, 'timing/model': 0.39310013900103513, 'learning_rate/qwen_vl_interface': 1.0542527632809783e-06, 'learning_rate/action_model': 1.0542527632809784e-05, 'epoch': 0.31}) +[2026-10-03 13:52:36,742][starVLA.training.train_starvla][INFO] - Step 4240, Loss: {'train/loss': 0.013622294180095196, 'train/grad_norm_pre_clip': 59.91971206665039, 'timing/data': 0.0003163599903928116, 'timing/model': 0.39395580800191965, 'learning_rate/qwen_vl_interface': 1.0528259945122925e-06, 'learning_rate/action_model': 1.0528259945122925e-05, 'epoch': 0.31}) +[2026-10-03 13:52:37,450][starVLA.training.train_starvla][INFO] - Step 4241, Loss: {'train/loss': 0.0008053656201809645, 'train/grad_norm_pre_clip': 4.463041305541992, 'timing/data': 0.00028731999918818474, 'timing/model': 0.39579193900863174, 'learning_rate/qwen_vl_interface': 1.0514009510432988e-06, 'learning_rate/action_model': 1.0514009510432988e-05, 'epoch': 0.31}) +[2026-10-03 13:52:38,355][starVLA.training.train_starvla][INFO] - Step 4242, Loss: {'train/loss': 0.7351310849189758, 'train/grad_norm_pre_clip': 38.31498718261719, 'timing/data': 0.00022153000463731587, 'timing/model': 0.5936977879900951, 'learning_rate/qwen_vl_interface': 1.0499776334597796e-06, 'learning_rate/action_model': 1.0499776334597797e-05, 'epoch': 0.31}) +[2026-10-03 13:52:39,060][starVLA.training.train_starvla][INFO] - Step 4243, Loss: {'train/loss': 0.00012162393250036985, 'train/grad_norm_pre_clip': 0.049637570977211, 'timing/data': 0.0003500500024529174, 'timing/model': 0.39346771899727173, 'learning_rate/qwen_vl_interface': 1.0485560423468064e-06, 'learning_rate/action_model': 1.0485560423468063e-05, 'epoch': 0.31}) +[2026-10-03 13:52:39,767][starVLA.training.train_starvla][INFO] - Step 4244, Loss: {'train/loss': 0.0002567248302511871, 'train/grad_norm_pre_clip': 0.21434491872787476, 'timing/data': 0.0003292499895906076, 'timing/model': 0.39526102899981197, 'learning_rate/qwen_vl_interface': 1.0471361782887421e-06, 'learning_rate/action_model': 1.0471361782887421e-05, 'epoch': 0.31}) +[2026-10-03 13:52:40,475][starVLA.training.train_starvla][INFO] - Step 4245, Loss: {'train/loss': 1.3611674148705788e-05, 'train/grad_norm_pre_clip': 0.003987019881606102, 'timing/data': 0.0003350999904796481, 'timing/model': 0.3954327590035973, 'learning_rate/qwen_vl_interface': 1.0457180418692385e-06, 'learning_rate/action_model': 1.0457180418692383e-05, 'epoch': 0.31}) +[2026-10-03 13:52:41,181][starVLA.training.train_starvla][INFO] - Step 4246, Loss: {'train/loss': 2.8092636057408527e-05, 'train/grad_norm_pre_clip': 0.006748624611645937, 'timing/data': 0.0003922000032616779, 'timing/model': 0.39393237899639644, 'learning_rate/qwen_vl_interface': 1.0443016336712395e-06, 'learning_rate/action_model': 1.0443016336712393e-05, 'epoch': 0.31}) +[2026-10-03 13:52:41,886][starVLA.training.train_starvla][INFO] - Step 4247, Loss: {'train/loss': 9.560264879837632e-05, 'train/grad_norm_pre_clip': 0.05595892667770386, 'timing/data': 0.00025055000151041895, 'timing/model': 0.3929017079935875, 'learning_rate/qwen_vl_interface': 1.0428869542769744e-06, 'learning_rate/action_model': 1.0428869542769744e-05, 'epoch': 0.31}) +[2026-10-03 13:52:42,803][starVLA.training.train_starvla][INFO] - Step 4248, Loss: {'train/loss': 0.0553908571600914, 'train/grad_norm_pre_clip': 143.51914978027344, 'timing/data': 0.00030029901245143265, 'timing/model': 0.39417686000524554, 'learning_rate/qwen_vl_interface': 1.0414740042679663e-06, 'learning_rate/action_model': 1.0414740042679662e-05, 'epoch': 0.31}) +[2026-10-03 13:52:43,510][starVLA.training.train_starvla][INFO] - Step 4249, Loss: {'train/loss': 0.3899087905883789, 'train/grad_norm_pre_clip': 38.66626739501953, 'timing/data': 0.0002525099989725277, 'timing/model': 0.39551539899548516, 'learning_rate/qwen_vl_interface': 1.040062784225025e-06, 'learning_rate/action_model': 1.0400627842250249e-05, 'epoch': 0.31}) +[2026-10-03 13:52:44,214][starVLA.training.train_starvla][INFO] - Step 4250, Loss: {'train/loss': 1.4237392861105036e-05, 'train/grad_norm_pre_clip': 0.0076452698558568954, 'timing/data': 0.0002488799946149811, 'timing/model': 0.39367396899615414, 'learning_rate/qwen_vl_interface': 1.0386532947282496e-06, 'learning_rate/action_model': 1.0386532947282495e-05, 'epoch': 0.31}) +[2026-10-03 13:52:44,920][starVLA.training.train_starvla][INFO] - Step 4251, Loss: {'train/loss': 0.30724960565567017, 'train/grad_norm_pre_clip': 38.34773254394531, 'timing/data': 0.00037306999729480594, 'timing/model': 0.39465647900942713, 'learning_rate/qwen_vl_interface': 1.0372455363570282e-06, 'learning_rate/action_model': 1.0372455363570283e-05, 'epoch': 0.31}) +[2026-10-03 13:52:45,627][starVLA.training.train_starvla][INFO] - Step 4252, Loss: {'train/loss': 0.00014375314640346915, 'train/grad_norm_pre_clip': 0.12801524996757507, 'timing/data': 0.00023584000882692635, 'timing/model': 0.3949462989985477, 'learning_rate/qwen_vl_interface': 1.035839509690039e-06, 'learning_rate/action_model': 1.0358395096900388e-05, 'epoch': 0.31}) +[2026-10-03 13:52:46,541][starVLA.training.train_starvla][INFO] - Step 4253, Loss: {'train/loss': 0.0024066357873380184, 'train/grad_norm_pre_clip': 11.854634284973145, 'timing/data': 0.0003838599950540811, 'timing/model': 0.39557856899045873, 'learning_rate/qwen_vl_interface': 1.0344352153052429e-06, 'learning_rate/action_model': 1.0344352153052427e-05, 'epoch': 0.31}) +[2026-10-03 13:52:47,246][starVLA.training.train_starvla][INFO] - Step 4254, Loss: {'train/loss': 0.5727830529212952, 'train/grad_norm_pre_clip': 58.18937301635742, 'timing/data': 0.00026125999283976853, 'timing/model': 0.39409509900724515, 'learning_rate/qwen_vl_interface': 1.033032653779894e-06, 'learning_rate/action_model': 1.033032653779894e-05, 'epoch': 0.31}) +[2026-10-03 13:52:47,954][starVLA.training.train_starvla][INFO] - Step 4255, Loss: {'train/loss': 0.265559583902359, 'train/grad_norm_pre_clip': 130.36221313476562, 'timing/data': 0.00035488999856170267, 'timing/model': 0.39468815800501034, 'learning_rate/qwen_vl_interface': 1.0316318256905325e-06, 'learning_rate/action_model': 1.0316318256905324e-05, 'epoch': 0.31}) +[2026-10-03 13:52:48,660][starVLA.training.train_starvla][INFO] - Step 4256, Loss: {'train/loss': 1.4498489690595306e-05, 'train/grad_norm_pre_clip': 0.004137673880904913, 'timing/data': 0.0002911100018536672, 'timing/model': 0.3947523089882452, 'learning_rate/qwen_vl_interface': 1.0302327316129856e-06, 'learning_rate/action_model': 1.0302327316129855e-05, 'epoch': 0.31}) +[2026-10-03 13:52:49,368][starVLA.training.train_starvla][INFO] - Step 4257, Loss: {'train/loss': 1.5615361917298287e-05, 'train/grad_norm_pre_clip': 0.0060732862912118435, 'timing/data': 0.00025893999554682523, 'timing/model': 0.39583677900373004, 'learning_rate/qwen_vl_interface': 1.028835372122368e-06, 'learning_rate/action_model': 1.0288353721223682e-05, 'epoch': 0.31}) +[2026-10-03 13:52:50,276][starVLA.training.train_starvla][INFO] - Step 4258, Loss: {'train/loss': 6.873108941363171e-05, 'train/grad_norm_pre_clip': 0.051603347063064575, 'timing/data': 0.00035457999911159277, 'timing/model': 0.5955497579998337, 'learning_rate/qwen_vl_interface': 1.0274397477930822e-06, 'learning_rate/action_model': 1.0274397477930822e-05, 'epoch': 0.31}) +[2026-10-03 13:52:50,982][starVLA.training.train_starvla][INFO] - Step 4259, Loss: {'train/loss': 0.0002177447750000283, 'train/grad_norm_pre_clip': 0.17141953110694885, 'timing/data': 0.00033518999407533556, 'timing/model': 0.39448056899709627, 'learning_rate/qwen_vl_interface': 1.0260458591988164e-06, 'learning_rate/action_model': 1.0260458591988165e-05, 'epoch': 0.31}) +[2026-10-03 13:52:51,689][starVLA.training.train_starvla][INFO] - Step 4260, Loss: {'train/loss': 2.505514567019418e-05, 'train/grad_norm_pre_clip': 0.009737079031765461, 'timing/data': 0.00033209999673999846, 'timing/model': 0.39461113899596967, 'learning_rate/qwen_vl_interface': 1.0246537069125445e-06, 'learning_rate/action_model': 1.0246537069125444e-05, 'epoch': 0.31}) +[2026-10-03 13:52:52,397][starVLA.training.train_starvla][INFO] - Step 4261, Loss: {'train/loss': 1.5437108231708407e-05, 'train/grad_norm_pre_clip': 0.005957415793091059, 'timing/data': 0.0003615199966588989, 'timing/model': 0.3959894289873773, 'learning_rate/qwen_vl_interface': 1.0232632915065286e-06, 'learning_rate/action_model': 1.0232632915065285e-05, 'epoch': 0.31}) +[2026-10-03 13:52:53,105][starVLA.training.train_starvla][INFO] - Step 4262, Loss: {'train/loss': 3.184083107043989e-05, 'train/grad_norm_pre_clip': 0.01025376282632351, 'timing/data': 0.00024266001128125936, 'timing/model': 0.3950305489997845, 'learning_rate/qwen_vl_interface': 1.021874613552316e-06, 'learning_rate/action_model': 1.0218746135523157e-05, 'epoch': 0.31}) +[2026-10-03 13:52:54,011][starVLA.training.train_starvla][INFO] - Step 4263, Loss: {'train/loss': 0.0005074458313174546, 'train/grad_norm_pre_clip': 0.31355252861976624, 'timing/data': 0.00024486001348122954, 'timing/model': 0.5945074079936603, 'learning_rate/qwen_vl_interface': 1.0204876736207384e-06, 'learning_rate/action_model': 1.0204876736207384e-05, 'epoch': 0.31}) +[2026-10-03 13:52:54,718][starVLA.training.train_starvla][INFO] - Step 4264, Loss: {'train/loss': 9.785524889593944e-05, 'train/grad_norm_pre_clip': 0.027583757415413857, 'timing/data': 0.0003712399920914322, 'timing/model': 0.3951001290115528, 'learning_rate/qwen_vl_interface': 1.0191024722819163e-06, 'learning_rate/action_model': 1.0191024722819162e-05, 'epoch': 0.31}) +[2026-10-03 13:52:55,426][starVLA.training.train_starvla][INFO] - Step 4265, Loss: {'train/loss': 1.0959689461742528e-05, 'train/grad_norm_pre_clip': 0.0029905156698077917, 'timing/data': 0.00025956000899896026, 'timing/model': 0.395480999010033, 'learning_rate/qwen_vl_interface': 1.017719010105253e-06, 'learning_rate/action_model': 1.017719010105253e-05, 'epoch': 0.31}) +[2026-10-03 13:52:56,133][starVLA.training.train_starvla][INFO] - Step 4266, Loss: {'train/loss': 7.461776840500534e-05, 'train/grad_norm_pre_clip': 0.023160360753536224, 'timing/data': 0.00022809900110587478, 'timing/model': 0.394751810003072, 'learning_rate/qwen_vl_interface': 1.016337287659437e-06, 'learning_rate/action_model': 1.0163372876594368e-05, 'epoch': 0.31}) +[2026-10-03 13:52:56,840][starVLA.training.train_starvla][INFO] - Step 4267, Loss: {'train/loss': 0.0013942864025011659, 'train/grad_norm_pre_clip': 1.6681541204452515, 'timing/data': 0.00022357000852935016, 'timing/model': 0.39470819900452625, 'learning_rate/qwen_vl_interface': 1.0149573055124427e-06, 'learning_rate/action_model': 1.0149573055124426e-05, 'epoch': 0.31}) +[2026-10-03 13:52:57,757][starVLA.training.train_starvla][INFO] - Step 4268, Loss: {'train/loss': 3.807869870797731e-05, 'train/grad_norm_pre_clip': 0.010627971030771732, 'timing/data': 0.0003690699959406629, 'timing/model': 0.6045534069999121, 'learning_rate/qwen_vl_interface': 1.013579064231529e-06, 'learning_rate/action_model': 1.013579064231529e-05, 'epoch': 0.31}) +[2026-10-03 13:52:58,466][starVLA.training.train_starvla][INFO] - Step 4269, Loss: {'train/loss': 0.000211769001907669, 'train/grad_norm_pre_clip': 0.14002037048339844, 'timing/data': 0.0002805399999488145, 'timing/model': 0.39622005900309887, 'learning_rate/qwen_vl_interface': 1.012202564383239e-06, 'learning_rate/action_model': 1.0122025643832388e-05, 'epoch': 0.31}) +[2026-10-03 13:52:59,172][starVLA.training.train_starvla][INFO] - Step 4270, Loss: {'train/loss': 6.370171831804328e-06, 'train/grad_norm_pre_clip': 0.0023989456240087748, 'timing/data': 0.00035447999835014343, 'timing/model': 0.393736809011898, 'learning_rate/qwen_vl_interface': 1.0108278065333978e-06, 'learning_rate/action_model': 1.0108278065333978e-05, 'epoch': 0.31}) +[2026-10-03 13:52:59,878][starVLA.training.train_starvla][INFO] - Step 4271, Loss: {'train/loss': 4.2915162339340895e-06, 'train/grad_norm_pre_clip': 0.0013302324805408716, 'timing/data': 0.0002874400088330731, 'timing/model': 0.3942290499981027, 'learning_rate/qwen_vl_interface': 1.0094547912471183e-06, 'learning_rate/action_model': 1.0094547912471184e-05, 'epoch': 0.31}) +[2026-10-03 13:53:00,584][starVLA.training.train_starvla][INFO] - Step 4272, Loss: {'train/loss': 0.000174697459442541, 'train/grad_norm_pre_clip': 0.06416483968496323, 'timing/data': 0.0003128400130663067, 'timing/model': 0.3953055790043436, 'learning_rate/qwen_vl_interface': 1.0080835190887949e-06, 'learning_rate/action_model': 1.0080835190887947e-05, 'epoch': 0.31}) +[2026-10-03 13:53:01,293][starVLA.training.train_starvla][INFO] - Step 4273, Loss: {'train/loss': 1.5377590898424387e-05, 'train/grad_norm_pre_clip': 0.004070437513291836, 'timing/data': 0.0002496700035408139, 'timing/model': 0.3965325489989482, 'learning_rate/qwen_vl_interface': 1.0067139906221052e-06, 'learning_rate/action_model': 1.0067139906221052e-05, 'epoch': 0.31}) +[2026-10-03 13:53:02,201][starVLA.training.train_starvla][INFO] - Step 4274, Loss: {'train/loss': 7.539882062701508e-06, 'train/grad_norm_pre_clip': 0.0019394485279917717, 'timing/data': 0.00039382999239023775, 'timing/model': 0.39269241099827923, 'learning_rate/qwen_vl_interface': 1.0053462064100112e-06, 'learning_rate/action_model': 1.0053462064100112e-05, 'epoch': 0.31}) +[2026-10-03 13:53:02,907][starVLA.training.train_starvla][INFO] - Step 4275, Loss: {'train/loss': 4.447981609700946e-06, 'train/grad_norm_pre_clip': 0.0016723503358662128, 'timing/data': 0.0003060099988942966, 'timing/model': 0.3939127100020414, 'learning_rate/qwen_vl_interface': 1.003980167014757e-06, 'learning_rate/action_model': 1.003980167014757e-05, 'epoch': 0.31}) +[2026-10-03 13:53:03,615][starVLA.training.train_starvla][INFO] - Step 4276, Loss: {'train/loss': 0.00015921912563499063, 'train/grad_norm_pre_clip': 0.12719137966632843, 'timing/data': 0.0003137500025331974, 'timing/model': 0.3954814390017418, 'learning_rate/qwen_vl_interface': 1.0026158729978702e-06, 'learning_rate/action_model': 1.0026158729978702e-05, 'epoch': 0.31}) +[2026-10-03 13:53:04,321][starVLA.training.train_starvla][INFO] - Step 4277, Loss: {'train/loss': 2.0070669052074663e-05, 'train/grad_norm_pre_clip': 0.010777830146253109, 'timing/data': 0.0002816199994413182, 'timing/model': 0.39538792899111286, 'learning_rate/qwen_vl_interface': 1.0012533249201604e-06, 'learning_rate/action_model': 1.0012533249201602e-05, 'epoch': 0.31}) +[2026-10-03 13:53:05,026][starVLA.training.train_starvla][INFO] - Step 4278, Loss: {'train/loss': 9.462140951654874e-06, 'train/grad_norm_pre_clip': 0.0019045586232095957, 'timing/data': 0.0003461600135779008, 'timing/model': 0.3932815290027065, 'learning_rate/qwen_vl_interface': 9.99892523341719e-07, 'learning_rate/action_model': 9.99892523341719e-06, 'epoch': 0.31}) +[2026-10-03 13:53:05,942][starVLA.training.train_starvla][INFO] - Step 4279, Loss: {'train/loss': 0.00018510916561353952, 'train/grad_norm_pre_clip': 0.07798837125301361, 'timing/data': 0.0002266590017825365, 'timing/model': 0.6047690880077425, 'learning_rate/qwen_vl_interface': 9.98533468821922e-07, 'learning_rate/action_model': 9.98533468821922e-06, 'epoch': 0.31}) +[2026-10-03 13:53:06,649][starVLA.training.train_starvla][INFO] - Step 4280, Loss: {'train/loss': 0.0004649391339626163, 'train/grad_norm_pre_clip': 0.9954567551612854, 'timing/data': 0.00022437000006902963, 'timing/model': 0.39429463900160044, 'learning_rate/qwen_vl_interface': 9.971761619194231e-07, 'learning_rate/action_model': 9.97176161919423e-06, 'epoch': 0.31}) +[2026-10-03 13:53:07,356][starVLA.training.train_starvla][INFO] - Step 4281, Loss: {'train/loss': 8.784164492681157e-06, 'train/grad_norm_pre_clip': 0.002808620920404792, 'timing/data': 0.0003424699971219525, 'timing/model': 0.39693032899231184, 'learning_rate/qwen_vl_interface': 9.958206031921606e-07, 'learning_rate/action_model': 9.958206031921604e-06, 'epoch': 0.31}) +[2026-10-03 13:53:08,064][starVLA.training.train_starvla][INFO] - Step 4282, Loss: {'train/loss': 0.04746324196457863, 'train/grad_norm_pre_clip': 70.77992248535156, 'timing/data': 0.00024020001001190394, 'timing/model': 0.3953170689928811, 'learning_rate/qwen_vl_interface': 9.944667931973537e-07, 'learning_rate/action_model': 9.944667931973536e-06, 'epoch': 0.31}) +[2026-10-03 13:53:08,771][starVLA.training.train_starvla][INFO] - Step 4283, Loss: {'train/loss': 2.3959135432960466e-05, 'train/grad_norm_pre_clip': 0.010322058573365211, 'timing/data': 0.00024016000679694116, 'timing/model': 0.39514439899357967, 'learning_rate/qwen_vl_interface': 9.93114732491503e-07, 'learning_rate/action_model': 9.93114732491503e-06, 'epoch': 0.31}) +[2026-10-03 13:53:09,477][starVLA.training.train_starvla][INFO] - Step 4284, Loss: {'train/loss': 3.547468804754317e-05, 'train/grad_norm_pre_clip': 0.03271505609154701, 'timing/data': 0.0003422599984332919, 'timing/model': 0.39365911899949424, 'learning_rate/qwen_vl_interface': 9.91764421630389e-07, 'learning_rate/action_model': 9.91764421630389e-06, 'epoch': 0.31}) +[2026-10-03 13:53:10,380][starVLA.training.train_starvla][INFO] - Step 4285, Loss: {'train/loss': 1.2315586900513154e-05, 'train/grad_norm_pre_clip': 0.0027128937654197216, 'timing/data': 0.00035777899029199034, 'timing/model': 0.3945365699910326, 'learning_rate/qwen_vl_interface': 9.904158611690732e-07, 'learning_rate/action_model': 9.90415861169073e-06, 'epoch': 0.31}) +[2026-10-03 13:53:11,087][starVLA.training.train_starvla][INFO] - Step 4286, Loss: {'train/loss': 0.00019714992959052324, 'train/grad_norm_pre_clip': 0.10108952969312668, 'timing/data': 0.0003718499938258901, 'timing/model': 0.39395328900718596, 'learning_rate/qwen_vl_interface': 9.890690516618987e-07, 'learning_rate/action_model': 9.890690516618985e-06, 'epoch': 0.31}) +[2026-10-03 13:53:11,791][starVLA.training.train_starvla][INFO] - Step 4287, Loss: {'train/loss': 1.0887222290039062, 'train/grad_norm_pre_clip': 78.39863586425781, 'timing/data': 0.00025274000654462725, 'timing/model': 0.39345961999788415, 'learning_rate/qwen_vl_interface': 9.877239936624875e-07, 'learning_rate/action_model': 9.877239936624875e-06, 'epoch': 0.31}) +[2026-10-03 13:53:12,498][starVLA.training.train_starvla][INFO] - Step 4288, Loss: {'train/loss': 1.4431076124310493e-05, 'train/grad_norm_pre_clip': 0.0026510453317314386, 'timing/data': 0.000311290001263842, 'timing/model': 0.39398599999549333, 'learning_rate/qwen_vl_interface': 9.863806877237421e-07, 'learning_rate/action_model': 9.863806877237422e-06, 'epoch': 0.31}) +[2026-10-03 13:53:13,206][starVLA.training.train_starvla][INFO] - Step 4289, Loss: {'train/loss': 8.70254953042604e-05, 'train/grad_norm_pre_clip': 0.03690990060567856, 'timing/data': 0.0003113900020252913, 'timing/model': 0.3960458399960771, 'learning_rate/qwen_vl_interface': 9.850391343978454e-07, 'learning_rate/action_model': 9.850391343978454e-06, 'epoch': 0.31}) +[2026-10-03 13:53:14,117][starVLA.training.train_starvla][INFO] - Step 4290, Loss: {'train/loss': 1.5183830328169279e-05, 'train/grad_norm_pre_clip': 0.003633422078564763, 'timing/data': 0.00037803000304847956, 'timing/model': 0.3932659690035507, 'learning_rate/qwen_vl_interface': 9.836993342362596e-07, 'learning_rate/action_model': 9.836993342362596e-06, 'epoch': 0.31}) +[2026-10-03 13:53:14,820][starVLA.training.train_starvla][INFO] - Step 4291, Loss: {'train/loss': 0.0003735707141458988, 'train/grad_norm_pre_clip': 0.8338561654090881, 'timing/data': 0.00022839001030661166, 'timing/model': 0.39256729000771884, 'learning_rate/qwen_vl_interface': 9.823612877897246e-07, 'learning_rate/action_model': 9.823612877897245e-06, 'epoch': 0.31}) +[2026-10-03 13:53:15,522][starVLA.training.train_starvla][INFO] - Step 4292, Loss: {'train/loss': 5.7775450841290876e-05, 'train/grad_norm_pre_clip': 0.01561444066464901, 'timing/data': 0.0002878190134651959, 'timing/model': 0.3924903300066944, 'learning_rate/qwen_vl_interface': 9.81024995608261e-07, 'learning_rate/action_model': 9.810249956082609e-06, 'epoch': 0.31}) +[2026-10-03 13:53:16,227][starVLA.training.train_starvla][INFO] - Step 4293, Loss: {'train/loss': 5.030297324992716e-05, 'train/grad_norm_pre_clip': 0.01716148853302002, 'timing/data': 0.0002827400021487847, 'timing/model': 0.3940496100112796, 'learning_rate/qwen_vl_interface': 9.796904582411685e-07, 'learning_rate/action_model': 9.796904582411684e-06, 'epoch': 0.31}) +[2026-10-03 13:53:16,929][starVLA.training.train_starvla][INFO] - Step 4294, Loss: {'train/loss': 2.085984124278184e-05, 'train/grad_norm_pre_clip': 0.030161038041114807, 'timing/data': 0.0003129999968223274, 'timing/model': 0.39241307000338566, 'learning_rate/qwen_vl_interface': 9.78357676237025e-07, 'learning_rate/action_model': 9.78357676237025e-06, 'epoch': 0.31}) +[2026-10-03 13:53:17,822][starVLA.training.train_starvla][INFO] - Step 4295, Loss: {'train/loss': 1.7098254829761572e-05, 'train/grad_norm_pre_clip': 0.005625431891530752, 'timing/data': 0.0002900300023611635, 'timing/model': 0.583099780997145, 'learning_rate/qwen_vl_interface': 9.77026650143687e-07, 'learning_rate/action_model': 9.77026650143687e-06, 'epoch': 0.31}) +[2026-10-03 13:53:18,526][starVLA.training.train_starvla][INFO] - Step 4296, Loss: {'train/loss': 2.556071376602631e-05, 'train/grad_norm_pre_clip': 0.012522108852863312, 'timing/data': 0.00024945000768639147, 'timing/model': 0.3929367899982026, 'learning_rate/qwen_vl_interface': 9.756973805082882e-07, 'learning_rate/action_model': 9.756973805082883e-06, 'epoch': 0.31}) +[2026-10-03 13:53:19,230][starVLA.training.train_starvla][INFO] - Step 4297, Loss: {'train/loss': 0.000182318952283822, 'train/grad_norm_pre_clip': 0.13997548818588257, 'timing/data': 0.0002841499954229221, 'timing/model': 0.39427894000255037, 'learning_rate/qwen_vl_interface': 9.743698678772434e-07, 'learning_rate/action_model': 9.743698678772433e-06, 'epoch': 0.31}) +[2026-10-03 13:53:19,933][starVLA.training.train_starvla][INFO] - Step 4298, Loss: {'train/loss': 0.0007813385454937816, 'train/grad_norm_pre_clip': 0.8376986980438232, 'timing/data': 0.0002570400101831183, 'timing/model': 0.39285190000373404, 'learning_rate/qwen_vl_interface': 9.730441127962395e-07, 'learning_rate/action_model': 9.730441127962395e-06, 'epoch': 0.31}) +[2026-10-03 13:53:20,635][starVLA.training.train_starvla][INFO] - Step 4299, Loss: {'train/loss': 0.0031789904460310936, 'train/grad_norm_pre_clip': 8.94129753112793, 'timing/data': 0.00032017000194173306, 'timing/model': 0.39251480900566094, 'learning_rate/qwen_vl_interface': 9.717201158102465e-07, 'learning_rate/action_model': 9.717201158102466e-06, 'epoch': 0.31}) +[2026-10-03 13:53:21,340][starVLA.training.train_starvla][INFO] - Step 4300, Loss: {'train/loss': 0.00012660352513194084, 'train/grad_norm_pre_clip': 0.11224819719791412, 'timing/data': 0.00026948000595439225, 'timing/model': 0.39354560000356287, 'learning_rate/qwen_vl_interface': 9.703978774635094e-07, 'learning_rate/action_model': 9.703978774635095e-06, 'epoch': 0.31}) +[2026-10-03 13:53:22,234][starVLA.training.train_starvla][INFO] - Step 4301, Loss: {'train/loss': 3.0239518309826963e-05, 'train/grad_norm_pre_clip': 0.011751673184335232, 'timing/data': 0.0002886000002035871, 'timing/model': 0.3949222289957106, 'learning_rate/qwen_vl_interface': 9.6907739829955e-07, 'learning_rate/action_model': 9.6907739829955e-06, 'epoch': 0.31}) +[2026-10-03 13:53:22,939][starVLA.training.train_starvla][INFO] - Step 4302, Loss: {'train/loss': 1.0224703550338745, 'train/grad_norm_pre_clip': 30.305889129638672, 'timing/data': 0.0002001389948418364, 'timing/model': 0.39339061999635305, 'learning_rate/qwen_vl_interface': 9.677586788611681e-07, 'learning_rate/action_model': 9.677586788611681e-06, 'epoch': 0.31}) +[2026-10-03 13:53:23,643][starVLA.training.train_starvla][INFO] - Step 4303, Loss: {'train/loss': 5.1117756811436266e-05, 'train/grad_norm_pre_clip': 0.022409696131944656, 'timing/data': 0.00029568000172730535, 'timing/model': 0.3929130300093675, 'learning_rate/qwen_vl_interface': 9.664417196904382e-07, 'learning_rate/action_model': 9.664417196904382e-06, 'epoch': 0.31}) +[2026-10-03 13:53:24,347][starVLA.training.train_starvla][INFO] - Step 4304, Loss: {'train/loss': 2.4406350348726846e-05, 'train/grad_norm_pre_clip': 0.005639656912535429, 'timing/data': 0.00024071999359875917, 'timing/model': 0.39288868999574333, 'learning_rate/qwen_vl_interface': 9.651265213287134e-07, 'learning_rate/action_model': 9.651265213287134e-06, 'epoch': 0.32}) +[2026-10-03 13:53:25,053][starVLA.training.train_starvla][INFO] - Step 4305, Loss: {'train/loss': 0.014096304774284363, 'train/grad_norm_pre_clip': 36.80453109741211, 'timing/data': 0.00031862000469118357, 'timing/model': 0.3946083000046201, 'learning_rate/qwen_vl_interface': 9.638130843166218e-07, 'learning_rate/action_model': 9.638130843166219e-06, 'epoch': 0.32}) +[2026-10-03 13:53:25,949][starVLA.training.train_starvla][INFO] - Step 4306, Loss: {'train/loss': 1.8267654013470747e-05, 'train/grad_norm_pre_clip': 0.00534952012822032, 'timing/data': 0.000218439003219828, 'timing/model': 0.3933773000026122, 'learning_rate/qwen_vl_interface': 9.62501409194068e-07, 'learning_rate/action_model': 9.625014091940678e-06, 'epoch': 0.32}) +[2026-10-03 13:53:26,651][starVLA.training.train_starvla][INFO] - Step 4307, Loss: {'train/loss': 0.0005365233519114554, 'train/grad_norm_pre_clip': 0.6238583922386169, 'timing/data': 0.00033239999902434647, 'timing/model': 0.3922312009963207, 'learning_rate/qwen_vl_interface': 9.611914965002317e-07, 'learning_rate/action_model': 9.611914965002316e-06, 'epoch': 0.32}) +[2026-10-03 13:53:27,355][starVLA.training.train_starvla][INFO] - Step 4308, Loss: {'train/loss': 3.989480319432914e-05, 'train/grad_norm_pre_clip': 0.015236249193549156, 'timing/data': 0.00020401000801939517, 'timing/model': 0.3935510499868542, 'learning_rate/qwen_vl_interface': 9.598833467735686e-07, 'learning_rate/action_model': 9.598833467735685e-06, 'epoch': 0.32}) +[2026-10-03 13:53:28,060][starVLA.training.train_starvla][INFO] - Step 4309, Loss: {'train/loss': 0.0036825600545853376, 'train/grad_norm_pre_clip': 11.411911010742188, 'timing/data': 0.00021757000649813563, 'timing/model': 0.3945068699977128, 'learning_rate/qwen_vl_interface': 9.585769605518097e-07, 'learning_rate/action_model': 9.585769605518097e-06, 'epoch': 0.32}) +[2026-10-03 13:53:28,763][starVLA.training.train_starvla][INFO] - Step 4310, Loss: {'train/loss': 1.6017591406125575e-05, 'train/grad_norm_pre_clip': 0.006466475315392017, 'timing/data': 0.0002930899936472997, 'timing/model': 0.3928076899901498, 'learning_rate/qwen_vl_interface': 9.572723383719611e-07, 'learning_rate/action_model': 9.572723383719612e-06, 'epoch': 0.32}) +[2026-10-03 13:53:29,657][starVLA.training.train_starvla][INFO] - Step 4311, Loss: {'train/loss': 0.41617393493652344, 'train/grad_norm_pre_clip': 62.91400146484375, 'timing/data': 0.00021451999782584608, 'timing/model': 0.5844931610045023, 'learning_rate/qwen_vl_interface': 9.559694807703041e-07, 'learning_rate/action_model': 9.55969480770304e-06, 'epoch': 0.32}) +[2026-10-03 13:53:30,361][starVLA.training.train_starvla][INFO] - Step 4312, Loss: {'train/loss': 0.24866005778312683, 'train/grad_norm_pre_clip': 43.869178771972656, 'timing/data': 0.00030471000354737043, 'timing/model': 0.3933142099995166, 'learning_rate/qwen_vl_interface': 9.546683882823945e-07, 'learning_rate/action_model': 9.546683882823945e-06, 'epoch': 0.32}) +[2026-10-03 13:53:31,066][starVLA.training.train_starvla][INFO] - Step 4313, Loss: {'train/loss': 6.742293044226244e-05, 'train/grad_norm_pre_clip': 0.030581612139940262, 'timing/data': 0.0002797799970721826, 'timing/model': 0.3945230199897196, 'learning_rate/qwen_vl_interface': 9.533690614430608e-07, 'learning_rate/action_model': 9.533690614430608e-06, 'epoch': 0.32}) +[2026-10-03 13:53:31,771][starVLA.training.train_starvla][INFO] - Step 4314, Loss: {'train/loss': 1.5988447557901964e-05, 'train/grad_norm_pre_clip': 0.008703436702489853, 'timing/data': 0.00031320999551098794, 'timing/model': 0.3936192800028948, 'learning_rate/qwen_vl_interface': 9.520715007864082e-07, 'learning_rate/action_model': 9.52071500786408e-06, 'epoch': 0.32}) +[2026-10-03 13:53:32,474][starVLA.training.train_starvla][INFO] - Step 4315, Loss: {'train/loss': 6.331435724860057e-05, 'train/grad_norm_pre_clip': 0.05259223282337189, 'timing/data': 0.00020094000501558185, 'timing/model': 0.3929870600113645, 'learning_rate/qwen_vl_interface': 9.50775706845814e-07, 'learning_rate/action_model': 9.50775706845814e-06, 'epoch': 0.32}) +[2026-10-03 13:53:33,370][starVLA.training.train_starvla][INFO] - Step 4316, Loss: {'train/loss': 0.00010422489140182734, 'train/grad_norm_pre_clip': 0.0452151820063591, 'timing/data': 0.00024896900868043303, 'timing/model': 0.5851961410080548, 'learning_rate/qwen_vl_interface': 9.49481680153931e-07, 'learning_rate/action_model': 9.494816801539309e-06, 'epoch': 0.32}) +[2026-10-03 13:53:34,076][starVLA.training.train_starvla][INFO] - Step 4317, Loss: {'train/loss': 0.11327212303876877, 'train/grad_norm_pre_clip': 90.91362762451172, 'timing/data': 0.0003023899917025119, 'timing/model': 0.394187170008081, 'learning_rate/qwen_vl_interface': 9.481894212426848e-07, 'learning_rate/action_model': 9.481894212426847e-06, 'epoch': 0.32}) +[2026-10-03 13:53:34,779][starVLA.training.train_starvla][INFO] - Step 4318, Loss: {'train/loss': 1.3783176655124407e-05, 'train/grad_norm_pre_clip': 0.0031160851940512657, 'timing/data': 0.0002442599943606183, 'timing/model': 0.39322314999299124, 'learning_rate/qwen_vl_interface': 9.468989306432738e-07, 'learning_rate/action_model': 9.468989306432738e-06, 'epoch': 0.32}) +[2026-10-03 13:53:35,482][starVLA.training.train_starvla][INFO] - Step 4319, Loss: {'train/loss': 5.399888686952181e-05, 'train/grad_norm_pre_clip': 0.014693533070385456, 'timing/data': 0.000287318995106034, 'timing/model': 0.39253401099995244, 'learning_rate/qwen_vl_interface': 9.456102088861702e-07, 'learning_rate/action_model': 9.456102088861701e-06, 'epoch': 0.32}) +[2026-10-03 13:53:36,185][starVLA.training.train_starvla][INFO] - Step 4320, Loss: {'train/loss': 2.8427763027139008e-05, 'train/grad_norm_pre_clip': 0.006249760743230581, 'timing/data': 0.00024805999419186264, 'timing/model': 0.393077059998177, 'learning_rate/qwen_vl_interface': 9.443232565011188e-07, 'learning_rate/action_model': 9.443232565011187e-06, 'epoch': 0.32}) +[2026-10-03 13:53:37,083][starVLA.training.train_starvla][INFO] - Step 4321, Loss: {'train/loss': 0.02442842163145542, 'train/grad_norm_pre_clip': 45.41364288330078, 'timing/data': 0.0004123800026718527, 'timing/model': 0.5873475810076343, 'learning_rate/qwen_vl_interface': 9.430380740171369e-07, 'learning_rate/action_model': 9.430380740171368e-06, 'epoch': 0.32}) +[2026-10-03 13:53:37,786][starVLA.training.train_starvla][INFO] - Step 4322, Loss: {'train/loss': 1.8826474843081087e-05, 'train/grad_norm_pre_clip': 0.007642998825758696, 'timing/data': 0.00019619999511633068, 'timing/model': 0.3926618499972392, 'learning_rate/qwen_vl_interface': 9.417546619625143e-07, 'learning_rate/action_model': 9.417546619625143e-06, 'epoch': 0.32}) +[2026-10-03 13:53:38,488][starVLA.training.train_starvla][INFO] - Step 4323, Loss: {'train/loss': 5.38669928573654e-06, 'train/grad_norm_pre_clip': 0.001576696871779859, 'timing/data': 0.00027847000455949455, 'timing/model': 0.392483789997641, 'learning_rate/qwen_vl_interface': 9.404730208648144e-07, 'learning_rate/action_model': 9.404730208648144e-06, 'epoch': 0.32}) +[2026-10-03 13:53:39,191][starVLA.training.train_starvla][INFO] - Step 4324, Loss: {'train/loss': 0.00011771792196668684, 'train/grad_norm_pre_clip': 0.1921718865633011, 'timing/data': 0.00024665000091772527, 'timing/model': 0.39294746100495104, 'learning_rate/qwen_vl_interface': 9.3919315125087e-07, 'learning_rate/action_model': 9.3919315125087e-06, 'epoch': 0.32}) +[2026-10-03 13:53:39,896][starVLA.training.train_starvla][INFO] - Step 4325, Loss: {'train/loss': 1.3201821275288239e-05, 'train/grad_norm_pre_clip': 0.004533806815743446, 'timing/data': 0.00028075999580323696, 'timing/model': 0.39404428099805955, 'learning_rate/qwen_vl_interface': 9.379150536467879e-07, 'learning_rate/action_model': 9.379150536467878e-06, 'epoch': 0.32}) +[2026-10-03 13:53:40,599][starVLA.training.train_starvla][INFO] - Step 4326, Loss: {'train/loss': 9.171485544356983e-06, 'train/grad_norm_pre_clip': 0.0028003051411360502, 'timing/data': 0.0001991500030271709, 'timing/model': 0.39262693098862655, 'learning_rate/qwen_vl_interface': 9.366387285779453e-07, 'learning_rate/action_model': 9.366387285779453e-06, 'epoch': 0.32}) +[2026-10-03 13:53:41,498][starVLA.training.train_starvla][INFO] - Step 4327, Loss: {'train/loss': 0.0009508828516118228, 'train/grad_norm_pre_clip': 0.7532260417938232, 'timing/data': 0.0003969100071117282, 'timing/model': 0.39425563999975566, 'learning_rate/qwen_vl_interface': 9.353641765689915e-07, 'learning_rate/action_model': 9.353641765689915e-06, 'epoch': 0.32}) +[2026-10-03 13:53:42,204][starVLA.training.train_starvla][INFO] - Step 4328, Loss: {'train/loss': 6.169027074065525e-06, 'train/grad_norm_pre_clip': 0.0021011934150010347, 'timing/data': 0.0002007000002777204, 'timing/model': 0.39429584000026807, 'learning_rate/qwen_vl_interface': 9.340913981438462e-07, 'learning_rate/action_model': 9.340913981438463e-06, 'epoch': 0.32}) +[2026-10-03 13:53:42,911][starVLA.training.train_starvla][INFO] - Step 4329, Loss: {'train/loss': 0.1882350593805313, 'train/grad_norm_pre_clip': 218.5670623779297, 'timing/data': 0.0002503400028217584, 'timing/model': 0.39558807999128476, 'learning_rate/qwen_vl_interface': 9.328203938257019e-07, 'learning_rate/action_model': 9.328203938257018e-06, 'epoch': 0.32}) +[2026-10-03 13:53:43,616][starVLA.training.train_starvla][INFO] - Step 4330, Loss: {'train/loss': 0.0007108363206498325, 'train/grad_norm_pre_clip': 0.4434356987476349, 'timing/data': 0.00024454000231344253, 'timing/model': 0.3940317700034939, 'learning_rate/qwen_vl_interface': 9.315511641370191e-07, 'learning_rate/action_model': 9.315511641370191e-06, 'epoch': 0.32}) +[2026-10-03 13:53:44,321][starVLA.training.train_starvla][INFO] - Step 4331, Loss: {'train/loss': 3.211646981071681e-05, 'train/grad_norm_pre_clip': 0.018062332645058632, 'timing/data': 0.0002007000002777204, 'timing/model': 0.39383684999484103, 'learning_rate/qwen_vl_interface': 9.302837095995306e-07, 'learning_rate/action_model': 9.302837095995306e-06, 'epoch': 0.32}) +[2026-10-03 13:53:45,217][starVLA.training.train_starvla][INFO] - Step 4332, Loss: {'train/loss': 0.11521903425455093, 'train/grad_norm_pre_clip': 75.6961898803711, 'timing/data': 0.00027319000218994915, 'timing/model': 0.5841530709876679, 'learning_rate/qwen_vl_interface': 9.290180307342392e-07, 'learning_rate/action_model': 9.290180307342391e-06, 'epoch': 0.32}) +[2026-10-03 13:53:45,924][starVLA.training.train_starvla][INFO] - Step 4333, Loss: {'train/loss': 1.0266694516758434e-05, 'train/grad_norm_pre_clip': 0.0024224971421062946, 'timing/data': 0.0003592099965317175, 'timing/model': 0.3945920800033491, 'learning_rate/qwen_vl_interface': 9.277541280614183e-07, 'learning_rate/action_model': 9.277541280614183e-06, 'epoch': 0.32}) +[2026-10-03 13:53:46,628][starVLA.training.train_starvla][INFO] - Step 4334, Loss: {'train/loss': 0.00018184266809839755, 'train/grad_norm_pre_clip': 0.06196651980280876, 'timing/data': 0.00020253000548109412, 'timing/model': 0.39313496100658085, 'learning_rate/qwen_vl_interface': 9.264920021006093e-07, 'learning_rate/action_model': 9.264920021006093e-06, 'epoch': 0.32}) +[2026-10-03 13:53:47,332][starVLA.training.train_starvla][INFO] - Step 4335, Loss: {'train/loss': 0.34513339400291443, 'train/grad_norm_pre_clip': 84.95236206054688, 'timing/data': 0.00020763000065926462, 'timing/model': 0.39362315999460407, 'learning_rate/qwen_vl_interface': 9.252316533706246e-07, 'learning_rate/action_model': 9.252316533706246e-06, 'epoch': 0.32}) +[2026-10-03 13:53:48,036][starVLA.training.train_starvla][INFO] - Step 4336, Loss: {'train/loss': 1.610005165275652e-05, 'train/grad_norm_pre_clip': 0.005177483893930912, 'timing/data': 0.00022525999520439655, 'timing/model': 0.39422056000330485, 'learning_rate/qwen_vl_interface': 9.239730823895469e-07, 'learning_rate/action_model': 9.23973082389547e-06, 'epoch': 0.32}) +[2026-10-03 13:53:48,742][starVLA.training.train_starvla][INFO] - Step 4337, Loss: {'train/loss': 2.365955151617527e-05, 'train/grad_norm_pre_clip': 0.008903449401259422, 'timing/data': 0.0002480799885233864, 'timing/model': 0.39503153099212795, 'learning_rate/qwen_vl_interface': 9.227162896747264e-07, 'learning_rate/action_model': 9.227162896747264e-06, 'epoch': 0.32}) +[2026-10-03 13:53:49,637][starVLA.training.train_starvla][INFO] - Step 4338, Loss: {'train/loss': 0.33603230118751526, 'train/grad_norm_pre_clip': 72.33604431152344, 'timing/data': 0.00022591000015381724, 'timing/model': 0.3936657099984586, 'learning_rate/qwen_vl_interface': 9.214612757427833e-07, 'learning_rate/action_model': 9.214612757427833e-06, 'epoch': 0.32}) +[2026-10-03 13:53:50,341][starVLA.training.train_starvla][INFO] - Step 4339, Loss: {'train/loss': 0.08257804065942764, 'train/grad_norm_pre_clip': 63.304832458496094, 'timing/data': 0.00029241999436635524, 'timing/model': 0.3928572100121528, 'learning_rate/qwen_vl_interface': 9.202080411096071e-07, 'learning_rate/action_model': 9.20208041109607e-06, 'epoch': 0.32}) +[2026-10-03 13:53:51,045][starVLA.training.train_starvla][INFO] - Step 4340, Loss: {'train/loss': 0.00020300415053498, 'train/grad_norm_pre_clip': 0.13293546438217163, 'timing/data': 0.00021290000586304814, 'timing/model': 0.39325913100037724, 'learning_rate/qwen_vl_interface': 9.189565862903543e-07, 'learning_rate/action_model': 9.189565862903543e-06, 'epoch': 0.32}) +[2026-10-03 13:53:51,750][starVLA.training.train_starvla][INFO] - Step 4341, Loss: {'train/loss': 9.692880666989367e-06, 'train/grad_norm_pre_clip': 0.004092643968760967, 'timing/data': 0.0002497100067557767, 'timing/model': 0.3942455799988238, 'learning_rate/qwen_vl_interface': 9.17706911799451e-07, 'learning_rate/action_model': 9.17706911799451e-06, 'epoch': 0.32}) +[2026-10-03 13:53:52,454][starVLA.training.train_starvla][INFO] - Step 4342, Loss: {'train/loss': 0.32587674260139465, 'train/grad_norm_pre_clip': 139.3218994140625, 'timing/data': 0.00020213000243529677, 'timing/model': 0.3924953910027398, 'learning_rate/qwen_vl_interface': 9.164590181505912e-07, 'learning_rate/action_model': 9.164590181505911e-06, 'epoch': 0.32}) +[2026-10-03 13:53:53,351][starVLA.training.train_starvla][INFO] - Step 4343, Loss: {'train/loss': 0.24221345782279968, 'train/grad_norm_pre_clip': 159.0476531982422, 'timing/data': 0.00030918000265955925, 'timing/model': 0.3931364910094999, 'learning_rate/qwen_vl_interface': 9.15212905856738e-07, 'learning_rate/action_model': 9.15212905856738e-06, 'epoch': 0.32}) +[2026-10-03 13:53:54,054][starVLA.training.train_starvla][INFO] - Step 4344, Loss: {'train/loss': 0.0003943772753700614, 'train/grad_norm_pre_clip': 0.31860676407814026, 'timing/data': 0.0003202900115866214, 'timing/model': 0.39311940100742504, 'learning_rate/qwen_vl_interface': 9.139685754301189e-07, 'learning_rate/action_model': 9.139685754301189e-06, 'epoch': 0.32}) +[2026-10-03 13:53:54,759][starVLA.training.train_starvla][INFO] - Step 4345, Loss: {'train/loss': 0.0007813764968886971, 'train/grad_norm_pre_clip': 1.2765623331069946, 'timing/data': 0.0002685799991013482, 'timing/model': 0.3947285100002773, 'learning_rate/qwen_vl_interface': 9.127260273822324e-07, 'learning_rate/action_model': 9.127260273822325e-06, 'epoch': 0.32}) +[2026-10-03 13:53:55,463][starVLA.training.train_starvla][INFO] - Step 4346, Loss: {'train/loss': 8.772326691541821e-05, 'train/grad_norm_pre_clip': 0.027492785826325417, 'timing/data': 0.00030591999529860914, 'timing/model': 0.39333637099480256, 'learning_rate/qwen_vl_interface': 9.114852622238435e-07, 'learning_rate/action_model': 9.114852622238435e-06, 'epoch': 0.32}) +[2026-10-03 13:53:56,167][starVLA.training.train_starvla][INFO] - Step 4347, Loss: {'train/loss': 2.213462721556425e-05, 'train/grad_norm_pre_clip': 0.005974426865577698, 'timing/data': 0.0002442900004098192, 'timing/model': 0.39282513099897187, 'learning_rate/qwen_vl_interface': 9.102462804649827e-07, 'learning_rate/action_model': 9.102462804649827e-06, 'epoch': 0.32}) +[2026-10-03 13:53:57,063][starVLA.training.train_starvla][INFO] - Step 4348, Loss: {'train/loss': 0.0014919491950422525, 'train/grad_norm_pre_clip': 2.0163581371307373, 'timing/data': 0.000222400005441159, 'timing/model': 0.5858620520011755, 'learning_rate/qwen_vl_interface': 9.090090826149489e-07, 'learning_rate/action_model': 9.09009082614949e-06, 'epoch': 0.32}) +[2026-10-03 13:53:57,769][starVLA.training.train_starvla][INFO] - Step 4349, Loss: {'train/loss': 0.00028850254602730274, 'train/grad_norm_pre_clip': 0.3174588680267334, 'timing/data': 0.000299129998893477, 'timing/model': 0.3944068609998794, 'learning_rate/qwen_vl_interface': 9.077736691823081e-07, 'learning_rate/action_model': 9.07773669182308e-06, 'epoch': 0.32}) +[2026-10-03 13:53:58,473][starVLA.training.train_starvla][INFO] - Step 4350, Loss: {'train/loss': 2.1605568690574728e-05, 'train/grad_norm_pre_clip': 0.006869788747280836, 'timing/data': 0.0002949900081148371, 'timing/model': 0.3936621799948625, 'learning_rate/qwen_vl_interface': 9.065400406748912e-07, 'learning_rate/action_model': 9.06540040674891e-06, 'epoch': 0.32}) +[2026-10-03 13:53:59,176][starVLA.training.train_starvla][INFO] - Step 4351, Loss: {'train/loss': 0.009150650352239609, 'train/grad_norm_pre_clip': 27.623050689697266, 'timing/data': 0.0002247500087833032, 'timing/model': 0.3927646009979071, 'learning_rate/qwen_vl_interface': 9.053081975997969e-07, 'learning_rate/action_model': 9.053081975997968e-06, 'epoch': 0.32}) +[2026-10-03 13:53:59,880][starVLA.training.train_starvla][INFO] - Step 4352, Loss: {'train/loss': 1.709758180368226e-05, 'train/grad_norm_pre_clip': 0.004028740804642439, 'timing/data': 0.0004190299951005727, 'timing/model': 0.3934435109986225, 'learning_rate/qwen_vl_interface': 9.040781404633894e-07, 'learning_rate/action_model': 9.040781404633895e-06, 'epoch': 0.32}) +[2026-10-03 13:54:00,587][starVLA.training.train_starvla][INFO] - Step 4353, Loss: {'train/loss': 5.477984814206138e-05, 'train/grad_norm_pre_clip': 0.01605687476694584, 'timing/data': 0.0002450400061206892, 'timing/model': 0.3949931410024874, 'learning_rate/qwen_vl_interface': 9.028498697712982e-07, 'learning_rate/action_model': 9.028498697712981e-06, 'epoch': 0.32}) +[2026-10-03 13:54:01,482][starVLA.training.train_starvla][INFO] - Step 4354, Loss: {'train/loss': 0.00013743234740104526, 'train/grad_norm_pre_clip': 0.1793769747018814, 'timing/data': 0.0002743399963947013, 'timing/model': 0.3938141300022835, 'learning_rate/qwen_vl_interface': 9.016233860284195e-07, 'learning_rate/action_model': 9.016233860284195e-06, 'epoch': 0.32}) +[2026-10-03 13:54:02,186][starVLA.training.train_starvla][INFO] - Step 4355, Loss: {'train/loss': 2.114225208060816e-05, 'train/grad_norm_pre_clip': 0.006179660093039274, 'timing/data': 0.0003165800007991493, 'timing/model': 0.39283070201054215, 'learning_rate/qwen_vl_interface': 9.003986897389142e-07, 'learning_rate/action_model': 9.003986897389141e-06, 'epoch': 0.32}) +[2026-10-03 13:54:02,890][starVLA.training.train_starvla][INFO] - Step 4356, Loss: {'train/loss': 0.545881450176239, 'train/grad_norm_pre_clip': 70.95219421386719, 'timing/data': 0.00029627999174408615, 'timing/model': 0.3934689210000215, 'learning_rate/qwen_vl_interface': 8.991757814062094e-07, 'learning_rate/action_model': 8.991757814062093e-06, 'epoch': 0.32}) +[2026-10-03 13:54:03,596][starVLA.training.train_starvla][INFO] - Step 4357, Loss: {'train/loss': 1.5011654795671348e-05, 'train/grad_norm_pre_clip': 0.004126856103539467, 'timing/data': 0.00026848999550566077, 'timing/model': 0.39443237100203987, 'learning_rate/qwen_vl_interface': 8.979546615329954e-07, 'learning_rate/action_model': 8.979546615329953e-06, 'epoch': 0.32}) +[2026-10-03 13:54:04,301][starVLA.training.train_starvla][INFO] - Step 4358, Loss: {'train/loss': 0.0016306074103340507, 'train/grad_norm_pre_clip': 1.2073684930801392, 'timing/data': 0.00026074999186675996, 'timing/model': 0.39252778100490104, 'learning_rate/qwen_vl_interface': 8.967353306212286e-07, 'learning_rate/action_model': 8.967353306212286e-06, 'epoch': 0.32}) +[2026-10-03 13:54:05,199][starVLA.training.train_starvla][INFO] - Step 4359, Loss: {'train/loss': 3.0398232411243953e-06, 'train/grad_norm_pre_clip': 0.0008761530625633895, 'timing/data': 0.0003604600060498342, 'timing/model': 0.3932148710009642, 'learning_rate/qwen_vl_interface': 8.955177891721303e-07, 'learning_rate/action_model': 8.955177891721303e-06, 'epoch': 0.32}) +[2026-10-03 13:54:05,903][starVLA.training.train_starvla][INFO] - Step 4360, Loss: {'train/loss': 0.5309967994689941, 'train/grad_norm_pre_clip': 78.2261962890625, 'timing/data': 0.0003241300000809133, 'timing/model': 0.3927101110020885, 'learning_rate/qwen_vl_interface': 8.943020376861858e-07, 'learning_rate/action_model': 8.943020376861857e-06, 'epoch': 0.32}) +[2026-10-03 13:54:06,609][starVLA.training.train_starvla][INFO] - Step 4361, Loss: {'train/loss': 5.595337825070601e-06, 'train/grad_norm_pre_clip': 0.0034152481239289045, 'timing/data': 0.0002497000095900148, 'timing/model': 0.394695501003298, 'learning_rate/qwen_vl_interface': 8.930880766631444e-07, 'learning_rate/action_model': 8.930880766631445e-06, 'epoch': 0.32}) +[2026-10-03 13:54:07,314][starVLA.training.train_starvla][INFO] - Step 4362, Loss: {'train/loss': 4.632929994841106e-05, 'train/grad_norm_pre_clip': 0.026318732649087906, 'timing/data': 0.0002926500019384548, 'timing/model': 0.39403870100795757, 'learning_rate/qwen_vl_interface': 8.918759066020201e-07, 'learning_rate/action_model': 8.9187590660202e-06, 'epoch': 0.32}) +[2026-10-03 13:54:08,018][starVLA.training.train_starvla][INFO] - Step 4363, Loss: {'train/loss': 1.1704509233823046e-05, 'train/grad_norm_pre_clip': 0.009756259620189667, 'timing/data': 0.0001981999957934022, 'timing/model': 0.3940584609954385, 'learning_rate/qwen_vl_interface': 8.906655280010895e-07, 'learning_rate/action_model': 8.906655280010896e-06, 'epoch': 0.32}) +[2026-10-03 13:54:08,918][starVLA.training.train_starvla][INFO] - Step 4364, Loss: {'train/loss': 2.5404211555724032e-05, 'train/grad_norm_pre_clip': 0.0064111449755728245, 'timing/data': 0.00035550999746192247, 'timing/model': 0.39393434999510646, 'learning_rate/qwen_vl_interface': 8.894569413578938e-07, 'learning_rate/action_model': 8.894569413578937e-06, 'epoch': 0.32}) +[2026-10-03 13:54:09,623][starVLA.training.train_starvla][INFO] - Step 4365, Loss: {'train/loss': 8.515749868820421e-06, 'train/grad_norm_pre_clip': 0.008327898569405079, 'timing/data': 0.0003321799886180088, 'timing/model': 0.3944566400023177, 'learning_rate/qwen_vl_interface': 8.882501471692379e-07, 'learning_rate/action_model': 8.882501471692378e-06, 'epoch': 0.32}) +[2026-10-03 13:54:10,327][starVLA.training.train_starvla][INFO] - Step 4366, Loss: {'train/loss': 0.00010818952432600781, 'train/grad_norm_pre_clip': 0.04482147470116615, 'timing/data': 0.0003070499951718375, 'timing/model': 0.3931700209941482, 'learning_rate/qwen_vl_interface': 8.870451459311894e-07, 'learning_rate/action_model': 8.870451459311893e-06, 'epoch': 0.32}) +[2026-10-03 13:54:11,030][starVLA.training.train_starvla][INFO] - Step 4367, Loss: {'train/loss': 0.003326001577079296, 'train/grad_norm_pre_clip': 3.206817626953125, 'timing/data': 0.000242260008235462, 'timing/model': 0.3931256910000229, 'learning_rate/qwen_vl_interface': 8.85841938139077e-07, 'learning_rate/action_model': 8.85841938139077e-06, 'epoch': 0.32}) +[2026-10-03 13:54:11,734][starVLA.training.train_starvla][INFO] - Step 4368, Loss: {'train/loss': 4.418165190145373e-06, 'train/grad_norm_pre_clip': 0.0013029526453465223, 'timing/data': 0.00023805999080650508, 'timing/model': 0.39370321099704597, 'learning_rate/qwen_vl_interface': 8.846405242874955e-07, 'learning_rate/action_model': 8.846405242874954e-06, 'epoch': 0.32}) +[2026-10-03 13:54:12,635][starVLA.training.train_starvla][INFO] - Step 4369, Loss: {'train/loss': 0.0224518571048975, 'train/grad_norm_pre_clip': 23.06086540222168, 'timing/data': 0.00028276999364607036, 'timing/model': 0.5887169719935628, 'learning_rate/qwen_vl_interface': 8.834409048703004e-07, 'learning_rate/action_model': 8.834409048703005e-06, 'epoch': 0.32}) +[2026-10-03 13:54:13,339][starVLA.training.train_starvla][INFO] - Step 4370, Loss: {'train/loss': 1.9235543732065707e-05, 'train/grad_norm_pre_clip': 0.022985685616731644, 'timing/data': 0.00025722000282257795, 'timing/model': 0.3935458110063337, 'learning_rate/qwen_vl_interface': 8.822430803806098e-07, 'learning_rate/action_model': 8.822430803806097e-06, 'epoch': 0.32}) +[2026-10-03 13:54:14,042][starVLA.training.train_starvla][INFO] - Step 4371, Loss: {'train/loss': 1.7083188140532002e-05, 'train/grad_norm_pre_clip': 0.004166598431766033, 'timing/data': 0.00032431998988613486, 'timing/model': 0.3924766910058679, 'learning_rate/qwen_vl_interface': 8.810470513108041e-07, 'learning_rate/action_model': 8.81047051310804e-06, 'epoch': 0.32}) +[2026-10-03 13:54:14,746][starVLA.training.train_starvla][INFO] - Step 4372, Loss: {'train/loss': 1.0736013791756704e-05, 'train/grad_norm_pre_clip': 0.003350460436195135, 'timing/data': 0.00025214000197593123, 'timing/model': 0.3932858110056259, 'learning_rate/qwen_vl_interface': 8.798528181525262e-07, 'learning_rate/action_model': 8.798528181525261e-06, 'epoch': 0.32}) +[2026-10-03 13:54:15,452][starVLA.training.train_starvla][INFO] - Step 4373, Loss: {'train/loss': 9.247042908100411e-05, 'train/grad_norm_pre_clip': 0.049769774079322815, 'timing/data': 0.0003489299997454509, 'timing/model': 0.3945175010012463, 'learning_rate/qwen_vl_interface': 8.786603813966798e-07, 'learning_rate/action_model': 8.786603813966798e-06, 'epoch': 0.32}) +[2026-10-03 13:54:16,349][starVLA.training.train_starvla][INFO] - Step 4374, Loss: {'train/loss': 7.830272807041183e-06, 'train/grad_norm_pre_clip': 0.004410612862557173, 'timing/data': 0.0002558100095484406, 'timing/model': 0.5854804410046199, 'learning_rate/qwen_vl_interface': 8.774697415334304e-07, 'learning_rate/action_model': 8.774697415334304e-06, 'epoch': 0.32}) +[2026-10-03 13:54:17,054][starVLA.training.train_starvla][INFO] - Step 4375, Loss: {'train/loss': 2.4912029402912594e-05, 'train/grad_norm_pre_clip': 0.007495960220694542, 'timing/data': 0.00022032999549992383, 'timing/model': 0.3939205810020212, 'learning_rate/qwen_vl_interface': 8.762808990522056e-07, 'learning_rate/action_model': 8.762808990522057e-06, 'epoch': 0.32}) +[2026-10-03 13:54:17,759][starVLA.training.train_starvla][INFO] - Step 4376, Loss: {'train/loss': 9.630213753553107e-05, 'train/grad_norm_pre_clip': 0.10659816116094589, 'timing/data': 0.00030931999208405614, 'timing/model': 0.39336736100085545, 'learning_rate/qwen_vl_interface': 8.750938544416934e-07, 'learning_rate/action_model': 8.750938544416934e-06, 'epoch': 0.32}) +[2026-10-03 13:54:18,475][starVLA.training.train_starvla][INFO] - Step 4377, Loss: {'train/loss': 0.17991721630096436, 'train/grad_norm_pre_clip': 193.91831970214844, 'timing/data': 0.00031569000566378236, 'timing/model': 0.40631555899744853, 'learning_rate/qwen_vl_interface': 8.739086081898437e-07, 'learning_rate/action_model': 8.739086081898437e-06, 'epoch': 0.32}) +[2026-10-03 13:54:19,177][starVLA.training.train_starvla][INFO] - Step 4378, Loss: {'train/loss': 1.3871745068172459e-05, 'train/grad_norm_pre_clip': 0.0038955514319241047, 'timing/data': 0.00031300999398808926, 'timing/model': 0.392552290999447, 'learning_rate/qwen_vl_interface': 8.727251607838647e-07, 'learning_rate/action_model': 8.727251607838647e-06, 'epoch': 0.32}) +[2026-10-03 13:54:19,878][starVLA.training.train_starvla][INFO] - Step 4379, Loss: {'train/loss': 1.7410482541890815e-05, 'train/grad_norm_pre_clip': 0.01012082677334547, 'timing/data': 0.0003342099953442812, 'timing/model': 0.3909105110069504, 'learning_rate/qwen_vl_interface': 8.715435127102281e-07, 'learning_rate/action_model': 8.715435127102281e-06, 'epoch': 0.32}) +[2026-10-03 13:54:20,780][starVLA.training.train_starvla][INFO] - Step 4380, Loss: {'train/loss': 0.0006510068196803331, 'train/grad_norm_pre_clip': 1.0771384239196777, 'timing/data': 0.00030174999847076833, 'timing/model': 0.39376209100009874, 'learning_rate/qwen_vl_interface': 8.703636644546651e-07, 'learning_rate/action_model': 8.703636644546651e-06, 'epoch': 0.32}) +[2026-10-03 13:54:21,498][starVLA.training.train_starvla][INFO] - Step 4381, Loss: {'train/loss': 3.967724842368625e-05, 'train/grad_norm_pre_clip': 0.018824229016900063, 'timing/data': 0.00020650999795179814, 'timing/model': 0.4066867590008769, 'learning_rate/qwen_vl_interface': 8.69185616502166e-07, 'learning_rate/action_model': 8.69185616502166e-06, 'epoch': 0.32}) +[2026-10-03 13:54:22,200][starVLA.training.train_starvla][INFO] - Step 4382, Loss: {'train/loss': 3.1143320029514143e-06, 'train/grad_norm_pre_clip': 0.0015836440725252032, 'timing/data': 0.0002051499905064702, 'timing/model': 0.39263809099793434, 'learning_rate/qwen_vl_interface': 8.680093693369821e-07, 'learning_rate/action_model': 8.68009369336982e-06, 'epoch': 0.32}) +[2026-10-03 13:54:22,902][starVLA.training.train_starvla][INFO] - Step 4383, Loss: {'train/loss': 3.9805709093343467e-05, 'train/grad_norm_pre_clip': 0.03670797869563103, 'timing/data': 0.00020079000387340784, 'timing/model': 0.39286323099804576, 'learning_rate/qwen_vl_interface': 8.66834923442624e-07, 'learning_rate/action_model': 8.66834923442624e-06, 'epoch': 0.32}) +[2026-10-03 13:54:23,603][starVLA.training.train_starvla][INFO] - Step 4384, Loss: {'train/loss': 7.353665750997607e-06, 'train/grad_norm_pre_clip': 0.001978835789486766, 'timing/data': 0.00029768999957013875, 'timing/model': 0.39222697100194637, 'learning_rate/qwen_vl_interface': 8.65662279301863e-07, 'learning_rate/action_model': 8.656622793018628e-06, 'epoch': 0.32}) +[2026-10-03 13:54:24,515][starVLA.training.train_starvla][INFO] - Step 4385, Loss: {'train/loss': 0.0004969379515387118, 'train/grad_norm_pre_clip': 0.7986654043197632, 'timing/data': 0.0004705400060629472, 'timing/model': 0.6011415200046031, 'learning_rate/qwen_vl_interface': 8.644914373967274e-07, 'learning_rate/action_model': 8.644914373967273e-06, 'epoch': 0.32}) +[2026-10-03 13:54:25,218][starVLA.training.train_starvla][INFO] - Step 4386, Loss: {'train/loss': 8.843697287375107e-06, 'train/grad_norm_pre_clip': 0.00231183017604053, 'timing/data': 0.00032993999775499105, 'timing/model': 0.39239581100991927, 'learning_rate/qwen_vl_interface': 8.633223982085068e-07, 'learning_rate/action_model': 8.633223982085068e-06, 'epoch': 0.32}) +[2026-10-03 13:54:25,921][starVLA.training.train_starvla][INFO] - Step 4387, Loss: {'train/loss': 6.0156227846164256e-05, 'train/grad_norm_pre_clip': 0.07198917865753174, 'timing/data': 0.0002488499885657802, 'timing/model': 0.3927890109916916, 'learning_rate/qwen_vl_interface': 8.621551622177499e-07, 'learning_rate/action_model': 8.621551622177499e-06, 'epoch': 0.32}) +[2026-10-03 13:54:26,624][starVLA.training.train_starvla][INFO] - Step 4388, Loss: {'train/loss': 5.1110082495142706e-06, 'train/grad_norm_pre_clip': 0.0023849839344620705, 'timing/data': 0.0002884290006477386, 'timing/model': 0.39311397199344356, 'learning_rate/qwen_vl_interface': 8.609897299042615e-07, 'learning_rate/action_model': 8.609897299042614e-06, 'epoch': 0.32}) +[2026-10-03 13:54:27,330][starVLA.training.train_starvla][INFO] - Step 4389, Loss: {'train/loss': 6.027456038282253e-06, 'train/grad_norm_pre_clip': 0.0014175361720845103, 'timing/data': 0.0002890299947466701, 'timing/model': 0.394998340998427, 'learning_rate/qwen_vl_interface': 8.598261017471088e-07, 'learning_rate/action_model': 8.598261017471088e-06, 'epoch': 0.32}) +[2026-10-03 13:54:28,034][starVLA.training.train_starvla][INFO] - Step 4390, Loss: {'train/loss': 0.0026834383606910706, 'train/grad_norm_pre_clip': 10.100785255432129, 'timing/data': 0.0003679899964481592, 'timing/model': 0.3927546519989846, 'learning_rate/qwen_vl_interface': 8.586642782246133e-07, 'learning_rate/action_model': 8.586642782246131e-06, 'epoch': 0.32}) +[2026-10-03 13:54:28,932][starVLA.training.train_starvla][INFO] - Step 4391, Loss: {'train/loss': 4.579769665724598e-05, 'train/grad_norm_pre_clip': 0.02082679234445095, 'timing/data': 0.0002761899959295988, 'timing/model': 0.3927058709959965, 'learning_rate/qwen_vl_interface': 8.57504259814358e-07, 'learning_rate/action_model': 8.57504259814358e-06, 'epoch': 0.32}) +[2026-10-03 13:54:29,635][starVLA.training.train_starvla][INFO] - Step 4392, Loss: {'train/loss': 0.0014676639111712575, 'train/grad_norm_pre_clip': 4.13571834564209, 'timing/data': 0.00024575000861659646, 'timing/model': 0.39282272101263516, 'learning_rate/qwen_vl_interface': 8.563460469931823e-07, 'learning_rate/action_model': 8.563460469931822e-06, 'epoch': 0.32}) +[2026-10-03 13:54:30,340][starVLA.training.train_starvla][INFO] - Step 4393, Loss: {'train/loss': 1.79695834958693e-05, 'train/grad_norm_pre_clip': 0.014414004981517792, 'timing/data': 0.000308920003590174, 'timing/model': 0.3944147510046605, 'learning_rate/qwen_vl_interface': 8.551896402371834e-07, 'learning_rate/action_model': 8.551896402371834e-06, 'epoch': 0.32}) +[2026-10-03 13:54:31,042][starVLA.training.train_starvla][INFO] - Step 4394, Loss: {'train/loss': 3.528138768160716e-05, 'train/grad_norm_pre_clip': 0.00885855033993721, 'timing/data': 0.000202240000362508, 'timing/model': 0.3914286420040298, 'learning_rate/qwen_vl_interface': 8.540350400217177e-07, 'learning_rate/action_model': 8.540350400217176e-06, 'epoch': 0.32}) +[2026-10-03 13:54:31,744][starVLA.training.train_starvla][INFO] - Step 4395, Loss: {'train/loss': 3.904469122062437e-05, 'train/grad_norm_pre_clip': 0.010214267298579216, 'timing/data': 0.0003090800018981099, 'timing/model': 0.3927969209908042, 'learning_rate/qwen_vl_interface': 8.528822468213964e-07, 'learning_rate/action_model': 8.528822468213964e-06, 'epoch': 0.32}) +[2026-10-03 13:54:32,641][starVLA.training.train_starvla][INFO] - Step 4396, Loss: {'train/loss': 2.4959392703749472e-06, 'train/grad_norm_pre_clip': 0.0006479938165284693, 'timing/data': 0.0003315700014354661, 'timing/model': 0.3928118109906791, 'learning_rate/qwen_vl_interface': 8.517312611100898e-07, 'learning_rate/action_model': 8.517312611100897e-06, 'epoch': 0.32}) +[2026-10-03 13:54:33,345][starVLA.training.train_starvla][INFO] - Step 4397, Loss: {'train/loss': 0.009863209910690784, 'train/grad_norm_pre_clip': 18.88743019104004, 'timing/data': 0.0003093599952990189, 'timing/model': 0.39412416199047584, 'learning_rate/qwen_vl_interface': 8.505820833609249e-07, 'learning_rate/action_model': 8.50582083360925e-06, 'epoch': 0.32}) +[2026-10-03 13:54:34,048][starVLA.training.train_starvla][INFO] - Step 4398, Loss: {'train/loss': 0.0011357683688402176, 'train/grad_norm_pre_clip': 2.3226752281188965, 'timing/data': 0.00029137000092305243, 'timing/model': 0.3925694620120339, 'learning_rate/qwen_vl_interface': 8.494347140462847e-07, 'learning_rate/action_model': 8.494347140462846e-06, 'epoch': 0.32}) +[2026-10-03 13:54:34,751][starVLA.training.train_starvla][INFO] - Step 4399, Loss: {'train/loss': 5.253689960227348e-05, 'train/grad_norm_pre_clip': 0.053166307508945465, 'timing/data': 0.00020788999972864985, 'timing/model': 0.3927733820019057, 'learning_rate/qwen_vl_interface': 8.482891536378105e-07, 'learning_rate/action_model': 8.482891536378105e-06, 'epoch': 0.32}) +[2026-10-03 13:54:35,454][starVLA.training.train_starvla][INFO] - Step 4400, Loss: {'train/loss': 6.4595069488859735e-06, 'train/grad_norm_pre_clip': 0.0037690524477511644, 'timing/data': 0.00031596999906469136, 'timing/model': 0.3929103009868413, 'learning_rate/qwen_vl_interface': 8.471454026063972e-07, 'learning_rate/action_model': 8.47145402606397e-06, 'epoch': 0.32}) +[2026-10-03 13:54:36,356][starVLA.training.train_starvla][INFO] - Step 4401, Loss: {'train/loss': 0.01265023648738861, 'train/grad_norm_pre_clip': 33.118743896484375, 'timing/data': 0.0003266300045652315, 'timing/model': 0.590926032004063, 'learning_rate/qwen_vl_interface': 8.460034614221988e-07, 'learning_rate/action_model': 8.460034614221988e-06, 'epoch': 0.32}) +[2026-10-03 13:54:37,059][starVLA.training.train_starvla][INFO] - Step 4402, Loss: {'train/loss': 6.608900730498135e-05, 'train/grad_norm_pre_clip': 0.0285336934030056, 'timing/data': 0.00020500000391621143, 'timing/model': 0.39254579099360853, 'learning_rate/qwen_vl_interface': 8.448633305546236e-07, 'learning_rate/action_model': 8.448633305546235e-06, 'epoch': 0.32}) +[2026-10-03 13:54:37,761][starVLA.training.train_starvla][INFO] - Step 4403, Loss: {'train/loss': 0.11941680312156677, 'train/grad_norm_pre_clip': 82.63526916503906, 'timing/data': 0.00022170000011101365, 'timing/model': 0.3929489319998538, 'learning_rate/qwen_vl_interface': 8.437250104723369e-07, 'learning_rate/action_model': 8.437250104723369e-06, 'epoch': 0.32}) +[2026-10-03 13:54:38,463][starVLA.training.train_starvla][INFO] - Step 4404, Loss: {'train/loss': 0.004892626311630011, 'train/grad_norm_pre_clip': 10.336644172668457, 'timing/data': 0.0003377499961061403, 'timing/model': 0.39204683199932333, 'learning_rate/qwen_vl_interface': 8.42588501643259e-07, 'learning_rate/action_model': 8.42588501643259e-06, 'epoch': 0.32}) +[2026-10-03 13:54:39,167][starVLA.training.train_starvla][INFO] - Step 4405, Loss: {'train/loss': 2.076256350846961e-05, 'train/grad_norm_pre_clip': 0.009242398664355278, 'timing/data': 0.00029024000104982406, 'timing/model': 0.39355015198816545, 'learning_rate/qwen_vl_interface': 8.414538045345655e-07, 'learning_rate/action_model': 8.414538045345654e-06, 'epoch': 0.32}) +[2026-10-03 13:54:39,869][starVLA.training.train_starvla][INFO] - Step 4406, Loss: {'train/loss': 0.5897672176361084, 'train/grad_norm_pre_clip': 86.94386291503906, 'timing/data': 0.00023346999660134315, 'timing/model': 0.39201783199678175, 'learning_rate/qwen_vl_interface': 8.403209196126871e-07, 'learning_rate/action_model': 8.403209196126871e-06, 'epoch': 0.32}) +[2026-10-03 13:54:40,761][starVLA.training.train_starvla][INFO] - Step 4407, Loss: {'train/loss': 1.0214484063908458e-05, 'train/grad_norm_pre_clip': 0.00414178567007184, 'timing/data': 0.00034435000270605087, 'timing/model': 0.39251717099978123, 'learning_rate/qwen_vl_interface': 8.391898473433111e-07, 'learning_rate/action_model': 8.39189847343311e-06, 'epoch': 0.32}) +[2026-10-03 13:54:41,464][starVLA.training.train_starvla][INFO] - Step 4408, Loss: {'train/loss': 0.00032078559161163867, 'train/grad_norm_pre_clip': 0.26236414909362793, 'timing/data': 0.00023470999440178275, 'timing/model': 0.3927728919952642, 'learning_rate/qwen_vl_interface': 8.380605881913778e-07, 'learning_rate/action_model': 8.380605881913779e-06, 'epoch': 0.32}) +[2026-10-03 13:54:42,167][starVLA.training.train_starvla][INFO] - Step 4409, Loss: {'train/loss': 1.0050596756627783e-05, 'train/grad_norm_pre_clip': 0.006440779194235802, 'timing/data': 0.00032575998920947313, 'timing/model': 0.3935963109979639, 'learning_rate/qwen_vl_interface': 8.369331426210832e-07, 'learning_rate/action_model': 8.369331426210832e-06, 'epoch': 0.32}) +[2026-10-03 13:54:42,871][starVLA.training.train_starvla][INFO] - Step 4410, Loss: {'train/loss': 0.6592028141021729, 'train/grad_norm_pre_clip': 48.513427734375, 'timing/data': 0.0003965299983974546, 'timing/model': 0.3925375120015815, 'learning_rate/qwen_vl_interface': 8.358075110958785e-07, 'learning_rate/action_model': 8.358075110958784e-06, 'epoch': 0.32}) +[2026-10-03 13:54:43,574][starVLA.training.train_starvla][INFO] - Step 4411, Loss: {'train/loss': 0.00040668409201316535, 'train/grad_norm_pre_clip': 1.0654157400131226, 'timing/data': 0.00026106998848263174, 'timing/model': 0.3930944809981156, 'learning_rate/qwen_vl_interface': 8.346836940784661e-07, 'learning_rate/action_model': 8.346836940784661e-06, 'epoch': 0.32}) +[2026-10-03 13:54:44,471][starVLA.training.train_starvla][INFO] - Step 4412, Loss: {'train/loss': 1.162257649411913e-05, 'train/grad_norm_pre_clip': 0.0029485749546438456, 'timing/data': 0.0002947300090454519, 'timing/model': 0.39298477199918125, 'learning_rate/qwen_vl_interface': 8.335616920308061e-07, 'learning_rate/action_model': 8.33561692030806e-06, 'epoch': 0.32}) +[2026-10-03 13:54:45,176][starVLA.training.train_starvla][INFO] - Step 4413, Loss: {'train/loss': 6.712872163916472e-06, 'train/grad_norm_pre_clip': 0.001606467878445983, 'timing/data': 0.0002866400027414784, 'timing/model': 0.39411013199423905, 'learning_rate/qwen_vl_interface': 8.324415054141111e-07, 'learning_rate/action_model': 8.32441505414111e-06, 'epoch': 0.32}) +[2026-10-03 13:54:45,880][starVLA.training.train_starvla][INFO] - Step 4414, Loss: {'train/loss': 1.962258102139458e-05, 'train/grad_norm_pre_clip': 0.006233228836208582, 'timing/data': 0.00020863999088760465, 'timing/model': 0.3932679419958731, 'learning_rate/qwen_vl_interface': 8.313231346888476e-07, 'learning_rate/action_model': 8.313231346888476e-06, 'epoch': 0.32}) +[2026-10-03 13:54:46,583][starVLA.training.train_starvla][INFO] - Step 4415, Loss: {'train/loss': 0.00036464424920268357, 'train/grad_norm_pre_clip': 0.29421043395996094, 'timing/data': 0.00024916000256780535, 'timing/model': 0.3927539909927873, 'learning_rate/qwen_vl_interface': 8.302065803147353e-07, 'learning_rate/action_model': 8.302065803147353e-06, 'epoch': 0.32}) +[2026-10-03 13:54:47,287][starVLA.training.train_starvla][INFO] - Step 4416, Loss: {'train/loss': 0.1971454620361328, 'train/grad_norm_pre_clip': 248.88308715820312, 'timing/data': 0.0002034799981629476, 'timing/model': 0.39366575100575574, 'learning_rate/qwen_vl_interface': 8.29091842750748e-07, 'learning_rate/action_model': 8.29091842750748e-06, 'epoch': 0.32}) +[2026-10-03 13:54:48,183][starVLA.training.train_starvla][INFO] - Step 4417, Loss: {'train/loss': 3.4719487302936614e-06, 'train/grad_norm_pre_clip': 0.000918080098927021, 'timing/data': 0.00025925000954885036, 'timing/model': 0.3941918210039148, 'learning_rate/qwen_vl_interface': 8.279789224551116e-07, 'learning_rate/action_model': 8.279789224551116e-06, 'epoch': 0.32}) +[2026-10-03 13:54:48,886][starVLA.training.train_starvla][INFO] - Step 4418, Loss: {'train/loss': 3.62510108971037e-05, 'train/grad_norm_pre_clip': 0.01485577691346407, 'timing/data': 0.00031655999191571027, 'timing/model': 0.39243714199983515, 'learning_rate/qwen_vl_interface': 8.268678198853058e-07, 'learning_rate/action_model': 8.268678198853059e-06, 'epoch': 0.32}) +[2026-10-03 13:54:49,589][starVLA.training.train_starvla][INFO] - Step 4419, Loss: {'train/loss': 2.6225966394122224e-06, 'train/grad_norm_pre_clip': 0.0005745694506913424, 'timing/data': 0.00024990999372676015, 'timing/model': 0.3928156420006417, 'learning_rate/qwen_vl_interface': 8.257585354980628e-07, 'learning_rate/action_model': 8.257585354980628e-06, 'epoch': 0.32}) +[2026-10-03 13:54:50,291][starVLA.training.train_starvla][INFO] - Step 4420, Loss: {'train/loss': 3.949948586523533e-05, 'train/grad_norm_pre_clip': 0.010376853868365288, 'timing/data': 0.00024808000307530165, 'timing/model': 0.39250460200128146, 'learning_rate/qwen_vl_interface': 8.246510697493681e-07, 'learning_rate/action_model': 8.246510697493682e-06, 'epoch': 0.32}) +[2026-10-03 13:54:50,995][starVLA.training.train_starvla][INFO] - Step 4421, Loss: {'train/loss': 0.40269723534584045, 'train/grad_norm_pre_clip': 65.72762298583984, 'timing/data': 0.00024681999639142305, 'timing/model': 0.39450885099358857, 'learning_rate/qwen_vl_interface': 8.235454230944584e-07, 'learning_rate/action_model': 8.235454230944584e-06, 'epoch': 0.32}) +[2026-10-03 13:54:51,893][starVLA.training.train_starvla][INFO] - Step 4422, Loss: {'train/loss': 4.947119577991543e-06, 'train/grad_norm_pre_clip': 0.001964143244549632, 'timing/data': 0.00029293000989127904, 'timing/model': 0.5883262429997558, 'learning_rate/qwen_vl_interface': 8.224415959878234e-07, 'learning_rate/action_model': 8.224415959878234e-06, 'epoch': 0.32}) +[2026-10-03 13:54:52,597][starVLA.training.train_starvla][INFO] - Step 4423, Loss: {'train/loss': 0.29211103916168213, 'train/grad_norm_pre_clip': 297.1026916503906, 'timing/data': 0.00024451001081615686, 'timing/model': 0.39289365099102724, 'learning_rate/qwen_vl_interface': 8.213395888832052e-07, 'learning_rate/action_model': 8.213395888832051e-06, 'epoch': 0.32}) +[2026-10-03 13:54:53,300][starVLA.training.train_starvla][INFO] - Step 4424, Loss: {'train/loss': 6.319476233329624e-05, 'train/grad_norm_pre_clip': 0.0169978104531765, 'timing/data': 0.00025407999055460095, 'timing/model': 0.39316767200944014, 'learning_rate/qwen_vl_interface': 8.202394022335958e-07, 'learning_rate/action_model': 8.202394022335957e-06, 'epoch': 0.32}) +[2026-10-03 13:54:54,005][starVLA.training.train_starvla][INFO] - Step 4425, Loss: {'train/loss': 1.6614767446299084e-06, 'train/grad_norm_pre_clip': 0.0004440430784597993, 'timing/data': 0.000309199996991083, 'timing/model': 0.3943908209912479, 'learning_rate/qwen_vl_interface': 8.191410364912424e-07, 'learning_rate/action_model': 8.191410364912424e-06, 'epoch': 0.32}) +[2026-10-03 13:54:54,709][starVLA.training.train_starvla][INFO] - Step 4426, Loss: {'train/loss': 0.00022564971004612744, 'train/grad_norm_pre_clip': 0.31484857201576233, 'timing/data': 0.00022251000336837023, 'timing/model': 0.3932616019883426, 'learning_rate/qwen_vl_interface': 8.1804449210764e-07, 'learning_rate/action_model': 8.1804449210764e-06, 'epoch': 0.32}) +[2026-10-03 13:54:55,607][starVLA.training.train_starvla][INFO] - Step 4427, Loss: {'train/loss': 7.025739705568412e-06, 'train/grad_norm_pre_clip': 0.003701156936585903, 'timing/data': 0.00020456999482121319, 'timing/model': 0.5875247520016273, 'learning_rate/qwen_vl_interface': 8.169497695335377e-07, 'learning_rate/action_model': 8.169497695335376e-06, 'epoch': 0.32}) +[2026-10-03 13:54:56,310][starVLA.training.train_starvla][INFO] - Step 4428, Loss: {'train/loss': 2.5257322704419494e-06, 'train/grad_norm_pre_clip': 0.0008373202290385962, 'timing/data': 0.00020294000569265336, 'timing/model': 0.3941167709999718, 'learning_rate/qwen_vl_interface': 8.158568692189342e-07, 'learning_rate/action_model': 8.158568692189342e-06, 'epoch': 0.32}) +[2026-10-03 13:54:57,014][starVLA.training.train_starvla][INFO] - Step 4429, Loss: {'train/loss': 2.775027314783074e-05, 'train/grad_norm_pre_clip': 0.00916543509811163, 'timing/data': 0.0002399600052740425, 'timing/model': 0.3943481710011838, 'learning_rate/qwen_vl_interface': 8.147657916130797e-07, 'learning_rate/action_model': 8.147657916130796e-06, 'epoch': 0.32}) +[2026-10-03 13:54:57,718][starVLA.training.train_starvla][INFO] - Step 4430, Loss: {'train/loss': 9.566198059474118e-06, 'train/grad_norm_pre_clip': 0.0020984173752367496, 'timing/data': 0.0003162900102324784, 'timing/model': 0.3930175619898364, 'learning_rate/qwen_vl_interface': 8.136765371644753e-07, 'learning_rate/action_model': 8.136765371644753e-06, 'epoch': 0.32}) +[2026-10-03 13:54:58,423][starVLA.training.train_starvla][INFO] - Step 4431, Loss: {'train/loss': 6.871859659440815e-05, 'train/grad_norm_pre_clip': 0.14016149938106537, 'timing/data': 0.00021149999520275742, 'timing/model': 0.3929332819971023, 'learning_rate/qwen_vl_interface': 8.125891063208713e-07, 'learning_rate/action_model': 8.125891063208712e-06, 'epoch': 0.32}) +[2026-10-03 13:54:59,127][starVLA.training.train_starvla][INFO] - Step 4432, Loss: {'train/loss': 5.289820364851039e-06, 'train/grad_norm_pre_clip': 0.0018895919201895595, 'timing/data': 0.0002954800002044067, 'timing/model': 0.39290214200445917, 'learning_rate/qwen_vl_interface': 8.115034995292698e-07, 'learning_rate/action_model': 8.115034995292698e-06, 'epoch': 0.32}) +[2026-10-03 13:55:00,024][starVLA.training.train_starvla][INFO] - Step 4433, Loss: {'train/loss': 2.1085093067085836e-06, 'train/grad_norm_pre_clip': 0.00048788284766487777, 'timing/data': 0.00022741999418940395, 'timing/model': 0.3939118619891815, 'learning_rate/qwen_vl_interface': 8.104197172359235e-07, 'learning_rate/action_model': 8.104197172359234e-06, 'epoch': 0.32}) +[2026-10-03 13:55:00,727][starVLA.training.train_starvla][INFO] - Step 4434, Loss: {'train/loss': 0.0001455827004974708, 'train/grad_norm_pre_clip': 0.42289090156555176, 'timing/data': 0.00020310000400058925, 'timing/model': 0.3925829420040827, 'learning_rate/qwen_vl_interface': 8.093377598863334e-07, 'learning_rate/action_model': 8.093377598863334e-06, 'epoch': 0.32}) +[2026-10-03 13:55:01,430][starVLA.training.train_starvla][INFO] - Step 4435, Loss: {'train/loss': 7.47319936635904e-05, 'train/grad_norm_pre_clip': 0.09997459501028061, 'timing/data': 0.0003064199991058558, 'timing/model': 0.39294178200361785, 'learning_rate/qwen_vl_interface': 8.082576279252519e-07, 'learning_rate/action_model': 8.082576279252518e-06, 'epoch': 0.32}) +[2026-10-03 13:55:02,135][starVLA.training.train_starvla][INFO] - Step 4436, Loss: {'train/loss': 0.0016917951870709658, 'train/grad_norm_pre_clip': 1.7965716123580933, 'timing/data': 0.00020555000810418278, 'timing/model': 0.3943528309901012, 'learning_rate/qwen_vl_interface': 8.071793217966803e-07, 'learning_rate/action_model': 8.071793217966802e-06, 'epoch': 0.32}) +[2026-10-03 13:55:02,840][starVLA.training.train_starvla][INFO] - Step 4437, Loss: {'train/loss': 0.0004598737577907741, 'train/grad_norm_pre_clip': 0.4585077464580536, 'timing/data': 0.0002515499945729971, 'timing/model': 0.3943483219918562, 'learning_rate/qwen_vl_interface': 8.061028419438691e-07, 'learning_rate/action_model': 8.06102841943869e-06, 'epoch': 0.32}) +[2026-10-03 13:55:03,734][starVLA.training.train_starvla][INFO] - Step 4438, Loss: {'train/loss': 8.985251042759046e-06, 'train/grad_norm_pre_clip': 0.002558434149250388, 'timing/data': 0.0002050299954134971, 'timing/model': 0.5832836039917311, 'learning_rate/qwen_vl_interface': 8.050281888093193e-07, 'learning_rate/action_model': 8.050281888093192e-06, 'epoch': 0.32}) +[2026-10-03 13:55:04,437][starVLA.training.train_starvla][INFO] - Step 4439, Loss: {'train/loss': 5.5133746172941756e-06, 'train/grad_norm_pre_clip': 0.0012803504941985011, 'timing/data': 0.00027681898791342974, 'timing/model': 0.39293279199046083, 'learning_rate/qwen_vl_interface': 8.039553628347794e-07, 'learning_rate/action_model': 8.039553628347793e-06, 'epoch': 0.32}) +[2026-10-03 13:55:05,141][starVLA.training.train_starvla][INFO] - Step 4440, Loss: {'train/loss': 2.996257899212651e-05, 'train/grad_norm_pre_clip': 0.017152149230241776, 'timing/data': 0.0003320499963592738, 'timing/model': 0.39301806199364364, 'learning_rate/qwen_vl_interface': 8.028843644612483e-07, 'learning_rate/action_model': 8.028843644612482e-06, 'epoch': 0.32}) +[2026-10-03 13:55:05,845][starVLA.training.train_starvla][INFO] - Step 4441, Loss: {'train/loss': 2.3690565285505727e-05, 'train/grad_norm_pre_clip': 0.00780270341783762, 'timing/data': 0.0003107800002908334, 'timing/model': 0.3941420520131942, 'learning_rate/qwen_vl_interface': 8.01815194128973e-07, 'learning_rate/action_model': 8.01815194128973e-06, 'epoch': 0.33}) +[2026-10-03 13:55:06,550][starVLA.training.train_starvla][INFO] - Step 4442, Loss: {'train/loss': 0.0001284978206967935, 'train/grad_norm_pre_clip': 0.1861834079027176, 'timing/data': 0.000370850000763312, 'timing/model': 0.3933385720010847, 'learning_rate/qwen_vl_interface': 8.007478522774481e-07, 'learning_rate/action_model': 8.00747852277448e-06, 'epoch': 0.33}) +[2026-10-03 13:55:07,253][starVLA.training.train_starvla][INFO] - Step 4443, Loss: {'train/loss': 0.5859529376029968, 'train/grad_norm_pre_clip': 148.85711669921875, 'timing/data': 0.00029584000003524125, 'timing/model': 0.3928721120028058, 'learning_rate/qwen_vl_interface': 7.996823393454197e-07, 'learning_rate/action_model': 7.996823393454196e-06, 'epoch': 0.33}) +[2026-10-03 13:55:08,147][starVLA.training.train_starvla][INFO] - Step 4444, Loss: {'train/loss': 1.1741541129595134e-05, 'train/grad_norm_pre_clip': 0.008692109026014805, 'timing/data': 0.0002598189894342795, 'timing/model': 0.3930903820000822, 'learning_rate/qwen_vl_interface': 7.986186557708773e-07, 'learning_rate/action_model': 7.986186557708771e-06, 'epoch': 0.33}) +[2026-10-03 13:55:08,852][starVLA.training.train_starvla][INFO] - Step 4445, Loss: {'train/loss': 1.7604341337573715e-05, 'train/grad_norm_pre_clip': 0.005341023672372103, 'timing/data': 0.00034980999771505594, 'timing/model': 0.3943281920073787, 'learning_rate/qwen_vl_interface': 7.975568019910617e-07, 'learning_rate/action_model': 7.975568019910617e-06, 'epoch': 0.33}) +[2026-10-03 13:55:09,556][starVLA.training.train_starvla][INFO] - Step 4446, Loss: {'train/loss': 1.814114693843294e-05, 'train/grad_norm_pre_clip': 0.006808682344853878, 'timing/data': 0.00020290999964345247, 'timing/model': 0.39302369199867826, 'learning_rate/qwen_vl_interface': 7.964967784424618e-07, 'learning_rate/action_model': 7.964967784424618e-06, 'epoch': 0.33}) +[2026-10-03 13:55:10,259][starVLA.training.train_starvla][INFO] - Step 4447, Loss: {'train/loss': 1.2039643479511142e-05, 'train/grad_norm_pre_clip': 0.002650781301781535, 'timing/data': 0.00029775001166854054, 'timing/model': 0.39241358199797105, 'learning_rate/qwen_vl_interface': 7.954385855608129e-07, 'learning_rate/action_model': 7.95438585560813e-06, 'epoch': 0.33}) +[2026-10-03 13:55:10,962][starVLA.training.train_starvla][INFO] - Step 4448, Loss: {'train/loss': 0.013802935369312763, 'train/grad_norm_pre_clip': 78.12879943847656, 'timing/data': 0.00030806999711785465, 'timing/model': 0.39286885199544486, 'learning_rate/qwen_vl_interface': 7.943822237810974e-07, 'learning_rate/action_model': 7.943822237810975e-06, 'epoch': 0.33}) +[2026-10-03 13:55:11,860][starVLA.training.train_starvla][INFO] - Step 4449, Loss: {'train/loss': 1.8423956134938635e-05, 'train/grad_norm_pre_clip': 0.00747543154284358, 'timing/data': 0.0002781000075628981, 'timing/model': 0.3939472019992536, 'learning_rate/qwen_vl_interface': 7.93327693537547e-07, 'learning_rate/action_model': 7.93327693537547e-06, 'epoch': 0.33}) +[2026-10-03 13:55:12,563][starVLA.training.train_starvla][INFO] - Step 4450, Loss: {'train/loss': 3.358339745318517e-05, 'train/grad_norm_pre_clip': 0.02121993899345398, 'timing/data': 0.0003371499915374443, 'timing/model': 0.3923618720000377, 'learning_rate/qwen_vl_interface': 7.92274995263638e-07, 'learning_rate/action_model': 7.92274995263638e-06, 'epoch': 0.33}) +[2026-10-03 13:55:13,265][starVLA.training.train_starvla][INFO] - Step 4451, Loss: {'train/loss': 0.3297211229801178, 'train/grad_norm_pre_clip': 195.86856079101562, 'timing/data': 0.00026234000688418746, 'timing/model': 0.3923484920087503, 'learning_rate/qwen_vl_interface': 7.912241293920958e-07, 'learning_rate/action_model': 7.912241293920959e-06, 'epoch': 0.33}) +[2026-10-03 13:55:13,968][starVLA.training.train_starvla][INFO] - Step 4452, Loss: {'train/loss': 5.200467967370059e-06, 'train/grad_norm_pre_clip': 0.001379946363158524, 'timing/data': 0.00034857999708037823, 'timing/model': 0.39271902199834585, 'learning_rate/qwen_vl_interface': 7.90175096354891e-07, 'learning_rate/action_model': 7.90175096354891e-06, 'epoch': 0.33}) +[2026-10-03 13:55:14,672][starVLA.training.train_starvla][INFO] - Step 4453, Loss: {'train/loss': 9.171443707600702e-06, 'train/grad_norm_pre_clip': 0.00350375147536397, 'timing/data': 0.0002525000018067658, 'timing/model': 0.3938577520020772, 'learning_rate/qwen_vl_interface': 7.891278965832431e-07, 'learning_rate/action_model': 7.891278965832432e-06, 'epoch': 0.33}) +[2026-10-03 13:55:15,574][starVLA.training.train_starvla][INFO] - Step 4454, Loss: {'train/loss': 0.22723513841629028, 'train/grad_norm_pre_clip': 460.01031494140625, 'timing/data': 0.0002870800090022385, 'timing/model': 0.5909737930051051, 'learning_rate/qwen_vl_interface': 7.880825305076138e-07, 'learning_rate/action_model': 7.880825305076139e-06, 'epoch': 0.33}) +[2026-10-03 13:55:16,277][starVLA.training.train_starvla][INFO] - Step 4455, Loss: {'train/loss': 7.673792424611747e-06, 'train/grad_norm_pre_clip': 0.002909277565777302, 'timing/data': 0.0002884400018956512, 'timing/model': 0.3929969419987174, 'learning_rate/qwen_vl_interface': 7.870389985577148e-07, 'learning_rate/action_model': 7.870389985577148e-06, 'epoch': 0.33}) +[2026-10-03 13:55:16,981][starVLA.training.train_starvla][INFO] - Step 4456, Loss: {'train/loss': 0.00011935883230762556, 'train/grad_norm_pre_clip': 0.08731845021247864, 'timing/data': 0.00026208900089841336, 'timing/model': 0.3938054820027901, 'learning_rate/qwen_vl_interface': 7.859973011625028e-07, 'learning_rate/action_model': 7.859973011625026e-06, 'epoch': 0.33}) +[2026-10-03 13:55:17,683][starVLA.training.train_starvla][INFO] - Step 4457, Loss: {'train/loss': 9.693041647551581e-06, 'train/grad_norm_pre_clip': 0.002470455365255475, 'timing/data': 0.00025004999770317227, 'timing/model': 0.39387346200237516, 'learning_rate/qwen_vl_interface': 7.849574387501801e-07, 'learning_rate/action_model': 7.8495743875018e-06, 'epoch': 0.33}) +[2026-10-03 13:55:18,387][starVLA.training.train_starvla][INFO] - Step 4458, Loss: {'train/loss': 0.0003089190577156842, 'train/grad_norm_pre_clip': 0.2796754837036133, 'timing/data': 0.00026079000963363796, 'timing/model': 0.3930667520035058, 'learning_rate/qwen_vl_interface': 7.839194117481952e-07, 'learning_rate/action_model': 7.839194117481952e-06, 'epoch': 0.33}) +[2026-10-03 13:55:19,092][starVLA.training.train_starvla][INFO] - Step 4459, Loss: {'train/loss': 9.387723025611194e-07, 'train/grad_norm_pre_clip': 0.0002428166480967775, 'timing/data': 0.00028985000972170383, 'timing/model': 0.39350605200161226, 'learning_rate/qwen_vl_interface': 7.828832205832418e-07, 'learning_rate/action_model': 7.828832205832418e-06, 'epoch': 0.33}) +[2026-10-03 13:55:19,989][starVLA.training.train_starvla][INFO] - Step 4460, Loss: {'train/loss': 1.362589500786271e-05, 'train/grad_norm_pre_clip': 0.01019354909658432, 'timing/data': 0.00022740999702364206, 'timing/model': 0.39443905199004803, 'learning_rate/qwen_vl_interface': 7.818488656812581e-07, 'learning_rate/action_model': 7.818488656812582e-06, 'epoch': 0.33}) +[2026-10-03 13:55:20,694][starVLA.training.train_starvla][INFO] - Step 4461, Loss: {'train/loss': 2.6164325390709564e-05, 'train/grad_norm_pre_clip': 0.008065723814070225, 'timing/data': 0.0002838299988070503, 'timing/model': 0.3945400119991973, 'learning_rate/qwen_vl_interface': 7.808163474674291e-07, 'learning_rate/action_model': 7.80816347467429e-06, 'epoch': 0.33}) +[2026-10-03 13:55:21,397][starVLA.training.train_starvla][INFO] - Step 4462, Loss: {'train/loss': 0.0006366704474203289, 'train/grad_norm_pre_clip': 1.3459231853485107, 'timing/data': 0.0003053800028283149, 'timing/model': 0.3928315019875299, 'learning_rate/qwen_vl_interface': 7.797856663661836e-07, 'learning_rate/action_model': 7.797856663661835e-06, 'epoch': 0.33}) +[2026-10-03 13:55:22,100][starVLA.training.train_starvla][INFO] - Step 4463, Loss: {'train/loss': 1.348553382740647e-06, 'train/grad_norm_pre_clip': 0.00024791996111162007, 'timing/data': 0.00032036000629886985, 'timing/model': 0.3926765620126389, 'learning_rate/qwen_vl_interface': 7.787568228011959e-07, 'learning_rate/action_model': 7.78756822801196e-06, 'epoch': 0.33}) +[2026-10-03 13:55:22,804][starVLA.training.train_starvla][INFO] - Step 4464, Loss: {'train/loss': 1.890801104309503e-05, 'train/grad_norm_pre_clip': 0.006379789672791958, 'timing/data': 0.00028905000363010913, 'timing/model': 0.3929336319997674, 'learning_rate/qwen_vl_interface': 7.777298171953858e-07, 'learning_rate/action_model': 7.777298171953857e-06, 'epoch': 0.33}) +[2026-10-03 13:55:23,703][starVLA.training.train_starvla][INFO] - Step 4465, Loss: {'train/loss': 0.0017585387686267495, 'train/grad_norm_pre_clip': 9.35890007019043, 'timing/data': 0.0002824499970301986, 'timing/model': 0.3939294520096155, 'learning_rate/qwen_vl_interface': 7.767046499709144e-07, 'learning_rate/action_model': 7.767046499709143e-06, 'epoch': 0.33}) +[2026-10-03 13:55:24,408][starVLA.training.train_starvla][INFO] - Step 4466, Loss: {'train/loss': 0.0009896665578708053, 'train/grad_norm_pre_clip': 0.8943018317222595, 'timing/data': 0.0002755399909801781, 'timing/model': 0.3923843720112927, 'learning_rate/qwen_vl_interface': 7.756813215491904e-07, 'learning_rate/action_model': 7.756813215491904e-06, 'epoch': 0.33}) +[2026-10-03 13:55:25,111][starVLA.training.train_starvla][INFO] - Step 4467, Loss: {'train/loss': 0.0004795624117832631, 'train/grad_norm_pre_clip': 0.9796932935714722, 'timing/data': 0.000208780009415932, 'timing/model': 0.39286791199992876, 'learning_rate/qwen_vl_interface': 7.746598323508655e-07, 'learning_rate/action_model': 7.746598323508655e-06, 'epoch': 0.33}) +[2026-10-03 13:55:25,815][starVLA.training.train_starvla][INFO] - Step 4468, Loss: {'train/loss': 2.754942761384882e-05, 'train/grad_norm_pre_clip': 0.010730414651334286, 'timing/data': 0.0003061699972022325, 'timing/model': 0.3932245619944297, 'learning_rate/qwen_vl_interface': 7.736401827958351e-07, 'learning_rate/action_model': 7.73640182795835e-06, 'epoch': 0.33}) +[2026-10-03 13:55:26,520][starVLA.training.train_starvla][INFO] - Step 4469, Loss: {'train/loss': 2.0629660866688937e-05, 'train/grad_norm_pre_clip': 0.010983132757246494, 'timing/data': 0.00023836000764276832, 'timing/model': 0.39509789198928047, 'learning_rate/qwen_vl_interface': 7.726223733032382e-07, 'learning_rate/action_model': 7.72622373303238e-06, 'epoch': 0.33}) +[2026-10-03 13:55:27,420][starVLA.training.train_starvla][INFO] - Step 4470, Loss: {'train/loss': 1.998856896534562e-05, 'train/grad_norm_pre_clip': 0.01668703369796276, 'timing/data': 0.00027581999893300235, 'timing/model': 0.39377252299163956, 'learning_rate/qwen_vl_interface': 7.716064042914584e-07, 'learning_rate/action_model': 7.716064042914584e-06, 'epoch': 0.33}) +[2026-10-03 13:55:28,123][starVLA.training.train_starvla][INFO] - Step 4471, Loss: {'train/loss': 0.0001743865868775174, 'train/grad_norm_pre_clip': 0.10902731865644455, 'timing/data': 0.00023116999363992363, 'timing/model': 0.3930551029916387, 'learning_rate/qwen_vl_interface': 7.705922761781226e-07, 'learning_rate/action_model': 7.705922761781226e-06, 'epoch': 0.33}) +[2026-10-03 13:55:28,830][starVLA.training.train_starvla][INFO] - Step 4472, Loss: {'train/loss': 4.400958641781472e-05, 'train/grad_norm_pre_clip': 0.02556147426366806, 'timing/data': 0.00021285000548232347, 'timing/model': 0.394629411995993, 'learning_rate/qwen_vl_interface': 7.695799893800996e-07, 'learning_rate/action_model': 7.695799893800996e-06, 'epoch': 0.33}) +[2026-10-03 13:55:29,538][starVLA.training.train_starvla][INFO] - Step 4473, Loss: {'train/loss': 1.2963993185621803e-06, 'train/grad_norm_pre_clip': 0.0002615377597976476, 'timing/data': 0.0003163999936077744, 'timing/model': 0.39585628200438805, 'learning_rate/qwen_vl_interface': 7.685695443135028e-07, 'learning_rate/action_model': 7.685695443135027e-06, 'epoch': 0.33}) +[2026-10-03 13:55:30,244][starVLA.training.train_starvla][INFO] - Step 4474, Loss: {'train/loss': 1.5042011000332423e-05, 'train/grad_norm_pre_clip': 0.003878904040902853, 'timing/data': 0.00027828999736811966, 'timing/model': 0.3942894520005211, 'learning_rate/qwen_vl_interface': 7.675609413936888e-07, 'learning_rate/action_model': 7.675609413936888e-06, 'epoch': 0.33}) +[2026-10-03 13:55:31,143][starVLA.training.train_starvla][INFO] - Step 4475, Loss: {'train/loss': 0.0017529047327116132, 'train/grad_norm_pre_clip': 2.554150342941284, 'timing/data': 0.00029132999770808965, 'timing/model': 0.5874247640022077, 'learning_rate/qwen_vl_interface': 7.665541810352541e-07, 'learning_rate/action_model': 7.66554181035254e-06, 'epoch': 0.33}) +[2026-10-03 13:55:31,850][starVLA.training.train_starvla][INFO] - Step 4476, Loss: {'train/loss': 6.285961717367172e-05, 'train/grad_norm_pre_clip': 0.02021072246134281, 'timing/data': 0.00027535000117495656, 'timing/model': 0.3953716319956584, 'learning_rate/qwen_vl_interface': 7.655492636520415e-07, 'learning_rate/action_model': 7.655492636520414e-06, 'epoch': 0.33}) +[2026-10-03 13:55:32,557][starVLA.training.train_starvla][INFO] - Step 4477, Loss: {'train/loss': 0.534828245639801, 'train/grad_norm_pre_clip': 44.09259796142578, 'timing/data': 0.0002827400021487847, 'timing/model': 0.3954093020001892, 'learning_rate/qwen_vl_interface': 7.64546189657134e-07, 'learning_rate/action_model': 7.645461896571339e-06, 'epoch': 0.33}) +[2026-10-03 13:55:33,259][starVLA.training.train_starvla][INFO] - Step 4478, Loss: {'train/loss': 1.5794728824403137e-05, 'train/grad_norm_pre_clip': 0.015352604910731316, 'timing/data': 0.0003338799870107323, 'timing/model': 0.39222949200484436, 'learning_rate/qwen_vl_interface': 7.635449594628571e-07, 'learning_rate/action_model': 7.635449594628571e-06, 'epoch': 0.33}) +[2026-10-03 13:55:33,963][starVLA.training.train_starvla][INFO] - Step 4479, Loss: {'train/loss': 1.4326754353533033e-05, 'train/grad_norm_pre_clip': 0.01187102310359478, 'timing/data': 0.00026777001039590687, 'timing/model': 0.3927687929972308, 'learning_rate/qwen_vl_interface': 7.625455734807803e-07, 'learning_rate/action_model': 7.625455734807803e-06, 'epoch': 0.33}) +[2026-10-03 13:55:34,861][starVLA.training.train_starvla][INFO] - Step 4480, Loss: {'train/loss': 2.4437822503386997e-06, 'train/grad_norm_pre_clip': 0.0005775059689767659, 'timing/data': 0.00028172999736852944, 'timing/model': 0.588644203991862, 'learning_rate/qwen_vl_interface': 7.615480321217116e-07, 'learning_rate/action_model': 7.615480321217117e-06, 'epoch': 0.33}) +[2026-10-03 13:55:35,566][starVLA.training.train_starvla][INFO] - Step 4481, Loss: {'train/loss': 0.0007544615073129535, 'train/grad_norm_pre_clip': 1.1696199178695679, 'timing/data': 0.00024508999194949865, 'timing/model': 0.39473697199719027, 'learning_rate/qwen_vl_interface': 7.605523357957028e-07, 'learning_rate/action_model': 7.605523357957028e-06, 'epoch': 0.33}) +[2026-10-03 13:55:36,268][starVLA.training.train_starvla][INFO] - Step 4482, Loss: {'train/loss': 1.5220694876916241e-05, 'train/grad_norm_pre_clip': 0.012218897230923176, 'timing/data': 0.0002073600044241175, 'timing/model': 0.39203838200774044, 'learning_rate/qwen_vl_interface': 7.59558484912048e-07, 'learning_rate/action_model': 7.59558484912048e-06, 'epoch': 0.33}) +[2026-10-03 13:55:36,971][starVLA.training.train_starvla][INFO] - Step 4483, Loss: {'train/loss': 7.3757678364927415e-06, 'train/grad_norm_pre_clip': 0.0025208136066794395, 'timing/data': 0.00025943000218831003, 'timing/model': 0.3925804629980121, 'learning_rate/qwen_vl_interface': 7.585664798792807e-07, 'learning_rate/action_model': 7.585664798792807e-06, 'epoch': 0.33}) +[2026-10-03 13:55:37,676][starVLA.training.train_starvla][INFO] - Step 4484, Loss: {'train/loss': 1.028121005219873e-05, 'train/grad_norm_pre_clip': 0.003085161093622446, 'timing/data': 0.00028727999597322196, 'timing/model': 0.3929631320061162, 'learning_rate/qwen_vl_interface': 7.575763211051779e-07, 'learning_rate/action_model': 7.575763211051778e-06, 'epoch': 0.33}) +[2026-10-03 13:55:38,380][starVLA.training.train_starvla][INFO] - Step 4485, Loss: {'train/loss': 0.00015351816546171904, 'train/grad_norm_pre_clip': 0.07043413072824478, 'timing/data': 0.00030772999161854386, 'timing/model': 0.3940423419990111, 'learning_rate/qwen_vl_interface': 7.565880089967553e-07, 'learning_rate/action_model': 7.565880089967553e-06, 'epoch': 0.33}) +[2026-10-03 13:55:39,276][starVLA.training.train_starvla][INFO] - Step 4486, Loss: {'train/loss': 2.250054876640206e-06, 'train/grad_norm_pre_clip': 0.0008965118322521448, 'timing/data': 0.00021795999782625586, 'timing/model': 0.39302089299599174, 'learning_rate/qwen_vl_interface': 7.556015439602725e-07, 'learning_rate/action_model': 7.556015439602725e-06, 'epoch': 0.33}) +[2026-10-03 13:55:39,979][starVLA.training.train_starvla][INFO] - Step 4487, Loss: {'train/loss': 1.4683689187222626e-05, 'train/grad_norm_pre_clip': 0.016483744606375694, 'timing/data': 0.0003364000003784895, 'timing/model': 0.39253758201084565, 'learning_rate/qwen_vl_interface': 7.546169264012252e-07, 'learning_rate/action_model': 7.5461692640122525e-06, 'epoch': 0.33}) +[2026-10-03 13:55:40,681][starVLA.training.train_starvla][INFO] - Step 4488, Loss: {'train/loss': 3.8275094993878156e-05, 'train/grad_norm_pre_clip': 0.03499313443899155, 'timing/data': 0.0002018000086536631, 'timing/model': 0.3927722230000654, 'learning_rate/qwen_vl_interface': 7.536341567243544e-07, 'learning_rate/action_model': 7.536341567243543e-06, 'epoch': 0.33}) +[2026-10-03 13:55:41,385][starVLA.training.train_starvla][INFO] - Step 4489, Loss: {'train/loss': 6.720159035467077e-06, 'train/grad_norm_pre_clip': 0.006074876524507999, 'timing/data': 0.0002039499959209934, 'timing/model': 0.39404524200654123, 'learning_rate/qwen_vl_interface': 7.526532353336395e-07, 'learning_rate/action_model': 7.526532353336395e-06, 'epoch': 0.33}) +[2026-10-03 13:55:42,088][starVLA.training.train_starvla][INFO] - Step 4490, Loss: {'train/loss': 2.2947683646634687e-06, 'train/grad_norm_pre_clip': 0.0008970748167484999, 'timing/data': 0.000252609999733977, 'timing/model': 0.392720742005622, 'learning_rate/qwen_vl_interface': 7.516741626322994e-07, 'learning_rate/action_model': 7.516741626322994e-06, 'epoch': 0.33}) +[2026-10-03 13:55:42,982][starVLA.training.train_starvla][INFO] - Step 4491, Loss: {'train/loss': 2.071255721602938e-06, 'train/grad_norm_pre_clip': 0.0006252839812077582, 'timing/data': 0.0002858700026990846, 'timing/model': 0.5844838240009267, 'learning_rate/qwen_vl_interface': 7.506969390227945e-07, 'learning_rate/action_model': 7.506969390227945e-06, 'epoch': 0.33}) +[2026-10-03 13:55:43,685][starVLA.training.train_starvla][INFO] - Step 4492, Loss: {'train/loss': 0.00013223866699263453, 'train/grad_norm_pre_clip': 0.15484555065631866, 'timing/data': 0.00020213999960105866, 'timing/model': 0.3928412120003486, 'learning_rate/qwen_vl_interface': 7.497215649068244e-07, 'learning_rate/action_model': 7.497215649068244e-06, 'epoch': 0.33}) +[2026-10-03 13:55:44,389][starVLA.training.train_starvla][INFO] - Step 4493, Loss: {'train/loss': 9.741295798448846e-05, 'train/grad_norm_pre_clip': 0.37800490856170654, 'timing/data': 0.00029907000134699047, 'timing/model': 0.39355767201050185, 'learning_rate/qwen_vl_interface': 7.48748040685329e-07, 'learning_rate/action_model': 7.487480406853289e-06, 'epoch': 0.33}) +[2026-10-03 13:55:45,092][starVLA.training.train_starvla][INFO] - Step 4494, Loss: {'train/loss': 0.00010780405136756599, 'train/grad_norm_pre_clip': 0.13184599578380585, 'timing/data': 0.000209060002816841, 'timing/model': 0.3932249220088124, 'learning_rate/qwen_vl_interface': 7.477763667584867e-07, 'learning_rate/action_model': 7.477763667584866e-06, 'epoch': 0.33}) +[2026-10-03 13:55:45,795][starVLA.training.train_starvla][INFO] - Step 4495, Loss: {'train/loss': 2.3026646886137314e-05, 'train/grad_norm_pre_clip': 0.01589449867606163, 'timing/data': 0.00031867000507190824, 'timing/model': 0.3923096130019985, 'learning_rate/qwen_vl_interface': 7.468065435257159e-07, 'learning_rate/action_model': 7.468065435257159e-06, 'epoch': 0.33}) +[2026-10-03 13:55:46,498][starVLA.training.train_starvla][INFO] - Step 4496, Loss: {'train/loss': 7.1225913416128606e-06, 'train/grad_norm_pre_clip': 0.0016880870098248124, 'timing/data': 0.00019610999152064323, 'timing/model': 0.39362461199925747, 'learning_rate/qwen_vl_interface': 7.458385713856745e-07, 'learning_rate/action_model': 7.458385713856744e-06, 'epoch': 0.33}) +[2026-10-03 13:55:47,392][starVLA.training.train_starvla][INFO] - Step 4497, Loss: {'train/loss': 0.0004123889375478029, 'train/grad_norm_pre_clip': 1.4767299890518188, 'timing/data': 0.0002620699960971251, 'timing/model': 0.3940761720004957, 'learning_rate/qwen_vl_interface': 7.448724507362605e-07, 'learning_rate/action_model': 7.448724507362605e-06, 'epoch': 0.33}) +[2026-10-03 13:55:48,095][starVLA.training.train_starvla][INFO] - Step 4498, Loss: {'train/loss': 0.444661021232605, 'train/grad_norm_pre_clip': 75.13783264160156, 'timing/data': 0.00021183000353630632, 'timing/model': 0.3927305420074845, 'learning_rate/qwen_vl_interface': 7.439081819746078e-07, 'learning_rate/action_model': 7.4390818197460775e-06, 'epoch': 0.33}) +[2026-10-03 13:55:48,798][starVLA.training.train_starvla][INFO] - Step 4499, Loss: {'train/loss': 2.1978255972499028e-05, 'train/grad_norm_pre_clip': 0.00702810101211071, 'timing/data': 0.0002757800102699548, 'timing/model': 0.3926115630019922, 'learning_rate/qwen_vl_interface': 7.42945765497091e-07, 'learning_rate/action_model': 7.42945765497091e-06, 'epoch': 0.33}) +[2026-10-03 13:55:49,500][starVLA.training.train_starvla][INFO] - Step 4500, Loss: {'train/loss': 2.824070725182537e-05, 'train/grad_norm_pre_clip': 0.011981697753071785, 'timing/data': 0.0002478300011716783, 'timing/model': 0.39254303299821913, 'learning_rate/qwen_vl_interface': 7.41985201699324e-07, 'learning_rate/action_model': 7.41985201699324e-06, 'epoch': 0.33}) +[2026-10-03 13:55:50,205][starVLA.training.train_starvla][INFO] - Step 4501, Loss: {'train/loss': 0.00035067935823462903, 'train/grad_norm_pre_clip': 0.5573644042015076, 'timing/data': 0.00024217000463977456, 'timing/model': 0.3941111930034822, 'learning_rate/qwen_vl_interface': 7.410264909761581e-07, 'learning_rate/action_model': 7.41026490976158e-06, 'epoch': 0.33}) +[2026-10-03 13:55:51,103][starVLA.training.train_starvla][INFO] - Step 4502, Loss: {'train/loss': 1.5191175407380797e-05, 'train/grad_norm_pre_clip': 0.004157084506005049, 'timing/data': 0.0003708000003825873, 'timing/model': 0.39257503299450036, 'learning_rate/qwen_vl_interface': 7.40069633721683e-07, 'learning_rate/action_model': 7.400696337216829e-06, 'epoch': 0.33}) +[2026-10-03 13:55:51,805][starVLA.training.train_starvla][INFO] - Step 4503, Loss: {'train/loss': 0.8629119992256165, 'train/grad_norm_pre_clip': 22.750093460083008, 'timing/data': 0.00025117000041063875, 'timing/model': 0.3925362129957648, 'learning_rate/qwen_vl_interface': 7.391146303292262e-07, 'learning_rate/action_model': 7.391146303292262e-06, 'epoch': 0.33}) +[2026-10-03 13:55:52,508][starVLA.training.train_starvla][INFO] - Step 4504, Loss: {'train/loss': 3.7401246117951814e-06, 'train/grad_norm_pre_clip': 0.0010813630651682615, 'timing/data': 0.0002013589983107522, 'timing/model': 0.39271568300318904, 'learning_rate/qwen_vl_interface': 7.381614811913538e-07, 'learning_rate/action_model': 7.381614811913537e-06, 'epoch': 0.33}) +[2026-10-03 13:55:53,212][starVLA.training.train_starvla][INFO] - Step 4505, Loss: {'train/loss': 1.1815945072157774e-05, 'train/grad_norm_pre_clip': 0.004144591279327869, 'timing/data': 0.0003177799953846261, 'timing/model': 0.393974912003614, 'learning_rate/qwen_vl_interface': 7.372101866998693e-07, 'learning_rate/action_model': 7.372101866998693e-06, 'epoch': 0.33}) +[2026-10-03 13:55:53,915][starVLA.training.train_starvla][INFO] - Step 4506, Loss: {'train/loss': 4.255972817190923e-05, 'train/grad_norm_pre_clip': 0.046375200152397156, 'timing/data': 0.00020583000150509179, 'timing/model': 0.39290066200192086, 'learning_rate/qwen_vl_interface': 7.362607472458147e-07, 'learning_rate/action_model': 7.362607472458147e-06, 'epoch': 0.33}) +[2026-10-03 13:55:54,811][starVLA.training.train_starvla][INFO] - Step 4507, Loss: {'train/loss': 0.00179194170050323, 'train/grad_norm_pre_clip': 7.4262285232543945, 'timing/data': 0.0002209999947808683, 'timing/model': 0.5859324440098135, 'learning_rate/qwen_vl_interface': 7.353131632194682e-07, 'learning_rate/action_model': 7.3531316321946816e-06, 'epoch': 0.33}) +[2026-10-03 13:55:55,513][starVLA.training.train_starvla][INFO] - Step 4508, Loss: {'train/loss': 1.6546782717341557e-05, 'train/grad_norm_pre_clip': 0.010660962201654911, 'timing/data': 0.00020893999317195266, 'timing/model': 0.39306131200282834, 'learning_rate/qwen_vl_interface': 7.343674350103455e-07, 'learning_rate/action_model': 7.343674350103455e-06, 'epoch': 0.33}) +[2026-10-03 13:55:56,217][starVLA.training.train_starvla][INFO] - Step 4509, Loss: {'train/loss': 2.8038357413606718e-05, 'train/grad_norm_pre_clip': 0.023369867354631424, 'timing/data': 0.00030371999309863895, 'timing/model': 0.3936662519990932, 'learning_rate/qwen_vl_interface': 7.334235630071995e-07, 'learning_rate/action_model': 7.334235630071995e-06, 'epoch': 0.33}) +[2026-10-03 13:55:56,919][starVLA.training.train_starvla][INFO] - Step 4510, Loss: {'train/loss': 9.810896153794602e-05, 'train/grad_norm_pre_clip': 0.16827256977558136, 'timing/data': 0.00030097000126261264, 'timing/model': 0.39245141299034003, 'learning_rate/qwen_vl_interface': 7.324815475980207e-07, 'learning_rate/action_model': 7.324815475980207e-06, 'epoch': 0.33}) +[2026-10-03 13:55:57,621][starVLA.training.train_starvla][INFO] - Step 4511, Loss: {'train/loss': 1.0452751666889526e-05, 'train/grad_norm_pre_clip': 0.0028449627570807934, 'timing/data': 0.0002884700079448521, 'timing/model': 0.3922081829950912, 'learning_rate/qwen_vl_interface': 7.315413891700363e-07, 'learning_rate/action_model': 7.315413891700362e-06, 'epoch': 0.33}) +[2026-10-03 13:55:58,323][starVLA.training.train_starvla][INFO] - Step 4512, Loss: {'train/loss': 1.7285303783864947e-06, 'train/grad_norm_pre_clip': 0.0004565175622701645, 'timing/data': 0.0002473899949109182, 'timing/model': 0.39245451299939305, 'learning_rate/qwen_vl_interface': 7.306030881097092e-07, 'learning_rate/action_model': 7.3060308810970925e-06, 'epoch': 0.33}) +[2026-10-03 13:55:59,217][starVLA.training.train_starvla][INFO] - Step 4513, Loss: {'train/loss': 0.4141312539577484, 'train/grad_norm_pre_clip': 35.39981460571289, 'timing/data': 0.00028326999745331705, 'timing/model': 0.3938503719982691, 'learning_rate/qwen_vl_interface': 7.296666448027406e-07, 'learning_rate/action_model': 7.296666448027406e-06, 'epoch': 0.33}) +[2026-10-03 13:55:59,920][starVLA.training.train_starvla][INFO] - Step 4514, Loss: {'train/loss': 2.0636383851524442e-05, 'train/grad_norm_pre_clip': 0.010978402569890022, 'timing/data': 0.0003061700117541477, 'timing/model': 0.3933839630044531, 'learning_rate/qwen_vl_interface': 7.287320596340664e-07, 'learning_rate/action_model': 7.287320596340663e-06, 'epoch': 0.33}) +[2026-10-03 13:56:00,622][starVLA.training.train_starvla][INFO] - Step 4515, Loss: {'train/loss': 3.941303930332651e-06, 'train/grad_norm_pre_clip': 0.0012184163788333535, 'timing/data': 0.00021227000979706645, 'timing/model': 0.39185204201203305, 'learning_rate/qwen_vl_interface': 7.277993329878593e-07, 'learning_rate/action_model': 7.277993329878594e-06, 'epoch': 0.33}) +[2026-10-03 13:56:01,326][starVLA.training.train_starvla][INFO] - Step 4516, Loss: {'train/loss': 0.3436664938926697, 'train/grad_norm_pre_clip': 129.07669067382812, 'timing/data': 0.00031908000528346747, 'timing/model': 0.3929020119976485, 'learning_rate/qwen_vl_interface': 7.268684652475277e-07, 'learning_rate/action_model': 7.268684652475277e-06, 'epoch': 0.33}) +[2026-10-03 13:56:02,031][starVLA.training.train_starvla][INFO] - Step 4517, Loss: {'train/loss': 0.0006881090812385082, 'train/grad_norm_pre_clip': 1.097900390625, 'timing/data': 0.00027303000388201326, 'timing/model': 0.3942072730133077, 'learning_rate/qwen_vl_interface': 7.259394567957169e-07, 'learning_rate/action_model': 7.259394567957169e-06, 'epoch': 0.33}) +[2026-10-03 13:56:02,928][starVLA.training.train_starvla][INFO] - Step 4518, Loss: {'train/loss': 0.19162914156913757, 'train/grad_norm_pre_clip': 315.6075134277344, 'timing/data': 0.0003112099948339164, 'timing/model': 0.39311566299875267, 'learning_rate/qwen_vl_interface': 7.250123080143066e-07, 'learning_rate/action_model': 7.2501230801430665e-06, 'epoch': 0.33}) +[2026-10-03 13:56:03,630][starVLA.training.train_starvla][INFO] - Step 4519, Loss: {'train/loss': 0.00010508134437259287, 'train/grad_norm_pre_clip': 0.08946281671524048, 'timing/data': 0.0003055499983020127, 'timing/model': 0.39133902300090995, 'learning_rate/qwen_vl_interface': 7.240870192844127e-07, 'learning_rate/action_model': 7.240870192844127e-06, 'epoch': 0.33}) +[2026-10-03 13:56:04,333][starVLA.training.train_starvla][INFO] - Step 4520, Loss: {'train/loss': 0.10962661355733871, 'train/grad_norm_pre_clip': 127.69290161132812, 'timing/data': 0.00030025899468455464, 'timing/model': 0.3929646629985655, 'learning_rate/qwen_vl_interface': 7.231635909863858e-07, 'learning_rate/action_model': 7.2316359098638585e-06, 'epoch': 0.33}) +[2026-10-03 13:56:05,037][starVLA.training.train_starvla][INFO] - Step 4521, Loss: {'train/loss': 5.189039802644402e-05, 'train/grad_norm_pre_clip': 0.030205005779862404, 'timing/data': 0.0003297200019005686, 'timing/model': 0.39413089300796855, 'learning_rate/qwen_vl_interface': 7.222420234998138e-07, 'learning_rate/action_model': 7.2224202349981385e-06, 'epoch': 0.33}) +[2026-10-03 13:56:05,741][starVLA.training.train_starvla][INFO] - Step 4522, Loss: {'train/loss': 0.466779887676239, 'train/grad_norm_pre_clip': 189.4572296142578, 'timing/data': 0.00023817000328563154, 'timing/model': 0.39320797300024424, 'learning_rate/qwen_vl_interface': 7.213223172035175e-07, 'learning_rate/action_model': 7.213223172035174e-06, 'epoch': 0.33}) +[2026-10-03 13:56:06,641][starVLA.training.train_starvla][INFO] - Step 4523, Loss: {'train/loss': 1.8349461242905818e-05, 'train/grad_norm_pre_clip': 0.010916456580162048, 'timing/data': 0.0002654190029716119, 'timing/model': 0.3933659729955252, 'learning_rate/qwen_vl_interface': 7.204044724755528e-07, 'learning_rate/action_model': 7.204044724755528e-06, 'epoch': 0.33}) +[2026-10-03 13:56:07,344][starVLA.training.train_starvla][INFO] - Step 4524, Loss: {'train/loss': 5.640011295326985e-06, 'train/grad_norm_pre_clip': 0.0024509502109140158, 'timing/data': 0.00019797999993897974, 'timing/model': 0.3932479420036543, 'learning_rate/qwen_vl_interface': 7.194884896932132e-07, 'learning_rate/action_model': 7.194884896932132e-06, 'epoch': 0.33}) +[2026-10-03 13:56:08,049][starVLA.training.train_starvla][INFO] - Step 4525, Loss: {'train/loss': 0.13641072809696198, 'train/grad_norm_pre_clip': 235.64427185058594, 'timing/data': 0.0002090699999826029, 'timing/model': 0.39441812300356105, 'learning_rate/qwen_vl_interface': 7.185743692330221e-07, 'learning_rate/action_model': 7.185743692330221e-06, 'epoch': 0.33}) +[2026-10-03 13:56:08,752][starVLA.training.train_starvla][INFO] - Step 4526, Loss: {'train/loss': 1.7022846805048175e-05, 'train/grad_norm_pre_clip': 0.007080282550305128, 'timing/data': 0.0002032899938058108, 'timing/model': 0.3925741729908623, 'learning_rate/qwen_vl_interface': 7.176621114707408e-07, 'learning_rate/action_model': 7.176621114707407e-06, 'epoch': 0.33}) +[2026-10-03 13:56:09,456][starVLA.training.train_starvla][INFO] - Step 4527, Loss: {'train/loss': 8.284800060209818e-06, 'train/grad_norm_pre_clip': 0.0023947700392454863, 'timing/data': 0.0001969399891095236, 'timing/model': 0.39298478300042916, 'learning_rate/qwen_vl_interface': 7.167517167813642e-07, 'learning_rate/action_model': 7.167517167813642e-06, 'epoch': 0.33}) +[2026-10-03 13:56:10,353][starVLA.training.train_starvla][INFO] - Step 4528, Loss: {'train/loss': 3.312988701509312e-05, 'train/grad_norm_pre_clip': 0.014536005444824696, 'timing/data': 0.0002005199930863455, 'timing/model': 0.5863711240090197, 'learning_rate/qwen_vl_interface': 7.15843185539121e-07, 'learning_rate/action_model': 7.15843185539121e-06, 'epoch': 0.33}) +[2026-10-03 13:56:11,071][starVLA.training.train_starvla][INFO] - Step 4529, Loss: {'train/loss': 3.3974508824030636e-06, 'train/grad_norm_pre_clip': 0.0007198130479082465, 'timing/data': 0.0003309399908175692, 'timing/model': 0.40669765100756194, 'learning_rate/qwen_vl_interface': 7.149365181174739e-07, 'learning_rate/action_model': 7.149365181174739e-06, 'epoch': 0.33}) +[2026-10-03 13:56:11,774][starVLA.training.train_starvla][INFO] - Step 4530, Loss: {'train/loss': 0.0002801230293698609, 'train/grad_norm_pre_clip': 0.25621381402015686, 'timing/data': 0.0002470199979143217, 'timing/model': 0.3925103729998227, 'learning_rate/qwen_vl_interface': 7.14031714889121e-07, 'learning_rate/action_model': 7.14031714889121e-06, 'epoch': 0.33}) +[2026-10-03 13:56:12,477][starVLA.training.train_starvla][INFO] - Step 4531, Loss: {'train/loss': 0.6249104738235474, 'train/grad_norm_pre_clip': 70.75228118896484, 'timing/data': 0.00028695999935735017, 'timing/model': 0.3922843229956925, 'learning_rate/qwen_vl_interface': 7.131287762259907e-07, 'learning_rate/action_model': 7.131287762259907e-06, 'epoch': 0.33}) +[2026-10-03 13:56:13,181][starVLA.training.train_starvla][INFO] - Step 4532, Loss: {'train/loss': 6.489384759333916e-06, 'train/grad_norm_pre_clip': 0.002423841506242752, 'timing/data': 0.00028094000299461186, 'timing/model': 0.3930769229918951, 'learning_rate/qwen_vl_interface': 7.12227702499248e-07, 'learning_rate/action_model': 7.12227702499248e-06, 'epoch': 0.33}) +[2026-10-03 13:56:14,092][starVLA.training.train_starvla][INFO] - Step 4533, Loss: {'train/loss': 6.528653466375545e-05, 'train/grad_norm_pre_clip': 0.029015781357884407, 'timing/data': 0.00024157999723684043, 'timing/model': 0.6003410429984797, 'learning_rate/qwen_vl_interface': 7.113284940792903e-07, 'learning_rate/action_model': 7.113284940792903e-06, 'epoch': 0.33}) +[2026-10-03 13:56:14,796][starVLA.training.train_starvla][INFO] - Step 4534, Loss: {'train/loss': 0.0025561521761119366, 'train/grad_norm_pre_clip': 8.203967094421387, 'timing/data': 0.00019964999228250235, 'timing/model': 0.39317501201003324, 'learning_rate/qwen_vl_interface': 7.104311513357476e-07, 'learning_rate/action_model': 7.1043115133574755e-06, 'epoch': 0.33}) +[2026-10-03 13:56:15,497][starVLA.training.train_starvla][INFO] - Step 4535, Loss: {'train/loss': 1.2889488516520942e-06, 'train/grad_norm_pre_clip': 0.00038543695700354874, 'timing/data': 0.00019607000285759568, 'timing/model': 0.3923215320101008, 'learning_rate/qwen_vl_interface': 7.095356746374845e-07, 'learning_rate/action_model': 7.095356746374844e-06, 'epoch': 0.33}) +[2026-10-03 13:56:16,201][starVLA.training.train_starvla][INFO] - Step 4536, Loss: {'train/loss': 0.06845464557409286, 'train/grad_norm_pre_clip': 28.776681900024414, 'timing/data': 0.00037665999843738973, 'timing/model': 0.39273307300754823, 'learning_rate/qwen_vl_interface': 7.086420643525973e-07, 'learning_rate/action_model': 7.086420643525972e-06, 'epoch': 0.33}) +[2026-10-03 13:56:16,919][starVLA.training.train_starvla][INFO] - Step 4537, Loss: {'train/loss': 0.0009637722396291792, 'train/grad_norm_pre_clip': 1.5356334447860718, 'timing/data': 0.0002973200025735423, 'timing/model': 0.4072268610034371, 'learning_rate/qwen_vl_interface': 7.077503208484151e-07, 'learning_rate/action_model': 7.0775032084841515e-06, 'epoch': 0.33}) +[2026-10-03 13:56:17,622][starVLA.training.train_starvla][INFO] - Step 4538, Loss: {'train/loss': 4.202041964163072e-06, 'train/grad_norm_pre_clip': 0.0009991347324103117, 'timing/data': 0.00022414998966269195, 'timing/model': 0.39229714300017804, 'learning_rate/qwen_vl_interface': 7.068604444915003e-07, 'learning_rate/action_model': 7.068604444915003e-06, 'epoch': 0.33}) +[2026-10-03 13:56:18,514][starVLA.training.train_starvla][INFO] - Step 4539, Loss: {'train/loss': 2.8203770852996968e-05, 'train/grad_norm_pre_clip': 0.017258305102586746, 'timing/data': 0.00020621999283321202, 'timing/model': 0.3927170929964632, 'learning_rate/qwen_vl_interface': 7.059724356476475e-07, 'learning_rate/action_model': 7.059724356476474e-06, 'epoch': 0.33}) +[2026-10-03 13:56:19,217][starVLA.training.train_starvla][INFO] - Step 4540, Loss: {'train/loss': 2.8694321372313425e-05, 'train/grad_norm_pre_clip': 0.036056049168109894, 'timing/data': 0.0002672300033736974, 'timing/model': 0.39284926300751977, 'learning_rate/qwen_vl_interface': 7.050862946818836e-07, 'learning_rate/action_model': 7.050862946818836e-06, 'epoch': 0.33}) +[2026-10-03 13:56:19,923][starVLA.training.train_starvla][INFO] - Step 4541, Loss: {'train/loss': 0.0752282664179802, 'train/grad_norm_pre_clip': 75.86060333251953, 'timing/data': 0.00025814000400714576, 'timing/model': 0.3953570919984486, 'learning_rate/qwen_vl_interface': 7.042020219584673e-07, 'learning_rate/action_model': 7.042020219584673e-06, 'epoch': 0.33}) +[2026-10-03 13:56:20,629][starVLA.training.train_starvla][INFO] - Step 4542, Loss: {'train/loss': 3.516615151966107e-06, 'train/grad_norm_pre_clip': 0.002318947808817029, 'timing/data': 0.00023964999127201736, 'timing/model': 0.39468123199185356, 'learning_rate/qwen_vl_interface': 7.0331961784089e-07, 'learning_rate/action_model': 7.033196178408899e-06, 'epoch': 0.33}) +[2026-10-03 13:56:21,335][starVLA.training.train_starvla][INFO] - Step 4543, Loss: {'train/loss': 0.0002230368263553828, 'train/grad_norm_pre_clip': 0.28968244791030884, 'timing/data': 0.00030066999897826463, 'timing/model': 0.39462117198854685, 'learning_rate/qwen_vl_interface': 7.024390826918748e-07, 'learning_rate/action_model': 7.024390826918747e-06, 'epoch': 0.33}) +[2026-10-03 13:56:22,231][starVLA.training.train_starvla][INFO] - Step 4544, Loss: {'train/loss': 2.548086740716826e-06, 'train/grad_norm_pre_clip': 0.0009233950404450297, 'timing/data': 0.0001965200062841177, 'timing/model': 0.5840887950034812, 'learning_rate/qwen_vl_interface': 7.01560416873376e-07, 'learning_rate/action_model': 7.015604168733759e-06, 'epoch': 0.33}) +[2026-10-03 13:56:22,939][starVLA.training.train_starvla][INFO] - Step 4545, Loss: {'train/loss': 0.3053378462791443, 'train/grad_norm_pre_clip': 78.64995574951172, 'timing/data': 0.0002476600056979805, 'timing/model': 0.39573035200010054, 'learning_rate/qwen_vl_interface': 7.006836207465803e-07, 'learning_rate/action_model': 7.006836207465803e-06, 'epoch': 0.33}) +[2026-10-03 13:56:23,648][starVLA.training.train_starvla][INFO] - Step 4546, Loss: {'train/loss': 3.79363373212982e-05, 'train/grad_norm_pre_clip': 0.04037624970078468, 'timing/data': 0.00032517001091036946, 'timing/model': 0.39353646300151013, 'learning_rate/qwen_vl_interface': 6.998086946719058e-07, 'learning_rate/action_model': 6.998086946719057e-06, 'epoch': 0.33}) +[2026-10-03 13:56:24,353][starVLA.training.train_starvla][INFO] - Step 4547, Loss: {'train/loss': 2.3904931367724203e-05, 'train/grad_norm_pre_clip': 0.023944487795233727, 'timing/data': 0.0003088100056629628, 'timing/model': 0.3942331630096305, 'learning_rate/qwen_vl_interface': 6.989356390090008e-07, 'learning_rate/action_model': 6.989356390090007e-06, 'epoch': 0.33}) +[2026-10-03 13:56:25,059][starVLA.training.train_starvla][INFO] - Step 4548, Loss: {'train/loss': 1.1860694939969108e-05, 'train/grad_norm_pre_clip': 0.005840320140123367, 'timing/data': 0.0001936799963004887, 'timing/model': 0.39431818200682756, 'learning_rate/qwen_vl_interface': 6.980644541167459e-07, 'learning_rate/action_model': 6.980644541167458e-06, 'epoch': 0.33}) +[2026-10-03 13:56:25,766][starVLA.training.train_starvla][INFO] - Step 4549, Loss: {'train/loss': 0.42746251821517944, 'train/grad_norm_pre_clip': 61.746116638183594, 'timing/data': 0.0002821200032485649, 'timing/model': 0.39521511299244594, 'learning_rate/qwen_vl_interface': 6.971951403532524e-07, 'learning_rate/action_model': 6.971951403532523e-06, 'epoch': 0.33}) +[2026-10-03 13:56:26,663][starVLA.training.train_starvla][INFO] - Step 4550, Loss: {'train/loss': 0.002472053049132228, 'train/grad_norm_pre_clip': 17.805795669555664, 'timing/data': 0.0002445599966449663, 'timing/model': 0.3948484029970132, 'learning_rate/qwen_vl_interface': 6.963276980758622e-07, 'learning_rate/action_model': 6.963276980758622e-06, 'epoch': 0.33}) +[2026-10-03 13:56:27,369][starVLA.training.train_starvla][INFO] - Step 4551, Loss: {'train/loss': 3.029819708899595e-05, 'train/grad_norm_pre_clip': 0.009841487742960453, 'timing/data': 0.0003164100053254515, 'timing/model': 0.394366471999092, 'learning_rate/qwen_vl_interface': 6.954621276411485e-07, 'learning_rate/action_model': 6.954621276411484e-06, 'epoch': 0.33}) +[2026-10-03 13:56:28,075][starVLA.training.train_starvla][INFO] - Step 4552, Loss: {'train/loss': 7.524899046984501e-06, 'train/grad_norm_pre_clip': 0.0014199865981936455, 'timing/data': 0.0002842800022335723, 'timing/model': 0.3942257730086567, 'learning_rate/qwen_vl_interface': 6.945984294049136e-07, 'learning_rate/action_model': 6.9459842940491364e-06, 'epoch': 0.33}) +[2026-10-03 13:56:28,782][starVLA.training.train_starvla][INFO] - Step 4553, Loss: {'train/loss': 3.574076617951505e-05, 'train/grad_norm_pre_clip': 0.07042905688285828, 'timing/data': 0.00028311999631114304, 'timing/model': 0.39571619300113525, 'learning_rate/qwen_vl_interface': 6.937366037221919e-07, 'learning_rate/action_model': 6.937366037221918e-06, 'epoch': 0.33}) +[2026-10-03 13:56:29,487][starVLA.training.train_starvla][INFO] - Step 4554, Loss: {'train/loss': 2.9280531634867657e-06, 'train/grad_norm_pre_clip': 0.0007959875511005521, 'timing/data': 0.0002699799952097237, 'timing/model': 0.3940457319986308, 'learning_rate/qwen_vl_interface': 6.928766509472471e-07, 'learning_rate/action_model': 6.92876650947247e-06, 'epoch': 0.33}) +[2026-10-03 13:56:30,384][starVLA.training.train_starvla][INFO] - Step 4555, Loss: {'train/loss': 5.550631613004953e-06, 'train/grad_norm_pre_clip': 0.0014958594692870975, 'timing/data': 0.0002765300014289096, 'timing/model': 0.3949830329947872, 'learning_rate/qwen_vl_interface': 6.920185714335737e-07, 'learning_rate/action_model': 6.9201857143357375e-06, 'epoch': 0.33}) +[2026-10-03 13:56:31,090][starVLA.training.train_starvla][INFO] - Step 4556, Loss: {'train/loss': 3.248434722991078e-06, 'train/grad_norm_pre_clip': 0.0012745244894176722, 'timing/data': 0.00019870999676641077, 'timing/model': 0.39468017300532665, 'learning_rate/qwen_vl_interface': 6.911623655338959e-07, 'learning_rate/action_model': 6.911623655338959e-06, 'epoch': 0.33}) +[2026-10-03 13:56:31,797][starVLA.training.train_starvla][INFO] - Step 4557, Loss: {'train/loss': 0.0003775399527512491, 'train/grad_norm_pre_clip': 0.6909911632537842, 'timing/data': 0.0002344800013815984, 'timing/model': 0.39544495299924165, 'learning_rate/qwen_vl_interface': 6.903080336001666e-07, 'learning_rate/action_model': 6.903080336001666e-06, 'epoch': 0.33}) +[2026-10-03 13:56:32,503][starVLA.training.train_starvla][INFO] - Step 4558, Loss: {'train/loss': 2.582960223662667e-05, 'train/grad_norm_pre_clip': 0.01282984483987093, 'timing/data': 0.00020855999900959432, 'timing/model': 0.39444455299235415, 'learning_rate/qwen_vl_interface': 6.894555759835707e-07, 'learning_rate/action_model': 6.894555759835707e-06, 'epoch': 0.33}) +[2026-10-03 13:56:33,207][starVLA.training.train_starvla][INFO] - Step 4559, Loss: {'train/loss': 1.5243308553181123e-05, 'train/grad_norm_pre_clip': 0.003763142740353942, 'timing/data': 0.00019382000027690083, 'timing/model': 0.3935850629932247, 'learning_rate/qwen_vl_interface': 6.886049930345204e-07, 'learning_rate/action_model': 6.8860499303452035e-06, 'epoch': 0.33}) +[2026-10-03 13:56:34,107][starVLA.training.train_starvla][INFO] - Step 4560, Loss: {'train/loss': 8.493245331919752e-06, 'train/grad_norm_pre_clip': 0.0063297138549387455, 'timing/data': 0.00019104000239167362, 'timing/model': 0.5881342149950797, 'learning_rate/qwen_vl_interface': 6.877562851026584e-07, 'learning_rate/action_model': 6.877562851026584e-06, 'epoch': 0.33}) +[2026-10-03 13:56:34,814][starVLA.training.train_starvla][INFO] - Step 4561, Loss: {'train/loss': 7.33751876396127e-05, 'train/grad_norm_pre_clip': 0.11047780513763428, 'timing/data': 0.000195720000192523, 'timing/model': 0.39610562300367747, 'learning_rate/qwen_vl_interface': 6.869094525368572e-07, 'learning_rate/action_model': 6.8690945253685715e-06, 'epoch': 0.33}) +[2026-10-03 13:56:35,521][starVLA.training.train_starvla][INFO] - Step 4562, Loss: {'train/loss': 1.4303887837741058e-05, 'train/grad_norm_pre_clip': 0.00495285727083683, 'timing/data': 0.00020166000467725098, 'timing/model': 0.39461458299774677, 'learning_rate/qwen_vl_interface': 6.860644956852168e-07, 'learning_rate/action_model': 6.860644956852168e-06, 'epoch': 0.33}) +[2026-10-03 13:56:36,225][starVLA.training.train_starvla][INFO] - Step 4563, Loss: {'train/loss': 3.5131994081893936e-05, 'train/grad_norm_pre_clip': 0.011987246572971344, 'timing/data': 0.00025659000675659627, 'timing/model': 0.3942010930040851, 'learning_rate/qwen_vl_interface': 6.852214148950667e-07, 'learning_rate/action_model': 6.852214148950667e-06, 'epoch': 0.33}) +[2026-10-03 13:56:36,931][starVLA.training.train_starvla][INFO] - Step 4564, Loss: {'train/loss': 0.4517703950405121, 'train/grad_norm_pre_clip': 72.85450744628906, 'timing/data': 0.00020003000099677593, 'timing/model': 0.3944768419896718, 'learning_rate/qwen_vl_interface': 6.843802105129667e-07, 'learning_rate/action_model': 6.843802105129666e-06, 'epoch': 0.33}) +[2026-10-03 13:56:37,639][starVLA.training.train_starvla][INFO] - Step 4565, Loss: {'train/loss': 0.00014568846381735057, 'train/grad_norm_pre_clip': 0.1662396788597107, 'timing/data': 0.00019672999042086303, 'timing/model': 0.39555638300953433, 'learning_rate/qwen_vl_interface': 6.835408828847035e-07, 'learning_rate/action_model': 6.8354088288470356e-06, 'epoch': 0.33}) +[2026-10-03 13:56:38,535][starVLA.training.train_starvla][INFO] - Step 4566, Loss: {'train/loss': 1.138393872679444e-05, 'train/grad_norm_pre_clip': 0.0050141760148108006, 'timing/data': 0.00021768000442534685, 'timing/model': 0.3948710230033612, 'learning_rate/qwen_vl_interface': 6.827034323552929e-07, 'learning_rate/action_model': 6.827034323552929e-06, 'epoch': 0.33}) +[2026-10-03 13:56:39,240][starVLA.training.train_starvla][INFO] - Step 4567, Loss: {'train/loss': 3.143835056107491e-05, 'train/grad_norm_pre_clip': 0.010864845477044582, 'timing/data': 0.00033237999014090747, 'timing/model': 0.393215374002466, 'learning_rate/qwen_vl_interface': 6.818678592689794e-07, 'learning_rate/action_model': 6.818678592689794e-06, 'epoch': 0.33}) +[2026-10-03 13:56:39,944][starVLA.training.train_starvla][INFO] - Step 4568, Loss: {'train/loss': 0.6299118399620056, 'train/grad_norm_pre_clip': 32.29083251953125, 'timing/data': 0.00029319999157451093, 'timing/model': 0.39312682299350854, 'learning_rate/qwen_vl_interface': 6.810341639692357e-07, 'learning_rate/action_model': 6.810341639692357e-06, 'epoch': 0.33}) +[2026-10-03 13:56:40,649][starVLA.training.train_starvla][INFO] - Step 4569, Loss: {'train/loss': 1.162217085948214e-05, 'train/grad_norm_pre_clip': 0.00968585629016161, 'timing/data': 0.00020881000091321766, 'timing/model': 0.394556053011911, 'learning_rate/qwen_vl_interface': 6.802023467987631e-07, 'learning_rate/action_model': 6.802023467987631e-06, 'epoch': 0.33}) +[2026-10-03 13:56:41,354][starVLA.training.train_starvla][INFO] - Step 4570, Loss: {'train/loss': 0.035455092787742615, 'train/grad_norm_pre_clip': 70.494873046875, 'timing/data': 0.00019916999735869467, 'timing/model': 0.3935719440050889, 'learning_rate/qwen_vl_interface': 6.793724080994884e-07, 'learning_rate/action_model': 6.793724080994884e-06, 'epoch': 0.33}) +[2026-10-03 13:56:42,249][starVLA.training.train_starvla][INFO] - Step 4571, Loss: {'train/loss': 3.115247454843484e-05, 'train/grad_norm_pre_clip': 0.022968966513872147, 'timing/data': 0.0002864499983843416, 'timing/model': 0.39339942298829556, 'learning_rate/qwen_vl_interface': 6.785443482125693e-07, 'learning_rate/action_model': 6.7854434821256915e-06, 'epoch': 0.33}) +[2026-10-03 13:56:42,954][starVLA.training.train_starvla][INFO] - Step 4572, Loss: {'train/loss': 1.0229263352812268e-05, 'train/grad_norm_pre_clip': 0.00345156854018569, 'timing/data': 0.00020150998898316175, 'timing/model': 0.3934154330054298, 'learning_rate/qwen_vl_interface': 6.777181674783895e-07, 'learning_rate/action_model': 6.777181674783895e-06, 'epoch': 0.33}) +[2026-10-03 13:56:43,659][starVLA.training.train_starvla][INFO] - Step 4573, Loss: {'train/loss': 2.510837475711014e-06, 'train/grad_norm_pre_clip': 0.0006956233992241323, 'timing/data': 0.0002042800042545423, 'timing/model': 0.39493945299182087, 'learning_rate/qwen_vl_interface': 6.76893866236561e-07, 'learning_rate/action_model': 6.7689386623656094e-06, 'epoch': 0.33}) +[2026-10-03 13:56:44,363][starVLA.training.train_starvla][INFO] - Step 4574, Loss: {'train/loss': 0.001461290754377842, 'train/grad_norm_pre_clip': 5.035315990447998, 'timing/data': 0.00020812900038436055, 'timing/model': 0.39313452401256654, 'learning_rate/qwen_vl_interface': 6.76071444825923e-07, 'learning_rate/action_model': 6.76071444825923e-06, 'epoch': 0.33}) +[2026-10-03 13:56:45,067][starVLA.training.train_starvla][INFO] - Step 4575, Loss: {'train/loss': 2.600764128146693e-05, 'train/grad_norm_pre_clip': 0.007267111446708441, 'timing/data': 0.0001972600002773106, 'timing/model': 0.39348617299401667, 'learning_rate/qwen_vl_interface': 6.752509035845413e-07, 'learning_rate/action_model': 6.7525090358454126e-06, 'epoch': 0.33}) +[2026-10-03 13:56:45,964][starVLA.training.train_starvla][INFO] - Step 4576, Loss: {'train/loss': 0.0002081031125271693, 'train/grad_norm_pre_clip': 0.20563295483589172, 'timing/data': 0.0003084099880652502, 'timing/model': 0.3938630529883085, 'learning_rate/qwen_vl_interface': 6.744322428497095e-07, 'learning_rate/action_model': 6.744322428497095e-06, 'epoch': 0.33}) +[2026-10-03 13:56:46,670][starVLA.training.train_starvla][INFO] - Step 4577, Loss: {'train/loss': 6.392472187144449e-06, 'train/grad_norm_pre_clip': 0.0017690606182441115, 'timing/data': 0.00023112999042496085, 'timing/model': 0.3946350830083247, 'learning_rate/qwen_vl_interface': 6.736154629579487e-07, 'learning_rate/action_model': 6.736154629579487e-06, 'epoch': 0.34}) +[2026-10-03 13:56:47,374][starVLA.training.train_starvla][INFO] - Step 4578, Loss: {'train/loss': 2.0957100787200034e-05, 'train/grad_norm_pre_clip': 0.013541078194975853, 'timing/data': 0.00031749998743180186, 'timing/model': 0.3937674729968421, 'learning_rate/qwen_vl_interface': 6.728005642450055e-07, 'learning_rate/action_model': 6.7280056424500554e-06, 'epoch': 0.34}) +[2026-10-03 13:56:48,078][starVLA.training.train_starvla][INFO] - Step 4579, Loss: {'train/loss': 4.2914962250506505e-06, 'train/grad_norm_pre_clip': 0.0015559825114905834, 'timing/data': 0.0003099499881500378, 'timing/model': 0.39262451299873646, 'learning_rate/qwen_vl_interface': 6.719875470458553e-07, 'learning_rate/action_model': 6.719875470458553e-06, 'epoch': 0.34}) +[2026-10-03 13:56:48,782][starVLA.training.train_starvla][INFO] - Step 4580, Loss: {'train/loss': 0.0005035296198911965, 'train/grad_norm_pre_clip': 0.8972007036209106, 'timing/data': 0.00020125000446569175, 'timing/model': 0.3939740830101073, 'learning_rate/qwen_vl_interface': 6.711764116946972e-07, 'learning_rate/action_model': 6.7117641169469715e-06, 'epoch': 0.34}) +[2026-10-03 13:56:49,680][starVLA.training.train_starvla][INFO] - Step 4581, Loss: {'train/loss': 0.4529491066932678, 'train/grad_norm_pre_clip': 74.35270690917969, 'timing/data': 0.0003060199960600585, 'timing/model': 0.5873239759966964, 'learning_rate/qwen_vl_interface': 6.703671585249587e-07, 'learning_rate/action_model': 6.703671585249586e-06, 'epoch': 0.34}) +[2026-10-03 13:56:50,384][starVLA.training.train_starvla][INFO] - Step 4582, Loss: {'train/loss': 0.0003472544194664806, 'train/grad_norm_pre_clip': 0.3473954200744629, 'timing/data': 0.00020228000357747078, 'timing/model': 0.3933654430002207, 'learning_rate/qwen_vl_interface': 6.69559787869294e-07, 'learning_rate/action_model': 6.69559787869294e-06, 'epoch': 0.34}) +[2026-10-03 13:56:51,087][starVLA.training.train_starvla][INFO] - Step 4583, Loss: {'train/loss': 1.1920911902052467e-06, 'train/grad_norm_pre_clip': 0.0003693452454172075, 'timing/data': 0.00026221000007353723, 'timing/model': 0.39264476401149295, 'learning_rate/qwen_vl_interface': 6.68754300059582e-07, 'learning_rate/action_model': 6.687543000595819e-06, 'epoch': 0.34}) +[2026-10-03 13:56:51,791][starVLA.training.train_starvla][INFO] - Step 4584, Loss: {'train/loss': 0.18057745695114136, 'train/grad_norm_pre_clip': 476.965087890625, 'timing/data': 0.00022665000869892538, 'timing/model': 0.393469782997272, 'learning_rate/qwen_vl_interface': 6.67950695426929e-07, 'learning_rate/action_model': 6.679506954269291e-06, 'epoch': 0.34}) +[2026-10-03 13:56:52,496][starVLA.training.train_starvla][INFO] - Step 4585, Loss: {'train/loss': 0.3430626690387726, 'train/grad_norm_pre_clip': 58.43600082397461, 'timing/data': 0.00020130000484641641, 'timing/model': 0.3945753330044681, 'learning_rate/qwen_vl_interface': 6.671489743016651e-07, 'learning_rate/action_model': 6.671489743016651e-06, 'epoch': 0.34}) +[2026-10-03 13:56:53,393][starVLA.training.train_starvla][INFO] - Step 4586, Loss: {'train/loss': 8.344644584212801e-07, 'train/grad_norm_pre_clip': 0.00026154526858590543, 'timing/data': 0.00024499000573996454, 'timing/model': 0.5863118459965335, 'learning_rate/qwen_vl_interface': 6.663491370133487e-07, 'learning_rate/action_model': 6.663491370133487e-06, 'epoch': 0.34}) +[2026-10-03 13:56:54,098][starVLA.training.train_starvla][INFO] - Step 4587, Loss: {'train/loss': 5.123583468957804e-05, 'train/grad_norm_pre_clip': 0.02608254738152027, 'timing/data': 0.00020024999685119838, 'timing/model': 0.3937166830000933, 'learning_rate/qwen_vl_interface': 6.655511838907629e-07, 'learning_rate/action_model': 6.655511838907629e-06, 'epoch': 0.34}) +[2026-10-03 13:56:54,799][starVLA.training.train_starvla][INFO] - Step 4588, Loss: {'train/loss': 2.0994060832890682e-05, 'train/grad_norm_pre_clip': 0.009599343873560429, 'timing/data': 0.00024013000074774027, 'timing/model': 0.391668292999384, 'learning_rate/qwen_vl_interface': 6.647551152619149e-07, 'learning_rate/action_model': 6.647551152619148e-06, 'epoch': 0.34}) +[2026-10-03 13:56:55,504][starVLA.training.train_starvla][INFO] - Step 4589, Loss: {'train/loss': 1.9973080270574428e-05, 'train/grad_norm_pre_clip': 0.01582973822951317, 'timing/data': 0.00024197899620048702, 'timing/model': 0.3939391439926112, 'learning_rate/qwen_vl_interface': 6.639609314540396e-07, 'learning_rate/action_model': 6.639609314540396e-06, 'epoch': 0.34}) +[2026-10-03 13:56:56,208][starVLA.training.train_starvla][INFO] - Step 4590, Loss: {'train/loss': 9.424778909306042e-06, 'train/grad_norm_pre_clip': 0.0025603179819881916, 'timing/data': 0.00020497999503277242, 'timing/model': 0.3929615339875454, 'learning_rate/qwen_vl_interface': 6.631686327935954e-07, 'learning_rate/action_model': 6.631686327935954e-06, 'epoch': 0.34}) +[2026-10-03 13:56:56,911][starVLA.training.train_starvla][INFO] - Step 4591, Loss: {'train/loss': 1.171181065728888e-05, 'train/grad_norm_pre_clip': 0.0031413263641297817, 'timing/data': 0.0001946700067492202, 'timing/model': 0.3933737530023791, 'learning_rate/qwen_vl_interface': 6.623782196062665e-07, 'learning_rate/action_model': 6.623782196062664e-06, 'epoch': 0.34}) +[2026-10-03 13:56:57,807][starVLA.training.train_starvla][INFO] - Step 4592, Loss: {'train/loss': 0.00019308259652461857, 'train/grad_norm_pre_clip': 0.3511720299720764, 'timing/data': 0.00025275000371038914, 'timing/model': 0.39254544299910776, 'learning_rate/qwen_vl_interface': 6.615896922169622e-07, 'learning_rate/action_model': 6.615896922169622e-06, 'epoch': 0.34}) +[2026-10-03 13:56:58,512][starVLA.training.train_starvla][INFO] - Step 4593, Loss: {'train/loss': 4.586645809467882e-05, 'train/grad_norm_pre_clip': 0.016193440183997154, 'timing/data': 0.00024576000578235835, 'timing/model': 0.39430644300591666, 'learning_rate/qwen_vl_interface': 6.608030509498156e-07, 'learning_rate/action_model': 6.608030509498156e-06, 'epoch': 0.34}) +[2026-10-03 13:56:59,216][starVLA.training.train_starvla][INFO] - Step 4594, Loss: {'train/loss': 0.5281615257263184, 'train/grad_norm_pre_clip': 35.640525817871094, 'timing/data': 0.0002811500016832724, 'timing/model': 0.39341008399787825, 'learning_rate/qwen_vl_interface': 6.600182961281865e-07, 'learning_rate/action_model': 6.600182961281864e-06, 'epoch': 0.34}) +[2026-10-03 13:56:59,919][starVLA.training.train_starvla][INFO] - Step 4595, Loss: {'train/loss': 0.14738865196704865, 'train/grad_norm_pre_clip': 186.89892578125, 'timing/data': 0.00026854001043830067, 'timing/model': 0.39311686300789006, 'learning_rate/qwen_vl_interface': 6.592354280746561e-07, 'learning_rate/action_model': 6.5923542807465595e-06, 'epoch': 0.34}) +[2026-10-03 13:57:00,623][starVLA.training.train_starvla][INFO] - Step 4596, Loss: {'train/loss': 5.586384213529527e-05, 'train/grad_norm_pre_clip': 0.07092835009098053, 'timing/data': 0.00023925999994389713, 'timing/model': 0.39330328301002737, 'learning_rate/qwen_vl_interface': 6.584544471110335e-07, 'learning_rate/action_model': 6.584544471110335e-06, 'epoch': 0.34}) +[2026-10-03 13:57:01,520][starVLA.training.train_starvla][INFO] - Step 4597, Loss: {'train/loss': 0.0008505024597980082, 'train/grad_norm_pre_clip': 2.503312826156616, 'timing/data': 0.0002658099983818829, 'timing/model': 0.5863139449938899, 'learning_rate/qwen_vl_interface': 6.576753535583495e-07, 'learning_rate/action_model': 6.576753535583495e-06, 'epoch': 0.34}) +[2026-10-03 13:57:02,224][starVLA.training.train_starvla][INFO] - Step 4598, Loss: {'train/loss': 6.131710506451782e-06, 'train/grad_norm_pre_clip': 0.0021747767459601164, 'timing/data': 0.0003200300125172362, 'timing/model': 0.39319728399277665, 'learning_rate/qwen_vl_interface': 6.5689814773686e-07, 'learning_rate/action_model': 6.568981477368599e-06, 'epoch': 0.34}) +[2026-10-03 13:57:02,928][starVLA.training.train_starvla][INFO] - Step 4599, Loss: {'train/loss': 1.2933722246089019e-05, 'train/grad_norm_pre_clip': 0.007067234255373478, 'timing/data': 0.0004175100038992241, 'timing/model': 0.3929969640012132, 'learning_rate/qwen_vl_interface': 6.56122829966046e-07, 'learning_rate/action_model': 6.56122829966046e-06, 'epoch': 0.34}) +[2026-10-03 13:57:03,632][starVLA.training.train_starvla][INFO] - Step 4600, Loss: {'train/loss': 4.7012363211251795e-06, 'train/grad_norm_pre_clip': 0.002013539196923375, 'timing/data': 0.00029028899734839797, 'timing/model': 0.3928359440033091, 'learning_rate/qwen_vl_interface': 6.553494005646101e-07, 'learning_rate/action_model': 6.553494005646101e-06, 'epoch': 0.34}) +[2026-10-03 13:57:04,336][starVLA.training.train_starvla][INFO] - Step 4601, Loss: {'train/loss': 0.00011241659376537427, 'train/grad_norm_pre_clip': 0.07283284515142441, 'timing/data': 0.0003188499977113679, 'timing/model': 0.394286362992716, 'learning_rate/qwen_vl_interface': 6.545778598504807e-07, 'learning_rate/action_model': 6.545778598504807e-06, 'epoch': 0.34}) +[2026-10-03 13:57:05,040][starVLA.training.train_starvla][INFO] - Step 4602, Loss: {'train/loss': 1.5995758076314814e-05, 'train/grad_norm_pre_clip': 0.008955751545727253, 'timing/data': 0.00019438000163063407, 'timing/model': 0.39327348400547635, 'learning_rate/qwen_vl_interface': 6.538082081408093e-07, 'learning_rate/action_model': 6.538082081408093e-06, 'epoch': 0.34}) +[2026-10-03 13:57:05,935][starVLA.training.train_starvla][INFO] - Step 4603, Loss: {'train/loss': 2.4437836145807523e-06, 'train/grad_norm_pre_clip': 0.0006235298351384699, 'timing/data': 0.0002754500019364059, 'timing/model': 0.39377571300428826, 'learning_rate/qwen_vl_interface': 6.530404457519697e-07, 'learning_rate/action_model': 6.530404457519697e-06, 'epoch': 0.34}) +[2026-10-03 13:57:06,640][starVLA.training.train_starvla][INFO] - Step 4604, Loss: {'train/loss': 0.0007142350077629089, 'train/grad_norm_pre_clip': 0.7774263620376587, 'timing/data': 0.00021872999786864966, 'timing/model': 0.39376494399039075, 'learning_rate/qwen_vl_interface': 6.522745729995614e-07, 'learning_rate/action_model': 6.522745729995613e-06, 'epoch': 0.34}) +[2026-10-03 13:57:07,346][starVLA.training.train_starvla][INFO] - Step 4605, Loss: {'train/loss': 1.0281171853421256e-05, 'train/grad_norm_pre_clip': 0.006288210395723581, 'timing/data': 0.0002385300031164661, 'timing/model': 0.3944602429983206, 'learning_rate/qwen_vl_interface': 6.515105901984037e-07, 'learning_rate/action_model': 6.515105901984037e-06, 'epoch': 0.34}) +[2026-10-03 13:57:08,050][starVLA.training.train_starvla][INFO] - Step 4606, Loss: {'train/loss': 0.00011955235095229, 'train/grad_norm_pre_clip': 0.12011899799108505, 'timing/data': 0.00020461999520193785, 'timing/model': 0.39340694299608003, 'learning_rate/qwen_vl_interface': 6.50748497662543e-07, 'learning_rate/action_model': 6.507484976625429e-06, 'epoch': 0.34}) +[2026-10-03 13:57:08,754][starVLA.training.train_starvla][INFO] - Step 4607, Loss: {'train/loss': 0.00012246159894857556, 'train/grad_norm_pre_clip': 0.11152555048465729, 'timing/data': 0.0001949400029843673, 'timing/model': 0.39344584400532767, 'learning_rate/qwen_vl_interface': 6.499882957052459e-07, 'learning_rate/action_model': 6.4998829570524586e-06, 'epoch': 0.34}) +[2026-10-03 13:57:09,653][starVLA.training.train_starvla][INFO] - Step 4608, Loss: {'train/loss': 8.158047421602532e-06, 'train/grad_norm_pre_clip': 0.002040537539869547, 'timing/data': 0.00031148000562097877, 'timing/model': 0.39373341300233733, 'learning_rate/qwen_vl_interface': 6.492299846390024e-07, 'learning_rate/action_model': 6.492299846390023e-06, 'epoch': 0.34}) +[2026-10-03 13:57:10,358][starVLA.training.train_starvla][INFO] - Step 4609, Loss: {'train/loss': 3.507976725813933e-05, 'train/grad_norm_pre_clip': 0.015122375451028347, 'timing/data': 0.00033416000951547176, 'timing/model': 0.39438560299458914, 'learning_rate/qwen_vl_interface': 6.484735647755259e-07, 'learning_rate/action_model': 6.48473564775526e-06, 'epoch': 0.34}) +[2026-10-03 13:57:11,062][starVLA.training.train_starvla][INFO] - Step 4610, Loss: {'train/loss': 0.2996670603752136, 'train/grad_norm_pre_clip': 115.9976577758789, 'timing/data': 0.00023968999448698014, 'timing/model': 0.39307603301131167, 'learning_rate/qwen_vl_interface': 6.477190364257524e-07, 'learning_rate/action_model': 6.477190364257524e-06, 'epoch': 0.34}) +[2026-10-03 13:57:11,764][starVLA.training.train_starvla][INFO] - Step 4611, Loss: {'train/loss': 0.0009676515473984182, 'train/grad_norm_pre_clip': 4.165294647216797, 'timing/data': 0.00030715999309904873, 'timing/model': 0.39241043299261946, 'learning_rate/qwen_vl_interface': 6.469663998998393e-07, 'learning_rate/action_model': 6.469663998998393e-06, 'epoch': 0.34}) +[2026-10-03 13:57:12,467][starVLA.training.train_starvla][INFO] - Step 4612, Loss: {'train/loss': 1.3142236639396288e-05, 'train/grad_norm_pre_clip': 0.0034407360944896936, 'timing/data': 0.000231869998970069, 'timing/model': 0.39273545300238766, 'learning_rate/qwen_vl_interface': 6.462156555071681e-07, 'learning_rate/action_model': 6.46215655507168e-06, 'epoch': 0.34}) +[2026-10-03 13:57:13,365][starVLA.training.train_starvla][INFO] - Step 4613, Loss: {'train/loss': 1.3932225556345657e-05, 'train/grad_norm_pre_clip': 0.004062323831021786, 'timing/data': 0.0002953200018964708, 'timing/model': 0.5870700160012348, 'learning_rate/qwen_vl_interface': 6.454668035563401e-07, 'learning_rate/action_model': 6.454668035563401e-06, 'epoch': 0.34}) +[2026-10-03 13:57:14,067][starVLA.training.train_starvla][INFO] - Step 4614, Loss: {'train/loss': 0.00011593720410019159, 'train/grad_norm_pre_clip': 0.19327032566070557, 'timing/data': 0.00020411999139469117, 'timing/model': 0.3928757929970743, 'learning_rate/qwen_vl_interface': 6.447198443551809e-07, 'learning_rate/action_model': 6.4471984435518075e-06, 'epoch': 0.34}) +[2026-10-03 13:57:14,770][starVLA.training.train_starvla][INFO] - Step 4615, Loss: {'train/loss': 1.2799247087968979e-05, 'train/grad_norm_pre_clip': 0.008082925342023373, 'timing/data': 0.0002347699919482693, 'timing/model': 0.39302977301122155, 'learning_rate/qwen_vl_interface': 6.439747782107373e-07, 'learning_rate/action_model': 6.4397477821073735e-06, 'epoch': 0.34}) +[2026-10-03 13:57:15,472][starVLA.training.train_starvla][INFO] - Step 4616, Loss: {'train/loss': 4.03818648919696e-06, 'train/grad_norm_pre_clip': 0.001023765653371811, 'timing/data': 0.00019705999875441194, 'timing/model': 0.3930852540070191, 'learning_rate/qwen_vl_interface': 6.432316054292779e-07, 'learning_rate/action_model': 6.43231605429278e-06, 'epoch': 0.34}) +[2026-10-03 13:57:16,178][starVLA.training.train_starvla][INFO] - Step 4617, Loss: {'train/loss': 1.7583329281478655e-06, 'train/grad_norm_pre_clip': 0.0005925255827605724, 'timing/data': 0.00029380999330896884, 'timing/model': 0.39452201400126796, 'learning_rate/qwen_vl_interface': 6.424903263162924e-07, 'learning_rate/action_model': 6.424903263162923e-06, 'epoch': 0.34}) +[2026-10-03 13:57:16,883][starVLA.training.train_starvla][INFO] - Step 4618, Loss: {'train/loss': 5.759236955782399e-06, 'train/grad_norm_pre_clip': 0.0020197867415845394, 'timing/data': 0.0002476700028637424, 'timing/model': 0.39326918400183786, 'learning_rate/qwen_vl_interface': 6.417509411764924e-07, 'learning_rate/action_model': 6.4175094117649244e-06, 'epoch': 0.34}) +[2026-10-03 13:57:17,777][starVLA.training.train_starvla][INFO] - Step 4619, Loss: {'train/loss': 1.6821486497065052e-05, 'train/grad_norm_pre_clip': 0.009503870271146297, 'timing/data': 0.0003122199996141717, 'timing/model': 0.39384650300780777, 'learning_rate/qwen_vl_interface': 6.410134503138122e-07, 'learning_rate/action_model': 6.4101345031381224e-06, 'epoch': 0.34}) +[2026-10-03 13:57:18,480][starVLA.training.train_starvla][INFO] - Step 4620, Loss: {'train/loss': 0.0017720788018777966, 'train/grad_norm_pre_clip': 10.663007736206055, 'timing/data': 0.00033308001002296805, 'timing/model': 0.39293325299513526, 'learning_rate/qwen_vl_interface': 6.402778540314054e-07, 'learning_rate/action_model': 6.402778540314054e-06, 'epoch': 0.34}) +[2026-10-03 13:57:19,185][starVLA.training.train_starvla][INFO] - Step 4621, Loss: {'train/loss': 2.8088484214094933e-06, 'train/grad_norm_pre_clip': 0.0011872191680595279, 'timing/data': 0.00029866000113543123, 'timing/model': 0.3941993730113609, 'learning_rate/qwen_vl_interface': 6.39544152631649e-07, 'learning_rate/action_model': 6.3954415263164895e-06, 'epoch': 0.34}) +[2026-10-03 13:57:19,889][starVLA.training.train_starvla][INFO] - Step 4622, Loss: {'train/loss': 1.6419489838881418e-05, 'train/grad_norm_pre_clip': 0.005711014848202467, 'timing/data': 0.00021708999702241272, 'timing/model': 0.39326386399625335, 'learning_rate/qwen_vl_interface': 6.388123464161388e-07, 'learning_rate/action_model': 6.388123464161389e-06, 'epoch': 0.34}) +[2026-10-03 13:57:20,591][starVLA.training.train_starvla][INFO] - Step 4623, Loss: {'train/loss': 0.1903616338968277, 'train/grad_norm_pre_clip': 311.0805969238281, 'timing/data': 0.00022081000497564673, 'timing/model': 0.39235582400579005, 'learning_rate/qwen_vl_interface': 6.380824356856936e-07, 'learning_rate/action_model': 6.380824356856936e-06, 'epoch': 0.34}) +[2026-10-03 13:57:21,487][starVLA.training.train_starvla][INFO] - Step 4624, Loss: {'train/loss': 0.018152544274926186, 'train/grad_norm_pre_clip': 38.26165771484375, 'timing/data': 0.0002685899962671101, 'timing/model': 0.3930806930002291, 'learning_rate/qwen_vl_interface': 6.373544207403513e-07, 'learning_rate/action_model': 6.373544207403513e-06, 'epoch': 0.34}) +[2026-10-03 13:57:22,192][starVLA.training.train_starvla][INFO] - Step 4625, Loss: {'train/loss': 5.342002623365261e-06, 'train/grad_norm_pre_clip': 0.0015047454508021474, 'timing/data': 0.00032623000151943415, 'timing/model': 0.3943442230083747, 'learning_rate/qwen_vl_interface': 6.366283018793723e-07, 'learning_rate/action_model': 6.366283018793723e-06, 'epoch': 0.34}) +[2026-10-03 13:57:22,895][starVLA.training.train_starvla][INFO] - Step 4626, Loss: {'train/loss': 0.00014596541586797684, 'train/grad_norm_pre_clip': 0.14722304046154022, 'timing/data': 0.00022262000129558146, 'timing/model': 0.39225766400340945, 'learning_rate/qwen_vl_interface': 6.359040794012358e-07, 'learning_rate/action_model': 6.359040794012358e-06, 'epoch': 0.34}) +[2026-10-03 13:57:23,596][starVLA.training.train_starvla][INFO] - Step 4627, Loss: {'train/loss': 2.6896507279161597e-06, 'train/grad_norm_pre_clip': 0.0005542528815567493, 'timing/data': 0.00021303999528754503, 'timing/model': 0.39234724300331436, 'learning_rate/qwen_vl_interface': 6.351817536036431e-07, 'learning_rate/action_model': 6.351817536036431e-06, 'epoch': 0.34}) +[2026-10-03 13:57:24,302][starVLA.training.train_starvla][INFO] - Step 4628, Loss: {'train/loss': 9.923728612193372e-06, 'train/grad_norm_pre_clip': 0.002618903759866953, 'timing/data': 0.0003247199929319322, 'timing/model': 0.39453219300776254, 'learning_rate/qwen_vl_interface': 6.344613247835135e-07, 'learning_rate/action_model': 6.344613247835135e-06, 'epoch': 0.34}) +[2026-10-03 13:57:25,199][starVLA.training.train_starvla][INFO] - Step 4629, Loss: {'train/loss': 0.0005489133181981742, 'train/grad_norm_pre_clip': 0.7891606092453003, 'timing/data': 0.00043684001138899475, 'timing/model': 0.39392358300392516, 'learning_rate/qwen_vl_interface': 6.337427932369898e-07, 'learning_rate/action_model': 6.337427932369898e-06, 'epoch': 0.34}) +[2026-10-03 13:57:25,902][starVLA.training.train_starvla][INFO] - Step 4630, Loss: {'train/loss': 3.606445170589723e-05, 'train/grad_norm_pre_clip': 0.011959055438637733, 'timing/data': 0.00022178998915478587, 'timing/model': 0.3927438130049268, 'learning_rate/qwen_vl_interface': 6.330261592594328e-07, 'learning_rate/action_model': 6.330261592594327e-06, 'epoch': 0.34}) +[2026-10-03 13:57:26,605][starVLA.training.train_starvla][INFO] - Step 4631, Loss: {'train/loss': 0.00046427847701124847, 'train/grad_norm_pre_clip': 0.36217111349105835, 'timing/data': 0.00031335999665316194, 'timing/model': 0.39232597399677616, 'learning_rate/qwen_vl_interface': 6.323114231454229e-07, 'learning_rate/action_model': 6.323114231454229e-06, 'epoch': 0.34}) +[2026-10-03 13:57:27,309][starVLA.training.train_starvla][INFO] - Step 4632, Loss: {'train/loss': 9.746836440172046e-05, 'train/grad_norm_pre_clip': 0.04761423543095589, 'timing/data': 0.0002792589948512614, 'timing/model': 0.39348807401256636, 'learning_rate/qwen_vl_interface': 6.315985851887618e-07, 'learning_rate/action_model': 6.315985851887618e-06, 'epoch': 0.34}) +[2026-10-03 13:57:28,015][starVLA.training.train_starvla][INFO] - Step 4633, Loss: {'train/loss': 0.0001348576624877751, 'train/grad_norm_pre_clip': 0.052577536553144455, 'timing/data': 0.0002530899946577847, 'timing/model': 0.3955778739909874, 'learning_rate/qwen_vl_interface': 6.308876456824697e-07, 'learning_rate/action_model': 6.3088764568246975e-06, 'epoch': 0.34}) +[2026-10-03 13:57:28,911][starVLA.training.train_starvla][INFO] - Step 4634, Loss: {'train/loss': 0.5408282279968262, 'train/grad_norm_pre_clip': 42.139862060546875, 'timing/data': 0.00022281000565271825, 'timing/model': 0.5841981859994121, 'learning_rate/qwen_vl_interface': 6.301786049187875e-07, 'learning_rate/action_model': 6.3017860491878745e-06, 'epoch': 0.34}) +[2026-10-03 13:57:29,617][starVLA.training.train_starvla][INFO] - Step 4635, Loss: {'train/loss': 0.05915847048163414, 'train/grad_norm_pre_clip': 190.7141571044922, 'timing/data': 0.00024284000392071903, 'timing/model': 0.3945593840035144, 'learning_rate/qwen_vl_interface': 6.294714631891755e-07, 'learning_rate/action_model': 6.294714631891754e-06, 'epoch': 0.34}) +[2026-10-03 13:57:30,323][starVLA.training.train_starvla][INFO] - Step 4636, Loss: {'train/loss': 0.49208182096481323, 'train/grad_norm_pre_clip': 51.54136657714844, 'timing/data': 0.00019561000226531178, 'timing/model': 0.3947100740042515, 'learning_rate/qwen_vl_interface': 6.287662207843115e-07, 'learning_rate/action_model': 6.287662207843114e-06, 'epoch': 0.34}) +[2026-10-03 13:57:31,030][starVLA.training.train_starvla][INFO] - Step 4637, Loss: {'train/loss': 5.602742476185085e-06, 'train/grad_norm_pre_clip': 0.0025944679509848356, 'timing/data': 0.0002542099973652512, 'timing/model': 0.39586402400163934, 'learning_rate/qwen_vl_interface': 6.280628779940956e-07, 'learning_rate/action_model': 6.280628779940956e-06, 'epoch': 0.34}) +[2026-10-03 13:57:31,735][starVLA.training.train_starvla][INFO] - Step 4638, Loss: {'train/loss': 4.082899977220222e-06, 'train/grad_norm_pre_clip': 0.0008729976252652705, 'timing/data': 0.00021674000890925527, 'timing/model': 0.39315310399979353, 'learning_rate/qwen_vl_interface': 6.273614351076445e-07, 'learning_rate/action_model': 6.273614351076445e-06, 'epoch': 0.34}) +[2026-10-03 13:57:32,632][starVLA.training.train_starvla][INFO] - Step 4639, Loss: {'train/loss': 8.746591447561514e-06, 'train/grad_norm_pre_clip': 0.0032468922436237335, 'timing/data': 0.0003205599932698533, 'timing/model': 0.5865454960003262, 'learning_rate/qwen_vl_interface': 6.266618924132951e-07, 'learning_rate/action_model': 6.266618924132951e-06, 'epoch': 0.34}) +[2026-10-03 13:57:33,336][starVLA.training.train_starvla][INFO] - Step 4640, Loss: {'train/loss': 2.414457776467316e-05, 'train/grad_norm_pre_clip': 0.0124056963250041, 'timing/data': 0.00020191000658087432, 'timing/model': 0.3934004029870266, 'learning_rate/qwen_vl_interface': 6.25964250198603e-07, 'learning_rate/action_model': 6.2596425019860294e-06, 'epoch': 0.34}) +[2026-10-03 13:57:34,043][starVLA.training.train_starvla][INFO] - Step 4641, Loss: {'train/loss': 0.2702995240688324, 'train/grad_norm_pre_clip': 91.32820129394531, 'timing/data': 0.0002526199968997389, 'timing/model': 0.3957388740091119, 'learning_rate/qwen_vl_interface': 6.25268508750342e-07, 'learning_rate/action_model': 6.25268508750342e-06, 'epoch': 0.34}) +[2026-10-03 13:57:34,750][starVLA.training.train_starvla][INFO] - Step 4642, Loss: {'train/loss': 4.231878847349435e-06, 'train/grad_norm_pre_clip': 0.002643536077812314, 'timing/data': 0.00031955000304151326, 'timing/model': 0.394579903993872, 'learning_rate/qwen_vl_interface': 6.245746683545062e-07, 'learning_rate/action_model': 6.245746683545062e-06, 'epoch': 0.34}) +[2026-10-03 13:57:35,455][starVLA.training.train_starvla][INFO] - Step 4643, Loss: {'train/loss': 9.916159797285218e-06, 'train/grad_norm_pre_clip': 0.002755113411694765, 'timing/data': 0.00027848999889101833, 'timing/model': 0.39418368399492465, 'learning_rate/qwen_vl_interface': 6.238827292963063e-07, 'learning_rate/action_model': 6.238827292963063e-06, 'epoch': 0.34}) +[2026-10-03 13:57:36,162][starVLA.training.train_starvla][INFO] - Step 4644, Loss: {'train/loss': 4.616325895767659e-05, 'train/grad_norm_pre_clip': 0.018465306609869003, 'timing/data': 0.00024002000282052904, 'timing/model': 0.3946266239945544, 'learning_rate/qwen_vl_interface': 6.231926918601731e-07, 'learning_rate/action_model': 6.2319269186017305e-06, 'epoch': 0.34}) +[2026-10-03 13:57:37,065][starVLA.training.train_starvla][INFO] - Step 4645, Loss: {'train/loss': 1.4542836652253754e-05, 'train/grad_norm_pre_clip': 0.0036441602278500795, 'timing/data': 0.00028457000735215843, 'timing/model': 0.39589604300272185, 'learning_rate/qwen_vl_interface': 6.22504556329754e-07, 'learning_rate/action_model': 6.22504556329754e-06, 'epoch': 0.34}) +[2026-10-03 13:57:37,772][starVLA.training.train_starvla][INFO] - Step 4646, Loss: {'train/loss': 2.403257167316042e-05, 'train/grad_norm_pre_clip': 0.014674066565930843, 'timing/data': 0.00020376000611577183, 'timing/model': 0.3945543629961321, 'learning_rate/qwen_vl_interface': 6.218183229879164e-07, 'learning_rate/action_model': 6.218183229879164e-06, 'epoch': 0.34}) +[2026-10-03 13:57:38,477][starVLA.training.train_starvla][INFO] - Step 4647, Loss: {'train/loss': 0.42048415541648865, 'train/grad_norm_pre_clip': 76.54867553710938, 'timing/data': 0.00021626999659929425, 'timing/model': 0.39437346399063244, 'learning_rate/qwen_vl_interface': 6.211339921167447e-07, 'learning_rate/action_model': 6.211339921167446e-06, 'epoch': 0.34}) +[2026-10-03 13:57:39,185][starVLA.training.train_starvla][INFO] - Step 4648, Loss: {'train/loss': 3.077057044720277e-06, 'train/grad_norm_pre_clip': 0.0016403869958594441, 'timing/data': 0.0003137199964839965, 'timing/model': 0.395429112992133, 'learning_rate/qwen_vl_interface': 6.204515639975415e-07, 'learning_rate/action_model': 6.204515639975414e-06, 'epoch': 0.34}) +[2026-10-03 13:57:39,894][starVLA.training.train_starvla][INFO] - Step 4649, Loss: {'train/loss': 4.121895108255558e-05, 'train/grad_norm_pre_clip': 0.028704728931188583, 'timing/data': 0.00022743000590708107, 'timing/model': 0.3963696139981039, 'learning_rate/qwen_vl_interface': 6.197710389108262e-07, 'learning_rate/action_model': 6.197710389108262e-06, 'epoch': 0.34}) +[2026-10-03 13:57:40,792][starVLA.training.train_starvla][INFO] - Step 4650, Loss: {'train/loss': 3.531877155182883e-05, 'train/grad_norm_pre_clip': 0.01745734177529812, 'timing/data': 0.00021969999943394214, 'timing/model': 0.585742517010658, 'learning_rate/qwen_vl_interface': 6.190924171363378e-07, 'learning_rate/action_model': 6.190924171363378e-06, 'epoch': 0.34}) +[2026-10-03 13:57:41,498][starVLA.training.train_starvla][INFO] - Step 4651, Loss: {'train/loss': 3.591377753764391e-05, 'train/grad_norm_pre_clip': 0.027500199154019356, 'timing/data': 0.0003063500043936074, 'timing/model': 0.39425553400360513, 'learning_rate/qwen_vl_interface': 6.184156989530326e-07, 'learning_rate/action_model': 6.184156989530325e-06, 'epoch': 0.34}) +[2026-10-03 13:57:42,205][starVLA.training.train_starvla][INFO] - Step 4652, Loss: {'train/loss': 0.00016445260553155094, 'train/grad_norm_pre_clip': 0.09752713143825531, 'timing/data': 0.00019739000708796084, 'timing/model': 0.3948970729979919, 'learning_rate/qwen_vl_interface': 6.17740884639083e-07, 'learning_rate/action_model': 6.177408846390829e-06, 'epoch': 0.34}) +[2026-10-03 13:57:42,913][starVLA.training.train_starvla][INFO] - Step 4653, Loss: {'train/loss': 0.5496013164520264, 'train/grad_norm_pre_clip': 79.90770721435547, 'timing/data': 0.00024901000142563134, 'timing/model': 0.3958504939946579, 'learning_rate/qwen_vl_interface': 6.170679744718798e-07, 'learning_rate/action_model': 6.1706797447187975e-06, 'epoch': 0.34}) +[2026-10-03 13:57:43,620][starVLA.training.train_starvla][INFO] - Step 4654, Loss: {'train/loss': 1.692590922175441e-05, 'train/grad_norm_pre_clip': 0.008454724214971066, 'timing/data': 0.000238540000282228, 'timing/model': 0.39442780399986077, 'learning_rate/qwen_vl_interface': 6.163969687280311e-07, 'learning_rate/action_model': 6.163969687280311e-06, 'epoch': 0.34}) +[2026-10-03 13:57:44,326][starVLA.training.train_starvla][INFO] - Step 4655, Loss: {'train/loss': 9.82709025265649e-06, 'train/grad_norm_pre_clip': 0.004102391190826893, 'timing/data': 0.00027766999846789986, 'timing/model': 0.39476919299340807, 'learning_rate/qwen_vl_interface': 6.157278676833627e-07, 'learning_rate/action_model': 6.157278676833627e-06, 'epoch': 0.34}) +[2026-10-03 13:57:45,223][starVLA.training.train_starvla][INFO] - Step 4656, Loss: {'train/loss': 4.1442250221734866e-05, 'train/grad_norm_pre_clip': 0.02788393571972847, 'timing/data': 0.00026225998590234667, 'timing/model': 0.39493174299423117, 'learning_rate/qwen_vl_interface': 6.150606716129155e-07, 'learning_rate/action_model': 6.150606716129154e-06, 'epoch': 0.34}) +[2026-10-03 13:57:45,932][starVLA.training.train_starvla][INFO] - Step 4657, Loss: {'train/loss': 0.06869499385356903, 'train/grad_norm_pre_clip': 154.70477294921875, 'timing/data': 0.00030834999051876366, 'timing/model': 0.3961342740076361, 'learning_rate/qwen_vl_interface': 6.143953807909497e-07, 'learning_rate/action_model': 6.143953807909496e-06, 'epoch': 0.34}) +[2026-10-03 13:57:46,640][starVLA.training.train_starvla][INFO] - Step 4658, Loss: {'train/loss': 1.3492090147337876e-05, 'train/grad_norm_pre_clip': 0.00395177723839879, 'timing/data': 0.0004158100055065006, 'timing/model': 0.3947824040078558, 'learning_rate/qwen_vl_interface': 6.137319954909407e-07, 'learning_rate/action_model': 6.137319954909406e-06, 'epoch': 0.34}) +[2026-10-03 13:57:47,346][starVLA.training.train_starvla][INFO] - Step 4659, Loss: {'train/loss': 4.529906618699897e-06, 'train/grad_norm_pre_clip': 0.001045366981998086, 'timing/data': 0.0003416099934838712, 'timing/model': 0.3943848839990096, 'learning_rate/qwen_vl_interface': 6.130705159855815e-07, 'learning_rate/action_model': 6.130705159855814e-06, 'epoch': 0.34}) +[2026-10-03 13:57:48,051][starVLA.training.train_starvla][INFO] - Step 4660, Loss: {'train/loss': 0.0001852086716098711, 'train/grad_norm_pre_clip': 0.21890778839588165, 'timing/data': 0.0002490499900886789, 'timing/model': 0.3929665440082317, 'learning_rate/qwen_vl_interface': 6.124109425467805e-07, 'learning_rate/action_model': 6.124109425467805e-06, 'epoch': 0.34}) +[2026-10-03 13:57:48,955][starVLA.training.train_starvla][INFO] - Step 4661, Loss: {'train/loss': 0.5959897637367249, 'train/grad_norm_pre_clip': 29.879295349121094, 'timing/data': 0.00031993001175578684, 'timing/model': 0.3937344839941943, 'learning_rate/qwen_vl_interface': 6.117532754456642e-07, 'learning_rate/action_model': 6.117532754456642e-06, 'epoch': 0.34}) +[2026-10-03 13:57:49,658][starVLA.training.train_starvla][INFO] - Step 4662, Loss: {'train/loss': 1.3492335710907355e-05, 'train/grad_norm_pre_clip': 0.003156630787998438, 'timing/data': 0.0003311100008431822, 'timing/model': 0.39260330398974475, 'learning_rate/qwen_vl_interface': 6.110975149525746e-07, 'learning_rate/action_model': 6.110975149525746e-06, 'epoch': 0.34}) +[2026-10-03 13:57:50,362][starVLA.training.train_starvla][INFO] - Step 4663, Loss: {'train/loss': 2.9802181416016538e-06, 'train/grad_norm_pre_clip': 0.0007796064601279795, 'timing/data': 0.00024356000358238816, 'timing/model': 0.3932396940072067, 'learning_rate/qwen_vl_interface': 6.104436613370692e-07, 'learning_rate/action_model': 6.104436613370692e-06, 'epoch': 0.34}) +[2026-10-03 13:57:51,065][starVLA.training.train_starvla][INFO] - Step 4664, Loss: {'train/loss': 1.8020766219706275e-05, 'train/grad_norm_pre_clip': 0.005483628250658512, 'timing/data': 0.00030110999068710953, 'timing/model': 0.3928054840071127, 'learning_rate/qwen_vl_interface': 6.09791714867924e-07, 'learning_rate/action_model': 6.0979171486792396e-06, 'epoch': 0.34}) +[2026-10-03 13:57:51,770][starVLA.training.train_starvla][INFO] - Step 4665, Loss: {'train/loss': 3.120813562418334e-05, 'train/grad_norm_pre_clip': 0.014560956507921219, 'timing/data': 0.00030051000067032874, 'timing/model': 0.39439255400793627, 'learning_rate/qwen_vl_interface': 6.091416758131287e-07, 'learning_rate/action_model': 6.091416758131286e-06, 'epoch': 0.34}) +[2026-10-03 13:57:52,669][starVLA.training.train_starvla][INFO] - Step 4666, Loss: {'train/loss': 5.1206934585934505e-05, 'train/grad_norm_pre_clip': 0.08868110179901123, 'timing/data': 0.0002408700092928484, 'timing/model': 0.5877454559959006, 'learning_rate/qwen_vl_interface': 6.084935444398895e-07, 'learning_rate/action_model': 6.084935444398894e-06, 'epoch': 0.34}) +[2026-10-03 13:57:53,371][starVLA.training.train_starvla][INFO] - Step 4667, Loss: {'train/loss': 1.8178490790887736e-05, 'train/grad_norm_pre_clip': 0.005549392197281122, 'timing/data': 0.00019994999456685036, 'timing/model': 0.3929434639867395, 'learning_rate/qwen_vl_interface': 6.078473210146294e-07, 'learning_rate/action_model': 6.078473210146294e-06, 'epoch': 0.34}) +[2026-10-03 13:57:54,075][starVLA.training.train_starvla][INFO] - Step 4668, Loss: {'train/loss': 3.9489394112024456e-05, 'train/grad_norm_pre_clip': 0.030681755393743515, 'timing/data': 0.00019977999909315258, 'timing/model': 0.39352071400207933, 'learning_rate/qwen_vl_interface': 6.072030058029862e-07, 'learning_rate/action_model': 6.072030058029861e-06, 'epoch': 0.34}) +[2026-10-03 13:57:54,779][starVLA.training.train_starvla][INFO] - Step 4669, Loss: {'train/loss': 1.2427004548953846e-05, 'train/grad_norm_pre_clip': 0.005690039601176977, 'timing/data': 0.00027137000870425254, 'timing/model': 0.3934850439982256, 'learning_rate/qwen_vl_interface': 6.065605990698136e-07, 'learning_rate/action_model': 6.065605990698136e-06, 'epoch': 0.34}) +[2026-10-03 13:57:55,482][starVLA.training.train_starvla][INFO] - Step 4670, Loss: {'train/loss': 1.1764052032958716e-05, 'train/grad_norm_pre_clip': 0.004927611444145441, 'timing/data': 0.00020083000708837062, 'timing/model': 0.39259477400628384, 'learning_rate/qwen_vl_interface': 6.059201010791811e-07, 'learning_rate/action_model': 6.059201010791811e-06, 'epoch': 0.34}) +[2026-10-03 13:57:56,185][starVLA.training.train_starvla][INFO] - Step 4671, Loss: {'train/loss': 0.0001453208242310211, 'train/grad_norm_pre_clip': 0.24235844612121582, 'timing/data': 0.00020634000247810036, 'timing/model': 0.3926464649994159, 'learning_rate/qwen_vl_interface': 6.052815120943721e-07, 'learning_rate/action_model': 6.052815120943721e-06, 'epoch': 0.34}) +[2026-10-03 13:57:57,079][starVLA.training.train_starvla][INFO] - Step 4672, Loss: {'train/loss': 1.6084573871921748e-05, 'train/grad_norm_pre_clip': 0.005715602543205023, 'timing/data': 0.00026277000142727047, 'timing/model': 0.393249934000778, 'learning_rate/qwen_vl_interface': 6.046448323778874e-07, 'learning_rate/action_model': 6.046448323778874e-06, 'epoch': 0.34}) +[2026-10-03 13:57:57,784][starVLA.training.train_starvla][INFO] - Step 4673, Loss: {'train/loss': 7.383442425634712e-05, 'train/grad_norm_pre_clip': 0.08467117697000504, 'timing/data': 0.0002937200042651966, 'timing/model': 0.39486265399318654, 'learning_rate/qwen_vl_interface': 6.040100621914419e-07, 'learning_rate/action_model': 6.040100621914418e-06, 'epoch': 0.34}) +[2026-10-03 13:57:58,487][starVLA.training.train_starvla][INFO] - Step 4674, Loss: {'train/loss': 2.243249946332071e-05, 'train/grad_norm_pre_clip': 0.011258579790592194, 'timing/data': 0.0003304999991087243, 'timing/model': 0.39240740399691276, 'learning_rate/qwen_vl_interface': 6.033772017959652e-07, 'learning_rate/action_model': 6.033772017959652e-06, 'epoch': 0.34}) +[2026-10-03 13:57:59,191][starVLA.training.train_starvla][INFO] - Step 4675, Loss: {'train/loss': 7.957017260196153e-06, 'train/grad_norm_pre_clip': 0.0017113772919401526, 'timing/data': 0.00031458999728783965, 'timing/model': 0.3927864739962388, 'learning_rate/qwen_vl_interface': 6.027462514516022e-07, 'learning_rate/action_model': 6.0274625145160214e-06, 'epoch': 0.34}) +[2026-10-03 13:57:59,894][starVLA.training.train_starvla][INFO] - Step 4676, Loss: {'train/loss': 0.3792558014392853, 'train/grad_norm_pre_clip': 77.11116790771484, 'timing/data': 0.0002699499891605228, 'timing/model': 0.3931538049946539, 'learning_rate/qwen_vl_interface': 6.021172114177139e-07, 'learning_rate/action_model': 6.021172114177139e-06, 'epoch': 0.34}) +[2026-10-03 13:58:00,792][starVLA.training.train_starvla][INFO] - Step 4677, Loss: {'train/loss': 3.7663216062355787e-05, 'train/grad_norm_pre_clip': 0.035224828869104385, 'timing/data': 0.0003407100011827424, 'timing/model': 0.3938304940093076, 'learning_rate/qwen_vl_interface': 6.014900819528734e-07, 'learning_rate/action_model': 6.0149008195287335e-06, 'epoch': 0.34}) +[2026-10-03 13:58:01,497][starVLA.training.train_starvla][INFO] - Step 4678, Loss: {'train/loss': 0.001862055971287191, 'train/grad_norm_pre_clip': 4.488716125488281, 'timing/data': 0.00025178000214509666, 'timing/model': 0.3937221140076872, 'learning_rate/qwen_vl_interface': 6.008648633148707e-07, 'learning_rate/action_model': 6.008648633148707e-06, 'epoch': 0.34}) +[2026-10-03 13:58:02,201][starVLA.training.train_starvla][INFO] - Step 4679, Loss: {'train/loss': 0.43455997109413147, 'train/grad_norm_pre_clip': 32.933616638183594, 'timing/data': 0.00031724000291433185, 'timing/model': 0.3927719939965755, 'learning_rate/qwen_vl_interface': 6.002415557607097e-07, 'learning_rate/action_model': 6.002415557607096e-06, 'epoch': 0.34}) +[2026-10-03 13:58:02,904][starVLA.training.train_starvla][INFO] - Step 4680, Loss: {'train/loss': 3.844657476292923e-05, 'train/grad_norm_pre_clip': 0.01472958829253912, 'timing/data': 0.0002578999992692843, 'timing/model': 0.39280470499943476, 'learning_rate/qwen_vl_interface': 5.996201595466088e-07, 'learning_rate/action_model': 5.996201595466088e-06, 'epoch': 0.34}) +[2026-10-03 13:58:03,608][starVLA.training.train_starvla][INFO] - Step 4681, Loss: {'train/loss': 5.006711489841109e-06, 'train/grad_norm_pre_clip': 0.002123909303918481, 'timing/data': 0.0002416299976175651, 'timing/model': 0.39375442400341853, 'learning_rate/qwen_vl_interface': 5.990006749279998e-07, 'learning_rate/action_model': 5.990006749279998e-06, 'epoch': 0.34}) +[2026-10-03 13:58:04,507][starVLA.training.train_starvla][INFO] - Step 4682, Loss: {'train/loss': 2.33938371820841e-05, 'train/grad_norm_pre_clip': 0.007522729691118002, 'timing/data': 0.0003939099988201633, 'timing/model': 0.3927981039887527, 'learning_rate/qwen_vl_interface': 5.983831021595296e-07, 'learning_rate/action_model': 5.9838310215952956e-06, 'epoch': 0.34}) +[2026-10-03 13:58:05,210][starVLA.training.train_starvla][INFO] - Step 4683, Loss: {'train/loss': 2.4426070012850687e-05, 'train/grad_norm_pre_clip': 0.026770802214741707, 'timing/data': 0.0002580399886937812, 'timing/model': 0.3923787740059197, 'learning_rate/qwen_vl_interface': 5.977674414950595e-07, 'learning_rate/action_model': 5.9776744149505955e-06, 'epoch': 0.34}) +[2026-10-03 13:58:05,913][starVLA.training.train_starvla][INFO] - Step 4684, Loss: {'train/loss': 0.3635820150375366, 'train/grad_norm_pre_clip': 63.52791976928711, 'timing/data': 0.000301070002024062, 'timing/model': 0.3930790150043322, 'learning_rate/qwen_vl_interface': 5.971536931876635e-07, 'learning_rate/action_model': 5.971536931876634e-06, 'epoch': 0.34}) +[2026-10-03 13:58:06,618][starVLA.training.train_starvla][INFO] - Step 4685, Loss: {'train/loss': 6.734136695740744e-05, 'train/grad_norm_pre_clip': 0.031603943556547165, 'timing/data': 0.00028000999009236693, 'timing/model': 0.39439277500787284, 'learning_rate/qwen_vl_interface': 5.965418574896315e-07, 'learning_rate/action_model': 5.9654185748963145e-06, 'epoch': 0.34}) +[2026-10-03 13:58:07,321][starVLA.training.train_starvla][INFO] - Step 4686, Loss: {'train/loss': 0.25026097893714905, 'train/grad_norm_pre_clip': 131.95350646972656, 'timing/data': 0.0003641700022853911, 'timing/model': 0.39235838399326894, 'learning_rate/qwen_vl_interface': 5.959319346524646e-07, 'learning_rate/action_model': 5.959319346524646e-06, 'epoch': 0.34}) +[2026-10-03 13:58:08,218][starVLA.training.train_starvla][INFO] - Step 4687, Loss: {'train/loss': 3.407475014682859e-05, 'train/grad_norm_pre_clip': 0.02079356648027897, 'timing/data': 0.00020970999321434647, 'timing/model': 0.5873260069929529, 'learning_rate/qwen_vl_interface': 5.953239249268804e-07, 'learning_rate/action_model': 5.953239249268804e-06, 'epoch': 0.34}) +[2026-10-03 13:58:08,923][starVLA.training.train_starvla][INFO] - Step 4688, Loss: {'train/loss': 9.864335879683495e-06, 'train/grad_norm_pre_clip': 0.0022439383901655674, 'timing/data': 0.0002296699967700988, 'timing/model': 0.39400921399646904, 'learning_rate/qwen_vl_interface': 5.947178285628081e-07, 'learning_rate/action_model': 5.94717828562808e-06, 'epoch': 0.34}) +[2026-10-03 13:58:09,629][starVLA.training.train_starvla][INFO] - Step 4689, Loss: {'train/loss': 1.961605084943585e-05, 'train/grad_norm_pre_clip': 0.003700332483276725, 'timing/data': 0.00029701998573727906, 'timing/model': 0.3956715239910409, 'learning_rate/qwen_vl_interface': 5.941136458093907e-07, 'learning_rate/action_model': 5.941136458093907e-06, 'epoch': 0.34}) +[2026-10-03 13:58:10,332][starVLA.training.train_starvla][INFO] - Step 4690, Loss: {'train/loss': 0.0008946435409598053, 'train/grad_norm_pre_clip': 5.104002952575684, 'timing/data': 0.00020276999566704035, 'timing/model': 0.3929383240028983, 'learning_rate/qwen_vl_interface': 5.935113769149855e-07, 'learning_rate/action_model': 5.935113769149856e-06, 'epoch': 0.34}) +[2026-10-03 13:58:11,036][starVLA.training.train_starvla][INFO] - Step 4691, Loss: {'train/loss': 0.33796244859695435, 'train/grad_norm_pre_clip': 82.2466049194336, 'timing/data': 0.0002565299946581945, 'timing/model': 0.3932315650017699, 'learning_rate/qwen_vl_interface': 5.92911022127163e-07, 'learning_rate/action_model': 5.929110221271629e-06, 'epoch': 0.34}) +[2026-10-03 13:58:11,936][starVLA.training.train_starvla][INFO] - Step 4692, Loss: {'train/loss': 5.875299757462926e-05, 'train/grad_norm_pre_clip': 0.11286775767803192, 'timing/data': 0.00029579999682027847, 'timing/model': 0.5900312359881354, 'learning_rate/qwen_vl_interface': 5.92312581692705e-07, 'learning_rate/action_model': 5.923125816927049e-06, 'epoch': 0.34}) +[2026-10-03 13:58:12,640][starVLA.training.train_starvla][INFO] - Step 4693, Loss: {'train/loss': 0.00018912339874077588, 'train/grad_norm_pre_clip': 0.1642788201570511, 'timing/data': 0.00024593000125605613, 'timing/model': 0.39396002401190344, 'learning_rate/qwen_vl_interface': 5.917160558576088e-07, 'learning_rate/action_model': 5.917160558576088e-06, 'epoch': 0.34}) +[2026-10-03 13:58:13,343][starVLA.training.train_starvla][INFO] - Step 4694, Loss: {'train/loss': 0.0006658266647718847, 'train/grad_norm_pre_clip': 1.9966245889663696, 'timing/data': 0.00020260999735910445, 'timing/model': 0.3928288850001991, 'learning_rate/qwen_vl_interface': 5.911214448670829e-07, 'learning_rate/action_model': 5.911214448670828e-06, 'epoch': 0.34}) +[2026-10-03 13:58:14,046][starVLA.training.train_starvla][INFO] - Step 4695, Loss: {'train/loss': 0.0024934562388807535, 'train/grad_norm_pre_clip': 5.797112464904785, 'timing/data': 0.0003329299943288788, 'timing/model': 0.39283149399852846, 'learning_rate/qwen_vl_interface': 5.905287489655499e-07, 'learning_rate/action_model': 5.905287489655499e-06, 'epoch': 0.34}) +[2026-10-03 13:58:14,749][starVLA.training.train_starvla][INFO] - Step 4696, Loss: {'train/loss': 3.1217832656693645e-06, 'train/grad_norm_pre_clip': 0.0006813836516812444, 'timing/data': 0.00022297000396065414, 'timing/model': 0.39310072499210946, 'learning_rate/qwen_vl_interface': 5.899379683966445e-07, 'learning_rate/action_model': 5.899379683966445e-06, 'epoch': 0.34}) +[2026-10-03 13:58:15,453][starVLA.training.train_starvla][INFO] - Step 4697, Loss: {'train/loss': 0.0002757965412456542, 'train/grad_norm_pre_clip': 0.26226484775543213, 'timing/data': 0.0003051800013054162, 'timing/model': 0.39423222400364466, 'learning_rate/qwen_vl_interface': 5.893491034032142e-07, 'learning_rate/action_model': 5.893491034032142e-06, 'epoch': 0.34}) +[2026-10-03 13:58:16,350][starVLA.training.train_starvla][INFO] - Step 4698, Loss: {'train/loss': 2.0301229596952908e-05, 'train/grad_norm_pre_clip': 0.013016230426728725, 'timing/data': 0.00022248999448493123, 'timing/model': 0.3932374139985768, 'learning_rate/qwen_vl_interface': 5.887621542273196e-07, 'learning_rate/action_model': 5.887621542273196e-06, 'epoch': 0.34}) +[2026-10-03 13:58:17,052][starVLA.training.train_starvla][INFO] - Step 4699, Loss: {'train/loss': 9.12671566766221e-06, 'train/grad_norm_pre_clip': 0.004342184402048588, 'timing/data': 0.0003302300028735772, 'timing/model': 0.3927227450039936, 'learning_rate/qwen_vl_interface': 5.881771211102327e-07, 'learning_rate/action_model': 5.881771211102327e-06, 'epoch': 0.34}) +[2026-10-03 13:58:17,755][starVLA.training.train_starvla][INFO] - Step 4700, Loss: {'train/loss': 2.1916413970757276e-05, 'train/grad_norm_pre_clip': 0.03298099339008331, 'timing/data': 0.00029857999470550567, 'timing/model': 0.39311027400253806, 'learning_rate/qwen_vl_interface': 5.875940042924391e-07, 'learning_rate/action_model': 5.875940042924391e-06, 'epoch': 0.34}) +[2026-10-03 13:58:18,460][starVLA.training.train_starvla][INFO] - Step 4701, Loss: {'train/loss': 2.592594682937488e-05, 'train/grad_norm_pre_clip': 0.006226486526429653, 'timing/data': 0.0002926199958892539, 'timing/model': 0.3943048849905608, 'learning_rate/qwen_vl_interface': 5.870128040136368e-07, 'learning_rate/action_model': 5.870128040136367e-06, 'epoch': 0.34}) +[2026-10-03 13:58:19,163][starVLA.training.train_starvla][INFO] - Step 4702, Loss: {'train/loss': 2.3489703380619176e-05, 'train/grad_norm_pre_clip': 0.009796197526156902, 'timing/data': 0.00033603000338189304, 'timing/model': 0.3927532340021571, 'learning_rate/qwen_vl_interface': 5.864335205127332e-07, 'learning_rate/action_model': 5.864335205127332e-06, 'epoch': 0.34}) +[2026-10-03 13:58:20,062][starVLA.training.train_starvla][INFO] - Step 4703, Loss: {'train/loss': 9.181098721455783e-05, 'train/grad_norm_pre_clip': 0.16899417340755463, 'timing/data': 0.0003483300097286701, 'timing/model': 0.5890066560095875, 'learning_rate/qwen_vl_interface': 5.858561540278513e-07, 'learning_rate/action_model': 5.858561540278513e-06, 'epoch': 0.34}) +[2026-10-03 13:58:20,765][starVLA.training.train_starvla][INFO] - Step 4704, Loss: {'train/loss': 3.4719482755463105e-06, 'train/grad_norm_pre_clip': 0.000970244116615504, 'timing/data': 0.00028668000595644116, 'timing/model': 0.3926427750120638, 'learning_rate/qwen_vl_interface': 5.852807047963244e-07, 'learning_rate/action_model': 5.852807047963243e-06, 'epoch': 0.34}) +[2026-10-03 13:58:21,469][starVLA.training.train_starvla][INFO] - Step 4705, Loss: {'train/loss': 0.0001581924589117989, 'train/grad_norm_pre_clip': 0.06691581755876541, 'timing/data': 0.0003059599985135719, 'timing/model': 0.39403089399274904, 'learning_rate/qwen_vl_interface': 5.847071730546974e-07, 'learning_rate/action_model': 5.847071730546974e-06, 'epoch': 0.34}) +[2026-10-03 13:58:22,172][starVLA.training.train_starvla][INFO] - Step 4706, Loss: {'train/loss': 0.00023306187358684838, 'train/grad_norm_pre_clip': 0.35531607270240784, 'timing/data': 0.0002498499961802736, 'timing/model': 0.39276595500996336, 'learning_rate/qwen_vl_interface': 5.841355590387291e-07, 'learning_rate/action_model': 5.841355590387291e-06, 'epoch': 0.34}) +[2026-10-03 13:58:22,874][starVLA.training.train_starvla][INFO] - Step 4707, Loss: {'train/loss': 3.0025721571291797e-06, 'train/grad_norm_pre_clip': 0.000903233652934432, 'timing/data': 0.00020414000027813017, 'timing/model': 0.39263897499768063, 'learning_rate/qwen_vl_interface': 5.835658629833866e-07, 'learning_rate/action_model': 5.8356586298338654e-06, 'epoch': 0.34}) +[2026-10-03 13:58:23,576][starVLA.training.train_starvla][INFO] - Step 4708, Loss: {'train/loss': 0.0008337743929587305, 'train/grad_norm_pre_clip': 3.9980053901672363, 'timing/data': 0.0002461300027789548, 'timing/model': 0.3926942050020443, 'learning_rate/qwen_vl_interface': 5.829980851228512e-07, 'learning_rate/action_model': 5.829980851228512e-06, 'epoch': 0.34}) +[2026-10-03 13:58:24,495][starVLA.training.train_starvla][INFO] - Step 4709, Loss: {'train/loss': 5.3846735681872815e-05, 'train/grad_norm_pre_clip': 0.04696030914783478, 'timing/data': 0.0003026899939868599, 'timing/model': 0.39368384500266984, 'learning_rate/qwen_vl_interface': 5.824322256905155e-07, 'learning_rate/action_model': 5.824322256905154e-06, 'epoch': 0.34}) +[2026-10-03 13:58:25,198][starVLA.training.train_starvla][INFO] - Step 4710, Loss: {'train/loss': 9.754046914167702e-05, 'train/grad_norm_pre_clip': 0.03681285306811333, 'timing/data': 0.0004310299991630018, 'timing/model': 0.3919989540008828, 'learning_rate/qwen_vl_interface': 5.818682849189818e-07, 'learning_rate/action_model': 5.818682849189817e-06, 'epoch': 0.34}) +[2026-10-03 13:58:25,901][starVLA.training.train_starvla][INFO] - Step 4711, Loss: {'train/loss': 0.20492041110992432, 'train/grad_norm_pre_clip': 208.21875, 'timing/data': 0.0002050500042969361, 'timing/model': 0.3928634850017261, 'learning_rate/qwen_vl_interface': 5.813062630400653e-07, 'learning_rate/action_model': 5.813062630400653e-06, 'epoch': 0.34}) +[2026-10-03 13:58:26,603][starVLA.training.train_starvla][INFO] - Step 4712, Loss: {'train/loss': 5.8336563597549684e-06, 'train/grad_norm_pre_clip': 0.0029986598528921604, 'timing/data': 0.0002505499869585037, 'timing/model': 0.3926082350080833, 'learning_rate/qwen_vl_interface': 5.807461602847922e-07, 'learning_rate/action_model': 5.807461602847922e-06, 'epoch': 0.34}) +[2026-10-03 13:58:27,308][starVLA.training.train_starvla][INFO] - Step 4713, Loss: {'train/loss': 1.543685357319191e-05, 'train/grad_norm_pre_clip': 0.008753455244004726, 'timing/data': 0.00028700999973807484, 'timing/model': 0.3946925740019651, 'learning_rate/qwen_vl_interface': 5.801879768833998e-07, 'learning_rate/action_model': 5.801879768833998e-06, 'epoch': 0.34}) +[2026-10-03 13:58:28,208][starVLA.training.train_starvla][INFO] - Step 4714, Loss: {'train/loss': 3.398537228349596e-05, 'train/grad_norm_pre_clip': 0.010324981063604355, 'timing/data': 0.0002828699944075197, 'timing/model': 0.39244977399357595, 'learning_rate/qwen_vl_interface': 5.796317130653363e-07, 'learning_rate/action_model': 5.796317130653363e-06, 'epoch': 0.35}) +[2026-10-03 13:58:28,910][starVLA.training.train_starvla][INFO] - Step 4715, Loss: {'train/loss': 9.603450052964035e-06, 'train/grad_norm_pre_clip': 0.0019422953482717276, 'timing/data': 0.00029394999728538096, 'timing/model': 0.3931496440054616, 'learning_rate/qwen_vl_interface': 5.790773690592592e-07, 'learning_rate/action_model': 5.7907736905925925e-06, 'epoch': 0.35}) +[2026-10-03 13:58:29,613][starVLA.training.train_starvla][INFO] - Step 4716, Loss: {'train/loss': 8.493651648677769e-07, 'train/grad_norm_pre_clip': 0.0002877035294659436, 'timing/data': 0.00024271001166198403, 'timing/model': 0.3926328350062249, 'learning_rate/qwen_vl_interface': 5.7852494509304e-07, 'learning_rate/action_model': 5.7852494509304e-06, 'epoch': 0.35}) +[2026-10-03 13:58:30,317][starVLA.training.train_starvla][INFO] - Step 4717, Loss: {'train/loss': 0.0002428079314995557, 'train/grad_norm_pre_clip': 0.3231566548347473, 'timing/data': 0.0003189999988535419, 'timing/model': 0.39393010499770753, 'learning_rate/qwen_vl_interface': 5.779744413937592e-07, 'learning_rate/action_model': 5.779744413937592e-06, 'epoch': 0.35}) +[2026-10-03 13:58:31,019][starVLA.training.train_starvla][INFO] - Step 4718, Loss: {'train/loss': 0.00021718029165640473, 'train/grad_norm_pre_clip': 0.22964966297149658, 'timing/data': 0.0002093799994327128, 'timing/model': 0.3925412940006936, 'learning_rate/qwen_vl_interface': 5.774258581877073e-07, 'learning_rate/action_model': 5.774258581877073e-06, 'epoch': 0.35}) +[2026-10-03 13:58:31,919][starVLA.training.train_starvla][INFO] - Step 4719, Loss: {'train/loss': 0.007038370706140995, 'train/grad_norm_pre_clip': 38.606658935546875, 'timing/data': 0.00031504999788012356, 'timing/model': 0.5898788270133082, 'learning_rate/qwen_vl_interface': 5.768791957003866e-07, 'learning_rate/action_model': 5.768791957003865e-06, 'epoch': 0.35}) +[2026-10-03 13:58:32,620][starVLA.training.train_starvla][INFO] - Step 4720, Loss: {'train/loss': 5.654943379340693e-06, 'train/grad_norm_pre_clip': 0.0016153826145455241, 'timing/data': 0.00020673000835813582, 'timing/model': 0.3931540839985246, 'learning_rate/qwen_vl_interface': 5.763344541565091e-07, 'learning_rate/action_model': 5.7633445415650905e-06, 'epoch': 0.35}) +[2026-10-03 13:58:33,323][starVLA.training.train_starvla][INFO] - Step 4721, Loss: {'train/loss': 0.0003018256393261254, 'train/grad_norm_pre_clip': 0.25514301657676697, 'timing/data': 0.00020178999693598598, 'timing/model': 0.39357908500824124, 'learning_rate/qwen_vl_interface': 5.75791633779998e-07, 'learning_rate/action_model': 5.75791633779998e-06, 'epoch': 0.35}) +[2026-10-03 13:58:34,025][starVLA.training.train_starvla][INFO] - Step 4722, Loss: {'train/loss': 2.38304946833523e-05, 'train/grad_norm_pre_clip': 0.022143933922052383, 'timing/data': 0.00019963999511674047, 'timing/model': 0.39250504499068484, 'learning_rate/qwen_vl_interface': 5.752507347939857e-07, 'learning_rate/action_model': 5.752507347939858e-06, 'epoch': 0.35}) +[2026-10-03 13:58:34,727][starVLA.training.train_starvla][INFO] - Step 4723, Loss: {'train/loss': 9.014930583362002e-06, 'train/grad_norm_pre_clip': 0.0022716510575264692, 'timing/data': 0.00020063000556547195, 'timing/model': 0.39244143399992026, 'learning_rate/qwen_vl_interface': 5.74711757420816e-07, 'learning_rate/action_model': 5.747117574208159e-06, 'epoch': 0.35}) +[2026-10-03 13:58:35,430][starVLA.training.train_starvla][INFO] - Step 4724, Loss: {'train/loss': 0.2803862690925598, 'train/grad_norm_pre_clip': 400.15728759765625, 'timing/data': 0.0002005300048040226, 'timing/model': 0.39251814399904106, 'learning_rate/qwen_vl_interface': 5.74174701882042e-07, 'learning_rate/action_model': 5.74174701882042e-06, 'epoch': 0.35}) +[2026-10-03 13:58:36,326][starVLA.training.train_starvla][INFO] - Step 4725, Loss: {'train/loss': 0.3819752037525177, 'train/grad_norm_pre_clip': 44.27903366088867, 'timing/data': 0.00021117999858688563, 'timing/model': 0.39422947401180863, 'learning_rate/qwen_vl_interface': 5.736395683984262e-07, 'learning_rate/action_model': 5.736395683984261e-06, 'epoch': 0.35}) +[2026-10-03 13:58:37,029][starVLA.training.train_starvla][INFO] - Step 4726, Loss: {'train/loss': 0.4436565935611725, 'train/grad_norm_pre_clip': 67.50634765625, 'timing/data': 0.0002051799965556711, 'timing/model': 0.3931219149962999, 'learning_rate/qwen_vl_interface': 5.731063571899421e-07, 'learning_rate/action_model': 5.731063571899421e-06, 'epoch': 0.35}) +[2026-10-03 13:58:37,731][starVLA.training.train_starvla][INFO] - Step 4727, Loss: {'train/loss': 6.891616067150608e-06, 'train/grad_norm_pre_clip': 0.0025525835808366537, 'timing/data': 0.0002718800096772611, 'timing/model': 0.39203409501351416, 'learning_rate/qwen_vl_interface': 5.72575068475773e-07, 'learning_rate/action_model': 5.72575068475773e-06, 'epoch': 0.35}) +[2026-10-03 13:58:38,433][starVLA.training.train_starvla][INFO] - Step 4728, Loss: {'train/loss': 1.8437925973557867e-05, 'train/grad_norm_pre_clip': 0.007754748687148094, 'timing/data': 0.00020110000332351774, 'timing/model': 0.39272155400249176, 'learning_rate/qwen_vl_interface': 5.720457024743115e-07, 'learning_rate/action_model': 5.720457024743115e-06, 'epoch': 0.35}) +[2026-10-03 13:58:39,137][starVLA.training.train_starvla][INFO] - Step 4729, Loss: {'train/loss': 2.8913167625432834e-05, 'train/grad_norm_pre_clip': 0.00801829807460308, 'timing/data': 0.0002015999925788492, 'timing/model': 0.3936739140044665, 'learning_rate/qwen_vl_interface': 5.715182594031601e-07, 'learning_rate/action_model': 5.7151825940316e-06, 'epoch': 0.35}) +[2026-10-03 13:58:40,034][starVLA.training.train_starvla][INFO] - Step 4730, Loss: {'train/loss': 4.731736044050194e-05, 'train/grad_norm_pre_clip': 0.011430620215833187, 'timing/data': 0.00023330999829340726, 'timing/model': 0.3925835349946283, 'learning_rate/qwen_vl_interface': 5.709927394791305e-07, 'learning_rate/action_model': 5.709927394791305e-06, 'epoch': 0.35}) +[2026-10-03 13:58:40,735][starVLA.training.train_starvla][INFO] - Step 4731, Loss: {'train/loss': 8.761637218412943e-06, 'train/grad_norm_pre_clip': 0.0025379115249961615, 'timing/data': 0.00020197000412736088, 'timing/model': 0.39180951500020456, 'learning_rate/qwen_vl_interface': 5.70469142918244e-07, 'learning_rate/action_model': 5.704691429182441e-06, 'epoch': 0.35}) +[2026-10-03 13:58:41,437][starVLA.training.train_starvla][INFO] - Step 4732, Loss: {'train/loss': 4.497340341913514e-05, 'train/grad_norm_pre_clip': 0.011562810279428959, 'timing/data': 0.00019957000040449202, 'timing/model': 0.3921731250011362, 'learning_rate/qwen_vl_interface': 5.699474699357317e-07, 'learning_rate/action_model': 5.699474699357317e-06, 'epoch': 0.35}) +[2026-10-03 13:58:42,141][starVLA.training.train_starvla][INFO] - Step 4733, Loss: {'train/loss': 8.277269444079138e-06, 'train/grad_norm_pre_clip': 0.005811589770019054, 'timing/data': 0.0002398999931756407, 'timing/model': 0.39335844400920905, 'learning_rate/qwen_vl_interface': 5.694277207460331e-07, 'learning_rate/action_model': 5.694277207460332e-06, 'epoch': 0.35}) +[2026-10-03 13:58:42,847][starVLA.training.train_starvla][INFO] - Step 4734, Loss: {'train/loss': 0.0004231202183291316, 'train/grad_norm_pre_clip': 0.3565821647644043, 'timing/data': 0.00022244999126996845, 'timing/model': 0.3928999250056222, 'learning_rate/qwen_vl_interface': 5.689098955627982e-07, 'learning_rate/action_model': 5.689098955627982e-06, 'epoch': 0.35}) +[2026-10-03 13:58:43,746][starVLA.training.train_starvla][INFO] - Step 4735, Loss: {'train/loss': 1.7381498764734715e-05, 'train/grad_norm_pre_clip': 0.009439648129045963, 'timing/data': 0.00030899999546818435, 'timing/model': 0.3931313950015465, 'learning_rate/qwen_vl_interface': 5.683939945988853e-07, 'learning_rate/action_model': 5.683939945988853e-06, 'epoch': 0.35}) +[2026-10-03 13:58:44,449][starVLA.training.train_starvla][INFO] - Step 4736, Loss: {'train/loss': 1.8014001398114488e-05, 'train/grad_norm_pre_clip': 0.011965461075305939, 'timing/data': 0.00021587000810541213, 'timing/model': 0.39274265500716865, 'learning_rate/qwen_vl_interface': 5.678800180663608e-07, 'learning_rate/action_model': 5.678800180663609e-06, 'epoch': 0.35}) +[2026-10-03 13:58:45,152][starVLA.training.train_starvla][INFO] - Step 4737, Loss: {'train/loss': 0.0004902834189124405, 'train/grad_norm_pre_clip': 1.693126916885376, 'timing/data': 0.00029337999876588583, 'timing/model': 0.3935737540014088, 'learning_rate/qwen_vl_interface': 5.673679661765018e-07, 'learning_rate/action_model': 5.6736796617650175e-06, 'epoch': 0.35}) +[2026-10-03 13:58:45,854][starVLA.training.train_starvla][INFO] - Step 4738, Loss: {'train/loss': 0.011188271455466747, 'train/grad_norm_pre_clip': 42.95156478881836, 'timing/data': 0.00031662000401411206, 'timing/model': 0.3929512339964276, 'learning_rate/qwen_vl_interface': 5.668578391397938e-07, 'learning_rate/action_model': 5.668578391397937e-06, 'epoch': 0.35}) +[2026-10-03 13:58:46,557][starVLA.training.train_starvla][INFO] - Step 4739, Loss: {'train/loss': 4.470756903174333e-05, 'train/grad_norm_pre_clip': 0.04164586961269379, 'timing/data': 0.0003075099957641214, 'timing/model': 0.3926275839912705, 'learning_rate/qwen_vl_interface': 5.663496371659303e-07, 'learning_rate/action_model': 5.663496371659303e-06, 'epoch': 0.35}) +[2026-10-03 13:58:47,456][starVLA.training.train_starvla][INFO] - Step 4740, Loss: {'train/loss': 0.0001872692082542926, 'train/grad_norm_pre_clip': 0.1671844720840454, 'timing/data': 0.0002621400053612888, 'timing/model': 0.5893961770052556, 'learning_rate/qwen_vl_interface': 5.658433604638141e-07, 'learning_rate/action_model': 5.65843360463814e-06, 'epoch': 0.35}) +[2026-10-03 13:58:48,161][starVLA.training.train_starvla][INFO] - Step 4741, Loss: {'train/loss': 0.12323198467493057, 'train/grad_norm_pre_clip': 222.0749053955078, 'timing/data': 0.00021210999693721533, 'timing/model': 0.39396632500574924, 'learning_rate/qwen_vl_interface': 5.653390092415559e-07, 'learning_rate/action_model': 5.653390092415559e-06, 'epoch': 0.35}) +[2026-10-03 13:58:48,863][starVLA.training.train_starvla][INFO] - Step 4742, Loss: {'train/loss': 5.12504338985309e-05, 'train/grad_norm_pre_clip': 0.016284167766571045, 'timing/data': 0.00029515998903661966, 'timing/model': 0.3920382350042928, 'learning_rate/qwen_vl_interface': 5.648365837064759e-07, 'learning_rate/action_model': 5.648365837064759e-06, 'epoch': 0.35}) +[2026-10-03 13:58:49,566][starVLA.training.train_starvla][INFO] - Step 4743, Loss: {'train/loss': 2.409998342045583e-05, 'train/grad_norm_pre_clip': 0.009271074086427689, 'timing/data': 0.00030656000308226794, 'timing/model': 0.3932881840009941, 'learning_rate/qwen_vl_interface': 5.643360840651025e-07, 'learning_rate/action_model': 5.643360840651025e-06, 'epoch': 0.35}) +[2026-10-03 13:58:50,269][starVLA.training.train_starvla][INFO] - Step 4744, Loss: {'train/loss': 9.134534047916532e-05, 'train/grad_norm_pre_clip': 0.07707300037145615, 'timing/data': 0.0002021690015681088, 'timing/model': 0.39222661600797437, 'learning_rate/qwen_vl_interface': 5.638375105231719e-07, 'learning_rate/action_model': 5.638375105231719e-06, 'epoch': 0.35}) +[2026-10-03 13:58:51,166][starVLA.training.train_starvla][INFO] - Step 4745, Loss: {'train/loss': 6.995906005613506e-06, 'train/grad_norm_pre_clip': 0.0016901063499972224, 'timing/data': 0.00024182000197470188, 'timing/model': 0.5867510969983414, 'learning_rate/qwen_vl_interface': 5.633408632856294e-07, 'learning_rate/action_model': 5.6334086328562945e-06, 'epoch': 0.35}) +[2026-10-03 13:58:51,869][starVLA.training.train_starvla][INFO] - Step 4746, Loss: {'train/loss': 0.7378814220428467, 'train/grad_norm_pre_clip': 27.078611373901367, 'timing/data': 0.00020809999841731042, 'timing/model': 0.39295215500169434, 'learning_rate/qwen_vl_interface': 5.628461425566268e-07, 'learning_rate/action_model': 5.6284614255662675e-06, 'epoch': 0.35}) +[2026-10-03 13:58:52,570][starVLA.training.train_starvla][INFO] - Step 4747, Loss: {'train/loss': 0.0001037069596350193, 'train/grad_norm_pre_clip': 0.035643305629491806, 'timing/data': 0.00031934998696669936, 'timing/model': 0.3918007949978346, 'learning_rate/qwen_vl_interface': 5.62353348539526e-07, 'learning_rate/action_model': 5.623533485395259e-06, 'epoch': 0.35}) +[2026-10-03 13:58:53,273][starVLA.training.train_starvla][INFO] - Step 4748, Loss: {'train/loss': 7.524946795456344e-06, 'train/grad_norm_pre_clip': 0.0024988213554024696, 'timing/data': 0.0002506299933884293, 'timing/model': 0.3929518939985428, 'learning_rate/qwen_vl_interface': 5.618624814368956e-07, 'learning_rate/action_model': 5.618624814368956e-06, 'epoch': 0.35}) +[2026-10-03 13:58:53,976][starVLA.training.train_starvla][INFO] - Step 4749, Loss: {'train/loss': 0.00023128789325710386, 'train/grad_norm_pre_clip': 0.2925160825252533, 'timing/data': 0.00025656999787315726, 'timing/model': 0.39355229400098324, 'learning_rate/qwen_vl_interface': 5.613735414505126e-07, 'learning_rate/action_model': 5.613735414505125e-06, 'epoch': 0.35}) +[2026-10-03 13:58:54,679][starVLA.training.train_starvla][INFO] - Step 4750, Loss: {'train/loss': 2.093604507535929e-06, 'train/grad_norm_pre_clip': 0.000558086670935154, 'timing/data': 0.0002693599963095039, 'timing/model': 0.3923987250018399, 'learning_rate/qwen_vl_interface': 5.608865287813617e-07, 'learning_rate/action_model': 5.608865287813616e-06, 'epoch': 0.35}) +[2026-10-03 13:58:55,573][starVLA.training.train_starvla][INFO] - Step 4751, Loss: {'train/loss': 0.1559622585773468, 'train/grad_norm_pre_clip': 196.91722106933594, 'timing/data': 0.00020788999972864985, 'timing/model': 0.3920832250005333, 'learning_rate/qwen_vl_interface': 5.604014436296362e-07, 'learning_rate/action_model': 5.604014436296362e-06, 'epoch': 0.35}) +[2026-10-03 13:58:56,275][starVLA.training.train_starvla][INFO] - Step 4752, Loss: {'train/loss': 9.015195701067569e-07, 'train/grad_norm_pre_clip': 0.00019161947420798242, 'timing/data': 0.0002530799974920228, 'timing/model': 0.39213699399260804, 'learning_rate/qwen_vl_interface': 5.599182861947348e-07, 'learning_rate/action_model': 5.599182861947347e-06, 'epoch': 0.35}) +[2026-10-03 13:58:56,979][starVLA.training.train_starvla][INFO] - Step 4753, Loss: {'train/loss': 6.1987248045625165e-06, 'train/grad_norm_pre_clip': 0.0013677983079105616, 'timing/data': 0.00024942999880295247, 'timing/model': 0.3939226849906845, 'learning_rate/qwen_vl_interface': 5.594370566752665e-07, 'learning_rate/action_model': 5.5943705667526645e-06, 'epoch': 0.35}) +[2026-10-03 13:58:57,681][starVLA.training.train_starvla][INFO] - Step 4754, Loss: {'train/loss': 9.089514605875593e-06, 'train/grad_norm_pre_clip': 0.0030790874734520912, 'timing/data': 0.0003247700078645721, 'timing/model': 0.3926429449929856, 'learning_rate/qwen_vl_interface': 5.589577552690456e-07, 'learning_rate/action_model': 5.589577552690456e-06, 'epoch': 0.35}) +[2026-10-03 13:58:58,383][starVLA.training.train_starvla][INFO] - Step 4755, Loss: {'train/loss': 0.00010318973363609985, 'train/grad_norm_pre_clip': 0.05653148144483566, 'timing/data': 0.0001992799952859059, 'timing/model': 0.39238015399314463, 'learning_rate/qwen_vl_interface': 5.584803821730961e-07, 'learning_rate/action_model': 5.58480382173096e-06, 'epoch': 0.35}) +[2026-10-03 13:58:59,280][starVLA.training.train_starvla][INFO] - Step 4756, Loss: {'train/loss': 0.2868223786354065, 'train/grad_norm_pre_clip': 287.28033447265625, 'timing/data': 0.00020676999702118337, 'timing/model': 0.5863291480054613, 'learning_rate/qwen_vl_interface': 5.580049375836466e-07, 'learning_rate/action_model': 5.580049375836466e-06, 'epoch': 0.35}) +[2026-10-03 13:58:59,984][starVLA.training.train_starvla][INFO] - Step 4757, Loss: {'train/loss': 3.1437346478924155e-05, 'train/grad_norm_pre_clip': 0.009991894476115704, 'timing/data': 0.0002361699880566448, 'timing/model': 0.39389931500772946, 'learning_rate/qwen_vl_interface': 5.575314216961356e-07, 'learning_rate/action_model': 5.575314216961356e-06, 'epoch': 0.35}) +[2026-10-03 13:59:00,687][starVLA.training.train_starvla][INFO] - Step 4758, Loss: {'train/loss': 2.302770189999137e-05, 'train/grad_norm_pre_clip': 0.02199653536081314, 'timing/data': 0.00023303000489249825, 'timing/model': 0.3921762550016865, 'learning_rate/qwen_vl_interface': 5.570598347052073e-07, 'learning_rate/action_model': 5.570598347052072e-06, 'epoch': 0.35}) +[2026-10-03 13:59:01,390][starVLA.training.train_starvla][INFO] - Step 4759, Loss: {'train/loss': 4.075414835824631e-06, 'train/grad_norm_pre_clip': 0.0015621279599145055, 'timing/data': 0.00020302999473642558, 'timing/model': 0.39228374599770177, 'learning_rate/qwen_vl_interface': 5.565901768047126e-07, 'learning_rate/action_model': 5.5659017680471255e-06, 'epoch': 0.35}) +[2026-10-03 13:59:02,093][starVLA.training.train_starvla][INFO] - Step 4760, Loss: {'train/loss': 0.000555386592168361, 'train/grad_norm_pre_clip': 0.2861383855342865, 'timing/data': 0.0002090699999826029, 'timing/model': 0.3924395950016333, 'learning_rate/qwen_vl_interface': 5.561224481877105e-07, 'learning_rate/action_model': 5.561224481877105e-06, 'epoch': 0.35}) +[2026-10-03 13:59:02,795][starVLA.training.train_starvla][INFO] - Step 4761, Loss: {'train/loss': 0.24832738935947418, 'train/grad_norm_pre_clip': 187.74185180664062, 'timing/data': 0.00024357899383176118, 'timing/model': 0.3934504959906917, 'learning_rate/qwen_vl_interface': 5.55656649046467e-07, 'learning_rate/action_model': 5.556566490464669e-06, 'epoch': 0.35}) +[2026-10-03 13:59:03,688][starVLA.training.train_starvla][INFO] - Step 4762, Loss: {'train/loss': 1.2218002666486427e-05, 'train/grad_norm_pre_clip': 0.007415162865072489, 'timing/data': 0.0002156489936169237, 'timing/model': 0.392491684993729, 'learning_rate/qwen_vl_interface': 5.551927795724544e-07, 'learning_rate/action_model': 5.551927795724543e-06, 'epoch': 0.35}) +[2026-10-03 13:59:04,389][starVLA.training.train_starvla][INFO] - Step 4763, Loss: {'train/loss': 1.840289996835054e-06, 'train/grad_norm_pre_clip': 0.0004329531511757523, 'timing/data': 0.00020244999905116856, 'timing/model': 0.39233102501020767, 'learning_rate/qwen_vl_interface': 5.547308399563515e-07, 'learning_rate/action_model': 5.547308399563514e-06, 'epoch': 0.35}) +[2026-10-03 13:59:05,092][starVLA.training.train_starvla][INFO] - Step 4764, Loss: {'train/loss': 5.525286906049587e-05, 'train/grad_norm_pre_clip': 0.05686936154961586, 'timing/data': 0.0002908899914473295, 'timing/model': 0.39273097600380424, 'learning_rate/qwen_vl_interface': 5.542708303880442e-07, 'learning_rate/action_model': 5.542708303880442e-06, 'epoch': 0.35}) +[2026-10-03 13:59:05,795][starVLA.training.train_starvla][INFO] - Step 4765, Loss: {'train/loss': 0.002575450111180544, 'train/grad_norm_pre_clip': 6.177606105804443, 'timing/data': 0.00021438999101519585, 'timing/model': 0.3937182549998397, 'learning_rate/qwen_vl_interface': 5.538127510566258e-07, 'learning_rate/action_model': 5.538127510566258e-06, 'epoch': 0.35}) +[2026-10-03 13:59:06,498][starVLA.training.train_starvla][INFO] - Step 4766, Loss: {'train/loss': 0.0004809980746358633, 'train/grad_norm_pre_clip': 0.4984286427497864, 'timing/data': 0.0004226000019116327, 'timing/model': 0.3928435650013853, 'learning_rate/qwen_vl_interface': 5.533566021503944e-07, 'learning_rate/action_model': 5.533566021503943e-06, 'epoch': 0.35}) +[2026-10-03 13:59:07,394][starVLA.training.train_starvla][INFO] - Step 4767, Loss: {'train/loss': 1.460311068512965e-06, 'train/grad_norm_pre_clip': 0.00028085598023608327, 'timing/data': 0.00027363999106455594, 'timing/model': 0.3931690850004088, 'learning_rate/qwen_vl_interface': 5.529023838568566e-07, 'learning_rate/action_model': 5.529023838568566e-06, 'epoch': 0.35}) +[2026-10-03 13:59:08,097][starVLA.training.train_starvla][INFO] - Step 4768, Loss: {'train/loss': 2.311649950570427e-05, 'train/grad_norm_pre_clip': 0.019775807857513428, 'timing/data': 0.00039057999674696475, 'timing/model': 0.39295798499369994, 'learning_rate/qwen_vl_interface': 5.524500963627239e-07, 'learning_rate/action_model': 5.524500963627239e-06, 'epoch': 0.35}) +[2026-10-03 13:59:08,800][starVLA.training.train_starvla][INFO] - Step 4769, Loss: {'train/loss': 4.656571036321111e-06, 'train/grad_norm_pre_clip': 0.001623435178771615, 'timing/data': 0.00025069000548683107, 'timing/model': 0.3934300649998477, 'learning_rate/qwen_vl_interface': 5.519997398539145e-07, 'learning_rate/action_model': 5.519997398539145e-06, 'epoch': 0.35}) +[2026-10-03 13:59:09,501][starVLA.training.train_starvla][INFO] - Step 4770, Loss: {'train/loss': 1.9854398487950675e-05, 'train/grad_norm_pre_clip': 0.006140848156064749, 'timing/data': 0.000198710011318326, 'timing/model': 0.39159648498753086, 'learning_rate/qwen_vl_interface': 5.515513145155531e-07, 'learning_rate/action_model': 5.515513145155531e-06, 'epoch': 0.35}) +[2026-10-03 13:59:10,202][starVLA.training.train_starvla][INFO] - Step 4771, Loss: {'train/loss': 0.43378594517707825, 'train/grad_norm_pre_clip': 58.047706604003906, 'timing/data': 0.000198540001292713, 'timing/model': 0.39172041400161106, 'learning_rate/qwen_vl_interface': 5.511048205319702e-07, 'learning_rate/action_model': 5.511048205319701e-06, 'epoch': 0.35}) +[2026-10-03 13:59:11,096][starVLA.training.train_starvla][INFO] - Step 4772, Loss: {'train/loss': 4.8092191718751565e-05, 'train/grad_norm_pre_clip': 0.05970008671283722, 'timing/data': 0.00020673999097198248, 'timing/model': 0.5845633979915874, 'learning_rate/qwen_vl_interface': 5.506602580867034e-07, 'learning_rate/action_model': 5.506602580867034e-06, 'epoch': 0.35}) +[2026-10-03 13:59:11,799][starVLA.training.train_starvla][INFO] - Step 4773, Loss: {'train/loss': 0.46604710817337036, 'train/grad_norm_pre_clip': 186.1243133544922, 'timing/data': 0.00025308001204393804, 'timing/model': 0.3935717449930962, 'learning_rate/qwen_vl_interface': 5.502176273624942e-07, 'learning_rate/action_model': 5.502176273624942e-06, 'epoch': 0.35}) +[2026-10-03 13:59:12,500][starVLA.training.train_starvla][INFO] - Step 4774, Loss: {'train/loss': 8.751280984142795e-05, 'train/grad_norm_pre_clip': 0.027729308232665062, 'timing/data': 0.0003123299975413829, 'timing/model': 0.39170566499524284, 'learning_rate/qwen_vl_interface': 5.497769285412921e-07, 'learning_rate/action_model': 5.497769285412921e-06, 'epoch': 0.35}) +[2026-10-03 13:59:13,200][starVLA.training.train_starvla][INFO] - Step 4775, Loss: {'train/loss': 6.7127512011211365e-06, 'train/grad_norm_pre_clip': 0.004187203943729401, 'timing/data': 0.0001975699997274205, 'timing/model': 0.3914434350008378, 'learning_rate/qwen_vl_interface': 5.493381618042519e-07, 'learning_rate/action_model': 5.493381618042518e-06, 'epoch': 0.35}) +[2026-10-03 13:59:13,902][starVLA.training.train_starvla][INFO] - Step 4776, Loss: {'train/loss': 6.613722507609054e-05, 'train/grad_norm_pre_clip': 0.04128820821642876, 'timing/data': 0.00020367000252008438, 'timing/model': 0.39247751400398556, 'learning_rate/qwen_vl_interface': 5.489013273317343e-07, 'learning_rate/action_model': 5.489013273317343e-06, 'epoch': 0.35}) +[2026-10-03 13:59:14,605][starVLA.training.train_starvla][INFO] - Step 4777, Loss: {'train/loss': 2.3348342438112013e-05, 'train/grad_norm_pre_clip': 0.014975055120885372, 'timing/data': 0.0002437599905533716, 'timing/model': 0.3930408550077118, 'learning_rate/qwen_vl_interface': 5.484664253033042e-07, 'learning_rate/action_model': 5.484664253033042e-06, 'epoch': 0.35}) +[2026-10-03 13:59:15,497][starVLA.training.train_starvla][INFO] - Step 4778, Loss: {'train/loss': 1.61523148562992e-05, 'train/grad_norm_pre_clip': 0.0044745756313204765, 'timing/data': 0.00021297000057529658, 'timing/model': 0.3923818549956195, 'learning_rate/qwen_vl_interface': 5.480334558977351e-07, 'learning_rate/action_model': 5.480334558977351e-06, 'epoch': 0.35}) +[2026-10-03 13:59:16,198][starVLA.training.train_starvla][INFO] - Step 4779, Loss: {'train/loss': 7.510829163948074e-05, 'train/grad_norm_pre_clip': 0.17828376591205597, 'timing/data': 0.00019536000036168844, 'timing/model': 0.3917732550034998, 'learning_rate/qwen_vl_interface': 5.476024192930029e-07, 'learning_rate/action_model': 5.47602419293003e-06, 'epoch': 0.35}) +[2026-10-03 13:59:16,900][starVLA.training.train_starvla][INFO] - Step 4780, Loss: {'train/loss': 5.207897629588842e-06, 'train/grad_norm_pre_clip': 0.0012644893722608685, 'timing/data': 0.00028078000468667597, 'timing/model': 0.3921003150026081, 'learning_rate/qwen_vl_interface': 5.471733156662914e-07, 'learning_rate/action_model': 5.4717331566629135e-06, 'epoch': 0.35}) +[2026-10-03 13:59:17,602][starVLA.training.train_starvla][INFO] - Step 4781, Loss: {'train/loss': 4.5746123760181945e-06, 'train/grad_norm_pre_clip': 0.0012544753262773156, 'timing/data': 0.000235269995755516, 'timing/model': 0.39312667498597875, 'learning_rate/qwen_vl_interface': 5.467461451939882e-07, 'learning_rate/action_model': 5.467461451939882e-06, 'epoch': 0.35}) +[2026-10-03 13:59:18,303][starVLA.training.train_starvla][INFO] - Step 4782, Loss: {'train/loss': 1.8030378896582988e-06, 'train/grad_norm_pre_clip': 0.00040593702578917146, 'timing/data': 0.00019987999985460192, 'timing/model': 0.39178172599349637, 'learning_rate/qwen_vl_interface': 5.463209080516873e-07, 'learning_rate/action_model': 5.463209080516873e-06, 'epoch': 0.35}) +[2026-10-03 13:59:19,197][starVLA.training.train_starvla][INFO] - Step 4783, Loss: {'train/loss': 1.2359442735032644e-05, 'train/grad_norm_pre_clip': 0.002977363532409072, 'timing/data': 0.00021456999820657074, 'timing/model': 0.3920227350026835, 'learning_rate/qwen_vl_interface': 5.458976044141875e-07, 'learning_rate/action_model': 5.458976044141875e-06, 'epoch': 0.35}) +[2026-10-03 13:59:19,899][starVLA.training.train_starvla][INFO] - Step 4784, Loss: {'train/loss': 0.5776509642601013, 'train/grad_norm_pre_clip': 87.15374755859375, 'timing/data': 0.00023581999994348735, 'timing/model': 0.39236471500771586, 'learning_rate/qwen_vl_interface': 5.454762344554931e-07, 'learning_rate/action_model': 5.4547623445549316e-06, 'epoch': 0.35}) +[2026-10-03 13:59:20,602][starVLA.training.train_starvla][INFO] - Step 4785, Loss: {'train/loss': 0.14146994054317474, 'train/grad_norm_pre_clip': 100.03706359863281, 'timing/data': 0.0002985399914905429, 'timing/model': 0.3933021050033858, 'learning_rate/qwen_vl_interface': 5.450567983488137e-07, 'learning_rate/action_model': 5.450567983488137e-06, 'epoch': 0.35}) +[2026-10-03 13:59:21,305][starVLA.training.train_starvla][INFO] - Step 4786, Loss: {'train/loss': 1.796196011127904e-05, 'train/grad_norm_pre_clip': 0.005031461827456951, 'timing/data': 0.0003103099879808724, 'timing/model': 0.3927994550031144, 'learning_rate/qwen_vl_interface': 5.446392962665629e-07, 'learning_rate/action_model': 5.4463929626656296e-06, 'epoch': 0.35}) +[2026-10-03 13:59:22,007][starVLA.training.train_starvla][INFO] - Step 4787, Loss: {'train/loss': 1.5071787856868468e-05, 'train/grad_norm_pre_clip': 0.0032695825211703777, 'timing/data': 0.00029075000202283263, 'timing/model': 0.39301756500208285, 'learning_rate/qwen_vl_interface': 5.442237283803605e-07, 'learning_rate/action_model': 5.442237283803604e-06, 'epoch': 0.35}) +[2026-10-03 13:59:22,907][starVLA.training.train_starvla][INFO] - Step 4788, Loss: {'train/loss': 1.4646492672909517e-05, 'train/grad_norm_pre_clip': 0.015780432149767876, 'timing/data': 0.0003206699911970645, 'timing/model': 0.39270938499248587, 'learning_rate/qwen_vl_interface': 5.43810094861031e-07, 'learning_rate/action_model': 5.43810094861031e-06, 'epoch': 0.35}) +[2026-10-03 13:59:23,609][starVLA.training.train_starvla][INFO] - Step 4789, Loss: {'train/loss': 5.820070146000944e-05, 'train/grad_norm_pre_clip': 0.03052273765206337, 'timing/data': 0.00034389999927952886, 'timing/model': 0.392960925004445, 'learning_rate/qwen_vl_interface': 5.433983958786027e-07, 'learning_rate/action_model': 5.433983958786027e-06, 'epoch': 0.35}) +[2026-10-03 13:59:24,310][starVLA.training.train_starvla][INFO] - Step 4790, Loss: {'train/loss': 0.00012371577031444758, 'train/grad_norm_pre_clip': 0.08329938352108002, 'timing/data': 0.0002098400000249967, 'timing/model': 0.39186295500257984, 'learning_rate/qwen_vl_interface': 5.429886316023102e-07, 'learning_rate/action_model': 5.429886316023101e-06, 'epoch': 0.35}) +[2026-10-03 13:59:25,010][starVLA.training.train_starvla][INFO] - Step 4791, Loss: {'train/loss': 0.00012921373127028346, 'train/grad_norm_pre_clip': 0.1370566189289093, 'timing/data': 0.00030717000481672585, 'timing/model': 0.39160109499061946, 'learning_rate/qwen_vl_interface': 5.425808022005916e-07, 'learning_rate/action_model': 5.425808022005916e-06, 'epoch': 0.35}) +[2026-10-03 13:59:25,713][starVLA.training.train_starvla][INFO] - Step 4792, Loss: {'train/loss': 0.00031484104692935944, 'train/grad_norm_pre_clip': 0.21514594554901123, 'timing/data': 0.0001963300019269809, 'timing/model': 0.39255748499999754, 'learning_rate/qwen_vl_interface': 5.421749078410903e-07, 'learning_rate/action_model': 5.421749078410903e-06, 'epoch': 0.35}) +[2026-10-03 13:59:26,609][starVLA.training.train_starvla][INFO] - Step 4793, Loss: {'train/loss': 0.4525090157985687, 'train/grad_norm_pre_clip': 38.73763656616211, 'timing/data': 0.00025384000036865473, 'timing/model': 0.5864345080044586, 'learning_rate/qwen_vl_interface': 5.417709486906547e-07, 'learning_rate/action_model': 5.417709486906547e-06, 'epoch': 0.35}) +[2026-10-03 13:59:27,314][starVLA.training.train_starvla][INFO] - Step 4794, Loss: {'train/loss': 0.28798267245292664, 'train/grad_norm_pre_clip': 108.8468246459961, 'timing/data': 0.00033792899921536446, 'timing/model': 0.39460166498611216, 'learning_rate/qwen_vl_interface': 5.413689249153367e-07, 'learning_rate/action_model': 5.413689249153366e-06, 'epoch': 0.35}) +[2026-10-03 13:59:28,015][starVLA.training.train_starvla][INFO] - Step 4795, Loss: {'train/loss': 4.5640401367563754e-05, 'train/grad_norm_pre_clip': 0.0377541184425354, 'timing/data': 0.00020588999905157834, 'timing/model': 0.3928043549967697, 'learning_rate/qwen_vl_interface': 5.409688366803931e-07, 'learning_rate/action_model': 5.409688366803931e-06, 'epoch': 0.35}) +[2026-10-03 13:59:28,716][starVLA.training.train_starvla][INFO] - Step 4796, Loss: {'train/loss': 4.716186595032923e-06, 'train/grad_norm_pre_clip': 0.0011404149699956179, 'timing/data': 0.00025831999664660543, 'timing/model': 0.3918512259988347, 'learning_rate/qwen_vl_interface': 5.405706841502853e-07, 'learning_rate/action_model': 5.4057068415028525e-06, 'epoch': 0.35}) +[2026-10-03 13:59:29,419][starVLA.training.train_starvla][INFO] - Step 4797, Loss: {'train/loss': 5.468641575134825e-06, 'train/grad_norm_pre_clip': 0.002679741708561778, 'timing/data': 0.00020543999562505633, 'timing/model': 0.39320324499567505, 'learning_rate/qwen_vl_interface': 5.401744674886781e-07, 'learning_rate/action_model': 5.40174467488678e-06, 'epoch': 0.35}) +[2026-10-03 13:59:30,313][starVLA.training.train_starvla][INFO] - Step 4798, Loss: {'train/loss': 3.5140274121658877e-05, 'train/grad_norm_pre_clip': 0.012395557016134262, 'timing/data': 0.00020061999384779483, 'timing/model': 0.5847055080084829, 'learning_rate/qwen_vl_interface': 5.397801868584418e-07, 'learning_rate/action_model': 5.397801868584417e-06, 'epoch': 0.35}) +[2026-10-03 13:59:31,017][starVLA.training.train_starvla][INFO] - Step 4799, Loss: {'train/loss': 0.46436089277267456, 'train/grad_norm_pre_clip': 42.14451217651367, 'timing/data': 0.0003333890053909272, 'timing/model': 0.39297719599562697, 'learning_rate/qwen_vl_interface': 5.393878424216502e-07, 'learning_rate/action_model': 5.393878424216501e-06, 'epoch': 0.35}) +[2026-10-03 13:59:31,721][starVLA.training.train_starvla][INFO] - Step 4800, Loss: {'train/loss': 0.0010854867286980152, 'train/grad_norm_pre_clip': 2.170283555984497, 'timing/data': 0.0002883600100176409, 'timing/model': 0.3932547950098524, 'learning_rate/qwen_vl_interface': 5.389974343395813e-07, 'learning_rate/action_model': 5.389974343395812e-06, 'epoch': 0.35}) +[2026-10-03 13:59:32,425][starVLA.training.train_starvla][INFO] - Step 4801, Loss: {'train/loss': 5.410854282672517e-05, 'train/grad_norm_pre_clip': 0.030611099675297737, 'timing/data': 0.00024546000349801034, 'timing/model': 0.39403785500326194, 'learning_rate/qwen_vl_interface': 5.38608962772717e-07, 'learning_rate/action_model': 5.38608962772717e-06, 'epoch': 0.35}) +[2026-10-03 13:59:33,128][starVLA.training.train_starvla][INFO] - Step 4802, Loss: {'train/loss': 8.09851371741388e-06, 'train/grad_norm_pre_clip': 0.004218302201479673, 'timing/data': 0.0002050299954134971, 'timing/model': 0.3925578449998284, 'learning_rate/qwen_vl_interface': 5.382224278807429e-07, 'learning_rate/action_model': 5.382224278807429e-06, 'epoch': 0.35}) +[2026-10-03 13:59:33,830][starVLA.training.train_starvla][INFO] - Step 4803, Loss: {'train/loss': 1.5987392544047907e-05, 'train/grad_norm_pre_clip': 0.005319433286786079, 'timing/data': 0.00019727999460883439, 'timing/model': 0.39289757500228006, 'learning_rate/qwen_vl_interface': 5.378378298225496e-07, 'learning_rate/action_model': 5.378378298225496e-06, 'epoch': 0.35}) +[2026-10-03 13:59:34,726][starVLA.training.train_starvla][INFO] - Step 4804, Loss: {'train/loss': 1.5497182630497264e-06, 'train/grad_norm_pre_clip': 0.00037027630605734885, 'timing/data': 0.00025400899176020175, 'timing/model': 0.3923908559954725, 'learning_rate/qwen_vl_interface': 5.374551687562306e-07, 'learning_rate/action_model': 5.374551687562306e-06, 'epoch': 0.35}) +[2026-10-03 13:59:35,430][starVLA.training.train_starvla][INFO] - Step 4805, Loss: {'train/loss': 0.00046704651322215796, 'train/grad_norm_pre_clip': 0.627692461013794, 'timing/data': 0.0002738900075200945, 'timing/model': 0.3937054849957349, 'learning_rate/qwen_vl_interface': 5.37074444839083e-07, 'learning_rate/action_model': 5.370744448390829e-06, 'epoch': 0.35}) +[2026-10-03 13:59:36,132][starVLA.training.train_starvla][INFO] - Step 4806, Loss: {'train/loss': 1.5719968359917402e-05, 'train/grad_norm_pre_clip': 0.0034668331500142813, 'timing/data': 0.00019962999795097858, 'timing/model': 0.39285327499965206, 'learning_rate/qwen_vl_interface': 5.366956582276085e-07, 'learning_rate/action_model': 5.366956582276084e-06, 'epoch': 0.35}) +[2026-10-03 13:59:36,834][starVLA.training.train_starvla][INFO] - Step 4807, Loss: {'train/loss': 1.1279757927695755e-05, 'train/grad_norm_pre_clip': 0.004017379134893417, 'timing/data': 0.0001936999906320125, 'timing/model': 0.3925564149976708, 'learning_rate/qwen_vl_interface': 5.363188090775122e-07, 'learning_rate/action_model': 5.363188090775122e-06, 'epoch': 0.35}) +[2026-10-03 13:59:37,537][starVLA.training.train_starvla][INFO] - Step 4808, Loss: {'train/loss': 8.828703357721679e-06, 'train/grad_norm_pre_clip': 0.005675899796187878, 'timing/data': 0.00025537000328768045, 'timing/model': 0.39322743499360513, 'learning_rate/qwen_vl_interface': 5.359438975437018e-07, 'learning_rate/action_model': 5.359438975437017e-06, 'epoch': 0.35}) +[2026-10-03 13:59:38,432][starVLA.training.train_starvla][INFO] - Step 4809, Loss: {'train/loss': 1.2300268281251192e-05, 'train/grad_norm_pre_clip': 0.005601043812930584, 'timing/data': 0.00024206000671256334, 'timing/model': 0.5854139090079116, 'learning_rate/qwen_vl_interface': 5.355709237802895e-07, 'learning_rate/action_model': 5.3557092378028955e-06, 'epoch': 0.35}) +[2026-10-03 13:59:39,135][starVLA.training.train_starvla][INFO] - Step 4810, Loss: {'train/loss': 1.827539199439343e-05, 'train/grad_norm_pre_clip': 0.006687117740511894, 'timing/data': 0.00024340000527445227, 'timing/model': 0.39332801499404013, 'learning_rate/qwen_vl_interface': 5.351998879405911e-07, 'learning_rate/action_model': 5.351998879405911e-06, 'epoch': 0.35}) +[2026-10-03 13:59:39,838][starVLA.training.train_starvla][INFO] - Step 4811, Loss: {'train/loss': 1.8192647985415533e-05, 'train/grad_norm_pre_clip': 0.004125649109482765, 'timing/data': 0.0003139000036753714, 'timing/model': 0.392440085997805, 'learning_rate/qwen_vl_interface': 5.348307901771256e-07, 'learning_rate/action_model': 5.348307901771256e-06, 'epoch': 0.35}) +[2026-10-03 13:59:40,540][starVLA.training.train_starvla][INFO] - Step 4812, Loss: {'train/loss': 2.8833583201048896e-06, 'train/grad_norm_pre_clip': 0.0009118564776144922, 'timing/data': 0.0002836200001183897, 'timing/model': 0.39266636600950733, 'learning_rate/qwen_vl_interface': 5.344636306416147e-07, 'learning_rate/action_model': 5.344636306416146e-06, 'epoch': 0.35}) +[2026-10-03 13:59:41,243][starVLA.training.train_starvla][INFO] - Step 4813, Loss: {'train/loss': 3.205717803211883e-05, 'train/grad_norm_pre_clip': 0.017159856855869293, 'timing/data': 0.00020132001372985542, 'timing/model': 0.3935079049988417, 'learning_rate/qwen_vl_interface': 5.340984094849836e-07, 'learning_rate/action_model': 5.340984094849836e-06, 'epoch': 0.35}) +[2026-10-03 13:59:41,946][starVLA.training.train_starvla][INFO] - Step 4814, Loss: {'train/loss': 0.1528477519750595, 'train/grad_norm_pre_clip': 154.42315673828125, 'timing/data': 0.0002391800080658868, 'timing/model': 0.3926425859972369, 'learning_rate/qwen_vl_interface': 5.337351268573606e-07, 'learning_rate/action_model': 5.337351268573606e-06, 'epoch': 0.35}) +[2026-10-03 13:59:42,838][starVLA.training.train_starvla][INFO] - Step 4815, Loss: {'train/loss': 4.819430978386663e-05, 'train/grad_norm_pre_clip': 0.03595464304089546, 'timing/data': 0.000335279997671023, 'timing/model': 0.3930037450045347, 'learning_rate/qwen_vl_interface': 5.33373782908079e-07, 'learning_rate/action_model': 5.333737829080789e-06, 'epoch': 0.35}) +[2026-10-03 13:59:43,540][starVLA.training.train_starvla][INFO] - Step 4816, Loss: {'train/loss': 3.6880192055832595e-06, 'train/grad_norm_pre_clip': 0.000951379828620702, 'timing/data': 0.00019970000721514225, 'timing/model': 0.3926122350094374, 'learning_rate/qwen_vl_interface': 5.330143777856718e-07, 'learning_rate/action_model': 5.330143777856718e-06, 'epoch': 0.35}) +[2026-10-03 13:59:44,245][starVLA.training.train_starvla][INFO] - Step 4817, Loss: {'train/loss': 3.354642103658989e-05, 'train/grad_norm_pre_clip': 0.020394202321767807, 'timing/data': 0.00019886999507434666, 'timing/model': 0.3937690449965885, 'learning_rate/qwen_vl_interface': 5.326569116378786e-07, 'learning_rate/action_model': 5.326569116378786e-06, 'epoch': 0.35}) +[2026-10-03 13:59:44,947][starVLA.training.train_starvla][INFO] - Step 4818, Loss: {'train/loss': 1.185959815979004, 'train/grad_norm_pre_clip': 78.83549499511719, 'timing/data': 0.0003118799941148609, 'timing/model': 0.39195944600214716, 'learning_rate/qwen_vl_interface': 5.323013846116387e-07, 'learning_rate/action_model': 5.323013846116386e-06, 'epoch': 0.35}) +[2026-10-03 13:59:45,649][starVLA.training.train_starvla][INFO] - Step 4819, Loss: {'train/loss': 3.7625043205480324e-06, 'train/grad_norm_pre_clip': 0.0010392442345619202, 'timing/data': 0.0002411699970252812, 'timing/model': 0.3921771349996561, 'learning_rate/qwen_vl_interface': 5.31947796853097e-07, 'learning_rate/action_model': 5.3194779685309696e-06, 'epoch': 0.35}) +[2026-10-03 13:59:46,544][starVLA.training.train_starvla][INFO] - Step 4820, Loss: {'train/loss': 0.00022584265389014035, 'train/grad_norm_pre_clip': 0.2627371549606323, 'timing/data': 0.0003409990022191778, 'timing/model': 0.3926396260067122, 'learning_rate/qwen_vl_interface': 5.315961485075992e-07, 'learning_rate/action_model': 5.315961485075993e-06, 'epoch': 0.35}) +[2026-10-03 13:59:47,248][starVLA.training.train_starvla][INFO] - Step 4821, Loss: {'train/loss': 1.134644935518736e-05, 'train/grad_norm_pre_clip': 0.002701835473999381, 'timing/data': 0.00031461000617127866, 'timing/model': 0.3936663550121011, 'learning_rate/qwen_vl_interface': 5.312464397196958e-07, 'learning_rate/action_model': 5.312464397196958e-06, 'epoch': 0.35}) +[2026-10-03 13:59:47,950][starVLA.training.train_starvla][INFO] - Step 4822, Loss: {'train/loss': 2.048893293249421e-06, 'train/grad_norm_pre_clip': 0.0007451014826074243, 'timing/data': 0.00019336900732014328, 'timing/model': 0.39244376600254327, 'learning_rate/qwen_vl_interface': 5.308986706331377e-07, 'learning_rate/action_model': 5.308986706331376e-06, 'epoch': 0.35}) +[2026-10-03 13:59:48,652][starVLA.training.train_starvla][INFO] - Step 4823, Loss: {'train/loss': 3.704652772285044e-05, 'train/grad_norm_pre_clip': 0.028038227930665016, 'timing/data': 0.00019378999422769994, 'timing/model': 0.39244714500091504, 'learning_rate/qwen_vl_interface': 5.305528413908797e-07, 'learning_rate/action_model': 5.305528413908798e-06, 'epoch': 0.35}) +[2026-10-03 13:59:49,354][starVLA.training.train_starvla][INFO] - Step 4824, Loss: {'train/loss': 6.325157301034778e-05, 'train/grad_norm_pre_clip': 0.05613916739821434, 'timing/data': 0.00019116001203656197, 'timing/model': 0.39240242600499187, 'learning_rate/qwen_vl_interface': 5.302089521350799e-07, 'learning_rate/action_model': 5.3020895213507996e-06, 'epoch': 0.35}) +[2026-10-03 13:59:50,251][starVLA.training.train_starvla][INFO] - Step 4825, Loss: {'train/loss': 2.7920588763663545e-05, 'train/grad_norm_pre_clip': 0.01720953732728958, 'timing/data': 0.00019992899615317583, 'timing/model': 0.5868468289991142, 'learning_rate/qwen_vl_interface': 5.298670030070977e-07, 'learning_rate/action_model': 5.298670030070976e-06, 'epoch': 0.35}) +[2026-10-03 13:59:50,954][starVLA.training.train_starvla][INFO] - Step 4826, Loss: {'train/loss': 2.1984533304930665e-05, 'train/grad_norm_pre_clip': 0.005465863738209009, 'timing/data': 0.0002453699999023229, 'timing/model': 0.3921098959981464, 'learning_rate/qwen_vl_interface': 5.295269941474957e-07, 'learning_rate/action_model': 5.295269941474957e-06, 'epoch': 0.35}) +[2026-10-03 13:59:51,656][starVLA.training.train_starvla][INFO] - Step 4827, Loss: {'train/loss': 9.705140109872445e-05, 'train/grad_norm_pre_clip': 0.08723005652427673, 'timing/data': 0.00033132999669760466, 'timing/model': 0.3924256559985224, 'learning_rate/qwen_vl_interface': 5.291889256960384e-07, 'learning_rate/action_model': 5.291889256960384e-06, 'epoch': 0.35}) +[2026-10-03 13:59:52,360][starVLA.training.train_starvla][INFO] - Step 4828, Loss: {'train/loss': 0.001038023503497243, 'train/grad_norm_pre_clip': 7.280486106872559, 'timing/data': 0.000202519993763417, 'timing/model': 0.39333600499958266, 'learning_rate/qwen_vl_interface': 5.288527977916929e-07, 'learning_rate/action_model': 5.288527977916929e-06, 'epoch': 0.35}) +[2026-10-03 13:59:53,063][starVLA.training.train_starvla][INFO] - Step 4829, Loss: {'train/loss': 4.887542672804557e-06, 'train/grad_norm_pre_clip': 0.0012843288714066148, 'timing/data': 0.00023828000121284276, 'timing/model': 0.39383953499782365, 'learning_rate/qwen_vl_interface': 5.28518610572629e-07, 'learning_rate/action_model': 5.28518610572629e-06, 'epoch': 0.35}) +[2026-10-03 13:59:53,765][starVLA.training.train_starvla][INFO] - Step 4830, Loss: {'train/loss': 0.00022002114565111697, 'train/grad_norm_pre_clip': 0.3494579493999481, 'timing/data': 0.00021540001034736633, 'timing/model': 0.39173201499215793, 'learning_rate/qwen_vl_interface': 5.281863641762184e-07, 'learning_rate/action_model': 5.281863641762184e-06, 'epoch': 0.35}) +[2026-10-03 13:59:54,655][starVLA.training.train_starvla][INFO] - Step 4831, Loss: {'train/loss': 9.536482139083091e-06, 'train/grad_norm_pre_clip': 0.002060645492747426, 'timing/data': 0.0003103100025327876, 'timing/model': 0.39238088599813636, 'learning_rate/qwen_vl_interface': 5.278560587390355e-07, 'learning_rate/action_model': 5.278560587390355e-06, 'epoch': 0.35}) +[2026-10-03 13:59:55,356][starVLA.training.train_starvla][INFO] - Step 4832, Loss: {'train/loss': 6.483840115834028e-05, 'train/grad_norm_pre_clip': 0.027743617072701454, 'timing/data': 0.00019517999317031354, 'timing/model': 0.392060455007595, 'learning_rate/qwen_vl_interface': 5.275276943968557e-07, 'learning_rate/action_model': 5.275276943968557e-06, 'epoch': 0.35}) +[2026-10-03 13:59:56,058][starVLA.training.train_starvla][INFO] - Step 4833, Loss: {'train/loss': 2.555541414039908e-06, 'train/grad_norm_pre_clip': 0.0008037196821533144, 'timing/data': 0.0001986800052691251, 'timing/model': 0.3930645749933319, 'learning_rate/qwen_vl_interface': 5.272012712846566e-07, 'learning_rate/action_model': 5.2720127128465654e-06, 'epoch': 0.35}) +[2026-10-03 13:59:56,761][starVLA.training.train_starvla][INFO] - Step 4834, Loss: {'train/loss': 8.644202898722142e-05, 'train/grad_norm_pre_clip': 0.04143587499856949, 'timing/data': 0.00024090000079013407, 'timing/model': 0.3925800050055841, 'learning_rate/qwen_vl_interface': 5.268767895366198e-07, 'learning_rate/action_model': 5.268767895366198e-06, 'epoch': 0.35}) +[2026-10-03 13:59:57,463][starVLA.training.train_starvla][INFO] - Step 4835, Loss: {'train/loss': 0.00014587918121833354, 'train/grad_norm_pre_clip': 0.1965063363313675, 'timing/data': 0.00024597998708486557, 'timing/model': 0.3919013550039381, 'learning_rate/qwen_vl_interface': 5.265542492861268e-07, 'learning_rate/action_model': 5.2655424928612676e-06, 'epoch': 0.35}) +[2026-10-03 13:59:58,359][starVLA.training.train_starvla][INFO] - Step 4836, Loss: {'train/loss': 7.033283964119619e-06, 'train/grad_norm_pre_clip': 0.0019180303206667304, 'timing/data': 0.00028203999681863934, 'timing/model': 0.3917217760026688, 'learning_rate/qwen_vl_interface': 5.262336506657617e-07, 'learning_rate/action_model': 5.2623365066576165e-06, 'epoch': 0.35}) +[2026-10-03 13:59:59,062][starVLA.training.train_starvla][INFO] - Step 4837, Loss: {'train/loss': 4.44695760961622e-05, 'train/grad_norm_pre_clip': 0.01836358569562435, 'timing/data': 0.0003430299984756857, 'timing/model': 0.3934820759895956, 'learning_rate/qwen_vl_interface': 5.259149938073103e-07, 'learning_rate/action_model': 5.259149938073102e-06, 'epoch': 0.35}) +[2026-10-03 13:59:59,763][starVLA.training.train_starvla][INFO] - Step 4838, Loss: {'train/loss': 2.536517786211334e-05, 'train/grad_norm_pre_clip': 0.009361617267131805, 'timing/data': 0.0002620900049805641, 'timing/model': 0.39146525599062443, 'learning_rate/qwen_vl_interface': 5.255982788417609e-07, 'learning_rate/action_model': 5.255982788417608e-06, 'epoch': 0.35}) +[2026-10-03 14:00:00,464][starVLA.training.train_starvla][INFO] - Step 4839, Loss: {'train/loss': 6.534065505547915e-06, 'train/grad_norm_pre_clip': 0.0019903373904526234, 'timing/data': 0.00027639999461825937, 'timing/model': 0.39197130598768126, 'learning_rate/qwen_vl_interface': 5.252835058993029e-07, 'learning_rate/action_model': 5.252835058993029e-06, 'epoch': 0.35}) +[2026-10-03 14:00:01,166][starVLA.training.train_starvla][INFO] - Step 4840, Loss: {'train/loss': 0.0005033489433117211, 'train/grad_norm_pre_clip': 0.9935675859451294, 'timing/data': 0.00029984000138938427, 'timing/model': 0.3921934450045228, 'learning_rate/qwen_vl_interface': 5.249706751093274e-07, 'learning_rate/action_model': 5.249706751093273e-06, 'epoch': 0.35}) +[2026-10-03 14:00:02,060][starVLA.training.train_starvla][INFO] - Step 4841, Loss: {'train/loss': 1.9883489585481584e-05, 'train/grad_norm_pre_clip': 0.008307361043989658, 'timing/data': 0.00021985000057611614, 'timing/model': 0.3929513750044862, 'learning_rate/qwen_vl_interface': 5.246597866004273e-07, 'learning_rate/action_model': 5.246597866004273e-06, 'epoch': 0.35}) +[2026-10-03 14:00:02,762][starVLA.training.train_starvla][INFO] - Step 4842, Loss: {'train/loss': 2.6550142138148658e-05, 'train/grad_norm_pre_clip': 0.007589387707412243, 'timing/data': 0.00029772000561933964, 'timing/model': 0.3916689860052429, 'learning_rate/qwen_vl_interface': 5.243508405003975e-07, 'learning_rate/action_model': 5.243508405003975e-06, 'epoch': 0.35}) +[2026-10-03 14:00:03,462][starVLA.training.train_starvla][INFO] - Step 4843, Loss: {'train/loss': 0.0005222191684879363, 'train/grad_norm_pre_clip': 0.7042592763900757, 'timing/data': 0.0002401999954599887, 'timing/model': 0.3912641560018528, 'learning_rate/qwen_vl_interface': 5.240438369362336e-07, 'learning_rate/action_model': 5.2404383693623366e-06, 'epoch': 0.35}) +[2026-10-03 14:00:04,162][starVLA.training.train_starvla][INFO] - Step 4844, Loss: {'train/loss': 0.00012065562623320147, 'train/grad_norm_pre_clip': 0.06012386083602905, 'timing/data': 0.0003109300014330074, 'timing/model': 0.3917378049955005, 'learning_rate/qwen_vl_interface': 5.237387760341335e-07, 'learning_rate/action_model': 5.237387760341335e-06, 'epoch': 0.35}) +[2026-10-03 14:00:04,864][starVLA.training.train_starvla][INFO] - Step 4845, Loss: {'train/loss': 3.380485577508807e-05, 'train/grad_norm_pre_clip': 0.011105993762612343, 'timing/data': 0.0002806900010909885, 'timing/model': 0.3930749950086465, 'learning_rate/qwen_vl_interface': 5.234356579194963e-07, 'learning_rate/action_model': 5.234356579194963e-06, 'epoch': 0.35}) +[2026-10-03 14:00:05,755][starVLA.training.train_starvla][INFO] - Step 4846, Loss: {'train/loss': 2.926768320321571e-05, 'train/grad_norm_pre_clip': 0.012681401334702969, 'timing/data': 0.00019861999317072332, 'timing/model': 0.5815641090011923, 'learning_rate/qwen_vl_interface': 5.231344827169218e-07, 'learning_rate/action_model': 5.231344827169217e-06, 'epoch': 0.35}) +[2026-10-03 14:00:06,456][starVLA.training.train_starvla][INFO] - Step 4847, Loss: {'train/loss': 4.984363840776496e-06, 'train/grad_norm_pre_clip': 0.0026984792202711105, 'timing/data': 0.00022057000023778528, 'timing/model': 0.39233956500538625, 'learning_rate/qwen_vl_interface': 5.228352505502119e-07, 'learning_rate/action_model': 5.228352505502119e-06, 'epoch': 0.35}) +[2026-10-03 14:00:07,158][starVLA.training.train_starvla][INFO] - Step 4848, Loss: {'train/loss': 2.4970813683466986e-05, 'train/grad_norm_pre_clip': 0.009510783478617668, 'timing/data': 0.0003027299972018227, 'timing/model': 0.3931769949995214, 'learning_rate/qwen_vl_interface': 5.225379615423701e-07, 'learning_rate/action_model': 5.2253796154237e-06, 'epoch': 0.35}) +[2026-10-03 14:00:07,861][starVLA.training.train_starvla][INFO] - Step 4849, Loss: {'train/loss': 7.5993725658918265e-06, 'train/grad_norm_pre_clip': 0.0026641071308404207, 'timing/data': 0.00027465001039672643, 'timing/model': 0.39317135501187295, 'learning_rate/qwen_vl_interface': 5.222426158156008e-07, 'learning_rate/action_model': 5.222426158156008e-06, 'epoch': 0.35}) +[2026-10-03 14:00:08,562][starVLA.training.train_starvla][INFO] - Step 4850, Loss: {'train/loss': 4.3606974941212684e-05, 'train/grad_norm_pre_clip': 0.018299341201782227, 'timing/data': 0.00023774000874254853, 'timing/model': 0.39158537599723786, 'learning_rate/qwen_vl_interface': 5.219492134913084e-07, 'learning_rate/action_model': 5.219492134913084e-06, 'epoch': 0.35}) +[2026-10-03 14:00:09,456][starVLA.training.train_starvla][INFO] - Step 4851, Loss: {'train/loss': 2.849684824468568e-05, 'train/grad_norm_pre_clip': 0.011265135370194912, 'timing/data': 0.00019627998699434102, 'timing/model': 0.5843499089969555, 'learning_rate/qwen_vl_interface': 5.216577546901008e-07, 'learning_rate/action_model': 5.216577546901007e-06, 'epoch': 0.36}) +[2026-10-03 14:00:10,158][starVLA.training.train_starvla][INFO] - Step 4852, Loss: {'train/loss': 5.550598416448338e-06, 'train/grad_norm_pre_clip': 0.0017906976863741875, 'timing/data': 0.00020652000966947526, 'timing/model': 0.3923962660046527, 'learning_rate/qwen_vl_interface': 5.213682395317852e-07, 'learning_rate/action_model': 5.213682395317851e-06, 'epoch': 0.36}) +[2026-10-03 14:00:10,861][starVLA.training.train_starvla][INFO] - Step 4853, Loss: {'train/loss': 1.6673147911205888e-05, 'train/grad_norm_pre_clip': 0.008744023740291595, 'timing/data': 0.00025299999106209725, 'timing/model': 0.39304410498880316, 'learning_rate/qwen_vl_interface': 5.2108066813537e-07, 'learning_rate/action_model': 5.2108066813537e-06, 'epoch': 0.36}) +[2026-10-03 14:00:11,562][starVLA.training.train_starvla][INFO] - Step 4854, Loss: {'train/loss': 6.213725555426208e-06, 'train/grad_norm_pre_clip': 0.0019466852536424994, 'timing/data': 0.00019260001135990024, 'timing/model': 0.3919988159905188, 'learning_rate/qwen_vl_interface': 5.207950406190654e-07, 'learning_rate/action_model': 5.207950406190653e-06, 'epoch': 0.36}) +[2026-10-03 14:00:12,263][starVLA.training.train_starvla][INFO] - Step 4855, Loss: {'train/loss': 0.0001074585088645108, 'train/grad_norm_pre_clip': 0.2058407962322235, 'timing/data': 0.00029942000401206315, 'timing/model': 0.3918687260011211, 'learning_rate/qwen_vl_interface': 5.205113571002818e-07, 'learning_rate/action_model': 5.205113571002818e-06, 'epoch': 0.36}) +[2026-10-03 14:00:12,965][starVLA.training.train_starvla][INFO] - Step 4856, Loss: {'train/loss': 0.00012038708518957719, 'train/grad_norm_pre_clip': 0.23072703182697296, 'timing/data': 0.00024079000286292285, 'timing/model': 0.39197292500466574, 'learning_rate/qwen_vl_interface': 5.202296176956308e-07, 'learning_rate/action_model': 5.202296176956308e-06, 'epoch': 0.36}) +[2026-10-03 14:00:13,861][starVLA.training.train_starvla][INFO] - Step 4857, Loss: {'train/loss': 1.2740472357108956e-06, 'train/grad_norm_pre_clip': 0.0003835257375612855, 'timing/data': 0.00031537999166175723, 'timing/model': 0.39323328499449417, 'learning_rate/qwen_vl_interface': 5.19949822520925e-07, 'learning_rate/action_model': 5.19949822520925e-06, 'epoch': 0.36}) +[2026-10-03 14:00:14,563][starVLA.training.train_starvla][INFO] - Step 4858, Loss: {'train/loss': 0.3585757315158844, 'train/grad_norm_pre_clip': 108.32550811767578, 'timing/data': 0.0004305500042391941, 'timing/model': 0.39181527600158006, 'learning_rate/qwen_vl_interface': 5.196719716911777e-07, 'learning_rate/action_model': 5.1967197169117764e-06, 'epoch': 0.36}) +[2026-10-03 14:00:15,264][starVLA.training.train_starvla][INFO] - Step 4859, Loss: {'train/loss': 0.2960803508758545, 'train/grad_norm_pre_clip': 103.8067855834961, 'timing/data': 0.00023438000062014908, 'timing/model': 0.3918548659858061, 'learning_rate/qwen_vl_interface': 5.193960653206029e-07, 'learning_rate/action_model': 5.193960653206029e-06, 'epoch': 0.36}) +[2026-10-03 14:00:15,966][starVLA.training.train_starvla][INFO] - Step 4860, Loss: {'train/loss': 9.34235577005893e-05, 'train/grad_norm_pre_clip': 0.025471309199929237, 'timing/data': 0.00024367999867536128, 'timing/model': 0.3929800360056106, 'learning_rate/qwen_vl_interface': 5.191221035226144e-07, 'learning_rate/action_model': 5.191221035226144e-06, 'epoch': 0.36}) +[2026-10-03 14:00:16,669][starVLA.training.train_starvla][INFO] - Step 4861, Loss: {'train/loss': 9.961097021005116e-06, 'train/grad_norm_pre_clip': 0.0052303471602499485, 'timing/data': 0.00031377999403048307, 'timing/model': 0.3932201159914257, 'learning_rate/qwen_vl_interface': 5.188500864098292e-07, 'learning_rate/action_model': 5.1885008640982915e-06, 'epoch': 0.36}) +[2026-10-03 14:00:17,563][starVLA.training.train_starvla][INFO] - Step 4862, Loss: {'train/loss': 5.174789475859143e-05, 'train/grad_norm_pre_clip': 0.01905294880270958, 'timing/data': 0.00020383000082802027, 'timing/model': 0.5840629189915489, 'learning_rate/qwen_vl_interface': 5.185800140940613e-07, 'learning_rate/action_model': 5.185800140940613e-06, 'epoch': 0.36}) +[2026-10-03 14:00:18,266][starVLA.training.train_starvla][INFO] - Step 4863, Loss: {'train/loss': 3.1428491638507694e-05, 'train/grad_norm_pre_clip': 0.030939852818846703, 'timing/data': 0.0002749400009633973, 'timing/model': 0.3926496159983799, 'learning_rate/qwen_vl_interface': 5.183118866863288e-07, 'learning_rate/action_model': 5.1831188668632875e-06, 'epoch': 0.36}) +[2026-10-03 14:00:18,970][starVLA.training.train_starvla][INFO] - Step 4864, Loss: {'train/loss': 0.5538250207901001, 'train/grad_norm_pre_clip': 29.002704620361328, 'timing/data': 0.0003064400079892948, 'timing/model': 0.39313981600571424, 'learning_rate/qwen_vl_interface': 5.180457042968484e-07, 'learning_rate/action_model': 5.1804570429684835e-06, 'epoch': 0.36}) +[2026-10-03 14:00:19,674][starVLA.training.train_starvla][INFO] - Step 4865, Loss: {'train/loss': 1.421516662958311e-05, 'train/grad_norm_pre_clip': 0.005438632797449827, 'timing/data': 0.00025471000117249787, 'timing/model': 0.3941334650007775, 'learning_rate/qwen_vl_interface': 5.17781467035037e-07, 'learning_rate/action_model': 5.177814670350371e-06, 'epoch': 0.36}) +[2026-10-03 14:00:20,377][starVLA.training.train_starvla][INFO] - Step 4866, Loss: {'train/loss': 7.273694791365415e-05, 'train/grad_norm_pre_clip': 0.07567249238491058, 'timing/data': 0.00020883000979665667, 'timing/model': 0.3922690559993498, 'learning_rate/qwen_vl_interface': 5.175191750095133e-07, 'learning_rate/action_model': 5.175191750095132e-06, 'epoch': 0.36}) +[2026-10-03 14:00:21,079][starVLA.training.train_starvla][INFO] - Step 4867, Loss: {'train/loss': 0.00017821257642935961, 'train/grad_norm_pre_clip': 0.5249231457710266, 'timing/data': 0.0002408599975751713, 'timing/model': 0.392860464999103, 'learning_rate/qwen_vl_interface': 5.172588283280953e-07, 'learning_rate/action_model': 5.172588283280953e-06, 'epoch': 0.36}) +[2026-10-03 14:00:21,970][starVLA.training.train_starvla][INFO] - Step 4868, Loss: {'train/loss': 2.2059210095903836e-05, 'train/grad_norm_pre_clip': 0.0121331587433815, 'timing/data': 0.00031153000600170344, 'timing/model': 0.39297284599160776, 'learning_rate/qwen_vl_interface': 5.170004270978019e-07, 'learning_rate/action_model': 5.170004270978018e-06, 'epoch': 0.36}) +[2026-10-03 14:00:22,672][starVLA.training.train_starvla][INFO] - Step 4869, Loss: {'train/loss': 5.1402275857981294e-05, 'train/grad_norm_pre_clip': 0.0151231549680233, 'timing/data': 0.0003042799944523722, 'timing/model': 0.39295454499369953, 'learning_rate/qwen_vl_interface': 5.167439714248512e-07, 'learning_rate/action_model': 5.1674397142485116e-06, 'epoch': 0.36}) +[2026-10-03 14:00:23,374][starVLA.training.train_starvla][INFO] - Step 4870, Loss: {'train/loss': 4.149275628151372e-05, 'train/grad_norm_pre_clip': 0.014715829864144325, 'timing/data': 0.0003211599978385493, 'timing/model': 0.39136825600871816, 'learning_rate/qwen_vl_interface': 5.164894614146633e-07, 'learning_rate/action_model': 5.164894614146634e-06, 'epoch': 0.36}) +[2026-10-03 14:00:24,076][starVLA.training.train_starvla][INFO] - Step 4871, Loss: {'train/loss': 2.2573754904442467e-05, 'train/grad_norm_pre_clip': 0.004784686956554651, 'timing/data': 0.0002445599966449663, 'timing/model': 0.3928582360094879, 'learning_rate/qwen_vl_interface': 5.16236897171858e-07, 'learning_rate/action_model': 5.16236897171858e-06, 'epoch': 0.36}) +[2026-10-03 14:00:24,778][starVLA.training.train_starvla][INFO] - Step 4872, Loss: {'train/loss': 4.682581129600294e-05, 'train/grad_norm_pre_clip': 0.01459544338285923, 'timing/data': 0.0001984799891943112, 'timing/model': 0.39243343500129413, 'learning_rate/qwen_vl_interface': 5.15986278800254e-07, 'learning_rate/action_model': 5.15986278800254e-06, 'epoch': 0.36}) +[2026-10-03 14:00:25,673][starVLA.training.train_starvla][INFO] - Step 4873, Loss: {'train/loss': 2.4568209482822567e-05, 'train/grad_norm_pre_clip': 0.02294563688337803, 'timing/data': 0.00027898000553250313, 'timing/model': 0.3932432349974988, 'learning_rate/qwen_vl_interface': 5.157376064028719e-07, 'learning_rate/action_model': 5.157376064028718e-06, 'epoch': 0.36}) +[2026-10-03 14:00:26,375][starVLA.training.train_starvla][INFO] - Step 4874, Loss: {'train/loss': 2.496505476301536e-05, 'train/grad_norm_pre_clip': 0.005676808767020702, 'timing/data': 0.00027134000265505165, 'timing/model': 0.3927686750103021, 'learning_rate/qwen_vl_interface': 5.154908800819307e-07, 'learning_rate/action_model': 5.154908800819306e-06, 'epoch': 0.36}) +[2026-10-03 14:00:27,077][starVLA.training.train_starvla][INFO] - Step 4875, Loss: {'train/loss': 2.499367474229075e-05, 'train/grad_norm_pre_clip': 0.09649597108364105, 'timing/data': 0.00024873000802472234, 'timing/model': 0.39214097599324305, 'learning_rate/qwen_vl_interface': 5.152460999388504e-07, 'learning_rate/action_model': 5.152460999388503e-06, 'epoch': 0.36}) +[2026-10-03 14:00:27,778][starVLA.training.train_starvla][INFO] - Step 4876, Loss: {'train/loss': 5.349432740331395e-06, 'train/grad_norm_pre_clip': 0.002676531206816435, 'timing/data': 0.00029085000278428197, 'timing/model': 0.39230051600316074, 'learning_rate/qwen_vl_interface': 5.150032660742517e-07, 'learning_rate/action_model': 5.150032660742517e-06, 'epoch': 0.36}) +[2026-10-03 14:00:28,482][starVLA.training.train_starvla][INFO] - Step 4877, Loss: {'train/loss': 1.2449399946490303e-05, 'train/grad_norm_pre_clip': 0.0064389342442154884, 'timing/data': 0.0002793199964798987, 'timing/model': 0.39368384500266984, 'learning_rate/qwen_vl_interface': 5.147623785879535e-07, 'learning_rate/action_model': 5.147623785879534e-06, 'epoch': 0.36}) +[2026-10-03 14:00:29,371][starVLA.training.train_starvla][INFO] - Step 4878, Loss: {'train/loss': 2.4486362235620618e-05, 'train/grad_norm_pre_clip': 0.009001776576042175, 'timing/data': 0.0002473299973644316, 'timing/model': 0.579190930002369, 'learning_rate/qwen_vl_interface': 5.145234375789758e-07, 'learning_rate/action_model': 5.145234375789758e-06, 'epoch': 0.36}) +[2026-10-03 14:00:30,072][starVLA.training.train_starvla][INFO] - Step 4879, Loss: {'train/loss': 0.0049483515322208405, 'train/grad_norm_pre_clip': 19.849689483642578, 'timing/data': 0.0003877700073644519, 'timing/model': 0.392215165993548, 'learning_rate/qwen_vl_interface': 5.142864431455387e-07, 'learning_rate/action_model': 5.1428644314553875e-06, 'epoch': 0.36}) +[2026-10-03 14:00:30,775][starVLA.training.train_starvla][INFO] - Step 4880, Loss: {'train/loss': 3.852901863865554e-05, 'train/grad_norm_pre_clip': 0.026303730905056, 'timing/data': 0.0002568199997767806, 'timing/model': 0.39317297498928383, 'learning_rate/qwen_vl_interface': 5.140513953850612e-07, 'learning_rate/action_model': 5.140513953850612e-06, 'epoch': 0.36}) +[2026-10-03 14:00:31,478][starVLA.training.train_starvla][INFO] - Step 4881, Loss: {'train/loss': 4.462834112928249e-06, 'train/grad_norm_pre_clip': 0.0013571364106610417, 'timing/data': 0.00026768899988383055, 'timing/model': 0.3932468859857181, 'learning_rate/qwen_vl_interface': 5.138182943941623e-07, 'learning_rate/action_model': 5.138182943941623e-06, 'epoch': 0.36}) +[2026-10-03 14:00:32,180][starVLA.training.train_starvla][INFO] - Step 4882, Loss: {'train/loss': 3.813339208136313e-05, 'train/grad_norm_pre_clip': 0.029350660741329193, 'timing/data': 0.0002793299936456606, 'timing/model': 0.3917750559921842, 'learning_rate/qwen_vl_interface': 5.135871402686615e-07, 'learning_rate/action_model': 5.135871402686615e-06, 'epoch': 0.36}) +[2026-10-03 14:00:32,882][starVLA.training.train_starvla][INFO] - Step 4883, Loss: {'train/loss': 2.197914454882266e-06, 'train/grad_norm_pre_clip': 0.0004803525807801634, 'timing/data': 0.00038187998870853335, 'timing/model': 0.39241916600440163, 'learning_rate/qwen_vl_interface': 5.133579331035777e-07, 'learning_rate/action_model': 5.133579331035777e-06, 'epoch': 0.36}) +[2026-10-03 14:00:33,772][starVLA.training.train_starvla][INFO] - Step 4884, Loss: {'train/loss': 0.0003616169560700655, 'train/grad_norm_pre_clip': 0.5099679827690125, 'timing/data': 0.00026898999931290746, 'timing/model': 0.39250702600111254, 'learning_rate/qwen_vl_interface': 5.131306729931284e-07, 'learning_rate/action_model': 5.1313067299312836e-06, 'epoch': 0.36}) +[2026-10-03 14:00:34,475][starVLA.training.train_starvla][INFO] - Step 4885, Loss: {'train/loss': 0.00019556966435629874, 'train/grad_norm_pre_clip': 0.09884192049503326, 'timing/data': 0.00028573999588843435, 'timing/model': 0.39371753500017803, 'learning_rate/qwen_vl_interface': 5.129053600307326e-07, 'learning_rate/action_model': 5.129053600307326e-06, 'epoch': 0.36}) +[2026-10-03 14:00:35,178][starVLA.training.train_starvla][INFO] - Step 4886, Loss: {'train/loss': 4.19452553614974e-05, 'train/grad_norm_pre_clip': 0.015700915828347206, 'timing/data': 0.00028237000515218824, 'timing/model': 0.39193647498905193, 'learning_rate/qwen_vl_interface': 5.126819943090078e-07, 'learning_rate/action_model': 5.1268199430900775e-06, 'epoch': 0.36}) +[2026-10-03 14:00:35,880][starVLA.training.train_starvla][INFO] - Step 4887, Loss: {'train/loss': 0.00010341302549932152, 'train/grad_norm_pre_clip': 0.0722486674785614, 'timing/data': 0.0002932099887402728, 'timing/model': 0.3925423260079697, 'learning_rate/qwen_vl_interface': 5.124605759197706e-07, 'learning_rate/action_model': 5.124605759197706e-06, 'epoch': 0.36}) +[2026-10-03 14:00:36,583][starVLA.training.train_starvla][INFO] - Step 4888, Loss: {'train/loss': 0.00017905395361594856, 'train/grad_norm_pre_clip': 0.1852884441614151, 'timing/data': 0.0002014700003201142, 'timing/model': 0.39249156598816626, 'learning_rate/qwen_vl_interface': 5.122411049540383e-07, 'learning_rate/action_model': 5.1224110495403825e-06, 'epoch': 0.36}) +[2026-10-03 14:00:37,478][starVLA.training.train_starvla][INFO] - Step 4889, Loss: {'train/loss': 4.240373891661875e-05, 'train/grad_norm_pre_clip': 0.0410926416516304, 'timing/data': 0.0003097299922956154, 'timing/model': 0.393098275992088, 'learning_rate/qwen_vl_interface': 5.120235815020269e-07, 'learning_rate/action_model': 5.120235815020269e-06, 'epoch': 0.36}) +[2026-10-03 14:00:38,181][starVLA.training.train_starvla][INFO] - Step 4890, Loss: {'train/loss': 3.06558795273304e-05, 'train/grad_norm_pre_clip': 0.008852578699588776, 'timing/data': 0.0003103700000792742, 'timing/model': 0.3917819460039027, 'learning_rate/qwen_vl_interface': 5.118080056531524e-07, 'learning_rate/action_model': 5.118080056531524e-06, 'epoch': 0.36}) +[2026-10-03 14:00:38,882][starVLA.training.train_starvla][INFO] - Step 4891, Loss: {'train/loss': 0.00022625485144089907, 'train/grad_norm_pre_clip': 0.2068219631910324, 'timing/data': 0.0003021300071850419, 'timing/model': 0.39211892599996645, 'learning_rate/qwen_vl_interface': 5.115943774960295e-07, 'learning_rate/action_model': 5.115943774960295e-06, 'epoch': 0.36}) +[2026-10-03 14:00:39,583][starVLA.training.train_starvla][INFO] - Step 4892, Loss: {'train/loss': 2.375802068854682e-05, 'train/grad_norm_pre_clip': 0.018980663269758224, 'timing/data': 0.00027735000185202807, 'timing/model': 0.39177959600056056, 'learning_rate/qwen_vl_interface': 5.113826971184727e-07, 'learning_rate/action_model': 5.113826971184727e-06, 'epoch': 0.36}) +[2026-10-03 14:00:40,287][starVLA.training.train_starvla][INFO] - Step 4893, Loss: {'train/loss': 6.54145514999982e-06, 'train/grad_norm_pre_clip': 0.002178965602070093, 'timing/data': 0.00031245000718627125, 'timing/model': 0.3936863760027336, 'learning_rate/qwen_vl_interface': 5.111729646074954e-07, 'learning_rate/action_model': 5.111729646074954e-06, 'epoch': 0.36}) +[2026-10-03 14:00:41,183][starVLA.training.train_starvla][INFO] - Step 4894, Loss: {'train/loss': 9.429414785699919e-05, 'train/grad_norm_pre_clip': 0.05223975330591202, 'timing/data': 0.00036187999648973346, 'timing/model': 0.3921570060047088, 'learning_rate/qwen_vl_interface': 5.10965180049312e-07, 'learning_rate/action_model': 5.10965180049312e-06, 'epoch': 0.36}) +[2026-10-03 14:00:41,885][starVLA.training.train_starvla][INFO] - Step 4895, Loss: {'train/loss': 5.2227810556360055e-06, 'train/grad_norm_pre_clip': 0.0023099437821656466, 'timing/data': 0.0002564799942774698, 'timing/model': 0.3922751649952261, 'learning_rate/qwen_vl_interface': 5.107593435293336e-07, 'learning_rate/action_model': 5.107593435293336e-06, 'epoch': 0.36}) +[2026-10-03 14:00:42,586][starVLA.training.train_starvla][INFO] - Step 4896, Loss: {'train/loss': 0.12127205729484558, 'train/grad_norm_pre_clip': 127.322509765625, 'timing/data': 0.00018698000349104404, 'timing/model': 0.39246647601248696, 'learning_rate/qwen_vl_interface': 5.105554551321724e-07, 'learning_rate/action_model': 5.105554551321723e-06, 'epoch': 0.36}) +[2026-10-03 14:00:43,290][starVLA.training.train_starvla][INFO] - Step 4897, Loss: {'train/loss': 2.7194537324248813e-06, 'train/grad_norm_pre_clip': 0.0005793632590211928, 'timing/data': 0.0002461399999447167, 'timing/model': 0.3933021860138979, 'learning_rate/qwen_vl_interface': 5.103535149416389e-07, 'learning_rate/action_model': 5.1035351494163885e-06, 'epoch': 0.36}) +[2026-10-03 14:00:43,991][starVLA.training.train_starvla][INFO] - Step 4898, Loss: {'train/loss': 2.5560873837093823e-05, 'train/grad_norm_pre_clip': 0.030074836686253548, 'timing/data': 0.00020119998953305185, 'timing/model': 0.39229653601069003, 'learning_rate/qwen_vl_interface': 5.101535230407436e-07, 'learning_rate/action_model': 5.101535230407436e-06, 'epoch': 0.36}) +[2026-10-03 14:00:44,886][starVLA.training.train_starvla][INFO] - Step 4899, Loss: {'train/loss': 9.484401743975468e-06, 'train/grad_norm_pre_clip': 0.0029069145675748587, 'timing/data': 0.00027604000933934003, 'timing/model': 0.5841577989922371, 'learning_rate/qwen_vl_interface': 5.099554795116956e-07, 'learning_rate/action_model': 5.0995547951169565e-06, 'epoch': 0.36}) +[2026-10-03 14:00:45,586][starVLA.training.train_starvla][INFO] - Step 4900, Loss: {'train/loss': 0.3383404612541199, 'train/grad_norm_pre_clip': 67.60468292236328, 'timing/data': 0.0002826500131050125, 'timing/model': 0.3925550559943076, 'learning_rate/qwen_vl_interface': 5.097593844359024e-07, 'learning_rate/action_model': 5.0975938443590245e-06, 'epoch': 0.36}) +[2026-10-03 14:00:46,288][starVLA.training.train_starvla][INFO] - Step 4901, Loss: {'train/loss': 0.6061214804649353, 'train/grad_norm_pre_clip': 169.733154296875, 'timing/data': 0.00022588000865653157, 'timing/model': 0.3924407960003009, 'learning_rate/qwen_vl_interface': 5.095652378939718e-07, 'learning_rate/action_model': 5.095652378939717e-06, 'epoch': 0.36}) +[2026-10-03 14:00:46,990][starVLA.training.train_starvla][INFO] - Step 4902, Loss: {'train/loss': 5.418801447376609e-05, 'train/grad_norm_pre_clip': 0.02640954591333866, 'timing/data': 0.00034550001146271825, 'timing/model': 0.3924256860045716, 'learning_rate/qwen_vl_interface': 5.093730399657101e-07, 'learning_rate/action_model': 5.0937303996571e-06, 'epoch': 0.36}) +[2026-10-03 14:00:47,691][starVLA.training.train_starvla][INFO] - Step 4903, Loss: {'train/loss': 1.1614674804150127e-05, 'train/grad_norm_pre_clip': 0.003594221780076623, 'timing/data': 0.00027771999884862453, 'timing/model': 0.3918906259932555, 'learning_rate/qwen_vl_interface': 5.091827907301227e-07, 'learning_rate/action_model': 5.091827907301227e-06, 'epoch': 0.36}) +[2026-10-03 14:00:48,582][starVLA.training.train_starvla][INFO] - Step 4904, Loss: {'train/loss': 1.0035294508270454e-05, 'train/grad_norm_pre_clip': 0.005026533734053373, 'timing/data': 0.00029768999957013875, 'timing/model': 0.5810989300080109, 'learning_rate/qwen_vl_interface': 5.089944902654134e-07, 'learning_rate/action_model': 5.089944902654134e-06, 'epoch': 0.36}) +[2026-10-03 14:00:49,286][starVLA.training.train_starvla][INFO] - Step 4905, Loss: {'train/loss': 2.9457536584232002e-05, 'train/grad_norm_pre_clip': 0.009795322082936764, 'timing/data': 0.00023883000540081412, 'timing/model': 0.3934741760022007, 'learning_rate/qwen_vl_interface': 5.088081386489856e-07, 'learning_rate/action_model': 5.088081386489857e-06, 'epoch': 0.36}) +[2026-10-03 14:00:49,986][starVLA.training.train_starvla][INFO] - Step 4906, Loss: {'train/loss': 4.408723907545209e-05, 'train/grad_norm_pre_clip': 0.01856612041592598, 'timing/data': 0.00023417000193148851, 'timing/model': 0.3903464260074543, 'learning_rate/qwen_vl_interface': 5.086237359574419e-07, 'learning_rate/action_model': 5.086237359574419e-06, 'epoch': 0.36}) +[2026-10-03 14:00:50,686][starVLA.training.train_starvla][INFO] - Step 4907, Loss: {'train/loss': 0.0023834765888750553, 'train/grad_norm_pre_clip': 11.943533897399902, 'timing/data': 0.0002327699912711978, 'timing/model': 0.3919524859957164, 'learning_rate/qwen_vl_interface': 5.084412822665825e-07, 'learning_rate/action_model': 5.084412822665825e-06, 'epoch': 0.36}) +[2026-10-03 14:00:51,388][starVLA.training.train_starvla][INFO] - Step 4908, Loss: {'train/loss': 1.6002290067262948e-05, 'train/grad_norm_pre_clip': 0.006184526719152927, 'timing/data': 0.00027368000883143395, 'timing/model': 0.39210699700925034, 'learning_rate/qwen_vl_interface': 5.082607776514081e-07, 'learning_rate/action_model': 5.082607776514081e-06, 'epoch': 0.36}) +[2026-10-03 14:00:52,091][starVLA.training.train_starvla][INFO] - Step 4909, Loss: {'train/loss': 0.017570126801729202, 'train/grad_norm_pre_clip': 76.05020141601562, 'timing/data': 0.0003190300049027428, 'timing/model': 0.39362521599105094, 'learning_rate/qwen_vl_interface': 5.080822221861167e-07, 'learning_rate/action_model': 5.080822221861166e-06, 'epoch': 0.36}) +[2026-10-03 14:00:52,985][starVLA.training.train_starvla][INFO] - Step 4910, Loss: {'train/loss': 0.008051695302128792, 'train/grad_norm_pre_clip': 42.625999450683594, 'timing/data': 0.0002101400023093447, 'timing/model': 0.3922523860092042, 'learning_rate/qwen_vl_interface': 5.07905615944106e-07, 'learning_rate/action_model': 5.0790561594410595e-06, 'epoch': 0.36}) +[2026-10-03 14:00:53,687][starVLA.training.train_starvla][INFO] - Step 4911, Loss: {'train/loss': 3.792299594351789e-06, 'train/grad_norm_pre_clip': 0.0018619068432599306, 'timing/data': 0.0002830300072673708, 'timing/model': 0.3923787459934829, 'learning_rate/qwen_vl_interface': 5.077309589979716e-07, 'learning_rate/action_model': 5.077309589979715e-06, 'epoch': 0.36}) +[2026-10-03 14:00:54,389][starVLA.training.train_starvla][INFO] - Step 4912, Loss: {'train/loss': 0.0002656121796462685, 'train/grad_norm_pre_clip': 0.18579792976379395, 'timing/data': 0.0003100400062976405, 'timing/model': 0.39235795600689016, 'learning_rate/qwen_vl_interface': 5.075582514195089e-07, 'learning_rate/action_model': 5.0755825141950895e-06, 'epoch': 0.36}) +[2026-10-03 14:00:55,091][starVLA.training.train_starvla][INFO] - Step 4913, Loss: {'train/loss': 8.366743713850155e-06, 'train/grad_norm_pre_clip': 0.0021009647753089666, 'timing/data': 0.00024437000683974475, 'timing/model': 0.39359522600716446, 'learning_rate/qwen_vl_interface': 5.073874932797115e-07, 'learning_rate/action_model': 5.0738749327971155e-06, 'epoch': 0.36}) +[2026-10-03 14:00:55,793][starVLA.training.train_starvla][INFO] - Step 4914, Loss: {'train/loss': 0.0003663546231109649, 'train/grad_norm_pre_clip': 0.23868131637573242, 'timing/data': 0.00024249000125564635, 'timing/model': 0.39243566599907354, 'learning_rate/qwen_vl_interface': 5.072186846487713e-07, 'learning_rate/action_model': 5.072186846487713e-06, 'epoch': 0.36}) +[2026-10-03 14:00:56,683][starVLA.training.train_starvla][INFO] - Step 4915, Loss: {'train/loss': 3.7476188481377903e-06, 'train/grad_norm_pre_clip': 0.0008498787065036595, 'timing/data': 0.00030078999407123774, 'timing/model': 0.580380259998492, 'learning_rate/qwen_vl_interface': 5.070518255960796e-07, 'learning_rate/action_model': 5.070518255960795e-06, 'epoch': 0.36}) +[2026-10-03 14:00:57,385][starVLA.training.train_starvla][INFO] - Step 4916, Loss: {'train/loss': 1.2851693099946715e-05, 'train/grad_norm_pre_clip': 0.006996706128120422, 'timing/data': 0.0002014900092035532, 'timing/model': 0.3931664660049137, 'learning_rate/qwen_vl_interface': 5.068869161902256e-07, 'learning_rate/action_model': 5.068869161902256e-06, 'epoch': 0.36}) +[2026-10-03 14:00:58,087][starVLA.training.train_starvla][INFO] - Step 4917, Loss: {'train/loss': 1.4379601225300576e-06, 'train/grad_norm_pre_clip': 0.00037232739850878716, 'timing/data': 0.00032499000371899456, 'timing/model': 0.39277800600393675, 'learning_rate/qwen_vl_interface': 5.067239564989968e-07, 'learning_rate/action_model': 5.067239564989968e-06, 'epoch': 0.36}) +[2026-10-03 14:00:58,789][starVLA.training.train_starvla][INFO] - Step 4918, Loss: {'train/loss': 2.0936090550094377e-06, 'train/grad_norm_pre_clip': 0.0005141552537679672, 'timing/data': 0.00024361000396311283, 'timing/model': 0.39225251600146294, 'learning_rate/qwen_vl_interface': 5.065629465893808e-07, 'learning_rate/action_model': 5.065629465893807e-06, 'epoch': 0.36}) +[2026-10-03 14:00:59,490][starVLA.training.train_starvla][INFO] - Step 4919, Loss: {'train/loss': 0.02189304120838642, 'train/grad_norm_pre_clip': 99.90435028076172, 'timing/data': 0.00021236999600660056, 'timing/model': 0.39174178600660525, 'learning_rate/qwen_vl_interface': 5.064038865275625e-07, 'learning_rate/action_model': 5.0640388652756245e-06, 'epoch': 0.36}) +[2026-10-03 14:01:00,192][starVLA.training.train_starvla][INFO] - Step 4920, Loss: {'train/loss': 0.46785637736320496, 'train/grad_norm_pre_clip': 29.075092315673828, 'timing/data': 0.0002617100108182058, 'timing/model': 0.39240126599906944, 'learning_rate/qwen_vl_interface': 5.062467763789242e-07, 'learning_rate/action_model': 5.062467763789242e-06, 'epoch': 0.36}) +[2026-10-03 14:01:01,088][starVLA.training.train_starvla][INFO] - Step 4921, Loss: {'train/loss': 0.001478303223848343, 'train/grad_norm_pre_clip': 3.5619254112243652, 'timing/data': 0.0002620999875944108, 'timing/model': 0.3932059460057644, 'learning_rate/qwen_vl_interface': 5.060916162080496e-07, 'learning_rate/action_model': 5.060916162080496e-06, 'epoch': 0.36}) +[2026-10-03 14:01:01,794][starVLA.training.train_starvla][INFO] - Step 4922, Loss: {'train/loss': 2.073284485959448e-05, 'train/grad_norm_pre_clip': 0.0114009203389287, 'timing/data': 0.00027567999495659024, 'timing/model': 0.3918875859963009, 'learning_rate/qwen_vl_interface': 5.059384060787182e-07, 'learning_rate/action_model': 5.059384060787183e-06, 'epoch': 0.36}) +[2026-10-03 14:01:02,495][starVLA.training.train_starvla][INFO] - Step 4923, Loss: {'train/loss': 3.582027056836523e-05, 'train/grad_norm_pre_clip': 0.007478578016161919, 'timing/data': 0.00020111999765504152, 'timing/model': 0.39144823601236567, 'learning_rate/qwen_vl_interface': 5.05787146053909e-07, 'learning_rate/action_model': 5.0578714605390904e-06, 'epoch': 0.36}) +[2026-10-03 14:01:03,197][starVLA.training.train_starvla][INFO] - Step 4924, Loss: {'train/loss': 0.000103842539829202, 'train/grad_norm_pre_clip': 0.07784208655357361, 'timing/data': 0.00019455999427009374, 'timing/model': 0.3919978760095546, 'learning_rate/qwen_vl_interface': 5.056378361957998e-07, 'learning_rate/action_model': 5.056378361957998e-06, 'epoch': 0.36}) +[2026-10-03 14:01:03,900][starVLA.training.train_starvla][INFO] - Step 4925, Loss: {'train/loss': 0.0003283326223026961, 'train/grad_norm_pre_clip': 0.33376115560531616, 'timing/data': 0.00031310001213569194, 'timing/model': 0.3932656160031911, 'learning_rate/qwen_vl_interface': 5.05490476565766e-07, 'learning_rate/action_model': 5.054904765657659e-06, 'epoch': 0.36}) +[2026-10-03 14:01:04,794][starVLA.training.train_starvla][INFO] - Step 4926, Loss: {'train/loss': 0.002447648672387004, 'train/grad_norm_pre_clip': 3.1095449924468994, 'timing/data': 0.00030162000621203333, 'timing/model': 0.3924216160085052, 'learning_rate/qwen_vl_interface': 5.053450672243811e-07, 'learning_rate/action_model': 5.0534506722438114e-06, 'epoch': 0.36}) +[2026-10-03 14:01:05,495][starVLA.training.train_starvla][INFO] - Step 4927, Loss: {'train/loss': 4.947110028297175e-06, 'train/grad_norm_pre_clip': 0.0011884081177413464, 'timing/data': 0.00020742000197060406, 'timing/model': 0.3917281659960281, 'learning_rate/qwen_vl_interface': 5.052016082314184e-07, 'learning_rate/action_model': 5.052016082314184e-06, 'epoch': 0.36}) +[2026-10-03 14:01:06,197][starVLA.training.train_starvla][INFO] - Step 4928, Loss: {'train/loss': 3.0319870347739197e-05, 'train/grad_norm_pre_clip': 0.012141785584390163, 'timing/data': 0.00023215998953673989, 'timing/model': 0.39239889600139577, 'learning_rate/qwen_vl_interface': 5.050600996458473e-07, 'learning_rate/action_model': 5.050600996458473e-06, 'epoch': 0.36}) +[2026-10-03 14:01:06,900][starVLA.training.train_starvla][INFO] - Step 4929, Loss: {'train/loss': 5.155712642590515e-06, 'train/grad_norm_pre_clip': 0.0020511511247605085, 'timing/data': 0.0002471699990564957, 'timing/model': 0.3932298150029965, 'learning_rate/qwen_vl_interface': 5.049205415258373e-07, 'learning_rate/action_model': 5.049205415258373e-06, 'epoch': 0.36}) +[2026-10-03 14:01:07,601][starVLA.training.train_starvla][INFO] - Step 4930, Loss: {'train/loss': 2.1150353859411553e-05, 'train/grad_norm_pre_clip': 0.006405199877917767, 'timing/data': 0.00020138999389018863, 'timing/model': 0.3920365959929768, 'learning_rate/qwen_vl_interface': 5.047829339287557e-07, 'learning_rate/action_model': 5.047829339287557e-06, 'epoch': 0.36}) +[2026-10-03 14:01:08,495][starVLA.training.train_starvla][INFO] - Step 4931, Loss: {'train/loss': 9.499319276073948e-06, 'train/grad_norm_pre_clip': 0.003025555517524481, 'timing/data': 0.0003149700060021132, 'timing/model': 0.5853786989900982, 'learning_rate/qwen_vl_interface': 5.046472769111668e-07, 'learning_rate/action_model': 5.046472769111667e-06, 'epoch': 0.36}) +[2026-10-03 14:01:09,196][starVLA.training.train_starvla][INFO] - Step 4932, Loss: {'train/loss': 1.9847797375405207e-05, 'train/grad_norm_pre_clip': 0.004945309832692146, 'timing/data': 0.00024046000908128917, 'timing/model': 0.3919957959878957, 'learning_rate/qwen_vl_interface': 5.045135705288353e-07, 'learning_rate/action_model': 5.0451357052883525e-06, 'epoch': 0.36}) +[2026-10-03 14:01:09,898][starVLA.training.train_starvla][INFO] - Step 4933, Loss: {'train/loss': 8.898350642994046e-05, 'train/grad_norm_pre_clip': 0.2093983292579651, 'timing/data': 0.0002689100074348971, 'timing/model': 0.39313179699820466, 'learning_rate/qwen_vl_interface': 5.043818148367221e-07, 'learning_rate/action_model': 5.043818148367221e-06, 'epoch': 0.36}) +[2026-10-03 14:01:10,600][starVLA.training.train_starvla][INFO] - Step 4934, Loss: {'train/loss': 4.596948656399036e-06, 'train/grad_norm_pre_clip': 0.0014432952739298344, 'timing/data': 0.00033337999775540084, 'timing/model': 0.39185030599765014, 'learning_rate/qwen_vl_interface': 5.042520098889869e-07, 'learning_rate/action_model': 5.0425200988898685e-06, 'epoch': 0.36}) +[2026-10-03 14:01:11,300][starVLA.training.train_starvla][INFO] - Step 4935, Loss: {'train/loss': 1.4528611700370675e-06, 'train/grad_norm_pre_clip': 0.00040584782254882157, 'timing/data': 0.00029626999457832426, 'timing/model': 0.3911116459930781, 'learning_rate/qwen_vl_interface': 5.041241557389879e-07, 'learning_rate/action_model': 5.041241557389879e-06, 'epoch': 0.36}) +[2026-10-03 14:01:12,002][starVLA.training.train_starvla][INFO] - Step 4936, Loss: {'train/loss': 0.04659735783934593, 'train/grad_norm_pre_clip': 356.9566345214844, 'timing/data': 0.0003210300055798143, 'timing/model': 0.39230267600214574, 'learning_rate/qwen_vl_interface': 5.039982524392814e-07, 'learning_rate/action_model': 5.039982524392814e-06, 'epoch': 0.36}) +[2026-10-03 14:01:12,894][starVLA.training.train_starvla][INFO] - Step 4937, Loss: {'train/loss': 0.3742743730545044, 'train/grad_norm_pre_clip': 125.89167022705078, 'timing/data': 0.0002690399996936321, 'timing/model': 0.3931312060012715, 'learning_rate/qwen_vl_interface': 5.038743000416207e-07, 'learning_rate/action_model': 5.0387430004162064e-06, 'epoch': 0.36}) +[2026-10-03 14:01:13,595][starVLA.training.train_starvla][INFO] - Step 4938, Loss: {'train/loss': 0.009002035483717918, 'train/grad_norm_pre_clip': 31.266183853149414, 'timing/data': 0.00019576000340748578, 'timing/model': 0.39177071599988267, 'learning_rate/qwen_vl_interface': 5.037522985969588e-07, 'learning_rate/action_model': 5.037522985969587e-06, 'epoch': 0.36}) +[2026-10-03 14:01:14,295][starVLA.training.train_starvla][INFO] - Step 4939, Loss: {'train/loss': 2.9181232093833387e-05, 'train/grad_norm_pre_clip': 0.007924581877887249, 'timing/data': 0.00019602000247687101, 'timing/model': 0.39131587601150386, 'learning_rate/qwen_vl_interface': 5.036322481554451e-07, 'learning_rate/action_model': 5.036322481554451e-06, 'epoch': 0.36}) +[2026-10-03 14:01:14,997][starVLA.training.train_starvla][INFO] - Step 4940, Loss: {'train/loss': 2.6300431272829883e-06, 'train/grad_norm_pre_clip': 0.0005999827408231795, 'timing/data': 0.0002053200005320832, 'timing/model': 0.39266295499692205, 'learning_rate/qwen_vl_interface': 5.035141487664285e-07, 'learning_rate/action_model': 5.035141487664285e-06, 'epoch': 0.36}) +[2026-10-03 14:01:15,699][starVLA.training.train_starvla][INFO] - Step 4941, Loss: {'train/loss': 8.530521881766617e-06, 'train/grad_norm_pre_clip': 0.008094682358205318, 'timing/data': 0.0002970100031234324, 'timing/model': 0.39342348600621335, 'learning_rate/qwen_vl_interface': 5.03398000478455e-07, 'learning_rate/action_model': 5.03398000478455e-06, 'epoch': 0.36}) +[2026-10-03 14:01:16,593][starVLA.training.train_starvla][INFO] - Step 4942, Loss: {'train/loss': 0.0002557964180596173, 'train/grad_norm_pre_clip': 0.4175308048725128, 'timing/data': 0.00021355999342631549, 'timing/model': 0.3923066769930301, 'learning_rate/qwen_vl_interface': 5.032838033392686e-07, 'learning_rate/action_model': 5.032838033392685e-06, 'epoch': 0.36}) +[2026-10-03 14:01:17,296][starVLA.training.train_starvla][INFO] - Step 4943, Loss: {'train/loss': 0.0003030174702871591, 'train/grad_norm_pre_clip': 0.6680824756622314, 'timing/data': 0.00029855000320822, 'timing/model': 0.3920477460051188, 'learning_rate/qwen_vl_interface': 5.031715573958111e-07, 'learning_rate/action_model': 5.031715573958111e-06, 'epoch': 0.36}) +[2026-10-03 14:01:17,998][starVLA.training.train_starvla][INFO] - Step 4944, Loss: {'train/loss': 2.2575143248104723e-06, 'train/grad_norm_pre_clip': 0.0007048863917589188, 'timing/data': 0.0003039900038857013, 'timing/model': 0.3926227849879069, 'learning_rate/qwen_vl_interface': 5.030612626942235e-07, 'learning_rate/action_model': 5.030612626942235e-06, 'epoch': 0.36}) +[2026-10-03 14:01:18,702][starVLA.training.train_starvla][INFO] - Step 4945, Loss: {'train/loss': 0.2189992070198059, 'train/grad_norm_pre_clip': 115.75572967529297, 'timing/data': 0.00036551000084728, 'timing/model': 0.3935510760056786, 'learning_rate/qwen_vl_interface': 5.029529192798433e-07, 'learning_rate/action_model': 5.029529192798433e-06, 'epoch': 0.36}) +[2026-10-03 14:01:19,404][starVLA.training.train_starvla][INFO] - Step 4946, Loss: {'train/loss': 0.10715783387422562, 'train/grad_norm_pre_clip': 182.861328125, 'timing/data': 0.0003841400030069053, 'timing/model': 0.3921772360044997, 'learning_rate/qwen_vl_interface': 5.028465271972058e-07, 'learning_rate/action_model': 5.028465271972058e-06, 'epoch': 0.36}) +[2026-10-03 14:01:20,299][starVLA.training.train_starvla][INFO] - Step 4947, Loss: {'train/loss': 1.2925746887049172e-05, 'train/grad_norm_pre_clip': 0.00834721140563488, 'timing/data': 0.00025351000658702105, 'timing/model': 0.39242894599738065, 'learning_rate/qwen_vl_interface': 5.027420864900457e-07, 'learning_rate/action_model': 5.027420864900456e-06, 'epoch': 0.36}) +[2026-10-03 14:01:21,002][starVLA.training.train_starvla][INFO] - Step 4948, Loss: {'train/loss': 1.779082049324643e-05, 'train/grad_norm_pre_clip': 0.004457545466721058, 'timing/data': 0.0002064099971903488, 'timing/model': 0.39266710700758267, 'learning_rate/qwen_vl_interface': 5.026395972012943e-07, 'learning_rate/action_model': 5.0263959720129436e-06, 'epoch': 0.36}) +[2026-10-03 14:01:21,705][starVLA.training.train_starvla][INFO] - Step 4949, Loss: {'train/loss': 5.4983820518828e-06, 'train/grad_norm_pre_clip': 0.004132134839892387, 'timing/data': 0.0003971599944634363, 'timing/model': 0.3936223469936522, 'learning_rate/qwen_vl_interface': 5.025390593730809e-07, 'learning_rate/action_model': 5.025390593730809e-06, 'epoch': 0.36}) +[2026-10-03 14:01:22,407][starVLA.training.train_starvla][INFO] - Step 4950, Loss: {'train/loss': 0.26547643542289734, 'train/grad_norm_pre_clip': 37.84031295776367, 'timing/data': 0.0002471899933880195, 'timing/model': 0.392403325997293, 'learning_rate/qwen_vl_interface': 5.024404730467325e-07, 'learning_rate/action_model': 5.024404730467325e-06, 'epoch': 0.36}) +[2026-10-03 14:01:23,108][starVLA.training.train_starvla][INFO] - Step 4951, Loss: {'train/loss': 2.8002948965877295e-05, 'train/grad_norm_pre_clip': 0.012522821314632893, 'timing/data': 0.00024496999685652554, 'timing/model': 0.3917061160027515, 'learning_rate/qwen_vl_interface': 5.023438382627749e-07, 'learning_rate/action_model': 5.0234383826277486e-06, 'epoch': 0.36}) +[2026-10-03 14:01:24,013][starVLA.training.train_starvla][INFO] - Step 4952, Loss: {'train/loss': 6.87845385982655e-05, 'train/grad_norm_pre_clip': 0.09763924032449722, 'timing/data': 0.0002405700070085004, 'timing/model': 0.5959222689998569, 'learning_rate/qwen_vl_interface': 5.022491550609311e-07, 'learning_rate/action_model': 5.022491550609311e-06, 'epoch': 0.36}) +[2026-10-03 14:01:24,715][starVLA.training.train_starvla][INFO] - Step 4953, Loss: {'train/loss': 0.43108469247817993, 'train/grad_norm_pre_clip': 51.04988098144531, 'timing/data': 0.00021380001271609217, 'timing/model': 0.3933470949996263, 'learning_rate/qwen_vl_interface': 5.021564234801213e-07, 'learning_rate/action_model': 5.021564234801213e-06, 'epoch': 0.36}) +[2026-10-03 14:01:25,416][starVLA.training.train_starvla][INFO] - Step 4954, Loss: {'train/loss': 1.6002240954549052e-05, 'train/grad_norm_pre_clip': 0.005519486498087645, 'timing/data': 0.00020046999270562083, 'timing/model': 0.3917088260059245, 'learning_rate/qwen_vl_interface': 5.020656435584636e-07, 'learning_rate/action_model': 5.020656435584635e-06, 'epoch': 0.36}) +[2026-10-03 14:01:26,118][starVLA.training.train_starvla][INFO] - Step 4955, Loss: {'train/loss': 8.381628504139371e-06, 'train/grad_norm_pre_clip': 0.0019333584932610393, 'timing/data': 0.0003396700049052015, 'timing/model': 0.39216214598855004, 'learning_rate/qwen_vl_interface': 5.019768153332752e-07, 'learning_rate/action_model': 5.019768153332752e-06, 'epoch': 0.36}) +[2026-10-03 14:01:26,822][starVLA.training.train_starvla][INFO] - Step 4956, Loss: {'train/loss': 7.912381988717243e-06, 'train/grad_norm_pre_clip': 0.003158869920298457, 'timing/data': 0.00022200000239536166, 'timing/model': 0.3929276859998936, 'learning_rate/qwen_vl_interface': 5.018899388410693e-07, 'learning_rate/action_model': 5.0188993884106935e-06, 'epoch': 0.36}) +[2026-10-03 14:01:27,715][starVLA.training.train_starvla][INFO] - Step 4957, Loss: {'train/loss': 2.721393684623763e-05, 'train/grad_norm_pre_clip': 0.018785662949085236, 'timing/data': 0.00029193000227678567, 'timing/model': 0.5830331890028901, 'learning_rate/qwen_vl_interface': 5.018050141175576e-07, 'learning_rate/action_model': 5.018050141175575e-06, 'epoch': 0.36}) +[2026-10-03 14:01:28,416][starVLA.training.train_starvla][INFO] - Step 4958, Loss: {'train/loss': 6.370704068103805e-05, 'train/grad_norm_pre_clip': 0.04064546525478363, 'timing/data': 0.00024202998611144722, 'timing/model': 0.3918931959924521, 'learning_rate/qwen_vl_interface': 5.017220411976496e-07, 'learning_rate/action_model': 5.0172204119764965e-06, 'epoch': 0.36}) +[2026-10-03 14:01:29,117][starVLA.training.train_starvla][INFO] - Step 4959, Loss: {'train/loss': 0.002355030970647931, 'train/grad_norm_pre_clip': 5.874264240264893, 'timing/data': 0.0002987299958476797, 'timing/model': 0.39176129701081663, 'learning_rate/qwen_vl_interface': 5.016410201154523e-07, 'learning_rate/action_model': 5.016410201154523e-06, 'epoch': 0.36}) +[2026-10-03 14:01:29,819][starVLA.training.train_starvla][INFO] - Step 4960, Loss: {'train/loss': 1.2970549505553208e-05, 'train/grad_norm_pre_clip': 0.011192115023732185, 'timing/data': 0.0002944899897556752, 'timing/model': 0.39237682601378765, 'learning_rate/qwen_vl_interface': 5.015619509042709e-07, 'learning_rate/action_model': 5.015619509042709e-06, 'epoch': 0.36}) +[2026-10-03 14:01:30,523][starVLA.training.train_starvla][INFO] - Step 4961, Loss: {'train/loss': 0.000856477243360132, 'train/grad_norm_pre_clip': 2.7910726070404053, 'timing/data': 0.0002597400016384199, 'timing/model': 0.3934674060001271, 'learning_rate/qwen_vl_interface': 5.014848335966067e-07, 'learning_rate/action_model': 5.014848335966067e-06, 'epoch': 0.36}) +[2026-10-03 14:01:31,224][starVLA.training.train_starvla][INFO] - Step 4962, Loss: {'train/loss': 2.8360802389215678e-05, 'train/grad_norm_pre_clip': 0.02916678786277771, 'timing/data': 0.0002442400000290945, 'timing/model': 0.39177392699639313, 'learning_rate/qwen_vl_interface': 5.014096682241608e-07, 'learning_rate/action_model': 5.014096682241608e-06, 'epoch': 0.36}) +[2026-10-03 14:01:32,118][starVLA.training.train_starvla][INFO] - Step 4963, Loss: {'train/loss': 0.01911124587059021, 'train/grad_norm_pre_clip': 87.47187042236328, 'timing/data': 0.0002617499994812533, 'timing/model': 0.3916997659980552, 'learning_rate/qwen_vl_interface': 5.013364548178299e-07, 'learning_rate/action_model': 5.013364548178299e-06, 'epoch': 0.36}) +[2026-10-03 14:01:32,819][starVLA.training.train_starvla][INFO] - Step 4964, Loss: {'train/loss': 0.00027159214369021356, 'train/grad_norm_pre_clip': 0.44656866788864136, 'timing/data': 0.0002974300005007535, 'timing/model': 0.39192914600425866, 'learning_rate/qwen_vl_interface': 5.012651934077101e-07, 'learning_rate/action_model': 5.0126519340771e-06, 'epoch': 0.36}) +[2026-10-03 14:01:33,521][starVLA.training.train_starvla][INFO] - Step 4965, Loss: {'train/loss': 5.7740930969885085e-06, 'train/grad_norm_pre_clip': 0.002117346040904522, 'timing/data': 0.0002960700076073408, 'timing/model': 0.3933866860024864, 'learning_rate/qwen_vl_interface': 5.01195884023094e-07, 'learning_rate/action_model': 5.0119588402309395e-06, 'epoch': 0.36}) +[2026-10-03 14:01:34,224][starVLA.training.train_starvla][INFO] - Step 4966, Loss: {'train/loss': 1.5227903531922493e-05, 'train/grad_norm_pre_clip': 0.011321129277348518, 'timing/data': 0.000258120009675622, 'timing/model': 0.39195429599203635, 'learning_rate/qwen_vl_interface': 5.011285266924714e-07, 'learning_rate/action_model': 5.011285266924714e-06, 'epoch': 0.36}) +[2026-10-03 14:01:34,925][starVLA.training.train_starvla][INFO] - Step 4967, Loss: {'train/loss': 3.330014806124382e-05, 'train/grad_norm_pre_clip': 0.02476346120238304, 'timing/data': 0.00020185999164823443, 'timing/model': 0.3921293260063976, 'learning_rate/qwen_vl_interface': 5.010631214435312e-07, 'learning_rate/action_model': 5.010631214435312e-06, 'epoch': 0.36}) +[2026-10-03 14:01:35,817][starVLA.training.train_starvla][INFO] - Step 4968, Loss: {'train/loss': 2.0636665794881992e-05, 'train/grad_norm_pre_clip': 0.005203263834118843, 'timing/data': 0.0003417900006752461, 'timing/model': 0.5827475400001276, 'learning_rate/qwen_vl_interface': 5.009996683031586e-07, 'learning_rate/action_model': 5.009996683031586e-06, 'epoch': 0.36}) +[2026-10-03 14:01:36,521][starVLA.training.train_starvla][INFO] - Step 4969, Loss: {'train/loss': 4.4479679672804195e-06, 'train/grad_norm_pre_clip': 0.0011153408559039235, 'timing/data': 0.00030186999356374145, 'timing/model': 0.39329203600937035, 'learning_rate/qwen_vl_interface': 5.009381672974375e-07, 'learning_rate/action_model': 5.009381672974375e-06, 'epoch': 0.36}) +[2026-10-03 14:01:37,222][starVLA.training.train_starvla][INFO] - Step 4970, Loss: {'train/loss': 2.1979142275085906e-06, 'train/grad_norm_pre_clip': 0.00046866279444657266, 'timing/data': 0.00024440001288894564, 'timing/model': 0.39181998700951226, 'learning_rate/qwen_vl_interface': 5.008786184516475e-07, 'learning_rate/action_model': 5.008786184516474e-06, 'epoch': 0.36}) +[2026-10-03 14:01:37,924][starVLA.training.train_starvla][INFO] - Step 4971, Loss: {'train/loss': 8.269998943433166e-06, 'train/grad_norm_pre_clip': 0.0024223672226071358, 'timing/data': 0.00025718999677337706, 'timing/model': 0.39237188700644765, 'learning_rate/qwen_vl_interface': 5.008210217902676e-07, 'learning_rate/action_model': 5.008210217902676e-06, 'epoch': 0.36}) +[2026-10-03 14:01:38,625][starVLA.training.train_starvla][INFO] - Step 4972, Loss: {'train/loss': 8.202878234442323e-06, 'train/grad_norm_pre_clip': 0.004079387057572603, 'timing/data': 0.00019581000378821045, 'timing/model': 0.39200696600892115, 'learning_rate/qwen_vl_interface': 5.007653773369737e-07, 'learning_rate/action_model': 5.007653773369736e-06, 'epoch': 0.36}) +[2026-10-03 14:01:39,327][starVLA.training.train_starvla][INFO] - Step 4973, Loss: {'train/loss': 6.230408325791359e-05, 'train/grad_norm_pre_clip': 0.03911348432302475, 'timing/data': 0.00028466999356169254, 'timing/model': 0.3930600460007554, 'learning_rate/qwen_vl_interface': 5.007116851146386e-07, 'learning_rate/action_model': 5.007116851146386e-06, 'epoch': 0.36}) +[2026-10-03 14:01:40,224][starVLA.training.train_starvla][INFO] - Step 4974, Loss: {'train/loss': 1.7560081687406637e-05, 'train/grad_norm_pre_clip': 0.0035559365060180426, 'timing/data': 0.00021259000641293824, 'timing/model': 0.3927344560070196, 'learning_rate/qwen_vl_interface': 5.006599451453334e-07, 'learning_rate/action_model': 5.006599451453334e-06, 'epoch': 0.36}) +[2026-10-03 14:01:40,926][starVLA.training.train_starvla][INFO] - Step 4975, Loss: {'train/loss': 4.446216917131096e-05, 'train/grad_norm_pre_clip': 0.01925494521856308, 'timing/data': 0.00021738000214099884, 'timing/model': 0.3920809560077032, 'learning_rate/qwen_vl_interface': 5.006101574503268e-07, 'learning_rate/action_model': 5.006101574503267e-06, 'epoch': 0.36}) +[2026-10-03 14:01:41,629][starVLA.training.train_starvla][INFO] - Step 4976, Loss: {'train/loss': 1.4401543921849225e-05, 'train/grad_norm_pre_clip': 0.006313875317573547, 'timing/data': 0.00020614000095520169, 'timing/model': 0.39246434600499924, 'learning_rate/qwen_vl_interface': 5.005623220500839e-07, 'learning_rate/action_model': 5.005623220500839e-06, 'epoch': 0.36}) +[2026-10-03 14:01:42,331][starVLA.training.train_starvla][INFO] - Step 4977, Loss: {'train/loss': 0.001162198488600552, 'train/grad_norm_pre_clip': 7.571991920471191, 'timing/data': 0.00020260000019334257, 'timing/model': 0.3928078359895153, 'learning_rate/qwen_vl_interface': 5.005164389642687e-07, 'learning_rate/action_model': 5.005164389642687e-06, 'epoch': 0.36}) +[2026-10-03 14:01:43,032][starVLA.training.train_starvla][INFO] - Step 4978, Loss: {'train/loss': 4.697122858488001e-05, 'train/grad_norm_pre_clip': 0.03829735144972801, 'timing/data': 0.00019858000450767577, 'timing/model': 0.39205627600313164, 'learning_rate/qwen_vl_interface': 5.004725082117423e-07, 'learning_rate/action_model': 5.004725082117422e-06, 'epoch': 0.36}) +[2026-10-03 14:01:43,929][starVLA.training.train_starvla][INFO] - Step 4979, Loss: {'train/loss': 1.0542251402512193e-05, 'train/grad_norm_pre_clip': 0.002088358858600259, 'timing/data': 0.0002884999994421378, 'timing/model': 0.39290912699652836, 'learning_rate/qwen_vl_interface': 5.004305298105616e-07, 'learning_rate/action_model': 5.004305298105616e-06, 'epoch': 0.36}) +[2026-10-03 14:01:44,631][starVLA.training.train_starvla][INFO] - Step 4980, Loss: {'train/loss': 1.6331077858922072e-05, 'train/grad_norm_pre_clip': 0.003113131271675229, 'timing/data': 0.00019956000323873013, 'timing/model': 0.39282625599298626, 'learning_rate/qwen_vl_interface': 5.003905037779834e-07, 'learning_rate/action_model': 5.0039050377798345e-06, 'epoch': 0.36}) +[2026-10-03 14:01:45,335][starVLA.training.train_starvla][INFO] - Step 4981, Loss: {'train/loss': 0.00010008728713728487, 'train/grad_norm_pre_clip': 0.09262792021036148, 'timing/data': 0.00026982001145370305, 'timing/model': 0.3940834859968163, 'learning_rate/qwen_vl_interface': 5.003524301304608e-07, 'learning_rate/action_model': 5.003524301304608e-06, 'epoch': 0.36}) +[2026-10-03 14:01:46,037][starVLA.training.train_starvla][INFO] - Step 4982, Loss: {'train/loss': 9.797011443879455e-06, 'train/grad_norm_pre_clip': 0.0023744176141917706, 'timing/data': 0.00033907999750226736, 'timing/model': 0.39226173600764014, 'learning_rate/qwen_vl_interface': 5.003163088836445e-07, 'learning_rate/action_model': 5.003163088836446e-06, 'epoch': 0.36}) +[2026-10-03 14:01:46,740][starVLA.training.train_starvla][INFO] - Step 4983, Loss: {'train/loss': 1.557165660415194e-06, 'train/grad_norm_pre_clip': 0.00044152032933197916, 'timing/data': 0.00020466999558266252, 'timing/model': 0.3924056260002544, 'learning_rate/qwen_vl_interface': 5.002821400523818e-07, 'learning_rate/action_model': 5.002821400523818e-06, 'epoch': 0.36}) +[2026-10-03 14:01:47,632][starVLA.training.train_starvla][INFO] - Step 4984, Loss: {'train/loss': 2.905716655732249e-06, 'train/grad_norm_pre_clip': 0.0006820460548624396, 'timing/data': 0.000331000002915971, 'timing/model': 0.5824648499983596, 'learning_rate/qwen_vl_interface': 5.002499236507192e-07, 'learning_rate/action_model': 5.0024992365071925e-06, 'epoch': 0.36}) +[2026-10-03 14:01:48,334][starVLA.training.train_starvla][INFO] - Step 4985, Loss: {'train/loss': 3.615296009229496e-05, 'train/grad_norm_pre_clip': 0.009612113237380981, 'timing/data': 0.00030016999517101794, 'timing/model': 0.3927068470075028, 'learning_rate/qwen_vl_interface': 5.002196596918993e-07, 'learning_rate/action_model': 5.002196596918993e-06, 'epoch': 0.36}) +[2026-10-03 14:01:49,044][starVLA.training.train_starvla][INFO] - Step 4986, Loss: {'train/loss': 0.4012347459793091, 'train/grad_norm_pre_clip': 69.51343536376953, 'timing/data': 0.0002977800031658262, 'timing/model': 0.4005860760080395, 'learning_rate/qwen_vl_interface': 5.001913481883623e-07, 'learning_rate/action_model': 5.001913481883623e-06, 'epoch': 0.36}) +[2026-10-03 14:01:49,746][starVLA.training.train_starvla][INFO] - Step 4987, Loss: {'train/loss': 6.486335041699931e-05, 'train/grad_norm_pre_clip': 0.021331196650862694, 'timing/data': 0.00031470999238081276, 'timing/model': 0.39246182500210125, 'learning_rate/qwen_vl_interface': 5.001649891517459e-07, 'learning_rate/action_model': 5.001649891517459e-06, 'epoch': 0.37}) +[2026-10-03 14:01:50,448][starVLA.training.train_starvla][INFO] - Step 4988, Loss: {'train/loss': 0.3971462547779083, 'train/grad_norm_pre_clip': 44.68531036376953, 'timing/data': 0.00039103000017348677, 'timing/model': 0.39190326700918376, 'learning_rate/qwen_vl_interface': 5.001405825928858e-07, 'learning_rate/action_model': 5.0014058259288575e-06, 'epoch': 0.37}) +[2026-10-03 14:01:51,151][starVLA.training.train_starvla][INFO] - Step 4989, Loss: {'train/loss': 1.0169612323807087e-05, 'train/grad_norm_pre_clip': 0.006048529874533415, 'timing/data': 0.0002493900101399049, 'timing/model': 0.3930546460032929, 'learning_rate/qwen_vl_interface': 5.001181285218143e-07, 'learning_rate/action_model': 5.0011812852181434e-06, 'epoch': 0.37}) +[2026-10-03 14:01:52,042][starVLA.training.train_starvla][INFO] - Step 4990, Loss: {'train/loss': 1.8692197045311332e-05, 'train/grad_norm_pre_clip': 0.004659953061491251, 'timing/data': 0.00033205900399480015, 'timing/model': 0.39178337701014243, 'learning_rate/qwen_vl_interface': 5.000976269477613e-07, 'learning_rate/action_model': 5.0009762694776125e-06, 'epoch': 0.37}) +[2026-10-03 14:01:52,743][starVLA.training.train_starvla][INFO] - Step 4991, Loss: {'train/loss': 0.00014590442879125476, 'train/grad_norm_pre_clip': 0.2116745263338089, 'timing/data': 0.000219000008655712, 'timing/model': 0.3916484569926979, 'learning_rate/qwen_vl_interface': 5.000790778791542e-07, 'learning_rate/action_model': 5.000790778791543e-06, 'epoch': 0.37}) +[2026-10-03 14:01:53,441][starVLA.training.train_starvla][INFO] - Step 4992, Loss: {'train/loss': 0.00012088051880709827, 'train/grad_norm_pre_clip': 0.05292405188083649, 'timing/data': 0.00023778999457135797, 'timing/model': 0.3907271659991238, 'learning_rate/qwen_vl_interface': 5.000624813236185e-07, 'learning_rate/action_model': 5.000624813236185e-06, 'epoch': 0.37}) +[2026-10-03 14:01:54,142][starVLA.training.train_starvla][INFO] - Step 4993, Loss: {'train/loss': 2.0434312318684533e-05, 'train/grad_norm_pre_clip': 0.013889612630009651, 'timing/data': 0.00030520001018885523, 'timing/model': 0.39233092599897645, 'learning_rate/qwen_vl_interface': 5.000478372879752e-07, 'learning_rate/action_model': 5.000478372879752e-06, 'epoch': 0.37}) +[2026-10-03 14:01:54,843][starVLA.training.train_starvla][INFO] - Step 4994, Loss: {'train/loss': 3.841768193524331e-05, 'train/grad_norm_pre_clip': 0.029048964381217957, 'timing/data': 0.0002261199988424778, 'timing/model': 0.39162978700187523, 'learning_rate/qwen_vl_interface': 5.000351457782452e-07, 'learning_rate/action_model': 5.000351457782451e-06, 'epoch': 0.37}) +[2026-10-03 14:01:55,733][starVLA.training.train_starvla][INFO] - Step 4995, Loss: {'train/loss': 9.871608199318871e-06, 'train/grad_norm_pre_clip': 0.004965235013514757, 'timing/data': 0.00021551999088842422, 'timing/model': 0.3917257760040229, 'learning_rate/qwen_vl_interface': 5.000244067996445e-07, 'learning_rate/action_model': 5.000244067996445e-06, 'epoch': 0.37}) +[2026-10-03 14:01:56,434][starVLA.training.train_starvla][INFO] - Step 4996, Loss: {'train/loss': 0.42053642868995667, 'train/grad_norm_pre_clip': 62.04148864746094, 'timing/data': 0.00020169999334029853, 'timing/model': 0.39157624600920826, 'learning_rate/qwen_vl_interface': 5.000156203565884e-07, 'learning_rate/action_model': 5.000156203565883e-06, 'epoch': 0.37}) +[2026-10-03 14:01:57,135][starVLA.training.train_starvla][INFO] - Step 4997, Loss: {'train/loss': 0.002274632453918457, 'train/grad_norm_pre_clip': 19.53311538696289, 'timing/data': 0.00024302999372594059, 'timing/model': 0.3927566869970178, 'learning_rate/qwen_vl_interface': 5.000087864526878e-07, 'learning_rate/action_model': 5.000087864526878e-06, 'epoch': 0.37}) +[2026-10-03 14:01:57,836][starVLA.training.train_starvla][INFO] - Step 4998, Loss: {'train/loss': 0.00010424750507809222, 'train/grad_norm_pre_clip': 0.09200233221054077, 'timing/data': 0.00030341999081429094, 'timing/model': 0.3916457659943262, 'learning_rate/qwen_vl_interface': 5.000039050907523e-07, 'learning_rate/action_model': 5.000039050907523e-06, 'epoch': 0.37}) +[2026-10-03 14:01:58,538][starVLA.training.train_starvla][INFO] - Step 4999, Loss: {'train/loss': 0.47276005148887634, 'train/grad_norm_pre_clip': 55.744144439697266, 'timing/data': 0.0001971700112335384, 'timing/model': 0.39207722700666636, 'learning_rate/qwen_vl_interface': 5.000009762727885e-07, 'learning_rate/action_model': 5.000009762727884e-06, 'epoch': 0.37}) +[2026-10-03 14:01:59,431][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 5000 +[2026-10-03 14:02:24,769][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 5000 +[2026-10-03 14:02:24,770][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 5000 +[2026-10-03 14:02:24,770][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 5000 +[2026-10-03 14:02:24,771][starVLA.training.train_starvla][INFO] - Step 5000, Loss: {'train/loss': 0.35525190830230713, 'train/grad_norm_pre_clip': 40.85348892211914, 'eval/action_loss/seconds': 25.336162989988225, 'eval/action_loss/samples': 120.0, 'eval/loss': 0.9833691120147705, 'eval/latency_0/loss': 1.2987778186798096, 'eval/latency_2/loss': 0.8945199847221375, 'eval/latency_4/loss': 0.7568093538284302, 'eval/inverted_pendulum_rgb_state/loss': 0.9833691120147705, 'eval/inverted_pendulum_rgb_state/latency_0/loss': 1.2987778186798096, 'eval/inverted_pendulum_rgb_state/latency_2/loss': 0.8945199847221375, 'eval/inverted_pendulum_rgb_state/latency_4/loss': 0.7568093538284302, 'timing/data': 0.00028153999301139265, 'timing/model': 0.3918095460103359, 'learning_rate/qwen_vl_interface': 5e-07, 'learning_rate/action_model': 4.9999999999999996e-06, 'epoch': 0.37}) +[2026-10-03 14:02:24,838][root][INFO] - /data/ubuntu/conda_envs/starvla_rl_games_openvla/bin/x86_64-conda-linux-gnu-cc -march=nocona -mtune=haswell -ftree-vectorize -fPIC -fstack-protector-strong -fno-plt -O2 -ffunction-sections -pipe -isystem /data/ubuntu/conda_envs/starvla_rl_games_openvla/include -I/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/include -L/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/lib -L/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/lib/stubs -DNDEBUG -D_FORTIFY_SOURCE=2 -O2 -isystem /data/ubuntu/conda_envs/starvla_rl_games_openvla/include -I/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/include -L/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/lib -L/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/lib/stubs -fPIC -march=nocona -mtune=haswell -ftree-vectorize -fPIC -fstack-protector-strong -fno-plt -O2 -ffunction-sections -pipe -isystem /data/ubuntu/conda_envs/starvla_rl_games_openvla/include -I/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/include -L/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/lib -L/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/lib/stubs -c /tmp/tmplcoka1u7/test.c -o /tmp/tmplcoka1u7/test.o +[2026-10-03 14:02:24,847][root][INFO] - /data/ubuntu/conda_envs/starvla_rl_games_openvla/bin/x86_64-conda-linux-gnu-cc -Wl,-O2 -Wl,--sort-common -Wl,--as-needed -Wl,-z,relro -Wl,-z,now -Wl,--disable-new-dtags -Wl,--gc-sections -Wl,--allow-shlib-undefined -Wl,-rpath,/data/ubuntu/conda_envs/starvla_rl_games_openvla/lib -Wl,-rpath-link,/data/ubuntu/conda_envs/starvla_rl_games_openvla/lib -L/data/ubuntu/conda_envs/starvla_rl_games_openvla/lib -L/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/lib -L/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/lib/stubs /tmp/tmplcoka1u7/test.o -laio -o /tmp/tmplcoka1u7/a.out +[2026-10-03 14:02:24,854][root][INFO] - /data/ubuntu/conda_envs/starvla_rl_games_openvla/bin/x86_64-conda-linux-gnu-cc -march=nocona -mtune=haswell -ftree-vectorize -fPIC -fstack-protector-strong -fno-plt -O2 -ffunction-sections -pipe -isystem /data/ubuntu/conda_envs/starvla_rl_games_openvla/include -I/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/include -L/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/lib -L/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/lib/stubs -DNDEBUG -D_FORTIFY_SOURCE=2 -O2 -isystem /data/ubuntu/conda_envs/starvla_rl_games_openvla/include -I/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/include -L/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/lib -L/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/lib/stubs -fPIC -march=nocona -mtune=haswell -ftree-vectorize -fPIC -fstack-protector-strong -fno-plt -O2 -ffunction-sections -pipe -isystem /data/ubuntu/conda_envs/starvla_rl_games_openvla/include -I/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/include -L/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/lib -L/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/lib/stubs -c /tmp/tmpppmsv_t3/test.c -o /tmp/tmpppmsv_t3/test.o +[2026-10-03 14:02:24,862][root][INFO] - /data/ubuntu/conda_envs/starvla_rl_games_openvla/bin/x86_64-conda-linux-gnu-cc -Wl,-O2 -Wl,--sort-common -Wl,--as-needed -Wl,-z,relro -Wl,-z,now -Wl,--disable-new-dtags -Wl,--gc-sections -Wl,--allow-shlib-undefined -Wl,-rpath,/data/ubuntu/conda_envs/starvla_rl_games_openvla/lib -Wl,-rpath-link,/data/ubuntu/conda_envs/starvla_rl_games_openvla/lib -L/data/ubuntu/conda_envs/starvla_rl_games_openvla/lib -L/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/lib -L/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/lib/stubs /tmp/tmpppmsv_t3/test.o -L/data/ubuntu/conda_envs/starvla_rl_games_openvla -L/data/ubuntu/conda_envs/starvla_rl_games_openvla/lib64 -lcufile -o /tmp/tmpppmsv_t3/a.out +[2026-10-03 14:02:24,875][root][INFO] - /data/ubuntu/conda_envs/starvla_rl_games_openvla/bin/x86_64-conda-linux-gnu-cc -march=nocona -mtune=haswell -ftree-vectorize -fPIC -fstack-protector-strong -fno-plt -O2 -ffunction-sections -pipe -isystem /data/ubuntu/conda_envs/starvla_rl_games_openvla/include -I/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/include -L/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/lib -L/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/lib/stubs -DNDEBUG -D_FORTIFY_SOURCE=2 -O2 -isystem /data/ubuntu/conda_envs/starvla_rl_games_openvla/include -I/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/include -L/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/lib -L/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/lib/stubs -fPIC -march=nocona -mtune=haswell -ftree-vectorize -fPIC -fstack-protector-strong -fno-plt -O2 -ffunction-sections -pipe -isystem /data/ubuntu/conda_envs/starvla_rl_games_openvla/include -I/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/include -L/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/lib -L/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/lib/stubs -c /tmp/tmp8cmlx4dv/test.c -o /tmp/tmp8cmlx4dv/test.o +[2026-10-03 14:02:24,883][root][INFO] - /data/ubuntu/conda_envs/starvla_rl_games_openvla/bin/x86_64-conda-linux-gnu-cc -Wl,-O2 -Wl,--sort-common -Wl,--as-needed -Wl,-z,relro -Wl,-z,now -Wl,--disable-new-dtags -Wl,--gc-sections -Wl,--allow-shlib-undefined -Wl,-rpath,/data/ubuntu/conda_envs/starvla_rl_games_openvla/lib -Wl,-rpath-link,/data/ubuntu/conda_envs/starvla_rl_games_openvla/lib -L/data/ubuntu/conda_envs/starvla_rl_games_openvla/lib -L/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/lib -L/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/lib/stubs /tmp/tmp8cmlx4dv/test.o -laio -o /tmp/tmp8cmlx4dv/a.out +[2026-10-03 14:02:29,189][starVLA.training.rl_games.checkpoint_sync][WARNING] - HF checkpoint sync enabled for talha15032/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2 but HF_TOKEN/HUGGINGFACE_HUB_TOKEN is not set; skipping upload +[2026-10-03 14:02:29,190][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-10-03 14:02:29,212][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-10-03 14:02:29,212][starVLA.training.train_starvla][INFO] - Training complete. Final training state is saved under /home/ubuntu/talha/results/Checkpoints/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2/checkpoints diff --git a/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2/summary.jsonl b/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2/summary.jsonl new file mode 100644 index 0000000000000000000000000000000000000000..04f41c132dfb80a0ccb65212db19ce96e77fd03c --- /dev/null +++ b/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp2/summary.jsonl @@ -0,0 +1 @@ +{"steps": 5000} diff --git a/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3/checkpoints/steps_5000_pytorch_model.pt b/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3/checkpoints/steps_5000_pytorch_model.pt new file mode 100644 index 0000000000000000000000000000000000000000..9d9c648fc9aec7fe7f5bfd48b784eb1beaed6bfd --- /dev/null +++ b/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3/checkpoints/steps_5000_pytorch_model.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2f41824572db70b0517fffee698dfbd6a2274494eb6aecac82e6f91edf47a219 +size 9138435024 diff --git a/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3/config.full.yaml b/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3/config.full.yaml new file mode 100644 index 0000000000000000000000000000000000000000..62864d215f9ef3c079e04b35d156bd9fb86fc0a4 --- /dev/null +++ b/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3/config.full.yaml @@ -0,0 +1,338 @@ +framework: + name: QwenOFT + qwenvl: + base_vlm: /home/ubuntu/talha/playground/Pretrained_models/Qwen3-VL-4B-Instruct + attn_implementation: flash_attention_2 + enable_gradient_checkpointing: true + action_model: + action_model_type: MLP + action_dim: 5 + state_dim: 4 + action_hidden_dim: 2560 + future_action_window_size: 0 + past_action_window_size: 0 + loss_type: discrete_ce + state_encoding: discretized_text + action_horizon: 1 + action_env_dim: 5 + kv_memory: + enabled: false + window: 4 + rollout_len: 8 + packed_train: false + rebased_sink: true +datasets: + vla_data: + dataset_py: lerobot_datasets + include_state: true + data_root_dir: /home/ubuntu/talha/playground/Datasets/rl_games + data_mix: openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3 + eval_data_mix: openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3__val + custom_mixtures_path: /home/ubuntu/talha/playground/Datasets/rl_games/_generated_mixtures/single_79881079067b9b1b.json + action_type: discrete + sequential_step_sampling: true + eval_sequential_step_sampling: null + num_workers: 8 + eval_num_workers: 8 + prefetch_factor: 4 + persistent_workers: true + pin_memory: true + shuffle: true + action_balance: + enabled: false + strategy: balanced_epoch + action_key: action_id + target_flap_fraction: 0.3 + noop_id: 0 + flap_id: 1 + latency_curriculum: + enabled: false + strategy: exclusive + latencies: null + phase_steps: null + phase_distributions: null + new_latency_passes: 1.0 + replay_passes: 0.25 + target_total_passes: 2.0 + final_equalization: true + step_budget_mode: auto + eval_at_phase_end: false + save_at_phase_end: false + computed_plan: null + per_device_batch_size: 8 + load_all_data_for_training: true + num_obs_frames: 1 + image_mode: single + prompt_mode: raw + stitch_grid: + - 2 + - 2 + obs_image_size: null + video_backend: torchvision_av + gymnasium_task_contract: + task_name: inverted_pendulum_rgb_state + env_id: LatencyBench/InvertedPendulumDiscrete-v0 + registration_imports: + - latency_bench.envs.gymnasium_mujoco + env_fps: 25.0 + obs_fps: 25.0 + frame_stack: 1 + base_prompt: 'Keep the pole upright by applying horizontal cart forces. Choose + exactly one action from: push_left_hard, push_left_soft, zero_force, push_right_soft, + push_right_hard.' + action_labels: + - push_left_hard + - push_left_soft + - zero_force + - push_right_soft + - push_right_hard + action_values: + - 0 + - 1 + - 2 + - 3 + - 4 + noop_action_id: 2 + make_kwargs: + base_env_id: InvertedPendulum-v4 + render_mode: rgb_array + force_values: + - -3.0 + - -1.5 + - 0.0 + - 1.5 + - 3.0 + base_make_kwargs: {} + state_space: + labels: + - cart_position + - pole_angle + - cart_velocity + - pole_angular_velocity + active_action_dim: 5 +dataset: + source_hf: latency-sensitive-bench/inverted_pendulum_200ep_2 + config_name: null + source_subdir: inverted_pendulum_fixed_latency_0_100ep_1ksteps_og,inverted_pendulum_fixed_latency_2_1000ep_7k2steps,inverted_pendulum_fixed_latency_4_1000ep_7k2steps + converted_name: openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3 + single_source_hf: '' + mixed_source_hf: '' + single_converted_name: flappy_train + mixed_converted_name: flappy_train + single_latency_filter: null + mixed_latency_filter: null + force_download: false + setup_force: true + skip_verification: true + target_latency_unit: observation_steps + verify_rows: 200 + max_episodes: null + max_steps_per_episode: 1000 + episodes_per_latency: null + episodes_per_latency_by_latency: 0:90,2:1000,4:1000 + latency_filter: + - 0 + - 2 + - 4 + debug_subset: + enabled: false + max_episodes: 5 + suffix: debug +base_model: + repo_id: Qwen/Qwen3-VL-4B-Instruct +initialization: + checkpoint_local_dir: null + checkpoint_hf_repo_id: null + checkpoint_filename: null +trainer: + max_train_steps: 5000 + num_warmup_steps: 100 + save_interval: 5000 + eval_interval: 5000 + eval_num_batches: 200 + per_latency_eval_num_batches: 5 + eval_action_classification: true + eval_action_classification_interval: null + cc_f1_tolerance: 1 + learning_rate: + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + action_model: 0.0001 + lr_scheduler_type: cosine_with_min_lr + scheduler_specific_kwargs: + min_lr: 1.0e-06 + freeze_modules: '' + freeze_vit: false + freeze_tied_embedding: false + freeze_llm_layers: [] + loss_scale: + vla: 1.0 + vlm: 0.1 + max_grad_norm: 1.0 + weight_decay: 0.0 + logging_frequency: 1 + profile_timing: + enabled: false + log_interval: 10 + gradient_clipping: 1.0 + gradient_accumulation_steps: 2 + distributed_backend: none + is_resume: false + pretrained_checkpoint: null + resume_step: 0 + reload_modules: null + optimizer: + name: AdamW + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + weight_decay: 1.0e-08 + fused: true + save_format: pt +workspace_dir: /home/ubuntu/talha +run_root_dir: /home/ubuntu/talha/results/Checkpoints +seed: 42 +wandb_entity: talha1503 +wandb_project: latency-sensitive-bench +wandb_group: openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3 +wandb_tags: [] +training_latency_condition: unspecified +auth: + env_file: null + hf_token_env: HF_TOKEN + wandb_api_key_env: WANDB_API_KEY +paths: + run_root_dir: results/Checkpoints + dataset_local_dir: playground/Datasets/rl_games + dataset_cache_dir: null + base_model_dir: /home/ubuntu/talha/playground/Pretrained_models/Qwen3-VL-4B-Instruct + accelerate_config: starVLA/config/deepseeds/deepspeed_zero2.yaml +launch: + use_accelerate: true + gpus: null + num_processes: 1 + dry_run: false +conda: + enabled: true + env_name: null +rl_games: + model_alias: openvla + env_eval: + image_size: 224 + frameskip: 1 + image_transform: raw_rgb + prompt_mode: raw + ghost_trail: + history_frames: 5 + gamma: 1.3 + min_alpha: 35 + scroll_px_per_step: 4.0 + ground_fraction: 0.22 + seed: 42 + fixed_episode_seeds: true + latency_seed_stride: 0 + task_seed_stride: 0 + task_description: '' + eval_parallel_envs: 5 + action_chunk_execution: + enabled: false + chunk_size: null + enabled: true + gymnasium: + make_kwargs: {} + mid_train: + enabled: false + interval_steps: 100 + latencies: + - 0 + - 2 + - 4 + num_episodes: 5 + max_steps_per_episode: 3600 + post_train: + enabled: true + latencies: + - 0 + latencies_from_prompt_map: false + num_episodes: 20 + max_steps_per_episode: 3600 + eval_backend: latency_bench + distributed_mode: none + vectorized: + enabled: false + batch_size: 1 + latency: + prompt_map_path: /home/ubuntu/talha/playground/Datasets/rl_games/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3/latency_prompt_map.json + mode: mixed + values: + - 0 + task: gymnasium + gymnasium: + task_contract: + task_name: inverted_pendulum_rgb_state + env_id: LatencyBench/InvertedPendulumDiscrete-v0 + registration_imports: + - latency_bench.envs.gymnasium_mujoco + env_fps: 25.0 + obs_fps: 25.0 + frame_stack: 1 + base_prompt: 'Keep the pole upright by applying horizontal cart forces. Choose + exactly one action from: push_left_hard, push_left_soft, zero_force, push_right_soft, + push_right_hard.' + action_labels: + - push_left_hard + - push_left_soft + - zero_force + - push_right_soft + - push_right_hard + action_values: + - 0 + - 1 + - 2 + - 3 + - 4 + noop_action_id: 2 + make_kwargs: + base_env_id: InvertedPendulum-v4 + render_mode: rgb_array + force_values: + - -3.0 + - -1.5 + - 0.0 + - 1.5 + - 3.0 + state_space: + labels: + - cart_position + - pole_angle + - cart_velocity + - pole_angular_velocity + task_name: inverted_pendulum_rgb_state + initialization_mode: scratch + action_carrier: native +model: openvla +env: gymnasium +init: scratch +mode: mixed_latency +checkpoint: + load: none + hf_repo_id: talha15032/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3 + save_best_model: false + save_final_model: true + save_pt_file: true + save_training_state: false + save_safetensors_file: false + local: + keep_last_n: 1 + sync: + enabled: true + repo_id: talha15032/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3 + keep_last_n: 0 + sync_every_n_checkpoints: 1 + resume_policy: local_latest +run_id: openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3 +output_dir: /home/ubuntu/talha/results/Checkpoints/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3 +config_yaml: null +is_debug: false +version_id: '0.21' diff --git a/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3/config.yaml b/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3/config.yaml new file mode 100644 index 0000000000000000000000000000000000000000..c7f2fcba8c7fb6a520d10509f5e098d3cfb3d41a --- /dev/null +++ b/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3/config.yaml @@ -0,0 +1,93 @@ +checkpoint: + local: + keep_last_n: 1 + save_best_model: false + save_final_model: true + save_pt_file: true + save_safetensors_file: false + save_training_state: false + sync: + enabled: true + keep_last_n: 0 + repo_id: talha15032/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3 +datasets: + vla_data: + data_mix: openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3 + dataset_py: lerobot_datasets + eval_data_mix: openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3__val + latency_curriculum: + enabled: false + per_device_batch_size: 8 +framework: + action_model: + action_dim: 5 + action_env_dim: 5 + action_hidden_dim: 2560 + action_horizon: 1 + action_model_type: MLP + loss_type: discrete_ce + state_encoding: discretized_text + kv_memory: + enabled: false + packed_train: false + rebased_sink: true + rollout_len: 8 + window: 4 + name: QwenOFT + qwenvl: + attn_implementation: flash_attention_2 + base_vlm: /home/ubuntu/talha/playground/Pretrained_models/Qwen3-VL-4B-Instruct + enable_gradient_checkpointing: true +output_dir: /home/ubuntu/talha/results/Checkpoints/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3 +rl_games: + env_eval: + enabled: true + eval_backend: latency_bench + mid_train: + enabled: false + interval_steps: 100 + prompt_mode: raw + task: gymnasium +run_id: openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3 +run_root_dir: /home/ubuntu/talha/results/Checkpoints +seed: 42 +trainer: + distributed_backend: none + eval_action_classification: true + eval_action_classification_interval: null + eval_interval: 5000 + eval_num_batches: 200 + freeze_llm_layers: [] + freeze_modules: '' + freeze_tied_embedding: false + freeze_vit: false + gradient_accumulation_steps: 2 + gradient_clipping: 1.0 + is_resume: false + learning_rate: + action_model: 0.0001 + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + logging_frequency: 1 + lr_scheduler_type: cosine_with_min_lr + max_train_steps: 5000 + num_warmup_steps: 100 + optimizer: + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + fused: true + weight_decay: 1.0e-08 + per_latency_eval_num_batches: 5 + pretrained_checkpoint: null + profile_timing: + enabled: false + save_interval: 5000 + scheduler_specific_kwargs: + min_lr: 1.0e-06 +training_latency_condition: unspecified +wandb_entity: talha1503 +wandb_group: openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3 +wandb_project: latency-sensitive-bench +wandb_tags: [] diff --git a/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3/dataset_statistics.json b/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3/dataset_statistics.json new file mode 100644 index 0000000000000000000000000000000000000000..48310ea5d5254896f48d5e8d123c1528088202a0 --- /dev/null +++ b/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3/dataset_statistics.json @@ -0,0 +1,95 @@ +{ + "new_embodiment": { + "action": { + "mean": [ + 0.027996724471449852, + 0.1508392095565796, + 0.6481470465660095, + 0.13600370287895203, + 0.03701331838965416 + ], + "std": [ + 0.16490621864795685, + 0.3581511974334717, + 0.4778982102870941, + 0.3424498736858368, + 0.18899837136268616 + ], + "max": [ + 1.0, + 1.0, + 1.0, + 1.0, + 1.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 1.0, + 1.0, + 1.0, + 1.0, + 1.0 + ], + "mask": [ + true, + true, + true, + true, + true + ] + }, + "state": { + "mean": [ + 0.143598273396492, + 0.004461213015019894, + 0.11432920396327972, + -0.006049089599400759 + ], + "std": [ + 0.31354930996894836, + 0.06130174919962883, + 0.4503016769886017, + 0.8575727343559265 + ], + "max": [ + 1.031674861907959, + 0.19994768500328064, + 2.3079874515533447, + 3.7387092113494873 + ], + "min": [ + -1.0259008407592773, + -0.1999799758195877, + -2.2849948406219482, + -4.038050174713135 + ], + "q01": [ + -0.3022584795951843, + -0.15527626156806945, + -0.9133486723899841, + -2.359431056976318 + ], + "q99": [ + 1.0026830291748048, + 0.16135655224323286, + 1.3828536510467528, + 2.107583684921267 + ] + }, + "num_transitions": 218597, + "num_trajectories": 2070 + } +} \ No newline at end of file diff --git a/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3/dataset_statistics_eval.json b/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3/dataset_statistics_eval.json new file mode 100644 index 0000000000000000000000000000000000000000..d8b9a06ce9d7f472c01600f4e742aa2411c89047 --- /dev/null +++ b/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3/dataset_statistics_eval.json @@ -0,0 +1,95 @@ +{ + "new_embodiment": { + "action": { + "mean": [ + 0.035599857568740845, + 0.1577073633670807, + 0.6541473865509033, + 0.07760769128799438, + 0.07493770122528076 + ], + "std": [ + 0.1853053718805313, + 0.3644678294658661, + 0.4756529927253723, + 0.26755642890930176, + 0.26328903436660767 + ], + "max": [ + 1.0, + 1.0, + 1.0, + 1.0, + 1.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 1.0, + 1.0, + 1.0, + 1.0, + 1.0 + ], + "mask": [ + true, + true, + true, + true, + true + ] + }, + "state": { + "mean": [ + -0.006056365091353655, + -0.001078602741472423, + 0.021814774721860886, + -0.004301219247281551 + ], + "std": [ + 0.15048037469387054, + 0.0324825718998909, + 0.33326858282089233, + 0.713808000087738 + ], + "max": [ + 1.0273027420043945, + 0.198622465133667, + 2.3032615184783936, + 3.4846131801605225 + ], + "min": [ + -0.41009554266929626, + -0.19869747757911682, + -1.3806304931640625, + -3.9696083068847656 + ], + "q01": [ + -0.08295003287494182, + -0.1170046754181385, + -0.977174437046051, + -2.3071367263793947 + ], + "q99": [ + 0.9016301155090326, + 0.11161688491702067, + 1.2677143096923813, + 2.1670001268386816 + ] + }, + "num_transitions": 11236, + "num_trajectories": 30 + } +} \ No newline at end of file diff --git a/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3/hydra/.hydra/config.yaml b/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3/hydra/.hydra/config.yaml new file mode 100644 index 0000000000000000000000000000000000000000..c25d06505474ed729291d209569cfa9f9b77906a --- /dev/null +++ b/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3/hydra/.hydra/config.yaml @@ -0,0 +1,338 @@ +framework: + qwenvl: + base_vlm: /home/ubuntu/talha/playground/Pretrained_models/Qwen3-VL-4B-Instruct + attn_implementation: flash_attention_2 + enable_gradient_checkpointing: true + action_model: + state_dim: 4 + loss_type: discrete_ce + action_horizon: 1 + future_action_window_size: 0 + past_action_window_size: 0 + action_dim: 5 + action_env_dim: 5 + state_encoding: discretized_text + action_model_type: MLP + action_hidden_dim: 2560 + kv_memory: + enabled: false + window: 4 + rollout_len: 8 + packed_train: false + rebased_sink: true + name: QwenOFT +datasets: + vla_data: + dataset_py: lerobot_datasets + include_state: true + data_root_dir: /home/ubuntu/talha/playground/Datasets/rl_games + data_mix: openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3 + eval_data_mix: openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3__val + custom_mixtures_path: /home/ubuntu/talha/playground/Datasets/rl_games/_generated_mixtures/single_79881079067b9b1b.json + action_type: discrete + sequential_step_sampling: true + eval_sequential_step_sampling: null + num_workers: 8 + eval_num_workers: 8 + prefetch_factor: 4 + persistent_workers: true + pin_memory: true + shuffle: true + action_balance: + enabled: false + strategy: balanced_epoch + action_key: action_id + target_flap_fraction: 0.3 + noop_id: 0 + flap_id: 1 + latency_curriculum: + enabled: false + strategy: exclusive + latencies: null + phase_steps: null + phase_distributions: null + new_latency_passes: 1.0 + replay_passes: 0.25 + target_total_passes: 2.0 + final_equalization: true + step_budget_mode: auto + eval_at_phase_end: false + save_at_phase_end: false + computed_plan: null + per_device_batch_size: 8 + load_all_data_for_training: true + num_obs_frames: 1 + image_mode: single + prompt_mode: raw + stitch_grid: + - 2 + - 2 + obs_image_size: null + video_backend: torchvision_av + gymnasium_task_contract: + task_name: inverted_pendulum_rgb_state + env_id: LatencyBench/InvertedPendulumDiscrete-v0 + registration_imports: + - latency_bench.envs.gymnasium_mujoco + env_fps: 25.0 + obs_fps: 25.0 + frame_stack: 1 + base_prompt: 'Keep the pole upright by applying horizontal cart forces. Choose + exactly one action from: push_left_hard, push_left_soft, zero_force, push_right_soft, + push_right_hard.' + action_labels: + - push_left_hard + - push_left_soft + - zero_force + - push_right_soft + - push_right_hard + action_values: + - 0 + - 1 + - 2 + - 3 + - 4 + noop_action_id: 2 + make_kwargs: + base_env_id: InvertedPendulum-v4 + render_mode: rgb_array + force_values: + - -3.0 + - -1.5 + - 0.0 + - 1.5 + - 3.0 + base_make_kwargs: {} + state_space: + labels: + - cart_position + - pole_angle + - cart_velocity + - pole_angular_velocity + active_action_dim: 5 +dataset: + source_hf: latency-sensitive-bench/inverted_pendulum_200ep_2 + config_name: null + source_subdir: inverted_pendulum_fixed_latency_0_100ep_1ksteps_og,inverted_pendulum_fixed_latency_2_1000ep_7k2steps,inverted_pendulum_fixed_latency_4_1000ep_7k2steps + converted_name: openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3 + single_source_hf: '' + mixed_source_hf: '' + single_converted_name: flappy_train + mixed_converted_name: flappy_train + single_latency_filter: null + mixed_latency_filter: null + force_download: false + setup_force: true + skip_verification: true + target_latency_unit: observation_steps + verify_rows: 200 + max_episodes: null + max_steps_per_episode: 1000 + episodes_per_latency: null + episodes_per_latency_by_latency: 0:90,2:1000,4:1000 + latency_filter: + - 0 + - 2 + - 4 + debug_subset: + enabled: false + max_episodes: 5 + suffix: debug +base_model: + repo_id: Qwen/Qwen3-VL-4B-Instruct +initialization: + checkpoint_local_dir: null + checkpoint_hf_repo_id: null + checkpoint_filename: null +trainer: + max_train_steps: 5000 + num_warmup_steps: 100 + save_interval: 5000 + eval_interval: 5000 + eval_num_batches: 200 + per_latency_eval_num_batches: 5 + eval_action_classification: true + eval_action_classification_interval: null + cc_f1_tolerance: 1 + learning_rate: + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + action_model: 0.0001 + lr_scheduler_type: cosine_with_min_lr + scheduler_specific_kwargs: + min_lr: 1.0e-06 + freeze_modules: '' + freeze_vit: false + freeze_tied_embedding: false + freeze_llm_layers: [] + loss_scale: + vla: 1.0 + vlm: 0.1 + max_grad_norm: 1.0 + weight_decay: 0.0 + logging_frequency: 1 + profile_timing: + enabled: false + log_interval: 10 + gradient_clipping: 1.0 + gradient_accumulation_steps: 2 + distributed_backend: none + is_resume: false + pretrained_checkpoint: null + resume_step: 0 + reload_modules: null + optimizer: + name: AdamW + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + weight_decay: 1.0e-08 + fused: true + save_format: pt +workspace_dir: /home/ubuntu/talha +run_root_dir: /home/ubuntu/talha/results/Checkpoints +seed: 42 +wandb_entity: ${oc.env:WANDB_ENTITY} +wandb_project: ${oc.env:WANDB_PROJECT,starVLA_rl_games} +wandb_group: openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3 +wandb_tags: [] +training_latency_condition: unspecified +auth: + env_file: null + hf_token_env: HF_TOKEN + wandb_api_key_env: WANDB_API_KEY +paths: + run_root_dir: results/Checkpoints + dataset_local_dir: playground/Datasets/rl_games + dataset_cache_dir: null + base_model_dir: /home/ubuntu/talha/playground/Pretrained_models/Qwen3-VL-4B-Instruct + accelerate_config: starVLA/config/deepseeds/deepspeed_zero2.yaml +launch: + use_accelerate: true + gpus: null + num_processes: 1 + dry_run: false +conda: + enabled: true + env_name: null +rl_games: + model_alias: openvla + env_eval: + image_size: 224 + frameskip: 1 + image_transform: raw_rgb + prompt_mode: raw + ghost_trail: + history_frames: 5 + gamma: 1.3 + min_alpha: 35 + scroll_px_per_step: 4.0 + ground_fraction: 0.22 + seed: 42 + fixed_episode_seeds: true + latency_seed_stride: 0 + task_seed_stride: 0 + task_description: '' + eval_parallel_envs: 5 + action_chunk_execution: + enabled: false + chunk_size: null + enabled: true + gymnasium: + make_kwargs: {} + mid_train: + enabled: false + interval_steps: 100 + latencies: + - 0 + - 2 + - 4 + num_episodes: 5 + max_steps_per_episode: 3600 + post_train: + enabled: true + latencies: + - 0 + latencies_from_prompt_map: false + num_episodes: 20 + max_steps_per_episode: 3600 + eval_backend: latency_bench + distributed_mode: none + vectorized: + enabled: false + batch_size: 1 + latency: + prompt_map_path: /home/ubuntu/talha/playground/Datasets/rl_games/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3/latency_prompt_map.json + mode: mixed + values: + - 0 + task: gymnasium + gymnasium: + task_contract: + task_name: inverted_pendulum_rgb_state + env_id: LatencyBench/InvertedPendulumDiscrete-v0 + registration_imports: + - latency_bench.envs.gymnasium_mujoco + env_fps: 25.0 + obs_fps: 25.0 + frame_stack: 1 + base_prompt: 'Keep the pole upright by applying horizontal cart forces. Choose + exactly one action from: push_left_hard, push_left_soft, zero_force, push_right_soft, + push_right_hard.' + action_labels: + - push_left_hard + - push_left_soft + - zero_force + - push_right_soft + - push_right_hard + action_values: + - 0 + - 1 + - 2 + - 3 + - 4 + noop_action_id: 2 + make_kwargs: + base_env_id: InvertedPendulum-v4 + render_mode: rgb_array + force_values: + - -3.0 + - -1.5 + - 0.0 + - 1.5 + - 3.0 + state_space: + labels: + - cart_position + - pole_angle + - cart_velocity + - pole_angular_velocity + task_name: inverted_pendulum_rgb_state + initialization_mode: scratch + action_carrier: native +model: openvla +env: gymnasium +init: scratch +mode: mixed_latency +checkpoint: + load: none + hf_repo_id: talha15032/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3 + save_best_model: false + save_final_model: true + save_pt_file: true + save_training_state: false + save_safetensors_file: false + local: + keep_last_n: 1 + sync: + enabled: true + repo_id: talha15032/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3 + keep_last_n: 0 + sync_every_n_checkpoints: 1 + resume_policy: local_latest +run_id: openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3 +output_dir: null +config_yaml: null +is_debug: false +version_id: 0.21 diff --git a/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3/hydra/.hydra/hydra.yaml b/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3/hydra/.hydra/hydra.yaml new file mode 100644 index 0000000000000000000000000000000000000000..43cef95d5d9551dcc821de3697684483dc678d86 --- /dev/null +++ b/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3/hydra/.hydra/hydra.yaml @@ -0,0 +1,394 @@ +hydra: + run: + dir: ${run_root_dir}/${run_id}/hydra + sweep: + dir: multirun/${now:%Y-%m-%d}/${now:%H-%M-%S} + subdir: ${hydra.job.num} + launcher: + _target_: hydra._internal.core_plugins.basic_launcher.BasicLauncher + sweeper: + _target_: hydra._internal.core_plugins.basic_sweeper.BasicSweeper + max_batch_size: null + params: null + help: + app_name: ${hydra.job.name} + header: '${hydra.help.app_name} is powered by Hydra. + + ' + footer: 'Powered by Hydra (https://hydra.cc) + + Use --hydra-help to view Hydra specific help + + ' + template: '${hydra.help.header} + + == Configuration groups == + + Compose your configuration from those groups (group=option) + + + $APP_CONFIG_GROUPS + + + == Config == + + Override anything in the config (foo.bar=value) + + + $CONFIG + + + ${hydra.help.footer} + + ' + hydra_help: + template: 'Hydra (${hydra.runtime.version}) + + See https://hydra.cc for more info. + + + == Flags == + + $FLAGS_HELP + + + == Configuration groups == + + Compose your configuration from those groups (For example, append hydra/job_logging=disabled + to command line) + + + $HYDRA_CONFIG_GROUPS + + + Use ''--cfg hydra'' to Show the Hydra config. + + ' + hydra_help: ??? + hydra_logging: + version: 1 + formatters: + simple: + format: '[%(asctime)s][HYDRA] %(message)s' + handlers: + console: + class: logging.StreamHandler + formatter: simple + stream: ext://sys.stdout + root: + level: INFO + handlers: + - console + loggers: + logging_example: + level: DEBUG + disable_existing_loggers: false + job_logging: + version: 1 + formatters: + simple: + format: '[%(asctime)s][%(name)s][%(levelname)s] - %(message)s' + handlers: + console: + class: logging.StreamHandler + formatter: simple + stream: ext://sys.stdout + file: + class: logging.FileHandler + formatter: simple + filename: ${hydra.runtime.output_dir}/${hydra.job.name}.log + root: + level: INFO + handlers: + - console + - file + disable_existing_loggers: false + env: {} + mode: RUN + searchpath: [] + callbacks: {} + output_subdir: .hydra + overrides: + hydra: + - hydra.mode=RUN + task: + - model=openvla + - env=gymnasium + - init=scratch + - mode=mixed_latency + - ++framework.qwenvl.base_vlm=/home/ubuntu/talha/playground/Pretrained_models/Qwen3-VL-4B-Instruct + - ++framework.qwenvl.attn_implementation=flash_attention_2 + - ++framework.qwenvl.enable_gradient_checkpointing=true + - ++framework.action_model.state_dim=4 + - ++framework.action_model.loss_type=discrete_ce + - ++framework.action_model.action_horizon=1 + - ++framework.action_model.future_action_window_size=0 + - ++framework.action_model.past_action_window_size=0 + - ++framework.action_model.action_dim=5 + - ++framework.action_model.action_env_dim=5 + - ++framework.action_model.state_encoding=discretized_text + - ++framework.action_model.action_model_type=MLP + - ++framework.action_model.action_hidden_dim=2560 + - ++framework.kv_memory.enabled=false + - ++framework.kv_memory.window=4 + - ++framework.kv_memory.rollout_len=8 + - ++framework.kv_memory.packed_train=false + - ++framework.kv_memory.rebased_sink=true + - ++framework.name=QwenOFT + - ++datasets.vla_data.dataset_py=lerobot_datasets + - ++datasets.vla_data.include_state=true + - ++datasets.vla_data.data_root_dir=playground/Datasets/rl_games + - ++datasets.vla_data.data_mix=openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3 + - ++datasets.vla_data.eval_data_mix=openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3__val + - ++datasets.vla_data.custom_mixtures_path=null + - ++datasets.vla_data.action_type=discrete + - ++datasets.vla_data.sequential_step_sampling=true + - ++datasets.vla_data.eval_sequential_step_sampling=null + - ++datasets.vla_data.num_workers=8 + - ++datasets.vla_data.eval_num_workers=8 + - ++datasets.vla_data.prefetch_factor=4 + - ++datasets.vla_data.persistent_workers=true + - ++datasets.vla_data.pin_memory=true + - ++datasets.vla_data.shuffle=true + - ++datasets.vla_data.action_balance.enabled=false + - ++datasets.vla_data.action_balance.strategy=balanced_epoch + - ++datasets.vla_data.action_balance.action_key=action_id + - ++datasets.vla_data.action_balance.target_flap_fraction=0.3 + - ++datasets.vla_data.action_balance.noop_id=0 + - ++datasets.vla_data.action_balance.flap_id=1 + - ++datasets.vla_data.latency_curriculum.enabled=false + - ++datasets.vla_data.latency_curriculum.strategy=exclusive + - ++datasets.vla_data.latency_curriculum.latencies=null + - ++datasets.vla_data.latency_curriculum.phase_steps=null + - ++datasets.vla_data.latency_curriculum.phase_distributions=null + - ++datasets.vla_data.latency_curriculum.new_latency_passes=1.0 + - ++datasets.vla_data.latency_curriculum.replay_passes=0.25 + - ++datasets.vla_data.latency_curriculum.target_total_passes=2.0 + - ++datasets.vla_data.latency_curriculum.final_equalization=true + - ++datasets.vla_data.latency_curriculum.step_budget_mode=auto + - ++datasets.vla_data.latency_curriculum.eval_at_phase_end=false + - ++datasets.vla_data.latency_curriculum.save_at_phase_end=false + - ++datasets.vla_data.latency_curriculum.computed_plan=null + - ++datasets.vla_data.per_device_batch_size=8 + - ++datasets.vla_data.load_all_data_for_training=true + - ++datasets.vla_data.num_obs_frames=1 + - ++datasets.vla_data.image_mode=single + - ++datasets.vla_data.prompt_mode=raw + - ++datasets.vla_data.stitch_grid=[2,2] + - ++datasets.vla_data.obs_image_size=null + - ++datasets.vla_data.video_backend=torchvision_av + - '++datasets.vla_data.gymnasium_task_contract={task_name:inverted_pendulum_rgb_state,env_id:LatencyBench/InvertedPendulumDiscrete-v0,registration_imports:[latency_bench.envs.gymnasium_mujoco],env_fps:25.0,obs_fps:25.0,frame_stack:1,base_prompt:''Keep + the pole upright by applying horizontal cart forces. Choose exactly one action + from: push_left_hard, push_left_soft, zero_force, push_right_soft, push_right_hard.'',action_labels:[push_left_hard,push_left_soft,zero_force,push_right_soft,push_right_hard],action_values:[0,1,2,3,4],noop_action_id:2,make_kwargs:{base_env_id:InvertedPendulum-v4,render_mode:rgb_array,force_values:[-3.0,-1.5,0.0,1.5,3.0],base_make_kwargs:{}},state_space:{labels:[cart_position,pole_angle,cart_velocity,pole_angular_velocity]}}' + - ++datasets.vla_data.active_action_dim=5 + - ++dataset.source_hf=latency-sensitive-bench/inverted_pendulum_200ep_2 + - ++dataset.config_name=null + - ++dataset.source_subdir='inverted_pendulum_fixed_latency_0_100ep_1ksteps_og,inverted_pendulum_fixed_latency_2_1000ep_7k2steps,inverted_pendulum_fixed_latency_4_1000ep_7k2steps' + - ++dataset.converted_name=openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3 + - ++dataset.single_source_hf= + - ++dataset.mixed_source_hf= + - ++dataset.single_converted_name=flappy_train + - ++dataset.mixed_converted_name=flappy_train + - ++dataset.single_latency_filter=null + - ++dataset.mixed_latency_filter=null + - ++dataset.force_download=false + - ++dataset.setup_force=true + - ++dataset.skip_verification=true + - ++dataset.target_latency_unit=observation_steps + - ++dataset.verify_rows=200 + - ++dataset.max_episodes=null + - ++dataset.max_steps_per_episode=1000 + - ++dataset.episodes_per_latency=null + - ++dataset.episodes_per_latency_by_latency='0:90,2:1000,4:1000' + - ++dataset.latency_filter=[0,2,4] + - ++dataset.debug_subset.enabled=false + - ++dataset.debug_subset.max_episodes=5 + - ++dataset.debug_subset.suffix=debug + - ++base_model.repo_id=Qwen/Qwen3-VL-4B-Instruct + - ++initialization.checkpoint_local_dir=null + - ++initialization.checkpoint_hf_repo_id=null + - ++initialization.checkpoint_filename=null + - ++trainer.max_train_steps=5000 + - ++trainer.num_warmup_steps=100 + - ++trainer.save_interval=5000 + - ++trainer.eval_interval=5000 + - ++trainer.eval_num_batches=200 + - ++trainer.per_latency_eval_num_batches=5 + - ++trainer.eval_action_classification=true + - ++trainer.eval_action_classification_interval=null + - ++trainer.cc_f1_tolerance=1 + - ++trainer.learning_rate.base=2e-05 + - ++trainer.learning_rate.qwen_vl_interface=1e-05 + - ++trainer.learning_rate.action_model=0.0001 + - ++trainer.lr_scheduler_type=cosine_with_min_lr + - ++trainer.scheduler_specific_kwargs.min_lr=1e-06 + - ++trainer.freeze_modules= + - ++trainer.freeze_vit=false + - ++trainer.freeze_tied_embedding=false + - ++trainer.freeze_llm_layers=[] + - ++trainer.loss_scale.vla=1.0 + - ++trainer.loss_scale.vlm=0.1 + - ++trainer.max_grad_norm=1.0 + - ++trainer.weight_decay=0.0 + - ++trainer.logging_frequency=1 + - ++trainer.profile_timing.enabled=false + - ++trainer.profile_timing.log_interval=10 + - ++trainer.gradient_clipping=1.0 + - ++trainer.gradient_accumulation_steps=2 + - ++trainer.distributed_backend=none + - ++trainer.is_resume=false + - ++trainer.pretrained_checkpoint=null + - ++trainer.resume_step=0 + - ++trainer.reload_modules=null + - ++trainer.optimizer.name=AdamW + - ++trainer.optimizer.betas=[0.9,0.95] + - ++trainer.optimizer.eps=1e-08 + - ++trainer.optimizer.weight_decay=1e-08 + - ++trainer.optimizer.fused=true + - ++trainer.save_format=pt + - ++workspace_dir=/home/ubuntu/talha + - ++run_root_dir=results/Checkpoints + - ++seed=42 + - ++wandb_group=openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3 + - ++wandb_tags=[] + - ++training_latency_condition=unspecified + - ++auth.env_file=null + - ++auth.hf_token_env=HF_TOKEN + - ++auth.wandb_api_key_env=WANDB_API_KEY + - ++paths.run_root_dir=results/Checkpoints + - ++paths.dataset_local_dir=playground/Datasets/rl_games + - ++paths.dataset_cache_dir=null + - ++paths.base_model_dir=/home/ubuntu/talha/playground/Pretrained_models/Qwen3-VL-4B-Instruct + - ++paths.accelerate_config=starVLA/config/deepseeds/deepspeed_zero2.yaml + - ++rl_games.model_alias=openvla + - ++rl_games.env_eval.image_size=224 + - ++rl_games.env_eval.frameskip=1 + - ++rl_games.env_eval.image_transform=raw_rgb + - ++rl_games.env_eval.prompt_mode=raw + - ++rl_games.env_eval.ghost_trail.history_frames=5 + - ++rl_games.env_eval.ghost_trail.gamma=1.3 + - ++rl_games.env_eval.ghost_trail.min_alpha=35 + - ++rl_games.env_eval.ghost_trail.scroll_px_per_step=4.0 + - ++rl_games.env_eval.ghost_trail.ground_fraction=0.22 + - ++rl_games.env_eval.seed=42 + - ++rl_games.env_eval.fixed_episode_seeds=true + - ++rl_games.env_eval.latency_seed_stride=0 + - ++rl_games.env_eval.task_seed_stride=0 + - ++rl_games.env_eval.task_description= + - ++rl_games.env_eval.eval_parallel_envs=5 + - ++rl_games.env_eval.action_chunk_execution.enabled=false + - ++rl_games.env_eval.action_chunk_execution.chunk_size=null + - ++rl_games.env_eval.enabled=true + - ++rl_games.env_eval.mid_train.enabled=false + - ++rl_games.env_eval.mid_train.interval_steps=100 + - ++rl_games.env_eval.mid_train.latencies=[0,2,4] + - ++rl_games.env_eval.mid_train.num_episodes=5 + - ++rl_games.env_eval.mid_train.max_steps_per_episode=3600 + - ++rl_games.env_eval.post_train.enabled=true + - ++rl_games.env_eval.post_train.latencies=[0] + - ++rl_games.env_eval.post_train.latencies_from_prompt_map=false + - ++rl_games.env_eval.post_train.num_episodes=20 + - ++rl_games.env_eval.post_train.max_steps_per_episode=3600 + - ++rl_games.env_eval.eval_backend=latency_bench + - ++rl_games.env_eval.distributed_mode=none + - ++rl_games.env_eval.vectorized.enabled=false + - ++rl_games.env_eval.vectorized.batch_size=1 + - ++rl_games.env_eval.latency.prompt_map_path=null + - ++rl_games.env_eval.latency.mode=mixed + - ++rl_games.env_eval.latency.values=[0] + - ++rl_games.task=gymnasium + - ++rl_games.gymnasium.task_contract.task_name=inverted_pendulum_rgb_state + - ++rl_games.gymnasium.task_contract.env_id=LatencyBench/InvertedPendulumDiscrete-v0 + - ++rl_games.gymnasium.task_contract.registration_imports=[latency_bench.envs.gymnasium_mujoco] + - ++rl_games.gymnasium.task_contract.env_fps=25.0 + - ++rl_games.gymnasium.task_contract.obs_fps=25.0 + - ++rl_games.gymnasium.task_contract.frame_stack=1 + - '++rl_games.gymnasium.task_contract.base_prompt=''Keep the pole upright by applying + horizontal cart forces. Choose exactly one action from: push_left_hard, push_left_soft, + zero_force, push_right_soft, push_right_hard.''' + - ++rl_games.gymnasium.task_contract.action_labels=[push_left_hard,push_left_soft,zero_force,push_right_soft,push_right_hard] + - ++rl_games.gymnasium.task_contract.action_values=[0,1,2,3,4] + - ++rl_games.gymnasium.task_contract.noop_action_id=2 + - ++rl_games.gymnasium.task_contract.make_kwargs.base_env_id=InvertedPendulum-v4 + - ++rl_games.gymnasium.task_contract.make_kwargs.render_mode=rgb_array + - ++rl_games.gymnasium.task_contract.make_kwargs.force_values=[-3.0,-1.5,0.0,1.5,3.0] + - ++rl_games.gymnasium.task_contract.state_space.labels=[cart_position,pole_angle,cart_velocity,pole_angular_velocity] + - ++rl_games.gymnasium.task_name=inverted_pendulum_rgb_state + - ++rl_games.initialization_mode=scratch + - ++rl_games.action_carrier=native + - ++checkpoint.load=none + - ++checkpoint.hf_repo_id=talha15032/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3 + - ++checkpoint.save_best_model=false + - ++checkpoint.save_final_model=true + - ++checkpoint.save_pt_file=true + - ++checkpoint.save_training_state=false + - ++checkpoint.save_safetensors_file=false + - ++checkpoint.local.keep_last_n=1 + - ++checkpoint.sync.enabled=true + - ++checkpoint.sync.repo_id=talha15032/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3 + - ++checkpoint.sync.keep_last_n=0 + - ++checkpoint.sync.sync_every_n_checkpoints=1 + - ++checkpoint.sync.resume_policy=local_latest + - ++run_id=openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3 + - ++output_dir=null + - ++config_yaml=null + - ++is_debug=false + - ++version_id=0.21 + - ++run_root_dir=/home/ubuntu/talha/results/Checkpoints + - ++trainer.is_resume=false + - ++datasets.vla_data.data_root_dir=/home/ubuntu/talha/playground/Datasets/rl_games + - ++datasets.vla_data.data_mix=openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3 + - ++datasets.vla_data.eval_data_mix=openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3__val + - ++datasets.vla_data.custom_mixtures_path=/home/ubuntu/talha/playground/Datasets/rl_games/_generated_mixtures/single_79881079067b9b1b.json + - ++framework.qwenvl.base_vlm=/home/ubuntu/talha/playground/Pretrained_models/Qwen3-VL-4B-Instruct + - ++rl_games.env_eval.latency.prompt_map_path=/home/ubuntu/talha/playground/Datasets/rl_games/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3/latency_prompt_map.json + job: + name: train_starvla_hydra + chdir: false + override_dirname: '++auth.env_file=null,++auth.hf_token_env=HF_TOKEN,++auth.wandb_api_key_env=WANDB_API_KEY,++base_model.repo_id=Qwen/Qwen3-VL-4B-Instruct,++checkpoint.hf_repo_id=talha15032/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3,++checkpoint.load=none,++checkpoint.local.keep_last_n=1,++checkpoint.save_best_model=false,++checkpoint.save_final_model=true,++checkpoint.save_pt_file=true,++checkpoint.save_safetensors_file=false,++checkpoint.save_training_state=false,++checkpoint.sync.enabled=true,++checkpoint.sync.keep_last_n=0,++checkpoint.sync.repo_id=talha15032/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3,++checkpoint.sync.resume_policy=local_latest,++checkpoint.sync.sync_every_n_checkpoints=1,++config_yaml=null,++dataset.config_name=null,++dataset.converted_name=openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3,++dataset.debug_subset.enabled=false,++dataset.debug_subset.max_episodes=5,++dataset.debug_subset.suffix=debug,++dataset.episodes_per_latency=null,++dataset.episodes_per_latency_by_latency=''0:90,2:1000,4:1000'',++dataset.force_download=false,++dataset.latency_filter=[0,2,4],++dataset.max_episodes=null,++dataset.max_steps_per_episode=1000,++dataset.mixed_converted_name=flappy_train,++dataset.mixed_latency_filter=null,++dataset.mixed_source_hf=,++dataset.setup_force=true,++dataset.single_converted_name=flappy_train,++dataset.single_latency_filter=null,++dataset.single_source_hf=,++dataset.skip_verification=true,++dataset.source_hf=latency-sensitive-bench/inverted_pendulum_200ep_2,++dataset.source_subdir=''inverted_pendulum_fixed_latency_0_100ep_1ksteps_og,inverted_pendulum_fixed_latency_2_1000ep_7k2steps,inverted_pendulum_fixed_latency_4_1000ep_7k2steps'',++dataset.target_latency_unit=observation_steps,++dataset.verify_rows=200,++datasets.vla_data.action_balance.action_key=action_id,++datasets.vla_data.action_balance.enabled=false,++datasets.vla_data.action_balance.flap_id=1,++datasets.vla_data.action_balance.noop_id=0,++datasets.vla_data.action_balance.strategy=balanced_epoch,++datasets.vla_data.action_balance.target_flap_fraction=0.3,++datasets.vla_data.action_type=discrete,++datasets.vla_data.active_action_dim=5,++datasets.vla_data.custom_mixtures_path=/home/ubuntu/talha/playground/Datasets/rl_games/_generated_mixtures/single_79881079067b9b1b.json,++datasets.vla_data.custom_mixtures_path=null,++datasets.vla_data.data_mix=openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3,++datasets.vla_data.data_mix=openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3,++datasets.vla_data.data_root_dir=/home/ubuntu/talha/playground/Datasets/rl_games,++datasets.vla_data.data_root_dir=playground/Datasets/rl_games,++datasets.vla_data.dataset_py=lerobot_datasets,++datasets.vla_data.eval_data_mix=openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3__val,++datasets.vla_data.eval_data_mix=openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3__val,++datasets.vla_data.eval_num_workers=8,++datasets.vla_data.eval_sequential_step_sampling=null,++datasets.vla_data.gymnasium_task_contract={task_name:inverted_pendulum_rgb_state,env_id:LatencyBench/InvertedPendulumDiscrete-v0,registration_imports:[latency_bench.envs.gymnasium_mujoco],env_fps:25.0,obs_fps:25.0,frame_stack:1,base_prompt:''Keep + the pole upright by applying horizontal cart forces. Choose exactly one action + from: push_left_hard, push_left_soft, zero_force, push_right_soft, push_right_hard.'',action_labels:[push_left_hard,push_left_soft,zero_force,push_right_soft,push_right_hard],action_values:[0,1,2,3,4],noop_action_id:2,make_kwargs:{base_env_id:InvertedPendulum-v4,render_mode:rgb_array,force_values:[-3.0,-1.5,0.0,1.5,3.0],base_make_kwargs:{}},state_space:{labels:[cart_position,pole_angle,cart_velocity,pole_angular_velocity]}},++datasets.vla_data.image_mode=single,++datasets.vla_data.include_state=true,++datasets.vla_data.latency_curriculum.computed_plan=null,++datasets.vla_data.latency_curriculum.enabled=false,++datasets.vla_data.latency_curriculum.eval_at_phase_end=false,++datasets.vla_data.latency_curriculum.final_equalization=true,++datasets.vla_data.latency_curriculum.latencies=null,++datasets.vla_data.latency_curriculum.new_latency_passes=1.0,++datasets.vla_data.latency_curriculum.phase_distributions=null,++datasets.vla_data.latency_curriculum.phase_steps=null,++datasets.vla_data.latency_curriculum.replay_passes=0.25,++datasets.vla_data.latency_curriculum.save_at_phase_end=false,++datasets.vla_data.latency_curriculum.step_budget_mode=auto,++datasets.vla_data.latency_curriculum.strategy=exclusive,++datasets.vla_data.latency_curriculum.target_total_passes=2.0,++datasets.vla_data.load_all_data_for_training=true,++datasets.vla_data.num_obs_frames=1,++datasets.vla_data.num_workers=8,++datasets.vla_data.obs_image_size=null,++datasets.vla_data.per_device_batch_size=8,++datasets.vla_data.persistent_workers=true,++datasets.vla_data.pin_memory=true,++datasets.vla_data.prefetch_factor=4,++datasets.vla_data.prompt_mode=raw,++datasets.vla_data.sequential_step_sampling=true,++datasets.vla_data.shuffle=true,++datasets.vla_data.stitch_grid=[2,2],++datasets.vla_data.video_backend=torchvision_av,++framework.action_model.action_dim=5,++framework.action_model.action_env_dim=5,++framework.action_model.action_hidden_dim=2560,++framework.action_model.action_horizon=1,++framework.action_model.action_model_type=MLP,++framework.action_model.future_action_window_size=0,++framework.action_model.loss_type=discrete_ce,++framework.action_model.past_action_window_size=0,++framework.action_model.state_dim=4,++framework.action_model.state_encoding=discretized_text,++framework.kv_memory.enabled=false,++framework.kv_memory.packed_train=false,++framework.kv_memory.rebased_sink=true,++framework.kv_memory.rollout_len=8,++framework.kv_memory.window=4,++framework.name=QwenOFT,++framework.qwenvl.attn_implementation=flash_attention_2,++framework.qwenvl.base_vlm=/home/ubuntu/talha/playground/Pretrained_models/Qwen3-VL-4B-Instruct,++framework.qwenvl.base_vlm=/home/ubuntu/talha/playground/Pretrained_models/Qwen3-VL-4B-Instruct,++framework.qwenvl.enable_gradient_checkpointing=true,++initialization.checkpoint_filename=null,++initialization.checkpoint_hf_repo_id=null,++initialization.checkpoint_local_dir=null,++is_debug=false,++output_dir=null,++paths.accelerate_config=starVLA/config/deepseeds/deepspeed_zero2.yaml,++paths.base_model_dir=/home/ubuntu/talha/playground/Pretrained_models/Qwen3-VL-4B-Instruct,++paths.dataset_cache_dir=null,++paths.dataset_local_dir=playground/Datasets/rl_games,++paths.run_root_dir=results/Checkpoints,++rl_games.action_carrier=native,++rl_games.env_eval.action_chunk_execution.chunk_size=null,++rl_games.env_eval.action_chunk_execution.enabled=false,++rl_games.env_eval.distributed_mode=none,++rl_games.env_eval.enabled=true,++rl_games.env_eval.eval_backend=latency_bench,++rl_games.env_eval.eval_parallel_envs=5,++rl_games.env_eval.fixed_episode_seeds=true,++rl_games.env_eval.frameskip=1,++rl_games.env_eval.ghost_trail.gamma=1.3,++rl_games.env_eval.ghost_trail.ground_fraction=0.22,++rl_games.env_eval.ghost_trail.history_frames=5,++rl_games.env_eval.ghost_trail.min_alpha=35,++rl_games.env_eval.ghost_trail.scroll_px_per_step=4.0,++rl_games.env_eval.image_size=224,++rl_games.env_eval.image_transform=raw_rgb,++rl_games.env_eval.latency.mode=mixed,++rl_games.env_eval.latency.prompt_map_path=/home/ubuntu/talha/playground/Datasets/rl_games/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3/latency_prompt_map.json,++rl_games.env_eval.latency.prompt_map_path=null,++rl_games.env_eval.latency.values=[0],++rl_games.env_eval.latency_seed_stride=0,++rl_games.env_eval.mid_train.enabled=false,++rl_games.env_eval.mid_train.interval_steps=100,++rl_games.env_eval.mid_train.latencies=[0,2,4],++rl_games.env_eval.mid_train.max_steps_per_episode=3600,++rl_games.env_eval.mid_train.num_episodes=5,++rl_games.env_eval.post_train.enabled=true,++rl_games.env_eval.post_train.latencies=[0],++rl_games.env_eval.post_train.latencies_from_prompt_map=false,++rl_games.env_eval.post_train.max_steps_per_episode=3600,++rl_games.env_eval.post_train.num_episodes=20,++rl_games.env_eval.prompt_mode=raw,++rl_games.env_eval.seed=42,++rl_games.env_eval.task_description=,++rl_games.env_eval.task_seed_stride=0,++rl_games.env_eval.vectorized.batch_size=1,++rl_games.env_eval.vectorized.enabled=false,++rl_games.gymnasium.task_contract.action_labels=[push_left_hard,push_left_soft,zero_force,push_right_soft,push_right_hard],++rl_games.gymnasium.task_contract.action_values=[0,1,2,3,4],++rl_games.gymnasium.task_contract.base_prompt=''Keep + the pole upright by applying horizontal cart forces. Choose exactly one action + from: push_left_hard, push_left_soft, zero_force, push_right_soft, push_right_hard.'',++rl_games.gymnasium.task_contract.env_fps=25.0,++rl_games.gymnasium.task_contract.env_id=LatencyBench/InvertedPendulumDiscrete-v0,++rl_games.gymnasium.task_contract.frame_stack=1,++rl_games.gymnasium.task_contract.make_kwargs.base_env_id=InvertedPendulum-v4,++rl_games.gymnasium.task_contract.make_kwargs.force_values=[-3.0,-1.5,0.0,1.5,3.0],++rl_games.gymnasium.task_contract.make_kwargs.render_mode=rgb_array,++rl_games.gymnasium.task_contract.noop_action_id=2,++rl_games.gymnasium.task_contract.obs_fps=25.0,++rl_games.gymnasium.task_contract.registration_imports=[latency_bench.envs.gymnasium_mujoco],++rl_games.gymnasium.task_contract.state_space.labels=[cart_position,pole_angle,cart_velocity,pole_angular_velocity],++rl_games.gymnasium.task_contract.task_name=inverted_pendulum_rgb_state,++rl_games.gymnasium.task_name=inverted_pendulum_rgb_state,++rl_games.initialization_mode=scratch,++rl_games.model_alias=openvla,++rl_games.task=gymnasium,++run_id=openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3,++run_root_dir=/home/ubuntu/talha/results/Checkpoints,++run_root_dir=results/Checkpoints,++seed=42,++trainer.cc_f1_tolerance=1,++trainer.distributed_backend=none,++trainer.eval_action_classification=true,++trainer.eval_action_classification_interval=null,++trainer.eval_interval=5000,++trainer.eval_num_batches=200,++trainer.freeze_llm_layers=[],++trainer.freeze_modules=,++trainer.freeze_tied_embedding=false,++trainer.freeze_vit=false,++trainer.gradient_accumulation_steps=2,++trainer.gradient_clipping=1.0,++trainer.is_resume=false,++trainer.is_resume=false,++trainer.learning_rate.action_model=0.0001,++trainer.learning_rate.base=2e-05,++trainer.learning_rate.qwen_vl_interface=1e-05,++trainer.logging_frequency=1,++trainer.loss_scale.vla=1.0,++trainer.loss_scale.vlm=0.1,++trainer.lr_scheduler_type=cosine_with_min_lr,++trainer.max_grad_norm=1.0,++trainer.max_train_steps=5000,++trainer.num_warmup_steps=100,++trainer.optimizer.betas=[0.9,0.95],++trainer.optimizer.eps=1e-08,++trainer.optimizer.fused=true,++trainer.optimizer.name=AdamW,++trainer.optimizer.weight_decay=1e-08,++trainer.per_latency_eval_num_batches=5,++trainer.pretrained_checkpoint=null,++trainer.profile_timing.enabled=false,++trainer.profile_timing.log_interval=10,++trainer.reload_modules=null,++trainer.resume_step=0,++trainer.save_format=pt,++trainer.save_interval=5000,++trainer.scheduler_specific_kwargs.min_lr=1e-06,++trainer.weight_decay=0.0,++training_latency_condition=unspecified,++version_id=0.21,++wandb_group=openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3,++wandb_tags=[],++workspace_dir=/home/ubuntu/talha,env=gymnasium,init=scratch,mode=mixed_latency,model=openvla' + id: ??? + num: ??? + config_name: train + env_set: {} + env_copy: [] + config: + override_dirname: + kv_sep: '=' + item_sep: ',' + exclude_keys: [] + runtime: + version: 1.3.7 + version_base: '1.1' + cwd: /home/ubuntu/talha/starVLA + config_sources: + - path: hydra.conf + schema: pkg + provider: hydra + - path: /home/ubuntu/talha/starVLA/examples/rl_games/config + schema: file + provider: main + - path: '' + schema: structured + provider: schema + output_dir: /home/ubuntu/talha/results/Checkpoints/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3/hydra + choices: + cross_task_setup: null + checkpoint: default + mode: mixed_latency + init: scratch + env: gymnasium + model: openvla + hydra/env: default + hydra/callbacks: null + hydra/job_logging: default + hydra/hydra_logging: default + hydra/hydra_help: default + hydra/help: default + hydra/sweeper: basic + hydra/launcher: basic + hydra/output: default + verbose: false diff --git a/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3/hydra/.hydra/overrides.yaml b/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3/hydra/.hydra/overrides.yaml new file mode 100644 index 0000000000000000000000000000000000000000..07d096fc7c395e27568f0f644da46dbd589ee43c --- /dev/null +++ b/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3/hydra/.hydra/overrides.yaml @@ -0,0 +1,230 @@ +- model=openvla +- env=gymnasium +- init=scratch +- mode=mixed_latency +- ++framework.qwenvl.base_vlm=/home/ubuntu/talha/playground/Pretrained_models/Qwen3-VL-4B-Instruct +- ++framework.qwenvl.attn_implementation=flash_attention_2 +- ++framework.qwenvl.enable_gradient_checkpointing=true +- ++framework.action_model.state_dim=4 +- ++framework.action_model.loss_type=discrete_ce +- ++framework.action_model.action_horizon=1 +- ++framework.action_model.future_action_window_size=0 +- ++framework.action_model.past_action_window_size=0 +- ++framework.action_model.action_dim=5 +- ++framework.action_model.action_env_dim=5 +- ++framework.action_model.state_encoding=discretized_text +- ++framework.action_model.action_model_type=MLP +- ++framework.action_model.action_hidden_dim=2560 +- ++framework.kv_memory.enabled=false +- ++framework.kv_memory.window=4 +- ++framework.kv_memory.rollout_len=8 +- ++framework.kv_memory.packed_train=false +- ++framework.kv_memory.rebased_sink=true +- ++framework.name=QwenOFT +- ++datasets.vla_data.dataset_py=lerobot_datasets +- ++datasets.vla_data.include_state=true +- ++datasets.vla_data.data_root_dir=playground/Datasets/rl_games +- ++datasets.vla_data.data_mix=openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3 +- ++datasets.vla_data.eval_data_mix=openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3__val +- ++datasets.vla_data.custom_mixtures_path=null +- ++datasets.vla_data.action_type=discrete +- ++datasets.vla_data.sequential_step_sampling=true +- ++datasets.vla_data.eval_sequential_step_sampling=null +- ++datasets.vla_data.num_workers=8 +- ++datasets.vla_data.eval_num_workers=8 +- ++datasets.vla_data.prefetch_factor=4 +- ++datasets.vla_data.persistent_workers=true +- ++datasets.vla_data.pin_memory=true +- ++datasets.vla_data.shuffle=true +- ++datasets.vla_data.action_balance.enabled=false +- ++datasets.vla_data.action_balance.strategy=balanced_epoch +- ++datasets.vla_data.action_balance.action_key=action_id +- ++datasets.vla_data.action_balance.target_flap_fraction=0.3 +- ++datasets.vla_data.action_balance.noop_id=0 +- ++datasets.vla_data.action_balance.flap_id=1 +- ++datasets.vla_data.latency_curriculum.enabled=false +- ++datasets.vla_data.latency_curriculum.strategy=exclusive +- ++datasets.vla_data.latency_curriculum.latencies=null +- ++datasets.vla_data.latency_curriculum.phase_steps=null +- ++datasets.vla_data.latency_curriculum.phase_distributions=null +- ++datasets.vla_data.latency_curriculum.new_latency_passes=1.0 +- ++datasets.vla_data.latency_curriculum.replay_passes=0.25 +- ++datasets.vla_data.latency_curriculum.target_total_passes=2.0 +- ++datasets.vla_data.latency_curriculum.final_equalization=true +- ++datasets.vla_data.latency_curriculum.step_budget_mode=auto +- ++datasets.vla_data.latency_curriculum.eval_at_phase_end=false +- ++datasets.vla_data.latency_curriculum.save_at_phase_end=false +- ++datasets.vla_data.latency_curriculum.computed_plan=null +- ++datasets.vla_data.per_device_batch_size=8 +- ++datasets.vla_data.load_all_data_for_training=true +- ++datasets.vla_data.num_obs_frames=1 +- ++datasets.vla_data.image_mode=single +- ++datasets.vla_data.prompt_mode=raw +- ++datasets.vla_data.stitch_grid=[2,2] +- ++datasets.vla_data.obs_image_size=null +- ++datasets.vla_data.video_backend=torchvision_av +- '++datasets.vla_data.gymnasium_task_contract={task_name:inverted_pendulum_rgb_state,env_id:LatencyBench/InvertedPendulumDiscrete-v0,registration_imports:[latency_bench.envs.gymnasium_mujoco],env_fps:25.0,obs_fps:25.0,frame_stack:1,base_prompt:''Keep + the pole upright by applying horizontal cart forces. Choose exactly one action from: + push_left_hard, push_left_soft, zero_force, push_right_soft, push_right_hard.'',action_labels:[push_left_hard,push_left_soft,zero_force,push_right_soft,push_right_hard],action_values:[0,1,2,3,4],noop_action_id:2,make_kwargs:{base_env_id:InvertedPendulum-v4,render_mode:rgb_array,force_values:[-3.0,-1.5,0.0,1.5,3.0],base_make_kwargs:{}},state_space:{labels:[cart_position,pole_angle,cart_velocity,pole_angular_velocity]}}' +- ++datasets.vla_data.active_action_dim=5 +- ++dataset.source_hf=latency-sensitive-bench/inverted_pendulum_200ep_2 +- ++dataset.config_name=null +- ++dataset.source_subdir='inverted_pendulum_fixed_latency_0_100ep_1ksteps_og,inverted_pendulum_fixed_latency_2_1000ep_7k2steps,inverted_pendulum_fixed_latency_4_1000ep_7k2steps' +- ++dataset.converted_name=openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3 +- ++dataset.single_source_hf= +- ++dataset.mixed_source_hf= +- ++dataset.single_converted_name=flappy_train +- ++dataset.mixed_converted_name=flappy_train +- ++dataset.single_latency_filter=null +- ++dataset.mixed_latency_filter=null +- ++dataset.force_download=false +- ++dataset.setup_force=true +- ++dataset.skip_verification=true +- ++dataset.target_latency_unit=observation_steps +- ++dataset.verify_rows=200 +- ++dataset.max_episodes=null +- ++dataset.max_steps_per_episode=1000 +- ++dataset.episodes_per_latency=null +- ++dataset.episodes_per_latency_by_latency='0:90,2:1000,4:1000' +- ++dataset.latency_filter=[0,2,4] +- ++dataset.debug_subset.enabled=false +- ++dataset.debug_subset.max_episodes=5 +- ++dataset.debug_subset.suffix=debug +- ++base_model.repo_id=Qwen/Qwen3-VL-4B-Instruct +- ++initialization.checkpoint_local_dir=null +- ++initialization.checkpoint_hf_repo_id=null +- ++initialization.checkpoint_filename=null +- ++trainer.max_train_steps=5000 +- ++trainer.num_warmup_steps=100 +- ++trainer.save_interval=5000 +- ++trainer.eval_interval=5000 +- ++trainer.eval_num_batches=200 +- ++trainer.per_latency_eval_num_batches=5 +- ++trainer.eval_action_classification=true +- ++trainer.eval_action_classification_interval=null +- ++trainer.cc_f1_tolerance=1 +- ++trainer.learning_rate.base=2e-05 +- ++trainer.learning_rate.qwen_vl_interface=1e-05 +- ++trainer.learning_rate.action_model=0.0001 +- ++trainer.lr_scheduler_type=cosine_with_min_lr +- ++trainer.scheduler_specific_kwargs.min_lr=1e-06 +- ++trainer.freeze_modules= +- ++trainer.freeze_vit=false +- ++trainer.freeze_tied_embedding=false +- ++trainer.freeze_llm_layers=[] +- ++trainer.loss_scale.vla=1.0 +- ++trainer.loss_scale.vlm=0.1 +- ++trainer.max_grad_norm=1.0 +- ++trainer.weight_decay=0.0 +- ++trainer.logging_frequency=1 +- ++trainer.profile_timing.enabled=false +- ++trainer.profile_timing.log_interval=10 +- ++trainer.gradient_clipping=1.0 +- ++trainer.gradient_accumulation_steps=2 +- ++trainer.distributed_backend=none +- ++trainer.is_resume=false +- ++trainer.pretrained_checkpoint=null +- ++trainer.resume_step=0 +- ++trainer.reload_modules=null +- ++trainer.optimizer.name=AdamW +- ++trainer.optimizer.betas=[0.9,0.95] +- ++trainer.optimizer.eps=1e-08 +- ++trainer.optimizer.weight_decay=1e-08 +- ++trainer.optimizer.fused=true +- ++trainer.save_format=pt +- ++workspace_dir=/home/ubuntu/talha +- ++run_root_dir=results/Checkpoints +- ++seed=42 +- ++wandb_group=openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3 +- ++wandb_tags=[] +- ++training_latency_condition=unspecified +- ++auth.env_file=null +- ++auth.hf_token_env=HF_TOKEN +- ++auth.wandb_api_key_env=WANDB_API_KEY +- ++paths.run_root_dir=results/Checkpoints +- ++paths.dataset_local_dir=playground/Datasets/rl_games +- ++paths.dataset_cache_dir=null +- ++paths.base_model_dir=/home/ubuntu/talha/playground/Pretrained_models/Qwen3-VL-4B-Instruct +- ++paths.accelerate_config=starVLA/config/deepseeds/deepspeed_zero2.yaml +- ++rl_games.model_alias=openvla +- ++rl_games.env_eval.image_size=224 +- ++rl_games.env_eval.frameskip=1 +- ++rl_games.env_eval.image_transform=raw_rgb +- ++rl_games.env_eval.prompt_mode=raw +- ++rl_games.env_eval.ghost_trail.history_frames=5 +- ++rl_games.env_eval.ghost_trail.gamma=1.3 +- ++rl_games.env_eval.ghost_trail.min_alpha=35 +- ++rl_games.env_eval.ghost_trail.scroll_px_per_step=4.0 +- ++rl_games.env_eval.ghost_trail.ground_fraction=0.22 +- ++rl_games.env_eval.seed=42 +- ++rl_games.env_eval.fixed_episode_seeds=true +- ++rl_games.env_eval.latency_seed_stride=0 +- ++rl_games.env_eval.task_seed_stride=0 +- ++rl_games.env_eval.task_description= +- ++rl_games.env_eval.eval_parallel_envs=5 +- ++rl_games.env_eval.action_chunk_execution.enabled=false +- ++rl_games.env_eval.action_chunk_execution.chunk_size=null +- ++rl_games.env_eval.enabled=true +- ++rl_games.env_eval.mid_train.enabled=false +- ++rl_games.env_eval.mid_train.interval_steps=100 +- ++rl_games.env_eval.mid_train.latencies=[0,2,4] +- ++rl_games.env_eval.mid_train.num_episodes=5 +- ++rl_games.env_eval.mid_train.max_steps_per_episode=3600 +- ++rl_games.env_eval.post_train.enabled=true +- ++rl_games.env_eval.post_train.latencies=[0] +- ++rl_games.env_eval.post_train.latencies_from_prompt_map=false +- ++rl_games.env_eval.post_train.num_episodes=20 +- ++rl_games.env_eval.post_train.max_steps_per_episode=3600 +- ++rl_games.env_eval.eval_backend=latency_bench +- ++rl_games.env_eval.distributed_mode=none +- ++rl_games.env_eval.vectorized.enabled=false +- ++rl_games.env_eval.vectorized.batch_size=1 +- ++rl_games.env_eval.latency.prompt_map_path=null +- ++rl_games.env_eval.latency.mode=mixed +- ++rl_games.env_eval.latency.values=[0] +- ++rl_games.task=gymnasium +- ++rl_games.gymnasium.task_contract.task_name=inverted_pendulum_rgb_state +- ++rl_games.gymnasium.task_contract.env_id=LatencyBench/InvertedPendulumDiscrete-v0 +- ++rl_games.gymnasium.task_contract.registration_imports=[latency_bench.envs.gymnasium_mujoco] +- ++rl_games.gymnasium.task_contract.env_fps=25.0 +- ++rl_games.gymnasium.task_contract.obs_fps=25.0 +- ++rl_games.gymnasium.task_contract.frame_stack=1 +- '++rl_games.gymnasium.task_contract.base_prompt=''Keep the pole upright by applying + horizontal cart forces. Choose exactly one action from: push_left_hard, push_left_soft, + zero_force, push_right_soft, push_right_hard.''' +- ++rl_games.gymnasium.task_contract.action_labels=[push_left_hard,push_left_soft,zero_force,push_right_soft,push_right_hard] +- ++rl_games.gymnasium.task_contract.action_values=[0,1,2,3,4] +- ++rl_games.gymnasium.task_contract.noop_action_id=2 +- ++rl_games.gymnasium.task_contract.make_kwargs.base_env_id=InvertedPendulum-v4 +- ++rl_games.gymnasium.task_contract.make_kwargs.render_mode=rgb_array +- ++rl_games.gymnasium.task_contract.make_kwargs.force_values=[-3.0,-1.5,0.0,1.5,3.0] +- ++rl_games.gymnasium.task_contract.state_space.labels=[cart_position,pole_angle,cart_velocity,pole_angular_velocity] +- ++rl_games.gymnasium.task_name=inverted_pendulum_rgb_state +- ++rl_games.initialization_mode=scratch +- ++rl_games.action_carrier=native +- ++checkpoint.load=none +- ++checkpoint.hf_repo_id=talha15032/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3 +- ++checkpoint.save_best_model=false +- ++checkpoint.save_final_model=true +- ++checkpoint.save_pt_file=true +- ++checkpoint.save_training_state=false +- ++checkpoint.save_safetensors_file=false +- ++checkpoint.local.keep_last_n=1 +- ++checkpoint.sync.enabled=true +- ++checkpoint.sync.repo_id=talha15032/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3 +- ++checkpoint.sync.keep_last_n=0 +- ++checkpoint.sync.sync_every_n_checkpoints=1 +- ++checkpoint.sync.resume_policy=local_latest +- ++run_id=openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3 +- ++output_dir=null +- ++config_yaml=null +- ++is_debug=false +- ++version_id=0.21 +- ++run_root_dir=/home/ubuntu/talha/results/Checkpoints +- ++trainer.is_resume=false +- ++datasets.vla_data.data_root_dir=/home/ubuntu/talha/playground/Datasets/rl_games +- ++datasets.vla_data.data_mix=openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3 +- ++datasets.vla_data.eval_data_mix=openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3__val +- ++datasets.vla_data.custom_mixtures_path=/home/ubuntu/talha/playground/Datasets/rl_games/_generated_mixtures/single_79881079067b9b1b.json +- ++framework.qwenvl.base_vlm=/home/ubuntu/talha/playground/Pretrained_models/Qwen3-VL-4B-Instruct +- ++rl_games.env_eval.latency.prompt_map_path=/home/ubuntu/talha/playground/Datasets/rl_games/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3/latency_prompt_map.json diff --git a/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3/hydra/train_starvla_hydra.log b/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3/hydra/train_starvla_hydra.log new file mode 100644 index 0000000000000000000000000000000000000000..fd29f2e91c8de66955cea663bad6e5ee64aa10b6 --- /dev/null +++ b/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3/hydra/train_starvla_hydra.log @@ -0,0 +1,5071 @@ +[2026-10-03 18:54:30,754][starVLA.model.framework.share_tools][INFO] - [*] [apply_config_compat] normalised config from version_id=0.21 to '0.21' +[2026-10-03 18:54:30,754][starVLA.training.train_starvla][INFO] - VLA Training :: Warming Up +[2026-10-03 18:54:30,844][huggingface_hub._login][WARNING] - Note: Environment variable`HF_TOKEN` is set and is the current active token independently from the token you've just configured. +[2026-10-03 18:54:30,844][starVLA.training.rl_games.auth][INFO] - Logged into Hugging Face Hub +[2026-10-03 18:54:30,858][starVLA.training.train_starvla][INFO] - ✅ Configuration wrapped for access tracking +[2026-10-03 18:54:31,800][starVLA.training.train_starvla][INFO] - Creating VLA Dataset with Mixture `openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3` +[2026-10-03 18:54:32,179][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Behavior (data_config): ['BEHAVIOR_challenge'] +[2026-10-03 18:54:32,179][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from DOMINO: ['robotwin'] +[2026-10-03 18:54:32,179][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from DOMINO (data_config): [] +[2026-10-03 18:54:32,179][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from DOMINO (data_config): ['domino', 'domino_clean', 'domino_random', 'domino_cotrain'] +[2026-10-03 18:54:32,179][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Franka: ['custom_robot_config', 'demo_sim_franka_delta_joints', 'SO101'] +[2026-10-03 18:54:32,179][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Franka (data_config): [] +[2026-10-03 18:54:32,179][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Franka (data_config): ['custom_dataset', 'custom_dataset_2', 'demo_sim_pick_place', 'SO101_pick'] +[2026-10-03 18:54:32,179][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from LIBERO: ['libero_franka'] +[2026-10-03 18:54:32,180][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from LIBERO (data_config): [] +[2026-10-03 18:54:32,180][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from LIBERO (data_config): ['libero_all', 'libero_goal', 'multi_robot'] +[2026-10-03 18:54:32,180][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from MIKASA: ['mikasa_franka_h1'] +[2026-10-03 18:54:32,180][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from MIKASA (data_config): ['mikasa_franka_h1'] +[2026-10-03 18:54:32,180][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from MIKASA (data_config): ['local/intercept_grab_fast_vla_v0_h1_train'] +[2026-10-03 18:54:32,180][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from RoboChallenge_table30v2: ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-10-03 18:54:32,180][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from RoboChallenge_table30v2 (data_config): ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-10-03 18:54:32,180][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from RoboChallenge_table30v2 (data_config): ['robochallenge_table30v2_shred_paper', 'robochallenge_table30v2_ur5_all', 'robochallenge_table30v2_arx5_all', 'robochallenge_table30v2_dosw1_all'] +[2026-10-03 18:54:32,180][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_365: ['panda_omron_robocasa365'] +[2026-10-03 18:54:32,180][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_365 (data_config): [] +[2026-10-03 18:54:32,180][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_365 (data_config): ['robocasa365_open_drawer_target_human', 'robocasa365_atomic_target_human_all', 'robocasa365_composite_target_human_all', 'robocasa365_target_human_all'] +[2026-10-03 18:54:32,180][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_tabletop: ['fourier_gr1_arms_waist'] +[2026-10-03 18:54:32,180][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_tabletop (data_config): [] +[2026-10-03 18:54:32,180][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_tabletop (data_config): ['fourier_gr1_unified_1000'] +[2026-10-03 18:54:32,180][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robotwin: ['robotwin', 'robotwin50', 'arx_x5'] +[2026-10-03 18:54:32,180][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robotwin (data_config): [] +[2026-10-03 18:54:32,180][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robotwin (data_config): ['robotwin_all', 'robotwin_all_50', 'robotwin', 'robotwin_task1', 'robotwin_task2', 'arx_x5'] +[2026-10-03 18:54:32,181][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from SimplerEnv: ['oxe_droid', 'oxe_bridge', 'oxe_rt1'] +[2026-10-03 18:54:32,181][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from SimplerEnv (data_config): [] +[2026-10-03 18:54:32,181][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from SimplerEnv (data_config): ['bridge', 'bridge_rt_1'] +[2026-10-03 18:54:32,181][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from VLA-Arena: ['vla_arena_franka'] +[2026-10-03 18:54:32,181][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from VLA-Arena (data_config): [] +[2026-10-03 18:54:32,181][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from VLA-Arena (data_config): ['vla_arena_L0_S', 'vla_arena_L0_M', 'vla_arena_L0_L'] +[2026-10-03 18:54:32,181][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from rl_games: ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_defend_the_line', 'rl_games_deadly_corridor', 'rl_games_asterix', 'rl_games_atlantis', 'rl_games_gymnasium', 'rl_games_gymnasium_discrete', 'rl_games_gymnasium_native'] +[2026-10-03 18:54:32,181][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from rl_games (data_config): ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_defend_the_line', 'rl_games_deadly_corridor', 'rl_games_asterix', 'rl_games_atlantis', 'rl_games_gymnasium', 'rl_games_gymnasium_discrete', 'rl_games_gymnasium_native'] +[2026-10-03 18:54:32,181][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from rl_games (data_config): ['flappy_train', 'flappy_train__bridge', 'flappy_mixed_latency_train', 'flappy_mixed_latency_train__bridge', 'demon_attack_train', 'demon_attack_train__bridge', 'demon_attack_mixed_latency_train', 'demon_attack_mixed_latency_train__bridge', 'defend_the_line_train', 'defend_the_line_train__bridge', 'defend_the_line_mixed_latency_train', 'defend_the_line_mixed_latency_train__bridge', 'deadly_corridor_train', 'deadly_corridor_train__bridge', 'deadly_corridor_mixed_latency_train', 'deadly_corridor_mixed_latency_train__bridge', 'asterix_train', 'asterix_train__bridge', 'asterix_mixed_latency_train', 'asterix_mixed_latency_train__bridge', 'atlantis_train', 'atlantis_train__bridge', 'atlantis_mixed_latency_train', 'atlantis_mixed_latency_train__bridge', 'inverted_pendulum_mixed_latency_train', 'inverted_pendulum_mixed_latency_train__bridge', 'inverted_pendulum_mixed_latency_train_no_latency_prompt', 'inverted_pendulum_mixed_latency_train_no_latency_prompt__bridge', 'h1hand_balance_hard'] +[2026-10-03 18:54:32,189][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded custom mixtures from /home/ubuntu/talha/playground/Datasets/rl_games/_generated_mixtures/single_79881079067b9b1b.json: ['openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3', 'openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3__val'] +[2026-10-03 18:54:32,222][starVLA.dataloader][INFO] - Created train VLA dataloader with num_workers=8, persistent_workers=True, cpu_only_workers=True +[2026-10-03 18:54:32,223][starVLA.training.train_starvla][INFO] - Creating VLA Eval Dataset with Mixture `openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3__val` +[2026-10-03 18:54:32,230][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded custom mixtures from /home/ubuntu/talha/playground/Datasets/rl_games/_generated_mixtures/single_79881079067b9b1b.json: ['openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3', 'openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3__val'] +[2026-10-03 18:54:32,235][starVLA.dataloader][INFO] - Created eval VLA dataloader with num_workers=8, persistent_workers=True, cpu_only_workers=True +[2026-10-03 18:54:32,289][starVLA.training.train_starvla][INFO] - 📝 Full config saved at /home/ubuntu/talha/results/Checkpoints/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3/config.full.yaml +[2026-10-03 18:54:32,298][starVLA.training.train_starvla][INFO] - 📊 Accessed config snapshot saved at /home/ubuntu/talha/results/Checkpoints/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3/config.yaml +[2026-10-03 18:54:32,298][starVLA.training.train_starvla][INFO] - No pretrained checkpoint provided. Starting training from scratch. +[2026-10-03 18:54:34,187][starVLA.training.train_starvla][INFO] - ***** Training Configuration ***** +[2026-10-03 18:54:34,188][starVLA.training.train_starvla][INFO] - Total optimization steps = 5000 +[2026-10-03 18:54:34,188][starVLA.training.train_starvla][INFO] - Per device batch size = 8 +[2026-10-03 18:54:34,188][starVLA.training.train_starvla][INFO] - Gradient accumulation steps = 2 +[2026-10-03 18:54:34,188][starVLA.training.train_starvla][INFO] - Total batch size = 16 +[2026-10-03 18:55:00,103][starVLA.training.train_starvla][INFO] - Step 1, Loss: {'train/loss': 1.731070876121521, 'train/grad_norm_pre_clip': 101.0099868774414, 'timing/data': 0.000501629983773455, 'timing/model': 0.48186834700754844, 'learning_rate/qwen_vl_interface': 1.0000000000000001e-07, 'learning_rate/action_model': 1.0000000000000002e-06, 'epoch': 0.0}) +[2026-10-03 18:55:00,821][starVLA.training.train_starvla][INFO] - Step 2, Loss: {'train/loss': 1.6959222555160522, 'train/grad_norm_pre_clip': 99.61383819580078, 'timing/data': 0.0004346799978520721, 'timing/model': 0.39920500601874664, 'learning_rate/qwen_vl_interface': 2.0000000000000002e-07, 'learning_rate/action_model': 2.0000000000000003e-06, 'epoch': 0.0}) +[2026-10-03 18:55:01,535][starVLA.training.train_starvla][INFO] - Step 3, Loss: {'train/loss': 1.4199117422103882, 'train/grad_norm_pre_clip': 145.0426025390625, 'timing/data': 0.0003421299916226417, 'timing/model': 0.3980540460033808, 'learning_rate/qwen_vl_interface': 3.0000000000000004e-07, 'learning_rate/action_model': 3e-06, 'epoch': 0.0}) +[2026-10-03 18:55:02,244][starVLA.training.train_starvla][INFO] - Step 4, Loss: {'train/loss': 1.1607365608215332, 'train/grad_norm_pre_clip': 83.6635971069336, 'timing/data': 0.00024299000506289303, 'timing/model': 0.3951603659952525, 'learning_rate/qwen_vl_interface': 4.0000000000000003e-07, 'learning_rate/action_model': 4.000000000000001e-06, 'epoch': 0.0}) +[2026-10-03 18:55:02,950][starVLA.training.train_starvla][INFO] - Step 5, Loss: {'train/loss': 0.6521099805831909, 'train/grad_norm_pre_clip': 52.29100799560547, 'timing/data': 0.0003045200137421489, 'timing/model': 0.39349776701419614, 'learning_rate/qwen_vl_interface': 5.000000000000001e-07, 'learning_rate/action_model': 5e-06, 'epoch': 0.0}) +[2026-10-03 18:55:03,855][starVLA.training.train_starvla][INFO] - Step 6, Loss: {'train/loss': 0.7113391160964966, 'train/grad_norm_pre_clip': 25.722610473632812, 'timing/data': 0.00028666999423876405, 'timing/model': 0.5948029250139371, 'learning_rate/qwen_vl_interface': 6.000000000000001e-07, 'learning_rate/action_model': 6e-06, 'epoch': 0.0}) +[2026-10-03 18:55:04,568][starVLA.training.train_starvla][INFO] - Step 7, Loss: {'train/loss': 1.1461814641952515, 'train/grad_norm_pre_clip': 89.96150207519531, 'timing/data': 0.0003609600244089961, 'timing/model': 0.3984189759939909, 'learning_rate/qwen_vl_interface': 7.000000000000001e-07, 'learning_rate/action_model': 7.000000000000001e-06, 'epoch': 0.0}) +[2026-10-03 18:55:05,276][starVLA.training.train_starvla][INFO] - Step 8, Loss: {'train/loss': 1.930991768836975, 'train/grad_norm_pre_clip': 92.71530151367188, 'timing/data': 0.00022012999397702515, 'timing/model': 0.3977579969796352, 'learning_rate/qwen_vl_interface': 8.000000000000001e-07, 'learning_rate/action_model': 8.000000000000001e-06, 'epoch': 0.0}) +[2026-10-03 18:55:05,983][starVLA.training.train_starvla][INFO] - Step 9, Loss: {'train/loss': 2.643430233001709, 'train/grad_norm_pre_clip': 80.11473083496094, 'timing/data': 0.0001983499969355762, 'timing/model': 0.3948570679931436, 'learning_rate/qwen_vl_interface': 9.000000000000001e-07, 'learning_rate/action_model': 9e-06, 'epoch': 0.0}) +[2026-10-03 18:55:06,689][starVLA.training.train_starvla][INFO] - Step 10, Loss: {'train/loss': 1.5373499393463135, 'train/grad_norm_pre_clip': 33.36509704589844, 'timing/data': 0.0003041800227947533, 'timing/model': 0.39387426601024345, 'learning_rate/qwen_vl_interface': 1.0000000000000002e-06, 'learning_rate/action_model': 1e-05, 'epoch': 0.0}) +[2026-10-03 18:55:07,597][starVLA.training.train_starvla][INFO] - Step 11, Loss: {'train/loss': 1.618821382522583, 'train/grad_norm_pre_clip': 57.513023376464844, 'timing/data': 0.0003488000074867159, 'timing/model': 0.5943233449943364, 'learning_rate/qwen_vl_interface': 1.1e-06, 'learning_rate/action_model': 1.1000000000000001e-05, 'epoch': 0.0}) +[2026-10-03 18:55:08,303][starVLA.training.train_starvla][INFO] - Step 12, Loss: {'train/loss': 1.8319751024246216, 'train/grad_norm_pre_clip': 74.84938049316406, 'timing/data': 0.0003285799757577479, 'timing/model': 0.39453363697975874, 'learning_rate/qwen_vl_interface': 1.2000000000000002e-06, 'learning_rate/action_model': 1.2e-05, 'epoch': 0.0}) +[2026-10-03 18:55:09,010][starVLA.training.train_starvla][INFO] - Step 13, Loss: {'train/loss': 1.5256803035736084, 'train/grad_norm_pre_clip': 106.13189697265625, 'timing/data': 0.00035148998722434044, 'timing/model': 0.3942428070004098, 'learning_rate/qwen_vl_interface': 1.3e-06, 'learning_rate/action_model': 1.3000000000000001e-05, 'epoch': 0.0}) +[2026-10-03 18:55:09,715][starVLA.training.train_starvla][INFO] - Step 14, Loss: {'train/loss': 2.2725934982299805, 'train/grad_norm_pre_clip': 133.13143920898438, 'timing/data': 0.00021396999363787472, 'timing/model': 0.39354398800060153, 'learning_rate/qwen_vl_interface': 1.4000000000000001e-06, 'learning_rate/action_model': 1.4000000000000001e-05, 'epoch': 0.0}) +[2026-10-03 18:55:10,427][starVLA.training.train_starvla][INFO] - Step 15, Loss: {'train/loss': 1.4799635410308838, 'train/grad_norm_pre_clip': 108.76581573486328, 'timing/data': 0.0003003699821420014, 'timing/model': 0.398980035999557, 'learning_rate/qwen_vl_interface': 1.5e-06, 'learning_rate/action_model': 1.5e-05, 'epoch': 0.0}) +[2026-10-03 18:55:11,135][starVLA.training.train_starvla][INFO] - Step 16, Loss: {'train/loss': 1.1939482688903809, 'train/grad_norm_pre_clip': 41.60868835449219, 'timing/data': 0.00035667000338435173, 'timing/model': 0.3958419559930917, 'learning_rate/qwen_vl_interface': 1.6000000000000001e-06, 'learning_rate/action_model': 1.6000000000000003e-05, 'epoch': 0.0}) +[2026-10-03 18:55:12,045][starVLA.training.train_starvla][INFO] - Step 17, Loss: {'train/loss': 1.1559138298034668, 'train/grad_norm_pre_clip': 66.51953125, 'timing/data': 0.0002730300184339285, 'timing/model': 0.395160037005553, 'learning_rate/qwen_vl_interface': 1.7000000000000002e-06, 'learning_rate/action_model': 1.7000000000000003e-05, 'epoch': 0.0}) +[2026-10-03 18:55:12,752][starVLA.training.train_starvla][INFO] - Step 18, Loss: {'train/loss': 1.4390403032302856, 'train/grad_norm_pre_clip': 49.60825729370117, 'timing/data': 0.00035955000203102827, 'timing/model': 0.39426596701378, 'learning_rate/qwen_vl_interface': 1.8000000000000001e-06, 'learning_rate/action_model': 1.8e-05, 'epoch': 0.0}) +[2026-10-03 18:55:13,465][starVLA.training.train_starvla][INFO] - Step 19, Loss: {'train/loss': 1.4735064506530762, 'train/grad_norm_pre_clip': 62.687801361083984, 'timing/data': 0.0002531500067561865, 'timing/model': 0.3978305670025293, 'learning_rate/qwen_vl_interface': 1.9000000000000002e-06, 'learning_rate/action_model': 1.9e-05, 'epoch': 0.0}) +[2026-10-03 18:55:14,172][starVLA.training.train_starvla][INFO] - Step 20, Loss: {'train/loss': 1.0616859197616577, 'train/grad_norm_pre_clip': 36.891700744628906, 'timing/data': 0.00028909998945891857, 'timing/model': 0.39576520601985976, 'learning_rate/qwen_vl_interface': 2.0000000000000003e-06, 'learning_rate/action_model': 2e-05, 'epoch': 0.0}) +[2026-10-03 18:55:14,879][starVLA.training.train_starvla][INFO] - Step 21, Loss: {'train/loss': 0.7642251253128052, 'train/grad_norm_pre_clip': 26.355493545532227, 'timing/data': 0.0003236600023228675, 'timing/model': 0.3940819480048958, 'learning_rate/qwen_vl_interface': 2.1000000000000002e-06, 'learning_rate/action_model': 2.1e-05, 'epoch': 0.0}) +[2026-10-03 18:55:15,782][starVLA.training.train_starvla][INFO] - Step 22, Loss: {'train/loss': 1.1152973175048828, 'train/grad_norm_pre_clip': 27.771989822387695, 'timing/data': 0.00035668001510202885, 'timing/model': 0.5914146350114606, 'learning_rate/qwen_vl_interface': 2.2e-06, 'learning_rate/action_model': 2.2000000000000003e-05, 'epoch': 0.0}) +[2026-10-03 18:55:16,489][starVLA.training.train_starvla][INFO] - Step 23, Loss: {'train/loss': 1.7852745056152344, 'train/grad_norm_pre_clip': 79.75607299804688, 'timing/data': 0.00031850000959821045, 'timing/model': 0.3957079370156862, 'learning_rate/qwen_vl_interface': 2.3000000000000004e-06, 'learning_rate/action_model': 2.3000000000000003e-05, 'epoch': 0.0}) +[2026-10-03 18:55:17,198][starVLA.training.train_starvla][INFO] - Step 24, Loss: {'train/loss': 2.1357624530792236, 'train/grad_norm_pre_clip': 77.93151092529297, 'timing/data': 0.00033780999365262687, 'timing/model': 0.39674014699994586, 'learning_rate/qwen_vl_interface': 2.4000000000000003e-06, 'learning_rate/action_model': 2.4e-05, 'epoch': 0.0}) +[2026-10-03 18:55:17,904][starVLA.training.train_starvla][INFO] - Step 25, Loss: {'train/loss': 1.5689797401428223, 'train/grad_norm_pre_clip': 75.91065216064453, 'timing/data': 0.0002531800128053874, 'timing/model': 0.395292736007832, 'learning_rate/qwen_vl_interface': 2.5e-06, 'learning_rate/action_model': 2.5e-05, 'epoch': 0.0}) +[2026-10-03 18:55:18,612][starVLA.training.train_starvla][INFO] - Step 26, Loss: {'train/loss': 1.1200392246246338, 'train/grad_norm_pre_clip': 93.8840103149414, 'timing/data': 0.00037236997741274536, 'timing/model': 0.3954434870101977, 'learning_rate/qwen_vl_interface': 2.6e-06, 'learning_rate/action_model': 2.6000000000000002e-05, 'epoch': 0.0}) +[2026-10-03 18:55:19,322][starVLA.training.train_starvla][INFO] - Step 27, Loss: {'train/loss': 1.521402359008789, 'train/grad_norm_pre_clip': 46.086875915527344, 'timing/data': 0.00026666000485420227, 'timing/model': 0.3956884959770832, 'learning_rate/qwen_vl_interface': 2.7000000000000004e-06, 'learning_rate/action_model': 2.7000000000000002e-05, 'epoch': 0.0}) +[2026-10-03 18:55:20,228][starVLA.training.train_starvla][INFO] - Step 28, Loss: {'train/loss': 1.21690833568573, 'train/grad_norm_pre_clip': 25.83342742919922, 'timing/data': 0.0002609299845062196, 'timing/model': 0.39688033700804226, 'learning_rate/qwen_vl_interface': 2.8000000000000003e-06, 'learning_rate/action_model': 2.8000000000000003e-05, 'epoch': 0.0}) +[2026-10-03 18:55:20,936][starVLA.training.train_starvla][INFO] - Step 29, Loss: {'train/loss': 1.1975409984588623, 'train/grad_norm_pre_clip': 32.90830993652344, 'timing/data': 0.00026893001631833613, 'timing/model': 0.39492798698483966, 'learning_rate/qwen_vl_interface': 2.9e-06, 'learning_rate/action_model': 2.9e-05, 'epoch': 0.0}) +[2026-10-03 18:55:21,643][starVLA.training.train_starvla][INFO] - Step 30, Loss: {'train/loss': 1.4540408849716187, 'train/grad_norm_pre_clip': 39.06976318359375, 'timing/data': 0.00034065000363625586, 'timing/model': 0.3951288160169497, 'learning_rate/qwen_vl_interface': 3e-06, 'learning_rate/action_model': 3e-05, 'epoch': 0.0}) +[2026-10-03 18:55:22,354][starVLA.training.train_starvla][INFO] - Step 31, Loss: {'train/loss': 1.0513240098953247, 'train/grad_norm_pre_clip': 16.87114906311035, 'timing/data': 0.00032007001573219895, 'timing/model': 0.39732601700234227, 'learning_rate/qwen_vl_interface': 3.1000000000000004e-06, 'learning_rate/action_model': 3.1e-05, 'epoch': 0.0}) +[2026-10-03 18:55:23,065][starVLA.training.train_starvla][INFO] - Step 32, Loss: {'train/loss': 1.1013355255126953, 'train/grad_norm_pre_clip': 38.973899841308594, 'timing/data': 0.00041480999789200723, 'timing/model': 0.3977734360087197, 'learning_rate/qwen_vl_interface': 3.2000000000000003e-06, 'learning_rate/action_model': 3.2000000000000005e-05, 'epoch': 0.0}) +[2026-10-03 18:55:23,968][starVLA.training.train_starvla][INFO] - Step 33, Loss: {'train/loss': 1.1706204414367676, 'train/grad_norm_pre_clip': 136.35919189453125, 'timing/data': 0.000272060016868636, 'timing/model': 0.394416547002038, 'learning_rate/qwen_vl_interface': 3.3000000000000006e-06, 'learning_rate/action_model': 3.3e-05, 'epoch': 0.0}) +[2026-10-03 18:55:24,674][starVLA.training.train_starvla][INFO] - Step 34, Loss: {'train/loss': 1.0225833654403687, 'train/grad_norm_pre_clip': 12.746820449829102, 'timing/data': 0.0002635600103531033, 'timing/model': 0.393710996984737, 'learning_rate/qwen_vl_interface': 3.4000000000000005e-06, 'learning_rate/action_model': 3.4000000000000007e-05, 'epoch': 0.0}) +[2026-10-03 18:55:25,384][starVLA.training.train_starvla][INFO] - Step 35, Loss: {'train/loss': 0.973082959651947, 'train/grad_norm_pre_clip': 25.449926376342773, 'timing/data': 0.0003778299724217504, 'timing/model': 0.39713715700781904, 'learning_rate/qwen_vl_interface': 3.5e-06, 'learning_rate/action_model': 3.5e-05, 'epoch': 0.0}) +[2026-10-03 18:55:26,093][starVLA.training.train_starvla][INFO] - Step 36, Loss: {'train/loss': 2.086127281188965, 'train/grad_norm_pre_clip': 64.02236938476562, 'timing/data': 0.0003684900002554059, 'timing/model': 0.39726656599668786, 'learning_rate/qwen_vl_interface': 3.6000000000000003e-06, 'learning_rate/action_model': 3.6e-05, 'epoch': 0.0}) +[2026-10-03 18:55:26,799][starVLA.training.train_starvla][INFO] - Step 37, Loss: {'train/loss': 0.7174515724182129, 'train/grad_norm_pre_clip': 99.11912536621094, 'timing/data': 0.00024493999080732465, 'timing/model': 0.39474321698071435, 'learning_rate/qwen_vl_interface': 3.7e-06, 'learning_rate/action_model': 3.7e-05, 'epoch': 0.0}) +[2026-10-03 18:55:27,698][starVLA.training.train_starvla][INFO] - Step 38, Loss: {'train/loss': 1.0513038635253906, 'train/grad_norm_pre_clip': 42.82645034790039, 'timing/data': 0.00029771000845357776, 'timing/model': 0.5869118559930939, 'learning_rate/qwen_vl_interface': 3.8000000000000005e-06, 'learning_rate/action_model': 3.8e-05, 'epoch': 0.0}) +[2026-10-03 18:55:28,407][starVLA.training.train_starvla][INFO] - Step 39, Loss: {'train/loss': 0.9701728224754333, 'train/grad_norm_pre_clip': 45.321807861328125, 'timing/data': 0.000391440000385046, 'timing/model': 0.39577876700786874, 'learning_rate/qwen_vl_interface': 3.900000000000001e-06, 'learning_rate/action_model': 3.9000000000000006e-05, 'epoch': 0.0}) +[2026-10-03 18:55:29,117][starVLA.training.train_starvla][INFO] - Step 40, Loss: {'train/loss': 0.74350506067276, 'train/grad_norm_pre_clip': 17.057899475097656, 'timing/data': 0.0002699799952097237, 'timing/model': 0.39768926601391286, 'learning_rate/qwen_vl_interface': 4.000000000000001e-06, 'learning_rate/action_model': 4e-05, 'epoch': 0.0}) +[2026-10-03 18:55:29,825][starVLA.training.train_starvla][INFO] - Step 41, Loss: {'train/loss': 1.0507798194885254, 'train/grad_norm_pre_clip': 38.58891677856445, 'timing/data': 0.0002222199982497841, 'timing/model': 0.395397806976689, 'learning_rate/qwen_vl_interface': 4.1e-06, 'learning_rate/action_model': 4.1e-05, 'epoch': 0.0}) +[2026-10-03 18:55:30,531][starVLA.training.train_starvla][INFO] - Step 42, Loss: {'train/loss': 1.8750879764556885, 'train/grad_norm_pre_clip': 113.28285217285156, 'timing/data': 0.0003515499993227422, 'timing/model': 0.3939704379881732, 'learning_rate/qwen_vl_interface': 4.2000000000000004e-06, 'learning_rate/action_model': 4.2e-05, 'epoch': 0.0}) +[2026-10-03 18:55:31,243][starVLA.training.train_starvla][INFO] - Step 43, Loss: {'train/loss': 0.9814528822898865, 'train/grad_norm_pre_clip': 49.284549713134766, 'timing/data': 0.00024845998268574476, 'timing/model': 0.39763870698516257, 'learning_rate/qwen_vl_interface': 4.3e-06, 'learning_rate/action_model': 4.3e-05, 'epoch': 0.0}) +[2026-10-03 18:55:32,148][starVLA.training.train_starvla][INFO] - Step 44, Loss: {'train/loss': 2.5733165740966797, 'train/grad_norm_pre_clip': 112.23020935058594, 'timing/data': 0.0004034000157844275, 'timing/model': 0.3968456370057538, 'learning_rate/qwen_vl_interface': 4.4e-06, 'learning_rate/action_model': 4.4000000000000006e-05, 'epoch': 0.0}) +[2026-10-03 18:55:32,855][starVLA.training.train_starvla][INFO] - Step 45, Loss: {'train/loss': 1.0917702913284302, 'train/grad_norm_pre_clip': 143.18460083007812, 'timing/data': 0.0003658000205177814, 'timing/model': 0.39579027600120753, 'learning_rate/qwen_vl_interface': 4.5e-06, 'learning_rate/action_model': 4.5e-05, 'epoch': 0.0}) +[2026-10-03 18:55:33,563][starVLA.training.train_starvla][INFO] - Step 46, Loss: {'train/loss': 1.7797904014587402, 'train/grad_norm_pre_clip': 174.06642150878906, 'timing/data': 0.00033420001273043454, 'timing/model': 0.39584379698499106, 'learning_rate/qwen_vl_interface': 4.600000000000001e-06, 'learning_rate/action_model': 4.600000000000001e-05, 'epoch': 0.0}) +[2026-10-03 18:55:34,275][starVLA.training.train_starvla][INFO] - Step 47, Loss: {'train/loss': 1.3211092948913574, 'train/grad_norm_pre_clip': 121.17952728271484, 'timing/data': 0.00029520000680349767, 'timing/model': 0.39749425699119456, 'learning_rate/qwen_vl_interface': 4.7e-06, 'learning_rate/action_model': 4.7e-05, 'epoch': 0.0}) +[2026-10-03 18:55:34,985][starVLA.training.train_starvla][INFO] - Step 48, Loss: {'train/loss': 1.2750964164733887, 'train/grad_norm_pre_clip': 36.06625747680664, 'timing/data': 0.00036260997876524925, 'timing/model': 0.3976083470042795, 'learning_rate/qwen_vl_interface': 4.800000000000001e-06, 'learning_rate/action_model': 4.8e-05, 'epoch': 0.0}) +[2026-10-03 18:55:35,890][starVLA.training.train_starvla][INFO] - Step 49, Loss: {'train/loss': 0.5199482440948486, 'train/grad_norm_pre_clip': 20.285728454589844, 'timing/data': 0.00037364999298006296, 'timing/model': 0.394785287004197, 'learning_rate/qwen_vl_interface': 4.9000000000000005e-06, 'learning_rate/action_model': 4.9e-05, 'epoch': 0.0}) +[2026-10-03 18:55:36,597][starVLA.training.train_starvla][INFO] - Step 50, Loss: {'train/loss': 1.0404489040374756, 'train/grad_norm_pre_clip': 75.14949798583984, 'timing/data': 0.0002546900068409741, 'timing/model': 0.3937185770191718, 'learning_rate/qwen_vl_interface': 5e-06, 'learning_rate/action_model': 5e-05, 'epoch': 0.0}) +[2026-10-03 18:55:37,309][starVLA.training.train_starvla][INFO] - Step 51, Loss: {'train/loss': 0.6957734823226929, 'train/grad_norm_pre_clip': 13.846778869628906, 'timing/data': 0.0002828899887390435, 'timing/model': 0.3977717270026915, 'learning_rate/qwen_vl_interface': 5.1e-06, 'learning_rate/action_model': 5.1000000000000006e-05, 'epoch': 0.0}) +[2026-10-03 18:55:38,021][starVLA.training.train_starvla][INFO] - Step 52, Loss: {'train/loss': 2.182025671005249, 'train/grad_norm_pre_clip': 87.96633911132812, 'timing/data': 0.0002821300004143268, 'timing/model': 0.3975314370181877, 'learning_rate/qwen_vl_interface': 5.2e-06, 'learning_rate/action_model': 5.2000000000000004e-05, 'epoch': 0.0}) +[2026-10-03 18:55:38,728][starVLA.training.train_starvla][INFO] - Step 53, Loss: {'train/loss': 2.2598042488098145, 'train/grad_norm_pre_clip': 43.546478271484375, 'timing/data': 0.00022238001110963523, 'timing/model': 0.39602805697359145, 'learning_rate/qwen_vl_interface': 5.300000000000001e-06, 'learning_rate/action_model': 5.300000000000001e-05, 'epoch': 0.0}) +[2026-10-03 18:55:39,628][starVLA.training.train_starvla][INFO] - Step 54, Loss: {'train/loss': 2.0924360752105713, 'train/grad_norm_pre_clip': 60.925933837890625, 'timing/data': 0.0002202100004069507, 'timing/model': 0.5893035850021988, 'learning_rate/qwen_vl_interface': 5.400000000000001e-06, 'learning_rate/action_model': 5.4000000000000005e-05, 'epoch': 0.0}) +[2026-10-03 18:55:40,338][starVLA.training.train_starvla][INFO] - Step 55, Loss: {'train/loss': 1.992997169494629, 'train/grad_norm_pre_clip': 56.45095443725586, 'timing/data': 0.0003779199905693531, 'timing/model': 0.39700288700987585, 'learning_rate/qwen_vl_interface': 5.500000000000001e-06, 'learning_rate/action_model': 5.500000000000001e-05, 'epoch': 0.0}) +[2026-10-03 18:55:41,047][starVLA.training.train_starvla][INFO] - Step 56, Loss: {'train/loss': 0.9241682291030884, 'train/grad_norm_pre_clip': 28.12437629699707, 'timing/data': 0.00022668001474812627, 'timing/model': 0.3954659969895147, 'learning_rate/qwen_vl_interface': 5.600000000000001e-06, 'learning_rate/action_model': 5.6000000000000006e-05, 'epoch': 0.0}) +[2026-10-03 18:55:41,752][starVLA.training.train_starvla][INFO] - Step 57, Loss: {'train/loss': 1.05050790309906, 'train/grad_norm_pre_clip': 36.464599609375, 'timing/data': 0.00035976001527160406, 'timing/model': 0.39406266700825654, 'learning_rate/qwen_vl_interface': 5.7e-06, 'learning_rate/action_model': 5.6999999999999996e-05, 'epoch': 0.0}) +[2026-10-03 18:55:42,459][starVLA.training.train_starvla][INFO] - Step 58, Loss: {'train/loss': 0.7998765707015991, 'train/grad_norm_pre_clip': 24.18086051940918, 'timing/data': 0.00025024000206030905, 'timing/model': 0.3953825759817846, 'learning_rate/qwen_vl_interface': 5.8e-06, 'learning_rate/action_model': 5.8e-05, 'epoch': 0.0}) +[2026-10-03 18:55:43,371][starVLA.training.train_starvla][INFO] - Step 59, Loss: {'train/loss': 1.0994011163711548, 'train/grad_norm_pre_clip': 36.95240783691406, 'timing/data': 0.00027951999800279737, 'timing/model': 0.6002289339958224, 'learning_rate/qwen_vl_interface': 5.9e-06, 'learning_rate/action_model': 5.9e-05, 'epoch': 0.0}) +[2026-10-03 18:55:44,081][starVLA.training.train_starvla][INFO] - Step 60, Loss: {'train/loss': 0.8553933501243591, 'train/grad_norm_pre_clip': 46.172088623046875, 'timing/data': 0.00026930999592877924, 'timing/model': 0.39681563701014966, 'learning_rate/qwen_vl_interface': 6e-06, 'learning_rate/action_model': 6e-05, 'epoch': 0.0}) +[2026-10-03 18:55:44,787][starVLA.training.train_starvla][INFO] - Step 61, Loss: {'train/loss': 0.959460973739624, 'train/grad_norm_pre_clip': 29.15613555908203, 'timing/data': 0.00029722999897785485, 'timing/model': 0.3951041070104111, 'learning_rate/qwen_vl_interface': 6.1e-06, 'learning_rate/action_model': 6.1e-05, 'epoch': 0.0}) +[2026-10-03 18:55:45,494][starVLA.training.train_starvla][INFO] - Step 62, Loss: {'train/loss': 1.6915099620819092, 'train/grad_norm_pre_clip': 56.91743850708008, 'timing/data': 0.00034915999276563525, 'timing/model': 0.3944423969951458, 'learning_rate/qwen_vl_interface': 6.200000000000001e-06, 'learning_rate/action_model': 6.2e-05, 'epoch': 0.0}) +[2026-10-03 18:55:46,200][starVLA.training.train_starvla][INFO] - Step 63, Loss: {'train/loss': 0.8868730068206787, 'train/grad_norm_pre_clip': 59.55645751953125, 'timing/data': 0.000264620000962168, 'timing/model': 0.3947337070130743, 'learning_rate/qwen_vl_interface': 6.300000000000001e-06, 'learning_rate/action_model': 6.3e-05, 'epoch': 0.0}) +[2026-10-03 18:55:47,109][starVLA.training.train_starvla][INFO] - Step 64, Loss: {'train/loss': 2.391446590423584, 'train/grad_norm_pre_clip': 56.5482063293457, 'timing/data': 0.0002212200197391212, 'timing/model': 0.59722655499354, 'learning_rate/qwen_vl_interface': 6.4000000000000006e-06, 'learning_rate/action_model': 6.400000000000001e-05, 'epoch': 0.0}) +[2026-10-03 18:55:47,816][starVLA.training.train_starvla][INFO] - Step 65, Loss: {'train/loss': 1.8154239654541016, 'train/grad_norm_pre_clip': 57.32501983642578, 'timing/data': 0.0003811499918811023, 'timing/model': 0.39733703699312173, 'learning_rate/qwen_vl_interface': 6.5000000000000004e-06, 'learning_rate/action_model': 6.500000000000001e-05, 'epoch': 0.0}) +[2026-10-03 18:55:48,524][starVLA.training.train_starvla][INFO] - Step 66, Loss: {'train/loss': 1.1084825992584229, 'train/grad_norm_pre_clip': 28.335601806640625, 'timing/data': 0.0002998200070578605, 'timing/model': 0.3953907669929322, 'learning_rate/qwen_vl_interface': 6.600000000000001e-06, 'learning_rate/action_model': 6.6e-05, 'epoch': 0.0}) +[2026-10-03 18:55:49,229][starVLA.training.train_starvla][INFO] - Step 67, Loss: {'train/loss': 1.9098373651504517, 'train/grad_norm_pre_clip': 60.080284118652344, 'timing/data': 0.0004906500107608736, 'timing/model': 0.39486906700767577, 'learning_rate/qwen_vl_interface': 6.700000000000001e-06, 'learning_rate/action_model': 6.7e-05, 'epoch': 0.0}) +[2026-10-03 18:55:49,932][starVLA.training.train_starvla][INFO] - Step 68, Loss: {'train/loss': 1.013007640838623, 'train/grad_norm_pre_clip': 51.70866012573242, 'timing/data': 0.00022256901138462126, 'timing/model': 0.39338509799563326, 'learning_rate/qwen_vl_interface': 6.800000000000001e-06, 'learning_rate/action_model': 6.800000000000001e-05, 'epoch': 0.0}) +[2026-10-03 18:55:50,638][starVLA.training.train_starvla][INFO] - Step 69, Loss: {'train/loss': 1.0689609050750732, 'train/grad_norm_pre_clip': 17.45017433166504, 'timing/data': 0.0002137400151696056, 'timing/model': 0.3964123670011759, 'learning_rate/qwen_vl_interface': 6.9e-06, 'learning_rate/action_model': 6.9e-05, 'epoch': 0.01}) +[2026-10-03 18:55:51,551][starVLA.training.train_starvla][INFO] - Step 70, Loss: {'train/loss': 1.4671359062194824, 'train/grad_norm_pre_clip': 34.08060836791992, 'timing/data': 0.0003106700023636222, 'timing/model': 0.39336579700466245, 'learning_rate/qwen_vl_interface': 7e-06, 'learning_rate/action_model': 7e-05, 'epoch': 0.01}) +[2026-10-03 18:55:52,256][starVLA.training.train_starvla][INFO] - Step 71, Loss: {'train/loss': 1.4032588005065918, 'train/grad_norm_pre_clip': 112.72706604003906, 'timing/data': 0.00022081000497564673, 'timing/model': 0.39289147799718194, 'learning_rate/qwen_vl_interface': 7.100000000000001e-06, 'learning_rate/action_model': 7.1e-05, 'epoch': 0.01}) +[2026-10-03 18:55:52,959][starVLA.training.train_starvla][INFO] - Step 72, Loss: {'train/loss': 1.4874346256256104, 'train/grad_norm_pre_clip': 62.2567253112793, 'timing/data': 0.0003194599994458258, 'timing/model': 0.3933069569757208, 'learning_rate/qwen_vl_interface': 7.2000000000000005e-06, 'learning_rate/action_model': 7.2e-05, 'epoch': 0.01}) +[2026-10-03 18:55:53,670][starVLA.training.train_starvla][INFO] - Step 73, Loss: {'train/loss': 1.172614336013794, 'train/grad_norm_pre_clip': 28.56315040588379, 'timing/data': 0.0003916399728041142, 'timing/model': 0.3978033869934734, 'learning_rate/qwen_vl_interface': 7.3e-06, 'learning_rate/action_model': 7.3e-05, 'epoch': 0.01}) +[2026-10-03 18:55:54,376][starVLA.training.train_starvla][INFO] - Step 74, Loss: {'train/loss': 1.1090208292007446, 'train/grad_norm_pre_clip': 57.93231964111328, 'timing/data': 0.0002296699967700988, 'timing/model': 0.39536806699470617, 'learning_rate/qwen_vl_interface': 7.4e-06, 'learning_rate/action_model': 7.4e-05, 'epoch': 0.01}) +[2026-10-03 18:55:55,277][starVLA.training.train_starvla][INFO] - Step 75, Loss: {'train/loss': 1.053713083267212, 'train/grad_norm_pre_clip': 255.7595672607422, 'timing/data': 0.00022203999105840921, 'timing/model': 0.5883142960083205, 'learning_rate/qwen_vl_interface': 7.500000000000001e-06, 'learning_rate/action_model': 7.500000000000001e-05, 'epoch': 0.01}) +[2026-10-03 18:55:55,981][starVLA.training.train_starvla][INFO] - Step 76, Loss: {'train/loss': 1.2716354131698608, 'train/grad_norm_pre_clip': 76.12156677246094, 'timing/data': 0.00022295000962913036, 'timing/model': 0.3933330569998361, 'learning_rate/qwen_vl_interface': 7.600000000000001e-06, 'learning_rate/action_model': 7.6e-05, 'epoch': 0.01}) +[2026-10-03 18:55:56,690][starVLA.training.train_starvla][INFO] - Step 77, Loss: {'train/loss': 1.0770689249038696, 'train/grad_norm_pre_clip': 110.74967193603516, 'timing/data': 0.000317530008032918, 'timing/model': 0.39751639700261876, 'learning_rate/qwen_vl_interface': 7.7e-06, 'learning_rate/action_model': 7.7e-05, 'epoch': 0.01}) +[2026-10-03 18:55:57,397][starVLA.training.train_starvla][INFO] - Step 78, Loss: {'train/loss': 1.0854015350341797, 'train/grad_norm_pre_clip': 43.87852478027344, 'timing/data': 0.00025896000443026423, 'timing/model': 0.3947413270070683, 'learning_rate/qwen_vl_interface': 7.800000000000002e-06, 'learning_rate/action_model': 7.800000000000001e-05, 'epoch': 0.01}) +[2026-10-03 18:55:58,102][starVLA.training.train_starvla][INFO] - Step 79, Loss: {'train/loss': 0.6656183004379272, 'train/grad_norm_pre_clip': 39.7022705078125, 'timing/data': 0.00025548000121489167, 'timing/model': 0.3933412469923496, 'learning_rate/qwen_vl_interface': 7.9e-06, 'learning_rate/action_model': 7.900000000000001e-05, 'epoch': 0.01}) +[2026-10-03 18:55:58,809][starVLA.training.train_starvla][INFO] - Step 80, Loss: {'train/loss': 0.5975042581558228, 'train/grad_norm_pre_clip': 22.676042556762695, 'timing/data': 0.0002225700009148568, 'timing/model': 0.39526934700552374, 'learning_rate/qwen_vl_interface': 8.000000000000001e-06, 'learning_rate/action_model': 8e-05, 'epoch': 0.01}) +[2026-10-03 18:55:59,717][starVLA.training.train_starvla][INFO] - Step 81, Loss: {'train/loss': 0.7543351650238037, 'train/grad_norm_pre_clip': 67.06368255615234, 'timing/data': 0.00030334899201989174, 'timing/model': 0.39742267699330114, 'learning_rate/qwen_vl_interface': 8.1e-06, 'learning_rate/action_model': 8.1e-05, 'epoch': 0.01}) +[2026-10-03 18:56:00,423][starVLA.training.train_starvla][INFO] - Step 82, Loss: {'train/loss': 1.1133644580841064, 'train/grad_norm_pre_clip': 75.22441864013672, 'timing/data': 0.0003699899825733155, 'timing/model': 0.39290389700909145, 'learning_rate/qwen_vl_interface': 8.2e-06, 'learning_rate/action_model': 8.2e-05, 'epoch': 0.01}) +[2026-10-03 18:56:01,128][starVLA.training.train_starvla][INFO] - Step 83, Loss: {'train/loss': 0.5723079442977905, 'train/grad_norm_pre_clip': 41.460121154785156, 'timing/data': 0.00032670999644324183, 'timing/model': 0.3942376780032646, 'learning_rate/qwen_vl_interface': 8.3e-06, 'learning_rate/action_model': 8.3e-05, 'epoch': 0.01}) +[2026-10-03 18:56:01,834][starVLA.training.train_starvla][INFO] - Step 84, Loss: {'train/loss': 0.509717583656311, 'train/grad_norm_pre_clip': 21.573951721191406, 'timing/data': 0.00022094001178629696, 'timing/model': 0.3940265869896393, 'learning_rate/qwen_vl_interface': 8.400000000000001e-06, 'learning_rate/action_model': 8.4e-05, 'epoch': 0.01}) +[2026-10-03 18:56:02,542][starVLA.training.train_starvla][INFO] - Step 85, Loss: {'train/loss': 0.7731541395187378, 'train/grad_norm_pre_clip': 50.20188903808594, 'timing/data': 0.0003240699879825115, 'timing/model': 0.3956473570142407, 'learning_rate/qwen_vl_interface': 8.5e-06, 'learning_rate/action_model': 8.5e-05, 'epoch': 0.01}) +[2026-10-03 18:56:03,445][starVLA.training.train_starvla][INFO] - Step 86, Loss: {'train/loss': 0.9223855137825012, 'train/grad_norm_pre_clip': 141.75018310546875, 'timing/data': 0.0002515199885237962, 'timing/model': 0.39488287700805813, 'learning_rate/qwen_vl_interface': 8.6e-06, 'learning_rate/action_model': 8.6e-05, 'epoch': 0.01}) +[2026-10-03 18:56:04,150][starVLA.training.train_starvla][INFO] - Step 87, Loss: {'train/loss': 0.7838836312294006, 'train/grad_norm_pre_clip': 122.07675170898438, 'timing/data': 0.00021254998864606023, 'timing/model': 0.39455348800402135, 'learning_rate/qwen_vl_interface': 8.700000000000001e-06, 'learning_rate/action_model': 8.7e-05, 'epoch': 0.01}) +[2026-10-03 18:56:04,856][starVLA.training.train_starvla][INFO] - Step 88, Loss: {'train/loss': 1.7859350442886353, 'train/grad_norm_pre_clip': 349.6282043457031, 'timing/data': 0.0002141699951607734, 'timing/model': 0.3936248279933352, 'learning_rate/qwen_vl_interface': 8.8e-06, 'learning_rate/action_model': 8.800000000000001e-05, 'epoch': 0.01}) +[2026-10-03 18:56:05,561][starVLA.training.train_starvla][INFO] - Step 89, Loss: {'train/loss': 1.3958544731140137, 'train/grad_norm_pre_clip': 174.93551635742188, 'timing/data': 0.0003789100155699998, 'timing/model': 0.39451352701871656, 'learning_rate/qwen_vl_interface': 8.900000000000001e-06, 'learning_rate/action_model': 8.900000000000001e-05, 'epoch': 0.01}) +[2026-10-03 18:56:06,266][starVLA.training.train_starvla][INFO] - Step 90, Loss: {'train/loss': 1.0025105476379395, 'train/grad_norm_pre_clip': 123.791259765625, 'timing/data': 0.00021120000747032464, 'timing/model': 0.3941044469829649, 'learning_rate/qwen_vl_interface': 9e-06, 'learning_rate/action_model': 9e-05, 'epoch': 0.01}) +[2026-10-03 18:56:07,160][starVLA.training.train_starvla][INFO] - Step 91, Loss: {'train/loss': 1.262411117553711, 'train/grad_norm_pre_clip': 158.37152099609375, 'timing/data': 0.00020710998796857893, 'timing/model': 0.5857675060105976, 'learning_rate/qwen_vl_interface': 9.100000000000001e-06, 'learning_rate/action_model': 9.1e-05, 'epoch': 0.01}) +[2026-10-03 18:56:07,864][starVLA.training.train_starvla][INFO] - Step 92, Loss: {'train/loss': 1.1401866674423218, 'train/grad_norm_pre_clip': 170.7518768310547, 'timing/data': 0.00020820999634452164, 'timing/model': 0.3947989969747141, 'learning_rate/qwen_vl_interface': 9.200000000000002e-06, 'learning_rate/action_model': 9.200000000000001e-05, 'epoch': 0.01}) +[2026-10-03 18:56:08,572][starVLA.training.train_starvla][INFO] - Step 93, Loss: {'train/loss': 0.6599053144454956, 'train/grad_norm_pre_clip': 37.37132263183594, 'timing/data': 0.0002730200067162514, 'timing/model': 0.39698538702214137, 'learning_rate/qwen_vl_interface': 9.3e-06, 'learning_rate/action_model': 9.300000000000001e-05, 'epoch': 0.01}) +[2026-10-03 18:56:09,278][starVLA.training.train_starvla][INFO] - Step 94, Loss: {'train/loss': 0.755685031414032, 'train/grad_norm_pre_clip': 220.91671752929688, 'timing/data': 0.00022839001030661166, 'timing/model': 0.39407400699565187, 'learning_rate/qwen_vl_interface': 9.4e-06, 'learning_rate/action_model': 9.4e-05, 'epoch': 0.01}) +[2026-10-03 18:56:09,985][starVLA.training.train_starvla][INFO] - Step 95, Loss: {'train/loss': 0.8120182156562805, 'train/grad_norm_pre_clip': 129.0868682861328, 'timing/data': 0.00023022000095807016, 'timing/model': 0.39519556699087843, 'learning_rate/qwen_vl_interface': 9.5e-06, 'learning_rate/action_model': 9.5e-05, 'epoch': 0.01}) +[2026-10-03 18:56:10,691][starVLA.training.train_starvla][INFO] - Step 96, Loss: {'train/loss': 1.433550238609314, 'train/grad_norm_pre_clip': 178.1361083984375, 'timing/data': 0.00022828998044133186, 'timing/model': 0.3947618270176463, 'learning_rate/qwen_vl_interface': 9.600000000000001e-06, 'learning_rate/action_model': 9.6e-05, 'epoch': 0.01}) +[2026-10-03 18:56:11,596][starVLA.training.train_starvla][INFO] - Step 97, Loss: {'train/loss': 1.3245041370391846, 'train/grad_norm_pre_clip': 403.0451354980469, 'timing/data': 0.00022995000472292304, 'timing/model': 0.3959822670149151, 'learning_rate/qwen_vl_interface': 9.7e-06, 'learning_rate/action_model': 9.7e-05, 'epoch': 0.01}) +[2026-10-03 18:56:12,302][starVLA.training.train_starvla][INFO] - Step 98, Loss: {'train/loss': 1.294966220855713, 'train/grad_norm_pre_clip': 392.28253173828125, 'timing/data': 0.00022126000840216875, 'timing/model': 0.3931143270165194, 'learning_rate/qwen_vl_interface': 9.800000000000001e-06, 'learning_rate/action_model': 9.8e-05, 'epoch': 0.01}) +[2026-10-03 18:56:13,009][starVLA.training.train_starvla][INFO] - Step 99, Loss: {'train/loss': 1.2648942470550537, 'train/grad_norm_pre_clip': 85.85420989990234, 'timing/data': 0.0003470099763944745, 'timing/model': 0.393436557991663, 'learning_rate/qwen_vl_interface': 9.9e-06, 'learning_rate/action_model': 9.900000000000001e-05, 'epoch': 0.01}) +[2026-10-03 18:56:13,714][starVLA.training.train_starvla][INFO] - Step 100, Loss: {'train/loss': 0.8904734253883362, 'train/grad_norm_pre_clip': 73.54515838623047, 'timing/data': 0.0002264800132252276, 'timing/model': 0.3939264469954651, 'learning_rate/qwen_vl_interface': 1e-05, 'learning_rate/action_model': 0.0001, 'epoch': 0.01}) +[2026-10-03 18:56:14,419][starVLA.training.train_starvla][INFO] - Step 101, Loss: {'train/loss': 1.0782759189605713, 'train/grad_norm_pre_clip': 104.0731201171875, 'timing/data': 0.00022478000028058887, 'timing/model': 0.3949179070186801, 'learning_rate/qwen_vl_interface': 9.999999023727212e-06, 'learning_rate/action_model': 9.99999902372721e-05, 'epoch': 0.01}) +[2026-10-03 18:56:15,322][starVLA.training.train_starvla][INFO] - Step 102, Loss: {'train/loss': 1.203352451324463, 'train/grad_norm_pre_clip': 85.52523803710938, 'timing/data': 0.00022171999444253743, 'timing/model': 0.39489732700167224, 'learning_rate/qwen_vl_interface': 9.999996094909247e-06, 'learning_rate/action_model': 9.999996094909248e-05, 'epoch': 0.01}) +[2026-10-03 18:56:16,001][starVLA.training.train_starvla][INFO] - Step 103, Loss: {'train/loss': 0.8324056267738342, 'train/grad_norm_pre_clip': 220.42572021484375, 'timing/data': 0.00022533998708240688, 'timing/model': 0.368720630009193, 'learning_rate/qwen_vl_interface': 9.999991213547314e-06, 'learning_rate/action_model': 9.999991213547313e-05, 'epoch': 0.01}) +[2026-10-03 18:56:16,707][starVLA.training.train_starvla][INFO] - Step 104, Loss: {'train/loss': 0.9505987167358398, 'train/grad_norm_pre_clip': 131.8980255126953, 'timing/data': 0.00030289997812360525, 'timing/model': 0.39406433698604815, 'learning_rate/qwen_vl_interface': 9.999984379643413e-06, 'learning_rate/action_model': 9.999984379643412e-05, 'epoch': 0.01}) +[2026-10-03 18:56:17,416][starVLA.training.train_starvla][INFO] - Step 105, Loss: {'train/loss': 1.3721610307693481, 'train/grad_norm_pre_clip': 109.85579681396484, 'timing/data': 0.00023155001690611243, 'timing/model': 0.39702497801044956, 'learning_rate/qwen_vl_interface': 9.999975593200357e-06, 'learning_rate/action_model': 9.999975593200357e-05, 'epoch': 0.01}) +[2026-10-03 18:56:18,123][starVLA.training.train_starvla][INFO] - Step 106, Loss: {'train/loss': 0.7381305694580078, 'train/grad_norm_pre_clip': 49.437477111816406, 'timing/data': 0.0002198000147473067, 'timing/model': 0.39499263701145537, 'learning_rate/qwen_vl_interface': 9.999964854221756e-06, 'learning_rate/action_model': 9.999964854221754e-05, 'epoch': 0.01}) +[2026-10-03 18:56:19,026][starVLA.training.train_starvla][INFO] - Step 107, Loss: {'train/loss': 1.6303036212921143, 'train/grad_norm_pre_clip': 59.887969970703125, 'timing/data': 0.00022198999067768455, 'timing/model': 0.3936092169897165, 'learning_rate/qwen_vl_interface': 9.999952162712027e-06, 'learning_rate/action_model': 9.999952162712026e-05, 'epoch': 0.01}) +[2026-10-03 18:56:19,731][starVLA.training.train_starvla][INFO] - Step 108, Loss: {'train/loss': 1.1955125331878662, 'train/grad_norm_pre_clip': 17.502531051635742, 'timing/data': 0.0002289700205437839, 'timing/model': 0.3929921470116824, 'learning_rate/qwen_vl_interface': 9.999937518676382e-06, 'learning_rate/action_model': 9.999937518676382e-05, 'epoch': 0.01}) +[2026-10-03 18:56:20,437][starVLA.training.train_starvla][INFO] - Step 109, Loss: {'train/loss': 0.9825674295425415, 'train/grad_norm_pre_clip': 30.70696258544922, 'timing/data': 0.00026962000993080437, 'timing/model': 0.39476700802333653, 'learning_rate/qwen_vl_interface': 9.999920922120846e-06, 'learning_rate/action_model': 9.999920922120846e-05, 'epoch': 0.01}) +[2026-10-03 18:56:21,143][starVLA.training.train_starvla][INFO] - Step 110, Loss: {'train/loss': 1.085928201675415, 'train/grad_norm_pre_clip': 169.4293975830078, 'timing/data': 0.00022714000078849494, 'timing/model': 0.3948834270122461, 'learning_rate/qwen_vl_interface': 9.999902373052239e-06, 'learning_rate/action_model': 9.999902373052239e-05, 'epoch': 0.01}) +[2026-10-03 18:56:21,851][starVLA.training.train_starvla][INFO] - Step 111, Loss: {'train/loss': 0.7388827800750732, 'train/grad_norm_pre_clip': 157.67149353027344, 'timing/data': 0.00023864000104367733, 'timing/model': 0.39534026800538413, 'learning_rate/qwen_vl_interface': 9.999881871478188e-06, 'learning_rate/action_model': 9.999881871478187e-05, 'epoch': 0.01}) +[2026-10-03 18:56:22,763][starVLA.training.train_starvla][INFO] - Step 112, Loss: {'train/loss': 1.2453405857086182, 'train/grad_norm_pre_clip': 153.56039428710938, 'timing/data': 0.00022671997430734336, 'timing/model': 0.6006494750035927, 'learning_rate/qwen_vl_interface': 9.999859417407115e-06, 'learning_rate/action_model': 9.999859417407115e-05, 'epoch': 0.01}) +[2026-10-03 18:56:23,469][starVLA.training.train_starvla][INFO] - Step 113, Loss: {'train/loss': 1.0799696445465088, 'train/grad_norm_pre_clip': 63.950096130371094, 'timing/data': 0.0002425599959678948, 'timing/model': 0.395558467018418, 'learning_rate/qwen_vl_interface': 9.999835010848254e-06, 'learning_rate/action_model': 9.999835010848254e-05, 'epoch': 0.01}) +[2026-10-03 18:56:24,176][starVLA.training.train_starvla][INFO] - Step 114, Loss: {'train/loss': 0.8076146841049194, 'train/grad_norm_pre_clip': 56.83189010620117, 'timing/data': 0.0002205700147897005, 'timing/model': 0.3960497369989753, 'learning_rate/qwen_vl_interface': 9.999808651811639e-06, 'learning_rate/action_model': 9.999808651811639e-05, 'epoch': 0.01}) +[2026-10-03 18:56:24,881][starVLA.training.train_starvla][INFO] - Step 115, Loss: {'train/loss': 1.265550971031189, 'train/grad_norm_pre_clip': 127.83312225341797, 'timing/data': 0.0003076189896091819, 'timing/model': 0.39315519799129106, 'learning_rate/qwen_vl_interface': 9.999780340308101e-06, 'learning_rate/action_model': 9.999780340308101e-05, 'epoch': 0.01}) +[2026-10-03 18:56:25,589][starVLA.training.train_starvla][INFO] - Step 116, Loss: {'train/loss': 1.1499090194702148, 'train/grad_norm_pre_clip': 65.31336212158203, 'timing/data': 0.00036174998967908323, 'timing/model': 0.3960731470142491, 'learning_rate/qwen_vl_interface': 9.999750076349282e-06, 'learning_rate/action_model': 9.999750076349282e-05, 'epoch': 0.01}) +[2026-10-03 18:56:26,298][starVLA.training.train_starvla][INFO] - Step 117, Loss: {'train/loss': 0.7160210609436035, 'train/grad_norm_pre_clip': 73.61445617675781, 'timing/data': 0.0002819299988914281, 'timing/model': 0.39568467799108475, 'learning_rate/qwen_vl_interface': 9.999717859947618e-06, 'learning_rate/action_model': 9.999717859947618e-05, 'epoch': 0.01}) +[2026-10-03 18:56:27,202][starVLA.training.train_starvla][INFO] - Step 118, Loss: {'train/loss': 1.077783465385437, 'train/grad_norm_pre_clip': 77.22748565673828, 'timing/data': 0.0004587199946399778, 'timing/model': 0.3922128470148891, 'learning_rate/qwen_vl_interface': 9.999683691116356e-06, 'learning_rate/action_model': 9.999683691116355e-05, 'epoch': 0.01}) +[2026-10-03 18:56:27,909][starVLA.training.train_starvla][INFO] - Step 119, Loss: {'train/loss': 0.8810418844223022, 'train/grad_norm_pre_clip': 114.06327819824219, 'timing/data': 0.0003010100044775754, 'timing/model': 0.39554359699832276, 'learning_rate/qwen_vl_interface': 9.999647569869539e-06, 'learning_rate/action_model': 9.99964756986954e-05, 'epoch': 0.01}) +[2026-10-03 18:56:28,613][starVLA.training.train_starvla][INFO] - Step 120, Loss: {'train/loss': 0.845542848110199, 'train/grad_norm_pre_clip': 157.54747009277344, 'timing/data': 0.00022712998907081783, 'timing/model': 0.3940498169977218, 'learning_rate/qwen_vl_interface': 9.999609496222017e-06, 'learning_rate/action_model': 9.999609496222017e-05, 'epoch': 0.01}) +[2026-10-03 18:56:29,322][starVLA.training.train_starvla][INFO] - Step 121, Loss: {'train/loss': 1.0210175514221191, 'train/grad_norm_pre_clip': 91.74131774902344, 'timing/data': 0.0002602099848445505, 'timing/model': 0.3972107669978868, 'learning_rate/qwen_vl_interface': 9.99956947018944e-06, 'learning_rate/action_model': 9.999569470189439e-05, 'epoch': 0.01}) +[2026-10-03 18:56:30,027][starVLA.training.train_starvla][INFO] - Step 122, Loss: {'train/loss': 1.2506535053253174, 'train/grad_norm_pre_clip': 149.6337890625, 'timing/data': 0.000229049997869879, 'timing/model': 0.39326760699623264, 'learning_rate/qwen_vl_interface': 9.99952749178826e-06, 'learning_rate/action_model': 9.999527491788259e-05, 'epoch': 0.01}) +[2026-10-03 18:56:30,932][starVLA.training.train_starvla][INFO] - Step 123, Loss: {'train/loss': 1.742193579673767, 'train/grad_norm_pre_clip': 219.38247680664062, 'timing/data': 0.00035357000888325274, 'timing/model': 0.39471684698946774, 'learning_rate/qwen_vl_interface': 9.999483561035732e-06, 'learning_rate/action_model': 9.999483561035732e-05, 'epoch': 0.01}) +[2026-10-03 18:56:31,638][starVLA.training.train_starvla][INFO] - Step 124, Loss: {'train/loss': 1.783390998840332, 'train/grad_norm_pre_clip': 202.42481994628906, 'timing/data': 0.00022757000988349319, 'timing/model': 0.39402837699162774, 'learning_rate/qwen_vl_interface': 9.999437677949917e-06, 'learning_rate/action_model': 9.999437677949917e-05, 'epoch': 0.01}) +[2026-10-03 18:56:32,346][starVLA.training.train_starvla][INFO] - Step 125, Loss: {'train/loss': 0.9959450960159302, 'train/grad_norm_pre_clip': 52.03142547607422, 'timing/data': 0.0002680300094652921, 'timing/model': 0.39615066701662727, 'learning_rate/qwen_vl_interface': 9.999389842549674e-06, 'learning_rate/action_model': 9.999389842549673e-05, 'epoch': 0.01}) +[2026-10-03 18:56:33,051][starVLA.training.train_starvla][INFO] - Step 126, Loss: {'train/loss': 1.465577483177185, 'train/grad_norm_pre_clip': 220.030029296875, 'timing/data': 0.00022464999346993864, 'timing/model': 0.3932164370198734, 'learning_rate/qwen_vl_interface': 9.999340054854669e-06, 'learning_rate/action_model': 9.999340054854667e-05, 'epoch': 0.01}) +[2026-10-03 18:56:33,758][starVLA.training.train_starvla][INFO] - Step 127, Loss: {'train/loss': 1.0198554992675781, 'train/grad_norm_pre_clip': 155.9659881591797, 'timing/data': 0.00022128000273369253, 'timing/model': 0.3958979180024471, 'learning_rate/qwen_vl_interface': 9.999288314885362e-06, 'learning_rate/action_model': 9.999288314885362e-05, 'epoch': 0.01}) +[2026-10-03 18:56:34,666][starVLA.training.train_starvla][INFO] - Step 128, Loss: {'train/loss': 2.2069215774536133, 'train/grad_norm_pre_clip': 343.6889343261719, 'timing/data': 0.00036007000016979873, 'timing/model': 0.5957456259930041, 'learning_rate/qwen_vl_interface': 9.999234622663027e-06, 'learning_rate/action_model': 9.999234622663028e-05, 'epoch': 0.01}) +[2026-10-03 18:56:35,372][starVLA.training.train_starvla][INFO] - Step 129, Loss: {'train/loss': 0.650088906288147, 'train/grad_norm_pre_clip': 109.3085708618164, 'timing/data': 0.00022751998039893806, 'timing/model': 0.3947998270159587, 'learning_rate/qwen_vl_interface': 9.999178978209732e-06, 'learning_rate/action_model': 9.999178978209732e-05, 'epoch': 0.01}) +[2026-10-03 18:56:36,077][starVLA.training.train_starvla][INFO] - Step 130, Loss: {'train/loss': 0.6724405884742737, 'train/grad_norm_pre_clip': 29.940685272216797, 'timing/data': 0.00022588000865653157, 'timing/model': 0.39323329797480255, 'learning_rate/qwen_vl_interface': 9.999121381548354e-06, 'learning_rate/action_model': 9.999121381548353e-05, 'epoch': 0.01}) +[2026-10-03 18:56:36,783][starVLA.training.train_starvla][INFO] - Step 131, Loss: {'train/loss': 2.552252769470215, 'train/grad_norm_pre_clip': 424.41680908203125, 'timing/data': 0.0002376000047661364, 'timing/model': 0.39365853799972683, 'learning_rate/qwen_vl_interface': 9.999061832702563e-06, 'learning_rate/action_model': 9.999061832702563e-05, 'epoch': 0.01}) +[2026-10-03 18:56:37,489][starVLA.training.train_starvla][INFO] - Step 132, Loss: {'train/loss': 2.4575071334838867, 'train/grad_norm_pre_clip': 418.8062744140625, 'timing/data': 0.0002303500077687204, 'timing/model': 0.3954884470149409, 'learning_rate/qwen_vl_interface': 9.999000331696842e-06, 'learning_rate/action_model': 9.999000331696841e-05, 'epoch': 0.01}) +[2026-10-03 18:56:38,395][starVLA.training.train_starvla][INFO] - Step 133, Loss: {'train/loss': 2.849900007247925, 'train/grad_norm_pre_clip': 385.82647705078125, 'timing/data': 0.00026749001699499786, 'timing/model': 0.5951816460001282, 'learning_rate/qwen_vl_interface': 9.99893687855647e-06, 'learning_rate/action_model': 9.99893687855647e-05, 'epoch': 0.01}) +[2026-10-03 18:56:39,103][starVLA.training.train_starvla][INFO] - Step 134, Loss: {'train/loss': 1.9236443042755127, 'train/grad_norm_pre_clip': 270.0727233886719, 'timing/data': 0.00021963001927360892, 'timing/model': 0.39512368699070066, 'learning_rate/qwen_vl_interface': 9.998871473307529e-06, 'learning_rate/action_model': 9.99887147330753e-05, 'epoch': 0.01}) +[2026-10-03 18:56:39,808][starVLA.training.train_starvla][INFO] - Step 135, Loss: {'train/loss': 1.6036698818206787, 'train/grad_norm_pre_clip': 261.1297302246094, 'timing/data': 0.0003316999936942011, 'timing/model': 0.39499699699808843, 'learning_rate/qwen_vl_interface': 9.998804115976907e-06, 'learning_rate/action_model': 9.998804115976908e-05, 'epoch': 0.01}) +[2026-10-03 18:56:40,515][starVLA.training.train_starvla][INFO] - Step 136, Loss: {'train/loss': 0.7610495090484619, 'train/grad_norm_pre_clip': 49.66461181640625, 'timing/data': 0.0002255300059914589, 'timing/model': 0.3950598869996611, 'learning_rate/qwen_vl_interface': 9.99873480659229e-06, 'learning_rate/action_model': 9.99873480659229e-05, 'epoch': 0.01}) +[2026-10-03 18:56:41,222][starVLA.training.train_starvla][INFO] - Step 137, Loss: {'train/loss': 0.686943769454956, 'train/grad_norm_pre_clip': 67.32028198242188, 'timing/data': 0.00027867997414432466, 'timing/model': 0.39434084700769745, 'learning_rate/qwen_vl_interface': 9.99866354518217e-06, 'learning_rate/action_model': 9.99866354518217e-05, 'epoch': 0.01}) +[2026-10-03 18:56:42,127][starVLA.training.train_starvla][INFO] - Step 138, Loss: {'train/loss': 0.8775046467781067, 'train/grad_norm_pre_clip': 34.80305862426758, 'timing/data': 0.0004809400124941021, 'timing/model': 0.5934594449936412, 'learning_rate/qwen_vl_interface': 9.99859033177584e-06, 'learning_rate/action_model': 9.99859033177584e-05, 'epoch': 0.01}) +[2026-10-03 18:56:42,833][starVLA.training.train_starvla][INFO] - Step 139, Loss: {'train/loss': 0.6058406829833984, 'train/grad_norm_pre_clip': 41.23152542114258, 'timing/data': 0.00022590000298805535, 'timing/model': 0.3948775070020929, 'learning_rate/qwen_vl_interface': 9.998515166403394e-06, 'learning_rate/action_model': 9.998515166403394e-05, 'epoch': 0.01}) +[2026-10-03 18:56:43,540][starVLA.training.train_starvla][INFO] - Step 140, Loss: {'train/loss': 1.202667474746704, 'train/grad_norm_pre_clip': 62.96140670776367, 'timing/data': 0.0003196400066372007, 'timing/model': 0.39526266799657606, 'learning_rate/qwen_vl_interface': 9.99843804909573e-06, 'learning_rate/action_model': 9.99843804909573e-05, 'epoch': 0.01}) +[2026-10-03 18:56:44,246][starVLA.training.train_starvla][INFO] - Step 141, Loss: {'train/loss': 1.0652720928192139, 'train/grad_norm_pre_clip': 59.24611282348633, 'timing/data': 0.0002723500074353069, 'timing/model': 0.3950115279876627, 'learning_rate/qwen_vl_interface': 9.998358979884548e-06, 'learning_rate/action_model': 9.998358979884547e-05, 'epoch': 0.01}) +[2026-10-03 18:56:44,953][starVLA.training.train_starvla][INFO] - Step 142, Loss: {'train/loss': 0.5868938565254211, 'train/grad_norm_pre_clip': 70.27610778808594, 'timing/data': 0.00022077999892644584, 'timing/model': 0.3946347270102706, 'learning_rate/qwen_vl_interface': 9.998277958802352e-06, 'learning_rate/action_model': 9.998277958802351e-05, 'epoch': 0.01}) +[2026-10-03 18:56:45,658][starVLA.training.train_starvla][INFO] - Step 143, Loss: {'train/loss': 1.1091099977493286, 'train/grad_norm_pre_clip': 70.69609832763672, 'timing/data': 0.00021186997764743865, 'timing/model': 0.39367441699141636, 'learning_rate/qwen_vl_interface': 9.998194985882442e-06, 'learning_rate/action_model': 9.998194985882442e-05, 'epoch': 0.01}) +[2026-10-03 18:56:46,563][starVLA.training.train_starvla][INFO] - Step 144, Loss: {'train/loss': 1.0492699146270752, 'train/grad_norm_pre_clip': 64.11244201660156, 'timing/data': 0.0002267800155095756, 'timing/model': 0.39532131800660864, 'learning_rate/qwen_vl_interface': 9.99811006115893e-06, 'learning_rate/action_model': 9.998110061158931e-05, 'epoch': 0.01}) +[2026-10-03 18:56:47,269][starVLA.training.train_starvla][INFO] - Step 145, Loss: {'train/loss': 1.5301978588104248, 'train/grad_norm_pre_clip': 97.2078628540039, 'timing/data': 0.00037294000503607094, 'timing/model': 0.39612696701078676, 'learning_rate/qwen_vl_interface': 9.998023184666726e-06, 'learning_rate/action_model': 9.998023184666726e-05, 'epoch': 0.01}) +[2026-10-03 18:56:47,973][starVLA.training.train_starvla][INFO] - Step 146, Loss: {'train/loss': 0.7483289837837219, 'train/grad_norm_pre_clip': 49.614540100097656, 'timing/data': 0.0002217399887740612, 'timing/model': 0.39386901701800525, 'learning_rate/qwen_vl_interface': 9.997934356441538e-06, 'learning_rate/action_model': 9.997934356441537e-05, 'epoch': 0.01}) +[2026-10-03 18:56:48,678][starVLA.training.train_starvla][INFO] - Step 147, Loss: {'train/loss': 0.5001579523086548, 'train/grad_norm_pre_clip': 183.6197967529297, 'timing/data': 0.0003067500074394047, 'timing/model': 0.39537080700392835, 'learning_rate/qwen_vl_interface': 9.997843576519879e-06, 'learning_rate/action_model': 9.997843576519878e-05, 'epoch': 0.01}) +[2026-10-03 18:56:49,385][starVLA.training.train_starvla][INFO] - Step 148, Loss: {'train/loss': 0.9633371829986572, 'train/grad_norm_pre_clip': 166.9734344482422, 'timing/data': 0.00022331002401188016, 'timing/model': 0.39585578700643964, 'learning_rate/qwen_vl_interface': 9.99775084493907e-06, 'learning_rate/action_model': 9.997750844939069e-05, 'epoch': 0.01}) +[2026-10-03 18:56:50,291][starVLA.training.train_starvla][INFO] - Step 149, Loss: {'train/loss': 0.9680429697036743, 'train/grad_norm_pre_clip': 264.52532958984375, 'timing/data': 0.00022715999512001872, 'timing/model': 0.39425698699778877, 'learning_rate/qwen_vl_interface': 9.997656161737227e-06, 'learning_rate/action_model': 9.997656161737226e-05, 'epoch': 0.01}) +[2026-10-03 18:56:50,998][starVLA.training.train_starvla][INFO] - Step 150, Loss: {'train/loss': 1.1038599014282227, 'train/grad_norm_pre_clip': 57.873435974121094, 'timing/data': 0.00022302998695522547, 'timing/model': 0.39500271700671874, 'learning_rate/qwen_vl_interface': 9.99755952695327e-06, 'learning_rate/action_model': 9.997559526953269e-05, 'epoch': 0.01}) +[2026-10-03 18:56:51,704][starVLA.training.train_starvla][INFO] - Step 151, Loss: {'train/loss': 0.609681248664856, 'train/grad_norm_pre_clip': 28.31294822692871, 'timing/data': 0.00022662998526357114, 'timing/model': 0.39541260700207204, 'learning_rate/qwen_vl_interface': 9.99746094062692e-06, 'learning_rate/action_model': 9.99746094062692e-05, 'epoch': 0.01}) +[2026-10-03 18:56:52,410][starVLA.training.train_starvla][INFO] - Step 152, Loss: {'train/loss': 0.6622722148895264, 'train/grad_norm_pre_clip': 24.39557647705078, 'timing/data': 0.00023951998446136713, 'timing/model': 0.39407541701802984, 'learning_rate/qwen_vl_interface': 9.997360402798706e-06, 'learning_rate/action_model': 9.997360402798705e-05, 'epoch': 0.01}) +[2026-10-03 18:56:53,118][starVLA.training.train_starvla][INFO] - Step 153, Loss: {'train/loss': 0.6491495966911316, 'train/grad_norm_pre_clip': 53.68665313720703, 'timing/data': 0.0003378300170879811, 'timing/model': 0.39662812699680217, 'learning_rate/qwen_vl_interface': 9.997257913509955e-06, 'learning_rate/action_model': 9.997257913509955e-05, 'epoch': 0.01}) +[2026-10-03 18:56:54,019][starVLA.training.train_starvla][INFO] - Step 154, Loss: {'train/loss': 0.7860960364341736, 'train/grad_norm_pre_clip': 54.1734504699707, 'timing/data': 0.00023366999812424183, 'timing/model': 0.5893765359942336, 'learning_rate/qwen_vl_interface': 9.997153472802794e-06, 'learning_rate/action_model': 9.997153472802795e-05, 'epoch': 0.01}) +[2026-10-03 18:56:54,723][starVLA.training.train_starvla][INFO] - Step 155, Loss: {'train/loss': 1.0578309297561646, 'train/grad_norm_pre_clip': 44.63021469116211, 'timing/data': 0.00022539999918080866, 'timing/model': 0.3936543879972305, 'learning_rate/qwen_vl_interface': 9.997047080720157e-06, 'learning_rate/action_model': 9.997047080720158e-05, 'epoch': 0.01}) +[2026-10-03 18:56:55,430][starVLA.training.train_starvla][INFO] - Step 156, Loss: {'train/loss': 0.7083994150161743, 'train/grad_norm_pre_clip': 56.99239730834961, 'timing/data': 0.0002245499927084893, 'timing/model': 0.3950117380009033, 'learning_rate/qwen_vl_interface': 9.996938737305777e-06, 'learning_rate/action_model': 9.996938737305776e-05, 'epoch': 0.01}) +[2026-10-03 18:56:56,138][starVLA.training.train_starvla][INFO] - Step 157, Loss: {'train/loss': 0.31248190999031067, 'train/grad_norm_pre_clip': 115.8623046875, 'timing/data': 0.0003321699914522469, 'timing/model': 0.3958544670022093, 'learning_rate/qwen_vl_interface': 9.996828442604189e-06, 'learning_rate/action_model': 9.99682844260419e-05, 'epoch': 0.01}) +[2026-10-03 18:56:56,846][starVLA.training.train_starvla][INFO] - Step 158, Loss: {'train/loss': 0.63639235496521, 'train/grad_norm_pre_clip': 52.68455123901367, 'timing/data': 0.0002321199863217771, 'timing/model': 0.3950416469888296, 'learning_rate/qwen_vl_interface': 9.996716196660733e-06, 'learning_rate/action_model': 9.996716196660731e-05, 'epoch': 0.01}) +[2026-10-03 18:56:57,752][starVLA.training.train_starvla][INFO] - Step 159, Loss: {'train/loss': 0.4355579614639282, 'train/grad_norm_pre_clip': 172.47206115722656, 'timing/data': 0.00021887998445890844, 'timing/model': 0.595887406001566, 'learning_rate/qwen_vl_interface': 9.996601999521546e-06, 'learning_rate/action_model': 9.996601999521546e-05, 'epoch': 0.01}) +[2026-10-03 18:56:58,456][starVLA.training.train_starvla][INFO] - Step 160, Loss: {'train/loss': 2.3837316036224365, 'train/grad_norm_pre_clip': 322.7324523925781, 'timing/data': 0.00022683999850414693, 'timing/model': 0.39353272700100206, 'learning_rate/qwen_vl_interface': 9.996485851233572e-06, 'learning_rate/action_model': 9.996485851233572e-05, 'epoch': 0.01}) +[2026-10-03 18:56:59,163][starVLA.training.train_starvla][INFO] - Step 161, Loss: {'train/loss': 1.1854033470153809, 'train/grad_norm_pre_clip': 226.34364318847656, 'timing/data': 0.0002856499922927469, 'timing/model': 0.3953843869967386, 'learning_rate/qwen_vl_interface': 9.996367751844556e-06, 'learning_rate/action_model': 9.996367751844555e-05, 'epoch': 0.01}) +[2026-10-03 18:56:59,869][starVLA.training.train_starvla][INFO] - Step 162, Loss: {'train/loss': 1.3497858047485352, 'train/grad_norm_pre_clip': 304.9833984375, 'timing/data': 0.0003166200185660273, 'timing/model': 0.39359146798960865, 'learning_rate/qwen_vl_interface': 9.996247701403043e-06, 'learning_rate/action_model': 9.996247701403042e-05, 'epoch': 0.01}) +[2026-10-03 18:57:00,577][starVLA.training.train_starvla][INFO] - Step 163, Loss: {'train/loss': 1.746792197227478, 'train/grad_norm_pre_clip': 109.42940521240234, 'timing/data': 0.00026178002008236945, 'timing/model': 0.3961914469837211, 'learning_rate/qwen_vl_interface': 9.99612569995838e-06, 'learning_rate/action_model': 9.99612569995838e-05, 'epoch': 0.01}) +[2026-10-03 18:57:01,489][starVLA.training.train_starvla][INFO] - Step 164, Loss: {'train/loss': 0.6286118626594543, 'train/grad_norm_pre_clip': 75.38379669189453, 'timing/data': 0.00037717001396231353, 'timing/model': 0.5997353349812329, 'learning_rate/qwen_vl_interface': 9.99600174756072e-06, 'learning_rate/action_model': 9.99600174756072e-05, 'epoch': 0.01}) +[2026-10-03 18:57:02,197][starVLA.training.train_starvla][INFO] - Step 165, Loss: {'train/loss': 0.7471814751625061, 'train/grad_norm_pre_clip': 56.624107360839844, 'timing/data': 0.0002994700043927878, 'timing/model': 0.39592639697366394, 'learning_rate/qwen_vl_interface': 9.995875844261012e-06, 'learning_rate/action_model': 9.995875844261013e-05, 'epoch': 0.01}) +[2026-10-03 18:57:02,904][starVLA.training.train_starvla][INFO] - Step 166, Loss: {'train/loss': 0.8213776350021362, 'train/grad_norm_pre_clip': 87.31420135498047, 'timing/data': 0.00022660999093204737, 'timing/model': 0.39467670701560564, 'learning_rate/qwen_vl_interface': 9.995747990111015e-06, 'learning_rate/action_model': 9.995747990111013e-05, 'epoch': 0.01}) +[2026-10-03 18:57:03,610][starVLA.training.train_starvla][INFO] - Step 167, Loss: {'train/loss': 1.2294080257415771, 'train/grad_norm_pre_clip': 141.72743225097656, 'timing/data': 0.00023324001813307405, 'timing/model': 0.3954127570032142, 'learning_rate/qwen_vl_interface': 9.99561818516328e-06, 'learning_rate/action_model': 9.99561818516328e-05, 'epoch': 0.01}) +[2026-10-03 18:57:04,317][starVLA.training.train_starvla][INFO] - Step 168, Loss: {'train/loss': 1.5854780673980713, 'train/grad_norm_pre_clip': 140.49002075195312, 'timing/data': 0.00025454998831264675, 'timing/model': 0.39479150701663457, 'learning_rate/qwen_vl_interface': 9.995486429471166e-06, 'learning_rate/action_model': 9.995486429471166e-05, 'epoch': 0.01}) +[2026-10-03 18:57:05,025][starVLA.training.train_starvla][INFO] - Step 169, Loss: {'train/loss': 0.5868701934814453, 'train/grad_norm_pre_clip': 44.7994499206543, 'timing/data': 0.00027663001674227417, 'timing/model': 0.39645494701107964, 'learning_rate/qwen_vl_interface': 9.995352723088833e-06, 'learning_rate/action_model': 9.995352723088834e-05, 'epoch': 0.01}) +[2026-10-03 18:57:05,939][starVLA.training.train_starvla][INFO] - Step 170, Loss: {'train/loss': 1.1737747192382812, 'train/grad_norm_pre_clip': 175.72703552246094, 'timing/data': 0.00024151999969035387, 'timing/model': 0.3953810980019625, 'learning_rate/qwen_vl_interface': 9.995217066071245e-06, 'learning_rate/action_model': 9.995217066071246e-05, 'epoch': 0.01}) +[2026-10-03 18:57:06,644][starVLA.training.train_starvla][INFO] - Step 171, Loss: {'train/loss': 1.4622700214385986, 'train/grad_norm_pre_clip': 93.64916229248047, 'timing/data': 0.00022320999414660037, 'timing/model': 0.3953192580083851, 'learning_rate/qwen_vl_interface': 9.995079458474163e-06, 'learning_rate/action_model': 9.995079458474164e-05, 'epoch': 0.01}) +[2026-10-03 18:57:07,348][starVLA.training.train_starvla][INFO] - Step 172, Loss: {'train/loss': 1.6030082702636719, 'train/grad_norm_pre_clip': 55.14155578613281, 'timing/data': 0.0003736600046977401, 'timing/model': 0.3936438579985406, 'learning_rate/qwen_vl_interface': 9.994939900354154e-06, 'learning_rate/action_model': 9.994939900354154e-05, 'epoch': 0.01}) +[2026-10-03 18:57:08,054][starVLA.training.train_starvla][INFO] - Step 173, Loss: {'train/loss': 0.7219539284706116, 'train/grad_norm_pre_clip': 20.63189697265625, 'timing/data': 0.0002759800117928535, 'timing/model': 0.39494448801269755, 'learning_rate/qwen_vl_interface': 9.994798391768582e-06, 'learning_rate/action_model': 9.994798391768582e-05, 'epoch': 0.01}) +[2026-10-03 18:57:08,760][starVLA.training.train_starvla][INFO] - Step 174, Loss: {'train/loss': 1.4485347270965576, 'train/grad_norm_pre_clip': 80.5140380859375, 'timing/data': 0.00038039000355638564, 'timing/model': 0.39484127701143734, 'learning_rate/qwen_vl_interface': 9.99465493277562e-06, 'learning_rate/action_model': 9.994654932775618e-05, 'epoch': 0.01}) +[2026-10-03 18:57:09,667][starVLA.training.train_starvla][INFO] - Step 175, Loss: {'train/loss': 0.880353569984436, 'train/grad_norm_pre_clip': 38.15639877319336, 'timing/data': 0.00022485002409666777, 'timing/model': 0.5953220550145488, 'learning_rate/qwen_vl_interface': 9.994509523434235e-06, 'learning_rate/action_model': 9.994509523434235e-05, 'epoch': 0.01}) +[2026-10-03 18:57:10,375][starVLA.training.train_starvla][INFO] - Step 176, Loss: {'train/loss': 0.7458232641220093, 'train/grad_norm_pre_clip': 32.4566764831543, 'timing/data': 0.0003502200124785304, 'timing/model': 0.39573571700020693, 'learning_rate/qwen_vl_interface': 9.994362163804202e-06, 'learning_rate/action_model': 9.9943621638042e-05, 'epoch': 0.01}) +[2026-10-03 18:57:11,082][starVLA.training.train_starvla][INFO] - Step 177, Loss: {'train/loss': 0.40774059295654297, 'train/grad_norm_pre_clip': 20.979949951171875, 'timing/data': 0.0002655099960975349, 'timing/model': 0.3946842679870315, 'learning_rate/qwen_vl_interface': 9.994212853946092e-06, 'learning_rate/action_model': 9.994212853946092e-05, 'epoch': 0.01}) +[2026-10-03 18:57:11,789][starVLA.training.train_starvla][INFO] - Step 178, Loss: {'train/loss': 0.6120840907096863, 'train/grad_norm_pre_clip': 35.86960983276367, 'timing/data': 0.0003552999987732619, 'timing/model': 0.39409363697632216, 'learning_rate/qwen_vl_interface': 9.994061593921283e-06, 'learning_rate/action_model': 9.994061593921282e-05, 'epoch': 0.01}) +[2026-10-03 18:57:12,495][starVLA.training.train_starvla][INFO] - Step 179, Loss: {'train/loss': 1.485304594039917, 'train/grad_norm_pre_clip': 136.46710205078125, 'timing/data': 0.0002476800000295043, 'timing/model': 0.3933849569875747, 'learning_rate/qwen_vl_interface': 9.993908383791951e-06, 'learning_rate/action_model': 9.993908383791951e-05, 'epoch': 0.01}) +[2026-10-03 18:57:13,199][starVLA.training.train_starvla][INFO] - Step 180, Loss: {'train/loss': 0.9494996666908264, 'train/grad_norm_pre_clip': 68.37062072753906, 'timing/data': 0.0002373100141994655, 'timing/model': 0.3934254380001221, 'learning_rate/qwen_vl_interface': 9.993753223621076e-06, 'learning_rate/action_model': 9.993753223621077e-05, 'epoch': 0.01}) +[2026-10-03 18:57:14,110][starVLA.training.train_starvla][INFO] - Step 181, Loss: {'train/loss': 0.9814857244491577, 'train/grad_norm_pre_clip': 71.02513885498047, 'timing/data': 0.0002747200196608901, 'timing/model': 0.3956359680159949, 'learning_rate/qwen_vl_interface': 9.99359611347244e-06, 'learning_rate/action_model': 9.993596113472439e-05, 'epoch': 0.01}) +[2026-10-03 18:57:14,821][starVLA.training.train_starvla][INFO] - Step 182, Loss: {'train/loss': 0.6546496152877808, 'train/grad_norm_pre_clip': 66.18085479736328, 'timing/data': 0.00023020000662654638, 'timing/model': 0.3977971870044712, 'learning_rate/qwen_vl_interface': 9.99343705341062e-06, 'learning_rate/action_model': 9.99343705341062e-05, 'epoch': 0.01}) +[2026-10-03 18:57:15,525][starVLA.training.train_starvla][INFO] - Step 183, Loss: {'train/loss': 0.4831809997558594, 'train/grad_norm_pre_clip': 44.0848274230957, 'timing/data': 0.00022067001555114985, 'timing/model': 0.3940916179853957, 'learning_rate/qwen_vl_interface': 9.993276043501004e-06, 'learning_rate/action_model': 9.993276043501004e-05, 'epoch': 0.01}) +[2026-10-03 18:57:16,228][starVLA.training.train_starvla][INFO] - Step 184, Loss: {'train/loss': 0.461311936378479, 'train/grad_norm_pre_clip': 55.73417282104492, 'timing/data': 0.00026898999931290746, 'timing/model': 0.3917998390097637, 'learning_rate/qwen_vl_interface': 9.993113083809776e-06, 'learning_rate/action_model': 9.993113083809775e-05, 'epoch': 0.01}) +[2026-10-03 18:57:16,932][starVLA.training.train_starvla][INFO] - Step 185, Loss: {'train/loss': 0.4322602450847626, 'train/grad_norm_pre_clip': 79.45449829101562, 'timing/data': 0.00030154999694786966, 'timing/model': 0.3953522969968617, 'learning_rate/qwen_vl_interface': 9.992948174403921e-06, 'learning_rate/action_model': 9.992948174403922e-05, 'epoch': 0.01}) +[2026-10-03 18:57:17,841][starVLA.training.train_starvla][INFO] - Step 186, Loss: {'train/loss': 1.3867870569229126, 'train/grad_norm_pre_clip': 210.86898803710938, 'timing/data': 0.00035632000071927905, 'timing/model': 0.3937385469907895, 'learning_rate/qwen_vl_interface': 9.99278131535123e-06, 'learning_rate/action_model': 9.992781315351229e-05, 'epoch': 0.01}) +[2026-10-03 18:57:18,543][starVLA.training.train_starvla][INFO] - Step 187, Loss: {'train/loss': 0.8290815353393555, 'train/grad_norm_pre_clip': 74.11489868164062, 'timing/data': 0.00037796999095007777, 'timing/model': 0.3923939279920887, 'learning_rate/qwen_vl_interface': 9.992612506720289e-06, 'learning_rate/action_model': 9.99261250672029e-05, 'epoch': 0.01}) +[2026-10-03 18:57:19,247][starVLA.training.train_starvla][INFO] - Step 188, Loss: {'train/loss': 0.6939188241958618, 'train/grad_norm_pre_clip': 69.61787414550781, 'timing/data': 0.00027971999952569604, 'timing/model': 0.3937724370043725, 'learning_rate/qwen_vl_interface': 9.992441748580492e-06, 'learning_rate/action_model': 9.99244174858049e-05, 'epoch': 0.01}) +[2026-10-03 18:57:19,953][starVLA.training.train_starvla][INFO] - Step 189, Loss: {'train/loss': 0.8000351190567017, 'train/grad_norm_pre_clip': 103.5423812866211, 'timing/data': 0.0002706100058276206, 'timing/model': 0.3940398970153183, 'learning_rate/qwen_vl_interface': 9.992269041002028e-06, 'learning_rate/action_model': 9.992269041002027e-05, 'epoch': 0.01}) +[2026-10-03 18:57:20,658][starVLA.training.train_starvla][INFO] - Step 190, Loss: {'train/loss': 1.2714307308197021, 'train/grad_norm_pre_clip': 205.35684204101562, 'timing/data': 0.0003620500210672617, 'timing/model': 0.39410507801221684, 'learning_rate/qwen_vl_interface': 9.992094384055894e-06, 'learning_rate/action_model': 9.992094384055895e-05, 'epoch': 0.01}) +[2026-10-03 18:57:21,564][starVLA.training.train_starvla][INFO] - Step 191, Loss: {'train/loss': 0.5993455648422241, 'train/grad_norm_pre_clip': 118.75083923339844, 'timing/data': 0.00025241999537684023, 'timing/model': 0.5949193060223479, 'learning_rate/qwen_vl_interface': 9.991917777813884e-06, 'learning_rate/action_model': 9.991917777813884e-05, 'epoch': 0.01}) +[2026-10-03 18:57:22,269][starVLA.training.train_starvla][INFO] - Step 192, Loss: {'train/loss': 1.3570011854171753, 'train/grad_norm_pre_clip': 154.79258728027344, 'timing/data': 0.0003333200002089143, 'timing/model': 0.39383373799500987, 'learning_rate/qwen_vl_interface': 9.991739222348593e-06, 'learning_rate/action_model': 9.991739222348592e-05, 'epoch': 0.01}) +[2026-10-03 18:57:22,974][starVLA.training.train_starvla][INFO] - Step 193, Loss: {'train/loss': 1.258622169494629, 'train/grad_norm_pre_clip': 120.95805358886719, 'timing/data': 0.0002316699828952551, 'timing/model': 0.39408483798615634, 'learning_rate/qwen_vl_interface': 9.991558717733418e-06, 'learning_rate/action_model': 9.991558717733417e-05, 'epoch': 0.01}) +[2026-10-03 18:57:23,679][starVLA.training.train_starvla][INFO] - Step 194, Loss: {'train/loss': 1.1718063354492188, 'train/grad_norm_pre_clip': 96.54759979248047, 'timing/data': 0.00024295001639984548, 'timing/model': 0.3940379680134356, 'learning_rate/qwen_vl_interface': 9.99137626404256e-06, 'learning_rate/action_model': 9.991376264042559e-05, 'epoch': 0.01}) +[2026-10-03 18:57:24,383][starVLA.training.train_starvla][INFO] - Step 195, Loss: {'train/loss': 0.4733092784881592, 'train/grad_norm_pre_clip': 210.55453491210938, 'timing/data': 0.0003370500053279102, 'timing/model': 0.39425054800813086, 'learning_rate/qwen_vl_interface': 9.991191861351015e-06, 'learning_rate/action_model': 9.991191861351014e-05, 'epoch': 0.01}) +[2026-10-03 18:57:25,085][starVLA.training.train_starvla][INFO] - Step 196, Loss: {'train/loss': 1.4936636686325073, 'train/grad_norm_pre_clip': 95.88373565673828, 'timing/data': 0.0003940790193155408, 'timing/model': 0.39248678798321635, 'learning_rate/qwen_vl_interface': 9.991005509734587e-06, 'learning_rate/action_model': 9.991005509734587e-05, 'epoch': 0.01}) +[2026-10-03 18:57:25,994][starVLA.training.train_starvla][INFO] - Step 197, Loss: {'train/loss': 0.8897584080696106, 'train/grad_norm_pre_clip': 56.66264724731445, 'timing/data': 0.0002818299981299788, 'timing/model': 0.3950616380025167, 'learning_rate/qwen_vl_interface': 9.990817209269878e-06, 'learning_rate/action_model': 9.990817209269878e-05, 'epoch': 0.01}) +[2026-10-03 18:57:26,698][starVLA.training.train_starvla][INFO] - Step 198, Loss: {'train/loss': 0.7265747785568237, 'train/grad_norm_pre_clip': 44.32122039794922, 'timing/data': 0.00036003999412059784, 'timing/model': 0.3930632669944316, 'learning_rate/qwen_vl_interface': 9.990626960034291e-06, 'learning_rate/action_model': 9.990626960034291e-05, 'epoch': 0.01}) +[2026-10-03 18:57:27,400][starVLA.training.train_starvla][INFO] - Step 199, Loss: {'train/loss': 0.7069977521896362, 'train/grad_norm_pre_clip': 134.3826446533203, 'timing/data': 0.00034594000317156315, 'timing/model': 0.39242755802115425, 'learning_rate/qwen_vl_interface': 9.990434762106029e-06, 'learning_rate/action_model': 9.990434762106028e-05, 'epoch': 0.01}) +[2026-10-03 18:57:28,104][starVLA.training.train_starvla][INFO] - Step 200, Loss: {'train/loss': 1.1200039386749268, 'train/grad_norm_pre_clip': 54.077369689941406, 'timing/data': 0.00030867999885231256, 'timing/model': 0.3947809870005585, 'learning_rate/qwen_vl_interface': 9.990240615564099e-06, 'learning_rate/action_model': 9.990240615564099e-05, 'epoch': 0.01}) +[2026-10-03 18:57:28,811][starVLA.training.train_starvla][INFO] - Step 201, Loss: {'train/loss': 0.4254142940044403, 'train/grad_norm_pre_clip': 16.458616256713867, 'timing/data': 0.00036452000495046377, 'timing/model': 0.3957437579811085, 'learning_rate/qwen_vl_interface': 9.990044520488305e-06, 'learning_rate/action_model': 9.990044520488305e-05, 'epoch': 0.01}) +[2026-10-03 18:57:29,717][starVLA.training.train_starvla][INFO] - Step 202, Loss: {'train/loss': 0.6412357687950134, 'train/grad_norm_pre_clip': 28.36476707458496, 'timing/data': 0.00037967998650856316, 'timing/model': 0.3941146179859061, 'learning_rate/qwen_vl_interface': 9.989846476959258e-06, 'learning_rate/action_model': 9.989846476959257e-05, 'epoch': 0.01}) +[2026-10-03 18:57:30,422][starVLA.training.train_starvla][INFO] - Step 203, Loss: {'train/loss': 1.162095546722412, 'train/grad_norm_pre_clip': 127.37583923339844, 'timing/data': 0.00036564000765793025, 'timing/model': 0.39292908800416626, 'learning_rate/qwen_vl_interface': 9.989646485058361e-06, 'learning_rate/action_model': 9.989646485058361e-05, 'epoch': 0.01}) +[2026-10-03 18:57:31,128][starVLA.training.train_starvla][INFO] - Step 204, Loss: {'train/loss': 1.116300106048584, 'train/grad_norm_pre_clip': 103.01158905029297, 'timing/data': 0.00036760998773388565, 'timing/model': 0.39452403801260516, 'learning_rate/qwen_vl_interface': 9.98944454486783e-06, 'learning_rate/action_model': 9.989444544867829e-05, 'epoch': 0.01}) +[2026-10-03 18:57:31,832][starVLA.training.train_starvla][INFO] - Step 205, Loss: {'train/loss': 1.537053108215332, 'train/grad_norm_pre_clip': 169.4678497314453, 'timing/data': 0.00026024001999758184, 'timing/model': 0.39373592799529433, 'learning_rate/qwen_vl_interface': 9.989240656470667e-06, 'learning_rate/action_model': 9.989240656470666e-05, 'epoch': 0.02}) +[2026-10-03 18:57:32,538][starVLA.training.train_starvla][INFO] - Step 206, Loss: {'train/loss': 0.7679258584976196, 'train/grad_norm_pre_clip': 102.37663269042969, 'timing/data': 0.00038717000279575586, 'timing/model': 0.3935795270081144, 'learning_rate/qwen_vl_interface': 9.98903481995069e-06, 'learning_rate/action_model': 9.989034819950689e-05, 'epoch': 0.02}) +[2026-10-03 18:57:33,445][starVLA.training.train_starvla][INFO] - Step 207, Loss: {'train/loss': 1.2930402755737305, 'train/grad_norm_pre_clip': 142.0616455078125, 'timing/data': 0.00037345002056099474, 'timing/model': 0.5955046159797348, 'learning_rate/qwen_vl_interface': 9.988827035392506e-06, 'learning_rate/action_model': 9.988827035392507e-05, 'epoch': 0.02}) +[2026-10-03 18:57:34,148][starVLA.training.train_starvla][INFO] - Step 208, Loss: {'train/loss': 1.2136521339416504, 'train/grad_norm_pre_clip': 204.58035278320312, 'timing/data': 0.00037928001256659627, 'timing/model': 0.3919103780062869, 'learning_rate/qwen_vl_interface': 9.988617302881529e-06, 'learning_rate/action_model': 9.988617302881529e-05, 'epoch': 0.02}) +[2026-10-03 18:57:34,866][starVLA.training.train_starvla][INFO] - Step 209, Loss: {'train/loss': 0.8472302556037903, 'train/grad_norm_pre_clip': 30.96863555908203, 'timing/data': 0.00027449001208879054, 'timing/model': 0.40790690600988455, 'learning_rate/qwen_vl_interface': 9.988405622503972e-06, 'learning_rate/action_model': 9.988405622503971e-05, 'epoch': 0.02}) +[2026-10-03 18:57:35,571][starVLA.training.train_starvla][INFO] - Step 210, Loss: {'train/loss': 0.5115779042243958, 'train/grad_norm_pre_clip': 25.48331642150879, 'timing/data': 0.0002523800067137927, 'timing/model': 0.39372198801720515, 'learning_rate/qwen_vl_interface': 9.988191994346848e-06, 'learning_rate/action_model': 9.988191994346848e-05, 'epoch': 0.02}) +[2026-10-03 18:57:36,274][starVLA.training.train_starvla][INFO] - Step 211, Loss: {'train/loss': 0.5981177091598511, 'train/grad_norm_pre_clip': 90.74729919433594, 'timing/data': 0.00026436999905854464, 'timing/model': 0.39194020797731355, 'learning_rate/qwen_vl_interface': 9.987976418497974e-06, 'learning_rate/action_model': 9.987976418497972e-05, 'epoch': 0.02}) +[2026-10-03 18:57:37,178][starVLA.training.train_starvla][INFO] - Step 212, Loss: {'train/loss': 0.6936525106430054, 'train/grad_norm_pre_clip': 106.28584289550781, 'timing/data': 0.0003594599838834256, 'timing/model': 0.5933058070077095, 'learning_rate/qwen_vl_interface': 9.987758895045964e-06, 'learning_rate/action_model': 9.987758895045963e-05, 'epoch': 0.02}) +[2026-10-03 18:57:37,899][starVLA.training.train_starvla][INFO] - Step 213, Loss: {'train/loss': 0.9334205985069275, 'train/grad_norm_pre_clip': 97.67769622802734, 'timing/data': 0.00033941000583581626, 'timing/model': 0.40804586600279436, 'learning_rate/qwen_vl_interface': 9.98753942408023e-06, 'learning_rate/action_model': 9.98753942408023e-05, 'epoch': 0.02}) +[2026-10-03 18:57:38,604][starVLA.training.train_starvla][INFO] - Step 214, Loss: {'train/loss': 1.0953011512756348, 'train/grad_norm_pre_clip': 108.3887939453125, 'timing/data': 0.0004095499753020704, 'timing/model': 0.39367596799274907, 'learning_rate/qwen_vl_interface': 9.987318005690992e-06, 'learning_rate/action_model': 9.987318005690992e-05, 'epoch': 0.02}) +[2026-10-03 18:57:39,306][starVLA.training.train_starvla][INFO] - Step 215, Loss: {'train/loss': 0.43967753648757935, 'train/grad_norm_pre_clip': 35.56039810180664, 'timing/data': 0.00021820998517796397, 'timing/model': 0.39228463801555336, 'learning_rate/qwen_vl_interface': 9.987094639969268e-06, 'learning_rate/action_model': 9.987094639969268e-05, 'epoch': 0.02}) +[2026-10-03 18:57:40,012][starVLA.training.train_starvla][INFO] - Step 216, Loss: {'train/loss': 0.765623927116394, 'train/grad_norm_pre_clip': 160.01780700683594, 'timing/data': 0.0003337599919177592, 'timing/model': 0.39380326701211743, 'learning_rate/qwen_vl_interface': 9.986869327006872e-06, 'learning_rate/action_model': 9.986869327006873e-05, 'epoch': 0.02}) +[2026-10-03 18:57:40,935][starVLA.training.train_starvla][INFO] - Step 217, Loss: {'train/loss': 0.8355971574783325, 'train/grad_norm_pre_clip': 69.19100952148438, 'timing/data': 0.000282669992884621, 'timing/model': 0.6112272439931985, 'learning_rate/qwen_vl_interface': 9.986642066896424e-06, 'learning_rate/action_model': 9.986642066896423e-05, 'epoch': 0.02}) +[2026-10-03 18:57:41,640][starVLA.training.train_starvla][INFO] - Step 218, Loss: {'train/loss': 1.1000819206237793, 'train/grad_norm_pre_clip': 42.9588737487793, 'timing/data': 0.0002562199952080846, 'timing/model': 0.39473373800865375, 'learning_rate/qwen_vl_interface': 9.986412859731341e-06, 'learning_rate/action_model': 9.98641285973134e-05, 'epoch': 0.02}) +[2026-10-03 18:57:42,344][starVLA.training.train_starvla][INFO] - Step 219, Loss: {'train/loss': 0.5861061215400696, 'train/grad_norm_pre_clip': 67.00469970703125, 'timing/data': 0.0003086100041400641, 'timing/model': 0.3924388279847335, 'learning_rate/qwen_vl_interface': 9.986181705605839e-06, 'learning_rate/action_model': 9.986181705605839e-05, 'epoch': 0.02}) +[2026-10-03 18:57:43,048][starVLA.training.train_starvla][INFO] - Step 220, Loss: {'train/loss': 0.2936898171901703, 'train/grad_norm_pre_clip': 31.539470672607422, 'timing/data': 0.00040415002149529755, 'timing/model': 0.3932283379836008, 'learning_rate/qwen_vl_interface': 9.98594860461494e-06, 'learning_rate/action_model': 9.98594860461494e-05, 'epoch': 0.02}) +[2026-10-03 18:57:43,766][starVLA.training.train_starvla][INFO] - Step 221, Loss: {'train/loss': 0.6056309938430786, 'train/grad_norm_pre_clip': 67.52163696289062, 'timing/data': 0.0002411099849268794, 'timing/model': 0.4063455570139922, 'learning_rate/qwen_vl_interface': 9.985713556854461e-06, 'learning_rate/action_model': 9.98571355685446e-05, 'epoch': 0.02}) +[2026-10-03 18:57:44,471][starVLA.training.train_starvla][INFO] - Step 222, Loss: {'train/loss': 1.055844783782959, 'train/grad_norm_pre_clip': 93.41343688964844, 'timing/data': 0.00027323001995682716, 'timing/model': 0.3938606780138798, 'learning_rate/qwen_vl_interface': 9.985476562421024e-06, 'learning_rate/action_model': 9.985476562421024e-05, 'epoch': 0.02}) +[2026-10-03 18:57:45,355][starVLA.training.train_starvla][INFO] - Step 223, Loss: {'train/loss': 0.8621150255203247, 'train/grad_norm_pre_clip': 113.86711883544922, 'timing/data': 0.00028604999533854425, 'timing/model': 0.36715543098398484, 'learning_rate/qwen_vl_interface': 9.985237621412047e-06, 'learning_rate/action_model': 9.985237621412046e-05, 'epoch': 0.02}) +[2026-10-03 18:57:46,059][starVLA.training.train_starvla][INFO] - Step 224, Loss: {'train/loss': 0.648861289024353, 'train/grad_norm_pre_clip': 124.59932708740234, 'timing/data': 0.0003616700123529881, 'timing/model': 0.3926280779996887, 'learning_rate/qwen_vl_interface': 9.984996733925749e-06, 'learning_rate/action_model': 9.984996733925748e-05, 'epoch': 0.02}) +[2026-10-03 18:57:46,775][starVLA.training.train_starvla][INFO] - Step 225, Loss: {'train/loss': 1.1004151105880737, 'train/grad_norm_pre_clip': 58.88912582397461, 'timing/data': 0.00026159000117331743, 'timing/model': 0.4046219270094298, 'learning_rate/qwen_vl_interface': 9.98475390006115e-06, 'learning_rate/action_model': 9.98475390006115e-05, 'epoch': 0.02}) +[2026-10-03 18:57:47,480][starVLA.training.train_starvla][INFO] - Step 226, Loss: {'train/loss': 0.3744993507862091, 'train/grad_norm_pre_clip': 36.98918914794922, 'timing/data': 0.00035855997703038156, 'timing/model': 0.39381823799340054, 'learning_rate/qwen_vl_interface': 9.98450911991807e-06, 'learning_rate/action_model': 9.98450911991807e-05, 'epoch': 0.02}) +[2026-10-03 18:57:48,185][starVLA.training.train_starvla][INFO] - Step 227, Loss: {'train/loss': 0.35943737626075745, 'train/grad_norm_pre_clip': 52.5987434387207, 'timing/data': 0.0003501900064293295, 'timing/model': 0.39497109802323394, 'learning_rate/qwen_vl_interface': 9.98426239359713e-06, 'learning_rate/action_model': 9.984262393597129e-05, 'epoch': 0.02}) +[2026-10-03 18:57:49,088][starVLA.training.train_starvla][INFO] - Step 228, Loss: {'train/loss': 0.7099690437316895, 'train/grad_norm_pre_clip': 81.17166137695312, 'timing/data': 0.00035188000765629113, 'timing/model': 0.5911386270017829, 'learning_rate/qwen_vl_interface': 9.984013721199747e-06, 'learning_rate/action_model': 9.984013721199747e-05, 'epoch': 0.02}) +[2026-10-03 18:57:49,804][starVLA.training.train_starvla][INFO] - Step 229, Loss: {'train/loss': 0.6885252594947815, 'train/grad_norm_pre_clip': 59.15983200073242, 'timing/data': 0.0003675399930216372, 'timing/model': 0.40486502699786797, 'learning_rate/qwen_vl_interface': 9.983763102828143e-06, 'learning_rate/action_model': 9.983763102828143e-05, 'epoch': 0.02}) +[2026-10-03 18:57:50,507][starVLA.training.train_starvla][INFO] - Step 230, Loss: {'train/loss': 0.7398610711097717, 'train/grad_norm_pre_clip': 49.07310104370117, 'timing/data': 0.0003040000156033784, 'timing/model': 0.39288828798453324, 'learning_rate/qwen_vl_interface': 9.983510538585336e-06, 'learning_rate/action_model': 9.983510538585337e-05, 'epoch': 0.02}) +[2026-10-03 18:57:51,209][starVLA.training.train_starvla][INFO] - Step 231, Loss: {'train/loss': 0.6863754987716675, 'train/grad_norm_pre_clip': 22.357820510864258, 'timing/data': 0.0003374800144229084, 'timing/model': 0.39151914801914245, 'learning_rate/qwen_vl_interface': 9.98325602857515e-06, 'learning_rate/action_model': 9.98325602857515e-05, 'epoch': 0.02}) +[2026-10-03 18:57:51,913][starVLA.training.train_starvla][INFO] - Step 232, Loss: {'train/loss': 0.9697451591491699, 'train/grad_norm_pre_clip': 199.0648651123047, 'timing/data': 0.00036030999035574496, 'timing/model': 0.392345217987895, 'learning_rate/qwen_vl_interface': 9.9829995729022e-06, 'learning_rate/action_model': 9.982999572902199e-05, 'epoch': 0.02}) +[2026-10-03 18:57:52,627][starVLA.training.train_starvla][INFO] - Step 233, Loss: {'train/loss': 0.4018901586532593, 'train/grad_norm_pre_clip': 59.98541259765625, 'timing/data': 0.0002105100138578564, 'timing/model': 0.4049392660090234, 'learning_rate/qwen_vl_interface': 9.982741171671905e-06, 'learning_rate/action_model': 9.982741171671906e-05, 'epoch': 0.02}) +[2026-10-03 18:57:53,524][starVLA.training.train_starvla][INFO] - Step 234, Loss: {'train/loss': 0.33166298270225525, 'train/grad_norm_pre_clip': 21.839134216308594, 'timing/data': 0.0003736600046977401, 'timing/model': 0.3930360880040098, 'learning_rate/qwen_vl_interface': 9.982480824990489e-06, 'learning_rate/action_model': 9.982480824990488e-05, 'epoch': 0.02}) +[2026-10-03 18:57:54,228][starVLA.training.train_starvla][INFO] - Step 235, Loss: {'train/loss': 0.7479602098464966, 'train/grad_norm_pre_clip': 40.08186721801758, 'timing/data': 0.00030744000105187297, 'timing/model': 0.39472777798073366, 'learning_rate/qwen_vl_interface': 9.982218532964963e-06, 'learning_rate/action_model': 9.982218532964964e-05, 'epoch': 0.02}) +[2026-10-03 18:57:54,934][starVLA.training.train_starvla][INFO] - Step 236, Loss: {'train/loss': 0.9667324423789978, 'train/grad_norm_pre_clip': 82.37887573242188, 'timing/data': 0.0003620599745772779, 'timing/model': 0.39453819798654877, 'learning_rate/qwen_vl_interface': 9.981954295703152e-06, 'learning_rate/action_model': 9.981954295703152e-05, 'epoch': 0.02}) +[2026-10-03 18:57:55,651][starVLA.training.train_starvla][INFO] - Step 237, Loss: {'train/loss': 0.5947191715240479, 'train/grad_norm_pre_clip': 28.819316864013672, 'timing/data': 0.000354110001353547, 'timing/model': 0.4047778369858861, 'learning_rate/qwen_vl_interface': 9.981688113313672e-06, 'learning_rate/action_model': 9.981688113313672e-05, 'epoch': 0.02}) +[2026-10-03 18:57:56,354][starVLA.training.train_starvla][INFO] - Step 238, Loss: {'train/loss': 0.5536900758743286, 'train/grad_norm_pre_clip': 36.63325119018555, 'timing/data': 0.0003755300131160766, 'timing/model': 0.39274135799496435, 'learning_rate/qwen_vl_interface': 9.98141998590594e-06, 'learning_rate/action_model': 9.981419985905939e-05, 'epoch': 0.02}) +[2026-10-03 18:57:57,259][starVLA.training.train_starvla][INFO] - Step 239, Loss: {'train/loss': 0.5639780759811401, 'train/grad_norm_pre_clip': 40.141788482666016, 'timing/data': 0.0003529100213199854, 'timing/model': 0.39461954799480736, 'learning_rate/qwen_vl_interface': 9.981149913590171e-06, 'learning_rate/action_model': 9.981149913590172e-05, 'epoch': 0.02}) +[2026-10-03 18:57:57,965][starVLA.training.train_starvla][INFO] - Step 240, Loss: {'train/loss': 0.3739948272705078, 'train/grad_norm_pre_clip': 45.83528518676758, 'timing/data': 0.00023549998877570033, 'timing/model': 0.3945538080006372, 'learning_rate/qwen_vl_interface': 9.980877896477386e-06, 'learning_rate/action_model': 9.980877896477386e-05, 'epoch': 0.02}) +[2026-10-03 18:57:58,684][starVLA.training.train_starvla][INFO] - Step 241, Loss: {'train/loss': 0.6363605260848999, 'train/grad_norm_pre_clip': 63.43999481201172, 'timing/data': 0.0002491599880158901, 'timing/model': 0.4076553159975447, 'learning_rate/qwen_vl_interface': 9.980603934679397e-06, 'learning_rate/action_model': 9.980603934679398e-05, 'epoch': 0.02}) +[2026-10-03 18:57:59,387][starVLA.training.train_starvla][INFO] - Step 242, Loss: {'train/loss': 0.5236027836799622, 'train/grad_norm_pre_clip': 23.77404022216797, 'timing/data': 0.0003843200101982802, 'timing/model': 0.3925241580000147, 'learning_rate/qwen_vl_interface': 9.980328028308824e-06, 'learning_rate/action_model': 9.980328028308824e-05, 'epoch': 0.02}) +[2026-10-03 18:58:00,092][starVLA.training.train_starvla][INFO] - Step 243, Loss: {'train/loss': 0.679953932762146, 'train/grad_norm_pre_clip': 24.815881729125977, 'timing/data': 0.0004304200119804591, 'timing/model': 0.3932579579995945, 'learning_rate/qwen_vl_interface': 9.980050177479075e-06, 'learning_rate/action_model': 9.980050177479076e-05, 'epoch': 0.02}) +[2026-10-03 18:58:00,997][starVLA.training.train_starvla][INFO] - Step 244, Loss: {'train/loss': 1.2418972253799438, 'train/grad_norm_pre_clip': 93.63764190673828, 'timing/data': 0.00035977899096906185, 'timing/model': 0.5937178070016671, 'learning_rate/qwen_vl_interface': 9.97977038230437e-06, 'learning_rate/action_model': 9.97977038230437e-05, 'epoch': 0.02}) +[2026-10-03 18:58:01,712][starVLA.training.train_starvla][INFO] - Step 245, Loss: {'train/loss': 0.253836065530777, 'train/grad_norm_pre_clip': 45.949806213378906, 'timing/data': 0.0002848999865818769, 'timing/model': 0.40527244700933807, 'learning_rate/qwen_vl_interface': 9.97948864289972e-06, 'learning_rate/action_model': 9.97948864289972e-05, 'epoch': 0.02}) +[2026-10-03 18:58:02,417][starVLA.training.train_starvla][INFO] - Step 246, Loss: {'train/loss': 0.7168911695480347, 'train/grad_norm_pre_clip': 78.6818618774414, 'timing/data': 0.0002212200197391212, 'timing/model': 0.39382793800905347, 'learning_rate/qwen_vl_interface': 9.979204959380936e-06, 'learning_rate/action_model': 9.979204959380936e-05, 'epoch': 0.02}) +[2026-10-03 18:58:03,124][starVLA.training.train_starvla][INFO] - Step 247, Loss: {'train/loss': 0.7239928841590881, 'train/grad_norm_pre_clip': 74.68294525146484, 'timing/data': 0.0003717700019478798, 'timing/model': 0.3950245080050081, 'learning_rate/qwen_vl_interface': 9.978919331864631e-06, 'learning_rate/action_model': 9.97891933186463e-05, 'epoch': 0.02}) +[2026-10-03 18:58:03,828][starVLA.training.train_starvla][INFO] - Step 248, Loss: {'train/loss': 1.307377815246582, 'train/grad_norm_pre_clip': 444.7956237792969, 'timing/data': 0.00031755000236444175, 'timing/model': 0.3944916670152452, 'learning_rate/qwen_vl_interface': 9.978631760468216e-06, 'learning_rate/action_model': 9.978631760468216e-05, 'epoch': 0.02}) +[2026-10-03 18:58:04,751][starVLA.training.train_starvla][INFO] - Step 249, Loss: {'train/loss': 1.705049753189087, 'train/grad_norm_pre_clip': 495.2105712890625, 'timing/data': 0.00039288998232223094, 'timing/model': 0.6131319249980152, 'learning_rate/qwen_vl_interface': 9.9783422453099e-06, 'learning_rate/action_model': 9.9783422453099e-05, 'epoch': 0.02}) +[2026-10-03 18:58:05,455][starVLA.training.train_starvla][INFO] - Step 250, Loss: {'train/loss': 1.7479526996612549, 'train/grad_norm_pre_clip': 393.7113952636719, 'timing/data': 0.00035401000059209764, 'timing/model': 0.3936793679895345, 'learning_rate/qwen_vl_interface': 9.978050786508693e-06, 'learning_rate/action_model': 9.978050786508692e-05, 'epoch': 0.02}) +[2026-10-03 18:58:06,161][starVLA.training.train_starvla][INFO] - Step 251, Loss: {'train/loss': 0.7972655296325684, 'train/grad_norm_pre_clip': 118.78123474121094, 'timing/data': 0.00037075000000186265, 'timing/model': 0.3945899379905313, 'learning_rate/qwen_vl_interface': 9.9777573841844e-06, 'learning_rate/action_model': 9.977757384184401e-05, 'epoch': 0.02}) +[2026-10-03 18:58:06,868][starVLA.training.train_starvla][INFO] - Step 252, Loss: {'train/loss': 1.4304035902023315, 'train/grad_norm_pre_clip': 55.12629699707031, 'timing/data': 0.00038507001590915024, 'timing/model': 0.3947884679946583, 'learning_rate/qwen_vl_interface': 9.97746203845763e-06, 'learning_rate/action_model': 9.97746203845763e-05, 'epoch': 0.02}) +[2026-10-03 18:58:07,588][starVLA.training.train_starvla][INFO] - Step 253, Loss: {'train/loss': 0.7663506269454956, 'train/grad_norm_pre_clip': 102.05309295654297, 'timing/data': 0.000279229978332296, 'timing/model': 0.40873918699799106, 'learning_rate/qwen_vl_interface': 9.97716474944979e-06, 'learning_rate/action_model': 9.977164749449788e-05, 'epoch': 0.02}) +[2026-10-03 18:58:08,266][starVLA.training.train_starvla][INFO] - Step 254, Loss: {'train/loss': 0.4904159903526306, 'train/grad_norm_pre_clip': 22.422441482543945, 'timing/data': 0.00022623001132160425, 'timing/model': 0.36807207099627703, 'learning_rate/qwen_vl_interface': 9.976865517283079e-06, 'learning_rate/action_model': 9.976865517283078e-05, 'epoch': 0.02}) +[2026-10-03 18:58:09,172][starVLA.training.train_starvla][INFO] - Step 255, Loss: {'train/loss': 0.9946461319923401, 'train/grad_norm_pre_clip': 85.11469268798828, 'timing/data': 0.00035236001713201404, 'timing/model': 0.5950227860012092, 'learning_rate/qwen_vl_interface': 9.976564342080505e-06, 'learning_rate/action_model': 9.976564342080504e-05, 'epoch': 0.02}) +[2026-10-03 18:58:09,876][starVLA.training.train_starvla][INFO] - Step 256, Loss: {'train/loss': 0.6391854882240295, 'train/grad_norm_pre_clip': 63.19303894042969, 'timing/data': 0.0004043400113005191, 'timing/model': 0.39330844799405895, 'learning_rate/qwen_vl_interface': 9.976261223965867e-06, 'learning_rate/action_model': 9.976261223965865e-05, 'epoch': 0.02}) +[2026-10-03 18:58:10,558][starVLA.training.train_starvla][INFO] - Step 257, Loss: {'train/loss': 0.908323347568512, 'train/grad_norm_pre_clip': 147.14398193359375, 'timing/data': 0.00026919000083580613, 'timing/model': 0.36969654099084437, 'learning_rate/qwen_vl_interface': 9.975956163063767e-06, 'learning_rate/action_model': 9.975956163063767e-05, 'epoch': 0.02}) +[2026-10-03 18:58:11,263][starVLA.training.train_starvla][INFO] - Step 258, Loss: {'train/loss': 0.8226618766784668, 'train/grad_norm_pre_clip': 125.59449768066406, 'timing/data': 0.0003733100020326674, 'timing/model': 0.39298989798408, 'learning_rate/qwen_vl_interface': 9.975649159499605e-06, 'learning_rate/action_model': 9.975649159499603e-05, 'epoch': 0.02}) +[2026-10-03 18:58:11,969][starVLA.training.train_starvla][INFO] - Step 259, Loss: {'train/loss': 0.8062177300453186, 'train/grad_norm_pre_clip': 31.223344802856445, 'timing/data': 0.000321789993904531, 'timing/model': 0.3934199179930147, 'learning_rate/qwen_vl_interface': 9.975340213399573e-06, 'learning_rate/action_model': 9.975340213399573e-05, 'epoch': 0.02}) +[2026-10-03 18:58:12,880][starVLA.training.train_starvla][INFO] - Step 260, Loss: {'train/loss': 0.6527661681175232, 'train/grad_norm_pre_clip': 40.8648796081543, 'timing/data': 0.0003531290276441723, 'timing/model': 0.6002108459942974, 'learning_rate/qwen_vl_interface': 9.975029324890674e-06, 'learning_rate/action_model': 9.975029324890674e-05, 'epoch': 0.02}) +[2026-10-03 18:58:13,586][starVLA.training.train_starvla][INFO] - Step 261, Loss: {'train/loss': 0.33885887265205383, 'train/grad_norm_pre_clip': 33.61109924316406, 'timing/data': 0.00033353999606333673, 'timing/model': 0.3961061979935039, 'learning_rate/qwen_vl_interface': 9.974716494100698e-06, 'learning_rate/action_model': 9.974716494100697e-05, 'epoch': 0.02}) +[2026-10-03 18:58:14,292][starVLA.training.train_starvla][INFO] - Step 262, Loss: {'train/loss': 0.7271348834037781, 'train/grad_norm_pre_clip': 74.08433532714844, 'timing/data': 0.0003759399987757206, 'timing/model': 0.3940682680113241, 'learning_rate/qwen_vl_interface': 9.97440172115824e-06, 'learning_rate/action_model': 9.974401721158239e-05, 'epoch': 0.02}) +[2026-10-03 18:58:14,997][starVLA.training.train_starvla][INFO] - Step 263, Loss: {'train/loss': 1.454532265663147, 'train/grad_norm_pre_clip': 114.82096099853516, 'timing/data': 0.00033171000541187823, 'timing/model': 0.3937563579820562, 'learning_rate/qwen_vl_interface': 9.97408500619269e-06, 'learning_rate/action_model': 9.97408500619269e-05, 'epoch': 0.02}) +[2026-10-03 18:58:15,703][starVLA.training.train_starvla][INFO] - Step 264, Loss: {'train/loss': 0.7566049695014954, 'train/grad_norm_pre_clip': 30.14661407470703, 'timing/data': 0.00037838000571355224, 'timing/model': 0.39420485799200833, 'learning_rate/qwen_vl_interface': 9.973766349334239e-06, 'learning_rate/action_model': 9.973766349334239e-05, 'epoch': 0.02}) +[2026-10-03 18:58:16,409][starVLA.training.train_starvla][INFO] - Step 265, Loss: {'train/loss': 0.8770293593406677, 'train/grad_norm_pre_clip': 32.31618881225586, 'timing/data': 0.00027598999440670013, 'timing/model': 0.3943288679874968, 'learning_rate/qwen_vl_interface': 9.973445750713875e-06, 'learning_rate/action_model': 9.973445750713875e-05, 'epoch': 0.02}) +[2026-10-03 18:58:17,317][starVLA.training.train_starvla][INFO] - Step 266, Loss: {'train/loss': 0.8566201329231262, 'train/grad_norm_pre_clip': 47.29364013671875, 'timing/data': 0.000232449994655326, 'timing/model': 0.39416026798426174, 'learning_rate/qwen_vl_interface': 9.97312321046338e-06, 'learning_rate/action_model': 9.97312321046338e-05, 'epoch': 0.02}) +[2026-10-03 18:58:18,021][starVLA.training.train_starvla][INFO] - Step 267, Loss: {'train/loss': 0.9895282983779907, 'train/grad_norm_pre_clip': 122.0699462890625, 'timing/data': 0.0003753800119739026, 'timing/model': 0.39440225801081397, 'learning_rate/qwen_vl_interface': 9.972798728715343e-06, 'learning_rate/action_model': 9.972798728715344e-05, 'epoch': 0.02}) +[2026-10-03 18:58:18,727][starVLA.training.train_starvla][INFO] - Step 268, Loss: {'train/loss': 1.0310120582580566, 'train/grad_norm_pre_clip': 66.67538452148438, 'timing/data': 0.00035883000236935914, 'timing/model': 0.3945491580234375, 'learning_rate/qwen_vl_interface': 9.972472305603145e-06, 'learning_rate/action_model': 9.972472305603144e-05, 'epoch': 0.02}) +[2026-10-03 18:58:19,433][starVLA.training.train_starvla][INFO] - Step 269, Loss: {'train/loss': 0.6353262662887573, 'train/grad_norm_pre_clip': 47.52810287475586, 'timing/data': 0.00024074999964796007, 'timing/model': 0.3955743169935886, 'learning_rate/qwen_vl_interface': 9.972143941260965e-06, 'learning_rate/action_model': 9.972143941260965e-05, 'epoch': 0.02}) +[2026-10-03 18:58:20,137][starVLA.training.train_starvla][INFO] - Step 270, Loss: {'train/loss': 0.9363000392913818, 'train/grad_norm_pre_clip': 58.995277404785156, 'timing/data': 0.0003688099968712777, 'timing/model': 0.39194869800121523, 'learning_rate/qwen_vl_interface': 9.971813635823781e-06, 'learning_rate/action_model': 9.971813635823781e-05, 'epoch': 0.02}) +[2026-10-03 18:58:21,044][starVLA.training.train_starvla][INFO] - Step 271, Loss: {'train/loss': 0.7966995239257812, 'train/grad_norm_pre_clip': 22.411296844482422, 'timing/data': 0.00029431001166813076, 'timing/model': 0.3941562380059622, 'learning_rate/qwen_vl_interface': 9.971481389427372e-06, 'learning_rate/action_model': 9.971481389427371e-05, 'epoch': 0.02}) +[2026-10-03 18:58:21,749][starVLA.training.train_starvla][INFO] - Step 272, Loss: {'train/loss': 0.4758540987968445, 'train/grad_norm_pre_clip': 25.270341873168945, 'timing/data': 0.00022996999905444682, 'timing/model': 0.3947414080030285, 'learning_rate/qwen_vl_interface': 9.971147202208308e-06, 'learning_rate/action_model': 9.971147202208308e-05, 'epoch': 0.02}) +[2026-10-03 18:58:22,454][starVLA.training.train_starvla][INFO] - Step 273, Loss: {'train/loss': 0.7667236328125, 'train/grad_norm_pre_clip': 47.15958786010742, 'timing/data': 0.00029754001297988, 'timing/model': 0.3937762879941147, 'learning_rate/qwen_vl_interface': 9.970811074303962e-06, 'learning_rate/action_model': 9.970811074303962e-05, 'epoch': 0.02}) +[2026-10-03 18:58:23,161][starVLA.training.train_starvla][INFO] - Step 274, Loss: {'train/loss': 0.6525188684463501, 'train/grad_norm_pre_clip': 52.44389343261719, 'timing/data': 0.00035677000414580107, 'timing/model': 0.3945393280009739, 'learning_rate/qwen_vl_interface': 9.970473005852505e-06, 'learning_rate/action_model': 9.970473005852505e-05, 'epoch': 0.02}) +[2026-10-03 18:58:23,865][starVLA.training.train_starvla][INFO] - Step 275, Loss: {'train/loss': 0.9112850427627563, 'train/grad_norm_pre_clip': 97.69207763671875, 'timing/data': 0.0003227400011382997, 'timing/model': 0.3927892280044034, 'learning_rate/qwen_vl_interface': 9.970132996992904e-06, 'learning_rate/action_model': 9.970132996992904e-05, 'epoch': 0.02}) +[2026-10-03 18:58:24,773][starVLA.training.train_starvla][INFO] - Step 276, Loss: {'train/loss': 0.5219314098358154, 'train/grad_norm_pre_clip': 88.34539794921875, 'timing/data': 0.0003690900048241019, 'timing/model': 0.5967395760235377, 'learning_rate/qwen_vl_interface': 9.969791047864921e-06, 'learning_rate/action_model': 9.969791047864921e-05, 'epoch': 0.02}) +[2026-10-03 18:58:25,477][starVLA.training.train_starvla][INFO] - Step 277, Loss: {'train/loss': 0.7479195594787598, 'train/grad_norm_pre_clip': 51.141151428222656, 'timing/data': 0.0002522599825169891, 'timing/model': 0.39535004799836315, 'learning_rate/qwen_vl_interface': 9.969447158609122e-06, 'learning_rate/action_model': 9.96944715860912e-05, 'epoch': 0.02}) +[2026-10-03 18:58:26,184][starVLA.training.train_starvla][INFO] - Step 278, Loss: {'train/loss': 0.9833189845085144, 'train/grad_norm_pre_clip': 74.73501586914062, 'timing/data': 0.0003306299913674593, 'timing/model': 0.3941638680116739, 'learning_rate/qwen_vl_interface': 9.969101329366864e-06, 'learning_rate/action_model': 9.969101329366863e-05, 'epoch': 0.02}) +[2026-10-03 18:58:26,887][starVLA.training.train_starvla][INFO] - Step 279, Loss: {'train/loss': 1.1501939296722412, 'train/grad_norm_pre_clip': 107.24794006347656, 'timing/data': 0.00037217000499367714, 'timing/model': 0.39375029801158234, 'learning_rate/qwen_vl_interface': 9.968753560280305e-06, 'learning_rate/action_model': 9.968753560280304e-05, 'epoch': 0.02}) +[2026-10-03 18:58:27,593][starVLA.training.train_starvla][INFO] - Step 280, Loss: {'train/loss': 0.8217226266860962, 'train/grad_norm_pre_clip': 44.2718505859375, 'timing/data': 0.00037709999014623463, 'timing/model': 0.3946587280079257, 'learning_rate/qwen_vl_interface': 9.968403851492402e-06, 'learning_rate/action_model': 9.968403851492401e-05, 'epoch': 0.02}) +[2026-10-03 18:58:28,501][starVLA.training.train_starvla][INFO] - Step 281, Loss: {'train/loss': 0.5887292623519897, 'train/grad_norm_pre_clip': 36.39091491699219, 'timing/data': 0.0002915700024459511, 'timing/model': 0.5958159069996327, 'learning_rate/qwen_vl_interface': 9.968052203146904e-06, 'learning_rate/action_model': 9.968052203146903e-05, 'epoch': 0.02}) +[2026-10-03 18:58:29,207][starVLA.training.train_starvla][INFO] - Step 282, Loss: {'train/loss': 0.2923385500907898, 'train/grad_norm_pre_clip': 21.69533920288086, 'timing/data': 0.00031648000003769994, 'timing/model': 0.3945754980086349, 'learning_rate/qwen_vl_interface': 9.967698615388363e-06, 'learning_rate/action_model': 9.967698615388362e-05, 'epoch': 0.02}) +[2026-10-03 18:58:29,913][starVLA.training.train_starvla][INFO] - Step 283, Loss: {'train/loss': 0.2388838529586792, 'train/grad_norm_pre_clip': 26.287212371826172, 'timing/data': 0.00023757998133078218, 'timing/model': 0.3950504869862925, 'learning_rate/qwen_vl_interface': 9.967343088362123e-06, 'learning_rate/action_model': 9.967343088362122e-05, 'epoch': 0.02}) +[2026-10-03 18:58:30,620][starVLA.training.train_starvla][INFO] - Step 284, Loss: {'train/loss': 1.0931673049926758, 'train/grad_norm_pre_clip': 50.89358139038086, 'timing/data': 0.00038593000499531627, 'timing/model': 0.3953935379977338, 'learning_rate/qwen_vl_interface': 9.966985622214328e-06, 'learning_rate/action_model': 9.966985622214329e-05, 'epoch': 0.02}) +[2026-10-03 18:58:31,328][starVLA.training.train_starvla][INFO] - Step 285, Loss: {'train/loss': 0.20159421861171722, 'train/grad_norm_pre_clip': 8.268369674682617, 'timing/data': 0.0003225100226700306, 'timing/model': 0.3968273279897403, 'learning_rate/qwen_vl_interface': 9.966626217091922e-06, 'learning_rate/action_model': 9.966626217091922e-05, 'epoch': 0.02}) +[2026-10-03 18:58:32,233][starVLA.training.train_starvla][INFO] - Step 286, Loss: {'train/loss': 0.8919315338134766, 'train/grad_norm_pre_clip': 54.666343688964844, 'timing/data': 0.00027627000235952437, 'timing/model': 0.5937677370093297, 'learning_rate/qwen_vl_interface': 9.966264873142641e-06, 'learning_rate/action_model': 9.96626487314264e-05, 'epoch': 0.02}) +[2026-10-03 18:58:32,940][starVLA.training.train_starvla][INFO] - Step 287, Loss: {'train/loss': 0.7679027915000916, 'train/grad_norm_pre_clip': 99.61734008789062, 'timing/data': 0.0004068400012329221, 'timing/model': 0.39487452802131884, 'learning_rate/qwen_vl_interface': 9.965901590515018e-06, 'learning_rate/action_model': 9.965901590515018e-05, 'epoch': 0.02}) +[2026-10-03 18:58:33,646][starVLA.training.train_starvla][INFO] - Step 288, Loss: {'train/loss': 1.29659104347229, 'train/grad_norm_pre_clip': 80.32438659667969, 'timing/data': 0.0003525800129864365, 'timing/model': 0.39525131799746305, 'learning_rate/qwen_vl_interface': 9.965536369358386e-06, 'learning_rate/action_model': 9.965536369358386e-05, 'epoch': 0.02}) +[2026-10-03 18:58:34,353][starVLA.training.train_starvla][INFO] - Step 289, Loss: {'train/loss': 0.6931328773498535, 'train/grad_norm_pre_clip': 126.83183288574219, 'timing/data': 0.0002641600149217993, 'timing/model': 0.394748888007598, 'learning_rate/qwen_vl_interface': 9.965169209822874e-06, 'learning_rate/action_model': 9.965169209822874e-05, 'epoch': 0.02}) +[2026-10-03 18:58:35,062][starVLA.training.train_starvla][INFO] - Step 290, Loss: {'train/loss': 0.8569834232330322, 'train/grad_norm_pre_clip': 246.54571533203125, 'timing/data': 0.00034056001459248364, 'timing/model': 0.3957026580173988, 'learning_rate/qwen_vl_interface': 9.96480011205941e-06, 'learning_rate/action_model': 9.964800112059409e-05, 'epoch': 0.02}) +[2026-10-03 18:58:35,770][starVLA.training.train_starvla][INFO] - Step 291, Loss: {'train/loss': 0.7076636552810669, 'train/grad_norm_pre_clip': 101.51571655273438, 'timing/data': 0.00046613000449724495, 'timing/model': 0.3953162770194467, 'learning_rate/qwen_vl_interface': 9.964429076219711e-06, 'learning_rate/action_model': 9.964429076219711e-05, 'epoch': 0.02}) +[2026-10-03 18:58:36,678][starVLA.training.train_starvla][INFO] - Step 292, Loss: {'train/loss': 0.6386409997940063, 'train/grad_norm_pre_clip': 112.3857421875, 'timing/data': 0.0003815899835899472, 'timing/model': 0.39411374798510224, 'learning_rate/qwen_vl_interface': 9.964056102456299e-06, 'learning_rate/action_model': 9.964056102456299e-05, 'epoch': 0.02}) +[2026-10-03 18:58:37,386][starVLA.training.train_starvla][INFO] - Step 293, Loss: {'train/loss': 0.2187618613243103, 'train/grad_norm_pre_clip': 26.81960105895996, 'timing/data': 0.0003351799969095737, 'timing/model': 0.3963784170045983, 'learning_rate/qwen_vl_interface': 9.963681190922489e-06, 'learning_rate/action_model': 9.963681190922488e-05, 'epoch': 0.02}) +[2026-10-03 18:58:38,094][starVLA.training.train_starvla][INFO] - Step 294, Loss: {'train/loss': 0.37070584297180176, 'train/grad_norm_pre_clip': 16.731748580932617, 'timing/data': 0.00035423997906036675, 'timing/model': 0.3950394180137664, 'learning_rate/qwen_vl_interface': 9.963304341772391e-06, 'learning_rate/action_model': 9.963304341772391e-05, 'epoch': 0.02}) +[2026-10-03 18:58:38,802][starVLA.training.train_starvla][INFO] - Step 295, Loss: {'train/loss': 0.5728359818458557, 'train/grad_norm_pre_clip': 40.97107696533203, 'timing/data': 0.00036452000495046377, 'timing/model': 0.3948830879817251, 'learning_rate/qwen_vl_interface': 9.962925555160919e-06, 'learning_rate/action_model': 9.962925555160918e-05, 'epoch': 0.02}) +[2026-10-03 18:58:39,509][starVLA.training.train_starvla][INFO] - Step 296, Loss: {'train/loss': 1.0240461826324463, 'train/grad_norm_pre_clip': 135.4504852294922, 'timing/data': 0.00032725001801736653, 'timing/model': 0.39556944801006466, 'learning_rate/qwen_vl_interface': 9.96254483124377e-06, 'learning_rate/action_model': 9.96254483124377e-05, 'epoch': 0.02}) +[2026-10-03 18:58:40,417][starVLA.training.train_starvla][INFO] - Step 297, Loss: {'train/loss': 0.8382622599601746, 'train/grad_norm_pre_clip': 34.38356399536133, 'timing/data': 0.00027007999597117305, 'timing/model': 0.5943728470010683, 'learning_rate/qwen_vl_interface': 9.96216217017745e-06, 'learning_rate/action_model': 9.962162170177451e-05, 'epoch': 0.02}) +[2026-10-03 18:58:41,125][starVLA.training.train_starvla][INFO] - Step 298, Loss: {'train/loss': 1.114964246749878, 'train/grad_norm_pre_clip': 50.71297836303711, 'timing/data': 0.00023249001242220402, 'timing/model': 0.3954678179870825, 'learning_rate/qwen_vl_interface': 9.961777572119258e-06, 'learning_rate/action_model': 9.961777572119258e-05, 'epoch': 0.02}) +[2026-10-03 18:58:41,833][starVLA.training.train_starvla][INFO] - Step 299, Loss: {'train/loss': 0.5610697865486145, 'train/grad_norm_pre_clip': 36.11716079711914, 'timing/data': 0.00022493000142276287, 'timing/model': 0.39540111800306477, 'learning_rate/qwen_vl_interface': 9.961391037227286e-06, 'learning_rate/action_model': 9.961391037227284e-05, 'epoch': 0.02}) +[2026-10-03 18:58:42,539][starVLA.training.train_starvla][INFO] - Step 300, Loss: {'train/loss': 1.0211989879608154, 'train/grad_norm_pre_clip': 74.45352172851562, 'timing/data': 0.0002530899946577847, 'timing/model': 0.3938671079813503, 'learning_rate/qwen_vl_interface': 9.961002565660419e-06, 'learning_rate/action_model': 9.961002565660419e-05, 'epoch': 0.02}) +[2026-10-03 18:58:43,248][starVLA.training.train_starvla][INFO] - Step 301, Loss: {'train/loss': 0.46101751923561096, 'train/grad_norm_pre_clip': 17.669384002685547, 'timing/data': 0.00026532998890616, 'timing/model': 0.3963189469941426, 'learning_rate/qwen_vl_interface': 9.96061215757835e-06, 'learning_rate/action_model': 9.960612157578349e-05, 'epoch': 0.02}) +[2026-10-03 18:58:44,155][starVLA.training.train_starvla][INFO] - Step 302, Loss: {'train/loss': 0.30084556341171265, 'train/grad_norm_pre_clip': 56.423179626464844, 'timing/data': 0.0002960299898404628, 'timing/model': 0.5966802859911695, 'learning_rate/qwen_vl_interface': 9.96021981314156e-06, 'learning_rate/action_model': 9.96021981314156e-05, 'epoch': 0.02}) +[2026-10-03 18:58:44,862][starVLA.training.train_starvla][INFO] - Step 303, Loss: {'train/loss': 0.4041271507740021, 'train/grad_norm_pre_clip': 19.955842971801758, 'timing/data': 0.0002668999950401485, 'timing/model': 0.3940614070161246, 'learning_rate/qwen_vl_interface': 9.959825532511322e-06, 'learning_rate/action_model': 9.959825532511322e-05, 'epoch': 0.02}) +[2026-10-03 18:58:45,572][starVLA.training.train_starvla][INFO] - Step 304, Loss: {'train/loss': 0.8340743780136108, 'train/grad_norm_pre_clip': 91.31771087646484, 'timing/data': 0.0003869800129905343, 'timing/model': 0.3963433580065612, 'learning_rate/qwen_vl_interface': 9.959429315849715e-06, 'learning_rate/action_model': 9.959429315849715e-05, 'epoch': 0.02}) +[2026-10-03 18:58:46,280][starVLA.training.train_starvla][INFO] - Step 305, Loss: {'train/loss': 0.468100905418396, 'train/grad_norm_pre_clip': 92.74420166015625, 'timing/data': 0.00032715999986976385, 'timing/model': 0.39502761798212305, 'learning_rate/qwen_vl_interface': 9.959031163319606e-06, 'learning_rate/action_model': 9.959031163319607e-05, 'epoch': 0.02}) +[2026-10-03 18:58:46,989][starVLA.training.train_starvla][INFO] - Step 306, Loss: {'train/loss': 0.49188143014907837, 'train/grad_norm_pre_clip': 19.002256393432617, 'timing/data': 0.00022317998809739947, 'timing/model': 0.3955830880149733, 'learning_rate/qwen_vl_interface': 9.958631075084664e-06, 'learning_rate/action_model': 9.958631075084664e-05, 'epoch': 0.02}) +[2026-10-03 18:58:47,695][starVLA.training.train_starvla][INFO] - Step 307, Loss: {'train/loss': 0.16627007722854614, 'train/grad_norm_pre_clip': 44.16700744628906, 'timing/data': 0.00037647999124601483, 'timing/model': 0.3957954680081457, 'learning_rate/qwen_vl_interface': 9.958229051309346e-06, 'learning_rate/action_model': 9.958229051309346e-05, 'epoch': 0.02}) +[2026-10-03 18:58:48,599][starVLA.training.train_starvla][INFO] - Step 308, Loss: {'train/loss': 0.913163959980011, 'train/grad_norm_pre_clip': 112.9192123413086, 'timing/data': 0.0003423099988140166, 'timing/model': 0.3956417879962828, 'learning_rate/qwen_vl_interface': 9.95782509215891e-06, 'learning_rate/action_model': 9.95782509215891e-05, 'epoch': 0.02}) +[2026-10-03 18:58:49,307][starVLA.training.train_starvla][INFO] - Step 309, Loss: {'train/loss': 1.0975348949432373, 'train/grad_norm_pre_clip': 125.35497283935547, 'timing/data': 0.0002633499971125275, 'timing/model': 0.3968552469741553, 'learning_rate/qwen_vl_interface': 9.95741919779941e-06, 'learning_rate/action_model': 9.95741919779941e-05, 'epoch': 0.02}) +[2026-10-03 18:58:50,014][starVLA.training.train_starvla][INFO] - Step 310, Loss: {'train/loss': 0.510273277759552, 'train/grad_norm_pre_clip': 54.31460189819336, 'timing/data': 0.0003719799860846251, 'timing/model': 0.3945496969972737, 'learning_rate/qwen_vl_interface': 9.957011368397692e-06, 'learning_rate/action_model': 9.95701136839769e-05, 'epoch': 0.02}) +[2026-10-03 18:58:50,722][starVLA.training.train_starvla][INFO] - Step 311, Loss: {'train/loss': 0.6478946208953857, 'train/grad_norm_pre_clip': 36.11037063598633, 'timing/data': 0.00033492001239210367, 'timing/model': 0.3950669880141504, 'learning_rate/qwen_vl_interface': 9.956601604121398e-06, 'learning_rate/action_model': 9.956601604121397e-05, 'epoch': 0.02}) +[2026-10-03 18:58:51,430][starVLA.training.train_starvla][INFO] - Step 312, Loss: {'train/loss': 1.7373313903808594, 'train/grad_norm_pre_clip': 264.25567626953125, 'timing/data': 0.00027802001568488777, 'timing/model': 0.394099617988104, 'learning_rate/qwen_vl_interface': 9.95618990513897e-06, 'learning_rate/action_model': 9.95618990513897e-05, 'epoch': 0.02}) +[2026-10-03 18:58:52,338][starVLA.training.train_starvla][INFO] - Step 313, Loss: {'train/loss': 1.2246851921081543, 'train/grad_norm_pre_clip': 210.03082275390625, 'timing/data': 0.00034282999695278704, 'timing/model': 0.3948985780007206, 'learning_rate/qwen_vl_interface': 9.95577627161964e-06, 'learning_rate/action_model': 9.95577627161964e-05, 'epoch': 0.02}) +[2026-10-03 18:58:53,045][starVLA.training.train_starvla][INFO] - Step 314, Loss: {'train/loss': 1.3803774118423462, 'train/grad_norm_pre_clip': 226.1035919189453, 'timing/data': 0.00036179000744596124, 'timing/model': 0.39399841797421686, 'learning_rate/qwen_vl_interface': 9.955360703733438e-06, 'learning_rate/action_model': 9.955360703733437e-05, 'epoch': 0.02}) +[2026-10-03 18:58:53,749][starVLA.training.train_starvla][INFO] - Step 315, Loss: {'train/loss': 1.869375228881836, 'train/grad_norm_pre_clip': 153.4903106689453, 'timing/data': 0.00026989998877979815, 'timing/model': 0.39372952800476924, 'learning_rate/qwen_vl_interface': 9.954943201651187e-06, 'learning_rate/action_model': 9.954943201651187e-05, 'epoch': 0.02}) +[2026-10-03 18:58:54,458][starVLA.training.train_starvla][INFO] - Step 316, Loss: {'train/loss': 0.6662396788597107, 'train/grad_norm_pre_clip': 78.59748077392578, 'timing/data': 0.000245789997279644, 'timing/model': 0.39612471699365415, 'learning_rate/qwen_vl_interface': 9.954523765544507e-06, 'learning_rate/action_model': 9.954523765544508e-05, 'epoch': 0.02}) +[2026-10-03 18:58:55,165][starVLA.training.train_starvla][INFO] - Step 317, Loss: {'train/loss': 0.5513752698898315, 'train/grad_norm_pre_clip': 45.486236572265625, 'timing/data': 0.00030849999166093767, 'timing/model': 0.3948534869996365, 'learning_rate/qwen_vl_interface': 9.954102395585814e-06, 'learning_rate/action_model': 9.954102395585813e-05, 'epoch': 0.02}) +[2026-10-03 18:58:56,074][starVLA.training.train_starvla][INFO] - Step 318, Loss: {'train/loss': 0.537044107913971, 'train/grad_norm_pre_clip': 39.33738327026367, 'timing/data': 0.0003332200285512954, 'timing/model': 0.595454927009996, 'learning_rate/qwen_vl_interface': 9.953679091948313e-06, 'learning_rate/action_model': 9.953679091948314e-05, 'epoch': 0.02}) +[2026-10-03 18:58:56,782][starVLA.training.train_starvla][INFO] - Step 319, Loss: {'train/loss': 0.21428203582763672, 'train/grad_norm_pre_clip': 17.96404457092285, 'timing/data': 0.00034755998058244586, 'timing/model': 0.39541032700799406, 'learning_rate/qwen_vl_interface': 9.953253854806013e-06, 'learning_rate/action_model': 9.953253854806012e-05, 'epoch': 0.02}) +[2026-10-03 18:58:57,488][starVLA.training.train_starvla][INFO] - Step 320, Loss: {'train/loss': 0.701533854007721, 'train/grad_norm_pre_clip': 73.6661148071289, 'timing/data': 0.0003266099956817925, 'timing/model': 0.39395529800094664, 'learning_rate/qwen_vl_interface': 9.95282668433371e-06, 'learning_rate/action_model': 9.952826684333709e-05, 'epoch': 0.02}) +[2026-10-03 18:58:58,196][starVLA.training.train_starvla][INFO] - Step 321, Loss: {'train/loss': 0.7392977476119995, 'train/grad_norm_pre_clip': 56.690635681152344, 'timing/data': 0.0002731800195761025, 'timing/model': 0.39622253898414783, 'learning_rate/qwen_vl_interface': 9.952397580706998e-06, 'learning_rate/action_model': 9.952397580706998e-05, 'epoch': 0.02}) +[2026-10-03 18:58:58,905][starVLA.training.train_starvla][INFO] - Step 322, Loss: {'train/loss': 0.5752724409103394, 'train/grad_norm_pre_clip': 93.76162719726562, 'timing/data': 0.00037148999399505556, 'timing/model': 0.3961968280200381, 'learning_rate/qwen_vl_interface': 9.951966544102267e-06, 'learning_rate/action_model': 9.951966544102266e-05, 'epoch': 0.02}) +[2026-10-03 18:58:59,816][starVLA.training.train_starvla][INFO] - Step 323, Loss: {'train/loss': 0.695387601852417, 'train/grad_norm_pre_clip': 44.29103088378906, 'timing/data': 0.0003679200017359108, 'timing/model': 0.5991101859835908, 'learning_rate/qwen_vl_interface': 9.951533574696696e-06, 'learning_rate/action_model': 9.951533574696697e-05, 'epoch': 0.02}) +[2026-10-03 18:59:00,522][starVLA.training.train_starvla][INFO] - Step 324, Loss: {'train/loss': 0.9183316826820374, 'train/grad_norm_pre_clip': 41.75634002685547, 'timing/data': 0.0003515099815558642, 'timing/model': 0.39548961800755933, 'learning_rate/qwen_vl_interface': 9.951098672668266e-06, 'learning_rate/action_model': 9.951098672668265e-05, 'epoch': 0.02}) +[2026-10-03 18:59:01,230][starVLA.training.train_starvla][INFO] - Step 325, Loss: {'train/loss': 0.5859010815620422, 'train/grad_norm_pre_clip': 26.024702072143555, 'timing/data': 0.0002805099938996136, 'timing/model': 0.3946018780115992, 'learning_rate/qwen_vl_interface': 9.95066183819575e-06, 'learning_rate/action_model': 9.950661838195749e-05, 'epoch': 0.02}) +[2026-10-03 18:59:01,936][starVLA.training.train_starvla][INFO] - Step 326, Loss: {'train/loss': 0.9855979681015015, 'train/grad_norm_pre_clip': 29.712997436523438, 'timing/data': 0.0003515200223773718, 'timing/model': 0.39352935799979605, 'learning_rate/qwen_vl_interface': 9.95022307145871e-06, 'learning_rate/action_model': 9.950223071458708e-05, 'epoch': 0.02}) +[2026-10-03 18:59:02,643][starVLA.training.train_starvla][INFO] - Step 327, Loss: {'train/loss': 0.7574514746665955, 'train/grad_norm_pre_clip': 55.98017501831055, 'timing/data': 0.0003655900072772056, 'timing/model': 0.3958200979977846, 'learning_rate/qwen_vl_interface': 9.949782372637507e-06, 'learning_rate/action_model': 9.949782372637506e-05, 'epoch': 0.02}) +[2026-10-03 18:59:03,349][starVLA.training.train_starvla][INFO] - Step 328, Loss: {'train/loss': 0.4106907844543457, 'train/grad_norm_pre_clip': 20.517826080322266, 'timing/data': 0.000236109975958243, 'timing/model': 0.3956278480181936, 'learning_rate/qwen_vl_interface': 9.949339741913297e-06, 'learning_rate/action_model': 9.949339741913297e-05, 'epoch': 0.02}) +[2026-10-03 18:59:04,257][starVLA.training.train_starvla][INFO] - Step 329, Loss: {'train/loss': 0.5512235164642334, 'train/grad_norm_pre_clip': 29.336307525634766, 'timing/data': 0.0002974800008814782, 'timing/model': 0.3956372389802709, 'learning_rate/qwen_vl_interface': 9.948895179468032e-06, 'learning_rate/action_model': 9.948895179468031e-05, 'epoch': 0.02}) +[2026-10-03 18:59:04,941][starVLA.training.train_starvla][INFO] - Step 330, Loss: {'train/loss': 1.0444916486740112, 'train/grad_norm_pre_clip': 36.23514175415039, 'timing/data': 0.0002258300082758069, 'timing/model': 0.3969079380040057, 'learning_rate/qwen_vl_interface': 9.948448685484448e-06, 'learning_rate/action_model': 9.948448685484447e-05, 'epoch': 0.02}) +[2026-10-03 18:59:05,649][starVLA.training.train_starvla][INFO] - Step 331, Loss: {'train/loss': 0.40095797181129456, 'train/grad_norm_pre_clip': 31.685405731201172, 'timing/data': 0.0004044699890073389, 'timing/model': 0.39608757800306194, 'learning_rate/qwen_vl_interface': 9.948000260146087e-06, 'learning_rate/action_model': 9.948000260146087e-05, 'epoch': 0.02}) +[2026-10-03 18:59:06,357][starVLA.training.train_starvla][INFO] - Step 332, Loss: {'train/loss': 0.5769180655479431, 'train/grad_norm_pre_clip': 58.4259147644043, 'timing/data': 0.00023422998492605984, 'timing/model': 0.39535945901297964, 'learning_rate/qwen_vl_interface': 9.947549903637277e-06, 'learning_rate/action_model': 9.947549903637278e-05, 'epoch': 0.02}) +[2026-10-03 18:59:07,066][starVLA.training.train_starvla][INFO] - Step 333, Loss: {'train/loss': 0.7268160581588745, 'train/grad_norm_pre_clip': 49.431495666503906, 'timing/data': 0.000423610006691888, 'timing/model': 0.3964601780171506, 'learning_rate/qwen_vl_interface': 9.947097616143145e-06, 'learning_rate/action_model': 9.947097616143144e-05, 'epoch': 0.02}) +[2026-10-03 18:59:07,971][starVLA.training.train_starvla][INFO] - Step 334, Loss: {'train/loss': 0.8945156335830688, 'train/grad_norm_pre_clip': 41.47315979003906, 'timing/data': 0.0004896199970971793, 'timing/model': 0.39361379799083807, 'learning_rate/qwen_vl_interface': 9.946643397849607e-06, 'learning_rate/action_model': 9.946643397849607e-05, 'epoch': 0.02}) +[2026-10-03 18:59:08,679][starVLA.training.train_starvla][INFO] - Step 335, Loss: {'train/loss': 0.3643186390399933, 'train/grad_norm_pre_clip': 15.249556541442871, 'timing/data': 0.00035493000177666545, 'timing/model': 0.3952036979899276, 'learning_rate/qwen_vl_interface': 9.946187248943375e-06, 'learning_rate/action_model': 9.946187248943374e-05, 'epoch': 0.02}) +[2026-10-03 18:59:09,386][starVLA.training.train_starvla][INFO] - Step 336, Loss: {'train/loss': 0.2938099801540375, 'train/grad_norm_pre_clip': 35.11990737915039, 'timing/data': 0.0003582900098990649, 'timing/model': 0.39498167802230455, 'learning_rate/qwen_vl_interface': 9.945729169611957e-06, 'learning_rate/action_model': 9.945729169611958e-05, 'epoch': 0.02}) +[2026-10-03 18:59:10,093][starVLA.training.train_starvla][INFO] - Step 337, Loss: {'train/loss': 0.20262859761714935, 'train/grad_norm_pre_clip': 10.870614051818848, 'timing/data': 0.0002603699977044016, 'timing/model': 0.39455987900146283, 'learning_rate/qwen_vl_interface': 9.94526916004365e-06, 'learning_rate/action_model': 9.94526916004365e-05, 'epoch': 0.02}) +[2026-10-03 18:59:10,799][starVLA.training.train_starvla][INFO] - Step 338, Loss: {'train/loss': 1.0286777019500732, 'train/grad_norm_pre_clip': 56.09912109375, 'timing/data': 0.0002571099903434515, 'timing/model': 0.3951922770065721, 'learning_rate/qwen_vl_interface': 9.944807220427549e-06, 'learning_rate/action_model': 9.944807220427548e-05, 'epoch': 0.02}) +[2026-10-03 18:59:11,701][starVLA.training.train_starvla][INFO] - Step 339, Loss: {'train/loss': 0.6166476607322693, 'train/grad_norm_pre_clip': 31.62401008605957, 'timing/data': 0.0002908199967350811, 'timing/model': 0.5912043470016215, 'learning_rate/qwen_vl_interface': 9.944343350953534e-06, 'learning_rate/action_model': 9.944343350953533e-05, 'epoch': 0.02}) +[2026-10-03 18:59:12,409][starVLA.training.train_starvla][INFO] - Step 340, Loss: {'train/loss': 1.0033479928970337, 'train/grad_norm_pre_clip': 33.07905960083008, 'timing/data': 0.00022873000125400722, 'timing/model': 0.39577139797620475, 'learning_rate/qwen_vl_interface': 9.94387755181229e-06, 'learning_rate/action_model': 9.94387755181229e-05, 'epoch': 0.02}) +[2026-10-03 18:59:13,115][starVLA.training.train_starvla][INFO] - Step 341, Loss: {'train/loss': 0.9955122470855713, 'train/grad_norm_pre_clip': 85.3208236694336, 'timing/data': 0.00029247000929899514, 'timing/model': 0.39432779801427387, 'learning_rate/qwen_vl_interface': 9.943409823195289e-06, 'learning_rate/action_model': 9.943409823195288e-05, 'epoch': 0.02}) +[2026-10-03 18:59:13,823][starVLA.training.train_starvla][INFO] - Step 342, Loss: {'train/loss': 0.6350969076156616, 'train/grad_norm_pre_clip': 86.1883544921875, 'timing/data': 0.0002442199911456555, 'timing/model': 0.39517618899117224, 'learning_rate/qwen_vl_interface': 9.942940165294794e-06, 'learning_rate/action_model': 9.942940165294793e-05, 'epoch': 0.03}) +[2026-10-03 18:59:14,532][starVLA.training.train_starvla][INFO] - Step 343, Loss: {'train/loss': 0.6400772929191589, 'train/grad_norm_pre_clip': 22.45589256286621, 'timing/data': 0.0002245799987576902, 'timing/model': 0.39682840899331495, 'learning_rate/qwen_vl_interface': 9.942468578303865e-06, 'learning_rate/action_model': 9.942468578303865e-05, 'epoch': 0.03}) +[2026-10-03 18:59:15,440][starVLA.training.train_starvla][INFO] - Step 344, Loss: {'train/loss': 0.32722020149230957, 'train/grad_norm_pre_clip': 33.8183479309082, 'timing/data': 0.000262120011029765, 'timing/model': 0.5958110470091924, 'learning_rate/qwen_vl_interface': 9.941995062416354e-06, 'learning_rate/action_model': 9.941995062416354e-05, 'epoch': 0.03}) +[2026-10-03 18:59:16,147][starVLA.training.train_starvla][INFO] - Step 345, Loss: {'train/loss': 1.3956313133239746, 'train/grad_norm_pre_clip': 201.78688049316406, 'timing/data': 0.00036358999204821885, 'timing/model': 0.394395698007429, 'learning_rate/qwen_vl_interface': 9.941519617826906e-06, 'learning_rate/action_model': 9.941519617826906e-05, 'epoch': 0.03}) +[2026-10-03 18:59:16,855][starVLA.training.train_starvla][INFO] - Step 346, Loss: {'train/loss': 1.622534990310669, 'train/grad_norm_pre_clip': 169.76185607910156, 'timing/data': 0.00035292000393383205, 'timing/model': 0.39525012800004333, 'learning_rate/qwen_vl_interface': 9.941042244730954e-06, 'learning_rate/action_model': 9.941042244730955e-05, 'epoch': 0.03}) +[2026-10-03 18:59:17,560][starVLA.training.train_starvla][INFO] - Step 347, Loss: {'train/loss': 1.1188538074493408, 'train/grad_norm_pre_clip': 116.05050659179688, 'timing/data': 0.0002150599902961403, 'timing/model': 0.39551248800125904, 'learning_rate/qwen_vl_interface': 9.940562943324736e-06, 'learning_rate/action_model': 9.940562943324735e-05, 'epoch': 0.03}) +[2026-10-03 18:59:18,267][starVLA.training.train_starvla][INFO] - Step 348, Loss: {'train/loss': 0.7342791557312012, 'train/grad_norm_pre_clip': 72.42252349853516, 'timing/data': 0.0003153200086671859, 'timing/model': 0.39519380798446946, 'learning_rate/qwen_vl_interface': 9.940081713805266e-06, 'learning_rate/action_model': 9.940081713805266e-05, 'epoch': 0.03}) +[2026-10-03 18:59:18,974][starVLA.training.train_starvla][INFO] - Step 349, Loss: {'train/loss': 0.5654780864715576, 'train/grad_norm_pre_clip': 51.42838668823242, 'timing/data': 0.0003169099800288677, 'timing/model': 0.3956085279933177, 'learning_rate/qwen_vl_interface': 9.939598556370366e-06, 'learning_rate/action_model': 9.939598556370365e-05, 'epoch': 0.03}) +[2026-10-03 18:59:19,874][starVLA.training.train_starvla][INFO] - Step 350, Loss: {'train/loss': 0.9612220525741577, 'train/grad_norm_pre_clip': 44.54777908325195, 'timing/data': 0.00023070001043379307, 'timing/model': 0.393808558001183, 'learning_rate/qwen_vl_interface': 9.93911347121864e-06, 'learning_rate/action_model': 9.939113471218639e-05, 'epoch': 0.03}) +[2026-10-03 18:59:20,582][starVLA.training.train_starvla][INFO] - Step 351, Loss: {'train/loss': 0.93289715051651, 'train/grad_norm_pre_clip': 104.30921173095703, 'timing/data': 0.000348920002579689, 'timing/model': 0.3962840879976284, 'learning_rate/qwen_vl_interface': 9.938626458549488e-06, 'learning_rate/action_model': 9.938626458549488e-05, 'epoch': 0.03}) +[2026-10-03 18:59:21,291][starVLA.training.train_starvla][INFO] - Step 352, Loss: {'train/loss': 0.3179379403591156, 'train/grad_norm_pre_clip': 17.436079025268555, 'timing/data': 0.00026634999085217714, 'timing/model': 0.39550449899979867, 'learning_rate/qwen_vl_interface': 9.938137518563106e-06, 'learning_rate/action_model': 9.938137518563106e-05, 'epoch': 0.03}) +[2026-10-03 18:59:21,998][starVLA.training.train_starvla][INFO] - Step 353, Loss: {'train/loss': 0.9304578304290771, 'train/grad_norm_pre_clip': 26.8675479888916, 'timing/data': 0.0003594699956011027, 'timing/model': 0.39480700899730437, 'learning_rate/qwen_vl_interface': 9.937646651460475e-06, 'learning_rate/action_model': 9.937646651460474e-05, 'epoch': 0.03}) +[2026-10-03 18:59:22,703][starVLA.training.train_starvla][INFO] - Step 354, Loss: {'train/loss': 0.5024014711380005, 'train/grad_norm_pre_clip': 19.91318702697754, 'timing/data': 0.0004187199810985476, 'timing/model': 0.39464352899813093, 'learning_rate/qwen_vl_interface': 9.937153857443375e-06, 'learning_rate/action_model': 9.937153857443374e-05, 'epoch': 0.03}) +[2026-10-03 18:59:23,604][starVLA.training.train_starvla][INFO] - Step 355, Loss: {'train/loss': 0.2302655577659607, 'train/grad_norm_pre_clip': 26.874563217163086, 'timing/data': 0.00036399997770786285, 'timing/model': 0.39331083800061606, 'learning_rate/qwen_vl_interface': 9.936659136714372e-06, 'learning_rate/action_model': 9.936659136714371e-05, 'epoch': 0.03}) +[2026-10-03 18:59:24,310][starVLA.training.train_starvla][INFO] - Step 356, Loss: {'train/loss': 0.6503304243087769, 'train/grad_norm_pre_clip': 38.65650177001953, 'timing/data': 0.00032163001014851034, 'timing/model': 0.3933786589768715, 'learning_rate/qwen_vl_interface': 9.936162489476827e-06, 'learning_rate/action_model': 9.936162489476828e-05, 'epoch': 0.03}) +[2026-10-03 18:59:25,015][starVLA.training.train_starvla][INFO] - Step 357, Loss: {'train/loss': 0.42432501912117004, 'train/grad_norm_pre_clip': 81.9867935180664, 'timing/data': 0.0002568300114944577, 'timing/model': 0.39432248900993727, 'learning_rate/qwen_vl_interface': 9.935663915934899e-06, 'learning_rate/action_model': 9.935663915934898e-05, 'epoch': 0.03}) +[2026-10-03 18:59:25,721][starVLA.training.train_starvla][INFO] - Step 358, Loss: {'train/loss': 0.7791372537612915, 'train/grad_norm_pre_clip': 19.15128517150879, 'timing/data': 0.0002243999915663153, 'timing/model': 0.39463396900100634, 'learning_rate/qwen_vl_interface': 9.935163416293525e-06, 'learning_rate/action_model': 9.935163416293525e-05, 'epoch': 0.03}) +[2026-10-03 18:59:26,426][starVLA.training.train_starvla][INFO] - Step 359, Loss: {'train/loss': 0.2827475070953369, 'train/grad_norm_pre_clip': 8.220967292785645, 'timing/data': 0.000311740004690364, 'timing/model': 0.3949263079848606, 'learning_rate/qwen_vl_interface': 9.934660990758445e-06, 'learning_rate/action_model': 9.934660990758445e-05, 'epoch': 0.03}) +[2026-10-03 18:59:27,331][starVLA.training.train_starvla][INFO] - Step 360, Loss: {'train/loss': 1.0693806409835815, 'train/grad_norm_pre_clip': 35.8637580871582, 'timing/data': 0.0003022699966095388, 'timing/model': 0.5937414070067462, 'learning_rate/qwen_vl_interface': 9.934156639536186e-06, 'learning_rate/action_model': 9.934156639536187e-05, 'epoch': 0.03}) +[2026-10-03 18:59:28,038][starVLA.training.train_starvla][INFO] - Step 361, Loss: {'train/loss': 0.13001900911331177, 'train/grad_norm_pre_clip': 32.199790954589844, 'timing/data': 0.000262529996689409, 'timing/model': 0.39569662799476646, 'learning_rate/qwen_vl_interface': 9.93365036283407e-06, 'learning_rate/action_model': 9.93365036283407e-05, 'epoch': 0.03}) +[2026-10-03 18:59:28,744][starVLA.training.train_starvla][INFO] - Step 362, Loss: {'train/loss': 0.4137026369571686, 'train/grad_norm_pre_clip': 20.670801162719727, 'timing/data': 0.00033297998015768826, 'timing/model': 0.39511825900990516, 'learning_rate/qwen_vl_interface': 9.933142160860208e-06, 'learning_rate/action_model': 9.933142160860208e-05, 'epoch': 0.03}) +[2026-10-03 18:59:29,450][starVLA.training.train_starvla][INFO] - Step 363, Loss: {'train/loss': 0.5416573882102966, 'train/grad_norm_pre_clip': 23.90521812438965, 'timing/data': 0.0003436499973759055, 'timing/model': 0.39483437899616547, 'learning_rate/qwen_vl_interface': 9.932632033823499e-06, 'learning_rate/action_model': 9.932632033823499e-05, 'epoch': 0.03}) +[2026-10-03 18:59:30,154][starVLA.training.train_starvla][INFO] - Step 364, Loss: {'train/loss': 0.31221315264701843, 'train/grad_norm_pre_clip': 26.425003051757812, 'timing/data': 0.00034994998713955283, 'timing/model': 0.3938248880149331, 'learning_rate/qwen_vl_interface': 9.93211998193364e-06, 'learning_rate/action_model': 9.932119981933639e-05, 'epoch': 0.03}) +[2026-10-03 18:59:31,058][starVLA.training.train_starvla][INFO] - Step 365, Loss: {'train/loss': 0.43457621335983276, 'train/grad_norm_pre_clip': 22.498960494995117, 'timing/data': 0.000225800002226606, 'timing/model': 0.5939002270170022, 'learning_rate/qwen_vl_interface': 9.931606005401117e-06, 'learning_rate/action_model': 9.931606005401116e-05, 'epoch': 0.03}) +[2026-10-03 18:59:31,763][starVLA.training.train_starvla][INFO] - Step 366, Loss: {'train/loss': 0.1374945044517517, 'train/grad_norm_pre_clip': 41.71653747558594, 'timing/data': 0.0003096100117545575, 'timing/model': 0.3933864389837254, 'learning_rate/qwen_vl_interface': 9.931090104437202e-06, 'learning_rate/action_model': 9.931090104437202e-05, 'epoch': 0.03}) +[2026-10-03 18:59:32,467][starVLA.training.train_starvla][INFO] - Step 367, Loss: {'train/loss': 0.4439888596534729, 'train/grad_norm_pre_clip': 21.64219856262207, 'timing/data': 0.00021115000708959997, 'timing/model': 0.39415968902176246, 'learning_rate/qwen_vl_interface': 9.930572279253967e-06, 'learning_rate/action_model': 9.930572279253967e-05, 'epoch': 0.03}) +[2026-10-03 18:59:33,171][starVLA.training.train_starvla][INFO] - Step 368, Loss: {'train/loss': 0.743319034576416, 'train/grad_norm_pre_clip': 93.84274291992188, 'timing/data': 0.00024641997879371047, 'timing/model': 0.3938567180011887, 'learning_rate/qwen_vl_interface': 9.930052530064269e-06, 'learning_rate/action_model': 9.93005253006427e-05, 'epoch': 0.03}) +[2026-10-03 18:59:33,877][starVLA.training.train_starvla][INFO] - Step 369, Loss: {'train/loss': 0.8464106321334839, 'train/grad_norm_pre_clip': 40.36113357543945, 'timing/data': 0.0002297000028192997, 'timing/model': 0.39417928899638355, 'learning_rate/qwen_vl_interface': 9.929530857081757e-06, 'learning_rate/action_model': 9.929530857081757e-05, 'epoch': 0.03}) +[2026-10-03 18:59:34,587][starVLA.training.train_starvla][INFO] - Step 370, Loss: {'train/loss': 0.48042383790016174, 'train/grad_norm_pre_clip': 183.1593780517578, 'timing/data': 0.0003320099785923958, 'timing/model': 0.39753345798817463, 'learning_rate/qwen_vl_interface': 9.92900726052087e-06, 'learning_rate/action_model': 9.929007260520869e-05, 'epoch': 0.03}) +[2026-10-03 18:59:35,494][starVLA.training.train_starvla][INFO] - Step 371, Loss: {'train/loss': 0.8187981843948364, 'train/grad_norm_pre_clip': 121.30441284179688, 'timing/data': 0.00028425001073628664, 'timing/model': 0.3939263079955708, 'learning_rate/qwen_vl_interface': 9.92848174059684e-06, 'learning_rate/action_model': 9.92848174059684e-05, 'epoch': 0.03}) +[2026-10-03 18:59:36,202][starVLA.training.train_starvla][INFO] - Step 372, Loss: {'train/loss': 0.9531823396682739, 'train/grad_norm_pre_clip': 52.49489974975586, 'timing/data': 0.0002734099980443716, 'timing/model': 0.3955062380118761, 'learning_rate/qwen_vl_interface': 9.92795429752569e-06, 'learning_rate/action_model': 9.927954297525689e-05, 'epoch': 0.03}) +[2026-10-03 18:59:36,911][starVLA.training.train_starvla][INFO] - Step 373, Loss: {'train/loss': 0.5480421185493469, 'train/grad_norm_pre_clip': 64.77228546142578, 'timing/data': 0.0003313800261821598, 'timing/model': 0.397082657989813, 'learning_rate/qwen_vl_interface': 9.927424931524228e-06, 'learning_rate/action_model': 9.927424931524228e-05, 'epoch': 0.03}) +[2026-10-03 18:59:37,619][starVLA.training.train_starvla][INFO] - Step 374, Loss: {'train/loss': 0.713473379611969, 'train/grad_norm_pre_clip': 40.87916564941406, 'timing/data': 0.0003166999958921224, 'timing/model': 0.3946761089900974, 'learning_rate/qwen_vl_interface': 9.926893642810058e-06, 'learning_rate/action_model': 9.926893642810059e-05, 'epoch': 0.03}) +[2026-10-03 18:59:38,326][starVLA.training.train_starvla][INFO] - Step 375, Loss: {'train/loss': 0.5282146334648132, 'train/grad_norm_pre_clip': 40.43770217895508, 'timing/data': 0.00024095000117085874, 'timing/model': 0.3948925189906731, 'learning_rate/qwen_vl_interface': 9.926360431601575e-06, 'learning_rate/action_model': 9.926360431601574e-05, 'epoch': 0.03}) +[2026-10-03 18:59:39,223][starVLA.training.train_starvla][INFO] - Step 376, Loss: {'train/loss': 0.7439874410629272, 'train/grad_norm_pre_clip': 75.9139175415039, 'timing/data': 0.0002636399876791984, 'timing/model': 0.5849975979945157, 'learning_rate/qwen_vl_interface': 9.92582529811796e-06, 'learning_rate/action_model': 9.925825298117959e-05, 'epoch': 0.03}) +[2026-10-03 18:59:39,928][starVLA.training.train_starvla][INFO] - Step 377, Loss: {'train/loss': 0.5828917026519775, 'train/grad_norm_pre_clip': 41.15427017211914, 'timing/data': 0.0002880200045183301, 'timing/model': 0.3944816790171899, 'learning_rate/qwen_vl_interface': 9.925288242579184e-06, 'learning_rate/action_model': 9.925288242579184e-05, 'epoch': 0.03}) +[2026-10-03 18:59:40,631][starVLA.training.train_starvla][INFO] - Step 378, Loss: {'train/loss': 0.4504983127117157, 'train/grad_norm_pre_clip': 64.28105163574219, 'timing/data': 0.0003012999950442463, 'timing/model': 0.3925316679815296, 'learning_rate/qwen_vl_interface': 9.924749265206015e-06, 'learning_rate/action_model': 9.924749265206015e-05, 'epoch': 0.03}) +[2026-10-03 18:59:41,337][starVLA.training.train_starvla][INFO] - Step 379, Loss: {'train/loss': 0.828992486000061, 'train/grad_norm_pre_clip': 73.15385437011719, 'timing/data': 0.0003343499847687781, 'timing/model': 0.3952062469907105, 'learning_rate/qwen_vl_interface': 9.924208366220003e-06, 'learning_rate/action_model': 9.924208366220002e-05, 'epoch': 0.03}) +[2026-10-03 18:59:42,044][starVLA.training.train_starvla][INFO] - Step 380, Loss: {'train/loss': 0.30496692657470703, 'train/grad_norm_pre_clip': 17.161054611206055, 'timing/data': 0.00032478998764418066, 'timing/model': 0.39496087800944224, 'learning_rate/qwen_vl_interface': 9.923665545843492e-06, 'learning_rate/action_model': 9.923665545843491e-05, 'epoch': 0.03}) +[2026-10-03 18:59:42,946][starVLA.training.train_starvla][INFO] - Step 381, Loss: {'train/loss': 0.23423726856708527, 'train/grad_norm_pre_clip': 30.710145950317383, 'timing/data': 0.000285919988527894, 'timing/model': 0.5894453980145045, 'learning_rate/qwen_vl_interface': 9.923120804299615e-06, 'learning_rate/action_model': 9.923120804299613e-05, 'epoch': 0.03}) +[2026-10-03 18:59:43,649][starVLA.training.train_starvla][INFO] - Step 382, Loss: {'train/loss': 0.723663330078125, 'train/grad_norm_pre_clip': 45.846092224121094, 'timing/data': 0.00024639000184834003, 'timing/model': 0.392883708991576, 'learning_rate/qwen_vl_interface': 9.922574141812295e-06, 'learning_rate/action_model': 9.922574141812294e-05, 'epoch': 0.03}) +[2026-10-03 18:59:44,357][starVLA.training.train_starvla][INFO] - Step 383, Loss: {'train/loss': 0.8801763653755188, 'train/grad_norm_pre_clip': 82.60401916503906, 'timing/data': 0.00042335002217441797, 'timing/model': 0.39554588799364865, 'learning_rate/qwen_vl_interface': 9.922025558606242e-06, 'learning_rate/action_model': 9.922025558606242e-05, 'epoch': 0.03}) +[2026-10-03 18:59:45,065][starVLA.training.train_starvla][INFO] - Step 384, Loss: {'train/loss': 0.8996556401252747, 'train/grad_norm_pre_clip': 51.944725036621094, 'timing/data': 0.0002656799915712327, 'timing/model': 0.3949996879964601, 'learning_rate/qwen_vl_interface': 9.92147505490696e-06, 'learning_rate/action_model': 9.92147505490696e-05, 'epoch': 0.03}) +[2026-10-03 18:59:45,771][starVLA.training.train_starvla][INFO] - Step 385, Loss: {'train/loss': 0.9978156089782715, 'train/grad_norm_pre_clip': 99.08885955810547, 'timing/data': 0.00039380998350679874, 'timing/model': 0.39424215801409446, 'learning_rate/qwen_vl_interface': 9.920922630940741e-06, 'learning_rate/action_model': 9.920922630940742e-05, 'epoch': 0.03}) +[2026-10-03 18:59:46,477][starVLA.training.train_starvla][INFO] - Step 386, Loss: {'train/loss': 0.7899414896965027, 'train/grad_norm_pre_clip': 70.56332397460938, 'timing/data': 0.00032881999504752457, 'timing/model': 0.3945160179864615, 'learning_rate/qwen_vl_interface': 9.920368286934664e-06, 'learning_rate/action_model': 9.920368286934663e-05, 'epoch': 0.03}) +[2026-10-03 18:59:47,371][starVLA.training.train_starvla][INFO] - Step 387, Loss: {'train/loss': 0.4746987223625183, 'train/grad_norm_pre_clip': 16.30019760131836, 'timing/data': 0.0002197900030296296, 'timing/model': 0.39554785800282843, 'learning_rate/qwen_vl_interface': 9.9198120231166e-06, 'learning_rate/action_model': 9.9198120231166e-05, 'epoch': 0.03}) +[2026-10-03 18:59:48,078][starVLA.training.train_starvla][INFO] - Step 388, Loss: {'train/loss': 0.20657548308372498, 'train/grad_norm_pre_clip': 7.568419933319092, 'timing/data': 0.0002173100074287504, 'timing/model': 0.3960679580050055, 'learning_rate/qwen_vl_interface': 9.919253839715208e-06, 'learning_rate/action_model': 9.919253839715208e-05, 'epoch': 0.03}) +[2026-10-03 18:59:48,783][starVLA.training.train_starvla][INFO] - Step 389, Loss: {'train/loss': 0.2635772228240967, 'train/grad_norm_pre_clip': 52.39072036743164, 'timing/data': 0.00022350999643094838, 'timing/model': 0.39392931800102815, 'learning_rate/qwen_vl_interface': 9.918693736959935e-06, 'learning_rate/action_model': 9.918693736959934e-05, 'epoch': 0.03}) +[2026-10-03 18:59:49,490][starVLA.training.train_starvla][INFO] - Step 390, Loss: {'train/loss': 0.14690209925174713, 'train/grad_norm_pre_clip': 13.056112289428711, 'timing/data': 0.00032309998641721904, 'timing/model': 0.3935816279845312, 'learning_rate/qwen_vl_interface': 9.918131715081019e-06, 'learning_rate/action_model': 9.918131715081018e-05, 'epoch': 0.03}) +[2026-10-03 18:59:50,197][starVLA.training.train_starvla][INFO] - Step 391, Loss: {'train/loss': 0.35097917914390564, 'train/grad_norm_pre_clip': 63.71731948852539, 'timing/data': 0.0003570099943317473, 'timing/model': 0.39508841797942296, 'learning_rate/qwen_vl_interface': 9.917567774309486e-06, 'learning_rate/action_model': 9.917567774309485e-05, 'epoch': 0.03}) +[2026-10-03 18:59:51,095][starVLA.training.train_starvla][INFO] - Step 392, Loss: {'train/loss': 0.4856780469417572, 'train/grad_norm_pre_clip': 52.83573913574219, 'timing/data': 0.00035518998629413545, 'timing/model': 0.395238477998646, 'learning_rate/qwen_vl_interface': 9.91700191487715e-06, 'learning_rate/action_model': 9.91700191487715e-05, 'epoch': 0.03}) +[2026-10-03 18:59:51,802][starVLA.training.train_starvla][INFO] - Step 393, Loss: {'train/loss': 0.17656177282333374, 'train/grad_norm_pre_clip': 15.226541519165039, 'timing/data': 0.00025132999871857464, 'timing/model': 0.39635117701254785, 'learning_rate/qwen_vl_interface': 9.916434137016614e-06, 'learning_rate/action_model': 9.916434137016613e-05, 'epoch': 0.03}) +[2026-10-03 18:59:52,506][starVLA.training.train_starvla][INFO] - Step 394, Loss: {'train/loss': 1.2391245365142822, 'train/grad_norm_pre_clip': 78.65998840332031, 'timing/data': 0.0003258200013078749, 'timing/model': 0.3929050680017099, 'learning_rate/qwen_vl_interface': 9.915864440961272e-06, 'learning_rate/action_model': 9.915864440961272e-05, 'epoch': 0.03}) +[2026-10-03 18:59:53,213][starVLA.training.train_starvla][INFO] - Step 395, Loss: {'train/loss': 0.7237114906311035, 'train/grad_norm_pre_clip': 84.04010772705078, 'timing/data': 0.0003968800010625273, 'timing/model': 0.3951400480000302, 'learning_rate/qwen_vl_interface': 9.915292826945303e-06, 'learning_rate/action_model': 9.915292826945303e-05, 'epoch': 0.03}) +[2026-10-03 18:59:53,919][starVLA.training.train_starvla][INFO] - Step 396, Loss: {'train/loss': 0.7362351417541504, 'train/grad_norm_pre_clip': 56.41345977783203, 'timing/data': 0.0003103199996985495, 'timing/model': 0.39594887898419984, 'learning_rate/qwen_vl_interface': 9.914719295203676e-06, 'learning_rate/action_model': 9.914719295203676e-05, 'epoch': 0.03}) +[2026-10-03 18:59:54,822][starVLA.training.train_starvla][INFO] - Step 397, Loss: {'train/loss': 0.8225464820861816, 'train/grad_norm_pre_clip': 83.68163299560547, 'timing/data': 0.000304230023175478, 'timing/model': 0.5916203680098988, 'learning_rate/qwen_vl_interface': 9.91414384597215e-06, 'learning_rate/action_model': 9.91414384597215e-05, 'epoch': 0.03}) +[2026-10-03 18:59:55,527][starVLA.training.train_starvla][INFO] - Step 398, Loss: {'train/loss': 1.0195143222808838, 'train/grad_norm_pre_clip': 89.2286376953125, 'timing/data': 0.00024268002016469836, 'timing/model': 0.39425349902012385, 'learning_rate/qwen_vl_interface': 9.913566479487267e-06, 'learning_rate/action_model': 9.913566479487267e-05, 'epoch': 0.03}) +[2026-10-03 18:59:56,233][starVLA.training.train_starvla][INFO] - Step 399, Loss: {'train/loss': 0.7199349403381348, 'train/grad_norm_pre_clip': 17.63644027709961, 'timing/data': 0.00035022999509237707, 'timing/model': 0.39341045898618177, 'learning_rate/qwen_vl_interface': 9.912987195986366e-06, 'learning_rate/action_model': 9.912987195986365e-05, 'epoch': 0.03}) +[2026-10-03 18:59:56,941][starVLA.training.train_starvla][INFO] - Step 400, Loss: {'train/loss': 0.5456047058105469, 'train/grad_norm_pre_clip': 36.39229965209961, 'timing/data': 0.00027315999614074826, 'timing/model': 0.3949413680238649, 'learning_rate/qwen_vl_interface': 9.912405995707562e-06, 'learning_rate/action_model': 9.912405995707561e-05, 'epoch': 0.03}) +[2026-10-03 18:59:57,650][starVLA.training.train_starvla][INFO] - Step 401, Loss: {'train/loss': 0.5964651107788086, 'train/grad_norm_pre_clip': 44.02945327758789, 'timing/data': 0.00029793998692184687, 'timing/model': 0.39684637798927724, 'learning_rate/qwen_vl_interface': 9.911822878889767e-06, 'learning_rate/action_model': 9.911822878889767e-05, 'epoch': 0.03}) +[2026-10-03 18:59:58,546][starVLA.training.train_starvla][INFO] - Step 402, Loss: {'train/loss': 0.9053735136985779, 'train/grad_norm_pre_clip': 80.80094146728516, 'timing/data': 0.00039436001679860055, 'timing/model': 0.5846577379852533, 'learning_rate/qwen_vl_interface': 9.911237845772681e-06, 'learning_rate/action_model': 9.911237845772681e-05, 'epoch': 0.03}) +[2026-10-03 18:59:59,253][starVLA.training.train_starvla][INFO] - Step 403, Loss: {'train/loss': 0.3960351049900055, 'train/grad_norm_pre_clip': 46.83041000366211, 'timing/data': 0.00023689999943599105, 'timing/model': 0.39490230701630935, 'learning_rate/qwen_vl_interface': 9.910650896596787e-06, 'learning_rate/action_model': 9.910650896596786e-05, 'epoch': 0.03}) +[2026-10-03 18:59:59,960][starVLA.training.train_starvla][INFO] - Step 404, Loss: {'train/loss': 0.326138973236084, 'train/grad_norm_pre_clip': 18.11465835571289, 'timing/data': 0.0003014499961864203, 'timing/model': 0.3953314179962035, 'learning_rate/qwen_vl_interface': 9.910062031603356e-06, 'learning_rate/action_model': 9.910062031603356e-05, 'epoch': 0.03}) +[2026-10-03 19:00:00,666][starVLA.training.train_starvla][INFO] - Step 405, Loss: {'train/loss': 0.5428674817085266, 'train/grad_norm_pre_clip': 51.78390121459961, 'timing/data': 0.000285169982817024, 'timing/model': 0.39614683800027706, 'learning_rate/qwen_vl_interface': 9.909471251034451e-06, 'learning_rate/action_model': 9.90947125103445e-05, 'epoch': 0.03}) +[2026-10-03 19:00:01,374][starVLA.training.train_starvla][INFO] - Step 406, Loss: {'train/loss': 0.692512035369873, 'train/grad_norm_pre_clip': 93.32624816894531, 'timing/data': 0.00038949999725446105, 'timing/model': 0.39488731799065135, 'learning_rate/qwen_vl_interface': 9.908878555132918e-06, 'learning_rate/action_model': 9.908878555132918e-05, 'epoch': 0.03}) +[2026-10-03 19:00:02,274][starVLA.training.train_starvla][INFO] - Step 407, Loss: {'train/loss': 0.840097188949585, 'train/grad_norm_pre_clip': 39.51530456542969, 'timing/data': 0.0003419900021981448, 'timing/model': 0.590169426985085, 'learning_rate/qwen_vl_interface': 9.908283944142391e-06, 'learning_rate/action_model': 9.908283944142391e-05, 'epoch': 0.03}) +[2026-10-03 19:00:02,978][starVLA.training.train_starvla][INFO] - Step 408, Loss: {'train/loss': 0.3905506134033203, 'train/grad_norm_pre_clip': 19.230022430419922, 'timing/data': 0.00031879000016488135, 'timing/model': 0.3927743480016943, 'learning_rate/qwen_vl_interface': 9.907687418307297e-06, 'learning_rate/action_model': 9.907687418307296e-05, 'epoch': 0.03}) +[2026-10-03 19:00:03,684][starVLA.training.train_starvla][INFO] - Step 409, Loss: {'train/loss': 0.34139180183410645, 'train/grad_norm_pre_clip': 52.2861213684082, 'timing/data': 0.0002519400150049478, 'timing/model': 0.39446801799931563, 'learning_rate/qwen_vl_interface': 9.907088977872837e-06, 'learning_rate/action_model': 9.907088977872837e-05, 'epoch': 0.03}) +[2026-10-03 19:00:04,391][starVLA.training.train_starvla][INFO] - Step 410, Loss: {'train/loss': 0.4687257409095764, 'train/grad_norm_pre_clip': 63.44733810424805, 'timing/data': 0.00027483998565003276, 'timing/model': 0.39509649900719523, 'learning_rate/qwen_vl_interface': 9.906488623085015e-06, 'learning_rate/action_model': 9.906488623085015e-05, 'epoch': 0.03}) +[2026-10-03 19:00:05,099][starVLA.training.train_starvla][INFO] - Step 411, Loss: {'train/loss': 0.30632877349853516, 'train/grad_norm_pre_clip': 54.091800689697266, 'timing/data': 0.0003551300032995641, 'timing/model': 0.39364038800704293, 'learning_rate/qwen_vl_interface': 9.905886354190611e-06, 'learning_rate/action_model': 9.90588635419061e-05, 'epoch': 0.03}) +[2026-10-03 19:00:05,808][starVLA.training.train_starvla][INFO] - Step 412, Loss: {'train/loss': 0.7339503765106201, 'train/grad_norm_pre_clip': 41.072689056396484, 'timing/data': 0.0003664699906948954, 'timing/model': 0.39627698800177313, 'learning_rate/qwen_vl_interface': 9.905282171437192e-06, 'learning_rate/action_model': 9.905282171437192e-05, 'epoch': 0.03}) +[2026-10-03 19:00:06,724][starVLA.training.train_starvla][INFO] - Step 413, Loss: {'train/loss': 0.42180800437927246, 'train/grad_norm_pre_clip': 79.99396514892578, 'timing/data': 0.00028984001255594194, 'timing/model': 0.39676121799857356, 'learning_rate/qwen_vl_interface': 9.904676075073121e-06, 'learning_rate/action_model': 9.904676075073121e-05, 'epoch': 0.03}) +[2026-10-03 19:00:07,428][starVLA.training.train_starvla][INFO] - Step 414, Loss: {'train/loss': 0.35776734352111816, 'train/grad_norm_pre_clip': 63.46087646484375, 'timing/data': 0.00037505998625420034, 'timing/model': 0.39265982899814844, 'learning_rate/qwen_vl_interface': 9.904068065347536e-06, 'learning_rate/action_model': 9.904068065347536e-05, 'epoch': 0.03}) +[2026-10-03 19:00:08,134][starVLA.training.train_starvla][INFO] - Step 415, Loss: {'train/loss': 0.45031070709228516, 'train/grad_norm_pre_clip': 46.864166259765625, 'timing/data': 0.00031477000447921455, 'timing/model': 0.39449398798751645, 'learning_rate/qwen_vl_interface': 9.90345814251037e-06, 'learning_rate/action_model': 9.90345814251037e-05, 'epoch': 0.03}) +[2026-10-03 19:00:08,838][starVLA.training.train_starvla][INFO] - Step 416, Loss: {'train/loss': 0.624313235282898, 'train/grad_norm_pre_clip': 44.55089569091797, 'timing/data': 0.0003736400103662163, 'timing/model': 0.39408929797355086, 'learning_rate/qwen_vl_interface': 9.902846306812339e-06, 'learning_rate/action_model': 9.902846306812338e-05, 'epoch': 0.03}) +[2026-10-03 19:00:09,543][starVLA.training.train_starvla][INFO] - Step 417, Loss: {'train/loss': 0.6465579867362976, 'train/grad_norm_pre_clip': 48.13271713256836, 'timing/data': 0.0002760099887382239, 'timing/model': 0.3949724289996084, 'learning_rate/qwen_vl_interface': 9.902232558504942e-06, 'learning_rate/action_model': 9.902232558504941e-05, 'epoch': 0.03}) +[2026-10-03 19:00:10,448][starVLA.training.train_starvla][INFO] - Step 418, Loss: {'train/loss': 1.096618413925171, 'train/grad_norm_pre_clip': 215.39669799804688, 'timing/data': 0.000370740017388016, 'timing/model': 0.5928365670260973, 'learning_rate/qwen_vl_interface': 9.901616897840471e-06, 'learning_rate/action_model': 9.90161689784047e-05, 'epoch': 0.03}) +[2026-10-03 19:00:11,150][starVLA.training.train_starvla][INFO] - Step 419, Loss: {'train/loss': 0.5601437091827393, 'train/grad_norm_pre_clip': 204.7035369873047, 'timing/data': 0.0003169899864587933, 'timing/model': 0.3928243890113663, 'learning_rate/qwen_vl_interface': 9.900999325072e-06, 'learning_rate/action_model': 9.900999325072001e-05, 'epoch': 0.03}) +[2026-10-03 19:00:11,855][starVLA.training.train_starvla][INFO] - Step 420, Loss: {'train/loss': 0.3418329358100891, 'train/grad_norm_pre_clip': 39.90938186645508, 'timing/data': 0.00041412998689338565, 'timing/model': 0.39372724801069126, 'learning_rate/qwen_vl_interface': 9.900379840453392e-06, 'learning_rate/action_model': 9.900379840453392e-05, 'epoch': 0.03}) +[2026-10-03 19:00:12,559][starVLA.training.train_starvla][INFO] - Step 421, Loss: {'train/loss': 0.31277552247047424, 'train/grad_norm_pre_clip': 86.75202941894531, 'timing/data': 0.00022916001034900546, 'timing/model': 0.39492194799822755, 'learning_rate/qwen_vl_interface': 9.89975844423929e-06, 'learning_rate/action_model': 9.899758444239289e-05, 'epoch': 0.03}) +[2026-10-03 19:00:13,263][starVLA.training.train_starvla][INFO] - Step 422, Loss: {'train/loss': 0.22192256152629852, 'train/grad_norm_pre_clip': 26.12906265258789, 'timing/data': 0.00034775998210534453, 'timing/model': 0.39252502800081857, 'learning_rate/qwen_vl_interface': 9.899135136685131e-06, 'learning_rate/action_model': 9.89913513668513e-05, 'epoch': 0.03}) +[2026-10-03 19:00:13,966][starVLA.training.train_starvla][INFO] - Step 423, Loss: {'train/loss': 0.43745893239974976, 'train/grad_norm_pre_clip': 29.053525924682617, 'timing/data': 0.00034507998498156667, 'timing/model': 0.3938751789974049, 'learning_rate/qwen_vl_interface': 9.898509918047127e-06, 'learning_rate/action_model': 9.898509918047126e-05, 'epoch': 0.03}) +[2026-10-03 19:00:14,876][starVLA.training.train_starvla][INFO] - Step 424, Loss: {'train/loss': 0.4734896123409271, 'train/grad_norm_pre_clip': 49.628196716308594, 'timing/data': 0.0003102900227531791, 'timing/model': 0.39415922900661826, 'learning_rate/qwen_vl_interface': 9.897882788582286e-06, 'learning_rate/action_model': 9.897882788582287e-05, 'epoch': 0.03}) +[2026-10-03 19:00:15,580][starVLA.training.train_starvla][INFO] - Step 425, Loss: {'train/loss': 0.4801468253135681, 'train/grad_norm_pre_clip': 51.91875457763672, 'timing/data': 0.0003115000145044178, 'timing/model': 0.3949553180136718, 'learning_rate/qwen_vl_interface': 9.897253748548398e-06, 'learning_rate/action_model': 9.897253748548399e-05, 'epoch': 0.03}) +[2026-10-03 19:00:16,285][starVLA.training.train_starvla][INFO] - Step 426, Loss: {'train/loss': 0.10173840820789337, 'train/grad_norm_pre_clip': 21.38607406616211, 'timing/data': 0.0002754800079856068, 'timing/model': 0.3935007580148522, 'learning_rate/qwen_vl_interface': 9.896622798204035e-06, 'learning_rate/action_model': 9.896622798204035e-05, 'epoch': 0.03}) +[2026-10-03 19:00:16,987][starVLA.training.train_starvla][INFO] - Step 427, Loss: {'train/loss': 0.675227165222168, 'train/grad_norm_pre_clip': 59.790035247802734, 'timing/data': 0.00029286000062711537, 'timing/model': 0.3930526089970954, 'learning_rate/qwen_vl_interface': 9.89598993780856e-06, 'learning_rate/action_model': 9.89598993780856e-05, 'epoch': 0.03}) +[2026-10-03 19:00:17,690][starVLA.training.train_starvla][INFO] - Step 428, Loss: {'train/loss': 0.14950332045555115, 'train/grad_norm_pre_clip': 33.668514251708984, 'timing/data': 0.0002277999883517623, 'timing/model': 0.39242868797737174, 'learning_rate/qwen_vl_interface': 9.895355167622113e-06, 'learning_rate/action_model': 9.895355167622113e-05, 'epoch': 0.03}) +[2026-10-03 19:00:18,597][starVLA.training.train_starvla][INFO] - Step 429, Loss: {'train/loss': 0.05103955790400505, 'train/grad_norm_pre_clip': 6.776398181915283, 'timing/data': 0.00028328000917099416, 'timing/model': 0.3946159390034154, 'learning_rate/qwen_vl_interface': 9.894718487905629e-06, 'learning_rate/action_model': 9.894718487905629e-05, 'epoch': 0.03}) +[2026-10-03 19:00:19,302][starVLA.training.train_starvla][INFO] - Step 430, Loss: {'train/loss': 0.5973893404006958, 'train/grad_norm_pre_clip': 25.012197494506836, 'timing/data': 0.0004920400097034872, 'timing/model': 0.39373072900343686, 'learning_rate/qwen_vl_interface': 9.89407989892082e-06, 'learning_rate/action_model': 9.894079898920821e-05, 'epoch': 0.03}) +[2026-10-03 19:00:20,010][starVLA.training.train_starvla][INFO] - Step 431, Loss: {'train/loss': 0.39364445209503174, 'train/grad_norm_pre_clip': 61.19294738769531, 'timing/data': 0.0002425599959678948, 'timing/model': 0.39502051897579804, 'learning_rate/qwen_vl_interface': 9.893439400930187e-06, 'learning_rate/action_model': 9.893439400930187e-05, 'epoch': 0.03}) +[2026-10-03 19:00:20,713][starVLA.training.train_starvla][INFO] - Step 432, Loss: {'train/loss': 0.41238582134246826, 'train/grad_norm_pre_clip': 66.18846130371094, 'timing/data': 0.0002272099955007434, 'timing/model': 0.39426015902427025, 'learning_rate/qwen_vl_interface': 9.892796994197014e-06, 'learning_rate/action_model': 9.892796994197014e-05, 'epoch': 0.03}) +[2026-10-03 19:00:21,421][starVLA.training.train_starvla][INFO] - Step 433, Loss: {'train/loss': 0.5431097149848938, 'train/grad_norm_pre_clip': 49.55630111694336, 'timing/data': 0.00030673001310788095, 'timing/model': 0.3954681590257678, 'learning_rate/qwen_vl_interface': 9.892152678985372e-06, 'learning_rate/action_model': 9.892152678985372e-05, 'epoch': 0.03}) +[2026-10-03 19:00:22,331][starVLA.training.train_starvla][INFO] - Step 434, Loss: {'train/loss': 0.3159342408180237, 'train/grad_norm_pre_clip': 22.257061004638672, 'timing/data': 0.00022988999262452126, 'timing/model': 0.6005722370173316, 'learning_rate/qwen_vl_interface': 9.891506455560111e-06, 'learning_rate/action_model': 9.891506455560111e-05, 'epoch': 0.03}) +[2026-10-03 19:00:23,037][starVLA.training.train_starvla][INFO] - Step 435, Loss: {'train/loss': 0.33650773763656616, 'train/grad_norm_pre_clip': 27.001365661621094, 'timing/data': 0.0003688099968712777, 'timing/model': 0.3940742590057198, 'learning_rate/qwen_vl_interface': 9.890858324186872e-06, 'learning_rate/action_model': 9.890858324186872e-05, 'epoch': 0.03}) +[2026-10-03 19:00:23,743][starVLA.training.train_starvla][INFO] - Step 436, Loss: {'train/loss': 0.3464984893798828, 'train/grad_norm_pre_clip': 31.85722541809082, 'timing/data': 0.00042158999713137746, 'timing/model': 0.39456899897777475, 'learning_rate/qwen_vl_interface': 9.890208285132077e-06, 'learning_rate/action_model': 9.890208285132077e-05, 'epoch': 0.03}) +[2026-10-03 19:00:24,450][starVLA.training.train_starvla][INFO] - Step 437, Loss: {'train/loss': 0.47390317916870117, 'train/grad_norm_pre_clip': 34.94309616088867, 'timing/data': 0.0002829899895004928, 'timing/model': 0.39481673898990266, 'learning_rate/qwen_vl_interface': 9.889556338662933e-06, 'learning_rate/action_model': 9.889556338662932e-05, 'epoch': 0.03}) +[2026-10-03 19:00:25,156][starVLA.training.train_starvla][INFO] - Step 438, Loss: {'train/loss': 0.1763758808374405, 'train/grad_norm_pre_clip': 37.217227935791016, 'timing/data': 0.0003672199964057654, 'timing/model': 0.39375506798387505, 'learning_rate/qwen_vl_interface': 9.888902485047427e-06, 'learning_rate/action_model': 9.888902485047427e-05, 'epoch': 0.03}) +[2026-10-03 19:00:25,861][starVLA.training.train_starvla][INFO] - Step 439, Loss: {'train/loss': 0.28397226333618164, 'train/grad_norm_pre_clip': 89.8661880493164, 'timing/data': 0.0003937099827453494, 'timing/model': 0.39254307898227125, 'learning_rate/qwen_vl_interface': 9.888246724554337e-06, 'learning_rate/action_model': 9.888246724554337e-05, 'epoch': 0.03}) +[2026-10-03 19:00:26,768][starVLA.training.train_starvla][INFO] - Step 440, Loss: {'train/loss': 0.6250180006027222, 'train/grad_norm_pre_clip': 47.20951843261719, 'timing/data': 0.0004350300005171448, 'timing/model': 0.39484475800418295, 'learning_rate/qwen_vl_interface': 9.887589057453222e-06, 'learning_rate/action_model': 9.88758905745322e-05, 'epoch': 0.03}) +[2026-10-03 19:00:27,475][starVLA.training.train_starvla][INFO] - Step 441, Loss: {'train/loss': 0.4030047357082367, 'train/grad_norm_pre_clip': 63.04163360595703, 'timing/data': 0.0002670300018507987, 'timing/model': 0.3960149670019746, 'learning_rate/qwen_vl_interface': 9.88692948401442e-06, 'learning_rate/action_model': 9.88692948401442e-05, 'epoch': 0.03}) +[2026-10-03 19:00:28,184][starVLA.training.train_starvla][INFO] - Step 442, Loss: {'train/loss': 1.2642693519592285, 'train/grad_norm_pre_clip': 138.96688842773438, 'timing/data': 0.0003716499777510762, 'timing/model': 0.395761769003002, 'learning_rate/qwen_vl_interface': 9.88626800450906e-06, 'learning_rate/action_model': 9.88626800450906e-05, 'epoch': 0.03}) +[2026-10-03 19:00:28,890][starVLA.training.train_starvla][INFO] - Step 443, Loss: {'train/loss': 1.1693737506866455, 'train/grad_norm_pre_clip': 95.26238250732422, 'timing/data': 0.00033926000469364226, 'timing/model': 0.39516961900517344, 'learning_rate/qwen_vl_interface': 9.885604619209052e-06, 'learning_rate/action_model': 9.885604619209051e-05, 'epoch': 0.03}) +[2026-10-03 19:00:29,598][starVLA.training.train_starvla][INFO] - Step 444, Loss: {'train/loss': 0.4354572892189026, 'train/grad_norm_pre_clip': 22.973169326782227, 'timing/data': 0.00035308999940752983, 'timing/model': 0.39436869899509475, 'learning_rate/qwen_vl_interface': 9.884939328387085e-06, 'learning_rate/action_model': 9.884939328387084e-05, 'epoch': 0.03}) +[2026-10-03 19:00:30,512][starVLA.training.train_starvla][INFO] - Step 445, Loss: {'train/loss': 0.2730296850204468, 'train/grad_norm_pre_clip': 18.632516860961914, 'timing/data': 0.00043471998651511967, 'timing/model': 0.394626277993666, 'learning_rate/qwen_vl_interface': 9.884272132316638e-06, 'learning_rate/action_model': 9.884272132316637e-05, 'epoch': 0.03}) +[2026-10-03 19:00:31,222][starVLA.training.train_starvla][INFO] - Step 446, Loss: {'train/loss': 0.2716154456138611, 'train/grad_norm_pre_clip': 21.46071434020996, 'timing/data': 0.00023053001496009529, 'timing/model': 0.3963665190094616, 'learning_rate/qwen_vl_interface': 9.883603031271968e-06, 'learning_rate/action_model': 9.883603031271969e-05, 'epoch': 0.03}) +[2026-10-03 19:00:31,929][starVLA.training.train_starvla][INFO] - Step 447, Loss: {'train/loss': 0.285616010427475, 'train/grad_norm_pre_clip': 38.84617614746094, 'timing/data': 0.0002329199924133718, 'timing/model': 0.39640052799950354, 'learning_rate/qwen_vl_interface': 9.882932025528122e-06, 'learning_rate/action_model': 9.882932025528121e-05, 'epoch': 0.03}) +[2026-10-03 19:00:32,636][starVLA.training.train_starvla][INFO] - Step 448, Loss: {'train/loss': 0.46391868591308594, 'train/grad_norm_pre_clip': 27.761077880859375, 'timing/data': 0.0003080699825659394, 'timing/model': 0.3939106180041563, 'learning_rate/qwen_vl_interface': 9.882259115360918e-06, 'learning_rate/action_model': 9.882259115360917e-05, 'epoch': 0.03}) +[2026-10-03 19:00:33,343][starVLA.training.train_starvla][INFO] - Step 449, Loss: {'train/loss': 0.5732054114341736, 'train/grad_norm_pre_clip': 54.195011138916016, 'timing/data': 0.00031977001344785094, 'timing/model': 0.3941975189954974, 'learning_rate/qwen_vl_interface': 9.881584301046968e-06, 'learning_rate/action_model': 9.881584301046968e-05, 'epoch': 0.03}) +[2026-10-03 19:00:34,250][starVLA.training.train_starvla][INFO] - Step 450, Loss: {'train/loss': 0.5352831482887268, 'train/grad_norm_pre_clip': 34.26829147338867, 'timing/data': 0.00036571998498402536, 'timing/model': 0.5942244869947899, 'learning_rate/qwen_vl_interface': 9.880907582863662e-06, 'learning_rate/action_model': 9.880907582863662e-05, 'epoch': 0.03}) +[2026-10-03 19:00:34,954][starVLA.training.train_starvla][INFO] - Step 451, Loss: {'train/loss': 0.21091698110103607, 'train/grad_norm_pre_clip': 28.107593536376953, 'timing/data': 0.00036017000093124807, 'timing/model': 0.39407260800362565, 'learning_rate/qwen_vl_interface': 9.880228961089175e-06, 'learning_rate/action_model': 9.880228961089175e-05, 'epoch': 0.03}) +[2026-10-03 19:00:35,661][starVLA.training.train_starvla][INFO] - Step 452, Loss: {'train/loss': 0.283392995595932, 'train/grad_norm_pre_clip': 18.54016876220703, 'timing/data': 0.0002572299854364246, 'timing/model': 0.39362826899741776, 'learning_rate/qwen_vl_interface': 9.87954843600246e-06, 'learning_rate/action_model': 9.879548436002459e-05, 'epoch': 0.03}) +[2026-10-03 19:00:36,368][starVLA.training.train_starvla][INFO] - Step 453, Loss: {'train/loss': 0.3412086069583893, 'train/grad_norm_pre_clip': 53.61717224121094, 'timing/data': 0.00036183997872285545, 'timing/model': 0.3952368779864628, 'learning_rate/qwen_vl_interface': 9.878866007883256e-06, 'learning_rate/action_model': 9.878866007883257e-05, 'epoch': 0.03}) +[2026-10-03 19:00:37,075][starVLA.training.train_starvla][INFO] - Step 454, Loss: {'train/loss': 0.08928880095481873, 'train/grad_norm_pre_clip': 17.340238571166992, 'timing/data': 0.00023771999985910952, 'timing/model': 0.39440439897589386, 'learning_rate/qwen_vl_interface': 9.878181677012084e-06, 'learning_rate/action_model': 9.878181677012085e-05, 'epoch': 0.03}) +[2026-10-03 19:00:37,983][starVLA.training.train_starvla][INFO] - Step 455, Loss: {'train/loss': 0.1358603686094284, 'train/grad_norm_pre_clip': 28.395328521728516, 'timing/data': 0.0003500800230540335, 'timing/model': 0.5952821570099331, 'learning_rate/qwen_vl_interface': 9.877495443670247e-06, 'learning_rate/action_model': 9.877495443670246e-05, 'epoch': 0.03}) +[2026-10-03 19:00:38,690][starVLA.training.train_starvla][INFO] - Step 456, Loss: {'train/loss': 0.8836333155632019, 'train/grad_norm_pre_clip': 108.94420623779297, 'timing/data': 0.00023030000738799572, 'timing/model': 0.39405348800937645, 'learning_rate/qwen_vl_interface': 9.876807308139828e-06, 'learning_rate/action_model': 9.876807308139829e-05, 'epoch': 0.03}) +[2026-10-03 19:00:39,398][starVLA.training.train_starvla][INFO] - Step 457, Loss: {'train/loss': 0.12355099618434906, 'train/grad_norm_pre_clip': 9.367878913879395, 'timing/data': 0.0004794189881067723, 'timing/model': 0.39518335898173973, 'learning_rate/qwen_vl_interface': 9.876117270703695e-06, 'learning_rate/action_model': 9.876117270703694e-05, 'epoch': 0.03}) +[2026-10-03 19:00:40,104][starVLA.training.train_starvla][INFO] - Step 458, Loss: {'train/loss': 0.46685099601745605, 'train/grad_norm_pre_clip': 35.21644973754883, 'timing/data': 0.0003336000081617385, 'timing/model': 0.39354015901335515, 'learning_rate/qwen_vl_interface': 9.875425331645495e-06, 'learning_rate/action_model': 9.875425331645495e-05, 'epoch': 0.03}) +[2026-10-03 19:00:40,811][starVLA.training.train_starvla][INFO] - Step 459, Loss: {'train/loss': 0.5429222583770752, 'train/grad_norm_pre_clip': 65.62112426757812, 'timing/data': 0.00022596001508645713, 'timing/model': 0.3972953089978546, 'learning_rate/qwen_vl_interface': 9.874731491249658e-06, 'learning_rate/action_model': 9.874731491249657e-05, 'epoch': 0.03}) +[2026-10-03 19:00:41,726][starVLA.training.train_starvla][INFO] - Step 460, Loss: {'train/loss': 0.15624314546585083, 'train/grad_norm_pre_clip': 29.804128646850586, 'timing/data': 0.00022702000569552183, 'timing/model': 0.6025077670055907, 'learning_rate/qwen_vl_interface': 9.874035749801398e-06, 'learning_rate/action_model': 9.874035749801399e-05, 'epoch': 0.03}) +[2026-10-03 19:00:42,432][starVLA.training.train_starvla][INFO] - Step 461, Loss: {'train/loss': 0.4352591633796692, 'train/grad_norm_pre_clip': 55.70398712158203, 'timing/data': 0.0002582200104370713, 'timing/model': 0.3948680589965079, 'learning_rate/qwen_vl_interface': 9.873338107586704e-06, 'learning_rate/action_model': 9.873338107586705e-05, 'epoch': 0.03}) +[2026-10-03 19:00:43,137][starVLA.training.train_starvla][INFO] - Step 462, Loss: {'train/loss': 0.3025035560131073, 'train/grad_norm_pre_clip': 28.794692993164062, 'timing/data': 0.0005019599921070039, 'timing/model': 0.39343633901444264, 'learning_rate/qwen_vl_interface': 9.872638564892358e-06, 'learning_rate/action_model': 9.872638564892357e-05, 'epoch': 0.03}) +[2026-10-03 19:00:43,844][starVLA.training.train_starvla][INFO] - Step 463, Loss: {'train/loss': 0.3401273190975189, 'train/grad_norm_pre_clip': 35.89738464355469, 'timing/data': 0.00038819998735561967, 'timing/model': 0.39488220898783766, 'learning_rate/qwen_vl_interface': 9.871937122005905e-06, 'learning_rate/action_model': 9.871937122005905e-05, 'epoch': 0.03}) +[2026-10-03 19:00:44,552][starVLA.training.train_starvla][INFO] - Step 464, Loss: {'train/loss': 0.6264460682868958, 'train/grad_norm_pre_clip': 43.113956451416016, 'timing/data': 0.00036668998654931784, 'timing/model': 0.39471000799676403, 'learning_rate/qwen_vl_interface': 9.87123377921569e-06, 'learning_rate/action_model': 9.871233779215689e-05, 'epoch': 0.03}) +[2026-10-03 19:00:45,257][starVLA.training.train_starvla][INFO] - Step 465, Loss: {'train/loss': 0.4387853741645813, 'train/grad_norm_pre_clip': 89.2013931274414, 'timing/data': 0.00037558001349680126, 'timing/model': 0.39375541801564395, 'learning_rate/qwen_vl_interface': 9.870528536810825e-06, 'learning_rate/action_model': 9.870528536810825e-05, 'epoch': 0.03}) +[2026-10-03 19:00:46,165][starVLA.training.train_starvla][INFO] - Step 466, Loss: {'train/loss': 0.05493208020925522, 'train/grad_norm_pre_clip': 10.304878234863281, 'timing/data': 0.00038415001472458243, 'timing/model': 0.39583338898955844, 'learning_rate/qwen_vl_interface': 9.869821395081214e-06, 'learning_rate/action_model': 9.869821395081214e-05, 'epoch': 0.03}) +[2026-10-03 19:00:46,873][starVLA.training.train_starvla][INFO] - Step 467, Loss: {'train/loss': 1.155991554260254, 'train/grad_norm_pre_clip': 51.722686767578125, 'timing/data': 0.00021758998627774417, 'timing/model': 0.39642484800424427, 'learning_rate/qwen_vl_interface': 9.869112354317532e-06, 'learning_rate/action_model': 9.869112354317532e-05, 'epoch': 0.03}) +[2026-10-03 19:00:47,583][starVLA.training.train_starvla][INFO] - Step 468, Loss: {'train/loss': 0.16069066524505615, 'train/grad_norm_pre_clip': 28.0585994720459, 'timing/data': 0.00031022998155094683, 'timing/model': 0.39534777801600285, 'learning_rate/qwen_vl_interface': 9.86840141481124e-06, 'learning_rate/action_model': 9.868401414811239e-05, 'epoch': 0.03}) +[2026-10-03 19:00:48,291][starVLA.training.train_starvla][INFO] - Step 469, Loss: {'train/loss': 0.39346909523010254, 'train/grad_norm_pre_clip': 20.384674072265625, 'timing/data': 0.0002705899823922664, 'timing/model': 0.39598680901690386, 'learning_rate/qwen_vl_interface': 9.867688576854578e-06, 'learning_rate/action_model': 9.867688576854578e-05, 'epoch': 0.03}) +[2026-10-03 19:00:48,999][starVLA.training.train_starvla][INFO] - Step 470, Loss: {'train/loss': 0.5030578374862671, 'train/grad_norm_pre_clip': 40.89628219604492, 'timing/data': 0.0003825699968729168, 'timing/model': 0.3954871679889038, 'learning_rate/qwen_vl_interface': 9.86697384074057e-06, 'learning_rate/action_model': 9.866973840740568e-05, 'epoch': 0.03}) +[2026-10-03 19:00:49,908][starVLA.training.train_starvla][INFO] - Step 471, Loss: {'train/loss': 0.6621909141540527, 'train/grad_norm_pre_clip': 56.66944885253906, 'timing/data': 0.0003760700055863708, 'timing/model': 0.5991666259942576, 'learning_rate/qwen_vl_interface': 9.86625720676301e-06, 'learning_rate/action_model': 9.86625720676301e-05, 'epoch': 0.03}) +[2026-10-03 19:00:50,615][starVLA.training.train_starvla][INFO] - Step 472, Loss: {'train/loss': 0.5258353352546692, 'train/grad_norm_pre_clip': 86.33223724365234, 'timing/data': 0.0002296699967700988, 'timing/model': 0.39513824798632413, 'learning_rate/qwen_vl_interface': 9.865538675216488e-06, 'learning_rate/action_model': 9.865538675216486e-05, 'epoch': 0.03}) +[2026-10-03 19:00:51,320][starVLA.training.train_starvla][INFO] - Step 473, Loss: {'train/loss': 0.38459348678588867, 'train/grad_norm_pre_clip': 94.97761535644531, 'timing/data': 0.00024274998577311635, 'timing/model': 0.3941690379870124, 'learning_rate/qwen_vl_interface': 9.864818246396358e-06, 'learning_rate/action_model': 9.864818246396358e-05, 'epoch': 0.03}) +[2026-10-03 19:00:52,028][starVLA.training.train_starvla][INFO] - Step 474, Loss: {'train/loss': 0.5008943676948547, 'train/grad_norm_pre_clip': 87.07445526123047, 'timing/data': 0.00043462999747134745, 'timing/model': 0.3962172679894138, 'learning_rate/qwen_vl_interface': 9.864095920598764e-06, 'learning_rate/action_model': 9.864095920598763e-05, 'epoch': 0.03}) +[2026-10-03 19:00:52,733][starVLA.training.train_starvla][INFO] - Step 475, Loss: {'train/loss': 0.6334571838378906, 'train/grad_norm_pre_clip': 49.85182189941406, 'timing/data': 0.0003609300183597952, 'timing/model': 0.3932149789761752, 'learning_rate/qwen_vl_interface': 9.86337169812063e-06, 'learning_rate/action_model': 9.863371698120629e-05, 'epoch': 0.03}) +[2026-10-03 19:00:53,441][starVLA.training.train_starvla][INFO] - Step 476, Loss: {'train/loss': 0.40139880776405334, 'train/grad_norm_pre_clip': 30.659381866455078, 'timing/data': 0.00022921001072973013, 'timing/model': 0.39432698901509866, 'learning_rate/qwen_vl_interface': 9.862645579259648e-06, 'learning_rate/action_model': 9.862645579259648e-05, 'epoch': 0.03}) +[2026-10-03 19:00:54,345][starVLA.training.train_starvla][INFO] - Step 477, Loss: {'train/loss': 0.4783574938774109, 'train/grad_norm_pre_clip': 45.67913818359375, 'timing/data': 0.00046825999743305147, 'timing/model': 0.394159037998179, 'learning_rate/qwen_vl_interface': 9.861917564314308e-06, 'learning_rate/action_model': 9.861917564314308e-05, 'epoch': 0.03}) +[2026-10-03 19:00:55,053][starVLA.training.train_starvla][INFO] - Step 478, Loss: {'train/loss': 1.1050457954406738, 'train/grad_norm_pre_clip': 100.64972686767578, 'timing/data': 0.0004296789993532002, 'timing/model': 0.3952836190001108, 'learning_rate/qwen_vl_interface': 9.861187653583862e-06, 'learning_rate/action_model': 9.861187653583862e-05, 'epoch': 0.03}) +[2026-10-03 19:00:55,759][starVLA.training.train_starvla][INFO] - Step 479, Loss: {'train/loss': 0.12452597916126251, 'train/grad_norm_pre_clip': 20.408771514892578, 'timing/data': 0.00023504000273533165, 'timing/model': 0.3953024189977441, 'learning_rate/qwen_vl_interface': 9.860455847368352e-06, 'learning_rate/action_model': 9.860455847368351e-05, 'epoch': 0.04}) +[2026-10-03 19:00:56,468][starVLA.training.train_starvla][INFO] - Step 480, Loss: {'train/loss': 0.7595841288566589, 'train/grad_norm_pre_clip': 60.75078582763672, 'timing/data': 0.00023104000138118863, 'timing/model': 0.39477562802494504, 'learning_rate/qwen_vl_interface': 9.859722145968595e-06, 'learning_rate/action_model': 9.859722145968595e-05, 'epoch': 0.04}) +[2026-10-03 19:00:57,173][starVLA.training.train_starvla][INFO] - Step 481, Loss: {'train/loss': 0.12917974591255188, 'train/grad_norm_pre_clip': 24.818626403808594, 'timing/data': 0.00021100998856127262, 'timing/model': 0.3943819680134766, 'learning_rate/qwen_vl_interface': 9.858986549686188e-06, 'learning_rate/action_model': 9.858986549686188e-05, 'epoch': 0.04}) +[2026-10-03 19:00:58,087][starVLA.training.train_starvla][INFO] - Step 482, Loss: {'train/loss': 0.104404978454113, 'train/grad_norm_pre_clip': 37.481773376464844, 'timing/data': 0.00036544998874887824, 'timing/model': 0.3943850490031764, 'learning_rate/qwen_vl_interface': 9.85824905882351e-06, 'learning_rate/action_model': 9.858249058823509e-05, 'epoch': 0.04}) +[2026-10-03 19:00:58,792][starVLA.training.train_starvla][INFO] - Step 483, Loss: {'train/loss': 1.376927137374878, 'train/grad_norm_pre_clip': 60.545162200927734, 'timing/data': 0.0003381400019861758, 'timing/model': 0.39517914899624884, 'learning_rate/qwen_vl_interface': 9.857509673683709e-06, 'learning_rate/action_model': 9.857509673683708e-05, 'epoch': 0.04}) +[2026-10-03 19:00:59,500][starVLA.training.train_starvla][INFO] - Step 484, Loss: {'train/loss': 0.3310266137123108, 'train/grad_norm_pre_clip': 28.741443634033203, 'timing/data': 0.0003179699997417629, 'timing/model': 0.39472769800340757, 'learning_rate/qwen_vl_interface': 9.856768394570725e-06, 'learning_rate/action_model': 9.856768394570723e-05, 'epoch': 0.04}) +[2026-10-03 19:01:00,206][starVLA.training.train_starvla][INFO] - Step 485, Loss: {'train/loss': 0.3759783208370209, 'train/grad_norm_pre_clip': 40.79759216308594, 'timing/data': 0.00022094999440014362, 'timing/model': 0.3946900680020917, 'learning_rate/qwen_vl_interface': 9.856025221789264e-06, 'learning_rate/action_model': 9.856025221789264e-05, 'epoch': 0.04}) +[2026-10-03 19:01:00,912][starVLA.training.train_starvla][INFO] - Step 486, Loss: {'train/loss': 0.6220488548278809, 'train/grad_norm_pre_clip': 52.68788146972656, 'timing/data': 0.00039098001434467733, 'timing/model': 0.39500022801803425, 'learning_rate/qwen_vl_interface': 9.85528015564482e-06, 'learning_rate/action_model': 9.85528015564482e-05, 'epoch': 0.04}) +[2026-10-03 19:01:01,825][starVLA.training.train_starvla][INFO] - Step 487, Loss: {'train/loss': 0.2302451878786087, 'train/grad_norm_pre_clip': 65.2438735961914, 'timing/data': 0.00031355000101029873, 'timing/model': 0.6010935970116407, 'learning_rate/qwen_vl_interface': 9.854533196443661e-06, 'learning_rate/action_model': 9.854533196443661e-05, 'epoch': 0.04}) +[2026-10-03 19:01:02,534][starVLA.training.train_starvla][INFO] - Step 488, Loss: {'train/loss': 0.530673086643219, 'train/grad_norm_pre_clip': 37.232845306396484, 'timing/data': 0.00023030999000184238, 'timing/model': 0.3958872279908974, 'learning_rate/qwen_vl_interface': 9.853784344492833e-06, 'learning_rate/action_model': 9.853784344492833e-05, 'epoch': 0.04}) +[2026-10-03 19:01:03,239][starVLA.training.train_starvla][INFO] - Step 489, Loss: {'train/loss': 0.5208432674407959, 'train/grad_norm_pre_clip': 43.65281677246094, 'timing/data': 0.00035506999120116234, 'timing/model': 0.39409121801145375, 'learning_rate/qwen_vl_interface': 9.853033600100162e-06, 'learning_rate/action_model': 9.853033600100162e-05, 'epoch': 0.04}) +[2026-10-03 19:01:03,944][starVLA.training.train_starvla][INFO] - Step 490, Loss: {'train/loss': 0.5119911432266235, 'train/grad_norm_pre_clip': 37.1629524230957, 'timing/data': 0.00023697997676208615, 'timing/model': 0.3949032079835888, 'learning_rate/qwen_vl_interface': 9.85228096357425e-06, 'learning_rate/action_model': 9.852280963574249e-05, 'epoch': 0.04}) +[2026-10-03 19:01:04,650][starVLA.training.train_starvla][INFO] - Step 491, Loss: {'train/loss': 0.4290541410446167, 'train/grad_norm_pre_clip': 20.11214828491211, 'timing/data': 0.0003883889876306057, 'timing/model': 0.39525026897899806, 'learning_rate/qwen_vl_interface': 9.851526435224475e-06, 'learning_rate/action_model': 9.851526435224474e-05, 'epoch': 0.04}) +[2026-10-03 19:01:05,362][starVLA.training.train_starvla][INFO] - Step 492, Loss: {'train/loss': 0.36543846130371094, 'train/grad_norm_pre_clip': 43.31602096557617, 'timing/data': 0.00027944002067670226, 'timing/model': 0.39779163800994866, 'learning_rate/qwen_vl_interface': 9.850770015360999e-06, 'learning_rate/action_model': 9.850770015360999e-05, 'epoch': 0.04}) +[2026-10-03 19:01:06,275][starVLA.training.train_starvla][INFO] - Step 493, Loss: {'train/loss': 0.24219778180122375, 'train/grad_norm_pre_clip': 43.181175231933594, 'timing/data': 0.0003747199953068048, 'timing/model': 0.3942220289900433, 'learning_rate/qwen_vl_interface': 9.850011704294755e-06, 'learning_rate/action_model': 9.850011704294755e-05, 'epoch': 0.04}) +[2026-10-03 19:01:06,984][starVLA.training.train_starvla][INFO] - Step 494, Loss: {'train/loss': 0.3299171030521393, 'train/grad_norm_pre_clip': 40.7352180480957, 'timing/data': 0.00038824998773634434, 'timing/model': 0.3960270780080464, 'learning_rate/qwen_vl_interface': 9.849251502337458e-06, 'learning_rate/action_model': 9.849251502337458e-05, 'epoch': 0.04}) +[2026-10-03 19:01:07,691][starVLA.training.train_starvla][INFO] - Step 495, Loss: {'train/loss': 0.06796428561210632, 'train/grad_norm_pre_clip': 14.733440399169922, 'timing/data': 0.00035819000913761556, 'timing/model': 0.3946068979857955, 'learning_rate/qwen_vl_interface': 9.848489409801598e-06, 'learning_rate/action_model': 9.848489409801598e-05, 'epoch': 0.04}) +[2026-10-03 19:01:08,400][starVLA.training.train_starvla][INFO] - Step 496, Loss: {'train/loss': 0.03381849452853203, 'train/grad_norm_pre_clip': 4.049185752868652, 'timing/data': 0.00026604998856782913, 'timing/model': 0.39574387899483554, 'learning_rate/qwen_vl_interface': 9.84772542700044e-06, 'learning_rate/action_model': 9.84772542700044e-05, 'epoch': 0.04}) +[2026-10-03 19:01:09,107][starVLA.training.train_starvla][INFO] - Step 497, Loss: {'train/loss': 0.4434564411640167, 'train/grad_norm_pre_clip': 146.61643981933594, 'timing/data': 0.0003365099837537855, 'timing/model': 0.3964876079990063, 'learning_rate/qwen_vl_interface': 9.846959554248032e-06, 'learning_rate/action_model': 9.84695955424803e-05, 'epoch': 0.04}) +[2026-10-03 19:01:10,021][starVLA.training.train_starvla][INFO] - Step 498, Loss: {'train/loss': 0.7605343461036682, 'train/grad_norm_pre_clip': 60.28109359741211, 'timing/data': 0.000404930004151538, 'timing/model': 0.39478741798666306, 'learning_rate/qwen_vl_interface': 9.846191791859192e-06, 'learning_rate/action_model': 9.846191791859192e-05, 'epoch': 0.04}) +[2026-10-03 19:01:10,726][starVLA.training.train_starvla][INFO] - Step 499, Loss: {'train/loss': 0.1295691877603531, 'train/grad_norm_pre_clip': 52.87202072143555, 'timing/data': 0.00031291000777855515, 'timing/model': 0.39391887799138203, 'learning_rate/qwen_vl_interface': 9.845422140149518e-06, 'learning_rate/action_model': 9.845422140149519e-05, 'epoch': 0.04}) +[2026-10-03 19:01:11,435][starVLA.training.train_starvla][INFO] - Step 500, Loss: {'train/loss': 0.5400652885437012, 'train/grad_norm_pre_clip': 69.83732604980469, 'timing/data': 0.0002758899936452508, 'timing/model': 0.3953733679954894, 'learning_rate/qwen_vl_interface': 9.84465059943539e-06, 'learning_rate/action_model': 9.84465059943539e-05, 'epoch': 0.04}) +[2026-10-03 19:01:12,142][starVLA.training.train_starvla][INFO] - Step 501, Loss: {'train/loss': 0.4321323037147522, 'train/grad_norm_pre_clip': 56.96788787841797, 'timing/data': 0.00027037999825552106, 'timing/model': 0.39588409900898114, 'learning_rate/qwen_vl_interface': 9.843877170033956e-06, 'learning_rate/action_model': 9.843877170033954e-05, 'epoch': 0.04}) +[2026-10-03 19:01:12,849][starVLA.training.train_starvla][INFO] - Step 502, Loss: {'train/loss': 0.9236301183700562, 'train/grad_norm_pre_clip': 47.68218994140625, 'timing/data': 0.00022375001572072506, 'timing/model': 0.3956113979802467, 'learning_rate/qwen_vl_interface': 9.84310185226314e-06, 'learning_rate/action_model': 9.84310185226314e-05, 'epoch': 0.04}) +[2026-10-03 19:01:13,760][starVLA.training.train_starvla][INFO] - Step 503, Loss: {'train/loss': 0.04215056821703911, 'train/grad_norm_pre_clip': 5.5121355056762695, 'timing/data': 0.00035748997470363975, 'timing/model': 0.5980953160033096, 'learning_rate/qwen_vl_interface': 9.842324646441652e-06, 'learning_rate/action_model': 9.842324646441652e-05, 'epoch': 0.04}) +[2026-10-03 19:01:14,467][starVLA.training.train_starvla][INFO] - Step 504, Loss: {'train/loss': 0.39848506450653076, 'train/grad_norm_pre_clip': 60.6224250793457, 'timing/data': 0.0002275500155519694, 'timing/model': 0.39504391798982397, 'learning_rate/qwen_vl_interface': 9.841545552888968e-06, 'learning_rate/action_model': 9.841545552888967e-05, 'epoch': 0.04}) +[2026-10-03 19:01:15,175][starVLA.training.train_starvla][INFO] - Step 505, Loss: {'train/loss': 0.5569813251495361, 'train/grad_norm_pre_clip': 91.16685485839844, 'timing/data': 0.0003273799957241863, 'timing/model': 0.3947071289876476, 'learning_rate/qwen_vl_interface': 9.840764571925346e-06, 'learning_rate/action_model': 9.840764571925345e-05, 'epoch': 0.04}) +[2026-10-03 19:01:15,885][starVLA.training.train_starvla][INFO] - Step 506, Loss: {'train/loss': 0.9995667934417725, 'train/grad_norm_pre_clip': 129.18675231933594, 'timing/data': 0.00039940999704413116, 'timing/model': 0.3963046580029186, 'learning_rate/qwen_vl_interface': 9.839981703871815e-06, 'learning_rate/action_model': 9.839981703871814e-05, 'epoch': 0.04}) +[2026-10-03 19:01:16,591][starVLA.training.train_starvla][INFO] - Step 507, Loss: {'train/loss': 0.7022499442100525, 'train/grad_norm_pre_clip': 132.8502197265625, 'timing/data': 0.0002468599996063858, 'timing/model': 0.39443826902424917, 'learning_rate/qwen_vl_interface': 9.839196949050185e-06, 'learning_rate/action_model': 9.839196949050185e-05, 'epoch': 0.04}) +[2026-10-03 19:01:17,497][starVLA.training.train_starvla][INFO] - Step 508, Loss: {'train/loss': 0.1365778148174286, 'train/grad_norm_pre_clip': 67.6093978881836, 'timing/data': 0.0003599699994083494, 'timing/model': 0.5914451779972296, 'learning_rate/qwen_vl_interface': 9.838410307783038e-06, 'learning_rate/action_model': 9.838410307783038e-05, 'epoch': 0.04}) +[2026-10-03 19:01:18,206][starVLA.training.train_starvla][INFO] - Step 509, Loss: {'train/loss': 0.7043026685714722, 'train/grad_norm_pre_clip': 155.18637084960938, 'timing/data': 0.0002759300114121288, 'timing/model': 0.3961499579891097, 'learning_rate/qwen_vl_interface': 9.837621780393734e-06, 'learning_rate/action_model': 9.837621780393733e-05, 'epoch': 0.04}) +[2026-10-03 19:01:18,917][starVLA.training.train_starvla][INFO] - Step 510, Loss: {'train/loss': 0.6473038196563721, 'train/grad_norm_pre_clip': 31.01538848876953, 'timing/data': 0.0003815589880105108, 'timing/model': 0.39717510898481123, 'learning_rate/qwen_vl_interface': 9.836831367206406e-06, 'learning_rate/action_model': 9.836831367206406e-05, 'epoch': 0.04}) +[2026-10-03 19:01:19,627][starVLA.training.train_starvla][INFO] - Step 511, Loss: {'train/loss': 0.4348006844520569, 'train/grad_norm_pre_clip': 14.871003150939941, 'timing/data': 0.00036596000427380204, 'timing/model': 0.39514499800861813, 'learning_rate/qwen_vl_interface': 9.836039068545962e-06, 'learning_rate/action_model': 9.836039068545961e-05, 'epoch': 0.04}) +[2026-10-03 19:01:20,338][starVLA.training.train_starvla][INFO] - Step 512, Loss: {'train/loss': 0.47750207781791687, 'train/grad_norm_pre_clip': 43.288719177246094, 'timing/data': 0.0004208000027574599, 'timing/model': 0.39567087800242007, 'learning_rate/qwen_vl_interface': 9.835244884738086e-06, 'learning_rate/action_model': 9.835244884738085e-05, 'epoch': 0.04}) +[2026-10-03 19:01:21,046][starVLA.training.train_starvla][INFO] - Step 513, Loss: {'train/loss': 0.7049373388290405, 'train/grad_norm_pre_clip': 51.031105041503906, 'timing/data': 0.0003830899950116873, 'timing/model': 0.3961982189794071, 'learning_rate/qwen_vl_interface': 9.83444881610924e-06, 'learning_rate/action_model': 9.834448816109238e-05, 'epoch': 0.04}) +[2026-10-03 19:01:21,959][starVLA.training.train_starvla][INFO] - Step 514, Loss: {'train/loss': 0.5256216526031494, 'train/grad_norm_pre_clip': 32.192054748535156, 'timing/data': 0.0003562099882401526, 'timing/model': 0.39626147801754996, 'learning_rate/qwen_vl_interface': 9.833650862986652e-06, 'learning_rate/action_model': 9.833650862986653e-05, 'epoch': 0.04}) +[2026-10-03 19:01:22,668][starVLA.training.train_starvla][INFO] - Step 515, Loss: {'train/loss': 0.7732625007629395, 'train/grad_norm_pre_clip': 61.03715896606445, 'timing/data': 0.00035893998574465513, 'timing/model': 0.3952519799931906, 'learning_rate/qwen_vl_interface': 9.832851025698335e-06, 'learning_rate/action_model': 9.832851025698335e-05, 'epoch': 0.04}) +[2026-10-03 19:01:23,380][starVLA.training.train_starvla][INFO] - Step 516, Loss: {'train/loss': 0.08955138176679611, 'train/grad_norm_pre_clip': 20.323394775390625, 'timing/data': 0.0004417200107127428, 'timing/model': 0.397278288990492, 'learning_rate/qwen_vl_interface': 9.832049304573072e-06, 'learning_rate/action_model': 9.832049304573072e-05, 'epoch': 0.04}) +[2026-10-03 19:01:24,090][starVLA.training.train_starvla][INFO] - Step 517, Loss: {'train/loss': 0.47502434253692627, 'train/grad_norm_pre_clip': 91.25798034667969, 'timing/data': 0.000289060000795871, 'timing/model': 0.3978631880017929, 'learning_rate/qwen_vl_interface': 9.831245699940419e-06, 'learning_rate/action_model': 9.831245699940418e-05, 'epoch': 0.04}) +[2026-10-03 19:01:24,801][starVLA.training.train_starvla][INFO] - Step 518, Loss: {'train/loss': 0.17807912826538086, 'train/grad_norm_pre_clip': 34.48475646972656, 'timing/data': 0.0002826999989338219, 'timing/model': 0.3978580579860136, 'learning_rate/qwen_vl_interface': 9.830440212130707e-06, 'learning_rate/action_model': 9.830440212130708e-05, 'epoch': 0.04}) +[2026-10-03 19:01:25,709][starVLA.training.train_starvla][INFO] - Step 519, Loss: {'train/loss': 0.6933399438858032, 'train/grad_norm_pre_clip': 55.60533905029297, 'timing/data': 0.00035233001108281314, 'timing/model': 0.3968225689895917, 'learning_rate/qwen_vl_interface': 9.829632841475041e-06, 'learning_rate/action_model': 9.829632841475041e-05, 'epoch': 0.04}) +[2026-10-03 19:01:26,421][starVLA.training.train_starvla][INFO] - Step 520, Loss: {'train/loss': 0.44526004791259766, 'train/grad_norm_pre_clip': 61.17167663574219, 'timing/data': 0.00022486000671051443, 'timing/model': 0.39750221799477004, 'learning_rate/qwen_vl_interface': 9.828823588305305e-06, 'learning_rate/action_model': 9.828823588305303e-05, 'epoch': 0.04}) +[2026-10-03 19:01:27,133][starVLA.training.train_starvla][INFO] - Step 521, Loss: {'train/loss': 0.6733171939849854, 'train/grad_norm_pre_clip': 84.72173309326172, 'timing/data': 0.0003926999925170094, 'timing/model': 0.3973321590165142, 'learning_rate/qwen_vl_interface': 9.828012452954146e-06, 'learning_rate/action_model': 9.828012452954145e-05, 'epoch': 0.04}) +[2026-10-03 19:01:27,839][starVLA.training.train_starvla][INFO] - Step 522, Loss: {'train/loss': 0.50798499584198, 'train/grad_norm_pre_clip': 76.4406967163086, 'timing/data': 0.00020723999477922916, 'timing/model': 0.39549108798382804, 'learning_rate/qwen_vl_interface': 9.827199435754995e-06, 'learning_rate/action_model': 9.827199435754995e-05, 'epoch': 0.04}) +[2026-10-03 19:01:28,542][starVLA.training.train_starvla][INFO] - Step 523, Loss: {'train/loss': 0.4613696336746216, 'train/grad_norm_pre_clip': 18.246957778930664, 'timing/data': 0.00022812001407146454, 'timing/model': 0.3935275390103925, 'learning_rate/qwen_vl_interface': 9.826384537042052e-06, 'learning_rate/action_model': 9.826384537042052e-05, 'epoch': 0.04}) +[2026-10-03 19:01:29,452][starVLA.training.train_starvla][INFO] - Step 524, Loss: {'train/loss': 0.6366938352584839, 'train/grad_norm_pre_clip': 36.07719802856445, 'timing/data': 0.00021471999934874475, 'timing/model': 0.5979127679893281, 'learning_rate/qwen_vl_interface': 9.825567757150291e-06, 'learning_rate/action_model': 9.82556775715029e-05, 'epoch': 0.04}) +[2026-10-03 19:01:30,162][starVLA.training.train_starvla][INFO] - Step 525, Loss: {'train/loss': 0.28351008892059326, 'train/grad_norm_pre_clip': 40.199806213378906, 'timing/data': 0.0003341900010127574, 'timing/model': 0.3958402090065647, 'learning_rate/qwen_vl_interface': 9.82474909641546e-06, 'learning_rate/action_model': 9.82474909641546e-05, 'epoch': 0.04}) +[2026-10-03 19:01:30,868][starVLA.training.train_starvla][INFO] - Step 526, Loss: {'train/loss': 1.0715677738189697, 'train/grad_norm_pre_clip': 73.54393005371094, 'timing/data': 0.0002726299862843007, 'timing/model': 0.3948257279989775, 'learning_rate/qwen_vl_interface': 9.823928555174078e-06, 'learning_rate/action_model': 9.823928555174077e-05, 'epoch': 0.04}) +[2026-10-03 19:01:31,575][starVLA.training.train_starvla][INFO] - Step 527, Loss: {'train/loss': 0.32072898745536804, 'train/grad_norm_pre_clip': 12.723005294799805, 'timing/data': 0.00022382999304682016, 'timing/model': 0.3949218979978468, 'learning_rate/qwen_vl_interface': 9.82310613376344e-06, 'learning_rate/action_model': 9.82310613376344e-05, 'epoch': 0.04}) +[2026-10-03 19:01:32,282][starVLA.training.train_starvla][INFO] - Step 528, Loss: {'train/loss': 0.34412920475006104, 'train/grad_norm_pre_clip': 66.47396087646484, 'timing/data': 0.00022157002240419388, 'timing/model': 0.3949183790246025, 'learning_rate/qwen_vl_interface': 9.822281832521613e-06, 'learning_rate/action_model': 9.822281832521613e-05, 'epoch': 0.04}) +[2026-10-03 19:01:33,189][starVLA.training.train_starvla][INFO] - Step 529, Loss: {'train/loss': 0.5184363722801208, 'train/grad_norm_pre_clip': 51.33700942993164, 'timing/data': 0.00028183002723380923, 'timing/model': 0.5952204880013596, 'learning_rate/qwen_vl_interface': 9.821455651787432e-06, 'learning_rate/action_model': 9.821455651787432e-05, 'epoch': 0.04}) +[2026-10-03 19:01:33,893][starVLA.training.train_starvla][INFO] - Step 530, Loss: {'train/loss': 0.26709121465682983, 'train/grad_norm_pre_clip': 53.50536346435547, 'timing/data': 0.00021994998678565025, 'timing/model': 0.3946141689957585, 'learning_rate/qwen_vl_interface': 9.820627591900514e-06, 'learning_rate/action_model': 9.820627591900513e-05, 'epoch': 0.04}) +[2026-10-03 19:01:34,598][starVLA.training.train_starvla][INFO] - Step 531, Loss: {'train/loss': 0.6074882745742798, 'train/grad_norm_pre_clip': 82.16798400878906, 'timing/data': 0.00022476998856291175, 'timing/model': 0.39507334897643887, 'learning_rate/qwen_vl_interface': 9.819797653201238e-06, 'learning_rate/action_model': 9.819797653201238e-05, 'epoch': 0.04}) +[2026-10-03 19:01:35,306][starVLA.training.train_starvla][INFO] - Step 532, Loss: {'train/loss': 0.547831118106842, 'train/grad_norm_pre_clip': 48.47599792480469, 'timing/data': 0.00025635998463258147, 'timing/model': 0.3942158779827878, 'learning_rate/qwen_vl_interface': 9.818965836030765e-06, 'learning_rate/action_model': 9.818965836030764e-05, 'epoch': 0.04}) +[2026-10-03 19:01:36,011][starVLA.training.train_starvla][INFO] - Step 533, Loss: {'train/loss': 0.32125353813171387, 'train/grad_norm_pre_clip': 201.50645446777344, 'timing/data': 0.00036315000033937395, 'timing/model': 0.39595714799361303, 'learning_rate/qwen_vl_interface': 9.81813214073102e-06, 'learning_rate/action_model': 9.81813214073102e-05, 'epoch': 0.04}) +[2026-10-03 19:01:36,716][starVLA.training.train_starvla][INFO] - Step 534, Loss: {'train/loss': 0.5770888328552246, 'train/grad_norm_pre_clip': 37.49291229248047, 'timing/data': 0.0004146699793636799, 'timing/model': 0.3943907780048903, 'learning_rate/qwen_vl_interface': 9.817296567644709e-06, 'learning_rate/action_model': 9.817296567644709e-05, 'epoch': 0.04}) +[2026-10-03 19:01:37,620][starVLA.training.train_starvla][INFO] - Step 535, Loss: {'train/loss': 0.5600716471672058, 'train/grad_norm_pre_clip': 29.06386375427246, 'timing/data': 0.00022302998695522547, 'timing/model': 0.39323205800610594, 'learning_rate/qwen_vl_interface': 9.816459117115296e-06, 'learning_rate/action_model': 9.816459117115297e-05, 'epoch': 0.04}) +[2026-10-03 19:01:38,326][starVLA.training.train_starvla][INFO] - Step 536, Loss: {'train/loss': 0.25263315439224243, 'train/grad_norm_pre_clip': 51.27813720703125, 'timing/data': 0.00036643000203184783, 'timing/model': 0.39288082797429524, 'learning_rate/qwen_vl_interface': 9.815619789487035e-06, 'learning_rate/action_model': 9.815619789487034e-05, 'epoch': 0.04}) +[2026-10-03 19:01:39,030][starVLA.training.train_starvla][INFO] - Step 537, Loss: {'train/loss': 0.4982997477054596, 'train/grad_norm_pre_clip': 41.84490966796875, 'timing/data': 0.00025506000383757055, 'timing/model': 0.3947217090171762, 'learning_rate/qwen_vl_interface': 9.814778585104933e-06, 'learning_rate/action_model': 9.814778585104933e-05, 'epoch': 0.04}) +[2026-10-03 19:01:39,737][starVLA.training.train_starvla][INFO] - Step 538, Loss: {'train/loss': 0.40371161699295044, 'train/grad_norm_pre_clip': 57.03574752807617, 'timing/data': 0.0002646000066306442, 'timing/model': 0.3941538280050736, 'learning_rate/qwen_vl_interface': 9.813935504314784e-06, 'learning_rate/action_model': 9.813935504314784e-05, 'epoch': 0.04}) +[2026-10-03 19:01:40,442][starVLA.training.train_starvla][INFO] - Step 539, Loss: {'train/loss': 0.05300373211503029, 'train/grad_norm_pre_clip': 18.7747859954834, 'timing/data': 0.00022832999820820987, 'timing/model': 0.39558269901317544, 'learning_rate/qwen_vl_interface': 9.813090547463144e-06, 'learning_rate/action_model': 9.813090547463143e-05, 'epoch': 0.04}) +[2026-10-03 19:01:41,347][starVLA.training.train_starvla][INFO] - Step 540, Loss: {'train/loss': 0.034758180379867554, 'train/grad_norm_pre_clip': 6.0837907791137695, 'timing/data': 0.0002305100206285715, 'timing/model': 0.3937403379823081, 'learning_rate/qwen_vl_interface': 9.812243714897342e-06, 'learning_rate/action_model': 9.812243714897343e-05, 'epoch': 0.04}) +[2026-10-03 19:01:42,053][starVLA.training.train_starvla][INFO] - Step 541, Loss: {'train/loss': 0.18047314882278442, 'train/grad_norm_pre_clip': 51.878318786621094, 'timing/data': 0.0002714300062507391, 'timing/model': 0.3943745089927688, 'learning_rate/qwen_vl_interface': 9.811395006965481e-06, 'learning_rate/action_model': 9.81139500696548e-05, 'epoch': 0.04}) +[2026-10-03 19:01:42,762][starVLA.training.train_starvla][INFO] - Step 542, Loss: {'train/loss': 0.47463977336883545, 'train/grad_norm_pre_clip': 71.37875366210938, 'timing/data': 0.0002430400054436177, 'timing/model': 0.396223738003755, 'learning_rate/qwen_vl_interface': 9.81054442401643e-06, 'learning_rate/action_model': 9.81054442401643e-05, 'epoch': 0.04}) +[2026-10-03 19:01:43,470][starVLA.training.train_starvla][INFO] - Step 543, Loss: {'train/loss': 0.4252459406852722, 'train/grad_norm_pre_clip': 96.2896499633789, 'timing/data': 0.0002065699954982847, 'timing/model': 0.39603570802137256, 'learning_rate/qwen_vl_interface': 9.809691966399834e-06, 'learning_rate/action_model': 9.809691966399835e-05, 'epoch': 0.04}) +[2026-10-03 19:01:44,178][starVLA.training.train_starvla][INFO] - Step 544, Loss: {'train/loss': 0.3480985164642334, 'train/grad_norm_pre_clip': 12.289254188537598, 'timing/data': 0.00022705999435856938, 'timing/model': 0.3960845390101895, 'learning_rate/qwen_vl_interface': 9.808837634466104e-06, 'learning_rate/action_model': 9.808837634466104e-05, 'epoch': 0.04}) +[2026-10-03 19:01:45,085][starVLA.training.train_starvla][INFO] - Step 545, Loss: {'train/loss': 0.26806676387786865, 'train/grad_norm_pre_clip': 16.370891571044922, 'timing/data': 0.0002505000156816095, 'timing/model': 0.5938802569871768, 'learning_rate/qwen_vl_interface': 9.807981428566428e-06, 'learning_rate/action_model': 9.807981428566427e-05, 'epoch': 0.04}) +[2026-10-03 19:01:45,794][starVLA.training.train_starvla][INFO] - Step 546, Loss: {'train/loss': 0.1572224348783493, 'train/grad_norm_pre_clip': 36.69334030151367, 'timing/data': 0.0002973400114569813, 'timing/model': 0.39675534900743514, 'learning_rate/qwen_vl_interface': 9.807123349052754e-06, 'learning_rate/action_model': 9.807123349052753e-05, 'epoch': 0.04}) +[2026-10-03 19:01:46,502][starVLA.training.train_starvla][INFO] - Step 547, Loss: {'train/loss': 1.4735801219940186, 'train/grad_norm_pre_clip': 53.85732650756836, 'timing/data': 0.00025613000616431236, 'timing/model': 0.3965905570075847, 'learning_rate/qwen_vl_interface': 9.80626339627781e-06, 'learning_rate/action_model': 9.806263396277809e-05, 'epoch': 0.04}) +[2026-10-03 19:01:47,209][starVLA.training.train_starvla][INFO] - Step 548, Loss: {'train/loss': 0.4615311026573181, 'train/grad_norm_pre_clip': 38.493011474609375, 'timing/data': 0.00021795000066049397, 'timing/model': 0.3971585280087311, 'learning_rate/qwen_vl_interface': 9.805401570595087e-06, 'learning_rate/action_model': 9.805401570595087e-05, 'epoch': 0.04}) +[2026-10-03 19:01:47,917][starVLA.training.train_starvla][INFO] - Step 549, Loss: {'train/loss': 0.09321622550487518, 'train/grad_norm_pre_clip': 18.621109008789062, 'timing/data': 0.0002553399826865643, 'timing/model': 0.3956969089922495, 'learning_rate/qwen_vl_interface': 9.804537872358853e-06, 'learning_rate/action_model': 9.804537872358853e-05, 'epoch': 0.04}) +[2026-10-03 19:01:48,829][starVLA.training.train_starvla][INFO] - Step 550, Loss: {'train/loss': 0.7959225177764893, 'train/grad_norm_pre_clip': 43.09586715698242, 'timing/data': 0.0003619499912019819, 'timing/model': 0.5971282980171964, 'learning_rate/qwen_vl_interface': 9.80367230192414e-06, 'learning_rate/action_model': 9.803672301924138e-05, 'epoch': 0.04}) +[2026-10-03 19:01:49,538][starVLA.training.train_starvla][INFO] - Step 551, Loss: {'train/loss': 1.0149056911468506, 'train/grad_norm_pre_clip': 101.08885955810547, 'timing/data': 0.0003478199942037463, 'timing/model': 0.39728585799457505, 'learning_rate/qwen_vl_interface': 9.80280485964675e-06, 'learning_rate/action_model': 9.80280485964675e-05, 'epoch': 0.04}) +[2026-10-03 19:01:50,249][starVLA.training.train_starvla][INFO] - Step 552, Loss: {'train/loss': 1.2312304973602295, 'train/grad_norm_pre_clip': 92.53710174560547, 'timing/data': 0.00036385998828336596, 'timing/model': 0.3952278280048631, 'learning_rate/qwen_vl_interface': 9.801935545883255e-06, 'learning_rate/action_model': 9.801935545883255e-05, 'epoch': 0.04}) +[2026-10-03 19:01:50,959][starVLA.training.train_starvla][INFO] - Step 553, Loss: {'train/loss': 0.3930165469646454, 'train/grad_norm_pre_clip': 41.887962341308594, 'timing/data': 0.0002241400070488453, 'timing/model': 0.39712584801600315, 'learning_rate/qwen_vl_interface': 9.801064360991e-06, 'learning_rate/action_model': 9.801064360991e-05, 'epoch': 0.04}) +[2026-10-03 19:01:51,665][starVLA.training.train_starvla][INFO] - Step 554, Loss: {'train/loss': 0.2335488200187683, 'train/grad_norm_pre_clip': 34.96184539794922, 'timing/data': 0.000258829997619614, 'timing/model': 0.3958187980169896, 'learning_rate/qwen_vl_interface': 9.800191305328095e-06, 'learning_rate/action_model': 9.800191305328094e-05, 'epoch': 0.04}) +[2026-10-03 19:01:52,373][starVLA.training.train_starvla][INFO] - Step 555, Loss: {'train/loss': 0.39793771505355835, 'train/grad_norm_pre_clip': 22.15890884399414, 'timing/data': 0.00035676901461556554, 'timing/model': 0.3952892789966427, 'learning_rate/qwen_vl_interface': 9.79931637925342e-06, 'learning_rate/action_model': 9.79931637925342e-05, 'epoch': 0.04}) +[2026-10-03 19:01:53,283][starVLA.training.train_starvla][INFO] - Step 556, Loss: {'train/loss': 0.305472731590271, 'train/grad_norm_pre_clip': 31.639644622802734, 'timing/data': 0.0002281600027345121, 'timing/model': 0.3952601390192285, 'learning_rate/qwen_vl_interface': 9.798439583126625e-06, 'learning_rate/action_model': 9.798439583126625e-05, 'epoch': 0.04}) +[2026-10-03 19:01:53,992][starVLA.training.train_starvla][INFO] - Step 557, Loss: {'train/loss': 0.4856805205345154, 'train/grad_norm_pre_clip': 67.96153259277344, 'timing/data': 0.00026974998763762414, 'timing/model': 0.39567942899884656, 'learning_rate/qwen_vl_interface': 9.797560917308126e-06, 'learning_rate/action_model': 9.797560917308126e-05, 'epoch': 0.04}) +[2026-10-03 19:01:54,699][starVLA.training.train_starvla][INFO] - Step 558, Loss: {'train/loss': 0.34642958641052246, 'train/grad_norm_pre_clip': 42.01808547973633, 'timing/data': 0.0002229799865745008, 'timing/model': 0.3948743380024098, 'learning_rate/qwen_vl_interface': 9.796680382159112e-06, 'learning_rate/action_model': 9.79668038215911e-05, 'epoch': 0.04}) +[2026-10-03 19:01:55,412][starVLA.training.train_starvla][INFO] - Step 559, Loss: {'train/loss': 0.1226559579372406, 'train/grad_norm_pre_clip': 10.91944694519043, 'timing/data': 0.00023047998547554016, 'timing/model': 0.3956718689878471, 'learning_rate/qwen_vl_interface': 9.795797978041533e-06, 'learning_rate/action_model': 9.795797978041533e-05, 'epoch': 0.04}) +[2026-10-03 19:01:56,124][starVLA.training.train_starvla][INFO] - Step 560, Loss: {'train/loss': 0.14789700508117676, 'train/grad_norm_pre_clip': 25.738380432128906, 'timing/data': 0.0003387600008863956, 'timing/model': 0.3966355280135758, 'learning_rate/qwen_vl_interface': 9.794913705318117e-06, 'learning_rate/action_model': 9.794913705318118e-05, 'epoch': 0.04}) +[2026-10-03 19:01:57,037][starVLA.training.train_starvla][INFO] - Step 561, Loss: {'train/loss': 0.17220357060432434, 'train/grad_norm_pre_clip': 21.910350799560547, 'timing/data': 0.00023718000738881528, 'timing/model': 0.5999195479962509, 'learning_rate/qwen_vl_interface': 9.794027564352352e-06, 'learning_rate/action_model': 9.794027564352352e-05, 'epoch': 0.04}) +[2026-10-03 19:01:57,744][starVLA.training.train_starvla][INFO] - Step 562, Loss: {'train/loss': 0.22561265528202057, 'train/grad_norm_pre_clip': 166.7161407470703, 'timing/data': 0.0003837200056295842, 'timing/model': 0.3956820780003909, 'learning_rate/qwen_vl_interface': 9.793139555508501e-06, 'learning_rate/action_model': 9.7931395555085e-05, 'epoch': 0.04}) +[2026-10-03 19:01:58,454][starVLA.training.train_starvla][INFO] - Step 563, Loss: {'train/loss': 0.7450805902481079, 'train/grad_norm_pre_clip': 30.16752815246582, 'timing/data': 0.00022424900089390576, 'timing/model': 0.3970282099908218, 'learning_rate/qwen_vl_interface': 9.792249679151585e-06, 'learning_rate/action_model': 9.792249679151585e-05, 'epoch': 0.04}) +[2026-10-03 19:01:59,163][starVLA.training.train_starvla][INFO] - Step 564, Loss: {'train/loss': 0.7188559174537659, 'train/grad_norm_pre_clip': 98.53633117675781, 'timing/data': 0.000398130010580644, 'timing/model': 0.39479963897611015, 'learning_rate/qwen_vl_interface': 9.791357935647403e-06, 'learning_rate/action_model': 9.791357935647403e-05, 'epoch': 0.04}) +[2026-10-03 19:01:59,873][starVLA.training.train_starvla][INFO] - Step 565, Loss: {'train/loss': 0.5488722324371338, 'train/grad_norm_pre_clip': 77.18206024169922, 'timing/data': 0.00033791997702792287, 'timing/model': 0.39707161899423227, 'learning_rate/qwen_vl_interface': 9.790464325362516e-06, 'learning_rate/action_model': 9.790464325362516e-05, 'epoch': 0.04}) +[2026-10-03 19:02:00,780][starVLA.training.train_starvla][INFO] - Step 566, Loss: {'train/loss': 0.14755071699619293, 'train/grad_norm_pre_clip': 15.470174789428711, 'timing/data': 0.00037351998616941273, 'timing/model': 0.5936031570017803, 'learning_rate/qwen_vl_interface': 9.789568848664254e-06, 'learning_rate/action_model': 9.789568848664253e-05, 'epoch': 0.04}) +[2026-10-03 19:02:01,489][starVLA.training.train_starvla][INFO] - Step 567, Loss: {'train/loss': 0.8121340870857239, 'train/grad_norm_pre_clip': 39.21778106689453, 'timing/data': 0.00029757001902908087, 'timing/model': 0.3967058979906142, 'learning_rate/qwen_vl_interface': 9.788671505920712e-06, 'learning_rate/action_model': 9.788671505920711e-05, 'epoch': 0.04}) +[2026-10-03 19:02:02,199][starVLA.training.train_starvla][INFO] - Step 568, Loss: {'train/loss': 0.9830581545829773, 'train/grad_norm_pre_clip': 118.65717315673828, 'timing/data': 0.00022381998132914305, 'timing/model': 0.39627577798091806, 'learning_rate/qwen_vl_interface': 9.787772297500754e-06, 'learning_rate/action_model': 9.787772297500754e-05, 'epoch': 0.04}) +[2026-10-03 19:02:02,907][starVLA.training.train_starvla][INFO] - Step 569, Loss: {'train/loss': 0.8287935256958008, 'train/grad_norm_pre_clip': 110.14070129394531, 'timing/data': 0.0003367600147612393, 'timing/model': 0.39662533800583333, 'learning_rate/qwen_vl_interface': 9.78687122377401e-06, 'learning_rate/action_model': 9.78687122377401e-05, 'epoch': 0.04}) +[2026-10-03 19:02:03,613][starVLA.training.train_starvla][INFO] - Step 570, Loss: {'train/loss': 0.3425423204898834, 'train/grad_norm_pre_clip': 60.58976364135742, 'timing/data': 0.0003389490011613816, 'timing/model': 0.39560462901135907, 'learning_rate/qwen_vl_interface': 9.78596828511088e-06, 'learning_rate/action_model': 9.78596828511088e-05, 'epoch': 0.04}) +[2026-10-03 19:02:04,324][starVLA.training.train_starvla][INFO] - Step 571, Loss: {'train/loss': 0.5172891616821289, 'train/grad_norm_pre_clip': 49.40409851074219, 'timing/data': 0.0002324800007045269, 'timing/model': 0.3967845189908985, 'learning_rate/qwen_vl_interface': 9.785063481882527e-06, 'learning_rate/action_model': 9.785063481882527e-05, 'epoch': 0.04}) +[2026-10-03 19:02:05,240][starVLA.training.train_starvla][INFO] - Step 572, Loss: {'train/loss': 0.20010247826576233, 'train/grad_norm_pre_clip': 143.2952880859375, 'timing/data': 0.00031374001991935074, 'timing/model': 0.3937431590165943, 'learning_rate/qwen_vl_interface': 9.78415681446088e-06, 'learning_rate/action_model': 9.784156814460879e-05, 'epoch': 0.04}) +[2026-10-03 19:02:05,948][starVLA.training.train_starvla][INFO] - Step 573, Loss: {'train/loss': 0.12326575815677643, 'train/grad_norm_pre_clip': 29.028352737426758, 'timing/data': 0.0002537099935580045, 'timing/model': 0.3963677470164839, 'learning_rate/qwen_vl_interface': 9.783248283218638e-06, 'learning_rate/action_model': 9.783248283218638e-05, 'epoch': 0.04}) +[2026-10-03 19:02:06,656][starVLA.training.train_starvla][INFO] - Step 574, Loss: {'train/loss': 0.5584583282470703, 'train/grad_norm_pre_clip': 93.55494689941406, 'timing/data': 0.00032330999965779483, 'timing/model': 0.39531144898501225, 'learning_rate/qwen_vl_interface': 9.782337888529261e-06, 'learning_rate/action_model': 9.782337888529261e-05, 'epoch': 0.04}) +[2026-10-03 19:02:07,364][starVLA.training.train_starvla][INFO] - Step 575, Loss: {'train/loss': 0.039654456079006195, 'train/grad_norm_pre_clip': 14.884711265563965, 'timing/data': 0.0002601400192361325, 'timing/model': 0.39596396798151545, 'learning_rate/qwen_vl_interface': 9.781425630766978e-06, 'learning_rate/action_model': 9.781425630766979e-05, 'epoch': 0.04}) +[2026-10-03 19:02:08,073][starVLA.training.train_starvla][INFO] - Step 576, Loss: {'train/loss': 0.21157391369342804, 'train/grad_norm_pre_clip': 14.795222282409668, 'timing/data': 0.00026713000261224806, 'timing/model': 0.39488205799716525, 'learning_rate/qwen_vl_interface': 9.780511510306789e-06, 'learning_rate/action_model': 9.780511510306788e-05, 'epoch': 0.04}) +[2026-10-03 19:02:08,971][starVLA.training.train_starvla][INFO] - Step 577, Loss: {'train/loss': 0.6671223640441895, 'train/grad_norm_pre_clip': 64.11542510986328, 'timing/data': 0.00021868001203984022, 'timing/model': 0.394714017980732, 'learning_rate/qwen_vl_interface': 9.779595527524447e-06, 'learning_rate/action_model': 9.779595527524447e-05, 'epoch': 0.04}) +[2026-10-03 19:02:09,675][starVLA.training.train_starvla][INFO] - Step 578, Loss: {'train/loss': 0.1834951639175415, 'train/grad_norm_pre_clip': 66.99356079101562, 'timing/data': 0.00031229000887833536, 'timing/model': 0.39434468900435604, 'learning_rate/qwen_vl_interface': 9.778677682796483e-06, 'learning_rate/action_model': 9.778677682796483e-05, 'epoch': 0.04}) +[2026-10-03 19:02:10,381][starVLA.training.train_starvla][INFO] - Step 579, Loss: {'train/loss': 0.2935582399368286, 'train/grad_norm_pre_clip': 38.170372009277344, 'timing/data': 0.0003324599820189178, 'timing/model': 0.39438576900283806, 'learning_rate/qwen_vl_interface': 9.777757976500185e-06, 'learning_rate/action_model': 9.777757976500186e-05, 'epoch': 0.04}) +[2026-10-03 19:02:11,091][starVLA.training.train_starvla][INFO] - Step 580, Loss: {'train/loss': 0.36574435234069824, 'train/grad_norm_pre_clip': 19.174959182739258, 'timing/data': 0.0003811699862126261, 'timing/model': 0.39566617901436985, 'learning_rate/qwen_vl_interface': 9.776836409013614e-06, 'learning_rate/action_model': 9.776836409013615e-05, 'epoch': 0.04}) +[2026-10-03 19:02:11,800][starVLA.training.train_starvla][INFO] - Step 581, Loss: {'train/loss': 0.42718595266342163, 'train/grad_norm_pre_clip': 33.653785705566406, 'timing/data': 0.00036538997665047646, 'timing/model': 0.39668380899820477, 'learning_rate/qwen_vl_interface': 9.775912980715589e-06, 'learning_rate/action_model': 9.775912980715588e-05, 'epoch': 0.04}) +[2026-10-03 19:02:12,713][starVLA.training.train_starvla][INFO] - Step 582, Loss: {'train/loss': 0.40775778889656067, 'train/grad_norm_pre_clip': 13.755298614501953, 'timing/data': 0.0003528700035531074, 'timing/model': 0.5996245679853018, 'learning_rate/qwen_vl_interface': 9.774987691985694e-06, 'learning_rate/action_model': 9.774987691985694e-05, 'epoch': 0.04}) +[2026-10-03 19:02:13,419][starVLA.training.train_starvla][INFO] - Step 583, Loss: {'train/loss': 0.16980904340744019, 'train/grad_norm_pre_clip': 17.020854949951172, 'timing/data': 0.00022333001834340394, 'timing/model': 0.3940463690087199, 'learning_rate/qwen_vl_interface': 9.774060543204283e-06, 'learning_rate/action_model': 9.774060543204283e-05, 'epoch': 0.04}) +[2026-10-03 19:02:14,128][starVLA.training.train_starvla][INFO] - Step 584, Loss: {'train/loss': 0.43529704213142395, 'train/grad_norm_pre_clip': 98.83291625976562, 'timing/data': 0.00023411001893691719, 'timing/model': 0.39609409798868, 'learning_rate/qwen_vl_interface': 9.773131534752472e-06, 'learning_rate/action_model': 9.773131534752472e-05, 'epoch': 0.04}) +[2026-10-03 19:02:14,836][starVLA.training.train_starvla][INFO] - Step 585, Loss: {'train/loss': 0.028255203738808632, 'train/grad_norm_pre_clip': 2.926292896270752, 'timing/data': 0.0003980000037699938, 'timing/model': 0.3968557380139828, 'learning_rate/qwen_vl_interface': 9.772200667012142e-06, 'learning_rate/action_model': 9.772200667012142e-05, 'epoch': 0.04}) +[2026-10-03 19:02:15,545][starVLA.training.train_starvla][INFO] - Step 586, Loss: {'train/loss': 0.6691295504570007, 'train/grad_norm_pre_clip': 42.79620361328125, 'timing/data': 0.0002267099916934967, 'timing/model': 0.3960306880180724, 'learning_rate/qwen_vl_interface': 9.771267940365934e-06, 'learning_rate/action_model': 9.771267940365935e-05, 'epoch': 0.04}) +[2026-10-03 19:02:16,453][starVLA.training.train_starvla][INFO] - Step 587, Loss: {'train/loss': 0.4643828570842743, 'train/grad_norm_pre_clip': 23.741863250732422, 'timing/data': 0.0003494099946692586, 'timing/model': 0.5964108670013957, 'learning_rate/qwen_vl_interface': 9.770333355197262e-06, 'learning_rate/action_model': 9.77033335519726e-05, 'epoch': 0.04}) +[2026-10-03 19:02:17,162][starVLA.training.train_starvla][INFO] - Step 588, Loss: {'train/loss': 0.11852788925170898, 'train/grad_norm_pre_clip': 15.546103477478027, 'timing/data': 0.00023129000328481197, 'timing/model': 0.39560516900382936, 'learning_rate/qwen_vl_interface': 9.769396911890291e-06, 'learning_rate/action_model': 9.769396911890291e-05, 'epoch': 0.04}) +[2026-10-03 19:02:17,871][starVLA.training.train_starvla][INFO] - Step 589, Loss: {'train/loss': 0.32946422696113586, 'train/grad_norm_pre_clip': 91.79811096191406, 'timing/data': 0.0002691200061235577, 'timing/model': 0.3947588480077684, 'learning_rate/qwen_vl_interface': 9.768458610829965e-06, 'learning_rate/action_model': 9.768458610829965e-05, 'epoch': 0.04}) +[2026-10-03 19:02:18,576][starVLA.training.train_starvla][INFO] - Step 590, Loss: {'train/loss': 0.30392056703567505, 'train/grad_norm_pre_clip': 185.86790466308594, 'timing/data': 0.0002604999754112214, 'timing/model': 0.3935350980027579, 'learning_rate/qwen_vl_interface': 9.767518452401981e-06, 'learning_rate/action_model': 9.76751845240198e-05, 'epoch': 0.04}) +[2026-10-03 19:02:19,284][starVLA.training.train_starvla][INFO] - Step 591, Loss: {'train/loss': 0.10232620686292648, 'train/grad_norm_pre_clip': 45.10839080810547, 'timing/data': 0.00021518999710679054, 'timing/model': 0.3959009689860977, 'learning_rate/qwen_vl_interface': 9.766576436992802e-06, 'learning_rate/action_model': 9.766576436992801e-05, 'epoch': 0.04}) +[2026-10-03 19:02:19,992][starVLA.training.train_starvla][INFO] - Step 592, Loss: {'train/loss': 0.178105890750885, 'train/grad_norm_pre_clip': 104.13330078125, 'timing/data': 0.0002583499881438911, 'timing/model': 0.3950911680003628, 'learning_rate/qwen_vl_interface': 9.765632564989657e-06, 'learning_rate/action_model': 9.765632564989656e-05, 'epoch': 0.04}) +[2026-10-03 19:02:20,903][starVLA.training.train_starvla][INFO] - Step 593, Loss: {'train/loss': 0.34594404697418213, 'train/grad_norm_pre_clip': 41.46097946166992, 'timing/data': 0.00037908999365754426, 'timing/model': 0.3965205170097761, 'learning_rate/qwen_vl_interface': 9.764686836780533e-06, 'learning_rate/action_model': 9.764686836780532e-05, 'epoch': 0.04}) +[2026-10-03 19:02:21,613][starVLA.training.train_starvla][INFO] - Step 594, Loss: {'train/loss': 0.1772761046886444, 'train/grad_norm_pre_clip': 31.128097534179688, 'timing/data': 0.00036769002326764166, 'timing/model': 0.39698499799123965, 'learning_rate/qwen_vl_interface': 9.763739252754187e-06, 'learning_rate/action_model': 9.763739252754186e-05, 'epoch': 0.04}) +[2026-10-03 19:02:22,325][starVLA.training.train_starvla][INFO] - Step 595, Loss: {'train/loss': 0.16541551053524017, 'train/grad_norm_pre_clip': 25.942699432373047, 'timing/data': 0.0004021200002171099, 'timing/model': 0.39696602898766287, 'learning_rate/qwen_vl_interface': 9.76278981330013e-06, 'learning_rate/action_model': 9.762789813300131e-05, 'epoch': 0.04}) +[2026-10-03 19:02:23,036][starVLA.training.train_starvla][INFO] - Step 596, Loss: {'train/loss': 0.02463480830192566, 'train/grad_norm_pre_clip': 4.438657283782959, 'timing/data': 0.00023197999689728022, 'timing/model': 0.3975074469926767, 'learning_rate/qwen_vl_interface': 9.761838518808647e-06, 'learning_rate/action_model': 9.761838518808646e-05, 'epoch': 0.04}) +[2026-10-03 19:02:23,750][starVLA.training.train_starvla][INFO] - Step 597, Loss: {'train/loss': 0.564703106880188, 'train/grad_norm_pre_clip': 95.05758666992188, 'timing/data': 0.00028447998920455575, 'timing/model': 0.3979648789972998, 'learning_rate/qwen_vl_interface': 9.760885369670776e-06, 'learning_rate/action_model': 9.760885369670776e-05, 'epoch': 0.04}) +[2026-10-03 19:02:24,667][starVLA.training.train_starvla][INFO] - Step 598, Loss: {'train/loss': 0.40554219484329224, 'train/grad_norm_pre_clip': 62.35682678222656, 'timing/data': 0.00039011999615468085, 'timing/model': 0.39758964799693786, 'learning_rate/qwen_vl_interface': 9.759930366278319e-06, 'learning_rate/action_model': 9.759930366278318e-05, 'epoch': 0.04}) +[2026-10-03 19:02:25,378][starVLA.training.train_starvla][INFO] - Step 599, Loss: {'train/loss': 0.5315341949462891, 'train/grad_norm_pre_clip': 119.06704711914062, 'timing/data': 0.00029563999851234257, 'timing/model': 0.3972297789878212, 'learning_rate/qwen_vl_interface': 9.758973509023842e-06, 'learning_rate/action_model': 9.758973509023843e-05, 'epoch': 0.04}) +[2026-10-03 19:02:26,090][starVLA.training.train_starvla][INFO] - Step 600, Loss: {'train/loss': 0.23953837156295776, 'train/grad_norm_pre_clip': 102.95053100585938, 'timing/data': 0.0004077100020367652, 'timing/model': 0.39661778902518563, 'learning_rate/qwen_vl_interface': 9.758014798300677e-06, 'learning_rate/action_model': 9.758014798300676e-05, 'epoch': 0.04}) +[2026-10-03 19:02:26,802][starVLA.training.train_starvla][INFO] - Step 601, Loss: {'train/loss': 0.8914152383804321, 'train/grad_norm_pre_clip': 65.14085388183594, 'timing/data': 0.0002620199811644852, 'timing/model': 0.3977144389937166, 'learning_rate/qwen_vl_interface': 9.757054234502909e-06, 'learning_rate/action_model': 9.757054234502909e-05, 'epoch': 0.04}) +[2026-10-03 19:02:27,512][starVLA.training.train_starvla][INFO] - Step 602, Loss: {'train/loss': 0.16794070601463318, 'train/grad_norm_pre_clip': 16.417770385742188, 'timing/data': 0.00035119999665766954, 'timing/model': 0.3972902790119406, 'learning_rate/qwen_vl_interface': 9.756091818025394e-06, 'learning_rate/action_model': 9.756091818025394e-05, 'epoch': 0.04}) +[2026-10-03 19:02:28,424][starVLA.training.train_starvla][INFO] - Step 603, Loss: {'train/loss': 0.08264138549566269, 'train/grad_norm_pre_clip': 18.46770477294922, 'timing/data': 0.00022412999533116817, 'timing/model': 0.5991662169981282, 'learning_rate/qwen_vl_interface': 9.75512754926374e-06, 'learning_rate/action_model': 9.755127549263741e-05, 'epoch': 0.04}) +[2026-10-03 19:02:29,134][starVLA.training.train_starvla][INFO] - Step 604, Loss: {'train/loss': 0.24059194326400757, 'train/grad_norm_pre_clip': 33.8706169128418, 'timing/data': 0.00030250000418163836, 'timing/model': 0.39588587800972164, 'learning_rate/qwen_vl_interface': 9.754161428614325e-06, 'learning_rate/action_model': 9.754161428614324e-05, 'epoch': 0.04}) +[2026-10-03 19:02:29,847][starVLA.training.train_starvla][INFO] - Step 605, Loss: {'train/loss': 0.7316373586654663, 'train/grad_norm_pre_clip': 114.38127899169922, 'timing/data': 0.0003755100187845528, 'timing/model': 0.3981013880111277, 'learning_rate/qwen_vl_interface': 9.753193456474284e-06, 'learning_rate/action_model': 9.753193456474284e-05, 'epoch': 0.04}) +[2026-10-03 19:02:30,557][starVLA.training.train_starvla][INFO] - Step 606, Loss: {'train/loss': 0.23998954892158508, 'train/grad_norm_pre_clip': 23.054384231567383, 'timing/data': 0.0002270700060762465, 'timing/model': 0.39672975800931454, 'learning_rate/qwen_vl_interface': 9.752223633241514e-06, 'learning_rate/action_model': 9.752223633241514e-05, 'epoch': 0.04}) +[2026-10-03 19:02:31,266][starVLA.training.train_starvla][INFO] - Step 607, Loss: {'train/loss': 0.32166948914527893, 'train/grad_norm_pre_clip': 99.4928970336914, 'timing/data': 0.00032338997698388994, 'timing/model': 0.39633720900746994, 'learning_rate/qwen_vl_interface': 9.751251959314671e-06, 'learning_rate/action_model': 9.751251959314672e-05, 'epoch': 0.04}) +[2026-10-03 19:02:32,182][starVLA.training.train_starvla][INFO] - Step 608, Loss: {'train/loss': 0.2598100006580353, 'train/grad_norm_pre_clip': 24.82045555114746, 'timing/data': 0.00037511001573875546, 'timing/model': 0.6007778380007949, 'learning_rate/qwen_vl_interface': 9.750278435093176e-06, 'learning_rate/action_model': 9.750278435093177e-05, 'epoch': 0.04}) +[2026-10-03 19:02:32,893][starVLA.training.train_starvla][INFO] - Step 609, Loss: {'train/loss': 0.11650925874710083, 'train/grad_norm_pre_clip': 16.386045455932617, 'timing/data': 0.0003557599848136306, 'timing/model': 0.39758569799596444, 'learning_rate/qwen_vl_interface': 9.749303060977205e-06, 'learning_rate/action_model': 9.749303060977205e-05, 'epoch': 0.04}) +[2026-10-03 19:02:33,601][starVLA.training.train_starvla][INFO] - Step 610, Loss: {'train/loss': 0.07455747574567795, 'train/grad_norm_pre_clip': 8.080687522888184, 'timing/data': 0.00022614997578784823, 'timing/model': 0.39475545001914725, 'learning_rate/qwen_vl_interface': 9.748325837367701e-06, 'learning_rate/action_model': 9.748325837367702e-05, 'epoch': 0.04}) +[2026-10-03 19:02:34,308][starVLA.training.train_starvla][INFO] - Step 611, Loss: {'train/loss': 0.2269599735736847, 'train/grad_norm_pre_clip': 89.79296112060547, 'timing/data': 0.00045948001206852496, 'timing/model': 0.39469532898510806, 'learning_rate/qwen_vl_interface': 9.747346764666362e-06, 'learning_rate/action_model': 9.747346764666361e-05, 'epoch': 0.04}) +[2026-10-03 19:02:35,017][starVLA.training.train_starvla][INFO] - Step 612, Loss: {'train/loss': 0.2513878345489502, 'train/grad_norm_pre_clip': 90.96492004394531, 'timing/data': 0.00022660999093204737, 'timing/model': 0.3957436269847676, 'learning_rate/qwen_vl_interface': 9.746365843275648e-06, 'learning_rate/action_model': 9.746365843275646e-05, 'epoch': 0.04}) +[2026-10-03 19:02:35,725][starVLA.training.train_starvla][INFO] - Step 613, Loss: {'train/loss': 0.5455420613288879, 'train/grad_norm_pre_clip': 117.17033386230469, 'timing/data': 0.00032474001636728644, 'timing/model': 0.3965883889759425, 'learning_rate/qwen_vl_interface': 9.745383073598776e-06, 'learning_rate/action_model': 9.745383073598775e-05, 'epoch': 0.04}) +[2026-10-03 19:02:36,632][starVLA.training.train_starvla][INFO] - Step 614, Loss: {'train/loss': 1.161057710647583, 'train/grad_norm_pre_clip': 58.69573974609375, 'timing/data': 0.0003720800159499049, 'timing/model': 0.3949376189848408, 'learning_rate/qwen_vl_interface': 9.744398456039729e-06, 'learning_rate/action_model': 9.744398456039729e-05, 'epoch': 0.04}) +[2026-10-03 19:02:37,339][starVLA.training.train_starvla][INFO] - Step 615, Loss: {'train/loss': 0.44004562497138977, 'train/grad_norm_pre_clip': 33.691383361816406, 'timing/data': 0.00022245000582188368, 'timing/model': 0.39465037899208255, 'learning_rate/qwen_vl_interface': 9.743411991003246e-06, 'learning_rate/action_model': 9.743411991003247e-05, 'epoch': 0.05}) +[2026-10-03 19:02:38,049][starVLA.training.train_starvla][INFO] - Step 616, Loss: {'train/loss': 1.2160367965698242, 'train/grad_norm_pre_clip': 135.61642456054688, 'timing/data': 0.0005138900014571846, 'timing/model': 0.39579751901328564, 'learning_rate/qwen_vl_interface': 9.742423678894823e-06, 'learning_rate/action_model': 9.742423678894823e-05, 'epoch': 0.05}) +[2026-10-03 19:02:38,759][starVLA.training.train_starvla][INFO] - Step 617, Loss: {'train/loss': 1.0449893474578857, 'train/grad_norm_pre_clip': 183.20188903808594, 'timing/data': 0.00022733001969754696, 'timing/model': 0.3969839890196454, 'learning_rate/qwen_vl_interface': 9.741433520120722e-06, 'learning_rate/action_model': 9.74143352012072e-05, 'epoch': 0.05}) +[2026-10-03 19:02:39,465][starVLA.training.train_starvla][INFO] - Step 618, Loss: {'train/loss': 0.7830173969268799, 'train/grad_norm_pre_clip': 55.25538635253906, 'timing/data': 0.00030496998806484044, 'timing/model': 0.39442795899230987, 'learning_rate/qwen_vl_interface': 9.740441515087953e-06, 'learning_rate/action_model': 9.740441515087953e-05, 'epoch': 0.05}) +[2026-10-03 19:02:40,376][starVLA.training.train_starvla][INFO] - Step 619, Loss: {'train/loss': 0.4123467803001404, 'train/grad_norm_pre_clip': 63.18170166015625, 'timing/data': 0.0002574700047262013, 'timing/model': 0.39538701801211573, 'learning_rate/qwen_vl_interface': 9.739447664204298e-06, 'learning_rate/action_model': 9.739447664204297e-05, 'epoch': 0.05}) +[2026-10-03 19:02:41,083][starVLA.training.train_starvla][INFO] - Step 620, Loss: {'train/loss': 0.21640263497829437, 'train/grad_norm_pre_clip': 52.19065856933594, 'timing/data': 0.00024440998095087707, 'timing/model': 0.39374981899163686, 'learning_rate/qwen_vl_interface': 9.73845196787829e-06, 'learning_rate/action_model': 9.738451967878289e-05, 'epoch': 0.05}) +[2026-10-03 19:02:41,791][starVLA.training.train_starvla][INFO] - Step 621, Loss: {'train/loss': 0.5453571081161499, 'train/grad_norm_pre_clip': 107.0473403930664, 'timing/data': 0.00039137000567279756, 'timing/model': 0.3956029480032157, 'learning_rate/qwen_vl_interface': 9.737454426519221e-06, 'learning_rate/action_model': 9.737454426519221e-05, 'epoch': 0.05}) +[2026-10-03 19:02:42,496][starVLA.training.train_starvla][INFO] - Step 622, Loss: {'train/loss': 0.2730371356010437, 'train/grad_norm_pre_clip': 110.49669647216797, 'timing/data': 0.00033921998692676425, 'timing/model': 0.3944520389777608, 'learning_rate/qwen_vl_interface': 9.736455040537145e-06, 'learning_rate/action_model': 9.736455040537144e-05, 'epoch': 0.05}) +[2026-10-03 19:02:43,201][starVLA.training.train_starvla][INFO] - Step 623, Loss: {'train/loss': 0.34917742013931274, 'train/grad_norm_pre_clip': 21.58981704711914, 'timing/data': 0.00022515997989103198, 'timing/model': 0.39514395798323676, 'learning_rate/qwen_vl_interface': 9.735453810342868e-06, 'learning_rate/action_model': 9.735453810342867e-05, 'epoch': 0.05}) +[2026-10-03 19:02:44,106][starVLA.training.train_starvla][INFO] - Step 624, Loss: {'train/loss': 0.10309084504842758, 'train/grad_norm_pre_clip': 10.8337984085083, 'timing/data': 0.00037215001066215336, 'timing/model': 0.5911237990076188, 'learning_rate/qwen_vl_interface': 9.734450736347958e-06, 'learning_rate/action_model': 9.73445073634796e-05, 'epoch': 0.05}) +[2026-10-03 19:02:44,813][starVLA.training.train_starvla][INFO] - Step 625, Loss: {'train/loss': 0.42373183369636536, 'train/grad_norm_pre_clip': 46.92658233642578, 'timing/data': 0.0003148999821860343, 'timing/model': 0.39555198900052346, 'learning_rate/qwen_vl_interface': 9.733445818964746e-06, 'learning_rate/action_model': 9.733445818964746e-05, 'epoch': 0.05}) +[2026-10-03 19:02:45,520][starVLA.training.train_starvla][INFO] - Step 626, Loss: {'train/loss': 0.4306497573852539, 'train/grad_norm_pre_clip': 33.685813903808594, 'timing/data': 0.00028482999186962843, 'timing/model': 0.3948314989975188, 'learning_rate/qwen_vl_interface': 9.732439058606311e-06, 'learning_rate/action_model': 9.732439058606312e-05, 'epoch': 0.05}) +[2026-10-03 19:02:46,224][starVLA.training.train_starvla][INFO] - Step 627, Loss: {'train/loss': 0.42515113949775696, 'train/grad_norm_pre_clip': 17.112714767456055, 'timing/data': 0.0002965400053653866, 'timing/model': 0.3941310379887, 'learning_rate/qwen_vl_interface': 9.731430455686499e-06, 'learning_rate/action_model': 9.731430455686498e-05, 'epoch': 0.05}) +[2026-10-03 19:02:46,929][starVLA.training.train_starvla][INFO] - Step 628, Loss: {'train/loss': 0.158058300614357, 'train/grad_norm_pre_clip': 16.09812355041504, 'timing/data': 0.0002236299915239215, 'timing/model': 0.3941069300053641, 'learning_rate/qwen_vl_interface': 9.730420010619901e-06, 'learning_rate/action_model': 9.7304200106199e-05, 'epoch': 0.05}) +[2026-10-03 19:02:47,838][starVLA.training.train_starvla][INFO] - Step 629, Loss: {'train/loss': 0.08521997928619385, 'train/grad_norm_pre_clip': 13.25466537475586, 'timing/data': 0.0003158999898005277, 'timing/model': 0.5982162180007435, 'learning_rate/qwen_vl_interface': 9.72940772382188e-06, 'learning_rate/action_model': 9.729407723821878e-05, 'epoch': 0.05}) +[2026-10-03 19:02:48,544][starVLA.training.train_starvla][INFO] - Step 630, Loss: {'train/loss': 0.08883348852396011, 'train/grad_norm_pre_clip': 36.06521224975586, 'timing/data': 0.000265929993474856, 'timing/model': 0.3952575580042321, 'learning_rate/qwen_vl_interface': 9.728393595708542e-06, 'learning_rate/action_model': 9.728393595708542e-05, 'epoch': 0.05}) +[2026-10-03 19:02:49,252][starVLA.training.train_starvla][INFO] - Step 631, Loss: {'train/loss': 0.7832975387573242, 'train/grad_norm_pre_clip': 111.77204895019531, 'timing/data': 0.0002292900171596557, 'timing/model': 0.3960907180153299, 'learning_rate/qwen_vl_interface': 9.727377626696762e-06, 'learning_rate/action_model': 9.727377626696762e-05, 'epoch': 0.05}) +[2026-10-03 19:02:49,961][starVLA.training.train_starvla][INFO] - Step 632, Loss: {'train/loss': 0.022403478622436523, 'train/grad_norm_pre_clip': 3.412767171859741, 'timing/data': 0.0003103199996985495, 'timing/model': 0.3960066789877601, 'learning_rate/qwen_vl_interface': 9.726359817204165e-06, 'learning_rate/action_model': 9.726359817204166e-05, 'epoch': 0.05}) +[2026-10-03 19:02:50,670][starVLA.training.train_starvla][INFO] - Step 633, Loss: {'train/loss': 0.03385854512453079, 'train/grad_norm_pre_clip': 20.918861389160156, 'timing/data': 0.0002206799981649965, 'timing/model': 0.39633918899926357, 'learning_rate/qwen_vl_interface': 9.725340167649136e-06, 'learning_rate/action_model': 9.725340167649136e-05, 'epoch': 0.05}) +[2026-10-03 19:02:51,379][starVLA.training.train_starvla][INFO] - Step 634, Loss: {'train/loss': 0.8893124461174011, 'train/grad_norm_pre_clip': 49.04048156738281, 'timing/data': 0.00022772999363951385, 'timing/model': 0.39625549799529836, 'learning_rate/qwen_vl_interface': 9.72431867845081e-06, 'learning_rate/action_model': 9.72431867845081e-05, 'epoch': 0.05}) +[2026-10-03 19:02:52,290][starVLA.training.train_starvla][INFO] - Step 635, Loss: {'train/loss': 0.05999673902988434, 'train/grad_norm_pre_clip': 28.469755172729492, 'timing/data': 0.00023055999190546572, 'timing/model': 0.39600564999273047, 'learning_rate/qwen_vl_interface': 9.723295350029086e-06, 'learning_rate/action_model': 9.723295350029086e-05, 'epoch': 0.05}) +[2026-10-03 19:02:52,998][starVLA.training.train_starvla][INFO] - Step 636, Loss: {'train/loss': 0.8599200248718262, 'train/grad_norm_pre_clip': 59.51528549194336, 'timing/data': 0.00036952001391910017, 'timing/model': 0.3936189789965283, 'learning_rate/qwen_vl_interface': 9.722270182804616e-06, 'learning_rate/action_model': 9.722270182804614e-05, 'epoch': 0.05}) +[2026-10-03 19:02:53,705][starVLA.training.train_starvla][INFO] - Step 637, Loss: {'train/loss': 0.006276056636124849, 'train/grad_norm_pre_clip': 0.9390423893928528, 'timing/data': 0.0003195200115442276, 'timing/model': 0.3939158779976424, 'learning_rate/qwen_vl_interface': 9.721243177198805e-06, 'learning_rate/action_model': 9.721243177198804e-05, 'epoch': 0.05}) +[2026-10-03 19:02:54,410][starVLA.training.train_starvla][INFO] - Step 638, Loss: {'train/loss': 0.5864028930664062, 'train/grad_norm_pre_clip': 37.6180419921875, 'timing/data': 0.0002192499814555049, 'timing/model': 0.3937095180153847, 'learning_rate/qwen_vl_interface': 9.720214333633817e-06, 'learning_rate/action_model': 9.720214333633817e-05, 'epoch': 0.05}) +[2026-10-03 19:02:55,117][starVLA.training.train_starvla][INFO] - Step 639, Loss: {'train/loss': 0.835720956325531, 'train/grad_norm_pre_clip': 27.669557571411133, 'timing/data': 0.00033614999847486615, 'timing/model': 0.3948893089836929, 'learning_rate/qwen_vl_interface': 9.719183652532572e-06, 'learning_rate/action_model': 9.719183652532572e-05, 'epoch': 0.05}) +[2026-10-03 19:02:56,024][starVLA.training.train_starvla][INFO] - Step 640, Loss: {'train/loss': 0.022540945559740067, 'train/grad_norm_pre_clip': 18.836040496826172, 'timing/data': 0.0003698999935295433, 'timing/model': 0.5941361580044031, 'learning_rate/qwen_vl_interface': 9.718151134318743e-06, 'learning_rate/action_model': 9.718151134318743e-05, 'epoch': 0.05}) +[2026-10-03 19:02:56,733][starVLA.training.train_starvla][INFO] - Step 641, Loss: {'train/loss': 0.8530685901641846, 'train/grad_norm_pre_clip': 38.662086486816406, 'timing/data': 0.0003809299960266799, 'timing/model': 0.39560610899934545, 'learning_rate/qwen_vl_interface': 9.71711677941676e-06, 'learning_rate/action_model': 9.717116779416758e-05, 'epoch': 0.05}) +[2026-10-03 19:02:57,438][starVLA.training.train_starvla][INFO] - Step 642, Loss: {'train/loss': 0.005311296787112951, 'train/grad_norm_pre_clip': 1.1808280944824219, 'timing/data': 0.0002186700003221631, 'timing/model': 0.3951760979834944, 'learning_rate/qwen_vl_interface': 9.716080588251805e-06, 'learning_rate/action_model': 9.716080588251805e-05, 'epoch': 0.05}) +[2026-10-03 19:02:58,145][starVLA.training.train_starvla][INFO] - Step 643, Loss: {'train/loss': 0.2716994285583496, 'train/grad_norm_pre_clip': 70.40912628173828, 'timing/data': 0.0002548999909777194, 'timing/model': 0.39517078798962757, 'learning_rate/qwen_vl_interface': 9.71504256124982e-06, 'learning_rate/action_model': 9.71504256124982e-05, 'epoch': 0.05}) +[2026-10-03 19:02:58,852][starVLA.training.train_starvla][INFO] - Step 644, Loss: {'train/loss': 0.6916481256484985, 'train/grad_norm_pre_clip': 44.0579719543457, 'timing/data': 0.0003450399963185191, 'timing/model': 0.39493482900434174, 'learning_rate/qwen_vl_interface': 9.714002698837499e-06, 'learning_rate/action_model': 9.714002698837498e-05, 'epoch': 0.05}) +[2026-10-03 19:02:59,763][starVLA.training.train_starvla][INFO] - Step 645, Loss: {'train/loss': 0.015212349593639374, 'train/grad_norm_pre_clip': 25.506702423095703, 'timing/data': 0.00026763000641949475, 'timing/model': 0.5991574680083431, 'learning_rate/qwen_vl_interface': 9.712961001442285e-06, 'learning_rate/action_model': 9.712961001442286e-05, 'epoch': 0.05}) +[2026-10-03 19:03:00,469][starVLA.training.train_starvla][INFO] - Step 646, Loss: {'train/loss': 0.29254812002182007, 'train/grad_norm_pre_clip': 12.205577850341797, 'timing/data': 0.00029853000887669623, 'timing/model': 0.39384197001345456, 'learning_rate/qwen_vl_interface': 9.711917469492388e-06, 'learning_rate/action_model': 9.711917469492388e-05, 'epoch': 0.05}) +[2026-10-03 19:03:01,176][starVLA.training.train_starvla][INFO] - Step 647, Loss: {'train/loss': 0.04536545276641846, 'train/grad_norm_pre_clip': 26.23948860168457, 'timing/data': 0.00022263001301325858, 'timing/model': 0.39563670798088424, 'learning_rate/qwen_vl_interface': 9.71087210341676e-06, 'learning_rate/action_model': 9.71087210341676e-05, 'epoch': 0.05}) +[2026-10-03 19:03:01,882][starVLA.training.train_starvla][INFO] - Step 648, Loss: {'train/loss': 0.6083651781082153, 'train/grad_norm_pre_clip': 59.020931243896484, 'timing/data': 0.0003727099974639714, 'timing/model': 0.395768528978806, 'learning_rate/qwen_vl_interface': 9.70982490364511e-06, 'learning_rate/action_model': 9.70982490364511e-05, 'epoch': 0.05}) +[2026-10-03 19:03:02,589][starVLA.training.train_starvla][INFO] - Step 649, Loss: {'train/loss': 0.41229334473609924, 'train/grad_norm_pre_clip': 44.912410736083984, 'timing/data': 0.00027362999389879405, 'timing/model': 0.3941536089987494, 'learning_rate/qwen_vl_interface': 9.708775870607905e-06, 'learning_rate/action_model': 9.708775870607904e-05, 'epoch': 0.05}) +[2026-10-03 19:03:03,297][starVLA.training.train_starvla][INFO] - Step 650, Loss: {'train/loss': 0.41144952178001404, 'train/grad_norm_pre_clip': 64.14180755615234, 'timing/data': 0.00024221002240665257, 'timing/model': 0.3943454690161161, 'learning_rate/qwen_vl_interface': 9.707725004736363e-06, 'learning_rate/action_model': 9.707725004736363e-05, 'epoch': 0.05}) +[2026-10-03 19:03:04,203][starVLA.training.train_starvla][INFO] - Step 651, Loss: {'train/loss': 0.2560499906539917, 'train/grad_norm_pre_clip': 165.727294921875, 'timing/data': 0.0003726799914147705, 'timing/model': 0.39484442898537964, 'learning_rate/qwen_vl_interface': 9.706672306462454e-06, 'learning_rate/action_model': 9.706672306462454e-05, 'epoch': 0.05}) +[2026-10-03 19:03:04,909][starVLA.training.train_starvla][INFO] - Step 652, Loss: {'train/loss': 0.726581335067749, 'train/grad_norm_pre_clip': 75.20939636230469, 'timing/data': 0.00025708001339808106, 'timing/model': 0.3961882580188103, 'learning_rate/qwen_vl_interface': 9.705617776218905e-06, 'learning_rate/action_model': 9.705617776218904e-05, 'epoch': 0.05}) +[2026-10-03 19:03:05,613][starVLA.training.train_starvla][INFO] - Step 653, Loss: {'train/loss': 0.6631326675415039, 'train/grad_norm_pre_clip': 29.400575637817383, 'timing/data': 0.00023239999427460134, 'timing/model': 0.39297667899518274, 'learning_rate/qwen_vl_interface': 9.70456141443919e-06, 'learning_rate/action_model': 9.704561414439188e-05, 'epoch': 0.05}) +[2026-10-03 19:03:06,316][starVLA.training.train_starvla][INFO] - Step 654, Loss: {'train/loss': 0.15075929462909698, 'train/grad_norm_pre_clip': 29.925514221191406, 'timing/data': 0.0003360999980941415, 'timing/model': 0.3926692499953788, 'learning_rate/qwen_vl_interface': 9.703503221557538e-06, 'learning_rate/action_model': 9.703503221557538e-05, 'epoch': 0.05}) +[2026-10-03 19:03:07,025][starVLA.training.train_starvla][INFO] - Step 655, Loss: {'train/loss': 0.1196320503950119, 'train/grad_norm_pre_clip': 24.94025993347168, 'timing/data': 0.0003131400153506547, 'timing/model': 0.3959042989881709, 'learning_rate/qwen_vl_interface': 9.70244319800894e-06, 'learning_rate/action_model': 9.70244319800894e-05, 'epoch': 0.05}) +[2026-10-03 19:03:07,928][starVLA.training.train_starvla][INFO] - Step 656, Loss: {'train/loss': 0.31888067722320557, 'train/grad_norm_pre_clip': 121.33882141113281, 'timing/data': 0.00022054000874049962, 'timing/model': 0.3950913290027529, 'learning_rate/qwen_vl_interface': 9.701381344229124e-06, 'learning_rate/action_model': 9.701381344229123e-05, 'epoch': 0.05}) +[2026-10-03 19:03:08,635][starVLA.training.train_starvla][INFO] - Step 657, Loss: {'train/loss': 0.14377576112747192, 'train/grad_norm_pre_clip': 49.046234130859375, 'timing/data': 0.00022209997405298054, 'timing/model': 0.39520320799783804, 'learning_rate/qwen_vl_interface': 9.700317660654581e-06, 'learning_rate/action_model': 9.700317660654581e-05, 'epoch': 0.05}) +[2026-10-03 19:03:09,339][starVLA.training.train_starvla][INFO] - Step 658, Loss: {'train/loss': 0.5989814400672913, 'train/grad_norm_pre_clip': 70.0839614868164, 'timing/data': 0.00021739001385867596, 'timing/model': 0.3937407180201262, 'learning_rate/qwen_vl_interface': 9.699252147722554e-06, 'learning_rate/action_model': 9.699252147722553e-05, 'epoch': 0.05}) +[2026-10-03 19:03:10,023][starVLA.training.train_starvla][INFO] - Step 659, Loss: {'train/loss': 0.5855584144592285, 'train/grad_norm_pre_clip': 46.21198272705078, 'timing/data': 0.00032815002487041056, 'timing/model': 0.39660493002156727, 'learning_rate/qwen_vl_interface': 9.698184805871028e-06, 'learning_rate/action_model': 9.698184805871028e-05, 'epoch': 0.05}) +[2026-10-03 19:03:10,729][starVLA.training.train_starvla][INFO] - Step 660, Loss: {'train/loss': 0.18299511075019836, 'train/grad_norm_pre_clip': 22.946674346923828, 'timing/data': 0.0002648199733812362, 'timing/model': 0.3954732489946764, 'learning_rate/qwen_vl_interface': 9.697115635538753e-06, 'learning_rate/action_model': 9.697115635538753e-05, 'epoch': 0.05}) +[2026-10-03 19:03:11,627][starVLA.training.train_starvla][INFO] - Step 661, Loss: {'train/loss': 0.20435944199562073, 'train/grad_norm_pre_clip': 17.782958984375, 'timing/data': 0.00025574001483619213, 'timing/model': 0.5879324089910369, 'learning_rate/qwen_vl_interface': 9.69604463716522e-06, 'learning_rate/action_model': 9.69604463716522e-05, 'epoch': 0.05}) +[2026-10-03 19:03:12,331][starVLA.training.train_starvla][INFO] - Step 662, Loss: {'train/loss': 0.10548504441976547, 'train/grad_norm_pre_clip': 24.037750244140625, 'timing/data': 0.00022722998983226717, 'timing/model': 0.3941482389927842, 'learning_rate/qwen_vl_interface': 9.69497181119068e-06, 'learning_rate/action_model': 9.694971811190681e-05, 'epoch': 0.05}) +[2026-10-03 19:03:13,039][starVLA.training.train_starvla][INFO] - Step 663, Loss: {'train/loss': 0.7329044342041016, 'train/grad_norm_pre_clip': 119.57035064697266, 'timing/data': 0.0002752599830273539, 'timing/model': 0.39673248800681904, 'learning_rate/qwen_vl_interface': 9.693897158056132e-06, 'learning_rate/action_model': 9.693897158056132e-05, 'epoch': 0.05}) +[2026-10-03 19:03:13,748][starVLA.training.train_starvla][INFO] - Step 664, Loss: {'train/loss': 0.11855489015579224, 'train/grad_norm_pre_clip': 93.12678527832031, 'timing/data': 0.00022814999101683497, 'timing/model': 0.3952384289877955, 'learning_rate/qwen_vl_interface': 9.692820678203322e-06, 'learning_rate/action_model': 9.692820678203322e-05, 'epoch': 0.05}) +[2026-10-03 19:03:14,455][starVLA.training.train_starvla][INFO] - Step 665, Loss: {'train/loss': 0.05419448763132095, 'train/grad_norm_pre_clip': 4.9556965827941895, 'timing/data': 0.00022484001237899065, 'timing/model': 0.3950865790247917, 'learning_rate/qwen_vl_interface': 9.691742372074749e-06, 'learning_rate/action_model': 9.691742372074749e-05, 'epoch': 0.05}) +[2026-10-03 19:03:15,368][starVLA.training.train_starvla][INFO] - Step 666, Loss: {'train/loss': 0.2548459470272064, 'train/grad_norm_pre_clip': 38.13417053222656, 'timing/data': 0.0004168800078332424, 'timing/model': 0.5995798980002292, 'learning_rate/qwen_vl_interface': 9.690662240113669e-06, 'learning_rate/action_model': 9.690662240113668e-05, 'epoch': 0.05}) +[2026-10-03 19:03:16,076][starVLA.training.train_starvla][INFO] - Step 667, Loss: {'train/loss': 0.13275064527988434, 'train/grad_norm_pre_clip': 86.31316375732422, 'timing/data': 0.00022914999863132834, 'timing/model': 0.39570760799688287, 'learning_rate/qwen_vl_interface': 9.689580282764077e-06, 'learning_rate/action_model': 9.689580282764077e-05, 'epoch': 0.05}) +[2026-10-03 19:03:16,784][starVLA.training.train_starvla][INFO] - Step 668, Loss: {'train/loss': 0.015219666063785553, 'train/grad_norm_pre_clip': 1.7609928846359253, 'timing/data': 0.00031747997854836285, 'timing/model': 0.39603764898492955, 'learning_rate/qwen_vl_interface': 9.688496500470731e-06, 'learning_rate/action_model': 9.68849650047073e-05, 'epoch': 0.05}) +[2026-10-03 19:03:17,490][starVLA.training.train_starvla][INFO] - Step 669, Loss: {'train/loss': 0.36557134985923767, 'train/grad_norm_pre_clip': 77.50462341308594, 'timing/data': 0.00022807001369073987, 'timing/model': 0.3946555299917236, 'learning_rate/qwen_vl_interface': 9.68741089367913e-06, 'learning_rate/action_model': 9.687410893679129e-05, 'epoch': 0.05}) +[2026-10-03 19:03:18,171][starVLA.training.train_starvla][INFO] - Step 670, Loss: {'train/loss': 0.2304389327764511, 'train/grad_norm_pre_clip': 29.300247192382812, 'timing/data': 0.0003432600060477853, 'timing/model': 0.3940283979754895, 'learning_rate/qwen_vl_interface': 9.686323462835527e-06, 'learning_rate/action_model': 9.686323462835526e-05, 'epoch': 0.05}) +[2026-10-03 19:03:18,878][starVLA.training.train_starvla][INFO] - Step 671, Loss: {'train/loss': 0.013455109670758247, 'train/grad_norm_pre_clip': 7.0377583503723145, 'timing/data': 0.00022648999583907425, 'timing/model': 0.39581509900745004, 'learning_rate/qwen_vl_interface': 9.685234208386922e-06, 'learning_rate/action_model': 9.685234208386921e-05, 'epoch': 0.05}) +[2026-10-03 19:03:19,757][starVLA.training.train_starvla][INFO] - Step 672, Loss: {'train/loss': 0.41886642575263977, 'train/grad_norm_pre_clip': 22.904104232788086, 'timing/data': 0.0002304399968124926, 'timing/model': 0.3959979980136268, 'learning_rate/qwen_vl_interface': 9.684143130781066e-06, 'learning_rate/action_model': 9.684143130781066e-05, 'epoch': 0.05}) +[2026-10-03 19:03:20,463][starVLA.training.train_starvla][INFO] - Step 673, Loss: {'train/loss': 0.8820215463638306, 'train/grad_norm_pre_clip': 293.15087890625, 'timing/data': 0.0002893499913625419, 'timing/model': 0.39505720799206756, 'learning_rate/qwen_vl_interface': 9.683050230466462e-06, 'learning_rate/action_model': 9.683050230466462e-05, 'epoch': 0.05}) +[2026-10-03 19:03:21,169][starVLA.training.train_starvla][INFO] - Step 674, Loss: {'train/loss': 0.008862805552780628, 'train/grad_norm_pre_clip': 3.510432481765747, 'timing/data': 0.0002156300179194659, 'timing/model': 0.3932378889876418, 'learning_rate/qwen_vl_interface': 9.681955507892361e-06, 'learning_rate/action_model': 9.68195550789236e-05, 'epoch': 0.05}) +[2026-10-03 19:03:21,878][starVLA.training.train_starvla][INFO] - Step 675, Loss: {'train/loss': 0.3362341821193695, 'train/grad_norm_pre_clip': 41.36903381347656, 'timing/data': 0.00024900000425986946, 'timing/model': 0.3962095190072432, 'learning_rate/qwen_vl_interface': 9.680858963508759e-06, 'learning_rate/action_model': 9.68085896350876e-05, 'epoch': 0.05}) +[2026-10-03 19:03:22,589][starVLA.training.train_starvla][INFO] - Step 676, Loss: {'train/loss': 0.3994177579879761, 'train/grad_norm_pre_clip': 79.98482513427734, 'timing/data': 0.00038760999450460076, 'timing/model': 0.3963485199783463, 'learning_rate/qwen_vl_interface': 9.679760597766405e-06, 'learning_rate/action_model': 9.679760597766405e-05, 'epoch': 0.05}) +[2026-10-03 19:03:23,500][starVLA.training.train_starvla][INFO] - Step 677, Loss: {'train/loss': 0.7079706788063049, 'train/grad_norm_pre_clip': 52.85532760620117, 'timing/data': 0.00021789001766592264, 'timing/model': 0.5966876779857557, 'learning_rate/qwen_vl_interface': 9.678660411116796e-06, 'learning_rate/action_model': 9.678660411116796e-05, 'epoch': 0.05}) +[2026-10-03 19:03:24,207][starVLA.training.train_starvla][INFO] - Step 678, Loss: {'train/loss': 0.002325313864275813, 'train/grad_norm_pre_clip': 0.442914754152298, 'timing/data': 0.00036155999987386167, 'timing/model': 0.39590714900987223, 'learning_rate/qwen_vl_interface': 9.677558404012178e-06, 'learning_rate/action_model': 9.677558404012177e-05, 'epoch': 0.05}) +[2026-10-03 19:03:24,919][starVLA.training.train_starvla][INFO] - Step 679, Loss: {'train/loss': 0.5787758231163025, 'train/grad_norm_pre_clip': 124.6980209350586, 'timing/data': 0.00025772899971343577, 'timing/model': 0.3968996389885433, 'learning_rate/qwen_vl_interface': 9.676454576905542e-06, 'learning_rate/action_model': 9.676454576905542e-05, 'epoch': 0.05}) +[2026-10-03 19:03:25,629][starVLA.training.train_starvla][INFO] - Step 680, Loss: {'train/loss': 0.13112007081508636, 'train/grad_norm_pre_clip': 71.309814453125, 'timing/data': 0.0003655190230347216, 'timing/model': 0.3958233399898745, 'learning_rate/qwen_vl_interface': 9.675348930250633e-06, 'learning_rate/action_model': 9.675348930250632e-05, 'epoch': 0.05}) +[2026-10-03 19:03:26,337][starVLA.training.train_starvla][INFO] - Step 681, Loss: {'train/loss': 0.021211128681898117, 'train/grad_norm_pre_clip': 27.90300178527832, 'timing/data': 0.00035909999860450625, 'timing/model': 0.3962844990019221, 'learning_rate/qwen_vl_interface': 9.674241464501938e-06, 'learning_rate/action_model': 9.674241464501937e-05, 'epoch': 0.05}) +[2026-10-03 19:03:27,238][starVLA.training.train_starvla][INFO] - Step 682, Loss: {'train/loss': 0.764156699180603, 'train/grad_norm_pre_clip': 69.18930053710938, 'timing/data': 0.00022536999313160777, 'timing/model': 0.5902415979944635, 'learning_rate/qwen_vl_interface': 9.673132180114695e-06, 'learning_rate/action_model': 9.673132180114695e-05, 'epoch': 0.05}) +[2026-10-03 19:03:27,949][starVLA.training.train_starvla][INFO] - Step 683, Loss: {'train/loss': 0.5881171822547913, 'train/grad_norm_pre_clip': 73.61528015136719, 'timing/data': 0.00023322997731156647, 'timing/model': 0.3960917879885528, 'learning_rate/qwen_vl_interface': 9.67202107754489e-06, 'learning_rate/action_model': 9.672021077544889e-05, 'epoch': 0.05}) +[2026-10-03 19:03:28,657][starVLA.training.train_starvla][INFO] - Step 684, Loss: {'train/loss': 0.7112506031990051, 'train/grad_norm_pre_clip': 52.72645950317383, 'timing/data': 0.0004648100002668798, 'timing/model': 0.3956644680001773, 'learning_rate/qwen_vl_interface': 9.670908157249254e-06, 'learning_rate/action_model': 9.670908157249253e-05, 'epoch': 0.05}) +[2026-10-03 19:03:29,365][starVLA.training.train_starvla][INFO] - Step 685, Loss: {'train/loss': 0.02929351106286049, 'train/grad_norm_pre_clip': 19.449424743652344, 'timing/data': 0.00034590999712236226, 'timing/model': 0.39418479800224304, 'learning_rate/qwen_vl_interface': 9.669793419685267e-06, 'learning_rate/action_model': 9.669793419685266e-05, 'epoch': 0.05}) +[2026-10-03 19:03:30,070][starVLA.training.train_starvla][INFO] - Step 686, Loss: {'train/loss': 1.4088058471679688, 'train/grad_norm_pre_clip': 66.03141021728516, 'timing/data': 0.00023207999765872955, 'timing/model': 0.39473452800302766, 'learning_rate/qwen_vl_interface': 9.668676865311155e-06, 'learning_rate/action_model': 9.668676865311154e-05, 'epoch': 0.05}) +[2026-10-03 19:03:30,781][starVLA.training.train_starvla][INFO] - Step 687, Loss: {'train/loss': 0.6476086974143982, 'train/grad_norm_pre_clip': 43.91069412231445, 'timing/data': 0.0003384700103197247, 'timing/model': 0.39728586800629273, 'learning_rate/qwen_vl_interface': 9.667558494585891e-06, 'learning_rate/action_model': 9.66755849458589e-05, 'epoch': 0.05}) +[2026-10-03 19:03:31,685][starVLA.training.train_starvla][INFO] - Step 688, Loss: {'train/loss': 0.004090078640729189, 'train/grad_norm_pre_clip': 2.5060596466064453, 'timing/data': 0.00034841999877244234, 'timing/model': 0.39420032998896204, 'learning_rate/qwen_vl_interface': 9.666438307969194e-06, 'learning_rate/action_model': 9.666438307969195e-05, 'epoch': 0.05}) +[2026-10-03 19:03:32,393][starVLA.training.train_starvla][INFO] - Step 689, Loss: {'train/loss': 0.7402056455612183, 'train/grad_norm_pre_clip': 76.00520324707031, 'timing/data': 0.0003245699917897582, 'timing/model': 0.39548165901214816, 'learning_rate/qwen_vl_interface': 9.665316305921535e-06, 'learning_rate/action_model': 9.665316305921535e-05, 'epoch': 0.05}) +[2026-10-03 19:03:33,098][starVLA.training.train_starvla][INFO] - Step 690, Loss: {'train/loss': 0.11606577783823013, 'train/grad_norm_pre_clip': 104.44107055664062, 'timing/data': 0.0003368199977558106, 'timing/model': 0.39442133900593035, 'learning_rate/qwen_vl_interface': 9.664192488904123e-06, 'learning_rate/action_model': 9.664192488904122e-05, 'epoch': 0.05}) +[2026-10-03 19:03:33,810][starVLA.training.train_starvla][INFO] - Step 691, Loss: {'train/loss': 0.037003908306360245, 'train/grad_norm_pre_clip': 23.487226486206055, 'timing/data': 0.0002641800092533231, 'timing/model': 0.3981274390243925, 'learning_rate/qwen_vl_interface': 9.663066857378917e-06, 'learning_rate/action_model': 9.663066857378917e-05, 'epoch': 0.05}) +[2026-10-03 19:03:34,518][starVLA.training.train_starvla][INFO] - Step 692, Loss: {'train/loss': 1.2829689979553223, 'train/grad_norm_pre_clip': 91.72758483886719, 'timing/data': 0.0003583790094126016, 'timing/model': 0.39547643999685533, 'learning_rate/qwen_vl_interface': 9.661939411808623e-06, 'learning_rate/action_model': 9.661939411808622e-05, 'epoch': 0.05}) +[2026-10-03 19:03:35,422][starVLA.training.train_starvla][INFO] - Step 693, Loss: {'train/loss': 0.5374611616134644, 'train/grad_norm_pre_clip': 74.4062271118164, 'timing/data': 0.0003060100134462118, 'timing/model': 0.5896016389888246, 'learning_rate/qwen_vl_interface': 9.66081015265669e-06, 'learning_rate/action_model': 9.660810152656689e-05, 'epoch': 0.05}) +[2026-10-03 19:03:36,128][starVLA.training.train_starvla][INFO] - Step 694, Loss: {'train/loss': 0.04000420868396759, 'train/grad_norm_pre_clip': 10.791617393493652, 'timing/data': 0.00024850000045262277, 'timing/model': 0.39588523900602013, 'learning_rate/qwen_vl_interface': 9.659679080387314e-06, 'learning_rate/action_model': 9.659679080387313e-05, 'epoch': 0.05}) +[2026-10-03 19:03:36,838][starVLA.training.train_starvla][INFO] - Step 695, Loss: {'train/loss': 0.36770519614219666, 'train/grad_norm_pre_clip': 28.63796043395996, 'timing/data': 0.00027860002592206, 'timing/model': 0.3965308479964733, 'learning_rate/qwen_vl_interface': 9.658546195465436e-06, 'learning_rate/action_model': 9.658546195465435e-05, 'epoch': 0.05}) +[2026-10-03 19:03:37,548][starVLA.training.train_starvla][INFO] - Step 696, Loss: {'train/loss': 0.3749583065509796, 'train/grad_norm_pre_clip': 55.80979919433594, 'timing/data': 0.00035434000892564654, 'timing/model': 0.3964871479838621, 'learning_rate/qwen_vl_interface': 9.65741149835674e-06, 'learning_rate/action_model': 9.657411498356741e-05, 'epoch': 0.05}) +[2026-10-03 19:03:38,255][starVLA.training.train_starvla][INFO] - Step 697, Loss: {'train/loss': 0.9813299179077148, 'train/grad_norm_pre_clip': 46.6014518737793, 'timing/data': 0.0003461199812591076, 'timing/model': 0.39542687899665907, 'learning_rate/qwen_vl_interface': 9.656274989527663e-06, 'learning_rate/action_model': 9.656274989527664e-05, 'epoch': 0.05}) +[2026-10-03 19:03:39,159][starVLA.training.train_starvla][INFO] - Step 698, Loss: {'train/loss': 0.11478036642074585, 'train/grad_norm_pre_clip': 26.191495895385742, 'timing/data': 0.000338789977831766, 'timing/model': 0.590893649001373, 'learning_rate/qwen_vl_interface': 9.655136669445378e-06, 'learning_rate/action_model': 9.655136669445378e-05, 'epoch': 0.05}) +[2026-10-03 19:03:39,868][starVLA.training.train_starvla][INFO] - Step 699, Loss: {'train/loss': 0.6734904050827026, 'train/grad_norm_pre_clip': 95.78166961669922, 'timing/data': 0.00025857999571599066, 'timing/model': 0.3960161590075586, 'learning_rate/qwen_vl_interface': 9.653996538577802e-06, 'learning_rate/action_model': 9.653996538577802e-05, 'epoch': 0.05}) +[2026-10-03 19:03:40,576][starVLA.training.train_starvla][INFO] - Step 700, Loss: {'train/loss': 0.4236547648906708, 'train/grad_norm_pre_clip': 33.819007873535156, 'timing/data': 0.00044068999704904854, 'timing/model': 0.3943666699924506, 'learning_rate/qwen_vl_interface': 9.652854597393604e-06, 'learning_rate/action_model': 9.652854597393603e-05, 'epoch': 0.05}) +[2026-10-03 19:03:41,284][starVLA.training.train_starvla][INFO] - Step 701, Loss: {'train/loss': 0.12982459366321564, 'train/grad_norm_pre_clip': 45.875953674316406, 'timing/data': 0.0004376900033093989, 'timing/model': 0.3952678290079348, 'learning_rate/qwen_vl_interface': 9.651710846362192e-06, 'learning_rate/action_model': 9.651710846362192e-05, 'epoch': 0.05}) +[2026-10-03 19:03:41,990][starVLA.training.train_starvla][INFO] - Step 702, Loss: {'train/loss': 0.2623789310455322, 'train/grad_norm_pre_clip': 55.116886138916016, 'timing/data': 0.0003417000116314739, 'timing/model': 0.3954206589842215, 'learning_rate/qwen_vl_interface': 9.650565285953716e-06, 'learning_rate/action_model': 9.650565285953715e-05, 'epoch': 0.05}) +[2026-10-03 19:03:42,702][starVLA.training.train_starvla][INFO] - Step 703, Loss: {'train/loss': 0.11963116377592087, 'train/grad_norm_pre_clip': 13.139337539672852, 'timing/data': 0.00031470999238081276, 'timing/model': 0.39773938700091094, 'learning_rate/qwen_vl_interface': 9.649417916639077e-06, 'learning_rate/action_model': 9.649417916639076e-05, 'epoch': 0.05}) +[2026-10-03 19:03:43,611][starVLA.training.train_starvla][INFO] - Step 704, Loss: {'train/loss': 0.2137051522731781, 'train/grad_norm_pre_clip': 16.596660614013672, 'timing/data': 0.0004230600025039166, 'timing/model': 0.3954470279859379, 'learning_rate/qwen_vl_interface': 9.648268738889911e-06, 'learning_rate/action_model': 9.648268738889911e-05, 'epoch': 0.05}) +[2026-10-03 19:03:44,320][starVLA.training.train_starvla][INFO] - Step 705, Loss: {'train/loss': 0.11117011308670044, 'train/grad_norm_pre_clip': 15.350709915161133, 'timing/data': 0.00037065899232402444, 'timing/model': 0.39477212898782454, 'learning_rate/qwen_vl_interface': 9.647117753178604e-06, 'learning_rate/action_model': 9.647117753178604e-05, 'epoch': 0.05}) +[2026-10-03 19:03:45,028][starVLA.training.train_starvla][INFO] - Step 706, Loss: {'train/loss': 0.3516700267791748, 'train/grad_norm_pre_clip': 31.224990844726562, 'timing/data': 0.00028763001319020987, 'timing/model': 0.39559460899909027, 'learning_rate/qwen_vl_interface': 9.645964959978284e-06, 'learning_rate/action_model': 9.645964959978284e-05, 'epoch': 0.05}) +[2026-10-03 19:03:45,740][starVLA.training.train_starvla][INFO] - Step 707, Loss: {'train/loss': 0.3570791482925415, 'train/grad_norm_pre_clip': 51.15742874145508, 'timing/data': 0.00023617001716047525, 'timing/model': 0.3972253490064759, 'learning_rate/qwen_vl_interface': 9.644810359762818e-06, 'learning_rate/action_model': 9.644810359762817e-05, 'epoch': 0.05}) +[2026-10-03 19:03:46,448][starVLA.training.train_starvla][INFO] - Step 708, Loss: {'train/loss': 0.4664706885814667, 'train/grad_norm_pre_clip': 30.625593185424805, 'timing/data': 0.00038777998997829854, 'timing/model': 0.3943635699979495, 'learning_rate/qwen_vl_interface': 9.64365395300682e-06, 'learning_rate/action_model': 9.643653953006819e-05, 'epoch': 0.05}) +[2026-10-03 19:03:47,361][starVLA.training.train_starvla][INFO] - Step 709, Loss: {'train/loss': 0.397502601146698, 'train/grad_norm_pre_clip': 30.343868255615234, 'timing/data': 0.0004122199898120016, 'timing/model': 0.395418547996087, 'learning_rate/qwen_vl_interface': 9.642495740185644e-06, 'learning_rate/action_model': 9.642495740185643e-05, 'epoch': 0.05}) +[2026-10-03 19:03:48,068][starVLA.training.train_starvla][INFO] - Step 710, Loss: {'train/loss': 0.04088149219751358, 'train/grad_norm_pre_clip': 5.6215128898620605, 'timing/data': 0.00020994999795220792, 'timing/model': 0.39559738797834143, 'learning_rate/qwen_vl_interface': 9.641335721775388e-06, 'learning_rate/action_model': 9.641335721775387e-05, 'epoch': 0.05}) +[2026-10-03 19:03:48,780][starVLA.training.train_starvla][INFO] - Step 711, Loss: {'train/loss': 0.3899814784526825, 'train/grad_norm_pre_clip': 67.89554595947266, 'timing/data': 0.00031601000227965415, 'timing/model': 0.397178287996212, 'learning_rate/qwen_vl_interface': 9.640173898252893e-06, 'learning_rate/action_model': 9.640173898252892e-05, 'epoch': 0.05}) +[2026-10-03 19:03:49,486][starVLA.training.train_starvla][INFO] - Step 712, Loss: {'train/loss': 0.030415454879403114, 'train/grad_norm_pre_clip': 7.11760950088501, 'timing/data': 0.00034363000304438174, 'timing/model': 0.3948294100118801, 'learning_rate/qwen_vl_interface': 9.639010270095738e-06, 'learning_rate/action_model': 9.639010270095739e-05, 'epoch': 0.05}) +[2026-10-03 19:03:50,195][starVLA.training.train_starvla][INFO] - Step 713, Loss: {'train/loss': 0.5708109140396118, 'train/grad_norm_pre_clip': 80.85869598388672, 'timing/data': 0.0003309900057502091, 'timing/model': 0.39550739899277687, 'learning_rate/qwen_vl_interface': 9.637844837782251e-06, 'learning_rate/action_model': 9.63784483778225e-05, 'epoch': 0.05}) +[2026-10-03 19:03:51,099][starVLA.training.train_starvla][INFO] - Step 714, Loss: {'train/loss': 0.4834427237510681, 'train/grad_norm_pre_clip': 28.76483917236328, 'timing/data': 0.00021485000615939498, 'timing/model': 0.5927672580000944, 'learning_rate/qwen_vl_interface': 9.636677601791494e-06, 'learning_rate/action_model': 9.636677601791493e-05, 'epoch': 0.05}) +[2026-10-03 19:03:51,809][starVLA.training.train_starvla][INFO] - Step 715, Loss: {'train/loss': 0.0692094936966896, 'train/grad_norm_pre_clip': 52.314788818359375, 'timing/data': 0.0003420500142965466, 'timing/model': 0.39758256799541414, 'learning_rate/qwen_vl_interface': 9.635508562603273e-06, 'learning_rate/action_model': 9.635508562603272e-05, 'epoch': 0.05}) +[2026-10-03 19:03:52,518][starVLA.training.train_starvla][INFO] - Step 716, Loss: {'train/loss': 0.0849691778421402, 'train/grad_norm_pre_clip': 54.57686233520508, 'timing/data': 0.0003480500017758459, 'timing/model': 0.3962020690087229, 'learning_rate/qwen_vl_interface': 9.634337720698139e-06, 'learning_rate/action_model': 9.634337720698137e-05, 'epoch': 0.05}) +[2026-10-03 19:03:53,226][starVLA.training.train_starvla][INFO] - Step 717, Loss: {'train/loss': 0.39640289545059204, 'train/grad_norm_pre_clip': 87.67288208007812, 'timing/data': 0.00021824997384101152, 'timing/model': 0.39632656902540475, 'learning_rate/qwen_vl_interface': 9.633165076557376e-06, 'learning_rate/action_model': 9.633165076557375e-05, 'epoch': 0.05}) +[2026-10-03 19:03:53,930][starVLA.training.train_starvla][INFO] - Step 718, Loss: {'train/loss': 0.022545522078871727, 'train/grad_norm_pre_clip': 11.72221565246582, 'timing/data': 0.00025666001602075994, 'timing/model': 0.39364655897952616, 'learning_rate/qwen_vl_interface': 9.631990630663019e-06, 'learning_rate/action_model': 9.631990630663018e-05, 'epoch': 0.05}) +[2026-10-03 19:03:54,641][starVLA.training.train_starvla][INFO] - Step 719, Loss: {'train/loss': 0.03560733422636986, 'train/grad_norm_pre_clip': 16.50608253479004, 'timing/data': 0.00031534998561255634, 'timing/model': 0.3960855290060863, 'learning_rate/qwen_vl_interface': 9.630814383497836e-06, 'learning_rate/action_model': 9.630814383497835e-05, 'epoch': 0.05}) +[2026-10-03 19:03:55,547][starVLA.training.train_starvla][INFO] - Step 720, Loss: {'train/loss': 0.38345202803611755, 'train/grad_norm_pre_clip': 48.91144943237305, 'timing/data': 0.00036837998777627945, 'timing/model': 0.39586707900161855, 'learning_rate/qwen_vl_interface': 9.629636335545336e-06, 'learning_rate/action_model': 9.629636335545335e-05, 'epoch': 0.05}) +[2026-10-03 19:03:56,255][starVLA.training.train_starvla][INFO] - Step 721, Loss: {'train/loss': 0.03111645206809044, 'train/grad_norm_pre_clip': 21.729408264160156, 'timing/data': 0.00024120000307448208, 'timing/model': 0.3952115490101278, 'learning_rate/qwen_vl_interface': 9.628456487289773e-06, 'learning_rate/action_model': 9.628456487289773e-05, 'epoch': 0.05}) +[2026-10-03 19:03:56,960][starVLA.training.train_starvla][INFO] - Step 722, Loss: {'train/loss': 0.07582908868789673, 'train/grad_norm_pre_clip': 53.77439880371094, 'timing/data': 0.00019329998758621514, 'timing/model': 0.39450552000198513, 'learning_rate/qwen_vl_interface': 9.627274839216135e-06, 'learning_rate/action_model': 9.627274839216136e-05, 'epoch': 0.05}) +[2026-10-03 19:03:57,668][starVLA.training.train_starvla][INFO] - Step 723, Loss: {'train/loss': 0.5316300988197327, 'train/grad_norm_pre_clip': 123.37461853027344, 'timing/data': 0.0002438799710944295, 'timing/model': 0.39656983900931664, 'learning_rate/qwen_vl_interface': 9.626091391810158e-06, 'learning_rate/action_model': 9.626091391810157e-05, 'epoch': 0.05}) +[2026-10-03 19:03:58,374][starVLA.training.train_starvla][INFO] - Step 724, Loss: {'train/loss': 0.814071774482727, 'train/grad_norm_pre_clip': 140.3119354248047, 'timing/data': 0.00029435998294502497, 'timing/model': 0.3950634280045051, 'learning_rate/qwen_vl_interface': 9.624906145558309e-06, 'learning_rate/action_model': 9.624906145558307e-05, 'epoch': 0.05}) +[2026-10-03 19:03:59,271][starVLA.training.train_starvla][INFO] - Step 725, Loss: {'train/loss': 1.2783687114715576, 'train/grad_norm_pre_clip': 182.51522827148438, 'timing/data': 0.0003739300009328872, 'timing/model': 0.39518074999796227, 'learning_rate/qwen_vl_interface': 9.623719100947795e-06, 'learning_rate/action_model': 9.623719100947795e-05, 'epoch': 0.05}) +[2026-10-03 19:03:59,979][starVLA.training.train_starvla][INFO] - Step 726, Loss: {'train/loss': 0.019260970875620842, 'train/grad_norm_pre_clip': 3.8375730514526367, 'timing/data': 0.0003152000135742128, 'timing/model': 0.3955972279945854, 'learning_rate/qwen_vl_interface': 9.622530258466571e-06, 'learning_rate/action_model': 9.62253025846657e-05, 'epoch': 0.05}) +[2026-10-03 19:04:00,689][starVLA.training.train_starvla][INFO] - Step 727, Loss: {'train/loss': 0.3577040135860443, 'train/grad_norm_pre_clip': 66.44133758544922, 'timing/data': 0.00030747000710107386, 'timing/model': 0.3961126990034245, 'learning_rate/qwen_vl_interface': 9.62133961860332e-06, 'learning_rate/action_model': 9.62133961860332e-05, 'epoch': 0.05}) +[2026-10-03 19:04:01,396][starVLA.training.train_starvla][INFO] - Step 728, Loss: {'train/loss': 0.4769497811794281, 'train/grad_norm_pre_clip': 54.21945571899414, 'timing/data': 0.0002968100016005337, 'timing/model': 0.3961443880107254, 'learning_rate/qwen_vl_interface': 9.620147181847474e-06, 'learning_rate/action_model': 9.620147181847475e-05, 'epoch': 0.05}) +[2026-10-03 19:04:02,107][starVLA.training.train_starvla][INFO] - Step 729, Loss: {'train/loss': 0.7598781585693359, 'train/grad_norm_pre_clip': 97.8541259765625, 'timing/data': 0.000264620000962168, 'timing/model': 0.3963144489971455, 'learning_rate/qwen_vl_interface': 9.618952948689196e-06, 'learning_rate/action_model': 9.618952948689196e-05, 'epoch': 0.05}) +[2026-10-03 19:04:03,009][starVLA.training.train_starvla][INFO] - Step 730, Loss: {'train/loss': 0.7311180233955383, 'train/grad_norm_pre_clip': 49.152347564697266, 'timing/data': 0.00033209999673999846, 'timing/model': 0.589631708979141, 'learning_rate/qwen_vl_interface': 9.617756919619393e-06, 'learning_rate/action_model': 9.617756919619392e-05, 'epoch': 0.05}) +[2026-10-03 19:04:03,718][starVLA.training.train_starvla][INFO] - Step 731, Loss: {'train/loss': 0.30827808380126953, 'train/grad_norm_pre_clip': 29.30130958557129, 'timing/data': 0.0003164699883200228, 'timing/model': 0.3954124089796096, 'learning_rate/qwen_vl_interface': 9.616559095129701e-06, 'learning_rate/action_model': 9.6165590951297e-05, 'epoch': 0.05}) +[2026-10-03 19:04:04,428][starVLA.training.train_starvla][INFO] - Step 732, Loss: {'train/loss': 0.31798622012138367, 'train/grad_norm_pre_clip': 57.87446594238281, 'timing/data': 0.00026748000527732074, 'timing/model': 0.39599067901144736, 'learning_rate/qwen_vl_interface': 9.615359475712506e-06, 'learning_rate/action_model': 9.615359475712506e-05, 'epoch': 0.05}) +[2026-10-03 19:04:05,136][starVLA.training.train_starvla][INFO] - Step 733, Loss: {'train/loss': 1.2107853889465332, 'train/grad_norm_pre_clip': 119.93914794921875, 'timing/data': 0.00022159999934956431, 'timing/model': 0.39549768800497986, 'learning_rate/qwen_vl_interface': 9.614158061860924e-06, 'learning_rate/action_model': 9.614158061860924e-05, 'epoch': 0.05}) +[2026-10-03 19:04:05,842][starVLA.training.train_starvla][INFO] - Step 734, Loss: {'train/loss': 0.10572123527526855, 'train/grad_norm_pre_clip': 26.682022094726562, 'timing/data': 0.0002597199927549809, 'timing/model': 0.3946199890051503, 'learning_rate/qwen_vl_interface': 9.612954854068813e-06, 'learning_rate/action_model': 9.612954854068812e-05, 'epoch': 0.05}) +[2026-10-03 19:04:06,749][starVLA.training.train_starvla][INFO] - Step 735, Loss: {'train/loss': 0.18008437752723694, 'train/grad_norm_pre_clip': 60.817955017089844, 'timing/data': 0.0003170600102748722, 'timing/model': 0.5948863980011083, 'learning_rate/qwen_vl_interface': 9.611749852830763e-06, 'learning_rate/action_model': 9.611749852830763e-05, 'epoch': 0.05}) +[2026-10-03 19:04:07,454][starVLA.training.train_starvla][INFO] - Step 736, Loss: {'train/loss': 0.07335267961025238, 'train/grad_norm_pre_clip': 36.9049186706543, 'timing/data': 0.0003661299997474998, 'timing/model': 0.39374347002012655, 'learning_rate/qwen_vl_interface': 9.610543058642107e-06, 'learning_rate/action_model': 9.610543058642107e-05, 'epoch': 0.05}) +[2026-10-03 19:04:08,157][starVLA.training.train_starvla][INFO] - Step 737, Loss: {'train/loss': 0.2479625940322876, 'train/grad_norm_pre_clip': 52.54216766357422, 'timing/data': 0.000252589990850538, 'timing/model': 0.39302746899193153, 'learning_rate/qwen_vl_interface': 9.609334471998911e-06, 'learning_rate/action_model': 9.609334471998912e-05, 'epoch': 0.05}) +[2026-10-03 19:04:08,860][starVLA.training.train_starvla][INFO] - Step 738, Loss: {'train/loss': 0.12982505559921265, 'train/grad_norm_pre_clip': 64.98128509521484, 'timing/data': 0.00031392998062074184, 'timing/model': 0.39269100801902823, 'learning_rate/qwen_vl_interface': 9.60812409339798e-06, 'learning_rate/action_model': 9.60812409339798e-05, 'epoch': 0.05}) +[2026-10-03 19:04:09,567][starVLA.training.train_starvla][INFO] - Step 739, Loss: {'train/loss': 0.5806539058685303, 'train/grad_norm_pre_clip': 67.33763122558594, 'timing/data': 0.0003141899942420423, 'timing/model': 0.3945886489818804, 'learning_rate/qwen_vl_interface': 9.606911923336856e-06, 'learning_rate/action_model': 9.606911923336856e-05, 'epoch': 0.05}) +[2026-10-03 19:04:10,483][starVLA.training.train_starvla][INFO] - Step 740, Loss: {'train/loss': 0.6045004725456238, 'train/grad_norm_pre_clip': 66.40941619873047, 'timing/data': 0.00037603999953716993, 'timing/model': 0.6048113069846295, 'learning_rate/qwen_vl_interface': 9.605697962313817e-06, 'learning_rate/action_model': 9.605697962313816e-05, 'epoch': 0.05}) +[2026-10-03 19:04:11,189][starVLA.training.train_starvla][INFO] - Step 741, Loss: {'train/loss': 0.20208747684955597, 'train/grad_norm_pre_clip': 62.46392822265625, 'timing/data': 0.0002988699998240918, 'timing/model': 0.39558402900001965, 'learning_rate/qwen_vl_interface': 9.604482210827871e-06, 'learning_rate/action_model': 9.604482210827871e-05, 'epoch': 0.05}) +[2026-10-03 19:04:11,893][starVLA.training.train_starvla][INFO] - Step 742, Loss: {'train/loss': 0.26485171914100647, 'train/grad_norm_pre_clip': 87.4035873413086, 'timing/data': 0.00022168998839333653, 'timing/model': 0.3937360389973037, 'learning_rate/qwen_vl_interface': 9.603264669378772e-06, 'learning_rate/action_model': 9.603264669378772e-05, 'epoch': 0.05}) +[2026-10-03 19:04:12,603][starVLA.training.train_starvla][INFO] - Step 743, Loss: {'train/loss': 0.19038113951683044, 'train/grad_norm_pre_clip': 24.618467330932617, 'timing/data': 0.0003139199980068952, 'timing/model': 0.3964718290080782, 'learning_rate/qwen_vl_interface': 9.602045338467007e-06, 'learning_rate/action_model': 9.602045338467006e-05, 'epoch': 0.05}) +[2026-10-03 19:04:13,310][starVLA.training.train_starvla][INFO] - Step 744, Loss: {'train/loss': 0.13850438594818115, 'train/grad_norm_pre_clip': 20.798810958862305, 'timing/data': 0.00034828000934794545, 'timing/model': 0.3951032890181523, 'learning_rate/qwen_vl_interface': 9.600824218593791e-06, 'learning_rate/action_model': 9.600824218593791e-05, 'epoch': 0.05}) +[2026-10-03 19:04:14,017][starVLA.training.train_starvla][INFO] - Step 745, Loss: {'train/loss': 0.4079051613807678, 'train/grad_norm_pre_clip': 84.7406005859375, 'timing/data': 0.00030957997660152614, 'timing/model': 0.39492479999898933, 'learning_rate/qwen_vl_interface': 9.599601310261087e-06, 'learning_rate/action_model': 9.599601310261085e-05, 'epoch': 0.05}) +[2026-10-03 19:04:14,934][starVLA.training.train_starvla][INFO] - Step 746, Loss: {'train/loss': 0.47004812955856323, 'train/grad_norm_pre_clip': 101.89254760742188, 'timing/data': 0.0003754700010176748, 'timing/model': 0.3958922990132123, 'learning_rate/qwen_vl_interface': 9.598376613971582e-06, 'learning_rate/action_model': 9.598376613971582e-05, 'epoch': 0.05}) +[2026-10-03 19:04:15,643][starVLA.training.train_starvla][INFO] - Step 747, Loss: {'train/loss': 0.15175104141235352, 'train/grad_norm_pre_clip': 53.36052703857422, 'timing/data': 0.0003128690004814416, 'timing/model': 0.39508831000421196, 'learning_rate/qwen_vl_interface': 9.597150130228703e-06, 'learning_rate/action_model': 9.597150130228703e-05, 'epoch': 0.05}) +[2026-10-03 19:04:16,349][starVLA.training.train_starvla][INFO] - Step 748, Loss: {'train/loss': 0.3965352475643158, 'train/grad_norm_pre_clip': 79.88948059082031, 'timing/data': 0.00027474999660626054, 'timing/model': 0.39552629899117164, 'learning_rate/qwen_vl_interface': 9.595921859536612e-06, 'learning_rate/action_model': 9.595921859536611e-05, 'epoch': 0.05}) +[2026-10-03 19:04:17,056][starVLA.training.train_starvla][INFO] - Step 749, Loss: {'train/loss': 0.42532646656036377, 'train/grad_norm_pre_clip': 68.31582641601562, 'timing/data': 0.00035878000198863447, 'timing/model': 0.3946238190110307, 'learning_rate/qwen_vl_interface': 9.594691802400203e-06, 'learning_rate/action_model': 9.594691802400202e-05, 'epoch': 0.05}) +[2026-10-03 19:04:17,762][starVLA.training.train_starvla][INFO] - Step 750, Loss: {'train/loss': 0.5850069522857666, 'train/grad_norm_pre_clip': 36.31503677368164, 'timing/data': 0.00022617998183704913, 'timing/model': 0.3946904100012034, 'learning_rate/qwen_vl_interface': 9.59345995932511e-06, 'learning_rate/action_model': 9.59345995932511e-05, 'epoch': 0.05}) +[2026-10-03 19:04:18,675][starVLA.training.train_starvla][INFO] - Step 751, Loss: {'train/loss': 0.10639092326164246, 'train/grad_norm_pre_clip': 17.773448944091797, 'timing/data': 0.0003133089921902865, 'timing/model': 0.5985278689768165, 'learning_rate/qwen_vl_interface': 9.592226330817693e-06, 'learning_rate/action_model': 9.592226330817693e-05, 'epoch': 0.05}) +[2026-10-03 19:04:19,381][starVLA.training.train_starvla][INFO] - Step 752, Loss: {'train/loss': 0.3802637457847595, 'train/grad_norm_pre_clip': 78.57366943359375, 'timing/data': 0.00031091002165339887, 'timing/model': 0.3952712990285363, 'learning_rate/qwen_vl_interface': 9.590990917385052e-06, 'learning_rate/action_model': 9.590990917385052e-05, 'epoch': 0.06}) +[2026-10-03 19:04:20,085][starVLA.training.train_starvla][INFO] - Step 753, Loss: {'train/loss': 0.7303856611251831, 'train/grad_norm_pre_clip': 97.52824401855469, 'timing/data': 0.00030968000646680593, 'timing/model': 0.39283171901479363, 'learning_rate/qwen_vl_interface': 9.58975371953502e-06, 'learning_rate/action_model': 9.589753719535018e-05, 'epoch': 0.06}) +[2026-10-03 19:04:20,789][starVLA.training.train_starvla][INFO] - Step 754, Loss: {'train/loss': 0.07754214107990265, 'train/grad_norm_pre_clip': 15.798175811767578, 'timing/data': 0.0003717399958986789, 'timing/model': 0.3926780099864118, 'learning_rate/qwen_vl_interface': 9.588514737776158e-06, 'learning_rate/action_model': 9.588514737776157e-05, 'epoch': 0.06}) +[2026-10-03 19:04:21,497][starVLA.training.train_starvla][INFO] - Step 755, Loss: {'train/loss': 0.5193129777908325, 'train/grad_norm_pre_clip': 30.119874954223633, 'timing/data': 0.0002838500076904893, 'timing/model': 0.39586845799931325, 'learning_rate/qwen_vl_interface': 9.587273972617769e-06, 'learning_rate/action_model': 9.587273972617768e-05, 'epoch': 0.06}) +[2026-10-03 19:04:22,202][starVLA.training.train_starvla][INFO] - Step 756, Loss: {'train/loss': 0.09442275762557983, 'train/grad_norm_pre_clip': 33.04471206665039, 'timing/data': 0.00026725998031906784, 'timing/model': 0.3950452800199855, 'learning_rate/qwen_vl_interface': 9.586031424569882e-06, 'learning_rate/action_model': 9.58603142456988e-05, 'epoch': 0.06}) +[2026-10-03 19:04:23,106][starVLA.training.train_starvla][INFO] - Step 757, Loss: {'train/loss': 0.09311061352491379, 'train/grad_norm_pre_clip': 16.8930606842041, 'timing/data': 0.00034903999767266214, 'timing/model': 0.3929529390006792, 'learning_rate/qwen_vl_interface': 9.584787094143265e-06, 'learning_rate/action_model': 9.584787094143264e-05, 'epoch': 0.06}) +[2026-10-03 19:04:23,809][starVLA.training.train_starvla][INFO] - Step 758, Loss: {'train/loss': 0.0666751116514206, 'train/grad_norm_pre_clip': 13.089202880859375, 'timing/data': 0.00036472000647336245, 'timing/model': 0.3942701189953368, 'learning_rate/qwen_vl_interface': 9.58354098184941e-06, 'learning_rate/action_model': 9.583540981849409e-05, 'epoch': 0.06}) +[2026-10-03 19:04:24,517][starVLA.training.train_starvla][INFO] - Step 759, Loss: {'train/loss': 0.15356585383415222, 'train/grad_norm_pre_clip': 5.564664363861084, 'timing/data': 0.00030918000265955925, 'timing/model': 0.3957927390001714, 'learning_rate/qwen_vl_interface': 9.582293088200549e-06, 'learning_rate/action_model': 9.582293088200549e-05, 'epoch': 0.06}) +[2026-10-03 19:04:25,223][starVLA.training.train_starvla][INFO] - Step 760, Loss: {'train/loss': 0.27428096532821655, 'train/grad_norm_pre_clip': 36.19474792480469, 'timing/data': 0.0003591190034057945, 'timing/model': 0.39325179997831583, 'learning_rate/qwen_vl_interface': 9.581043413709646e-06, 'learning_rate/action_model': 9.581043413709646e-05, 'epoch': 0.06}) +[2026-10-03 19:04:25,928][starVLA.training.train_starvla][INFO] - Step 761, Loss: {'train/loss': 0.40642082691192627, 'train/grad_norm_pre_clip': 32.57697677612305, 'timing/data': 0.00027166001382283866, 'timing/model': 0.39374204899650067, 'learning_rate/qwen_vl_interface': 9.579791958890393e-06, 'learning_rate/action_model': 9.579791958890394e-05, 'epoch': 0.06}) +[2026-10-03 19:04:26,838][starVLA.training.train_starvla][INFO] - Step 762, Loss: {'train/loss': 0.23939359188079834, 'train/grad_norm_pre_clip': 59.79725646972656, 'timing/data': 0.00023040999076329172, 'timing/model': 0.3945477989909705, 'learning_rate/qwen_vl_interface': 9.578538724257218e-06, 'learning_rate/action_model': 9.578538724257216e-05, 'epoch': 0.06}) +[2026-10-03 19:04:27,546][starVLA.training.train_starvla][INFO] - Step 763, Loss: {'train/loss': 0.030176391825079918, 'train/grad_norm_pre_clip': 4.542939186096191, 'timing/data': 0.0002830589946825057, 'timing/model': 0.3948240799945779, 'learning_rate/qwen_vl_interface': 9.577283710325274e-06, 'learning_rate/action_model': 9.577283710325274e-05, 'epoch': 0.06}) +[2026-10-03 19:04:28,250][starVLA.training.train_starvla][INFO] - Step 764, Loss: {'train/loss': 0.014762095175683498, 'train/grad_norm_pre_clip': 3.2199957370758057, 'timing/data': 0.00021368000307120383, 'timing/model': 0.3927032589854207, 'learning_rate/qwen_vl_interface': 9.576026917610455e-06, 'learning_rate/action_model': 9.576026917610455e-05, 'epoch': 0.06}) +[2026-10-03 19:04:28,955][starVLA.training.train_starvla][INFO] - Step 765, Loss: {'train/loss': 0.006203384138643742, 'train/grad_norm_pre_clip': 0.5882250666618347, 'timing/data': 0.00022447999799624085, 'timing/model': 0.3930384800187312, 'learning_rate/qwen_vl_interface': 9.574768346629376e-06, 'learning_rate/action_model': 9.574768346629376e-05, 'epoch': 0.06}) +[2026-10-03 19:04:29,659][starVLA.training.train_starvla][INFO] - Step 766, Loss: {'train/loss': 0.3134464919567108, 'train/grad_norm_pre_clip': 108.23165130615234, 'timing/data': 0.0002676699950825423, 'timing/model': 0.39344124001218006, 'learning_rate/qwen_vl_interface': 9.573507997899391e-06, 'learning_rate/action_model': 9.573507997899392e-05, 'epoch': 0.06}) +[2026-10-03 19:04:30,572][starVLA.training.train_starvla][INFO] - Step 767, Loss: {'train/loss': 0.6152026057243347, 'train/grad_norm_pre_clip': 87.86006927490234, 'timing/data': 0.00021841999841853976, 'timing/model': 0.6007866480213124, 'learning_rate/qwen_vl_interface': 9.572245871938582e-06, 'learning_rate/action_model': 9.572245871938582e-05, 'epoch': 0.06}) +[2026-10-03 19:04:31,277][starVLA.training.train_starvla][INFO] - Step 768, Loss: {'train/loss': 0.3634048104286194, 'train/grad_norm_pre_clip': 56.03968811035156, 'timing/data': 0.00022792999516241252, 'timing/model': 0.39350415900116786, 'learning_rate/qwen_vl_interface': 9.570981969265761e-06, 'learning_rate/action_model': 9.570981969265761e-05, 'epoch': 0.06}) +[2026-10-03 19:04:31,981][starVLA.training.train_starvla][INFO] - Step 769, Loss: {'train/loss': 0.4884806275367737, 'train/grad_norm_pre_clip': 17.22901153564453, 'timing/data': 0.00021204998483881354, 'timing/model': 0.3945355890027713, 'learning_rate/qwen_vl_interface': 9.56971629040047e-06, 'learning_rate/action_model': 9.56971629040047e-05, 'epoch': 0.06}) +[2026-10-03 19:04:32,685][starVLA.training.train_starvla][INFO] - Step 770, Loss: {'train/loss': 1.2812138795852661, 'train/grad_norm_pre_clip': 62.70072937011719, 'timing/data': 0.00033109000651165843, 'timing/model': 0.3942113090015482, 'learning_rate/qwen_vl_interface': 9.56844883586298e-06, 'learning_rate/action_model': 9.568448835862981e-05, 'epoch': 0.06}) +[2026-10-03 19:04:33,392][starVLA.training.train_starvla][INFO] - Step 771, Loss: {'train/loss': 0.05971125140786171, 'train/grad_norm_pre_clip': 56.17900848388672, 'timing/data': 0.00032540998654440045, 'timing/model': 0.3965974989987444, 'learning_rate/qwen_vl_interface': 9.567179606174299e-06, 'learning_rate/action_model': 9.567179606174299e-05, 'epoch': 0.06}) +[2026-10-03 19:04:34,100][starVLA.training.train_starvla][INFO] - Step 772, Loss: {'train/loss': 0.26088881492614746, 'train/grad_norm_pre_clip': 110.05168914794922, 'timing/data': 0.00037883000914007425, 'timing/model': 0.3953986689739395, 'learning_rate/qwen_vl_interface': 9.565908601856155e-06, 'learning_rate/action_model': 9.565908601856154e-05, 'epoch': 0.06}) +[2026-10-03 19:04:35,006][starVLA.training.train_starvla][INFO] - Step 773, Loss: {'train/loss': 0.48680204153060913, 'train/grad_norm_pre_clip': 187.97201538085938, 'timing/data': 0.0003700300003401935, 'timing/model': 0.39283172000432387, 'learning_rate/qwen_vl_interface': 9.56463582343101e-06, 'learning_rate/action_model': 9.56463582343101e-05, 'epoch': 0.06}) +[2026-10-03 19:04:35,708][starVLA.training.train_starvla][INFO] - Step 774, Loss: {'train/loss': 1.4325584173202515, 'train/grad_norm_pre_clip': 58.85686111450195, 'timing/data': 0.0003248089924454689, 'timing/model': 0.39389182897866704, 'learning_rate/qwen_vl_interface': 9.563361271422056e-06, 'learning_rate/action_model': 9.563361271422056e-05, 'epoch': 0.06}) +[2026-10-03 19:04:36,415][starVLA.training.train_starvla][INFO] - Step 775, Loss: {'train/loss': 0.7670959234237671, 'train/grad_norm_pre_clip': 59.372276306152344, 'timing/data': 0.0003689400036819279, 'timing/model': 0.39523674998781644, 'learning_rate/qwen_vl_interface': 9.562084946353212e-06, 'learning_rate/action_model': 9.562084946353213e-05, 'epoch': 0.06}) +[2026-10-03 19:04:37,119][starVLA.training.train_starvla][INFO] - Step 776, Loss: {'train/loss': 0.7614457011222839, 'train/grad_norm_pre_clip': 49.298606872558594, 'timing/data': 0.00035332998959347606, 'timing/model': 0.39481579800485633, 'learning_rate/qwen_vl_interface': 9.560806848749132e-06, 'learning_rate/action_model': 9.560806848749132e-05, 'epoch': 0.06}) +[2026-10-03 19:04:37,824][starVLA.training.train_starvla][INFO] - Step 777, Loss: {'train/loss': 0.9349851608276367, 'train/grad_norm_pre_clip': 203.3649139404297, 'timing/data': 0.0003733899793587625, 'timing/model': 0.3925523989892099, 'learning_rate/qwen_vl_interface': 9.559526979135186e-06, 'learning_rate/action_model': 9.559526979135187e-05, 'epoch': 0.06}) +[2026-10-03 19:04:38,730][starVLA.training.train_starvla][INFO] - Step 778, Loss: {'train/loss': 0.9190964698791504, 'train/grad_norm_pre_clip': 67.7879638671875, 'timing/data': 0.00034740002593025565, 'timing/model': 0.3939556580153294, 'learning_rate/qwen_vl_interface': 9.558245338037486e-06, 'learning_rate/action_model': 9.558245338037486e-05, 'epoch': 0.06}) +[2026-10-03 19:04:39,439][starVLA.training.train_starvla][INFO] - Step 779, Loss: {'train/loss': 0.5453459620475769, 'train/grad_norm_pre_clip': 66.58192443847656, 'timing/data': 0.0002636600111145526, 'timing/model': 0.39554315799614415, 'learning_rate/qwen_vl_interface': 9.556961925982864e-06, 'learning_rate/action_model': 9.556961925982863e-05, 'epoch': 0.06}) +[2026-10-03 19:04:40,144][starVLA.training.train_starvla][INFO] - Step 780, Loss: {'train/loss': 0.2332042157649994, 'train/grad_norm_pre_clip': 86.92548370361328, 'timing/data': 0.0003567300154827535, 'timing/model': 0.39363498997408897, 'learning_rate/qwen_vl_interface': 9.555676743498882e-06, 'learning_rate/action_model': 9.555676743498881e-05, 'epoch': 0.06}) +[2026-10-03 19:04:40,848][starVLA.training.train_starvla][INFO] - Step 781, Loss: {'train/loss': 0.26430901885032654, 'train/grad_norm_pre_clip': 42.90449905395508, 'timing/data': 0.0002586300252005458, 'timing/model': 0.3936487100145314, 'learning_rate/qwen_vl_interface': 9.55438979111383e-06, 'learning_rate/action_model': 9.554389791113831e-05, 'epoch': 0.06}) +[2026-10-03 19:04:41,554][starVLA.training.train_starvla][INFO] - Step 782, Loss: {'train/loss': 0.30281946063041687, 'train/grad_norm_pre_clip': 17.970842361450195, 'timing/data': 0.00021341998944990337, 'timing/model': 0.3945285289955791, 'learning_rate/qwen_vl_interface': 9.553101069356728e-06, 'learning_rate/action_model': 9.553101069356727e-05, 'epoch': 0.06}) +[2026-10-03 19:04:42,458][starVLA.training.train_starvla][INFO] - Step 783, Loss: {'train/loss': 0.02501550316810608, 'train/grad_norm_pre_clip': 4.842860221862793, 'timing/data': 0.00030829900060780346, 'timing/model': 0.5930722799967043, 'learning_rate/qwen_vl_interface': 9.551810578757316e-06, 'learning_rate/action_model': 9.551810578757315e-05, 'epoch': 0.06}) +[2026-10-03 19:04:43,163][starVLA.training.train_starvla][INFO] - Step 784, Loss: {'train/loss': 0.31115803122520447, 'train/grad_norm_pre_clip': 28.023853302001953, 'timing/data': 0.00026569998590275645, 'timing/model': 0.39450603901059367, 'learning_rate/qwen_vl_interface': 9.550518319846069e-06, 'learning_rate/action_model': 9.550518319846069e-05, 'epoch': 0.06}) +[2026-10-03 19:04:43,867][starVLA.training.train_starvla][INFO] - Step 785, Loss: {'train/loss': 0.09240299463272095, 'train/grad_norm_pre_clip': 67.99842071533203, 'timing/data': 0.0002475100045558065, 'timing/model': 0.3925671989854891, 'learning_rate/qwen_vl_interface': 9.549224293154187e-06, 'learning_rate/action_model': 9.549224293154187e-05, 'epoch': 0.06}) +[2026-10-03 19:04:44,571][starVLA.training.train_starvla][INFO] - Step 786, Loss: {'train/loss': 0.020885691046714783, 'train/grad_norm_pre_clip': 2.636953115463257, 'timing/data': 0.0002379500074312091, 'timing/model': 0.39393470899085514, 'learning_rate/qwen_vl_interface': 9.547928499213593e-06, 'learning_rate/action_model': 9.547928499213593e-05, 'epoch': 0.06}) +[2026-10-03 19:04:45,282][starVLA.training.train_starvla][INFO] - Step 787, Loss: {'train/loss': 0.4201313853263855, 'train/grad_norm_pre_clip': 31.551496505737305, 'timing/data': 0.00028107999241910875, 'timing/model': 0.3974897090229206, 'learning_rate/qwen_vl_interface': 9.54663093855694e-06, 'learning_rate/action_model': 9.54663093855694e-05, 'epoch': 0.06}) +[2026-10-03 19:04:46,196][starVLA.training.train_starvla][INFO] - Step 788, Loss: {'train/loss': 0.7056897878646851, 'train/grad_norm_pre_clip': 16.574861526489258, 'timing/data': 0.00044956000056117773, 'timing/model': 0.6013726870005485, 'learning_rate/qwen_vl_interface': 9.545331611717607e-06, 'learning_rate/action_model': 9.545331611717608e-05, 'epoch': 0.06}) +[2026-10-03 19:04:46,904][starVLA.training.train_starvla][INFO] - Step 789, Loss: {'train/loss': 0.09583267569541931, 'train/grad_norm_pre_clip': 12.415258407592773, 'timing/data': 0.0002624999906402081, 'timing/model': 0.395486689027166, 'learning_rate/qwen_vl_interface': 9.544030519229697e-06, 'learning_rate/action_model': 9.544030519229697e-05, 'epoch': 0.06}) +[2026-10-03 19:04:47,609][starVLA.training.train_starvla][INFO] - Step 790, Loss: {'train/loss': 0.3595634400844574, 'train/grad_norm_pre_clip': 22.393905639648438, 'timing/data': 0.0003384099982213229, 'timing/model': 0.39468802901683375, 'learning_rate/qwen_vl_interface': 9.54272766162804e-06, 'learning_rate/action_model': 9.54272766162804e-05, 'epoch': 0.06}) +[2026-10-03 19:04:48,320][starVLA.training.train_starvla][INFO] - Step 791, Loss: {'train/loss': 0.3630731999874115, 'train/grad_norm_pre_clip': 19.809242248535156, 'timing/data': 0.00023202999727800488, 'timing/model': 0.39714745999663137, 'learning_rate/qwen_vl_interface': 9.54142303944819e-06, 'learning_rate/action_model': 9.541423039448191e-05, 'epoch': 0.06}) +[2026-10-03 19:04:49,028][starVLA.training.train_starvla][INFO] - Step 792, Loss: {'train/loss': 0.05709417909383774, 'train/grad_norm_pre_clip': 25.02501678466797, 'timing/data': 0.0003428399795666337, 'timing/model': 0.3957909879973158, 'learning_rate/qwen_vl_interface': 9.540116653226433e-06, 'learning_rate/action_model': 9.540116653226431e-05, 'epoch': 0.06}) +[2026-10-03 19:04:49,943][starVLA.training.train_starvla][INFO] - Step 793, Loss: {'train/loss': 0.1294308304786682, 'train/grad_norm_pre_clip': 24.950178146362305, 'timing/data': 0.000258639978710562, 'timing/model': 0.6011265270062722, 'learning_rate/qwen_vl_interface': 9.53880850349977e-06, 'learning_rate/action_model': 9.538808503499768e-05, 'epoch': 0.06}) +[2026-10-03 19:04:50,648][starVLA.training.train_starvla][INFO] - Step 794, Loss: {'train/loss': 0.722675085067749, 'train/grad_norm_pre_clip': 50.823890686035156, 'timing/data': 0.0003323800046928227, 'timing/model': 0.39514178899116814, 'learning_rate/qwen_vl_interface': 9.537498590805934e-06, 'learning_rate/action_model': 9.537498590805933e-05, 'epoch': 0.06}) +[2026-10-03 19:04:51,358][starVLA.training.train_starvla][INFO] - Step 795, Loss: {'train/loss': 0.10093259811401367, 'train/grad_norm_pre_clip': 23.356782913208008, 'timing/data': 0.00038168998435139656, 'timing/model': 0.3959499889751896, 'learning_rate/qwen_vl_interface': 9.536186915683379e-06, 'learning_rate/action_model': 9.53618691568338e-05, 'epoch': 0.06}) +[2026-10-03 19:04:52,066][starVLA.training.train_starvla][INFO] - Step 796, Loss: {'train/loss': 0.30681106448173523, 'train/grad_norm_pre_clip': 19.798091888427734, 'timing/data': 0.0004206900193821639, 'timing/model': 0.3951161690056324, 'learning_rate/qwen_vl_interface': 9.534873478671289e-06, 'learning_rate/action_model': 9.534873478671288e-05, 'epoch': 0.06}) +[2026-10-03 19:04:52,772][starVLA.training.train_starvla][INFO] - Step 797, Loss: {'train/loss': 0.08465379476547241, 'train/grad_norm_pre_clip': 20.631736755371094, 'timing/data': 0.0003534099960234016, 'timing/model': 0.39500606898218393, 'learning_rate/qwen_vl_interface': 9.533558280309563e-06, 'learning_rate/action_model': 9.533558280309563e-05, 'epoch': 0.06}) +[2026-10-03 19:04:53,479][starVLA.training.train_starvla][INFO] - Step 798, Loss: {'train/loss': 0.09946228563785553, 'train/grad_norm_pre_clip': 16.9464168548584, 'timing/data': 0.00033412998891435564, 'timing/model': 0.39502689000801183, 'learning_rate/qwen_vl_interface': 9.532241321138832e-06, 'learning_rate/action_model': 9.532241321138833e-05, 'epoch': 0.06}) +[2026-10-03 19:04:54,394][starVLA.training.train_starvla][INFO] - Step 799, Loss: {'train/loss': 0.04966909438371658, 'train/grad_norm_pre_clip': 11.5238676071167, 'timing/data': 0.00031991000287234783, 'timing/model': 0.39797314899624325, 'learning_rate/qwen_vl_interface': 9.530922601700451e-06, 'learning_rate/action_model': 9.530922601700452e-05, 'epoch': 0.06}) +[2026-10-03 19:04:55,102][starVLA.training.train_starvla][INFO] - Step 800, Loss: {'train/loss': 0.08431163430213928, 'train/grad_norm_pre_clip': 24.212871551513672, 'timing/data': 0.0003029999788850546, 'timing/model': 0.3964965689810924, 'learning_rate/qwen_vl_interface': 9.52960212253649e-06, 'learning_rate/action_model': 9.529602122536491e-05, 'epoch': 0.06}) +[2026-10-03 19:04:55,813][starVLA.training.train_starvla][INFO] - Step 801, Loss: {'train/loss': 0.08280070871114731, 'train/grad_norm_pre_clip': 28.883464813232422, 'timing/data': 0.0003414899983908981, 'timing/model': 0.39685677899979055, 'learning_rate/qwen_vl_interface': 9.528279884189754e-06, 'learning_rate/action_model': 9.528279884189754e-05, 'epoch': 0.06}) +[2026-10-03 19:04:56,521][starVLA.training.train_starvla][INFO] - Step 802, Loss: {'train/loss': 0.6859904527664185, 'train/grad_norm_pre_clip': 94.1891098022461, 'timing/data': 0.0003406799805816263, 'timing/model': 0.3962605689885095, 'learning_rate/qwen_vl_interface': 9.526955887203762e-06, 'learning_rate/action_model': 9.526955887203761e-05, 'epoch': 0.06}) +[2026-10-03 19:04:57,231][starVLA.training.train_starvla][INFO] - Step 803, Loss: {'train/loss': 0.16418668627738953, 'train/grad_norm_pre_clip': 75.03750610351562, 'timing/data': 0.0003148099931422621, 'timing/model': 0.39585185900796205, 'learning_rate/qwen_vl_interface': 9.525630132122758e-06, 'learning_rate/action_model': 9.525630132122758e-05, 'epoch': 0.06}) +[2026-10-03 19:04:58,135][starVLA.training.train_starvla][INFO] - Step 804, Loss: {'train/loss': 0.2665145993232727, 'train/grad_norm_pre_clip': 33.0402717590332, 'timing/data': 0.0002492800122126937, 'timing/model': 0.5916422479785979, 'learning_rate/qwen_vl_interface': 9.524302619491712e-06, 'learning_rate/action_model': 9.524302619491712e-05, 'epoch': 0.06}) +[2026-10-03 19:04:58,839][starVLA.training.train_starvla][INFO] - Step 805, Loss: {'train/loss': 0.18020454049110413, 'train/grad_norm_pre_clip': 46.97608184814453, 'timing/data': 0.00033306999830529094, 'timing/model': 0.3951139499840792, 'learning_rate/qwen_vl_interface': 9.522973349856315e-06, 'learning_rate/action_model': 9.522973349856315e-05, 'epoch': 0.06}) +[2026-10-03 19:04:59,541][starVLA.training.train_starvla][INFO] - Step 806, Loss: {'train/loss': 0.08371235430240631, 'train/grad_norm_pre_clip': 133.06651306152344, 'timing/data': 0.0002155900001525879, 'timing/model': 0.393144570000004, 'learning_rate/qwen_vl_interface': 9.521642323762976e-06, 'learning_rate/action_model': 9.521642323762976e-05, 'epoch': 0.06}) +[2026-10-03 19:05:00,250][starVLA.training.train_starvla][INFO] - Step 807, Loss: {'train/loss': 0.5535193681716919, 'train/grad_norm_pre_clip': 13.357866287231445, 'timing/data': 0.0003257000062149018, 'timing/model': 0.396087739005452, 'learning_rate/qwen_vl_interface': 9.520309541758833e-06, 'learning_rate/action_model': 9.520309541758833e-05, 'epoch': 0.06}) +[2026-10-03 19:05:00,955][starVLA.training.train_starvla][INFO] - Step 808, Loss: {'train/loss': 0.06679022312164307, 'train/grad_norm_pre_clip': 12.761847496032715, 'timing/data': 0.00034142000367864966, 'timing/model': 0.3926979000098072, 'learning_rate/qwen_vl_interface': 9.51897500439174e-06, 'learning_rate/action_model': 9.51897500439174e-05, 'epoch': 0.06}) +[2026-10-03 19:05:01,660][starVLA.training.train_starvla][INFO] - Step 809, Loss: {'train/loss': 0.6340866684913635, 'train/grad_norm_pre_clip': 41.78004837036133, 'timing/data': 0.000294740020763129, 'timing/model': 0.3942544989986345, 'learning_rate/qwen_vl_interface': 9.517638712210277e-06, 'learning_rate/action_model': 9.517638712210276e-05, 'epoch': 0.06}) +[2026-10-03 19:05:02,565][starVLA.training.train_starvla][INFO] - Step 810, Loss: {'train/loss': 0.5263795256614685, 'train/grad_norm_pre_clip': 91.59954833984375, 'timing/data': 0.00046733999624848366, 'timing/model': 0.3957588589983061, 'learning_rate/qwen_vl_interface': 9.51630066576374e-06, 'learning_rate/action_model': 9.51630066576374e-05, 'epoch': 0.06}) +[2026-10-03 19:05:03,276][starVLA.training.train_starvla][INFO] - Step 811, Loss: {'train/loss': 0.27857398986816406, 'train/grad_norm_pre_clip': 23.203954696655273, 'timing/data': 0.00028884998755529523, 'timing/model': 0.3970771089952905, 'learning_rate/qwen_vl_interface': 9.514960865602154e-06, 'learning_rate/action_model': 9.514960865602153e-05, 'epoch': 0.06}) +[2026-10-03 19:05:03,984][starVLA.training.train_starvla][INFO] - Step 812, Loss: {'train/loss': 0.25684142112731934, 'train/grad_norm_pre_clip': 31.50811004638672, 'timing/data': 0.0003570300177671015, 'timing/model': 0.3950730090145953, 'learning_rate/qwen_vl_interface': 9.513619312276259e-06, 'learning_rate/action_model': 9.513619312276258e-05, 'epoch': 0.06}) +[2026-10-03 19:05:04,694][starVLA.training.train_starvla][INFO] - Step 813, Loss: {'train/loss': 0.0375436469912529, 'train/grad_norm_pre_clip': 8.457547187805176, 'timing/data': 0.00032901999657042325, 'timing/model': 0.39656769900466315, 'learning_rate/qwen_vl_interface': 9.512276006337514e-06, 'learning_rate/action_model': 9.512276006337513e-05, 'epoch': 0.06}) +[2026-10-03 19:05:05,402][starVLA.training.train_starvla][INFO] - Step 814, Loss: {'train/loss': 0.12111244350671768, 'train/grad_norm_pre_clip': 61.81389617919922, 'timing/data': 0.0003551800036802888, 'timing/model': 0.3959990589937661, 'learning_rate/qwen_vl_interface': 9.510930948338102e-06, 'learning_rate/action_model': 9.510930948338102e-05, 'epoch': 0.06}) +[2026-10-03 19:05:06,313][starVLA.training.train_starvla][INFO] - Step 815, Loss: {'train/loss': 0.040039680898189545, 'train/grad_norm_pre_clip': 6.10148811340332, 'timing/data': 0.0003093699924647808, 'timing/model': 0.3956017989839893, 'learning_rate/qwen_vl_interface': 9.509584138830929e-06, 'learning_rate/action_model': 9.509584138830928e-05, 'epoch': 0.06}) +[2026-10-03 19:05:07,022][starVLA.training.train_starvla][INFO] - Step 816, Loss: {'train/loss': 0.1153678297996521, 'train/grad_norm_pre_clip': 66.03269958496094, 'timing/data': 0.0003481500025372952, 'timing/model': 0.3963903389812913, 'learning_rate/qwen_vl_interface': 9.508235578369613e-06, 'learning_rate/action_model': 9.508235578369612e-05, 'epoch': 0.06}) +[2026-10-03 19:05:07,732][starVLA.training.train_starvla][INFO] - Step 817, Loss: {'train/loss': 0.046294957399368286, 'train/grad_norm_pre_clip': 9.459186553955078, 'timing/data': 0.00022898000315763056, 'timing/model': 0.39702515900717117, 'learning_rate/qwen_vl_interface': 9.506885267508498e-06, 'learning_rate/action_model': 9.506885267508498e-05, 'epoch': 0.06}) +[2026-10-03 19:05:08,440][starVLA.training.train_starvla][INFO] - Step 818, Loss: {'train/loss': 0.02049154043197632, 'train/grad_norm_pre_clip': 2.6226799488067627, 'timing/data': 0.0002683099883142859, 'timing/model': 0.3958725390257314, 'learning_rate/qwen_vl_interface': 9.505533206802648e-06, 'learning_rate/action_model': 9.505533206802647e-05, 'epoch': 0.06}) +[2026-10-03 19:05:09,153][starVLA.training.train_starvla][INFO] - Step 819, Loss: {'train/loss': 0.27526742219924927, 'train/grad_norm_pre_clip': 19.80865478515625, 'timing/data': 0.0003583899815566838, 'timing/model': 0.39799988898448646, 'learning_rate/qwen_vl_interface': 9.50417939680784e-06, 'learning_rate/action_model': 9.504179396807839e-05, 'epoch': 0.06}) +[2026-10-03 19:05:10,062][starVLA.training.train_starvla][INFO] - Step 820, Loss: {'train/loss': 0.27380627393722534, 'train/grad_norm_pre_clip': 70.83035278320312, 'timing/data': 0.0003436999977566302, 'timing/model': 0.5965612080181018, 'learning_rate/qwen_vl_interface': 9.502823838080578e-06, 'learning_rate/action_model': 9.502823838080578e-05, 'epoch': 0.06}) +[2026-10-03 19:05:10,770][starVLA.training.train_starvla][INFO] - Step 821, Loss: {'train/loss': 0.12163320928812027, 'train/grad_norm_pre_clip': 16.2275447845459, 'timing/data': 0.00036504000308923423, 'timing/model': 0.39489051801501773, 'learning_rate/qwen_vl_interface': 9.501466531178079e-06, 'learning_rate/action_model': 9.501466531178079e-05, 'epoch': 0.06}) +[2026-10-03 19:05:11,480][starVLA.training.train_starvla][INFO] - Step 822, Loss: {'train/loss': 0.00477618258446455, 'train/grad_norm_pre_clip': 0.5429174900054932, 'timing/data': 0.0003340700059197843, 'timing/model': 0.3965032889973372, 'learning_rate/qwen_vl_interface': 9.50010747665828e-06, 'learning_rate/action_model': 9.500107476658281e-05, 'epoch': 0.06}) +[2026-10-03 19:05:12,191][starVLA.training.train_starvla][INFO] - Step 823, Loss: {'train/loss': 0.29364195466041565, 'train/grad_norm_pre_clip': 43.72276306152344, 'timing/data': 0.00033522999729029834, 'timing/model': 0.39802724798209965, 'learning_rate/qwen_vl_interface': 9.498746675079842e-06, 'learning_rate/action_model': 9.49874667507984e-05, 'epoch': 0.06}) +[2026-10-03 19:05:12,900][starVLA.training.train_starvla][INFO] - Step 824, Loss: {'train/loss': 0.05417831987142563, 'train/grad_norm_pre_clip': 142.31326293945312, 'timing/data': 0.00041614999645389616, 'timing/model': 0.39481170900398865, 'learning_rate/qwen_vl_interface': 9.49738412700213e-06, 'learning_rate/action_model': 9.49738412700213e-05, 'epoch': 0.06}) +[2026-10-03 19:05:13,609][starVLA.training.train_starvla][INFO] - Step 825, Loss: {'train/loss': 1.0089493989944458, 'train/grad_norm_pre_clip': 190.9387969970703, 'timing/data': 0.00032610000926069915, 'timing/model': 0.3958016689866781, 'learning_rate/qwen_vl_interface': 9.496019832985243e-06, 'learning_rate/action_model': 9.496019832985243e-05, 'epoch': 0.06}) +[2026-10-03 19:05:14,514][starVLA.training.train_starvla][INFO] - Step 826, Loss: {'train/loss': 0.004614945501089096, 'train/grad_norm_pre_clip': 1.2646597623825073, 'timing/data': 0.0003562299825716764, 'timing/model': 0.39493028898141347, 'learning_rate/qwen_vl_interface': 9.49465379358999e-06, 'learning_rate/action_model': 9.49465379358999e-05, 'epoch': 0.06}) +[2026-10-03 19:05:15,227][starVLA.training.train_starvla][INFO] - Step 827, Loss: {'train/loss': 0.34318363666534424, 'train/grad_norm_pre_clip': 59.12657928466797, 'timing/data': 0.0003107899974565953, 'timing/model': 0.39780953899025917, 'learning_rate/qwen_vl_interface': 9.493286009377896e-06, 'learning_rate/action_model': 9.493286009377896e-05, 'epoch': 0.06}) +[2026-10-03 19:05:15,937][starVLA.training.train_starvla][INFO] - Step 828, Loss: {'train/loss': 0.21800744533538818, 'train/grad_norm_pre_clip': 24.945402145385742, 'timing/data': 0.0003630499995779246, 'timing/model': 0.3962269190233201, 'learning_rate/qwen_vl_interface': 9.491916480911208e-06, 'learning_rate/action_model': 9.491916480911207e-05, 'epoch': 0.06}) +[2026-10-03 19:05:16,645][starVLA.training.train_starvla][INFO] - Step 829, Loss: {'train/loss': 0.2795524597167969, 'train/grad_norm_pre_clip': 20.143810272216797, 'timing/data': 0.0003343889839015901, 'timing/model': 0.3940635500184726, 'learning_rate/qwen_vl_interface': 9.490545208752884e-06, 'learning_rate/action_model': 9.490545208752883e-05, 'epoch': 0.06}) +[2026-10-03 19:05:17,352][starVLA.training.train_starvla][INFO] - Step 830, Loss: {'train/loss': 0.4369506537914276, 'train/grad_norm_pre_clip': 50.71342849731445, 'timing/data': 0.00021366000873968005, 'timing/model': 0.3965388590004295, 'learning_rate/qwen_vl_interface': 9.489172193466603e-06, 'learning_rate/action_model': 9.489172193466603e-05, 'epoch': 0.06}) +[2026-10-03 19:05:18,264][starVLA.training.train_starvla][INFO] - Step 831, Loss: {'train/loss': 0.012998783960938454, 'train/grad_norm_pre_clip': 6.940393447875977, 'timing/data': 0.0002884299901779741, 'timing/model': 0.3956961189978756, 'learning_rate/qwen_vl_interface': 9.487797435616762e-06, 'learning_rate/action_model': 9.487797435616761e-05, 'epoch': 0.06}) +[2026-10-03 19:05:18,973][starVLA.training.train_starvla][INFO] - Step 832, Loss: {'train/loss': 0.010262706317007542, 'train/grad_norm_pre_clip': 20.687400817871094, 'timing/data': 0.00021488999482244253, 'timing/model': 0.3954700390168, 'learning_rate/qwen_vl_interface': 9.486420935768472e-06, 'learning_rate/action_model': 9.486420935768472e-05, 'epoch': 0.06}) +[2026-10-03 19:05:19,677][starVLA.training.train_starvla][INFO] - Step 833, Loss: {'train/loss': 0.5859488248825073, 'train/grad_norm_pre_clip': 63.001216888427734, 'timing/data': 0.00034677001531235874, 'timing/model': 0.39423914899816737, 'learning_rate/qwen_vl_interface': 9.485042694487558e-06, 'learning_rate/action_model': 9.485042694487559e-05, 'epoch': 0.06}) +[2026-10-03 19:05:20,383][starVLA.training.train_starvla][INFO] - Step 834, Loss: {'train/loss': 0.24047411978244781, 'train/grad_norm_pre_clip': 68.4939193725586, 'timing/data': 0.00044624999281950295, 'timing/model': 0.3949703389953356, 'learning_rate/qwen_vl_interface': 9.483662712340565e-06, 'learning_rate/action_model': 9.483662712340564e-05, 'epoch': 0.06}) +[2026-10-03 19:05:21,093][starVLA.training.train_starvla][INFO] - Step 835, Loss: {'train/loss': 0.007076149806380272, 'train/grad_norm_pre_clip': 1.7588047981262207, 'timing/data': 0.0003161400090903044, 'timing/model': 0.3969850589928683, 'learning_rate/qwen_vl_interface': 9.482280989894748e-06, 'learning_rate/action_model': 9.482280989894748e-05, 'epoch': 0.06}) +[2026-10-03 19:05:22,001][starVLA.training.train_starvla][INFO] - Step 836, Loss: {'train/loss': 0.3346574008464813, 'train/grad_norm_pre_clip': 37.156681060791016, 'timing/data': 0.0003505100030452013, 'timing/model': 0.39370017999317497, 'learning_rate/qwen_vl_interface': 9.480897527718083e-06, 'learning_rate/action_model': 9.480897527718084e-05, 'epoch': 0.06}) +[2026-10-03 19:05:22,707][starVLA.training.train_starvla][INFO] - Step 837, Loss: {'train/loss': 0.002595195546746254, 'train/grad_norm_pre_clip': 0.9774930477142334, 'timing/data': 0.0003542900085449219, 'timing/model': 0.39459720900049433, 'learning_rate/qwen_vl_interface': 9.479512326379263e-06, 'learning_rate/action_model': 9.479512326379264e-05, 'epoch': 0.06}) +[2026-10-03 19:05:23,414][starVLA.training.train_starvla][INFO] - Step 838, Loss: {'train/loss': 0.1565575748682022, 'train/grad_norm_pre_clip': 75.13461303710938, 'timing/data': 0.0003073700063396245, 'timing/model': 0.39498401901801117, 'learning_rate/qwen_vl_interface': 9.478125386447686e-06, 'learning_rate/action_model': 9.478125386447686e-05, 'epoch': 0.06}) +[2026-10-03 19:05:24,126][starVLA.training.train_starvla][INFO] - Step 839, Loss: {'train/loss': 0.4681857228279114, 'train/grad_norm_pre_clip': 87.02072143554688, 'timing/data': 0.00028919000760652125, 'timing/model': 0.3974535090092104, 'learning_rate/qwen_vl_interface': 9.476736708493472e-06, 'learning_rate/action_model': 9.476736708493472e-05, 'epoch': 0.06}) +[2026-10-03 19:05:24,834][starVLA.training.train_starvla][INFO] - Step 840, Loss: {'train/loss': 0.1422138512134552, 'train/grad_norm_pre_clip': 88.70085906982422, 'timing/data': 0.0003577799943741411, 'timing/model': 0.3949630389979575, 'learning_rate/qwen_vl_interface': 9.475346293087457e-06, 'learning_rate/action_model': 9.475346293087458e-05, 'epoch': 0.06}) +[2026-10-03 19:05:25,743][starVLA.training.train_starvla][INFO] - Step 841, Loss: {'train/loss': 0.332181453704834, 'train/grad_norm_pre_clip': 32.20848083496094, 'timing/data': 0.00035749998642131686, 'timing/model': 0.5964739090122748, 'learning_rate/qwen_vl_interface': 9.473954140801184e-06, 'learning_rate/action_model': 9.473954140801185e-05, 'epoch': 0.06}) +[2026-10-03 19:05:26,448][starVLA.training.train_starvla][INFO] - Step 842, Loss: {'train/loss': 0.4742565155029297, 'train/grad_norm_pre_clip': 24.31898307800293, 'timing/data': 0.00023617001716047525, 'timing/model': 0.3948535489907954, 'learning_rate/qwen_vl_interface': 9.472560252206918e-06, 'learning_rate/action_model': 9.472560252206918e-05, 'epoch': 0.06}) +[2026-10-03 19:05:27,158][starVLA.training.train_starvla][INFO] - Step 843, Loss: {'train/loss': 0.032050419598817825, 'train/grad_norm_pre_clip': 7.0511345863342285, 'timing/data': 0.00035311997635290027, 'timing/model': 0.39674967899918556, 'learning_rate/qwen_vl_interface': 9.471164627877633e-06, 'learning_rate/action_model': 9.471164627877633e-05, 'epoch': 0.06}) +[2026-10-03 19:05:27,865][starVLA.training.train_starvla][INFO] - Step 844, Loss: {'train/loss': 0.24356457591056824, 'train/grad_norm_pre_clip': 221.57601928710938, 'timing/data': 0.00044497998896986246, 'timing/model': 0.3949931990064215, 'learning_rate/qwen_vl_interface': 9.469767268387016e-06, 'learning_rate/action_model': 9.469767268387016e-05, 'epoch': 0.06}) +[2026-10-03 19:05:28,572][starVLA.training.train_starvla][INFO] - Step 845, Loss: {'train/loss': 0.04882363975048065, 'train/grad_norm_pre_clip': 10.178200721740723, 'timing/data': 0.00031114998273551464, 'timing/model': 0.3949322790140286, 'learning_rate/qwen_vl_interface': 9.468368174309468e-06, 'learning_rate/action_model': 9.468368174309469e-05, 'epoch': 0.06}) +[2026-10-03 19:05:29,473][starVLA.training.train_starvla][INFO] - Step 846, Loss: {'train/loss': 1.082902193069458, 'train/grad_norm_pre_clip': 125.53781127929688, 'timing/data': 0.0003563989885151386, 'timing/model': 0.5904210190055892, 'learning_rate/qwen_vl_interface': 9.466967346220107e-06, 'learning_rate/action_model': 9.466967346220107e-05, 'epoch': 0.06}) +[2026-10-03 19:05:30,181][starVLA.training.train_starvla][INFO] - Step 847, Loss: {'train/loss': 0.047901444137096405, 'train/grad_norm_pre_clip': 15.96051025390625, 'timing/data': 0.0002855600032489747, 'timing/model': 0.3938811089901719, 'learning_rate/qwen_vl_interface': 9.465564784694759e-06, 'learning_rate/action_model': 9.465564784694759e-05, 'epoch': 0.06}) +[2026-10-03 19:05:30,890][starVLA.training.train_starvla][INFO] - Step 848, Loss: {'train/loss': 0.21699923276901245, 'train/grad_norm_pre_clip': 60.306339263916016, 'timing/data': 0.0003205800021532923, 'timing/model': 0.3955810590123292, 'learning_rate/qwen_vl_interface': 9.464160490309962e-06, 'learning_rate/action_model': 9.464160490309962e-05, 'epoch': 0.06}) +[2026-10-03 19:05:31,598][starVLA.training.train_starvla][INFO] - Step 849, Loss: {'train/loss': 0.0058503225445747375, 'train/grad_norm_pre_clip': 7.66614294052124, 'timing/data': 0.0003566199738997966, 'timing/model': 0.3951487389858812, 'learning_rate/qwen_vl_interface': 9.462754463642972e-06, 'learning_rate/action_model': 9.46275446364297e-05, 'epoch': 0.06}) +[2026-10-03 19:05:32,303][starVLA.training.train_starvla][INFO] - Step 850, Loss: {'train/loss': 0.18788325786590576, 'train/grad_norm_pre_clip': 26.616235733032227, 'timing/data': 0.000318689999403432, 'timing/model': 0.3933790900046006, 'learning_rate/qwen_vl_interface': 9.461346705271753e-06, 'learning_rate/action_model': 9.461346705271752e-05, 'epoch': 0.06}) +[2026-10-03 19:05:33,011][starVLA.training.train_starvla][INFO] - Step 851, Loss: {'train/loss': 0.028235647827386856, 'train/grad_norm_pre_clip': 13.927868843078613, 'timing/data': 0.0003128300013486296, 'timing/model': 0.3966475189954508, 'learning_rate/qwen_vl_interface': 9.459937215774976e-06, 'learning_rate/action_model': 9.459937215774976e-05, 'epoch': 0.06}) +[2026-10-03 19:05:33,920][starVLA.training.train_starvla][INFO] - Step 852, Loss: {'train/loss': 0.01630483940243721, 'train/grad_norm_pre_clip': 9.478875160217285, 'timing/data': 0.00028302002465352416, 'timing/model': 0.3948857989744283, 'learning_rate/qwen_vl_interface': 9.458525995732035e-06, 'learning_rate/action_model': 9.458525995732035e-05, 'epoch': 0.06}) +[2026-10-03 19:05:34,627][starVLA.training.train_starvla][INFO] - Step 853, Loss: {'train/loss': 0.32180818915367126, 'train/grad_norm_pre_clip': 23.1496524810791, 'timing/data': 0.00029170000925660133, 'timing/model': 0.39497334900079295, 'learning_rate/qwen_vl_interface': 9.457113045723027e-06, 'learning_rate/action_model': 9.457113045723026e-05, 'epoch': 0.06}) +[2026-10-03 19:05:35,332][starVLA.training.train_starvla][INFO] - Step 854, Loss: {'train/loss': 0.7664493322372437, 'train/grad_norm_pre_clip': 107.25271606445312, 'timing/data': 0.00032493998878635466, 'timing/model': 0.39452663899282925, 'learning_rate/qwen_vl_interface': 9.455698366328762e-06, 'learning_rate/action_model': 9.455698366328762e-05, 'epoch': 0.06}) +[2026-10-03 19:05:36,041][starVLA.training.train_starvla][INFO] - Step 855, Loss: {'train/loss': 0.42787155508995056, 'train/grad_norm_pre_clip': 35.891845703125, 'timing/data': 0.0003415999817661941, 'timing/model': 0.39604208897799253, 'learning_rate/qwen_vl_interface': 9.454281958130761e-06, 'learning_rate/action_model': 9.454281958130761e-05, 'epoch': 0.06}) +[2026-10-03 19:05:36,749][starVLA.training.train_starvla][INFO] - Step 856, Loss: {'train/loss': 0.009071161970496178, 'train/grad_norm_pre_clip': 2.605998992919922, 'timing/data': 0.00033645000075921416, 'timing/model': 0.3946699089719914, 'learning_rate/qwen_vl_interface': 9.452863821711257e-06, 'learning_rate/action_model': 9.452863821711258e-05, 'epoch': 0.06}) +[2026-10-03 19:05:37,654][starVLA.training.train_starvla][INFO] - Step 857, Loss: {'train/loss': 0.36031877994537354, 'train/grad_norm_pre_clip': 28.69661521911621, 'timing/data': 0.0003011499939020723, 'timing/model': 0.5939124389842618, 'learning_rate/qwen_vl_interface': 9.451443957653195e-06, 'learning_rate/action_model': 9.451443957653195e-05, 'epoch': 0.06}) +[2026-10-03 19:05:38,361][starVLA.training.train_starvla][INFO] - Step 858, Loss: {'train/loss': 0.26583167910575867, 'train/grad_norm_pre_clip': 64.35417175292969, 'timing/data': 0.0002281299966853112, 'timing/model': 0.3947443490033038, 'learning_rate/qwen_vl_interface': 9.450022366540221e-06, 'learning_rate/action_model': 9.45002236654022e-05, 'epoch': 0.06}) +[2026-10-03 19:05:39,070][starVLA.training.train_starvla][INFO] - Step 859, Loss: {'train/loss': 0.27083107829093933, 'train/grad_norm_pre_clip': 43.71963882446289, 'timing/data': 0.0003096899890806526, 'timing/model': 0.3952872890222352, 'learning_rate/qwen_vl_interface': 9.448599048956702e-06, 'learning_rate/action_model': 9.448599048956702e-05, 'epoch': 0.06}) +[2026-10-03 19:05:39,776][starVLA.training.train_starvla][INFO] - Step 860, Loss: {'train/loss': 0.2983773648738861, 'train/grad_norm_pre_clip': 24.810745239257812, 'timing/data': 0.0003392200160305947, 'timing/model': 0.39503429899923503, 'learning_rate/qwen_vl_interface': 9.447174005487708e-06, 'learning_rate/action_model': 9.447174005487708e-05, 'epoch': 0.06}) +[2026-10-03 19:05:40,482][starVLA.training.train_starvla][INFO] - Step 861, Loss: {'train/loss': 0.03831038624048233, 'train/grad_norm_pre_clip': 22.656856536865234, 'timing/data': 0.00028551000286825, 'timing/model': 0.39489097899058834, 'learning_rate/qwen_vl_interface': 9.445747236719022e-06, 'learning_rate/action_model': 9.445747236719022e-05, 'epoch': 0.06}) +[2026-10-03 19:05:41,186][starVLA.training.train_starvla][INFO] - Step 862, Loss: {'train/loss': 0.6857366561889648, 'train/grad_norm_pre_clip': 71.53343963623047, 'timing/data': 0.0003335599903948605, 'timing/model': 0.39421723899431527, 'learning_rate/qwen_vl_interface': 9.444318743237135e-06, 'learning_rate/action_model': 9.444318743237134e-05, 'epoch': 0.06}) +[2026-10-03 19:05:42,089][starVLA.training.train_starvla][INFO] - Step 863, Loss: {'train/loss': 0.02531937137246132, 'train/grad_norm_pre_clip': 12.354541778564453, 'timing/data': 0.00029609998455271125, 'timing/model': 0.39656163900508545, 'learning_rate/qwen_vl_interface': 9.442888525629244e-06, 'learning_rate/action_model': 9.442888525629245e-05, 'epoch': 0.06}) +[2026-10-03 19:05:42,792][starVLA.training.train_starvla][INFO] - Step 864, Loss: {'train/loss': 1.7481977939605713, 'train/grad_norm_pre_clip': 126.22510528564453, 'timing/data': 0.00033263000659644604, 'timing/model': 0.3927909299964085, 'learning_rate/qwen_vl_interface': 9.441456584483261e-06, 'learning_rate/action_model': 9.441456584483261e-05, 'epoch': 0.06}) +[2026-10-03 19:05:43,498][starVLA.training.train_starvla][INFO] - Step 865, Loss: {'train/loss': 0.08801118284463882, 'train/grad_norm_pre_clip': 62.81428909301758, 'timing/data': 0.00020054998458363116, 'timing/model': 0.3941531890013721, 'learning_rate/qwen_vl_interface': 9.440022920387801e-06, 'learning_rate/action_model': 9.440022920387801e-05, 'epoch': 0.06}) +[2026-10-03 19:05:44,200][starVLA.training.train_starvla][INFO] - Step 866, Loss: {'train/loss': 0.09245811402797699, 'train/grad_norm_pre_clip': 36.22929000854492, 'timing/data': 0.00024434999795630574, 'timing/model': 0.3931213190080598, 'learning_rate/qwen_vl_interface': 9.438587533932191e-06, 'learning_rate/action_model': 9.43858753393219e-05, 'epoch': 0.06}) +[2026-10-03 19:05:44,913][starVLA.training.train_starvla][INFO] - Step 867, Loss: {'train/loss': 0.17713305354118347, 'train/grad_norm_pre_clip': 23.54300880432129, 'timing/data': 0.00030504001188091934, 'timing/model': 0.39784525899449363, 'learning_rate/qwen_vl_interface': 9.43715042570646e-06, 'learning_rate/action_model': 9.43715042570646e-05, 'epoch': 0.06}) +[2026-10-03 19:05:45,824][starVLA.training.train_starvla][INFO] - Step 868, Loss: {'train/loss': 0.42178794741630554, 'train/grad_norm_pre_clip': 39.22943115234375, 'timing/data': 0.0004054100136272609, 'timing/model': 0.39505776899750344, 'learning_rate/qwen_vl_interface': 9.435711596301353e-06, 'learning_rate/action_model': 9.435711596301352e-05, 'epoch': 0.06}) +[2026-10-03 19:05:46,530][starVLA.training.train_starvla][INFO] - Step 869, Loss: {'train/loss': 0.30163970589637756, 'train/grad_norm_pre_clip': 55.03876876831055, 'timing/data': 0.00022747000912204385, 'timing/model': 0.39449710998451337, 'learning_rate/qwen_vl_interface': 9.434271046308317e-06, 'learning_rate/action_model': 9.434271046308317e-05, 'epoch': 0.06}) +[2026-10-03 19:05:47,237][starVLA.training.train_starvla][INFO] - Step 870, Loss: {'train/loss': 0.057183533906936646, 'train/grad_norm_pre_clip': 9.733490943908691, 'timing/data': 0.0002816299966070801, 'timing/model': 0.39514344901544973, 'learning_rate/qwen_vl_interface': 9.432828776319507e-06, 'learning_rate/action_model': 9.432828776319505e-05, 'epoch': 0.06}) +[2026-10-03 19:05:47,948][starVLA.training.train_starvla][INFO] - Step 871, Loss: {'train/loss': 0.14989949762821198, 'train/grad_norm_pre_clip': 36.552223205566406, 'timing/data': 0.00033822000841610134, 'timing/model': 0.3972437189950142, 'learning_rate/qwen_vl_interface': 9.431384786927785e-06, 'learning_rate/action_model': 9.431384786927786e-05, 'epoch': 0.06}) +[2026-10-03 19:05:48,656][starVLA.training.train_starvla][INFO] - Step 872, Loss: {'train/loss': 0.10124234110116959, 'train/grad_norm_pre_clip': 20.033418655395508, 'timing/data': 0.00045629998203366995, 'timing/model': 0.3952635890163947, 'learning_rate/qwen_vl_interface': 9.429939078726723e-06, 'learning_rate/action_model': 9.429939078726722e-05, 'epoch': 0.06}) +[2026-10-03 19:05:49,577][starVLA.training.train_starvla][INFO] - Step 873, Loss: {'train/loss': 0.8455622792243958, 'train/grad_norm_pre_clip': 31.710174560546875, 'timing/data': 0.00023574999067932367, 'timing/model': 0.60455168801127, 'learning_rate/qwen_vl_interface': 9.428491652310598e-06, 'learning_rate/action_model': 9.428491652310597e-05, 'epoch': 0.06}) +[2026-10-03 19:05:50,287][starVLA.training.train_starvla][INFO] - Step 874, Loss: {'train/loss': 0.017869174480438232, 'train/grad_norm_pre_clip': 3.289470672607422, 'timing/data': 0.0002342300140298903, 'timing/model': 0.39579567898181267, 'learning_rate/qwen_vl_interface': 9.427042508274386e-06, 'learning_rate/action_model': 9.427042508274386e-05, 'epoch': 0.06}) +[2026-10-03 19:05:50,994][starVLA.training.train_starvla][INFO] - Step 875, Loss: {'train/loss': 0.09021048992872238, 'train/grad_norm_pre_clip': 50.159400939941406, 'timing/data': 0.00022628999431617558, 'timing/model': 0.39440119900973514, 'learning_rate/qwen_vl_interface': 9.425591647213783e-06, 'learning_rate/action_model': 9.425591647213782e-05, 'epoch': 0.06}) +[2026-10-03 19:05:51,698][starVLA.training.train_starvla][INFO] - Step 876, Loss: {'train/loss': 0.4258338212966919, 'train/grad_norm_pre_clip': 41.904476165771484, 'timing/data': 0.00020971999038010836, 'timing/model': 0.3930547799973283, 'learning_rate/qwen_vl_interface': 9.424139069725177e-06, 'learning_rate/action_model': 9.424139069725178e-05, 'epoch': 0.06}) +[2026-10-03 19:05:52,400][starVLA.training.train_starvla][INFO] - Step 877, Loss: {'train/loss': 0.3186200261116028, 'train/grad_norm_pre_clip': 61.830604553222656, 'timing/data': 0.00020773999858647585, 'timing/model': 0.39361831900896505, 'learning_rate/qwen_vl_interface': 9.422684776405673e-06, 'learning_rate/action_model': 9.422684776405672e-05, 'epoch': 0.06}) +[2026-10-03 19:05:53,099][starVLA.training.train_starvla][INFO] - Step 878, Loss: {'train/loss': 0.030569210648536682, 'train/grad_norm_pre_clip': 13.764185905456543, 'timing/data': 0.00034689001040533185, 'timing/model': 0.39104236898128875, 'learning_rate/qwen_vl_interface': 9.421228767853072e-06, 'learning_rate/action_model': 9.421228767853072e-05, 'epoch': 0.06}) +[2026-10-03 19:05:53,999][starVLA.training.train_starvla][INFO] - Step 879, Loss: {'train/loss': 0.233928382396698, 'train/grad_norm_pre_clip': 51.0118293762207, 'timing/data': 0.0003505299973767251, 'timing/model': 0.3906152790004853, 'learning_rate/qwen_vl_interface': 9.419771044665887e-06, 'learning_rate/action_model': 9.419771044665887e-05, 'epoch': 0.06}) +[2026-10-03 19:05:54,699][starVLA.training.train_starvla][INFO] - Step 880, Loss: {'train/loss': 0.017938116565346718, 'train/grad_norm_pre_clip': 9.393064498901367, 'timing/data': 0.0003445699985604733, 'timing/model': 0.3919647900038399, 'learning_rate/qwen_vl_interface': 9.41831160744333e-06, 'learning_rate/action_model': 9.41831160744333e-05, 'epoch': 0.06}) +[2026-10-03 19:05:55,397][starVLA.training.train_starvla][INFO] - Step 881, Loss: {'train/loss': 0.011402837932109833, 'train/grad_norm_pre_clip': 3.0090246200561523, 'timing/data': 0.0002790400176309049, 'timing/model': 0.39146554900798947, 'learning_rate/qwen_vl_interface': 9.416850456785322e-06, 'learning_rate/action_model': 9.416850456785321e-05, 'epoch': 0.06}) +[2026-10-03 19:05:56,095][starVLA.training.train_starvla][INFO] - Step 882, Loss: {'train/loss': 0.8371974229812622, 'train/grad_norm_pre_clip': 81.55235290527344, 'timing/data': 0.0003603200020734221, 'timing/model': 0.3913551389996428, 'learning_rate/qwen_vl_interface': 9.415387593292489e-06, 'learning_rate/action_model': 9.415387593292488e-05, 'epoch': 0.06}) +[2026-10-03 19:05:56,796][starVLA.training.train_starvla][INFO] - Step 883, Loss: {'train/loss': 0.865615725517273, 'train/grad_norm_pre_clip': 55.29622268676758, 'timing/data': 0.0003489399969112128, 'timing/model': 0.3927386599825695, 'learning_rate/qwen_vl_interface': 9.413923017566155e-06, 'learning_rate/action_model': 9.413923017566155e-05, 'epoch': 0.06}) +[2026-10-03 19:05:57,704][starVLA.training.train_starvla][INFO] - Step 884, Loss: {'train/loss': 0.2620506286621094, 'train/grad_norm_pre_clip': 50.34252166748047, 'timing/data': 0.00021283997921273112, 'timing/model': 0.39274745000875555, 'learning_rate/qwen_vl_interface': 9.412456730208352e-06, 'learning_rate/action_model': 9.412456730208352e-05, 'epoch': 0.06}) +[2026-10-03 19:05:58,404][starVLA.training.train_starvla][INFO] - Step 885, Loss: {'train/loss': 0.23980700969696045, 'train/grad_norm_pre_clip': 88.39344024658203, 'timing/data': 0.00024969998048618436, 'timing/model': 0.3923283390176948, 'learning_rate/qwen_vl_interface': 9.41098873182182e-06, 'learning_rate/action_model': 9.41098873182182e-05, 'epoch': 0.06}) +[2026-10-03 19:05:59,106][starVLA.training.train_starvla][INFO] - Step 886, Loss: {'train/loss': 0.11560200154781342, 'train/grad_norm_pre_clip': 52.37449264526367, 'timing/data': 0.00021448000916279852, 'timing/model': 0.3927762089879252, 'learning_rate/qwen_vl_interface': 9.409519023009993e-06, 'learning_rate/action_model': 9.409519023009993e-05, 'epoch': 0.06}) +[2026-10-03 19:05:59,807][starVLA.training.train_starvla][INFO] - Step 887, Loss: {'train/loss': 0.20444048941135406, 'train/grad_norm_pre_clip': 35.99671936035156, 'timing/data': 0.0003374499792698771, 'timing/model': 0.3923927790019661, 'learning_rate/qwen_vl_interface': 9.408047604377014e-06, 'learning_rate/action_model': 9.408047604377014e-05, 'epoch': 0.06}) +[2026-10-03 19:06:00,506][starVLA.training.train_starvla][INFO] - Step 888, Loss: {'train/loss': 0.6492548584938049, 'train/grad_norm_pre_clip': 83.51988220214844, 'timing/data': 0.00021415000082924962, 'timing/model': 0.39014023999334313, 'learning_rate/qwen_vl_interface': 9.40657447652773e-06, 'learning_rate/action_model': 9.406574476527731e-05, 'epoch': 0.06}) +[2026-10-03 19:06:01,406][starVLA.training.train_starvla][INFO] - Step 889, Loss: {'train/loss': 0.180630624294281, 'train/grad_norm_pre_clip': 67.33829498291016, 'timing/data': 0.0002529800112824887, 'timing/model': 0.5934628079994582, 'learning_rate/qwen_vl_interface': 9.405099640067685e-06, 'learning_rate/action_model': 9.405099640067685e-05, 'epoch': 0.07}) +[2026-10-03 19:06:02,105][starVLA.training.train_starvla][INFO] - Step 890, Loss: {'train/loss': 0.1103011965751648, 'train/grad_norm_pre_clip': 46.24199676513672, 'timing/data': 0.00020897999638691545, 'timing/model': 0.3925355999963358, 'learning_rate/qwen_vl_interface': 9.40362309560313e-06, 'learning_rate/action_model': 9.40362309560313e-05, 'epoch': 0.07}) +[2026-10-03 19:06:02,806][starVLA.training.train_starvla][INFO] - Step 891, Loss: {'train/loss': 0.23192445933818817, 'train/grad_norm_pre_clip': 40.48964309692383, 'timing/data': 0.0004888400144409388, 'timing/model': 0.3920757890155073, 'learning_rate/qwen_vl_interface': 9.402144843741016e-06, 'learning_rate/action_model': 9.402144843741017e-05, 'epoch': 0.07}) +[2026-10-03 19:06:03,506][starVLA.training.train_starvla][INFO] - Step 892, Loss: {'train/loss': 0.11980178207159042, 'train/grad_norm_pre_clip': 30.239822387695312, 'timing/data': 0.00035782999475486577, 'timing/model': 0.3898685500025749, 'learning_rate/qwen_vl_interface': 9.400664885088999e-06, 'learning_rate/action_model': 9.400664885088999e-05, 'epoch': 0.07}) +[2026-10-03 19:06:04,209][starVLA.training.train_starvla][INFO] - Step 893, Loss: {'train/loss': 0.018651966005563736, 'train/grad_norm_pre_clip': 10.143851280212402, 'timing/data': 0.00025316901155747473, 'timing/model': 0.393870350002544, 'learning_rate/qwen_vl_interface': 9.39918322025543e-06, 'learning_rate/action_model': 9.39918322025543e-05, 'epoch': 0.07}) +[2026-10-03 19:06:05,116][starVLA.training.train_starvla][INFO] - Step 894, Loss: {'train/loss': 0.5726338624954224, 'train/grad_norm_pre_clip': 276.458984375, 'timing/data': 0.00027175998548045754, 'timing/model': 0.5994445690012071, 'learning_rate/qwen_vl_interface': 9.397699849849368e-06, 'learning_rate/action_model': 9.397699849849367e-05, 'epoch': 0.07}) +[2026-10-03 19:06:05,817][starVLA.training.train_starvla][INFO] - Step 895, Loss: {'train/loss': 0.3815617561340332, 'train/grad_norm_pre_clip': 74.5344009399414, 'timing/data': 0.00020379998022690415, 'timing/model': 0.39233984900056385, 'learning_rate/qwen_vl_interface': 9.39621477448057e-06, 'learning_rate/action_model': 9.39621477448057e-05, 'epoch': 0.07}) +[2026-10-03 19:06:06,517][starVLA.training.train_starvla][INFO] - Step 896, Loss: {'train/loss': 0.2862154543399811, 'train/grad_norm_pre_clip': 44.855567932128906, 'timing/data': 0.00022851000539958477, 'timing/model': 0.39038242999231443, 'learning_rate/qwen_vl_interface': 9.394727994759495e-06, 'learning_rate/action_model': 9.394727994759495e-05, 'epoch': 0.07}) +[2026-10-03 19:06:07,218][starVLA.training.train_starvla][INFO] - Step 897, Loss: {'train/loss': 0.31642070412635803, 'train/grad_norm_pre_clip': 85.17436981201172, 'timing/data': 0.0002631000243127346, 'timing/model': 0.39349283001502044, 'learning_rate/qwen_vl_interface': 9.3932395112973e-06, 'learning_rate/action_model': 9.3932395112973e-05, 'epoch': 0.07}) +[2026-10-03 19:06:07,914][starVLA.training.train_starvla][INFO] - Step 898, Loss: {'train/loss': 0.250530481338501, 'train/grad_norm_pre_clip': 36.042022705078125, 'timing/data': 0.0003688099968712777, 'timing/model': 0.38971693100756966, 'learning_rate/qwen_vl_interface': 9.391749324705846e-06, 'learning_rate/action_model': 9.391749324705846e-05, 'epoch': 0.07}) +[2026-10-03 19:06:08,818][starVLA.training.train_starvla][INFO] - Step 899, Loss: {'train/loss': 0.16238951683044434, 'train/grad_norm_pre_clip': 29.52223777770996, 'timing/data': 0.0003419799904804677, 'timing/model': 0.5947964889928699, 'learning_rate/qwen_vl_interface': 9.390257435597692e-06, 'learning_rate/action_model': 9.390257435597692e-05, 'epoch': 0.07}) +[2026-10-03 19:06:09,516][starVLA.training.train_starvla][INFO] - Step 900, Loss: {'train/loss': 0.3594960868358612, 'train/grad_norm_pre_clip': 26.771841049194336, 'timing/data': 0.00026780000189319253, 'timing/model': 0.39116397002362646, 'learning_rate/qwen_vl_interface': 9.3887638445861e-06, 'learning_rate/action_model': 9.3887638445861e-05, 'epoch': 0.07}) +[2026-10-03 19:06:10,213][starVLA.training.train_starvla][INFO] - Step 901, Loss: {'train/loss': 0.23518432676792145, 'train/grad_norm_pre_clip': 46.19290542602539, 'timing/data': 0.00020963000133633614, 'timing/model': 0.39239265900687315, 'learning_rate/qwen_vl_interface': 9.387268552285024e-06, 'learning_rate/action_model': 9.387268552285025e-05, 'epoch': 0.07}) +[2026-10-03 19:06:10,912][starVLA.training.train_starvla][INFO] - Step 902, Loss: {'train/loss': 0.05568516626954079, 'train/grad_norm_pre_clip': 27.444246292114258, 'timing/data': 0.00029599000117741525, 'timing/model': 0.3909607599780429, 'learning_rate/qwen_vl_interface': 9.385771559309125e-06, 'learning_rate/action_model': 9.385771559309125e-05, 'epoch': 0.07}) +[2026-10-03 19:06:11,607][starVLA.training.train_starvla][INFO] - Step 903, Loss: {'train/loss': 1.0694963932037354, 'train/grad_norm_pre_clip': 81.78602600097656, 'timing/data': 0.0003672489838209003, 'timing/model': 0.3891243500111159, 'learning_rate/qwen_vl_interface': 9.38427286627376e-06, 'learning_rate/action_model': 9.38427286627376e-05, 'epoch': 0.07}) +[2026-10-03 19:06:12,308][starVLA.training.train_starvla][INFO] - Step 904, Loss: {'train/loss': 0.165788933634758, 'train/grad_norm_pre_clip': 49.40330505371094, 'timing/data': 0.0003621700161602348, 'timing/model': 0.3911624500178732, 'learning_rate/qwen_vl_interface': 9.382772473794985e-06, 'learning_rate/action_model': 9.382772473794985e-05, 'epoch': 0.07}) +[2026-10-03 19:06:13,216][starVLA.training.train_starvla][INFO] - Step 905, Loss: {'train/loss': 0.03753936290740967, 'train/grad_norm_pre_clip': 35.294864654541016, 'timing/data': 0.0003825200255960226, 'timing/model': 0.3907513100130018, 'learning_rate/qwen_vl_interface': 9.381270382489554e-06, 'learning_rate/action_model': 9.381270382489554e-05, 'epoch': 0.07}) +[2026-10-03 19:06:13,915][starVLA.training.train_starvla][INFO] - Step 906, Loss: {'train/loss': 0.2997909188270569, 'train/grad_norm_pre_clip': 98.34754180908203, 'timing/data': 0.00020754997967742383, 'timing/model': 0.39138118899427354, 'learning_rate/qwen_vl_interface': 9.37976659297492e-06, 'learning_rate/action_model': 9.379766592974921e-05, 'epoch': 0.07}) +[2026-10-03 19:06:14,612][starVLA.training.train_starvla][INFO] - Step 907, Loss: {'train/loss': 0.42337632179260254, 'train/grad_norm_pre_clip': 60.30643844604492, 'timing/data': 0.0002476699883118272, 'timing/model': 0.39051560999359936, 'learning_rate/qwen_vl_interface': 9.378261105869237e-06, 'learning_rate/action_model': 9.378261105869237e-05, 'epoch': 0.07}) +[2026-10-03 19:06:15,311][starVLA.training.train_starvla][INFO] - Step 908, Loss: {'train/loss': 0.011696217581629753, 'train/grad_norm_pre_clip': 11.955140113830566, 'timing/data': 0.0003450100193731487, 'timing/model': 0.39141348900739104, 'learning_rate/qwen_vl_interface': 9.376753921791348e-06, 'learning_rate/action_model': 9.376753921791348e-05, 'epoch': 0.07}) +[2026-10-03 19:06:16,012][starVLA.training.train_starvla][INFO] - Step 909, Loss: {'train/loss': 0.08242081850767136, 'train/grad_norm_pre_clip': 18.078472137451172, 'timing/data': 0.00021075998665764928, 'timing/model': 0.3926580190018285, 'learning_rate/qwen_vl_interface': 9.375245041360805e-06, 'learning_rate/action_model': 9.375245041360804e-05, 'epoch': 0.07}) +[2026-10-03 19:06:16,916][starVLA.training.train_starvla][INFO] - Step 910, Loss: {'train/loss': 0.0765417218208313, 'train/grad_norm_pre_clip': 14.087811470031738, 'timing/data': 0.00032869999995455146, 'timing/model': 0.596713037986774, 'learning_rate/qwen_vl_interface': 9.37373446519785e-06, 'learning_rate/action_model': 9.373734465197849e-05, 'epoch': 0.07}) +[2026-10-03 19:06:17,615][starVLA.training.train_starvla][INFO] - Step 911, Loss: {'train/loss': 0.12030410021543503, 'train/grad_norm_pre_clip': 58.32060241699219, 'timing/data': 0.0002343399974051863, 'timing/model': 0.39092255898867734, 'learning_rate/qwen_vl_interface': 9.372222193923421e-06, 'learning_rate/action_model': 9.372222193923421e-05, 'epoch': 0.07}) +[2026-10-03 19:06:18,314][starVLA.training.train_starvla][INFO] - Step 912, Loss: {'train/loss': 0.039958175271749496, 'train/grad_norm_pre_clip': 14.22160530090332, 'timing/data': 0.0002104399900417775, 'timing/model': 0.3897535800060723, 'learning_rate/qwen_vl_interface': 9.370708228159158e-06, 'learning_rate/action_model': 9.370708228159159e-05, 'epoch': 0.07}) +[2026-10-03 19:06:19,015][starVLA.training.train_starvla][INFO] - Step 913, Loss: {'train/loss': 0.05631078779697418, 'train/grad_norm_pre_clip': 25.19222640991211, 'timing/data': 0.00024973999825306237, 'timing/model': 0.3928262389963493, 'learning_rate/qwen_vl_interface': 9.369192568527398e-06, 'learning_rate/action_model': 9.369192568527398e-05, 'epoch': 0.07}) +[2026-10-03 19:06:19,714][starVLA.training.train_starvla][INFO] - Step 914, Loss: {'train/loss': 0.18820056319236755, 'train/grad_norm_pre_clip': 95.40737915039062, 'timing/data': 0.0003705700219143182, 'timing/model': 0.39054509898414835, 'learning_rate/qwen_vl_interface': 9.367675215651168e-06, 'learning_rate/action_model': 9.367675215651168e-05, 'epoch': 0.07}) +[2026-10-03 19:06:20,411][starVLA.training.train_starvla][INFO] - Step 915, Loss: {'train/loss': 0.03247319906949997, 'train/grad_norm_pre_clip': 11.027836799621582, 'timing/data': 0.00026418999186716974, 'timing/model': 0.3900343799905386, 'learning_rate/qwen_vl_interface': 9.366156170154194e-06, 'learning_rate/action_model': 9.366156170154194e-05, 'epoch': 0.07}) +[2026-10-03 19:06:21,311][starVLA.training.train_starvla][INFO] - Step 916, Loss: {'train/loss': 0.43765193223953247, 'train/grad_norm_pre_clip': 72.07316589355469, 'timing/data': 0.00020305000361986458, 'timing/model': 0.3892315900011454, 'learning_rate/qwen_vl_interface': 9.3646354326609e-06, 'learning_rate/action_model': 9.364635432660899e-05, 'epoch': 0.07}) +[2026-10-03 19:06:22,011][starVLA.training.train_starvla][INFO] - Step 917, Loss: {'train/loss': 0.013526328839361668, 'train/grad_norm_pre_clip': 8.08215045928955, 'timing/data': 0.00020524000865407288, 'timing/model': 0.39096697900095023, 'learning_rate/qwen_vl_interface': 9.3631130037964e-06, 'learning_rate/action_model': 9.363113003796401e-05, 'epoch': 0.07}) +[2026-10-03 19:06:22,707][starVLA.training.train_starvla][INFO] - Step 918, Loss: {'train/loss': 0.004316404461860657, 'train/grad_norm_pre_clip': 1.0276979207992554, 'timing/data': 0.00020165002206340432, 'timing/model': 0.3903656309994403, 'learning_rate/qwen_vl_interface': 9.361588884186513e-06, 'learning_rate/action_model': 9.361588884186513e-05, 'epoch': 0.07}) +[2026-10-03 19:06:23,406][starVLA.training.train_starvla][INFO] - Step 919, Loss: {'train/loss': 0.021917063742876053, 'train/grad_norm_pre_clip': 15.168447494506836, 'timing/data': 0.0002353399759158492, 'timing/model': 0.3901983889809344, 'learning_rate/qwen_vl_interface': 9.360063074457744e-06, 'learning_rate/action_model': 9.360063074457743e-05, 'epoch': 0.07}) +[2026-10-03 19:06:24,110][starVLA.training.train_starvla][INFO] - Step 920, Loss: {'train/loss': 0.09395534545183182, 'train/grad_norm_pre_clip': 46.4607048034668, 'timing/data': 0.00021620001643896103, 'timing/model': 0.3932969289890025, 'learning_rate/qwen_vl_interface': 9.358535575237296e-06, 'learning_rate/action_model': 9.358535575237295e-05, 'epoch': 0.07}) +[2026-10-03 19:06:25,018][starVLA.training.train_starvla][INFO] - Step 921, Loss: {'train/loss': 0.0038115112110972404, 'train/grad_norm_pre_clip': 1.5516360998153687, 'timing/data': 0.00030949999927543104, 'timing/model': 0.3918386499863118, 'learning_rate/qwen_vl_interface': 9.357006387153065e-06, 'learning_rate/action_model': 9.357006387153065e-05, 'epoch': 0.07}) +[2026-10-03 19:06:25,716][starVLA.training.train_starvla][INFO] - Step 922, Loss: {'train/loss': 0.19772188365459442, 'train/grad_norm_pre_clip': 48.54584503173828, 'timing/data': 0.0003056699933949858, 'timing/model': 0.39095003900001757, 'learning_rate/qwen_vl_interface': 9.355475510833643e-06, 'learning_rate/action_model': 9.355475510833643e-05, 'epoch': 0.07}) +[2026-10-03 19:06:26,414][starVLA.training.train_starvla][INFO] - Step 923, Loss: {'train/loss': 1.3474736213684082, 'train/grad_norm_pre_clip': 149.0211639404297, 'timing/data': 0.0003273999900557101, 'timing/model': 0.39010779999080114, 'learning_rate/qwen_vl_interface': 9.353942946908318e-06, 'learning_rate/action_model': 9.353942946908317e-05, 'epoch': 0.07}) +[2026-10-03 19:06:27,115][starVLA.training.train_starvla][INFO] - Step 924, Loss: {'train/loss': 0.18088091909885406, 'train/grad_norm_pre_clip': 47.818206787109375, 'timing/data': 0.0003269500157330185, 'timing/model': 0.39126952999504283, 'learning_rate/qwen_vl_interface': 9.352408696007065e-06, 'learning_rate/action_model': 9.352408696007065e-05, 'epoch': 0.07}) +[2026-10-03 19:06:27,818][starVLA.training.train_starvla][INFO] - Step 925, Loss: {'train/loss': 0.7466477155685425, 'train/grad_norm_pre_clip': 137.6560516357422, 'timing/data': 0.0003005000180564821, 'timing/model': 0.394137530005537, 'learning_rate/qwen_vl_interface': 9.350872758760558e-06, 'learning_rate/action_model': 9.350872758760559e-05, 'epoch': 0.07}) +[2026-10-03 19:06:28,724][starVLA.training.train_starvla][INFO] - Step 926, Loss: {'train/loss': 0.16782529652118683, 'train/grad_norm_pre_clip': 215.59620666503906, 'timing/data': 0.00020838002092204988, 'timing/model': 0.59872659802204, 'learning_rate/qwen_vl_interface': 9.349335135800164e-06, 'learning_rate/action_model': 9.349335135800163e-05, 'epoch': 0.07}) +[2026-10-03 19:06:29,424][starVLA.training.train_starvla][INFO] - Step 927, Loss: {'train/loss': 0.5224926471710205, 'train/grad_norm_pre_clip': 43.31730651855469, 'timing/data': 0.00047374999849125743, 'timing/model': 0.39225727002485655, 'learning_rate/qwen_vl_interface': 9.347795827757942e-06, 'learning_rate/action_model': 9.347795827757941e-05, 'epoch': 0.07}) +[2026-10-03 19:06:30,126][starVLA.training.train_starvla][INFO] - Step 928, Loss: {'train/loss': 1.1712634563446045, 'train/grad_norm_pre_clip': 46.60813522338867, 'timing/data': 0.00021764999837614596, 'timing/model': 0.3923732800176367, 'learning_rate/qwen_vl_interface': 9.346254835266643e-06, 'learning_rate/action_model': 9.346254835266642e-05, 'epoch': 0.07}) +[2026-10-03 19:06:30,828][starVLA.training.train_starvla][INFO] - Step 929, Loss: {'train/loss': 0.2326037734746933, 'train/grad_norm_pre_clip': 72.64826202392578, 'timing/data': 0.00020095001673325896, 'timing/model': 0.3938611700141337, 'learning_rate/qwen_vl_interface': 9.344712158959709e-06, 'learning_rate/action_model': 9.344712158959709e-05, 'epoch': 0.07}) +[2026-10-03 19:06:31,529][starVLA.training.train_starvla][INFO] - Step 930, Loss: {'train/loss': 0.5945841670036316, 'train/grad_norm_pre_clip': 226.44906616210938, 'timing/data': 0.00020372000290080905, 'timing/model': 0.39325243001803756, 'learning_rate/qwen_vl_interface': 9.343167799471278e-06, 'learning_rate/action_model': 9.343167799471277e-05, 'epoch': 0.07}) +[2026-10-03 19:06:32,229][starVLA.training.train_starvla][INFO] - Step 931, Loss: {'train/loss': 0.35915085673332214, 'train/grad_norm_pre_clip': 76.91250610351562, 'timing/data': 0.00022976999753154814, 'timing/model': 0.3923914290207904, 'learning_rate/qwen_vl_interface': 9.341621757436177e-06, 'learning_rate/action_model': 9.341621757436177e-05, 'epoch': 0.07}) +[2026-10-03 19:06:33,132][starVLA.training.train_starvla][INFO] - Step 932, Loss: {'train/loss': 0.1942753940820694, 'train/grad_norm_pre_clip': 133.2504119873047, 'timing/data': 0.0002062699932139367, 'timing/model': 0.39177307998761535, 'learning_rate/qwen_vl_interface': 9.340074033489924e-06, 'learning_rate/action_model': 9.340074033489923e-05, 'epoch': 0.07}) +[2026-10-03 19:06:33,832][starVLA.training.train_starvla][INFO] - Step 933, Loss: {'train/loss': 0.2762216627597809, 'train/grad_norm_pre_clip': 86.02222442626953, 'timing/data': 0.00020405001123435795, 'timing/model': 0.39176532000419684, 'learning_rate/qwen_vl_interface': 9.338524628268734e-06, 'learning_rate/action_model': 9.338524628268733e-05, 'epoch': 0.07}) +[2026-10-03 19:06:34,532][starVLA.training.train_starvla][INFO] - Step 934, Loss: {'train/loss': 0.7095680832862854, 'train/grad_norm_pre_clip': 74.24534606933594, 'timing/data': 0.0003365199954714626, 'timing/model': 0.3903570600086823, 'learning_rate/qwen_vl_interface': 9.336973542409505e-06, 'learning_rate/action_model': 9.336973542409504e-05, 'epoch': 0.07}) +[2026-10-03 19:06:35,235][starVLA.training.train_starvla][INFO] - Step 935, Loss: {'train/loss': 0.7131884098052979, 'train/grad_norm_pre_clip': 110.6914291381836, 'timing/data': 0.00020314997527748346, 'timing/model': 0.3924026800086722, 'learning_rate/qwen_vl_interface': 9.33542077654983e-06, 'learning_rate/action_model': 9.33542077654983e-05, 'epoch': 0.07}) +[2026-10-03 19:06:35,935][starVLA.training.train_starvla][INFO] - Step 936, Loss: {'train/loss': 0.00407965574413538, 'train/grad_norm_pre_clip': 2.5001676082611084, 'timing/data': 0.00036885999725200236, 'timing/model': 0.389993949997006, 'learning_rate/qwen_vl_interface': 9.333866331327995e-06, 'learning_rate/action_model': 9.333866331327995e-05, 'epoch': 0.07}) +[2026-10-03 19:06:36,836][starVLA.training.train_starvla][INFO] - Step 937, Loss: {'train/loss': 0.09423738718032837, 'train/grad_norm_pre_clip': 44.242576599121094, 'timing/data': 0.00020542999845929444, 'timing/model': 0.39292055999976583, 'learning_rate/qwen_vl_interface': 9.332310207382972e-06, 'learning_rate/action_model': 9.332310207382972e-05, 'epoch': 0.07}) +[2026-10-03 19:06:37,537][starVLA.training.train_starvla][INFO] - Step 938, Loss: {'train/loss': 0.0047423564828932285, 'train/grad_norm_pre_clip': 2.930494785308838, 'timing/data': 0.0003447099879849702, 'timing/model': 0.39243490001535974, 'learning_rate/qwen_vl_interface': 9.330752405354426e-06, 'learning_rate/action_model': 9.330752405354426e-05, 'epoch': 0.07}) +[2026-10-03 19:06:38,237][starVLA.training.train_starvla][INFO] - Step 939, Loss: {'train/loss': 0.4028133153915405, 'train/grad_norm_pre_clip': 61.83051681518555, 'timing/data': 0.0002441900142002851, 'timing/model': 0.39073909900616854, 'learning_rate/qwen_vl_interface': 9.32919292588271e-06, 'learning_rate/action_model': 9.32919292588271e-05, 'epoch': 0.07}) +[2026-10-03 19:06:38,941][starVLA.training.train_starvla][INFO] - Step 940, Loss: {'train/loss': 0.01521463692188263, 'train/grad_norm_pre_clip': 13.135504722595215, 'timing/data': 0.0003577900060918182, 'timing/model': 0.39214132897905074, 'learning_rate/qwen_vl_interface': 9.327631769608864e-06, 'learning_rate/action_model': 9.327631769608864e-05, 'epoch': 0.07}) +[2026-10-03 19:06:39,644][starVLA.training.train_starvla][INFO] - Step 941, Loss: {'train/loss': 0.30834734439849854, 'train/grad_norm_pre_clip': 44.39357376098633, 'timing/data': 0.000209019985049963, 'timing/model': 0.3944257090042811, 'learning_rate/qwen_vl_interface': 9.326068937174627e-06, 'learning_rate/action_model': 9.326068937174626e-05, 'epoch': 0.07}) +[2026-10-03 19:06:40,545][starVLA.training.train_starvla][INFO] - Step 942, Loss: {'train/loss': 0.21493452787399292, 'train/grad_norm_pre_clip': 61.163185119628906, 'timing/data': 0.00033465001615695655, 'timing/model': 0.5919875780236907, 'learning_rate/qwen_vl_interface': 9.324504429222412e-06, 'learning_rate/action_model': 9.324504429222412e-05, 'epoch': 0.07}) +[2026-10-03 19:06:41,247][starVLA.training.train_starvla][INFO] - Step 943, Loss: {'train/loss': 0.017772572115063667, 'train/grad_norm_pre_clip': 11.542170524597168, 'timing/data': 0.00029848000849597156, 'timing/model': 0.39186046898248605, 'learning_rate/qwen_vl_interface': 9.322938246395336e-06, 'learning_rate/action_model': 9.322938246395337e-05, 'epoch': 0.07}) +[2026-10-03 19:06:41,950][starVLA.training.train_starvla][INFO] - Step 944, Loss: {'train/loss': 0.029212772846221924, 'train/grad_norm_pre_clip': 7.924095630645752, 'timing/data': 0.00021130999084562063, 'timing/model': 0.3934207590064034, 'learning_rate/qwen_vl_interface': 9.321370389337196e-06, 'learning_rate/action_model': 9.321370389337196e-05, 'epoch': 0.07}) +[2026-10-03 19:06:42,650][starVLA.training.train_starvla][INFO] - Step 945, Loss: {'train/loss': 0.24179047346115112, 'train/grad_norm_pre_clip': 173.77972412109375, 'timing/data': 0.0003132399870082736, 'timing/model': 0.39300519999233074, 'learning_rate/qwen_vl_interface': 9.319800858692478e-06, 'learning_rate/action_model': 9.319800858692476e-05, 'epoch': 0.07}) +[2026-10-03 19:06:43,346][starVLA.training.train_starvla][INFO] - Step 946, Loss: {'train/loss': 0.0521479956805706, 'train/grad_norm_pre_clip': 18.165691375732422, 'timing/data': 0.00023174998932518065, 'timing/model': 0.3896786799887195, 'learning_rate/qwen_vl_interface': 9.318229655106357e-06, 'learning_rate/action_model': 9.318229655106356e-05, 'epoch': 0.07}) +[2026-10-03 19:06:44,248][starVLA.training.train_starvla][INFO] - Step 947, Loss: {'train/loss': 0.21344488859176636, 'train/grad_norm_pre_clip': 35.922828674316406, 'timing/data': 0.00035072999889962375, 'timing/model': 0.5938713780196849, 'learning_rate/qwen_vl_interface': 9.316656779224693e-06, 'learning_rate/action_model': 9.316656779224693e-05, 'epoch': 0.07}) +[2026-10-03 19:06:44,947][starVLA.training.train_starvla][INFO] - Step 948, Loss: {'train/loss': 0.1892716884613037, 'train/grad_norm_pre_clip': 49.25798416137695, 'timing/data': 0.0003705100098159164, 'timing/model': 0.39054013998247683, 'learning_rate/qwen_vl_interface': 9.31508223169404e-06, 'learning_rate/action_model': 9.31508223169404e-05, 'epoch': 0.07}) +[2026-10-03 19:06:45,647][starVLA.training.train_starvla][INFO] - Step 949, Loss: {'train/loss': 0.4190417528152466, 'train/grad_norm_pre_clip': 34.786746978759766, 'timing/data': 0.00020631999359466136, 'timing/model': 0.39322990001528524, 'learning_rate/qwen_vl_interface': 9.313506013161633e-06, 'learning_rate/action_model': 9.313506013161633e-05, 'epoch': 0.07}) +[2026-10-03 19:06:46,347][starVLA.training.train_starvla][INFO] - Step 950, Loss: {'train/loss': 0.06057795509696007, 'train/grad_norm_pre_clip': 26.725736618041992, 'timing/data': 0.00020885001868009567, 'timing/model': 0.3906115090067033, 'learning_rate/qwen_vl_interface': 9.311928124275396e-06, 'learning_rate/action_model': 9.311928124275395e-05, 'epoch': 0.07}) +[2026-10-03 19:06:47,046][starVLA.training.train_starvla][INFO] - Step 951, Loss: {'train/loss': 0.03989331051707268, 'train/grad_norm_pre_clip': 11.21744441986084, 'timing/data': 0.0003535799914970994, 'timing/model': 0.39029049000237137, 'learning_rate/qwen_vl_interface': 9.310348565683938e-06, 'learning_rate/action_model': 9.310348565683939e-05, 'epoch': 0.07}) +[2026-10-03 19:06:47,952][starVLA.training.train_starvla][INFO] - Step 952, Loss: {'train/loss': 0.23698918521404266, 'train/grad_norm_pre_clip': 263.0676574707031, 'timing/data': 0.00027042999863624573, 'timing/model': 0.5986330379964784, 'learning_rate/qwen_vl_interface': 9.30876733803656e-06, 'learning_rate/action_model': 9.308767338036559e-05, 'epoch': 0.07}) +[2026-10-03 19:06:48,654][starVLA.training.train_starvla][INFO] - Step 953, Loss: {'train/loss': 0.35889172554016113, 'train/grad_norm_pre_clip': 37.4633903503418, 'timing/data': 0.00020449000294320285, 'timing/model': 0.3934266389987897, 'learning_rate/qwen_vl_interface': 9.30718444198324e-06, 'learning_rate/action_model': 9.307184441983239e-05, 'epoch': 0.07}) +[2026-10-03 19:06:49,353][starVLA.training.train_starvla][INFO] - Step 954, Loss: {'train/loss': 1.015275478363037, 'train/grad_norm_pre_clip': 90.72689056396484, 'timing/data': 0.00020462999236769974, 'timing/model': 0.3909743610129226, 'learning_rate/qwen_vl_interface': 9.30559987817465e-06, 'learning_rate/action_model': 9.30559987817465e-05, 'epoch': 0.07}) +[2026-10-03 19:06:50,052][starVLA.training.train_starvla][INFO] - Step 955, Loss: {'train/loss': 0.030765263363718987, 'train/grad_norm_pre_clip': 6.09912109375, 'timing/data': 0.00020655000116676092, 'timing/model': 0.3920881189987995, 'learning_rate/qwen_vl_interface': 9.304013647262143e-06, 'learning_rate/action_model': 9.304013647262143e-05, 'epoch': 0.07}) +[2026-10-03 19:06:50,751][starVLA.training.train_starvla][INFO] - Step 956, Loss: {'train/loss': 0.009086010977625847, 'train/grad_norm_pre_clip': 4.2407307624816895, 'timing/data': 0.000328630005242303, 'timing/model': 0.39169562997994944, 'learning_rate/qwen_vl_interface': 9.302425749897761e-06, 'learning_rate/action_model': 9.302425749897761e-05, 'epoch': 0.07}) +[2026-10-03 19:06:51,451][starVLA.training.train_starvla][INFO] - Step 957, Loss: {'train/loss': 0.5145227909088135, 'train/grad_norm_pre_clip': 56.52785873413086, 'timing/data': 0.00029712001560255885, 'timing/model': 0.39267308998387307, 'learning_rate/qwen_vl_interface': 9.300836186734225e-06, 'learning_rate/action_model': 9.300836186734226e-05, 'epoch': 0.07}) +[2026-10-03 19:06:52,358][starVLA.training.train_starvla][INFO] - Step 958, Loss: {'train/loss': 0.7864190936088562, 'train/grad_norm_pre_clip': 64.35986328125, 'timing/data': 0.000215259991819039, 'timing/model': 0.3910041000053752, 'learning_rate/qwen_vl_interface': 9.299244958424947e-06, 'learning_rate/action_model': 9.299244958424946e-05, 'epoch': 0.07}) +[2026-10-03 19:06:53,059][starVLA.training.train_starvla][INFO] - Step 959, Loss: {'train/loss': 0.2820861041545868, 'train/grad_norm_pre_clip': 48.44212341308594, 'timing/data': 0.00021152000408619642, 'timing/model': 0.3920644389872905, 'learning_rate/qwen_vl_interface': 9.29765206562402e-06, 'learning_rate/action_model': 9.297652065624018e-05, 'epoch': 0.07}) +[2026-10-03 19:06:53,758][starVLA.training.train_starvla][INFO] - Step 960, Loss: {'train/loss': 0.07211006432771683, 'train/grad_norm_pre_clip': 89.02960968017578, 'timing/data': 0.0003386500175110996, 'timing/model': 0.391867819998879, 'learning_rate/qwen_vl_interface': 9.296057508986221e-06, 'learning_rate/action_model': 9.296057508986221e-05, 'epoch': 0.07}) +[2026-10-03 19:06:54,458][starVLA.training.train_starvla][INFO] - Step 961, Loss: {'train/loss': 0.24273966252803802, 'train/grad_norm_pre_clip': 20.2846736907959, 'timing/data': 0.00028802998713217676, 'timing/model': 0.3932775490102358, 'learning_rate/qwen_vl_interface': 9.294461289167013e-06, 'learning_rate/action_model': 9.294461289167014e-05, 'epoch': 0.07}) +[2026-10-03 19:06:55,157][starVLA.training.train_starvla][INFO] - Step 962, Loss: {'train/loss': 0.03529834747314453, 'train/grad_norm_pre_clip': 23.497350692749023, 'timing/data': 0.000315290002617985, 'timing/model': 0.39064264902845025, 'learning_rate/qwen_vl_interface': 9.292863406822543e-06, 'learning_rate/action_model': 9.292863406822542e-05, 'epoch': 0.07}) +[2026-10-03 19:06:56,054][starVLA.training.train_starvla][INFO] - Step 963, Loss: {'train/loss': 0.7147592306137085, 'train/grad_norm_pre_clip': 45.100311279296875, 'timing/data': 0.00021692001610063016, 'timing/model': 0.5878052900079638, 'learning_rate/qwen_vl_interface': 9.291263862609638e-06, 'learning_rate/action_model': 9.291263862609638e-05, 'epoch': 0.07}) +[2026-10-03 19:06:56,753][starVLA.training.train_starvla][INFO] - Step 964, Loss: {'train/loss': 0.005903773941099644, 'train/grad_norm_pre_clip': 1.457188367843628, 'timing/data': 0.00033161998726427555, 'timing/model': 0.3896338590129744, 'learning_rate/qwen_vl_interface': 9.28966265718581e-06, 'learning_rate/action_model': 9.28966265718581e-05, 'epoch': 0.07}) +[2026-10-03 19:06:57,454][starVLA.training.train_starvla][INFO] - Step 965, Loss: {'train/loss': 0.02184847928583622, 'train/grad_norm_pre_clip': 6.287976264953613, 'timing/data': 0.0002076700038742274, 'timing/model': 0.39262430102098733, 'learning_rate/qwen_vl_interface': 9.288059791209256e-06, 'learning_rate/action_model': 9.288059791209256e-05, 'epoch': 0.07}) +[2026-10-03 19:06:58,152][starVLA.training.train_starvla][INFO] - Step 966, Loss: {'train/loss': 0.011526334099471569, 'train/grad_norm_pre_clip': 5.981556415557861, 'timing/data': 0.00023920999956317246, 'timing/model': 0.38976939997519366, 'learning_rate/qwen_vl_interface': 9.286455265338853e-06, 'learning_rate/action_model': 9.286455265338853e-05, 'epoch': 0.07}) +[2026-10-03 19:06:58,850][starVLA.training.train_starvla][INFO] - Step 967, Loss: {'train/loss': 0.36764195561408997, 'train/grad_norm_pre_clip': 38.83686065673828, 'timing/data': 0.0003339399991091341, 'timing/model': 0.3912431600037962, 'learning_rate/qwen_vl_interface': 9.284849080234161e-06, 'learning_rate/action_model': 9.28484908023416e-05, 'epoch': 0.07}) +[2026-10-03 19:06:59,551][starVLA.training.train_starvla][INFO] - Step 968, Loss: {'train/loss': 0.06241445615887642, 'train/grad_norm_pre_clip': 46.42253875732422, 'timing/data': 0.0003299000090919435, 'timing/model': 0.3906331299804151, 'learning_rate/qwen_vl_interface': 9.283241236555421e-06, 'learning_rate/action_model': 9.283241236555421e-05, 'epoch': 0.07}) +[2026-10-03 19:07:00,454][starVLA.training.train_starvla][INFO] - Step 969, Loss: {'train/loss': 0.004967098124325275, 'train/grad_norm_pre_clip': 1.646695852279663, 'timing/data': 0.00022899001487530768, 'timing/model': 0.392447460006224, 'learning_rate/qwen_vl_interface': 9.281631734963557e-06, 'learning_rate/action_model': 9.281631734963557e-05, 'epoch': 0.07}) +[2026-10-03 19:07:01,154][starVLA.training.train_starvla][INFO] - Step 970, Loss: {'train/loss': 0.22403042018413544, 'train/grad_norm_pre_clip': 42.063663482666016, 'timing/data': 0.00031132998992688954, 'timing/model': 0.39098356998874806, 'learning_rate/qwen_vl_interface': 9.280020576120176e-06, 'learning_rate/action_model': 9.280020576120176e-05, 'epoch': 0.07}) +[2026-10-03 19:07:01,855][starVLA.training.train_starvla][INFO] - Step 971, Loss: {'train/loss': 0.30859196186065674, 'train/grad_norm_pre_clip': 15.260150909423828, 'timing/data': 0.00034955900628119707, 'timing/model': 0.39145791000919417, 'learning_rate/qwen_vl_interface': 9.278407760687562e-06, 'learning_rate/action_model': 9.278407760687562e-05, 'epoch': 0.07}) +[2026-10-03 19:07:02,554][starVLA.training.train_starvla][INFO] - Step 972, Loss: {'train/loss': 0.34239134192466736, 'train/grad_norm_pre_clip': 207.82571411132812, 'timing/data': 0.0003243100072722882, 'timing/model': 0.3918223800137639, 'learning_rate/qwen_vl_interface': 9.276793289328686e-06, 'learning_rate/action_model': 9.276793289328685e-05, 'epoch': 0.07}) +[2026-10-03 19:07:03,256][starVLA.training.train_starvla][INFO] - Step 973, Loss: {'train/loss': 0.8743209838867188, 'train/grad_norm_pre_clip': 31.594127655029297, 'timing/data': 0.000360800011549145, 'timing/model': 0.39286049097427167, 'learning_rate/qwen_vl_interface': 9.27517716270719e-06, 'learning_rate/action_model': 9.275177162707191e-05, 'epoch': 0.07}) +[2026-10-03 19:07:04,160][starVLA.training.train_starvla][INFO] - Step 974, Loss: {'train/loss': 0.3772420287132263, 'train/grad_norm_pre_clip': 55.047542572021484, 'timing/data': 0.00020429998403415084, 'timing/model': 0.39176663898979314, 'learning_rate/qwen_vl_interface': 9.273559381487409e-06, 'learning_rate/action_model': 9.273559381487409e-05, 'epoch': 0.07}) +[2026-10-03 19:07:04,859][starVLA.training.train_starvla][INFO] - Step 975, Loss: {'train/loss': 0.6479250192642212, 'train/grad_norm_pre_clip': 124.99771118164062, 'timing/data': 0.00021552000544033945, 'timing/model': 0.3920487799914554, 'learning_rate/qwen_vl_interface': 9.27193994633435e-06, 'learning_rate/action_model': 9.27193994633435e-05, 'epoch': 0.07}) +[2026-10-03 19:07:05,559][starVLA.training.train_starvla][INFO] - Step 976, Loss: {'train/loss': 0.0012129865353927016, 'train/grad_norm_pre_clip': 0.21152405440807343, 'timing/data': 0.00032761000329628587, 'timing/model': 0.39042451899149455, 'learning_rate/qwen_vl_interface': 9.2703188579137e-06, 'learning_rate/action_model': 9.2703188579137e-05, 'epoch': 0.07}) +[2026-10-03 19:07:06,258][starVLA.training.train_starvla][INFO] - Step 977, Loss: {'train/loss': 0.20257796347141266, 'train/grad_norm_pre_clip': 75.18855285644531, 'timing/data': 0.00027541001327335835, 'timing/model': 0.39245436000055633, 'learning_rate/qwen_vl_interface': 9.26869611689183e-06, 'learning_rate/action_model': 9.26869611689183e-05, 'epoch': 0.07}) +[2026-10-03 19:07:06,955][starVLA.training.train_starvla][INFO] - Step 978, Loss: {'train/loss': 0.0415203720331192, 'train/grad_norm_pre_clip': 23.260547637939453, 'timing/data': 0.0003572799905668944, 'timing/model': 0.3899945890007075, 'learning_rate/qwen_vl_interface': 9.267071723935784e-06, 'learning_rate/action_model': 9.267071723935784e-05, 'epoch': 0.07}) +[2026-10-03 19:07:07,847][starVLA.training.train_starvla][INFO] - Step 979, Loss: {'train/loss': 0.4834533929824829, 'train/grad_norm_pre_clip': 77.45233917236328, 'timing/data': 0.000354900024831295, 'timing/model': 0.5857272499997634, 'learning_rate/qwen_vl_interface': 9.26544567971329e-06, 'learning_rate/action_model': 9.26544567971329e-05, 'epoch': 0.07}) +[2026-10-03 19:07:08,543][starVLA.training.train_starvla][INFO] - Step 980, Loss: {'train/loss': 0.33653390407562256, 'train/grad_norm_pre_clip': 16.12618064880371, 'timing/data': 0.0002493700012564659, 'timing/model': 0.3882160309876781, 'learning_rate/qwen_vl_interface': 9.263817984892756e-06, 'learning_rate/action_model': 9.263817984892757e-05, 'epoch': 0.07}) +[2026-10-03 19:07:09,239][starVLA.training.train_starvla][INFO] - Step 981, Loss: {'train/loss': 0.0060323672369122505, 'train/grad_norm_pre_clip': 0.7453797459602356, 'timing/data': 0.000297840015264228, 'timing/model': 0.39290787000209093, 'learning_rate/qwen_vl_interface': 9.262188640143262e-06, 'learning_rate/action_model': 9.262188640143262e-05, 'epoch': 0.07}) +[2026-10-03 19:07:09,936][starVLA.training.train_starvla][INFO] - Step 982, Loss: {'train/loss': 0.20087400078773499, 'train/grad_norm_pre_clip': 71.24871826171875, 'timing/data': 0.0003330899926368147, 'timing/model': 0.3891098010062706, 'learning_rate/qwen_vl_interface': 9.260557646134572e-06, 'learning_rate/action_model': 9.260557646134572e-05, 'epoch': 0.07}) +[2026-10-03 19:07:10,635][starVLA.training.train_starvla][INFO] - Step 983, Loss: {'train/loss': 0.4346734285354614, 'train/grad_norm_pre_clip': 55.55470275878906, 'timing/data': 0.00033860999974422157, 'timing/model': 0.3907137700007297, 'learning_rate/qwen_vl_interface': 9.258925003537127e-06, 'learning_rate/action_model': 9.258925003537126e-05, 'epoch': 0.07}) +[2026-10-03 19:07:11,338][starVLA.training.train_starvla][INFO] - Step 984, Loss: {'train/loss': 0.2286197692155838, 'train/grad_norm_pre_clip': 24.7890625, 'timing/data': 0.0003366499731782824, 'timing/model': 0.39076123002450913, 'learning_rate/qwen_vl_interface': 9.257290713022041e-06, 'learning_rate/action_model': 9.257290713022041e-05, 'epoch': 0.07}) +[2026-10-03 19:07:12,227][starVLA.training.train_starvla][INFO] - Step 985, Loss: {'train/loss': 0.5013623833656311, 'train/grad_norm_pre_clip': 52.13047790527344, 'timing/data': 0.0003069700032938272, 'timing/model': 0.3921238790208008, 'learning_rate/qwen_vl_interface': 9.255654775261112e-06, 'learning_rate/action_model': 9.255654775261112e-05, 'epoch': 0.07}) +[2026-10-03 19:07:12,924][starVLA.training.train_starvla][INFO] - Step 986, Loss: {'train/loss': 0.13183319568634033, 'train/grad_norm_pre_clip': 43.77733612060547, 'timing/data': 0.00034093999420292675, 'timing/model': 0.3889245500031393, 'learning_rate/qwen_vl_interface': 9.254017190926813e-06, 'learning_rate/action_model': 9.254017190926813e-05, 'epoch': 0.07}) +[2026-10-03 19:07:13,626][starVLA.training.train_starvla][INFO] - Step 987, Loss: {'train/loss': 0.09686599671840668, 'train/grad_norm_pre_clip': 29.288000106811523, 'timing/data': 0.00036142999306321144, 'timing/model': 0.3917253389954567, 'learning_rate/qwen_vl_interface': 9.25237796069229e-06, 'learning_rate/action_model': 9.25237796069229e-05, 'epoch': 0.07}) +[2026-10-03 19:07:14,326][starVLA.training.train_starvla][INFO] - Step 988, Loss: {'train/loss': 0.05707132816314697, 'train/grad_norm_pre_clip': 19.064756393432617, 'timing/data': 0.0003885200130753219, 'timing/model': 0.3923330399848055, 'learning_rate/qwen_vl_interface': 9.250737085231371e-06, 'learning_rate/action_model': 9.250737085231372e-05, 'epoch': 0.07}) +[2026-10-03 19:07:15,026][starVLA.training.train_starvla][INFO] - Step 989, Loss: {'train/loss': 0.7362357378005981, 'train/grad_norm_pre_clip': 86.66069030761719, 'timing/data': 0.0003132499987259507, 'timing/model': 0.3914267100044526, 'learning_rate/qwen_vl_interface': 9.24909456521856e-06, 'learning_rate/action_model': 9.249094565218559e-05, 'epoch': 0.07}) +[2026-10-03 19:07:15,925][starVLA.training.train_starvla][INFO] - Step 990, Loss: {'train/loss': 0.39289867877960205, 'train/grad_norm_pre_clip': 126.00401306152344, 'timing/data': 0.00036529998760670424, 'timing/model': 0.39031243001227267, 'learning_rate/qwen_vl_interface': 9.247450401329027e-06, 'learning_rate/action_model': 9.247450401329027e-05, 'epoch': 0.07}) +[2026-10-03 19:07:16,625][starVLA.training.train_starvla][INFO] - Step 991, Loss: {'train/loss': 0.36680930852890015, 'train/grad_norm_pre_clip': 35.56673812866211, 'timing/data': 0.00020916000357829034, 'timing/model': 0.3912196689925622, 'learning_rate/qwen_vl_interface': 9.245804594238637e-06, 'learning_rate/action_model': 9.245804594238636e-05, 'epoch': 0.07}) +[2026-10-03 19:07:17,324][starVLA.training.train_starvla][INFO] - Step 992, Loss: {'train/loss': 0.2978089451789856, 'train/grad_norm_pre_clip': 65.8998031616211, 'timing/data': 0.00036819998058490455, 'timing/model': 0.3919564600219019, 'learning_rate/qwen_vl_interface': 9.244157144623909e-06, 'learning_rate/action_model': 9.244157144623909e-05, 'epoch': 0.07}) +[2026-10-03 19:07:18,024][starVLA.training.train_starvla][INFO] - Step 993, Loss: {'train/loss': 0.038791950792074203, 'train/grad_norm_pre_clip': 6.51194429397583, 'timing/data': 0.0003257500065956265, 'timing/model': 0.39233561902074143, 'learning_rate/qwen_vl_interface': 9.242508053162051e-06, 'learning_rate/action_model': 9.242508053162051e-05, 'epoch': 0.07}) +[2026-10-03 19:07:18,723][starVLA.training.train_starvla][INFO] - Step 994, Loss: {'train/loss': 0.14771708846092224, 'train/grad_norm_pre_clip': 38.3768424987793, 'timing/data': 0.00023331999545916915, 'timing/model': 0.39005540098878555, 'learning_rate/qwen_vl_interface': 9.240857320530944e-06, 'learning_rate/action_model': 9.240857320530943e-05, 'epoch': 0.07}) +[2026-10-03 19:07:19,624][starVLA.training.train_starvla][INFO] - Step 995, Loss: {'train/loss': 0.00744378287345171, 'train/grad_norm_pre_clip': 2.161301612854004, 'timing/data': 0.00037573001463897526, 'timing/model': 0.5922290500020608, 'learning_rate/qwen_vl_interface': 9.23920494740914e-06, 'learning_rate/action_model': 9.23920494740914e-05, 'epoch': 0.07}) +[2026-10-03 19:07:20,321][starVLA.training.train_starvla][INFO] - Step 996, Loss: {'train/loss': 0.005113230086863041, 'train/grad_norm_pre_clip': 0.788853108882904, 'timing/data': 0.00020640000002458692, 'timing/model': 0.3893656799918972, 'learning_rate/qwen_vl_interface': 9.237550934475868e-06, 'learning_rate/action_model': 9.237550934475867e-05, 'epoch': 0.07}) +[2026-10-03 19:07:21,020][starVLA.training.train_starvla][INFO] - Step 997, Loss: {'train/loss': 0.28735366463661194, 'train/grad_norm_pre_clip': 52.571388244628906, 'timing/data': 0.00021601002663373947, 'timing/model': 0.3929297900176607, 'learning_rate/qwen_vl_interface': 9.235895282411029e-06, 'learning_rate/action_model': 9.235895282411029e-05, 'epoch': 0.07}) +[2026-10-03 19:07:21,693][starVLA.training.train_starvla][INFO] - Step 998, Loss: {'train/loss': 0.39961010217666626, 'train/grad_norm_pre_clip': 37.373260498046875, 'timing/data': 0.00020541000412777066, 'timing/model': 0.3659654130169656, 'learning_rate/qwen_vl_interface': 9.2342379918952e-06, 'learning_rate/action_model': 9.2342379918952e-05, 'epoch': 0.07}) +[2026-10-03 19:07:22,393][starVLA.training.train_starvla][INFO] - Step 999, Loss: {'train/loss': 0.24900415539741516, 'train/grad_norm_pre_clip': 22.945383071899414, 'timing/data': 0.0003835099923890084, 'timing/model': 0.38992608900298364, 'learning_rate/qwen_vl_interface': 9.232579063609631e-06, 'learning_rate/action_model': 9.232579063609631e-05, 'epoch': 0.07}) +[2026-10-03 19:07:23,294][starVLA.training.train_starvla][INFO] - Step 1000, Loss: {'train/loss': 0.3996189832687378, 'train/grad_norm_pre_clip': 26.9201602935791, 'timing/data': 0.0003659899812191725, 'timing/model': 0.5915169689978939, 'learning_rate/qwen_vl_interface': 9.230918498236243e-06, 'learning_rate/action_model': 9.230918498236244e-05, 'epoch': 0.07}) +[2026-10-03 19:07:23,994][starVLA.training.train_starvla][INFO] - Step 1001, Loss: {'train/loss': 0.23632939159870148, 'train/grad_norm_pre_clip': 81.32282257080078, 'timing/data': 0.00024120000307448208, 'timing/model': 0.3925260399992112, 'learning_rate/qwen_vl_interface': 9.229256296457635e-06, 'learning_rate/action_model': 9.229256296457634e-05, 'epoch': 0.07}) +[2026-10-03 19:07:24,695][starVLA.training.train_starvla][INFO] - Step 1002, Loss: {'train/loss': 0.14134204387664795, 'train/grad_norm_pre_clip': 14.945032119750977, 'timing/data': 0.0003681899979710579, 'timing/model': 0.39144014002522454, 'learning_rate/qwen_vl_interface': 9.22759245895707e-06, 'learning_rate/action_model': 9.227592458957071e-05, 'epoch': 0.07}) +[2026-10-03 19:07:25,394][starVLA.training.train_starvla][INFO] - Step 1003, Loss: {'train/loss': 0.033183932304382324, 'train/grad_norm_pre_clip': 6.49548864364624, 'timing/data': 0.00020701999892480671, 'timing/model': 0.3894427089835517, 'learning_rate/qwen_vl_interface': 9.225926986418494e-06, 'learning_rate/action_model': 9.225926986418494e-05, 'epoch': 0.07}) +[2026-10-03 19:07:26,094][starVLA.training.train_starvla][INFO] - Step 1004, Loss: {'train/loss': 0.23993244767189026, 'train/grad_norm_pre_clip': 55.55423355102539, 'timing/data': 0.0005000700184609741, 'timing/model': 0.39181850000750273, 'learning_rate/qwen_vl_interface': 9.224259879526519e-06, 'learning_rate/action_model': 9.224259879526519e-05, 'epoch': 0.07}) +[2026-10-03 19:07:26,996][starVLA.training.train_starvla][INFO] - Step 1005, Loss: {'train/loss': 0.18951871991157532, 'train/grad_norm_pre_clip': 57.62648391723633, 'timing/data': 0.00032235000981017947, 'timing/model': 0.5929422999906819, 'learning_rate/qwen_vl_interface': 9.222591138966427e-06, 'learning_rate/action_model': 9.222591138966426e-05, 'epoch': 0.07}) +[2026-10-03 19:07:27,696][starVLA.training.train_starvla][INFO] - Step 1006, Loss: {'train/loss': 0.36220046877861023, 'train/grad_norm_pre_clip': 47.69625473022461, 'timing/data': 0.00020506998407654464, 'timing/model': 0.3913285300077405, 'learning_rate/qwen_vl_interface': 9.220920765424175e-06, 'learning_rate/action_model': 9.220920765424174e-05, 'epoch': 0.07}) +[2026-10-03 19:07:28,395][starVLA.training.train_starvla][INFO] - Step 1007, Loss: {'train/loss': 0.016433972865343094, 'train/grad_norm_pre_clip': 4.371526718139648, 'timing/data': 0.00033183000050485134, 'timing/model': 0.3916344789904542, 'learning_rate/qwen_vl_interface': 9.219248759586392e-06, 'learning_rate/action_model': 9.219248759586391e-05, 'epoch': 0.07}) +[2026-10-03 19:07:29,097][starVLA.training.train_starvla][INFO] - Step 1008, Loss: {'train/loss': 0.022448452189564705, 'train/grad_norm_pre_clip': 12.926243782043457, 'timing/data': 0.000361769984010607, 'timing/model': 0.39124487101798877, 'learning_rate/qwen_vl_interface': 9.217575122140375e-06, 'learning_rate/action_model': 9.217575122140375e-05, 'epoch': 0.07}) +[2026-10-03 19:07:29,798][starVLA.training.train_starvla][INFO] - Step 1009, Loss: {'train/loss': 0.22273802757263184, 'train/grad_norm_pre_clip': 245.32981872558594, 'timing/data': 0.00032277998980134726, 'timing/model': 0.3930143300094642, 'learning_rate/qwen_vl_interface': 9.215899853774094e-06, 'learning_rate/action_model': 9.215899853774093e-05, 'epoch': 0.07}) +[2026-10-03 19:07:30,698][starVLA.training.train_starvla][INFO] - Step 1010, Loss: {'train/loss': 0.4176689088344574, 'train/grad_norm_pre_clip': 43.148284912109375, 'timing/data': 0.00035151999327354133, 'timing/model': 0.5927502980048303, 'learning_rate/qwen_vl_interface': 9.214222955176189e-06, 'learning_rate/action_model': 9.214222955176189e-05, 'epoch': 0.07}) +[2026-10-03 19:07:31,398][starVLA.training.train_starvla][INFO] - Step 1011, Loss: {'train/loss': 0.5050190687179565, 'train/grad_norm_pre_clip': 65.4216537475586, 'timing/data': 0.00035327000659890473, 'timing/model': 0.3911013489996549, 'learning_rate/qwen_vl_interface': 9.212544427035967e-06, 'learning_rate/action_model': 9.212544427035967e-05, 'epoch': 0.07}) +[2026-10-03 19:07:32,096][starVLA.training.train_starvla][INFO] - Step 1012, Loss: {'train/loss': 0.02211952768266201, 'train/grad_norm_pre_clip': 11.482026100158691, 'timing/data': 0.00024163001216948032, 'timing/model': 0.3909076710115187, 'learning_rate/qwen_vl_interface': 9.21086427004341e-06, 'learning_rate/action_model': 9.210864270043411e-05, 'epoch': 0.07}) +[2026-10-03 19:07:32,797][starVLA.training.train_starvla][INFO] - Step 1013, Loss: {'train/loss': 0.3453056812286377, 'train/grad_norm_pre_clip': 24.645925521850586, 'timing/data': 0.0003170889976900071, 'timing/model': 0.3920279199955985, 'learning_rate/qwen_vl_interface': 9.209182484889168e-06, 'learning_rate/action_model': 9.209182484889167e-05, 'epoch': 0.07}) +[2026-10-03 19:07:33,497][starVLA.training.train_starvla][INFO] - Step 1014, Loss: {'train/loss': 0.008184504695236683, 'train/grad_norm_pre_clip': 2.911877155303955, 'timing/data': 0.000209319987334311, 'timing/model': 0.3920860590005759, 'learning_rate/qwen_vl_interface': 9.207499072264557e-06, 'learning_rate/action_model': 9.207499072264558e-05, 'epoch': 0.07}) +[2026-10-03 19:07:34,195][starVLA.training.train_starvla][INFO] - Step 1015, Loss: {'train/loss': 0.3634050488471985, 'train/grad_norm_pre_clip': 35.01382827758789, 'timing/data': 0.00030242998036555946, 'timing/model': 0.3911638400168158, 'learning_rate/qwen_vl_interface': 9.205814032861566e-06, 'learning_rate/action_model': 9.205814032861566e-05, 'epoch': 0.07}) +[2026-10-03 19:07:35,084][starVLA.training.train_starvla][INFO] - Step 1016, Loss: {'train/loss': 0.08479559421539307, 'train/grad_norm_pre_clip': 58.149436950683594, 'timing/data': 0.00034680901444517076, 'timing/model': 0.390602820989443, 'learning_rate/qwen_vl_interface': 9.204127367372849e-06, 'learning_rate/action_model': 9.204127367372849e-05, 'epoch': 0.07}) +[2026-10-03 19:07:35,784][starVLA.training.train_starvla][INFO] - Step 1017, Loss: {'train/loss': 0.05698276683688164, 'train/grad_norm_pre_clip': 26.263198852539062, 'timing/data': 0.00025548000121489167, 'timing/model': 0.39216931999544613, 'learning_rate/qwen_vl_interface': 9.202439076491732e-06, 'learning_rate/action_model': 9.202439076491732e-05, 'epoch': 0.07}) +[2026-10-03 19:07:36,480][starVLA.training.train_starvla][INFO] - Step 1018, Loss: {'train/loss': 0.3182201087474823, 'train/grad_norm_pre_clip': 67.0940170288086, 'timing/data': 0.0003561300109140575, 'timing/model': 0.38899271900299937, 'learning_rate/qwen_vl_interface': 9.200749160912208e-06, 'learning_rate/action_model': 9.200749160912208e-05, 'epoch': 0.07}) +[2026-10-03 19:07:37,178][starVLA.training.train_starvla][INFO] - Step 1019, Loss: {'train/loss': 0.03749299421906471, 'train/grad_norm_pre_clip': 176.5663604736328, 'timing/data': 0.00043485997593961656, 'timing/model': 0.3894032899988815, 'learning_rate/qwen_vl_interface': 9.199057621328937e-06, 'learning_rate/action_model': 9.199057621328935e-05, 'epoch': 0.07}) +[2026-10-03 19:07:37,875][starVLA.training.train_starvla][INFO] - Step 1020, Loss: {'train/loss': 0.31608715653419495, 'train/grad_norm_pre_clip': 111.56534576416016, 'timing/data': 0.0002046799927484244, 'timing/model': 0.3891864199831616, 'learning_rate/qwen_vl_interface': 9.197364458437244e-06, 'learning_rate/action_model': 9.197364458437245e-05, 'epoch': 0.07}) +[2026-10-03 19:07:38,771][starVLA.training.train_starvla][INFO] - Step 1021, Loss: {'train/loss': 0.36400794982910156, 'train/grad_norm_pre_clip': 19.689350128173828, 'timing/data': 0.0003466500202193856, 'timing/model': 0.3928091289999429, 'learning_rate/qwen_vl_interface': 9.19566967293313e-06, 'learning_rate/action_model': 9.19566967293313e-05, 'epoch': 0.07}) +[2026-10-03 19:07:39,471][starVLA.training.train_starvla][INFO] - Step 1022, Loss: {'train/loss': 0.0018718986539170146, 'train/grad_norm_pre_clip': 0.5483462810516357, 'timing/data': 0.0003615400055423379, 'timing/model': 0.3920245699991938, 'learning_rate/qwen_vl_interface': 9.193973265513253e-06, 'learning_rate/action_model': 9.193973265513253e-05, 'epoch': 0.07}) +[2026-10-03 19:07:40,174][starVLA.training.train_starvla][INFO] - Step 1023, Loss: {'train/loss': 0.2794005274772644, 'train/grad_norm_pre_clip': 42.9171028137207, 'timing/data': 0.00024252000730484724, 'timing/model': 0.39309899002546445, 'learning_rate/qwen_vl_interface': 9.192275236874946e-06, 'learning_rate/action_model': 9.192275236874946e-05, 'epoch': 0.07}) +[2026-10-03 19:07:40,874][starVLA.training.train_starvla][INFO] - Step 1024, Loss: {'train/loss': 0.3299195170402527, 'train/grad_norm_pre_clip': 107.6661148071289, 'timing/data': 0.00027642998611554503, 'timing/model': 0.3913197299989406, 'learning_rate/qwen_vl_interface': 9.190575587716199e-06, 'learning_rate/action_model': 9.190575587716199e-05, 'epoch': 0.07}) +[2026-10-03 19:07:41,575][starVLA.training.train_starvla][INFO] - Step 1025, Loss: {'train/loss': 0.2449733167886734, 'train/grad_norm_pre_clip': 15.9373140335083, 'timing/data': 0.0002355000178795308, 'timing/model': 0.39383565899333917, 'learning_rate/qwen_vl_interface': 9.188874318735677e-06, 'learning_rate/action_model': 9.188874318735676e-05, 'epoch': 0.08}) +[2026-10-03 19:07:42,481][starVLA.training.train_starvla][INFO] - Step 1026, Loss: {'train/loss': 0.014887306839227676, 'train/grad_norm_pre_clip': 6.448786735534668, 'timing/data': 0.0002391700109001249, 'timing/model': 0.5979500990069937, 'learning_rate/qwen_vl_interface': 9.187171430632707e-06, 'learning_rate/action_model': 9.187171430632707e-05, 'epoch': 0.08}) +[2026-10-03 19:07:43,180][starVLA.training.train_starvla][INFO] - Step 1027, Loss: {'train/loss': 0.015378614887595177, 'train/grad_norm_pre_clip': 7.83750581741333, 'timing/data': 0.0003216299810446799, 'timing/model': 0.3926772299746517, 'learning_rate/qwen_vl_interface': 9.185466924107283e-06, 'learning_rate/action_model': 9.185466924107282e-05, 'epoch': 0.08}) +[2026-10-03 19:07:43,879][starVLA.training.train_starvla][INFO] - Step 1028, Loss: {'train/loss': 0.01198143046349287, 'train/grad_norm_pre_clip': 12.929701805114746, 'timing/data': 0.0002175699919462204, 'timing/model': 0.3913441399927251, 'learning_rate/qwen_vl_interface': 9.18376079986006e-06, 'learning_rate/action_model': 9.183760799860061e-05, 'epoch': 0.08}) +[2026-10-03 19:07:44,576][starVLA.training.train_starvla][INFO] - Step 1029, Loss: {'train/loss': 0.7175105810165405, 'train/grad_norm_pre_clip': 115.61666870117188, 'timing/data': 0.0002057200181297958, 'timing/model': 0.391277759015793, 'learning_rate/qwen_vl_interface': 9.182053058592365e-06, 'learning_rate/action_model': 9.182053058592365e-05, 'epoch': 0.08}) +[2026-10-03 19:07:45,275][starVLA.training.train_starvla][INFO] - Step 1030, Loss: {'train/loss': 0.0225311741232872, 'train/grad_norm_pre_clip': 9.733704566955566, 'timing/data': 0.0002039399987552315, 'timing/model': 0.39084864099277183, 'learning_rate/qwen_vl_interface': 9.180343701006183e-06, 'learning_rate/action_model': 9.180343701006183e-05, 'epoch': 0.08}) +[2026-10-03 19:07:45,974][starVLA.training.train_starvla][INFO] - Step 1031, Loss: {'train/loss': 0.012066696770489216, 'train/grad_norm_pre_clip': 9.64238452911377, 'timing/data': 0.0003190200077369809, 'timing/model': 0.3912285290134605, 'learning_rate/qwen_vl_interface': 9.17863272780417e-06, 'learning_rate/action_model': 9.178632727804168e-05, 'epoch': 0.08}) +[2026-10-03 19:07:46,874][starVLA.training.train_starvla][INFO] - Step 1032, Loss: {'train/loss': 0.2391887605190277, 'train/grad_norm_pre_clip': 104.2641372680664, 'timing/data': 0.00036504000308923423, 'timing/model': 0.3890268999966793, 'learning_rate/qwen_vl_interface': 9.176920139689638e-06, 'learning_rate/action_model': 9.176920139689637e-05, 'epoch': 0.08}) +[2026-10-03 19:07:47,570][starVLA.training.train_starvla][INFO] - Step 1033, Loss: {'train/loss': 0.03186480328440666, 'train/grad_norm_pre_clip': 38.98765563964844, 'timing/data': 0.0003210300055798143, 'timing/model': 0.39136095900903456, 'learning_rate/qwen_vl_interface': 9.175205937366568e-06, 'learning_rate/action_model': 9.175205937366569e-05, 'epoch': 0.08}) +[2026-10-03 19:07:48,268][starVLA.training.train_starvla][INFO] - Step 1034, Loss: {'train/loss': 0.3818192481994629, 'train/grad_norm_pre_clip': 54.0413932800293, 'timing/data': 0.0002047099987976253, 'timing/model': 0.38929696998093277, 'learning_rate/qwen_vl_interface': 9.173490121539608e-06, 'learning_rate/action_model': 9.173490121539608e-05, 'epoch': 0.08}) +[2026-10-03 19:07:48,967][starVLA.training.train_starvla][INFO] - Step 1035, Loss: {'train/loss': 0.5299456715583801, 'train/grad_norm_pre_clip': 161.472412109375, 'timing/data': 0.0002449599851388484, 'timing/model': 0.3910569889994804, 'learning_rate/qwen_vl_interface': 9.171772692914062e-06, 'learning_rate/action_model': 9.171772692914061e-05, 'epoch': 0.08}) +[2026-10-03 19:07:49,666][starVLA.training.train_starvla][INFO] - Step 1036, Loss: {'train/loss': 0.011754537001252174, 'train/grad_norm_pre_clip': 13.522838592529297, 'timing/data': 0.00021079997532069683, 'timing/model': 0.38991206997889094, 'learning_rate/qwen_vl_interface': 9.1700536521959e-06, 'learning_rate/action_model': 9.170053652195899e-05, 'epoch': 0.08}) +[2026-10-03 19:07:50,571][starVLA.training.train_starvla][INFO] - Step 1037, Loss: {'train/loss': 0.5240305662155151, 'train/grad_norm_pre_clip': 36.30958938598633, 'timing/data': 0.00027538000722415745, 'timing/model': 0.392601878993446, 'learning_rate/qwen_vl_interface': 9.168333000091754e-06, 'learning_rate/action_model': 9.168333000091754e-05, 'epoch': 0.08}) +[2026-10-03 19:07:51,267][starVLA.training.train_starvla][INFO] - Step 1038, Loss: {'train/loss': 0.21674394607543945, 'train/grad_norm_pre_clip': 39.30439376831055, 'timing/data': 0.0002406999992672354, 'timing/model': 0.38892359999590553, 'learning_rate/qwen_vl_interface': 9.16661073730892e-06, 'learning_rate/action_model': 9.166610737308919e-05, 'epoch': 0.08}) +[2026-10-03 19:07:51,975][starVLA.training.train_starvla][INFO] - Step 1039, Loss: {'train/loss': 0.3131166100502014, 'train/grad_norm_pre_clip': 44.827430725097656, 'timing/data': 0.00037672900361940265, 'timing/model': 0.3976924809976481, 'learning_rate/qwen_vl_interface': 9.164886864555355e-06, 'learning_rate/action_model': 9.164886864555355e-05, 'epoch': 0.08}) +[2026-10-03 19:07:52,687][starVLA.training.train_starvla][INFO] - Step 1040, Loss: {'train/loss': 0.17352215945720673, 'train/grad_norm_pre_clip': 51.90510177612305, 'timing/data': 0.0002345300163142383, 'timing/model': 0.39792684800340794, 'learning_rate/qwen_vl_interface': 9.163161382539678e-06, 'learning_rate/action_model': 9.163161382539678e-05, 'epoch': 0.08}) +[2026-10-03 19:07:53,396][starVLA.training.train_starvla][INFO] - Step 1041, Loss: {'train/loss': 0.002403753809630871, 'train/grad_norm_pre_clip': 0.7703501582145691, 'timing/data': 0.00031276000663638115, 'timing/model': 0.3968144789978396, 'learning_rate/qwen_vl_interface': 9.16143429197117e-06, 'learning_rate/action_model': 9.16143429197117e-05, 'epoch': 0.08}) +[2026-10-03 19:07:54,308][starVLA.training.train_starvla][INFO] - Step 1042, Loss: {'train/loss': 0.16849887371063232, 'train/grad_norm_pre_clip': 62.51833724975586, 'timing/data': 0.00035362900234758854, 'timing/model': 0.6011037979915272, 'learning_rate/qwen_vl_interface': 9.159705593559773e-06, 'learning_rate/action_model': 9.159705593559773e-05, 'epoch': 0.08}) +[2026-10-03 19:07:55,019][starVLA.training.train_starvla][INFO] - Step 1043, Loss: {'train/loss': 0.9710197448730469, 'train/grad_norm_pre_clip': 54.82889175415039, 'timing/data': 0.00024528001085855067, 'timing/model': 0.3981964789854828, 'learning_rate/qwen_vl_interface': 9.157975288016089e-06, 'learning_rate/action_model': 9.157975288016088e-05, 'epoch': 0.08}) +[2026-10-03 19:07:55,727][starVLA.training.train_starvla][INFO] - Step 1044, Loss: {'train/loss': 0.21448422968387604, 'train/grad_norm_pre_clip': 49.78771209716797, 'timing/data': 0.00023768001119606197, 'timing/model': 0.3959542390075512, 'learning_rate/qwen_vl_interface': 9.156243376051381e-06, 'learning_rate/action_model': 9.156243376051381e-05, 'epoch': 0.08}) +[2026-10-03 19:07:56,435][starVLA.training.train_starvla][INFO] - Step 1045, Loss: {'train/loss': 0.285001665353775, 'train/grad_norm_pre_clip': 22.006488800048828, 'timing/data': 0.00035290999221615493, 'timing/model': 0.39674259000457823, 'learning_rate/qwen_vl_interface': 9.154509858377574e-06, 'learning_rate/action_model': 9.154509858377574e-05, 'epoch': 0.08}) +[2026-10-03 19:07:57,143][starVLA.training.train_starvla][INFO] - Step 1046, Loss: {'train/loss': 0.3290759027004242, 'train/grad_norm_pre_clip': 63.7216796875, 'timing/data': 0.0003733700141310692, 'timing/model': 0.39466041899868287, 'learning_rate/qwen_vl_interface': 9.152774735707248e-06, 'learning_rate/action_model': 9.152774735707247e-05, 'epoch': 0.08}) +[2026-10-03 19:07:58,062][starVLA.training.train_starvla][INFO] - Step 1047, Loss: {'train/loss': 0.588743269443512, 'train/grad_norm_pre_clip': 200.04342651367188, 'timing/data': 0.0003336500085424632, 'timing/model': 0.604625758016482, 'learning_rate/qwen_vl_interface': 9.15103800875365e-06, 'learning_rate/action_model': 9.15103800875365e-05, 'epoch': 0.08}) +[2026-10-03 19:07:58,744][starVLA.training.train_starvla][INFO] - Step 1048, Loss: {'train/loss': 0.0663057416677475, 'train/grad_norm_pre_clip': 31.540773391723633, 'timing/data': 0.00044023001100867987, 'timing/model': 0.36966120201395825, 'learning_rate/qwen_vl_interface': 9.149299678230681e-06, 'learning_rate/action_model': 9.149299678230681e-05, 'epoch': 0.08}) +[2026-10-03 19:07:59,453][starVLA.training.train_starvla][INFO] - Step 1049, Loss: {'train/loss': 0.19344007968902588, 'train/grad_norm_pre_clip': 107.55059814453125, 'timing/data': 0.0002492800122126937, 'timing/model': 0.39525842902367003, 'learning_rate/qwen_vl_interface': 9.147559744852905e-06, 'learning_rate/action_model': 9.147559744852905e-05, 'epoch': 0.08}) +[2026-10-03 19:08:00,160][starVLA.training.train_starvla][INFO] - Step 1050, Loss: {'train/loss': 0.03939398378133774, 'train/grad_norm_pre_clip': 12.805520057678223, 'timing/data': 0.0002894100034609437, 'timing/model': 0.39668904899735935, 'learning_rate/qwen_vl_interface': 9.14581820933554e-06, 'learning_rate/action_model': 9.14581820933554e-05, 'epoch': 0.08}) +[2026-10-03 19:08:00,870][starVLA.training.train_starvla][INFO] - Step 1051, Loss: {'train/loss': 0.026678845286369324, 'train/grad_norm_pre_clip': 5.634906768798828, 'timing/data': 0.0002798899949993938, 'timing/model': 0.3959236500086263, 'learning_rate/qwen_vl_interface': 9.14407507239447e-06, 'learning_rate/action_model': 9.14407507239447e-05, 'epoch': 0.08}) +[2026-10-03 19:08:01,579][starVLA.training.train_starvla][INFO] - Step 1052, Loss: {'train/loss': 0.10943350940942764, 'train/grad_norm_pre_clip': 38.998592376708984, 'timing/data': 0.00024518001009710133, 'timing/model': 0.39545248899958096, 'learning_rate/qwen_vl_interface': 9.142330334746227e-06, 'learning_rate/action_model': 9.142330334746227e-05, 'epoch': 0.08}) +[2026-10-03 19:08:02,487][starVLA.training.train_starvla][INFO] - Step 1053, Loss: {'train/loss': 0.6931825280189514, 'train/grad_norm_pre_clip': 88.2925796508789, 'timing/data': 0.0002355000178795308, 'timing/model': 0.39632118100416847, 'learning_rate/qwen_vl_interface': 9.14058399710801e-06, 'learning_rate/action_model': 9.14058399710801e-05, 'epoch': 0.08}) +[2026-10-03 19:08:03,196][starVLA.training.train_starvla][INFO] - Step 1054, Loss: {'train/loss': 0.03863315284252167, 'train/grad_norm_pre_clip': 7.417735576629639, 'timing/data': 0.0003561000048648566, 'timing/model': 0.39596914901630953, 'learning_rate/qwen_vl_interface': 9.138836060197673e-06, 'learning_rate/action_model': 9.138836060197672e-05, 'epoch': 0.08}) +[2026-10-03 19:08:03,906][starVLA.training.train_starvla][INFO] - Step 1055, Loss: {'train/loss': 0.31483954191207886, 'train/grad_norm_pre_clip': 111.71857452392578, 'timing/data': 0.0002695400035008788, 'timing/model': 0.3956179889792111, 'learning_rate/qwen_vl_interface': 9.137086524733725e-06, 'learning_rate/action_model': 9.137086524733724e-05, 'epoch': 0.08}) +[2026-10-03 19:08:04,613][starVLA.training.train_starvla][INFO] - Step 1056, Loss: {'train/loss': 0.005308790132403374, 'train/grad_norm_pre_clip': 0.9602925181388855, 'timing/data': 0.00043213999015279114, 'timing/model': 0.39525745902210474, 'learning_rate/qwen_vl_interface': 9.135335391435336e-06, 'learning_rate/action_model': 9.135335391435335e-05, 'epoch': 0.08}) +[2026-10-03 19:08:05,318][starVLA.training.train_starvla][INFO] - Step 1057, Loss: {'train/loss': 0.4419252872467041, 'train/grad_norm_pre_clip': 107.9109878540039, 'timing/data': 0.00025983998784795403, 'timing/model': 0.39324413100257516, 'learning_rate/qwen_vl_interface': 9.13358266102233e-06, 'learning_rate/action_model': 9.133582661022328e-05, 'epoch': 0.08}) +[2026-10-03 19:08:06,227][starVLA.training.train_starvla][INFO] - Step 1058, Loss: {'train/loss': 0.13832923769950867, 'train/grad_norm_pre_clip': 276.71075439453125, 'timing/data': 0.00030457001412287354, 'timing/model': 0.3926641500147525, 'learning_rate/qwen_vl_interface': 9.131828334215185e-06, 'learning_rate/action_model': 9.131828334215185e-05, 'epoch': 0.08}) +[2026-10-03 19:08:06,933][starVLA.training.train_starvla][INFO] - Step 1059, Loss: {'train/loss': 0.061861153692007065, 'train/grad_norm_pre_clip': 130.13824462890625, 'timing/data': 0.00027278001653030515, 'timing/model': 0.3957931190088857, 'learning_rate/qwen_vl_interface': 9.130072411735043e-06, 'learning_rate/action_model': 9.130072411735043e-05, 'epoch': 0.08}) +[2026-10-03 19:08:07,635][starVLA.training.train_starvla][INFO] - Step 1060, Loss: {'train/loss': 0.12342166155576706, 'train/grad_norm_pre_clip': 98.14361572265625, 'timing/data': 0.00021898001432418823, 'timing/model': 0.39230848898296244, 'learning_rate/qwen_vl_interface': 9.128314894303693e-06, 'learning_rate/action_model': 9.128314894303693e-05, 'epoch': 0.08}) +[2026-10-03 19:08:08,338][starVLA.training.train_starvla][INFO] - Step 1061, Loss: {'train/loss': 0.4402398467063904, 'train/grad_norm_pre_clip': 24.5454044342041, 'timing/data': 0.00032309998641721904, 'timing/model': 0.39303547897725366, 'learning_rate/qwen_vl_interface': 9.126555782643592e-06, 'learning_rate/action_model': 9.126555782643591e-05, 'epoch': 0.08}) +[2026-10-03 19:08:09,041][starVLA.training.train_starvla][INFO] - Step 1062, Loss: {'train/loss': 0.03133891150355339, 'train/grad_norm_pre_clip': 48.554752349853516, 'timing/data': 0.00022409000666812062, 'timing/model': 0.39358026001718827, 'learning_rate/qwen_vl_interface': 9.124795077477835e-06, 'learning_rate/action_model': 9.124795077477834e-05, 'epoch': 0.08}) +[2026-10-03 19:08:09,948][starVLA.training.train_starvla][INFO] - Step 1063, Loss: {'train/loss': 0.28218144178390503, 'train/grad_norm_pre_clip': 69.18569946289062, 'timing/data': 0.0003899600123986602, 'timing/model': 0.597229178994894, 'learning_rate/qwen_vl_interface': 9.123032779530186e-06, 'learning_rate/action_model': 9.123032779530187e-05, 'epoch': 0.08}) +[2026-10-03 19:08:10,656][starVLA.training.train_starvla][INFO] - Step 1064, Loss: {'train/loss': 0.26597678661346436, 'train/grad_norm_pre_clip': 63.853458404541016, 'timing/data': 0.0002346700057387352, 'timing/model': 0.3945927789900452, 'learning_rate/qwen_vl_interface': 9.12126888952506e-06, 'learning_rate/action_model': 9.12126888952506e-05, 'epoch': 0.08}) +[2026-10-03 19:08:11,365][starVLA.training.train_starvla][INFO] - Step 1065, Loss: {'train/loss': 0.017617180943489075, 'train/grad_norm_pre_clip': 8.694062232971191, 'timing/data': 0.0003390299971215427, 'timing/model': 0.3959029790130444, 'learning_rate/qwen_vl_interface': 9.119503408187521e-06, 'learning_rate/action_model': 9.119503408187522e-05, 'epoch': 0.08}) +[2026-10-03 19:08:12,073][starVLA.training.train_starvla][INFO] - Step 1066, Loss: {'train/loss': 0.007238238118588924, 'train/grad_norm_pre_clip': 4.022193431854248, 'timing/data': 0.00022386998170986772, 'timing/model': 0.39425419000326656, 'learning_rate/qwen_vl_interface': 9.117736336243299e-06, 'learning_rate/action_model': 9.117736336243298e-05, 'epoch': 0.08}) +[2026-10-03 19:08:12,782][starVLA.training.train_starvla][INFO] - Step 1067, Loss: {'train/loss': 0.41067731380462646, 'train/grad_norm_pre_clip': 21.949087142944336, 'timing/data': 0.00046308001037687063, 'timing/model': 0.39710002901847474, 'learning_rate/qwen_vl_interface': 9.115967674418765e-06, 'learning_rate/action_model': 9.115967674418765e-05, 'epoch': 0.08}) +[2026-10-03 19:08:13,694][starVLA.training.train_starvla][INFO] - Step 1068, Loss: {'train/loss': 0.2788412570953369, 'train/grad_norm_pre_clip': 42.78832244873047, 'timing/data': 0.0004084899846930057, 'timing/model': 0.5981004489876796, 'learning_rate/qwen_vl_interface': 9.114197423440949e-06, 'learning_rate/action_model': 9.11419742344095e-05, 'epoch': 0.08}) +[2026-10-03 19:08:14,403][starVLA.training.train_starvla][INFO] - Step 1069, Loss: {'train/loss': 0.2579483687877655, 'train/grad_norm_pre_clip': 72.38432312011719, 'timing/data': 0.00023365000379271805, 'timing/model': 0.3963320390030276, 'learning_rate/qwen_vl_interface': 9.112425584037536e-06, 'learning_rate/action_model': 9.112425584037536e-05, 'epoch': 0.08}) +[2026-10-03 19:08:15,109][starVLA.training.train_starvla][INFO] - Step 1070, Loss: {'train/loss': 0.08845821768045425, 'train/grad_norm_pre_clip': 54.83282470703125, 'timing/data': 0.00021975001436658204, 'timing/model': 0.39405799901578575, 'learning_rate/qwen_vl_interface': 9.110652156936863e-06, 'learning_rate/action_model': 9.110652156936863e-05, 'epoch': 0.08}) +[2026-10-03 19:08:15,818][starVLA.training.train_starvla][INFO] - Step 1071, Loss: {'train/loss': 0.42480209469795227, 'train/grad_norm_pre_clip': 69.49546813964844, 'timing/data': 0.00034155999310314655, 'timing/model': 0.39530754898441955, 'learning_rate/qwen_vl_interface': 9.108877142867918e-06, 'learning_rate/action_model': 9.108877142867918e-05, 'epoch': 0.08}) +[2026-10-03 19:08:16,525][starVLA.training.train_starvla][INFO] - Step 1072, Loss: {'train/loss': 0.46493709087371826, 'train/grad_norm_pre_clip': 209.26109313964844, 'timing/data': 0.0003802000137511641, 'timing/model': 0.3951228189980611, 'learning_rate/qwen_vl_interface': 9.107100542560342e-06, 'learning_rate/action_model': 9.107100542560342e-05, 'epoch': 0.08}) +[2026-10-03 19:08:17,233][starVLA.training.train_starvla][INFO] - Step 1073, Loss: {'train/loss': 0.5828611254692078, 'train/grad_norm_pre_clip': 84.12873840332031, 'timing/data': 0.00032885000109672546, 'timing/model': 0.39465532900067046, 'learning_rate/qwen_vl_interface': 9.105322356744429e-06, 'learning_rate/action_model': 9.105322356744429e-05, 'epoch': 0.08}) +[2026-10-03 19:08:18,146][starVLA.training.train_starvla][INFO] - Step 1074, Loss: {'train/loss': 0.4846440255641937, 'train/grad_norm_pre_clip': 27.39493179321289, 'timing/data': 0.0004434300062712282, 'timing/model': 0.3953452389978338, 'learning_rate/qwen_vl_interface': 9.103542586151121e-06, 'learning_rate/action_model': 9.10354258615112e-05, 'epoch': 0.08}) +[2026-10-03 19:08:18,856][starVLA.training.train_starvla][INFO] - Step 1075, Loss: {'train/loss': 0.20965316891670227, 'train/grad_norm_pre_clip': 89.52676391601562, 'timing/data': 0.00034410000080242753, 'timing/model': 0.396281489985995, 'learning_rate/qwen_vl_interface': 9.101761231512017e-06, 'learning_rate/action_model': 9.101761231512017e-05, 'epoch': 0.08}) +[2026-10-03 19:08:19,566][starVLA.training.train_starvla][INFO] - Step 1076, Loss: {'train/loss': 0.0330660454928875, 'train/grad_norm_pre_clip': 16.479257583618164, 'timing/data': 0.00043476998689584434, 'timing/model': 0.3952918600116391, 'learning_rate/qwen_vl_interface': 9.099978293559365e-06, 'learning_rate/action_model': 9.099978293559366e-05, 'epoch': 0.08}) +[2026-10-03 19:08:20,274][starVLA.training.train_starvla][INFO] - Step 1077, Loss: {'train/loss': 0.0052714599296450615, 'train/grad_norm_pre_clip': 2.1081628799438477, 'timing/data': 0.00031706999288871884, 'timing/model': 0.3957038290100172, 'learning_rate/qwen_vl_interface': 9.098193773026065e-06, 'learning_rate/action_model': 9.098193773026064e-05, 'epoch': 0.08}) +[2026-10-03 19:08:20,982][starVLA.training.train_starvla][INFO] - Step 1078, Loss: {'train/loss': 0.096236452460289, 'train/grad_norm_pre_clip': 19.772188186645508, 'timing/data': 0.0002242499904241413, 'timing/model': 0.39599109897972085, 'learning_rate/qwen_vl_interface': 9.096407670645663e-06, 'learning_rate/action_model': 9.096407670645662e-05, 'epoch': 0.08}) +[2026-10-03 19:08:21,893][starVLA.training.train_starvla][INFO] - Step 1079, Loss: {'train/loss': 0.07135316729545593, 'train/grad_norm_pre_clip': 59.79193115234375, 'timing/data': 0.00037096001324243844, 'timing/model': 0.5987995179893915, 'learning_rate/qwen_vl_interface': 9.094619987152357e-06, 'learning_rate/action_model': 9.094619987152358e-05, 'epoch': 0.08}) +[2026-10-03 19:08:22,600][starVLA.training.train_starvla][INFO] - Step 1080, Loss: {'train/loss': 0.2982107400894165, 'train/grad_norm_pre_clip': 67.07682037353516, 'timing/data': 0.00022733001969754696, 'timing/model': 0.39558692899299785, 'learning_rate/qwen_vl_interface': 9.092830723281e-06, 'learning_rate/action_model': 9.092830723281e-05, 'epoch': 0.08}) +[2026-10-03 19:08:23,308][starVLA.training.train_starvla][INFO] - Step 1081, Loss: {'train/loss': 0.1807764768600464, 'train/grad_norm_pre_clip': 53.562843322753906, 'timing/data': 0.00023925999994389713, 'timing/model': 0.39586246898397803, 'learning_rate/qwen_vl_interface': 9.09103987976709e-06, 'learning_rate/action_model': 9.09103987976709e-05, 'epoch': 0.08}) +[2026-10-03 19:08:24,017][starVLA.training.train_starvla][INFO] - Step 1082, Loss: {'train/loss': 0.4504023790359497, 'train/grad_norm_pre_clip': 24.26823616027832, 'timing/data': 0.0002236299915239215, 'timing/model': 0.3963857589988038, 'learning_rate/qwen_vl_interface': 9.089247457346774e-06, 'learning_rate/action_model': 9.089247457346773e-05, 'epoch': 0.08}) +[2026-10-03 19:08:24,727][starVLA.training.train_starvla][INFO] - Step 1083, Loss: {'train/loss': 0.030789626762270927, 'train/grad_norm_pre_clip': 5.656724452972412, 'timing/data': 0.00029147000168450177, 'timing/model': 0.396331440017093, 'learning_rate/qwen_vl_interface': 9.087453456756849e-06, 'learning_rate/action_model': 9.087453456756849e-05, 'epoch': 0.08}) +[2026-10-03 19:08:25,639][starVLA.training.train_starvla][INFO] - Step 1084, Loss: {'train/loss': 0.4314571022987366, 'train/grad_norm_pre_clip': 64.62291717529297, 'timing/data': 0.00022767001064494252, 'timing/model': 0.6004385389969684, 'learning_rate/qwen_vl_interface': 9.085657878734759e-06, 'learning_rate/action_model': 9.085657878734758e-05, 'epoch': 0.08}) +[2026-10-03 19:08:26,346][starVLA.training.train_starvla][INFO] - Step 1085, Loss: {'train/loss': 0.042337559163570404, 'train/grad_norm_pre_clip': 6.045306205749512, 'timing/data': 0.00027990000671707094, 'timing/model': 0.3951902390108444, 'learning_rate/qwen_vl_interface': 9.083860724018603e-06, 'learning_rate/action_model': 9.083860724018602e-05, 'epoch': 0.08}) +[2026-10-03 19:08:27,051][starVLA.training.train_starvla][INFO] - Step 1086, Loss: {'train/loss': 0.20857086777687073, 'train/grad_norm_pre_clip': 81.31293487548828, 'timing/data': 0.00035701997694559395, 'timing/model': 0.3942882099945564, 'learning_rate/qwen_vl_interface': 9.08206199334712e-06, 'learning_rate/action_model': 9.082061993347119e-05, 'epoch': 0.08}) +[2026-10-03 19:08:27,759][starVLA.training.train_starvla][INFO] - Step 1087, Loss: {'train/loss': 0.5607834458351135, 'train/grad_norm_pre_clip': 23.270090103149414, 'timing/data': 0.00028737998218275607, 'timing/model': 0.3955505290068686, 'learning_rate/qwen_vl_interface': 9.0802616874597e-06, 'learning_rate/action_model': 9.0802616874597e-05, 'epoch': 0.08}) +[2026-10-03 19:08:28,467][starVLA.training.train_starvla][INFO] - Step 1088, Loss: {'train/loss': 0.5016367435455322, 'train/grad_norm_pre_clip': 42.3316535949707, 'timing/data': 0.0003618900082074106, 'timing/model': 0.3960563889995683, 'learning_rate/qwen_vl_interface': 9.078459807096385e-06, 'learning_rate/action_model': 9.078459807096385e-05, 'epoch': 0.08}) +[2026-10-03 19:08:29,173][starVLA.training.train_starvla][INFO] - Step 1089, Loss: {'train/loss': 0.26903361082077026, 'train/grad_norm_pre_clip': 17.487857818603516, 'timing/data': 0.0002999900025315583, 'timing/model': 0.3946916190034244, 'learning_rate/qwen_vl_interface': 9.076656352997854e-06, 'learning_rate/action_model': 9.076656352997853e-05, 'epoch': 0.08}) +[2026-10-03 19:08:30,086][starVLA.training.train_starvla][INFO] - Step 1090, Loss: {'train/loss': 0.10983078181743622, 'train/grad_norm_pre_clip': 20.508888244628906, 'timing/data': 0.00039356001070700586, 'timing/model': 0.395953308994649, 'learning_rate/qwen_vl_interface': 9.074851325905443e-06, 'learning_rate/action_model': 9.074851325905442e-05, 'epoch': 0.08}) +[2026-10-03 19:08:30,798][starVLA.training.train_starvla][INFO] - Step 1091, Loss: {'train/loss': 0.2789972424507141, 'train/grad_norm_pre_clip': 39.245086669921875, 'timing/data': 0.00027563999174162745, 'timing/model': 0.39766190899536014, 'learning_rate/qwen_vl_interface': 9.073044726561128e-06, 'learning_rate/action_model': 9.073044726561129e-05, 'epoch': 0.08}) +[2026-10-03 19:08:31,505][starVLA.training.train_starvla][INFO] - Step 1092, Loss: {'train/loss': 0.008982186205685139, 'train/grad_norm_pre_clip': 2.95807147026062, 'timing/data': 0.00035735999699682, 'timing/model': 0.3962205690040719, 'learning_rate/qwen_vl_interface': 9.071236555707536e-06, 'learning_rate/action_model': 9.071236555707537e-05, 'epoch': 0.08}) +[2026-10-03 19:08:32,215][starVLA.training.train_starvla][INFO] - Step 1093, Loss: {'train/loss': 0.5810883641242981, 'train/grad_norm_pre_clip': 63.990779876708984, 'timing/data': 0.0002574500103946775, 'timing/model': 0.3965756979887374, 'learning_rate/qwen_vl_interface': 9.069426814087937e-06, 'learning_rate/action_model': 9.069426814087937e-05, 'epoch': 0.08}) +[2026-10-03 19:08:32,924][starVLA.training.train_starvla][INFO] - Step 1094, Loss: {'train/loss': 0.010794623754918575, 'train/grad_norm_pre_clip': 3.4796245098114014, 'timing/data': 0.00036994999391026795, 'timing/model': 0.3959370099764783, 'learning_rate/qwen_vl_interface': 9.067615502446247e-06, 'learning_rate/action_model': 9.067615502446246e-05, 'epoch': 0.08}) +[2026-10-03 19:08:33,840][starVLA.training.train_starvla][INFO] - Step 1095, Loss: {'train/loss': 0.4748668968677521, 'train/grad_norm_pre_clip': 65.57482147216797, 'timing/data': 0.00023053999757394195, 'timing/model': 0.3969103190174792, 'learning_rate/qwen_vl_interface': 9.06580262152703e-06, 'learning_rate/action_model': 9.065802621527028e-05, 'epoch': 0.08}) +[2026-10-03 19:08:34,548][starVLA.training.train_starvla][INFO] - Step 1096, Loss: {'train/loss': 0.02968153916299343, 'train/grad_norm_pre_clip': 14.318585395812988, 'timing/data': 0.00022926999372430146, 'timing/model': 0.3970625090005342, 'learning_rate/qwen_vl_interface': 9.063988172075486e-06, 'learning_rate/action_model': 9.063988172075486e-05, 'epoch': 0.08}) +[2026-10-03 19:08:35,259][starVLA.training.train_starvla][INFO] - Step 1097, Loss: {'train/loss': 0.25563281774520874, 'train/grad_norm_pre_clip': 16.89678192138672, 'timing/data': 0.00031356001272797585, 'timing/model': 0.39710567900328897, 'learning_rate/qwen_vl_interface': 9.062172154837475e-06, 'learning_rate/action_model': 9.062172154837475e-05, 'epoch': 0.08}) +[2026-10-03 19:08:35,967][starVLA.training.train_starvla][INFO] - Step 1098, Loss: {'train/loss': 0.016699498519301414, 'train/grad_norm_pre_clip': 6.401893138885498, 'timing/data': 0.0003619499912019819, 'timing/model': 0.3955931300006341, 'learning_rate/qwen_vl_interface': 9.060354570559487e-06, 'learning_rate/action_model': 9.060354570559487e-05, 'epoch': 0.08}) +[2026-10-03 19:08:36,678][starVLA.training.train_starvla][INFO] - Step 1099, Loss: {'train/loss': 0.14520138502120972, 'train/grad_norm_pre_clip': 60.49773025512695, 'timing/data': 0.00026913901092484593, 'timing/model': 0.39697690901812166, 'learning_rate/qwen_vl_interface': 9.058535419988667e-06, 'learning_rate/action_model': 9.058535419988667e-05, 'epoch': 0.08}) +[2026-10-03 19:08:37,595][starVLA.training.train_starvla][INFO] - Step 1100, Loss: {'train/loss': 0.009528877213597298, 'train/grad_norm_pre_clip': 2.0365846157073975, 'timing/data': 0.0003517500008456409, 'timing/model': 0.6038271379948128, 'learning_rate/qwen_vl_interface': 9.056714703872794e-06, 'learning_rate/action_model': 9.056714703872794e-05, 'epoch': 0.08}) +[2026-10-03 19:08:38,303][starVLA.training.train_starvla][INFO] - Step 1101, Loss: {'train/loss': 0.24911606311798096, 'train/grad_norm_pre_clip': 21.3762149810791, 'timing/data': 0.00031452998518943787, 'timing/model': 0.39747884901589714, 'learning_rate/qwen_vl_interface': 9.0548924229603e-06, 'learning_rate/action_model': 9.054892422960301e-05, 'epoch': 0.08}) +[2026-10-03 19:08:39,012][starVLA.training.train_starvla][INFO] - Step 1102, Loss: {'train/loss': 0.2296113520860672, 'train/grad_norm_pre_clip': 55.202552795410156, 'timing/data': 0.00031368000782094896, 'timing/model': 0.39724223999655806, 'learning_rate/qwen_vl_interface': 9.053068578000256e-06, 'learning_rate/action_model': 9.053068578000256e-05, 'epoch': 0.08}) +[2026-10-03 19:08:39,720][starVLA.training.train_starvla][INFO] - Step 1103, Loss: {'train/loss': 0.07359612733125687, 'train/grad_norm_pre_clip': 33.471195220947266, 'timing/data': 0.0002771800209302455, 'timing/model': 0.39543609900283627, 'learning_rate/qwen_vl_interface': 9.051243169742372e-06, 'learning_rate/action_model': 9.051243169742372e-05, 'epoch': 0.08}) +[2026-10-03 19:08:40,431][starVLA.training.train_starvla][INFO] - Step 1104, Loss: {'train/loss': 0.009241579100489616, 'train/grad_norm_pre_clip': 1.9310157299041748, 'timing/data': 0.0003559500037226826, 'timing/model': 0.39643703898764215, 'learning_rate/qwen_vl_interface': 9.049416198937008e-06, 'learning_rate/action_model': 9.049416198937007e-05, 'epoch': 0.08}) +[2026-10-03 19:08:41,140][starVLA.training.train_starvla][INFO] - Step 1105, Loss: {'train/loss': 0.037795983254909515, 'train/grad_norm_pre_clip': 14.77550983428955, 'timing/data': 0.00035127002047374845, 'timing/model': 0.39526353901601397, 'learning_rate/qwen_vl_interface': 9.047587666335159e-06, 'learning_rate/action_model': 9.047587666335159e-05, 'epoch': 0.08}) +[2026-10-03 19:08:42,081][starVLA.training.train_starvla][INFO] - Step 1106, Loss: {'train/loss': 0.031520672142505646, 'train/grad_norm_pre_clip': 6.345688819885254, 'timing/data': 0.00039255002047866583, 'timing/model': 0.39650550900842063, 'learning_rate/qwen_vl_interface': 9.045757572688469e-06, 'learning_rate/action_model': 9.04575757268847e-05, 'epoch': 0.08}) +[2026-10-03 19:08:42,793][starVLA.training.train_starvla][INFO] - Step 1107, Loss: {'train/loss': 0.031839434057474136, 'train/grad_norm_pre_clip': 16.207441329956055, 'timing/data': 0.00026867000269703567, 'timing/model': 0.3979736510082148, 'learning_rate/qwen_vl_interface': 9.043925918749218e-06, 'learning_rate/action_model': 9.043925918749218e-05, 'epoch': 0.08}) +[2026-10-03 19:08:43,504][starVLA.training.train_starvla][INFO] - Step 1108, Loss: {'train/loss': 0.08752226829528809, 'train/grad_norm_pre_clip': 131.92091369628906, 'timing/data': 0.00026053900364786386, 'timing/model': 0.3963165589957498, 'learning_rate/qwen_vl_interface': 9.042092705270333e-06, 'learning_rate/action_model': 9.042092705270333e-05, 'epoch': 0.08}) +[2026-10-03 19:08:44,212][starVLA.training.train_starvla][INFO] - Step 1109, Loss: {'train/loss': 0.032050080597400665, 'train/grad_norm_pre_clip': 64.21076965332031, 'timing/data': 0.00025812999228946865, 'timing/model': 0.395472909003729, 'learning_rate/qwen_vl_interface': 9.040257933005375e-06, 'learning_rate/action_model': 9.040257933005374e-05, 'epoch': 0.08}) +[2026-10-03 19:08:44,924][starVLA.training.train_starvla][INFO] - Step 1110, Loss: {'train/loss': 0.0011310920817777514, 'train/grad_norm_pre_clip': 0.23327171802520752, 'timing/data': 0.0002381899976171553, 'timing/model': 0.3976355089980643, 'learning_rate/qwen_vl_interface': 9.038421602708552e-06, 'learning_rate/action_model': 9.038421602708552e-05, 'epoch': 0.08}) +[2026-10-03 19:08:45,841][starVLA.training.train_starvla][INFO] - Step 1111, Loss: {'train/loss': 0.5102654099464417, 'train/grad_norm_pre_clip': 64.9286117553711, 'timing/data': 0.0002738799958024174, 'timing/model': 0.39764079099404626, 'learning_rate/qwen_vl_interface': 9.036583715134709e-06, 'learning_rate/action_model': 9.036583715134708e-05, 'epoch': 0.08}) +[2026-10-03 19:08:46,550][starVLA.training.train_starvla][INFO] - Step 1112, Loss: {'train/loss': 0.6391457915306091, 'train/grad_norm_pre_clip': 194.95143127441406, 'timing/data': 0.0002439199888613075, 'timing/model': 0.3965372790116817, 'learning_rate/qwen_vl_interface': 9.034744271039332e-06, 'learning_rate/action_model': 9.034744271039332e-05, 'epoch': 0.08}) +[2026-10-03 19:08:47,259][starVLA.training.train_starvla][INFO] - Step 1113, Loss: {'train/loss': 0.0724564790725708, 'train/grad_norm_pre_clip': 178.11558532714844, 'timing/data': 0.0003502399777062237, 'timing/model': 0.3955765290011186, 'learning_rate/qwen_vl_interface': 9.03290327117855e-06, 'learning_rate/action_model': 9.032903271178549e-05, 'epoch': 0.08}) +[2026-10-03 19:08:47,969][starVLA.training.train_starvla][INFO] - Step 1114, Loss: {'train/loss': 0.2605675458908081, 'train/grad_norm_pre_clip': 39.302215576171875, 'timing/data': 0.0003769999893847853, 'timing/model': 0.396221389004495, 'learning_rate/qwen_vl_interface': 9.031060716309122e-06, 'learning_rate/action_model': 9.031060716309121e-05, 'epoch': 0.08}) +[2026-10-03 19:08:48,682][starVLA.training.train_starvla][INFO] - Step 1115, Loss: {'train/loss': 0.09273992478847504, 'train/grad_norm_pre_clip': 30.357059478759766, 'timing/data': 0.0002703100035432726, 'timing/model': 0.397899907984538, 'learning_rate/qwen_vl_interface': 9.02921660718846e-06, 'learning_rate/action_model': 9.029216607188459e-05, 'epoch': 0.08}) +[2026-10-03 19:08:49,595][starVLA.training.train_starvla][INFO] - Step 1116, Loss: {'train/loss': 0.001426431117579341, 'train/grad_norm_pre_clip': 0.2979922592639923, 'timing/data': 0.00021902000298723578, 'timing/model': 0.6013146279728971, 'learning_rate/qwen_vl_interface': 9.027370944574604e-06, 'learning_rate/action_model': 9.027370944574604e-05, 'epoch': 0.08}) +[2026-10-03 19:08:50,305][starVLA.training.train_starvla][INFO] - Step 1117, Loss: {'train/loss': 0.01703951507806778, 'train/grad_norm_pre_clip': 37.9666748046875, 'timing/data': 0.0005142000154592097, 'timing/model': 0.3975689880026039, 'learning_rate/qwen_vl_interface': 9.025523729226236e-06, 'learning_rate/action_model': 9.025523729226236e-05, 'epoch': 0.08}) +[2026-10-03 19:08:51,012][starVLA.training.train_starvla][INFO] - Step 1118, Loss: {'train/loss': 0.07199613004922867, 'train/grad_norm_pre_clip': 22.025657653808594, 'timing/data': 0.0003709500015247613, 'timing/model': 0.3958854190132115, 'learning_rate/qwen_vl_interface': 9.023674961902677e-06, 'learning_rate/action_model': 9.023674961902676e-05, 'epoch': 0.08}) +[2026-10-03 19:08:51,724][starVLA.training.train_starvla][INFO] - Step 1119, Loss: {'train/loss': 0.004527579061686993, 'train/grad_norm_pre_clip': 9.05324649810791, 'timing/data': 0.0002686000079847872, 'timing/model': 0.39799070899607614, 'learning_rate/qwen_vl_interface': 9.021824643363883e-06, 'learning_rate/action_model': 9.021824643363883e-05, 'epoch': 0.08}) +[2026-10-03 19:08:52,431][starVLA.training.train_starvla][INFO] - Step 1120, Loss: {'train/loss': 0.36338919401168823, 'train/grad_norm_pre_clip': 63.28094482421875, 'timing/data': 0.0003469600051175803, 'timing/model': 0.39514514899929054, 'learning_rate/qwen_vl_interface': 9.019972774370456e-06, 'learning_rate/action_model': 9.019972774370455e-05, 'epoch': 0.08}) +[2026-10-03 19:08:53,339][starVLA.training.train_starvla][INFO] - Step 1121, Loss: {'train/loss': 0.0011271751718595624, 'train/grad_norm_pre_clip': 0.3572145700454712, 'timing/data': 0.00026006001280620694, 'timing/model': 0.5951226089964621, 'learning_rate/qwen_vl_interface': 9.018119355683623e-06, 'learning_rate/action_model': 9.018119355683623e-05, 'epoch': 0.08}) +[2026-10-03 19:08:54,049][starVLA.training.train_starvla][INFO] - Step 1122, Loss: {'train/loss': 0.015657396987080574, 'train/grad_norm_pre_clip': 13.757272720336914, 'timing/data': 0.00039139000000432134, 'timing/model': 0.39498689901665784, 'learning_rate/qwen_vl_interface': 9.016264388065258e-06, 'learning_rate/action_model': 9.016264388065258e-05, 'epoch': 0.08}) +[2026-10-03 19:08:54,765][starVLA.training.train_starvla][INFO] - Step 1123, Loss: {'train/loss': 0.2667171359062195, 'train/grad_norm_pre_clip': 156.84849548339844, 'timing/data': 0.0002830800076480955, 'timing/model': 0.39815797901246697, 'learning_rate/qwen_vl_interface': 9.014407872277869e-06, 'learning_rate/action_model': 9.014407872277869e-05, 'epoch': 0.08}) +[2026-10-03 19:08:55,476][starVLA.training.train_starvla][INFO] - Step 1124, Loss: {'train/loss': 0.010094226337969303, 'train/grad_norm_pre_clip': 8.778669357299805, 'timing/data': 0.00022923998767510056, 'timing/model': 0.39713774897973053, 'learning_rate/qwen_vl_interface': 9.012549809084595e-06, 'learning_rate/action_model': 9.012549809084595e-05, 'epoch': 0.08}) +[2026-10-03 19:08:56,186][starVLA.training.train_starvla][INFO] - Step 1125, Loss: {'train/loss': 0.4940807819366455, 'train/grad_norm_pre_clip': 169.14149475097656, 'timing/data': 0.00035307000507600605, 'timing/model': 0.3958028689958155, 'learning_rate/qwen_vl_interface': 9.01069019924922e-06, 'learning_rate/action_model': 9.010690199249219e-05, 'epoch': 0.08}) +[2026-10-03 19:08:57,100][starVLA.training.train_starvla][INFO] - Step 1126, Loss: {'train/loss': 0.5224469304084778, 'train/grad_norm_pre_clip': 150.07235717773438, 'timing/data': 0.00035225998726673424, 'timing/model': 0.6013133380038198, 'learning_rate/qwen_vl_interface': 9.008829043536157e-06, 'learning_rate/action_model': 9.008829043536156e-05, 'epoch': 0.08}) +[2026-10-03 19:08:57,813][starVLA.training.train_starvla][INFO] - Step 1127, Loss: {'train/loss': 0.915809154510498, 'train/grad_norm_pre_clip': 69.15858459472656, 'timing/data': 0.00024566997308284044, 'timing/model': 0.398483088996727, 'learning_rate/qwen_vl_interface': 9.006966342710458e-06, 'learning_rate/action_model': 9.006966342710457e-05, 'epoch': 0.08}) +[2026-10-03 19:08:58,524][starVLA.training.train_starvla][INFO] - Step 1128, Loss: {'train/loss': 0.5767691135406494, 'train/grad_norm_pre_clip': 159.1990203857422, 'timing/data': 0.00023045999114401639, 'timing/model': 0.3967421790002845, 'learning_rate/qwen_vl_interface': 9.005102097537809e-06, 'learning_rate/action_model': 9.005102097537809e-05, 'epoch': 0.08}) +[2026-10-03 19:08:59,234][starVLA.training.train_starvla][INFO] - Step 1129, Loss: {'train/loss': 0.48319390416145325, 'train/grad_norm_pre_clip': 24.97064971923828, 'timing/data': 0.0003273199836257845, 'timing/model': 0.396789007994812, 'learning_rate/qwen_vl_interface': 9.00323630878453e-06, 'learning_rate/action_model': 9.00323630878453e-05, 'epoch': 0.08}) +[2026-10-03 19:08:59,942][starVLA.training.train_starvla][INFO] - Step 1130, Loss: {'train/loss': 0.02389254793524742, 'train/grad_norm_pre_clip': 56.9410400390625, 'timing/data': 0.00048067999887280166, 'timing/model': 0.3954169990029186, 'learning_rate/qwen_vl_interface': 9.001368977217577e-06, 'learning_rate/action_model': 9.001368977217577e-05, 'epoch': 0.08}) +[2026-10-03 19:09:00,674][starVLA.training.train_starvla][INFO] - Step 1131, Loss: {'train/loss': 0.0004592777113430202, 'train/grad_norm_pre_clip': 0.1562642753124237, 'timing/data': 0.00027720999787561595, 'timing/model': 0.407558978011366, 'learning_rate/qwen_vl_interface': 8.999500103604538e-06, 'learning_rate/action_model': 8.999500103604536e-05, 'epoch': 0.08}) +[2026-10-03 19:09:01,586][starVLA.training.train_starvla][INFO] - Step 1132, Loss: {'train/loss': 0.5260682702064514, 'train/grad_norm_pre_clip': 15.795037269592285, 'timing/data': 0.0003301599936094135, 'timing/model': 0.3937500390165951, 'learning_rate/qwen_vl_interface': 8.997629688713639e-06, 'learning_rate/action_model': 8.997629688713638e-05, 'epoch': 0.08}) +[2026-10-03 19:09:02,293][starVLA.training.train_starvla][INFO] - Step 1133, Loss: {'train/loss': 0.3609617054462433, 'train/grad_norm_pre_clip': 114.34195709228516, 'timing/data': 0.0002140600117854774, 'timing/model': 0.3955937790160533, 'learning_rate/qwen_vl_interface': 8.995757733313733e-06, 'learning_rate/action_model': 8.995757733313732e-05, 'epoch': 0.08}) +[2026-10-03 19:09:03,000][starVLA.training.train_starvla][INFO] - Step 1134, Loss: {'train/loss': 0.19184914231300354, 'train/grad_norm_pre_clip': 69.6975326538086, 'timing/data': 0.00045965000754222274, 'timing/model': 0.39489927899558097, 'learning_rate/qwen_vl_interface': 8.993884238174314e-06, 'learning_rate/action_model': 8.993884238174314e-05, 'epoch': 0.08}) +[2026-10-03 19:09:03,729][starVLA.training.train_starvla][INFO] - Step 1135, Loss: {'train/loss': 0.2920641005039215, 'train/grad_norm_pre_clip': 88.91390228271484, 'timing/data': 0.00034057997982017696, 'timing/model': 0.4057607580034528, 'learning_rate/qwen_vl_interface': 8.992009204065505e-06, 'learning_rate/action_model': 8.992009204065505e-05, 'epoch': 0.08}) +[2026-10-03 19:09:04,436][starVLA.training.train_starvla][INFO] - Step 1136, Loss: {'train/loss': 0.1548369824886322, 'train/grad_norm_pre_clip': 37.008792877197266, 'timing/data': 0.0003730790049303323, 'timing/model': 0.39399314002366737, 'learning_rate/qwen_vl_interface': 8.990132631758058e-06, 'learning_rate/action_model': 8.990132631758059e-05, 'epoch': 0.08}) +[2026-10-03 19:09:05,345][starVLA.training.train_starvla][INFO] - Step 1137, Loss: {'train/loss': 1.355877161026001, 'train/grad_norm_pre_clip': 135.33160400390625, 'timing/data': 0.00021389999892562628, 'timing/model': 0.5975100089854095, 'learning_rate/qwen_vl_interface': 8.988254522023364e-06, 'learning_rate/action_model': 8.988254522023363e-05, 'epoch': 0.08}) +[2026-10-03 19:09:06,051][starVLA.training.train_starvla][INFO] - Step 1138, Loss: {'train/loss': 0.5923457741737366, 'train/grad_norm_pre_clip': 213.67465209960938, 'timing/data': 0.00022668999736197293, 'timing/model': 0.3951253790000919, 'learning_rate/qwen_vl_interface': 8.986374875633441e-06, 'learning_rate/action_model': 8.98637487563344e-05, 'epoch': 0.08}) +[2026-10-03 19:09:06,762][starVLA.training.train_starvla][INFO] - Step 1139, Loss: {'train/loss': 0.14599105715751648, 'train/grad_norm_pre_clip': 80.22929382324219, 'timing/data': 0.0002332199946977198, 'timing/model': 0.39803551899967715, 'learning_rate/qwen_vl_interface': 8.984493693360944e-06, 'learning_rate/action_model': 8.984493693360943e-05, 'epoch': 0.08}) +[2026-10-03 19:09:07,470][starVLA.training.train_starvla][INFO] - Step 1140, Loss: {'train/loss': 0.10989277809858322, 'train/grad_norm_pre_clip': 33.851806640625, 'timing/data': 0.00036777000059373677, 'timing/model': 0.3951559989945963, 'learning_rate/qwen_vl_interface': 8.982610975979153e-06, 'learning_rate/action_model': 8.982610975979152e-05, 'epoch': 0.08}) +[2026-10-03 19:09:08,180][starVLA.training.train_starvla][INFO] - Step 1141, Loss: {'train/loss': 0.039131809026002884, 'train/grad_norm_pre_clip': 15.720549583435059, 'timing/data': 0.0003274699847679585, 'timing/model': 0.3971441889880225, 'learning_rate/qwen_vl_interface': 8.98072672426198e-06, 'learning_rate/action_model': 8.980726724261981e-05, 'epoch': 0.08}) +[2026-10-03 19:09:08,889][starVLA.training.train_starvla][INFO] - Step 1142, Loss: {'train/loss': 0.1806013137102127, 'train/grad_norm_pre_clip': 122.39141082763672, 'timing/data': 0.0003403400187380612, 'timing/model': 0.39591929901507683, 'learning_rate/qwen_vl_interface': 8.978840938983977e-06, 'learning_rate/action_model': 8.978840938983977e-05, 'epoch': 0.08}) +[2026-10-03 19:09:09,802][starVLA.training.train_starvla][INFO] - Step 1143, Loss: {'train/loss': 0.7010778188705444, 'train/grad_norm_pre_clip': 20.782466888427734, 'timing/data': 0.00034696998773142695, 'timing/model': 0.396336108999094, 'learning_rate/qwen_vl_interface': 8.976953620920313e-06, 'learning_rate/action_model': 8.976953620920312e-05, 'epoch': 0.08}) +[2026-10-03 19:09:10,510][starVLA.training.train_starvla][INFO] - Step 1144, Loss: {'train/loss': 0.03818674758076668, 'train/grad_norm_pre_clip': 11.845294952392578, 'timing/data': 0.0002272200072184205, 'timing/model': 0.39667368901427835, 'learning_rate/qwen_vl_interface': 8.975064770846794e-06, 'learning_rate/action_model': 8.975064770846794e-05, 'epoch': 0.08}) +[2026-10-03 19:09:11,218][starVLA.training.train_starvla][INFO] - Step 1145, Loss: {'train/loss': 0.38480472564697266, 'train/grad_norm_pre_clip': 101.679443359375, 'timing/data': 0.00025551000726409256, 'timing/model': 0.39504980901256204, 'learning_rate/qwen_vl_interface': 8.973174389539856e-06, 'learning_rate/action_model': 8.973174389539855e-05, 'epoch': 0.08}) +[2026-10-03 19:09:11,923][starVLA.training.train_starvla][INFO] - Step 1146, Loss: {'train/loss': 0.4074643850326538, 'train/grad_norm_pre_clip': 30.65159034729004, 'timing/data': 0.0003758799866773188, 'timing/model': 0.3946673200116493, 'learning_rate/qwen_vl_interface': 8.971282477776563e-06, 'learning_rate/action_model': 8.971282477776563e-05, 'epoch': 0.08}) +[2026-10-03 19:09:12,635][starVLA.training.train_starvla][INFO] - Step 1147, Loss: {'train/loss': 0.01184280589222908, 'train/grad_norm_pre_clip': 3.210960626602173, 'timing/data': 0.0002735000161919743, 'timing/model': 0.39789319899864495, 'learning_rate/qwen_vl_interface': 8.969389036334609e-06, 'learning_rate/action_model': 8.969389036334609e-05, 'epoch': 0.08}) +[2026-10-03 19:09:13,555][starVLA.training.train_starvla][INFO] - Step 1148, Loss: {'train/loss': 0.002711860928684473, 'train/grad_norm_pre_clip': 0.7929675579071045, 'timing/data': 0.00027593999402597547, 'timing/model': 0.39688980899518356, 'learning_rate/qwen_vl_interface': 8.967494065992315e-06, 'learning_rate/action_model': 8.967494065992315e-05, 'epoch': 0.08}) +[2026-10-03 19:09:14,263][starVLA.training.train_starvla][INFO] - Step 1149, Loss: {'train/loss': 0.05518285185098648, 'train/grad_norm_pre_clip': 15.228886604309082, 'timing/data': 0.00035322000621818006, 'timing/model': 0.3950927789846901, 'learning_rate/qwen_vl_interface': 8.965597567528634e-06, 'learning_rate/action_model': 8.965597567528633e-05, 'epoch': 0.08}) +[2026-10-03 19:09:14,971][starVLA.training.train_starvla][INFO] - Step 1150, Loss: {'train/loss': 0.2654130458831787, 'train/grad_norm_pre_clip': 95.62947082519531, 'timing/data': 0.0003500989987514913, 'timing/model': 0.3948558599804528, 'learning_rate/qwen_vl_interface': 8.963699541723143e-06, 'learning_rate/action_model': 8.963699541723142e-05, 'epoch': 0.08}) +[2026-10-03 19:09:15,681][starVLA.training.train_starvla][INFO] - Step 1151, Loss: {'train/loss': 0.24996408820152283, 'train/grad_norm_pre_clip': 142.0355224609375, 'timing/data': 0.00039209998794831336, 'timing/model': 0.396497209003428, 'learning_rate/qwen_vl_interface': 8.961799989356048e-06, 'learning_rate/action_model': 8.961799989356048e-05, 'epoch': 0.08}) +[2026-10-03 19:09:16,389][starVLA.training.train_starvla][INFO] - Step 1152, Loss: {'train/loss': 0.13104234635829926, 'train/grad_norm_pre_clip': 14.158289909362793, 'timing/data': 0.00026524998247623444, 'timing/model': 0.3953767089988105, 'learning_rate/qwen_vl_interface': 8.959898911208185e-06, 'learning_rate/action_model': 8.959898911208185e-05, 'epoch': 0.08}) +[2026-10-03 19:09:17,301][starVLA.training.train_starvla][INFO] - Step 1153, Loss: {'train/loss': 0.44863128662109375, 'train/grad_norm_pre_clip': 43.41545486450195, 'timing/data': 0.000265110022155568, 'timing/model': 0.5985868090065196, 'learning_rate/qwen_vl_interface': 8.957996308061015e-06, 'learning_rate/action_model': 8.957996308061014e-05, 'epoch': 0.08}) +[2026-10-03 19:09:18,007][starVLA.training.train_starvla][INFO] - Step 1154, Loss: {'train/loss': 0.004086917266249657, 'train/grad_norm_pre_clip': 1.6195244789123535, 'timing/data': 0.00022548000561073422, 'timing/model': 0.39553992901346646, 'learning_rate/qwen_vl_interface': 8.956092180696626e-06, 'learning_rate/action_model': 8.956092180696625e-05, 'epoch': 0.08}) +[2026-10-03 19:09:18,718][starVLA.training.train_starvla][INFO] - Step 1155, Loss: {'train/loss': 0.033632270991802216, 'train/grad_norm_pre_clip': 6.753813743591309, 'timing/data': 0.0002448099839966744, 'timing/model': 0.3967234590090811, 'learning_rate/qwen_vl_interface': 8.95418652989773e-06, 'learning_rate/action_model': 8.95418652989773e-05, 'epoch': 0.08}) +[2026-10-03 19:09:19,425][starVLA.training.train_starvla][INFO] - Step 1156, Loss: {'train/loss': 0.26087507605552673, 'train/grad_norm_pre_clip': 65.36089324951172, 'timing/data': 0.0003581299970392138, 'timing/model': 0.39409901900216937, 'learning_rate/qwen_vl_interface': 8.952279356447675e-06, 'learning_rate/action_model': 8.952279356447675e-05, 'epoch': 0.08}) +[2026-10-03 19:09:20,134][starVLA.training.train_starvla][INFO] - Step 1157, Loss: {'train/loss': 0.49879300594329834, 'train/grad_norm_pre_clip': 80.80522155761719, 'timing/data': 0.0003338599926792085, 'timing/model': 0.39586431899806485, 'learning_rate/qwen_vl_interface': 8.950370661130421e-06, 'learning_rate/action_model': 8.950370661130422e-05, 'epoch': 0.08}) +[2026-10-03 19:09:20,841][starVLA.training.train_starvla][INFO] - Step 1158, Loss: {'train/loss': 0.0034102832432836294, 'train/grad_norm_pre_clip': 0.5726668238639832, 'timing/data': 0.0003330300096422434, 'timing/model': 0.3957179090066347, 'learning_rate/qwen_vl_interface': 8.948460444730566e-06, 'learning_rate/action_model': 8.948460444730566e-05, 'epoch': 0.08}) +[2026-10-03 19:09:21,748][starVLA.training.train_starvla][INFO] - Step 1159, Loss: {'train/loss': 0.05132207274436951, 'train/grad_norm_pre_clip': 7.330568313598633, 'timing/data': 0.00022617998183704913, 'timing/model': 0.3960266299836803, 'learning_rate/qwen_vl_interface': 8.946548708033324e-06, 'learning_rate/action_model': 8.946548708033323e-05, 'epoch': 0.08}) +[2026-10-03 19:09:22,458][starVLA.training.train_starvla][INFO] - Step 1160, Loss: {'train/loss': 0.7631276845932007, 'train/grad_norm_pre_clip': 121.32366180419922, 'timing/data': 0.00037455998244695365, 'timing/model': 0.3958547090005595, 'learning_rate/qwen_vl_interface': 8.944635451824538e-06, 'learning_rate/action_model': 8.944635451824538e-05, 'epoch': 0.08}) +[2026-10-03 19:09:23,167][starVLA.training.train_starvla][INFO] - Step 1161, Loss: {'train/loss': 0.40646108984947205, 'train/grad_norm_pre_clip': 21.31399917602539, 'timing/data': 0.00023241000599227846, 'timing/model': 0.39431164899724536, 'learning_rate/qwen_vl_interface': 8.942720676890677e-06, 'learning_rate/action_model': 8.942720676890677e-05, 'epoch': 0.08}) +[2026-10-03 19:09:23,875][starVLA.training.train_starvla][INFO] - Step 1162, Loss: {'train/loss': 0.5455158948898315, 'train/grad_norm_pre_clip': 75.71043395996094, 'timing/data': 0.00036559998989105225, 'timing/model': 0.39630733800004236, 'learning_rate/qwen_vl_interface': 8.940804384018833e-06, 'learning_rate/action_model': 8.940804384018832e-05, 'epoch': 0.09}) +[2026-10-03 19:09:24,586][starVLA.training.train_starvla][INFO] - Step 1163, Loss: {'train/loss': 0.002593918237835169, 'train/grad_norm_pre_clip': 0.7447713613510132, 'timing/data': 0.0002738900075200945, 'timing/model': 0.3973144789924845, 'learning_rate/qwen_vl_interface': 8.93888657399672e-06, 'learning_rate/action_model': 8.93888657399672e-05, 'epoch': 0.09}) +[2026-10-03 19:09:25,489][starVLA.training.train_starvla][INFO] - Step 1164, Loss: {'train/loss': 0.42992064356803894, 'train/grad_norm_pre_clip': 151.27883911132812, 'timing/data': 0.00029872998129576445, 'timing/model': 0.39636133101885207, 'learning_rate/qwen_vl_interface': 8.936967247612678e-06, 'learning_rate/action_model': 8.936967247612678e-05, 'epoch': 0.09}) +[2026-10-03 19:09:26,196][starVLA.training.train_starvla][INFO] - Step 1165, Loss: {'train/loss': 0.00426825974136591, 'train/grad_norm_pre_clip': 1.2385920286178589, 'timing/data': 0.0002276199811603874, 'timing/model': 0.3948355900065508, 'learning_rate/qwen_vl_interface': 8.93504640565567e-06, 'learning_rate/action_model': 8.93504640565567e-05, 'epoch': 0.09}) +[2026-10-03 19:09:26,902][starVLA.training.train_starvla][INFO] - Step 1166, Loss: {'train/loss': 0.46106982231140137, 'train/grad_norm_pre_clip': 29.97422218322754, 'timing/data': 0.0003101699985563755, 'timing/model': 0.3945243190100882, 'learning_rate/qwen_vl_interface': 8.933124048915278e-06, 'learning_rate/action_model': 8.933124048915278e-05, 'epoch': 0.09}) +[2026-10-03 19:09:27,612][starVLA.training.train_starvla][INFO] - Step 1167, Loss: {'train/loss': 0.001279608579352498, 'train/grad_norm_pre_clip': 0.5472304821014404, 'timing/data': 0.0003769300237763673, 'timing/model': 0.3966534190112725, 'learning_rate/qwen_vl_interface': 8.931200178181715e-06, 'learning_rate/action_model': 8.931200178181715e-05, 'epoch': 0.09}) +[2026-10-03 19:09:28,322][starVLA.training.train_starvla][INFO] - Step 1168, Loss: {'train/loss': 0.06436112523078918, 'train/grad_norm_pre_clip': 40.12813949584961, 'timing/data': 0.00022641001851297915, 'timing/model': 0.3965472790005151, 'learning_rate/qwen_vl_interface': 8.929274794245812e-06, 'learning_rate/action_model': 8.929274794245812e-05, 'epoch': 0.09}) +[2026-10-03 19:09:29,231][starVLA.training.train_starvla][INFO] - Step 1169, Loss: {'train/loss': 0.003993701655417681, 'train/grad_norm_pre_clip': 12.619011878967285, 'timing/data': 0.00023373001022264361, 'timing/model': 0.5956356190145016, 'learning_rate/qwen_vl_interface': 8.927347897899018e-06, 'learning_rate/action_model': 8.927347897899016e-05, 'epoch': 0.09}) +[2026-10-03 19:09:29,938][starVLA.training.train_starvla][INFO] - Step 1170, Loss: {'train/loss': 0.10586342215538025, 'train/grad_norm_pre_clip': 63.19761657714844, 'timing/data': 0.0004102600214537233, 'timing/model': 0.3951107200118713, 'learning_rate/qwen_vl_interface': 8.925419489933407e-06, 'learning_rate/action_model': 8.925419489933408e-05, 'epoch': 0.09}) +[2026-10-03 19:09:30,646][starVLA.training.train_starvla][INFO] - Step 1171, Loss: {'train/loss': 0.3857244551181793, 'train/grad_norm_pre_clip': 71.55892181396484, 'timing/data': 0.00033179999445565045, 'timing/model': 0.39532212997437455, 'learning_rate/qwen_vl_interface': 8.923489571141679e-06, 'learning_rate/action_model': 8.923489571141679e-05, 'epoch': 0.09}) +[2026-10-03 19:09:31,353][starVLA.training.train_starvla][INFO] - Step 1172, Loss: {'train/loss': 0.0014902940019965172, 'train/grad_norm_pre_clip': 0.6059622168540955, 'timing/data': 0.00026752002304419875, 'timing/model': 0.3956065689853858, 'learning_rate/qwen_vl_interface': 8.921558142317147e-06, 'learning_rate/action_model': 8.921558142317147e-05, 'epoch': 0.09}) +[2026-10-03 19:09:32,059][starVLA.training.train_starvla][INFO] - Step 1173, Loss: {'train/loss': 0.3973238170146942, 'train/grad_norm_pre_clip': 142.65383911132812, 'timing/data': 0.0002294000005349517, 'timing/model': 0.39456875901669264, 'learning_rate/qwen_vl_interface': 8.91962520425375e-06, 'learning_rate/action_model': 8.91962520425375e-05, 'epoch': 0.09}) +[2026-10-03 19:09:32,968][starVLA.training.train_starvla][INFO] - Step 1174, Loss: {'train/loss': 0.0009891968220472336, 'train/grad_norm_pre_clip': 0.2568637430667877, 'timing/data': 0.00023725000210106373, 'timing/model': 0.5958697290043347, 'learning_rate/qwen_vl_interface': 8.917690757746044e-06, 'learning_rate/action_model': 8.917690757746044e-05, 'epoch': 0.09}) +[2026-10-03 19:09:33,674][starVLA.training.train_starvla][INFO] - Step 1175, Loss: {'train/loss': 0.3967478573322296, 'train/grad_norm_pre_clip': 37.995811462402344, 'timing/data': 0.0002312700089532882, 'timing/model': 0.3960378589981701, 'learning_rate/qwen_vl_interface': 8.915754803589208e-06, 'learning_rate/action_model': 8.915754803589209e-05, 'epoch': 0.09}) +[2026-10-03 19:09:34,378][starVLA.training.train_starvla][INFO] - Step 1176, Loss: {'train/loss': 0.0017092667985707521, 'train/grad_norm_pre_clip': 0.6655009388923645, 'timing/data': 0.00031650002347305417, 'timing/model': 0.39293102000374347, 'learning_rate/qwen_vl_interface': 8.913817342579042e-06, 'learning_rate/action_model': 8.913817342579042e-05, 'epoch': 0.09}) +[2026-10-03 19:09:35,085][starVLA.training.train_starvla][INFO] - Step 1177, Loss: {'train/loss': 0.4230719208717346, 'train/grad_norm_pre_clip': 77.89031982421875, 'timing/data': 0.0003512800030875951, 'timing/model': 0.39605988998664543, 'learning_rate/qwen_vl_interface': 8.91187837551196e-06, 'learning_rate/action_model': 8.911878375511961e-05, 'epoch': 0.09}) +[2026-10-03 19:09:35,793][starVLA.training.train_starvla][INFO] - Step 1178, Loss: {'train/loss': 0.2572762668132782, 'train/grad_norm_pre_clip': 34.59458923339844, 'timing/data': 0.000357509998138994, 'timing/model': 0.3945239490130916, 'learning_rate/qwen_vl_interface': 8.909937903185e-06, 'learning_rate/action_model': 8.909937903185e-05, 'epoch': 0.09}) +[2026-10-03 19:09:36,697][starVLA.training.train_starvla][INFO] - Step 1179, Loss: {'train/loss': 0.2816605567932129, 'train/grad_norm_pre_clip': 76.00468444824219, 'timing/data': 0.00022047001402825117, 'timing/model': 0.5937638889881782, 'learning_rate/qwen_vl_interface': 8.907995926395814e-06, 'learning_rate/action_model': 8.907995926395815e-05, 'epoch': 0.09}) +[2026-10-03 19:09:37,403][starVLA.training.train_starvla][INFO] - Step 1180, Loss: {'train/loss': 0.00043267273576930165, 'train/grad_norm_pre_clip': 0.10326366871595383, 'timing/data': 0.00032627000473439693, 'timing/model': 0.39382933898014016, 'learning_rate/qwen_vl_interface': 8.90605244594268e-06, 'learning_rate/action_model': 8.906052445942679e-05, 'epoch': 0.09}) +[2026-10-03 19:09:38,110][starVLA.training.train_starvla][INFO] - Step 1181, Loss: {'train/loss': 0.0029632942751049995, 'train/grad_norm_pre_clip': 1.647283911705017, 'timing/data': 0.00034790998324751854, 'timing/model': 0.394926109991502, 'learning_rate/qwen_vl_interface': 8.904107462624484e-06, 'learning_rate/action_model': 8.904107462624483e-05, 'epoch': 0.09}) +[2026-10-03 19:09:38,814][starVLA.training.train_starvla][INFO] - Step 1182, Loss: {'train/loss': 0.24783378839492798, 'train/grad_norm_pre_clip': 64.56890869140625, 'timing/data': 0.00036281999200582504, 'timing/model': 0.39339673100039363, 'learning_rate/qwen_vl_interface': 8.902160977240739e-06, 'learning_rate/action_model': 8.902160977240739e-05, 'epoch': 0.09}) +[2026-10-03 19:09:39,520][starVLA.training.train_starvla][INFO] - Step 1183, Loss: {'train/loss': 0.014267266727983952, 'train/grad_norm_pre_clip': 9.358463287353516, 'timing/data': 0.0002563699963502586, 'timing/model': 0.3937486599897966, 'learning_rate/qwen_vl_interface': 8.90021299059157e-06, 'learning_rate/action_model': 8.90021299059157e-05, 'epoch': 0.09}) +[2026-10-03 19:09:40,227][starVLA.training.train_starvla][INFO] - Step 1184, Loss: {'train/loss': 0.04613209143280983, 'train/grad_norm_pre_clip': 47.13779067993164, 'timing/data': 0.00023428001441061497, 'timing/model': 0.39421436001430266, 'learning_rate/qwen_vl_interface': 8.898263503477722e-06, 'learning_rate/action_model': 8.898263503477721e-05, 'epoch': 0.09}) +[2026-10-03 19:09:41,137][starVLA.training.train_starvla][INFO] - Step 1185, Loss: {'train/loss': 0.42157939076423645, 'train/grad_norm_pre_clip': 183.35809326171875, 'timing/data': 0.0003654400061350316, 'timing/model': 0.3959571390005294, 'learning_rate/qwen_vl_interface': 8.896312516700553e-06, 'learning_rate/action_model': 8.896312516700554e-05, 'epoch': 0.09}) +[2026-10-03 19:09:41,840][starVLA.training.train_starvla][INFO] - Step 1186, Loss: {'train/loss': 0.553441047668457, 'train/grad_norm_pre_clip': 92.38706970214844, 'timing/data': 0.00024756998755037785, 'timing/model': 0.3933632289990783, 'learning_rate/qwen_vl_interface': 8.894360031062043e-06, 'learning_rate/action_model': 8.894360031062043e-05, 'epoch': 0.09}) +[2026-10-03 19:09:42,546][starVLA.training.train_starvla][INFO] - Step 1187, Loss: {'train/loss': 0.2735770046710968, 'train/grad_norm_pre_clip': 40.92743682861328, 'timing/data': 0.00023554000654257834, 'timing/model': 0.3941293790121563, 'learning_rate/qwen_vl_interface': 8.892406047364783e-06, 'learning_rate/action_model': 8.892406047364784e-05, 'epoch': 0.09}) +[2026-10-03 19:09:43,253][starVLA.training.train_starvla][INFO] - Step 1188, Loss: {'train/loss': 0.02721797488629818, 'train/grad_norm_pre_clip': 23.642589569091797, 'timing/data': 0.00035031000152230263, 'timing/model': 0.39436270101577975, 'learning_rate/qwen_vl_interface': 8.890450566411984e-06, 'learning_rate/action_model': 8.890450566411984e-05, 'epoch': 0.09}) +[2026-10-03 19:09:43,959][starVLA.training.train_starvla][INFO] - Step 1189, Loss: {'train/loss': 0.4522659182548523, 'train/grad_norm_pre_clip': 69.42845916748047, 'timing/data': 0.00029383000219240785, 'timing/model': 0.3944450900016818, 'learning_rate/qwen_vl_interface': 8.888493589007467e-06, 'learning_rate/action_model': 8.888493589007468e-05, 'epoch': 0.09}) +[2026-10-03 19:09:44,870][starVLA.training.train_starvla][INFO] - Step 1190, Loss: {'train/loss': 0.29916369915008545, 'train/grad_norm_pre_clip': 60.30366134643555, 'timing/data': 0.00023547999444417655, 'timing/model': 0.5993258579983376, 'learning_rate/qwen_vl_interface': 8.886535115955676e-06, 'learning_rate/action_model': 8.886535115955675e-05, 'epoch': 0.09}) +[2026-10-03 19:09:45,578][starVLA.training.train_starvla][INFO] - Step 1191, Loss: {'train/loss': 0.029634693637490273, 'train/grad_norm_pre_clip': 13.859667778015137, 'timing/data': 0.00038444000529125333, 'timing/model': 0.39544423899496906, 'learning_rate/qwen_vl_interface': 8.884575148061662e-06, 'learning_rate/action_model': 8.88457514806166e-05, 'epoch': 0.09}) +[2026-10-03 19:09:46,284][starVLA.training.train_starvla][INFO] - Step 1192, Loss: {'train/loss': 0.5440398454666138, 'train/grad_norm_pre_clip': 72.31603240966797, 'timing/data': 0.00023179998970590532, 'timing/model': 0.3947802290203981, 'learning_rate/qwen_vl_interface': 8.882613686131095e-06, 'learning_rate/action_model': 8.882613686131093e-05, 'epoch': 0.09}) +[2026-10-03 19:09:46,991][starVLA.training.train_starvla][INFO] - Step 1193, Loss: {'train/loss': 0.18935589492321014, 'train/grad_norm_pre_clip': 85.32222747802734, 'timing/data': 0.0003279100055806339, 'timing/model': 0.3955460290017072, 'learning_rate/qwen_vl_interface': 8.880650730970254e-06, 'learning_rate/action_model': 8.880650730970254e-05, 'epoch': 0.09}) +[2026-10-03 19:09:47,698][starVLA.training.train_starvla][INFO] - Step 1194, Loss: {'train/loss': 0.035758066922426224, 'train/grad_norm_pre_clip': 17.761751174926758, 'timing/data': 0.0003556099836714566, 'timing/model': 0.3952060989977326, 'learning_rate/qwen_vl_interface': 8.878686283386042e-06, 'learning_rate/action_model': 8.878686283386042e-05, 'epoch': 0.09}) +[2026-10-03 19:09:48,406][starVLA.training.train_starvla][INFO] - Step 1195, Loss: {'train/loss': 0.02316523715853691, 'train/grad_norm_pre_clip': 20.99867820739746, 'timing/data': 0.0003674199979286641, 'timing/model': 0.39555793901672587, 'learning_rate/qwen_vl_interface': 8.876720344185967e-06, 'learning_rate/action_model': 8.876720344185967e-05, 'epoch': 0.09}) +[2026-10-03 19:09:49,318][starVLA.training.train_starvla][INFO] - Step 1196, Loss: {'train/loss': 0.38520199060440063, 'train/grad_norm_pre_clip': 35.13353729248047, 'timing/data': 0.00035606001620180905, 'timing/model': 0.3953671100025531, 'learning_rate/qwen_vl_interface': 8.874752914178149e-06, 'learning_rate/action_model': 8.87475291417815e-05, 'epoch': 0.09}) +[2026-10-03 19:09:50,025][starVLA.training.train_starvla][INFO] - Step 1197, Loss: {'train/loss': 0.4008138179779053, 'train/grad_norm_pre_clip': 162.4038543701172, 'timing/data': 0.00025744998129084706, 'timing/model': 0.395780308987014, 'learning_rate/qwen_vl_interface': 8.87278399417133e-06, 'learning_rate/action_model': 8.872783994171329e-05, 'epoch': 0.09}) +[2026-10-03 19:09:50,730][starVLA.training.train_starvla][INFO] - Step 1198, Loss: {'train/loss': 0.030077073723077774, 'train/grad_norm_pre_clip': 13.199786186218262, 'timing/data': 0.00022585998522117734, 'timing/model': 0.3940045600174926, 'learning_rate/qwen_vl_interface': 8.870813584974852e-06, 'learning_rate/action_model': 8.870813584974852e-05, 'epoch': 0.09}) +[2026-10-03 19:09:51,436][starVLA.training.train_starvla][INFO] - Step 1199, Loss: {'train/loss': 0.022894226014614105, 'train/grad_norm_pre_clip': 32.77558135986328, 'timing/data': 0.00023184000747278333, 'timing/model': 0.39495589898433536, 'learning_rate/qwen_vl_interface': 8.868841687398683e-06, 'learning_rate/action_model': 8.868841687398683e-05, 'epoch': 0.09}) +[2026-10-03 19:09:52,141][starVLA.training.train_starvla][INFO] - Step 1200, Loss: {'train/loss': 0.006587055511772633, 'train/grad_norm_pre_clip': 2.7734768390655518, 'timing/data': 0.00021415000082924962, 'timing/model': 0.39491233899025247, 'learning_rate/qwen_vl_interface': 8.86686830225339e-06, 'learning_rate/action_model': 8.866868302253389e-05, 'epoch': 0.09}) +[2026-10-03 19:09:53,049][starVLA.training.train_starvla][INFO] - Step 1201, Loss: {'train/loss': 0.3622859716415405, 'train/grad_norm_pre_clip': 78.9913558959961, 'timing/data': 0.0003540600009728223, 'timing/model': 0.3940007009950932, 'learning_rate/qwen_vl_interface': 8.864893430350159e-06, 'learning_rate/action_model': 8.864893430350159e-05, 'epoch': 0.09}) +[2026-10-03 19:09:53,755][starVLA.training.train_starvla][INFO] - Step 1202, Loss: {'train/loss': 0.1273648589849472, 'train/grad_norm_pre_clip': 73.32545471191406, 'timing/data': 0.0002216600114479661, 'timing/model': 0.3952595900045708, 'learning_rate/qwen_vl_interface': 8.862917072500785e-06, 'learning_rate/action_model': 8.862917072500784e-05, 'epoch': 0.09}) +[2026-10-03 19:09:54,461][starVLA.training.train_starvla][INFO] - Step 1203, Loss: {'train/loss': 0.25510549545288086, 'train/grad_norm_pre_clip': 53.877933502197266, 'timing/data': 0.00032612000359222293, 'timing/model': 0.3947716699913144, 'learning_rate/qwen_vl_interface': 8.860939229517674e-06, 'learning_rate/action_model': 8.860939229517674e-05, 'epoch': 0.09}) +[2026-10-03 19:09:55,141][starVLA.training.train_starvla][INFO] - Step 1204, Loss: {'train/loss': 0.19448384642601013, 'train/grad_norm_pre_clip': 119.76566314697266, 'timing/data': 0.00024883999140001833, 'timing/model': 0.39356149997911416, 'learning_rate/qwen_vl_interface': 8.858959902213844e-06, 'learning_rate/action_model': 8.858959902213845e-05, 'epoch': 0.09}) +[2026-10-03 19:09:55,849][starVLA.training.train_starvla][INFO] - Step 1205, Loss: {'train/loss': 0.35507917404174805, 'train/grad_norm_pre_clip': 51.78270721435547, 'timing/data': 0.0003787199966609478, 'timing/model': 0.3955505790072493, 'learning_rate/qwen_vl_interface': 8.856979091402921e-06, 'learning_rate/action_model': 8.85697909140292e-05, 'epoch': 0.09}) +[2026-10-03 19:09:56,761][starVLA.training.train_starvla][INFO] - Step 1206, Loss: {'train/loss': 0.005530126858502626, 'train/grad_norm_pre_clip': 7.076172351837158, 'timing/data': 0.00022982998052611947, 'timing/model': 0.5996211279998533, 'learning_rate/qwen_vl_interface': 8.854996797899139e-06, 'learning_rate/action_model': 8.854996797899139e-05, 'epoch': 0.09}) +[2026-10-03 19:09:57,467][starVLA.training.train_starvla][INFO] - Step 1207, Loss: {'train/loss': 0.03128787502646446, 'train/grad_norm_pre_clip': 10.648090362548828, 'timing/data': 0.0003183599910698831, 'timing/model': 0.393804918974638, 'learning_rate/qwen_vl_interface': 8.853013022517348e-06, 'learning_rate/action_model': 8.853013022517348e-05, 'epoch': 0.09}) +[2026-10-03 19:09:58,175][starVLA.training.train_starvla][INFO] - Step 1208, Loss: {'train/loss': 0.31291884183883667, 'train/grad_norm_pre_clip': 126.05592346191406, 'timing/data': 0.00039824002305977046, 'timing/model': 0.39524779000203125, 'learning_rate/qwen_vl_interface': 8.851027766073002e-06, 'learning_rate/action_model': 8.851027766073001e-05, 'epoch': 0.09}) +[2026-10-03 19:09:58,882][starVLA.training.train_starvla][INFO] - Step 1209, Loss: {'train/loss': 0.7450188994407654, 'train/grad_norm_pre_clip': 46.73051071166992, 'timing/data': 0.0003311400068923831, 'timing/model': 0.39474051000433974, 'learning_rate/qwen_vl_interface': 8.84904102938216e-06, 'learning_rate/action_model': 8.84904102938216e-05, 'epoch': 0.09}) +[2026-10-03 19:09:59,589][starVLA.training.train_starvla][INFO] - Step 1210, Loss: {'train/loss': 0.006510578561574221, 'train/grad_norm_pre_clip': 3.8035709857940674, 'timing/data': 0.00036926998291164637, 'timing/model': 0.3952734209888149, 'learning_rate/qwen_vl_interface': 8.8470528132615e-06, 'learning_rate/action_model': 8.8470528132615e-05, 'epoch': 0.09}) +[2026-10-03 19:10:00,295][starVLA.training.train_starvla][INFO] - Step 1211, Loss: {'train/loss': 0.20410284399986267, 'train/grad_norm_pre_clip': 88.68988037109375, 'timing/data': 0.0004345499910414219, 'timing/model': 0.3951929309987463, 'learning_rate/qwen_vl_interface': 8.845063118528303e-06, 'learning_rate/action_model': 8.845063118528304e-05, 'epoch': 0.09}) +[2026-10-03 19:10:01,207][starVLA.training.train_starvla][INFO] - Step 1212, Loss: {'train/loss': 0.386088490486145, 'train/grad_norm_pre_clip': 44.7659912109375, 'timing/data': 0.0003994799917563796, 'timing/model': 0.39565834900713526, 'learning_rate/qwen_vl_interface': 8.843071946000455e-06, 'learning_rate/action_model': 8.843071946000456e-05, 'epoch': 0.09}) +[2026-10-03 19:10:01,913][starVLA.training.train_starvla][INFO] - Step 1213, Loss: {'train/loss': 0.18802540004253387, 'train/grad_norm_pre_clip': 58.98603439331055, 'timing/data': 0.0003349499893374741, 'timing/model': 0.39434638898819685, 'learning_rate/qwen_vl_interface': 8.841079296496451e-06, 'learning_rate/action_model': 8.84107929649645e-05, 'epoch': 0.09}) +[2026-10-03 19:10:02,620][starVLA.training.train_starvla][INFO] - Step 1214, Loss: {'train/loss': 0.6504797339439392, 'train/grad_norm_pre_clip': 26.072420120239258, 'timing/data': 0.00027511001098901033, 'timing/model': 0.3948588089842815, 'learning_rate/qwen_vl_interface': 8.839085170835396e-06, 'learning_rate/action_model': 8.839085170835395e-05, 'epoch': 0.09}) +[2026-10-03 19:10:03,326][starVLA.training.train_starvla][INFO] - Step 1215, Loss: {'train/loss': 0.2495294213294983, 'train/grad_norm_pre_clip': 93.01789855957031, 'timing/data': 0.00041903997771441936, 'timing/model': 0.3942094399826601, 'learning_rate/qwen_vl_interface': 8.837089569837e-06, 'learning_rate/action_model': 8.837089569836998e-05, 'epoch': 0.09}) +[2026-10-03 19:10:04,031][starVLA.training.train_starvla][INFO] - Step 1216, Loss: {'train/loss': 0.19606125354766846, 'train/grad_norm_pre_clip': 67.23758697509766, 'timing/data': 0.0003511100076138973, 'timing/model': 0.3951439590018708, 'learning_rate/qwen_vl_interface': 8.835092494321575e-06, 'learning_rate/action_model': 8.835092494321575e-05, 'epoch': 0.09}) +[2026-10-03 19:10:04,936][starVLA.training.train_starvla][INFO] - Step 1217, Loss: {'train/loss': 0.6184173822402954, 'train/grad_norm_pre_clip': 75.96660614013672, 'timing/data': 0.00022866000654175878, 'timing/model': 0.5948429889976978, 'learning_rate/qwen_vl_interface': 8.833093945110048e-06, 'learning_rate/action_model': 8.833093945110047e-05, 'epoch': 0.09}) +[2026-10-03 19:10:05,640][starVLA.training.train_starvla][INFO] - Step 1218, Loss: {'train/loss': 0.03993440419435501, 'train/grad_norm_pre_clip': 13.099563598632812, 'timing/data': 0.0002898699895013124, 'timing/model': 0.39432148000923917, 'learning_rate/qwen_vl_interface': 8.831093923023946e-06, 'learning_rate/action_model': 8.831093923023945e-05, 'epoch': 0.09}) +[2026-10-03 19:10:06,345][starVLA.training.train_starvla][INFO] - Step 1219, Loss: {'train/loss': 0.0455666184425354, 'train/grad_norm_pre_clip': 12.46094036102295, 'timing/data': 0.0003752100165002048, 'timing/model': 0.3925362190057058, 'learning_rate/qwen_vl_interface': 8.8290924288854e-06, 'learning_rate/action_model': 8.8290924288854e-05, 'epoch': 0.09}) +[2026-10-03 19:10:07,051][starVLA.training.train_starvla][INFO] - Step 1220, Loss: {'train/loss': 0.12200166285037994, 'train/grad_norm_pre_clip': 46.370113372802734, 'timing/data': 0.00023226000485010445, 'timing/model': 0.3946279799856711, 'learning_rate/qwen_vl_interface': 8.827089463517153e-06, 'learning_rate/action_model': 8.827089463517153e-05, 'epoch': 0.09}) +[2026-10-03 19:10:07,756][starVLA.training.train_starvla][INFO] - Step 1221, Loss: {'train/loss': 0.09035561233758926, 'train/grad_norm_pre_clip': 31.34065818786621, 'timing/data': 0.0003416100225877017, 'timing/model': 0.3934568500262685, 'learning_rate/qwen_vl_interface': 8.825085027742545e-06, 'learning_rate/action_model': 8.825085027742544e-05, 'epoch': 0.09}) +[2026-10-03 19:10:08,660][starVLA.training.train_starvla][INFO] - Step 1222, Loss: {'train/loss': 0.15486282110214233, 'train/grad_norm_pre_clip': 58.76179504394531, 'timing/data': 0.00035891999141313136, 'timing/model': 0.592295100010233, 'learning_rate/qwen_vl_interface': 8.823079122385523e-06, 'learning_rate/action_model': 8.823079122385523e-05, 'epoch': 0.09}) +[2026-10-03 19:10:09,363][starVLA.training.train_starvla][INFO] - Step 1223, Loss: {'train/loss': 0.04341200739145279, 'train/grad_norm_pre_clip': 15.162439346313477, 'timing/data': 0.00021649998961947858, 'timing/model': 0.3926202300062869, 'learning_rate/qwen_vl_interface': 8.821071748270644e-06, 'learning_rate/action_model': 8.821071748270644e-05, 'epoch': 0.09}) +[2026-10-03 19:10:10,066][starVLA.training.train_starvla][INFO] - Step 1224, Loss: {'train/loss': 0.31960350275039673, 'train/grad_norm_pre_clip': 79.53196716308594, 'timing/data': 0.0002252700214739889, 'timing/model': 0.39252348098671064, 'learning_rate/qwen_vl_interface': 8.81906290622306e-06, 'learning_rate/action_model': 8.81906290622306e-05, 'epoch': 0.09}) +[2026-10-03 19:10:10,772][starVLA.training.train_starvla][INFO] - Step 1225, Loss: {'train/loss': 0.46838101744651794, 'train/grad_norm_pre_clip': 107.03179931640625, 'timing/data': 0.00022190998424775898, 'timing/model': 0.3952265600091778, 'learning_rate/qwen_vl_interface': 8.81705259706853e-06, 'learning_rate/action_model': 8.817052597068529e-05, 'epoch': 0.09}) +[2026-10-03 19:10:11,475][starVLA.training.train_starvla][INFO] - Step 1226, Loss: {'train/loss': 0.11020699143409729, 'train/grad_norm_pre_clip': 78.01949310302734, 'timing/data': 0.00022201999672688544, 'timing/model': 0.3937106710218359, 'learning_rate/qwen_vl_interface': 8.815040821633416e-06, 'learning_rate/action_model': 8.815040821633416e-05, 'epoch': 0.09}) +[2026-10-03 19:10:12,180][starVLA.training.train_starvla][INFO] - Step 1227, Loss: {'train/loss': 0.2745608985424042, 'train/grad_norm_pre_clip': 38.435630798339844, 'timing/data': 0.00022914001601748168, 'timing/model': 0.39456717000575736, 'learning_rate/qwen_vl_interface': 8.813027580744685e-06, 'learning_rate/action_model': 8.813027580744684e-05, 'epoch': 0.09}) +[2026-10-03 19:10:13,082][starVLA.training.train_starvla][INFO] - Step 1228, Loss: {'train/loss': 0.6336370706558228, 'train/grad_norm_pre_clip': 70.92259979248047, 'timing/data': 0.00022312000510282815, 'timing/model': 0.3953996089985594, 'learning_rate/qwen_vl_interface': 8.8110128752299e-06, 'learning_rate/action_model': 8.8110128752299e-05, 'epoch': 0.09}) +[2026-10-03 19:10:13,787][starVLA.training.train_starvla][INFO] - Step 1229, Loss: {'train/loss': 0.6966819167137146, 'train/grad_norm_pre_clip': 124.34930419921875, 'timing/data': 0.00035172997741028666, 'timing/model': 0.39502498900401406, 'learning_rate/qwen_vl_interface': 8.808996705917235e-06, 'learning_rate/action_model': 8.808996705917235e-05, 'epoch': 0.09}) +[2026-10-03 19:10:14,491][starVLA.training.train_starvla][INFO] - Step 1230, Loss: {'train/loss': 0.09431207180023193, 'train/grad_norm_pre_clip': 22.972152709960938, 'timing/data': 0.000318280013743788, 'timing/model': 0.3942149789945688, 'learning_rate/qwen_vl_interface': 8.80697907363546e-06, 'learning_rate/action_model': 8.80697907363546e-05, 'epoch': 0.09}) +[2026-10-03 19:10:15,198][starVLA.training.train_starvla][INFO] - Step 1231, Loss: {'train/loss': 0.010637899860739708, 'train/grad_norm_pre_clip': 6.806448459625244, 'timing/data': 0.0004075199831277132, 'timing/model': 0.3945263689965941, 'learning_rate/qwen_vl_interface': 8.804959979213945e-06, 'learning_rate/action_model': 8.804959979213945e-05, 'epoch': 0.09}) +[2026-10-03 19:10:15,906][starVLA.training.train_starvla][INFO] - Step 1232, Loss: {'train/loss': 0.20187745988368988, 'train/grad_norm_pre_clip': 24.8539981842041, 'timing/data': 0.00035488998400978744, 'timing/model': 0.39505947899306193, 'learning_rate/qwen_vl_interface': 8.802939423482667e-06, 'learning_rate/action_model': 8.802939423482666e-05, 'epoch': 0.09}) +[2026-10-03 19:10:16,816][starVLA.training.train_starvla][INFO] - Step 1233, Loss: {'train/loss': 0.07526537775993347, 'train/grad_norm_pre_clip': 23.759960174560547, 'timing/data': 0.0003281999961473048, 'timing/model': 0.3957541710115038, 'learning_rate/qwen_vl_interface': 8.800917407272198e-06, 'learning_rate/action_model': 8.800917407272198e-05, 'epoch': 0.09}) +[2026-10-03 19:10:17,524][starVLA.training.train_starvla][INFO] - Step 1234, Loss: {'train/loss': 0.01386365108191967, 'train/grad_norm_pre_clip': 1.8174834251403809, 'timing/data': 0.0002216300053987652, 'timing/model': 0.39528651000000536, 'learning_rate/qwen_vl_interface': 8.798893931413711e-06, 'learning_rate/action_model': 8.79889393141371e-05, 'epoch': 0.09}) +[2026-10-03 19:10:18,229][starVLA.training.train_starvla][INFO] - Step 1235, Loss: {'train/loss': 0.08337652683258057, 'train/grad_norm_pre_clip': 31.41539192199707, 'timing/data': 0.0003612400032579899, 'timing/model': 0.3930382110120263, 'learning_rate/qwen_vl_interface': 8.796868996738985e-06, 'learning_rate/action_model': 8.796868996738984e-05, 'epoch': 0.09}) +[2026-10-03 19:10:18,933][starVLA.training.train_starvla][INFO] - Step 1236, Loss: {'train/loss': 0.04152775928378105, 'train/grad_norm_pre_clip': 11.228706359863281, 'timing/data': 0.0003046899801120162, 'timing/model': 0.3930811410245951, 'learning_rate/qwen_vl_interface': 8.794842604080389e-06, 'learning_rate/action_model': 8.794842604080389e-05, 'epoch': 0.09}) +[2026-10-03 19:10:19,640][starVLA.training.train_starvla][INFO] - Step 1237, Loss: {'train/loss': 0.17119726538658142, 'train/grad_norm_pre_clip': 35.20166015625, 'timing/data': 0.0003499200101941824, 'timing/model': 0.3955455399991479, 'learning_rate/qwen_vl_interface': 8.7928147542709e-06, 'learning_rate/action_model': 8.7928147542709e-05, 'epoch': 0.09}) +[2026-10-03 19:10:20,542][starVLA.training.train_starvla][INFO] - Step 1238, Loss: {'train/loss': 0.17488035559654236, 'train/grad_norm_pre_clip': 35.85245895385742, 'timing/data': 0.00022580998484045267, 'timing/model': 0.5910353700164706, 'learning_rate/qwen_vl_interface': 8.790785448144088e-06, 'learning_rate/action_model': 8.790785448144088e-05, 'epoch': 0.09}) +[2026-10-03 19:10:21,251][starVLA.training.train_starvla][INFO] - Step 1239, Loss: {'train/loss': 0.9218772053718567, 'train/grad_norm_pre_clip': 93.861572265625, 'timing/data': 0.00036372002796269953, 'timing/model': 0.39492900899495, 'learning_rate/qwen_vl_interface': 8.788754686534126e-06, 'learning_rate/action_model': 8.788754686534127e-05, 'epoch': 0.09}) +[2026-10-03 19:10:21,958][starVLA.training.train_starvla][INFO] - Step 1240, Loss: {'train/loss': 0.2595551311969757, 'train/grad_norm_pre_clip': 83.21951293945312, 'timing/data': 0.0003569599939510226, 'timing/model': 0.39472079899860546, 'learning_rate/qwen_vl_interface': 8.786722470275784e-06, 'learning_rate/action_model': 8.786722470275784e-05, 'epoch': 0.09}) +[2026-10-03 19:10:22,665][starVLA.training.train_starvla][INFO] - Step 1241, Loss: {'train/loss': 0.48352885246276855, 'train/grad_norm_pre_clip': 27.187009811401367, 'timing/data': 0.0003365500015206635, 'timing/model': 0.3950519790232647, 'learning_rate/qwen_vl_interface': 8.784688800204427e-06, 'learning_rate/action_model': 8.784688800204426e-05, 'epoch': 0.09}) +[2026-10-03 19:10:23,372][starVLA.training.train_starvla][INFO] - Step 1242, Loss: {'train/loss': 0.38282662630081177, 'train/grad_norm_pre_clip': 61.33849334716797, 'timing/data': 0.00022950998391024768, 'timing/model': 0.395845228980761, 'learning_rate/qwen_vl_interface': 8.782653677156022e-06, 'learning_rate/action_model': 8.782653677156022e-05, 'epoch': 0.09}) +[2026-10-03 19:10:24,277][starVLA.training.train_starvla][INFO] - Step 1243, Loss: {'train/loss': 0.13286623358726501, 'train/grad_norm_pre_clip': 34.777713775634766, 'timing/data': 0.0002977800031658262, 'timing/model': 0.5919988399837166, 'learning_rate/qwen_vl_interface': 8.780617101967129e-06, 'learning_rate/action_model': 8.780617101967129e-05, 'epoch': 0.09}) +[2026-10-03 19:10:24,983][starVLA.training.train_starvla][INFO] - Step 1244, Loss: {'train/loss': 0.6286635398864746, 'train/grad_norm_pre_clip': 116.34796142578125, 'timing/data': 0.00033128997893072665, 'timing/model': 0.3939994100073818, 'learning_rate/qwen_vl_interface': 8.77857907547491e-06, 'learning_rate/action_model': 8.77857907547491e-05, 'epoch': 0.09}) +[2026-10-03 19:10:25,689][starVLA.training.train_starvla][INFO] - Step 1245, Loss: {'train/loss': 0.0859500989317894, 'train/grad_norm_pre_clip': 13.070870399475098, 'timing/data': 0.00035749998642131686, 'timing/model': 0.39509661999181844, 'learning_rate/qwen_vl_interface': 8.776539598517119e-06, 'learning_rate/action_model': 8.776539598517118e-05, 'epoch': 0.09}) +[2026-10-03 19:10:26,396][starVLA.training.train_starvla][INFO] - Step 1246, Loss: {'train/loss': 0.014552813023328781, 'train/grad_norm_pre_clip': 6.553140163421631, 'timing/data': 0.00021869002375751734, 'timing/model': 0.39558673999272287, 'learning_rate/qwen_vl_interface': 8.774498671932108e-06, 'learning_rate/action_model': 8.774498671932107e-05, 'epoch': 0.09}) +[2026-10-03 19:10:27,102][starVLA.training.train_starvla][INFO] - Step 1247, Loss: {'train/loss': 0.22578728199005127, 'train/grad_norm_pre_clip': 177.69215393066406, 'timing/data': 0.0003413399972487241, 'timing/model': 0.3933927899925038, 'learning_rate/qwen_vl_interface': 8.772456296558824e-06, 'learning_rate/action_model': 8.772456296558824e-05, 'epoch': 0.09}) +[2026-10-03 19:10:27,808][starVLA.training.train_starvla][INFO] - Step 1248, Loss: {'train/loss': 0.13847145438194275, 'train/grad_norm_pre_clip': 39.8918571472168, 'timing/data': 0.00034466999932192266, 'timing/model': 0.3949936699937098, 'learning_rate/qwen_vl_interface': 8.77041247323681e-06, 'learning_rate/action_model': 8.770412473236811e-05, 'epoch': 0.09}) +[2026-10-03 19:10:28,711][starVLA.training.train_starvla][INFO] - Step 1249, Loss: {'train/loss': 0.16699707508087158, 'train/grad_norm_pre_clip': 99.6413803100586, 'timing/data': 0.00037473998963832855, 'timing/model': 0.3959668099996634, 'learning_rate/qwen_vl_interface': 8.76836720280621e-06, 'learning_rate/action_model': 8.768367202806209e-05, 'epoch': 0.09}) +[2026-10-03 19:10:29,417][starVLA.training.train_starvla][INFO] - Step 1250, Loss: {'train/loss': 0.11376708000898361, 'train/grad_norm_pre_clip': 41.33256149291992, 'timing/data': 0.0003664699906948954, 'timing/model': 0.3933474399964325, 'learning_rate/qwen_vl_interface': 8.766320486107749e-06, 'learning_rate/action_model': 8.766320486107749e-05, 'epoch': 0.09}) +[2026-10-03 19:10:30,124][starVLA.training.train_starvla][INFO] - Step 1251, Loss: {'train/loss': 0.10379774868488312, 'train/grad_norm_pre_clip': 30.794675827026367, 'timing/data': 0.00039080000715330243, 'timing/model': 0.3933309400163125, 'learning_rate/qwen_vl_interface': 8.76427232398276e-06, 'learning_rate/action_model': 8.76427232398276e-05, 'epoch': 0.09}) +[2026-10-03 19:10:30,830][starVLA.training.train_starvla][INFO] - Step 1252, Loss: {'train/loss': 0.033323273062705994, 'train/grad_norm_pre_clip': 25.06438636779785, 'timing/data': 0.00022072001593187451, 'timing/model': 0.3950120499939658, 'learning_rate/qwen_vl_interface': 8.762222717273167e-06, 'learning_rate/action_model': 8.762222717273167e-05, 'epoch': 0.09}) +[2026-10-03 19:10:31,538][starVLA.training.train_starvla][INFO] - Step 1253, Loss: {'train/loss': 0.013878394849598408, 'train/grad_norm_pre_clip': 3.4966578483581543, 'timing/data': 0.00034254000638611615, 'timing/model': 0.3951476710208226, 'learning_rate/qwen_vl_interface': 8.76017166682148e-06, 'learning_rate/action_model': 8.760171666821481e-05, 'epoch': 0.09}) +[2026-10-03 19:10:32,443][starVLA.training.train_starvla][INFO] - Step 1254, Loss: {'train/loss': 0.48179560899734497, 'train/grad_norm_pre_clip': 114.01544189453125, 'timing/data': 0.00036639999598264694, 'timing/model': 0.39486284099984914, 'learning_rate/qwen_vl_interface': 8.758119173470815e-06, 'learning_rate/action_model': 8.758119173470815e-05, 'epoch': 0.09}) +[2026-10-03 19:10:33,150][starVLA.training.train_starvla][INFO] - Step 1255, Loss: {'train/loss': 0.045698944479227066, 'train/grad_norm_pre_clip': 15.4952392578125, 'timing/data': 0.00035295000998303294, 'timing/model': 0.3952607009850908, 'learning_rate/qwen_vl_interface': 8.75606523806487e-06, 'learning_rate/action_model': 8.756065238064868e-05, 'epoch': 0.09}) +[2026-10-03 19:10:33,857][starVLA.training.train_starvla][INFO] - Step 1256, Loss: {'train/loss': 0.0071523599326610565, 'train/grad_norm_pre_clip': 2.4678945541381836, 'timing/data': 0.00034918999881483614, 'timing/model': 0.39519645998370834, 'learning_rate/qwen_vl_interface': 8.75400986144794e-06, 'learning_rate/action_model': 8.75400986144794e-05, 'epoch': 0.09}) +[2026-10-03 19:10:34,565][starVLA.training.train_starvla][INFO] - Step 1257, Loss: {'train/loss': 0.11078602820634842, 'train/grad_norm_pre_clip': 31.757219314575195, 'timing/data': 0.0003459599975030869, 'timing/model': 0.3945285700028762, 'learning_rate/qwen_vl_interface': 8.751953044464913e-06, 'learning_rate/action_model': 8.751953044464913e-05, 'epoch': 0.09}) +[2026-10-03 19:10:35,272][starVLA.training.train_starvla][INFO] - Step 1258, Loss: {'train/loss': 0.03107457235455513, 'train/grad_norm_pre_clip': 14.812227249145508, 'timing/data': 0.0003960600006394088, 'timing/model': 0.39466245999210514, 'learning_rate/qwen_vl_interface': 8.749894787961272e-06, 'learning_rate/action_model': 8.749894787961271e-05, 'epoch': 0.09}) +[2026-10-03 19:10:36,173][starVLA.training.train_starvla][INFO] - Step 1259, Loss: {'train/loss': 0.014750897884368896, 'train/grad_norm_pre_clip': 14.798388481140137, 'timing/data': 0.00034035000135190785, 'timing/model': 0.5898113989969715, 'learning_rate/qwen_vl_interface': 8.747835092783086e-06, 'learning_rate/action_model': 8.747835092783085e-05, 'epoch': 0.09}) +[2026-10-03 19:10:36,880][starVLA.training.train_starvla][INFO] - Step 1260, Loss: {'train/loss': 0.7198728322982788, 'train/grad_norm_pre_clip': 53.68868637084961, 'timing/data': 0.00042057997779920697, 'timing/model': 0.39471976898494177, 'learning_rate/qwen_vl_interface': 8.745773959777018e-06, 'learning_rate/action_model': 8.745773959777017e-05, 'epoch': 0.09}) +[2026-10-03 19:10:37,588][starVLA.training.train_starvla][INFO] - Step 1261, Loss: {'train/loss': 0.377179354429245, 'train/grad_norm_pre_clip': 27.143810272216797, 'timing/data': 0.0002272099955007434, 'timing/model': 0.3957072600023821, 'learning_rate/qwen_vl_interface': 8.743711389790321e-06, 'learning_rate/action_model': 8.74371138979032e-05, 'epoch': 0.09}) +[2026-10-03 19:10:38,296][starVLA.training.train_starvla][INFO] - Step 1262, Loss: {'train/loss': 0.39667028188705444, 'train/grad_norm_pre_clip': 66.44644927978516, 'timing/data': 0.00023472000611945987, 'timing/model': 0.39461026902426966, 'learning_rate/qwen_vl_interface': 8.741647383670839e-06, 'learning_rate/action_model': 8.741647383670839e-05, 'epoch': 0.09}) +[2026-10-03 19:10:39,005][starVLA.training.train_starvla][INFO] - Step 1263, Loss: {'train/loss': 0.4245740473270416, 'train/grad_norm_pre_clip': 24.849275588989258, 'timing/data': 0.00030158000299707055, 'timing/model': 0.39607415901264176, 'learning_rate/qwen_vl_interface': 8.739581942267011e-06, 'learning_rate/action_model': 8.739581942267011e-05, 'epoch': 0.09}) +[2026-10-03 19:10:39,924][starVLA.training.train_starvla][INFO] - Step 1264, Loss: {'train/loss': 0.5094281435012817, 'train/grad_norm_pre_clip': 128.81298828125, 'timing/data': 0.00029154997901059687, 'timing/model': 0.6067434480064549, 'learning_rate/qwen_vl_interface': 8.737515066427861e-06, 'learning_rate/action_model': 8.73751506642786e-05, 'epoch': 0.09}) +[2026-10-03 19:10:40,632][starVLA.training.train_starvla][INFO] - Step 1265, Loss: {'train/loss': 0.4477195143699646, 'train/grad_norm_pre_clip': 137.7659912109375, 'timing/data': 0.00022390999947674572, 'timing/model': 0.395598819013685, 'learning_rate/qwen_vl_interface': 8.735446757003e-06, 'learning_rate/action_model': 8.735446757003e-05, 'epoch': 0.09}) +[2026-10-03 19:10:41,336][starVLA.training.train_starvla][INFO] - Step 1266, Loss: {'train/loss': 0.007381942123174667, 'train/grad_norm_pre_clip': 3.135561466217041, 'timing/data': 0.0002277999883517623, 'timing/model': 0.3923904299736023, 'learning_rate/qwen_vl_interface': 8.733377014842635e-06, 'learning_rate/action_model': 8.733377014842635e-05, 'epoch': 0.09}) +[2026-10-03 19:10:42,046][starVLA.training.train_starvla][INFO] - Step 1267, Loss: {'train/loss': 0.2992156445980072, 'train/grad_norm_pre_clip': 31.700681686401367, 'timing/data': 0.000311030016746372, 'timing/model': 0.39627750002546236, 'learning_rate/qwen_vl_interface': 8.731305840797557e-06, 'learning_rate/action_model': 8.731305840797558e-05, 'epoch': 0.09}) +[2026-10-03 19:10:42,761][starVLA.training.train_starvla][INFO] - Step 1268, Loss: {'train/loss': 0.016999300569295883, 'train/grad_norm_pre_clip': 4.45558500289917, 'timing/data': 0.00022055997396819293, 'timing/model': 0.40413221801281907, 'learning_rate/qwen_vl_interface': 8.72923323571915e-06, 'learning_rate/action_model': 8.72923323571915e-05, 'epoch': 0.09}) +[2026-10-03 19:10:43,470][starVLA.training.train_starvla][INFO] - Step 1269, Loss: {'train/loss': 0.021504506468772888, 'train/grad_norm_pre_clip': 13.357156753540039, 'timing/data': 0.0002652300172485411, 'timing/model': 0.3963062999828253, 'learning_rate/qwen_vl_interface': 8.727159200459381e-06, 'learning_rate/action_model': 8.727159200459381e-05, 'epoch': 0.09}) +[2026-10-03 19:10:44,373][starVLA.training.train_starvla][INFO] - Step 1270, Loss: {'train/loss': 0.10185556858778, 'train/grad_norm_pre_clip': 62.38141632080078, 'timing/data': 0.0004367700021248311, 'timing/model': 0.3927604590135161, 'learning_rate/qwen_vl_interface': 8.725083735870807e-06, 'learning_rate/action_model': 8.725083735870807e-05, 'epoch': 0.09}) +[2026-10-03 19:10:45,082][starVLA.training.train_starvla][INFO] - Step 1271, Loss: {'train/loss': 0.0013671987690031528, 'train/grad_norm_pre_clip': 0.2871154546737671, 'timing/data': 0.00027183000929653645, 'timing/model': 0.3942193499824498, 'learning_rate/qwen_vl_interface': 8.723006842806575e-06, 'learning_rate/action_model': 8.723006842806576e-05, 'epoch': 0.09}) +[2026-10-03 19:10:45,794][starVLA.training.train_starvla][INFO] - Step 1272, Loss: {'train/loss': 0.2701665461063385, 'train/grad_norm_pre_clip': 58.479766845703125, 'timing/data': 0.0003664800024125725, 'timing/model': 0.3989811089995783, 'learning_rate/qwen_vl_interface': 8.720928522120419e-06, 'learning_rate/action_model': 8.720928522120418e-05, 'epoch': 0.09}) +[2026-10-03 19:10:46,503][starVLA.training.train_starvla][INFO] - Step 1273, Loss: {'train/loss': 0.05641606077551842, 'train/grad_norm_pre_clip': 38.08818054199219, 'timing/data': 0.0002541799913160503, 'timing/model': 0.3952709900040645, 'learning_rate/qwen_vl_interface': 8.718848774666655e-06, 'learning_rate/action_model': 8.718848774666655e-05, 'epoch': 0.09}) +[2026-10-03 19:10:47,208][starVLA.training.train_starvla][INFO] - Step 1274, Loss: {'train/loss': 0.03620322048664093, 'train/grad_norm_pre_clip': 29.642452239990234, 'timing/data': 0.00028757998370565474, 'timing/model': 0.3953557799977716, 'learning_rate/qwen_vl_interface': 8.716767601300189e-06, 'learning_rate/action_model': 8.716767601300188e-05, 'epoch': 0.09}) +[2026-10-03 19:10:48,115][starVLA.training.train_starvla][INFO] - Step 1275, Loss: {'train/loss': 0.0637814924120903, 'train/grad_norm_pre_clip': 83.79849243164062, 'timing/data': 0.00025799000286497176, 'timing/model': 0.594360820017755, 'learning_rate/qwen_vl_interface': 8.714685002876514e-06, 'learning_rate/action_model': 8.714685002876513e-05, 'epoch': 0.09}) +[2026-10-03 19:10:48,833][starVLA.training.train_starvla][INFO] - Step 1276, Loss: {'train/loss': 0.06633790582418442, 'train/grad_norm_pre_clip': 32.07099533081055, 'timing/data': 0.0002871600154321641, 'timing/model': 0.4047403190052137, 'learning_rate/qwen_vl_interface': 8.712600980251706e-06, 'learning_rate/action_model': 8.712600980251705e-05, 'epoch': 0.09}) +[2026-10-03 19:10:49,540][starVLA.training.train_starvla][INFO] - Step 1277, Loss: {'train/loss': 0.09239310026168823, 'train/grad_norm_pre_clip': 148.16712951660156, 'timing/data': 0.00022352999076247215, 'timing/model': 0.3954841600207146, 'learning_rate/qwen_vl_interface': 8.710515534282428e-06, 'learning_rate/action_model': 8.710515534282428e-05, 'epoch': 0.09}) +[2026-10-03 19:10:50,246][starVLA.training.train_starvla][INFO] - Step 1278, Loss: {'train/loss': 0.0008537949179299176, 'train/grad_norm_pre_clip': 0.1585664004087448, 'timing/data': 0.0002251299738418311, 'timing/model': 0.39503446899470873, 'learning_rate/qwen_vl_interface': 8.708428665825931e-06, 'learning_rate/action_model': 8.708428665825931e-05, 'epoch': 0.09}) +[2026-10-03 19:10:50,956][starVLA.training.train_starvla][INFO] - Step 1279, Loss: {'train/loss': 0.0005812199087813497, 'train/grad_norm_pre_clip': 0.13948403298854828, 'timing/data': 0.000266039976850152, 'timing/model': 0.3962992390152067, 'learning_rate/qwen_vl_interface': 8.706340375740045e-06, 'learning_rate/action_model': 8.706340375740045e-05, 'epoch': 0.09}) +[2026-10-03 19:10:51,870][starVLA.training.train_starvla][INFO] - Step 1280, Loss: {'train/loss': 0.03731982409954071, 'train/grad_norm_pre_clip': 12.321989059448242, 'timing/data': 0.00022362000891007483, 'timing/model': 0.6038766580168158, 'learning_rate/qwen_vl_interface': 8.704250664883188e-06, 'learning_rate/action_model': 8.704250664883188e-05, 'epoch': 0.09}) +[2026-10-03 19:10:52,579][starVLA.training.train_starvla][INFO] - Step 1281, Loss: {'train/loss': 0.997315526008606, 'train/grad_norm_pre_clip': 72.30712890625, 'timing/data': 0.000352129980456084, 'timing/model': 0.3951897899969481, 'learning_rate/qwen_vl_interface': 8.70215953411436e-06, 'learning_rate/action_model': 8.70215953411436e-05, 'epoch': 0.09}) +[2026-10-03 19:10:53,284][starVLA.training.train_starvla][INFO] - Step 1282, Loss: {'train/loss': 0.25025367736816406, 'train/grad_norm_pre_clip': 25.869413375854492, 'timing/data': 0.0002503099967725575, 'timing/model': 0.39539154001977295, 'learning_rate/qwen_vl_interface': 8.70006698429315e-06, 'learning_rate/action_model': 8.700066984293149e-05, 'epoch': 0.09}) +[2026-10-03 19:10:53,993][starVLA.training.train_starvla][INFO] - Step 1283, Loss: {'train/loss': 0.0013569581788033247, 'train/grad_norm_pre_clip': 0.9225509762763977, 'timing/data': 0.0002837100182659924, 'timing/model': 0.39473676899797283, 'learning_rate/qwen_vl_interface': 8.697973016279722e-06, 'learning_rate/action_model': 8.697973016279722e-05, 'epoch': 0.09}) +[2026-10-03 19:10:54,708][starVLA.training.train_starvla][INFO] - Step 1284, Loss: {'train/loss': 0.03432885929942131, 'train/grad_norm_pre_clip': 42.335289001464844, 'timing/data': 0.00022117001935839653, 'timing/model': 0.4027178389951587, 'learning_rate/qwen_vl_interface': 8.695877630934827e-06, 'learning_rate/action_model': 8.695877630934826e-05, 'epoch': 0.09}) +[2026-10-03 19:10:55,415][starVLA.training.train_starvla][INFO] - Step 1285, Loss: {'train/loss': 0.011720821261405945, 'train/grad_norm_pre_clip': 8.811084747314453, 'timing/data': 0.00022784998873248696, 'timing/model': 0.39561916000093333, 'learning_rate/qwen_vl_interface': 8.693780829119801e-06, 'learning_rate/action_model': 8.693780829119801e-05, 'epoch': 0.09}) +[2026-10-03 19:10:56,314][starVLA.training.train_starvla][INFO] - Step 1286, Loss: {'train/loss': 0.33228054642677307, 'train/grad_norm_pre_clip': 100.17855072021484, 'timing/data': 0.00038435999886132777, 'timing/model': 0.3928955100127496, 'learning_rate/qwen_vl_interface': 8.69168261169656e-06, 'learning_rate/action_model': 8.691682611696559e-05, 'epoch': 0.09}) +[2026-10-03 19:10:57,026][starVLA.training.train_starvla][INFO] - Step 1287, Loss: {'train/loss': 0.2727791368961334, 'train/grad_norm_pre_clip': 108.3025894165039, 'timing/data': 0.0003980400215368718, 'timing/model': 0.39746557900798507, 'learning_rate/qwen_vl_interface': 8.689582979527599e-06, 'learning_rate/action_model': 8.689582979527599e-05, 'epoch': 0.09}) +[2026-10-03 19:10:57,708][starVLA.training.train_starvla][INFO] - Step 1288, Loss: {'train/loss': 0.05999327078461647, 'train/grad_norm_pre_clip': 107.05209350585938, 'timing/data': 0.000221690017497167, 'timing/model': 0.3696729619987309, 'learning_rate/qwen_vl_interface': 8.687481933476e-06, 'learning_rate/action_model': 8.687481933476e-05, 'epoch': 0.09}) +[2026-10-03 19:10:58,423][starVLA.training.train_starvla][INFO] - Step 1289, Loss: {'train/loss': 0.361472487449646, 'train/grad_norm_pre_clip': 64.88847351074219, 'timing/data': 0.00025817001005634665, 'timing/model': 0.39490044899866916, 'learning_rate/qwen_vl_interface': 8.685379474405423e-06, 'learning_rate/action_model': 8.685379474405423e-05, 'epoch': 0.09}) +[2026-10-03 19:10:59,127][starVLA.training.train_starvla][INFO] - Step 1290, Loss: {'train/loss': 0.0014099334366619587, 'train/grad_norm_pre_clip': 0.4375740885734558, 'timing/data': 0.00037066900404170156, 'timing/model': 0.39317341000423767, 'learning_rate/qwen_vl_interface': 8.68327560318011e-06, 'learning_rate/action_model': 8.68327560318011e-05, 'epoch': 0.09}) +[2026-10-03 19:11:00,033][starVLA.training.train_starvla][INFO] - Step 1291, Loss: {'train/loss': 0.0010050369892269373, 'train/grad_norm_pre_clip': 0.21768593788146973, 'timing/data': 0.0003053199907299131, 'timing/model': 0.3955485000042245, 'learning_rate/qwen_vl_interface': 8.681170320664883e-06, 'learning_rate/action_model': 8.681170320664883e-05, 'epoch': 0.09}) +[2026-10-03 19:11:00,741][starVLA.training.train_starvla][INFO] - Step 1292, Loss: {'train/loss': 0.20730899274349213, 'train/grad_norm_pre_clip': 41.48189926147461, 'timing/data': 0.00028491998091340065, 'timing/model': 0.39581056000315584, 'learning_rate/qwen_vl_interface': 8.679063627725143e-06, 'learning_rate/action_model': 8.679063627725144e-05, 'epoch': 0.09}) +[2026-10-03 19:11:01,436][starVLA.training.train_starvla][INFO] - Step 1293, Loss: {'train/loss': 0.08749063313007355, 'train/grad_norm_pre_clip': 29.711214065551758, 'timing/data': 0.0002727799874264747, 'timing/model': 0.39507189000141807, 'learning_rate/qwen_vl_interface': 8.676955525226876e-06, 'learning_rate/action_model': 8.676955525226875e-05, 'epoch': 0.09}) +[2026-10-03 19:11:02,142][starVLA.training.train_starvla][INFO] - Step 1294, Loss: {'train/loss': 0.025434238836169243, 'train/grad_norm_pre_clip': 28.179269790649414, 'timing/data': 0.00027776998467743397, 'timing/model': 0.39330601002438925, 'learning_rate/qwen_vl_interface': 8.674846014036636e-06, 'learning_rate/action_model': 8.674846014036636e-05, 'epoch': 0.09}) +[2026-10-03 19:11:02,849][starVLA.training.train_starvla][INFO] - Step 1295, Loss: {'train/loss': 0.020493054762482643, 'train/grad_norm_pre_clip': 15.832356452941895, 'timing/data': 0.0003153400029987097, 'timing/model': 0.3960436299967114, 'learning_rate/qwen_vl_interface': 8.672735095021569e-06, 'learning_rate/action_model': 8.672735095021569e-05, 'epoch': 0.09}) +[2026-10-03 19:11:03,753][starVLA.training.train_starvla][INFO] - Step 1296, Loss: {'train/loss': 0.0060831839218735695, 'train/grad_norm_pre_clip': 7.4625372886657715, 'timing/data': 0.00027171999681741, 'timing/model': 0.5932826590142213, 'learning_rate/qwen_vl_interface': 8.670622769049393e-06, 'learning_rate/action_model': 8.670622769049392e-05, 'epoch': 0.09}) +[2026-10-03 19:11:04,469][starVLA.training.train_starvla][INFO] - Step 1297, Loss: {'train/loss': 0.541566014289856, 'train/grad_norm_pre_clip': 40.376136779785156, 'timing/data': 0.0003588699910324067, 'timing/model': 0.3949690500157885, 'learning_rate/qwen_vl_interface': 8.668509036988406e-06, 'learning_rate/action_model': 8.668509036988405e-05, 'epoch': 0.09}) +[2026-10-03 19:11:05,175][starVLA.training.train_starvla][INFO] - Step 1298, Loss: {'train/loss': 1.098812222480774, 'train/grad_norm_pre_clip': 223.74874877929688, 'timing/data': 0.00037089001853019, 'timing/model': 0.3938893300073687, 'learning_rate/qwen_vl_interface': 8.66639389970748e-06, 'learning_rate/action_model': 8.666393899707479e-05, 'epoch': 0.1}) +[2026-10-03 19:11:05,882][starVLA.training.train_starvla][INFO] - Step 1299, Loss: {'train/loss': 0.1333465427160263, 'train/grad_norm_pre_clip': 68.94593811035156, 'timing/data': 0.00025584999821148813, 'timing/model': 0.39381097999284975, 'learning_rate/qwen_vl_interface': 8.66427735807607e-06, 'learning_rate/action_model': 8.664277358076069e-05, 'epoch': 0.1}) +[2026-10-03 19:11:06,589][starVLA.training.train_starvla][INFO] - Step 1300, Loss: {'train/loss': 0.4315159320831299, 'train/grad_norm_pre_clip': 31.581716537475586, 'timing/data': 0.0002938399848062545, 'timing/model': 0.39512368000578135, 'learning_rate/qwen_vl_interface': 8.662159412964208e-06, 'learning_rate/action_model': 8.662159412964207e-05, 'epoch': 0.1}) +[2026-10-03 19:11:07,505][starVLA.training.train_starvla][INFO] - Step 1301, Loss: {'train/loss': 0.0010324589675292373, 'train/grad_norm_pre_clip': 0.28671005368232727, 'timing/data': 0.00033457999234087765, 'timing/model': 0.5921710290131159, 'learning_rate/qwen_vl_interface': 8.660040065242499e-06, 'learning_rate/action_model': 8.660040065242498e-05, 'epoch': 0.1}) +[2026-10-03 19:11:08,210][starVLA.training.train_starvla][INFO] - Step 1302, Loss: {'train/loss': 0.5646741986274719, 'train/grad_norm_pre_clip': 52.745513916015625, 'timing/data': 0.000254979997407645, 'timing/model': 0.3943162990035489, 'learning_rate/qwen_vl_interface': 8.657919315782127e-06, 'learning_rate/action_model': 8.657919315782127e-05, 'epoch': 0.1}) +[2026-10-03 19:11:08,917][starVLA.training.train_starvla][INFO] - Step 1303, Loss: {'train/loss': 0.0013211587211117148, 'train/grad_norm_pre_clip': 0.5082918405532837, 'timing/data': 0.00035803898936137557, 'timing/model': 0.3951059099927079, 'learning_rate/qwen_vl_interface': 8.655797165454853e-06, 'learning_rate/action_model': 8.655797165454853e-05, 'epoch': 0.1}) +[2026-10-03 19:11:09,625][starVLA.training.train_starvla][INFO] - Step 1304, Loss: {'train/loss': 0.009988109581172466, 'train/grad_norm_pre_clip': 2.1098036766052246, 'timing/data': 0.00031150999711826444, 'timing/model': 0.39554542998666875, 'learning_rate/qwen_vl_interface': 8.653673615133013e-06, 'learning_rate/action_model': 8.653673615133013e-05, 'epoch': 0.1}) +[2026-10-03 19:11:10,344][starVLA.training.train_starvla][INFO] - Step 1305, Loss: {'train/loss': 0.17632772028446198, 'train/grad_norm_pre_clip': 69.70565032958984, 'timing/data': 0.0003723700065165758, 'timing/model': 0.3951798100024462, 'learning_rate/qwen_vl_interface': 8.651548665689519e-06, 'learning_rate/action_model': 8.651548665689518e-05, 'epoch': 0.1}) +[2026-10-03 19:11:11,050][starVLA.training.train_starvla][INFO] - Step 1306, Loss: {'train/loss': 0.23717987537384033, 'train/grad_norm_pre_clip': 112.98553466796875, 'timing/data': 0.0002239599998574704, 'timing/model': 0.39408018000540324, 'learning_rate/qwen_vl_interface': 8.649422317997855e-06, 'learning_rate/action_model': 8.649422317997856e-05, 'epoch': 0.1}) +[2026-10-03 19:11:11,952][starVLA.training.train_starvla][INFO] - Step 1307, Loss: {'train/loss': 0.4612323045730591, 'train/grad_norm_pre_clip': 57.01483154296875, 'timing/data': 0.0002928500180132687, 'timing/model': 0.39387746900320053, 'learning_rate/qwen_vl_interface': 8.647294572932082e-06, 'learning_rate/action_model': 8.647294572932082e-05, 'epoch': 0.1}) +[2026-10-03 19:11:12,657][starVLA.training.train_starvla][INFO] - Step 1308, Loss: {'train/loss': 0.0022544548846781254, 'train/grad_norm_pre_clip': 0.7669726610183716, 'timing/data': 0.0003100800095126033, 'timing/model': 0.3943966290098615, 'learning_rate/qwen_vl_interface': 8.64516543136684e-06, 'learning_rate/action_model': 8.64516543136684e-05, 'epoch': 0.1}) +[2026-10-03 19:11:13,377][starVLA.training.train_starvla][INFO] - Step 1309, Loss: {'train/loss': 0.00629454106092453, 'train/grad_norm_pre_clip': 5.047893047332764, 'timing/data': 0.00032587998430244625, 'timing/model': 0.39441415001056157, 'learning_rate/qwen_vl_interface': 8.643034894177335e-06, 'learning_rate/action_model': 8.643034894177335e-05, 'epoch': 0.1}) +[2026-10-03 19:11:14,082][starVLA.training.train_starvla][INFO] - Step 1310, Loss: {'train/loss': 0.007450764067471027, 'train/grad_norm_pre_clip': 4.705685615539551, 'timing/data': 0.0003307399747427553, 'timing/model': 0.3935610599874053, 'learning_rate/qwen_vl_interface': 8.64090296223935e-06, 'learning_rate/action_model': 8.64090296223935e-05, 'epoch': 0.1}) +[2026-10-03 19:11:14,791][starVLA.training.train_starvla][INFO] - Step 1311, Loss: {'train/loss': 0.0013694192748516798, 'train/grad_norm_pre_clip': 0.9220221638679504, 'timing/data': 0.00029931002063676715, 'timing/model': 0.3940243989927694, 'learning_rate/qwen_vl_interface': 8.638769636429245e-06, 'learning_rate/action_model': 8.638769636429245e-05, 'epoch': 0.1}) +[2026-10-03 19:11:15,687][starVLA.training.train_starvla][INFO] - Step 1312, Loss: {'train/loss': 0.000691050139721483, 'train/grad_norm_pre_clip': 0.3809947967529297, 'timing/data': 0.00031408999348059297, 'timing/model': 0.5847858710039873, 'learning_rate/qwen_vl_interface': 8.636634917623946e-06, 'learning_rate/action_model': 8.636634917623944e-05, 'epoch': 0.1}) +[2026-10-03 19:11:16,404][starVLA.training.train_starvla][INFO] - Step 1313, Loss: {'train/loss': 0.48029646277427673, 'train/grad_norm_pre_clip': 32.93857955932617, 'timing/data': 0.00024628001847304404, 'timing/model': 0.39446902001509443, 'learning_rate/qwen_vl_interface': 8.634498806700955e-06, 'learning_rate/action_model': 8.634498806700954e-05, 'epoch': 0.1}) +[2026-10-03 19:11:17,109][starVLA.training.train_starvla][INFO] - Step 1314, Loss: {'train/loss': 0.1891772598028183, 'train/grad_norm_pre_clip': 73.13972473144531, 'timing/data': 0.0004157799994572997, 'timing/model': 0.3942741799983196, 'learning_rate/qwen_vl_interface': 8.63236130453835e-06, 'learning_rate/action_model': 8.63236130453835e-05, 'epoch': 0.1}) +[2026-10-03 19:11:17,817][starVLA.training.train_starvla][INFO] - Step 1315, Loss: {'train/loss': 0.014627219177782536, 'train/grad_norm_pre_clip': 17.36280632019043, 'timing/data': 0.00030946999322623014, 'timing/model': 0.3975314889976289, 'learning_rate/qwen_vl_interface': 8.630222412014777e-06, 'learning_rate/action_model': 8.630222412014776e-05, 'epoch': 0.1}) +[2026-10-03 19:11:18,525][starVLA.training.train_starvla][INFO] - Step 1316, Loss: {'train/loss': 0.8509948253631592, 'train/grad_norm_pre_clip': 189.4587860107422, 'timing/data': 0.0003430299984756857, 'timing/model': 0.3955087100039236, 'learning_rate/qwen_vl_interface': 8.628082130009451e-06, 'learning_rate/action_model': 8.628082130009451e-05, 'epoch': 0.1}) +[2026-10-03 19:11:19,441][starVLA.training.train_starvla][INFO] - Step 1317, Loss: {'train/loss': 0.23616524040699005, 'train/grad_norm_pre_clip': 235.9680938720703, 'timing/data': 0.00022066000383347273, 'timing/model': 0.5919651699950919, 'learning_rate/qwen_vl_interface': 8.62594045940216e-06, 'learning_rate/action_model': 8.62594045940216e-05, 'epoch': 0.1}) +[2026-10-03 19:11:20,147][starVLA.training.train_starvla][INFO] - Step 1318, Loss: {'train/loss': 1.0577270984649658, 'train/grad_norm_pre_clip': 87.6847152709961, 'timing/data': 0.00039508999907411635, 'timing/model': 0.39440469999681227, 'learning_rate/qwen_vl_interface': 8.623797401073272e-06, 'learning_rate/action_model': 8.623797401073271e-05, 'epoch': 0.1}) +[2026-10-03 19:11:20,855][starVLA.training.train_starvla][INFO] - Step 1319, Loss: {'train/loss': 1.1294852495193481, 'train/grad_norm_pre_clip': 123.45068359375, 'timing/data': 0.00028083001961931586, 'timing/model': 0.3946094299899414, 'learning_rate/qwen_vl_interface': 8.62165295590371e-06, 'learning_rate/action_model': 8.621652955903709e-05, 'epoch': 0.1}) +[2026-10-03 19:11:21,576][starVLA.training.train_starvla][INFO] - Step 1320, Loss: {'train/loss': 0.00037936665466986597, 'train/grad_norm_pre_clip': 0.15473616123199463, 'timing/data': 0.00035082001704722643, 'timing/model': 0.4073616990062874, 'learning_rate/qwen_vl_interface': 8.619507124774975e-06, 'learning_rate/action_model': 8.619507124774975e-05, 'epoch': 0.1}) +[2026-10-03 19:11:22,284][starVLA.training.train_starvla][INFO] - Step 1321, Loss: {'train/loss': 1.082936406135559, 'train/grad_norm_pre_clip': 110.45709228515625, 'timing/data': 0.00026403999072499573, 'timing/model': 0.39477008001995273, 'learning_rate/qwen_vl_interface': 8.61735990856914e-06, 'learning_rate/action_model': 8.617359908569139e-05, 'epoch': 0.1}) +[2026-10-03 19:11:22,989][starVLA.training.train_starvla][INFO] - Step 1322, Loss: {'train/loss': 0.6555776000022888, 'train/grad_norm_pre_clip': 134.9517059326172, 'timing/data': 0.0002478400128893554, 'timing/model': 0.39529716898687184, 'learning_rate/qwen_vl_interface': 8.615211308168842e-06, 'learning_rate/action_model': 8.61521130816884e-05, 'epoch': 0.1}) +[2026-10-03 19:11:23,893][starVLA.training.train_starvla][INFO] - Step 1323, Loss: {'train/loss': 0.11563378572463989, 'train/grad_norm_pre_clip': 335.48052978515625, 'timing/data': 0.0003279999946244061, 'timing/model': 0.39466503998846747, 'learning_rate/qwen_vl_interface': 8.613061324457289e-06, 'learning_rate/action_model': 8.613061324457288e-05, 'epoch': 0.1}) +[2026-10-03 19:11:24,614][starVLA.training.train_starvla][INFO] - Step 1324, Loss: {'train/loss': 0.29289451241493225, 'train/grad_norm_pre_clip': 150.80032348632812, 'timing/data': 0.00028352998197078705, 'timing/model': 0.4083655180002097, 'learning_rate/qwen_vl_interface': 8.610909958318262e-06, 'learning_rate/action_model': 8.610909958318261e-05, 'epoch': 0.1}) +[2026-10-03 19:11:25,323][starVLA.training.train_starvla][INFO] - Step 1325, Loss: {'train/loss': 0.9815530180931091, 'train/grad_norm_pre_clip': 86.68216705322266, 'timing/data': 0.000244670023676008, 'timing/model': 0.3951458400115371, 'learning_rate/qwen_vl_interface': 8.608757210636101e-06, 'learning_rate/action_model': 8.608757210636101e-05, 'epoch': 0.1}) +[2026-10-03 19:11:26,028][starVLA.training.train_starvla][INFO] - Step 1326, Loss: {'train/loss': 0.004875651560723782, 'train/grad_norm_pre_clip': 6.5489349365234375, 'timing/data': 0.00034517000312916934, 'timing/model': 0.3938178390089888, 'learning_rate/qwen_vl_interface': 8.606603082295724e-06, 'learning_rate/action_model': 8.606603082295724e-05, 'epoch': 0.1}) +[2026-10-03 19:11:26,736][starVLA.training.train_starvla][INFO] - Step 1327, Loss: {'train/loss': 0.0067911772057414055, 'train/grad_norm_pre_clip': 2.3355774879455566, 'timing/data': 0.0003148600226268172, 'timing/model': 0.39468307900824584, 'learning_rate/qwen_vl_interface': 8.60444757418261e-06, 'learning_rate/action_model': 8.604447574182609e-05, 'epoch': 0.1}) +[2026-10-03 19:11:27,652][starVLA.training.train_starvla][INFO] - Step 1328, Loss: {'train/loss': 0.0037696659564971924, 'train/grad_norm_pre_clip': 1.6469910144805908, 'timing/data': 0.0003643399977590889, 'timing/model': 0.40636878900113516, 'learning_rate/qwen_vl_interface': 8.602290687182805e-06, 'learning_rate/action_model': 8.602290687182806e-05, 'epoch': 0.1}) +[2026-10-03 19:11:28,361][starVLA.training.train_starvla][INFO] - Step 1329, Loss: {'train/loss': 0.02066858485341072, 'train/grad_norm_pre_clip': 30.679777145385742, 'timing/data': 0.00022652000188827515, 'timing/model': 0.39488809899194166, 'learning_rate/qwen_vl_interface': 8.600132422182929e-06, 'learning_rate/action_model': 8.600132422182928e-05, 'epoch': 0.1}) +[2026-10-03 19:11:29,067][starVLA.training.train_starvla][INFO] - Step 1330, Loss: {'train/loss': 0.0030507268384099007, 'train/grad_norm_pre_clip': 0.6049741506576538, 'timing/data': 0.0002931689959950745, 'timing/model': 0.3936785999976564, 'learning_rate/qwen_vl_interface': 8.597972780070156e-06, 'learning_rate/action_model': 8.597972780070155e-05, 'epoch': 0.1}) +[2026-10-03 19:11:29,778][starVLA.training.train_starvla][INFO] - Step 1331, Loss: {'train/loss': 0.25267529487609863, 'train/grad_norm_pre_clip': 84.79365539550781, 'timing/data': 0.00039919998380355537, 'timing/model': 0.3970321599917952, 'learning_rate/qwen_vl_interface': 8.595811761732239e-06, 'learning_rate/action_model': 8.595811761732238e-05, 'epoch': 0.1}) +[2026-10-03 19:11:30,496][starVLA.training.train_starvla][INFO] - Step 1332, Loss: {'train/loss': 0.06378236413002014, 'train/grad_norm_pre_clip': 72.36127471923828, 'timing/data': 0.0003059899900108576, 'timing/model': 0.4075313379871659, 'learning_rate/qwen_vl_interface': 8.593649368057489e-06, 'learning_rate/action_model': 8.59364936805749e-05, 'epoch': 0.1}) +[2026-10-03 19:11:31,407][starVLA.training.train_starvla][INFO] - Step 1333, Loss: {'train/loss': 0.3532010316848755, 'train/grad_norm_pre_clip': 39.20970153808594, 'timing/data': 0.00030413002241402864, 'timing/model': 0.5962515189894475, 'learning_rate/qwen_vl_interface': 8.591485599934786e-06, 'learning_rate/action_model': 8.591485599934785e-05, 'epoch': 0.1}) +[2026-10-03 19:11:32,113][starVLA.training.train_starvla][INFO] - Step 1334, Loss: {'train/loss': 0.001276311930269003, 'train/grad_norm_pre_clip': 0.4212628901004791, 'timing/data': 0.00036813001497648656, 'timing/model': 0.39417055001831613, 'learning_rate/qwen_vl_interface': 8.58932045825357e-06, 'learning_rate/action_model': 8.589320458253569e-05, 'epoch': 0.1}) +[2026-10-03 19:11:32,821][starVLA.training.train_starvla][INFO] - Step 1335, Loss: {'train/loss': 0.4461347162723541, 'train/grad_norm_pre_clip': 10.661569595336914, 'timing/data': 0.00031586000113748014, 'timing/model': 0.395569059997797, 'learning_rate/qwen_vl_interface': 8.58715394390385e-06, 'learning_rate/action_model': 8.58715394390385e-05, 'epoch': 0.1}) +[2026-10-03 19:11:33,542][starVLA.training.train_starvla][INFO] - Step 1336, Loss: {'train/loss': 0.2683447301387787, 'train/grad_norm_pre_clip': 155.6148223876953, 'timing/data': 0.00028730000485666096, 'timing/model': 0.40810969899757765, 'learning_rate/qwen_vl_interface': 8.5849860577762e-06, 'learning_rate/action_model': 8.5849860577762e-05, 'epoch': 0.1}) +[2026-10-03 19:11:34,252][starVLA.training.train_starvla][INFO] - Step 1337, Loss: {'train/loss': 0.22741541266441345, 'train/grad_norm_pre_clip': 171.1488800048828, 'timing/data': 0.0003715100174304098, 'timing/model': 0.3958098489965778, 'learning_rate/qwen_vl_interface': 8.582816800761756e-06, 'learning_rate/action_model': 8.582816800761755e-05, 'epoch': 0.1}) +[2026-10-03 19:11:35,158][starVLA.training.train_starvla][INFO] - Step 1338, Loss: {'train/loss': 0.015771489590406418, 'train/grad_norm_pre_clip': 21.28585433959961, 'timing/data': 0.00035983999259769917, 'timing/model': 0.5936592500074767, 'learning_rate/qwen_vl_interface': 8.580646173752214e-06, 'learning_rate/action_model': 8.580646173752213e-05, 'epoch': 0.1}) +[2026-10-03 19:11:35,866][starVLA.training.train_starvla][INFO] - Step 1339, Loss: {'train/loss': 0.0026588812470436096, 'train/grad_norm_pre_clip': 0.7547945380210876, 'timing/data': 0.0003322300035506487, 'timing/model': 0.39438815900939517, 'learning_rate/qwen_vl_interface': 8.578474177639841e-06, 'learning_rate/action_model': 8.578474177639842e-05, 'epoch': 0.1}) +[2026-10-03 19:11:36,581][starVLA.training.train_starvla][INFO] - Step 1340, Loss: {'train/loss': 0.33714988827705383, 'train/grad_norm_pre_clip': 50.452598571777344, 'timing/data': 0.0003090800018981099, 'timing/model': 0.4014999780047219, 'learning_rate/qwen_vl_interface': 8.576300813317463e-06, 'learning_rate/action_model': 8.576300813317462e-05, 'epoch': 0.1}) +[2026-10-03 19:11:37,289][starVLA.training.train_starvla][INFO] - Step 1341, Loss: {'train/loss': 0.0011094148503616452, 'train/grad_norm_pre_clip': 0.3699067533016205, 'timing/data': 0.0002523900184314698, 'timing/model': 0.39586858998518437, 'learning_rate/qwen_vl_interface': 8.574126081678463e-06, 'learning_rate/action_model': 8.574126081678462e-05, 'epoch': 0.1}) +[2026-10-03 19:11:37,994][starVLA.training.train_starvla][INFO] - Step 1342, Loss: {'train/loss': 0.007340666837990284, 'train/grad_norm_pre_clip': 3.561540126800537, 'timing/data': 0.00021959998412057757, 'timing/model': 0.3953851200058125, 'learning_rate/qwen_vl_interface': 8.571949983616796e-06, 'learning_rate/action_model': 8.571949983616795e-05, 'epoch': 0.1}) +[2026-10-03 19:11:38,703][starVLA.training.train_starvla][INFO] - Step 1343, Loss: {'train/loss': 0.4998801648616791, 'train/grad_norm_pre_clip': 51.256431579589844, 'timing/data': 0.00028184999246150255, 'timing/model': 0.3955070590018295, 'learning_rate/qwen_vl_interface': 8.569772520026971e-06, 'learning_rate/action_model': 8.569772520026971e-05, 'epoch': 0.1}) +[2026-10-03 19:11:39,622][starVLA.training.train_starvla][INFO] - Step 1344, Loss: {'train/loss': 0.4060441255569458, 'train/grad_norm_pre_clip': 28.06621742248535, 'timing/data': 0.00032259998260997236, 'timing/model': 0.40940394898643717, 'learning_rate/qwen_vl_interface': 8.567593691804061e-06, 'learning_rate/action_model': 8.567593691804061e-05, 'epoch': 0.1}) +[2026-10-03 19:11:40,330][starVLA.training.train_starvla][INFO] - Step 1345, Loss: {'train/loss': 0.043878816068172455, 'train/grad_norm_pre_clip': 20.97849464416504, 'timing/data': 0.00035484001273289323, 'timing/model': 0.3939200899912976, 'learning_rate/qwen_vl_interface': 8.565413499843703e-06, 'learning_rate/action_model': 8.565413499843703e-05, 'epoch': 0.1}) +[2026-10-03 19:11:41,034][starVLA.training.train_starvla][INFO] - Step 1346, Loss: {'train/loss': 0.33290916681289673, 'train/grad_norm_pre_clip': 25.038286209106445, 'timing/data': 0.00023339001927524805, 'timing/model': 0.3936554699903354, 'learning_rate/qwen_vl_interface': 8.563231945042087e-06, 'learning_rate/action_model': 8.563231945042088e-05, 'epoch': 0.1}) +[2026-10-03 19:11:41,741][starVLA.training.train_starvla][INFO] - Step 1347, Loss: {'train/loss': 0.6824798583984375, 'train/grad_norm_pre_clip': 142.52066040039062, 'timing/data': 0.0002781899820547551, 'timing/model': 0.3946618099871557, 'learning_rate/qwen_vl_interface': 8.56104902829597e-06, 'learning_rate/action_model': 8.56104902829597e-05, 'epoch': 0.1}) +[2026-10-03 19:11:42,462][starVLA.training.train_starvla][INFO] - Step 1348, Loss: {'train/loss': 0.5920994281768799, 'train/grad_norm_pre_clip': 16.739370346069336, 'timing/data': 0.0003480300074443221, 'timing/model': 0.4086054080107715, 'learning_rate/qwen_vl_interface': 8.558864750502669e-06, 'learning_rate/action_model': 8.558864750502668e-05, 'epoch': 0.1}) +[2026-10-03 19:11:43,370][starVLA.training.train_starvla][INFO] - Step 1349, Loss: {'train/loss': 0.27138832211494446, 'train/grad_norm_pre_clip': 178.75990295410156, 'timing/data': 0.0003398300032131374, 'timing/model': 0.39490753901191056, 'learning_rate/qwen_vl_interface': 8.556679112560055e-06, 'learning_rate/action_model': 8.556679112560054e-05, 'epoch': 0.1}) +[2026-10-03 19:11:44,076][starVLA.training.train_starvla][INFO] - Step 1350, Loss: {'train/loss': 0.0669095367193222, 'train/grad_norm_pre_clip': 57.2277717590332, 'timing/data': 0.0003570899716578424, 'timing/model': 0.39451840898254886, 'learning_rate/qwen_vl_interface': 8.554492115366562e-06, 'learning_rate/action_model': 8.554492115366561e-05, 'epoch': 0.1}) +[2026-10-03 19:11:44,783][starVLA.training.train_starvla][INFO] - Step 1351, Loss: {'train/loss': 0.0010692928917706013, 'train/grad_norm_pre_clip': 0.26392364501953125, 'timing/data': 0.0003631100116763264, 'timing/model': 0.39476008899509907, 'learning_rate/qwen_vl_interface': 8.55230375982118e-06, 'learning_rate/action_model': 8.55230375982118e-05, 'epoch': 0.1}) +[2026-10-03 19:11:45,504][starVLA.training.train_starvla][INFO] - Step 1352, Loss: {'train/loss': 0.003252476453781128, 'train/grad_norm_pre_clip': 1.9652283191680908, 'timing/data': 0.0002711899869609624, 'timing/model': 0.4078784989833366, 'learning_rate/qwen_vl_interface': 8.550114046823463e-06, 'learning_rate/action_model': 8.550114046823463e-05, 'epoch': 0.1}) +[2026-10-03 19:11:46,213][starVLA.training.train_starvla][INFO] - Step 1353, Loss: {'train/loss': 0.43907690048217773, 'train/grad_norm_pre_clip': 105.29084014892578, 'timing/data': 0.000264920003246516, 'timing/model': 0.39560391000122763, 'learning_rate/qwen_vl_interface': 8.547922977273518e-06, 'learning_rate/action_model': 8.547922977273518e-05, 'epoch': 0.1}) +[2026-10-03 19:11:47,113][starVLA.training.train_starvla][INFO] - Step 1354, Loss: {'train/loss': 0.07138685882091522, 'train/grad_norm_pre_clip': 30.886018753051758, 'timing/data': 0.0003891300002578646, 'timing/model': 0.5905057190102525, 'learning_rate/qwen_vl_interface': 8.54573055207201e-06, 'learning_rate/action_model': 8.545730552072008e-05, 'epoch': 0.1}) +[2026-10-03 19:11:47,797][starVLA.training.train_starvla][INFO] - Step 1355, Loss: {'train/loss': 0.1444711983203888, 'train/grad_norm_pre_clip': 39.57167434692383, 'timing/data': 0.00030149001395329833, 'timing/model': 0.3704718920052983, 'learning_rate/qwen_vl_interface': 8.543536772120161e-06, 'learning_rate/action_model': 8.54353677212016e-05, 'epoch': 0.1}) +[2026-10-03 19:11:48,512][starVLA.training.train_starvla][INFO] - Step 1356, Loss: {'train/loss': 0.5419940948486328, 'train/grad_norm_pre_clip': 133.69871520996094, 'timing/data': 0.0002756899921223521, 'timing/model': 0.40455360800842755, 'learning_rate/qwen_vl_interface': 8.541341638319754e-06, 'learning_rate/action_model': 8.541341638319754e-05, 'epoch': 0.1}) +[2026-10-03 19:11:49,220][starVLA.training.train_starvla][INFO] - Step 1357, Loss: {'train/loss': 0.00453867856413126, 'train/grad_norm_pre_clip': 1.0218886137008667, 'timing/data': 0.00034764999873004854, 'timing/model': 0.3957815099856816, 'learning_rate/qwen_vl_interface': 8.539145151573126e-06, 'learning_rate/action_model': 8.539145151573125e-05, 'epoch': 0.1}) +[2026-10-03 19:11:49,926][starVLA.training.train_starvla][INFO] - Step 1358, Loss: {'train/loss': 0.6748642325401306, 'train/grad_norm_pre_clip': 68.65647888183594, 'timing/data': 0.00024068998754955828, 'timing/model': 0.3939514800149482, 'learning_rate/qwen_vl_interface': 8.536947312783167e-06, 'learning_rate/action_model': 8.536947312783166e-05, 'epoch': 0.1}) +[2026-10-03 19:11:50,839][starVLA.training.train_starvla][INFO] - Step 1359, Loss: {'train/loss': 0.4784497022628784, 'train/grad_norm_pre_clip': 25.697025299072266, 'timing/data': 0.00029878999339416623, 'timing/model': 0.5987313590012491, 'learning_rate/qwen_vl_interface': 8.534748122853326e-06, 'learning_rate/action_model': 8.534748122853325e-05, 'epoch': 0.1}) +[2026-10-03 19:11:51,560][starVLA.training.train_starvla][INFO] - Step 1360, Loss: {'train/loss': 0.13273459672927856, 'train/grad_norm_pre_clip': 27.211809158325195, 'timing/data': 0.00028142001247033477, 'timing/model': 0.40928994902060367, 'learning_rate/qwen_vl_interface': 8.532547582687609e-06, 'learning_rate/action_model': 8.532547582687608e-05, 'epoch': 0.1}) +[2026-10-03 19:11:52,281][starVLA.training.train_starvla][INFO] - Step 1361, Loss: {'train/loss': 0.02215558849275112, 'train/grad_norm_pre_clip': 8.414055824279785, 'timing/data': 0.00024240001221187413, 'timing/model': 0.3963076090149116, 'learning_rate/qwen_vl_interface': 8.530345693190573e-06, 'learning_rate/action_model': 8.530345693190572e-05, 'epoch': 0.1}) +[2026-10-03 19:11:52,983][starVLA.training.train_starvla][INFO] - Step 1362, Loss: {'train/loss': 0.009587274864315987, 'train/grad_norm_pre_clip': 3.948476791381836, 'timing/data': 0.00023965002037584782, 'timing/model': 0.3928193899919279, 'learning_rate/qwen_vl_interface': 8.528142455267328e-06, 'learning_rate/action_model': 8.528142455267329e-05, 'epoch': 0.1}) +[2026-10-03 19:11:53,689][starVLA.training.train_starvla][INFO] - Step 1363, Loss: {'train/loss': 0.012119866907596588, 'train/grad_norm_pre_clip': 8.963658332824707, 'timing/data': 0.0003749399911612272, 'timing/model': 0.39460214899736457, 'learning_rate/qwen_vl_interface': 8.52593786982355e-06, 'learning_rate/action_model': 8.525937869823551e-05, 'epoch': 0.1}) +[2026-10-03 19:11:54,395][starVLA.training.train_starvla][INFO] - Step 1364, Loss: {'train/loss': 0.0853537991642952, 'train/grad_norm_pre_clip': 81.6041259765625, 'timing/data': 0.00022570000146515667, 'timing/model': 0.39516765999724157, 'learning_rate/qwen_vl_interface': 8.523731937765456e-06, 'learning_rate/action_model': 8.523731937765456e-05, 'epoch': 0.1}) +[2026-10-03 19:11:55,305][starVLA.training.train_starvla][INFO] - Step 1365, Loss: {'train/loss': 0.09387282282114029, 'train/grad_norm_pre_clip': 36.828792572021484, 'timing/data': 0.00023879899526946247, 'timing/model': 0.3945619500009343, 'learning_rate/qwen_vl_interface': 8.521524659999823e-06, 'learning_rate/action_model': 8.521524659999822e-05, 'epoch': 0.1}) +[2026-10-03 19:11:56,010][starVLA.training.train_starvla][INFO] - Step 1366, Loss: {'train/loss': 0.008509626612067223, 'train/grad_norm_pre_clip': 4.26170539855957, 'timing/data': 0.00037094001891091466, 'timing/model': 0.3935932199819945, 'learning_rate/qwen_vl_interface': 8.519316037433976e-06, 'learning_rate/action_model': 8.519316037433977e-05, 'epoch': 0.1}) +[2026-10-03 19:11:56,715][starVLA.training.train_starvla][INFO] - Step 1367, Loss: {'train/loss': 0.4591980576515198, 'train/grad_norm_pre_clip': 51.512351989746094, 'timing/data': 0.0003121399786323309, 'timing/model': 0.3942901100090239, 'learning_rate/qwen_vl_interface': 8.517106070975801e-06, 'learning_rate/action_model': 8.5171060709758e-05, 'epoch': 0.1}) +[2026-10-03 19:11:57,420][starVLA.training.train_starvla][INFO] - Step 1368, Loss: {'train/loss': 0.4787147045135498, 'train/grad_norm_pre_clip': 76.0771484375, 'timing/data': 0.0002617499849293381, 'timing/model': 0.395166649017483, 'learning_rate/qwen_vl_interface': 8.514894761533729e-06, 'learning_rate/action_model': 8.514894761533728e-05, 'epoch': 0.1}) +[2026-10-03 19:11:58,127][starVLA.training.train_starvla][INFO] - Step 1369, Loss: {'train/loss': 0.0008736252784729004, 'train/grad_norm_pre_clip': 0.3071592450141907, 'timing/data': 0.00033288999111391604, 'timing/model': 0.3950453790021129, 'learning_rate/qwen_vl_interface': 8.512682110016747e-06, 'learning_rate/action_model': 8.512682110016747e-05, 'epoch': 0.1}) +[2026-10-03 19:11:59,030][starVLA.training.train_starvla][INFO] - Step 1370, Loss: {'train/loss': 0.3551199436187744, 'train/grad_norm_pre_clip': 257.0177001953125, 'timing/data': 0.000364869978511706, 'timing/model': 0.3940975400037132, 'learning_rate/qwen_vl_interface': 8.51046811733439e-06, 'learning_rate/action_model': 8.51046811733439e-05, 'epoch': 0.1}) +[2026-10-03 19:11:59,736][starVLA.training.train_starvla][INFO] - Step 1371, Loss: {'train/loss': 0.010598496533930302, 'train/grad_norm_pre_clip': 1.5766570568084717, 'timing/data': 0.0003067700017709285, 'timing/model': 0.3951376490003895, 'learning_rate/qwen_vl_interface': 8.50825278439675e-06, 'learning_rate/action_model': 8.508252784396749e-05, 'epoch': 0.1}) +[2026-10-03 19:12:00,441][starVLA.training.train_starvla][INFO] - Step 1372, Loss: {'train/loss': 0.39361727237701416, 'train/grad_norm_pre_clip': 46.991756439208984, 'timing/data': 0.00022308999905362725, 'timing/model': 0.3954195200058166, 'learning_rate/qwen_vl_interface': 8.506036112114465e-06, 'learning_rate/action_model': 8.506036112114464e-05, 'epoch': 0.1}) +[2026-10-03 19:12:01,145][starVLA.training.train_starvla][INFO] - Step 1373, Loss: {'train/loss': 0.17403766512870789, 'train/grad_norm_pre_clip': 32.3963508605957, 'timing/data': 0.00032513999030925333, 'timing/model': 0.3928708400053438, 'learning_rate/qwen_vl_interface': 8.503818101398725e-06, 'learning_rate/action_model': 8.503818101398725e-05, 'epoch': 0.1}) +[2026-10-03 19:12:01,849][starVLA.training.train_starvla][INFO] - Step 1374, Loss: {'train/loss': 0.047792840749025345, 'train/grad_norm_pre_clip': 22.835346221923828, 'timing/data': 0.00037612998858094215, 'timing/model': 0.3937650099978782, 'learning_rate/qwen_vl_interface': 8.501598753161273e-06, 'learning_rate/action_model': 8.501598753161272e-05, 'epoch': 0.1}) +[2026-10-03 19:12:02,757][starVLA.training.train_starvla][INFO] - Step 1375, Loss: {'train/loss': 0.09762491285800934, 'train/grad_norm_pre_clip': 32.830078125, 'timing/data': 0.00030819998937658966, 'timing/model': 0.5971008289780002, 'learning_rate/qwen_vl_interface': 8.499378068314395e-06, 'learning_rate/action_model': 8.499378068314395e-05, 'epoch': 0.1}) +[2026-10-03 19:12:03,463][starVLA.training.train_starvla][INFO] - Step 1376, Loss: {'train/loss': 0.3236203193664551, 'train/grad_norm_pre_clip': 44.71861267089844, 'timing/data': 0.00027966999914497137, 'timing/model': 0.39489060000050813, 'learning_rate/qwen_vl_interface': 8.497156047770932e-06, 'learning_rate/action_model': 8.497156047770933e-05, 'epoch': 0.1}) +[2026-10-03 19:12:04,168][starVLA.training.train_starvla][INFO] - Step 1377, Loss: {'train/loss': 0.00192068621981889, 'train/grad_norm_pre_clip': 0.5624910593032837, 'timing/data': 0.0003504699852783233, 'timing/model': 0.39373829000396654, 'learning_rate/qwen_vl_interface': 8.494932692444276e-06, 'learning_rate/action_model': 8.494932692444276e-05, 'epoch': 0.1}) +[2026-10-03 19:12:04,876][starVLA.training.train_starvla][INFO] - Step 1378, Loss: {'train/loss': 0.3529602289199829, 'train/grad_norm_pre_clip': 70.77274322509766, 'timing/data': 0.000410310021834448, 'timing/model': 0.39560101000824943, 'learning_rate/qwen_vl_interface': 8.492708003248362e-06, 'learning_rate/action_model': 8.492708003248362e-05, 'epoch': 0.1}) +[2026-10-03 19:12:05,585][starVLA.training.train_starvla][INFO] - Step 1379, Loss: {'train/loss': 0.541571855545044, 'train/grad_norm_pre_clip': 124.37435913085938, 'timing/data': 0.00033845999860204756, 'timing/model': 0.39701972898910753, 'learning_rate/qwen_vl_interface': 8.490481981097674e-06, 'learning_rate/action_model': 8.490481981097674e-05, 'epoch': 0.1}) +[2026-10-03 19:12:06,293][starVLA.training.train_starvla][INFO] - Step 1380, Loss: {'train/loss': 0.2812959551811218, 'train/grad_norm_pre_clip': 69.07521057128906, 'timing/data': 0.00021674999152310193, 'timing/model': 0.3957926889997907, 'learning_rate/qwen_vl_interface': 8.488254626907249e-06, 'learning_rate/action_model': 8.488254626907249e-05, 'epoch': 0.1}) +[2026-10-03 19:12:07,195][starVLA.training.train_starvla][INFO] - Step 1381, Loss: {'train/loss': 0.7825406193733215, 'train/grad_norm_pre_clip': 96.13715362548828, 'timing/data': 0.00022742000874131918, 'timing/model': 0.395170350006083, 'learning_rate/qwen_vl_interface': 8.486025941592664e-06, 'learning_rate/action_model': 8.486025941592664e-05, 'epoch': 0.1}) +[2026-10-03 19:12:07,900][starVLA.training.train_starvla][INFO] - Step 1382, Loss: {'train/loss': 1.5434563159942627, 'train/grad_norm_pre_clip': 131.7100830078125, 'timing/data': 0.000493219995405525, 'timing/model': 0.3935352099942975, 'learning_rate/qwen_vl_interface': 8.483795926070054e-06, 'learning_rate/action_model': 8.483795926070052e-05, 'epoch': 0.1}) +[2026-10-03 19:12:08,606][starVLA.training.train_starvla][INFO] - Step 1383, Loss: {'train/loss': 0.016890177503228188, 'train/grad_norm_pre_clip': 14.806816101074219, 'timing/data': 0.0003255099873058498, 'timing/model': 0.39559245001873933, 'learning_rate/qwen_vl_interface': 8.481564581256087e-06, 'learning_rate/action_model': 8.481564581256087e-05, 'epoch': 0.1}) +[2026-10-03 19:12:09,313][starVLA.training.train_starvla][INFO] - Step 1384, Loss: {'train/loss': 0.2852188050746918, 'train/grad_norm_pre_clip': 62.22821044921875, 'timing/data': 0.0003098500019405037, 'timing/model': 0.3953517700138036, 'learning_rate/qwen_vl_interface': 8.479331908067987e-06, 'learning_rate/action_model': 8.479331908067987e-05, 'epoch': 0.1}) +[2026-10-03 19:12:10,021][starVLA.training.train_starvla][INFO] - Step 1385, Loss: {'train/loss': 0.3791343867778778, 'train/grad_norm_pre_clip': 120.75065612792969, 'timing/data': 0.00024478998966515064, 'timing/model': 0.395027388993185, 'learning_rate/qwen_vl_interface': 8.477097907423524e-06, 'learning_rate/action_model': 8.477097907423524e-05, 'epoch': 0.1}) +[2026-10-03 19:12:10,925][starVLA.training.train_starvla][INFO] - Step 1386, Loss: {'train/loss': 0.40917646884918213, 'train/grad_norm_pre_clip': 88.1917495727539, 'timing/data': 0.0002202800242230296, 'timing/model': 0.5922661290096585, 'learning_rate/qwen_vl_interface': 8.474862580241006e-06, 'learning_rate/action_model': 8.474862580241006e-05, 'epoch': 0.1}) +[2026-10-03 19:12:11,634][starVLA.training.train_starvla][INFO] - Step 1387, Loss: {'train/loss': 0.0009110692190006375, 'train/grad_norm_pre_clip': 0.32953619956970215, 'timing/data': 0.00022635000641457736, 'timing/model': 0.3960409800056368, 'learning_rate/qwen_vl_interface': 8.472625927439299e-06, 'learning_rate/action_model': 8.472625927439299e-05, 'epoch': 0.1}) +[2026-10-03 19:12:12,342][starVLA.training.train_starvla][INFO] - Step 1388, Loss: {'train/loss': 0.02461809106171131, 'train/grad_norm_pre_clip': 18.816028594970703, 'timing/data': 0.0003514400159474462, 'timing/model': 0.3950385490024928, 'learning_rate/qwen_vl_interface': 8.4703879499378e-06, 'learning_rate/action_model': 8.4703879499378e-05, 'epoch': 0.1}) +[2026-10-03 19:12:13,047][starVLA.training.train_starvla][INFO] - Step 1389, Loss: {'train/loss': 0.04284144192934036, 'train/grad_norm_pre_clip': 31.313152313232422, 'timing/data': 0.0003461499873083085, 'timing/model': 0.3941476699837949, 'learning_rate/qwen_vl_interface': 8.46814864865646e-06, 'learning_rate/action_model': 8.46814864865646e-05, 'epoch': 0.1}) +[2026-10-03 19:12:13,752][starVLA.training.train_starvla][INFO] - Step 1390, Loss: {'train/loss': 0.000890772498678416, 'train/grad_norm_pre_clip': 0.4307819902896881, 'timing/data': 0.0003622999938670546, 'timing/model': 0.3939621699973941, 'learning_rate/qwen_vl_interface': 8.465908024515771e-06, 'learning_rate/action_model': 8.46590802451577e-05, 'epoch': 0.1}) +[2026-10-03 19:12:14,656][starVLA.training.train_starvla][INFO] - Step 1391, Loss: {'train/loss': 0.003011152846738696, 'train/grad_norm_pre_clip': 1.7982929944992065, 'timing/data': 0.000338829995598644, 'timing/model': 0.5930107389867771, 'learning_rate/qwen_vl_interface': 8.463666078436769e-06, 'learning_rate/action_model': 8.463666078436769e-05, 'epoch': 0.1}) +[2026-10-03 19:12:15,365][starVLA.training.train_starvla][INFO] - Step 1392, Loss: {'train/loss': 0.0037227619905024767, 'train/grad_norm_pre_clip': 2.3569424152374268, 'timing/data': 0.0003340399998705834, 'timing/model': 0.39531399999395944, 'learning_rate/qwen_vl_interface': 8.46142281134103e-06, 'learning_rate/action_model': 8.461422811341031e-05, 'epoch': 0.1}) +[2026-10-03 19:12:16,071][starVLA.training.train_starvla][INFO] - Step 1393, Loss: {'train/loss': 0.9547927379608154, 'train/grad_norm_pre_clip': 85.12345123291016, 'timing/data': 0.00034331000642850995, 'timing/model': 0.39531418998376466, 'learning_rate/qwen_vl_interface': 8.459178224150683e-06, 'learning_rate/action_model': 8.459178224150683e-05, 'epoch': 0.1}) +[2026-10-03 19:12:16,780][starVLA.training.train_starvla][INFO] - Step 1394, Loss: {'train/loss': 0.3311881721019745, 'train/grad_norm_pre_clip': 62.771446228027344, 'timing/data': 0.0003244300023652613, 'timing/model': 0.3961811890185345, 'learning_rate/qwen_vl_interface': 8.456932317788389e-06, 'learning_rate/action_model': 8.456932317788389e-05, 'epoch': 0.1}) +[2026-10-03 19:12:17,489][starVLA.training.train_starvla][INFO] - Step 1395, Loss: {'train/loss': 0.42000964283943176, 'train/grad_norm_pre_clip': 43.14463806152344, 'timing/data': 0.00034138999762944877, 'timing/model': 0.39556175901088864, 'learning_rate/qwen_vl_interface': 8.454685093177356e-06, 'learning_rate/action_model': 8.454685093177354e-05, 'epoch': 0.1}) +[2026-10-03 19:12:18,198][starVLA.training.train_starvla][INFO] - Step 1396, Loss: {'train/loss': 0.0037041273899376392, 'train/grad_norm_pre_clip': 3.7625365257263184, 'timing/data': 0.00021204000222496688, 'timing/model': 0.396276340004988, 'learning_rate/qwen_vl_interface': 8.452436551241328e-06, 'learning_rate/action_model': 8.452436551241328e-05, 'epoch': 0.1}) +[2026-10-03 19:12:19,107][starVLA.training.train_starvla][INFO] - Step 1397, Loss: {'train/loss': 0.0009051007218658924, 'train/grad_norm_pre_clip': 0.24608805775642395, 'timing/data': 0.0003503600019030273, 'timing/model': 0.39649474000907503, 'learning_rate/qwen_vl_interface': 8.450186692904605e-06, 'learning_rate/action_model': 8.450186692904606e-05, 'epoch': 0.1}) +[2026-10-03 19:12:19,815][starVLA.training.train_starvla][INFO] - Step 1398, Loss: {'train/loss': 0.2608572542667389, 'train/grad_norm_pre_clip': 121.91859436035156, 'timing/data': 0.0003720800159499049, 'timing/model': 0.39569963997928426, 'learning_rate/qwen_vl_interface': 8.447935519092014e-06, 'learning_rate/action_model': 8.447935519092014e-05, 'epoch': 0.1}) +[2026-10-03 19:12:20,524][starVLA.training.train_starvla][INFO] - Step 1399, Loss: {'train/loss': 0.2567692697048187, 'train/grad_norm_pre_clip': 22.675018310546875, 'timing/data': 0.0003226500120945275, 'timing/model': 0.3960054999915883, 'learning_rate/qwen_vl_interface': 8.445683030728925e-06, 'learning_rate/action_model': 8.445683030728925e-05, 'epoch': 0.1}) +[2026-10-03 19:12:21,231][starVLA.training.train_starvla][INFO] - Step 1400, Loss: {'train/loss': 0.0048749810084700584, 'train/grad_norm_pre_clip': 4.737948417663574, 'timing/data': 0.00022022001212462783, 'timing/model': 0.3940527600061614, 'learning_rate/qwen_vl_interface': 8.443429228741256e-06, 'learning_rate/action_model': 8.443429228741255e-05, 'epoch': 0.1}) +[2026-10-03 19:12:21,938][starVLA.training.train_starvla][INFO] - Step 1401, Loss: {'train/loss': 0.03852624446153641, 'train/grad_norm_pre_clip': 34.2741813659668, 'timing/data': 0.00022029000683687627, 'timing/model': 0.3953299499989953, 'learning_rate/qwen_vl_interface': 8.441174114055457e-06, 'learning_rate/action_model': 8.441174114055456e-05, 'epoch': 0.1}) +[2026-10-03 19:12:22,845][starVLA.training.train_starvla][INFO] - Step 1402, Loss: {'train/loss': 0.07345587015151978, 'train/grad_norm_pre_clip': 47.92667007446289, 'timing/data': 0.0003577799943741411, 'timing/model': 0.3941137490037363, 'learning_rate/qwen_vl_interface': 8.43891768759852e-06, 'learning_rate/action_model': 8.43891768759852e-05, 'epoch': 0.1}) +[2026-10-03 19:12:23,552][starVLA.training.train_starvla][INFO] - Step 1403, Loss: {'train/loss': 0.017795028164982796, 'train/grad_norm_pre_clip': 6.059598922729492, 'timing/data': 0.0003282699908595532, 'timing/model': 0.39456652000080794, 'learning_rate/qwen_vl_interface': 8.436659950297978e-06, 'learning_rate/action_model': 8.436659950297977e-05, 'epoch': 0.1}) +[2026-10-03 19:12:24,257][starVLA.training.train_starvla][INFO] - Step 1404, Loss: {'train/loss': 0.024269167333841324, 'train/grad_norm_pre_clip': 8.41649341583252, 'timing/data': 0.0003741999971680343, 'timing/model': 0.39323163000517525, 'learning_rate/qwen_vl_interface': 8.4344009030819e-06, 'learning_rate/action_model': 8.4344009030819e-05, 'epoch': 0.1}) +[2026-10-03 19:12:24,966][starVLA.training.train_starvla][INFO] - Step 1405, Loss: {'train/loss': 0.40198829770088196, 'train/grad_norm_pre_clip': 106.44378662109375, 'timing/data': 0.0002160299918614328, 'timing/model': 0.39644688001135364, 'learning_rate/qwen_vl_interface': 8.432140546878899e-06, 'learning_rate/action_model': 8.432140546878898e-05, 'epoch': 0.1}) +[2026-10-03 19:12:25,672][starVLA.training.train_starvla][INFO] - Step 1406, Loss: {'train/loss': 0.024440687149763107, 'train/grad_norm_pre_clip': 18.40157699584961, 'timing/data': 0.000353850016836077, 'timing/model': 0.3953032790159341, 'learning_rate/qwen_vl_interface': 8.429878882618116e-06, 'learning_rate/action_model': 8.429878882618116e-05, 'epoch': 0.1}) +[2026-10-03 19:12:26,577][starVLA.training.train_starvla][INFO] - Step 1407, Loss: {'train/loss': 0.017339637503027916, 'train/grad_norm_pre_clip': 6.493662357330322, 'timing/data': 0.0002200100279878825, 'timing/model': 0.5920441990019754, 'learning_rate/qwen_vl_interface': 8.42761591122924e-06, 'learning_rate/action_model': 8.427615911229241e-05, 'epoch': 0.1}) +[2026-10-03 19:12:27,284][starVLA.training.train_starvla][INFO] - Step 1408, Loss: {'train/loss': 0.16977521777153015, 'train/grad_norm_pre_clip': 84.37834167480469, 'timing/data': 0.00021930999355390668, 'timing/model': 0.3957942790002562, 'learning_rate/qwen_vl_interface': 8.425351633642491e-06, 'learning_rate/action_model': 8.425351633642492e-05, 'epoch': 0.1}) +[2026-10-03 19:12:27,991][starVLA.training.train_starvla][INFO] - Step 1409, Loss: {'train/loss': 0.03989889472723007, 'train/grad_norm_pre_clip': 19.643461227416992, 'timing/data': 0.0003550600085873157, 'timing/model': 0.39436167999519967, 'learning_rate/qwen_vl_interface': 8.423086050788632e-06, 'learning_rate/action_model': 8.423086050788631e-05, 'epoch': 0.1}) +[2026-10-03 19:12:28,699][starVLA.training.train_starvla][INFO] - Step 1410, Loss: {'train/loss': 0.006468844600021839, 'train/grad_norm_pre_clip': 4.197065353393555, 'timing/data': 0.0003547400119714439, 'timing/model': 0.395410669996636, 'learning_rate/qwen_vl_interface': 8.420819163598953e-06, 'learning_rate/action_model': 8.420819163598952e-05, 'epoch': 0.1}) +[2026-10-03 19:12:29,409][starVLA.training.train_starvla][INFO] - Step 1411, Loss: {'train/loss': 0.000504366063978523, 'train/grad_norm_pre_clip': 0.13308140635490417, 'timing/data': 0.0002716499730013311, 'timing/model': 0.397268160013482, 'learning_rate/qwen_vl_interface': 8.418550973005289e-06, 'learning_rate/action_model': 8.418550973005289e-05, 'epoch': 0.1}) +[2026-10-03 19:12:30,318][starVLA.training.train_starvla][INFO] - Step 1412, Loss: {'train/loss': 0.0008070026524364948, 'train/grad_norm_pre_clip': 0.2687835991382599, 'timing/data': 0.0003518399898894131, 'timing/model': 0.5979963680147193, 'learning_rate/qwen_vl_interface': 8.416281479940005e-06, 'learning_rate/action_model': 8.416281479940005e-05, 'epoch': 0.1}) +[2026-10-03 19:12:31,025][starVLA.training.train_starvla][INFO] - Step 1413, Loss: {'train/loss': 0.5633260011672974, 'train/grad_norm_pre_clip': 54.45302963256836, 'timing/data': 0.0003325300058349967, 'timing/model': 0.3959321889851708, 'learning_rate/qwen_vl_interface': 8.414010685336008e-06, 'learning_rate/action_model': 8.414010685336008e-05, 'epoch': 0.1}) +[2026-10-03 19:12:31,731][starVLA.training.train_starvla][INFO] - Step 1414, Loss: {'train/loss': 0.0023954377975314856, 'train/grad_norm_pre_clip': 0.39383652806282043, 'timing/data': 0.0003562399942893535, 'timing/model': 0.393903069983935, 'learning_rate/qwen_vl_interface': 8.411738590126733e-06, 'learning_rate/action_model': 8.411738590126732e-05, 'epoch': 0.1}) +[2026-10-03 19:12:32,440][starVLA.training.train_starvla][INFO] - Step 1415, Loss: {'train/loss': 0.00021401586127467453, 'train/grad_norm_pre_clip': 0.08825317770242691, 'timing/data': 0.00033974999678321183, 'timing/model': 0.39554001999204047, 'learning_rate/qwen_vl_interface': 8.409465195246152e-06, 'learning_rate/action_model': 8.409465195246152e-05, 'epoch': 0.1}) +[2026-10-03 19:12:33,148][starVLA.training.train_starvla][INFO] - Step 1416, Loss: {'train/loss': 0.2629537582397461, 'train/grad_norm_pre_clip': 114.41071319580078, 'timing/data': 0.00035396998282521963, 'timing/model': 0.39612338901497424, 'learning_rate/qwen_vl_interface': 8.407190501628772e-06, 'learning_rate/action_model': 8.407190501628772e-05, 'epoch': 0.1}) +[2026-10-03 19:12:33,854][starVLA.training.train_starvla][INFO] - Step 1417, Loss: {'train/loss': 0.21086817979812622, 'train/grad_norm_pre_clip': 83.19781494140625, 'timing/data': 0.0003693499893415719, 'timing/model': 0.39502662999439053, 'learning_rate/qwen_vl_interface': 8.404914510209635e-06, 'learning_rate/action_model': 8.404914510209635e-05, 'epoch': 0.1}) +[2026-10-03 19:12:34,761][starVLA.training.train_starvla][INFO] - Step 1418, Loss: {'train/loss': 0.6210628747940063, 'train/grad_norm_pre_clip': 285.5239562988281, 'timing/data': 0.00037193999742157757, 'timing/model': 0.395816010015551, 'learning_rate/qwen_vl_interface': 8.402637221924315e-06, 'learning_rate/action_model': 8.402637221924314e-05, 'epoch': 0.1}) +[2026-10-03 19:12:35,443][starVLA.training.train_starvla][INFO] - Step 1419, Loss: {'train/loss': 0.24687007069587708, 'train/grad_norm_pre_clip': 49.284488677978516, 'timing/data': 0.0003553899878170341, 'timing/model': 0.3942696799931582, 'learning_rate/qwen_vl_interface': 8.400358637708919e-06, 'learning_rate/action_model': 8.400358637708918e-05, 'epoch': 0.1}) +[2026-10-03 19:12:36,150][starVLA.training.train_starvla][INFO] - Step 1420, Loss: {'train/loss': 0.0004372904950287193, 'train/grad_norm_pre_clip': 0.12293779104948044, 'timing/data': 0.0003571700071915984, 'timing/model': 0.3946453589887824, 'learning_rate/qwen_vl_interface': 8.398078758500086e-06, 'learning_rate/action_model': 8.398078758500086e-05, 'epoch': 0.1}) +[2026-10-03 19:12:36,860][starVLA.training.train_starvla][INFO] - Step 1421, Loss: {'train/loss': 0.0006963650812394917, 'train/grad_norm_pre_clip': 0.2375783622264862, 'timing/data': 0.00035343001945875585, 'timing/model': 0.3960853600001428, 'learning_rate/qwen_vl_interface': 8.395797585234987e-06, 'learning_rate/action_model': 8.395797585234988e-05, 'epoch': 0.1}) +[2026-10-03 19:12:37,568][starVLA.training.train_starvla][INFO] - Step 1422, Loss: {'train/loss': 0.0002525242744013667, 'train/grad_norm_pre_clip': 0.1335919350385666, 'timing/data': 0.00035605000448413193, 'timing/model': 0.39538027899106964, 'learning_rate/qwen_vl_interface': 8.39351511885133e-06, 'learning_rate/action_model': 8.393515118851329e-05, 'epoch': 0.1}) +[2026-10-03 19:12:38,474][starVLA.training.train_starvla][INFO] - Step 1423, Loss: {'train/loss': 0.0028198091313242912, 'train/grad_norm_pre_clip': 1.4383578300476074, 'timing/data': 0.0003526800137478858, 'timing/model': 0.3943007799971383, 'learning_rate/qwen_vl_interface': 8.391231360287348e-06, 'learning_rate/action_model': 8.391231360287347e-05, 'epoch': 0.1}) +[2026-10-03 19:12:39,184][starVLA.training.train_starvla][INFO] - Step 1424, Loss: {'train/loss': 0.00042114287498407066, 'train/grad_norm_pre_clip': 0.17130379378795624, 'timing/data': 0.00036403999547474086, 'timing/model': 0.39606359900790267, 'learning_rate/qwen_vl_interface': 8.388946310481807e-06, 'learning_rate/action_model': 8.388946310481807e-05, 'epoch': 0.1}) +[2026-10-03 19:12:39,890][starVLA.training.train_starvla][INFO] - Step 1425, Loss: {'train/loss': 0.5001744627952576, 'train/grad_norm_pre_clip': 44.152000427246094, 'timing/data': 0.00022255998919717968, 'timing/model': 0.3955132899864111, 'learning_rate/qwen_vl_interface': 8.386659970374007e-06, 'learning_rate/action_model': 8.386659970374006e-05, 'epoch': 0.1}) +[2026-10-03 19:12:40,597][starVLA.training.train_starvla][INFO] - Step 1426, Loss: {'train/loss': 0.3136887550354004, 'train/grad_norm_pre_clip': 57.59804916381836, 'timing/data': 0.0003619900089688599, 'timing/model': 0.394341349019669, 'learning_rate/qwen_vl_interface': 8.384372340903775e-06, 'learning_rate/action_model': 8.384372340903775e-05, 'epoch': 0.1}) +[2026-10-03 19:12:41,308][starVLA.training.train_starvla][INFO] - Step 1427, Loss: {'train/loss': 0.0002868190058507025, 'train/grad_norm_pre_clip': 0.16099070012569427, 'timing/data': 0.0003518899902701378, 'timing/model': 0.39737136001349427, 'learning_rate/qwen_vl_interface': 8.382083423011472e-06, 'learning_rate/action_model': 8.382083423011471e-05, 'epoch': 0.1}) +[2026-10-03 19:12:42,214][starVLA.training.train_starvla][INFO] - Step 1428, Loss: {'train/loss': 0.0008314505685120821, 'train/grad_norm_pre_clip': 0.6282627582550049, 'timing/data': 0.00035684998147189617, 'timing/model': 0.5924675390124321, 'learning_rate/qwen_vl_interface': 8.379793217637979e-06, 'learning_rate/action_model': 8.379793217637979e-05, 'epoch': 0.1}) +[2026-10-03 19:12:42,921][starVLA.training.train_starvla][INFO] - Step 1429, Loss: {'train/loss': 0.0009177627507597208, 'train/grad_norm_pre_clip': 0.45363298058509827, 'timing/data': 0.0003229000139981508, 'timing/model': 0.39468448000843637, 'learning_rate/qwen_vl_interface': 8.377501725724719e-06, 'learning_rate/action_model': 8.377501725724719e-05, 'epoch': 0.1}) +[2026-10-03 19:12:43,629][starVLA.training.train_starvla][INFO] - Step 1430, Loss: {'train/loss': 0.00023599124688189477, 'train/grad_norm_pre_clip': 0.07040978968143463, 'timing/data': 0.0003786899906117469, 'timing/model': 0.3954733800201211, 'learning_rate/qwen_vl_interface': 8.375208948213636e-06, 'learning_rate/action_model': 8.375208948213635e-05, 'epoch': 0.1}) +[2026-10-03 19:12:44,339][starVLA.training.train_starvla][INFO] - Step 1431, Loss: {'train/loss': 0.353925883769989, 'train/grad_norm_pre_clip': 86.05744934082031, 'timing/data': 0.00023106997832655907, 'timing/model': 0.39663582999492064, 'learning_rate/qwen_vl_interface': 8.372914886047203e-06, 'learning_rate/action_model': 8.372914886047203e-05, 'epoch': 0.1}) +[2026-10-03 19:12:45,048][starVLA.training.train_starvla][INFO] - Step 1432, Loss: {'train/loss': 0.3325978219509125, 'train/grad_norm_pre_clip': 56.773353576660156, 'timing/data': 0.00024641997879371047, 'timing/model': 0.3972098090162035, 'learning_rate/qwen_vl_interface': 8.370619540168422e-06, 'learning_rate/action_model': 8.370619540168423e-05, 'epoch': 0.1}) +[2026-10-03 19:12:45,959][starVLA.training.train_starvla][INFO] - Step 1433, Loss: {'train/loss': 0.0001104569309973158, 'train/grad_norm_pre_clip': 0.07449087500572205, 'timing/data': 0.00022096000611782074, 'timing/model': 0.5978968389972579, 'learning_rate/qwen_vl_interface': 8.368322911520826e-06, 'learning_rate/action_model': 8.368322911520826e-05, 'epoch': 0.1}) +[2026-10-03 19:12:46,668][starVLA.training.train_starvla][INFO] - Step 1434, Loss: {'train/loss': 0.3034616708755493, 'train/grad_norm_pre_clip': 12.295441627502441, 'timing/data': 0.00038053898606449366, 'timing/model': 0.39607398997759447, 'learning_rate/qwen_vl_interface': 8.36602500104847e-06, 'learning_rate/action_model': 8.36602500104847e-05, 'epoch': 0.1}) +[2026-10-03 19:12:47,377][starVLA.training.train_starvla][INFO] - Step 1435, Loss: {'train/loss': 0.11326112598180771, 'train/grad_norm_pre_clip': 70.87382507324219, 'timing/data': 0.0003440200234763324, 'timing/model': 0.3961270799918566, 'learning_rate/qwen_vl_interface': 8.36372580969594e-06, 'learning_rate/action_model': 8.36372580969594e-05, 'epoch': 0.11}) +[2026-10-03 19:12:48,087][starVLA.training.train_starvla][INFO] - Step 1436, Loss: {'train/loss': 0.0005570445209741592, 'train/grad_norm_pre_clip': 0.2266925573348999, 'timing/data': 0.0003033299872186035, 'timing/model': 0.395912230014801, 'learning_rate/qwen_vl_interface': 8.361425338408344e-06, 'learning_rate/action_model': 8.361425338408343e-05, 'epoch': 0.11}) +[2026-10-03 19:12:48,769][starVLA.training.train_starvla][INFO] - Step 1437, Loss: {'train/loss': 0.1224399283528328, 'train/grad_norm_pre_clip': 35.31425857543945, 'timing/data': 0.0003336400259286165, 'timing/model': 0.3945081990095787, 'learning_rate/qwen_vl_interface': 8.359123588131319e-06, 'learning_rate/action_model': 8.359123588131319e-05, 'epoch': 0.11}) +[2026-10-03 19:12:49,475][starVLA.training.train_starvla][INFO] - Step 1438, Loss: {'train/loss': 0.552478015422821, 'train/grad_norm_pre_clip': 125.31428527832031, 'timing/data': 0.00027771000168286264, 'timing/model': 0.3942637099826243, 'learning_rate/qwen_vl_interface': 8.35682055981103e-06, 'learning_rate/action_model': 8.35682055981103e-05, 'epoch': 0.11}) +[2026-10-03 19:12:50,378][starVLA.training.train_starvla][INFO] - Step 1439, Loss: {'train/loss': 0.12921547889709473, 'train/grad_norm_pre_clip': 120.61764526367188, 'timing/data': 0.00021959000150673091, 'timing/model': 0.396284148999257, 'learning_rate/qwen_vl_interface': 8.354516254394163e-06, 'learning_rate/action_model': 8.354516254394162e-05, 'epoch': 0.11}) +[2026-10-03 19:12:51,086][starVLA.training.train_starvla][INFO] - Step 1440, Loss: {'train/loss': 0.0001990755117731169, 'train/grad_norm_pre_clip': 0.14694681763648987, 'timing/data': 0.00033571000676602125, 'timing/model': 0.3955082190223038, 'learning_rate/qwen_vl_interface': 8.352210672827932e-06, 'learning_rate/action_model': 8.352210672827931e-05, 'epoch': 0.11}) +[2026-10-03 19:12:51,793][starVLA.training.train_starvla][INFO] - Step 1441, Loss: {'train/loss': 0.00023185140162240714, 'train/grad_norm_pre_clip': 0.1343371868133545, 'timing/data': 0.00032926001586019993, 'timing/model': 0.39616128997295164, 'learning_rate/qwen_vl_interface': 8.349903816060074e-06, 'learning_rate/action_model': 8.349903816060075e-05, 'epoch': 0.11}) +[2026-10-03 19:12:52,501][starVLA.training.train_starvla][INFO] - Step 1442, Loss: {'train/loss': 0.027715085074305534, 'train/grad_norm_pre_clip': 55.54775619506836, 'timing/data': 0.00034338998375460505, 'timing/model': 0.39547932901768945, 'learning_rate/qwen_vl_interface': 8.34759568503885e-06, 'learning_rate/action_model': 8.347595685038851e-05, 'epoch': 0.11}) +[2026-10-03 19:12:53,209][starVLA.training.train_starvla][INFO] - Step 1443, Loss: {'train/loss': 0.3712925910949707, 'train/grad_norm_pre_clip': 55.69013595581055, 'timing/data': 0.00033151000388897955, 'timing/model': 0.39405742997769266, 'learning_rate/qwen_vl_interface': 8.345286280713049e-06, 'learning_rate/action_model': 8.345286280713048e-05, 'epoch': 0.11}) +[2026-10-03 19:12:54,116][starVLA.training.train_starvla][INFO] - Step 1444, Loss: {'train/loss': 0.06095145642757416, 'train/grad_norm_pre_clip': 39.244232177734375, 'timing/data': 0.00035063002724200487, 'timing/model': 0.39488400900154375, 'learning_rate/qwen_vl_interface': 8.342975604031974e-06, 'learning_rate/action_model': 8.342975604031974e-05, 'epoch': 0.11}) +[2026-10-03 19:12:54,825][starVLA.training.train_starvla][INFO] - Step 1445, Loss: {'train/loss': 0.00047245676978491247, 'train/grad_norm_pre_clip': 0.7317140102386475, 'timing/data': 0.0003517900186125189, 'timing/model': 0.39646792999701574, 'learning_rate/qwen_vl_interface': 8.340663655945462e-06, 'learning_rate/action_model': 8.340663655945462e-05, 'epoch': 0.11}) +[2026-10-03 19:12:55,532][starVLA.training.train_starvla][INFO] - Step 1446, Loss: {'train/loss': 0.3797561228275299, 'train/grad_norm_pre_clip': 81.48251342773438, 'timing/data': 0.00021843001013621688, 'timing/model': 0.39434636902296916, 'learning_rate/qwen_vl_interface': 8.338350437403866e-06, 'learning_rate/action_model': 8.338350437403865e-05, 'epoch': 0.11}) +[2026-10-03 19:12:56,242][starVLA.training.train_starvla][INFO] - Step 1447, Loss: {'train/loss': 0.014749809168279171, 'train/grad_norm_pre_clip': 10.02828598022461, 'timing/data': 0.0003935899876523763, 'timing/model': 0.39616884899442084, 'learning_rate/qwen_vl_interface': 8.336035949358063e-06, 'learning_rate/action_model': 8.336035949358062e-05, 'epoch': 0.11}) +[2026-10-03 19:12:56,951][starVLA.training.train_starvla][INFO] - Step 1448, Loss: {'train/loss': 0.2828426957130432, 'train/grad_norm_pre_clip': 65.66629791259766, 'timing/data': 0.00031057000160217285, 'timing/model': 0.3959118099883199, 'learning_rate/qwen_vl_interface': 8.333720192759448e-06, 'learning_rate/action_model': 8.333720192759448e-05, 'epoch': 0.11}) +[2026-10-03 19:12:57,851][starVLA.training.train_starvla][INFO] - Step 1449, Loss: {'train/loss': 0.025172317400574684, 'train/grad_norm_pre_clip': 14.234477043151855, 'timing/data': 0.0003004099999088794, 'timing/model': 0.5889703500142787, 'learning_rate/qwen_vl_interface': 8.331403168559946e-06, 'learning_rate/action_model': 8.331403168559945e-05, 'epoch': 0.11}) +[2026-10-03 19:12:58,558][starVLA.training.train_starvla][INFO] - Step 1450, Loss: {'train/loss': 0.4326697289943695, 'train/grad_norm_pre_clip': 66.82106018066406, 'timing/data': 0.00034637999488040805, 'timing/model': 0.39517418001196347, 'learning_rate/qwen_vl_interface': 8.329084877711995e-06, 'learning_rate/action_model': 8.329084877711996e-05, 'epoch': 0.11}) +[2026-10-03 19:12:59,262][starVLA.training.train_starvla][INFO] - Step 1451, Loss: {'train/loss': 0.0011303918436169624, 'train/grad_norm_pre_clip': 0.5822813510894775, 'timing/data': 0.0002530700003262609, 'timing/model': 0.39363289999892004, 'learning_rate/qwen_vl_interface': 8.32676532116856e-06, 'learning_rate/action_model': 8.326765321168559e-05, 'epoch': 0.11}) +[2026-10-03 19:12:59,966][starVLA.training.train_starvla][INFO] - Step 1452, Loss: {'train/loss': 0.9739091992378235, 'train/grad_norm_pre_clip': 66.17550659179688, 'timing/data': 0.0003114999854005873, 'timing/model': 0.39351958900806494, 'learning_rate/qwen_vl_interface': 8.324444499883119e-06, 'learning_rate/action_model': 8.324444499883119e-05, 'epoch': 0.11}) +[2026-10-03 19:13:00,673][starVLA.training.train_starvla][INFO] - Step 1453, Loss: {'train/loss': 0.0046860831789672375, 'train/grad_norm_pre_clip': 2.76933217048645, 'timing/data': 0.00030278999474830925, 'timing/model': 0.39530589000787586, 'learning_rate/qwen_vl_interface': 8.322122414809678e-06, 'learning_rate/action_model': 8.322122414809677e-05, 'epoch': 0.11}) +[2026-10-03 19:13:01,567][starVLA.training.train_starvla][INFO] - Step 1454, Loss: {'train/loss': 0.00807034783065319, 'train/grad_norm_pre_clip': 2.4515981674194336, 'timing/data': 0.00020551000488922, 'timing/model': 0.5841708410007413, 'learning_rate/qwen_vl_interface': 8.319799066902755e-06, 'learning_rate/action_model': 8.319799066902755e-05, 'epoch': 0.11}) +[2026-10-03 19:13:02,273][starVLA.training.train_starvla][INFO] - Step 1455, Loss: {'train/loss': 0.0008360572392120957, 'train/grad_norm_pre_clip': 0.7216512560844421, 'timing/data': 0.00031096002203412354, 'timing/model': 0.394916760007618, 'learning_rate/qwen_vl_interface': 8.317474457117393e-06, 'learning_rate/action_model': 8.317474457117392e-05, 'epoch': 0.11}) +[2026-10-03 19:13:02,980][starVLA.training.train_starvla][INFO] - Step 1456, Loss: {'train/loss': 0.46685367822647095, 'train/grad_norm_pre_clip': 44.77251434326172, 'timing/data': 0.0003237300261389464, 'timing/model': 0.39372314998763613, 'learning_rate/qwen_vl_interface': 8.315148586409148e-06, 'learning_rate/action_model': 8.315148586409148e-05, 'epoch': 0.11}) +[2026-10-03 19:13:03,687][starVLA.training.train_starvla][INFO] - Step 1457, Loss: {'train/loss': 0.15870429575443268, 'train/grad_norm_pre_clip': 193.54762268066406, 'timing/data': 0.00021928001660853624, 'timing/model': 0.39531558999442495, 'learning_rate/qwen_vl_interface': 8.312821455734104e-06, 'learning_rate/action_model': 8.312821455734103e-05, 'epoch': 0.11}) +[2026-10-03 19:13:04,394][starVLA.training.train_starvla][INFO] - Step 1458, Loss: {'train/loss': 1.2406771183013916, 'train/grad_norm_pre_clip': 123.88882446289062, 'timing/data': 0.0003527499793563038, 'timing/model': 0.39404397999169305, 'learning_rate/qwen_vl_interface': 8.31049306604885e-06, 'learning_rate/action_model': 8.310493066048849e-05, 'epoch': 0.11}) +[2026-10-03 19:13:05,102][starVLA.training.train_starvla][INFO] - Step 1459, Loss: {'train/loss': 0.09138675034046173, 'train/grad_norm_pre_clip': 41.51155471801758, 'timing/data': 0.0003256100171711296, 'timing/model': 0.3946965300128795, 'learning_rate/qwen_vl_interface': 8.3081634183105e-06, 'learning_rate/action_model': 8.308163418310501e-05, 'epoch': 0.11}) +[2026-10-03 19:13:06,010][starVLA.training.train_starvla][INFO] - Step 1460, Loss: {'train/loss': 0.010899187065660954, 'train/grad_norm_pre_clip': 6.2788262367248535, 'timing/data': 0.0003460099978838116, 'timing/model': 0.39461228999425657, 'learning_rate/qwen_vl_interface': 8.305832513476689e-06, 'learning_rate/action_model': 8.305832513476688e-05, 'epoch': 0.11}) +[2026-10-03 19:13:06,719][starVLA.training.train_starvla][INFO] - Step 1461, Loss: {'train/loss': 0.005196747370064259, 'train/grad_norm_pre_clip': 6.032904148101807, 'timing/data': 0.0004535999905783683, 'timing/model': 0.3961987200018484, 'learning_rate/qwen_vl_interface': 8.303500352505557e-06, 'learning_rate/action_model': 8.303500352505557e-05, 'epoch': 0.11}) +[2026-10-03 19:13:07,427][starVLA.training.train_starvla][INFO] - Step 1462, Loss: {'train/loss': 0.5158186554908752, 'train/grad_norm_pre_clip': 172.18455505371094, 'timing/data': 0.00036153997643850744, 'timing/model': 0.39509309999994, 'learning_rate/qwen_vl_interface': 8.301166936355773e-06, 'learning_rate/action_model': 8.301166936355773e-05, 'epoch': 0.11}) +[2026-10-03 19:13:08,136][starVLA.training.train_starvla][INFO] - Step 1463, Loss: {'train/loss': 0.5137336254119873, 'train/grad_norm_pre_clip': 108.8263168334961, 'timing/data': 0.0003839799901470542, 'timing/model': 0.3960045900021214, 'learning_rate/qwen_vl_interface': 8.298832265986514e-06, 'learning_rate/action_model': 8.298832265986513e-05, 'epoch': 0.11}) +[2026-10-03 19:13:08,845][starVLA.training.train_starvla][INFO] - Step 1464, Loss: {'train/loss': 0.20592352747917175, 'train/grad_norm_pre_clip': 113.12225341796875, 'timing/data': 0.0003559199976734817, 'timing/model': 0.3965395599952899, 'learning_rate/qwen_vl_interface': 8.296496342357475e-06, 'learning_rate/action_model': 8.296496342357474e-05, 'epoch': 0.11}) +[2026-10-03 19:13:09,755][starVLA.training.train_starvla][INFO] - Step 1465, Loss: {'train/loss': 0.2790561616420746, 'train/grad_norm_pre_clip': 9.673503875732422, 'timing/data': 0.0002566699986346066, 'timing/model': 0.395834390015807, 'learning_rate/qwen_vl_interface': 8.294159166428864e-06, 'learning_rate/action_model': 8.294159166428864e-05, 'epoch': 0.11}) +[2026-10-03 19:13:10,464][starVLA.training.train_starvla][INFO] - Step 1466, Loss: {'train/loss': 0.001120769651606679, 'train/grad_norm_pre_clip': 0.7570364475250244, 'timing/data': 0.00035387001116760075, 'timing/model': 0.39569411901175044, 'learning_rate/qwen_vl_interface': 8.29182073916141e-06, 'learning_rate/action_model': 8.29182073916141e-05, 'epoch': 0.11}) +[2026-10-03 19:13:11,170][starVLA.training.train_starvla][INFO] - Step 1467, Loss: {'train/loss': 0.31045001745224, 'train/grad_norm_pre_clip': 55.75690460205078, 'timing/data': 0.00022730001364834607, 'timing/model': 0.39440072898287326, 'learning_rate/qwen_vl_interface': 8.28948106151635e-06, 'learning_rate/action_model': 8.28948106151635e-05, 'epoch': 0.11}) +[2026-10-03 19:13:11,875][starVLA.training.train_starvla][INFO] - Step 1468, Loss: {'train/loss': 0.0032981405965983868, 'train/grad_norm_pre_clip': 0.9765739440917969, 'timing/data': 0.0002633900148794055, 'timing/model': 0.39406130000134, 'learning_rate/qwen_vl_interface': 8.287140134455435e-06, 'learning_rate/action_model': 8.287140134455435e-05, 'epoch': 0.11}) +[2026-10-03 19:13:12,583][starVLA.training.train_starvla][INFO] - Step 1469, Loss: {'train/loss': 0.0012898932909592986, 'train/grad_norm_pre_clip': 0.6827514171600342, 'timing/data': 0.00034583001979626715, 'timing/model': 0.3962651689944323, 'learning_rate/qwen_vl_interface': 8.284797958940934e-06, 'learning_rate/action_model': 8.284797958940935e-05, 'epoch': 0.11}) +[2026-10-03 19:13:13,490][starVLA.training.train_starvla][INFO] - Step 1470, Loss: {'train/loss': 0.00691515626385808, 'train/grad_norm_pre_clip': 0.6336362957954407, 'timing/data': 0.00033750999136827886, 'timing/model': 0.5937582089973148, 'learning_rate/qwen_vl_interface': 8.282454535935626e-06, 'learning_rate/action_model': 8.282454535935626e-05, 'epoch': 0.11}) +[2026-10-03 19:13:14,198][starVLA.training.train_starvla][INFO] - Step 1471, Loss: {'train/loss': 0.12040278315544128, 'train/grad_norm_pre_clip': 69.96795654296875, 'timing/data': 0.0002564699971117079, 'timing/model': 0.39622921898262575, 'learning_rate/qwen_vl_interface': 8.280109866402805e-06, 'learning_rate/action_model': 8.280109866402806e-05, 'epoch': 0.11}) +[2026-10-03 19:13:14,907][starVLA.training.train_starvla][INFO] - Step 1472, Loss: {'train/loss': 0.03375573083758354, 'train/grad_norm_pre_clip': 16.346811294555664, 'timing/data': 0.0002232600236311555, 'timing/model': 0.39527372899465263, 'learning_rate/qwen_vl_interface': 8.277763951306277e-06, 'learning_rate/action_model': 8.277763951306277e-05, 'epoch': 0.11}) +[2026-10-03 19:13:15,615][starVLA.training.train_starvla][INFO] - Step 1473, Loss: {'train/loss': 0.021331537514925003, 'train/grad_norm_pre_clip': 14.414993286132812, 'timing/data': 0.0003256899944972247, 'timing/model': 0.3967158900049981, 'learning_rate/qwen_vl_interface': 8.275416791610356e-06, 'learning_rate/action_model': 8.275416791610355e-05, 'epoch': 0.11}) +[2026-10-03 19:13:16,320][starVLA.training.train_starvla][INFO] - Step 1474, Loss: {'train/loss': 0.003505488159134984, 'train/grad_norm_pre_clip': 1.6475571393966675, 'timing/data': 0.00036679001641459763, 'timing/model': 0.3933982700109482, 'learning_rate/qwen_vl_interface': 8.27306838827987e-06, 'learning_rate/action_model': 8.273068388279871e-05, 'epoch': 0.11}) +[2026-10-03 19:13:17,030][starVLA.training.train_starvla][INFO] - Step 1475, Loss: {'train/loss': 0.07458841800689697, 'train/grad_norm_pre_clip': 22.67460823059082, 'timing/data': 0.00034090998815372586, 'timing/model': 0.39633114999742247, 'learning_rate/qwen_vl_interface': 8.270718742280162e-06, 'learning_rate/action_model': 8.270718742280162e-05, 'epoch': 0.11}) +[2026-10-03 19:13:17,951][starVLA.training.train_starvla][INFO] - Step 1476, Loss: {'train/loss': 0.3540855646133423, 'train/grad_norm_pre_clip': 57.75710678100586, 'timing/data': 0.00024322999524883926, 'timing/model': 0.39672633999725804, 'learning_rate/qwen_vl_interface': 8.268367854577084e-06, 'learning_rate/action_model': 8.268367854577082e-05, 'epoch': 0.11}) +[2026-10-03 19:13:18,669][starVLA.training.train_starvla][INFO] - Step 1477, Loss: {'train/loss': 0.32277974486351013, 'train/grad_norm_pre_clip': 201.0731658935547, 'timing/data': 0.00034383998718112707, 'timing/model': 0.39494523001485504, 'learning_rate/qwen_vl_interface': 8.266015726136991e-06, 'learning_rate/action_model': 8.266015726136991e-05, 'epoch': 0.11}) +[2026-10-03 19:13:19,372][starVLA.training.train_starvla][INFO] - Step 1478, Loss: {'train/loss': 0.1528984159231186, 'train/grad_norm_pre_clip': 168.68087768554688, 'timing/data': 0.00035523000406101346, 'timing/model': 0.3935041499789804, 'learning_rate/qwen_vl_interface': 8.26366235792676e-06, 'learning_rate/action_model': 8.26366235792676e-05, 'epoch': 0.11}) +[2026-10-03 19:13:20,081][starVLA.training.train_starvla][INFO] - Step 1479, Loss: {'train/loss': 0.11245924979448318, 'train/grad_norm_pre_clip': 18.07501220703125, 'timing/data': 0.00023365000379271805, 'timing/model': 0.3963416999904439, 'learning_rate/qwen_vl_interface': 8.261307750913769e-06, 'learning_rate/action_model': 8.261307750913769e-05, 'epoch': 0.11}) +[2026-10-03 19:13:20,791][starVLA.training.train_starvla][INFO] - Step 1480, Loss: {'train/loss': 0.15453067421913147, 'train/grad_norm_pre_clip': 121.79940795898438, 'timing/data': 0.0002859499945770949, 'timing/model': 0.39704106899444014, 'learning_rate/qwen_vl_interface': 8.258951906065908e-06, 'learning_rate/action_model': 8.258951906065908e-05, 'epoch': 0.11}) +[2026-10-03 19:13:21,737][starVLA.training.train_starvla][INFO] - Step 1481, Loss: {'train/loss': 0.00046779142576269805, 'train/grad_norm_pre_clip': 1.4724372625350952, 'timing/data': 0.00021853999351151288, 'timing/model': 0.3949811700149439, 'learning_rate/qwen_vl_interface': 8.256594824351577e-06, 'learning_rate/action_model': 8.256594824351576e-05, 'epoch': 0.11}) +[2026-10-03 19:13:22,438][starVLA.training.train_starvla][INFO] - Step 1482, Loss: {'train/loss': 0.006057095713913441, 'train/grad_norm_pre_clip': 6.487123489379883, 'timing/data': 0.0003697199863381684, 'timing/model': 0.39156286101206206, 'learning_rate/qwen_vl_interface': 8.254236506739683e-06, 'learning_rate/action_model': 8.254236506739681e-05, 'epoch': 0.11}) +[2026-10-03 19:13:23,140][starVLA.training.train_starvla][INFO] - Step 1483, Loss: {'train/loss': 0.005825492553412914, 'train/grad_norm_pre_clip': 3.247786521911621, 'timing/data': 0.00020546000450849533, 'timing/model': 0.3922910499968566, 'learning_rate/qwen_vl_interface': 8.25187695419964e-06, 'learning_rate/action_model': 8.25187695419964e-05, 'epoch': 0.11}) +[2026-10-03 19:13:23,844][starVLA.training.train_starvla][INFO] - Step 1484, Loss: {'train/loss': 0.8574695587158203, 'train/grad_norm_pre_clip': 89.72081756591797, 'timing/data': 0.00033847999293357134, 'timing/model': 0.39159266999922693, 'learning_rate/qwen_vl_interface': 8.249516167701373e-06, 'learning_rate/action_model': 8.249516167701374e-05, 'epoch': 0.11}) +[2026-10-03 19:13:24,543][starVLA.training.train_starvla][INFO] - Step 1485, Loss: {'train/loss': 0.99869704246521, 'train/grad_norm_pre_clip': 37.87066650390625, 'timing/data': 0.0002016299986280501, 'timing/model': 0.39241366999340244, 'learning_rate/qwen_vl_interface': 8.247154148215311e-06, 'learning_rate/action_model': 8.24715414821531e-05, 'epoch': 0.11}) +[2026-10-03 19:13:25,449][starVLA.training.train_starvla][INFO] - Step 1486, Loss: {'train/loss': 0.00036455594818107784, 'train/grad_norm_pre_clip': 0.04862173646688461, 'timing/data': 0.00034276998485438526, 'timing/model': 0.3902768910047598, 'learning_rate/qwen_vl_interface': 8.24479089671239e-06, 'learning_rate/action_model': 8.24479089671239e-05, 'epoch': 0.11}) +[2026-10-03 19:13:26,147][starVLA.training.train_starvla][INFO] - Step 1487, Loss: {'train/loss': 9.406438039150089e-05, 'train/grad_norm_pre_clip': 0.027237923815846443, 'timing/data': 0.0002707000239752233, 'timing/model': 0.392422759992769, 'learning_rate/qwen_vl_interface': 8.242426414164056e-06, 'learning_rate/action_model': 8.242426414164056e-05, 'epoch': 0.11}) +[2026-10-03 19:13:26,845][starVLA.training.train_starvla][INFO] - Step 1488, Loss: {'train/loss': 0.032557204365730286, 'train/grad_norm_pre_clip': 46.5159912109375, 'timing/data': 0.00035195000236853957, 'timing/model': 0.389055939973332, 'learning_rate/qwen_vl_interface': 8.240060701542257e-06, 'learning_rate/action_model': 8.240060701542257e-05, 'epoch': 0.11}) +[2026-10-03 19:13:27,545][starVLA.training.train_starvla][INFO] - Step 1489, Loss: {'train/loss': 0.4382830262184143, 'train/grad_norm_pre_clip': 81.65214538574219, 'timing/data': 0.0002524500014260411, 'timing/model': 0.39228043000912294, 'learning_rate/qwen_vl_interface': 8.237693759819447e-06, 'learning_rate/action_model': 8.237693759819448e-05, 'epoch': 0.11}) +[2026-10-03 19:13:28,243][starVLA.training.train_starvla][INFO] - Step 1490, Loss: {'train/loss': 0.0005681489128619432, 'train/grad_norm_pre_clip': 0.9830617904663086, 'timing/data': 0.000345520005794242, 'timing/model': 0.3903691410087049, 'learning_rate/qwen_vl_interface': 8.235325589968587e-06, 'learning_rate/action_model': 8.235325589968587e-05, 'epoch': 0.11}) +[2026-10-03 19:13:29,147][starVLA.training.train_starvla][INFO] - Step 1491, Loss: {'train/loss': 0.5744841694831848, 'train/grad_norm_pre_clip': 252.82215881347656, 'timing/data': 0.0003300700045656413, 'timing/model': 0.5956720900139771, 'learning_rate/qwen_vl_interface': 8.232956192963143e-06, 'learning_rate/action_model': 8.232956192963144e-05, 'epoch': 0.11}) +[2026-10-03 19:13:29,844][starVLA.training.train_starvla][INFO] - Step 1492, Loss: {'train/loss': 0.9397204518318176, 'train/grad_norm_pre_clip': 22.091976165771484, 'timing/data': 0.0003472600074019283, 'timing/model': 0.39018452100572176, 'learning_rate/qwen_vl_interface': 8.230585569777086e-06, 'learning_rate/action_model': 8.230585569777085e-05, 'epoch': 0.11}) +[2026-10-03 19:13:30,541][starVLA.training.train_starvla][INFO] - Step 1493, Loss: {'train/loss': 0.6984772682189941, 'train/grad_norm_pre_clip': 182.14134216308594, 'timing/data': 0.0003062399919144809, 'timing/model': 0.3916500399936922, 'learning_rate/qwen_vl_interface': 8.228213721384885e-06, 'learning_rate/action_model': 8.228213721384885e-05, 'epoch': 0.11}) +[2026-10-03 19:13:31,239][starVLA.training.train_starvla][INFO] - Step 1494, Loss: {'train/loss': 0.00360680790618062, 'train/grad_norm_pre_clip': 2.3304288387298584, 'timing/data': 0.0003886700142174959, 'timing/model': 0.39044898000429384, 'learning_rate/qwen_vl_interface': 8.225840648761523e-06, 'learning_rate/action_model': 8.225840648761523e-05, 'epoch': 0.11}) +[2026-10-03 19:13:31,935][starVLA.training.train_starvla][INFO] - Step 1495, Loss: {'train/loss': 0.010826430283486843, 'train/grad_norm_pre_clip': 8.169282913208008, 'timing/data': 0.0002612300158943981, 'timing/model': 0.38902413100004196, 'learning_rate/qwen_vl_interface': 8.223466352882476e-06, 'learning_rate/action_model': 8.223466352882475e-05, 'epoch': 0.11}) +[2026-10-03 19:13:32,832][starVLA.training.train_starvla][INFO] - Step 1496, Loss: {'train/loss': 0.4372308850288391, 'train/grad_norm_pre_clip': 473.07269287109375, 'timing/data': 0.00031026999931782484, 'timing/model': 0.5907689999730792, 'learning_rate/qwen_vl_interface': 8.22109083472373e-06, 'learning_rate/action_model': 8.22109083472373e-05, 'epoch': 0.11}) +[2026-10-03 19:13:33,528][starVLA.training.train_starvla][INFO] - Step 1497, Loss: {'train/loss': 0.5965943932533264, 'train/grad_norm_pre_clip': 315.3377380371094, 'timing/data': 0.00020159000996500254, 'timing/model': 0.39074678998440504, 'learning_rate/qwen_vl_interface': 8.218714095261769e-06, 'learning_rate/action_model': 8.218714095261768e-05, 'epoch': 0.11}) +[2026-10-03 19:13:34,227][starVLA.training.train_starvla][INFO] - Step 1498, Loss: {'train/loss': 0.013471598736941814, 'train/grad_norm_pre_clip': 21.99961280822754, 'timing/data': 0.0003363790165167302, 'timing/model': 0.39030607099994086, 'learning_rate/qwen_vl_interface': 8.21633613547358e-06, 'learning_rate/action_model': 8.21633613547358e-05, 'epoch': 0.11}) +[2026-10-03 19:13:34,933][starVLA.training.train_starvla][INFO] - Step 1499, Loss: {'train/loss': 0.0012531657703220844, 'train/grad_norm_pre_clip': 0.15462371706962585, 'timing/data': 0.00040782001451589167, 'timing/model': 0.3929689899960067, 'learning_rate/qwen_vl_interface': 8.213956956336656e-06, 'learning_rate/action_model': 8.213956956336656e-05, 'epoch': 0.11}) +[2026-10-03 19:13:35,632][starVLA.training.train_starvla][INFO] - Step 1500, Loss: {'train/loss': 0.4514576196670532, 'train/grad_norm_pre_clip': 101.47073364257812, 'timing/data': 0.0003668600111268461, 'timing/model': 0.39109975000610575, 'learning_rate/qwen_vl_interface': 8.211576558828985e-06, 'learning_rate/action_model': 8.211576558828985e-05, 'epoch': 0.11}) +[2026-10-03 19:13:36,330][starVLA.training.train_starvla][INFO] - Step 1501, Loss: {'train/loss': 0.5125983953475952, 'train/grad_norm_pre_clip': 56.86886215209961, 'timing/data': 0.00020352000137791038, 'timing/model': 0.39272543002152815, 'learning_rate/qwen_vl_interface': 8.209194943929059e-06, 'learning_rate/action_model': 8.20919494392906e-05, 'epoch': 0.11}) +[2026-10-03 19:13:37,237][starVLA.training.train_starvla][INFO] - Step 1502, Loss: {'train/loss': 0.5313374996185303, 'train/grad_norm_pre_clip': 42.56161880493164, 'timing/data': 0.0002690900000743568, 'timing/model': 0.391060190013377, 'learning_rate/qwen_vl_interface': 8.20681211261587e-06, 'learning_rate/action_model': 8.20681211261587e-05, 'epoch': 0.11}) +[2026-10-03 19:13:37,934][starVLA.training.train_starvla][INFO] - Step 1503, Loss: {'train/loss': 0.0013081978540867567, 'train/grad_norm_pre_clip': 0.43806323409080505, 'timing/data': 0.00020671900711022317, 'timing/model': 0.3911572310025804, 'learning_rate/qwen_vl_interface': 8.204428065868911e-06, 'learning_rate/action_model': 8.204428065868911e-05, 'epoch': 0.11}) +[2026-10-03 19:13:38,637][starVLA.training.train_starvla][INFO] - Step 1504, Loss: {'train/loss': 0.7562647461891174, 'train/grad_norm_pre_clip': 57.08604049682617, 'timing/data': 0.00020100001711398363, 'timing/model': 0.39335018998826854, 'learning_rate/qwen_vl_interface': 8.202042804668173e-06, 'learning_rate/action_model': 8.202042804668172e-05, 'epoch': 0.11}) +[2026-10-03 19:13:39,337][starVLA.training.train_starvla][INFO] - Step 1505, Loss: {'train/loss': 0.3079472780227661, 'train/grad_norm_pre_clip': 36.885040283203125, 'timing/data': 0.0002080600243061781, 'timing/model': 0.3912511600065045, 'learning_rate/qwen_vl_interface': 8.199656329994145e-06, 'learning_rate/action_model': 8.199656329994145e-05, 'epoch': 0.11}) +[2026-10-03 19:13:40,035][starVLA.training.train_starvla][INFO] - Step 1506, Loss: {'train/loss': 0.700697660446167, 'train/grad_norm_pre_clip': 77.3668441772461, 'timing/data': 0.00036251998972147703, 'timing/model': 0.3892239210254047, 'learning_rate/qwen_vl_interface': 8.19726864282782e-06, 'learning_rate/action_model': 8.197268642827818e-05, 'epoch': 0.11}) +[2026-10-03 19:13:40,937][starVLA.training.train_starvla][INFO] - Step 1507, Loss: {'train/loss': 0.27234044671058655, 'train/grad_norm_pre_clip': 80.55599212646484, 'timing/data': 0.0003150799893774092, 'timing/model': 0.5953978599864058, 'learning_rate/qwen_vl_interface': 8.194879744150682e-06, 'learning_rate/action_model': 8.194879744150681e-05, 'epoch': 0.11}) +[2026-10-03 19:13:41,633][starVLA.training.train_starvla][INFO] - Step 1508, Loss: {'train/loss': 1.0356868505477905, 'train/grad_norm_pre_clip': 88.35562896728516, 'timing/data': 0.0002062399871647358, 'timing/model': 0.3892428409890272, 'learning_rate/qwen_vl_interface': 8.192489634944718e-06, 'learning_rate/action_model': 8.192489634944718e-05, 'epoch': 0.11}) +[2026-10-03 19:13:42,332][starVLA.training.train_starvla][INFO] - Step 1509, Loss: {'train/loss': 0.0031390213407576084, 'train/grad_norm_pre_clip': 0.7527622580528259, 'timing/data': 0.0002856299979612231, 'timing/model': 0.39128383999923244, 'learning_rate/qwen_vl_interface': 8.190098316192416e-06, 'learning_rate/action_model': 8.190098316192416e-05, 'epoch': 0.11}) +[2026-10-03 19:13:43,029][starVLA.training.train_starvla][INFO] - Step 1510, Loss: {'train/loss': 0.002443811157718301, 'train/grad_norm_pre_clip': 1.2668898105621338, 'timing/data': 0.00020424998365342617, 'timing/model': 0.3909795109939296, 'learning_rate/qwen_vl_interface': 8.18770578887675e-06, 'learning_rate/action_model': 8.18770578887675e-05, 'epoch': 0.11}) +[2026-10-03 19:13:43,728][starVLA.training.train_starvla][INFO] - Step 1511, Loss: {'train/loss': 0.4035114049911499, 'train/grad_norm_pre_clip': 115.20795440673828, 'timing/data': 0.00046944001223891973, 'timing/model': 0.3910102900117636, 'learning_rate/qwen_vl_interface': 8.185312053981204e-06, 'learning_rate/action_model': 8.185312053981204e-05, 'epoch': 0.11}) +[2026-10-03 19:13:44,426][starVLA.training.train_starvla][INFO] - Step 1512, Loss: {'train/loss': 0.40048056840896606, 'train/grad_norm_pre_clip': 76.95741271972656, 'timing/data': 0.00021055000252090394, 'timing/model': 0.39123812000616454, 'learning_rate/qwen_vl_interface': 8.182917112489747e-06, 'learning_rate/action_model': 8.182917112489746e-05, 'epoch': 0.11}) +[2026-10-03 19:13:45,332][starVLA.training.train_starvla][INFO] - Step 1513, Loss: {'train/loss': 0.04943254217505455, 'train/grad_norm_pre_clip': 42.30460739135742, 'timing/data': 0.0002648700028657913, 'timing/model': 0.3922990999999456, 'learning_rate/qwen_vl_interface': 8.18052096538685e-06, 'learning_rate/action_model': 8.18052096538685e-05, 'epoch': 0.11}) +[2026-10-03 19:13:46,031][starVLA.training.train_starvla][INFO] - Step 1514, Loss: {'train/loss': 0.4326792061328888, 'train/grad_norm_pre_clip': 34.363075256347656, 'timing/data': 0.000357919983798638, 'timing/model': 0.3910097490006592, 'learning_rate/qwen_vl_interface': 8.178123613657483e-06, 'learning_rate/action_model': 8.178123613657482e-05, 'epoch': 0.11}) +[2026-10-03 19:13:46,728][starVLA.training.train_starvla][INFO] - Step 1515, Loss: {'train/loss': 0.7532875537872314, 'train/grad_norm_pre_clip': 90.96290588378906, 'timing/data': 0.0002594800025690347, 'timing/model': 0.39056786097353324, 'learning_rate/qwen_vl_interface': 8.175725058287102e-06, 'learning_rate/action_model': 8.175725058287102e-05, 'epoch': 0.11}) +[2026-10-03 19:13:47,427][starVLA.training.train_starvla][INFO] - Step 1516, Loss: {'train/loss': 0.036590538918972015, 'train/grad_norm_pre_clip': 25.910539627075195, 'timing/data': 0.00021222999203018844, 'timing/model': 0.38969973000348546, 'learning_rate/qwen_vl_interface': 8.173325300261664e-06, 'learning_rate/action_model': 8.173325300261663e-05, 'epoch': 0.11}) +[2026-10-03 19:13:48,127][starVLA.training.train_starvla][INFO] - Step 1517, Loss: {'train/loss': 0.1660229116678238, 'train/grad_norm_pre_clip': 95.68841552734375, 'timing/data': 0.00027792001492343843, 'timing/model': 0.39276626001810655, 'learning_rate/qwen_vl_interface': 8.170924340567617e-06, 'learning_rate/action_model': 8.170924340567616e-05, 'epoch': 0.11}) +[2026-10-03 19:13:49,032][starVLA.training.train_starvla][INFO] - Step 1518, Loss: {'train/loss': 0.004952894523739815, 'train/grad_norm_pre_clip': 4.453093528747559, 'timing/data': 0.0004360000020824373, 'timing/model': 0.39038160000927746, 'learning_rate/qwen_vl_interface': 8.168522180191909e-06, 'learning_rate/action_model': 8.168522180191909e-05, 'epoch': 0.11}) +[2026-10-03 19:13:49,729][starVLA.training.train_starvla][INFO] - Step 1519, Loss: {'train/loss': 0.20746871829032898, 'train/grad_norm_pre_clip': 35.4145393371582, 'timing/data': 0.0003355790104251355, 'timing/model': 0.38942127098562196, 'learning_rate/qwen_vl_interface': 8.166118820121972e-06, 'learning_rate/action_model': 8.166118820121973e-05, 'epoch': 0.11}) +[2026-10-03 19:13:50,429][starVLA.training.train_starvla][INFO] - Step 1520, Loss: {'train/loss': 0.054094098508358, 'train/grad_norm_pre_clip': 14.221867561340332, 'timing/data': 0.0002677899901755154, 'timing/model': 0.39130183999077417, 'learning_rate/qwen_vl_interface': 8.163714261345744e-06, 'learning_rate/action_model': 8.163714261345742e-05, 'epoch': 0.11}) +[2026-10-03 19:13:51,131][starVLA.training.train_starvla][INFO] - Step 1521, Loss: {'train/loss': 0.48151397705078125, 'train/grad_norm_pre_clip': 43.16963195800781, 'timing/data': 0.0002959800185635686, 'timing/model': 0.39330501001677476, 'learning_rate/qwen_vl_interface': 8.16130850485164e-06, 'learning_rate/action_model': 8.161308504851639e-05, 'epoch': 0.11}) +[2026-10-03 19:13:51,828][starVLA.training.train_starvla][INFO] - Step 1522, Loss: {'train/loss': 0.3374789357185364, 'train/grad_norm_pre_clip': 101.50414276123047, 'timing/data': 0.00037240999517962337, 'timing/model': 0.39038209000136703, 'learning_rate/qwen_vl_interface': 8.158901551628579e-06, 'learning_rate/action_model': 8.158901551628579e-05, 'epoch': 0.11}) +[2026-10-03 19:13:52,732][starVLA.training.train_starvla][INFO] - Step 1523, Loss: {'train/loss': 0.014416954480111599, 'train/grad_norm_pre_clip': 11.984870910644531, 'timing/data': 0.00024016000679694116, 'timing/model': 0.5977623779908754, 'learning_rate/qwen_vl_interface': 8.156493402665971e-06, 'learning_rate/action_model': 8.15649340266597e-05, 'epoch': 0.11}) +[2026-10-03 19:13:53,432][starVLA.training.train_starvla][INFO] - Step 1524, Loss: {'train/loss': 0.1272907704114914, 'train/grad_norm_pre_clip': 146.47222900390625, 'timing/data': 0.00031889998354017735, 'timing/model': 0.3922939600015525, 'learning_rate/qwen_vl_interface': 8.154084058953713e-06, 'learning_rate/action_model': 8.154084058953713e-05, 'epoch': 0.11}) +[2026-10-03 19:13:54,133][starVLA.training.train_starvla][INFO] - Step 1525, Loss: {'train/loss': 0.09179887920618057, 'train/grad_norm_pre_clip': 82.70061492919922, 'timing/data': 0.00028214999474585056, 'timing/model': 0.3919672899937723, 'learning_rate/qwen_vl_interface': 8.151673521482193e-06, 'learning_rate/action_model': 8.151673521482194e-05, 'epoch': 0.11}) +[2026-10-03 19:13:54,833][starVLA.training.train_starvla][INFO] - Step 1526, Loss: {'train/loss': 0.17856086790561676, 'train/grad_norm_pre_clip': 16.574365615844727, 'timing/data': 0.0003383599978405982, 'timing/model': 0.3906565800134558, 'learning_rate/qwen_vl_interface': 8.149261791242297e-06, 'learning_rate/action_model': 8.149261791242297e-05, 'epoch': 0.11}) +[2026-10-03 19:13:55,533][starVLA.training.train_starvla][INFO] - Step 1527, Loss: {'train/loss': 0.4566539227962494, 'train/grad_norm_pre_clip': 17.168851852416992, 'timing/data': 0.00025849000667221844, 'timing/model': 0.39094480001949705, 'learning_rate/qwen_vl_interface': 8.146848869225391e-06, 'learning_rate/action_model': 8.146848869225391e-05, 'epoch': 0.11}) +[2026-10-03 19:13:56,235][starVLA.training.train_starvla][INFO] - Step 1528, Loss: {'train/loss': 0.002992910798639059, 'train/grad_norm_pre_clip': 0.34666162729263306, 'timing/data': 0.00028709997422993183, 'timing/model': 0.3920267000212334, 'learning_rate/qwen_vl_interface': 8.14443475642334e-06, 'learning_rate/action_model': 8.144434756423339e-05, 'epoch': 0.11}) +[2026-10-03 19:13:57,136][starVLA.training.train_starvla][INFO] - Step 1529, Loss: {'train/loss': 0.008910742588341236, 'train/grad_norm_pre_clip': 3.1634135246276855, 'timing/data': 0.00032264000037685037, 'timing/model': 0.3937253199983388, 'learning_rate/qwen_vl_interface': 8.142019453828493e-06, 'learning_rate/action_model': 8.142019453828493e-05, 'epoch': 0.11}) +[2026-10-03 19:13:57,833][starVLA.training.train_starvla][INFO] - Step 1530, Loss: {'train/loss': 0.05944710597395897, 'train/grad_norm_pre_clip': 22.009296417236328, 'timing/data': 0.0003875299880746752, 'timing/model': 0.38980200997320935, 'learning_rate/qwen_vl_interface': 8.139602962433688e-06, 'learning_rate/action_model': 8.139602962433688e-05, 'epoch': 0.11}) +[2026-10-03 19:13:58,530][starVLA.training.train_starvla][INFO] - Step 1531, Loss: {'train/loss': 0.0063255238346755505, 'train/grad_norm_pre_clip': 0.832542359828949, 'timing/data': 0.00022933000582270324, 'timing/model': 0.39059908999479376, 'learning_rate/qwen_vl_interface': 8.137185283232255e-06, 'learning_rate/action_model': 8.137185283232256e-05, 'epoch': 0.11}) +[2026-10-03 19:13:59,225][starVLA.training.train_starvla][INFO] - Step 1532, Loss: {'train/loss': 0.010430918075144291, 'train/grad_norm_pre_clip': 1.2931636571884155, 'timing/data': 0.0002067199966404587, 'timing/model': 0.38892805099021643, 'learning_rate/qwen_vl_interface': 8.134766417218014e-06, 'learning_rate/action_model': 8.134766417218012e-05, 'epoch': 0.11}) +[2026-10-03 19:13:59,923][starVLA.training.train_starvla][INFO] - Step 1533, Loss: {'train/loss': 0.3168831467628479, 'train/grad_norm_pre_clip': 53.97322082519531, 'timing/data': 0.0003147699753753841, 'timing/model': 0.3919376500125509, 'learning_rate/qwen_vl_interface': 8.132346365385262e-06, 'learning_rate/action_model': 8.132346365385261e-05, 'epoch': 0.11}) +[2026-10-03 19:14:00,817][starVLA.training.train_starvla][INFO] - Step 1534, Loss: {'train/loss': 0.0076308827847242355, 'train/grad_norm_pre_clip': 1.3976373672485352, 'timing/data': 0.00023585002054460347, 'timing/model': 0.3881564809998963, 'learning_rate/qwen_vl_interface': 8.129925128728796e-06, 'learning_rate/action_model': 8.129925128728796e-05, 'epoch': 0.11}) +[2026-10-03 19:14:01,513][starVLA.training.train_starvla][INFO] - Step 1535, Loss: {'train/loss': 0.0620955154299736, 'train/grad_norm_pre_clip': 58.16193389892578, 'timing/data': 0.00023298998712562025, 'timing/model': 0.38867949100676924, 'learning_rate/qwen_vl_interface': 8.127502708243895e-06, 'learning_rate/action_model': 8.127502708243895e-05, 'epoch': 0.11}) +[2026-10-03 19:14:02,210][starVLA.training.train_starvla][INFO] - Step 1536, Loss: {'train/loss': 0.014158264733850956, 'train/grad_norm_pre_clip': 7.342172622680664, 'timing/data': 0.00022906900267116725, 'timing/model': 0.38904926102259196, 'learning_rate/qwen_vl_interface': 8.125079104926324e-06, 'learning_rate/action_model': 8.125079104926323e-05, 'epoch': 0.11}) +[2026-10-03 19:14:02,906][starVLA.training.train_starvla][INFO] - Step 1537, Loss: {'train/loss': 0.7177304625511169, 'train/grad_norm_pre_clip': 36.84468078613281, 'timing/data': 0.00020369997946545482, 'timing/model': 0.39182658001664095, 'learning_rate/qwen_vl_interface': 8.122654319772332e-06, 'learning_rate/action_model': 8.122654319772332e-05, 'epoch': 0.11}) +[2026-10-03 19:14:03,602][starVLA.training.train_starvla][INFO] - Step 1538, Loss: {'train/loss': 0.27594298124313354, 'train/grad_norm_pre_clip': 41.131080627441406, 'timing/data': 0.00019936999888159335, 'timing/model': 0.3908745000080671, 'learning_rate/qwen_vl_interface': 8.120228353778661e-06, 'learning_rate/action_model': 8.12022835377866e-05, 'epoch': 0.11}) +[2026-10-03 19:14:04,502][starVLA.training.train_starvla][INFO] - Step 1539, Loss: {'train/loss': 0.00234434986487031, 'train/grad_norm_pre_clip': 0.7882985472679138, 'timing/data': 0.0003438799758441746, 'timing/model': 0.5930873489996884, 'learning_rate/qwen_vl_interface': 8.117801207942528e-06, 'learning_rate/action_model': 8.117801207942528e-05, 'epoch': 0.11}) +[2026-10-03 19:14:05,201][starVLA.training.train_starvla][INFO] - Step 1540, Loss: {'train/loss': 0.003573998576030135, 'train/grad_norm_pre_clip': 1.0661561489105225, 'timing/data': 0.0002032799820881337, 'timing/model': 0.3912266799889039, 'learning_rate/qwen_vl_interface': 8.115372883261648e-06, 'learning_rate/action_model': 8.115372883261647e-05, 'epoch': 0.11}) +[2026-10-03 19:14:05,900][starVLA.training.train_starvla][INFO] - Step 1541, Loss: {'train/loss': 0.4484400451183319, 'train/grad_norm_pre_clip': 18.023096084594727, 'timing/data': 0.0002422299876343459, 'timing/model': 0.3929349899990484, 'learning_rate/qwen_vl_interface': 8.11294338073421e-06, 'learning_rate/action_model': 8.11294338073421e-05, 'epoch': 0.11}) +[2026-10-03 19:14:06,598][starVLA.training.train_starvla][INFO] - Step 1542, Loss: {'train/loss': 0.7615936994552612, 'train/grad_norm_pre_clip': 65.68456268310547, 'timing/data': 0.00021485998877324164, 'timing/model': 0.3900880209985189, 'learning_rate/qwen_vl_interface': 8.110512701358892e-06, 'learning_rate/action_model': 8.110512701358891e-05, 'epoch': 0.11}) +[2026-10-03 19:14:07,295][starVLA.training.train_starvla][INFO] - Step 1543, Loss: {'train/loss': 0.11350958049297333, 'train/grad_norm_pre_clip': 64.64704132080078, 'timing/data': 0.00024147998192347586, 'timing/model': 0.3908712089760229, 'learning_rate/qwen_vl_interface': 8.108080846134852e-06, 'learning_rate/action_model': 8.108080846134852e-05, 'epoch': 0.11}) +[2026-10-03 19:14:08,196][starVLA.training.train_starvla][INFO] - Step 1544, Loss: {'train/loss': 0.5421085357666016, 'train/grad_norm_pre_clip': 131.4855194091797, 'timing/data': 0.00022976001491770148, 'timing/model': 0.5931512399984058, 'learning_rate/qwen_vl_interface': 8.105647816061734e-06, 'learning_rate/action_model': 8.105647816061735e-05, 'epoch': 0.11}) +[2026-10-03 19:14:08,895][starVLA.training.train_starvla][INFO] - Step 1545, Loss: {'train/loss': 0.497123658657074, 'train/grad_norm_pre_clip': 62.49565124511719, 'timing/data': 0.0002140600117854774, 'timing/model': 0.39122826998936944, 'learning_rate/qwen_vl_interface': 8.103213612139667e-06, 'learning_rate/action_model': 8.103213612139667e-05, 'epoch': 0.11}) +[2026-10-03 19:14:09,591][starVLA.training.train_starvla][INFO] - Step 1546, Loss: {'train/loss': 0.22732850909233093, 'train/grad_norm_pre_clip': 55.54789352416992, 'timing/data': 0.0003268499858677387, 'timing/model': 0.3906630700221285, 'learning_rate/qwen_vl_interface': 8.100778235369262e-06, 'learning_rate/action_model': 8.10077823536926e-05, 'epoch': 0.11}) +[2026-10-03 19:14:10,287][starVLA.training.train_starvla][INFO] - Step 1547, Loss: {'train/loss': 0.011885630898177624, 'train/grad_norm_pre_clip': 4.545822620391846, 'timing/data': 0.0002388299908488989, 'timing/model': 0.3899566309992224, 'learning_rate/qwen_vl_interface': 8.098341686751605e-06, 'learning_rate/action_model': 8.098341686751605e-05, 'epoch': 0.11}) +[2026-10-03 19:14:10,986][starVLA.training.train_starvla][INFO] - Step 1548, Loss: {'train/loss': 0.010059862397611141, 'train/grad_norm_pre_clip': 4.391934871673584, 'timing/data': 0.0003291300090495497, 'timing/model': 0.39038311000331305, 'learning_rate/qwen_vl_interface': 8.095903967288273e-06, 'learning_rate/action_model': 8.095903967288272e-05, 'epoch': 0.11}) +[2026-10-03 19:14:11,892][starVLA.training.train_starvla][INFO] - Step 1549, Loss: {'train/loss': 0.648190975189209, 'train/grad_norm_pre_clip': 43.199867248535156, 'timing/data': 0.00032479999936185777, 'timing/model': 0.5987953500007279, 'learning_rate/qwen_vl_interface': 8.093465077981318e-06, 'learning_rate/action_model': 8.093465077981318e-05, 'epoch': 0.11}) +[2026-10-03 19:14:12,588][starVLA.training.train_starvla][INFO] - Step 1550, Loss: {'train/loss': 0.6033987402915955, 'train/grad_norm_pre_clip': 22.987789154052734, 'timing/data': 0.0003361400158610195, 'timing/model': 0.38929506999556907, 'learning_rate/qwen_vl_interface': 8.091025019833276e-06, 'learning_rate/action_model': 8.091025019833276e-05, 'epoch': 0.11}) +[2026-10-03 19:14:13,286][starVLA.training.train_starvla][INFO] - Step 1551, Loss: {'train/loss': 0.28262558579444885, 'train/grad_norm_pre_clip': 172.12796020507812, 'timing/data': 0.0002607999776955694, 'timing/model': 0.39013426000019535, 'learning_rate/qwen_vl_interface': 8.088583793847165e-06, 'learning_rate/action_model': 8.088583793847164e-05, 'epoch': 0.11}) +[2026-10-03 19:14:13,984][starVLA.training.train_starvla][INFO] - Step 1552, Loss: {'train/loss': 0.04144534096121788, 'train/grad_norm_pre_clip': 45.01343536376953, 'timing/data': 0.0002668699889909476, 'timing/model': 0.39033600100083277, 'learning_rate/qwen_vl_interface': 8.086141401026478e-06, 'learning_rate/action_model': 8.086141401026478e-05, 'epoch': 0.11}) +[2026-10-03 19:14:14,683][starVLA.training.train_starvla][INFO] - Step 1553, Loss: {'train/loss': 0.07584425061941147, 'train/grad_norm_pre_clip': 42.100284576416016, 'timing/data': 0.00030614997376687825, 'timing/model': 0.3915570499957539, 'learning_rate/qwen_vl_interface': 8.083697842375193e-06, 'learning_rate/action_model': 8.083697842375193e-05, 'epoch': 0.11}) +[2026-10-03 19:14:15,379][starVLA.training.train_starvla][INFO] - Step 1554, Loss: {'train/loss': 0.0008572254446335137, 'train/grad_norm_pre_clip': 0.3369573652744293, 'timing/data': 0.0004721999866887927, 'timing/model': 0.3907885200169403, 'learning_rate/qwen_vl_interface': 8.081253118897762e-06, 'learning_rate/action_model': 8.081253118897761e-05, 'epoch': 0.11}) +[2026-10-03 19:14:16,284][starVLA.training.train_starvla][INFO] - Step 1555, Loss: {'train/loss': 0.00414181686937809, 'train/grad_norm_pre_clip': 0.7051767110824585, 'timing/data': 0.0002539300185162574, 'timing/model': 0.3904558609938249, 'learning_rate/qwen_vl_interface': 8.07880723159912e-06, 'learning_rate/action_model': 8.07880723159912e-05, 'epoch': 0.11}) +[2026-10-03 19:14:16,982][starVLA.training.train_starvla][INFO] - Step 1556, Loss: {'train/loss': 0.0023151931818574667, 'train/grad_norm_pre_clip': 0.8347938656806946, 'timing/data': 0.00029532998451031744, 'timing/model': 0.38904735998949036, 'learning_rate/qwen_vl_interface': 8.076360181484676e-06, 'learning_rate/action_model': 8.076360181484676e-05, 'epoch': 0.11}) +[2026-10-03 19:14:17,683][starVLA.training.train_starvla][INFO] - Step 1557, Loss: {'train/loss': 0.27383291721343994, 'train/grad_norm_pre_clip': 102.538330078125, 'timing/data': 0.00034046999644488096, 'timing/model': 0.3933562299935147, 'learning_rate/qwen_vl_interface': 8.073911969560326e-06, 'learning_rate/action_model': 8.073911969560326e-05, 'epoch': 0.11}) +[2026-10-03 19:14:18,382][starVLA.training.train_starvla][INFO] - Step 1558, Loss: {'train/loss': 0.00046948814997449517, 'train/grad_norm_pre_clip': 0.20081326365470886, 'timing/data': 0.00032411000574938953, 'timing/model': 0.38905520999105647, 'learning_rate/qwen_vl_interface': 8.071462596832432e-06, 'learning_rate/action_model': 8.071462596832432e-05, 'epoch': 0.11}) +[2026-10-03 19:14:19,080][starVLA.training.train_starvla][INFO] - Step 1559, Loss: {'train/loss': 0.08467298746109009, 'train/grad_norm_pre_clip': 172.77220153808594, 'timing/data': 0.00022575000184588134, 'timing/model': 0.39101111999480054, 'learning_rate/qwen_vl_interface': 8.06901206430784e-06, 'learning_rate/action_model': 8.06901206430784e-05, 'epoch': 0.11}) +[2026-10-03 19:14:19,981][starVLA.training.train_starvla][INFO] - Step 1560, Loss: {'train/loss': 0.30327489972114563, 'train/grad_norm_pre_clip': 79.1749038696289, 'timing/data': 0.00032943999394774437, 'timing/model': 0.5919297600048594, 'learning_rate/qwen_vl_interface': 8.066560372993873e-06, 'learning_rate/action_model': 8.066560372993874e-05, 'epoch': 0.11}) +[2026-10-03 19:14:20,680][starVLA.training.train_starvla][INFO] - Step 1561, Loss: {'train/loss': 0.990275502204895, 'train/grad_norm_pre_clip': 114.14178466796875, 'timing/data': 0.00029953999910503626, 'timing/model': 0.3914415600011125, 'learning_rate/qwen_vl_interface': 8.064107523898327e-06, 'learning_rate/action_model': 8.064107523898326e-05, 'epoch': 0.11}) +[2026-10-03 19:14:21,377][starVLA.training.train_starvla][INFO] - Step 1562, Loss: {'train/loss': 0.5870308876037598, 'train/grad_norm_pre_clip': 33.37674331665039, 'timing/data': 0.0003030299849342555, 'timing/model': 0.38910591998137534, 'learning_rate/qwen_vl_interface': 8.061653518029476e-06, 'learning_rate/action_model': 8.061653518029476e-05, 'epoch': 0.11}) +[2026-10-03 19:14:22,077][starVLA.training.train_starvla][INFO] - Step 1563, Loss: {'train/loss': 0.005136486608535051, 'train/grad_norm_pre_clip': 1.31513249874115, 'timing/data': 0.00029848000849597156, 'timing/model': 0.39177431000280194, 'learning_rate/qwen_vl_interface': 8.059198356396069e-06, 'learning_rate/action_model': 8.059198356396068e-05, 'epoch': 0.11}) +[2026-10-03 19:14:22,776][starVLA.training.train_starvla][INFO] - Step 1564, Loss: {'train/loss': 0.3189639449119568, 'train/grad_norm_pre_clip': 167.46702575683594, 'timing/data': 0.00020743999630212784, 'timing/model': 0.3898939109931234, 'learning_rate/qwen_vl_interface': 8.05674204000733e-06, 'learning_rate/action_model': 8.05674204000733e-05, 'epoch': 0.11}) +[2026-10-03 19:14:23,474][starVLA.training.train_starvla][INFO] - Step 1565, Loss: {'train/loss': 0.0014975897502154112, 'train/grad_norm_pre_clip': 0.2329343557357788, 'timing/data': 0.0002100899873767048, 'timing/model': 0.3926622200233396, 'learning_rate/qwen_vl_interface': 8.05428456987296e-06, 'learning_rate/action_model': 8.05428456987296e-05, 'epoch': 0.11}) +[2026-10-03 19:14:24,377][starVLA.training.train_starvla][INFO] - Step 1566, Loss: {'train/loss': 0.0019136385526508093, 'train/grad_norm_pre_clip': 1.7777822017669678, 'timing/data': 0.0002155900001525879, 'timing/model': 0.390961850003805, 'learning_rate/qwen_vl_interface': 8.05182594700313e-06, 'learning_rate/action_model': 8.05182594700313e-05, 'epoch': 0.11}) +[2026-10-03 19:14:25,078][starVLA.training.train_starvla][INFO] - Step 1567, Loss: {'train/loss': 0.986414909362793, 'train/grad_norm_pre_clip': 83.1103286743164, 'timing/data': 0.00032376000308431685, 'timing/model': 0.3913600399973802, 'learning_rate/qwen_vl_interface': 8.049366172408484e-06, 'learning_rate/action_model': 8.049366172408485e-05, 'epoch': 0.11}) +[2026-10-03 19:14:25,775][starVLA.training.train_starvla][INFO] - Step 1568, Loss: {'train/loss': 0.5039783120155334, 'train/grad_norm_pre_clip': 64.33659362792969, 'timing/data': 0.00024064001627266407, 'timing/model': 0.3898269000055734, 'learning_rate/qwen_vl_interface': 8.04690524710015e-06, 'learning_rate/action_model': 8.046905247100151e-05, 'epoch': 0.11}) +[2026-10-03 19:14:26,475][starVLA.training.train_starvla][INFO] - Step 1569, Loss: {'train/loss': 0.01836434379220009, 'train/grad_norm_pre_clip': 22.425128936767578, 'timing/data': 0.00030126000638119876, 'timing/model': 0.392596229998162, 'learning_rate/qwen_vl_interface': 8.044443172089715e-06, 'learning_rate/action_model': 8.044443172089714e-05, 'epoch': 0.11}) +[2026-10-03 19:14:27,175][starVLA.training.train_starvla][INFO] - Step 1570, Loss: {'train/loss': 0.1796511709690094, 'train/grad_norm_pre_clip': 60.64904022216797, 'timing/data': 0.00034324999433010817, 'timing/model': 0.39124561098287813, 'learning_rate/qwen_vl_interface': 8.041979948389248e-06, 'learning_rate/action_model': 8.041979948389248e-05, 'epoch': 0.11}) +[2026-10-03 19:14:28,079][starVLA.training.train_starvla][INFO] - Step 1571, Loss: {'train/loss': 0.28810638189315796, 'train/grad_norm_pre_clip': 23.334163665771484, 'timing/data': 0.00042239000322297215, 'timing/model': 0.3896434610069264, 'learning_rate/qwen_vl_interface': 8.039515577011287e-06, 'learning_rate/action_model': 8.039515577011288e-05, 'epoch': 0.11}) +[2026-10-03 19:14:28,779][starVLA.training.train_starvla][INFO] - Step 1572, Loss: {'train/loss': 0.0011767063988372684, 'train/grad_norm_pre_clip': 0.2632234990596771, 'timing/data': 0.00021718998323194683, 'timing/model': 0.39186425899970345, 'learning_rate/qwen_vl_interface': 8.037050058968841e-06, 'learning_rate/action_model': 8.03705005896884e-05, 'epoch': 0.12}) +[2026-10-03 19:14:29,476][starVLA.training.train_starvla][INFO] - Step 1573, Loss: {'train/loss': 0.3902508318424225, 'train/grad_norm_pre_clip': 31.47641944885254, 'timing/data': 0.00020286999642848969, 'timing/model': 0.3907206400181167, 'learning_rate/qwen_vl_interface': 8.034583395275392e-06, 'learning_rate/action_model': 8.034583395275392e-05, 'epoch': 0.12}) +[2026-10-03 19:14:30,149][starVLA.training.train_starvla][INFO] - Step 1574, Loss: {'train/loss': 0.011993108317255974, 'train/grad_norm_pre_clip': 11.370100021362305, 'timing/data': 0.00021513999672606587, 'timing/model': 0.3654043229762465, 'learning_rate/qwen_vl_interface': 8.032115586944893e-06, 'learning_rate/action_model': 8.032115586944892e-05, 'epoch': 0.12}) +[2026-10-03 19:14:30,845][starVLA.training.train_starvla][INFO] - Step 1575, Loss: {'train/loss': 0.003557062242180109, 'train/grad_norm_pre_clip': 2.6771771907806396, 'timing/data': 0.00021345997811295092, 'timing/model': 0.39110480999806896, 'learning_rate/qwen_vl_interface': 8.029646634991762e-06, 'learning_rate/action_model': 8.029646634991762e-05, 'epoch': 0.12}) +[2026-10-03 19:14:31,749][starVLA.training.train_starvla][INFO] - Step 1576, Loss: {'train/loss': 0.0018181806663051248, 'train/grad_norm_pre_clip': 0.29429247975349426, 'timing/data': 0.0004465200181584805, 'timing/model': 0.5955909089825582, 'learning_rate/qwen_vl_interface': 8.0271765404309e-06, 'learning_rate/action_model': 8.0271765404309e-05, 'epoch': 0.12}) +[2026-10-03 19:14:32,449][starVLA.training.train_starvla][INFO] - Step 1577, Loss: {'train/loss': 0.7443780303001404, 'train/grad_norm_pre_clip': 34.1747932434082, 'timing/data': 0.00023453999892808497, 'timing/model': 0.3926282199972775, 'learning_rate/qwen_vl_interface': 8.024705304277664e-06, 'learning_rate/action_model': 8.024705304277664e-05, 'epoch': 0.12}) +[2026-10-03 19:14:33,149][starVLA.training.train_starvla][INFO] - Step 1578, Loss: {'train/loss': 0.08027156442403793, 'train/grad_norm_pre_clip': 21.973155975341797, 'timing/data': 0.00032454001484438777, 'timing/model': 0.39159563000430353, 'learning_rate/qwen_vl_interface': 8.022232927547888e-06, 'learning_rate/action_model': 8.022232927547886e-05, 'epoch': 0.12}) +[2026-10-03 19:14:33,848][starVLA.training.train_starvla][INFO] - Step 1579, Loss: {'train/loss': 0.024711504578590393, 'train/grad_norm_pre_clip': 42.195770263671875, 'timing/data': 0.0002604799810796976, 'timing/model': 0.39076831997954287, 'learning_rate/qwen_vl_interface': 8.01975941125787e-06, 'learning_rate/action_model': 8.01975941125787e-05, 'epoch': 0.12}) +[2026-10-03 19:14:34,548][starVLA.training.train_starvla][INFO] - Step 1580, Loss: {'train/loss': 0.012945758178830147, 'train/grad_norm_pre_clip': 6.418487071990967, 'timing/data': 0.000248589989496395, 'timing/model': 0.3916072699939832, 'learning_rate/qwen_vl_interface': 8.017284756424381e-06, 'learning_rate/action_model': 8.017284756424381e-05, 'epoch': 0.12}) +[2026-10-03 19:14:35,250][starVLA.training.train_starvla][INFO] - Step 1581, Loss: {'train/loss': 0.19110944867134094, 'train/grad_norm_pre_clip': 23.501001358032227, 'timing/data': 0.00022829999215900898, 'timing/model': 0.392316400015261, 'learning_rate/qwen_vl_interface': 8.01480896406466e-06, 'learning_rate/action_model': 8.01480896406466e-05, 'epoch': 0.12}) +[2026-10-03 19:14:36,157][starVLA.training.train_starvla][INFO] - Step 1582, Loss: {'train/loss': 0.5743089914321899, 'train/grad_norm_pre_clip': 25.789031982421875, 'timing/data': 0.000273999001365155, 'timing/model': 0.3907642609847244, 'learning_rate/qwen_vl_interface': 8.01233203519641e-06, 'learning_rate/action_model': 8.012332035196409e-05, 'epoch': 0.12}) +[2026-10-03 19:14:36,856][starVLA.training.train_starvla][INFO] - Step 1583, Loss: {'train/loss': 0.07086498290300369, 'train/grad_norm_pre_clip': 101.44723510742188, 'timing/data': 0.00023747998056933284, 'timing/model': 0.39211623001028784, 'learning_rate/qwen_vl_interface': 8.009853970837801e-06, 'learning_rate/action_model': 8.009853970837801e-05, 'epoch': 0.12}) +[2026-10-03 19:14:37,555][starVLA.training.train_starvla][INFO] - Step 1584, Loss: {'train/loss': 0.4045727550983429, 'train/grad_norm_pre_clip': 143.02793884277344, 'timing/data': 0.000322499981848523, 'timing/model': 0.39186326100025326, 'learning_rate/qwen_vl_interface': 8.007374772007474e-06, 'learning_rate/action_model': 8.007374772007474e-05, 'epoch': 0.12}) +[2026-10-03 19:14:38,256][starVLA.training.train_starvla][INFO] - Step 1585, Loss: {'train/loss': 0.022431282326579094, 'train/grad_norm_pre_clip': 4.653777599334717, 'timing/data': 0.00023904998670332134, 'timing/model': 0.39271252902108245, 'learning_rate/qwen_vl_interface': 8.004894439724533e-06, 'learning_rate/action_model': 8.004894439724533e-05, 'epoch': 0.12}) +[2026-10-03 19:14:38,956][starVLA.training.train_starvla][INFO] - Step 1586, Loss: {'train/loss': 0.7480354905128479, 'train/grad_norm_pre_clip': 53.8765983581543, 'timing/data': 0.00040422999882139266, 'timing/model': 0.39176743998541497, 'learning_rate/qwen_vl_interface': 8.002412975008552e-06, 'learning_rate/action_model': 8.002412975008552e-05, 'epoch': 0.12}) +[2026-10-03 19:14:39,857][starVLA.training.train_starvla][INFO] - Step 1587, Loss: {'train/loss': 0.22150738537311554, 'train/grad_norm_pre_clip': 55.240875244140625, 'timing/data': 0.0002565199974924326, 'timing/model': 0.5921851400053129, 'learning_rate/qwen_vl_interface': 7.999930378879562e-06, 'learning_rate/action_model': 7.999930378879561e-05, 'epoch': 0.12}) +[2026-10-03 19:14:40,556][starVLA.training.train_starvla][INFO] - Step 1588, Loss: {'train/loss': 0.008786787278950214, 'train/grad_norm_pre_clip': 1.5398181676864624, 'timing/data': 0.0002053700154647231, 'timing/model': 0.39160045998869464, 'learning_rate/qwen_vl_interface': 7.997446652358067e-06, 'learning_rate/action_model': 7.997446652358068e-05, 'epoch': 0.12}) +[2026-10-03 19:14:41,256][starVLA.training.train_starvla][INFO] - Step 1589, Loss: {'train/loss': 0.05689522624015808, 'train/grad_norm_pre_clip': 22.826799392700195, 'timing/data': 0.0003365600132383406, 'timing/model': 0.39257651998195797, 'learning_rate/qwen_vl_interface': 7.994961796465035e-06, 'learning_rate/action_model': 7.994961796465035e-05, 'epoch': 0.12}) +[2026-10-03 19:14:41,954][starVLA.training.train_starvla][INFO] - Step 1590, Loss: {'train/loss': 0.013645443134009838, 'train/grad_norm_pre_clip': 6.556158542633057, 'timing/data': 0.0003683399991132319, 'timing/model': 0.3911592210060917, 'learning_rate/qwen_vl_interface': 7.992475812221894e-06, 'learning_rate/action_model': 7.992475812221894e-05, 'epoch': 0.12}) +[2026-10-03 19:14:42,654][starVLA.training.train_starvla][INFO] - Step 1591, Loss: {'train/loss': 0.01961921527981758, 'train/grad_norm_pre_clip': 14.510953903198242, 'timing/data': 0.00030638001044280827, 'timing/model': 0.39166242998908274, 'learning_rate/qwen_vl_interface': 7.989988700650538e-06, 'learning_rate/action_model': 7.989988700650538e-05, 'epoch': 0.12}) +[2026-10-03 19:14:43,550][starVLA.training.train_starvla][INFO] - Step 1592, Loss: {'train/loss': 0.06184898689389229, 'train/grad_norm_pre_clip': 25.341644287109375, 'timing/data': 0.0003421599976718426, 'timing/model': 0.5889144199900329, 'learning_rate/qwen_vl_interface': 7.987500462773326e-06, 'learning_rate/action_model': 7.987500462773326e-05, 'epoch': 0.12}) +[2026-10-03 19:14:44,248][starVLA.training.train_starvla][INFO] - Step 1593, Loss: {'train/loss': 0.003368678968399763, 'train/grad_norm_pre_clip': 0.7830829620361328, 'timing/data': 0.00024279998615384102, 'timing/model': 0.3915404500148725, 'learning_rate/qwen_vl_interface': 7.985011099613078e-06, 'learning_rate/action_model': 7.985011099613077e-05, 'epoch': 0.12}) +[2026-10-03 19:14:44,944][starVLA.training.train_starvla][INFO] - Step 1594, Loss: {'train/loss': 0.001981351524591446, 'train/grad_norm_pre_clip': 1.6645948886871338, 'timing/data': 0.0003846299950964749, 'timing/model': 0.3884291000140365, 'learning_rate/qwen_vl_interface': 7.98252061219308e-06, 'learning_rate/action_model': 7.982520612193078e-05, 'epoch': 0.12}) +[2026-10-03 19:14:45,638][starVLA.training.train_starvla][INFO] - Step 1595, Loss: {'train/loss': 0.003315265756100416, 'train/grad_norm_pre_clip': 6.809188365936279, 'timing/data': 0.000258120009675622, 'timing/model': 0.38871974099311046, 'learning_rate/qwen_vl_interface': 7.980029001537073e-06, 'learning_rate/action_model': 7.980029001537072e-05, 'epoch': 0.12}) +[2026-10-03 19:14:46,331][starVLA.training.train_starvla][INFO] - Step 1596, Loss: {'train/loss': 0.6226426362991333, 'train/grad_norm_pre_clip': 25.22583770751953, 'timing/data': 0.0002421200042590499, 'timing/model': 0.3888800799904857, 'learning_rate/qwen_vl_interface': 7.977536268669264e-06, 'learning_rate/action_model': 7.977536268669264e-05, 'epoch': 0.12}) +[2026-10-03 19:14:47,026][starVLA.training.train_starvla][INFO] - Step 1597, Loss: {'train/loss': 0.5171301364898682, 'train/grad_norm_pre_clip': 239.48561096191406, 'timing/data': 0.00023624999448657036, 'timing/model': 0.38998703999095596, 'learning_rate/qwen_vl_interface': 7.975042414614324e-06, 'learning_rate/action_model': 7.975042414614324e-05, 'epoch': 0.12}) +[2026-10-03 19:14:47,918][starVLA.training.train_starvla][INFO] - Step 1598, Loss: {'train/loss': 0.19862641394138336, 'train/grad_norm_pre_clip': 86.52960205078125, 'timing/data': 0.0005051300104241818, 'timing/model': 0.38949788999161683, 'learning_rate/qwen_vl_interface': 7.972547440397382e-06, 'learning_rate/action_model': 7.972547440397383e-05, 'epoch': 0.12}) +[2026-10-03 19:14:48,614][starVLA.training.train_starvla][INFO] - Step 1599, Loss: {'train/loss': 0.006388127338141203, 'train/grad_norm_pre_clip': 15.794270515441895, 'timing/data': 0.00037552000139839947, 'timing/model': 0.3900720390083734, 'learning_rate/qwen_vl_interface': 7.970051347044026e-06, 'learning_rate/action_model': 7.970051347044025e-05, 'epoch': 0.12}) +[2026-10-03 19:14:49,313][starVLA.training.train_starvla][INFO] - Step 1600, Loss: {'train/loss': 0.2670672535896301, 'train/grad_norm_pre_clip': 83.08836364746094, 'timing/data': 0.0003972999984398484, 'timing/model': 0.3910538099880796, 'learning_rate/qwen_vl_interface': 7.967554135580307e-06, 'learning_rate/action_model': 7.967554135580307e-05, 'epoch': 0.12}) +[2026-10-03 19:14:50,012][starVLA.training.train_starvla][INFO] - Step 1601, Loss: {'train/loss': 0.03456306457519531, 'train/grad_norm_pre_clip': 51.644187927246094, 'timing/data': 0.00021202999050728977, 'timing/model': 0.39151744000264443, 'learning_rate/qwen_vl_interface': 7.965055807032734e-06, 'learning_rate/action_model': 7.965055807032735e-05, 'epoch': 0.12}) +[2026-10-03 19:14:50,713][starVLA.training.train_starvla][INFO] - Step 1602, Loss: {'train/loss': 0.0040235985070466995, 'train/grad_norm_pre_clip': 1.9801822900772095, 'timing/data': 0.00030523998429998755, 'timing/model': 0.39239840998197906, 'learning_rate/qwen_vl_interface': 7.962556362428275e-06, 'learning_rate/action_model': 7.962556362428274e-05, 'epoch': 0.12}) +[2026-10-03 19:14:51,619][starVLA.training.train_starvla][INFO] - Step 1603, Loss: {'train/loss': 7.550480222562328e-05, 'train/grad_norm_pre_clip': 0.01455469150096178, 'timing/data': 0.0002139200223609805, 'timing/model': 0.3924135999986902, 'learning_rate/qwen_vl_interface': 7.960055802794358e-06, 'learning_rate/action_model': 7.960055802794357e-05, 'epoch': 0.12}) +[2026-10-03 19:14:52,317][starVLA.training.train_starvla][INFO] - Step 1604, Loss: {'train/loss': 0.006546338088810444, 'train/grad_norm_pre_clip': 15.730731010437012, 'timing/data': 0.00020801002392545342, 'timing/model': 0.39044982998166233, 'learning_rate/qwen_vl_interface': 7.957554129158866e-06, 'learning_rate/action_model': 7.957554129158866e-05, 'epoch': 0.12}) +[2026-10-03 19:14:53,017][starVLA.training.train_starvla][INFO] - Step 1605, Loss: {'train/loss': 0.005731530487537384, 'train/grad_norm_pre_clip': 2.7635796070098877, 'timing/data': 0.00034381001023575664, 'timing/model': 0.3915702899976168, 'learning_rate/qwen_vl_interface': 7.955051342550148e-06, 'learning_rate/action_model': 7.955051342550148e-05, 'epoch': 0.12}) +[2026-10-03 19:14:53,716][starVLA.training.train_starvla][INFO] - Step 1606, Loss: {'train/loss': 0.28650376200675964, 'train/grad_norm_pre_clip': 38.66332244873047, 'timing/data': 0.00036294999881647527, 'timing/model': 0.3920279199955985, 'learning_rate/qwen_vl_interface': 7.952547443997e-06, 'learning_rate/action_model': 7.952547443997e-05, 'epoch': 0.12}) +[2026-10-03 19:14:54,416][starVLA.training.train_starvla][INFO] - Step 1607, Loss: {'train/loss': 0.004119806457310915, 'train/grad_norm_pre_clip': 0.5655640959739685, 'timing/data': 0.0003166800015605986, 'timing/model': 0.39147147999028675, 'learning_rate/qwen_vl_interface': 7.950042434528681e-06, 'learning_rate/action_model': 7.950042434528682e-05, 'epoch': 0.12}) +[2026-10-03 19:14:55,324][starVLA.training.train_starvla][INFO] - Step 1608, Loss: {'train/loss': 0.2804999351501465, 'train/grad_norm_pre_clip': 78.4814453125, 'timing/data': 0.00036679001641459763, 'timing/model': 0.5982577100221533, 'learning_rate/qwen_vl_interface': 7.94753631517491e-06, 'learning_rate/action_model': 7.947536315174909e-05, 'epoch': 0.12}) +[2026-10-03 19:14:56,025][starVLA.training.train_starvla][INFO] - Step 1609, Loss: {'train/loss': 0.03493121266365051, 'train/grad_norm_pre_clip': 22.227943420410156, 'timing/data': 0.00029272999381646514, 'timing/model': 0.39284737000707537, 'learning_rate/qwen_vl_interface': 7.945029086965854e-06, 'learning_rate/action_model': 7.945029086965853e-05, 'epoch': 0.12}) +[2026-10-03 19:14:56,725][starVLA.training.train_starvla][INFO] - Step 1610, Loss: {'train/loss': 0.8343842625617981, 'train/grad_norm_pre_clip': 49.413028717041016, 'timing/data': 0.00043139999615959823, 'timing/model': 0.3913202599796932, 'learning_rate/qwen_vl_interface': 7.942520750932138e-06, 'learning_rate/action_model': 7.942520750932138e-05, 'epoch': 0.12}) +[2026-10-03 19:14:57,424][starVLA.training.train_starvla][INFO] - Step 1611, Loss: {'train/loss': 0.3122459053993225, 'train/grad_norm_pre_clip': 73.92931365966797, 'timing/data': 0.00022273999638855457, 'timing/model': 0.39009849101421423, 'learning_rate/qwen_vl_interface': 7.940011308104848e-06, 'learning_rate/action_model': 7.940011308104848e-05, 'epoch': 0.12}) +[2026-10-03 19:14:58,122][starVLA.training.train_starvla][INFO] - Step 1612, Loss: {'train/loss': 0.2418859452009201, 'train/grad_norm_pre_clip': 96.55318450927734, 'timing/data': 0.00029408998670987785, 'timing/model': 0.38924052999936976, 'learning_rate/qwen_vl_interface': 7.93750075951552e-06, 'learning_rate/action_model': 7.93750075951552e-05, 'epoch': 0.12}) +[2026-10-03 19:14:59,030][starVLA.training.train_starvla][INFO] - Step 1613, Loss: {'train/loss': 0.34038233757019043, 'train/grad_norm_pre_clip': 37.13970184326172, 'timing/data': 0.00034328000037930906, 'timing/model': 0.6007277790049557, 'learning_rate/qwen_vl_interface': 7.934989106196145e-06, 'learning_rate/action_model': 7.934989106196146e-05, 'epoch': 0.12}) +[2026-10-03 19:14:59,727][starVLA.training.train_starvla][INFO] - Step 1614, Loss: {'train/loss': 0.30655407905578613, 'train/grad_norm_pre_clip': 11.838040351867676, 'timing/data': 0.00025286999880336225, 'timing/model': 0.39106885000364855, 'learning_rate/qwen_vl_interface': 7.93247634917917e-06, 'learning_rate/action_model': 7.93247634917917e-05, 'epoch': 0.12}) +[2026-10-03 19:15:00,426][starVLA.training.train_starvla][INFO] - Step 1615, Loss: {'train/loss': 0.008958397433161736, 'train/grad_norm_pre_clip': 4.6503071784973145, 'timing/data': 0.0003370099875610322, 'timing/model': 0.3903945299971383, 'learning_rate/qwen_vl_interface': 7.929962489497494e-06, 'learning_rate/action_model': 7.929962489497493e-05, 'epoch': 0.12}) +[2026-10-03 19:15:01,125][starVLA.training.train_starvla][INFO] - Step 1616, Loss: {'train/loss': 0.3025551438331604, 'train/grad_norm_pre_clip': 97.16060638427734, 'timing/data': 0.0003917499852832407, 'timing/model': 0.3904260199924465, 'learning_rate/qwen_vl_interface': 7.927447528184469e-06, 'learning_rate/action_model': 7.927447528184468e-05, 'epoch': 0.12}) +[2026-10-03 19:15:01,825][starVLA.training.train_starvla][INFO] - Step 1617, Loss: {'train/loss': 0.13493824005126953, 'train/grad_norm_pre_clip': 38.99038314819336, 'timing/data': 0.0002442999975755811, 'timing/model': 0.39200968999648467, 'learning_rate/qwen_vl_interface': 7.924931466273901e-06, 'learning_rate/action_model': 7.924931466273902e-05, 'epoch': 0.12}) +[2026-10-03 19:15:02,525][starVLA.training.train_starvla][INFO] - Step 1618, Loss: {'train/loss': 0.03019549697637558, 'train/grad_norm_pre_clip': 11.178238868713379, 'timing/data': 0.0002399600052740425, 'timing/model': 0.3914741509943269, 'learning_rate/qwen_vl_interface': 7.92241430480005e-06, 'learning_rate/action_model': 7.922414304800049e-05, 'epoch': 0.12}) +[2026-10-03 19:15:03,428][starVLA.training.train_starvla][INFO] - Step 1619, Loss: {'train/loss': 0.5236009955406189, 'train/grad_norm_pre_clip': 146.69696044921875, 'timing/data': 0.00031897000735625625, 'timing/model': 0.39308314002119005, 'learning_rate/qwen_vl_interface': 7.919896044797622e-06, 'learning_rate/action_model': 7.91989604479762e-05, 'epoch': 0.12}) +[2026-10-03 19:15:04,130][starVLA.training.train_starvla][INFO] - Step 1620, Loss: {'train/loss': 0.4872153699398041, 'train/grad_norm_pre_clip': 33.93320083618164, 'timing/data': 0.00032423000084236264, 'timing/model': 0.39140535998740233, 'learning_rate/qwen_vl_interface': 7.91737668730178e-06, 'learning_rate/action_model': 7.91737668730178e-05, 'epoch': 0.12}) +[2026-10-03 19:15:04,828][starVLA.training.train_starvla][INFO] - Step 1621, Loss: {'train/loss': 0.33583587408065796, 'train/grad_norm_pre_clip': 81.34005737304688, 'timing/data': 0.00024856001255102456, 'timing/model': 0.3913846899813507, 'learning_rate/qwen_vl_interface': 7.91485623334814e-06, 'learning_rate/action_model': 7.91485623334814e-05, 'epoch': 0.12}) +[2026-10-03 19:15:05,527][starVLA.training.train_starvla][INFO] - Step 1622, Loss: {'train/loss': 0.5162484049797058, 'train/grad_norm_pre_clip': 21.785274505615234, 'timing/data': 0.0002449000021442771, 'timing/model': 0.3921383999986574, 'learning_rate/qwen_vl_interface': 7.91233468397276e-06, 'learning_rate/action_model': 7.912334683972761e-05, 'epoch': 0.12}) +[2026-10-03 19:15:06,225][starVLA.training.train_starvla][INFO] - Step 1623, Loss: {'train/loss': 0.26610565185546875, 'train/grad_norm_pre_clip': 110.4852294921875, 'timing/data': 0.000248780008405447, 'timing/model': 0.39142766999430023, 'learning_rate/qwen_vl_interface': 7.90981204021216e-06, 'learning_rate/action_model': 7.90981204021216e-05, 'epoch': 0.12}) +[2026-10-03 19:15:07,127][starVLA.training.train_starvla][INFO] - Step 1624, Loss: {'train/loss': 0.6630194783210754, 'train/grad_norm_pre_clip': 32.78912353515625, 'timing/data': 0.0002147899940609932, 'timing/model': 0.3912583499914035, 'learning_rate/qwen_vl_interface': 7.907288303103298e-06, 'learning_rate/action_model': 7.907288303103298e-05, 'epoch': 0.12}) +[2026-10-03 19:15:07,826][starVLA.training.train_starvla][INFO] - Step 1625, Loss: {'train/loss': 0.0033497766125947237, 'train/grad_norm_pre_clip': 2.3888750076293945, 'timing/data': 0.00023060000967234373, 'timing/model': 0.3920193499943707, 'learning_rate/qwen_vl_interface': 7.90476347368359e-06, 'learning_rate/action_model': 7.90476347368359e-05, 'epoch': 0.12}) +[2026-10-03 19:15:08,524][starVLA.training.train_starvla][INFO] - Step 1626, Loss: {'train/loss': 0.03195341303944588, 'train/grad_norm_pre_clip': 23.11224365234375, 'timing/data': 0.000244739989284426, 'timing/model': 0.3911019299994223, 'learning_rate/qwen_vl_interface': 7.902237552990897e-06, 'learning_rate/action_model': 7.902237552990897e-05, 'epoch': 0.12}) +[2026-10-03 19:15:09,221][starVLA.training.train_starvla][INFO] - Step 1627, Loss: {'train/loss': 0.5005162954330444, 'train/grad_norm_pre_clip': 308.9063415527344, 'timing/data': 0.00030578000587411225, 'timing/model': 0.3909632099966984, 'learning_rate/qwen_vl_interface': 7.89971054206353e-06, 'learning_rate/action_model': 7.89971054206353e-05, 'epoch': 0.12}) +[2026-10-03 19:15:09,922][starVLA.training.train_starvla][INFO] - Step 1628, Loss: {'train/loss': 0.021642951294779778, 'train/grad_norm_pre_clip': 8.966803550720215, 'timing/data': 0.00027328997384756804, 'timing/model': 0.39099890997749753, 'learning_rate/qwen_vl_interface': 7.897182441940247e-06, 'learning_rate/action_model': 7.897182441940247e-05, 'epoch': 0.12}) +[2026-10-03 19:15:10,821][starVLA.training.train_starvla][INFO] - Step 1629, Loss: {'train/loss': 0.5464867949485779, 'train/grad_norm_pre_clip': 95.64860534667969, 'timing/data': 0.0002870299795176834, 'timing/model': 0.5935788500064518, 'learning_rate/qwen_vl_interface': 7.894653253660256e-06, 'learning_rate/action_model': 7.894653253660256e-05, 'epoch': 0.12}) +[2026-10-03 19:15:11,521][starVLA.training.train_starvla][INFO] - Step 1630, Loss: {'train/loss': 0.0020152374636381865, 'train/grad_norm_pre_clip': 0.4245477616786957, 'timing/data': 0.00039391001337207854, 'timing/model': 0.3911270200042054, 'learning_rate/qwen_vl_interface': 7.89212297826321e-06, 'learning_rate/action_model': 7.892122978263209e-05, 'epoch': 0.12}) +[2026-10-03 19:15:12,220][starVLA.training.train_starvla][INFO] - Step 1631, Loss: {'train/loss': 0.0022706345189362764, 'train/grad_norm_pre_clip': 0.28985482454299927, 'timing/data': 0.0002017799997702241, 'timing/model': 0.39183141000103205, 'learning_rate/qwen_vl_interface': 7.889591616789206e-06, 'learning_rate/action_model': 7.889591616789206e-05, 'epoch': 0.12}) +[2026-10-03 19:15:12,923][starVLA.training.train_starvla][INFO] - Step 1632, Loss: {'train/loss': 0.46566444635391235, 'train/grad_norm_pre_clip': 68.68962860107422, 'timing/data': 0.00027709000278264284, 'timing/model': 0.39304603001801297, 'learning_rate/qwen_vl_interface': 7.887059170278799e-06, 'learning_rate/action_model': 7.887059170278797e-05, 'epoch': 0.12}) +[2026-10-03 19:15:13,621][starVLA.training.train_starvla][INFO] - Step 1633, Loss: {'train/loss': 0.1290763020515442, 'train/grad_norm_pre_clip': 54.30384063720703, 'timing/data': 0.0002652399998623878, 'timing/model': 0.39134203002322465, 'learning_rate/qwen_vl_interface': 7.884525639772974e-06, 'learning_rate/action_model': 7.884525639772974e-05, 'epoch': 0.12}) +[2026-10-03 19:15:14,527][starVLA.training.train_starvla][INFO] - Step 1634, Loss: {'train/loss': 0.010602263733744621, 'train/grad_norm_pre_clip': 3.211991786956787, 'timing/data': 0.00020586000755429268, 'timing/model': 0.599903728987556, 'learning_rate/qwen_vl_interface': 7.881991026313176e-06, 'learning_rate/action_model': 7.881991026313175e-05, 'epoch': 0.12}) +[2026-10-03 19:15:15,225][starVLA.training.train_starvla][INFO] - Step 1635, Loss: {'train/loss': 0.5093170404434204, 'train/grad_norm_pre_clip': 27.450881958007812, 'timing/data': 0.00030751898884773254, 'timing/model': 0.39045770099619403, 'learning_rate/qwen_vl_interface': 7.879455330941283e-06, 'learning_rate/action_model': 7.879455330941282e-05, 'epoch': 0.12}) +[2026-10-03 19:15:15,924][starVLA.training.train_starvla][INFO] - Step 1636, Loss: {'train/loss': 0.0032928078435361385, 'train/grad_norm_pre_clip': 1.3670272827148438, 'timing/data': 0.0003329500032123178, 'timing/model': 0.3910301699943375, 'learning_rate/qwen_vl_interface': 7.876918554699627e-06, 'learning_rate/action_model': 7.876918554699627e-05, 'epoch': 0.12}) +[2026-10-03 19:15:16,623][starVLA.training.train_starvla][INFO] - Step 1637, Loss: {'train/loss': 0.0022121407091617584, 'train/grad_norm_pre_clip': 0.31670695543289185, 'timing/data': 0.0002402900136075914, 'timing/model': 0.3919050310214516, 'learning_rate/qwen_vl_interface': 7.874380698630982e-06, 'learning_rate/action_model': 7.874380698630982e-05, 'epoch': 0.12}) +[2026-10-03 19:15:17,321][starVLA.training.train_starvla][INFO] - Step 1638, Loss: {'train/loss': 0.0036800613161176443, 'train/grad_norm_pre_clip': 0.40218982100486755, 'timing/data': 0.00033275000168941915, 'timing/model': 0.38937536999583244, 'learning_rate/qwen_vl_interface': 7.871841763778562e-06, 'learning_rate/action_model': 7.871841763778563e-05, 'epoch': 0.12}) +[2026-10-03 19:15:18,016][starVLA.training.train_starvla][INFO] - Step 1639, Loss: {'train/loss': 0.3643287420272827, 'train/grad_norm_pre_clip': 38.78214645385742, 'timing/data': 0.00030437999521382153, 'timing/model': 0.3898343100154307, 'learning_rate/qwen_vl_interface': 7.86930175118603e-06, 'learning_rate/action_model': 7.86930175118603e-05, 'epoch': 0.12}) +[2026-10-03 19:15:18,915][starVLA.training.train_starvla][INFO] - Step 1640, Loss: {'train/loss': 0.5997245907783508, 'train/grad_norm_pre_clip': 44.0698356628418, 'timing/data': 0.0003447399940341711, 'timing/model': 0.39012162998551503, 'learning_rate/qwen_vl_interface': 7.866760661897487e-06, 'learning_rate/action_model': 7.866760661897487e-05, 'epoch': 0.12}) +[2026-10-03 19:15:19,614][starVLA.training.train_starvla][INFO] - Step 1641, Loss: {'train/loss': 0.0058873919770121574, 'train/grad_norm_pre_clip': 4.470987319946289, 'timing/data': 0.00020941000548191369, 'timing/model': 0.3921656210150104, 'learning_rate/qwen_vl_interface': 7.86421849695748e-06, 'learning_rate/action_model': 7.864218496957481e-05, 'epoch': 0.12}) +[2026-10-03 19:15:20,312][starVLA.training.train_starvla][INFO] - Step 1642, Loss: {'train/loss': 0.0009633747395128012, 'train/grad_norm_pre_clip': 0.15834380686283112, 'timing/data': 0.00020222997409291565, 'timing/model': 0.39061668000067584, 'learning_rate/qwen_vl_interface': 7.861675257410998e-06, 'learning_rate/action_model': 7.861675257410997e-05, 'epoch': 0.12}) +[2026-10-03 19:15:21,010][starVLA.training.train_starvla][INFO] - Step 1643, Loss: {'train/loss': 0.002011539414525032, 'train/grad_norm_pre_clip': 1.216989517211914, 'timing/data': 0.00023877000785432756, 'timing/model': 0.3902157400152646, 'learning_rate/qwen_vl_interface': 7.859130944303468e-06, 'learning_rate/action_model': 7.859130944303467e-05, 'epoch': 0.12}) +[2026-10-03 19:15:21,707][starVLA.training.train_starvla][INFO] - Step 1644, Loss: {'train/loss': 0.33778730034828186, 'train/grad_norm_pre_clip': 28.964359283447266, 'timing/data': 0.00032143000862561166, 'timing/model': 0.39065215998562053, 'learning_rate/qwen_vl_interface': 7.856585558680761e-06, 'learning_rate/action_model': 7.856585558680761e-05, 'epoch': 0.12}) +[2026-10-03 19:15:22,606][starVLA.training.train_starvla][INFO] - Step 1645, Loss: {'train/loss': 0.4136624038219452, 'train/grad_norm_pre_clip': 28.31222915649414, 'timing/data': 0.0002927600289694965, 'timing/model': 0.5926897999888752, 'learning_rate/qwen_vl_interface': 7.854039101589191e-06, 'learning_rate/action_model': 7.854039101589191e-05, 'epoch': 0.12}) +[2026-10-03 19:15:23,302][starVLA.training.train_starvla][INFO] - Step 1646, Loss: {'train/loss': 0.4708957076072693, 'train/grad_norm_pre_clip': 81.28292083740234, 'timing/data': 0.00020154999219812453, 'timing/model': 0.3906096509890631, 'learning_rate/qwen_vl_interface': 7.851491574075509e-06, 'learning_rate/action_model': 7.851491574075509e-05, 'epoch': 0.12}) +[2026-10-03 19:15:24,001][starVLA.training.train_starvla][INFO] - Step 1647, Loss: {'train/loss': 0.29677140712738037, 'train/grad_norm_pre_clip': 31.848281860351562, 'timing/data': 0.00034610999864526093, 'timing/model': 0.39181413999176584, 'learning_rate/qwen_vl_interface': 7.848942977186909e-06, 'learning_rate/action_model': 7.848942977186908e-05, 'epoch': 0.12}) +[2026-10-03 19:15:24,698][starVLA.training.train_starvla][INFO] - Step 1648, Loss: {'train/loss': 0.001219258876517415, 'train/grad_norm_pre_clip': 0.24446111917495728, 'timing/data': 0.00028010999085381627, 'timing/model': 0.38924342099926434, 'learning_rate/qwen_vl_interface': 7.84639331197102e-06, 'learning_rate/action_model': 7.84639331197102e-05, 'epoch': 0.12}) +[2026-10-03 19:15:25,399][starVLA.training.train_starvla][INFO] - Step 1649, Loss: {'train/loss': 0.0057898093946278095, 'train/grad_norm_pre_clip': 3.0871989727020264, 'timing/data': 0.00029691000236198306, 'timing/model': 0.3923241500160657, 'learning_rate/qwen_vl_interface': 7.843842579475918e-06, 'learning_rate/action_model': 7.843842579475918e-05, 'epoch': 0.12}) +[2026-10-03 19:15:26,313][starVLA.training.train_starvla][INFO] - Step 1650, Loss: {'train/loss': 0.0037758122198283672, 'train/grad_norm_pre_clip': 1.5743200778961182, 'timing/data': 0.0003834200033452362, 'timing/model': 0.6024504190136213, 'learning_rate/qwen_vl_interface': 7.841290780750107e-06, 'learning_rate/action_model': 7.841290780750107e-05, 'epoch': 0.12}) +[2026-10-03 19:15:27,009][starVLA.training.train_starvla][INFO] - Step 1651, Loss: {'train/loss': 0.05932158976793289, 'train/grad_norm_pre_clip': 66.46174621582031, 'timing/data': 0.00024250001297332346, 'timing/model': 0.3903254200122319, 'learning_rate/qwen_vl_interface': 7.838737916842538e-06, 'learning_rate/action_model': 7.838737916842538e-05, 'epoch': 0.12}) +[2026-10-03 19:15:27,708][starVLA.training.train_starvla][INFO] - Step 1652, Loss: {'train/loss': 0.0038023493252694607, 'train/grad_norm_pre_clip': 1.1727677583694458, 'timing/data': 0.0003420890134293586, 'timing/model': 0.3915222409996204, 'learning_rate/qwen_vl_interface': 7.836183988802596e-06, 'learning_rate/action_model': 7.836183988802595e-05, 'epoch': 0.12}) +[2026-10-03 19:15:28,408][starVLA.training.train_starvla][INFO] - Step 1653, Loss: {'train/loss': 0.16276247799396515, 'train/grad_norm_pre_clip': 179.10296630859375, 'timing/data': 0.0002392799942754209, 'timing/model': 0.39170802998705767, 'learning_rate/qwen_vl_interface': 7.833628997680107e-06, 'learning_rate/action_model': 7.833628997680107e-05, 'epoch': 0.12}) +[2026-10-03 19:15:29,105][starVLA.training.train_starvla][INFO] - Step 1654, Loss: {'train/loss': 0.5521660447120667, 'train/grad_norm_pre_clip': 99.93586730957031, 'timing/data': 0.00037253001937642694, 'timing/model': 0.39001156098674983, 'learning_rate/qwen_vl_interface': 7.831072944525329e-06, 'learning_rate/action_model': 7.831072944525328e-05, 'epoch': 0.12}) +[2026-10-03 19:15:29,804][starVLA.training.train_starvla][INFO] - Step 1655, Loss: {'train/loss': 0.10512498021125793, 'train/grad_norm_pre_clip': 166.54640197753906, 'timing/data': 0.00020464000408537686, 'timing/model': 0.3918048509804066, 'learning_rate/qwen_vl_interface': 7.828515830388956e-06, 'learning_rate/action_model': 7.828515830388956e-05, 'epoch': 0.12}) +[2026-10-03 19:15:30,706][starVLA.training.train_starvla][INFO] - Step 1656, Loss: {'train/loss': 0.1320529580116272, 'train/grad_norm_pre_clip': 16.538570404052734, 'timing/data': 0.0002572000084910542, 'timing/model': 0.3913487000099849, 'learning_rate/qwen_vl_interface': 7.825957656322129e-06, 'learning_rate/action_model': 7.825957656322128e-05, 'epoch': 0.12}) +[2026-10-03 19:15:31,404][starVLA.training.train_starvla][INFO] - Step 1657, Loss: {'train/loss': 0.0018943515606224537, 'train/grad_norm_pre_clip': 0.6315521597862244, 'timing/data': 0.0002741200150921941, 'timing/model': 0.39160704097594135, 'learning_rate/qwen_vl_interface': 7.823398423376406e-06, 'learning_rate/action_model': 7.823398423376406e-05, 'epoch': 0.12}) +[2026-10-03 19:15:32,101][starVLA.training.train_starvla][INFO] - Step 1658, Loss: {'train/loss': 0.6683547496795654, 'train/grad_norm_pre_clip': 58.58354949951172, 'timing/data': 0.00021224998636171222, 'timing/model': 0.3909970799868461, 'learning_rate/qwen_vl_interface': 7.820838132603801e-06, 'learning_rate/action_model': 7.8208381326038e-05, 'epoch': 0.12}) +[2026-10-03 19:15:32,799][starVLA.training.train_starvla][INFO] - Step 1659, Loss: {'train/loss': 0.23700346052646637, 'train/grad_norm_pre_clip': 22.281461715698242, 'timing/data': 0.00024091999512165785, 'timing/model': 0.39192834999994375, 'learning_rate/qwen_vl_interface': 7.818276785056748e-06, 'learning_rate/action_model': 7.818276785056749e-05, 'epoch': 0.12}) +[2026-10-03 19:15:33,500][starVLA.training.train_starvla][INFO] - Step 1660, Loss: {'train/loss': 0.4668988287448883, 'train/grad_norm_pre_clip': 73.07605743408203, 'timing/data': 0.00023941902327351272, 'timing/model': 0.391681120992871, 'learning_rate/qwen_vl_interface': 7.815714381788123e-06, 'learning_rate/action_model': 7.815714381788122e-05, 'epoch': 0.12}) +[2026-10-03 19:15:34,402][starVLA.training.train_starvla][INFO] - Step 1661, Loss: {'train/loss': 0.003468951443210244, 'train/grad_norm_pre_clip': 2.1412603855133057, 'timing/data': 0.0002868100127670914, 'timing/model': 0.39292779998504557, 'learning_rate/qwen_vl_interface': 7.813150923851227e-06, 'learning_rate/action_model': 7.813150923851227e-05, 'epoch': 0.12}) +[2026-10-03 19:15:35,099][starVLA.training.train_starvla][INFO] - Step 1662, Loss: {'train/loss': 0.09883063286542892, 'train/grad_norm_pre_clip': 18.86981773376465, 'timing/data': 0.00039605001802556217, 'timing/model': 0.3896525510062929, 'learning_rate/qwen_vl_interface': 7.810586412299807e-06, 'learning_rate/action_model': 7.810586412299806e-05, 'epoch': 0.12}) +[2026-10-03 19:15:35,798][starVLA.training.train_starvla][INFO] - Step 1663, Loss: {'train/loss': 0.48363298177719116, 'train/grad_norm_pre_clip': 118.22046661376953, 'timing/data': 0.000206179975066334, 'timing/model': 0.39121231000171974, 'learning_rate/qwen_vl_interface': 7.808020848188034e-06, 'learning_rate/action_model': 7.808020848188034e-05, 'epoch': 0.12}) +[2026-10-03 19:15:36,497][starVLA.training.train_starvla][INFO] - Step 1664, Loss: {'train/loss': 0.9605964422225952, 'train/grad_norm_pre_clip': 91.36759185791016, 'timing/data': 0.00024538001162, 'timing/model': 0.3898375399876386, 'learning_rate/qwen_vl_interface': 7.805454232570517e-06, 'learning_rate/action_model': 7.805454232570517e-05, 'epoch': 0.12}) +[2026-10-03 19:15:37,196][starVLA.training.train_starvla][INFO] - Step 1665, Loss: {'train/loss': 0.0022573708556592464, 'train/grad_norm_pre_clip': 0.9094911217689514, 'timing/data': 0.0003478199942037463, 'timing/model': 0.3912693010061048, 'learning_rate/qwen_vl_interface': 7.80288656650229e-06, 'learning_rate/action_model': 7.80288656650229e-05, 'epoch': 0.12}) +[2026-10-03 19:15:38,111][starVLA.training.train_starvla][INFO] - Step 1666, Loss: {'train/loss': 0.14003115892410278, 'train/grad_norm_pre_clip': 69.2149658203125, 'timing/data': 0.00035303999902680516, 'timing/model': 0.6066609679837711, 'learning_rate/qwen_vl_interface': 7.800317851038829e-06, 'learning_rate/action_model': 7.800317851038828e-05, 'epoch': 0.12}) +[2026-10-03 19:15:38,810][starVLA.training.train_starvla][INFO] - Step 1667, Loss: {'train/loss': 0.002670040586963296, 'train/grad_norm_pre_clip': 0.6415619254112244, 'timing/data': 0.00032248999923467636, 'timing/model': 0.39069801999721676, 'learning_rate/qwen_vl_interface': 7.797748087236033e-06, 'learning_rate/action_model': 7.797748087236033e-05, 'epoch': 0.12}) +[2026-10-03 19:15:39,509][starVLA.training.train_starvla][INFO] - Step 1668, Loss: {'train/loss': 0.0009975855937227607, 'train/grad_norm_pre_clip': 0.16647766530513763, 'timing/data': 0.000281209999229759, 'timing/model': 0.39001266000559554, 'learning_rate/qwen_vl_interface': 7.795177276150234e-06, 'learning_rate/action_model': 7.795177276150234e-05, 'epoch': 0.12}) +[2026-10-03 19:15:40,206][starVLA.training.train_starvla][INFO] - Step 1669, Loss: {'train/loss': 0.006564225070178509, 'train/grad_norm_pre_clip': 2.285212278366089, 'timing/data': 0.0002774499880615622, 'timing/model': 0.3915425500017591, 'learning_rate/qwen_vl_interface': 7.792605418838196e-06, 'learning_rate/action_model': 7.792605418838197e-05, 'epoch': 0.12}) +[2026-10-03 19:15:40,904][starVLA.training.train_starvla][INFO] - Step 1670, Loss: {'train/loss': 0.060307763516902924, 'train/grad_norm_pre_clip': 14.293248176574707, 'timing/data': 0.000302100001135841, 'timing/model': 0.3915154100104701, 'learning_rate/qwen_vl_interface': 7.790032516357116e-06, 'learning_rate/action_model': 7.790032516357117e-05, 'epoch': 0.12}) +[2026-10-03 19:15:41,809][starVLA.training.train_starvla][INFO] - Step 1671, Loss: {'train/loss': 0.0866599977016449, 'train/grad_norm_pre_clip': 30.396846771240234, 'timing/data': 0.00021214998560026288, 'timing/model': 0.5957962590036914, 'learning_rate/qwen_vl_interface': 7.787458569764613e-06, 'learning_rate/action_model': 7.787458569764612e-05, 'epoch': 0.12}) +[2026-10-03 19:15:42,507][starVLA.training.train_starvla][INFO] - Step 1672, Loss: {'train/loss': 0.0689820796251297, 'train/grad_norm_pre_clip': 113.93560791015625, 'timing/data': 0.0002152700035367161, 'timing/model': 0.38986913999542594, 'learning_rate/qwen_vl_interface': 7.784883580118738e-06, 'learning_rate/action_model': 7.784883580118738e-05, 'epoch': 0.12}) +[2026-10-03 19:15:43,207][starVLA.training.train_starvla][INFO] - Step 1673, Loss: {'train/loss': 0.08919452875852585, 'train/grad_norm_pre_clip': 34.574546813964844, 'timing/data': 0.0002871600154321641, 'timing/model': 0.3926855300087482, 'learning_rate/qwen_vl_interface': 7.782307548477975e-06, 'learning_rate/action_model': 7.782307548477975e-05, 'epoch': 0.12}) +[2026-10-03 19:15:43,905][starVLA.training.train_starvla][INFO] - Step 1674, Loss: {'train/loss': 0.12463420629501343, 'train/grad_norm_pre_clip': 188.7953643798828, 'timing/data': 0.0003364000003784895, 'timing/model': 0.38935792999109253, 'learning_rate/qwen_vl_interface': 7.779730475901233e-06, 'learning_rate/action_model': 7.779730475901232e-05, 'epoch': 0.12}) +[2026-10-03 19:15:44,605][starVLA.training.train_starvla][INFO] - Step 1675, Loss: {'train/loss': 0.5233395099639893, 'train/grad_norm_pre_clip': 27.604215621948242, 'timing/data': 0.00027288999990560114, 'timing/model': 0.39152251099585555, 'learning_rate/qwen_vl_interface': 7.777152363447849e-06, 'learning_rate/action_model': 7.77715236344785e-05, 'epoch': 0.12}) +[2026-10-03 19:15:45,509][starVLA.training.train_starvla][INFO] - Step 1676, Loss: {'train/loss': 0.018622437492012978, 'train/grad_norm_pre_clip': 17.80839729309082, 'timing/data': 0.0002880200045183301, 'timing/model': 0.596606798004359, 'learning_rate/qwen_vl_interface': 7.774573212177587e-06, 'learning_rate/action_model': 7.774573212177588e-05, 'epoch': 0.12}) +[2026-10-03 19:15:46,208][starVLA.training.train_starvla][INFO] - Step 1677, Loss: {'train/loss': 0.49570250511169434, 'train/grad_norm_pre_clip': 36.9979362487793, 'timing/data': 0.0003008399799000472, 'timing/model': 0.39181296998867765, 'learning_rate/qwen_vl_interface': 7.77199302315064e-06, 'learning_rate/action_model': 7.77199302315064e-05, 'epoch': 0.12}) +[2026-10-03 19:15:46,907][starVLA.training.train_starvla][INFO] - Step 1678, Loss: {'train/loss': 0.4150112569332123, 'train/grad_norm_pre_clip': 36.020381927490234, 'timing/data': 0.00032158999238163233, 'timing/model': 0.39149320000433363, 'learning_rate/qwen_vl_interface': 7.769411797427622e-06, 'learning_rate/action_model': 7.769411797427623e-05, 'epoch': 0.12}) +[2026-10-03 19:15:47,607][starVLA.training.train_starvla][INFO] - Step 1679, Loss: {'train/loss': 1.1907179355621338, 'train/grad_norm_pre_clip': 89.0807876586914, 'timing/data': 0.00035153000499121845, 'timing/model': 0.39139912999235094, 'learning_rate/qwen_vl_interface': 7.766829536069584e-06, 'learning_rate/action_model': 7.766829536069584e-05, 'epoch': 0.12}) +[2026-10-03 19:15:48,307][starVLA.training.train_starvla][INFO] - Step 1680, Loss: {'train/loss': 0.0006211434956640005, 'train/grad_norm_pre_clip': 0.11700417101383209, 'timing/data': 0.00025024000206030905, 'timing/model': 0.391221880010562, 'learning_rate/qwen_vl_interface': 7.764246240137991e-06, 'learning_rate/action_model': 7.764246240137991e-05, 'epoch': 0.12}) +[2026-10-03 19:15:49,006][starVLA.training.train_starvla][INFO] - Step 1681, Loss: {'train/loss': 0.0015211477875709534, 'train/grad_norm_pre_clip': 0.26167991757392883, 'timing/data': 0.00029552000341936946, 'timing/model': 0.3908939399989322, 'learning_rate/qwen_vl_interface': 7.761661910694741e-06, 'learning_rate/action_model': 7.761661910694741e-05, 'epoch': 0.12}) +[2026-10-03 19:15:49,910][starVLA.training.train_starvla][INFO] - Step 1682, Loss: {'train/loss': 0.008547384291887283, 'train/grad_norm_pre_clip': 8.45814037322998, 'timing/data': 0.00024623001809231937, 'timing/model': 0.38913661101832986, 'learning_rate/qwen_vl_interface': 7.759076548802153e-06, 'learning_rate/action_model': 7.759076548802152e-05, 'epoch': 0.12}) +[2026-10-03 19:15:50,608][starVLA.training.train_starvla][INFO] - Step 1683, Loss: {'train/loss': 0.0054236832074820995, 'train/grad_norm_pre_clip': 2.0538928508758545, 'timing/data': 0.00020388999837450683, 'timing/model': 0.39249007002217695, 'learning_rate/qwen_vl_interface': 7.75649015552297e-06, 'learning_rate/action_model': 7.75649015552297e-05, 'epoch': 0.12}) +[2026-10-03 19:15:51,304][starVLA.training.train_starvla][INFO] - Step 1684, Loss: {'train/loss': 0.5318600535392761, 'train/grad_norm_pre_clip': 50.23517608642578, 'timing/data': 0.0003638399939518422, 'timing/model': 0.3891563299985137, 'learning_rate/qwen_vl_interface': 7.753902731920364e-06, 'learning_rate/action_model': 7.753902731920364e-05, 'epoch': 0.12}) +[2026-10-03 19:15:52,003][starVLA.training.train_starvla][INFO] - Step 1685, Loss: {'train/loss': 0.016575314104557037, 'train/grad_norm_pre_clip': 27.635845184326172, 'timing/data': 0.0002874100173357874, 'timing/model': 0.391829920001328, 'learning_rate/qwen_vl_interface': 7.751314279057925e-06, 'learning_rate/action_model': 7.751314279057923e-05, 'epoch': 0.12}) +[2026-10-03 19:15:52,706][starVLA.training.train_starvla][INFO] - Step 1686, Loss: {'train/loss': 0.3578343391418457, 'train/grad_norm_pre_clip': 57.524967193603516, 'timing/data': 0.0003460989973973483, 'timing/model': 0.39271332099451683, 'learning_rate/qwen_vl_interface': 7.74872479799967e-06, 'learning_rate/action_model': 7.74872479799967e-05, 'epoch': 0.12}) +[2026-10-03 19:15:53,617][starVLA.training.train_starvla][INFO] - Step 1687, Loss: {'train/loss': 0.001358124427497387, 'train/grad_norm_pre_clip': 0.3631455600261688, 'timing/data': 0.00026220001745969057, 'timing/model': 0.39122965998831205, 'learning_rate/qwen_vl_interface': 7.746134289810034e-06, 'learning_rate/action_model': 7.746134289810033e-05, 'epoch': 0.12}) +[2026-10-03 19:15:54,315][starVLA.training.train_starvla][INFO] - Step 1688, Loss: {'train/loss': 0.2153531312942505, 'train/grad_norm_pre_clip': 19.4244327545166, 'timing/data': 0.0002077300159726292, 'timing/model': 0.39046389001305215, 'learning_rate/qwen_vl_interface': 7.74354275555388e-06, 'learning_rate/action_model': 7.74354275555388e-05, 'epoch': 0.12}) +[2026-10-03 19:15:55,015][starVLA.training.train_starvla][INFO] - Step 1689, Loss: {'train/loss': 0.31793951988220215, 'train/grad_norm_pre_clip': 33.35020065307617, 'timing/data': 0.00020855001639574766, 'timing/model': 0.3923744499916211, 'learning_rate/qwen_vl_interface': 7.740950196296488e-06, 'learning_rate/action_model': 7.740950196296488e-05, 'epoch': 0.12}) +[2026-10-03 19:15:55,715][starVLA.training.train_starvla][INFO] - Step 1690, Loss: {'train/loss': 0.007586610037833452, 'train/grad_norm_pre_clip': 5.956806182861328, 'timing/data': 0.00020723999477922916, 'timing/model': 0.39153713002451696, 'learning_rate/qwen_vl_interface': 7.738356613103562e-06, 'learning_rate/action_model': 7.73835661310356e-05, 'epoch': 0.12}) +[2026-10-03 19:15:56,410][starVLA.training.train_starvla][INFO] - Step 1691, Loss: {'train/loss': 0.5366998910903931, 'train/grad_norm_pre_clip': 60.50909423828125, 'timing/data': 0.0002803699753712863, 'timing/model': 0.3893650000100024, 'learning_rate/qwen_vl_interface': 7.735762007041226e-06, 'learning_rate/action_model': 7.735762007041226e-05, 'epoch': 0.12}) +[2026-10-03 19:15:57,306][starVLA.training.train_starvla][INFO] - Step 1692, Loss: {'train/loss': 0.007891742512583733, 'train/grad_norm_pre_clip': 1.9873725175857544, 'timing/data': 0.00033266001264564693, 'timing/model': 0.5865654009976424, 'learning_rate/qwen_vl_interface': 7.733166379176023e-06, 'learning_rate/action_model': 7.733166379176023e-05, 'epoch': 0.12}) +[2026-10-03 19:15:58,003][starVLA.training.train_starvla][INFO] - Step 1693, Loss: {'train/loss': 0.26629728078842163, 'train/grad_norm_pre_clip': 56.91162872314453, 'timing/data': 0.00025865997304208577, 'timing/model': 0.3914953400089871, 'learning_rate/qwen_vl_interface': 7.730569730574921e-06, 'learning_rate/action_model': 7.73056973057492e-05, 'epoch': 0.12}) +[2026-10-03 19:15:58,698][starVLA.training.train_starvla][INFO] - Step 1694, Loss: {'train/loss': 0.42129838466644287, 'train/grad_norm_pre_clip': 39.9024772644043, 'timing/data': 0.0003170600102748722, 'timing/model': 0.3903920300072059, 'learning_rate/qwen_vl_interface': 7.727972062305299e-06, 'learning_rate/action_model': 7.727972062305299e-05, 'epoch': 0.12}) +[2026-10-03 19:15:59,393][starVLA.training.train_starvla][INFO] - Step 1695, Loss: {'train/loss': 0.15203383564949036, 'train/grad_norm_pre_clip': 22.44932746887207, 'timing/data': 0.00021088001085445285, 'timing/model': 0.3885276799846906, 'learning_rate/qwen_vl_interface': 7.725373375434965e-06, 'learning_rate/action_model': 7.725373375434965e-05, 'epoch': 0.12}) +[2026-10-03 19:16:00,089][starVLA.training.train_starvla][INFO] - Step 1696, Loss: {'train/loss': 0.025629941374063492, 'train/grad_norm_pre_clip': 9.136748313903809, 'timing/data': 0.00025198000366799533, 'timing/model': 0.3884529009810649, 'learning_rate/qwen_vl_interface': 7.722773671032141e-06, 'learning_rate/action_model': 7.722773671032142e-05, 'epoch': 0.12}) +[2026-10-03 19:16:00,991][starVLA.training.train_starvla][INFO] - Step 1697, Loss: {'train/loss': 0.05982957407832146, 'train/grad_norm_pre_clip': 87.98600769042969, 'timing/data': 0.00024104901240207255, 'timing/model': 0.5961689199903049, 'learning_rate/qwen_vl_interface': 7.720172950165466e-06, 'learning_rate/action_model': 7.720172950165465e-05, 'epoch': 0.12}) +[2026-10-03 19:16:01,689][starVLA.training.train_starvla][INFO] - Step 1698, Loss: {'train/loss': 0.0009681778610683978, 'train/grad_norm_pre_clip': 0.1812344640493393, 'timing/data': 0.0003589100087992847, 'timing/model': 0.39087766999728046, 'learning_rate/qwen_vl_interface': 7.717571213903996e-06, 'learning_rate/action_model': 7.717571213903996e-05, 'epoch': 0.12}) +[2026-10-03 19:16:02,386][starVLA.training.train_starvla][INFO] - Step 1699, Loss: {'train/loss': 0.0024767620489001274, 'train/grad_norm_pre_clip': 0.6971406936645508, 'timing/data': 0.00020537999807856977, 'timing/model': 0.3913607900030911, 'learning_rate/qwen_vl_interface': 7.714968463317207e-06, 'learning_rate/action_model': 7.714968463317207e-05, 'epoch': 0.12}) +[2026-10-03 19:16:03,085][starVLA.training.train_starvla][INFO] - Step 1700, Loss: {'train/loss': 0.2663215398788452, 'train/grad_norm_pre_clip': 95.03119659423828, 'timing/data': 0.00020947001758031547, 'timing/model': 0.39198765001492575, 'learning_rate/qwen_vl_interface': 7.712364699474996e-06, 'learning_rate/action_model': 7.712364699474996e-05, 'epoch': 0.12}) +[2026-10-03 19:16:03,782][starVLA.training.train_starvla][INFO] - Step 1701, Loss: {'train/loss': 0.6284536719322205, 'train/grad_norm_pre_clip': 132.37225341796875, 'timing/data': 0.00020891998428851366, 'timing/model': 0.39099969001836143, 'learning_rate/qwen_vl_interface': 7.709759923447667e-06, 'learning_rate/action_model': 7.709759923447665e-05, 'epoch': 0.12}) +[2026-10-03 19:16:04,686][starVLA.training.train_starvla][INFO] - Step 1702, Loss: {'train/loss': 0.10406465828418732, 'train/grad_norm_pre_clip': 32.534610748291016, 'timing/data': 0.0003517300065141171, 'timing/model': 0.5978359889995772, 'learning_rate/qwen_vl_interface': 7.707154136305945e-06, 'learning_rate/action_model': 7.707154136305945e-05, 'epoch': 0.12}) +[2026-10-03 19:16:05,385][starVLA.training.train_starvla][INFO] - Step 1703, Loss: {'train/loss': 0.0010530744912102818, 'train/grad_norm_pre_clip': 0.2892287075519562, 'timing/data': 0.00028080001357011497, 'timing/model': 0.391543029982131, 'learning_rate/qwen_vl_interface': 7.704547339120974e-06, 'learning_rate/action_model': 7.704547339120974e-05, 'epoch': 0.12}) +[2026-10-03 19:16:06,084][starVLA.training.train_starvla][INFO] - Step 1704, Loss: {'train/loss': 0.0006825380842201412, 'train/grad_norm_pre_clip': 0.1778523027896881, 'timing/data': 0.00021276000188663602, 'timing/model': 0.3899255910073407, 'learning_rate/qwen_vl_interface': 7.70193953296431e-06, 'learning_rate/action_model': 7.70193953296431e-05, 'epoch': 0.12}) +[2026-10-03 19:16:06,783][starVLA.training.train_starvla][INFO] - Step 1705, Loss: {'train/loss': 0.11700131744146347, 'train/grad_norm_pre_clip': 166.51686096191406, 'timing/data': 0.00020137999672442675, 'timing/model': 0.39194413999211974, 'learning_rate/qwen_vl_interface': 7.69933071890792e-06, 'learning_rate/action_model': 7.699330718907919e-05, 'epoch': 0.12}) +[2026-10-03 19:16:07,484][starVLA.training.train_starvla][INFO] - Step 1706, Loss: {'train/loss': 0.1023196130990982, 'train/grad_norm_pre_clip': 41.54671096801758, 'timing/data': 0.00021723000099882483, 'timing/model': 0.3920918710064143, 'learning_rate/qwen_vl_interface': 7.696720898024188e-06, 'learning_rate/action_model': 7.69672089802419e-05, 'epoch': 0.12}) +[2026-10-03 19:16:08,183][starVLA.training.train_starvla][INFO] - Step 1707, Loss: {'train/loss': 0.0864209309220314, 'train/grad_norm_pre_clip': 43.97331237792969, 'timing/data': 0.00020603000302799046, 'timing/model': 0.3915225409728009, 'learning_rate/qwen_vl_interface': 7.69411007138592e-06, 'learning_rate/action_model': 7.694110071385919e-05, 'epoch': 0.12}) +[2026-10-03 19:16:09,088][starVLA.training.train_starvla][INFO] - Step 1708, Loss: {'train/loss': 0.43637973070144653, 'train/grad_norm_pre_clip': 56.94160842895508, 'timing/data': 0.00020916000357829034, 'timing/model': 0.3886759700253606, 'learning_rate/qwen_vl_interface': 7.691498240066322e-06, 'learning_rate/action_model': 7.691498240066322e-05, 'epoch': 0.13}) +[2026-10-03 19:16:09,786][starVLA.training.train_starvla][INFO] - Step 1709, Loss: {'train/loss': 0.19109675288200378, 'train/grad_norm_pre_clip': 52.63288116455078, 'timing/data': 0.00020574001246131957, 'timing/model': 0.3920172200014349, 'learning_rate/qwen_vl_interface': 7.688885405139022e-06, 'learning_rate/action_model': 7.688885405139021e-05, 'epoch': 0.13}) +[2026-10-03 19:16:10,483][starVLA.training.train_starvla][INFO] - Step 1710, Loss: {'train/loss': 0.42380520701408386, 'train/grad_norm_pre_clip': 73.28614807128906, 'timing/data': 0.00036827000440098345, 'timing/model': 0.38923328000237234, 'learning_rate/qwen_vl_interface': 7.686271567678056e-06, 'learning_rate/action_model': 7.686271567678054e-05, 'epoch': 0.13}) +[2026-10-03 19:16:11,183][starVLA.training.train_starvla][INFO] - Step 1711, Loss: {'train/loss': 0.010407201014459133, 'train/grad_norm_pre_clip': 16.0930233001709, 'timing/data': 0.0003137199964839965, 'timing/model': 0.39142051100498065, 'learning_rate/qwen_vl_interface': 7.683656728757873e-06, 'learning_rate/action_model': 7.683656728757872e-05, 'epoch': 0.13}) +[2026-10-03 19:16:11,882][starVLA.training.train_starvla][INFO] - Step 1712, Loss: {'train/loss': 0.00890444591641426, 'train/grad_norm_pre_clip': 9.785892486572266, 'timing/data': 0.00021515999105758965, 'timing/model': 0.3913040900079068, 'learning_rate/qwen_vl_interface': 7.681040889453337e-06, 'learning_rate/action_model': 7.681040889453336e-05, 'epoch': 0.13}) +[2026-10-03 19:16:12,788][starVLA.training.train_starvla][INFO] - Step 1713, Loss: {'train/loss': 0.4438927173614502, 'train/grad_norm_pre_clip': 12.268332481384277, 'timing/data': 0.0002560699940659106, 'timing/model': 0.5969406899821479, 'learning_rate/qwen_vl_interface': 7.678424050839722e-06, 'learning_rate/action_model': 7.67842405083972e-05, 'epoch': 0.13}) +[2026-10-03 19:16:13,495][starVLA.training.train_starvla][INFO] - Step 1714, Loss: {'train/loss': 0.33824416995048523, 'train/grad_norm_pre_clip': 86.113525390625, 'timing/data': 0.00023053001496009529, 'timing/model': 0.3964811890036799, 'learning_rate/qwen_vl_interface': 7.675806213992706e-06, 'learning_rate/action_model': 7.675806213992707e-05, 'epoch': 0.13}) +[2026-10-03 19:16:14,203][starVLA.training.train_starvla][INFO] - Step 1715, Loss: {'train/loss': 0.4749962389469147, 'train/grad_norm_pre_clip': 112.8840103149414, 'timing/data': 0.000357509998138994, 'timing/model': 0.3956017089949455, 'learning_rate/qwen_vl_interface': 7.673187379988388e-06, 'learning_rate/action_model': 7.673187379988388e-05, 'epoch': 0.13}) +[2026-10-03 19:16:14,912][starVLA.training.train_starvla][INFO] - Step 1716, Loss: {'train/loss': 0.006824925541877747, 'train/grad_norm_pre_clip': 1.774533748626709, 'timing/data': 0.00034220999805256724, 'timing/model': 0.39597670899820514, 'learning_rate/qwen_vl_interface': 7.67056754990327e-06, 'learning_rate/action_model': 7.670567549903269e-05, 'epoch': 0.13}) +[2026-10-03 19:16:15,623][starVLA.training.train_starvla][INFO] - Step 1717, Loss: {'train/loss': 0.01695336401462555, 'train/grad_norm_pre_clip': 22.112045288085938, 'timing/data': 0.00030872999923303723, 'timing/model': 0.3972497890063096, 'learning_rate/qwen_vl_interface': 7.667946724814264e-06, 'learning_rate/action_model': 7.667946724814265e-05, 'epoch': 0.13}) +[2026-10-03 19:16:16,330][starVLA.training.train_starvla][INFO] - Step 1718, Loss: {'train/loss': 0.003431684337556362, 'train/grad_norm_pre_clip': 0.7810242176055908, 'timing/data': 0.00036877000820823014, 'timing/model': 0.3945984000165481, 'learning_rate/qwen_vl_interface': 7.665324905798694e-06, 'learning_rate/action_model': 7.665324905798695e-05, 'epoch': 0.13}) +[2026-10-03 19:16:17,235][starVLA.training.train_starvla][INFO] - Step 1719, Loss: {'train/loss': 0.1107686385512352, 'train/grad_norm_pre_clip': 54.704315185546875, 'timing/data': 0.0003482890024315566, 'timing/model': 0.39440598999499343, 'learning_rate/qwen_vl_interface': 7.66270209393429e-06, 'learning_rate/action_model': 7.66270209393429e-05, 'epoch': 0.13}) +[2026-10-03 19:16:17,943][starVLA.training.train_starvla][INFO] - Step 1720, Loss: {'train/loss': 0.09257995337247849, 'train/grad_norm_pre_clip': 127.71015167236328, 'timing/data': 0.0003139199980068952, 'timing/model': 0.3942146199988201, 'learning_rate/qwen_vl_interface': 7.66007829029919e-06, 'learning_rate/action_model': 7.660078290299191e-05, 'epoch': 0.13}) +[2026-10-03 19:16:18,651][starVLA.training.train_starvla][INFO] - Step 1721, Loss: {'train/loss': 0.1682269275188446, 'train/grad_norm_pre_clip': 166.3529510498047, 'timing/data': 0.00024659998598508537, 'timing/model': 0.39668150900979526, 'learning_rate/qwen_vl_interface': 7.657453495971942e-06, 'learning_rate/action_model': 7.657453495971943e-05, 'epoch': 0.13}) +[2026-10-03 19:16:19,357][starVLA.training.train_starvla][INFO] - Step 1722, Loss: {'train/loss': 0.004076431505382061, 'train/grad_norm_pre_clip': 3.9040119647979736, 'timing/data': 0.00025497001479379833, 'timing/model': 0.39449907999369316, 'learning_rate/qwen_vl_interface': 7.6548277120315e-06, 'learning_rate/action_model': 7.6548277120315e-05, 'epoch': 0.13}) +[2026-10-03 19:16:20,064][starVLA.training.train_starvla][INFO] - Step 1723, Loss: {'train/loss': 0.4441456198692322, 'train/grad_norm_pre_clip': 52.26716995239258, 'timing/data': 0.00022481000632978976, 'timing/model': 0.39601044001756236, 'learning_rate/qwen_vl_interface': 7.652200939557225e-06, 'learning_rate/action_model': 7.652200939557225e-05, 'epoch': 0.13}) +[2026-10-03 19:16:20,975][starVLA.training.train_starvla][INFO] - Step 1724, Loss: {'train/loss': 0.001113151665776968, 'train/grad_norm_pre_clip': 2.212322950363159, 'timing/data': 0.0003912700049113482, 'timing/model': 0.394533960003173, 'learning_rate/qwen_vl_interface': 7.649573179628881e-06, 'learning_rate/action_model': 7.649573179628881e-05, 'epoch': 0.13}) +[2026-10-03 19:16:21,684][starVLA.training.train_starvla][INFO] - Step 1725, Loss: {'train/loss': 0.005013167392462492, 'train/grad_norm_pre_clip': 7.199469089508057, 'timing/data': 0.0003398600092623383, 'timing/model': 0.3969022399978712, 'learning_rate/qwen_vl_interface': 7.646944433326645e-06, 'learning_rate/action_model': 7.646944433326643e-05, 'epoch': 0.13}) +[2026-10-03 19:16:22,392][starVLA.training.train_starvla][INFO] - Step 1726, Loss: {'train/loss': 0.8740624189376831, 'train/grad_norm_pre_clip': 37.00333023071289, 'timing/data': 0.0003225299878977239, 'timing/model': 0.39596300999983214, 'learning_rate/qwen_vl_interface': 7.64431470173109e-06, 'learning_rate/action_model': 7.64431470173109e-05, 'epoch': 0.13}) +[2026-10-03 19:16:23,099][starVLA.training.train_starvla][INFO] - Step 1727, Loss: {'train/loss': 0.00010071317228721455, 'train/grad_norm_pre_clip': 0.017530690878629684, 'timing/data': 0.0002978699922095984, 'timing/model': 0.3943051600072067, 'learning_rate/qwen_vl_interface': 7.641683985923203e-06, 'learning_rate/action_model': 7.641683985923203e-05, 'epoch': 0.13}) +[2026-10-03 19:16:23,805][starVLA.training.train_starvla][INFO] - Step 1728, Loss: {'train/loss': 0.05858834460377693, 'train/grad_norm_pre_clip': 86.86888885498047, 'timing/data': 0.0003413399972487241, 'timing/model': 0.3952597389870789, 'learning_rate/qwen_vl_interface': 7.63905228698437e-06, 'learning_rate/action_model': 7.63905228698437e-05, 'epoch': 0.13}) +[2026-10-03 19:16:24,712][starVLA.training.train_starvla][INFO] - Step 1729, Loss: {'train/loss': 0.022288396954536438, 'train/grad_norm_pre_clip': 47.22692108154297, 'timing/data': 0.00035536999348551035, 'timing/model': 0.5937239499762654, 'learning_rate/qwen_vl_interface': 7.636419605996385e-06, 'learning_rate/action_model': 7.636419605996384e-05, 'epoch': 0.13}) +[2026-10-03 19:16:25,420][starVLA.training.train_starvla][INFO] - Step 1730, Loss: {'train/loss': 0.2955608665943146, 'train/grad_norm_pre_clip': 36.4898567199707, 'timing/data': 0.0003306900034658611, 'timing/model': 0.3953317900013644, 'learning_rate/qwen_vl_interface': 7.633785944041444e-06, 'learning_rate/action_model': 7.633785944041444e-05, 'epoch': 0.13}) +[2026-10-03 19:16:26,128][starVLA.training.train_starvla][INFO] - Step 1731, Loss: {'train/loss': 0.12008155137300491, 'train/grad_norm_pre_clip': 72.21868133544922, 'timing/data': 0.0003231499867979437, 'timing/model': 0.3954977300018072, 'learning_rate/qwen_vl_interface': 7.631151302202145e-06, 'learning_rate/action_model': 7.631151302202144e-05, 'epoch': 0.13}) +[2026-10-03 19:16:26,834][starVLA.training.train_starvla][INFO] - Step 1732, Loss: {'train/loss': 0.2656039893627167, 'train/grad_norm_pre_clip': 176.5287628173828, 'timing/data': 0.00031743000727146864, 'timing/model': 0.394506820011884, 'learning_rate/qwen_vl_interface': 7.628515681561488e-06, 'learning_rate/action_model': 7.628515681561487e-05, 'epoch': 0.13}) +[2026-10-03 19:16:27,541][starVLA.training.train_starvla][INFO] - Step 1733, Loss: {'train/loss': 0.3763216435909271, 'train/grad_norm_pre_clip': 149.81385803222656, 'timing/data': 0.00022829000954516232, 'timing/model': 0.3961429490009323, 'learning_rate/qwen_vl_interface': 7.62587908320288e-06, 'learning_rate/action_model': 7.62587908320288e-05, 'epoch': 0.13}) +[2026-10-03 19:16:28,248][starVLA.training.train_starvla][INFO] - Step 1734, Loss: {'train/loss': 0.05161343887448311, 'train/grad_norm_pre_clip': 64.01626586914062, 'timing/data': 0.0004594599886331707, 'timing/model': 0.3937192700104788, 'learning_rate/qwen_vl_interface': 7.6232415082101255e-06, 'learning_rate/action_model': 7.623241508210124e-05, 'epoch': 0.13}) +[2026-10-03 19:16:29,154][starVLA.training.train_starvla][INFO] - Step 1735, Loss: {'train/loss': 0.0995800793170929, 'train/grad_norm_pre_clip': 58.56524658203125, 'timing/data': 0.00022633999469690025, 'timing/model': 0.3944845189980697, 'learning_rate/qwen_vl_interface': 7.620602957667432e-06, 'learning_rate/action_model': 7.620602957667432e-05, 'epoch': 0.13}) +[2026-10-03 19:16:29,861][starVLA.training.train_starvla][INFO] - Step 1736, Loss: {'train/loss': 0.21935723721981049, 'train/grad_norm_pre_clip': 65.67862701416016, 'timing/data': 0.0003409000055398792, 'timing/model': 0.39388406998477876, 'learning_rate/qwen_vl_interface': 7.617963432659408e-06, 'learning_rate/action_model': 7.617963432659408e-05, 'epoch': 0.13}) +[2026-10-03 19:16:30,567][starVLA.training.train_starvla][INFO] - Step 1737, Loss: {'train/loss': 0.005357168149203062, 'train/grad_norm_pre_clip': 1.0860052108764648, 'timing/data': 0.0003342400013934821, 'timing/model': 0.39405369001906365, 'learning_rate/qwen_vl_interface': 7.615322934271063e-06, 'learning_rate/action_model': 7.615322934271064e-05, 'epoch': 0.13}) +[2026-10-03 19:16:31,274][starVLA.training.train_starvla][INFO] - Step 1738, Loss: {'train/loss': 0.01950220949947834, 'train/grad_norm_pre_clip': 49.46854782104492, 'timing/data': 0.000342229992384091, 'timing/model': 0.3941379900206812, 'learning_rate/qwen_vl_interface': 7.6126814635878055e-06, 'learning_rate/action_model': 7.612681463587805e-05, 'epoch': 0.13}) +[2026-10-03 19:16:31,982][starVLA.training.train_starvla][INFO] - Step 1739, Loss: {'train/loss': 0.007447049953043461, 'train/grad_norm_pre_clip': 3.841219663619995, 'timing/data': 0.00032277998980134726, 'timing/model': 0.3955149099929258, 'learning_rate/qwen_vl_interface': 7.6100390216954445e-06, 'learning_rate/action_model': 7.610039021695445e-05, 'epoch': 0.13}) +[2026-10-03 19:16:32,878][starVLA.training.train_starvla][INFO] - Step 1740, Loss: {'train/loss': 0.3829221725463867, 'train/grad_norm_pre_clip': 167.51206970214844, 'timing/data': 0.000331689981976524, 'timing/model': 0.3946215300238691, 'learning_rate/qwen_vl_interface': 7.607395609680189e-06, 'learning_rate/action_model': 7.60739560968019e-05, 'epoch': 0.13}) +[2026-10-03 19:16:33,584][starVLA.training.train_starvla][INFO] - Step 1741, Loss: {'train/loss': 0.002435168717056513, 'train/grad_norm_pre_clip': 0.8388519883155823, 'timing/data': 0.00033209999673999846, 'timing/model': 0.39599767001345754, 'learning_rate/qwen_vl_interface': 7.604751228628644e-06, 'learning_rate/action_model': 7.604751228628644e-05, 'epoch': 0.13}) +[2026-10-03 19:16:34,292][starVLA.training.train_starvla][INFO] - Step 1742, Loss: {'train/loss': 0.6461505889892578, 'train/grad_norm_pre_clip': 62.76655578613281, 'timing/data': 0.0003464600013103336, 'timing/model': 0.39562306000152603, 'learning_rate/qwen_vl_interface': 7.602105879627814e-06, 'learning_rate/action_model': 7.602105879627814e-05, 'epoch': 0.13}) +[2026-10-03 19:16:35,001][starVLA.training.train_starvla][INFO] - Step 1743, Loss: {'train/loss': 0.006852488964796066, 'train/grad_norm_pre_clip': 8.831512451171875, 'timing/data': 0.0002442399854771793, 'timing/model': 0.3960264399938751, 'learning_rate/qwen_vl_interface': 7.599459563765107e-06, 'learning_rate/action_model': 7.599459563765106e-05, 'epoch': 0.13}) +[2026-10-03 19:16:35,708][starVLA.training.train_starvla][INFO] - Step 1744, Loss: {'train/loss': 0.12075802683830261, 'train/grad_norm_pre_clip': 68.91266632080078, 'timing/data': 0.0003273600013926625, 'timing/model': 0.3949735500209499, 'learning_rate/qwen_vl_interface': 7.596812282128319e-06, 'learning_rate/action_model': 7.596812282128319e-05, 'epoch': 0.13}) +[2026-10-03 19:16:36,617][starVLA.training.train_starvla][INFO] - Step 1745, Loss: {'train/loss': 0.012203452177345753, 'train/grad_norm_pre_clip': 9.161746978759766, 'timing/data': 0.0003340399998705834, 'timing/model': 0.5961988480121363, 'learning_rate/qwen_vl_interface': 7.594164035805648e-06, 'learning_rate/action_model': 7.594164035805648e-05, 'epoch': 0.13}) +[2026-10-03 19:16:37,324][starVLA.training.train_starvla][INFO] - Step 1746, Loss: {'train/loss': 0.0006653012242168188, 'train/grad_norm_pre_clip': 0.23810459673404694, 'timing/data': 0.00035136000951752067, 'timing/model': 0.396015289996285, 'learning_rate/qwen_vl_interface': 7.59151482588569e-06, 'learning_rate/action_model': 7.59151482588569e-05, 'epoch': 0.13}) +[2026-10-03 19:16:38,033][starVLA.training.train_starvla][INFO] - Step 1747, Loss: {'train/loss': 0.004981473553925753, 'train/grad_norm_pre_clip': 7.74740743637085, 'timing/data': 0.0003514800046104938, 'timing/model': 0.39597318999585696, 'learning_rate/qwen_vl_interface': 7.588864653457433e-06, 'learning_rate/action_model': 7.588864653457432e-05, 'epoch': 0.13}) +[2026-10-03 19:16:38,739][starVLA.training.train_starvla][INFO] - Step 1748, Loss: {'train/loss': 0.255372017621994, 'train/grad_norm_pre_clip': 25.649736404418945, 'timing/data': 0.0002294000005349517, 'timing/model': 0.39477882999926805, 'learning_rate/qwen_vl_interface': 7.5862135196102634e-06, 'learning_rate/action_model': 7.586213519610263e-05, 'epoch': 0.13}) +[2026-10-03 19:16:39,446][starVLA.training.train_starvla][INFO] - Step 1749, Loss: {'train/loss': 0.006934755481779575, 'train/grad_norm_pre_clip': 21.70730209350586, 'timing/data': 0.000355829979525879, 'timing/model': 0.396228230005363, 'learning_rate/qwen_vl_interface': 7.583561425433962e-06, 'learning_rate/action_model': 7.583561425433962e-05, 'epoch': 0.13}) +[2026-10-03 19:16:40,353][starVLA.training.train_starvla][INFO] - Step 1750, Loss: {'train/loss': 0.18309511244297028, 'train/grad_norm_pre_clip': 199.39968872070312, 'timing/data': 0.0003462300228420645, 'timing/model': 0.5936518200032879, 'learning_rate/qwen_vl_interface': 7.580908372018706e-06, 'learning_rate/action_model': 7.580908372018705e-05, 'epoch': 0.13}) +[2026-10-03 19:16:41,061][starVLA.training.train_starvla][INFO] - Step 1751, Loss: {'train/loss': 0.00038649674388580024, 'train/grad_norm_pre_clip': 0.03978308290243149, 'timing/data': 0.0003228199784643948, 'timing/model': 0.39574643000378273, 'learning_rate/qwen_vl_interface': 7.5782543604550655e-06, 'learning_rate/action_model': 7.578254360455064e-05, 'epoch': 0.13}) +[2026-10-03 19:16:41,770][starVLA.training.train_starvla][INFO] - Step 1752, Loss: {'train/loss': 0.36864155530929565, 'train/grad_norm_pre_clip': 78.4771499633789, 'timing/data': 0.00033271999564021826, 'timing/model': 0.3957439100195188, 'learning_rate/qwen_vl_interface': 7.575599391834e-06, 'learning_rate/action_model': 7.575599391834e-05, 'epoch': 0.13}) +[2026-10-03 19:16:42,477][starVLA.training.train_starvla][INFO] - Step 1753, Loss: {'train/loss': 0.410131573677063, 'train/grad_norm_pre_clip': 128.506591796875, 'timing/data': 0.0003341799892950803, 'timing/model': 0.3946900899754837, 'learning_rate/qwen_vl_interface': 7.572943467246871e-06, 'learning_rate/action_model': 7.57294346724687e-05, 'epoch': 0.13}) +[2026-10-03 19:16:43,185][starVLA.training.train_starvla][INFO] - Step 1754, Loss: {'train/loss': 0.008591067977249622, 'train/grad_norm_pre_clip': 27.665977478027344, 'timing/data': 0.00022494897712022066, 'timing/model': 0.39550680000684224, 'learning_rate/qwen_vl_interface': 7.570286587785428e-06, 'learning_rate/action_model': 7.570286587785428e-05, 'epoch': 0.13}) +[2026-10-03 19:16:44,093][starVLA.training.train_starvla][INFO] - Step 1755, Loss: {'train/loss': 0.3667818605899811, 'train/grad_norm_pre_clip': 51.110862731933594, 'timing/data': 0.0002881899999920279, 'timing/model': 0.5961642000183929, 'learning_rate/qwen_vl_interface': 7.567628754541815e-06, 'learning_rate/action_model': 7.567628754541814e-05, 'epoch': 0.13}) +[2026-10-03 19:16:44,800][starVLA.training.train_starvla][INFO] - Step 1756, Loss: {'train/loss': 0.0004436464514583349, 'train/grad_norm_pre_clip': 0.18753735721111298, 'timing/data': 0.0003097199951298535, 'timing/model': 0.3958854300144594, 'learning_rate/qwen_vl_interface': 7.564969968608565e-06, 'learning_rate/action_model': 7.564969968608565e-05, 'epoch': 0.13}) +[2026-10-03 19:16:45,508][starVLA.training.train_starvla][INFO] - Step 1757, Loss: {'train/loss': 0.06642676144838333, 'train/grad_norm_pre_clip': 193.28622436523438, 'timing/data': 0.0003002499870490283, 'timing/model': 0.39670580899110064, 'learning_rate/qwen_vl_interface': 7.562310231078605e-06, 'learning_rate/action_model': 7.562310231078604e-05, 'epoch': 0.13}) +[2026-10-03 19:16:46,216][starVLA.training.train_starvla][INFO] - Step 1758, Loss: {'train/loss': 0.38708221912384033, 'train/grad_norm_pre_clip': 127.94548797607422, 'timing/data': 0.0003740499960258603, 'timing/model': 0.39580565999494866, 'learning_rate/qwen_vl_interface': 7.559649543045252e-06, 'learning_rate/action_model': 7.559649543045252e-05, 'epoch': 0.13}) +[2026-10-03 19:16:46,924][starVLA.training.train_starvla][INFO] - Step 1759, Loss: {'train/loss': 0.3258870542049408, 'train/grad_norm_pre_clip': 15.28333568572998, 'timing/data': 0.0002288299729116261, 'timing/model': 0.396007550007198, 'learning_rate/qwen_vl_interface': 7.556987905602215e-06, 'learning_rate/action_model': 7.556987905602215e-05, 'epoch': 0.13}) +[2026-10-03 19:16:47,631][starVLA.training.train_starvla][INFO] - Step 1760, Loss: {'train/loss': 1.096831202507019, 'train/grad_norm_pre_clip': 68.329833984375, 'timing/data': 0.00033680000342428684, 'timing/model': 0.3957216389826499, 'learning_rate/qwen_vl_interface': 7.554325319843593e-06, 'learning_rate/action_model': 7.554325319843592e-05, 'epoch': 0.13}) +[2026-10-03 19:16:48,553][starVLA.training.train_starvla][INFO] - Step 1761, Loss: {'train/loss': 0.33235862851142883, 'train/grad_norm_pre_clip': 57.256507873535156, 'timing/data': 0.00038102001417428255, 'timing/model': 0.39666969000245444, 'learning_rate/qwen_vl_interface': 7.551661786863874e-06, 'learning_rate/action_model': 7.551661786863874e-05, 'epoch': 0.13}) +[2026-10-03 19:16:49,259][starVLA.training.train_starvla][INFO] - Step 1762, Loss: {'train/loss': 0.36692142486572266, 'train/grad_norm_pre_clip': 11.848873138427734, 'timing/data': 0.0004181099939160049, 'timing/model': 0.3943736390210688, 'learning_rate/qwen_vl_interface': 7.548997307757936e-06, 'learning_rate/action_model': 7.548997307757935e-05, 'epoch': 0.13}) +[2026-10-03 19:16:49,967][starVLA.training.train_starvla][INFO] - Step 1763, Loss: {'train/loss': 0.07968181371688843, 'train/grad_norm_pre_clip': 67.70537567138672, 'timing/data': 0.00023319001775234938, 'timing/model': 0.39447292999830097, 'learning_rate/qwen_vl_interface': 7.546331883621044e-06, 'learning_rate/action_model': 7.546331883621044e-05, 'epoch': 0.13}) +[2026-10-03 19:16:50,675][starVLA.training.train_starvla][INFO] - Step 1764, Loss: {'train/loss': 0.002492932602763176, 'train/grad_norm_pre_clip': 2.933847188949585, 'timing/data': 0.00022258999524638057, 'timing/model': 0.39526027999818325, 'learning_rate/qwen_vl_interface': 7.543665515548857e-06, 'learning_rate/action_model': 7.543665515548856e-05, 'epoch': 0.13}) +[2026-10-03 19:16:51,384][starVLA.training.train_starvla][INFO] - Step 1765, Loss: {'train/loss': 0.0011900949757546186, 'train/grad_norm_pre_clip': 0.2830606997013092, 'timing/data': 0.00030525997863151133, 'timing/model': 0.39632960001472384, 'learning_rate/qwen_vl_interface': 7.540998204637413e-06, 'learning_rate/action_model': 7.540998204637413e-05, 'epoch': 0.13}) +[2026-10-03 19:16:52,290][starVLA.training.train_starvla][INFO] - Step 1766, Loss: {'train/loss': 0.1881403625011444, 'train/grad_norm_pre_clip': 80.06536102294922, 'timing/data': 0.00023108001914806664, 'timing/model': 0.5942022099916358, 'learning_rate/qwen_vl_interface': 7.538329951983148e-06, 'learning_rate/action_model': 7.538329951983147e-05, 'epoch': 0.13}) +[2026-10-03 19:16:52,998][starVLA.training.train_starvla][INFO] - Step 1767, Loss: {'train/loss': 0.010366052389144897, 'train/grad_norm_pre_clip': 4.704258918762207, 'timing/data': 0.0003492899995762855, 'timing/model': 0.39548231998924166, 'learning_rate/qwen_vl_interface': 7.535660758682877e-06, 'learning_rate/action_model': 7.535660758682876e-05, 'epoch': 0.13}) +[2026-10-03 19:16:53,705][starVLA.training.train_starvla][INFO] - Step 1768, Loss: {'train/loss': 0.00032706034835428, 'train/grad_norm_pre_clip': 0.07443300634622574, 'timing/data': 0.0002860899840015918, 'timing/model': 0.3955224499804899, 'learning_rate/qwen_vl_interface': 7.532990625833806e-06, 'learning_rate/action_model': 7.532990625833805e-05, 'epoch': 0.13}) +[2026-10-03 19:16:54,413][starVLA.training.train_starvla][INFO] - Step 1769, Loss: {'train/loss': 0.004261215683072805, 'train/grad_norm_pre_clip': 0.43109455704689026, 'timing/data': 0.0003887699858751148, 'timing/model': 0.39619271000265144, 'learning_rate/qwen_vl_interface': 7.5303195545335225e-06, 'learning_rate/action_model': 7.530319554533522e-05, 'epoch': 0.13}) +[2026-10-03 19:16:55,119][starVLA.training.train_starvla][INFO] - Step 1770, Loss: {'train/loss': 0.5187441110610962, 'train/grad_norm_pre_clip': 111.81321716308594, 'timing/data': 0.00037675999919883907, 'timing/model': 0.3945087699976284, 'learning_rate/qwen_vl_interface': 7.527647545880007e-06, 'learning_rate/action_model': 7.527647545880007e-05, 'epoch': 0.13}) +[2026-10-03 19:16:55,825][starVLA.training.train_starvla][INFO] - Step 1771, Loss: {'train/loss': 0.9525946378707886, 'train/grad_norm_pre_clip': 63.46907424926758, 'timing/data': 0.00022570000146515667, 'timing/model': 0.3956510489806533, 'learning_rate/qwen_vl_interface': 7.52497460097162e-06, 'learning_rate/action_model': 7.52497460097162e-05, 'epoch': 0.13}) +[2026-10-03 19:16:56,731][starVLA.training.train_starvla][INFO] - Step 1772, Loss: {'train/loss': 0.010714640840888023, 'train/grad_norm_pre_clip': 4.027042388916016, 'timing/data': 0.0003633599844761193, 'timing/model': 0.3945370299916249, 'learning_rate/qwen_vl_interface': 7.522300720907105e-06, 'learning_rate/action_model': 7.522300720907105e-05, 'epoch': 0.13}) +[2026-10-03 19:16:57,439][starVLA.training.train_starvla][INFO] - Step 1773, Loss: {'train/loss': 0.15911166369915009, 'train/grad_norm_pre_clip': 51.286415100097656, 'timing/data': 0.0003622799995355308, 'timing/model': 0.39673319898429327, 'learning_rate/qwen_vl_interface': 7.519625906785601e-06, 'learning_rate/action_model': 7.5196259067856e-05, 'epoch': 0.13}) +[2026-10-03 19:16:58,146][starVLA.training.train_starvla][INFO] - Step 1774, Loss: {'train/loss': 0.1339731514453888, 'train/grad_norm_pre_clip': 35.73607635498047, 'timing/data': 0.0003925600030925125, 'timing/model': 0.39477590000024065, 'learning_rate/qwen_vl_interface': 7.516950159706614e-06, 'learning_rate/action_model': 7.516950159706614e-05, 'epoch': 0.13}) +[2026-10-03 19:16:58,853][starVLA.training.train_starvla][INFO] - Step 1775, Loss: {'train/loss': 0.06991080939769745, 'train/grad_norm_pre_clip': 38.44429016113281, 'timing/data': 0.0003318000235594809, 'timing/model': 0.3939123910095077, 'learning_rate/qwen_vl_interface': 7.514273480770047e-06, 'learning_rate/action_model': 7.514273480770047e-05, 'epoch': 0.13}) +[2026-10-03 19:16:59,560][starVLA.training.train_starvla][INFO] - Step 1776, Loss: {'train/loss': 0.19936545193195343, 'train/grad_norm_pre_clip': 88.35394287109375, 'timing/data': 0.000341029983246699, 'timing/model': 0.39571593000437133, 'learning_rate/qwen_vl_interface': 7.511595871076181e-06, 'learning_rate/action_model': 7.511595871076181e-05, 'epoch': 0.13}) +[2026-10-03 19:17:00,470][starVLA.training.train_starvla][INFO] - Step 1777, Loss: {'train/loss': 0.4156281054019928, 'train/grad_norm_pre_clip': 106.05278015136719, 'timing/data': 0.00027253999724052846, 'timing/model': 0.39597538000089116, 'learning_rate/qwen_vl_interface': 7.50891733172568e-06, 'learning_rate/action_model': 7.50891733172568e-05, 'epoch': 0.13}) +[2026-10-03 19:17:01,178][starVLA.training.train_starvla][INFO] - Step 1778, Loss: {'train/loss': 0.0005905821453779936, 'train/grad_norm_pre_clip': 0.2805701196193695, 'timing/data': 0.00031457998557016253, 'timing/model': 0.3953449609980453, 'learning_rate/qwen_vl_interface': 7.506237863819593e-06, 'learning_rate/action_model': 7.506237863819593e-05, 'epoch': 0.13}) +[2026-10-03 19:17:01,884][starVLA.training.train_starvla][INFO] - Step 1779, Loss: {'train/loss': 0.5000544190406799, 'train/grad_norm_pre_clip': 61.21047592163086, 'timing/data': 0.00021952998940832913, 'timing/model': 0.39388776998384856, 'learning_rate/qwen_vl_interface': 7.503557468459345e-06, 'learning_rate/action_model': 7.503557468459344e-05, 'epoch': 0.13}) +[2026-10-03 19:17:02,590][starVLA.training.train_starvla][INFO] - Step 1780, Loss: {'train/loss': 0.1833556890487671, 'train/grad_norm_pre_clip': 141.1942596435547, 'timing/data': 0.000334270007442683, 'timing/model': 0.39538901997730136, 'learning_rate/qwen_vl_interface': 7.5008761467467435e-06, 'learning_rate/action_model': 7.500876146746743e-05, 'epoch': 0.13}) +[2026-10-03 19:17:03,298][starVLA.training.train_starvla][INFO] - Step 1781, Loss: {'train/loss': 0.04291949048638344, 'train/grad_norm_pre_clip': 44.5673713684082, 'timing/data': 0.0003421599976718426, 'timing/model': 0.3946963300113566, 'learning_rate/qwen_vl_interface': 7.498193899783983e-06, 'learning_rate/action_model': 7.498193899783983e-05, 'epoch': 0.13}) +[2026-10-03 19:17:04,203][starVLA.training.train_starvla][INFO] - Step 1782, Loss: {'train/loss': 0.0069367713294923306, 'train/grad_norm_pre_clip': 6.449868679046631, 'timing/data': 0.0002408500004094094, 'timing/model': 0.5932285100279842, 'learning_rate/qwen_vl_interface': 7.495510728673631e-06, 'learning_rate/action_model': 7.495510728673632e-05, 'epoch': 0.13}) +[2026-10-03 19:17:04,908][starVLA.training.train_starvla][INFO] - Step 1783, Loss: {'train/loss': 0.004013126716017723, 'train/grad_norm_pre_clip': 4.19597864151001, 'timing/data': 0.000222209986532107, 'timing/model': 0.39338613898144104, 'learning_rate/qwen_vl_interface': 7.492826634518642e-06, 'learning_rate/action_model': 7.492826634518641e-05, 'epoch': 0.13}) +[2026-10-03 19:17:05,612][starVLA.training.train_starvla][INFO] - Step 1784, Loss: {'train/loss': 0.003175978548824787, 'train/grad_norm_pre_clip': 1.2496639490127563, 'timing/data': 0.0003600000054575503, 'timing/model': 0.3949059099832084, 'learning_rate/qwen_vl_interface': 7.4901416184223404e-06, 'learning_rate/action_model': 7.49014161842234e-05, 'epoch': 0.13}) +[2026-10-03 19:17:06,317][starVLA.training.train_starvla][INFO] - Step 1785, Loss: {'train/loss': 0.22652597725391388, 'train/grad_norm_pre_clip': 13.127731323242188, 'timing/data': 0.00034465998760424554, 'timing/model': 0.39394762000301853, 'learning_rate/qwen_vl_interface': 7.487455681488437e-06, 'learning_rate/action_model': 7.487455681488436e-05, 'epoch': 0.13}) +[2026-10-03 19:17:07,021][starVLA.training.train_starvla][INFO] - Step 1786, Loss: {'train/loss': 0.028053047135472298, 'train/grad_norm_pre_clip': 17.949926376342773, 'timing/data': 0.000365059997420758, 'timing/model': 0.39348351999069564, 'learning_rate/qwen_vl_interface': 7.484768824821018e-06, 'learning_rate/action_model': 7.484768824821018e-05, 'epoch': 0.13}) +[2026-10-03 19:17:07,727][starVLA.training.train_starvla][INFO] - Step 1787, Loss: {'train/loss': 0.012793798930943012, 'train/grad_norm_pre_clip': 4.36764669418335, 'timing/data': 0.00044423999497666955, 'timing/model': 0.39383292000275105, 'learning_rate/qwen_vl_interface': 7.482081049524551e-06, 'learning_rate/action_model': 7.482081049524551e-05, 'epoch': 0.13}) +[2026-10-03 19:17:08,629][starVLA.training.train_starvla][INFO] - Step 1788, Loss: {'train/loss': 0.6727064847946167, 'train/grad_norm_pre_clip': 33.33132553100586, 'timing/data': 0.00044890999561175704, 'timing/model': 0.39285285997902974, 'learning_rate/qwen_vl_interface': 7.479392356703878e-06, 'learning_rate/action_model': 7.479392356703878e-05, 'epoch': 0.13}) +[2026-10-03 19:17:09,335][starVLA.training.train_starvla][INFO] - Step 1789, Loss: {'train/loss': 0.0049317460507154465, 'train/grad_norm_pre_clip': 0.8557764887809753, 'timing/data': 0.0003419799904804677, 'timing/model': 0.3942926099989563, 'learning_rate/qwen_vl_interface': 7.476702747464219e-06, 'learning_rate/action_model': 7.476702747464219e-05, 'epoch': 0.13}) +[2026-10-03 19:17:10,039][starVLA.training.train_starvla][INFO] - Step 1790, Loss: {'train/loss': 0.2924951910972595, 'train/grad_norm_pre_clip': 55.963623046875, 'timing/data': 0.0002442999975755811, 'timing/model': 0.3937032200046815, 'learning_rate/qwen_vl_interface': 7.474012222911169e-06, 'learning_rate/action_model': 7.474012222911169e-05, 'epoch': 0.13}) +[2026-10-03 19:17:10,744][starVLA.training.train_starvla][INFO] - Step 1791, Loss: {'train/loss': 0.39425408840179443, 'train/grad_norm_pre_clip': 65.7855224609375, 'timing/data': 0.0003829399938695133, 'timing/model': 0.3938819899922237, 'learning_rate/qwen_vl_interface': 7.471320784150703e-06, 'learning_rate/action_model': 7.471320784150703e-05, 'epoch': 0.13}) +[2026-10-03 19:17:11,450][starVLA.training.train_starvla][INFO] - Step 1792, Loss: {'train/loss': 0.0009116394794546068, 'train/grad_norm_pre_clip': 0.4485008120536804, 'timing/data': 0.0003192000149283558, 'timing/model': 0.39420899999095127, 'learning_rate/qwen_vl_interface': 7.468628432289169e-06, 'learning_rate/action_model': 7.468628432289168e-05, 'epoch': 0.13}) +[2026-10-03 19:17:12,358][starVLA.training.train_starvla][INFO] - Step 1793, Loss: {'train/loss': 0.009720830246806145, 'train/grad_norm_pre_clip': 5.463156223297119, 'timing/data': 0.00037093000719323754, 'timing/model': 0.3944166100118309, 'learning_rate/qwen_vl_interface': 7.465935168433292e-06, 'learning_rate/action_model': 7.465935168433292e-05, 'epoch': 0.13}) +[2026-10-03 19:17:13,063][starVLA.training.train_starvla][INFO] - Step 1794, Loss: {'train/loss': 0.18087154626846313, 'train/grad_norm_pre_clip': 128.9498748779297, 'timing/data': 0.00035757001023739576, 'timing/model': 0.3938202589924913, 'learning_rate/qwen_vl_interface': 7.463240993690168e-06, 'learning_rate/action_model': 7.463240993690168e-05, 'epoch': 0.13}) +[2026-10-03 19:17:13,768][starVLA.training.train_starvla][INFO] - Step 1795, Loss: {'train/loss': 0.13579992949962616, 'train/grad_norm_pre_clip': 19.506168365478516, 'timing/data': 0.00021539998124353588, 'timing/model': 0.39385110000148416, 'learning_rate/qwen_vl_interface': 7.460545909167276e-06, 'learning_rate/action_model': 7.460545909167275e-05, 'epoch': 0.13}) +[2026-10-03 19:17:14,473][starVLA.training.train_starvla][INFO] - Step 1796, Loss: {'train/loss': 0.1478017270565033, 'train/grad_norm_pre_clip': 15.812479019165039, 'timing/data': 0.00022767999325878918, 'timing/model': 0.3938080700172577, 'learning_rate/qwen_vl_interface': 7.457849915972459e-06, 'learning_rate/action_model': 7.457849915972458e-05, 'epoch': 0.13}) +[2026-10-03 19:17:15,179][starVLA.training.train_starvla][INFO] - Step 1797, Loss: {'train/loss': 0.011010291986167431, 'train/grad_norm_pre_clip': 7.981098175048828, 'timing/data': 0.0003725500137079507, 'timing/model': 0.3946661800146103, 'learning_rate/qwen_vl_interface': 7.455153015213941e-06, 'learning_rate/action_model': 7.45515301521394e-05, 'epoch': 0.13}) +[2026-10-03 19:17:16,058][starVLA.training.train_starvla][INFO] - Step 1798, Loss: {'train/loss': 0.3090977370738983, 'train/grad_norm_pre_clip': 77.20655822753906, 'timing/data': 0.0003674199979286641, 'timing/model': 0.5938862299954053, 'learning_rate/qwen_vl_interface': 7.452455208000311e-06, 'learning_rate/action_model': 7.452455208000311e-05, 'epoch': 0.13}) +[2026-10-03 19:17:16,763][starVLA.training.train_starvla][INFO] - Step 1799, Loss: {'train/loss': 0.0003727069706656039, 'train/grad_norm_pre_clip': 0.14547380805015564, 'timing/data': 0.00032804999500513077, 'timing/model': 0.39346513999043964, 'learning_rate/qwen_vl_interface': 7.449756495440541e-06, 'learning_rate/action_model': 7.44975649544054e-05, 'epoch': 0.13}) +[2026-10-03 19:17:17,470][starVLA.training.train_starvla][INFO] - Step 1800, Loss: {'train/loss': 0.00011659700976451859, 'train/grad_norm_pre_clip': 0.03918188065290451, 'timing/data': 0.00044523997348733246, 'timing/model': 0.39542599002015777, 'learning_rate/qwen_vl_interface': 7.447056878643969e-06, 'learning_rate/action_model': 7.447056878643968e-05, 'epoch': 0.13}) +[2026-10-03 19:17:18,175][starVLA.training.train_starvla][INFO] - Step 1801, Loss: {'train/loss': 0.0018056367989629507, 'train/grad_norm_pre_clip': 2.8600876331329346, 'timing/data': 0.00033775001065805554, 'timing/model': 0.39417812900501303, 'learning_rate/qwen_vl_interface': 7.444356358720303e-06, 'learning_rate/action_model': 7.444356358720302e-05, 'epoch': 0.13}) +[2026-10-03 19:17:18,881][starVLA.training.train_starvla][INFO] - Step 1802, Loss: {'train/loss': 0.33432745933532715, 'train/grad_norm_pre_clip': 198.2742156982422, 'timing/data': 0.0004089099820703268, 'timing/model': 0.39384705998236313, 'learning_rate/qwen_vl_interface': 7.4416549367796255e-06, 'learning_rate/action_model': 7.441654936779625e-05, 'epoch': 0.13}) +[2026-10-03 19:17:19,791][starVLA.training.train_starvla][INFO] - Step 1803, Loss: {'train/loss': 0.000603343010880053, 'train/grad_norm_pre_clip': 0.22816452383995056, 'timing/data': 0.00037290999898687005, 'timing/model': 0.5969888490217272, 'learning_rate/qwen_vl_interface': 7.4389526139323895e-06, 'learning_rate/action_model': 7.438952613932389e-05, 'epoch': 0.13}) +[2026-10-03 19:17:20,499][starVLA.training.train_starvla][INFO] - Step 1804, Loss: {'train/loss': 0.3332657814025879, 'train/grad_norm_pre_clip': 64.07211303710938, 'timing/data': 0.000335429998813197, 'timing/model': 0.3955193189904094, 'learning_rate/qwen_vl_interface': 7.436249391289419e-06, 'learning_rate/action_model': 7.436249391289418e-05, 'epoch': 0.13}) +[2026-10-03 19:17:21,210][starVLA.training.train_starvla][INFO] - Step 1805, Loss: {'train/loss': 0.01452445238828659, 'train/grad_norm_pre_clip': 17.662845611572266, 'timing/data': 0.0003469600051175803, 'timing/model': 0.3978133189957589, 'learning_rate/qwen_vl_interface': 7.433545269961906e-06, 'learning_rate/action_model': 7.433545269961906e-05, 'epoch': 0.13}) +[2026-10-03 19:17:21,919][starVLA.training.train_starvla][INFO] - Step 1806, Loss: {'train/loss': 0.03366389870643616, 'train/grad_norm_pre_clip': 16.716999053955078, 'timing/data': 0.000385090010240674, 'timing/model': 0.3955893800011836, 'learning_rate/qwen_vl_interface': 7.430840251061411e-06, 'learning_rate/action_model': 7.430840251061411e-05, 'epoch': 0.13}) +[2026-10-03 19:17:22,627][starVLA.training.train_starvla][INFO] - Step 1807, Loss: {'train/loss': 0.0006389283807948232, 'train/grad_norm_pre_clip': 0.7881811261177063, 'timing/data': 0.00030727998819202185, 'timing/model': 0.39532440999755636, 'learning_rate/qwen_vl_interface': 7.428134335699866e-06, 'learning_rate/action_model': 7.428134335699865e-05, 'epoch': 0.13}) +[2026-10-03 19:17:23,336][starVLA.training.train_starvla][INFO] - Step 1808, Loss: {'train/loss': 0.0013473446015268564, 'train/grad_norm_pre_clip': 0.9683735370635986, 'timing/data': 0.0003833599912468344, 'timing/model': 0.39581390999956056, 'learning_rate/qwen_vl_interface': 7.425427524989572e-06, 'learning_rate/action_model': 7.425427524989572e-05, 'epoch': 0.13}) +[2026-10-03 19:17:24,240][starVLA.training.train_starvla][INFO] - Step 1809, Loss: {'train/loss': 0.005353583488613367, 'train/grad_norm_pre_clip': 7.360487937927246, 'timing/data': 0.0003527999797370285, 'timing/model': 0.39537914999527857, 'learning_rate/qwen_vl_interface': 7.422719820043194e-06, 'learning_rate/action_model': 7.422719820043193e-05, 'epoch': 0.13}) +[2026-10-03 19:17:24,948][starVLA.training.train_starvla][INFO] - Step 1810, Loss: {'train/loss': 0.011590281501412392, 'train/grad_norm_pre_clip': 34.72544860839844, 'timing/data': 0.00023006999981589615, 'timing/model': 0.39625347999390215, 'learning_rate/qwen_vl_interface': 7.420011221973771e-06, 'learning_rate/action_model': 7.420011221973771e-05, 'epoch': 0.13}) +[2026-10-03 19:17:25,656][starVLA.training.train_starvla][INFO] - Step 1811, Loss: {'train/loss': 0.006792902015149593, 'train/grad_norm_pre_clip': 2.604623794555664, 'timing/data': 0.00033353999606333673, 'timing/model': 0.3954029799788259, 'learning_rate/qwen_vl_interface': 7.417301731894702e-06, 'learning_rate/action_model': 7.417301731894701e-05, 'epoch': 0.13}) +[2026-10-03 19:17:26,363][starVLA.training.train_starvla][INFO] - Step 1812, Loss: {'train/loss': 0.010516648180782795, 'train/grad_norm_pre_clip': 7.340145111083984, 'timing/data': 0.0002194700064137578, 'timing/model': 0.3959134489996359, 'learning_rate/qwen_vl_interface': 7.414591350919756e-06, 'learning_rate/action_model': 7.414591350919756e-05, 'epoch': 0.13}) +[2026-10-03 19:17:27,070][starVLA.training.train_starvla][INFO] - Step 1813, Loss: {'train/loss': 0.02986900880932808, 'train/grad_norm_pre_clip': 100.89476776123047, 'timing/data': 0.00036277002072893083, 'timing/model': 0.3945624600164592, 'learning_rate/qwen_vl_interface': 7.4118800801630715e-06, 'learning_rate/action_model': 7.411880080163072e-05, 'epoch': 0.13}) +[2026-10-03 19:17:27,974][starVLA.training.train_starvla][INFO] - Step 1814, Loss: {'train/loss': 0.009364530444145203, 'train/grad_norm_pre_clip': 12.996415138244629, 'timing/data': 0.00028124000527895987, 'timing/model': 0.3942860999959521, 'learning_rate/qwen_vl_interface': 7.409167920739146e-06, 'learning_rate/action_model': 7.409167920739147e-05, 'epoch': 0.13}) +[2026-10-03 19:17:28,682][starVLA.training.train_starvla][INFO] - Step 1815, Loss: {'train/loss': 0.0019086380489170551, 'train/grad_norm_pre_clip': 3.141162872314453, 'timing/data': 0.00033030001213774085, 'timing/model': 0.3941027000255417, 'learning_rate/qwen_vl_interface': 7.406454873762848e-06, 'learning_rate/action_model': 7.406454873762848e-05, 'epoch': 0.13}) +[2026-10-03 19:17:29,389][starVLA.training.train_starvla][INFO] - Step 1816, Loss: {'train/loss': 0.0018864944577217102, 'train/grad_norm_pre_clip': 0.22355270385742188, 'timing/data': 0.0003481800085864961, 'timing/model': 0.39552854900830425, 'learning_rate/qwen_vl_interface': 7.403740940349408e-06, 'learning_rate/action_model': 7.403740940349408e-05, 'epoch': 0.13}) +[2026-10-03 19:17:30,097][starVLA.training.train_starvla][INFO] - Step 1817, Loss: {'train/loss': 0.3390275835990906, 'train/grad_norm_pre_clip': 12.447307586669922, 'timing/data': 0.0003039000148419291, 'timing/model': 0.39600788897951134, 'learning_rate/qwen_vl_interface': 7.401026121614421e-06, 'learning_rate/action_model': 7.40102612161442e-05, 'epoch': 0.13}) +[2026-10-03 19:17:30,804][starVLA.training.train_starvla][INFO] - Step 1818, Loss: {'train/loss': 0.6896248459815979, 'train/grad_norm_pre_clip': 150.97103881835938, 'timing/data': 0.00033078002161346376, 'timing/model': 0.3942962799919769, 'learning_rate/qwen_vl_interface': 7.398310418673846e-06, 'learning_rate/action_model': 7.398310418673846e-05, 'epoch': 0.13}) +[2026-10-03 19:17:31,709][starVLA.training.train_starvla][INFO] - Step 1819, Loss: {'train/loss': 0.11547248065471649, 'train/grad_norm_pre_clip': 155.7595672607422, 'timing/data': 0.00022188900038599968, 'timing/model': 0.5918847110006027, 'learning_rate/qwen_vl_interface': 7.395593832644007e-06, 'learning_rate/action_model': 7.395593832644007e-05, 'epoch': 0.13}) +[2026-10-03 19:17:32,416][starVLA.training.train_starvla][INFO] - Step 1820, Loss: {'train/loss': 0.34162530303001404, 'train/grad_norm_pre_clip': 79.01918029785156, 'timing/data': 0.00027982998290099204, 'timing/model': 0.39536645999760367, 'learning_rate/qwen_vl_interface': 7.39287636464159e-06, 'learning_rate/action_model': 7.39287636464159e-05, 'epoch': 0.13}) +[2026-10-03 19:17:33,125][starVLA.training.train_starvla][INFO] - Step 1821, Loss: {'train/loss': 0.0004730869841296226, 'train/grad_norm_pre_clip': 0.4790368676185608, 'timing/data': 0.0003841599973384291, 'timing/model': 0.3963437300117221, 'learning_rate/qwen_vl_interface': 7.3901580157836425e-06, 'learning_rate/action_model': 7.390158015783643e-05, 'epoch': 0.13}) +[2026-10-03 19:17:33,835][starVLA.training.train_starvla][INFO] - Step 1822, Loss: {'train/loss': 0.1542329043149948, 'train/grad_norm_pre_clip': 149.905029296875, 'timing/data': 0.0003900600131601095, 'timing/model': 0.39482706002308987, 'learning_rate/qwen_vl_interface': 7.387438787187576e-06, 'learning_rate/action_model': 7.387438787187576e-05, 'epoch': 0.13}) +[2026-10-03 19:17:34,543][starVLA.training.train_starvla][INFO] - Step 1823, Loss: {'train/loss': 0.0008344700909219682, 'train/grad_norm_pre_clip': 0.7894110679626465, 'timing/data': 0.00033459998667240143, 'timing/model': 0.3956435190048069, 'learning_rate/qwen_vl_interface': 7.3847186799711625e-06, 'learning_rate/action_model': 7.384718679971162e-05, 'epoch': 0.13}) +[2026-10-03 19:17:35,449][starVLA.training.train_starvla][INFO] - Step 1824, Loss: {'train/loss': 0.009018543176352978, 'train/grad_norm_pre_clip': 6.434994697570801, 'timing/data': 0.00030351002351380885, 'timing/model': 0.5948057099885773, 'learning_rate/qwen_vl_interface': 7.381997695252535e-06, 'learning_rate/action_model': 7.381997695252535e-05, 'epoch': 0.13}) +[2026-10-03 19:17:36,157][starVLA.training.train_starvla][INFO] - Step 1825, Loss: {'train/loss': 0.005891742184758186, 'train/grad_norm_pre_clip': 3.355541706085205, 'timing/data': 0.0002552600053604692, 'timing/model': 0.3966532499762252, 'learning_rate/qwen_vl_interface': 7.379275834150186e-06, 'learning_rate/action_model': 7.379275834150186e-05, 'epoch': 0.13}) +[2026-10-03 19:17:36,864][starVLA.training.train_starvla][INFO] - Step 1826, Loss: {'train/loss': 1.4602868759538978e-05, 'train/grad_norm_pre_clip': 0.0027298382483422756, 'timing/data': 0.00021696000476367772, 'timing/model': 0.3952885699982289, 'learning_rate/qwen_vl_interface': 7.376553097782972e-06, 'learning_rate/action_model': 7.376553097782972e-05, 'epoch': 0.13}) +[2026-10-03 19:17:37,573][starVLA.training.train_starvla][INFO] - Step 1827, Loss: {'train/loss': 0.7714670896530151, 'train/grad_norm_pre_clip': 68.60726165771484, 'timing/data': 0.000218659988604486, 'timing/model': 0.3965646600117907, 'learning_rate/qwen_vl_interface': 7.373829487270109e-06, 'learning_rate/action_model': 7.373829487270109e-05, 'epoch': 0.13}) +[2026-10-03 19:17:38,282][starVLA.training.train_starvla][INFO] - Step 1828, Loss: {'train/loss': 0.07849878817796707, 'train/grad_norm_pre_clip': 167.4608917236328, 'timing/data': 0.00022889001411385834, 'timing/model': 0.39573104001465254, 'learning_rate/qwen_vl_interface': 7.371105003731165e-06, 'learning_rate/action_model': 7.371105003731164e-05, 'epoch': 0.13}) +[2026-10-03 19:17:39,190][starVLA.training.train_starvla][INFO] - Step 1829, Loss: {'train/loss': 0.39580628275871277, 'train/grad_norm_pre_clip': 197.01171875, 'timing/data': 0.0002520499983802438, 'timing/model': 0.5969655099906959, 'learning_rate/qwen_vl_interface': 7.3683796482860755e-06, 'learning_rate/action_model': 7.368379648286075e-05, 'epoch': 0.13}) +[2026-10-03 19:17:39,896][starVLA.training.train_starvla][INFO] - Step 1830, Loss: {'train/loss': 0.002444982063025236, 'train/grad_norm_pre_clip': 0.7059723734855652, 'timing/data': 0.00033907999750226736, 'timing/model': 0.39388602998224087, 'learning_rate/qwen_vl_interface': 7.365653422055131e-06, 'learning_rate/action_model': 7.365653422055131e-05, 'epoch': 0.13}) +[2026-10-03 19:17:40,605][starVLA.training.train_starvla][INFO] - Step 1831, Loss: {'train/loss': 0.00012310445890761912, 'train/grad_norm_pre_clip': 0.11082460731267929, 'timing/data': 0.00032796000596135855, 'timing/model': 0.3957913600024767, 'learning_rate/qwen_vl_interface': 7.3629263261589816e-06, 'learning_rate/action_model': 7.362926326158981e-05, 'epoch': 0.13}) +[2026-10-03 19:17:41,312][starVLA.training.train_starvla][INFO] - Step 1832, Loss: {'train/loss': 0.12438224256038666, 'train/grad_norm_pre_clip': 57.616188049316406, 'timing/data': 0.0002990200009662658, 'timing/model': 0.3939116599794943, 'learning_rate/qwen_vl_interface': 7.36019836171863e-06, 'learning_rate/action_model': 7.360198361718629e-05, 'epoch': 0.13}) +[2026-10-03 19:17:42,021][starVLA.training.train_starvla][INFO] - Step 1833, Loss: {'train/loss': 0.3711111843585968, 'train/grad_norm_pre_clip': 95.77177429199219, 'timing/data': 0.00032706899219192564, 'timing/model': 0.3962806100025773, 'learning_rate/qwen_vl_interface': 7.357469529855441e-06, 'learning_rate/action_model': 7.35746952985544e-05, 'epoch': 0.13}) +[2026-10-03 19:17:42,727][starVLA.training.train_starvla][INFO] - Step 1834, Loss: {'train/loss': 0.8257461786270142, 'train/grad_norm_pre_clip': 75.00777435302734, 'timing/data': 0.00031728000612929463, 'timing/model': 0.395375519990921, 'learning_rate/qwen_vl_interface': 7.354739831691134e-06, 'learning_rate/action_model': 7.354739831691135e-05, 'epoch': 0.13}) +[2026-10-03 19:17:43,635][starVLA.training.train_starvla][INFO] - Step 1835, Loss: {'train/loss': 4.990580418962054e-05, 'train/grad_norm_pre_clip': 0.018353130668401718, 'timing/data': 0.00028408001526258886, 'timing/model': 0.39587346999906003, 'learning_rate/qwen_vl_interface': 7.352009268347786e-06, 'learning_rate/action_model': 7.352009268347786e-05, 'epoch': 0.13}) +[2026-10-03 19:17:44,342][starVLA.training.train_starvla][INFO] - Step 1836, Loss: {'train/loss': 4.3253017793176696e-05, 'train/grad_norm_pre_clip': 0.016890408471226692, 'timing/data': 0.00033181000617332757, 'timing/model': 0.3942045000148937, 'learning_rate/qwen_vl_interface': 7.349277840947828e-06, 'learning_rate/action_model': 7.349277840947828e-05, 'epoch': 0.13}) +[2026-10-03 19:17:45,052][starVLA.training.train_starvla][INFO] - Step 1837, Loss: {'train/loss': 0.06765135377645493, 'train/grad_norm_pre_clip': 82.40921020507812, 'timing/data': 0.0003388399782124907, 'timing/model': 0.39652962898253463, 'learning_rate/qwen_vl_interface': 7.346545550614046e-06, 'learning_rate/action_model': 7.346545550614046e-05, 'epoch': 0.13}) +[2026-10-03 19:17:45,762][starVLA.training.train_starvla][INFO] - Step 1838, Loss: {'train/loss': 0.40010809898376465, 'train/grad_norm_pre_clip': 45.45726776123047, 'timing/data': 0.00032267000642605126, 'timing/model': 0.39712633899762295, 'learning_rate/qwen_vl_interface': 7.343812398469582e-06, 'learning_rate/action_model': 7.343812398469581e-05, 'epoch': 0.13}) +[2026-10-03 19:17:46,470][starVLA.training.train_starvla][INFO] - Step 1839, Loss: {'train/loss': 0.0001720967993605882, 'train/grad_norm_pre_clip': 0.05892961099743843, 'timing/data': 0.00035689998185262084, 'timing/model': 0.3944100599910598, 'learning_rate/qwen_vl_interface': 7.34107838563793e-06, 'learning_rate/action_model': 7.341078385637929e-05, 'epoch': 0.13}) +[2026-10-03 19:17:47,376][starVLA.training.train_starvla][INFO] - Step 1840, Loss: {'train/loss': 0.0037588803097605705, 'train/grad_norm_pre_clip': 7.582969665527344, 'timing/data': 0.0002228700031992048, 'timing/model': 0.5952053499931935, 'learning_rate/qwen_vl_interface': 7.338343513242941e-06, 'learning_rate/action_model': 7.338343513242941e-05, 'epoch': 0.13}) +[2026-10-03 19:17:48,081][starVLA.training.train_starvla][INFO] - Step 1841, Loss: {'train/loss': 0.7852926254272461, 'train/grad_norm_pre_clip': 86.12821960449219, 'timing/data': 0.00028854000265710056, 'timing/model': 0.39425133998156525, 'learning_rate/qwen_vl_interface': 7.335607782408818e-06, 'learning_rate/action_model': 7.335607782408818e-05, 'epoch': 0.13}) +[2026-10-03 19:17:48,788][starVLA.training.train_starvla][INFO] - Step 1842, Loss: {'train/loss': 0.005536421202123165, 'train/grad_norm_pre_clip': 5.644298553466797, 'timing/data': 0.0003562099882401526, 'timing/model': 0.39470698998775333, 'learning_rate/qwen_vl_interface': 7.3328711942601145e-06, 'learning_rate/action_model': 7.332871194260115e-05, 'epoch': 0.13}) +[2026-10-03 19:17:49,494][starVLA.training.train_starvla][INFO] - Step 1843, Loss: {'train/loss': 0.004313337616622448, 'train/grad_norm_pre_clip': 6.169033527374268, 'timing/data': 0.0003379800182301551, 'timing/model': 0.39450275999843143, 'learning_rate/qwen_vl_interface': 7.33013374992174e-06, 'learning_rate/action_model': 7.33013374992174e-05, 'epoch': 0.13}) +[2026-10-03 19:17:50,198][starVLA.training.train_starvla][INFO] - Step 1844, Loss: {'train/loss': 0.3055402338504791, 'train/grad_norm_pre_clip': 65.30669403076172, 'timing/data': 0.0002957599936053157, 'timing/model': 0.3943078400043305, 'learning_rate/qwen_vl_interface': 7.327395450518956e-06, 'learning_rate/action_model': 7.327395450518956e-05, 'epoch': 0.13}) +[2026-10-03 19:17:50,906][starVLA.training.train_starvla][INFO] - Step 1845, Loss: {'train/loss': 0.0028857667930424213, 'train/grad_norm_pre_clip': 1.1504734754562378, 'timing/data': 0.00037838000571355224, 'timing/model': 0.39577861901489086, 'learning_rate/qwen_vl_interface': 7.324656297177372e-06, 'learning_rate/action_model': 7.324656297177371e-05, 'epoch': 0.14}) +[2026-10-03 19:17:51,810][starVLA.training.train_starvla][INFO] - Step 1846, Loss: {'train/loss': 0.3677736818790436, 'train/grad_norm_pre_clip': 52.59485626220703, 'timing/data': 0.0003577799943741411, 'timing/model': 0.3932157810195349, 'learning_rate/qwen_vl_interface': 7.32191629102295e-06, 'learning_rate/action_model': 7.32191629102295e-05, 'epoch': 0.14}) +[2026-10-03 19:17:52,516][starVLA.training.train_starvla][INFO] - Step 1847, Loss: {'train/loss': 0.39000338315963745, 'train/grad_norm_pre_clip': 60.94755554199219, 'timing/data': 0.00024247000692412257, 'timing/model': 0.39417816000059247, 'learning_rate/qwen_vl_interface': 7.319175433182004e-06, 'learning_rate/action_model': 7.319175433182004e-05, 'epoch': 0.14}) +[2026-10-03 19:17:53,222][starVLA.training.train_starvla][INFO] - Step 1848, Loss: {'train/loss': 0.049560122191905975, 'train/grad_norm_pre_clip': 49.30352020263672, 'timing/data': 0.00022329000057652593, 'timing/model': 0.3938782700279262, 'learning_rate/qwen_vl_interface': 7.316433724781198e-06, 'learning_rate/action_model': 7.316433724781197e-05, 'epoch': 0.14}) +[2026-10-03 19:17:53,929][starVLA.training.train_starvla][INFO] - Step 1849, Loss: {'train/loss': 0.09484502673149109, 'train/grad_norm_pre_clip': 60.677276611328125, 'timing/data': 0.0003522600163705647, 'timing/model': 0.39468381999176927, 'learning_rate/qwen_vl_interface': 7.313691166947543e-06, 'learning_rate/action_model': 7.313691166947543e-05, 'epoch': 0.14}) +[2026-10-03 19:17:54,635][starVLA.training.train_starvla][INFO] - Step 1850, Loss: {'train/loss': 0.6824548244476318, 'train/grad_norm_pre_clip': 73.20402526855469, 'timing/data': 0.00022770999930799007, 'timing/model': 0.3948475299985148, 'learning_rate/qwen_vl_interface': 7.310947760808401e-06, 'learning_rate/action_model': 7.3109477608084e-05, 'epoch': 0.14}) +[2026-10-03 19:17:55,539][starVLA.training.train_starvla][INFO] - Step 1851, Loss: {'train/loss': 0.07628201693296432, 'train/grad_norm_pre_clip': 61.168617248535156, 'timing/data': 0.00022281997371464968, 'timing/model': 0.39302757001132704, 'learning_rate/qwen_vl_interface': 7.308203507491485e-06, 'learning_rate/action_model': 7.308203507491484e-05, 'epoch': 0.14}) +[2026-10-03 19:17:56,242][starVLA.training.train_starvla][INFO] - Step 1852, Loss: {'train/loss': 0.0016602715477347374, 'train/grad_norm_pre_clip': 2.6829001903533936, 'timing/data': 0.0003372300125192851, 'timing/model': 0.3931267900043167, 'learning_rate/qwen_vl_interface': 7.3054584081248515e-06, 'learning_rate/action_model': 7.305458408124852e-05, 'epoch': 0.14}) +[2026-10-03 19:17:56,949][starVLA.training.train_starvla][INFO] - Step 1853, Loss: {'train/loss': 8.78343780641444e-05, 'train/grad_norm_pre_clip': 0.01365023571997881, 'timing/data': 0.0003512900148052722, 'timing/model': 0.39472782000666484, 'learning_rate/qwen_vl_interface': 7.302712463836907e-06, 'learning_rate/action_model': 7.302712463836907e-05, 'epoch': 0.14}) +[2026-10-03 19:17:57,655][starVLA.training.train_starvla][INFO] - Step 1854, Loss: {'train/loss': 0.10422861576080322, 'train/grad_norm_pre_clip': 94.37322235107422, 'timing/data': 0.0003547999949660152, 'timing/model': 0.39484306998201646, 'learning_rate/qwen_vl_interface': 7.2999656757564086e-06, 'learning_rate/action_model': 7.299965675756409e-05, 'epoch': 0.14}) +[2026-10-03 19:17:58,360][starVLA.training.train_starvla][INFO] - Step 1855, Loss: {'train/loss': 0.0773952379822731, 'train/grad_norm_pre_clip': 71.53026580810547, 'timing/data': 0.00042158999713137746, 'timing/model': 0.393028210994089, 'learning_rate/qwen_vl_interface': 7.297218045012453e-06, 'learning_rate/action_model': 7.297218045012453e-05, 'epoch': 0.14}) +[2026-10-03 19:17:59,266][starVLA.training.train_starvla][INFO] - Step 1856, Loss: {'train/loss': 0.0004963534884154797, 'train/grad_norm_pre_clip': 0.26359453797340393, 'timing/data': 0.0002874899946618825, 'timing/model': 0.5945312599942554, 'learning_rate/qwen_vl_interface': 7.29446957273449e-06, 'learning_rate/action_model': 7.294469572734489e-05, 'epoch': 0.14}) +[2026-10-03 19:17:59,974][starVLA.training.train_starvla][INFO] - Step 1857, Loss: {'train/loss': 0.0020103775896131992, 'train/grad_norm_pre_clip': 1.2394095659255981, 'timing/data': 0.0003452399978414178, 'timing/model': 0.39561189999221824, 'learning_rate/qwen_vl_interface': 7.2917202600523104e-06, 'learning_rate/action_model': 7.29172026005231e-05, 'epoch': 0.14}) +[2026-10-03 19:18:00,678][starVLA.training.train_starvla][INFO] - Step 1858, Loss: {'train/loss': 0.0031182379461824894, 'train/grad_norm_pre_clip': 0.9104740619659424, 'timing/data': 0.0003712099860422313, 'timing/model': 0.393581729993457, 'learning_rate/qwen_vl_interface': 7.2889701080960565e-06, 'learning_rate/action_model': 7.288970108096056e-05, 'epoch': 0.14}) +[2026-10-03 19:18:01,382][starVLA.training.train_starvla][INFO] - Step 1859, Loss: {'train/loss': 0.3070005774497986, 'train/grad_norm_pre_clip': 44.23052215576172, 'timing/data': 0.0003354000218678266, 'timing/model': 0.39267170999664813, 'learning_rate/qwen_vl_interface': 7.286219117996209e-06, 'learning_rate/action_model': 7.286219117996207e-05, 'epoch': 0.14}) +[2026-10-03 19:18:02,090][starVLA.training.train_starvla][INFO] - Step 1860, Loss: {'train/loss': 0.515217125415802, 'train/grad_norm_pre_clip': 79.10031127929688, 'timing/data': 0.0003826199972536415, 'timing/model': 0.3954392900050152, 'learning_rate/qwen_vl_interface': 7.283467290883598e-06, 'learning_rate/action_model': 7.283467290883598e-05, 'epoch': 0.14}) +[2026-10-03 19:18:02,800][starVLA.training.train_starvla][INFO] - Step 1861, Loss: {'train/loss': 0.004167541861534119, 'train/grad_norm_pre_clip': 5.74988317489624, 'timing/data': 0.0002346399996895343, 'timing/model': 0.39528672900632955, 'learning_rate/qwen_vl_interface': 7.2807146278893925e-06, 'learning_rate/action_model': 7.280714627889393e-05, 'epoch': 0.14}) +[2026-10-03 19:18:03,721][starVLA.training.train_starvla][INFO] - Step 1862, Loss: {'train/loss': 0.4310607314109802, 'train/grad_norm_pre_clip': 84.02552795410156, 'timing/data': 0.00034958997275680304, 'timing/model': 0.3957269200182054, 'learning_rate/qwen_vl_interface': 7.277961130145111e-06, 'learning_rate/action_model': 7.277961130145111e-05, 'epoch': 0.14}) +[2026-10-03 19:18:04,431][starVLA.training.train_starvla][INFO] - Step 1863, Loss: {'train/loss': 0.0022237745579332113, 'train/grad_norm_pre_clip': 0.8751505017280579, 'timing/data': 0.00024164997739717364, 'timing/model': 0.39737729899934493, 'learning_rate/qwen_vl_interface': 7.2752067987826116e-06, 'learning_rate/action_model': 7.275206798782611e-05, 'epoch': 0.14}) +[2026-10-03 19:18:05,139][starVLA.training.train_starvla][INFO] - Step 1864, Loss: {'train/loss': 0.0005130420904606581, 'train/grad_norm_pre_clip': 0.23078878223896027, 'timing/data': 0.0003667499986477196, 'timing/model': 0.39441359997726977, 'learning_rate/qwen_vl_interface': 7.272451634934096e-06, 'learning_rate/action_model': 7.272451634934096e-05, 'epoch': 0.14}) +[2026-10-03 19:18:05,850][starVLA.training.train_starvla][INFO] - Step 1865, Loss: {'train/loss': 0.0003346819430589676, 'train/grad_norm_pre_clip': 0.1344158798456192, 'timing/data': 0.0003829500055871904, 'timing/model': 0.3971897499868646, 'learning_rate/qwen_vl_interface': 7.269695639732108e-06, 'learning_rate/action_model': 7.269695639732108e-05, 'epoch': 0.14}) +[2026-10-03 19:18:06,559][starVLA.training.train_starvla][INFO] - Step 1866, Loss: {'train/loss': 0.01507624052464962, 'train/grad_norm_pre_clip': 11.935884475708008, 'timing/data': 0.00021868999465368688, 'timing/model': 0.39587392000248656, 'learning_rate/qwen_vl_interface': 7.266938814309533e-06, 'learning_rate/action_model': 7.266938814309533e-05, 'epoch': 0.14}) +[2026-10-03 19:18:07,437][starVLA.training.train_starvla][INFO] - Step 1867, Loss: {'train/loss': 0.5934782028198242, 'train/grad_norm_pre_clip': 187.70306396484375, 'timing/data': 0.00034507000236772, 'timing/model': 0.36942821301636286, 'learning_rate/qwen_vl_interface': 7.264181159799598e-06, 'learning_rate/action_model': 7.264181159799597e-05, 'epoch': 0.14}) +[2026-10-03 19:18:08,143][starVLA.training.train_starvla][INFO] - Step 1868, Loss: {'train/loss': 0.0003041054296772927, 'train/grad_norm_pre_clip': 0.2417486608028412, 'timing/data': 0.0003381099959369749, 'timing/model': 0.3941619700053707, 'learning_rate/qwen_vl_interface': 7.261422677335871e-06, 'learning_rate/action_model': 7.261422677335871e-05, 'epoch': 0.14}) +[2026-10-03 19:18:08,849][starVLA.training.train_starvla][INFO] - Step 1869, Loss: {'train/loss': 0.6242448687553406, 'train/grad_norm_pre_clip': 35.32722091674805, 'timing/data': 0.0002682000049389899, 'timing/model': 0.39484612000524066, 'learning_rate/qwen_vl_interface': 7.258663368052259e-06, 'learning_rate/action_model': 7.258663368052259e-05, 'epoch': 0.14}) +[2026-10-03 19:18:09,557][starVLA.training.train_starvla][INFO] - Step 1870, Loss: {'train/loss': 6.8920329795219e-05, 'train/grad_norm_pre_clip': 0.011328957043588161, 'timing/data': 0.00025409998488612473, 'timing/model': 0.39404290000675246, 'learning_rate/qwen_vl_interface': 7.2559032330830105e-06, 'learning_rate/action_model': 7.25590323308301e-05, 'epoch': 0.14}) +[2026-10-03 19:18:10,263][starVLA.training.train_starvla][INFO] - Step 1871, Loss: {'train/loss': 0.0002725763770285994, 'train/grad_norm_pre_clip': 0.11292275786399841, 'timing/data': 0.0002212599792983383, 'timing/model': 0.39602894001291133, 'learning_rate/qwen_vl_interface': 7.253142273562712e-06, 'learning_rate/action_model': 7.253142273562712e-05, 'epoch': 0.14}) +[2026-10-03 19:18:11,171][starVLA.training.train_starvla][INFO] - Step 1872, Loss: {'train/loss': 0.0007617121445946395, 'train/grad_norm_pre_clip': 0.23364250361919403, 'timing/data': 0.0003129800024908036, 'timing/model': 0.5941060890036169, 'learning_rate/qwen_vl_interface': 7.25038049062629e-06, 'learning_rate/action_model': 7.250380490626289e-05, 'epoch': 0.14}) +[2026-10-03 19:18:11,879][starVLA.training.train_starvla][INFO] - Step 1873, Loss: {'train/loss': 0.02338395267724991, 'train/grad_norm_pre_clip': 48.21775436401367, 'timing/data': 0.0003181500069331378, 'timing/model': 0.39525868999771774, 'learning_rate/qwen_vl_interface': 7.24761788540901e-06, 'learning_rate/action_model': 7.24761788540901e-05, 'epoch': 0.14}) +[2026-10-03 19:18:12,586][starVLA.training.train_starvla][INFO] - Step 1874, Loss: {'train/loss': 0.34137725830078125, 'train/grad_norm_pre_clip': 71.3032455444336, 'timing/data': 0.00033628998789936304, 'timing/model': 0.3941123700060416, 'learning_rate/qwen_vl_interface': 7.244854459046473e-06, 'learning_rate/action_model': 7.244854459046473e-05, 'epoch': 0.14}) +[2026-10-03 19:18:13,295][starVLA.training.train_starvla][INFO] - Step 1875, Loss: {'train/loss': 0.008643822744488716, 'train/grad_norm_pre_clip': 31.342443466186523, 'timing/data': 0.0004633499775081873, 'timing/model': 0.3932933500036597, 'learning_rate/qwen_vl_interface': 7.24209021267462e-06, 'learning_rate/action_model': 7.24209021267462e-05, 'epoch': 0.14}) +[2026-10-03 19:18:14,001][starVLA.training.train_starvla][INFO] - Step 1876, Loss: {'train/loss': 0.0011993303196504712, 'train/grad_norm_pre_clip': 0.4713554382324219, 'timing/data': 0.00021569000091403723, 'timing/model': 0.39518662900081836, 'learning_rate/qwen_vl_interface': 7.23932514742973e-06, 'learning_rate/action_model': 7.23932514742973e-05, 'epoch': 0.14}) +[2026-10-03 19:18:14,707][starVLA.training.train_starvla][INFO] - Step 1877, Loss: {'train/loss': 0.4127357602119446, 'train/grad_norm_pre_clip': 9.216127395629883, 'timing/data': 0.00026470000739209354, 'timing/model': 0.39567160001024604, 'learning_rate/qwen_vl_interface': 7.236559264448413e-06, 'learning_rate/action_model': 7.236559264448412e-05, 'epoch': 0.14}) +[2026-10-03 19:18:15,606][starVLA.training.train_starvla][INFO] - Step 1878, Loss: {'train/loss': 7.393583655357361e-05, 'train/grad_norm_pre_clip': 0.016139784827828407, 'timing/data': 0.00021547998767346144, 'timing/model': 0.3954604390019085, 'learning_rate/qwen_vl_interface': 7.233792564867623e-06, 'learning_rate/action_model': 7.233792564867623e-05, 'epoch': 0.14}) +[2026-10-03 19:18:16,314][starVLA.training.train_starvla][INFO] - Step 1879, Loss: {'train/loss': 0.0003585627127904445, 'train/grad_norm_pre_clip': 0.23427468538284302, 'timing/data': 0.00021231998107396066, 'timing/model': 0.3957084389985539, 'learning_rate/qwen_vl_interface': 7.231025049824644e-06, 'learning_rate/action_model': 7.231025049824643e-05, 'epoch': 0.14}) +[2026-10-03 19:18:17,018][starVLA.training.train_starvla][INFO] - Step 1880, Loss: {'train/loss': 0.0014485333813354373, 'train/grad_norm_pre_clip': 0.6846610307693481, 'timing/data': 0.00021502000163309276, 'timing/model': 0.3934495700232219, 'learning_rate/qwen_vl_interface': 7.228256720457097e-06, 'learning_rate/action_model': 7.228256720457096e-05, 'epoch': 0.14}) +[2026-10-03 19:18:17,727][starVLA.training.train_starvla][INFO] - Step 1881, Loss: {'train/loss': 0.0004579632659442723, 'train/grad_norm_pre_clip': 0.07637114077806473, 'timing/data': 0.0003698099753819406, 'timing/model': 0.396475218993146, 'learning_rate/qwen_vl_interface': 7.225487577902937e-06, 'learning_rate/action_model': 7.225487577902936e-05, 'epoch': 0.14}) +[2026-10-03 19:18:18,435][starVLA.training.train_starvla][INFO] - Step 1882, Loss: {'train/loss': 0.3696083128452301, 'train/grad_norm_pre_clip': 22.19546890258789, 'timing/data': 0.0003478199942037463, 'timing/model': 0.39604039999539964, 'learning_rate/qwen_vl_interface': 7.222717623300455e-06, 'learning_rate/action_model': 7.222717623300455e-05, 'epoch': 0.14}) +[2026-10-03 19:18:19,346][starVLA.training.train_starvla][INFO] - Step 1883, Loss: {'train/loss': 1.4090019464492798, 'train/grad_norm_pre_clip': 58.37103271484375, 'timing/data': 0.00025504999211989343, 'timing/model': 0.39514101899112575, 'learning_rate/qwen_vl_interface': 7.219946857788273e-06, 'learning_rate/action_model': 7.219946857788272e-05, 'epoch': 0.14}) +[2026-10-03 19:18:20,054][starVLA.training.train_starvla][INFO] - Step 1884, Loss: {'train/loss': 0.0010823967168107629, 'train/grad_norm_pre_clip': 0.2509209215641022, 'timing/data': 0.0002707800013013184, 'timing/model': 0.39583545000641607, 'learning_rate/qwen_vl_interface': 7.21717528250535e-06, 'learning_rate/action_model': 7.21717528250535e-05, 'epoch': 0.14}) +[2026-10-03 19:18:20,763][starVLA.training.train_starvla][INFO] - Step 1885, Loss: {'train/loss': 0.0030819731764495373, 'train/grad_norm_pre_clip': 3.767500162124634, 'timing/data': 0.0003271200112067163, 'timing/model': 0.39626603000215255, 'learning_rate/qwen_vl_interface': 7.214402898590974e-06, 'learning_rate/action_model': 7.214402898590973e-05, 'epoch': 0.14}) +[2026-10-03 19:18:21,471][starVLA.training.train_starvla][INFO] - Step 1886, Loss: {'train/loss': 0.007018962409347296, 'train/grad_norm_pre_clip': 5.940585136413574, 'timing/data': 0.00036030999035574496, 'timing/model': 0.39464841000153683, 'learning_rate/qwen_vl_interface': 7.211629707184768e-06, 'learning_rate/action_model': 7.211629707184768e-05, 'epoch': 0.14}) +[2026-10-03 19:18:22,178][starVLA.training.train_starvla][INFO] - Step 1887, Loss: {'train/loss': 0.877090334892273, 'train/grad_norm_pre_clip': 66.35017395019531, 'timing/data': 0.00023211000370793045, 'timing/model': 0.39652273899991997, 'learning_rate/qwen_vl_interface': 7.2088557094266885e-06, 'learning_rate/action_model': 7.208855709426688e-05, 'epoch': 0.14}) +[2026-10-03 19:18:23,084][starVLA.training.train_starvla][INFO] - Step 1888, Loss: {'train/loss': 0.38345208764076233, 'train/grad_norm_pre_clip': 49.030696868896484, 'timing/data': 0.00022973999148234725, 'timing/model': 0.5944015699787997, 'learning_rate/qwen_vl_interface': 7.2060809064570195e-06, 'learning_rate/action_model': 7.20608090645702e-05, 'epoch': 0.14}) +[2026-10-03 19:18:23,794][starVLA.training.train_starvla][INFO] - Step 1889, Loss: {'train/loss': 0.0001476240431657061, 'train/grad_norm_pre_clip': 0.12311901152133942, 'timing/data': 0.0004227900062687695, 'timing/model': 0.39625424999394454, 'learning_rate/qwen_vl_interface': 7.203305299416375e-06, 'learning_rate/action_model': 7.203305299416375e-05, 'epoch': 0.14}) +[2026-10-03 19:18:24,502][starVLA.training.train_starvla][INFO] - Step 1890, Loss: {'train/loss': 0.016662873327732086, 'train/grad_norm_pre_clip': 32.77676010131836, 'timing/data': 0.0003688399738166481, 'timing/model': 0.39527814998291433, 'learning_rate/qwen_vl_interface': 7.200528889445707e-06, 'learning_rate/action_model': 7.200528889445706e-05, 'epoch': 0.14}) +[2026-10-03 19:18:25,212][starVLA.training.train_starvla][INFO] - Step 1891, Loss: {'train/loss': 0.16329695284366608, 'train/grad_norm_pre_clip': 48.86083221435547, 'timing/data': 0.00022211001487448812, 'timing/model': 0.39697544899536297, 'learning_rate/qwen_vl_interface': 7.19775167768629e-06, 'learning_rate/action_model': 7.19775167768629e-05, 'epoch': 0.14}) +[2026-10-03 19:18:25,918][starVLA.training.train_starvla][INFO] - Step 1892, Loss: {'train/loss': 0.0038844977971166372, 'train/grad_norm_pre_clip': 6.490099906921387, 'timing/data': 0.00022355999681167305, 'timing/model': 0.3945814890030306, 'learning_rate/qwen_vl_interface': 7.194973665279731e-06, 'learning_rate/action_model': 7.19497366527973e-05, 'epoch': 0.14}) +[2026-10-03 19:18:26,823][starVLA.training.train_starvla][INFO] - Step 1893, Loss: {'train/loss': 1.0588330030441284, 'train/grad_norm_pre_clip': 50.029808044433594, 'timing/data': 0.0002250199904665351, 'timing/model': 0.5944328189943917, 'learning_rate/qwen_vl_interface': 7.1921948533679666e-06, 'learning_rate/action_model': 7.192194853367967e-05, 'epoch': 0.14}) +[2026-10-03 19:18:27,530][starVLA.training.train_starvla][INFO] - Step 1894, Loss: {'train/loss': 0.5253769159317017, 'train/grad_norm_pre_clip': 71.31425476074219, 'timing/data': 0.0004116999916732311, 'timing/model': 0.3960607390035875, 'learning_rate/qwen_vl_interface': 7.1894152430932605e-06, 'learning_rate/action_model': 7.189415243093261e-05, 'epoch': 0.14}) +[2026-10-03 19:18:28,238][starVLA.training.train_starvla][INFO] - Step 1895, Loss: {'train/loss': 0.0627746433019638, 'train/grad_norm_pre_clip': 125.77198028564453, 'timing/data': 0.0003295399947091937, 'timing/model': 0.39510339000844397, 'learning_rate/qwen_vl_interface': 7.186634835598205e-06, 'learning_rate/action_model': 7.186634835598205e-05, 'epoch': 0.14}) +[2026-10-03 19:18:28,947][starVLA.training.train_starvla][INFO] - Step 1896, Loss: {'train/loss': 0.09068704396486282, 'train/grad_norm_pre_clip': 24.54720687866211, 'timing/data': 0.00023031001910567284, 'timing/model': 0.39646173000801355, 'learning_rate/qwen_vl_interface': 7.183853632025724e-06, 'learning_rate/action_model': 7.183853632025723e-05, 'epoch': 0.14}) +[2026-10-03 19:18:29,655][starVLA.training.train_starvla][INFO] - Step 1897, Loss: {'train/loss': 0.0005084027070552111, 'train/grad_norm_pre_clip': 0.12019266188144684, 'timing/data': 0.00030375001369975507, 'timing/model': 0.3951250499812886, 'learning_rate/qwen_vl_interface': 7.181071633519062e-06, 'learning_rate/action_model': 7.181071633519062e-05, 'epoch': 0.14}) +[2026-10-03 19:18:30,360][starVLA.training.train_starvla][INFO] - Step 1898, Loss: {'train/loss': 0.1821768581867218, 'train/grad_norm_pre_clip': 71.87187194824219, 'timing/data': 0.00023432000307366252, 'timing/model': 0.3944640289992094, 'learning_rate/qwen_vl_interface': 7.178288841221795e-06, 'learning_rate/action_model': 7.178288841221794e-05, 'epoch': 0.14}) +[2026-10-03 19:18:31,282][starVLA.training.train_starvla][INFO] - Step 1899, Loss: {'train/loss': 0.15104150772094727, 'train/grad_norm_pre_clip': 129.44859313964844, 'timing/data': 0.00031269999453797936, 'timing/model': 0.39703036000719294, 'learning_rate/qwen_vl_interface': 7.1755052562778215e-06, 'learning_rate/action_model': 7.175505256277822e-05, 'epoch': 0.14}) +[2026-10-03 19:18:31,989][starVLA.training.train_starvla][INFO] - Step 1900, Loss: {'train/loss': 0.02406943589448929, 'train/grad_norm_pre_clip': 52.77016067504883, 'timing/data': 0.00023046997375786304, 'timing/model': 0.3949089100060519, 'learning_rate/qwen_vl_interface': 7.1727208798313704e-06, 'learning_rate/action_model': 7.172720879831371e-05, 'epoch': 0.14}) +[2026-10-03 19:18:32,698][starVLA.training.train_starvla][INFO] - Step 1901, Loss: {'train/loss': 0.0009053725516423583, 'train/grad_norm_pre_clip': 0.28659185767173767, 'timing/data': 0.000298850005492568, 'timing/model': 0.3964952199894469, 'learning_rate/qwen_vl_interface': 7.169935713026994e-06, 'learning_rate/action_model': 7.169935713026994e-05, 'epoch': 0.14}) +[2026-10-03 19:18:33,408][starVLA.training.train_starvla][INFO] - Step 1902, Loss: {'train/loss': 0.0839502289891243, 'train/grad_norm_pre_clip': 36.14149856567383, 'timing/data': 0.00035499001387506723, 'timing/model': 0.39587964001111686, 'learning_rate/qwen_vl_interface': 7.1671497570095684e-06, 'learning_rate/action_model': 7.167149757009568e-05, 'epoch': 0.14}) +[2026-10-03 19:18:34,113][starVLA.training.train_starvla][INFO] - Step 1903, Loss: {'train/loss': 0.004877431783825159, 'train/grad_norm_pre_clip': 6.311208248138428, 'timing/data': 0.00034361000871285796, 'timing/model': 0.39395050998427905, 'learning_rate/qwen_vl_interface': 7.164363012924295e-06, 'learning_rate/action_model': 7.164363012924294e-05, 'epoch': 0.14}) +[2026-10-03 19:18:35,021][starVLA.training.train_starvla][INFO] - Step 1904, Loss: {'train/loss': 0.2046203464269638, 'train/grad_norm_pre_clip': 119.2140121459961, 'timing/data': 0.00037866001366637647, 'timing/model': 0.39597340900218114, 'learning_rate/qwen_vl_interface': 7.161575481916699e-06, 'learning_rate/action_model': 7.1615754819167e-05, 'epoch': 0.14}) +[2026-10-03 19:18:35,728][starVLA.training.train_starvla][INFO] - Step 1905, Loss: {'train/loss': 0.04371856898069382, 'train/grad_norm_pre_clip': 16.328575134277344, 'timing/data': 0.0003004400059580803, 'timing/model': 0.3950214210199192, 'learning_rate/qwen_vl_interface': 7.158787165132627e-06, 'learning_rate/action_model': 7.158787165132626e-05, 'epoch': 0.14}) +[2026-10-03 19:18:36,434][starVLA.training.train_starvla][INFO] - Step 1906, Loss: {'train/loss': 0.2585109770298004, 'train/grad_norm_pre_clip': 56.38255310058594, 'timing/data': 0.0002261199988424778, 'timing/model': 0.3943016300036106, 'learning_rate/qwen_vl_interface': 7.1559980637182535e-06, 'learning_rate/action_model': 7.155998063718254e-05, 'epoch': 0.14}) +[2026-10-03 19:18:37,142][starVLA.training.train_starvla][INFO] - Step 1907, Loss: {'train/loss': 0.07803492993116379, 'train/grad_norm_pre_clip': 29.297435760498047, 'timing/data': 0.00023334001889452338, 'timing/model': 0.39450761998887174, 'learning_rate/qwen_vl_interface': 7.153208178820071e-06, 'learning_rate/action_model': 7.153208178820072e-05, 'epoch': 0.14}) +[2026-10-03 19:18:37,849][starVLA.training.train_starvla][INFO] - Step 1908, Loss: {'train/loss': 0.03264419734477997, 'train/grad_norm_pre_clip': 169.11599731445312, 'timing/data': 0.0003105100186076015, 'timing/model': 0.39400251998449676, 'learning_rate/qwen_vl_interface': 7.150417511584898e-06, 'learning_rate/action_model': 7.150417511584898e-05, 'epoch': 0.14}) +[2026-10-03 19:18:38,754][starVLA.training.train_starvla][INFO] - Step 1909, Loss: {'train/loss': 0.35779881477355957, 'train/grad_norm_pre_clip': 90.67526245117188, 'timing/data': 0.0004376900033093989, 'timing/model': 0.5939937690272927, 'learning_rate/qwen_vl_interface': 7.147626063159872e-06, 'learning_rate/action_model': 7.147626063159872e-05, 'epoch': 0.14}) +[2026-10-03 19:18:39,464][starVLA.training.train_starvla][INFO] - Step 1910, Loss: {'train/loss': 0.00011837389320135117, 'train/grad_norm_pre_clip': 0.02332519181072712, 'timing/data': 0.0003812100039795041, 'timing/model': 0.39624683899455704, 'learning_rate/qwen_vl_interface': 7.14483383469245e-06, 'learning_rate/action_model': 7.144833834692449e-05, 'epoch': 0.14}) +[2026-10-03 19:18:40,172][starVLA.training.train_starvla][INFO] - Step 1911, Loss: {'train/loss': 0.00037507922388613224, 'train/grad_norm_pre_clip': 0.10292401164770126, 'timing/data': 0.00025900002219714224, 'timing/model': 0.39472806997946464, 'learning_rate/qwen_vl_interface': 7.142040827330411e-06, 'learning_rate/action_model': 7.142040827330411e-05, 'epoch': 0.14}) +[2026-10-03 19:18:40,881][starVLA.training.train_starvla][INFO] - Step 1912, Loss: {'train/loss': 0.28927579522132874, 'train/grad_norm_pre_clip': 20.59638023376465, 'timing/data': 0.0003334200009703636, 'timing/model': 0.3961445610038936, 'learning_rate/qwen_vl_interface': 7.139247042221859e-06, 'learning_rate/action_model': 7.139247042221858e-05, 'epoch': 0.14}) +[2026-10-03 19:18:41,590][starVLA.training.train_starvla][INFO] - Step 1913, Loss: {'train/loss': 0.2596729099750519, 'train/grad_norm_pre_clip': 114.46067810058594, 'timing/data': 0.00034048999077640474, 'timing/model': 0.39596648098086007, 'learning_rate/qwen_vl_interface': 7.136452480515208e-06, 'learning_rate/action_model': 7.136452480515208e-05, 'epoch': 0.14}) +[2026-10-03 19:18:42,500][starVLA.training.train_starvla][INFO] - Step 1914, Loss: {'train/loss': 0.9169444441795349, 'train/grad_norm_pre_clip': 205.0323944091797, 'timing/data': 0.0003434999962337315, 'timing/model': 0.5968516800203361, 'learning_rate/qwen_vl_interface': 7.133657143359203e-06, 'learning_rate/action_model': 7.133657143359202e-05, 'epoch': 0.14}) +[2026-10-03 19:18:43,208][starVLA.training.train_starvla][INFO] - Step 1915, Loss: {'train/loss': 0.08007858693599701, 'train/grad_norm_pre_clip': 19.952863693237305, 'timing/data': 0.00022792999516241252, 'timing/model': 0.3959821189928334, 'learning_rate/qwen_vl_interface': 7.130861031902897e-06, 'learning_rate/action_model': 7.130861031902897e-05, 'epoch': 0.14}) +[2026-10-03 19:18:43,915][starVLA.training.train_starvla][INFO] - Step 1916, Loss: {'train/loss': 0.008275558240711689, 'train/grad_norm_pre_clip': 2.4915730953216553, 'timing/data': 0.0003019699943251908, 'timing/model': 0.3960028990113642, 'learning_rate/qwen_vl_interface': 7.128064147295667e-06, 'learning_rate/action_model': 7.128064147295667e-05, 'epoch': 0.14}) +[2026-10-03 19:18:44,622][starVLA.training.train_starvla][INFO] - Step 1917, Loss: {'train/loss': 0.07793641835451126, 'train/grad_norm_pre_clip': 23.50821304321289, 'timing/data': 0.0003762700071092695, 'timing/model': 0.3962915899755899, 'learning_rate/qwen_vl_interface': 7.125266490687206e-06, 'learning_rate/action_model': 7.125266490687206e-05, 'epoch': 0.14}) +[2026-10-03 19:18:45,330][starVLA.training.train_starvla][INFO] - Step 1918, Loss: {'train/loss': 0.021268241107463837, 'train/grad_norm_pre_clip': 8.577676773071289, 'timing/data': 0.0003311999898869544, 'timing/model': 0.3949489099904895, 'learning_rate/qwen_vl_interface': 7.1224680632275265e-06, 'learning_rate/action_model': 7.122468063227527e-05, 'epoch': 0.14}) +[2026-10-03 19:18:46,038][starVLA.training.train_starvla][INFO] - Step 1919, Loss: {'train/loss': 0.6613045930862427, 'train/grad_norm_pre_clip': 38.74702453613281, 'timing/data': 0.00045702900388278067, 'timing/model': 0.39575517998309806, 'learning_rate/qwen_vl_interface': 7.1196688660669555e-06, 'learning_rate/action_model': 7.119668866066955e-05, 'epoch': 0.14}) +[2026-10-03 19:18:46,944][starVLA.training.train_starvla][INFO] - Step 1920, Loss: {'train/loss': 0.0003224175306968391, 'train/grad_norm_pre_clip': 0.10959351062774658, 'timing/data': 0.0003161200147587806, 'timing/model': 0.39584735000971705, 'learning_rate/qwen_vl_interface': 7.116868900356138e-06, 'learning_rate/action_model': 7.116868900356138e-05, 'epoch': 0.14}) +[2026-10-03 19:18:47,649][starVLA.training.train_starvla][INFO] - Step 1921, Loss: {'train/loss': 0.4230678975582123, 'train/grad_norm_pre_clip': 211.39669799804688, 'timing/data': 0.00037033899570815265, 'timing/model': 0.39446040999609977, 'learning_rate/qwen_vl_interface': 7.114068167246033e-06, 'learning_rate/action_model': 7.114068167246033e-05, 'epoch': 0.14}) +[2026-10-03 19:18:48,355][starVLA.training.train_starvla][INFO] - Step 1922, Loss: {'train/loss': 0.0955086350440979, 'train/grad_norm_pre_clip': 24.284486770629883, 'timing/data': 0.00035748002119362354, 'timing/model': 0.39370988999144174, 'learning_rate/qwen_vl_interface': 7.111266667887916e-06, 'learning_rate/action_model': 7.111266667887915e-05, 'epoch': 0.14}) +[2026-10-03 19:18:49,058][starVLA.training.train_starvla][INFO] - Step 1923, Loss: {'train/loss': 0.18888314068317413, 'train/grad_norm_pre_clip': 37.47285842895508, 'timing/data': 0.00033537999843247235, 'timing/model': 0.39390226898831315, 'learning_rate/qwen_vl_interface': 7.108464403433378e-06, 'learning_rate/action_model': 7.108464403433378e-05, 'epoch': 0.14}) +[2026-10-03 19:18:49,760][starVLA.training.train_starvla][INFO] - Step 1924, Loss: {'train/loss': 0.004753127694129944, 'train/grad_norm_pre_clip': 2.105083703994751, 'timing/data': 0.0003609699779190123, 'timing/model': 0.39237961001344956, 'learning_rate/qwen_vl_interface': 7.105661375034324e-06, 'learning_rate/action_model': 7.105661375034324e-05, 'epoch': 0.14}) +[2026-10-03 19:18:50,666][starVLA.training.train_starvla][INFO] - Step 1925, Loss: {'train/loss': 0.0012803857680410147, 'train/grad_norm_pre_clip': 0.7050541043281555, 'timing/data': 0.00034152998705394566, 'timing/model': 0.5940968400100246, 'learning_rate/qwen_vl_interface': 7.102857583842973e-06, 'learning_rate/action_model': 7.102857583842974e-05, 'epoch': 0.14}) +[2026-10-03 19:18:51,371][starVLA.training.train_starvla][INFO] - Step 1926, Loss: {'train/loss': 0.7024484276771545, 'train/grad_norm_pre_clip': 59.05070877075195, 'timing/data': 0.0003739600069820881, 'timing/model': 0.39384567001252435, 'learning_rate/qwen_vl_interface': 7.100053031011859e-06, 'learning_rate/action_model': 7.100053031011859e-05, 'epoch': 0.14}) +[2026-10-03 19:18:52,076][starVLA.training.train_starvla][INFO] - Step 1927, Loss: {'train/loss': 0.028173254802823067, 'train/grad_norm_pre_clip': 16.1544189453125, 'timing/data': 0.0003310500178486109, 'timing/model': 0.3940030800004024, 'learning_rate/qwen_vl_interface': 7.097247717693827e-06, 'learning_rate/action_model': 7.097247717693827e-05, 'epoch': 0.14}) +[2026-10-03 19:18:52,782][starVLA.training.train_starvla][INFO] - Step 1928, Loss: {'train/loss': 0.0005379646318033338, 'train/grad_norm_pre_clip': 0.16507266461849213, 'timing/data': 0.00036537001142278314, 'timing/model': 0.39374046999728307, 'learning_rate/qwen_vl_interface': 7.094441645042033e-06, 'learning_rate/action_model': 7.094441645042032e-05, 'epoch': 0.14}) +[2026-10-03 19:18:53,485][starVLA.training.train_starvla][INFO] - Step 1929, Loss: {'train/loss': 0.3647969663143158, 'train/grad_norm_pre_clip': 34.88416290283203, 'timing/data': 0.00036023001302964985, 'timing/model': 0.3932443000085186, 'learning_rate/qwen_vl_interface': 7.09163481420995e-06, 'learning_rate/action_model': 7.09163481420995e-05, 'epoch': 0.14}) +[2026-10-03 19:18:54,393][starVLA.training.train_starvla][INFO] - Step 1930, Loss: {'train/loss': 0.00022675094078294933, 'train/grad_norm_pre_clip': 0.0397646389901638, 'timing/data': 0.0003455299884080887, 'timing/model': 0.5969467800168786, 'learning_rate/qwen_vl_interface': 7.0888272263513595e-06, 'learning_rate/action_model': 7.088827226351359e-05, 'epoch': 0.14}) +[2026-10-03 19:18:55,098][starVLA.training.train_starvla][INFO] - Step 1931, Loss: {'train/loss': 0.38815733790397644, 'train/grad_norm_pre_clip': 35.72425842285156, 'timing/data': 0.0003042600001208484, 'timing/model': 0.3938920799992047, 'learning_rate/qwen_vl_interface': 7.086018882620354e-06, 'learning_rate/action_model': 7.086018882620354e-05, 'epoch': 0.14}) +[2026-10-03 19:18:55,804][starVLA.training.train_starvla][INFO] - Step 1932, Loss: {'train/loss': 0.021684952080249786, 'train/grad_norm_pre_clip': 4.94477653503418, 'timing/data': 0.000222659989958629, 'timing/model': 0.39342831997782923, 'learning_rate/qwen_vl_interface': 7.083209784171341e-06, 'learning_rate/action_model': 7.083209784171341e-05, 'epoch': 0.14}) +[2026-10-03 19:18:56,508][starVLA.training.train_starvla][INFO] - Step 1933, Loss: {'train/loss': 0.00026902960962615907, 'train/grad_norm_pre_clip': 0.08996611088514328, 'timing/data': 0.00029781999182887375, 'timing/model': 0.3927659800101537, 'learning_rate/qwen_vl_interface': 7.080399932159029e-06, 'learning_rate/action_model': 7.080399932159029e-05, 'epoch': 0.14}) +[2026-10-03 19:18:57,213][starVLA.training.train_starvla][INFO] - Step 1934, Loss: {'train/loss': 0.02281617745757103, 'train/grad_norm_pre_clip': 20.296436309814453, 'timing/data': 0.00022397999418899417, 'timing/model': 0.3940203300153371, 'learning_rate/qwen_vl_interface': 7.077589327738448e-06, 'learning_rate/action_model': 7.077589327738447e-05, 'epoch': 0.14}) +[2026-10-03 19:18:57,915][starVLA.training.train_starvla][INFO] - Step 1935, Loss: {'train/loss': 0.014348194003105164, 'train/grad_norm_pre_clip': 2.467925548553467, 'timing/data': 0.000215409992961213, 'timing/model': 0.391945431008935, 'learning_rate/qwen_vl_interface': 7.0747779720649265e-06, 'learning_rate/action_model': 7.074777972064927e-05, 'epoch': 0.14}) +[2026-10-03 19:18:58,817][starVLA.training.train_starvla][INFO] - Step 1936, Loss: {'train/loss': 0.0012287348508834839, 'train/grad_norm_pre_clip': 1.5395910739898682, 'timing/data': 0.00033964001340791583, 'timing/model': 0.39399622997734696, 'learning_rate/qwen_vl_interface': 7.071965866294108e-06, 'learning_rate/action_model': 7.071965866294108e-05, 'epoch': 0.14}) +[2026-10-03 19:18:59,522][starVLA.training.train_starvla][INFO] - Step 1937, Loss: {'train/loss': 0.0006591043202206492, 'train/grad_norm_pre_clip': 0.26430022716522217, 'timing/data': 0.0003508899826556444, 'timing/model': 0.3932213789958041, 'learning_rate/qwen_vl_interface': 7.0691530115819454e-06, 'learning_rate/action_model': 7.069153011581945e-05, 'epoch': 0.14}) +[2026-10-03 19:19:00,226][starVLA.training.train_starvla][INFO] - Step 1938, Loss: {'train/loss': 0.008322094567120075, 'train/grad_norm_pre_clip': 1.2081276178359985, 'timing/data': 0.00032533000921830535, 'timing/model': 0.39371293000294827, 'learning_rate/qwen_vl_interface': 7.066339409084694e-06, 'learning_rate/action_model': 7.066339409084694e-05, 'epoch': 0.14}) +[2026-10-03 19:19:00,931][starVLA.training.train_starvla][INFO] - Step 1939, Loss: {'train/loss': 0.2772563397884369, 'train/grad_norm_pre_clip': 56.21189498901367, 'timing/data': 0.0002504599979147315, 'timing/model': 0.3940661709930282, 'learning_rate/qwen_vl_interface': 7.0635250599589215e-06, 'learning_rate/action_model': 7.06352505995892e-05, 'epoch': 0.14}) +[2026-10-03 19:19:01,634][starVLA.training.train_starvla][INFO] - Step 1940, Loss: {'train/loss': 0.4326590597629547, 'train/grad_norm_pre_clip': 47.795284271240234, 'timing/data': 0.0004452900029718876, 'timing/model': 0.39254801999777555, 'learning_rate/qwen_vl_interface': 7.060709965361499e-06, 'learning_rate/action_model': 7.060709965361499e-05, 'epoch': 0.14}) +[2026-10-03 19:19:02,540][starVLA.training.train_starvla][INFO] - Step 1941, Loss: {'train/loss': 1.1665199995040894, 'train/grad_norm_pre_clip': 271.3399353027344, 'timing/data': 0.0003212899900972843, 'timing/model': 0.3945151199877728, 'learning_rate/qwen_vl_interface': 7.057894126449606e-06, 'learning_rate/action_model': 7.057894126449606e-05, 'epoch': 0.14}) +[2026-10-03 19:19:03,243][starVLA.training.train_starvla][INFO] - Step 1942, Loss: {'train/loss': 1.514108657836914, 'train/grad_norm_pre_clip': 107.72706604003906, 'timing/data': 0.0004167099832557142, 'timing/model': 0.3940510600223206, 'learning_rate/qwen_vl_interface': 7.055077544380728e-06, 'learning_rate/action_model': 7.055077544380728e-05, 'epoch': 0.14}) +[2026-10-03 19:19:03,946][starVLA.training.train_starvla][INFO] - Step 1943, Loss: {'train/loss': 0.21736598014831543, 'train/grad_norm_pre_clip': 15.871604919433594, 'timing/data': 0.0003503700136207044, 'timing/model': 0.39261588000226766, 'learning_rate/qwen_vl_interface': 7.0522602203126554e-06, 'learning_rate/action_model': 7.052260220312656e-05, 'epoch': 0.14}) +[2026-10-03 19:19:04,650][starVLA.training.train_starvla][INFO] - Step 1944, Loss: {'train/loss': 0.23582027852535248, 'train/grad_norm_pre_clip': 32.46940612792969, 'timing/data': 0.0002273900026921183, 'timing/model': 0.3949444099853281, 'learning_rate/qwen_vl_interface': 7.0494421554034845e-06, 'learning_rate/action_model': 7.049442155403484e-05, 'epoch': 0.14}) +[2026-10-03 19:19:05,358][starVLA.training.train_starvla][INFO] - Step 1945, Loss: {'train/loss': 0.3093196153640747, 'train/grad_norm_pre_clip': 130.62051391601562, 'timing/data': 0.0002257299784105271, 'timing/model': 0.3967098699940834, 'learning_rate/qwen_vl_interface': 7.046623350811612e-06, 'learning_rate/action_model': 7.046623350811612e-05, 'epoch': 0.14}) +[2026-10-03 19:19:06,269][starVLA.training.train_starvla][INFO] - Step 1946, Loss: {'train/loss': 0.0015535589773207903, 'train/grad_norm_pre_clip': 0.21988897025585175, 'timing/data': 0.0003694900078698993, 'timing/model': 0.5983859489788301, 'learning_rate/qwen_vl_interface': 7.043803807695747e-06, 'learning_rate/action_model': 7.043803807695747e-05, 'epoch': 0.14}) +[2026-10-03 19:19:06,973][starVLA.training.train_starvla][INFO] - Step 1947, Loss: {'train/loss': 0.0009919266449287534, 'train/grad_norm_pre_clip': 0.15892130136489868, 'timing/data': 0.0003632000007200986, 'timing/model': 0.3927375600032974, 'learning_rate/qwen_vl_interface': 7.040983527214893e-06, 'learning_rate/action_model': 7.040983527214893e-05, 'epoch': 0.14}) +[2026-10-03 19:19:07,680][starVLA.training.train_starvla][INFO] - Step 1948, Loss: {'train/loss': 0.21136733889579773, 'train/grad_norm_pre_clip': 91.8949966430664, 'timing/data': 0.0002261199988424778, 'timing/model': 0.3947271590004675, 'learning_rate/qwen_vl_interface': 7.038162510528363e-06, 'learning_rate/action_model': 7.038162510528362e-05, 'epoch': 0.14}) +[2026-10-03 19:19:08,386][starVLA.training.train_starvla][INFO] - Step 1949, Loss: {'train/loss': 0.8696120381355286, 'train/grad_norm_pre_clip': 46.929786682128906, 'timing/data': 0.00022888998501002789, 'timing/model': 0.39535996998893097, 'learning_rate/qwen_vl_interface': 7.03534075879577e-06, 'learning_rate/action_model': 7.03534075879577e-05, 'epoch': 0.14}) +[2026-10-03 19:19:09,091][starVLA.training.train_starvla][INFO] - Step 1950, Loss: {'train/loss': 0.009895049035549164, 'train/grad_norm_pre_clip': 13.132078170776367, 'timing/data': 0.0002907299785874784, 'timing/model': 0.3943636189796962, 'learning_rate/qwen_vl_interface': 7.032518273177028e-06, 'learning_rate/action_model': 7.032518273177028e-05, 'epoch': 0.14}) +[2026-10-03 19:19:09,999][starVLA.training.train_starvla][INFO] - Step 1951, Loss: {'train/loss': 0.0028443399351090193, 'train/grad_norm_pre_clip': 1.5879335403442383, 'timing/data': 0.000338789977831766, 'timing/model': 0.5965770889888518, 'learning_rate/qwen_vl_interface': 7.029695054832357e-06, 'learning_rate/action_model': 7.029695054832357e-05, 'epoch': 0.14}) +[2026-10-03 19:19:10,703][starVLA.training.train_starvla][INFO] - Step 1952, Loss: {'train/loss': 0.4819667637348175, 'train/grad_norm_pre_clip': 154.056884765625, 'timing/data': 0.0003266790008638054, 'timing/model': 0.3927934809762519, 'learning_rate/qwen_vl_interface': 7.026871104922273e-06, 'learning_rate/action_model': 7.026871104922273e-05, 'epoch': 0.14}) +[2026-10-03 19:19:11,405][starVLA.training.train_starvla][INFO] - Step 1953, Loss: {'train/loss': 0.6719655394554138, 'train/grad_norm_pre_clip': 77.5765609741211, 'timing/data': 0.0002581999870017171, 'timing/model': 0.3933371000166517, 'learning_rate/qwen_vl_interface': 7.024046424607596e-06, 'learning_rate/action_model': 7.024046424607595e-05, 'epoch': 0.14}) +[2026-10-03 19:19:12,112][starVLA.training.train_starvla][INFO] - Step 1954, Loss: {'train/loss': 0.42945951223373413, 'train/grad_norm_pre_clip': 62.90530014038086, 'timing/data': 0.0003543199854902923, 'timing/model': 0.3942842000105884, 'learning_rate/qwen_vl_interface': 7.021221015049445e-06, 'learning_rate/action_model': 7.021221015049446e-05, 'epoch': 0.14}) +[2026-10-03 19:19:12,818][starVLA.training.train_starvla][INFO] - Step 1955, Loss: {'train/loss': 0.08017753064632416, 'train/grad_norm_pre_clip': 73.8990707397461, 'timing/data': 0.000328330002957955, 'timing/model': 0.39446524999220856, 'learning_rate/qwen_vl_interface': 7.01839487740924e-06, 'learning_rate/action_model': 7.01839487740924e-05, 'epoch': 0.14}) +[2026-10-03 19:19:13,724][starVLA.training.train_starvla][INFO] - Step 1956, Loss: {'train/loss': 0.0016727359034121037, 'train/grad_norm_pre_clip': 0.310308575630188, 'timing/data': 0.00023407000117003918, 'timing/model': 0.5946280790085439, 'learning_rate/qwen_vl_interface': 7.015568012848699e-06, 'learning_rate/action_model': 7.015568012848699e-05, 'epoch': 0.14}) +[2026-10-03 19:19:14,430][starVLA.training.train_starvla][INFO] - Step 1957, Loss: {'train/loss': 0.25773629546165466, 'train/grad_norm_pre_clip': 30.171762466430664, 'timing/data': 0.00029085000278428197, 'timing/model': 0.3949662400118541, 'learning_rate/qwen_vl_interface': 7.01274042252984e-06, 'learning_rate/action_model': 7.01274042252984e-05, 'epoch': 0.14}) +[2026-10-03 19:19:15,135][starVLA.training.train_starvla][INFO] - Step 1958, Loss: {'train/loss': 0.004841410554945469, 'train/grad_norm_pre_clip': 4.273614883422852, 'timing/data': 0.0003619699855335057, 'timing/model': 0.3940583299845457, 'learning_rate/qwen_vl_interface': 7.009912107614978e-06, 'learning_rate/action_model': 7.009912107614978e-05, 'epoch': 0.14}) +[2026-10-03 19:19:15,841][starVLA.training.train_starvla][INFO] - Step 1959, Loss: {'train/loss': 0.3467969596385956, 'train/grad_norm_pre_clip': 129.2413787841797, 'timing/data': 0.000222659989958629, 'timing/model': 0.3949193400039803, 'learning_rate/qwen_vl_interface': 7.0070830692667244e-06, 'learning_rate/action_model': 7.007083069266724e-05, 'epoch': 0.14}) +[2026-10-03 19:19:16,545][starVLA.training.train_starvla][INFO] - Step 1960, Loss: {'train/loss': 0.009183710440993309, 'train/grad_norm_pre_clip': 8.134739875793457, 'timing/data': 0.00035208999179303646, 'timing/model': 0.39445063000312075, 'learning_rate/qwen_vl_interface': 7.004253308647993e-06, 'learning_rate/action_model': 7.004253308647994e-05, 'epoch': 0.14}) +[2026-10-03 19:19:17,252][starVLA.training.train_starvla][INFO] - Step 1961, Loss: {'train/loss': 0.36392998695373535, 'train/grad_norm_pre_clip': 11.92253589630127, 'timing/data': 0.00033991000964306295, 'timing/model': 0.39499925999552943, 'learning_rate/qwen_vl_interface': 7.001422826921991e-06, 'learning_rate/action_model': 7.00142282692199e-05, 'epoch': 0.14}) +[2026-10-03 19:19:18,168][starVLA.training.train_starvla][INFO] - Step 1962, Loss: {'train/loss': 0.23178008198738098, 'train/grad_norm_pre_clip': 18.460433959960938, 'timing/data': 0.0003760599938686937, 'timing/model': 0.39364165998995304, 'learning_rate/qwen_vl_interface': 6.9985916252522225e-06, 'learning_rate/action_model': 6.998591625252222e-05, 'epoch': 0.14}) +[2026-10-03 19:19:18,875][starVLA.training.train_starvla][INFO] - Step 1963, Loss: {'train/loss': 0.0003340461407788098, 'train/grad_norm_pre_clip': 0.1557730734348297, 'timing/data': 0.0003696400090120733, 'timing/model': 0.3961807500163559, 'learning_rate/qwen_vl_interface': 6.995759704802485e-06, 'learning_rate/action_model': 6.995759704802485e-05, 'epoch': 0.14}) +[2026-10-03 19:19:19,581][starVLA.training.train_starvla][INFO] - Step 1964, Loss: {'train/loss': 0.2229425013065338, 'train/grad_norm_pre_clip': 234.6658477783203, 'timing/data': 0.00023351999698206782, 'timing/model': 0.39423850999446586, 'learning_rate/qwen_vl_interface': 6.992927066736876e-06, 'learning_rate/action_model': 6.992927066736875e-05, 'epoch': 0.14}) +[2026-10-03 19:19:20,287][starVLA.training.train_starvla][INFO] - Step 1965, Loss: {'train/loss': 0.004059765487909317, 'train/grad_norm_pre_clip': 2.4381062984466553, 'timing/data': 0.0003182900254614651, 'timing/model': 0.39535868001985364, 'learning_rate/qwen_vl_interface': 6.990093712219787e-06, 'learning_rate/action_model': 6.990093712219787e-05, 'epoch': 0.14}) +[2026-10-03 19:19:20,992][starVLA.training.train_starvla][INFO] - Step 1966, Loss: {'train/loss': 0.09001865237951279, 'train/grad_norm_pre_clip': 43.72418975830078, 'timing/data': 0.00034426001366227865, 'timing/model': 0.39411937099066563, 'learning_rate/qwen_vl_interface': 6.9872596424159e-06, 'learning_rate/action_model': 6.9872596424159e-05, 'epoch': 0.14}) +[2026-10-03 19:19:21,899][starVLA.training.train_starvla][INFO] - Step 1967, Loss: {'train/loss': 0.5058922171592712, 'train/grad_norm_pre_clip': 24.840892791748047, 'timing/data': 0.00034063999191857874, 'timing/model': 0.5971382200077642, 'learning_rate/qwen_vl_interface': 6.9844248584902e-06, 'learning_rate/action_model': 6.984424858490199e-05, 'epoch': 0.14}) +[2026-10-03 19:19:22,604][starVLA.training.train_starvla][INFO] - Step 1968, Loss: {'train/loss': 0.2565608322620392, 'train/grad_norm_pre_clip': 18.334047317504883, 'timing/data': 0.0003497999859973788, 'timing/model': 0.3936367900168989, 'learning_rate/qwen_vl_interface': 6.981589361607952e-06, 'learning_rate/action_model': 6.981589361607953e-05, 'epoch': 0.14}) +[2026-10-03 19:19:23,308][starVLA.training.train_starvla][INFO] - Step 1969, Loss: {'train/loss': 0.1908147782087326, 'train/grad_norm_pre_clip': 36.4857177734375, 'timing/data': 0.00027560998569242656, 'timing/model': 0.3942731999850366, 'learning_rate/qwen_vl_interface': 6.9787531529347274e-06, 'learning_rate/action_model': 6.978753152934727e-05, 'epoch': 0.14}) +[2026-10-03 19:19:24,013][starVLA.training.train_starvla][INFO] - Step 1970, Loss: {'train/loss': 0.18446089327335358, 'train/grad_norm_pre_clip': 19.598323822021484, 'timing/data': 0.0003410399949643761, 'timing/model': 0.3938591899932362, 'learning_rate/qwen_vl_interface': 6.97591623363638e-06, 'learning_rate/action_model': 6.97591623363638e-05, 'epoch': 0.14}) +[2026-10-03 19:19:24,716][starVLA.training.train_starvla][INFO] - Step 1971, Loss: {'train/loss': 0.08475811779499054, 'train/grad_norm_pre_clip': 71.33606719970703, 'timing/data': 0.00024894002126529813, 'timing/model': 0.3920944200071972, 'learning_rate/qwen_vl_interface': 6.9730786048790644e-06, 'learning_rate/action_model': 6.973078604879064e-05, 'epoch': 0.14}) +[2026-10-03 19:19:25,420][starVLA.training.train_starvla][INFO] - Step 1972, Loss: {'train/loss': 0.001975785009562969, 'train/grad_norm_pre_clip': 0.20310699939727783, 'timing/data': 0.00033055999665521085, 'timing/model': 0.39360904099885374, 'learning_rate/qwen_vl_interface': 6.970240267829221e-06, 'learning_rate/action_model': 6.970240267829222e-05, 'epoch': 0.14}) +[2026-10-03 19:19:26,330][starVLA.training.train_starvla][INFO] - Step 1973, Loss: {'train/loss': 0.0032064435072243214, 'train/grad_norm_pre_clip': 0.39869558811187744, 'timing/data': 0.00022767001064494252, 'timing/model': 0.39559461001772434, 'learning_rate/qwen_vl_interface': 6.967401223653582e-06, 'learning_rate/action_model': 6.967401223653581e-05, 'epoch': 0.14}) +[2026-10-03 19:19:27,037][starVLA.training.train_starvla][INFO] - Step 1974, Loss: {'train/loss': 0.004507087171077728, 'train/grad_norm_pre_clip': 0.8413785696029663, 'timing/data': 0.00022399998852051795, 'timing/model': 0.3951938599930145, 'learning_rate/qwen_vl_interface': 6.9645614735191724e-06, 'learning_rate/action_model': 6.964561473519172e-05, 'epoch': 0.14}) +[2026-10-03 19:19:27,746][starVLA.training.train_starvla][INFO] - Step 1975, Loss: {'train/loss': 0.3036763370037079, 'train/grad_norm_pre_clip': 29.771726608276367, 'timing/data': 0.0003389899793546647, 'timing/model': 0.39524233900010586, 'learning_rate/qwen_vl_interface': 6.961721018593305e-06, 'learning_rate/action_model': 6.961721018593305e-05, 'epoch': 0.14}) +[2026-10-03 19:19:28,452][starVLA.training.train_starvla][INFO] - Step 1976, Loss: {'train/loss': 0.00693520437926054, 'train/grad_norm_pre_clip': 3.965670108795166, 'timing/data': 0.0003457400016486645, 'timing/model': 0.39380458000232466, 'learning_rate/qwen_vl_interface': 6.958879860043585e-06, 'learning_rate/action_model': 6.958879860043585e-05, 'epoch': 0.14}) +[2026-10-03 19:19:29,157][starVLA.training.train_starvla][INFO] - Step 1977, Loss: {'train/loss': 0.10725826770067215, 'train/grad_norm_pre_clip': 23.78419303894043, 'timing/data': 0.0002866799768526107, 'timing/model': 0.39428330000373535, 'learning_rate/qwen_vl_interface': 6.9560379990379025e-06, 'learning_rate/action_model': 6.956037999037902e-05, 'epoch': 0.14}) +[2026-10-03 19:19:30,067][starVLA.training.train_starvla][INFO] - Step 1978, Loss: {'train/loss': 0.001856536604464054, 'train/grad_norm_pre_clip': 0.33757272362709045, 'timing/data': 0.0003321699914522469, 'timing/model': 0.39580189998378046, 'learning_rate/qwen_vl_interface': 6.953195436744442e-06, 'learning_rate/action_model': 6.953195436744442e-05, 'epoch': 0.14}) +[2026-10-03 19:19:30,777][starVLA.training.train_starvla][INFO] - Step 1979, Loss: {'train/loss': 0.0005697954911738634, 'train/grad_norm_pre_clip': 0.12493660300970078, 'timing/data': 0.0002819000219460577, 'timing/model': 0.3964145600039046, 'learning_rate/qwen_vl_interface': 6.950352174331676e-06, 'learning_rate/action_model': 6.950352174331675e-05, 'epoch': 0.14}) +[2026-10-03 19:19:31,482][starVLA.training.train_starvla][INFO] - Step 1980, Loss: {'train/loss': 0.009952791035175323, 'train/grad_norm_pre_clip': 1.8996469974517822, 'timing/data': 0.00024041999131441116, 'timing/model': 0.3946286909922492, 'learning_rate/qwen_vl_interface': 6.947508212968356e-06, 'learning_rate/action_model': 6.947508212968356e-05, 'epoch': 0.14}) +[2026-10-03 19:19:32,187][starVLA.training.train_starvla][INFO] - Step 1981, Loss: {'train/loss': 0.09862722456455231, 'train/grad_norm_pre_clip': 66.60508728027344, 'timing/data': 0.00033706999965943396, 'timing/model': 0.39423184999031946, 'learning_rate/qwen_vl_interface': 6.9446635538235295e-06, 'learning_rate/action_model': 6.944663553823529e-05, 'epoch': 0.14}) +[2026-10-03 19:19:32,889][starVLA.training.train_starvla][INFO] - Step 1982, Loss: {'train/loss': 0.11699559539556503, 'train/grad_norm_pre_clip': 88.78919982910156, 'timing/data': 0.000281469983747229, 'timing/model': 0.3916930599953048, 'learning_rate/qwen_vl_interface': 6.94181819806653e-06, 'learning_rate/action_model': 6.94181819806653e-05, 'epoch': 0.15}) +[2026-10-03 19:19:33,797][starVLA.training.train_starvla][INFO] - Step 1983, Loss: {'train/loss': 0.09512145072221756, 'train/grad_norm_pre_clip': 48.20305252075195, 'timing/data': 0.00025988000561483204, 'timing/model': 0.595831698999973, 'learning_rate/qwen_vl_interface': 6.938972146866975e-06, 'learning_rate/action_model': 6.938972146866974e-05, 'epoch': 0.15}) +[2026-10-03 19:19:34,500][starVLA.training.train_starvla][INFO] - Step 1984, Loss: {'train/loss': 0.8754096031188965, 'train/grad_norm_pre_clip': 88.39989471435547, 'timing/data': 0.00022431998513638973, 'timing/model': 0.3934726700244937, 'learning_rate/qwen_vl_interface': 6.936125401394769e-06, 'learning_rate/action_model': 6.936125401394768e-05, 'epoch': 0.15}) +[2026-10-03 19:19:35,207][starVLA.training.train_starvla][INFO] - Step 1985, Loss: {'train/loss': 0.014529025182127953, 'train/grad_norm_pre_clip': 4.252630710601807, 'timing/data': 0.00034065000363625586, 'timing/model': 0.3958554400014691, 'learning_rate/qwen_vl_interface': 6.933277962820101e-06, 'learning_rate/action_model': 6.9332779628201e-05, 'epoch': 0.15}) +[2026-10-03 19:19:35,910][starVLA.training.train_starvla][INFO] - Step 1986, Loss: {'train/loss': 0.017934169620275497, 'train/grad_norm_pre_clip': 21.474027633666992, 'timing/data': 0.0003368199977558106, 'timing/model': 0.3943610700080171, 'learning_rate/qwen_vl_interface': 6.930429832313444e-06, 'learning_rate/action_model': 6.930429832313444e-05, 'epoch': 0.15}) +[2026-10-03 19:19:36,614][starVLA.training.train_starvla][INFO] - Step 1987, Loss: {'train/loss': 0.0018359936075285077, 'train/grad_norm_pre_clip': 1.1979435682296753, 'timing/data': 0.00025342998560518026, 'timing/model': 0.39232337099383585, 'learning_rate/qwen_vl_interface': 6.92758101104556e-06, 'learning_rate/action_model': 6.92758101104556e-05, 'epoch': 0.15}) +[2026-10-03 19:19:37,320][starVLA.training.train_starvla][INFO] - Step 1988, Loss: {'train/loss': 0.23355750739574432, 'train/grad_norm_pre_clip': 45.3841438293457, 'timing/data': 0.00030274997698143125, 'timing/model': 0.39409587997943163, 'learning_rate/qwen_vl_interface': 6.9247315001874885e-06, 'learning_rate/action_model': 6.924731500187488e-05, 'epoch': 0.15}) +[2026-10-03 19:19:38,223][starVLA.training.train_starvla][INFO] - Step 1989, Loss: {'train/loss': 0.3866540193557739, 'train/grad_norm_pre_clip': 93.0422592163086, 'timing/data': 0.00028368999483063817, 'timing/model': 0.393158909981139, 'learning_rate/qwen_vl_interface': 6.9218813009105605e-06, 'learning_rate/action_model': 6.92188130091056e-05, 'epoch': 0.15}) +[2026-10-03 19:19:38,929][starVLA.training.train_starvla][INFO] - Step 1990, Loss: {'train/loss': 0.3566374182701111, 'train/grad_norm_pre_clip': 29.90156364440918, 'timing/data': 0.00034566002432256937, 'timing/model': 0.3945224000199232, 'learning_rate/qwen_vl_interface': 6.919030414386378e-06, 'learning_rate/action_model': 6.919030414386379e-05, 'epoch': 0.15}) +[2026-10-03 19:19:39,633][starVLA.training.train_starvla][INFO] - Step 1991, Loss: {'train/loss': 0.20994530618190765, 'train/grad_norm_pre_clip': 120.79447937011719, 'timing/data': 0.00036756001645699143, 'timing/model': 0.3938213690125849, 'learning_rate/qwen_vl_interface': 6.916178841786839e-06, 'learning_rate/action_model': 6.916178841786838e-05, 'epoch': 0.15}) +[2026-10-03 19:19:40,337][starVLA.training.train_starvla][INFO] - Step 1992, Loss: {'train/loss': 0.018097490072250366, 'train/grad_norm_pre_clip': 5.193547248840332, 'timing/data': 0.00033430999610573053, 'timing/model': 0.3941933099995367, 'learning_rate/qwen_vl_interface': 6.913326584284115e-06, 'learning_rate/action_model': 6.913326584284115e-05, 'epoch': 0.15}) +[2026-10-03 19:19:41,017][starVLA.training.train_starvla][INFO] - Step 1993, Loss: {'train/loss': 1.5214691162109375, 'train/grad_norm_pre_clip': 154.80589294433594, 'timing/data': 0.00034776999382302165, 'timing/model': 0.3687742329784669, 'learning_rate/qwen_vl_interface': 6.910473643050662e-06, 'learning_rate/action_model': 6.910473643050662e-05, 'epoch': 0.15}) +[2026-10-03 19:19:41,928][starVLA.training.train_starvla][INFO] - Step 1994, Loss: {'train/loss': 0.006575245875865221, 'train/grad_norm_pre_clip': 1.2267849445343018, 'timing/data': 0.0002475100045558065, 'timing/model': 0.39444999999250285, 'learning_rate/qwen_vl_interface': 6.907620019259215e-06, 'learning_rate/action_model': 6.907620019259216e-05, 'epoch': 0.15}) +[2026-10-03 19:19:42,633][starVLA.training.train_starvla][INFO] - Step 1995, Loss: {'train/loss': 0.02465003728866577, 'train/grad_norm_pre_clip': 13.031733512878418, 'timing/data': 0.0003025599871762097, 'timing/model': 0.39386832001036964, 'learning_rate/qwen_vl_interface': 6.9047657140827935e-06, 'learning_rate/action_model': 6.904765714082794e-05, 'epoch': 0.15}) +[2026-10-03 19:19:43,337][starVLA.training.train_starvla][INFO] - Step 1996, Loss: {'train/loss': 0.010452684946358204, 'train/grad_norm_pre_clip': 11.174349784851074, 'timing/data': 0.00032871999428607523, 'timing/model': 0.3944319699949119, 'learning_rate/qwen_vl_interface': 6.901910728694694e-06, 'learning_rate/action_model': 6.901910728694694e-05, 'epoch': 0.15}) +[2026-10-03 19:19:44,041][starVLA.training.train_starvla][INFO] - Step 1997, Loss: {'train/loss': 0.002863002475351095, 'train/grad_norm_pre_clip': 1.3336793184280396, 'timing/data': 0.0003227099950890988, 'timing/model': 0.39344422999420203, 'learning_rate/qwen_vl_interface': 6.899055064268489e-06, 'learning_rate/action_model': 6.899055064268489e-05, 'epoch': 0.15}) +[2026-10-03 19:19:44,745][starVLA.training.train_starvla][INFO] - Step 1998, Loss: {'train/loss': 0.9984976053237915, 'train/grad_norm_pre_clip': 78.59590148925781, 'timing/data': 0.00033169001108035445, 'timing/model': 0.39425605998258106, 'learning_rate/qwen_vl_interface': 6.896198721978038e-06, 'learning_rate/action_model': 6.896198721978038e-05, 'epoch': 0.15}) +[2026-10-03 19:19:45,653][starVLA.training.train_starvla][INFO] - Step 1999, Loss: {'train/loss': 0.1626460701227188, 'train/grad_norm_pre_clip': 23.433948516845703, 'timing/data': 0.00023034997866488993, 'timing/model': 0.5972956189943943, 'learning_rate/qwen_vl_interface': 6.893341702997475e-06, 'learning_rate/action_model': 6.893341702997475e-05, 'epoch': 0.15}) +[2026-10-03 19:19:46,355][starVLA.training.train_starvla][INFO] - Step 2000, Loss: {'train/loss': 0.4440203905105591, 'train/grad_norm_pre_clip': 203.2284393310547, 'timing/data': 0.0003283499972894788, 'timing/model': 0.392154989996925, 'learning_rate/qwen_vl_interface': 6.890484008501211e-06, 'learning_rate/action_model': 6.890484008501211e-05, 'epoch': 0.15}) +[2026-10-03 19:19:47,059][starVLA.training.train_starvla][INFO] - Step 2001, Loss: {'train/loss': 0.0006787549937143922, 'train/grad_norm_pre_clip': 0.17687903344631195, 'timing/data': 0.00030236999737098813, 'timing/model': 0.39470051997341216, 'learning_rate/qwen_vl_interface': 6.887625639663941e-06, 'learning_rate/action_model': 6.88762563966394e-05, 'epoch': 0.15}) +[2026-10-03 19:19:47,764][starVLA.training.train_starvla][INFO] - Step 2002, Loss: {'train/loss': 0.001206682762131095, 'train/grad_norm_pre_clip': 0.6704118251800537, 'timing/data': 0.00032811000710353255, 'timing/model': 0.3945029790047556, 'learning_rate/qwen_vl_interface': 6.884766597660626e-06, 'learning_rate/action_model': 6.884766597660626e-05, 'epoch': 0.15}) +[2026-10-03 19:19:48,467][starVLA.training.train_starvla][INFO] - Step 2003, Loss: {'train/loss': 0.19056124985218048, 'train/grad_norm_pre_clip': 46.004798889160156, 'timing/data': 0.0003772999916691333, 'timing/model': 0.39296964000095613, 'learning_rate/qwen_vl_interface': 6.881906883666515e-06, 'learning_rate/action_model': 6.881906883666515e-05, 'epoch': 0.15}) +[2026-10-03 19:19:49,374][starVLA.training.train_starvla][INFO] - Step 2004, Loss: {'train/loss': 0.07392019778490067, 'train/grad_norm_pre_clip': 62.32738494873047, 'timing/data': 0.00023303998750634491, 'timing/model': 0.597434239985887, 'learning_rate/qwen_vl_interface': 6.879046498857125e-06, 'learning_rate/action_model': 6.879046498857124e-05, 'epoch': 0.15}) +[2026-10-03 19:19:50,079][starVLA.training.train_starvla][INFO] - Step 2005, Loss: {'train/loss': 0.2838782072067261, 'train/grad_norm_pre_clip': 105.50190734863281, 'timing/data': 0.0003617999900598079, 'timing/model': 0.39429038000525907, 'learning_rate/qwen_vl_interface': 6.876185444408255e-06, 'learning_rate/action_model': 6.876185444408255e-05, 'epoch': 0.15}) +[2026-10-03 19:19:50,784][starVLA.training.train_starvla][INFO] - Step 2006, Loss: {'train/loss': 0.007203117944300175, 'train/grad_norm_pre_clip': 5.909570693969727, 'timing/data': 0.00034200999652966857, 'timing/model': 0.39486468001268804, 'learning_rate/qwen_vl_interface': 6.873323721495976e-06, 'learning_rate/action_model': 6.873323721495976e-05, 'epoch': 0.15}) +[2026-10-03 19:19:51,489][starVLA.training.train_starvla][INFO] - Step 2007, Loss: {'train/loss': 0.8003162741661072, 'train/grad_norm_pre_clip': 58.269710540771484, 'timing/data': 0.00026760000037029386, 'timing/model': 0.39414957899134606, 'learning_rate/qwen_vl_interface': 6.870461331296632e-06, 'learning_rate/action_model': 6.870461331296632e-05, 'epoch': 0.15}) +[2026-10-03 19:19:52,193][starVLA.training.train_starvla][INFO] - Step 2008, Loss: {'train/loss': 0.5367609858512878, 'train/grad_norm_pre_clip': 164.56455993652344, 'timing/data': 0.00033526000333949924, 'timing/model': 0.39379539000219665, 'learning_rate/qwen_vl_interface': 6.8675982749868455e-06, 'learning_rate/action_model': 6.867598274986845e-05, 'epoch': 0.15}) +[2026-10-03 19:19:52,899][starVLA.training.train_starvla][INFO] - Step 2009, Loss: {'train/loss': 0.0013444104697555304, 'train/grad_norm_pre_clip': 0.8354361057281494, 'timing/data': 0.00031013000989332795, 'timing/model': 0.3943340289988555, 'learning_rate/qwen_vl_interface': 6.864734553743509e-06, 'learning_rate/action_model': 6.86473455374351e-05, 'epoch': 0.15}) +[2026-10-03 19:19:53,803][starVLA.training.train_starvla][INFO] - Step 2010, Loss: {'train/loss': 0.010355781763792038, 'train/grad_norm_pre_clip': 3.517132043838501, 'timing/data': 0.0002600599837023765, 'timing/model': 0.39421877000131644, 'learning_rate/qwen_vl_interface': 6.861870168743793e-06, 'learning_rate/action_model': 6.861870168743793e-05, 'epoch': 0.15}) +[2026-10-03 19:19:54,509][starVLA.training.train_starvla][INFO] - Step 2011, Loss: {'train/loss': 0.012225804850459099, 'train/grad_norm_pre_clip': 4.905484676361084, 'timing/data': 0.00027118000434711576, 'timing/model': 0.3944599399983417, 'learning_rate/qwen_vl_interface': 6.859005121165135e-06, 'learning_rate/action_model': 6.859005121165135e-05, 'epoch': 0.15}) +[2026-10-03 19:19:55,214][starVLA.training.train_starvla][INFO] - Step 2012, Loss: {'train/loss': 0.5671034455299377, 'train/grad_norm_pre_clip': 28.90537452697754, 'timing/data': 0.0003277899813838303, 'timing/model': 0.3932288409851026, 'learning_rate/qwen_vl_interface': 6.85613941218525e-06, 'learning_rate/action_model': 6.85613941218525e-05, 'epoch': 0.15}) +[2026-10-03 19:19:55,920][starVLA.training.train_starvla][INFO] - Step 2013, Loss: {'train/loss': 1.0371936559677124, 'train/grad_norm_pre_clip': 146.0992889404297, 'timing/data': 0.0003434100071899593, 'timing/model': 0.39506218998576514, 'learning_rate/qwen_vl_interface': 6.8532730429821215e-06, 'learning_rate/action_model': 6.853273042982122e-05, 'epoch': 0.15}) +[2026-10-03 19:19:56,624][starVLA.training.train_starvla][INFO] - Step 2014, Loss: {'train/loss': 0.2445879578590393, 'train/grad_norm_pre_clip': 48.32467269897461, 'timing/data': 0.0003728199808392674, 'timing/model': 0.39266724101617, 'learning_rate/qwen_vl_interface': 6.850406014734005e-06, 'learning_rate/action_model': 6.850406014734005e-05, 'epoch': 0.15}) +[2026-10-03 19:19:57,525][starVLA.training.train_starvla][INFO] - Step 2015, Loss: {'train/loss': 0.014759663492441177, 'train/grad_norm_pre_clip': 6.877967834472656, 'timing/data': 0.0003100000030826777, 'timing/model': 0.3942763500090223, 'learning_rate/qwen_vl_interface': 6.8475383286194295e-06, 'learning_rate/action_model': 6.84753832861943e-05, 'epoch': 0.15}) +[2026-10-03 19:19:58,230][starVLA.training.train_starvla][INFO] - Step 2016, Loss: {'train/loss': 0.19842202961444855, 'train/grad_norm_pre_clip': 59.21070861816406, 'timing/data': 0.0003423999878577888, 'timing/model': 0.39402917999541387, 'learning_rate/qwen_vl_interface': 6.844669985817192e-06, 'learning_rate/action_model': 6.844669985817192e-05, 'epoch': 0.15}) +[2026-10-03 19:19:58,935][starVLA.training.train_starvla][INFO] - Step 2017, Loss: {'train/loss': 0.04018070921301842, 'train/grad_norm_pre_clip': 38.702659606933594, 'timing/data': 0.0003533890121616423, 'timing/model': 0.39459164001164027, 'learning_rate/qwen_vl_interface': 6.8418009875063594e-06, 'learning_rate/action_model': 6.84180098750636e-05, 'epoch': 0.15}) +[2026-10-03 19:19:59,640][starVLA.training.train_starvla][INFO] - Step 2018, Loss: {'train/loss': 0.00038825214141979814, 'train/grad_norm_pre_clip': 0.11314927041530609, 'timing/data': 0.0003484100161585957, 'timing/model': 0.39415810999344103, 'learning_rate/qwen_vl_interface': 6.838931334866267e-06, 'learning_rate/action_model': 6.838931334866266e-05, 'epoch': 0.15}) +[2026-10-03 19:20:00,345][starVLA.training.train_starvla][INFO] - Step 2019, Loss: {'train/loss': 0.21967831254005432, 'train/grad_norm_pre_clip': 28.830760955810547, 'timing/data': 0.00024386000586673617, 'timing/model': 0.395195088989567, 'learning_rate/qwen_vl_interface': 6.836061029076523e-06, 'learning_rate/action_model': 6.836061029076522e-05, 'epoch': 0.15}) +[2026-10-03 19:20:01,250][starVLA.training.train_starvla][INFO] - Step 2020, Loss: {'train/loss': 0.9768253564834595, 'train/grad_norm_pre_clip': 82.05553436279297, 'timing/data': 0.0003399800043553114, 'timing/model': 0.5937707189877983, 'learning_rate/qwen_vl_interface': 6.833190071317e-06, 'learning_rate/action_model': 6.833190071316999e-05, 'epoch': 0.15}) +[2026-10-03 19:20:01,957][starVLA.training.train_starvla][INFO] - Step 2021, Loss: {'train/loss': 0.7085724472999573, 'train/grad_norm_pre_clip': 24.09535789489746, 'timing/data': 0.00038993998896330595, 'timing/model': 0.3966998899995815, 'learning_rate/qwen_vl_interface': 6.830318462767841e-06, 'learning_rate/action_model': 6.830318462767841e-05, 'epoch': 0.15}) +[2026-10-03 19:20:02,662][starVLA.training.train_starvla][INFO] - Step 2022, Loss: {'train/loss': 0.009030426852405071, 'train/grad_norm_pre_clip': 3.7810943126678467, 'timing/data': 0.0003417200059629977, 'timing/model': 0.39337766999960877, 'learning_rate/qwen_vl_interface': 6.827446204609456e-06, 'learning_rate/action_model': 6.827446204609456e-05, 'epoch': 0.15}) +[2026-10-03 19:20:03,369][starVLA.training.train_starvla][INFO] - Step 2023, Loss: {'train/loss': 0.0018119437154382467, 'train/grad_norm_pre_clip': 0.18355520069599152, 'timing/data': 0.0003585900121834129, 'timing/model': 0.394208839017665, 'learning_rate/qwen_vl_interface': 6.824573298022522e-06, 'learning_rate/action_model': 6.824573298022522e-05, 'epoch': 0.15}) +[2026-10-03 19:20:04,076][starVLA.training.train_starvla][INFO] - Step 2024, Loss: {'train/loss': 0.0018471208168193698, 'train/grad_norm_pre_clip': 0.633431077003479, 'timing/data': 0.00022717000683769584, 'timing/model': 0.39549717999761924, 'learning_rate/qwen_vl_interface': 6.8216997441879805e-06, 'learning_rate/action_model': 6.821699744187981e-05, 'epoch': 0.15}) +[2026-10-03 19:20:04,987][starVLA.training.train_starvla][INFO] - Step 2025, Loss: {'train/loss': 0.021527068689465523, 'train/grad_norm_pre_clip': 11.863412857055664, 'timing/data': 0.0002792899904306978, 'timing/model': 0.6001761789957527, 'learning_rate/qwen_vl_interface': 6.818825544287045e-06, 'learning_rate/action_model': 6.818825544287045e-05, 'epoch': 0.15}) +[2026-10-03 19:20:05,692][starVLA.training.train_starvla][INFO] - Step 2026, Loss: {'train/loss': 0.25295311212539673, 'train/grad_norm_pre_clip': 72.6593246459961, 'timing/data': 0.00023367002722807229, 'timing/model': 0.3939055700029712, 'learning_rate/qwen_vl_interface': 6.815950699501186e-06, 'learning_rate/action_model': 6.815950699501186e-05, 'epoch': 0.15}) +[2026-10-03 19:20:06,398][starVLA.training.train_starvla][INFO] - Step 2027, Loss: {'train/loss': 0.01839354634284973, 'train/grad_norm_pre_clip': 6.171688556671143, 'timing/data': 0.000377390009816736, 'timing/model': 0.3952628989936784, 'learning_rate/qwen_vl_interface': 6.813075211012148e-06, 'learning_rate/action_model': 6.813075211012147e-05, 'epoch': 0.15}) +[2026-10-03 19:20:07,106][starVLA.training.train_starvla][INFO] - Step 2028, Loss: {'train/loss': 0.4241098165512085, 'train/grad_norm_pre_clip': 67.67314910888672, 'timing/data': 0.00037152000004425645, 'timing/model': 0.3949901599844452, 'learning_rate/qwen_vl_interface': 6.81019908000193e-06, 'learning_rate/action_model': 6.81019908000193e-05, 'epoch': 0.15}) +[2026-10-03 19:20:07,813][starVLA.training.train_starvla][INFO] - Step 2029, Loss: {'train/loss': 0.01468963734805584, 'train/grad_norm_pre_clip': 7.156789302825928, 'timing/data': 0.0003374399966560304, 'timing/model': 0.3941356400027871, 'learning_rate/qwen_vl_interface': 6.807322307652806e-06, 'learning_rate/action_model': 6.807322307652806e-05, 'epoch': 0.15}) +[2026-10-03 19:20:08,723][starVLA.training.train_starvla][INFO] - Step 2030, Loss: {'train/loss': 0.08173967897891998, 'train/grad_norm_pre_clip': 74.72564697265625, 'timing/data': 0.00028097000904381275, 'timing/model': 0.5979672299872618, 'learning_rate/qwen_vl_interface': 6.804444895147306e-06, 'learning_rate/action_model': 6.804444895147306e-05, 'epoch': 0.15}) +[2026-10-03 19:20:09,428][starVLA.training.train_starvla][INFO] - Step 2031, Loss: {'train/loss': 0.19394560158252716, 'train/grad_norm_pre_clip': 32.70941162109375, 'timing/data': 0.000328589987475425, 'timing/model': 0.3936726899992209, 'learning_rate/qwen_vl_interface': 6.8015668436682255e-06, 'learning_rate/action_model': 6.801566843668226e-05, 'epoch': 0.15}) +[2026-10-03 19:20:10,135][starVLA.training.train_starvla][INFO] - Step 2032, Loss: {'train/loss': 0.25682270526885986, 'train/grad_norm_pre_clip': 43.059165954589844, 'timing/data': 0.0003069700032938272, 'timing/model': 0.39477406098740175, 'learning_rate/qwen_vl_interface': 6.7986881543986245e-06, 'learning_rate/action_model': 6.798688154398624e-05, 'epoch': 0.15}) +[2026-10-03 19:20:10,841][starVLA.training.train_starvla][INFO] - Step 2033, Loss: {'train/loss': 0.3784776031970978, 'train/grad_norm_pre_clip': 41.59736251831055, 'timing/data': 0.00039381999522447586, 'timing/model': 0.3959356300183572, 'learning_rate/qwen_vl_interface': 6.795808828521822e-06, 'learning_rate/action_model': 6.795808828521821e-05, 'epoch': 0.15}) +[2026-10-03 19:20:11,549][starVLA.training.train_starvla][INFO] - Step 2034, Loss: {'train/loss': 0.40815815329551697, 'train/grad_norm_pre_clip': 42.74638366699219, 'timing/data': 0.00023319001775234938, 'timing/model': 0.39608375998795964, 'learning_rate/qwen_vl_interface': 6.792928867221401e-06, 'learning_rate/action_model': 6.792928867221401e-05, 'epoch': 0.15}) +[2026-10-03 19:20:12,254][starVLA.training.train_starvla][INFO] - Step 2035, Loss: {'train/loss': 0.27733102440834045, 'train/grad_norm_pre_clip': 44.170719146728516, 'timing/data': 0.00022020001779310405, 'timing/model': 0.39455758102121763, 'learning_rate/qwen_vl_interface': 6.790048271681202e-06, 'learning_rate/action_model': 6.790048271681202e-05, 'epoch': 0.15}) +[2026-10-03 19:20:13,176][starVLA.training.train_starvla][INFO] - Step 2036, Loss: {'train/loss': 0.020139455795288086, 'train/grad_norm_pre_clip': 20.30918312072754, 'timing/data': 0.0005116800020914525, 'timing/model': 0.39477136998903006, 'learning_rate/qwen_vl_interface': 6.7871670430853335e-06, 'learning_rate/action_model': 6.787167043085333e-05, 'epoch': 0.15}) +[2026-10-03 19:20:13,884][starVLA.training.train_starvla][INFO] - Step 2037, Loss: {'train/loss': 0.006055788137018681, 'train/grad_norm_pre_clip': 6.821628093719482, 'timing/data': 0.0003852299996651709, 'timing/model': 0.39552826998988166, 'learning_rate/qwen_vl_interface': 6.784285182618156e-06, 'learning_rate/action_model': 6.784285182618156e-05, 'epoch': 0.15}) +[2026-10-03 19:20:14,594][starVLA.training.train_starvla][INFO] - Step 2038, Loss: {'train/loss': 0.0017841863445937634, 'train/grad_norm_pre_clip': 0.5752818584442139, 'timing/data': 0.0002649500092957169, 'timing/model': 0.39643684998736717, 'learning_rate/qwen_vl_interface': 6.781402691464297e-06, 'learning_rate/action_model': 6.781402691464297e-05, 'epoch': 0.15}) +[2026-10-03 19:20:15,302][starVLA.training.train_starvla][INFO] - Step 2039, Loss: {'train/loss': 0.21639342606067657, 'train/grad_norm_pre_clip': 59.843467712402344, 'timing/data': 0.00023262001923285425, 'timing/model': 0.39584005100186914, 'learning_rate/qwen_vl_interface': 6.778519570808636e-06, 'learning_rate/action_model': 6.778519570808636e-05, 'epoch': 0.15}) +[2026-10-03 19:20:16,008][starVLA.training.train_starvla][INFO] - Step 2040, Loss: {'train/loss': 0.0004769845400005579, 'train/grad_norm_pre_clip': 0.0935763493180275, 'timing/data': 0.0003592899884097278, 'timing/model': 0.39329372099018656, 'learning_rate/qwen_vl_interface': 6.775635821836317e-06, 'learning_rate/action_model': 6.775635821836317e-05, 'epoch': 0.15}) +[2026-10-03 19:20:16,917][starVLA.training.train_starvla][INFO] - Step 2041, Loss: {'train/loss': 0.012015272863209248, 'train/grad_norm_pre_clip': 10.365117073059082, 'timing/data': 0.00032323997584171593, 'timing/model': 0.5966234689985868, 'learning_rate/qwen_vl_interface': 6.77275144573274e-06, 'learning_rate/action_model': 6.77275144573274e-05, 'epoch': 0.15}) +[2026-10-03 19:20:17,625][starVLA.training.train_starvla][INFO] - Step 2042, Loss: {'train/loss': 0.5849155187606812, 'train/grad_norm_pre_clip': 40.88404846191406, 'timing/data': 0.00023202999727800488, 'timing/model': 0.3958679600036703, 'learning_rate/qwen_vl_interface': 6.769866443683564e-06, 'learning_rate/action_model': 6.769866443683563e-05, 'epoch': 0.15}) +[2026-10-03 19:20:18,332][starVLA.training.train_starvla][INFO] - Step 2043, Loss: {'train/loss': 0.007966307923197746, 'train/grad_norm_pre_clip': 9.227574348449707, 'timing/data': 0.00022740999702364206, 'timing/model': 0.3951553000079002, 'learning_rate/qwen_vl_interface': 6.766980816874701e-06, 'learning_rate/action_model': 6.7669808168747e-05, 'epoch': 0.15}) +[2026-10-03 19:20:19,039][starVLA.training.train_starvla][INFO] - Step 2044, Loss: {'train/loss': 0.3089917302131653, 'train/grad_norm_pre_clip': 69.63117218017578, 'timing/data': 0.00034795000101439655, 'timing/model': 0.3949496600253042, 'learning_rate/qwen_vl_interface': 6.764094566492327e-06, 'learning_rate/action_model': 6.764094566492328e-05, 'epoch': 0.15}) +[2026-10-03 19:20:19,745][starVLA.training.train_starvla][INFO] - Step 2045, Loss: {'train/loss': 0.11130592226982117, 'train/grad_norm_pre_clip': 42.0151252746582, 'timing/data': 0.0003535599971655756, 'timing/model': 0.3934039000014309, 'learning_rate/qwen_vl_interface': 6.7612076937228685e-06, 'learning_rate/action_model': 6.761207693722869e-05, 'epoch': 0.15}) +[2026-10-03 19:20:20,451][starVLA.training.train_starvla][INFO] - Step 2046, Loss: {'train/loss': 0.0015822268323972821, 'train/grad_norm_pre_clip': 0.5016398429870605, 'timing/data': 0.0002251099795103073, 'timing/model': 0.3940950899850577, 'learning_rate/qwen_vl_interface': 6.7583201997530126e-06, 'learning_rate/action_model': 6.758320199753012e-05, 'epoch': 0.15}) +[2026-10-03 19:20:21,359][starVLA.training.train_starvla][INFO] - Step 2047, Loss: {'train/loss': 0.017846640199422836, 'train/grad_norm_pre_clip': 7.941978454589844, 'timing/data': 0.0002291100099682808, 'timing/model': 0.3949198600021191, 'learning_rate/qwen_vl_interface': 6.755432085769693e-06, 'learning_rate/action_model': 6.755432085769694e-05, 'epoch': 0.15}) +[2026-10-03 19:20:22,065][starVLA.training.train_starvla][INFO] - Step 2048, Loss: {'train/loss': 0.2836840748786926, 'train/grad_norm_pre_clip': 54.15821838378906, 'timing/data': 0.0003503499901853502, 'timing/model': 0.3940229889994953, 'learning_rate/qwen_vl_interface': 6.75254335296011e-06, 'learning_rate/action_model': 6.75254335296011e-05, 'epoch': 0.15}) +[2026-10-03 19:20:22,772][starVLA.training.train_starvla][INFO] - Step 2049, Loss: {'train/loss': 0.2662695050239563, 'train/grad_norm_pre_clip': 33.691402435302734, 'timing/data': 0.00025912001729011536, 'timing/model': 0.3957650800002739, 'learning_rate/qwen_vl_interface': 6.749654002511712e-06, 'learning_rate/action_model': 6.749654002511712e-05, 'epoch': 0.15}) +[2026-10-03 19:20:23,481][starVLA.training.train_starvla][INFO] - Step 2050, Loss: {'train/loss': 0.3903970718383789, 'train/grad_norm_pre_clip': 55.06399154663086, 'timing/data': 0.0004489000129979104, 'timing/model': 0.39528639998752624, 'learning_rate/qwen_vl_interface': 6.7467640356122e-06, 'learning_rate/action_model': 6.746764035612199e-05, 'epoch': 0.15}) +[2026-10-03 19:20:24,187][starVLA.training.train_starvla][INFO] - Step 2051, Loss: {'train/loss': 0.0013560921652242541, 'train/grad_norm_pre_clip': 0.4461256265640259, 'timing/data': 0.0004787500074598938, 'timing/model': 0.3937489599920809, 'learning_rate/qwen_vl_interface': 6.74387345344953e-06, 'learning_rate/action_model': 6.74387345344953e-05, 'epoch': 0.15}) +[2026-10-03 19:20:25,096][starVLA.training.train_starvla][INFO] - Step 2052, Loss: {'train/loss': 0.3392255902290344, 'train/grad_norm_pre_clip': 35.319061279296875, 'timing/data': 0.0003126800002064556, 'timing/model': 0.39444669999647886, 'learning_rate/qwen_vl_interface': 6.740982257211911e-06, 'learning_rate/action_model': 6.740982257211911e-05, 'epoch': 0.15}) +[2026-10-03 19:20:25,800][starVLA.training.train_starvla][INFO] - Step 2053, Loss: {'train/loss': 0.24675804376602173, 'train/grad_norm_pre_clip': 131.3734893798828, 'timing/data': 0.0003150200063828379, 'timing/model': 0.39398940099636093, 'learning_rate/qwen_vl_interface': 6.738090448087806e-06, 'learning_rate/action_model': 6.738090448087806e-05, 'epoch': 0.15}) +[2026-10-03 19:20:26,506][starVLA.training.train_starvla][INFO] - Step 2054, Loss: {'train/loss': 0.06291461735963821, 'train/grad_norm_pre_clip': 41.36903762817383, 'timing/data': 0.00028000999009236693, 'timing/model': 0.39373655000235885, 'learning_rate/qwen_vl_interface': 6.735198027265928e-06, 'learning_rate/action_model': 6.735198027265928e-05, 'epoch': 0.15}) +[2026-10-03 19:20:27,213][starVLA.training.train_starvla][INFO] - Step 2055, Loss: {'train/loss': 0.008737311698496342, 'train/grad_norm_pre_clip': 4.554786682128906, 'timing/data': 0.0003106099902652204, 'timing/model': 0.39484002100653015, 'learning_rate/qwen_vl_interface': 6.7323049959352424e-06, 'learning_rate/action_model': 6.732304995935243e-05, 'epoch': 0.15}) +[2026-10-03 19:20:27,921][starVLA.training.train_starvla][INFO] - Step 2056, Loss: {'train/loss': 0.003200998529791832, 'train/grad_norm_pre_clip': 1.2184456586837769, 'timing/data': 0.00030719899223186076, 'timing/model': 0.3952052209933754, 'learning_rate/qwen_vl_interface': 6.729411355284962e-06, 'learning_rate/action_model': 6.729411355284963e-05, 'epoch': 0.15}) +[2026-10-03 19:20:28,831][starVLA.training.train_starvla][INFO] - Step 2057, Loss: {'train/loss': 0.005667274817824364, 'train/grad_norm_pre_clip': 0.977534294128418, 'timing/data': 0.00025063002249225974, 'timing/model': 0.6000837690080516, 'learning_rate/qwen_vl_interface': 6.726517106504559e-06, 'learning_rate/action_model': 6.726517106504558e-05, 'epoch': 0.15}) +[2026-10-03 19:20:29,539][starVLA.training.train_starvla][INFO] - Step 2058, Loss: {'train/loss': 0.1309431493282318, 'train/grad_norm_pre_clip': 34.874969482421875, 'timing/data': 0.00036500897840596735, 'timing/model': 0.3947425399965141, 'learning_rate/qwen_vl_interface': 6.723622250783745e-06, 'learning_rate/action_model': 6.723622250783744e-05, 'epoch': 0.15}) +[2026-10-03 19:20:30,245][starVLA.training.train_starvla][INFO] - Step 2059, Loss: {'train/loss': 0.18199792504310608, 'train/grad_norm_pre_clip': 146.13754272460938, 'timing/data': 0.0002580000145826489, 'timing/model': 0.3956975400215015, 'learning_rate/qwen_vl_interface': 6.720726789312487e-06, 'learning_rate/action_model': 6.720726789312487e-05, 'epoch': 0.15}) +[2026-10-03 19:20:30,953][starVLA.training.train_starvla][INFO] - Step 2060, Loss: {'train/loss': 0.40906310081481934, 'train/grad_norm_pre_clip': 84.37332916259766, 'timing/data': 0.00032655001268722117, 'timing/model': 0.39530115999514237, 'learning_rate/qwen_vl_interface': 6.7178307232810015e-06, 'learning_rate/action_model': 6.717830723281e-05, 'epoch': 0.15}) +[2026-10-03 19:20:31,657][starVLA.training.train_starvla][INFO] - Step 2061, Loss: {'train/loss': 0.002883772598579526, 'train/grad_norm_pre_clip': 1.1195216178894043, 'timing/data': 0.0003372899955138564, 'timing/model': 0.3940803089935798, 'learning_rate/qwen_vl_interface': 6.714934053879749e-06, 'learning_rate/action_model': 6.71493405387975e-05, 'epoch': 0.15}) +[2026-10-03 19:20:32,364][starVLA.training.train_starvla][INFO] - Step 2062, Loss: {'train/loss': 0.0030878859106451273, 'train/grad_norm_pre_clip': 0.6894927024841309, 'timing/data': 0.0003245500265620649, 'timing/model': 0.3938957900099922, 'learning_rate/qwen_vl_interface': 6.712036782299442e-06, 'learning_rate/action_model': 6.712036782299443e-05, 'epoch': 0.15}) +[2026-10-03 19:20:33,273][starVLA.training.train_starvla][INFO] - Step 2063, Loss: {'train/loss': 0.08508815616369247, 'train/grad_norm_pre_clip': 123.18694305419922, 'timing/data': 0.000255089980782941, 'timing/model': 0.39351547000114806, 'learning_rate/qwen_vl_interface': 6.709138909731041e-06, 'learning_rate/action_model': 6.709138909731041e-05, 'epoch': 0.15}) +[2026-10-03 19:20:33,979][starVLA.training.train_starvla][INFO] - Step 2064, Loss: {'train/loss': 0.011460868641734123, 'train/grad_norm_pre_clip': 5.141106605529785, 'timing/data': 0.00035632000071927905, 'timing/model': 0.39333111999440007, 'learning_rate/qwen_vl_interface': 6.706240437365751e-06, 'learning_rate/action_model': 6.70624043736575e-05, 'epoch': 0.15}) +[2026-10-03 19:20:34,687][starVLA.training.train_starvla][INFO] - Step 2065, Loss: {'train/loss': 0.002269295509904623, 'train/grad_norm_pre_clip': 1.6170337200164795, 'timing/data': 0.00026308000087738037, 'timing/model': 0.39573163900058717, 'learning_rate/qwen_vl_interface': 6.703341366395023e-06, 'learning_rate/action_model': 6.703341366395023e-05, 'epoch': 0.15}) +[2026-10-03 19:20:35,393][starVLA.training.train_starvla][INFO] - Step 2066, Loss: {'train/loss': 0.017912331968545914, 'train/grad_norm_pre_clip': 9.77602481842041, 'timing/data': 0.00037260999670252204, 'timing/model': 0.3946250599983614, 'learning_rate/qwen_vl_interface': 6.700441698010558e-06, 'learning_rate/action_model': 6.700441698010558e-05, 'epoch': 0.15}) +[2026-10-03 19:20:36,103][starVLA.training.train_starvla][INFO] - Step 2067, Loss: {'train/loss': 0.17680302262306213, 'train/grad_norm_pre_clip': 87.86128997802734, 'timing/data': 0.0002708300016820431, 'timing/model': 0.3957760799967218, 'learning_rate/qwen_vl_interface': 6.697541433404299e-06, 'learning_rate/action_model': 6.697541433404299e-05, 'epoch': 0.15}) +[2026-10-03 19:20:37,013][starVLA.training.train_starvla][INFO] - Step 2068, Loss: {'train/loss': 0.008635892532765865, 'train/grad_norm_pre_clip': 11.655402183532715, 'timing/data': 0.00039525001193396747, 'timing/model': 0.39468928999849595, 'learning_rate/qwen_vl_interface': 6.6946405737684325e-06, 'learning_rate/action_model': 6.694640573768433e-05, 'epoch': 0.15}) +[2026-10-03 19:20:37,721][starVLA.training.train_starvla][INFO] - Step 2069, Loss: {'train/loss': 0.0002702846541069448, 'train/grad_norm_pre_clip': 0.049977388232946396, 'timing/data': 0.00023611000506207347, 'timing/model': 0.3950927400146611, 'learning_rate/qwen_vl_interface': 6.691739120295397e-06, 'learning_rate/action_model': 6.691739120295397e-05, 'epoch': 0.15}) +[2026-10-03 19:20:38,427][starVLA.training.train_starvla][INFO] - Step 2070, Loss: {'train/loss': 0.00011456424545031041, 'train/grad_norm_pre_clip': 0.016535842791199684, 'timing/data': 0.00022000999888405204, 'timing/model': 0.39535780998994596, 'learning_rate/qwen_vl_interface': 6.688837074177868e-06, 'learning_rate/action_model': 6.688837074177868e-05, 'epoch': 0.15}) +[2026-10-03 19:20:39,135][starVLA.training.train_starvla][INFO] - Step 2071, Loss: {'train/loss': 0.00012835237430408597, 'train/grad_norm_pre_clip': 0.020481405779719353, 'timing/data': 0.0003096100117545575, 'timing/model': 0.39508096902864054, 'learning_rate/qwen_vl_interface': 6.68593443660877e-06, 'learning_rate/action_model': 6.685934436608769e-05, 'epoch': 0.15}) +[2026-10-03 19:20:39,844][starVLA.training.train_starvla][INFO] - Step 2072, Loss: {'train/loss': 1.0265529155731201, 'train/grad_norm_pre_clip': 67.58899688720703, 'timing/data': 0.000341220002155751, 'timing/model': 0.39596701998380013, 'learning_rate/qwen_vl_interface': 6.6830312087812625e-06, 'learning_rate/action_model': 6.683031208781262e-05, 'epoch': 0.15}) +[2026-10-03 19:20:40,754][starVLA.training.train_starvla][INFO] - Step 2073, Loss: {'train/loss': 0.65807044506073, 'train/grad_norm_pre_clip': 25.116924285888672, 'timing/data': 0.00031885001226328313, 'timing/model': 0.3957594399980735, 'learning_rate/qwen_vl_interface': 6.680127391888754e-06, 'learning_rate/action_model': 6.680127391888754e-05, 'epoch': 0.15}) +[2026-10-03 19:20:41,460][starVLA.training.train_starvla][INFO] - Step 2074, Loss: {'train/loss': 0.40332216024398804, 'train/grad_norm_pre_clip': 60.949684143066406, 'timing/data': 0.0002735900052357465, 'timing/model': 0.3953517010086216, 'learning_rate/qwen_vl_interface': 6.677222987124897e-06, 'learning_rate/action_model': 6.677222987124897e-05, 'epoch': 0.15}) +[2026-10-03 19:20:42,165][starVLA.training.train_starvla][INFO] - Step 2075, Loss: {'train/loss': 0.2665204703807831, 'train/grad_norm_pre_clip': 82.78956604003906, 'timing/data': 0.00033213000278919935, 'timing/model': 0.39359267000691034, 'learning_rate/qwen_vl_interface': 6.67431799568358e-06, 'learning_rate/action_model': 6.67431799568358e-05, 'epoch': 0.15}) +[2026-10-03 19:20:42,873][starVLA.training.train_starvla][INFO] - Step 2076, Loss: {'train/loss': 0.003749894443899393, 'train/grad_norm_pre_clip': 6.153796195983887, 'timing/data': 0.0002730900014284998, 'timing/model': 0.3952890409855172, 'learning_rate/qwen_vl_interface': 6.671412418758939e-06, 'learning_rate/action_model': 6.671412418758938e-05, 'epoch': 0.15}) +[2026-10-03 19:20:43,580][starVLA.training.train_starvla][INFO] - Step 2077, Loss: {'train/loss': 0.004385038744658232, 'train/grad_norm_pre_clip': 2.1992406845092773, 'timing/data': 0.00027905000024475157, 'timing/model': 0.3955787699960638, 'learning_rate/qwen_vl_interface': 6.668506257545341e-06, 'learning_rate/action_model': 6.668506257545341e-05, 'epoch': 0.15}) +[2026-10-03 19:20:44,487][starVLA.training.train_starvla][INFO] - Step 2078, Loss: {'train/loss': 0.00018200560589320958, 'train/grad_norm_pre_clip': 0.030324582010507584, 'timing/data': 0.00035072999889962375, 'timing/model': 0.5946028499747626, 'learning_rate/qwen_vl_interface': 6.665599513237403e-06, 'learning_rate/action_model': 6.665599513237402e-05, 'epoch': 0.15}) +[2026-10-03 19:20:45,192][starVLA.training.train_starvla][INFO] - Step 2079, Loss: {'train/loss': 0.0008287348318845034, 'train/grad_norm_pre_clip': 0.26871877908706665, 'timing/data': 0.0002413300098851323, 'timing/model': 0.39571427000919357, 'learning_rate/qwen_vl_interface': 6.662692187029978e-06, 'learning_rate/action_model': 6.662692187029978e-05, 'epoch': 0.15}) +[2026-10-03 19:20:45,898][starVLA.training.train_starvla][INFO] - Step 2080, Loss: {'train/loss': 0.009355887770652771, 'train/grad_norm_pre_clip': 20.242767333984375, 'timing/data': 0.00022245000582188368, 'timing/model': 0.3944056989857927, 'learning_rate/qwen_vl_interface': 6.659784280118156e-06, 'learning_rate/action_model': 6.659784280118156e-05, 'epoch': 0.15}) +[2026-10-03 19:20:46,604][starVLA.training.train_starvla][INFO] - Step 2081, Loss: {'train/loss': 0.05870332568883896, 'train/grad_norm_pre_clip': 62.86627960205078, 'timing/data': 0.000327579997247085, 'timing/model': 0.3939930299820844, 'learning_rate/qwen_vl_interface': 6.65687579369727e-06, 'learning_rate/action_model': 6.656875793697271e-05, 'epoch': 0.15}) +[2026-10-03 19:20:47,312][starVLA.training.train_starvla][INFO] - Step 2082, Loss: {'train/loss': 0.23824536800384521, 'train/grad_norm_pre_clip': 45.25221252441406, 'timing/data': 0.0004295299877412617, 'timing/model': 0.3957357899926137, 'learning_rate/qwen_vl_interface': 6.653966728962885e-06, 'learning_rate/action_model': 6.653966728962885e-05, 'epoch': 0.15}) +[2026-10-03 19:20:48,228][starVLA.training.train_starvla][INFO] - Step 2083, Loss: {'train/loss': 0.4051215648651123, 'train/grad_norm_pre_clip': 94.3115234375, 'timing/data': 0.00034065000363625586, 'timing/model': 0.6020183189830277, 'learning_rate/qwen_vl_interface': 6.651057087110813e-06, 'learning_rate/action_model': 6.651057087110813e-05, 'epoch': 0.15}) +[2026-10-03 19:20:48,934][starVLA.training.train_starvla][INFO] - Step 2084, Loss: {'train/loss': 0.37224119901657104, 'train/grad_norm_pre_clip': 15.072258949279785, 'timing/data': 0.00030004000291228294, 'timing/model': 0.3935325000202283, 'learning_rate/qwen_vl_interface': 6.6481468693370924e-06, 'learning_rate/action_model': 6.648146869337091e-05, 'epoch': 0.15}) +[2026-10-03 19:20:49,640][starVLA.training.train_starvla][INFO] - Step 2085, Loss: {'train/loss': 0.2963372766971588, 'train/grad_norm_pre_clip': 39.40303421020508, 'timing/data': 0.0003241199883632362, 'timing/model': 0.3957912500191014, 'learning_rate/qwen_vl_interface': 6.645236076838009e-06, 'learning_rate/action_model': 6.645236076838009e-05, 'epoch': 0.15}) +[2026-10-03 19:20:50,347][starVLA.training.train_starvla][INFO] - Step 2086, Loss: {'train/loss': 0.3926119804382324, 'train/grad_norm_pre_clip': 35.32530975341797, 'timing/data': 0.00028998000198043883, 'timing/model': 0.3953898790059611, 'learning_rate/qwen_vl_interface': 6.642324710810075e-06, 'learning_rate/action_model': 6.642324710810076e-05, 'epoch': 0.15}) +[2026-10-03 19:20:51,053][starVLA.training.train_starvla][INFO] - Step 2087, Loss: {'train/loss': 0.22819405794143677, 'train/grad_norm_pre_clip': 65.7405776977539, 'timing/data': 0.00035277000279165804, 'timing/model': 0.3939512299839407, 'learning_rate/qwen_vl_interface': 6.639412772450047e-06, 'learning_rate/action_model': 6.639412772450047e-05, 'epoch': 0.15}) +[2026-10-03 19:20:51,759][starVLA.training.train_starvla][INFO] - Step 2088, Loss: {'train/loss': 0.28480932116508484, 'train/grad_norm_pre_clip': 32.39543151855469, 'timing/data': 0.0003291900211479515, 'timing/model': 0.39340709999669343, 'learning_rate/qwen_vl_interface': 6.6365002629549104e-06, 'learning_rate/action_model': 6.63650026295491e-05, 'epoch': 0.15}) +[2026-10-03 19:20:52,674][starVLA.training.train_starvla][INFO] - Step 2089, Loss: {'train/loss': 0.2632191777229309, 'train/grad_norm_pre_clip': 38.986961364746094, 'timing/data': 0.0003692600002977997, 'timing/model': 0.3934025199851021, 'learning_rate/qwen_vl_interface': 6.633587183521887e-06, 'learning_rate/action_model': 6.633587183521887e-05, 'epoch': 0.15}) +[2026-10-03 19:20:53,381][starVLA.training.train_starvla][INFO] - Step 2090, Loss: {'train/loss': 0.2526487410068512, 'train/grad_norm_pre_clip': 15.913872718811035, 'timing/data': 0.00021883001318201423, 'timing/model': 0.39518582000164315, 'learning_rate/qwen_vl_interface': 6.6306735353484365e-06, 'learning_rate/action_model': 6.630673535348436e-05, 'epoch': 0.15}) +[2026-10-03 19:20:54,061][starVLA.training.train_starvla][INFO] - Step 2091, Loss: {'train/loss': 0.00785374827682972, 'train/grad_norm_pre_clip': 6.479789733886719, 'timing/data': 0.00030051000067032874, 'timing/model': 0.36884683300741017, 'learning_rate/qwen_vl_interface': 6.627759319632246e-06, 'learning_rate/action_model': 6.627759319632247e-05, 'epoch': 0.15}) +[2026-10-03 19:20:54,765][starVLA.training.train_starvla][INFO] - Step 2092, Loss: {'train/loss': 0.2918049395084381, 'train/grad_norm_pre_clip': 79.0373306274414, 'timing/data': 0.00022258999524638057, 'timing/model': 0.3935200599953532, 'learning_rate/qwen_vl_interface': 6.624844537571245e-06, 'learning_rate/action_model': 6.624844537571245e-05, 'epoch': 0.15}) +[2026-10-03 19:20:55,470][starVLA.training.train_starvla][INFO] - Step 2093, Loss: {'train/loss': 0.28355222940444946, 'train/grad_norm_pre_clip': 25.707534790039062, 'timing/data': 0.00027694000164046884, 'timing/model': 0.3951428289874457, 'learning_rate/qwen_vl_interface': 6.621929190363585e-06, 'learning_rate/action_model': 6.621929190363585e-05, 'epoch': 0.15}) +[2026-10-03 19:20:56,377][starVLA.training.train_starvla][INFO] - Step 2094, Loss: {'train/loss': 0.36676517128944397, 'train/grad_norm_pre_clip': 69.56169891357422, 'timing/data': 0.0003773600037675351, 'timing/model': 0.5949192890257109, 'learning_rate/qwen_vl_interface': 6.619013279207657e-06, 'learning_rate/action_model': 6.619013279207656e-05, 'epoch': 0.15}) +[2026-10-03 19:20:57,083][starVLA.training.train_starvla][INFO] - Step 2095, Loss: {'train/loss': 0.0014717010781168938, 'train/grad_norm_pre_clip': 0.43619802594184875, 'timing/data': 0.0004206400190014392, 'timing/model': 0.3936390109884087, 'learning_rate/qwen_vl_interface': 6.616096805302083e-06, 'learning_rate/action_model': 6.616096805302082e-05, 'epoch': 0.15}) +[2026-10-03 19:20:57,790][starVLA.training.train_starvla][INFO] - Step 2096, Loss: {'train/loss': 0.344511479139328, 'train/grad_norm_pre_clip': 25.954322814941406, 'timing/data': 0.0003408399934414774, 'timing/model': 0.39507369001512416, 'learning_rate/qwen_vl_interface': 6.6131797698457135e-06, 'learning_rate/action_model': 6.613179769845713e-05, 'epoch': 0.15}) +[2026-10-03 19:20:58,498][starVLA.training.train_starvla][INFO] - Step 2097, Loss: {'train/loss': 0.014445550739765167, 'train/grad_norm_pre_clip': 20.45345115661621, 'timing/data': 0.0003474300028756261, 'timing/model': 0.39530066002043895, 'learning_rate/qwen_vl_interface': 6.610262174037633e-06, 'learning_rate/action_model': 6.610262174037633e-05, 'epoch': 0.15}) +[2026-10-03 19:20:59,205][starVLA.training.train_starvla][INFO] - Step 2098, Loss: {'train/loss': 0.0051278844475746155, 'train/grad_norm_pre_clip': 1.402547001838684, 'timing/data': 0.00036724001984111965, 'timing/model': 0.3949460510048084, 'learning_rate/qwen_vl_interface': 6.607344019077155e-06, 'learning_rate/action_model': 6.607344019077156e-05, 'epoch': 0.15}) +[2026-10-03 19:20:59,912][starVLA.training.train_starvla][INFO] - Step 2099, Loss: {'train/loss': 0.002600878942757845, 'train/grad_norm_pre_clip': 1.1210880279541016, 'timing/data': 0.0003530799876898527, 'timing/model': 0.39607180998427793, 'learning_rate/qwen_vl_interface': 6.604425306163822e-06, 'learning_rate/action_model': 6.604425306163822e-05, 'epoch': 0.15}) +[2026-10-03 19:21:00,825][starVLA.training.train_starvla][INFO] - Step 2100, Loss: {'train/loss': 0.32462403178215027, 'train/grad_norm_pre_clip': 46.51613235473633, 'timing/data': 0.000292269978672266, 'timing/model': 0.3948247809894383, 'learning_rate/qwen_vl_interface': 6.6015060364974045e-06, 'learning_rate/action_model': 6.601506036497404e-05, 'epoch': 0.15}) +[2026-10-03 19:21:01,533][starVLA.training.train_starvla][INFO] - Step 2101, Loss: {'train/loss': 0.0017028020229190588, 'train/grad_norm_pre_clip': 0.450168639421463, 'timing/data': 0.0002789100108202547, 'timing/model': 0.39529443098581396, 'learning_rate/qwen_vl_interface': 6.598586211277907e-06, 'learning_rate/action_model': 6.598586211277907e-05, 'epoch': 0.15}) +[2026-10-03 19:21:02,240][starVLA.training.train_starvla][INFO] - Step 2102, Loss: {'train/loss': 0.6412203311920166, 'train/grad_norm_pre_clip': 118.0496826171875, 'timing/data': 0.00022062999778427184, 'timing/model': 0.3956270499911625, 'learning_rate/qwen_vl_interface': 6.59566583170556e-06, 'learning_rate/action_model': 6.59566583170556e-05, 'epoch': 0.15}) +[2026-10-03 19:21:02,943][starVLA.training.train_starvla][INFO] - Step 2103, Loss: {'train/loss': 0.01320634875446558, 'train/grad_norm_pre_clip': 5.947711944580078, 'timing/data': 0.00022504999651573598, 'timing/model': 0.39316768100252375, 'learning_rate/qwen_vl_interface': 6.59274489898082e-06, 'learning_rate/action_model': 6.592744898980819e-05, 'epoch': 0.15}) +[2026-10-03 19:21:03,649][starVLA.training.train_starvla][INFO] - Step 2104, Loss: {'train/loss': 0.12797033786773682, 'train/grad_norm_pre_clip': 76.38188171386719, 'timing/data': 0.00033429000177420676, 'timing/model': 0.394259749999037, 'learning_rate/qwen_vl_interface': 6.5898234143043716e-06, 'learning_rate/action_model': 6.589823414304371e-05, 'epoch': 0.15}) +[2026-10-03 19:21:04,553][starVLA.training.train_starvla][INFO] - Step 2105, Loss: {'train/loss': 0.3759949505329132, 'train/grad_norm_pre_clip': 111.5682601928711, 'timing/data': 0.00033655998413451016, 'timing/model': 0.393452949996572, 'learning_rate/qwen_vl_interface': 6.5869013788771264e-06, 'learning_rate/action_model': 6.586901378877126e-05, 'epoch': 0.15}) +[2026-10-03 19:21:05,259][starVLA.training.train_starvla][INFO] - Step 2106, Loss: {'train/loss': 0.3537415564060211, 'train/grad_norm_pre_clip': 36.30770492553711, 'timing/data': 0.0003008699859492481, 'timing/model': 0.39438509999308735, 'learning_rate/qwen_vl_interface': 6.583978793900223e-06, 'learning_rate/action_model': 6.583978793900223e-05, 'epoch': 0.15}) +[2026-10-03 19:21:05,965][starVLA.training.train_starvla][INFO] - Step 2107, Loss: {'train/loss': 0.25251421332359314, 'train/grad_norm_pre_clip': 36.20164489746094, 'timing/data': 0.0002479999966453761, 'timing/model': 0.39335108001250774, 'learning_rate/qwen_vl_interface': 6.581055660575028e-06, 'learning_rate/action_model': 6.581055660575027e-05, 'epoch': 0.15}) +[2026-10-03 19:21:06,668][starVLA.training.train_starvla][INFO] - Step 2108, Loss: {'train/loss': 0.2731321156024933, 'train/grad_norm_pre_clip': 139.24749755859375, 'timing/data': 0.000354000017978251, 'timing/model': 0.3922568300040439, 'learning_rate/qwen_vl_interface': 6.578131980103128e-06, 'learning_rate/action_model': 6.578131980103128e-05, 'epoch': 0.15}) +[2026-10-03 19:21:07,372][starVLA.training.train_starvla][INFO] - Step 2109, Loss: {'train/loss': 0.8809519410133362, 'train/grad_norm_pre_clip': 86.82816314697266, 'timing/data': 0.00022237998200580478, 'timing/model': 0.39515777898486704, 'learning_rate/qwen_vl_interface': 6.5752077536863404e-06, 'learning_rate/action_model': 6.57520775368634e-05, 'epoch': 0.15}) +[2026-10-03 19:21:08,279][starVLA.training.train_starvla][INFO] - Step 2110, Loss: {'train/loss': 0.0006666806293651462, 'train/grad_norm_pre_clip': 0.17194680869579315, 'timing/data': 0.00021168001694604754, 'timing/model': 0.5954909099964425, 'learning_rate/qwen_vl_interface': 6.572282982526702e-06, 'learning_rate/action_model': 6.572282982526702e-05, 'epoch': 0.15}) +[2026-10-03 19:21:08,981][starVLA.training.train_starvla][INFO] - Step 2111, Loss: {'train/loss': 0.0038182688876986504, 'train/grad_norm_pre_clip': 1.159079670906067, 'timing/data': 0.00030555998091585934, 'timing/model': 0.3924710999999661, 'learning_rate/qwen_vl_interface': 6.5693576678264755e-06, 'learning_rate/action_model': 6.569357667826476e-05, 'epoch': 0.15}) +[2026-10-03 19:21:09,687][starVLA.training.train_starvla][INFO] - Step 2112, Loss: {'train/loss': 0.0014948963653296232, 'train/grad_norm_pre_clip': 0.33047720789909363, 'timing/data': 0.0003266200073994696, 'timing/model': 0.39434425000217743, 'learning_rate/qwen_vl_interface': 6.566431810788148e-06, 'learning_rate/action_model': 6.566431810788148e-05, 'epoch': 0.15}) +[2026-10-03 19:21:10,392][starVLA.training.train_starvla][INFO] - Step 2113, Loss: {'train/loss': 0.028688548132777214, 'train/grad_norm_pre_clip': 56.93221664428711, 'timing/data': 0.0002540500136092305, 'timing/model': 0.39586147901718505, 'learning_rate/qwen_vl_interface': 6.5635054126144306e-06, 'learning_rate/action_model': 6.56350541261443e-05, 'epoch': 0.15}) +[2026-10-03 19:21:11,098][starVLA.training.train_starvla][INFO] - Step 2114, Loss: {'train/loss': 0.009896809235215187, 'train/grad_norm_pre_clip': 3.2415032386779785, 'timing/data': 0.000348169996868819, 'timing/model': 0.39437752999947406, 'learning_rate/qwen_vl_interface': 6.560578474508254e-06, 'learning_rate/action_model': 6.560578474508253e-05, 'epoch': 0.15}) +[2026-10-03 19:21:12,009][starVLA.training.train_starvla][INFO] - Step 2115, Loss: {'train/loss': 0.5346501469612122, 'train/grad_norm_pre_clip': 119.66497039794922, 'timing/data': 0.0002539499837439507, 'timing/model': 0.5984323099837638, 'learning_rate/qwen_vl_interface': 6.5576509976727685e-06, 'learning_rate/action_model': 6.557650997672769e-05, 'epoch': 0.15}) +[2026-10-03 19:21:12,716][starVLA.training.train_starvla][INFO] - Step 2116, Loss: {'train/loss': 0.10451410710811615, 'train/grad_norm_pre_clip': 38.1619987487793, 'timing/data': 0.00022965000243857503, 'timing/model': 0.3946928499790374, 'learning_rate/qwen_vl_interface': 6.554722983311354e-06, 'learning_rate/action_model': 6.554722983311355e-05, 'epoch': 0.15}) +[2026-10-03 19:21:13,422][starVLA.training.train_starvla][INFO] - Step 2117, Loss: {'train/loss': 0.06485780328512192, 'train/grad_norm_pre_clip': 110.31305694580078, 'timing/data': 0.0002633300027810037, 'timing/model': 0.39484985900344327, 'learning_rate/qwen_vl_interface': 6.551794432627605e-06, 'learning_rate/action_model': 6.551794432627605e-05, 'epoch': 0.15}) +[2026-10-03 19:21:14,124][starVLA.training.train_starvla][INFO] - Step 2118, Loss: {'train/loss': 0.0127848656848073, 'train/grad_norm_pre_clip': 15.678678512573242, 'timing/data': 0.00034228997537866235, 'timing/model': 0.3930146400234662, 'learning_rate/qwen_vl_interface': 6.548865346825338e-06, 'learning_rate/action_model': 6.548865346825338e-05, 'epoch': 0.16}) +[2026-10-03 19:21:14,830][starVLA.training.train_starvla][INFO] - Step 2119, Loss: {'train/loss': 0.08834896236658096, 'train/grad_norm_pre_clip': 116.4772720336914, 'timing/data': 0.0002393100003246218, 'timing/model': 0.3942950899945572, 'learning_rate/qwen_vl_interface': 6.545935727108589e-06, 'learning_rate/action_model': 6.545935727108588e-05, 'epoch': 0.16}) +[2026-10-03 19:21:15,537][starVLA.training.train_starvla][INFO] - Step 2120, Loss: {'train/loss': 0.002052213763818145, 'train/grad_norm_pre_clip': 1.4462798833847046, 'timing/data': 0.0003320099785923958, 'timing/model': 0.3950615799985826, 'learning_rate/qwen_vl_interface': 6.543005574681613e-06, 'learning_rate/action_model': 6.543005574681613e-05, 'epoch': 0.16}) +[2026-10-03 19:21:16,442][starVLA.training.train_starvla][INFO] - Step 2121, Loss: {'train/loss': 0.007752592209726572, 'train/grad_norm_pre_clip': 2.2901790142059326, 'timing/data': 0.0003586399834603071, 'timing/model': 0.3949215899920091, 'learning_rate/qwen_vl_interface': 6.5400748907488886e-06, 'learning_rate/action_model': 6.540074890748888e-05, 'epoch': 0.16}) +[2026-10-03 19:21:17,148][starVLA.training.train_starvla][INFO] - Step 2122, Loss: {'train/loss': 0.006612330675125122, 'train/grad_norm_pre_clip': 2.0731754302978516, 'timing/data': 0.0002302499779034406, 'timing/model': 0.39464201097143814, 'learning_rate/qwen_vl_interface': 6.537143676515105e-06, 'learning_rate/action_model': 6.537143676515105e-05, 'epoch': 0.16}) +[2026-10-03 19:21:17,854][starVLA.training.train_starvla][INFO] - Step 2123, Loss: {'train/loss': 0.45061904191970825, 'train/grad_norm_pre_clip': 118.54535675048828, 'timing/data': 0.00033528998028486967, 'timing/model': 0.3942379599902779, 'learning_rate/qwen_vl_interface': 6.534211933185176e-06, 'learning_rate/action_model': 6.534211933185176e-05, 'epoch': 0.16}) +[2026-10-03 19:21:18,558][starVLA.training.train_starvla][INFO] - Step 2124, Loss: {'train/loss': 0.3493373990058899, 'train/grad_norm_pre_clip': 49.031646728515625, 'timing/data': 0.00032729998929426074, 'timing/model': 0.39445329099544324, 'learning_rate/qwen_vl_interface': 6.531279661964231e-06, 'learning_rate/action_model': 6.531279661964231e-05, 'epoch': 0.16}) +[2026-10-03 19:21:19,264][starVLA.training.train_starvla][INFO] - Step 2125, Loss: {'train/loss': 0.007402552291750908, 'train/grad_norm_pre_clip': 4.119410991668701, 'timing/data': 0.0003306700091343373, 'timing/model': 0.3950349799997639, 'learning_rate/qwen_vl_interface': 6.528346864057616e-06, 'learning_rate/action_model': 6.528346864057615e-05, 'epoch': 0.16}) +[2026-10-03 19:21:20,170][starVLA.training.train_starvla][INFO] - Step 2126, Loss: {'train/loss': 0.005766131449490786, 'train/grad_norm_pre_clip': 4.002026081085205, 'timing/data': 0.00035668001510202885, 'timing/model': 0.5955636700091418, 'learning_rate/qwen_vl_interface': 6.5254135406708906e-06, 'learning_rate/action_model': 6.52541354067089e-05, 'epoch': 0.16}) +[2026-10-03 19:21:20,875][starVLA.training.train_starvla][INFO] - Step 2127, Loss: {'train/loss': 0.0016044669318944216, 'train/grad_norm_pre_clip': 1.2733097076416016, 'timing/data': 0.000332180003169924, 'timing/model': 0.39267674001166597, 'learning_rate/qwen_vl_interface': 6.522479693009836e-06, 'learning_rate/action_model': 6.522479693009837e-05, 'epoch': 0.16}) +[2026-10-03 19:21:21,579][starVLA.training.train_starvla][INFO] - Step 2128, Loss: {'train/loss': 0.0002461746335029602, 'train/grad_norm_pre_clip': 0.03735562413930893, 'timing/data': 0.00029364999500103295, 'timing/model': 0.3942961200082209, 'learning_rate/qwen_vl_interface': 6.519545322280446e-06, 'learning_rate/action_model': 6.519545322280446e-05, 'epoch': 0.16}) +[2026-10-03 19:21:22,285][starVLA.training.train_starvla][INFO] - Step 2129, Loss: {'train/loss': 0.0017282336484640837, 'train/grad_norm_pre_clip': 1.075161337852478, 'timing/data': 0.00032631997601129115, 'timing/model': 0.3936964000167791, 'learning_rate/qwen_vl_interface': 6.5166104296889275e-06, 'learning_rate/action_model': 6.516610429688927e-05, 'epoch': 0.16}) +[2026-10-03 19:21:22,991][starVLA.training.train_starvla][INFO] - Step 2130, Loss: {'train/loss': 0.34231632947921753, 'train/grad_norm_pre_clip': 48.16495895385742, 'timing/data': 0.0003281600074842572, 'timing/model': 0.3947039499762468, 'learning_rate/qwen_vl_interface': 6.513675016441707e-06, 'learning_rate/action_model': 6.513675016441707e-05, 'epoch': 0.16}) +[2026-10-03 19:21:23,899][starVLA.training.train_starvla][INFO] - Step 2131, Loss: {'train/loss': 0.1331864595413208, 'train/grad_norm_pre_clip': 78.33809661865234, 'timing/data': 0.00032631002250127494, 'timing/model': 0.5963247189938556, 'learning_rate/qwen_vl_interface': 6.510739083745422e-06, 'learning_rate/action_model': 6.510739083745422e-05, 'epoch': 0.16}) +[2026-10-03 19:21:24,605][starVLA.training.train_starvla][INFO] - Step 2132, Loss: {'train/loss': 0.4377526342868805, 'train/grad_norm_pre_clip': 41.24173355102539, 'timing/data': 0.0003798999823629856, 'timing/model': 0.3956418000161648, 'learning_rate/qwen_vl_interface': 6.507802632806919e-06, 'learning_rate/action_model': 6.507802632806919e-05, 'epoch': 0.16}) +[2026-10-03 19:21:25,311][starVLA.training.train_starvla][INFO] - Step 2133, Loss: {'train/loss': 0.0034339390695095062, 'train/grad_norm_pre_clip': 1.2611421346664429, 'timing/data': 0.0003011499939020723, 'timing/model': 0.395429049996892, 'learning_rate/qwen_vl_interface': 6.5048656648332646e-06, 'learning_rate/action_model': 6.504865664833264e-05, 'epoch': 0.16}) +[2026-10-03 19:21:26,018][starVLA.training.train_starvla][INFO] - Step 2134, Loss: {'train/loss': 0.5571991205215454, 'train/grad_norm_pre_clip': 48.59706497192383, 'timing/data': 0.0004136200004722923, 'timing/model': 0.3949746199941728, 'learning_rate/qwen_vl_interface': 6.501928181031737e-06, 'learning_rate/action_model': 6.501928181031737e-05, 'epoch': 0.16}) +[2026-10-03 19:21:26,725][starVLA.training.train_starvla][INFO] - Step 2135, Loss: {'train/loss': 0.002769051119685173, 'train/grad_norm_pre_clip': 2.881652355194092, 'timing/data': 0.00026331000844947994, 'timing/model': 0.3949972999980673, 'learning_rate/qwen_vl_interface': 6.498990182609823e-06, 'learning_rate/action_model': 6.498990182609823e-05, 'epoch': 0.16}) +[2026-10-03 19:21:27,429][starVLA.training.train_starvla][INFO] - Step 2136, Loss: {'train/loss': 0.03358469903469086, 'train/grad_norm_pre_clip': 26.81893539428711, 'timing/data': 0.00031001001480035484, 'timing/model': 0.3933206500078086, 'learning_rate/qwen_vl_interface': 6.4960516707752255e-06, 'learning_rate/action_model': 6.496051670775226e-05, 'epoch': 0.16}) +[2026-10-03 19:21:28,335][starVLA.training.train_starvla][INFO] - Step 2137, Loss: {'train/loss': 0.1080145388841629, 'train/grad_norm_pre_clip': 160.61508178710938, 'timing/data': 0.00029302999610081315, 'timing/model': 0.3948685089999344, 'learning_rate/qwen_vl_interface': 6.493112646735851e-06, 'learning_rate/action_model': 6.493112646735851e-05, 'epoch': 0.16}) +[2026-10-03 19:21:29,040][starVLA.training.train_starvla][INFO] - Step 2138, Loss: {'train/loss': 0.03879128023982048, 'train/grad_norm_pre_clip': 36.410579681396484, 'timing/data': 0.00028311999631114304, 'timing/model': 0.392998510011239, 'learning_rate/qwen_vl_interface': 6.490173111699824e-06, 'learning_rate/action_model': 6.490173111699825e-05, 'epoch': 0.16}) +[2026-10-03 19:21:29,744][starVLA.training.train_starvla][INFO] - Step 2139, Loss: {'train/loss': 0.0006381103303283453, 'train/grad_norm_pre_clip': 0.15807294845581055, 'timing/data': 0.00036589999217540026, 'timing/model': 0.3944567299913615, 'learning_rate/qwen_vl_interface': 6.487233066875479e-06, 'learning_rate/action_model': 6.487233066875479e-05, 'epoch': 0.16}) +[2026-10-03 19:21:30,450][starVLA.training.train_starvla][INFO] - Step 2140, Loss: {'train/loss': 0.0002960474230349064, 'train/grad_norm_pre_clip': 0.14442236721515656, 'timing/data': 0.0003296100185252726, 'timing/model': 0.3939280199992936, 'learning_rate/qwen_vl_interface': 6.484292513471352e-06, 'learning_rate/action_model': 6.484292513471352e-05, 'epoch': 0.16}) +[2026-10-03 19:21:31,157][starVLA.training.train_starvla][INFO] - Step 2141, Loss: {'train/loss': 0.007974009029567242, 'train/grad_norm_pre_clip': 10.182917594909668, 'timing/data': 0.00031984999077394605, 'timing/model': 0.3952467089984566, 'learning_rate/qwen_vl_interface': 6.4813514526961975e-06, 'learning_rate/action_model': 6.481351452696198e-05, 'epoch': 0.16}) +[2026-10-03 19:21:32,069][starVLA.training.train_starvla][INFO] - Step 2142, Loss: {'train/loss': 0.5030403733253479, 'train/grad_norm_pre_clip': 92.08124542236328, 'timing/data': 0.00047951002488844097, 'timing/model': 0.3945943599974271, 'learning_rate/qwen_vl_interface': 6.478409885758972e-06, 'learning_rate/action_model': 6.478409885758971e-05, 'epoch': 0.16}) +[2026-10-03 19:21:32,776][starVLA.training.train_starvla][INFO] - Step 2143, Loss: {'train/loss': 0.000762954237870872, 'train/grad_norm_pre_clip': 0.17418202757835388, 'timing/data': 0.0003985690127592534, 'timing/model': 0.3949613500153646, 'learning_rate/qwen_vl_interface': 6.475467813868844e-06, 'learning_rate/action_model': 6.475467813868844e-05, 'epoch': 0.16}) +[2026-10-03 19:21:33,481][starVLA.training.train_starvla][INFO] - Step 2144, Loss: {'train/loss': 0.0021089587826281786, 'train/grad_norm_pre_clip': 1.3315554857254028, 'timing/data': 0.00032203999580815434, 'timing/model': 0.39407366001978517, 'learning_rate/qwen_vl_interface': 6.472525238235187e-06, 'learning_rate/action_model': 6.472525238235186e-05, 'epoch': 0.16}) +[2026-10-03 19:21:34,188][starVLA.training.train_starvla][INFO] - Step 2145, Loss: {'train/loss': 0.2542673349380493, 'train/grad_norm_pre_clip': 45.39560317993164, 'timing/data': 0.00044289001380093396, 'timing/model': 0.39515462997951545, 'learning_rate/qwen_vl_interface': 6.4695821600675834e-06, 'learning_rate/action_model': 6.469582160067583e-05, 'epoch': 0.16}) +[2026-10-03 19:21:34,892][starVLA.training.train_starvla][INFO] - Step 2146, Loss: {'train/loss': 0.30839481949806213, 'train/grad_norm_pre_clip': 96.12454986572266, 'timing/data': 0.00029982998967170715, 'timing/model': 0.39355119998799637, 'learning_rate/qwen_vl_interface': 6.466638580575824e-06, 'learning_rate/action_model': 6.466638580575824e-05, 'epoch': 0.16}) +[2026-10-03 19:21:35,798][starVLA.training.train_starvla][INFO] - Step 2147, Loss: {'train/loss': 0.0053303842432796955, 'train/grad_norm_pre_clip': 19.42271614074707, 'timing/data': 0.0002507900062482804, 'timing/model': 0.5949335600016639, 'learning_rate/qwen_vl_interface': 6.463694500969898e-06, 'learning_rate/action_model': 6.463694500969898e-05, 'epoch': 0.16}) +[2026-10-03 19:21:36,502][starVLA.training.train_starvla][INFO] - Step 2148, Loss: {'train/loss': 0.9119235873222351, 'train/grad_norm_pre_clip': 64.31366729736328, 'timing/data': 0.0003063000040128827, 'timing/model': 0.3929261199955363, 'learning_rate/qwen_vl_interface': 6.460749922460008e-06, 'learning_rate/action_model': 6.460749922460008e-05, 'epoch': 0.16}) +[2026-10-03 19:21:37,206][starVLA.training.train_starvla][INFO] - Step 2149, Loss: {'train/loss': 0.000269152078544721, 'train/grad_norm_pre_clip': 0.11342859268188477, 'timing/data': 0.0003513200208544731, 'timing/model': 0.39319206998334266, 'learning_rate/qwen_vl_interface': 6.457804846256559e-06, 'learning_rate/action_model': 6.457804846256559e-05, 'epoch': 0.16}) +[2026-10-03 19:21:37,911][starVLA.training.train_starvla][INFO] - Step 2150, Loss: {'train/loss': 0.05017843469977379, 'train/grad_norm_pre_clip': 157.1945343017578, 'timing/data': 0.00038005999522283673, 'timing/model': 0.3930396100040525, 'learning_rate/qwen_vl_interface': 6.45485927357016e-06, 'learning_rate/action_model': 6.454859273570159e-05, 'epoch': 0.16}) +[2026-10-03 19:21:38,617][starVLA.training.train_starvla][INFO] - Step 2151, Loss: {'train/loss': 0.0006406266475096345, 'train/grad_norm_pre_clip': 0.7031802535057068, 'timing/data': 0.0003976890002377331, 'timing/model': 0.3947681800054852, 'learning_rate/qwen_vl_interface': 6.451913205611626e-06, 'learning_rate/action_model': 6.451913205611626e-05, 'epoch': 0.16}) +[2026-10-03 19:21:39,520][starVLA.training.train_starvla][INFO] - Step 2152, Loss: {'train/loss': 0.047199808061122894, 'train/grad_norm_pre_clip': 67.68159484863281, 'timing/data': 0.00034584998502396047, 'timing/model': 0.5907573800068349, 'learning_rate/qwen_vl_interface': 6.448966643591974e-06, 'learning_rate/action_model': 6.448966643591974e-05, 'epoch': 0.16}) +[2026-10-03 19:21:40,227][starVLA.training.train_starvla][INFO] - Step 2153, Loss: {'train/loss': 0.20656424760818481, 'train/grad_norm_pre_clip': 46.318267822265625, 'timing/data': 0.00021999000455252826, 'timing/model': 0.3956017000018619, 'learning_rate/qwen_vl_interface': 6.446019588722423e-06, 'learning_rate/action_model': 6.446019588722422e-05, 'epoch': 0.16}) +[2026-10-03 19:21:40,931][starVLA.training.train_starvla][INFO] - Step 2154, Loss: {'train/loss': 0.22336377203464508, 'train/grad_norm_pre_clip': 62.431331634521484, 'timing/data': 0.0002800700021907687, 'timing/model': 0.39384954000706784, 'learning_rate/qwen_vl_interface': 6.443072042214398e-06, 'learning_rate/action_model': 6.443072042214397e-05, 'epoch': 0.16}) +[2026-10-03 19:21:41,634][starVLA.training.train_starvla][INFO] - Step 2155, Loss: {'train/loss': 0.5714136362075806, 'train/grad_norm_pre_clip': 43.6549186706543, 'timing/data': 0.00021058000857010484, 'timing/model': 0.3930339600192383, 'learning_rate/qwen_vl_interface': 6.440124005279521e-06, 'learning_rate/action_model': 6.440124005279521e-05, 'epoch': 0.16}) +[2026-10-03 19:21:42,341][starVLA.training.train_starvla][INFO] - Step 2156, Loss: {'train/loss': 0.0528983473777771, 'train/grad_norm_pre_clip': 208.0263214111328, 'timing/data': 0.00022635998902842402, 'timing/model': 0.3945407799910754, 'learning_rate/qwen_vl_interface': 6.43717547912962e-06, 'learning_rate/action_model': 6.437175479129621e-05, 'epoch': 0.16}) +[2026-10-03 19:21:43,238][starVLA.training.train_starvla][INFO] - Step 2157, Loss: {'train/loss': 1.3868892192840576, 'train/grad_norm_pre_clip': 77.28253173828125, 'timing/data': 0.00028716999804601073, 'timing/model': 0.5875788600242231, 'learning_rate/qwen_vl_interface': 6.434226464976725e-06, 'learning_rate/action_model': 6.434226464976724e-05, 'epoch': 0.16}) +[2026-10-03 19:21:43,946][starVLA.training.train_starvla][INFO] - Step 2158, Loss: {'train/loss': 0.8336524963378906, 'train/grad_norm_pre_clip': 227.05406188964844, 'timing/data': 0.0002456500078551471, 'timing/model': 0.3955172009882517, 'learning_rate/qwen_vl_interface': 6.431276964033062e-06, 'learning_rate/action_model': 6.431276964033062e-05, 'epoch': 0.16}) +[2026-10-03 19:21:44,652][starVLA.training.train_starvla][INFO] - Step 2159, Loss: {'train/loss': 0.47340014576911926, 'train/grad_norm_pre_clip': 73.18113708496094, 'timing/data': 0.000330870010657236, 'timing/model': 0.39401045101112686, 'learning_rate/qwen_vl_interface': 6.428326977511059e-06, 'learning_rate/action_model': 6.428326977511058e-05, 'epoch': 0.16}) +[2026-10-03 19:21:45,357][starVLA.training.train_starvla][INFO] - Step 2160, Loss: {'train/loss': 0.0006483497563749552, 'train/grad_norm_pre_clip': 0.09974367171525955, 'timing/data': 0.00033683000947348773, 'timing/model': 0.39390027002082206, 'learning_rate/qwen_vl_interface': 6.425376506623344e-06, 'learning_rate/action_model': 6.425376506623344e-05, 'epoch': 0.16}) +[2026-10-03 19:21:46,063][starVLA.training.train_starvla][INFO] - Step 2161, Loss: {'train/loss': 0.00770129170268774, 'train/grad_norm_pre_clip': 6.092097759246826, 'timing/data': 0.00026664999313652515, 'timing/model': 0.3944280110008549, 'learning_rate/qwen_vl_interface': 6.422425552582747e-06, 'learning_rate/action_model': 6.422425552582747e-05, 'epoch': 0.16}) +[2026-10-03 19:21:46,767][starVLA.training.train_starvla][INFO] - Step 2162, Loss: {'train/loss': 0.5167365074157715, 'train/grad_norm_pre_clip': 95.08940124511719, 'timing/data': 0.0003373200015630573, 'timing/model': 0.3937518009915948, 'learning_rate/qwen_vl_interface': 6.41947411660229e-06, 'learning_rate/action_model': 6.41947411660229e-05, 'epoch': 0.16}) +[2026-10-03 19:21:47,682][starVLA.training.train_starvla][INFO] - Step 2163, Loss: {'train/loss': 0.00037616718327626586, 'train/grad_norm_pre_clip': 0.2397524118423462, 'timing/data': 0.0005147499905433506, 'timing/model': 0.39346120099071413, 'learning_rate/qwen_vl_interface': 6.4165221998951995e-06, 'learning_rate/action_model': 6.4165221998952e-05, 'epoch': 0.16}) +[2026-10-03 19:21:48,388][starVLA.training.train_starvla][INFO] - Step 2164, Loss: {'train/loss': 0.0009166104719042778, 'train/grad_norm_pre_clip': 1.2955771684646606, 'timing/data': 0.0003072399995289743, 'timing/model': 0.395540350000374, 'learning_rate/qwen_vl_interface': 6.413569803674896e-06, 'learning_rate/action_model': 6.413569803674896e-05, 'epoch': 0.16}) +[2026-10-03 19:21:49,095][starVLA.training.train_starvla][INFO] - Step 2165, Loss: {'train/loss': 0.04421908035874367, 'train/grad_norm_pre_clip': 21.478727340698242, 'timing/data': 0.00030931999208405614, 'timing/model': 0.39430704098776914, 'learning_rate/qwen_vl_interface': 6.410616929154997e-06, 'learning_rate/action_model': 6.410616929154996e-05, 'epoch': 0.16}) +[2026-10-03 19:21:49,800][starVLA.training.train_starvla][INFO] - Step 2166, Loss: {'train/loss': 0.002330436371266842, 'train/grad_norm_pre_clip': 0.5360031723976135, 'timing/data': 0.00035930000012740493, 'timing/model': 0.3943292400217615, 'learning_rate/qwen_vl_interface': 6.40766357754932e-06, 'learning_rate/action_model': 6.407663577549319e-05, 'epoch': 0.16}) +[2026-10-03 19:21:50,507][starVLA.training.train_starvla][INFO] - Step 2167, Loss: {'train/loss': 0.24040235579013824, 'train/grad_norm_pre_clip': 279.0812072753906, 'timing/data': 0.00026107000303454697, 'timing/model': 0.39531489901128225, 'learning_rate/qwen_vl_interface': 6.404709750071875e-06, 'learning_rate/action_model': 6.404709750071874e-05, 'epoch': 0.16}) +[2026-10-03 19:21:51,414][starVLA.training.train_starvla][INFO] - Step 2168, Loss: {'train/loss': 0.0027794816996902227, 'train/grad_norm_pre_clip': 0.947662889957428, 'timing/data': 0.0003274400078225881, 'timing/model': 0.5958052400092129, 'learning_rate/qwen_vl_interface': 6.401755447936869e-06, 'learning_rate/action_model': 6.401755447936869e-05, 'epoch': 0.16}) +[2026-10-03 19:21:52,119][starVLA.training.train_starvla][INFO] - Step 2169, Loss: {'train/loss': 0.004265733063220978, 'train/grad_norm_pre_clip': 2.0180246829986572, 'timing/data': 0.00026169000193476677, 'timing/model': 0.39531635999446735, 'learning_rate/qwen_vl_interface': 6.3988006723587036e-06, 'learning_rate/action_model': 6.398800672358703e-05, 'epoch': 0.16}) +[2026-10-03 19:21:52,824][starVLA.training.train_starvla][INFO] - Step 2170, Loss: {'train/loss': 0.0009919009171426296, 'train/grad_norm_pre_clip': 0.718358039855957, 'timing/data': 0.0003217000048607588, 'timing/model': 0.3934682709805202, 'learning_rate/qwen_vl_interface': 6.395845424551975e-06, 'learning_rate/action_model': 6.395845424551975e-05, 'epoch': 0.16}) +[2026-10-03 19:21:53,532][starVLA.training.train_starvla][INFO] - Step 2171, Loss: {'train/loss': 0.5389916896820068, 'train/grad_norm_pre_clip': 36.29062271118164, 'timing/data': 0.0003423999878577888, 'timing/model': 0.39500386000145227, 'learning_rate/qwen_vl_interface': 6.392889705731475e-06, 'learning_rate/action_model': 6.392889705731476e-05, 'epoch': 0.16}) +[2026-10-03 19:21:54,238][starVLA.training.train_starvla][INFO] - Step 2172, Loss: {'train/loss': 0.03242485225200653, 'train/grad_norm_pre_clip': 35.968963623046875, 'timing/data': 0.0002327999973203987, 'timing/model': 0.39440029999241233, 'learning_rate/qwen_vl_interface': 6.389933517112188e-06, 'learning_rate/action_model': 6.389933517112187e-05, 'epoch': 0.16}) +[2026-10-03 19:21:54,941][starVLA.training.train_starvla][INFO] - Step 2173, Loss: {'train/loss': 0.0006576559389941394, 'train/grad_norm_pre_clip': 0.23434551060199738, 'timing/data': 0.0002962800208479166, 'timing/model': 0.393462450010702, 'learning_rate/qwen_vl_interface': 6.386976859909293e-06, 'learning_rate/action_model': 6.386976859909292e-05, 'epoch': 0.16}) +[2026-10-03 19:21:55,838][starVLA.training.train_starvla][INFO] - Step 2174, Loss: {'train/loss': 0.0033857880625873804, 'train/grad_norm_pre_clip': 1.504388451576233, 'timing/data': 0.00036307002301327884, 'timing/model': 0.39436549998936243, 'learning_rate/qwen_vl_interface': 6.384019735338156e-06, 'learning_rate/action_model': 6.384019735338156e-05, 'epoch': 0.16}) +[2026-10-03 19:21:56,544][starVLA.training.train_starvla][INFO] - Step 2175, Loss: {'train/loss': 0.11140428483486176, 'train/grad_norm_pre_clip': 56.29863739013672, 'timing/data': 0.00024901001597754657, 'timing/model': 0.3942630710080266, 'learning_rate/qwen_vl_interface': 6.381062144614342e-06, 'learning_rate/action_model': 6.381062144614342e-05, 'epoch': 0.16}) +[2026-10-03 19:21:57,249][starVLA.training.train_starvla][INFO] - Step 2176, Loss: {'train/loss': 0.0011632745154201984, 'train/grad_norm_pre_clip': 0.24336956441402435, 'timing/data': 0.00020322998170740902, 'timing/model': 0.3945039700192865, 'learning_rate/qwen_vl_interface': 6.378104088953605e-06, 'learning_rate/action_model': 6.378104088953605e-05, 'epoch': 0.16}) +[2026-10-03 19:21:57,954][starVLA.training.train_starvla][INFO] - Step 2177, Loss: {'train/loss': 0.03305936977267265, 'train/grad_norm_pre_clip': 13.016068458557129, 'timing/data': 0.0002446400176268071, 'timing/model': 0.3937198809871916, 'learning_rate/qwen_vl_interface': 6.375145569571889e-06, 'learning_rate/action_model': 6.375145569571889e-05, 'epoch': 0.16}) +[2026-10-03 19:21:58,662][starVLA.training.train_starvla][INFO] - Step 2178, Loss: {'train/loss': 0.001038375194184482, 'train/grad_norm_pre_clip': 0.5692172646522522, 'timing/data': 0.0003393900115042925, 'timing/model': 0.39574646000983194, 'learning_rate/qwen_vl_interface': 6.372186587685329e-06, 'learning_rate/action_model': 6.372186587685329e-05, 'epoch': 0.16}) +[2026-10-03 19:21:59,577][starVLA.training.train_starvla][INFO] - Step 2179, Loss: {'train/loss': 0.21759159862995148, 'train/grad_norm_pre_clip': 31.717695236206055, 'timing/data': 0.00023318000603467226, 'timing/model': 0.3970788199803792, 'learning_rate/qwen_vl_interface': 6.369227144510252e-06, 'learning_rate/action_model': 6.369227144510252e-05, 'epoch': 0.16}) +[2026-10-03 19:22:00,286][starVLA.training.train_starvla][INFO] - Step 2180, Loss: {'train/loss': 0.0004407476808410138, 'train/grad_norm_pre_clip': 0.16520607471466064, 'timing/data': 0.0003508899826556444, 'timing/model': 0.3960310800175648, 'learning_rate/qwen_vl_interface': 6.3662672412631736e-06, 'learning_rate/action_model': 6.366267241263173e-05, 'epoch': 0.16}) +[2026-10-03 19:22:00,994][starVLA.training.train_starvla][INFO] - Step 2181, Loss: {'train/loss': 0.33406156301498413, 'train/grad_norm_pre_clip': 85.30989837646484, 'timing/data': 0.00030297000193968415, 'timing/model': 0.39525650002178736, 'learning_rate/qwen_vl_interface': 6.3633068791607954e-06, 'learning_rate/action_model': 6.363306879160795e-05, 'epoch': 0.16}) +[2026-10-03 19:22:01,705][starVLA.training.train_starvla][INFO] - Step 2182, Loss: {'train/loss': 0.6635932922363281, 'train/grad_norm_pre_clip': 88.68806457519531, 'timing/data': 0.00035473998286761343, 'timing/model': 0.3976426290173549, 'learning_rate/qwen_vl_interface': 6.360346059420014e-06, 'learning_rate/action_model': 6.360346059420014e-05, 'epoch': 0.16}) +[2026-10-03 19:22:02,416][starVLA.training.train_starvla][INFO] - Step 2183, Loss: {'train/loss': 0.04723529890179634, 'train/grad_norm_pre_clip': 77.19734191894531, 'timing/data': 0.00040483998600393534, 'timing/model': 0.39619080899865367, 'learning_rate/qwen_vl_interface': 6.357384783257909e-06, 'learning_rate/action_model': 6.357384783257908e-05, 'epoch': 0.16}) +[2026-10-03 19:22:03,329][starVLA.training.train_starvla][INFO] - Step 2184, Loss: {'train/loss': 0.0011733982246369123, 'train/grad_norm_pre_clip': 0.4957186281681061, 'timing/data': 0.00034986998070962727, 'timing/model': 0.6018575089983642, 'learning_rate/qwen_vl_interface': 6.354423051891748e-06, 'learning_rate/action_model': 6.354423051891749e-05, 'epoch': 0.16}) +[2026-10-03 19:22:04,039][starVLA.training.train_starvla][INFO] - Step 2185, Loss: {'train/loss': 0.42198142409324646, 'train/grad_norm_pre_clip': 46.50231170654297, 'timing/data': 0.0003466100024525076, 'timing/model': 0.3965018499875441, 'learning_rate/qwen_vl_interface': 6.35146086653899e-06, 'learning_rate/action_model': 6.35146086653899e-05, 'epoch': 0.16}) +[2026-10-03 19:22:04,747][starVLA.training.train_starvla][INFO] - Step 2186, Loss: {'train/loss': 0.0012740090023726225, 'train/grad_norm_pre_clip': 0.6119276285171509, 'timing/data': 0.00028083001961931586, 'timing/model': 0.3958755299972836, 'learning_rate/qwen_vl_interface': 6.348498228417273e-06, 'learning_rate/action_model': 6.348498228417273e-05, 'epoch': 0.16}) +[2026-10-03 19:22:05,456][starVLA.training.train_starvla][INFO] - Step 2187, Loss: {'train/loss': 0.7693472504615784, 'train/grad_norm_pre_clip': 121.25482940673828, 'timing/data': 0.000255580001976341, 'timing/model': 0.3961250190041028, 'learning_rate/qwen_vl_interface': 6.34553513874443e-06, 'learning_rate/action_model': 6.34553513874443e-05, 'epoch': 0.16}) +[2026-10-03 19:22:06,164][starVLA.training.train_starvla][INFO] - Step 2188, Loss: {'train/loss': 0.026226650923490524, 'train/grad_norm_pre_clip': 54.88538360595703, 'timing/data': 0.0003097500011790544, 'timing/model': 0.3945642600010615, 'learning_rate/qwen_vl_interface': 6.342571598738473e-06, 'learning_rate/action_model': 6.342571598738472e-05, 'epoch': 0.16}) +[2026-10-03 19:22:06,871][starVLA.training.train_starvla][INFO] - Step 2189, Loss: {'train/loss': 0.30311325192451477, 'train/grad_norm_pre_clip': 102.04691314697266, 'timing/data': 0.0003424400056246668, 'timing/model': 0.39498522999929264, 'learning_rate/qwen_vl_interface': 6.339607609617603e-06, 'learning_rate/action_model': 6.339607609617603e-05, 'epoch': 0.16}) +[2026-10-03 19:22:07,780][starVLA.training.train_starvla][INFO] - Step 2190, Loss: {'train/loss': 0.6003843545913696, 'train/grad_norm_pre_clip': 34.11924743652344, 'timing/data': 0.00023003999376669526, 'timing/model': 0.3947661900019739, 'learning_rate/qwen_vl_interface': 6.336643172600202e-06, 'learning_rate/action_model': 6.336643172600202e-05, 'epoch': 0.16}) +[2026-10-03 19:22:08,488][starVLA.training.train_starvla][INFO] - Step 2191, Loss: {'train/loss': 0.22928446531295776, 'train/grad_norm_pre_clip': 56.82023239135742, 'timing/data': 0.00023152001085691154, 'timing/model': 0.3964910999929998, 'learning_rate/qwen_vl_interface': 6.333678288904838e-06, 'learning_rate/action_model': 6.333678288904837e-05, 'epoch': 0.16}) +[2026-10-03 19:22:09,197][starVLA.training.train_starvla][INFO] - Step 2192, Loss: {'train/loss': 0.009733515791594982, 'train/grad_norm_pre_clip': 4.060074806213379, 'timing/data': 0.0003035600238945335, 'timing/model': 0.3957685010100249, 'learning_rate/qwen_vl_interface': 6.330712959750264e-06, 'learning_rate/action_model': 6.330712959750263e-05, 'epoch': 0.16}) +[2026-10-03 19:22:09,905][starVLA.training.train_starvla][INFO] - Step 2193, Loss: {'train/loss': 0.005427396856248379, 'train/grad_norm_pre_clip': 7.140961170196533, 'timing/data': 0.00021987000945955515, 'timing/model': 0.39506428100867197, 'learning_rate/qwen_vl_interface': 6.3277471863554145e-06, 'learning_rate/action_model': 6.327747186355414e-05, 'epoch': 0.16}) +[2026-10-03 19:22:10,616][starVLA.training.train_starvla][INFO] - Step 2194, Loss: {'train/loss': 0.001327157486230135, 'train/grad_norm_pre_clip': 0.7281147241592407, 'timing/data': 0.00036325998371466994, 'timing/model': 0.39709584999945946, 'learning_rate/qwen_vl_interface': 6.324780969939406e-06, 'learning_rate/action_model': 6.324780969939405e-05, 'epoch': 0.16}) +[2026-10-03 19:22:11,531][starVLA.training.train_starvla][INFO] - Step 2195, Loss: {'train/loss': 0.0016104229725897312, 'train/grad_norm_pre_clip': 0.39253827929496765, 'timing/data': 0.00022877001902088523, 'timing/model': 0.3962917400058359, 'learning_rate/qwen_vl_interface': 6.32181431172154e-06, 'learning_rate/action_model': 6.32181431172154e-05, 'epoch': 0.16}) +[2026-10-03 19:22:12,240][starVLA.training.train_starvla][INFO] - Step 2196, Loss: {'train/loss': 0.3880006670951843, 'train/grad_norm_pre_clip': 45.75297164916992, 'timing/data': 0.00033042000723071396, 'timing/model': 0.3957313100108877, 'learning_rate/qwen_vl_interface': 6.318847212921295e-06, 'learning_rate/action_model': 6.318847212921295e-05, 'epoch': 0.16}) +[2026-10-03 19:22:12,948][starVLA.training.train_starvla][INFO] - Step 2197, Loss: {'train/loss': 0.009918715804815292, 'train/grad_norm_pre_clip': 5.618372440338135, 'timing/data': 0.00032212998485192657, 'timing/model': 0.3950322500022594, 'learning_rate/qwen_vl_interface': 6.3158796747583364e-06, 'learning_rate/action_model': 6.315879674758336e-05, 'epoch': 0.16}) +[2026-10-03 19:22:13,657][starVLA.training.train_starvla][INFO] - Step 2198, Loss: {'train/loss': 0.0002085970190819353, 'train/grad_norm_pre_clip': 0.06895247846841812, 'timing/data': 0.00035314998240210116, 'timing/model': 0.3965786990011111, 'learning_rate/qwen_vl_interface': 6.312911698452503e-06, 'learning_rate/action_model': 6.312911698452502e-05, 'epoch': 0.16}) +[2026-10-03 19:22:14,365][starVLA.training.train_starvla][INFO] - Step 2199, Loss: {'train/loss': 0.5881393551826477, 'train/grad_norm_pre_clip': 152.93630981445312, 'timing/data': 0.00031210997258313, 'timing/model': 0.3944110790034756, 'learning_rate/qwen_vl_interface': 6.309943285223824e-06, 'learning_rate/action_model': 6.309943285223824e-05, 'epoch': 0.16}) +[2026-10-03 19:22:15,270][starVLA.training.train_starvla][INFO] - Step 2200, Loss: {'train/loss': 0.003266551997512579, 'train/grad_norm_pre_clip': 2.947511911392212, 'timing/data': 0.00030978998984210193, 'timing/model': 0.592242519982392, 'learning_rate/qwen_vl_interface': 6.306974436292493e-06, 'learning_rate/action_model': 6.306974436292494e-05, 'epoch': 0.16}) +[2026-10-03 19:22:15,978][starVLA.training.train_starvla][INFO] - Step 2201, Loss: {'train/loss': 0.00016616471111774445, 'train/grad_norm_pre_clip': 0.033101920038461685, 'timing/data': 0.00038253000820986927, 'timing/model': 0.39609143999405205, 'learning_rate/qwen_vl_interface': 6.304005152878902e-06, 'learning_rate/action_model': 6.304005152878901e-05, 'epoch': 0.16}) +[2026-10-03 19:22:16,686][starVLA.training.train_starvla][INFO] - Step 2202, Loss: {'train/loss': 0.00025178404757753015, 'train/grad_norm_pre_clip': 0.09501088410615921, 'timing/data': 0.00026821999927051365, 'timing/model': 0.39606778899906203, 'learning_rate/qwen_vl_interface': 6.301035436203602e-06, 'learning_rate/action_model': 6.301035436203602e-05, 'epoch': 0.16}) +[2026-10-03 19:22:17,394][starVLA.training.train_starvla][INFO] - Step 2203, Loss: {'train/loss': 0.21379604935646057, 'train/grad_norm_pre_clip': 40.62726593017578, 'timing/data': 0.00023020000662654638, 'timing/model': 0.3945111899811309, 'learning_rate/qwen_vl_interface': 6.298065287487337e-06, 'learning_rate/action_model': 6.298065287487337e-05, 'epoch': 0.16}) +[2026-10-03 19:22:18,100][starVLA.training.train_starvla][INFO] - Step 2204, Loss: {'train/loss': 0.5315817594528198, 'train/grad_norm_pre_clip': 108.66688537597656, 'timing/data': 0.0003003900055773556, 'timing/model': 0.3955337000079453, 'learning_rate/qwen_vl_interface': 6.295094707951019e-06, 'learning_rate/action_model': 6.295094707951019e-05, 'epoch': 0.16}) +[2026-10-03 19:22:19,009][starVLA.training.train_starvla][INFO] - Step 2205, Loss: {'train/loss': 0.1711873561143875, 'train/grad_norm_pre_clip': 98.77156066894531, 'timing/data': 0.00029553999775089324, 'timing/model': 0.5969855099974666, 'learning_rate/qwen_vl_interface': 6.292123698815745e-06, 'learning_rate/action_model': 6.292123698815745e-05, 'epoch': 0.16}) +[2026-10-03 19:22:19,718][starVLA.training.train_starvla][INFO] - Step 2206, Loss: {'train/loss': 0.8669663667678833, 'train/grad_norm_pre_clip': 186.57772827148438, 'timing/data': 0.00021780998213216662, 'timing/model': 0.39626138101448305, 'learning_rate/qwen_vl_interface': 6.289152261302783e-06, 'learning_rate/action_model': 6.289152261302784e-05, 'epoch': 0.16}) +[2026-10-03 19:22:20,424][starVLA.training.train_starvla][INFO] - Step 2207, Loss: {'train/loss': 0.013451449573040009, 'train/grad_norm_pre_clip': 20.521345138549805, 'timing/data': 0.00021945001208223403, 'timing/model': 0.3939868510060478, 'learning_rate/qwen_vl_interface': 6.286180396633579e-06, 'learning_rate/action_model': 6.286180396633579e-05, 'epoch': 0.16}) +[2026-10-03 19:22:21,131][starVLA.training.train_starvla][INFO] - Step 2208, Loss: {'train/loss': 0.002355787204578519, 'train/grad_norm_pre_clip': 2.347227096557617, 'timing/data': 0.0003209799760952592, 'timing/model': 0.3953873399877921, 'learning_rate/qwen_vl_interface': 6.283208106029753e-06, 'learning_rate/action_model': 6.283208106029752e-05, 'epoch': 0.16}) +[2026-10-03 19:22:21,840][starVLA.training.train_starvla][INFO] - Step 2209, Loss: {'train/loss': 0.03127724677324295, 'train/grad_norm_pre_clip': 37.056297302246094, 'timing/data': 0.0004603099951054901, 'timing/model': 0.3962223500129767, 'learning_rate/qwen_vl_interface': 6.280235390713102e-06, 'learning_rate/action_model': 6.280235390713102e-05, 'epoch': 0.16}) +[2026-10-03 19:22:22,768][starVLA.training.train_starvla][INFO] - Step 2210, Loss: {'train/loss': 0.09454569965600967, 'train/grad_norm_pre_clip': 192.49916076660156, 'timing/data': 0.000328070018440485, 'timing/model': 0.6137868570222054, 'learning_rate/qwen_vl_interface': 6.277262251905597e-06, 'learning_rate/action_model': 6.277262251905597e-05, 'epoch': 0.16}) +[2026-10-03 19:22:23,475][starVLA.training.train_starvla][INFO] - Step 2211, Loss: {'train/loss': 0.22096586227416992, 'train/grad_norm_pre_clip': 123.34043884277344, 'timing/data': 0.00041599999531172216, 'timing/model': 0.3957170200010296, 'learning_rate/qwen_vl_interface': 6.274288690829385e-06, 'learning_rate/action_model': 6.274288690829384e-05, 'epoch': 0.16}) +[2026-10-03 19:22:24,184][starVLA.training.train_starvla][INFO] - Step 2212, Loss: {'train/loss': 0.0003675897023640573, 'train/grad_norm_pre_clip': 0.08394142985343933, 'timing/data': 0.00029644998721778393, 'timing/model': 0.3957314500003122, 'learning_rate/qwen_vl_interface': 6.271314708706781e-06, 'learning_rate/action_model': 6.271314708706782e-05, 'epoch': 0.16}) +[2026-10-03 19:22:24,893][starVLA.training.train_starvla][INFO] - Step 2213, Loss: {'train/loss': 0.272929310798645, 'train/grad_norm_pre_clip': 46.20651626586914, 'timing/data': 0.00037857997813262045, 'timing/model': 0.39607236001756974, 'learning_rate/qwen_vl_interface': 6.26834030676028e-06, 'learning_rate/action_model': 6.26834030676028e-05, 'epoch': 0.16}) +[2026-10-03 19:22:25,602][starVLA.training.train_starvla][INFO] - Step 2214, Loss: {'train/loss': 0.20035289227962494, 'train/grad_norm_pre_clip': 43.476539611816406, 'timing/data': 0.00040520000038668513, 'timing/model': 0.39584216999355704, 'learning_rate/qwen_vl_interface': 6.265365486212545e-06, 'learning_rate/action_model': 6.265365486212545e-05, 'epoch': 0.16}) +[2026-10-03 19:22:26,310][starVLA.training.train_starvla][INFO] - Step 2215, Loss: {'train/loss': 0.003599404124543071, 'train/grad_norm_pre_clip': 3.2420873641967773, 'timing/data': 0.00027062001754529774, 'timing/model': 0.39600744101335295, 'learning_rate/qwen_vl_interface': 6.26239024828641e-06, 'learning_rate/action_model': 6.26239024828641e-05, 'epoch': 0.16}) +[2026-10-03 19:22:27,223][starVLA.training.train_starvla][INFO] - Step 2216, Loss: {'train/loss': 0.31650790572166443, 'train/grad_norm_pre_clip': 100.99673461914062, 'timing/data': 0.000318949983920902, 'timing/model': 0.3962634999770671, 'learning_rate/qwen_vl_interface': 6.259414594204888e-06, 'learning_rate/action_model': 6.259414594204888e-05, 'epoch': 0.16}) +[2026-10-03 19:22:27,930][starVLA.training.train_starvla][INFO] - Step 2217, Loss: {'train/loss': 0.27710404992103577, 'train/grad_norm_pre_clip': 76.08171081542969, 'timing/data': 0.00034042997867800295, 'timing/model': 0.39458257000660524, 'learning_rate/qwen_vl_interface': 6.256438525191155e-06, 'learning_rate/action_model': 6.256438525191156e-05, 'epoch': 0.16}) +[2026-10-03 19:22:28,636][starVLA.training.train_starvla][INFO] - Step 2218, Loss: {'train/loss': 0.003067068289965391, 'train/grad_norm_pre_clip': 1.6697579622268677, 'timing/data': 0.0003975899890065193, 'timing/model': 0.3947049909911584, 'learning_rate/qwen_vl_interface': 6.253462042468561e-06, 'learning_rate/action_model': 6.25346204246856e-05, 'epoch': 0.16}) +[2026-10-03 19:22:29,346][starVLA.training.train_starvla][INFO] - Step 2219, Loss: {'train/loss': 0.004317650105804205, 'train/grad_norm_pre_clip': 5.842769622802734, 'timing/data': 0.00030827001319266856, 'timing/model': 0.3967450800118968, 'learning_rate/qwen_vl_interface': 6.2504851472606246e-06, 'learning_rate/action_model': 6.250485147260625e-05, 'epoch': 0.16}) +[2026-10-03 19:22:30,051][starVLA.training.train_starvla][INFO] - Step 2220, Loss: {'train/loss': 0.00028580264188349247, 'train/grad_norm_pre_clip': 0.2103872150182724, 'timing/data': 0.0002666700165718794, 'timing/model': 0.3936697600001935, 'learning_rate/qwen_vl_interface': 6.247507840791035e-06, 'learning_rate/action_model': 6.247507840791035e-05, 'epoch': 0.16}) +[2026-10-03 19:22:30,959][starVLA.training.train_starvla][INFO] - Step 2221, Loss: {'train/loss': 0.0009898218559101224, 'train/grad_norm_pre_clip': 1.5133835077285767, 'timing/data': 0.00031848999788053334, 'timing/model': 0.5969173789781053, 'learning_rate/qwen_vl_interface': 6.244530124283651e-06, 'learning_rate/action_model': 6.24453012428365e-05, 'epoch': 0.16}) +[2026-10-03 19:22:31,667][starVLA.training.train_starvla][INFO] - Step 2222, Loss: {'train/loss': 0.0014460600214079022, 'train/grad_norm_pre_clip': 0.8216686844825745, 'timing/data': 0.0004503399832174182, 'timing/model': 0.39600174000952393, 'learning_rate/qwen_vl_interface': 6.2415519989625016e-06, 'learning_rate/action_model': 6.241551998962501e-05, 'epoch': 0.16}) +[2026-10-03 19:22:32,372][starVLA.training.train_starvla][INFO] - Step 2223, Loss: {'train/loss': 0.1068953275680542, 'train/grad_norm_pre_clip': 116.451904296875, 'timing/data': 0.0002528300101403147, 'timing/model': 0.3951534499938134, 'learning_rate/qwen_vl_interface': 6.238573466051781e-06, 'learning_rate/action_model': 6.23857346605178e-05, 'epoch': 0.16}) +[2026-10-03 19:22:33,078][starVLA.training.train_starvla][INFO] - Step 2224, Loss: {'train/loss': 0.4974607825279236, 'train/grad_norm_pre_clip': 144.4217987060547, 'timing/data': 0.0002174199908040464, 'timing/model': 0.39404180998099037, 'learning_rate/qwen_vl_interface': 6.2355945267758465e-06, 'learning_rate/action_model': 6.235594526775846e-05, 'epoch': 0.16}) +[2026-10-03 19:22:33,786][starVLA.training.train_starvla][INFO] - Step 2225, Loss: {'train/loss': 0.0011356205213814974, 'train/grad_norm_pre_clip': 0.6279087662696838, 'timing/data': 0.0003051800013054162, 'timing/model': 0.39536843000678346, 'learning_rate/qwen_vl_interface': 6.232615182359231e-06, 'learning_rate/action_model': 6.232615182359231e-05, 'epoch': 0.16}) +[2026-10-03 19:22:34,492][starVLA.training.train_starvla][INFO] - Step 2226, Loss: {'train/loss': 0.005737426690757275, 'train/grad_norm_pre_clip': 12.0380859375, 'timing/data': 0.00040273001650348306, 'timing/model': 0.3945970600179862, 'learning_rate/qwen_vl_interface': 6.229635434026634e-06, 'learning_rate/action_model': 6.229635434026634e-05, 'epoch': 0.16}) +[2026-10-03 19:22:35,416][starVLA.training.train_starvla][INFO] - Step 2227, Loss: {'train/loss': 0.37334781885147095, 'train/grad_norm_pre_clip': 53.364139556884766, 'timing/data': 0.0003885899786837399, 'timing/model': 0.39535053897998296, 'learning_rate/qwen_vl_interface': 6.226655283002912e-06, 'learning_rate/action_model': 6.226655283002911e-05, 'epoch': 0.16}) +[2026-10-03 19:22:36,123][starVLA.training.train_starvla][INFO] - Step 2228, Loss: {'train/loss': 0.0006488057551905513, 'train/grad_norm_pre_clip': 0.6359397172927856, 'timing/data': 0.00026436999905854464, 'timing/model': 0.3930476999958046, 'learning_rate/qwen_vl_interface': 6.223674730513096e-06, 'learning_rate/action_model': 6.223674730513096e-05, 'epoch': 0.16}) +[2026-10-03 19:22:36,829][starVLA.training.train_starvla][INFO] - Step 2229, Loss: {'train/loss': 0.00010839973401743919, 'train/grad_norm_pre_clip': 0.03855939209461212, 'timing/data': 0.0003407399926800281, 'timing/model': 0.394984751008451, 'learning_rate/qwen_vl_interface': 6.220693777782375e-06, 'learning_rate/action_model': 6.220693777782375e-05, 'epoch': 0.16}) +[2026-10-03 19:22:37,533][starVLA.training.train_starvla][INFO] - Step 2230, Loss: {'train/loss': 0.38915008306503296, 'train/grad_norm_pre_clip': 184.7821807861328, 'timing/data': 0.00036551000084728, 'timing/model': 0.3944658100081142, 'learning_rate/qwen_vl_interface': 6.21771242603611e-06, 'learning_rate/action_model': 6.21771242603611e-05, 'epoch': 0.16}) +[2026-10-03 19:22:38,238][starVLA.training.train_starvla][INFO] - Step 2231, Loss: {'train/loss': 0.11838862299919128, 'train/grad_norm_pre_clip': 49.90711975097656, 'timing/data': 0.0003271099994890392, 'timing/model': 0.3931180909858085, 'learning_rate/qwen_vl_interface': 6.21473067649982e-06, 'learning_rate/action_model': 6.21473067649982e-05, 'epoch': 0.16}) +[2026-10-03 19:22:39,147][starVLA.training.train_starvla][INFO] - Step 2232, Loss: {'train/loss': 0.2493281066417694, 'train/grad_norm_pre_clip': 65.05516052246094, 'timing/data': 0.0003615899768192321, 'timing/model': 0.3950705399911385, 'learning_rate/qwen_vl_interface': 6.211748530399189e-06, 'learning_rate/action_model': 6.211748530399189e-05, 'epoch': 0.16}) +[2026-10-03 19:22:39,850][starVLA.training.train_starvla][INFO] - Step 2233, Loss: {'train/loss': 0.00015378631360363215, 'train/grad_norm_pre_clip': 0.046080660074949265, 'timing/data': 0.00021969000226818025, 'timing/model': 0.39364466100232676, 'learning_rate/qwen_vl_interface': 6.208765988960067e-06, 'learning_rate/action_model': 6.208765988960066e-05, 'epoch': 0.16}) +[2026-10-03 19:22:40,556][starVLA.training.train_starvla][INFO] - Step 2234, Loss: {'train/loss': 0.22768396139144897, 'train/grad_norm_pre_clip': 193.46002197265625, 'timing/data': 0.0003598600160330534, 'timing/model': 0.3942284199874848, 'learning_rate/qwen_vl_interface': 6.2057830534084615e-06, 'learning_rate/action_model': 6.205783053408461e-05, 'epoch': 0.16}) +[2026-10-03 19:22:41,262][starVLA.training.train_starvla][INFO] - Step 2235, Loss: {'train/loss': 0.4099642038345337, 'train/grad_norm_pre_clip': 18.196420669555664, 'timing/data': 0.00022079001064412296, 'timing/model': 0.3945719310140703, 'learning_rate/qwen_vl_interface': 6.202799724970546e-06, 'learning_rate/action_model': 6.202799724970546e-05, 'epoch': 0.16}) +[2026-10-03 19:22:41,967][starVLA.training.train_starvla][INFO] - Step 2236, Loss: {'train/loss': 0.04530748352408409, 'train/grad_norm_pre_clip': 26.08931541442871, 'timing/data': 0.0003225299878977239, 'timing/model': 0.39406359000713564, 'learning_rate/qwen_vl_interface': 6.199816004872655e-06, 'learning_rate/action_model': 6.199816004872655e-05, 'epoch': 0.16}) +[2026-10-03 19:22:42,872][starVLA.training.train_starvla][INFO] - Step 2237, Loss: {'train/loss': 0.00015591034025419503, 'train/grad_norm_pre_clip': 0.04011409729719162, 'timing/data': 0.00032873000600375235, 'timing/model': 0.5937398700043559, 'learning_rate/qwen_vl_interface': 6.196831894341281e-06, 'learning_rate/action_model': 6.19683189434128e-05, 'epoch': 0.16}) +[2026-10-03 19:22:43,578][starVLA.training.train_starvla][INFO] - Step 2238, Loss: {'train/loss': 0.0011552763171494007, 'train/grad_norm_pre_clip': 2.036360263824463, 'timing/data': 0.0003741289838217199, 'timing/model': 0.3941867310204543, 'learning_rate/qwen_vl_interface': 6.1938473946030795e-06, 'learning_rate/action_model': 6.193847394603079e-05, 'epoch': 0.16}) +[2026-10-03 19:22:44,284][starVLA.training.train_starvla][INFO] - Step 2239, Loss: {'train/loss': 0.002228334080427885, 'train/grad_norm_pre_clip': 4.264377593994141, 'timing/data': 0.00034838999272324145, 'timing/model': 0.39441804902162403, 'learning_rate/qwen_vl_interface': 6.1908625068848695e-06, 'learning_rate/action_model': 6.19086250688487e-05, 'epoch': 0.16}) +[2026-10-03 19:22:44,989][starVLA.training.train_starvla][INFO] - Step 2240, Loss: {'train/loss': 0.15527671575546265, 'train/grad_norm_pre_clip': 40.64552307128906, 'timing/data': 0.0002183800097554922, 'timing/model': 0.3946488199871965, 'learning_rate/qwen_vl_interface': 6.187877232413619e-06, 'learning_rate/action_model': 6.18787723241362e-05, 'epoch': 0.16}) +[2026-10-03 19:22:45,692][starVLA.training.train_starvla][INFO] - Step 2241, Loss: {'train/loss': 0.0001860659394878894, 'train/grad_norm_pre_clip': 0.048705339431762695, 'timing/data': 0.0003233600000385195, 'timing/model': 0.39327813999261707, 'learning_rate/qwen_vl_interface': 6.184891572416467e-06, 'learning_rate/action_model': 6.184891572416466e-05, 'epoch': 0.16}) +[2026-10-03 19:22:46,396][starVLA.training.train_starvla][INFO] - Step 2242, Loss: {'train/loss': 0.0012022459413856268, 'train/grad_norm_pre_clip': 0.5122218132019043, 'timing/data': 0.00023214000975713134, 'timing/model': 0.3934039300074801, 'learning_rate/qwen_vl_interface': 6.181905528120703e-06, 'learning_rate/action_model': 6.181905528120703e-05, 'epoch': 0.16}) +[2026-10-03 19:22:47,302][starVLA.training.train_starvla][INFO] - Step 2243, Loss: {'train/loss': 0.00010128493886440992, 'train/grad_norm_pre_clip': 0.02934306114912033, 'timing/data': 0.00036707002436742187, 'timing/model': 0.395607079990441, 'learning_rate/qwen_vl_interface': 6.178919100753778e-06, 'learning_rate/action_model': 6.178919100753778e-05, 'epoch': 0.16}) +[2026-10-03 19:22:48,009][starVLA.training.train_starvla][INFO] - Step 2244, Loss: {'train/loss': 0.13017702102661133, 'train/grad_norm_pre_clip': 57.41140365600586, 'timing/data': 0.0004317999992053956, 'timing/model': 0.39476839001872577, 'learning_rate/qwen_vl_interface': 6.175932291543299e-06, 'learning_rate/action_model': 6.175932291543298e-05, 'epoch': 0.16}) +[2026-10-03 19:22:48,712][starVLA.training.train_starvla][INFO] - Step 2245, Loss: {'train/loss': 0.6605882048606873, 'train/grad_norm_pre_clip': 42.36871337890625, 'timing/data': 0.0002798600180540234, 'timing/model': 0.39309379999758676, 'learning_rate/qwen_vl_interface': 6.172945101717029e-06, 'learning_rate/action_model': 6.17294510171703e-05, 'epoch': 0.16}) +[2026-10-03 19:22:49,417][starVLA.training.train_starvla][INFO] - Step 2246, Loss: {'train/loss': 1.2673472166061401, 'train/grad_norm_pre_clip': 83.6395263671875, 'timing/data': 0.00032550000469200313, 'timing/model': 0.3951918189995922, 'learning_rate/qwen_vl_interface': 6.169957532502892e-06, 'learning_rate/action_model': 6.169957532502892e-05, 'epoch': 0.16}) +[2026-10-03 19:22:50,121][starVLA.training.train_starvla][INFO] - Step 2247, Loss: {'train/loss': 0.026598099619150162, 'train/grad_norm_pre_clip': 25.33021354675293, 'timing/data': 0.0003240699879825115, 'timing/model': 0.39444541998091154, 'learning_rate/qwen_vl_interface': 6.166969585128962e-06, 'learning_rate/action_model': 6.166969585128962e-05, 'epoch': 0.16}) +[2026-10-03 19:22:51,028][starVLA.training.train_starvla][INFO] - Step 2248, Loss: {'train/loss': 0.0006432172376662493, 'train/grad_norm_pre_clip': 0.4151197075843811, 'timing/data': 0.0003275600029155612, 'timing/model': 0.3935938699869439, 'learning_rate/qwen_vl_interface': 6.163981260823473e-06, 'learning_rate/action_model': 6.163981260823473e-05, 'epoch': 0.16}) +[2026-10-03 19:22:51,735][starVLA.training.train_starvla][INFO] - Step 2249, Loss: {'train/loss': 0.27866724133491516, 'train/grad_norm_pre_clip': 37.55439376831055, 'timing/data': 0.0003325500001665205, 'timing/model': 0.39485492900712416, 'learning_rate/qwen_vl_interface': 6.16099256081481e-06, 'learning_rate/action_model': 6.16099256081481e-05, 'epoch': 0.16}) +[2026-10-03 19:22:52,441][starVLA.training.train_starvla][INFO] - Step 2250, Loss: {'train/loss': 0.12187634408473969, 'train/grad_norm_pre_clip': 68.65890502929688, 'timing/data': 0.00022690999321639538, 'timing/model': 0.39498680000542663, 'learning_rate/qwen_vl_interface': 6.15800348633152e-06, 'learning_rate/action_model': 6.15800348633152e-05, 'epoch': 0.16}) +[2026-10-03 19:22:53,151][starVLA.training.train_starvla][INFO] - Step 2251, Loss: {'train/loss': 0.00030563600012101233, 'train/grad_norm_pre_clip': 0.19137218594551086, 'timing/data': 0.00022411998361349106, 'timing/model': 0.3944941000081599, 'learning_rate/qwen_vl_interface': 6.155014038602293e-06, 'learning_rate/action_model': 6.155014038602292e-05, 'epoch': 0.16}) +[2026-10-03 19:22:53,857][starVLA.training.train_starvla][INFO] - Step 2252, Loss: {'train/loss': 0.00164203776512295, 'train/grad_norm_pre_clip': 0.42730939388275146, 'timing/data': 0.00032916999771259725, 'timing/model': 0.3935691800143104, 'learning_rate/qwen_vl_interface': 6.152024218855979e-06, 'learning_rate/action_model': 6.15202421885598e-05, 'epoch': 0.16}) +[2026-10-03 19:22:54,765][starVLA.training.train_starvla][INFO] - Step 2253, Loss: {'train/loss': 0.08496367186307907, 'train/grad_norm_pre_clip': 41.17396926879883, 'timing/data': 0.00028018999728374183, 'timing/model': 0.597822759009432, 'learning_rate/qwen_vl_interface': 6.149034028321583e-06, 'learning_rate/action_model': 6.149034028321583e-05, 'epoch': 0.16}) +[2026-10-03 19:22:55,469][starVLA.training.train_starvla][INFO] - Step 2254, Loss: {'train/loss': 0.08267050236463547, 'train/grad_norm_pre_clip': 48.727962493896484, 'timing/data': 0.00038263999158516526, 'timing/model': 0.3921746500127483, 'learning_rate/qwen_vl_interface': 6.1460434682282566e-06, 'learning_rate/action_model': 6.146043468228256e-05, 'epoch': 0.16}) +[2026-10-03 19:22:56,169][starVLA.training.train_starvla][INFO] - Step 2255, Loss: {'train/loss': 0.002039553364738822, 'train/grad_norm_pre_clip': 5.078982830047607, 'timing/data': 0.00020106000010855496, 'timing/model': 0.391604489996098, 'learning_rate/qwen_vl_interface': 6.143052539805308e-06, 'learning_rate/action_model': 6.143052539805308e-05, 'epoch': 0.17}) +[2026-10-03 19:22:56,844][starVLA.training.train_starvla][INFO] - Step 2256, Loss: {'train/loss': 0.028243103995919228, 'train/grad_norm_pre_clip': 33.92829513549805, 'timing/data': 0.00029240999720059335, 'timing/model': 0.3657541229913477, 'learning_rate/qwen_vl_interface': 6.140061244282194e-06, 'learning_rate/action_model': 6.140061244282194e-05, 'epoch': 0.17}) +[2026-10-03 19:22:57,546][starVLA.training.train_starvla][INFO] - Step 2257, Loss: {'train/loss': 0.006824810989201069, 'train/grad_norm_pre_clip': 6.537906169891357, 'timing/data': 0.00033867001184262335, 'timing/model': 0.39312616002280265, 'learning_rate/qwen_vl_interface': 6.137069582888521e-06, 'learning_rate/action_model': 6.137069582888521e-05, 'epoch': 0.17}) +[2026-10-03 19:22:58,450][starVLA.training.train_starvla][INFO] - Step 2258, Loss: {'train/loss': 0.2886798679828644, 'train/grad_norm_pre_clip': 60.710601806640625, 'timing/data': 0.00041418999899178743, 'timing/model': 0.5957783290068619, 'learning_rate/qwen_vl_interface': 6.13407755685405e-06, 'learning_rate/action_model': 6.13407755685405e-05, 'epoch': 0.17}) +[2026-10-03 19:22:59,150][starVLA.training.train_starvla][INFO] - Step 2259, Loss: {'train/loss': 0.20529992878437042, 'train/grad_norm_pre_clip': 13.076125144958496, 'timing/data': 0.00020308000966906548, 'timing/model': 0.39196148098562844, 'learning_rate/qwen_vl_interface': 6.131085167408692e-06, 'learning_rate/action_model': 6.131085167408692e-05, 'epoch': 0.17}) +[2026-10-03 19:22:59,850][starVLA.training.train_starvla][INFO] - Step 2260, Loss: {'train/loss': 0.054083678871393204, 'train/grad_norm_pre_clip': 37.001644134521484, 'timing/data': 0.00022019000607542694, 'timing/model': 0.39127313101198524, 'learning_rate/qwen_vl_interface': 6.128092415782502e-06, 'learning_rate/action_model': 6.128092415782501e-05, 'epoch': 0.17}) +[2026-10-03 19:23:00,551][starVLA.training.train_starvla][INFO] - Step 2261, Loss: {'train/loss': 0.0018141858745366335, 'train/grad_norm_pre_clip': 0.37811943888664246, 'timing/data': 0.0003151500131934881, 'timing/model': 0.39257991101476364, 'learning_rate/qwen_vl_interface': 6.125099303205688e-06, 'learning_rate/action_model': 6.125099303205688e-05, 'epoch': 0.17}) +[2026-10-03 19:23:01,249][starVLA.training.train_starvla][INFO] - Step 2262, Loss: {'train/loss': 0.0032493118196725845, 'train/grad_norm_pre_clip': 4.023082256317139, 'timing/data': 0.000384790007956326, 'timing/model': 0.3914323110075202, 'learning_rate/qwen_vl_interface': 6.122105830908607e-06, 'learning_rate/action_model': 6.122105830908606e-05, 'epoch': 0.17}) +[2026-10-03 19:23:02,152][starVLA.training.train_starvla][INFO] - Step 2263, Loss: {'train/loss': 0.8843556046485901, 'train/grad_norm_pre_clip': 261.2634582519531, 'timing/data': 0.00020529000903479755, 'timing/model': 0.595562369999243, 'learning_rate/qwen_vl_interface': 6.11911200012176e-06, 'learning_rate/action_model': 6.11911200012176e-05, 'epoch': 0.17}) +[2026-10-03 19:23:02,850][starVLA.training.train_starvla][INFO] - Step 2264, Loss: {'train/loss': 0.8109616637229919, 'train/grad_norm_pre_clip': 157.0209197998047, 'timing/data': 0.0002114000089932233, 'timing/model': 0.3896809100115206, 'learning_rate/qwen_vl_interface': 6.116117812075799e-06, 'learning_rate/action_model': 6.116117812075799e-05, 'epoch': 0.17}) +[2026-10-03 19:23:03,549][starVLA.training.train_starvla][INFO] - Step 2265, Loss: {'train/loss': 0.2521176338195801, 'train/grad_norm_pre_clip': 50.794124603271484, 'timing/data': 0.00022825002088211477, 'timing/model': 0.391321450995747, 'learning_rate/qwen_vl_interface': 6.113123268001524e-06, 'learning_rate/action_model': 6.113123268001524e-05, 'epoch': 0.17}) +[2026-10-03 19:23:04,247][starVLA.training.train_starvla][INFO] - Step 2266, Loss: {'train/loss': 0.00018059037392959, 'train/grad_norm_pre_clip': 0.05736785754561424, 'timing/data': 0.0002297799801453948, 'timing/model': 0.391090909979539, 'learning_rate/qwen_vl_interface': 6.110128369129873e-06, 'learning_rate/action_model': 6.110128369129873e-05, 'epoch': 0.17}) +[2026-10-03 19:23:04,946][starVLA.training.train_starvla][INFO] - Step 2267, Loss: {'train/loss': 0.08354885131120682, 'train/grad_norm_pre_clip': 50.695220947265625, 'timing/data': 0.00024376000510528684, 'timing/model': 0.39147343998774886, 'learning_rate/qwen_vl_interface': 6.107133116691943e-06, 'learning_rate/action_model': 6.107133116691942e-05, 'epoch': 0.17}) +[2026-10-03 19:23:05,643][starVLA.training.train_starvla][INFO] - Step 2268, Loss: {'train/loss': 0.0008548162295483053, 'train/grad_norm_pre_clip': 0.8622835874557495, 'timing/data': 0.00027508000493980944, 'timing/model': 0.38981088998843916, 'learning_rate/qwen_vl_interface': 6.104137511918963e-06, 'learning_rate/action_model': 6.104137511918963e-05, 'epoch': 0.17}) +[2026-10-03 19:23:06,544][starVLA.training.train_starvla][INFO] - Step 2269, Loss: {'train/loss': 0.00025012536207214, 'train/grad_norm_pre_clip': 0.10642226785421371, 'timing/data': 0.0003126199881080538, 'timing/model': 0.3902767699910328, 'learning_rate/qwen_vl_interface': 6.1011415560423155e-06, 'learning_rate/action_model': 6.101141556042316e-05, 'epoch': 0.17}) +[2026-10-03 19:23:07,241][starVLA.training.train_starvla][INFO] - Step 2270, Loss: {'train/loss': 0.07881820201873779, 'train/grad_norm_pre_clip': 110.01090240478516, 'timing/data': 0.0003879499854519963, 'timing/model': 0.3908978599938564, 'learning_rate/qwen_vl_interface': 6.098145250293525e-06, 'learning_rate/action_model': 6.0981452502935246e-05, 'epoch': 0.17}) +[2026-10-03 19:23:07,914][starVLA.training.train_starvla][INFO] - Step 2271, Loss: {'train/loss': 0.051432542502880096, 'train/grad_norm_pre_clip': 42.787391662597656, 'timing/data': 0.0002813900064211339, 'timing/model': 0.3896505099837668, 'learning_rate/qwen_vl_interface': 6.09514859590426e-06, 'learning_rate/action_model': 6.09514859590426e-05, 'epoch': 0.17}) +[2026-10-03 19:23:08,613][starVLA.training.train_starvla][INFO] - Step 2272, Loss: {'train/loss': 0.2551420032978058, 'train/grad_norm_pre_clip': 36.98834991455078, 'timing/data': 0.00030030999914743006, 'timing/model': 0.39136303099803627, 'learning_rate/qwen_vl_interface': 6.09215159410633e-06, 'learning_rate/action_model': 6.09215159410633e-05, 'epoch': 0.17}) +[2026-10-03 19:23:09,312][starVLA.training.train_starvla][INFO] - Step 2273, Loss: {'train/loss': 0.0005047554150223732, 'train/grad_norm_pre_clip': 0.20678870379924774, 'timing/data': 0.00021748998551629484, 'timing/model': 0.39169990000664257, 'learning_rate/qwen_vl_interface': 6.089154246131692e-06, 'learning_rate/action_model': 6.089154246131692e-05, 'epoch': 0.17}) +[2026-10-03 19:23:10,214][starVLA.training.train_starvla][INFO] - Step 2274, Loss: {'train/loss': 0.003892134176567197, 'train/grad_norm_pre_clip': 0.7315009236335754, 'timing/data': 0.0003068499790970236, 'timing/model': 0.5934720399964135, 'learning_rate/qwen_vl_interface': 6.08615655321244e-06, 'learning_rate/action_model': 6.08615655321244e-05, 'epoch': 0.17}) +[2026-10-03 19:23:10,912][starVLA.training.train_starvla][INFO] - Step 2275, Loss: {'train/loss': 0.00046340550761669874, 'train/grad_norm_pre_clip': 0.31947097182273865, 'timing/data': 0.00024203999782912433, 'timing/model': 0.3913760800205637, 'learning_rate/qwen_vl_interface': 6.083158516580814e-06, 'learning_rate/action_model': 6.083158516580814e-05, 'epoch': 0.17}) +[2026-10-03 19:23:11,612][starVLA.training.train_starvla][INFO] - Step 2276, Loss: {'train/loss': 0.11145684123039246, 'train/grad_norm_pre_clip': 44.2640495300293, 'timing/data': 0.00021579998428933322, 'timing/model': 0.39159841000218876, 'learning_rate/qwen_vl_interface': 6.080160137469192e-06, 'learning_rate/action_model': 6.0801601374691916e-05, 'epoch': 0.17}) +[2026-10-03 19:23:12,313][starVLA.training.train_starvla][INFO] - Step 2277, Loss: {'train/loss': 0.0006915905978530645, 'train/grad_norm_pre_clip': 0.4472350776195526, 'timing/data': 0.0003091600083280355, 'timing/model': 0.3927576999994926, 'learning_rate/qwen_vl_interface': 6.077161417110097e-06, 'learning_rate/action_model': 6.0771614171100965e-05, 'epoch': 0.17}) +[2026-10-03 19:23:13,014][starVLA.training.train_starvla][INFO] - Step 2278, Loss: {'train/loss': 1.0298717021942139, 'train/grad_norm_pre_clip': 172.19590759277344, 'timing/data': 0.00034525999217294157, 'timing/model': 0.39415079000173137, 'learning_rate/qwen_vl_interface': 6.0741623567361864e-06, 'learning_rate/action_model': 6.074162356736186e-05, 'epoch': 0.17}) +[2026-10-03 19:23:13,922][starVLA.training.train_starvla][INFO] - Step 2279, Loss: {'train/loss': 0.13260991871356964, 'train/grad_norm_pre_clip': 70.29175567626953, 'timing/data': 0.00023833999875932932, 'timing/model': 0.5992355390044395, 'learning_rate/qwen_vl_interface': 6.071162957580262e-06, 'learning_rate/action_model': 6.071162957580262e-05, 'epoch': 0.17}) +[2026-10-03 19:23:14,622][starVLA.training.train_starvla][INFO] - Step 2280, Loss: {'train/loss': 0.3023912310600281, 'train/grad_norm_pre_clip': 91.20916748046875, 'timing/data': 0.00024356998619623482, 'timing/model': 0.39133934999699704, 'learning_rate/qwen_vl_interface': 6.068163220875266e-06, 'learning_rate/action_model': 6.068163220875266e-05, 'epoch': 0.17}) +[2026-10-03 19:23:15,320][starVLA.training.train_starvla][INFO] - Step 2281, Loss: {'train/loss': 0.2531423270702362, 'train/grad_norm_pre_clip': 71.71464538574219, 'timing/data': 0.00031622001552022994, 'timing/model': 0.3914744599896949, 'learning_rate/qwen_vl_interface': 6.065163147854273e-06, 'learning_rate/action_model': 6.065163147854273e-05, 'epoch': 0.17}) +[2026-10-03 19:23:16,020][starVLA.training.train_starvla][INFO] - Step 2282, Loss: {'train/loss': 4.545280535239726e-05, 'train/grad_norm_pre_clip': 0.015162362717092037, 'timing/data': 0.00020748001406900585, 'timing/model': 0.39179628097917885, 'learning_rate/qwen_vl_interface': 6.062162739750502e-06, 'learning_rate/action_model': 6.062162739750502e-05, 'epoch': 0.17}) +[2026-10-03 19:23:16,716][starVLA.training.train_starvla][INFO] - Step 2283, Loss: {'train/loss': 0.0732371062040329, 'train/grad_norm_pre_clip': 46.29690933227539, 'timing/data': 0.0002982000005431473, 'timing/model': 0.39025645999936387, 'learning_rate/qwen_vl_interface': 6.059161997797306e-06, 'learning_rate/action_model': 6.0591619977973056e-05, 'epoch': 0.17}) +[2026-10-03 19:23:17,417][starVLA.training.train_starvla][INFO] - Step 2284, Loss: {'train/loss': 0.2663005590438843, 'train/grad_norm_pre_clip': 98.16339111328125, 'timing/data': 0.00028060999466106296, 'timing/model': 0.3911136499955319, 'learning_rate/qwen_vl_interface': 6.056160923228178e-06, 'learning_rate/action_model': 6.056160923228178e-05, 'epoch': 0.17}) +[2026-10-03 19:23:18,318][starVLA.training.train_starvla][INFO] - Step 2285, Loss: {'train/loss': 0.0014811739092692733, 'train/grad_norm_pre_clip': 0.45577272772789, 'timing/data': 0.00035662000300362706, 'timing/model': 0.39145653098239563, 'learning_rate/qwen_vl_interface': 6.0531595172767455e-06, 'learning_rate/action_model': 6.0531595172767455e-05, 'epoch': 0.17}) +[2026-10-03 19:23:19,019][starVLA.training.train_starvla][INFO] - Step 2286, Loss: {'train/loss': 0.00010204524733126163, 'train/grad_norm_pre_clip': 0.029058562591671944, 'timing/data': 0.00038865001988597214, 'timing/model': 0.39166418099193834, 'learning_rate/qwen_vl_interface': 6.050157781176774e-06, 'learning_rate/action_model': 6.050157781176774e-05, 'epoch': 0.17}) +[2026-10-03 19:23:19,717][starVLA.training.train_starvla][INFO] - Step 2287, Loss: {'train/loss': 0.14898928999900818, 'train/grad_norm_pre_clip': 107.06182861328125, 'timing/data': 0.00021684999228455126, 'timing/model': 0.3904702009749599, 'learning_rate/qwen_vl_interface': 6.047155716162164e-06, 'learning_rate/action_model': 6.047155716162164e-05, 'epoch': 0.17}) +[2026-10-03 19:23:20,415][starVLA.training.train_starvla][INFO] - Step 2288, Loss: {'train/loss': 0.057099565863609314, 'train/grad_norm_pre_clip': 43.95812225341797, 'timing/data': 0.0003834000090137124, 'timing/model': 0.3897772609780077, 'learning_rate/qwen_vl_interface': 6.044153323466948e-06, 'learning_rate/action_model': 6.044153323466948e-05, 'epoch': 0.17}) +[2026-10-03 19:23:21,117][starVLA.training.train_starvla][INFO] - Step 2289, Loss: {'train/loss': 4.0140701457858086e-05, 'train/grad_norm_pre_clip': 0.012027682736515999, 'timing/data': 0.0003555100120138377, 'timing/model': 0.3928092299902346, 'learning_rate/qwen_vl_interface': 6.0411506043253e-06, 'learning_rate/action_model': 6.0411506043253e-05, 'epoch': 0.17}) +[2026-10-03 19:23:22,020][starVLA.training.train_starvla][INFO] - Step 2290, Loss: {'train/loss': 0.5923588871955872, 'train/grad_norm_pre_clip': 185.2301025390625, 'timing/data': 0.0003826300089713186, 'timing/model': 0.39113271099631675, 'learning_rate/qwen_vl_interface': 6.038147559971523e-06, 'learning_rate/action_model': 6.038147559971523e-05, 'epoch': 0.17}) +[2026-10-03 19:23:22,718][starVLA.training.train_starvla][INFO] - Step 2291, Loss: {'train/loss': 0.0002239760651718825, 'train/grad_norm_pre_clip': 0.36014437675476074, 'timing/data': 0.00022605000413022935, 'timing/model': 0.39197235999745317, 'learning_rate/qwen_vl_interface': 6.035144191640055e-06, 'learning_rate/action_model': 6.035144191640055e-05, 'epoch': 0.17}) +[2026-10-03 19:23:23,418][starVLA.training.train_starvla][INFO] - Step 2292, Loss: {'train/loss': 0.0001434045407222584, 'train/grad_norm_pre_clip': 0.07237118482589722, 'timing/data': 0.00033817999064922333, 'timing/model': 0.3915632799908053, 'learning_rate/qwen_vl_interface': 6.032140500565467e-06, 'learning_rate/action_model': 6.0321405005654665e-05, 'epoch': 0.17}) +[2026-10-03 19:23:24,118][starVLA.training.train_starvla][INFO] - Step 2293, Loss: {'train/loss': 0.13766871392726898, 'train/grad_norm_pre_clip': 92.26191711425781, 'timing/data': 0.000361250014975667, 'timing/model': 0.3928544209920801, 'learning_rate/qwen_vl_interface': 6.029136487982463e-06, 'learning_rate/action_model': 6.029136487982463e-05, 'epoch': 0.17}) +[2026-10-03 19:23:24,817][starVLA.training.train_starvla][INFO] - Step 2294, Loss: {'train/loss': 0.08159592002630234, 'train/grad_norm_pre_clip': 46.6771125793457, 'timing/data': 0.00029166997410357, 'timing/model': 0.39077344100223854, 'learning_rate/qwen_vl_interface': 6.02613215512588e-06, 'learning_rate/action_model': 6.0261321551258796e-05, 'epoch': 0.17}) +[2026-10-03 19:23:25,718][starVLA.training.train_starvla][INFO] - Step 2295, Loss: {'train/loss': 0.04129428043961525, 'train/grad_norm_pre_clip': 32.63792037963867, 'timing/data': 0.00019463998614810407, 'timing/model': 0.5926620800164528, 'learning_rate/qwen_vl_interface': 6.023127503230684e-06, 'learning_rate/action_model': 6.023127503230683e-05, 'epoch': 0.17}) +[2026-10-03 19:23:26,416][starVLA.training.train_starvla][INFO] - Step 2296, Loss: {'train/loss': 0.37935706973075867, 'train/grad_norm_pre_clip': 94.30693054199219, 'timing/data': 0.00020702998153865337, 'timing/model': 0.39055525002186187, 'learning_rate/qwen_vl_interface': 6.020122533531976e-06, 'learning_rate/action_model': 6.0201225335319755e-05, 'epoch': 0.17}) +[2026-10-03 19:23:27,116][starVLA.training.train_starvla][INFO] - Step 2297, Loss: {'train/loss': 0.0560830794274807, 'train/grad_norm_pre_clip': 85.34402465820312, 'timing/data': 0.00020465999841690063, 'timing/model': 0.3928998000046704, 'learning_rate/qwen_vl_interface': 6.017117247264983e-06, 'learning_rate/action_model': 6.017117247264983e-05, 'epoch': 0.17}) +[2026-10-03 19:23:27,817][starVLA.training.train_starvla][INFO] - Step 2298, Loss: {'train/loss': 0.001197519013658166, 'train/grad_norm_pre_clip': 1.122632384300232, 'timing/data': 0.00020324901561252773, 'timing/model': 0.3927077610278502, 'learning_rate/qwen_vl_interface': 6.014111645665069e-06, 'learning_rate/action_model': 6.0141116456650684e-05, 'epoch': 0.17}) +[2026-10-03 19:23:28,513][starVLA.training.train_starvla][INFO] - Step 2299, Loss: {'train/loss': 0.000156241818331182, 'train/grad_norm_pre_clip': 0.06067151203751564, 'timing/data': 0.0001976999919861555, 'timing/model': 0.3908454799966421, 'learning_rate/qwen_vl_interface': 6.011105729967719e-06, 'learning_rate/action_model': 6.0111057299677184e-05, 'epoch': 0.17}) +[2026-10-03 19:23:29,403][starVLA.training.train_starvla][INFO] - Step 2300, Loss: {'train/loss': 0.30768445134162903, 'train/grad_norm_pre_clip': 48.3348274230957, 'timing/data': 0.0002467499871272594, 'timing/model': 0.5812779609987047, 'learning_rate/qwen_vl_interface': 6.008099501408552e-06, 'learning_rate/action_model': 6.008099501408552e-05, 'epoch': 0.17}) +[2026-10-03 19:23:30,103][starVLA.training.train_starvla][INFO] - Step 2301, Loss: {'train/loss': 0.008349570445716381, 'train/grad_norm_pre_clip': 21.027442932128906, 'timing/data': 0.00021026001195423305, 'timing/model': 0.3924376099894289, 'learning_rate/qwen_vl_interface': 6.0050929612233165e-06, 'learning_rate/action_model': 6.005092961223316e-05, 'epoch': 0.17}) +[2026-10-03 19:23:30,799][starVLA.training.train_starvla][INFO] - Step 2302, Loss: {'train/loss': 0.4104219675064087, 'train/grad_norm_pre_clip': 41.524234771728516, 'timing/data': 0.0003036199777852744, 'timing/model': 0.3901272110233549, 'learning_rate/qwen_vl_interface': 6.002086110647888e-06, 'learning_rate/action_model': 6.002086110647888e-05, 'epoch': 0.17}) +[2026-10-03 19:23:31,495][starVLA.training.train_starvla][INFO] - Step 2303, Loss: {'train/loss': 6.985482468735427e-05, 'train/grad_norm_pre_clip': 0.04029575362801552, 'timing/data': 0.00019764000899158418, 'timing/model': 0.3886527310241945, 'learning_rate/qwen_vl_interface': 5.999078950918267e-06, 'learning_rate/action_model': 5.999078950918267e-05, 'epoch': 0.17}) +[2026-10-03 19:23:32,194][starVLA.training.train_starvla][INFO] - Step 2304, Loss: {'train/loss': 0.022503256797790527, 'train/grad_norm_pre_clip': 18.63116455078125, 'timing/data': 0.0002030600153375417, 'timing/model': 0.3899356609908864, 'learning_rate/qwen_vl_interface': 5.996071483270585e-06, 'learning_rate/action_model': 5.996071483270585e-05, 'epoch': 0.17}) +[2026-10-03 19:23:32,890][starVLA.training.train_starvla][INFO] - Step 2305, Loss: {'train/loss': 0.6124982237815857, 'train/grad_norm_pre_clip': 113.94084167480469, 'timing/data': 0.00020334002329036593, 'timing/model': 0.39086676097940654, 'learning_rate/qwen_vl_interface': 5.993063708941098e-06, 'learning_rate/action_model': 5.993063708941098e-05, 'epoch': 0.17}) +[2026-10-03 19:23:33,790][starVLA.training.train_starvla][INFO] - Step 2306, Loss: {'train/loss': 0.000522680813446641, 'train/grad_norm_pre_clip': 0.19853492081165314, 'timing/data': 0.00024061999283730984, 'timing/model': 0.3902808799757622, 'learning_rate/qwen_vl_interface': 5.990055629166187e-06, 'learning_rate/action_model': 5.990055629166187e-05, 'epoch': 0.17}) +[2026-10-03 19:23:34,490][starVLA.training.train_starvla][INFO] - Step 2307, Loss: {'train/loss': 0.490030437707901, 'train/grad_norm_pre_clip': 43.8463020324707, 'timing/data': 0.0002072000061161816, 'timing/model': 0.39130678100627847, 'learning_rate/qwen_vl_interface': 5.987047245182363e-06, 'learning_rate/action_model': 5.987047245182362e-05, 'epoch': 0.17}) +[2026-10-03 19:23:35,189][starVLA.training.train_starvla][INFO] - Step 2308, Loss: {'train/loss': 0.0009280047961510718, 'train/grad_norm_pre_clip': 0.7206370234489441, 'timing/data': 0.00037204002728685737, 'timing/model': 0.38892503100214526, 'learning_rate/qwen_vl_interface': 5.984038558226255e-06, 'learning_rate/action_model': 5.9840385582262546e-05, 'epoch': 0.17}) +[2026-10-03 19:23:35,888][starVLA.training.train_starvla][INFO] - Step 2309, Loss: {'train/loss': 0.22870729863643646, 'train/grad_norm_pre_clip': 56.93632125854492, 'timing/data': 0.0002274699800182134, 'timing/model': 0.39152159000514075, 'learning_rate/qwen_vl_interface': 5.981029569534625e-06, 'learning_rate/action_model': 5.981029569534625e-05, 'epoch': 0.17}) +[2026-10-03 19:23:36,588][starVLA.training.train_starvla][INFO] - Step 2310, Loss: {'train/loss': 0.2074805647134781, 'train/grad_norm_pre_clip': 253.6063690185547, 'timing/data': 0.0003897499991580844, 'timing/model': 0.3914484000124503, 'learning_rate/qwen_vl_interface': 5.978020280344351e-06, 'learning_rate/action_model': 5.9780202803443506e-05, 'epoch': 0.17}) +[2026-10-03 19:23:37,493][starVLA.training.train_starvla][INFO] - Step 2311, Loss: {'train/loss': 0.5196331143379211, 'train/grad_norm_pre_clip': 56.43751525878906, 'timing/data': 0.00020866998238489032, 'timing/model': 0.39230639100424014, 'learning_rate/qwen_vl_interface': 5.97501069189244e-06, 'learning_rate/action_model': 5.97501069189244e-05, 'epoch': 0.17}) +[2026-10-03 19:23:38,192][starVLA.training.train_starvla][INFO] - Step 2312, Loss: {'train/loss': 0.00493835611268878, 'train/grad_norm_pre_clip': 1.3307644128799438, 'timing/data': 0.00021562998881563544, 'timing/model': 0.39107770001282915, 'learning_rate/qwen_vl_interface': 5.97200080541602e-06, 'learning_rate/action_model': 5.97200080541602e-05, 'epoch': 0.17}) +[2026-10-03 19:23:38,892][starVLA.training.train_starvla][INFO] - Step 2313, Loss: {'train/loss': 0.19740505516529083, 'train/grad_norm_pre_clip': 28.893327713012695, 'timing/data': 0.00021133999689482152, 'timing/model': 0.39254026001435705, 'learning_rate/qwen_vl_interface': 5.968990622152341e-06, 'learning_rate/action_model': 5.9689906221523404e-05, 'epoch': 0.17}) +[2026-10-03 19:23:39,589][starVLA.training.train_starvla][INFO] - Step 2314, Loss: {'train/loss': 0.000740922405384481, 'train/grad_norm_pre_clip': 0.333311527967453, 'timing/data': 0.0002075600204989314, 'timing/model': 0.3903582900238689, 'learning_rate/qwen_vl_interface': 5.965980143338774e-06, 'learning_rate/action_model': 5.965980143338774e-05, 'epoch': 0.17}) +[2026-10-03 19:23:40,288][starVLA.training.train_starvla][INFO] - Step 2315, Loss: {'train/loss': 0.0007363994373008609, 'train/grad_norm_pre_clip': 0.35875844955444336, 'timing/data': 0.00020201000734232366, 'timing/model': 0.3917578200052958, 'learning_rate/qwen_vl_interface': 5.962969370212819e-06, 'learning_rate/action_model': 5.9629693702128185e-05, 'epoch': 0.17}) +[2026-10-03 19:23:41,191][starVLA.training.train_starvla][INFO] - Step 2316, Loss: {'train/loss': 0.03261430189013481, 'train/grad_norm_pre_clip': 50.131439208984375, 'timing/data': 0.00021373899653553963, 'timing/model': 0.5954331699758768, 'learning_rate/qwen_vl_interface': 5.959958304012085e-06, 'learning_rate/action_model': 5.9599583040120845e-05, 'epoch': 0.17}) +[2026-10-03 19:23:41,890][starVLA.training.train_starvla][INFO] - Step 2317, Loss: {'train/loss': 0.0523885078728199, 'train/grad_norm_pre_clip': 65.50625610351562, 'timing/data': 0.00021013998775742948, 'timing/model': 0.3914342600037344, 'learning_rate/qwen_vl_interface': 5.956946945974309e-06, 'learning_rate/action_model': 5.9569469459743085e-05, 'epoch': 0.17}) +[2026-10-03 19:23:42,589][starVLA.training.train_starvla][INFO] - Step 2318, Loss: {'train/loss': 0.0003005427133757621, 'train/grad_norm_pre_clip': 0.13537509739398956, 'timing/data': 0.0002050299954134971, 'timing/model': 0.39006225098273717, 'learning_rate/qwen_vl_interface': 5.953935297337347e-06, 'learning_rate/action_model': 5.953935297337347e-05, 'epoch': 0.17}) +[2026-10-03 19:23:43,285][starVLA.training.train_starvla][INFO] - Step 2319, Loss: {'train/loss': 0.10509320348501205, 'train/grad_norm_pre_clip': 170.25173950195312, 'timing/data': 0.00020201998995617032, 'timing/model': 0.38977049000095576, 'learning_rate/qwen_vl_interface': 5.950923359339175e-06, 'learning_rate/action_model': 5.950923359339174e-05, 'epoch': 0.17}) +[2026-10-03 19:23:43,987][starVLA.training.train_starvla][INFO] - Step 2320, Loss: {'train/loss': 0.5516176223754883, 'train/grad_norm_pre_clip': 70.5721206665039, 'timing/data': 0.00021359999664127827, 'timing/model': 0.39229812001576647, 'learning_rate/qwen_vl_interface': 5.947911133217886e-06, 'learning_rate/action_model': 5.947911133217886e-05, 'epoch': 0.17}) +[2026-10-03 19:23:44,898][starVLA.training.train_starvla][INFO] - Step 2321, Loss: {'train/loss': 0.36544111371040344, 'train/grad_norm_pre_clip': 47.09711456298828, 'timing/data': 0.00021033998928032815, 'timing/model': 0.6027027789968997, 'learning_rate/qwen_vl_interface': 5.944898620211693e-06, 'learning_rate/action_model': 5.944898620211693e-05, 'epoch': 0.17}) +[2026-10-03 19:23:45,595][starVLA.training.train_starvla][INFO] - Step 2322, Loss: {'train/loss': 0.21860970556735992, 'train/grad_norm_pre_clip': 100.52049255371094, 'timing/data': 0.00020638000569306314, 'timing/model': 0.3902064400026575, 'learning_rate/qwen_vl_interface': 5.941885821558925e-06, 'learning_rate/action_model': 5.9418858215589245e-05, 'epoch': 0.17}) +[2026-10-03 19:23:46,292][starVLA.training.train_starvla][INFO] - Step 2323, Loss: {'train/loss': 0.0008349508279934525, 'train/grad_norm_pre_clip': 0.5519665479660034, 'timing/data': 0.00033618000452406704, 'timing/model': 0.39001113100675866, 'learning_rate/qwen_vl_interface': 5.938872738498031e-06, 'learning_rate/action_model': 5.93887273849803e-05, 'epoch': 0.17}) +[2026-10-03 19:23:46,991][starVLA.training.train_starvla][INFO] - Step 2324, Loss: {'train/loss': 0.000891669187694788, 'train/grad_norm_pre_clip': 0.23618441820144653, 'timing/data': 0.00022021998302079737, 'timing/model': 0.39005375001579523, 'learning_rate/qwen_vl_interface': 5.935859372267575e-06, 'learning_rate/action_model': 5.935859372267575e-05, 'epoch': 0.17}) +[2026-10-03 19:23:47,693][starVLA.training.train_starvla][INFO] - Step 2325, Loss: {'train/loss': 0.039741016924381256, 'train/grad_norm_pre_clip': 42.331119537353516, 'timing/data': 0.00027007999597117305, 'timing/model': 0.39293024002108723, 'learning_rate/qwen_vl_interface': 5.9328457241062396e-06, 'learning_rate/action_model': 5.9328457241062394e-05, 'epoch': 0.17}) +[2026-10-03 19:23:48,390][starVLA.training.train_starvla][INFO] - Step 2326, Loss: {'train/loss': 0.016268257051706314, 'train/grad_norm_pre_clip': 15.027263641357422, 'timing/data': 0.0002092400100082159, 'timing/model': 0.3909744299890008, 'learning_rate/qwen_vl_interface': 5.929831795252819e-06, 'learning_rate/action_model': 5.9298317952528184e-05, 'epoch': 0.17}) +[2026-10-03 19:23:49,297][starVLA.training.train_starvla][INFO] - Step 2327, Loss: {'train/loss': 0.0003785750886891037, 'train/grad_norm_pre_clip': 0.12649071216583252, 'timing/data': 0.0002171300002373755, 'timing/model': 0.3910077500040643, 'learning_rate/qwen_vl_interface': 5.926817586946227e-06, 'learning_rate/action_model': 5.926817586946227e-05, 'epoch': 0.17}) +[2026-10-03 19:23:50,004][starVLA.training.train_starvla][INFO] - Step 2328, Loss: {'train/loss': 0.22434966266155243, 'train/grad_norm_pre_clip': 73.0003662109375, 'timing/data': 0.00022991999867372215, 'timing/model': 0.39350900999852456, 'learning_rate/qwen_vl_interface': 5.923803100425491e-06, 'learning_rate/action_model': 5.92380310042549e-05, 'epoch': 0.17}) +[2026-10-03 19:23:50,711][starVLA.training.train_starvla][INFO] - Step 2329, Loss: {'train/loss': 0.0011577795958146453, 'train/grad_norm_pre_clip': 1.4438661336898804, 'timing/data': 0.00022504001390188932, 'timing/model': 0.3951620099833235, 'learning_rate/qwen_vl_interface': 5.92078833692975e-06, 'learning_rate/action_model': 5.92078833692975e-05, 'epoch': 0.17}) +[2026-10-03 19:23:51,416][starVLA.training.train_starvla][INFO] - Step 2330, Loss: {'train/loss': 0.009728691540658474, 'train/grad_norm_pre_clip': 5.671268463134766, 'timing/data': 0.0003422199806664139, 'timing/model': 0.3946565490041394, 'learning_rate/qwen_vl_interface': 5.917773297698264e-06, 'learning_rate/action_model': 5.9177732976982636e-05, 'epoch': 0.17}) +[2026-10-03 19:23:52,123][starVLA.training.train_starvla][INFO] - Step 2331, Loss: {'train/loss': 0.23786067962646484, 'train/grad_norm_pre_clip': 103.85784149169922, 'timing/data': 0.00022792001254856586, 'timing/model': 0.394448420003755, 'learning_rate/qwen_vl_interface': 5.9147579839704e-06, 'learning_rate/action_model': 5.9147579839704e-05, 'epoch': 0.17}) +[2026-10-03 19:23:53,028][starVLA.training.train_starvla][INFO] - Step 2332, Loss: {'train/loss': 0.00032201287103816867, 'train/grad_norm_pre_clip': 0.05933203548192978, 'timing/data': 0.0003461400046944618, 'timing/model': 0.393351859995164, 'learning_rate/qwen_vl_interface': 5.911742396985637e-06, 'learning_rate/action_model': 5.9117423969856364e-05, 'epoch': 0.17}) +[2026-10-03 19:23:53,734][starVLA.training.train_starvla][INFO] - Step 2333, Loss: {'train/loss': 0.0004685093299485743, 'train/grad_norm_pre_clip': 0.2972160577774048, 'timing/data': 0.00022601999808102846, 'timing/model': 0.39439823999418877, 'learning_rate/qwen_vl_interface': 5.908726537983571e-06, 'learning_rate/action_model': 5.908726537983571e-05, 'epoch': 0.17}) +[2026-10-03 19:23:54,441][starVLA.training.train_starvla][INFO] - Step 2334, Loss: {'train/loss': 0.06426525115966797, 'train/grad_norm_pre_clip': 59.6877326965332, 'timing/data': 0.0003664499963633716, 'timing/model': 0.39455807098420337, 'learning_rate/qwen_vl_interface': 5.9057104082039085e-06, 'learning_rate/action_model': 5.905710408203908e-05, 'epoch': 0.17}) +[2026-10-03 19:23:55,146][starVLA.training.train_starvla][INFO] - Step 2335, Loss: {'train/loss': 0.38988393545150757, 'train/grad_norm_pre_clip': 85.55574798583984, 'timing/data': 0.0003266700077801943, 'timing/model': 0.39443618000950664, 'learning_rate/qwen_vl_interface': 5.902694008886466e-06, 'learning_rate/action_model': 5.902694008886466e-05, 'epoch': 0.17}) +[2026-10-03 19:23:55,853][starVLA.training.train_starvla][INFO] - Step 2336, Loss: {'train/loss': 0.02748784050345421, 'train/grad_norm_pre_clip': 14.20176887512207, 'timing/data': 0.0003261299862060696, 'timing/model': 0.39638809001189657, 'learning_rate/qwen_vl_interface': 5.8996773412711694e-06, 'learning_rate/action_model': 5.899677341271169e-05, 'epoch': 0.17}) +[2026-10-03 19:23:56,765][starVLA.training.train_starvla][INFO] - Step 2337, Loss: {'train/loss': 0.0010372252436354756, 'train/grad_norm_pre_clip': 0.693397045135498, 'timing/data': 0.0003663400129880756, 'timing/model': 0.5997125589929055, 'learning_rate/qwen_vl_interface': 5.896660406598059e-06, 'learning_rate/action_model': 5.896660406598059e-05, 'epoch': 0.17}) +[2026-10-03 19:23:57,470][starVLA.training.train_starvla][INFO] - Step 2338, Loss: {'train/loss': 0.030294595286250114, 'train/grad_norm_pre_clip': 16.342374801635742, 'timing/data': 0.0003599699994083494, 'timing/model': 0.3929977600055281, 'learning_rate/qwen_vl_interface': 5.893643206107283e-06, 'learning_rate/action_model': 5.893643206107282e-05, 'epoch': 0.17}) +[2026-10-03 19:23:58,173][starVLA.training.train_starvla][INFO] - Step 2339, Loss: {'train/loss': 0.5836334824562073, 'train/grad_norm_pre_clip': 105.55584716796875, 'timing/data': 0.0003505899803712964, 'timing/model': 0.39285107000614516, 'learning_rate/qwen_vl_interface': 5.890625741039096e-06, 'learning_rate/action_model': 5.890625741039095e-05, 'epoch': 0.17}) +[2026-10-03 19:23:58,878][starVLA.training.train_starvla][INFO] - Step 2340, Loss: {'train/loss': 0.03240064159035683, 'train/grad_norm_pre_clip': 58.090492248535156, 'timing/data': 0.0003537900047376752, 'timing/model': 0.3942951200006064, 'learning_rate/qwen_vl_interface': 5.887608012633864e-06, 'learning_rate/action_model': 5.887608012633864e-05, 'epoch': 0.17}) +[2026-10-03 19:23:59,582][starVLA.training.train_starvla][INFO] - Step 2341, Loss: {'train/loss': 0.03510706499218941, 'train/grad_norm_pre_clip': 31.37120819091797, 'timing/data': 0.0003487899957690388, 'timing/model': 0.39438947997405194, 'learning_rate/qwen_vl_interface': 5.884590022132064e-06, 'learning_rate/action_model': 5.884590022132064e-05, 'epoch': 0.17}) +[2026-10-03 19:24:00,490][starVLA.training.train_starvla][INFO] - Step 2342, Loss: {'train/loss': 0.14849470555782318, 'train/grad_norm_pre_clip': 79.07969665527344, 'timing/data': 0.00022201001411303878, 'timing/model': 0.5958934399823193, 'learning_rate/qwen_vl_interface': 5.881571770774278e-06, 'learning_rate/action_model': 5.881571770774278e-05, 'epoch': 0.17}) +[2026-10-03 19:24:01,194][starVLA.training.train_starvla][INFO] - Step 2343, Loss: {'train/loss': 0.00018595474830362946, 'train/grad_norm_pre_clip': 0.1218293309211731, 'timing/data': 0.00031434997799806297, 'timing/model': 0.3927900999842677, 'learning_rate/qwen_vl_interface': 5.878553259801192e-06, 'learning_rate/action_model': 5.878553259801192e-05, 'epoch': 0.17}) +[2026-10-03 19:24:01,900][starVLA.training.train_starvla][INFO] - Step 2344, Loss: {'train/loss': 2.6806326786754653e-05, 'train/grad_norm_pre_clip': 0.00810870248824358, 'timing/data': 0.000314689998049289, 'timing/model': 0.39445592000265606, 'learning_rate/qwen_vl_interface': 5.8755344904536025e-06, 'learning_rate/action_model': 5.8755344904536025e-05, 'epoch': 0.17}) +[2026-10-03 19:24:02,605][starVLA.training.train_starvla][INFO] - Step 2345, Loss: {'train/loss': 0.16986988484859467, 'train/grad_norm_pre_clip': 170.68954467773438, 'timing/data': 0.0003567200037650764, 'timing/model': 0.3928852409881074, 'learning_rate/qwen_vl_interface': 5.872515463972415e-06, 'learning_rate/action_model': 5.8725154639724146e-05, 'epoch': 0.17}) +[2026-10-03 19:24:03,310][starVLA.training.train_starvla][INFO] - Step 2346, Loss: {'train/loss': 6.924881017766893e-05, 'train/grad_norm_pre_clip': 0.02567126601934433, 'timing/data': 0.0003746699949260801, 'timing/model': 0.39378720000968315, 'learning_rate/qwen_vl_interface': 5.869496181598633e-06, 'learning_rate/action_model': 5.869496181598633e-05, 'epoch': 0.17}) +[2026-10-03 19:24:04,015][starVLA.training.train_starvla][INFO] - Step 2347, Loss: {'train/loss': 0.1178816631436348, 'train/grad_norm_pre_clip': 296.1957092285156, 'timing/data': 0.0003401300054974854, 'timing/model': 0.39405609099776484, 'learning_rate/qwen_vl_interface': 5.866476644573371e-06, 'learning_rate/action_model': 5.866476644573371e-05, 'epoch': 0.17}) +[2026-10-03 19:24:04,922][starVLA.training.train_starvla][INFO] - Step 2348, Loss: {'train/loss': 0.0036854275967925787, 'train/grad_norm_pre_clip': 3.604236364364624, 'timing/data': 0.000305049994494766, 'timing/model': 0.39457900999695994, 'learning_rate/qwen_vl_interface': 5.863456854137848e-06, 'learning_rate/action_model': 5.8634568541378474e-05, 'epoch': 0.17}) +[2026-10-03 19:24:05,627][starVLA.training.train_starvla][INFO] - Step 2349, Loss: {'train/loss': 0.37795865535736084, 'train/grad_norm_pre_clip': 21.98727798461914, 'timing/data': 0.00035188000765629113, 'timing/model': 0.39471605099970475, 'learning_rate/qwen_vl_interface': 5.860436811533383e-06, 'learning_rate/action_model': 5.860436811533383e-05, 'epoch': 0.17}) +[2026-10-03 19:24:06,333][starVLA.training.train_starvla][INFO] - Step 2350, Loss: {'train/loss': 0.6285697221755981, 'train/grad_norm_pre_clip': 207.6068878173828, 'timing/data': 0.00045046000741422176, 'timing/model': 0.395020719995955, 'learning_rate/qwen_vl_interface': 5.857416518001404e-06, 'learning_rate/action_model': 5.857416518001404e-05, 'epoch': 0.17}) +[2026-10-03 19:24:07,038][starVLA.training.train_starvla][INFO] - Step 2351, Loss: {'train/loss': 0.0011785845272243023, 'train/grad_norm_pre_clip': 0.49681875109672546, 'timing/data': 0.0002258600143250078, 'timing/model': 0.3933125000039581, 'learning_rate/qwen_vl_interface': 5.854395974783439e-06, 'learning_rate/action_model': 5.8543959747834386e-05, 'epoch': 0.17}) +[2026-10-03 19:24:07,746][starVLA.training.train_starvla][INFO] - Step 2352, Loss: {'train/loss': 0.00019904487999156117, 'train/grad_norm_pre_clip': 0.13472598791122437, 'timing/data': 0.00029726000502705574, 'timing/model': 0.3953455699956976, 'learning_rate/qwen_vl_interface': 5.851375183121119e-06, 'learning_rate/action_model': 5.8513751831211186e-05, 'epoch': 0.17}) +[2026-10-03 19:24:08,652][starVLA.training.train_starvla][INFO] - Step 2353, Loss: {'train/loss': 0.26298731565475464, 'train/grad_norm_pre_clip': 20.66529655456543, 'timing/data': 0.0002802399976644665, 'timing/model': 0.5945251699886285, 'learning_rate/qwen_vl_interface': 5.848354144256175e-06, 'learning_rate/action_model': 5.8483541442561754e-05, 'epoch': 0.17}) +[2026-10-03 19:24:09,361][starVLA.training.train_starvla][INFO] - Step 2354, Loss: {'train/loss': 0.0018844980513677, 'train/grad_norm_pre_clip': 0.9123480319976807, 'timing/data': 0.00044134000199846923, 'timing/model': 0.3957501290133223, 'learning_rate/qwen_vl_interface': 5.845332859430446e-06, 'learning_rate/action_model': 5.845332859430446e-05, 'epoch': 0.17}) +[2026-10-03 19:24:10,068][starVLA.training.train_starvla][INFO] - Step 2355, Loss: {'train/loss': 0.0007045659003779292, 'train/grad_norm_pre_clip': 0.2909628748893738, 'timing/data': 0.00022493000142276287, 'timing/model': 0.3953299399872776, 'learning_rate/qwen_vl_interface': 5.842311329885866e-06, 'learning_rate/action_model': 5.842311329885866e-05, 'epoch': 0.17}) +[2026-10-03 19:24:10,774][starVLA.training.train_starvla][INFO] - Step 2356, Loss: {'train/loss': 0.19100627303123474, 'train/grad_norm_pre_clip': 20.034208297729492, 'timing/data': 0.00023617001716047525, 'timing/model': 0.3953504810051527, 'learning_rate/qwen_vl_interface': 5.839289556864472e-06, 'learning_rate/action_model': 5.8392895568644714e-05, 'epoch': 0.17}) +[2026-10-03 19:24:11,478][starVLA.training.train_starvla][INFO] - Step 2357, Loss: {'train/loss': 0.002571204211562872, 'train/grad_norm_pre_clip': 1.2631924152374268, 'timing/data': 0.0003669399884529412, 'timing/model': 0.39362864999566227, 'learning_rate/qwen_vl_interface': 5.836267541608398e-06, 'learning_rate/action_model': 5.836267541608398e-05, 'epoch': 0.17}) +[2026-10-03 19:24:12,383][starVLA.training.train_starvla][INFO] - Step 2358, Loss: {'train/loss': 0.0430687740445137, 'train/grad_norm_pre_clip': 51.65287399291992, 'timing/data': 0.00036230002297088504, 'timing/model': 0.5933575200033374, 'learning_rate/qwen_vl_interface': 5.833245285359887e-06, 'learning_rate/action_model': 5.833245285359886e-05, 'epoch': 0.17}) +[2026-10-03 19:24:13,088][starVLA.training.train_starvla][INFO] - Step 2359, Loss: {'train/loss': 0.0015914608957245946, 'train/grad_norm_pre_clip': 1.2263436317443848, 'timing/data': 0.00033599999733269215, 'timing/model': 0.3946394300146494, 'learning_rate/qwen_vl_interface': 5.830222789361268e-06, 'learning_rate/action_model': 5.8302227893612684e-05, 'epoch': 0.17}) +[2026-10-03 19:24:13,794][starVLA.training.train_starvla][INFO] - Step 2360, Loss: {'train/loss': 0.13437271118164062, 'train/grad_norm_pre_clip': 202.5653839111328, 'timing/data': 0.00032603999716229737, 'timing/model': 0.3947679199918639, 'learning_rate/qwen_vl_interface': 5.8272000548549776e-06, 'learning_rate/action_model': 5.827200054854978e-05, 'epoch': 0.17}) +[2026-10-03 19:24:14,498][starVLA.training.train_starvla][INFO] - Step 2361, Loss: {'train/loss': 0.03045796975493431, 'train/grad_norm_pre_clip': 63.38047409057617, 'timing/data': 0.0003691299934871495, 'timing/model': 0.3944076999905519, 'learning_rate/qwen_vl_interface': 5.824177083083547e-06, 'learning_rate/action_model': 5.8241770830835475e-05, 'epoch': 0.17}) +[2026-10-03 19:24:15,203][starVLA.training.train_starvla][INFO] - Step 2362, Loss: {'train/loss': 0.17620046436786652, 'train/grad_norm_pre_clip': 74.28202056884766, 'timing/data': 0.0003722999827004969, 'timing/model': 0.3925485000072513, 'learning_rate/qwen_vl_interface': 5.8211538752896056e-06, 'learning_rate/action_model': 5.821153875289606e-05, 'epoch': 0.17}) +[2026-10-03 19:24:15,909][starVLA.training.train_starvla][INFO] - Step 2363, Loss: {'train/loss': 0.0001858184114098549, 'train/grad_norm_pre_clip': 0.048140112310647964, 'timing/data': 0.000235140003496781, 'timing/model': 0.39440739998826757, 'learning_rate/qwen_vl_interface': 5.8181304327158814e-06, 'learning_rate/action_model': 5.8181304327158814e-05, 'epoch': 0.17}) +[2026-10-03 19:24:16,810][starVLA.training.train_starvla][INFO] - Step 2364, Loss: {'train/loss': 0.014017635025084019, 'train/grad_norm_pre_clip': 14.351414680480957, 'timing/data': 0.0003318900126032531, 'timing/model': 0.3939782809757162, 'learning_rate/qwen_vl_interface': 5.8151067566051965e-06, 'learning_rate/action_model': 5.815106756605196e-05, 'epoch': 0.17}) +[2026-10-03 19:24:17,515][starVLA.training.train_starvla][INFO] - Step 2365, Loss: {'train/loss': 0.7607118487358093, 'train/grad_norm_pre_clip': 47.82688903808594, 'timing/data': 0.00035620000562630594, 'timing/model': 0.3949014889949467, 'learning_rate/qwen_vl_interface': 5.81208284820047e-06, 'learning_rate/action_model': 5.81208284820047e-05, 'epoch': 0.17}) +[2026-10-03 19:24:18,220][starVLA.training.train_starvla][INFO] - Step 2366, Loss: {'train/loss': 0.0013511436991393566, 'train/grad_norm_pre_clip': 3.698758840560913, 'timing/data': 0.0003569900000002235, 'timing/model': 0.3944682599976659, 'learning_rate/qwen_vl_interface': 5.809058708744716e-06, 'learning_rate/action_model': 5.809058708744716e-05, 'epoch': 0.17}) +[2026-10-03 19:24:18,926][starVLA.training.train_starvla][INFO] - Step 2367, Loss: {'train/loss': 0.0008960814448073506, 'train/grad_norm_pre_clip': 1.9429527521133423, 'timing/data': 0.0003206200199201703, 'timing/model': 0.39455878001172096, 'learning_rate/qwen_vl_interface': 5.8060343394810445e-06, 'learning_rate/action_model': 5.806034339481045e-05, 'epoch': 0.17}) +[2026-10-03 19:24:19,631][starVLA.training.train_starvla][INFO] - Step 2368, Loss: {'train/loss': 0.46663954854011536, 'train/grad_norm_pre_clip': 57.9472541809082, 'timing/data': 0.00034887000219896436, 'timing/model': 0.3953416399890557, 'learning_rate/qwen_vl_interface': 5.803009741652659e-06, 'learning_rate/action_model': 5.8030097416526585e-05, 'epoch': 0.17}) +[2026-10-03 19:24:20,538][starVLA.training.train_starvla][INFO] - Step 2369, Loss: {'train/loss': 0.00010153656330658123, 'train/grad_norm_pre_clip': 0.04334524646401405, 'timing/data': 0.00022367999190464616, 'timing/model': 0.3949073789990507, 'learning_rate/qwen_vl_interface': 5.799984916502859e-06, 'learning_rate/action_model': 5.799984916502859e-05, 'epoch': 0.17}) +[2026-10-03 19:24:21,249][starVLA.training.train_starvla][INFO] - Step 2370, Loss: {'train/loss': 0.24044515192508698, 'train/grad_norm_pre_clip': 172.24807739257812, 'timing/data': 0.00036056997487321496, 'timing/model': 0.39687145000789315, 'learning_rate/qwen_vl_interface': 5.7969598652750334e-06, 'learning_rate/action_model': 5.796959865275033e-05, 'epoch': 0.17}) +[2026-10-03 19:24:21,960][starVLA.training.train_starvla][INFO] - Step 2371, Loss: {'train/loss': 0.1548098921775818, 'train/grad_norm_pre_clip': 144.1497039794922, 'timing/data': 0.0002733799919951707, 'timing/model': 0.39665817900095135, 'learning_rate/qwen_vl_interface': 5.793934589212666e-06, 'learning_rate/action_model': 5.793934589212666e-05, 'epoch': 0.17}) +[2026-10-03 19:24:22,670][starVLA.training.train_starvla][INFO] - Step 2372, Loss: {'train/loss': 0.0005761944921687245, 'train/grad_norm_pre_clip': 0.3251604735851288, 'timing/data': 0.0003425199829507619, 'timing/model': 0.39636534999590367, 'learning_rate/qwen_vl_interface': 5.7909090895593375e-06, 'learning_rate/action_model': 5.790909089559338e-05, 'epoch': 0.17}) +[2026-10-03 19:24:23,377][starVLA.training.train_starvla][INFO] - Step 2373, Loss: {'train/loss': 1.2159116522525437e-05, 'train/grad_norm_pre_clip': 0.004691298119723797, 'timing/data': 0.00035358002060092986, 'timing/model': 0.39611435899860226, 'learning_rate/qwen_vl_interface': 5.7878833675587135e-06, 'learning_rate/action_model': 5.787883367558713e-05, 'epoch': 0.17}) +[2026-10-03 19:24:24,279][starVLA.training.train_starvla][INFO] - Step 2374, Loss: {'train/loss': 0.6636319160461426, 'train/grad_norm_pre_clip': 79.21918487548828, 'timing/data': 0.0003693800244946033, 'timing/model': 0.588607924990356, 'learning_rate/qwen_vl_interface': 5.784857424454556e-06, 'learning_rate/action_model': 5.7848574244545563e-05, 'epoch': 0.17}) +[2026-10-03 19:24:24,987][starVLA.training.train_starvla][INFO] - Step 2375, Loss: {'train/loss': 3.6200028262101114e-05, 'train/grad_norm_pre_clip': 0.016436735168099403, 'timing/data': 0.0002604199980851263, 'timing/model': 0.3960874500044156, 'learning_rate/qwen_vl_interface': 5.7818312614907125e-06, 'learning_rate/action_model': 5.781831261490712e-05, 'epoch': 0.17}) +[2026-10-03 19:24:25,697][starVLA.training.train_starvla][INFO] - Step 2376, Loss: {'train/loss': 0.0908999890089035, 'train/grad_norm_pre_clip': 182.78790283203125, 'timing/data': 0.00031632001628167927, 'timing/model': 0.39696089999051765, 'learning_rate/qwen_vl_interface': 5.778804879911129e-06, 'learning_rate/action_model': 5.778804879911128e-05, 'epoch': 0.17}) +[2026-10-03 19:24:26,406][starVLA.training.train_starvla][INFO] - Step 2377, Loss: {'train/loss': 0.5495430827140808, 'train/grad_norm_pre_clip': 48.40645217895508, 'timing/data': 0.0002313799923285842, 'timing/model': 0.3963487499859184, 'learning_rate/qwen_vl_interface': 5.775778280959832e-06, 'learning_rate/action_model': 5.7757782809598315e-05, 'epoch': 0.17}) +[2026-10-03 19:24:27,116][starVLA.training.train_starvla][INFO] - Step 2378, Loss: {'train/loss': 0.0777541920542717, 'train/grad_norm_pre_clip': 92.6341323852539, 'timing/data': 0.0002263000060338527, 'timing/model': 0.39674924899009056, 'learning_rate/qwen_vl_interface': 5.7727514658809435e-06, 'learning_rate/action_model': 5.772751465880944e-05, 'epoch': 0.17}) +[2026-10-03 19:24:28,021][starVLA.training.train_starvla][INFO] - Step 2379, Loss: {'train/loss': 0.012105826288461685, 'train/grad_norm_pre_clip': 9.477893829345703, 'timing/data': 0.00022636001813225448, 'timing/model': 0.5926966450060718, 'learning_rate/qwen_vl_interface': 5.769724435918676e-06, 'learning_rate/action_model': 5.7697244359186755e-05, 'epoch': 0.17}) +[2026-10-03 19:24:28,730][starVLA.training.train_starvla][INFO] - Step 2380, Loss: {'train/loss': 0.019791534170508385, 'train/grad_norm_pre_clip': 13.859967231750488, 'timing/data': 0.00024105998454615474, 'timing/model': 0.3968067589739803, 'learning_rate/qwen_vl_interface': 5.766697192317323e-06, 'learning_rate/action_model': 5.766697192317323e-05, 'epoch': 0.17}) +[2026-10-03 19:24:29,439][starVLA.training.train_starvla][INFO] - Step 2381, Loss: {'train/loss': 0.0007611030014231801, 'train/grad_norm_pre_clip': 0.798952043056488, 'timing/data': 0.00026815998717211187, 'timing/model': 0.3951204199984204, 'learning_rate/qwen_vl_interface': 5.763669736321271e-06, 'learning_rate/action_model': 5.763669736321271e-05, 'epoch': 0.17}) +[2026-10-03 19:24:30,149][starVLA.training.train_starvla][INFO] - Step 2382, Loss: {'train/loss': 0.3255554437637329, 'train/grad_norm_pre_clip': 105.48918914794922, 'timing/data': 0.0003159600018989295, 'timing/model': 0.39616145900799893, 'learning_rate/qwen_vl_interface': 5.760642069174994e-06, 'learning_rate/action_model': 5.760642069174994e-05, 'epoch': 0.17}) +[2026-10-03 19:24:30,857][starVLA.training.train_starvla][INFO] - Step 2383, Loss: {'train/loss': 8.1483151006978e-05, 'train/grad_norm_pre_clip': 0.028676409274339676, 'timing/data': 0.00023718000738881528, 'timing/model': 0.3943261399981566, 'learning_rate/qwen_vl_interface': 5.757614192123051e-06, 'learning_rate/action_model': 5.757614192123051e-05, 'epoch': 0.17}) +[2026-10-03 19:24:31,765][starVLA.training.train_starvla][INFO] - Step 2384, Loss: {'train/loss': 0.0008587233023718, 'train/grad_norm_pre_clip': 0.273835688829422, 'timing/data': 0.0002224300114903599, 'timing/model': 0.5953216740163043, 'learning_rate/qwen_vl_interface': 5.754586106410089e-06, 'learning_rate/action_model': 5.754586106410088e-05, 'epoch': 0.17}) +[2026-10-03 19:24:32,475][starVLA.training.train_starvla][INFO] - Step 2385, Loss: {'train/loss': 0.004836873151361942, 'train/grad_norm_pre_clip': 1.902601718902588, 'timing/data': 0.00035641001886688173, 'timing/model': 0.3964591989933979, 'learning_rate/qwen_vl_interface': 5.751557813280838e-06, 'learning_rate/action_model': 5.751557813280838e-05, 'epoch': 0.17}) +[2026-10-03 19:24:33,185][starVLA.training.train_starvla][INFO] - Step 2386, Loss: {'train/loss': 0.05436209961771965, 'train/grad_norm_pre_clip': 85.10047149658203, 'timing/data': 0.00035000001662410796, 'timing/model': 0.3961067299824208, 'learning_rate/qwen_vl_interface': 5.7485293139801146e-06, 'learning_rate/action_model': 5.7485293139801146e-05, 'epoch': 0.17}) +[2026-10-03 19:24:33,895][starVLA.training.train_starvla][INFO] - Step 2387, Loss: {'train/loss': 0.023455731570720673, 'train/grad_norm_pre_clip': 19.983566284179688, 'timing/data': 0.00036315000033937395, 'timing/model': 0.3969463389948942, 'learning_rate/qwen_vl_interface': 5.745500609752821e-06, 'learning_rate/action_model': 5.745500609752821e-05, 'epoch': 0.17}) +[2026-10-03 19:24:34,605][starVLA.training.train_starvla][INFO] - Step 2388, Loss: {'train/loss': 4.0161816286854446e-05, 'train/grad_norm_pre_clip': 0.018806923180818558, 'timing/data': 0.00036691001150757074, 'timing/model': 0.39632775899372064, 'learning_rate/qwen_vl_interface': 5.742471701843943e-06, 'learning_rate/action_model': 5.742471701843943e-05, 'epoch': 0.17}) +[2026-10-03 19:24:35,312][starVLA.training.train_starvla][INFO] - Step 2389, Loss: {'train/loss': 0.00030789009178988636, 'train/grad_norm_pre_clip': 0.3088359236717224, 'timing/data': 0.00029242999153211713, 'timing/model': 0.39489997999044135, 'learning_rate/qwen_vl_interface': 5.739442591498553e-06, 'learning_rate/action_model': 5.739442591498553e-05, 'epoch': 0.17}) +[2026-10-03 19:24:36,218][starVLA.training.train_starvla][INFO] - Step 2390, Loss: {'train/loss': 1.2539123417809606e-05, 'train/grad_norm_pre_clip': 0.003352411324158311, 'timing/data': 0.00023713998962193727, 'timing/model': 0.3949448399944231, 'learning_rate/qwen_vl_interface': 5.7364132799617985e-06, 'learning_rate/action_model': 5.7364132799617984e-05, 'epoch': 0.17}) +[2026-10-03 19:24:36,927][starVLA.training.train_starvla][INFO] - Step 2391, Loss: {'train/loss': 0.000537816493306309, 'train/grad_norm_pre_clip': 0.2032884955406189, 'timing/data': 0.00028925001970492303, 'timing/model': 0.3960957690142095, 'learning_rate/qwen_vl_interface': 5.733383768478923e-06, 'learning_rate/action_model': 5.7333837684789225e-05, 'epoch': 0.18}) +[2026-10-03 19:24:37,636][starVLA.training.train_starvla][INFO] - Step 2392, Loss: {'train/loss': 0.016904419288039207, 'train/grad_norm_pre_clip': 54.87671661376953, 'timing/data': 0.0003481500025372952, 'timing/model': 0.3962546490074601, 'learning_rate/qwen_vl_interface': 5.730354058295234e-06, 'learning_rate/action_model': 5.730354058295234e-05, 'epoch': 0.18}) +[2026-10-03 19:24:38,346][starVLA.training.train_starvla][INFO] - Step 2393, Loss: {'train/loss': 0.4807363450527191, 'train/grad_norm_pre_clip': 79.30042266845703, 'timing/data': 0.00045133999083191156, 'timing/model': 0.3965810200024862, 'learning_rate/qwen_vl_interface': 5.72732415065614e-06, 'learning_rate/action_model': 5.727324150656139e-05, 'epoch': 0.18}) +[2026-10-03 19:24:39,054][starVLA.training.train_starvla][INFO] - Step 2394, Loss: {'train/loss': 0.015353362075984478, 'train/grad_norm_pre_clip': 19.69601058959961, 'timing/data': 0.00037290999898687005, 'timing/model': 0.39581739000277594, 'learning_rate/qwen_vl_interface': 5.724294046807115e-06, 'learning_rate/action_model': 5.724294046807115e-05, 'epoch': 0.18}) +[2026-10-03 19:24:39,956][starVLA.training.train_starvla][INFO] - Step 2395, Loss: {'train/loss': 0.00021692027803510427, 'train/grad_norm_pre_clip': 0.06057910993695259, 'timing/data': 0.0003431200166232884, 'timing/model': 0.587981454009423, 'learning_rate/qwen_vl_interface': 5.7212637479937234e-06, 'learning_rate/action_model': 5.721263747993724e-05, 'epoch': 0.18}) +[2026-10-03 19:24:40,665][starVLA.training.train_starvla][INFO] - Step 2396, Loss: {'train/loss': 0.0003439459251239896, 'train/grad_norm_pre_clip': 0.062027834355831146, 'timing/data': 0.00032541999826207757, 'timing/model': 0.3954983590228949, 'learning_rate/qwen_vl_interface': 5.718233255461607e-06, 'learning_rate/action_model': 5.718233255461607e-05, 'epoch': 0.18}) +[2026-10-03 19:24:41,374][starVLA.training.train_starvla][INFO] - Step 2397, Loss: {'train/loss': 1.2650545613723807e-05, 'train/grad_norm_pre_clip': 0.0035488875582814217, 'timing/data': 0.0003768599999602884, 'timing/model': 0.39636035001603886, 'learning_rate/qwen_vl_interface': 5.715202570456486e-06, 'learning_rate/action_model': 5.715202570456486e-05, 'epoch': 0.18}) +[2026-10-03 19:24:42,084][starVLA.training.train_starvla][INFO] - Step 2398, Loss: {'train/loss': 0.016874223947525024, 'train/grad_norm_pre_clip': 44.85325622558594, 'timing/data': 0.00036653000279329717, 'timing/model': 0.3957029790035449, 'learning_rate/qwen_vl_interface': 5.7121716942241615e-06, 'learning_rate/action_model': 5.712171694224161e-05, 'epoch': 0.18}) +[2026-10-03 19:24:42,789][starVLA.training.train_starvla][INFO] - Step 2399, Loss: {'train/loss': 0.0001316852285526693, 'train/grad_norm_pre_clip': 0.06751297414302826, 'timing/data': 0.0003334600187372416, 'timing/model': 0.39439776897779666, 'learning_rate/qwen_vl_interface': 5.709140628010509e-06, 'learning_rate/action_model': 5.7091406280105085e-05, 'epoch': 0.18}) +[2026-10-03 19:24:43,497][starVLA.training.train_starvla][INFO] - Step 2400, Loss: {'train/loss': 0.12313482910394669, 'train/grad_norm_pre_clip': 36.85369110107422, 'timing/data': 0.00034319001133553684, 'timing/model': 0.3943147390091326, 'learning_rate/qwen_vl_interface': 5.706109373061489e-06, 'learning_rate/action_model': 5.706109373061489e-05, 'epoch': 0.18}) +[2026-10-03 19:24:44,403][starVLA.training.train_starvla][INFO] - Step 2401, Loss: {'train/loss': 0.06238511949777603, 'train/grad_norm_pre_clip': 83.79653930664062, 'timing/data': 0.00022555000032298267, 'timing/model': 0.39507374897948466, 'learning_rate/qwen_vl_interface': 5.703077930623134e-06, 'learning_rate/action_model': 5.7030779306231344e-05, 'epoch': 0.18}) +[2026-10-03 19:24:45,111][starVLA.training.train_starvla][INFO] - Step 2402, Loss: {'train/loss': 0.6896592378616333, 'train/grad_norm_pre_clip': 305.9800109863281, 'timing/data': 0.00037395002436824143, 'timing/model': 0.39441390900174156, 'learning_rate/qwen_vl_interface': 5.7000463019415565e-06, 'learning_rate/action_model': 5.7000463019415563e-05, 'epoch': 0.18}) +[2026-10-03 19:24:45,820][starVLA.training.train_starvla][INFO] - Step 2403, Loss: {'train/loss': 9.846237662713975e-05, 'train/grad_norm_pre_clip': 0.03960069268941879, 'timing/data': 0.00021965999621897936, 'timing/model': 0.39455423899926245, 'learning_rate/qwen_vl_interface': 5.697014488262945e-06, 'learning_rate/action_model': 5.697014488262945e-05, 'epoch': 0.18}) +[2026-10-03 19:24:46,531][starVLA.training.train_starvla][INFO] - Step 2404, Loss: {'train/loss': 0.05836670473217964, 'train/grad_norm_pre_clip': 138.2897491455078, 'timing/data': 0.00038608998875133693, 'timing/model': 0.3974838290014304, 'learning_rate/qwen_vl_interface': 5.693982490833561e-06, 'learning_rate/action_model': 5.693982490833561e-05, 'epoch': 0.18}) +[2026-10-03 19:24:47,240][starVLA.training.train_starvla][INFO] - Step 2405, Loss: {'train/loss': 4.20463147747796e-05, 'train/grad_norm_pre_clip': 0.020341653376817703, 'timing/data': 0.0002923900028690696, 'timing/model': 0.3967270989960525, 'learning_rate/qwen_vl_interface': 5.690950310899745e-06, 'learning_rate/action_model': 5.6909503108997456e-05, 'epoch': 0.18}) +[2026-10-03 19:24:48,142][starVLA.training.train_starvla][INFO] - Step 2406, Loss: {'train/loss': 0.6889377236366272, 'train/grad_norm_pre_clip': 365.52972412109375, 'timing/data': 0.00036209000973030925, 'timing/model': 0.3945411390159279, 'learning_rate/qwen_vl_interface': 5.687917949707913e-06, 'learning_rate/action_model': 5.687917949707913e-05, 'epoch': 0.18}) +[2026-10-03 19:24:48,850][starVLA.training.train_starvla][INFO] - Step 2407, Loss: {'train/loss': 0.5597540736198425, 'train/grad_norm_pre_clip': 61.30159378051758, 'timing/data': 0.00033024000003933907, 'timing/model': 0.3944034790038131, 'learning_rate/qwen_vl_interface': 5.684885408504555e-06, 'learning_rate/action_model': 5.684885408504554e-05, 'epoch': 0.18}) +[2026-10-03 19:24:49,557][starVLA.training.train_starvla][INFO] - Step 2408, Loss: {'train/loss': 0.5890904664993286, 'train/grad_norm_pre_clip': 118.01312255859375, 'timing/data': 0.000308859976939857, 'timing/model': 0.3947693289956078, 'learning_rate/qwen_vl_interface': 5.681852688536231e-06, 'learning_rate/action_model': 5.681852688536231e-05, 'epoch': 0.18}) +[2026-10-03 19:24:50,264][starVLA.training.train_starvla][INFO] - Step 2409, Loss: {'train/loss': 0.27658140659332275, 'train/grad_norm_pre_clip': 194.94631958007812, 'timing/data': 0.00022525998065248132, 'timing/model': 0.3943042989994865, 'learning_rate/qwen_vl_interface': 5.67881979104958e-06, 'learning_rate/action_model': 5.67881979104958e-05, 'epoch': 0.18}) +[2026-10-03 19:24:50,974][starVLA.training.train_starvla][INFO] - Step 2410, Loss: {'train/loss': 1.2300407886505127, 'train/grad_norm_pre_clip': 96.40122985839844, 'timing/data': 0.0002960299898404628, 'timing/model': 0.3965049389807973, 'learning_rate/qwen_vl_interface': 5.67578671729131e-06, 'learning_rate/action_model': 5.675786717291309e-05, 'epoch': 0.18}) +[2026-10-03 19:24:51,882][starVLA.training.train_starvla][INFO] - Step 2411, Loss: {'train/loss': 0.21128569543361664, 'train/grad_norm_pre_clip': 156.40647888183594, 'timing/data': 0.0003761200059670955, 'timing/model': 0.5937869129993487, 'learning_rate/qwen_vl_interface': 5.672753468508204e-06, 'learning_rate/action_model': 5.672753468508204e-05, 'epoch': 0.18}) +[2026-10-03 19:24:52,588][starVLA.training.train_starvla][INFO] - Step 2412, Loss: {'train/loss': 0.02821950614452362, 'train/grad_norm_pre_clip': 39.0528678894043, 'timing/data': 0.00024619000032544136, 'timing/model': 0.39495143899694085, 'learning_rate/qwen_vl_interface': 5.6697200459471165e-06, 'learning_rate/action_model': 5.669720045947117e-05, 'epoch': 0.18}) +[2026-10-03 19:24:53,298][starVLA.training.train_starvla][INFO] - Step 2413, Loss: {'train/loss': 0.3349418044090271, 'train/grad_norm_pre_clip': 86.94367218017578, 'timing/data': 0.00021882998407818377, 'timing/model': 0.39692427901900373, 'learning_rate/qwen_vl_interface': 5.666686450854973e-06, 'learning_rate/action_model': 5.666686450854973e-05, 'epoch': 0.18}) +[2026-10-03 19:24:54,008][starVLA.training.train_starvla][INFO] - Step 2414, Loss: {'train/loss': 0.0003412200603634119, 'train/grad_norm_pre_clip': 0.254657119512558, 'timing/data': 0.00021963001927360892, 'timing/model': 0.3963843889941927, 'learning_rate/qwen_vl_interface': 5.663652684478768e-06, 'learning_rate/action_model': 5.6636526844787676e-05, 'epoch': 0.18}) +[2026-10-03 19:24:54,715][starVLA.training.train_starvla][INFO] - Step 2415, Loss: {'train/loss': 0.3325756788253784, 'train/grad_norm_pre_clip': 54.955787658691406, 'timing/data': 0.0003589400148484856, 'timing/model': 0.3948770400020294, 'learning_rate/qwen_vl_interface': 5.660618748065569e-06, 'learning_rate/action_model': 5.6606187480655695e-05, 'epoch': 0.18}) +[2026-10-03 19:24:55,424][starVLA.training.train_starvla][INFO] - Step 2416, Loss: {'train/loss': 4.1424827941227704e-06, 'train/grad_norm_pre_clip': 0.002460726071149111, 'timing/data': 0.00032398998155258596, 'timing/model': 0.3960860199877061, 'learning_rate/qwen_vl_interface': 5.657584642862514e-06, 'learning_rate/action_model': 5.657584642862513e-05, 'epoch': 0.18}) +[2026-10-03 19:24:56,332][starVLA.training.train_starvla][INFO] - Step 2417, Loss: {'train/loss': 0.0024008245673030615, 'train/grad_norm_pre_clip': 1.4325364828109741, 'timing/data': 0.0002715199952945113, 'timing/model': 0.3967977790161967, 'learning_rate/qwen_vl_interface': 5.654550370116808e-06, 'learning_rate/action_model': 5.654550370116808e-05, 'epoch': 0.18}) +[2026-10-03 19:24:57,040][starVLA.training.train_starvla][INFO] - Step 2418, Loss: {'train/loss': 0.22590072453022003, 'train/grad_norm_pre_clip': 123.90970611572266, 'timing/data': 0.00030611001420766115, 'timing/model': 0.396482439013198, 'learning_rate/qwen_vl_interface': 5.651515931075726e-06, 'learning_rate/action_model': 5.651515931075726e-05, 'epoch': 0.18}) +[2026-10-03 19:24:57,748][starVLA.training.train_starvla][INFO] - Step 2419, Loss: {'train/loss': 0.006246831268072128, 'train/grad_norm_pre_clip': 4.685561656951904, 'timing/data': 0.000225499999942258, 'timing/model': 0.39590278902323917, 'learning_rate/qwen_vl_interface': 5.6484813269866126e-06, 'learning_rate/action_model': 5.648481326986612e-05, 'epoch': 0.18}) +[2026-10-03 19:24:58,458][starVLA.training.train_starvla][INFO] - Step 2420, Loss: {'train/loss': 1.1507583856582642, 'train/grad_norm_pre_clip': 81.89705657958984, 'timing/data': 0.0002817499917000532, 'timing/model': 0.3975808489776682, 'learning_rate/qwen_vl_interface': 5.645446559096875e-06, 'learning_rate/action_model': 5.645446559096875e-05, 'epoch': 0.18}) +[2026-10-03 19:24:59,166][starVLA.training.train_starvla][INFO] - Step 2421, Loss: {'train/loss': 0.5937093496322632, 'train/grad_norm_pre_clip': 160.27218627929688, 'timing/data': 0.00027327900170348585, 'timing/model': 0.3962129800056573, 'learning_rate/qwen_vl_interface': 5.6424116286539965e-06, 'learning_rate/action_model': 5.642411628653996e-05, 'epoch': 0.18}) +[2026-10-03 19:25:00,071][starVLA.training.train_starvla][INFO] - Step 2422, Loss: {'train/loss': 0.19870592653751373, 'train/grad_norm_pre_clip': 69.39006805419922, 'timing/data': 0.00026078999508172274, 'timing/model': 0.3958555090066511, 'learning_rate/qwen_vl_interface': 5.639376536905517e-06, 'learning_rate/action_model': 5.6393765369055175e-05, 'epoch': 0.18}) +[2026-10-03 19:25:00,779][starVLA.training.train_starvla][INFO] - Step 2423, Loss: {'train/loss': 0.01085631176829338, 'train/grad_norm_pre_clip': 7.642410755157471, 'timing/data': 0.0003145300142932683, 'timing/model': 0.3942045499861706, 'learning_rate/qwen_vl_interface': 5.636341285099055e-06, 'learning_rate/action_model': 5.636341285099055e-05, 'epoch': 0.18}) +[2026-10-03 19:25:01,485][starVLA.training.train_starvla][INFO] - Step 2424, Loss: {'train/loss': 0.3954834043979645, 'train/grad_norm_pre_clip': 60.611473083496094, 'timing/data': 0.00030798002262599766, 'timing/model': 0.3939312190050259, 'learning_rate/qwen_vl_interface': 5.6333058744822835e-06, 'learning_rate/action_model': 5.633305874482283e-05, 'epoch': 0.18}) +[2026-10-03 19:25:02,196][starVLA.training.train_starvla][INFO] - Step 2425, Loss: {'train/loss': 0.08002778887748718, 'train/grad_norm_pre_clip': 91.97993469238281, 'timing/data': 0.00039360998198390007, 'timing/model': 0.39750998897943646, 'learning_rate/qwen_vl_interface': 5.630270306302944e-06, 'learning_rate/action_model': 5.6302703063029436e-05, 'epoch': 0.18}) +[2026-10-03 19:25:02,904][starVLA.training.train_starvla][INFO] - Step 2426, Loss: {'train/loss': 0.4329831898212433, 'train/grad_norm_pre_clip': 25.432889938354492, 'timing/data': 0.00021933999960310757, 'timing/model': 0.3965711090131663, 'learning_rate/qwen_vl_interface': 5.627234581808848e-06, 'learning_rate/action_model': 5.627234581808848e-05, 'epoch': 0.18}) +[2026-10-03 19:25:03,811][starVLA.training.train_starvla][INFO] - Step 2427, Loss: {'train/loss': 0.0010395245626568794, 'train/grad_norm_pre_clip': 0.7105548977851868, 'timing/data': 0.00022782001178711653, 'timing/model': 0.5952346439880785, 'learning_rate/qwen_vl_interface': 5.624198702247863e-06, 'learning_rate/action_model': 5.624198702247863e-05, 'epoch': 0.18}) +[2026-10-03 19:25:04,522][starVLA.training.train_starvla][INFO] - Step 2428, Loss: {'train/loss': 0.020957041531801224, 'train/grad_norm_pre_clip': 27.41356086730957, 'timing/data': 0.00036418999661691487, 'timing/model': 0.39666798899997957, 'learning_rate/qwen_vl_interface': 5.621162668867925e-06, 'learning_rate/action_model': 5.621162668867925e-05, 'epoch': 0.18}) +[2026-10-03 19:25:05,231][starVLA.training.train_starvla][INFO] - Step 2429, Loss: {'train/loss': 0.1965513974428177, 'train/grad_norm_pre_clip': 86.51432800292969, 'timing/data': 0.00035754000418819487, 'timing/model': 0.39594630801002495, 'learning_rate/qwen_vl_interface': 5.618126482917037e-06, 'learning_rate/action_model': 5.618126482917036e-05, 'epoch': 0.18}) +[2026-10-03 19:25:05,943][starVLA.training.train_starvla][INFO] - Step 2430, Loss: {'train/loss': 0.00038342599873431027, 'train/grad_norm_pre_clip': 0.2988036572933197, 'timing/data': 0.00034997001057490706, 'timing/model': 0.39682274899678305, 'learning_rate/qwen_vl_interface': 5.6150901456432546e-06, 'learning_rate/action_model': 5.615090145643255e-05, 'epoch': 0.18}) +[2026-10-03 19:25:06,646][starVLA.training.train_starvla][INFO] - Step 2431, Loss: {'train/loss': 0.00033054346567951143, 'train/grad_norm_pre_clip': 0.12250107526779175, 'timing/data': 0.00034893001429736614, 'timing/model': 0.3928781889844686, 'learning_rate/qwen_vl_interface': 5.612053658294705e-06, 'learning_rate/action_model': 5.6120536582947045e-05, 'epoch': 0.18}) +[2026-10-03 19:25:07,552][starVLA.training.train_starvla][INFO] - Step 2432, Loss: {'train/loss': 0.2512317895889282, 'train/grad_norm_pre_clip': 127.12071228027344, 'timing/data': 0.0003684799885377288, 'timing/model': 0.5957961439853534, 'learning_rate/qwen_vl_interface': 5.609017022119572e-06, 'learning_rate/action_model': 5.609017022119573e-05, 'epoch': 0.18}) +[2026-10-03 19:25:08,255][starVLA.training.train_starvla][INFO] - Step 2433, Loss: {'train/loss': 0.002498288406059146, 'train/grad_norm_pre_clip': 1.2642971277236938, 'timing/data': 0.00037483000778593123, 'timing/model': 0.3932918589853216, 'learning_rate/qwen_vl_interface': 5.605980238366104e-06, 'learning_rate/action_model': 5.605980238366103e-05, 'epoch': 0.18}) +[2026-10-03 19:25:08,963][starVLA.training.train_starvla][INFO] - Step 2434, Loss: {'train/loss': 0.00018212900613434613, 'train/grad_norm_pre_clip': 0.04134557768702507, 'timing/data': 0.0003329500032123178, 'timing/model': 0.39636985899414867, 'learning_rate/qwen_vl_interface': 5.602943308282606e-06, 'learning_rate/action_model': 5.602943308282605e-05, 'epoch': 0.18}) +[2026-10-03 19:25:09,669][starVLA.training.train_starvla][INFO] - Step 2435, Loss: {'train/loss': 0.007997437380254269, 'train/grad_norm_pre_clip': 8.898299217224121, 'timing/data': 0.00030993999098427594, 'timing/model': 0.39414668901008554, 'learning_rate/qwen_vl_interface': 5.599906233117445e-06, 'learning_rate/action_model': 5.5999062331174446e-05, 'epoch': 0.18}) +[2026-10-03 19:25:10,374][starVLA.training.train_starvla][INFO] - Step 2436, Loss: {'train/loss': 0.000677434029057622, 'train/grad_norm_pre_clip': 0.3916107416152954, 'timing/data': 0.0003334200009703636, 'timing/model': 0.39443688897881657, 'learning_rate/qwen_vl_interface': 5.596869014119048e-06, 'learning_rate/action_model': 5.596869014119048e-05, 'epoch': 0.18}) +[2026-10-03 19:25:11,278][starVLA.training.train_starvla][INFO] - Step 2437, Loss: {'train/loss': 0.3579580783843994, 'train/grad_norm_pre_clip': 109.7728500366211, 'timing/data': 0.00038345999200828373, 'timing/model': 0.5941752529761288, 'learning_rate/qwen_vl_interface': 5.593831652535902e-06, 'learning_rate/action_model': 5.593831652535902e-05, 'epoch': 0.18}) +[2026-10-03 19:25:11,986][starVLA.training.train_starvla][INFO] - Step 2438, Loss: {'train/loss': 7.97272368799895e-05, 'train/grad_norm_pre_clip': 0.02465754747390747, 'timing/data': 0.0003335499786771834, 'timing/model': 0.3958935380214825, 'learning_rate/qwen_vl_interface': 5.590794149616552e-06, 'learning_rate/action_model': 5.590794149616552e-05, 'epoch': 0.18}) +[2026-10-03 19:25:12,695][starVLA.training.train_starvla][INFO] - Step 2439, Loss: {'train/loss': 0.09790550917387009, 'train/grad_norm_pre_clip': 47.03013229370117, 'timing/data': 0.00023094998323358595, 'timing/model': 0.3921068589843344, 'learning_rate/qwen_vl_interface': 5.587756506609601e-06, 'learning_rate/action_model': 5.5877565066096006e-05, 'epoch': 0.18}) +[2026-10-03 19:25:13,398][starVLA.training.train_starvla][INFO] - Step 2440, Loss: {'train/loss': 0.0017060147365555167, 'train/grad_norm_pre_clip': 1.193080186843872, 'timing/data': 0.0003205499961040914, 'timing/model': 0.3924369690066669, 'learning_rate/qwen_vl_interface': 5.584718724763707e-06, 'learning_rate/action_model': 5.584718724763707e-05, 'epoch': 0.18}) +[2026-10-03 19:25:14,102][starVLA.training.train_starvla][INFO] - Step 2441, Loss: {'train/loss': 0.03179595619440079, 'train/grad_norm_pre_clip': 33.75865173339844, 'timing/data': 0.0002272500132676214, 'timing/model': 0.3927048289915547, 'learning_rate/qwen_vl_interface': 5.581680805327588e-06, 'learning_rate/action_model': 5.5816808053275885e-05, 'epoch': 0.18}) +[2026-10-03 19:25:14,809][starVLA.training.train_starvla][INFO] - Step 2442, Loss: {'train/loss': 0.26538199186325073, 'train/grad_norm_pre_clip': 35.70375442504883, 'timing/data': 0.0002498899993952364, 'timing/model': 0.39545658900169656, 'learning_rate/qwen_vl_interface': 5.578642749550019e-06, 'learning_rate/action_model': 5.578642749550019e-05, 'epoch': 0.18}) +[2026-10-03 19:25:15,711][starVLA.training.train_starvla][INFO] - Step 2443, Loss: {'train/loss': 0.0052261557430028915, 'train/grad_norm_pre_clip': 3.500619649887085, 'timing/data': 0.00033271999564021826, 'timing/model': 0.3920068900042679, 'learning_rate/qwen_vl_interface': 5.575604558679831e-06, 'learning_rate/action_model': 5.57560455867983e-05, 'epoch': 0.18}) +[2026-10-03 19:25:16,414][starVLA.training.train_starvla][INFO] - Step 2444, Loss: {'train/loss': 0.14754483103752136, 'train/grad_norm_pre_clip': 93.28707885742188, 'timing/data': 0.0003168499970342964, 'timing/model': 0.393227588996524, 'learning_rate/qwen_vl_interface': 5.5725662339659085e-06, 'learning_rate/action_model': 5.572566233965908e-05, 'epoch': 0.18}) +[2026-10-03 19:25:17,120][starVLA.training.train_starvla][INFO] - Step 2445, Loss: {'train/loss': 0.6879157423973083, 'train/grad_norm_pre_clip': 98.2213134765625, 'timing/data': 0.00035347999073565006, 'timing/model': 0.39542389899725094, 'learning_rate/qwen_vl_interface': 5.56952777665719e-06, 'learning_rate/action_model': 5.56952777665719e-05, 'epoch': 0.18}) +[2026-10-03 19:25:17,828][starVLA.training.train_starvla][INFO] - Step 2446, Loss: {'train/loss': 0.5903201103210449, 'train/grad_norm_pre_clip': 130.4129638671875, 'timing/data': 0.00025723999715410173, 'timing/model': 0.39520096901105717, 'learning_rate/qwen_vl_interface': 5.566489188002671e-06, 'learning_rate/action_model': 5.56648918800267e-05, 'epoch': 0.18}) +[2026-10-03 19:25:18,534][starVLA.training.train_starvla][INFO] - Step 2447, Loss: {'train/loss': 0.0018031811341643333, 'train/grad_norm_pre_clip': 1.2376731634140015, 'timing/data': 0.0003324699937365949, 'timing/model': 0.3938690679788124, 'learning_rate/qwen_vl_interface': 5.563450469251403e-06, 'learning_rate/action_model': 5.563450469251402e-05, 'epoch': 0.18}) +[2026-10-03 19:25:19,444][starVLA.training.train_starvla][INFO] - Step 2448, Loss: {'train/loss': 0.0006938448059372604, 'train/grad_norm_pre_clip': 0.26122188568115234, 'timing/data': 0.0003654289757832885, 'timing/model': 0.5984478729951661, 'learning_rate/qwen_vl_interface': 5.560411621652488e-06, 'learning_rate/action_model': 5.560411621652488e-05, 'epoch': 0.18}) +[2026-10-03 19:25:20,151][starVLA.training.train_starvla][INFO] - Step 2449, Loss: {'train/loss': 0.017707835882902145, 'train/grad_norm_pre_clip': 13.023289680480957, 'timing/data': 0.000308190006762743, 'timing/model': 0.3955783789861016, 'learning_rate/qwen_vl_interface': 5.557372646455081e-06, 'learning_rate/action_model': 5.55737264645508e-05, 'epoch': 0.18}) +[2026-10-03 19:25:20,860][starVLA.training.train_starvla][INFO] - Step 2450, Loss: {'train/loss': 0.1113039180636406, 'train/grad_norm_pre_clip': 60.77423095703125, 'timing/data': 0.0002721899945754558, 'timing/model': 0.3965726589958649, 'learning_rate/qwen_vl_interface': 5.554333544908388e-06, 'learning_rate/action_model': 5.5543335449083875e-05, 'epoch': 0.18}) +[2026-10-03 19:25:21,566][starVLA.training.train_starvla][INFO] - Step 2451, Loss: {'train/loss': 0.011657534167170525, 'train/grad_norm_pre_clip': 3.1216635704040527, 'timing/data': 0.0003454300167504698, 'timing/model': 0.39384496901766397, 'learning_rate/qwen_vl_interface': 5.551294318261671e-06, 'learning_rate/action_model': 5.5512943182616705e-05, 'epoch': 0.18}) +[2026-10-03 19:25:22,273][starVLA.training.train_starvla][INFO] - Step 2452, Loss: {'train/loss': 6.950107490411028e-05, 'train/grad_norm_pre_clip': 0.017969105392694473, 'timing/data': 0.00033614999847486615, 'timing/model': 0.3946974390128162, 'learning_rate/qwen_vl_interface': 5.548254967764239e-06, 'learning_rate/action_model': 5.5482549677642394e-05, 'epoch': 0.18}) +[2026-10-03 19:25:22,980][starVLA.training.train_starvla][INFO] - Step 2453, Loss: {'train/loss': 0.37607643008232117, 'train/grad_norm_pre_clip': 44.40915298461914, 'timing/data': 0.00035506999120116234, 'timing/model': 0.3959696880192496, 'learning_rate/qwen_vl_interface': 5.545215494665458e-06, 'learning_rate/action_model': 5.545215494665458e-05, 'epoch': 0.18}) +[2026-10-03 19:25:23,887][starVLA.training.train_starvla][INFO] - Step 2454, Loss: {'train/loss': 0.10986728221178055, 'train/grad_norm_pre_clip': 118.8534164428711, 'timing/data': 0.000258079991908744, 'timing/model': 0.39618731901282445, 'learning_rate/qwen_vl_interface': 5.5421759002147376e-06, 'learning_rate/action_model': 5.542175900214738e-05, 'epoch': 0.18}) +[2026-10-03 19:25:24,593][starVLA.training.train_starvla][INFO] - Step 2455, Loss: {'train/loss': 0.00016720070561859757, 'train/grad_norm_pre_clip': 0.14974628388881683, 'timing/data': 0.00021903999731875956, 'timing/model': 0.39468430902343243, 'learning_rate/qwen_vl_interface': 5.53913618566154e-06, 'learning_rate/action_model': 5.53913618566154e-05, 'epoch': 0.18}) +[2026-10-03 19:25:25,297][starVLA.training.train_starvla][INFO] - Step 2456, Loss: {'train/loss': 0.35268422961235046, 'train/grad_norm_pre_clip': 167.2010040283203, 'timing/data': 0.00022186999558471143, 'timing/model': 0.3930278990010265, 'learning_rate/qwen_vl_interface': 5.536096352255381e-06, 'learning_rate/action_model': 5.536096352255381e-05, 'epoch': 0.18}) +[2026-10-03 19:25:25,999][starVLA.training.train_starvla][INFO] - Step 2457, Loss: {'train/loss': 0.002204910619184375, 'train/grad_norm_pre_clip': 4.423135757446289, 'timing/data': 0.0002578900021035224, 'timing/model': 0.39318784899660386, 'learning_rate/qwen_vl_interface': 5.533056401245817e-06, 'learning_rate/action_model': 5.5330564012458164e-05, 'epoch': 0.18}) +[2026-10-03 19:25:26,707][starVLA.training.train_starvla][INFO] - Step 2458, Loss: {'train/loss': 0.21909034252166748, 'train/grad_norm_pre_clip': 106.01654815673828, 'timing/data': 0.00036924000596627593, 'timing/model': 0.39576629898510873, 'learning_rate/qwen_vl_interface': 5.530016333882456e-06, 'learning_rate/action_model': 5.5300163338824565e-05, 'epoch': 0.18}) +[2026-10-03 19:25:27,615][starVLA.training.train_starvla][INFO] - Step 2459, Loss: {'train/loss': 0.5705838203430176, 'train/grad_norm_pre_clip': 37.97814178466797, 'timing/data': 0.0002538900007493794, 'timing/model': 0.39515234899590723, 'learning_rate/qwen_vl_interface': 5.52697615141496e-06, 'learning_rate/action_model': 5.52697615141496e-05, 'epoch': 0.18}) +[2026-10-03 19:25:28,321][starVLA.training.train_starvla][INFO] - Step 2460, Loss: {'train/loss': 7.588112930534407e-05, 'train/grad_norm_pre_clip': 0.07272334396839142, 'timing/data': 0.0003564999788068235, 'timing/model': 0.3934668790025171, 'learning_rate/qwen_vl_interface': 5.52393585509303e-06, 'learning_rate/action_model': 5.52393585509303e-05, 'epoch': 0.18}) +[2026-10-03 19:25:29,028][starVLA.training.train_starvla][INFO] - Step 2461, Loss: {'train/loss': 0.24394726753234863, 'train/grad_norm_pre_clip': 101.4326400756836, 'timing/data': 0.00026483001420274377, 'timing/model': 0.3954563190054614, 'learning_rate/qwen_vl_interface': 5.520895446166418e-06, 'learning_rate/action_model': 5.5208954461664175e-05, 'epoch': 0.18}) +[2026-10-03 19:25:29,735][starVLA.training.train_starvla][INFO] - Step 2462, Loss: {'train/loss': 0.6885462999343872, 'train/grad_norm_pre_clip': 82.22168731689453, 'timing/data': 0.00034216998028568923, 'timing/model': 0.3962358490098268, 'learning_rate/qwen_vl_interface': 5.517854925884919e-06, 'learning_rate/action_model': 5.517854925884919e-05, 'epoch': 0.18}) +[2026-10-03 19:25:30,439][starVLA.training.train_starvla][INFO] - Step 2463, Loss: {'train/loss': 0.001124270143918693, 'train/grad_norm_pre_clip': 0.8423361778259277, 'timing/data': 0.0003450399963185191, 'timing/model': 0.39253579900832847, 'learning_rate/qwen_vl_interface': 5.514814295498379e-06, 'learning_rate/action_model': 5.514814295498378e-05, 'epoch': 0.18}) +[2026-10-03 19:25:31,360][starVLA.training.train_starvla][INFO] - Step 2464, Loss: {'train/loss': 0.03957901895046234, 'train/grad_norm_pre_clip': 44.60606384277344, 'timing/data': 0.0003739600069820881, 'timing/model': 0.6082387509813998, 'learning_rate/qwen_vl_interface': 5.5117735562566846e-06, 'learning_rate/action_model': 5.5117735562566844e-05, 'epoch': 0.18}) +[2026-10-03 19:25:32,066][starVLA.training.train_starvla][INFO] - Step 2465, Loss: {'train/loss': 0.13562743365764618, 'train/grad_norm_pre_clip': 41.849124908447266, 'timing/data': 0.0003783799766097218, 'timing/model': 0.3953227690071799, 'learning_rate/qwen_vl_interface': 5.5087327094097685e-06, 'learning_rate/action_model': 5.508732709409769e-05, 'epoch': 0.18}) +[2026-10-03 19:25:32,747][starVLA.training.train_starvla][INFO] - Step 2466, Loss: {'train/loss': 0.30181288719177246, 'train/grad_norm_pre_clip': 29.017709732055664, 'timing/data': 0.000348620000295341, 'timing/model': 0.39463893900392577, 'learning_rate/qwen_vl_interface': 5.505691756207609e-06, 'learning_rate/action_model': 5.505691756207609e-05, 'epoch': 0.18}) +[2026-10-03 19:25:33,453][starVLA.training.train_starvla][INFO] - Step 2467, Loss: {'train/loss': 0.6680575013160706, 'train/grad_norm_pre_clip': 34.994625091552734, 'timing/data': 0.00034224000410176814, 'timing/model': 0.3957938290259335, 'learning_rate/qwen_vl_interface': 5.502650697900228e-06, 'learning_rate/action_model': 5.502650697900228e-05, 'epoch': 0.18}) +[2026-10-03 19:25:34,158][starVLA.training.train_starvla][INFO] - Step 2468, Loss: {'train/loss': 0.5703153014183044, 'train/grad_norm_pre_clip': 56.45729064941406, 'timing/data': 0.00035172002390027046, 'timing/model': 0.39500229799887165, 'learning_rate/qwen_vl_interface': 5.499609535737688e-06, 'learning_rate/action_model': 5.4996095357376875e-05, 'epoch': 0.18}) +[2026-10-03 19:25:34,865][starVLA.training.train_starvla][INFO] - Step 2469, Loss: {'train/loss': 0.32239189743995667, 'train/grad_norm_pre_clip': 31.79378890991211, 'timing/data': 0.000275419995887205, 'timing/model': 0.39518299797782674, 'learning_rate/qwen_vl_interface': 5.496568270970096e-06, 'learning_rate/action_model': 5.496568270970096e-05, 'epoch': 0.18}) +[2026-10-03 19:25:35,773][starVLA.training.train_starvla][INFO] - Step 2470, Loss: {'train/loss': 0.004332616925239563, 'train/grad_norm_pre_clip': 8.249336242675781, 'timing/data': 0.00034382997546344995, 'timing/model': 0.39481927899760194, 'learning_rate/qwen_vl_interface': 5.4935269048476015e-06, 'learning_rate/action_model': 5.493526904847601e-05, 'epoch': 0.18}) +[2026-10-03 19:25:36,475][starVLA.training.train_starvla][INFO] - Step 2471, Loss: {'train/loss': 0.25686508417129517, 'train/grad_norm_pre_clip': 33.340110778808594, 'timing/data': 0.00022832999820820987, 'timing/model': 0.3922582489904016, 'learning_rate/qwen_vl_interface': 5.490485438620397e-06, 'learning_rate/action_model': 5.4904854386203965e-05, 'epoch': 0.18}) +[2026-10-03 19:25:37,181][starVLA.training.train_starvla][INFO] - Step 2472, Loss: {'train/loss': 0.0014446747954934835, 'train/grad_norm_pre_clip': 1.807952642440796, 'timing/data': 0.00034840000444091856, 'timing/model': 0.39278148900484666, 'learning_rate/qwen_vl_interface': 5.487443873538712e-06, 'learning_rate/action_model': 5.4874438735387115e-05, 'epoch': 0.18}) +[2026-10-03 19:25:37,886][starVLA.training.train_starvla][INFO] - Step 2473, Loss: {'train/loss': 0.0003717317886184901, 'train/grad_norm_pre_clip': 0.06122744828462601, 'timing/data': 0.00022852999973110855, 'timing/model': 0.39379659900441766, 'learning_rate/qwen_vl_interface': 5.484402210852819e-06, 'learning_rate/action_model': 5.484402210852818e-05, 'epoch': 0.18}) +[2026-10-03 19:25:38,594][starVLA.training.train_starvla][INFO] - Step 2474, Loss: {'train/loss': 0.0038079002406448126, 'train/grad_norm_pre_clip': 2.279855728149414, 'timing/data': 0.0002670900139492005, 'timing/model': 0.39608917900477536, 'learning_rate/qwen_vl_interface': 5.48136045181303e-06, 'learning_rate/action_model': 5.48136045181303e-05, 'epoch': 0.18}) +[2026-10-03 19:25:39,500][starVLA.training.train_starvla][INFO] - Step 2475, Loss: {'train/loss': 0.0047117844223976135, 'train/grad_norm_pre_clip': 4.366096019744873, 'timing/data': 0.00022347999038174748, 'timing/model': 0.5946096519764978, 'learning_rate/qwen_vl_interface': 5.478318597669699e-06, 'learning_rate/action_model': 5.4783185976696995e-05, 'epoch': 0.18}) +[2026-10-03 19:25:40,208][starVLA.training.train_starvla][INFO] - Step 2476, Loss: {'train/loss': 0.4219337999820709, 'train/grad_norm_pre_clip': 46.4656867980957, 'timing/data': 0.00022421000176109374, 'timing/model': 0.3961661979847122, 'learning_rate/qwen_vl_interface': 5.475276649673218e-06, 'learning_rate/action_model': 5.4752766496732174e-05, 'epoch': 0.18}) +[2026-10-03 19:25:40,915][starVLA.training.train_starvla][INFO] - Step 2477, Loss: {'train/loss': 0.552380383014679, 'train/grad_norm_pre_clip': 91.84040832519531, 'timing/data': 0.0003402199945412576, 'timing/model': 0.39510241901734844, 'learning_rate/qwen_vl_interface': 5.472234609074014e-06, 'learning_rate/action_model': 5.472234609074014e-05, 'epoch': 0.18}) +[2026-10-03 19:25:41,620][starVLA.training.train_starvla][INFO] - Step 2478, Loss: {'train/loss': 0.2581457793712616, 'train/grad_norm_pre_clip': 49.805938720703125, 'timing/data': 0.00032233999809250236, 'timing/model': 0.39581654898938723, 'learning_rate/qwen_vl_interface': 5.46919247712256e-06, 'learning_rate/action_model': 5.46919247712256e-05, 'epoch': 0.18}) +[2026-10-03 19:25:42,328][starVLA.training.train_starvla][INFO] - Step 2479, Loss: {'train/loss': 0.06877533346414566, 'train/grad_norm_pre_clip': 29.31841278076172, 'timing/data': 0.00022682000417262316, 'timing/model': 0.3953287180047482, 'learning_rate/qwen_vl_interface': 5.466150255069356e-06, 'learning_rate/action_model': 5.4661502550693555e-05, 'epoch': 0.18}) +[2026-10-03 19:25:43,232][starVLA.training.train_starvla][INFO] - Step 2480, Loss: {'train/loss': 0.0164340790361166, 'train/grad_norm_pre_clip': 16.55284881591797, 'timing/data': 0.0002298800100106746, 'timing/model': 0.5923763629980385, 'learning_rate/qwen_vl_interface': 5.463107944164947e-06, 'learning_rate/action_model': 5.4631079441649466e-05, 'epoch': 0.18}) +[2026-10-03 19:25:43,936][starVLA.training.train_starvla][INFO] - Step 2481, Loss: {'train/loss': 0.03074493259191513, 'train/grad_norm_pre_clip': 45.48845291137695, 'timing/data': 0.00040267998701892793, 'timing/model': 0.3933988490025513, 'learning_rate/qwen_vl_interface': 5.46006554565991e-06, 'learning_rate/action_model': 5.4600655456599105e-05, 'epoch': 0.18}) +[2026-10-03 19:25:44,644][starVLA.training.train_starvla][INFO] - Step 2482, Loss: {'train/loss': 0.2683175504207611, 'train/grad_norm_pre_clip': 41.579505920410156, 'timing/data': 0.0003297500079497695, 'timing/model': 0.3964354479976464, 'learning_rate/qwen_vl_interface': 5.457023060804865e-06, 'learning_rate/action_model': 5.457023060804864e-05, 'epoch': 0.18}) +[2026-10-03 19:25:45,349][starVLA.training.train_starvla][INFO] - Step 2483, Loss: {'train/loss': 0.21233956515789032, 'train/grad_norm_pre_clip': 26.003799438476562, 'timing/data': 0.0003108799865003675, 'timing/model': 0.3939342579978984, 'learning_rate/qwen_vl_interface': 5.453980490850459e-06, 'learning_rate/action_model': 5.453980490850459e-05, 'epoch': 0.18}) +[2026-10-03 19:25:46,056][starVLA.training.train_starvla][INFO] - Step 2484, Loss: {'train/loss': 0.000955030438490212, 'train/grad_norm_pre_clip': 0.6979094743728638, 'timing/data': 0.0002639900194481015, 'timing/model': 0.39420742899528705, 'learning_rate/qwen_vl_interface': 5.4509378370473785e-06, 'learning_rate/action_model': 5.450937837047378e-05, 'epoch': 0.18}) +[2026-10-03 19:25:46,762][starVLA.training.train_starvla][INFO] - Step 2485, Loss: {'train/loss': 0.0015386762097477913, 'train/grad_norm_pre_clip': 0.6188929080963135, 'timing/data': 0.0004058399936184287, 'timing/model': 0.39451120799640194, 'learning_rate/qwen_vl_interface': 5.4478951006463424e-06, 'learning_rate/action_model': 5.447895100646342e-05, 'epoch': 0.18}) +[2026-10-03 19:25:47,667][starVLA.training.train_starvla][INFO] - Step 2486, Loss: {'train/loss': 0.008608809672296047, 'train/grad_norm_pre_clip': 15.516626358032227, 'timing/data': 0.0003497099969536066, 'timing/model': 0.39530033798655495, 'learning_rate/qwen_vl_interface': 5.444852282898106e-06, 'learning_rate/action_model': 5.444852282898106e-05, 'epoch': 0.18}) +[2026-10-03 19:25:48,372][starVLA.training.train_starvla][INFO] - Step 2487, Loss: {'train/loss': 0.2843179702758789, 'train/grad_norm_pre_clip': 65.36376953125, 'timing/data': 0.00033509000786580145, 'timing/model': 0.39392582900472917, 'learning_rate/qwen_vl_interface': 5.441809385053456e-06, 'learning_rate/action_model': 5.4418093850534555e-05, 'epoch': 0.18}) +[2026-10-03 19:25:49,078][starVLA.training.train_starvla][INFO] - Step 2488, Loss: {'train/loss': 0.31265851855278015, 'train/grad_norm_pre_clip': 139.34844970703125, 'timing/data': 0.000335279997671023, 'timing/model': 0.39438332800637, 'learning_rate/qwen_vl_interface': 5.438766408363214e-06, 'learning_rate/action_model': 5.438766408363214e-05, 'epoch': 0.18}) +[2026-10-03 19:25:49,784][starVLA.training.train_starvla][INFO] - Step 2489, Loss: {'train/loss': 0.04585197567939758, 'train/grad_norm_pre_clip': 23.09530258178711, 'timing/data': 0.0004484899982344359, 'timing/model': 0.39514540901291184, 'learning_rate/qwen_vl_interface': 5.43572335407823e-06, 'learning_rate/action_model': 5.43572335407823e-05, 'epoch': 0.18}) +[2026-10-03 19:25:50,489][starVLA.training.train_starvla][INFO] - Step 2490, Loss: {'train/loss': 0.0008842717506922781, 'train/grad_norm_pre_clip': 0.37181001901626587, 'timing/data': 0.0003092800034210086, 'timing/model': 0.393892407999374, 'learning_rate/qwen_vl_interface': 5.432680223449392e-06, 'learning_rate/action_model': 5.4326802234493924e-05, 'epoch': 0.18}) +[2026-10-03 19:25:51,397][starVLA.training.train_starvla][INFO] - Step 2491, Loss: {'train/loss': 0.0004349332593847066, 'train/grad_norm_pre_clip': 0.15815801918506622, 'timing/data': 0.0003795099910348654, 'timing/model': 0.3950430980185047, 'learning_rate/qwen_vl_interface': 5.429637017727614e-06, 'learning_rate/action_model': 5.4296370177276144e-05, 'epoch': 0.18}) +[2026-10-03 19:25:52,103][starVLA.training.train_starvla][INFO] - Step 2492, Loss: {'train/loss': 0.00020296784350648522, 'train/grad_norm_pre_clip': 0.04574138671159744, 'timing/data': 0.00028080999618396163, 'timing/model': 0.3944782879843842, 'learning_rate/qwen_vl_interface': 5.426593738163845e-06, 'learning_rate/action_model': 5.4265937381638446e-05, 'epoch': 0.18}) +[2026-10-03 19:25:52,809][starVLA.training.train_starvla][INFO] - Step 2493, Loss: {'train/loss': 0.002251153811812401, 'train/grad_norm_pre_clip': 0.5609415769577026, 'timing/data': 0.0003182500076945871, 'timing/model': 0.3954346090031322, 'learning_rate/qwen_vl_interface': 5.42355038600906e-06, 'learning_rate/action_model': 5.42355038600906e-05, 'epoch': 0.18}) +[2026-10-03 19:25:53,517][starVLA.training.train_starvla][INFO] - Step 2494, Loss: {'train/loss': 0.248798668384552, 'train/grad_norm_pre_clip': 34.760257720947266, 'timing/data': 0.0003023000026587397, 'timing/model': 0.39591697801370174, 'learning_rate/qwen_vl_interface': 5.420506962514267e-06, 'learning_rate/action_model': 5.420506962514267e-05, 'epoch': 0.18}) +[2026-10-03 19:25:54,224][starVLA.training.train_starvla][INFO] - Step 2495, Loss: {'train/loss': 0.24713890254497528, 'train/grad_norm_pre_clip': 39.898956298828125, 'timing/data': 0.0003397100081201643, 'timing/model': 0.3944709490169771, 'learning_rate/qwen_vl_interface': 5.4174634689305015e-06, 'learning_rate/action_model': 5.417463468930501e-05, 'epoch': 0.18}) +[2026-10-03 19:25:55,128][starVLA.training.train_starvla][INFO] - Step 2496, Loss: {'train/loss': 0.18157997727394104, 'train/grad_norm_pre_clip': 76.25777435302734, 'timing/data': 0.00026865999097935855, 'timing/model': 0.594657002016902, 'learning_rate/qwen_vl_interface': 5.41441990650883e-06, 'learning_rate/action_model': 5.4144199065088295e-05, 'epoch': 0.18}) +[2026-10-03 19:25:55,831][starVLA.training.train_starvla][INFO] - Step 2497, Loss: {'train/loss': 0.005526583641767502, 'train/grad_norm_pre_clip': 2.6568400859832764, 'timing/data': 0.00021518001449294388, 'timing/model': 0.3930607179936487, 'learning_rate/qwen_vl_interface': 5.411376276500345e-06, 'learning_rate/action_model': 5.411376276500345e-05, 'epoch': 0.18}) +[2026-10-03 19:25:56,535][starVLA.training.train_starvla][INFO] - Step 2498, Loss: {'train/loss': 0.025869708508253098, 'train/grad_norm_pre_clip': 21.689678192138672, 'timing/data': 0.00031849899096414447, 'timing/model': 0.3929984589922242, 'learning_rate/qwen_vl_interface': 5.408332580156168e-06, 'learning_rate/action_model': 5.408332580156168e-05, 'epoch': 0.18}) +[2026-10-03 19:25:57,239][starVLA.training.train_starvla][INFO] - Step 2499, Loss: {'train/loss': 0.3644176125526428, 'train/grad_norm_pre_clip': 62.95793151855469, 'timing/data': 0.00031023999326862395, 'timing/model': 0.3935024389938917, 'learning_rate/qwen_vl_interface': 5.40528881872745e-06, 'learning_rate/action_model': 5.40528881872745e-05, 'epoch': 0.18}) +[2026-10-03 19:25:57,945][starVLA.training.train_starvla][INFO] - Step 2500, Loss: {'train/loss': 0.005795700009912252, 'train/grad_norm_pre_clip': 5.532949447631836, 'timing/data': 0.0003327699960209429, 'timing/model': 0.3945994279929437, 'learning_rate/qwen_vl_interface': 5.402244993465364e-06, 'learning_rate/action_model': 5.402244993465364e-05, 'epoch': 0.18}) +[2026-10-03 19:25:58,857][starVLA.training.train_starvla][INFO] - Step 2501, Loss: {'train/loss': 0.001035899855196476, 'train/grad_norm_pre_clip': 0.39384186267852783, 'timing/data': 0.00022775999968871474, 'timing/model': 0.6009850809932686, 'learning_rate/qwen_vl_interface': 5.399201105621111e-06, 'learning_rate/action_model': 5.39920110562111e-05, 'epoch': 0.18}) +[2026-10-03 19:25:59,562][starVLA.training.train_starvla][INFO] - Step 2502, Loss: {'train/loss': 0.140734001994133, 'train/grad_norm_pre_clip': 44.516849517822266, 'timing/data': 0.0003603999793995172, 'timing/model': 0.3945796480111312, 'learning_rate/qwen_vl_interface': 5.396157156445918e-06, 'learning_rate/action_model': 5.396157156445918e-05, 'epoch': 0.18}) +[2026-10-03 19:26:00,268][starVLA.training.train_starvla][INFO] - Step 2503, Loss: {'train/loss': 0.051887281239032745, 'train/grad_norm_pre_clip': 79.82466125488281, 'timing/data': 0.00022146000992506742, 'timing/model': 0.3945985180034768, 'learning_rate/qwen_vl_interface': 5.39311314719104e-06, 'learning_rate/action_model': 5.39311314719104e-05, 'epoch': 0.18}) +[2026-10-03 19:26:00,970][starVLA.training.train_starvla][INFO] - Step 2504, Loss: {'train/loss': 0.0017106609884649515, 'train/grad_norm_pre_clip': 0.8472751379013062, 'timing/data': 0.0003366100136190653, 'timing/model': 0.39305980899371207, 'learning_rate/qwen_vl_interface': 5.390069079107755e-06, 'learning_rate/action_model': 5.390069079107755e-05, 'epoch': 0.18}) +[2026-10-03 19:26:01,674][starVLA.training.train_starvla][INFO] - Step 2505, Loss: {'train/loss': 0.00018054925021715462, 'train/grad_norm_pre_clip': 0.09914000332355499, 'timing/data': 0.00029955999343656003, 'timing/model': 0.39492339800926857, 'learning_rate/qwen_vl_interface': 5.387024953447361e-06, 'learning_rate/action_model': 5.387024953447361e-05, 'epoch': 0.18}) +[2026-10-03 19:26:02,379][starVLA.training.train_starvla][INFO] - Step 2506, Loss: {'train/loss': 6.157396273920313e-05, 'train/grad_norm_pre_clip': 0.011487728916108608, 'timing/data': 0.000362069986294955, 'timing/model': 0.3946672589809168, 'learning_rate/qwen_vl_interface': 5.383980771461184e-06, 'learning_rate/action_model': 5.3839807714611846e-05, 'epoch': 0.18}) +[2026-10-03 19:26:03,286][starVLA.training.train_starvla][INFO] - Step 2507, Loss: {'train/loss': 0.27235305309295654, 'train/grad_norm_pre_clip': 97.53860473632812, 'timing/data': 0.00022943998919799924, 'timing/model': 0.3936695879965555, 'learning_rate/qwen_vl_interface': 5.380936534400572e-06, 'learning_rate/action_model': 5.3809365344005726e-05, 'epoch': 0.18}) +[2026-10-03 19:26:03,991][starVLA.training.train_starvla][INFO] - Step 2508, Loss: {'train/loss': 0.010554742999374866, 'train/grad_norm_pre_clip': 4.191257953643799, 'timing/data': 0.00030114001128822565, 'timing/model': 0.3935346379876137, 'learning_rate/qwen_vl_interface': 5.377892243516899e-06, 'learning_rate/action_model': 5.3778922435168985e-05, 'epoch': 0.18}) +[2026-10-03 19:26:04,694][starVLA.training.train_starvla][INFO] - Step 2509, Loss: {'train/loss': 0.0771440863609314, 'train/grad_norm_pre_clip': 158.31768798828125, 'timing/data': 0.0002241400070488453, 'timing/model': 0.39350721798837185, 'learning_rate/qwen_vl_interface': 5.3748479000615525e-06, 'learning_rate/action_model': 5.374847900061553e-05, 'epoch': 0.18}) +[2026-10-03 19:26:05,400][starVLA.training.train_starvla][INFO] - Step 2510, Loss: {'train/loss': 0.10809154808521271, 'train/grad_norm_pre_clip': 24.1557674407959, 'timing/data': 0.00043290999019518495, 'timing/model': 0.3938314889965113, 'learning_rate/qwen_vl_interface': 5.371803505285954e-06, 'learning_rate/action_model': 5.371803505285954e-05, 'epoch': 0.18}) +[2026-10-03 19:26:06,106][starVLA.training.train_starvla][INFO] - Step 2511, Loss: {'train/loss': 0.0002134406240656972, 'train/grad_norm_pre_clip': 0.08059526234865189, 'timing/data': 0.00021667999681085348, 'timing/model': 0.3950840280158445, 'learning_rate/qwen_vl_interface': 5.368759060441534e-06, 'learning_rate/action_model': 5.368759060441534e-05, 'epoch': 0.18}) +[2026-10-03 19:26:07,032][starVLA.training.train_starvla][INFO] - Step 2512, Loss: {'train/loss': 0.7382007837295532, 'train/grad_norm_pre_clip': 63.41746520996094, 'timing/data': 0.0003446099872235209, 'timing/model': 0.39516719797393307, 'learning_rate/qwen_vl_interface': 5.365714566779749e-06, 'learning_rate/action_model': 5.3657145667797484e-05, 'epoch': 0.18}) +[2026-10-03 19:26:07,739][starVLA.training.train_starvla][INFO] - Step 2513, Loss: {'train/loss': 0.09945910423994064, 'train/grad_norm_pre_clip': 85.69805908203125, 'timing/data': 0.00023955001961439848, 'timing/model': 0.395380807982292, 'learning_rate/qwen_vl_interface': 5.362670025552077e-06, 'learning_rate/action_model': 5.362670025552077e-05, 'epoch': 0.18}) +[2026-10-03 19:26:08,445][starVLA.training.train_starvla][INFO] - Step 2514, Loss: {'train/loss': 0.0018293559551239014, 'train/grad_norm_pre_clip': 0.5426062345504761, 'timing/data': 0.00031756999669596553, 'timing/model': 0.3957943779823836, 'learning_rate/qwen_vl_interface': 5.359625438010013e-06, 'learning_rate/action_model': 5.359625438010013e-05, 'epoch': 0.18}) +[2026-10-03 19:26:09,152][starVLA.training.train_starvla][INFO] - Step 2515, Loss: {'train/loss': 0.0014666840434074402, 'train/grad_norm_pre_clip': 0.5550246238708496, 'timing/data': 0.0002562600129749626, 'timing/model': 0.3946649679855909, 'learning_rate/qwen_vl_interface': 5.356580805405075e-06, 'learning_rate/action_model': 5.356580805405075e-05, 'epoch': 0.18}) +[2026-10-03 19:26:09,857][starVLA.training.train_starvla][INFO] - Step 2516, Loss: {'train/loss': 0.003445695387199521, 'train/grad_norm_pre_clip': 1.451316475868225, 'timing/data': 0.00022176001220941544, 'timing/model': 0.39430130799883045, 'learning_rate/qwen_vl_interface': 5.353536128988791e-06, 'learning_rate/action_model': 5.353536128988791e-05, 'epoch': 0.18}) +[2026-10-03 19:26:10,766][starVLA.training.train_starvla][INFO] - Step 2517, Loss: {'train/loss': 0.004769005347043276, 'train/grad_norm_pre_clip': 5.749854564666748, 'timing/data': 0.00030362000688910484, 'timing/model': 0.5973025819985196, 'learning_rate/qwen_vl_interface': 5.350491410012717e-06, 'learning_rate/action_model': 5.350491410012717e-05, 'epoch': 0.18}) +[2026-10-03 19:26:11,475][starVLA.training.train_starvla][INFO] - Step 2518, Loss: {'train/loss': 0.062285929918289185, 'train/grad_norm_pre_clip': 29.300758361816406, 'timing/data': 0.0002204999909736216, 'timing/model': 0.39573525797459297, 'learning_rate/qwen_vl_interface': 5.347446649728419e-06, 'learning_rate/action_model': 5.347446649728418e-05, 'epoch': 0.18}) +[2026-10-03 19:26:12,178][starVLA.training.train_starvla][INFO] - Step 2519, Loss: {'train/loss': 0.0019148981664329767, 'train/grad_norm_pre_clip': 0.7559702396392822, 'timing/data': 0.00021284000831656158, 'timing/model': 0.3932606680027675, 'learning_rate/qwen_vl_interface': 5.344401849387484e-06, 'learning_rate/action_model': 5.344401849387484e-05, 'epoch': 0.18}) +[2026-10-03 19:26:12,881][starVLA.training.train_starvla][INFO] - Step 2520, Loss: {'train/loss': 0.13482394814491272, 'train/grad_norm_pre_clip': 49.03018569946289, 'timing/data': 0.00024602998746559024, 'timing/model': 0.3931928880047053, 'learning_rate/qwen_vl_interface': 5.3413570102415144e-06, 'learning_rate/action_model': 5.3413570102415144e-05, 'epoch': 0.18}) +[2026-10-03 19:26:13,584][starVLA.training.train_starvla][INFO] - Step 2521, Loss: {'train/loss': 0.00036106916377320886, 'train/grad_norm_pre_clip': 0.136857271194458, 'timing/data': 0.0002134400128852576, 'timing/model': 0.39410936800413765, 'learning_rate/qwen_vl_interface': 5.338312133542131e-06, 'learning_rate/action_model': 5.3383121335421305e-05, 'epoch': 0.18}) +[2026-10-03 19:26:14,491][starVLA.training.train_starvla][INFO] - Step 2522, Loss: {'train/loss': 0.00012193724978715181, 'train/grad_norm_pre_clip': 0.07531770318746567, 'timing/data': 0.00034163001691922545, 'timing/model': 0.5955006320145912, 'learning_rate/qwen_vl_interface': 5.335267220540963e-06, 'learning_rate/action_model': 5.335267220540963e-05, 'epoch': 0.18}) +[2026-10-03 19:26:15,194][starVLA.training.train_starvla][INFO] - Step 2523, Loss: {'train/loss': 0.540153443813324, 'train/grad_norm_pre_clip': 80.80786895751953, 'timing/data': 0.00032262998865917325, 'timing/model': 0.3937901080062147, 'learning_rate/qwen_vl_interface': 5.332222272489662e-06, 'learning_rate/action_model': 5.332222272489662e-05, 'epoch': 0.18}) +[2026-10-03 19:26:15,899][starVLA.training.train_starvla][INFO] - Step 2524, Loss: {'train/loss': 0.00019185090786777437, 'train/grad_norm_pre_clip': 0.07458394765853882, 'timing/data': 0.0003332499763928354, 'timing/model': 0.3939490480115637, 'learning_rate/qwen_vl_interface': 5.3291772906398895e-06, 'learning_rate/action_model': 5.32917729063989e-05, 'epoch': 0.18}) +[2026-10-03 19:26:16,604][starVLA.training.train_starvla][INFO] - Step 2525, Loss: {'train/loss': 0.05214860290288925, 'train/grad_norm_pre_clip': 20.44580078125, 'timing/data': 0.00034381001023575664, 'timing/model': 0.39384526800131425, 'learning_rate/qwen_vl_interface': 5.326132276243327e-06, 'learning_rate/action_model': 5.3261322762433266e-05, 'epoch': 0.18}) +[2026-10-03 19:26:17,306][starVLA.training.train_starvla][INFO] - Step 2526, Loss: {'train/loss': 0.5007067322731018, 'train/grad_norm_pre_clip': 44.16910171508789, 'timing/data': 0.0003021789889317006, 'timing/model': 0.39316670899279416, 'learning_rate/qwen_vl_interface': 5.323087230551659e-06, 'learning_rate/action_model': 5.323087230551659e-05, 'epoch': 0.18}) +[2026-10-03 19:26:18,009][starVLA.training.train_starvla][INFO] - Step 2527, Loss: {'train/loss': 0.8349846601486206, 'train/grad_norm_pre_clip': 103.8849868774414, 'timing/data': 0.0002538299886509776, 'timing/model': 0.39397419898887165, 'learning_rate/qwen_vl_interface': 5.3200421548165916e-06, 'learning_rate/action_model': 5.3200421548165914e-05, 'epoch': 0.18}) +[2026-10-03 19:26:18,919][starVLA.training.train_starvla][INFO] - Step 2528, Loss: {'train/loss': 0.0005047953454777598, 'train/grad_norm_pre_clip': 0.3062240779399872, 'timing/data': 0.0003138400206808001, 'timing/model': 0.393670867983019, 'learning_rate/qwen_vl_interface': 5.3169970502898405e-06, 'learning_rate/action_model': 5.31699705028984e-05, 'epoch': 0.19}) +[2026-10-03 19:26:19,620][starVLA.training.train_starvla][INFO] - Step 2529, Loss: {'train/loss': 0.34484827518463135, 'train/grad_norm_pre_clip': 62.828346252441406, 'timing/data': 0.0003097600128967315, 'timing/model': 0.3920705479977187, 'learning_rate/qwen_vl_interface': 5.31395191822313e-06, 'learning_rate/action_model': 5.31395191822313e-05, 'epoch': 0.19}) +[2026-10-03 19:26:20,328][starVLA.training.train_starvla][INFO] - Step 2530, Loss: {'train/loss': 0.2691381871700287, 'train/grad_norm_pre_clip': 78.36546325683594, 'timing/data': 0.0003193700104020536, 'timing/model': 0.395170178002445, 'learning_rate/qwen_vl_interface': 5.310906759868205e-06, 'learning_rate/action_model': 5.310906759868205e-05, 'epoch': 0.19}) +[2026-10-03 19:26:21,033][starVLA.training.train_starvla][INFO] - Step 2531, Loss: {'train/loss': 0.00017475646745879203, 'train/grad_norm_pre_clip': 0.07606729120016098, 'timing/data': 0.00037675999919883907, 'timing/model': 0.3940245470148511, 'learning_rate/qwen_vl_interface': 5.30786157647681e-06, 'learning_rate/action_model': 5.30786157647681e-05, 'epoch': 0.19}) +[2026-10-03 19:26:21,738][starVLA.training.train_starvla][INFO] - Step 2532, Loss: {'train/loss': 0.03864797204732895, 'train/grad_norm_pre_clip': 62.313350677490234, 'timing/data': 0.0003126900119241327, 'timing/model': 0.393957327993121, 'learning_rate/qwen_vl_interface': 5.304816369300709e-06, 'learning_rate/action_model': 5.304816369300709e-05, 'epoch': 0.19}) +[2026-10-03 19:26:22,636][starVLA.training.train_starvla][INFO] - Step 2533, Loss: {'train/loss': 0.20481622219085693, 'train/grad_norm_pre_clip': 167.8332977294922, 'timing/data': 0.00029461001395247877, 'timing/model': 0.5888439120026305, 'learning_rate/qwen_vl_interface': 5.301771139591666e-06, 'learning_rate/action_model': 5.301771139591666e-05, 'epoch': 0.19}) +[2026-10-03 19:26:23,341][starVLA.training.train_starvla][INFO] - Step 2534, Loss: {'train/loss': 0.0009492994868196547, 'train/grad_norm_pre_clip': 0.3528423607349396, 'timing/data': 0.00034339999547228217, 'timing/model': 0.3944709880161099, 'learning_rate/qwen_vl_interface': 5.298725888601465e-06, 'learning_rate/action_model': 5.298725888601465e-05, 'epoch': 0.19}) +[2026-10-03 19:26:24,046][starVLA.training.train_starvla][INFO] - Step 2535, Loss: {'train/loss': 0.015540366061031818, 'train/grad_norm_pre_clip': 17.150100708007812, 'timing/data': 0.00030194001737982035, 'timing/model': 0.39541272798669524, 'learning_rate/qwen_vl_interface': 5.295680617581891e-06, 'learning_rate/action_model': 5.2956806175818904e-05, 'epoch': 0.19}) +[2026-10-03 19:26:24,750][starVLA.training.train_starvla][INFO] - Step 2536, Loss: {'train/loss': 0.4118689298629761, 'train/grad_norm_pre_clip': 62.08186721801758, 'timing/data': 0.00026639000861905515, 'timing/model': 0.3934602479857858, 'learning_rate/qwen_vl_interface': 5.2926353277847395e-06, 'learning_rate/action_model': 5.292635327784739e-05, 'epoch': 0.19}) +[2026-10-03 19:26:25,455][starVLA.training.train_starvla][INFO] - Step 2537, Loss: {'train/loss': 0.3841794729232788, 'train/grad_norm_pre_clip': 95.23253631591797, 'timing/data': 0.0003396800020709634, 'timing/model': 0.39438584801973775, 'learning_rate/qwen_vl_interface': 5.2895900204618175e-06, 'learning_rate/action_model': 5.289590020461818e-05, 'epoch': 0.19}) +[2026-10-03 19:26:26,364][starVLA.training.train_starvla][INFO] - Step 2538, Loss: {'train/loss': 0.00028839652077294886, 'train/grad_norm_pre_clip': 0.08496331423521042, 'timing/data': 0.00029252999229356647, 'timing/model': 0.5982273220142815, 'learning_rate/qwen_vl_interface': 5.286544696864933e-06, 'learning_rate/action_model': 5.2865446968649324e-05, 'epoch': 0.19}) +[2026-10-03 19:26:27,067][starVLA.training.train_starvla][INFO] - Step 2539, Loss: {'train/loss': 0.4876171052455902, 'train/grad_norm_pre_clip': 55.06654357910156, 'timing/data': 0.0002859800006262958, 'timing/model': 0.39334242799668573, 'learning_rate/qwen_vl_interface': 5.283499358245904e-06, 'learning_rate/action_model': 5.2834993582459034e-05, 'epoch': 0.19}) +[2026-10-03 19:26:27,774][starVLA.training.train_starvla][INFO] - Step 2540, Loss: {'train/loss': 0.2991465628147125, 'train/grad_norm_pre_clip': 104.57675170898438, 'timing/data': 0.00031616000342182815, 'timing/model': 0.39489934701123275, 'learning_rate/qwen_vl_interface': 5.280454005856553e-06, 'learning_rate/action_model': 5.2804540058565534e-05, 'epoch': 0.19}) +[2026-10-03 19:26:28,480][starVLA.training.train_starvla][INFO] - Step 2541, Loss: {'train/loss': 0.16889570653438568, 'train/grad_norm_pre_clip': 54.06768798828125, 'timing/data': 0.00026504998095333576, 'timing/model': 0.39527820798684843, 'learning_rate/qwen_vl_interface': 5.277408640948711e-06, 'learning_rate/action_model': 5.277408640948711e-05, 'epoch': 0.19}) +[2026-10-03 19:26:29,185][starVLA.training.train_starvla][INFO] - Step 2542, Loss: {'train/loss': 0.10393457114696503, 'train/grad_norm_pre_clip': 47.64955520629883, 'timing/data': 0.00037692999467253685, 'timing/model': 0.3953789680090267, 'learning_rate/qwen_vl_interface': 5.274363264774213e-06, 'learning_rate/action_model': 5.274363264774212e-05, 'epoch': 0.19}) +[2026-10-03 19:26:29,893][starVLA.training.train_starvla][INFO] - Step 2543, Loss: {'train/loss': 0.017286619171500206, 'train/grad_norm_pre_clip': 50.240135192871094, 'timing/data': 0.0003216199984308332, 'timing/model': 0.39479871798539534, 'learning_rate/qwen_vl_interface': 5.271317878584897e-06, 'learning_rate/action_model': 5.2713178785848974e-05, 'epoch': 0.19}) +[2026-10-03 19:26:30,800][starVLA.training.train_starvla][INFO] - Step 2544, Loss: {'train/loss': 0.6104320883750916, 'train/grad_norm_pre_clip': 59.37712860107422, 'timing/data': 0.00025110002025030553, 'timing/model': 0.39581190698663704, 'learning_rate/qwen_vl_interface': 5.268272483632606e-06, 'learning_rate/action_model': 5.268272483632606e-05, 'epoch': 0.19}) +[2026-10-03 19:26:31,506][starVLA.training.train_starvla][INFO] - Step 2545, Loss: {'train/loss': 0.21459726989269257, 'train/grad_norm_pre_clip': 57.92522048950195, 'timing/data': 0.0003647400008048862, 'timing/model': 0.39446425798814744, 'learning_rate/qwen_vl_interface': 5.265227081169187e-06, 'learning_rate/action_model': 5.265227081169186e-05, 'epoch': 0.19}) +[2026-10-03 19:26:32,214][starVLA.training.train_starvla][INFO] - Step 2546, Loss: {'train/loss': 0.013258565217256546, 'train/grad_norm_pre_clip': 4.13801383972168, 'timing/data': 0.00032531999750062823, 'timing/model': 0.3956160270026885, 'learning_rate/qwen_vl_interface': 5.2621816724464895e-06, 'learning_rate/action_model': 5.262181672446489e-05, 'epoch': 0.19}) +[2026-10-03 19:26:32,920][starVLA.training.train_starvla][INFO] - Step 2547, Loss: {'train/loss': 0.010853445157408714, 'train/grad_norm_pre_clip': 5.537743091583252, 'timing/data': 0.0003167400136590004, 'timing/model': 0.39454193800338544, 'learning_rate/qwen_vl_interface': 5.259136258716367e-06, 'learning_rate/action_model': 5.2591362587163675e-05, 'epoch': 0.19}) +[2026-10-03 19:26:33,625][starVLA.training.train_starvla][INFO] - Step 2548, Loss: {'train/loss': 0.0002471663465257734, 'train/grad_norm_pre_clip': 0.06021248549222946, 'timing/data': 0.00028843001928180456, 'timing/model': 0.3944038180052303, 'learning_rate/qwen_vl_interface': 5.256090841230675e-06, 'learning_rate/action_model': 5.256090841230674e-05, 'epoch': 0.19}) +[2026-10-03 19:26:34,537][starVLA.training.train_starvla][INFO] - Step 2549, Loss: {'train/loss': 0.06987594813108444, 'train/grad_norm_pre_clip': 35.06999588012695, 'timing/data': 0.00029747001826763153, 'timing/model': 0.3953905880043749, 'learning_rate/qwen_vl_interface': 5.253045421241266e-06, 'learning_rate/action_model': 5.253045421241266e-05, 'epoch': 0.19}) +[2026-10-03 19:26:35,244][starVLA.training.train_starvla][INFO] - Step 2550, Loss: {'train/loss': 0.0004065979737788439, 'train/grad_norm_pre_clip': 0.11357401311397552, 'timing/data': 0.00030461998539976776, 'timing/model': 0.39444762899074703, 'learning_rate/qwen_vl_interface': 5.2500000000000006e-06, 'learning_rate/action_model': 5.25e-05, 'epoch': 0.19}) +[2026-10-03 19:26:35,948][starVLA.training.train_starvla][INFO] - Step 2551, Loss: {'train/loss': 0.0037814718671143055, 'train/grad_norm_pre_clip': 0.7400572896003723, 'timing/data': 0.00021813999046571553, 'timing/model': 0.3938942780077923, 'learning_rate/qwen_vl_interface': 5.246954578758735e-06, 'learning_rate/action_model': 5.246954578758735e-05, 'epoch': 0.19}) +[2026-10-03 19:26:36,655][starVLA.training.train_starvla][INFO] - Step 2552, Loss: {'train/loss': 0.0019093353766947985, 'train/grad_norm_pre_clip': 1.202154278755188, 'timing/data': 0.00022144001559354365, 'timing/model': 0.39541979698697105, 'learning_rate/qwen_vl_interface': 5.243909158769327e-06, 'learning_rate/action_model': 5.2439091587693264e-05, 'epoch': 0.19}) +[2026-10-03 19:26:37,363][starVLA.training.train_starvla][INFO] - Step 2553, Loss: {'train/loss': 0.028001222759485245, 'train/grad_norm_pre_clip': 16.76594352722168, 'timing/data': 0.0002700400073081255, 'timing/model': 0.3957341180066578, 'learning_rate/qwen_vl_interface': 5.240863741283634e-06, 'learning_rate/action_model': 5.2408637412836336e-05, 'epoch': 0.19}) +[2026-10-03 19:26:38,268][starVLA.training.train_starvla][INFO] - Step 2554, Loss: {'train/loss': 0.0674225464463234, 'train/grad_norm_pre_clip': 34.206398010253906, 'timing/data': 0.00031165999826043844, 'timing/model': 0.5937243119988125, 'learning_rate/qwen_vl_interface': 5.2378183275535125e-06, 'learning_rate/action_model': 5.237818327553512e-05, 'epoch': 0.19}) +[2026-10-03 19:26:38,974][starVLA.training.train_starvla][INFO] - Step 2555, Loss: {'train/loss': 0.15862169861793518, 'train/grad_norm_pre_clip': 76.2203598022461, 'timing/data': 0.0002898800012189895, 'timing/model': 0.39520513798925094, 'learning_rate/qwen_vl_interface': 5.234772918830815e-06, 'learning_rate/action_model': 5.234772918830815e-05, 'epoch': 0.19}) +[2026-10-03 19:26:39,678][starVLA.training.train_starvla][INFO] - Step 2556, Loss: {'train/loss': 0.26914289593696594, 'train/grad_norm_pre_clip': 38.43190383911133, 'timing/data': 0.000330720009515062, 'timing/model': 0.39333403899217956, 'learning_rate/qwen_vl_interface': 5.231727516367395e-06, 'learning_rate/action_model': 5.231727516367395e-05, 'epoch': 0.19}) +[2026-10-03 19:26:40,385][starVLA.training.train_starvla][INFO] - Step 2557, Loss: {'train/loss': 0.12713880836963654, 'train/grad_norm_pre_clip': 154.00901794433594, 'timing/data': 0.00036970002111047506, 'timing/model': 0.39453791797859594, 'learning_rate/qwen_vl_interface': 5.228682121415104e-06, 'learning_rate/action_model': 5.228682121415104e-05, 'epoch': 0.19}) +[2026-10-03 19:26:41,092][starVLA.training.train_starvla][INFO] - Step 2558, Loss: {'train/loss': 0.09543512761592865, 'train/grad_norm_pre_clip': 42.175045013427734, 'timing/data': 0.0003475499979685992, 'timing/model': 0.39468150699394755, 'learning_rate/qwen_vl_interface': 5.225636735225789e-06, 'learning_rate/action_model': 5.225636735225788e-05, 'epoch': 0.19}) +[2026-10-03 19:26:41,997][starVLA.training.train_starvla][INFO] - Step 2559, Loss: {'train/loss': 0.0677104964852333, 'train/grad_norm_pre_clip': 64.57434844970703, 'timing/data': 0.0003307599981781095, 'timing/model': 0.5932453020068351, 'learning_rate/qwen_vl_interface': 5.22259135905129e-06, 'learning_rate/action_model': 5.2225913590512895e-05, 'epoch': 0.19}) +[2026-10-03 19:26:42,704][starVLA.training.train_starvla][INFO] - Step 2560, Loss: {'train/loss': 0.3478355407714844, 'train/grad_norm_pre_clip': 32.60725784301758, 'timing/data': 0.00029098999220877886, 'timing/model': 0.3949655869801063, 'learning_rate/qwen_vl_interface': 5.219545994143448e-06, 'learning_rate/action_model': 5.219545994143448e-05, 'epoch': 0.19}) +[2026-10-03 19:26:43,407][starVLA.training.train_starvla][INFO] - Step 2561, Loss: {'train/loss': 0.7787854671478271, 'train/grad_norm_pre_clip': 62.52663803100586, 'timing/data': 0.00021892000222578645, 'timing/model': 0.39414057799149305, 'learning_rate/qwen_vl_interface': 5.216500641754097e-06, 'learning_rate/action_model': 5.216500641754097e-05, 'epoch': 0.19}) +[2026-10-03 19:26:44,116][starVLA.training.train_starvla][INFO] - Step 2562, Loss: {'train/loss': 0.5824146866798401, 'train/grad_norm_pre_clip': 77.96454620361328, 'timing/data': 0.0002972699876409024, 'timing/model': 0.39468161802506074, 'learning_rate/qwen_vl_interface': 5.213455303135067e-06, 'learning_rate/action_model': 5.2134553031350673e-05, 'epoch': 0.19}) +[2026-10-03 19:26:44,823][starVLA.training.train_starvla][INFO] - Step 2563, Loss: {'train/loss': 0.005130986217409372, 'train/grad_norm_pre_clip': 2.7411603927612305, 'timing/data': 0.0002352599985897541, 'timing/model': 0.39484600801370107, 'learning_rate/qwen_vl_interface': 5.2104099795381836e-06, 'learning_rate/action_model': 5.210409979538183e-05, 'epoch': 0.19}) +[2026-10-03 19:26:45,732][starVLA.training.train_starvla][INFO] - Step 2564, Loss: {'train/loss': 0.20252938568592072, 'train/grad_norm_pre_clip': 54.20920944213867, 'timing/data': 0.00034634998883120716, 'timing/model': 0.59683449100703, 'learning_rate/qwen_vl_interface': 5.207364672215261e-06, 'learning_rate/action_model': 5.2073646722152604e-05, 'epoch': 0.19}) +[2026-10-03 19:26:46,438][starVLA.training.train_starvla][INFO] - Step 2565, Loss: {'train/loss': 0.0017310235416516662, 'train/grad_norm_pre_clip': 0.5163535475730896, 'timing/data': 0.00026736001018434763, 'timing/model': 0.39430443799938075, 'learning_rate/qwen_vl_interface': 5.2043193824181095e-06, 'learning_rate/action_model': 5.2043193824181094e-05, 'epoch': 0.19}) +[2026-10-03 19:26:47,145][starVLA.training.train_starvla][INFO] - Step 2566, Loss: {'train/loss': 0.1722986400127411, 'train/grad_norm_pre_clip': 70.5914306640625, 'timing/data': 0.0002928100002463907, 'timing/model': 0.3945463380077854, 'learning_rate/qwen_vl_interface': 5.201274111398535e-06, 'learning_rate/action_model': 5.2012741113985356e-05, 'epoch': 0.19}) +[2026-10-03 19:26:47,849][starVLA.training.train_starvla][INFO] - Step 2567, Loss: {'train/loss': 0.0037051215767860413, 'train/grad_norm_pre_clip': 1.0048727989196777, 'timing/data': 0.00021839002147316933, 'timing/model': 0.3945030869799666, 'learning_rate/qwen_vl_interface': 5.198228860408334e-06, 'learning_rate/action_model': 5.198228860408334e-05, 'epoch': 0.19}) +[2026-10-03 19:26:48,553][starVLA.training.train_starvla][INFO] - Step 2568, Loss: {'train/loss': 0.34286248683929443, 'train/grad_norm_pre_clip': 47.704837799072266, 'timing/data': 0.00021815000218339264, 'timing/model': 0.39337339700432494, 'learning_rate/qwen_vl_interface': 5.195183630699292e-06, 'learning_rate/action_model': 5.195183630699292e-05, 'epoch': 0.19}) +[2026-10-03 19:26:49,259][starVLA.training.train_starvla][INFO] - Step 2569, Loss: {'train/loss': 0.0021659107878804207, 'train/grad_norm_pre_clip': 1.6202921867370605, 'timing/data': 0.0003071000101044774, 'timing/model': 0.3938704879838042, 'learning_rate/qwen_vl_interface': 5.192138423523189e-06, 'learning_rate/action_model': 5.1921384235231895e-05, 'epoch': 0.19}) +[2026-10-03 19:26:50,181][starVLA.training.train_starvla][INFO] - Step 2570, Loss: {'train/loss': 0.02374539151787758, 'train/grad_norm_pre_clip': 26.613496780395508, 'timing/data': 0.00032715999986976385, 'timing/model': 0.39389721798943356, 'learning_rate/qwen_vl_interface': 5.189093240131795e-06, 'learning_rate/action_model': 5.189093240131795e-05, 'epoch': 0.19}) +[2026-10-03 19:26:50,885][starVLA.training.train_starvla][INFO] - Step 2571, Loss: {'train/loss': 0.0007409048266708851, 'train/grad_norm_pre_clip': 0.20599599182605743, 'timing/data': 0.00022033997811377048, 'timing/model': 0.3934262080001645, 'learning_rate/qwen_vl_interface': 5.186048081776869e-06, 'learning_rate/action_model': 5.1860480817768685e-05, 'epoch': 0.19}) +[2026-10-03 19:26:51,591][starVLA.training.train_starvla][INFO] - Step 2572, Loss: {'train/loss': 0.17058195173740387, 'train/grad_norm_pre_clip': 150.89805603027344, 'timing/data': 0.00033968998468481004, 'timing/model': 0.39434760800213553, 'learning_rate/qwen_vl_interface': 5.183002949710161e-06, 'learning_rate/action_model': 5.18300294971016e-05, 'epoch': 0.19}) +[2026-10-03 19:26:52,294][starVLA.training.train_starvla][INFO] - Step 2573, Loss: {'train/loss': 0.003940375987440348, 'train/grad_norm_pre_clip': 2.2760987281799316, 'timing/data': 0.00026394001906737685, 'timing/model': 0.3929762880143244, 'learning_rate/qwen_vl_interface': 5.179957845183411e-06, 'learning_rate/action_model': 5.179957845183411e-05, 'epoch': 0.19}) +[2026-10-03 19:26:52,998][starVLA.training.train_starvla][INFO] - Step 2574, Loss: {'train/loss': 0.007676251698285341, 'train/grad_norm_pre_clip': 15.777488708496094, 'timing/data': 0.0002469099999871105, 'timing/model': 0.3934178780182265, 'learning_rate/qwen_vl_interface': 5.176912769448343e-06, 'learning_rate/action_model': 5.176912769448343e-05, 'epoch': 0.19}) +[2026-10-03 19:26:53,900][starVLA.training.train_starvla][INFO] - Step 2575, Loss: {'train/loss': 0.0003155411686748266, 'train/grad_norm_pre_clip': 0.16298645734786987, 'timing/data': 0.00022360999719239771, 'timing/model': 0.5899986520234961, 'learning_rate/qwen_vl_interface': 5.173867723756676e-06, 'learning_rate/action_model': 5.173867723756676e-05, 'epoch': 0.19}) +[2026-10-03 19:26:54,604][starVLA.training.train_starvla][INFO] - Step 2576, Loss: {'train/loss': 0.003861218923702836, 'train/grad_norm_pre_clip': 2.788583517074585, 'timing/data': 0.0003293899935670197, 'timing/model': 0.3942241079930682, 'learning_rate/qwen_vl_interface': 5.170822709360112e-06, 'learning_rate/action_model': 5.170822709360111e-05, 'epoch': 0.19}) +[2026-10-03 19:26:55,306][starVLA.training.train_starvla][INFO] - Step 2577, Loss: {'train/loss': 0.1637374609708786, 'train/grad_norm_pre_clip': 66.18154907226562, 'timing/data': 0.00020634999964386225, 'timing/model': 0.39282562799053267, 'learning_rate/qwen_vl_interface': 5.16777772751034e-06, 'learning_rate/action_model': 5.16777772751034e-05, 'epoch': 0.19}) +[2026-10-03 19:26:56,010][starVLA.training.train_starvla][INFO] - Step 2578, Loss: {'train/loss': 0.1916685253381729, 'train/grad_norm_pre_clip': 30.17135238647461, 'timing/data': 0.00025877999723888934, 'timing/model': 0.3943512469995767, 'learning_rate/qwen_vl_interface': 5.16473277945904e-06, 'learning_rate/action_model': 5.1647327794590396e-05, 'epoch': 0.19}) +[2026-10-03 19:26:56,712][starVLA.training.train_starvla][INFO] - Step 2579, Loss: {'train/loss': 0.03328828513622284, 'train/grad_norm_pre_clip': 21.0822696685791, 'timing/data': 0.00034119997872039676, 'timing/model': 0.392318197991699, 'learning_rate/qwen_vl_interface': 5.161687866457872e-06, 'learning_rate/action_model': 5.161687866457871e-05, 'epoch': 0.19}) +[2026-10-03 19:26:57,415][starVLA.training.train_starvla][INFO] - Step 2580, Loss: {'train/loss': 0.0004016732273157686, 'train/grad_norm_pre_clip': 0.15913093090057373, 'timing/data': 0.00020845001563429832, 'timing/model': 0.3930206980148796, 'learning_rate/qwen_vl_interface': 5.158642989758487e-06, 'learning_rate/action_model': 5.158642989758487e-05, 'epoch': 0.19}) +[2026-10-03 19:26:58,321][starVLA.training.train_starvla][INFO] - Step 2581, Loss: {'train/loss': 0.004156123846769333, 'train/grad_norm_pre_clip': 2.913721799850464, 'timing/data': 0.0002630499948281795, 'timing/model': 0.39496109797619283, 'learning_rate/qwen_vl_interface': 5.155598150612519e-06, 'learning_rate/action_model': 5.1555981506125185e-05, 'epoch': 0.19}) +[2026-10-03 19:26:59,025][starVLA.training.train_starvla][INFO] - Step 2582, Loss: {'train/loss': 0.45567941665649414, 'train/grad_norm_pre_clip': 37.88605880737305, 'timing/data': 0.0002499900001566857, 'timing/model': 0.39288570699864067, 'learning_rate/qwen_vl_interface': 5.152553350271584e-06, 'learning_rate/action_model': 5.152553350271584e-05, 'epoch': 0.19}) +[2026-10-03 19:26:59,729][starVLA.training.train_starvla][INFO] - Step 2583, Loss: {'train/loss': 0.008918305858969688, 'train/grad_norm_pre_clip': 6.314220428466797, 'timing/data': 0.00024957998539321125, 'timing/model': 0.39360394701361656, 'learning_rate/qwen_vl_interface': 5.149508589987286e-06, 'learning_rate/action_model': 5.1495085899872855e-05, 'epoch': 0.19}) +[2026-10-03 19:27:00,433][starVLA.training.train_starvla][INFO] - Step 2584, Loss: {'train/loss': 0.011148715391755104, 'train/grad_norm_pre_clip': 11.613740921020508, 'timing/data': 0.0002936599776148796, 'timing/model': 0.3934481379983481, 'learning_rate/qwen_vl_interface': 5.14646387101121e-06, 'learning_rate/action_model': 5.14646387101121e-05, 'epoch': 0.19}) +[2026-10-03 19:27:01,136][starVLA.training.train_starvla][INFO] - Step 2585, Loss: {'train/loss': 0.059882521629333496, 'train/grad_norm_pre_clip': 92.34850311279297, 'timing/data': 0.00022851000539958477, 'timing/model': 0.393787898006849, 'learning_rate/qwen_vl_interface': 5.143419194594928e-06, 'learning_rate/action_model': 5.143419194594927e-05, 'epoch': 0.19}) +[2026-10-03 19:27:02,034][starVLA.training.train_starvla][INFO] - Step 2586, Loss: {'train/loss': 0.13135533034801483, 'train/grad_norm_pre_clip': 52.381160736083984, 'timing/data': 0.00025497001479379833, 'timing/model': 0.3938863080111332, 'learning_rate/qwen_vl_interface': 5.140374561989988e-06, 'learning_rate/action_model': 5.140374561989988e-05, 'epoch': 0.19}) +[2026-10-03 19:27:02,735][starVLA.training.train_starvla][INFO] - Step 2587, Loss: {'train/loss': 0.6731899380683899, 'train/grad_norm_pre_clip': 212.3142547607422, 'timing/data': 0.0003262699756305665, 'timing/model': 0.3925223379919771, 'learning_rate/qwen_vl_interface': 5.137329974447924e-06, 'learning_rate/action_model': 5.137329974447924e-05, 'epoch': 0.19}) +[2026-10-03 19:27:03,440][starVLA.training.train_starvla][INFO] - Step 2588, Loss: {'train/loss': 0.06972882151603699, 'train/grad_norm_pre_clip': 41.19728088378906, 'timing/data': 0.0003202499938197434, 'timing/model': 0.392919398000231, 'learning_rate/qwen_vl_interface': 5.1342854332202516e-06, 'learning_rate/action_model': 5.1342854332202514e-05, 'epoch': 0.19}) +[2026-10-03 19:27:04,147][starVLA.training.train_starvla][INFO] - Step 2589, Loss: {'train/loss': 0.11198627948760986, 'train/grad_norm_pre_clip': 47.76884841918945, 'timing/data': 0.00041024998063221574, 'timing/model': 0.3951781179930549, 'learning_rate/qwen_vl_interface': 5.131240939558467e-06, 'learning_rate/action_model': 5.131240939558467e-05, 'epoch': 0.19}) +[2026-10-03 19:27:04,853][starVLA.training.train_starvla][INFO] - Step 2590, Loss: {'train/loss': 0.0054048942402005196, 'train/grad_norm_pre_clip': 13.423993110656738, 'timing/data': 0.000304340006550774, 'timing/model': 0.39445983699988574, 'learning_rate/qwen_vl_interface': 5.128196494714048e-06, 'learning_rate/action_model': 5.128196494714047e-05, 'epoch': 0.19}) +[2026-10-03 19:27:05,758][starVLA.training.train_starvla][INFO] - Step 2591, Loss: {'train/loss': 0.025938350707292557, 'train/grad_norm_pre_clip': 12.356569290161133, 'timing/data': 0.0003257300122641027, 'timing/model': 0.5943886710156221, 'learning_rate/qwen_vl_interface': 5.125152099938448e-06, 'learning_rate/action_model': 5.125152099938447e-05, 'epoch': 0.19}) +[2026-10-03 19:27:06,464][starVLA.training.train_starvla][INFO] - Step 2592, Loss: {'train/loss': 7.82735733082518e-05, 'train/grad_norm_pre_clip': 0.02767724171280861, 'timing/data': 0.00021703998208977282, 'timing/model': 0.3938548779988196, 'learning_rate/qwen_vl_interface': 5.122107756483104e-06, 'learning_rate/action_model': 5.122107756483103e-05, 'epoch': 0.19}) +[2026-10-03 19:27:07,170][starVLA.training.train_starvla][INFO] - Step 2593, Loss: {'train/loss': 0.04410418123006821, 'train/grad_norm_pre_clip': 47.798465728759766, 'timing/data': 0.00022763002198189497, 'timing/model': 0.3951315679878462, 'learning_rate/qwen_vl_interface': 5.119063465599429e-06, 'learning_rate/action_model': 5.1190634655994285e-05, 'epoch': 0.19}) +[2026-10-03 19:27:07,875][starVLA.training.train_starvla][INFO] - Step 2594, Loss: {'train/loss': 0.30673307180404663, 'train/grad_norm_pre_clip': 69.84695434570312, 'timing/data': 0.00031659999513067305, 'timing/model': 0.39504369799396954, 'learning_rate/qwen_vl_interface': 5.116019228538818e-06, 'learning_rate/action_model': 5.116019228538818e-05, 'epoch': 0.19}) +[2026-10-03 19:27:08,581][starVLA.training.train_starvla][INFO] - Step 2595, Loss: {'train/loss': 0.011200129054486752, 'train/grad_norm_pre_clip': 12.548883438110352, 'timing/data': 0.00023420999059453607, 'timing/model': 0.39518072799546644, 'learning_rate/qwen_vl_interface': 5.112975046552641e-06, 'learning_rate/action_model': 5.1129750465526404e-05, 'epoch': 0.19}) +[2026-10-03 19:27:09,286][starVLA.training.train_starvla][INFO] - Step 2596, Loss: {'train/loss': 0.5063183307647705, 'train/grad_norm_pre_clip': 43.84904479980469, 'timing/data': 0.0003527100197970867, 'timing/model': 0.3935004769882653, 'learning_rate/qwen_vl_interface': 5.109930920892246e-06, 'learning_rate/action_model': 5.1099309208922464e-05, 'epoch': 0.19}) +[2026-10-03 19:27:10,192][starVLA.training.train_starvla][INFO] - Step 2597, Loss: {'train/loss': 0.0005991349462419748, 'train/grad_norm_pre_clip': 0.645994246006012, 'timing/data': 0.00023895001504570246, 'timing/model': 0.39592921698931605, 'learning_rate/qwen_vl_interface': 5.10688685280896e-06, 'learning_rate/action_model': 5.10688685280896e-05, 'epoch': 0.19}) +[2026-10-03 19:27:10,897][starVLA.training.train_starvla][INFO] - Step 2598, Loss: {'train/loss': 0.044239360839128494, 'train/grad_norm_pre_clip': 95.3212661743164, 'timing/data': 0.0003719099913723767, 'timing/model': 0.3946158070175443, 'learning_rate/qwen_vl_interface': 5.103842843554082e-06, 'learning_rate/action_model': 5.103842843554082e-05, 'epoch': 0.19}) +[2026-10-03 19:27:11,600][starVLA.training.train_starvla][INFO] - Step 2599, Loss: {'train/loss': 0.0008290471159853041, 'train/grad_norm_pre_clip': 0.5360466837882996, 'timing/data': 0.0002098300028592348, 'timing/model': 0.393115597020369, 'learning_rate/qwen_vl_interface': 5.100798894378891e-06, 'learning_rate/action_model': 5.100798894378891e-05, 'epoch': 0.19}) +[2026-10-03 19:27:12,303][starVLA.training.train_starvla][INFO] - Step 2600, Loss: {'train/loss': 0.02369537390768528, 'train/grad_norm_pre_clip': 38.61875915527344, 'timing/data': 0.00020482999389059842, 'timing/model': 0.39290340698789805, 'learning_rate/qwen_vl_interface': 5.097755006534638e-06, 'learning_rate/action_model': 5.097755006534638e-05, 'epoch': 0.19}) +[2026-10-03 19:27:13,006][starVLA.training.train_starvla][INFO] - Step 2601, Loss: {'train/loss': 0.09406594932079315, 'train/grad_norm_pre_clip': 163.46157836914062, 'timing/data': 0.00021025000023655593, 'timing/model': 0.3948438969964627, 'learning_rate/qwen_vl_interface': 5.094711181272551e-06, 'learning_rate/action_model': 5.094711181272551e-05, 'epoch': 0.19}) +[2026-10-03 19:27:13,908][starVLA.training.train_starvla][INFO] - Step 2602, Loss: {'train/loss': 0.11102838814258575, 'train/grad_norm_pre_clip': 95.2452621459961, 'timing/data': 0.0003570300177671015, 'timing/model': 0.3905347580148373, 'learning_rate/qwen_vl_interface': 5.091667419843833e-06, 'learning_rate/action_model': 5.091667419843833e-05, 'epoch': 0.19}) +[2026-10-03 19:27:14,609][starVLA.training.train_starvla][INFO] - Step 2603, Loss: {'train/loss': 0.230580672621727, 'train/grad_norm_pre_clip': 52.19829177856445, 'timing/data': 0.00021855998784303665, 'timing/model': 0.3926294870034326, 'learning_rate/qwen_vl_interface': 5.088623723499657e-06, 'learning_rate/action_model': 5.088623723499657e-05, 'epoch': 0.19}) +[2026-10-03 19:27:15,310][starVLA.training.train_starvla][INFO] - Step 2604, Loss: {'train/loss': 0.07355352491140366, 'train/grad_norm_pre_clip': 65.951416015625, 'timing/data': 0.00039852000190876424, 'timing/model': 0.3928255169885233, 'learning_rate/qwen_vl_interface': 5.085580093491171e-06, 'learning_rate/action_model': 5.085580093491171e-05, 'epoch': 0.19}) +[2026-10-03 19:27:16,017][starVLA.training.train_starvla][INFO] - Step 2605, Loss: {'train/loss': 0.0009509561350569129, 'train/grad_norm_pre_clip': 0.3556566536426544, 'timing/data': 0.00022221001563593745, 'timing/model': 0.3964512070233468, 'learning_rate/qwen_vl_interface': 5.0825365310695e-06, 'learning_rate/action_model': 5.082536531069499e-05, 'epoch': 0.19}) +[2026-10-03 19:27:16,719][starVLA.training.train_starvla][INFO] - Step 2606, Loss: {'train/loss': 0.0462266281247139, 'train/grad_norm_pre_clip': 40.130531311035156, 'timing/data': 0.00021009999909438193, 'timing/model': 0.39399657800095156, 'learning_rate/qwen_vl_interface': 5.079493037485735e-06, 'learning_rate/action_model': 5.0794930374857344e-05, 'epoch': 0.19}) +[2026-10-03 19:27:17,624][starVLA.training.train_starvla][INFO] - Step 2607, Loss: {'train/loss': 0.0006756479851901531, 'train/grad_norm_pre_clip': 0.15322807431221008, 'timing/data': 0.00021295002079568803, 'timing/model': 0.5953148710250389, 'learning_rate/qwen_vl_interface': 5.076449613990941e-06, 'learning_rate/action_model': 5.076449613990941e-05, 'epoch': 0.19}) +[2026-10-03 19:27:18,326][starVLA.training.train_starvla][INFO] - Step 2608, Loss: {'train/loss': 0.15338970720767975, 'train/grad_norm_pre_clip': 100.7817153930664, 'timing/data': 0.00021507000201381743, 'timing/model': 0.39174172800267115, 'learning_rate/qwen_vl_interface': 5.073406261836157e-06, 'learning_rate/action_model': 5.0734062618361565e-05, 'epoch': 0.19}) +[2026-10-03 19:27:19,030][starVLA.training.train_starvla][INFO] - Step 2609, Loss: {'train/loss': 0.2866400182247162, 'train/grad_norm_pre_clip': 90.50286865234375, 'timing/data': 0.00021553001715801656, 'timing/model': 0.3959452470007818, 'learning_rate/qwen_vl_interface': 5.070362982272386e-06, 'learning_rate/action_model': 5.070362982272385e-05, 'epoch': 0.19}) +[2026-10-03 19:27:19,735][starVLA.training.train_starvla][INFO] - Step 2610, Loss: {'train/loss': 0.4800873398780823, 'train/grad_norm_pre_clip': 38.90647506713867, 'timing/data': 0.0002765190147329122, 'timing/model': 0.39332526799989864, 'learning_rate/qwen_vl_interface': 5.067319776550607e-06, 'learning_rate/action_model': 5.067319776550607e-05, 'epoch': 0.19}) +[2026-10-03 19:27:20,437][starVLA.training.train_starvla][INFO] - Step 2611, Loss: {'train/loss': 0.12736321985721588, 'train/grad_norm_pre_clip': 165.4132080078125, 'timing/data': 0.0002083699801005423, 'timing/model': 0.3935814269934781, 'learning_rate/qwen_vl_interface': 5.064276645921768e-06, 'learning_rate/action_model': 5.064276645921768e-05, 'epoch': 0.19}) +[2026-10-03 19:27:21,347][starVLA.training.train_starvla][INFO] - Step 2612, Loss: {'train/loss': 0.531112551689148, 'train/grad_norm_pre_clip': 98.75965118408203, 'timing/data': 0.00021557000582106411, 'timing/model': 0.5986916399851907, 'learning_rate/qwen_vl_interface': 5.061233591636787e-06, 'learning_rate/action_model': 5.061233591636787e-05, 'epoch': 0.19}) +[2026-10-03 19:27:22,050][starVLA.training.train_starvla][INFO] - Step 2613, Loss: {'train/loss': 0.651306688785553, 'train/grad_norm_pre_clip': 85.46676635742188, 'timing/data': 0.00021006001043133438, 'timing/model': 0.39456620599958114, 'learning_rate/qwen_vl_interface': 5.058190614946544e-06, 'learning_rate/action_model': 5.058190614946544e-05, 'epoch': 0.19}) +[2026-10-03 19:27:22,750][starVLA.training.train_starvla][INFO] - Step 2614, Loss: {'train/loss': 0.0008511393680237234, 'train/grad_norm_pre_clip': 0.42444413900375366, 'timing/data': 0.000330720009515062, 'timing/model': 0.39139671699376777, 'learning_rate/qwen_vl_interface': 5.055147717101895e-06, 'learning_rate/action_model': 5.055147717101894e-05, 'epoch': 0.19}) +[2026-10-03 19:27:23,450][starVLA.training.train_starvla][INFO] - Step 2615, Loss: {'train/loss': 0.09893766045570374, 'train/grad_norm_pre_clip': 82.1723403930664, 'timing/data': 0.00020298000890761614, 'timing/model': 0.3913989970169496, 'learning_rate/qwen_vl_interface': 5.052104899353658e-06, 'learning_rate/action_model': 5.0521048993536577e-05, 'epoch': 0.19}) +[2026-10-03 19:27:24,153][starVLA.training.train_starvla][INFO] - Step 2616, Loss: {'train/loss': 0.010403775610029697, 'train/grad_norm_pre_clip': 18.989177703857422, 'timing/data': 0.000264729984337464, 'timing/model': 0.3916851870017126, 'learning_rate/qwen_vl_interface': 5.049062162952622e-06, 'learning_rate/action_model': 5.0490621629526214e-05, 'epoch': 0.19}) +[2026-10-03 19:27:25,064][starVLA.training.train_starvla][INFO] - Step 2617, Loss: {'train/loss': 0.0006713871262036264, 'train/grad_norm_pre_clip': 0.3657882809638977, 'timing/data': 0.00019982000230811536, 'timing/model': 0.6019036710204091, 'learning_rate/qwen_vl_interface': 5.046019509149541e-06, 'learning_rate/action_model': 5.046019509149541e-05, 'epoch': 0.19}) +[2026-10-03 19:27:25,766][starVLA.training.train_starvla][INFO] - Step 2618, Loss: {'train/loss': 0.4844910204410553, 'train/grad_norm_pre_clip': 98.05619049072266, 'timing/data': 0.0003842189908027649, 'timing/model': 0.3920267579960637, 'learning_rate/qwen_vl_interface': 5.042976939195135e-06, 'learning_rate/action_model': 5.0429769391951355e-05, 'epoch': 0.19}) +[2026-10-03 19:27:26,470][starVLA.training.train_starvla][INFO] - Step 2619, Loss: {'train/loss': 0.00016882273484952748, 'train/grad_norm_pre_clip': 0.08461638540029526, 'timing/data': 0.0002112099900841713, 'timing/model': 0.3937705270072911, 'learning_rate/qwen_vl_interface': 5.039934454340091e-06, 'learning_rate/action_model': 5.0399344543400907e-05, 'epoch': 0.19}) +[2026-10-03 19:27:27,172][starVLA.training.train_starvla][INFO] - Step 2620, Loss: {'train/loss': 0.12652678787708282, 'train/grad_norm_pre_clip': 165.7439727783203, 'timing/data': 0.00021342997206375003, 'timing/model': 0.39203725699917413, 'learning_rate/qwen_vl_interface': 5.036892055835056e-06, 'learning_rate/action_model': 5.0368920558350566e-05, 'epoch': 0.19}) +[2026-10-03 19:27:27,877][starVLA.training.train_starvla][INFO] - Step 2621, Loss: {'train/loss': 0.4166147708892822, 'train/grad_norm_pre_clip': 41.63926315307617, 'timing/data': 0.0002169499930460006, 'timing/model': 0.3960505059803836, 'learning_rate/qwen_vl_interface': 5.033849744930647e-06, 'learning_rate/action_model': 5.033849744930646e-05, 'epoch': 0.19}) +[2026-10-03 19:27:28,582][starVLA.training.train_starvla][INFO] - Step 2622, Loss: {'train/loss': 0.002167660277336836, 'train/grad_norm_pre_clip': 0.959078311920166, 'timing/data': 0.00022091000573709607, 'timing/model': 0.3939976569963619, 'learning_rate/qwen_vl_interface': 5.030807522877443e-06, 'learning_rate/action_model': 5.030807522877443e-05, 'epoch': 0.19}) +[2026-10-03 19:27:29,493][starVLA.training.train_starvla][INFO] - Step 2623, Loss: {'train/loss': 0.00018181520863436162, 'train/grad_norm_pre_clip': 0.09916027635335922, 'timing/data': 0.0002196200075559318, 'timing/model': 0.3932099470112007, 'learning_rate/qwen_vl_interface': 5.027765390925987e-06, 'learning_rate/action_model': 5.027765390925987e-05, 'epoch': 0.19}) +[2026-10-03 19:27:30,195][starVLA.training.train_starvla][INFO] - Step 2624, Loss: {'train/loss': 0.00020986725576221943, 'train/grad_norm_pre_clip': 0.08359333127737045, 'timing/data': 0.00025246001314371824, 'timing/model': 0.3932530879974365, 'learning_rate/qwen_vl_interface': 5.024723350326783e-06, 'learning_rate/action_model': 5.024723350326783e-05, 'epoch': 0.19}) +[2026-10-03 19:27:30,900][starVLA.training.train_starvla][INFO] - Step 2625, Loss: {'train/loss': 0.1889994591474533, 'train/grad_norm_pre_clip': 63.603843688964844, 'timing/data': 0.00031415000557899475, 'timing/model': 0.39469316697795875, 'learning_rate/qwen_vl_interface': 5.021681402330303e-06, 'learning_rate/action_model': 5.021681402330303e-05, 'epoch': 0.19}) +[2026-10-03 19:27:31,604][starVLA.training.train_starvla][INFO] - Step 2626, Loss: {'train/loss': 0.0004192385822534561, 'train/grad_norm_pre_clip': 0.27468883991241455, 'timing/data': 0.00025252997875213623, 'timing/model': 0.3931994769955054, 'learning_rate/qwen_vl_interface': 5.018639548186972e-06, 'learning_rate/action_model': 5.0186395481869716e-05, 'epoch': 0.19}) +[2026-10-03 19:27:32,309][starVLA.training.train_starvla][INFO] - Step 2627, Loss: {'train/loss': 0.0004526556294877082, 'train/grad_norm_pre_clip': 0.32668524980545044, 'timing/data': 0.000315180019242689, 'timing/model': 0.3931612070009578, 'learning_rate/qwen_vl_interface': 5.015597789147184e-06, 'learning_rate/action_model': 5.015597789147184e-05, 'epoch': 0.19}) +[2026-10-03 19:27:33,211][starVLA.training.train_starvla][INFO] - Step 2628, Loss: {'train/loss': 0.14317667484283447, 'train/grad_norm_pre_clip': 36.5018196105957, 'timing/data': 0.0003556700248736888, 'timing/model': 0.5911645710002631, 'learning_rate/qwen_vl_interface': 5.012556126461291e-06, 'learning_rate/action_model': 5.012556126461291e-05, 'epoch': 0.19}) +[2026-10-03 19:27:33,912][starVLA.training.train_starvla][INFO] - Step 2629, Loss: {'train/loss': 0.0004385260690469295, 'train/grad_norm_pre_clip': 0.13487683236598969, 'timing/data': 0.0002565000031609088, 'timing/model': 0.3933762270025909, 'learning_rate/qwen_vl_interface': 5.0095145613796044e-06, 'learning_rate/action_model': 5.0095145613796046e-05, 'epoch': 0.19}) +[2026-10-03 19:27:34,615][starVLA.training.train_starvla][INFO] - Step 2630, Loss: {'train/loss': 0.0030392135959118605, 'train/grad_norm_pre_clip': 4.703305244445801, 'timing/data': 0.00037748998147435486, 'timing/model': 0.39208084699930623, 'learning_rate/qwen_vl_interface': 5.0064730951524e-06, 'learning_rate/action_model': 5.0064730951524e-05, 'epoch': 0.19}) +[2026-10-03 19:27:35,317][starVLA.training.train_starvla][INFO] - Step 2631, Loss: {'train/loss': 0.0004960206570103765, 'train/grad_norm_pre_clip': 0.11567176878452301, 'timing/data': 0.00034853999386541545, 'timing/model': 0.39363453700207174, 'learning_rate/qwen_vl_interface': 5.003431729029905e-06, 'learning_rate/action_model': 5.003431729029905e-05, 'epoch': 0.19}) +[2026-10-03 19:27:36,022][starVLA.training.train_starvla][INFO] - Step 2632, Loss: {'train/loss': 2.7886413590749726e-05, 'train/grad_norm_pre_clip': 0.009847983717918396, 'timing/data': 0.00038372998824343085, 'timing/model': 0.3945336870092433, 'learning_rate/qwen_vl_interface': 5.000390464262314e-06, 'learning_rate/action_model': 5.0003904642623136e-05, 'epoch': 0.19}) +[2026-10-03 19:27:36,725][starVLA.training.train_starvla][INFO] - Step 2633, Loss: {'train/loss': 0.0929817259311676, 'train/grad_norm_pre_clip': 84.73223876953125, 'timing/data': 0.00030682000215165317, 'timing/model': 0.3947537069907412, 'learning_rate/qwen_vl_interface': 4.997349302099774e-06, 'learning_rate/action_model': 4.997349302099773e-05, 'epoch': 0.19}) +[2026-10-03 19:27:37,625][starVLA.training.train_starvla][INFO] - Step 2634, Loss: {'train/loss': 0.21307718753814697, 'train/grad_norm_pre_clip': 87.25554656982422, 'timing/data': 0.0002037599915638566, 'timing/model': 0.3915279580105562, 'learning_rate/qwen_vl_interface': 4.994308243792391e-06, 'learning_rate/action_model': 4.994308243792391e-05, 'epoch': 0.19}) +[2026-10-03 19:27:38,322][starVLA.training.train_starvla][INFO] - Step 2635, Loss: {'train/loss': 0.00044457046897150576, 'train/grad_norm_pre_clip': 0.7001916766166687, 'timing/data': 0.00020246999338269234, 'timing/model': 0.3895089270081371, 'learning_rate/qwen_vl_interface': 4.991267290590233e-06, 'learning_rate/action_model': 4.991267290590233e-05, 'epoch': 0.19}) +[2026-10-03 19:27:39,021][starVLA.training.train_starvla][INFO] - Step 2636, Loss: {'train/loss': 6.501619645860046e-05, 'train/grad_norm_pre_clip': 0.03132473677396774, 'timing/data': 0.00033358001383021474, 'timing/model': 0.3896857179934159, 'learning_rate/qwen_vl_interface': 4.988226443743317e-06, 'learning_rate/action_model': 4.9882264437433174e-05, 'epoch': 0.19}) +[2026-10-03 19:27:39,722][starVLA.training.train_starvla][INFO] - Step 2637, Loss: {'train/loss': 0.0024171234108507633, 'train/grad_norm_pre_clip': 1.2096396684646606, 'timing/data': 0.00020282997866161168, 'timing/model': 0.39367590798065066, 'learning_rate/qwen_vl_interface': 4.985185704501623e-06, 'learning_rate/action_model': 4.985185704501623e-05, 'epoch': 0.19}) +[2026-10-03 19:27:40,428][starVLA.training.train_starvla][INFO] - Step 2638, Loss: {'train/loss': 1.0357756614685059, 'train/grad_norm_pre_clip': 115.35322570800781, 'timing/data': 0.00022106998949311674, 'timing/model': 0.3942400169908069, 'learning_rate/qwen_vl_interface': 4.982145074115082e-06, 'learning_rate/action_model': 4.982145074115082e-05, 'epoch': 0.19}) +[2026-10-03 19:27:41,339][starVLA.training.train_starvla][INFO] - Step 2639, Loss: {'train/loss': 0.0011680935276672244, 'train/grad_norm_pre_clip': 1.2505173683166504, 'timing/data': 0.0003643800155259669, 'timing/model': 0.39248091698391363, 'learning_rate/qwen_vl_interface': 4.979104553833584e-06, 'learning_rate/action_model': 4.979104553833584e-05, 'epoch': 0.19}) +[2026-10-03 19:27:42,044][starVLA.training.train_starvla][INFO] - Step 2640, Loss: {'train/loss': 0.002592545235529542, 'train/grad_norm_pre_clip': 4.358904838562012, 'timing/data': 0.00034417997812852263, 'timing/model': 0.39319055702071637, 'learning_rate/qwen_vl_interface': 4.9760641449069724e-06, 'learning_rate/action_model': 4.976064144906972e-05, 'epoch': 0.19}) +[2026-10-03 19:27:42,747][starVLA.training.train_starvla][INFO] - Step 2641, Loss: {'train/loss': 0.027502767741680145, 'train/grad_norm_pre_clip': 30.06403923034668, 'timing/data': 0.0003329099854454398, 'timing/model': 0.3936163370090071, 'learning_rate/qwen_vl_interface': 4.973023848585042e-06, 'learning_rate/action_model': 4.973023848585042e-05, 'epoch': 0.19}) +[2026-10-03 19:27:43,448][starVLA.training.train_starvla][INFO] - Step 2642, Loss: {'train/loss': 6.32507071713917e-05, 'train/grad_norm_pre_clip': 0.034862205386161804, 'timing/data': 0.00020161998691037297, 'timing/model': 0.3918864869920071, 'learning_rate/qwen_vl_interface': 4.969983666117545e-06, 'learning_rate/action_model': 4.9699836661175446e-05, 'epoch': 0.19}) +[2026-10-03 19:27:44,149][starVLA.training.train_starvla][INFO] - Step 2643, Loss: {'train/loss': 0.7481375336647034, 'train/grad_norm_pre_clip': 150.09988403320312, 'timing/data': 0.00021828999160788953, 'timing/model': 0.3926149779872503, 'learning_rate/qwen_vl_interface': 4.966943598754185e-06, 'learning_rate/action_model': 4.9669435987541854e-05, 'epoch': 0.19}) +[2026-10-03 19:27:45,059][starVLA.training.train_starvla][INFO] - Step 2644, Loss: {'train/loss': 0.31365832686424255, 'train/grad_norm_pre_clip': 91.62035369873047, 'timing/data': 0.00022084999363869429, 'timing/model': 0.5980380309920292, 'learning_rate/qwen_vl_interface': 4.963903647744621e-06, 'learning_rate/action_model': 4.96390364774462e-05, 'epoch': 0.19}) +[2026-10-03 19:27:45,762][starVLA.training.train_starvla][INFO] - Step 2645, Loss: {'train/loss': 0.34968623518943787, 'train/grad_norm_pre_clip': 90.39091491699219, 'timing/data': 0.00020968998433090746, 'timing/model': 0.3952014370006509, 'learning_rate/qwen_vl_interface': 4.9608638143384604e-06, 'learning_rate/action_model': 4.960863814338461e-05, 'epoch': 0.19}) +[2026-10-03 19:27:46,463][starVLA.training.train_starvla][INFO] - Step 2646, Loss: {'train/loss': 6.336084334179759e-05, 'train/grad_norm_pre_clip': 0.02385820634663105, 'timing/data': 0.0003317900118418038, 'timing/model': 0.3914035070047248, 'learning_rate/qwen_vl_interface': 4.9578240997852636e-06, 'learning_rate/action_model': 4.957824099785264e-05, 'epoch': 0.19}) +[2026-10-03 19:27:47,166][starVLA.training.train_starvla][INFO] - Step 2647, Loss: {'train/loss': 3.835380630334839e-05, 'train/grad_norm_pre_clip': 0.010564261116087437, 'timing/data': 0.0003413800150156021, 'timing/model': 0.3930300369975157, 'learning_rate/qwen_vl_interface': 4.954784505334544e-06, 'learning_rate/action_model': 4.9547845053345434e-05, 'epoch': 0.19}) +[2026-10-03 19:27:47,870][starVLA.training.train_starvla][INFO] - Step 2648, Loss: {'train/loss': 0.013353154063224792, 'train/grad_norm_pre_clip': 12.313642501831055, 'timing/data': 0.0002284299989696592, 'timing/model': 0.3929747870133724, 'learning_rate/qwen_vl_interface': 4.951745032235762e-06, 'learning_rate/action_model': 4.951745032235762e-05, 'epoch': 0.19}) +[2026-10-03 19:27:48,574][starVLA.training.train_starvla][INFO] - Step 2649, Loss: {'train/loss': 0.292243093252182, 'train/grad_norm_pre_clip': 28.660400390625, 'timing/data': 0.00020049000158905983, 'timing/model': 0.39347411700873636, 'learning_rate/qwen_vl_interface': 4.94870568173833e-06, 'learning_rate/action_model': 4.9487056817383307e-05, 'epoch': 0.19}) +[2026-10-03 19:27:49,474][starVLA.training.train_starvla][INFO] - Step 2650, Loss: {'train/loss': 0.23435193300247192, 'train/grad_norm_pre_clip': 55.85382080078125, 'timing/data': 0.0003621299983933568, 'timing/model': 0.39303867702255957, 'learning_rate/qwen_vl_interface': 4.945666455091614e-06, 'learning_rate/action_model': 4.9456664550916136e-05, 'epoch': 0.19}) +[2026-10-03 19:27:50,174][starVLA.training.train_starvla][INFO] - Step 2651, Loss: {'train/loss': 0.035238608717918396, 'train/grad_norm_pre_clip': 119.01948547363281, 'timing/data': 0.000235140003496781, 'timing/model': 0.39174486699630506, 'learning_rate/qwen_vl_interface': 4.942627353544921e-06, 'learning_rate/action_model': 4.942627353544921e-05, 'epoch': 0.19}) +[2026-10-03 19:27:50,878][starVLA.training.train_starvla][INFO] - Step 2652, Loss: {'train/loss': 0.0017533490899950266, 'train/grad_norm_pre_clip': 1.3272422552108765, 'timing/data': 0.00021160999312996864, 'timing/model': 0.3922981770010665, 'learning_rate/qwen_vl_interface': 4.939588378347513e-06, 'learning_rate/action_model': 4.9395883783475125e-05, 'epoch': 0.19}) +[2026-10-03 19:27:51,582][starVLA.training.train_starvla][INFO] - Step 2653, Loss: {'train/loss': 0.034909557551145554, 'train/grad_norm_pre_clip': 25.139387130737305, 'timing/data': 0.00025359998107887805, 'timing/model': 0.3954268369998317, 'learning_rate/qwen_vl_interface': 4.936549530748597e-06, 'learning_rate/action_model': 4.9365495307485963e-05, 'epoch': 0.19}) +[2026-10-03 19:27:52,285][starVLA.training.train_starvla][INFO] - Step 2654, Loss: {'train/loss': 0.0013005641521885991, 'train/grad_norm_pre_clip': 0.8823345303535461, 'timing/data': 0.0003681399975903332, 'timing/model': 0.39349847700214013, 'learning_rate/qwen_vl_interface': 4.9335108119973286e-06, 'learning_rate/action_model': 4.933510811997328e-05, 'epoch': 0.19}) +[2026-10-03 19:27:53,199][starVLA.training.train_starvla][INFO] - Step 2655, Loss: {'train/loss': 0.004561624024063349, 'train/grad_norm_pre_clip': 4.115696907043457, 'timing/data': 0.00021855998784303665, 'timing/model': 0.3910722370201256, 'learning_rate/qwen_vl_interface': 4.930472223342812e-06, 'learning_rate/action_model': 4.930472223342812e-05, 'epoch': 0.19}) +[2026-10-03 19:27:53,902][starVLA.training.train_starvla][INFO] - Step 2656, Loss: {'train/loss': 0.00027952075470238924, 'train/grad_norm_pre_clip': 0.6127448081970215, 'timing/data': 0.0003405399911571294, 'timing/model': 0.3915414270013571, 'learning_rate/qwen_vl_interface': 4.927433766034093e-06, 'learning_rate/action_model': 4.9274337660340934e-05, 'epoch': 0.19}) +[2026-10-03 19:27:54,603][starVLA.training.train_starvla][INFO] - Step 2657, Loss: {'train/loss': 0.0004484052478801459, 'train/grad_norm_pre_clip': 0.20073546469211578, 'timing/data': 0.0003065300115849823, 'timing/model': 0.39396862700232305, 'learning_rate/qwen_vl_interface': 4.92439544132017e-06, 'learning_rate/action_model': 4.92439544132017e-05, 'epoch': 0.19}) +[2026-10-03 19:27:55,303][starVLA.training.train_starvla][INFO] - Step 2658, Loss: {'train/loss': 2.1121795725775883e-05, 'train/grad_norm_pre_clip': 0.006026925519108772, 'timing/data': 0.00036133898538537323, 'timing/model': 0.3911393980088178, 'learning_rate/qwen_vl_interface': 4.92135725044998e-06, 'learning_rate/action_model': 4.92135725044998e-05, 'epoch': 0.19}) +[2026-10-03 19:27:56,005][starVLA.training.train_starvla][INFO] - Step 2659, Loss: {'train/loss': 0.40285947918891907, 'train/grad_norm_pre_clip': 79.44174194335938, 'timing/data': 0.00022511999122798443, 'timing/model': 0.3929267569910735, 'learning_rate/qwen_vl_interface': 4.918319194672412e-06, 'learning_rate/action_model': 4.918319194672412e-05, 'epoch': 0.19}) +[2026-10-03 19:27:56,910][starVLA.training.train_starvla][INFO] - Step 2660, Loss: {'train/loss': 1.4304890100902412e-05, 'train/grad_norm_pre_clip': 0.0027943523600697517, 'timing/data': 0.00032140000257641077, 'timing/model': 0.5955068899784237, 'learning_rate/qwen_vl_interface': 4.9152812752362935e-06, 'learning_rate/action_model': 4.915281275236293e-05, 'epoch': 0.19}) +[2026-10-03 19:27:57,611][starVLA.training.train_starvla][INFO] - Step 2661, Loss: {'train/loss': 0.0001417621097061783, 'train/grad_norm_pre_clip': 0.10123971104621887, 'timing/data': 0.0002475900109857321, 'timing/model': 0.3941490869910922, 'learning_rate/qwen_vl_interface': 4.9122434933904e-06, 'learning_rate/action_model': 4.9122434933904e-05, 'epoch': 0.19}) +[2026-10-03 19:27:58,315][starVLA.training.train_starvla][INFO] - Step 2662, Loss: {'train/loss': 0.0007814277778379619, 'train/grad_norm_pre_clip': 1.2170816659927368, 'timing/data': 0.00022238001110963523, 'timing/model': 0.39313783700345084, 'learning_rate/qwen_vl_interface': 4.909205850383449e-06, 'learning_rate/action_model': 4.9092058503834485e-05, 'epoch': 0.19}) +[2026-10-03 19:27:59,016][starVLA.training.train_starvla][INFO] - Step 2663, Loss: {'train/loss': 0.001961074536666274, 'train/grad_norm_pre_clip': 0.7790243625640869, 'timing/data': 0.0003334299835842103, 'timing/model': 0.391402397013735, 'learning_rate/qwen_vl_interface': 4.906168347464098e-06, 'learning_rate/action_model': 4.906168347464098e-05, 'epoch': 0.19}) +[2026-10-03 19:27:59,719][starVLA.training.train_starvla][INFO] - Step 2664, Loss: {'train/loss': 0.0002666453365236521, 'train/grad_norm_pre_clip': 0.041329167783260345, 'timing/data': 0.00028832998941652477, 'timing/model': 0.39236838699434884, 'learning_rate/qwen_vl_interface': 4.903130985880955e-06, 'learning_rate/action_model': 4.903130985880954e-05, 'epoch': 0.19}) +[2026-10-03 19:28:00,617][starVLA.training.train_starvla][INFO] - Step 2665, Loss: {'train/loss': 0.0004993798211216927, 'train/grad_norm_pre_clip': 0.4884897768497467, 'timing/data': 0.00026306998915970325, 'timing/model': 0.5912728800030891, 'learning_rate/qwen_vl_interface': 4.900093766882558e-06, 'learning_rate/action_model': 4.900093766882558e-05, 'epoch': 0.2}) +[2026-10-03 19:28:01,316][starVLA.training.train_starvla][INFO] - Step 2666, Loss: {'train/loss': 0.5241783261299133, 'train/grad_norm_pre_clip': 37.30204391479492, 'timing/data': 0.00023194999084807932, 'timing/model': 0.39022567798383534, 'learning_rate/qwen_vl_interface': 4.897056691717397e-06, 'learning_rate/action_model': 4.8970566917173966e-05, 'epoch': 0.2}) +[2026-10-03 19:28:02,013][starVLA.training.train_starvla][INFO] - Step 2667, Loss: {'train/loss': 0.36932581663131714, 'train/grad_norm_pre_clip': 47.36309814453125, 'timing/data': 0.00031759002013131976, 'timing/model': 0.39024782698834315, 'learning_rate/qwen_vl_interface': 4.894019761633898e-06, 'learning_rate/action_model': 4.894019761633898e-05, 'epoch': 0.2}) +[2026-10-03 19:28:02,711][starVLA.training.train_starvla][INFO] - Step 2668, Loss: {'train/loss': 0.12338467687368393, 'train/grad_norm_pre_clip': 57.634464263916016, 'timing/data': 0.00020889998995698988, 'timing/model': 0.3898440669872798, 'learning_rate/qwen_vl_interface': 4.8909829778804304e-06, 'learning_rate/action_model': 4.89098297788043e-05, 'epoch': 0.2}) +[2026-10-03 19:28:03,413][starVLA.training.train_starvla][INFO] - Step 2669, Loss: {'train/loss': 0.5123375058174133, 'train/grad_norm_pre_clip': 55.09625244140625, 'timing/data': 0.0003052300016861409, 'timing/model': 0.39331472699996084, 'learning_rate/qwen_vl_interface': 4.887946341705298e-06, 'learning_rate/action_model': 4.887946341705297e-05, 'epoch': 0.2}) +[2026-10-03 19:28:04,314][starVLA.training.train_starvla][INFO] - Step 2670, Loss: {'train/loss': 0.0005281957564875484, 'train/grad_norm_pre_clip': 0.16596674919128418, 'timing/data': 0.00026116002118214965, 'timing/model': 0.5922343910206109, 'learning_rate/qwen_vl_interface': 4.8849098543567465e-06, 'learning_rate/action_model': 4.884909854356747e-05, 'epoch': 0.2}) +[2026-10-03 19:28:05,013][starVLA.training.train_starvla][INFO] - Step 2671, Loss: {'train/loss': 0.00012256407353561372, 'train/grad_norm_pre_clip': 0.049766965210437775, 'timing/data': 0.00020848997519351542, 'timing/model': 0.39040890699834563, 'learning_rate/qwen_vl_interface': 4.881873517082965e-06, 'learning_rate/action_model': 4.8818735170829655e-05, 'epoch': 0.2}) +[2026-10-03 19:28:05,713][starVLA.training.train_starvla][INFO] - Step 2672, Loss: {'train/loss': 4.859748878516257e-05, 'train/grad_norm_pre_clip': 0.014269324019551277, 'timing/data': 0.00032009001006372273, 'timing/model': 0.3903379979892634, 'learning_rate/qwen_vl_interface': 4.878837331132076e-06, 'learning_rate/action_model': 4.8788373311320754e-05, 'epoch': 0.2}) +[2026-10-03 19:28:06,418][starVLA.training.train_starvla][INFO] - Step 2673, Loss: {'train/loss': 0.00042446955922059715, 'train/grad_norm_pre_clip': 0.7200726270675659, 'timing/data': 0.0002185999765060842, 'timing/model': 0.3954269459936768, 'learning_rate/qwen_vl_interface': 4.87580129775214e-06, 'learning_rate/action_model': 4.87580129775214e-05, 'epoch': 0.2}) +[2026-10-03 19:28:07,118][starVLA.training.train_starvla][INFO] - Step 2674, Loss: {'train/loss': 0.0011585571337491274, 'train/grad_norm_pre_clip': 0.9375646710395813, 'timing/data': 0.00036527999327518046, 'timing/model': 0.3910193670017179, 'learning_rate/qwen_vl_interface': 4.872765418191156e-06, 'learning_rate/action_model': 4.872765418191155e-05, 'epoch': 0.2}) +[2026-10-03 19:28:07,818][starVLA.training.train_starvla][INFO] - Step 2675, Loss: {'train/loss': 0.0016790564404800534, 'train/grad_norm_pre_clip': 1.8102965354919434, 'timing/data': 0.0003333799832034856, 'timing/model': 0.3913195770001039, 'learning_rate/qwen_vl_interface': 4.8697296936970575e-06, 'learning_rate/action_model': 4.869729693697057e-05, 'epoch': 0.2}) +[2026-10-03 19:28:08,718][starVLA.training.train_starvla][INFO] - Step 2676, Loss: {'train/loss': 0.00019847317889798433, 'train/grad_norm_pre_clip': 0.06871666759252548, 'timing/data': 0.0003110000106971711, 'timing/model': 0.39148067697533406, 'learning_rate/qwen_vl_interface': 4.8666941255177185e-06, 'learning_rate/action_model': 4.8666941255177185e-05, 'epoch': 0.2}) +[2026-10-03 19:28:09,418][starVLA.training.train_starvla][INFO] - Step 2677, Loss: {'train/loss': 0.00023417506599798799, 'train/grad_norm_pre_clip': 0.0612909160554409, 'timing/data': 0.0003132099809590727, 'timing/model': 0.3912582869816106, 'learning_rate/qwen_vl_interface': 4.863658714900945e-06, 'learning_rate/action_model': 4.863658714900945e-05, 'epoch': 0.2}) +[2026-10-03 19:28:10,116][starVLA.training.train_starvla][INFO] - Step 2678, Loss: {'train/loss': 0.67435622215271, 'train/grad_norm_pre_clip': 93.67594146728516, 'timing/data': 0.00030559999868273735, 'timing/model': 0.3895055580069311, 'learning_rate/qwen_vl_interface': 4.860623463094483e-06, 'learning_rate/action_model': 4.860623463094483e-05, 'epoch': 0.2}) +[2026-10-03 19:28:10,813][starVLA.training.train_starvla][INFO] - Step 2679, Loss: {'train/loss': 0.0103664705529809, 'train/grad_norm_pre_clip': 11.147177696228027, 'timing/data': 0.0003083500196225941, 'timing/model': 0.39044666799600236, 'learning_rate/qwen_vl_interface': 4.857588371346006e-06, 'learning_rate/action_model': 4.8575883713460057e-05, 'epoch': 0.2}) +[2026-10-03 19:28:11,513][starVLA.training.train_starvla][INFO] - Step 2680, Loss: {'train/loss': 0.0008281386690214276, 'train/grad_norm_pre_clip': 0.4216703772544861, 'timing/data': 0.00023794997832737863, 'timing/model': 0.3901198369858321, 'learning_rate/qwen_vl_interface': 4.854553440903127e-06, 'learning_rate/action_model': 4.854553440903126e-05, 'epoch': 0.2}) +[2026-10-03 19:28:12,416][starVLA.training.train_starvla][INFO] - Step 2681, Loss: {'train/loss': 0.00039778801146894693, 'train/grad_norm_pre_clip': 0.16785438358783722, 'timing/data': 0.00033899000845849514, 'timing/model': 0.3923256369889714, 'learning_rate/qwen_vl_interface': 4.85151867301339e-06, 'learning_rate/action_model': 4.85151867301339e-05, 'epoch': 0.2}) +[2026-10-03 19:28:13,117][starVLA.training.train_starvla][INFO] - Step 2682, Loss: {'train/loss': 0.011112350039184093, 'train/grad_norm_pre_clip': 12.598456382751465, 'timing/data': 0.00033747998531907797, 'timing/model': 0.3915020670101512, 'learning_rate/qwen_vl_interface': 4.848484068924274e-06, 'learning_rate/action_model': 4.848484068924275e-05, 'epoch': 0.2}) +[2026-10-03 19:28:13,817][starVLA.training.train_starvla][INFO] - Step 2683, Loss: {'train/loss': 0.3554251492023468, 'train/grad_norm_pre_clip': 124.2756576538086, 'timing/data': 0.0002061600098386407, 'timing/model': 0.3916383270116057, 'learning_rate/qwen_vl_interface': 4.845449629883193e-06, 'learning_rate/action_model': 4.8454496298831934e-05, 'epoch': 0.2}) +[2026-10-03 19:28:14,517][starVLA.training.train_starvla][INFO] - Step 2684, Loss: {'train/loss': 0.00010129374277312309, 'train/grad_norm_pre_clip': 0.027305779978632927, 'timing/data': 0.00020512001356109977, 'timing/model': 0.3904391569958534, 'learning_rate/qwen_vl_interface': 4.842415357137487e-06, 'learning_rate/action_model': 4.842415357137487e-05, 'epoch': 0.2}) +[2026-10-03 19:28:15,219][starVLA.training.train_starvla][INFO] - Step 2685, Loss: {'train/loss': 5.211146344663575e-05, 'train/grad_norm_pre_clip': 0.016055505722761154, 'timing/data': 0.00030446998425759375, 'timing/model': 0.3933739860076457, 'learning_rate/qwen_vl_interface': 4.839381251934432e-06, 'learning_rate/action_model': 4.8393812519344316e-05, 'epoch': 0.2}) +[2026-10-03 19:28:16,121][starVLA.training.train_starvla][INFO] - Step 2686, Loss: {'train/loss': 0.4769923686981201, 'train/grad_norm_pre_clip': 38.232704162597656, 'timing/data': 0.0002085799933411181, 'timing/model': 0.5945901089871768, 'learning_rate/qwen_vl_interface': 4.836347315521233e-06, 'learning_rate/action_model': 4.836347315521233e-05, 'epoch': 0.2}) +[2026-10-03 19:28:16,820][starVLA.training.train_starvla][INFO] - Step 2687, Loss: {'train/loss': 0.47532185912132263, 'train/grad_norm_pre_clip': 88.00914001464844, 'timing/data': 0.00022475002333521843, 'timing/model': 0.3906472969974857, 'learning_rate/qwen_vl_interface': 4.833313549145029e-06, 'learning_rate/action_model': 4.833313549145029e-05, 'epoch': 0.2}) +[2026-10-03 19:28:17,523][starVLA.training.train_starvla][INFO] - Step 2688, Loss: {'train/loss': 6.758196832379326e-05, 'train/grad_norm_pre_clip': 0.021607553586363792, 'timing/data': 0.00021671000286005437, 'timing/model': 0.3928556270257104, 'learning_rate/qwen_vl_interface': 4.830279954052884e-06, 'learning_rate/action_model': 4.8302799540528836e-05, 'epoch': 0.2}) +[2026-10-03 19:28:18,224][starVLA.training.train_starvla][INFO] - Step 2689, Loss: {'train/loss': 0.00014941538393031806, 'train/grad_norm_pre_clip': 0.04556925594806671, 'timing/data': 0.0003055600100196898, 'timing/model': 0.39294224602053873, 'learning_rate/qwen_vl_interface': 4.827246531491797e-06, 'learning_rate/action_model': 4.827246531491797e-05, 'epoch': 0.2}) +[2026-10-03 19:28:18,923][starVLA.training.train_starvla][INFO] - Step 2690, Loss: {'train/loss': 0.04988698661327362, 'train/grad_norm_pre_clip': 115.8843765258789, 'timing/data': 0.00024085998302325606, 'timing/model': 0.3920605360181071, 'learning_rate/qwen_vl_interface': 4.824213282708691e-06, 'learning_rate/action_model': 4.824213282708691e-05, 'epoch': 0.2}) +[2026-10-03 19:28:19,823][starVLA.training.train_starvla][INFO] - Step 2691, Loss: {'train/loss': 0.36959341168403625, 'train/grad_norm_pre_clip': 35.32330322265625, 'timing/data': 0.00020701001631096005, 'timing/model': 0.591956959979143, 'learning_rate/qwen_vl_interface': 4.821180208950422e-06, 'learning_rate/action_model': 4.821180208950421e-05, 'epoch': 0.2}) +[2026-10-03 19:28:20,523][starVLA.training.train_starvla][INFO] - Step 2692, Loss: {'train/loss': 0.00021603095228783786, 'train/grad_norm_pre_clip': 0.08940932899713516, 'timing/data': 0.00021396001102402806, 'timing/model': 0.391525287006516, 'learning_rate/qwen_vl_interface': 4.81814731146377e-06, 'learning_rate/action_model': 4.8181473114637695e-05, 'epoch': 0.2}) +[2026-10-03 19:28:21,222][starVLA.training.train_starvla][INFO] - Step 2693, Loss: {'train/loss': 0.0001640058180782944, 'train/grad_norm_pre_clip': 0.047665778547525406, 'timing/data': 0.00024478998966515064, 'timing/model': 0.3919093259901274, 'learning_rate/qwen_vl_interface': 4.8151145914954465e-06, 'learning_rate/action_model': 4.815114591495446e-05, 'epoch': 0.2}) +[2026-10-03 19:28:21,922][starVLA.training.train_starvla][INFO] - Step 2694, Loss: {'train/loss': 0.000286282564047724, 'train/grad_norm_pre_clip': 0.10251589119434357, 'timing/data': 0.00037935900036245584, 'timing/model': 0.3915449880005326, 'learning_rate/qwen_vl_interface': 4.812082050292088e-06, 'learning_rate/action_model': 4.8120820502920876e-05, 'epoch': 0.2}) +[2026-10-03 19:28:22,622][starVLA.training.train_starvla][INFO] - Step 2695, Loss: {'train/loss': 0.000614636461250484, 'train/grad_norm_pre_clip': 0.19499193131923676, 'timing/data': 0.0003259499790146947, 'timing/model': 0.3911326770030428, 'learning_rate/qwen_vl_interface': 4.809049689100256e-06, 'learning_rate/action_model': 4.8090496891002555e-05, 'epoch': 0.2}) +[2026-10-03 19:28:23,525][starVLA.training.train_starvla][INFO] - Step 2696, Loss: {'train/loss': 0.0003102736664004624, 'train/grad_norm_pre_clip': 0.16176214814186096, 'timing/data': 0.00033830999745987356, 'timing/model': 0.5939754999999423, 'learning_rate/qwen_vl_interface': 4.80601750916644e-06, 'learning_rate/action_model': 4.8060175091664404e-05, 'epoch': 0.2}) +[2026-10-03 19:28:24,225][starVLA.training.train_starvla][INFO] - Step 2697, Loss: {'train/loss': 0.5712825059890747, 'train/grad_norm_pre_clip': 240.3505401611328, 'timing/data': 0.0003045390185434371, 'timing/model': 0.3927163280022796, 'learning_rate/qwen_vl_interface': 4.802985511737057e-06, 'learning_rate/action_model': 4.802985511737057e-05, 'epoch': 0.2}) +[2026-10-03 19:28:24,921][starVLA.training.train_starvla][INFO] - Step 2698, Loss: {'train/loss': 0.00019702690769918263, 'train/grad_norm_pre_clip': 0.06329580396413803, 'timing/data': 0.0002175999979954213, 'timing/model': 0.389700008003274, 'learning_rate/qwen_vl_interface': 4.799953698058444e-06, 'learning_rate/action_model': 4.7999536980584434e-05, 'epoch': 0.2}) +[2026-10-03 19:28:25,618][starVLA.training.train_starvla][INFO] - Step 2699, Loss: {'train/loss': 0.0665120780467987, 'train/grad_norm_pre_clip': 59.76898193359375, 'timing/data': 0.0003445299807935953, 'timing/model': 0.390151057014009, 'learning_rate/qwen_vl_interface': 4.796922069376867e-06, 'learning_rate/action_model': 4.796922069376867e-05, 'epoch': 0.2}) +[2026-10-03 19:28:26,318][starVLA.training.train_starvla][INFO] - Step 2700, Loss: {'train/loss': 0.00017918687080964446, 'train/grad_norm_pre_clip': 0.055799562484025955, 'timing/data': 0.00033720998908393085, 'timing/model': 0.39115148698328994, 'learning_rate/qwen_vl_interface': 4.793890626938511e-06, 'learning_rate/action_model': 4.793890626938511e-05, 'epoch': 0.2}) +[2026-10-03 19:28:27,018][starVLA.training.train_starvla][INFO] - Step 2701, Loss: {'train/loss': 6.983356433920562e-05, 'train/grad_norm_pre_clip': 0.020600954070687294, 'timing/data': 0.00033581999014131725, 'timing/model': 0.3916641859977972, 'learning_rate/qwen_vl_interface': 4.7908593719894914e-06, 'learning_rate/action_model': 4.790859371989491e-05, 'epoch': 0.2}) +[2026-10-03 19:28:27,925][starVLA.training.train_starvla][INFO] - Step 2702, Loss: {'train/loss': 0.35116180777549744, 'train/grad_norm_pre_clip': 35.38420867919922, 'timing/data': 0.00032042001839727163, 'timing/model': 0.3909477459965274, 'learning_rate/qwen_vl_interface': 4.78782830577584e-06, 'learning_rate/action_model': 4.7878283057758396e-05, 'epoch': 0.2}) +[2026-10-03 19:28:28,625][starVLA.training.train_starvla][INFO] - Step 2703, Loss: {'train/loss': 0.0414111390709877, 'train/grad_norm_pre_clip': 55.18815994262695, 'timing/data': 0.00033136000274680555, 'timing/model': 0.3921572059916798, 'learning_rate/qwen_vl_interface': 4.784797429543513e-06, 'learning_rate/action_model': 4.784797429543513e-05, 'epoch': 0.2}) +[2026-10-03 19:28:29,330][starVLA.training.train_starvla][INFO] - Step 2704, Loss: {'train/loss': 0.0019263388821855187, 'train/grad_norm_pre_clip': 2.08975887298584, 'timing/data': 0.00036467998870648444, 'timing/model': 0.39483024599030614, 'learning_rate/qwen_vl_interface': 4.781766744538393e-06, 'learning_rate/action_model': 4.781766744538393e-05, 'epoch': 0.2}) +[2026-10-03 19:28:30,031][starVLA.training.train_starvla][INFO] - Step 2705, Loss: {'train/loss': 0.013638945296406746, 'train/grad_norm_pre_clip': 19.947961807250977, 'timing/data': 0.00031887998920865357, 'timing/model': 0.39273035697988234, 'learning_rate/qwen_vl_interface': 4.778736252006277e-06, 'learning_rate/action_model': 4.7787362520062765e-05, 'epoch': 0.2}) +[2026-10-03 19:28:30,732][starVLA.training.train_starvla][INFO] - Step 2706, Loss: {'train/loss': 9.703248360892758e-05, 'train/grad_norm_pre_clip': 0.03101220540702343, 'timing/data': 0.00036567001370713115, 'timing/model': 0.3914576160022989, 'learning_rate/qwen_vl_interface': 4.775705953192885e-06, 'learning_rate/action_model': 4.7757059531928856e-05, 'epoch': 0.2}) +[2026-10-03 19:28:31,636][starVLA.training.train_starvla][INFO] - Step 2707, Loss: {'train/loss': 1.9237029846408404e-05, 'train/grad_norm_pre_clip': 0.00353893986903131, 'timing/data': 0.00034795000101439655, 'timing/model': 0.5948423499939963, 'learning_rate/qwen_vl_interface': 4.7726758493438615e-06, 'learning_rate/action_model': 4.7726758493438613e-05, 'epoch': 0.2}) +[2026-10-03 19:28:32,338][starVLA.training.train_starvla][INFO] - Step 2708, Loss: {'train/loss': 0.06287925690412521, 'train/grad_norm_pre_clip': 80.41373443603516, 'timing/data': 0.00034012002288363874, 'timing/model': 0.3910185170243494, 'learning_rate/qwen_vl_interface': 4.769645941704766e-06, 'learning_rate/action_model': 4.769645941704766e-05, 'epoch': 0.2}) +[2026-10-03 19:28:33,040][starVLA.training.train_starvla][INFO] - Step 2709, Loss: {'train/loss': 0.11417149007320404, 'train/grad_norm_pre_clip': 181.26556396484375, 'timing/data': 0.00032208001357503235, 'timing/model': 0.39370511600282043, 'learning_rate/qwen_vl_interface': 4.766616231521081e-06, 'learning_rate/action_model': 4.7666162315210806e-05, 'epoch': 0.2}) +[2026-10-03 19:28:33,743][starVLA.training.train_starvla][INFO] - Step 2710, Loss: {'train/loss': 0.004447292536497116, 'train/grad_norm_pre_clip': 4.502542018890381, 'timing/data': 0.00036795000778511167, 'timing/model': 0.39334113599034026, 'learning_rate/qwen_vl_interface': 4.763586720038202e-06, 'learning_rate/action_model': 4.7635867200382014e-05, 'epoch': 0.2}) +[2026-10-03 19:28:34,445][starVLA.training.train_starvla][INFO] - Step 2711, Loss: {'train/loss': 0.6241809129714966, 'train/grad_norm_pre_clip': 49.2314567565918, 'timing/data': 0.00020655000116676092, 'timing/model': 0.3944968060241081, 'learning_rate/qwen_vl_interface': 4.76055740850145e-06, 'learning_rate/action_model': 4.76055740850145e-05, 'epoch': 0.2}) +[2026-10-03 19:28:35,149][starVLA.training.train_starvla][INFO] - Step 2712, Loss: {'train/loss': 1.087289571762085, 'train/grad_norm_pre_clip': 132.63323974609375, 'timing/data': 0.0002154200046788901, 'timing/model': 0.39282068598549813, 'learning_rate/qwen_vl_interface': 4.757528298156058e-06, 'learning_rate/action_model': 4.757528298156058e-05, 'epoch': 0.2}) +[2026-10-03 19:28:36,060][starVLA.training.train_starvla][INFO] - Step 2713, Loss: {'train/loss': 4.3656160414684564e-05, 'train/grad_norm_pre_clip': 0.012401306070387363, 'timing/data': 0.00031171899172477424, 'timing/model': 0.3951320670021232, 'learning_rate/qwen_vl_interface': 4.754499390247181e-06, 'learning_rate/action_model': 4.754499390247181e-05, 'epoch': 0.2}) +[2026-10-03 19:28:36,760][starVLA.training.train_starvla][INFO] - Step 2714, Loss: {'train/loss': 8.105623419396579e-05, 'train/grad_norm_pre_clip': 0.03630496561527252, 'timing/data': 0.00034113001311197877, 'timing/model': 0.3926101560064126, 'learning_rate/qwen_vl_interface': 4.751470686019887e-06, 'learning_rate/action_model': 4.751470686019888e-05, 'epoch': 0.2}) +[2026-10-03 19:28:37,462][starVLA.training.train_starvla][INFO] - Step 2715, Loss: {'train/loss': 0.0011782492510974407, 'train/grad_norm_pre_clip': 3.497774124145508, 'timing/data': 0.0003083600022364408, 'timing/model': 0.39238936599576846, 'learning_rate/qwen_vl_interface': 4.7484421867191645e-06, 'learning_rate/action_model': 4.748442186719165e-05, 'epoch': 0.2}) +[2026-10-03 19:28:38,165][starVLA.training.train_starvla][INFO] - Step 2716, Loss: {'train/loss': 0.44626760482788086, 'train/grad_norm_pre_clip': 64.76597595214844, 'timing/data': 0.00032873000600375235, 'timing/model': 0.3913487770187203, 'learning_rate/qwen_vl_interface': 4.745413893589913e-06, 'learning_rate/action_model': 4.7454138935899125e-05, 'epoch': 0.2}) +[2026-10-03 19:28:38,866][starVLA.training.train_starvla][INFO] - Step 2717, Loss: {'train/loss': 0.5406795740127563, 'train/grad_norm_pre_clip': 53.97539520263672, 'timing/data': 0.00028629001462832093, 'timing/model': 0.3924807459989097, 'learning_rate/qwen_vl_interface': 4.7423858078769506e-06, 'learning_rate/action_model': 4.742385807876951e-05, 'epoch': 0.2}) +[2026-10-03 19:28:39,770][starVLA.training.train_starvla][INFO] - Step 2718, Loss: {'train/loss': 9.792857599677518e-05, 'train/grad_norm_pre_clip': 0.08489358425140381, 'timing/data': 0.0003596500027924776, 'timing/model': 0.39097548701101914, 'learning_rate/qwen_vl_interface': 4.739357930825008e-06, 'learning_rate/action_model': 4.739357930825008e-05, 'epoch': 0.2}) +[2026-10-03 19:28:40,467][starVLA.training.train_starvla][INFO] - Step 2719, Loss: {'train/loss': 2.128585401806049e-05, 'train/grad_norm_pre_clip': 0.0043589333072304726, 'timing/data': 0.0001969599979929626, 'timing/model': 0.3907504869857803, 'learning_rate/qwen_vl_interface': 4.736330263678731e-06, 'learning_rate/action_model': 4.736330263678731e-05, 'epoch': 0.2}) +[2026-10-03 19:28:41,165][starVLA.training.train_starvla][INFO] - Step 2720, Loss: {'train/loss': 0.0020643442403525114, 'train/grad_norm_pre_clip': 0.5614073276519775, 'timing/data': 0.00031700002728030086, 'timing/model': 0.389219797012629, 'learning_rate/qwen_vl_interface': 4.733302807682679e-06, 'learning_rate/action_model': 4.733302807682679e-05, 'epoch': 0.2}) +[2026-10-03 19:28:41,866][starVLA.training.train_starvla][INFO] - Step 2721, Loss: {'train/loss': 0.0005553349619731307, 'train/grad_norm_pre_clip': 0.45365479588508606, 'timing/data': 0.00023469998268410563, 'timing/model': 0.39345379700534977, 'learning_rate/qwen_vl_interface': 4.7302755640813255e-06, 'learning_rate/action_model': 4.7302755640813256e-05, 'epoch': 0.2}) +[2026-10-03 19:28:42,565][starVLA.training.train_starvla][INFO] - Step 2722, Loss: {'train/loss': 3.84872400900349e-05, 'train/grad_norm_pre_clip': 0.00948242750018835, 'timing/data': 0.0001962900278158486, 'timing/model': 0.3910635670181364, 'learning_rate/qwen_vl_interface': 4.727248534119058e-06, 'learning_rate/action_model': 4.7272485341190574e-05, 'epoch': 0.2}) +[2026-10-03 19:28:43,470][starVLA.training.train_starvla][INFO] - Step 2723, Loss: {'train/loss': 0.00010473006113898009, 'train/grad_norm_pre_clip': 0.031524334102869034, 'timing/data': 0.00030744000105187297, 'timing/model': 0.5973849690053612, 'learning_rate/qwen_vl_interface': 4.72422171904017e-06, 'learning_rate/action_model': 4.7242217190401697e-05, 'epoch': 0.2}) +[2026-10-03 19:28:44,169][starVLA.training.train_starvla][INFO] - Step 2724, Loss: {'train/loss': 0.3118022680282593, 'train/grad_norm_pre_clip': 292.9669494628906, 'timing/data': 0.00035174001823179424, 'timing/model': 0.3916644259879831, 'learning_rate/qwen_vl_interface': 4.721195120088873e-06, 'learning_rate/action_model': 4.721195120088873e-05, 'epoch': 0.2}) +[2026-10-03 19:28:44,866][starVLA.training.train_starvla][INFO] - Step 2725, Loss: {'train/loss': 0.5996203422546387, 'train/grad_norm_pre_clip': 76.49787902832031, 'timing/data': 0.0002928100002463907, 'timing/model': 0.39212941701407544, 'learning_rate/qwen_vl_interface': 4.718168738509288e-06, 'learning_rate/action_model': 4.718168738509288e-05, 'epoch': 0.2}) +[2026-10-03 19:28:45,561][starVLA.training.train_starvla][INFO] - Step 2726, Loss: {'train/loss': 0.16594506800174713, 'train/grad_norm_pre_clip': 84.42979431152344, 'timing/data': 0.0003424099995754659, 'timing/model': 0.38938501599477604, 'learning_rate/qwen_vl_interface': 4.715142575545445e-06, 'learning_rate/action_model': 4.7151425755454454e-05, 'epoch': 0.2}) +[2026-10-03 19:28:46,260][starVLA.training.train_starvla][INFO] - Step 2727, Loss: {'train/loss': 0.009677248075604439, 'train/grad_norm_pre_clip': 8.235246658325195, 'timing/data': 0.00033546998747624457, 'timing/model': 0.390797467000084, 'learning_rate/qwen_vl_interface': 4.712116632441287e-06, 'learning_rate/action_model': 4.7121166324412866e-05, 'epoch': 0.2}) +[2026-10-03 19:28:46,959][starVLA.training.train_starvla][INFO] - Step 2728, Loss: {'train/loss': 0.4727649390697479, 'train/grad_norm_pre_clip': 87.04354095458984, 'timing/data': 0.0002211800019722432, 'timing/model': 0.390467096993234, 'learning_rate/qwen_vl_interface': 4.709090910440664e-06, 'learning_rate/action_model': 4.7090909104406634e-05, 'epoch': 0.2}) +[2026-10-03 19:28:47,852][starVLA.training.train_starvla][INFO] - Step 2729, Loss: {'train/loss': 0.00010305467003490776, 'train/grad_norm_pre_clip': 0.04033168405294418, 'timing/data': 0.00024797001970000565, 'timing/model': 0.39015465701231733, 'learning_rate/qwen_vl_interface': 4.7060654107873335e-06, 'learning_rate/action_model': 4.7060654107873335e-05, 'epoch': 0.2}) +[2026-10-03 19:28:48,549][starVLA.training.train_starvla][INFO] - Step 2730, Loss: {'train/loss': 0.004311354830861092, 'train/grad_norm_pre_clip': 4.524438381195068, 'timing/data': 0.00020775001030415297, 'timing/model': 0.38913252600468695, 'learning_rate/qwen_vl_interface': 4.7030401347249685e-06, 'learning_rate/action_model': 4.703040134724968e-05, 'epoch': 0.2}) +[2026-10-03 19:28:49,247][starVLA.training.train_starvla][INFO] - Step 2731, Loss: {'train/loss': 0.001691454672254622, 'train/grad_norm_pre_clip': 0.7877746224403381, 'timing/data': 0.00029846999677829444, 'timing/model': 0.39095598700805567, 'learning_rate/qwen_vl_interface': 4.700015083497143e-06, 'learning_rate/action_model': 4.700015083497142e-05, 'epoch': 0.2}) +[2026-10-03 19:28:49,946][starVLA.training.train_starvla][INFO] - Step 2732, Loss: {'train/loss': 0.016443368047475815, 'train/grad_norm_pre_clip': 31.398714065551758, 'timing/data': 0.00023398999474011362, 'timing/model': 0.3918065859761555, 'learning_rate/qwen_vl_interface': 4.696990258347342e-06, 'learning_rate/action_model': 4.696990258347342e-05, 'epoch': 0.2}) +[2026-10-03 19:28:50,646][starVLA.training.train_starvla][INFO] - Step 2733, Loss: {'train/loss': 0.3443346321582794, 'train/grad_norm_pre_clip': 107.74585723876953, 'timing/data': 0.0003101399925071746, 'timing/model': 0.39319095702376217, 'learning_rate/qwen_vl_interface': 4.693965660518957e-06, 'learning_rate/action_model': 4.693965660518957e-05, 'epoch': 0.2}) +[2026-10-03 19:28:51,514][starVLA.training.train_starvla][INFO] - Step 2734, Loss: {'train/loss': 8.959414844866842e-05, 'train/grad_norm_pre_clip': 0.018882321193814278, 'timing/data': 0.00035004998790100217, 'timing/model': 0.3913254260260146, 'learning_rate/qwen_vl_interface': 4.690941291255285e-06, 'learning_rate/action_model': 4.6909412912552844e-05, 'epoch': 0.2}) +[2026-10-03 19:28:52,212][starVLA.training.train_starvla][INFO] - Step 2735, Loss: {'train/loss': 0.002983488840982318, 'train/grad_norm_pre_clip': 3.8577044010162354, 'timing/data': 0.00032342999475076795, 'timing/model': 0.3908993169898167, 'learning_rate/qwen_vl_interface': 4.68791715179953e-06, 'learning_rate/action_model': 4.68791715179953e-05, 'epoch': 0.2}) +[2026-10-03 19:28:52,912][starVLA.training.train_starvla][INFO] - Step 2736, Loss: {'train/loss': 0.26423802971839905, 'train/grad_norm_pre_clip': 67.47705078125, 'timing/data': 0.0003237599739804864, 'timing/model': 0.3915625269873999, 'learning_rate/qwen_vl_interface': 4.684893243394804e-06, 'learning_rate/action_model': 4.6848932433948036e-05, 'epoch': 0.2}) +[2026-10-03 19:28:53,613][starVLA.training.train_starvla][INFO] - Step 2737, Loss: {'train/loss': 0.0024920545984059572, 'train/grad_norm_pre_clip': 1.4085407257080078, 'timing/data': 0.00020529000903479755, 'timing/model': 0.39318965599522926, 'learning_rate/qwen_vl_interface': 4.68186956728412e-06, 'learning_rate/action_model': 4.68186956728412e-05, 'epoch': 0.2}) +[2026-10-03 19:28:54,315][starVLA.training.train_starvla][INFO] - Step 2738, Loss: {'train/loss': 0.593267560005188, 'train/grad_norm_pre_clip': 129.6757354736328, 'timing/data': 0.00023326001246459782, 'timing/model': 0.3918312569730915, 'learning_rate/qwen_vl_interface': 4.678846124710395e-06, 'learning_rate/action_model': 4.678846124710394e-05, 'epoch': 0.2}) +[2026-10-03 19:28:55,216][starVLA.training.train_starvla][INFO] - Step 2739, Loss: {'train/loss': 0.00022283186262939125, 'train/grad_norm_pre_clip': 0.07397758960723877, 'timing/data': 0.00021507000201381743, 'timing/model': 0.5935215899953619, 'learning_rate/qwen_vl_interface': 4.675822916916454e-06, 'learning_rate/action_model': 4.6758229169164536e-05, 'epoch': 0.2}) +[2026-10-03 19:28:55,917][starVLA.training.train_starvla][INFO] - Step 2740, Loss: {'train/loss': 0.6308801770210266, 'train/grad_norm_pre_clip': 97.5850830078125, 'timing/data': 0.0003331099869683385, 'timing/model': 0.3916546370019205, 'learning_rate/qwen_vl_interface': 4.6727999451450236e-06, 'learning_rate/action_model': 4.672799945145023e-05, 'epoch': 0.2}) +[2026-10-03 19:28:56,618][starVLA.training.train_starvla][INFO] - Step 2741, Loss: {'train/loss': 0.016051076352596283, 'train/grad_norm_pre_clip': 28.361066818237305, 'timing/data': 0.0002751700230874121, 'timing/model': 0.3938177469826769, 'learning_rate/qwen_vl_interface': 4.669777210638733e-06, 'learning_rate/action_model': 4.669777210638733e-05, 'epoch': 0.2}) +[2026-10-03 19:28:57,316][starVLA.training.train_starvla][INFO] - Step 2742, Loss: {'train/loss': 0.0021028893534094095, 'train/grad_norm_pre_clip': 1.6594603061676025, 'timing/data': 0.00020450001466087997, 'timing/model': 0.3905680959869642, 'learning_rate/qwen_vl_interface': 4.666754714640115e-06, 'learning_rate/action_model': 4.666754714640115e-05, 'epoch': 0.2}) +[2026-10-03 19:28:58,012][starVLA.training.train_starvla][INFO] - Step 2743, Loss: {'train/loss': 4.2689374822657555e-05, 'train/grad_norm_pre_clip': 0.008790283463895321, 'timing/data': 0.00024116001441143453, 'timing/model': 0.38843051600269973, 'learning_rate/qwen_vl_interface': 4.663732458391602e-06, 'learning_rate/action_model': 4.663732458391601e-05, 'epoch': 0.2}) +[2026-10-03 19:28:58,922][starVLA.training.train_starvla][INFO] - Step 2744, Loss: {'train/loss': 0.005840175319463015, 'train/grad_norm_pre_clip': 4.918740272521973, 'timing/data': 0.0003164899826515466, 'timing/model': 0.6028669480001554, 'learning_rate/qwen_vl_interface': 4.6607104431355295e-06, 'learning_rate/action_model': 4.66071044313553e-05, 'epoch': 0.2}) +[2026-10-03 19:28:59,621][starVLA.training.train_starvla][INFO] - Step 2745, Loss: {'train/loss': 0.010462790727615356, 'train/grad_norm_pre_clip': 19.60886573791504, 'timing/data': 0.0002574800164438784, 'timing/model': 0.3922794460086152, 'learning_rate/qwen_vl_interface': 4.657688670114135e-06, 'learning_rate/action_model': 4.657688670114135e-05, 'epoch': 0.2}) +[2026-10-03 19:29:00,320][starVLA.training.train_starvla][INFO] - Step 2746, Loss: {'train/loss': 0.019851915538311005, 'train/grad_norm_pre_clip': 41.725341796875, 'timing/data': 0.00024936997215263546, 'timing/model': 0.39162831602152437, 'learning_rate/qwen_vl_interface': 4.654667140569555e-06, 'learning_rate/action_model': 4.6546671405695544e-05, 'epoch': 0.2}) +[2026-10-03 19:29:01,016][starVLA.training.train_starvla][INFO] - Step 2747, Loss: {'train/loss': 0.0004851192352361977, 'train/grad_norm_pre_clip': 0.4886131286621094, 'timing/data': 0.00030901000718586147, 'timing/model': 0.3896293459984008, 'learning_rate/qwen_vl_interface': 4.651645855743825e-06, 'learning_rate/action_model': 4.651645855743825e-05, 'epoch': 0.2}) +[2026-10-03 19:29:01,713][starVLA.training.train_starvla][INFO] - Step 2748, Loss: {'train/loss': 0.0023220761213451624, 'train/grad_norm_pre_clip': 5.393869876861572, 'timing/data': 0.00034692001645453274, 'timing/model': 0.38884717697510496, 'learning_rate/qwen_vl_interface': 4.648624816878883e-06, 'learning_rate/action_model': 4.6486248168788825e-05, 'epoch': 0.2}) +[2026-10-03 19:29:02,413][starVLA.training.train_starvla][INFO] - Step 2749, Loss: {'train/loss': 0.0011558008845895529, 'train/grad_norm_pre_clip': 1.7138614654541016, 'timing/data': 0.00027210000553168356, 'timing/model': 0.39199957699747756, 'learning_rate/qwen_vl_interface': 4.6456040252165615e-06, 'learning_rate/action_model': 4.645604025216562e-05, 'epoch': 0.2}) +[2026-10-03 19:29:03,316][starVLA.training.train_starvla][INFO] - Step 2750, Loss: {'train/loss': 0.2755794823169708, 'train/grad_norm_pre_clip': 59.299537658691406, 'timing/data': 0.0004422499914653599, 'timing/model': 0.3914128159813117, 'learning_rate/qwen_vl_interface': 4.642583481998596e-06, 'learning_rate/action_model': 4.642583481998596e-05, 'epoch': 0.2}) +[2026-10-03 19:29:04,019][starVLA.training.train_starvla][INFO] - Step 2751, Loss: {'train/loss': 0.38267359137535095, 'train/grad_norm_pre_clip': 140.19898986816406, 'timing/data': 0.0003551399859134108, 'timing/model': 0.39314726600423455, 'learning_rate/qwen_vl_interface': 4.639563188466616e-06, 'learning_rate/action_model': 4.6395631884666165e-05, 'epoch': 0.2}) +[2026-10-03 19:29:04,718][starVLA.training.train_starvla][INFO] - Step 2752, Loss: {'train/loss': 0.0005921729025430977, 'train/grad_norm_pre_clip': 0.3878740072250366, 'timing/data': 0.00034652999602258205, 'timing/model': 0.3908590759965591, 'learning_rate/qwen_vl_interface': 4.636543145862152e-06, 'learning_rate/action_model': 4.6365431458621523e-05, 'epoch': 0.2}) +[2026-10-03 19:29:05,420][starVLA.training.train_starvla][INFO] - Step 2753, Loss: {'train/loss': 0.003950216807425022, 'train/grad_norm_pre_clip': 8.380220413208008, 'timing/data': 0.00029853000887669623, 'timing/model': 0.39295780600514263, 'learning_rate/qwen_vl_interface': 4.633523355426629e-06, 'learning_rate/action_model': 4.6335233554266295e-05, 'epoch': 0.2}) +[2026-10-03 19:29:06,121][starVLA.training.train_starvla][INFO] - Step 2754, Loss: {'train/loss': 5.370801954995841e-05, 'train/grad_norm_pre_clip': 0.014331522397696972, 'timing/data': 0.0002054210053756833, 'timing/model': 0.3922955050074961, 'learning_rate/qwen_vl_interface': 4.6305038184013675e-06, 'learning_rate/action_model': 4.6305038184013675e-05, 'epoch': 0.2}) +[2026-10-03 19:29:07,024][starVLA.training.train_starvla][INFO] - Step 2755, Loss: {'train/loss': 3.158201798214577e-05, 'train/grad_norm_pre_clip': 0.006586848758161068, 'timing/data': 0.0003568899992387742, 'timing/model': 0.3898491260188166, 'learning_rate/qwen_vl_interface': 4.627484536027589e-06, 'learning_rate/action_model': 4.627484536027588e-05, 'epoch': 0.2}) +[2026-10-03 19:29:07,724][starVLA.training.train_starvla][INFO] - Step 2756, Loss: {'train/loss': 0.6651669144630432, 'train/grad_norm_pre_clip': 142.24703979492188, 'timing/data': 0.0003382599970791489, 'timing/model': 0.3916843860060908, 'learning_rate/qwen_vl_interface': 4.624465509546399e-06, 'learning_rate/action_model': 4.6244655095463986e-05, 'epoch': 0.2}) +[2026-10-03 19:29:08,423][starVLA.training.train_starvla][INFO] - Step 2757, Loss: {'train/loss': 0.016716349869966507, 'train/grad_norm_pre_clip': 39.77177429199219, 'timing/data': 0.0002616200072225183, 'timing/model': 0.39282955598901026, 'learning_rate/qwen_vl_interface': 4.6214467401988106e-06, 'learning_rate/action_model': 4.621446740198811e-05, 'epoch': 0.2}) +[2026-10-03 19:29:09,123][starVLA.training.train_starvla][INFO] - Step 2758, Loss: {'train/loss': 0.0051942612044513226, 'train/grad_norm_pre_clip': 9.849599838256836, 'timing/data': 0.00020311999833211303, 'timing/model': 0.3915268660057336, 'learning_rate/qwen_vl_interface': 4.6184282292257245e-06, 'learning_rate/action_model': 4.618428229225725e-05, 'epoch': 0.2}) +[2026-10-03 19:29:09,820][starVLA.training.train_starvla][INFO] - Step 2759, Loss: {'train/loss': 0.22246053814888, 'train/grad_norm_pre_clip': 147.3346405029297, 'timing/data': 0.0002112200018018484, 'timing/model': 0.39157767602591775, 'learning_rate/qwen_vl_interface': 4.615409977867938e-06, 'learning_rate/action_model': 4.615409977867937e-05, 'epoch': 0.2}) +[2026-10-03 19:29:10,719][starVLA.training.train_starvla][INFO] - Step 2760, Loss: {'train/loss': 0.3255932331085205, 'train/grad_norm_pre_clip': 56.86714553833008, 'timing/data': 0.00034066999796777964, 'timing/model': 0.5909295989840757, 'learning_rate/qwen_vl_interface': 4.612391987366138e-06, 'learning_rate/action_model': 4.612391987366138e-05, 'epoch': 0.2}) +[2026-10-03 19:29:11,417][starVLA.training.train_starvla][INFO] - Step 2761, Loss: {'train/loss': 0.2928721010684967, 'train/grad_norm_pre_clip': 70.29644775390625, 'timing/data': 0.0003041099989786744, 'timing/model': 0.392346926004393, 'learning_rate/qwen_vl_interface': 4.609374258960908e-06, 'learning_rate/action_model': 4.6093742589609074e-05, 'epoch': 0.2}) +[2026-10-03 19:29:12,113][starVLA.training.train_starvla][INFO] - Step 2762, Loss: {'train/loss': 0.37653109431266785, 'train/grad_norm_pre_clip': 50.30667495727539, 'timing/data': 0.0002137300034519285, 'timing/model': 0.3893584859906696, 'learning_rate/qwen_vl_interface': 4.60635679389272e-06, 'learning_rate/action_model': 4.60635679389272e-05, 'epoch': 0.2}) +[2026-10-03 19:29:12,811][starVLA.training.train_starvla][INFO] - Step 2763, Loss: {'train/loss': 8.220141899073496e-05, 'train/grad_norm_pre_clip': 0.03926762193441391, 'timing/data': 0.0003040000156033784, 'timing/model': 0.3906252259912435, 'learning_rate/qwen_vl_interface': 4.603339593401943e-06, 'learning_rate/action_model': 4.603339593401942e-05, 'epoch': 0.2}) +[2026-10-03 19:29:13,488][starVLA.training.train_starvla][INFO] - Step 2764, Loss: {'train/loss': 9.728906297823414e-05, 'train/grad_norm_pre_clip': 0.032860249280929565, 'timing/data': 0.00021933999960310757, 'timing/model': 0.3922520760097541, 'learning_rate/qwen_vl_interface': 4.600322658728832e-06, 'learning_rate/action_model': 4.600322658728832e-05, 'epoch': 0.2}) +[2026-10-03 19:29:14,388][starVLA.training.train_starvla][INFO] - Step 2765, Loss: {'train/loss': 0.0016075873281806707, 'train/grad_norm_pre_clip': 1.6453920602798462, 'timing/data': 0.00020720998873002827, 'timing/model': 0.5941692300257273, 'learning_rate/qwen_vl_interface': 4.597305991113537e-06, 'learning_rate/action_model': 4.597305991113537e-05, 'epoch': 0.2}) +[2026-10-03 19:29:15,086][starVLA.training.train_starvla][INFO] - Step 2766, Loss: {'train/loss': 0.0001332075917162001, 'train/grad_norm_pre_clip': 0.056773796677589417, 'timing/data': 0.00020172999938949943, 'timing/model': 0.39070075601921417, 'learning_rate/qwen_vl_interface': 4.5942895917960935e-06, 'learning_rate/action_model': 4.5942895917960935e-05, 'epoch': 0.2}) +[2026-10-03 19:29:15,780][starVLA.training.train_starvla][INFO] - Step 2767, Loss: {'train/loss': 0.31635767221450806, 'train/grad_norm_pre_clip': 92.1391372680664, 'timing/data': 0.00021009001648053527, 'timing/model': 0.38900948700029403, 'learning_rate/qwen_vl_interface': 4.59127346201643e-06, 'learning_rate/action_model': 4.59127346201643e-05, 'epoch': 0.2}) +[2026-10-03 19:29:16,479][starVLA.training.train_starvla][INFO] - Step 2768, Loss: {'train/loss': 0.03279023990035057, 'train/grad_norm_pre_clip': 55.514320373535156, 'timing/data': 0.0002156600239686668, 'timing/model': 0.39098925702273846, 'learning_rate/qwen_vl_interface': 4.588257603014364e-06, 'learning_rate/action_model': 4.588257603014364e-05, 'epoch': 0.2}) +[2026-10-03 19:29:17,178][starVLA.training.train_starvla][INFO] - Step 2769, Loss: {'train/loss': 0.0014235047856345773, 'train/grad_norm_pre_clip': 1.4519177675247192, 'timing/data': 0.00019754000823013484, 'timing/model': 0.39087345599546097, 'learning_rate/qwen_vl_interface': 4.585242016029602e-06, 'learning_rate/action_model': 4.5852420160296015e-05, 'epoch': 0.2}) +[2026-10-03 19:29:18,078][starVLA.training.train_starvla][INFO] - Step 2770, Loss: {'train/loss': 0.08341608941555023, 'train/grad_norm_pre_clip': 55.440216064453125, 'timing/data': 0.0002827199932653457, 'timing/model': 0.5933260789897759, 'learning_rate/qwen_vl_interface': 4.582226702301737e-06, 'learning_rate/action_model': 4.5822267023017375e-05, 'epoch': 0.2}) +[2026-10-03 19:29:18,774][starVLA.training.train_starvla][INFO] - Step 2771, Loss: {'train/loss': 7.667172758374363e-05, 'train/grad_norm_pre_clip': 0.027492133900523186, 'timing/data': 0.0003759800165425986, 'timing/model': 0.3885693369957153, 'learning_rate/qwen_vl_interface': 4.57921166307025e-06, 'learning_rate/action_model': 4.57921166307025e-05, 'epoch': 0.2}) +[2026-10-03 19:29:19,472][starVLA.training.train_starvla][INFO] - Step 2772, Loss: {'train/loss': 0.0004035299352835864, 'train/grad_norm_pre_clip': 0.28381770849227905, 'timing/data': 0.00021118999575264752, 'timing/model': 0.39003658600267954, 'learning_rate/qwen_vl_interface': 4.576196899574511e-06, 'learning_rate/action_model': 4.576196899574511e-05, 'epoch': 0.2}) +[2026-10-03 19:29:20,170][starVLA.training.train_starvla][INFO] - Step 2773, Loss: {'train/loss': 0.6120569109916687, 'train/grad_norm_pre_clip': 50.56446075439453, 'timing/data': 0.00024519002181477845, 'timing/model': 0.38990107699646614, 'learning_rate/qwen_vl_interface': 4.573182413053775e-06, 'learning_rate/action_model': 4.5731824130537746e-05, 'epoch': 0.2}) +[2026-10-03 19:29:20,865][starVLA.training.train_starvla][INFO] - Step 2774, Loss: {'train/loss': 0.0002637215657159686, 'train/grad_norm_pre_clip': 0.44571250677108765, 'timing/data': 0.0003207199915777892, 'timing/model': 0.38957921601831913, 'learning_rate/qwen_vl_interface': 4.570168204747183e-06, 'learning_rate/action_model': 4.570168204747183e-05, 'epoch': 0.2}) +[2026-10-03 19:29:21,564][starVLA.training.train_starvla][INFO] - Step 2775, Loss: {'train/loss': 0.2691911458969116, 'train/grad_norm_pre_clip': 72.87586975097656, 'timing/data': 0.00027434001094661653, 'timing/model': 0.3912545460043475, 'learning_rate/qwen_vl_interface': 4.567154275893762e-06, 'learning_rate/action_model': 4.5671542758937624e-05, 'epoch': 0.2}) +[2026-10-03 19:29:22,463][starVLA.training.train_starvla][INFO] - Step 2776, Loss: {'train/loss': 0.22076216340065002, 'train/grad_norm_pre_clip': 253.19725036621094, 'timing/data': 0.0003441600129008293, 'timing/model': 0.39071737599442713, 'learning_rate/qwen_vl_interface': 4.564140627732426e-06, 'learning_rate/action_model': 4.564140627732426e-05, 'epoch': 0.2}) +[2026-10-03 19:29:23,166][starVLA.training.train_starvla][INFO] - Step 2777, Loss: {'train/loss': 0.5362188816070557, 'train/grad_norm_pre_clip': 45.74858093261719, 'timing/data': 0.00021272001322358847, 'timing/model': 0.3971940859919414, 'learning_rate/qwen_vl_interface': 4.56112726150197e-06, 'learning_rate/action_model': 4.56112726150197e-05, 'epoch': 0.2}) +[2026-10-03 19:29:23,866][starVLA.training.train_starvla][INFO] - Step 2778, Loss: {'train/loss': 0.0010314283426851034, 'train/grad_norm_pre_clip': 1.537627935409546, 'timing/data': 0.00020122999558225274, 'timing/model': 0.3925985360110644, 'learning_rate/qwen_vl_interface': 4.558114178441076e-06, 'learning_rate/action_model': 4.558114178441076e-05, 'epoch': 0.2}) +[2026-10-03 19:29:24,566][starVLA.training.train_starvla][INFO] - Step 2779, Loss: {'train/loss': 0.4610154628753662, 'train/grad_norm_pre_clip': 121.10391235351562, 'timing/data': 0.0003366299788467586, 'timing/model': 0.39128730600350536, 'learning_rate/qwen_vl_interface': 4.555101379788308e-06, 'learning_rate/action_model': 4.555101379788308e-05, 'epoch': 0.2}) +[2026-10-03 19:29:25,266][starVLA.training.train_starvla][INFO] - Step 2780, Loss: {'train/loss': 0.00012116684229113162, 'train/grad_norm_pre_clip': 0.05734531208872795, 'timing/data': 0.00031015000422485173, 'timing/model': 0.3929827870160807, 'learning_rate/qwen_vl_interface': 4.552088866782115e-06, 'learning_rate/action_model': 4.552088866782115e-05, 'epoch': 0.2}) +[2026-10-03 19:29:26,171][starVLA.training.train_starvla][INFO] - Step 2781, Loss: {'train/loss': 0.0892086997628212, 'train/grad_norm_pre_clip': 76.06444549560547, 'timing/data': 0.00043701002141460776, 'timing/model': 0.5974217490002047, 'learning_rate/qwen_vl_interface': 4.5490766406608265e-06, 'learning_rate/action_model': 4.549076640660827e-05, 'epoch': 0.2}) +[2026-10-03 19:29:26,870][starVLA.training.train_starvla][INFO] - Step 2782, Loss: {'train/loss': 0.0100718317553401, 'train/grad_norm_pre_clip': 9.340327262878418, 'timing/data': 0.00040590998833067715, 'timing/model': 0.3911110660119448, 'learning_rate/qwen_vl_interface': 4.546064702662655e-06, 'learning_rate/action_model': 4.546064702662655e-05, 'epoch': 0.2}) +[2026-10-03 19:29:27,566][starVLA.training.train_starvla][INFO] - Step 2783, Loss: {'train/loss': 0.0007006676169112325, 'train/grad_norm_pre_clip': 0.5273521542549133, 'timing/data': 0.00020397998741827905, 'timing/model': 0.38968877599108964, 'learning_rate/qwen_vl_interface': 4.543053054025693e-06, 'learning_rate/action_model': 4.5430530540256926e-05, 'epoch': 0.2}) +[2026-10-03 19:29:28,266][starVLA.training.train_starvla][INFO] - Step 2784, Loss: {'train/loss': 0.1527739316225052, 'train/grad_norm_pre_clip': 119.45880126953125, 'timing/data': 0.0003228499845135957, 'timing/model': 0.3912928460049443, 'learning_rate/qwen_vl_interface': 4.540041695987917e-06, 'learning_rate/action_model': 4.540041695987917e-05, 'epoch': 0.2}) +[2026-10-03 19:29:28,965][starVLA.training.train_starvla][INFO] - Step 2785, Loss: {'train/loss': 0.002637968398630619, 'train/grad_norm_pre_clip': 2.376234292984009, 'timing/data': 0.0002799399953801185, 'timing/model': 0.39138451701728627, 'learning_rate/qwen_vl_interface': 4.537030629787182e-06, 'learning_rate/action_model': 4.5370306297871826e-05, 'epoch': 0.2}) +[2026-10-03 19:29:29,863][starVLA.training.train_starvla][INFO] - Step 2786, Loss: {'train/loss': 0.2982770502567291, 'train/grad_norm_pre_clip': 90.17188262939453, 'timing/data': 0.00041258998680859804, 'timing/model': 0.5899949290032964, 'learning_rate/qwen_vl_interface': 4.534019856661226e-06, 'learning_rate/action_model': 4.534019856661226e-05, 'epoch': 0.2}) +[2026-10-03 19:29:30,562][starVLA.training.train_starvla][INFO] - Step 2787, Loss: {'train/loss': 0.00015756275388412178, 'train/grad_norm_pre_clip': 0.05437653511762619, 'timing/data': 0.0003294199996162206, 'timing/model': 0.39159781602211297, 'learning_rate/qwen_vl_interface': 4.531009377847661e-06, 'learning_rate/action_model': 4.531009377847661e-05, 'epoch': 0.2}) +[2026-10-03 19:29:31,262][starVLA.training.train_starvla][INFO] - Step 2788, Loss: {'train/loss': 0.0007400544127449393, 'train/grad_norm_pre_clip': 0.5583238005638123, 'timing/data': 0.0003358899848535657, 'timing/model': 0.3907898259931244, 'learning_rate/qwen_vl_interface': 4.527999194583982e-06, 'learning_rate/action_model': 4.527999194583981e-05, 'epoch': 0.2}) +[2026-10-03 19:29:31,960][starVLA.training.train_starvla][INFO] - Step 2789, Loss: {'train/loss': 0.004365681204944849, 'train/grad_norm_pre_clip': 7.858012676239014, 'timing/data': 0.00026164998416788876, 'timing/model': 0.3905828870192636, 'learning_rate/qwen_vl_interface': 4.52498930810756e-06, 'learning_rate/action_model': 4.52498930810756e-05, 'epoch': 0.2}) +[2026-10-03 19:29:32,656][starVLA.training.train_starvla][INFO] - Step 2790, Loss: {'train/loss': 0.7161562442779541, 'train/grad_norm_pre_clip': 37.03352355957031, 'timing/data': 0.0002413600159343332, 'timing/model': 0.38984693700331263, 'learning_rate/qwen_vl_interface': 4.521979719655648e-06, 'learning_rate/action_model': 4.5219797196556484e-05, 'epoch': 0.2}) +[2026-10-03 19:29:33,554][starVLA.training.train_starvla][INFO] - Step 2791, Loss: {'train/loss': 0.4399956464767456, 'train/grad_norm_pre_clip': 114.12583923339844, 'timing/data': 0.0003146399976685643, 'timing/model': 0.5929469190014061, 'learning_rate/qwen_vl_interface': 4.518970430465377e-06, 'learning_rate/action_model': 4.518970430465376e-05, 'epoch': 0.2}) +[2026-10-03 19:29:34,253][starVLA.training.train_starvla][INFO] - Step 2792, Loss: {'train/loss': 0.030506014823913574, 'train/grad_norm_pre_clip': 34.10919189453125, 'timing/data': 0.00040983999497257173, 'timing/model': 0.39075330601190217, 'learning_rate/qwen_vl_interface': 4.515961441773745e-06, 'learning_rate/action_model': 4.515961441773745e-05, 'epoch': 0.2}) +[2026-10-03 19:29:34,949][starVLA.training.train_starvla][INFO] - Step 2793, Loss: {'train/loss': 0.0001150093594333157, 'train/grad_norm_pre_clip': 0.03202712535858154, 'timing/data': 0.0002985699975397438, 'timing/model': 0.3906565560027957, 'learning_rate/qwen_vl_interface': 4.5129527548176385e-06, 'learning_rate/action_model': 4.512952754817638e-05, 'epoch': 0.2}) +[2026-10-03 19:29:35,645][starVLA.training.train_starvla][INFO] - Step 2794, Loss: {'train/loss': 0.4906032383441925, 'train/grad_norm_pre_clip': 32.54891586303711, 'timing/data': 0.00039368000579997897, 'timing/model': 0.3904732560040429, 'learning_rate/qwen_vl_interface': 4.509944370833813e-06, 'learning_rate/action_model': 4.509944370833813e-05, 'epoch': 0.2}) +[2026-10-03 19:29:36,343][starVLA.training.train_starvla][INFO] - Step 2795, Loss: {'train/loss': 0.06879056990146637, 'train/grad_norm_pre_clip': 83.47154235839844, 'timing/data': 0.00021329999435693026, 'timing/model': 0.38982662599300966, 'learning_rate/qwen_vl_interface': 4.506936291058902e-06, 'learning_rate/action_model': 4.506936291058902e-05, 'epoch': 0.2}) +[2026-10-03 19:29:37,041][starVLA.training.train_starvla][INFO] - Step 2796, Loss: {'train/loss': 0.00010518314957153052, 'train/grad_norm_pre_clip': 0.046493008732795715, 'timing/data': 0.0003314600035082549, 'timing/model': 0.38932232599472627, 'learning_rate/qwen_vl_interface': 4.503928516729416e-06, 'learning_rate/action_model': 4.503928516729416e-05, 'epoch': 0.2}) +[2026-10-03 19:29:37,938][starVLA.training.train_starvla][INFO] - Step 2797, Loss: {'train/loss': 0.00035527549334801733, 'train/grad_norm_pre_clip': 0.131069615483284, 'timing/data': 0.00020887001301161945, 'timing/model': 0.39249339600792155, 'learning_rate/qwen_vl_interface': 4.500921049081734e-06, 'learning_rate/action_model': 4.500921049081734e-05, 'epoch': 0.2}) +[2026-10-03 19:29:38,636][starVLA.training.train_starvla][INFO] - Step 2798, Loss: {'train/loss': 0.0002562800655141473, 'train/grad_norm_pre_clip': 0.07823409885168076, 'timing/data': 0.0003668500285129994, 'timing/model': 0.39010805601719767, 'learning_rate/qwen_vl_interface': 4.497913889352113e-06, 'learning_rate/action_model': 4.4979138893521126e-05, 'epoch': 0.2}) +[2026-10-03 19:29:39,334][starVLA.training.train_starvla][INFO] - Step 2799, Loss: {'train/loss': 0.0064722271636128426, 'train/grad_norm_pre_clip': 22.171342849731445, 'timing/data': 0.00032952000037766993, 'timing/model': 0.3906401960121002, 'learning_rate/qwen_vl_interface': 4.494907038776684e-06, 'learning_rate/action_model': 4.4949070387766836e-05, 'epoch': 0.2}) +[2026-10-03 19:29:40,032][starVLA.training.train_starvla][INFO] - Step 2800, Loss: {'train/loss': 0.3981403708457947, 'train/grad_norm_pre_clip': 135.15872192382812, 'timing/data': 0.00033581999014131725, 'timing/model': 0.3888855059922207, 'learning_rate/qwen_vl_interface': 4.49190049859145e-06, 'learning_rate/action_model': 4.49190049859145e-05, 'epoch': 0.2}) +[2026-10-03 19:29:40,730][starVLA.training.train_starvla][INFO] - Step 2801, Loss: {'train/loss': 5.661074101226404e-05, 'train/grad_norm_pre_clip': 0.024512045085430145, 'timing/data': 0.00028015999123454094, 'timing/model': 0.39220661600120366, 'learning_rate/qwen_vl_interface': 4.488894270032284e-06, 'learning_rate/action_model': 4.4888942700322834e-05, 'epoch': 0.21}) +[2026-10-03 19:29:41,634][starVLA.training.train_starvla][INFO] - Step 2802, Loss: {'train/loss': 7.988127617863938e-05, 'train/grad_norm_pre_clip': 0.03683030977845192, 'timing/data': 0.00021441999706439674, 'timing/model': 0.3898327359929681, 'learning_rate/qwen_vl_interface': 4.485888354334933e-06, 'learning_rate/action_model': 4.4858883543349334e-05, 'epoch': 0.21}) +[2026-10-03 19:29:42,330][starVLA.training.train_starvla][INFO] - Step 2803, Loss: {'train/loss': 0.012537202797830105, 'train/grad_norm_pre_clip': 16.8199462890625, 'timing/data': 0.00032374999136663973, 'timing/model': 0.3878047360049095, 'learning_rate/qwen_vl_interface': 4.482882752735017e-06, 'learning_rate/action_model': 4.482882752735017e-05, 'epoch': 0.21}) +[2026-10-03 19:29:43,030][starVLA.training.train_starvla][INFO] - Step 2804, Loss: {'train/loss': 0.02951272763311863, 'train/grad_norm_pre_clip': 46.35071563720703, 'timing/data': 0.00035802999627776444, 'timing/model': 0.3905150259961374, 'learning_rate/qwen_vl_interface': 4.479877466468027e-06, 'learning_rate/action_model': 4.479877466468026e-05, 'epoch': 0.21}) +[2026-10-03 19:29:43,729][starVLA.training.train_starvla][INFO] - Step 2805, Loss: {'train/loss': 1.0541162490844727, 'train/grad_norm_pre_clip': 84.59297943115234, 'timing/data': 0.00022031000116840005, 'timing/model': 0.3918880760029424, 'learning_rate/qwen_vl_interface': 4.476872496769318e-06, 'learning_rate/action_model': 4.476872496769318e-05, 'epoch': 0.21}) +[2026-10-03 19:29:44,428][starVLA.training.train_starvla][INFO] - Step 2806, Loss: {'train/loss': 1.641321614442859e-05, 'train/grad_norm_pre_clip': 0.0038019532803446054, 'timing/data': 0.00032616997486911714, 'timing/model': 0.3904761860030703, 'learning_rate/qwen_vl_interface': 4.473867844874122e-06, 'learning_rate/action_model': 4.473867844874122e-05, 'epoch': 0.21}) +[2026-10-03 19:29:45,323][starVLA.training.train_starvla][INFO] - Step 2807, Loss: {'train/loss': 0.06551273167133331, 'train/grad_norm_pre_clip': 29.21934700012207, 'timing/data': 0.00021031999494880438, 'timing/model': 0.5892003590124659, 'learning_rate/qwen_vl_interface': 4.470863512017539e-06, 'learning_rate/action_model': 4.470863512017539e-05, 'epoch': 0.21}) +[2026-10-03 19:29:46,021][starVLA.training.train_starvla][INFO] - Step 2808, Loss: {'train/loss': 0.0006031164666637778, 'train/grad_norm_pre_clip': 0.27935686707496643, 'timing/data': 0.0002562799782026559, 'timing/model': 0.39130853599635884, 'learning_rate/qwen_vl_interface': 4.4678594994345344e-06, 'learning_rate/action_model': 4.467859499434534e-05, 'epoch': 0.21}) +[2026-10-03 19:29:46,721][starVLA.training.train_starvla][INFO] - Step 2809, Loss: {'train/loss': 0.37846168875694275, 'train/grad_norm_pre_clip': 146.6192626953125, 'timing/data': 0.00020233000395819545, 'timing/model': 0.3931607859849464, 'learning_rate/qwen_vl_interface': 4.4648558083599475e-06, 'learning_rate/action_model': 4.464855808359947e-05, 'epoch': 0.21}) +[2026-10-03 19:29:47,418][starVLA.training.train_starvla][INFO] - Step 2810, Loss: {'train/loss': 0.00011812368029495701, 'train/grad_norm_pre_clip': 0.11741255968809128, 'timing/data': 0.00032401998760178685, 'timing/model': 0.39078796599642374, 'learning_rate/qwen_vl_interface': 4.461852440028479e-06, 'learning_rate/action_model': 4.461852440028479e-05, 'epoch': 0.21}) +[2026-10-03 19:29:48,115][starVLA.training.train_starvla][INFO] - Step 2811, Loss: {'train/loss': 0.015353042632341385, 'train/grad_norm_pre_clip': 37.555171966552734, 'timing/data': 0.00032702001044526696, 'timing/model': 0.3882504270004574, 'learning_rate/qwen_vl_interface': 4.458849395674702e-06, 'learning_rate/action_model': 4.458849395674701e-05, 'epoch': 0.21}) +[2026-10-03 19:29:49,020][starVLA.training.train_starvla][INFO] - Step 2812, Loss: {'train/loss': 0.1034688651561737, 'train/grad_norm_pre_clip': 52.68897247314453, 'timing/data': 0.00027563999174162745, 'timing/model': 0.5965173779986799, 'learning_rate/qwen_vl_interface': 4.455846676533052e-06, 'learning_rate/action_model': 4.4558466765330525e-05, 'epoch': 0.21}) +[2026-10-03 19:29:49,720][starVLA.training.train_starvla][INFO] - Step 2813, Loss: {'train/loss': 0.0064113205298781395, 'train/grad_norm_pre_clip': 3.640623092651367, 'timing/data': 0.0002993500092998147, 'timing/model': 0.39300770600675605, 'learning_rate/qwen_vl_interface': 4.452844283837838e-06, 'learning_rate/action_model': 4.452844283837838e-05, 'epoch': 0.21}) +[2026-10-03 19:29:50,418][starVLA.training.train_starvla][INFO] - Step 2814, Loss: {'train/loss': 0.022898171097040176, 'train/grad_norm_pre_clip': 27.128673553466797, 'timing/data': 0.0002329499984625727, 'timing/model': 0.3910642559931148, 'learning_rate/qwen_vl_interface': 4.449842218823228e-06, 'learning_rate/action_model': 4.449842218823228e-05, 'epoch': 0.21}) +[2026-10-03 19:29:51,123][starVLA.training.train_starvla][INFO] - Step 2815, Loss: {'train/loss': 0.2996821105480194, 'train/grad_norm_pre_clip': 90.2151107788086, 'timing/data': 0.00020867001148872077, 'timing/model': 0.3919771960063372, 'learning_rate/qwen_vl_interface': 4.446840482723256e-06, 'learning_rate/action_model': 4.446840482723255e-05, 'epoch': 0.21}) +[2026-10-03 19:29:51,821][starVLA.training.train_starvla][INFO] - Step 2816, Loss: {'train/loss': 4.98024164699018e-05, 'train/grad_norm_pre_clip': 0.013138956390321255, 'timing/data': 0.00034338998375460505, 'timing/model': 0.3901443560025655, 'learning_rate/qwen_vl_interface': 4.443839076771823e-06, 'learning_rate/action_model': 4.443839076771823e-05, 'epoch': 0.21}) +[2026-10-03 19:29:52,722][starVLA.training.train_starvla][INFO] - Step 2817, Loss: {'train/loss': 0.29132649302482605, 'train/grad_norm_pre_clip': 56.41694259643555, 'timing/data': 0.0002770099963527173, 'timing/model': 0.5934733879985288, 'learning_rate/qwen_vl_interface': 4.4408380022026945e-06, 'learning_rate/action_model': 4.440838002202694e-05, 'epoch': 0.21}) +[2026-10-03 19:29:53,424][starVLA.training.train_starvla][INFO] - Step 2818, Loss: {'train/loss': 0.0003508449881337583, 'train/grad_norm_pre_clip': 0.28978168964385986, 'timing/data': 0.0004038400074932724, 'timing/model': 0.39345562501694076, 'learning_rate/qwen_vl_interface': 4.437837260249499e-06, 'learning_rate/action_model': 4.437837260249499e-05, 'epoch': 0.21}) +[2026-10-03 19:29:54,126][starVLA.training.train_starvla][INFO] - Step 2819, Loss: {'train/loss': 0.022273028269410133, 'train/grad_norm_pre_clip': 11.440869331359863, 'timing/data': 0.00021005998132750392, 'timing/model': 0.39234222599770874, 'learning_rate/qwen_vl_interface': 4.434836852145728e-06, 'learning_rate/action_model': 4.434836852145728e-05, 'epoch': 0.21}) +[2026-10-03 19:29:54,829][starVLA.training.train_starvla][INFO] - Step 2820, Loss: {'train/loss': 0.020281808450818062, 'train/grad_norm_pre_clip': 19.221839904785156, 'timing/data': 0.00021784999989904463, 'timing/model': 0.39252468600170687, 'learning_rate/qwen_vl_interface': 4.431836779124736e-06, 'learning_rate/action_model': 4.431836779124735e-05, 'epoch': 0.21}) +[2026-10-03 19:29:55,528][starVLA.training.train_starvla][INFO] - Step 2821, Loss: {'train/loss': 8.589294156990945e-05, 'train/grad_norm_pre_clip': 0.03615758940577507, 'timing/data': 0.00032364000799134374, 'timing/model': 0.39197927599889226, 'learning_rate/qwen_vl_interface': 4.428837042419738e-06, 'learning_rate/action_model': 4.428837042419738e-05, 'epoch': 0.21}) +[2026-10-03 19:29:56,228][starVLA.training.train_starvla][INFO] - Step 2822, Loss: {'train/loss': 0.9162539839744568, 'train/grad_norm_pre_clip': 52.35649490356445, 'timing/data': 0.00030990998493507504, 'timing/model': 0.3915678259800188, 'learning_rate/qwen_vl_interface': 4.425837643263815e-06, 'learning_rate/action_model': 4.425837643263814e-05, 'epoch': 0.21}) +[2026-10-03 19:29:57,109][starVLA.training.train_starvla][INFO] - Step 2823, Loss: {'train/loss': 0.44706594944000244, 'train/grad_norm_pre_clip': 35.859188079833984, 'timing/data': 0.000277960003586486, 'timing/model': 0.3655012979870662, 'learning_rate/qwen_vl_interface': 4.4228385828899045e-06, 'learning_rate/action_model': 4.4228385828899046e-05, 'epoch': 0.21}) +[2026-10-03 19:29:57,807][starVLA.training.train_starvla][INFO] - Step 2824, Loss: {'train/loss': 0.014636920765042305, 'train/grad_norm_pre_clip': 17.541866302490234, 'timing/data': 0.0003297600196674466, 'timing/model': 0.3901711259968579, 'learning_rate/qwen_vl_interface': 4.419839862530809e-06, 'learning_rate/action_model': 4.419839862530809e-05, 'epoch': 0.21}) +[2026-10-03 19:29:58,506][starVLA.training.train_starvla][INFO] - Step 2825, Loss: {'train/loss': 0.00014897112851031125, 'train/grad_norm_pre_clip': 0.032710861414670944, 'timing/data': 0.00021023998851887882, 'timing/model': 0.39084484599879943, 'learning_rate/qwen_vl_interface': 4.416841483419188e-06, 'learning_rate/action_model': 4.416841483419188e-05, 'epoch': 0.21}) +[2026-10-03 19:29:59,207][starVLA.training.train_starvla][INFO] - Step 2826, Loss: {'train/loss': 0.0012585934018716216, 'train/grad_norm_pre_clip': 1.3378289937973022, 'timing/data': 0.00034922000486403704, 'timing/model': 0.39179027601494454, 'learning_rate/qwen_vl_interface': 4.413843446787561e-06, 'learning_rate/action_model': 4.413843446787561e-05, 'epoch': 0.21}) +[2026-10-03 19:29:59,906][starVLA.training.train_starvla][INFO] - Step 2827, Loss: {'train/loss': 0.9718942642211914, 'train/grad_norm_pre_clip': 177.7881622314453, 'timing/data': 0.0002978299744427204, 'timing/model': 0.39052705600624904, 'learning_rate/qwen_vl_interface': 4.41084575386831e-06, 'learning_rate/action_model': 4.4108457538683093e-05, 'epoch': 0.21}) +[2026-10-03 19:30:00,777][starVLA.training.train_starvla][INFO] - Step 2828, Loss: {'train/loss': 0.0017225719057023525, 'train/grad_norm_pre_clip': 0.7895541191101074, 'timing/data': 0.00038281999877654016, 'timing/model': 0.5599367709946819, 'learning_rate/qwen_vl_interface': 4.40784840589367e-06, 'learning_rate/action_model': 4.4078484058936706e-05, 'epoch': 0.21}) +[2026-10-03 19:30:01,476][starVLA.training.train_starvla][INFO] - Step 2829, Loss: {'train/loss': 0.07969799637794495, 'train/grad_norm_pre_clip': 105.77951049804688, 'timing/data': 0.0002015999925788492, 'timing/model': 0.3923639949935023, 'learning_rate/qwen_vl_interface': 4.404851404095742e-06, 'learning_rate/action_model': 4.404851404095742e-05, 'epoch': 0.21}) +[2026-10-03 19:30:02,174][starVLA.training.train_starvla][INFO] - Step 2830, Loss: {'train/loss': 0.43431252241134644, 'train/grad_norm_pre_clip': 115.22039794921875, 'timing/data': 0.0003365500015206635, 'timing/model': 0.3915651559946127, 'learning_rate/qwen_vl_interface': 4.4018547497064765e-06, 'learning_rate/action_model': 4.401854749706476e-05, 'epoch': 0.21}) +[2026-10-03 19:30:02,874][starVLA.training.train_starvla][INFO] - Step 2831, Loss: {'train/loss': 0.0011543332366272807, 'train/grad_norm_pre_clip': 3.339165210723877, 'timing/data': 0.0003344499855302274, 'timing/model': 0.3902239360031672, 'learning_rate/qwen_vl_interface': 4.398858443957686e-06, 'learning_rate/action_model': 4.398858443957686e-05, 'epoch': 0.21}) +[2026-10-03 19:30:03,576][starVLA.training.train_starvla][INFO] - Step 2832, Loss: {'train/loss': 0.0006866760086268187, 'train/grad_norm_pre_clip': 0.2534487843513489, 'timing/data': 0.00034587999107316136, 'timing/model': 0.3939704349904787, 'learning_rate/qwen_vl_interface': 4.395862488081038e-06, 'learning_rate/action_model': 4.395862488081038e-05, 'epoch': 0.21}) +[2026-10-03 19:30:04,276][starVLA.training.train_starvla][INFO] - Step 2833, Loss: {'train/loss': 0.05975470691919327, 'train/grad_norm_pre_clip': 62.48103713989258, 'timing/data': 0.0002928400062955916, 'timing/model': 0.3932666059990879, 'learning_rate/qwen_vl_interface': 4.392866883308059e-06, 'learning_rate/action_model': 4.392866883308059e-05, 'epoch': 0.21}) +[2026-10-03 19:30:05,178][starVLA.training.train_starvla][INFO] - Step 2834, Loss: {'train/loss': 0.0010385566856712103, 'train/grad_norm_pre_clip': 0.24308599531650543, 'timing/data': 0.000360759993782267, 'timing/model': 0.39097634598147124, 'learning_rate/qwen_vl_interface': 4.389871630870127e-06, 'learning_rate/action_model': 4.389871630870127e-05, 'epoch': 0.21}) +[2026-10-03 19:30:05,879][starVLA.training.train_starvla][INFO] - Step 2835, Loss: {'train/loss': 0.0005798599449917674, 'train/grad_norm_pre_clip': 0.45141828060150146, 'timing/data': 0.0003116799925919622, 'timing/model': 0.3920752559788525, 'learning_rate/qwen_vl_interface': 4.386876731998478e-06, 'learning_rate/action_model': 4.386876731998478e-05, 'epoch': 0.21}) +[2026-10-03 19:30:06,580][starVLA.training.train_starvla][INFO] - Step 2836, Loss: {'train/loss': 0.0027340776287019253, 'train/grad_norm_pre_clip': 2.036372661590576, 'timing/data': 0.00023510001483373344, 'timing/model': 0.39236865600105375, 'learning_rate/qwen_vl_interface': 4.3838821879242006e-06, 'learning_rate/action_model': 4.3838821879242e-05, 'epoch': 0.21}) +[2026-10-03 19:30:07,283][starVLA.training.train_starvla][INFO] - Step 2837, Loss: {'train/loss': 0.09972091019153595, 'train/grad_norm_pre_clip': 118.64228057861328, 'timing/data': 0.00030173000413924456, 'timing/model': 0.3932896259939298, 'learning_rate/qwen_vl_interface': 4.38088799987824e-06, 'learning_rate/action_model': 4.38088799987824e-05, 'epoch': 0.21}) +[2026-10-03 19:30:07,982][starVLA.training.train_starvla][INFO] - Step 2838, Loss: {'train/loss': 0.011327741667628288, 'train/grad_norm_pre_clip': 33.85007858276367, 'timing/data': 0.00034234998747706413, 'timing/model': 0.3913581660017371, 'learning_rate/qwen_vl_interface': 4.3778941690913935e-06, 'learning_rate/action_model': 4.3778941690913934e-05, 'epoch': 0.21}) +[2026-10-03 19:30:08,882][starVLA.training.train_starvla][INFO] - Step 2839, Loss: {'train/loss': 0.0003975460131186992, 'train/grad_norm_pre_clip': 0.591548502445221, 'timing/data': 0.0003472199896350503, 'timing/model': 0.3914466660062317, 'learning_rate/qwen_vl_interface': 4.374900696794311e-06, 'learning_rate/action_model': 4.374900696794311e-05, 'epoch': 0.21}) +[2026-10-03 19:30:09,583][starVLA.training.train_starvla][INFO] - Step 2840, Loss: {'train/loss': 0.0006199250346980989, 'train/grad_norm_pre_clip': 0.35966578125953674, 'timing/data': 0.00023551000049337745, 'timing/model': 0.3924080349970609, 'learning_rate/qwen_vl_interface': 4.371907584217499e-06, 'learning_rate/action_model': 4.371907584217498e-05, 'epoch': 0.21}) +[2026-10-03 19:30:10,284][starVLA.training.train_starvla][INFO] - Step 2841, Loss: {'train/loss': 0.14451396465301514, 'train/grad_norm_pre_clip': 53.19240188598633, 'timing/data': 0.0003054400149267167, 'timing/model': 0.3920123159768991, 'learning_rate/qwen_vl_interface': 4.368914832591309e-06, 'learning_rate/action_model': 4.368914832591309e-05, 'epoch': 0.21}) +[2026-10-03 19:30:10,984][starVLA.training.train_starvla][INFO] - Step 2842, Loss: {'train/loss': 0.6189756393432617, 'train/grad_norm_pre_clip': 188.8260040283203, 'timing/data': 0.0003598000039346516, 'timing/model': 0.3908478160155937, 'learning_rate/qwen_vl_interface': 4.365922443145949e-06, 'learning_rate/action_model': 4.365922443145949e-05, 'epoch': 0.21}) +[2026-10-03 19:30:11,681][starVLA.training.train_starvla][INFO] - Step 2843, Loss: {'train/loss': 0.902135968208313, 'train/grad_norm_pre_clip': 95.40174102783203, 'timing/data': 0.0002748300030361861, 'timing/model': 0.3909007959882729, 'learning_rate/qwen_vl_interface': 4.362930417111479e-06, 'learning_rate/action_model': 4.362930417111479e-05, 'epoch': 0.21}) +[2026-10-03 19:30:12,581][starVLA.training.train_starvla][INFO] - Step 2844, Loss: {'train/loss': 0.0009885802865028381, 'train/grad_norm_pre_clip': 0.49085479974746704, 'timing/data': 0.0003479700244497508, 'timing/model': 0.5947462980111595, 'learning_rate/qwen_vl_interface': 4.359938755717807e-06, 'learning_rate/action_model': 4.359938755717807e-05, 'epoch': 0.21}) +[2026-10-03 19:30:13,280][starVLA.training.train_starvla][INFO] - Step 2845, Loss: {'train/loss': 0.0012729109730571508, 'train/grad_norm_pre_clip': 1.4856685400009155, 'timing/data': 0.000325599976349622, 'timing/model': 0.3915199860057328, 'learning_rate/qwen_vl_interface': 4.3569474601946934e-06, 'learning_rate/action_model': 4.356947460194694e-05, 'epoch': 0.21}) +[2026-10-03 19:30:13,979][starVLA.training.train_starvla][INFO] - Step 2846, Loss: {'train/loss': 0.46318408846855164, 'train/grad_norm_pre_clip': 43.00282287597656, 'timing/data': 0.0002543700102251023, 'timing/model': 0.3911577459948603, 'learning_rate/qwen_vl_interface': 4.3539565317717446e-06, 'learning_rate/action_model': 4.353956531771745e-05, 'epoch': 0.21}) +[2026-10-03 19:30:14,679][starVLA.training.train_starvla][INFO] - Step 2847, Loss: {'train/loss': 6.852393562439829e-05, 'train/grad_norm_pre_clip': 0.030817760154604912, 'timing/data': 0.00031676999060437083, 'timing/model': 0.39136745600262657, 'learning_rate/qwen_vl_interface': 4.350965971678419e-06, 'learning_rate/action_model': 4.3509659716784185e-05, 'epoch': 0.21}) +[2026-10-03 19:30:15,380][starVLA.training.train_starvla][INFO] - Step 2848, Loss: {'train/loss': 0.00020486622815951705, 'train/grad_norm_pre_clip': 0.1422034651041031, 'timing/data': 0.00034068000968545675, 'timing/model': 0.3914226659981068, 'learning_rate/qwen_vl_interface': 4.3479757811440225e-06, 'learning_rate/action_model': 4.347975781144022e-05, 'epoch': 0.21}) +[2026-10-03 19:30:16,282][starVLA.training.train_starvla][INFO] - Step 2849, Loss: {'train/loss': 0.014144180342555046, 'train/grad_norm_pre_clip': 10.159496307373047, 'timing/data': 0.0003015600086655468, 'timing/model': 0.5958127079939004, 'learning_rate/qwen_vl_interface': 4.34498596139771e-06, 'learning_rate/action_model': 4.34498596139771e-05, 'epoch': 0.21}) +[2026-10-03 19:30:16,985][starVLA.training.train_starvla][INFO] - Step 2850, Loss: {'train/loss': 0.005080518312752247, 'train/grad_norm_pre_clip': 4.273676872253418, 'timing/data': 0.00023519000387750566, 'timing/model': 0.39289549499517307, 'learning_rate/qwen_vl_interface': 4.341996513668483e-06, 'learning_rate/action_model': 4.3419965136684824e-05, 'epoch': 0.21}) +[2026-10-03 19:30:17,685][starVLA.training.train_starvla][INFO] - Step 2851, Loss: {'train/loss': 7.988418656168506e-05, 'train/grad_norm_pre_clip': 0.03235766291618347, 'timing/data': 0.00021185001241974533, 'timing/model': 0.3917388360132463, 'learning_rate/qwen_vl_interface': 4.33900743918519e-06, 'learning_rate/action_model': 4.33900743918519e-05, 'epoch': 0.21}) +[2026-10-03 19:30:18,386][starVLA.training.train_starvla][INFO] - Step 2852, Loss: {'train/loss': 0.0007245044689625502, 'train/grad_norm_pre_clip': 0.8950101137161255, 'timing/data': 0.00021133999689482152, 'timing/model': 0.39189702601288445, 'learning_rate/qwen_vl_interface': 4.336018739176529e-06, 'learning_rate/action_model': 4.33601873917653e-05, 'epoch': 0.21}) +[2026-10-03 19:30:19,087][starVLA.training.train_starvla][INFO] - Step 2853, Loss: {'train/loss': 7.979352085385472e-05, 'train/grad_norm_pre_clip': 0.030637696385383606, 'timing/data': 0.00020551000488922, 'timing/model': 0.3921831860207021, 'learning_rate/qwen_vl_interface': 4.333030414871041e-06, 'learning_rate/action_model': 4.33303041487104e-05, 'epoch': 0.21}) +[2026-10-03 19:30:19,787][starVLA.training.train_starvla][INFO] - Step 2854, Loss: {'train/loss': 2.7871501515619457e-05, 'train/grad_norm_pre_clip': 0.00909603200852871, 'timing/data': 0.00020124000729992986, 'timing/model': 0.3920202750014141, 'learning_rate/qwen_vl_interface': 4.330042467497111e-06, 'learning_rate/action_model': 4.330042467497111e-05, 'epoch': 0.21}) +[2026-10-03 19:30:20,683][starVLA.training.train_starvla][INFO] - Step 2855, Loss: {'train/loss': 0.3025919497013092, 'train/grad_norm_pre_clip': 73.08759307861328, 'timing/data': 0.00031954998848959804, 'timing/model': 0.3900827459874563, 'learning_rate/qwen_vl_interface': 4.327054898282973e-06, 'learning_rate/action_model': 4.3270548982829727e-05, 'epoch': 0.21}) +[2026-10-03 19:30:21,382][starVLA.training.train_starvla][INFO] - Step 2856, Loss: {'train/loss': 0.3982863426208496, 'train/grad_norm_pre_clip': 47.18324279785156, 'timing/data': 0.00020783001673407853, 'timing/model': 0.390280844992958, 'learning_rate/qwen_vl_interface': 4.324067708456703e-06, 'learning_rate/action_model': 4.3240677084567036e-05, 'epoch': 0.21}) +[2026-10-03 19:30:22,084][starVLA.training.train_starvla][INFO] - Step 2857, Loss: {'train/loss': 0.008855264633893967, 'train/grad_norm_pre_clip': 7.80233097076416, 'timing/data': 0.00021019999985583127, 'timing/model': 0.39362244500080124, 'learning_rate/qwen_vl_interface': 4.321080899246225e-06, 'learning_rate/action_model': 4.321080899246225e-05, 'epoch': 0.21}) +[2026-10-03 19:30:22,786][starVLA.training.train_starvla][INFO] - Step 2858, Loss: {'train/loss': 6.324616697384045e-05, 'train/grad_norm_pre_clip': 0.026691969484090805, 'timing/data': 0.00020533998031169176, 'timing/model': 0.39460131598752923, 'learning_rate/qwen_vl_interface': 4.318094471879299e-06, 'learning_rate/action_model': 4.3180944718792985e-05, 'epoch': 0.21}) +[2026-10-03 19:30:23,488][starVLA.training.train_starvla][INFO] - Step 2859, Loss: {'train/loss': 0.11729896813631058, 'train/grad_norm_pre_clip': 88.7015151977539, 'timing/data': 0.00021132000256329775, 'timing/model': 0.3929600059927907, 'learning_rate/qwen_vl_interface': 4.315108427583535e-06, 'learning_rate/action_model': 4.315108427583535e-05, 'epoch': 0.21}) +[2026-10-03 19:30:24,401][starVLA.training.train_starvla][INFO] - Step 2860, Loss: {'train/loss': 0.03171823173761368, 'train/grad_norm_pre_clip': 54.96543502807617, 'timing/data': 0.0004088900168426335, 'timing/model': 0.39047677599592134, 'learning_rate/qwen_vl_interface': 4.312122767586381e-06, 'learning_rate/action_model': 4.3121227675863815e-05, 'epoch': 0.21}) +[2026-10-03 19:30:25,103][starVLA.training.train_starvla][INFO] - Step 2861, Loss: {'train/loss': 0.127613827586174, 'train/grad_norm_pre_clip': 221.71188354492188, 'timing/data': 0.00020614999812096357, 'timing/model': 0.3940363959991373, 'learning_rate/qwen_vl_interface': 4.3091374931151324e-06, 'learning_rate/action_model': 4.309137493115132e-05, 'epoch': 0.21}) +[2026-10-03 19:30:25,802][starVLA.training.train_starvla][INFO] - Step 2862, Loss: {'train/loss': 0.38749074935913086, 'train/grad_norm_pre_clip': 180.69825744628906, 'timing/data': 0.00035027999547310174, 'timing/model': 0.39191230601863936, 'learning_rate/qwen_vl_interface': 4.306152605396921e-06, 'learning_rate/action_model': 4.306152605396921e-05, 'epoch': 0.21}) +[2026-10-03 19:30:26,504][starVLA.training.train_starvla][INFO] - Step 2863, Loss: {'train/loss': 5.7667009969009086e-05, 'train/grad_norm_pre_clip': 0.016767257824540138, 'timing/data': 0.0003527899971231818, 'timing/model': 0.3926250649965368, 'learning_rate/qwen_vl_interface': 4.30316810565872e-06, 'learning_rate/action_model': 4.3031681056587206e-05, 'epoch': 0.21}) +[2026-10-03 19:30:27,206][starVLA.training.train_starvla][INFO] - Step 2864, Loss: {'train/loss': 0.5066652297973633, 'train/grad_norm_pre_clip': 114.17363739013672, 'timing/data': 0.00021098999422974885, 'timing/model': 0.3932378659956157, 'learning_rate/qwen_vl_interface': 4.300183995127347e-06, 'learning_rate/action_model': 4.300183995127346e-05, 'epoch': 0.21}) +[2026-10-03 19:30:28,113][starVLA.training.train_starvla][INFO] - Step 2865, Loss: {'train/loss': 8.769521082285792e-05, 'train/grad_norm_pre_clip': 0.03355732560157776, 'timing/data': 0.0002619099977891892, 'timing/model': 0.5988340970070567, 'learning_rate/qwen_vl_interface': 4.297200275029454e-06, 'learning_rate/action_model': 4.297200275029454e-05, 'epoch': 0.21}) +[2026-10-03 19:30:28,815][starVLA.training.train_starvla][INFO] - Step 2866, Loss: {'train/loss': 0.00017856500926427543, 'train/grad_norm_pre_clip': 0.0813942402601242, 'timing/data': 0.00020055999630130827, 'timing/model': 0.39332520600873977, 'learning_rate/qwen_vl_interface': 4.29421694659154e-06, 'learning_rate/action_model': 4.294216946591539e-05, 'epoch': 0.21}) +[2026-10-03 19:30:29,517][starVLA.training.train_starvla][INFO] - Step 2867, Loss: {'train/loss': 0.2372114360332489, 'train/grad_norm_pre_clip': 412.7889404296875, 'timing/data': 0.0003345200093463063, 'timing/model': 0.3942196550196968, 'learning_rate/qwen_vl_interface': 4.2912340110399345e-06, 'learning_rate/action_model': 4.291234011039934e-05, 'epoch': 0.21}) +[2026-10-03 19:30:30,219][starVLA.training.train_starvla][INFO] - Step 2868, Loss: {'train/loss': 0.01579952798783779, 'train/grad_norm_pre_clip': 32.513092041015625, 'timing/data': 0.0003284199920017272, 'timing/model': 0.3923685259942431, 'learning_rate/qwen_vl_interface': 4.2882514696008125e-06, 'learning_rate/action_model': 4.2882514696008125e-05, 'epoch': 0.21}) +[2026-10-03 19:30:30,922][starVLA.training.train_starvla][INFO] - Step 2869, Loss: {'train/loss': 0.17938724160194397, 'train/grad_norm_pre_clip': 139.0662384033203, 'timing/data': 0.00025626999558880925, 'timing/model': 0.3948965150048025, 'learning_rate/qwen_vl_interface': 4.2852693235001815e-06, 'learning_rate/action_model': 4.2852693235001815e-05, 'epoch': 0.21}) +[2026-10-03 19:30:31,835][starVLA.training.train_starvla][INFO] - Step 2870, Loss: {'train/loss': 0.060807038098573685, 'train/grad_norm_pre_clip': 55.37917709350586, 'timing/data': 0.0003797899989876896, 'timing/model': 0.6024528570123948, 'learning_rate/qwen_vl_interface': 4.2822875739638915e-06, 'learning_rate/action_model': 4.282287573963891e-05, 'epoch': 0.21}) +[2026-10-03 19:30:32,536][starVLA.training.train_starvla][INFO] - Step 2871, Loss: {'train/loss': 0.45818713307380676, 'train/grad_norm_pre_clip': 63.386016845703125, 'timing/data': 0.00020037000649608672, 'timing/model': 0.3921843949938193, 'learning_rate/qwen_vl_interface': 4.279306222217625e-06, 'learning_rate/action_model': 4.279306222217626e-05, 'epoch': 0.21}) +[2026-10-03 19:30:33,234][starVLA.training.train_starvla][INFO] - Step 2872, Loss: {'train/loss': 0.736072838306427, 'train/grad_norm_pre_clip': 145.21957397460938, 'timing/data': 0.00020865999977104366, 'timing/model': 0.38887234599678777, 'learning_rate/qwen_vl_interface': 4.276325269486906e-06, 'learning_rate/action_model': 4.276325269486906e-05, 'epoch': 0.21}) +[2026-10-03 19:30:33,932][starVLA.training.train_starvla][INFO] - Step 2873, Loss: {'train/loss': 0.33107370138168335, 'train/grad_norm_pre_clip': 172.74600219726562, 'timing/data': 0.00026200999855063856, 'timing/model': 0.39023836600244977, 'learning_rate/qwen_vl_interface': 4.273344716997089e-06, 'learning_rate/action_model': 4.273344716997089e-05, 'epoch': 0.21}) +[2026-10-03 19:30:34,626][starVLA.training.train_starvla][INFO] - Step 2874, Loss: {'train/loss': 0.0002543542068451643, 'train/grad_norm_pre_clip': 0.10162189602851868, 'timing/data': 0.00020413001766428351, 'timing/model': 0.38872821599943563, 'learning_rate/qwen_vl_interface': 4.270364565973368e-06, 'learning_rate/action_model': 4.270364565973367e-05, 'epoch': 0.21}) +[2026-10-03 19:30:35,320][starVLA.training.train_starvla][INFO] - Step 2875, Loss: {'train/loss': 0.0011673006229102612, 'train/grad_norm_pre_clip': 0.5263858437538147, 'timing/data': 0.0003126900119241327, 'timing/model': 0.3882210259907879, 'learning_rate/qwen_vl_interface': 4.267384817640768e-06, 'learning_rate/action_model': 4.2673848176407685e-05, 'epoch': 0.21}) +[2026-10-03 19:30:36,218][starVLA.training.train_starvla][INFO] - Step 2876, Loss: {'train/loss': 0.08123213797807693, 'train/grad_norm_pre_clip': 66.4179916381836, 'timing/data': 0.00020752998534590006, 'timing/model': 0.3917635249963496, 'learning_rate/qwen_vl_interface': 4.264405473224155e-06, 'learning_rate/action_model': 4.2644054732241546e-05, 'epoch': 0.21}) +[2026-10-03 19:30:36,918][starVLA.training.train_starvla][INFO] - Step 2877, Loss: {'train/loss': 4.9610709538683295e-05, 'train/grad_norm_pre_clip': 0.011716632172465324, 'timing/data': 0.00021118999575264752, 'timing/model': 0.3927996059937868, 'learning_rate/qwen_vl_interface': 4.261426533948221e-06, 'learning_rate/action_model': 4.2614265339482213e-05, 'epoch': 0.21}) +[2026-10-03 19:30:37,615][starVLA.training.train_starvla][INFO] - Step 2878, Loss: {'train/loss': 0.03429612144827843, 'train/grad_norm_pre_clip': 43.05918502807617, 'timing/data': 0.00020683000911958516, 'timing/model': 0.39036679500713944, 'learning_rate/qwen_vl_interface': 4.258448001037499e-06, 'learning_rate/action_model': 4.2584480010374985e-05, 'epoch': 0.21}) +[2026-10-03 19:30:38,314][starVLA.training.train_starvla][INFO] - Step 2879, Loss: {'train/loss': 0.0006405740277841687, 'train/grad_norm_pre_clip': 0.5443968176841736, 'timing/data': 0.00032466999255120754, 'timing/model': 0.39023709497996606, 'learning_rate/qwen_vl_interface': 4.255469875716348e-06, 'learning_rate/action_model': 4.255469875716348e-05, 'epoch': 0.21}) +[2026-10-03 19:30:39,014][starVLA.training.train_starvla][INFO] - Step 2880, Loss: {'train/loss': 0.0011146035976707935, 'train/grad_norm_pre_clip': 0.725180983543396, 'timing/data': 0.0002141699951607734, 'timing/model': 0.3903183059883304, 'learning_rate/qwen_vl_interface': 4.252492159208966e-06, 'learning_rate/action_model': 4.252492159208966e-05, 'epoch': 0.21}) +[2026-10-03 19:30:39,914][starVLA.training.train_starvla][INFO] - Step 2881, Loss: {'train/loss': 0.00020170996140222996, 'train/grad_norm_pre_clip': 0.10736895352602005, 'timing/data': 0.00024496999685652554, 'timing/model': 0.5942438170022797, 'learning_rate/qwen_vl_interface': 4.2495148527393766e-06, 'learning_rate/action_model': 4.2495148527393766e-05, 'epoch': 0.21}) +[2026-10-03 19:30:40,613][starVLA.training.train_starvla][INFO] - Step 2882, Loss: {'train/loss': 0.0029895994812250137, 'train/grad_norm_pre_clip': 4.0909528732299805, 'timing/data': 0.0002007999864872545, 'timing/model': 0.39192052601720206, 'learning_rate/qwen_vl_interface': 4.246537957531441e-06, 'learning_rate/action_model': 4.2465379575314404e-05, 'epoch': 0.21}) +[2026-10-03 19:30:41,311][starVLA.training.train_starvla][INFO] - Step 2883, Loss: {'train/loss': 0.00019555956532713026, 'train/grad_norm_pre_clip': 0.17412222921848297, 'timing/data': 0.00020727000082843006, 'timing/model': 0.3900127859960776, 'learning_rate/qwen_vl_interface': 4.243561474808846e-06, 'learning_rate/action_model': 4.2435614748088456e-05, 'epoch': 0.21}) +[2026-10-03 19:30:42,009][starVLA.training.train_starvla][INFO] - Step 2884, Loss: {'train/loss': 4.870928023592569e-05, 'train/grad_norm_pre_clip': 0.014610414393246174, 'timing/data': 0.0003616700123529881, 'timing/model': 0.3895398570166435, 'learning_rate/qwen_vl_interface': 4.240585405795112e-06, 'learning_rate/action_model': 4.240585405795112e-05, 'epoch': 0.21}) +[2026-10-03 19:30:42,709][starVLA.training.train_starvla][INFO] - Step 2885, Loss: {'train/loss': 0.691058874130249, 'train/grad_norm_pre_clip': 55.228702545166016, 'timing/data': 0.0002590500225778669, 'timing/model': 0.39310305498656817, 'learning_rate/qwen_vl_interface': 4.237609751713589e-06, 'learning_rate/action_model': 4.2376097517135896e-05, 'epoch': 0.21}) +[2026-10-03 19:30:43,616][starVLA.training.train_starvla][INFO] - Step 2886, Loss: {'train/loss': 7.658503454877064e-05, 'train/grad_norm_pre_clip': 0.02858077362179756, 'timing/data': 0.0002835500054061413, 'timing/model': 0.5999132269935217, 'learning_rate/qwen_vl_interface': 4.234634513787455e-06, 'learning_rate/action_model': 4.2346345137874556e-05, 'epoch': 0.21}) +[2026-10-03 19:30:44,315][starVLA.training.train_starvla][INFO] - Step 2887, Loss: {'train/loss': 0.007405213546007872, 'train/grad_norm_pre_clip': 9.437823295593262, 'timing/data': 0.00031470999238081276, 'timing/model': 0.39078431500820443, 'learning_rate/qwen_vl_interface': 4.2316596932397195e-06, 'learning_rate/action_model': 4.2316596932397195e-05, 'epoch': 0.21}) +[2026-10-03 19:30:45,012][starVLA.training.train_starvla][INFO] - Step 2888, Loss: {'train/loss': 0.32424235343933105, 'train/grad_norm_pre_clip': 119.10269165039062, 'timing/data': 0.00033214999712072313, 'timing/model': 0.39008277698303573, 'learning_rate/qwen_vl_interface': 4.22868529129322e-06, 'learning_rate/action_model': 4.22868529129322e-05, 'epoch': 0.21}) +[2026-10-03 19:30:45,710][starVLA.training.train_starvla][INFO] - Step 2889, Loss: {'train/loss': 0.00039411860052496195, 'train/grad_norm_pre_clip': 0.2614402770996094, 'timing/data': 0.0002004699781537056, 'timing/model': 0.3913797659915872, 'learning_rate/qwen_vl_interface': 4.2257113091706165e-06, 'learning_rate/action_model': 4.2257113091706165e-05, 'epoch': 0.21}) +[2026-10-03 19:30:46,406][starVLA.training.train_starvla][INFO] - Step 2890, Loss: {'train/loss': 0.0007990596350282431, 'train/grad_norm_pre_clip': 0.5810508728027344, 'timing/data': 0.000324589986121282, 'timing/model': 0.3890891349874437, 'learning_rate/qwen_vl_interface': 4.222737748094404e-06, 'learning_rate/action_model': 4.222737748094404e-05, 'epoch': 0.21}) +[2026-10-03 19:30:47,107][starVLA.training.train_starvla][INFO] - Step 2891, Loss: {'train/loss': 0.23212409019470215, 'train/grad_norm_pre_clip': 81.80233764648438, 'timing/data': 0.0003359300026204437, 'timing/model': 0.3920818160113413, 'learning_rate/qwen_vl_interface': 4.219764609286901e-06, 'learning_rate/action_model': 4.219764609286901e-05, 'epoch': 0.21}) +[2026-10-03 19:30:48,015][starVLA.training.train_starvla][INFO] - Step 2892, Loss: {'train/loss': 0.008786756545305252, 'train/grad_norm_pre_clip': 11.203017234802246, 'timing/data': 0.00021659000776708126, 'timing/model': 0.3915083759929985, 'learning_rate/qwen_vl_interface': 4.21679189397025e-06, 'learning_rate/action_model': 4.21679189397025e-05, 'epoch': 0.21}) +[2026-10-03 19:30:48,713][starVLA.training.train_starvla][INFO] - Step 2893, Loss: {'train/loss': 0.033435508608818054, 'train/grad_norm_pre_clip': 51.96417236328125, 'timing/data': 0.00020198000129312277, 'timing/model': 0.3920655859983526, 'learning_rate/qwen_vl_interface': 4.213819603366424e-06, 'learning_rate/action_model': 4.213819603366424e-05, 'epoch': 0.21}) +[2026-10-03 19:30:49,413][starVLA.training.train_starvla][INFO] - Step 2894, Loss: {'train/loss': 0.00924227200448513, 'train/grad_norm_pre_clip': 5.190800189971924, 'timing/data': 0.0002628999936860055, 'timing/model': 0.39157653599977493, 'learning_rate/qwen_vl_interface': 4.210847738697219e-06, 'learning_rate/action_model': 4.210847738697218e-05, 'epoch': 0.21}) +[2026-10-03 19:30:50,113][starVLA.training.train_starvla][INFO] - Step 2895, Loss: {'train/loss': 0.14654892683029175, 'train/grad_norm_pre_clip': 195.14266967773438, 'timing/data': 0.00031849998049438, 'timing/model': 0.39281312498496845, 'learning_rate/qwen_vl_interface': 4.207876301184256e-06, 'learning_rate/action_model': 4.207876301184256e-05, 'epoch': 0.21}) +[2026-10-03 19:30:50,815][starVLA.training.train_starvla][INFO] - Step 2896, Loss: {'train/loss': 0.03696611523628235, 'train/grad_norm_pre_clip': 113.7008056640625, 'timing/data': 0.0003592100110836327, 'timing/model': 0.3923409049748443, 'learning_rate/qwen_vl_interface': 4.2049052920489825e-06, 'learning_rate/action_model': 4.2049052920489825e-05, 'epoch': 0.21}) +[2026-10-03 19:30:51,728][starVLA.training.train_starvla][INFO] - Step 2897, Loss: {'train/loss': 9.646409307606518e-05, 'train/grad_norm_pre_clip': 0.03188485652208328, 'timing/data': 0.00021018998813815415, 'timing/model': 0.39307883600122295, 'learning_rate/qwen_vl_interface': 4.201934712512666e-06, 'learning_rate/action_model': 4.2019347125126656e-05, 'epoch': 0.21}) +[2026-10-03 19:30:52,426][starVLA.training.train_starvla][INFO] - Step 2898, Loss: {'train/loss': 3.77574069716502e-05, 'train/grad_norm_pre_clip': 0.01328292116522789, 'timing/data': 0.00020740000763908029, 'timing/model': 0.3897574959846679, 'learning_rate/qwen_vl_interface': 4.1989645637964e-06, 'learning_rate/action_model': 4.1989645637964e-05, 'epoch': 0.21}) +[2026-10-03 19:30:53,126][starVLA.training.train_starvla][INFO] - Step 2899, Loss: {'train/loss': 0.20967251062393188, 'train/grad_norm_pre_clip': 76.16040802001953, 'timing/data': 0.0002196200075559318, 'timing/model': 0.39131085597909987, 'learning_rate/qwen_vl_interface': 4.195994847121101e-06, 'learning_rate/action_model': 4.195994847121101e-05, 'epoch': 0.21}) +[2026-10-03 19:30:53,828][starVLA.training.train_starvla][INFO] - Step 2900, Loss: {'train/loss': 0.0033515726681798697, 'train/grad_norm_pre_clip': 3.354340076446533, 'timing/data': 0.00030578000587411225, 'timing/model': 0.392496246000519, 'learning_rate/qwen_vl_interface': 4.193025563707507e-06, 'learning_rate/action_model': 4.1930255637075066e-05, 'epoch': 0.21}) +[2026-10-03 19:30:54,528][starVLA.training.train_starvla][INFO] - Step 2901, Loss: {'train/loss': 4.457305112737231e-05, 'train/grad_norm_pre_clip': 0.015252883546054363, 'timing/data': 0.0002841500099748373, 'timing/model': 0.39228078597807325, 'learning_rate/qwen_vl_interface': 4.19005671477618e-06, 'learning_rate/action_model': 4.1900567147761794e-05, 'epoch': 0.21}) +[2026-10-03 19:30:55,434][starVLA.training.train_starvla][INFO] - Step 2902, Loss: {'train/loss': 0.30686429142951965, 'train/grad_norm_pre_clip': 151.6440887451172, 'timing/data': 0.00021005998132750392, 'timing/model': 0.3908621049777139, 'learning_rate/qwen_vl_interface': 4.1870883015474975e-06, 'learning_rate/action_model': 4.187088301547497e-05, 'epoch': 0.21}) +[2026-10-03 19:30:56,131][starVLA.training.train_starvla][INFO] - Step 2903, Loss: {'train/loss': 0.00019890081603080034, 'train/grad_norm_pre_clip': 0.3490179777145386, 'timing/data': 0.00020793001749552786, 'timing/model': 0.3908760359918233, 'learning_rate/qwen_vl_interface': 4.1841203252416655e-06, 'learning_rate/action_model': 4.1841203252416655e-05, 'epoch': 0.21}) +[2026-10-03 19:30:56,831][starVLA.training.train_starvla][INFO] - Step 2904, Loss: {'train/loss': 0.000980588374659419, 'train/grad_norm_pre_clip': 1.6118688583374023, 'timing/data': 0.0002095300005748868, 'timing/model': 0.39105832600034773, 'learning_rate/qwen_vl_interface': 4.181152787078705e-06, 'learning_rate/action_model': 4.1811527870787056e-05, 'epoch': 0.21}) +[2026-10-03 19:30:57,532][starVLA.training.train_starvla][INFO] - Step 2905, Loss: {'train/loss': 0.1856689304113388, 'train/grad_norm_pre_clip': 110.34504699707031, 'timing/data': 0.00031920999754220247, 'timing/model': 0.39345123499515466, 'learning_rate/qwen_vl_interface': 4.178185688278462e-06, 'learning_rate/action_model': 4.1781856882784616e-05, 'epoch': 0.21}) +[2026-10-03 19:30:58,229][starVLA.training.train_starvla][INFO] - Step 2906, Loss: {'train/loss': 0.2734561264514923, 'train/grad_norm_pre_clip': 181.58352661132812, 'timing/data': 0.0002382999809924513, 'timing/model': 0.3910098460037261, 'learning_rate/qwen_vl_interface': 4.175219030060595e-06, 'learning_rate/action_model': 4.175219030060595e-05, 'epoch': 0.21}) +[2026-10-03 19:30:59,128][starVLA.training.train_starvla][INFO] - Step 2907, Loss: {'train/loss': 0.752206563949585, 'train/grad_norm_pre_clip': 38.591827392578125, 'timing/data': 0.00020882001263089478, 'timing/model': 0.5944024679774884, 'learning_rate/qwen_vl_interface': 4.1722528136445875e-06, 'learning_rate/action_model': 4.172252813644588e-05, 'epoch': 0.21}) +[2026-10-03 19:30:59,803][starVLA.training.train_starvla][INFO] - Step 2908, Loss: {'train/loss': 0.40140411257743835, 'train/grad_norm_pre_clip': 17.54194450378418, 'timing/data': 0.00021769999875687063, 'timing/model': 0.39151425601448864, 'learning_rate/qwen_vl_interface': 4.169287040249738e-06, 'learning_rate/action_model': 4.169287040249738e-05, 'epoch': 0.21}) +[2026-10-03 19:31:00,501][starVLA.training.train_starvla][INFO] - Step 2909, Loss: {'train/loss': 0.00012939076987095177, 'train/grad_norm_pre_clip': 0.09217751771211624, 'timing/data': 0.0003121799963992089, 'timing/model': 0.3900900750013534, 'learning_rate/qwen_vl_interface': 4.166321711095164e-06, 'learning_rate/action_model': 4.166321711095164e-05, 'epoch': 0.21}) +[2026-10-03 19:31:01,197][starVLA.training.train_starvla][INFO] - Step 2910, Loss: {'train/loss': 0.010906200855970383, 'train/grad_norm_pre_clip': 16.810895919799805, 'timing/data': 0.00023696001153439283, 'timing/model': 0.3897438260028139, 'learning_rate/qwen_vl_interface': 4.1633568273998e-06, 'learning_rate/action_model': 4.1633568273998e-05, 'epoch': 0.21}) +[2026-10-03 19:31:01,895][starVLA.training.train_starvla][INFO] - Step 2911, Loss: {'train/loss': 2.601606320240535e-05, 'train/grad_norm_pre_clip': 0.009269959293305874, 'timing/data': 0.00021341998944990337, 'timing/model': 0.3894364050065633, 'learning_rate/qwen_vl_interface': 4.1603923903823995e-06, 'learning_rate/action_model': 4.160392390382399e-05, 'epoch': 0.21}) +[2026-10-03 19:31:02,593][starVLA.training.train_starvla][INFO] - Step 2912, Loss: {'train/loss': 0.6842735409736633, 'train/grad_norm_pre_clip': 354.6656188964844, 'timing/data': 0.0003262000100221485, 'timing/model': 0.39168380497721955, 'learning_rate/qwen_vl_interface': 4.1574284012615276e-06, 'learning_rate/action_model': 4.157428401261528e-05, 'epoch': 0.21}) +[2026-10-03 19:31:03,499][starVLA.training.train_starvla][INFO] - Step 2913, Loss: {'train/loss': 0.04312810301780701, 'train/grad_norm_pre_clip': 59.34871292114258, 'timing/data': 0.0003284900158178061, 'timing/model': 0.390981214994099, 'learning_rate/qwen_vl_interface': 4.154464861255571e-06, 'learning_rate/action_model': 4.154464861255571e-05, 'epoch': 0.21}) +[2026-10-03 19:31:04,196][starVLA.training.train_starvla][INFO] - Step 2914, Loss: {'train/loss': 0.0024509718641638756, 'train/grad_norm_pre_clip': 2.0233676433563232, 'timing/data': 0.0003339389804750681, 'timing/model': 0.3911948459863197, 'learning_rate/qwen_vl_interface': 4.151501771582728e-06, 'learning_rate/action_model': 4.151501771582728e-05, 'epoch': 0.21}) +[2026-10-03 19:31:04,894][starVLA.training.train_starvla][INFO] - Step 2915, Loss: {'train/loss': 0.0023817401379346848, 'train/grad_norm_pre_clip': 3.2352428436279297, 'timing/data': 0.00031339001725427806, 'timing/model': 0.39032536500599235, 'learning_rate/qwen_vl_interface': 4.148539133461012e-06, 'learning_rate/action_model': 4.1485391334610116e-05, 'epoch': 0.21}) +[2026-10-03 19:31:05,594][starVLA.training.train_starvla][INFO] - Step 2916, Loss: {'train/loss': 0.00011497756349854171, 'train/grad_norm_pre_clip': 0.07846119999885559, 'timing/data': 0.0003300500102341175, 'timing/model': 0.3918398950190749, 'learning_rate/qwen_vl_interface': 4.145576948108254e-06, 'learning_rate/action_model': 4.145576948108253e-05, 'epoch': 0.21}) +[2026-10-03 19:31:06,293][starVLA.training.train_starvla][INFO] - Step 2917, Loss: {'train/loss': 7.22298864275217e-05, 'train/grad_norm_pre_clip': 0.04528458043932915, 'timing/data': 0.00020693999249488115, 'timing/model': 0.39187416501226835, 'learning_rate/qwen_vl_interface': 4.142615216742093e-06, 'learning_rate/action_model': 4.142615216742093e-05, 'epoch': 0.21}) +[2026-10-03 19:31:07,201][starVLA.training.train_starvla][INFO] - Step 2918, Loss: {'train/loss': 5.130837962497026e-05, 'train/grad_norm_pre_clip': 0.01975986361503601, 'timing/data': 0.00021035998361185193, 'timing/model': 0.3897426649928093, 'learning_rate/qwen_vl_interface': 4.139653940579988e-06, 'learning_rate/action_model': 4.1396539405799874e-05, 'epoch': 0.21}) +[2026-10-03 19:31:07,901][starVLA.training.train_starvla][INFO] - Step 2919, Loss: {'train/loss': 0.4748845100402832, 'train/grad_norm_pre_clip': 36.1795768737793, 'timing/data': 0.0002075600204989314, 'timing/model': 0.39156491501489654, 'learning_rate/qwen_vl_interface': 4.136693120839206e-06, 'learning_rate/action_model': 4.136693120839205e-05, 'epoch': 0.21}) +[2026-10-03 19:31:08,598][starVLA.training.train_starvla][INFO] - Step 2920, Loss: {'train/loss': 2.571851473476272e-05, 'train/grad_norm_pre_clip': 0.009330440312623978, 'timing/data': 0.00020807000692002475, 'timing/model': 0.3900171960121952, 'learning_rate/qwen_vl_interface': 4.133732758736828e-06, 'learning_rate/action_model': 4.133732758736828e-05, 'epoch': 0.21}) +[2026-10-03 19:31:09,298][starVLA.training.train_starvla][INFO] - Step 2921, Loss: {'train/loss': 0.9869953393936157, 'train/grad_norm_pre_clip': 61.821876525878906, 'timing/data': 0.000282110006082803, 'timing/model': 0.3928862860193476, 'learning_rate/qwen_vl_interface': 4.13077285548975e-06, 'learning_rate/action_model': 4.13077285548975e-05, 'epoch': 0.21}) +[2026-10-03 19:31:09,996][starVLA.training.train_starvla][INFO] - Step 2922, Loss: {'train/loss': 0.040769051760435104, 'train/grad_norm_pre_clip': 60.23540496826172, 'timing/data': 0.00020271999528631568, 'timing/model': 0.3900219459901564, 'learning_rate/qwen_vl_interface': 4.127813412314672e-06, 'learning_rate/action_model': 4.1278134123146725e-05, 'epoch': 0.21}) +[2026-10-03 19:31:10,896][starVLA.training.train_starvla][INFO] - Step 2923, Loss: {'train/loss': 0.02437366172671318, 'train/grad_norm_pre_clip': 34.813533782958984, 'timing/data': 0.00020402000518515706, 'timing/model': 0.5916021080047358, 'learning_rate/qwen_vl_interface': 4.124854430428113e-06, 'learning_rate/action_model': 4.1248544304281126e-05, 'epoch': 0.21}) +[2026-10-03 19:31:11,596][starVLA.training.train_starvla][INFO] - Step 2924, Loss: {'train/loss': 0.2849057614803314, 'train/grad_norm_pre_clip': 47.23659896850586, 'timing/data': 0.0002368300047237426, 'timing/model': 0.3917363850050606, 'learning_rate/qwen_vl_interface': 4.121895911046397e-06, 'learning_rate/action_model': 4.121895911046396e-05, 'epoch': 0.21}) +[2026-10-03 19:31:12,300][starVLA.training.train_starvla][INFO] - Step 2925, Loss: {'train/loss': 0.0013916823081672192, 'train/grad_norm_pre_clip': 0.6018142104148865, 'timing/data': 0.00034216998028568923, 'timing/model': 0.39366284498828463, 'learning_rate/qwen_vl_interface': 4.118937855385659e-06, 'learning_rate/action_model': 4.1189378553856584e-05, 'epoch': 0.21}) +[2026-10-03 19:31:12,999][starVLA.training.train_starvla][INFO] - Step 2926, Loss: {'train/loss': 0.4721095860004425, 'train/grad_norm_pre_clip': 184.8933563232422, 'timing/data': 0.00021458000992424786, 'timing/model': 0.3922625549894292, 'learning_rate/qwen_vl_interface': 4.115980264661844e-06, 'learning_rate/action_model': 4.115980264661844e-05, 'epoch': 0.21}) +[2026-10-03 19:31:13,697][starVLA.training.train_starvla][INFO] - Step 2927, Loss: {'train/loss': 0.00019076380704063922, 'train/grad_norm_pre_clip': 0.1900337040424347, 'timing/data': 0.0002134899841621518, 'timing/model': 0.38994365502730943, 'learning_rate/qwen_vl_interface': 4.113023140090709e-06, 'learning_rate/action_model': 4.113023140090709e-05, 'epoch': 0.21}) +[2026-10-03 19:31:14,603][starVLA.training.train_starvla][INFO] - Step 2928, Loss: {'train/loss': 0.0051928493194282055, 'train/grad_norm_pre_clip': 8.151473045349121, 'timing/data': 0.00034826999763026834, 'timing/model': 0.5976807769911829, 'learning_rate/qwen_vl_interface': 4.110066482887812e-06, 'learning_rate/action_model': 4.110066482887812e-05, 'epoch': 0.21}) +[2026-10-03 19:31:15,300][starVLA.training.train_starvla][INFO] - Step 2929, Loss: {'train/loss': 0.006080209743231535, 'train/grad_norm_pre_clip': 3.8393375873565674, 'timing/data': 0.00029178999830037355, 'timing/model': 0.39024131500627846, 'learning_rate/qwen_vl_interface': 4.107110294268525e-06, 'learning_rate/action_model': 4.107110294268525e-05, 'epoch': 0.21}) +[2026-10-03 19:31:15,995][starVLA.training.train_starvla][INFO] - Step 2930, Loss: {'train/loss': 0.00022860683384351432, 'train/grad_norm_pre_clip': 0.2141978144645691, 'timing/data': 0.0003737199876923114, 'timing/model': 0.38900955600547604, 'learning_rate/qwen_vl_interface': 4.104154575448026e-06, 'learning_rate/action_model': 4.104154575448025e-05, 'epoch': 0.21}) +[2026-10-03 19:31:16,691][starVLA.training.train_starvla][INFO] - Step 2931, Loss: {'train/loss': 0.00029174500377848744, 'train/grad_norm_pre_clip': 0.0814826563000679, 'timing/data': 0.0003338899987284094, 'timing/model': 0.38940873500541784, 'learning_rate/qwen_vl_interface': 4.101199327641297e-06, 'learning_rate/action_model': 4.101199327641297e-05, 'epoch': 0.21}) +[2026-10-03 19:31:17,392][starVLA.training.train_starvla][INFO] - Step 2932, Loss: {'train/loss': 0.13098520040512085, 'train/grad_norm_pre_clip': 123.44590759277344, 'timing/data': 0.00021324001136235893, 'timing/model': 0.39176044598571025, 'learning_rate/qwen_vl_interface': 4.098244552063133e-06, 'learning_rate/action_model': 4.098244552063132e-05, 'epoch': 0.21}) +[2026-10-03 19:31:18,297][starVLA.training.train_starvla][INFO] - Step 2933, Loss: {'train/loss': 0.0015603736974298954, 'train/grad_norm_pre_clip': 1.3175462484359741, 'timing/data': 0.00030865997541695833, 'timing/model': 0.5987911169941071, 'learning_rate/qwen_vl_interface': 4.0952902499281254e-06, 'learning_rate/action_model': 4.095290249928125e-05, 'epoch': 0.21}) +[2026-10-03 19:31:18,994][starVLA.training.train_starvla][INFO] - Step 2934, Loss: {'train/loss': 0.6720494627952576, 'train/grad_norm_pre_clip': 30.397890090942383, 'timing/data': 0.00021222000941634178, 'timing/model': 0.38994677501614206, 'learning_rate/qwen_vl_interface': 4.092336422450681e-06, 'learning_rate/action_model': 4.0923364224506804e-05, 'epoch': 0.21}) +[2026-10-03 19:31:19,695][starVLA.training.train_starvla][INFO] - Step 2935, Loss: {'train/loss': 0.00015862256987020373, 'train/grad_norm_pre_clip': 0.02629978582262993, 'timing/data': 0.00033676999737508595, 'timing/model': 0.3918629060208332, 'learning_rate/qwen_vl_interface': 4.089383070845003e-06, 'learning_rate/action_model': 4.089383070845003e-05, 'epoch': 0.21}) +[2026-10-03 19:31:20,394][starVLA.training.train_starvla][INFO] - Step 2936, Loss: {'train/loss': 0.0007710736826993525, 'train/grad_norm_pre_clip': 0.4933050572872162, 'timing/data': 0.00021430998458527029, 'timing/model': 0.39110211600200273, 'learning_rate/qwen_vl_interface': 4.086430196325106e-06, 'learning_rate/action_model': 4.086430196325106e-05, 'epoch': 0.21}) +[2026-10-03 19:31:21,096][starVLA.training.train_starvla][INFO] - Step 2937, Loss: {'train/loss': 0.0008293322753161192, 'train/grad_norm_pre_clip': 1.1677367687225342, 'timing/data': 0.00038508998113684356, 'timing/model': 0.39313812501495704, 'learning_rate/qwen_vl_interface': 4.083477800104802e-06, 'learning_rate/action_model': 4.083477800104802e-05, 'epoch': 0.21}) +[2026-10-03 19:31:21,793][starVLA.training.train_starvla][INFO] - Step 2938, Loss: {'train/loss': 0.00013296279939822853, 'train/grad_norm_pre_clip': 0.05288991332054138, 'timing/data': 0.0003936300054192543, 'timing/model': 0.3902795049943961, 'learning_rate/qwen_vl_interface': 4.080525883397711e-06, 'learning_rate/action_model': 4.080525883397711e-05, 'epoch': 0.22}) +[2026-10-03 19:31:22,697][starVLA.training.train_starvla][INFO] - Step 2939, Loss: {'train/loss': 0.0447983555495739, 'train/grad_norm_pre_clip': 35.062686920166016, 'timing/data': 0.0003614300221670419, 'timing/model': 0.3923821060161572, 'learning_rate/qwen_vl_interface': 4.077574447417256e-06, 'learning_rate/action_model': 4.077574447417256e-05, 'epoch': 0.22}) +[2026-10-03 19:31:23,397][starVLA.training.train_starvla][INFO] - Step 2940, Loss: {'train/loss': 0.0003210062859579921, 'train/grad_norm_pre_clip': 0.15376080572605133, 'timing/data': 0.000209210003959015, 'timing/model': 0.3908077759842854, 'learning_rate/qwen_vl_interface': 4.074623493376657e-06, 'learning_rate/action_model': 4.074623493376657e-05, 'epoch': 0.22}) +[2026-10-03 19:31:24,100][starVLA.training.train_starvla][INFO] - Step 2941, Loss: {'train/loss': 0.0003242433595005423, 'train/grad_norm_pre_clip': 0.3480803668498993, 'timing/data': 0.00022520998027175665, 'timing/model': 0.39440377498976886, 'learning_rate/qwen_vl_interface': 4.071673022488944e-06, 'learning_rate/action_model': 4.071673022488944e-05, 'epoch': 0.22}) +[2026-10-03 19:31:24,801][starVLA.training.train_starvla][INFO] - Step 2942, Loss: {'train/loss': 0.00018893563537858427, 'train/grad_norm_pre_clip': 0.08470100909471512, 'timing/data': 0.00020688999211415648, 'timing/model': 0.39283683500252664, 'learning_rate/qwen_vl_interface': 4.06872303596694e-06, 'learning_rate/action_model': 4.06872303596694e-05, 'epoch': 0.22}) +[2026-10-03 19:31:25,500][starVLA.training.train_starvla][INFO] - Step 2943, Loss: {'train/loss': 0.0005389763973653316, 'train/grad_norm_pre_clip': 0.8040160536766052, 'timing/data': 0.0003330599865876138, 'timing/model': 0.39251605499885045, 'learning_rate/qwen_vl_interface': 4.065773535023277e-06, 'learning_rate/action_model': 4.0657735350232766e-05, 'epoch': 0.22}) +[2026-10-03 19:31:26,405][starVLA.training.train_starvla][INFO] - Step 2944, Loss: {'train/loss': 0.00010146284330403432, 'train/grad_norm_pre_clip': 0.03644377738237381, 'timing/data': 0.0002099999983329326, 'timing/model': 0.5947334670054261, 'learning_rate/qwen_vl_interface': 4.062824520870382e-06, 'learning_rate/action_model': 4.062824520870381e-05, 'epoch': 0.22}) +[2026-10-03 19:31:27,107][starVLA.training.train_starvla][INFO] - Step 2945, Loss: {'train/loss': 0.01057711523026228, 'train/grad_norm_pre_clip': 8.291478157043457, 'timing/data': 0.00025270000332966447, 'timing/model': 0.39235264499438927, 'learning_rate/qwen_vl_interface': 4.059875994720481e-06, 'learning_rate/action_model': 4.059875994720481e-05, 'epoch': 0.22}) +[2026-10-03 19:31:27,807][starVLA.training.train_starvla][INFO] - Step 2946, Loss: {'train/loss': 0.019587211310863495, 'train/grad_norm_pre_clip': 33.36252212524414, 'timing/data': 0.0002731700078584254, 'timing/model': 0.39190534499357454, 'learning_rate/qwen_vl_interface': 4.056927957785605e-06, 'learning_rate/action_model': 4.056927957785605e-05, 'epoch': 0.22}) +[2026-10-03 19:31:28,507][starVLA.training.train_starvla][INFO] - Step 2947, Loss: {'train/loss': 4.2822059185709804e-05, 'train/grad_norm_pre_clip': 0.012615525163710117, 'timing/data': 0.0002085799933411181, 'timing/model': 0.39119023497914895, 'learning_rate/qwen_vl_interface': 4.0539804112775785e-06, 'learning_rate/action_model': 4.053980411277578e-05, 'epoch': 0.22}) +[2026-10-03 19:31:29,209][starVLA.training.train_starvla][INFO] - Step 2948, Loss: {'train/loss': 0.0013851531548425555, 'train/grad_norm_pre_clip': 1.019822120666504, 'timing/data': 0.00026063999393954873, 'timing/model': 0.3919303650036454, 'learning_rate/qwen_vl_interface': 4.051033356408027e-06, 'learning_rate/action_model': 4.0510333564080264e-05, 'epoch': 0.22}) +[2026-10-03 19:31:29,911][starVLA.training.train_starvla][INFO] - Step 2949, Loss: {'train/loss': 0.22585424780845642, 'train/grad_norm_pre_clip': 39.139984130859375, 'timing/data': 0.0002098300028592348, 'timing/model': 0.39447477500652894, 'learning_rate/qwen_vl_interface': 4.048086794388375e-06, 'learning_rate/action_model': 4.048086794388375e-05, 'epoch': 0.22}) +[2026-10-03 19:31:30,813][starVLA.training.train_starvla][INFO] - Step 2950, Loss: {'train/loss': 0.03377178683876991, 'train/grad_norm_pre_clip': 44.14974594116211, 'timing/data': 0.0004113409959245473, 'timing/model': 0.3901230749906972, 'learning_rate/qwen_vl_interface': 4.045140726429841e-06, 'learning_rate/action_model': 4.0451407264298406e-05, 'epoch': 0.22}) +[2026-10-03 19:31:31,512][starVLA.training.train_starvla][INFO] - Step 2951, Loss: {'train/loss': 0.3865062892436981, 'train/grad_norm_pre_clip': 117.84624481201172, 'timing/data': 0.0003124799986835569, 'timing/model': 0.3909624849911779, 'learning_rate/qwen_vl_interface': 4.042195153743443e-06, 'learning_rate/action_model': 4.042195153743443e-05, 'epoch': 0.22}) +[2026-10-03 19:31:32,215][starVLA.training.train_starvla][INFO] - Step 2952, Loss: {'train/loss': 0.001517973723821342, 'train/grad_norm_pre_clip': 0.5812886953353882, 'timing/data': 0.00035814999137073755, 'timing/model': 0.39277838499401696, 'learning_rate/qwen_vl_interface': 4.039250077539993e-06, 'learning_rate/action_model': 4.039250077539993e-05, 'epoch': 0.22}) +[2026-10-03 19:31:32,918][starVLA.training.train_starvla][INFO] - Step 2953, Loss: {'train/loss': 2.5763074518181384e-05, 'train/grad_norm_pre_clip': 0.005317742936313152, 'timing/data': 0.00021345997811295092, 'timing/model': 0.39439052500529215, 'learning_rate/qwen_vl_interface': 4.036305499030104e-06, 'learning_rate/action_model': 4.0363054990301036e-05, 'epoch': 0.22}) +[2026-10-03 19:31:33,621][starVLA.training.train_starvla][INFO] - Step 2954, Loss: {'train/loss': 0.0534091517329216, 'train/grad_norm_pre_clip': 220.10501098632812, 'timing/data': 0.0003221200022380799, 'timing/model': 0.39276144601171836, 'learning_rate/qwen_vl_interface': 4.033361419424178e-06, 'learning_rate/action_model': 4.033361419424178e-05, 'epoch': 0.22}) +[2026-10-03 19:31:34,525][starVLA.training.train_starvla][INFO] - Step 2955, Loss: {'train/loss': 0.09186941385269165, 'train/grad_norm_pre_clip': 122.97911071777344, 'timing/data': 0.0002903800050262362, 'timing/model': 0.39130456498241983, 'learning_rate/qwen_vl_interface': 4.030417839932418e-06, 'learning_rate/action_model': 4.030417839932417e-05, 'epoch': 0.22}) +[2026-10-03 19:31:35,226][starVLA.training.train_starvla][INFO] - Step 2956, Loss: {'train/loss': 1.9445576981524937e-05, 'train/grad_norm_pre_clip': 0.005714672617614269, 'timing/data': 0.00032418998307548463, 'timing/model': 0.3913253060018178, 'learning_rate/qwen_vl_interface': 4.027474761764815e-06, 'learning_rate/action_model': 4.027474761764814e-05, 'epoch': 0.22}) +[2026-10-03 19:31:35,930][starVLA.training.train_starvla][INFO] - Step 2957, Loss: {'train/loss': 0.12422304600477219, 'train/grad_norm_pre_clip': 134.52435302734375, 'timing/data': 0.0004044900124426931, 'timing/model': 0.3938078850042075, 'learning_rate/qwen_vl_interface': 4.024532186131158e-06, 'learning_rate/action_model': 4.0245321861311575e-05, 'epoch': 0.22}) +[2026-10-03 19:31:36,630][starVLA.training.train_starvla][INFO] - Step 2958, Loss: {'train/loss': 0.0086422860622406, 'train/grad_norm_pre_clip': 25.477115631103516, 'timing/data': 0.0002077799872495234, 'timing/model': 0.39189603601698764, 'learning_rate/qwen_vl_interface': 4.021590114241029e-06, 'learning_rate/action_model': 4.021590114241029e-05, 'epoch': 0.22}) +[2026-10-03 19:31:37,330][starVLA.training.train_starvla][INFO] - Step 2959, Loss: {'train/loss': 7.957252819323912e-05, 'train/grad_norm_pre_clip': 0.028691818937659264, 'timing/data': 0.0003203699889127165, 'timing/model': 0.3925987640104722, 'learning_rate/qwen_vl_interface': 4.018648547303804e-06, 'learning_rate/action_model': 4.0186485473038036e-05, 'epoch': 0.22}) +[2026-10-03 19:31:38,240][starVLA.training.train_starvla][INFO] - Step 2960, Loss: {'train/loss': 0.0016961815999820828, 'train/grad_norm_pre_clip': 1.34250009059906, 'timing/data': 0.0003522800107020885, 'timing/model': 0.5990399069851264, 'learning_rate/qwen_vl_interface': 4.015707486528649e-06, 'learning_rate/action_model': 4.015707486528649e-05, 'epoch': 0.22}) +[2026-10-03 19:31:38,937][starVLA.training.train_starvla][INFO] - Step 2961, Loss: {'train/loss': 3.7019577575847507e-05, 'train/grad_norm_pre_clip': 0.013348299078643322, 'timing/data': 0.00038178000249899924, 'timing/model': 0.3917293449921999, 'learning_rate/qwen_vl_interface': 4.012766933124523e-06, 'learning_rate/action_model': 4.012766933124523e-05, 'epoch': 0.22}) +[2026-10-03 19:31:39,635][starVLA.training.train_starvla][INFO] - Step 2962, Loss: {'train/loss': 0.00023521771072410047, 'train/grad_norm_pre_clip': 0.1681714951992035, 'timing/data': 0.000202669994905591, 'timing/model': 0.39084323501447216, 'learning_rate/qwen_vl_interface': 4.009826888300176e-06, 'learning_rate/action_model': 4.0098268883001755e-05, 'epoch': 0.22}) +[2026-10-03 19:31:40,329][starVLA.training.train_starvla][INFO] - Step 2963, Loss: {'train/loss': 0.7009850740432739, 'train/grad_norm_pre_clip': 30.77733039855957, 'timing/data': 0.00021147000370547175, 'timing/model': 0.38867350498912856, 'learning_rate/qwen_vl_interface': 4.00688735326415e-06, 'learning_rate/action_model': 4.0068873532641495e-05, 'epoch': 0.22}) +[2026-10-03 19:31:41,028][starVLA.training.train_starvla][INFO] - Step 2964, Loss: {'train/loss': 6.551312981173396e-05, 'train/grad_norm_pre_clip': 0.034180376678705215, 'timing/data': 0.00021341998944990337, 'timing/model': 0.39077887500752695, 'learning_rate/qwen_vl_interface': 4.0039483292247765e-06, 'learning_rate/action_model': 4.003948329224776e-05, 'epoch': 0.22}) +[2026-10-03 19:31:41,727][starVLA.training.train_starvla][INFO] - Step 2965, Loss: {'train/loss': 0.14307257533073425, 'train/grad_norm_pre_clip': 53.52224349975586, 'timing/data': 0.00025509000988677144, 'timing/model': 0.39169487499748357, 'learning_rate/qwen_vl_interface': 4.001009817390178e-06, 'learning_rate/action_model': 4.001009817390178e-05, 'epoch': 0.22}) +[2026-10-03 19:31:42,627][starVLA.training.train_starvla][INFO] - Step 2966, Loss: {'train/loss': 0.00587312038987875, 'train/grad_norm_pre_clip': 4.9666666984558105, 'timing/data': 0.000205660006031394, 'timing/model': 0.3910421950276941, 'learning_rate/qwen_vl_interface': 3.998071818968264e-06, 'learning_rate/action_model': 3.998071818968264e-05, 'epoch': 0.22}) +[2026-10-03 19:31:43,323][starVLA.training.train_starvla][INFO] - Step 2967, Loss: {'train/loss': 0.5833852291107178, 'train/grad_norm_pre_clip': 96.28902435302734, 'timing/data': 0.00021068999194540083, 'timing/model': 0.3887457159871701, 'learning_rate/qwen_vl_interface': 3.995134335166736e-06, 'learning_rate/action_model': 3.995134335166736e-05, 'epoch': 0.22}) +[2026-10-03 19:31:44,019][starVLA.training.train_starvla][INFO] - Step 2968, Loss: {'train/loss': 0.21798011660575867, 'train/grad_norm_pre_clip': 191.38876342773438, 'timing/data': 0.0002101800055243075, 'timing/model': 0.3891228660068009, 'learning_rate/qwen_vl_interface': 3.992197367193083e-06, 'learning_rate/action_model': 3.992197367193082e-05, 'epoch': 0.22}) +[2026-10-03 19:31:44,717][starVLA.training.train_starvla][INFO] - Step 2969, Loss: {'train/loss': 3.882988676195964e-05, 'train/grad_norm_pre_clip': 0.010373642668128014, 'timing/data': 0.00036063999868929386, 'timing/model': 0.3907045849773567, 'learning_rate/qwen_vl_interface': 3.98926091625458e-06, 'learning_rate/action_model': 3.9892609162545795e-05, 'epoch': 0.22}) +[2026-10-03 19:31:45,411][starVLA.training.train_starvla][INFO] - Step 2970, Loss: {'train/loss': 0.604547917842865, 'train/grad_norm_pre_clip': 51.988319396972656, 'timing/data': 0.000241830013692379, 'timing/model': 0.38840960498782806, 'learning_rate/qwen_vl_interface': 3.986324983558294e-06, 'learning_rate/action_model': 3.9863249835582934e-05, 'epoch': 0.22}) +[2026-10-03 19:31:46,315][starVLA.training.train_starvla][INFO] - Step 2971, Loss: {'train/loss': 0.20554442703723907, 'train/grad_norm_pre_clip': 219.51806640625, 'timing/data': 0.00021341998944990337, 'timing/model': 0.39087282499531284, 'learning_rate/qwen_vl_interface': 3.983389570311073e-06, 'learning_rate/action_model': 3.9833895703110724e-05, 'epoch': 0.22}) +[2026-10-03 19:31:47,015][starVLA.training.train_starvla][INFO] - Step 2972, Loss: {'train/loss': 0.010853744111955166, 'train/grad_norm_pre_clip': 18.15928840637207, 'timing/data': 0.0002181899908464402, 'timing/model': 0.39084192499285564, 'learning_rate/qwen_vl_interface': 3.9804546777195546e-06, 'learning_rate/action_model': 3.980454677719554e-05, 'epoch': 0.22}) +[2026-10-03 19:31:47,712][starVLA.training.train_starvla][INFO] - Step 2973, Loss: {'train/loss': 0.00032328363158740103, 'train/grad_norm_pre_clip': 0.20984092354774475, 'timing/data': 0.00020460999803617597, 'timing/model': 0.3908686439972371, 'learning_rate/qwen_vl_interface': 3.977520306990164e-06, 'learning_rate/action_model': 3.977520306990163e-05, 'epoch': 0.22}) +[2026-10-03 19:31:48,411][starVLA.training.train_starvla][INFO] - Step 2974, Loss: {'train/loss': 0.39658698439598083, 'train/grad_norm_pre_clip': 29.218008041381836, 'timing/data': 0.0002515300002414733, 'timing/model': 0.3914331950072665, 'learning_rate/qwen_vl_interface': 3.974586459329109e-06, 'learning_rate/action_model': 3.9745864593291085e-05, 'epoch': 0.22}) +[2026-10-03 19:31:49,111][starVLA.training.train_starvla][INFO] - Step 2975, Loss: {'train/loss': 0.00020504670101217926, 'train/grad_norm_pre_clip': 0.2176317274570465, 'timing/data': 0.00028788999770767987, 'timing/model': 0.3915220150083769, 'learning_rate/qwen_vl_interface': 3.9716531359423855e-06, 'learning_rate/action_model': 3.971653135942385e-05, 'epoch': 0.22}) +[2026-10-03 19:31:50,010][starVLA.training.train_starvla][INFO] - Step 2976, Loss: {'train/loss': 8.308719407068565e-05, 'train/grad_norm_pre_clip': 0.08951034396886826, 'timing/data': 0.0003397699911147356, 'timing/model': 0.5902590179757681, 'learning_rate/qwen_vl_interface': 3.968720338035769e-06, 'learning_rate/action_model': 3.968720338035769e-05, 'epoch': 0.22}) +[2026-10-03 19:31:50,710][starVLA.training.train_starvla][INFO] - Step 2977, Loss: {'train/loss': 0.37103697657585144, 'train/grad_norm_pre_clip': 95.33378601074219, 'timing/data': 0.00032086001010611653, 'timing/model': 0.39298378498642705, 'learning_rate/qwen_vl_interface': 3.965788066814824e-06, 'learning_rate/action_model': 3.965788066814823e-05, 'epoch': 0.22}) +[2026-10-03 19:31:51,408][starVLA.training.train_starvla][INFO] - Step 2978, Loss: {'train/loss': 0.0006535969441756606, 'train/grad_norm_pre_clip': 0.7500406503677368, 'timing/data': 0.00020107999444007874, 'timing/model': 0.3902344749949407, 'learning_rate/qwen_vl_interface': 3.962856323484895e-06, 'learning_rate/action_model': 3.9628563234848945e-05, 'epoch': 0.22}) +[2026-10-03 19:31:52,105][starVLA.training.train_starvla][INFO] - Step 2979, Loss: {'train/loss': 0.00013707509788218886, 'train/grad_norm_pre_clip': 0.07740332931280136, 'timing/data': 0.0002046290028374642, 'timing/model': 0.38950806501088664, 'learning_rate/qwen_vl_interface': 3.959925109251112e-06, 'learning_rate/action_model': 3.959925109251112e-05, 'epoch': 0.22}) +[2026-10-03 19:31:52,806][starVLA.training.train_starvla][INFO] - Step 2980, Loss: {'train/loss': 0.004013577941805124, 'train/grad_norm_pre_clip': 3.5407047271728516, 'timing/data': 0.0003436599799897522, 'timing/model': 0.3917963950079866, 'learning_rate/qwen_vl_interface': 3.956994425318387e-06, 'learning_rate/action_model': 3.956994425318387e-05, 'epoch': 0.22}) +[2026-10-03 19:31:53,727][starVLA.training.train_starvla][INFO] - Step 2981, Loss: {'train/loss': 0.6673429012298584, 'train/grad_norm_pre_clip': 196.8096160888672, 'timing/data': 0.00031115999445319176, 'timing/model': 0.6129770449770149, 'learning_rate/qwen_vl_interface': 3.9540642728914115e-06, 'learning_rate/action_model': 3.9540642728914115e-05, 'epoch': 0.22}) +[2026-10-03 19:31:54,423][starVLA.training.train_starvla][INFO] - Step 2982, Loss: {'train/loss': 0.0003757129597943276, 'train/grad_norm_pre_clip': 0.33711522817611694, 'timing/data': 0.00020491000032052398, 'timing/model': 0.390523754991591, 'learning_rate/qwen_vl_interface': 3.951134653174664e-06, 'learning_rate/action_model': 3.951134653174663e-05, 'epoch': 0.22}) +[2026-10-03 19:31:55,119][starVLA.training.train_starvla][INFO] - Step 2983, Loss: {'train/loss': 0.17745064198970795, 'train/grad_norm_pre_clip': 337.89373779296875, 'timing/data': 0.0002788190031424165, 'timing/model': 0.3908844559919089, 'learning_rate/qwen_vl_interface': 3.948205567372397e-06, 'learning_rate/action_model': 3.948205567372397e-05, 'epoch': 0.22}) +[2026-10-03 19:31:55,829][starVLA.training.train_starvla][INFO] - Step 2984, Loss: {'train/loss': 0.31074419617652893, 'train/grad_norm_pre_clip': 75.32440948486328, 'timing/data': 0.00030143000185489655, 'timing/model': 0.39556090501719154, 'learning_rate/qwen_vl_interface': 3.945277016688647e-06, 'learning_rate/action_model': 3.945277016688647e-05, 'epoch': 0.22}) +[2026-10-03 19:31:56,537][starVLA.training.train_starvla][INFO] - Step 2985, Loss: {'train/loss': 0.00046229796134866774, 'train/grad_norm_pre_clip': 0.10832121968269348, 'timing/data': 0.0003436399856582284, 'timing/model': 0.3961812350025866, 'learning_rate/qwen_vl_interface': 3.9423490023272326e-06, 'learning_rate/action_model': 3.9423490023272324e-05, 'epoch': 0.22}) +[2026-10-03 19:31:57,446][starVLA.training.train_starvla][INFO] - Step 2986, Loss: {'train/loss': 0.0008174008689820766, 'train/grad_norm_pre_clip': 0.4028390645980835, 'timing/data': 0.00024878999101929367, 'timing/model': 0.597806486999616, 'learning_rate/qwen_vl_interface': 3.939421525491748e-06, 'learning_rate/action_model': 3.9394215254917483e-05, 'epoch': 0.22}) +[2026-10-03 19:31:58,152][starVLA.training.train_starvla][INFO] - Step 2987, Loss: {'train/loss': 0.0004851724661421031, 'train/grad_norm_pre_clip': 0.18084251880645752, 'timing/data': 0.0003137199964839965, 'timing/model': 0.3946153040160425, 'learning_rate/qwen_vl_interface': 3.9364945873855705e-06, 'learning_rate/action_model': 3.93649458738557e-05, 'epoch': 0.22}) +[2026-10-03 19:31:58,858][starVLA.training.train_starvla][INFO] - Step 2988, Loss: {'train/loss': 0.00029152771458029747, 'train/grad_norm_pre_clip': 0.1007106825709343, 'timing/data': 0.0002731400018092245, 'timing/model': 0.3953331550001167, 'learning_rate/qwen_vl_interface': 3.9335681892118536e-06, 'learning_rate/action_model': 3.933568189211854e-05, 'epoch': 0.22}) +[2026-10-03 19:31:59,563][starVLA.training.train_starvla][INFO] - Step 2989, Loss: {'train/loss': 0.006478828378021717, 'train/grad_norm_pre_clip': 7.596314430236816, 'timing/data': 0.00031720998231321573, 'timing/model': 0.3946970950055402, 'learning_rate/qwen_vl_interface': 3.930642332173527e-06, 'learning_rate/action_model': 3.930642332173527e-05, 'epoch': 0.22}) +[2026-10-03 19:32:00,269][starVLA.training.train_starvla][INFO] - Step 2990, Loss: {'train/loss': 7.325423939619213e-05, 'train/grad_norm_pre_clip': 0.040396325290203094, 'timing/data': 0.00025705000734888017, 'timing/model': 0.39571709497249685, 'learning_rate/qwen_vl_interface': 3.927717017473301e-06, 'learning_rate/action_model': 3.9277170174733006e-05, 'epoch': 0.22}) +[2026-10-03 19:32:00,977][starVLA.training.train_starvla][INFO] - Step 2991, Loss: {'train/loss': 0.00010406223009340465, 'train/grad_norm_pre_clip': 0.027066193521022797, 'timing/data': 0.0002504000149201602, 'timing/model': 0.39454066500184126, 'learning_rate/qwen_vl_interface': 3.9247922463136615e-06, 'learning_rate/action_model': 3.924792246313662e-05, 'epoch': 0.22}) +[2026-10-03 19:32:01,890][starVLA.training.train_starvla][INFO] - Step 2992, Loss: {'train/loss': 0.011877976357936859, 'train/grad_norm_pre_clip': 18.496091842651367, 'timing/data': 0.0003420000139158219, 'timing/model': 0.39499944500857964, 'learning_rate/qwen_vl_interface': 3.921868019896873e-06, 'learning_rate/action_model': 3.921868019896873e-05, 'epoch': 0.22}) +[2026-10-03 19:32:02,598][starVLA.training.train_starvla][INFO] - Step 2993, Loss: {'train/loss': 0.08056176453828812, 'train/grad_norm_pre_clip': 102.11370849609375, 'timing/data': 0.0003227500128559768, 'timing/model': 0.3971957140020095, 'learning_rate/qwen_vl_interface': 3.918944339424974e-06, 'learning_rate/action_model': 3.9189443394249736e-05, 'epoch': 0.22}) +[2026-10-03 19:32:03,304][starVLA.training.train_starvla][INFO] - Step 2994, Loss: {'train/loss': 7.747807103442028e-05, 'train/grad_norm_pre_clip': 0.025780173018574715, 'timing/data': 0.0004686900065280497, 'timing/model': 0.3939481939887628, 'learning_rate/qwen_vl_interface': 3.9160212060997774e-06, 'learning_rate/action_model': 3.9160212060997774e-05, 'epoch': 0.22}) +[2026-10-03 19:32:04,012][starVLA.training.train_starvla][INFO] - Step 2995, Loss: {'train/loss': 0.5636937618255615, 'train/grad_norm_pre_clip': 56.827091217041016, 'timing/data': 0.00024262000806629658, 'timing/model': 0.3945544850139413, 'learning_rate/qwen_vl_interface': 3.913098621122875e-06, 'learning_rate/action_model': 3.913098621122875e-05, 'epoch': 0.22}) +[2026-10-03 19:32:04,721][starVLA.training.train_starvla][INFO] - Step 2996, Loss: {'train/loss': 0.047527968883514404, 'train/grad_norm_pre_clip': 61.6279182434082, 'timing/data': 0.000305909983580932, 'timing/model': 0.3962549850111827, 'learning_rate/qwen_vl_interface': 3.91017658569563e-06, 'learning_rate/action_model': 3.91017658569563e-05, 'epoch': 0.22}) +[2026-10-03 19:32:05,635][starVLA.training.train_starvla][INFO] - Step 2997, Loss: {'train/loss': 0.00029425875982269645, 'train/grad_norm_pre_clip': 0.5157894492149353, 'timing/data': 0.0003129199903924018, 'timing/model': 0.6005563060170971, 'learning_rate/qwen_vl_interface': 3.907255101019181e-06, 'learning_rate/action_model': 3.907255101019181e-05, 'epoch': 0.22}) +[2026-10-03 19:32:06,343][starVLA.training.train_starvla][INFO] - Step 2998, Loss: {'train/loss': 0.43504849076271057, 'train/grad_norm_pre_clip': 131.78353881835938, 'timing/data': 0.00035759000456891954, 'timing/model': 0.3966185739845969, 'learning_rate/qwen_vl_interface': 3.904334168294442e-06, 'learning_rate/action_model': 3.9043341682944415e-05, 'epoch': 0.22}) +[2026-10-03 19:32:07,049][starVLA.training.train_starvla][INFO] - Step 2999, Loss: {'train/loss': 0.05058540776371956, 'train/grad_norm_pre_clip': 31.41670799255371, 'timing/data': 0.00041491998126730323, 'timing/model': 0.3942187549837399, 'learning_rate/qwen_vl_interface': 3.901413788722093e-06, 'learning_rate/action_model': 3.901413788722093e-05, 'epoch': 0.22}) +[2026-10-03 19:32:07,755][starVLA.training.train_starvla][INFO] - Step 3000, Loss: {'train/loss': 0.025194691494107246, 'train/grad_norm_pre_clip': 66.41486358642578, 'timing/data': 0.00023236998822540045, 'timing/model': 0.39338224500534125, 'learning_rate/qwen_vl_interface': 3.8984939635025966e-06, 'learning_rate/action_model': 3.898493963502596e-05, 'epoch': 0.22}) +[2026-10-03 19:32:08,465][starVLA.training.train_starvla][INFO] - Step 3001, Loss: {'train/loss': 3.898601426044479e-05, 'train/grad_norm_pre_clip': 0.015820764005184174, 'timing/data': 0.00027929997304454446, 'timing/model': 0.39623569397372194, 'learning_rate/qwen_vl_interface': 3.895574693836181e-06, 'learning_rate/action_model': 3.8955746938361804e-05, 'epoch': 0.22}) +[2026-10-03 19:32:09,175][starVLA.training.train_starvla][INFO] - Step 3002, Loss: {'train/loss': 0.00046785513404756784, 'train/grad_norm_pre_clip': 0.22715918719768524, 'timing/data': 0.00042597000719979405, 'timing/model': 0.3956823149928823, 'learning_rate/qwen_vl_interface': 3.892655980922847e-06, 'learning_rate/action_model': 3.8926559809228466e-05, 'epoch': 0.22}) +[2026-10-03 19:32:10,088][starVLA.training.train_starvla][INFO] - Step 3003, Loss: {'train/loss': 1.420722246170044, 'train/grad_norm_pre_clip': 183.08230590820312, 'timing/data': 0.0002527599863242358, 'timing/model': 0.3955081349995453, 'learning_rate/qwen_vl_interface': 3.889737825962368e-06, 'learning_rate/action_model': 3.889737825962368e-05, 'epoch': 0.22}) +[2026-10-03 19:32:10,795][starVLA.training.train_starvla][INFO] - Step 3004, Loss: {'train/loss': 0.2634142339229584, 'train/grad_norm_pre_clip': 33.89347839355469, 'timing/data': 0.00041164999129250646, 'timing/model': 0.3961639449989889, 'learning_rate/qwen_vl_interface': 3.886820230154288e-06, 'learning_rate/action_model': 3.886820230154287e-05, 'epoch': 0.22}) +[2026-10-03 19:32:11,505][starVLA.training.train_starvla][INFO] - Step 3005, Loss: {'train/loss': 0.31444215774536133, 'train/grad_norm_pre_clip': 31.311357498168945, 'timing/data': 0.0003558400203473866, 'timing/model': 0.39492262501153164, 'learning_rate/qwen_vl_interface': 3.8839031946979184e-06, 'learning_rate/action_model': 3.883903194697918e-05, 'epoch': 0.22}) +[2026-10-03 19:32:12,212][starVLA.training.train_starvla][INFO] - Step 3006, Loss: {'train/loss': 0.5324532985687256, 'train/grad_norm_pre_clip': 159.4637451171875, 'timing/data': 0.00028150001890026033, 'timing/model': 0.3943568749818951, 'learning_rate/qwen_vl_interface': 3.880986720792344e-06, 'learning_rate/action_model': 3.880986720792344e-05, 'epoch': 0.22}) +[2026-10-03 19:32:12,919][starVLA.training.train_starvla][INFO] - Step 3007, Loss: {'train/loss': 0.07195619493722916, 'train/grad_norm_pre_clip': 91.89990234375, 'timing/data': 0.0002753300068434328, 'timing/model': 0.3948483349813614, 'learning_rate/qwen_vl_interface': 3.878070809636416e-06, 'learning_rate/action_model': 3.878070809636416e-05, 'epoch': 0.22}) +[2026-10-03 19:32:13,836][starVLA.training.train_starvla][INFO] - Step 3008, Loss: {'train/loss': 0.0010230621555820107, 'train/grad_norm_pre_clip': 2.191521167755127, 'timing/data': 0.0003822699945885688, 'timing/model': 0.396121044002939, 'learning_rate/qwen_vl_interface': 3.875155462428758e-06, 'learning_rate/action_model': 3.875155462428758e-05, 'epoch': 0.22}) +[2026-10-03 19:32:14,521][starVLA.training.train_starvla][INFO] - Step 3009, Loss: {'train/loss': 0.0002904291613958776, 'train/grad_norm_pre_clip': 0.11940518021583557, 'timing/data': 0.0002811900048982352, 'timing/model': 0.3968351839866955, 'learning_rate/qwen_vl_interface': 3.872240680367754e-06, 'learning_rate/action_model': 3.872240680367754e-05, 'epoch': 0.22}) +[2026-10-03 19:32:15,226][starVLA.training.train_starvla][INFO] - Step 3010, Loss: {'train/loss': 0.5101096034049988, 'train/grad_norm_pre_clip': 217.4972381591797, 'timing/data': 0.0003702899848576635, 'timing/model': 0.39349720400059596, 'learning_rate/qwen_vl_interface': 3.869326464651566e-06, 'learning_rate/action_model': 3.869326464651566e-05, 'epoch': 0.22}) +[2026-10-03 19:32:15,931][starVLA.training.train_starvla][INFO] - Step 3011, Loss: {'train/loss': 0.3391464650630951, 'train/grad_norm_pre_clip': 95.79353332519531, 'timing/data': 0.00021694001043215394, 'timing/model': 0.3937646239937749, 'learning_rate/qwen_vl_interface': 3.866412816478114e-06, 'learning_rate/action_model': 3.8664128164781144e-05, 'epoch': 0.22}) +[2026-10-03 19:32:16,638][starVLA.training.train_starvla][INFO] - Step 3012, Loss: {'train/loss': 0.03233621269464493, 'train/grad_norm_pre_clip': 45.03102111816406, 'timing/data': 0.00038155997754074633, 'timing/model': 0.3947870749980211, 'learning_rate/qwen_vl_interface': 3.8634997370450915e-06, 'learning_rate/action_model': 3.8634997370450914e-05, 'epoch': 0.22}) +[2026-10-03 19:32:17,544][starVLA.training.train_starvla][INFO] - Step 3013, Loss: {'train/loss': 6.488606595667079e-05, 'train/grad_norm_pre_clip': 0.017116446048021317, 'timing/data': 0.0002629599766805768, 'timing/model': 0.5937871770001948, 'learning_rate/qwen_vl_interface': 3.860587227549956e-06, 'learning_rate/action_model': 3.860587227549955e-05, 'epoch': 0.22}) +[2026-10-03 19:32:18,249][starVLA.training.train_starvla][INFO] - Step 3014, Loss: {'train/loss': 0.43434712290763855, 'train/grad_norm_pre_clip': 67.95645904541016, 'timing/data': 0.0003581400087568909, 'timing/model': 0.39311531401472166, 'learning_rate/qwen_vl_interface': 3.857675289189926e-06, 'learning_rate/action_model': 3.857675289189926e-05, 'epoch': 0.22}) +[2026-10-03 19:32:18,956][starVLA.training.train_starvla][INFO] - Step 3015, Loss: {'train/loss': 0.15337377786636353, 'train/grad_norm_pre_clip': 94.71246337890625, 'timing/data': 0.0003141699999105185, 'timing/model': 0.3953068550035823, 'learning_rate/qwen_vl_interface': 3.854763923161993e-06, 'learning_rate/action_model': 3.854763923161992e-05, 'epoch': 0.22}) +[2026-10-03 19:32:19,665][starVLA.training.train_starvla][INFO] - Step 3016, Loss: {'train/loss': 2.3341824999079108e-05, 'train/grad_norm_pre_clip': 0.004962105769664049, 'timing/data': 0.0002345200045965612, 'timing/model': 0.3961765150015708, 'learning_rate/qwen_vl_interface': 3.851853130662908e-06, 'learning_rate/action_model': 3.8518531306629077e-05, 'epoch': 0.22}) +[2026-10-03 19:32:20,373][starVLA.training.train_starvla][INFO] - Step 3017, Loss: {'train/loss': 0.03032035380601883, 'train/grad_norm_pre_clip': 45.3255729675293, 'timing/data': 0.000298850005492568, 'timing/model': 0.3963972149940673, 'learning_rate/qwen_vl_interface': 3.8489429128891884e-06, 'learning_rate/action_model': 3.848942912889188e-05, 'epoch': 0.22}) +[2026-10-03 19:32:21,080][starVLA.training.train_starvla][INFO] - Step 3018, Loss: {'train/loss': 0.001704551512375474, 'train/grad_norm_pre_clip': 0.8203654885292053, 'timing/data': 0.00023316001170314848, 'timing/model': 0.3953818139852956, 'learning_rate/qwen_vl_interface': 3.846033271037114e-06, 'learning_rate/action_model': 3.8460332710371136e-05, 'epoch': 0.22}) +[2026-10-03 19:32:21,989][starVLA.training.train_starvla][INFO] - Step 3019, Loss: {'train/loss': 0.0002627870417200029, 'train/grad_norm_pre_clip': 0.1757795512676239, 'timing/data': 0.0002371199952904135, 'timing/model': 0.39477717500994913, 'learning_rate/qwen_vl_interface': 3.843124206302732e-06, 'learning_rate/action_model': 3.8431242063027316e-05, 'epoch': 0.22}) +[2026-10-03 19:32:22,697][starVLA.training.train_starvla][INFO] - Step 3020, Loss: {'train/loss': 0.0006862084846943617, 'train/grad_norm_pre_clip': 1.4645817279815674, 'timing/data': 0.00033098002313636243, 'timing/model': 0.39629044401226565, 'learning_rate/qwen_vl_interface': 3.840215719881844e-06, 'learning_rate/action_model': 3.8402157198818445e-05, 'epoch': 0.22}) +[2026-10-03 19:32:23,408][starVLA.training.train_starvla][INFO] - Step 3021, Loss: {'train/loss': 0.1718342900276184, 'train/grad_norm_pre_clip': 79.01470184326172, 'timing/data': 0.00037707999581471086, 'timing/model': 0.3973614939895924, 'learning_rate/qwen_vl_interface': 3.837307812970022e-06, 'learning_rate/action_model': 3.837307812970022e-05, 'epoch': 0.22}) +[2026-10-03 19:32:24,115][starVLA.training.train_starvla][INFO] - Step 3022, Loss: {'train/loss': 0.004631672985851765, 'train/grad_norm_pre_clip': 12.577481269836426, 'timing/data': 0.00043999997433274984, 'timing/model': 0.39431775399134494, 'learning_rate/qwen_vl_interface': 3.834400486762596e-06, 'learning_rate/action_model': 3.834400486762596e-05, 'epoch': 0.22}) +[2026-10-03 19:32:24,823][starVLA.training.train_starvla][INFO] - Step 3023, Loss: {'train/loss': 4.4481090299086645e-05, 'train/grad_norm_pre_clip': 0.015060150995850563, 'timing/data': 0.00034915999276563525, 'timing/model': 0.39630524400854483, 'learning_rate/qwen_vl_interface': 3.8314937424546585e-06, 'learning_rate/action_model': 3.8314937424546584e-05, 'epoch': 0.22}) +[2026-10-03 19:32:25,734][starVLA.training.train_starvla][INFO] - Step 3024, Loss: {'train/loss': 0.002844247967004776, 'train/grad_norm_pre_clip': 0.6781934499740601, 'timing/data': 0.0003560599870979786, 'timing/model': 0.5971582670172211, 'learning_rate/qwen_vl_interface': 3.828587581241064e-06, 'learning_rate/action_model': 3.8285875812410635e-05, 'epoch': 0.22}) +[2026-10-03 19:32:26,445][starVLA.training.train_starvla][INFO] - Step 3025, Loss: {'train/loss': 0.00038164760917425156, 'train/grad_norm_pre_clip': 0.7422235608100891, 'timing/data': 0.0003326500009279698, 'timing/model': 0.3977145650133025, 'learning_rate/qwen_vl_interface': 3.825682004316419e-06, 'learning_rate/action_model': 3.825682004316419e-05, 'epoch': 0.22}) +[2026-10-03 19:32:27,153][starVLA.training.train_starvla][INFO] - Step 3026, Loss: {'train/loss': 0.006145148072391748, 'train/grad_norm_pre_clip': 5.705296516418457, 'timing/data': 0.0004143299884162843, 'timing/model': 0.39579073400818743, 'learning_rate/qwen_vl_interface': 3.822777012875104e-06, 'learning_rate/action_model': 3.822777012875103e-05, 'epoch': 0.22}) +[2026-10-03 19:32:27,858][starVLA.training.train_starvla][INFO] - Step 3027, Loss: {'train/loss': 0.056571077555418015, 'train/grad_norm_pre_clip': 72.72978210449219, 'timing/data': 0.00021145999198779464, 'timing/model': 0.3942926650051959, 'learning_rate/qwen_vl_interface': 3.819872608111248e-06, 'learning_rate/action_model': 3.819872608111248e-05, 'epoch': 0.22}) +[2026-10-03 19:32:28,568][starVLA.training.train_starvla][INFO] - Step 3028, Loss: {'train/loss': 0.4955928921699524, 'train/grad_norm_pre_clip': 62.621646881103516, 'timing/data': 0.00035898000351153314, 'timing/model': 0.39652882399968803, 'learning_rate/qwen_vl_interface': 3.816968791218741e-06, 'learning_rate/action_model': 3.816968791218741e-05, 'epoch': 0.22}) +[2026-10-03 19:32:29,452][starVLA.training.train_starvla][INFO] - Step 3029, Loss: {'train/loss': 0.18956540524959564, 'train/grad_norm_pre_clip': 123.6484146118164, 'timing/data': 0.0003378799883648753, 'timing/model': 0.5975910759880207, 'learning_rate/qwen_vl_interface': 3.8140655633912337e-06, 'learning_rate/action_model': 3.814065563391234e-05, 'epoch': 0.22}) +[2026-10-03 19:32:30,162][starVLA.training.train_starvla][INFO] - Step 3030, Loss: {'train/loss': 2.5756064133020118e-05, 'train/grad_norm_pre_clip': 0.0058281272649765015, 'timing/data': 0.00045305999810807407, 'timing/model': 0.3958483840106055, 'learning_rate/qwen_vl_interface': 3.811162925822133e-06, 'learning_rate/action_model': 3.811162925822132e-05, 'epoch': 0.22}) +[2026-10-03 19:32:30,869][starVLA.training.train_starvla][INFO] - Step 3031, Loss: {'train/loss': 0.6273947954177856, 'train/grad_norm_pre_clip': 58.111175537109375, 'timing/data': 0.0003422000154387206, 'timing/model': 0.39478418498765677, 'learning_rate/qwen_vl_interface': 3.8082608797046045e-06, 'learning_rate/action_model': 3.8082608797046045e-05, 'epoch': 0.22}) +[2026-10-03 19:32:31,578][starVLA.training.train_starvla][INFO] - Step 3032, Loss: {'train/loss': 0.017119836062192917, 'train/grad_norm_pre_clip': 30.98973846435547, 'timing/data': 0.0003650000144261867, 'timing/model': 0.39423378399806097, 'learning_rate/qwen_vl_interface': 3.8053594262315687e-06, 'learning_rate/action_model': 3.805359426231568e-05, 'epoch': 0.22}) +[2026-10-03 19:32:32,285][starVLA.training.train_starvla][INFO] - Step 3033, Loss: {'train/loss': 0.05093259736895561, 'train/grad_norm_pre_clip': 65.14753723144531, 'timing/data': 0.000328630005242303, 'timing/model': 0.39603140400140546, 'learning_rate/qwen_vl_interface': 3.802458566595704e-06, 'learning_rate/action_model': 3.802458566595704e-05, 'epoch': 0.22}) +[2026-10-03 19:32:32,994][starVLA.training.train_starvla][INFO] - Step 3034, Loss: {'train/loss': 0.000299967301543802, 'train/grad_norm_pre_clip': 0.15451288223266602, 'timing/data': 0.0003476999991107732, 'timing/model': 0.39522328501334414, 'learning_rate/qwen_vl_interface': 3.7995583019894444e-06, 'learning_rate/action_model': 3.7995583019894445e-05, 'epoch': 0.22}) +[2026-10-03 19:32:33,913][starVLA.training.train_starvla][INFO] - Step 3035, Loss: {'train/loss': 0.00010893339640460908, 'train/grad_norm_pre_clip': 0.03983436897397041, 'timing/data': 0.00040427001658827066, 'timing/model': 0.39539541499107145, 'learning_rate/qwen_vl_interface': 3.7966586336049782e-06, 'learning_rate/action_model': 3.796658633604978e-05, 'epoch': 0.22}) +[2026-10-03 19:32:34,622][starVLA.training.train_starvla][INFO] - Step 3036, Loss: {'train/loss': 8.265409269370139e-05, 'train/grad_norm_pre_clip': 0.028424551710486412, 'timing/data': 0.0004907900001853704, 'timing/model': 0.39556832500966266, 'learning_rate/qwen_vl_interface': 3.7937595626342516e-06, 'learning_rate/action_model': 3.793759562634252e-05, 'epoch': 0.22}) +[2026-10-03 19:32:35,330][starVLA.training.train_starvla][INFO] - Step 3037, Loss: {'train/loss': 0.0008136951364576817, 'train/grad_norm_pre_clip': 0.43680340051651, 'timing/data': 0.0002850299933925271, 'timing/model': 0.3957420640217606, 'learning_rate/qwen_vl_interface': 3.79086109026896e-06, 'learning_rate/action_model': 3.7908610902689604e-05, 'epoch': 0.22}) +[2026-10-03 19:32:36,038][starVLA.training.train_starvla][INFO] - Step 3038, Loss: {'train/loss': 0.0014731755945831537, 'train/grad_norm_pre_clip': 1.6340831518173218, 'timing/data': 0.0004054900200571865, 'timing/model': 0.39562852401286364, 'learning_rate/qwen_vl_interface': 3.7879632177005584e-06, 'learning_rate/action_model': 3.787963217700558e-05, 'epoch': 0.22}) +[2026-10-03 19:32:36,746][starVLA.training.train_starvla][INFO] - Step 3039, Loss: {'train/loss': 0.22385264933109283, 'train/grad_norm_pre_clip': 202.906982421875, 'timing/data': 0.00032098000519908965, 'timing/model': 0.39434023399371654, 'learning_rate/qwen_vl_interface': 3.7850659461202523e-06, 'learning_rate/action_model': 3.785065946120252e-05, 'epoch': 0.22}) +[2026-10-03 19:32:37,650][starVLA.training.train_starvla][INFO] - Step 3040, Loss: {'train/loss': 0.00046986923553049564, 'train/grad_norm_pre_clip': 0.13336853682994843, 'timing/data': 0.0003641500079538673, 'timing/model': 0.5932868559903, 'learning_rate/qwen_vl_interface': 3.782169276719e-06, 'learning_rate/action_model': 3.782169276719e-05, 'epoch': 0.22}) +[2026-10-03 19:32:38,358][starVLA.training.train_starvla][INFO] - Step 3041, Loss: {'train/loss': 0.00025245401775464416, 'train/grad_norm_pre_clip': 0.1051352322101593, 'timing/data': 0.0003136000013910234, 'timing/model': 0.39589267401606776, 'learning_rate/qwen_vl_interface': 3.779273210687514e-06, 'learning_rate/action_model': 3.779273210687514e-05, 'epoch': 0.22}) +[2026-10-03 19:32:39,067][starVLA.training.train_starvla][INFO] - Step 3042, Loss: {'train/loss': 0.00010444979125168175, 'train/grad_norm_pre_clip': 0.017725229263305664, 'timing/data': 0.00023442998644895852, 'timing/model': 0.39613296499010175, 'learning_rate/qwen_vl_interface': 3.776377749216257e-06, 'learning_rate/action_model': 3.776377749216257e-05, 'epoch': 0.22}) +[2026-10-03 19:32:39,772][starVLA.training.train_starvla][INFO] - Step 3043, Loss: {'train/loss': 0.00043062836630269885, 'train/grad_norm_pre_clip': 0.09854480624198914, 'timing/data': 0.000318949983920902, 'timing/model': 0.39378288498846814, 'learning_rate/qwen_vl_interface': 3.773482893495442e-06, 'learning_rate/action_model': 3.773482893495442e-05, 'epoch': 0.22}) +[2026-10-03 19:32:40,477][starVLA.training.train_starvla][INFO] - Step 3044, Loss: {'train/loss': 9.937900904333219e-05, 'train/grad_norm_pre_clip': 0.040301013737916946, 'timing/data': 0.0002621500170789659, 'timing/model': 0.39401146498858, 'learning_rate/qwen_vl_interface': 3.7705886447150375e-06, 'learning_rate/action_model': 3.7705886447150376e-05, 'epoch': 0.22}) +[2026-10-03 19:32:41,385][starVLA.training.train_starvla][INFO] - Step 3045, Loss: {'train/loss': 0.11171630024909973, 'train/grad_norm_pre_clip': 113.82989501953125, 'timing/data': 0.000373838993255049, 'timing/model': 0.5963579170056619, 'learning_rate/qwen_vl_interface': 3.767695004064759e-06, 'learning_rate/action_model': 3.767695004064759e-05, 'epoch': 0.22}) +[2026-10-03 19:32:42,092][starVLA.training.train_starvla][INFO] - Step 3046, Loss: {'train/loss': 0.00023707453510724008, 'train/grad_norm_pre_clip': 0.12906387448310852, 'timing/data': 0.00022541999351233244, 'timing/model': 0.394849753996823, 'learning_rate/qwen_vl_interface': 3.7648019727340725e-06, 'learning_rate/action_model': 3.764801972734072e-05, 'epoch': 0.22}) +[2026-10-03 19:32:42,797][starVLA.training.train_starvla][INFO] - Step 3047, Loss: {'train/loss': 0.0030524658504873514, 'train/grad_norm_pre_clip': 8.862992286682129, 'timing/data': 0.0003415099927224219, 'timing/model': 0.3932220540009439, 'learning_rate/qwen_vl_interface': 3.761909551912196e-06, 'learning_rate/action_model': 3.761909551912196e-05, 'epoch': 0.22}) +[2026-10-03 19:32:43,505][starVLA.training.train_starvla][INFO] - Step 3048, Loss: {'train/loss': 0.00012949583469890058, 'train/grad_norm_pre_clip': 0.06147965416312218, 'timing/data': 0.00033372000325471163, 'timing/model': 0.39542064399574883, 'learning_rate/qwen_vl_interface': 3.75901774278809e-06, 'learning_rate/action_model': 3.7590177427880904e-05, 'epoch': 0.22}) +[2026-10-03 19:32:44,210][starVLA.training.train_starvla][INFO] - Step 3049, Loss: {'train/loss': 3.508943336782977e-05, 'train/grad_norm_pre_clip': 0.01452925056219101, 'timing/data': 0.00033060001442208886, 'timing/model': 0.3942309039994143, 'learning_rate/qwen_vl_interface': 3.756126546550472e-06, 'learning_rate/action_model': 3.756126546550472e-05, 'epoch': 0.22}) +[2026-10-03 19:32:44,914][starVLA.training.train_starvla][INFO] - Step 3050, Loss: {'train/loss': 0.0010700422571972013, 'train/grad_norm_pre_clip': 0.8211019039154053, 'timing/data': 0.00034338998375460505, 'timing/model': 0.39408789499429986, 'learning_rate/qwen_vl_interface': 3.753235964387802e-06, 'learning_rate/action_model': 3.753235964387802e-05, 'epoch': 0.22}) +[2026-10-03 19:32:45,816][starVLA.training.train_starvla][INFO] - Step 3051, Loss: {'train/loss': 0.36519792675971985, 'train/grad_norm_pre_clip': 39.85564422607422, 'timing/data': 0.0002565199974924326, 'timing/model': 0.39554484499967657, 'learning_rate/qwen_vl_interface': 3.750345997488289e-06, 'learning_rate/action_model': 3.7503459974882886e-05, 'epoch': 0.22}) +[2026-10-03 19:32:46,523][starVLA.training.train_starvla][INFO] - Step 3052, Loss: {'train/loss': 4.726936458609998e-05, 'train/grad_norm_pre_clip': 0.022360414266586304, 'timing/data': 0.00022515899036079645, 'timing/model': 0.3954790350107942, 'learning_rate/qwen_vl_interface': 3.7474566470398903e-06, 'learning_rate/action_model': 3.74745664703989e-05, 'epoch': 0.22}) +[2026-10-03 19:32:47,229][starVLA.training.train_starvla][INFO] - Step 3053, Loss: {'train/loss': 0.3466154634952545, 'train/grad_norm_pre_clip': 162.32655334472656, 'timing/data': 0.000324928987538442, 'timing/model': 0.3945077550015412, 'learning_rate/qwen_vl_interface': 3.744567914230307e-06, 'learning_rate/action_model': 3.744567914230307e-05, 'epoch': 0.22}) +[2026-10-03 19:32:47,936][starVLA.training.train_starvla][INFO] - Step 3054, Loss: {'train/loss': 0.38583600521087646, 'train/grad_norm_pre_clip': 48.735477447509766, 'timing/data': 0.000222400005441159, 'timing/model': 0.3951389139983803, 'learning_rate/qwen_vl_interface': 3.7416798002469902e-06, 'learning_rate/action_model': 3.74167980024699e-05, 'epoch': 0.22}) +[2026-10-03 19:32:48,641][starVLA.training.train_starvla][INFO] - Step 3055, Loss: {'train/loss': 0.0001488176640123129, 'train/grad_norm_pre_clip': 0.0543009415268898, 'timing/data': 0.0002149399952031672, 'timing/model': 0.395093375002034, 'learning_rate/qwen_vl_interface': 3.7387923062771313e-06, 'learning_rate/action_model': 3.738792306277131e-05, 'epoch': 0.22}) +[2026-10-03 19:32:49,555][starVLA.training.train_starvla][INFO] - Step 3056, Loss: {'train/loss': 0.00011139699927298352, 'train/grad_norm_pre_clip': 0.03465180844068527, 'timing/data': 0.00036576000275090337, 'timing/model': 0.3948201239982154, 'learning_rate/qwen_vl_interface': 3.7359054335076732e-06, 'learning_rate/action_model': 3.7359054335076734e-05, 'epoch': 0.22}) +[2026-10-03 19:32:50,262][starVLA.training.train_starvla][INFO] - Step 3057, Loss: {'train/loss': 0.0011367190163582563, 'train/grad_norm_pre_clip': 1.2619881629943848, 'timing/data': 0.0002659900055732578, 'timing/model': 0.3952254939940758, 'learning_rate/qwen_vl_interface': 3.7330191831253006e-06, 'learning_rate/action_model': 3.7330191831253e-05, 'epoch': 0.22}) +[2026-10-03 19:32:50,972][starVLA.training.train_starvla][INFO] - Step 3058, Loss: {'train/loss': 0.22652725875377655, 'train/grad_norm_pre_clip': 192.80992126464844, 'timing/data': 0.0002340999781154096, 'timing/model': 0.3959977649792563, 'learning_rate/qwen_vl_interface': 3.7301335563164382e-06, 'learning_rate/action_model': 3.730133556316438e-05, 'epoch': 0.22}) +[2026-10-03 19:32:51,677][starVLA.training.train_starvla][INFO] - Step 3059, Loss: {'train/loss': 0.3401132822036743, 'train/grad_norm_pre_clip': 70.05303955078125, 'timing/data': 0.00031952999415807426, 'timing/model': 0.39334022399270907, 'learning_rate/qwen_vl_interface': 3.727248554267261e-06, 'learning_rate/action_model': 3.727248554267261e-05, 'epoch': 0.22}) +[2026-10-03 19:32:52,384][starVLA.training.train_starvla][INFO] - Step 3060, Loss: {'train/loss': 0.00015271097072400153, 'train/grad_norm_pre_clip': 0.036233656108379364, 'timing/data': 0.00038636999670416117, 'timing/model': 0.3954617739946116, 'learning_rate/qwen_vl_interface': 3.7243641781636838e-06, 'learning_rate/action_model': 3.7243641781636836e-05, 'epoch': 0.22}) +[2026-10-03 19:32:53,293][starVLA.training.train_starvla][INFO] - Step 3061, Loss: {'train/loss': 0.00014490851026494056, 'train/grad_norm_pre_clip': 0.06950525939464569, 'timing/data': 0.00027446000603958964, 'timing/model': 0.5958857460063882, 'learning_rate/qwen_vl_interface': 3.721480429191364e-06, 'learning_rate/action_model': 3.721480429191364e-05, 'epoch': 0.22}) +[2026-10-03 19:32:53,998][starVLA.training.train_starvla][INFO] - Step 3062, Loss: {'train/loss': 0.22410163283348083, 'train/grad_norm_pre_clip': 80.824951171875, 'timing/data': 0.00036638000165112317, 'timing/model': 0.3950018049799837, 'learning_rate/qwen_vl_interface': 3.7185973085357045e-06, 'learning_rate/action_model': 3.718597308535705e-05, 'epoch': 0.22}) +[2026-10-03 19:32:54,704][starVLA.training.train_starvla][INFO] - Step 3063, Loss: {'train/loss': 0.23524631559848785, 'train/grad_norm_pre_clip': 129.20274353027344, 'timing/data': 0.00022358002024702728, 'timing/model': 0.394872384989867, 'learning_rate/qwen_vl_interface': 3.7157148173818443e-06, 'learning_rate/action_model': 3.715714817381844e-05, 'epoch': 0.22}) +[2026-10-03 19:32:55,413][starVLA.training.train_starvla][INFO] - Step 3064, Loss: {'train/loss': 0.0017349000554531813, 'train/grad_norm_pre_clip': 2.179168224334717, 'timing/data': 0.0002278100000694394, 'timing/model': 0.39509657499729656, 'learning_rate/qwen_vl_interface': 3.7128329569146672e-06, 'learning_rate/action_model': 3.7128329569146666e-05, 'epoch': 0.22}) +[2026-10-03 19:32:56,119][starVLA.training.train_starvla][INFO] - Step 3065, Loss: {'train/loss': 0.00029830477433279157, 'train/grad_norm_pre_clip': 0.07201338559389114, 'timing/data': 0.00026421999791637063, 'timing/model': 0.39432974500232376, 'learning_rate/qwen_vl_interface': 3.7099517283187977e-06, 'learning_rate/action_model': 3.7099517283187976e-05, 'epoch': 0.22}) +[2026-10-03 19:32:56,825][starVLA.training.train_starvla][INFO] - Step 3066, Loss: {'train/loss': 0.000339826918207109, 'train/grad_norm_pre_clip': 0.13399772346019745, 'timing/data': 0.0003115599974989891, 'timing/model': 0.39469329398707487, 'learning_rate/qwen_vl_interface': 3.7070711327786e-06, 'learning_rate/action_model': 3.7070711327785997e-05, 'epoch': 0.22}) +[2026-10-03 19:32:57,721][starVLA.training.train_starvla][INFO] - Step 3067, Loss: {'train/loss': 0.001615422428585589, 'train/grad_norm_pre_clip': 1.9035266637802124, 'timing/data': 0.0002300000051036477, 'timing/model': 0.3960237340070307, 'learning_rate/qwen_vl_interface': 3.704191171478179e-06, 'learning_rate/action_model': 3.704191171478179e-05, 'epoch': 0.22}) +[2026-10-03 19:32:58,428][starVLA.training.train_starvla][INFO] - Step 3068, Loss: {'train/loss': 0.09975320845842361, 'train/grad_norm_pre_clip': 97.23837280273438, 'timing/data': 0.0002197599969804287, 'timing/model': 0.3962026239896659, 'learning_rate/qwen_vl_interface': 3.701311845601376e-06, 'learning_rate/action_model': 3.701311845601376e-05, 'epoch': 0.22}) +[2026-10-03 19:32:59,137][starVLA.training.train_starvla][INFO] - Step 3069, Loss: {'train/loss': 0.003876976203173399, 'train/grad_norm_pre_clip': 5.019524097442627, 'timing/data': 0.00040476000867784023, 'timing/model': 0.3962620839884039, 'learning_rate/qwen_vl_interface': 3.6984331563317744e-06, 'learning_rate/action_model': 3.698433156331774e-05, 'epoch': 0.22}) +[2026-10-03 19:32:59,845][starVLA.training.train_starvla][INFO] - Step 3070, Loss: {'train/loss': 0.000598590646404773, 'train/grad_norm_pre_clip': 1.9115989208221436, 'timing/data': 0.00026030998560599983, 'timing/model': 0.3961192040005699, 'learning_rate/qwen_vl_interface': 3.695555104852695e-06, 'learning_rate/action_model': 3.6955551048526946e-05, 'epoch': 0.22}) +[2026-10-03 19:33:00,552][starVLA.training.train_starvla][INFO] - Step 3071, Loss: {'train/loss': 0.0006312591722235084, 'train/grad_norm_pre_clip': 0.6368982791900635, 'timing/data': 0.00031783999293111265, 'timing/model': 0.3949714039918035, 'learning_rate/qwen_vl_interface': 3.6926776923471948e-06, 'learning_rate/action_model': 3.6926776923471946e-05, 'epoch': 0.22}) +[2026-10-03 19:33:01,451][starVLA.training.train_starvla][INFO] - Step 3072, Loss: {'train/loss': 9.498938743490726e-05, 'train/grad_norm_pre_clip': 0.24378833174705505, 'timing/data': 0.00026861001970246434, 'timing/model': 0.3937615239992738, 'learning_rate/qwen_vl_interface': 3.6898009199980715e-06, 'learning_rate/action_model': 3.6898009199980714e-05, 'epoch': 0.22}) +[2026-10-03 19:33:02,158][starVLA.training.train_starvla][INFO] - Step 3073, Loss: {'train/loss': 0.3251788318157196, 'train/grad_norm_pre_clip': 228.8269805908203, 'timing/data': 0.000245939998421818, 'timing/model': 0.3946423839952331, 'learning_rate/qwen_vl_interface': 3.686924788987855e-06, 'learning_rate/action_model': 3.686924788987855e-05, 'epoch': 0.22}) +[2026-10-03 19:33:02,865][starVLA.training.train_starvla][INFO] - Step 3074, Loss: {'train/loss': 0.062466852366924286, 'train/grad_norm_pre_clip': 213.3473663330078, 'timing/data': 0.00028465999639593065, 'timing/model': 0.39463873399654403, 'learning_rate/qwen_vl_interface': 3.6840493004988155e-06, 'learning_rate/action_model': 3.684049300498815e-05, 'epoch': 0.22}) +[2026-10-03 19:33:03,573][starVLA.training.train_starvla][INFO] - Step 3075, Loss: {'train/loss': 1.626420271350071e-05, 'train/grad_norm_pre_clip': 0.004319703672081232, 'timing/data': 0.00029349999385885894, 'timing/model': 0.39580336402286775, 'learning_rate/qwen_vl_interface': 3.681174455712958e-06, 'learning_rate/action_model': 3.681174455712958e-05, 'epoch': 0.23}) +[2026-10-03 19:33:04,280][starVLA.training.train_starvla][INFO] - Step 3076, Loss: {'train/loss': 0.0012394553050398827, 'train/grad_norm_pre_clip': 0.5361152291297913, 'timing/data': 0.00029336000443436205, 'timing/model': 0.3958170039986726, 'learning_rate/qwen_vl_interface': 3.6783002558120206e-06, 'learning_rate/action_model': 3.67830025581202e-05, 'epoch': 0.23}) +[2026-10-03 19:33:05,187][starVLA.training.train_starvla][INFO] - Step 3077, Loss: {'train/loss': 0.00014229025691747665, 'train/grad_norm_pre_clip': 0.027603020891547203, 'timing/data': 0.00030203000642359257, 'timing/model': 0.5942540459800512, 'learning_rate/qwen_vl_interface': 3.6754267019774804e-06, 'learning_rate/action_model': 3.67542670197748e-05, 'epoch': 0.23}) +[2026-10-03 19:33:05,892][starVLA.training.train_starvla][INFO] - Step 3078, Loss: {'train/loss': 2.780957947834395e-05, 'train/grad_norm_pre_clip': 0.024554843083024025, 'timing/data': 0.0003582299978006631, 'timing/model': 0.395168104005279, 'learning_rate/qwen_vl_interface': 3.6725537953905453e-06, 'learning_rate/action_model': 3.672553795390545e-05, 'epoch': 0.23}) +[2026-10-03 19:33:06,600][starVLA.training.train_starvla][INFO] - Step 3079, Loss: {'train/loss': 0.04216604679822922, 'train/grad_norm_pre_clip': 46.68015670776367, 'timing/data': 0.00020473901531659067, 'timing/model': 0.39452146500116214, 'learning_rate/qwen_vl_interface': 3.6696815372321603e-06, 'learning_rate/action_model': 3.6696815372321604e-05, 'epoch': 0.23}) +[2026-10-03 19:33:07,309][starVLA.training.train_starvla][INFO] - Step 3080, Loss: {'train/loss': 0.03540334850549698, 'train/grad_norm_pre_clip': 93.25287628173828, 'timing/data': 0.00039684001239947975, 'timing/model': 0.39550040499307215, 'learning_rate/qwen_vl_interface': 3.6668099286830023e-06, 'learning_rate/action_model': 3.666809928683002e-05, 'epoch': 0.23}) +[2026-10-03 19:33:08,018][starVLA.training.train_starvla][INFO] - Step 3081, Loss: {'train/loss': 7.398360594379483e-06, 'train/grad_norm_pre_clip': 0.0014525760198011994, 'timing/data': 0.0002266300143674016, 'timing/model': 0.3964088349894155, 'learning_rate/qwen_vl_interface': 3.66393897092348e-06, 'learning_rate/action_model': 3.66393897092348e-05, 'epoch': 0.23}) +[2026-10-03 19:33:08,918][starVLA.training.train_starvla][INFO] - Step 3082, Loss: {'train/loss': 6.393232615664601e-05, 'train/grad_norm_pre_clip': 0.14717413485050201, 'timing/data': 0.0003429299977142364, 'timing/model': 0.5897578269941732, 'learning_rate/qwen_vl_interface': 3.661068665133735e-06, 'learning_rate/action_model': 3.661068665133735e-05, 'epoch': 0.23}) +[2026-10-03 19:33:09,625][starVLA.training.train_starvla][INFO] - Step 3083, Loss: {'train/loss': 0.0010796557180583477, 'train/grad_norm_pre_clip': 2.3444058895111084, 'timing/data': 0.00025474000722169876, 'timing/model': 0.39411867398303, 'learning_rate/qwen_vl_interface': 3.6581990124936434e-06, 'learning_rate/action_model': 3.6581990124936434e-05, 'epoch': 0.23}) +[2026-10-03 19:33:10,331][starVLA.training.train_starvla][INFO] - Step 3084, Loss: {'train/loss': 0.00013134549953974783, 'train/grad_norm_pre_clip': 0.12931540608406067, 'timing/data': 0.00022941999486647546, 'timing/model': 0.3944996740028728, 'learning_rate/qwen_vl_interface': 3.65533001418281e-06, 'learning_rate/action_model': 3.655330014182809e-05, 'epoch': 0.23}) +[2026-10-03 19:33:11,038][starVLA.training.train_starvla][INFO] - Step 3085, Loss: {'train/loss': 1.735901605570689e-05, 'train/grad_norm_pre_clip': 0.013464681804180145, 'timing/data': 0.00023929000599309802, 'timing/model': 0.3947188339952845, 'learning_rate/qwen_vl_interface': 3.652461671380572e-06, 'learning_rate/action_model': 3.652461671380572e-05, 'epoch': 0.23}) +[2026-10-03 19:33:11,746][starVLA.training.train_starvla][INFO] - Step 3086, Loss: {'train/loss': 0.0009920658776536584, 'train/grad_norm_pre_clip': 1.5502172708511353, 'timing/data': 0.00037460000021383166, 'timing/model': 0.39588337499299087, 'learning_rate/qwen_vl_interface': 3.6495939852659966e-06, 'learning_rate/action_model': 3.649593985265997e-05, 'epoch': 0.23}) +[2026-10-03 19:33:12,651][starVLA.training.train_starvla][INFO] - Step 3087, Loss: {'train/loss': 6.997708260314539e-05, 'train/grad_norm_pre_clip': 0.12072201073169708, 'timing/data': 0.00031856997520662844, 'timing/model': 0.592874016991118, 'learning_rate/qwen_vl_interface': 3.646726957017881e-06, 'learning_rate/action_model': 3.646726957017881e-05, 'epoch': 0.23}) +[2026-10-03 19:33:13,358][starVLA.training.train_starvla][INFO] - Step 3088, Loss: {'train/loss': 0.000696543778758496, 'train/grad_norm_pre_clip': 0.7433342933654785, 'timing/data': 0.0002962999860756099, 'timing/model': 0.3956193439953495, 'learning_rate/qwen_vl_interface': 3.643860587814752e-06, 'learning_rate/action_model': 3.643860587814752e-05, 'epoch': 0.23}) +[2026-10-03 19:33:14,066][starVLA.training.train_starvla][INFO] - Step 3089, Loss: {'train/loss': 0.06564450263977051, 'train/grad_norm_pre_clip': 335.2510681152344, 'timing/data': 0.0003521799808368087, 'timing/model': 0.39669531397521496, 'learning_rate/qwen_vl_interface': 3.640994878834866e-06, 'learning_rate/action_model': 3.640994878834866e-05, 'epoch': 0.23}) +[2026-10-03 19:33:14,773][starVLA.training.train_starvla][INFO] - Step 3090, Loss: {'train/loss': 7.024747901596129e-05, 'train/grad_norm_pre_clip': 0.04097418487071991, 'timing/data': 0.00022419000742956996, 'timing/model': 0.39430109498789534, 'learning_rate/qwen_vl_interface': 3.6381298312562095e-06, 'learning_rate/action_model': 3.6381298312562094e-05, 'epoch': 0.23}) +[2026-10-03 19:33:15,478][starVLA.training.train_starvla][INFO] - Step 3091, Loss: {'train/loss': 1.1264859494986013e-05, 'train/grad_norm_pre_clip': 0.009366057813167572, 'timing/data': 0.000291559990728274, 'timing/model': 0.3954721340269316, 'learning_rate/qwen_vl_interface': 3.635265446256493e-06, 'learning_rate/action_model': 3.6352654462564926e-05, 'epoch': 0.23}) +[2026-10-03 19:33:16,187][starVLA.training.train_starvla][INFO] - Step 3092, Loss: {'train/loss': 0.0014479694655165076, 'train/grad_norm_pre_clip': 3.736908197402954, 'timing/data': 0.00042762901284731925, 'timing/model': 0.3951209149963688, 'learning_rate/qwen_vl_interface': 3.6324017250131565e-06, 'learning_rate/action_model': 3.6324017250131567e-05, 'epoch': 0.23}) +[2026-10-03 19:33:17,098][starVLA.training.train_starvla][INFO] - Step 3093, Loss: {'train/loss': 7.465320322808111e-06, 'train/grad_norm_pre_clip': 0.001570576918311417, 'timing/data': 0.00031688998569734395, 'timing/model': 0.3954269440146163, 'learning_rate/qwen_vl_interface': 3.6295386687033696e-06, 'learning_rate/action_model': 3.6295386687033695e-05, 'epoch': 0.23}) +[2026-10-03 19:33:17,804][starVLA.training.train_starvla][INFO] - Step 3094, Loss: {'train/loss': 9.888396743917838e-05, 'train/grad_norm_pre_clip': 0.27181997895240784, 'timing/data': 0.00021157998708076775, 'timing/model': 0.39474002399947494, 'learning_rate/qwen_vl_interface': 3.6266762785040253e-06, 'learning_rate/action_model': 3.626676278504025e-05, 'epoch': 0.23}) +[2026-10-03 19:33:18,510][starVLA.training.train_starvla][INFO] - Step 3095, Loss: {'train/loss': 0.5570693612098694, 'train/grad_norm_pre_clip': 51.8780632019043, 'timing/data': 0.0003223799867555499, 'timing/model': 0.394574953999836, 'learning_rate/qwen_vl_interface': 3.6238145555917464e-06, 'learning_rate/action_model': 3.6238145555917465e-05, 'epoch': 0.23}) +[2026-10-03 19:33:19,217][starVLA.training.train_starvla][INFO] - Step 3096, Loss: {'train/loss': 5.565214087255299e-05, 'train/grad_norm_pre_clip': 0.016329843550920486, 'timing/data': 0.0002245499927084893, 'timing/model': 0.39594867400592193, 'learning_rate/qwen_vl_interface': 3.6209535011428757e-06, 'learning_rate/action_model': 3.620953501142876e-05, 'epoch': 0.23}) +[2026-10-03 19:33:19,925][starVLA.training.train_starvla][INFO] - Step 3097, Loss: {'train/loss': 0.24612079560756683, 'train/grad_norm_pre_clip': 70.12556457519531, 'timing/data': 0.00022776899277232587, 'timing/model': 0.39479236499755643, 'learning_rate/qwen_vl_interface': 3.6180931163334863e-06, 'learning_rate/action_model': 3.6180931163334863e-05, 'epoch': 0.23}) +[2026-10-03 19:33:20,826][starVLA.training.train_starvla][INFO] - Step 3098, Loss: {'train/loss': 0.008674297481775284, 'train/grad_norm_pre_clip': 75.18135833740234, 'timing/data': 0.00021645001834258437, 'timing/model': 0.5905505560222082, 'learning_rate/qwen_vl_interface': 3.6152334023393747e-06, 'learning_rate/action_model': 3.6152334023393745e-05, 'epoch': 0.23}) +[2026-10-03 19:33:21,532][starVLA.training.train_starvla][INFO] - Step 3099, Loss: {'train/loss': 7.897433533798903e-06, 'train/grad_norm_pre_clip': 0.0032140135299414396, 'timing/data': 0.00022059999173507094, 'timing/model': 0.3932666039909236, 'learning_rate/qwen_vl_interface': 3.6123743603360613e-06, 'learning_rate/action_model': 3.612374360336061e-05, 'epoch': 0.23}) +[2026-10-03 19:33:22,239][starVLA.training.train_starvla][INFO] - Step 3100, Loss: {'train/loss': 0.08843530714511871, 'train/grad_norm_pre_clip': 182.63800048828125, 'timing/data': 0.00022857901058159769, 'timing/model': 0.3949046549969353, 'learning_rate/qwen_vl_interface': 3.6095159914987893e-06, 'learning_rate/action_model': 3.6095159914987894e-05, 'epoch': 0.23}) +[2026-10-03 19:33:22,920][starVLA.training.train_starvla][INFO] - Step 3101, Loss: {'train/loss': 0.24017517268657684, 'train/grad_norm_pre_clip': 89.68748474121094, 'timing/data': 0.00030665998929180205, 'timing/model': 0.3950343340111431, 'learning_rate/qwen_vl_interface': 3.6066582970025266e-06, 'learning_rate/action_model': 3.6066582970025264e-05, 'epoch': 0.23}) +[2026-10-03 19:33:23,623][starVLA.training.train_starvla][INFO] - Step 3102, Loss: {'train/loss': 0.08949856460094452, 'train/grad_norm_pre_clip': 381.7270812988281, 'timing/data': 0.00030906000756658614, 'timing/model': 0.39323750499170274, 'learning_rate/qwen_vl_interface': 3.603801278021964e-06, 'learning_rate/action_model': 3.6038012780219635e-05, 'epoch': 0.23}) +[2026-10-03 19:33:24,326][starVLA.training.train_starvla][INFO] - Step 3103, Loss: {'train/loss': 3.1952517019817606e-05, 'train/grad_norm_pre_clip': 0.0196553822606802, 'timing/data': 0.00023502999101765454, 'timing/model': 0.39222315402003005, 'learning_rate/qwen_vl_interface': 3.6009449357315122e-06, 'learning_rate/action_model': 3.6009449357315125e-05, 'epoch': 0.23}) +[2026-10-03 19:33:25,230][starVLA.training.train_starvla][INFO] - Step 3104, Loss: {'train/loss': 3.954407657147385e-05, 'train/grad_norm_pre_clip': 0.020208492875099182, 'timing/data': 0.00022517002071253955, 'timing/model': 0.39350182499038056, 'learning_rate/qwen_vl_interface': 3.5980892713053084e-06, 'learning_rate/action_model': 3.598089271305308e-05, 'epoch': 0.23}) +[2026-10-03 19:33:25,937][starVLA.training.train_starvla][INFO] - Step 3105, Loss: {'train/loss': 7.591676694573835e-05, 'train/grad_norm_pre_clip': 0.05005582794547081, 'timing/data': 0.00033621001057326794, 'timing/model': 0.39573615399422124, 'learning_rate/qwen_vl_interface': 3.5952342859172063e-06, 'learning_rate/action_model': 3.595234285917206e-05, 'epoch': 0.23}) +[2026-10-03 19:33:26,643][starVLA.training.train_starvla][INFO] - Step 3106, Loss: {'train/loss': 2.175564532080898e-06, 'train/grad_norm_pre_clip': 0.00040477767470292747, 'timing/data': 0.00027088000206276774, 'timing/model': 0.3947512840095442, 'learning_rate/qwen_vl_interface': 3.592379980740785e-06, 'learning_rate/action_model': 3.5923799807407845e-05, 'epoch': 0.23}) +[2026-10-03 19:33:27,348][starVLA.training.train_starvla][INFO] - Step 3107, Loss: {'train/loss': 1.0599225759506226, 'train/grad_norm_pre_clip': 75.66152954101562, 'timing/data': 0.00020691001554951072, 'timing/model': 0.39393466399633326, 'learning_rate/qwen_vl_interface': 3.5895263569493384e-06, 'learning_rate/action_model': 3.589526356949338e-05, 'epoch': 0.23}) +[2026-10-03 19:33:28,054][starVLA.training.train_starvla][INFO] - Step 3108, Loss: {'train/loss': 0.5481341481208801, 'train/grad_norm_pre_clip': 141.95777893066406, 'timing/data': 0.0002912700001616031, 'timing/model': 0.39375467400532216, 'learning_rate/qwen_vl_interface': 3.586673415715885e-06, 'learning_rate/action_model': 3.586673415715885e-05, 'epoch': 0.23}) +[2026-10-03 19:33:28,962][starVLA.training.train_starvla][INFO] - Step 3109, Loss: {'train/loss': 0.5938272476196289, 'train/grad_norm_pre_clip': 182.1601104736328, 'timing/data': 0.0003495599958114326, 'timing/model': 0.394727225007955, 'learning_rate/qwen_vl_interface': 3.5838211582131603e-06, 'learning_rate/action_model': 3.58382115821316e-05, 'epoch': 0.23}) +[2026-10-03 19:33:29,668][starVLA.training.train_starvla][INFO] - Step 3110, Loss: {'train/loss': 0.00015905870532151312, 'train/grad_norm_pre_clip': 0.04808128625154495, 'timing/data': 0.00030600998434238136, 'timing/model': 0.3938597640080843, 'learning_rate/qwen_vl_interface': 3.580969585613622e-06, 'learning_rate/action_model': 3.580969585613622e-05, 'epoch': 0.23}) +[2026-10-03 19:33:30,371][starVLA.training.train_starvla][INFO] - Step 3111, Loss: {'train/loss': 0.14673517644405365, 'train/grad_norm_pre_clip': 93.16786193847656, 'timing/data': 0.00021574998390860856, 'timing/model': 0.39267241398920305, 'learning_rate/qwen_vl_interface': 3.5781186990894423e-06, 'learning_rate/action_model': 3.578118699089442e-05, 'epoch': 0.23}) +[2026-10-03 19:33:31,079][starVLA.training.train_starvla][INFO] - Step 3112, Loss: {'train/loss': 0.0003392978978808969, 'train/grad_norm_pre_clip': 0.1256350427865982, 'timing/data': 0.0003588800027500838, 'timing/model': 0.39552789498702623, 'learning_rate/qwen_vl_interface': 3.5752684998125113e-06, 'learning_rate/action_model': 3.5752684998125116e-05, 'epoch': 0.23}) +[2026-10-03 19:33:31,785][starVLA.training.train_starvla][INFO] - Step 3113, Loss: {'train/loss': 0.02403760701417923, 'train/grad_norm_pre_clip': 33.39318084716797, 'timing/data': 0.0003695300256367773, 'timing/model': 0.3946991440025158, 'learning_rate/qwen_vl_interface': 3.5724189889544407e-06, 'learning_rate/action_model': 3.572418988954441e-05, 'epoch': 0.23}) +[2026-10-03 19:33:32,692][starVLA.training.train_starvla][INFO] - Step 3114, Loss: {'train/loss': 0.00012534897541627288, 'train/grad_norm_pre_clip': 0.0691186785697937, 'timing/data': 0.00023032998433336616, 'timing/model': 0.5944909160025418, 'learning_rate/qwen_vl_interface': 3.569570167686556e-06, 'learning_rate/action_model': 3.569570167686556e-05, 'epoch': 0.23}) +[2026-10-03 19:33:33,397][starVLA.training.train_starvla][INFO] - Step 3115, Loss: {'train/loss': 0.41082367300987244, 'train/grad_norm_pre_clip': 44.40130615234375, 'timing/data': 0.00031785000464878976, 'timing/model': 0.39448039399576373, 'learning_rate/qwen_vl_interface': 3.5667220371798995e-06, 'learning_rate/action_model': 3.566722037179899e-05, 'epoch': 0.23}) +[2026-10-03 19:33:34,105][starVLA.training.train_starvla][INFO] - Step 3116, Loss: {'train/loss': 6.388527253875509e-05, 'train/grad_norm_pre_clip': 0.03458930924534798, 'timing/data': 0.00037426999188028276, 'timing/model': 0.3950369849917479, 'learning_rate/qwen_vl_interface': 3.563874598605232e-06, 'learning_rate/action_model': 3.5638745986052316e-05, 'epoch': 0.23}) +[2026-10-03 19:33:34,811][starVLA.training.train_starvla][INFO] - Step 3117, Loss: {'train/loss': 0.19223017990589142, 'train/grad_norm_pre_clip': 258.7411193847656, 'timing/data': 0.00032970000756904483, 'timing/model': 0.396556234016316, 'learning_rate/qwen_vl_interface': 3.561027853133025e-06, 'learning_rate/action_model': 3.561027853133025e-05, 'epoch': 0.23}) +[2026-10-03 19:33:35,518][starVLA.training.train_starvla][INFO] - Step 3118, Loss: {'train/loss': 0.00026555830845609307, 'train/grad_norm_pre_clip': 0.19947092235088348, 'timing/data': 0.0003007500199601054, 'timing/model': 0.3948092039790936, 'learning_rate/qwen_vl_interface': 3.558181801933472e-06, 'learning_rate/action_model': 3.5581818019334714e-05, 'epoch': 0.23}) +[2026-10-03 19:33:36,422][starVLA.training.train_starvla][INFO] - Step 3119, Loss: {'train/loss': 0.0014512604102492332, 'train/grad_norm_pre_clip': 1.2544971704483032, 'timing/data': 0.00032410002313554287, 'timing/model': 0.5921488659805618, 'learning_rate/qwen_vl_interface': 3.555336446176473e-06, 'learning_rate/action_model': 3.5553364461764726e-05, 'epoch': 0.23}) +[2026-10-03 19:33:37,130][starVLA.training.train_starvla][INFO] - Step 3120, Loss: {'train/loss': 0.15822988748550415, 'train/grad_norm_pre_clip': 69.39056396484375, 'timing/data': 0.0002269200049340725, 'timing/model': 0.39516053299303167, 'learning_rate/qwen_vl_interface': 3.552491787031647e-06, 'learning_rate/action_model': 3.552491787031647e-05, 'epoch': 0.23}) +[2026-10-03 19:33:37,836][starVLA.training.train_starvla][INFO] - Step 3121, Loss: {'train/loss': 0.664206862449646, 'train/grad_norm_pre_clip': 44.049583435058594, 'timing/data': 0.00021734999609179795, 'timing/model': 0.3957450950110797, 'learning_rate/qwen_vl_interface': 3.5496478256683277e-06, 'learning_rate/action_model': 3.549647825668327e-05, 'epoch': 0.23}) +[2026-10-03 19:33:38,542][starVLA.training.train_starvla][INFO] - Step 3122, Loss: {'train/loss': 1.7582891814527102e-05, 'train/grad_norm_pre_clip': 0.004053314216434956, 'timing/data': 0.00025059000472538173, 'timing/model': 0.394693004025612, 'learning_rate/qwen_vl_interface': 3.5468045632555585e-06, 'learning_rate/action_model': 3.546804563255559e-05, 'epoch': 0.23}) +[2026-10-03 19:33:39,249][starVLA.training.train_starvla][INFO] - Step 3123, Loss: {'train/loss': 0.00016356540436390787, 'train/grad_norm_pre_clip': 0.05820750072598457, 'timing/data': 0.00021361999097280204, 'timing/model': 0.39456418401096016, 'learning_rate/qwen_vl_interface': 3.5439620009620995e-06, 'learning_rate/action_model': 3.543962000962099e-05, 'epoch': 0.23}) +[2026-10-03 19:33:39,955][starVLA.training.train_starvla][INFO] - Step 3124, Loss: {'train/loss': 0.43024393916130066, 'train/grad_norm_pre_clip': 31.987382888793945, 'timing/data': 0.0002167790080420673, 'timing/model': 0.39531031498336233, 'learning_rate/qwen_vl_interface': 3.541120139956419e-06, 'learning_rate/action_model': 3.5411201399564184e-05, 'epoch': 0.23}) +[2026-10-03 19:33:40,880][starVLA.training.train_starvla][INFO] - Step 3125, Loss: {'train/loss': 0.10436547547578812, 'train/grad_norm_pre_clip': 83.66192626953125, 'timing/data': 0.00024031999055296183, 'timing/model': 0.3948891250183806, 'learning_rate/qwen_vl_interface': 3.538278981406697e-06, 'learning_rate/action_model': 3.5382789814066966e-05, 'epoch': 0.23}) +[2026-10-03 19:33:41,585][starVLA.training.train_starvla][INFO] - Step 3126, Loss: {'train/loss': 0.0004464062221813947, 'train/grad_norm_pre_clip': 0.20073920488357544, 'timing/data': 0.00037796999095007777, 'timing/model': 0.39366663500550203, 'learning_rate/qwen_vl_interface': 3.5354385264808295e-06, 'learning_rate/action_model': 3.535438526480829e-05, 'epoch': 0.23}) +[2026-10-03 19:33:42,291][starVLA.training.train_starvla][INFO] - Step 3127, Loss: {'train/loss': 0.0006062707398086786, 'train/grad_norm_pre_clip': 0.3123324513435364, 'timing/data': 0.00021838999236933887, 'timing/model': 0.3943993939901702, 'learning_rate/qwen_vl_interface': 3.5325987763464196e-06, 'learning_rate/action_model': 3.5325987763464194e-05, 'epoch': 0.23}) +[2026-10-03 19:33:42,998][starVLA.training.train_starvla][INFO] - Step 3128, Loss: {'train/loss': 0.00023449675063602626, 'train/grad_norm_pre_clip': 0.09030753374099731, 'timing/data': 0.0002823399845510721, 'timing/model': 0.3944578140217345, 'learning_rate/qwen_vl_interface': 3.529759732170781e-06, 'learning_rate/action_model': 3.529759732170781e-05, 'epoch': 0.23}) +[2026-10-03 19:33:43,705][starVLA.training.train_starvla][INFO] - Step 3129, Loss: {'train/loss': 0.00042917171958833933, 'train/grad_norm_pre_clip': 1.1233965158462524, 'timing/data': 0.00021383000421337783, 'timing/model': 0.39614672400057316, 'learning_rate/qwen_vl_interface': 3.5269213951209375e-06, 'learning_rate/action_model': 3.5269213951209374e-05, 'epoch': 0.23}) +[2026-10-03 19:33:44,608][starVLA.training.train_starvla][INFO] - Step 3130, Loss: {'train/loss': 9.220931679010391e-05, 'train/grad_norm_pre_clip': 0.060779377818107605, 'timing/data': 0.0002229499805252999, 'timing/model': 0.5928959759767167, 'learning_rate/qwen_vl_interface': 3.5240837663636207e-06, 'learning_rate/action_model': 3.5240837663636204e-05, 'epoch': 0.23}) +[2026-10-03 19:33:45,313][starVLA.training.train_starvla][INFO] - Step 3131, Loss: {'train/loss': 5.396638880483806e-05, 'train/grad_norm_pre_clip': 0.011494654230773449, 'timing/data': 0.00022325001191347837, 'timing/model': 0.3942003439879045, 'learning_rate/qwen_vl_interface': 3.5212468470652745e-06, 'learning_rate/action_model': 3.521246847065274e-05, 'epoch': 0.23}) +[2026-10-03 19:33:46,018][starVLA.training.train_starvla][INFO] - Step 3132, Loss: {'train/loss': 0.3129175007343292, 'train/grad_norm_pre_clip': 50.473121643066406, 'timing/data': 0.00022645998978987336, 'timing/model': 0.3946472739917226, 'learning_rate/qwen_vl_interface': 3.5184106383920487e-06, 'learning_rate/action_model': 3.5184106383920486e-05, 'epoch': 0.23}) +[2026-10-03 19:33:46,725][starVLA.training.train_starvla][INFO] - Step 3133, Loss: {'train/loss': 0.001044359989464283, 'train/grad_norm_pre_clip': 1.0455281734466553, 'timing/data': 0.0003761000116355717, 'timing/model': 0.3960765539959539, 'learning_rate/qwen_vl_interface': 3.5155751415098022e-06, 'learning_rate/action_model': 3.515575141509802e-05, 'epoch': 0.23}) +[2026-10-03 19:33:47,429][starVLA.training.train_starvla][INFO] - Step 3134, Loss: {'train/loss': 5.154300743015483e-05, 'train/grad_norm_pre_clip': 0.019513148814439774, 'timing/data': 0.00021065998589619994, 'timing/model': 0.39461734300130047, 'learning_rate/qwen_vl_interface': 3.5127403575841e-06, 'learning_rate/action_model': 3.5127403575841e-05, 'epoch': 0.23}) +[2026-10-03 19:33:48,338][starVLA.training.train_starvla][INFO] - Step 3135, Loss: {'train/loss': 0.6522786021232605, 'train/grad_norm_pre_clip': 89.54710388183594, 'timing/data': 0.00021333000040613115, 'timing/model': 0.5984617960057221, 'learning_rate/qwen_vl_interface': 3.509906287780215e-06, 'learning_rate/action_model': 3.5099062877802144e-05, 'epoch': 0.23}) +[2026-10-03 19:33:49,045][starVLA.training.train_starvla][INFO] - Step 3136, Loss: {'train/loss': 0.36071932315826416, 'train/grad_norm_pre_clip': 30.64407730102539, 'timing/data': 0.00022966001415625215, 'timing/model': 0.39375467397621833, 'learning_rate/qwen_vl_interface': 3.5070729332631247e-06, 'learning_rate/action_model': 3.507072933263125e-05, 'epoch': 0.23}) +[2026-10-03 19:33:49,752][starVLA.training.train_starvla][INFO] - Step 3137, Loss: {'train/loss': 0.01453342940658331, 'train/grad_norm_pre_clip': 17.828222274780273, 'timing/data': 0.0003036900016013533, 'timing/model': 0.3959786639898084, 'learning_rate/qwen_vl_interface': 3.5042402951975167e-06, 'learning_rate/action_model': 3.5042402951975167e-05, 'epoch': 0.23}) +[2026-10-03 19:33:50,459][starVLA.training.train_starvla][INFO] - Step 3138, Loss: {'train/loss': 0.7463863492012024, 'train/grad_norm_pre_clip': 93.97107696533203, 'timing/data': 0.00036171998362988234, 'timing/model': 0.3945463339914568, 'learning_rate/qwen_vl_interface': 3.50140837474778e-06, 'learning_rate/action_model': 3.50140837474778e-05, 'epoch': 0.23}) +[2026-10-03 19:33:51,164][starVLA.training.train_starvla][INFO] - Step 3139, Loss: {'train/loss': 0.0006098365411162376, 'train/grad_norm_pre_clip': 0.7999417185783386, 'timing/data': 0.0003969999961555004, 'timing/model': 0.3935909240099136, 'learning_rate/qwen_vl_interface': 3.498577173078011e-06, 'learning_rate/action_model': 3.498577173078011e-05, 'epoch': 0.23}) +[2026-10-03 19:33:51,870][starVLA.training.train_starvla][INFO] - Step 3140, Loss: {'train/loss': 2.6441359295859e-05, 'train/grad_norm_pre_clip': 0.006294284947216511, 'timing/data': 0.0002546700125094503, 'timing/model': 0.39522690398734994, 'learning_rate/qwen_vl_interface': 3.495746691352008e-06, 'learning_rate/action_model': 3.495746691352008e-05, 'epoch': 0.23}) +[2026-10-03 19:33:52,767][starVLA.training.train_starvla][INFO] - Step 3141, Loss: {'train/loss': 5.676980072166771e-05, 'train/grad_norm_pre_clip': 0.009725131094455719, 'timing/data': 0.0003024399920832366, 'timing/model': 0.3924084039754234, 'learning_rate/qwen_vl_interface': 3.4929169307332763e-06, 'learning_rate/action_model': 3.492916930733276e-05, 'epoch': 0.23}) +[2026-10-03 19:33:53,471][starVLA.training.train_starvla][INFO] - Step 3142, Loss: {'train/loss': 0.004988388624042273, 'train/grad_norm_pre_clip': 4.555001258850098, 'timing/data': 0.0003690900048241019, 'timing/model': 0.3928496539883781, 'learning_rate/qwen_vl_interface': 3.4900878923850234e-06, 'learning_rate/action_model': 3.490087892385023e-05, 'epoch': 0.23}) +[2026-10-03 19:33:54,175][starVLA.training.train_starvla][INFO] - Step 3143, Loss: {'train/loss': 0.009094597771763802, 'train/grad_norm_pre_clip': 15.532383918762207, 'timing/data': 0.00036333000753074884, 'timing/model': 0.39343742400524206, 'learning_rate/qwen_vl_interface': 3.4872595774701607e-06, 'learning_rate/action_model': 3.487259577470161e-05, 'epoch': 0.23}) +[2026-10-03 19:33:54,880][starVLA.training.train_starvla][INFO] - Step 3144, Loss: {'train/loss': 0.14404459297657013, 'train/grad_norm_pre_clip': 24.80645751953125, 'timing/data': 0.00027553000836633146, 'timing/model': 0.3955643030058127, 'learning_rate/qwen_vl_interface': 3.4844319871513014e-06, 'learning_rate/action_model': 3.484431987151301e-05, 'epoch': 0.23}) +[2026-10-03 19:33:55,586][starVLA.training.train_starvla][INFO] - Step 3145, Loss: {'train/loss': 0.014605140313506126, 'train/grad_norm_pre_clip': 12.827598571777344, 'timing/data': 0.0003489700029604137, 'timing/model': 0.3942031250044238, 'learning_rate/qwen_vl_interface': 3.4816051225907606e-06, 'learning_rate/action_model': 3.48160512259076e-05, 'epoch': 0.23}) +[2026-10-03 19:33:56,492][starVLA.training.train_starvla][INFO] - Step 3146, Loss: {'train/loss': 7.463218207703903e-05, 'train/grad_norm_pre_clip': 0.0268479622900486, 'timing/data': 0.0002578100247774273, 'timing/model': 0.39341047502239235, 'learning_rate/qwen_vl_interface': 3.478778984950556e-06, 'learning_rate/action_model': 3.478778984950556e-05, 'epoch': 0.23}) +[2026-10-03 19:33:57,200][starVLA.training.train_starvla][INFO] - Step 3147, Loss: {'train/loss': 0.00022015292779542506, 'train/grad_norm_pre_clip': 0.13722513616085052, 'timing/data': 0.00027052999939769506, 'timing/model': 0.3961687340051867, 'learning_rate/qwen_vl_interface': 3.475953575392405e-06, 'learning_rate/action_model': 3.475953575392405e-05, 'epoch': 0.23}) +[2026-10-03 19:33:57,907][starVLA.training.train_starvla][INFO] - Step 3148, Loss: {'train/loss': 0.0011379505740478635, 'train/grad_norm_pre_clip': 0.8713392019271851, 'timing/data': 0.0003333200002089143, 'timing/model': 0.3947374240087811, 'learning_rate/qwen_vl_interface': 3.473128895077728e-06, 'learning_rate/action_model': 3.473128895077728e-05, 'epoch': 0.23}) +[2026-10-03 19:33:58,612][starVLA.training.train_starvla][INFO] - Step 3149, Loss: {'train/loss': 2.416892493783962e-05, 'train/grad_norm_pre_clip': 0.005593909882009029, 'timing/data': 0.0003753600176423788, 'timing/model': 0.3931018440052867, 'learning_rate/qwen_vl_interface': 3.4703049451676453e-06, 'learning_rate/action_model': 3.4703049451676454e-05, 'epoch': 0.23}) +[2026-10-03 19:33:59,316][starVLA.training.train_starvla][INFO] - Step 3150, Loss: {'train/loss': 0.025480791926383972, 'train/grad_norm_pre_clip': 53.069801330566406, 'timing/data': 0.00021023000590503216, 'timing/model': 0.3930334039905574, 'learning_rate/qwen_vl_interface': 3.4674817268229732e-06, 'learning_rate/action_model': 3.467481726822973e-05, 'epoch': 0.23}) +[2026-10-03 19:34:00,215][starVLA.training.train_starvla][INFO] - Step 3151, Loss: {'train/loss': 3.7315479858079925e-05, 'train/grad_norm_pre_clip': 0.011017443612217903, 'timing/data': 0.00027278001653030515, 'timing/model': 0.588439446000848, 'learning_rate/qwen_vl_interface': 3.464659241204232e-06, 'learning_rate/action_model': 3.4646592412042314e-05, 'epoch': 0.23}) +[2026-10-03 19:34:00,920][starVLA.training.train_starvla][INFO] - Step 3152, Loss: {'train/loss': 0.0005094681400805712, 'train/grad_norm_pre_clip': 1.4370577335357666, 'timing/data': 0.0002625499910209328, 'timing/model': 0.3949223239906132, 'learning_rate/qwen_vl_interface': 3.4618374894716377e-06, 'learning_rate/action_model': 3.461837489471638e-05, 'epoch': 0.23}) +[2026-10-03 19:34:01,626][starVLA.training.train_starvla][INFO] - Step 3153, Loss: {'train/loss': 0.4403753876686096, 'train/grad_norm_pre_clip': 96.94444274902344, 'timing/data': 0.00031912000849843025, 'timing/model': 0.393472853989806, 'learning_rate/qwen_vl_interface': 3.459016472785107e-06, 'learning_rate/action_model': 3.459016472785107e-05, 'epoch': 0.23}) +[2026-10-03 19:34:02,331][starVLA.training.train_starvla][INFO] - Step 3154, Loss: {'train/loss': 0.011470980942249298, 'train/grad_norm_pre_clip': 25.350614547729492, 'timing/data': 0.0003695900086313486, 'timing/model': 0.39430568399257027, 'learning_rate/qwen_vl_interface': 3.456196192304254e-06, 'learning_rate/action_model': 3.456196192304254e-05, 'epoch': 0.23}) +[2026-10-03 19:34:03,038][starVLA.training.train_starvla][INFO] - Step 3155, Loss: {'train/loss': 0.00012170446279924363, 'train/grad_norm_pre_clip': 0.06861459463834763, 'timing/data': 0.00027427999884821475, 'timing/model': 0.39408370497403666, 'learning_rate/qwen_vl_interface': 3.4533766491883887e-06, 'learning_rate/action_model': 3.453376649188388e-05, 'epoch': 0.23}) +[2026-10-03 19:34:03,940][starVLA.training.train_starvla][INFO] - Step 3156, Loss: {'train/loss': 0.0012807202292606235, 'train/grad_norm_pre_clip': 2.675213575363159, 'timing/data': 0.00020989999757148325, 'timing/model': 0.591162025986705, 'learning_rate/qwen_vl_interface': 3.450557844596517e-06, 'learning_rate/action_model': 3.450557844596517e-05, 'epoch': 0.23}) +[2026-10-03 19:34:04,647][starVLA.training.train_starvla][INFO] - Step 3157, Loss: {'train/loss': 0.0001134017511503771, 'train/grad_norm_pre_clip': 0.029907386749982834, 'timing/data': 0.0003745199937839061, 'timing/model': 0.39475271399714984, 'learning_rate/qwen_vl_interface': 3.447739779687345e-06, 'learning_rate/action_model': 3.4477397796873446e-05, 'epoch': 0.23}) +[2026-10-03 19:34:05,351][starVLA.training.train_starvla][INFO] - Step 3158, Loss: {'train/loss': 0.00014873551845084876, 'train/grad_norm_pre_clip': 0.104977086186409, 'timing/data': 0.00020857001072727144, 'timing/model': 0.3946933140105102, 'learning_rate/qwen_vl_interface': 3.4449224556192722e-06, 'learning_rate/action_model': 3.444922455619272e-05, 'epoch': 0.23}) +[2026-10-03 19:34:06,055][starVLA.training.train_starvla][INFO] - Step 3159, Loss: {'train/loss': 0.0018206509994342923, 'train/grad_norm_pre_clip': 1.7042605876922607, 'timing/data': 0.0003038500144612044, 'timing/model': 0.3938704640022479, 'learning_rate/qwen_vl_interface': 3.442105873550395e-06, 'learning_rate/action_model': 3.442105873550395e-05, 'epoch': 0.23}) +[2026-10-03 19:34:06,763][starVLA.training.train_starvla][INFO] - Step 3160, Loss: {'train/loss': 5.4438434744952247e-05, 'train/grad_norm_pre_clip': 0.03485820069909096, 'timing/data': 0.0002565999748185277, 'timing/model': 0.39528822401189245, 'learning_rate/qwen_vl_interface': 3.4392900346385022e-06, 'learning_rate/action_model': 3.4392900346385024e-05, 'epoch': 0.23}) +[2026-10-03 19:34:07,663][starVLA.training.train_starvla][INFO] - Step 3161, Loss: {'train/loss': 0.0008559153648093343, 'train/grad_norm_pre_clip': 0.5499804019927979, 'timing/data': 0.000370590016245842, 'timing/model': 0.5870403760054614, 'learning_rate/qwen_vl_interface': 3.4364749400410796e-06, 'learning_rate/action_model': 3.43647494004108e-05, 'epoch': 0.23}) +[2026-10-03 19:34:08,366][starVLA.training.train_starvla][INFO] - Step 3162, Loss: {'train/loss': 0.0041590966284275055, 'train/grad_norm_pre_clip': 2.8612496852874756, 'timing/data': 0.0002463300188537687, 'timing/model': 0.39352008400601335, 'learning_rate/qwen_vl_interface': 3.4336605909153066e-06, 'learning_rate/action_model': 3.433660590915306e-05, 'epoch': 0.23}) +[2026-10-03 19:34:09,074][starVLA.training.train_starvla][INFO] - Step 3163, Loss: {'train/loss': 0.3293685019016266, 'train/grad_norm_pre_clip': 65.16471099853516, 'timing/data': 0.0003209100104868412, 'timing/model': 0.39593489302205853, 'learning_rate/qwen_vl_interface': 3.430846988418057e-06, 'learning_rate/action_model': 3.430846988418057e-05, 'epoch': 0.23}) +[2026-10-03 19:34:09,780][starVLA.training.train_starvla][INFO] - Step 3164, Loss: {'train/loss': 7.194736826932058e-05, 'train/grad_norm_pre_clip': 0.029127437621355057, 'timing/data': 0.00029237999115139246, 'timing/model': 0.395189473987557, 'learning_rate/qwen_vl_interface': 3.428034133705893e-06, 'learning_rate/action_model': 3.428034133705893e-05, 'epoch': 0.23}) +[2026-10-03 19:34:10,487][starVLA.training.train_starvla][INFO] - Step 3165, Loss: {'train/loss': 0.5506710410118103, 'train/grad_norm_pre_clip': 216.18296813964844, 'timing/data': 0.0002694099966902286, 'timing/model': 0.3940064039779827, 'learning_rate/qwen_vl_interface': 3.4252220279350763e-06, 'learning_rate/action_model': 3.425222027935076e-05, 'epoch': 0.23}) +[2026-10-03 19:34:11,191][starVLA.training.train_starvla][INFO] - Step 3166, Loss: {'train/loss': 0.3103552758693695, 'train/grad_norm_pre_clip': 42.4334716796875, 'timing/data': 0.0003691400052048266, 'timing/model': 0.39431576401693746, 'learning_rate/qwen_vl_interface': 3.4224106722615547e-06, 'learning_rate/action_model': 3.422410672261554e-05, 'epoch': 0.23}) +[2026-10-03 19:34:12,100][starVLA.training.train_starvla][INFO] - Step 3167, Loss: {'train/loss': 0.007486530113965273, 'train/grad_norm_pre_clip': 11.109604835510254, 'timing/data': 0.00020796002354472876, 'timing/model': 0.4067292929976247, 'learning_rate/qwen_vl_interface': 3.419600067840973e-06, 'learning_rate/action_model': 3.419600067840973e-05, 'epoch': 0.23}) +[2026-10-03 19:34:12,810][starVLA.training.train_starvla][INFO] - Step 3168, Loss: {'train/loss': 0.05228078365325928, 'train/grad_norm_pre_clip': 57.99448776245117, 'timing/data': 0.00035061000380665064, 'timing/model': 0.39695982402190566, 'learning_rate/qwen_vl_interface': 3.4167902158286627e-06, 'learning_rate/action_model': 3.4167902158286626e-05, 'epoch': 0.23}) +[2026-10-03 19:34:13,515][starVLA.training.train_starvla][INFO] - Step 3169, Loss: {'train/loss': 0.00035897994530387223, 'train/grad_norm_pre_clip': 0.2406141757965088, 'timing/data': 0.00020944001153111458, 'timing/model': 0.39348879299359396, 'learning_rate/qwen_vl_interface': 3.4139811173796468e-06, 'learning_rate/action_model': 3.413981117379647e-05, 'epoch': 0.23}) +[2026-10-03 19:34:14,221][starVLA.training.train_starvla][INFO] - Step 3170, Loss: {'train/loss': 0.000389357766835019, 'train/grad_norm_pre_clip': 0.5658361911773682, 'timing/data': 0.00041949001024477184, 'timing/model': 0.39423740399070084, 'learning_rate/qwen_vl_interface': 3.411172773648642e-06, 'learning_rate/action_model': 3.411172773648642e-05, 'epoch': 0.23}) +[2026-10-03 19:34:14,927][starVLA.training.train_starvla][INFO] - Step 3171, Loss: {'train/loss': 0.00579100102186203, 'train/grad_norm_pre_clip': 6.441896915435791, 'timing/data': 0.0003072099934797734, 'timing/model': 0.39356239399057813, 'learning_rate/qwen_vl_interface': 3.408365185790052e-06, 'learning_rate/action_model': 3.4083651857900515e-05, 'epoch': 0.23}) +[2026-10-03 19:34:15,832][starVLA.training.train_starvla][INFO] - Step 3172, Loss: {'train/loss': 4.4312117097433656e-05, 'train/grad_norm_pre_clip': 0.020346254110336304, 'timing/data': 0.0002370899892412126, 'timing/model': 0.39401598399854265, 'learning_rate/qwen_vl_interface': 3.40555835495797e-06, 'learning_rate/action_model': 3.40555835495797e-05, 'epoch': 0.23}) +[2026-10-03 19:34:16,538][starVLA.training.train_starvla][INFO] - Step 3173, Loss: {'train/loss': 0.0003019121941179037, 'train/grad_norm_pre_clip': 0.33586516976356506, 'timing/data': 0.0003242499951738864, 'timing/model': 0.39492744399467483, 'learning_rate/qwen_vl_interface': 3.4027522823061763e-06, 'learning_rate/action_model': 3.4027522823061765e-05, 'epoch': 0.23}) +[2026-10-03 19:34:17,244][starVLA.training.train_starvla][INFO] - Step 3174, Loss: {'train/loss': 0.0001837332674767822, 'train/grad_norm_pre_clip': 0.10511227697134018, 'timing/data': 0.00025079998886212707, 'timing/model': 0.3962188530131243, 'learning_rate/qwen_vl_interface': 3.399946968988143e-06, 'learning_rate/action_model': 3.399946968988143e-05, 'epoch': 0.23}) +[2026-10-03 19:34:17,949][starVLA.training.train_starvla][INFO] - Step 3175, Loss: {'train/loss': 0.3759002387523651, 'train/grad_norm_pre_clip': 40.12638854980469, 'timing/data': 0.00020740000763908029, 'timing/model': 0.39471319399308413, 'learning_rate/qwen_vl_interface': 3.397142416157028e-06, 'learning_rate/action_model': 3.397142416157028e-05, 'epoch': 0.23}) +[2026-10-03 19:34:18,655][starVLA.training.train_starvla][INFO] - Step 3176, Loss: {'train/loss': 0.364438533782959, 'train/grad_norm_pre_clip': 143.0120391845703, 'timing/data': 0.0002392799942754209, 'timing/model': 0.39576851300080307, 'learning_rate/qwen_vl_interface': 3.3943386249656766e-06, 'learning_rate/action_model': 3.394338624965677e-05, 'epoch': 0.23}) +[2026-10-03 19:34:19,558][starVLA.training.train_starvla][INFO] - Step 3177, Loss: {'train/loss': 0.006745829246938229, 'train/grad_norm_pre_clip': 9.447013854980469, 'timing/data': 0.0003260900266468525, 'timing/model': 0.5917196150112431, 'learning_rate/qwen_vl_interface': 3.391535596566624e-06, 'learning_rate/action_model': 3.391535596566624e-05, 'epoch': 0.23}) +[2026-10-03 19:34:20,264][starVLA.training.train_starvla][INFO] - Step 3178, Loss: {'train/loss': 0.8697904348373413, 'train/grad_norm_pre_clip': 111.08317565917969, 'timing/data': 0.00024806000874377787, 'timing/model': 0.39574885397451, 'learning_rate/qwen_vl_interface': 3.388733332112086e-06, 'learning_rate/action_model': 3.3887333321120856e-05, 'epoch': 0.23}) +[2026-10-03 19:34:20,967][starVLA.training.train_starvla][INFO] - Step 3179, Loss: {'train/loss': 0.0006311277393251657, 'train/grad_norm_pre_clip': 0.3081853687763214, 'timing/data': 0.0002157400012947619, 'timing/model': 0.392932954011485, 'learning_rate/qwen_vl_interface': 3.3859318327539693e-06, 'learning_rate/action_model': 3.385931832753969e-05, 'epoch': 0.23}) +[2026-10-03 19:34:21,673][starVLA.training.train_starvla][INFO] - Step 3180, Loss: {'train/loss': 0.4169294238090515, 'train/grad_norm_pre_clip': 79.13629150390625, 'timing/data': 0.000277960003586486, 'timing/model': 0.3938742740137968, 'learning_rate/qwen_vl_interface': 3.3831310996438633e-06, 'learning_rate/action_model': 3.383131099643863e-05, 'epoch': 0.23}) +[2026-10-03 19:34:22,379][starVLA.training.train_starvla][INFO] - Step 3181, Loss: {'train/loss': 0.07123838365077972, 'train/grad_norm_pre_clip': 84.83360290527344, 'timing/data': 0.00020862999372184277, 'timing/model': 0.39528633499867283, 'learning_rate/qwen_vl_interface': 3.3803311339330447e-06, 'learning_rate/action_model': 3.3803311339330444e-05, 'epoch': 0.23}) +[2026-10-03 19:34:23,280][starVLA.training.train_starvla][INFO] - Step 3182, Loss: {'train/loss': 0.0014575751265510917, 'train/grad_norm_pre_clip': 0.8218607902526855, 'timing/data': 0.0002389100263826549, 'timing/model': 0.5891812860209029, 'learning_rate/qwen_vl_interface': 3.3775319367724746e-06, 'learning_rate/action_model': 3.3775319367724745e-05, 'epoch': 0.23}) +[2026-10-03 19:34:23,988][starVLA.training.train_starvla][INFO] - Step 3183, Loss: {'train/loss': 0.0007769763469696045, 'train/grad_norm_pre_clip': 0.6368224024772644, 'timing/data': 0.0002604899927973747, 'timing/model': 0.3959706339810509, 'learning_rate/qwen_vl_interface': 3.374733509312795e-06, 'learning_rate/action_model': 3.374733509312795e-05, 'epoch': 0.23}) +[2026-10-03 19:34:24,695][starVLA.training.train_starvla][INFO] - Step 3184, Loss: {'train/loss': 0.031413283199071884, 'train/grad_norm_pre_clip': 113.03654479980469, 'timing/data': 0.0003021300071850419, 'timing/model': 0.3951779950002674, 'learning_rate/qwen_vl_interface': 3.3719358527043345e-06, 'learning_rate/action_model': 3.371935852704334e-05, 'epoch': 0.23}) +[2026-10-03 19:34:25,399][starVLA.training.train_starvla][INFO] - Step 3185, Loss: {'train/loss': 0.053991954773664474, 'train/grad_norm_pre_clip': 64.94774627685547, 'timing/data': 0.00021104002371430397, 'timing/model': 0.39313398400554433, 'learning_rate/qwen_vl_interface': 3.3691389680971047e-06, 'learning_rate/action_model': 3.3691389680971044e-05, 'epoch': 0.23}) +[2026-10-03 19:34:26,107][starVLA.training.train_starvla][INFO] - Step 3186, Loss: {'train/loss': 1.7955277144210413e-05, 'train/grad_norm_pre_clip': 0.006181271281093359, 'timing/data': 0.00032252000528387725, 'timing/model': 0.3955420040001627, 'learning_rate/qwen_vl_interface': 3.3663428566407985e-06, 'learning_rate/action_model': 3.366342856640798e-05, 'epoch': 0.23}) +[2026-10-03 19:34:27,007][starVLA.training.train_starvla][INFO] - Step 3187, Loss: {'train/loss': 0.0008310035336762667, 'train/grad_norm_pre_clip': 1.304728627204895, 'timing/data': 0.00021426999592222273, 'timing/model': 0.5890093260095455, 'learning_rate/qwen_vl_interface': 3.3635475194847927e-06, 'learning_rate/action_model': 3.363547519484793e-05, 'epoch': 0.23}) +[2026-10-03 19:34:27,712][starVLA.training.train_starvla][INFO] - Step 3188, Loss: {'train/loss': 0.45210742950439453, 'train/grad_norm_pre_clip': 110.46936798095703, 'timing/data': 0.0002959200064651668, 'timing/model': 0.39354034399730153, 'learning_rate/qwen_vl_interface': 3.3607529577781433e-06, 'learning_rate/action_model': 3.360752957778143e-05, 'epoch': 0.23}) +[2026-10-03 19:34:28,417][starVLA.training.train_starvla][INFO] - Step 3189, Loss: {'train/loss': 0.015286232344806194, 'train/grad_norm_pre_clip': 36.18115997314453, 'timing/data': 0.0002833500038832426, 'timing/model': 0.3954162930021994, 'learning_rate/qwen_vl_interface': 3.3579591726695906e-06, 'learning_rate/action_model': 3.35795917266959e-05, 'epoch': 0.23}) +[2026-10-03 19:34:29,127][starVLA.training.train_starvla][INFO] - Step 3190, Loss: {'train/loss': 0.29220935702323914, 'train/grad_norm_pre_clip': 101.92343139648438, 'timing/data': 0.0002568199997767806, 'timing/model': 0.39604093399248086, 'learning_rate/qwen_vl_interface': 3.3551661653075522e-06, 'learning_rate/action_model': 3.355166165307552e-05, 'epoch': 0.23}) +[2026-10-03 19:34:29,839][starVLA.training.train_starvla][INFO] - Step 3191, Loss: {'train/loss': 0.40992632508277893, 'train/grad_norm_pre_clip': 41.936058044433594, 'timing/data': 0.0002507000172045082, 'timing/model': 0.3948771239956841, 'learning_rate/qwen_vl_interface': 3.35237393684013e-06, 'learning_rate/action_model': 3.35237393684013e-05, 'epoch': 0.23}) +[2026-10-03 19:34:30,543][starVLA.training.train_starvla][INFO] - Step 3192, Loss: {'train/loss': 0.011729374527931213, 'train/grad_norm_pre_clip': 41.00034713745117, 'timing/data': 0.0003143299836665392, 'timing/model': 0.39428220299305394, 'learning_rate/qwen_vl_interface': 3.349582488415103e-06, 'learning_rate/action_model': 3.3495824884151026e-05, 'epoch': 0.23}) +[2026-10-03 19:34:31,471][starVLA.training.train_starvla][INFO] - Step 3193, Loss: {'train/loss': 0.5236086249351501, 'train/grad_norm_pre_clip': 25.276647567749023, 'timing/data': 0.00030352998874150217, 'timing/model': 0.3933051440108102, 'learning_rate/qwen_vl_interface': 3.3467918211799294e-06, 'learning_rate/action_model': 3.3467918211799295e-05, 'epoch': 0.23}) +[2026-10-03 19:34:32,178][starVLA.training.train_starvla][INFO] - Step 3194, Loss: {'train/loss': 0.33264562487602234, 'train/grad_norm_pre_clip': 53.22453308105469, 'timing/data': 0.00027338002109900117, 'timing/model': 0.3949241240043193, 'learning_rate/qwen_vl_interface': 3.3440019362817476e-06, 'learning_rate/action_model': 3.344001936281747e-05, 'epoch': 0.23}) +[2026-10-03 19:34:32,884][starVLA.training.train_starvla][INFO] - Step 3195, Loss: {'train/loss': 2.332694930373691e-05, 'train/grad_norm_pre_clip': 0.007108539808541536, 'timing/data': 0.0003104299830738455, 'timing/model': 0.3944580639945343, 'learning_rate/qwen_vl_interface': 3.3412128348673745e-06, 'learning_rate/action_model': 3.341212834867374e-05, 'epoch': 0.23}) +[2026-10-03 19:34:33,589][starVLA.training.train_starvla][INFO] - Step 3196, Loss: {'train/loss': 0.22228965163230896, 'train/grad_norm_pre_clip': 52.79352569580078, 'timing/data': 0.0002962999860756099, 'timing/model': 0.39579290300025605, 'learning_rate/qwen_vl_interface': 3.338424518083303e-06, 'learning_rate/action_model': 3.338424518083303e-05, 'epoch': 0.23}) +[2026-10-03 19:34:34,296][starVLA.training.train_starvla][INFO] - Step 3197, Loss: {'train/loss': 0.46562659740448, 'train/grad_norm_pre_clip': 33.91200256347656, 'timing/data': 0.00026815998717211187, 'timing/model': 0.3952653239830397, 'learning_rate/qwen_vl_interface': 3.3356369870757056e-06, 'learning_rate/action_model': 3.335636987075705e-05, 'epoch': 0.23}) +[2026-10-03 19:34:35,201][starVLA.training.train_starvla][INFO] - Step 3198, Loss: {'train/loss': 1.4260184798331466e-05, 'train/grad_norm_pre_clip': 0.0036753525491803885, 'timing/data': 0.00020832999143749475, 'timing/model': 0.5932453350105789, 'learning_rate/qwen_vl_interface': 3.3328502429904327e-06, 'learning_rate/action_model': 3.3328502429904325e-05, 'epoch': 0.23}) +[2026-10-03 19:34:35,907][starVLA.training.train_starvla][INFO] - Step 3199, Loss: {'train/loss': 0.00036797011853195727, 'train/grad_norm_pre_clip': 0.22645264863967896, 'timing/data': 0.0003172899887431413, 'timing/model': 0.39459556399378926, 'learning_rate/qwen_vl_interface': 3.3300642869730066e-06, 'learning_rate/action_model': 3.330064286973006e-05, 'epoch': 0.23}) +[2026-10-03 19:34:36,613][starVLA.training.train_starvla][INFO] - Step 3200, Loss: {'train/loss': 0.0030444569420069456, 'train/grad_norm_pre_clip': 9.465536117553711, 'timing/data': 0.0002154200046788901, 'timing/model': 0.39291942401905544, 'learning_rate/qwen_vl_interface': 3.3272791201686294e-06, 'learning_rate/action_model': 3.327279120168629e-05, 'epoch': 0.23}) +[2026-10-03 19:34:37,320][starVLA.training.train_starvla][INFO] - Step 3201, Loss: {'train/loss': 0.32067766785621643, 'train/grad_norm_pre_clip': 68.12950897216797, 'timing/data': 0.0002726700040511787, 'timing/model': 0.3954234030097723, 'learning_rate/qwen_vl_interface': 3.324494743722179e-06, 'learning_rate/action_model': 3.3244947437221785e-05, 'epoch': 0.23}) +[2026-10-03 19:34:38,028][starVLA.training.train_starvla][INFO] - Step 3202, Loss: {'train/loss': 0.00017243623733520508, 'train/grad_norm_pre_clip': 0.17349183559417725, 'timing/data': 0.00027609997778199613, 'timing/model': 0.39584756299154833, 'learning_rate/qwen_vl_interface': 3.321711158778205e-06, 'learning_rate/action_model': 3.321711158778205e-05, 'epoch': 0.23}) +[2026-10-03 19:34:38,734][starVLA.training.train_starvla][INFO] - Step 3203, Loss: {'train/loss': 0.00011723303759936243, 'train/grad_norm_pre_clip': 0.04001122713088989, 'timing/data': 0.0003395899839233607, 'timing/model': 0.3954402639938053, 'learning_rate/qwen_vl_interface': 3.3189283664809393e-06, 'learning_rate/action_model': 3.318928366480939e-05, 'epoch': 0.23}) +[2026-10-03 19:34:39,631][starVLA.training.train_starvla][INFO] - Step 3204, Loss: {'train/loss': 0.02776370197534561, 'train/grad_norm_pre_clip': 96.26930236816406, 'timing/data': 0.0003484800108708441, 'timing/model': 0.394667443993967, 'learning_rate/qwen_vl_interface': 3.316146367974276e-06, 'learning_rate/action_model': 3.316146367974276e-05, 'epoch': 0.23}) +[2026-10-03 19:34:40,337][starVLA.training.train_starvla][INFO] - Step 3205, Loss: {'train/loss': 0.10390719026327133, 'train/grad_norm_pre_clip': 366.1587829589844, 'timing/data': 0.000334270007442683, 'timing/model': 0.39536950300680473, 'learning_rate/qwen_vl_interface': 3.3133651644017945e-06, 'learning_rate/action_model': 3.3133651644017944e-05, 'epoch': 0.23}) +[2026-10-03 19:34:41,043][starVLA.training.train_starvla][INFO] - Step 3206, Loss: {'train/loss': 0.24053043127059937, 'train/grad_norm_pre_clip': 69.41860961914062, 'timing/data': 0.0002593100070953369, 'timing/model': 0.3941542239917908, 'learning_rate/qwen_vl_interface': 3.31058475690674e-06, 'learning_rate/action_model': 3.31058475690674e-05, 'epoch': 0.23}) +[2026-10-03 19:34:41,747][starVLA.training.train_starvla][INFO] - Step 3207, Loss: {'train/loss': 0.0005223890184424818, 'train/grad_norm_pre_clip': 0.21062208712100983, 'timing/data': 0.00032714998815208673, 'timing/model': 0.3931339039991144, 'learning_rate/qwen_vl_interface': 3.3078051466320337e-06, 'learning_rate/action_model': 3.307805146632034e-05, 'epoch': 0.23}) +[2026-10-03 19:34:42,451][starVLA.training.train_starvla][INFO] - Step 3208, Loss: {'train/loss': 0.037628624588251114, 'train/grad_norm_pre_clip': 127.23916625976562, 'timing/data': 0.00020357000175863504, 'timing/model': 0.395037594018504, 'learning_rate/qwen_vl_interface': 3.3050263347202712e-06, 'learning_rate/action_model': 3.305026334720271e-05, 'epoch': 0.23}) +[2026-10-03 19:34:43,360][starVLA.training.train_starvla][INFO] - Step 3209, Loss: {'train/loss': 1.4155741155263968e-05, 'train/grad_norm_pre_clip': 0.003578626085072756, 'timing/data': 0.00034220999805256724, 'timing/model': 0.3952312830078881, 'learning_rate/qwen_vl_interface': 3.302248322313712e-06, 'learning_rate/action_model': 3.302248322313712e-05, 'epoch': 0.23}) +[2026-10-03 19:34:44,069][starVLA.training.train_starvla][INFO] - Step 3210, Loss: {'train/loss': 9.537747246213257e-05, 'train/grad_norm_pre_clip': 0.056439585983753204, 'timing/data': 0.00027057999977841973, 'timing/model': 0.39633792400127277, 'learning_rate/qwen_vl_interface': 3.299471110554295e-06, 'learning_rate/action_model': 3.299471110554295e-05, 'epoch': 0.23}) +[2026-10-03 19:34:44,775][starVLA.training.train_starvla][INFO] - Step 3211, Loss: {'train/loss': 0.11664912104606628, 'train/grad_norm_pre_clip': 44.96022033691406, 'timing/data': 0.00020931000472046435, 'timing/model': 0.394726983999135, 'learning_rate/qwen_vl_interface': 3.296694700583626e-06, 'learning_rate/action_model': 3.296694700583626e-05, 'epoch': 0.24}) +[2026-10-03 19:34:45,481][starVLA.training.train_starvla][INFO] - Step 3212, Loss: {'train/loss': 1.027526617050171, 'train/grad_norm_pre_clip': 118.709716796875, 'timing/data': 0.00031244001002050936, 'timing/model': 0.3931147840048652, 'learning_rate/qwen_vl_interface': 3.2939190935429833e-06, 'learning_rate/action_model': 3.293919093542983e-05, 'epoch': 0.24}) +[2026-10-03 19:34:46,184][starVLA.training.train_starvla][INFO] - Step 3213, Loss: {'train/loss': 4.735711627290584e-05, 'train/grad_norm_pre_clip': 0.03225475177168846, 'timing/data': 0.00021242001093924046, 'timing/model': 0.3939434440108016, 'learning_rate/qwen_vl_interface': 3.2911442905733126e-06, 'learning_rate/action_model': 3.291144290573312e-05, 'epoch': 0.24}) +[2026-10-03 19:34:47,087][starVLA.training.train_starvla][INFO] - Step 3214, Loss: {'train/loss': 0.5468560457229614, 'train/grad_norm_pre_clip': 38.26615905761719, 'timing/data': 0.00027557002613320947, 'timing/model': 0.5926294059900101, 'learning_rate/qwen_vl_interface': 3.2883702928152323e-06, 'learning_rate/action_model': 3.288370292815232e-05, 'epoch': 0.24}) +[2026-10-03 19:34:47,791][starVLA.training.train_starvla][INFO] - Step 3215, Loss: {'train/loss': 0.00011454337072791532, 'train/grad_norm_pre_clip': 0.027544837445020676, 'timing/data': 0.00030683999648317695, 'timing/model': 0.3933229240064975, 'learning_rate/qwen_vl_interface': 3.2855971014090275e-06, 'learning_rate/action_model': 3.285597101409028e-05, 'epoch': 0.24}) +[2026-10-03 19:34:48,495][starVLA.training.train_starvla][INFO] - Step 3216, Loss: {'train/loss': 0.00015298707876354456, 'train/grad_norm_pre_clip': 0.0345248281955719, 'timing/data': 0.0002440699900034815, 'timing/model': 0.3949099339952227, 'learning_rate/qwen_vl_interface': 3.2828247174946528e-06, 'learning_rate/action_model': 3.2828247174946524e-05, 'epoch': 0.24}) +[2026-10-03 19:34:49,202][starVLA.training.train_starvla][INFO] - Step 3217, Loss: {'train/loss': 0.13748976588249207, 'train/grad_norm_pre_clip': 37.810813903808594, 'timing/data': 0.00029337999876588583, 'timing/model': 0.39526595300412737, 'learning_rate/qwen_vl_interface': 3.280053142211729e-06, 'learning_rate/action_model': 3.280053142211729e-05, 'epoch': 0.24}) +[2026-10-03 19:34:49,906][starVLA.training.train_starvla][INFO] - Step 3218, Loss: {'train/loss': 0.1616731435060501, 'train/grad_norm_pre_clip': 171.258056640625, 'timing/data': 0.00034411001252010465, 'timing/model': 0.3943010039802175, 'learning_rate/qwen_vl_interface': 3.2772823766995475e-06, 'learning_rate/action_model': 3.277282376699548e-05, 'epoch': 0.24}) +[2026-10-03 19:34:50,611][starVLA.training.train_starvla][INFO] - Step 3219, Loss: {'train/loss': 0.00022594130132347345, 'train/grad_norm_pre_clip': 0.039111580699682236, 'timing/data': 0.00031167000997811556, 'timing/model': 0.3924532239907421, 'learning_rate/qwen_vl_interface': 3.2745124220970646e-06, 'learning_rate/action_model': 3.274512422097065e-05, 'epoch': 0.24}) +[2026-10-03 19:34:51,512][starVLA.training.train_starvla][INFO] - Step 3220, Loss: {'train/loss': 0.0020810605492442846, 'train/grad_norm_pre_clip': 1.2361267805099487, 'timing/data': 0.0003147600218653679, 'timing/model': 0.3941958929935936, 'learning_rate/qwen_vl_interface': 3.2717432795429057e-06, 'learning_rate/action_model': 3.271743279542905e-05, 'epoch': 0.24}) +[2026-10-03 19:34:52,218][starVLA.training.train_starvla][INFO] - Step 3221, Loss: {'train/loss': 0.40230804681777954, 'train/grad_norm_pre_clip': 123.53730010986328, 'timing/data': 0.0002819700166583061, 'timing/model': 0.3952327029837761, 'learning_rate/qwen_vl_interface': 3.2689749501753578e-06, 'learning_rate/action_model': 3.2689749501753574e-05, 'epoch': 0.24}) +[2026-10-03 19:34:52,926][starVLA.training.train_starvla][INFO] - Step 3222, Loss: {'train/loss': 0.0008185183396562934, 'train/grad_norm_pre_clip': 0.27382516860961914, 'timing/data': 0.00031060000765137374, 'timing/model': 0.3951484739955049, 'learning_rate/qwen_vl_interface': 3.2662074351323774e-06, 'learning_rate/action_model': 3.266207435132378e-05, 'epoch': 0.24}) +[2026-10-03 19:34:53,632][starVLA.training.train_starvla][INFO] - Step 3223, Loss: {'train/loss': 0.004504944663494825, 'train/grad_norm_pre_clip': 4.468513011932373, 'timing/data': 0.00021321000531315804, 'timing/model': 0.3946506030042656, 'learning_rate/qwen_vl_interface': 3.2634407355515875e-06, 'learning_rate/action_model': 3.2634407355515875e-05, 'epoch': 0.24}) +[2026-10-03 19:34:54,337][starVLA.training.train_starvla][INFO] - Step 3224, Loss: {'train/loss': 0.00017713330453261733, 'train/grad_norm_pre_clip': 0.04543226212263107, 'timing/data': 0.00020403999951668084, 'timing/model': 0.3941310129885096, 'learning_rate/qwen_vl_interface': 3.2606748525702715e-06, 'learning_rate/action_model': 3.260674852570271e-05, 'epoch': 0.24}) +[2026-10-03 19:34:55,242][starVLA.training.train_starvla][INFO] - Step 3225, Loss: {'train/loss': 0.000871954660397023, 'train/grad_norm_pre_clip': 0.23244608938694, 'timing/data': 0.00022437999723479152, 'timing/model': 0.3937665640260093, 'learning_rate/qwen_vl_interface': 3.2579097873253803e-06, 'learning_rate/action_model': 3.25790978732538e-05, 'epoch': 0.24}) +[2026-10-03 19:34:55,948][starVLA.training.train_starvla][INFO] - Step 3226, Loss: {'train/loss': 8.620205335319042e-06, 'train/grad_norm_pre_clip': 0.002825211500748992, 'timing/data': 0.00031451001996174455, 'timing/model': 0.39582558401161805, 'learning_rate/qwen_vl_interface': 3.2551455409535274e-06, 'learning_rate/action_model': 3.255145540953528e-05, 'epoch': 0.24}) +[2026-10-03 19:34:56,653][starVLA.training.train_starvla][INFO] - Step 3227, Loss: {'train/loss': 0.2734382450580597, 'train/grad_norm_pre_clip': 81.25462341308594, 'timing/data': 0.0003109399985987693, 'timing/model': 0.39414868399035186, 'learning_rate/qwen_vl_interface': 3.2523821145909907e-06, 'learning_rate/action_model': 3.252382114590991e-05, 'epoch': 0.24}) +[2026-10-03 19:34:57,358][starVLA.training.train_starvla][INFO] - Step 3228, Loss: {'train/loss': 0.0004889961564913392, 'train/grad_norm_pre_clip': 0.21089893579483032, 'timing/data': 0.00031969998963177204, 'timing/model': 0.3949783240095712, 'learning_rate/qwen_vl_interface': 3.249619509373712e-06, 'learning_rate/action_model': 3.249619509373712e-05, 'epoch': 0.24}) +[2026-10-03 19:34:58,064][starVLA.training.train_starvla][INFO] - Step 3229, Loss: {'train/loss': 0.0006501198513433337, 'train/grad_norm_pre_clip': 0.8473877906799316, 'timing/data': 0.0003270599991083145, 'timing/model': 0.3947800840251148, 'learning_rate/qwen_vl_interface': 3.2468577264372905e-06, 'learning_rate/action_model': 3.24685772643729e-05, 'epoch': 0.24}) +[2026-10-03 19:34:58,967][starVLA.training.train_starvla][INFO] - Step 3230, Loss: {'train/loss': 0.008756721392273903, 'train/grad_norm_pre_clip': 9.295968055725098, 'timing/data': 0.00020728001254610717, 'timing/model': 0.5927617950073909, 'learning_rate/qwen_vl_interface': 3.244096766916992e-06, 'learning_rate/action_model': 3.244096766916992e-05, 'epoch': 0.24}) +[2026-10-03 19:34:59,673][starVLA.training.train_starvla][INFO] - Step 3231, Loss: {'train/loss': 0.0017630475340411067, 'train/grad_norm_pre_clip': 1.692860722541809, 'timing/data': 0.0003318199887871742, 'timing/model': 0.3944350540114101, 'learning_rate/qwen_vl_interface': 3.241336631947743e-06, 'learning_rate/action_model': 3.2413366319477427e-05, 'epoch': 0.24}) +[2026-10-03 19:35:00,377][starVLA.training.train_starvla][INFO] - Step 3232, Loss: {'train/loss': 0.06932257860898972, 'train/grad_norm_pre_clip': 58.38446044921875, 'timing/data': 0.0003242000238969922, 'timing/model': 0.3932527840079274, 'learning_rate/qwen_vl_interface': 3.238577322664131e-06, 'learning_rate/action_model': 3.238577322664131e-05, 'epoch': 0.24}) +[2026-10-03 19:35:01,083][starVLA.training.train_starvla][INFO] - Step 3233, Loss: {'train/loss': 7.033349538687617e-05, 'train/grad_norm_pre_clip': 0.024105018004775047, 'timing/data': 0.0003533699782565236, 'timing/model': 0.39410780399339274, 'learning_rate/qwen_vl_interface': 3.2358188402004026e-06, 'learning_rate/action_model': 3.2358188402004024e-05, 'epoch': 0.24}) +[2026-10-03 19:35:01,789][starVLA.training.train_starvla][INFO] - Step 3234, Loss: {'train/loss': 0.005517688114196062, 'train/grad_norm_pre_clip': 10.604795455932617, 'timing/data': 0.00038667998160235584, 'timing/model': 0.39497099397704005, 'learning_rate/qwen_vl_interface': 3.2330611856904675e-06, 'learning_rate/action_model': 3.233061185690468e-05, 'epoch': 0.24}) +[2026-10-03 19:35:02,692][starVLA.training.train_starvla][INFO] - Step 3235, Loss: {'train/loss': 0.00014406113768927753, 'train/grad_norm_pre_clip': 0.046576905995607376, 'timing/data': 0.0003597899922169745, 'timing/model': 0.5927134649828076, 'learning_rate/qwen_vl_interface': 3.2303043602678928e-06, 'learning_rate/action_model': 3.230304360267893e-05, 'epoch': 0.24}) +[2026-10-03 19:35:03,398][starVLA.training.train_starvla][INFO] - Step 3236, Loss: {'train/loss': 5.623425749945454e-05, 'train/grad_norm_pre_clip': 0.0338624082505703, 'timing/data': 0.00021496997214853764, 'timing/model': 0.39470255401101895, 'learning_rate/qwen_vl_interface': 3.227548365065905e-06, 'learning_rate/action_model': 3.227548365065905e-05, 'epoch': 0.24}) +[2026-10-03 19:35:04,104][starVLA.training.train_starvla][INFO] - Step 3237, Loss: {'train/loss': 0.958073616027832, 'train/grad_norm_pre_clip': 139.5542449951172, 'timing/data': 0.00021627001115120947, 'timing/model': 0.39478461400722153, 'learning_rate/qwen_vl_interface': 3.2247932012173887e-06, 'learning_rate/action_model': 3.2247932012173884e-05, 'epoch': 0.24}) +[2026-10-03 19:35:04,810][starVLA.training.train_starvla][INFO] - Step 3238, Loss: {'train/loss': 0.0012149709509685636, 'train/grad_norm_pre_clip': 2.1060688495635986, 'timing/data': 0.00026606000028550625, 'timing/model': 0.39374218400917016, 'learning_rate/qwen_vl_interface': 3.2220388698548897e-06, 'learning_rate/action_model': 3.22203886985489e-05, 'epoch': 0.24}) +[2026-10-03 19:35:05,515][starVLA.training.train_starvla][INFO] - Step 3239, Loss: {'train/loss': 0.3534613847732544, 'train/grad_norm_pre_clip': 62.67387390136719, 'timing/data': 0.0002841800160240382, 'timing/model': 0.3945032139890827, 'learning_rate/qwen_vl_interface': 3.21928537211061e-06, 'learning_rate/action_model': 3.21928537211061e-05, 'epoch': 0.24}) +[2026-10-03 19:35:06,421][starVLA.training.train_starvla][INFO] - Step 3240, Loss: {'train/loss': 0.0004193316271994263, 'train/grad_norm_pre_clip': 0.22236581146717072, 'timing/data': 0.00031786999898031354, 'timing/model': 0.5968662360101007, 'learning_rate/qwen_vl_interface': 3.216532709116405e-06, 'learning_rate/action_model': 3.216532709116405e-05, 'epoch': 0.24}) +[2026-10-03 19:35:07,128][starVLA.training.train_starvla][INFO] - Step 3241, Loss: {'train/loss': 0.13528205454349518, 'train/grad_norm_pre_clip': 49.298622131347656, 'timing/data': 0.00024034001398831606, 'timing/model': 0.3960029129812028, 'learning_rate/qwen_vl_interface': 3.2137808820037925e-06, 'learning_rate/action_model': 3.2137808820037924e-05, 'epoch': 0.24}) +[2026-10-03 19:35:07,833][starVLA.training.train_starvla][INFO] - Step 3242, Loss: {'train/loss': 0.21000362932682037, 'train/grad_norm_pre_clip': 278.2444763183594, 'timing/data': 0.0002908799797296524, 'timing/model': 0.393763942993246, 'learning_rate/qwen_vl_interface': 3.211029891903945e-06, 'learning_rate/action_model': 3.211029891903945e-05, 'epoch': 0.24}) +[2026-10-03 19:35:08,536][starVLA.training.train_starvla][INFO] - Step 3243, Loss: {'train/loss': 0.533460259437561, 'train/grad_norm_pre_clip': 40.26860809326172, 'timing/data': 0.00020454998593777418, 'timing/model': 0.3927965140028391, 'learning_rate/qwen_vl_interface': 3.2082797399476894e-06, 'learning_rate/action_model': 3.2082797399476895e-05, 'epoch': 0.24}) +[2026-10-03 19:35:09,239][starVLA.training.train_starvla][INFO] - Step 3244, Loss: {'train/loss': 0.0007838743622414768, 'train/grad_norm_pre_clip': 0.6784621477127075, 'timing/data': 0.00033690998679958284, 'timing/model': 0.3927740440121852, 'learning_rate/qwen_vl_interface': 3.2055304272655114e-06, 'learning_rate/action_model': 3.205530427265511e-05, 'epoch': 0.24}) +[2026-10-03 19:35:09,944][starVLA.training.train_starvla][INFO] - Step 3245, Loss: {'train/loss': 0.002695465227589011, 'train/grad_norm_pre_clip': 7.515289306640625, 'timing/data': 0.00030591999529860914, 'timing/model': 0.3946985730144661, 'learning_rate/qwen_vl_interface': 3.2027819549875477e-06, 'learning_rate/action_model': 3.202781954987548e-05, 'epoch': 0.24}) +[2026-10-03 19:35:10,853][starVLA.training.train_starvla][INFO] - Step 3246, Loss: {'train/loss': 0.0004122350364923477, 'train/grad_norm_pre_clip': 0.15666784346103668, 'timing/data': 0.00022596001508645713, 'timing/model': 0.39357491399277933, 'learning_rate/qwen_vl_interface': 3.2000343242435926e-06, 'learning_rate/action_model': 3.2000343242435924e-05, 'epoch': 0.24}) +[2026-10-03 19:35:11,556][starVLA.training.train_starvla][INFO] - Step 3247, Loss: {'train/loss': 0.025883734226226807, 'train/grad_norm_pre_clip': 19.76703453063965, 'timing/data': 0.00020878002396784723, 'timing/model': 0.3926157430105377, 'learning_rate/qwen_vl_interface': 3.1972875361630923e-06, 'learning_rate/action_model': 3.1972875361630923e-05, 'epoch': 0.24}) +[2026-10-03 19:35:12,260][starVLA.training.train_starvla][INFO] - Step 3248, Loss: {'train/loss': 6.247520650504157e-05, 'train/grad_norm_pre_clip': 0.024474943056702614, 'timing/data': 0.00021053999080322683, 'timing/model': 0.39369391297805123, 'learning_rate/qwen_vl_interface': 3.1945415918751484e-06, 'learning_rate/action_model': 3.194541591875148e-05, 'epoch': 0.24}) +[2026-10-03 19:35:12,966][starVLA.training.train_starvla][INFO] - Step 3249, Loss: {'train/loss': 0.017505813390016556, 'train/grad_norm_pre_clip': 13.735958099365234, 'timing/data': 0.00021142000332474709, 'timing/model': 0.3954974630032666, 'learning_rate/qwen_vl_interface': 3.191796492508516e-06, 'learning_rate/action_model': 3.191796492508516e-05, 'epoch': 0.24}) +[2026-10-03 19:35:13,670][starVLA.training.train_starvla][INFO] - Step 3250, Loss: {'train/loss': 0.00012284213153179735, 'train/grad_norm_pre_clip': 0.04265643656253815, 'timing/data': 0.00029883001116104424, 'timing/model': 0.39455425401683897, 'learning_rate/qwen_vl_interface': 3.189052239191599e-06, 'learning_rate/action_model': 3.189052239191599e-05, 'epoch': 0.24}) +[2026-10-03 19:35:14,564][starVLA.training.train_starvla][INFO] - Step 3251, Loss: {'train/loss': 0.007273517083376646, 'train/grad_norm_pre_clip': 9.692584991455078, 'timing/data': 0.00032480998197570443, 'timing/model': 0.5849822260206565, 'learning_rate/qwen_vl_interface': 3.1863088330524577e-06, 'learning_rate/action_model': 3.186308833052458e-05, 'epoch': 0.24}) +[2026-10-03 19:35:15,270][starVLA.training.train_starvla][INFO] - Step 3252, Loss: {'train/loss': 0.16469886898994446, 'train/grad_norm_pre_clip': 78.5615463256836, 'timing/data': 0.0001965500123333186, 'timing/model': 0.39582571299979463, 'learning_rate/qwen_vl_interface': 3.183566275218803e-06, 'learning_rate/action_model': 3.1835662752188026e-05, 'epoch': 0.24}) +[2026-10-03 19:35:15,976][starVLA.training.train_starvla][INFO] - Step 3253, Loss: {'train/loss': 0.211532324552536, 'train/grad_norm_pre_clip': 171.09693908691406, 'timing/data': 0.0003443899913690984, 'timing/model': 0.3952925839985255, 'learning_rate/qwen_vl_interface': 3.180824566817996e-06, 'learning_rate/action_model': 3.1808245668179956e-05, 'epoch': 0.24}) +[2026-10-03 19:35:16,687][starVLA.training.train_starvla][INFO] - Step 3254, Loss: {'train/loss': 6.111462425906211e-05, 'train/grad_norm_pre_clip': 0.017064293846488, 'timing/data': 0.0003058500005863607, 'timing/model': 0.39733976300340146, 'learning_rate/qwen_vl_interface': 3.1780837089770524e-06, 'learning_rate/action_model': 3.1780837089770524e-05, 'epoch': 0.24}) +[2026-10-03 19:35:17,394][starVLA.training.train_starvla][INFO] - Step 3255, Loss: {'train/loss': 0.0004818551242351532, 'train/grad_norm_pre_clip': 0.14090560376644135, 'timing/data': 0.00037754001095891, 'timing/model': 0.39406027397490107, 'learning_rate/qwen_vl_interface': 3.17534370282263e-06, 'learning_rate/action_model': 3.1753437028226303e-05, 'epoch': 0.24}) +[2026-10-03 19:35:18,102][starVLA.training.train_starvla][INFO] - Step 3256, Loss: {'train/loss': 0.00031187961576506495, 'train/grad_norm_pre_clip': 0.1605706363916397, 'timing/data': 0.00027887002215720713, 'timing/model': 0.39466820398229174, 'learning_rate/qwen_vl_interface': 3.172604549481046e-06, 'learning_rate/action_model': 3.1726045494810455e-05, 'epoch': 0.24}) +[2026-10-03 19:35:18,996][starVLA.training.train_starvla][INFO] - Step 3257, Loss: {'train/loss': 0.0022902819328010082, 'train/grad_norm_pre_clip': 2.6423568725585938, 'timing/data': 0.0003186299873050302, 'timing/model': 0.3970789040031377, 'learning_rate/qwen_vl_interface': 3.169866250078261e-06, 'learning_rate/action_model': 3.169866250078261e-05, 'epoch': 0.24}) +[2026-10-03 19:35:19,705][starVLA.training.train_starvla][INFO] - Step 3258, Loss: {'train/loss': 0.0005854640039615333, 'train/grad_norm_pre_clip': 0.27109000086784363, 'timing/data': 0.0002601899905130267, 'timing/model': 0.3967391230107751, 'learning_rate/qwen_vl_interface': 3.167128805739887e-06, 'learning_rate/action_model': 3.1671288057398866e-05, 'epoch': 0.24}) +[2026-10-03 19:35:20,412][starVLA.training.train_starvla][INFO] - Step 3259, Loss: {'train/loss': 0.028144409880042076, 'train/grad_norm_pre_clip': 58.48466873168945, 'timing/data': 0.00024945000768639147, 'timing/model': 0.3949151440174319, 'learning_rate/qwen_vl_interface': 3.1643922175911847e-06, 'learning_rate/action_model': 3.164392217591185e-05, 'epoch': 0.24}) +[2026-10-03 19:35:21,120][starVLA.training.train_starvla][INFO] - Step 3260, Loss: {'train/loss': 8.059518586378545e-05, 'train/grad_norm_pre_clip': 0.03085452876985073, 'timing/data': 0.00019088000408373773, 'timing/model': 0.3951475939829834, 'learning_rate/qwen_vl_interface': 3.1616564867570603e-06, 'learning_rate/action_model': 3.16165648675706e-05, 'epoch': 0.24}) +[2026-10-03 19:35:21,826][starVLA.training.train_starvla][INFO] - Step 3261, Loss: {'train/loss': 0.04171648249030113, 'train/grad_norm_pre_clip': 63.16997146606445, 'timing/data': 0.00023869998403824866, 'timing/model': 0.394549613003619, 'learning_rate/qwen_vl_interface': 3.1589216143620733e-06, 'learning_rate/action_model': 3.1589216143620734e-05, 'epoch': 0.24}) +[2026-10-03 19:35:22,725][starVLA.training.train_starvla][INFO] - Step 3262, Loss: {'train/loss': 0.15487658977508545, 'train/grad_norm_pre_clip': 137.68121337890625, 'timing/data': 0.00028544000815600157, 'timing/model': 0.3973017029929906, 'learning_rate/qwen_vl_interface': 3.1561876015304215e-06, 'learning_rate/action_model': 3.156187601530421e-05, 'epoch': 0.24}) +[2026-10-03 19:35:23,433][starVLA.training.train_starvla][INFO] - Step 3263, Loss: {'train/loss': 0.0004550245066639036, 'train/grad_norm_pre_clip': 0.3045993447303772, 'timing/data': 0.0002238999877590686, 'timing/model': 0.39549322400125675, 'learning_rate/qwen_vl_interface': 3.1534544493859563e-06, 'learning_rate/action_model': 3.1534544493859566e-05, 'epoch': 0.24}) +[2026-10-03 19:35:24,143][starVLA.training.train_starvla][INFO] - Step 3264, Loss: {'train/loss': 0.00013039268378634006, 'train/grad_norm_pre_clip': 0.057766132056713104, 'timing/data': 0.00039977001142688096, 'timing/model': 0.3952099430025555, 'learning_rate/qwen_vl_interface': 3.150722159052174e-06, 'learning_rate/action_model': 3.150722159052174e-05, 'epoch': 0.24}) +[2026-10-03 19:35:24,851][starVLA.training.train_starvla][INFO] - Step 3265, Loss: {'train/loss': 0.2522502839565277, 'train/grad_norm_pre_clip': 113.08988952636719, 'timing/data': 0.0002993500092998147, 'timing/model': 0.39713715299149044, 'learning_rate/qwen_vl_interface': 3.147990731652215e-06, 'learning_rate/action_model': 3.147990731652215e-05, 'epoch': 0.24}) +[2026-10-03 19:35:25,560][starVLA.training.train_starvla][INFO] - Step 3266, Loss: {'train/loss': 0.0704311728477478, 'train/grad_norm_pre_clip': 179.4091339111328, 'timing/data': 0.0002600699954200536, 'timing/model': 0.39703139298944734, 'learning_rate/qwen_vl_interface': 3.1452601683088673e-06, 'learning_rate/action_model': 3.145260168308867e-05, 'epoch': 0.24}) +[2026-10-03 19:35:26,471][starVLA.training.train_starvla][INFO] - Step 3267, Loss: {'train/loss': 0.0003113322309218347, 'train/grad_norm_pre_clip': 0.21857604384422302, 'timing/data': 0.00021831999765709043, 'timing/model': 0.5967191649833694, 'learning_rate/qwen_vl_interface': 3.1425304701445608e-06, 'learning_rate/action_model': 3.1425304701445605e-05, 'epoch': 0.24}) +[2026-10-03 19:35:27,181][starVLA.training.train_starvla][INFO] - Step 3268, Loss: {'train/loss': 0.4604354798793793, 'train/grad_norm_pre_clip': 34.19560241699219, 'timing/data': 0.0003391699865460396, 'timing/model': 0.3972117629891727, 'learning_rate/qwen_vl_interface': 3.139801638281372e-06, 'learning_rate/action_model': 3.139801638281372e-05, 'epoch': 0.24}) +[2026-10-03 19:35:27,892][starVLA.training.train_starvla][INFO] - Step 3269, Loss: {'train/loss': 0.18621300160884857, 'train/grad_norm_pre_clip': 93.95358276367188, 'timing/data': 0.0003879800206050277, 'timing/model': 0.39853088298696093, 'learning_rate/qwen_vl_interface': 3.13707367384102e-06, 'learning_rate/action_model': 3.13707367384102e-05, 'epoch': 0.24}) +[2026-10-03 19:35:28,604][starVLA.training.train_starvla][INFO] - Step 3270, Loss: {'train/loss': 0.00010721480066422373, 'train/grad_norm_pre_clip': 0.08906003832817078, 'timing/data': 0.0003440100117586553, 'timing/model': 0.3964018639817368, 'learning_rate/qwen_vl_interface': 3.134346577944869e-06, 'learning_rate/action_model': 3.134346577944869e-05, 'epoch': 0.24}) +[2026-10-03 19:35:29,314][starVLA.training.train_starvla][INFO] - Step 3271, Loss: {'train/loss': 0.006179814226925373, 'train/grad_norm_pre_clip': 7.01702880859375, 'timing/data': 0.00033753999741747975, 'timing/model': 0.39687564299674705, 'learning_rate/qwen_vl_interface': 3.1316203517139243e-06, 'learning_rate/action_model': 3.131620351713924e-05, 'epoch': 0.24}) +[2026-10-03 19:35:30,025][starVLA.training.train_starvla][INFO] - Step 3272, Loss: {'train/loss': 0.00039485874003730714, 'train/grad_norm_pre_clip': 0.08047392964363098, 'timing/data': 0.0003969500248786062, 'timing/model': 0.39724527299404144, 'learning_rate/qwen_vl_interface': 3.128894996268837e-06, 'learning_rate/action_model': 3.128894996268837e-05, 'epoch': 0.24}) +[2026-10-03 19:35:30,934][starVLA.training.train_starvla][INFO] - Step 3273, Loss: {'train/loss': 6.902236782480031e-05, 'train/grad_norm_pre_clip': 0.021159613505005836, 'timing/data': 0.0003401400172151625, 'timing/model': 0.39735690399538726, 'learning_rate/qwen_vl_interface': 3.1261705127298934e-06, 'learning_rate/action_model': 3.1261705127298935e-05, 'epoch': 0.24}) +[2026-10-03 19:35:31,644][starVLA.training.train_starvla][INFO] - Step 3274, Loss: {'train/loss': 8.965706365415826e-05, 'train/grad_norm_pre_clip': 0.03439462557435036, 'timing/data': 0.00031284999568015337, 'timing/model': 0.3965631540049799, 'learning_rate/qwen_vl_interface': 3.123446902217028e-06, 'learning_rate/action_model': 3.1234469022170283e-05, 'epoch': 0.24}) +[2026-10-03 19:35:32,356][starVLA.training.train_starvla][INFO] - Step 3275, Loss: {'train/loss': 4.723290476249531e-05, 'train/grad_norm_pre_clip': 0.0216597281396389, 'timing/data': 0.00022524999803863466, 'timing/model': 0.396901323023485, 'learning_rate/qwen_vl_interface': 3.1207241658498145e-06, 'learning_rate/action_model': 3.1207241658498145e-05, 'epoch': 0.24}) +[2026-10-03 19:35:33,067][starVLA.training.train_starvla][INFO] - Step 3276, Loss: {'train/loss': 0.00014027136785443872, 'train/grad_norm_pre_clip': 0.052749909460544586, 'timing/data': 0.00021194000146351755, 'timing/model': 0.39739248299156316, 'learning_rate/qwen_vl_interface': 3.1180023047474667e-06, 'learning_rate/action_model': 3.1180023047474664e-05, 'epoch': 0.24}) +[2026-10-03 19:35:33,775][starVLA.training.train_starvla][INFO] - Step 3277, Loss: {'train/loss': 0.54069584608078, 'train/grad_norm_pre_clip': 146.83828735351562, 'timing/data': 0.0003371399943716824, 'timing/model': 0.3954019340162631, 'learning_rate/qwen_vl_interface': 3.1152813200288386e-06, 'learning_rate/action_model': 3.115281320028838e-05, 'epoch': 0.24}) +[2026-10-03 19:35:34,682][starVLA.training.train_starvla][INFO] - Step 3278, Loss: {'train/loss': 4.711230212706141e-05, 'train/grad_norm_pre_clip': 0.015675637871026993, 'timing/data': 0.00028039998142048717, 'timing/model': 0.3956081430078484, 'learning_rate/qwen_vl_interface': 3.1125612128124244e-06, 'learning_rate/action_model': 3.112561212812424e-05, 'epoch': 0.24}) +[2026-10-03 19:35:35,393][starVLA.training.train_starvla][INFO] - Step 3279, Loss: {'train/loss': 0.06854014098644257, 'train/grad_norm_pre_clip': 74.38471984863281, 'timing/data': 0.0004457600007299334, 'timing/model': 0.3964322830142919, 'learning_rate/qwen_vl_interface': 3.1098419842163573e-06, 'learning_rate/action_model': 3.109841984216357e-05, 'epoch': 0.24}) +[2026-10-03 19:35:36,102][starVLA.training.train_starvla][INFO] - Step 3280, Loss: {'train/loss': 0.004312367178499699, 'train/grad_norm_pre_clip': 4.981791973114014, 'timing/data': 0.00032010997529141605, 'timing/model': 0.3967638139729388, 'learning_rate/qwen_vl_interface': 3.10712363535841e-06, 'learning_rate/action_model': 3.10712363535841e-05, 'epoch': 0.24}) +[2026-10-03 19:35:36,816][starVLA.training.train_starvla][INFO] - Step 3281, Loss: {'train/loss': 0.34011468291282654, 'train/grad_norm_pre_clip': 164.3616485595703, 'timing/data': 0.0003254790208302438, 'timing/model': 0.3985071839997545, 'learning_rate/qwen_vl_interface': 3.1044061673559926e-06, 'learning_rate/action_model': 3.1044061673559925e-05, 'epoch': 0.24}) +[2026-10-03 19:35:37,528][starVLA.training.train_starvla][INFO] - Step 3282, Loss: {'train/loss': 0.0023059069644659758, 'train/grad_norm_pre_clip': 3.6048378944396973, 'timing/data': 0.0003654400061350316, 'timing/model': 0.3971668730082456, 'learning_rate/qwen_vl_interface': 3.1016895813261556e-06, 'learning_rate/action_model': 3.101689581326155e-05, 'epoch': 0.24}) +[2026-10-03 19:35:38,434][starVLA.training.train_starvla][INFO] - Step 3283, Loss: {'train/loss': 4.4495136535260826e-05, 'train/grad_norm_pre_clip': 0.017303714528679848, 'timing/data': 0.00032676002592779696, 'timing/model': 0.5924922149861231, 'learning_rate/qwen_vl_interface': 3.098973878385581e-06, 'learning_rate/action_model': 3.098973878385581e-05, 'epoch': 0.24}) +[2026-10-03 19:35:39,143][starVLA.training.train_starvla][INFO] - Step 3284, Loss: {'train/loss': 0.00013233294885139912, 'train/grad_norm_pre_clip': 0.07126947492361069, 'timing/data': 0.0002305499801877886, 'timing/model': 0.39693629200337455, 'learning_rate/qwen_vl_interface': 3.096259059650594e-06, 'learning_rate/action_model': 3.096259059650594e-05, 'epoch': 0.24}) +[2026-10-03 19:35:39,856][starVLA.training.train_starvla][INFO] - Step 3285, Loss: {'train/loss': 0.07333043217658997, 'train/grad_norm_pre_clip': 85.68441009521484, 'timing/data': 0.0003036099951714277, 'timing/model': 0.3986126729869284, 'learning_rate/qwen_vl_interface': 3.093545126237153e-06, 'learning_rate/action_model': 3.0935451262371527e-05, 'epoch': 0.24}) +[2026-10-03 19:35:40,567][starVLA.training.train_starvla][INFO] - Step 3286, Loss: {'train/loss': 0.0004155647475272417, 'train/grad_norm_pre_clip': 0.11784067749977112, 'timing/data': 0.0004437800089363009, 'timing/model': 0.39626649298588745, 'learning_rate/qwen_vl_interface': 3.090832079260855e-06, 'learning_rate/action_model': 3.0908320792608544e-05, 'epoch': 0.24}) +[2026-10-03 19:35:41,277][starVLA.training.train_starvla][INFO] - Step 3287, Loss: {'train/loss': 0.3610876798629761, 'train/grad_norm_pre_clip': 59.64076232910156, 'timing/data': 0.00035339000169187784, 'timing/model': 0.3964772829785943, 'learning_rate/qwen_vl_interface': 3.088119919836929e-06, 'learning_rate/action_model': 3.088119919836929e-05, 'epoch': 0.24}) +[2026-10-03 19:35:42,185][starVLA.training.train_starvla][INFO] - Step 3288, Loss: {'train/loss': 0.0011664601042866707, 'train/grad_norm_pre_clip': 2.2346432209014893, 'timing/data': 0.0002586999908089638, 'timing/model': 0.5941312549985014, 'learning_rate/qwen_vl_interface': 3.0854086490802433e-06, 'learning_rate/action_model': 3.085408649080243e-05, 'epoch': 0.24}) +[2026-10-03 19:35:42,896][starVLA.training.train_starvla][INFO] - Step 3289, Loss: {'train/loss': 0.13007162511348724, 'train/grad_norm_pre_clip': 117.53359985351562, 'timing/data': 0.0003406300093047321, 'timing/model': 0.39749487198423594, 'learning_rate/qwen_vl_interface': 3.082698268105299e-06, 'learning_rate/action_model': 3.082698268105299e-05, 'epoch': 0.24}) +[2026-10-03 19:35:43,604][starVLA.training.train_starvla][INFO] - Step 3290, Loss: {'train/loss': 5.5499964219052345e-05, 'train/grad_norm_pre_clip': 0.019050419330596924, 'timing/data': 0.0003219989885110408, 'timing/model': 0.39587105400278233, 'learning_rate/qwen_vl_interface': 3.0799887780262295e-06, 'learning_rate/action_model': 3.0799887780262296e-05, 'epoch': 0.24}) +[2026-10-03 19:35:44,312][starVLA.training.train_starvla][INFO] - Step 3291, Loss: {'train/loss': 0.00017796136671677232, 'train/grad_norm_pre_clip': 0.060247115790843964, 'timing/data': 0.00022029000683687627, 'timing/model': 0.3948923240241129, 'learning_rate/qwen_vl_interface': 3.0772801799568054e-06, 'learning_rate/action_model': 3.0772801799568054e-05, 'epoch': 0.24}) +[2026-10-03 19:35:45,021][starVLA.training.train_starvla][INFO] - Step 3292, Loss: {'train/loss': 7.007941167103127e-05, 'train/grad_norm_pre_clip': 0.023412013426423073, 'timing/data': 0.0003264099941588938, 'timing/model': 0.3965378330030944, 'learning_rate/qwen_vl_interface': 3.07457247501043e-06, 'learning_rate/action_model': 3.0745724750104296e-05, 'epoch': 0.24}) +[2026-10-03 19:35:45,931][starVLA.training.train_starvla][INFO] - Step 3293, Loss: {'train/loss': 0.7766827344894409, 'train/grad_norm_pre_clip': 253.50450134277344, 'timing/data': 0.00034813000820577145, 'timing/model': 0.5977121049945708, 'learning_rate/qwen_vl_interface': 3.0718656643001353e-06, 'learning_rate/action_model': 3.071865664300135e-05, 'epoch': 0.24}) +[2026-10-03 19:35:46,642][starVLA.training.train_starvla][INFO] - Step 3294, Loss: {'train/loss': 0.00014796876348555088, 'train/grad_norm_pre_clip': 0.04772622883319855, 'timing/data': 0.0003153200086671859, 'timing/model': 0.3962533930025529, 'learning_rate/qwen_vl_interface': 3.069159748938591e-06, 'learning_rate/action_model': 3.069159748938591e-05, 'epoch': 0.24}) +[2026-10-03 19:35:47,350][starVLA.training.train_starvla][INFO] - Step 3295, Loss: {'train/loss': 0.000924754305742681, 'train/grad_norm_pre_clip': 0.7996272444725037, 'timing/data': 0.00033773001632653177, 'timing/model': 0.3945966040191706, 'learning_rate/qwen_vl_interface': 3.066454730038096e-06, 'learning_rate/action_model': 3.0664547300380956e-05, 'epoch': 0.24}) +[2026-10-03 19:35:48,061][starVLA.training.train_starvla][INFO] - Step 3296, Loss: {'train/loss': 0.36365532875061035, 'train/grad_norm_pre_clip': 84.99129486083984, 'timing/data': 0.000284719979390502, 'timing/model': 0.3968775929824915, 'learning_rate/qwen_vl_interface': 3.0637506087105816e-06, 'learning_rate/action_model': 3.0637506087105815e-05, 'epoch': 0.24}) +[2026-10-03 19:35:48,772][starVLA.training.train_starvla][INFO] - Step 3297, Loss: {'train/loss': 0.004175823647528887, 'train/grad_norm_pre_clip': 6.5774126052856445, 'timing/data': 0.00022556001204065979, 'timing/model': 0.3981088830041699, 'learning_rate/qwen_vl_interface': 3.0610473860676103e-06, 'learning_rate/action_model': 3.06104738606761e-05, 'epoch': 0.24}) +[2026-10-03 19:35:49,485][starVLA.training.train_starvla][INFO] - Step 3298, Loss: {'train/loss': 0.0008379651117138565, 'train/grad_norm_pre_clip': 0.7409101724624634, 'timing/data': 0.00032246997579932213, 'timing/model': 0.3977251629985403, 'learning_rate/qwen_vl_interface': 3.0583450632203748e-06, 'learning_rate/action_model': 3.058345063220375e-05, 'epoch': 0.24}) +[2026-10-03 19:35:50,399][starVLA.training.train_starvla][INFO] - Step 3299, Loss: {'train/loss': 0.004069442395120859, 'train/grad_norm_pre_clip': 25.41741371154785, 'timing/data': 0.0002675800060387701, 'timing/model': 0.3932993430062197, 'learning_rate/qwen_vl_interface': 3.055643641279699e-06, 'learning_rate/action_model': 3.055643641279699e-05, 'epoch': 0.24}) +[2026-10-03 19:35:51,110][starVLA.training.train_starvla][INFO] - Step 3300, Loss: {'train/loss': 0.3473317325115204, 'train/grad_norm_pre_clip': 50.430747985839844, 'timing/data': 0.0002920800179708749, 'timing/model': 0.39676260299165733, 'learning_rate/qwen_vl_interface': 3.0529431213560334e-06, 'learning_rate/action_model': 3.0529431213560334e-05, 'epoch': 0.24}) +[2026-10-03 19:35:51,820][starVLA.training.train_starvla][INFO] - Step 3301, Loss: {'train/loss': 0.0002700348268263042, 'train/grad_norm_pre_clip': 0.24008242785930634, 'timing/data': 0.00032960998942144215, 'timing/model': 0.3971665630233474, 'learning_rate/qwen_vl_interface': 3.0502435045594596e-06, 'learning_rate/action_model': 3.0502435045594595e-05, 'epoch': 0.24}) +[2026-10-03 19:35:52,528][starVLA.training.train_starvla][INFO] - Step 3302, Loss: {'train/loss': 0.00020193519594613463, 'train/grad_norm_pre_clip': 0.09196139127016068, 'timing/data': 0.00030372999026440084, 'timing/model': 0.3958252629963681, 'learning_rate/qwen_vl_interface': 3.047544791999689e-06, 'learning_rate/action_model': 3.0475447919996892e-05, 'epoch': 0.24}) +[2026-10-03 19:35:53,236][starVLA.training.train_starvla][INFO] - Step 3303, Loss: {'train/loss': 9.329533349955454e-05, 'train/grad_norm_pre_clip': 0.034760817885398865, 'timing/data': 0.00021428000763989985, 'timing/model': 0.39662391401361674, 'learning_rate/qwen_vl_interface': 3.044846984786061e-06, 'learning_rate/action_model': 3.044846984786061e-05, 'epoch': 0.24}) +[2026-10-03 19:35:54,142][starVLA.training.train_starvla][INFO] - Step 3304, Loss: {'train/loss': 0.00403933459892869, 'train/grad_norm_pre_clip': 16.303796768188477, 'timing/data': 0.00020689002121798694, 'timing/model': 0.5918291839770973, 'learning_rate/qwen_vl_interface': 3.0421500840275413e-06, 'learning_rate/action_model': 3.0421500840275413e-05, 'epoch': 0.24}) +[2026-10-03 19:35:54,852][starVLA.training.train_starvla][INFO] - Step 3305, Loss: {'train/loss': 4.238258406985551e-05, 'train/grad_norm_pre_clip': 0.024667903780937195, 'timing/data': 0.0003555000002961606, 'timing/model': 0.3964098739961628, 'learning_rate/qwen_vl_interface': 3.039454090832727e-06, 'learning_rate/action_model': 3.0394540908327268e-05, 'epoch': 0.24}) +[2026-10-03 19:35:55,564][starVLA.training.train_starvla][INFO] - Step 3306, Loss: {'train/loss': 0.001901280484162271, 'train/grad_norm_pre_clip': 1.5470390319824219, 'timing/data': 0.00021282999659888446, 'timing/model': 0.3978437439945992, 'learning_rate/qwen_vl_interface': 3.0367590063098333e-06, 'learning_rate/action_model': 3.0367590063098333e-05, 'epoch': 0.24}) +[2026-10-03 19:35:56,274][starVLA.training.train_starvla][INFO] - Step 3307, Loss: {'train/loss': 0.5562531352043152, 'train/grad_norm_pre_clip': 46.63765335083008, 'timing/data': 0.0003556600131560117, 'timing/model': 0.39654594301828183, 'learning_rate/qwen_vl_interface': 3.0340648315667115e-06, 'learning_rate/action_model': 3.0340648315667113e-05, 'epoch': 0.24}) +[2026-10-03 19:35:56,982][starVLA.training.train_starvla][INFO] - Step 3308, Loss: {'train/loss': 0.0001471435825806111, 'train/grad_norm_pre_clip': 0.10914618521928787, 'timing/data': 0.00036094000097364187, 'timing/model': 0.3961512829991989, 'learning_rate/qwen_vl_interface': 3.0313715677108336e-06, 'learning_rate/action_model': 3.0313715677108335e-05, 'epoch': 0.24}) +[2026-10-03 19:35:57,693][starVLA.training.train_starvla][INFO] - Step 3309, Loss: {'train/loss': 7.768448267597705e-05, 'train/grad_norm_pre_clip': 0.02317381650209427, 'timing/data': 0.000352060014847666, 'timing/model': 0.3975268529902678, 'learning_rate/qwen_vl_interface': 3.0286792158492996e-06, 'learning_rate/action_model': 3.0286792158492994e-05, 'epoch': 0.24}) +[2026-10-03 19:35:58,599][starVLA.training.train_starvla][INFO] - Step 3310, Loss: {'train/loss': 0.0002761297218967229, 'train/grad_norm_pre_clip': 0.14027003943920135, 'timing/data': 0.00027752999449148774, 'timing/model': 0.3963159330014605, 'learning_rate/qwen_vl_interface': 3.0259877770888334e-06, 'learning_rate/action_model': 3.0259877770888333e-05, 'epoch': 0.24}) +[2026-10-03 19:35:59,310][starVLA.training.train_starvla][INFO] - Step 3311, Loss: {'train/loss': 0.0003705398121383041, 'train/grad_norm_pre_clip': 0.33985424041748047, 'timing/data': 0.00022980000358074903, 'timing/model': 0.39653733401792124, 'learning_rate/qwen_vl_interface': 3.0232972525357836e-06, 'learning_rate/action_model': 3.0232972525357834e-05, 'epoch': 0.24}) +[2026-10-03 19:36:00,020][starVLA.training.train_starvla][INFO] - Step 3312, Loss: {'train/loss': 5.263331695459783e-05, 'train/grad_norm_pre_clip': 0.01349370926618576, 'timing/data': 0.00033599999733269215, 'timing/model': 0.39509376301430166, 'learning_rate/qwen_vl_interface': 3.020607643296123e-06, 'learning_rate/action_model': 3.0206076432961226e-05, 'epoch': 0.24}) +[2026-10-03 19:36:00,729][starVLA.training.train_starvla][INFO] - Step 3313, Loss: {'train/loss': 2.343881715205498e-05, 'train/grad_norm_pre_clip': 0.007101358845829964, 'timing/data': 0.0003265999839641154, 'timing/model': 0.3969692629761994, 'learning_rate/qwen_vl_interface': 3.017918950475449e-06, 'learning_rate/action_model': 3.017918950475449e-05, 'epoch': 0.24}) +[2026-10-03 19:36:01,438][starVLA.training.train_starvla][INFO] - Step 3314, Loss: {'train/loss': 0.0024169995449483395, 'train/grad_norm_pre_clip': 12.44542407989502, 'timing/data': 0.00026515001081861556, 'timing/model': 0.3957241339958273, 'learning_rate/qwen_vl_interface': 3.0152311751789815e-06, 'learning_rate/action_model': 3.0152311751789814e-05, 'epoch': 0.24}) +[2026-10-03 19:36:02,345][starVLA.training.train_starvla][INFO] - Step 3315, Loss: {'train/loss': 1.564632534980774, 'train/grad_norm_pre_clip': 146.94529724121094, 'timing/data': 0.00021773998741991818, 'timing/model': 0.39372831300715916, 'learning_rate/qwen_vl_interface': 3.0125443185115655e-06, 'learning_rate/action_model': 3.0125443185115655e-05, 'epoch': 0.24}) +[2026-10-03 19:36:03,047][starVLA.training.train_starvla][INFO] - Step 3316, Loss: {'train/loss': 0.5019740462303162, 'train/grad_norm_pre_clip': 52.6799201965332, 'timing/data': 0.00021569000091403723, 'timing/model': 0.3919500140182208, 'learning_rate/qwen_vl_interface': 3.0098583815776615e-06, 'learning_rate/action_model': 3.0098583815776615e-05, 'epoch': 0.24}) +[2026-10-03 19:36:03,749][starVLA.training.train_starvla][INFO] - Step 3317, Loss: {'train/loss': 8.771780994720757e-05, 'train/grad_norm_pre_clip': 0.0462309792637825, 'timing/data': 0.00020345000666566193, 'timing/model': 0.393581003008876, 'learning_rate/qwen_vl_interface': 3.0071733654813595e-06, 'learning_rate/action_model': 3.0071733654813595e-05, 'epoch': 0.24}) +[2026-10-03 19:36:04,449][starVLA.training.train_starvla][INFO] - Step 3318, Loss: {'train/loss': 0.00012424694432411343, 'train/grad_norm_pre_clip': 0.03238794207572937, 'timing/data': 0.00020735000725835562, 'timing/model': 0.3912228439876344, 'learning_rate/qwen_vl_interface': 3.0044892713263685e-06, 'learning_rate/action_model': 3.0044892713263682e-05, 'epoch': 0.24}) +[2026-10-03 19:36:05,147][starVLA.training.train_starvla][INFO] - Step 3319, Loss: {'train/loss': 0.0009751708712428808, 'train/grad_norm_pre_clip': 0.9748944640159607, 'timing/data': 0.00020359997870400548, 'timing/model': 0.39115977397887036, 'learning_rate/qwen_vl_interface': 3.001806100216017e-06, 'learning_rate/action_model': 3.001806100216017e-05, 'epoch': 0.24}) +[2026-10-03 19:36:06,056][starVLA.training.train_starvla][INFO] - Step 3320, Loss: {'train/loss': 0.0003683876129798591, 'train/grad_norm_pre_clip': 0.19999800622463226, 'timing/data': 0.00021232999279163778, 'timing/model': 0.6006220239796676, 'learning_rate/qwen_vl_interface': 2.9991238532532567e-06, 'learning_rate/action_model': 2.9991238532532567e-05, 'epoch': 0.24}) +[2026-10-03 19:36:06,758][starVLA.training.train_starvla][INFO] - Step 3321, Loss: {'train/loss': 4.641332634491846e-05, 'train/grad_norm_pre_clip': 0.021498562768101692, 'timing/data': 0.0002384109830018133, 'timing/model': 0.3939958230184857, 'learning_rate/qwen_vl_interface': 2.996442531540657e-06, 'learning_rate/action_model': 2.9964425315406568e-05, 'epoch': 0.24}) +[2026-10-03 19:36:07,461][starVLA.training.train_starvla][INFO] - Step 3322, Loss: {'train/loss': 0.39866629242897034, 'train/grad_norm_pre_clip': 48.39597702026367, 'timing/data': 0.0002034700009971857, 'timing/model': 0.3928530330013018, 'learning_rate/qwen_vl_interface': 2.993762136180408e-06, 'learning_rate/action_model': 2.993762136180408e-05, 'epoch': 0.24}) +[2026-10-03 19:36:08,158][starVLA.training.train_starvla][INFO] - Step 3323, Loss: {'train/loss': 0.7932736277580261, 'train/grad_norm_pre_clip': 75.93868255615234, 'timing/data': 0.00020511000184342265, 'timing/model': 0.3908457839861512, 'learning_rate/qwen_vl_interface': 2.9910826682743195e-06, 'learning_rate/action_model': 2.9910826682743192e-05, 'epoch': 0.24}) +[2026-10-03 19:36:08,859][starVLA.training.train_starvla][INFO] - Step 3324, Loss: {'train/loss': 0.1530323475599289, 'train/grad_norm_pre_clip': 60.65876770019531, 'timing/data': 0.00033683000947348773, 'timing/model': 0.3913463839853648, 'learning_rate/qwen_vl_interface': 2.9884041289238192e-06, 'learning_rate/action_model': 2.988404128923819e-05, 'epoch': 0.24}) +[2026-10-03 19:36:09,556][starVLA.training.train_starvla][INFO] - Step 3325, Loss: {'train/loss': 0.6042513251304626, 'train/grad_norm_pre_clip': 71.15082550048828, 'timing/data': 0.00030676997266709805, 'timing/model': 0.3912016130052507, 'learning_rate/qwen_vl_interface': 2.9857265192299557e-06, 'learning_rate/action_model': 2.9857265192299556e-05, 'epoch': 0.24}) +[2026-10-03 19:36:10,456][starVLA.training.train_starvla][INFO] - Step 3326, Loss: {'train/loss': 0.31873273849487305, 'train/grad_norm_pre_clip': 152.6055450439453, 'timing/data': 0.00024228999973274767, 'timing/model': 0.3913590040174313, 'learning_rate/qwen_vl_interface': 2.9830498402933894e-06, 'learning_rate/action_model': 2.9830498402933894e-05, 'epoch': 0.24}) +[2026-10-03 19:36:11,156][starVLA.training.train_starvla][INFO] - Step 3327, Loss: {'train/loss': 0.001263098674826324, 'train/grad_norm_pre_clip': 2.4515230655670166, 'timing/data': 0.0002124900056514889, 'timing/model': 0.3914621239819098, 'learning_rate/qwen_vl_interface': 2.9803740932144025e-06, 'learning_rate/action_model': 2.9803740932144024e-05, 'epoch': 0.24}) +[2026-10-03 19:36:11,856][starVLA.training.train_starvla][INFO] - Step 3328, Loss: {'train/loss': 2.0987681637052447e-05, 'train/grad_norm_pre_clip': 0.005924217868596315, 'timing/data': 0.0003489099908620119, 'timing/model': 0.39121999399503693, 'learning_rate/qwen_vl_interface': 2.9776992790928953e-06, 'learning_rate/action_model': 2.9776992790928952e-05, 'epoch': 0.24}) +[2026-10-03 19:36:12,556][starVLA.training.train_starvla][INFO] - Step 3329, Loss: {'train/loss': 2.705206316022668e-05, 'train/grad_norm_pre_clip': 0.007122999057173729, 'timing/data': 0.0002940900158137083, 'timing/model': 0.392551054013893, 'learning_rate/qwen_vl_interface': 2.975025399028382e-06, 'learning_rate/action_model': 2.975025399028382e-05, 'epoch': 0.24}) +[2026-10-03 19:36:13,256][starVLA.training.train_starvla][INFO] - Step 3330, Loss: {'train/loss': 5.609802610706538e-05, 'train/grad_norm_pre_clip': 0.020469505339860916, 'timing/data': 0.0003411700017750263, 'timing/model': 0.39148627399117686, 'learning_rate/qwen_vl_interface': 2.9723524541199943e-06, 'learning_rate/action_model': 2.972352454119994e-05, 'epoch': 0.24}) +[2026-10-03 19:36:14,159][starVLA.training.train_starvla][INFO] - Step 3331, Loss: {'train/loss': 0.16200488805770874, 'train/grad_norm_pre_clip': 126.12921905517578, 'timing/data': 0.00036489000194706023, 'timing/model': 0.38915646402165294, 'learning_rate/qwen_vl_interface': 2.969680445466478e-06, 'learning_rate/action_model': 2.969680445466478e-05, 'epoch': 0.24}) +[2026-10-03 19:36:14,862][starVLA.training.train_starvla][INFO] - Step 3332, Loss: {'train/loss': 0.0011080215917900205, 'train/grad_norm_pre_clip': 1.5515450239181519, 'timing/data': 0.0003683399991132319, 'timing/model': 0.3926901940139942, 'learning_rate/qwen_vl_interface': 2.9670093741661955e-06, 'learning_rate/action_model': 2.9670093741661954e-05, 'epoch': 0.24}) +[2026-10-03 19:36:15,562][starVLA.training.train_starvla][INFO] - Step 3333, Loss: {'train/loss': 0.001326155150309205, 'train/grad_norm_pre_clip': 0.6558641791343689, 'timing/data': 0.0003177500038873404, 'timing/model': 0.39210158301284537, 'learning_rate/qwen_vl_interface': 2.964339241317123e-06, 'learning_rate/action_model': 2.964339241317123e-05, 'epoch': 0.24}) +[2026-10-03 19:36:16,266][starVLA.training.train_starvla][INFO] - Step 3334, Loss: {'train/loss': 0.0121116042137146, 'train/grad_norm_pre_clip': 15.097990036010742, 'timing/data': 0.0003674499748740345, 'timing/model': 0.3934000129811466, 'learning_rate/qwen_vl_interface': 2.961670048016852e-06, 'learning_rate/action_model': 2.9616700480168518e-05, 'epoch': 0.24}) +[2026-10-03 19:36:16,969][starVLA.training.train_starvla][INFO] - Step 3335, Loss: {'train/loss': 0.10773623734712601, 'train/grad_norm_pre_clip': 144.8065948486328, 'timing/data': 0.00030958998831920326, 'timing/model': 0.39309704399784096, 'learning_rate/qwen_vl_interface': 2.9590017953625854e-06, 'learning_rate/action_model': 2.9590017953625853e-05, 'epoch': 0.24}) +[2026-10-03 19:36:17,869][starVLA.training.train_starvla][INFO] - Step 3336, Loss: {'train/loss': 2.1099578589200974e-05, 'train/grad_norm_pre_clip': 0.005044218618422747, 'timing/data': 0.00020549001055769622, 'timing/model': 0.5899789750110358, 'learning_rate/qwen_vl_interface': 2.9563344844511455e-06, 'learning_rate/action_model': 2.9563344844511455e-05, 'epoch': 0.24}) +[2026-10-03 19:36:18,572][starVLA.training.train_starvla][INFO] - Step 3337, Loss: {'train/loss': 0.00013468829274643213, 'train/grad_norm_pre_clip': 0.07475944608449936, 'timing/data': 0.00022258999524638057, 'timing/model': 0.3947400630277116, 'learning_rate/qwen_vl_interface': 2.9536681163789568e-06, 'learning_rate/action_model': 2.9536681163789564e-05, 'epoch': 0.24}) +[2026-10-03 19:36:19,269][starVLA.training.train_starvla][INFO] - Step 3338, Loss: {'train/loss': 0.4018368124961853, 'train/grad_norm_pre_clip': 117.2143783569336, 'timing/data': 0.0002417299838270992, 'timing/model': 0.38976193399867043, 'learning_rate/qwen_vl_interface': 2.9510026922420653e-06, 'learning_rate/action_model': 2.9510026922420653e-05, 'epoch': 0.24}) +[2026-10-03 19:36:19,968][starVLA.training.train_starvla][INFO] - Step 3339, Loss: {'train/loss': 0.2840629816055298, 'train/grad_norm_pre_clip': 47.607425689697266, 'timing/data': 0.00020494998898357153, 'timing/model': 0.3921330840094015, 'learning_rate/qwen_vl_interface': 2.9483382131361273e-06, 'learning_rate/action_model': 2.948338213136127e-05, 'epoch': 0.24}) +[2026-10-03 19:36:20,672][starVLA.training.train_starvla][INFO] - Step 3340, Loss: {'train/loss': 0.10025637596845627, 'train/grad_norm_pre_clip': 214.75296020507812, 'timing/data': 0.00043969997204840183, 'timing/model': 0.39413549401797354, 'learning_rate/qwen_vl_interface': 2.9456746801564072e-06, 'learning_rate/action_model': 2.945674680156407e-05, 'epoch': 0.24}) +[2026-10-03 19:36:21,577][starVLA.training.train_starvla][INFO] - Step 3341, Loss: {'train/loss': 2.1576361177721992e-05, 'train/grad_norm_pre_clip': 0.007124042604118586, 'timing/data': 0.00030067001353017986, 'timing/model': 0.5942386849783361, 'learning_rate/qwen_vl_interface': 2.9430120943977854e-06, 'learning_rate/action_model': 2.9430120943977853e-05, 'epoch': 0.24}) +[2026-10-03 19:36:22,278][starVLA.training.train_starvla][INFO] - Step 3342, Loss: {'train/loss': 0.0005176024860702455, 'train/grad_norm_pre_clip': 0.4826301336288452, 'timing/data': 0.00020527999731712043, 'timing/model': 0.3915620239858981, 'learning_rate/qwen_vl_interface': 2.940350456954748e-06, 'learning_rate/action_model': 2.940350456954748e-05, 'epoch': 0.24}) +[2026-10-03 19:36:22,981][starVLA.training.train_starvla][INFO] - Step 3343, Loss: {'train/loss': 7.341748278122395e-05, 'train/grad_norm_pre_clip': 0.024657154455780983, 'timing/data': 0.0002489000034984201, 'timing/model': 0.39305388400680386, 'learning_rate/qwen_vl_interface': 2.9376897689213952e-06, 'learning_rate/action_model': 2.937689768921395e-05, 'epoch': 0.24}) +[2026-10-03 19:36:23,683][starVLA.training.train_starvla][INFO] - Step 3344, Loss: {'train/loss': 0.08717517554759979, 'train/grad_norm_pre_clip': 88.07820892333984, 'timing/data': 0.0002081099955830723, 'timing/model': 0.391621643997496, 'learning_rate/qwen_vl_interface': 2.9350300313914364e-06, 'learning_rate/action_model': 2.9350300313914363e-05, 'epoch': 0.24}) +[2026-10-03 19:36:24,386][starVLA.training.train_starvla][INFO] - Step 3345, Loss: {'train/loss': 0.2403298169374466, 'train/grad_norm_pre_clip': 117.21605682373047, 'timing/data': 0.00024285001563839614, 'timing/model': 0.3925068039970938, 'learning_rate/qwen_vl_interface': 2.9323712454581867e-06, 'learning_rate/action_model': 2.9323712454581866e-05, 'epoch': 0.24}) +[2026-10-03 19:36:25,289][starVLA.training.train_starvla][INFO] - Step 3346, Loss: {'train/loss': 0.5010045170783997, 'train/grad_norm_pre_clip': 174.45626831054688, 'timing/data': 0.00033991000964306295, 'timing/model': 0.5948205050081015, 'learning_rate/qwen_vl_interface': 2.929713412214572e-06, 'learning_rate/action_model': 2.929713412214572e-05, 'epoch': 0.24}) +[2026-10-03 19:36:25,989][starVLA.training.train_starvla][INFO] - Step 3347, Loss: {'train/loss': 0.000269200507318601, 'train/grad_norm_pre_clip': 0.49841177463531494, 'timing/data': 0.00027716998010873795, 'timing/model': 0.3918825739820022, 'learning_rate/qwen_vl_interface': 2.9270565327531294e-06, 'learning_rate/action_model': 2.927056532753129e-05, 'epoch': 0.24}) +[2026-10-03 19:36:26,696][starVLA.training.train_starvla][INFO] - Step 3348, Loss: {'train/loss': 0.0011610998772084713, 'train/grad_norm_pre_clip': 0.7891661524772644, 'timing/data': 0.00025169001310132444, 'timing/model': 0.396090483001899, 'learning_rate/qwen_vl_interface': 2.924400608166003e-06, 'learning_rate/action_model': 2.924400608166003e-05, 'epoch': 0.25}) +[2026-10-03 19:36:27,398][starVLA.training.train_starvla][INFO] - Step 3349, Loss: {'train/loss': 0.011424504220485687, 'train/grad_norm_pre_clip': 29.545061111450195, 'timing/data': 0.0003580900083761662, 'timing/model': 0.39271042399923317, 'learning_rate/qwen_vl_interface': 2.921745639544939e-06, 'learning_rate/action_model': 2.921745639544939e-05, 'epoch': 0.25}) +[2026-10-03 19:36:28,100][starVLA.training.train_starvla][INFO] - Step 3350, Loss: {'train/loss': 0.21475033462047577, 'train/grad_norm_pre_clip': 64.76932525634766, 'timing/data': 0.0003517800068948418, 'timing/model': 0.39111131298705004, 'learning_rate/qwen_vl_interface': 2.919091627981297e-06, 'learning_rate/action_model': 2.919091627981297e-05, 'epoch': 0.25}) +[2026-10-03 19:36:28,799][starVLA.training.train_starvla][INFO] - Step 3351, Loss: {'train/loss': 0.0015702113742008805, 'train/grad_norm_pre_clip': 2.1605582237243652, 'timing/data': 0.0003077699802815914, 'timing/model': 0.39141749299596995, 'learning_rate/qwen_vl_interface': 2.9164385745660396e-06, 'learning_rate/action_model': 2.9164385745660394e-05, 'epoch': 0.25}) +[2026-10-03 19:36:29,709][starVLA.training.train_starvla][INFO] - Step 3352, Loss: {'train/loss': 0.04835965111851692, 'train/grad_norm_pre_clip': 51.47692108154297, 'timing/data': 0.0003549599787220359, 'timing/model': 0.39069333401857875, 'learning_rate/qwen_vl_interface': 2.9137864803897385e-06, 'learning_rate/action_model': 2.9137864803897387e-05, 'epoch': 0.25}) +[2026-10-03 19:36:30,406][starVLA.training.train_starvla][INFO] - Step 3353, Loss: {'train/loss': 0.0007519841310568154, 'train/grad_norm_pre_clip': 0.9911649227142334, 'timing/data': 0.00021182000637054443, 'timing/model': 0.3912661240028683, 'learning_rate/qwen_vl_interface': 2.9111353465425696e-06, 'learning_rate/action_model': 2.9111353465425694e-05, 'epoch': 0.25}) +[2026-10-03 19:36:31,103][starVLA.training.train_starvla][INFO] - Step 3354, Loss: {'train/loss': 0.0011786336544901133, 'train/grad_norm_pre_clip': 3.483661413192749, 'timing/data': 0.0003429999924264848, 'timing/model': 0.3901623839919921, 'learning_rate/qwen_vl_interface': 2.908485174114312e-06, 'learning_rate/action_model': 2.908485174114312e-05, 'epoch': 0.25}) +[2026-10-03 19:36:31,802][starVLA.training.train_starvla][INFO] - Step 3355, Loss: {'train/loss': 0.00014050303434487432, 'train/grad_norm_pre_clip': 0.039076823741197586, 'timing/data': 0.00030051000067032874, 'timing/model': 0.3905563139996957, 'learning_rate/qwen_vl_interface': 2.9058359641943523e-06, 'learning_rate/action_model': 2.9058359641943523e-05, 'epoch': 0.25}) +[2026-10-03 19:36:32,502][starVLA.training.train_starvla][INFO] - Step 3356, Loss: {'train/loss': 0.002591958036646247, 'train/grad_norm_pre_clip': 4.4548492431640625, 'timing/data': 0.00033475999953225255, 'timing/model': 0.39139651399455033, 'learning_rate/qwen_vl_interface': 2.9031877178716823e-06, 'learning_rate/action_model': 2.903187717871682e-05, 'epoch': 0.25}) +[2026-10-03 19:36:33,401][starVLA.training.train_starvla][INFO] - Step 3357, Loss: {'train/loss': 0.011878982186317444, 'train/grad_norm_pre_clip': 28.106727600097656, 'timing/data': 0.00021805998403578997, 'timing/model': 0.5917254660162143, 'learning_rate/qwen_vl_interface': 2.900540436234894e-06, 'learning_rate/action_model': 2.9005404362348937e-05, 'epoch': 0.25}) +[2026-10-03 19:36:34,101][starVLA.training.train_starvla][INFO] - Step 3358, Loss: {'train/loss': 0.0003213999734725803, 'train/grad_norm_pre_clip': 0.07933378964662552, 'timing/data': 0.0003311200125608593, 'timing/model': 0.3909093840047717, 'learning_rate/qwen_vl_interface': 2.897894120372185e-06, 'learning_rate/action_model': 2.897894120372185e-05, 'epoch': 0.25}) +[2026-10-03 19:36:34,800][starVLA.training.train_starvla][INFO] - Step 3359, Loss: {'train/loss': 0.0018830600893124938, 'train/grad_norm_pre_clip': 0.6672521829605103, 'timing/data': 0.00031463001505471766, 'timing/model': 0.38978679399588145, 'learning_rate/qwen_vl_interface': 2.8952487713713585e-06, 'learning_rate/action_model': 2.8952487713713582e-05, 'epoch': 0.25}) +[2026-10-03 19:36:35,500][starVLA.training.train_starvla][INFO] - Step 3360, Loss: {'train/loss': 0.011754072271287441, 'train/grad_norm_pre_clip': 10.062440872192383, 'timing/data': 0.00020914000924676657, 'timing/model': 0.3904788439976983, 'learning_rate/qwen_vl_interface': 2.8926043903198135e-06, 'learning_rate/action_model': 2.8926043903198135e-05, 'epoch': 0.25}) +[2026-10-03 19:36:36,199][starVLA.training.train_starvla][INFO] - Step 3361, Loss: {'train/loss': 0.30171918869018555, 'train/grad_norm_pre_clip': 69.06161499023438, 'timing/data': 0.0002012300246860832, 'timing/model': 0.3929929540026933, 'learning_rate/qwen_vl_interface': 2.8899609783045574e-06, 'learning_rate/action_model': 2.8899609783045573e-05, 'epoch': 0.25}) +[2026-10-03 19:36:36,900][starVLA.training.train_starvla][INFO] - Step 3362, Loss: {'train/loss': 0.22499607503414154, 'train/grad_norm_pre_clip': 190.83831787109375, 'timing/data': 0.00034935001167468727, 'timing/model': 0.39159117301460356, 'learning_rate/qwen_vl_interface': 2.8873185364121965e-06, 'learning_rate/action_model': 2.8873185364121963e-05, 'epoch': 0.25}) +[2026-10-03 19:36:37,798][starVLA.training.train_starvla][INFO] - Step 3363, Loss: {'train/loss': 0.00037335383240133524, 'train/grad_norm_pre_clip': 0.0914447158575058, 'timing/data': 0.0004817000008188188, 'timing/model': 0.3920935729984194, 'learning_rate/qwen_vl_interface': 2.8846770657289386e-06, 'learning_rate/action_model': 2.8846770657289386e-05, 'epoch': 0.25}) +[2026-10-03 19:36:38,496][starVLA.training.train_starvla][INFO] - Step 3364, Loss: {'train/loss': 0.00027019804110750556, 'train/grad_norm_pre_clip': 0.12486831098794937, 'timing/data': 0.0003545889921952039, 'timing/model': 0.39059505399200134, 'learning_rate/qwen_vl_interface': 2.882036567340593e-06, 'learning_rate/action_model': 2.8820365673405932e-05, 'epoch': 0.25}) +[2026-10-03 19:36:39,199][starVLA.training.train_starvla][INFO] - Step 3365, Loss: {'train/loss': 0.09725847095251083, 'train/grad_norm_pre_clip': 133.2859649658203, 'timing/data': 0.00020392000442370772, 'timing/model': 0.3946488529909402, 'learning_rate/qwen_vl_interface': 2.8793970423325684e-06, 'learning_rate/action_model': 2.8793970423325685e-05, 'epoch': 0.25}) +[2026-10-03 19:36:39,899][starVLA.training.train_starvla][INFO] - Step 3366, Loss: {'train/loss': 0.041637782007455826, 'train/grad_norm_pre_clip': 125.73297119140625, 'timing/data': 0.00020552999922074378, 'timing/model': 0.39176979300100356, 'learning_rate/qwen_vl_interface': 2.8767584917898756e-06, 'learning_rate/action_model': 2.8767584917898754e-05, 'epoch': 0.25}) +[2026-10-03 19:36:40,600][starVLA.training.train_starvla][INFO] - Step 3367, Loss: {'train/loss': 0.0006437054253183305, 'train/grad_norm_pre_clip': 2.5859882831573486, 'timing/data': 0.0003023899917025119, 'timing/model': 0.3925700429826975, 'learning_rate/qwen_vl_interface': 2.874120916797121e-06, 'learning_rate/action_model': 2.8741209167971207e-05, 'epoch': 0.25}) +[2026-10-03 19:36:41,504][starVLA.training.train_starvla][INFO] - Step 3368, Loss: {'train/loss': 0.003694788785651326, 'train/grad_norm_pre_clip': 13.869385719299316, 'timing/data': 0.0002094999945256859, 'timing/model': 0.39144655398558825, 'learning_rate/qwen_vl_interface': 2.8714843184385124e-06, 'learning_rate/action_model': 2.871484318438512e-05, 'epoch': 0.25}) +[2026-10-03 19:36:42,205][starVLA.training.train_starvla][INFO] - Step 3369, Loss: {'train/loss': 0.18951500952243805, 'train/grad_norm_pre_clip': 84.05878448486328, 'timing/data': 0.0002532700018491596, 'timing/model': 0.39368444398860447, 'learning_rate/qwen_vl_interface': 2.8688486977978575e-06, 'learning_rate/action_model': 2.8688486977978572e-05, 'epoch': 0.25}) +[2026-10-03 19:36:42,906][starVLA.training.train_starvla][INFO] - Step 3370, Loss: {'train/loss': 0.07261485606431961, 'train/grad_norm_pre_clip': 168.55087280273438, 'timing/data': 0.00021139002637937665, 'timing/model': 0.3923486529965885, 'learning_rate/qwen_vl_interface': 2.8662140559585584e-06, 'learning_rate/action_model': 2.866214055958558e-05, 'epoch': 0.25}) +[2026-10-03 19:36:43,604][starVLA.training.train_starvla][INFO] - Step 3371, Loss: {'train/loss': 0.012071099132299423, 'train/grad_norm_pre_clip': 19.858572006225586, 'timing/data': 0.00031689999741502106, 'timing/model': 0.3916266330052167, 'learning_rate/qwen_vl_interface': 2.8635803940036162e-06, 'learning_rate/action_model': 2.863580394003616e-05, 'epoch': 0.25}) +[2026-10-03 19:36:44,303][starVLA.training.train_starvla][INFO] - Step 3372, Loss: {'train/loss': 0.03412491828203201, 'train/grad_norm_pre_clip': 33.503639221191406, 'timing/data': 0.0002039399987552315, 'timing/model': 0.39193082400015555, 'learning_rate/qwen_vl_interface': 2.8609477130156317e-06, 'learning_rate/action_model': 2.8609477130156314e-05, 'epoch': 0.25}) +[2026-10-03 19:36:45,202][starVLA.training.train_starvla][INFO] - Step 3373, Loss: {'train/loss': 0.27767714858055115, 'train/grad_norm_pre_clip': 112.56383514404297, 'timing/data': 0.000295709993224591, 'timing/model': 0.5916891359956935, 'learning_rate/qwen_vl_interface': 2.8583160140767997e-06, 'learning_rate/action_model': 2.8583160140767994e-05, 'epoch': 0.25}) +[2026-10-03 19:36:45,902][starVLA.training.train_starvla][INFO] - Step 3374, Loss: {'train/loss': 0.016852593049407005, 'train/grad_norm_pre_clip': 24.72663688659668, 'timing/data': 0.0003423400048632175, 'timing/model': 0.3906907539931126, 'learning_rate/qwen_vl_interface': 2.8556852982689125e-06, 'learning_rate/action_model': 2.8556852982689126e-05, 'epoch': 0.25}) +[2026-10-03 19:36:46,602][starVLA.training.train_starvla][INFO] - Step 3375, Loss: {'train/loss': 0.2670447826385498, 'train/grad_norm_pre_clip': 65.2386703491211, 'timing/data': 0.00033518997952342033, 'timing/model': 0.3922423530020751, 'learning_rate/qwen_vl_interface': 2.8530555666733584e-06, 'learning_rate/action_model': 2.8530555666733583e-05, 'epoch': 0.25}) +[2026-10-03 19:36:47,302][starVLA.training.train_starvla][INFO] - Step 3376, Loss: {'train/loss': 0.0023022356908768415, 'train/grad_norm_pre_clip': 5.105158805847168, 'timing/data': 0.0003060300077777356, 'timing/model': 0.3907454439904541, 'learning_rate/qwen_vl_interface': 2.85042682037112e-06, 'learning_rate/action_model': 2.85042682037112e-05, 'epoch': 0.25}) +[2026-10-03 19:36:48,005][starVLA.training.train_starvla][INFO] - Step 3377, Loss: {'train/loss': 0.00014017622743267566, 'train/grad_norm_pre_clip': 0.22316984832286835, 'timing/data': 0.00021028000628575683, 'timing/model': 0.39339481299975887, 'learning_rate/qwen_vl_interface': 2.8477990604427764e-06, 'learning_rate/action_model': 2.8477990604427763e-05, 'epoch': 0.25}) +[2026-10-03 19:36:48,709][starVLA.training.train_starvla][INFO] - Step 3378, Loss: {'train/loss': 0.5284356474876404, 'train/grad_norm_pre_clip': 163.70681762695312, 'timing/data': 0.00020460999803617597, 'timing/model': 0.3933380130038131, 'learning_rate/qwen_vl_interface': 2.8451722879684994e-06, 'learning_rate/action_model': 2.8451722879684994e-05, 'epoch': 0.25}) +[2026-10-03 19:36:49,616][starVLA.training.train_starvla][INFO] - Step 3379, Loss: {'train/loss': 9.700547707325313e-06, 'train/grad_norm_pre_clip': 0.002789006568491459, 'timing/data': 0.0002067199966404587, 'timing/model': 0.3915825930016581, 'learning_rate/qwen_vl_interface': 2.842546504028059e-06, 'learning_rate/action_model': 2.8425465040280586e-05, 'epoch': 0.25}) +[2026-10-03 19:36:50,313][starVLA.training.train_starvla][INFO] - Step 3380, Loss: {'train/loss': 0.00023468371364288032, 'train/grad_norm_pre_clip': 0.05887357518076897, 'timing/data': 0.0004203599819447845, 'timing/model': 0.3896284630172886, 'learning_rate/qwen_vl_interface': 2.83992170970081e-06, 'learning_rate/action_model': 2.8399217097008102e-05, 'epoch': 0.25}) +[2026-10-03 19:36:51,013][starVLA.training.train_starvla][INFO] - Step 3381, Loss: {'train/loss': 0.00010743118764366955, 'train/grad_norm_pre_clip': 0.05750624090433121, 'timing/data': 0.00026169000193476677, 'timing/model': 0.39298093400429934, 'learning_rate/qwen_vl_interface': 2.837297906065711e-06, 'learning_rate/action_model': 2.8372979060657107e-05, 'epoch': 0.25}) +[2026-10-03 19:36:51,715][starVLA.training.train_starvla][INFO] - Step 3382, Loss: {'train/loss': 0.24928447604179382, 'train/grad_norm_pre_clip': 70.1532974243164, 'timing/data': 0.00021257001208141446, 'timing/model': 0.3944147330184933, 'learning_rate/qwen_vl_interface': 2.834675094201306e-06, 'learning_rate/action_model': 2.8346750942013063e-05, 'epoch': 0.25}) +[2026-10-03 19:36:52,415][starVLA.training.train_starvla][INFO] - Step 3383, Loss: {'train/loss': 0.004158978350460529, 'train/grad_norm_pre_clip': 4.409975051879883, 'timing/data': 0.00021148001542314887, 'timing/model': 0.39138574301614426, 'learning_rate/qwen_vl_interface': 2.8320532751857365e-06, 'learning_rate/action_model': 2.8320532751857366e-05, 'epoch': 0.25}) +[2026-10-03 19:36:53,314][starVLA.training.train_starvla][INFO] - Step 3384, Loss: {'train/loss': 0.02136143110692501, 'train/grad_norm_pre_clip': 63.698265075683594, 'timing/data': 0.00034391001099720597, 'timing/model': 0.39089203299954534, 'learning_rate/qwen_vl_interface': 2.829432450096731e-06, 'learning_rate/action_model': 2.8294324500967306e-05, 'epoch': 0.25}) +[2026-10-03 19:36:54,014][starVLA.training.train_starvla][INFO] - Step 3385, Loss: {'train/loss': 3.572200148482807e-05, 'train/grad_norm_pre_clip': 0.025121619924902916, 'timing/data': 0.00026570999762043357, 'timing/model': 0.392365012987284, 'learning_rate/qwen_vl_interface': 2.8268126200116118e-06, 'learning_rate/action_model': 2.8268126200116117e-05, 'epoch': 0.25}) +[2026-10-03 19:36:54,713][starVLA.training.train_starvla][INFO] - Step 3386, Loss: {'train/loss': 0.18523527681827545, 'train/grad_norm_pre_clip': 64.42594146728516, 'timing/data': 0.0002027600130531937, 'timing/model': 0.39066800399450585, 'learning_rate/qwen_vl_interface': 2.824193786007294e-06, 'learning_rate/action_model': 2.8241937860072937e-05, 'epoch': 0.25}) +[2026-10-03 19:36:55,412][starVLA.training.train_starvla][INFO] - Step 3387, Loss: {'train/loss': 1.636113302083686e-05, 'train/grad_norm_pre_clip': 0.005585698876529932, 'timing/data': 0.0003546799998730421, 'timing/model': 0.39161108300322667, 'learning_rate/qwen_vl_interface': 2.821575949160279e-06, 'learning_rate/action_model': 2.8215759491602787e-05, 'epoch': 0.25}) +[2026-10-03 19:36:56,112][starVLA.training.train_starvla][INFO] - Step 3388, Loss: {'train/loss': 0.00014249078230932355, 'train/grad_norm_pre_clip': 0.08258774131536484, 'timing/data': 0.0003181799838785082, 'timing/model': 0.3898769529769197, 'learning_rate/qwen_vl_interface': 2.8189591105466625e-06, 'learning_rate/action_model': 2.8189591105466624e-05, 'epoch': 0.25}) +[2026-10-03 19:36:57,019][starVLA.training.train_starvla][INFO] - Step 3389, Loss: {'train/loss': 4.1102495742961764e-05, 'train/grad_norm_pre_clip': 0.012430660426616669, 'timing/data': 0.00024761000531725585, 'timing/model': 0.5980961640016176, 'learning_rate/qwen_vl_interface': 2.816343271242126e-06, 'learning_rate/action_model': 2.816343271242126e-05, 'epoch': 0.25}) +[2026-10-03 19:36:57,720][starVLA.training.train_starvla][INFO] - Step 3390, Loss: {'train/loss': 1.9922004867112264e-05, 'train/grad_norm_pre_clip': 0.007343010511249304, 'timing/data': 0.0002741199859883636, 'timing/model': 0.3926458730129525, 'learning_rate/qwen_vl_interface': 2.8137284323219467e-06, 'learning_rate/action_model': 2.8137284323219464e-05, 'epoch': 0.25}) +[2026-10-03 19:36:58,417][starVLA.training.train_starvla][INFO] - Step 3391, Loss: {'train/loss': 0.00032794199069030583, 'train/grad_norm_pre_clip': 0.4251875579357147, 'timing/data': 0.00024705001851543784, 'timing/model': 0.3895980539964512, 'learning_rate/qwen_vl_interface': 2.81111459486098e-06, 'learning_rate/action_model': 2.81111459486098e-05, 'epoch': 0.25}) +[2026-10-03 19:36:59,115][starVLA.training.train_starvla][INFO] - Step 3392, Loss: {'train/loss': 0.006193012464791536, 'train/grad_norm_pre_clip': 13.83568286895752, 'timing/data': 0.00021482998272404075, 'timing/model': 0.39126905298326164, 'learning_rate/qwen_vl_interface': 2.8085017599336806e-06, 'learning_rate/action_model': 2.8085017599336806e-05, 'epoch': 0.25}) +[2026-10-03 19:36:59,814][starVLA.training.train_starvla][INFO] - Step 3393, Loss: {'train/loss': 0.09082315117120743, 'train/grad_norm_pre_clip': 230.1256561279297, 'timing/data': 0.00029083999106660485, 'timing/model': 0.3925428130023647, 'learning_rate/qwen_vl_interface': 2.8058899286140828e-06, 'learning_rate/action_model': 2.805889928614083e-05, 'epoch': 0.25}) +[2026-10-03 19:37:00,718][starVLA.training.train_starvla][INFO] - Step 3394, Loss: {'train/loss': 0.4240908920764923, 'train/grad_norm_pre_clip': 257.031005859375, 'timing/data': 0.000348429981386289, 'timing/model': 0.5984262450074311, 'learning_rate/qwen_vl_interface': 2.8032791019758127e-06, 'learning_rate/action_model': 2.8032791019758127e-05, 'epoch': 0.25}) +[2026-10-03 19:37:01,416][starVLA.training.train_starvla][INFO] - Step 3395, Loss: {'train/loss': 0.002363244304433465, 'train/grad_norm_pre_clip': 9.888726234436035, 'timing/data': 0.0002976300020236522, 'timing/model': 0.3888614229799714, 'learning_rate/qwen_vl_interface': 2.800669281092084e-06, 'learning_rate/action_model': 2.8006692810920838e-05, 'epoch': 0.25}) +[2026-10-03 19:37:02,112][starVLA.training.train_starvla][INFO] - Step 3396, Loss: {'train/loss': 0.0005351351574063301, 'train/grad_norm_pre_clip': 0.15438230335712433, 'timing/data': 0.00020834000315517187, 'timing/model': 0.38898809399688616, 'learning_rate/qwen_vl_interface': 2.7980604670356938e-06, 'learning_rate/action_model': 2.7980604670356937e-05, 'epoch': 0.25}) +[2026-10-03 19:37:02,810][starVLA.training.train_starvla][INFO] - Step 3397, Loss: {'train/loss': 0.10353562980890274, 'train/grad_norm_pre_clip': 70.0645980834961, 'timing/data': 0.00033286999678239226, 'timing/model': 0.390338402998168, 'learning_rate/qwen_vl_interface': 2.795452660879027e-06, 'learning_rate/action_model': 2.7954526608790273e-05, 'epoch': 0.25}) +[2026-10-03 19:37:03,510][starVLA.training.train_starvla][INFO] - Step 3398, Loss: {'train/loss': 6.69730652589351e-05, 'train/grad_norm_pre_clip': 0.026675749570131302, 'timing/data': 0.00032359000761061907, 'timing/model': 0.3926927830034401, 'learning_rate/qwen_vl_interface': 2.7928458636940565e-06, 'learning_rate/action_model': 2.7928458636940564e-05, 'epoch': 0.25}) +[2026-10-03 19:37:04,207][starVLA.training.train_starvla][INFO] - Step 3399, Loss: {'train/loss': 0.0041366443037986755, 'train/grad_norm_pre_clip': 8.29647445678711, 'timing/data': 0.00028984100208617747, 'timing/model': 0.39113873298629187, 'learning_rate/qwen_vl_interface': 2.790240076552336e-06, 'learning_rate/action_model': 2.7902400765523355e-05, 'epoch': 0.25}) +[2026-10-03 19:37:05,102][starVLA.training.train_starvla][INFO] - Step 3400, Loss: {'train/loss': 0.4694366455078125, 'train/grad_norm_pre_clip': 210.96884155273438, 'timing/data': 0.0004746000049635768, 'timing/model': 0.39013228300609626, 'learning_rate/qwen_vl_interface': 2.787635300525006e-06, 'learning_rate/action_model': 2.787635300525006e-05, 'epoch': 0.25}) +[2026-10-03 19:37:05,803][starVLA.training.train_starvla][INFO] - Step 3401, Loss: {'train/loss': 0.3495701849460602, 'train/grad_norm_pre_clip': 67.30322265625, 'timing/data': 0.0003035699774045497, 'timing/model': 0.39264511299552396, 'learning_rate/qwen_vl_interface': 2.7850315366827933e-06, 'learning_rate/action_model': 2.7850315366827934e-05, 'epoch': 0.25}) +[2026-10-03 19:37:06,503][starVLA.training.train_starvla][INFO] - Step 3402, Loss: {'train/loss': 1.882698416011408e-05, 'train/grad_norm_pre_clip': 0.005656288005411625, 'timing/data': 0.00036037000245414674, 'timing/model': 0.391214642993873, 'learning_rate/qwen_vl_interface': 2.7824287860960076e-06, 'learning_rate/action_model': 2.7824287860960075e-05, 'epoch': 0.25}) +[2026-10-03 19:37:07,201][starVLA.training.train_starvla][INFO] - Step 3403, Loss: {'train/loss': 8.629041258245707e-05, 'train/grad_norm_pre_clip': 0.0876721665263176, 'timing/data': 0.0002961900027003139, 'timing/model': 0.3900816229870543, 'learning_rate/qwen_vl_interface': 2.7798270498345377e-06, 'learning_rate/action_model': 2.7798270498345376e-05, 'epoch': 0.25}) +[2026-10-03 19:37:07,898][starVLA.training.train_starvla][INFO] - Step 3404, Loss: {'train/loss': 0.02828369103372097, 'train/grad_norm_pre_clip': 58.066341400146484, 'timing/data': 0.00020733001292683184, 'timing/model': 0.3909276629856322, 'learning_rate/qwen_vl_interface': 2.777226328967861e-06, 'learning_rate/action_model': 2.7772263289678607e-05, 'epoch': 0.25}) +[2026-10-03 19:37:08,793][starVLA.training.train_starvla][INFO] - Step 3405, Loss: {'train/loss': 3.1610383302904665e-05, 'train/grad_norm_pre_clip': 0.011635148897767067, 'timing/data': 0.00031360998400487006, 'timing/model': 0.3920617740077432, 'learning_rate/qwen_vl_interface': 2.774626624565035e-06, 'learning_rate/action_model': 2.774626624565035e-05, 'epoch': 0.25}) +[2026-10-03 19:37:09,490][starVLA.training.train_starvla][INFO] - Step 3406, Loss: {'train/loss': 0.00278270710259676, 'train/grad_norm_pre_clip': 1.6577414274215698, 'timing/data': 0.0003050600062124431, 'timing/model': 0.3906456940167118, 'learning_rate/qwen_vl_interface': 2.772027937694701e-06, 'learning_rate/action_model': 2.772027937694701e-05, 'epoch': 0.25}) +[2026-10-03 19:37:10,187][starVLA.training.train_starvla][INFO] - Step 3407, Loss: {'train/loss': 0.0030208101961761713, 'train/grad_norm_pre_clip': 2.355656862258911, 'timing/data': 0.00032689000363461673, 'timing/model': 0.3897805730230175, 'learning_rate/qwen_vl_interface': 2.7694302694250806e-06, 'learning_rate/action_model': 2.7694302694250808e-05, 'epoch': 0.25}) +[2026-10-03 19:37:10,886][starVLA.training.train_starvla][INFO] - Step 3408, Loss: {'train/loss': 0.0003376050153747201, 'train/grad_norm_pre_clip': 0.308054655790329, 'timing/data': 0.00031869998201727867, 'timing/model': 0.39100410300306976, 'learning_rate/qwen_vl_interface': 2.7668336208239775e-06, 'learning_rate/action_model': 2.7668336208239776e-05, 'epoch': 0.25}) +[2026-10-03 19:37:11,589][starVLA.training.train_starvla][INFO] - Step 3409, Loss: {'train/loss': 0.0056796809658408165, 'train/grad_norm_pre_clip': 10.687383651733398, 'timing/data': 0.00025099999038502574, 'timing/model': 0.3933110029902309, 'learning_rate/qwen_vl_interface': 2.7642379929587755e-06, 'learning_rate/action_model': 2.7642379929587754e-05, 'epoch': 0.25}) +[2026-10-03 19:37:12,492][starVLA.training.train_starvla][INFO] - Step 3410, Loss: {'train/loss': 0.0007923313532955945, 'train/grad_norm_pre_clip': 0.8001490831375122, 'timing/data': 0.0002031800104305148, 'timing/model': 0.59483493398875, 'learning_rate/qwen_vl_interface': 2.7616433868964387e-06, 'learning_rate/action_model': 2.7616433868964388e-05, 'epoch': 0.25}) +[2026-10-03 19:37:13,189][starVLA.training.train_starvla][INFO] - Step 3411, Loss: {'train/loss': 0.442354291677475, 'train/grad_norm_pre_clip': 59.47430419921875, 'timing/data': 0.0003313300258014351, 'timing/model': 0.3900626940012444, 'learning_rate/qwen_vl_interface': 2.7590498037035122e-06, 'learning_rate/action_model': 2.7590498037035124e-05, 'epoch': 0.25}) +[2026-10-03 19:37:13,890][starVLA.training.train_starvla][INFO] - Step 3412, Loss: {'train/loss': 0.22117140889167786, 'train/grad_norm_pre_clip': 85.15574645996094, 'timing/data': 0.00033536000410094857, 'timing/model': 0.39176384298480116, 'learning_rate/qwen_vl_interface': 2.7564572444461218e-06, 'learning_rate/action_model': 2.7564572444461217e-05, 'epoch': 0.25}) +[2026-10-03 19:37:14,592][starVLA.training.train_starvla][INFO] - Step 3413, Loss: {'train/loss': 4.741828524856828e-05, 'train/grad_norm_pre_clip': 0.023293662816286087, 'timing/data': 0.00029783998616039753, 'timing/model': 0.3935870029963553, 'learning_rate/qwen_vl_interface': 2.7538657101899674e-06, 'learning_rate/action_model': 2.7538657101899673e-05, 'epoch': 0.25}) +[2026-10-03 19:37:15,292][starVLA.training.train_starvla][INFO] - Step 3414, Loss: {'train/loss': 0.01581541635096073, 'train/grad_norm_pre_clip': 45.622581481933594, 'timing/data': 0.0003404799790587276, 'timing/model': 0.3909327930014115, 'learning_rate/qwen_vl_interface': 2.751275202000332e-06, 'learning_rate/action_model': 2.7512752020003317e-05, 'epoch': 0.25}) +[2026-10-03 19:37:16,193][starVLA.training.train_starvla][INFO] - Step 3415, Loss: {'train/loss': 6.91679451847449e-05, 'train/grad_norm_pre_clip': 0.04219247028231621, 'timing/data': 0.00032909002038650215, 'timing/model': 0.5941338739939965, 'learning_rate/qwen_vl_interface': 2.7486857209420763e-06, 'learning_rate/action_model': 2.7486857209420763e-05, 'epoch': 0.25}) +[2026-10-03 19:37:16,892][starVLA.training.train_starvla][INFO] - Step 3416, Loss: {'train/loss': 0.6259296536445618, 'train/grad_norm_pre_clip': 34.171566009521484, 'timing/data': 0.00021487998310476542, 'timing/model': 0.3909496240084991, 'learning_rate/qwen_vl_interface': 2.7460972680796377e-06, 'learning_rate/action_model': 2.7460972680796375e-05, 'epoch': 0.25}) +[2026-10-03 19:37:17,593][starVLA.training.train_starvla][INFO] - Step 3417, Loss: {'train/loss': 0.202296182513237, 'train/grad_norm_pre_clip': 45.189910888671875, 'timing/data': 0.00029654998797923326, 'timing/model': 0.3936803330143448, 'learning_rate/qwen_vl_interface': 2.7435098444770314e-06, 'learning_rate/action_model': 2.7435098444770315e-05, 'epoch': 0.25}) +[2026-10-03 19:37:18,290][starVLA.training.train_starvla][INFO] - Step 3418, Loss: {'train/loss': 0.4200262427330017, 'train/grad_norm_pre_clip': 46.255252838134766, 'timing/data': 0.00020214900723658502, 'timing/model': 0.3900538140151184, 'learning_rate/qwen_vl_interface': 2.7409234511978495e-06, 'learning_rate/action_model': 2.7409234511978492e-05, 'epoch': 0.25}) +[2026-10-03 19:37:18,989][starVLA.training.train_starvla][INFO] - Step 3419, Loss: {'train/loss': 0.0005584932514466345, 'train/grad_norm_pre_clip': 0.8177355527877808, 'timing/data': 0.00021066999761387706, 'timing/model': 0.3917603029985912, 'learning_rate/qwen_vl_interface': 2.7383380893052606e-06, 'learning_rate/action_model': 2.7383380893052608e-05, 'epoch': 0.25}) +[2026-10-03 19:37:19,689][starVLA.training.train_starvla][INFO] - Step 3420, Loss: {'train/loss': 0.00020181199943181127, 'train/grad_norm_pre_clip': 0.11188945919275284, 'timing/data': 0.00027523000608198345, 'timing/model': 0.3911174129752908, 'learning_rate/qwen_vl_interface': 2.73575375986201e-06, 'learning_rate/action_model': 2.7357537598620096e-05, 'epoch': 0.25}) +[2026-10-03 19:37:20,586][starVLA.training.train_starvla][INFO] - Step 3421, Loss: {'train/loss': 9.819743354455568e-06, 'train/grad_norm_pre_clip': 0.002660615835338831, 'timing/data': 0.00032102002296596766, 'timing/model': 0.3921233139990363, 'learning_rate/qwen_vl_interface': 2.733170463930417e-06, 'learning_rate/action_model': 2.7331704639304167e-05, 'epoch': 0.25}) +[2026-10-03 19:37:21,285][starVLA.training.train_starvla][INFO] - Step 3422, Loss: {'train/loss': 5.5571228585904464e-05, 'train/grad_norm_pre_clip': 0.029592519626021385, 'timing/data': 0.0003347799938637763, 'timing/model': 0.39245201300946064, 'learning_rate/qwen_vl_interface': 2.7305882025723774e-06, 'learning_rate/action_model': 2.7305882025723774e-05, 'epoch': 0.25}) +[2026-10-03 19:37:21,984][starVLA.training.train_starvla][INFO] - Step 3423, Loss: {'train/loss': 0.4892693758010864, 'train/grad_norm_pre_clip': 114.72773742675781, 'timing/data': 0.0002141399891115725, 'timing/model': 0.3921216339804232, 'learning_rate/qwen_vl_interface': 2.7280069768493626e-06, 'learning_rate/action_model': 2.7280069768493628e-05, 'epoch': 0.25}) +[2026-10-03 19:37:22,684][starVLA.training.train_starvla][INFO] - Step 3424, Loss: {'train/loss': 5.867153959115967e-05, 'train/grad_norm_pre_clip': 0.035578370094299316, 'timing/data': 0.00021729999571107328, 'timing/model': 0.392313052987447, 'learning_rate/qwen_vl_interface': 2.7254267878224144e-06, 'learning_rate/action_model': 2.7254267878224143e-05, 'epoch': 0.25}) +[2026-10-03 19:37:23,383][starVLA.training.train_starvla][INFO] - Step 3425, Loss: {'train/loss': 0.2673960030078888, 'train/grad_norm_pre_clip': 129.33140563964844, 'timing/data': 0.00024135998683050275, 'timing/model': 0.3926272139942739, 'learning_rate/qwen_vl_interface': 2.7228476365521524e-06, 'learning_rate/action_model': 2.722847636552152e-05, 'epoch': 0.25}) +[2026-10-03 19:37:24,283][starVLA.training.train_starvla][INFO] - Step 3426, Loss: {'train/loss': 0.00019739857816603035, 'train/grad_norm_pre_clip': 0.1426706612110138, 'timing/data': 0.000301049993140623, 'timing/model': 0.39164245399297215, 'learning_rate/qwen_vl_interface': 2.7202695240987674e-06, 'learning_rate/action_model': 2.7202695240987676e-05, 'epoch': 0.25}) +[2026-10-03 19:37:24,984][starVLA.training.train_starvla][INFO] - Step 3427, Loss: {'train/loss': 0.768399715423584, 'train/grad_norm_pre_clip': 196.703857421875, 'timing/data': 0.00022630998864769936, 'timing/model': 0.3920270729868207, 'learning_rate/qwen_vl_interface': 2.717692451522025e-06, 'learning_rate/action_model': 2.7176924515220248e-05, 'epoch': 0.25}) +[2026-10-03 19:37:25,684][starVLA.training.train_starvla][INFO] - Step 3428, Loss: {'train/loss': 2.326690810150467e-05, 'train/grad_norm_pre_clip': 0.014051830396056175, 'timing/data': 0.00021425000159069896, 'timing/model': 0.3906516639981419, 'learning_rate/qwen_vl_interface': 2.7151164198812624e-06, 'learning_rate/action_model': 2.7151164198812622e-05, 'epoch': 0.25}) +[2026-10-03 19:37:26,382][starVLA.training.train_starvla][INFO] - Step 3429, Loss: {'train/loss': 3.9880691474536434e-05, 'train/grad_norm_pre_clip': 0.012961658649146557, 'timing/data': 0.0003127899835817516, 'timing/model': 0.39021659301943146, 'learning_rate/qwen_vl_interface': 2.712541430235388e-06, 'learning_rate/action_model': 2.712541430235388e-05, 'epoch': 0.25}) +[2026-10-03 19:37:27,081][starVLA.training.train_starvla][INFO] - Step 3430, Loss: {'train/loss': 0.0008419531513936818, 'train/grad_norm_pre_clip': 0.8108813166618347, 'timing/data': 0.0003632199950516224, 'timing/model': 0.39177776299766265, 'learning_rate/qwen_vl_interface': 2.7099674836428842e-06, 'learning_rate/action_model': 2.709967483642884e-05, 'epoch': 0.25}) +[2026-10-03 19:37:27,980][starVLA.training.train_starvla][INFO] - Step 3431, Loss: {'train/loss': 0.2274002581834793, 'train/grad_norm_pre_clip': 83.70455169677734, 'timing/data': 0.00020619999850168824, 'timing/model': 0.5912669849931262, 'learning_rate/qwen_vl_interface': 2.7073945811618018e-06, 'learning_rate/action_model': 2.7073945811618017e-05, 'epoch': 0.25}) +[2026-10-03 19:37:28,679][starVLA.training.train_starvla][INFO] - Step 3432, Loss: {'train/loss': 0.31126952171325684, 'train/grad_norm_pre_clip': 175.6354217529297, 'timing/data': 0.00020870001753792167, 'timing/model': 0.390645212988602, 'learning_rate/qwen_vl_interface': 2.704822723849765e-06, 'learning_rate/action_model': 2.7048227238497648e-05, 'epoch': 0.25}) +[2026-10-03 19:37:29,377][starVLA.training.train_starvla][INFO] - Step 3433, Loss: {'train/loss': 2.0615387256839313e-05, 'train/grad_norm_pre_clip': 0.0055666700936853886, 'timing/data': 0.00030651999986730516, 'timing/model': 0.39165872300509363, 'learning_rate/qwen_vl_interface': 2.7022519127639687e-06, 'learning_rate/action_model': 2.7022519127639682e-05, 'epoch': 0.25}) +[2026-10-03 19:37:30,075][starVLA.training.train_starvla][INFO] - Step 3434, Loss: {'train/loss': 0.2636394202709198, 'train/grad_norm_pre_clip': 71.79949188232422, 'timing/data': 0.0002704600046854466, 'timing/model': 0.3908903539995663, 'learning_rate/qwen_vl_interface': 2.6996821489611713e-06, 'learning_rate/action_model': 2.699682148961171e-05, 'epoch': 0.25}) +[2026-10-03 19:37:30,774][starVLA.training.train_starvla][INFO] - Step 3435, Loss: {'train/loss': 0.0001138927400461398, 'train/grad_norm_pre_clip': 0.05176302790641785, 'timing/data': 0.0002396400086581707, 'timing/model': 0.3913709330081474, 'learning_rate/qwen_vl_interface': 2.697113433497712e-06, 'learning_rate/action_model': 2.6971134334977115e-05, 'epoch': 0.25}) +[2026-10-03 19:37:31,677][starVLA.training.train_starvla][INFO] - Step 3436, Loss: {'train/loss': 1.0236948583042249e-05, 'train/grad_norm_pre_clip': 0.002482163952663541, 'timing/data': 0.0003353790089022368, 'timing/model': 0.5932583250105381, 'learning_rate/qwen_vl_interface': 2.6945457674294863e-06, 'learning_rate/action_model': 2.694545767429486e-05, 'epoch': 0.25}) +[2026-10-03 19:37:32,377][starVLA.training.train_starvla][INFO] - Step 3437, Loss: {'train/loss': 5.66606831853278e-05, 'train/grad_norm_pre_clip': 0.03127032890915871, 'timing/data': 0.0002060899860225618, 'timing/model': 0.39273511301144026, 'learning_rate/qwen_vl_interface': 2.6919791518119676e-06, 'learning_rate/action_model': 2.6919791518119674e-05, 'epoch': 0.25}) +[2026-10-03 19:37:33,075][starVLA.training.train_starvla][INFO] - Step 3438, Loss: {'train/loss': 0.0016918466426432133, 'train/grad_norm_pre_clip': 1.9313827753067017, 'timing/data': 0.00035806000232696533, 'timing/model': 0.38904140298836865, 'learning_rate/qwen_vl_interface': 2.6894135877001946e-06, 'learning_rate/action_model': 2.689413587700195e-05, 'epoch': 0.25}) +[2026-10-03 19:37:33,776][starVLA.training.train_starvla][INFO] - Step 3439, Loss: {'train/loss': 0.3458592891693115, 'train/grad_norm_pre_clip': 39.528411865234375, 'timing/data': 0.0003357600071467459, 'timing/model': 0.39143272300134413, 'learning_rate/qwen_vl_interface': 2.686849076148775e-06, 'learning_rate/action_model': 2.6868490761487747e-05, 'epoch': 0.25}) +[2026-10-03 19:37:34,478][starVLA.training.train_starvla][INFO] - Step 3440, Loss: {'train/loss': 0.0001352257968392223, 'train/grad_norm_pre_clip': 0.028890207409858704, 'timing/data': 0.000334870012011379, 'timing/model': 0.39156067301519215, 'learning_rate/qwen_vl_interface': 2.68428561821188e-06, 'learning_rate/action_model': 2.6842856182118802e-05, 'epoch': 0.25}) +[2026-10-03 19:37:35,179][starVLA.training.train_starvla][INFO] - Step 3441, Loss: {'train/loss': 0.005182633642107248, 'train/grad_norm_pre_clip': 7.889349460601807, 'timing/data': 0.0003185099922120571, 'timing/model': 0.3935601329721976, 'learning_rate/qwen_vl_interface': 2.681723214943253e-06, 'learning_rate/action_model': 2.681723214943253e-05, 'epoch': 0.25}) +[2026-10-03 19:37:36,080][starVLA.training.train_starvla][INFO] - Step 3442, Loss: {'train/loss': 0.3064940869808197, 'train/grad_norm_pre_clip': 165.91148376464844, 'timing/data': 0.00025498998002149165, 'timing/model': 0.391264723992208, 'learning_rate/qwen_vl_interface': 2.6791618673961996e-06, 'learning_rate/action_model': 2.6791618673961994e-05, 'epoch': 0.25}) +[2026-10-03 19:37:36,779][starVLA.training.train_starvla][INFO] - Step 3443, Loss: {'train/loss': 0.00018129586533177644, 'train/grad_norm_pre_clip': 0.15530729293823242, 'timing/data': 0.00030071998480707407, 'timing/model': 0.3906396040110849, 'learning_rate/qwen_vl_interface': 2.676601576623594e-06, 'learning_rate/action_model': 2.6766015766235937e-05, 'epoch': 0.25}) +[2026-10-03 19:37:37,478][starVLA.training.train_starvla][INFO] - Step 3444, Loss: {'train/loss': 0.00018136115977540612, 'train/grad_norm_pre_clip': 0.05256832763552666, 'timing/data': 0.00033879000693559647, 'timing/model': 0.3904536339978222, 'learning_rate/qwen_vl_interface': 2.674042343677874e-06, 'learning_rate/action_model': 2.6740423436778738e-05, 'epoch': 0.25}) +[2026-10-03 19:37:38,179][starVLA.training.train_starvla][INFO] - Step 3445, Loss: {'train/loss': 0.7014300227165222, 'train/grad_norm_pre_clip': 144.1423797607422, 'timing/data': 0.0003236799966543913, 'timing/model': 0.3930133630055934, 'learning_rate/qwen_vl_interface': 2.6714841696110456e-06, 'learning_rate/action_model': 2.6714841696110456e-05, 'epoch': 0.25}) +[2026-10-03 19:37:38,878][starVLA.training.train_starvla][INFO] - Step 3446, Loss: {'train/loss': 0.0006275714840739965, 'train/grad_norm_pre_clip': 0.4841780662536621, 'timing/data': 0.00036176000139676034, 'timing/model': 0.39144602298620157, 'learning_rate/qwen_vl_interface': 2.6689270554746743e-06, 'learning_rate/action_model': 2.6689270554746743e-05, 'epoch': 0.25}) +[2026-10-03 19:37:39,776][starVLA.training.train_starvla][INFO] - Step 3447, Loss: {'train/loss': 6.535254215123132e-05, 'train/grad_norm_pre_clip': 0.04316559061408043, 'timing/data': 0.0002047099987976253, 'timing/model': 0.5897778349753935, 'learning_rate/qwen_vl_interface': 2.6663710023198956e-06, 'learning_rate/action_model': 2.6663710023198956e-05, 'epoch': 0.25}) +[2026-10-03 19:37:40,476][starVLA.training.train_starvla][INFO] - Step 3448, Loss: {'train/loss': 0.0003841441939584911, 'train/grad_norm_pre_clip': 0.3625550866127014, 'timing/data': 0.0003157590108457953, 'timing/model': 0.391190924012335, 'learning_rate/qwen_vl_interface': 2.663816011197405e-06, 'learning_rate/action_model': 2.6638160111974048e-05, 'epoch': 0.25}) +[2026-10-03 19:37:41,175][starVLA.training.train_starvla][INFO] - Step 3449, Loss: {'train/loss': 0.002134176203981042, 'train/grad_norm_pre_clip': 6.125611782073975, 'timing/data': 0.00019982000230811536, 'timing/model': 0.39233763300580904, 'learning_rate/qwen_vl_interface': 2.661262083157463e-06, 'learning_rate/action_model': 2.661262083157463e-05, 'epoch': 0.25}) +[2026-10-03 19:37:41,873][starVLA.training.train_starvla][INFO] - Step 3450, Loss: {'train/loss': 0.7086119651794434, 'train/grad_norm_pre_clip': 42.855716705322266, 'timing/data': 0.0003481800085864961, 'timing/model': 0.38967553398106247, 'learning_rate/qwen_vl_interface': 2.6587092192498943e-06, 'learning_rate/action_model': 2.658709219249894e-05, 'epoch': 0.25}) +[2026-10-03 19:37:42,570][starVLA.training.train_starvla][INFO] - Step 3451, Loss: {'train/loss': 6.377637873811182e-06, 'train/grad_norm_pre_clip': 0.0011165185132995248, 'timing/data': 0.00020429998403415084, 'timing/model': 0.3906440529972315, 'learning_rate/qwen_vl_interface': 2.656157420524084e-06, 'learning_rate/action_model': 2.6561574205240842e-05, 'epoch': 0.25}) +[2026-10-03 19:37:43,476][starVLA.training.train_starvla][INFO] - Step 3452, Loss: {'train/loss': 0.48280060291290283, 'train/grad_norm_pre_clip': 127.5276107788086, 'timing/data': 0.0002889900060836226, 'timing/model': 0.5966597940132488, 'learning_rate/qwen_vl_interface': 2.6536066880289795e-06, 'learning_rate/action_model': 2.6536066880289794e-05, 'epoch': 0.25}) +[2026-10-03 19:37:44,179][starVLA.training.train_starvla][INFO] - Step 3453, Loss: {'train/loss': 0.14384952187538147, 'train/grad_norm_pre_clip': 131.02101135253906, 'timing/data': 0.00020403999951668084, 'timing/model': 0.393510594003601, 'learning_rate/qwen_vl_interface': 2.6510570228130916e-06, 'learning_rate/action_model': 2.6510570228130914e-05, 'epoch': 0.25}) +[2026-10-03 19:37:44,881][starVLA.training.train_starvla][INFO] - Step 3454, Loss: {'train/loss': 0.3393062949180603, 'train/grad_norm_pre_clip': 87.19740295410156, 'timing/data': 0.0003432899829931557, 'timing/model': 0.3926768030214589, 'learning_rate/qwen_vl_interface': 2.648508425924491e-06, 'learning_rate/action_model': 2.6485084259244907e-05, 'epoch': 0.25}) +[2026-10-03 19:37:45,582][starVLA.training.train_starvla][INFO] - Step 3455, Loss: {'train/loss': 0.007823873311281204, 'train/grad_norm_pre_clip': 19.473283767700195, 'timing/data': 0.0003364400181453675, 'timing/model': 0.39152496302267537, 'learning_rate/qwen_vl_interface': 2.6459608984108093e-06, 'learning_rate/action_model': 2.645960898410809e-05, 'epoch': 0.25}) +[2026-10-03 19:37:46,285][starVLA.training.train_starvla][INFO] - Step 3456, Loss: {'train/loss': 1.0400837709312327e-05, 'train/grad_norm_pre_clip': 0.0019715260714292526, 'timing/data': 0.00021308998111635447, 'timing/model': 0.3933619129820727, 'learning_rate/qwen_vl_interface': 2.6434144413192414e-06, 'learning_rate/action_model': 2.6434144413192414e-05, 'epoch': 0.25}) +[2026-10-03 19:37:46,990][starVLA.training.train_starvla][INFO] - Step 3457, Loss: {'train/loss': 0.08057564496994019, 'train/grad_norm_pre_clip': 115.14950561523438, 'timing/data': 0.0002525000018067658, 'timing/model': 0.39535171299939975, 'learning_rate/qwen_vl_interface': 2.6408690556965347e-06, 'learning_rate/action_model': 2.640869055696535e-05, 'epoch': 0.25}) +[2026-10-03 19:37:47,901][starVLA.training.train_starvla][INFO] - Step 3458, Loss: {'train/loss': 0.7808446884155273, 'train/grad_norm_pre_clip': 146.34878540039062, 'timing/data': 0.0002083999861497432, 'timing/model': 0.3912400839908514, 'learning_rate/qwen_vl_interface': 2.6383247425890047e-06, 'learning_rate/action_model': 2.6383247425890047e-05, 'epoch': 0.25}) +[2026-10-03 19:37:48,603][starVLA.training.train_starvla][INFO] - Step 3459, Loss: {'train/loss': 6.332918019325007e-06, 'train/grad_norm_pre_clip': 0.0011436892673373222, 'timing/data': 0.0002233099949080497, 'timing/model': 0.3934812429943122, 'learning_rate/qwen_vl_interface': 2.6357815030425204e-06, 'learning_rate/action_model': 2.6357815030425205e-05, 'epoch': 0.25}) +[2026-10-03 19:37:49,305][starVLA.training.train_starvla][INFO] - Step 3460, Loss: {'train/loss': 2.1694464521715418e-05, 'train/grad_norm_pre_clip': 0.009848861023783684, 'timing/data': 0.0003135399892926216, 'timing/model': 0.3919785230245907, 'learning_rate/qwen_vl_interface': 2.6332393381025132e-06, 'learning_rate/action_model': 2.6332393381025133e-05, 'epoch': 0.25}) +[2026-10-03 19:37:50,009][starVLA.training.train_starvla][INFO] - Step 3461, Loss: {'train/loss': 0.6312150359153748, 'train/grad_norm_pre_clip': 70.63683319091797, 'timing/data': 0.0002756600151769817, 'timing/model': 0.39308425298077054, 'learning_rate/qwen_vl_interface': 2.6306982488139704e-06, 'learning_rate/action_model': 2.6306982488139702e-05, 'epoch': 0.25}) +[2026-10-03 19:37:50,707][starVLA.training.train_starvla][INFO] - Step 3462, Loss: {'train/loss': 0.4893959164619446, 'train/grad_norm_pre_clip': 148.95867919921875, 'timing/data': 0.00033581999014131725, 'timing/model': 0.3914978739921935, 'learning_rate/qwen_vl_interface': 2.628158236221438e-06, 'learning_rate/action_model': 2.628158236221438e-05, 'epoch': 0.25}) +[2026-10-03 19:37:51,612][starVLA.training.train_starvla][INFO] - Step 3463, Loss: {'train/loss': 0.007415202911943197, 'train/grad_norm_pre_clip': 33.05890655517578, 'timing/data': 0.00020900997333228588, 'timing/model': 0.3921571040118579, 'learning_rate/qwen_vl_interface': 2.6256193013690183e-06, 'learning_rate/action_model': 2.6256193013690182e-05, 'epoch': 0.25}) +[2026-10-03 19:37:52,311][starVLA.training.train_starvla][INFO] - Step 3464, Loss: {'train/loss': 0.3687479794025421, 'train/grad_norm_pre_clip': 91.60895538330078, 'timing/data': 0.0002096099779009819, 'timing/model': 0.3912442029977683, 'learning_rate/qwen_vl_interface': 2.623081445300373e-06, 'learning_rate/action_model': 2.6230814453003728e-05, 'epoch': 0.25}) +[2026-10-03 19:37:53,013][starVLA.training.train_starvla][INFO] - Step 3465, Loss: {'train/loss': 0.3221816122531891, 'train/grad_norm_pre_clip': 117.03642272949219, 'timing/data': 0.00032294998527504504, 'timing/model': 0.39292581399786286, 'learning_rate/qwen_vl_interface': 2.6205446690587183e-06, 'learning_rate/action_model': 2.6205446690587182e-05, 'epoch': 0.25}) +[2026-10-03 19:37:53,712][starVLA.training.train_starvla][INFO] - Step 3466, Loss: {'train/loss': 0.00936099886894226, 'train/grad_norm_pre_clip': 8.330512046813965, 'timing/data': 0.0002571800141595304, 'timing/model': 0.389872023020871, 'learning_rate/qwen_vl_interface': 2.618008973686828e-06, 'learning_rate/action_model': 2.618008973686828e-05, 'epoch': 0.25}) +[2026-10-03 19:37:54,411][starVLA.training.train_starvla][INFO] - Step 3467, Loss: {'train/loss': 0.8447899222373962, 'train/grad_norm_pre_clip': 118.09474182128906, 'timing/data': 0.00044626000453718007, 'timing/model': 0.39075748302275315, 'learning_rate/qwen_vl_interface': 2.615474360227028e-06, 'learning_rate/action_model': 2.615474360227028e-05, 'epoch': 0.25}) +[2026-10-03 19:37:55,307][starVLA.training.train_starvla][INFO] - Step 3468, Loss: {'train/loss': 0.00023495833738707006, 'train/grad_norm_pre_clip': 0.10378028452396393, 'timing/data': 0.000395249982830137, 'timing/model': 0.5877086149994284, 'learning_rate/qwen_vl_interface': 2.612940829721204e-06, 'learning_rate/action_model': 2.612940829721204e-05, 'epoch': 0.25}) +[2026-10-03 19:37:56,006][starVLA.training.train_starvla][INFO] - Step 3469, Loss: {'train/loss': 0.004525552503764629, 'train/grad_norm_pre_clip': 2.6295299530029297, 'timing/data': 0.00027416000375524163, 'timing/model': 0.3920798939943779, 'learning_rate/qwen_vl_interface': 2.6104083832107948e-06, 'learning_rate/action_model': 2.6104083832107945e-05, 'epoch': 0.25}) +[2026-10-03 19:37:56,702][starVLA.training.train_starvla][INFO] - Step 3470, Loss: {'train/loss': 8.428459841525182e-05, 'train/grad_norm_pre_clip': 0.06108569726347923, 'timing/data': 0.00040155998431146145, 'timing/model': 0.3895164630084764, 'learning_rate/qwen_vl_interface': 2.607877021736793e-06, 'learning_rate/action_model': 2.6078770217367926e-05, 'epoch': 0.25}) +[2026-10-03 19:37:57,400][starVLA.training.train_starvla][INFO] - Step 3471, Loss: {'train/loss': 2.0197305275360122e-05, 'train/grad_norm_pre_clip': 0.01350098941475153, 'timing/data': 0.00031800000579096377, 'timing/model': 0.38922266400186345, 'learning_rate/qwen_vl_interface': 2.605346746339745e-06, 'learning_rate/action_model': 2.605346746339745e-05, 'epoch': 0.25}) +[2026-10-03 19:37:58,099][starVLA.training.train_starvla][INFO] - Step 3472, Loss: {'train/loss': 4.408811219036579e-05, 'train/grad_norm_pre_clip': 0.01724189892411232, 'timing/data': 0.0002101499994751066, 'timing/model': 0.3911569329793565, 'learning_rate/qwen_vl_interface': 2.602817558059753e-06, 'learning_rate/action_model': 2.6028175580597532e-05, 'epoch': 0.25}) +[2026-10-03 19:37:59,002][starVLA.training.train_starvla][INFO] - Step 3473, Loss: {'train/loss': 0.00043604563688859344, 'train/grad_norm_pre_clip': 0.8789710998535156, 'timing/data': 0.0002385500119999051, 'timing/model': 0.5956487740040757, 'learning_rate/qwen_vl_interface': 2.60028945793647e-06, 'learning_rate/action_model': 2.60028945793647e-05, 'epoch': 0.25}) +[2026-10-03 19:37:59,700][starVLA.training.train_starvla][INFO] - Step 3474, Loss: {'train/loss': 0.21679145097732544, 'train/grad_norm_pre_clip': 51.135887145996094, 'timing/data': 0.0002787899866234511, 'timing/model': 0.3924420629919041, 'learning_rate/qwen_vl_interface': 2.5977624470091024e-06, 'learning_rate/action_model': 2.5977624470091023e-05, 'epoch': 0.25}) +[2026-10-03 19:38:00,400][starVLA.training.train_starvla][INFO] - Step 3475, Loss: {'train/loss': 0.02166852541267872, 'train/grad_norm_pre_clip': 37.00373458862305, 'timing/data': 0.00020891998428851366, 'timing/model': 0.3913524429954123, 'learning_rate/qwen_vl_interface': 2.59523652631641e-06, 'learning_rate/action_model': 2.5952365263164097e-05, 'epoch': 0.25}) +[2026-10-03 19:38:01,100][starVLA.training.train_starvla][INFO] - Step 3476, Loss: {'train/loss': 0.0007114653708413243, 'train/grad_norm_pre_clip': 0.30675411224365234, 'timing/data': 0.00036403999547474086, 'timing/model': 0.3905897130025551, 'learning_rate/qwen_vl_interface': 2.5927116968967024e-06, 'learning_rate/action_model': 2.5927116968967025e-05, 'epoch': 0.25}) +[2026-10-03 19:38:01,800][starVLA.training.train_starvla][INFO] - Step 3477, Loss: {'train/loss': 2.057011988654267e-05, 'train/grad_norm_pre_clip': 0.00565495528280735, 'timing/data': 0.0002560699940659106, 'timing/model': 0.393490402988391, 'learning_rate/qwen_vl_interface': 2.5901879597878417e-06, 'learning_rate/action_model': 2.5901879597878415e-05, 'epoch': 0.25}) +[2026-10-03 19:38:02,703][starVLA.training.train_starvla][INFO] - Step 3478, Loss: {'train/loss': 0.29089850187301636, 'train/grad_norm_pre_clip': 487.25128173828125, 'timing/data': 0.0003502100007608533, 'timing/model': 0.5956759339896962, 'learning_rate/qwen_vl_interface': 2.5876653160272395e-06, 'learning_rate/action_model': 2.5876653160272397e-05, 'epoch': 0.25}) +[2026-10-03 19:38:03,402][starVLA.training.train_starvla][INFO] - Step 3479, Loss: {'train/loss': 0.7107388973236084, 'train/grad_norm_pre_clip': 92.74666595458984, 'timing/data': 0.0003329899918753654, 'timing/model': 0.390629784000339, 'learning_rate/qwen_vl_interface': 2.5851437666518607e-06, 'learning_rate/action_model': 2.5851437666518607e-05, 'epoch': 0.25}) +[2026-10-03 19:38:04,100][starVLA.training.train_starvla][INFO] - Step 3480, Loss: {'train/loss': 9.352782944915816e-05, 'train/grad_norm_pre_clip': 0.040932394564151764, 'timing/data': 0.00032974002533592284, 'timing/model': 0.3891119939798955, 'learning_rate/qwen_vl_interface': 2.5826233126982198e-06, 'learning_rate/action_model': 2.5826233126982196e-05, 'epoch': 0.25}) +[2026-10-03 19:38:04,803][starVLA.training.train_starvla][INFO] - Step 3481, Loss: {'train/loss': 0.0014310587430372834, 'train/grad_norm_pre_clip': 0.8012916445732117, 'timing/data': 0.00041181998676620424, 'timing/model': 0.3938706229964737, 'learning_rate/qwen_vl_interface': 2.580103955202381e-06, 'learning_rate/action_model': 2.580103955202381e-05, 'epoch': 0.25}) +[2026-10-03 19:38:05,508][starVLA.training.train_starvla][INFO] - Step 3482, Loss: {'train/loss': 0.025144677609205246, 'train/grad_norm_pre_clip': 42.22285461425781, 'timing/data': 0.00034350002533756196, 'timing/model': 0.3940362430003006, 'learning_rate/qwen_vl_interface': 2.577585695199954e-06, 'learning_rate/action_model': 2.5775856951999538e-05, 'epoch': 0.25}) +[2026-10-03 19:38:06,213][starVLA.training.train_starvla][INFO] - Step 3483, Loss: {'train/loss': 0.0009456141269765794, 'train/grad_norm_pre_clip': 1.7326985597610474, 'timing/data': 0.0003171199932694435, 'timing/model': 0.39381697299540974, 'learning_rate/qwen_vl_interface': 2.575068533726101e-06, 'learning_rate/action_model': 2.575068533726101e-05, 'epoch': 0.25}) +[2026-10-03 19:38:07,131][starVLA.training.train_starvla][INFO] - Step 3484, Loss: {'train/loss': 6.050489537301473e-05, 'train/grad_norm_pre_clip': 0.023468196392059326, 'timing/data': 0.0003570389817468822, 'timing/model': 0.39507481301552616, 'learning_rate/qwen_vl_interface': 2.5725524718155333e-06, 'learning_rate/action_model': 2.572552471815533e-05, 'epoch': 0.26}) +[2026-10-03 19:38:07,838][starVLA.training.train_starvla][INFO] - Step 3485, Loss: {'train/loss': 3.656441913335584e-05, 'train/grad_norm_pre_clip': 0.012138061225414276, 'timing/data': 0.0002726500097196549, 'timing/model': 0.39674273299169727, 'learning_rate/qwen_vl_interface': 2.570037510502508e-06, 'learning_rate/action_model': 2.5700375105025083e-05, 'epoch': 0.26}) +[2026-10-03 19:38:08,546][starVLA.training.train_starvla][INFO] - Step 3486, Loss: {'train/loss': 0.024607479572296143, 'train/grad_norm_pre_clip': 39.603878021240234, 'timing/data': 0.00036517000989988446, 'timing/model': 0.3959551230072975, 'learning_rate/qwen_vl_interface': 2.5675236508208313e-06, 'learning_rate/action_model': 2.5675236508208313e-05, 'epoch': 0.26}) +[2026-10-03 19:38:09,255][starVLA.training.train_starvla][INFO] - Step 3487, Loss: {'train/loss': 0.10891193896532059, 'train/grad_norm_pre_clip': 75.95247650146484, 'timing/data': 0.0003342099953442812, 'timing/model': 0.39604140300070867, 'learning_rate/qwen_vl_interface': 2.565010893803856e-06, 'learning_rate/action_model': 2.565010893803856e-05, 'epoch': 0.26}) +[2026-10-03 19:38:09,963][starVLA.training.train_starvla][INFO] - Step 3488, Loss: {'train/loss': 0.00116860365960747, 'train/grad_norm_pre_clip': 0.3905424177646637, 'timing/data': 0.0002994899987243116, 'timing/model': 0.3954025420243852, 'learning_rate/qwen_vl_interface': 2.562499240484482e-06, 'learning_rate/action_model': 2.5624992404844816e-05, 'epoch': 0.26}) +[2026-10-03 19:38:10,869][starVLA.training.train_starvla][INFO] - Step 3489, Loss: {'train/loss': 0.0002674766874406487, 'train/grad_norm_pre_clip': 0.11059203743934631, 'timing/data': 0.0002672600094228983, 'timing/model': 0.5931266240077093, 'learning_rate/qwen_vl_interface': 2.559988691895155e-06, 'learning_rate/action_model': 2.559988691895155e-05, 'epoch': 0.26}) +[2026-10-03 19:38:11,577][starVLA.training.train_starvla][INFO] - Step 3490, Loss: {'train/loss': 6.301650137174875e-05, 'train/grad_norm_pre_clip': 0.027375901117920876, 'timing/data': 0.0002230099926237017, 'timing/model': 0.3955661430081818, 'learning_rate/qwen_vl_interface': 2.5574792490678648e-06, 'learning_rate/action_model': 2.5574792490678644e-05, 'epoch': 0.26}) +[2026-10-03 19:38:12,283][starVLA.training.train_starvla][INFO] - Step 3491, Loss: {'train/loss': 0.0005015849601477385, 'train/grad_norm_pre_clip': 0.13638447225093842, 'timing/data': 0.00022265000734478235, 'timing/model': 0.3937140830093995, 'learning_rate/qwen_vl_interface': 2.5549709130341493e-06, 'learning_rate/action_model': 2.5549709130341493e-05, 'epoch': 0.26}) +[2026-10-03 19:38:12,992][starVLA.training.train_starvla][INFO] - Step 3492, Loss: {'train/loss': 2.6827277906704694e-05, 'train/grad_norm_pre_clip': 0.011555436067283154, 'timing/data': 0.00030910997884348035, 'timing/model': 0.3959625120041892, 'learning_rate/qwen_vl_interface': 2.5524636848250924e-06, 'learning_rate/action_model': 2.552463684825092e-05, 'epoch': 0.26}) +[2026-10-03 19:38:13,702][starVLA.training.train_starvla][INFO] - Step 3493, Loss: {'train/loss': 0.1148974597454071, 'train/grad_norm_pre_clip': 165.92849731445312, 'timing/data': 0.00026581002748571336, 'timing/model': 0.3966548930038698, 'learning_rate/qwen_vl_interface': 2.5499575654713186e-06, 'learning_rate/action_model': 2.5499575654713185e-05, 'epoch': 0.26}) +[2026-10-03 19:38:14,411][starVLA.training.train_starvla][INFO] - Step 3494, Loss: {'train/loss': 0.2995500862598419, 'train/grad_norm_pre_clip': 484.3099060058594, 'timing/data': 0.0003978400200139731, 'timing/model': 0.395812973001739, 'learning_rate/qwen_vl_interface': 2.547452556003001e-06, 'learning_rate/action_model': 2.547452556003001e-05, 'epoch': 0.26}) +[2026-10-03 19:38:15,324][starVLA.training.train_starvla][INFO] - Step 3495, Loss: {'train/loss': 0.00013412456610240042, 'train/grad_norm_pre_clip': 0.06469220668077469, 'timing/data': 0.0003131600096821785, 'timing/model': 0.39371852300246246, 'learning_rate/qwen_vl_interface': 2.5449486574498535e-06, 'learning_rate/action_model': 2.5449486574498534e-05, 'epoch': 0.26}) +[2026-10-03 19:38:16,032][starVLA.training.train_starvla][INFO] - Step 3496, Loss: {'train/loss': 1.3962117009214126e-05, 'train/grad_norm_pre_clip': 0.0035403163637965918, 'timing/data': 0.00038959001540206373, 'timing/model': 0.3952758130035363, 'learning_rate/qwen_vl_interface': 2.5424458708411335e-06, 'learning_rate/action_model': 2.5424458708411335e-05, 'epoch': 0.26}) +[2026-10-03 19:38:16,739][starVLA.training.train_starvla][INFO] - Step 3497, Loss: {'train/loss': 0.7431833148002625, 'train/grad_norm_pre_clip': 420.0908508300781, 'timing/data': 0.0002226899960078299, 'timing/model': 0.3943855820107274, 'learning_rate/qwen_vl_interface': 2.5399441972056437e-06, 'learning_rate/action_model': 2.5399441972056438e-05, 'epoch': 0.26}) +[2026-10-03 19:38:17,446][starVLA.training.train_starvla][INFO] - Step 3498, Loss: {'train/loss': 8.678816084284335e-05, 'train/grad_norm_pre_clip': 0.023318931460380554, 'timing/data': 0.00034406999475322664, 'timing/model': 0.3951571329962462, 'learning_rate/qwen_vl_interface': 2.537443637571726e-06, 'learning_rate/action_model': 2.537443637571726e-05, 'epoch': 0.26}) +[2026-10-03 19:38:18,152][starVLA.training.train_starvla][INFO] - Step 3499, Loss: {'train/loss': 0.1497659683227539, 'train/grad_norm_pre_clip': 177.42889404296875, 'timing/data': 0.0003213699965272099, 'timing/model': 0.3931858030264266, 'learning_rate/qwen_vl_interface': 2.534944192967269e-06, 'learning_rate/action_model': 2.534944192967269e-05, 'epoch': 0.26}) +[2026-10-03 19:38:19,063][starVLA.training.train_starvla][INFO] - Step 3500, Loss: {'train/loss': 1.5936420822981745e-05, 'train/grad_norm_pre_clip': 0.002719730604439974, 'timing/data': 0.0002693599963095039, 'timing/model': 0.3938173330097925, 'learning_rate/qwen_vl_interface': 2.532445864419695e-06, 'learning_rate/action_model': 2.5324458644196952e-05, 'epoch': 0.26}) +[2026-10-03 19:38:19,769][starVLA.training.train_starvla][INFO] - Step 3501, Loss: {'train/loss': 0.2002444863319397, 'train/grad_norm_pre_clip': 365.83367919921875, 'timing/data': 0.00025522001669742167, 'timing/model': 0.3963297629961744, 'learning_rate/qwen_vl_interface': 2.5299486529559758e-06, 'learning_rate/action_model': 2.5299486529559756e-05, 'epoch': 0.26}) +[2026-10-03 19:38:20,475][starVLA.training.train_starvla][INFO] - Step 3502, Loss: {'train/loss': 0.0011030370369553566, 'train/grad_norm_pre_clip': 0.9785115718841553, 'timing/data': 0.0003431999939493835, 'timing/model': 0.39446222299011424, 'learning_rate/qwen_vl_interface': 2.5274525596026202e-06, 'learning_rate/action_model': 2.52745255960262e-05, 'epoch': 0.26}) +[2026-10-03 19:38:21,181][starVLA.training.train_starvla][INFO] - Step 3503, Loss: {'train/loss': 9.662711818236858e-05, 'train/grad_norm_pre_clip': 0.03964896872639656, 'timing/data': 0.00037703002453781664, 'timing/model': 0.3944589130114764, 'learning_rate/qwen_vl_interface': 2.524957585385677e-06, 'learning_rate/action_model': 2.524957585385677e-05, 'epoch': 0.26}) +[2026-10-03 19:38:21,888][starVLA.training.train_starvla][INFO] - Step 3504, Loss: {'train/loss': 0.00033973142853938043, 'train/grad_norm_pre_clip': 0.3718125820159912, 'timing/data': 0.00024083000607788563, 'timing/model': 0.39460289300768636, 'learning_rate/qwen_vl_interface': 2.522463731330738e-06, 'learning_rate/action_model': 2.5224637313307374e-05, 'epoch': 0.26}) +[2026-10-03 19:38:22,794][starVLA.training.train_starvla][INFO] - Step 3505, Loss: {'train/loss': 0.00012745724234264344, 'train/grad_norm_pre_clip': 0.06789149343967438, 'timing/data': 0.0003713100159075111, 'timing/model': 0.5935020740143955, 'learning_rate/qwen_vl_interface': 2.5199709984629294e-06, 'learning_rate/action_model': 2.5199709984629292e-05, 'epoch': 0.26}) +[2026-10-03 19:38:23,502][starVLA.training.train_starvla][INFO] - Step 3506, Loss: {'train/loss': 0.00023881910601630807, 'train/grad_norm_pre_clip': 0.20328520238399506, 'timing/data': 0.0003481899912003428, 'timing/model': 0.3953655319928657, 'learning_rate/qwen_vl_interface': 2.517479387806922e-06, 'learning_rate/action_model': 2.517479387806922e-05, 'epoch': 0.26}) +[2026-10-03 19:38:24,208][starVLA.training.train_starvla][INFO] - Step 3507, Loss: {'train/loss': 0.013750048354268074, 'train/grad_norm_pre_clip': 55.294403076171875, 'timing/data': 0.0003607500111684203, 'timing/model': 0.3954984029987827, 'learning_rate/qwen_vl_interface': 2.5149889003869217e-06, 'learning_rate/action_model': 2.5149889003869214e-05, 'epoch': 0.26}) +[2026-10-03 19:38:24,915][starVLA.training.train_starvla][INFO] - Step 3508, Loss: {'train/loss': 8.70986987138167e-05, 'train/grad_norm_pre_clip': 0.04272489622235298, 'timing/data': 0.00035837001632899046, 'timing/model': 0.3942903119896073, 'learning_rate/qwen_vl_interface': 2.5124995372266736e-06, 'learning_rate/action_model': 2.5124995372266735e-05, 'epoch': 0.26}) +[2026-10-03 19:38:25,622][starVLA.training.train_starvla][INFO] - Step 3509, Loss: {'train/loss': 0.02007582224905491, 'train/grad_norm_pre_clip': 38.935577392578125, 'timing/data': 0.00027106000925414264, 'timing/model': 0.3952865329920314, 'learning_rate/qwen_vl_interface': 2.5100112993494626e-06, 'learning_rate/action_model': 2.5100112993494624e-05, 'epoch': 0.26}) +[2026-10-03 19:38:26,330][starVLA.training.train_starvla][INFO] - Step 3510, Loss: {'train/loss': 2.2507321773446165e-05, 'train/grad_norm_pre_clip': 0.006034247111529112, 'timing/data': 0.0002890799951273948, 'timing/model': 0.3938417629979085, 'learning_rate/qwen_vl_interface': 2.5075241877781085e-06, 'learning_rate/action_model': 2.5075241877781086e-05, 'epoch': 0.26}) +[2026-10-03 19:38:27,234][starVLA.training.train_starvla][INFO] - Step 3511, Loss: {'train/loss': 0.0003889909712597728, 'train/grad_norm_pre_clip': 0.14146225154399872, 'timing/data': 0.0003153899742756039, 'timing/model': 0.39412744200672023, 'learning_rate/qwen_vl_interface': 2.5050382035349673e-06, 'learning_rate/action_model': 2.5050382035349674e-05, 'epoch': 0.26}) +[2026-10-03 19:38:27,941][starVLA.training.train_starvla][INFO] - Step 3512, Loss: {'train/loss': 0.0006032941746525466, 'train/grad_norm_pre_clip': 0.4904108941555023, 'timing/data': 0.00039937999099493027, 'timing/model': 0.3953506730031222, 'learning_rate/qwen_vl_interface': 2.5025533476419343e-06, 'learning_rate/action_model': 2.5025533476419343e-05, 'epoch': 0.26}) +[2026-10-03 19:38:28,647][starVLA.training.train_starvla][INFO] - Step 3513, Loss: {'train/loss': 2.0786348613910377e-05, 'train/grad_norm_pre_clip': 0.005487852264195681, 'timing/data': 0.0002889299939852208, 'timing/model': 0.3958090420055669, 'learning_rate/qwen_vl_interface': 2.5000696211204394e-06, 'learning_rate/action_model': 2.5000696211204394e-05, 'epoch': 0.26}) +[2026-10-03 19:38:29,350][starVLA.training.train_starvla][INFO] - Step 3514, Loss: {'train/loss': 3.624878445407376e-05, 'train/grad_norm_pre_clip': 0.02882927842438221, 'timing/data': 0.0003600100171752274, 'timing/model': 0.39321790300891735, 'learning_rate/qwen_vl_interface': 2.4975870249914495e-06, 'learning_rate/action_model': 2.4975870249914497e-05, 'epoch': 0.26}) +[2026-10-03 19:38:30,053][starVLA.training.train_starvla][INFO] - Step 3515, Loss: {'train/loss': 0.0012936923885717988, 'train/grad_norm_pre_clip': 5.258073806762695, 'timing/data': 0.0003733299963641912, 'timing/model': 0.39283479301957414, 'learning_rate/qwen_vl_interface': 2.4951055602754664e-06, 'learning_rate/action_model': 2.4951055602754665e-05, 'epoch': 0.26}) +[2026-10-03 19:38:30,961][starVLA.training.train_starvla][INFO] - Step 3516, Loss: {'train/loss': 1.566078390169423e-05, 'train/grad_norm_pre_clip': 0.0028819204308092594, 'timing/data': 0.00042725002276711166, 'timing/model': 0.3942156629927922, 'learning_rate/qwen_vl_interface': 2.4926252279925255e-06, 'learning_rate/action_model': 2.4926252279925254e-05, 'epoch': 0.26}) +[2026-10-03 19:38:31,666][starVLA.training.train_starvla][INFO] - Step 3517, Loss: {'train/loss': 8.564525342080742e-05, 'train/grad_norm_pre_clip': 0.029961982741951942, 'timing/data': 0.00027285999385640025, 'timing/model': 0.3934732429916039, 'learning_rate/qwen_vl_interface': 2.490146029162199e-06, 'learning_rate/action_model': 2.4901460291621988e-05, 'epoch': 0.26}) +[2026-10-03 19:38:32,373][starVLA.training.train_starvla][INFO] - Step 3518, Loss: {'train/loss': 0.22521279752254486, 'train/grad_norm_pre_clip': 80.26708984375, 'timing/data': 0.0003745300055015832, 'timing/model': 0.394548903015675, 'learning_rate/qwen_vl_interface': 2.4876679648035905e-06, 'learning_rate/action_model': 2.4876679648035903e-05, 'epoch': 0.26}) +[2026-10-03 19:38:33,078][starVLA.training.train_starvla][INFO] - Step 3519, Loss: {'train/loss': 0.007884010672569275, 'train/grad_norm_pre_clip': 17.41504669189453, 'timing/data': 0.00033944001188501716, 'timing/model': 0.3945833520265296, 'learning_rate/qwen_vl_interface': 2.4851910359353392e-06, 'learning_rate/action_model': 2.485191035935339e-05, 'epoch': 0.26}) +[2026-10-03 19:38:33,784][starVLA.training.train_starvla][INFO] - Step 3520, Loss: {'train/loss': 0.001596457208506763, 'train/grad_norm_pre_clip': 1.830029010772705, 'timing/data': 0.0003615299938246608, 'timing/model': 0.39507197300554253, 'learning_rate/qwen_vl_interface': 2.4827152435756185e-06, 'learning_rate/action_model': 2.4827152435756185e-05, 'epoch': 0.26}) +[2026-10-03 19:38:34,690][starVLA.training.train_starvla][INFO] - Step 3521, Loss: {'train/loss': 0.05812274292111397, 'train/grad_norm_pre_clip': 71.43341827392578, 'timing/data': 0.00022032001288607717, 'timing/model': 0.5944989949930459, 'learning_rate/qwen_vl_interface': 2.4802405887421307e-06, 'learning_rate/action_model': 2.4802405887421306e-05, 'epoch': 0.26}) +[2026-10-03 19:38:35,394][starVLA.training.train_starvla][INFO] - Step 3522, Loss: {'train/loss': 2.8839453079854138e-05, 'train/grad_norm_pre_clip': 0.016270402818918228, 'timing/data': 0.00033735999022610486, 'timing/model': 0.3933853130147327, 'learning_rate/qwen_vl_interface': 2.4777670724521134e-06, 'learning_rate/action_model': 2.4777670724521134e-05, 'epoch': 0.26}) +[2026-10-03 19:38:36,097][starVLA.training.train_starvla][INFO] - Step 3523, Loss: {'train/loss': 6.603531073778868e-05, 'train/grad_norm_pre_clip': 0.05711517482995987, 'timing/data': 0.0003212899900972843, 'timing/model': 0.39304291299777105, 'learning_rate/qwen_vl_interface': 2.475294695722337e-06, 'learning_rate/action_model': 2.475294695722337e-05, 'epoch': 0.26}) +[2026-10-03 19:38:36,803][starVLA.training.train_starvla][INFO] - Step 3524, Loss: {'train/loss': 3.94133894587867e-06, 'train/grad_norm_pre_clip': 0.0009336938383057714, 'timing/data': 0.00034108999534510076, 'timing/model': 0.3937331030028872, 'learning_rate/qwen_vl_interface': 2.4728234595691006e-06, 'learning_rate/action_model': 2.4728234595691005e-05, 'epoch': 0.26}) +[2026-10-03 19:38:37,508][starVLA.training.train_starvla][INFO] - Step 3525, Loss: {'train/loss': 0.00030471617355942726, 'train/grad_norm_pre_clip': 0.5595524311065674, 'timing/data': 0.0003590199921745807, 'timing/model': 0.39555604199995287, 'learning_rate/qwen_vl_interface': 2.4703533650082368e-06, 'learning_rate/action_model': 2.4703533650082366e-05, 'epoch': 0.26}) +[2026-10-03 19:38:38,417][starVLA.training.train_starvla][INFO] - Step 3526, Loss: {'train/loss': 1.131697177886963, 'train/grad_norm_pre_clip': 126.70332336425781, 'timing/data': 0.000309009978082031, 'timing/model': 0.5980322639807127, 'learning_rate/qwen_vl_interface': 2.4678844130551103e-06, 'learning_rate/action_model': 2.4678844130551103e-05, 'epoch': 0.26}) +[2026-10-03 19:38:39,126][starVLA.training.train_starvla][INFO] - Step 3527, Loss: {'train/loss': 0.6125240325927734, 'train/grad_norm_pre_clip': 199.5067901611328, 'timing/data': 0.000251580000622198, 'timing/model': 0.39636130299186334, 'learning_rate/qwen_vl_interface': 2.4654166047246102e-06, 'learning_rate/action_model': 2.46541660472461e-05, 'epoch': 0.26}) +[2026-10-03 19:38:39,831][starVLA.training.train_starvla][INFO] - Step 3528, Loss: {'train/loss': 0.00011583552259253338, 'train/grad_norm_pre_clip': 0.07690462470054626, 'timing/data': 0.000348879984812811, 'timing/model': 0.393612372019561, 'learning_rate/qwen_vl_interface': 2.4629499410311606e-06, 'learning_rate/action_model': 2.462949941031161e-05, 'epoch': 0.26}) +[2026-10-03 19:38:40,538][starVLA.training.train_starvla][INFO] - Step 3529, Loss: {'train/loss': 2.9802158678648993e-06, 'train/grad_norm_pre_clip': 0.0009504922199994326, 'timing/data': 0.00032148000900633633, 'timing/model': 0.3957462229882367, 'learning_rate/qwen_vl_interface': 2.460484422988715e-06, 'learning_rate/action_model': 2.4604844229887146e-05, 'epoch': 0.26}) +[2026-10-03 19:38:41,245][starVLA.training.train_starvla][INFO] - Step 3530, Loss: {'train/loss': 0.16958552598953247, 'train/grad_norm_pre_clip': 296.41387939453125, 'timing/data': 0.0003677300119306892, 'timing/model': 0.39365247200476006, 'learning_rate/qwen_vl_interface': 2.458020051610753e-06, 'learning_rate/action_model': 2.4580200516107528e-05, 'epoch': 0.26}) +[2026-10-03 19:38:42,149][starVLA.training.train_starvla][INFO] - Step 3531, Loss: {'train/loss': 0.024677755311131477, 'train/grad_norm_pre_clip': 23.792072296142578, 'timing/data': 0.00033152999822050333, 'timing/model': 0.5936815339955501, 'learning_rate/qwen_vl_interface': 2.455556827910286e-06, 'learning_rate/action_model': 2.455556827910286e-05, 'epoch': 0.26}) +[2026-10-03 19:38:42,856][starVLA.training.train_starvla][INFO] - Step 3532, Loss: {'train/loss': 0.000713805144187063, 'train/grad_norm_pre_clip': 1.1979204416275024, 'timing/data': 0.0003227400011382997, 'timing/model': 0.39386846401612274, 'learning_rate/qwen_vl_interface': 2.453094752899853e-06, 'learning_rate/action_model': 2.453094752899853e-05, 'epoch': 0.26}) +[2026-10-03 19:38:43,564][starVLA.training.train_starvla][INFO] - Step 3533, Loss: {'train/loss': 0.6791936755180359, 'train/grad_norm_pre_clip': 99.1327133178711, 'timing/data': 0.0002865399874281138, 'timing/model': 0.3956374030094594, 'learning_rate/qwen_vl_interface': 2.4506338275915173e-06, 'learning_rate/action_model': 2.4506338275915172e-05, 'epoch': 0.26}) +[2026-10-03 19:38:44,270][starVLA.training.train_starvla][INFO] - Step 3534, Loss: {'train/loss': 5.270889596431516e-05, 'train/grad_norm_pre_clip': 0.06817710399627686, 'timing/data': 0.0004819899913854897, 'timing/model': 0.3935449039854575, 'learning_rate/qwen_vl_interface': 2.4481740529968735e-06, 'learning_rate/action_model': 2.4481740529968734e-05, 'epoch': 0.26}) +[2026-10-03 19:38:44,975][starVLA.training.train_starvla][INFO] - Step 3535, Loss: {'train/loss': 9.566318112774752e-06, 'train/grad_norm_pre_clip': 0.0038864221423864365, 'timing/data': 0.0003622999938670546, 'timing/model': 0.39381154300644994, 'learning_rate/qwen_vl_interface': 2.4457154301270427e-06, 'learning_rate/action_model': 2.4457154301270427e-05, 'epoch': 0.26}) +[2026-10-03 19:38:45,680][starVLA.training.train_starvla][INFO] - Step 3536, Loss: {'train/loss': 7.22161348676309e-05, 'train/grad_norm_pre_clip': 0.02163272351026535, 'timing/data': 0.0002343700034543872, 'timing/model': 0.39413625199813396, 'learning_rate/qwen_vl_interface': 2.443257959992671e-06, 'learning_rate/action_model': 2.443257959992671e-05, 'epoch': 0.26}) +[2026-10-03 19:38:46,587][starVLA.training.train_starvla][INFO] - Step 3537, Loss: {'train/loss': 0.5636310577392578, 'train/grad_norm_pre_clip': 138.54434204101562, 'timing/data': 0.0002935699885711074, 'timing/model': 0.39428265299648046, 'learning_rate/qwen_vl_interface': 2.440801643603933e-06, 'learning_rate/action_model': 2.440801643603933e-05, 'epoch': 0.26}) +[2026-10-03 19:38:47,291][starVLA.training.train_starvla][INFO] - Step 3538, Loss: {'train/loss': 0.0014539975672960281, 'train/grad_norm_pre_clip': 3.064014196395874, 'timing/data': 0.00036823999835178256, 'timing/model': 0.39337532297940925, 'learning_rate/qwen_vl_interface': 2.4383464819705252e-06, 'learning_rate/action_model': 2.438346481970525e-05, 'epoch': 0.26}) +[2026-10-03 19:38:47,994][starVLA.training.train_starvla][INFO] - Step 3539, Loss: {'train/loss': 0.0006588456453755498, 'train/grad_norm_pre_clip': 0.42412567138671875, 'timing/data': 0.00034367997432127595, 'timing/model': 0.39325614398694597, 'learning_rate/qwen_vl_interface': 2.435892476101674e-06, 'learning_rate/action_model': 2.435892476101674e-05, 'epoch': 0.26}) +[2026-10-03 19:38:48,703][starVLA.training.train_starvla][INFO] - Step 3540, Loss: {'train/loss': 0.7001065015792847, 'train/grad_norm_pre_clip': 142.69613647460938, 'timing/data': 0.0003453899989835918, 'timing/model': 0.3954821719962638, 'learning_rate/qwen_vl_interface': 2.433439627006128e-06, 'learning_rate/action_model': 2.4334396270061277e-05, 'epoch': 0.26}) +[2026-10-03 19:38:49,410][starVLA.training.train_starvla][INFO] - Step 3541, Loss: {'train/loss': 1.484100448578829e-05, 'train/grad_norm_pre_clip': 0.009543790481984615, 'timing/data': 0.00030852999771013856, 'timing/model': 0.3961260329815559, 'learning_rate/qwen_vl_interface': 2.4309879356921602e-06, 'learning_rate/action_model': 2.43098793569216e-05, 'epoch': 0.26}) +[2026-10-03 19:38:50,319][starVLA.training.train_starvla][INFO] - Step 3542, Loss: {'train/loss': 0.00085784075781703, 'train/grad_norm_pre_clip': 1.4273490905761719, 'timing/data': 0.0002971900103148073, 'timing/model': 0.5954149230092298, 'learning_rate/qwen_vl_interface': 2.428537403167569e-06, 'learning_rate/action_model': 2.4285374031675687e-05, 'epoch': 0.26}) +[2026-10-03 19:38:51,024][starVLA.training.train_starvla][INFO] - Step 3543, Loss: {'train/loss': 0.2125588208436966, 'train/grad_norm_pre_clip': 108.60775756835938, 'timing/data': 0.000342120009008795, 'timing/model': 0.3928019629966002, 'learning_rate/qwen_vl_interface': 2.4260880304396765e-06, 'learning_rate/action_model': 2.4260880304396765e-05, 'epoch': 0.26}) +[2026-10-03 19:38:51,727][starVLA.training.train_starvla][INFO] - Step 3544, Loss: {'train/loss': 0.11001724004745483, 'train/grad_norm_pre_clip': 139.32757568359375, 'timing/data': 0.00022067001555114985, 'timing/model': 0.3928642940009013, 'learning_rate/qwen_vl_interface': 2.4236398185153245e-06, 'learning_rate/action_model': 2.4236398185153245e-05, 'epoch': 0.26}) +[2026-10-03 19:38:52,430][starVLA.training.train_starvla][INFO] - Step 3545, Loss: {'train/loss': 1.866207821876742e-05, 'train/grad_norm_pre_clip': 0.007630415726453066, 'timing/data': 0.00031730899354442954, 'timing/model': 0.39467795300879516, 'learning_rate/qwen_vl_interface': 2.4211927684008823e-06, 'learning_rate/action_model': 2.4211927684008824e-05, 'epoch': 0.26}) +[2026-10-03 19:38:53,136][starVLA.training.train_starvla][INFO] - Step 3546, Loss: {'train/loss': 0.0012616891181096435, 'train/grad_norm_pre_clip': 6.764000415802002, 'timing/data': 0.0003267400024924427, 'timing/model': 0.3942344019887969, 'learning_rate/qwen_vl_interface': 2.4187468811022398e-06, 'learning_rate/action_model': 2.4187468811022396e-05, 'epoch': 0.26}) +[2026-10-03 19:38:53,842][starVLA.training.train_starvla][INFO] - Step 3547, Loss: {'train/loss': 6.325495633063838e-06, 'train/grad_norm_pre_clip': 0.0025750743225216866, 'timing/data': 0.000354710005922243, 'timing/model': 0.3944181630213279, 'learning_rate/qwen_vl_interface': 2.416302157624809e-06, 'learning_rate/action_model': 2.4163021576248088e-05, 'epoch': 0.26}) +[2026-10-03 19:38:54,751][starVLA.training.train_starvla][INFO] - Step 3548, Loss: {'train/loss': 0.588080644607544, 'train/grad_norm_pre_clip': 135.94137573242188, 'timing/data': 0.000354300020262599, 'timing/model': 0.3943123829958495, 'learning_rate/qwen_vl_interface': 2.4138585989735227e-06, 'learning_rate/action_model': 2.4138585989735226e-05, 'epoch': 0.26}) +[2026-10-03 19:38:55,454][starVLA.training.train_starvla][INFO] - Step 3549, Loss: {'train/loss': 0.2006394863128662, 'train/grad_norm_pre_clip': 134.47418212890625, 'timing/data': 0.00024975999258458614, 'timing/model': 0.39435495299403556, 'learning_rate/qwen_vl_interface': 2.4114162061528357e-06, 'learning_rate/action_model': 2.4114162061528354e-05, 'epoch': 0.26}) +[2026-10-03 19:38:56,159][starVLA.training.train_starvla][INFO] - Step 3550, Loss: {'train/loss': 0.0008707226370461285, 'train/grad_norm_pre_clip': 1.8031420707702637, 'timing/data': 0.00025853997794911265, 'timing/model': 0.393977471976541, 'learning_rate/qwen_vl_interface': 2.408974980166724e-06, 'learning_rate/action_model': 2.408974980166724e-05, 'epoch': 0.26}) +[2026-10-03 19:38:56,862][starVLA.training.train_starvla][INFO] - Step 3551, Loss: {'train/loss': 0.23281654715538025, 'train/grad_norm_pre_clip': 144.84512329101562, 'timing/data': 0.00032374999136663973, 'timing/model': 0.3936896329978481, 'learning_rate/qwen_vl_interface': 2.406534922018683e-06, 'learning_rate/action_model': 2.4065349220186827e-05, 'epoch': 0.26}) +[2026-10-03 19:38:57,569][starVLA.training.train_starvla][INFO] - Step 3552, Loss: {'train/loss': 2.9876741791667882e-06, 'train/grad_norm_pre_clip': 0.0006792477797716856, 'timing/data': 0.0003524000057950616, 'timing/model': 0.3948225929925684, 'learning_rate/qwen_vl_interface': 2.4040960327117287e-06, 'learning_rate/action_model': 2.4040960327117283e-05, 'epoch': 0.26}) +[2026-10-03 19:38:58,470][starVLA.training.train_starvla][INFO] - Step 3553, Loss: {'train/loss': 1.0311472578905523e-05, 'train/grad_norm_pre_clip': 0.002328425645828247, 'timing/data': 0.0003239999932702631, 'timing/model': 0.39581801198073663, 'learning_rate/qwen_vl_interface': 2.4016583132483976e-06, 'learning_rate/action_model': 2.4016583132483976e-05, 'epoch': 0.26}) +[2026-10-03 19:38:59,179][starVLA.training.train_starvla][INFO] - Step 3554, Loss: {'train/loss': 0.0016712212236598134, 'train/grad_norm_pre_clip': 9.06877613067627, 'timing/data': 0.00035041000228375196, 'timing/model': 0.3956686629971955, 'learning_rate/qwen_vl_interface': 2.399221764630741e-06, 'learning_rate/action_model': 2.3992217646307406e-05, 'epoch': 0.26}) +[2026-10-03 19:38:59,887][starVLA.training.train_starvla][INFO] - Step 3555, Loss: {'train/loss': 2.0369065168779343e-05, 'train/grad_norm_pre_clip': 0.009117549285292625, 'timing/data': 0.0003540900070220232, 'timing/model': 0.39526000298792496, 'learning_rate/qwen_vl_interface': 2.3967863878603334e-06, 'learning_rate/action_model': 2.3967863878603332e-05, 'epoch': 0.26}) +[2026-10-03 19:39:00,596][starVLA.training.train_starvla][INFO] - Step 3556, Loss: {'train/loss': 0.513114869594574, 'train/grad_norm_pre_clip': 59.41360855102539, 'timing/data': 0.0003722399997059256, 'timing/model': 0.3960219130094629, 'learning_rate/qwen_vl_interface': 2.394352183938266e-06, 'learning_rate/action_model': 2.3943521839382664e-05, 'epoch': 0.26}) +[2026-10-03 19:39:01,303][starVLA.training.train_starvla][INFO] - Step 3557, Loss: {'train/loss': 0.3000052869319916, 'train/grad_norm_pre_clip': 88.39106750488281, 'timing/data': 0.0003278199874330312, 'timing/model': 0.39483938200282864, 'learning_rate/qwen_vl_interface': 2.39191915386515e-06, 'learning_rate/action_model': 2.39191915386515e-05, 'epoch': 0.26}) +[2026-10-03 19:39:02,214][starVLA.training.train_starvla][INFO] - Step 3558, Loss: {'train/loss': 0.6938180923461914, 'train/grad_norm_pre_clip': 41.657901763916016, 'timing/data': 0.00045162998139858246, 'timing/model': 0.5980205139785539, 'learning_rate/qwen_vl_interface': 2.3894872986411095e-06, 'learning_rate/action_model': 2.3894872986411094e-05, 'epoch': 0.26}) +[2026-10-03 19:39:02,920][starVLA.training.train_starvla][INFO] - Step 3559, Loss: {'train/loss': 0.015419499948620796, 'train/grad_norm_pre_clip': 32.0976448059082, 'timing/data': 0.0004474199959076941, 'timing/model': 0.3954986030003056, 'learning_rate/qwen_vl_interface': 2.3870566192657903e-06, 'learning_rate/action_model': 2.38705661926579e-05, 'epoch': 0.26}) +[2026-10-03 19:39:03,630][starVLA.training.train_starvla][INFO] - Step 3560, Loss: {'train/loss': 3.342125273775309e-05, 'train/grad_norm_pre_clip': 0.008608895353972912, 'timing/data': 0.00025821998133324087, 'timing/model': 0.3966214520041831, 'learning_rate/qwen_vl_interface': 2.3846271167383515e-06, 'learning_rate/action_model': 2.3846271167383515e-05, 'epoch': 0.26}) +[2026-10-03 19:39:04,337][starVLA.training.train_starvla][INFO] - Step 3561, Loss: {'train/loss': 0.0001530532754259184, 'train/grad_norm_pre_clip': 0.13142597675323486, 'timing/data': 0.0002960700076073408, 'timing/model': 0.39472647302318364, 'learning_rate/qwen_vl_interface': 2.382198792057471e-06, 'learning_rate/action_model': 2.3821987920574707e-05, 'epoch': 0.26}) +[2026-10-03 19:39:05,043][starVLA.training.train_starvla][INFO] - Step 3562, Loss: {'train/loss': 5.9907026297878474e-05, 'train/grad_norm_pre_clip': 0.04093199223279953, 'timing/data': 0.0003717200015671551, 'timing/model': 0.3947692629881203, 'learning_rate/qwen_vl_interface': 2.3797716462213402e-06, 'learning_rate/action_model': 2.3797716462213402e-05, 'epoch': 0.26}) +[2026-10-03 19:39:05,748][starVLA.training.train_starvla][INFO] - Step 3563, Loss: {'train/loss': 0.00022702720889355987, 'train/grad_norm_pre_clip': 0.14033295214176178, 'timing/data': 0.00043742998968809843, 'timing/model': 0.39434801298193634, 'learning_rate/qwen_vl_interface': 2.37734568022767e-06, 'learning_rate/action_model': 2.3773456802276697e-05, 'epoch': 0.26}) +[2026-10-03 19:39:06,651][starVLA.training.train_starvla][INFO] - Step 3564, Loss: {'train/loss': 0.00023119484831113368, 'train/grad_norm_pre_clip': 0.08934791386127472, 'timing/data': 0.0002456899965181947, 'timing/model': 0.3927268830011599, 'learning_rate/qwen_vl_interface': 2.3749208950736783e-06, 'learning_rate/action_model': 2.374920895073678e-05, 'epoch': 0.26}) +[2026-10-03 19:39:07,357][starVLA.training.train_starvla][INFO] - Step 3565, Loss: {'train/loss': 3.283150363131426e-05, 'train/grad_norm_pre_clip': 0.020726874470710754, 'timing/data': 0.0002570899960119277, 'timing/model': 0.39528482302557677, 'learning_rate/qwen_vl_interface': 2.372497291756106e-06, 'learning_rate/action_model': 2.3724972917561058e-05, 'epoch': 0.26}) +[2026-10-03 19:39:08,064][starVLA.training.train_starvla][INFO] - Step 3566, Loss: {'train/loss': 7.681782881263644e-05, 'train/grad_norm_pre_clip': 0.022830188274383545, 'timing/data': 0.00041900001815520227, 'timing/model': 0.3929542129917536, 'learning_rate/qwen_vl_interface': 2.3700748712712043e-06, 'learning_rate/action_model': 2.370074871271204e-05, 'epoch': 0.26}) +[2026-10-03 19:39:08,776][starVLA.training.train_starvla][INFO] - Step 3567, Loss: {'train/loss': 0.0023736616130918264, 'train/grad_norm_pre_clip': 1.6130125522613525, 'timing/data': 0.00028929999098181725, 'timing/model': 0.39450456300983205, 'learning_rate/qwen_vl_interface': 2.3676536346147388e-06, 'learning_rate/action_model': 2.3676536346147388e-05, 'epoch': 0.26}) +[2026-10-03 19:39:09,483][starVLA.training.train_starvla][INFO] - Step 3568, Loss: {'train/loss': 5.182105815038085e-05, 'train/grad_norm_pre_clip': 0.01797768659889698, 'timing/data': 0.00023735998547635972, 'timing/model': 0.39362224299111404, 'learning_rate/qwen_vl_interface': 2.365233582781988e-06, 'learning_rate/action_model': 2.365233582781988e-05, 'epoch': 0.26}) +[2026-10-03 19:39:10,398][starVLA.training.train_starvla][INFO] - Step 3569, Loss: {'train/loss': 2.7564641641220078e-05, 'train/grad_norm_pre_clip': 0.017968380823731422, 'timing/data': 0.00033954999526031315, 'timing/model': 0.3935899630014319, 'learning_rate/qwen_vl_interface': 2.3628147167677444e-06, 'learning_rate/action_model': 2.3628147167677443e-05, 'epoch': 0.26}) +[2026-10-03 19:39:11,102][starVLA.training.train_starvla][INFO] - Step 3570, Loss: {'train/loss': 0.00018636527238413692, 'train/grad_norm_pre_clip': 0.35692375898361206, 'timing/data': 0.0003606200043577701, 'timing/model': 0.39294270201935433, 'learning_rate/qwen_vl_interface': 2.3603970375663123e-06, 'learning_rate/action_model': 2.3603970375663125e-05, 'epoch': 0.26}) +[2026-10-03 19:39:11,807][starVLA.training.train_starvla][INFO] - Step 3571, Loss: {'train/loss': 0.02862315997481346, 'train/grad_norm_pre_clip': 69.83456420898438, 'timing/data': 0.00024873000802472234, 'timing/model': 0.3933986730116885, 'learning_rate/qwen_vl_interface': 2.35798054617151e-06, 'learning_rate/action_model': 2.3579805461715096e-05, 'epoch': 0.26}) +[2026-10-03 19:39:12,515][starVLA.training.train_starvla][INFO] - Step 3572, Loss: {'train/loss': 0.00020407381816767156, 'train/grad_norm_pre_clip': 0.25113046169281006, 'timing/data': 0.00042580999433994293, 'timing/model': 0.3956226529844571, 'learning_rate/qwen_vl_interface': 2.355565243576662e-06, 'learning_rate/action_model': 2.3555652435766617e-05, 'epoch': 0.26}) +[2026-10-03 19:39:13,219][starVLA.training.train_starvla][INFO] - Step 3573, Loss: {'train/loss': 0.222030371427536, 'train/grad_norm_pre_clip': 313.2944030761719, 'timing/data': 0.0002654699783306569, 'timing/model': 0.394369341986021, 'learning_rate/qwen_vl_interface': 2.35315113077461e-06, 'learning_rate/action_model': 2.3531511307746098e-05, 'epoch': 0.26}) +[2026-10-03 19:39:14,128][starVLA.training.train_starvla][INFO] - Step 3574, Loss: {'train/loss': 0.3637068271636963, 'train/grad_norm_pre_clip': 156.73333740234375, 'timing/data': 0.0003583999932743609, 'timing/model': 0.5965618540067226, 'learning_rate/qwen_vl_interface': 2.3507382087577047e-06, 'learning_rate/action_model': 2.350738208757705e-05, 'epoch': 0.26}) +[2026-10-03 19:39:14,825][starVLA.training.train_starvla][INFO] - Step 3575, Loss: {'train/loss': 0.0011640606680884957, 'train/grad_norm_pre_clip': 1.455520749092102, 'timing/data': 0.00020193998352624476, 'timing/model': 0.38968660301179625, 'learning_rate/qwen_vl_interface': 2.348326478517807e-06, 'learning_rate/action_model': 2.3483264785178072e-05, 'epoch': 0.26}) +[2026-10-03 19:39:15,524][starVLA.training.train_starvla][INFO] - Step 3576, Loss: {'train/loss': 0.027239929884672165, 'train/grad_norm_pre_clip': 46.863990783691406, 'timing/data': 0.00020491000032052398, 'timing/model': 0.39127100200857967, 'learning_rate/qwen_vl_interface': 2.3459159410462907e-06, 'learning_rate/action_model': 2.3459159410462905e-05, 'epoch': 0.26}) +[2026-10-03 19:39:16,222][starVLA.training.train_starvla][INFO] - Step 3577, Loss: {'train/loss': 0.00015397617244161665, 'train/grad_norm_pre_clip': 0.19082480669021606, 'timing/data': 0.00021287001436576247, 'timing/model': 0.3901219239924103, 'learning_rate/qwen_vl_interface': 2.343506597334031e-06, 'learning_rate/action_model': 2.343506597334031e-05, 'epoch': 0.26}) +[2026-10-03 19:39:16,920][starVLA.training.train_starvla][INFO] - Step 3578, Loss: {'train/loss': 0.00016470885020680726, 'train/grad_norm_pre_clip': 0.1553640514612198, 'timing/data': 0.00041040999349206686, 'timing/model': 0.39085762298782356, 'learning_rate/qwen_vl_interface': 2.341098448371423e-06, 'learning_rate/action_model': 2.3410984483714227e-05, 'epoch': 0.26}) +[2026-10-03 19:39:17,822][starVLA.training.train_starvla][INFO] - Step 3579, Loss: {'train/loss': 8.260775939561427e-05, 'train/grad_norm_pre_clip': 0.03922102600336075, 'timing/data': 0.0003174500016029924, 'timing/model': 0.5954545130080078, 'learning_rate/qwen_vl_interface': 2.338691495148363e-06, 'learning_rate/action_model': 2.3386914951483626e-05, 'epoch': 0.26}) +[2026-10-03 19:39:18,518][starVLA.training.train_starvla][INFO] - Step 3580, Loss: {'train/loss': 1.93561445485102e-05, 'train/grad_norm_pre_clip': 0.003394651459529996, 'timing/data': 0.00035121000837534666, 'timing/model': 0.38914580398704857, 'learning_rate/qwen_vl_interface': 2.33628573865426e-06, 'learning_rate/action_model': 2.3362857386542596e-05, 'epoch': 0.26}) +[2026-10-03 19:39:19,216][starVLA.training.train_starvla][INFO] - Step 3581, Loss: {'train/loss': 3.8254893297562376e-05, 'train/grad_norm_pre_clip': 0.01667671650648117, 'timing/data': 0.00037833000533282757, 'timing/model': 0.39081162298680283, 'learning_rate/qwen_vl_interface': 2.3338811798780282e-06, 'learning_rate/action_model': 2.333881179878028e-05, 'epoch': 0.26}) +[2026-10-03 19:39:19,917][starVLA.training.train_starvla][INFO] - Step 3582, Loss: {'train/loss': 0.39774009585380554, 'train/grad_norm_pre_clip': 131.93881225585938, 'timing/data': 0.00021115998970344663, 'timing/model': 0.39147242199396715, 'learning_rate/qwen_vl_interface': 2.3314778198080933e-06, 'learning_rate/action_model': 2.331477819808093e-05, 'epoch': 0.26}) +[2026-10-03 19:39:20,615][starVLA.training.train_starvla][INFO] - Step 3583, Loss: {'train/loss': 2.240971298306249e-05, 'train/grad_norm_pre_clip': 0.01758246310055256, 'timing/data': 0.0003091600083280355, 'timing/model': 0.39054110398865305, 'learning_rate/qwen_vl_interface': 2.3290756594323834e-06, 'learning_rate/action_model': 2.3290756594323836e-05, 'epoch': 0.26}) +[2026-10-03 19:39:21,526][starVLA.training.train_starvla][INFO] - Step 3584, Loss: {'train/loss': 0.041451092809438705, 'train/grad_norm_pre_clip': 62.20275115966797, 'timing/data': 0.00035913000465370715, 'timing/model': 0.6017184919910505, 'learning_rate/qwen_vl_interface': 2.3266746997383378e-06, 'learning_rate/action_model': 2.3266746997383377e-05, 'epoch': 0.26}) +[2026-10-03 19:39:22,223][starVLA.training.train_starvla][INFO] - Step 3585, Loss: {'train/loss': 0.00034653174225240946, 'train/grad_norm_pre_clip': 0.7969905734062195, 'timing/data': 0.00020333999418653548, 'timing/model': 0.39040203401236795, 'learning_rate/qwen_vl_interface': 2.324274941712899e-06, 'learning_rate/action_model': 2.324274941712899e-05, 'epoch': 0.26}) +[2026-10-03 19:39:22,920][starVLA.training.train_starvla][INFO] - Step 3586, Loss: {'train/loss': 3.232631570426747e-05, 'train/grad_norm_pre_clip': 0.011418570764362812, 'timing/data': 0.00020235998090356588, 'timing/model': 0.38963598301052116, 'learning_rate/qwen_vl_interface': 2.321876386342519e-06, 'learning_rate/action_model': 2.3218763863425187e-05, 'epoch': 0.26}) +[2026-10-03 19:39:23,616][starVLA.training.train_starvla][INFO] - Step 3587, Loss: {'train/loss': 0.00026416531181894243, 'train/grad_norm_pre_clip': 0.30047211050987244, 'timing/data': 0.00039979000575840473, 'timing/model': 0.3879185730183963, 'learning_rate/qwen_vl_interface': 2.31947903461315e-06, 'learning_rate/action_model': 2.3194790346131502e-05, 'epoch': 0.26}) +[2026-10-03 19:39:24,317][starVLA.training.train_starvla][INFO] - Step 3588, Loss: {'train/loss': 0.00017395395843777806, 'train/grad_norm_pre_clip': 0.12366397678852081, 'timing/data': 0.00032130000181496143, 'timing/model': 0.390291674004402, 'learning_rate/qwen_vl_interface': 2.3170828875102544e-06, 'learning_rate/action_model': 2.3170828875102543e-05, 'epoch': 0.26}) +[2026-10-03 19:39:25,015][starVLA.training.train_starvla][INFO] - Step 3589, Loss: {'train/loss': 0.022259864956140518, 'train/grad_norm_pre_clip': 14.279363632202148, 'timing/data': 0.0002523600123822689, 'timing/model': 0.39061331399716437, 'learning_rate/qwen_vl_interface': 2.3146879460187984e-06, 'learning_rate/action_model': 2.3146879460187982e-05, 'epoch': 0.26}) +[2026-10-03 19:39:25,915][starVLA.training.train_starvla][INFO] - Step 3590, Loss: {'train/loss': 0.0018933687824755907, 'train/grad_norm_pre_clip': 3.5335581302642822, 'timing/data': 0.00036194000858813524, 'timing/model': 0.3901147530123126, 'learning_rate/qwen_vl_interface': 2.3122942111232506e-06, 'learning_rate/action_model': 2.3122942111232506e-05, 'epoch': 0.26}) +[2026-10-03 19:39:26,612][starVLA.training.train_starvla][INFO] - Step 3591, Loss: {'train/loss': 5.4241230827756226e-05, 'train/grad_norm_pre_clip': 0.025603918358683586, 'timing/data': 0.00030848997994326055, 'timing/model': 0.3899062230193522, 'learning_rate/qwen_vl_interface': 2.3099016838075857e-06, 'learning_rate/action_model': 2.3099016838075856e-05, 'epoch': 0.26}) +[2026-10-03 19:39:27,309][starVLA.training.train_starvla][INFO] - Step 3592, Loss: {'train/loss': 0.0184915941208601, 'train/grad_norm_pre_clip': 35.937381744384766, 'timing/data': 0.0003034999826923013, 'timing/model': 0.38939803300308995, 'learning_rate/qwen_vl_interface': 2.307510365055282e-06, 'learning_rate/action_model': 2.307510365055282e-05, 'epoch': 0.26}) +[2026-10-03 19:39:28,004][starVLA.training.train_starvla][INFO] - Step 3593, Loss: {'train/loss': 2.666439104359597e-05, 'train/grad_norm_pre_clip': 0.009733340702950954, 'timing/data': 0.0002056900120805949, 'timing/model': 0.38886639301199466, 'learning_rate/qwen_vl_interface': 2.3051202558493196e-06, 'learning_rate/action_model': 2.3051202558493192e-05, 'epoch': 0.26}) +[2026-10-03 19:39:28,702][starVLA.training.train_starvla][INFO] - Step 3594, Loss: {'train/loss': 6.136453157523647e-05, 'train/grad_norm_pre_clip': 0.04405888170003891, 'timing/data': 0.00026614998932927847, 'timing/model': 0.3891455830016639, 'learning_rate/qwen_vl_interface': 2.302731357172182e-06, 'learning_rate/action_model': 2.302731357172182e-05, 'epoch': 0.26}) +[2026-10-03 19:39:29,598][starVLA.training.train_starvla][INFO] - Step 3595, Loss: {'train/loss': 6.880919681861997e-05, 'train/grad_norm_pre_clip': 0.03534961864352226, 'timing/data': 0.00024530000519007444, 'timing/model': 0.5900413939962164, 'learning_rate/qwen_vl_interface': 2.3003436700058556e-06, 'learning_rate/action_model': 2.3003436700058554e-05, 'epoch': 0.26}) +[2026-10-03 19:39:30,293][starVLA.training.train_starvla][INFO] - Step 3596, Loss: {'train/loss': 1.3418252819974441e-05, 'train/grad_norm_pre_clip': 0.0028019805904477835, 'timing/data': 0.0002602300082799047, 'timing/model': 0.3892861630010884, 'learning_rate/qwen_vl_interface': 2.297957195331828e-06, 'learning_rate/action_model': 2.2979571953318276e-05, 'epoch': 0.26}) +[2026-10-03 19:39:30,990][starVLA.training.train_starvla][INFO] - Step 3597, Loss: {'train/loss': 0.0016947820549830794, 'train/grad_norm_pre_clip': 3.426758289337158, 'timing/data': 0.0002465900033712387, 'timing/model': 0.3911011930031236, 'learning_rate/qwen_vl_interface': 2.2955719341310903e-06, 'learning_rate/action_model': 2.2955719341310903e-05, 'epoch': 0.26}) +[2026-10-03 19:39:31,688][starVLA.training.train_starvla][INFO] - Step 3598, Loss: {'train/loss': 2.5985640604631044e-05, 'train/grad_norm_pre_clip': 0.02752009779214859, 'timing/data': 0.00034917000448331237, 'timing/model': 0.3893749340204522, 'learning_rate/qwen_vl_interface': 2.2931878873841304e-06, 'learning_rate/action_model': 2.2931878873841302e-05, 'epoch': 0.26}) +[2026-10-03 19:39:32,384][starVLA.training.train_starvla][INFO] - Step 3599, Loss: {'train/loss': 0.375343382358551, 'train/grad_norm_pre_clip': 50.81404113769531, 'timing/data': 0.00022192997857928276, 'timing/model': 0.39031751299626194, 'learning_rate/qwen_vl_interface': 2.290805056070942e-06, 'learning_rate/action_model': 2.290805056070942e-05, 'epoch': 0.26}) +[2026-10-03 19:39:33,080][starVLA.training.train_starvla][INFO] - Step 3600, Loss: {'train/loss': 5.919523391639814e-05, 'train/grad_norm_pre_clip': 0.023948898538947105, 'timing/data': 0.00024662999203428626, 'timing/model': 0.3905881430255249, 'learning_rate/qwen_vl_interface': 2.288423441171016e-06, 'learning_rate/action_model': 2.288423441171016e-05, 'epoch': 0.26}) +[2026-10-03 19:39:33,981][starVLA.training.train_starvla][INFO] - Step 3601, Loss: {'train/loss': 0.00010598695371299982, 'train/grad_norm_pre_clip': 0.05722005292773247, 'timing/data': 0.0002554900129325688, 'timing/model': 0.3918455030070618, 'learning_rate/qwen_vl_interface': 2.2860430436633447e-06, 'learning_rate/action_model': 2.2860430436633447e-05, 'epoch': 0.26}) +[2026-10-03 19:39:34,677][starVLA.training.train_starvla][INFO] - Step 3602, Loss: {'train/loss': 0.7608237862586975, 'train/grad_norm_pre_clip': 30.83600616455078, 'timing/data': 0.00020710000535473228, 'timing/model': 0.38939757298794575, 'learning_rate/qwen_vl_interface': 2.28366386452642e-06, 'learning_rate/action_model': 2.28366386452642e-05, 'epoch': 0.26}) +[2026-10-03 19:39:35,371][starVLA.training.train_starvla][INFO] - Step 3603, Loss: {'train/loss': 0.8917872905731201, 'train/grad_norm_pre_clip': 210.7704620361328, 'timing/data': 0.00020248000510036945, 'timing/model': 0.38848981299088337, 'learning_rate/qwen_vl_interface': 2.2812859047382326e-06, 'learning_rate/action_model': 2.2812859047382323e-05, 'epoch': 0.26}) +[2026-10-03 19:39:36,068][starVLA.training.train_starvla][INFO] - Step 3604, Loss: {'train/loss': 0.00022160340449772775, 'train/grad_norm_pre_clip': 0.3306821584701538, 'timing/data': 0.00024226997629739344, 'timing/model': 0.3905125329911243, 'learning_rate/qwen_vl_interface': 2.2789091652762712e-06, 'learning_rate/action_model': 2.2789091652762712e-05, 'epoch': 0.26}) +[2026-10-03 19:39:36,765][starVLA.training.train_starvla][INFO] - Step 3605, Loss: {'train/loss': 0.0015629074769094586, 'train/grad_norm_pre_clip': 6.053590297698975, 'timing/data': 0.00039608997758477926, 'timing/model': 0.39015443401876837, 'learning_rate/qwen_vl_interface': 2.276533647117524e-06, 'learning_rate/action_model': 2.2765336471175242e-05, 'epoch': 0.26}) +[2026-10-03 19:39:37,663][starVLA.training.train_starvla][INFO] - Step 3606, Loss: {'train/loss': 0.0002826685959007591, 'train/grad_norm_pre_clip': 0.1824423372745514, 'timing/data': 0.00020297997980378568, 'timing/model': 0.3889171839982737, 'learning_rate/qwen_vl_interface': 2.2741593512384773e-06, 'learning_rate/action_model': 2.2741593512384772e-05, 'epoch': 0.26}) +[2026-10-03 19:39:38,360][starVLA.training.train_starvla][INFO] - Step 3607, Loss: {'train/loss': 4.2174768168479204e-05, 'train/grad_norm_pre_clip': 0.02276010997593403, 'timing/data': 0.00023766999947838485, 'timing/model': 0.39015583400032483, 'learning_rate/qwen_vl_interface': 2.2717862786151153e-06, 'learning_rate/action_model': 2.271786278615115e-05, 'epoch': 0.26}) +[2026-10-03 19:39:39,057][starVLA.training.train_starvla][INFO] - Step 3608, Loss: {'train/loss': 0.0011117964750155807, 'train/grad_norm_pre_clip': 0.8734262585639954, 'timing/data': 0.00028300000121816993, 'timing/model': 0.39101700301398523, 'learning_rate/qwen_vl_interface': 2.2694144302229152e-06, 'learning_rate/action_model': 2.2694144302229152e-05, 'epoch': 0.26}) +[2026-10-03 19:39:39,754][starVLA.training.train_starvla][INFO] - Step 3609, Loss: {'train/loss': 0.0002118675474775955, 'train/grad_norm_pre_clip': 0.3959008455276489, 'timing/data': 0.00031152000883594155, 'timing/model': 0.39051972297602333, 'learning_rate/qwen_vl_interface': 2.2670438070368568e-06, 'learning_rate/action_model': 2.2670438070368566e-05, 'epoch': 0.26}) +[2026-10-03 19:39:40,451][starVLA.training.train_starvla][INFO] - Step 3610, Loss: {'train/loss': 0.0009440529975108802, 'train/grad_norm_pre_clip': 1.6613223552703857, 'timing/data': 0.0003174399898853153, 'timing/model': 0.3901994320040103, 'learning_rate/qwen_vl_interface': 2.264674410031413e-06, 'learning_rate/action_model': 2.2646744100314127e-05, 'epoch': 0.26}) +[2026-10-03 19:39:41,349][starVLA.training.train_starvla][INFO] - Step 3611, Loss: {'train/loss': 0.00046858208952471614, 'train/grad_norm_pre_clip': 0.4406564235687256, 'timing/data': 0.00020974999642930925, 'timing/model': 0.592214103002334, 'learning_rate/qwen_vl_interface': 2.2623062401805537e-06, 'learning_rate/action_model': 2.2623062401805535e-05, 'epoch': 0.26}) +[2026-10-03 19:39:42,044][starVLA.training.train_starvla][INFO] - Step 3612, Loss: {'train/loss': 4.222617644700222e-05, 'train/grad_norm_pre_clip': 0.010006285272538662, 'timing/data': 0.0002556100080255419, 'timing/model': 0.3897144430084154, 'learning_rate/qwen_vl_interface': 2.2599392984577436e-06, 'learning_rate/action_model': 2.2599392984577438e-05, 'epoch': 0.26}) +[2026-10-03 19:39:42,740][starVLA.training.train_starvla][INFO] - Step 3613, Loss: {'train/loss': 0.00021917051344644278, 'train/grad_norm_pre_clip': 0.17182187736034393, 'timing/data': 0.00033368999720551074, 'timing/model': 0.39038088300731033, 'learning_rate/qwen_vl_interface': 2.2575735858359436e-06, 'learning_rate/action_model': 2.257573585835944e-05, 'epoch': 0.26}) +[2026-10-03 19:39:43,439][starVLA.training.train_starvla][INFO] - Step 3614, Loss: {'train/loss': 0.0004049318376928568, 'train/grad_norm_pre_clip': 1.6035867929458618, 'timing/data': 0.000319099985063076, 'timing/model': 0.3908027329889592, 'learning_rate/qwen_vl_interface': 2.255209103287609e-06, 'learning_rate/action_model': 2.255209103287609e-05, 'epoch': 0.26}) +[2026-10-03 19:39:44,136][starVLA.training.train_starvla][INFO] - Step 3615, Loss: {'train/loss': 4.6363107685465366e-05, 'train/grad_norm_pre_clip': 0.06599954515695572, 'timing/data': 0.00020154000958427787, 'timing/model': 0.38999814400449395, 'learning_rate/qwen_vl_interface': 2.2528458517846886e-06, 'learning_rate/action_model': 2.252845851784689e-05, 'epoch': 0.26}) +[2026-10-03 19:39:44,833][starVLA.training.train_starvla][INFO] - Step 3616, Loss: {'train/loss': 0.0001811720139812678, 'train/grad_norm_pre_clip': 0.09874261170625687, 'timing/data': 0.0003045200137421489, 'timing/model': 0.39049178300774656, 'learning_rate/qwen_vl_interface': 2.250483832298627e-06, 'learning_rate/action_model': 2.250483832298627e-05, 'epoch': 0.26}) +[2026-10-03 19:39:45,736][starVLA.training.train_starvla][INFO] - Step 3617, Loss: {'train/loss': 3.608068072935566e-05, 'train/grad_norm_pre_clip': 0.011840499937534332, 'timing/data': 0.00039646998629905283, 'timing/model': 0.3919329430209473, 'learning_rate/qwen_vl_interface': 2.2481230458003608e-06, 'learning_rate/action_model': 2.2481230458003606e-05, 'epoch': 0.26}) +[2026-10-03 19:39:46,436][starVLA.training.train_starvla][INFO] - Step 3618, Loss: {'train/loss': 3.4806347684934735e-05, 'train/grad_norm_pre_clip': 0.01758931763470173, 'timing/data': 0.0002137300034519285, 'timing/model': 0.391194692987483, 'learning_rate/qwen_vl_interface': 2.2457634932603184e-06, 'learning_rate/action_model': 2.2457634932603183e-05, 'epoch': 0.26}) +[2026-10-03 19:39:47,135][starVLA.training.train_starvla][INFO] - Step 3619, Loss: {'train/loss': 2.0264764316380024e-05, 'train/grad_norm_pre_clip': 0.00813804566860199, 'timing/data': 0.00029754001297988, 'timing/model': 0.3904787430074066, 'learning_rate/qwen_vl_interface': 2.2434051756484255e-06, 'learning_rate/action_model': 2.2434051756484253e-05, 'epoch': 0.26}) +[2026-10-03 19:39:47,829][starVLA.training.train_starvla][INFO] - Step 3620, Loss: {'train/loss': 6.011134246364236e-05, 'train/grad_norm_pre_clip': 0.03423561900854111, 'timing/data': 0.00030236999737098813, 'timing/model': 0.3898417629825417, 'learning_rate/qwen_vl_interface': 2.2410480939340945e-06, 'learning_rate/action_model': 2.2410480939340945e-05, 'epoch': 0.26}) +[2026-10-03 19:39:48,527][starVLA.training.train_starvla][INFO] - Step 3621, Loss: {'train/loss': 0.3433167636394501, 'train/grad_norm_pre_clip': 88.56114959716797, 'timing/data': 0.00024639000184834003, 'timing/model': 0.39084087300579995, 'learning_rate/qwen_vl_interface': 2.238692249086234e-06, 'learning_rate/action_model': 2.2386922490862336e-05, 'epoch': 0.27}) +[2026-10-03 19:39:49,428][starVLA.training.train_starvla][INFO] - Step 3622, Loss: {'train/loss': 0.0005317626637406647, 'train/grad_norm_pre_clip': 0.6015580892562866, 'timing/data': 0.00024046999169513583, 'timing/model': 0.3899106119934004, 'learning_rate/qwen_vl_interface': 2.2363376420732426e-06, 'learning_rate/action_model': 2.2363376420732423e-05, 'epoch': 0.27}) +[2026-10-03 19:39:50,124][starVLA.training.train_starvla][INFO] - Step 3623, Loss: {'train/loss': 0.40015023946762085, 'train/grad_norm_pre_clip': 65.38365936279297, 'timing/data': 0.00034383998718112707, 'timing/model': 0.3899474930076394, 'learning_rate/qwen_vl_interface': 2.233984273863011e-06, 'learning_rate/action_model': 2.233984273863011e-05, 'epoch': 0.27}) +[2026-10-03 19:39:50,820][starVLA.training.train_starvla][INFO] - Step 3624, Loss: {'train/loss': 3.528384695528075e-05, 'train/grad_norm_pre_clip': 0.012146681547164917, 'timing/data': 0.0002340200007893145, 'timing/model': 0.39154851299826987, 'learning_rate/qwen_vl_interface': 2.231632145422919e-06, 'learning_rate/action_model': 2.231632145422919e-05, 'epoch': 0.27}) +[2026-10-03 19:39:51,516][starVLA.training.train_starvla][INFO] - Step 3625, Loss: {'train/loss': 0.44574299454689026, 'train/grad_norm_pre_clip': 64.18498229980469, 'timing/data': 0.0002067900204565376, 'timing/model': 0.3905239430023357, 'learning_rate/qwen_vl_interface': 2.2292812577198383e-06, 'learning_rate/action_model': 2.229281257719838e-05, 'epoch': 0.27}) +[2026-10-03 19:39:52,214][starVLA.training.train_starvla][INFO] - Step 3626, Loss: {'train/loss': 0.5481470227241516, 'train/grad_norm_pre_clip': 213.01626586914062, 'timing/data': 0.00033407998853363097, 'timing/model': 0.3903255829936825, 'learning_rate/qwen_vl_interface': 2.2269316117201307e-06, 'learning_rate/action_model': 2.2269316117201307e-05, 'epoch': 0.27}) +[2026-10-03 19:39:53,110][starVLA.training.train_starvla][INFO] - Step 3627, Loss: {'train/loss': 7.981267845025286e-05, 'train/grad_norm_pre_clip': 0.028155319392681122, 'timing/data': 0.0003036099951714277, 'timing/model': 0.5906764440005645, 'learning_rate/qwen_vl_interface': 2.224583208389646e-06, 'learning_rate/action_model': 2.224583208389646e-05, 'epoch': 0.27}) +[2026-10-03 19:39:53,806][starVLA.training.train_starvla][INFO] - Step 3628, Loss: {'train/loss': 4.81763418065384e-05, 'train/grad_norm_pre_clip': 0.03854449465870857, 'timing/data': 0.00024244000087492168, 'timing/model': 0.3895551130117383, 'learning_rate/qwen_vl_interface': 2.2222360486937238e-06, 'learning_rate/action_model': 2.2222360486937235e-05, 'epoch': 0.27}) +[2026-10-03 19:39:54,505][starVLA.training.train_starvla][INFO] - Step 3629, Loss: {'train/loss': 0.4735579788684845, 'train/grad_norm_pre_clip': 61.41532516479492, 'timing/data': 0.00028863997431471944, 'timing/model': 0.39120805301354267, 'learning_rate/qwen_vl_interface': 2.219890133597194e-06, 'learning_rate/action_model': 2.219890133597194e-05, 'epoch': 0.27}) +[2026-10-03 19:39:55,205][starVLA.training.train_starvla][INFO] - Step 3630, Loss: {'train/loss': 0.00012078104191459715, 'train/grad_norm_pre_clip': 0.10125315189361572, 'timing/data': 0.00020441997912712395, 'timing/model': 0.38951867300784215, 'learning_rate/qwen_vl_interface': 2.217545464064374e-06, 'learning_rate/action_model': 2.217545464064374e-05, 'epoch': 0.27}) +[2026-10-03 19:39:55,902][starVLA.training.train_starvla][INFO] - Step 3631, Loss: {'train/loss': 2.9972576157888398e-05, 'train/grad_norm_pre_clip': 0.010052424855530262, 'timing/data': 0.0003638800117187202, 'timing/model': 0.3905717229936272, 'learning_rate/qwen_vl_interface': 2.215202041059067e-06, 'learning_rate/action_model': 2.215202041059067e-05, 'epoch': 0.27}) +[2026-10-03 19:39:56,798][starVLA.training.train_starvla][INFO] - Step 3632, Loss: {'train/loss': 0.048104360699653625, 'train/grad_norm_pre_clip': 100.12606048583984, 'timing/data': 0.00025275000371038914, 'timing/model': 0.5902850140118971, 'learning_rate/qwen_vl_interface': 2.212859865544566e-06, 'learning_rate/action_model': 2.212859865544566e-05, 'epoch': 0.27}) +[2026-10-03 19:39:57,498][starVLA.training.train_starvla][INFO] - Step 3633, Loss: {'train/loss': 0.00019188561418559402, 'train/grad_norm_pre_clip': 0.1610344499349594, 'timing/data': 0.0002548600023146719, 'timing/model': 0.39281023302464746, 'learning_rate/qwen_vl_interface': 2.210518938483652e-06, 'learning_rate/action_model': 2.210518938483652e-05, 'epoch': 0.27}) +[2026-10-03 19:39:58,200][starVLA.training.train_starvla][INFO] - Step 3634, Loss: {'train/loss': 0.0017212892416864634, 'train/grad_norm_pre_clip': 6.724288463592529, 'timing/data': 0.0002467099984642118, 'timing/model': 0.3914217120036483, 'learning_rate/qwen_vl_interface': 2.20817926083859e-06, 'learning_rate/action_model': 2.20817926083859e-05, 'epoch': 0.27}) +[2026-10-03 19:39:58,898][starVLA.training.train_starvla][INFO] - Step 3635, Loss: {'train/loss': 0.7994565367698669, 'train/grad_norm_pre_clip': 82.99925231933594, 'timing/data': 0.00021117000142112374, 'timing/model': 0.39046665298519656, 'learning_rate/qwen_vl_interface': 2.205840833571135e-06, 'learning_rate/action_model': 2.205840833571135e-05, 'epoch': 0.27}) +[2026-10-03 19:39:59,597][starVLA.training.train_starvla][INFO] - Step 3636, Loss: {'train/loss': 0.0008068200550042093, 'train/grad_norm_pre_clip': 1.0200623273849487, 'timing/data': 0.00020570997730828822, 'timing/model': 0.39152777299750596, 'learning_rate/qwen_vl_interface': 2.2035036576425264e-06, 'learning_rate/action_model': 2.203503657642526e-05, 'epoch': 0.27}) +[2026-10-03 19:40:00,494][starVLA.training.train_starvla][INFO] - Step 3637, Loss: {'train/loss': 0.48260489106178284, 'train/grad_norm_pre_clip': 69.50037384033203, 'timing/data': 0.00037836001138202846, 'timing/model': 0.58843919399078, 'learning_rate/qwen_vl_interface': 2.2011677340134867e-06, 'learning_rate/action_model': 2.2011677340134867e-05, 'epoch': 0.27}) +[2026-10-03 19:40:01,193][starVLA.training.train_starvla][INFO] - Step 3638, Loss: {'train/loss': 0.04340456798672676, 'train/grad_norm_pre_clip': 71.33429718017578, 'timing/data': 0.0002474699867889285, 'timing/model': 0.3915920730214566, 'learning_rate/qwen_vl_interface': 2.198833063644227e-06, 'learning_rate/action_model': 2.198833063644227e-05, 'epoch': 0.27}) +[2026-10-03 19:40:01,890][starVLA.training.train_starvla][INFO] - Step 3639, Loss: {'train/loss': 0.4062257409095764, 'train/grad_norm_pre_clip': 169.76417541503906, 'timing/data': 0.00031750998459756374, 'timing/model': 0.38945588300703093, 'learning_rate/qwen_vl_interface': 2.196499647494443e-06, 'learning_rate/action_model': 2.196499647494443e-05, 'epoch': 0.27}) +[2026-10-03 19:40:02,589][starVLA.training.train_starvla][INFO] - Step 3640, Loss: {'train/loss': 0.0003851760702673346, 'train/grad_norm_pre_clip': 0.5116719603538513, 'timing/data': 0.00033668000833131373, 'timing/model': 0.39144223200855777, 'learning_rate/qwen_vl_interface': 2.1941674865233144e-06, 'learning_rate/action_model': 2.1941674865233143e-05, 'epoch': 0.27}) +[2026-10-03 19:40:03,285][starVLA.training.train_starvla][INFO] - Step 3641, Loss: {'train/loss': 0.1370890587568283, 'train/grad_norm_pre_clip': 217.45608520507812, 'timing/data': 0.0003655990003608167, 'timing/model': 0.3896438730007503, 'learning_rate/qwen_vl_interface': 2.1918365816895013e-06, 'learning_rate/action_model': 2.191836581689501e-05, 'epoch': 0.27}) +[2026-10-03 19:40:03,984][starVLA.training.train_starvla][INFO] - Step 3642, Loss: {'train/loss': 0.0008538140100426972, 'train/grad_norm_pre_clip': 1.3060352802276611, 'timing/data': 0.000318089994834736, 'timing/model': 0.39130027301143855, 'learning_rate/qwen_vl_interface': 2.189506933951153e-06, 'learning_rate/action_model': 2.1895069339511525e-05, 'epoch': 0.27}) +[2026-10-03 19:40:04,885][starVLA.training.train_starvla][INFO] - Step 3643, Loss: {'train/loss': 5.3443320211954415e-05, 'train/grad_norm_pre_clip': 0.02119467779994011, 'timing/data': 0.00032682999153621495, 'timing/model': 0.38986190300784074, 'learning_rate/qwen_vl_interface': 2.187178544265898e-06, 'learning_rate/action_model': 2.1871785442658983e-05, 'epoch': 0.27}) +[2026-10-03 19:40:05,582][starVLA.training.train_starvla][INFO] - Step 3644, Loss: {'train/loss': 2.4846864107530564e-05, 'train/grad_norm_pre_clip': 0.004813373554497957, 'timing/data': 0.00033070999779738486, 'timing/model': 0.3897241429949645, 'learning_rate/qwen_vl_interface': 2.184851413590852e-06, 'learning_rate/action_model': 2.1848514135908518e-05, 'epoch': 0.27}) +[2026-10-03 19:40:06,278][starVLA.training.train_starvla][INFO] - Step 3645, Loss: {'train/loss': 0.00025528663536533713, 'train/grad_norm_pre_clip': 0.469198614358902, 'timing/data': 0.0003507200162857771, 'timing/model': 0.38988272298593074, 'learning_rate/qwen_vl_interface': 2.1825255428826087e-06, 'learning_rate/action_model': 2.182525542882609e-05, 'epoch': 0.27}) +[2026-10-03 19:40:06,980][starVLA.training.train_starvla][INFO] - Step 3646, Loss: {'train/loss': 0.011474594473838806, 'train/grad_norm_pre_clip': 25.780635833740234, 'timing/data': 0.00020785999367944896, 'timing/model': 0.39176520198816434, 'learning_rate/qwen_vl_interface': 2.180200933097246e-06, 'learning_rate/action_model': 2.180200933097246e-05, 'epoch': 0.27}) +[2026-10-03 19:40:07,679][starVLA.training.train_starvla][INFO] - Step 3647, Loss: {'train/loss': 0.41913995146751404, 'train/grad_norm_pre_clip': 45.71455001831055, 'timing/data': 0.0002028800081461668, 'timing/model': 0.3911332929856144, 'learning_rate/qwen_vl_interface': 2.1778775851903235e-06, 'learning_rate/action_model': 2.177877585190323e-05, 'epoch': 0.27}) +[2026-10-03 19:40:08,576][starVLA.training.train_starvla][INFO] - Step 3648, Loss: {'train/loss': 0.014098802581429482, 'train/grad_norm_pre_clip': 48.29910659790039, 'timing/data': 0.00021433000802062452, 'timing/model': 0.5888363039994147, 'learning_rate/qwen_vl_interface': 2.175555500116881e-06, 'learning_rate/action_model': 2.175555500116881e-05, 'epoch': 0.27}) +[2026-10-03 19:40:09,273][starVLA.training.train_starvla][INFO] - Step 3649, Loss: {'train/loss': 0.25496694445610046, 'train/grad_norm_pre_clip': 109.6278076171875, 'timing/data': 0.0002043300191871822, 'timing/model': 0.3898719330027234, 'learning_rate/qwen_vl_interface': 2.1732346788314404e-06, 'learning_rate/action_model': 2.1732346788314402e-05, 'epoch': 0.27}) +[2026-10-03 19:40:09,977][starVLA.training.train_starvla][INFO] - Step 3650, Loss: {'train/loss': 0.05245712026953697, 'train/grad_norm_pre_clip': 106.4576644897461, 'timing/data': 0.0003628599806688726, 'timing/model': 0.3933827219880186, 'learning_rate/qwen_vl_interface': 2.170915122288006e-06, 'learning_rate/action_model': 2.170915122288006e-05, 'epoch': 0.27}) +[2026-10-03 19:40:10,676][starVLA.training.train_starvla][INFO] - Step 3651, Loss: {'train/loss': 0.9585624933242798, 'train/grad_norm_pre_clip': 386.85491943359375, 'timing/data': 0.00046671999734826386, 'timing/model': 0.39105879300041124, 'learning_rate/qwen_vl_interface': 2.168596831440055e-06, 'learning_rate/action_model': 2.1685968314400552e-05, 'epoch': 0.27}) +[2026-10-03 19:40:11,373][starVLA.training.train_starvla][INFO] - Step 3652, Loss: {'train/loss': 0.5897908806800842, 'train/grad_norm_pre_clip': 37.29372787475586, 'timing/data': 0.00022177002392709255, 'timing/model': 0.3912932329985779, 'learning_rate/qwen_vl_interface': 2.166279807240553e-06, 'learning_rate/action_model': 2.1662798072405528e-05, 'epoch': 0.27}) +[2026-10-03 19:40:12,073][starVLA.training.train_starvla][INFO] - Step 3653, Loss: {'train/loss': 3.7221536331344396e-05, 'train/grad_norm_pre_clip': 0.013600556179881096, 'timing/data': 0.0004402699996717274, 'timing/model': 0.3914353420259431, 'learning_rate/qwen_vl_interface': 2.1639640506419394e-06, 'learning_rate/action_model': 2.163964050641939e-05, 'epoch': 0.27}) +[2026-10-03 19:40:12,970][starVLA.training.train_starvla][INFO] - Step 3654, Loss: {'train/loss': 4.084921965841204e-05, 'train/grad_norm_pre_clip': 0.010680267587304115, 'timing/data': 0.00021831001504324377, 'timing/model': 0.39038588298717514, 'learning_rate/qwen_vl_interface': 2.161649562596135e-06, 'learning_rate/action_model': 2.161649562596135e-05, 'epoch': 0.27}) +[2026-10-03 19:40:13,666][starVLA.training.train_starvla][INFO] - Step 3655, Loss: {'train/loss': 0.00010455230221850798, 'train/grad_norm_pre_clip': 0.048278216272592545, 'timing/data': 0.00031703998683951795, 'timing/model': 0.38869571301620454, 'learning_rate/qwen_vl_interface': 2.1593363440545383e-06, 'learning_rate/action_model': 2.1593363440545384e-05, 'epoch': 0.27}) +[2026-10-03 19:40:14,361][starVLA.training.train_starvla][INFO] - Step 3656, Loss: {'train/loss': 0.3429659307003021, 'train/grad_norm_pre_clip': 80.04228973388672, 'timing/data': 0.0003417499829083681, 'timing/model': 0.38918518301215954, 'learning_rate/qwen_vl_interface': 2.157024395968025e-06, 'learning_rate/action_model': 2.1570243959680252e-05, 'epoch': 0.27}) +[2026-10-03 19:40:15,058][starVLA.training.train_starvla][INFO] - Step 3657, Loss: {'train/loss': 3.081453905906528e-05, 'train/grad_norm_pre_clip': 0.007006218656897545, 'timing/data': 0.00020776997553184628, 'timing/model': 0.3905170730140526, 'learning_rate/qwen_vl_interface': 2.154713719286952e-06, 'learning_rate/action_model': 2.154713719286952e-05, 'epoch': 0.27}) +[2026-10-03 19:40:15,758][starVLA.training.train_starvla][INFO] - Step 3658, Loss: {'train/loss': 0.01936403289437294, 'train/grad_norm_pre_clip': 16.704830169677734, 'timing/data': 0.00030156999127939343, 'timing/model': 0.391490773006808, 'learning_rate/qwen_vl_interface': 2.152404314961149e-06, 'learning_rate/action_model': 2.152404314961149e-05, 'epoch': 0.27}) +[2026-10-03 19:40:16,657][starVLA.training.train_starvla][INFO] - Step 3659, Loss: {'train/loss': 0.006054736208170652, 'train/grad_norm_pre_clip': 20.69689178466797, 'timing/data': 0.00034924002829939127, 'timing/model': 0.388852684001904, 'learning_rate/qwen_vl_interface': 2.1500961839399264e-06, 'learning_rate/action_model': 2.1500961839399262e-05, 'epoch': 0.27}) +[2026-10-03 19:40:17,353][starVLA.training.train_starvla][INFO] - Step 3660, Loss: {'train/loss': 0.0001614507636986673, 'train/grad_norm_pre_clip': 0.07498334348201752, 'timing/data': 0.0003107899974565953, 'timing/model': 0.39083019297686405, 'learning_rate/qwen_vl_interface': 2.1477893271720683e-06, 'learning_rate/action_model': 2.1477893271720682e-05, 'epoch': 0.27}) +[2026-10-03 19:40:18,049][starVLA.training.train_starvla][INFO] - Step 3661, Loss: {'train/loss': 5.757030885433778e-05, 'train/grad_norm_pre_clip': 0.02148616500198841, 'timing/data': 0.0002393400063738227, 'timing/model': 0.3893293329747394, 'learning_rate/qwen_vl_interface': 2.1454837456058385e-06, 'learning_rate/action_model': 2.1454837456058386e-05, 'epoch': 0.27}) +[2026-10-03 19:40:18,747][starVLA.training.train_starvla][INFO] - Step 3662, Loss: {'train/loss': 0.027326339855790138, 'train/grad_norm_pre_clip': 78.76937103271484, 'timing/data': 0.0003428900090511888, 'timing/model': 0.39136614301241934, 'learning_rate/qwen_vl_interface': 2.1431794401889722e-06, 'learning_rate/action_model': 2.143179440188972e-05, 'epoch': 0.27}) +[2026-10-03 19:40:19,442][starVLA.training.train_starvla][INFO] - Step 3663, Loss: {'train/loss': 0.28260135650634766, 'train/grad_norm_pre_clip': 76.15047454833984, 'timing/data': 0.0002949299814645201, 'timing/model': 0.38888059300370514, 'learning_rate/qwen_vl_interface': 2.1408764118686837e-06, 'learning_rate/action_model': 2.1408764118686838e-05, 'epoch': 0.27}) +[2026-10-03 19:40:20,340][starVLA.training.train_starvla][INFO] - Step 3664, Loss: {'train/loss': 1.9617000361904502e-05, 'train/grad_norm_pre_clip': 0.004228958394378424, 'timing/data': 0.00020758999744430184, 'timing/model': 0.5911065840045922, 'learning_rate/qwen_vl_interface': 2.1385746615916598e-06, 'learning_rate/action_model': 2.1385746615916596e-05, 'epoch': 0.27}) +[2026-10-03 19:40:21,038][starVLA.training.train_starvla][INFO] - Step 3665, Loss: {'train/loss': 0.00037627099663950503, 'train/grad_norm_pre_clip': 0.2212892472743988, 'timing/data': 0.00029185900348238647, 'timing/model': 0.3909999930183403, 'learning_rate/qwen_vl_interface': 2.1362741903040636e-06, 'learning_rate/action_model': 2.1362741903040633e-05, 'epoch': 0.27}) +[2026-10-03 19:40:21,735][starVLA.training.train_starvla][INFO] - Step 3666, Loss: {'train/loss': 0.0005499423132278025, 'train/grad_norm_pre_clip': 1.0382353067398071, 'timing/data': 0.0002315599995199591, 'timing/model': 0.38948299302137457, 'learning_rate/qwen_vl_interface': 2.1339749989515318e-06, 'learning_rate/action_model': 2.1339749989515316e-05, 'epoch': 0.27}) +[2026-10-03 19:40:22,435][starVLA.training.train_starvla][INFO] - Step 3667, Loss: {'train/loss': 0.00013503333320841193, 'train/grad_norm_pre_clip': 0.07546643167734146, 'timing/data': 0.0003448299830779433, 'timing/model': 0.3911903730186168, 'learning_rate/qwen_vl_interface': 2.131677088479175e-06, 'learning_rate/action_model': 2.1316770884791752e-05, 'epoch': 0.27}) +[2026-10-03 19:40:23,133][starVLA.training.train_starvla][INFO] - Step 3668, Loss: {'train/loss': 2.761122232186608e-05, 'train/grad_norm_pre_clip': 0.005616630893200636, 'timing/data': 0.00020852999296039343, 'timing/model': 0.38957439301884733, 'learning_rate/qwen_vl_interface': 2.1293804598315787e-06, 'learning_rate/action_model': 2.1293804598315787e-05, 'epoch': 0.27}) +[2026-10-03 19:40:23,834][starVLA.training.train_starvla][INFO] - Step 3669, Loss: {'train/loss': 0.00013990039587952197, 'train/grad_norm_pre_clip': 0.07104542851448059, 'timing/data': 0.000284650013782084, 'timing/model': 0.39351291200728156, 'learning_rate/qwen_vl_interface': 2.1270851139527986e-06, 'learning_rate/action_model': 2.1270851139527983e-05, 'epoch': 0.27}) +[2026-10-03 19:40:24,728][starVLA.training.train_starvla][INFO] - Step 3670, Loss: {'train/loss': 0.3701668679714203, 'train/grad_norm_pre_clip': 133.63931274414062, 'timing/data': 0.00021150000975467265, 'timing/model': 0.3918899630079977, 'learning_rate/qwen_vl_interface': 2.1247910517863653e-06, 'learning_rate/action_model': 2.124791051786365e-05, 'epoch': 0.27}) +[2026-10-03 19:40:25,424][starVLA.training.train_starvla][INFO] - Step 3671, Loss: {'train/loss': 0.008240234106779099, 'train/grad_norm_pre_clip': 27.554996490478516, 'timing/data': 0.00020142001449130476, 'timing/model': 0.38916906301165, 'learning_rate/qwen_vl_interface': 2.1224982742752814e-06, 'learning_rate/action_model': 2.1224982742752812e-05, 'epoch': 0.27}) +[2026-10-03 19:40:26,124][starVLA.training.train_starvla][INFO] - Step 3672, Loss: {'train/loss': 0.001552610076032579, 'train/grad_norm_pre_clip': 2.453843593597412, 'timing/data': 0.00021214000298641622, 'timing/model': 0.39194473301176913, 'learning_rate/qwen_vl_interface': 2.1202067823620213e-06, 'learning_rate/action_model': 2.1202067823620212e-05, 'epoch': 0.27}) +[2026-10-03 19:40:26,820][starVLA.training.train_starvla][INFO] - Step 3673, Loss: {'train/loss': 0.0005186191410757601, 'train/grad_norm_pre_clip': 0.2798495590686798, 'timing/data': 0.00021157998708076775, 'timing/model': 0.38973022301797755, 'learning_rate/qwen_vl_interface': 2.117916576988531e-06, 'learning_rate/action_model': 2.117916576988531e-05, 'epoch': 0.27}) +[2026-10-03 19:40:27,524][starVLA.training.train_starvla][INFO] - Step 3674, Loss: {'train/loss': 7.471411663573235e-05, 'train/grad_norm_pre_clip': 0.023441409692168236, 'timing/data': 0.00020814998424611986, 'timing/model': 0.3943666330014821, 'learning_rate/qwen_vl_interface': 2.1156276590962257e-06, 'learning_rate/action_model': 2.1156276590962257e-05, 'epoch': 0.27}) +[2026-10-03 19:40:28,433][starVLA.training.train_starvla][INFO] - Step 3675, Loss: {'train/loss': 0.0008335004677064717, 'train/grad_norm_pre_clip': 2.7313079833984375, 'timing/data': 0.00020766002126038074, 'timing/model': 0.39381367299938574, 'learning_rate/qwen_vl_interface': 2.113340029625994e-06, 'learning_rate/action_model': 2.1133400296259938e-05, 'epoch': 0.27}) +[2026-10-03 19:40:29,133][starVLA.training.train_starvla][INFO] - Step 3676, Loss: {'train/loss': 0.3842220902442932, 'train/grad_norm_pre_clip': 64.47821807861328, 'timing/data': 0.00031512000714428723, 'timing/model': 0.390802733018063, 'learning_rate/qwen_vl_interface': 2.111053689518194e-06, 'learning_rate/action_model': 2.111053689518194e-05, 'epoch': 0.27}) +[2026-10-03 19:40:29,838][starVLA.training.train_starvla][INFO] - Step 3677, Loss: {'train/loss': 0.6706598401069641, 'train/grad_norm_pre_clip': 105.58859252929688, 'timing/data': 0.0003324400167912245, 'timing/model': 0.3946319630194921, 'learning_rate/qwen_vl_interface': 2.1087686397126544e-06, 'learning_rate/action_model': 2.1087686397126545e-05, 'epoch': 0.27}) +[2026-10-03 19:40:30,546][starVLA.training.train_starvla][INFO] - Step 3678, Loss: {'train/loss': 0.000742102216463536, 'train/grad_norm_pre_clip': 0.4994260370731354, 'timing/data': 0.0003662100061774254, 'timing/model': 0.39489430299727246, 'learning_rate/qwen_vl_interface': 2.106484881148672e-06, 'learning_rate/action_model': 2.1064848811486718e-05, 'epoch': 0.27}) +[2026-10-03 19:40:31,247][starVLA.training.train_starvla][INFO] - Step 3679, Loss: {'train/loss': 0.020379869267344475, 'train/grad_norm_pre_clip': 31.79457664489746, 'timing/data': 0.0003004400059580803, 'timing/model': 0.39289459300925955, 'learning_rate/qwen_vl_interface': 2.1042024147650137e-06, 'learning_rate/action_model': 2.1042024147650137e-05, 'epoch': 0.27}) +[2026-10-03 19:40:32,142][starVLA.training.train_starvla][INFO] - Step 3680, Loss: {'train/loss': 0.0001959326327778399, 'train/grad_norm_pre_clip': 0.09723588824272156, 'timing/data': 0.00021264998940750957, 'timing/model': 0.5877027839887887, 'learning_rate/qwen_vl_interface': 2.1019212414999163e-06, 'learning_rate/action_model': 2.101921241499916e-05, 'epoch': 0.27}) +[2026-10-03 19:40:32,840][starVLA.training.train_starvla][INFO] - Step 3681, Loss: {'train/loss': 0.47095808386802673, 'train/grad_norm_pre_clip': 43.50251388549805, 'timing/data': 0.00020578998373821378, 'timing/model': 0.3905065430153627, 'learning_rate/qwen_vl_interface': 2.099641362291082e-06, 'learning_rate/action_model': 2.099641362291082e-05, 'epoch': 0.27}) +[2026-10-03 19:40:33,539][starVLA.training.train_starvla][INFO] - Step 3682, Loss: {'train/loss': 0.22078260779380798, 'train/grad_norm_pre_clip': 180.4174041748047, 'timing/data': 0.00033809998421929777, 'timing/model': 0.3900306829891633, 'learning_rate/qwen_vl_interface': 2.097362778075685e-06, 'learning_rate/action_model': 2.097362778075685e-05, 'epoch': 0.27}) +[2026-10-03 19:40:34,238][starVLA.training.train_starvla][INFO] - Step 3683, Loss: {'train/loss': 0.6498112678527832, 'train/grad_norm_pre_clip': 118.81945037841797, 'timing/data': 0.00020186000620014966, 'timing/model': 0.39081892298418097, 'learning_rate/qwen_vl_interface': 2.0950854897903663e-06, 'learning_rate/action_model': 2.095085489790366e-05, 'epoch': 0.27}) +[2026-10-03 19:40:34,935][starVLA.training.train_starvla][INFO] - Step 3684, Loss: {'train/loss': 0.08790159225463867, 'train/grad_norm_pre_clip': 128.0165557861328, 'timing/data': 0.00020573000074364245, 'timing/model': 0.3915245730022434, 'learning_rate/qwen_vl_interface': 2.092809498371229e-06, 'learning_rate/action_model': 2.0928094983712292e-05, 'epoch': 0.27}) +[2026-10-03 19:40:35,837][starVLA.training.train_starvla][INFO] - Step 3685, Loss: {'train/loss': 0.0030171829275786877, 'train/grad_norm_pre_clip': 12.815932273864746, 'timing/data': 0.00019918999169021845, 'timing/model': 0.5947279829997569, 'learning_rate/qwen_vl_interface': 2.0905348047538498e-06, 'learning_rate/action_model': 2.0905348047538497e-05, 'epoch': 0.27}) +[2026-10-03 19:40:36,535][starVLA.training.train_starvla][INFO] - Step 3686, Loss: {'train/loss': 0.0007063081720843911, 'train/grad_norm_pre_clip': 0.4224352240562439, 'timing/data': 0.00033075001556426287, 'timing/model': 0.39085145300487056, 'learning_rate/qwen_vl_interface': 2.0882614098732686e-06, 'learning_rate/action_model': 2.0882614098732685e-05, 'epoch': 0.27}) +[2026-10-03 19:40:37,228][starVLA.training.train_starvla][INFO] - Step 3687, Loss: {'train/loss': 0.05727805569767952, 'train/grad_norm_pre_clip': 162.7876739501953, 'timing/data': 0.00020474998746067286, 'timing/model': 0.3887763530074153, 'learning_rate/qwen_vl_interface': 2.085989314663993e-06, 'learning_rate/action_model': 2.0859893146639927e-05, 'epoch': 0.27}) +[2026-10-03 19:40:37,925][starVLA.training.train_starvla][INFO] - Step 3688, Loss: {'train/loss': 6.801769632147625e-05, 'train/grad_norm_pre_clip': 0.0770854651927948, 'timing/data': 0.0002705899823922664, 'timing/model': 0.3907163729891181, 'learning_rate/qwen_vl_interface': 2.083718520059995e-06, 'learning_rate/action_model': 2.0837185200599946e-05, 'epoch': 0.27}) +[2026-10-03 19:40:38,621][starVLA.training.train_starvla][INFO] - Step 3689, Loss: {'train/loss': 0.00021290895529091358, 'train/grad_norm_pre_clip': 0.5544334053993225, 'timing/data': 0.0002788399870041758, 'timing/model': 0.38939075300004333, 'learning_rate/qwen_vl_interface': 2.0814490269947117e-06, 'learning_rate/action_model': 2.0814490269947118e-05, 'epoch': 0.27}) +[2026-10-03 19:40:39,516][starVLA.training.train_starvla][INFO] - Step 3690, Loss: {'train/loss': 9.398038673680276e-05, 'train/grad_norm_pre_clip': 0.050334688276052475, 'timing/data': 0.0003326699952594936, 'timing/model': 0.5864065150090028, 'learning_rate/qwen_vl_interface': 2.079180836401047e-06, 'learning_rate/action_model': 2.079180836401047e-05, 'epoch': 0.27}) +[2026-10-03 19:40:40,211][starVLA.training.train_starvla][INFO] - Step 3691, Loss: {'train/loss': 7.680034468648955e-05, 'train/grad_norm_pre_clip': 0.021906642243266106, 'timing/data': 0.00031025000498630106, 'timing/model': 0.3883416330209002, 'learning_rate/qwen_vl_interface': 2.0769139492113687e-06, 'learning_rate/action_model': 2.076913949211369e-05, 'epoch': 0.27}) +[2026-10-03 19:40:40,908][starVLA.training.train_starvla][INFO] - Step 3692, Loss: {'train/loss': 0.005953164771199226, 'train/grad_norm_pre_clip': 10.14195442199707, 'timing/data': 0.0003394800005480647, 'timing/model': 0.38997718298924156, 'learning_rate/qwen_vl_interface': 2.0746483663575075e-06, 'learning_rate/action_model': 2.0746483663575076e-05, 'epoch': 0.27}) +[2026-10-03 19:40:41,605][starVLA.training.train_starvla][INFO] - Step 3693, Loss: {'train/loss': 0.001071425387635827, 'train/grad_norm_pre_clip': 2.0136032104492188, 'timing/data': 0.00032892997842282057, 'timing/model': 0.3914104830182623, 'learning_rate/qwen_vl_interface': 2.0723840887707595e-06, 'learning_rate/action_model': 2.0723840887707594e-05, 'epoch': 0.27}) +[2026-10-03 19:40:42,306][starVLA.training.train_starvla][INFO] - Step 3694, Loss: {'train/loss': 0.4569273293018341, 'train/grad_norm_pre_clip': 32.86922836303711, 'timing/data': 0.0002067800087388605, 'timing/model': 0.39169924301677383, 'learning_rate/qwen_vl_interface': 2.0701211173818854e-06, 'learning_rate/action_model': 2.0701211173818853e-05, 'epoch': 0.27}) +[2026-10-03 19:40:43,005][starVLA.training.train_starvla][INFO] - Step 3695, Loss: {'train/loss': 0.0033345166593790054, 'train/grad_norm_pre_clip': 5.423076629638672, 'timing/data': 0.00021235999884083867, 'timing/model': 0.392225472984137, 'learning_rate/qwen_vl_interface': 2.0678594531211036e-06, 'learning_rate/action_model': 2.0678594531211033e-05, 'epoch': 0.27}) +[2026-10-03 19:40:43,915][starVLA.training.train_starvla][INFO] - Step 3696, Loss: {'train/loss': 0.16442139446735382, 'train/grad_norm_pre_clip': 112.55590057373047, 'timing/data': 0.00021659000776708126, 'timing/model': 0.3900122230115812, 'learning_rate/qwen_vl_interface': 2.0655990969181e-06, 'learning_rate/action_model': 2.0655990969181003e-05, 'epoch': 0.27}) +[2026-10-03 19:40:44,614][starVLA.training.train_starvla][INFO] - Step 3697, Loss: {'train/loss': 4.1631486965343356e-05, 'train/grad_norm_pre_clip': 0.010284164920449257, 'timing/data': 0.00021093001123517752, 'timing/model': 0.3907146129931789, 'learning_rate/qwen_vl_interface': 2.0633400497020236e-06, 'learning_rate/action_model': 2.0633400497020232e-05, 'epoch': 0.27}) +[2026-10-03 19:40:45,314][starVLA.training.train_starvla][INFO] - Step 3698, Loss: {'train/loss': 0.00806211493909359, 'train/grad_norm_pre_clip': 38.47687530517578, 'timing/data': 0.00021274000755511224, 'timing/model': 0.3926285430206917, 'learning_rate/qwen_vl_interface': 2.061082312401481e-06, 'learning_rate/action_model': 2.061082312401481e-05, 'epoch': 0.27}) +[2026-10-03 19:40:46,016][starVLA.training.train_starvla][INFO] - Step 3699, Loss: {'train/loss': 0.0002596129197627306, 'train/grad_norm_pre_clip': 0.13723735511302948, 'timing/data': 0.00021269998978823423, 'timing/model': 0.3911921829858329, 'learning_rate/qwen_vl_interface': 2.058825885944544e-06, 'learning_rate/action_model': 2.0588258859445442e-05, 'epoch': 0.27}) +[2026-10-03 19:40:46,720][starVLA.training.train_starvla][INFO] - Step 3700, Loss: {'train/loss': 0.00041598177631385624, 'train/grad_norm_pre_clip': 0.27363666892051697, 'timing/data': 0.0002101600111927837, 'timing/model': 0.3951172019878868, 'learning_rate/qwen_vl_interface': 2.0565707712587447e-06, 'learning_rate/action_model': 2.0565707712587444e-05, 'epoch': 0.27}) +[2026-10-03 19:40:47,628][starVLA.training.train_starvla][INFO] - Step 3701, Loss: {'train/loss': 0.00017625477630645037, 'train/grad_norm_pre_clip': 0.12357019633054733, 'timing/data': 0.00020859998767264187, 'timing/model': 0.5956740330148023, 'learning_rate/qwen_vl_interface': 2.0543169692710747e-06, 'learning_rate/action_model': 2.054316969271075e-05, 'epoch': 0.27}) +[2026-10-03 19:40:48,332][starVLA.training.train_starvla][INFO] - Step 3702, Loss: {'train/loss': 0.0008606552728451788, 'train/grad_norm_pre_clip': 1.0401538610458374, 'timing/data': 0.00020585002494044602, 'timing/model': 0.3937663729884662, 'learning_rate/qwen_vl_interface': 2.0520644809079873e-06, 'learning_rate/action_model': 2.0520644809079873e-05, 'epoch': 0.27}) +[2026-10-03 19:40:49,037][starVLA.training.train_starvla][INFO] - Step 3703, Loss: {'train/loss': 0.633639395236969, 'train/grad_norm_pre_clip': 192.78469848632812, 'timing/data': 0.0002148200001101941, 'timing/model': 0.3930482729920186, 'learning_rate/qwen_vl_interface': 2.049813307095395e-06, 'learning_rate/action_model': 2.0498133070953947e-05, 'epoch': 0.27}) +[2026-10-03 19:40:49,741][starVLA.training.train_starvla][INFO] - Step 3704, Loss: {'train/loss': 0.03718758374452591, 'train/grad_norm_pre_clip': 38.86822509765625, 'timing/data': 0.00036145999911241233, 'timing/model': 0.3931822829763405, 'learning_rate/qwen_vl_interface': 2.047563448758672e-06, 'learning_rate/action_model': 2.0475634487586717e-05, 'epoch': 0.27}) +[2026-10-03 19:40:50,438][starVLA.training.train_starvla][INFO] - Step 3705, Loss: {'train/loss': 0.006158990319818258, 'train/grad_norm_pre_clip': 26.53691864013672, 'timing/data': 0.00020669001969508827, 'timing/model': 0.3899040230026003, 'learning_rate/qwen_vl_interface': 2.045314906822647e-06, 'learning_rate/action_model': 2.0453149068226467e-05, 'epoch': 0.27}) +[2026-10-03 19:40:51,137][starVLA.training.train_starvla][INFO] - Step 3706, Loss: {'train/loss': 0.1642218679189682, 'train/grad_norm_pre_clip': 171.30511474609375, 'timing/data': 0.00023877999046817422, 'timing/model': 0.3905504930007737, 'learning_rate/qwen_vl_interface': 2.0430676822116124e-06, 'learning_rate/action_model': 2.0430676822116123e-05, 'epoch': 0.27}) +[2026-10-03 19:40:52,033][starVLA.training.train_starvla][INFO] - Step 3707, Loss: {'train/loss': 2.071194103336893e-05, 'train/grad_norm_pre_clip': 0.005661497358232737, 'timing/data': 0.0002080300182569772, 'timing/model': 0.3895122829999309, 'learning_rate/qwen_vl_interface': 2.040821775849319e-06, 'learning_rate/action_model': 2.0408217758493188e-05, 'epoch': 0.27}) +[2026-10-03 19:40:52,732][starVLA.training.train_starvla][INFO] - Step 3708, Loss: {'train/loss': 0.04738491028547287, 'train/grad_norm_pre_clip': 115.12071990966797, 'timing/data': 0.0002599500003270805, 'timing/model': 0.39071277299080975, 'learning_rate/qwen_vl_interface': 2.03857718865897e-06, 'learning_rate/action_model': 2.0385771886589702e-05, 'epoch': 0.27}) +[2026-10-03 19:40:53,433][starVLA.training.train_starvla][INFO] - Step 3709, Loss: {'train/loss': 2.2201380488695577e-05, 'train/grad_norm_pre_clip': 0.008120587095618248, 'timing/data': 0.00020286001381464303, 'timing/model': 0.39272707200143486, 'learning_rate/qwen_vl_interface': 2.0363339215632337e-06, 'learning_rate/action_model': 2.0363339215632334e-05, 'epoch': 0.27}) +[2026-10-03 19:40:54,135][starVLA.training.train_starvla][INFO] - Step 3710, Loss: {'train/loss': 0.0027901516295969486, 'train/grad_norm_pre_clip': 3.9776113033294678, 'timing/data': 0.0002723500074353069, 'timing/model': 0.39236692199483514, 'learning_rate/qwen_vl_interface': 2.034091975484231e-06, 'learning_rate/action_model': 2.0340919754842307e-05, 'epoch': 0.27}) +[2026-10-03 19:40:54,838][starVLA.training.train_starvla][INFO] - Step 3711, Loss: {'train/loss': 0.0005346113466657698, 'train/grad_norm_pre_clip': 1.4329506158828735, 'timing/data': 0.00021602000924758613, 'timing/model': 0.39283086301293224, 'learning_rate/qwen_vl_interface': 2.0318513513435415e-06, 'learning_rate/action_model': 2.0318513513435415e-05, 'epoch': 0.27}) +[2026-10-03 19:40:55,741][starVLA.training.train_starvla][INFO] - Step 3712, Loss: {'train/loss': 0.5509366989135742, 'train/grad_norm_pre_clip': 206.0917510986328, 'timing/data': 0.0003822100115939975, 'timing/model': 0.3927022629941348, 'learning_rate/qwen_vl_interface': 2.0296120500622018e-06, 'learning_rate/action_model': 2.0296120500622017e-05, 'epoch': 0.27}) +[2026-10-03 19:40:56,442][starVLA.training.train_starvla][INFO] - Step 3713, Loss: {'train/loss': 1.822342528612353e-05, 'train/grad_norm_pre_clip': 0.005736239720135927, 'timing/data': 0.0002371500013396144, 'timing/model': 0.3914799029880669, 'learning_rate/qwen_vl_interface': 2.0273740725607034e-06, 'learning_rate/action_model': 2.0273740725607032e-05, 'epoch': 0.27}) +[2026-10-03 19:40:57,146][starVLA.training.train_starvla][INFO] - Step 3714, Loss: {'train/loss': 7.897515388322063e-06, 'train/grad_norm_pre_clip': 0.002464084420353174, 'timing/data': 0.0003704100090544671, 'timing/model': 0.3914117130043451, 'learning_rate/qwen_vl_interface': 2.0251374197589942e-06, 'learning_rate/action_model': 2.0251374197589943e-05, 'epoch': 0.27}) +[2026-10-03 19:40:57,843][starVLA.training.train_starvla][INFO] - Step 3715, Loss: {'train/loss': 8.57901104609482e-05, 'train/grad_norm_pre_clip': 0.04788217321038246, 'timing/data': 0.0003431300283409655, 'timing/model': 0.39139188299304806, 'learning_rate/qwen_vl_interface': 2.0229020925764785e-06, 'learning_rate/action_model': 2.0229020925764785e-05, 'epoch': 0.27}) +[2026-10-03 19:40:58,540][starVLA.training.train_starvla][INFO] - Step 3716, Loss: {'train/loss': 0.041687462478876114, 'train/grad_norm_pre_clip': 88.39984130859375, 'timing/data': 0.0003231900045648217, 'timing/model': 0.3895673730003182, 'learning_rate/qwen_vl_interface': 2.0206680919320145e-06, 'learning_rate/action_model': 2.020668091932014e-05, 'epoch': 0.27}) +[2026-10-03 19:40:59,444][starVLA.training.train_starvla][INFO] - Step 3717, Loss: {'train/loss': 1.7433396351407282e-05, 'train/grad_norm_pre_clip': 0.0036310655996203423, 'timing/data': 0.0003708800068125129, 'timing/model': 0.5945818939944729, 'learning_rate/qwen_vl_interface': 2.0184354187439164e-06, 'learning_rate/action_model': 2.0184354187439162e-05, 'epoch': 0.27}) +[2026-10-03 19:41:00,141][starVLA.training.train_starvla][INFO] - Step 3718, Loss: {'train/loss': 0.6482955813407898, 'train/grad_norm_pre_clip': 49.39596176147461, 'timing/data': 0.00036286000977270305, 'timing/model': 0.39033035302418284, 'learning_rate/qwen_vl_interface': 2.016204073929949e-06, 'learning_rate/action_model': 2.016204073929949e-05, 'epoch': 0.27}) +[2026-10-03 19:41:00,841][starVLA.training.train_starvla][INFO] - Step 3719, Loss: {'train/loss': 0.030971817672252655, 'train/grad_norm_pre_clip': 48.06439971923828, 'timing/data': 0.00027903998852707446, 'timing/model': 0.3914303130004555, 'learning_rate/qwen_vl_interface': 2.0139740584073363e-06, 'learning_rate/action_model': 2.0139740584073364e-05, 'epoch': 0.27}) +[2026-10-03 19:41:01,538][starVLA.training.train_starvla][INFO] - Step 3720, Loss: {'train/loss': 0.7481565475463867, 'train/grad_norm_pre_clip': 42.886383056640625, 'timing/data': 0.0003546199877746403, 'timing/model': 0.39046654300182126, 'learning_rate/qwen_vl_interface': 2.011745373092753e-06, 'learning_rate/action_model': 2.011745373092753e-05, 'epoch': 0.27}) +[2026-10-03 19:41:02,234][starVLA.training.train_starvla][INFO] - Step 3721, Loss: {'train/loss': 0.00012118681479478255, 'train/grad_norm_pre_clip': 0.12159765511751175, 'timing/data': 0.0002419099910184741, 'timing/model': 0.38917357299942523, 'learning_rate/qwen_vl_interface': 2.0095180189023267e-06, 'learning_rate/action_model': 2.0095180189023265e-05, 'epoch': 0.27}) +[2026-10-03 19:41:02,930][starVLA.training.train_starvla][INFO] - Step 3722, Loss: {'train/loss': 6.129717075964436e-05, 'train/grad_norm_pre_clip': 0.052153926342725754, 'timing/data': 0.000337860023137182, 'timing/model': 0.3889912629965693, 'learning_rate/qwen_vl_interface': 2.0072919967516393e-06, 'learning_rate/action_model': 2.0072919967516392e-05, 'epoch': 0.27}) +[2026-10-03 19:41:03,825][starVLA.training.train_starvla][INFO] - Step 3723, Loss: {'train/loss': 0.00019387264910619706, 'train/grad_norm_pre_clip': 0.166132852435112, 'timing/data': 0.0002352599985897541, 'timing/model': 0.3901427329983562, 'learning_rate/qwen_vl_interface': 2.0050673075557245e-06, 'learning_rate/action_model': 2.0050673075557243e-05, 'epoch': 0.27}) +[2026-10-03 19:41:04,523][starVLA.training.train_starvla][INFO] - Step 3724, Loss: {'train/loss': 0.12621894478797913, 'train/grad_norm_pre_clip': 114.46973419189453, 'timing/data': 0.00021987000945955515, 'timing/model': 0.39132835299824364, 'learning_rate/qwen_vl_interface': 2.0028439522290677e-06, 'learning_rate/action_model': 2.0028439522290676e-05, 'epoch': 0.27}) +[2026-10-03 19:41:05,220][starVLA.training.train_starvla][INFO] - Step 3725, Loss: {'train/loss': 1.9951416106778197e-05, 'train/grad_norm_pre_clip': 0.017020517960190773, 'timing/data': 0.0003321899857837707, 'timing/model': 0.38982060298440047, 'learning_rate/qwen_vl_interface': 2.000621931685606e-06, 'learning_rate/action_model': 2.000621931685606e-05, 'epoch': 0.27}) +[2026-10-03 19:41:05,917][starVLA.training.train_starvla][INFO] - Step 3726, Loss: {'train/loss': 2.4474427846143954e-05, 'train/grad_norm_pre_clip': 0.009776785038411617, 'timing/data': 0.0002135499962605536, 'timing/model': 0.3908767029934097, 'learning_rate/qwen_vl_interface': 1.9984012468387285e-06, 'learning_rate/action_model': 1.9984012468387285e-05, 'epoch': 0.27}) +[2026-10-03 19:41:06,614][starVLA.training.train_starvla][INFO] - Step 3727, Loss: {'train/loss': 0.3874381184577942, 'train/grad_norm_pre_clip': 452.5688171386719, 'timing/data': 0.00023013001191429794, 'timing/model': 0.3906799830147065, 'learning_rate/qwen_vl_interface': 1.9961818986012767e-06, 'learning_rate/action_model': 1.9961818986012763e-05, 'epoch': 0.27}) +[2026-10-03 19:41:07,513][starVLA.training.train_starvla][INFO] - Step 3728, Loss: {'train/loss': 1.338099900749512e-05, 'train/grad_norm_pre_clip': 0.0032965978607535362, 'timing/data': 0.0003707700234372169, 'timing/model': 0.38812022301135585, 'learning_rate/qwen_vl_interface': 1.9939638878855372e-06, 'learning_rate/action_model': 1.993963887885537e-05, 'epoch': 0.27}) +[2026-10-03 19:41:08,210][starVLA.training.train_starvla][INFO] - Step 3729, Loss: {'train/loss': 0.6545901298522949, 'train/grad_norm_pre_clip': 58.58735656738281, 'timing/data': 0.00040796000394038856, 'timing/model': 0.3897936529829167, 'learning_rate/qwen_vl_interface': 1.991747215603252e-06, 'learning_rate/action_model': 1.991747215603252e-05, 'epoch': 0.27}) +[2026-10-03 19:41:08,908][starVLA.training.train_starvla][INFO] - Step 3730, Loss: {'train/loss': 0.5396513938903809, 'train/grad_norm_pre_clip': 149.82493591308594, 'timing/data': 0.0003308589803054929, 'timing/model': 0.3900689629954286, 'learning_rate/qwen_vl_interface': 1.9895318826656116e-06, 'learning_rate/action_model': 1.9895318826656114e-05, 'epoch': 0.27}) +[2026-10-03 19:41:09,604][starVLA.training.train_starvla][INFO] - Step 3731, Loss: {'train/loss': 0.0029706694185733795, 'train/grad_norm_pre_clip': 3.682637929916382, 'timing/data': 0.0003242699895054102, 'timing/model': 0.38928728297469206, 'learning_rate/qwen_vl_interface': 1.9873178899832553e-06, 'learning_rate/action_model': 1.9873178899832554e-05, 'epoch': 0.27}) +[2026-10-03 19:41:10,301][starVLA.training.train_starvla][INFO] - Step 3732, Loss: {'train/loss': 0.36034566164016724, 'train/grad_norm_pre_clip': 174.68380737304688, 'timing/data': 0.00020668000797741115, 'timing/model': 0.3899977429828141, 'learning_rate/qwen_vl_interface': 1.985105238466272e-06, 'learning_rate/action_model': 1.9851052384662722e-05, 'epoch': 0.27}) +[2026-10-03 19:41:11,211][starVLA.training.train_starvla][INFO] - Step 3733, Loss: {'train/loss': 8.73995159054175e-05, 'train/grad_norm_pre_clip': 0.03866375982761383, 'timing/data': 0.00031539000337943435, 'timing/model': 0.6025508820021059, 'learning_rate/qwen_vl_interface': 1.9828939290242e-06, 'learning_rate/action_model': 1.9828939290241998e-05, 'epoch': 0.27}) +[2026-10-03 19:41:11,907][starVLA.training.train_starvla][INFO] - Step 3734, Loss: {'train/loss': 6.749755993951112e-05, 'train/grad_norm_pre_clip': 0.05098988488316536, 'timing/data': 0.00020589999621734023, 'timing/model': 0.3899922829878051, 'learning_rate/qwen_vl_interface': 1.980683962566024e-06, 'learning_rate/action_model': 1.980683962566024e-05, 'epoch': 0.27}) +[2026-10-03 19:41:12,604][starVLA.training.train_starvla][INFO] - Step 3735, Loss: {'train/loss': 0.13807366788387299, 'train/grad_norm_pre_clip': 106.92357635498047, 'timing/data': 0.00032685999758541584, 'timing/model': 0.3903696830093395, 'learning_rate/qwen_vl_interface': 1.9784753400001784e-06, 'learning_rate/action_model': 1.978475340000178e-05, 'epoch': 0.27}) +[2026-10-03 19:41:13,298][starVLA.training.train_starvla][INFO] - Step 3736, Loss: {'train/loss': 0.00014910816389601678, 'train/grad_norm_pre_clip': 0.1051146611571312, 'timing/data': 0.00023608998162671924, 'timing/model': 0.38832804301637225, 'learning_rate/qwen_vl_interface': 1.9762680622345443e-06, 'learning_rate/action_model': 1.9762680622345442e-05, 'epoch': 0.27}) +[2026-10-03 19:41:13,991][starVLA.training.train_starvla][INFO] - Step 3737, Loss: {'train/loss': 5.156093175173737e-05, 'train/grad_norm_pre_clip': 0.01787356287240982, 'timing/data': 0.00020240998128429055, 'timing/model': 0.38974976199097, 'learning_rate/qwen_vl_interface': 1.9740621301764513e-06, 'learning_rate/action_model': 1.9740621301764513e-05, 'epoch': 0.27}) +[2026-10-03 19:41:14,891][starVLA.training.train_starvla][INFO] - Step 3738, Loss: {'train/loss': 0.00036363821709528565, 'train/grad_norm_pre_clip': 0.2841174900531769, 'timing/data': 0.0003151599958073348, 'timing/model': 0.5935190640157089, 'learning_rate/qwen_vl_interface': 1.971857544732672e-06, 'learning_rate/action_model': 1.971857544732672e-05, 'epoch': 0.27}) +[2026-10-03 19:41:15,587][starVLA.training.train_starvla][INFO] - Step 3739, Loss: {'train/loss': 0.00012873306695837528, 'train/grad_norm_pre_clip': 0.07642548531293869, 'timing/data': 0.00020429998403415084, 'timing/model': 0.39101735199801624, 'learning_rate/qwen_vl_interface': 1.96965430680943e-06, 'learning_rate/action_model': 1.96965430680943e-05, 'epoch': 0.27}) +[2026-10-03 19:41:16,284][starVLA.training.train_starvla][INFO] - Step 3740, Loss: {'train/loss': 0.602689802646637, 'train/grad_norm_pre_clip': 37.71021270751953, 'timing/data': 0.00033114998950622976, 'timing/model': 0.39001716300845146, 'learning_rate/qwen_vl_interface': 1.967452417312393e-06, 'learning_rate/action_model': 1.9674524173123927e-05, 'epoch': 0.27}) +[2026-10-03 19:41:16,980][starVLA.training.train_starvla][INFO] - Step 3741, Loss: {'train/loss': 5.1767761760856956e-05, 'train/grad_norm_pre_clip': 0.023934563621878624, 'timing/data': 0.00020561000565066934, 'timing/model': 0.3894838829874061, 'learning_rate/qwen_vl_interface': 1.9652518771466746e-06, 'learning_rate/action_model': 1.9652518771466747e-05, 'epoch': 0.27}) +[2026-10-03 19:41:17,678][starVLA.training.train_starvla][INFO] - Step 3742, Loss: {'train/loss': 0.23275035619735718, 'train/grad_norm_pre_clip': 94.08073425292969, 'timing/data': 0.0003451300144661218, 'timing/model': 0.3903146619850304, 'learning_rate/qwen_vl_interface': 1.963052687216834e-06, 'learning_rate/action_model': 1.963052687216834e-05, 'epoch': 0.27}) +[2026-10-03 19:41:18,570][starVLA.training.train_starvla][INFO] - Step 3743, Loss: {'train/loss': 0.06257310509681702, 'train/grad_norm_pre_clip': 123.21733856201172, 'timing/data': 0.00034941997728310525, 'timing/model': 0.5856193840154447, 'learning_rate/qwen_vl_interface': 1.960854848426874e-06, 'learning_rate/action_model': 1.960854848426874e-05, 'epoch': 0.27}) +[2026-10-03 19:41:19,265][starVLA.training.train_starvla][INFO] - Step 3744, Loss: {'train/loss': 0.00040100570186041296, 'train/grad_norm_pre_clip': 1.4643226861953735, 'timing/data': 0.00033995998091995716, 'timing/model': 0.3897070430102758, 'learning_rate/qwen_vl_interface': 1.9586583616802446e-06, 'learning_rate/action_model': 1.9586583616802446e-05, 'epoch': 0.27}) +[2026-10-03 19:41:19,961][starVLA.training.train_starvla][INFO] - Step 3745, Loss: {'train/loss': 0.0005439710221253335, 'train/grad_norm_pre_clip': 1.2350775003433228, 'timing/data': 0.00020739997853524983, 'timing/model': 0.38900825299788266, 'learning_rate/qwen_vl_interface': 1.956463227879838e-06, 'learning_rate/action_model': 1.9564632278798377e-05, 'epoch': 0.27}) +[2026-10-03 19:41:20,659][starVLA.training.train_starvla][INFO] - Step 3746, Loss: {'train/loss': 0.0002420425007585436, 'train/grad_norm_pre_clip': 0.47495025396347046, 'timing/data': 0.00034770998172461987, 'timing/model': 0.3905149119964335, 'learning_rate/qwen_vl_interface': 1.95426944792799e-06, 'learning_rate/action_model': 1.95426944792799e-05, 'epoch': 0.27}) +[2026-10-03 19:41:21,353][starVLA.training.train_starvla][INFO] - Step 3747, Loss: {'train/loss': 0.0006032343953847885, 'train/grad_norm_pre_clip': 0.47231778502464294, 'timing/data': 0.00032578001264482737, 'timing/model': 0.38917411302099936, 'learning_rate/qwen_vl_interface': 1.9520770227264817e-06, 'learning_rate/action_model': 1.9520770227264817e-05, 'epoch': 0.27}) +[2026-10-03 19:41:22,050][starVLA.training.train_starvla][INFO] - Step 3748, Loss: {'train/loss': 0.0012668982381001115, 'train/grad_norm_pre_clip': 1.2630447149276733, 'timing/data': 0.0003201299987267703, 'timing/model': 0.3902763530204538, 'learning_rate/qwen_vl_interface': 1.9498859531765373e-06, 'learning_rate/action_model': 1.9498859531765372e-05, 'epoch': 0.27}) +[2026-10-03 19:41:22,950][starVLA.training.train_starvla][INFO] - Step 3749, Loss: {'train/loss': 3.0620532925240695e-05, 'train/grad_norm_pre_clip': 0.005039367824792862, 'timing/data': 0.0003580199845600873, 'timing/model': 0.39010059300926514, 'learning_rate/qwen_vl_interface': 1.9476962401788204e-06, 'learning_rate/action_model': 1.9476962401788204e-05, 'epoch': 0.27}) +[2026-10-03 19:41:23,648][starVLA.training.train_starvla][INFO] - Step 3750, Loss: {'train/loss': 0.00017151073552668095, 'train/grad_norm_pre_clip': 0.1746618151664734, 'timing/data': 0.00020871998276561499, 'timing/model': 0.3913793920073658, 'learning_rate/qwen_vl_interface': 1.9455078846334397e-06, 'learning_rate/action_model': 1.9455078846334395e-05, 'epoch': 0.27}) +[2026-10-03 19:41:24,345][starVLA.training.train_starvla][INFO] - Step 3751, Loss: {'train/loss': 0.003346313489601016, 'train/grad_norm_pre_clip': 4.861609935760498, 'timing/data': 0.0003102100163232535, 'timing/model': 0.38798054298968054, 'learning_rate/qwen_vl_interface': 1.9433208874399465e-06, 'learning_rate/action_model': 1.9433208874399464e-05, 'epoch': 0.27}) +[2026-10-03 19:41:25,038][starVLA.training.train_starvla][INFO] - Step 3752, Loss: {'train/loss': 8.150697976816446e-05, 'train/grad_norm_pre_clip': 0.0684787780046463, 'timing/data': 0.00032637000549584627, 'timing/model': 0.387639783002669, 'learning_rate/qwen_vl_interface': 1.9411352494973314e-06, 'learning_rate/action_model': 1.9411352494973315e-05, 'epoch': 0.27}) +[2026-10-03 19:41:25,735][starVLA.training.train_starvla][INFO] - Step 3753, Loss: {'train/loss': 8.675811841385439e-05, 'train/grad_norm_pre_clip': 0.027277342975139618, 'timing/data': 0.0003519600140862167, 'timing/model': 0.38995331298792735, 'learning_rate/qwen_vl_interface': 1.938950971704031e-06, 'learning_rate/action_model': 1.9389509717040308e-05, 'epoch': 0.27}) +[2026-10-03 19:41:26,637][starVLA.training.train_starvla][INFO] - Step 3754, Loss: {'train/loss': 0.09063995629549026, 'train/grad_norm_pre_clip': 139.52008056640625, 'timing/data': 0.00037183999666012824, 'timing/model': 0.5921380030049477, 'learning_rate/qwen_vl_interface': 1.936768054957915e-06, 'learning_rate/action_model': 1.9367680549579147e-05, 'epoch': 0.27}) +[2026-10-03 19:41:27,342][starVLA.training.train_starvla][INFO] - Step 3755, Loss: {'train/loss': 0.00010990776354447007, 'train/grad_norm_pre_clip': 0.028262551873922348, 'timing/data': 0.00020551998750306666, 'timing/model': 0.3927315920009278, 'learning_rate/qwen_vl_interface': 1.9345865001562994e-06, 'learning_rate/action_model': 1.9345865001562994e-05, 'epoch': 0.27}) +[2026-10-03 19:41:28,045][starVLA.training.train_starvla][INFO] - Step 3756, Loss: {'train/loss': 0.00024807811132632196, 'train/grad_norm_pre_clip': 0.2027129828929901, 'timing/data': 0.0002782900119200349, 'timing/model': 0.3907295919780154, 'learning_rate/qwen_vl_interface': 1.9324063081959394e-06, 'learning_rate/action_model': 1.9324063081959395e-05, 'epoch': 0.27}) +[2026-10-03 19:41:28,743][starVLA.training.train_starvla][INFO] - Step 3757, Loss: {'train/loss': 0.00032928341533988714, 'train/grad_norm_pre_clip': 0.5215176939964294, 'timing/data': 0.0003462199820205569, 'timing/model': 0.3915974929986987, 'learning_rate/qwen_vl_interface': 1.9302274799730304e-06, 'learning_rate/action_model': 1.9302274799730306e-05, 'epoch': 0.27}) +[2026-10-03 19:41:29,446][starVLA.training.train_starvla][INFO] - Step 3758, Loss: {'train/loss': 0.0001757068239385262, 'train/grad_norm_pre_clip': 0.09111446142196655, 'timing/data': 0.00038363999919965863, 'timing/model': 0.393862902012188, 'learning_rate/qwen_vl_interface': 1.9280500163832055e-06, 'learning_rate/action_model': 1.9280500163832055e-05, 'epoch': 0.28}) +[2026-10-03 19:41:30,151][starVLA.training.train_starvla][INFO] - Step 3759, Loss: {'train/loss': 8.486940350849181e-05, 'train/grad_norm_pre_clip': 0.024736931547522545, 'timing/data': 0.0003297099901828915, 'timing/model': 0.3933340719959233, 'learning_rate/qwen_vl_interface': 1.925873918321537e-06, 'learning_rate/action_model': 1.925873918321537e-05, 'epoch': 0.28}) +[2026-10-03 19:41:31,049][starVLA.training.train_starvla][INFO] - Step 3760, Loss: {'train/loss': 0.002262513153254986, 'train/grad_norm_pre_clip': 4.078240871429443, 'timing/data': 0.0002489099861122668, 'timing/model': 0.3914056830108166, 'learning_rate/qwen_vl_interface': 1.92369918668254e-06, 'learning_rate/action_model': 1.92369918668254e-05, 'epoch': 0.28}) +[2026-10-03 19:41:31,752][starVLA.training.train_starvla][INFO] - Step 3761, Loss: {'train/loss': 0.9628828167915344, 'train/grad_norm_pre_clip': 214.51278686523438, 'timing/data': 0.00034602999221533537, 'timing/model': 0.3919055529986508, 'learning_rate/qwen_vl_interface': 1.9215258223601605e-06, 'learning_rate/action_model': 1.9215258223601605e-05, 'epoch': 0.28}) +[2026-10-03 19:41:32,452][starVLA.training.train_starvla][INFO] - Step 3762, Loss: {'train/loss': 7.811532850610092e-05, 'train/grad_norm_pre_clip': 0.031998857855796814, 'timing/data': 0.00020750000840052962, 'timing/model': 0.3913785630138591, 'learning_rate/qwen_vl_interface': 1.9193538262477874e-06, 'learning_rate/action_model': 1.9193538262477874e-05, 'epoch': 0.28}) +[2026-10-03 19:41:33,148][starVLA.training.train_starvla][INFO] - Step 3763, Loss: {'train/loss': 4.6578865294577554e-05, 'train/grad_norm_pre_clip': 0.01039796695113182, 'timing/data': 0.00020762000349350274, 'timing/model': 0.3890931429923512, 'learning_rate/qwen_vl_interface': 1.917183199238247e-06, 'learning_rate/action_model': 1.917183199238247e-05, 'epoch': 0.28}) +[2026-10-03 19:41:33,844][starVLA.training.train_starvla][INFO] - Step 3764, Loss: {'train/loss': 0.720357358455658, 'train/grad_norm_pre_clip': 51.488563537597656, 'timing/data': 0.00020987997413612902, 'timing/model': 0.391887511999812, 'learning_rate/qwen_vl_interface': 1.9150139422238016e-06, 'learning_rate/action_model': 1.9150139422238018e-05, 'epoch': 0.28}) +[2026-10-03 19:41:34,749][starVLA.training.train_starvla][INFO] - Step 3765, Loss: {'train/loss': 0.4262513220310211, 'train/grad_norm_pre_clip': 210.66461181640625, 'timing/data': 0.00021361000835895538, 'timing/model': 0.3922430529783014, 'learning_rate/qwen_vl_interface': 1.9128460560961513e-06, 'learning_rate/action_model': 1.9128460560961514e-05, 'epoch': 0.28}) +[2026-10-03 19:41:35,447][starVLA.training.train_starvla][INFO] - Step 3766, Loss: {'train/loss': 0.00016812951071187854, 'train/grad_norm_pre_clip': 0.08422789722681046, 'timing/data': 0.00022000001627020538, 'timing/model': 0.3924657519964967, 'learning_rate/qwen_vl_interface': 1.910679541746432e-06, 'learning_rate/action_model': 1.9106795417464323e-05, 'epoch': 0.28}) +[2026-10-03 19:41:36,145][starVLA.training.train_starvla][INFO] - Step 3767, Loss: {'train/loss': 0.07574121654033661, 'train/grad_norm_pre_clip': 48.545589447021484, 'timing/data': 0.0003196100005879998, 'timing/model': 0.3895756129932124, 'learning_rate/qwen_vl_interface': 1.9085144000652166e-06, 'learning_rate/action_model': 1.9085144000652166e-05, 'epoch': 0.28}) +[2026-10-03 19:41:36,842][starVLA.training.train_starvla][INFO] - Step 3768, Loss: {'train/loss': 0.0003122433554381132, 'train/grad_norm_pre_clip': 0.3788367807865143, 'timing/data': 0.0003031599917449057, 'timing/model': 0.3910609220038168, 'learning_rate/qwen_vl_interface': 1.9063506319425116e-06, 'learning_rate/action_model': 1.9063506319425114e-05, 'epoch': 0.28}) +[2026-10-03 19:41:37,539][starVLA.training.train_starvla][INFO] - Step 3769, Loss: {'train/loss': 0.0001609167957212776, 'train/grad_norm_pre_clip': 0.1144789457321167, 'timing/data': 0.00022248999448493123, 'timing/model': 0.3914425020047929, 'learning_rate/qwen_vl_interface': 1.9041882382677613e-06, 'learning_rate/action_model': 1.9041882382677612e-05, 'epoch': 0.28}) +[2026-10-03 19:41:38,439][starVLA.training.train_starvla][INFO] - Step 3770, Loss: {'train/loss': 0.0008848691359162331, 'train/grad_norm_pre_clip': 1.1337052583694458, 'timing/data': 0.0003287389990873635, 'timing/model': 0.5914430039993022, 'learning_rate/qwen_vl_interface': 1.9020272199298452e-06, 'learning_rate/action_model': 1.902027219929845e-05, 'epoch': 0.28}) +[2026-10-03 19:41:39,135][starVLA.training.train_starvla][INFO] - Step 3771, Loss: {'train/loss': 0.1203937903046608, 'train/grad_norm_pre_clip': 104.192138671875, 'timing/data': 0.0002017900114879012, 'timing/model': 0.3905783530208282, 'learning_rate/qwen_vl_interface': 1.8998675778170736e-06, 'learning_rate/action_model': 1.8998675778170734e-05, 'epoch': 0.28}) +[2026-10-03 19:41:39,833][starVLA.training.train_starvla][INFO] - Step 3772, Loss: {'train/loss': 0.5243703722953796, 'train/grad_norm_pre_clip': 190.24205017089844, 'timing/data': 0.00031432000105269253, 'timing/model': 0.3907819830055814, 'learning_rate/qwen_vl_interface': 1.8977093128171955e-06, 'learning_rate/action_model': 1.8977093128171954e-05, 'epoch': 0.28}) +[2026-10-03 19:41:40,531][starVLA.training.train_starvla][INFO] - Step 3773, Loss: {'train/loss': 0.39022862911224365, 'train/grad_norm_pre_clip': 52.981201171875, 'timing/data': 0.00034471999970264733, 'timing/model': 0.3899954220105428, 'learning_rate/qwen_vl_interface': 1.8955524258173919e-06, 'learning_rate/action_model': 1.895552425817392e-05, 'epoch': 0.28}) +[2026-10-03 19:41:41,226][starVLA.training.train_starvla][INFO] - Step 3774, Loss: {'train/loss': 0.0676024854183197, 'train/grad_norm_pre_clip': 276.94891357421875, 'timing/data': 0.0003198599733877927, 'timing/model': 0.38900587300304323, 'learning_rate/qwen_vl_interface': 1.8933969177042769e-06, 'learning_rate/action_model': 1.8933969177042768e-05, 'epoch': 0.28}) +[2026-10-03 19:41:41,921][starVLA.training.train_starvla][INFO] - Step 3775, Loss: {'train/loss': 6.749356543878093e-05, 'train/grad_norm_pre_clip': 0.025277262553572655, 'timing/data': 0.00031909000244922936, 'timing/model': 0.38989727300941013, 'learning_rate/qwen_vl_interface': 1.8912427893638996e-06, 'learning_rate/action_model': 1.8912427893638996e-05, 'epoch': 0.28}) +[2026-10-03 19:41:42,813][starVLA.training.train_starvla][INFO] - Step 3776, Loss: {'train/loss': 0.06876353174448013, 'train/grad_norm_pre_clip': 96.80644989013672, 'timing/data': 0.0002101199934259057, 'timing/model': 0.38925221300451085, 'learning_rate/qwen_vl_interface': 1.8890900416817398e-06, 'learning_rate/action_model': 1.8890900416817397e-05, 'epoch': 0.28}) +[2026-10-03 19:41:43,509][starVLA.training.train_starvla][INFO] - Step 3777, Loss: {'train/loss': 0.00029146962333470583, 'train/grad_norm_pre_clip': 0.09041871130466461, 'timing/data': 0.00032330999965779483, 'timing/model': 0.38933381298556924, 'learning_rate/qwen_vl_interface': 1.8869386755427117e-06, 'learning_rate/action_model': 1.8869386755427114e-05, 'epoch': 0.28}) +[2026-10-03 19:41:44,209][starVLA.training.train_starvla][INFO] - Step 3778, Loss: {'train/loss': 6.94215705152601e-05, 'train/grad_norm_pre_clip': 0.03595719486474991, 'timing/data': 0.00021196997840888798, 'timing/model': 0.39160464200540446, 'learning_rate/qwen_vl_interface': 1.8847886918311597e-06, 'learning_rate/action_model': 1.8847886918311598e-05, 'epoch': 0.28}) +[2026-10-03 19:41:44,906][starVLA.training.train_starvla][INFO] - Step 3779, Loss: {'train/loss': 0.06281868368387222, 'train/grad_norm_pre_clip': 138.9867706298828, 'timing/data': 0.0003190700081177056, 'timing/model': 0.38943499300512485, 'learning_rate/qwen_vl_interface': 1.8826400914308616e-06, 'learning_rate/action_model': 1.8826400914308613e-05, 'epoch': 0.28}) +[2026-10-03 19:41:45,606][starVLA.training.train_starvla][INFO] - Step 3780, Loss: {'train/loss': 0.45206505060195923, 'train/grad_norm_pre_clip': 35.750213623046875, 'timing/data': 0.00033991000964306295, 'timing/model': 0.3924151819956023, 'learning_rate/qwen_vl_interface': 1.880492875225025e-06, 'learning_rate/action_model': 1.8804928752250252e-05, 'epoch': 0.28}) +[2026-10-03 19:41:46,508][starVLA.training.train_starvla][INFO] - Step 3781, Loss: {'train/loss': 0.04557153582572937, 'train/grad_norm_pre_clip': 153.04774475097656, 'timing/data': 0.0002412100147921592, 'timing/model': 0.3931708419986535, 'learning_rate/qwen_vl_interface': 1.8783470440962922e-06, 'learning_rate/action_model': 1.878347044096292e-05, 'epoch': 0.28}) +[2026-10-03 19:41:47,212][starVLA.training.train_starvla][INFO] - Step 3782, Loss: {'train/loss': 0.0031674120109528303, 'train/grad_norm_pre_clip': 1.8782811164855957, 'timing/data': 0.000261779990978539, 'timing/model': 0.39298429200425744, 'learning_rate/qwen_vl_interface': 1.8762025989267293e-06, 'learning_rate/action_model': 1.876202598926729e-05, 'epoch': 0.28}) +[2026-10-03 19:41:47,910][starVLA.training.train_starvla][INFO] - Step 3783, Loss: {'train/loss': 7.136362546589226e-05, 'train/grad_norm_pre_clip': 0.01307766791433096, 'timing/data': 0.0003333200002089143, 'timing/model': 0.3914697629807051, 'learning_rate/qwen_vl_interface': 1.874059540597839e-06, 'learning_rate/action_model': 1.874059540597839e-05, 'epoch': 0.28}) +[2026-10-03 19:41:48,614][starVLA.training.train_starvla][INFO] - Step 3784, Loss: {'train/loss': 0.003068850841373205, 'train/grad_norm_pre_clip': 2.7506628036499023, 'timing/data': 0.00029014001484028995, 'timing/model': 0.39346186199691147, 'learning_rate/qwen_vl_interface': 1.8719178699905508e-06, 'learning_rate/action_model': 1.8719178699905508e-05, 'epoch': 0.28}) +[2026-10-03 19:41:49,315][starVLA.training.train_starvla][INFO] - Step 3785, Loss: {'train/loss': 0.0003515988646540791, 'train/grad_norm_pre_clip': 0.14698217809200287, 'timing/data': 0.0004012499994132668, 'timing/model': 0.39178627301589586, 'learning_rate/qwen_vl_interface': 1.8697775879852242e-06, 'learning_rate/action_model': 1.8697775879852242e-05, 'epoch': 0.28}) +[2026-10-03 19:41:50,213][starVLA.training.train_starvla][INFO] - Step 3786, Loss: {'train/loss': 0.0013313287636265159, 'train/grad_norm_pre_clip': 2.300319194793701, 'timing/data': 0.0003606200043577701, 'timing/model': 0.5886041940248106, 'learning_rate/qwen_vl_interface': 1.86763869546165e-06, 'learning_rate/action_model': 1.86763869546165e-05, 'epoch': 0.28}) +[2026-10-03 19:41:50,909][starVLA.training.train_starvla][INFO] - Step 3787, Loss: {'train/loss': 0.46470969915390015, 'train/grad_norm_pre_clip': 44.38088607788086, 'timing/data': 0.00019724000594578683, 'timing/model': 0.39135699201142415, 'learning_rate/qwen_vl_interface': 1.865501193299045e-06, 'learning_rate/action_model': 1.8655011932990447e-05, 'epoch': 0.28}) +[2026-10-03 19:41:51,610][starVLA.training.train_starvla][INFO] - Step 3788, Loss: {'train/loss': 0.006267928518354893, 'train/grad_norm_pre_clip': 36.75703430175781, 'timing/data': 0.00021791900508105755, 'timing/model': 0.3913166629790794, 'learning_rate/qwen_vl_interface': 1.8633650823760555e-06, 'learning_rate/action_model': 1.8633650823760555e-05, 'epoch': 0.28}) +[2026-10-03 19:41:52,306][starVLA.training.train_starvla][INFO] - Step 3789, Loss: {'train/loss': 0.14528799057006836, 'train/grad_norm_pre_clip': 118.41020202636719, 'timing/data': 0.0002220490132458508, 'timing/model': 0.3908246130158659, 'learning_rate/qwen_vl_interface': 1.8612303635707562e-06, 'learning_rate/action_model': 1.861230363570756e-05, 'epoch': 0.28}) +[2026-10-03 19:41:53,007][starVLA.training.train_starvla][INFO] - Step 3790, Loss: {'train/loss': 7.179325621109456e-05, 'train/grad_norm_pre_clip': 0.01843436062335968, 'timing/data': 0.0003084699856117368, 'timing/model': 0.39110900298692286, 'learning_rate/qwen_vl_interface': 1.8590970377606494e-06, 'learning_rate/action_model': 1.859097037760649e-05, 'epoch': 0.28}) +[2026-10-03 19:41:53,892][starVLA.training.train_starvla][INFO] - Step 3791, Loss: {'train/loss': 0.06759054958820343, 'train/grad_norm_pre_clip': 127.4535140991211, 'timing/data': 0.00025187002029269934, 'timing/model': 0.5793864339939319, 'learning_rate/qwen_vl_interface': 1.8569651058226667e-06, 'learning_rate/action_model': 1.8569651058226666e-05, 'epoch': 0.28}) +[2026-10-03 19:41:54,588][starVLA.training.train_starvla][INFO] - Step 3792, Loss: {'train/loss': 0.0003033060929737985, 'train/grad_norm_pre_clip': 0.11557856947183609, 'timing/data': 0.00019932998111471534, 'timing/model': 0.3892156029760372, 'learning_rate/qwen_vl_interface': 1.854834568633161e-06, 'learning_rate/action_model': 1.854834568633161e-05, 'epoch': 0.28}) +[2026-10-03 19:41:55,286][starVLA.training.train_starvla][INFO] - Step 3793, Loss: {'train/loss': 0.002693269867449999, 'train/grad_norm_pre_clip': 7.42569637298584, 'timing/data': 0.0003699100052472204, 'timing/model': 0.3899105930177029, 'learning_rate/qwen_vl_interface': 1.8527054270679174e-06, 'learning_rate/action_model': 1.8527054270679172e-05, 'epoch': 0.28}) +[2026-10-03 19:41:55,984][starVLA.training.train_starvla][INFO] - Step 3794, Loss: {'train/loss': 0.00011138278205180541, 'train/grad_norm_pre_clip': 0.17786213755607605, 'timing/data': 0.0003510999958962202, 'timing/model': 0.3885330030170735, 'learning_rate/qwen_vl_interface': 1.850577682002146e-06, 'learning_rate/action_model': 1.850577682002146e-05, 'epoch': 0.28}) +[2026-10-03 19:41:56,680][starVLA.training.train_starvla][INFO] - Step 3795, Loss: {'train/loss': 0.0006542488699778914, 'train/grad_norm_pre_clip': 0.6518750190734863, 'timing/data': 0.0001990899909287691, 'timing/model': 0.38911751299747266, 'learning_rate/qwen_vl_interface': 1.848451334310482e-06, 'learning_rate/action_model': 1.848451334310482e-05, 'epoch': 0.28}) +[2026-10-03 19:41:57,378][starVLA.training.train_starvla][INFO] - Step 3796, Loss: {'train/loss': 0.7225598692893982, 'train/grad_norm_pre_clip': 139.9322052001953, 'timing/data': 0.0002969499910250306, 'timing/model': 0.3906788819876965, 'learning_rate/qwen_vl_interface': 1.8463263848669865e-06, 'learning_rate/action_model': 1.8463263848669864e-05, 'epoch': 0.28}) +[2026-10-03 19:41:58,277][starVLA.training.train_starvla][INFO] - Step 3797, Loss: {'train/loss': 0.028606507927179337, 'train/grad_norm_pre_clip': 32.87984085083008, 'timing/data': 0.00020281999604776502, 'timing/model': 0.39050515298731625, 'learning_rate/qwen_vl_interface': 1.8442028345451468e-06, 'learning_rate/action_model': 1.8442028345451468e-05, 'epoch': 0.28}) +[2026-10-03 19:41:58,973][starVLA.training.train_starvla][INFO] - Step 3798, Loss: {'train/loss': 0.0001622233394300565, 'train/grad_norm_pre_clip': 0.07973890751600266, 'timing/data': 0.00020842000958509743, 'timing/model': 0.38932771200779825, 'learning_rate/qwen_vl_interface': 1.8420806842178746e-06, 'learning_rate/action_model': 1.8420806842178745e-05, 'epoch': 0.28}) +[2026-10-03 19:41:59,669][starVLA.training.train_starvla][INFO] - Step 3799, Loss: {'train/loss': 0.0004035085439682007, 'train/grad_norm_pre_clip': 0.7912034392356873, 'timing/data': 0.00030968000646680593, 'timing/model': 0.3895131930185016, 'learning_rate/qwen_vl_interface': 1.8399599347575034e-06, 'learning_rate/action_model': 1.8399599347575033e-05, 'epoch': 0.28}) +[2026-10-03 19:42:00,363][starVLA.training.train_starvla][INFO] - Step 3800, Loss: {'train/loss': 0.0011665893252938986, 'train/grad_norm_pre_clip': 0.6096089482307434, 'timing/data': 0.0002035000070463866, 'timing/model': 0.38926900300430134, 'learning_rate/qwen_vl_interface': 1.8378405870357943e-06, 'learning_rate/action_model': 1.837840587035794e-05, 'epoch': 0.28}) +[2026-10-03 19:42:01,059][starVLA.training.train_starvla][INFO] - Step 3801, Loss: {'train/loss': 0.0006773728528060019, 'train/grad_norm_pre_clip': 0.44019240140914917, 'timing/data': 0.00020289001986384392, 'timing/model': 0.38857547301449813, 'learning_rate/qwen_vl_interface': 1.8357226419239316e-06, 'learning_rate/action_model': 1.8357226419239315e-05, 'epoch': 0.28}) +[2026-10-03 19:42:01,949][starVLA.training.train_starvla][INFO] - Step 3802, Loss: {'train/loss': 0.3700149357318878, 'train/grad_norm_pre_clip': 51.52214431762695, 'timing/data': 0.00020762000349350274, 'timing/model': 0.3907649620086886, 'learning_rate/qwen_vl_interface': 1.8336061002925224e-06, 'learning_rate/action_model': 1.8336061002925224e-05, 'epoch': 0.28}) +[2026-10-03 19:42:02,648][starVLA.training.train_starvla][INFO] - Step 3803, Loss: {'train/loss': 0.00011361524229869246, 'train/grad_norm_pre_clip': 0.04512134566903114, 'timing/data': 0.0002403699909336865, 'timing/model': 0.3907468320103362, 'learning_rate/qwen_vl_interface': 1.8314909630115985e-06, 'learning_rate/action_model': 1.8314909630115983e-05, 'epoch': 0.28}) +[2026-10-03 19:42:03,344][starVLA.training.train_starvla][INFO] - Step 3804, Loss: {'train/loss': 0.00031070297700352967, 'train/grad_norm_pre_clip': 0.24985598027706146, 'timing/data': 0.00037120000342838466, 'timing/model': 0.3889850229898002, 'learning_rate/qwen_vl_interface': 1.8293772309506092e-06, 'learning_rate/action_model': 1.829377230950609e-05, 'epoch': 0.28}) +[2026-10-03 19:42:04,014][starVLA.training.train_starvla][INFO] - Step 3805, Loss: {'train/loss': 0.5294880270957947, 'train/grad_norm_pre_clip': 73.5389633178711, 'timing/data': 0.0002043799904640764, 'timing/model': 0.36487665501772426, 'learning_rate/qwen_vl_interface': 1.8272649049784326e-06, 'learning_rate/action_model': 1.8272649049784323e-05, 'epoch': 0.28}) +[2026-10-03 19:42:04,714][starVLA.training.train_starvla][INFO] - Step 3806, Loss: {'train/loss': 0.00012911944941151887, 'train/grad_norm_pre_clip': 0.07942458987236023, 'timing/data': 0.00021147998631931841, 'timing/model': 0.39151767297880724, 'learning_rate/qwen_vl_interface': 1.8251539859633653e-06, 'learning_rate/action_model': 1.825153985963365e-05, 'epoch': 0.28}) +[2026-10-03 19:42:05,605][starVLA.training.train_starvla][INFO] - Step 3807, Loss: {'train/loss': 0.006335081532597542, 'train/grad_norm_pre_clip': 11.079083442687988, 'timing/data': 0.00020228998619131744, 'timing/model': 0.5841420039942022, 'learning_rate/qwen_vl_interface': 1.8230444747731265e-06, 'learning_rate/action_model': 1.8230444747731264e-05, 'epoch': 0.28}) +[2026-10-03 19:42:06,298][starVLA.training.train_starvla][INFO] - Step 3808, Loss: {'train/loss': 0.000849252101033926, 'train/grad_norm_pre_clip': 0.34664222598075867, 'timing/data': 0.0003002499870490283, 'timing/model': 0.38877594299265184, 'learning_rate/qwen_vl_interface': 1.820936372274857e-06, 'learning_rate/action_model': 1.820936372274857e-05, 'epoch': 0.28}) +[2026-10-03 19:42:06,996][starVLA.training.train_starvla][INFO] - Step 3809, Loss: {'train/loss': 0.3202149569988251, 'train/grad_norm_pre_clip': 46.08831787109375, 'timing/data': 0.00022759000421501696, 'timing/model': 0.3910599620139692, 'learning_rate/qwen_vl_interface': 1.8188296793351178e-06, 'learning_rate/action_model': 1.818829679335118e-05, 'epoch': 0.28}) +[2026-10-03 19:42:07,696][starVLA.training.train_starvla][INFO] - Step 3810, Loss: {'train/loss': 8.086169691523537e-05, 'train/grad_norm_pre_clip': 0.018486758694052696, 'timing/data': 0.00031839899020269513, 'timing/model': 0.390575422992697, 'learning_rate/qwen_vl_interface': 1.8167243968198905e-06, 'learning_rate/action_model': 1.8167243968198905e-05, 'epoch': 0.28}) +[2026-10-03 19:42:08,393][starVLA.training.train_starvla][INFO] - Step 3811, Loss: {'train/loss': 0.0002158564457204193, 'train/grad_norm_pre_clip': 0.09094277024269104, 'timing/data': 0.00019571001757867634, 'timing/model': 0.3902938030078076, 'learning_rate/qwen_vl_interface': 1.8146205255945775e-06, 'learning_rate/action_model': 1.8146205255945777e-05, 'epoch': 0.28}) +[2026-10-03 19:42:09,279][starVLA.training.train_starvla][INFO] - Step 3812, Loss: {'train/loss': 0.32282790541648865, 'train/grad_norm_pre_clip': 189.2477569580078, 'timing/data': 0.0002746199897956103, 'timing/model': 0.5779727449989878, 'learning_rate/qwen_vl_interface': 1.8125180665240007e-06, 'learning_rate/action_model': 1.8125180665240008e-05, 'epoch': 0.28}) +[2026-10-03 19:42:09,979][starVLA.training.train_starvla][INFO] - Step 3813, Loss: {'train/loss': 0.0305250883102417, 'train/grad_norm_pre_clip': 70.49352264404297, 'timing/data': 0.00020730000687763095, 'timing/model': 0.3909818519896362, 'learning_rate/qwen_vl_interface': 1.8104170204724025e-06, 'learning_rate/action_model': 1.8104170204724024e-05, 'epoch': 0.28}) +[2026-10-03 19:42:10,679][starVLA.training.train_starvla][INFO] - Step 3814, Loss: {'train/loss': 0.00025949825067073107, 'train/grad_norm_pre_clip': 0.1453293412923813, 'timing/data': 0.0003366399905644357, 'timing/model': 0.39173567201942205, 'learning_rate/qwen_vl_interface': 1.8083173883034435e-06, 'learning_rate/action_model': 1.8083173883034435e-05, 'epoch': 0.28}) +[2026-10-03 19:42:11,375][starVLA.training.train_starvla][INFO] - Step 3815, Loss: {'train/loss': 0.04496314376592636, 'train/grad_norm_pre_clip': 96.698974609375, 'timing/data': 0.00028236000798642635, 'timing/model': 0.3904688929906115, 'learning_rate/qwen_vl_interface': 1.8062191708802013e-06, 'learning_rate/action_model': 1.8062191708802012e-05, 'epoch': 0.28}) +[2026-10-03 19:42:12,073][starVLA.training.train_starvla][INFO] - Step 3816, Loss: {'train/loss': 0.001072149956598878, 'train/grad_norm_pre_clip': 0.758346438407898, 'timing/data': 0.00031319999834522605, 'timing/model': 0.38959105298272334, 'learning_rate/qwen_vl_interface': 1.804122369065175e-06, 'learning_rate/action_model': 1.8041223690651748e-05, 'epoch': 0.28}) +[2026-10-03 19:42:12,770][starVLA.training.train_starvla][INFO] - Step 3817, Loss: {'train/loss': 7.324274338316172e-05, 'train/grad_norm_pre_clip': 0.02082665078341961, 'timing/data': 0.00020346001838333905, 'timing/model': 0.39121431298553944, 'learning_rate/qwen_vl_interface': 1.8020269837202806e-06, 'learning_rate/action_model': 1.8020269837202804e-05, 'epoch': 0.28}) +[2026-10-03 19:42:13,664][starVLA.training.train_starvla][INFO] - Step 3818, Loss: {'train/loss': 0.0007694673840887845, 'train/grad_norm_pre_clip': 1.1966438293457031, 'timing/data': 0.00020352998399175704, 'timing/model': 0.39126384299015626, 'learning_rate/qwen_vl_interface': 1.7999330157068517e-06, 'learning_rate/action_model': 1.7999330157068515e-05, 'epoch': 0.28}) +[2026-10-03 19:42:14,361][starVLA.training.train_starvla][INFO] - Step 3819, Loss: {'train/loss': 0.0002648755325935781, 'train/grad_norm_pre_clip': 0.12119023501873016, 'timing/data': 0.0002857399813365191, 'timing/model': 0.38966111300396733, 'learning_rate/qwen_vl_interface': 1.7978404658856393e-06, 'learning_rate/action_model': 1.7978404658856394e-05, 'epoch': 0.28}) +[2026-10-03 19:42:15,062][starVLA.training.train_starvla][INFO] - Step 3820, Loss: {'train/loss': 0.2763669490814209, 'train/grad_norm_pre_clip': 81.09007263183594, 'timing/data': 0.0002086999884340912, 'timing/model': 0.39440001198090613, 'learning_rate/qwen_vl_interface': 1.7957493351168126e-06, 'learning_rate/action_model': 1.7957493351168126e-05, 'epoch': 0.28}) +[2026-10-03 19:42:15,764][starVLA.training.train_starvla][INFO] - Step 3821, Loss: {'train/loss': 0.539489209651947, 'train/grad_norm_pre_clip': 36.92119598388672, 'timing/data': 0.00036348000867292285, 'timing/model': 0.39308367201010697, 'learning_rate/qwen_vl_interface': 1.7936596242599552e-06, 'learning_rate/action_model': 1.7936596242599553e-05, 'epoch': 0.28}) +[2026-10-03 19:42:16,465][starVLA.training.train_starvla][INFO] - Step 3822, Loss: {'train/loss': 0.019029082730412483, 'train/grad_norm_pre_clip': 28.478782653808594, 'timing/data': 0.00020480999955907464, 'timing/model': 0.3921793820045423, 'learning_rate/qwen_vl_interface': 1.7915713341740692e-06, 'learning_rate/action_model': 1.791571334174069e-05, 'epoch': 0.28}) +[2026-10-03 19:42:17,361][starVLA.training.train_starvla][INFO] - Step 3823, Loss: {'train/loss': 0.00880814902484417, 'train/grad_norm_pre_clip': 14.515515327453613, 'timing/data': 0.00021202999050728977, 'timing/model': 0.39196258201263845, 'learning_rate/qwen_vl_interface': 1.7894844657175716e-06, 'learning_rate/action_model': 1.7894844657175717e-05, 'epoch': 0.28}) +[2026-10-03 19:42:18,061][starVLA.training.train_starvla][INFO] - Step 3824, Loss: {'train/loss': 4.8528003389947116e-05, 'train/grad_norm_pre_clip': 0.015562309883534908, 'timing/data': 0.0002079599944408983, 'timing/model': 0.3936593219987117, 'learning_rate/qwen_vl_interface': 1.7873990197482962e-06, 'learning_rate/action_model': 1.787399019748296e-05, 'epoch': 0.28}) +[2026-10-03 19:42:18,762][starVLA.training.train_starvla][INFO] - Step 3825, Loss: {'train/loss': 0.0001335306151304394, 'train/grad_norm_pre_clip': 0.040543705224990845, 'timing/data': 0.00030417999369092286, 'timing/model': 0.3940626620023977, 'learning_rate/qwen_vl_interface': 1.7853149971234883e-06, 'learning_rate/action_model': 1.785314997123488e-05, 'epoch': 0.28}) +[2026-10-03 19:42:19,471][starVLA.training.train_starvla][INFO] - Step 3826, Loss: {'train/loss': 0.0002107897453242913, 'train/grad_norm_pre_clip': 0.11523647606372833, 'timing/data': 0.0003470499941613525, 'timing/model': 0.3895494230091572, 'learning_rate/qwen_vl_interface': 1.7832323986998128e-06, 'learning_rate/action_model': 1.783232398699813e-05, 'epoch': 0.28}) +[2026-10-03 19:42:20,171][starVLA.training.train_starvla][INFO] - Step 3827, Loss: {'train/loss': 0.0051473514176905155, 'train/grad_norm_pre_clip': 24.548511505126953, 'timing/data': 0.0002027000009547919, 'timing/model': 0.39120131300296634, 'learning_rate/qwen_vl_interface': 1.7811512253333468e-06, 'learning_rate/action_model': 1.7811512253333467e-05, 'epoch': 0.28}) +[2026-10-03 19:42:21,068][starVLA.training.train_starvla][INFO] - Step 3828, Loss: {'train/loss': 0.030230747535824776, 'train/grad_norm_pre_clip': 63.860069274902344, 'timing/data': 0.0003268799919169396, 'timing/model': 0.5890158739930484, 'learning_rate/qwen_vl_interface': 1.7790714778795823e-06, 'learning_rate/action_model': 1.779071477879582e-05, 'epoch': 0.28}) +[2026-10-03 19:42:21,768][starVLA.training.train_starvla][INFO] - Step 3829, Loss: {'train/loss': 0.021667027845978737, 'train/grad_norm_pre_clip': 53.59701919555664, 'timing/data': 0.00024393000057898462, 'timing/model': 0.3905427330173552, 'learning_rate/qwen_vl_interface': 1.7769931571934246e-06, 'learning_rate/action_model': 1.7769931571934246e-05, 'epoch': 0.28}) +[2026-10-03 19:42:22,466][starVLA.training.train_starvla][INFO] - Step 3830, Loss: {'train/loss': 0.328014075756073, 'train/grad_norm_pre_clip': 45.50991439819336, 'timing/data': 0.0003775500226765871, 'timing/model': 0.3908369619748555, 'learning_rate/qwen_vl_interface': 1.7749162641291933e-06, 'learning_rate/action_model': 1.7749162641291933e-05, 'epoch': 0.28}) +[2026-10-03 19:42:23,164][starVLA.training.train_starvla][INFO] - Step 3831, Loss: {'train/loss': 0.5586900115013123, 'train/grad_norm_pre_clip': 168.1418914794922, 'timing/data': 0.00028944999212399125, 'timing/model': 0.3912917319976259, 'learning_rate/qwen_vl_interface': 1.7728407995406206e-06, 'learning_rate/action_model': 1.7728407995406205e-05, 'epoch': 0.28}) +[2026-10-03 19:42:23,861][starVLA.training.train_starvla][INFO] - Step 3832, Loss: {'train/loss': 0.3652530014514923, 'train/grad_norm_pre_clip': 145.1513214111328, 'timing/data': 0.00021038000704720616, 'timing/model': 0.39107179298298433, 'learning_rate/qwen_vl_interface': 1.7707667642808513e-06, 'learning_rate/action_model': 1.7707667642808512e-05, 'epoch': 0.28}) +[2026-10-03 19:42:24,755][starVLA.training.train_starvla][INFO] - Step 3833, Loss: {'train/loss': 0.0014274836285039783, 'train/grad_norm_pre_clip': 1.4797470569610596, 'timing/data': 0.000202669994905591, 'timing/model': 0.586937563988613, 'learning_rate/qwen_vl_interface': 1.7686941592024423e-06, 'learning_rate/action_model': 1.7686941592024423e-05, 'epoch': 0.28}) +[2026-10-03 19:42:25,452][starVLA.training.train_starvla][INFO] - Step 3834, Loss: {'train/loss': 0.001066792057827115, 'train/grad_norm_pre_clip': 1.3461006879806519, 'timing/data': 0.00020770999253727496, 'timing/model': 0.3895447730028536, 'learning_rate/qwen_vl_interface': 1.7666229851573662e-06, 'learning_rate/action_model': 1.7666229851573664e-05, 'epoch': 0.28}) +[2026-10-03 19:42:26,150][starVLA.training.train_starvla][INFO] - Step 3835, Loss: {'train/loss': 0.000365953310392797, 'train/grad_norm_pre_clip': 0.3422390818595886, 'timing/data': 0.0002041200059466064, 'timing/model': 0.3902417229837738, 'learning_rate/qwen_vl_interface': 1.7645532429970006e-06, 'learning_rate/action_model': 1.7645532429970004e-05, 'epoch': 0.28}) +[2026-10-03 19:42:26,848][starVLA.training.train_starvla][INFO] - Step 3836, Loss: {'train/loss': 0.00013599336671177298, 'train/grad_norm_pre_clip': 0.05158739164471626, 'timing/data': 0.00022844999330118299, 'timing/model': 0.39021784299984574, 'learning_rate/qwen_vl_interface': 1.7624849335721402e-06, 'learning_rate/action_model': 1.7624849335721402e-05, 'epoch': 0.28}) +[2026-10-03 19:42:27,551][starVLA.training.train_starvla][INFO] - Step 3837, Loss: {'train/loss': 0.0004805075586773455, 'train/grad_norm_pre_clip': 0.35991668701171875, 'timing/data': 0.0002981300058308989, 'timing/model': 0.3932619529950898, 'learning_rate/qwen_vl_interface': 1.7604180577329883e-06, 'learning_rate/action_model': 1.7604180577329883e-05, 'epoch': 0.28}) +[2026-10-03 19:42:28,253][starVLA.training.train_starvla][INFO] - Step 3838, Loss: {'train/loss': 9.327768202638254e-05, 'train/grad_norm_pre_clip': 0.025683514773845673, 'timing/data': 0.0003661800001282245, 'timing/model': 0.39162879198556766, 'learning_rate/qwen_vl_interface': 1.7583526163291604e-06, 'learning_rate/action_model': 1.7583526163291604e-05, 'epoch': 0.28}) +[2026-10-03 19:42:29,155][starVLA.training.train_starvla][INFO] - Step 3839, Loss: {'train/loss': 0.0002911875199060887, 'train/grad_norm_pre_clip': 1.4936013221740723, 'timing/data': 0.00041921998490579426, 'timing/model': 0.39332044299226254, 'learning_rate/qwen_vl_interface': 1.7562886102096807e-06, 'learning_rate/action_model': 1.7562886102096807e-05, 'epoch': 0.28}) +[2026-10-03 19:42:29,856][starVLA.training.train_starvla][INFO] - Step 3840, Loss: {'train/loss': 0.0012726217973977327, 'train/grad_norm_pre_clip': 2.016263961791992, 'timing/data': 0.0002374800096731633, 'timing/model': 0.3930557419953402, 'learning_rate/qwen_vl_interface': 1.7542260402229838e-06, 'learning_rate/action_model': 1.7542260402229838e-05, 'epoch': 0.28}) +[2026-10-03 19:42:30,557][starVLA.training.train_starvla][INFO] - Step 3841, Loss: {'train/loss': 0.00010080602078232914, 'train/grad_norm_pre_clip': 0.05612102523446083, 'timing/data': 0.00020305000361986458, 'timing/model': 0.39328816201305017, 'learning_rate/qwen_vl_interface': 1.7521649072169152e-06, 'learning_rate/action_model': 1.752164907216915e-05, 'epoch': 0.28}) +[2026-10-03 19:42:31,257][starVLA.training.train_starvla][INFO] - Step 3842, Loss: {'train/loss': 0.15716566145420074, 'train/grad_norm_pre_clip': 120.23800659179688, 'timing/data': 0.000370699999621138, 'timing/model': 0.39163403201382607, 'learning_rate/qwen_vl_interface': 1.7501052120387285e-06, 'learning_rate/action_model': 1.7501052120387285e-05, 'epoch': 0.28}) +[2026-10-03 19:42:31,956][starVLA.training.train_starvla][INFO] - Step 3843, Loss: {'train/loss': 4.6041437599342316e-05, 'train/grad_norm_pre_clip': 0.011134916916489601, 'timing/data': 0.0003137800085823983, 'timing/model': 0.3903831719944719, 'learning_rate/qwen_vl_interface': 1.7480469555350866e-06, 'learning_rate/action_model': 1.7480469555350865e-05, 'epoch': 0.28}) +[2026-10-03 19:42:32,857][starVLA.training.train_starvla][INFO] - Step 3844, Loss: {'train/loss': 8.996044925879687e-05, 'train/grad_norm_pre_clip': 0.026156628504395485, 'timing/data': 0.0002035590005107224, 'timing/model': 0.39017227300792, 'learning_rate/qwen_vl_interface': 1.7459901385520619e-06, 'learning_rate/action_model': 1.745990138552062e-05, 'epoch': 0.28}) +[2026-10-03 19:42:33,553][starVLA.training.train_starvla][INFO] - Step 3845, Loss: {'train/loss': 0.006776716560125351, 'train/grad_norm_pre_clip': 13.786145210266113, 'timing/data': 0.00034515999141149223, 'timing/model': 0.38931602297816426, 'learning_rate/qwen_vl_interface': 1.7439347619351324e-06, 'learning_rate/action_model': 1.7439347619351324e-05, 'epoch': 0.28}) +[2026-10-03 19:42:34,251][starVLA.training.train_starvla][INFO] - Step 3846, Loss: {'train/loss': 6.16097531747073e-05, 'train/grad_norm_pre_clip': 0.019831396639347076, 'timing/data': 0.0002640699967741966, 'timing/model': 0.39091826300136745, 'learning_rate/qwen_vl_interface': 1.741880826529186e-06, 'learning_rate/action_model': 1.741880826529186e-05, 'epoch': 0.28}) +[2026-10-03 19:42:34,946][starVLA.training.train_starvla][INFO] - Step 3847, Loss: {'train/loss': 0.003238063072785735, 'train/grad_norm_pre_clip': 5.799049377441406, 'timing/data': 0.0003321100084576756, 'timing/model': 0.38787342299474403, 'learning_rate/qwen_vl_interface': 1.7398283331785205e-06, 'learning_rate/action_model': 1.7398283331785203e-05, 'epoch': 0.28}) +[2026-10-03 19:42:35,644][starVLA.training.train_starvla][INFO] - Step 3848, Loss: {'train/loss': 0.00015090336091816425, 'train/grad_norm_pre_clip': 0.06267905235290527, 'timing/data': 0.00023315998259931803, 'timing/model': 0.39126695299637504, 'learning_rate/qwen_vl_interface': 1.737777282726835e-06, 'learning_rate/action_model': 1.737777282726835e-05, 'epoch': 0.28}) +[2026-10-03 19:42:36,539][starVLA.training.train_starvla][INFO] - Step 3849, Loss: {'train/loss': 0.3169390559196472, 'train/grad_norm_pre_clip': 82.0715560913086, 'timing/data': 0.000258120009675622, 'timing/model': 0.5879490739898756, 'learning_rate/qwen_vl_interface': 1.7357276760172403e-06, 'learning_rate/action_model': 1.7357276760172404e-05, 'epoch': 0.28}) +[2026-10-03 19:42:37,237][starVLA.training.train_starvla][INFO] - Step 3850, Loss: {'train/loss': 9.930936357704923e-05, 'train/grad_norm_pre_clip': 0.03701025992631912, 'timing/data': 0.0003856000257655978, 'timing/model': 0.39041632297448814, 'learning_rate/qwen_vl_interface': 1.733679513892252e-06, 'learning_rate/action_model': 1.7336795138922517e-05, 'epoch': 0.28}) +[2026-10-03 19:42:37,934][starVLA.training.train_starvla][INFO] - Step 3851, Loss: {'train/loss': 0.0003561134508345276, 'train/grad_norm_pre_clip': 0.15700633823871613, 'timing/data': 0.0002952399954665452, 'timing/model': 0.39063291301135905, 'learning_rate/qwen_vl_interface': 1.7316327971937926e-06, 'learning_rate/action_model': 1.7316327971937926e-05, 'epoch': 0.28}) +[2026-10-03 19:42:38,630][starVLA.training.train_starvla][INFO] - Step 3852, Loss: {'train/loss': 0.7117434144020081, 'train/grad_norm_pre_clip': 249.27304077148438, 'timing/data': 0.00031126997782848775, 'timing/model': 0.39084204300888814, 'learning_rate/qwen_vl_interface': 1.7295875267631904e-06, 'learning_rate/action_model': 1.7295875267631904e-05, 'epoch': 0.28}) +[2026-10-03 19:42:39,329][starVLA.training.train_starvla][INFO] - Step 3853, Loss: {'train/loss': 2.078650868497789e-05, 'train/grad_norm_pre_clip': 0.004533842206001282, 'timing/data': 0.0002827700227499008, 'timing/model': 0.3912276720220689, 'learning_rate/qwen_vl_interface': 1.7275437034411778e-06, 'learning_rate/action_model': 1.7275437034411776e-05, 'epoch': 0.28}) +[2026-10-03 19:42:40,227][starVLA.training.train_starvla][INFO] - Step 3854, Loss: {'train/loss': 0.025692738592624664, 'train/grad_norm_pre_clip': 50.86847686767578, 'timing/data': 0.00020871000015176833, 'timing/model': 0.5915321529901121, 'learning_rate/qwen_vl_interface': 1.7255013280678941e-06, 'learning_rate/action_model': 1.725501328067894e-05, 'epoch': 0.28}) +[2026-10-03 19:42:40,924][starVLA.training.train_starvla][INFO] - Step 3855, Loss: {'train/loss': 0.4457361102104187, 'train/grad_norm_pre_clip': 42.447288513183594, 'timing/data': 0.0002165599726140499, 'timing/model': 0.3907198020024225, 'learning_rate/qwen_vl_interface': 1.7234604014828824e-06, 'learning_rate/action_model': 1.7234604014828823e-05, 'epoch': 0.28}) +[2026-10-03 19:42:41,620][starVLA.training.train_starvla][INFO] - Step 3856, Loss: {'train/loss': 0.00013857030717190355, 'train/grad_norm_pre_clip': 0.09558892995119095, 'timing/data': 0.00020353999570943415, 'timing/model': 0.38873875301214866, 'learning_rate/qwen_vl_interface': 1.7214209245250906e-06, 'learning_rate/action_model': 1.7214209245250905e-05, 'epoch': 0.28}) +[2026-10-03 19:42:42,317][starVLA.training.train_starvla][INFO] - Step 3857, Loss: {'train/loss': 0.0004269468190614134, 'train/grad_norm_pre_clip': 0.38082388043403625, 'timing/data': 0.00020527999731712043, 'timing/model': 0.3905459629895631, 'learning_rate/qwen_vl_interface': 1.7193828980328723e-06, 'learning_rate/action_model': 1.719382898032872e-05, 'epoch': 0.28}) +[2026-10-03 19:42:43,012][starVLA.training.train_starvla][INFO] - Step 3858, Loss: {'train/loss': 0.1290919929742813, 'train/grad_norm_pre_clip': 105.42137908935547, 'timing/data': 0.0002055389923043549, 'timing/model': 0.38861594401532784, 'learning_rate/qwen_vl_interface': 1.71734632284398e-06, 'learning_rate/action_model': 1.7173463228439798e-05, 'epoch': 0.28}) +[2026-10-03 19:42:43,710][starVLA.training.train_starvla][INFO] - Step 3859, Loss: {'train/loss': 3.379373447387479e-05, 'train/grad_norm_pre_clip': 0.00824009720236063, 'timing/data': 0.0002009300224017352, 'timing/model': 0.39076217298861593, 'learning_rate/qwen_vl_interface': 1.7153111997955744e-06, 'learning_rate/action_model': 1.7153111997955743e-05, 'epoch': 0.28}) +[2026-10-03 19:42:44,608][starVLA.training.train_starvla][INFO] - Step 3860, Loss: {'train/loss': 0.00010358197323512286, 'train/grad_norm_pre_clip': 0.034369468688964844, 'timing/data': 0.0002129299973603338, 'timing/model': 0.3887981429870706, 'learning_rate/qwen_vl_interface': 1.7132775297242175e-06, 'learning_rate/action_model': 1.7132775297242172e-05, 'epoch': 0.28}) +[2026-10-03 19:42:45,306][starVLA.training.train_starvla][INFO] - Step 3861, Loss: {'train/loss': 3.255735646234825e-05, 'train/grad_norm_pre_clip': 0.021901655942201614, 'timing/data': 0.00019707001047208905, 'timing/model': 0.3915868120093364, 'learning_rate/qwen_vl_interface': 1.7112453134658747e-06, 'learning_rate/action_model': 1.7112453134658747e-05, 'epoch': 0.28}) +[2026-10-03 19:42:46,004][starVLA.training.train_starvla][INFO] - Step 3862, Loss: {'train/loss': 0.0004899641498923302, 'train/grad_norm_pre_clip': 0.6219779849052429, 'timing/data': 0.00021865000599063933, 'timing/model': 0.3905813519959338, 'learning_rate/qwen_vl_interface': 1.7092145518559129e-06, 'learning_rate/action_model': 1.709214551855913e-05, 'epoch': 0.28}) +[2026-10-03 19:42:46,702][starVLA.training.train_starvla][INFO] - Step 3863, Loss: {'train/loss': 0.5194048881530762, 'train/grad_norm_pre_clip': 44.22383499145508, 'timing/data': 0.00020310000400058925, 'timing/model': 0.38985266198869795, 'learning_rate/qwen_vl_interface': 1.7071852457291013e-06, 'learning_rate/action_model': 1.707185245729101e-05, 'epoch': 0.28}) +[2026-10-03 19:42:47,401][starVLA.training.train_starvla][INFO] - Step 3864, Loss: {'train/loss': 0.007644009776413441, 'train/grad_norm_pre_clip': 8.896599769592285, 'timing/data': 0.00022822999744676054, 'timing/model': 0.3913634619966615, 'learning_rate/qwen_vl_interface': 1.7051573959196122e-06, 'learning_rate/action_model': 1.7051573959196122e-05, 'epoch': 0.28}) +[2026-10-03 19:42:48,297][starVLA.training.train_starvla][INFO] - Step 3865, Loss: {'train/loss': 0.33467918634414673, 'train/grad_norm_pre_clip': 100.01066589355469, 'timing/data': 0.0003596800088416785, 'timing/model': 0.5872870240127668, 'learning_rate/qwen_vl_interface': 1.7031310032610166e-06, 'learning_rate/action_model': 1.7031310032610163e-05, 'epoch': 0.28}) +[2026-10-03 19:42:48,998][starVLA.training.train_starvla][INFO] - Step 3866, Loss: {'train/loss': 0.5016462802886963, 'train/grad_norm_pre_clip': 52.8573112487793, 'timing/data': 0.00022656997316516936, 'timing/model': 0.3911680630117189, 'learning_rate/qwen_vl_interface': 1.701106068586289e-06, 'learning_rate/action_model': 1.701106068586289e-05, 'epoch': 0.28}) +[2026-10-03 19:42:49,696][starVLA.training.train_starvla][INFO] - Step 3867, Loss: {'train/loss': 0.00012431407230906188, 'train/grad_norm_pre_clip': 0.08070749789476395, 'timing/data': 0.000205210002604872, 'timing/model': 0.3907114729809109, 'learning_rate/qwen_vl_interface': 1.6990825927278032e-06, 'learning_rate/action_model': 1.699082592727803e-05, 'epoch': 0.28}) +[2026-10-03 19:42:50,392][starVLA.training.train_starvla][INFO] - Step 3868, Loss: {'train/loss': 0.10485474020242691, 'train/grad_norm_pre_clip': 111.30419158935547, 'timing/data': 0.00024796000798232853, 'timing/model': 0.3896614529949147, 'learning_rate/qwen_vl_interface': 1.6970605765173344e-06, 'learning_rate/action_model': 1.6970605765173346e-05, 'epoch': 0.28}) +[2026-10-03 19:42:51,091][starVLA.training.train_starvla][INFO] - Step 3869, Loss: {'train/loss': 0.6321613788604736, 'train/grad_norm_pre_clip': 101.66397094726562, 'timing/data': 0.00021959899459034204, 'timing/model': 0.3906235529866535, 'learning_rate/qwen_vl_interface': 1.6950400207860557e-06, 'learning_rate/action_model': 1.6950400207860557e-05, 'epoch': 0.28}) +[2026-10-03 19:42:51,992][starVLA.training.train_starvla][INFO] - Step 3870, Loss: {'train/loss': 2.558470623625908e-05, 'train/grad_norm_pre_clip': 0.00501753855496645, 'timing/data': 0.0002067500026896596, 'timing/model': 0.5926233129866887, 'learning_rate/qwen_vl_interface': 1.6930209263645412e-06, 'learning_rate/action_model': 1.6930209263645412e-05, 'epoch': 0.28}) +[2026-10-03 19:42:52,689][starVLA.training.train_starvla][INFO] - Step 3871, Loss: {'train/loss': 0.0011498542735353112, 'train/grad_norm_pre_clip': 1.1820635795593262, 'timing/data': 0.0003449800133239478, 'timing/model': 0.39142782299313694, 'learning_rate/qwen_vl_interface': 1.6910032940827656e-06, 'learning_rate/action_model': 1.6910032940827657e-05, 'epoch': 0.28}) +[2026-10-03 19:42:53,385][starVLA.training.train_starvla][INFO] - Step 3872, Loss: {'train/loss': 0.00042741233482956886, 'train/grad_norm_pre_clip': 0.2070537507534027, 'timing/data': 0.00020422000670805573, 'timing/model': 0.39089415199123323, 'learning_rate/qwen_vl_interface': 1.6889871247701003e-06, 'learning_rate/action_model': 1.6889871247701002e-05, 'epoch': 0.28}) +[2026-10-03 19:42:54,083][starVLA.training.train_starvla][INFO] - Step 3873, Loss: {'train/loss': 0.002471755724400282, 'train/grad_norm_pre_clip': 2.2733004093170166, 'timing/data': 0.0002793000021483749, 'timing/model': 0.391334112995537, 'learning_rate/qwen_vl_interface': 1.6869724192553168e-06, 'learning_rate/action_model': 1.686972419255317e-05, 'epoch': 0.28}) +[2026-10-03 19:42:54,783][starVLA.training.train_starvla][INFO] - Step 3874, Loss: {'train/loss': 0.3667266070842743, 'train/grad_norm_pre_clip': 101.30523681640625, 'timing/data': 0.00032258001738227904, 'timing/model': 0.39003759299521334, 'learning_rate/qwen_vl_interface': 1.6849591783665847e-06, 'learning_rate/action_model': 1.6849591783665847e-05, 'epoch': 0.28}) +[2026-10-03 19:42:55,486][starVLA.training.train_starvla][INFO] - Step 3875, Loss: {'train/loss': 0.9652909636497498, 'train/grad_norm_pre_clip': 118.63224792480469, 'timing/data': 0.00031065999064594507, 'timing/model': 0.39660043199546635, 'learning_rate/qwen_vl_interface': 1.6829474029314713e-06, 'learning_rate/action_model': 1.6829474029314713e-05, 'epoch': 0.28}) +[2026-10-03 19:42:56,392][starVLA.training.train_starvla][INFO] - Step 3876, Loss: {'train/loss': 0.21852582693099976, 'train/grad_norm_pre_clip': 133.23944091796875, 'timing/data': 0.0003197600017301738, 'timing/model': 0.39581847202498466, 'learning_rate/qwen_vl_interface': 1.6809370937769408e-06, 'learning_rate/action_model': 1.680937093776941e-05, 'epoch': 0.28}) +[2026-10-03 19:42:57,098][starVLA.training.train_starvla][INFO] - Step 3877, Loss: {'train/loss': 0.08258996158838272, 'train/grad_norm_pre_clip': 208.96612548828125, 'timing/data': 0.00022513000294566154, 'timing/model': 0.3936742329970002, 'learning_rate/qwen_vl_interface': 1.678928251729356e-06, 'learning_rate/action_model': 1.6789282517293557e-05, 'epoch': 0.28}) +[2026-10-03 19:42:57,804][starVLA.training.train_starvla][INFO] - Step 3878, Loss: {'train/loss': 0.3668426275253296, 'train/grad_norm_pre_clip': 87.76953125, 'timing/data': 0.00037455998244695365, 'timing/model': 0.3938729720248375, 'learning_rate/qwen_vl_interface': 1.6769208776144767e-06, 'learning_rate/action_model': 1.6769208776144764e-05, 'epoch': 0.28}) +[2026-10-03 19:42:58,511][starVLA.training.train_starvla][INFO] - Step 3879, Loss: {'train/loss': 0.40385937690734863, 'train/grad_norm_pre_clip': 183.18333435058594, 'timing/data': 0.00027329998556524515, 'timing/model': 0.394802922004601, 'learning_rate/qwen_vl_interface': 1.6749149722574564e-06, 'learning_rate/action_model': 1.6749149722574564e-05, 'epoch': 0.28}) +[2026-10-03 19:42:59,218][starVLA.training.train_starvla][INFO] - Step 3880, Loss: {'train/loss': 0.0001638189860386774, 'train/grad_norm_pre_clip': 0.10225378721952438, 'timing/data': 0.0002617599966470152, 'timing/model': 0.3952134730061516, 'learning_rate/qwen_vl_interface': 1.6729105364828486e-06, 'learning_rate/action_model': 1.6729105364828484e-05, 'epoch': 0.28}) +[2026-10-03 19:43:00,122][starVLA.training.train_starvla][INFO] - Step 3881, Loss: {'train/loss': 0.09233680367469788, 'train/grad_norm_pre_clip': 91.91468811035156, 'timing/data': 0.00034826999763026834, 'timing/model': 0.3948503919818904, 'learning_rate/qwen_vl_interface': 1.6709075711146e-06, 'learning_rate/action_model': 1.6709075711146e-05, 'epoch': 0.28}) +[2026-10-03 19:43:00,829][starVLA.training.train_starvla][INFO] - Step 3882, Loss: {'train/loss': 0.0003728411393240094, 'train/grad_norm_pre_clip': 0.4829341769218445, 'timing/data': 0.00032341000041924417, 'timing/model': 0.3946658619970549, 'learning_rate/qwen_vl_interface': 1.6689060769760556e-06, 'learning_rate/action_model': 1.6689060769760557e-05, 'epoch': 0.28}) +[2026-10-03 19:43:01,534][starVLA.training.train_starvla][INFO] - Step 3883, Loss: {'train/loss': 0.0003166850656270981, 'train/grad_norm_pre_clip': 0.20591291785240173, 'timing/data': 0.0003167400136590004, 'timing/model': 0.3949689529836178, 'learning_rate/qwen_vl_interface': 1.6669060548899528e-06, 'learning_rate/action_model': 1.6669060548899525e-05, 'epoch': 0.28}) +[2026-10-03 19:43:02,241][starVLA.training.train_starvla][INFO] - Step 3884, Loss: {'train/loss': 0.036955852061510086, 'train/grad_norm_pre_clip': 57.78571319580078, 'timing/data': 0.0002976100076921284, 'timing/model': 0.39563065199763514, 'learning_rate/qwen_vl_interface': 1.6649075056784253e-06, 'learning_rate/action_model': 1.6649075056784253e-05, 'epoch': 0.28}) +[2026-10-03 19:43:02,946][starVLA.training.train_starvla][INFO] - Step 3885, Loss: {'train/loss': 0.0009143486386165023, 'train/grad_norm_pre_clip': 0.9845335483551025, 'timing/data': 0.00036298000486567616, 'timing/model': 0.394620572013082, 'learning_rate/qwen_vl_interface': 1.6629104301630019e-06, 'learning_rate/action_model': 1.662910430163002e-05, 'epoch': 0.28}) +[2026-10-03 19:43:03,841][starVLA.training.train_starvla][INFO] - Step 3886, Loss: {'train/loss': 0.19513404369354248, 'train/grad_norm_pre_clip': 112.43550872802734, 'timing/data': 0.00033566998899914324, 'timing/model': 0.3947952019807417, 'learning_rate/qwen_vl_interface': 1.6609148291646043e-06, 'learning_rate/action_model': 1.6609148291646042e-05, 'epoch': 0.28}) +[2026-10-03 19:43:04,549][starVLA.training.train_starvla][INFO] - Step 3887, Loss: {'train/loss': 0.00011970661580562592, 'train/grad_norm_pre_clip': 0.024935998022556305, 'timing/data': 0.0003032099921256304, 'timing/model': 0.39559303200803697, 'learning_rate/qwen_vl_interface': 1.658920703503549e-06, 'learning_rate/action_model': 1.6589207035035488e-05, 'epoch': 0.28}) +[2026-10-03 19:43:05,253][starVLA.training.train_starvla][INFO] - Step 3888, Loss: {'train/loss': 0.5324674844741821, 'train/grad_norm_pre_clip': 107.7247543334961, 'timing/data': 0.0003535300202202052, 'timing/model': 0.39398542197886854, 'learning_rate/qwen_vl_interface': 1.6569280539995467e-06, 'learning_rate/action_model': 1.6569280539995467e-05, 'epoch': 0.28}) +[2026-10-03 19:43:05,959][starVLA.training.train_starvla][INFO] - Step 3889, Loss: {'train/loss': 0.0001410233962815255, 'train/grad_norm_pre_clip': 0.05828479304909706, 'timing/data': 0.00023308998788706958, 'timing/model': 0.3938693029922433, 'learning_rate/qwen_vl_interface': 1.6549368814716977e-06, 'learning_rate/action_model': 1.6549368814716977e-05, 'epoch': 0.28}) +[2026-10-03 19:43:06,667][starVLA.training.train_starvla][INFO] - Step 3890, Loss: {'train/loss': 7.244524749694392e-05, 'train/grad_norm_pre_clip': 0.03979633376002312, 'timing/data': 0.0003725099959410727, 'timing/model': 0.395083612005692, 'learning_rate/qwen_vl_interface': 1.6529471867385009e-06, 'learning_rate/action_model': 1.6529471867385007e-05, 'epoch': 0.28}) +[2026-10-03 19:43:07,566][starVLA.training.train_starvla][INFO] - Step 3891, Loss: {'train/loss': 0.022105522453784943, 'train/grad_norm_pre_clip': 90.24922943115234, 'timing/data': 0.00032975999056361616, 'timing/model': 0.5886250440089498, 'learning_rate/qwen_vl_interface': 1.6509589706178417e-06, 'learning_rate/action_model': 1.6509589706178417e-05, 'epoch': 0.28}) +[2026-10-03 19:43:08,274][starVLA.training.train_starvla][INFO] - Step 3892, Loss: {'train/loss': 0.00018872706277761608, 'train/grad_norm_pre_clip': 0.15111292898654938, 'timing/data': 0.00035951900645159185, 'timing/model': 0.3958398929971736, 'learning_rate/qwen_vl_interface': 1.648972233927002e-06, 'learning_rate/action_model': 1.648972233927002e-05, 'epoch': 0.28}) +[2026-10-03 19:43:08,980][starVLA.training.train_starvla][INFO] - Step 3893, Loss: {'train/loss': 0.00014832198212388903, 'train/grad_norm_pre_clip': 0.03425286337733269, 'timing/data': 0.00048367000999860466, 'timing/model': 0.3951753620058298, 'learning_rate/qwen_vl_interface': 1.646986977482654e-06, 'learning_rate/action_model': 1.6469869774826538e-05, 'epoch': 0.28}) +[2026-10-03 19:43:09,686][starVLA.training.train_starvla][INFO] - Step 3894, Loss: {'train/loss': 0.0048722922801971436, 'train/grad_norm_pre_clip': 7.002060890197754, 'timing/data': 0.00022160998196341097, 'timing/model': 0.39399100199807435, 'learning_rate/qwen_vl_interface': 1.6450032021008618e-06, 'learning_rate/action_model': 1.6450032021008617e-05, 'epoch': 0.29}) +[2026-10-03 19:43:10,390][starVLA.training.train_starvla][INFO] - Step 3895, Loss: {'train/loss': 0.011980458162724972, 'train/grad_norm_pre_clip': 18.763093948364258, 'timing/data': 0.00029056999483145773, 'timing/model': 0.39372101199114695, 'learning_rate/qwen_vl_interface': 1.6430209085970812e-06, 'learning_rate/action_model': 1.643020908597081e-05, 'epoch': 0.29}) +[2026-10-03 19:43:11,290][starVLA.training.train_starvla][INFO] - Step 3896, Loss: {'train/loss': 0.7569738626480103, 'train/grad_norm_pre_clip': 139.23907470703125, 'timing/data': 0.00026006001280620694, 'timing/model': 0.5879865239840001, 'learning_rate/qwen_vl_interface': 1.641040097786157e-06, 'learning_rate/action_model': 1.641040097786157e-05, 'epoch': 0.29}) +[2026-10-03 19:43:11,994][starVLA.training.train_starvla][INFO] - Step 3897, Loss: {'train/loss': 1.2314339876174927, 'train/grad_norm_pre_clip': 198.1141815185547, 'timing/data': 0.0003648199781309813, 'timing/model': 0.3922385920013767, 'learning_rate/qwen_vl_interface': 1.6390607704823267e-06, 'learning_rate/action_model': 1.6390607704823265e-05, 'epoch': 0.29}) +[2026-10-03 19:43:12,697][starVLA.training.train_starvla][INFO] - Step 3898, Loss: {'train/loss': 2.2112133592600003e-05, 'train/grad_norm_pre_clip': 0.009769316762685776, 'timing/data': 0.00026125999283976853, 'timing/model': 0.39239362298394553, 'learning_rate/qwen_vl_interface': 1.6370829274992162e-06, 'learning_rate/action_model': 1.6370829274992162e-05, 'epoch': 0.29}) +[2026-10-03 19:43:13,404][starVLA.training.train_starvla][INFO] - Step 3899, Loss: {'train/loss': 3.460625157458708e-05, 'train/grad_norm_pre_clip': 0.0078042903915047646, 'timing/data': 0.0002964800223708153, 'timing/model': 0.39510373200755566, 'learning_rate/qwen_vl_interface': 1.6351065696498422e-06, 'learning_rate/action_model': 1.6351065696498422e-05, 'epoch': 0.29}) +[2026-10-03 19:43:14,109][starVLA.training.train_starvla][INFO] - Step 3900, Loss: {'train/loss': 0.03584263101220131, 'train/grad_norm_pre_clip': 56.63397979736328, 'timing/data': 0.00029948001611046493, 'timing/model': 0.3948911629850045, 'learning_rate/qwen_vl_interface': 1.6331316977466117e-06, 'learning_rate/action_model': 1.6331316977466115e-05, 'epoch': 0.29}) +[2026-10-03 19:43:14,816][starVLA.training.train_starvla][INFO] - Step 3901, Loss: {'train/loss': 6.646498513873667e-05, 'train/grad_norm_pre_clip': 0.03283870592713356, 'timing/data': 0.0002670300018507987, 'timing/model': 0.39458081297925673, 'learning_rate/qwen_vl_interface': 1.6311583126013198e-06, 'learning_rate/action_model': 1.6311583126013198e-05, 'epoch': 0.29}) +[2026-10-03 19:43:15,721][starVLA.training.train_starvla][INFO] - Step 3902, Loss: {'train/loss': 0.0013463383074849844, 'train/grad_norm_pre_clip': 4.107420444488525, 'timing/data': 0.0003597899922169745, 'timing/model': 0.39459369200631045, 'learning_rate/qwen_vl_interface': 1.6291864150251488e-06, 'learning_rate/action_model': 1.629186415025149e-05, 'epoch': 0.29}) +[2026-10-03 19:43:16,432][starVLA.training.train_starvla][INFO] - Step 3903, Loss: {'train/loss': 0.4140688180923462, 'train/grad_norm_pre_clip': 42.12629699707031, 'timing/data': 0.00032096999348141253, 'timing/model': 0.39649575101793744, 'learning_rate/qwen_vl_interface': 1.6272160058286727e-06, 'learning_rate/action_model': 1.6272160058286726e-05, 'epoch': 0.29}) +[2026-10-03 19:43:17,144][starVLA.training.train_starvla][INFO] - Step 3904, Loss: {'train/loss': 2.255226718261838e-05, 'train/grad_norm_pre_clip': 0.0049634031020104885, 'timing/data': 0.0002555200189817697, 'timing/model': 0.39736555199488066, 'learning_rate/qwen_vl_interface': 1.6252470858218519e-06, 'learning_rate/action_model': 1.6252470858218517e-05, 'epoch': 0.29}) +[2026-10-03 19:43:17,855][starVLA.training.train_starvla][INFO] - Step 3905, Loss: {'train/loss': 0.000251890771323815, 'train/grad_norm_pre_clip': 0.08486920595169067, 'timing/data': 0.000268169998889789, 'timing/model': 0.3969049520092085, 'learning_rate/qwen_vl_interface': 1.6232796558140357e-06, 'learning_rate/action_model': 1.6232796558140357e-05, 'epoch': 0.29}) +[2026-10-03 19:43:18,564][starVLA.training.train_starvla][INFO] - Step 3906, Loss: {'train/loss': 3.72516515199095e-05, 'train/grad_norm_pre_clip': 0.008814549073576927, 'timing/data': 0.00025418002041988075, 'timing/model': 0.39526928201667033, 'learning_rate/qwen_vl_interface': 1.6213137166139592e-06, 'learning_rate/action_model': 1.621313716613959e-05, 'epoch': 0.29}) +[2026-10-03 19:43:19,476][starVLA.training.train_starvla][INFO] - Step 3907, Loss: {'train/loss': 6.740288517903537e-05, 'train/grad_norm_pre_clip': 0.01822364144027233, 'timing/data': 0.00025882001500576735, 'timing/model': 0.5978254929941613, 'learning_rate/qwen_vl_interface': 1.6193492690297463e-06, 'learning_rate/action_model': 1.6193492690297464e-05, 'epoch': 0.29}) +[2026-10-03 19:43:20,183][starVLA.training.train_starvla][INFO] - Step 3908, Loss: {'train/loss': 0.0011263805208727717, 'train/grad_norm_pre_clip': 1.1054911613464355, 'timing/data': 0.00026544000138528645, 'timing/model': 0.3949690219887998, 'learning_rate/qwen_vl_interface': 1.6173863138689078e-06, 'learning_rate/action_model': 1.617386313868908e-05, 'epoch': 0.29}) +[2026-10-03 19:43:20,892][starVLA.training.train_starvla][INFO] - Step 3909, Loss: {'train/loss': 0.00013094338646624237, 'train/grad_norm_pre_clip': 0.08329760283231735, 'timing/data': 0.0002801200025714934, 'timing/model': 0.39528927198261954, 'learning_rate/qwen_vl_interface': 1.6154248519383396e-06, 'learning_rate/action_model': 1.6154248519383395e-05, 'epoch': 0.29}) +[2026-10-03 19:43:21,604][starVLA.training.train_starvla][INFO] - Step 3910, Loss: {'train/loss': 0.30228570103645325, 'train/grad_norm_pre_clip': 75.28382873535156, 'timing/data': 0.00041494000470265746, 'timing/model': 0.39776489199721254, 'learning_rate/qwen_vl_interface': 1.6134648840443248e-06, 'learning_rate/action_model': 1.613464884044325e-05, 'epoch': 0.29}) +[2026-10-03 19:43:22,314][starVLA.training.train_starvla][INFO] - Step 3911, Loss: {'train/loss': 3.9120219298638403e-05, 'train/grad_norm_pre_clip': 0.02121960185468197, 'timing/data': 0.00022616999922320247, 'timing/model': 0.39688797199050896, 'learning_rate/qwen_vl_interface': 1.6115064109925334e-06, 'learning_rate/action_model': 1.6115064109925336e-05, 'epoch': 0.29}) +[2026-10-03 19:43:23,025][starVLA.training.train_starvla][INFO] - Step 3912, Loss: {'train/loss': 1.934839201567229e-05, 'train/grad_norm_pre_clip': 0.006253828294575214, 'timing/data': 0.00021572000696323812, 'timing/model': 0.3969213420059532, 'learning_rate/qwen_vl_interface': 1.6095494335880173e-06, 'learning_rate/action_model': 1.6095494335880174e-05, 'epoch': 0.29}) +[2026-10-03 19:43:23,936][starVLA.training.train_starvla][INFO] - Step 3913, Loss: {'train/loss': 0.5084406137466431, 'train/grad_norm_pre_clip': 65.69255065917969, 'timing/data': 0.00039507998735643923, 'timing/model': 0.3969113220227882, 'learning_rate/qwen_vl_interface': 1.6075939526352175e-06, 'learning_rate/action_model': 1.6075939526352175e-05, 'epoch': 0.29}) +[2026-10-03 19:43:24,645][starVLA.training.train_starvla][INFO] - Step 3914, Loss: {'train/loss': 0.00014104189176578075, 'train/grad_norm_pre_clip': 0.05517585203051567, 'timing/data': 0.0003948899975512177, 'timing/model': 0.39588199197896756, 'learning_rate/qwen_vl_interface': 1.6056399689379576e-06, 'learning_rate/action_model': 1.6056399689379575e-05, 'epoch': 0.29}) +[2026-10-03 19:43:25,354][starVLA.training.train_starvla][INFO] - Step 3915, Loss: {'train/loss': 0.04484947770833969, 'train/grad_norm_pre_clip': 137.71446228027344, 'timing/data': 0.00022901000920683146, 'timing/model': 0.3965164620021824, 'learning_rate/qwen_vl_interface': 1.6036874832994473e-06, 'learning_rate/action_model': 1.603687483299447e-05, 'epoch': 0.29}) +[2026-10-03 19:43:26,064][starVLA.training.train_starvla][INFO] - Step 3916, Loss: {'train/loss': 0.5089595317840576, 'train/grad_norm_pre_clip': 28.296354293823242, 'timing/data': 0.0002140600117854774, 'timing/model': 0.39624936200561933, 'learning_rate/qwen_vl_interface': 1.6017364965222796e-06, 'learning_rate/action_model': 1.6017364965222793e-05, 'epoch': 0.29}) +[2026-10-03 19:43:26,774][starVLA.training.train_starvla][INFO] - Step 3917, Loss: {'train/loss': 0.3682258427143097, 'train/grad_norm_pre_clip': 64.6542739868164, 'timing/data': 0.00035792001290246844, 'timing/model': 0.3967264319944661, 'learning_rate/qwen_vl_interface': 1.5997870094084305e-06, 'learning_rate/action_model': 1.5997870094084306e-05, 'epoch': 0.29}) +[2026-10-03 19:43:27,677][starVLA.training.train_starvla][INFO] - Step 3918, Loss: {'train/loss': 0.2592073976993561, 'train/grad_norm_pre_clip': 86.73611450195312, 'timing/data': 0.00032036000629886985, 'timing/model': 0.39607215198338963, 'learning_rate/qwen_vl_interface': 1.5978390227592616e-06, 'learning_rate/action_model': 1.5978390227592615e-05, 'epoch': 0.29}) +[2026-10-03 19:43:28,387][starVLA.training.train_starvla][INFO] - Step 3919, Loss: {'train/loss': 0.38529443740844727, 'train/grad_norm_pre_clip': 93.5804443359375, 'timing/data': 0.0003540700126904994, 'timing/model': 0.39596602201345377, 'learning_rate/qwen_vl_interface': 1.5958925373755164e-06, 'learning_rate/action_model': 1.5958925373755164e-05, 'epoch': 0.29}) +[2026-10-03 19:43:29,096][starVLA.training.train_starvla][INFO] - Step 3920, Loss: {'train/loss': 0.29008597135543823, 'train/grad_norm_pre_clip': 54.57312774658203, 'timing/data': 0.0002661400067154318, 'timing/model': 0.39603195199742913, 'learning_rate/qwen_vl_interface': 1.5939475540573213e-06, 'learning_rate/action_model': 1.5939475540573212e-05, 'epoch': 0.29}) +[2026-10-03 19:43:29,805][starVLA.training.train_starvla][INFO] - Step 3921, Loss: {'train/loss': 0.009071798995137215, 'train/grad_norm_pre_clip': 22.260286331176758, 'timing/data': 0.00035616999957710505, 'timing/model': 0.39574975200230256, 'learning_rate/qwen_vl_interface': 1.5920040736041867e-06, 'learning_rate/action_model': 1.5920040736041866e-05, 'epoch': 0.29}) +[2026-10-03 19:43:30,517][starVLA.training.train_starvla][INFO] - Step 3922, Loss: {'train/loss': 0.0003909636288881302, 'train/grad_norm_pre_clip': 0.17534886300563812, 'timing/data': 0.0002983800077345222, 'timing/model': 0.3975354919966776, 'learning_rate/qwen_vl_interface': 1.5900620968150015e-06, 'learning_rate/action_model': 1.5900620968150015e-05, 'epoch': 0.29}) +[2026-10-03 19:43:31,430][starVLA.training.train_starvla][INFO] - Step 3923, Loss: {'train/loss': 5.983364462736063e-05, 'train/grad_norm_pre_clip': 0.018766572698950768, 'timing/data': 0.00022437001462094486, 'timing/model': 0.5994558420206886, 'learning_rate/qwen_vl_interface': 1.5881216244880407e-06, 'learning_rate/action_model': 1.5881216244880407e-05, 'epoch': 0.29}) +[2026-10-03 19:43:32,139][starVLA.training.train_starvla][INFO] - Step 3924, Loss: {'train/loss': 0.6115034222602844, 'train/grad_norm_pre_clip': 50.951332092285156, 'timing/data': 0.0002246599760837853, 'timing/model': 0.39644736199988984, 'learning_rate/qwen_vl_interface': 1.5861826574209587e-06, 'learning_rate/action_model': 1.5861826574209587e-05, 'epoch': 0.29}) +[2026-10-03 19:43:32,849][starVLA.training.train_starvla][INFO] - Step 3925, Loss: {'train/loss': 0.00011153993546031415, 'train/grad_norm_pre_clip': 0.028689641505479813, 'timing/data': 0.0003661199880298227, 'timing/model': 0.3959623720147647, 'learning_rate/qwen_vl_interface': 1.5842451964107916e-06, 'learning_rate/action_model': 1.5842451964107916e-05, 'epoch': 0.29}) +[2026-10-03 19:43:33,559][starVLA.training.train_starvla][INFO] - Step 3926, Loss: {'train/loss': 0.0017140634590759873, 'train/grad_norm_pre_clip': 2.3663365840911865, 'timing/data': 0.000272019999101758, 'timing/model': 0.39647151200915687, 'learning_rate/qwen_vl_interface': 1.5823092422539567e-06, 'learning_rate/action_model': 1.582309242253957e-05, 'epoch': 0.29}) +[2026-10-03 19:43:34,269][starVLA.training.train_starvla][INFO] - Step 3927, Loss: {'train/loss': 0.0002217401925008744, 'train/grad_norm_pre_clip': 0.053483281284570694, 'timing/data': 0.0003294500056654215, 'timing/model': 0.39626306100399233, 'learning_rate/qwen_vl_interface': 1.5803747957462514e-06, 'learning_rate/action_model': 1.5803747957462514e-05, 'epoch': 0.29}) +[2026-10-03 19:43:34,978][starVLA.training.train_starvla][INFO] - Step 3928, Loss: {'train/loss': 0.03092067316174507, 'train/grad_norm_pre_clip': 149.41493225097656, 'timing/data': 0.00022401998285204172, 'timing/model': 0.39554368099197745, 'learning_rate/qwen_vl_interface': 1.5784418576828541e-06, 'learning_rate/action_model': 1.5784418576828543e-05, 'epoch': 0.29}) +[2026-10-03 19:43:35,887][starVLA.training.train_starvla][INFO] - Step 3929, Loss: {'train/loss': 0.00012628686090465635, 'train/grad_norm_pre_clip': 0.02880726382136345, 'timing/data': 0.0003511100076138973, 'timing/model': 0.3943526829825714, 'learning_rate/qwen_vl_interface': 1.576510428858322e-06, 'learning_rate/action_model': 1.5765104288583217e-05, 'epoch': 0.29}) +[2026-10-03 19:43:36,599][starVLA.training.train_starvla][INFO] - Step 3930, Loss: {'train/loss': 0.0004278937412891537, 'train/grad_norm_pre_clip': 0.10004337877035141, 'timing/data': 0.000322239997331053, 'timing/model': 0.3982233020069543, 'learning_rate/qwen_vl_interface': 1.5745805100665922e-06, 'learning_rate/action_model': 1.5745805100665923e-05, 'epoch': 0.29}) +[2026-10-03 19:43:37,308][starVLA.training.train_starvla][INFO] - Step 3931, Loss: {'train/loss': 0.19334954023361206, 'train/grad_norm_pre_clip': 99.20905303955078, 'timing/data': 0.0003568899992387742, 'timing/model': 0.39629183200304396, 'learning_rate/qwen_vl_interface': 1.572652102100983e-06, 'learning_rate/action_model': 1.572652102100983e-05, 'epoch': 0.29}) +[2026-10-03 19:43:38,017][starVLA.training.train_starvla][INFO] - Step 3932, Loss: {'train/loss': 0.16256462037563324, 'train/grad_norm_pre_clip': 53.74079132080078, 'timing/data': 0.00023057000362314284, 'timing/model': 0.3967876319948118, 'learning_rate/qwen_vl_interface': 1.5707252057541898e-06, 'learning_rate/action_model': 1.5707252057541896e-05, 'epoch': 0.29}) +[2026-10-03 19:43:38,726][starVLA.training.train_starvla][INFO] - Step 3933, Loss: {'train/loss': 0.06480534374713898, 'train/grad_norm_pre_clip': 253.20054626464844, 'timing/data': 0.00025554001331329346, 'timing/model': 0.396840921981493, 'learning_rate/qwen_vl_interface': 1.5687998218182849e-06, 'learning_rate/action_model': 1.5687998218182847e-05, 'epoch': 0.29}) +[2026-10-03 19:43:39,638][starVLA.training.train_starvla][INFO] - Step 3934, Loss: {'train/loss': 0.0006909083458594978, 'train/grad_norm_pre_clip': 0.41105908155441284, 'timing/data': 0.00036617001751437783, 'timing/model': 0.39732855098554865, 'learning_rate/qwen_vl_interface': 1.5668759510847235e-06, 'learning_rate/action_model': 1.5668759510847233e-05, 'epoch': 0.29}) +[2026-10-03 19:43:40,349][starVLA.training.train_starvla][INFO] - Step 3935, Loss: {'train/loss': 0.045742105692625046, 'train/grad_norm_pre_clip': 97.41496276855469, 'timing/data': 0.00035050002043135464, 'timing/model': 0.3959895720181521, 'learning_rate/qwen_vl_interface': 1.5649535943443331e-06, 'learning_rate/action_model': 1.564953594344333e-05, 'epoch': 0.29}) +[2026-10-03 19:43:41,057][starVLA.training.train_starvla][INFO] - Step 3936, Loss: {'train/loss': 0.17865818738937378, 'train/grad_norm_pre_clip': 61.21501159667969, 'timing/data': 0.00021711000590585172, 'timing/model': 0.3966621319996193, 'learning_rate/qwen_vl_interface': 1.5630327523873244e-06, 'learning_rate/action_model': 1.563032752387324e-05, 'epoch': 0.29}) +[2026-10-03 19:43:41,766][starVLA.training.train_starvla][INFO] - Step 3937, Loss: {'train/loss': 0.002539980225265026, 'train/grad_norm_pre_clip': 3.2301337718963623, 'timing/data': 0.0003426700131967664, 'timing/model': 0.39680895197670907, 'learning_rate/qwen_vl_interface': 1.561113426003282e-06, 'learning_rate/action_model': 1.561113426003282e-05, 'epoch': 0.29}) +[2026-10-03 19:43:42,478][starVLA.training.train_starvla][INFO] - Step 3938, Loss: {'train/loss': 0.024044474586844444, 'train/grad_norm_pre_clip': 49.5072135925293, 'timing/data': 0.00042640999890863895, 'timing/model': 0.39703719201497734, 'learning_rate/qwen_vl_interface': 1.5591956159811686e-06, 'learning_rate/action_model': 1.5591956159811687e-05, 'epoch': 0.29}) +[2026-10-03 19:43:43,385][starVLA.training.train_starvla][INFO] - Step 3939, Loss: {'train/loss': 0.6798179745674133, 'train/grad_norm_pre_clip': 69.86392211914062, 'timing/data': 0.00032341000041924417, 'timing/model': 0.5944822329911403, 'learning_rate/qwen_vl_interface': 1.5572793231093235e-06, 'learning_rate/action_model': 1.5572793231093236e-05, 'epoch': 0.29}) +[2026-10-03 19:43:44,095][starVLA.training.train_starvla][INFO] - Step 3940, Loss: {'train/loss': 0.00017642509192228317, 'train/grad_norm_pre_clip': 0.06653770804405212, 'timing/data': 0.0002268899988848716, 'timing/model': 0.3966549430042505, 'learning_rate/qwen_vl_interface': 1.5553645481754626e-06, 'learning_rate/action_model': 1.5553645481754626e-05, 'epoch': 0.29}) +[2026-10-03 19:43:44,803][starVLA.training.train_starvla][INFO] - Step 3941, Loss: {'train/loss': 0.0003257781208958477, 'train/grad_norm_pre_clip': 0.15158645808696747, 'timing/data': 0.00035774000571109354, 'timing/model': 0.3962249629839789, 'learning_rate/qwen_vl_interface': 1.5534512919666775e-06, 'learning_rate/action_model': 1.5534512919666776e-05, 'epoch': 0.29}) +[2026-10-03 19:43:45,514][starVLA.training.train_starvla][INFO] - Step 3942, Loss: {'train/loss': 0.00023888458963483572, 'train/grad_norm_pre_clip': 0.07452059537172318, 'timing/data': 0.00030109999352134764, 'timing/model': 0.3964745330158621, 'learning_rate/qwen_vl_interface': 1.551539555269435e-06, 'learning_rate/action_model': 1.5515395552694348e-05, 'epoch': 0.29}) +[2026-10-03 19:43:46,228][starVLA.training.train_starvla][INFO] - Step 3943, Loss: {'train/loss': 0.0001690222998149693, 'train/grad_norm_pre_clip': 0.1858380287885666, 'timing/data': 0.0002242390182800591, 'timing/model': 0.39584360300796106, 'learning_rate/qwen_vl_interface': 1.5496293388695787e-06, 'learning_rate/action_model': 1.5496293388695786e-05, 'epoch': 0.29}) +[2026-10-03 19:43:47,129][starVLA.training.train_starvla][INFO] - Step 3944, Loss: {'train/loss': 0.0023716390132904053, 'train/grad_norm_pre_clip': 4.5441789627075195, 'timing/data': 0.00032905899570323527, 'timing/model': 0.5889134439930785, 'learning_rate/qwen_vl_interface': 1.547720643552327e-06, 'learning_rate/action_model': 1.547720643552327e-05, 'epoch': 0.29}) +[2026-10-03 19:43:47,839][starVLA.training.train_starvla][INFO] - Step 3945, Loss: {'train/loss': 0.0005550610949285328, 'train/grad_norm_pre_clip': 0.6315599679946899, 'timing/data': 0.00033531998633407056, 'timing/model': 0.3951700119941961, 'learning_rate/qwen_vl_interface': 1.5458134701022706e-06, 'learning_rate/action_model': 1.5458134701022705e-05, 'epoch': 0.29}) +[2026-10-03 19:43:48,547][starVLA.training.train_starvla][INFO] - Step 3946, Loss: {'train/loss': 0.05218610540032387, 'train/grad_norm_pre_clip': 82.80618286132812, 'timing/data': 0.0002578900021035224, 'timing/model': 0.3962573619792238, 'learning_rate/qwen_vl_interface': 1.5439078193033765e-06, 'learning_rate/action_model': 1.5439078193033766e-05, 'epoch': 0.29}) +[2026-10-03 19:43:49,254][starVLA.training.train_starvla][INFO] - Step 3947, Loss: {'train/loss': 0.036639709025621414, 'train/grad_norm_pre_clip': 98.91468811035156, 'timing/data': 0.0002310900017619133, 'timing/model': 0.39596857197466306, 'learning_rate/qwen_vl_interface': 1.5420036919389872e-06, 'learning_rate/action_model': 1.5420036919389873e-05, 'epoch': 0.29}) +[2026-10-03 19:43:49,964][starVLA.training.train_starvla][INFO] - Step 3948, Loss: {'train/loss': 6.476304406533018e-05, 'train/grad_norm_pre_clip': 0.012125394307076931, 'timing/data': 0.0003463299944996834, 'timing/model': 0.39615691200015135, 'learning_rate/qwen_vl_interface': 1.540101088791817e-06, 'learning_rate/action_model': 1.5401010887918168e-05, 'epoch': 0.29}) +[2026-10-03 19:43:50,673][starVLA.training.train_starvla][INFO] - Step 3949, Loss: {'train/loss': 0.02496459148824215, 'train/grad_norm_pre_clip': 18.391931533813477, 'timing/data': 0.00021612999262288213, 'timing/model': 0.396639761980623, 'learning_rate/qwen_vl_interface': 1.5382000106439532e-06, 'learning_rate/action_model': 1.5382000106439532e-05, 'epoch': 0.29}) +[2026-10-03 19:43:51,580][starVLA.training.train_starvla][INFO] - Step 3950, Loss: {'train/loss': 0.43790915608406067, 'train/grad_norm_pre_clip': 55.694557189941406, 'timing/data': 0.0003396600077394396, 'timing/model': 0.3965564620157238, 'learning_rate/qwen_vl_interface': 1.5363004582768584e-06, 'learning_rate/action_model': 1.5363004582768584e-05, 'epoch': 0.29}) +[2026-10-03 19:43:52,288][starVLA.training.train_starvla][INFO] - Step 3951, Loss: {'train/loss': 0.12344610691070557, 'train/grad_norm_pre_clip': 100.43094635009766, 'timing/data': 0.00034713998320512474, 'timing/model': 0.39471747298375703, 'learning_rate/qwen_vl_interface': 1.534402432471367e-06, 'learning_rate/action_model': 1.534402432471367e-05, 'epoch': 0.29}) +[2026-10-03 19:43:52,996][starVLA.training.train_starvla][INFO] - Step 3952, Loss: {'train/loss': 0.00018150564574170858, 'train/grad_norm_pre_clip': 0.04988572746515274, 'timing/data': 0.00032836999162100255, 'timing/model': 0.39564144299947657, 'learning_rate/qwen_vl_interface': 1.532505934007685e-06, 'learning_rate/action_model': 1.532505934007685e-05, 'epoch': 0.29}) +[2026-10-03 19:43:53,704][starVLA.training.train_starvla][INFO] - Step 3953, Loss: {'train/loss': 0.26893943548202515, 'train/grad_norm_pre_clip': 39.98720932006836, 'timing/data': 0.0003494999837130308, 'timing/model': 0.3960726219811477, 'learning_rate/qwen_vl_interface': 1.5306109636653915e-06, 'learning_rate/action_model': 1.5306109636653914e-05, 'epoch': 0.29}) +[2026-10-03 19:43:54,416][starVLA.training.train_starvla][INFO] - Step 3954, Loss: {'train/loss': 0.02176455222070217, 'train/grad_norm_pre_clip': 28.155139923095703, 'timing/data': 0.0002667700173333287, 'timing/model': 0.39757050201296806, 'learning_rate/qwen_vl_interface': 1.5287175222234383e-06, 'learning_rate/action_model': 1.528717522223438e-05, 'epoch': 0.29}) +[2026-10-03 19:43:55,325][starVLA.training.train_starvla][INFO] - Step 3955, Loss: {'train/loss': 0.00024645341909490526, 'train/grad_norm_pre_clip': 0.07561547309160233, 'timing/data': 0.0002575799881014973, 'timing/model': 0.39633832199615426, 'learning_rate/qwen_vl_interface': 1.5268256104601457e-06, 'learning_rate/action_model': 1.5268256104601454e-05, 'epoch': 0.29}) +[2026-10-03 19:43:56,035][starVLA.training.train_starvla][INFO] - Step 3956, Loss: {'train/loss': 0.00018394700600765646, 'train/grad_norm_pre_clip': 0.05927810072898865, 'timing/data': 0.00036191000253893435, 'timing/model': 0.3965507720131427, 'learning_rate/qwen_vl_interface': 1.5249352291532075e-06, 'learning_rate/action_model': 1.5249352291532073e-05, 'epoch': 0.29}) +[2026-10-03 19:43:56,746][starVLA.training.train_starvla][INFO] - Step 3957, Loss: {'train/loss': 0.000110609908006154, 'train/grad_norm_pre_clip': 0.0254568699747324, 'timing/data': 0.00024709998979233205, 'timing/model': 0.3969908309809398, 'learning_rate/qwen_vl_interface': 1.523046379079689e-06, 'learning_rate/action_model': 1.523046379079689e-05, 'epoch': 0.29}) +[2026-10-03 19:43:57,456][starVLA.training.train_starvla][INFO] - Step 3958, Loss: {'train/loss': 0.0008265565265901387, 'train/grad_norm_pre_clip': 0.7898777723312378, 'timing/data': 0.00032473000464960933, 'timing/model': 0.3958002319850493, 'learning_rate/qwen_vl_interface': 1.5211590610160244e-06, 'learning_rate/action_model': 1.5211590610160242e-05, 'epoch': 0.29}) +[2026-10-03 19:43:58,165][starVLA.training.train_starvla][INFO] - Step 3959, Loss: {'train/loss': 0.002866524737328291, 'train/grad_norm_pre_clip': 1.9584170579910278, 'timing/data': 0.0003384999872650951, 'timing/model': 0.39600972199696116, 'learning_rate/qwen_vl_interface': 1.5192732757380193e-06, 'learning_rate/action_model': 1.5192732757380194e-05, 'epoch': 0.29}) +[2026-10-03 19:43:59,076][starVLA.training.train_starvla][INFO] - Step 3960, Loss: {'train/loss': 0.16234515607357025, 'train/grad_norm_pre_clip': 115.33138275146484, 'timing/data': 0.00031067998497746885, 'timing/model': 0.5981376119889319, 'learning_rate/qwen_vl_interface': 1.517389024020849e-06, 'learning_rate/action_model': 1.5173890240208488e-05, 'epoch': 0.29}) +[2026-10-03 19:43:59,784][starVLA.training.train_starvla][INFO] - Step 3961, Loss: {'train/loss': 0.00038341476465575397, 'train/grad_norm_pre_clip': 0.255104660987854, 'timing/data': 0.00022730999626219273, 'timing/model': 0.39632266099215485, 'learning_rate/qwen_vl_interface': 1.515506306639057e-06, 'learning_rate/action_model': 1.515506306639057e-05, 'epoch': 0.29}) +[2026-10-03 19:44:00,494][starVLA.training.train_starvla][INFO] - Step 3962, Loss: {'train/loss': 0.1946410983800888, 'train/grad_norm_pre_clip': 75.36964416503906, 'timing/data': 0.0003809299960266799, 'timing/model': 0.3960551409982145, 'learning_rate/qwen_vl_interface': 1.5136251243665589e-06, 'learning_rate/action_model': 1.5136251243665589e-05, 'epoch': 0.29}) +[2026-10-03 19:44:01,199][starVLA.training.train_starvla][INFO] - Step 3963, Loss: {'train/loss': 0.017501721158623695, 'train/grad_norm_pre_clip': 57.75340270996094, 'timing/data': 0.0003399199922569096, 'timing/model': 0.3941882030048873, 'learning_rate/qwen_vl_interface': 1.5117454779766372e-06, 'learning_rate/action_model': 1.5117454779766371e-05, 'epoch': 0.29}) +[2026-10-03 19:44:01,906][starVLA.training.train_starvla][INFO] - Step 3964, Loss: {'train/loss': 0.007521914783865213, 'train/grad_norm_pre_clip': 27.711565017700195, 'timing/data': 0.00031786999898031354, 'timing/model': 0.395484202017542, 'learning_rate/qwen_vl_interface': 1.5098673682419428e-06, 'learning_rate/action_model': 1.5098673682419429e-05, 'epoch': 0.29}) +[2026-10-03 19:44:02,810][starVLA.training.train_starvla][INFO] - Step 3965, Loss: {'train/loss': 0.33909735083580017, 'train/grad_norm_pre_clip': 168.04551696777344, 'timing/data': 0.0003525399952195585, 'timing/model': 0.5947498829918914, 'learning_rate/qwen_vl_interface': 1.5079907959344977e-06, 'learning_rate/action_model': 1.5079907959344978e-05, 'epoch': 0.29}) +[2026-10-03 19:44:03,517][starVLA.training.train_starvla][INFO] - Step 3966, Loss: {'train/loss': 0.0012722169049084187, 'train/grad_norm_pre_clip': 0.726632297039032, 'timing/data': 0.0002914999786298722, 'timing/model': 0.3940628920099698, 'learning_rate/qwen_vl_interface': 1.5061157618256873e-06, 'learning_rate/action_model': 1.5061157618256871e-05, 'epoch': 0.29}) +[2026-10-03 19:44:04,222][starVLA.training.train_starvla][INFO] - Step 3967, Loss: {'train/loss': 0.30926746129989624, 'train/grad_norm_pre_clip': 100.9386215209961, 'timing/data': 0.00036025000736117363, 'timing/model': 0.3931430329976138, 'learning_rate/qwen_vl_interface': 1.5042422666862675e-06, 'learning_rate/action_model': 1.5042422666862673e-05, 'epoch': 0.29}) +[2026-10-03 19:44:04,928][starVLA.training.train_starvla][INFO] - Step 3968, Loss: {'train/loss': 0.0002631777024362236, 'train/grad_norm_pre_clip': 0.09632555395364761, 'timing/data': 0.0003268200089223683, 'timing/model': 0.39458630201988854, 'learning_rate/qwen_vl_interface': 1.5023703112863628e-06, 'learning_rate/action_model': 1.5023703112863626e-05, 'epoch': 0.29}) +[2026-10-03 19:44:05,632][starVLA.training.train_starvla][INFO] - Step 3969, Loss: {'train/loss': 0.001136837643571198, 'train/grad_norm_pre_clip': 1.0022096633911133, 'timing/data': 0.00036763999378308654, 'timing/model': 0.39461139199556783, 'learning_rate/qwen_vl_interface': 1.5004998963954629e-06, 'learning_rate/action_model': 1.5004998963954628e-05, 'epoch': 0.29}) +[2026-10-03 19:44:06,340][starVLA.training.train_starvla][INFO] - Step 3970, Loss: {'train/loss': 8.053278725128621e-05, 'train/grad_norm_pre_clip': 0.02847730740904808, 'timing/data': 0.00026616998366080225, 'timing/model': 0.3955888530181255, 'learning_rate/qwen_vl_interface': 1.498631022782424e-06, 'learning_rate/action_model': 1.4986310227824241e-05, 'epoch': 0.29}) +[2026-10-03 19:44:07,240][starVLA.training.train_starvla][INFO] - Step 3971, Loss: {'train/loss': 7.228091999422759e-05, 'train/grad_norm_pre_clip': 0.015677858144044876, 'timing/data': 0.00028916000155732036, 'timing/model': 0.39461973201832734, 'learning_rate/qwen_vl_interface': 1.4967636912154707e-06, 'learning_rate/action_model': 1.4967636912154706e-05, 'epoch': 0.29}) +[2026-10-03 19:44:07,944][starVLA.training.train_starvla][INFO] - Step 3972, Loss: {'train/loss': 0.00043883558828383684, 'train/grad_norm_pre_clip': 0.15837542712688446, 'timing/data': 0.0002399600052740425, 'timing/model': 0.39451751302112825, 'learning_rate/qwen_vl_interface': 1.4948979024621915e-06, 'learning_rate/action_model': 1.4948979024621914e-05, 'epoch': 0.29}) +[2026-10-03 19:44:08,651][starVLA.training.train_starvla][INFO] - Step 3973, Loss: {'train/loss': 9.867527114693075e-05, 'train/grad_norm_pre_clip': 0.07073713839054108, 'timing/data': 0.0003761799889616668, 'timing/model': 0.3946406130271498, 'learning_rate/qwen_vl_interface': 1.493033657289542e-06, 'learning_rate/action_model': 1.493033657289542e-05, 'epoch': 0.29}) +[2026-10-03 19:44:09,357][starVLA.training.train_starvla][INFO] - Step 3974, Loss: {'train/loss': 0.007329039741307497, 'train/grad_norm_pre_clip': 11.285405158996582, 'timing/data': 0.0002760599891189486, 'timing/model': 0.39446035100263543, 'learning_rate/qwen_vl_interface': 1.4911709564638428e-06, 'learning_rate/action_model': 1.4911709564638427e-05, 'epoch': 0.29}) +[2026-10-03 19:44:10,063][starVLA.training.train_starvla][INFO] - Step 3975, Loss: {'train/loss': 0.3504977226257324, 'train/grad_norm_pre_clip': 94.15660858154297, 'timing/data': 0.0003072200051974505, 'timing/model': 0.3940789019980002, 'learning_rate/qwen_vl_interface': 1.4893098007507817e-06, 'learning_rate/action_model': 1.4893098007507816e-05, 'epoch': 0.29}) +[2026-10-03 19:44:10,966][starVLA.training.train_starvla][INFO] - Step 3976, Loss: {'train/loss': 0.10098619014024734, 'train/grad_norm_pre_clip': 213.62164306640625, 'timing/data': 0.00028775999089702964, 'timing/model': 0.39355243201134726, 'learning_rate/qwen_vl_interface': 1.487450190915406e-06, 'learning_rate/action_model': 1.487450190915406e-05, 'epoch': 0.29}) +[2026-10-03 19:44:11,672][starVLA.training.train_starvla][INFO] - Step 3977, Loss: {'train/loss': 0.3114554286003113, 'train/grad_norm_pre_clip': 105.76495361328125, 'timing/data': 0.00034766001044772565, 'timing/model': 0.39458222201210447, 'learning_rate/qwen_vl_interface': 1.4855921277221336e-06, 'learning_rate/action_model': 1.4855921277221336e-05, 'epoch': 0.29}) +[2026-10-03 19:44:12,381][starVLA.training.train_starvla][INFO] - Step 3978, Loss: {'train/loss': 0.36985400319099426, 'train/grad_norm_pre_clip': 57.4793586730957, 'timing/data': 0.0003251599846407771, 'timing/model': 0.39563691199873574, 'learning_rate/qwen_vl_interface': 1.4837356119347423e-06, 'learning_rate/action_model': 1.4837356119347423e-05, 'epoch': 0.29}) +[2026-10-03 19:44:13,088][starVLA.training.train_starvla][INFO] - Step 3979, Loss: {'train/loss': 0.0005490302573889494, 'train/grad_norm_pre_clip': 1.12098228931427, 'timing/data': 0.00036985997576266527, 'timing/model': 0.394001892011147, 'learning_rate/qwen_vl_interface': 1.4818806443163774e-06, 'learning_rate/action_model': 1.4818806443163775e-05, 'epoch': 0.29}) +[2026-10-03 19:44:13,794][starVLA.training.train_starvla][INFO] - Step 3980, Loss: {'train/loss': 9.455691179027781e-05, 'train/grad_norm_pre_clip': 0.02991299331188202, 'timing/data': 0.00034018998849205673, 'timing/model': 0.3955318419903051, 'learning_rate/qwen_vl_interface': 1.480027225629547e-06, 'learning_rate/action_model': 1.480027225629547e-05, 'epoch': 0.29}) +[2026-10-03 19:44:14,706][starVLA.training.train_starvla][INFO] - Step 3981, Loss: {'train/loss': 0.00011587180779315531, 'train/grad_norm_pre_clip': 0.05338835343718529, 'timing/data': 0.00037016000715084374, 'timing/model': 0.5985195430112071, 'learning_rate/qwen_vl_interface': 1.4781753566361182e-06, 'learning_rate/action_model': 1.4781753566361182e-05, 'epoch': 0.29}) +[2026-10-03 19:44:15,415][starVLA.training.train_starvla][INFO] - Step 3982, Loss: {'train/loss': 5.355848406907171e-05, 'train/grad_norm_pre_clip': 0.013659879565238953, 'timing/data': 0.00030490997596643865, 'timing/model': 0.39792927098460495, 'learning_rate/qwen_vl_interface': 1.4763250380973255e-06, 'learning_rate/action_model': 1.4763250380973254e-05, 'epoch': 0.29}) +[2026-10-03 19:44:16,121][starVLA.training.train_starvla][INFO] - Step 3983, Loss: {'train/loss': 0.003455079160630703, 'train/grad_norm_pre_clip': 3.7955589294433594, 'timing/data': 0.0003327899903524667, 'timing/model': 0.3939052509958856, 'learning_rate/qwen_vl_interface': 1.4744762707737662e-06, 'learning_rate/action_model': 1.4744762707737662e-05, 'epoch': 0.29}) +[2026-10-03 19:44:16,826][starVLA.training.train_starvla][INFO] - Step 3984, Loss: {'train/loss': 0.00019440960022620857, 'train/grad_norm_pre_clip': 0.14054204523563385, 'timing/data': 0.0002225700009148568, 'timing/model': 0.39482921300805174, 'learning_rate/qwen_vl_interface': 1.4726290554253977e-06, 'learning_rate/action_model': 1.4726290554253977e-05, 'epoch': 0.29}) +[2026-10-03 19:44:17,529][starVLA.training.train_starvla][INFO] - Step 3985, Loss: {'train/loss': 0.0022260197438299656, 'train/grad_norm_pre_clip': 3.1405770778656006, 'timing/data': 0.00035701002343557775, 'timing/model': 0.392463422002038, 'learning_rate/qwen_vl_interface': 1.4707833928115406e-06, 'learning_rate/action_model': 1.4707833928115405e-05, 'epoch': 0.29}) +[2026-10-03 19:44:18,438][starVLA.training.train_starvla][INFO] - Step 3986, Loss: {'train/loss': 0.00013309167115949094, 'train/grad_norm_pre_clip': 0.05543803796172142, 'timing/data': 0.0002163600001949817, 'timing/model': 0.6002709919994231, 'learning_rate/qwen_vl_interface': 1.4689392836908775e-06, 'learning_rate/action_model': 1.4689392836908774e-05, 'epoch': 0.29}) +[2026-10-03 19:44:19,143][starVLA.training.train_starvla][INFO] - Step 3987, Loss: {'train/loss': 0.012939882464706898, 'train/grad_norm_pre_clip': 40.775428771972656, 'timing/data': 0.0002503400028217584, 'timing/model': 0.39427897200221196, 'learning_rate/qwen_vl_interface': 1.4670967288214514e-06, 'learning_rate/action_model': 1.4670967288214513e-05, 'epoch': 0.29}) +[2026-10-03 19:44:19,847][starVLA.training.train_starvla][INFO] - Step 3988, Loss: {'train/loss': 0.0001393976854160428, 'train/grad_norm_pre_clip': 0.04316622018814087, 'timing/data': 0.0003519900201354176, 'timing/model': 0.3943159719929099, 'learning_rate/qwen_vl_interface': 1.4652557289606687e-06, 'learning_rate/action_model': 1.4652557289606685e-05, 'epoch': 0.29}) +[2026-10-03 19:44:20,554][starVLA.training.train_starvla][INFO] - Step 3989, Loss: {'train/loss': 0.005056845024228096, 'train/grad_norm_pre_clip': 4.619510650634766, 'timing/data': 0.000321680010529235, 'timing/model': 0.39641028200276196, 'learning_rate/qwen_vl_interface': 1.4634162848652918e-06, 'learning_rate/action_model': 1.4634162848652919e-05, 'epoch': 0.29}) +[2026-10-03 19:44:21,257][starVLA.training.train_starvla][INFO] - Step 3990, Loss: {'train/loss': 4.927935515297577e-05, 'train/grad_norm_pre_clip': 0.02323049306869507, 'timing/data': 0.0002730699779931456, 'timing/model': 0.39316284298547544, 'learning_rate/qwen_vl_interface': 1.4615783972914494e-06, 'learning_rate/action_model': 1.4615783972914493e-05, 'epoch': 0.29}) +[2026-10-03 19:44:21,963][starVLA.training.train_starvla][INFO] - Step 3991, Loss: {'train/loss': 0.0013765997719019651, 'train/grad_norm_pre_clip': 1.090011477470398, 'timing/data': 0.00034686000435613096, 'timing/model': 0.3941508720163256, 'learning_rate/qwen_vl_interface': 1.459742066994626e-06, 'learning_rate/action_model': 1.4597420669946261e-05, 'epoch': 0.29}) +[2026-10-03 19:44:22,870][starVLA.training.train_starvla][INFO] - Step 3992, Loss: {'train/loss': 0.0015006324974820018, 'train/grad_norm_pre_clip': 1.2811429500579834, 'timing/data': 0.0003703399852383882, 'timing/model': 0.39279508200706914, 'learning_rate/qwen_vl_interface': 1.4579072947296687e-06, 'learning_rate/action_model': 1.4579072947296687e-05, 'epoch': 0.29}) +[2026-10-03 19:44:23,574][starVLA.training.train_starvla][INFO] - Step 3993, Loss: {'train/loss': 5.313371366355568e-05, 'train/grad_norm_pre_clip': 0.017592694610357285, 'timing/data': 0.0003567200037650764, 'timing/model': 0.39296019199537113, 'learning_rate/qwen_vl_interface': 1.4560740812507823e-06, 'learning_rate/action_model': 1.4560740812507822e-05, 'epoch': 0.29}) +[2026-10-03 19:44:24,280][starVLA.training.train_starvla][INFO] - Step 3994, Loss: {'train/loss': 5.052302367403172e-05, 'train/grad_norm_pre_clip': 0.02289518155157566, 'timing/data': 0.00029664000612683594, 'timing/model': 0.3950167720031459, 'learning_rate/qwen_vl_interface': 1.4542424273115322e-06, 'learning_rate/action_model': 1.454242427311532e-05, 'epoch': 0.29}) +[2026-10-03 19:44:24,985][starVLA.training.train_starvla][INFO] - Step 3995, Loss: {'train/loss': 0.001616372261196375, 'train/grad_norm_pre_clip': 3.428898334503174, 'timing/data': 0.00022036998416297138, 'timing/model': 0.39427439199062064, 'learning_rate/qwen_vl_interface': 1.4524123336648419e-06, 'learning_rate/action_model': 1.4524123336648418e-05, 'epoch': 0.29}) +[2026-10-03 19:44:25,690][starVLA.training.train_starvla][INFO] - Step 3996, Loss: {'train/loss': 0.00043499242747202516, 'train/grad_norm_pre_clip': 0.5085089802742004, 'timing/data': 0.0002574689860921353, 'timing/model': 0.3947343720064964, 'learning_rate/qwen_vl_interface': 1.4505838010629942e-06, 'learning_rate/action_model': 1.4505838010629941e-05, 'epoch': 0.29}) +[2026-10-03 19:44:26,593][starVLA.training.train_starvla][INFO] - Step 3997, Loss: {'train/loss': 0.5462943911552429, 'train/grad_norm_pre_clip': 52.419044494628906, 'timing/data': 0.00023018999490886927, 'timing/model': 0.5934134529961739, 'learning_rate/qwen_vl_interface': 1.4487568302576284e-06, 'learning_rate/action_model': 1.4487568302576282e-05, 'epoch': 0.29}) +[2026-10-03 19:44:27,300][starVLA.training.train_starvla][INFO] - Step 3998, Loss: {'train/loss': 0.020909762009978294, 'train/grad_norm_pre_clip': 42.58430862426758, 'timing/data': 0.00026557999080978334, 'timing/model': 0.3956690030172467, 'learning_rate/qwen_vl_interface': 1.446931421999746e-06, 'learning_rate/action_model': 1.4469314219997459e-05, 'epoch': 0.29}) +[2026-10-03 19:44:28,005][starVLA.training.train_starvla][INFO] - Step 3999, Loss: {'train/loss': 0.49985024333000183, 'train/grad_norm_pre_clip': 57.33533477783203, 'timing/data': 0.0002181899908464402, 'timing/model': 0.39392781199421734, 'learning_rate/qwen_vl_interface': 1.4451075770397004e-06, 'learning_rate/action_model': 1.4451075770397004e-05, 'epoch': 0.29}) +[2026-10-03 19:44:28,709][starVLA.training.train_starvla][INFO] - Step 4000, Loss: {'train/loss': 3.844898674287833e-05, 'train/grad_norm_pre_clip': 0.0156747754663229, 'timing/data': 0.00038492001476697624, 'timing/model': 0.394302692002384, 'learning_rate/qwen_vl_interface': 1.4432852961272068e-06, 'learning_rate/action_model': 1.4432852961272066e-05, 'epoch': 0.29}) +[2026-10-03 19:44:29,415][starVLA.training.train_starvla][INFO] - Step 4001, Loss: {'train/loss': 0.0015979285817593336, 'train/grad_norm_pre_clip': 2.6538803577423096, 'timing/data': 0.0002980800054501742, 'timing/model': 0.39424854202661663, 'learning_rate/qwen_vl_interface': 1.4414645800113353e-06, 'learning_rate/action_model': 1.4414645800113352e-05, 'epoch': 0.29}) +[2026-10-03 19:44:30,319][starVLA.training.train_starvla][INFO] - Step 4002, Loss: {'train/loss': 0.3594333827495575, 'train/grad_norm_pre_clip': 86.82691192626953, 'timing/data': 0.00021687999833375216, 'timing/model': 0.5947003629989922, 'learning_rate/qwen_vl_interface': 1.4396454294405138e-06, 'learning_rate/action_model': 1.4396454294405136e-05, 'epoch': 0.29}) +[2026-10-03 19:44:31,023][starVLA.training.train_starvla][INFO] - Step 4003, Loss: {'train/loss': 1.567578874528408e-05, 'train/grad_norm_pre_clip': 0.003567151026800275, 'timing/data': 0.00036381001700647175, 'timing/model': 0.3925760420097504, 'learning_rate/qwen_vl_interface': 1.4378278451625271e-06, 'learning_rate/action_model': 1.4378278451625271e-05, 'epoch': 0.29}) +[2026-10-03 19:44:31,728][starVLA.training.train_starvla][INFO] - Step 4004, Loss: {'train/loss': 0.45951154828071594, 'train/grad_norm_pre_clip': 53.12871170043945, 'timing/data': 0.00030533000244759023, 'timing/model': 0.39476413099328056, 'learning_rate/qwen_vl_interface': 1.4360118279245143e-06, 'learning_rate/action_model': 1.4360118279245144e-05, 'epoch': 0.29}) +[2026-10-03 19:44:32,433][starVLA.training.train_starvla][INFO] - Step 4005, Loss: {'train/loss': 0.0002882186381611973, 'train/grad_norm_pre_clip': 1.0902458429336548, 'timing/data': 0.000364459992852062, 'timing/model': 0.3937586030224338, 'learning_rate/qwen_vl_interface': 1.434197378472973e-06, 'learning_rate/action_model': 1.4341973784729729e-05, 'epoch': 0.29}) +[2026-10-03 19:44:33,139][starVLA.training.train_starvla][INFO] - Step 4006, Loss: {'train/loss': 0.00017809598648454994, 'train/grad_norm_pre_clip': 0.09740498661994934, 'timing/data': 0.00028167999698780477, 'timing/model': 0.39539169101044536, 'learning_rate/qwen_vl_interface': 1.4323844975537535e-06, 'learning_rate/action_model': 1.4323844975537534e-05, 'epoch': 0.29}) +[2026-10-03 19:44:33,844][starVLA.training.train_starvla][INFO] - Step 4007, Loss: {'train/loss': 0.00011472689220681787, 'train/grad_norm_pre_clip': 0.05254325270652771, 'timing/data': 0.0002670499961823225, 'timing/model': 0.39247672201599926, 'learning_rate/qwen_vl_interface': 1.4305731859120628e-06, 'learning_rate/action_model': 1.4305731859120628e-05, 'epoch': 0.29}) +[2026-10-03 19:44:34,748][starVLA.training.train_starvla][INFO] - Step 4008, Loss: {'train/loss': 6.986810331000015e-05, 'train/grad_norm_pre_clip': 0.07657401263713837, 'timing/data': 0.00025365001056343317, 'timing/model': 0.39267599198501557, 'learning_rate/qwen_vl_interface': 1.4287634442924647e-06, 'learning_rate/action_model': 1.4287634442924646e-05, 'epoch': 0.29}) +[2026-10-03 19:44:35,452][starVLA.training.train_starvla][INFO] - Step 4009, Loss: {'train/loss': 0.0001094800463761203, 'train/grad_norm_pre_clip': 0.042731933295726776, 'timing/data': 0.0003744999994523823, 'timing/model': 0.3928305820154492, 'learning_rate/qwen_vl_interface': 1.4269552734388725e-06, 'learning_rate/action_model': 1.4269552734388724e-05, 'epoch': 0.29}) +[2026-10-03 19:44:36,158][starVLA.training.train_starvla][INFO] - Step 4010, Loss: {'train/loss': 9.559328464092687e-05, 'train/grad_norm_pre_clip': 0.04849642142653465, 'timing/data': 0.00029580999398604035, 'timing/model': 0.39393489199574105, 'learning_rate/qwen_vl_interface': 1.4251486740945584e-06, 'learning_rate/action_model': 1.4251486740945583e-05, 'epoch': 0.29}) +[2026-10-03 19:44:36,865][starVLA.training.train_starvla][INFO] - Step 4011, Loss: {'train/loss': 3.0278097256086767e-05, 'train/grad_norm_pre_clip': 0.007752405479550362, 'timing/data': 0.0002281600027345121, 'timing/model': 0.39494289201684296, 'learning_rate/qwen_vl_interface': 1.423343647002147e-06, 'learning_rate/action_model': 1.423343647002147e-05, 'epoch': 0.29}) +[2026-10-03 19:44:37,568][starVLA.training.train_starvla][INFO] - Step 4012, Loss: {'train/loss': 0.0015037371776998043, 'train/grad_norm_pre_clip': 1.201311707496643, 'timing/data': 0.00035563998972065747, 'timing/model': 0.39419824301148765, 'learning_rate/qwen_vl_interface': 1.421540192903617e-06, 'learning_rate/action_model': 1.421540192903617e-05, 'epoch': 0.29}) +[2026-10-03 19:44:38,473][starVLA.training.train_starvla][INFO] - Step 4013, Loss: {'train/loss': 2.831128949765116e-05, 'train/grad_norm_pre_clip': 0.00605803495272994, 'timing/data': 0.0003816699900198728, 'timing/model': 0.3947008620016277, 'learning_rate/qwen_vl_interface': 1.4197383125402997e-06, 'learning_rate/action_model': 1.4197383125402997e-05, 'epoch': 0.29}) +[2026-10-03 19:44:39,178][starVLA.training.train_starvla][INFO] - Step 4014, Loss: {'train/loss': 0.0034958466421812773, 'train/grad_norm_pre_clip': 5.405916690826416, 'timing/data': 0.00027084999601356685, 'timing/model': 0.3933456710074097, 'learning_rate/qwen_vl_interface': 1.4179380066528807e-06, 'learning_rate/action_model': 1.4179380066528807e-05, 'epoch': 0.29}) +[2026-10-03 19:44:39,884][starVLA.training.train_starvla][INFO] - Step 4015, Loss: {'train/loss': 0.0008795904577709734, 'train/grad_norm_pre_clip': 1.125253677368164, 'timing/data': 0.0003512999974191189, 'timing/model': 0.39382787299109623, 'learning_rate/qwen_vl_interface': 1.4161392759813984e-06, 'learning_rate/action_model': 1.4161392759813983e-05, 'epoch': 0.29}) +[2026-10-03 19:44:40,590][starVLA.training.train_starvla][INFO] - Step 4016, Loss: {'train/loss': 0.1120888739824295, 'train/grad_norm_pre_clip': 125.38119506835938, 'timing/data': 0.0003358999965712428, 'timing/model': 0.3949877110135276, 'learning_rate/qwen_vl_interface': 1.4143421212652411e-06, 'learning_rate/action_model': 1.414342121265241e-05, 'epoch': 0.29}) +[2026-10-03 19:44:41,292][starVLA.training.train_starvla][INFO] - Step 4017, Loss: {'train/loss': 0.5625101327896118, 'train/grad_norm_pre_clip': 87.87935638427734, 'timing/data': 0.0003349999897181988, 'timing/model': 0.3933118920249399, 'learning_rate/qwen_vl_interface': 1.4125465432431529e-06, 'learning_rate/action_model': 1.4125465432431529e-05, 'epoch': 0.29}) +[2026-10-03 19:44:42,204][starVLA.training.train_starvla][INFO] - Step 4018, Loss: {'train/loss': 0.00012009304919047281, 'train/grad_norm_pre_clip': 0.11462363600730896, 'timing/data': 0.0003042900061700493, 'timing/model': 0.598662772012176, 'learning_rate/qwen_vl_interface': 1.4107525426532273e-06, 'learning_rate/action_model': 1.4107525426532272e-05, 'epoch': 0.29}) +[2026-10-03 19:44:42,909][starVLA.training.train_starvla][INFO] - Step 4019, Loss: {'train/loss': 0.00015231814177241176, 'train/grad_norm_pre_clip': 0.09030860662460327, 'timing/data': 0.00029352999990805984, 'timing/model': 0.3949503830226604, 'learning_rate/qwen_vl_interface': 1.4089601202329116e-06, 'learning_rate/action_model': 1.4089601202329114e-05, 'epoch': 0.29}) +[2026-10-03 19:44:43,614][starVLA.training.train_starvla][INFO] - Step 4020, Loss: {'train/loss': 0.00016036542365327477, 'train/grad_norm_pre_clip': 0.057735808193683624, 'timing/data': 0.00037473998963832855, 'timing/model': 0.39362290099961683, 'learning_rate/qwen_vl_interface': 1.407169276719e-06, 'learning_rate/action_model': 1.4071692767190001e-05, 'epoch': 0.29}) +[2026-10-03 19:44:44,321][starVLA.training.train_starvla][INFO] - Step 4021, Loss: {'train/loss': 0.00018572145199868828, 'train/grad_norm_pre_clip': 0.15617339313030243, 'timing/data': 0.0003100000030826777, 'timing/model': 0.39531006300239824, 'learning_rate/qwen_vl_interface': 1.4053800128476426e-06, 'learning_rate/action_model': 1.4053800128476427e-05, 'epoch': 0.29}) +[2026-10-03 19:44:45,028][starVLA.training.train_starvla][INFO] - Step 4022, Loss: {'train/loss': 0.0008496647933498025, 'train/grad_norm_pre_clip': 1.1014618873596191, 'timing/data': 0.0002950599882751703, 'timing/model': 0.39436946299974807, 'learning_rate/qwen_vl_interface': 1.403592329354338e-06, 'learning_rate/action_model': 1.403592329354338e-05, 'epoch': 0.29}) +[2026-10-03 19:44:45,933][starVLA.training.train_starvla][INFO] - Step 4023, Loss: {'train/loss': 0.00520137045532465, 'train/grad_norm_pre_clip': 8.503584861755371, 'timing/data': 0.0003391499922145158, 'timing/model': 0.593311402015388, 'learning_rate/qwen_vl_interface': 1.4018062269739354e-06, 'learning_rate/action_model': 1.4018062269739355e-05, 'epoch': 0.29}) +[2026-10-03 19:44:46,640][starVLA.training.train_starvla][INFO] - Step 4024, Loss: {'train/loss': 0.00011893865303136408, 'train/grad_norm_pre_clip': 0.04961690679192543, 'timing/data': 0.00028577999910339713, 'timing/model': 0.3953511410218198, 'learning_rate/qwen_vl_interface': 1.4000217064406339e-06, 'learning_rate/action_model': 1.4000217064406337e-05, 'epoch': 0.29}) +[2026-10-03 19:44:47,345][starVLA.training.train_starvla][INFO] - Step 4025, Loss: {'train/loss': 0.006630103103816509, 'train/grad_norm_pre_clip': 31.048870086669922, 'timing/data': 0.00035594997461885214, 'timing/model': 0.3944869329861831, 'learning_rate/qwen_vl_interface': 1.3982387684879836e-06, 'learning_rate/action_model': 1.3982387684879836e-05, 'epoch': 0.29}) +[2026-10-03 19:44:48,052][starVLA.training.train_starvla][INFO] - Step 4026, Loss: {'train/loss': 0.00013159892114344984, 'train/grad_norm_pre_clip': 0.034350842237472534, 'timing/data': 0.0003042600001208484, 'timing/model': 0.3956564209947828, 'learning_rate/qwen_vl_interface': 1.3964574138488804e-06, 'learning_rate/action_model': 1.3964574138488803e-05, 'epoch': 0.29}) +[2026-10-03 19:44:48,760][starVLA.training.train_starvla][INFO] - Step 4027, Loss: {'train/loss': 0.5633115768432617, 'train/grad_norm_pre_clip': 71.35003662109375, 'timing/data': 0.0004894699959550053, 'timing/model': 0.3950457919854671, 'learning_rate/qwen_vl_interface': 1.3946776432555743e-06, 'learning_rate/action_model': 1.3946776432555741e-05, 'epoch': 0.29}) +[2026-10-03 19:44:49,674][starVLA.training.train_starvla][INFO] - Step 4028, Loss: {'train/loss': 0.3097992539405823, 'train/grad_norm_pre_clip': 68.20576477050781, 'timing/data': 0.0005030000174883753, 'timing/model': 0.602198161999695, 'learning_rate/qwen_vl_interface': 1.3928994574396595e-06, 'learning_rate/action_model': 1.3928994574396594e-05, 'epoch': 0.29}) +[2026-10-03 19:44:50,379][starVLA.training.train_starvla][INFO] - Step 4029, Loss: {'train/loss': 0.025038916617631912, 'train/grad_norm_pre_clip': 71.8936996459961, 'timing/data': 0.0002601899905130267, 'timing/model': 0.3942974029923789, 'learning_rate/qwen_vl_interface': 1.391122857132083e-06, 'learning_rate/action_model': 1.3911228571320831e-05, 'epoch': 0.29}) +[2026-10-03 19:44:51,084][starVLA.training.train_starvla][INFO] - Step 4030, Loss: {'train/loss': 0.0002142237062798813, 'train/grad_norm_pre_clip': 0.1553862988948822, 'timing/data': 0.00024607000523246825, 'timing/model': 0.3952022809826303, 'learning_rate/qwen_vl_interface': 1.3893478430631376e-06, 'learning_rate/action_model': 1.3893478430631376e-05, 'epoch': 0.29}) +[2026-10-03 19:44:51,790][starVLA.training.train_starvla][INFO] - Step 4031, Loss: {'train/loss': 0.000270368269411847, 'train/grad_norm_pre_clip': 0.47985753417015076, 'timing/data': 0.00029356000595726073, 'timing/model': 0.39398503201664425, 'learning_rate/qwen_vl_interface': 1.3875744159624652e-06, 'learning_rate/action_model': 1.3875744159624652e-05, 'epoch': 0.3}) +[2026-10-03 19:44:52,492][starVLA.training.train_starvla][INFO] - Step 4032, Loss: {'train/loss': 7.517078483942896e-05, 'train/grad_norm_pre_clip': 0.022790387272834778, 'timing/data': 0.0003244499966967851, 'timing/model': 0.39230325201060623, 'learning_rate/qwen_vl_interface': 1.3858025765590525e-06, 'learning_rate/action_model': 1.3858025765590525e-05, 'epoch': 0.3}) +[2026-10-03 19:44:53,197][starVLA.training.train_starvla][INFO] - Step 4033, Loss: {'train/loss': 0.00023305992363020778, 'train/grad_norm_pre_clip': 0.11053918302059174, 'timing/data': 0.00021893999655731022, 'timing/model': 0.3938239220005926, 'learning_rate/qwen_vl_interface': 1.384032325581237e-06, 'learning_rate/action_model': 1.3840323255812371e-05, 'epoch': 0.3}) +[2026-10-03 19:44:54,106][starVLA.training.train_starvla][INFO] - Step 4034, Loss: {'train/loss': 0.46778926253318787, 'train/grad_norm_pre_clip': 73.38433837890625, 'timing/data': 0.00028793999808840454, 'timing/model': 0.39557982300175354, 'learning_rate/qwen_vl_interface': 1.3822636637567025e-06, 'learning_rate/action_model': 1.3822636637567024e-05, 'epoch': 0.3}) +[2026-10-03 19:44:54,811][starVLA.training.train_starvla][INFO] - Step 4035, Loss: {'train/loss': 3.806963286479004e-05, 'train/grad_norm_pre_clip': 0.00852199736982584, 'timing/data': 0.0003373900253791362, 'timing/model': 0.3944539520016406, 'learning_rate/qwen_vl_interface': 1.3804965918124788e-06, 'learning_rate/action_model': 1.3804965918124787e-05, 'epoch': 0.3}) +[2026-10-03 19:44:55,515][starVLA.training.train_starvla][INFO] - Step 4036, Loss: {'train/loss': 0.026232009753584862, 'train/grad_norm_pre_clip': 62.93022155761719, 'timing/data': 0.00029756000731140375, 'timing/model': 0.39299501199275255, 'learning_rate/qwen_vl_interface': 1.3787311104749418e-06, 'learning_rate/action_model': 1.3787311104749418e-05, 'epoch': 0.3}) +[2026-10-03 19:44:56,219][starVLA.training.train_starvla][INFO] - Step 4037, Loss: {'train/loss': 0.15518608689308167, 'train/grad_norm_pre_clip': 274.2632751464844, 'timing/data': 0.0003465699846856296, 'timing/model': 0.39308464198256843, 'learning_rate/qwen_vl_interface': 1.3769672204698149e-06, 'learning_rate/action_model': 1.3769672204698147e-05, 'epoch': 0.3}) +[2026-10-03 19:44:56,925][starVLA.training.train_starvla][INFO] - Step 4038, Loss: {'train/loss': 0.4481825530529022, 'train/grad_norm_pre_clip': 108.28294372558594, 'timing/data': 0.000245189992710948, 'timing/model': 0.395456792000914, 'learning_rate/qwen_vl_interface': 1.3752049225221664e-06, 'learning_rate/action_model': 1.3752049225221663e-05, 'epoch': 0.3}) +[2026-10-03 19:44:57,831][starVLA.training.train_starvla][INFO] - Step 4039, Loss: {'train/loss': 0.544274091720581, 'train/grad_norm_pre_clip': 121.02765655517578, 'timing/data': 0.0002346700057387352, 'timing/model': 0.5951461929944344, 'learning_rate/qwen_vl_interface': 1.3734442173564103e-06, 'learning_rate/action_model': 1.3734442173564102e-05, 'epoch': 0.3}) +[2026-10-03 19:44:58,536][starVLA.training.train_starvla][INFO] - Step 4040, Loss: {'train/loss': 2.3200336727313697e-05, 'train/grad_norm_pre_clip': 0.005277788266539574, 'timing/data': 0.00028630997985601425, 'timing/model': 0.3941621219855733, 'learning_rate/qwen_vl_interface': 1.3716851056963062e-06, 'learning_rate/action_model': 1.3716851056963062e-05, 'epoch': 0.3}) +[2026-10-03 19:44:59,243][starVLA.training.train_starvla][INFO] - Step 4041, Loss: {'train/loss': 0.0033929862547665834, 'train/grad_norm_pre_clip': 7.146058559417725, 'timing/data': 0.00039053999353200197, 'timing/model': 0.39361745200585574, 'learning_rate/qwen_vl_interface': 1.3699275882649597e-06, 'learning_rate/action_model': 1.3699275882649596e-05, 'epoch': 0.3}) +[2026-10-03 19:44:59,950][starVLA.training.train_starvla][INFO] - Step 4042, Loss: {'train/loss': 0.07717939466238022, 'train/grad_norm_pre_clip': 122.40568542480469, 'timing/data': 0.00032034001196734607, 'timing/model': 0.39630544200190343, 'learning_rate/qwen_vl_interface': 1.368171665784817e-06, 'learning_rate/action_model': 1.3681716657848171e-05, 'epoch': 0.3}) +[2026-10-03 19:45:00,655][starVLA.training.train_starvla][INFO] - Step 4043, Loss: {'train/loss': 0.00012171099660918117, 'train/grad_norm_pre_clip': 0.05353228747844696, 'timing/data': 0.00021878999541513622, 'timing/model': 0.39387522198376246, 'learning_rate/qwen_vl_interface': 1.3664173389776735e-06, 'learning_rate/action_model': 1.3664173389776735e-05, 'epoch': 0.3}) +[2026-10-03 19:45:01,361][starVLA.training.train_starvla][INFO] - Step 4044, Loss: {'train/loss': 0.6632144451141357, 'train/grad_norm_pre_clip': 111.72080993652344, 'timing/data': 0.0002561699948273599, 'timing/model': 0.39525541200418957, 'learning_rate/qwen_vl_interface': 1.3646646085646658e-06, 'learning_rate/action_model': 1.3646646085646659e-05, 'epoch': 0.3}) +[2026-10-03 19:45:02,261][starVLA.training.train_starvla][INFO] - Step 4045, Loss: {'train/loss': 0.41166627407073975, 'train/grad_norm_pre_clip': 53.935752868652344, 'timing/data': 0.0003165199887007475, 'timing/model': 0.3942620920133777, 'learning_rate/qwen_vl_interface': 1.3629134752662756e-06, 'learning_rate/action_model': 1.3629134752662756e-05, 'epoch': 0.3}) +[2026-10-03 19:45:02,972][starVLA.training.train_starvla][INFO] - Step 4046, Loss: {'train/loss': 0.001519629848189652, 'train/grad_norm_pre_clip': 1.6928443908691406, 'timing/data': 0.0003119100001640618, 'timing/model': 0.39722812199033797, 'learning_rate/qwen_vl_interface': 1.3611639398023282e-06, 'learning_rate/action_model': 1.3611639398023283e-05, 'epoch': 0.3}) +[2026-10-03 19:45:03,683][starVLA.training.train_starvla][INFO] - Step 4047, Loss: {'train/loss': 0.000291601027129218, 'train/grad_norm_pre_clip': 0.1413513422012329, 'timing/data': 0.00034365002647973597, 'timing/model': 0.39482892202795483, 'learning_rate/qwen_vl_interface': 1.35941600289199e-06, 'learning_rate/action_model': 1.35941600289199e-05, 'epoch': 0.3}) +[2026-10-03 19:45:04,366][starVLA.training.train_starvla][INFO] - Step 4048, Loss: {'train/loss': 0.00036304653622210026, 'train/grad_norm_pre_clip': 0.41319602727890015, 'timing/data': 0.0003158799954690039, 'timing/model': 0.3701910840172786, 'learning_rate/qwen_vl_interface': 1.3576696652537735e-06, 'learning_rate/action_model': 1.3576696652537736e-05, 'epoch': 0.3}) +[2026-10-03 19:45:05,074][starVLA.training.train_starvla][INFO] - Step 4049, Loss: {'train/loss': 0.5416203737258911, 'train/grad_norm_pre_clip': 40.303932189941406, 'timing/data': 0.00037143001100048423, 'timing/model': 0.39635853201616555, 'learning_rate/qwen_vl_interface': 1.355924927605531e-06, 'learning_rate/action_model': 1.3559249276055311e-05, 'epoch': 0.3}) +[2026-10-03 19:45:05,981][starVLA.training.train_starvla][INFO] - Step 4050, Loss: {'train/loss': 0.15988478064537048, 'train/grad_norm_pre_clip': 226.7659454345703, 'timing/data': 0.0003351400082465261, 'timing/model': 0.3957818920025602, 'learning_rate/qwen_vl_interface': 1.3541817906644591e-06, 'learning_rate/action_model': 1.3541817906644591e-05, 'epoch': 0.3}) +[2026-10-03 19:45:06,690][starVLA.training.train_starvla][INFO] - Step 4051, Loss: {'train/loss': 0.00011637101124506444, 'train/grad_norm_pre_clip': 0.08237401396036148, 'timing/data': 0.00023289001546800137, 'timing/model': 0.3956999419897329, 'learning_rate/qwen_vl_interface': 1.3524402551470955e-06, 'learning_rate/action_model': 1.3524402551470956e-05, 'epoch': 0.3}) +[2026-10-03 19:45:07,399][starVLA.training.train_starvla][INFO] - Step 4052, Loss: {'train/loss': 0.0002234236162621528, 'train/grad_norm_pre_clip': 0.1685362160205841, 'timing/data': 0.0002391299931332469, 'timing/model': 0.39546015197993256, 'learning_rate/qwen_vl_interface': 1.3507003217693196e-06, 'learning_rate/action_model': 1.3507003217693196e-05, 'epoch': 0.3}) +[2026-10-03 19:45:08,107][starVLA.training.train_starvla][INFO] - Step 4053, Loss: {'train/loss': 0.00016390447854064405, 'train/grad_norm_pre_clip': 0.057079706341028214, 'timing/data': 0.0003545299987308681, 'timing/model': 0.3946088920056354, 'learning_rate/qwen_vl_interface': 1.3489619912463515e-06, 'learning_rate/action_model': 1.3489619912463514e-05, 'epoch': 0.3}) +[2026-10-03 19:45:08,818][starVLA.training.train_starvla][INFO] - Step 4054, Loss: {'train/loss': 0.9211411476135254, 'train/grad_norm_pre_clip': 81.08973693847656, 'timing/data': 0.00026699999580159783, 'timing/model': 0.3973806629946921, 'learning_rate/qwen_vl_interface': 1.3472252642927534e-06, 'learning_rate/action_model': 1.3472252642927533e-05, 'epoch': 0.3}) +[2026-10-03 19:45:09,724][starVLA.training.train_starvla][INFO] - Step 4055, Loss: {'train/loss': 6.352942727971822e-05, 'train/grad_norm_pre_clip': 0.025022802874445915, 'timing/data': 0.0002726099919527769, 'timing/model': 0.59225330298068, 'learning_rate/qwen_vl_interface': 1.3454901416224284e-06, 'learning_rate/action_model': 1.3454901416224284e-05, 'epoch': 0.3}) +[2026-10-03 19:45:10,433][starVLA.training.train_starvla][INFO] - Step 4056, Loss: {'train/loss': 9.176269668387249e-05, 'train/grad_norm_pre_clip': 0.03511536866426468, 'timing/data': 0.00021964998450130224, 'timing/model': 0.39593597198836505, 'learning_rate/qwen_vl_interface': 1.34375662394862e-06, 'learning_rate/action_model': 1.34375662394862e-05, 'epoch': 0.3}) +[2026-10-03 19:45:11,140][starVLA.training.train_starvla][INFO] - Step 4057, Loss: {'train/loss': 0.0003864853351842612, 'train/grad_norm_pre_clip': 0.35661718249320984, 'timing/data': 0.0002214499982073903, 'timing/model': 0.39515008198213764, 'learning_rate/qwen_vl_interface': 1.3420247119839116e-06, 'learning_rate/action_model': 1.3420247119839117e-05, 'epoch': 0.3}) +[2026-10-03 19:45:11,851][starVLA.training.train_starvla][INFO] - Step 4058, Loss: {'train/loss': 0.00017848175775725394, 'train/grad_norm_pre_clip': 0.054351598024368286, 'timing/data': 0.00029100998654030263, 'timing/model': 0.3976636610168498, 'learning_rate/qwen_vl_interface': 1.340294406440227e-06, 'learning_rate/action_model': 1.340294406440227e-05, 'epoch': 0.3}) +[2026-10-03 19:45:12,559][starVLA.training.train_starvla][INFO] - Step 4059, Loss: {'train/loss': 0.00011108480975963175, 'train/grad_norm_pre_clip': 0.14110565185546875, 'timing/data': 0.00022431998513638973, 'timing/model': 0.3950753020180855, 'learning_rate/qwen_vl_interface': 1.3385657080288297e-06, 'learning_rate/action_model': 1.3385657080288297e-05, 'epoch': 0.3}) +[2026-10-03 19:45:13,477][starVLA.training.train_starvla][INFO] - Step 4060, Loss: {'train/loss': 0.9187124967575073, 'train/grad_norm_pre_clip': 49.388492584228516, 'timing/data': 0.0002371600130572915, 'timing/model': 0.6047829909948632, 'learning_rate/qwen_vl_interface': 1.336838617460322e-06, 'learning_rate/action_model': 1.336838617460322e-05, 'epoch': 0.3}) +[2026-10-03 19:45:14,186][starVLA.training.train_starvla][INFO] - Step 4061, Loss: {'train/loss': 0.0005229402449913323, 'train/grad_norm_pre_clip': 0.3148951232433319, 'timing/data': 0.00030417999369092286, 'timing/model': 0.39710544201079756, 'learning_rate/qwen_vl_interface': 1.335113135444646e-06, 'learning_rate/action_model': 1.335113135444646e-05, 'epoch': 0.3}) +[2026-10-03 19:45:14,897][starVLA.training.train_starvla][INFO] - Step 4062, Loss: {'train/loss': 9.190863784169778e-05, 'train/grad_norm_pre_clip': 0.043797727674245834, 'timing/data': 0.00023239999427460134, 'timing/model': 0.39712793100625277, 'learning_rate/qwen_vl_interface': 1.3333892626910823e-06, 'learning_rate/action_model': 1.3333892626910823e-05, 'epoch': 0.3}) +[2026-10-03 19:45:15,607][starVLA.training.train_starvla][INFO] - Step 4063, Loss: {'train/loss': 0.00021511917293537408, 'train/grad_norm_pre_clip': 0.10651951283216476, 'timing/data': 0.0002313499862793833, 'timing/model': 0.3961820419935975, 'learning_rate/qwen_vl_interface': 1.3316669999082486e-06, 'learning_rate/action_model': 1.3316669999082484e-05, 'epoch': 0.3}) +[2026-10-03 19:45:16,316][starVLA.training.train_starvla][INFO] - Step 4064, Loss: {'train/loss': 0.16279304027557373, 'train/grad_norm_pre_clip': 113.05525207519531, 'timing/data': 0.0003291700268164277, 'timing/model': 0.3972329520038329, 'learning_rate/qwen_vl_interface': 1.3299463478041024e-06, 'learning_rate/action_model': 1.3299463478041024e-05, 'epoch': 0.3}) +[2026-10-03 19:45:17,027][starVLA.training.train_starvla][INFO] - Step 4065, Loss: {'train/loss': 0.0020784332882612944, 'train/grad_norm_pre_clip': 4.314663887023926, 'timing/data': 0.00037351998616941273, 'timing/model': 0.3963388009869959, 'learning_rate/qwen_vl_interface': 1.3282273070859396e-06, 'learning_rate/action_model': 1.3282273070859396e-05, 'epoch': 0.3}) +[2026-10-03 19:45:17,942][starVLA.training.train_starvla][INFO] - Step 4066, Loss: {'train/loss': 4.3442167225293815e-05, 'train/grad_norm_pre_clip': 0.013767645694315434, 'timing/data': 0.0003891700180247426, 'timing/model': 0.3972769009997137, 'learning_rate/qwen_vl_interface': 1.3265098784603924e-06, 'learning_rate/action_model': 1.3265098784603924e-05, 'epoch': 0.3}) +[2026-10-03 19:45:18,649][starVLA.training.train_starvla][INFO] - Step 4067, Loss: {'train/loss': 0.0003000509459525347, 'train/grad_norm_pre_clip': 0.17437686026096344, 'timing/data': 0.0002307599934283644, 'timing/model': 0.3953392520197667, 'learning_rate/qwen_vl_interface': 1.324794062633431e-06, 'learning_rate/action_model': 1.324794062633431e-05, 'epoch': 0.3}) +[2026-10-03 19:45:19,357][starVLA.training.train_starvla][INFO] - Step 4068, Loss: {'train/loss': 4.3404103053035215e-05, 'train/grad_norm_pre_clip': 0.0071126134134829044, 'timing/data': 0.0003762199776247144, 'timing/model': 0.3946402220171876, 'learning_rate/qwen_vl_interface': 1.323079860310363e-06, 'learning_rate/action_model': 1.3230798603103628e-05, 'epoch': 0.3}) +[2026-10-03 19:45:20,064][starVLA.training.train_starvla][INFO] - Step 4069, Loss: {'train/loss': 0.13597546517848969, 'train/grad_norm_pre_clip': 95.07049560546875, 'timing/data': 0.0002987300103995949, 'timing/model': 0.3960579220147338, 'learning_rate/qwen_vl_interface': 1.3213672721958309e-06, 'learning_rate/action_model': 1.321367272195831e-05, 'epoch': 0.3}) +[2026-10-03 19:45:20,774][starVLA.training.train_starvla][INFO] - Step 4070, Loss: {'train/loss': 0.0005080661503598094, 'train/grad_norm_pre_clip': 0.4355555474758148, 'timing/data': 0.0003365300071891397, 'timing/model': 0.39702459098771214, 'learning_rate/qwen_vl_interface': 1.3196562989938175e-06, 'learning_rate/action_model': 1.3196562989938174e-05, 'epoch': 0.3}) +[2026-10-03 19:45:21,686][starVLA.training.train_starvla][INFO] - Step 4071, Loss: {'train/loss': 0.03939361125230789, 'train/grad_norm_pre_clip': 95.95167541503906, 'timing/data': 0.00034030998358502984, 'timing/model': 0.5984880120086018, 'learning_rate/qwen_vl_interface': 1.3179469414076357e-06, 'learning_rate/action_model': 1.3179469414076356e-05, 'epoch': 0.3}) +[2026-10-03 19:45:22,394][starVLA.training.train_starvla][INFO] - Step 4072, Loss: {'train/loss': 0.2628938853740692, 'train/grad_norm_pre_clip': 142.26376342773438, 'timing/data': 0.00038629001937806606, 'timing/model': 0.394835832004901, 'learning_rate/qwen_vl_interface': 1.3162392001399401e-06, 'learning_rate/action_model': 1.3162392001399401e-05, 'epoch': 0.3}) +[2026-10-03 19:45:23,102][starVLA.training.train_starvla][INFO] - Step 4073, Loss: {'train/loss': 0.00010477781324880198, 'train/grad_norm_pre_clip': 0.04266197606921196, 'timing/data': 0.00036376999923959374, 'timing/model': 0.39647926200996153, 'learning_rate/qwen_vl_interface': 1.3145330758927182e-06, 'learning_rate/action_model': 1.3145330758927182e-05, 'epoch': 0.3}) +[2026-10-03 19:45:23,813][starVLA.training.train_starvla][INFO] - Step 4074, Loss: {'train/loss': 0.0007849277462810278, 'train/grad_norm_pre_clip': 1.4829005002975464, 'timing/data': 0.0002551400102674961, 'timing/model': 0.3976999119913671, 'learning_rate/qwen_vl_interface': 1.3128285693672935e-06, 'learning_rate/action_model': 1.3128285693672934e-05, 'epoch': 0.3}) +[2026-10-03 19:45:24,520][starVLA.training.train_starvla][INFO] - Step 4075, Loss: {'train/loss': 0.8440214991569519, 'train/grad_norm_pre_clip': 149.74729919433594, 'timing/data': 0.0004538699868135154, 'timing/model': 0.3958103419863619, 'learning_rate/qwen_vl_interface': 1.311125681264325e-06, 'learning_rate/action_model': 1.311125681264325e-05, 'epoch': 0.3}) +[2026-10-03 19:45:25,429][starVLA.training.train_starvla][INFO] - Step 4076, Loss: {'train/loss': 0.0004934673197567463, 'train/grad_norm_pre_clip': 0.3489975333213806, 'timing/data': 0.00032705001649446785, 'timing/model': 0.5961029220197815, 'learning_rate/qwen_vl_interface': 1.309424412283803e-06, 'learning_rate/action_model': 1.3094244122838029e-05, 'epoch': 0.3}) +[2026-10-03 19:45:26,139][starVLA.training.train_starvla][INFO] - Step 4077, Loss: {'train/loss': 7.687650213483721e-05, 'train/grad_norm_pre_clip': 0.01933799684047699, 'timing/data': 0.00033204001374542713, 'timing/model': 0.3953638020029757, 'learning_rate/qwen_vl_interface': 1.307724763125057e-06, 'learning_rate/action_model': 1.307724763125057e-05, 'epoch': 0.3}) +[2026-10-03 19:45:26,846][starVLA.training.train_starvla][INFO] - Step 4078, Loss: {'train/loss': 3.547035885276273e-05, 'train/grad_norm_pre_clip': 0.008188284933567047, 'timing/data': 0.0002721100172493607, 'timing/model': 0.3953486419923138, 'learning_rate/qwen_vl_interface': 1.3060267344867478e-06, 'learning_rate/action_model': 1.3060267344867478e-05, 'epoch': 0.3}) +[2026-10-03 19:45:27,553][starVLA.training.train_starvla][INFO] - Step 4079, Loss: {'train/loss': 0.0053192428313195705, 'train/grad_norm_pre_clip': 15.0296630859375, 'timing/data': 0.00032449999707750976, 'timing/model': 0.39582485100254416, 'learning_rate/qwen_vl_interface': 1.3043303270668715e-06, 'learning_rate/action_model': 1.3043303270668713e-05, 'epoch': 0.3}) +[2026-10-03 19:45:28,262][starVLA.training.train_starvla][INFO] - Step 4080, Loss: {'train/loss': 0.00022067921236157417, 'train/grad_norm_pre_clip': 0.10655289143323898, 'timing/data': 0.00034945999504998326, 'timing/model': 0.39571389099000953, 'learning_rate/qwen_vl_interface': 1.3026355415627565e-06, 'learning_rate/action_model': 1.3026355415627564e-05, 'epoch': 0.3}) +[2026-10-03 19:45:28,968][starVLA.training.train_starvla][INFO] - Step 4081, Loss: {'train/loss': 0.00018862014985643327, 'train/grad_norm_pre_clip': 0.09815511107444763, 'timing/data': 0.00034748000325635076, 'timing/model': 0.3939008819870651, 'learning_rate/qwen_vl_interface': 1.3009423786710653e-06, 'learning_rate/action_model': 1.3009423786710653e-05, 'epoch': 0.3}) +[2026-10-03 19:45:29,881][starVLA.training.train_starvla][INFO] - Step 4082, Loss: {'train/loss': 0.00015600284677930176, 'train/grad_norm_pre_clip': 0.29563239216804504, 'timing/data': 0.00028608000138774514, 'timing/model': 0.3957569320045877, 'learning_rate/qwen_vl_interface': 1.2992508390877933e-06, 'learning_rate/action_model': 1.2992508390877933e-05, 'epoch': 0.3}) +[2026-10-03 19:45:30,588][starVLA.training.train_starvla][INFO] - Step 4083, Loss: {'train/loss': 0.3464263677597046, 'train/grad_norm_pre_clip': 102.91751098632812, 'timing/data': 0.0003284499980509281, 'timing/model': 0.3951303719950374, 'learning_rate/qwen_vl_interface': 1.297560923508269e-06, 'learning_rate/action_model': 1.297560923508269e-05, 'epoch': 0.3}) +[2026-10-03 19:45:31,296][starVLA.training.train_starvla][INFO] - Step 4084, Loss: {'train/loss': 0.027551278471946716, 'train/grad_norm_pre_clip': 46.40520477294922, 'timing/data': 0.0002597999991849065, 'timing/model': 0.3955384909932036, 'learning_rate/qwen_vl_interface': 1.295872632627152e-06, 'learning_rate/action_model': 1.2958726326271519e-05, 'epoch': 0.3}) +[2026-10-03 19:45:32,003][starVLA.training.train_starvla][INFO] - Step 4085, Loss: {'train/loss': 5.6110391597030684e-05, 'train/grad_norm_pre_clip': 0.021634556353092194, 'timing/data': 0.0002191700041294098, 'timing/model': 0.39437771300436, 'learning_rate/qwen_vl_interface': 1.2941859671384364e-06, 'learning_rate/action_model': 1.2941859671384362e-05, 'epoch': 0.3}) +[2026-10-03 19:45:32,711][starVLA.training.train_starvla][INFO] - Step 4086, Loss: {'train/loss': 0.6261768341064453, 'train/grad_norm_pre_clip': 114.61749267578125, 'timing/data': 0.00025368001661263406, 'timing/model': 0.3953682720020879, 'learning_rate/qwen_vl_interface': 1.2925009277354444e-06, 'learning_rate/action_model': 1.2925009277354444e-05, 'epoch': 0.3}) +[2026-10-03 19:45:33,622][starVLA.training.train_starvla][INFO] - Step 4087, Loss: {'train/loss': 0.5803653597831726, 'train/grad_norm_pre_clip': 131.48622131347656, 'timing/data': 0.0002700400073081255, 'timing/model': 0.3951759409974329, 'learning_rate/qwen_vl_interface': 1.2908175151108335e-06, 'learning_rate/action_model': 1.2908175151108334e-05, 'epoch': 0.3}) +[2026-10-03 19:45:34,331][starVLA.training.train_starvla][INFO] - Step 4088, Loss: {'train/loss': 0.41797739267349243, 'train/grad_norm_pre_clip': 314.8045349121094, 'timing/data': 0.00026011999580077827, 'timing/model': 0.3954706520016771, 'learning_rate/qwen_vl_interface': 1.2891357299565899e-06, 'learning_rate/action_model': 1.2891357299565898e-05, 'epoch': 0.3}) +[2026-10-03 19:45:35,039][starVLA.training.train_starvla][INFO] - Step 4089, Loss: {'train/loss': 0.004984408151358366, 'train/grad_norm_pre_clip': 6.738125801086426, 'timing/data': 0.00022087999968789518, 'timing/model': 0.39579655198031105, 'learning_rate/qwen_vl_interface': 1.2874555729640333e-06, 'learning_rate/action_model': 1.2874555729640334e-05, 'epoch': 0.3}) +[2026-10-03 19:45:35,746][starVLA.training.train_starvla][INFO] - Step 4090, Loss: {'train/loss': 0.0683555081486702, 'train/grad_norm_pre_clip': 74.14531707763672, 'timing/data': 0.00035975902574136853, 'timing/model': 0.3949808019970078, 'learning_rate/qwen_vl_interface': 1.2857770448238124e-06, 'learning_rate/action_model': 1.2857770448238123e-05, 'epoch': 0.3}) +[2026-10-03 19:45:36,455][starVLA.training.train_starvla][INFO] - Step 4091, Loss: {'train/loss': 0.007777602877467871, 'train/grad_norm_pre_clip': 33.5555534362793, 'timing/data': 0.00029858999187126756, 'timing/model': 0.3969506509893108, 'learning_rate/qwen_vl_interface': 1.2841001462259067e-06, 'learning_rate/action_model': 1.2841001462259067e-05, 'epoch': 0.3}) +[2026-10-03 19:45:37,363][starVLA.training.train_starvla][INFO] - Step 4092, Loss: {'train/loss': 9.306536230724305e-05, 'train/grad_norm_pre_clip': 0.07856758683919907, 'timing/data': 0.00021697001648135483, 'timing/model': 0.5973698529996909, 'learning_rate/qwen_vl_interface': 1.282424877859626e-06, 'learning_rate/action_model': 1.282424877859626e-05, 'epoch': 0.3}) +[2026-10-03 19:45:38,070][starVLA.training.train_starvla][INFO] - Step 4093, Loss: {'train/loss': 0.0007812962867319584, 'train/grad_norm_pre_clip': 1.843083143234253, 'timing/data': 0.00022292000357992947, 'timing/model': 0.3962683909921907, 'learning_rate/qwen_vl_interface': 1.2807512404136088e-06, 'learning_rate/action_model': 1.2807512404136087e-05, 'epoch': 0.3}) +[2026-10-03 19:45:38,779][starVLA.training.train_starvla][INFO] - Step 4094, Loss: {'train/loss': 0.0017676937859505415, 'train/grad_norm_pre_clip': 4.177374362945557, 'timing/data': 0.000245939998421818, 'timing/model': 0.395159292005701, 'learning_rate/qwen_vl_interface': 1.2790792345758255e-06, 'learning_rate/action_model': 1.2790792345758254e-05, 'epoch': 0.3}) +[2026-10-03 19:45:39,486][starVLA.training.train_starvla][INFO] - Step 4095, Loss: {'train/loss': 0.028619281947612762, 'train/grad_norm_pre_clip': 196.65093994140625, 'timing/data': 0.0003104399947915226, 'timing/model': 0.3950691520003602, 'learning_rate/qwen_vl_interface': 1.2774088610335752e-06, 'learning_rate/action_model': 1.2774088610335752e-05, 'epoch': 0.3}) +[2026-10-03 19:45:40,192][starVLA.training.train_starvla][INFO] - Step 4096, Loss: {'train/loss': 4.733452806249261e-05, 'train/grad_norm_pre_clip': 0.045913998037576675, 'timing/data': 0.00022057999740354717, 'timing/model': 0.39580396201927215, 'learning_rate/qwen_vl_interface': 1.2757401204734827e-06, 'learning_rate/action_model': 1.2757401204734826e-05, 'epoch': 0.3}) +[2026-10-03 19:45:41,102][starVLA.training.train_starvla][INFO] - Step 4097, Loss: {'train/loss': 0.01322069950401783, 'train/grad_norm_pre_clip': 26.77637481689453, 'timing/data': 0.00021860998822376132, 'timing/model': 0.5986365620046854, 'learning_rate/qwen_vl_interface': 1.2740730135815064e-06, 'learning_rate/action_model': 1.2740730135815062e-05, 'epoch': 0.3}) +[2026-10-03 19:45:41,811][starVLA.training.train_starvla][INFO] - Step 4098, Loss: {'train/loss': 0.6492446064949036, 'train/grad_norm_pre_clip': 62.70882797241211, 'timing/data': 0.0003013600071426481, 'timing/model': 0.3968932220013812, 'learning_rate/qwen_vl_interface': 1.2724075410429298e-06, 'learning_rate/action_model': 1.2724075410429299e-05, 'epoch': 0.3}) +[2026-10-03 19:45:42,517][starVLA.training.train_starvla][INFO] - Step 4099, Loss: {'train/loss': 0.6757563352584839, 'train/grad_norm_pre_clip': 296.4862365722656, 'timing/data': 0.00021630999981425703, 'timing/model': 0.39586580099421553, 'learning_rate/qwen_vl_interface': 1.2707437035423664e-06, 'learning_rate/action_model': 1.2707437035423664e-05, 'epoch': 0.3}) +[2026-10-03 19:45:43,220][starVLA.training.train_starvla][INFO] - Step 4100, Loss: {'train/loss': 0.033554546535015106, 'train/grad_norm_pre_clip': 62.93437576293945, 'timing/data': 0.00035571001353673637, 'timing/model': 0.3929928220168222, 'learning_rate/qwen_vl_interface': 1.2690815017637573e-06, 'learning_rate/action_model': 1.2690815017637572e-05, 'epoch': 0.3}) +[2026-10-03 19:45:43,927][starVLA.training.train_starvla][INFO] - Step 4101, Loss: {'train/loss': 0.00030298615456558764, 'train/grad_norm_pre_clip': 0.261740505695343, 'timing/data': 0.00021698998170904815, 'timing/model': 0.39426393099711277, 'learning_rate/qwen_vl_interface': 1.2674209363903694e-06, 'learning_rate/action_model': 1.2674209363903693e-05, 'epoch': 0.3}) +[2026-10-03 19:45:44,838][starVLA.training.train_starvla][INFO] - Step 4102, Loss: {'train/loss': 0.134898379445076, 'train/grad_norm_pre_clip': 113.88005828857422, 'timing/data': 0.0003383499861229211, 'timing/model': 0.6011851329822093, 'learning_rate/qwen_vl_interface': 1.2657620081048005e-06, 'learning_rate/action_model': 1.2657620081048006e-05, 'epoch': 0.3}) +[2026-10-03 19:45:45,545][starVLA.training.train_starvla][INFO] - Step 4103, Loss: {'train/loss': 0.2422635853290558, 'train/grad_norm_pre_clip': 146.04991149902344, 'timing/data': 0.00022691002232022583, 'timing/model': 0.39450583199504763, 'learning_rate/qwen_vl_interface': 1.2641047175889714e-06, 'learning_rate/action_model': 1.2641047175889714e-05, 'epoch': 0.3}) +[2026-10-03 19:45:46,254][starVLA.training.train_starvla][INFO] - Step 4104, Loss: {'train/loss': 0.00011892613838426769, 'train/grad_norm_pre_clip': 0.12069043517112732, 'timing/data': 0.0004725600010715425, 'timing/model': 0.3950748119968921, 'learning_rate/qwen_vl_interface': 1.2624490655241332e-06, 'learning_rate/action_model': 1.262449065524133e-05, 'epoch': 0.3}) +[2026-10-03 19:45:46,962][starVLA.training.train_starvla][INFO] - Step 4105, Loss: {'train/loss': 1.598869130248204e-05, 'train/grad_norm_pre_clip': 0.0033210921101272106, 'timing/data': 0.00025534999440424144, 'timing/model': 0.396043311979156, 'learning_rate/qwen_vl_interface': 1.2607950525908602e-06, 'learning_rate/action_model': 1.2607950525908601e-05, 'epoch': 0.3}) +[2026-10-03 19:45:47,671][starVLA.training.train_starvla][INFO] - Step 4106, Loss: {'train/loss': 0.5933148264884949, 'train/grad_norm_pre_clip': 42.65811538696289, 'timing/data': 0.00026956998044624925, 'timing/model': 0.39686802198411897, 'learning_rate/qwen_vl_interface': 1.2591426794690575e-06, 'learning_rate/action_model': 1.2591426794690576e-05, 'epoch': 0.3}) +[2026-10-03 19:45:48,378][starVLA.training.train_starvla][INFO] - Step 4107, Loss: {'train/loss': 0.13086025416851044, 'train/grad_norm_pre_clip': 242.60369873046875, 'timing/data': 0.00022108000121079385, 'timing/model': 0.3954965920129325, 'learning_rate/qwen_vl_interface': 1.25749194683795e-06, 'learning_rate/action_model': 1.25749194683795e-05, 'epoch': 0.3}) +[2026-10-03 19:45:49,285][starVLA.training.train_starvla][INFO] - Step 4108, Loss: {'train/loss': 7.913757872302085e-05, 'train/grad_norm_pre_clip': 0.0316668376326561, 'timing/data': 0.00023425999097526073, 'timing/model': 0.3942690920084715, 'learning_rate/qwen_vl_interface': 1.2558428553760933e-06, 'learning_rate/action_model': 1.2558428553760932e-05, 'epoch': 0.3}) +[2026-10-03 19:45:49,991][starVLA.training.train_starvla][INFO] - Step 4109, Loss: {'train/loss': 0.010396508499979973, 'train/grad_norm_pre_clip': 15.241480827331543, 'timing/data': 0.00023397000040858984, 'timing/model': 0.39514063199749216, 'learning_rate/qwen_vl_interface': 1.2541954057613659e-06, 'learning_rate/action_model': 1.2541954057613658e-05, 'epoch': 0.3}) +[2026-10-03 19:45:50,700][starVLA.training.train_starvla][INFO] - Step 4110, Loss: {'train/loss': 0.543953537940979, 'train/grad_norm_pre_clip': 42.71406173706055, 'timing/data': 0.00037028000224381685, 'timing/model': 0.39698019198840484, 'learning_rate/qwen_vl_interface': 1.2525495986709727e-06, 'learning_rate/action_model': 1.2525495986709726e-05, 'epoch': 0.3}) +[2026-10-03 19:45:51,410][starVLA.training.train_starvla][INFO] - Step 4111, Loss: {'train/loss': 0.0034431100357323885, 'train/grad_norm_pre_clip': 7.124040126800537, 'timing/data': 0.0003486500063445419, 'timing/model': 0.39632264198735356, 'learning_rate/qwen_vl_interface': 1.2509054347814423e-06, 'learning_rate/action_model': 1.2509054347814423e-05, 'epoch': 0.3}) +[2026-10-03 19:45:52,118][starVLA.training.train_starvla][INFO] - Step 4112, Loss: {'train/loss': 9.4492883363273e-05, 'train/grad_norm_pre_clip': 0.063865527510643, 'timing/data': 0.00022560000070370734, 'timing/model': 0.3953267019824125, 'learning_rate/qwen_vl_interface': 1.2492629147686288e-06, 'learning_rate/action_model': 1.2492629147686289e-05, 'epoch': 0.3}) +[2026-10-03 19:45:53,021][starVLA.training.train_starvla][INFO] - Step 4113, Loss: {'train/loss': 9.941938333213329e-05, 'train/grad_norm_pre_clip': 0.03388388454914093, 'timing/data': 0.0003550700203049928, 'timing/model': 0.5907890030066483, 'learning_rate/qwen_vl_interface': 1.2476220393077098e-06, 'learning_rate/action_model': 1.2476220393077098e-05, 'epoch': 0.3}) +[2026-10-03 19:45:53,729][starVLA.training.train_starvla][INFO] - Step 4114, Loss: {'train/loss': 7.6927593909204e-05, 'train/grad_norm_pre_clip': 0.023636596277356148, 'timing/data': 0.0002984600141644478, 'timing/model': 0.3951825620024465, 'learning_rate/qwen_vl_interface': 1.2459828090731877e-06, 'learning_rate/action_model': 1.2459828090731876e-05, 'epoch': 0.3}) +[2026-10-03 19:45:54,436][starVLA.training.train_starvla][INFO] - Step 4115, Loss: {'train/loss': 0.31884101033210754, 'train/grad_norm_pre_clip': 275.44598388671875, 'timing/data': 0.00029091001488268375, 'timing/model': 0.3949843120062724, 'learning_rate/qwen_vl_interface': 1.2443452247388876e-06, 'learning_rate/action_model': 1.2443452247388877e-05, 'epoch': 0.3}) +[2026-10-03 19:45:55,143][starVLA.training.train_starvla][INFO] - Step 4116, Loss: {'train/loss': 7.110206934157759e-05, 'train/grad_norm_pre_clip': 0.02172914892435074, 'timing/data': 0.000244739989284426, 'timing/model': 0.39573686200310476, 'learning_rate/qwen_vl_interface': 1.2427092869779603e-06, 'learning_rate/action_model': 1.2427092869779602e-05, 'epoch': 0.3}) +[2026-10-03 19:45:55,847][starVLA.training.train_starvla][INFO] - Step 4117, Loss: {'train/loss': 0.7354474663734436, 'train/grad_norm_pre_clip': 102.35113525390625, 'timing/data': 0.0002112200018018484, 'timing/model': 0.3936802020180039, 'learning_rate/qwen_vl_interface': 1.2410749964628751e-06, 'learning_rate/action_model': 1.2410749964628751e-05, 'epoch': 0.3}) +[2026-10-03 19:45:56,556][starVLA.training.train_starvla][INFO] - Step 4118, Loss: {'train/loss': 0.0005480422987602651, 'train/grad_norm_pre_clip': 0.323127806186676, 'timing/data': 0.00031094998121261597, 'timing/model': 0.39654574100859463, 'learning_rate/qwen_vl_interface': 1.2394423538654296e-06, 'learning_rate/action_model': 1.2394423538654296e-05, 'epoch': 0.3}) +[2026-10-03 19:45:57,463][starVLA.training.train_starvla][INFO] - Step 4119, Loss: {'train/loss': 0.058426808565855026, 'train/grad_norm_pre_clip': 77.1476821899414, 'timing/data': 0.0002612199750728905, 'timing/model': 0.3941930319997482, 'learning_rate/qwen_vl_interface': 1.2378113598567394e-06, 'learning_rate/action_model': 1.2378113598567392e-05, 'epoch': 0.3}) +[2026-10-03 19:45:58,170][starVLA.training.train_starvla][INFO] - Step 4120, Loss: {'train/loss': 4.964685285813175e-05, 'train/grad_norm_pre_clip': 0.014615165069699287, 'timing/data': 0.00030958998831920326, 'timing/model': 0.39395069199963473, 'learning_rate/qwen_vl_interface': 1.2361820151072454e-06, 'learning_rate/action_model': 1.2361820151072453e-05, 'epoch': 0.3}) +[2026-10-03 19:45:58,878][starVLA.training.train_starvla][INFO] - Step 4121, Loss: {'train/loss': 0.0006274838233366609, 'train/grad_norm_pre_clip': 0.31931328773498535, 'timing/data': 0.00035754000418819487, 'timing/model': 0.39532962199882604, 'learning_rate/qwen_vl_interface': 1.23455432028671e-06, 'learning_rate/action_model': 1.23455432028671e-05, 'epoch': 0.3}) +[2026-10-03 19:45:59,587][starVLA.training.train_starvla][INFO] - Step 4122, Loss: {'train/loss': 0.5918706655502319, 'train/grad_norm_pre_clip': 45.165733337402344, 'timing/data': 0.0003102289920207113, 'timing/model': 0.39674654201371595, 'learning_rate/qwen_vl_interface': 1.2329282760642167e-06, 'learning_rate/action_model': 1.2329282760642166e-05, 'epoch': 0.3}) +[2026-10-03 19:46:00,294][starVLA.training.train_starvla][INFO] - Step 4123, Loss: {'train/loss': 5.383168172556907e-05, 'train/grad_norm_pre_clip': 0.01603655517101288, 'timing/data': 0.0002397600037511438, 'timing/model': 0.3958817620004993, 'learning_rate/qwen_vl_interface': 1.2313038831081716e-06, 'learning_rate/action_model': 1.2313038831081717e-05, 'epoch': 0.3}) +[2026-10-03 19:46:01,198][starVLA.training.train_starvla][INFO] - Step 4124, Loss: {'train/loss': 4.4097490899730474e-05, 'train/grad_norm_pre_clip': 0.015610978938639164, 'timing/data': 0.00032663001911714673, 'timing/model': 0.3953984020045027, 'learning_rate/qwen_vl_interface': 1.2296811420863002e-06, 'learning_rate/action_model': 1.2296811420863003e-05, 'epoch': 0.3}) +[2026-10-03 19:46:01,906][starVLA.training.train_starvla][INFO] - Step 4125, Loss: {'train/loss': 0.0011864359257742763, 'train/grad_norm_pre_clip': 1.6629244089126587, 'timing/data': 0.0003035400004591793, 'timing/model': 0.3956742520094849, 'learning_rate/qwen_vl_interface': 1.2280600536656504e-06, 'learning_rate/action_model': 1.2280600536656503e-05, 'epoch': 0.3}) +[2026-10-03 19:46:02,613][starVLA.training.train_starvla][INFO] - Step 4126, Loss: {'train/loss': 5.732504723710008e-05, 'train/grad_norm_pre_clip': 0.02047952264547348, 'timing/data': 0.0002915700024459511, 'timing/model': 0.3950121619855054, 'learning_rate/qwen_vl_interface': 1.2264406185125907e-06, 'learning_rate/action_model': 1.2264406185125907e-05, 'epoch': 0.3}) +[2026-10-03 19:46:03,321][starVLA.training.train_starvla][INFO] - Step 4127, Loss: {'train/loss': 0.7378389835357666, 'train/grad_norm_pre_clip': 177.6047821044922, 'timing/data': 0.00022208999143913388, 'timing/model': 0.3961384819995146, 'learning_rate/qwen_vl_interface': 1.2248228372928098e-06, 'learning_rate/action_model': 1.2248228372928097e-05, 'epoch': 0.3}) +[2026-10-03 19:46:04,028][starVLA.training.train_starvla][INFO] - Step 4128, Loss: {'train/loss': 0.5672357082366943, 'train/grad_norm_pre_clip': 154.88821411132812, 'timing/data': 0.0002996500115841627, 'timing/model': 0.39551066199783236, 'learning_rate/qwen_vl_interface': 1.2232067106713174e-06, 'learning_rate/action_model': 1.2232067106713174e-05, 'epoch': 0.3}) +[2026-10-03 19:46:04,936][starVLA.training.train_starvla][INFO] - Step 4129, Loss: {'train/loss': 4.5460772525984794e-05, 'train/grad_norm_pre_clip': 0.014826416969299316, 'timing/data': 0.0002260100154671818, 'timing/model': 0.5946944930183236, 'learning_rate/qwen_vl_interface': 1.22159223931244e-06, 'learning_rate/action_model': 1.2215922393124399e-05, 'epoch': 0.3}) +[2026-10-03 19:46:05,645][starVLA.training.train_starvla][INFO] - Step 4130, Loss: {'train/loss': 4.282883310224861e-05, 'train/grad_norm_pre_clip': 0.017385244369506836, 'timing/data': 0.0002125999890267849, 'timing/model': 0.3971911810222082, 'learning_rate/qwen_vl_interface': 1.2199794238798266e-06, 'learning_rate/action_model': 1.2199794238798264e-05, 'epoch': 0.3}) +[2026-10-03 19:46:06,358][starVLA.training.train_starvla][INFO] - Step 4131, Loss: {'train/loss': 0.000899420352652669, 'train/grad_norm_pre_clip': 0.558430016040802, 'timing/data': 0.00047581997932866216, 'timing/model': 0.395801291015232, 'learning_rate/qwen_vl_interface': 1.2183682650364447e-06, 'learning_rate/action_model': 1.2183682650364447e-05, 'epoch': 0.3}) +[2026-10-03 19:46:07,068][starVLA.training.train_starvla][INFO] - Step 4132, Loss: {'train/loss': 0.14310549199581146, 'train/grad_norm_pre_clip': 189.55398559570312, 'timing/data': 0.00036462998832575977, 'timing/model': 0.3958817319944501, 'learning_rate/qwen_vl_interface': 1.216758763444581e-06, 'learning_rate/action_model': 1.2167587634445809e-05, 'epoch': 0.3}) +[2026-10-03 19:46:07,776][starVLA.training.train_starvla][INFO] - Step 4133, Loss: {'train/loss': 0.3028716444969177, 'train/grad_norm_pre_clip': 188.46206665039062, 'timing/data': 0.00023479899391531944, 'timing/model': 0.3961494829854928, 'learning_rate/qwen_vl_interface': 1.215150919765841e-06, 'learning_rate/action_model': 1.215150919765841e-05, 'epoch': 0.3}) +[2026-10-03 19:46:08,485][starVLA.training.train_starvla][INFO] - Step 4134, Loss: {'train/loss': 0.00023295002756640315, 'train/grad_norm_pre_clip': 0.5855926871299744, 'timing/data': 0.0004326899943407625, 'timing/model': 0.39616769200074486, 'learning_rate/qwen_vl_interface': 1.213544734661148e-06, 'learning_rate/action_model': 1.2135447346611478e-05, 'epoch': 0.3}) +[2026-10-03 19:46:09,391][starVLA.training.train_starvla][INFO] - Step 4135, Loss: {'train/loss': 0.00016579314251430333, 'train/grad_norm_pre_clip': 0.0614892803132534, 'timing/data': 0.0003968999953940511, 'timing/model': 0.395614071021555, 'learning_rate/qwen_vl_interface': 1.2119402087907448e-06, 'learning_rate/action_model': 1.2119402087907448e-05, 'epoch': 0.3}) +[2026-10-03 19:46:10,097][starVLA.training.train_starvla][INFO] - Step 4136, Loss: {'train/loss': 7.64586657169275e-05, 'train/grad_norm_pre_clip': 0.01825781539082527, 'timing/data': 0.0003632300067692995, 'timing/model': 0.3956012609996833, 'learning_rate/qwen_vl_interface': 1.2103373428141903e-06, 'learning_rate/action_model': 1.2103373428141903e-05, 'epoch': 0.3}) +[2026-10-03 19:46:10,805][starVLA.training.train_starvla][INFO] - Step 4137, Loss: {'train/loss': 0.5242038369178772, 'train/grad_norm_pre_clip': 131.20428466796875, 'timing/data': 0.0003436499973759055, 'timing/model': 0.39549465200980194, 'learning_rate/qwen_vl_interface': 1.2087361373903633e-06, 'learning_rate/action_model': 1.2087361373903633e-05, 'epoch': 0.3}) +[2026-10-03 19:46:11,514][starVLA.training.train_starvla][INFO] - Step 4138, Loss: {'train/loss': 0.5248193740844727, 'train/grad_norm_pre_clip': 40.22433853149414, 'timing/data': 0.00032070997986011207, 'timing/model': 0.39532410202082247, 'learning_rate/qwen_vl_interface': 1.2071365931774584e-06, 'learning_rate/action_model': 1.2071365931774584e-05, 'epoch': 0.3}) +[2026-10-03 19:46:12,222][starVLA.training.train_starvla][INFO] - Step 4139, Loss: {'train/loss': 2.6716963475337252e-05, 'train/grad_norm_pre_clip': 0.006118598394095898, 'timing/data': 0.00032192000071518123, 'timing/model': 0.3952503210166469, 'learning_rate/qwen_vl_interface': 1.2055387108329885e-06, 'learning_rate/action_model': 1.2055387108329886e-05, 'epoch': 0.3}) +[2026-10-03 19:46:13,132][starVLA.training.train_starvla][INFO] - Step 4140, Loss: {'train/loss': 0.0007131661986932158, 'train/grad_norm_pre_clip': 0.5556585192680359, 'timing/data': 0.0003340599942021072, 'timing/model': 0.3955324820126407, 'learning_rate/qwen_vl_interface': 1.2039424910137807e-06, 'learning_rate/action_model': 1.2039424910137806e-05, 'epoch': 0.3}) +[2026-10-03 19:46:13,838][starVLA.training.train_starvla][INFO] - Step 4141, Loss: {'train/loss': 0.1926024854183197, 'train/grad_norm_pre_clip': 99.93775177001953, 'timing/data': 0.0003370099875610322, 'timing/model': 0.3943833220109809, 'learning_rate/qwen_vl_interface': 1.2023479343759822e-06, 'learning_rate/action_model': 1.2023479343759823e-05, 'epoch': 0.3}) +[2026-10-03 19:46:14,547][starVLA.training.train_starvla][INFO] - Step 4142, Loss: {'train/loss': 0.017317622900009155, 'train/grad_norm_pre_clip': 26.83816909790039, 'timing/data': 0.0003510399837978184, 'timing/model': 0.3969780520128552, 'learning_rate/qwen_vl_interface': 1.2007550415750546e-06, 'learning_rate/action_model': 1.2007550415750545e-05, 'epoch': 0.3}) +[2026-10-03 19:46:15,253][starVLA.training.train_starvla][INFO] - Step 4143, Loss: {'train/loss': 0.0001950328005477786, 'train/grad_norm_pre_clip': 0.14810872077941895, 'timing/data': 0.00028305998421274126, 'timing/model': 0.3946709719893988, 'learning_rate/qwen_vl_interface': 1.1991638132657759e-06, 'learning_rate/action_model': 1.199163813265776e-05, 'epoch': 0.3}) +[2026-10-03 19:46:15,959][starVLA.training.train_starvla][INFO] - Step 4144, Loss: {'train/loss': 3.4599499485921115e-05, 'train/grad_norm_pre_clip': 0.00957939401268959, 'timing/data': 0.00022156999330036342, 'timing/model': 0.3951217819994781, 'learning_rate/qwen_vl_interface': 1.1975742501022406e-06, 'learning_rate/action_model': 1.1975742501022405e-05, 'epoch': 0.3}) +[2026-10-03 19:46:16,863][starVLA.training.train_starvla][INFO] - Step 4145, Loss: {'train/loss': 3.1537358154309914e-05, 'train/grad_norm_pre_clip': 0.0057548354379832745, 'timing/data': 0.0003344290016684681, 'timing/model': 0.5938611330057029, 'learning_rate/qwen_vl_interface': 1.1959863527378569e-06, 'learning_rate/action_model': 1.1959863527378567e-05, 'epoch': 0.3}) +[2026-10-03 19:46:17,571][starVLA.training.train_starvla][INFO] - Step 4146, Loss: {'train/loss': 9.373610373586416e-05, 'train/grad_norm_pre_clip': 0.03491642326116562, 'timing/data': 0.00036179000744596124, 'timing/model': 0.3952234909811523, 'learning_rate/qwen_vl_interface': 1.1944001218253505e-06, 'learning_rate/action_model': 1.1944001218253503e-05, 'epoch': 0.3}) +[2026-10-03 19:46:18,277][starVLA.training.train_starvla][INFO] - Step 4147, Loss: {'train/loss': 0.34812313318252563, 'train/grad_norm_pre_clip': 76.4378890991211, 'timing/data': 0.000265370006673038, 'timing/model': 0.39538397299475037, 'learning_rate/qwen_vl_interface': 1.1928155580167608e-06, 'learning_rate/action_model': 1.1928155580167608e-05, 'epoch': 0.3}) +[2026-10-03 19:46:18,985][starVLA.training.train_starvla][INFO] - Step 4148, Loss: {'train/loss': 0.04593721032142639, 'train/grad_norm_pre_clip': 34.92633819580078, 'timing/data': 0.00021155999274924397, 'timing/model': 0.3959004320204258, 'learning_rate/qwen_vl_interface': 1.1912326619634412e-06, 'learning_rate/action_model': 1.1912326619634412e-05, 'epoch': 0.3}) +[2026-10-03 19:46:19,691][starVLA.training.train_starvla][INFO] - Step 4149, Loss: {'train/loss': 0.004471374209970236, 'train/grad_norm_pre_clip': 13.878578186035156, 'timing/data': 0.00022355001419782639, 'timing/model': 0.3944329720106907, 'learning_rate/qwen_vl_interface': 1.1896514343160632e-06, 'learning_rate/action_model': 1.189651434316063e-05, 'epoch': 0.3}) +[2026-10-03 19:46:20,600][starVLA.training.train_starvla][INFO] - Step 4150, Loss: {'train/loss': 0.00010085792746394873, 'train/grad_norm_pre_clip': 0.05197541043162346, 'timing/data': 0.00036330000148154795, 'timing/model': 0.5985707420040853, 'learning_rate/qwen_vl_interface': 1.1880718757246062e-06, 'learning_rate/action_model': 1.1880718757246063e-05, 'epoch': 0.3}) +[2026-10-03 19:46:21,309][starVLA.training.train_starvla][INFO] - Step 4151, Loss: {'train/loss': 0.000195069020264782, 'train/grad_norm_pre_clip': 0.0853644609451294, 'timing/data': 0.0003966399817727506, 'timing/model': 0.395663880975917, 'learning_rate/qwen_vl_interface': 1.186493986838369e-06, 'learning_rate/action_model': 1.186493986838369e-05, 'epoch': 0.3}) +[2026-10-03 19:46:22,017][starVLA.training.train_starvla][INFO] - Step 4152, Loss: {'train/loss': 0.0005537717952392995, 'train/grad_norm_pre_clip': 0.5450380444526672, 'timing/data': 0.0003281999961473048, 'timing/model': 0.395493711985182, 'learning_rate/qwen_vl_interface': 1.1849177683059612e-06, 'learning_rate/action_model': 1.1849177683059612e-05, 'epoch': 0.3}) +[2026-10-03 19:46:22,725][starVLA.training.train_starvla][INFO] - Step 4153, Loss: {'train/loss': 0.000189808604773134, 'train/grad_norm_pre_clip': 0.1497529000043869, 'timing/data': 0.00021853001089766622, 'timing/model': 0.395462841988774, 'learning_rate/qwen_vl_interface': 1.1833432207753075e-06, 'learning_rate/action_model': 1.1833432207753076e-05, 'epoch': 0.3}) +[2026-10-03 19:46:23,432][starVLA.training.train_starvla][INFO] - Step 4154, Loss: {'train/loss': 0.011279577389359474, 'train/grad_norm_pre_clip': 17.44499397277832, 'timing/data': 0.00025958000333048403, 'timing/model': 0.39487786099198274, 'learning_rate/qwen_vl_interface': 1.1817703448936448e-06, 'learning_rate/action_model': 1.181770344893645e-05, 'epoch': 0.3}) +[2026-10-03 19:46:24,343][starVLA.training.train_starvla][INFO] - Step 4155, Loss: {'train/loss': 0.07726816833019257, 'train/grad_norm_pre_clip': 58.41753387451172, 'timing/data': 0.0003766200097743422, 'timing/model': 0.599221572978422, 'learning_rate/qwen_vl_interface': 1.1801991413075228e-06, 'learning_rate/action_model': 1.1801991413075227e-05, 'epoch': 0.3}) +[2026-10-03 19:46:25,048][starVLA.training.train_starvla][INFO] - Step 4156, Loss: {'train/loss': 0.0001358309673378244, 'train/grad_norm_pre_clip': 0.042273540049791336, 'timing/data': 0.00021807997836731374, 'timing/model': 0.39435122098075226, 'learning_rate/qwen_vl_interface': 1.1786296106628038e-06, 'learning_rate/action_model': 1.1786296106628036e-05, 'epoch': 0.3}) +[2026-10-03 19:46:25,756][starVLA.training.train_starvla][INFO] - Step 4157, Loss: {'train/loss': 0.009027383290231228, 'train/grad_norm_pre_clip': 20.018478393554688, 'timing/data': 0.0002305000089108944, 'timing/model': 0.3966487619909458, 'learning_rate/qwen_vl_interface': 1.1770617536046633e-06, 'learning_rate/action_model': 1.1770617536046632e-05, 'epoch': 0.3}) +[2026-10-03 19:46:26,463][starVLA.training.train_starvla][INFO] - Step 4158, Loss: {'train/loss': 0.23050181567668915, 'train/grad_norm_pre_clip': 74.71896362304688, 'timing/data': 0.0003833499795291573, 'timing/model': 0.3964337310171686, 'learning_rate/qwen_vl_interface': 1.1754955707775873e-06, 'learning_rate/action_model': 1.1754955707775872e-05, 'epoch': 0.3}) +[2026-10-03 19:46:27,172][starVLA.training.train_starvla][INFO] - Step 4159, Loss: {'train/loss': 0.00016230129403993487, 'train/grad_norm_pre_clip': 0.0586317777633667, 'timing/data': 0.00022298999829217792, 'timing/model': 0.39540184201905504, 'learning_rate/qwen_vl_interface': 1.1739310628253753e-06, 'learning_rate/action_model': 1.1739310628253753e-05, 'epoch': 0.3}) +[2026-10-03 19:46:27,879][starVLA.training.train_starvla][INFO] - Step 4160, Loss: {'train/loss': 3.250528607168235e-05, 'train/grad_norm_pre_clip': 0.007588977459818125, 'timing/data': 0.0002016100042965263, 'timing/model': 0.39518187200883403, 'learning_rate/qwen_vl_interface': 1.172368230391136e-06, 'learning_rate/action_model': 1.1723682303911359e-05, 'epoch': 0.3}) +[2026-10-03 19:46:28,785][starVLA.training.train_starvla][INFO] - Step 4161, Loss: {'train/loss': 0.0004062006191816181, 'train/grad_norm_pre_clip': 0.2868254780769348, 'timing/data': 0.00034602999221533537, 'timing/model': 0.39428676199167967, 'learning_rate/qwen_vl_interface': 1.1708070741172928e-06, 'learning_rate/action_model': 1.1708070741172928e-05, 'epoch': 0.3}) +[2026-10-03 19:46:29,494][starVLA.training.train_starvla][INFO] - Step 4162, Loss: {'train/loss': 0.6143734455108643, 'train/grad_norm_pre_clip': 123.28486633300781, 'timing/data': 0.0003118599997833371, 'timing/model': 0.39648737202514894, 'learning_rate/qwen_vl_interface': 1.1692475946455755e-06, 'learning_rate/action_model': 1.1692475946455755e-05, 'epoch': 0.3}) +[2026-10-03 19:46:30,200][starVLA.training.train_starvla][INFO] - Step 4163, Loss: {'train/loss': 0.0009211665019392967, 'train/grad_norm_pre_clip': 0.44831642508506775, 'timing/data': 0.00024354999186471105, 'timing/model': 0.39414332201704383, 'learning_rate/qwen_vl_interface': 1.1676897926170289e-06, 'learning_rate/action_model': 1.1676897926170289e-05, 'epoch': 0.3}) +[2026-10-03 19:46:30,904][starVLA.training.train_starvla][INFO] - Step 4164, Loss: {'train/loss': 0.00028840621234849095, 'train/grad_norm_pre_clip': 0.10244902223348618, 'timing/data': 0.00020797998877242208, 'timing/model': 0.3942412409815006, 'learning_rate/qwen_vl_interface': 1.166133668672006e-06, 'learning_rate/action_model': 1.166133668672006e-05, 'epoch': 0.3}) +[2026-10-03 19:46:31,612][starVLA.training.train_starvla][INFO] - Step 4165, Loss: {'train/loss': 0.018499379977583885, 'train/grad_norm_pre_clip': 23.009489059448242, 'timing/data': 0.0002932900097221136, 'timing/model': 0.39578990198788233, 'learning_rate/qwen_vl_interface': 1.1645792234501706e-06, 'learning_rate/action_model': 1.1645792234501706e-05, 'epoch': 0.3}) +[2026-10-03 19:46:32,522][starVLA.training.train_starvla][INFO] - Step 4166, Loss: {'train/loss': 0.3442886769771576, 'train/grad_norm_pre_clip': 37.7684211730957, 'timing/data': 0.00029406900284811854, 'timing/model': 0.3944792319962289, 'learning_rate/qwen_vl_interface': 1.1630264575904972e-06, 'learning_rate/action_model': 1.1630264575904973e-05, 'epoch': 0.3}) +[2026-10-03 19:46:33,229][starVLA.training.train_starvla][INFO] - Step 4167, Loss: {'train/loss': 0.003249971428886056, 'train/grad_norm_pre_clip': 4.921529293060303, 'timing/data': 0.0002330999996047467, 'timing/model': 0.39448904199525714, 'learning_rate/qwen_vl_interface': 1.1614753717312683e-06, 'learning_rate/action_model': 1.1614753717312681e-05, 'epoch': 0.3}) +[2026-10-03 19:46:33,936][starVLA.training.train_starvla][INFO] - Step 4168, Loss: {'train/loss': 0.00016394587873946875, 'train/grad_norm_pre_clip': 0.06301615387201309, 'timing/data': 0.0002435100032016635, 'timing/model': 0.39490383098018356, 'learning_rate/qwen_vl_interface': 1.1599259665100767e-06, 'learning_rate/action_model': 1.1599259665100767e-05, 'epoch': 0.31}) +[2026-10-03 19:46:34,642][starVLA.training.train_starvla][INFO] - Step 4169, Loss: {'train/loss': 0.19346807897090912, 'train/grad_norm_pre_clip': 133.30551147460938, 'timing/data': 0.00023540001711808145, 'timing/model': 0.3962971919972915, 'learning_rate/qwen_vl_interface': 1.1583782425638252e-06, 'learning_rate/action_model': 1.158378242563825e-05, 'epoch': 0.31}) +[2026-10-03 19:46:35,348][starVLA.training.train_starvla][INFO] - Step 4170, Loss: {'train/loss': 0.00043777143582701683, 'train/grad_norm_pre_clip': 0.4597432017326355, 'timing/data': 0.0002685899962671101, 'timing/model': 0.3957701420004014, 'learning_rate/qwen_vl_interface': 1.1568322005287237e-06, 'learning_rate/action_model': 1.1568322005287235e-05, 'epoch': 0.31}) +[2026-10-03 19:46:36,256][starVLA.training.train_starvla][INFO] - Step 4171, Loss: {'train/loss': 0.06243448704481125, 'train/grad_norm_pre_clip': 80.19255828857422, 'timing/data': 0.00023033999605104327, 'timing/model': 0.39370884202071466, 'learning_rate/qwen_vl_interface': 1.1552878410402923e-06, 'learning_rate/action_model': 1.1552878410402924e-05, 'epoch': 0.31}) +[2026-10-03 19:46:36,964][starVLA.training.train_starvla][INFO] - Step 4172, Loss: {'train/loss': 0.0027592896949499846, 'train/grad_norm_pre_clip': 11.838150024414062, 'timing/data': 0.0002262399939354509, 'timing/model': 0.3954945419973228, 'learning_rate/qwen_vl_interface': 1.1537451647333594e-06, 'learning_rate/action_model': 1.1537451647333593e-05, 'epoch': 0.31}) +[2026-10-03 19:46:37,670][starVLA.training.train_starvla][INFO] - Step 4173, Loss: {'train/loss': 0.00039763402310200036, 'train/grad_norm_pre_clip': 0.24647189676761627, 'timing/data': 0.00022440002067014575, 'timing/model': 0.3941231430217158, 'learning_rate/qwen_vl_interface': 1.1522041722420594e-06, 'learning_rate/action_model': 1.1522041722420594e-05, 'epoch': 0.31}) +[2026-10-03 19:46:38,375][starVLA.training.train_starvla][INFO] - Step 4174, Loss: {'train/loss': 0.0655330941081047, 'train/grad_norm_pre_clip': 58.57701873779297, 'timing/data': 0.0002693599963095039, 'timing/model': 0.3947836020088289, 'learning_rate/qwen_vl_interface': 1.1506648641998365e-06, 'learning_rate/action_model': 1.1506648641998365e-05, 'epoch': 0.31}) +[2026-10-03 19:46:39,083][starVLA.training.train_starvla][INFO] - Step 4175, Loss: {'train/loss': 3.898753857356496e-05, 'train/grad_norm_pre_clip': 0.0076492950320243835, 'timing/data': 0.0003225899999961257, 'timing/model': 0.39524028199957684, 'learning_rate/qwen_vl_interface': 1.1491272412394433e-06, 'learning_rate/action_model': 1.1491272412394433e-05, 'epoch': 0.31}) +[2026-10-03 19:46:39,998][starVLA.training.train_starvla][INFO] - Step 4176, Loss: {'train/loss': 0.1585828959941864, 'train/grad_norm_pre_clip': 51.12998962402344, 'timing/data': 0.0002629400114528835, 'timing/model': 0.6035470420029014, 'learning_rate/qwen_vl_interface': 1.1475913039929373e-06, 'learning_rate/action_model': 1.1475913039929372e-05, 'epoch': 0.31}) +[2026-10-03 19:46:40,677][starVLA.training.train_starvla][INFO] - Step 4177, Loss: {'train/loss': 0.0002616338315419853, 'train/grad_norm_pre_clip': 0.14619435369968414, 'timing/data': 0.00030991999665275216, 'timing/model': 0.3697319939965382, 'learning_rate/qwen_vl_interface': 1.1460570530916844e-06, 'learning_rate/action_model': 1.1460570530916844e-05, 'epoch': 0.31}) +[2026-10-03 19:46:41,386][starVLA.training.train_starvla][INFO] - Step 4178, Loss: {'train/loss': 0.0008334409212693572, 'train/grad_norm_pre_clip': 2.6128180027008057, 'timing/data': 0.0002730900014284998, 'timing/model': 0.39746743100113235, 'learning_rate/qwen_vl_interface': 1.1445244891663581e-06, 'learning_rate/action_model': 1.1445244891663581e-05, 'epoch': 0.31}) +[2026-10-03 19:46:42,091][starVLA.training.train_starvla][INFO] - Step 4179, Loss: {'train/loss': 4.148282459937036e-05, 'train/grad_norm_pre_clip': 0.010469292290508747, 'timing/data': 0.00020502001279965043, 'timing/model': 0.39552195198484696, 'learning_rate/qwen_vl_interface': 1.1429936128469366e-06, 'learning_rate/action_model': 1.1429936128469366e-05, 'epoch': 0.31}) +[2026-10-03 19:46:42,797][starVLA.training.train_starvla][INFO] - Step 4180, Loss: {'train/loss': 0.6530698537826538, 'train/grad_norm_pre_clip': 22.091941833496094, 'timing/data': 0.00021754001500084996, 'timing/model': 0.39551132201449946, 'learning_rate/qwen_vl_interface': 1.1414644247627056e-06, 'learning_rate/action_model': 1.1414644247627058e-05, 'epoch': 0.31}) +[2026-10-03 19:46:43,504][starVLA.training.train_starvla][INFO] - Step 4181, Loss: {'train/loss': 0.0004919534549117088, 'train/grad_norm_pre_clip': 0.4089203476905823, 'timing/data': 0.0003117599990218878, 'timing/model': 0.3941072419984266, 'learning_rate/qwen_vl_interface': 1.1399369255422568e-06, 'learning_rate/action_model': 1.1399369255422568e-05, 'epoch': 0.31}) +[2026-10-03 19:46:44,416][starVLA.training.train_starvla][INFO] - Step 4182, Loss: {'train/loss': 0.01268666610121727, 'train/grad_norm_pre_clip': 25.948467254638672, 'timing/data': 0.00031581000075675547, 'timing/model': 0.3956233409990091, 'learning_rate/qwen_vl_interface': 1.138411115813488e-06, 'learning_rate/action_model': 1.138411115813488e-05, 'epoch': 0.31}) +[2026-10-03 19:46:45,124][starVLA.training.train_starvla][INFO] - Step 4183, Loss: {'train/loss': 0.42385444045066833, 'train/grad_norm_pre_clip': 50.554656982421875, 'timing/data': 0.0002248099772259593, 'timing/model': 0.39580837200628594, 'learning_rate/qwen_vl_interface': 1.1368869962036006e-06, 'learning_rate/action_model': 1.1368869962036005e-05, 'epoch': 0.31}) +[2026-10-03 19:46:45,830][starVLA.training.train_starvla][INFO] - Step 4184, Loss: {'train/loss': 0.0006209457642398775, 'train/grad_norm_pre_clip': 0.7380477786064148, 'timing/data': 0.0003172200231347233, 'timing/model': 0.3955133320123423, 'learning_rate/qwen_vl_interface': 1.1353645673391025e-06, 'learning_rate/action_model': 1.1353645673391026e-05, 'epoch': 0.31}) +[2026-10-03 19:46:46,536][starVLA.training.train_starvla][INFO] - Step 4185, Loss: {'train/loss': 5.9538557252381e-05, 'train/grad_norm_pre_clip': 0.01918753795325756, 'timing/data': 0.00022578000789508224, 'timing/model': 0.3947463210206479, 'learning_rate/qwen_vl_interface': 1.1338438298458078e-06, 'learning_rate/action_model': 1.1338438298458078e-05, 'epoch': 0.31}) +[2026-10-03 19:46:47,244][starVLA.training.train_starvla][INFO] - Step 4186, Loss: {'train/loss': 0.16245988011360168, 'train/grad_norm_pre_clip': 167.32041931152344, 'timing/data': 0.0002476800000295043, 'timing/model': 0.39575561197125353, 'learning_rate/qwen_vl_interface': 1.1323247843488345e-06, 'learning_rate/action_model': 1.1323247843488344e-05, 'epoch': 0.31}) +[2026-10-03 19:46:48,157][starVLA.training.train_starvla][INFO] - Step 4187, Loss: {'train/loss': 0.00012689660070464015, 'train/grad_norm_pre_clip': 0.037752415984869, 'timing/data': 0.00032933001057244837, 'timing/model': 0.6001161919848528, 'learning_rate/qwen_vl_interface': 1.130807431472603e-06, 'learning_rate/action_model': 1.130807431472603e-05, 'epoch': 0.31}) +[2026-10-03 19:46:48,863][starVLA.training.train_starvla][INFO] - Step 4188, Loss: {'train/loss': 0.7066912651062012, 'train/grad_norm_pre_clip': 111.7477798461914, 'timing/data': 0.00025296001695096493, 'timing/model': 0.3954662120086141, 'learning_rate/qwen_vl_interface': 1.1292917718408416e-06, 'learning_rate/action_model': 1.1292917718408416e-05, 'epoch': 0.31}) +[2026-10-03 19:46:49,567][starVLA.training.train_starvla][INFO] - Step 4189, Loss: {'train/loss': 0.0005164516624063253, 'train/grad_norm_pre_clip': 0.6044593453407288, 'timing/data': 0.0002492099883966148, 'timing/model': 0.39467121099005453, 'learning_rate/qwen_vl_interface': 1.12777780607658e-06, 'learning_rate/action_model': 1.12777780607658e-05, 'epoch': 0.31}) +[2026-10-03 19:46:50,275][starVLA.training.train_starvla][INFO] - Step 4190, Loss: {'train/loss': 0.00012775062350556254, 'train/grad_norm_pre_clip': 0.049078624695539474, 'timing/data': 0.00022211999748833477, 'timing/model': 0.3959362919849809, 'learning_rate/qwen_vl_interface': 1.1262655348021523e-06, 'learning_rate/action_model': 1.1262655348021523e-05, 'epoch': 0.31}) +[2026-10-03 19:46:50,979][starVLA.training.train_starvla][INFO] - Step 4191, Loss: {'train/loss': 0.8455135822296143, 'train/grad_norm_pre_clip': 225.81687927246094, 'timing/data': 0.0002274000144097954, 'timing/model': 0.3949981620244216, 'learning_rate/qwen_vl_interface': 1.1247549586391966e-06, 'learning_rate/action_model': 1.1247549586391964e-05, 'epoch': 0.31}) +[2026-10-03 19:46:51,873][starVLA.training.train_starvla][INFO] - Step 4192, Loss: {'train/loss': 0.27563953399658203, 'train/grad_norm_pre_clip': 159.87046813964844, 'timing/data': 0.00032040997757576406, 'timing/model': 0.5850768329983111, 'learning_rate/qwen_vl_interface': 1.1232460782086534e-06, 'learning_rate/action_model': 1.1232460782086534e-05, 'epoch': 0.31}) +[2026-10-03 19:46:52,580][starVLA.training.train_starvla][INFO] - Step 4193, Loss: {'train/loss': 9.987100202124566e-05, 'train/grad_norm_pre_clip': 0.032077495008707047, 'timing/data': 0.00022475002333521843, 'timing/model': 0.3951253220147919, 'learning_rate/qwen_vl_interface': 1.1217388941307658e-06, 'learning_rate/action_model': 1.1217388941307658e-05, 'epoch': 0.31}) +[2026-10-03 19:46:53,283][starVLA.training.train_starvla][INFO] - Step 4194, Loss: {'train/loss': 0.00010485603706911206, 'train/grad_norm_pre_clip': 0.04494999721646309, 'timing/data': 0.00022031000116840005, 'timing/model': 0.3940058120060712, 'learning_rate/qwen_vl_interface': 1.1202334070250807e-06, 'learning_rate/action_model': 1.1202334070250806e-05, 'epoch': 0.31}) +[2026-10-03 19:46:53,990][starVLA.training.train_starvla][INFO] - Step 4195, Loss: {'train/loss': 0.993800163269043, 'train/grad_norm_pre_clip': 143.31117248535156, 'timing/data': 0.00021413000649772584, 'timing/model': 0.39420525199966505, 'learning_rate/qwen_vl_interface': 1.1187296175104472e-06, 'learning_rate/action_model': 1.1187296175104471e-05, 'epoch': 0.31}) +[2026-10-03 19:46:54,698][starVLA.training.train_starvla][INFO] - Step 4196, Loss: {'train/loss': 0.04193885624408722, 'train/grad_norm_pre_clip': 61.90269088745117, 'timing/data': 0.0003360499977134168, 'timing/model': 0.395515420997981, 'learning_rate/qwen_vl_interface': 1.1172275262050158e-06, 'learning_rate/action_model': 1.1172275262050159e-05, 'epoch': 0.31}) +[2026-10-03 19:46:55,405][starVLA.training.train_starvla][INFO] - Step 4197, Loss: {'train/loss': 0.24829591810703278, 'train/grad_norm_pre_clip': 110.77632904052734, 'timing/data': 0.000354969990439713, 'timing/model': 0.3955315920175053, 'learning_rate/qwen_vl_interface': 1.1157271337262401e-06, 'learning_rate/action_model': 1.11572713372624e-05, 'epoch': 0.31}) +[2026-10-03 19:46:56,316][starVLA.training.train_starvla][INFO] - Step 4198, Loss: {'train/loss': 0.0014506488805636764, 'train/grad_norm_pre_clip': 1.2480956315994263, 'timing/data': 0.0002282799978274852, 'timing/model': 0.3968987010011915, 'learning_rate/qwen_vl_interface': 1.1142284406908751e-06, 'learning_rate/action_model': 1.114228440690875e-05, 'epoch': 0.31}) +[2026-10-03 19:46:57,025][starVLA.training.train_starvla][INFO] - Step 4199, Loss: {'train/loss': 0.0001493009622208774, 'train/grad_norm_pre_clip': 0.05498502776026726, 'timing/data': 0.0002210899838246405, 'timing/model': 0.3956168729928322, 'learning_rate/qwen_vl_interface': 1.1127314477149758e-06, 'learning_rate/action_model': 1.1127314477149759e-05, 'epoch': 0.31}) +[2026-10-03 19:46:57,731][starVLA.training.train_starvla][INFO] - Step 4200, Loss: {'train/loss': 0.02691328153014183, 'train/grad_norm_pre_clip': 31.546842575073242, 'timing/data': 0.0003435199905652553, 'timing/model': 0.3954310319968499, 'learning_rate/qwen_vl_interface': 1.1112361554139003e-06, 'learning_rate/action_model': 1.1112361554139004e-05, 'epoch': 0.31}) +[2026-10-03 19:46:58,439][starVLA.training.train_starvla][INFO] - Step 4201, Loss: {'train/loss': 0.45978495478630066, 'train/grad_norm_pre_clip': 34.235069274902344, 'timing/data': 0.0003258200013078749, 'timing/model': 0.3950012710120063, 'learning_rate/qwen_vl_interface': 1.1097425644023065e-06, 'learning_rate/action_model': 1.1097425644023065e-05, 'epoch': 0.31}) +[2026-10-03 19:46:59,148][starVLA.training.train_starvla][INFO] - Step 4202, Loss: {'train/loss': 7.737967098364606e-05, 'train/grad_norm_pre_clip': 0.033506814390420914, 'timing/data': 0.00026033000904135406, 'timing/model': 0.3963222620077431, 'learning_rate/qwen_vl_interface': 1.1082506752941529e-06, 'learning_rate/action_model': 1.1082506752941528e-05, 'epoch': 0.31}) +[2026-10-03 19:47:00,054][starVLA.training.train_starvla][INFO] - Step 4203, Loss: {'train/loss': 0.28464365005493164, 'train/grad_norm_pre_clip': 49.5576171875, 'timing/data': 0.00022593000903725624, 'timing/model': 0.3947210320038721, 'learning_rate/qwen_vl_interface': 1.1067604887027004e-06, 'learning_rate/action_model': 1.1067604887027004e-05, 'epoch': 0.31}) +[2026-10-03 19:47:00,760][starVLA.training.train_starvla][INFO] - Step 4204, Loss: {'train/loss': 0.00021036938414908946, 'train/grad_norm_pre_clip': 0.13622748851776123, 'timing/data': 0.00032377001480199397, 'timing/model': 0.39518536199466325, 'learning_rate/qwen_vl_interface': 1.1052720052405066e-06, 'learning_rate/action_model': 1.1052720052405064e-05, 'epoch': 0.31}) +[2026-10-03 19:47:01,467][starVLA.training.train_starvla][INFO] - Step 4205, Loss: {'train/loss': 0.0005446205614134669, 'train/grad_norm_pre_clip': 0.5612907409667969, 'timing/data': 0.00035162997664883733, 'timing/model': 0.39415750198531896, 'learning_rate/qwen_vl_interface': 1.1037852255194309e-06, 'learning_rate/action_model': 1.1037852255194308e-05, 'epoch': 0.31}) +[2026-10-03 19:47:02,173][starVLA.training.train_starvla][INFO] - Step 4206, Loss: {'train/loss': 5.566063555306755e-05, 'train/grad_norm_pre_clip': 0.01187640056014061, 'timing/data': 0.00031723000574856997, 'timing/model': 0.394251231016824, 'learning_rate/qwen_vl_interface': 1.1023001501506329e-06, 'learning_rate/action_model': 1.102300150150633e-05, 'epoch': 0.31}) +[2026-10-03 19:47:02,880][starVLA.training.train_starvla][INFO] - Step 4207, Loss: {'train/loss': 0.004163403995335102, 'train/grad_norm_pre_clip': 5.792568206787109, 'timing/data': 0.0002022900152951479, 'timing/model': 0.3951948820031248, 'learning_rate/qwen_vl_interface': 1.1008167797445717e-06, 'learning_rate/action_model': 1.1008167797445718e-05, 'epoch': 0.31}) +[2026-10-03 19:47:03,790][starVLA.training.train_starvla][INFO] - Step 4208, Loss: {'train/loss': 0.0006949875387363136, 'train/grad_norm_pre_clip': 0.9936733841896057, 'timing/data': 0.00034110998967662454, 'timing/model': 0.59580466200714, 'learning_rate/qwen_vl_interface': 1.099335114911003e-06, 'learning_rate/action_model': 1.0993351149110031e-05, 'epoch': 0.31}) +[2026-10-03 19:47:04,498][starVLA.training.train_starvla][INFO] - Step 4209, Loss: {'train/loss': 0.00015841939602978528, 'train/grad_norm_pre_clip': 0.05049072578549385, 'timing/data': 0.00026316000730730593, 'timing/model': 0.3947364519990515, 'learning_rate/qwen_vl_interface': 1.0978551562589846e-06, 'learning_rate/action_model': 1.0978551562589845e-05, 'epoch': 0.31}) +[2026-10-03 19:47:05,206][starVLA.training.train_starvla][INFO] - Step 4210, Loss: {'train/loss': 0.00015483511378988624, 'train/grad_norm_pre_clip': 0.07308708876371384, 'timing/data': 0.0003118799941148609, 'timing/model': 0.3972185620223172, 'learning_rate/qwen_vl_interface': 1.0963769043968708e-06, 'learning_rate/action_model': 1.0963769043968708e-05, 'epoch': 0.31}) +[2026-10-03 19:47:05,914][starVLA.training.train_starvla][INFO] - Step 4211, Loss: {'train/loss': 0.23307569324970245, 'train/grad_norm_pre_clip': 70.51251220703125, 'timing/data': 0.00033799000084400177, 'timing/model': 0.3945254419813864, 'learning_rate/qwen_vl_interface': 1.0949003599323158e-06, 'learning_rate/action_model': 1.0949003599323157e-05, 'epoch': 0.31}) +[2026-10-03 19:47:06,623][starVLA.training.train_starvla][INFO] - Step 4212, Loss: {'train/loss': 6.547120574396104e-05, 'train/grad_norm_pre_clip': 0.01105770468711853, 'timing/data': 0.0003478500002529472, 'timing/model': 0.3966769620019477, 'learning_rate/qwen_vl_interface': 1.0934255234722708e-06, 'learning_rate/action_model': 1.0934255234722706e-05, 'epoch': 0.31}) +[2026-10-03 19:47:07,552][starVLA.training.train_starvla][INFO] - Step 4213, Loss: {'train/loss': 0.0822291374206543, 'train/grad_norm_pre_clip': 95.671875, 'timing/data': 0.00022621999960392714, 'timing/model': 0.6160705299989786, 'learning_rate/qwen_vl_interface': 1.0919523956229858e-06, 'learning_rate/action_model': 1.0919523956229857e-05, 'epoch': 0.31}) +[2026-10-03 19:47:08,262][starVLA.training.train_starvla][INFO] - Step 4214, Loss: {'train/loss': 0.0001228028122568503, 'train/grad_norm_pre_clip': 0.044096093624830246, 'timing/data': 0.00032255001133307815, 'timing/model': 0.39765992100001313, 'learning_rate/qwen_vl_interface': 1.0904809769900075e-06, 'learning_rate/action_model': 1.0904809769900073e-05, 'epoch': 0.31}) +[2026-10-03 19:47:08,967][starVLA.training.train_starvla][INFO] - Step 4215, Loss: {'train/loss': 0.00020981044508516788, 'train/grad_norm_pre_clip': 0.15845055878162384, 'timing/data': 0.0002283100038766861, 'timing/model': 0.3939829019946046, 'learning_rate/qwen_vl_interface': 1.0890112681781821e-06, 'learning_rate/action_model': 1.089011268178182e-05, 'epoch': 0.31}) +[2026-10-03 19:47:09,673][starVLA.training.train_starvla][INFO] - Step 4216, Loss: {'train/loss': 0.42248329520225525, 'train/grad_norm_pre_clip': 48.02216720581055, 'timing/data': 0.00030058002448640764, 'timing/model': 0.3956544219981879, 'learning_rate/qwen_vl_interface': 1.0875432697916492e-06, 'learning_rate/action_model': 1.087543269791649e-05, 'epoch': 0.31}) +[2026-10-03 19:47:10,380][starVLA.training.train_starvla][INFO] - Step 4217, Loss: {'train/loss': 0.0001835281727835536, 'train/grad_norm_pre_clip': 0.14644105732440948, 'timing/data': 0.00035497997305355966, 'timing/model': 0.39528494101250544, 'learning_rate/qwen_vl_interface': 1.0860769824338477e-06, 'learning_rate/action_model': 1.0860769824338478e-05, 'epoch': 0.31}) +[2026-10-03 19:47:11,089][starVLA.training.train_starvla][INFO] - Step 4218, Loss: {'train/loss': 0.7607516050338745, 'train/grad_norm_pre_clip': 104.65227508544922, 'timing/data': 0.00029989000177010894, 'timing/model': 0.39616417200886644, 'learning_rate/qwen_vl_interface': 1.084612406707514e-06, 'learning_rate/action_model': 1.084612406707514e-05, 'epoch': 0.31}) +[2026-10-03 19:47:11,997][starVLA.training.train_starvla][INFO] - Step 4219, Loss: {'train/loss': 0.07800821214914322, 'train/grad_norm_pre_clip': 71.59477233886719, 'timing/data': 0.00023010000586509705, 'timing/model': 0.3940620319917798, 'learning_rate/qwen_vl_interface': 1.0831495432146795e-06, 'learning_rate/action_model': 1.0831495432146795e-05, 'epoch': 0.31}) +[2026-10-03 19:47:12,703][starVLA.training.train_starvla][INFO] - Step 4220, Loss: {'train/loss': 0.00014864577678963542, 'train/grad_norm_pre_clip': 0.07320544868707657, 'timing/data': 0.00034470000537112355, 'timing/model': 0.39439768201555125, 'learning_rate/qwen_vl_interface': 1.0816883925566717e-06, 'learning_rate/action_model': 1.0816883925566716e-05, 'epoch': 0.31}) +[2026-10-03 19:47:13,412][starVLA.training.train_starvla][INFO] - Step 4221, Loss: {'train/loss': 0.30404573678970337, 'train/grad_norm_pre_clip': 45.793697357177734, 'timing/data': 0.0002624299959279597, 'timing/model': 0.39665892100310884, 'learning_rate/qwen_vl_interface': 1.0802289553341153e-06, 'learning_rate/action_model': 1.0802289553341152e-05, 'epoch': 0.31}) +[2026-10-03 19:47:14,119][starVLA.training.train_starvla][INFO] - Step 4222, Loss: {'train/loss': 0.35905805230140686, 'train/grad_norm_pre_clip': 46.8476448059082, 'timing/data': 0.00027339000371284783, 'timing/model': 0.3966749120154418, 'learning_rate/qwen_vl_interface': 1.0787712321469289e-06, 'learning_rate/action_model': 1.0787712321469288e-05, 'epoch': 0.31}) +[2026-10-03 19:47:14,828][starVLA.training.train_starvla][INFO] - Step 4223, Loss: {'train/loss': 0.00019259366672486067, 'train/grad_norm_pre_clip': 0.08879218250513077, 'timing/data': 0.00033650000113993883, 'timing/model': 0.3959229319880251, 'learning_rate/qwen_vl_interface': 1.077315223594328e-06, 'learning_rate/action_model': 1.077315223594328e-05, 'epoch': 0.31}) +[2026-10-03 19:47:15,740][starVLA.training.train_starvla][INFO] - Step 4224, Loss: {'train/loss': 0.0003021692391484976, 'train/grad_norm_pre_clip': 0.26013222336769104, 'timing/data': 0.00023297997540794313, 'timing/model': 0.5996424820041284, 'learning_rate/qwen_vl_interface': 1.0758609302748227e-06, 'learning_rate/action_model': 1.0758609302748227e-05, 'epoch': 0.31}) +[2026-10-03 19:47:16,450][starVLA.training.train_starvla][INFO] - Step 4225, Loss: {'train/loss': 0.0009622220532037318, 'train/grad_norm_pre_clip': 0.7976402044296265, 'timing/data': 0.0002770799910649657, 'timing/model': 0.3979078709962778, 'learning_rate/qwen_vl_interface': 1.0744083527862182e-06, 'learning_rate/action_model': 1.074408352786218e-05, 'epoch': 0.31}) +[2026-10-03 19:47:17,159][starVLA.training.train_starvla][INFO] - Step 4226, Loss: {'train/loss': 0.0020972685888409615, 'train/grad_norm_pre_clip': 1.4835662841796875, 'timing/data': 0.00027735999901778996, 'timing/model': 0.39526647102320567, 'learning_rate/qwen_vl_interface': 1.0729574917256144e-06, 'learning_rate/action_model': 1.0729574917256145e-05, 'epoch': 0.31}) +[2026-10-03 19:47:17,868][starVLA.training.train_starvla][INFO] - Step 4227, Loss: {'train/loss': 0.00022474885918200016, 'train/grad_norm_pre_clip': 0.1304941326379776, 'timing/data': 0.0002497799869161099, 'timing/model': 0.3957161919970531, 'learning_rate/qwen_vl_interface': 1.0715083476894047e-06, 'learning_rate/action_model': 1.0715083476894047e-05, 'epoch': 0.31}) +[2026-10-03 19:47:18,575][starVLA.training.train_starvla][INFO] - Step 4228, Loss: {'train/loss': 1.2362362146377563, 'train/grad_norm_pre_clip': 116.89289855957031, 'timing/data': 0.00022792001254856586, 'timing/model': 0.39415579100023024, 'learning_rate/qwen_vl_interface': 1.0700609212732773e-06, 'learning_rate/action_model': 1.0700609212732774e-05, 'epoch': 0.31}) +[2026-10-03 19:47:19,488][starVLA.training.train_starvla][INFO] - Step 4229, Loss: {'train/loss': 0.2566660940647125, 'train/grad_norm_pre_clip': 201.37548828125, 'timing/data': 0.0003060600138269365, 'timing/model': 0.6004706519888714, 'learning_rate/qwen_vl_interface': 1.0686152130722155e-06, 'learning_rate/action_model': 1.0686152130722154e-05, 'epoch': 0.31}) +[2026-10-03 19:47:20,196][starVLA.training.train_starvla][INFO] - Step 4230, Loss: {'train/loss': 0.001127969822846353, 'train/grad_norm_pre_clip': 1.2563518285751343, 'timing/data': 0.00036597001599147916, 'timing/model': 0.3954265420034062, 'learning_rate/qwen_vl_interface': 1.0671712236804947e-06, 'learning_rate/action_model': 1.0671712236804946e-05, 'epoch': 0.31}) +[2026-10-03 19:47:20,904][starVLA.training.train_starvla][INFO] - Step 4231, Loss: {'train/loss': 9.137474989984185e-05, 'train/grad_norm_pre_clip': 0.03903914988040924, 'timing/data': 0.0003368290199432522, 'timing/model': 0.39574727197759785, 'learning_rate/qwen_vl_interface': 1.0657289536916845e-06, 'learning_rate/action_model': 1.0657289536916845e-05, 'epoch': 0.31}) +[2026-10-03 19:47:21,612][starVLA.training.train_starvla][INFO] - Step 4232, Loss: {'train/loss': 0.016513532027602196, 'train/grad_norm_pre_clip': 45.66428756713867, 'timing/data': 0.00025974001619033515, 'timing/model': 0.39526606199797243, 'learning_rate/qwen_vl_interface': 1.0642884036986481e-06, 'learning_rate/action_model': 1.064288403698648e-05, 'epoch': 0.31}) +[2026-10-03 19:47:22,319][starVLA.training.train_starvla][INFO] - Step 4233, Loss: {'train/loss': 0.012354380451142788, 'train/grad_norm_pre_clip': 22.619918823242188, 'timing/data': 0.00021535000996664166, 'timing/model': 0.39533261200995184, 'learning_rate/qwen_vl_interface': 1.0628495742935404e-06, 'learning_rate/action_model': 1.0628495742935405e-05, 'epoch': 0.31}) +[2026-10-03 19:47:23,027][starVLA.training.train_starvla][INFO] - Step 4234, Loss: {'train/loss': 0.35361921787261963, 'train/grad_norm_pre_clip': 52.90790557861328, 'timing/data': 0.00031981998472474515, 'timing/model': 0.3965753710072022, 'learning_rate/qwen_vl_interface': 1.0614124660678103e-06, 'learning_rate/action_model': 1.0614124660678102e-05, 'epoch': 0.31}) +[2026-10-03 19:47:23,933][starVLA.training.train_starvla][INFO] - Step 4235, Loss: {'train/loss': 0.00020412463345564902, 'train/grad_norm_pre_clip': 0.07770030945539474, 'timing/data': 0.0003423099988140166, 'timing/model': 0.39463685199734755, 'learning_rate/qwen_vl_interface': 1.0599770796121988e-06, 'learning_rate/action_model': 1.0599770796121987e-05, 'epoch': 0.31}) +[2026-10-03 19:47:24,639][starVLA.training.train_starvla][INFO] - Step 4236, Loss: {'train/loss': 0.25436654686927795, 'train/grad_norm_pre_clip': 68.58897399902344, 'timing/data': 0.0003493999829515815, 'timing/model': 0.39513928099768236, 'learning_rate/qwen_vl_interface': 1.0585434155167394e-06, 'learning_rate/action_model': 1.0585434155167393e-05, 'epoch': 0.31}) +[2026-10-03 19:47:25,345][starVLA.training.train_starvla][INFO] - Step 4237, Loss: {'train/loss': 8.050879841903225e-05, 'train/grad_norm_pre_clip': 0.021973000839352608, 'timing/data': 0.00030278999474830925, 'timing/model': 0.3942866120196413, 'learning_rate/qwen_vl_interface': 1.0571114743707556e-06, 'learning_rate/action_model': 1.0571114743707555e-05, 'epoch': 0.31}) +[2026-10-03 19:47:26,051][starVLA.training.train_starvla][INFO] - Step 4238, Loss: {'train/loss': 4.455997986951843e-05, 'train/grad_norm_pre_clip': 0.008051536045968533, 'timing/data': 0.00030563000473193824, 'timing/model': 0.39492430197424255, 'learning_rate/qwen_vl_interface': 1.0556812567628661e-06, 'learning_rate/action_model': 1.055681256762866e-05, 'epoch': 0.31}) +[2026-10-03 19:47:26,758][starVLA.training.train_starvla][INFO] - Step 4239, Loss: {'train/loss': 8.68720089783892e-05, 'train/grad_norm_pre_clip': 0.02463417500257492, 'timing/data': 0.00033255902235396206, 'timing/model': 0.39600315201096237, 'learning_rate/qwen_vl_interface': 1.0542527632809783e-06, 'learning_rate/action_model': 1.0542527632809784e-05, 'epoch': 0.31}) +[2026-10-03 19:47:27,663][starVLA.training.train_starvla][INFO] - Step 4240, Loss: {'train/loss': 0.00015971205721143633, 'train/grad_norm_pre_clip': 0.05554009974002838, 'timing/data': 0.00031783999293111265, 'timing/model': 0.39376474198070355, 'learning_rate/qwen_vl_interface': 1.0528259945122925e-06, 'learning_rate/action_model': 1.0528259945122925e-05, 'epoch': 0.31}) +[2026-10-03 19:47:28,370][starVLA.training.train_starvla][INFO] - Step 4241, Loss: {'train/loss': 0.000993665773421526, 'train/grad_norm_pre_clip': 0.9119192361831665, 'timing/data': 0.0002187209902331233, 'timing/model': 0.39455716201337054, 'learning_rate/qwen_vl_interface': 1.0514009510432988e-06, 'learning_rate/action_model': 1.0514009510432988e-05, 'epoch': 0.31}) +[2026-10-03 19:47:29,080][starVLA.training.train_starvla][INFO] - Step 4242, Loss: {'train/loss': 0.022655384615063667, 'train/grad_norm_pre_clip': 45.2072639465332, 'timing/data': 0.00028457000735215843, 'timing/model': 0.3973262020153925, 'learning_rate/qwen_vl_interface': 1.0499776334597796e-06, 'learning_rate/action_model': 1.0499776334597797e-05, 'epoch': 0.31}) +[2026-10-03 19:47:29,788][starVLA.training.train_starvla][INFO] - Step 4243, Loss: {'train/loss': 0.0005086688906885684, 'train/grad_norm_pre_clip': 0.6354902982711792, 'timing/data': 0.00034918999881483614, 'timing/model': 0.3944742519815918, 'learning_rate/qwen_vl_interface': 1.0485560423468064e-06, 'learning_rate/action_model': 1.0485560423468063e-05, 'epoch': 0.31}) +[2026-10-03 19:47:30,493][starVLA.training.train_starvla][INFO] - Step 4244, Loss: {'train/loss': 0.0031906496733427048, 'train/grad_norm_pre_clip': 13.948532104492188, 'timing/data': 0.00025110002025030553, 'timing/model': 0.3958155419968534, 'learning_rate/qwen_vl_interface': 1.0471361782887421e-06, 'learning_rate/action_model': 1.0471361782887421e-05, 'epoch': 0.31}) +[2026-10-03 19:47:31,402][starVLA.training.train_starvla][INFO] - Step 4245, Loss: {'train/loss': 0.26119017601013184, 'train/grad_norm_pre_clip': 232.67135620117188, 'timing/data': 0.0002490899933036417, 'timing/model': 0.5965559920005035, 'learning_rate/qwen_vl_interface': 1.0457180418692385e-06, 'learning_rate/action_model': 1.0457180418692383e-05, 'epoch': 0.31}) +[2026-10-03 19:47:32,109][starVLA.training.train_starvla][INFO] - Step 4246, Loss: {'train/loss': 0.5000311136245728, 'train/grad_norm_pre_clip': 118.02030181884766, 'timing/data': 0.00032052001915872097, 'timing/model': 0.39492407100624405, 'learning_rate/qwen_vl_interface': 1.0443016336712395e-06, 'learning_rate/action_model': 1.0443016336712393e-05, 'epoch': 0.31}) +[2026-10-03 19:47:32,815][starVLA.training.train_starvla][INFO] - Step 4247, Loss: {'train/loss': 0.13694171607494354, 'train/grad_norm_pre_clip': 168.32745361328125, 'timing/data': 0.0003241600061301142, 'timing/model': 0.39397424200433306, 'learning_rate/qwen_vl_interface': 1.0428869542769744e-06, 'learning_rate/action_model': 1.0428869542769744e-05, 'epoch': 0.31}) +[2026-10-03 19:47:33,524][starVLA.training.train_starvla][INFO] - Step 4248, Loss: {'train/loss': 4.937165431329049e-05, 'train/grad_norm_pre_clip': 0.013217789120972157, 'timing/data': 0.000225240015424788, 'timing/model': 0.3958250009745825, 'learning_rate/qwen_vl_interface': 1.0414740042679663e-06, 'learning_rate/action_model': 1.0414740042679662e-05, 'epoch': 0.31}) +[2026-10-03 19:47:34,232][starVLA.training.train_starvla][INFO] - Step 4249, Loss: {'train/loss': 0.0030804332345724106, 'train/grad_norm_pre_clip': 2.893650531768799, 'timing/data': 0.0003837500116787851, 'timing/model': 0.3956315719988197, 'learning_rate/qwen_vl_interface': 1.040062784225025e-06, 'learning_rate/action_model': 1.0400627842250249e-05, 'epoch': 0.31}) +[2026-10-03 19:47:35,141][starVLA.training.train_starvla][INFO] - Step 4250, Loss: {'train/loss': 0.4449012875556946, 'train/grad_norm_pre_clip': 54.63479995727539, 'timing/data': 0.00035819000913761556, 'timing/model': 0.5955611430108547, 'learning_rate/qwen_vl_interface': 1.0386532947282496e-06, 'learning_rate/action_model': 1.0386532947282495e-05, 'epoch': 0.31}) +[2026-10-03 19:47:35,847][starVLA.training.train_starvla][INFO] - Step 4251, Loss: {'train/loss': 0.0005027918377891183, 'train/grad_norm_pre_clip': 0.3029114603996277, 'timing/data': 0.00022032001288607717, 'timing/model': 0.3959679009858519, 'learning_rate/qwen_vl_interface': 1.0372455363570282e-06, 'learning_rate/action_model': 1.0372455363570283e-05, 'epoch': 0.31}) +[2026-10-03 19:47:36,553][starVLA.training.train_starvla][INFO] - Step 4252, Loss: {'train/loss': 0.6007943749427795, 'train/grad_norm_pre_clip': 86.15489196777344, 'timing/data': 0.0002384699764661491, 'timing/model': 0.3953411219990812, 'learning_rate/qwen_vl_interface': 1.035839509690039e-06, 'learning_rate/action_model': 1.0358395096900388e-05, 'epoch': 0.31}) +[2026-10-03 19:47:37,258][starVLA.training.train_starvla][INFO] - Step 4253, Loss: {'train/loss': 0.7274122834205627, 'train/grad_norm_pre_clip': 85.91070556640625, 'timing/data': 0.0003340599942021072, 'timing/model': 0.3947589920135215, 'learning_rate/qwen_vl_interface': 1.0344352153052429e-06, 'learning_rate/action_model': 1.0344352153052427e-05, 'epoch': 0.31}) +[2026-10-03 19:47:37,967][starVLA.training.train_starvla][INFO] - Step 4254, Loss: {'train/loss': 0.0016398635925725102, 'train/grad_norm_pre_clip': 4.335334777832031, 'timing/data': 0.00022553998860530555, 'timing/model': 0.39688660198589787, 'learning_rate/qwen_vl_interface': 1.033032653779894e-06, 'learning_rate/action_model': 1.033032653779894e-05, 'epoch': 0.31}) +[2026-10-03 19:47:38,878][starVLA.training.train_starvla][INFO] - Step 4255, Loss: {'train/loss': 9.637336188461632e-05, 'train/grad_norm_pre_clip': 0.045492593199014664, 'timing/data': 0.0003376099921297282, 'timing/model': 0.5983208419929724, 'learning_rate/qwen_vl_interface': 1.0316318256905325e-06, 'learning_rate/action_model': 1.0316318256905324e-05, 'epoch': 0.31}) +[2026-10-03 19:47:39,585][starVLA.training.train_starvla][INFO] - Step 4256, Loss: {'train/loss': 0.0003485121123958379, 'train/grad_norm_pre_clip': 0.16896384954452515, 'timing/data': 0.00022421000176109374, 'timing/model': 0.39537085202755406, 'learning_rate/qwen_vl_interface': 1.0302327316129856e-06, 'learning_rate/action_model': 1.0302327316129855e-05, 'epoch': 0.31}) +[2026-10-03 19:47:40,293][starVLA.training.train_starvla][INFO] - Step 4257, Loss: {'train/loss': 0.016907336190342903, 'train/grad_norm_pre_clip': 34.16057205200195, 'timing/data': 0.00025986999389715493, 'timing/model': 0.39498835199628957, 'learning_rate/qwen_vl_interface': 1.028835372122368e-06, 'learning_rate/action_model': 1.0288353721223682e-05, 'epoch': 0.31}) +[2026-10-03 19:47:41,003][starVLA.training.train_starvla][INFO] - Step 4258, Loss: {'train/loss': 0.00038616443634964526, 'train/grad_norm_pre_clip': 0.5346240401268005, 'timing/data': 0.00029768000240437686, 'timing/model': 0.3968232310144231, 'learning_rate/qwen_vl_interface': 1.0274397477930822e-06, 'learning_rate/action_model': 1.0274397477930822e-05, 'epoch': 0.31}) +[2026-10-03 19:47:41,711][starVLA.training.train_starvla][INFO] - Step 4259, Loss: {'train/loss': 0.3727300465106964, 'train/grad_norm_pre_clip': 135.8606719970703, 'timing/data': 0.000235399988014251, 'timing/model': 0.3956991520244628, 'learning_rate/qwen_vl_interface': 1.0260458591988164e-06, 'learning_rate/action_model': 1.0260458591988165e-05, 'epoch': 0.31}) +[2026-10-03 19:47:42,419][starVLA.training.train_starvla][INFO] - Step 4260, Loss: {'train/loss': 0.0006999836186878383, 'train/grad_norm_pre_clip': 0.4865838289260864, 'timing/data': 0.0003246400156058371, 'timing/model': 0.39578911202261224, 'learning_rate/qwen_vl_interface': 1.0246537069125445e-06, 'learning_rate/action_model': 1.0246537069125444e-05, 'epoch': 0.31}) +[2026-10-03 19:47:43,333][starVLA.training.train_starvla][INFO] - Step 4261, Loss: {'train/loss': 0.00020343594951555133, 'train/grad_norm_pre_clip': 0.09511356055736542, 'timing/data': 0.0002897300000768155, 'timing/model': 0.3962368310021702, 'learning_rate/qwen_vl_interface': 1.0232632915065286e-06, 'learning_rate/action_model': 1.0232632915065285e-05, 'epoch': 0.31}) +[2026-10-03 19:47:44,041][starVLA.training.train_starvla][INFO] - Step 4262, Loss: {'train/loss': 0.03346491977572441, 'train/grad_norm_pre_clip': 67.11943817138672, 'timing/data': 0.00022459999308921397, 'timing/model': 0.3964222520007752, 'learning_rate/qwen_vl_interface': 1.021874613552316e-06, 'learning_rate/action_model': 1.0218746135523157e-05, 'epoch': 0.31}) +[2026-10-03 19:47:44,745][starVLA.training.train_starvla][INFO] - Step 4263, Loss: {'train/loss': 5.378925197874196e-05, 'train/grad_norm_pre_clip': 0.013190248981118202, 'timing/data': 0.0002166600024793297, 'timing/model': 0.39440041201305576, 'learning_rate/qwen_vl_interface': 1.0204876736207384e-06, 'learning_rate/action_model': 1.0204876736207384e-05, 'epoch': 0.31}) +[2026-10-03 19:47:45,453][starVLA.training.train_starvla][INFO] - Step 4264, Loss: {'train/loss': 0.00045351681183092296, 'train/grad_norm_pre_clip': 0.3034062087535858, 'timing/data': 0.000292309996439144, 'timing/model': 0.3959596320055425, 'learning_rate/qwen_vl_interface': 1.0191024722819163e-06, 'learning_rate/action_model': 1.0191024722819162e-05, 'epoch': 0.31}) +[2026-10-03 19:47:46,162][starVLA.training.train_starvla][INFO] - Step 4265, Loss: {'train/loss': 0.46810659766197205, 'train/grad_norm_pre_clip': 66.25851440429688, 'timing/data': 0.00022051000269129872, 'timing/model': 0.3967057110276073, 'learning_rate/qwen_vl_interface': 1.017719010105253e-06, 'learning_rate/action_model': 1.017719010105253e-05, 'epoch': 0.31}) +[2026-10-03 19:47:47,069][starVLA.training.train_starvla][INFO] - Step 4266, Loss: {'train/loss': 0.00020716914150398225, 'train/grad_norm_pre_clip': 0.12914958596229553, 'timing/data': 0.00035615998785942793, 'timing/model': 0.5954333329864312, 'learning_rate/qwen_vl_interface': 1.016337287659437e-06, 'learning_rate/action_model': 1.0163372876594368e-05, 'epoch': 0.31}) +[2026-10-03 19:47:47,774][starVLA.training.train_starvla][INFO] - Step 4267, Loss: {'train/loss': 0.06842797249555588, 'train/grad_norm_pre_clip': 110.90142822265625, 'timing/data': 0.00023089000023901463, 'timing/model': 0.39412066200748086, 'learning_rate/qwen_vl_interface': 1.0149573055124427e-06, 'learning_rate/action_model': 1.0149573055124426e-05, 'epoch': 0.31}) +[2026-10-03 19:47:48,479][starVLA.training.train_starvla][INFO] - Step 4268, Loss: {'train/loss': 0.31577691435813904, 'train/grad_norm_pre_clip': 67.06391143798828, 'timing/data': 0.00034121001954190433, 'timing/model': 0.394567011971958, 'learning_rate/qwen_vl_interface': 1.013579064231529e-06, 'learning_rate/action_model': 1.013579064231529e-05, 'epoch': 0.31}) +[2026-10-03 19:47:49,184][starVLA.training.train_starvla][INFO] - Step 4269, Loss: {'train/loss': 5.6583736295579e-05, 'train/grad_norm_pre_clip': 0.010767776519060135, 'timing/data': 0.0002602099848445505, 'timing/model': 0.39448266199906357, 'learning_rate/qwen_vl_interface': 1.012202564383239e-06, 'learning_rate/action_model': 1.0122025643832388e-05, 'epoch': 0.31}) +[2026-10-03 19:47:49,889][starVLA.training.train_starvla][INFO] - Step 4270, Loss: {'train/loss': 0.0012856008252128959, 'train/grad_norm_pre_clip': 0.7390028834342957, 'timing/data': 0.0002610989904496819, 'timing/model': 0.3938987220171839, 'learning_rate/qwen_vl_interface': 1.0108278065333978e-06, 'learning_rate/action_model': 1.0108278065333978e-05, 'epoch': 0.31}) +[2026-10-03 19:47:50,596][starVLA.training.train_starvla][INFO] - Step 4271, Loss: {'train/loss': 0.0832262635231018, 'train/grad_norm_pre_clip': 102.0462875366211, 'timing/data': 0.0002510900085326284, 'timing/model': 0.3949402120197192, 'learning_rate/qwen_vl_interface': 1.0094547912471183e-06, 'learning_rate/action_model': 1.0094547912471184e-05, 'epoch': 0.31}) +[2026-10-03 19:47:51,501][starVLA.training.train_starvla][INFO] - Step 4272, Loss: {'train/loss': 0.00046329363249242306, 'train/grad_norm_pre_clip': 0.6146654486656189, 'timing/data': 0.00029090000316500664, 'timing/model': 0.39461200198275037, 'learning_rate/qwen_vl_interface': 1.0080835190887949e-06, 'learning_rate/action_model': 1.0080835190887947e-05, 'epoch': 0.31}) +[2026-10-03 19:47:52,209][starVLA.training.train_starvla][INFO] - Step 4273, Loss: {'train/loss': 0.029695920646190643, 'train/grad_norm_pre_clip': 83.95320129394531, 'timing/data': 0.0002155299880541861, 'timing/model': 0.3953992819879204, 'learning_rate/qwen_vl_interface': 1.0067139906221052e-06, 'learning_rate/action_model': 1.0067139906221052e-05, 'epoch': 0.31}) +[2026-10-03 19:47:52,919][starVLA.training.train_starvla][INFO] - Step 4274, Loss: {'train/loss': 3.775033837882802e-05, 'train/grad_norm_pre_clip': 0.008362507447600365, 'timing/data': 0.00031291000777855515, 'timing/model': 0.39614735200302675, 'learning_rate/qwen_vl_interface': 1.0053462064100112e-06, 'learning_rate/action_model': 1.0053462064100112e-05, 'epoch': 0.31}) +[2026-10-03 19:47:53,622][starVLA.training.train_starvla][INFO] - Step 4275, Loss: {'train/loss': 0.04562218114733696, 'train/grad_norm_pre_clip': 46.4173698425293, 'timing/data': 0.00022923998767510056, 'timing/model': 0.393323202006286, 'learning_rate/qwen_vl_interface': 1.003980167014757e-06, 'learning_rate/action_model': 1.003980167014757e-05, 'epoch': 0.31}) +[2026-10-03 19:47:54,325][starVLA.training.train_starvla][INFO] - Step 4276, Loss: {'train/loss': 0.0008158026030287147, 'train/grad_norm_pre_clip': 0.7877386808395386, 'timing/data': 0.00021547000505961478, 'timing/model': 0.39348024199716747, 'learning_rate/qwen_vl_interface': 1.0026158729978702e-06, 'learning_rate/action_model': 1.0026158729978702e-05, 'epoch': 0.31}) +[2026-10-03 19:47:55,232][starVLA.training.train_starvla][INFO] - Step 4277, Loss: {'train/loss': 0.4432545304298401, 'train/grad_norm_pre_clip': 76.28704071044922, 'timing/data': 0.0003002499870490283, 'timing/model': 0.3929001720098313, 'learning_rate/qwen_vl_interface': 1.0012533249201604e-06, 'learning_rate/action_model': 1.0012533249201602e-05, 'epoch': 0.31}) +[2026-10-03 19:47:55,936][starVLA.training.train_starvla][INFO] - Step 4278, Loss: {'train/loss': 0.25971630215644836, 'train/grad_norm_pre_clip': 66.69068908691406, 'timing/data': 0.0002245799987576902, 'timing/model': 0.3936577219865285, 'learning_rate/qwen_vl_interface': 9.99892523341719e-07, 'learning_rate/action_model': 9.99892523341719e-06, 'epoch': 0.31}) +[2026-10-03 19:47:56,640][starVLA.training.train_starvla][INFO] - Step 4279, Loss: {'train/loss': 0.4901551902294159, 'train/grad_norm_pre_clip': 47.890716552734375, 'timing/data': 0.00023998998221941292, 'timing/model': 0.3947511709993705, 'learning_rate/qwen_vl_interface': 9.98533468821922e-07, 'learning_rate/action_model': 9.98533468821922e-06, 'epoch': 0.31}) +[2026-10-03 19:47:57,348][starVLA.training.train_starvla][INFO] - Step 4280, Loss: {'train/loss': 7.319131691474468e-05, 'train/grad_norm_pre_clip': 0.043101150542497635, 'timing/data': 0.0002281700144521892, 'timing/model': 0.3951747019891627, 'learning_rate/qwen_vl_interface': 9.971761619194231e-07, 'learning_rate/action_model': 9.97176161919423e-06, 'epoch': 0.31}) +[2026-10-03 19:47:58,055][starVLA.training.train_starvla][INFO] - Step 4281, Loss: {'train/loss': 0.03306369110941887, 'train/grad_norm_pre_clip': 63.857269287109375, 'timing/data': 0.00022806000197306275, 'timing/model': 0.3956022310012486, 'learning_rate/qwen_vl_interface': 9.958206031921606e-07, 'learning_rate/action_model': 9.958206031921604e-06, 'epoch': 0.31}) +[2026-10-03 19:47:58,968][starVLA.training.train_starvla][INFO] - Step 4282, Loss: {'train/loss': 0.0001150718962890096, 'train/grad_norm_pre_clip': 0.03634088113903999, 'timing/data': 0.0002763099910225719, 'timing/model': 0.6021028819959611, 'learning_rate/qwen_vl_interface': 9.944667931973537e-07, 'learning_rate/action_model': 9.944667931973536e-06, 'epoch': 0.31}) +[2026-10-03 19:47:59,673][starVLA.training.train_starvla][INFO] - Step 4283, Loss: {'train/loss': 0.005285865627229214, 'train/grad_norm_pre_clip': 12.288557052612305, 'timing/data': 0.00031646000570617616, 'timing/model': 0.39338186199893244, 'learning_rate/qwen_vl_interface': 9.93114732491503e-07, 'learning_rate/action_model': 9.93114732491503e-06, 'epoch': 0.31}) +[2026-10-03 19:48:00,381][starVLA.training.train_starvla][INFO] - Step 4284, Loss: {'train/loss': 9.871025395113975e-05, 'train/grad_norm_pre_clip': 0.04443979635834694, 'timing/data': 0.00021748000290244818, 'timing/model': 0.39511520200176165, 'learning_rate/qwen_vl_interface': 9.91764421630389e-07, 'learning_rate/action_model': 9.91764421630389e-06, 'epoch': 0.31}) +[2026-10-03 19:48:01,088][starVLA.training.train_starvla][INFO] - Step 4285, Loss: {'train/loss': 0.00022990036814007908, 'train/grad_norm_pre_clip': 0.18084131181240082, 'timing/data': 0.0002280500193592161, 'timing/model': 0.3949557419982739, 'learning_rate/qwen_vl_interface': 9.904158611690732e-07, 'learning_rate/action_model': 9.90415861169073e-06, 'epoch': 0.31}) +[2026-10-03 19:48:01,795][starVLA.training.train_starvla][INFO] - Step 4286, Loss: {'train/loss': 8.169846114469692e-05, 'train/grad_norm_pre_clip': 0.023762758821249008, 'timing/data': 0.00036174998967908323, 'timing/model': 0.394826642004773, 'learning_rate/qwen_vl_interface': 9.890690516618987e-07, 'learning_rate/action_model': 9.890690516618985e-06, 'epoch': 0.31}) +[2026-10-03 19:48:02,501][starVLA.training.train_starvla][INFO] - Step 4287, Loss: {'train/loss': 0.1254393309354782, 'train/grad_norm_pre_clip': 92.84835052490234, 'timing/data': 0.0003339000104460865, 'timing/model': 0.39433423199807294, 'learning_rate/qwen_vl_interface': 9.877239936624875e-07, 'learning_rate/action_model': 9.877239936624875e-06, 'epoch': 0.31}) +[2026-10-03 19:48:03,407][starVLA.training.train_starvla][INFO] - Step 4288, Loss: {'train/loss': 0.371699720621109, 'train/grad_norm_pre_clip': 74.48094177246094, 'timing/data': 0.00023062998661771417, 'timing/model': 0.3954367819824256, 'learning_rate/qwen_vl_interface': 9.863806877237421e-07, 'learning_rate/action_model': 9.863806877237422e-06, 'epoch': 0.31}) +[2026-10-03 19:48:04,113][starVLA.training.train_starvla][INFO] - Step 4289, Loss: {'train/loss': 0.00014035881031304598, 'train/grad_norm_pre_clip': 0.05384662002325058, 'timing/data': 0.00029949998133815825, 'timing/model': 0.39520864200312644, 'learning_rate/qwen_vl_interface': 9.850391343978454e-07, 'learning_rate/action_model': 9.850391343978454e-06, 'epoch': 0.31}) +[2026-10-03 19:48:04,817][starVLA.training.train_starvla][INFO] - Step 4290, Loss: {'train/loss': 0.0010726707987487316, 'train/grad_norm_pre_clip': 1.2130814790725708, 'timing/data': 0.00022523998632095754, 'timing/model': 0.3946459010185208, 'learning_rate/qwen_vl_interface': 9.836993342362596e-07, 'learning_rate/action_model': 9.836993342362596e-06, 'epoch': 0.31}) +[2026-10-03 19:48:05,521][starVLA.training.train_starvla][INFO] - Step 4291, Loss: {'train/loss': 0.0003197167825419456, 'train/grad_norm_pre_clip': 0.16654901206493378, 'timing/data': 0.00033346901182085276, 'timing/model': 0.39409899202291854, 'learning_rate/qwen_vl_interface': 9.823612877897246e-07, 'learning_rate/action_model': 9.823612877897245e-06, 'epoch': 0.31}) +[2026-10-03 19:48:06,228][starVLA.training.train_starvla][INFO] - Step 4292, Loss: {'train/loss': 0.21917790174484253, 'train/grad_norm_pre_clip': 69.83492279052734, 'timing/data': 0.0003435900143813342, 'timing/model': 0.3952860420104116, 'learning_rate/qwen_vl_interface': 9.81024995608261e-07, 'learning_rate/action_model': 9.810249956082609e-06, 'epoch': 0.31}) +[2026-10-03 19:48:07,146][starVLA.training.train_starvla][INFO] - Step 4293, Loss: {'train/loss': 0.0007663697469979525, 'train/grad_norm_pre_clip': 0.5781930685043335, 'timing/data': 0.00023534000501967967, 'timing/model': 0.3950249420013279, 'learning_rate/qwen_vl_interface': 9.796904582411685e-07, 'learning_rate/action_model': 9.796904582411684e-06, 'epoch': 0.31}) +[2026-10-03 19:48:07,852][starVLA.training.train_starvla][INFO] - Step 4294, Loss: {'train/loss': 0.00028010623645968735, 'train/grad_norm_pre_clip': 0.1765388697385788, 'timing/data': 0.0002156799891963601, 'timing/model': 0.3946473010000773, 'learning_rate/qwen_vl_interface': 9.78357676237025e-07, 'learning_rate/action_model': 9.78357676237025e-06, 'epoch': 0.31}) +[2026-10-03 19:48:08,557][starVLA.training.train_starvla][INFO] - Step 4295, Loss: {'train/loss': 7.182229455793276e-05, 'train/grad_norm_pre_clip': 0.02600770629942417, 'timing/data': 0.00038394000148400664, 'timing/model': 0.39429100198321976, 'learning_rate/qwen_vl_interface': 9.77026650143687e-07, 'learning_rate/action_model': 9.77026650143687e-06, 'epoch': 0.31}) +[2026-10-03 19:48:09,263][starVLA.training.train_starvla][INFO] - Step 4296, Loss: {'train/loss': 6.066917921998538e-05, 'train/grad_norm_pre_clip': 0.01885036565363407, 'timing/data': 0.00033296001492999494, 'timing/model': 0.39415859201108105, 'learning_rate/qwen_vl_interface': 9.756973805082882e-07, 'learning_rate/action_model': 9.756973805082883e-06, 'epoch': 0.31}) +[2026-10-03 19:48:09,968][starVLA.training.train_starvla][INFO] - Step 4297, Loss: {'train/loss': 0.000796052161604166, 'train/grad_norm_pre_clip': 1.1496508121490479, 'timing/data': 0.0002279799955431372, 'timing/model': 0.3936335210164543, 'learning_rate/qwen_vl_interface': 9.743698678772434e-07, 'learning_rate/action_model': 9.743698678772433e-06, 'epoch': 0.31}) +[2026-10-03 19:48:10,877][starVLA.training.train_starvla][INFO] - Step 4298, Loss: {'train/loss': 0.7032904624938965, 'train/grad_norm_pre_clip': 360.3911437988281, 'timing/data': 0.00024061999283730984, 'timing/model': 0.5992653120192699, 'learning_rate/qwen_vl_interface': 9.730441127962395e-07, 'learning_rate/action_model': 9.730441127962395e-06, 'epoch': 0.31}) +[2026-10-03 19:48:11,583][starVLA.training.train_starvla][INFO] - Step 4299, Loss: {'train/loss': 0.0001064081079675816, 'train/grad_norm_pre_clip': 0.05389107018709183, 'timing/data': 0.00022501000785268843, 'timing/model': 0.39495432199328206, 'learning_rate/qwen_vl_interface': 9.717201158102465e-07, 'learning_rate/action_model': 9.717201158102466e-06, 'epoch': 0.31}) +[2026-10-03 19:48:12,264][starVLA.training.train_starvla][INFO] - Step 4300, Loss: {'train/loss': 0.4250538647174835, 'train/grad_norm_pre_clip': 118.73178100585938, 'timing/data': 0.00023436002084054053, 'timing/model': 0.393657892011106, 'learning_rate/qwen_vl_interface': 9.703978774635094e-07, 'learning_rate/action_model': 9.703978774635095e-06, 'epoch': 0.31}) +[2026-10-03 19:48:12,971][starVLA.training.train_starvla][INFO] - Step 4301, Loss: {'train/loss': 0.32631227374076843, 'train/grad_norm_pre_clip': 185.0152587890625, 'timing/data': 0.00034242001129314303, 'timing/model': 0.3951067409943789, 'learning_rate/qwen_vl_interface': 9.6907739829955e-07, 'learning_rate/action_model': 9.6907739829955e-06, 'epoch': 0.31}) +[2026-10-03 19:48:13,675][starVLA.training.train_starvla][INFO] - Step 4302, Loss: {'train/loss': 0.00024851574562489986, 'train/grad_norm_pre_clip': 0.10329879075288773, 'timing/data': 0.0003563700011000037, 'timing/model': 0.39379026199458167, 'learning_rate/qwen_vl_interface': 9.677586788611681e-07, 'learning_rate/action_model': 9.677586788611681e-06, 'epoch': 0.31}) +[2026-10-03 19:48:14,585][starVLA.training.train_starvla][INFO] - Step 4303, Loss: {'train/loss': 0.2862197458744049, 'train/grad_norm_pre_clip': 194.4877471923828, 'timing/data': 0.0002281600027345121, 'timing/model': 0.599703143001534, 'learning_rate/qwen_vl_interface': 9.664417196904382e-07, 'learning_rate/action_model': 9.664417196904382e-06, 'epoch': 0.31}) +[2026-10-03 19:48:15,289][starVLA.training.train_starvla][INFO] - Step 4304, Loss: {'train/loss': 0.13624343276023865, 'train/grad_norm_pre_clip': 219.7519073486328, 'timing/data': 0.00022567002451978624, 'timing/model': 0.39408146298956126, 'learning_rate/qwen_vl_interface': 9.651265213287134e-07, 'learning_rate/action_model': 9.651265213287134e-06, 'epoch': 0.32}) +[2026-10-03 19:48:15,994][starVLA.training.train_starvla][INFO] - Step 4305, Loss: {'train/loss': 0.40841495990753174, 'train/grad_norm_pre_clip': 196.06414794921875, 'timing/data': 0.0003917499852832407, 'timing/model': 0.3950052119907923, 'learning_rate/qwen_vl_interface': 9.638130843166218e-07, 'learning_rate/action_model': 9.638130843166219e-06, 'epoch': 0.32}) +[2026-10-03 19:48:16,698][starVLA.training.train_starvla][INFO] - Step 4306, Loss: {'train/loss': 0.0019137172494083643, 'train/grad_norm_pre_clip': 2.2651097774505615, 'timing/data': 0.00034888999653048813, 'timing/model': 0.3932161819830071, 'learning_rate/qwen_vl_interface': 9.62501409194068e-07, 'learning_rate/action_model': 9.625014091940678e-06, 'epoch': 0.32}) +[2026-10-03 19:48:17,403][starVLA.training.train_starvla][INFO] - Step 4307, Loss: {'train/loss': 0.4101273715496063, 'train/grad_norm_pre_clip': 274.6514587402344, 'timing/data': 0.000335429998813197, 'timing/model': 0.3945566109905485, 'learning_rate/qwen_vl_interface': 9.611914965002317e-07, 'learning_rate/action_model': 9.611914965002316e-06, 'epoch': 0.32}) +[2026-10-03 19:48:18,111][starVLA.training.train_starvla][INFO] - Step 4308, Loss: {'train/loss': 0.5117395520210266, 'train/grad_norm_pre_clip': 139.5673828125, 'timing/data': 0.00029259000439196825, 'timing/model': 0.3938007520046085, 'learning_rate/qwen_vl_interface': 9.598833467735686e-07, 'learning_rate/action_model': 9.598833467735685e-06, 'epoch': 0.32}) +[2026-10-03 19:48:19,033][starVLA.training.train_starvla][INFO] - Step 4309, Loss: {'train/loss': 6.564151408383623e-05, 'train/grad_norm_pre_clip': 0.01865490712225437, 'timing/data': 0.0002392890164628625, 'timing/model': 0.39522303198464215, 'learning_rate/qwen_vl_interface': 9.585769605518097e-07, 'learning_rate/action_model': 9.585769605518097e-06, 'epoch': 0.32}) +[2026-10-03 19:48:19,740][starVLA.training.train_starvla][INFO] - Step 4310, Loss: {'train/loss': 0.00016684771981090307, 'train/grad_norm_pre_clip': 0.06106436252593994, 'timing/data': 0.00035735999699682, 'timing/model': 0.39444912198814563, 'learning_rate/qwen_vl_interface': 9.572723383719611e-07, 'learning_rate/action_model': 9.572723383719612e-06, 'epoch': 0.32}) +[2026-10-03 19:48:20,442][starVLA.training.train_starvla][INFO] - Step 4311, Loss: {'train/loss': 0.41264915466308594, 'train/grad_norm_pre_clip': 83.78507995605469, 'timing/data': 0.00032577000092715025, 'timing/model': 0.3921603919879999, 'learning_rate/qwen_vl_interface': 9.559694807703041e-07, 'learning_rate/action_model': 9.55969480770304e-06, 'epoch': 0.32}) +[2026-10-03 19:48:21,141][starVLA.training.train_starvla][INFO] - Step 4312, Loss: {'train/loss': 0.0004101158119738102, 'train/grad_norm_pre_clip': 0.3008413314819336, 'timing/data': 0.00020243998733349144, 'timing/model': 0.3906611019920092, 'learning_rate/qwen_vl_interface': 9.546683882823945e-07, 'learning_rate/action_model': 9.546683882823945e-06, 'epoch': 0.32}) +[2026-10-03 19:48:21,843][starVLA.training.train_starvla][INFO] - Step 4313, Loss: {'train/loss': 0.5113937258720398, 'train/grad_norm_pre_clip': 29.599794387817383, 'timing/data': 0.00027792001492343843, 'timing/model': 0.39341999200405553, 'learning_rate/qwen_vl_interface': 9.533690614430608e-07, 'learning_rate/action_model': 9.533690614430608e-06, 'epoch': 0.32}) +[2026-10-03 19:48:22,739][starVLA.training.train_starvla][INFO] - Step 4314, Loss: {'train/loss': 0.06436522305011749, 'train/grad_norm_pre_clip': 66.16911315917969, 'timing/data': 0.00031873001717031, 'timing/model': 0.38977454201085493, 'learning_rate/qwen_vl_interface': 9.520715007864082e-07, 'learning_rate/action_model': 9.52071500786408e-06, 'epoch': 0.32}) +[2026-10-03 19:48:23,438][starVLA.training.train_starvla][INFO] - Step 4315, Loss: {'train/loss': 0.4185853600502014, 'train/grad_norm_pre_clip': 24.59577178955078, 'timing/data': 0.000205810007173568, 'timing/model': 0.3900128420209512, 'learning_rate/qwen_vl_interface': 9.50775706845814e-07, 'learning_rate/action_model': 9.50775706845814e-06, 'epoch': 0.32}) +[2026-10-03 19:48:24,139][starVLA.training.train_starvla][INFO] - Step 4316, Loss: {'train/loss': 6.959360325708985e-05, 'train/grad_norm_pre_clip': 0.024455664679408073, 'timing/data': 0.0003525000065565109, 'timing/model': 0.3921354319900274, 'learning_rate/qwen_vl_interface': 9.49481680153931e-07, 'learning_rate/action_model': 9.494816801539309e-06, 'epoch': 0.32}) +[2026-10-03 19:48:24,840][starVLA.training.train_starvla][INFO] - Step 4317, Loss: {'train/loss': 0.020114917308092117, 'train/grad_norm_pre_clip': 42.63827896118164, 'timing/data': 0.00027305001276545227, 'timing/model': 0.39314766199095175, 'learning_rate/qwen_vl_interface': 9.481894212426848e-07, 'learning_rate/action_model': 9.481894212426847e-06, 'epoch': 0.32}) +[2026-10-03 19:48:25,542][starVLA.training.train_starvla][INFO] - Step 4318, Loss: {'train/loss': 0.294046014547348, 'train/grad_norm_pre_clip': 47.13212203979492, 'timing/data': 0.00027377999504096806, 'timing/model': 0.3925081719935406, 'learning_rate/qwen_vl_interface': 9.468989306432738e-07, 'learning_rate/action_model': 9.468989306432738e-06, 'epoch': 0.32}) +[2026-10-03 19:48:26,445][starVLA.training.train_starvla][INFO] - Step 4319, Loss: {'train/loss': 0.0013296552933752537, 'train/grad_norm_pre_clip': 0.9150893688201904, 'timing/data': 0.0003239900106564164, 'timing/model': 0.5901275629876181, 'learning_rate/qwen_vl_interface': 9.456102088861702e-07, 'learning_rate/action_model': 9.456102088861701e-06, 'epoch': 0.32}) +[2026-10-03 19:48:27,146][starVLA.training.train_starvla][INFO] - Step 4320, Loss: {'train/loss': 0.00011164595343871042, 'train/grad_norm_pre_clip': 0.02321166731417179, 'timing/data': 0.0002314600278623402, 'timing/model': 0.3921407920133788, 'learning_rate/qwen_vl_interface': 9.443232565011188e-07, 'learning_rate/action_model': 9.443232565011187e-06, 'epoch': 0.32}) +[2026-10-03 19:48:27,845][starVLA.training.train_starvla][INFO] - Step 4321, Loss: {'train/loss': 0.3218607008457184, 'train/grad_norm_pre_clip': 48.631465911865234, 'timing/data': 0.00026935001369565725, 'timing/model': 0.39309243200113997, 'learning_rate/qwen_vl_interface': 9.430380740171369e-07, 'learning_rate/action_model': 9.430380740171368e-06, 'epoch': 0.32}) +[2026-10-03 19:48:28,542][starVLA.training.train_starvla][INFO] - Step 4322, Loss: {'train/loss': 0.0003568493411876261, 'train/grad_norm_pre_clip': 0.11908270418643951, 'timing/data': 0.00043886000639759004, 'timing/model': 0.3882956020243, 'learning_rate/qwen_vl_interface': 9.417546619625143e-07, 'learning_rate/action_model': 9.417546619625143e-06, 'epoch': 0.32}) +[2026-10-03 19:48:29,243][starVLA.training.train_starvla][INFO] - Step 4323, Loss: {'train/loss': 0.0003638913622125983, 'train/grad_norm_pre_clip': 0.3474937081336975, 'timing/data': 0.00026421999791637063, 'timing/model': 0.3924083420133684, 'learning_rate/qwen_vl_interface': 9.404730208648144e-07, 'learning_rate/action_model': 9.404730208648144e-06, 'epoch': 0.32}) +[2026-10-03 19:48:30,139][starVLA.training.train_starvla][INFO] - Step 4324, Loss: {'train/loss': 0.00017077371012419462, 'train/grad_norm_pre_clip': 0.06546967476606369, 'timing/data': 0.0003041500167455524, 'timing/model': 0.5880198430095334, 'learning_rate/qwen_vl_interface': 9.3919315125087e-07, 'learning_rate/action_model': 9.3919315125087e-06, 'epoch': 0.32}) +[2026-10-03 19:48:30,842][starVLA.training.train_starvla][INFO] - Step 4325, Loss: {'train/loss': 0.0011177221313118935, 'train/grad_norm_pre_clip': 0.805834949016571, 'timing/data': 0.00043085002107545733, 'timing/model': 0.3931691120087635, 'learning_rate/qwen_vl_interface': 9.379150536467879e-07, 'learning_rate/action_model': 9.379150536467878e-06, 'epoch': 0.32}) +[2026-10-03 19:48:31,540][starVLA.training.train_starvla][INFO] - Step 4326, Loss: {'train/loss': 0.3493512272834778, 'train/grad_norm_pre_clip': 116.91167449951172, 'timing/data': 0.00020348001271486282, 'timing/model': 0.3916734819940757, 'learning_rate/qwen_vl_interface': 9.366387285779453e-07, 'learning_rate/action_model': 9.366387285779453e-06, 'epoch': 0.32}) +[2026-10-03 19:48:32,241][starVLA.training.train_starvla][INFO] - Step 4327, Loss: {'train/loss': 4.4044922105968e-05, 'train/grad_norm_pre_clip': 0.009852644987404346, 'timing/data': 0.0003350800252519548, 'timing/model': 0.39199914201162755, 'learning_rate/qwen_vl_interface': 9.353641765689915e-07, 'learning_rate/action_model': 9.353641765689915e-06, 'epoch': 0.32}) +[2026-10-03 19:48:32,940][starVLA.training.train_starvla][INFO] - Step 4328, Loss: {'train/loss': 0.36807316541671753, 'train/grad_norm_pre_clip': 146.27072143554688, 'timing/data': 0.00024793000193312764, 'timing/model': 0.3897832820075564, 'learning_rate/qwen_vl_interface': 9.340913981438462e-07, 'learning_rate/action_model': 9.340913981438463e-06, 'epoch': 0.32}) +[2026-10-03 19:48:33,638][starVLA.training.train_starvla][INFO] - Step 4329, Loss: {'train/loss': 0.8221083879470825, 'train/grad_norm_pre_clip': 81.4122543334961, 'timing/data': 0.0004218399990350008, 'timing/model': 0.3916982419905253, 'learning_rate/qwen_vl_interface': 9.328203938257019e-07, 'learning_rate/action_model': 9.328203938257018e-06, 'epoch': 0.32}) +[2026-10-03 19:48:34,536][starVLA.training.train_starvla][INFO] - Step 4330, Loss: {'train/loss': 0.05069736763834953, 'train/grad_norm_pre_clip': 145.84970092773438, 'timing/data': 0.0003188599948771298, 'timing/model': 0.3918122029863298, 'learning_rate/qwen_vl_interface': 9.315511641370191e-07, 'learning_rate/action_model': 9.315511641370191e-06, 'epoch': 0.32}) +[2026-10-03 19:48:35,232][starVLA.training.train_starvla][INFO] - Step 4331, Loss: {'train/loss': 0.00019738139235414565, 'train/grad_norm_pre_clip': 0.07087261229753494, 'timing/data': 0.00027444001170806587, 'timing/model': 0.38956130301812664, 'learning_rate/qwen_vl_interface': 9.302837095995306e-07, 'learning_rate/action_model': 9.302837095995306e-06, 'epoch': 0.32}) +[2026-10-03 19:48:35,929][starVLA.training.train_starvla][INFO] - Step 4332, Loss: {'train/loss': 0.00018678136984817684, 'train/grad_norm_pre_clip': 0.05951128154993057, 'timing/data': 0.00023316001170314848, 'timing/model': 0.3912576219881885, 'learning_rate/qwen_vl_interface': 9.290180307342392e-07, 'learning_rate/action_model': 9.290180307342391e-06, 'epoch': 0.32}) +[2026-10-03 19:48:36,631][starVLA.training.train_starvla][INFO] - Step 4333, Loss: {'train/loss': 8.821376832202077e-05, 'train/grad_norm_pre_clip': 0.026247290894389153, 'timing/data': 0.0003658199857454747, 'timing/model': 0.3923932820034679, 'learning_rate/qwen_vl_interface': 9.277541280614183e-07, 'learning_rate/action_model': 9.277541280614183e-06, 'epoch': 0.32}) +[2026-10-03 19:48:37,330][starVLA.training.train_starvla][INFO] - Step 4334, Loss: {'train/loss': 9.00240265764296e-05, 'train/grad_norm_pre_clip': 0.01969095505774021, 'timing/data': 0.00036620997707359493, 'timing/model': 0.3911002220120281, 'learning_rate/qwen_vl_interface': 9.264920021006093e-07, 'learning_rate/action_model': 9.264920021006093e-06, 'epoch': 0.32}) +[2026-10-03 19:48:38,231][starVLA.training.train_starvla][INFO] - Step 4335, Loss: {'train/loss': 0.2384636253118515, 'train/grad_norm_pre_clip': 256.36871337890625, 'timing/data': 0.0002330299757886678, 'timing/model': 0.39138621199526824, 'learning_rate/qwen_vl_interface': 9.252316533706246e-07, 'learning_rate/action_model': 9.252316533706246e-06, 'epoch': 0.32}) +[2026-10-03 19:48:38,931][starVLA.training.train_starvla][INFO] - Step 4336, Loss: {'train/loss': 0.00018034508684650064, 'train/grad_norm_pre_clip': 0.18620269000530243, 'timing/data': 0.00032592000206932425, 'timing/model': 0.39181240199832246, 'learning_rate/qwen_vl_interface': 9.239730823895469e-07, 'learning_rate/action_model': 9.23973082389547e-06, 'epoch': 0.32}) +[2026-10-03 19:48:39,629][starVLA.training.train_starvla][INFO] - Step 4337, Loss: {'train/loss': 5.50117802049499e-05, 'train/grad_norm_pre_clip': 0.013280863873660564, 'timing/data': 0.00035761998151429, 'timing/model': 0.39064456202322617, 'learning_rate/qwen_vl_interface': 9.227162896747264e-07, 'learning_rate/action_model': 9.227162896747264e-06, 'epoch': 0.32}) +[2026-10-03 19:48:40,329][starVLA.training.train_starvla][INFO] - Step 4338, Loss: {'train/loss': 0.3156984746456146, 'train/grad_norm_pre_clip': 58.81891632080078, 'timing/data': 0.0003435600083321333, 'timing/model': 0.3916971219878178, 'learning_rate/qwen_vl_interface': 9.214612757427833e-07, 'learning_rate/action_model': 9.214612757427833e-06, 'epoch': 0.32}) +[2026-10-03 19:48:41,029][starVLA.training.train_starvla][INFO] - Step 4339, Loss: {'train/loss': 0.000157447240781039, 'train/grad_norm_pre_clip': 0.060385994613170624, 'timing/data': 0.0003261899983044714, 'timing/model': 0.39037835199269466, 'learning_rate/qwen_vl_interface': 9.202080411096071e-07, 'learning_rate/action_model': 9.20208041109607e-06, 'epoch': 0.32}) +[2026-10-03 19:48:41,933][starVLA.training.train_starvla][INFO] - Step 4340, Loss: {'train/loss': 0.00026830629212781787, 'train/grad_norm_pre_clip': 0.07768327742815018, 'timing/data': 0.00032237000414170325, 'timing/model': 0.5954682530136779, 'learning_rate/qwen_vl_interface': 9.189565862903543e-07, 'learning_rate/action_model': 9.189565862903543e-06, 'epoch': 0.32}) +[2026-10-03 19:48:42,642][starVLA.training.train_starvla][INFO] - Step 4341, Loss: {'train/loss': 0.003143078414723277, 'train/grad_norm_pre_clip': 4.109273433685303, 'timing/data': 0.00023249001242220402, 'timing/model': 0.39583096202113666, 'learning_rate/qwen_vl_interface': 9.17706911799451e-07, 'learning_rate/action_model': 9.17706911799451e-06, 'epoch': 0.32}) +[2026-10-03 19:48:43,348][starVLA.training.train_starvla][INFO] - Step 4342, Loss: {'train/loss': 0.017005229368805885, 'train/grad_norm_pre_clip': 39.42374038696289, 'timing/data': 0.0002622799947857857, 'timing/model': 0.395435731014004, 'learning_rate/qwen_vl_interface': 9.164590181505912e-07, 'learning_rate/action_model': 9.164590181505911e-06, 'epoch': 0.32}) +[2026-10-03 19:48:44,059][starVLA.training.train_starvla][INFO] - Step 4343, Loss: {'train/loss': 0.30242738127708435, 'train/grad_norm_pre_clip': 83.4273910522461, 'timing/data': 0.0003174599842168391, 'timing/model': 0.3972271819948219, 'learning_rate/qwen_vl_interface': 9.15212905856738e-07, 'learning_rate/action_model': 9.15212905856738e-06, 'epoch': 0.32}) +[2026-10-03 19:48:44,767][starVLA.training.train_starvla][INFO] - Step 4344, Loss: {'train/loss': 0.010595206171274185, 'train/grad_norm_pre_clip': 38.10087203979492, 'timing/data': 0.00030174999847076833, 'timing/model': 0.3967756920028478, 'learning_rate/qwen_vl_interface': 9.139685754301189e-07, 'learning_rate/action_model': 9.139685754301189e-06, 'epoch': 0.32}) +[2026-10-03 19:48:45,671][starVLA.training.train_starvla][INFO] - Step 4345, Loss: {'train/loss': 0.0521404854953289, 'train/grad_norm_pre_clip': 92.05303192138672, 'timing/data': 0.00037280001561157405, 'timing/model': 0.5907802220026497, 'learning_rate/qwen_vl_interface': 9.127260273822324e-07, 'learning_rate/action_model': 9.127260273822325e-06, 'epoch': 0.32}) +[2026-10-03 19:48:46,377][starVLA.training.train_starvla][INFO] - Step 4346, Loss: {'train/loss': 2.7387613954488188e-05, 'train/grad_norm_pre_clip': 0.004370121750980616, 'timing/data': 0.000357550015905872, 'timing/model': 0.39430549199460074, 'learning_rate/qwen_vl_interface': 9.114852622238435e-07, 'learning_rate/action_model': 9.114852622238435e-06, 'epoch': 0.32}) +[2026-10-03 19:48:47,082][starVLA.training.train_starvla][INFO] - Step 4347, Loss: {'train/loss': 0.002010773401707411, 'train/grad_norm_pre_clip': 2.5931191444396973, 'timing/data': 0.00022156999330036342, 'timing/model': 0.39451132097747177, 'learning_rate/qwen_vl_interface': 9.102462804649827e-07, 'learning_rate/action_model': 9.102462804649827e-06, 'epoch': 0.32}) +[2026-10-03 19:48:47,789][starVLA.training.train_starvla][INFO] - Step 4348, Loss: {'train/loss': 0.00020767809473909438, 'train/grad_norm_pre_clip': 0.03951941058039665, 'timing/data': 0.00030857001547701657, 'timing/model': 0.3940039720037021, 'learning_rate/qwen_vl_interface': 9.090090826149489e-07, 'learning_rate/action_model': 9.09009082614949e-06, 'epoch': 0.32}) +[2026-10-03 19:48:48,496][starVLA.training.train_starvla][INFO] - Step 4349, Loss: {'train/loss': 0.0004071810108143836, 'train/grad_norm_pre_clip': 0.23326140642166138, 'timing/data': 0.0002162099990528077, 'timing/model': 0.3951395720068831, 'learning_rate/qwen_vl_interface': 9.077736691823081e-07, 'learning_rate/action_model': 9.07773669182308e-06, 'epoch': 0.32}) +[2026-10-03 19:48:49,201][starVLA.training.train_starvla][INFO] - Step 4350, Loss: {'train/loss': 0.0017027792055159807, 'train/grad_norm_pre_clip': 1.655275583267212, 'timing/data': 0.0002020499960053712, 'timing/model': 0.3941318620054517, 'learning_rate/qwen_vl_interface': 9.065400406748912e-07, 'learning_rate/action_model': 9.06540040674891e-06, 'epoch': 0.32}) +[2026-10-03 19:48:50,108][starVLA.training.train_starvla][INFO] - Step 4351, Loss: {'train/loss': 0.22791258990764618, 'train/grad_norm_pre_clip': 63.25507736206055, 'timing/data': 0.0002805099938996136, 'timing/model': 0.397219431004487, 'learning_rate/qwen_vl_interface': 9.053081975997969e-07, 'learning_rate/action_model': 9.053081975997968e-06, 'epoch': 0.32}) +[2026-10-03 19:48:50,815][starVLA.training.train_starvla][INFO] - Step 4352, Loss: {'train/loss': 0.1370028704404831, 'train/grad_norm_pre_clip': 301.2027587890625, 'timing/data': 0.00024313997710123658, 'timing/model': 0.39584798098076135, 'learning_rate/qwen_vl_interface': 9.040781404633894e-07, 'learning_rate/action_model': 9.040781404633895e-06, 'epoch': 0.32}) +[2026-10-03 19:48:51,522][starVLA.training.train_starvla][INFO] - Step 4353, Loss: {'train/loss': 7.873629510868341e-05, 'train/grad_norm_pre_clip': 0.019066961482167244, 'timing/data': 0.00023206000332720578, 'timing/model': 0.39521377100027166, 'learning_rate/qwen_vl_interface': 9.028498697712982e-07, 'learning_rate/action_model': 9.028498697712981e-06, 'epoch': 0.32}) +[2026-10-03 19:48:52,230][starVLA.training.train_starvla][INFO] - Step 4354, Loss: {'train/loss': 0.0008488581515848637, 'train/grad_norm_pre_clip': 0.8086890578269958, 'timing/data': 0.0003469199873507023, 'timing/model': 0.39484761201310903, 'learning_rate/qwen_vl_interface': 9.016233860284195e-07, 'learning_rate/action_model': 9.016233860284195e-06, 'epoch': 0.32}) +[2026-10-03 19:48:52,939][starVLA.training.train_starvla][INFO] - Step 4355, Loss: {'train/loss': 0.00018703227397054434, 'train/grad_norm_pre_clip': 0.08089106529951096, 'timing/data': 0.0003741690015885979, 'timing/model': 0.3964648319815751, 'learning_rate/qwen_vl_interface': 9.003986897389142e-07, 'learning_rate/action_model': 9.003986897389141e-06, 'epoch': 0.32}) +[2026-10-03 19:48:53,855][starVLA.training.train_starvla][INFO] - Step 4356, Loss: {'train/loss': 0.4921548068523407, 'train/grad_norm_pre_clip': 38.69350051879883, 'timing/data': 0.0004075299948453903, 'timing/model': 0.3947082619997673, 'learning_rate/qwen_vl_interface': 8.991757814062094e-07, 'learning_rate/action_model': 8.991757814062093e-06, 'epoch': 0.32}) +[2026-10-03 19:48:54,560][starVLA.training.train_starvla][INFO] - Step 4357, Loss: {'train/loss': 0.00010479147749720141, 'train/grad_norm_pre_clip': 0.02950950339436531, 'timing/data': 0.00036485999589785933, 'timing/model': 0.39345356199191883, 'learning_rate/qwen_vl_interface': 8.979546615329954e-07, 'learning_rate/action_model': 8.979546615329953e-06, 'epoch': 0.32}) +[2026-10-03 19:48:55,268][starVLA.training.train_starvla][INFO] - Step 4358, Loss: {'train/loss': 0.00027110299561172724, 'train/grad_norm_pre_clip': 0.5406211614608765, 'timing/data': 0.00034448999213054776, 'timing/model': 0.3950398009910714, 'learning_rate/qwen_vl_interface': 8.967353306212286e-07, 'learning_rate/action_model': 8.967353306212286e-06, 'epoch': 0.32}) +[2026-10-03 19:48:55,974][starVLA.training.train_starvla][INFO] - Step 4359, Loss: {'train/loss': 0.0006034470861777663, 'train/grad_norm_pre_clip': 1.362098217010498, 'timing/data': 0.0003398600092623383, 'timing/model': 0.3938860920025036, 'learning_rate/qwen_vl_interface': 8.955177891721303e-07, 'learning_rate/action_model': 8.955177891721303e-06, 'epoch': 0.32}) +[2026-10-03 19:48:56,678][starVLA.training.train_starvla][INFO] - Step 4360, Loss: {'train/loss': 7.172206824179739e-05, 'train/grad_norm_pre_clip': 0.010450633242726326, 'timing/data': 0.00024779001250863075, 'timing/model': 0.39440441198530607, 'learning_rate/qwen_vl_interface': 8.943020376861858e-07, 'learning_rate/action_model': 8.943020376861857e-06, 'epoch': 0.32}) +[2026-10-03 19:48:57,589][starVLA.training.train_starvla][INFO] - Step 4361, Loss: {'train/loss': 0.007381372153759003, 'train/grad_norm_pre_clip': 14.00466537475586, 'timing/data': 0.00033621001057326794, 'timing/model': 0.5987299219996203, 'learning_rate/qwen_vl_interface': 8.930880766631444e-07, 'learning_rate/action_model': 8.930880766631445e-06, 'epoch': 0.32}) +[2026-10-03 19:48:58,293][starVLA.training.train_starvla][INFO] - Step 4362, Loss: {'train/loss': 0.43815353512763977, 'train/grad_norm_pre_clip': 129.2913818359375, 'timing/data': 0.0002255300059914589, 'timing/model': 0.39364220199058764, 'learning_rate/qwen_vl_interface': 8.918759066020201e-07, 'learning_rate/action_model': 8.9187590660202e-06, 'epoch': 0.32}) +[2026-10-03 19:48:59,001][starVLA.training.train_starvla][INFO] - Step 4363, Loss: {'train/loss': 5.27320516994223e-05, 'train/grad_norm_pre_clip': 0.012657465413212776, 'timing/data': 0.00025226999423466623, 'timing/model': 0.3953851820260752, 'learning_rate/qwen_vl_interface': 8.906655280010895e-07, 'learning_rate/action_model': 8.906655280010896e-06, 'epoch': 0.32}) +[2026-10-03 19:48:59,705][starVLA.training.train_starvla][INFO] - Step 4364, Loss: {'train/loss': 0.0017008405411615968, 'train/grad_norm_pre_clip': 3.072082757949829, 'timing/data': 0.0003293699992354959, 'timing/model': 0.3939810320152901, 'learning_rate/qwen_vl_interface': 8.894569413578938e-07, 'learning_rate/action_model': 8.894569413578937e-06, 'epoch': 0.32}) +[2026-10-03 19:49:00,413][starVLA.training.train_starvla][INFO] - Step 4365, Loss: {'train/loss': 0.06233983114361763, 'train/grad_norm_pre_clip': 55.0513801574707, 'timing/data': 0.00040944002103060484, 'timing/model': 0.39491154201095924, 'learning_rate/qwen_vl_interface': 8.882501471692379e-07, 'learning_rate/action_model': 8.882501471692378e-06, 'epoch': 0.32}) +[2026-10-03 19:49:01,322][starVLA.training.train_starvla][INFO] - Step 4366, Loss: {'train/loss': 8.775213791523129e-05, 'train/grad_norm_pre_clip': 0.03946348652243614, 'timing/data': 0.0003236199845559895, 'timing/model': 0.5990493609861005, 'learning_rate/qwen_vl_interface': 8.870451459311894e-07, 'learning_rate/action_model': 8.870451459311893e-06, 'epoch': 0.32}) +[2026-10-03 19:49:02,030][starVLA.training.train_starvla][INFO] - Step 4367, Loss: {'train/loss': 0.0010026437230408192, 'train/grad_norm_pre_clip': 1.1088547706604004, 'timing/data': 0.0004912200092803687, 'timing/model': 0.39477092199376784, 'learning_rate/qwen_vl_interface': 8.85841938139077e-07, 'learning_rate/action_model': 8.85841938139077e-06, 'epoch': 0.32}) +[2026-10-03 19:49:02,737][starVLA.training.train_starvla][INFO] - Step 4368, Loss: {'train/loss': 0.0006291994941420853, 'train/grad_norm_pre_clip': 0.3900763988494873, 'timing/data': 0.00032990999170579016, 'timing/model': 0.3951557209948078, 'learning_rate/qwen_vl_interface': 8.846405242874955e-07, 'learning_rate/action_model': 8.846405242874954e-06, 'epoch': 0.32}) +[2026-10-03 19:49:03,442][starVLA.training.train_starvla][INFO] - Step 4369, Loss: {'train/loss': 0.15503956377506256, 'train/grad_norm_pre_clip': 122.38745880126953, 'timing/data': 0.0003712100151460618, 'timing/model': 0.39457246198435314, 'learning_rate/qwen_vl_interface': 8.834409048703004e-07, 'learning_rate/action_model': 8.834409048703005e-06, 'epoch': 0.32}) +[2026-10-03 19:49:04,148][starVLA.training.train_starvla][INFO] - Step 4370, Loss: {'train/loss': 1.3105424841342028e-05, 'train/grad_norm_pre_clip': 0.0024922110605984926, 'timing/data': 0.0003203699889127165, 'timing/model': 0.39489135099574924, 'learning_rate/qwen_vl_interface': 8.822430803806098e-07, 'learning_rate/action_model': 8.822430803806097e-06, 'epoch': 0.32}) +[2026-10-03 19:49:04,853][starVLA.training.train_starvla][INFO] - Step 4371, Loss: {'train/loss': 0.2732265591621399, 'train/grad_norm_pre_clip': 73.90706634521484, 'timing/data': 0.00036070001078769565, 'timing/model': 0.3951375119795557, 'learning_rate/qwen_vl_interface': 8.810470513108041e-07, 'learning_rate/action_model': 8.81047051310804e-06, 'epoch': 0.32}) +[2026-10-03 19:49:05,762][starVLA.training.train_starvla][INFO] - Step 4372, Loss: {'train/loss': 6.307836156338453e-05, 'train/grad_norm_pre_clip': 0.01438403595238924, 'timing/data': 0.0003408800112083554, 'timing/model': 0.3950749220093712, 'learning_rate/qwen_vl_interface': 8.798528181525262e-07, 'learning_rate/action_model': 8.798528181525261e-06, 'epoch': 0.32}) +[2026-10-03 19:49:06,469][starVLA.training.train_starvla][INFO] - Step 4373, Loss: {'train/loss': 0.38527366518974304, 'train/grad_norm_pre_clip': 96.7188720703125, 'timing/data': 0.00036499000270850956, 'timing/model': 0.3951429110020399, 'learning_rate/qwen_vl_interface': 8.786603813966798e-07, 'learning_rate/action_model': 8.786603813966798e-06, 'epoch': 0.32}) +[2026-10-03 19:49:07,176][starVLA.training.train_starvla][INFO] - Step 4374, Loss: {'train/loss': 0.0001592650660313666, 'train/grad_norm_pre_clip': 0.09504760056734085, 'timing/data': 0.0003332499763928354, 'timing/model': 0.3947139520023484, 'learning_rate/qwen_vl_interface': 8.774697415334304e-07, 'learning_rate/action_model': 8.774697415334304e-06, 'epoch': 0.32}) +[2026-10-03 19:49:07,881][starVLA.training.train_starvla][INFO] - Step 4375, Loss: {'train/loss': 0.8720591068267822, 'train/grad_norm_pre_clip': 250.36538696289062, 'timing/data': 0.00021893999655731022, 'timing/model': 0.3952118019806221, 'learning_rate/qwen_vl_interface': 8.762808990522056e-07, 'learning_rate/action_model': 8.762808990522057e-06, 'epoch': 0.32}) +[2026-10-03 19:49:08,586][starVLA.training.train_starvla][INFO] - Step 4376, Loss: {'train/loss': 0.03585704043507576, 'train/grad_norm_pre_clip': 76.09541320800781, 'timing/data': 0.0003305099962744862, 'timing/model': 0.39511907199630514, 'learning_rate/qwen_vl_interface': 8.750938544416934e-07, 'learning_rate/action_model': 8.750938544416934e-06, 'epoch': 0.32}) +[2026-10-03 19:49:09,492][starVLA.training.train_starvla][INFO] - Step 4377, Loss: {'train/loss': 0.0005595519905909896, 'train/grad_norm_pre_clip': 0.47232183814048767, 'timing/data': 0.00037951002013869584, 'timing/model': 0.5938526020036079, 'learning_rate/qwen_vl_interface': 8.739086081898437e-07, 'learning_rate/action_model': 8.739086081898437e-06, 'epoch': 0.32}) +[2026-10-03 19:49:10,195][starVLA.training.train_starvla][INFO] - Step 4378, Loss: {'train/loss': 8.74917532200925e-05, 'train/grad_norm_pre_clip': 0.015847912058234215, 'timing/data': 0.0003578800242394209, 'timing/model': 0.3928090510016773, 'learning_rate/qwen_vl_interface': 8.727251607838647e-07, 'learning_rate/action_model': 8.727251607838647e-06, 'epoch': 0.32}) +[2026-10-03 19:49:10,899][starVLA.training.train_starvla][INFO] - Step 4379, Loss: {'train/loss': 0.013333499431610107, 'train/grad_norm_pre_clip': 39.259605407714844, 'timing/data': 0.0002566599869169295, 'timing/model': 0.3928344520099927, 'learning_rate/qwen_vl_interface': 8.715435127102281e-07, 'learning_rate/action_model': 8.715435127102281e-06, 'epoch': 0.32}) +[2026-10-03 19:49:11,601][starVLA.training.train_starvla][INFO] - Step 4380, Loss: {'train/loss': 0.00016144385153893381, 'train/grad_norm_pre_clip': 0.13541944324970245, 'timing/data': 0.00032578001264482737, 'timing/model': 0.3926719519949984, 'learning_rate/qwen_vl_interface': 8.703636644546651e-07, 'learning_rate/action_model': 8.703636644546651e-06, 'epoch': 0.32}) +[2026-10-03 19:49:12,305][starVLA.training.train_starvla][INFO] - Step 4381, Loss: {'train/loss': 0.48545247316360474, 'train/grad_norm_pre_clip': 111.85984802246094, 'timing/data': 0.00035601999843493104, 'timing/model': 0.39313241199124604, 'learning_rate/qwen_vl_interface': 8.69185616502166e-07, 'learning_rate/action_model': 8.69185616502166e-06, 'epoch': 0.32}) +[2026-10-03 19:49:13,214][starVLA.training.train_starvla][INFO] - Step 4382, Loss: {'train/loss': 0.0002542957663536072, 'train/grad_norm_pre_clip': 0.17574669420719147, 'timing/data': 0.0003248100110795349, 'timing/model': 0.5979397220071405, 'learning_rate/qwen_vl_interface': 8.680093693369821e-07, 'learning_rate/action_model': 8.68009369336982e-06, 'epoch': 0.32}) +[2026-10-03 19:49:13,919][starVLA.training.train_starvla][INFO] - Step 4383, Loss: {'train/loss': 0.00024558152654208243, 'train/grad_norm_pre_clip': 0.254379004240036, 'timing/data': 0.00037703997804783285, 'timing/model': 0.39286171199637465, 'learning_rate/qwen_vl_interface': 8.66834923442624e-07, 'learning_rate/action_model': 8.66834923442624e-06, 'epoch': 0.32}) +[2026-10-03 19:49:14,623][starVLA.training.train_starvla][INFO] - Step 4384, Loss: {'train/loss': 0.0036583684850484133, 'train/grad_norm_pre_clip': 4.728716850280762, 'timing/data': 0.0003212599840480834, 'timing/model': 0.3948215320124291, 'learning_rate/qwen_vl_interface': 8.65662279301863e-07, 'learning_rate/action_model': 8.656622793018628e-06, 'epoch': 0.32}) +[2026-10-03 19:49:15,330][starVLA.training.train_starvla][INFO] - Step 4385, Loss: {'train/loss': 0.016274310648441315, 'train/grad_norm_pre_clip': 43.535125732421875, 'timing/data': 0.0003843200101982802, 'timing/model': 0.3935987509903498, 'learning_rate/qwen_vl_interface': 8.644914373967274e-07, 'learning_rate/action_model': 8.644914373967273e-06, 'epoch': 0.32}) +[2026-10-03 19:49:16,035][starVLA.training.train_starvla][INFO] - Step 4386, Loss: {'train/loss': 0.56284499168396, 'train/grad_norm_pre_clip': 122.5053939819336, 'timing/data': 0.000361209997208789, 'timing/model': 0.39455547102261335, 'learning_rate/qwen_vl_interface': 8.633223982085068e-07, 'learning_rate/action_model': 8.633223982085068e-06, 'epoch': 0.32}) +[2026-10-03 19:49:16,739][starVLA.training.train_starvla][INFO] - Step 4387, Loss: {'train/loss': 0.4781518876552582, 'train/grad_norm_pre_clip': 40.48139953613281, 'timing/data': 0.00032815997838042676, 'timing/model': 0.3928231720055919, 'learning_rate/qwen_vl_interface': 8.621551622177499e-07, 'learning_rate/action_model': 8.621551622177499e-06, 'epoch': 0.32}) +[2026-10-03 19:49:17,646][starVLA.training.train_starvla][INFO] - Step 4388, Loss: {'train/loss': 7.943109812913463e-05, 'train/grad_norm_pre_clip': 0.02918098308146, 'timing/data': 0.00033113997778855264, 'timing/model': 0.3939723609946668, 'learning_rate/qwen_vl_interface': 8.609897299042615e-07, 'learning_rate/action_model': 8.609897299042614e-06, 'epoch': 0.32}) +[2026-10-03 19:49:18,350][starVLA.training.train_starvla][INFO] - Step 4389, Loss: {'train/loss': 0.0005470763426274061, 'train/grad_norm_pre_clip': 0.5464783310890198, 'timing/data': 0.000354860007064417, 'timing/model': 0.3944502119848039, 'learning_rate/qwen_vl_interface': 8.598261017471088e-07, 'learning_rate/action_model': 8.598261017471088e-06, 'epoch': 0.32}) +[2026-10-03 19:49:19,054][starVLA.training.train_starvla][INFO] - Step 4390, Loss: {'train/loss': 0.00015258134226314723, 'train/grad_norm_pre_clip': 0.042329538613557816, 'timing/data': 0.00034027997753582895, 'timing/model': 0.39419653100776486, 'learning_rate/qwen_vl_interface': 8.586642782246133e-07, 'learning_rate/action_model': 8.586642782246131e-06, 'epoch': 0.32}) +[2026-10-03 19:49:19,759][starVLA.training.train_starvla][INFO] - Step 4391, Loss: {'train/loss': 0.006846697069704533, 'train/grad_norm_pre_clip': 17.45383644104004, 'timing/data': 0.0003196899779140949, 'timing/model': 0.3946436010010075, 'learning_rate/qwen_vl_interface': 8.57504259814358e-07, 'learning_rate/action_model': 8.57504259814358e-06, 'epoch': 0.32}) +[2026-10-03 19:49:20,464][starVLA.training.train_starvla][INFO] - Step 4392, Loss: {'train/loss': 0.0005074227810837328, 'train/grad_norm_pre_clip': 0.6378100514411926, 'timing/data': 0.00023130999761633575, 'timing/model': 0.39414177200524136, 'learning_rate/qwen_vl_interface': 8.563460469931823e-07, 'learning_rate/action_model': 8.563460469931822e-06, 'epoch': 0.32}) +[2026-10-03 19:49:21,369][starVLA.training.train_starvla][INFO] - Step 4393, Loss: {'train/loss': 0.017534829676151276, 'train/grad_norm_pre_clip': 44.997249603271484, 'timing/data': 0.00022781998268328607, 'timing/model': 0.3928295220248401, 'learning_rate/qwen_vl_interface': 8.551896402371834e-07, 'learning_rate/action_model': 8.551896402371834e-06, 'epoch': 0.32}) +[2026-10-03 19:49:22,072][starVLA.training.train_starvla][INFO] - Step 4394, Loss: {'train/loss': 0.0010925477836281061, 'train/grad_norm_pre_clip': 1.1434186697006226, 'timing/data': 0.0004041000211145729, 'timing/model': 0.3934296419902239, 'learning_rate/qwen_vl_interface': 8.540350400217177e-07, 'learning_rate/action_model': 8.540350400217176e-06, 'epoch': 0.32}) +[2026-10-03 19:49:22,777][starVLA.training.train_starvla][INFO] - Step 4395, Loss: {'train/loss': 0.005611357744783163, 'train/grad_norm_pre_clip': 19.110273361206055, 'timing/data': 0.00022663999698124826, 'timing/model': 0.39426443100092, 'learning_rate/qwen_vl_interface': 8.528822468213964e-07, 'learning_rate/action_model': 8.528822468213964e-06, 'epoch': 0.32}) +[2026-10-03 19:49:23,483][starVLA.training.train_starvla][INFO] - Step 4396, Loss: {'train/loss': 0.0005274967988952994, 'train/grad_norm_pre_clip': 0.3492520749568939, 'timing/data': 0.00048034999053925276, 'timing/model': 0.39404655099497177, 'learning_rate/qwen_vl_interface': 8.517312611100898e-07, 'learning_rate/action_model': 8.517312611100897e-06, 'epoch': 0.32}) +[2026-10-03 19:49:24,188][starVLA.training.train_starvla][INFO] - Step 4397, Loss: {'train/loss': 0.0045713502913713455, 'train/grad_norm_pre_clip': 7.497995853424072, 'timing/data': 0.00032877997728064656, 'timing/model': 0.39341715202317573, 'learning_rate/qwen_vl_interface': 8.505820833609249e-07, 'learning_rate/action_model': 8.50582083360925e-06, 'epoch': 0.32}) +[2026-10-03 19:49:25,094][starVLA.training.train_starvla][INFO] - Step 4398, Loss: {'train/loss': 0.0002054496726486832, 'train/grad_norm_pre_clip': 0.0749906674027443, 'timing/data': 0.00022625000565312803, 'timing/model': 0.5954185519949533, 'learning_rate/qwen_vl_interface': 8.494347140462847e-07, 'learning_rate/action_model': 8.494347140462846e-06, 'epoch': 0.32}) +[2026-10-03 19:49:25,796][starVLA.training.train_starvla][INFO] - Step 4399, Loss: {'train/loss': 0.0004995364579372108, 'train/grad_norm_pre_clip': 0.2536137104034424, 'timing/data': 0.0003286199935246259, 'timing/model': 0.39288026199210435, 'learning_rate/qwen_vl_interface': 8.482891536378105e-07, 'learning_rate/action_model': 8.482891536378105e-06, 'epoch': 0.32}) +[2026-10-03 19:49:26,501][starVLA.training.train_starvla][INFO] - Step 4400, Loss: {'train/loss': 0.1550057828426361, 'train/grad_norm_pre_clip': 115.3203353881836, 'timing/data': 0.0003300399985164404, 'timing/model': 0.3941550419840496, 'learning_rate/qwen_vl_interface': 8.471454026063972e-07, 'learning_rate/action_model': 8.47145402606397e-06, 'epoch': 0.32}) +[2026-10-03 19:49:27,206][starVLA.training.train_starvla][INFO] - Step 4401, Loss: {'train/loss': 0.00019254455401096493, 'train/grad_norm_pre_clip': 0.06537376344203949, 'timing/data': 0.00024148999364115298, 'timing/model': 0.3951388109999243, 'learning_rate/qwen_vl_interface': 8.460034614221988e-07, 'learning_rate/action_model': 8.460034614221988e-06, 'epoch': 0.32}) +[2026-10-03 19:49:27,913][starVLA.training.train_starvla][INFO] - Step 4402, Loss: {'train/loss': 0.40778663754463196, 'train/grad_norm_pre_clip': 40.338687896728516, 'timing/data': 0.0003914299886673689, 'timing/model': 0.39413114098715596, 'learning_rate/qwen_vl_interface': 8.448633305546236e-07, 'learning_rate/action_model': 8.448633305546235e-06, 'epoch': 0.32}) +[2026-10-03 19:49:28,820][starVLA.training.train_starvla][INFO] - Step 4403, Loss: {'train/loss': 0.007843287661671638, 'train/grad_norm_pre_clip': 9.144942283630371, 'timing/data': 0.0004852399870287627, 'timing/model': 0.5943551320233382, 'learning_rate/qwen_vl_interface': 8.437250104723369e-07, 'learning_rate/action_model': 8.437250104723369e-06, 'epoch': 0.32}) +[2026-10-03 19:49:29,526][starVLA.training.train_starvla][INFO] - Step 4404, Loss: {'train/loss': 0.02204830013215542, 'train/grad_norm_pre_clip': 51.94169235229492, 'timing/data': 0.00031883001793175936, 'timing/model': 0.39460911200148985, 'learning_rate/qwen_vl_interface': 8.42588501643259e-07, 'learning_rate/action_model': 8.42588501643259e-06, 'epoch': 0.32}) +[2026-10-03 19:49:30,229][starVLA.training.train_starvla][INFO] - Step 4405, Loss: {'train/loss': 0.018032077699899673, 'train/grad_norm_pre_clip': 34.31309509277344, 'timing/data': 0.0003657200140878558, 'timing/model': 0.3934564520022832, 'learning_rate/qwen_vl_interface': 8.414538045345655e-07, 'learning_rate/action_model': 8.414538045345654e-06, 'epoch': 0.32}) +[2026-10-03 19:49:30,937][starVLA.training.train_starvla][INFO] - Step 4406, Loss: {'train/loss': 0.45979586243629456, 'train/grad_norm_pre_clip': 166.64175415039062, 'timing/data': 0.00026288998196832836, 'timing/model': 0.39574293201440014, 'learning_rate/qwen_vl_interface': 8.403209196126871e-07, 'learning_rate/action_model': 8.403209196126871e-06, 'epoch': 0.32}) +[2026-10-03 19:49:31,643][starVLA.training.train_starvla][INFO] - Step 4407, Loss: {'train/loss': 0.10500726848840714, 'train/grad_norm_pre_clip': 162.98316955566406, 'timing/data': 0.00023345000226981938, 'timing/model': 0.39462451098370366, 'learning_rate/qwen_vl_interface': 8.391898473433111e-07, 'learning_rate/action_model': 8.39189847343311e-06, 'epoch': 0.32}) +[2026-10-03 19:49:32,348][starVLA.training.train_starvla][INFO] - Step 4408, Loss: {'train/loss': 0.0018289610743522644, 'train/grad_norm_pre_clip': 1.8525371551513672, 'timing/data': 0.000328330002957955, 'timing/model': 0.3938580120156985, 'learning_rate/qwen_vl_interface': 8.380605881913778e-07, 'learning_rate/action_model': 8.380605881913779e-06, 'epoch': 0.32}) +[2026-10-03 19:49:33,253][starVLA.training.train_starvla][INFO] - Step 4409, Loss: {'train/loss': 9.887907071970403e-05, 'train/grad_norm_pre_clip': 0.05935024470090866, 'timing/data': 0.00022602002718485892, 'timing/model': 0.3944128819857724, 'learning_rate/qwen_vl_interface': 8.369331426210832e-07, 'learning_rate/action_model': 8.369331426210832e-06, 'epoch': 0.32}) +[2026-10-03 19:49:33,961][starVLA.training.train_starvla][INFO] - Step 4410, Loss: {'train/loss': 0.010749894194304943, 'train/grad_norm_pre_clip': 34.419715881347656, 'timing/data': 0.00022456000442616642, 'timing/model': 0.3964715510082897, 'learning_rate/qwen_vl_interface': 8.358075110958785e-07, 'learning_rate/action_model': 8.358075110958784e-06, 'epoch': 0.32}) +[2026-10-03 19:49:34,669][starVLA.training.train_starvla][INFO] - Step 4411, Loss: {'train/loss': 0.00011954112414969131, 'train/grad_norm_pre_clip': 0.03635013476014137, 'timing/data': 0.00023268000222742558, 'timing/model': 0.3957116309902631, 'learning_rate/qwen_vl_interface': 8.346836940784661e-07, 'learning_rate/action_model': 8.346836940784661e-06, 'epoch': 0.32}) +[2026-10-03 19:49:35,378][starVLA.training.train_starvla][INFO] - Step 4412, Loss: {'train/loss': 4.235150117892772e-05, 'train/grad_norm_pre_clip': 0.013397871516644955, 'timing/data': 0.0003371599887032062, 'timing/model': 0.3954928720195312, 'learning_rate/qwen_vl_interface': 8.335616920308061e-07, 'learning_rate/action_model': 8.33561692030806e-06, 'epoch': 0.32}) +[2026-10-03 19:49:36,084][starVLA.training.train_starvla][INFO] - Step 4413, Loss: {'train/loss': 0.373497873544693, 'train/grad_norm_pre_clip': 30.890836715698242, 'timing/data': 0.0003845000173896551, 'timing/model': 0.39457670200499706, 'learning_rate/qwen_vl_interface': 8.324415054141111e-07, 'learning_rate/action_model': 8.32441505414111e-06, 'epoch': 0.32}) +[2026-10-03 19:49:36,994][starVLA.training.train_starvla][INFO] - Step 4414, Loss: {'train/loss': 0.6369800567626953, 'train/grad_norm_pre_clip': 42.82494354248047, 'timing/data': 0.0004323100147303194, 'timing/model': 0.39596465101931244, 'learning_rate/qwen_vl_interface': 8.313231346888476e-07, 'learning_rate/action_model': 8.313231346888476e-06, 'epoch': 0.32}) +[2026-10-03 19:49:37,701][starVLA.training.train_starvla][INFO] - Step 4415, Loss: {'train/loss': 0.00027613426209427416, 'train/grad_norm_pre_clip': 0.12863977253437042, 'timing/data': 0.00023231000523082912, 'timing/model': 0.39381315201171674, 'learning_rate/qwen_vl_interface': 8.302065803147353e-07, 'learning_rate/action_model': 8.302065803147353e-06, 'epoch': 0.32}) +[2026-10-03 19:49:38,410][starVLA.training.train_starvla][INFO] - Step 4416, Loss: {'train/loss': 0.004770122468471527, 'train/grad_norm_pre_clip': 14.989237785339355, 'timing/data': 0.00034560999483801425, 'timing/model': 0.3957006709824782, 'learning_rate/qwen_vl_interface': 8.29091842750748e-07, 'learning_rate/action_model': 8.29091842750748e-06, 'epoch': 0.32}) +[2026-10-03 19:49:39,118][starVLA.training.train_starvla][INFO] - Step 4417, Loss: {'train/loss': 0.0009204952511936426, 'train/grad_norm_pre_clip': 0.5991413593292236, 'timing/data': 0.00023082000552676618, 'timing/model': 0.39626692098681815, 'learning_rate/qwen_vl_interface': 8.279789224551116e-07, 'learning_rate/action_model': 8.279789224551116e-06, 'epoch': 0.32}) +[2026-10-03 19:49:39,826][starVLA.training.train_starvla][INFO] - Step 4418, Loss: {'train/loss': 0.00036227714736014605, 'train/grad_norm_pre_clip': 0.2290523797273636, 'timing/data': 0.00022851998801343143, 'timing/model': 0.3949290910095442, 'learning_rate/qwen_vl_interface': 8.268678198853058e-07, 'learning_rate/action_model': 8.268678198853059e-06, 'epoch': 0.32}) +[2026-10-03 19:49:40,732][starVLA.training.train_starvla][INFO] - Step 4419, Loss: {'train/loss': 0.0008974920492619276, 'train/grad_norm_pre_clip': 0.5664656162261963, 'timing/data': 0.00030135002452880144, 'timing/model': 0.5950317929964513, 'learning_rate/qwen_vl_interface': 8.257585354980628e-07, 'learning_rate/action_model': 8.257585354980628e-06, 'epoch': 0.32}) +[2026-10-03 19:49:41,442][starVLA.training.train_starvla][INFO] - Step 4420, Loss: {'train/loss': 0.20519432425498962, 'train/grad_norm_pre_clip': 158.38209533691406, 'timing/data': 0.0002957800170406699, 'timing/model': 0.39728502099751495, 'learning_rate/qwen_vl_interface': 8.246510697493681e-07, 'learning_rate/action_model': 8.246510697493682e-06, 'epoch': 0.32}) +[2026-10-03 19:49:42,151][starVLA.training.train_starvla][INFO] - Step 4421, Loss: {'train/loss': 0.0722009688615799, 'train/grad_norm_pre_clip': 106.94020080566406, 'timing/data': 0.0003930300008505583, 'timing/model': 0.39558302200748585, 'learning_rate/qwen_vl_interface': 8.235454230944584e-07, 'learning_rate/action_model': 8.235454230944584e-06, 'epoch': 0.32}) +[2026-10-03 19:49:42,859][starVLA.training.train_starvla][INFO] - Step 4422, Loss: {'train/loss': 0.3209967315196991, 'train/grad_norm_pre_clip': 64.9709243774414, 'timing/data': 0.0003428000200074166, 'timing/model': 0.39576158198178746, 'learning_rate/qwen_vl_interface': 8.224415959878234e-07, 'learning_rate/action_model': 8.224415959878234e-06, 'epoch': 0.32}) +[2026-10-03 19:49:43,567][starVLA.training.train_starvla][INFO] - Step 4423, Loss: {'train/loss': 0.006663590669631958, 'train/grad_norm_pre_clip': 12.160758018493652, 'timing/data': 0.0003254599869251251, 'timing/model': 0.39581390100647695, 'learning_rate/qwen_vl_interface': 8.213395888832052e-07, 'learning_rate/action_model': 8.213395888832051e-06, 'epoch': 0.32}) +[2026-10-03 19:49:44,473][starVLA.training.train_starvla][INFO] - Step 4424, Loss: {'train/loss': 0.09852585196495056, 'train/grad_norm_pre_clip': 108.86475372314453, 'timing/data': 0.0002834699989762157, 'timing/model': 0.595060892024776, 'learning_rate/qwen_vl_interface': 8.202394022335958e-07, 'learning_rate/action_model': 8.202394022335957e-06, 'epoch': 0.32}) +[2026-10-03 19:49:45,180][starVLA.training.train_starvla][INFO] - Step 4425, Loss: {'train/loss': 6.259206566028297e-05, 'train/grad_norm_pre_clip': 0.02789061889052391, 'timing/data': 0.00031863999902270734, 'timing/model': 0.396298070991179, 'learning_rate/qwen_vl_interface': 8.191410364912424e-07, 'learning_rate/action_model': 8.191410364912424e-06, 'epoch': 0.32}) +[2026-10-03 19:49:45,887][starVLA.training.train_starvla][INFO] - Step 4426, Loss: {'train/loss': 0.1362350434064865, 'train/grad_norm_pre_clip': 95.2477035522461, 'timing/data': 0.00037590001011267304, 'timing/model': 0.39428706202306785, 'learning_rate/qwen_vl_interface': 8.1804449210764e-07, 'learning_rate/action_model': 8.1804449210764e-06, 'epoch': 0.32}) +[2026-10-03 19:49:46,597][starVLA.training.train_starvla][INFO] - Step 4427, Loss: {'train/loss': 0.4837610721588135, 'train/grad_norm_pre_clip': 210.54656982421875, 'timing/data': 0.0002639699960127473, 'timing/model': 0.3964083010214381, 'learning_rate/qwen_vl_interface': 8.169497695335377e-07, 'learning_rate/action_model': 8.169497695335376e-06, 'epoch': 0.32}) +[2026-10-03 19:49:47,304][starVLA.training.train_starvla][INFO] - Step 4428, Loss: {'train/loss': 2.9875925974920392e-05, 'train/grad_norm_pre_clip': 0.00706133060157299, 'timing/data': 0.0002193200052715838, 'timing/model': 0.3947092119778972, 'learning_rate/qwen_vl_interface': 8.158568692189342e-07, 'learning_rate/action_model': 8.158568692189342e-06, 'epoch': 0.32}) +[2026-10-03 19:49:48,011][starVLA.training.train_starvla][INFO] - Step 4429, Loss: {'train/loss': 0.06899811327457428, 'train/grad_norm_pre_clip': 70.96832275390625, 'timing/data': 0.0003689700097311288, 'timing/model': 0.39568213099846616, 'learning_rate/qwen_vl_interface': 8.147657916130797e-07, 'learning_rate/action_model': 8.147657916130796e-06, 'epoch': 0.32}) +[2026-10-03 19:49:48,919][starVLA.training.train_starvla][INFO] - Step 4430, Loss: {'train/loss': 0.18241509795188904, 'train/grad_norm_pre_clip': 102.58432006835938, 'timing/data': 0.00034625999978743494, 'timing/model': 0.39547885101637803, 'learning_rate/qwen_vl_interface': 8.136765371644753e-07, 'learning_rate/action_model': 8.136765371644753e-06, 'epoch': 0.32}) +[2026-10-03 19:49:49,627][starVLA.training.train_starvla][INFO] - Step 4431, Loss: {'train/loss': 0.0731717050075531, 'train/grad_norm_pre_clip': 72.07135772705078, 'timing/data': 0.0003475300036370754, 'timing/model': 0.3959688709874172, 'learning_rate/qwen_vl_interface': 8.125891063208713e-07, 'learning_rate/action_model': 8.125891063208712e-06, 'epoch': 0.32}) +[2026-10-03 19:49:50,332][starVLA.training.train_starvla][INFO] - Step 4432, Loss: {'train/loss': 0.00026273546973243356, 'train/grad_norm_pre_clip': 0.13546279072761536, 'timing/data': 0.0003388799959793687, 'timing/model': 0.39391270198393613, 'learning_rate/qwen_vl_interface': 8.115034995292698e-07, 'learning_rate/action_model': 8.115034995292698e-06, 'epoch': 0.32}) +[2026-10-03 19:49:51,039][starVLA.training.train_starvla][INFO] - Step 4433, Loss: {'train/loss': 0.0004251308855600655, 'train/grad_norm_pre_clip': 0.2926393151283264, 'timing/data': 0.00027735999901778996, 'timing/model': 0.3960087610175833, 'learning_rate/qwen_vl_interface': 8.104197172359235e-07, 'learning_rate/action_model': 8.104197172359234e-06, 'epoch': 0.32}) +[2026-10-03 19:49:51,748][starVLA.training.train_starvla][INFO] - Step 4434, Loss: {'train/loss': 0.0010457595344632864, 'train/grad_norm_pre_clip': 0.5066961050033569, 'timing/data': 0.000494420004542917, 'timing/model': 0.3963473909825552, 'learning_rate/qwen_vl_interface': 8.093377598863334e-07, 'learning_rate/action_model': 8.093377598863334e-06, 'epoch': 0.32}) +[2026-10-03 19:49:52,656][starVLA.training.train_starvla][INFO] - Step 4435, Loss: {'train/loss': 0.0002792391460388899, 'train/grad_norm_pre_clip': 0.3010876476764679, 'timing/data': 0.0003310000174678862, 'timing/model': 0.3928955020091962, 'learning_rate/qwen_vl_interface': 8.082576279252519e-07, 'learning_rate/action_model': 8.082576279252518e-06, 'epoch': 0.32}) +[2026-10-03 19:49:53,364][starVLA.training.train_starvla][INFO] - Step 4436, Loss: {'train/loss': 0.007230035960674286, 'train/grad_norm_pre_clip': 47.38405227661133, 'timing/data': 0.0002626000205054879, 'timing/model': 0.3971651419997215, 'learning_rate/qwen_vl_interface': 8.071793217966803e-07, 'learning_rate/action_model': 8.071793217966802e-06, 'epoch': 0.32}) +[2026-10-03 19:49:54,072][starVLA.training.train_starvla][INFO] - Step 4437, Loss: {'train/loss': 3.4501754271332175e-05, 'train/grad_norm_pre_clip': 0.00954457651823759, 'timing/data': 0.0003605899983085692, 'timing/model': 0.39566151102189906, 'learning_rate/qwen_vl_interface': 8.061028419438691e-07, 'learning_rate/action_model': 8.06102841943869e-06, 'epoch': 0.32}) +[2026-10-03 19:49:54,780][starVLA.training.train_starvla][INFO] - Step 4438, Loss: {'train/loss': 3.6579738662112504e-05, 'train/grad_norm_pre_clip': 0.021181948482990265, 'timing/data': 0.0003505100030452013, 'timing/model': 0.39528227198752575, 'learning_rate/qwen_vl_interface': 8.050281888093193e-07, 'learning_rate/action_model': 8.050281888093192e-06, 'epoch': 0.32}) +[2026-10-03 19:49:55,488][starVLA.training.train_starvla][INFO] - Step 4439, Loss: {'train/loss': 0.0003054897242691368, 'train/grad_norm_pre_clip': 0.11608381569385529, 'timing/data': 0.00033794998307712376, 'timing/model': 0.3952989620156586, 'learning_rate/qwen_vl_interface': 8.039553628347794e-07, 'learning_rate/action_model': 8.039553628347793e-06, 'epoch': 0.32}) +[2026-10-03 19:49:56,392][starVLA.training.train_starvla][INFO] - Step 4440, Loss: {'train/loss': 0.006352166645228863, 'train/grad_norm_pre_clip': 8.454010009765625, 'timing/data': 0.00028253000346012414, 'timing/model': 0.5911965720006265, 'learning_rate/qwen_vl_interface': 8.028843644612483e-07, 'learning_rate/action_model': 8.028843644612482e-06, 'epoch': 0.32}) +[2026-10-03 19:49:57,100][starVLA.training.train_starvla][INFO] - Step 4441, Loss: {'train/loss': 0.02350206859409809, 'train/grad_norm_pre_clip': 29.847301483154297, 'timing/data': 0.0003577999887056649, 'timing/model': 0.3945069909968879, 'learning_rate/qwen_vl_interface': 8.01815194128973e-07, 'learning_rate/action_model': 8.01815194128973e-06, 'epoch': 0.33}) +[2026-10-03 19:49:57,807][starVLA.training.train_starvla][INFO] - Step 4442, Loss: {'train/loss': 0.002580805914476514, 'train/grad_norm_pre_clip': 5.126919269561768, 'timing/data': 0.00022515997989103198, 'timing/model': 0.3964047819899861, 'learning_rate/qwen_vl_interface': 8.007478522774481e-07, 'learning_rate/action_model': 8.00747852277448e-06, 'epoch': 0.33}) +[2026-10-03 19:49:58,512][starVLA.training.train_starvla][INFO] - Step 4443, Loss: {'train/loss': 6.319160456769168e-05, 'train/grad_norm_pre_clip': 0.026466740295290947, 'timing/data': 0.000341110018780455, 'timing/model': 0.3948551820067223, 'learning_rate/qwen_vl_interface': 7.996823393454197e-07, 'learning_rate/action_model': 7.996823393454196e-06, 'epoch': 0.33}) +[2026-10-03 19:49:59,220][starVLA.training.train_starvla][INFO] - Step 4444, Loss: {'train/loss': 0.00013699819101020694, 'train/grad_norm_pre_clip': 0.045760318636894226, 'timing/data': 0.0003119800239801407, 'timing/model': 0.39564762098598294, 'learning_rate/qwen_vl_interface': 7.986186557708773e-07, 'learning_rate/action_model': 7.986186557708771e-06, 'epoch': 0.33}) +[2026-10-03 19:50:00,129][starVLA.training.train_starvla][INFO] - Step 4445, Loss: {'train/loss': 0.0002305889647686854, 'train/grad_norm_pre_clip': 0.279319703578949, 'timing/data': 0.00023666000925004482, 'timing/model': 0.5983866519818548, 'learning_rate/qwen_vl_interface': 7.975568019910617e-07, 'learning_rate/action_model': 7.975568019910617e-06, 'epoch': 0.33}) +[2026-10-03 19:50:00,836][starVLA.training.train_starvla][INFO] - Step 4446, Loss: {'train/loss': 0.049754366278648376, 'train/grad_norm_pre_clip': 86.4483871459961, 'timing/data': 0.00022002999321557581, 'timing/model': 0.3944853020075243, 'learning_rate/qwen_vl_interface': 7.964967784424618e-07, 'learning_rate/action_model': 7.964967784424618e-06, 'epoch': 0.33}) +[2026-10-03 19:50:01,544][starVLA.training.train_starvla][INFO] - Step 4447, Loss: {'train/loss': 2.45856172114145e-05, 'train/grad_norm_pre_clip': 0.00967442151159048, 'timing/data': 0.0002239599998574704, 'timing/model': 0.39512730098795146, 'learning_rate/qwen_vl_interface': 7.954385855608129e-07, 'learning_rate/action_model': 7.95438585560813e-06, 'epoch': 0.33}) +[2026-10-03 19:50:02,253][starVLA.training.train_starvla][INFO] - Step 4448, Loss: {'train/loss': 0.0005400560330599546, 'train/grad_norm_pre_clip': 0.5687418580055237, 'timing/data': 0.00025147999986074865, 'timing/model': 0.3959459720063023, 'learning_rate/qwen_vl_interface': 7.943822237810974e-07, 'learning_rate/action_model': 7.943822237810975e-06, 'epoch': 0.33}) +[2026-10-03 19:50:02,962][starVLA.training.train_starvla][INFO] - Step 4449, Loss: {'train/loss': 0.0004076345940120518, 'train/grad_norm_pre_clip': 0.21877793967723846, 'timing/data': 0.000252180005190894, 'timing/model': 0.3967965509800706, 'learning_rate/qwen_vl_interface': 7.93327693537547e-07, 'learning_rate/action_model': 7.93327693537547e-06, 'epoch': 0.33}) +[2026-10-03 19:50:03,670][starVLA.training.train_starvla][INFO] - Step 4450, Loss: {'train/loss': 0.00014090152399148792, 'train/grad_norm_pre_clip': 0.030128270387649536, 'timing/data': 0.0002360199869144708, 'timing/model': 0.3959982419910375, 'learning_rate/qwen_vl_interface': 7.92274995263638e-07, 'learning_rate/action_model': 7.92274995263638e-06, 'epoch': 0.33}) +[2026-10-03 19:50:04,577][starVLA.training.train_starvla][INFO] - Step 4451, Loss: {'train/loss': 0.0002500299015082419, 'train/grad_norm_pre_clip': 0.1934601068496704, 'timing/data': 0.00022789000649936497, 'timing/model': 0.39407565200235695, 'learning_rate/qwen_vl_interface': 7.912241293920958e-07, 'learning_rate/action_model': 7.912241293920959e-06, 'epoch': 0.33}) +[2026-10-03 19:50:05,282][starVLA.training.train_starvla][INFO] - Step 4452, Loss: {'train/loss': 0.39948147535324097, 'train/grad_norm_pre_clip': 96.78549194335938, 'timing/data': 0.00022534001618623734, 'timing/model': 0.39411054202355444, 'learning_rate/qwen_vl_interface': 7.90175096354891e-07, 'learning_rate/action_model': 7.90175096354891e-06, 'epoch': 0.33}) +[2026-10-03 19:50:05,991][starVLA.training.train_starvla][INFO] - Step 4453, Loss: {'train/loss': 0.00010829276288859546, 'train/grad_norm_pre_clip': 0.0632631704211235, 'timing/data': 0.00022986999829299748, 'timing/model': 0.3956030709960032, 'learning_rate/qwen_vl_interface': 7.891278965832431e-07, 'learning_rate/action_model': 7.891278965832432e-06, 'epoch': 0.33}) +[2026-10-03 19:50:06,699][starVLA.training.train_starvla][INFO] - Step 4454, Loss: {'train/loss': 0.000266915449174121, 'train/grad_norm_pre_clip': 0.2193741500377655, 'timing/data': 0.0003225000109523535, 'timing/model': 0.3944539520016406, 'learning_rate/qwen_vl_interface': 7.880825305076138e-07, 'learning_rate/action_model': 7.880825305076139e-06, 'epoch': 0.33}) +[2026-10-03 19:50:07,407][starVLA.training.train_starvla][INFO] - Step 4455, Loss: {'train/loss': 0.00026431630249135196, 'train/grad_norm_pre_clip': 0.16057516634464264, 'timing/data': 0.00024309000582434237, 'timing/model': 0.3973945809993893, 'learning_rate/qwen_vl_interface': 7.870389985577148e-07, 'learning_rate/action_model': 7.870389985577148e-06, 'epoch': 0.33}) +[2026-10-03 19:50:08,313][starVLA.training.train_starvla][INFO] - Step 4456, Loss: {'train/loss': 0.41719362139701843, 'train/grad_norm_pre_clip': 63.98016357421875, 'timing/data': 0.00022663999698124826, 'timing/model': 0.5953006030176766, 'learning_rate/qwen_vl_interface': 7.859973011625028e-07, 'learning_rate/action_model': 7.859973011625026e-06, 'epoch': 0.33}) +[2026-10-03 19:50:09,020][starVLA.training.train_starvla][INFO] - Step 4457, Loss: {'train/loss': 0.0001553975307615474, 'train/grad_norm_pre_clip': 0.09211728721857071, 'timing/data': 0.00022876998991705477, 'timing/model': 0.3964166610094253, 'learning_rate/qwen_vl_interface': 7.849574387501801e-07, 'learning_rate/action_model': 7.8495743875018e-06, 'epoch': 0.33}) +[2026-10-03 19:50:09,727][starVLA.training.train_starvla][INFO] - Step 4458, Loss: {'train/loss': 0.00013304516323842108, 'train/grad_norm_pre_clip': 0.023903168737888336, 'timing/data': 0.00022655000793747604, 'timing/model': 0.39475229202071205, 'learning_rate/qwen_vl_interface': 7.839194117481952e-07, 'learning_rate/action_model': 7.839194117481952e-06, 'epoch': 0.33}) +[2026-10-03 19:50:10,433][starVLA.training.train_starvla][INFO] - Step 4459, Loss: {'train/loss': 0.0002463949494995177, 'train/grad_norm_pre_clip': 0.18811342120170593, 'timing/data': 0.00023187001352198422, 'timing/model': 0.39479400200070813, 'learning_rate/qwen_vl_interface': 7.828832205832418e-07, 'learning_rate/action_model': 7.828832205832418e-06, 'epoch': 0.33}) +[2026-10-03 19:50:11,135][starVLA.training.train_starvla][INFO] - Step 4460, Loss: {'train/loss': 0.2279323786497116, 'train/grad_norm_pre_clip': 164.85755920410156, 'timing/data': 0.00023556000087410212, 'timing/model': 0.3930446119920816, 'learning_rate/qwen_vl_interface': 7.818488656812581e-07, 'learning_rate/action_model': 7.818488656812582e-06, 'epoch': 0.33}) +[2026-10-03 19:50:12,047][starVLA.training.train_starvla][INFO] - Step 4461, Loss: {'train/loss': 2.3833779778215103e-05, 'train/grad_norm_pre_clip': 0.005039994139224291, 'timing/data': 0.00022252000053413212, 'timing/model': 0.5998435619985685, 'learning_rate/qwen_vl_interface': 7.808163474674291e-07, 'learning_rate/action_model': 7.80816347467429e-06, 'epoch': 0.33}) +[2026-10-03 19:50:12,752][starVLA.training.train_starvla][INFO] - Step 4462, Loss: {'train/loss': 0.0006396239041350782, 'train/grad_norm_pre_clip': 0.9569821357727051, 'timing/data': 0.00031046001822687685, 'timing/model': 0.3935403309878893, 'learning_rate/qwen_vl_interface': 7.797856663661836e-07, 'learning_rate/action_model': 7.797856663661835e-06, 'epoch': 0.33}) +[2026-10-03 19:50:13,457][starVLA.training.train_starvla][INFO] - Step 4463, Loss: {'train/loss': 0.4591355621814728, 'train/grad_norm_pre_clip': 54.03703308105469, 'timing/data': 0.0002352199808228761, 'timing/model': 0.39544993100571446, 'learning_rate/qwen_vl_interface': 7.787568228011959e-07, 'learning_rate/action_model': 7.78756822801196e-06, 'epoch': 0.33}) +[2026-10-03 19:50:14,164][starVLA.training.train_starvla][INFO] - Step 4464, Loss: {'train/loss': 0.00010985061817336828, 'train/grad_norm_pre_clip': 0.04951269179582596, 'timing/data': 0.00046431997907347977, 'timing/model': 0.39478670200333, 'learning_rate/qwen_vl_interface': 7.777298171953858e-07, 'learning_rate/action_model': 7.777298171953857e-06, 'epoch': 0.33}) +[2026-10-03 19:50:14,869][starVLA.training.train_starvla][INFO] - Step 4465, Loss: {'train/loss': 6.411060167010874e-05, 'train/grad_norm_pre_clip': 0.024518132209777832, 'timing/data': 0.00022586999693885446, 'timing/model': 0.3948125409951899, 'learning_rate/qwen_vl_interface': 7.767046499709144e-07, 'learning_rate/action_model': 7.767046499709143e-06, 'epoch': 0.33}) +[2026-10-03 19:50:15,549][starVLA.training.train_starvla][INFO] - Step 4466, Loss: {'train/loss': 0.33600980043411255, 'train/grad_norm_pre_clip': 47.5692024230957, 'timing/data': 0.0004777300055138767, 'timing/model': 0.3945174010004848, 'learning_rate/qwen_vl_interface': 7.756813215491904e-07, 'learning_rate/action_model': 7.756813215491904e-06, 'epoch': 0.33}) +[2026-10-03 19:50:16,452][starVLA.training.train_starvla][INFO] - Step 4467, Loss: {'train/loss': 0.00019486932433210313, 'train/grad_norm_pre_clip': 0.12889811396598816, 'timing/data': 0.0004408499808050692, 'timing/model': 0.39426919200923294, 'learning_rate/qwen_vl_interface': 7.746598323508655e-07, 'learning_rate/action_model': 7.746598323508655e-06, 'epoch': 0.33}) +[2026-10-03 19:50:17,158][starVLA.training.train_starvla][INFO] - Step 4468, Loss: {'train/loss': 0.00012290179438423365, 'train/grad_norm_pre_clip': 0.05054566264152527, 'timing/data': 0.00034806999610736966, 'timing/model': 0.39534034198732115, 'learning_rate/qwen_vl_interface': 7.736401827958351e-07, 'learning_rate/action_model': 7.73640182795835e-06, 'epoch': 0.33}) +[2026-10-03 19:50:17,862][starVLA.training.train_starvla][INFO] - Step 4469, Loss: {'train/loss': 0.0004117100907023996, 'train/grad_norm_pre_clip': 0.3560031056404114, 'timing/data': 0.00037546001840382814, 'timing/model': 0.3934901010070462, 'learning_rate/qwen_vl_interface': 7.726223733032382e-07, 'learning_rate/action_model': 7.72622373303238e-06, 'epoch': 0.33}) +[2026-10-03 19:50:18,567][starVLA.training.train_starvla][INFO] - Step 4470, Loss: {'train/loss': 0.00015503645408898592, 'train/grad_norm_pre_clip': 0.06288839876651764, 'timing/data': 0.0003528600209392607, 'timing/model': 0.3929937719949521, 'learning_rate/qwen_vl_interface': 7.716064042914584e-07, 'learning_rate/action_model': 7.716064042914584e-06, 'epoch': 0.33}) +[2026-10-03 19:50:19,274][starVLA.training.train_starvla][INFO] - Step 4471, Loss: {'train/loss': 0.2114410549402237, 'train/grad_norm_pre_clip': 130.00071716308594, 'timing/data': 0.00024966002092696726, 'timing/model': 0.39498503200593404, 'learning_rate/qwen_vl_interface': 7.705922761781226e-07, 'learning_rate/action_model': 7.705922761781226e-06, 'epoch': 0.33}) +[2026-10-03 19:50:20,184][starVLA.training.train_starvla][INFO] - Step 4472, Loss: {'train/loss': 0.38342899084091187, 'train/grad_norm_pre_clip': 143.7147979736328, 'timing/data': 0.00028219001251272857, 'timing/model': 0.5971998719905969, 'learning_rate/qwen_vl_interface': 7.695799893800996e-07, 'learning_rate/action_model': 7.695799893800996e-06, 'epoch': 0.33}) +[2026-10-03 19:50:20,888][starVLA.training.train_starvla][INFO] - Step 4473, Loss: {'train/loss': 0.06470475345849991, 'train/grad_norm_pre_clip': 215.0840301513672, 'timing/data': 0.0002690599940251559, 'timing/model': 0.3948019219969865, 'learning_rate/qwen_vl_interface': 7.685695443135028e-07, 'learning_rate/action_model': 7.685695443135027e-06, 'epoch': 0.33}) +[2026-10-03 19:50:21,596][starVLA.training.train_starvla][INFO] - Step 4474, Loss: {'train/loss': 3.687843491206877e-05, 'train/grad_norm_pre_clip': 0.014806878753006458, 'timing/data': 0.000265179987763986, 'timing/model': 0.3956089109997265, 'learning_rate/qwen_vl_interface': 7.675609413936888e-07, 'learning_rate/action_model': 7.675609413936888e-06, 'epoch': 0.33}) +[2026-10-03 19:50:22,303][starVLA.training.train_starvla][INFO] - Step 4475, Loss: {'train/loss': 0.00020877289352938533, 'train/grad_norm_pre_clip': 0.11065857857465744, 'timing/data': 0.000274710007943213, 'timing/model': 0.3938380620093085, 'learning_rate/qwen_vl_interface': 7.665541810352541e-07, 'learning_rate/action_model': 7.66554181035254e-06, 'epoch': 0.33}) +[2026-10-03 19:50:23,008][starVLA.training.train_starvla][INFO] - Step 4476, Loss: {'train/loss': 0.3730262815952301, 'train/grad_norm_pre_clip': 209.80181884765625, 'timing/data': 0.000357550015905872, 'timing/model': 0.3945159319846425, 'learning_rate/qwen_vl_interface': 7.655492636520415e-07, 'learning_rate/action_model': 7.655492636520414e-06, 'epoch': 0.33}) +[2026-10-03 19:50:23,914][starVLA.training.train_starvla][INFO] - Step 4477, Loss: {'train/loss': 0.5496792793273926, 'train/grad_norm_pre_clip': 62.62570571899414, 'timing/data': 0.00021600999752990901, 'timing/model': 0.5944343620212749, 'learning_rate/qwen_vl_interface': 7.64546189657134e-07, 'learning_rate/action_model': 7.645461896571339e-06, 'epoch': 0.33}) +[2026-10-03 19:50:24,619][starVLA.training.train_starvla][INFO] - Step 4478, Loss: {'train/loss': 0.0007544634863734245, 'train/grad_norm_pre_clip': 0.5260373950004578, 'timing/data': 0.00035493000177666545, 'timing/model': 0.3941734409891069, 'learning_rate/qwen_vl_interface': 7.635449594628571e-07, 'learning_rate/action_model': 7.635449594628571e-06, 'epoch': 0.33}) +[2026-10-03 19:50:25,326][starVLA.training.train_starvla][INFO] - Step 4479, Loss: {'train/loss': 0.02854849025607109, 'train/grad_norm_pre_clip': 83.5508041381836, 'timing/data': 0.00029511001775972545, 'timing/model': 0.39431084200623445, 'learning_rate/qwen_vl_interface': 7.625455734807803e-07, 'learning_rate/action_model': 7.625455734807803e-06, 'epoch': 0.33}) +[2026-10-03 19:50:26,030][starVLA.training.train_starvla][INFO] - Step 4480, Loss: {'train/loss': 0.369545578956604, 'train/grad_norm_pre_clip': 111.93547821044922, 'timing/data': 0.0002539200067985803, 'timing/model': 0.3938842710049357, 'learning_rate/qwen_vl_interface': 7.615480321217116e-07, 'learning_rate/action_model': 7.615480321217117e-06, 'epoch': 0.33}) +[2026-10-03 19:50:26,735][starVLA.training.train_starvla][INFO] - Step 4481, Loss: {'train/loss': 0.00013748723722528666, 'train/grad_norm_pre_clip': 0.05105801671743393, 'timing/data': 0.00022285000886768103, 'timing/model': 0.3927928420016542, 'learning_rate/qwen_vl_interface': 7.605523357957028e-07, 'learning_rate/action_model': 7.605523357957028e-06, 'epoch': 0.33}) +[2026-10-03 19:50:27,439][starVLA.training.train_starvla][INFO] - Step 4482, Loss: {'train/loss': 4.0394821553491056e-05, 'train/grad_norm_pre_clip': 0.009829080663621426, 'timing/data': 0.0003487799840513617, 'timing/model': 0.39443155101616867, 'learning_rate/qwen_vl_interface': 7.59558484912048e-07, 'learning_rate/action_model': 7.59558484912048e-06, 'epoch': 0.33}) +[2026-10-03 19:50:28,345][starVLA.training.train_starvla][INFO] - Step 4483, Loss: {'train/loss': 0.13475193083286285, 'train/grad_norm_pre_clip': 98.9533920288086, 'timing/data': 0.00028577999910339713, 'timing/model': 0.3949050210067071, 'learning_rate/qwen_vl_interface': 7.585664798792807e-07, 'learning_rate/action_model': 7.585664798792807e-06, 'epoch': 0.33}) +[2026-10-03 19:50:29,051][starVLA.training.train_starvla][INFO] - Step 4484, Loss: {'train/loss': 0.0017160087591037154, 'train/grad_norm_pre_clip': 3.948793411254883, 'timing/data': 0.00033389998134225607, 'timing/model': 0.3954766719834879, 'learning_rate/qwen_vl_interface': 7.575763211051779e-07, 'learning_rate/action_model': 7.575763211051778e-06, 'epoch': 0.33}) +[2026-10-03 19:50:29,758][starVLA.training.train_starvla][INFO] - Step 4485, Loss: {'train/loss': 0.3698839843273163, 'train/grad_norm_pre_clip': 43.33955764770508, 'timing/data': 0.00023168002371676266, 'timing/model': 0.3945564920140896, 'learning_rate/qwen_vl_interface': 7.565880089967553e-07, 'learning_rate/action_model': 7.565880089967553e-06, 'epoch': 0.33}) +[2026-10-03 19:50:30,462][starVLA.training.train_starvla][INFO] - Step 4486, Loss: {'train/loss': 0.29595258831977844, 'train/grad_norm_pre_clip': 123.5428695678711, 'timing/data': 0.00022494999575428665, 'timing/model': 0.39286017199628986, 'learning_rate/qwen_vl_interface': 7.556015439602725e-07, 'learning_rate/action_model': 7.556015439602725e-06, 'epoch': 0.33}) +[2026-10-03 19:50:31,169][starVLA.training.train_starvla][INFO] - Step 4487, Loss: {'train/loss': 2.6917692593997344e-05, 'train/grad_norm_pre_clip': 0.007952158339321613, 'timing/data': 0.0003460099978838116, 'timing/model': 0.39480057201581076, 'learning_rate/qwen_vl_interface': 7.546169264012252e-07, 'learning_rate/action_model': 7.5461692640122525e-06, 'epoch': 0.33}) +[2026-10-03 19:50:32,075][starVLA.training.train_starvla][INFO] - Step 4488, Loss: {'train/loss': 0.00013141371891833842, 'train/grad_norm_pre_clip': 0.040619295090436935, 'timing/data': 0.00023344997316598892, 'timing/model': 0.3935608920000959, 'learning_rate/qwen_vl_interface': 7.536341567243544e-07, 'learning_rate/action_model': 7.536341567243543e-06, 'epoch': 0.33}) +[2026-10-03 19:50:32,781][starVLA.training.train_starvla][INFO] - Step 4489, Loss: {'train/loss': 0.00013481402129400522, 'train/grad_norm_pre_clip': 0.04502221569418907, 'timing/data': 0.00038949999725446105, 'timing/model': 0.39360046200454235, 'learning_rate/qwen_vl_interface': 7.526532353336395e-07, 'learning_rate/action_model': 7.526532353336395e-06, 'epoch': 0.33}) +[2026-10-03 19:50:33,486][starVLA.training.train_starvla][INFO] - Step 4490, Loss: {'train/loss': 0.5729933977127075, 'train/grad_norm_pre_clip': 66.4122543334961, 'timing/data': 0.00037220001104287803, 'timing/model': 0.39303174198721536, 'learning_rate/qwen_vl_interface': 7.516741626322994e-07, 'learning_rate/action_model': 7.516741626322994e-06, 'epoch': 0.33}) +[2026-10-03 19:50:34,192][starVLA.training.train_starvla][INFO] - Step 4491, Loss: {'train/loss': 0.2095661610364914, 'train/grad_norm_pre_clip': 93.70201873779297, 'timing/data': 0.00034072998096235096, 'timing/model': 0.39442760101519525, 'learning_rate/qwen_vl_interface': 7.506969390227945e-07, 'learning_rate/action_model': 7.506969390227945e-06, 'epoch': 0.33}) +[2026-10-03 19:50:34,897][starVLA.training.train_starvla][INFO] - Step 4492, Loss: {'train/loss': 0.00038689811481162906, 'train/grad_norm_pre_clip': 0.5996798276901245, 'timing/data': 0.000324328982969746, 'timing/model': 0.3947439120092895, 'learning_rate/qwen_vl_interface': 7.497215649068244e-07, 'learning_rate/action_model': 7.497215649068244e-06, 'epoch': 0.33}) +[2026-10-03 19:50:35,801][starVLA.training.train_starvla][INFO] - Step 4493, Loss: {'train/loss': 0.35883986949920654, 'train/grad_norm_pre_clip': 30.539216995239258, 'timing/data': 0.00039860999095253646, 'timing/model': 0.5925093819969334, 'learning_rate/qwen_vl_interface': 7.48748040685329e-07, 'learning_rate/action_model': 7.487480406853289e-06, 'epoch': 0.33}) +[2026-10-03 19:50:36,506][starVLA.training.train_starvla][INFO] - Step 4494, Loss: {'train/loss': 0.07263394445180893, 'train/grad_norm_pre_clip': 59.866432189941406, 'timing/data': 0.00048283999785780907, 'timing/model': 0.39454028199543245, 'learning_rate/qwen_vl_interface': 7.477763667584867e-07, 'learning_rate/action_model': 7.477763667584866e-06, 'epoch': 0.33}) +[2026-10-03 19:50:37,213][starVLA.training.train_starvla][INFO] - Step 4495, Loss: {'train/loss': 9.612251596990973e-05, 'train/grad_norm_pre_clip': 0.028155028820037842, 'timing/data': 0.00040421998710371554, 'timing/model': 0.394785751006566, 'learning_rate/qwen_vl_interface': 7.468065435257159e-07, 'learning_rate/action_model': 7.468065435257159e-06, 'epoch': 0.33}) +[2026-10-03 19:50:37,917][starVLA.training.train_starvla][INFO] - Step 4496, Loss: {'train/loss': 0.012143898755311966, 'train/grad_norm_pre_clip': 18.850727081298828, 'timing/data': 0.00023372998111881316, 'timing/model': 0.3944066020194441, 'learning_rate/qwen_vl_interface': 7.458385713856745e-07, 'learning_rate/action_model': 7.458385713856744e-06, 'epoch': 0.33}) +[2026-10-03 19:50:38,624][starVLA.training.train_starvla][INFO] - Step 4497, Loss: {'train/loss': 3.6745106626767665e-05, 'train/grad_norm_pre_clip': 0.007358130998909473, 'timing/data': 0.0002317100006621331, 'timing/model': 0.39517788199009374, 'learning_rate/qwen_vl_interface': 7.448724507362605e-07, 'learning_rate/action_model': 7.448724507362605e-06, 'epoch': 0.33}) +[2026-10-03 19:50:39,530][starVLA.training.train_starvla][INFO] - Step 4498, Loss: {'train/loss': 4.032005381304771e-05, 'train/grad_norm_pre_clip': 0.009087181650102139, 'timing/data': 0.0002251800033263862, 'timing/model': 0.5947436429851223, 'learning_rate/qwen_vl_interface': 7.439081819746078e-07, 'learning_rate/action_model': 7.4390818197460775e-06, 'epoch': 0.33}) +[2026-10-03 19:50:40,236][starVLA.training.train_starvla][INFO] - Step 4499, Loss: {'train/loss': 8.740639896132052e-05, 'train/grad_norm_pre_clip': 0.029948310926556587, 'timing/data': 0.0003651099978014827, 'timing/model': 0.3941580319951754, 'learning_rate/qwen_vl_interface': 7.42945765497091e-07, 'learning_rate/action_model': 7.42945765497091e-06, 'epoch': 0.33}) +[2026-10-03 19:50:40,941][starVLA.training.train_starvla][INFO] - Step 4500, Loss: {'train/loss': 3.4598524507600814e-05, 'train/grad_norm_pre_clip': 0.009342645294964314, 'timing/data': 0.0003353299980517477, 'timing/model': 0.3928792920196429, 'learning_rate/qwen_vl_interface': 7.41985201699324e-07, 'learning_rate/action_model': 7.41985201699324e-06, 'epoch': 0.33}) +[2026-10-03 19:50:41,648][starVLA.training.train_starvla][INFO] - Step 4501, Loss: {'train/loss': 0.027968836948275566, 'train/grad_norm_pre_clip': 44.8998908996582, 'timing/data': 0.00037420002627186477, 'timing/model': 0.3949556620209478, 'learning_rate/qwen_vl_interface': 7.410264909761581e-07, 'learning_rate/action_model': 7.41026490976158e-06, 'epoch': 0.33}) +[2026-10-03 19:50:42,353][starVLA.training.train_starvla][INFO] - Step 4502, Loss: {'train/loss': 0.0015398747054859996, 'train/grad_norm_pre_clip': 1.5316640138626099, 'timing/data': 0.00039420000393874943, 'timing/model': 0.3929635619861074, 'learning_rate/qwen_vl_interface': 7.40069633721683e-07, 'learning_rate/action_model': 7.400696337216829e-06, 'epoch': 0.33}) +[2026-10-03 19:50:43,058][starVLA.training.train_starvla][INFO] - Step 4503, Loss: {'train/loss': 0.0003402883594390005, 'train/grad_norm_pre_clip': 0.2388886958360672, 'timing/data': 0.00033776002237573266, 'timing/model': 0.3941351019893773, 'learning_rate/qwen_vl_interface': 7.391146303292262e-07, 'learning_rate/action_model': 7.391146303292262e-06, 'epoch': 0.33}) +[2026-10-03 19:50:43,964][starVLA.training.train_starvla][INFO] - Step 4504, Loss: {'train/loss': 0.005538955330848694, 'train/grad_norm_pre_clip': 9.532195091247559, 'timing/data': 0.00033236900344491005, 'timing/model': 0.39405425201402977, 'learning_rate/qwen_vl_interface': 7.381614811913538e-07, 'learning_rate/action_model': 7.381614811913537e-06, 'epoch': 0.33}) +[2026-10-03 19:50:44,667][starVLA.training.train_starvla][INFO] - Step 4505, Loss: {'train/loss': 0.00037148394039832056, 'train/grad_norm_pre_clip': 0.19998979568481445, 'timing/data': 0.0003971800033468753, 'timing/model': 0.3938051819859538, 'learning_rate/qwen_vl_interface': 7.372101866998693e-07, 'learning_rate/action_model': 7.372101866998693e-06, 'epoch': 0.33}) +[2026-10-03 19:50:45,372][starVLA.training.train_starvla][INFO] - Step 4506, Loss: {'train/loss': 0.508626401424408, 'train/grad_norm_pre_clip': 69.11605072021484, 'timing/data': 0.000352019997080788, 'timing/model': 0.3938633909856435, 'learning_rate/qwen_vl_interface': 7.362607472458147e-07, 'learning_rate/action_model': 7.362607472458147e-06, 'epoch': 0.33}) +[2026-10-03 19:50:46,080][starVLA.training.train_starvla][INFO] - Step 4507, Loss: {'train/loss': 0.011166608892381191, 'train/grad_norm_pre_clip': 15.514880180358887, 'timing/data': 0.00023196000256575644, 'timing/model': 0.3936368419963401, 'learning_rate/qwen_vl_interface': 7.353131632194682e-07, 'learning_rate/action_model': 7.3531316321946816e-06, 'epoch': 0.33}) +[2026-10-03 19:50:46,785][starVLA.training.train_starvla][INFO] - Step 4508, Loss: {'train/loss': 0.0032067573629319668, 'train/grad_norm_pre_clip': 7.259341716766357, 'timing/data': 0.00023115999647416174, 'timing/model': 0.3941111720050685, 'learning_rate/qwen_vl_interface': 7.343674350103455e-07, 'learning_rate/action_model': 7.343674350103455e-06, 'epoch': 0.33}) +[2026-10-03 19:50:47,692][starVLA.training.train_starvla][INFO] - Step 4509, Loss: {'train/loss': 0.001542324898764491, 'train/grad_norm_pre_clip': 2.9017999172210693, 'timing/data': 0.00023107999004423618, 'timing/model': 0.3943282119871583, 'learning_rate/qwen_vl_interface': 7.334235630071995e-07, 'learning_rate/action_model': 7.334235630071995e-06, 'epoch': 0.33}) +[2026-10-03 19:50:48,397][starVLA.training.train_starvla][INFO] - Step 4510, Loss: {'train/loss': 0.3783162832260132, 'train/grad_norm_pre_clip': 104.86860656738281, 'timing/data': 0.00022682998678646982, 'timing/model': 0.39332856200053357, 'learning_rate/qwen_vl_interface': 7.324815475980207e-07, 'learning_rate/action_model': 7.324815475980207e-06, 'epoch': 0.33}) +[2026-10-03 19:50:49,102][starVLA.training.train_starvla][INFO] - Step 4511, Loss: {'train/loss': 1.8402626665192656e-05, 'train/grad_norm_pre_clip': 0.0035695726983249187, 'timing/data': 0.0003631699946708977, 'timing/model': 0.39421692202449776, 'learning_rate/qwen_vl_interface': 7.315413891700363e-07, 'learning_rate/action_model': 7.315413891700362e-06, 'epoch': 0.33}) +[2026-10-03 19:50:49,804][starVLA.training.train_starvla][INFO] - Step 4512, Loss: {'train/loss': 0.00019995577167719603, 'train/grad_norm_pre_clip': 0.13723871111869812, 'timing/data': 0.0003208399866707623, 'timing/model': 0.39255348200094886, 'learning_rate/qwen_vl_interface': 7.306030881097092e-07, 'learning_rate/action_model': 7.3060308810970925e-06, 'epoch': 0.33}) +[2026-10-03 19:50:50,507][starVLA.training.train_starvla][INFO] - Step 4513, Loss: {'train/loss': 0.07992153614759445, 'train/grad_norm_pre_clip': 64.7684555053711, 'timing/data': 0.00038628000766038895, 'timing/model': 0.39220226198085584, 'learning_rate/qwen_vl_interface': 7.296666448027406e-07, 'learning_rate/action_model': 7.296666448027406e-06, 'epoch': 0.33}) +[2026-10-03 19:50:51,418][starVLA.training.train_starvla][INFO] - Step 4514, Loss: {'train/loss': 0.0277686957269907, 'train/grad_norm_pre_clip': 44.805381774902344, 'timing/data': 0.00028299001860432327, 'timing/model': 0.5992956119880546, 'learning_rate/qwen_vl_interface': 7.287320596340664e-07, 'learning_rate/action_model': 7.287320596340663e-06, 'epoch': 0.33}) +[2026-10-03 19:50:52,120][starVLA.training.train_starvla][INFO] - Step 4515, Loss: {'train/loss': 0.10736909508705139, 'train/grad_norm_pre_clip': 162.4440460205078, 'timing/data': 0.0003285300044808537, 'timing/model': 0.3922882020124234, 'learning_rate/qwen_vl_interface': 7.277993329878593e-07, 'learning_rate/action_model': 7.277993329878594e-06, 'epoch': 0.33}) +[2026-10-03 19:50:52,826][starVLA.training.train_starvla][INFO] - Step 4516, Loss: {'train/loss': 0.0036129264626652002, 'train/grad_norm_pre_clip': 6.517487049102783, 'timing/data': 0.00030463997973129153, 'timing/model': 0.39436760201351717, 'learning_rate/qwen_vl_interface': 7.268684652475277e-07, 'learning_rate/action_model': 7.268684652475277e-06, 'epoch': 0.33}) +[2026-10-03 19:50:53,531][starVLA.training.train_starvla][INFO] - Step 4517, Loss: {'train/loss': 2.6567884560790844e-05, 'train/grad_norm_pre_clip': 0.006446118466556072, 'timing/data': 0.0003669300058390945, 'timing/model': 0.3925370519864373, 'learning_rate/qwen_vl_interface': 7.259394567957169e-07, 'learning_rate/action_model': 7.259394567957169e-06, 'epoch': 0.33}) +[2026-10-03 19:50:54,233][starVLA.training.train_starvla][INFO] - Step 4518, Loss: {'train/loss': 0.09199640154838562, 'train/grad_norm_pre_clip': 120.4047622680664, 'timing/data': 0.0003406899922993034, 'timing/model': 0.39344478098792024, 'learning_rate/qwen_vl_interface': 7.250123080143066e-07, 'learning_rate/action_model': 7.2501230801430665e-06, 'epoch': 0.33}) +[2026-10-03 19:50:55,137][starVLA.training.train_starvla][INFO] - Step 4519, Loss: {'train/loss': 0.0008892606711015105, 'train/grad_norm_pre_clip': 0.7857425212860107, 'timing/data': 0.0002950100169982761, 'timing/model': 0.5940498110139742, 'learning_rate/qwen_vl_interface': 7.240870192844127e-07, 'learning_rate/action_model': 7.240870192844127e-06, 'epoch': 0.33}) +[2026-10-03 19:50:55,839][starVLA.training.train_starvla][INFO] - Step 4520, Loss: {'train/loss': 9.025494364323094e-05, 'train/grad_norm_pre_clip': 0.02341667376458645, 'timing/data': 0.00022437001462094486, 'timing/model': 0.3917520319810137, 'learning_rate/qwen_vl_interface': 7.231635909863858e-07, 'learning_rate/action_model': 7.2316359098638585e-06, 'epoch': 0.33}) +[2026-10-03 19:50:56,543][starVLA.training.train_starvla][INFO] - Step 4521, Loss: {'train/loss': 0.041930340230464935, 'train/grad_norm_pre_clip': 75.18946838378906, 'timing/data': 0.00035416000173427165, 'timing/model': 0.392799531982746, 'learning_rate/qwen_vl_interface': 7.222420234998138e-07, 'learning_rate/action_model': 7.2224202349981385e-06, 'epoch': 0.33}) +[2026-10-03 19:50:57,247][starVLA.training.train_starvla][INFO] - Step 4522, Loss: {'train/loss': 0.17751623690128326, 'train/grad_norm_pre_clip': 104.5130386352539, 'timing/data': 0.00032724998891353607, 'timing/model': 0.3939769210119266, 'learning_rate/qwen_vl_interface': 7.213223172035175e-07, 'learning_rate/action_model': 7.213223172035174e-06, 'epoch': 0.33}) +[2026-10-03 19:50:57,952][starVLA.training.train_starvla][INFO] - Step 4523, Loss: {'train/loss': 0.08954397588968277, 'train/grad_norm_pre_clip': 111.36624145507812, 'timing/data': 0.0003444800095167011, 'timing/model': 0.39372531199478544, 'learning_rate/qwen_vl_interface': 7.204044724755528e-07, 'learning_rate/action_model': 7.204044724755528e-06, 'epoch': 0.33}) +[2026-10-03 19:50:58,656][starVLA.training.train_starvla][INFO] - Step 4524, Loss: {'train/loss': 0.1719682365655899, 'train/grad_norm_pre_clip': 69.47989654541016, 'timing/data': 0.0002467099984642118, 'timing/model': 0.39429944098810665, 'learning_rate/qwen_vl_interface': 7.194884896932132e-07, 'learning_rate/action_model': 7.194884896932132e-06, 'epoch': 0.33}) +[2026-10-03 19:50:59,566][starVLA.training.train_starvla][INFO] - Step 4525, Loss: {'train/loss': 0.00010460427438374609, 'train/grad_norm_pre_clip': 0.04029197245836258, 'timing/data': 0.00037846999475732446, 'timing/model': 0.3939529309864156, 'learning_rate/qwen_vl_interface': 7.185743692330221e-07, 'learning_rate/action_model': 7.185743692330221e-06, 'epoch': 0.33}) +[2026-10-03 19:51:00,272][starVLA.training.train_starvla][INFO] - Step 4526, Loss: {'train/loss': 4.876915772911161e-05, 'train/grad_norm_pre_clip': 0.010751570574939251, 'timing/data': 0.0002837599895428866, 'timing/model': 0.3942766419786494, 'learning_rate/qwen_vl_interface': 7.176621114707408e-07, 'learning_rate/action_model': 7.176621114707407e-06, 'epoch': 0.33}) +[2026-10-03 19:51:00,975][starVLA.training.train_starvla][INFO] - Step 4527, Loss: {'train/loss': 8.416196214966476e-05, 'train/grad_norm_pre_clip': 0.03825636953115463, 'timing/data': 0.00022451000404544175, 'timing/model': 0.39171742100734264, 'learning_rate/qwen_vl_interface': 7.167517167813642e-07, 'learning_rate/action_model': 7.167517167813642e-06, 'epoch': 0.33}) +[2026-10-03 19:51:01,679][starVLA.training.train_starvla][INFO] - Step 4528, Loss: {'train/loss': 3.5038450732827187e-05, 'train/grad_norm_pre_clip': 0.00878182239830494, 'timing/data': 0.00028593000024557114, 'timing/model': 0.3936410620226525, 'learning_rate/qwen_vl_interface': 7.15843185539121e-07, 'learning_rate/action_model': 7.15843185539121e-06, 'epoch': 0.33}) +[2026-10-03 19:51:02,385][starVLA.training.train_starvla][INFO] - Step 4529, Loss: {'train/loss': 0.5975102782249451, 'train/grad_norm_pre_clip': 69.95391845703125, 'timing/data': 0.00022990998695604503, 'timing/model': 0.39443781101726927, 'learning_rate/qwen_vl_interface': 7.149365181174739e-07, 'learning_rate/action_model': 7.149365181174739e-06, 'epoch': 0.33}) +[2026-10-03 19:51:03,289][starVLA.training.train_starvla][INFO] - Step 4530, Loss: {'train/loss': 5.4731008276576176e-05, 'train/grad_norm_pre_clip': 0.024134857580065727, 'timing/data': 0.00021830000332556665, 'timing/model': 0.5929478129837662, 'learning_rate/qwen_vl_interface': 7.14031714889121e-07, 'learning_rate/action_model': 7.14031714889121e-06, 'epoch': 0.33}) +[2026-10-03 19:51:03,994][starVLA.training.train_starvla][INFO] - Step 4531, Loss: {'train/loss': 0.008488702587783337, 'train/grad_norm_pre_clip': 13.48253059387207, 'timing/data': 0.0002211499959230423, 'timing/model': 0.39438879198860377, 'learning_rate/qwen_vl_interface': 7.131287762259907e-07, 'learning_rate/action_model': 7.131287762259907e-06, 'epoch': 0.33}) +[2026-10-03 19:51:04,698][starVLA.training.train_starvla][INFO] - Step 4532, Loss: {'train/loss': 0.022146612405776978, 'train/grad_norm_pre_clip': 53.063106536865234, 'timing/data': 0.00022990000434219837, 'timing/model': 0.39455490099499, 'learning_rate/qwen_vl_interface': 7.12227702499248e-07, 'learning_rate/action_model': 7.12227702499248e-06, 'epoch': 0.33}) +[2026-10-03 19:51:05,401][starVLA.training.train_starvla][INFO] - Step 4533, Loss: {'train/loss': 0.09385071694850922, 'train/grad_norm_pre_clip': 96.07948303222656, 'timing/data': 0.00022007999359630048, 'timing/model': 0.3930803920084145, 'learning_rate/qwen_vl_interface': 7.113284940792903e-07, 'learning_rate/action_model': 7.113284940792903e-06, 'epoch': 0.33}) +[2026-10-03 19:51:06,106][starVLA.training.train_starvla][INFO] - Step 4534, Loss: {'train/loss': 4.326013004174456e-05, 'train/grad_norm_pre_clip': 0.01653101295232773, 'timing/data': 0.00022255000658333302, 'timing/model': 0.3942148909845855, 'learning_rate/qwen_vl_interface': 7.104311513357476e-07, 'learning_rate/action_model': 7.1043115133574755e-06, 'epoch': 0.33}) +[2026-10-03 19:51:07,015][starVLA.training.train_starvla][INFO] - Step 4535, Loss: {'train/loss': 0.0661170557141304, 'train/grad_norm_pre_clip': 54.56324005126953, 'timing/data': 0.00024160000612027943, 'timing/model': 0.5979979120020289, 'learning_rate/qwen_vl_interface': 7.095356746374845e-07, 'learning_rate/action_model': 7.095356746374844e-06, 'epoch': 0.33}) +[2026-10-03 19:51:07,722][starVLA.training.train_starvla][INFO] - Step 4536, Loss: {'train/loss': 0.001372964819893241, 'train/grad_norm_pre_clip': 2.1439428329467773, 'timing/data': 0.00036204999196343124, 'timing/model': 0.3944292220112402, 'learning_rate/qwen_vl_interface': 7.086420643525973e-07, 'learning_rate/action_model': 7.086420643525972e-06, 'epoch': 0.33}) +[2026-10-03 19:51:08,428][starVLA.training.train_starvla][INFO] - Step 4537, Loss: {'train/loss': 0.06354868412017822, 'train/grad_norm_pre_clip': 56.554054260253906, 'timing/data': 0.0003774499928113073, 'timing/model': 0.3948402820096817, 'learning_rate/qwen_vl_interface': 7.077503208484151e-07, 'learning_rate/action_model': 7.0775032084841515e-06, 'epoch': 0.33}) +[2026-10-03 19:51:09,135][starVLA.training.train_starvla][INFO] - Step 4538, Loss: {'train/loss': 0.00019428580708336085, 'train/grad_norm_pre_clip': 0.1315731406211853, 'timing/data': 0.0002890799951273948, 'timing/model': 0.39484795098542236, 'learning_rate/qwen_vl_interface': 7.068604444915003e-07, 'learning_rate/action_model': 7.068604444915003e-06, 'epoch': 0.33}) +[2026-10-03 19:51:09,839][starVLA.training.train_starvla][INFO] - Step 4539, Loss: {'train/loss': 0.03833871707320213, 'train/grad_norm_pre_clip': 42.59377670288086, 'timing/data': 0.0004059000057168305, 'timing/model': 0.3932955519994721, 'learning_rate/qwen_vl_interface': 7.059724356476475e-07, 'learning_rate/action_model': 7.059724356476474e-06, 'epoch': 0.33}) +[2026-10-03 19:51:10,542][starVLA.training.train_starvla][INFO] - Step 4540, Loss: {'train/loss': 8.440859528491274e-05, 'train/grad_norm_pre_clip': 0.02429286204278469, 'timing/data': 0.00033070999779738486, 'timing/model': 0.3930934419913683, 'learning_rate/qwen_vl_interface': 7.050862946818836e-07, 'learning_rate/action_model': 7.050862946818836e-06, 'epoch': 0.33}) +[2026-10-03 19:51:11,445][starVLA.training.train_starvla][INFO] - Step 4541, Loss: {'train/loss': 0.0002744471130426973, 'train/grad_norm_pre_clip': 0.6575590968132019, 'timing/data': 0.0003205299726687372, 'timing/model': 0.3946734210185241, 'learning_rate/qwen_vl_interface': 7.042020219584673e-07, 'learning_rate/action_model': 7.042020219584673e-06, 'epoch': 0.33}) +[2026-10-03 19:51:12,149][starVLA.training.train_starvla][INFO] - Step 4542, Loss: {'train/loss': 0.0007129217847250402, 'train/grad_norm_pre_clip': 0.7179794311523438, 'timing/data': 0.00033373001497238874, 'timing/model': 0.3924871220078785, 'learning_rate/qwen_vl_interface': 7.0331961784089e-07, 'learning_rate/action_model': 7.033196178408899e-06, 'epoch': 0.33}) +[2026-10-03 19:51:12,856][starVLA.training.train_starvla][INFO] - Step 4543, Loss: {'train/loss': 6.380486593116075e-05, 'train/grad_norm_pre_clip': 0.023067805916070938, 'timing/data': 0.000341629987815395, 'timing/model': 0.3949226610129699, 'learning_rate/qwen_vl_interface': 7.024390826918748e-07, 'learning_rate/action_model': 7.024390826918747e-06, 'epoch': 0.33}) +[2026-10-03 19:51:13,561][starVLA.training.train_starvla][INFO] - Step 4544, Loss: {'train/loss': 0.21903720498085022, 'train/grad_norm_pre_clip': 99.21558380126953, 'timing/data': 0.00036358999204821885, 'timing/model': 0.395284931990318, 'learning_rate/qwen_vl_interface': 7.01560416873376e-07, 'learning_rate/action_model': 7.015604168733759e-06, 'epoch': 0.33}) +[2026-10-03 19:51:14,267][starVLA.training.train_starvla][INFO] - Step 4545, Loss: {'train/loss': 0.6352496147155762, 'train/grad_norm_pre_clip': 100.88369750976562, 'timing/data': 0.00025093997828662395, 'timing/model': 0.39566111101885326, 'learning_rate/qwen_vl_interface': 7.006836207465803e-07, 'learning_rate/action_model': 7.006836207465803e-06, 'epoch': 0.33}) +[2026-10-03 19:51:15,170][starVLA.training.train_starvla][INFO] - Step 4546, Loss: {'train/loss': 5.8392997743794695e-05, 'train/grad_norm_pre_clip': 0.019416116178035736, 'timing/data': 0.0004979500081390142, 'timing/model': 0.3950112509774044, 'learning_rate/qwen_vl_interface': 6.998086946719058e-07, 'learning_rate/action_model': 6.998086946719057e-06, 'epoch': 0.33}) +[2026-10-03 19:51:15,876][starVLA.training.train_starvla][INFO] - Step 4547, Loss: {'train/loss': 0.0005427830619737506, 'train/grad_norm_pre_clip': 0.37155473232269287, 'timing/data': 0.0003566000086721033, 'timing/model': 0.39334711199626327, 'learning_rate/qwen_vl_interface': 6.989356390090008e-07, 'learning_rate/action_model': 6.989356390090007e-06, 'epoch': 0.33}) +[2026-10-03 19:51:16,580][starVLA.training.train_starvla][INFO] - Step 4548, Loss: {'train/loss': 9.958860755432397e-05, 'train/grad_norm_pre_clip': 0.025039682164788246, 'timing/data': 0.00031756001408211887, 'timing/model': 0.3931221519887913, 'learning_rate/qwen_vl_interface': 6.980644541167459e-07, 'learning_rate/action_model': 6.980644541167458e-06, 'epoch': 0.33}) +[2026-10-03 19:51:17,287][starVLA.training.train_starvla][INFO] - Step 4549, Loss: {'train/loss': 0.0076391133479774, 'train/grad_norm_pre_clip': 11.886037826538086, 'timing/data': 0.00023494998458772898, 'timing/model': 0.3957920209795702, 'learning_rate/qwen_vl_interface': 6.971951403532524e-07, 'learning_rate/action_model': 6.971951403532523e-06, 'epoch': 0.33}) +[2026-10-03 19:51:17,993][starVLA.training.train_starvla][INFO] - Step 4550, Loss: {'train/loss': 0.00024211722484324127, 'train/grad_norm_pre_clip': 0.2709214687347412, 'timing/data': 0.0003430499928072095, 'timing/model': 0.3948047620069701, 'learning_rate/qwen_vl_interface': 6.963276980758622e-07, 'learning_rate/action_model': 6.963276980758622e-06, 'epoch': 0.33}) +[2026-10-03 19:51:18,897][starVLA.training.train_starvla][INFO] - Step 4551, Loss: {'train/loss': 0.0009869533823803067, 'train/grad_norm_pre_clip': 1.94789719581604, 'timing/data': 0.00034488001256249845, 'timing/model': 0.5914212029892951, 'learning_rate/qwen_vl_interface': 6.954621276411485e-07, 'learning_rate/action_model': 6.954621276411484e-06, 'epoch': 0.33}) +[2026-10-03 19:51:19,604][starVLA.training.train_starvla][INFO] - Step 4552, Loss: {'train/loss': 4.134786649956368e-05, 'train/grad_norm_pre_clip': 0.010079698637127876, 'timing/data': 0.00022515002638101578, 'timing/model': 0.395270712004276, 'learning_rate/qwen_vl_interface': 6.945984294049136e-07, 'learning_rate/action_model': 6.9459842940491364e-06, 'epoch': 0.33}) +[2026-10-03 19:51:20,309][starVLA.training.train_starvla][INFO] - Step 4553, Loss: {'train/loss': 0.29067301750183105, 'train/grad_norm_pre_clip': 117.1363754272461, 'timing/data': 0.00034604998654685915, 'timing/model': 0.3938103520195, 'learning_rate/qwen_vl_interface': 6.937366037221919e-07, 'learning_rate/action_model': 6.937366037221918e-06, 'epoch': 0.33}) +[2026-10-03 19:51:21,015][starVLA.training.train_starvla][INFO] - Step 4554, Loss: {'train/loss': 0.5124385356903076, 'train/grad_norm_pre_clip': 44.487342834472656, 'timing/data': 0.00034965999657288194, 'timing/model': 0.39422891099820845, 'learning_rate/qwen_vl_interface': 6.928766509472471e-07, 'learning_rate/action_model': 6.92876650947247e-06, 'epoch': 0.33}) +[2026-10-03 19:51:21,720][starVLA.training.train_starvla][INFO] - Step 4555, Loss: {'train/loss': 2.9048853321000934e-05, 'train/grad_norm_pre_clip': 0.005487182177603245, 'timing/data': 0.00033802000689320266, 'timing/model': 0.39384074200643227, 'learning_rate/qwen_vl_interface': 6.920185714335737e-07, 'learning_rate/action_model': 6.9201857143357375e-06, 'epoch': 0.33}) +[2026-10-03 19:51:22,624][starVLA.training.train_starvla][INFO] - Step 4556, Loss: {'train/loss': 0.0003217137127649039, 'train/grad_norm_pre_clip': 0.3137674331665039, 'timing/data': 0.0003382900031283498, 'timing/model': 0.5918239330057986, 'learning_rate/qwen_vl_interface': 6.911623655338959e-07, 'learning_rate/action_model': 6.911623655338959e-06, 'epoch': 0.33}) +[2026-10-03 19:51:23,331][starVLA.training.train_starvla][INFO] - Step 4557, Loss: {'train/loss': 0.0045978301204741, 'train/grad_norm_pre_clip': 13.929947853088379, 'timing/data': 0.0003925299970433116, 'timing/model': 0.39594976100488566, 'learning_rate/qwen_vl_interface': 6.903080336001666e-07, 'learning_rate/action_model': 6.903080336001666e-06, 'epoch': 0.33}) +[2026-10-03 19:51:24,041][starVLA.training.train_starvla][INFO] - Step 4558, Loss: {'train/loss': 0.00036599827581085265, 'train/grad_norm_pre_clip': 0.13298942148685455, 'timing/data': 0.0003299000090919435, 'timing/model': 0.39935194098507054, 'learning_rate/qwen_vl_interface': 6.894555759835707e-07, 'learning_rate/action_model': 6.894555759835707e-06, 'epoch': 0.33}) +[2026-10-03 19:51:24,750][starVLA.training.train_starvla][INFO] - Step 4559, Loss: {'train/loss': 0.0008345753303728998, 'train/grad_norm_pre_clip': 2.3952016830444336, 'timing/data': 0.00032676002592779696, 'timing/model': 0.3966116110095754, 'learning_rate/qwen_vl_interface': 6.886049930345204e-07, 'learning_rate/action_model': 6.8860499303452035e-06, 'epoch': 0.33}) +[2026-10-03 19:51:25,460][starVLA.training.train_starvla][INFO] - Step 4560, Loss: {'train/loss': 0.00020757937454618514, 'train/grad_norm_pre_clip': 0.5110226273536682, 'timing/data': 0.00037791000795550644, 'timing/model': 0.39644542199675925, 'learning_rate/qwen_vl_interface': 6.877562851026584e-07, 'learning_rate/action_model': 6.877562851026584e-06, 'epoch': 0.33}) +[2026-10-03 19:51:26,379][starVLA.training.train_starvla][INFO] - Step 4561, Loss: {'train/loss': 0.0005509139155037701, 'train/grad_norm_pre_clip': 0.2909369170665741, 'timing/data': 0.0003871200024150312, 'timing/model': 0.6083018200006336, 'learning_rate/qwen_vl_interface': 6.869094525368572e-07, 'learning_rate/action_model': 6.8690945253685715e-06, 'epoch': 0.33}) +[2026-10-03 19:51:27,087][starVLA.training.train_starvla][INFO] - Step 4562, Loss: {'train/loss': 0.6568776369094849, 'train/grad_norm_pre_clip': 168.0667266845703, 'timing/data': 0.0003369599871803075, 'timing/model': 0.3954369409766514, 'learning_rate/qwen_vl_interface': 6.860644956852168e-07, 'learning_rate/action_model': 6.860644956852168e-06, 'epoch': 0.33}) +[2026-10-03 19:51:27,796][starVLA.training.train_starvla][INFO] - Step 4563, Loss: {'train/loss': 0.0002295648882864043, 'train/grad_norm_pre_clip': 0.10360059887170792, 'timing/data': 0.00035707998904399574, 'timing/model': 0.3954463610134553, 'learning_rate/qwen_vl_interface': 6.852214148950667e-07, 'learning_rate/action_model': 6.852214148950667e-06, 'epoch': 0.33}) +[2026-10-03 19:51:28,503][starVLA.training.train_starvla][INFO] - Step 4564, Loss: {'train/loss': 0.10210776329040527, 'train/grad_norm_pre_clip': 97.60160827636719, 'timing/data': 0.00025052999262697995, 'timing/model': 0.3936297210166231, 'learning_rate/qwen_vl_interface': 6.843802105129667e-07, 'learning_rate/action_model': 6.843802105129666e-06, 'epoch': 0.33}) +[2026-10-03 19:51:29,211][starVLA.training.train_starvla][INFO] - Step 4565, Loss: {'train/loss': 0.3325103521347046, 'train/grad_norm_pre_clip': 60.08552551269531, 'timing/data': 0.00023301999317482114, 'timing/model': 0.39497885201126337, 'learning_rate/qwen_vl_interface': 6.835408828847035e-07, 'learning_rate/action_model': 6.8354088288470356e-06, 'epoch': 0.33}) +[2026-10-03 19:51:29,920][starVLA.training.train_starvla][INFO] - Step 4566, Loss: {'train/loss': 0.015325324609875679, 'train/grad_norm_pre_clip': 57.24391174316406, 'timing/data': 0.00036130999797023833, 'timing/model': 0.39616671099793166, 'learning_rate/qwen_vl_interface': 6.827034323552929e-07, 'learning_rate/action_model': 6.827034323552929e-06, 'epoch': 0.33}) +[2026-10-03 19:51:30,833][starVLA.training.train_starvla][INFO] - Step 4567, Loss: {'train/loss': 0.0005365047254599631, 'train/grad_norm_pre_clip': 1.0429747104644775, 'timing/data': 0.0003784900181926787, 'timing/model': 0.3948319919873029, 'learning_rate/qwen_vl_interface': 6.818678592689794e-07, 'learning_rate/action_model': 6.818678592689794e-06, 'epoch': 0.33}) +[2026-10-03 19:51:31,539][starVLA.training.train_starvla][INFO] - Step 4568, Loss: {'train/loss': 0.30938220024108887, 'train/grad_norm_pre_clip': 75.63152313232422, 'timing/data': 0.00023614999372512102, 'timing/model': 0.39498837100109085, 'learning_rate/qwen_vl_interface': 6.810341639692357e-07, 'learning_rate/action_model': 6.810341639692357e-06, 'epoch': 0.33}) +[2026-10-03 19:51:32,248][starVLA.training.train_starvla][INFO] - Step 4569, Loss: {'train/loss': 8.903731941245496e-05, 'train/grad_norm_pre_clip': 0.02071554586291313, 'timing/data': 0.0002335699973627925, 'timing/model': 0.3959439320024103, 'learning_rate/qwen_vl_interface': 6.802023467987631e-07, 'learning_rate/action_model': 6.802023467987631e-06, 'epoch': 0.33}) +[2026-10-03 19:51:32,958][starVLA.training.train_starvla][INFO] - Step 4570, Loss: {'train/loss': 0.00013662621495313942, 'train/grad_norm_pre_clip': 0.10465356707572937, 'timing/data': 0.0003171600110363215, 'timing/model': 0.39621599100064486, 'learning_rate/qwen_vl_interface': 6.793724080994884e-07, 'learning_rate/action_model': 6.793724080994884e-06, 'epoch': 0.33}) +[2026-10-03 19:51:33,665][starVLA.training.train_starvla][INFO] - Step 4571, Loss: {'train/loss': 0.13345201313495636, 'train/grad_norm_pre_clip': 85.4679183959961, 'timing/data': 0.00031725000008009374, 'timing/model': 0.3944668619951699, 'learning_rate/qwen_vl_interface': 6.785443482125693e-07, 'learning_rate/action_model': 6.7854434821256915e-06, 'epoch': 0.33}) +[2026-10-03 19:51:34,577][starVLA.training.train_starvla][INFO] - Step 4572, Loss: {'train/loss': 0.00012121653708163649, 'train/grad_norm_pre_clip': 0.04658389091491699, 'timing/data': 0.00037218001671135426, 'timing/model': 0.39471879100892693, 'learning_rate/qwen_vl_interface': 6.777181674783895e-07, 'learning_rate/action_model': 6.777181674783895e-06, 'epoch': 0.33}) +[2026-10-03 19:51:35,287][starVLA.training.train_starvla][INFO] - Step 4573, Loss: {'train/loss': 0.11533671617507935, 'train/grad_norm_pre_clip': 67.40093994140625, 'timing/data': 0.0003588699910324067, 'timing/model': 0.39614853099919856, 'learning_rate/qwen_vl_interface': 6.76893866236561e-07, 'learning_rate/action_model': 6.7689386623656094e-06, 'epoch': 0.33}) +[2026-10-03 19:51:35,994][starVLA.training.train_starvla][INFO] - Step 4574, Loss: {'train/loss': 0.12067852169275284, 'train/grad_norm_pre_clip': 147.48223876953125, 'timing/data': 0.00036775998887605965, 'timing/model': 0.3937747309973929, 'learning_rate/qwen_vl_interface': 6.76071444825923e-07, 'learning_rate/action_model': 6.76071444825923e-06, 'epoch': 0.33}) +[2026-10-03 19:51:36,703][starVLA.training.train_starvla][INFO] - Step 4575, Loss: {'train/loss': 3.480729355942458e-05, 'train/grad_norm_pre_clip': 0.008458913303911686, 'timing/data': 0.00037598999915644526, 'timing/model': 0.39615026098908857, 'learning_rate/qwen_vl_interface': 6.752509035845413e-07, 'learning_rate/action_model': 6.7525090358454126e-06, 'epoch': 0.33}) +[2026-10-03 19:51:37,411][starVLA.training.train_starvla][INFO] - Step 4576, Loss: {'train/loss': 0.00045891033369116485, 'train/grad_norm_pre_clip': 0.6668076515197754, 'timing/data': 0.0003790800110436976, 'timing/model': 0.39494159098831005, 'learning_rate/qwen_vl_interface': 6.744322428497095e-07, 'learning_rate/action_model': 6.744322428497095e-06, 'epoch': 0.33}) +[2026-10-03 19:51:38,323][starVLA.training.train_starvla][INFO] - Step 4577, Loss: {'train/loss': 0.00965591799467802, 'train/grad_norm_pre_clip': 21.297103881835938, 'timing/data': 0.0003520500031299889, 'timing/model': 0.5987179719959386, 'learning_rate/qwen_vl_interface': 6.736154629579487e-07, 'learning_rate/action_model': 6.736154629579487e-06, 'epoch': 0.34}) +[2026-10-03 19:51:39,032][starVLA.training.train_starvla][INFO] - Step 4578, Loss: {'train/loss': 5.635691195493564e-05, 'train/grad_norm_pre_clip': 0.016699504107236862, 'timing/data': 0.0003413500089664012, 'timing/model': 0.39600066098500974, 'learning_rate/qwen_vl_interface': 6.728005642450055e-07, 'learning_rate/action_model': 6.7280056424500554e-06, 'epoch': 0.34}) +[2026-10-03 19:51:39,741][starVLA.training.train_starvla][INFO] - Step 4579, Loss: {'train/loss': 0.00018956864369101822, 'train/grad_norm_pre_clip': 0.07175382226705551, 'timing/data': 0.00034410000080242753, 'timing/model': 0.3959144520049449, 'learning_rate/qwen_vl_interface': 6.719875470458553e-07, 'learning_rate/action_model': 6.719875470458553e-06, 'epoch': 0.34}) +[2026-10-03 19:51:40,449][starVLA.training.train_starvla][INFO] - Step 4580, Loss: {'train/loss': 8.287948730867356e-05, 'train/grad_norm_pre_clip': 0.020558608695864677, 'timing/data': 0.00025949999690055847, 'timing/model': 0.3964850010233931, 'learning_rate/qwen_vl_interface': 6.711764116946972e-07, 'learning_rate/action_model': 6.7117641169469715e-06, 'epoch': 0.34}) +[2026-10-03 19:51:41,155][starVLA.training.train_starvla][INFO] - Step 4581, Loss: {'train/loss': 0.0002047376474365592, 'train/grad_norm_pre_clip': 0.08197175711393356, 'timing/data': 0.000532530015334487, 'timing/model': 0.3952143719943706, 'learning_rate/qwen_vl_interface': 6.703671585249587e-07, 'learning_rate/action_model': 6.703671585249586e-06, 'epoch': 0.34}) +[2026-10-03 19:51:42,067][starVLA.training.train_starvla][INFO] - Step 4582, Loss: {'train/loss': 0.5575665831565857, 'train/grad_norm_pre_clip': 115.63720703125, 'timing/data': 0.00034077998134307563, 'timing/model': 0.5988126919837669, 'learning_rate/qwen_vl_interface': 6.69559787869294e-07, 'learning_rate/action_model': 6.69559787869294e-06, 'epoch': 0.34}) +[2026-10-03 19:51:42,774][starVLA.training.train_starvla][INFO] - Step 4583, Loss: {'train/loss': 0.00022563817037735134, 'train/grad_norm_pre_clip': 0.12689660489559174, 'timing/data': 0.000238840002566576, 'timing/model': 0.3963157009857241, 'learning_rate/qwen_vl_interface': 6.68754300059582e-07, 'learning_rate/action_model': 6.687543000595819e-06, 'epoch': 0.34}) +[2026-10-03 19:51:43,483][starVLA.training.train_starvla][INFO] - Step 4584, Loss: {'train/loss': 0.0004649603215511888, 'train/grad_norm_pre_clip': 0.4213297367095947, 'timing/data': 0.0003272299945820123, 'timing/model': 0.3960241420136299, 'learning_rate/qwen_vl_interface': 6.67950695426929e-07, 'learning_rate/action_model': 6.679506954269291e-06, 'epoch': 0.34}) +[2026-10-03 19:51:44,192][starVLA.training.train_starvla][INFO] - Step 4585, Loss: {'train/loss': 0.32248497009277344, 'train/grad_norm_pre_clip': 224.09146118164062, 'timing/data': 0.0002645399945322424, 'timing/model': 0.3962947110121604, 'learning_rate/qwen_vl_interface': 6.671489743016651e-07, 'learning_rate/action_model': 6.671489743016651e-06, 'epoch': 0.34}) +[2026-10-03 19:51:44,901][starVLA.training.train_starvla][INFO] - Step 4586, Loss: {'train/loss': 0.0004633234057109803, 'train/grad_norm_pre_clip': 0.5102117657661438, 'timing/data': 0.00034281000262126327, 'timing/model': 0.39599846201599576, 'learning_rate/qwen_vl_interface': 6.663491370133487e-07, 'learning_rate/action_model': 6.663491370133487e-06, 'epoch': 0.34}) +[2026-10-03 19:51:45,816][starVLA.training.train_starvla][INFO] - Step 4587, Loss: {'train/loss': 0.012316658161580563, 'train/grad_norm_pre_clip': 28.016542434692383, 'timing/data': 0.00032918000943027437, 'timing/model': 0.6025554410007317, 'learning_rate/qwen_vl_interface': 6.655511838907629e-07, 'learning_rate/action_model': 6.655511838907629e-06, 'epoch': 0.34}) +[2026-10-03 19:51:46,523][starVLA.training.train_starvla][INFO] - Step 4588, Loss: {'train/loss': 0.00016384158516302705, 'train/grad_norm_pre_clip': 0.07836513966321945, 'timing/data': 0.0002362100058235228, 'timing/model': 0.394320052000694, 'learning_rate/qwen_vl_interface': 6.647551152619149e-07, 'learning_rate/action_model': 6.647551152619148e-06, 'epoch': 0.34}) +[2026-10-03 19:51:47,233][starVLA.training.train_starvla][INFO] - Step 4589, Loss: {'train/loss': 0.5110213756561279, 'train/grad_norm_pre_clip': 174.14903259277344, 'timing/data': 0.00036361999809741974, 'timing/model': 0.39637029101140797, 'learning_rate/qwen_vl_interface': 6.639609314540396e-07, 'learning_rate/action_model': 6.639609314540396e-06, 'epoch': 0.34}) +[2026-10-03 19:51:47,942][starVLA.training.train_starvla][INFO] - Step 4590, Loss: {'train/loss': 9.572721319273114e-05, 'train/grad_norm_pre_clip': 0.043261829763650894, 'timing/data': 0.00023281999165192246, 'timing/model': 0.3965104419912677, 'learning_rate/qwen_vl_interface': 6.631686327935954e-07, 'learning_rate/action_model': 6.631686327935954e-06, 'epoch': 0.34}) +[2026-10-03 19:51:48,652][starVLA.training.train_starvla][INFO] - Step 4591, Loss: {'train/loss': 0.00014395045582205057, 'train/grad_norm_pre_clip': 0.09763013571500778, 'timing/data': 0.00023499899543821812, 'timing/model': 0.3963093020138331, 'learning_rate/qwen_vl_interface': 6.623782196062665e-07, 'learning_rate/action_model': 6.623782196062664e-06, 'epoch': 0.34}) +[2026-10-03 19:51:49,361][starVLA.training.train_starvla][INFO] - Step 4592, Loss: {'train/loss': 6.741909601259977e-05, 'train/grad_norm_pre_clip': 0.021797088906168938, 'timing/data': 0.0003247900167480111, 'timing/model': 0.3958467619959265, 'learning_rate/qwen_vl_interface': 6.615896922169622e-07, 'learning_rate/action_model': 6.615896922169622e-06, 'epoch': 0.34}) +[2026-10-03 19:51:50,268][starVLA.training.train_starvla][INFO] - Step 4593, Loss: {'train/loss': 0.0001660672714933753, 'train/grad_norm_pre_clip': 0.050109587609767914, 'timing/data': 0.0003232999879401177, 'timing/model': 0.39421676099300385, 'learning_rate/qwen_vl_interface': 6.608030509498156e-07, 'learning_rate/action_model': 6.608030509498156e-06, 'epoch': 0.34}) +[2026-10-03 19:51:50,975][starVLA.training.train_starvla][INFO] - Step 4594, Loss: {'train/loss': 0.3452536165714264, 'train/grad_norm_pre_clip': 83.15233612060547, 'timing/data': 0.0003947600198443979, 'timing/model': 0.39600306199281476, 'learning_rate/qwen_vl_interface': 6.600182961281865e-07, 'learning_rate/action_model': 6.600182961281864e-06, 'epoch': 0.34}) +[2026-10-03 19:51:51,682][starVLA.training.train_starvla][INFO] - Step 4595, Loss: {'train/loss': 6.178973853820935e-05, 'train/grad_norm_pre_clip': 0.020483599975705147, 'timing/data': 0.0002968700136989355, 'timing/model': 0.3941634419898037, 'learning_rate/qwen_vl_interface': 6.592354280746561e-07, 'learning_rate/action_model': 6.5923542807465595e-06, 'epoch': 0.34}) +[2026-10-03 19:51:52,391][starVLA.training.train_starvla][INFO] - Step 4596, Loss: {'train/loss': 4.496138353715651e-05, 'train/grad_norm_pre_clip': 0.010228173807263374, 'timing/data': 0.00030591999529860914, 'timing/model': 0.3959315220126882, 'learning_rate/qwen_vl_interface': 6.584544471110335e-07, 'learning_rate/action_model': 6.584544471110335e-06, 'epoch': 0.34}) +[2026-10-03 19:51:53,098][starVLA.training.train_starvla][INFO] - Step 4597, Loss: {'train/loss': 4.937653648084961e-05, 'train/grad_norm_pre_clip': 0.014657808467745781, 'timing/data': 0.00023260997841134667, 'timing/model': 0.39410364200011827, 'learning_rate/qwen_vl_interface': 6.576753535583495e-07, 'learning_rate/action_model': 6.576753535583495e-06, 'epoch': 0.34}) +[2026-10-03 19:51:54,005][starVLA.training.train_starvla][INFO] - Step 4598, Loss: {'train/loss': 0.38824662566185, 'train/grad_norm_pre_clip': 76.4871597290039, 'timing/data': 0.00036399002419784665, 'timing/model': 0.39595817201188765, 'learning_rate/qwen_vl_interface': 6.5689814773686e-07, 'learning_rate/action_model': 6.568981477368599e-06, 'epoch': 0.34}) +[2026-10-03 19:51:54,714][starVLA.training.train_starvla][INFO] - Step 4599, Loss: {'train/loss': 0.4961976706981659, 'train/grad_norm_pre_clip': 96.4256820678711, 'timing/data': 0.000348019995726645, 'timing/model': 0.3956480110064149, 'learning_rate/qwen_vl_interface': 6.56122829966046e-07, 'learning_rate/action_model': 6.56122829966046e-06, 'epoch': 0.34}) +[2026-10-03 19:51:55,424][starVLA.training.train_starvla][INFO] - Step 4600, Loss: {'train/loss': 0.008373457938432693, 'train/grad_norm_pre_clip': 21.0660457611084, 'timing/data': 0.0002666999935172498, 'timing/model': 0.3959750510111917, 'learning_rate/qwen_vl_interface': 6.553494005646101e-07, 'learning_rate/action_model': 6.553494005646101e-06, 'epoch': 0.34}) +[2026-10-03 19:51:56,133][starVLA.training.train_starvla][INFO] - Step 4601, Loss: {'train/loss': 7.926065882202238e-05, 'train/grad_norm_pre_clip': 0.030591728165745735, 'timing/data': 0.0003574900038074702, 'timing/model': 0.39630195099744014, 'learning_rate/qwen_vl_interface': 6.545778598504807e-07, 'learning_rate/action_model': 6.545778598504807e-06, 'epoch': 0.34}) +[2026-10-03 19:51:56,843][starVLA.training.train_starvla][INFO] - Step 4602, Loss: {'train/loss': 0.0003744241548702121, 'train/grad_norm_pre_clip': 0.18585126101970673, 'timing/data': 0.00047788998926989734, 'timing/model': 0.3965402310132049, 'learning_rate/qwen_vl_interface': 6.538082081408093e-07, 'learning_rate/action_model': 6.538082081408093e-06, 'epoch': 0.34}) +[2026-10-03 19:51:57,748][starVLA.training.train_starvla][INFO] - Step 4603, Loss: {'train/loss': 6.793565989937633e-05, 'train/grad_norm_pre_clip': 0.016881495714187622, 'timing/data': 0.00035186001332476735, 'timing/model': 0.5919610419950914, 'learning_rate/qwen_vl_interface': 6.530404457519697e-07, 'learning_rate/action_model': 6.530404457519697e-06, 'epoch': 0.34}) +[2026-10-03 19:51:58,453][starVLA.training.train_starvla][INFO] - Step 4604, Loss: {'train/loss': 9.84409634838812e-05, 'train/grad_norm_pre_clip': 0.027542719617486, 'timing/data': 0.0003422300214879215, 'timing/model': 0.39457863199640997, 'learning_rate/qwen_vl_interface': 6.522745729995614e-07, 'learning_rate/action_model': 6.522745729995613e-06, 'epoch': 0.34}) +[2026-10-03 19:51:59,164][starVLA.training.train_starvla][INFO] - Step 4605, Loss: {'train/loss': 0.27563926577568054, 'train/grad_norm_pre_clip': 173.49432373046875, 'timing/data': 0.0004033100267406553, 'timing/model': 0.3965282409917563, 'learning_rate/qwen_vl_interface': 6.515105901984037e-07, 'learning_rate/action_model': 6.515105901984037e-06, 'epoch': 0.34}) +[2026-10-03 19:51:59,871][starVLA.training.train_starvla][INFO] - Step 4606, Loss: {'train/loss': 4.551297752186656e-05, 'train/grad_norm_pre_clip': 0.011449581943452358, 'timing/data': 0.0003691899764817208, 'timing/model': 0.3944085919938516, 'learning_rate/qwen_vl_interface': 6.50748497662543e-07, 'learning_rate/action_model': 6.507484976625429e-06, 'epoch': 0.34}) +[2026-10-03 19:52:00,581][starVLA.training.train_starvla][INFO] - Step 4607, Loss: {'train/loss': 0.00012668431736528873, 'train/grad_norm_pre_clip': 0.08120369911193848, 'timing/data': 0.0003324599820189178, 'timing/model': 0.3963488910230808, 'learning_rate/qwen_vl_interface': 6.499882957052459e-07, 'learning_rate/action_model': 6.4998829570524586e-06, 'epoch': 0.34}) +[2026-10-03 19:52:01,487][starVLA.training.train_starvla][INFO] - Step 4608, Loss: {'train/loss': 0.45574751496315, 'train/grad_norm_pre_clip': 43.035728454589844, 'timing/data': 0.00022178998915478587, 'timing/model': 0.5938738329859916, 'learning_rate/qwen_vl_interface': 6.492299846390024e-07, 'learning_rate/action_model': 6.492299846390023e-06, 'epoch': 0.34}) +[2026-10-03 19:52:02,197][starVLA.training.train_starvla][INFO] - Step 4609, Loss: {'train/loss': 0.00022881918994244188, 'train/grad_norm_pre_clip': 0.2312251329421997, 'timing/data': 0.00039324999670498073, 'timing/model': 0.39623397099785507, 'learning_rate/qwen_vl_interface': 6.484735647755259e-07, 'learning_rate/action_model': 6.48473564775526e-06, 'epoch': 0.34}) +[2026-10-03 19:52:02,905][starVLA.training.train_starvla][INFO] - Step 4610, Loss: {'train/loss': 0.0001073286184691824, 'train/grad_norm_pre_clip': 0.04616091027855873, 'timing/data': 0.00037501001497730613, 'timing/model': 0.3957796419854276, 'learning_rate/qwen_vl_interface': 6.477190364257524e-07, 'learning_rate/action_model': 6.477190364257524e-06, 'epoch': 0.34}) +[2026-10-03 19:52:03,612][starVLA.training.train_starvla][INFO] - Step 4611, Loss: {'train/loss': 0.1486412137746811, 'train/grad_norm_pre_clip': 218.29183959960938, 'timing/data': 0.0003054100088775158, 'timing/model': 0.39414211199618876, 'learning_rate/qwen_vl_interface': 6.469663998998393e-07, 'learning_rate/action_model': 6.469663998998393e-06, 'epoch': 0.34}) +[2026-10-03 19:52:04,319][starVLA.training.train_starvla][INFO] - Step 4612, Loss: {'train/loss': 0.001990174176171422, 'train/grad_norm_pre_clip': 2.157998561859131, 'timing/data': 0.0003319700190331787, 'timing/model': 0.3961625610245392, 'learning_rate/qwen_vl_interface': 6.462156555071681e-07, 'learning_rate/action_model': 6.46215655507168e-06, 'epoch': 0.34}) +[2026-10-03 19:52:05,227][starVLA.training.train_starvla][INFO] - Step 4613, Loss: {'train/loss': 5.4600492148892954e-05, 'train/grad_norm_pre_clip': 0.020281465724110603, 'timing/data': 0.0003688400029204786, 'timing/model': 0.5948875730100553, 'learning_rate/qwen_vl_interface': 6.454668035563401e-07, 'learning_rate/action_model': 6.454668035563401e-06, 'epoch': 0.34}) +[2026-10-03 19:52:05,936][starVLA.training.train_starvla][INFO] - Step 4614, Loss: {'train/loss': 8.189754589693621e-05, 'train/grad_norm_pre_clip': 0.04247252270579338, 'timing/data': 0.00031385000329464674, 'timing/model': 0.3962259509717114, 'learning_rate/qwen_vl_interface': 6.447198443551809e-07, 'learning_rate/action_model': 6.4471984435518075e-06, 'epoch': 0.34}) +[2026-10-03 19:52:06,641][starVLA.training.train_starvla][INFO] - Step 4615, Loss: {'train/loss': 0.0007417135639116168, 'train/grad_norm_pre_clip': 0.32118839025497437, 'timing/data': 0.00030617997981607914, 'timing/model': 0.3947025620145723, 'learning_rate/qwen_vl_interface': 6.439747782107373e-07, 'learning_rate/action_model': 6.4397477821073735e-06, 'epoch': 0.34}) +[2026-10-03 19:52:07,352][starVLA.training.train_starvla][INFO] - Step 4616, Loss: {'train/loss': 5.990170029690489e-05, 'train/grad_norm_pre_clip': 0.01845606602728367, 'timing/data': 0.00023428999702446163, 'timing/model': 0.39668310098932125, 'learning_rate/qwen_vl_interface': 6.432316054292779e-07, 'learning_rate/action_model': 6.43231605429278e-06, 'epoch': 0.34}) +[2026-10-03 19:52:08,061][starVLA.training.train_starvla][INFO] - Step 4617, Loss: {'train/loss': 0.1969103366136551, 'train/grad_norm_pre_clip': 122.85511016845703, 'timing/data': 0.0005296599993016571, 'timing/model': 0.39592835199437104, 'learning_rate/qwen_vl_interface': 6.424903263162924e-07, 'learning_rate/action_model': 6.424903263162923e-06, 'epoch': 0.34}) +[2026-10-03 19:52:08,768][starVLA.training.train_starvla][INFO] - Step 4618, Loss: {'train/loss': 0.5467762351036072, 'train/grad_norm_pre_clip': 71.94774627685547, 'timing/data': 0.00035187002504244447, 'timing/model': 0.3939824419794604, 'learning_rate/qwen_vl_interface': 6.417509411764924e-07, 'learning_rate/action_model': 6.4175094117649244e-06, 'epoch': 0.34}) +[2026-10-03 19:52:09,670][starVLA.training.train_starvla][INFO] - Step 4619, Loss: {'train/loss': 0.13112251460552216, 'train/grad_norm_pre_clip': 79.45683288574219, 'timing/data': 0.0003084299969486892, 'timing/model': 0.3955433120136149, 'learning_rate/qwen_vl_interface': 6.410134503138122e-07, 'learning_rate/action_model': 6.4101345031381224e-06, 'epoch': 0.34}) +[2026-10-03 19:52:10,376][starVLA.training.train_starvla][INFO] - Step 4620, Loss: {'train/loss': 0.0013543280074372888, 'train/grad_norm_pre_clip': 2.6344411373138428, 'timing/data': 0.00030278999474830925, 'timing/model': 0.3953106410044711, 'learning_rate/qwen_vl_interface': 6.402778540314054e-07, 'learning_rate/action_model': 6.402778540314054e-06, 'epoch': 0.34}) +[2026-10-03 19:52:11,082][starVLA.training.train_starvla][INFO] - Step 4621, Loss: {'train/loss': 0.0070945387706160545, 'train/grad_norm_pre_clip': 6.52711296081543, 'timing/data': 0.00020761002087965608, 'timing/model': 0.3942575120017864, 'learning_rate/qwen_vl_interface': 6.39544152631649e-07, 'learning_rate/action_model': 6.3954415263164895e-06, 'epoch': 0.34}) +[2026-10-03 19:52:11,790][starVLA.training.train_starvla][INFO] - Step 4622, Loss: {'train/loss': 8.86893249116838e-05, 'train/grad_norm_pre_clip': 0.021075552329421043, 'timing/data': 0.00026247999630868435, 'timing/model': 0.3953884110087529, 'learning_rate/qwen_vl_interface': 6.388123464161388e-07, 'learning_rate/action_model': 6.388123464161389e-06, 'epoch': 0.34}) +[2026-10-03 19:52:12,497][starVLA.training.train_starvla][INFO] - Step 4623, Loss: {'train/loss': 0.00040675082709640265, 'train/grad_norm_pre_clip': 0.31803232431411743, 'timing/data': 0.00029445000109262764, 'timing/model': 0.39606958100921474, 'learning_rate/qwen_vl_interface': 6.380824356856936e-07, 'learning_rate/action_model': 6.380824356856936e-06, 'epoch': 0.34}) +[2026-10-03 19:52:13,408][starVLA.training.train_starvla][INFO] - Step 4624, Loss: {'train/loss': 0.007644995581358671, 'train/grad_norm_pre_clip': 18.33833885192871, 'timing/data': 0.00033462001010775566, 'timing/model': 0.39586682099616155, 'learning_rate/qwen_vl_interface': 6.373544207403513e-07, 'learning_rate/action_model': 6.373544207403513e-06, 'epoch': 0.34}) +[2026-10-03 19:52:14,116][starVLA.training.train_starvla][INFO] - Step 4625, Loss: {'train/loss': 0.00010477144678588957, 'train/grad_norm_pre_clip': 0.04564850777387619, 'timing/data': 0.00022700999397784472, 'timing/model': 0.3964427319879178, 'learning_rate/qwen_vl_interface': 6.366283018793723e-07, 'learning_rate/action_model': 6.366283018793723e-06, 'epoch': 0.34}) +[2026-10-03 19:52:14,824][starVLA.training.train_starvla][INFO] - Step 4626, Loss: {'train/loss': 0.0003053608816117048, 'train/grad_norm_pre_clip': 0.09710216522216797, 'timing/data': 0.0002344899985473603, 'timing/model': 0.3962556019832846, 'learning_rate/qwen_vl_interface': 6.359040794012358e-07, 'learning_rate/action_model': 6.359040794012358e-06, 'epoch': 0.34}) +[2026-10-03 19:52:15,532][starVLA.training.train_starvla][INFO] - Step 4627, Loss: {'train/loss': 0.03808770328760147, 'train/grad_norm_pre_clip': 38.802242279052734, 'timing/data': 0.0003399800043553114, 'timing/model': 0.39439516101265326, 'learning_rate/qwen_vl_interface': 6.351817536036431e-07, 'learning_rate/action_model': 6.351817536036431e-06, 'epoch': 0.34}) +[2026-10-03 19:52:16,241][starVLA.training.train_starvla][INFO] - Step 4628, Loss: {'train/loss': 0.0005326824611984193, 'train/grad_norm_pre_clip': 0.86958247423172, 'timing/data': 0.0004623900167644024, 'timing/model': 0.39651494202553295, 'learning_rate/qwen_vl_interface': 6.344613247835135e-07, 'learning_rate/action_model': 6.344613247835135e-06, 'epoch': 0.34}) +[2026-10-03 19:52:17,152][starVLA.training.train_starvla][INFO] - Step 4629, Loss: {'train/loss': 0.00011190552322659642, 'train/grad_norm_pre_clip': 0.04141164943575859, 'timing/data': 0.00034625999978743494, 'timing/model': 0.5978374219848774, 'learning_rate/qwen_vl_interface': 6.337427932369898e-07, 'learning_rate/action_model': 6.337427932369898e-06, 'epoch': 0.34}) +[2026-10-03 19:52:17,859][starVLA.training.train_starvla][INFO] - Step 4630, Loss: {'train/loss': 0.5186707377433777, 'train/grad_norm_pre_clip': 55.353694915771484, 'timing/data': 0.0003713899932336062, 'timing/model': 0.39416710200021043, 'learning_rate/qwen_vl_interface': 6.330261592594328e-07, 'learning_rate/action_model': 6.330261592594327e-06, 'epoch': 0.34}) +[2026-10-03 19:52:18,568][starVLA.training.train_starvla][INFO] - Step 4631, Loss: {'train/loss': 8.189386426238343e-05, 'train/grad_norm_pre_clip': 0.022530078887939453, 'timing/data': 0.000312190008116886, 'timing/model': 0.39597868200507946, 'learning_rate/qwen_vl_interface': 6.323114231454229e-07, 'learning_rate/action_model': 6.323114231454229e-06, 'epoch': 0.34}) +[2026-10-03 19:52:19,277][starVLA.training.train_starvla][INFO] - Step 4632, Loss: {'train/loss': 0.0003293730551376939, 'train/grad_norm_pre_clip': 0.16771958768367767, 'timing/data': 0.0003447500057518482, 'timing/model': 0.3961898809939157, 'learning_rate/qwen_vl_interface': 6.315985851887618e-07, 'learning_rate/action_model': 6.315985851887618e-06, 'epoch': 0.34}) +[2026-10-03 19:52:19,986][starVLA.training.train_starvla][INFO] - Step 4633, Loss: {'train/loss': 4.9228008720092475e-05, 'train/grad_norm_pre_clip': 0.011160585097968578, 'timing/data': 0.0002285699883941561, 'timing/model': 0.39682472200365737, 'learning_rate/qwen_vl_interface': 6.308876456824697e-07, 'learning_rate/action_model': 6.3088764568246975e-06, 'epoch': 0.34}) +[2026-10-03 19:52:20,892][starVLA.training.train_starvla][INFO] - Step 4634, Loss: {'train/loss': 0.35407158732414246, 'train/grad_norm_pre_clip': 52.58274841308594, 'timing/data': 0.0003548800013959408, 'timing/model': 0.5945835019811057, 'learning_rate/qwen_vl_interface': 6.301786049187875e-07, 'learning_rate/action_model': 6.3017860491878745e-06, 'epoch': 0.34}) +[2026-10-03 19:52:21,596][starVLA.training.train_starvla][INFO] - Step 4635, Loss: {'train/loss': 0.4594559371471405, 'train/grad_norm_pre_clip': 139.57521057128906, 'timing/data': 0.0002972799993585795, 'timing/model': 0.39405453199287876, 'learning_rate/qwen_vl_interface': 6.294714631891755e-07, 'learning_rate/action_model': 6.294714631891754e-06, 'epoch': 0.34}) +[2026-10-03 19:52:22,305][starVLA.training.train_starvla][INFO] - Step 4636, Loss: {'train/loss': 0.003061847062781453, 'train/grad_norm_pre_clip': 9.245115280151367, 'timing/data': 0.00022592002642340958, 'timing/model': 0.39576656199642457, 'learning_rate/qwen_vl_interface': 6.287662207843115e-07, 'learning_rate/action_model': 6.287662207843114e-06, 'epoch': 0.34}) +[2026-10-03 19:52:23,012][starVLA.training.train_starvla][INFO] - Step 4637, Loss: {'train/loss': 0.00015721989620942622, 'train/grad_norm_pre_clip': 0.03239516168832779, 'timing/data': 0.00032199997804127634, 'timing/model': 0.3957521620031912, 'learning_rate/qwen_vl_interface': 6.280628779940956e-07, 'learning_rate/action_model': 6.280628779940956e-06, 'epoch': 0.34}) +[2026-10-03 19:52:23,720][starVLA.training.train_starvla][INFO] - Step 4638, Loss: {'train/loss': 0.10810215771198273, 'train/grad_norm_pre_clip': 178.67015075683594, 'timing/data': 0.00034141901414841413, 'timing/model': 0.395288932020776, 'learning_rate/qwen_vl_interface': 6.273614351076445e-07, 'learning_rate/action_model': 6.273614351076445e-06, 'epoch': 0.34}) +[2026-10-03 19:52:24,629][starVLA.training.train_starvla][INFO] - Step 4639, Loss: {'train/loss': 0.002267814241349697, 'train/grad_norm_pre_clip': 3.5416760444641113, 'timing/data': 0.00021704999380744994, 'timing/model': 0.5955052419740241, 'learning_rate/qwen_vl_interface': 6.266618924132951e-07, 'learning_rate/action_model': 6.266618924132951e-06, 'epoch': 0.34}) +[2026-10-03 19:52:25,337][starVLA.training.train_starvla][INFO] - Step 4640, Loss: {'train/loss': 0.00048396523925475776, 'train/grad_norm_pre_clip': 0.47209882736206055, 'timing/data': 0.00031413999386131763, 'timing/model': 0.39476319099776447, 'learning_rate/qwen_vl_interface': 6.25964250198603e-07, 'learning_rate/action_model': 6.2596425019860294e-06, 'epoch': 0.34}) +[2026-10-03 19:52:26,044][starVLA.training.train_starvla][INFO] - Step 4641, Loss: {'train/loss': 0.0016521316720172763, 'train/grad_norm_pre_clip': 2.278181552886963, 'timing/data': 0.0002265299845021218, 'timing/model': 0.39482140200561844, 'learning_rate/qwen_vl_interface': 6.25268508750342e-07, 'learning_rate/action_model': 6.25268508750342e-06, 'epoch': 0.34}) +[2026-10-03 19:52:26,750][starVLA.training.train_starvla][INFO] - Step 4642, Loss: {'train/loss': 0.0017989602638408542, 'train/grad_norm_pre_clip': 1.7755353450775146, 'timing/data': 0.00034193001920357347, 'timing/model': 0.39471250199130736, 'learning_rate/qwen_vl_interface': 6.245746683545062e-07, 'learning_rate/action_model': 6.245746683545062e-06, 'epoch': 0.34}) +[2026-10-03 19:52:27,463][starVLA.training.train_starvla][INFO] - Step 4643, Loss: {'train/loss': 0.17507480084896088, 'train/grad_norm_pre_clip': 100.2159194946289, 'timing/data': 0.00033758001518435776, 'timing/model': 0.39814816100988537, 'learning_rate/qwen_vl_interface': 6.238827292963063e-07, 'learning_rate/action_model': 6.238827292963063e-06, 'epoch': 0.34}) +[2026-10-03 19:52:28,174][starVLA.training.train_starvla][INFO] - Step 4644, Loss: {'train/loss': 4.0206388803198934e-05, 'train/grad_norm_pre_clip': 0.010673736222088337, 'timing/data': 0.00022332000662572682, 'timing/model': 0.3972577120002825, 'learning_rate/qwen_vl_interface': 6.231926918601731e-07, 'learning_rate/action_model': 6.2319269186017305e-06, 'epoch': 0.34}) +[2026-10-03 19:52:29,090][starVLA.training.train_starvla][INFO] - Step 4645, Loss: {'train/loss': 3.934458800358698e-05, 'train/grad_norm_pre_clip': 0.010365938767790794, 'timing/data': 0.0002484199940226972, 'timing/model': 0.3945632209943142, 'learning_rate/qwen_vl_interface': 6.22504556329754e-07, 'learning_rate/action_model': 6.22504556329754e-06, 'epoch': 0.34}) +[2026-10-03 19:52:29,800][starVLA.training.train_starvla][INFO] - Step 4646, Loss: {'train/loss': 1.9140243239235133e-05, 'train/grad_norm_pre_clip': 0.0036750202998518944, 'timing/data': 0.00035215000389143825, 'timing/model': 0.39650987199274823, 'learning_rate/qwen_vl_interface': 6.218183229879164e-07, 'learning_rate/action_model': 6.218183229879164e-06, 'epoch': 0.34}) +[2026-10-03 19:52:30,511][starVLA.training.train_starvla][INFO] - Step 4647, Loss: {'train/loss': 3.3667849493213e-05, 'train/grad_norm_pre_clip': 0.005880562588572502, 'timing/data': 0.00026358000468462706, 'timing/model': 0.398233720974531, 'learning_rate/qwen_vl_interface': 6.211339921167447e-07, 'learning_rate/action_model': 6.211339921167446e-06, 'epoch': 0.34}) +[2026-10-03 19:52:31,221][starVLA.training.train_starvla][INFO] - Step 4648, Loss: {'train/loss': 0.514075756072998, 'train/grad_norm_pre_clip': 46.2225227355957, 'timing/data': 0.00027680001221597195, 'timing/model': 0.3974814410030376, 'learning_rate/qwen_vl_interface': 6.204515639975415e-07, 'learning_rate/action_model': 6.204515639975414e-06, 'epoch': 0.34}) +[2026-10-03 19:52:31,931][starVLA.training.train_starvla][INFO] - Step 4649, Loss: {'train/loss': 0.001229041488841176, 'train/grad_norm_pre_clip': 0.5504767298698425, 'timing/data': 0.00030926999170333147, 'timing/model': 0.39688447199296206, 'learning_rate/qwen_vl_interface': 6.197710389108262e-07, 'learning_rate/action_model': 6.197710389108262e-06, 'epoch': 0.34}) +[2026-10-03 19:52:32,840][starVLA.training.train_starvla][INFO] - Step 4650, Loss: {'train/loss': 5.669569509336725e-05, 'train/grad_norm_pre_clip': 0.011130346916615963, 'timing/data': 0.0003544399805832654, 'timing/model': 0.5973909110180102, 'learning_rate/qwen_vl_interface': 6.190924171363378e-07, 'learning_rate/action_model': 6.190924171363378e-06, 'epoch': 0.34}) +[2026-10-03 19:52:33,550][starVLA.training.train_starvla][INFO] - Step 4651, Loss: {'train/loss': 0.000233110215049237, 'train/grad_norm_pre_clip': 0.13127601146697998, 'timing/data': 0.0003047400095965713, 'timing/model': 0.3971877910080366, 'learning_rate/qwen_vl_interface': 6.184156989530326e-07, 'learning_rate/action_model': 6.184156989530325e-06, 'epoch': 0.34}) +[2026-10-03 19:52:34,261][starVLA.training.train_starvla][INFO] - Step 4652, Loss: {'train/loss': 0.0002444107667542994, 'train/grad_norm_pre_clip': 0.1608767807483673, 'timing/data': 0.00035205998574383557, 'timing/model': 0.39703308197204024, 'learning_rate/qwen_vl_interface': 6.17740884639083e-07, 'learning_rate/action_model': 6.177408846390829e-06, 'epoch': 0.34}) +[2026-10-03 19:52:34,972][starVLA.training.train_starvla][INFO] - Step 4653, Loss: {'train/loss': 0.23865379393100739, 'train/grad_norm_pre_clip': 158.054931640625, 'timing/data': 0.00033735999022610486, 'timing/model': 0.39551277199643664, 'learning_rate/qwen_vl_interface': 6.170679744718798e-07, 'learning_rate/action_model': 6.1706797447187975e-06, 'epoch': 0.34}) +[2026-10-03 19:52:35,679][starVLA.training.train_starvla][INFO] - Step 4654, Loss: {'train/loss': 0.6907945275306702, 'train/grad_norm_pre_clip': 119.6182632446289, 'timing/data': 0.00035568999010138214, 'timing/model': 0.39494495099643245, 'learning_rate/qwen_vl_interface': 6.163969687280311e-07, 'learning_rate/action_model': 6.163969687280311e-06, 'epoch': 0.34}) +[2026-10-03 19:52:36,392][starVLA.training.train_starvla][INFO] - Step 4655, Loss: {'train/loss': 0.5351694822311401, 'train/grad_norm_pre_clip': 70.92485809326172, 'timing/data': 0.00027435002266429365, 'timing/model': 0.39791340101510286, 'learning_rate/qwen_vl_interface': 6.157278676833627e-07, 'learning_rate/action_model': 6.157278676833627e-06, 'epoch': 0.34}) +[2026-10-03 19:52:37,306][starVLA.training.train_starvla][INFO] - Step 4656, Loss: {'train/loss': 7.201785774668679e-05, 'train/grad_norm_pre_clip': 0.0239887572824955, 'timing/data': 0.0002989000058732927, 'timing/model': 0.3964580220053904, 'learning_rate/qwen_vl_interface': 6.150606716129155e-07, 'learning_rate/action_model': 6.150606716129154e-06, 'epoch': 0.34}) +[2026-10-03 19:52:38,017][starVLA.training.train_starvla][INFO] - Step 4657, Loss: {'train/loss': 4.893147706752643e-05, 'train/grad_norm_pre_clip': 0.011223883368074894, 'timing/data': 0.000308859976939857, 'timing/model': 0.39670141000533476, 'learning_rate/qwen_vl_interface': 6.143953807909497e-07, 'learning_rate/action_model': 6.143953807909496e-06, 'epoch': 0.34}) +[2026-10-03 19:52:38,725][starVLA.training.train_starvla][INFO] - Step 4658, Loss: {'train/loss': 4.9076112190959975e-05, 'train/grad_norm_pre_clip': 0.019916092976927757, 'timing/data': 0.00022464999346993864, 'timing/model': 0.39534982200711966, 'learning_rate/qwen_vl_interface': 6.137319954909407e-07, 'learning_rate/action_model': 6.137319954909406e-06, 'epoch': 0.34}) +[2026-10-03 19:52:39,434][starVLA.training.train_starvla][INFO] - Step 4659, Loss: {'train/loss': 0.046073537319898605, 'train/grad_norm_pre_clip': 126.42494201660156, 'timing/data': 0.00025444000493735075, 'timing/model': 0.3954421009984799, 'learning_rate/qwen_vl_interface': 6.130705159855815e-07, 'learning_rate/action_model': 6.130705159855814e-06, 'epoch': 0.34}) +[2026-10-03 19:52:40,139][starVLA.training.train_starvla][INFO] - Step 4660, Loss: {'train/loss': 0.04185321927070618, 'train/grad_norm_pre_clip': 76.2336196899414, 'timing/data': 0.0003981500049121678, 'timing/model': 0.3936075620003976, 'learning_rate/qwen_vl_interface': 6.124109425467805e-07, 'learning_rate/action_model': 6.124109425467805e-06, 'epoch': 0.34}) +[2026-10-03 19:52:41,053][starVLA.training.train_starvla][INFO] - Step 4661, Loss: {'train/loss': 0.006055149715393782, 'train/grad_norm_pre_clip': 18.15518569946289, 'timing/data': 0.00026159000117331743, 'timing/model': 0.39586771198082715, 'learning_rate/qwen_vl_interface': 6.117532754456642e-07, 'learning_rate/action_model': 6.117532754456642e-06, 'epoch': 0.34}) +[2026-10-03 19:52:41,760][starVLA.training.train_starvla][INFO] - Step 4662, Loss: {'train/loss': 0.0002023210981860757, 'train/grad_norm_pre_clip': 0.11536647379398346, 'timing/data': 0.000222509988816455, 'timing/model': 0.3960784510127269, 'learning_rate/qwen_vl_interface': 6.110975149525746e-07, 'learning_rate/action_model': 6.110975149525746e-06, 'epoch': 0.34}) +[2026-10-03 19:52:42,465][starVLA.training.train_starvla][INFO] - Step 4663, Loss: {'train/loss': 0.15562638640403748, 'train/grad_norm_pre_clip': 135.50064086914062, 'timing/data': 0.00022175998310558498, 'timing/model': 0.3945349710120354, 'learning_rate/qwen_vl_interface': 6.104436613370692e-07, 'learning_rate/action_model': 6.104436613370692e-06, 'epoch': 0.34}) +[2026-10-03 19:52:43,172][starVLA.training.train_starvla][INFO] - Step 4664, Loss: {'train/loss': 0.00010809989180415869, 'train/grad_norm_pre_clip': 0.0402492955327034, 'timing/data': 0.00022745001479052007, 'timing/model': 0.39410051101003774, 'learning_rate/qwen_vl_interface': 6.09791714867924e-07, 'learning_rate/action_model': 6.0979171486792396e-06, 'epoch': 0.34}) +[2026-10-03 19:52:43,878][starVLA.training.train_starvla][INFO] - Step 4665, Loss: {'train/loss': 0.002438728231936693, 'train/grad_norm_pre_clip': 3.711796760559082, 'timing/data': 0.00042021999252028763, 'timing/model': 0.39476746099535376, 'learning_rate/qwen_vl_interface': 6.091416758131287e-07, 'learning_rate/action_model': 6.091416758131286e-06, 'epoch': 0.34}) +[2026-10-03 19:52:44,793][starVLA.training.train_starvla][INFO] - Step 4666, Loss: {'train/loss': 0.00021242591901682317, 'train/grad_norm_pre_clip': 0.36822932958602905, 'timing/data': 0.0003682300157379359, 'timing/model': 0.6017176410241518, 'learning_rate/qwen_vl_interface': 6.084935444398895e-07, 'learning_rate/action_model': 6.084935444398894e-06, 'epoch': 0.34}) +[2026-10-03 19:52:45,499][starVLA.training.train_starvla][INFO] - Step 4667, Loss: {'train/loss': 0.000766645884141326, 'train/grad_norm_pre_clip': 0.33040520548820496, 'timing/data': 0.0003392400103621185, 'timing/model': 0.3941954410111066, 'learning_rate/qwen_vl_interface': 6.078473210146294e-07, 'learning_rate/action_model': 6.078473210146294e-06, 'epoch': 0.34}) +[2026-10-03 19:52:46,206][starVLA.training.train_starvla][INFO] - Step 4668, Loss: {'train/loss': 8.13525912235491e-05, 'train/grad_norm_pre_clip': 0.05063510313630104, 'timing/data': 0.00025850001838989556, 'timing/model': 0.39589228099794127, 'learning_rate/qwen_vl_interface': 6.072030058029862e-07, 'learning_rate/action_model': 6.072030058029861e-06, 'epoch': 0.34}) +[2026-10-03 19:52:46,915][starVLA.training.train_starvla][INFO] - Step 4669, Loss: {'train/loss': 8.752835856284946e-05, 'train/grad_norm_pre_clip': 0.02353806048631668, 'timing/data': 0.00025996999465860426, 'timing/model': 0.39612251200014725, 'learning_rate/qwen_vl_interface': 6.065605990698136e-07, 'learning_rate/action_model': 6.065605990698136e-06, 'epoch': 0.34}) +[2026-10-03 19:52:47,621][starVLA.training.train_starvla][INFO] - Step 4670, Loss: {'train/loss': 0.00036126270424574614, 'train/grad_norm_pre_clip': 0.44988784193992615, 'timing/data': 0.000258979998761788, 'timing/model': 0.3945703119970858, 'learning_rate/qwen_vl_interface': 6.059201010791811e-07, 'learning_rate/action_model': 6.059201010791811e-06, 'epoch': 0.34}) +[2026-10-03 19:52:48,327][starVLA.training.train_starvla][INFO] - Step 4671, Loss: {'train/loss': 0.00010691185889299959, 'train/grad_norm_pre_clip': 0.03892253711819649, 'timing/data': 0.00025459000607952476, 'timing/model': 0.3946183520019986, 'learning_rate/qwen_vl_interface': 6.052815120943721e-07, 'learning_rate/action_model': 6.052815120943721e-06, 'epoch': 0.34}) +[2026-10-03 19:52:49,243][starVLA.training.train_starvla][INFO] - Step 4672, Loss: {'train/loss': 0.022306138649582863, 'train/grad_norm_pre_clip': 29.417936325073242, 'timing/data': 0.000262229994405061, 'timing/model': 0.39559656201163307, 'learning_rate/qwen_vl_interface': 6.046448323778874e-07, 'learning_rate/action_model': 6.046448323778874e-06, 'epoch': 0.34}) +[2026-10-03 19:52:49,950][starVLA.training.train_starvla][INFO] - Step 4673, Loss: {'train/loss': 0.5915254354476929, 'train/grad_norm_pre_clip': 82.29570770263672, 'timing/data': 0.0002246500225737691, 'timing/model': 0.394594022014644, 'learning_rate/qwen_vl_interface': 6.040100621914419e-07, 'learning_rate/action_model': 6.040100621914418e-06, 'epoch': 0.34}) +[2026-10-03 19:52:50,654][starVLA.training.train_starvla][INFO] - Step 4674, Loss: {'train/loss': 0.0010219236137345433, 'train/grad_norm_pre_clip': 2.018695831298828, 'timing/data': 0.00044106997665949166, 'timing/model': 0.3931381120055448, 'learning_rate/qwen_vl_interface': 6.033772017959652e-07, 'learning_rate/action_model': 6.033772017959652e-06, 'epoch': 0.34}) +[2026-10-03 19:52:51,362][starVLA.training.train_starvla][INFO] - Step 4675, Loss: {'train/loss': 0.5181441307067871, 'train/grad_norm_pre_clip': 95.53509521484375, 'timing/data': 0.00031087000388652086, 'timing/model': 0.39450357199530117, 'learning_rate/qwen_vl_interface': 6.027462514516022e-07, 'learning_rate/action_model': 6.0274625145160214e-06, 'epoch': 0.34}) +[2026-10-03 19:52:52,071][starVLA.training.train_starvla][INFO] - Step 4676, Loss: {'train/loss': 4.317941784393042e-05, 'train/grad_norm_pre_clip': 0.018257616087794304, 'timing/data': 0.0002914099895861, 'timing/model': 0.39627940198988654, 'learning_rate/qwen_vl_interface': 6.021172114177139e-07, 'learning_rate/action_model': 6.021172114177139e-06, 'epoch': 0.34}) +[2026-10-03 19:52:52,985][starVLA.training.train_starvla][INFO] - Step 4677, Loss: {'train/loss': 0.0004428017418831587, 'train/grad_norm_pre_clip': 0.4884450137615204, 'timing/data': 0.0003494600241538137, 'timing/model': 0.39592050001374446, 'learning_rate/qwen_vl_interface': 6.014900819528734e-07, 'learning_rate/action_model': 6.0149008195287335e-06, 'epoch': 0.34}) +[2026-10-03 19:52:53,692][starVLA.training.train_starvla][INFO] - Step 4678, Loss: {'train/loss': 0.020286155864596367, 'train/grad_norm_pre_clip': 21.764097213745117, 'timing/data': 0.00039277999894693494, 'timing/model': 0.3939105219906196, 'learning_rate/qwen_vl_interface': 6.008648633148707e-07, 'learning_rate/action_model': 6.008648633148707e-06, 'epoch': 0.34}) +[2026-10-03 19:52:54,397][starVLA.training.train_starvla][INFO] - Step 4679, Loss: {'train/loss': 0.3614277243614197, 'train/grad_norm_pre_clip': 396.4270324707031, 'timing/data': 0.00034115000744350255, 'timing/model': 0.39381536201108247, 'learning_rate/qwen_vl_interface': 6.002415557607097e-07, 'learning_rate/action_model': 6.002415557607096e-06, 'epoch': 0.34}) +[2026-10-03 19:52:55,106][starVLA.training.train_starvla][INFO] - Step 4680, Loss: {'train/loss': 0.00011011396418325603, 'train/grad_norm_pre_clip': 0.08472130447626114, 'timing/data': 0.0003535599971655756, 'timing/model': 0.3956974509928841, 'learning_rate/qwen_vl_interface': 5.996201595466088e-07, 'learning_rate/action_model': 5.996201595466088e-06, 'epoch': 0.34}) +[2026-10-03 19:52:55,815][starVLA.training.train_starvla][INFO] - Step 4681, Loss: {'train/loss': 0.6169313788414001, 'train/grad_norm_pre_clip': 141.1463165283203, 'timing/data': 0.00030449999030679464, 'timing/model': 0.3965059409965761, 'learning_rate/qwen_vl_interface': 5.990006749279998e-07, 'learning_rate/action_model': 5.990006749279998e-06, 'epoch': 0.34}) +[2026-10-03 19:52:56,723][starVLA.training.train_starvla][INFO] - Step 4682, Loss: {'train/loss': 4.416676165419631e-05, 'train/grad_norm_pre_clip': 0.01793370582163334, 'timing/data': 0.0003892699896823615, 'timing/model': 0.5951360219914932, 'learning_rate/qwen_vl_interface': 5.983831021595296e-07, 'learning_rate/action_model': 5.9838310215952956e-06, 'epoch': 0.34}) +[2026-10-03 19:52:57,431][starVLA.training.train_starvla][INFO] - Step 4683, Loss: {'train/loss': 3.4695607610046864e-05, 'train/grad_norm_pre_clip': 0.008662722073495388, 'timing/data': 0.00026953997439704835, 'timing/model': 0.395626011013519, 'learning_rate/qwen_vl_interface': 5.977674414950595e-07, 'learning_rate/action_model': 5.9776744149505955e-06, 'epoch': 0.34}) +[2026-10-03 19:52:58,138][starVLA.training.train_starvla][INFO] - Step 4684, Loss: {'train/loss': 0.2740001976490021, 'train/grad_norm_pre_clip': 142.82772827148438, 'timing/data': 0.0003029499785043299, 'timing/model': 0.3943085409991909, 'learning_rate/qwen_vl_interface': 5.971536931876635e-07, 'learning_rate/action_model': 5.971536931876634e-06, 'epoch': 0.34}) +[2026-10-03 19:52:58,845][starVLA.training.train_starvla][INFO] - Step 4685, Loss: {'train/loss': 0.00017600232968106866, 'train/grad_norm_pre_clip': 0.38580453395843506, 'timing/data': 0.00031691999174654484, 'timing/model': 0.39589455199893564, 'learning_rate/qwen_vl_interface': 5.965418574896315e-07, 'learning_rate/action_model': 5.9654185748963145e-06, 'epoch': 0.34}) +[2026-10-03 19:52:59,550][starVLA.training.train_starvla][INFO] - Step 4686, Loss: {'train/loss': 6.749994645360857e-05, 'train/grad_norm_pre_clip': 0.02036539651453495, 'timing/data': 0.00023564998991787434, 'timing/model': 0.3945030920149293, 'learning_rate/qwen_vl_interface': 5.959319346524646e-07, 'learning_rate/action_model': 5.959319346524646e-06, 'epoch': 0.34}) +[2026-10-03 19:53:00,466][starVLA.training.train_starvla][INFO] - Step 4687, Loss: {'train/loss': 0.0035090085584670305, 'train/grad_norm_pre_clip': 4.264312744140625, 'timing/data': 0.00039366999408230186, 'timing/model': 0.6014311710023321, 'learning_rate/qwen_vl_interface': 5.953239249268804e-07, 'learning_rate/action_model': 5.953239249268804e-06, 'epoch': 0.34}) +[2026-10-03 19:53:01,174][starVLA.training.train_starvla][INFO] - Step 4688, Loss: {'train/loss': 0.4307583272457123, 'train/grad_norm_pre_clip': 46.84206008911133, 'timing/data': 0.00036929998896084726, 'timing/model': 0.39525009199860506, 'learning_rate/qwen_vl_interface': 5.947178285628081e-07, 'learning_rate/action_model': 5.94717828562808e-06, 'epoch': 0.34}) +[2026-10-03 19:53:01,881][starVLA.training.train_starvla][INFO] - Step 4689, Loss: {'train/loss': 0.5746077299118042, 'train/grad_norm_pre_clip': 210.5477752685547, 'timing/data': 0.0002643899933900684, 'timing/model': 0.3936271010024939, 'learning_rate/qwen_vl_interface': 5.941136458093907e-07, 'learning_rate/action_model': 5.941136458093907e-06, 'epoch': 0.34}) +[2026-10-03 19:53:02,587][starVLA.training.train_starvla][INFO] - Step 4690, Loss: {'train/loss': 0.5498631596565247, 'train/grad_norm_pre_clip': 58.18844985961914, 'timing/data': 0.00022252998314797878, 'timing/model': 0.3949338409875054, 'learning_rate/qwen_vl_interface': 5.935113769149855e-07, 'learning_rate/action_model': 5.935113769149856e-06, 'epoch': 0.34}) +[2026-10-03 19:53:03,296][starVLA.training.train_starvla][INFO] - Step 4691, Loss: {'train/loss': 7.37286900402978e-05, 'train/grad_norm_pre_clip': 0.05948970839381218, 'timing/data': 0.00033415001234970987, 'timing/model': 0.3951627020142041, 'learning_rate/qwen_vl_interface': 5.92911022127163e-07, 'learning_rate/action_model': 5.929110221271629e-06, 'epoch': 0.34}) +[2026-10-03 19:53:04,210][starVLA.training.train_starvla][INFO] - Step 4692, Loss: {'train/loss': 0.0001262545702047646, 'train/grad_norm_pre_clip': 0.0863659456372261, 'timing/data': 0.0003518599842209369, 'timing/model': 0.6023714309849311, 'learning_rate/qwen_vl_interface': 5.92312581692705e-07, 'learning_rate/action_model': 5.923125816927049e-06, 'epoch': 0.34}) +[2026-10-03 19:53:04,919][starVLA.training.train_starvla][INFO] - Step 4693, Loss: {'train/loss': 0.12203573435544968, 'train/grad_norm_pre_clip': 126.00918579101562, 'timing/data': 0.00022434999118559062, 'timing/model': 0.39640688098734245, 'learning_rate/qwen_vl_interface': 5.917160558576088e-07, 'learning_rate/action_model': 5.917160558576088e-06, 'epoch': 0.34}) +[2026-10-03 19:53:05,627][starVLA.training.train_starvla][INFO] - Step 4694, Loss: {'train/loss': 0.0005421029054559767, 'train/grad_norm_pre_clip': 0.4408060312271118, 'timing/data': 0.00022186999558471143, 'timing/model': 0.3953356810088735, 'learning_rate/qwen_vl_interface': 5.911214448670829e-07, 'learning_rate/action_model': 5.911214448670828e-06, 'epoch': 0.34}) +[2026-10-03 19:53:06,340][starVLA.training.train_starvla][INFO] - Step 4695, Loss: {'train/loss': 0.0007464903173968196, 'train/grad_norm_pre_clip': 0.9719954133033752, 'timing/data': 0.00034565001260489225, 'timing/model': 0.39533002200187184, 'learning_rate/qwen_vl_interface': 5.905287489655499e-07, 'learning_rate/action_model': 5.905287489655499e-06, 'epoch': 0.34}) +[2026-10-03 19:53:07,050][starVLA.training.train_starvla][INFO] - Step 4696, Loss: {'train/loss': 3.141026900266297e-05, 'train/grad_norm_pre_clip': 0.01024689618498087, 'timing/data': 0.00030940998112782836, 'timing/model': 0.3961002119758632, 'learning_rate/qwen_vl_interface': 5.899379683966445e-07, 'learning_rate/action_model': 5.899379683966445e-06, 'epoch': 0.34}) +[2026-10-03 19:53:07,757][starVLA.training.train_starvla][INFO] - Step 4697, Loss: {'train/loss': 0.000316446676151827, 'train/grad_norm_pre_clip': 0.21738331019878387, 'timing/data': 0.0003609799896366894, 'timing/model': 0.39427798200631514, 'learning_rate/qwen_vl_interface': 5.893491034032142e-07, 'learning_rate/action_model': 5.893491034032142e-06, 'epoch': 0.34}) +[2026-10-03 19:53:08,670][starVLA.training.train_starvla][INFO] - Step 4698, Loss: {'train/loss': 0.3543047308921814, 'train/grad_norm_pre_clip': 54.09003829956055, 'timing/data': 0.0004045100067742169, 'timing/model': 0.3932470509898849, 'learning_rate/qwen_vl_interface': 5.887621542273196e-07, 'learning_rate/action_model': 5.887621542273196e-06, 'epoch': 0.34}) +[2026-10-03 19:53:09,378][starVLA.training.train_starvla][INFO] - Step 4699, Loss: {'train/loss': 0.00035056378692388535, 'train/grad_norm_pre_clip': 0.20273159444332123, 'timing/data': 0.0003335799847263843, 'timing/model': 0.39579126102034934, 'learning_rate/qwen_vl_interface': 5.881771211102327e-07, 'learning_rate/action_model': 5.881771211102327e-06, 'epoch': 0.34}) +[2026-10-03 19:53:10,086][starVLA.training.train_starvla][INFO] - Step 4700, Loss: {'train/loss': 0.00019575897022150457, 'train/grad_norm_pre_clip': 0.14366044104099274, 'timing/data': 0.00026172000798396766, 'timing/model': 0.3944226719904691, 'learning_rate/qwen_vl_interface': 5.875940042924391e-07, 'learning_rate/action_model': 5.875940042924391e-06, 'epoch': 0.34}) +[2026-10-03 19:53:10,794][starVLA.training.train_starvla][INFO] - Step 4701, Loss: {'train/loss': 0.00015066290507093072, 'train/grad_norm_pre_clip': 0.06323889642953873, 'timing/data': 0.00037626997800543904, 'timing/model': 0.3962119810166769, 'learning_rate/qwen_vl_interface': 5.870128040136368e-07, 'learning_rate/action_model': 5.870128040136367e-06, 'epoch': 0.34}) +[2026-10-03 19:53:11,502][starVLA.training.train_starvla][INFO] - Step 4702, Loss: {'train/loss': 0.00018086913041770458, 'train/grad_norm_pre_clip': 0.06928481161594391, 'timing/data': 0.0003376200038474053, 'timing/model': 0.39523646098678, 'learning_rate/qwen_vl_interface': 5.864335205127332e-07, 'learning_rate/action_model': 5.864335205127332e-06, 'epoch': 0.34}) +[2026-10-03 19:53:12,415][starVLA.training.train_starvla][INFO] - Step 4703, Loss: {'train/loss': 0.00011707007797667757, 'train/grad_norm_pre_clip': 0.030678531154990196, 'timing/data': 0.00034180001239292324, 'timing/model': 0.5999931510013994, 'learning_rate/qwen_vl_interface': 5.858561540278513e-07, 'learning_rate/action_model': 5.858561540278513e-06, 'epoch': 0.34}) +[2026-10-03 19:53:13,123][starVLA.training.train_starvla][INFO] - Step 4704, Loss: {'train/loss': 6.43436869722791e-05, 'train/grad_norm_pre_clip': 0.014752395451068878, 'timing/data': 0.00033899000845849514, 'timing/model': 0.39514828100800514, 'learning_rate/qwen_vl_interface': 5.852807047963244e-07, 'learning_rate/action_model': 5.852807047963243e-06, 'epoch': 0.34}) +[2026-10-03 19:53:13,830][starVLA.training.train_starvla][INFO] - Step 4705, Loss: {'train/loss': 0.0065976884216070175, 'train/grad_norm_pre_clip': 9.838277816772461, 'timing/data': 0.00035547002335079014, 'timing/model': 0.3963696010177955, 'learning_rate/qwen_vl_interface': 5.847071730546974e-07, 'learning_rate/action_model': 5.847071730546974e-06, 'epoch': 0.34}) +[2026-10-03 19:53:14,537][starVLA.training.train_starvla][INFO] - Step 4706, Loss: {'train/loss': 0.0006642374792136252, 'train/grad_norm_pre_clip': 0.7532781362533569, 'timing/data': 0.00035592998028732836, 'timing/model': 0.39392912198673, 'learning_rate/qwen_vl_interface': 5.841355590387291e-07, 'learning_rate/action_model': 5.841355590387291e-06, 'epoch': 0.34}) +[2026-10-03 19:53:15,242][starVLA.training.train_starvla][INFO] - Step 4707, Loss: {'train/loss': 3.894981637131423e-05, 'train/grad_norm_pre_clip': 0.007495657075196505, 'timing/data': 0.0002413900219835341, 'timing/model': 0.3929768119996879, 'learning_rate/qwen_vl_interface': 5.835658629833866e-07, 'learning_rate/action_model': 5.8356586298338654e-06, 'epoch': 0.34}) +[2026-10-03 19:53:15,950][starVLA.training.train_starvla][INFO] - Step 4708, Loss: {'train/loss': 0.006646136753261089, 'train/grad_norm_pre_clip': 25.158004760742188, 'timing/data': 0.0003287099825683981, 'timing/model': 0.3953609620220959, 'learning_rate/qwen_vl_interface': 5.829980851228512e-07, 'learning_rate/action_model': 5.829980851228512e-06, 'epoch': 0.34}) +[2026-10-03 19:53:16,861][starVLA.training.train_starvla][INFO] - Step 4709, Loss: {'train/loss': 0.00013601289538200945, 'train/grad_norm_pre_clip': 0.11320193111896515, 'timing/data': 0.00024128000950440764, 'timing/model': 0.3972847709956113, 'learning_rate/qwen_vl_interface': 5.824322256905155e-07, 'learning_rate/action_model': 5.824322256905154e-06, 'epoch': 0.34}) +[2026-10-03 19:53:17,568][starVLA.training.train_starvla][INFO] - Step 4710, Loss: {'train/loss': 0.00011675491987261921, 'train/grad_norm_pre_clip': 0.07856070250272751, 'timing/data': 0.000235590006923303, 'timing/model': 0.39414502200088464, 'learning_rate/qwen_vl_interface': 5.818682849189818e-07, 'learning_rate/action_model': 5.818682849189817e-06, 'epoch': 0.34}) +[2026-10-03 19:53:18,275][starVLA.training.train_starvla][INFO] - Step 4711, Loss: {'train/loss': 0.045019540935754776, 'train/grad_norm_pre_clip': 117.58350372314453, 'timing/data': 0.0002269999822601676, 'timing/model': 0.3940444219915662, 'learning_rate/qwen_vl_interface': 5.813062630400653e-07, 'learning_rate/action_model': 5.813062630400653e-06, 'epoch': 0.34}) +[2026-10-03 19:53:18,982][starVLA.training.train_starvla][INFO] - Step 4712, Loss: {'train/loss': 6.581131310667843e-05, 'train/grad_norm_pre_clip': 0.020275434479117393, 'timing/data': 0.00033577997237443924, 'timing/model': 0.39575904101366177, 'learning_rate/qwen_vl_interface': 5.807461602847922e-07, 'learning_rate/action_model': 5.807461602847922e-06, 'epoch': 0.34}) +[2026-10-03 19:53:19,690][starVLA.training.train_starvla][INFO] - Step 4713, Loss: {'train/loss': 0.4074794352054596, 'train/grad_norm_pre_clip': 87.51446533203125, 'timing/data': 0.00022026000078767538, 'timing/model': 0.39575292100198567, 'learning_rate/qwen_vl_interface': 5.801879768833998e-07, 'learning_rate/action_model': 5.801879768833998e-06, 'epoch': 0.34}) +[2026-10-03 19:53:20,605][starVLA.training.train_starvla][INFO] - Step 4714, Loss: {'train/loss': 4.303205059841275e-05, 'train/grad_norm_pre_clip': 0.011884443461894989, 'timing/data': 0.00024717001360841095, 'timing/model': 0.39527152202208526, 'learning_rate/qwen_vl_interface': 5.796317130653363e-07, 'learning_rate/action_model': 5.796317130653363e-06, 'epoch': 0.35}) +[2026-10-03 19:53:21,313][starVLA.training.train_starvla][INFO] - Step 4715, Loss: {'train/loss': 0.709119439125061, 'train/grad_norm_pre_clip': 66.1050796508789, 'timing/data': 0.0002248189994134009, 'timing/model': 0.39524639199953526, 'learning_rate/qwen_vl_interface': 5.790773690592592e-07, 'learning_rate/action_model': 5.7907736905925925e-06, 'epoch': 0.35}) +[2026-10-03 19:53:22,022][starVLA.training.train_starvla][INFO] - Step 4716, Loss: {'train/loss': 0.03971332684159279, 'train/grad_norm_pre_clip': 55.78351974487305, 'timing/data': 0.0003259100194554776, 'timing/model': 0.395861272001639, 'learning_rate/qwen_vl_interface': 5.7852494509304e-07, 'learning_rate/action_model': 5.7852494509304e-06, 'epoch': 0.35}) +[2026-10-03 19:53:22,730][starVLA.training.train_starvla][INFO] - Step 4717, Loss: {'train/loss': 0.0008083871798589826, 'train/grad_norm_pre_clip': 0.7751573324203491, 'timing/data': 0.000324180000461638, 'timing/model': 0.3945940520206932, 'learning_rate/qwen_vl_interface': 5.779744413937592e-07, 'learning_rate/action_model': 5.779744413937592e-06, 'epoch': 0.35}) +[2026-10-03 19:53:23,436][starVLA.training.train_starvla][INFO] - Step 4718, Loss: {'train/loss': 8.207947394112125e-05, 'train/grad_norm_pre_clip': 0.027399394661188126, 'timing/data': 0.0003412799851503223, 'timing/model': 0.3943551520060282, 'learning_rate/qwen_vl_interface': 5.774258581877073e-07, 'learning_rate/action_model': 5.774258581877073e-06, 'epoch': 0.35}) +[2026-10-03 19:53:24,347][starVLA.training.train_starvla][INFO] - Step 4719, Loss: {'train/loss': 0.010646930895745754, 'train/grad_norm_pre_clip': 13.73459243774414, 'timing/data': 0.0004717199772130698, 'timing/model': 0.5980925020121504, 'learning_rate/qwen_vl_interface': 5.768791957003866e-07, 'learning_rate/action_model': 5.768791957003865e-06, 'epoch': 0.35}) +[2026-10-03 19:53:25,054][starVLA.training.train_starvla][INFO] - Step 4720, Loss: {'train/loss': 5.318623152561486e-05, 'train/grad_norm_pre_clip': 0.01213168166577816, 'timing/data': 0.0003543900093063712, 'timing/model': 0.3943245419941377, 'learning_rate/qwen_vl_interface': 5.763344541565091e-07, 'learning_rate/action_model': 5.7633445415650905e-06, 'epoch': 0.35}) +[2026-10-03 19:53:25,760][starVLA.training.train_starvla][INFO] - Step 4721, Loss: {'train/loss': 0.043666958808898926, 'train/grad_norm_pre_clip': 65.73161315917969, 'timing/data': 0.0002842799876816571, 'timing/model': 0.3960115119989496, 'learning_rate/qwen_vl_interface': 5.75791633779998e-07, 'learning_rate/action_model': 5.75791633779998e-06, 'epoch': 0.35}) +[2026-10-03 19:53:26,468][starVLA.training.train_starvla][INFO] - Step 4722, Loss: {'train/loss': 9.633527952246368e-05, 'train/grad_norm_pre_clip': 0.02945244126021862, 'timing/data': 0.00036078999983146787, 'timing/model': 0.3953632620105054, 'learning_rate/qwen_vl_interface': 5.752507347939857e-07, 'learning_rate/action_model': 5.752507347939858e-06, 'epoch': 0.35}) +[2026-10-03 19:53:27,178][starVLA.training.train_starvla][INFO] - Step 4723, Loss: {'train/loss': 0.12238038331270218, 'train/grad_norm_pre_clip': 83.87860107421875, 'timing/data': 0.00022437999723479152, 'timing/model': 0.3956311009824276, 'learning_rate/qwen_vl_interface': 5.74711757420816e-07, 'learning_rate/action_model': 5.747117574208159e-06, 'epoch': 0.35}) +[2026-10-03 19:53:27,887][starVLA.training.train_starvla][INFO] - Step 4724, Loss: {'train/loss': 5.3362000471679494e-05, 'train/grad_norm_pre_clip': 0.016875606030225754, 'timing/data': 0.00035821000346913934, 'timing/model': 0.3956050819833763, 'learning_rate/qwen_vl_interface': 5.74174701882042e-07, 'learning_rate/action_model': 5.74174701882042e-06, 'epoch': 0.35}) +[2026-10-03 19:53:28,792][starVLA.training.train_starvla][INFO] - Step 4725, Loss: {'train/loss': 0.00016754670650698245, 'train/grad_norm_pre_clip': 0.08149336278438568, 'timing/data': 0.0002945800079032779, 'timing/model': 0.3941830910043791, 'learning_rate/qwen_vl_interface': 5.736395683984262e-07, 'learning_rate/action_model': 5.736395683984261e-06, 'epoch': 0.35}) +[2026-10-03 19:53:29,500][starVLA.training.train_starvla][INFO] - Step 4726, Loss: {'train/loss': 0.00014075092622078955, 'train/grad_norm_pre_clip': 0.05650194361805916, 'timing/data': 0.00022891999105922878, 'timing/model': 0.3957097709935624, 'learning_rate/qwen_vl_interface': 5.731063571899421e-07, 'learning_rate/action_model': 5.731063571899421e-06, 'epoch': 0.35}) +[2026-10-03 19:53:30,209][starVLA.training.train_starvla][INFO] - Step 4727, Loss: {'train/loss': 0.0001387293596053496, 'train/grad_norm_pre_clip': 0.02899652160704136, 'timing/data': 0.00033948998316191137, 'timing/model': 0.39568015100667253, 'learning_rate/qwen_vl_interface': 5.72575068475773e-07, 'learning_rate/action_model': 5.72575068475773e-06, 'epoch': 0.35}) +[2026-10-03 19:53:30,917][starVLA.training.train_starvla][INFO] - Step 4728, Loss: {'train/loss': 0.0027222284115850925, 'train/grad_norm_pre_clip': 6.885603427886963, 'timing/data': 0.0002933490031864494, 'timing/model': 0.3955827119934838, 'learning_rate/qwen_vl_interface': 5.720457024743115e-07, 'learning_rate/action_model': 5.720457024743115e-06, 'epoch': 0.35}) +[2026-10-03 19:53:31,624][starVLA.training.train_starvla][INFO] - Step 4729, Loss: {'train/loss': 0.0001916642941068858, 'train/grad_norm_pre_clip': 0.10937964916229248, 'timing/data': 0.00022091000573709607, 'timing/model': 0.3956982620002236, 'learning_rate/qwen_vl_interface': 5.715182594031601e-07, 'learning_rate/action_model': 5.7151825940316e-06, 'epoch': 0.35}) +[2026-10-03 19:53:32,534][starVLA.training.train_starvla][INFO] - Step 4730, Loss: {'train/loss': 0.03599178418517113, 'train/grad_norm_pre_clip': 72.04936218261719, 'timing/data': 0.00029336000443436205, 'timing/model': 0.3948403109970968, 'learning_rate/qwen_vl_interface': 5.709927394791305e-07, 'learning_rate/action_model': 5.709927394791305e-06, 'epoch': 0.35}) +[2026-10-03 19:53:33,242][starVLA.training.train_starvla][INFO] - Step 4731, Loss: {'train/loss': 0.00021332362666726112, 'train/grad_norm_pre_clip': 0.09357038885354996, 'timing/data': 0.00033552999957464635, 'timing/model': 0.39562516097794287, 'learning_rate/qwen_vl_interface': 5.70469142918244e-07, 'learning_rate/action_model': 5.704691429182441e-06, 'epoch': 0.35}) +[2026-10-03 19:53:33,951][starVLA.training.train_starvla][INFO] - Step 4732, Loss: {'train/loss': 9.265440166927874e-05, 'train/grad_norm_pre_clip': 0.03174653276801109, 'timing/data': 0.0003355699882376939, 'timing/model': 0.3955944010231178, 'learning_rate/qwen_vl_interface': 5.699474699357317e-07, 'learning_rate/action_model': 5.699474699357317e-06, 'epoch': 0.35}) +[2026-10-03 19:53:34,659][starVLA.training.train_starvla][INFO] - Step 4733, Loss: {'train/loss': 0.0001746055786497891, 'train/grad_norm_pre_clip': 0.2306271344423294, 'timing/data': 0.0004074600001331419, 'timing/model': 0.39502143202116713, 'learning_rate/qwen_vl_interface': 5.694277207460331e-07, 'learning_rate/action_model': 5.694277207460332e-06, 'epoch': 0.35}) +[2026-10-03 19:53:35,368][starVLA.training.train_starvla][INFO] - Step 4734, Loss: {'train/loss': 0.4421796202659607, 'train/grad_norm_pre_clip': 125.6305923461914, 'timing/data': 0.0002866400172933936, 'timing/model': 0.39602113101864234, 'learning_rate/qwen_vl_interface': 5.689098955627982e-07, 'learning_rate/action_model': 5.689098955627982e-06, 'epoch': 0.35}) +[2026-10-03 19:53:36,277][starVLA.training.train_starvla][INFO] - Step 4735, Loss: {'train/loss': 2.6515350327827036e-05, 'train/grad_norm_pre_clip': 0.007616425398737192, 'timing/data': 0.00023904998670332134, 'timing/model': 0.5953081720217597, 'learning_rate/qwen_vl_interface': 5.683939945988853e-07, 'learning_rate/action_model': 5.683939945988853e-06, 'epoch': 0.35}) +[2026-10-03 19:53:36,983][starVLA.training.train_starvla][INFO] - Step 4736, Loss: {'train/loss': 0.0003543109050951898, 'train/grad_norm_pre_clip': 0.21773265302181244, 'timing/data': 0.0002326600078959018, 'timing/model': 0.3947090709989425, 'learning_rate/qwen_vl_interface': 5.678800180663608e-07, 'learning_rate/action_model': 5.678800180663609e-06, 'epoch': 0.35}) +[2026-10-03 19:53:37,692][starVLA.training.train_starvla][INFO] - Step 4737, Loss: {'train/loss': 0.004730544053018093, 'train/grad_norm_pre_clip': 11.089129447937012, 'timing/data': 0.00023195001995190978, 'timing/model': 0.3956695410015527, 'learning_rate/qwen_vl_interface': 5.673679661765018e-07, 'learning_rate/action_model': 5.6736796617650175e-06, 'epoch': 0.35}) +[2026-10-03 19:53:38,396][starVLA.training.train_starvla][INFO] - Step 4738, Loss: {'train/loss': 0.13990238308906555, 'train/grad_norm_pre_clip': 97.71006774902344, 'timing/data': 0.0002194400003645569, 'timing/model': 0.3940735810028855, 'learning_rate/qwen_vl_interface': 5.668578391397938e-07, 'learning_rate/action_model': 5.668578391397937e-06, 'epoch': 0.35}) +[2026-10-03 19:53:39,105][starVLA.training.train_starvla][INFO] - Step 4739, Loss: {'train/loss': 0.004869017284363508, 'train/grad_norm_pre_clip': 13.602994918823242, 'timing/data': 0.0003285700222477317, 'timing/model': 0.3956188620068133, 'learning_rate/qwen_vl_interface': 5.663496371659303e-07, 'learning_rate/action_model': 5.663496371659303e-06, 'epoch': 0.35}) +[2026-10-03 19:53:40,013][starVLA.training.train_starvla][INFO] - Step 4740, Loss: {'train/loss': 0.9601997137069702, 'train/grad_norm_pre_clip': 119.62879943847656, 'timing/data': 0.0003757399972528219, 'timing/model': 0.5972014020080678, 'learning_rate/qwen_vl_interface': 5.658433604638141e-07, 'learning_rate/action_model': 5.65843360463814e-06, 'epoch': 0.35}) +[2026-10-03 19:53:40,720][starVLA.training.train_starvla][INFO] - Step 4741, Loss: {'train/loss': 0.014409830793738365, 'train/grad_norm_pre_clip': 23.24163246154785, 'timing/data': 0.0004026500100735575, 'timing/model': 0.3964612520067021, 'learning_rate/qwen_vl_interface': 5.653390092415559e-07, 'learning_rate/action_model': 5.653390092415559e-06, 'epoch': 0.35}) +[2026-10-03 19:53:41,428][starVLA.training.train_starvla][INFO] - Step 4742, Loss: {'train/loss': 8.577661355957389e-05, 'train/grad_norm_pre_clip': 0.037982214242219925, 'timing/data': 0.0003549599787220359, 'timing/model': 0.395995782018872, 'learning_rate/qwen_vl_interface': 5.648365837064759e-07, 'learning_rate/action_model': 5.648365837064759e-06, 'epoch': 0.35}) +[2026-10-03 19:53:42,132][starVLA.training.train_starvla][INFO] - Step 4743, Loss: {'train/loss': 0.040000636130571365, 'train/grad_norm_pre_clip': 115.42552185058594, 'timing/data': 0.00033729997812770307, 'timing/model': 0.3934664709959179, 'learning_rate/qwen_vl_interface': 5.643360840651025e-07, 'learning_rate/action_model': 5.643360840651025e-06, 'epoch': 0.35}) +[2026-10-03 19:53:42,838][starVLA.training.train_starvla][INFO] - Step 4744, Loss: {'train/loss': 0.002968007232993841, 'train/grad_norm_pre_clip': 7.110556125640869, 'timing/data': 0.00021624000510200858, 'timing/model': 0.39429585202015005, 'learning_rate/qwen_vl_interface': 5.638375105231719e-07, 'learning_rate/action_model': 5.638375105231719e-06, 'epoch': 0.35}) +[2026-10-03 19:53:43,747][starVLA.training.train_starvla][INFO] - Step 4745, Loss: {'train/loss': 0.8411812782287598, 'train/grad_norm_pre_clip': 142.314697265625, 'timing/data': 0.0002516800013836473, 'timing/model': 0.5976393320015632, 'learning_rate/qwen_vl_interface': 5.633408632856294e-07, 'learning_rate/action_model': 5.6334086328562945e-06, 'epoch': 0.35}) +[2026-10-03 19:53:44,453][starVLA.training.train_starvla][INFO] - Step 4746, Loss: {'train/loss': 0.0003480361483525485, 'train/grad_norm_pre_clip': 0.9268644452095032, 'timing/data': 0.00023129000328481197, 'timing/model': 0.39352728202356957, 'learning_rate/qwen_vl_interface': 5.628461425566268e-07, 'learning_rate/action_model': 5.6284614255662675e-06, 'epoch': 0.35}) +[2026-10-03 19:53:45,159][starVLA.training.train_starvla][INFO] - Step 4747, Loss: {'train/loss': 0.00039790934533812106, 'train/grad_norm_pre_clip': 0.5499740242958069, 'timing/data': 0.00021659999038092792, 'timing/model': 0.394672691996675, 'learning_rate/qwen_vl_interface': 5.62353348539526e-07, 'learning_rate/action_model': 5.623533485395259e-06, 'epoch': 0.35}) +[2026-10-03 19:53:45,862][starVLA.training.train_starvla][INFO] - Step 4748, Loss: {'train/loss': 0.3131547272205353, 'train/grad_norm_pre_clip': 190.41213989257812, 'timing/data': 0.0003102300106547773, 'timing/model': 0.3926872709998861, 'learning_rate/qwen_vl_interface': 5.618624814368956e-07, 'learning_rate/action_model': 5.618624814368956e-06, 'epoch': 0.35}) +[2026-10-03 19:53:46,566][starVLA.training.train_starvla][INFO] - Step 4749, Loss: {'train/loss': 0.00011704616917995736, 'train/grad_norm_pre_clip': 0.04254563897848129, 'timing/data': 0.0003365399898029864, 'timing/model': 0.3930098109995015, 'learning_rate/qwen_vl_interface': 5.613735414505126e-07, 'learning_rate/action_model': 5.613735414505125e-06, 'epoch': 0.35}) +[2026-10-03 19:53:47,272][starVLA.training.train_starvla][INFO] - Step 4750, Loss: {'train/loss': 0.024215197190642357, 'train/grad_norm_pre_clip': 145.07749938964844, 'timing/data': 0.00032568001188337803, 'timing/model': 0.39429110201308504, 'learning_rate/qwen_vl_interface': 5.608865287813617e-07, 'learning_rate/action_model': 5.608865287813616e-06, 'epoch': 0.35}) +[2026-10-03 19:53:48,174][starVLA.training.train_starvla][INFO] - Step 4751, Loss: {'train/loss': 0.00019055417214985937, 'train/grad_norm_pre_clip': 0.11918539553880692, 'timing/data': 0.00036050999187864363, 'timing/model': 0.39298739199875854, 'learning_rate/qwen_vl_interface': 5.604014436296362e-07, 'learning_rate/action_model': 5.604014436296362e-06, 'epoch': 0.35}) +[2026-10-03 19:53:48,880][starVLA.training.train_starvla][INFO] - Step 4752, Loss: {'train/loss': 0.0006548690726049244, 'train/grad_norm_pre_clip': 2.3581130504608154, 'timing/data': 0.0003471899835858494, 'timing/model': 0.3935925920086447, 'learning_rate/qwen_vl_interface': 5.599182861947348e-07, 'learning_rate/action_model': 5.599182861947347e-06, 'epoch': 0.35}) +[2026-10-03 19:53:49,588][starVLA.training.train_starvla][INFO] - Step 4753, Loss: {'train/loss': 0.0004167906299699098, 'train/grad_norm_pre_clip': 0.20806194841861725, 'timing/data': 0.00033412998891435564, 'timing/model': 0.3959088219853584, 'learning_rate/qwen_vl_interface': 5.594370566752665e-07, 'learning_rate/action_model': 5.5943705667526645e-06, 'epoch': 0.35}) +[2026-10-03 19:53:50,293][starVLA.training.train_starvla][INFO] - Step 4754, Loss: {'train/loss': 0.0004434232832863927, 'train/grad_norm_pre_clip': 0.515149712562561, 'timing/data': 0.0003557500021997839, 'timing/model': 0.3946703409892507, 'learning_rate/qwen_vl_interface': 5.589577552690456e-07, 'learning_rate/action_model': 5.589577552690456e-06, 'epoch': 0.35}) +[2026-10-03 19:53:50,997][starVLA.training.train_starvla][INFO] - Step 4755, Loss: {'train/loss': 3.0352597605087794e-05, 'train/grad_norm_pre_clip': 0.00810216087847948, 'timing/data': 0.0003578999894671142, 'timing/model': 0.3934088009991683, 'learning_rate/qwen_vl_interface': 5.584803821730961e-07, 'learning_rate/action_model': 5.58480382173096e-06, 'epoch': 0.35}) +[2026-10-03 19:53:51,902][starVLA.training.train_starvla][INFO] - Step 4756, Loss: {'train/loss': 0.48949822783470154, 'train/grad_norm_pre_clip': 51.025787353515625, 'timing/data': 0.0003466899797786027, 'timing/model': 0.5947675920033362, 'learning_rate/qwen_vl_interface': 5.580049375836466e-07, 'learning_rate/action_model': 5.580049375836466e-06, 'epoch': 0.35}) +[2026-10-03 19:53:52,607][starVLA.training.train_starvla][INFO] - Step 4757, Loss: {'train/loss': 0.00025130639551207423, 'train/grad_norm_pre_clip': 0.10052412003278732, 'timing/data': 0.00029110998730175197, 'timing/model': 0.3945648609951604, 'learning_rate/qwen_vl_interface': 5.575314216961356e-07, 'learning_rate/action_model': 5.575314216961356e-06, 'epoch': 0.35}) +[2026-10-03 19:53:53,312][starVLA.training.train_starvla][INFO] - Step 4758, Loss: {'train/loss': 0.16505353152751923, 'train/grad_norm_pre_clip': 94.2733154296875, 'timing/data': 0.0003202799998689443, 'timing/model': 0.3950783709879033, 'learning_rate/qwen_vl_interface': 5.570598347052073e-07, 'learning_rate/action_model': 5.570598347052072e-06, 'epoch': 0.35}) +[2026-10-03 19:53:54,019][starVLA.training.train_starvla][INFO] - Step 4759, Loss: {'train/loss': 0.15504153072834015, 'train/grad_norm_pre_clip': 291.72918701171875, 'timing/data': 0.0003839799901470542, 'timing/model': 0.39463220100151375, 'learning_rate/qwen_vl_interface': 5.565901768047126e-07, 'learning_rate/action_model': 5.5659017680471255e-06, 'epoch': 0.35}) +[2026-10-03 19:53:54,724][starVLA.training.train_starvla][INFO] - Step 4760, Loss: {'train/loss': 0.02665039151906967, 'train/grad_norm_pre_clip': 55.726318359375, 'timing/data': 0.00036460001138038933, 'timing/model': 0.3953583510010503, 'learning_rate/qwen_vl_interface': 5.561224481877105e-07, 'learning_rate/action_model': 5.561224481877105e-06, 'epoch': 0.35}) +[2026-10-03 19:53:55,431][starVLA.training.train_starvla][INFO] - Step 4761, Loss: {'train/loss': 2.63371694018133e-05, 'train/grad_norm_pre_clip': 0.005797352641820908, 'timing/data': 0.00035658999695442617, 'timing/model': 0.39532701097778045, 'learning_rate/qwen_vl_interface': 5.55656649046467e-07, 'learning_rate/action_model': 5.556566490464669e-06, 'epoch': 0.35}) +[2026-10-03 19:53:56,337][starVLA.training.train_starvla][INFO] - Step 4762, Loss: {'train/loss': 8.909318421501666e-05, 'train/grad_norm_pre_clip': 0.10728698968887329, 'timing/data': 0.00024935000692494214, 'timing/model': 0.39627002098131925, 'learning_rate/qwen_vl_interface': 5.551927795724544e-07, 'learning_rate/action_model': 5.551927795724543e-06, 'epoch': 0.35}) +[2026-10-03 19:53:57,044][starVLA.training.train_starvla][INFO] - Step 4763, Loss: {'train/loss': 0.00020043518452439457, 'train/grad_norm_pre_clip': 0.08878827095031738, 'timing/data': 0.00027238999609835446, 'timing/model': 0.3958041910082102, 'learning_rate/qwen_vl_interface': 5.547308399563515e-07, 'learning_rate/action_model': 5.547308399563514e-06, 'epoch': 0.35}) +[2026-10-03 19:53:57,753][starVLA.training.train_starvla][INFO] - Step 4764, Loss: {'train/loss': 0.2318516969680786, 'train/grad_norm_pre_clip': 130.06988525390625, 'timing/data': 0.0003337800153531134, 'timing/model': 0.3953746919869445, 'learning_rate/qwen_vl_interface': 5.542708303880442e-07, 'learning_rate/action_model': 5.542708303880442e-06, 'epoch': 0.35}) +[2026-10-03 19:53:58,460][starVLA.training.train_starvla][INFO] - Step 4765, Loss: {'train/loss': 0.3467956781387329, 'train/grad_norm_pre_clip': 120.01187896728516, 'timing/data': 0.0003239600046072155, 'timing/model': 0.3946938709996175, 'learning_rate/qwen_vl_interface': 5.538127510566258e-07, 'learning_rate/action_model': 5.538127510566258e-06, 'epoch': 0.35}) +[2026-10-03 19:53:59,168][starVLA.training.train_starvla][INFO] - Step 4766, Loss: {'train/loss': 0.5640716552734375, 'train/grad_norm_pre_clip': 47.914493560791016, 'timing/data': 0.00025615000049583614, 'timing/model': 0.39478220100863837, 'learning_rate/qwen_vl_interface': 5.533566021503944e-07, 'learning_rate/action_model': 5.533566021503943e-06, 'epoch': 0.35}) +[2026-10-03 19:54:00,082][starVLA.training.train_starvla][INFO] - Step 4767, Loss: {'train/loss': 0.06216876208782196, 'train/grad_norm_pre_clip': 100.44771575927734, 'timing/data': 0.0003370500053279102, 'timing/model': 0.3940845520119183, 'learning_rate/qwen_vl_interface': 5.529023838568566e-07, 'learning_rate/action_model': 5.529023838568566e-06, 'epoch': 0.35}) +[2026-10-03 19:54:00,791][starVLA.training.train_starvla][INFO] - Step 4768, Loss: {'train/loss': 0.005207367707043886, 'train/grad_norm_pre_clip': 8.975069999694824, 'timing/data': 0.00036727997940033674, 'timing/model': 0.3965555519971531, 'learning_rate/qwen_vl_interface': 5.524500963627239e-07, 'learning_rate/action_model': 5.524500963627239e-06, 'epoch': 0.35}) +[2026-10-03 19:54:01,499][starVLA.training.train_starvla][INFO] - Step 4769, Loss: {'train/loss': 0.0001666708558332175, 'train/grad_norm_pre_clip': 0.3135126829147339, 'timing/data': 0.0003635000030044466, 'timing/model': 0.3945633309776895, 'learning_rate/qwen_vl_interface': 5.519997398539145e-07, 'learning_rate/action_model': 5.519997398539145e-06, 'epoch': 0.35}) +[2026-10-03 19:54:02,208][starVLA.training.train_starvla][INFO] - Step 4770, Loss: {'train/loss': 0.13360615074634552, 'train/grad_norm_pre_clip': 107.7562255859375, 'timing/data': 0.0003590199921745807, 'timing/model': 0.39588728098897263, 'learning_rate/qwen_vl_interface': 5.515513145155531e-07, 'learning_rate/action_model': 5.515513145155531e-06, 'epoch': 0.35}) +[2026-10-03 19:54:02,918][starVLA.training.train_starvla][INFO] - Step 4771, Loss: {'train/loss': 0.00011051766341552138, 'train/grad_norm_pre_clip': 0.0835178941488266, 'timing/data': 0.0002707800013013184, 'timing/model': 0.3964885910099838, 'learning_rate/qwen_vl_interface': 5.511048205319702e-07, 'learning_rate/action_model': 5.511048205319701e-06, 'epoch': 0.35}) +[2026-10-03 19:54:03,834][starVLA.training.train_starvla][INFO] - Step 4772, Loss: {'train/loss': 4.877436003880575e-05, 'train/grad_norm_pre_clip': 0.01555600855499506, 'timing/data': 0.00036452000495046377, 'timing/model': 0.6029171419795603, 'learning_rate/qwen_vl_interface': 5.506602580867034e-07, 'learning_rate/action_model': 5.506602580867034e-06, 'epoch': 0.35}) +[2026-10-03 19:54:04,541][starVLA.training.train_starvla][INFO] - Step 4773, Loss: {'train/loss': 0.0003219528589397669, 'train/grad_norm_pre_clip': 0.20896150171756744, 'timing/data': 0.0003583100042305887, 'timing/model': 0.39521328097907826, 'learning_rate/qwen_vl_interface': 5.502176273624942e-07, 'learning_rate/action_model': 5.502176273624942e-06, 'epoch': 0.35}) +[2026-10-03 19:54:05,251][starVLA.training.train_starvla][INFO] - Step 4774, Loss: {'train/loss': 0.2036937177181244, 'train/grad_norm_pre_clip': 140.43246459960938, 'timing/data': 0.0003144599904771894, 'timing/model': 0.3967418209940661, 'learning_rate/qwen_vl_interface': 5.497769285412921e-07, 'learning_rate/action_model': 5.497769285412921e-06, 'epoch': 0.35}) +[2026-10-03 19:54:05,963][starVLA.training.train_starvla][INFO] - Step 4775, Loss: {'train/loss': 0.0012157537275925279, 'train/grad_norm_pre_clip': 1.71185302734375, 'timing/data': 0.00033672997960820794, 'timing/model': 0.39797682099742815, 'learning_rate/qwen_vl_interface': 5.493381618042519e-07, 'learning_rate/action_model': 5.493381618042518e-06, 'epoch': 0.35}) +[2026-10-03 19:54:06,674][starVLA.training.train_starvla][INFO] - Step 4776, Loss: {'train/loss': 0.13947322964668274, 'train/grad_norm_pre_clip': 138.4783477783203, 'timing/data': 0.00037161100772209466, 'timing/model': 0.39652200997807086, 'learning_rate/qwen_vl_interface': 5.489013273317343e-07, 'learning_rate/action_model': 5.489013273317343e-06, 'epoch': 0.35}) +[2026-10-03 19:54:07,385][starVLA.training.train_starvla][INFO] - Step 4777, Loss: {'train/loss': 0.00012413787771947682, 'train/grad_norm_pre_clip': 0.09214085340499878, 'timing/data': 0.0003133799764327705, 'timing/model': 0.39730886099277996, 'learning_rate/qwen_vl_interface': 5.484664253033042e-07, 'learning_rate/action_model': 5.484664253033042e-06, 'epoch': 0.35}) +[2026-10-03 19:54:08,293][starVLA.training.train_starvla][INFO] - Step 4778, Loss: {'train/loss': 0.0017191823571920395, 'train/grad_norm_pre_clip': 2.0380988121032715, 'timing/data': 0.00022729000193066895, 'timing/model': 0.39611084197531454, 'learning_rate/qwen_vl_interface': 5.480334558977351e-07, 'learning_rate/action_model': 5.480334558977351e-06, 'epoch': 0.35}) +[2026-10-03 19:54:09,004][starVLA.training.train_starvla][INFO] - Step 4779, Loss: {'train/loss': 0.00010605472198221833, 'train/grad_norm_pre_clip': 0.03420088812708855, 'timing/data': 0.00035026000114157796, 'timing/model': 0.39784189098281786, 'learning_rate/qwen_vl_interface': 5.476024192930029e-07, 'learning_rate/action_model': 5.47602419293003e-06, 'epoch': 0.35}) +[2026-10-03 19:54:09,712][starVLA.training.train_starvla][INFO] - Step 4780, Loss: {'train/loss': 0.000376377982320264, 'train/grad_norm_pre_clip': 0.3583757281303406, 'timing/data': 0.0003705199924297631, 'timing/model': 0.39593805201002397, 'learning_rate/qwen_vl_interface': 5.471733156662914e-07, 'learning_rate/action_model': 5.4717331566629135e-06, 'epoch': 0.35}) +[2026-10-03 19:54:10,422][starVLA.training.train_starvla][INFO] - Step 4781, Loss: {'train/loss': 0.19639931619167328, 'train/grad_norm_pre_clip': 72.37188720703125, 'timing/data': 0.0004288300115149468, 'timing/model': 0.3965922000061255, 'learning_rate/qwen_vl_interface': 5.467461451939882e-07, 'learning_rate/action_model': 5.467461451939882e-06, 'epoch': 0.35}) +[2026-10-03 19:54:11,130][starVLA.training.train_starvla][INFO] - Step 4782, Loss: {'train/loss': 0.02541174553334713, 'train/grad_norm_pre_clip': 67.90889739990234, 'timing/data': 0.0003542100021149963, 'timing/model': 0.39578559200163, 'learning_rate/qwen_vl_interface': 5.463209080516873e-07, 'learning_rate/action_model': 5.463209080516873e-06, 'epoch': 0.35}) +[2026-10-03 19:54:12,038][starVLA.training.train_starvla][INFO] - Step 4783, Loss: {'train/loss': 0.00025394666590727866, 'train/grad_norm_pre_clip': 0.459774911403656, 'timing/data': 0.00028360000578686595, 'timing/model': 0.39576328100520186, 'learning_rate/qwen_vl_interface': 5.458976044141875e-07, 'learning_rate/action_model': 5.458976044141875e-06, 'epoch': 0.35}) +[2026-10-03 19:54:12,747][starVLA.training.train_starvla][INFO] - Step 4784, Loss: {'train/loss': 2.9606417228933424e-05, 'train/grad_norm_pre_clip': 0.011052481830120087, 'timing/data': 0.00026903001707978547, 'timing/model': 0.39519859198480844, 'learning_rate/qwen_vl_interface': 5.454762344554931e-07, 'learning_rate/action_model': 5.4547623445549316e-06, 'epoch': 0.35}) +[2026-10-03 19:54:13,456][starVLA.training.train_starvla][INFO] - Step 4785, Loss: {'train/loss': 6.34175303275697e-05, 'train/grad_norm_pre_clip': 0.036940593272447586, 'timing/data': 0.00035509001463651657, 'timing/model': 0.3962583420216106, 'learning_rate/qwen_vl_interface': 5.450567983488137e-07, 'learning_rate/action_model': 5.450567983488137e-06, 'epoch': 0.35}) +[2026-10-03 19:54:14,163][starVLA.training.train_starvla][INFO] - Step 4786, Loss: {'train/loss': 0.43050867319107056, 'train/grad_norm_pre_clip': 90.2253189086914, 'timing/data': 0.00023412000155076385, 'timing/model': 0.3959692310017999, 'learning_rate/qwen_vl_interface': 5.446392962665629e-07, 'learning_rate/action_model': 5.4463929626656296e-06, 'epoch': 0.35}) +[2026-10-03 19:54:14,873][starVLA.training.train_starvla][INFO] - Step 4787, Loss: {'train/loss': 0.09721086174249649, 'train/grad_norm_pre_clip': 150.4619598388672, 'timing/data': 0.0002399299992248416, 'timing/model': 0.3959055719897151, 'learning_rate/qwen_vl_interface': 5.442237283803605e-07, 'learning_rate/action_model': 5.442237283803604e-06, 'epoch': 0.35}) +[2026-10-03 19:54:15,780][starVLA.training.train_starvla][INFO] - Step 4788, Loss: {'train/loss': 0.00034382170997560024, 'train/grad_norm_pre_clip': 0.2272835373878479, 'timing/data': 0.00024921001750044525, 'timing/model': 0.5943485619791318, 'learning_rate/qwen_vl_interface': 5.43810094861031e-07, 'learning_rate/action_model': 5.43810094861031e-06, 'epoch': 0.35}) +[2026-10-03 19:54:16,488][starVLA.training.train_starvla][INFO] - Step 4789, Loss: {'train/loss': 0.0033713344018906355, 'train/grad_norm_pre_clip': 5.9455885887146, 'timing/data': 0.00022755999816581607, 'timing/model': 0.3968731419881806, 'learning_rate/qwen_vl_interface': 5.433983958786027e-07, 'learning_rate/action_model': 5.433983958786027e-06, 'epoch': 0.35}) +[2026-10-03 19:54:17,198][starVLA.training.train_starvla][INFO] - Step 4790, Loss: {'train/loss': 0.018252182751893997, 'train/grad_norm_pre_clip': 20.818269729614258, 'timing/data': 0.00034984000376425683, 'timing/model': 0.39588822101359256, 'learning_rate/qwen_vl_interface': 5.429886316023102e-07, 'learning_rate/action_model': 5.429886316023101e-06, 'epoch': 0.35}) +[2026-10-03 19:54:17,910][starVLA.training.train_starvla][INFO] - Step 4791, Loss: {'train/loss': 0.0004652117204386741, 'train/grad_norm_pre_clip': 0.7275385856628418, 'timing/data': 0.0003565299848560244, 'timing/model': 0.39792823101743124, 'learning_rate/qwen_vl_interface': 5.425808022005916e-07, 'learning_rate/action_model': 5.425808022005916e-06, 'epoch': 0.35}) +[2026-10-03 19:54:18,621][starVLA.training.train_starvla][INFO] - Step 4792, Loss: {'train/loss': 0.0007565531996078789, 'train/grad_norm_pre_clip': 1.073708176612854, 'timing/data': 0.0002622400061227381, 'timing/model': 0.39749644199036993, 'learning_rate/qwen_vl_interface': 5.421749078410903e-07, 'learning_rate/action_model': 5.421749078410903e-06, 'epoch': 0.35}) +[2026-10-03 19:54:19,534][starVLA.training.train_starvla][INFO] - Step 4793, Loss: {'train/loss': 9.38639568630606e-05, 'train/grad_norm_pre_clip': 0.06234682723879814, 'timing/data': 0.0003496799909044057, 'timing/model': 0.5986530919908546, 'learning_rate/qwen_vl_interface': 5.417709486906547e-07, 'learning_rate/action_model': 5.417709486906547e-06, 'epoch': 0.35}) +[2026-10-03 19:54:20,242][starVLA.training.train_starvla][INFO] - Step 4794, Loss: {'train/loss': 0.00012420385610312223, 'train/grad_norm_pre_clip': 0.05532440170645714, 'timing/data': 0.00025378999998793006, 'timing/model': 0.39666802098508924, 'learning_rate/qwen_vl_interface': 5.413689249153367e-07, 'learning_rate/action_model': 5.413689249153366e-06, 'epoch': 0.35}) +[2026-10-03 19:54:20,953][starVLA.training.train_starvla][INFO] - Step 4795, Loss: {'train/loss': 0.4054257273674011, 'train/grad_norm_pre_clip': 57.23634338378906, 'timing/data': 0.0003159100015182048, 'timing/model': 0.39710912099690177, 'learning_rate/qwen_vl_interface': 5.409688366803931e-07, 'learning_rate/action_model': 5.409688366803931e-06, 'epoch': 0.35}) +[2026-10-03 19:54:21,663][starVLA.training.train_starvla][INFO] - Step 4796, Loss: {'train/loss': 0.5852504372596741, 'train/grad_norm_pre_clip': 195.90927124023438, 'timing/data': 0.00034240001696161926, 'timing/model': 0.3972104119893629, 'learning_rate/qwen_vl_interface': 5.405706841502853e-07, 'learning_rate/action_model': 5.4057068415028525e-06, 'epoch': 0.35}) +[2026-10-03 19:54:22,376][starVLA.training.train_starvla][INFO] - Step 4797, Loss: {'train/loss': 6.609363481402397e-05, 'train/grad_norm_pre_clip': 0.01385075505822897, 'timing/data': 0.00035697000566869974, 'timing/model': 0.3969251609814819, 'learning_rate/qwen_vl_interface': 5.401744674886781e-07, 'learning_rate/action_model': 5.40174467488678e-06, 'epoch': 0.35}) +[2026-10-03 19:54:23,288][starVLA.training.train_starvla][INFO] - Step 4798, Loss: {'train/loss': 0.00018380094843450934, 'train/grad_norm_pre_clip': 0.14131174981594086, 'timing/data': 0.00021889997879043221, 'timing/model': 0.5990142009977717, 'learning_rate/qwen_vl_interface': 5.397801868584418e-07, 'learning_rate/action_model': 5.397801868584417e-06, 'epoch': 0.35}) +[2026-10-03 19:54:23,998][starVLA.training.train_starvla][INFO] - Step 4799, Loss: {'train/loss': 0.00046403694432228804, 'train/grad_norm_pre_clip': 0.6173579096794128, 'timing/data': 0.00034158999915234745, 'timing/model': 0.39811868997639976, 'learning_rate/qwen_vl_interface': 5.393878424216502e-07, 'learning_rate/action_model': 5.393878424216501e-06, 'epoch': 0.35}) +[2026-10-03 19:54:24,711][starVLA.training.train_starvla][INFO] - Step 4800, Loss: {'train/loss': 0.00012221794167999178, 'train/grad_norm_pre_clip': 0.10906988382339478, 'timing/data': 0.00033007998717948794, 'timing/model': 0.3980680009990465, 'learning_rate/qwen_vl_interface': 5.389974343395813e-07, 'learning_rate/action_model': 5.389974343395812e-06, 'epoch': 0.35}) +[2026-10-03 19:54:25,420][starVLA.training.train_starvla][INFO] - Step 4801, Loss: {'train/loss': 4.1364044591318816e-05, 'train/grad_norm_pre_clip': 0.008255597203969955, 'timing/data': 0.00022717000683769584, 'timing/model': 0.3972920609812718, 'learning_rate/qwen_vl_interface': 5.38608962772717e-07, 'learning_rate/action_model': 5.38608962772717e-06, 'epoch': 0.35}) +[2026-10-03 19:54:26,128][starVLA.training.train_starvla][INFO] - Step 4802, Loss: {'train/loss': 0.0001289768988499418, 'train/grad_norm_pre_clip': 0.13176508247852325, 'timing/data': 0.00022099001216702163, 'timing/model': 0.39614182099467143, 'learning_rate/qwen_vl_interface': 5.382224278807429e-07, 'learning_rate/action_model': 5.382224278807429e-06, 'epoch': 0.35}) +[2026-10-03 19:54:26,840][starVLA.training.train_starvla][INFO] - Step 4803, Loss: {'train/loss': 0.0002165359619539231, 'train/grad_norm_pre_clip': 0.21207404136657715, 'timing/data': 0.00036872000782750547, 'timing/model': 0.3975700409791898, 'learning_rate/qwen_vl_interface': 5.378378298225496e-07, 'learning_rate/action_model': 5.378378298225496e-06, 'epoch': 0.35}) +[2026-10-03 19:54:27,754][starVLA.training.train_starvla][INFO] - Step 4804, Loss: {'train/loss': 3.073858533753082e-05, 'train/grad_norm_pre_clip': 0.011751527898013592, 'timing/data': 0.0002802800154313445, 'timing/model': 0.39702545100590214, 'learning_rate/qwen_vl_interface': 5.374551687562306e-07, 'learning_rate/action_model': 5.374551687562306e-06, 'epoch': 0.35}) +[2026-10-03 19:54:28,462][starVLA.training.train_starvla][INFO] - Step 4805, Loss: {'train/loss': 0.05454646423459053, 'train/grad_norm_pre_clip': 69.22628021240234, 'timing/data': 0.00034759999834932387, 'timing/model': 0.39450016198679805, 'learning_rate/qwen_vl_interface': 5.37074444839083e-07, 'learning_rate/action_model': 5.370744448390829e-06, 'epoch': 0.35}) +[2026-10-03 19:54:29,169][starVLA.training.train_starvla][INFO] - Step 4806, Loss: {'train/loss': 6.661712541244924e-05, 'train/grad_norm_pre_clip': 0.02111157216131687, 'timing/data': 0.00023947999579831958, 'timing/model': 0.39519108200329356, 'learning_rate/qwen_vl_interface': 5.366956582276085e-07, 'learning_rate/action_model': 5.366956582276084e-06, 'epoch': 0.35}) +[2026-10-03 19:54:29,881][starVLA.training.train_starvla][INFO] - Step 4807, Loss: {'train/loss': 0.00016662939742673188, 'train/grad_norm_pre_clip': 0.09561215341091156, 'timing/data': 0.00022464001085609198, 'timing/model': 0.39820341000449844, 'learning_rate/qwen_vl_interface': 5.363188090775122e-07, 'learning_rate/action_model': 5.363188090775122e-06, 'epoch': 0.35}) +[2026-10-03 19:54:30,592][starVLA.training.train_starvla][INFO] - Step 4808, Loss: {'train/loss': 7.491232099710032e-05, 'train/grad_norm_pre_clip': 0.0640505850315094, 'timing/data': 0.0002630600065458566, 'timing/model': 0.3972385709930677, 'learning_rate/qwen_vl_interface': 5.359438975437018e-07, 'learning_rate/action_model': 5.359438975437017e-06, 'epoch': 0.35}) +[2026-10-03 19:54:31,501][starVLA.training.train_starvla][INFO] - Step 4809, Loss: {'train/loss': 0.5218616127967834, 'train/grad_norm_pre_clip': 102.5524673461914, 'timing/data': 0.0002699699834920466, 'timing/model': 0.5954905530088581, 'learning_rate/qwen_vl_interface': 5.355709237802895e-07, 'learning_rate/action_model': 5.3557092378028955e-06, 'epoch': 0.35}) +[2026-10-03 19:54:32,210][starVLA.training.train_starvla][INFO] - Step 4810, Loss: {'train/loss': 8.012980833882466e-05, 'train/grad_norm_pre_clip': 0.05328911915421486, 'timing/data': 0.00023626998881809413, 'timing/model': 0.395803021005122, 'learning_rate/qwen_vl_interface': 5.351998879405911e-07, 'learning_rate/action_model': 5.351998879405911e-06, 'epoch': 0.35}) +[2026-10-03 19:54:32,919][starVLA.training.train_starvla][INFO] - Step 4811, Loss: {'train/loss': 0.012118835933506489, 'train/grad_norm_pre_clip': 18.88071632385254, 'timing/data': 0.00023640997824259102, 'timing/model': 0.39546923100715503, 'learning_rate/qwen_vl_interface': 5.348307901771256e-07, 'learning_rate/action_model': 5.348307901771256e-06, 'epoch': 0.35}) +[2026-10-03 19:54:33,627][starVLA.training.train_starvla][INFO] - Step 4812, Loss: {'train/loss': 0.0001275393005926162, 'train/grad_norm_pre_clip': 0.08725620061159134, 'timing/data': 0.0002872399927582592, 'timing/model': 0.39531472098315135, 'learning_rate/qwen_vl_interface': 5.344636306416147e-07, 'learning_rate/action_model': 5.344636306416146e-06, 'epoch': 0.35}) +[2026-10-03 19:54:34,335][starVLA.training.train_starvla][INFO] - Step 4813, Loss: {'train/loss': 0.16942501068115234, 'train/grad_norm_pre_clip': 62.70575714111328, 'timing/data': 0.00022421000176109374, 'timing/model': 0.3947802720067557, 'learning_rate/qwen_vl_interface': 5.340984094849836e-07, 'learning_rate/action_model': 5.340984094849836e-06, 'epoch': 0.35}) +[2026-10-03 19:54:35,042][starVLA.training.train_starvla][INFO] - Step 4814, Loss: {'train/loss': 0.5198941826820374, 'train/grad_norm_pre_clip': 35.553104400634766, 'timing/data': 0.00022116999025456607, 'timing/model': 0.3960803020163439, 'learning_rate/qwen_vl_interface': 5.337351268573606e-07, 'learning_rate/action_model': 5.337351268573606e-06, 'epoch': 0.35}) +[2026-10-03 19:54:35,953][starVLA.training.train_starvla][INFO] - Step 4815, Loss: {'train/loss': 0.0276488047093153, 'train/grad_norm_pre_clip': 57.943973541259766, 'timing/data': 0.0002808900026138872, 'timing/model': 0.3963314220018219, 'learning_rate/qwen_vl_interface': 5.33373782908079e-07, 'learning_rate/action_model': 5.333737829080789e-06, 'epoch': 0.35}) +[2026-10-03 19:54:36,665][starVLA.training.train_starvla][INFO] - Step 4816, Loss: {'train/loss': 0.00031296355882659554, 'train/grad_norm_pre_clip': 0.2992592751979828, 'timing/data': 0.0002275099977850914, 'timing/model': 0.39704514099867083, 'learning_rate/qwen_vl_interface': 5.330143777856718e-07, 'learning_rate/action_model': 5.330143777856718e-06, 'epoch': 0.35}) +[2026-10-03 19:54:37,376][starVLA.training.train_starvla][INFO] - Step 4817, Loss: {'train/loss': 3.15885481541045e-05, 'train/grad_norm_pre_clip': 0.012916713021695614, 'timing/data': 0.00027039999258704484, 'timing/model': 0.3962912709976081, 'learning_rate/qwen_vl_interface': 5.326569116378786e-07, 'learning_rate/action_model': 5.326569116378786e-06, 'epoch': 0.35}) +[2026-10-03 19:54:38,083][starVLA.training.train_starvla][INFO] - Step 4818, Loss: {'train/loss': 0.28953710198402405, 'train/grad_norm_pre_clip': 88.39667510986328, 'timing/data': 0.00035332900006324053, 'timing/model': 0.39512678200844675, 'learning_rate/qwen_vl_interface': 5.323013846116387e-07, 'learning_rate/action_model': 5.323013846116386e-06, 'epoch': 0.35}) +[2026-10-03 19:54:38,793][starVLA.training.train_starvla][INFO] - Step 4819, Loss: {'train/loss': 2.1375279175117612e-05, 'train/grad_norm_pre_clip': 0.00418677506968379, 'timing/data': 0.0004088300047442317, 'timing/model': 0.3951149019994773, 'learning_rate/qwen_vl_interface': 5.31947796853097e-07, 'learning_rate/action_model': 5.3194779685309696e-06, 'epoch': 0.35}) +[2026-10-03 19:54:39,702][starVLA.training.train_starvla][INFO] - Step 4820, Loss: {'train/loss': 0.00011867284047184512, 'train/grad_norm_pre_clip': 0.05141384154558182, 'timing/data': 0.0002998999843839556, 'timing/model': 0.39675508200889453, 'learning_rate/qwen_vl_interface': 5.315961485075992e-07, 'learning_rate/action_model': 5.315961485075993e-06, 'epoch': 0.35}) +[2026-10-03 19:54:40,412][starVLA.training.train_starvla][INFO] - Step 4821, Loss: {'train/loss': 0.00011428730795159936, 'train/grad_norm_pre_clip': 0.0965154692530632, 'timing/data': 0.0003661900118459016, 'timing/model': 0.39624971101875417, 'learning_rate/qwen_vl_interface': 5.312464397196958e-07, 'learning_rate/action_model': 5.312464397196958e-06, 'epoch': 0.35}) +[2026-10-03 19:54:41,119][starVLA.training.train_starvla][INFO] - Step 4822, Loss: {'train/loss': 0.2548632025718689, 'train/grad_norm_pre_clip': 214.404541015625, 'timing/data': 0.0002518499968573451, 'timing/model': 0.39595047099282965, 'learning_rate/qwen_vl_interface': 5.308986706331377e-07, 'learning_rate/action_model': 5.308986706331376e-06, 'epoch': 0.35}) +[2026-10-03 19:54:41,831][starVLA.training.train_starvla][INFO] - Step 4823, Loss: {'train/loss': 0.0010483920341357589, 'train/grad_norm_pre_clip': 2.3133106231689453, 'timing/data': 0.00035099999513477087, 'timing/model': 0.39721822101273574, 'learning_rate/qwen_vl_interface': 5.305528413908797e-07, 'learning_rate/action_model': 5.305528413908798e-06, 'epoch': 0.35}) +[2026-10-03 19:54:42,540][starVLA.training.train_starvla][INFO] - Step 4824, Loss: {'train/loss': 0.00021243556693661958, 'train/grad_norm_pre_clip': 0.21786558628082275, 'timing/data': 0.00023087000590749085, 'timing/model': 0.39781774100265466, 'learning_rate/qwen_vl_interface': 5.302089521350799e-07, 'learning_rate/action_model': 5.3020895213507996e-06, 'epoch': 0.35}) +[2026-10-03 19:54:43,449][starVLA.training.train_starvla][INFO] - Step 4825, Loss: {'train/loss': 0.0026047867722809315, 'train/grad_norm_pre_clip': 5.340409755706787, 'timing/data': 0.00031629999284632504, 'timing/model': 0.5954579919925891, 'learning_rate/qwen_vl_interface': 5.298670030070977e-07, 'learning_rate/action_model': 5.298670030070976e-06, 'epoch': 0.35}) +[2026-10-03 19:54:44,155][starVLA.training.train_starvla][INFO] - Step 4826, Loss: {'train/loss': 0.00017549761105328798, 'train/grad_norm_pre_clip': 0.08452083170413971, 'timing/data': 0.0004902099899481982, 'timing/model': 0.3942317019973416, 'learning_rate/qwen_vl_interface': 5.295269941474957e-07, 'learning_rate/action_model': 5.295269941474957e-06, 'epoch': 0.35}) +[2026-10-03 19:54:44,866][starVLA.training.train_starvla][INFO] - Step 4827, Loss: {'train/loss': 0.5176212787628174, 'train/grad_norm_pre_clip': 64.29730987548828, 'timing/data': 0.00030482999864034355, 'timing/model': 0.39714894199278206, 'learning_rate/qwen_vl_interface': 5.291889256960384e-07, 'learning_rate/action_model': 5.291889256960384e-06, 'epoch': 0.35}) +[2026-10-03 19:54:45,574][starVLA.training.train_starvla][INFO] - Step 4828, Loss: {'train/loss': 3.5291035601403564e-05, 'train/grad_norm_pre_clip': 0.010048901662230492, 'timing/data': 0.0003649700083769858, 'timing/model': 0.39473693098989315, 'learning_rate/qwen_vl_interface': 5.288527977916929e-07, 'learning_rate/action_model': 5.288527977916929e-06, 'epoch': 0.35}) +[2026-10-03 19:54:46,281][starVLA.training.train_starvla][INFO] - Step 4829, Loss: {'train/loss': 0.0019593730103224516, 'train/grad_norm_pre_clip': 4.621424198150635, 'timing/data': 0.00035429999115876853, 'timing/model': 0.39458092098357156, 'learning_rate/qwen_vl_interface': 5.28518610572629e-07, 'learning_rate/action_model': 5.28518610572629e-06, 'epoch': 0.35}) +[2026-10-03 19:54:46,987][starVLA.training.train_starvla][INFO] - Step 4830, Loss: {'train/loss': 0.1310890018939972, 'train/grad_norm_pre_clip': 100.31614685058594, 'timing/data': 0.00026738998712971807, 'timing/model': 0.39416552102193236, 'learning_rate/qwen_vl_interface': 5.281863641762184e-07, 'learning_rate/action_model': 5.281863641762184e-06, 'epoch': 0.35}) +[2026-10-03 19:54:47,893][starVLA.training.train_starvla][INFO] - Step 4831, Loss: {'train/loss': 0.05881176516413689, 'train/grad_norm_pre_clip': 124.0206069946289, 'timing/data': 0.00037747001624666154, 'timing/model': 0.39692125099827535, 'learning_rate/qwen_vl_interface': 5.278560587390355e-07, 'learning_rate/action_model': 5.278560587390355e-06, 'epoch': 0.35}) +[2026-10-03 19:54:48,604][starVLA.training.train_starvla][INFO] - Step 4832, Loss: {'train/loss': 2.842908725142479e-05, 'train/grad_norm_pre_clip': 0.004794499836862087, 'timing/data': 0.00034234998747706413, 'timing/model': 0.3971385220065713, 'learning_rate/qwen_vl_interface': 5.275276943968557e-07, 'learning_rate/action_model': 5.275276943968557e-06, 'epoch': 0.35}) +[2026-10-03 19:54:49,313][starVLA.training.train_starvla][INFO] - Step 4833, Loss: {'train/loss': 0.00011995782551821321, 'train/grad_norm_pre_clip': 0.07609445601701736, 'timing/data': 0.00022672000341117382, 'timing/model': 0.3973704510135576, 'learning_rate/qwen_vl_interface': 5.272012712846566e-07, 'learning_rate/action_model': 5.2720127128465654e-06, 'epoch': 0.35}) +[2026-10-03 19:54:50,021][starVLA.training.train_starvla][INFO] - Step 4834, Loss: {'train/loss': 0.028844865038990974, 'train/grad_norm_pre_clip': 24.62842559814453, 'timing/data': 0.00022438002633862197, 'timing/model': 0.3956057709874585, 'learning_rate/qwen_vl_interface': 5.268767895366198e-07, 'learning_rate/action_model': 5.268767895366198e-06, 'epoch': 0.35}) +[2026-10-03 19:54:50,730][starVLA.training.train_starvla][INFO] - Step 4835, Loss: {'train/loss': 0.5097013711929321, 'train/grad_norm_pre_clip': 42.942359924316406, 'timing/data': 0.00023361999774351716, 'timing/model': 0.3972808010003064, 'learning_rate/qwen_vl_interface': 5.265542492861268e-07, 'learning_rate/action_model': 5.2655424928612676e-06, 'epoch': 0.35}) +[2026-10-03 19:54:51,640][starVLA.training.train_starvla][INFO] - Step 4836, Loss: {'train/loss': 9.880457218969241e-05, 'train/grad_norm_pre_clip': 0.055124253034591675, 'timing/data': 0.00029100998654030263, 'timing/model': 0.3966575710219331, 'learning_rate/qwen_vl_interface': 5.262336506657617e-07, 'learning_rate/action_model': 5.2623365066576165e-06, 'epoch': 0.35}) +[2026-10-03 19:54:52,349][starVLA.training.train_starvla][INFO] - Step 4837, Loss: {'train/loss': 0.0002217822620877996, 'train/grad_norm_pre_clip': 0.07808300852775574, 'timing/data': 0.00022648999583907425, 'timing/model': 0.39681125100469217, 'learning_rate/qwen_vl_interface': 5.259149938073103e-07, 'learning_rate/action_model': 5.259149938073102e-06, 'epoch': 0.35}) +[2026-10-03 19:54:53,055][starVLA.training.train_starvla][INFO] - Step 4838, Loss: {'train/loss': 0.000168416605447419, 'train/grad_norm_pre_clip': 0.10196571052074432, 'timing/data': 0.00023210901417769492, 'timing/model': 0.3939273419964593, 'learning_rate/qwen_vl_interface': 5.255982788417609e-07, 'learning_rate/action_model': 5.255982788417608e-06, 'epoch': 0.35}) +[2026-10-03 19:54:53,763][starVLA.training.train_starvla][INFO] - Step 4839, Loss: {'train/loss': 2.518919063732028e-05, 'train/grad_norm_pre_clip': 0.0064323665574193, 'timing/data': 0.00026560999685898423, 'timing/model': 0.39411616200231947, 'learning_rate/qwen_vl_interface': 5.252835058993029e-07, 'learning_rate/action_model': 5.252835058993029e-06, 'epoch': 0.35}) +[2026-10-03 19:54:54,471][starVLA.training.train_starvla][INFO] - Step 4840, Loss: {'train/loss': 0.09323427081108093, 'train/grad_norm_pre_clip': 124.91910552978516, 'timing/data': 0.0002170999941881746, 'timing/model': 0.39600636198883876, 'learning_rate/qwen_vl_interface': 5.249706751093274e-07, 'learning_rate/action_model': 5.249706751093273e-06, 'epoch': 0.35}) +[2026-10-03 19:54:55,377][starVLA.training.train_starvla][INFO] - Step 4841, Loss: {'train/loss': 4.6301327529363334e-05, 'train/grad_norm_pre_clip': 0.020470313727855682, 'timing/data': 0.00021507000201381743, 'timing/model': 0.5954446109826677, 'learning_rate/qwen_vl_interface': 5.246597866004273e-07, 'learning_rate/action_model': 5.246597866004273e-06, 'epoch': 0.35}) +[2026-10-03 19:54:56,083][starVLA.training.train_starvla][INFO] - Step 4842, Loss: {'train/loss': 0.1997823864221573, 'train/grad_norm_pre_clip': 105.67218780517578, 'timing/data': 0.0002191399980802089, 'timing/model': 0.39430710099986754, 'learning_rate/qwen_vl_interface': 5.243508405003975e-07, 'learning_rate/action_model': 5.243508405003975e-06, 'epoch': 0.35}) +[2026-10-03 19:54:56,791][starVLA.training.train_starvla][INFO] - Step 4843, Loss: {'train/loss': 0.0009032384841702878, 'train/grad_norm_pre_clip': 1.9323378801345825, 'timing/data': 0.00032717001158744097, 'timing/model': 0.3955499310104642, 'learning_rate/qwen_vl_interface': 5.240438369362336e-07, 'learning_rate/action_model': 5.2404383693623366e-06, 'epoch': 0.35}) +[2026-10-03 19:54:57,499][starVLA.training.train_starvla][INFO] - Step 4844, Loss: {'train/loss': 5.2203678933437914e-05, 'train/grad_norm_pre_clip': 0.010648814029991627, 'timing/data': 0.0003136899904347956, 'timing/model': 0.39620358098181896, 'learning_rate/qwen_vl_interface': 5.237387760341335e-07, 'learning_rate/action_model': 5.237387760341335e-06, 'epoch': 0.35}) +[2026-10-03 19:54:58,206][starVLA.training.train_starvla][INFO] - Step 4845, Loss: {'train/loss': 3.6423200072022155e-05, 'train/grad_norm_pre_clip': 0.01191254984587431, 'timing/data': 0.00026892998721450567, 'timing/model': 0.3938705410109833, 'learning_rate/qwen_vl_interface': 5.234356579194963e-07, 'learning_rate/action_model': 5.234356579194963e-06, 'epoch': 0.35}) +[2026-10-03 19:54:59,106][starVLA.training.train_starvla][INFO] - Step 4846, Loss: {'train/loss': 0.000680626486428082, 'train/grad_norm_pre_clip': 1.1069831848144531, 'timing/data': 0.00035459999344311655, 'timing/model': 0.589135781978257, 'learning_rate/qwen_vl_interface': 5.231344827169218e-07, 'learning_rate/action_model': 5.231344827169217e-06, 'epoch': 0.35}) +[2026-10-03 19:54:59,815][starVLA.training.train_starvla][INFO] - Step 4847, Loss: {'train/loss': 0.0009060127194970846, 'train/grad_norm_pre_clip': 1.1436222791671753, 'timing/data': 0.0003273900074418634, 'timing/model': 0.39612589101307094, 'learning_rate/qwen_vl_interface': 5.228352505502119e-07, 'learning_rate/action_model': 5.228352505502119e-06, 'epoch': 0.35}) +[2026-10-03 19:55:00,522][starVLA.training.train_starvla][INFO] - Step 4848, Loss: {'train/loss': 0.00016180097009055316, 'train/grad_norm_pre_clip': 0.12580004334449768, 'timing/data': 0.00030976999551057816, 'timing/model': 0.3937001319718547, 'learning_rate/qwen_vl_interface': 5.225379615423701e-07, 'learning_rate/action_model': 5.2253796154237e-06, 'epoch': 0.35}) +[2026-10-03 19:55:01,229][starVLA.training.train_starvla][INFO] - Step 4849, Loss: {'train/loss': 0.00010394060518592596, 'train/grad_norm_pre_clip': 0.030841641128063202, 'timing/data': 0.00035204997402615845, 'timing/model': 0.3937181619985495, 'learning_rate/qwen_vl_interface': 5.222426158156008e-07, 'learning_rate/action_model': 5.222426158156008e-06, 'epoch': 0.35}) +[2026-10-03 19:55:01,935][starVLA.training.train_starvla][INFO] - Step 4850, Loss: {'train/loss': 0.0016126943519338965, 'train/grad_norm_pre_clip': 1.7734310626983643, 'timing/data': 0.0003418400010559708, 'timing/model': 0.3947906020039227, 'learning_rate/qwen_vl_interface': 5.219492134913084e-07, 'learning_rate/action_model': 5.219492134913084e-06, 'epoch': 0.35}) +[2026-10-03 19:55:02,644][starVLA.training.train_starvla][INFO] - Step 4851, Loss: {'train/loss': 0.3840606212615967, 'train/grad_norm_pre_clip': 33.291053771972656, 'timing/data': 0.0004183699784334749, 'timing/model': 0.3962074820010457, 'learning_rate/qwen_vl_interface': 5.216577546901008e-07, 'learning_rate/action_model': 5.216577546901007e-06, 'epoch': 0.36}) +[2026-10-03 19:55:03,549][starVLA.training.train_starvla][INFO] - Step 4852, Loss: {'train/loss': 1.7068543456844054e-05, 'train/grad_norm_pre_clip': 0.0025812999811023474, 'timing/data': 0.0002746600075624883, 'timing/model': 0.39393280100193806, 'learning_rate/qwen_vl_interface': 5.213682395317852e-07, 'learning_rate/action_model': 5.213682395317851e-06, 'epoch': 0.36}) +[2026-10-03 19:55:04,258][starVLA.training.train_starvla][INFO] - Step 4853, Loss: {'train/loss': 6.549942190758884e-05, 'train/grad_norm_pre_clip': 0.026593534275889397, 'timing/data': 0.00023526998120360076, 'timing/model': 0.3961459610145539, 'learning_rate/qwen_vl_interface': 5.2108066813537e-07, 'learning_rate/action_model': 5.2108066813537e-06, 'epoch': 0.36}) +[2026-10-03 19:55:04,965][starVLA.training.train_starvla][INFO] - Step 4854, Loss: {'train/loss': 0.0038105230778455734, 'train/grad_norm_pre_clip': 3.887269973754883, 'timing/data': 0.0002707800013013184, 'timing/model': 0.39512722098152153, 'learning_rate/qwen_vl_interface': 5.207950406190654e-07, 'learning_rate/action_model': 5.207950406190653e-06, 'epoch': 0.36}) +[2026-10-03 19:55:05,674][starVLA.training.train_starvla][INFO] - Step 4855, Loss: {'train/loss': 3.969837780459784e-05, 'train/grad_norm_pre_clip': 0.01970195770263672, 'timing/data': 0.00028572999872267246, 'timing/model': 0.3959240209951531, 'learning_rate/qwen_vl_interface': 5.205113571002818e-07, 'learning_rate/action_model': 5.205113571002818e-06, 'epoch': 0.36}) +[2026-10-03 19:55:06,380][starVLA.training.train_starvla][INFO] - Step 4856, Loss: {'train/loss': 0.020413517951965332, 'train/grad_norm_pre_clip': 57.730899810791016, 'timing/data': 0.00023209999199025333, 'timing/model': 0.39447062197723426, 'learning_rate/qwen_vl_interface': 5.202296176956308e-07, 'learning_rate/action_model': 5.202296176956308e-06, 'epoch': 0.36}) +[2026-10-03 19:55:07,286][starVLA.training.train_starvla][INFO] - Step 4857, Loss: {'train/loss': 9.646471153246239e-05, 'train/grad_norm_pre_clip': 0.08114581555128098, 'timing/data': 0.00023331999545916915, 'timing/model': 0.39451796098728664, 'learning_rate/qwen_vl_interface': 5.19949822520925e-07, 'learning_rate/action_model': 5.19949822520925e-06, 'epoch': 0.36}) +[2026-10-03 19:55:07,992][starVLA.training.train_starvla][INFO] - Step 4858, Loss: {'train/loss': 0.1835084855556488, 'train/grad_norm_pre_clip': 103.02230834960938, 'timing/data': 0.0002703100035432726, 'timing/model': 0.39417099198908545, 'learning_rate/qwen_vl_interface': 5.196719716911777e-07, 'learning_rate/action_model': 5.1967197169117764e-06, 'epoch': 0.36}) +[2026-10-03 19:55:08,698][starVLA.training.train_starvla][INFO] - Step 4859, Loss: {'train/loss': 0.0545678474009037, 'train/grad_norm_pre_clip': 63.35786056518555, 'timing/data': 0.00040755001828074455, 'timing/model': 0.3948325819801539, 'learning_rate/qwen_vl_interface': 5.193960653206029e-07, 'learning_rate/action_model': 5.193960653206029e-06, 'epoch': 0.36}) +[2026-10-03 19:55:09,406][starVLA.training.train_starvla][INFO] - Step 4860, Loss: {'train/loss': 3.275094786658883e-05, 'train/grad_norm_pre_clip': 0.010999266989529133, 'timing/data': 0.0002673199924174696, 'timing/model': 0.3960129510087427, 'learning_rate/qwen_vl_interface': 5.191221035226144e-07, 'learning_rate/action_model': 5.191221035226144e-06, 'epoch': 0.36}) +[2026-10-03 19:55:10,113][starVLA.training.train_starvla][INFO] - Step 4861, Loss: {'train/loss': 0.06109601631760597, 'train/grad_norm_pre_clip': 123.98262786865234, 'timing/data': 0.00023594999220222235, 'timing/model': 0.3959987109992653, 'learning_rate/qwen_vl_interface': 5.188500864098292e-07, 'learning_rate/action_model': 5.1885008640982915e-06, 'epoch': 0.36}) +[2026-10-03 19:55:11,019][starVLA.training.train_starvla][INFO] - Step 4862, Loss: {'train/loss': 0.22328977286815643, 'train/grad_norm_pre_clip': 131.26734924316406, 'timing/data': 0.0002380700025241822, 'timing/model': 0.5937698420020752, 'learning_rate/qwen_vl_interface': 5.185800140940613e-07, 'learning_rate/action_model': 5.185800140940613e-06, 'epoch': 0.36}) +[2026-10-03 19:55:11,727][starVLA.training.train_starvla][INFO] - Step 4863, Loss: {'train/loss': 2.9837909096386284e-05, 'train/grad_norm_pre_clip': 0.007692517712712288, 'timing/data': 0.00030020001577213407, 'timing/model': 0.39605102202040143, 'learning_rate/qwen_vl_interface': 5.183118866863288e-07, 'learning_rate/action_model': 5.1831188668632875e-06, 'epoch': 0.36}) +[2026-10-03 19:55:12,435][starVLA.training.train_starvla][INFO] - Step 4864, Loss: {'train/loss': 0.00011157547123730183, 'train/grad_norm_pre_clip': 0.03071022965013981, 'timing/data': 0.00029227000777609646, 'timing/model': 0.3955893309903331, 'learning_rate/qwen_vl_interface': 5.180457042968484e-07, 'learning_rate/action_model': 5.1804570429684835e-06, 'epoch': 0.36}) +[2026-10-03 19:55:13,142][starVLA.training.train_starvla][INFO] - Step 4865, Loss: {'train/loss': 0.002852685982361436, 'train/grad_norm_pre_clip': 7.581406116485596, 'timing/data': 0.00043916000868193805, 'timing/model': 0.3938925009861123, 'learning_rate/qwen_vl_interface': 5.17781467035037e-07, 'learning_rate/action_model': 5.177814670350371e-06, 'epoch': 0.36}) +[2026-10-03 19:55:13,848][starVLA.training.train_starvla][INFO] - Step 4866, Loss: {'train/loss': 0.00046751368790864944, 'train/grad_norm_pre_clip': 0.641405463218689, 'timing/data': 0.0003300399985164404, 'timing/model': 0.39370521099772304, 'learning_rate/qwen_vl_interface': 5.175191750095133e-07, 'learning_rate/action_model': 5.175191750095132e-06, 'epoch': 0.36}) +[2026-10-03 19:55:14,762][starVLA.training.train_starvla][INFO] - Step 4867, Loss: {'train/loss': 5.475091893458739e-05, 'train/grad_norm_pre_clip': 0.06572473794221878, 'timing/data': 0.00033059000270441175, 'timing/model': 0.5999887009966187, 'learning_rate/qwen_vl_interface': 5.172588283280953e-07, 'learning_rate/action_model': 5.172588283280953e-06, 'epoch': 0.36}) +[2026-10-03 19:55:15,472][starVLA.training.train_starvla][INFO] - Step 4868, Loss: {'train/loss': 0.0001995251077460125, 'train/grad_norm_pre_clip': 0.09219034761190414, 'timing/data': 0.00023061002139002085, 'timing/model': 0.3974355809914414, 'learning_rate/qwen_vl_interface': 5.170004270978019e-07, 'learning_rate/action_model': 5.170004270978018e-06, 'epoch': 0.36}) +[2026-10-03 19:55:16,181][starVLA.training.train_starvla][INFO] - Step 4869, Loss: {'train/loss': 0.002464170102030039, 'train/grad_norm_pre_clip': 4.314425468444824, 'timing/data': 0.000434669986134395, 'timing/model': 0.3962688519968651, 'learning_rate/qwen_vl_interface': 5.167439714248512e-07, 'learning_rate/action_model': 5.1674397142485116e-06, 'epoch': 0.36}) +[2026-10-03 19:55:16,892][starVLA.training.train_starvla][INFO] - Step 4870, Loss: {'train/loss': 0.000377665797714144, 'train/grad_norm_pre_clip': 0.29053741693496704, 'timing/data': 0.0003347300225868821, 'timing/model': 0.39699091101647355, 'learning_rate/qwen_vl_interface': 5.164894614146633e-07, 'learning_rate/action_model': 5.164894614146634e-06, 'epoch': 0.36}) +[2026-10-03 19:55:17,603][starVLA.training.train_starvla][INFO] - Step 4871, Loss: {'train/loss': 0.0002669260429684073, 'train/grad_norm_pre_clip': 0.12732648849487305, 'timing/data': 0.00032137997914105654, 'timing/model': 0.39709009099169634, 'learning_rate/qwen_vl_interface': 5.16236897171858e-07, 'learning_rate/action_model': 5.16236897171858e-06, 'epoch': 0.36}) +[2026-10-03 19:55:18,312][starVLA.training.train_starvla][INFO] - Step 4872, Loss: {'train/loss': 0.00012237335613463074, 'train/grad_norm_pre_clip': 0.051441919058561325, 'timing/data': 0.0002993099915329367, 'timing/model': 0.39562887101783417, 'learning_rate/qwen_vl_interface': 5.15986278800254e-07, 'learning_rate/action_model': 5.15986278800254e-06, 'epoch': 0.36}) +[2026-10-03 19:55:19,216][starVLA.training.train_starvla][INFO] - Step 4873, Loss: {'train/loss': 6.376306555466726e-05, 'train/grad_norm_pre_clip': 0.020568156614899635, 'timing/data': 0.0003316399815957993, 'timing/model': 0.3960967409948353, 'learning_rate/qwen_vl_interface': 5.157376064028719e-07, 'learning_rate/action_model': 5.157376064028718e-06, 'epoch': 0.36}) +[2026-10-03 19:55:19,924][starVLA.training.train_starvla][INFO] - Step 4874, Loss: {'train/loss': 0.00021707237465307117, 'train/grad_norm_pre_clip': 0.13960376381874084, 'timing/data': 0.0002574599930085242, 'timing/model': 0.39535020099719986, 'learning_rate/qwen_vl_interface': 5.154908800819307e-07, 'learning_rate/action_model': 5.154908800819306e-06, 'epoch': 0.36}) +[2026-10-03 19:55:20,634][starVLA.training.train_starvla][INFO] - Step 4875, Loss: {'train/loss': 8.969771442934871e-05, 'train/grad_norm_pre_clip': 0.09641537815332413, 'timing/data': 0.0002209699887316674, 'timing/model': 0.3958533619879745, 'learning_rate/qwen_vl_interface': 5.152460999388504e-07, 'learning_rate/action_model': 5.152460999388503e-06, 'epoch': 0.36}) +[2026-10-03 19:55:21,343][starVLA.training.train_starvla][INFO] - Step 4876, Loss: {'train/loss': 0.4631775915622711, 'train/grad_norm_pre_clip': 115.71794128417969, 'timing/data': 0.00026544000138528645, 'timing/model': 0.3963763520005159, 'learning_rate/qwen_vl_interface': 5.150032660742517e-07, 'learning_rate/action_model': 5.150032660742517e-06, 'epoch': 0.36}) +[2026-10-03 19:55:22,052][starVLA.training.train_starvla][INFO] - Step 4877, Loss: {'train/loss': 0.027201475575566292, 'train/grad_norm_pre_clip': 47.33090591430664, 'timing/data': 0.0003711300087161362, 'timing/model': 0.39586431099451147, 'learning_rate/qwen_vl_interface': 5.147623785879535e-07, 'learning_rate/action_model': 5.147623785879534e-06, 'epoch': 0.36}) +[2026-10-03 19:55:22,959][starVLA.training.train_starvla][INFO] - Step 4878, Loss: {'train/loss': 0.04248614236712456, 'train/grad_norm_pre_clip': 65.86898803710938, 'timing/data': 0.00036246998934075236, 'timing/model': 0.3954917319933884, 'learning_rate/qwen_vl_interface': 5.145234375789758e-07, 'learning_rate/action_model': 5.145234375789758e-06, 'epoch': 0.36}) +[2026-10-03 19:55:23,668][starVLA.training.train_starvla][INFO] - Step 4879, Loss: {'train/loss': 0.00017918666708283126, 'train/grad_norm_pre_clip': 0.11792287230491638, 'timing/data': 0.0003068300138693303, 'timing/model': 0.396083161002025, 'learning_rate/qwen_vl_interface': 5.142864431455387e-07, 'learning_rate/action_model': 5.1428644314553875e-06, 'epoch': 0.36}) +[2026-10-03 19:55:24,377][starVLA.training.train_starvla][INFO] - Step 4880, Loss: {'train/loss': 0.0005854432820342481, 'train/grad_norm_pre_clip': 0.3886829912662506, 'timing/data': 0.0003238100034650415, 'timing/model': 0.3958923910104204, 'learning_rate/qwen_vl_interface': 5.140513953850612e-07, 'learning_rate/action_model': 5.140513953850612e-06, 'epoch': 0.36}) +[2026-10-03 19:55:25,085][starVLA.training.train_starvla][INFO] - Step 4881, Loss: {'train/loss': 0.43333396315574646, 'train/grad_norm_pre_clip': 59.88094711303711, 'timing/data': 0.0003934399865102023, 'timing/model': 0.3954998510016594, 'learning_rate/qwen_vl_interface': 5.138182943941623e-07, 'learning_rate/action_model': 5.138182943941623e-06, 'epoch': 0.36}) +[2026-10-03 19:55:25,794][starVLA.training.train_starvla][INFO] - Step 4882, Loss: {'train/loss': 0.30127865076065063, 'train/grad_norm_pre_clip': 206.82321166992188, 'timing/data': 0.0003190800198353827, 'timing/model': 0.39579381098155864, 'learning_rate/qwen_vl_interface': 5.135871402686615e-07, 'learning_rate/action_model': 5.135871402686615e-06, 'epoch': 0.36}) +[2026-10-03 19:55:26,708][starVLA.training.train_starvla][INFO] - Step 4883, Loss: {'train/loss': 0.05885808542370796, 'train/grad_norm_pre_clip': 114.25469207763672, 'timing/data': 0.00023902999237179756, 'timing/model': 0.5988174620142672, 'learning_rate/qwen_vl_interface': 5.133579331035777e-07, 'learning_rate/action_model': 5.133579331035777e-06, 'epoch': 0.36}) +[2026-10-03 19:55:27,416][starVLA.training.train_starvla][INFO] - Step 4884, Loss: {'train/loss': 0.033427223563194275, 'train/grad_norm_pre_clip': 54.860042572021484, 'timing/data': 0.00035767999361269176, 'timing/model': 0.39508467199630104, 'learning_rate/qwen_vl_interface': 5.131306729931284e-07, 'learning_rate/action_model': 5.1313067299312836e-06, 'epoch': 0.36}) +[2026-10-03 19:55:28,123][starVLA.training.train_starvla][INFO] - Step 4885, Loss: {'train/loss': 0.008399059064686298, 'train/grad_norm_pre_clip': 31.129844665527344, 'timing/data': 0.00037691998295485973, 'timing/model': 0.39563473098678514, 'learning_rate/qwen_vl_interface': 5.129053600307326e-07, 'learning_rate/action_model': 5.129053600307326e-06, 'epoch': 0.36}) +[2026-10-03 19:55:28,830][starVLA.training.train_starvla][INFO] - Step 4886, Loss: {'train/loss': 3.5067838325630873e-05, 'train/grad_norm_pre_clip': 0.011101314797997475, 'timing/data': 0.00022262000129558146, 'timing/model': 0.39564879101817496, 'learning_rate/qwen_vl_interface': 5.126819943090078e-07, 'learning_rate/action_model': 5.1268199430900775e-06, 'epoch': 0.36}) +[2026-10-03 19:55:29,538][starVLA.training.train_starvla][INFO] - Step 4887, Loss: {'train/loss': 0.0018617452587932348, 'train/grad_norm_pre_clip': 1.6079412698745728, 'timing/data': 0.00034339999547228217, 'timing/model': 0.39527456197538413, 'learning_rate/qwen_vl_interface': 5.124605759197706e-07, 'learning_rate/action_model': 5.124605759197706e-06, 'epoch': 0.36}) +[2026-10-03 19:55:30,457][starVLA.training.train_starvla][INFO] - Step 4888, Loss: {'train/loss': 9.686416888143867e-05, 'train/grad_norm_pre_clip': 0.04374171793460846, 'timing/data': 0.00021853001089766622, 'timing/model': 0.6076909709954634, 'learning_rate/qwen_vl_interface': 5.122411049540383e-07, 'learning_rate/action_model': 5.1224110495403825e-06, 'epoch': 0.36}) +[2026-10-03 19:55:31,138][starVLA.training.train_starvla][INFO] - Step 4889, Loss: {'train/loss': 0.2579537332057953, 'train/grad_norm_pre_clip': 128.7186737060547, 'timing/data': 0.00037580999196507037, 'timing/model': 0.36898897401988506, 'learning_rate/qwen_vl_interface': 5.120235815020269e-07, 'learning_rate/action_model': 5.120235815020269e-06, 'epoch': 0.36}) +[2026-10-03 19:55:31,843][starVLA.training.train_starvla][INFO] - Step 4890, Loss: {'train/loss': 0.04418540000915527, 'train/grad_norm_pre_clip': 84.63103485107422, 'timing/data': 0.0004077800258528441, 'timing/model': 0.39347812201594934, 'learning_rate/qwen_vl_interface': 5.118080056531524e-07, 'learning_rate/action_model': 5.118080056531524e-06, 'epoch': 0.36}) +[2026-10-03 19:55:32,549][starVLA.training.train_starvla][INFO] - Step 4891, Loss: {'train/loss': 0.03857104852795601, 'train/grad_norm_pre_clip': 74.30513763427734, 'timing/data': 0.0002702799974940717, 'timing/model': 0.39337921200785786, 'learning_rate/qwen_vl_interface': 5.115943774960295e-07, 'learning_rate/action_model': 5.115943774960295e-06, 'epoch': 0.36}) +[2026-10-03 19:55:33,261][starVLA.training.train_starvla][INFO] - Step 4892, Loss: {'train/loss': 0.126011922955513, 'train/grad_norm_pre_clip': 221.88381958007812, 'timing/data': 0.00023153997608460486, 'timing/model': 0.39264357098727487, 'learning_rate/qwen_vl_interface': 5.113826971184727e-07, 'learning_rate/action_model': 5.113826971184727e-06, 'epoch': 0.36}) +[2026-10-03 19:55:33,960][starVLA.training.train_starvla][INFO] - Step 4893, Loss: {'train/loss': 4.578291918733157e-05, 'train/grad_norm_pre_clip': 0.013582375831902027, 'timing/data': 0.00024440998095087707, 'timing/model': 0.39256591198500246, 'learning_rate/qwen_vl_interface': 5.111729646074954e-07, 'learning_rate/action_model': 5.111729646074954e-06, 'epoch': 0.36}) +[2026-10-03 19:55:34,863][starVLA.training.train_starvla][INFO] - Step 4894, Loss: {'train/loss': 0.00037166368565522134, 'train/grad_norm_pre_clip': 0.16276167333126068, 'timing/data': 0.00021499002468772233, 'timing/model': 0.5942074820050038, 'learning_rate/qwen_vl_interface': 5.10965180049312e-07, 'learning_rate/action_model': 5.10965180049312e-06, 'epoch': 0.36}) +[2026-10-03 19:55:35,563][starVLA.training.train_starvla][INFO] - Step 4895, Loss: {'train/loss': 0.2983596622943878, 'train/grad_norm_pre_clip': 54.199527740478516, 'timing/data': 0.0002467099984642118, 'timing/model': 0.3923399720224552, 'learning_rate/qwen_vl_interface': 5.107593435293336e-07, 'learning_rate/action_model': 5.107593435293336e-06, 'epoch': 0.36}) +[2026-10-03 19:55:36,275][starVLA.training.train_starvla][INFO] - Step 4896, Loss: {'train/loss': 0.3476971983909607, 'train/grad_norm_pre_clip': 127.24703216552734, 'timing/data': 0.00030986001365818083, 'timing/model': 0.3922105919918977, 'learning_rate/qwen_vl_interface': 5.105554551321724e-07, 'learning_rate/action_model': 5.105554551321723e-06, 'epoch': 0.36}) +[2026-10-03 19:55:36,971][starVLA.training.train_starvla][INFO] - Step 4897, Loss: {'train/loss': 0.0012145454529672861, 'train/grad_norm_pre_clip': 1.6901010274887085, 'timing/data': 0.0003639899950940162, 'timing/model': 0.3902544819866307, 'learning_rate/qwen_vl_interface': 5.103535149416389e-07, 'learning_rate/action_model': 5.1035351494163885e-06, 'epoch': 0.36}) +[2026-10-03 19:55:37,667][starVLA.training.train_starvla][INFO] - Step 4898, Loss: {'train/loss': 0.021697208285331726, 'train/grad_norm_pre_clip': 52.02423858642578, 'timing/data': 0.0004550600133370608, 'timing/model': 0.38951852201716974, 'learning_rate/qwen_vl_interface': 5.101535230407436e-07, 'learning_rate/action_model': 5.101535230407436e-06, 'epoch': 0.36}) +[2026-10-03 19:55:38,570][starVLA.training.train_starvla][INFO] - Step 4899, Loss: {'train/loss': 0.26544076204299927, 'train/grad_norm_pre_clip': 153.7588348388672, 'timing/data': 0.00020926998695358634, 'timing/model': 0.5945508420118131, 'learning_rate/qwen_vl_interface': 5.099554795116956e-07, 'learning_rate/action_model': 5.0995547951169565e-06, 'epoch': 0.36}) +[2026-10-03 19:55:39,266][starVLA.training.train_starvla][INFO] - Step 4900, Loss: {'train/loss': 0.2563591003417969, 'train/grad_norm_pre_clip': 206.5619354248047, 'timing/data': 0.0002922200073953718, 'timing/model': 0.3896272720012348, 'learning_rate/qwen_vl_interface': 5.097593844359024e-07, 'learning_rate/action_model': 5.0975938443590245e-06, 'epoch': 0.36}) +[2026-10-03 19:55:39,967][starVLA.training.train_starvla][INFO] - Step 4901, Loss: {'train/loss': 0.028440406545996666, 'train/grad_norm_pre_clip': 45.83694839477539, 'timing/data': 0.0003490899980533868, 'timing/model': 0.39134994099731557, 'learning_rate/qwen_vl_interface': 5.095652378939718e-07, 'learning_rate/action_model': 5.095652378939717e-06, 'epoch': 0.36}) +[2026-10-03 19:55:40,666][starVLA.training.train_starvla][INFO] - Step 4902, Loss: {'train/loss': 0.0003832425572909415, 'train/grad_norm_pre_clip': 0.2738164961338043, 'timing/data': 0.00033633000566624105, 'timing/model': 0.39125494202016853, 'learning_rate/qwen_vl_interface': 5.093730399657101e-07, 'learning_rate/action_model': 5.0937303996571e-06, 'epoch': 0.36}) +[2026-10-03 19:55:41,366][starVLA.training.train_starvla][INFO] - Step 4903, Loss: {'train/loss': 0.0001724222383927554, 'train/grad_norm_pre_clip': 0.14766597747802734, 'timing/data': 0.0003066500066779554, 'timing/model': 0.3917491720058024, 'learning_rate/qwen_vl_interface': 5.091827907301227e-07, 'learning_rate/action_model': 5.091827907301227e-06, 'epoch': 0.36}) +[2026-10-03 19:55:42,272][starVLA.training.train_starvla][INFO] - Step 4904, Loss: {'train/loss': 0.17996735870838165, 'train/grad_norm_pre_clip': 125.7995834350586, 'timing/data': 0.00038646001485176384, 'timing/model': 0.59734294199734, 'learning_rate/qwen_vl_interface': 5.089944902654134e-07, 'learning_rate/action_model': 5.089944902654134e-06, 'epoch': 0.36}) +[2026-10-03 19:55:42,973][starVLA.training.train_starvla][INFO] - Step 4905, Loss: {'train/loss': 0.0010173136834055185, 'train/grad_norm_pre_clip': 2.582773447036743, 'timing/data': 0.00035482001840136945, 'timing/model': 0.3924413920030929, 'learning_rate/qwen_vl_interface': 5.088081386489856e-07, 'learning_rate/action_model': 5.088081386489857e-06, 'epoch': 0.36}) +[2026-10-03 19:55:43,672][starVLA.training.train_starvla][INFO] - Step 4906, Loss: {'train/loss': 0.3490433096885681, 'train/grad_norm_pre_clip': 63.70478820800781, 'timing/data': 0.0003325500001665205, 'timing/model': 0.392465930985054, 'learning_rate/qwen_vl_interface': 5.086237359574419e-07, 'learning_rate/action_model': 5.086237359574419e-06, 'epoch': 0.36}) +[2026-10-03 19:55:44,370][starVLA.training.train_starvla][INFO] - Step 4907, Loss: {'train/loss': 5.3490130085265264e-05, 'train/grad_norm_pre_clip': 0.018250884488224983, 'timing/data': 0.0003653799940366298, 'timing/model': 0.39123754200409167, 'learning_rate/qwen_vl_interface': 5.084412822665825e-07, 'learning_rate/action_model': 5.084412822665825e-06, 'epoch': 0.36}) +[2026-10-03 19:55:45,070][starVLA.training.train_starvla][INFO] - Step 4908, Loss: {'train/loss': 0.4781140685081482, 'train/grad_norm_pre_clip': 178.53970336914062, 'timing/data': 0.0003340599942021072, 'timing/model': 0.3915550210222136, 'learning_rate/qwen_vl_interface': 5.082607776514081e-07, 'learning_rate/action_model': 5.082607776514081e-06, 'epoch': 0.36}) +[2026-10-03 19:55:45,768][starVLA.training.train_starvla][INFO] - Step 4909, Loss: {'train/loss': 0.0009021925507113338, 'train/grad_norm_pre_clip': 0.6254653930664062, 'timing/data': 0.00034783000592142344, 'timing/model': 0.3902040309912991, 'learning_rate/qwen_vl_interface': 5.080822221861167e-07, 'learning_rate/action_model': 5.080822221861166e-06, 'epoch': 0.36}) +[2026-10-03 19:55:46,673][starVLA.training.train_starvla][INFO] - Step 4910, Loss: {'train/loss': 9.1064881416969e-05, 'train/grad_norm_pre_clip': 0.045070864260196686, 'timing/data': 0.00023831002181395888, 'timing/model': 0.391215921001276, 'learning_rate/qwen_vl_interface': 5.07905615944106e-07, 'learning_rate/action_model': 5.0790561594410595e-06, 'epoch': 0.36}) +[2026-10-03 19:55:47,369][starVLA.training.train_starvla][INFO] - Step 4911, Loss: {'train/loss': 0.10691435635089874, 'train/grad_norm_pre_clip': 52.51963424682617, 'timing/data': 0.00020544999279081821, 'timing/model': 0.38965818198630586, 'learning_rate/qwen_vl_interface': 5.077309589979716e-07, 'learning_rate/action_model': 5.077309589979715e-06, 'epoch': 0.36}) +[2026-10-03 19:55:48,069][starVLA.training.train_starvla][INFO] - Step 4912, Loss: {'train/loss': 2.429473533993587e-05, 'train/grad_norm_pre_clip': 0.011163820512592793, 'timing/data': 0.00034207000862807035, 'timing/model': 0.3922421619936358, 'learning_rate/qwen_vl_interface': 5.075582514195089e-07, 'learning_rate/action_model': 5.0755825141950895e-06, 'epoch': 0.36}) +[2026-10-03 19:55:48,765][starVLA.training.train_starvla][INFO] - Step 4913, Loss: {'train/loss': 3.165638190694153e-05, 'train/grad_norm_pre_clip': 0.007943812757730484, 'timing/data': 0.00035165000008419156, 'timing/model': 0.3901543320098426, 'learning_rate/qwen_vl_interface': 5.073874932797115e-07, 'learning_rate/action_model': 5.0738749327971155e-06, 'epoch': 0.36}) +[2026-10-03 19:55:49,466][starVLA.training.train_starvla][INFO] - Step 4914, Loss: {'train/loss': 0.17534427344799042, 'train/grad_norm_pre_clip': 280.1578674316406, 'timing/data': 0.000330979994032532, 'timing/model': 0.3931468109949492, 'learning_rate/qwen_vl_interface': 5.072186846487713e-07, 'learning_rate/action_model': 5.072186846487713e-06, 'epoch': 0.36}) +[2026-10-03 19:55:50,370][starVLA.training.train_starvla][INFO] - Step 4915, Loss: {'train/loss': 5.200316809350625e-05, 'train/grad_norm_pre_clip': 0.020133182406425476, 'timing/data': 0.00028922001365572214, 'timing/model': 0.5954010620189365, 'learning_rate/qwen_vl_interface': 5.070518255960796e-07, 'learning_rate/action_model': 5.070518255960795e-06, 'epoch': 0.36}) +[2026-10-03 19:55:51,079][starVLA.training.train_starvla][INFO] - Step 4916, Loss: {'train/loss': 0.3396960198879242, 'train/grad_norm_pre_clip': 52.68098831176758, 'timing/data': 0.0002472100022714585, 'timing/model': 0.39232932101003826, 'learning_rate/qwen_vl_interface': 5.068869161902256e-07, 'learning_rate/action_model': 5.068869161902256e-06, 'epoch': 0.36}) +[2026-10-03 19:55:51,778][starVLA.training.train_starvla][INFO] - Step 4917, Loss: {'train/loss': 0.0009333681664429605, 'train/grad_norm_pre_clip': 0.5752252340316772, 'timing/data': 0.0003878599964082241, 'timing/model': 0.39225261099636555, 'learning_rate/qwen_vl_interface': 5.067239564989968e-07, 'learning_rate/action_model': 5.067239564989968e-06, 'epoch': 0.36}) +[2026-10-03 19:55:52,478][starVLA.training.train_starvla][INFO] - Step 4918, Loss: {'train/loss': 1.7746864614309743e-05, 'train/grad_norm_pre_clip': 0.002800219925120473, 'timing/data': 0.00033629999961704016, 'timing/model': 0.390425462013809, 'learning_rate/qwen_vl_interface': 5.065629465893808e-07, 'learning_rate/action_model': 5.065629465893807e-06, 'epoch': 0.36}) +[2026-10-03 19:55:53,177][starVLA.training.train_starvla][INFO] - Step 4919, Loss: {'train/loss': 1.846217128331773e-05, 'train/grad_norm_pre_clip': 0.0040755621157586575, 'timing/data': 0.000292009994154796, 'timing/model': 0.39013289200374857, 'learning_rate/qwen_vl_interface': 5.064038865275625e-07, 'learning_rate/action_model': 5.0640388652756245e-06, 'epoch': 0.36}) +[2026-10-03 19:55:54,091][starVLA.training.train_starvla][INFO] - Step 4920, Loss: {'train/loss': 0.0011478479718789458, 'train/grad_norm_pre_clip': 3.450899600982666, 'timing/data': 0.00035370999830774963, 'timing/model': 0.5949047309986781, 'learning_rate/qwen_vl_interface': 5.062467763789242e-07, 'learning_rate/action_model': 5.062467763789242e-06, 'epoch': 0.36}) +[2026-10-03 19:55:54,789][starVLA.training.train_starvla][INFO] - Step 4921, Loss: {'train/loss': 0.3586276173591614, 'train/grad_norm_pre_clip': 99.45684051513672, 'timing/data': 0.0003587300016079098, 'timing/model': 0.39015651200315915, 'learning_rate/qwen_vl_interface': 5.060916162080496e-07, 'learning_rate/action_model': 5.060916162080496e-06, 'epoch': 0.36}) +[2026-10-03 19:55:55,489][starVLA.training.train_starvla][INFO] - Step 4922, Loss: {'train/loss': 0.006108643487095833, 'train/grad_norm_pre_clip': 14.341178894042969, 'timing/data': 0.00031987999682314694, 'timing/model': 0.3900911920063663, 'learning_rate/qwen_vl_interface': 5.059384060787182e-07, 'learning_rate/action_model': 5.059384060787183e-06, 'epoch': 0.36}) +[2026-10-03 19:55:56,185][starVLA.training.train_starvla][INFO] - Step 4923, Loss: {'train/loss': 9.855336975306273e-05, 'train/grad_norm_pre_clip': 0.05133579671382904, 'timing/data': 0.0002067199966404587, 'timing/model': 0.39015270199161023, 'learning_rate/qwen_vl_interface': 5.05787146053909e-07, 'learning_rate/action_model': 5.0578714605390904e-06, 'epoch': 0.36}) +[2026-10-03 19:55:56,897][starVLA.training.train_starvla][INFO] - Step 4924, Loss: {'train/loss': 2.2231821276363917e-05, 'train/grad_norm_pre_clip': 0.006567818578332663, 'timing/data': 0.0003536200092639774, 'timing/model': 0.39109050200204365, 'learning_rate/qwen_vl_interface': 5.056378361957998e-07, 'learning_rate/action_model': 5.056378361957998e-06, 'epoch': 0.36}) +[2026-10-03 19:55:57,595][starVLA.training.train_starvla][INFO] - Step 4925, Loss: {'train/loss': 0.03862537443637848, 'train/grad_norm_pre_clip': 28.562414169311523, 'timing/data': 0.0003336000081617385, 'timing/model': 0.390696221991675, 'learning_rate/qwen_vl_interface': 5.05490476565766e-07, 'learning_rate/action_model': 5.054904765657659e-06, 'epoch': 0.36}) +[2026-10-03 19:55:58,497][starVLA.training.train_starvla][INFO] - Step 4926, Loss: {'train/loss': 0.2771243453025818, 'train/grad_norm_pre_clip': 54.76576614379883, 'timing/data': 0.0003313799970783293, 'timing/model': 0.3906163719948381, 'learning_rate/qwen_vl_interface': 5.053450672243811e-07, 'learning_rate/action_model': 5.0534506722438114e-06, 'epoch': 0.36}) +[2026-10-03 19:55:59,197][starVLA.training.train_starvla][INFO] - Step 4927, Loss: {'train/loss': 5.0776488933479413e-05, 'train/grad_norm_pre_clip': 0.012578283436596394, 'timing/data': 0.00029714000993408263, 'timing/model': 0.3912010009808, 'learning_rate/qwen_vl_interface': 5.052016082314184e-07, 'learning_rate/action_model': 5.052016082314184e-06, 'epoch': 0.36}) +[2026-10-03 19:55:59,905][starVLA.training.train_starvla][INFO] - Step 4928, Loss: {'train/loss': 0.017608074471354485, 'train/grad_norm_pre_clip': 46.546417236328125, 'timing/data': 0.00030256001628004014, 'timing/model': 0.3903000419959426, 'learning_rate/qwen_vl_interface': 5.050600996458473e-07, 'learning_rate/action_model': 5.050600996458473e-06, 'epoch': 0.36}) +[2026-10-03 19:56:00,602][starVLA.training.train_starvla][INFO] - Step 4929, Loss: {'train/loss': 7.310931687243283e-05, 'train/grad_norm_pre_clip': 0.11446349322795868, 'timing/data': 0.00024300001678057015, 'timing/model': 0.3903745219868142, 'learning_rate/qwen_vl_interface': 5.049205415258373e-07, 'learning_rate/action_model': 5.049205415258373e-06, 'epoch': 0.36}) +[2026-10-03 19:56:01,303][starVLA.training.train_starvla][INFO] - Step 4930, Loss: {'train/loss': 0.7482119798660278, 'train/grad_norm_pre_clip': 30.46407699584961, 'timing/data': 0.00032228001509793103, 'timing/model': 0.3913865209906362, 'learning_rate/qwen_vl_interface': 5.047829339287557e-07, 'learning_rate/action_model': 5.047829339287557e-06, 'epoch': 0.36}) +[2026-10-03 19:56:02,212][starVLA.training.train_starvla][INFO] - Step 4931, Loss: {'train/loss': 0.00822598859667778, 'train/grad_norm_pre_clip': 7.620881080627441, 'timing/data': 0.0003520500031299889, 'timing/model': 0.3915306220005732, 'learning_rate/qwen_vl_interface': 5.046472769111668e-07, 'learning_rate/action_model': 5.046472769111667e-06, 'epoch': 0.36}) +[2026-10-03 19:56:02,924][starVLA.training.train_starvla][INFO] - Step 4932, Loss: {'train/loss': 0.12878161668777466, 'train/grad_norm_pre_clip': 113.313720703125, 'timing/data': 0.00021078999270685017, 'timing/model': 0.3919803809840232, 'learning_rate/qwen_vl_interface': 5.045135705288353e-07, 'learning_rate/action_model': 5.0451357052883525e-06, 'epoch': 0.36}) +[2026-10-03 19:56:03,623][starVLA.training.train_starvla][INFO] - Step 4933, Loss: {'train/loss': 0.7245005369186401, 'train/grad_norm_pre_clip': 56.257022857666016, 'timing/data': 0.0003490799863357097, 'timing/model': 0.39161881202016957, 'learning_rate/qwen_vl_interface': 5.043818148367221e-07, 'learning_rate/action_model': 5.043818148367221e-06, 'epoch': 0.36}) +[2026-10-03 19:56:04,321][starVLA.training.train_starvla][INFO] - Step 4934, Loss: {'train/loss': 0.3983534872531891, 'train/grad_norm_pre_clip': 42.96725082397461, 'timing/data': 0.00035320999450050294, 'timing/model': 0.39094960200600326, 'learning_rate/qwen_vl_interface': 5.042520098889869e-07, 'learning_rate/action_model': 5.0425200988898685e-06, 'epoch': 0.36}) +[2026-10-03 19:56:05,021][starVLA.training.train_starvla][INFO] - Step 4935, Loss: {'train/loss': 0.05826572701334953, 'train/grad_norm_pre_clip': 116.1872787475586, 'timing/data': 0.0003568399988580495, 'timing/model': 0.3914007019775454, 'learning_rate/qwen_vl_interface': 5.041241557389879e-07, 'learning_rate/action_model': 5.041241557389879e-06, 'epoch': 0.36}) +[2026-10-03 19:56:05,935][starVLA.training.train_starvla][INFO] - Step 4936, Loss: {'train/loss': 3.4627169952727854e-05, 'train/grad_norm_pre_clip': 0.02095235325396061, 'timing/data': 0.0003418500127736479, 'timing/model': 0.5956465420022141, 'learning_rate/qwen_vl_interface': 5.039982524392814e-07, 'learning_rate/action_model': 5.039982524392814e-06, 'epoch': 0.36}) +[2026-10-03 19:56:06,632][starVLA.training.train_starvla][INFO] - Step 4937, Loss: {'train/loss': 0.0006902776076458395, 'train/grad_norm_pre_clip': 1.1029618978500366, 'timing/data': 0.0003472800017334521, 'timing/model': 0.3910069320118055, 'learning_rate/qwen_vl_interface': 5.038743000416207e-07, 'learning_rate/action_model': 5.0387430004162064e-06, 'epoch': 0.36}) +[2026-10-03 19:56:07,331][starVLA.training.train_starvla][INFO] - Step 4938, Loss: {'train/loss': 0.2869027554988861, 'train/grad_norm_pre_clip': 69.05191802978516, 'timing/data': 0.0002188800135627389, 'timing/model': 0.3906684719840996, 'learning_rate/qwen_vl_interface': 5.037522985969588e-07, 'learning_rate/action_model': 5.037522985969587e-06, 'epoch': 0.36}) +[2026-10-03 19:56:08,031][starVLA.training.train_starvla][INFO] - Step 4939, Loss: {'train/loss': 0.5916091799736023, 'train/grad_norm_pre_clip': 53.9852409362793, 'timing/data': 0.0003308999876026064, 'timing/model': 0.3909877619880717, 'learning_rate/qwen_vl_interface': 5.036322481554451e-07, 'learning_rate/action_model': 5.036322481554451e-06, 'epoch': 0.36}) +[2026-10-03 19:56:08,730][starVLA.training.train_starvla][INFO] - Step 4940, Loss: {'train/loss': 4.5457905798684806e-05, 'train/grad_norm_pre_clip': 0.011188676580786705, 'timing/data': 0.0003750500036403537, 'timing/model': 0.3914732320117764, 'learning_rate/qwen_vl_interface': 5.035141487664285e-07, 'learning_rate/action_model': 5.035141487664285e-06, 'epoch': 0.36}) +[2026-10-03 19:56:09,632][starVLA.training.train_starvla][INFO] - Step 4941, Loss: {'train/loss': 0.0010947933187708259, 'train/grad_norm_pre_clip': 1.410569429397583, 'timing/data': 0.00039177999133244157, 'timing/model': 0.5957690319919493, 'learning_rate/qwen_vl_interface': 5.03398000478455e-07, 'learning_rate/action_model': 5.03398000478455e-06, 'epoch': 0.36}) +[2026-10-03 19:56:10,334][starVLA.training.train_starvla][INFO] - Step 4942, Loss: {'train/loss': 4.3632404413074255e-05, 'train/grad_norm_pre_clip': 0.015737636014819145, 'timing/data': 0.00025058002211153507, 'timing/model': 0.3923784409998916, 'learning_rate/qwen_vl_interface': 5.032838033392686e-07, 'learning_rate/action_model': 5.032838033392685e-06, 'epoch': 0.36}) +[2026-10-03 19:56:11,031][starVLA.training.train_starvla][INFO] - Step 4943, Loss: {'train/loss': 0.0008039783569984138, 'train/grad_norm_pre_clip': 0.6679079532623291, 'timing/data': 0.00025363999884575605, 'timing/model': 0.3878975819970947, 'learning_rate/qwen_vl_interface': 5.031715573958111e-07, 'learning_rate/action_model': 5.031715573958111e-06, 'epoch': 0.36}) +[2026-10-03 19:56:11,731][starVLA.training.train_starvla][INFO] - Step 4944, Loss: {'train/loss': 0.563129186630249, 'train/grad_norm_pre_clip': 136.19866943359375, 'timing/data': 0.000334118987666443, 'timing/model': 0.3902215030102525, 'learning_rate/qwen_vl_interface': 5.030612626942235e-07, 'learning_rate/action_model': 5.030612626942235e-06, 'epoch': 0.36}) +[2026-10-03 19:56:12,430][starVLA.training.train_starvla][INFO] - Step 4945, Loss: {'train/loss': 9.487572242505848e-05, 'train/grad_norm_pre_clip': 0.034924376755952835, 'timing/data': 0.0003435499966144562, 'timing/model': 0.39135872200131416, 'learning_rate/qwen_vl_interface': 5.029529192798433e-07, 'learning_rate/action_model': 5.029529192798433e-06, 'epoch': 0.36}) +[2026-10-03 19:56:13,130][starVLA.training.train_starvla][INFO] - Step 4946, Loss: {'train/loss': 0.00030846719164401293, 'train/grad_norm_pre_clip': 0.12344847619533539, 'timing/data': 0.00025703999563120306, 'timing/model': 0.39196206198539585, 'learning_rate/qwen_vl_interface': 5.028465271972058e-07, 'learning_rate/action_model': 5.028465271972058e-06, 'epoch': 0.36}) +[2026-10-03 19:56:14,029][starVLA.training.train_starvla][INFO] - Step 4947, Loss: {'train/loss': 0.0004998494987376034, 'train/grad_norm_pre_clip': 0.45446789264678955, 'timing/data': 0.0003541799960657954, 'timing/model': 0.39098616200499237, 'learning_rate/qwen_vl_interface': 5.027420864900457e-07, 'learning_rate/action_model': 5.027420864900456e-06, 'epoch': 0.36}) +[2026-10-03 19:56:14,729][starVLA.training.train_starvla][INFO] - Step 4948, Loss: {'train/loss': 0.0006811491912230849, 'train/grad_norm_pre_clip': 0.8589609265327454, 'timing/data': 0.0002397099742665887, 'timing/model': 0.3913408009975683, 'learning_rate/qwen_vl_interface': 5.026395972012943e-07, 'learning_rate/action_model': 5.0263959720129436e-06, 'epoch': 0.36}) +[2026-10-03 19:56:15,427][starVLA.training.train_starvla][INFO] - Step 4949, Loss: {'train/loss': 0.3287317156791687, 'train/grad_norm_pre_clip': 41.98879623413086, 'timing/data': 0.00034994000452570617, 'timing/model': 0.3915025120077189, 'learning_rate/qwen_vl_interface': 5.025390593730809e-07, 'learning_rate/action_model': 5.025390593730809e-06, 'epoch': 0.36}) +[2026-10-03 19:56:16,129][starVLA.training.train_starvla][INFO] - Step 4950, Loss: {'train/loss': 1.9721297576325014e-05, 'train/grad_norm_pre_clip': 0.0050508808344602585, 'timing/data': 0.0003140199987683445, 'timing/model': 0.39324835198931396, 'learning_rate/qwen_vl_interface': 5.024404730467325e-07, 'learning_rate/action_model': 5.024404730467325e-06, 'epoch': 0.36}) +[2026-10-03 19:56:16,827][starVLA.training.train_starvla][INFO] - Step 4951, Loss: {'train/loss': 0.00024501586449332535, 'train/grad_norm_pre_clip': 0.11302196979522705, 'timing/data': 0.0003227390116080642, 'timing/model': 0.3902596529806033, 'learning_rate/qwen_vl_interface': 5.023438382627749e-07, 'learning_rate/action_model': 5.0234383826277486e-06, 'epoch': 0.36}) +[2026-10-03 19:56:17,728][starVLA.training.train_starvla][INFO] - Step 4952, Loss: {'train/loss': 0.018185878172516823, 'train/grad_norm_pre_clip': 30.644227981567383, 'timing/data': 0.00021328000002540648, 'timing/model': 0.3918773609912023, 'learning_rate/qwen_vl_interface': 5.022491550609311e-07, 'learning_rate/action_model': 5.022491550609311e-06, 'epoch': 0.36}) +[2026-10-03 19:56:18,427][starVLA.training.train_starvla][INFO] - Step 4953, Loss: {'train/loss': 1.843228528741747e-05, 'train/grad_norm_pre_clip': 0.004740193020552397, 'timing/data': 0.0002824600087478757, 'timing/model': 0.391194551018998, 'learning_rate/qwen_vl_interface': 5.021564234801213e-07, 'learning_rate/action_model': 5.021564234801213e-06, 'epoch': 0.36}) +[2026-10-03 19:56:19,127][starVLA.training.train_starvla][INFO] - Step 4954, Loss: {'train/loss': 0.00017093766655307263, 'train/grad_norm_pre_clip': 0.23803284764289856, 'timing/data': 0.00025267002638429403, 'timing/model': 0.39181510198977776, 'learning_rate/qwen_vl_interface': 5.020656435584636e-07, 'learning_rate/action_model': 5.020656435584635e-06, 'epoch': 0.36}) +[2026-10-03 19:56:19,828][starVLA.training.train_starvla][INFO] - Step 4955, Loss: {'train/loss': 0.0007165874121710658, 'train/grad_norm_pre_clip': 1.0878868103027344, 'timing/data': 0.0003194799937773496, 'timing/model': 0.39251649199286476, 'learning_rate/qwen_vl_interface': 5.019768153332752e-07, 'learning_rate/action_model': 5.019768153332752e-06, 'epoch': 0.36}) +[2026-10-03 19:56:20,529][starVLA.training.train_starvla][INFO] - Step 4956, Loss: {'train/loss': 0.00013711945211980492, 'train/grad_norm_pre_clip': 0.14151252806186676, 'timing/data': 0.00021860000560991466, 'timing/model': 0.3919299319968559, 'learning_rate/qwen_vl_interface': 5.018899388410693e-07, 'learning_rate/action_model': 5.0188993884106935e-06, 'epoch': 0.36}) +[2026-10-03 19:56:21,429][starVLA.training.train_starvla][INFO] - Step 4957, Loss: {'train/loss': 0.010562680661678314, 'train/grad_norm_pre_clip': 21.379552841186523, 'timing/data': 0.00020711999968625605, 'timing/model': 0.5932157110073604, 'learning_rate/qwen_vl_interface': 5.018050141175576e-07, 'learning_rate/action_model': 5.018050141175575e-06, 'epoch': 0.36}) +[2026-10-03 19:56:22,129][starVLA.training.train_starvla][INFO] - Step 4958, Loss: {'train/loss': 0.00011107302270829678, 'train/grad_norm_pre_clip': 0.16429515182971954, 'timing/data': 0.00025164999533444643, 'timing/model': 0.39164725100272335, 'learning_rate/qwen_vl_interface': 5.017220411976496e-07, 'learning_rate/action_model': 5.0172204119764965e-06, 'epoch': 0.36}) +[2026-10-03 19:56:22,828][starVLA.training.train_starvla][INFO] - Step 4959, Loss: {'train/loss': 0.0005698568420484662, 'train/grad_norm_pre_clip': 0.494263619184494, 'timing/data': 0.00032183999428525567, 'timing/model': 0.3924364019767381, 'learning_rate/qwen_vl_interface': 5.016410201154523e-07, 'learning_rate/action_model': 5.016410201154523e-06, 'epoch': 0.36}) +[2026-10-03 19:56:23,503][starVLA.training.train_starvla][INFO] - Step 4960, Loss: {'train/loss': 4.6494315029121935e-05, 'train/grad_norm_pre_clip': 0.017183389514684677, 'timing/data': 0.00020574001246131957, 'timing/model': 0.3904387620277703, 'learning_rate/qwen_vl_interface': 5.015619509042709e-07, 'learning_rate/action_model': 5.015619509042709e-06, 'epoch': 0.36}) +[2026-10-03 19:56:24,202][starVLA.training.train_starvla][INFO] - Step 4961, Loss: {'train/loss': 0.0005023753037676215, 'train/grad_norm_pre_clip': 0.5776850581169128, 'timing/data': 0.00034470000537112355, 'timing/model': 0.38986621200456284, 'learning_rate/qwen_vl_interface': 5.014848335966067e-07, 'learning_rate/action_model': 5.014848335966067e-06, 'epoch': 0.36}) +[2026-10-03 19:56:25,104][starVLA.training.train_starvla][INFO] - Step 4962, Loss: {'train/loss': 2.8921731427544728e-05, 'train/grad_norm_pre_clip': 0.011057605035603046, 'timing/data': 0.00024838000535964966, 'timing/model': 0.5933219120197464, 'learning_rate/qwen_vl_interface': 5.014096682241608e-07, 'learning_rate/action_model': 5.014096682241608e-06, 'epoch': 0.36}) +[2026-10-03 19:56:25,803][starVLA.training.train_starvla][INFO] - Step 4963, Loss: {'train/loss': 0.0008057808154262602, 'train/grad_norm_pre_clip': 1.1855632066726685, 'timing/data': 0.0003132499987259507, 'timing/model': 0.39023664200794883, 'learning_rate/qwen_vl_interface': 5.013364548178299e-07, 'learning_rate/action_model': 5.013364548178299e-06, 'epoch': 0.36}) +[2026-10-03 19:56:26,504][starVLA.training.train_starvla][INFO] - Step 4964, Loss: {'train/loss': 0.6588546633720398, 'train/grad_norm_pre_clip': 323.1987609863281, 'timing/data': 0.0003545400104485452, 'timing/model': 0.3910244519938715, 'learning_rate/qwen_vl_interface': 5.012651934077101e-07, 'learning_rate/action_model': 5.0126519340771e-06, 'epoch': 0.36}) +[2026-10-03 19:56:27,201][starVLA.training.train_starvla][INFO] - Step 4965, Loss: {'train/loss': 7.5650452345144e-05, 'train/grad_norm_pre_clip': 0.0242412481456995, 'timing/data': 0.00035143899731338024, 'timing/model': 0.38955062298919074, 'learning_rate/qwen_vl_interface': 5.01195884023094e-07, 'learning_rate/action_model': 5.0119588402309395e-06, 'epoch': 0.36}) +[2026-10-03 19:56:27,901][starVLA.training.train_starvla][INFO] - Step 4966, Loss: {'train/loss': 0.03818053379654884, 'train/grad_norm_pre_clip': 77.81665802001953, 'timing/data': 0.00026157902902923524, 'timing/model': 0.3923787919920869, 'learning_rate/qwen_vl_interface': 5.011285266924714e-07, 'learning_rate/action_model': 5.011285266924714e-06, 'epoch': 0.36}) +[2026-10-03 19:56:28,600][starVLA.training.train_starvla][INFO] - Step 4967, Loss: {'train/loss': 0.0011768621625378728, 'train/grad_norm_pre_clip': 2.2050600051879883, 'timing/data': 0.0003259499790146947, 'timing/model': 0.3899957220128272, 'learning_rate/qwen_vl_interface': 5.010631214435312e-07, 'learning_rate/action_model': 5.010631214435312e-06, 'epoch': 0.36}) +[2026-10-03 19:56:29,504][starVLA.training.train_starvla][INFO] - Step 4968, Loss: {'train/loss': 5.9152967878617346e-05, 'train/grad_norm_pre_clip': 0.06480801850557327, 'timing/data': 0.0003606100217439234, 'timing/model': 0.39213093102443963, 'learning_rate/qwen_vl_interface': 5.009996683031586e-07, 'learning_rate/action_model': 5.009996683031586e-06, 'epoch': 0.36}) +[2026-10-03 19:56:30,200][starVLA.training.train_starvla][INFO] - Step 4969, Loss: {'train/loss': 0.01018134318292141, 'train/grad_norm_pre_clip': 23.394922256469727, 'timing/data': 0.00034239000524394214, 'timing/model': 0.38932343301712535, 'learning_rate/qwen_vl_interface': 5.009381672974375e-07, 'learning_rate/action_model': 5.009381672974375e-06, 'epoch': 0.36}) +[2026-10-03 19:56:30,900][starVLA.training.train_starvla][INFO] - Step 4970, Loss: {'train/loss': 0.8218673467636108, 'train/grad_norm_pre_clip': 262.29437255859375, 'timing/data': 0.00020541000412777066, 'timing/model': 0.3928440720192157, 'learning_rate/qwen_vl_interface': 5.008786184516475e-07, 'learning_rate/action_model': 5.008786184516474e-06, 'epoch': 0.36}) +[2026-10-03 19:56:31,597][starVLA.training.train_starvla][INFO] - Step 4971, Loss: {'train/loss': 0.00013158276851754636, 'train/grad_norm_pre_clip': 0.14805789291858673, 'timing/data': 0.00020601000869646668, 'timing/model': 0.3893594320106786, 'learning_rate/qwen_vl_interface': 5.008210217902676e-07, 'learning_rate/action_model': 5.008210217902676e-06, 'epoch': 0.36}) +[2026-10-03 19:56:32,294][starVLA.training.train_starvla][INFO] - Step 4972, Loss: {'train/loss': 5.946323290118016e-05, 'train/grad_norm_pre_clip': 0.029816066846251488, 'timing/data': 0.00020582001889124513, 'timing/model': 0.3897065919882152, 'learning_rate/qwen_vl_interface': 5.007653773369737e-07, 'learning_rate/action_model': 5.007653773369736e-06, 'epoch': 0.36}) +[2026-10-03 19:56:33,202][starVLA.training.train_starvla][INFO] - Step 4973, Loss: {'train/loss': 0.3922092616558075, 'train/grad_norm_pre_clip': 51.66990661621094, 'timing/data': 0.00023304001661017537, 'timing/model': 0.3917620409920346, 'learning_rate/qwen_vl_interface': 5.007116851146386e-07, 'learning_rate/action_model': 5.007116851146386e-06, 'epoch': 0.36}) +[2026-10-03 19:56:33,903][starVLA.training.train_starvla][INFO] - Step 4974, Loss: {'train/loss': 0.15400011837482452, 'train/grad_norm_pre_clip': 96.0085220336914, 'timing/data': 0.00026105999131686985, 'timing/model': 0.39291447200230323, 'learning_rate/qwen_vl_interface': 5.006599451453334e-07, 'learning_rate/action_model': 5.006599451453334e-06, 'epoch': 0.36}) +[2026-10-03 19:56:34,599][starVLA.training.train_starvla][INFO] - Step 4975, Loss: {'train/loss': 0.1256202757358551, 'train/grad_norm_pre_clip': 111.66326141357422, 'timing/data': 0.000255580001976341, 'timing/model': 0.3901558030047454, 'learning_rate/qwen_vl_interface': 5.006101574503268e-07, 'learning_rate/action_model': 5.006101574503267e-06, 'epoch': 0.36}) +[2026-10-03 19:56:35,298][starVLA.training.train_starvla][INFO] - Step 4976, Loss: {'train/loss': 0.18981163203716278, 'train/grad_norm_pre_clip': 270.6300354003906, 'timing/data': 0.0002044200082309544, 'timing/model': 0.39146667200839147, 'learning_rate/qwen_vl_interface': 5.005623220500839e-07, 'learning_rate/action_model': 5.005623220500839e-06, 'epoch': 0.36}) +[2026-10-03 19:56:35,995][starVLA.training.train_starvla][INFO] - Step 4977, Loss: {'train/loss': 0.0001937704801093787, 'train/grad_norm_pre_clip': 0.2333446592092514, 'timing/data': 0.0002104500017594546, 'timing/model': 0.39129593200050294, 'learning_rate/qwen_vl_interface': 5.005164389642687e-07, 'learning_rate/action_model': 5.005164389642687e-06, 'epoch': 0.36}) +[2026-10-03 19:56:36,898][starVLA.training.train_starvla][INFO] - Step 4978, Loss: {'train/loss': 8.152799273375422e-05, 'train/grad_norm_pre_clip': 0.05152561515569687, 'timing/data': 0.00019988001440651715, 'timing/model': 0.5970913219789509, 'learning_rate/qwen_vl_interface': 5.004725082117423e-07, 'learning_rate/action_model': 5.004725082117422e-06, 'epoch': 0.36}) +[2026-10-03 19:56:37,595][starVLA.training.train_starvla][INFO] - Step 4979, Loss: {'train/loss': 0.006624843459576368, 'train/grad_norm_pre_clip': 14.32257080078125, 'timing/data': 0.00030568998772650957, 'timing/model': 0.389870023005642, 'learning_rate/qwen_vl_interface': 5.004305298105616e-07, 'learning_rate/action_model': 5.004305298105616e-06, 'epoch': 0.36}) +[2026-10-03 19:56:38,294][starVLA.training.train_starvla][INFO] - Step 4980, Loss: {'train/loss': 0.06548380106687546, 'train/grad_norm_pre_clip': 93.01667022705078, 'timing/data': 0.00021195999579504132, 'timing/model': 0.39083912200294435, 'learning_rate/qwen_vl_interface': 5.003905037779834e-07, 'learning_rate/action_model': 5.0039050377798345e-06, 'epoch': 0.36}) +[2026-10-03 19:56:38,992][starVLA.training.train_starvla][INFO] - Step 4981, Loss: {'train/loss': 0.003983946982771158, 'train/grad_norm_pre_clip': 6.278971195220947, 'timing/data': 0.0002160299918614328, 'timing/model': 0.3904403619817458, 'learning_rate/qwen_vl_interface': 5.003524301304608e-07, 'learning_rate/action_model': 5.003524301304608e-06, 'epoch': 0.36}) +[2026-10-03 19:56:39,694][starVLA.training.train_starvla][INFO] - Step 4982, Loss: {'train/loss': 0.022275783121585846, 'train/grad_norm_pre_clip': 54.45200729370117, 'timing/data': 0.000195720000192523, 'timing/model': 0.3949524109775666, 'learning_rate/qwen_vl_interface': 5.003163088836445e-07, 'learning_rate/action_model': 5.003163088836446e-06, 'epoch': 0.36}) +[2026-10-03 19:56:40,390][starVLA.training.train_starvla][INFO] - Step 4983, Loss: {'train/loss': 0.0002725682861637324, 'train/grad_norm_pre_clip': 0.2046278715133667, 'timing/data': 0.00020087999291718006, 'timing/model': 0.38978900099755265, 'learning_rate/qwen_vl_interface': 5.002821400523818e-07, 'learning_rate/action_model': 5.002821400523818e-06, 'epoch': 0.36}) +[2026-10-03 19:56:41,320][starVLA.training.train_starvla][INFO] - Step 4984, Loss: {'train/loss': 1.826840525609441e-05, 'train/grad_norm_pre_clip': 0.004130191169679165, 'timing/data': 0.0003241300000809133, 'timing/model': 0.39256702101556584, 'learning_rate/qwen_vl_interface': 5.002499236507192e-07, 'learning_rate/action_model': 5.0024992365071925e-06, 'epoch': 0.36}) +[2026-10-03 19:56:42,019][starVLA.training.train_starvla][INFO] - Step 4985, Loss: {'train/loss': 2.8780263164662756e-05, 'train/grad_norm_pre_clip': 0.007846294902265072, 'timing/data': 0.0002139100106433034, 'timing/model': 0.39093011201475747, 'learning_rate/qwen_vl_interface': 5.002196596918993e-07, 'learning_rate/action_model': 5.002196596918993e-06, 'epoch': 0.36}) +[2026-10-03 19:56:42,718][starVLA.training.train_starvla][INFO] - Step 4986, Loss: {'train/loss': 4.970479858457111e-05, 'train/grad_norm_pre_clip': 0.012223988771438599, 'timing/data': 0.00030662998324260116, 'timing/model': 0.3919266820012126, 'learning_rate/qwen_vl_interface': 5.001913481883623e-07, 'learning_rate/action_model': 5.001913481883623e-06, 'epoch': 0.36}) +[2026-10-03 19:56:43,416][starVLA.training.train_starvla][INFO] - Step 4987, Loss: {'train/loss': 5.488894385052845e-05, 'train/grad_norm_pre_clip': 0.02293529361486435, 'timing/data': 0.0003289500018581748, 'timing/model': 0.3907920529891271, 'learning_rate/qwen_vl_interface': 5.001649891517459e-07, 'learning_rate/action_model': 5.001649891517459e-06, 'epoch': 0.37}) +[2026-10-03 19:56:44,118][starVLA.training.train_starvla][INFO] - Step 4988, Loss: {'train/loss': 0.00013100016803946346, 'train/grad_norm_pre_clip': 0.13190752267837524, 'timing/data': 0.0002719300100579858, 'timing/model': 0.39250495200394653, 'learning_rate/qwen_vl_interface': 5.001405825928858e-07, 'learning_rate/action_model': 5.0014058259288575e-06, 'epoch': 0.37}) +[2026-10-03 19:56:45,012][starVLA.training.train_starvla][INFO] - Step 4989, Loss: {'train/loss': 0.0009824137669056654, 'train/grad_norm_pre_clip': 1.0371363162994385, 'timing/data': 0.0003281600074842572, 'timing/model': 0.5883095519966446, 'learning_rate/qwen_vl_interface': 5.001181285218143e-07, 'learning_rate/action_model': 5.0011812852181434e-06, 'epoch': 0.37}) +[2026-10-03 19:56:45,711][starVLA.training.train_starvla][INFO] - Step 4990, Loss: {'train/loss': 0.00011627695494098589, 'train/grad_norm_pre_clip': 0.07658010721206665, 'timing/data': 0.0002369800058659166, 'timing/model': 0.3924439410038758, 'learning_rate/qwen_vl_interface': 5.000976269477613e-07, 'learning_rate/action_model': 5.0009762694776125e-06, 'epoch': 0.37}) +[2026-10-03 19:56:46,409][starVLA.training.train_starvla][INFO] - Step 4991, Loss: {'train/loss': 0.00041359246824868023, 'train/grad_norm_pre_clip': 0.4411000609397888, 'timing/data': 0.0003246100095566362, 'timing/model': 0.39039711200166494, 'learning_rate/qwen_vl_interface': 5.000790778791542e-07, 'learning_rate/action_model': 5.000790778791543e-06, 'epoch': 0.37}) +[2026-10-03 19:56:47,113][starVLA.training.train_starvla][INFO] - Step 4992, Loss: {'train/loss': 0.0008421832462772727, 'train/grad_norm_pre_clip': 1.0387643575668335, 'timing/data': 0.00020733001292683184, 'timing/model': 0.39369111199630424, 'learning_rate/qwen_vl_interface': 5.000624813236185e-07, 'learning_rate/action_model': 5.000624813236185e-06, 'epoch': 0.37}) +[2026-10-03 19:56:47,812][starVLA.training.train_starvla][INFO] - Step 4993, Loss: {'train/loss': 0.5859835743904114, 'train/grad_norm_pre_clip': 70.50439453125, 'timing/data': 0.00033044000156223774, 'timing/model': 0.39234874199610204, 'learning_rate/qwen_vl_interface': 5.000478372879752e-07, 'learning_rate/action_model': 5.000478372879752e-06, 'epoch': 0.37}) +[2026-10-03 19:56:48,718][starVLA.training.train_starvla][INFO] - Step 4994, Loss: {'train/loss': 0.0004175815265625715, 'train/grad_norm_pre_clip': 0.8992804288864136, 'timing/data': 0.00023536899243481457, 'timing/model': 0.5969683320145123, 'learning_rate/qwen_vl_interface': 5.000351457782452e-07, 'learning_rate/action_model': 5.000351457782451e-06, 'epoch': 0.37}) +[2026-10-03 19:56:49,418][starVLA.training.train_starvla][INFO] - Step 4995, Loss: {'train/loss': 4.77213689009659e-05, 'train/grad_norm_pre_clip': 0.0256284661591053, 'timing/data': 0.00021050000214017928, 'timing/model': 0.3906943920010235, 'learning_rate/qwen_vl_interface': 5.000244067996445e-07, 'learning_rate/action_model': 5.000244067996445e-06, 'epoch': 0.37}) +[2026-10-03 19:56:50,119][starVLA.training.train_starvla][INFO] - Step 4996, Loss: {'train/loss': 0.0010221862467005849, 'train/grad_norm_pre_clip': 0.9279869198799133, 'timing/data': 0.000328739988617599, 'timing/model': 0.39188997098244727, 'learning_rate/qwen_vl_interface': 5.000156203565884e-07, 'learning_rate/action_model': 5.000156203565883e-06, 'epoch': 0.37}) +[2026-10-03 19:56:50,819][starVLA.training.train_starvla][INFO] - Step 4997, Loss: {'train/loss': 0.0002515072701498866, 'train/grad_norm_pre_clip': 0.06617984175682068, 'timing/data': 0.00021374999778345227, 'timing/model': 0.3932003419904504, 'learning_rate/qwen_vl_interface': 5.000087864526878e-07, 'learning_rate/action_model': 5.000087864526878e-06, 'epoch': 0.37}) +[2026-10-03 19:56:51,521][starVLA.training.train_starvla][INFO] - Step 4998, Loss: {'train/loss': 0.7288081049919128, 'train/grad_norm_pre_clip': 79.82188415527344, 'timing/data': 0.0003505299973767251, 'timing/model': 0.39371428202139214, 'learning_rate/qwen_vl_interface': 5.000039050907523e-07, 'learning_rate/action_model': 5.000039050907523e-06, 'epoch': 0.37}) +[2026-10-03 19:56:52,221][starVLA.training.train_starvla][INFO] - Step 4999, Loss: {'train/loss': 0.0004811378021258861, 'train/grad_norm_pre_clip': 0.5317917466163635, 'timing/data': 0.0002704399812500924, 'timing/model': 0.3917869620199781, 'learning_rate/qwen_vl_interface': 5.000009762727885e-07, 'learning_rate/action_model': 5.000009762727884e-06, 'epoch': 0.37}) +[2026-10-03 19:56:53,127][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 5000 +[2026-10-03 19:57:20,349][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 5000 +[2026-10-03 19:57:20,349][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 5000 +[2026-10-03 19:57:20,349][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 5000 +[2026-10-03 19:57:20,350][starVLA.training.train_starvla][INFO] - Step 5000, Loss: {'train/loss': 0.00012759873061440885, 'train/grad_norm_pre_clip': 0.0777597427368164, 'eval/action_loss/seconds': 27.2197511580016, 'eval/action_loss/samples': 120.0, 'eval/loss': 0.897685170173645, 'eval/latency_0/loss': 2.5731334686279297, 'eval/latency_2/loss': 0.00010004248179029673, 'eval/latency_4/loss': 0.11982198059558868, 'eval/inverted_pendulum_rgb_state/loss': 0.897685170173645, 'eval/inverted_pendulum_rgb_state/latency_0/loss': 2.5731334686279297, 'eval/inverted_pendulum_rgb_state/latency_2/loss': 0.00010004248179029673, 'eval/inverted_pendulum_rgb_state/latency_4/loss': 0.11982198059558868, 'timing/data': 0.0003056290152017027, 'timing/model': 0.3931595419999212, 'learning_rate/qwen_vl_interface': 5e-07, 'learning_rate/action_model': 4.9999999999999996e-06, 'epoch': 0.37}) +[2026-10-03 19:57:20,417][root][INFO] - /data/ubuntu/conda_envs/starvla_rl_games_openvla/bin/x86_64-conda-linux-gnu-cc -march=nocona -mtune=haswell -ftree-vectorize -fPIC -fstack-protector-strong -fno-plt -O2 -ffunction-sections -pipe -isystem /data/ubuntu/conda_envs/starvla_rl_games_openvla/include -I/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/include -L/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/lib -L/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/lib/stubs -DNDEBUG -D_FORTIFY_SOURCE=2 -O2 -isystem /data/ubuntu/conda_envs/starvla_rl_games_openvla/include -I/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/include -L/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/lib -L/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/lib/stubs -fPIC -march=nocona -mtune=haswell -ftree-vectorize -fPIC -fstack-protector-strong -fno-plt -O2 -ffunction-sections -pipe -isystem /data/ubuntu/conda_envs/starvla_rl_games_openvla/include -I/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/include -L/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/lib -L/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/lib/stubs -c /tmp/tmptpp7hh0i/test.c -o /tmp/tmptpp7hh0i/test.o +[2026-10-03 19:57:20,426][root][INFO] - /data/ubuntu/conda_envs/starvla_rl_games_openvla/bin/x86_64-conda-linux-gnu-cc -Wl,-O2 -Wl,--sort-common -Wl,--as-needed -Wl,-z,relro -Wl,-z,now -Wl,--disable-new-dtags -Wl,--gc-sections -Wl,--allow-shlib-undefined -Wl,-rpath,/data/ubuntu/conda_envs/starvla_rl_games_openvla/lib -Wl,-rpath-link,/data/ubuntu/conda_envs/starvla_rl_games_openvla/lib -L/data/ubuntu/conda_envs/starvla_rl_games_openvla/lib -L/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/lib -L/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/lib/stubs /tmp/tmptpp7hh0i/test.o -laio -o /tmp/tmptpp7hh0i/a.out +[2026-10-03 19:57:20,434][root][INFO] - /data/ubuntu/conda_envs/starvla_rl_games_openvla/bin/x86_64-conda-linux-gnu-cc -march=nocona -mtune=haswell -ftree-vectorize -fPIC -fstack-protector-strong -fno-plt -O2 -ffunction-sections -pipe -isystem /data/ubuntu/conda_envs/starvla_rl_games_openvla/include -I/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/include -L/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/lib -L/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/lib/stubs -DNDEBUG -D_FORTIFY_SOURCE=2 -O2 -isystem /data/ubuntu/conda_envs/starvla_rl_games_openvla/include -I/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/include -L/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/lib -L/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/lib/stubs -fPIC -march=nocona -mtune=haswell -ftree-vectorize -fPIC -fstack-protector-strong -fno-plt -O2 -ffunction-sections -pipe -isystem /data/ubuntu/conda_envs/starvla_rl_games_openvla/include -I/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/include -L/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/lib -L/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/lib/stubs -c /tmp/tmpa7_tzt6f/test.c -o /tmp/tmpa7_tzt6f/test.o +[2026-10-03 19:57:20,443][root][INFO] - /data/ubuntu/conda_envs/starvla_rl_games_openvla/bin/x86_64-conda-linux-gnu-cc -Wl,-O2 -Wl,--sort-common -Wl,--as-needed -Wl,-z,relro -Wl,-z,now -Wl,--disable-new-dtags -Wl,--gc-sections -Wl,--allow-shlib-undefined -Wl,-rpath,/data/ubuntu/conda_envs/starvla_rl_games_openvla/lib -Wl,-rpath-link,/data/ubuntu/conda_envs/starvla_rl_games_openvla/lib -L/data/ubuntu/conda_envs/starvla_rl_games_openvla/lib -L/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/lib -L/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/lib/stubs /tmp/tmpa7_tzt6f/test.o -L/data/ubuntu/conda_envs/starvla_rl_games_openvla -L/data/ubuntu/conda_envs/starvla_rl_games_openvla/lib64 -lcufile -o /tmp/tmpa7_tzt6f/a.out +[2026-10-03 19:57:20,456][root][INFO] - /data/ubuntu/conda_envs/starvla_rl_games_openvla/bin/x86_64-conda-linux-gnu-cc -march=nocona -mtune=haswell -ftree-vectorize -fPIC -fstack-protector-strong -fno-plt -O2 -ffunction-sections -pipe -isystem /data/ubuntu/conda_envs/starvla_rl_games_openvla/include -I/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/include -L/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/lib -L/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/lib/stubs -DNDEBUG -D_FORTIFY_SOURCE=2 -O2 -isystem /data/ubuntu/conda_envs/starvla_rl_games_openvla/include -I/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/include -L/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/lib -L/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/lib/stubs -fPIC -march=nocona -mtune=haswell -ftree-vectorize -fPIC -fstack-protector-strong -fno-plt -O2 -ffunction-sections -pipe -isystem /data/ubuntu/conda_envs/starvla_rl_games_openvla/include -I/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/include -L/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/lib -L/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/lib/stubs -c /tmp/tmpqexcrtnz/test.c -o /tmp/tmpqexcrtnz/test.o +[2026-10-03 19:57:20,464][root][INFO] - /data/ubuntu/conda_envs/starvla_rl_games_openvla/bin/x86_64-conda-linux-gnu-cc -Wl,-O2 -Wl,--sort-common -Wl,--as-needed -Wl,-z,relro -Wl,-z,now -Wl,--disable-new-dtags -Wl,--gc-sections -Wl,--allow-shlib-undefined -Wl,-rpath,/data/ubuntu/conda_envs/starvla_rl_games_openvla/lib -Wl,-rpath-link,/data/ubuntu/conda_envs/starvla_rl_games_openvla/lib -L/data/ubuntu/conda_envs/starvla_rl_games_openvla/lib -L/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/lib -L/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/lib/stubs /tmp/tmpqexcrtnz/test.o -laio -o /tmp/tmpqexcrtnz/a.out +[2026-10-03 19:57:24,953][starVLA.training.rl_games.checkpoint_sync][WARNING] - HF checkpoint sync failed for repo talha15032/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3: (Request ID: Root=1-6ac15e24-3cccd5a0440d241776803252;7e662bf4-0d1f-4569-9a8f-e0b7dcbfab9a) + +403 Forbidden: You don't have the rights to create a model under the namespace "talha15032". +Cannot access content at: https://huggingface.co/api/repos/create. +Make sure your token has the correct permissions. +[2026-10-03 19:57:24,954][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-10-03 19:57:24,976][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-10-03 19:57:24,977][starVLA.training.train_starvla][INFO] - Training complete. Final training state is saved under /home/ubuntu/talha/results/Checkpoints/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3/checkpoints diff --git a/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3/summary.jsonl b/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3/summary.jsonl new file mode 100644 index 0000000000000000000000000000000000000000..04f41c132dfb80a0ccb65212db19ce96e77fd03c --- /dev/null +++ b/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_exp3/summary.jsonl @@ -0,0 +1 @@ +{"steps": 5000} diff --git a/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2/checkpoints/steps_5000_pytorch_model.pt b/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2/checkpoints/steps_5000_pytorch_model.pt new file mode 100644 index 0000000000000000000000000000000000000000..51017672a429ce9d6fe585058cb932624498f25f --- /dev/null +++ b/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2/checkpoints/steps_5000_pytorch_model.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d275016d512fe0fe071d5c2c9f7362fba939c89cb357e461d8b59e96c31ba62f +size 9138435024 diff --git a/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2/config.full.yaml b/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2/config.full.yaml new file mode 100644 index 0000000000000000000000000000000000000000..4a2f90dfcbb2c0280b8a12df9ab816d4e17d21a9 --- /dev/null +++ b/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2/config.full.yaml @@ -0,0 +1,338 @@ +framework: + name: QwenOFT + qwenvl: + base_vlm: /home/ubuntu/talha/playground/Pretrained_models/Qwen3-VL-4B-Instruct + attn_implementation: flash_attention_2 + enable_gradient_checkpointing: true + action_model: + action_model_type: MLP + action_dim: 5 + state_dim: 4 + action_hidden_dim: 2560 + future_action_window_size: 0 + past_action_window_size: 0 + loss_type: discrete_ce + state_encoding: discretized_text + action_horizon: 1 + action_env_dim: 5 + kv_memory: + enabled: false + window: 4 + rollout_len: 8 + packed_train: false + rebased_sink: true +datasets: + vla_data: + dataset_py: lerobot_datasets + include_state: true + data_root_dir: /home/ubuntu/talha/playground/Datasets/rl_games + data_mix: openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2 + eval_data_mix: openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2__val + custom_mixtures_path: /home/ubuntu/talha/playground/Datasets/rl_games/_generated_mixtures/single_622b136e13e57595.json + action_type: discrete + sequential_step_sampling: false + eval_sequential_step_sampling: null + num_workers: 8 + eval_num_workers: 8 + prefetch_factor: 4 + persistent_workers: true + pin_memory: true + shuffle: true + action_balance: + enabled: false + strategy: balanced_epoch + action_key: action_id + target_flap_fraction: 0.3 + noop_id: 0 + flap_id: 1 + latency_curriculum: + enabled: false + strategy: exclusive + latencies: null + phase_steps: null + phase_distributions: null + new_latency_passes: 1.0 + replay_passes: 0.25 + target_total_passes: 2.0 + final_equalization: true + step_budget_mode: auto + eval_at_phase_end: false + save_at_phase_end: false + computed_plan: null + per_device_batch_size: 8 + load_all_data_for_training: true + num_obs_frames: 1 + image_mode: single + prompt_mode: latency_neutral + stitch_grid: + - 2 + - 2 + obs_image_size: null + video_backend: torchvision_av + gymnasium_task_contract: + task_name: inverted_pendulum_rgb_state + env_id: LatencyBench/InvertedPendulumDiscrete-v0 + registration_imports: + - latency_bench.envs.gymnasium_mujoco + env_fps: 25.0 + obs_fps: 25.0 + frame_stack: 1 + base_prompt: 'Keep the pole upright by applying horizontal cart forces. Choose + exactly one action from: push_left_hard, push_left_soft, zero_force, push_right_soft, + push_right_hard.' + action_labels: + - push_left_hard + - push_left_soft + - zero_force + - push_right_soft + - push_right_hard + action_values: + - 0 + - 1 + - 2 + - 3 + - 4 + noop_action_id: 2 + make_kwargs: + base_env_id: InvertedPendulum-v4 + render_mode: rgb_array + force_values: + - -3.0 + - -1.5 + - 0.0 + - 1.5 + - 3.0 + base_make_kwargs: {} + state_space: + labels: + - cart_position + - pole_angle + - cart_velocity + - pole_angular_velocity + active_action_dim: 5 +dataset: + source_hf: latency-sensitive-bench/inverted_pendulum_200ep_2 + config_name: null + source_subdir: inverted_pendulum_fixed_latency_0_100ep_1ksteps_og,inverted_pendulum_fixed_latency_2_1000ep_7k2steps,inverted_pendulum_fixed_latency_4_1000ep_7k2steps + converted_name: openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2 + single_source_hf: '' + mixed_source_hf: '' + single_converted_name: flappy_train + mixed_converted_name: flappy_train + single_latency_filter: null + mixed_latency_filter: null + force_download: false + setup_force: true + skip_verification: true + target_latency_unit: observation_steps + verify_rows: 200 + max_episodes: null + max_steps_per_episode: 1000 + episodes_per_latency: null + episodes_per_latency_by_latency: 0:90,2:1000,4:1000 + latency_filter: + - 0 + - 2 + - 4 + debug_subset: + enabled: false + max_episodes: 5 + suffix: debug +base_model: + repo_id: Qwen/Qwen3-VL-4B-Instruct +initialization: + checkpoint_local_dir: null + checkpoint_hf_repo_id: null + checkpoint_filename: null +trainer: + max_train_steps: 5000 + num_warmup_steps: 100 + save_interval: 5000 + eval_interval: 5000 + eval_num_batches: 200 + per_latency_eval_num_batches: 5 + eval_action_classification: true + eval_action_classification_interval: null + cc_f1_tolerance: 1 + learning_rate: + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + action_model: 0.0001 + lr_scheduler_type: cosine_with_min_lr + scheduler_specific_kwargs: + min_lr: 1.0e-06 + freeze_modules: '' + freeze_vit: false + freeze_tied_embedding: false + freeze_llm_layers: [] + loss_scale: + vla: 1.0 + vlm: 0.1 + max_grad_norm: 1.0 + weight_decay: 0.0 + logging_frequency: 1 + profile_timing: + enabled: false + log_interval: 10 + gradient_clipping: 1.0 + gradient_accumulation_steps: 2 + distributed_backend: none + is_resume: false + pretrained_checkpoint: null + resume_step: 0 + reload_modules: null + optimizer: + name: AdamW + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + weight_decay: 1.0e-08 + fused: true + save_format: pt +workspace_dir: /home/ubuntu/talha +run_root_dir: /home/ubuntu/talha/results/Checkpoints +seed: 42 +wandb_entity: talha1503 +wandb_project: latency-sensitive-bench +wandb_group: openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2 +wandb_tags: [] +training_latency_condition: unspecified +auth: + env_file: null + hf_token_env: HF_TOKEN + wandb_api_key_env: WANDB_API_KEY +paths: + run_root_dir: results/Checkpoints + dataset_local_dir: playground/Datasets/rl_games + dataset_cache_dir: null + base_model_dir: /home/ubuntu/talha/playground/Pretrained_models/Qwen3-VL-4B-Instruct + accelerate_config: starVLA/config/deepseeds/deepspeed_zero2.yaml +launch: + use_accelerate: true + gpus: null + num_processes: 1 + dry_run: false +conda: + enabled: true + env_name: null +rl_games: + model_alias: openvla + env_eval: + image_size: 224 + frameskip: 1 + image_transform: raw_rgb + prompt_mode: latency_neutral + ghost_trail: + history_frames: 5 + gamma: 1.3 + min_alpha: 35 + scroll_px_per_step: 4.0 + ground_fraction: 0.22 + seed: 42 + fixed_episode_seeds: true + latency_seed_stride: 0 + task_seed_stride: 0 + task_description: '' + eval_parallel_envs: 5 + action_chunk_execution: + enabled: false + chunk_size: null + enabled: true + gymnasium: + make_kwargs: {} + mid_train: + enabled: false + interval_steps: 100 + latencies: + - 0 + - 2 + - 4 + num_episodes: 5 + max_steps_per_episode: 3600 + post_train: + enabled: true + latencies: + - 0 + latencies_from_prompt_map: false + num_episodes: 20 + max_steps_per_episode: 3600 + eval_backend: latency_bench + distributed_mode: none + vectorized: + enabled: false + batch_size: 1 + latency: + prompt_map_path: /home/ubuntu/talha/playground/Datasets/rl_games/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2/latency_prompt_map.json + mode: mixed + values: + - 0 + task: gymnasium + gymnasium: + task_contract: + task_name: inverted_pendulum_rgb_state + env_id: LatencyBench/InvertedPendulumDiscrete-v0 + registration_imports: + - latency_bench.envs.gymnasium_mujoco + env_fps: 25.0 + obs_fps: 25.0 + frame_stack: 1 + base_prompt: 'Keep the pole upright by applying horizontal cart forces. Choose + exactly one action from: push_left_hard, push_left_soft, zero_force, push_right_soft, + push_right_hard.' + action_labels: + - push_left_hard + - push_left_soft + - zero_force + - push_right_soft + - push_right_hard + action_values: + - 0 + - 1 + - 2 + - 3 + - 4 + noop_action_id: 2 + make_kwargs: + base_env_id: InvertedPendulum-v4 + render_mode: rgb_array + force_values: + - -3.0 + - -1.5 + - 0.0 + - 1.5 + - 3.0 + state_space: + labels: + - cart_position + - pole_angle + - cart_velocity + - pole_angular_velocity + task_name: inverted_pendulum_rgb_state + initialization_mode: scratch + action_carrier: native +model: openvla +env: gymnasium +init: scratch +mode: mixed_latency +checkpoint: + load: none + hf_repo_id: talha15032/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2 + save_best_model: false + save_final_model: true + save_pt_file: true + save_training_state: false + save_safetensors_file: false + local: + keep_last_n: 1 + sync: + enabled: true + repo_id: talha15032/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2 + keep_last_n: 0 + sync_every_n_checkpoints: 1 + resume_policy: local_latest +run_id: openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2 +output_dir: /home/ubuntu/talha/results/Checkpoints/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2 +config_yaml: null +is_debug: false +version_id: '0.21' diff --git a/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2/config.yaml b/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2/config.yaml new file mode 100644 index 0000000000000000000000000000000000000000..dd534731e35149de5e61c097c77a17b7cd2cfcd3 --- /dev/null +++ b/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2/config.yaml @@ -0,0 +1,93 @@ +checkpoint: + local: + keep_last_n: 1 + save_best_model: false + save_final_model: true + save_pt_file: true + save_safetensors_file: false + save_training_state: false + sync: + enabled: true + keep_last_n: 0 + repo_id: talha15032/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2 +datasets: + vla_data: + data_mix: openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2 + dataset_py: lerobot_datasets + eval_data_mix: openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2__val + latency_curriculum: + enabled: false + per_device_batch_size: 8 +framework: + action_model: + action_dim: 5 + action_env_dim: 5 + action_hidden_dim: 2560 + action_horizon: 1 + action_model_type: MLP + loss_type: discrete_ce + state_encoding: discretized_text + kv_memory: + enabled: false + packed_train: false + rebased_sink: true + rollout_len: 8 + window: 4 + name: QwenOFT + qwenvl: + attn_implementation: flash_attention_2 + base_vlm: /home/ubuntu/talha/playground/Pretrained_models/Qwen3-VL-4B-Instruct + enable_gradient_checkpointing: true +output_dir: /home/ubuntu/talha/results/Checkpoints/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2 +rl_games: + env_eval: + enabled: true + eval_backend: latency_bench + mid_train: + enabled: false + interval_steps: 100 + prompt_mode: latency_neutral + task: gymnasium +run_id: openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2 +run_root_dir: /home/ubuntu/talha/results/Checkpoints +seed: 42 +trainer: + distributed_backend: none + eval_action_classification: true + eval_action_classification_interval: null + eval_interval: 5000 + eval_num_batches: 200 + freeze_llm_layers: [] + freeze_modules: '' + freeze_tied_embedding: false + freeze_vit: false + gradient_accumulation_steps: 2 + gradient_clipping: 1.0 + is_resume: false + learning_rate: + action_model: 0.0001 + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + logging_frequency: 1 + lr_scheduler_type: cosine_with_min_lr + max_train_steps: 5000 + num_warmup_steps: 100 + optimizer: + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + fused: true + weight_decay: 1.0e-08 + per_latency_eval_num_batches: 5 + pretrained_checkpoint: null + profile_timing: + enabled: false + save_interval: 5000 + scheduler_specific_kwargs: + min_lr: 1.0e-06 +training_latency_condition: unspecified +wandb_entity: talha1503 +wandb_group: openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2 +wandb_project: latency-sensitive-bench +wandb_tags: [] diff --git a/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2/dataset_statistics.json b/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2/dataset_statistics.json new file mode 100644 index 0000000000000000000000000000000000000000..48310ea5d5254896f48d5e8d123c1528088202a0 --- /dev/null +++ b/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2/dataset_statistics.json @@ -0,0 +1,95 @@ +{ + "new_embodiment": { + "action": { + "mean": [ + 0.027996724471449852, + 0.1508392095565796, + 0.6481470465660095, + 0.13600370287895203, + 0.03701331838965416 + ], + "std": [ + 0.16490621864795685, + 0.3581511974334717, + 0.4778982102870941, + 0.3424498736858368, + 0.18899837136268616 + ], + "max": [ + 1.0, + 1.0, + 1.0, + 1.0, + 1.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 1.0, + 1.0, + 1.0, + 1.0, + 1.0 + ], + "mask": [ + true, + true, + true, + true, + true + ] + }, + "state": { + "mean": [ + 0.143598273396492, + 0.004461213015019894, + 0.11432920396327972, + -0.006049089599400759 + ], + "std": [ + 0.31354930996894836, + 0.06130174919962883, + 0.4503016769886017, + 0.8575727343559265 + ], + "max": [ + 1.031674861907959, + 0.19994768500328064, + 2.3079874515533447, + 3.7387092113494873 + ], + "min": [ + -1.0259008407592773, + -0.1999799758195877, + -2.2849948406219482, + -4.038050174713135 + ], + "q01": [ + -0.3022584795951843, + -0.15527626156806945, + -0.9133486723899841, + -2.359431056976318 + ], + "q99": [ + 1.0026830291748048, + 0.16135655224323286, + 1.3828536510467528, + 2.107583684921267 + ] + }, + "num_transitions": 218597, + "num_trajectories": 2070 + } +} \ No newline at end of file diff --git a/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2/dataset_statistics_eval.json b/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2/dataset_statistics_eval.json new file mode 100644 index 0000000000000000000000000000000000000000..d8b9a06ce9d7f472c01600f4e742aa2411c89047 --- /dev/null +++ b/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2/dataset_statistics_eval.json @@ -0,0 +1,95 @@ +{ + "new_embodiment": { + "action": { + "mean": [ + 0.035599857568740845, + 0.1577073633670807, + 0.6541473865509033, + 0.07760769128799438, + 0.07493770122528076 + ], + "std": [ + 0.1853053718805313, + 0.3644678294658661, + 0.4756529927253723, + 0.26755642890930176, + 0.26328903436660767 + ], + "max": [ + 1.0, + 1.0, + 1.0, + 1.0, + 1.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 1.0, + 1.0, + 1.0, + 1.0, + 1.0 + ], + "mask": [ + true, + true, + true, + true, + true + ] + }, + "state": { + "mean": [ + -0.006056365091353655, + -0.001078602741472423, + 0.021814774721860886, + -0.004301219247281551 + ], + "std": [ + 0.15048037469387054, + 0.0324825718998909, + 0.33326858282089233, + 0.713808000087738 + ], + "max": [ + 1.0273027420043945, + 0.198622465133667, + 2.3032615184783936, + 3.4846131801605225 + ], + "min": [ + -0.41009554266929626, + -0.19869747757911682, + -1.3806304931640625, + -3.9696083068847656 + ], + "q01": [ + -0.08295003287494182, + -0.1170046754181385, + -0.977174437046051, + -2.3071367263793947 + ], + "q99": [ + 0.9016301155090326, + 0.11161688491702067, + 1.2677143096923813, + 2.1670001268386816 + ] + }, + "num_transitions": 11236, + "num_trajectories": 30 + } +} \ No newline at end of file diff --git a/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2/hydra/.hydra/config.yaml b/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2/hydra/.hydra/config.yaml new file mode 100644 index 0000000000000000000000000000000000000000..5952e0b8c6b14b9de57744bd4b567af13baf8ada --- /dev/null +++ b/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2/hydra/.hydra/config.yaml @@ -0,0 +1,338 @@ +framework: + qwenvl: + base_vlm: /home/ubuntu/talha/playground/Pretrained_models/Qwen3-VL-4B-Instruct + attn_implementation: flash_attention_2 + enable_gradient_checkpointing: true + action_model: + state_dim: 4 + loss_type: discrete_ce + action_horizon: 1 + future_action_window_size: 0 + past_action_window_size: 0 + action_dim: 5 + action_env_dim: 5 + state_encoding: discretized_text + action_model_type: MLP + action_hidden_dim: 2560 + kv_memory: + enabled: false + window: 4 + rollout_len: 8 + packed_train: false + rebased_sink: true + name: QwenOFT +datasets: + vla_data: + dataset_py: lerobot_datasets + include_state: true + data_root_dir: /home/ubuntu/talha/playground/Datasets/rl_games + data_mix: openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2 + eval_data_mix: openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2__val + custom_mixtures_path: /home/ubuntu/talha/playground/Datasets/rl_games/_generated_mixtures/single_622b136e13e57595.json + action_type: discrete + sequential_step_sampling: false + eval_sequential_step_sampling: null + num_workers: 8 + eval_num_workers: 8 + prefetch_factor: 4 + persistent_workers: true + pin_memory: true + shuffle: true + action_balance: + enabled: false + strategy: balanced_epoch + action_key: action_id + target_flap_fraction: 0.3 + noop_id: 0 + flap_id: 1 + latency_curriculum: + enabled: false + strategy: exclusive + latencies: null + phase_steps: null + phase_distributions: null + new_latency_passes: 1.0 + replay_passes: 0.25 + target_total_passes: 2.0 + final_equalization: true + step_budget_mode: auto + eval_at_phase_end: false + save_at_phase_end: false + computed_plan: null + per_device_batch_size: 8 + load_all_data_for_training: true + num_obs_frames: 1 + image_mode: single + prompt_mode: latency_neutral + stitch_grid: + - 2 + - 2 + obs_image_size: null + video_backend: torchvision_av + gymnasium_task_contract: + task_name: inverted_pendulum_rgb_state + env_id: LatencyBench/InvertedPendulumDiscrete-v0 + registration_imports: + - latency_bench.envs.gymnasium_mujoco + env_fps: 25.0 + obs_fps: 25.0 + frame_stack: 1 + base_prompt: 'Keep the pole upright by applying horizontal cart forces. Choose + exactly one action from: push_left_hard, push_left_soft, zero_force, push_right_soft, + push_right_hard.' + action_labels: + - push_left_hard + - push_left_soft + - zero_force + - push_right_soft + - push_right_hard + action_values: + - 0 + - 1 + - 2 + - 3 + - 4 + noop_action_id: 2 + make_kwargs: + base_env_id: InvertedPendulum-v4 + render_mode: rgb_array + force_values: + - -3.0 + - -1.5 + - 0.0 + - 1.5 + - 3.0 + base_make_kwargs: {} + state_space: + labels: + - cart_position + - pole_angle + - cart_velocity + - pole_angular_velocity + active_action_dim: 5 +dataset: + source_hf: latency-sensitive-bench/inverted_pendulum_200ep_2 + config_name: null + source_subdir: inverted_pendulum_fixed_latency_0_100ep_1ksteps_og,inverted_pendulum_fixed_latency_2_1000ep_7k2steps,inverted_pendulum_fixed_latency_4_1000ep_7k2steps + converted_name: openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2 + single_source_hf: '' + mixed_source_hf: '' + single_converted_name: flappy_train + mixed_converted_name: flappy_train + single_latency_filter: null + mixed_latency_filter: null + force_download: false + setup_force: true + skip_verification: true + target_latency_unit: observation_steps + verify_rows: 200 + max_episodes: null + max_steps_per_episode: 1000 + episodes_per_latency: null + episodes_per_latency_by_latency: 0:90,2:1000,4:1000 + latency_filter: + - 0 + - 2 + - 4 + debug_subset: + enabled: false + max_episodes: 5 + suffix: debug +base_model: + repo_id: Qwen/Qwen3-VL-4B-Instruct +initialization: + checkpoint_local_dir: null + checkpoint_hf_repo_id: null + checkpoint_filename: null +trainer: + max_train_steps: 5000 + num_warmup_steps: 100 + save_interval: 5000 + eval_interval: 5000 + eval_num_batches: 200 + per_latency_eval_num_batches: 5 + eval_action_classification: true + eval_action_classification_interval: null + cc_f1_tolerance: 1 + learning_rate: + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + action_model: 0.0001 + lr_scheduler_type: cosine_with_min_lr + scheduler_specific_kwargs: + min_lr: 1.0e-06 + freeze_modules: '' + freeze_vit: false + freeze_tied_embedding: false + freeze_llm_layers: [] + loss_scale: + vla: 1.0 + vlm: 0.1 + max_grad_norm: 1.0 + weight_decay: 0.0 + logging_frequency: 1 + profile_timing: + enabled: false + log_interval: 10 + gradient_clipping: 1.0 + gradient_accumulation_steps: 2 + distributed_backend: none + is_resume: false + pretrained_checkpoint: null + resume_step: 0 + reload_modules: null + optimizer: + name: AdamW + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + weight_decay: 1.0e-08 + fused: true + save_format: pt +workspace_dir: /home/ubuntu/talha +run_root_dir: /home/ubuntu/talha/results/Checkpoints +seed: 42 +wandb_entity: ${oc.env:WANDB_ENTITY} +wandb_project: ${oc.env:WANDB_PROJECT,starVLA_rl_games} +wandb_group: openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2 +wandb_tags: [] +training_latency_condition: unspecified +auth: + env_file: null + hf_token_env: HF_TOKEN + wandb_api_key_env: WANDB_API_KEY +paths: + run_root_dir: results/Checkpoints + dataset_local_dir: playground/Datasets/rl_games + dataset_cache_dir: null + base_model_dir: /home/ubuntu/talha/playground/Pretrained_models/Qwen3-VL-4B-Instruct + accelerate_config: starVLA/config/deepseeds/deepspeed_zero2.yaml +launch: + use_accelerate: true + gpus: null + num_processes: 1 + dry_run: false +conda: + enabled: true + env_name: null +rl_games: + model_alias: openvla + env_eval: + image_size: 224 + frameskip: 1 + image_transform: raw_rgb + prompt_mode: latency_neutral + ghost_trail: + history_frames: 5 + gamma: 1.3 + min_alpha: 35 + scroll_px_per_step: 4.0 + ground_fraction: 0.22 + seed: 42 + fixed_episode_seeds: true + latency_seed_stride: 0 + task_seed_stride: 0 + task_description: '' + eval_parallel_envs: 5 + action_chunk_execution: + enabled: false + chunk_size: null + enabled: true + gymnasium: + make_kwargs: {} + mid_train: + enabled: false + interval_steps: 100 + latencies: + - 0 + - 2 + - 4 + num_episodes: 5 + max_steps_per_episode: 3600 + post_train: + enabled: true + latencies: + - 0 + latencies_from_prompt_map: false + num_episodes: 20 + max_steps_per_episode: 3600 + eval_backend: latency_bench + distributed_mode: none + vectorized: + enabled: false + batch_size: 1 + latency: + prompt_map_path: /home/ubuntu/talha/playground/Datasets/rl_games/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2/latency_prompt_map.json + mode: mixed + values: + - 0 + task: gymnasium + gymnasium: + task_contract: + task_name: inverted_pendulum_rgb_state + env_id: LatencyBench/InvertedPendulumDiscrete-v0 + registration_imports: + - latency_bench.envs.gymnasium_mujoco + env_fps: 25.0 + obs_fps: 25.0 + frame_stack: 1 + base_prompt: 'Keep the pole upright by applying horizontal cart forces. Choose + exactly one action from: push_left_hard, push_left_soft, zero_force, push_right_soft, + push_right_hard.' + action_labels: + - push_left_hard + - push_left_soft + - zero_force + - push_right_soft + - push_right_hard + action_values: + - 0 + - 1 + - 2 + - 3 + - 4 + noop_action_id: 2 + make_kwargs: + base_env_id: InvertedPendulum-v4 + render_mode: rgb_array + force_values: + - -3.0 + - -1.5 + - 0.0 + - 1.5 + - 3.0 + state_space: + labels: + - cart_position + - pole_angle + - cart_velocity + - pole_angular_velocity + task_name: inverted_pendulum_rgb_state + initialization_mode: scratch + action_carrier: native +model: openvla +env: gymnasium +init: scratch +mode: mixed_latency +checkpoint: + load: none + hf_repo_id: talha15032/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2 + save_best_model: false + save_final_model: true + save_pt_file: true + save_training_state: false + save_safetensors_file: false + local: + keep_last_n: 1 + sync: + enabled: true + repo_id: talha15032/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2 + keep_last_n: 0 + sync_every_n_checkpoints: 1 + resume_policy: local_latest +run_id: openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2 +output_dir: null +config_yaml: null +is_debug: false +version_id: 0.21 diff --git a/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2/hydra/.hydra/hydra.yaml b/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2/hydra/.hydra/hydra.yaml new file mode 100644 index 0000000000000000000000000000000000000000..1855661615107bfd856e62ac76d69a2fd7635303 --- /dev/null +++ b/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2/hydra/.hydra/hydra.yaml @@ -0,0 +1,394 @@ +hydra: + run: + dir: ${run_root_dir}/${run_id}/hydra + sweep: + dir: multirun/${now:%Y-%m-%d}/${now:%H-%M-%S} + subdir: ${hydra.job.num} + launcher: + _target_: hydra._internal.core_plugins.basic_launcher.BasicLauncher + sweeper: + _target_: hydra._internal.core_plugins.basic_sweeper.BasicSweeper + max_batch_size: null + params: null + help: + app_name: ${hydra.job.name} + header: '${hydra.help.app_name} is powered by Hydra. + + ' + footer: 'Powered by Hydra (https://hydra.cc) + + Use --hydra-help to view Hydra specific help + + ' + template: '${hydra.help.header} + + == Configuration groups == + + Compose your configuration from those groups (group=option) + + + $APP_CONFIG_GROUPS + + + == Config == + + Override anything in the config (foo.bar=value) + + + $CONFIG + + + ${hydra.help.footer} + + ' + hydra_help: + template: 'Hydra (${hydra.runtime.version}) + + See https://hydra.cc for more info. + + + == Flags == + + $FLAGS_HELP + + + == Configuration groups == + + Compose your configuration from those groups (For example, append hydra/job_logging=disabled + to command line) + + + $HYDRA_CONFIG_GROUPS + + + Use ''--cfg hydra'' to Show the Hydra config. + + ' + hydra_help: ??? + hydra_logging: + version: 1 + formatters: + simple: + format: '[%(asctime)s][HYDRA] %(message)s' + handlers: + console: + class: logging.StreamHandler + formatter: simple + stream: ext://sys.stdout + root: + level: INFO + handlers: + - console + loggers: + logging_example: + level: DEBUG + disable_existing_loggers: false + job_logging: + version: 1 + formatters: + simple: + format: '[%(asctime)s][%(name)s][%(levelname)s] - %(message)s' + handlers: + console: + class: logging.StreamHandler + formatter: simple + stream: ext://sys.stdout + file: + class: logging.FileHandler + formatter: simple + filename: ${hydra.runtime.output_dir}/${hydra.job.name}.log + root: + level: INFO + handlers: + - console + - file + disable_existing_loggers: false + env: {} + mode: RUN + searchpath: [] + callbacks: {} + output_subdir: .hydra + overrides: + hydra: + - hydra.mode=RUN + task: + - model=openvla + - env=gymnasium + - init=scratch + - mode=mixed_latency + - ++framework.qwenvl.base_vlm=/home/ubuntu/talha/playground/Pretrained_models/Qwen3-VL-4B-Instruct + - ++framework.qwenvl.attn_implementation=flash_attention_2 + - ++framework.qwenvl.enable_gradient_checkpointing=true + - ++framework.action_model.state_dim=4 + - ++framework.action_model.loss_type=discrete_ce + - ++framework.action_model.action_horizon=1 + - ++framework.action_model.future_action_window_size=0 + - ++framework.action_model.past_action_window_size=0 + - ++framework.action_model.action_dim=5 + - ++framework.action_model.action_env_dim=5 + - ++framework.action_model.state_encoding=discretized_text + - ++framework.action_model.action_model_type=MLP + - ++framework.action_model.action_hidden_dim=2560 + - ++framework.kv_memory.enabled=false + - ++framework.kv_memory.window=4 + - ++framework.kv_memory.rollout_len=8 + - ++framework.kv_memory.packed_train=false + - ++framework.kv_memory.rebased_sink=true + - ++framework.name=QwenOFT + - ++datasets.vla_data.dataset_py=lerobot_datasets + - ++datasets.vla_data.include_state=true + - ++datasets.vla_data.data_root_dir=playground/Datasets/rl_games + - ++datasets.vla_data.data_mix=openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2 + - ++datasets.vla_data.eval_data_mix=openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2__val + - ++datasets.vla_data.custom_mixtures_path=null + - ++datasets.vla_data.action_type=discrete + - ++datasets.vla_data.sequential_step_sampling=false + - ++datasets.vla_data.eval_sequential_step_sampling=null + - ++datasets.vla_data.num_workers=8 + - ++datasets.vla_data.eval_num_workers=8 + - ++datasets.vla_data.prefetch_factor=4 + - ++datasets.vla_data.persistent_workers=true + - ++datasets.vla_data.pin_memory=true + - ++datasets.vla_data.shuffle=true + - ++datasets.vla_data.action_balance.enabled=false + - ++datasets.vla_data.action_balance.strategy=balanced_epoch + - ++datasets.vla_data.action_balance.action_key=action_id + - ++datasets.vla_data.action_balance.target_flap_fraction=0.3 + - ++datasets.vla_data.action_balance.noop_id=0 + - ++datasets.vla_data.action_balance.flap_id=1 + - ++datasets.vla_data.latency_curriculum.enabled=false + - ++datasets.vla_data.latency_curriculum.strategy=exclusive + - ++datasets.vla_data.latency_curriculum.latencies=null + - ++datasets.vla_data.latency_curriculum.phase_steps=null + - ++datasets.vla_data.latency_curriculum.phase_distributions=null + - ++datasets.vla_data.latency_curriculum.new_latency_passes=1.0 + - ++datasets.vla_data.latency_curriculum.replay_passes=0.25 + - ++datasets.vla_data.latency_curriculum.target_total_passes=2.0 + - ++datasets.vla_data.latency_curriculum.final_equalization=true + - ++datasets.vla_data.latency_curriculum.step_budget_mode=auto + - ++datasets.vla_data.latency_curriculum.eval_at_phase_end=false + - ++datasets.vla_data.latency_curriculum.save_at_phase_end=false + - ++datasets.vla_data.latency_curriculum.computed_plan=null + - ++datasets.vla_data.per_device_batch_size=8 + - ++datasets.vla_data.load_all_data_for_training=true + - ++datasets.vla_data.num_obs_frames=1 + - ++datasets.vla_data.image_mode=single + - ++datasets.vla_data.prompt_mode=latency_neutral + - ++datasets.vla_data.stitch_grid=[2,2] + - ++datasets.vla_data.obs_image_size=null + - ++datasets.vla_data.video_backend=torchvision_av + - '++datasets.vla_data.gymnasium_task_contract={task_name:inverted_pendulum_rgb_state,env_id:LatencyBench/InvertedPendulumDiscrete-v0,registration_imports:[latency_bench.envs.gymnasium_mujoco],env_fps:25.0,obs_fps:25.0,frame_stack:1,base_prompt:''Keep + the pole upright by applying horizontal cart forces. Choose exactly one action + from: push_left_hard, push_left_soft, zero_force, push_right_soft, push_right_hard.'',action_labels:[push_left_hard,push_left_soft,zero_force,push_right_soft,push_right_hard],action_values:[0,1,2,3,4],noop_action_id:2,make_kwargs:{base_env_id:InvertedPendulum-v4,render_mode:rgb_array,force_values:[-3.0,-1.5,0.0,1.5,3.0],base_make_kwargs:{}},state_space:{labels:[cart_position,pole_angle,cart_velocity,pole_angular_velocity]}}' + - ++datasets.vla_data.active_action_dim=5 + - ++dataset.source_hf=latency-sensitive-bench/inverted_pendulum_200ep_2 + - ++dataset.config_name=null + - ++dataset.source_subdir='inverted_pendulum_fixed_latency_0_100ep_1ksteps_og,inverted_pendulum_fixed_latency_2_1000ep_7k2steps,inverted_pendulum_fixed_latency_4_1000ep_7k2steps' + - ++dataset.converted_name=openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2 + - ++dataset.single_source_hf= + - ++dataset.mixed_source_hf= + - ++dataset.single_converted_name=flappy_train + - ++dataset.mixed_converted_name=flappy_train + - ++dataset.single_latency_filter=null + - ++dataset.mixed_latency_filter=null + - ++dataset.force_download=false + - ++dataset.setup_force=true + - ++dataset.skip_verification=true + - ++dataset.target_latency_unit=observation_steps + - ++dataset.verify_rows=200 + - ++dataset.max_episodes=null + - ++dataset.max_steps_per_episode=1000 + - ++dataset.episodes_per_latency=null + - ++dataset.episodes_per_latency_by_latency='0:90,2:1000,4:1000' + - ++dataset.latency_filter=[0,2,4] + - ++dataset.debug_subset.enabled=false + - ++dataset.debug_subset.max_episodes=5 + - ++dataset.debug_subset.suffix=debug + - ++base_model.repo_id=Qwen/Qwen3-VL-4B-Instruct + - ++initialization.checkpoint_local_dir=null + - ++initialization.checkpoint_hf_repo_id=null + - ++initialization.checkpoint_filename=null + - ++trainer.max_train_steps=5000 + - ++trainer.num_warmup_steps=100 + - ++trainer.save_interval=5000 + - ++trainer.eval_interval=5000 + - ++trainer.eval_num_batches=200 + - ++trainer.per_latency_eval_num_batches=5 + - ++trainer.eval_action_classification=true + - ++trainer.eval_action_classification_interval=null + - ++trainer.cc_f1_tolerance=1 + - ++trainer.learning_rate.base=2e-05 + - ++trainer.learning_rate.qwen_vl_interface=1e-05 + - ++trainer.learning_rate.action_model=0.0001 + - ++trainer.lr_scheduler_type=cosine_with_min_lr + - ++trainer.scheduler_specific_kwargs.min_lr=1e-06 + - ++trainer.freeze_modules= + - ++trainer.freeze_vit=false + - ++trainer.freeze_tied_embedding=false + - ++trainer.freeze_llm_layers=[] + - ++trainer.loss_scale.vla=1.0 + - ++trainer.loss_scale.vlm=0.1 + - ++trainer.max_grad_norm=1.0 + - ++trainer.weight_decay=0.0 + - ++trainer.logging_frequency=1 + - ++trainer.profile_timing.enabled=false + - ++trainer.profile_timing.log_interval=10 + - ++trainer.gradient_clipping=1.0 + - ++trainer.gradient_accumulation_steps=2 + - ++trainer.distributed_backend=none + - ++trainer.is_resume=false + - ++trainer.pretrained_checkpoint=null + - ++trainer.resume_step=0 + - ++trainer.reload_modules=null + - ++trainer.optimizer.name=AdamW + - ++trainer.optimizer.betas=[0.9,0.95] + - ++trainer.optimizer.eps=1e-08 + - ++trainer.optimizer.weight_decay=1e-08 + - ++trainer.optimizer.fused=true + - ++trainer.save_format=pt + - ++workspace_dir=/home/ubuntu/talha + - ++run_root_dir=results/Checkpoints + - ++seed=42 + - ++wandb_group=openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2 + - ++wandb_tags=[] + - ++training_latency_condition=unspecified + - ++auth.env_file=null + - ++auth.hf_token_env=HF_TOKEN + - ++auth.wandb_api_key_env=WANDB_API_KEY + - ++paths.run_root_dir=results/Checkpoints + - ++paths.dataset_local_dir=playground/Datasets/rl_games + - ++paths.dataset_cache_dir=null + - ++paths.base_model_dir=/home/ubuntu/talha/playground/Pretrained_models/Qwen3-VL-4B-Instruct + - ++paths.accelerate_config=starVLA/config/deepseeds/deepspeed_zero2.yaml + - ++rl_games.model_alias=openvla + - ++rl_games.env_eval.image_size=224 + - ++rl_games.env_eval.frameskip=1 + - ++rl_games.env_eval.image_transform=raw_rgb + - ++rl_games.env_eval.prompt_mode=latency_neutral + - ++rl_games.env_eval.ghost_trail.history_frames=5 + - ++rl_games.env_eval.ghost_trail.gamma=1.3 + - ++rl_games.env_eval.ghost_trail.min_alpha=35 + - ++rl_games.env_eval.ghost_trail.scroll_px_per_step=4.0 + - ++rl_games.env_eval.ghost_trail.ground_fraction=0.22 + - ++rl_games.env_eval.seed=42 + - ++rl_games.env_eval.fixed_episode_seeds=true + - ++rl_games.env_eval.latency_seed_stride=0 + - ++rl_games.env_eval.task_seed_stride=0 + - ++rl_games.env_eval.task_description= + - ++rl_games.env_eval.eval_parallel_envs=5 + - ++rl_games.env_eval.action_chunk_execution.enabled=false + - ++rl_games.env_eval.action_chunk_execution.chunk_size=null + - ++rl_games.env_eval.enabled=true + - ++rl_games.env_eval.mid_train.enabled=false + - ++rl_games.env_eval.mid_train.interval_steps=100 + - ++rl_games.env_eval.mid_train.latencies=[0,2,4] + - ++rl_games.env_eval.mid_train.num_episodes=5 + - ++rl_games.env_eval.mid_train.max_steps_per_episode=3600 + - ++rl_games.env_eval.post_train.enabled=true + - ++rl_games.env_eval.post_train.latencies=[0] + - ++rl_games.env_eval.post_train.latencies_from_prompt_map=false + - ++rl_games.env_eval.post_train.num_episodes=20 + - ++rl_games.env_eval.post_train.max_steps_per_episode=3600 + - ++rl_games.env_eval.eval_backend=latency_bench + - ++rl_games.env_eval.distributed_mode=none + - ++rl_games.env_eval.vectorized.enabled=false + - ++rl_games.env_eval.vectorized.batch_size=1 + - ++rl_games.env_eval.latency.prompt_map_path=null + - ++rl_games.env_eval.latency.mode=mixed + - ++rl_games.env_eval.latency.values=[0] + - ++rl_games.task=gymnasium + - ++rl_games.gymnasium.task_contract.task_name=inverted_pendulum_rgb_state + - ++rl_games.gymnasium.task_contract.env_id=LatencyBench/InvertedPendulumDiscrete-v0 + - ++rl_games.gymnasium.task_contract.registration_imports=[latency_bench.envs.gymnasium_mujoco] + - ++rl_games.gymnasium.task_contract.env_fps=25.0 + - ++rl_games.gymnasium.task_contract.obs_fps=25.0 + - ++rl_games.gymnasium.task_contract.frame_stack=1 + - '++rl_games.gymnasium.task_contract.base_prompt=''Keep the pole upright by applying + horizontal cart forces. Choose exactly one action from: push_left_hard, push_left_soft, + zero_force, push_right_soft, push_right_hard.''' + - ++rl_games.gymnasium.task_contract.action_labels=[push_left_hard,push_left_soft,zero_force,push_right_soft,push_right_hard] + - ++rl_games.gymnasium.task_contract.action_values=[0,1,2,3,4] + - ++rl_games.gymnasium.task_contract.noop_action_id=2 + - ++rl_games.gymnasium.task_contract.make_kwargs.base_env_id=InvertedPendulum-v4 + - ++rl_games.gymnasium.task_contract.make_kwargs.render_mode=rgb_array + - ++rl_games.gymnasium.task_contract.make_kwargs.force_values=[-3.0,-1.5,0.0,1.5,3.0] + - ++rl_games.gymnasium.task_contract.state_space.labels=[cart_position,pole_angle,cart_velocity,pole_angular_velocity] + - ++rl_games.gymnasium.task_name=inverted_pendulum_rgb_state + - ++rl_games.initialization_mode=scratch + - ++rl_games.action_carrier=native + - ++checkpoint.load=none + - ++checkpoint.hf_repo_id=talha15032/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2 + - ++checkpoint.save_best_model=false + - ++checkpoint.save_final_model=true + - ++checkpoint.save_pt_file=true + - ++checkpoint.save_training_state=false + - ++checkpoint.save_safetensors_file=false + - ++checkpoint.local.keep_last_n=1 + - ++checkpoint.sync.enabled=true + - ++checkpoint.sync.repo_id=talha15032/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2 + - ++checkpoint.sync.keep_last_n=0 + - ++checkpoint.sync.sync_every_n_checkpoints=1 + - ++checkpoint.sync.resume_policy=local_latest + - ++run_id=openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2 + - ++output_dir=null + - ++config_yaml=null + - ++is_debug=false + - ++version_id=0.21 + - ++run_root_dir=/home/ubuntu/talha/results/Checkpoints + - ++trainer.is_resume=false + - ++datasets.vla_data.data_root_dir=/home/ubuntu/talha/playground/Datasets/rl_games + - ++datasets.vla_data.data_mix=openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2 + - ++datasets.vla_data.eval_data_mix=openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2__val + - ++datasets.vla_data.custom_mixtures_path=/home/ubuntu/talha/playground/Datasets/rl_games/_generated_mixtures/single_622b136e13e57595.json + - ++framework.qwenvl.base_vlm=/home/ubuntu/talha/playground/Pretrained_models/Qwen3-VL-4B-Instruct + - ++rl_games.env_eval.latency.prompt_map_path=/home/ubuntu/talha/playground/Datasets/rl_games/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2/latency_prompt_map.json + job: + name: train_starvla_hydra + chdir: false + override_dirname: '++auth.env_file=null,++auth.hf_token_env=HF_TOKEN,++auth.wandb_api_key_env=WANDB_API_KEY,++base_model.repo_id=Qwen/Qwen3-VL-4B-Instruct,++checkpoint.hf_repo_id=talha15032/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2,++checkpoint.load=none,++checkpoint.local.keep_last_n=1,++checkpoint.save_best_model=false,++checkpoint.save_final_model=true,++checkpoint.save_pt_file=true,++checkpoint.save_safetensors_file=false,++checkpoint.save_training_state=false,++checkpoint.sync.enabled=true,++checkpoint.sync.keep_last_n=0,++checkpoint.sync.repo_id=talha15032/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2,++checkpoint.sync.resume_policy=local_latest,++checkpoint.sync.sync_every_n_checkpoints=1,++config_yaml=null,++dataset.config_name=null,++dataset.converted_name=openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2,++dataset.debug_subset.enabled=false,++dataset.debug_subset.max_episodes=5,++dataset.debug_subset.suffix=debug,++dataset.episodes_per_latency=null,++dataset.episodes_per_latency_by_latency=''0:90,2:1000,4:1000'',++dataset.force_download=false,++dataset.latency_filter=[0,2,4],++dataset.max_episodes=null,++dataset.max_steps_per_episode=1000,++dataset.mixed_converted_name=flappy_train,++dataset.mixed_latency_filter=null,++dataset.mixed_source_hf=,++dataset.setup_force=true,++dataset.single_converted_name=flappy_train,++dataset.single_latency_filter=null,++dataset.single_source_hf=,++dataset.skip_verification=true,++dataset.source_hf=latency-sensitive-bench/inverted_pendulum_200ep_2,++dataset.source_subdir=''inverted_pendulum_fixed_latency_0_100ep_1ksteps_og,inverted_pendulum_fixed_latency_2_1000ep_7k2steps,inverted_pendulum_fixed_latency_4_1000ep_7k2steps'',++dataset.target_latency_unit=observation_steps,++dataset.verify_rows=200,++datasets.vla_data.action_balance.action_key=action_id,++datasets.vla_data.action_balance.enabled=false,++datasets.vla_data.action_balance.flap_id=1,++datasets.vla_data.action_balance.noop_id=0,++datasets.vla_data.action_balance.strategy=balanced_epoch,++datasets.vla_data.action_balance.target_flap_fraction=0.3,++datasets.vla_data.action_type=discrete,++datasets.vla_data.active_action_dim=5,++datasets.vla_data.custom_mixtures_path=/home/ubuntu/talha/playground/Datasets/rl_games/_generated_mixtures/single_622b136e13e57595.json,++datasets.vla_data.custom_mixtures_path=null,++datasets.vla_data.data_mix=openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2,++datasets.vla_data.data_mix=openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2,++datasets.vla_data.data_root_dir=/home/ubuntu/talha/playground/Datasets/rl_games,++datasets.vla_data.data_root_dir=playground/Datasets/rl_games,++datasets.vla_data.dataset_py=lerobot_datasets,++datasets.vla_data.eval_data_mix=openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2__val,++datasets.vla_data.eval_data_mix=openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2__val,++datasets.vla_data.eval_num_workers=8,++datasets.vla_data.eval_sequential_step_sampling=null,++datasets.vla_data.gymnasium_task_contract={task_name:inverted_pendulum_rgb_state,env_id:LatencyBench/InvertedPendulumDiscrete-v0,registration_imports:[latency_bench.envs.gymnasium_mujoco],env_fps:25.0,obs_fps:25.0,frame_stack:1,base_prompt:''Keep + the pole upright by applying horizontal cart forces. Choose exactly one action + from: push_left_hard, push_left_soft, zero_force, push_right_soft, push_right_hard.'',action_labels:[push_left_hard,push_left_soft,zero_force,push_right_soft,push_right_hard],action_values:[0,1,2,3,4],noop_action_id:2,make_kwargs:{base_env_id:InvertedPendulum-v4,render_mode:rgb_array,force_values:[-3.0,-1.5,0.0,1.5,3.0],base_make_kwargs:{}},state_space:{labels:[cart_position,pole_angle,cart_velocity,pole_angular_velocity]}},++datasets.vla_data.image_mode=single,++datasets.vla_data.include_state=true,++datasets.vla_data.latency_curriculum.computed_plan=null,++datasets.vla_data.latency_curriculum.enabled=false,++datasets.vla_data.latency_curriculum.eval_at_phase_end=false,++datasets.vla_data.latency_curriculum.final_equalization=true,++datasets.vla_data.latency_curriculum.latencies=null,++datasets.vla_data.latency_curriculum.new_latency_passes=1.0,++datasets.vla_data.latency_curriculum.phase_distributions=null,++datasets.vla_data.latency_curriculum.phase_steps=null,++datasets.vla_data.latency_curriculum.replay_passes=0.25,++datasets.vla_data.latency_curriculum.save_at_phase_end=false,++datasets.vla_data.latency_curriculum.step_budget_mode=auto,++datasets.vla_data.latency_curriculum.strategy=exclusive,++datasets.vla_data.latency_curriculum.target_total_passes=2.0,++datasets.vla_data.load_all_data_for_training=true,++datasets.vla_data.num_obs_frames=1,++datasets.vla_data.num_workers=8,++datasets.vla_data.obs_image_size=null,++datasets.vla_data.per_device_batch_size=8,++datasets.vla_data.persistent_workers=true,++datasets.vla_data.pin_memory=true,++datasets.vla_data.prefetch_factor=4,++datasets.vla_data.prompt_mode=latency_neutral,++datasets.vla_data.sequential_step_sampling=false,++datasets.vla_data.shuffle=true,++datasets.vla_data.stitch_grid=[2,2],++datasets.vla_data.video_backend=torchvision_av,++framework.action_model.action_dim=5,++framework.action_model.action_env_dim=5,++framework.action_model.action_hidden_dim=2560,++framework.action_model.action_horizon=1,++framework.action_model.action_model_type=MLP,++framework.action_model.future_action_window_size=0,++framework.action_model.loss_type=discrete_ce,++framework.action_model.past_action_window_size=0,++framework.action_model.state_dim=4,++framework.action_model.state_encoding=discretized_text,++framework.kv_memory.enabled=false,++framework.kv_memory.packed_train=false,++framework.kv_memory.rebased_sink=true,++framework.kv_memory.rollout_len=8,++framework.kv_memory.window=4,++framework.name=QwenOFT,++framework.qwenvl.attn_implementation=flash_attention_2,++framework.qwenvl.base_vlm=/home/ubuntu/talha/playground/Pretrained_models/Qwen3-VL-4B-Instruct,++framework.qwenvl.base_vlm=/home/ubuntu/talha/playground/Pretrained_models/Qwen3-VL-4B-Instruct,++framework.qwenvl.enable_gradient_checkpointing=true,++initialization.checkpoint_filename=null,++initialization.checkpoint_hf_repo_id=null,++initialization.checkpoint_local_dir=null,++is_debug=false,++output_dir=null,++paths.accelerate_config=starVLA/config/deepseeds/deepspeed_zero2.yaml,++paths.base_model_dir=/home/ubuntu/talha/playground/Pretrained_models/Qwen3-VL-4B-Instruct,++paths.dataset_cache_dir=null,++paths.dataset_local_dir=playground/Datasets/rl_games,++paths.run_root_dir=results/Checkpoints,++rl_games.action_carrier=native,++rl_games.env_eval.action_chunk_execution.chunk_size=null,++rl_games.env_eval.action_chunk_execution.enabled=false,++rl_games.env_eval.distributed_mode=none,++rl_games.env_eval.enabled=true,++rl_games.env_eval.eval_backend=latency_bench,++rl_games.env_eval.eval_parallel_envs=5,++rl_games.env_eval.fixed_episode_seeds=true,++rl_games.env_eval.frameskip=1,++rl_games.env_eval.ghost_trail.gamma=1.3,++rl_games.env_eval.ghost_trail.ground_fraction=0.22,++rl_games.env_eval.ghost_trail.history_frames=5,++rl_games.env_eval.ghost_trail.min_alpha=35,++rl_games.env_eval.ghost_trail.scroll_px_per_step=4.0,++rl_games.env_eval.image_size=224,++rl_games.env_eval.image_transform=raw_rgb,++rl_games.env_eval.latency.mode=mixed,++rl_games.env_eval.latency.prompt_map_path=/home/ubuntu/talha/playground/Datasets/rl_games/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2/latency_prompt_map.json,++rl_games.env_eval.latency.prompt_map_path=null,++rl_games.env_eval.latency.values=[0],++rl_games.env_eval.latency_seed_stride=0,++rl_games.env_eval.mid_train.enabled=false,++rl_games.env_eval.mid_train.interval_steps=100,++rl_games.env_eval.mid_train.latencies=[0,2,4],++rl_games.env_eval.mid_train.max_steps_per_episode=3600,++rl_games.env_eval.mid_train.num_episodes=5,++rl_games.env_eval.post_train.enabled=true,++rl_games.env_eval.post_train.latencies=[0],++rl_games.env_eval.post_train.latencies_from_prompt_map=false,++rl_games.env_eval.post_train.max_steps_per_episode=3600,++rl_games.env_eval.post_train.num_episodes=20,++rl_games.env_eval.prompt_mode=latency_neutral,++rl_games.env_eval.seed=42,++rl_games.env_eval.task_description=,++rl_games.env_eval.task_seed_stride=0,++rl_games.env_eval.vectorized.batch_size=1,++rl_games.env_eval.vectorized.enabled=false,++rl_games.gymnasium.task_contract.action_labels=[push_left_hard,push_left_soft,zero_force,push_right_soft,push_right_hard],++rl_games.gymnasium.task_contract.action_values=[0,1,2,3,4],++rl_games.gymnasium.task_contract.base_prompt=''Keep + the pole upright by applying horizontal cart forces. Choose exactly one action + from: push_left_hard, push_left_soft, zero_force, push_right_soft, push_right_hard.'',++rl_games.gymnasium.task_contract.env_fps=25.0,++rl_games.gymnasium.task_contract.env_id=LatencyBench/InvertedPendulumDiscrete-v0,++rl_games.gymnasium.task_contract.frame_stack=1,++rl_games.gymnasium.task_contract.make_kwargs.base_env_id=InvertedPendulum-v4,++rl_games.gymnasium.task_contract.make_kwargs.force_values=[-3.0,-1.5,0.0,1.5,3.0],++rl_games.gymnasium.task_contract.make_kwargs.render_mode=rgb_array,++rl_games.gymnasium.task_contract.noop_action_id=2,++rl_games.gymnasium.task_contract.obs_fps=25.0,++rl_games.gymnasium.task_contract.registration_imports=[latency_bench.envs.gymnasium_mujoco],++rl_games.gymnasium.task_contract.state_space.labels=[cart_position,pole_angle,cart_velocity,pole_angular_velocity],++rl_games.gymnasium.task_contract.task_name=inverted_pendulum_rgb_state,++rl_games.gymnasium.task_name=inverted_pendulum_rgb_state,++rl_games.initialization_mode=scratch,++rl_games.model_alias=openvla,++rl_games.task=gymnasium,++run_id=openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2,++run_root_dir=/home/ubuntu/talha/results/Checkpoints,++run_root_dir=results/Checkpoints,++seed=42,++trainer.cc_f1_tolerance=1,++trainer.distributed_backend=none,++trainer.eval_action_classification=true,++trainer.eval_action_classification_interval=null,++trainer.eval_interval=5000,++trainer.eval_num_batches=200,++trainer.freeze_llm_layers=[],++trainer.freeze_modules=,++trainer.freeze_tied_embedding=false,++trainer.freeze_vit=false,++trainer.gradient_accumulation_steps=2,++trainer.gradient_clipping=1.0,++trainer.is_resume=false,++trainer.is_resume=false,++trainer.learning_rate.action_model=0.0001,++trainer.learning_rate.base=2e-05,++trainer.learning_rate.qwen_vl_interface=1e-05,++trainer.logging_frequency=1,++trainer.loss_scale.vla=1.0,++trainer.loss_scale.vlm=0.1,++trainer.lr_scheduler_type=cosine_with_min_lr,++trainer.max_grad_norm=1.0,++trainer.max_train_steps=5000,++trainer.num_warmup_steps=100,++trainer.optimizer.betas=[0.9,0.95],++trainer.optimizer.eps=1e-08,++trainer.optimizer.fused=true,++trainer.optimizer.name=AdamW,++trainer.optimizer.weight_decay=1e-08,++trainer.per_latency_eval_num_batches=5,++trainer.pretrained_checkpoint=null,++trainer.profile_timing.enabled=false,++trainer.profile_timing.log_interval=10,++trainer.reload_modules=null,++trainer.resume_step=0,++trainer.save_format=pt,++trainer.save_interval=5000,++trainer.scheduler_specific_kwargs.min_lr=1e-06,++trainer.weight_decay=0.0,++training_latency_condition=unspecified,++version_id=0.21,++wandb_group=openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2,++wandb_tags=[],++workspace_dir=/home/ubuntu/talha,env=gymnasium,init=scratch,mode=mixed_latency,model=openvla' + id: ??? + num: ??? + config_name: train + env_set: {} + env_copy: [] + config: + override_dirname: + kv_sep: '=' + item_sep: ',' + exclude_keys: [] + runtime: + version: 1.3.7 + version_base: '1.1' + cwd: /home/ubuntu/talha/starVLA + config_sources: + - path: hydra.conf + schema: pkg + provider: hydra + - path: /home/ubuntu/talha/starVLA/examples/rl_games/config + schema: file + provider: main + - path: '' + schema: structured + provider: schema + output_dir: /home/ubuntu/talha/results/Checkpoints/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2/hydra + choices: + cross_task_setup: null + checkpoint: default + mode: mixed_latency + init: scratch + env: gymnasium + model: openvla + hydra/env: default + hydra/callbacks: null + hydra/job_logging: default + hydra/hydra_logging: default + hydra/hydra_help: default + hydra/help: default + hydra/sweeper: basic + hydra/launcher: basic + hydra/output: default + verbose: false diff --git a/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2/hydra/.hydra/overrides.yaml b/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2/hydra/.hydra/overrides.yaml new file mode 100644 index 0000000000000000000000000000000000000000..9cd4a2abae349c78f1b54d430202930e9ecb5c47 --- /dev/null +++ b/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2/hydra/.hydra/overrides.yaml @@ -0,0 +1,230 @@ +- model=openvla +- env=gymnasium +- init=scratch +- mode=mixed_latency +- ++framework.qwenvl.base_vlm=/home/ubuntu/talha/playground/Pretrained_models/Qwen3-VL-4B-Instruct +- ++framework.qwenvl.attn_implementation=flash_attention_2 +- ++framework.qwenvl.enable_gradient_checkpointing=true +- ++framework.action_model.state_dim=4 +- ++framework.action_model.loss_type=discrete_ce +- ++framework.action_model.action_horizon=1 +- ++framework.action_model.future_action_window_size=0 +- ++framework.action_model.past_action_window_size=0 +- ++framework.action_model.action_dim=5 +- ++framework.action_model.action_env_dim=5 +- ++framework.action_model.state_encoding=discretized_text +- ++framework.action_model.action_model_type=MLP +- ++framework.action_model.action_hidden_dim=2560 +- ++framework.kv_memory.enabled=false +- ++framework.kv_memory.window=4 +- ++framework.kv_memory.rollout_len=8 +- ++framework.kv_memory.packed_train=false +- ++framework.kv_memory.rebased_sink=true +- ++framework.name=QwenOFT +- ++datasets.vla_data.dataset_py=lerobot_datasets +- ++datasets.vla_data.include_state=true +- ++datasets.vla_data.data_root_dir=playground/Datasets/rl_games +- ++datasets.vla_data.data_mix=openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2 +- ++datasets.vla_data.eval_data_mix=openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2__val +- ++datasets.vla_data.custom_mixtures_path=null +- ++datasets.vla_data.action_type=discrete +- ++datasets.vla_data.sequential_step_sampling=false +- ++datasets.vla_data.eval_sequential_step_sampling=null +- ++datasets.vla_data.num_workers=8 +- ++datasets.vla_data.eval_num_workers=8 +- ++datasets.vla_data.prefetch_factor=4 +- ++datasets.vla_data.persistent_workers=true +- ++datasets.vla_data.pin_memory=true +- ++datasets.vla_data.shuffle=true +- ++datasets.vla_data.action_balance.enabled=false +- ++datasets.vla_data.action_balance.strategy=balanced_epoch +- ++datasets.vla_data.action_balance.action_key=action_id +- ++datasets.vla_data.action_balance.target_flap_fraction=0.3 +- ++datasets.vla_data.action_balance.noop_id=0 +- ++datasets.vla_data.action_balance.flap_id=1 +- ++datasets.vla_data.latency_curriculum.enabled=false +- ++datasets.vla_data.latency_curriculum.strategy=exclusive +- ++datasets.vla_data.latency_curriculum.latencies=null +- ++datasets.vla_data.latency_curriculum.phase_steps=null +- ++datasets.vla_data.latency_curriculum.phase_distributions=null +- ++datasets.vla_data.latency_curriculum.new_latency_passes=1.0 +- ++datasets.vla_data.latency_curriculum.replay_passes=0.25 +- ++datasets.vla_data.latency_curriculum.target_total_passes=2.0 +- ++datasets.vla_data.latency_curriculum.final_equalization=true +- ++datasets.vla_data.latency_curriculum.step_budget_mode=auto +- ++datasets.vla_data.latency_curriculum.eval_at_phase_end=false +- ++datasets.vla_data.latency_curriculum.save_at_phase_end=false +- ++datasets.vla_data.latency_curriculum.computed_plan=null +- ++datasets.vla_data.per_device_batch_size=8 +- ++datasets.vla_data.load_all_data_for_training=true +- ++datasets.vla_data.num_obs_frames=1 +- ++datasets.vla_data.image_mode=single +- ++datasets.vla_data.prompt_mode=latency_neutral +- ++datasets.vla_data.stitch_grid=[2,2] +- ++datasets.vla_data.obs_image_size=null +- ++datasets.vla_data.video_backend=torchvision_av +- '++datasets.vla_data.gymnasium_task_contract={task_name:inverted_pendulum_rgb_state,env_id:LatencyBench/InvertedPendulumDiscrete-v0,registration_imports:[latency_bench.envs.gymnasium_mujoco],env_fps:25.0,obs_fps:25.0,frame_stack:1,base_prompt:''Keep + the pole upright by applying horizontal cart forces. Choose exactly one action from: + push_left_hard, push_left_soft, zero_force, push_right_soft, push_right_hard.'',action_labels:[push_left_hard,push_left_soft,zero_force,push_right_soft,push_right_hard],action_values:[0,1,2,3,4],noop_action_id:2,make_kwargs:{base_env_id:InvertedPendulum-v4,render_mode:rgb_array,force_values:[-3.0,-1.5,0.0,1.5,3.0],base_make_kwargs:{}},state_space:{labels:[cart_position,pole_angle,cart_velocity,pole_angular_velocity]}}' +- ++datasets.vla_data.active_action_dim=5 +- ++dataset.source_hf=latency-sensitive-bench/inverted_pendulum_200ep_2 +- ++dataset.config_name=null +- ++dataset.source_subdir='inverted_pendulum_fixed_latency_0_100ep_1ksteps_og,inverted_pendulum_fixed_latency_2_1000ep_7k2steps,inverted_pendulum_fixed_latency_4_1000ep_7k2steps' +- ++dataset.converted_name=openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2 +- ++dataset.single_source_hf= +- ++dataset.mixed_source_hf= +- ++dataset.single_converted_name=flappy_train +- ++dataset.mixed_converted_name=flappy_train +- ++dataset.single_latency_filter=null +- ++dataset.mixed_latency_filter=null +- ++dataset.force_download=false +- ++dataset.setup_force=true +- ++dataset.skip_verification=true +- ++dataset.target_latency_unit=observation_steps +- ++dataset.verify_rows=200 +- ++dataset.max_episodes=null +- ++dataset.max_steps_per_episode=1000 +- ++dataset.episodes_per_latency=null +- ++dataset.episodes_per_latency_by_latency='0:90,2:1000,4:1000' +- ++dataset.latency_filter=[0,2,4] +- ++dataset.debug_subset.enabled=false +- ++dataset.debug_subset.max_episodes=5 +- ++dataset.debug_subset.suffix=debug +- ++base_model.repo_id=Qwen/Qwen3-VL-4B-Instruct +- ++initialization.checkpoint_local_dir=null +- ++initialization.checkpoint_hf_repo_id=null +- ++initialization.checkpoint_filename=null +- ++trainer.max_train_steps=5000 +- ++trainer.num_warmup_steps=100 +- ++trainer.save_interval=5000 +- ++trainer.eval_interval=5000 +- ++trainer.eval_num_batches=200 +- ++trainer.per_latency_eval_num_batches=5 +- ++trainer.eval_action_classification=true +- ++trainer.eval_action_classification_interval=null +- ++trainer.cc_f1_tolerance=1 +- ++trainer.learning_rate.base=2e-05 +- ++trainer.learning_rate.qwen_vl_interface=1e-05 +- ++trainer.learning_rate.action_model=0.0001 +- ++trainer.lr_scheduler_type=cosine_with_min_lr +- ++trainer.scheduler_specific_kwargs.min_lr=1e-06 +- ++trainer.freeze_modules= +- ++trainer.freeze_vit=false +- ++trainer.freeze_tied_embedding=false +- ++trainer.freeze_llm_layers=[] +- ++trainer.loss_scale.vla=1.0 +- ++trainer.loss_scale.vlm=0.1 +- ++trainer.max_grad_norm=1.0 +- ++trainer.weight_decay=0.0 +- ++trainer.logging_frequency=1 +- ++trainer.profile_timing.enabled=false +- ++trainer.profile_timing.log_interval=10 +- ++trainer.gradient_clipping=1.0 +- ++trainer.gradient_accumulation_steps=2 +- ++trainer.distributed_backend=none +- ++trainer.is_resume=false +- ++trainer.pretrained_checkpoint=null +- ++trainer.resume_step=0 +- ++trainer.reload_modules=null +- ++trainer.optimizer.name=AdamW +- ++trainer.optimizer.betas=[0.9,0.95] +- ++trainer.optimizer.eps=1e-08 +- ++trainer.optimizer.weight_decay=1e-08 +- ++trainer.optimizer.fused=true +- ++trainer.save_format=pt +- ++workspace_dir=/home/ubuntu/talha +- ++run_root_dir=results/Checkpoints +- ++seed=42 +- ++wandb_group=openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2 +- ++wandb_tags=[] +- ++training_latency_condition=unspecified +- ++auth.env_file=null +- ++auth.hf_token_env=HF_TOKEN +- ++auth.wandb_api_key_env=WANDB_API_KEY +- ++paths.run_root_dir=results/Checkpoints +- ++paths.dataset_local_dir=playground/Datasets/rl_games +- ++paths.dataset_cache_dir=null +- ++paths.base_model_dir=/home/ubuntu/talha/playground/Pretrained_models/Qwen3-VL-4B-Instruct +- ++paths.accelerate_config=starVLA/config/deepseeds/deepspeed_zero2.yaml +- ++rl_games.model_alias=openvla +- ++rl_games.env_eval.image_size=224 +- ++rl_games.env_eval.frameskip=1 +- ++rl_games.env_eval.image_transform=raw_rgb +- ++rl_games.env_eval.prompt_mode=latency_neutral +- ++rl_games.env_eval.ghost_trail.history_frames=5 +- ++rl_games.env_eval.ghost_trail.gamma=1.3 +- ++rl_games.env_eval.ghost_trail.min_alpha=35 +- ++rl_games.env_eval.ghost_trail.scroll_px_per_step=4.0 +- ++rl_games.env_eval.ghost_trail.ground_fraction=0.22 +- ++rl_games.env_eval.seed=42 +- ++rl_games.env_eval.fixed_episode_seeds=true +- ++rl_games.env_eval.latency_seed_stride=0 +- ++rl_games.env_eval.task_seed_stride=0 +- ++rl_games.env_eval.task_description= +- ++rl_games.env_eval.eval_parallel_envs=5 +- ++rl_games.env_eval.action_chunk_execution.enabled=false +- ++rl_games.env_eval.action_chunk_execution.chunk_size=null +- ++rl_games.env_eval.enabled=true +- ++rl_games.env_eval.mid_train.enabled=false +- ++rl_games.env_eval.mid_train.interval_steps=100 +- ++rl_games.env_eval.mid_train.latencies=[0,2,4] +- ++rl_games.env_eval.mid_train.num_episodes=5 +- ++rl_games.env_eval.mid_train.max_steps_per_episode=3600 +- ++rl_games.env_eval.post_train.enabled=true +- ++rl_games.env_eval.post_train.latencies=[0] +- ++rl_games.env_eval.post_train.latencies_from_prompt_map=false +- ++rl_games.env_eval.post_train.num_episodes=20 +- ++rl_games.env_eval.post_train.max_steps_per_episode=3600 +- ++rl_games.env_eval.eval_backend=latency_bench +- ++rl_games.env_eval.distributed_mode=none +- ++rl_games.env_eval.vectorized.enabled=false +- ++rl_games.env_eval.vectorized.batch_size=1 +- ++rl_games.env_eval.latency.prompt_map_path=null +- ++rl_games.env_eval.latency.mode=mixed +- ++rl_games.env_eval.latency.values=[0] +- ++rl_games.task=gymnasium +- ++rl_games.gymnasium.task_contract.task_name=inverted_pendulum_rgb_state +- ++rl_games.gymnasium.task_contract.env_id=LatencyBench/InvertedPendulumDiscrete-v0 +- ++rl_games.gymnasium.task_contract.registration_imports=[latency_bench.envs.gymnasium_mujoco] +- ++rl_games.gymnasium.task_contract.env_fps=25.0 +- ++rl_games.gymnasium.task_contract.obs_fps=25.0 +- ++rl_games.gymnasium.task_contract.frame_stack=1 +- '++rl_games.gymnasium.task_contract.base_prompt=''Keep the pole upright by applying + horizontal cart forces. Choose exactly one action from: push_left_hard, push_left_soft, + zero_force, push_right_soft, push_right_hard.''' +- ++rl_games.gymnasium.task_contract.action_labels=[push_left_hard,push_left_soft,zero_force,push_right_soft,push_right_hard] +- ++rl_games.gymnasium.task_contract.action_values=[0,1,2,3,4] +- ++rl_games.gymnasium.task_contract.noop_action_id=2 +- ++rl_games.gymnasium.task_contract.make_kwargs.base_env_id=InvertedPendulum-v4 +- ++rl_games.gymnasium.task_contract.make_kwargs.render_mode=rgb_array +- ++rl_games.gymnasium.task_contract.make_kwargs.force_values=[-3.0,-1.5,0.0,1.5,3.0] +- ++rl_games.gymnasium.task_contract.state_space.labels=[cart_position,pole_angle,cart_velocity,pole_angular_velocity] +- ++rl_games.gymnasium.task_name=inverted_pendulum_rgb_state +- ++rl_games.initialization_mode=scratch +- ++rl_games.action_carrier=native +- ++checkpoint.load=none +- ++checkpoint.hf_repo_id=talha15032/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2 +- ++checkpoint.save_best_model=false +- ++checkpoint.save_final_model=true +- ++checkpoint.save_pt_file=true +- ++checkpoint.save_training_state=false +- ++checkpoint.save_safetensors_file=false +- ++checkpoint.local.keep_last_n=1 +- ++checkpoint.sync.enabled=true +- ++checkpoint.sync.repo_id=talha15032/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2 +- ++checkpoint.sync.keep_last_n=0 +- ++checkpoint.sync.sync_every_n_checkpoints=1 +- ++checkpoint.sync.resume_policy=local_latest +- ++run_id=openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2 +- ++output_dir=null +- ++config_yaml=null +- ++is_debug=false +- ++version_id=0.21 +- ++run_root_dir=/home/ubuntu/talha/results/Checkpoints +- ++trainer.is_resume=false +- ++datasets.vla_data.data_root_dir=/home/ubuntu/talha/playground/Datasets/rl_games +- ++datasets.vla_data.data_mix=openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2 +- ++datasets.vla_data.eval_data_mix=openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2__val +- ++datasets.vla_data.custom_mixtures_path=/home/ubuntu/talha/playground/Datasets/rl_games/_generated_mixtures/single_622b136e13e57595.json +- ++framework.qwenvl.base_vlm=/home/ubuntu/talha/playground/Pretrained_models/Qwen3-VL-4B-Instruct +- ++rl_games.env_eval.latency.prompt_map_path=/home/ubuntu/talha/playground/Datasets/rl_games/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2/latency_prompt_map.json diff --git a/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2/hydra/train_starvla_hydra.log b/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2/hydra/train_starvla_hydra.log new file mode 100644 index 0000000000000000000000000000000000000000..365b61e9dd63c48afa427c633f2a7ecbf4147284 --- /dev/null +++ b/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2/hydra/train_starvla_hydra.log @@ -0,0 +1,5071 @@ +[2026-10-03 14:37:16,986][starVLA.model.framework.share_tools][INFO] - [*] [apply_config_compat] normalised config from version_id=0.21 to '0.21' +[2026-10-03 14:37:16,986][starVLA.training.train_starvla][INFO] - VLA Training :: Warming Up +[2026-10-03 14:37:17,060][huggingface_hub._login][WARNING] - Note: Environment variable`HF_TOKEN` is set and is the current active token independently from the token you've just configured. +[2026-10-03 14:37:17,060][starVLA.training.rl_games.auth][INFO] - Logged into Hugging Face Hub +[2026-10-03 14:37:17,074][starVLA.training.train_starvla][INFO] - ✅ Configuration wrapped for access tracking +[2026-10-03 14:37:18,007][starVLA.training.train_starvla][INFO] - Creating VLA Dataset with Mixture `openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2` +[2026-10-03 14:37:18,378][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Behavior (data_config): ['BEHAVIOR_challenge'] +[2026-10-03 14:37:18,378][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from DOMINO: ['robotwin'] +[2026-10-03 14:37:18,378][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from DOMINO (data_config): [] +[2026-10-03 14:37:18,378][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from DOMINO (data_config): ['domino', 'domino_clean', 'domino_random', 'domino_cotrain'] +[2026-10-03 14:37:18,378][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Franka: ['custom_robot_config', 'demo_sim_franka_delta_joints', 'SO101'] +[2026-10-03 14:37:18,378][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Franka (data_config): [] +[2026-10-03 14:37:18,378][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Franka (data_config): ['custom_dataset', 'custom_dataset_2', 'demo_sim_pick_place', 'SO101_pick'] +[2026-10-03 14:37:18,378][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from LIBERO: ['libero_franka'] +[2026-10-03 14:37:18,378][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from LIBERO (data_config): [] +[2026-10-03 14:37:18,378][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from LIBERO (data_config): ['libero_all', 'libero_goal', 'multi_robot'] +[2026-10-03 14:37:18,378][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from MIKASA: ['mikasa_franka_h1'] +[2026-10-03 14:37:18,379][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from MIKASA (data_config): ['mikasa_franka_h1'] +[2026-10-03 14:37:18,379][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from MIKASA (data_config): ['local/intercept_grab_fast_vla_v0_h1_train'] +[2026-10-03 14:37:18,379][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from RoboChallenge_table30v2: ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-10-03 14:37:18,379][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from RoboChallenge_table30v2 (data_config): ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-10-03 14:37:18,379][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from RoboChallenge_table30v2 (data_config): ['robochallenge_table30v2_shred_paper', 'robochallenge_table30v2_ur5_all', 'robochallenge_table30v2_arx5_all', 'robochallenge_table30v2_dosw1_all'] +[2026-10-03 14:37:18,379][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_365: ['panda_omron_robocasa365'] +[2026-10-03 14:37:18,379][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_365 (data_config): [] +[2026-10-03 14:37:18,379][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_365 (data_config): ['robocasa365_open_drawer_target_human', 'robocasa365_atomic_target_human_all', 'robocasa365_composite_target_human_all', 'robocasa365_target_human_all'] +[2026-10-03 14:37:18,379][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_tabletop: ['fourier_gr1_arms_waist'] +[2026-10-03 14:37:18,379][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_tabletop (data_config): [] +[2026-10-03 14:37:18,379][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_tabletop (data_config): ['fourier_gr1_unified_1000'] +[2026-10-03 14:37:18,379][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robotwin: ['robotwin', 'robotwin50', 'arx_x5'] +[2026-10-03 14:37:18,379][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robotwin (data_config): [] +[2026-10-03 14:37:18,379][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robotwin (data_config): ['robotwin_all', 'robotwin_all_50', 'robotwin', 'robotwin_task1', 'robotwin_task2', 'arx_x5'] +[2026-10-03 14:37:18,379][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from SimplerEnv: ['oxe_droid', 'oxe_bridge', 'oxe_rt1'] +[2026-10-03 14:37:18,379][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from SimplerEnv (data_config): [] +[2026-10-03 14:37:18,379][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from SimplerEnv (data_config): ['bridge', 'bridge_rt_1'] +[2026-10-03 14:37:18,380][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from VLA-Arena: ['vla_arena_franka'] +[2026-10-03 14:37:18,380][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from VLA-Arena (data_config): [] +[2026-10-03 14:37:18,380][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from VLA-Arena (data_config): ['vla_arena_L0_S', 'vla_arena_L0_M', 'vla_arena_L0_L'] +[2026-10-03 14:37:18,380][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from rl_games: ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_defend_the_line', 'rl_games_deadly_corridor', 'rl_games_asterix', 'rl_games_atlantis', 'rl_games_gymnasium', 'rl_games_gymnasium_discrete', 'rl_games_gymnasium_native'] +[2026-10-03 14:37:18,380][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from rl_games (data_config): ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_defend_the_line', 'rl_games_deadly_corridor', 'rl_games_asterix', 'rl_games_atlantis', 'rl_games_gymnasium', 'rl_games_gymnasium_discrete', 'rl_games_gymnasium_native'] +[2026-10-03 14:37:18,380][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from rl_games (data_config): ['flappy_train', 'flappy_train__bridge', 'flappy_mixed_latency_train', 'flappy_mixed_latency_train__bridge', 'demon_attack_train', 'demon_attack_train__bridge', 'demon_attack_mixed_latency_train', 'demon_attack_mixed_latency_train__bridge', 'defend_the_line_train', 'defend_the_line_train__bridge', 'defend_the_line_mixed_latency_train', 'defend_the_line_mixed_latency_train__bridge', 'deadly_corridor_train', 'deadly_corridor_train__bridge', 'deadly_corridor_mixed_latency_train', 'deadly_corridor_mixed_latency_train__bridge', 'asterix_train', 'asterix_train__bridge', 'asterix_mixed_latency_train', 'asterix_mixed_latency_train__bridge', 'atlantis_train', 'atlantis_train__bridge', 'atlantis_mixed_latency_train', 'atlantis_mixed_latency_train__bridge', 'inverted_pendulum_mixed_latency_train', 'inverted_pendulum_mixed_latency_train__bridge', 'inverted_pendulum_mixed_latency_train_no_latency_prompt', 'inverted_pendulum_mixed_latency_train_no_latency_prompt__bridge', 'h1hand_balance_hard'] +[2026-10-03 14:37:18,388][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded custom mixtures from /home/ubuntu/talha/playground/Datasets/rl_games/_generated_mixtures/single_622b136e13e57595.json: ['openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2', 'openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2__val'] +[2026-10-03 14:37:18,407][starVLA.dataloader][INFO] - Created train VLA dataloader with num_workers=8, persistent_workers=True, cpu_only_workers=True +[2026-10-03 14:37:18,408][starVLA.training.train_starvla][INFO] - Creating VLA Eval Dataset with Mixture `openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2__val` +[2026-10-03 14:37:18,415][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded custom mixtures from /home/ubuntu/talha/playground/Datasets/rl_games/_generated_mixtures/single_622b136e13e57595.json: ['openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2', 'openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2__val'] +[2026-10-03 14:37:18,419][starVLA.dataloader][INFO] - Created eval VLA dataloader with num_workers=8, persistent_workers=True, cpu_only_workers=True +[2026-10-03 14:37:18,473][starVLA.training.train_starvla][INFO] - 📝 Full config saved at /home/ubuntu/talha/results/Checkpoints/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2/config.full.yaml +[2026-10-03 14:37:18,483][starVLA.training.train_starvla][INFO] - 📊 Accessed config snapshot saved at /home/ubuntu/talha/results/Checkpoints/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2/config.yaml +[2026-10-03 14:37:18,483][starVLA.training.train_starvla][INFO] - No pretrained checkpoint provided. Starting training from scratch. +[2026-10-03 14:37:20,504][starVLA.training.train_starvla][INFO] - ***** Training Configuration ***** +[2026-10-03 14:37:20,504][starVLA.training.train_starvla][INFO] - Total optimization steps = 5000 +[2026-10-03 14:37:20,504][starVLA.training.train_starvla][INFO] - Per device batch size = 8 +[2026-10-03 14:37:20,504][starVLA.training.train_starvla][INFO] - Gradient accumulation steps = 2 +[2026-10-03 14:37:20,505][starVLA.training.train_starvla][INFO] - Total batch size = 16 +[2026-10-03 14:37:45,731][starVLA.training.train_starvla][INFO] - Step 1, Loss: {'train/loss': 1.5983362197875977, 'train/grad_norm_pre_clip': 126.70796966552734, 'timing/data': 0.00025752000510692596, 'timing/model': 0.4591663679893827, 'learning_rate/qwen_vl_interface': 1.0000000000000001e-07, 'learning_rate/action_model': 1.0000000000000002e-06, 'epoch': 0.0}) +[2026-10-03 14:37:46,406][starVLA.training.train_starvla][INFO] - Step 2, Loss: {'train/loss': 1.5224967002868652, 'train/grad_norm_pre_clip': 187.89666748046875, 'timing/data': 0.00039431999903172255, 'timing/model': 0.3764917669905117, 'learning_rate/qwen_vl_interface': 2.0000000000000002e-07, 'learning_rate/action_model': 2.0000000000000003e-06, 'epoch': 0.0}) +[2026-10-03 14:37:47,071][starVLA.training.train_starvla][INFO] - Step 3, Loss: {'train/loss': 1.401871681213379, 'train/grad_norm_pre_clip': 106.14061737060547, 'timing/data': 0.0002968699991470203, 'timing/model': 0.373760927002877, 'learning_rate/qwen_vl_interface': 3.0000000000000004e-07, 'learning_rate/action_model': 3e-06, 'epoch': 0.0}) +[2026-10-03 14:37:47,740][starVLA.training.train_starvla][INFO] - Step 4, Loss: {'train/loss': 1.4422523975372314, 'train/grad_norm_pre_clip': 68.97527313232422, 'timing/data': 0.00032874000316951424, 'timing/model': 0.3754186970036244, 'learning_rate/qwen_vl_interface': 4.0000000000000003e-07, 'learning_rate/action_model': 4.000000000000001e-06, 'epoch': 0.0}) +[2026-10-03 14:37:48,401][starVLA.training.train_starvla][INFO] - Step 5, Loss: {'train/loss': 1.1336774826049805, 'train/grad_norm_pre_clip': 38.47635269165039, 'timing/data': 0.0002664700004970655, 'timing/model': 0.37058428800082766, 'learning_rate/qwen_vl_interface': 5.000000000000001e-07, 'learning_rate/action_model': 5e-06, 'epoch': 0.0}) +[2026-10-03 14:37:49,260][starVLA.training.train_starvla][INFO] - Step 6, Loss: {'train/loss': 1.0921242237091064, 'train/grad_norm_pre_clip': 33.14817810058594, 'timing/data': 0.00020774001313839108, 'timing/model': 0.5700990549958078, 'learning_rate/qwen_vl_interface': 6.000000000000001e-07, 'learning_rate/action_model': 6e-06, 'epoch': 0.0}) +[2026-10-03 14:37:49,920][starVLA.training.train_starvla][INFO] - Step 7, Loss: {'train/loss': 1.4135209321975708, 'train/grad_norm_pre_clip': 63.508338928222656, 'timing/data': 0.000308620001305826, 'timing/model': 0.37013789800403174, 'learning_rate/qwen_vl_interface': 7.000000000000001e-07, 'learning_rate/action_model': 7.000000000000001e-06, 'epoch': 0.0}) +[2026-10-03 14:37:50,581][starVLA.training.train_starvla][INFO] - Step 8, Loss: {'train/loss': 0.651614785194397, 'train/grad_norm_pre_clip': 22.687360763549805, 'timing/data': 0.0002535700041335076, 'timing/model': 0.3730796579911839, 'learning_rate/qwen_vl_interface': 8.000000000000001e-07, 'learning_rate/action_model': 8.000000000000001e-06, 'epoch': 0.0}) +[2026-10-03 14:37:51,239][starVLA.training.train_starvla][INFO] - Step 9, Loss: {'train/loss': 1.0451624393463135, 'train/grad_norm_pre_clip': 45.4612922668457, 'timing/data': 0.00025406999338883907, 'timing/model': 0.36909609798749443, 'learning_rate/qwen_vl_interface': 9.000000000000001e-07, 'learning_rate/action_model': 9e-06, 'epoch': 0.0}) +[2026-10-03 14:37:51,895][starVLA.training.train_starvla][INFO] - Step 10, Loss: {'train/loss': 1.095975399017334, 'train/grad_norm_pre_clip': 28.035179138183594, 'timing/data': 0.0003171399876009673, 'timing/model': 0.36886648800282273, 'learning_rate/qwen_vl_interface': 1.0000000000000002e-06, 'learning_rate/action_model': 1e-05, 'epoch': 0.0}) +[2026-10-03 14:37:52,760][starVLA.training.train_starvla][INFO] - Step 11, Loss: {'train/loss': 0.9108079671859741, 'train/grad_norm_pre_clip': 27.028881072998047, 'timing/data': 0.00031781999859958887, 'timing/model': 0.5755309240048518, 'learning_rate/qwen_vl_interface': 1.1e-06, 'learning_rate/action_model': 1.1000000000000001e-05, 'epoch': 0.0}) +[2026-10-03 14:37:53,396][starVLA.training.train_starvla][INFO] - Step 12, Loss: {'train/loss': 0.6740413904190063, 'train/grad_norm_pre_clip': 35.46213150024414, 'timing/data': 0.00030835998768452555, 'timing/model': 0.3708457070024451, 'learning_rate/qwen_vl_interface': 1.2000000000000002e-06, 'learning_rate/action_model': 1.2e-05, 'epoch': 0.0}) +[2026-10-03 14:37:54,054][starVLA.training.train_starvla][INFO] - Step 13, Loss: {'train/loss': 1.1720242500305176, 'train/grad_norm_pre_clip': 57.752750396728516, 'timing/data': 0.0003163900109939277, 'timing/model': 0.36934313901292626, 'learning_rate/qwen_vl_interface': 1.3e-06, 'learning_rate/action_model': 1.3000000000000001e-05, 'epoch': 0.0}) +[2026-10-03 14:37:54,709][starVLA.training.train_starvla][INFO] - Step 14, Loss: {'train/loss': 0.5707615613937378, 'train/grad_norm_pre_clip': 26.50067710876465, 'timing/data': 0.0003194699966115877, 'timing/model': 0.3683993190061301, 'learning_rate/qwen_vl_interface': 1.4000000000000001e-06, 'learning_rate/action_model': 1.4000000000000001e-05, 'epoch': 0.0}) +[2026-10-03 14:37:55,367][starVLA.training.train_starvla][INFO] - Step 15, Loss: {'train/loss': 1.4341309070587158, 'train/grad_norm_pre_clip': 69.18634796142578, 'timing/data': 0.00028279999969527125, 'timing/model': 0.36934804801421706, 'learning_rate/qwen_vl_interface': 1.5e-06, 'learning_rate/action_model': 1.5e-05, 'epoch': 0.0}) +[2026-10-03 14:37:56,024][starVLA.training.train_starvla][INFO] - Step 16, Loss: {'train/loss': 0.8252453207969666, 'train/grad_norm_pre_clip': 31.285749435424805, 'timing/data': 0.00019437000446487218, 'timing/model': 0.37014697800623253, 'learning_rate/qwen_vl_interface': 1.6000000000000001e-06, 'learning_rate/action_model': 1.6000000000000003e-05, 'epoch': 0.0}) +[2026-10-03 14:37:56,889][starVLA.training.train_starvla][INFO] - Step 17, Loss: {'train/loss': 1.48454749584198, 'train/grad_norm_pre_clip': 67.35362243652344, 'timing/data': 0.00039159000152722, 'timing/model': 0.3690917989879381, 'learning_rate/qwen_vl_interface': 1.7000000000000002e-06, 'learning_rate/action_model': 1.7000000000000003e-05, 'epoch': 0.0}) +[2026-10-03 14:37:57,545][starVLA.training.train_starvla][INFO] - Step 18, Loss: {'train/loss': 1.318634033203125, 'train/grad_norm_pre_clip': 108.9542236328125, 'timing/data': 0.00030464999144896865, 'timing/model': 0.36853023900766857, 'learning_rate/qwen_vl_interface': 1.8000000000000001e-06, 'learning_rate/action_model': 1.8e-05, 'epoch': 0.0}) +[2026-10-03 14:37:58,202][starVLA.training.train_starvla][INFO] - Step 19, Loss: {'train/loss': 1.4408341646194458, 'train/grad_norm_pre_clip': 108.62138366699219, 'timing/data': 0.0003692299942485988, 'timing/model': 0.3692933980055386, 'learning_rate/qwen_vl_interface': 1.9000000000000002e-06, 'learning_rate/action_model': 1.9e-05, 'epoch': 0.0}) +[2026-10-03 14:37:58,860][starVLA.training.train_starvla][INFO] - Step 20, Loss: {'train/loss': 1.1127935647964478, 'train/grad_norm_pre_clip': 101.77255249023438, 'timing/data': 0.00034389999927952886, 'timing/model': 0.3708390780084301, 'learning_rate/qwen_vl_interface': 2.0000000000000003e-06, 'learning_rate/action_model': 2e-05, 'epoch': 0.0}) +[2026-10-03 14:37:59,494][starVLA.training.train_starvla][INFO] - Step 21, Loss: {'train/loss': 0.7773816585540771, 'train/grad_norm_pre_clip': 52.67823791503906, 'timing/data': 0.0002950100024463609, 'timing/model': 0.3689550590061117, 'learning_rate/qwen_vl_interface': 2.1000000000000002e-06, 'learning_rate/action_model': 2.1e-05, 'epoch': 0.0}) +[2026-10-03 14:38:00,296][starVLA.training.train_starvla][INFO] - Step 22, Loss: {'train/loss': 1.5297874212265015, 'train/grad_norm_pre_clip': 91.79676818847656, 'timing/data': 0.00021408000611700118, 'timing/model': 0.5370263079967117, 'learning_rate/qwen_vl_interface': 2.2e-06, 'learning_rate/action_model': 2.2000000000000003e-05, 'epoch': 0.0}) +[2026-10-03 14:38:00,953][starVLA.training.train_starvla][INFO] - Step 23, Loss: {'train/loss': 1.1817950010299683, 'train/grad_norm_pre_clip': 67.04298400878906, 'timing/data': 0.0002714500005822629, 'timing/model': 0.3687273279938381, 'learning_rate/qwen_vl_interface': 2.3000000000000004e-06, 'learning_rate/action_model': 2.3000000000000003e-05, 'epoch': 0.0}) +[2026-10-03 14:38:01,610][starVLA.training.train_starvla][INFO] - Step 24, Loss: {'train/loss': 1.5838255882263184, 'train/grad_norm_pre_clip': 86.61519622802734, 'timing/data': 0.00039095000829547644, 'timing/model': 0.3693056279880693, 'learning_rate/qwen_vl_interface': 2.4000000000000003e-06, 'learning_rate/action_model': 2.4e-05, 'epoch': 0.0}) +[2026-10-03 14:38:02,267][starVLA.training.train_starvla][INFO] - Step 25, Loss: {'train/loss': 1.1286571025848389, 'train/grad_norm_pre_clip': 46.91501235961914, 'timing/data': 0.00020576000679284334, 'timing/model': 0.36956102898693644, 'learning_rate/qwen_vl_interface': 2.5e-06, 'learning_rate/action_model': 2.5e-05, 'epoch': 0.0}) +[2026-10-03 14:38:02,923][starVLA.training.train_starvla][INFO] - Step 26, Loss: {'train/loss': 2.3528504371643066, 'train/grad_norm_pre_clip': 56.91714859008789, 'timing/data': 0.00034664999111555517, 'timing/model': 0.36851551799918525, 'learning_rate/qwen_vl_interface': 2.6e-06, 'learning_rate/action_model': 2.6000000000000002e-05, 'epoch': 0.0}) +[2026-10-03 14:38:03,580][starVLA.training.train_starvla][INFO] - Step 27, Loss: {'train/loss': 1.303906798362732, 'train/grad_norm_pre_clip': 82.55596160888672, 'timing/data': 0.0002630200033308938, 'timing/model': 0.3689525079971645, 'learning_rate/qwen_vl_interface': 2.7000000000000004e-06, 'learning_rate/action_model': 2.7000000000000002e-05, 'epoch': 0.0}) +[2026-10-03 14:38:04,431][starVLA.training.train_starvla][INFO] - Step 28, Loss: {'train/loss': 0.9477073550224304, 'train/grad_norm_pre_clip': 45.2249755859375, 'timing/data': 0.00038952000613790005, 'timing/model': 0.37105110799893737, 'learning_rate/qwen_vl_interface': 2.8000000000000003e-06, 'learning_rate/action_model': 2.8000000000000003e-05, 'epoch': 0.0}) +[2026-10-03 14:38:05,092][starVLA.training.train_starvla][INFO] - Step 29, Loss: {'train/loss': 0.8740396499633789, 'train/grad_norm_pre_clip': 50.36904525756836, 'timing/data': 0.0004204300057608634, 'timing/model': 0.37039502800325863, 'learning_rate/qwen_vl_interface': 2.9e-06, 'learning_rate/action_model': 2.9e-05, 'epoch': 0.0}) +[2026-10-03 14:38:05,752][starVLA.training.train_starvla][INFO] - Step 30, Loss: {'train/loss': 0.35181641578674316, 'train/grad_norm_pre_clip': 46.01908493041992, 'timing/data': 0.00029579999682027847, 'timing/model': 0.37236758701328654, 'learning_rate/qwen_vl_interface': 3e-06, 'learning_rate/action_model': 3e-05, 'epoch': 0.0}) +[2026-10-03 14:38:06,413][starVLA.training.train_starvla][INFO] - Step 31, Loss: {'train/loss': 0.8797425627708435, 'train/grad_norm_pre_clip': 47.86246109008789, 'timing/data': 0.0002526299940655008, 'timing/model': 0.37147545799962245, 'learning_rate/qwen_vl_interface': 3.1000000000000004e-06, 'learning_rate/action_model': 3.1e-05, 'epoch': 0.0}) +[2026-10-03 14:38:07,072][starVLA.training.train_starvla][INFO] - Step 32, Loss: {'train/loss': 1.9891042709350586, 'train/grad_norm_pre_clip': 113.51661682128906, 'timing/data': 0.0002591600059531629, 'timing/model': 0.3711846279911697, 'learning_rate/qwen_vl_interface': 3.2000000000000003e-06, 'learning_rate/action_model': 3.2000000000000005e-05, 'epoch': 0.0}) +[2026-10-03 14:38:07,923][starVLA.training.train_starvla][INFO] - Step 33, Loss: {'train/loss': 2.2689921855926514, 'train/grad_norm_pre_clip': 127.23387908935547, 'timing/data': 0.0003717200015671551, 'timing/model': 0.3709035780047998, 'learning_rate/qwen_vl_interface': 3.3000000000000006e-06, 'learning_rate/action_model': 3.3e-05, 'epoch': 0.0}) +[2026-10-03 14:38:08,584][starVLA.training.train_starvla][INFO] - Step 34, Loss: {'train/loss': 1.3548076152801514, 'train/grad_norm_pre_clip': 77.36807250976562, 'timing/data': 0.00024778000079095364, 'timing/model': 0.37196522699377965, 'learning_rate/qwen_vl_interface': 3.4000000000000005e-06, 'learning_rate/action_model': 3.4000000000000007e-05, 'epoch': 0.0}) +[2026-10-03 14:38:09,244][starVLA.training.train_starvla][INFO] - Step 35, Loss: {'train/loss': 1.1037737131118774, 'train/grad_norm_pre_clip': 46.39503860473633, 'timing/data': 0.0003265999985160306, 'timing/model': 0.3714624480053317, 'learning_rate/qwen_vl_interface': 3.5e-06, 'learning_rate/action_model': 3.5e-05, 'epoch': 0.0}) +[2026-10-03 14:38:09,903][starVLA.training.train_starvla][INFO] - Step 36, Loss: {'train/loss': 0.9623947143554688, 'train/grad_norm_pre_clip': 71.95903778076172, 'timing/data': 0.00021223000658210367, 'timing/model': 0.37081814800330903, 'learning_rate/qwen_vl_interface': 3.6000000000000003e-06, 'learning_rate/action_model': 3.6e-05, 'epoch': 0.0}) +[2026-10-03 14:38:10,560][starVLA.training.train_starvla][INFO] - Step 37, Loss: {'train/loss': 1.1385934352874756, 'train/grad_norm_pre_clip': 98.92405700683594, 'timing/data': 0.00020620999566745013, 'timing/model': 0.3693698279967066, 'learning_rate/qwen_vl_interface': 3.7e-06, 'learning_rate/action_model': 3.7e-05, 'epoch': 0.0}) +[2026-10-03 14:38:11,408][starVLA.training.train_starvla][INFO] - Step 38, Loss: {'train/loss': 1.8812021017074585, 'train/grad_norm_pre_clip': 104.25028991699219, 'timing/data': 0.0002986399922519922, 'timing/model': 0.5604303260042798, 'learning_rate/qwen_vl_interface': 3.8000000000000005e-06, 'learning_rate/action_model': 3.8e-05, 'epoch': 0.0}) +[2026-10-03 14:38:12,066][starVLA.training.train_starvla][INFO] - Step 39, Loss: {'train/loss': 0.9655648469924927, 'train/grad_norm_pre_clip': 39.63518142700195, 'timing/data': 0.0002148200001101941, 'timing/model': 0.36962249799398705, 'learning_rate/qwen_vl_interface': 3.900000000000001e-06, 'learning_rate/action_model': 3.9000000000000006e-05, 'epoch': 0.0}) +[2026-10-03 14:38:12,726][starVLA.training.train_starvla][INFO] - Step 40, Loss: {'train/loss': 1.2627443075180054, 'train/grad_norm_pre_clip': 58.09628677368164, 'timing/data': 0.00036502000875771046, 'timing/model': 0.37090956800966524, 'learning_rate/qwen_vl_interface': 4.000000000000001e-06, 'learning_rate/action_model': 4e-05, 'epoch': 0.0}) +[2026-10-03 14:38:13,383][starVLA.training.train_starvla][INFO] - Step 41, Loss: {'train/loss': 1.1863154172897339, 'train/grad_norm_pre_clip': 51.763431549072266, 'timing/data': 0.0002045900037046522, 'timing/model': 0.3696262979938183, 'learning_rate/qwen_vl_interface': 4.1e-06, 'learning_rate/action_model': 4.1e-05, 'epoch': 0.0}) +[2026-10-03 14:38:14,040][starVLA.training.train_starvla][INFO] - Step 42, Loss: {'train/loss': 1.2109038829803467, 'train/grad_norm_pre_clip': 44.75571823120117, 'timing/data': 0.0002617599966470152, 'timing/model': 0.3689458779990673, 'learning_rate/qwen_vl_interface': 4.2000000000000004e-06, 'learning_rate/action_model': 4.2e-05, 'epoch': 0.0}) +[2026-10-03 14:38:14,893][starVLA.training.train_starvla][INFO] - Step 43, Loss: {'train/loss': 1.228513479232788, 'train/grad_norm_pre_clip': 40.9153938293457, 'timing/data': 0.00036639999598264694, 'timing/model': 0.5642172649968415, 'learning_rate/qwen_vl_interface': 4.3e-06, 'learning_rate/action_model': 4.3e-05, 'epoch': 0.0}) +[2026-10-03 14:38:15,530][starVLA.training.train_starvla][INFO] - Step 44, Loss: {'train/loss': 1.1224591732025146, 'train/grad_norm_pre_clip': 53.61534881591797, 'timing/data': 0.0003352499916218221, 'timing/model': 0.3713720480009215, 'learning_rate/qwen_vl_interface': 4.4e-06, 'learning_rate/action_model': 4.4000000000000006e-05, 'epoch': 0.0}) +[2026-10-03 14:38:16,164][starVLA.training.train_starvla][INFO] - Step 45, Loss: {'train/loss': 1.6489070653915405, 'train/grad_norm_pre_clip': 94.9553451538086, 'timing/data': 0.00038067999412305653, 'timing/model': 0.3464300199993886, 'learning_rate/qwen_vl_interface': 4.5e-06, 'learning_rate/action_model': 4.5e-05, 'epoch': 0.0}) +[2026-10-03 14:38:16,821][starVLA.training.train_starvla][INFO] - Step 46, Loss: {'train/loss': 1.1080775260925293, 'train/grad_norm_pre_clip': 101.06356048583984, 'timing/data': 0.0002091999922413379, 'timing/model': 0.3690619879926089, 'learning_rate/qwen_vl_interface': 4.600000000000001e-06, 'learning_rate/action_model': 4.600000000000001e-05, 'epoch': 0.0}) +[2026-10-03 14:38:17,480][starVLA.training.train_starvla][INFO] - Step 47, Loss: {'train/loss': 1.3310813903808594, 'train/grad_norm_pre_clip': 123.82513427734375, 'timing/data': 0.00029121000261511654, 'timing/model': 0.3698407589981798, 'learning_rate/qwen_vl_interface': 4.7e-06, 'learning_rate/action_model': 4.7e-05, 'epoch': 0.0}) +[2026-10-03 14:38:18,140][starVLA.training.train_starvla][INFO] - Step 48, Loss: {'train/loss': 1.5373096466064453, 'train/grad_norm_pre_clip': 45.05059814453125, 'timing/data': 0.00032560000545345247, 'timing/model': 0.37129164799989667, 'learning_rate/qwen_vl_interface': 4.800000000000001e-06, 'learning_rate/action_model': 4.8e-05, 'epoch': 0.0}) +[2026-10-03 14:38:18,987][starVLA.training.train_starvla][INFO] - Step 49, Loss: {'train/loss': 1.1811177730560303, 'train/grad_norm_pre_clip': 46.154842376708984, 'timing/data': 0.00030906000756658614, 'timing/model': 0.3700263079954311, 'learning_rate/qwen_vl_interface': 4.9000000000000005e-06, 'learning_rate/action_model': 4.9e-05, 'epoch': 0.0}) +[2026-10-03 14:38:19,643][starVLA.training.train_starvla][INFO] - Step 50, Loss: {'train/loss': 0.6948363184928894, 'train/grad_norm_pre_clip': 17.818845748901367, 'timing/data': 0.0002549600030761212, 'timing/model': 0.36897805900662206, 'learning_rate/qwen_vl_interface': 5e-06, 'learning_rate/action_model': 5e-05, 'epoch': 0.0}) +[2026-10-03 14:38:20,302][starVLA.training.train_starvla][INFO] - Step 51, Loss: {'train/loss': 1.0192844867706299, 'train/grad_norm_pre_clip': 90.90779876708984, 'timing/data': 0.00027129999944008887, 'timing/model': 0.369404448007117, 'learning_rate/qwen_vl_interface': 5.1e-06, 'learning_rate/action_model': 5.1000000000000006e-05, 'epoch': 0.0}) +[2026-10-03 14:38:20,939][starVLA.training.train_starvla][INFO] - Step 52, Loss: {'train/loss': 0.6260676383972168, 'train/grad_norm_pre_clip': 25.251157760620117, 'timing/data': 0.0002704499929677695, 'timing/model': 0.3483290309959557, 'learning_rate/qwen_vl_interface': 5.2e-06, 'learning_rate/action_model': 5.2000000000000004e-05, 'epoch': 0.0}) +[2026-10-03 14:38:21,596][starVLA.training.train_starvla][INFO] - Step 53, Loss: {'train/loss': 1.6245918273925781, 'train/grad_norm_pre_clip': 205.43605041503906, 'timing/data': 0.00043557000753935426, 'timing/model': 0.36917770899890456, 'learning_rate/qwen_vl_interface': 5.300000000000001e-06, 'learning_rate/action_model': 5.300000000000001e-05, 'epoch': 0.0}) +[2026-10-03 14:38:22,450][starVLA.training.train_starvla][INFO] - Step 54, Loss: {'train/loss': 1.5099581480026245, 'train/grad_norm_pre_clip': 194.89739990234375, 'timing/data': 0.00022108999837655574, 'timing/model': 0.36931506800465286, 'learning_rate/qwen_vl_interface': 5.400000000000001e-06, 'learning_rate/action_model': 5.4000000000000005e-05, 'epoch': 0.0}) +[2026-10-03 14:38:23,108][starVLA.training.train_starvla][INFO] - Step 55, Loss: {'train/loss': 2.2865493297576904, 'train/grad_norm_pre_clip': 228.94007873535156, 'timing/data': 0.00022834999253973365, 'timing/model': 0.3692124880035408, 'learning_rate/qwen_vl_interface': 5.500000000000001e-06, 'learning_rate/action_model': 5.500000000000001e-05, 'epoch': 0.0}) +[2026-10-03 14:38:23,768][starVLA.training.train_starvla][INFO] - Step 56, Loss: {'train/loss': 1.333808422088623, 'train/grad_norm_pre_clip': 43.964786529541016, 'timing/data': 0.0002773000014713034, 'timing/model': 0.37166435799736064, 'learning_rate/qwen_vl_interface': 5.600000000000001e-06, 'learning_rate/action_model': 5.6000000000000006e-05, 'epoch': 0.0}) +[2026-10-03 14:38:24,424][starVLA.training.train_starvla][INFO] - Step 57, Loss: {'train/loss': 0.6271026134490967, 'train/grad_norm_pre_clip': 108.54220581054688, 'timing/data': 0.00020858900097664446, 'timing/model': 0.369139518996235, 'learning_rate/qwen_vl_interface': 5.7e-06, 'learning_rate/action_model': 5.6999999999999996e-05, 'epoch': 0.0}) +[2026-10-03 14:38:25,081][starVLA.training.train_starvla][INFO] - Step 58, Loss: {'train/loss': 1.4338191747665405, 'train/grad_norm_pre_clip': 132.80918884277344, 'timing/data': 0.00039282899524550885, 'timing/model': 0.3688903590082191, 'learning_rate/qwen_vl_interface': 5.8e-06, 'learning_rate/action_model': 5.8e-05, 'epoch': 0.0}) +[2026-10-03 14:38:25,912][starVLA.training.train_starvla][INFO] - Step 59, Loss: {'train/loss': 0.43988165259361267, 'train/grad_norm_pre_clip': 34.971351623535156, 'timing/data': 0.0002771599974948913, 'timing/model': 0.5653162550006527, 'learning_rate/qwen_vl_interface': 5.9e-06, 'learning_rate/action_model': 5.9e-05, 'epoch': 0.0}) +[2026-10-03 14:38:26,571][starVLA.training.train_starvla][INFO] - Step 60, Loss: {'train/loss': 0.7415488958358765, 'train/grad_norm_pre_clip': 59.5223388671875, 'timing/data': 0.0002056700031971559, 'timing/model': 0.3710578269965481, 'learning_rate/qwen_vl_interface': 6e-06, 'learning_rate/action_model': 6e-05, 'epoch': 0.0}) +[2026-10-03 14:38:27,227][starVLA.training.train_starvla][INFO] - Step 61, Loss: {'train/loss': 1.7105728387832642, 'train/grad_norm_pre_clip': 65.22473907470703, 'timing/data': 0.0003610600106185302, 'timing/model': 0.3687778579915175, 'learning_rate/qwen_vl_interface': 6.1e-06, 'learning_rate/action_model': 6.1e-05, 'epoch': 0.0}) +[2026-10-03 14:38:27,883][starVLA.training.train_starvla][INFO] - Step 62, Loss: {'train/loss': 1.8695588111877441, 'train/grad_norm_pre_clip': 151.2830352783203, 'timing/data': 0.0002516800013836473, 'timing/model': 0.3687657869886607, 'learning_rate/qwen_vl_interface': 6.200000000000001e-06, 'learning_rate/action_model': 6.2e-05, 'epoch': 0.0}) +[2026-10-03 14:38:28,540][starVLA.training.train_starvla][INFO] - Step 63, Loss: {'train/loss': 1.163393497467041, 'train/grad_norm_pre_clip': 80.2948226928711, 'timing/data': 0.0003340599942021072, 'timing/model': 0.36929729700204916, 'learning_rate/qwen_vl_interface': 6.300000000000001e-06, 'learning_rate/action_model': 6.3e-05, 'epoch': 0.0}) +[2026-10-03 14:38:29,178][starVLA.training.train_starvla][INFO] - Step 64, Loss: {'train/loss': 0.5222198367118835, 'train/grad_norm_pre_clip': 149.99119567871094, 'timing/data': 0.0003413500089664012, 'timing/model': 0.3475509800045984, 'learning_rate/qwen_vl_interface': 6.4000000000000006e-06, 'learning_rate/action_model': 6.400000000000001e-05, 'epoch': 0.0}) +[2026-10-03 14:38:30,007][starVLA.training.train_starvla][INFO] - Step 65, Loss: {'train/loss': 0.06579403579235077, 'train/grad_norm_pre_clip': 41.0629768371582, 'timing/data': 0.0003623100055847317, 'timing/model': 0.3471140900073806, 'learning_rate/qwen_vl_interface': 6.5000000000000004e-06, 'learning_rate/action_model': 6.500000000000001e-05, 'epoch': 0.0}) +[2026-10-03 14:38:30,666][starVLA.training.train_starvla][INFO] - Step 66, Loss: {'train/loss': 1.8178752660751343, 'train/grad_norm_pre_clip': 126.49604797363281, 'timing/data': 0.0003016800037585199, 'timing/model': 0.36973663700337056, 'learning_rate/qwen_vl_interface': 6.600000000000001e-06, 'learning_rate/action_model': 6.6e-05, 'epoch': 0.0}) +[2026-10-03 14:38:31,325][starVLA.training.train_starvla][INFO] - Step 67, Loss: {'train/loss': 1.6548033952713013, 'train/grad_norm_pre_clip': 143.3114471435547, 'timing/data': 0.00024945000768639147, 'timing/model': 0.37047177700151224, 'learning_rate/qwen_vl_interface': 6.700000000000001e-06, 'learning_rate/action_model': 6.7e-05, 'epoch': 0.0}) +[2026-10-03 14:38:31,985][starVLA.training.train_starvla][INFO] - Step 68, Loss: {'train/loss': 2.185718059539795, 'train/grad_norm_pre_clip': 169.1294403076172, 'timing/data': 0.0003109300014330074, 'timing/model': 0.37111307799932547, 'learning_rate/qwen_vl_interface': 6.800000000000001e-06, 'learning_rate/action_model': 6.800000000000001e-05, 'epoch': 0.0}) +[2026-10-03 14:38:32,645][starVLA.training.train_starvla][INFO] - Step 69, Loss: {'train/loss': 0.14878873527050018, 'train/grad_norm_pre_clip': 22.608198165893555, 'timing/data': 0.0003435399994486943, 'timing/model': 0.37026020800112747, 'learning_rate/qwen_vl_interface': 6.9e-06, 'learning_rate/action_model': 6.9e-05, 'epoch': 0.01}) +[2026-10-03 14:38:33,508][starVLA.training.train_starvla][INFO] - Step 70, Loss: {'train/loss': 1.1696544885635376, 'train/grad_norm_pre_clip': 84.67569732666016, 'timing/data': 0.00027455999224912375, 'timing/model': 0.3704201770015061, 'learning_rate/qwen_vl_interface': 7e-06, 'learning_rate/action_model': 7e-05, 'epoch': 0.01}) +[2026-10-03 14:38:34,167][starVLA.training.train_starvla][INFO] - Step 71, Loss: {'train/loss': 0.85934978723526, 'train/grad_norm_pre_clip': 52.804927825927734, 'timing/data': 0.00022294999507721514, 'timing/model': 0.3701721769903088, 'learning_rate/qwen_vl_interface': 7.100000000000001e-06, 'learning_rate/action_model': 7.1e-05, 'epoch': 0.01}) +[2026-10-03 14:38:34,827][starVLA.training.train_starvla][INFO] - Step 72, Loss: {'train/loss': 0.7031638622283936, 'train/grad_norm_pre_clip': 26.91241455078125, 'timing/data': 0.0003340499970363453, 'timing/model': 0.3705309479992138, 'learning_rate/qwen_vl_interface': 7.2000000000000005e-06, 'learning_rate/action_model': 7.2e-05, 'epoch': 0.01}) +[2026-10-03 14:38:35,486][starVLA.training.train_starvla][INFO] - Step 73, Loss: {'train/loss': 1.596402645111084, 'train/grad_norm_pre_clip': 103.78450775146484, 'timing/data': 0.00024633998691570014, 'timing/model': 0.3704643870005384, 'learning_rate/qwen_vl_interface': 7.3e-06, 'learning_rate/action_model': 7.3e-05, 'epoch': 0.01}) +[2026-10-03 14:38:36,123][starVLA.training.train_starvla][INFO] - Step 74, Loss: {'train/loss': 0.6099042892456055, 'train/grad_norm_pre_clip': 76.1552963256836, 'timing/data': 0.0002689300017664209, 'timing/model': 0.3471862099977443, 'learning_rate/qwen_vl_interface': 7.4e-06, 'learning_rate/action_model': 7.4e-05, 'epoch': 0.01}) +[2026-10-03 14:38:36,971][starVLA.training.train_starvla][INFO] - Step 75, Loss: {'train/loss': 0.9149000644683838, 'train/grad_norm_pre_clip': 44.01177978515625, 'timing/data': 0.0003179199993610382, 'timing/model': 0.5595634060009615, 'learning_rate/qwen_vl_interface': 7.500000000000001e-06, 'learning_rate/action_model': 7.500000000000001e-05, 'epoch': 0.01}) +[2026-10-03 14:38:37,631][starVLA.training.train_starvla][INFO] - Step 76, Loss: {'train/loss': 1.2050752639770508, 'train/grad_norm_pre_clip': 90.1473159790039, 'timing/data': 0.00025035999715328217, 'timing/model': 0.3711900079943007, 'learning_rate/qwen_vl_interface': 7.600000000000001e-06, 'learning_rate/action_model': 7.6e-05, 'epoch': 0.01}) +[2026-10-03 14:38:38,290][starVLA.training.train_starvla][INFO] - Step 77, Loss: {'train/loss': 1.3851096630096436, 'train/grad_norm_pre_clip': 128.613525390625, 'timing/data': 0.00024094000400509685, 'timing/model': 0.3702828179957578, 'learning_rate/qwen_vl_interface': 7.7e-06, 'learning_rate/action_model': 7.7e-05, 'epoch': 0.01}) +[2026-10-03 14:38:38,925][starVLA.training.train_starvla][INFO] - Step 78, Loss: {'train/loss': 0.5729761719703674, 'train/grad_norm_pre_clip': 82.75448608398438, 'timing/data': 0.0002501700073480606, 'timing/model': 0.3465012899978319, 'learning_rate/qwen_vl_interface': 7.800000000000002e-06, 'learning_rate/action_model': 7.800000000000001e-05, 'epoch': 0.01}) +[2026-10-03 14:38:39,584][starVLA.training.train_starvla][INFO] - Step 79, Loss: {'train/loss': 1.0006475448608398, 'train/grad_norm_pre_clip': 55.03159713745117, 'timing/data': 0.00031581000075675547, 'timing/model': 0.3701486380014103, 'learning_rate/qwen_vl_interface': 7.9e-06, 'learning_rate/action_model': 7.900000000000001e-05, 'epoch': 0.01}) +[2026-10-03 14:38:40,222][starVLA.training.train_starvla][INFO] - Step 80, Loss: {'train/loss': 0.3455579876899719, 'train/grad_norm_pre_clip': 38.47232437133789, 'timing/data': 0.00020568999752867967, 'timing/model': 0.34853153000585735, 'learning_rate/qwen_vl_interface': 8.000000000000001e-06, 'learning_rate/action_model': 8e-05, 'epoch': 0.01}) +[2026-10-03 14:38:41,079][starVLA.training.train_starvla][INFO] - Step 81, Loss: {'train/loss': 1.3445005416870117, 'train/grad_norm_pre_clip': 79.94340515136719, 'timing/data': 0.00035146999289281666, 'timing/model': 0.3708579579979414, 'learning_rate/qwen_vl_interface': 8.1e-06, 'learning_rate/action_model': 8.1e-05, 'epoch': 0.01}) +[2026-10-03 14:38:41,736][starVLA.training.train_starvla][INFO] - Step 82, Loss: {'train/loss': 1.4151607751846313, 'train/grad_norm_pre_clip': 108.46654510498047, 'timing/data': 0.0003178400074830279, 'timing/model': 0.3693172369967215, 'learning_rate/qwen_vl_interface': 8.2e-06, 'learning_rate/action_model': 8.2e-05, 'epoch': 0.01}) +[2026-10-03 14:38:42,396][starVLA.training.train_starvla][INFO] - Step 83, Loss: {'train/loss': 1.3008477687835693, 'train/grad_norm_pre_clip': 50.87040328979492, 'timing/data': 0.0003433800011407584, 'timing/model': 0.37060344699420966, 'learning_rate/qwen_vl_interface': 8.3e-06, 'learning_rate/action_model': 8.3e-05, 'epoch': 0.01}) +[2026-10-03 14:38:43,056][starVLA.training.train_starvla][INFO] - Step 84, Loss: {'train/loss': 0.5746757388114929, 'train/grad_norm_pre_clip': 55.11724090576172, 'timing/data': 0.00035976000071968883, 'timing/model': 0.37077841800055467, 'learning_rate/qwen_vl_interface': 8.400000000000001e-06, 'learning_rate/action_model': 8.4e-05, 'epoch': 0.01}) +[2026-10-03 14:38:43,716][starVLA.training.train_starvla][INFO] - Step 85, Loss: {'train/loss': 0.7613362669944763, 'train/grad_norm_pre_clip': 19.353029251098633, 'timing/data': 0.0002107699983753264, 'timing/model': 0.3707702070096275, 'learning_rate/qwen_vl_interface': 8.5e-06, 'learning_rate/action_model': 8.5e-05, 'epoch': 0.01}) +[2026-10-03 14:38:44,546][starVLA.training.train_starvla][INFO] - Step 86, Loss: {'train/loss': 0.8921604156494141, 'train/grad_norm_pre_clip': 49.644107818603516, 'timing/data': 0.00023778999457135797, 'timing/model': 0.564825954992557, 'learning_rate/qwen_vl_interface': 8.6e-06, 'learning_rate/action_model': 8.6e-05, 'epoch': 0.01}) +[2026-10-03 14:38:45,206][starVLA.training.train_starvla][INFO] - Step 87, Loss: {'train/loss': 0.7312204837799072, 'train/grad_norm_pre_clip': 31.281814575195312, 'timing/data': 0.0002696900046430528, 'timing/model': 0.37074308800220024, 'learning_rate/qwen_vl_interface': 8.700000000000001e-06, 'learning_rate/action_model': 8.7e-05, 'epoch': 0.01}) +[2026-10-03 14:38:45,865][starVLA.training.train_starvla][INFO] - Step 88, Loss: {'train/loss': 1.4052066802978516, 'train/grad_norm_pre_clip': 59.0046272277832, 'timing/data': 0.00033522000012453645, 'timing/model': 0.37012820699601434, 'learning_rate/qwen_vl_interface': 8.8e-06, 'learning_rate/action_model': 8.800000000000001e-05, 'epoch': 0.01}) +[2026-10-03 14:38:46,526][starVLA.training.train_starvla][INFO] - Step 89, Loss: {'train/loss': 0.9860546588897705, 'train/grad_norm_pre_clip': 51.88747024536133, 'timing/data': 0.0002861200046027079, 'timing/model': 0.37073433700425085, 'learning_rate/qwen_vl_interface': 8.900000000000001e-06, 'learning_rate/action_model': 8.900000000000001e-05, 'epoch': 0.01}) +[2026-10-03 14:38:47,184][starVLA.training.train_starvla][INFO] - Step 90, Loss: {'train/loss': 1.191230058670044, 'train/grad_norm_pre_clip': 29.514991760253906, 'timing/data': 0.00020751998818013817, 'timing/model': 0.37027557801047806, 'learning_rate/qwen_vl_interface': 9e-06, 'learning_rate/action_model': 9e-05, 'epoch': 0.01}) +[2026-10-03 14:38:47,843][starVLA.training.train_starvla][INFO] - Step 91, Loss: {'train/loss': 0.9148887395858765, 'train/grad_norm_pre_clip': 67.67318725585938, 'timing/data': 0.0003737499937415123, 'timing/model': 0.3699926279950887, 'learning_rate/qwen_vl_interface': 9.100000000000001e-06, 'learning_rate/action_model': 9.1e-05, 'epoch': 0.01}) +[2026-10-03 14:38:48,694][starVLA.training.train_starvla][INFO] - Step 92, Loss: {'train/loss': 0.9273525476455688, 'train/grad_norm_pre_clip': 86.83252716064453, 'timing/data': 0.0003410190111026168, 'timing/model': 0.37102385800972115, 'learning_rate/qwen_vl_interface': 9.200000000000002e-06, 'learning_rate/action_model': 9.200000000000001e-05, 'epoch': 0.01}) +[2026-10-03 14:38:49,354][starVLA.training.train_starvla][INFO] - Step 93, Loss: {'train/loss': 1.00528883934021, 'train/grad_norm_pre_clip': 29.893827438354492, 'timing/data': 0.00021009000192862004, 'timing/model': 0.37089680798817426, 'learning_rate/qwen_vl_interface': 9.3e-06, 'learning_rate/action_model': 9.300000000000001e-05, 'epoch': 0.01}) +[2026-10-03 14:38:50,012][starVLA.training.train_starvla][INFO] - Step 94, Loss: {'train/loss': 1.2626644372940063, 'train/grad_norm_pre_clip': 44.0699348449707, 'timing/data': 0.00023808999685570598, 'timing/model': 0.369936807997874, 'learning_rate/qwen_vl_interface': 9.4e-06, 'learning_rate/action_model': 9.4e-05, 'epoch': 0.01}) +[2026-10-03 14:38:50,673][starVLA.training.train_starvla][INFO] - Step 95, Loss: {'train/loss': 1.372997760772705, 'train/grad_norm_pre_clip': 86.60287475585938, 'timing/data': 0.00022218999220058322, 'timing/model': 0.37149171700002626, 'learning_rate/qwen_vl_interface': 9.5e-06, 'learning_rate/action_model': 9.5e-05, 'epoch': 0.01}) +[2026-10-03 14:38:51,335][starVLA.training.train_starvla][INFO] - Step 96, Loss: {'train/loss': 0.7356007099151611, 'train/grad_norm_pre_clip': 108.70858001708984, 'timing/data': 0.00034347000473644584, 'timing/model': 0.37180943800194655, 'learning_rate/qwen_vl_interface': 9.600000000000001e-06, 'learning_rate/action_model': 9.6e-05, 'epoch': 0.01}) +[2026-10-03 14:38:52,188][starVLA.training.train_starvla][INFO] - Step 97, Loss: {'train/loss': 1.318172812461853, 'train/grad_norm_pre_clip': 94.7997055053711, 'timing/data': 0.0002665499923750758, 'timing/model': 0.3707052880054107, 'learning_rate/qwen_vl_interface': 9.7e-06, 'learning_rate/action_model': 9.7e-05, 'epoch': 0.01}) +[2026-10-03 14:38:52,848][starVLA.training.train_starvla][INFO] - Step 98, Loss: {'train/loss': 0.6611580848693848, 'train/grad_norm_pre_clip': 35.37529373168945, 'timing/data': 0.00020546000450849533, 'timing/model': 0.37048659700667486, 'learning_rate/qwen_vl_interface': 9.800000000000001e-06, 'learning_rate/action_model': 9.8e-05, 'epoch': 0.01}) +[2026-10-03 14:38:53,508][starVLA.training.train_starvla][INFO] - Step 99, Loss: {'train/loss': 0.7764983177185059, 'train/grad_norm_pre_clip': 52.35749435424805, 'timing/data': 0.00032178001129068434, 'timing/model': 0.3715477480000118, 'learning_rate/qwen_vl_interface': 9.9e-06, 'learning_rate/action_model': 9.900000000000001e-05, 'epoch': 0.01}) +[2026-10-03 14:38:54,170][starVLA.training.train_starvla][INFO] - Step 100, Loss: {'train/loss': 0.678391695022583, 'train/grad_norm_pre_clip': 39.89354705810547, 'timing/data': 0.0002092200011247769, 'timing/model': 0.3716670779976994, 'learning_rate/qwen_vl_interface': 1e-05, 'learning_rate/action_model': 0.0001, 'epoch': 0.01}) +[2026-10-03 14:38:54,830][starVLA.training.train_starvla][INFO] - Step 101, Loss: {'train/loss': 1.0304595232009888, 'train/grad_norm_pre_clip': 46.031959533691406, 'timing/data': 0.00024634999863337725, 'timing/model': 0.37074468699574936, 'learning_rate/qwen_vl_interface': 9.999999023727212e-06, 'learning_rate/action_model': 9.99999902372721e-05, 'epoch': 0.01}) +[2026-10-03 14:38:55,654][starVLA.training.train_starvla][INFO] - Step 102, Loss: {'train/loss': 0.5650225281715393, 'train/grad_norm_pre_clip': 32.93065643310547, 'timing/data': 0.0002455099893268198, 'timing/model': 0.5349402889987687, 'learning_rate/qwen_vl_interface': 9.999996094909247e-06, 'learning_rate/action_model': 9.999996094909248e-05, 'epoch': 0.01}) +[2026-10-03 14:38:56,314][starVLA.training.train_starvla][INFO] - Step 103, Loss: {'train/loss': 0.7995538711547852, 'train/grad_norm_pre_clip': 87.0214614868164, 'timing/data': 0.00033909999183379114, 'timing/model': 0.3710674769972684, 'learning_rate/qwen_vl_interface': 9.999991213547314e-06, 'learning_rate/action_model': 9.999991213547313e-05, 'epoch': 0.01}) +[2026-10-03 14:38:56,975][starVLA.training.train_starvla][INFO] - Step 104, Loss: {'train/loss': 0.8577485084533691, 'train/grad_norm_pre_clip': 27.907224655151367, 'timing/data': 0.00034512999991420656, 'timing/model': 0.37136946800455917, 'learning_rate/qwen_vl_interface': 9.999984379643413e-06, 'learning_rate/action_model': 9.999984379643412e-05, 'epoch': 0.01}) +[2026-10-03 14:38:57,637][starVLA.training.train_starvla][INFO] - Step 105, Loss: {'train/loss': 0.5924840569496155, 'train/grad_norm_pre_clip': 24.531827926635742, 'timing/data': 0.00033681000059004873, 'timing/model': 0.3715361079957802, 'learning_rate/qwen_vl_interface': 9.999975593200357e-06, 'learning_rate/action_model': 9.999975593200357e-05, 'epoch': 0.01}) +[2026-10-03 14:38:58,297][starVLA.training.train_starvla][INFO] - Step 106, Loss: {'train/loss': 1.004246711730957, 'train/grad_norm_pre_clip': 27.96446990966797, 'timing/data': 0.0003261499950895086, 'timing/model': 0.3705761470046127, 'learning_rate/qwen_vl_interface': 9.999964854221756e-06, 'learning_rate/action_model': 9.999964854221754e-05, 'epoch': 0.01}) +[2026-10-03 14:38:59,151][starVLA.training.train_starvla][INFO] - Step 107, Loss: {'train/loss': 0.626396656036377, 'train/grad_norm_pre_clip': 32.119590759277344, 'timing/data': 0.00024316000053659081, 'timing/model': 0.5653173459868412, 'learning_rate/qwen_vl_interface': 9.999952162712027e-06, 'learning_rate/action_model': 9.999952162712026e-05, 'epoch': 0.01}) +[2026-10-03 14:38:59,811][starVLA.training.train_starvla][INFO] - Step 108, Loss: {'train/loss': 0.47631019353866577, 'train/grad_norm_pre_clip': 17.38913345336914, 'timing/data': 0.0002590500080259517, 'timing/model': 0.3708315779949771, 'learning_rate/qwen_vl_interface': 9.999937518676382e-06, 'learning_rate/action_model': 9.999937518676382e-05, 'epoch': 0.01}) +[2026-10-03 14:39:00,472][starVLA.training.train_starvla][INFO] - Step 109, Loss: {'train/loss': 0.6651278138160706, 'train/grad_norm_pre_clip': 49.32651138305664, 'timing/data': 0.00026313000125810504, 'timing/model': 0.3709470670000883, 'learning_rate/qwen_vl_interface': 9.999920922120846e-06, 'learning_rate/action_model': 9.999920922120846e-05, 'epoch': 0.01}) +[2026-10-03 14:39:01,132][starVLA.training.train_starvla][INFO] - Step 110, Loss: {'train/loss': 0.6901067495346069, 'train/grad_norm_pre_clip': 26.993167877197266, 'timing/data': 0.00027303000388201326, 'timing/model': 0.37054880800133105, 'learning_rate/qwen_vl_interface': 9.999902373052239e-06, 'learning_rate/action_model': 9.999902373052239e-05, 'epoch': 0.01}) +[2026-10-03 14:39:01,793][starVLA.training.train_starvla][INFO] - Step 111, Loss: {'train/loss': 0.4842754304409027, 'train/grad_norm_pre_clip': 47.27065658569336, 'timing/data': 0.00027088000206276774, 'timing/model': 0.3713099879969377, 'learning_rate/qwen_vl_interface': 9.999881871478188e-06, 'learning_rate/action_model': 9.999881871478187e-05, 'epoch': 0.01}) +[2026-10-03 14:39:02,454][starVLA.training.train_starvla][INFO] - Step 112, Loss: {'train/loss': 0.8280261754989624, 'train/grad_norm_pre_clip': 60.42290115356445, 'timing/data': 0.0003691800084197894, 'timing/model': 0.3716169770050328, 'learning_rate/qwen_vl_interface': 9.999859417407115e-06, 'learning_rate/action_model': 9.999859417407115e-05, 'epoch': 0.01}) +[2026-10-03 14:39:03,305][starVLA.training.train_starvla][INFO] - Step 113, Loss: {'train/loss': 0.8522121906280518, 'train/grad_norm_pre_clip': 58.32837677001953, 'timing/data': 0.00028156999906059355, 'timing/model': 0.371186686999863, 'learning_rate/qwen_vl_interface': 9.999835010848254e-06, 'learning_rate/action_model': 9.999835010848254e-05, 'epoch': 0.01}) +[2026-10-03 14:39:03,964][starVLA.training.train_starvla][INFO] - Step 114, Loss: {'train/loss': 0.9873600006103516, 'train/grad_norm_pre_clip': 148.0504913330078, 'timing/data': 0.00026342000637669116, 'timing/model': 0.3702710469951853, 'learning_rate/qwen_vl_interface': 9.999808651811639e-06, 'learning_rate/action_model': 9.999808651811639e-05, 'epoch': 0.01}) +[2026-10-03 14:39:04,625][starVLA.training.train_starvla][INFO] - Step 115, Loss: {'train/loss': 0.8540295362472534, 'train/grad_norm_pre_clip': 73.79306030273438, 'timing/data': 0.00034505000803619623, 'timing/model': 0.3711197969969362, 'learning_rate/qwen_vl_interface': 9.999780340308101e-06, 'learning_rate/action_model': 9.999780340308101e-05, 'epoch': 0.01}) +[2026-10-03 14:39:05,263][starVLA.training.train_starvla][INFO] - Step 116, Loss: {'train/loss': 0.9307715892791748, 'train/grad_norm_pre_clip': 70.98358917236328, 'timing/data': 0.0003362199931871146, 'timing/model': 0.3712087669991888, 'learning_rate/qwen_vl_interface': 9.999750076349282e-06, 'learning_rate/action_model': 9.999750076349282e-05, 'epoch': 0.01}) +[2026-10-03 14:39:05,925][starVLA.training.train_starvla][INFO] - Step 117, Loss: {'train/loss': 1.1030648946762085, 'train/grad_norm_pre_clip': 137.7822265625, 'timing/data': 0.00022279001132119447, 'timing/model': 0.37140333699062467, 'learning_rate/qwen_vl_interface': 9.999717859947618e-06, 'learning_rate/action_model': 9.999717859947618e-05, 'epoch': 0.01}) +[2026-10-03 14:39:06,778][starVLA.training.train_starvla][INFO] - Step 118, Loss: {'train/loss': 0.8927538394927979, 'train/grad_norm_pre_clip': 54.190460205078125, 'timing/data': 0.00034332000359427184, 'timing/model': 0.3705231380008627, 'learning_rate/qwen_vl_interface': 9.999683691116356e-06, 'learning_rate/action_model': 9.999683691116355e-05, 'epoch': 0.01}) +[2026-10-03 14:39:07,415][starVLA.training.train_starvla][INFO] - Step 119, Loss: {'train/loss': 0.7182632088661194, 'train/grad_norm_pre_clip': 87.61982727050781, 'timing/data': 0.0003322500124340877, 'timing/model': 0.37090157700004056, 'learning_rate/qwen_vl_interface': 9.999647569869539e-06, 'learning_rate/action_model': 9.99964756986954e-05, 'epoch': 0.01}) +[2026-10-03 14:39:08,077][starVLA.training.train_starvla][INFO] - Step 120, Loss: {'train/loss': 0.8042715191841125, 'train/grad_norm_pre_clip': 49.460044860839844, 'timing/data': 0.00033152000105474144, 'timing/model': 0.3714386479987297, 'learning_rate/qwen_vl_interface': 9.999609496222017e-06, 'learning_rate/action_model': 9.999609496222017e-05, 'epoch': 0.01}) +[2026-10-03 14:39:08,737][starVLA.training.train_starvla][INFO] - Step 121, Loss: {'train/loss': 0.8308725357055664, 'train/grad_norm_pre_clip': 41.75037384033203, 'timing/data': 0.00023601000430062413, 'timing/model': 0.37086064799223095, 'learning_rate/qwen_vl_interface': 9.99956947018944e-06, 'learning_rate/action_model': 9.999569470189439e-05, 'epoch': 0.01}) +[2026-10-03 14:39:09,397][starVLA.training.train_starvla][INFO] - Step 122, Loss: {'train/loss': 0.5072201490402222, 'train/grad_norm_pre_clip': 13.421092987060547, 'timing/data': 0.00022813001123722643, 'timing/model': 0.37089924700558186, 'learning_rate/qwen_vl_interface': 9.99952749178826e-06, 'learning_rate/action_model': 9.999527491788259e-05, 'epoch': 0.01}) +[2026-10-03 14:39:10,257][starVLA.training.train_starvla][INFO] - Step 123, Loss: {'train/loss': 0.8927779197692871, 'train/grad_norm_pre_clip': 25.870372772216797, 'timing/data': 0.0003299200034234673, 'timing/model': 0.37104875801014714, 'learning_rate/qwen_vl_interface': 9.999483561035732e-06, 'learning_rate/action_model': 9.999483561035732e-05, 'epoch': 0.01}) +[2026-10-03 14:39:10,919][starVLA.training.train_starvla][INFO] - Step 124, Loss: {'train/loss': 0.9689150452613831, 'train/grad_norm_pre_clip': 49.675479888916016, 'timing/data': 0.0003270699962740764, 'timing/model': 0.3718323670036625, 'learning_rate/qwen_vl_interface': 9.999437677949917e-06, 'learning_rate/action_model': 9.999437677949917e-05, 'epoch': 0.01}) +[2026-10-03 14:39:11,581][starVLA.training.train_starvla][INFO] - Step 125, Loss: {'train/loss': 0.8942798972129822, 'train/grad_norm_pre_clip': 30.954696655273438, 'timing/data': 0.0003784700093092397, 'timing/model': 0.37150788799044676, 'learning_rate/qwen_vl_interface': 9.999389842549674e-06, 'learning_rate/action_model': 9.999389842549673e-05, 'epoch': 0.01}) +[2026-10-03 14:39:12,240][starVLA.training.train_starvla][INFO] - Step 126, Loss: {'train/loss': 0.6805202960968018, 'train/grad_norm_pre_clip': 52.46930694580078, 'timing/data': 0.00033845999860204756, 'timing/model': 0.37069216700911056, 'learning_rate/qwen_vl_interface': 9.999340054854669e-06, 'learning_rate/action_model': 9.999340054854667e-05, 'epoch': 0.01}) +[2026-10-03 14:39:12,901][starVLA.training.train_starvla][INFO] - Step 127, Loss: {'train/loss': 0.7023024559020996, 'train/grad_norm_pre_clip': 28.810195922851562, 'timing/data': 0.00028888000815641135, 'timing/model': 0.3712112079956569, 'learning_rate/qwen_vl_interface': 9.999288314885362e-06, 'learning_rate/action_model': 9.999288314885362e-05, 'epoch': 0.01}) +[2026-10-03 14:39:13,759][starVLA.training.train_starvla][INFO] - Step 128, Loss: {'train/loss': 1.1765851974487305, 'train/grad_norm_pre_clip': 35.45575714111328, 'timing/data': 0.00026822999643627554, 'timing/model': 0.5682155250106007, 'learning_rate/qwen_vl_interface': 9.999234622663027e-06, 'learning_rate/action_model': 9.999234622663028e-05, 'epoch': 0.01}) +[2026-10-03 14:39:14,420][starVLA.training.train_starvla][INFO] - Step 129, Loss: {'train/loss': 0.30106982588768005, 'train/grad_norm_pre_clip': 29.269119262695312, 'timing/data': 0.00033799000084400177, 'timing/model': 0.37113033699279185, 'learning_rate/qwen_vl_interface': 9.999178978209732e-06, 'learning_rate/action_model': 9.999178978209732e-05, 'epoch': 0.01}) +[2026-10-03 14:39:15,080][starVLA.training.train_starvla][INFO] - Step 130, Loss: {'train/loss': 0.5238974690437317, 'train/grad_norm_pre_clip': 58.14055633544922, 'timing/data': 0.00032136900699697435, 'timing/model': 0.37047773800441064, 'learning_rate/qwen_vl_interface': 9.999121381548354e-06, 'learning_rate/action_model': 9.999121381548353e-05, 'epoch': 0.01}) +[2026-10-03 14:39:15,718][starVLA.training.train_starvla][INFO] - Step 131, Loss: {'train/loss': 0.8360869884490967, 'train/grad_norm_pre_clip': 57.75654983520508, 'timing/data': 0.0002779000060399994, 'timing/model': 0.37144038699625526, 'learning_rate/qwen_vl_interface': 9.999061832702563e-06, 'learning_rate/action_model': 9.999061832702563e-05, 'epoch': 0.01}) +[2026-10-03 14:39:16,380][starVLA.training.train_starvla][INFO] - Step 132, Loss: {'train/loss': 0.849166989326477, 'train/grad_norm_pre_clip': 28.63115882873535, 'timing/data': 0.00030868999601807445, 'timing/model': 0.3718628369970247, 'learning_rate/qwen_vl_interface': 9.999000331696842e-06, 'learning_rate/action_model': 9.999000331696841e-05, 'epoch': 0.01}) +[2026-10-03 14:39:17,237][starVLA.training.train_starvla][INFO] - Step 133, Loss: {'train/loss': 0.46058568358421326, 'train/grad_norm_pre_clip': 82.36724853515625, 'timing/data': 0.00020787998801097274, 'timing/model': 0.5675428350077709, 'learning_rate/qwen_vl_interface': 9.99893687855647e-06, 'learning_rate/action_model': 9.99893687855647e-05, 'epoch': 0.01}) +[2026-10-03 14:39:17,897][starVLA.training.train_starvla][INFO] - Step 134, Loss: {'train/loss': 0.6114727258682251, 'train/grad_norm_pre_clip': 32.87967300415039, 'timing/data': 0.0002101800055243075, 'timing/model': 0.37034109700471163, 'learning_rate/qwen_vl_interface': 9.998871473307529e-06, 'learning_rate/action_model': 9.99887147330753e-05, 'epoch': 0.01}) +[2026-10-03 14:39:18,557][starVLA.training.train_starvla][INFO] - Step 135, Loss: {'train/loss': 0.7155361175537109, 'train/grad_norm_pre_clip': 31.874576568603516, 'timing/data': 0.00028484000358730555, 'timing/model': 0.37136119698698167, 'learning_rate/qwen_vl_interface': 9.998804115976907e-06, 'learning_rate/action_model': 9.998804115976908e-05, 'epoch': 0.01}) +[2026-10-03 14:39:19,218][starVLA.training.train_starvla][INFO] - Step 136, Loss: {'train/loss': 0.5596016645431519, 'train/grad_norm_pre_clip': 83.15575408935547, 'timing/data': 0.00035720900632441044, 'timing/model': 0.37145949799742084, 'learning_rate/qwen_vl_interface': 9.99873480659229e-06, 'learning_rate/action_model': 9.99873480659229e-05, 'epoch': 0.01}) +[2026-10-03 14:39:19,856][starVLA.training.train_starvla][INFO] - Step 137, Loss: {'train/loss': 1.0761216878890991, 'train/grad_norm_pre_clip': 49.63500213623047, 'timing/data': 0.0002513700019335374, 'timing/model': 0.3709904270072002, 'learning_rate/qwen_vl_interface': 9.99866354518217e-06, 'learning_rate/action_model': 9.99866354518217e-05, 'epoch': 0.01}) +[2026-10-03 14:39:20,516][starVLA.training.train_starvla][INFO] - Step 138, Loss: {'train/loss': 0.3633262515068054, 'train/grad_norm_pre_clip': 24.12058448791504, 'timing/data': 0.0002048599999397993, 'timing/model': 0.3710654769965913, 'learning_rate/qwen_vl_interface': 9.99859033177584e-06, 'learning_rate/action_model': 9.99859033177584e-05, 'epoch': 0.01}) +[2026-10-03 14:39:21,372][starVLA.training.train_starvla][INFO] - Step 139, Loss: {'train/loss': 0.7133402824401855, 'train/grad_norm_pre_clip': 112.4291763305664, 'timing/data': 0.00034487000084482133, 'timing/model': 0.37133301798894536, 'learning_rate/qwen_vl_interface': 9.998515166403394e-06, 'learning_rate/action_model': 9.998515166403394e-05, 'epoch': 0.01}) +[2026-10-03 14:39:22,032][starVLA.training.train_starvla][INFO] - Step 140, Loss: {'train/loss': 1.2086942195892334, 'train/grad_norm_pre_clip': 45.81009292602539, 'timing/data': 0.0003765400033444166, 'timing/model': 0.3708799069863744, 'learning_rate/qwen_vl_interface': 9.99843804909573e-06, 'learning_rate/action_model': 9.99843804909573e-05, 'epoch': 0.01}) +[2026-10-03 14:39:22,695][starVLA.training.train_starvla][INFO] - Step 141, Loss: {'train/loss': 0.5319421291351318, 'train/grad_norm_pre_clip': 113.72115325927734, 'timing/data': 0.0004004600050393492, 'timing/model': 0.37163506700017024, 'learning_rate/qwen_vl_interface': 9.998358979884548e-06, 'learning_rate/action_model': 9.998358979884547e-05, 'epoch': 0.01}) +[2026-10-03 14:39:23,355][starVLA.training.train_starvla][INFO] - Step 142, Loss: {'train/loss': 1.0858240127563477, 'train/grad_norm_pre_clip': 69.46916198730469, 'timing/data': 0.00021152998670004308, 'timing/model': 0.37085890700109303, 'learning_rate/qwen_vl_interface': 9.998277958802352e-06, 'learning_rate/action_model': 9.998277958802351e-05, 'epoch': 0.01}) +[2026-10-03 14:39:24,016][starVLA.training.train_starvla][INFO] - Step 143, Loss: {'train/loss': 1.3764108419418335, 'train/grad_norm_pre_clip': 148.57095336914062, 'timing/data': 0.0002943400031654164, 'timing/model': 0.3713311669998802, 'learning_rate/qwen_vl_interface': 9.998194985882442e-06, 'learning_rate/action_model': 9.998194985882442e-05, 'epoch': 0.01}) +[2026-10-03 14:39:24,872][starVLA.training.train_starvla][INFO] - Step 144, Loss: {'train/loss': 1.6487085819244385, 'train/grad_norm_pre_clip': 136.76243591308594, 'timing/data': 0.0004887099930783734, 'timing/model': 0.5660955450002803, 'learning_rate/qwen_vl_interface': 9.99811006115893e-06, 'learning_rate/action_model': 9.998110061158931e-05, 'epoch': 0.01}) +[2026-10-03 14:39:25,534][starVLA.training.train_starvla][INFO] - Step 145, Loss: {'train/loss': 1.474853277206421, 'train/grad_norm_pre_clip': 76.89398193359375, 'timing/data': 0.00021183000353630632, 'timing/model': 0.3718896469945321, 'learning_rate/qwen_vl_interface': 9.998023184666726e-06, 'learning_rate/action_model': 9.998023184666726e-05, 'epoch': 0.01}) +[2026-10-03 14:39:26,195][starVLA.training.train_starvla][INFO] - Step 146, Loss: {'train/loss': 0.5461261868476868, 'train/grad_norm_pre_clip': 26.835586547851562, 'timing/data': 0.0003273499896749854, 'timing/model': 0.37095993699040264, 'learning_rate/qwen_vl_interface': 9.997934356441538e-06, 'learning_rate/action_model': 9.997934356441537e-05, 'epoch': 0.01}) +[2026-10-03 14:39:26,856][starVLA.training.train_starvla][INFO] - Step 147, Loss: {'train/loss': 0.9838473200798035, 'train/grad_norm_pre_clip': 111.94515991210938, 'timing/data': 0.00033537000126671046, 'timing/model': 0.37125271701370366, 'learning_rate/qwen_vl_interface': 9.997843576519879e-06, 'learning_rate/action_model': 9.997843576519878e-05, 'epoch': 0.01}) +[2026-10-03 14:39:27,517][starVLA.training.train_starvla][INFO] - Step 148, Loss: {'train/loss': 0.9418525695800781, 'train/grad_norm_pre_clip': 69.24268341064453, 'timing/data': 0.00028467000811360776, 'timing/model': 0.3714951279980596, 'learning_rate/qwen_vl_interface': 9.99775084493907e-06, 'learning_rate/action_model': 9.997750844939069e-05, 'epoch': 0.01}) +[2026-10-03 14:39:28,376][starVLA.training.train_starvla][INFO] - Step 149, Loss: {'train/loss': 1.1342369318008423, 'train/grad_norm_pre_clip': 22.425600051879883, 'timing/data': 0.0002667200024006888, 'timing/model': 0.5683722050016513, 'learning_rate/qwen_vl_interface': 9.997656161737227e-06, 'learning_rate/action_model': 9.997656161737226e-05, 'epoch': 0.01}) +[2026-10-03 14:39:29,036][starVLA.training.train_starvla][INFO] - Step 150, Loss: {'train/loss': 0.42619767785072327, 'train/grad_norm_pre_clip': 24.756027221679688, 'timing/data': 0.0002933100040536374, 'timing/model': 0.3706439169909572, 'learning_rate/qwen_vl_interface': 9.99755952695327e-06, 'learning_rate/action_model': 9.997559526953269e-05, 'epoch': 0.01}) +[2026-10-03 14:39:29,697][starVLA.training.train_starvla][INFO] - Step 151, Loss: {'train/loss': 0.6336764097213745, 'train/grad_norm_pre_clip': 19.22859001159668, 'timing/data': 0.0003093800041824579, 'timing/model': 0.3717263879952952, 'learning_rate/qwen_vl_interface': 9.99746094062692e-06, 'learning_rate/action_model': 9.99746094062692e-05, 'epoch': 0.01}) +[2026-10-03 14:39:30,358][starVLA.training.train_starvla][INFO] - Step 152, Loss: {'train/loss': 0.5116496086120605, 'train/grad_norm_pre_clip': 25.347532272338867, 'timing/data': 0.00032088998705148697, 'timing/model': 0.3715393570018932, 'learning_rate/qwen_vl_interface': 9.997360402798706e-06, 'learning_rate/action_model': 9.997360402798705e-05, 'epoch': 0.01}) +[2026-10-03 14:39:31,019][starVLA.training.train_starvla][INFO] - Step 153, Loss: {'train/loss': 1.0129659175872803, 'train/grad_norm_pre_clip': 92.44942474365234, 'timing/data': 0.00021368000307120383, 'timing/model': 0.3707846269971924, 'learning_rate/qwen_vl_interface': 9.997257913509955e-06, 'learning_rate/action_model': 9.997257913509955e-05, 'epoch': 0.01}) +[2026-10-03 14:39:31,680][starVLA.training.train_starvla][INFO] - Step 154, Loss: {'train/loss': 0.965764582157135, 'train/grad_norm_pre_clip': 40.698402404785156, 'timing/data': 0.0002901400002883747, 'timing/model': 0.3711442670028191, 'learning_rate/qwen_vl_interface': 9.997153472802794e-06, 'learning_rate/action_model': 9.997153472802795e-05, 'epoch': 0.01}) +[2026-10-03 14:39:32,536][starVLA.training.train_starvla][INFO] - Step 155, Loss: {'train/loss': 0.7034187912940979, 'train/grad_norm_pre_clip': 108.25586700439453, 'timing/data': 0.00032171000202652067, 'timing/model': 0.37131910800235346, 'learning_rate/qwen_vl_interface': 9.997047080720157e-06, 'learning_rate/action_model': 9.997047080720158e-05, 'epoch': 0.01}) +[2026-10-03 14:39:33,197][starVLA.training.train_starvla][INFO] - Step 156, Loss: {'train/loss': 1.0388615131378174, 'train/grad_norm_pre_clip': 25.186046600341797, 'timing/data': 0.0002119799901265651, 'timing/model': 0.3712777470063884, 'learning_rate/qwen_vl_interface': 9.996938737305777e-06, 'learning_rate/action_model': 9.996938737305776e-05, 'epoch': 0.01}) +[2026-10-03 14:39:33,859][starVLA.training.train_starvla][INFO] - Step 157, Loss: {'train/loss': 0.49588465690612793, 'train/grad_norm_pre_clip': 68.34342956542969, 'timing/data': 0.0003315600042697042, 'timing/model': 0.37164397700689733, 'learning_rate/qwen_vl_interface': 9.996828442604189e-06, 'learning_rate/action_model': 9.99682844260419e-05, 'epoch': 0.01}) +[2026-10-03 14:39:34,519][starVLA.training.train_starvla][INFO] - Step 158, Loss: {'train/loss': 0.7469547986984253, 'train/grad_norm_pre_clip': 40.56638717651367, 'timing/data': 0.00021838999236933887, 'timing/model': 0.37056138800107874, 'learning_rate/qwen_vl_interface': 9.996716196660733e-06, 'learning_rate/action_model': 9.996716196660731e-05, 'epoch': 0.01}) +[2026-10-03 14:39:35,180][starVLA.training.train_starvla][INFO] - Step 159, Loss: {'train/loss': 1.0694260597229004, 'train/grad_norm_pre_clip': 61.29869079589844, 'timing/data': 0.00030519999563694, 'timing/model': 0.3713213480077684, 'learning_rate/qwen_vl_interface': 9.996601999521546e-06, 'learning_rate/action_model': 9.996601999521546e-05, 'epoch': 0.01}) +[2026-10-03 14:39:36,032][starVLA.training.train_starvla][INFO] - Step 160, Loss: {'train/loss': 1.2262122631072998, 'train/grad_norm_pre_clip': 92.30522918701172, 'timing/data': 0.00020904000848531723, 'timing/model': 0.562008025997784, 'learning_rate/qwen_vl_interface': 9.996485851233572e-06, 'learning_rate/action_model': 9.996485851233572e-05, 'epoch': 0.01}) +[2026-10-03 14:39:36,693][starVLA.training.train_starvla][INFO] - Step 161, Loss: {'train/loss': 0.48579642176628113, 'train/grad_norm_pre_clip': 50.85639572143555, 'timing/data': 0.00020839000353589654, 'timing/model': 0.371489298006054, 'learning_rate/qwen_vl_interface': 9.996367751844556e-06, 'learning_rate/action_model': 9.996367751844555e-05, 'epoch': 0.01}) +[2026-10-03 14:39:37,354][starVLA.training.train_starvla][INFO] - Step 162, Loss: {'train/loss': 0.6040613651275635, 'train/grad_norm_pre_clip': 99.88011169433594, 'timing/data': 0.00021749999723397195, 'timing/model': 0.37098068799241446, 'learning_rate/qwen_vl_interface': 9.996247701403043e-06, 'learning_rate/action_model': 9.996247701403042e-05, 'epoch': 0.01}) +[2026-10-03 14:39:37,992][starVLA.training.train_starvla][INFO] - Step 163, Loss: {'train/loss': 0.33388566970825195, 'train/grad_norm_pre_clip': 114.52666473388672, 'timing/data': 0.0002719300100579858, 'timing/model': 0.3717795379925519, 'learning_rate/qwen_vl_interface': 9.99612569995838e-06, 'learning_rate/action_model': 9.99612569995838e-05, 'epoch': 0.01}) +[2026-10-03 14:39:38,653][starVLA.training.train_starvla][INFO] - Step 164, Loss: {'train/loss': 0.8269542455673218, 'train/grad_norm_pre_clip': 67.09097290039062, 'timing/data': 0.0003353099891683087, 'timing/model': 0.371317477998673, 'learning_rate/qwen_vl_interface': 9.99600174756072e-06, 'learning_rate/action_model': 9.99600174756072e-05, 'epoch': 0.01}) +[2026-10-03 14:39:39,510][starVLA.training.train_starvla][INFO] - Step 165, Loss: {'train/loss': 0.2122384011745453, 'train/grad_norm_pre_clip': 117.37791442871094, 'timing/data': 0.0002976399991894141, 'timing/model': 0.5665139050106518, 'learning_rate/qwen_vl_interface': 9.995875844261012e-06, 'learning_rate/action_model': 9.995875844261013e-05, 'epoch': 0.01}) +[2026-10-03 14:39:40,171][starVLA.training.train_starvla][INFO] - Step 166, Loss: {'train/loss': 0.703199028968811, 'train/grad_norm_pre_clip': 46.65547180175781, 'timing/data': 0.0002514400111977011, 'timing/model': 0.37057898700004444, 'learning_rate/qwen_vl_interface': 9.995747990111015e-06, 'learning_rate/action_model': 9.995747990111013e-05, 'epoch': 0.01}) +[2026-10-03 14:39:40,832][starVLA.training.train_starvla][INFO] - Step 167, Loss: {'train/loss': 0.4596133232116699, 'train/grad_norm_pre_clip': 40.43849563598633, 'timing/data': 0.0002676699950825423, 'timing/model': 0.37162180700397585, 'learning_rate/qwen_vl_interface': 9.99561818516328e-06, 'learning_rate/action_model': 9.99561818516328e-05, 'epoch': 0.01}) +[2026-10-03 14:39:41,493][starVLA.training.train_starvla][INFO] - Step 168, Loss: {'train/loss': 1.0826196670532227, 'train/grad_norm_pre_clip': 70.13838195800781, 'timing/data': 0.0002725100057432428, 'timing/model': 0.3716622570063919, 'learning_rate/qwen_vl_interface': 9.995486429471166e-06, 'learning_rate/action_model': 9.995486429471166e-05, 'epoch': 0.01}) +[2026-10-03 14:39:42,154][starVLA.training.train_starvla][INFO] - Step 169, Loss: {'train/loss': 1.8914811611175537, 'train/grad_norm_pre_clip': 167.4744873046875, 'timing/data': 0.00036235999141354114, 'timing/model': 0.3706992369989166, 'learning_rate/qwen_vl_interface': 9.995352723088833e-06, 'learning_rate/action_model': 9.995352723088834e-05, 'epoch': 0.01}) +[2026-10-03 14:39:42,814][starVLA.training.train_starvla][INFO] - Step 170, Loss: {'train/loss': 1.1680806875228882, 'train/grad_norm_pre_clip': 140.7357635498047, 'timing/data': 0.00032715999986976385, 'timing/model': 0.37032310699578375, 'learning_rate/qwen_vl_interface': 9.995217066071245e-06, 'learning_rate/action_model': 9.995217066071246e-05, 'epoch': 0.01}) +[2026-10-03 14:39:43,668][starVLA.training.train_starvla][INFO] - Step 171, Loss: {'train/loss': 0.43176424503326416, 'train/grad_norm_pre_clip': 21.986852645874023, 'timing/data': 0.0003390100027900189, 'timing/model': 0.3712454170017736, 'learning_rate/qwen_vl_interface': 9.995079458474163e-06, 'learning_rate/action_model': 9.995079458474164e-05, 'epoch': 0.01}) +[2026-10-03 14:39:44,329][starVLA.training.train_starvla][INFO] - Step 172, Loss: {'train/loss': 0.5360775589942932, 'train/grad_norm_pre_clip': 70.35259246826172, 'timing/data': 0.00034625999978743494, 'timing/model': 0.3716974669951014, 'learning_rate/qwen_vl_interface': 9.994939900354154e-06, 'learning_rate/action_model': 9.994939900354154e-05, 'epoch': 0.01}) +[2026-10-03 14:39:44,991][starVLA.training.train_starvla][INFO] - Step 173, Loss: {'train/loss': 0.30887940526008606, 'train/grad_norm_pre_clip': 43.0133171081543, 'timing/data': 0.0002633200056152418, 'timing/model': 0.37139576800109353, 'learning_rate/qwen_vl_interface': 9.994798391768582e-06, 'learning_rate/action_model': 9.994798391768582e-05, 'epoch': 0.01}) +[2026-10-03 14:39:45,651][starVLA.training.train_starvla][INFO] - Step 174, Loss: {'train/loss': 0.4305135905742645, 'train/grad_norm_pre_clip': 76.07510375976562, 'timing/data': 0.00026996000087819993, 'timing/model': 0.37027469799795654, 'learning_rate/qwen_vl_interface': 9.99465493277562e-06, 'learning_rate/action_model': 9.994654932775618e-05, 'epoch': 0.01}) +[2026-10-03 14:39:46,311][starVLA.training.train_starvla][INFO] - Step 175, Loss: {'train/loss': 0.42500007152557373, 'train/grad_norm_pre_clip': 39.607765197753906, 'timing/data': 0.0003464599867584184, 'timing/model': 0.37101222699857317, 'learning_rate/qwen_vl_interface': 9.994509523434235e-06, 'learning_rate/action_model': 9.994509523434235e-05, 'epoch': 0.01}) +[2026-10-03 14:39:47,166][starVLA.training.train_starvla][INFO] - Step 176, Loss: {'train/loss': 0.7797141075134277, 'train/grad_norm_pre_clip': 45.44475173950195, 'timing/data': 0.00033028000325430185, 'timing/model': 0.5648908050061436, 'learning_rate/qwen_vl_interface': 9.994362163804202e-06, 'learning_rate/action_model': 9.9943621638042e-05, 'epoch': 0.01}) +[2026-10-03 14:39:47,827][starVLA.training.train_starvla][INFO] - Step 177, Loss: {'train/loss': 0.8426671624183655, 'train/grad_norm_pre_clip': 160.90919494628906, 'timing/data': 0.00031957999453879893, 'timing/model': 0.370804947000579, 'learning_rate/qwen_vl_interface': 9.994212853946092e-06, 'learning_rate/action_model': 9.994212853946092e-05, 'epoch': 0.01}) +[2026-10-03 14:39:48,487][starVLA.training.train_starvla][INFO] - Step 178, Loss: {'train/loss': 0.3163374662399292, 'train/grad_norm_pre_clip': 22.468154907226562, 'timing/data': 0.0003366900054970756, 'timing/model': 0.3707966369984206, 'learning_rate/qwen_vl_interface': 9.994061593921283e-06, 'learning_rate/action_model': 9.994061593921282e-05, 'epoch': 0.01}) +[2026-10-03 14:39:49,149][starVLA.training.train_starvla][INFO] - Step 179, Loss: {'train/loss': 0.4658673107624054, 'train/grad_norm_pre_clip': 21.490507125854492, 'timing/data': 0.000288480005110614, 'timing/model': 0.3715692969999509, 'learning_rate/qwen_vl_interface': 9.993908383791951e-06, 'learning_rate/action_model': 9.993908383791951e-05, 'epoch': 0.01}) +[2026-10-03 14:39:49,812][starVLA.training.train_starvla][INFO] - Step 180, Loss: {'train/loss': 0.38106387853622437, 'train/grad_norm_pre_clip': 68.56985473632812, 'timing/data': 0.00032923000981099904, 'timing/model': 0.37209241700475104, 'learning_rate/qwen_vl_interface': 9.993753223621076e-06, 'learning_rate/action_model': 9.993753223621077e-05, 'epoch': 0.01}) +[2026-10-03 14:39:50,662][starVLA.training.train_starvla][INFO] - Step 181, Loss: {'train/loss': 0.6773210763931274, 'train/grad_norm_pre_clip': 63.50666427612305, 'timing/data': 0.0002479899994796142, 'timing/model': 0.5607633750041714, 'learning_rate/qwen_vl_interface': 9.99359611347244e-06, 'learning_rate/action_model': 9.993596113472439e-05, 'epoch': 0.01}) +[2026-10-03 14:39:51,301][starVLA.training.train_starvla][INFO] - Step 182, Loss: {'train/loss': 0.5388972163200378, 'train/grad_norm_pre_clip': 64.93555450439453, 'timing/data': 0.0003403699956834316, 'timing/model': 0.3474882699956652, 'learning_rate/qwen_vl_interface': 9.99343705341062e-06, 'learning_rate/action_model': 9.99343705341062e-05, 'epoch': 0.01}) +[2026-10-03 14:39:51,961][starVLA.training.train_starvla][INFO] - Step 183, Loss: {'train/loss': 0.6793222427368164, 'train/grad_norm_pre_clip': 77.19988250732422, 'timing/data': 0.0002659399906406179, 'timing/model': 0.37099333699734416, 'learning_rate/qwen_vl_interface': 9.993276043501004e-06, 'learning_rate/action_model': 9.993276043501004e-05, 'epoch': 0.01}) +[2026-10-03 14:39:52,624][starVLA.training.train_starvla][INFO] - Step 184, Loss: {'train/loss': 0.4877973794937134, 'train/grad_norm_pre_clip': 58.693687438964844, 'timing/data': 0.00021636999736074358, 'timing/model': 0.37223125800665, 'learning_rate/qwen_vl_interface': 9.993113083809776e-06, 'learning_rate/action_model': 9.993113083809775e-05, 'epoch': 0.01}) +[2026-10-03 14:39:53,285][starVLA.training.train_starvla][INFO] - Step 185, Loss: {'train/loss': 0.8369776606559753, 'train/grad_norm_pre_clip': 83.40055847167969, 'timing/data': 0.00023239999427460134, 'timing/model': 0.37092411699995864, 'learning_rate/qwen_vl_interface': 9.992948174403921e-06, 'learning_rate/action_model': 9.992948174403922e-05, 'epoch': 0.01}) +[2026-10-03 14:39:53,944][starVLA.training.train_starvla][INFO] - Step 186, Loss: {'train/loss': 0.6295688152313232, 'train/grad_norm_pre_clip': 69.44652557373047, 'timing/data': 0.0003458900027908385, 'timing/model': 0.37008924801193643, 'learning_rate/qwen_vl_interface': 9.99278131535123e-06, 'learning_rate/action_model': 9.992781315351229e-05, 'epoch': 0.01}) +[2026-10-03 14:39:54,797][starVLA.training.train_starvla][INFO] - Step 187, Loss: {'train/loss': 1.52666175365448, 'train/grad_norm_pre_clip': 216.8976287841797, 'timing/data': 0.00028541999927256256, 'timing/model': 0.3706893480120925, 'learning_rate/qwen_vl_interface': 9.992612506720289e-06, 'learning_rate/action_model': 9.99261250672029e-05, 'epoch': 0.01}) +[2026-10-03 14:39:55,458][starVLA.training.train_starvla][INFO] - Step 188, Loss: {'train/loss': 1.1532442569732666, 'train/grad_norm_pre_clip': 324.14349365234375, 'timing/data': 0.0002700400073081255, 'timing/model': 0.3716832969948882, 'learning_rate/qwen_vl_interface': 9.992441748580492e-06, 'learning_rate/action_model': 9.99244174858049e-05, 'epoch': 0.01}) +[2026-10-03 14:39:56,118][starVLA.training.train_starvla][INFO] - Step 189, Loss: {'train/loss': 0.8980607390403748, 'train/grad_norm_pre_clip': 159.49583435058594, 'timing/data': 0.00025911000557243824, 'timing/model': 0.3709111870120978, 'learning_rate/qwen_vl_interface': 9.992269041002028e-06, 'learning_rate/action_model': 9.992269041002027e-05, 'epoch': 0.01}) +[2026-10-03 14:39:56,777][starVLA.training.train_starvla][INFO] - Step 190, Loss: {'train/loss': 0.5811969041824341, 'train/grad_norm_pre_clip': 37.74040603637695, 'timing/data': 0.00026498999795876443, 'timing/model': 0.36979086798965, 'learning_rate/qwen_vl_interface': 9.992094384055894e-06, 'learning_rate/action_model': 9.992094384055895e-05, 'epoch': 0.01}) +[2026-10-03 14:39:57,413][starVLA.training.train_starvla][INFO] - Step 191, Loss: {'train/loss': 0.998254656791687, 'train/grad_norm_pre_clip': 68.01127624511719, 'timing/data': 0.0002663400082383305, 'timing/model': 0.37030448800942395, 'learning_rate/qwen_vl_interface': 9.991917777813884e-06, 'learning_rate/action_model': 9.991917777813884e-05, 'epoch': 0.01}) +[2026-10-03 14:39:58,261][starVLA.training.train_starvla][INFO] - Step 192, Loss: {'train/loss': 0.9334641695022583, 'train/grad_norm_pre_clip': 44.81489944458008, 'timing/data': 0.00034281999978702515, 'timing/model': 0.5587009560113074, 'learning_rate/qwen_vl_interface': 9.991739222348593e-06, 'learning_rate/action_model': 9.991739222348592e-05, 'epoch': 0.01}) +[2026-10-03 14:39:58,927][starVLA.training.train_starvla][INFO] - Step 193, Loss: {'train/loss': 0.5826798677444458, 'train/grad_norm_pre_clip': 33.75642776489258, 'timing/data': 0.00025645999994594604, 'timing/model': 0.37447672699636314, 'learning_rate/qwen_vl_interface': 9.991558717733418e-06, 'learning_rate/action_model': 9.991558717733417e-05, 'epoch': 0.01}) +[2026-10-03 14:39:59,584][starVLA.training.train_starvla][INFO] - Step 194, Loss: {'train/loss': 0.6015084981918335, 'train/grad_norm_pre_clip': 69.80558013916016, 'timing/data': 0.0003815600066445768, 'timing/model': 0.3698142969951732, 'learning_rate/qwen_vl_interface': 9.99137626404256e-06, 'learning_rate/action_model': 9.991376264042559e-05, 'epoch': 0.01}) +[2026-10-03 14:40:00,245][starVLA.training.train_starvla][INFO] - Step 195, Loss: {'train/loss': 0.9326499104499817, 'train/grad_norm_pre_clip': 29.583438873291016, 'timing/data': 0.0003288200095994398, 'timing/model': 0.3709241370088421, 'learning_rate/qwen_vl_interface': 9.991191861351015e-06, 'learning_rate/action_model': 9.991191861351014e-05, 'epoch': 0.01}) +[2026-10-03 14:40:00,884][starVLA.training.train_starvla][INFO] - Step 196, Loss: {'train/loss': 0.5881235003471375, 'train/grad_norm_pre_clip': 27.129398345947266, 'timing/data': 0.00033379999513272196, 'timing/model': 0.3720681269915076, 'learning_rate/qwen_vl_interface': 9.991005509734587e-06, 'learning_rate/action_model': 9.991005509734587e-05, 'epoch': 0.01}) +[2026-10-03 14:40:01,734][starVLA.training.train_starvla][INFO] - Step 197, Loss: {'train/loss': 1.0866763591766357, 'train/grad_norm_pre_clip': 51.99720764160156, 'timing/data': 0.00022987999545875937, 'timing/model': 0.5599913249898236, 'learning_rate/qwen_vl_interface': 9.990817209269878e-06, 'learning_rate/action_model': 9.990817209269878e-05, 'epoch': 0.01}) +[2026-10-03 14:40:02,369][starVLA.training.train_starvla][INFO] - Step 198, Loss: {'train/loss': 0.2872384488582611, 'train/grad_norm_pre_clip': 15.375273704528809, 'timing/data': 0.0002895200013881549, 'timing/model': 0.36995223799021915, 'learning_rate/qwen_vl_interface': 9.990626960034291e-06, 'learning_rate/action_model': 9.990626960034291e-05, 'epoch': 0.01}) +[2026-10-03 14:40:03,029][starVLA.training.train_starvla][INFO] - Step 199, Loss: {'train/loss': 1.5224876403808594, 'train/grad_norm_pre_clip': 71.6623306274414, 'timing/data': 0.0002690399996936321, 'timing/model': 0.3706068080064142, 'learning_rate/qwen_vl_interface': 9.990434762106029e-06, 'learning_rate/action_model': 9.990434762106028e-05, 'epoch': 0.01}) +[2026-10-03 14:40:03,689][starVLA.training.train_starvla][INFO] - Step 200, Loss: {'train/loss': 1.116758942604065, 'train/grad_norm_pre_clip': 44.46213150024414, 'timing/data': 0.0002674999996088445, 'timing/model': 0.3707636469916906, 'learning_rate/qwen_vl_interface': 9.990240615564099e-06, 'learning_rate/action_model': 9.990240615564099e-05, 'epoch': 0.01}) +[2026-10-03 14:40:04,327][starVLA.training.train_starvla][INFO] - Step 201, Loss: {'train/loss': 1.055572748184204, 'train/grad_norm_pre_clip': 48.124046325683594, 'timing/data': 0.00034683999547269195, 'timing/model': 0.3476195999974152, 'learning_rate/qwen_vl_interface': 9.990044520488305e-06, 'learning_rate/action_model': 9.990044520488305e-05, 'epoch': 0.01}) +[2026-10-03 14:40:04,985][starVLA.training.train_starvla][INFO] - Step 202, Loss: {'train/loss': 0.6811264753341675, 'train/grad_norm_pre_clip': 20.23171043395996, 'timing/data': 0.00025702999846544117, 'timing/model': 0.37023703700106125, 'learning_rate/qwen_vl_interface': 9.989846476959258e-06, 'learning_rate/action_model': 9.989846476959257e-05, 'epoch': 0.01}) +[2026-10-03 14:40:05,840][starVLA.training.train_starvla][INFO] - Step 203, Loss: {'train/loss': 0.48589611053466797, 'train/grad_norm_pre_clip': 31.93616485595703, 'timing/data': 0.0003360400005476549, 'timing/model': 0.3708262870059116, 'learning_rate/qwen_vl_interface': 9.989646485058361e-06, 'learning_rate/action_model': 9.989646485058361e-05, 'epoch': 0.01}) +[2026-10-03 14:40:06,501][starVLA.training.train_starvla][INFO] - Step 204, Loss: {'train/loss': 0.5922815203666687, 'train/grad_norm_pre_clip': 48.416473388671875, 'timing/data': 0.00032636999094393104, 'timing/model': 0.3713106569921365, 'learning_rate/qwen_vl_interface': 9.98944454486783e-06, 'learning_rate/action_model': 9.989444544867829e-05, 'epoch': 0.01}) +[2026-10-03 14:40:07,162][starVLA.training.train_starvla][INFO] - Step 205, Loss: {'train/loss': 0.7310882210731506, 'train/grad_norm_pre_clip': 43.58491897583008, 'timing/data': 0.00037413999962154776, 'timing/model': 0.3712211980018765, 'learning_rate/qwen_vl_interface': 9.989240656470667e-06, 'learning_rate/action_model': 9.989240656470666e-05, 'epoch': 0.02}) +[2026-10-03 14:40:07,822][starVLA.training.train_starvla][INFO] - Step 206, Loss: {'train/loss': 0.6244391798973083, 'train/grad_norm_pre_clip': 28.147432327270508, 'timing/data': 0.00021755001216661185, 'timing/model': 0.3707039779983461, 'learning_rate/qwen_vl_interface': 9.98903481995069e-06, 'learning_rate/action_model': 9.989034819950689e-05, 'epoch': 0.02}) +[2026-10-03 14:40:08,483][starVLA.training.train_starvla][INFO] - Step 207, Loss: {'train/loss': 0.3382120132446289, 'train/grad_norm_pre_clip': 102.2702865600586, 'timing/data': 0.00026635000540409237, 'timing/model': 0.3713830880005844, 'learning_rate/qwen_vl_interface': 9.988827035392506e-06, 'learning_rate/action_model': 9.988827035392507e-05, 'epoch': 0.02}) +[2026-10-03 14:40:09,334][starVLA.training.train_starvla][INFO] - Step 208, Loss: {'train/loss': 0.3794410824775696, 'train/grad_norm_pre_clip': 15.731239318847656, 'timing/data': 0.0003336200024932623, 'timing/model': 0.3718764469958842, 'learning_rate/qwen_vl_interface': 9.988617302881529e-06, 'learning_rate/action_model': 9.988617302881529e-05, 'epoch': 0.02}) +[2026-10-03 14:40:09,993][starVLA.training.train_starvla][INFO] - Step 209, Loss: {'train/loss': 0.5573194622993469, 'train/grad_norm_pre_clip': 18.173595428466797, 'timing/data': 0.0003205299872206524, 'timing/model': 0.3703356670011999, 'learning_rate/qwen_vl_interface': 9.988405622503972e-06, 'learning_rate/action_model': 9.988405622503971e-05, 'epoch': 0.02}) +[2026-10-03 14:40:10,652][starVLA.training.train_starvla][INFO] - Step 210, Loss: {'train/loss': 0.2551460862159729, 'train/grad_norm_pre_clip': 14.253714561462402, 'timing/data': 0.00023883000540081412, 'timing/model': 0.3699884369998472, 'learning_rate/qwen_vl_interface': 9.988191994346848e-06, 'learning_rate/action_model': 9.988191994346848e-05, 'epoch': 0.02}) +[2026-10-03 14:40:11,311][starVLA.training.train_starvla][INFO] - Step 211, Loss: {'train/loss': 0.43077677488327026, 'train/grad_norm_pre_clip': 30.38359832763672, 'timing/data': 0.00032762999762780964, 'timing/model': 0.3705253970110789, 'learning_rate/qwen_vl_interface': 9.987976418497974e-06, 'learning_rate/action_model': 9.987976418497972e-05, 'epoch': 0.02}) +[2026-10-03 14:40:11,971][starVLA.training.train_starvla][INFO] - Step 212, Loss: {'train/loss': 0.29372525215148926, 'train/grad_norm_pre_clip': 21.642221450805664, 'timing/data': 0.0003350199986016378, 'timing/model': 0.37106579598912504, 'learning_rate/qwen_vl_interface': 9.987758895045964e-06, 'learning_rate/action_model': 9.987758895045963e-05, 'epoch': 0.02}) +[2026-10-03 14:40:12,824][starVLA.training.train_starvla][INFO] - Step 213, Loss: {'train/loss': 0.4027504026889801, 'train/grad_norm_pre_clip': 21.869007110595703, 'timing/data': 0.00025958000333048403, 'timing/model': 0.5627928549947683, 'learning_rate/qwen_vl_interface': 9.98753942408023e-06, 'learning_rate/action_model': 9.98753942408023e-05, 'epoch': 0.02}) +[2026-10-03 14:40:13,483][starVLA.training.train_starvla][INFO] - Step 214, Loss: {'train/loss': 1.6760990619659424, 'train/grad_norm_pre_clip': 109.90349578857422, 'timing/data': 0.0002532199869165197, 'timing/model': 0.37052096799016, 'learning_rate/qwen_vl_interface': 9.987318005690992e-06, 'learning_rate/action_model': 9.987318005690992e-05, 'epoch': 0.02}) +[2026-10-03 14:40:14,143][starVLA.training.train_starvla][INFO] - Step 215, Loss: {'train/loss': 0.633600115776062, 'train/grad_norm_pre_clip': 64.01139831542969, 'timing/data': 0.00026773998979479074, 'timing/model': 0.37119633800466545, 'learning_rate/qwen_vl_interface': 9.987094639969268e-06, 'learning_rate/action_model': 9.987094639969268e-05, 'epoch': 0.02}) +[2026-10-03 14:40:14,781][starVLA.training.train_starvla][INFO] - Step 216, Loss: {'train/loss': 0.2734591066837311, 'train/grad_norm_pre_clip': 24.845458984375, 'timing/data': 0.00031879999733064324, 'timing/model': 0.3478403100016294, 'learning_rate/qwen_vl_interface': 9.986869327006872e-06, 'learning_rate/action_model': 9.986869327006873e-05, 'epoch': 0.02}) +[2026-10-03 14:40:15,441][starVLA.training.train_starvla][INFO] - Step 217, Loss: {'train/loss': 0.39064961671829224, 'train/grad_norm_pre_clip': 16.541913986206055, 'timing/data': 0.00026127000455744565, 'timing/model': 0.3708750669902656, 'learning_rate/qwen_vl_interface': 9.986642066896424e-06, 'learning_rate/action_model': 9.986642066896423e-05, 'epoch': 0.02}) +[2026-10-03 14:40:16,295][starVLA.training.train_starvla][INFO] - Step 218, Loss: {'train/loss': 0.8805784583091736, 'train/grad_norm_pre_clip': 41.441184997558594, 'timing/data': 0.0003347799938637763, 'timing/model': 0.5652178640011698, 'learning_rate/qwen_vl_interface': 9.986412859731341e-06, 'learning_rate/action_model': 9.98641285973134e-05, 'epoch': 0.02}) +[2026-10-03 14:40:16,955][starVLA.training.train_starvla][INFO] - Step 219, Loss: {'train/loss': 0.5485224723815918, 'train/grad_norm_pre_clip': 27.812875747680664, 'timing/data': 0.0002700100012589246, 'timing/model': 0.37060462700901553, 'learning_rate/qwen_vl_interface': 9.986181705605839e-06, 'learning_rate/action_model': 9.986181705605839e-05, 'epoch': 0.02}) +[2026-10-03 14:40:17,615][starVLA.training.train_starvla][INFO] - Step 220, Loss: {'train/loss': 0.7892169952392578, 'train/grad_norm_pre_clip': 37.388099670410156, 'timing/data': 0.0002639199956320226, 'timing/model': 0.3709497969975928, 'learning_rate/qwen_vl_interface': 9.98594860461494e-06, 'learning_rate/action_model': 9.98594860461494e-05, 'epoch': 0.02}) +[2026-10-03 14:40:18,275][starVLA.training.train_starvla][INFO] - Step 221, Loss: {'train/loss': 0.49175184965133667, 'train/grad_norm_pre_clip': 102.01783752441406, 'timing/data': 0.0003020100120920688, 'timing/model': 0.3709239570016507, 'learning_rate/qwen_vl_interface': 9.985713556854461e-06, 'learning_rate/action_model': 9.98571355685446e-05, 'epoch': 0.02}) +[2026-10-03 14:40:18,934][starVLA.training.train_starvla][INFO] - Step 222, Loss: {'train/loss': 0.9906257390975952, 'train/grad_norm_pre_clip': 57.48919677734375, 'timing/data': 0.00025414000265300274, 'timing/model': 0.3700881569966441, 'learning_rate/qwen_vl_interface': 9.985476562421024e-06, 'learning_rate/action_model': 9.985476562421024e-05, 'epoch': 0.02}) +[2026-10-03 14:40:19,593][starVLA.training.train_starvla][INFO] - Step 223, Loss: {'train/loss': 0.34860527515411377, 'train/grad_norm_pre_clip': 71.62010192871094, 'timing/data': 0.0003245699917897582, 'timing/model': 0.3706919870019192, 'learning_rate/qwen_vl_interface': 9.985237621412047e-06, 'learning_rate/action_model': 9.985237621412046e-05, 'epoch': 0.02}) +[2026-10-03 14:40:20,453][starVLA.training.train_starvla][INFO] - Step 224, Loss: {'train/loss': 0.4438958764076233, 'train/grad_norm_pre_clip': 30.063711166381836, 'timing/data': 0.0004022389912279323, 'timing/model': 0.37146663801104296, 'learning_rate/qwen_vl_interface': 9.984996733925749e-06, 'learning_rate/action_model': 9.984996733925748e-05, 'epoch': 0.02}) +[2026-10-03 14:40:21,114][starVLA.training.train_starvla][INFO] - Step 225, Loss: {'train/loss': 0.41177913546562195, 'train/grad_norm_pre_clip': 36.10358428955078, 'timing/data': 0.0003055099950870499, 'timing/model': 0.37080908700590953, 'learning_rate/qwen_vl_interface': 9.98475390006115e-06, 'learning_rate/action_model': 9.98475390006115e-05, 'epoch': 0.02}) +[2026-10-03 14:40:21,749][starVLA.training.train_starvla][INFO] - Step 226, Loss: {'train/loss': 0.6491707563400269, 'train/grad_norm_pre_clip': 49.83846664428711, 'timing/data': 0.0002983900049002841, 'timing/model': 0.3473353400040651, 'learning_rate/qwen_vl_interface': 9.98450911991807e-06, 'learning_rate/action_model': 9.98450911991807e-05, 'epoch': 0.02}) +[2026-10-03 14:40:22,409][starVLA.training.train_starvla][INFO] - Step 227, Loss: {'train/loss': 0.267929345369339, 'train/grad_norm_pre_clip': 13.604085922241211, 'timing/data': 0.00037658998917322606, 'timing/model': 0.37075205698783975, 'learning_rate/qwen_vl_interface': 9.98426239359713e-06, 'learning_rate/action_model': 9.984262393597129e-05, 'epoch': 0.02}) +[2026-10-03 14:40:23,046][starVLA.training.train_starvla][INFO] - Step 228, Loss: {'train/loss': 0.3142428994178772, 'train/grad_norm_pre_clip': 13.786069869995117, 'timing/data': 0.0002662699989741668, 'timing/model': 0.3481266690068878, 'learning_rate/qwen_vl_interface': 9.984013721199747e-06, 'learning_rate/action_model': 9.984013721199747e-05, 'epoch': 0.02}) +[2026-10-03 14:40:23,896][starVLA.training.train_starvla][INFO] - Step 229, Loss: {'train/loss': 0.2303813248872757, 'train/grad_norm_pre_clip': 15.16182804107666, 'timing/data': 0.00025750999338924885, 'timing/model': 0.559338115999708, 'learning_rate/qwen_vl_interface': 9.983763102828143e-06, 'learning_rate/action_model': 9.983763102828143e-05, 'epoch': 0.02}) +[2026-10-03 14:40:24,555][starVLA.training.train_starvla][INFO] - Step 230, Loss: {'train/loss': 0.3174842298030853, 'train/grad_norm_pre_clip': 64.25175476074219, 'timing/data': 0.0002467700105626136, 'timing/model': 0.3705558569927234, 'learning_rate/qwen_vl_interface': 9.983510538585336e-06, 'learning_rate/action_model': 9.983510538585337e-05, 'epoch': 0.02}) +[2026-10-03 14:40:25,214][starVLA.training.train_starvla][INFO] - Step 231, Loss: {'train/loss': 0.4138867259025574, 'train/grad_norm_pre_clip': 29.249794006347656, 'timing/data': 0.00035942000977229327, 'timing/model': 0.3706613669928629, 'learning_rate/qwen_vl_interface': 9.98325602857515e-06, 'learning_rate/action_model': 9.98325602857515e-05, 'epoch': 0.02}) +[2026-10-03 14:40:25,875][starVLA.training.train_starvla][INFO] - Step 232, Loss: {'train/loss': 0.6202434301376343, 'train/grad_norm_pre_clip': 55.34011459350586, 'timing/data': 0.0003340600087540224, 'timing/model': 0.3716655169992009, 'learning_rate/qwen_vl_interface': 9.9829995729022e-06, 'learning_rate/action_model': 9.982999572902199e-05, 'epoch': 0.02}) +[2026-10-03 14:40:26,537][starVLA.training.train_starvla][INFO] - Step 233, Loss: {'train/loss': 0.7050630450248718, 'train/grad_norm_pre_clip': 62.93335723876953, 'timing/data': 0.0003519099991535768, 'timing/model': 0.3715306569938548, 'learning_rate/qwen_vl_interface': 9.982741171671905e-06, 'learning_rate/action_model': 9.982741171671906e-05, 'epoch': 0.02}) +[2026-10-03 14:40:27,197][starVLA.training.train_starvla][INFO] - Step 234, Loss: {'train/loss': 0.34887710213661194, 'train/grad_norm_pre_clip': 28.952407836914062, 'timing/data': 0.00030962999153416604, 'timing/model': 0.3706953370128758, 'learning_rate/qwen_vl_interface': 9.982480824990489e-06, 'learning_rate/action_model': 9.982480824990488e-05, 'epoch': 0.02}) +[2026-10-03 14:40:28,049][starVLA.training.train_starvla][INFO] - Step 235, Loss: {'train/loss': 0.8397675156593323, 'train/grad_norm_pre_clip': 160.3447723388672, 'timing/data': 0.00029893999453634024, 'timing/model': 0.37119885699939914, 'learning_rate/qwen_vl_interface': 9.982218532964963e-06, 'learning_rate/action_model': 9.982218532964964e-05, 'epoch': 0.02}) +[2026-10-03 14:40:28,712][starVLA.training.train_starvla][INFO] - Step 236, Loss: {'train/loss': 0.2964070737361908, 'train/grad_norm_pre_clip': 29.52802276611328, 'timing/data': 0.00033707999682519585, 'timing/model': 0.37191643699770793, 'learning_rate/qwen_vl_interface': 9.981954295703152e-06, 'learning_rate/action_model': 9.981954295703152e-05, 'epoch': 0.02}) +[2026-10-03 14:40:29,373][starVLA.training.train_starvla][INFO] - Step 237, Loss: {'train/loss': 0.16279280185699463, 'train/grad_norm_pre_clip': 22.314319610595703, 'timing/data': 0.00031615000625606626, 'timing/model': 0.3717913170112297, 'learning_rate/qwen_vl_interface': 9.981688113313672e-06, 'learning_rate/action_model': 9.981688113313672e-05, 'epoch': 0.02}) +[2026-10-03 14:40:30,033][starVLA.training.train_starvla][INFO] - Step 238, Loss: {'train/loss': 0.44482868909835815, 'train/grad_norm_pre_clip': 61.761505126953125, 'timing/data': 0.0003283990081399679, 'timing/model': 0.3705277080007363, 'learning_rate/qwen_vl_interface': 9.98141998590594e-06, 'learning_rate/action_model': 9.981419985905939e-05, 'epoch': 0.02}) +[2026-10-03 14:40:30,695][starVLA.training.train_starvla][INFO] - Step 239, Loss: {'train/loss': 0.6598262786865234, 'train/grad_norm_pre_clip': 208.44305419921875, 'timing/data': 0.0003927200014004484, 'timing/model': 0.37193041799764615, 'learning_rate/qwen_vl_interface': 9.981149913590171e-06, 'learning_rate/action_model': 9.981149913590172e-05, 'epoch': 0.02}) +[2026-10-03 14:40:31,552][starVLA.training.train_starvla][INFO] - Step 240, Loss: {'train/loss': 0.2702314257621765, 'train/grad_norm_pre_clip': 45.763092041015625, 'timing/data': 0.00040624999382998794, 'timing/model': 0.3718849569995655, 'learning_rate/qwen_vl_interface': 9.980877896477386e-06, 'learning_rate/action_model': 9.980877896477386e-05, 'epoch': 0.02}) +[2026-10-03 14:40:32,214][starVLA.training.train_starvla][INFO] - Step 241, Loss: {'train/loss': 0.7449960708618164, 'train/grad_norm_pre_clip': 29.636455535888672, 'timing/data': 0.00021705000835936517, 'timing/model': 0.37155289700604044, 'learning_rate/qwen_vl_interface': 9.980603934679397e-06, 'learning_rate/action_model': 9.980603934679398e-05, 'epoch': 0.02}) +[2026-10-03 14:40:32,873][starVLA.training.train_starvla][INFO] - Step 242, Loss: {'train/loss': 0.4136035442352295, 'train/grad_norm_pre_clip': 62.28675079345703, 'timing/data': 0.0002634699922055006, 'timing/model': 0.37049956699775066, 'learning_rate/qwen_vl_interface': 9.980328028308824e-06, 'learning_rate/action_model': 9.980328028308824e-05, 'epoch': 0.02}) +[2026-10-03 14:40:33,534][starVLA.training.train_starvla][INFO] - Step 243, Loss: {'train/loss': 0.6384734511375427, 'train/grad_norm_pre_clip': 139.02874755859375, 'timing/data': 0.0002622600004542619, 'timing/model': 0.37153832700278144, 'learning_rate/qwen_vl_interface': 9.980050177479075e-06, 'learning_rate/action_model': 9.980050177479076e-05, 'epoch': 0.02}) +[2026-10-03 14:40:34,172][starVLA.training.train_starvla][INFO] - Step 244, Loss: {'train/loss': 0.21707679331302643, 'train/grad_norm_pre_clip': 21.04198455810547, 'timing/data': 0.00030758998764213175, 'timing/model': 0.3478281599964248, 'learning_rate/qwen_vl_interface': 9.97977038230437e-06, 'learning_rate/action_model': 9.97977038230437e-05, 'epoch': 0.02}) +[2026-10-03 14:40:35,028][starVLA.training.train_starvla][INFO] - Step 245, Loss: {'train/loss': 0.5471516251564026, 'train/grad_norm_pre_clip': 135.4677734375, 'timing/data': 0.0003343599964864552, 'timing/model': 0.5659027350047836, 'learning_rate/qwen_vl_interface': 9.97948864289972e-06, 'learning_rate/action_model': 9.97948864289972e-05, 'epoch': 0.02}) +[2026-10-03 14:40:35,689][starVLA.training.train_starvla][INFO] - Step 246, Loss: {'train/loss': 0.9414273500442505, 'train/grad_norm_pre_clip': 47.6802864074707, 'timing/data': 0.00021958998695481569, 'timing/model': 0.3714775480038952, 'learning_rate/qwen_vl_interface': 9.979204959380936e-06, 'learning_rate/action_model': 9.979204959380936e-05, 'epoch': 0.02}) +[2026-10-03 14:40:36,350][starVLA.training.train_starvla][INFO] - Step 247, Loss: {'train/loss': 0.15874269604682922, 'train/grad_norm_pre_clip': 47.92660903930664, 'timing/data': 0.00035382999340072274, 'timing/model': 0.37101369799347594, 'learning_rate/qwen_vl_interface': 9.978919331864631e-06, 'learning_rate/action_model': 9.97891933186463e-05, 'epoch': 0.02}) +[2026-10-03 14:40:37,011][starVLA.training.train_starvla][INFO] - Step 248, Loss: {'train/loss': 0.1999008059501648, 'train/grad_norm_pre_clip': 18.69122886657715, 'timing/data': 0.0002658300072653219, 'timing/model': 0.37207829700491857, 'learning_rate/qwen_vl_interface': 9.978631760468216e-06, 'learning_rate/action_model': 9.978631760468216e-05, 'epoch': 0.02}) +[2026-10-03 14:40:37,673][starVLA.training.train_starvla][INFO] - Step 249, Loss: {'train/loss': 0.4345836341381073, 'train/grad_norm_pre_clip': 40.83747482299805, 'timing/data': 0.0003579699987312779, 'timing/model': 0.37120843699085526, 'learning_rate/qwen_vl_interface': 9.9783422453099e-06, 'learning_rate/action_model': 9.9783422453099e-05, 'epoch': 0.02}) +[2026-10-03 14:40:38,521][starVLA.training.train_starvla][INFO] - Step 250, Loss: {'train/loss': 0.5954380631446838, 'train/grad_norm_pre_clip': 73.44300842285156, 'timing/data': 0.00032002000079955906, 'timing/model': 0.557905485999072, 'learning_rate/qwen_vl_interface': 9.978050786508693e-06, 'learning_rate/action_model': 9.978050786508692e-05, 'epoch': 0.02}) +[2026-10-03 14:40:39,181][starVLA.training.train_starvla][INFO] - Step 251, Loss: {'train/loss': 0.684288740158081, 'train/grad_norm_pre_clip': 90.80101776123047, 'timing/data': 0.00023157999385148287, 'timing/model': 0.37127049699483905, 'learning_rate/qwen_vl_interface': 9.9777573841844e-06, 'learning_rate/action_model': 9.977757384184401e-05, 'epoch': 0.02}) +[2026-10-03 14:40:39,842][starVLA.training.train_starvla][INFO] - Step 252, Loss: {'train/loss': 0.40452563762664795, 'train/grad_norm_pre_clip': 31.00112533569336, 'timing/data': 0.0003306199942016974, 'timing/model': 0.3713117470033467, 'learning_rate/qwen_vl_interface': 9.97746203845763e-06, 'learning_rate/action_model': 9.97746203845763e-05, 'epoch': 0.02}) +[2026-10-03 14:40:40,503][starVLA.training.train_starvla][INFO] - Step 253, Loss: {'train/loss': 0.4128672778606415, 'train/grad_norm_pre_clip': 76.22119903564453, 'timing/data': 0.00025753999943844974, 'timing/model': 0.3709507069870597, 'learning_rate/qwen_vl_interface': 9.97716474944979e-06, 'learning_rate/action_model': 9.977164749449788e-05, 'epoch': 0.02}) +[2026-10-03 14:40:41,164][starVLA.training.train_starvla][INFO] - Step 254, Loss: {'train/loss': 0.4009576141834259, 'train/grad_norm_pre_clip': 44.2381706237793, 'timing/data': 0.0003241399972466752, 'timing/model': 0.3708065669925418, 'learning_rate/qwen_vl_interface': 9.976865517283079e-06, 'learning_rate/action_model': 9.976865517283078e-05, 'epoch': 0.02}) +[2026-10-03 14:40:41,824][starVLA.training.train_starvla][INFO] - Step 255, Loss: {'train/loss': 0.03831208869814873, 'train/grad_norm_pre_clip': 5.174824237823486, 'timing/data': 0.00025986001128330827, 'timing/model': 0.3708195869985502, 'learning_rate/qwen_vl_interface': 9.976564342080505e-06, 'learning_rate/action_model': 9.976564342080504e-05, 'epoch': 0.02}) +[2026-10-03 14:40:42,656][starVLA.training.train_starvla][INFO] - Step 256, Loss: {'train/loss': 0.3187958598136902, 'train/grad_norm_pre_clip': 32.82518768310547, 'timing/data': 0.00034783000592142344, 'timing/model': 0.37145867699291557, 'learning_rate/qwen_vl_interface': 9.976261223965867e-06, 'learning_rate/action_model': 9.976261223965865e-05, 'epoch': 0.02}) +[2026-10-03 14:40:43,293][starVLA.training.train_starvla][INFO] - Step 257, Loss: {'train/loss': 0.5866192579269409, 'train/grad_norm_pre_clip': 54.26447296142578, 'timing/data': 0.0003444500034675002, 'timing/model': 0.3477933599933749, 'learning_rate/qwen_vl_interface': 9.975956163063767e-06, 'learning_rate/action_model': 9.975956163063767e-05, 'epoch': 0.02}) +[2026-10-03 14:40:43,953][starVLA.training.train_starvla][INFO] - Step 258, Loss: {'train/loss': 0.4046175479888916, 'train/grad_norm_pre_clip': 39.426631927490234, 'timing/data': 0.00034698999661486596, 'timing/model': 0.3709658770094393, 'learning_rate/qwen_vl_interface': 9.975649159499605e-06, 'learning_rate/action_model': 9.975649159499603e-05, 'epoch': 0.02}) +[2026-10-03 14:40:44,614][starVLA.training.train_starvla][INFO] - Step 259, Loss: {'train/loss': 0.4177672863006592, 'train/grad_norm_pre_clip': 75.69773864746094, 'timing/data': 0.0002632000105222687, 'timing/model': 0.3712628179928288, 'learning_rate/qwen_vl_interface': 9.975340213399573e-06, 'learning_rate/action_model': 9.975340213399573e-05, 'epoch': 0.02}) +[2026-10-03 14:40:45,276][starVLA.training.train_starvla][INFO] - Step 260, Loss: {'train/loss': 0.9016870260238647, 'train/grad_norm_pre_clip': 55.63126754760742, 'timing/data': 0.00035453999589663, 'timing/model': 0.37120213799062185, 'learning_rate/qwen_vl_interface': 9.975029324890674e-06, 'learning_rate/action_model': 9.975029324890674e-05, 'epoch': 0.02}) +[2026-10-03 14:40:46,128][starVLA.training.train_starvla][INFO] - Step 261, Loss: {'train/loss': 0.23526416718959808, 'train/grad_norm_pre_clip': 18.57213592529297, 'timing/data': 0.0002779000060399994, 'timing/model': 0.5617945550038712, 'learning_rate/qwen_vl_interface': 9.974716494100698e-06, 'learning_rate/action_model': 9.974716494100697e-05, 'epoch': 0.02}) +[2026-10-03 14:40:46,764][starVLA.training.train_starvla][INFO] - Step 262, Loss: {'train/loss': 0.5489764213562012, 'train/grad_norm_pre_clip': 30.045644760131836, 'timing/data': 0.00025738999829627573, 'timing/model': 0.34731286999885924, 'learning_rate/qwen_vl_interface': 9.97440172115824e-06, 'learning_rate/action_model': 9.974401721158239e-05, 'epoch': 0.02}) +[2026-10-03 14:40:47,425][starVLA.training.train_starvla][INFO] - Step 263, Loss: {'train/loss': 0.9925881028175354, 'train/grad_norm_pre_clip': 49.055259704589844, 'timing/data': 0.0002594399993540719, 'timing/model': 0.37107165799534414, 'learning_rate/qwen_vl_interface': 9.97408500619269e-06, 'learning_rate/action_model': 9.97408500619269e-05, 'epoch': 0.02}) +[2026-10-03 14:40:48,086][starVLA.training.train_starvla][INFO] - Step 264, Loss: {'train/loss': 0.5660033822059631, 'train/grad_norm_pre_clip': 127.40436553955078, 'timing/data': 0.000262809990090318, 'timing/model': 0.3717250470072031, 'learning_rate/qwen_vl_interface': 9.973766349334239e-06, 'learning_rate/action_model': 9.973766349334239e-05, 'epoch': 0.02}) +[2026-10-03 14:40:48,748][starVLA.training.train_starvla][INFO] - Step 265, Loss: {'train/loss': 0.46528953313827515, 'train/grad_norm_pre_clip': 22.521682739257812, 'timing/data': 0.00036365000414662063, 'timing/model': 0.371694818008109, 'learning_rate/qwen_vl_interface': 9.973445750713875e-06, 'learning_rate/action_model': 9.973445750713875e-05, 'epoch': 0.02}) +[2026-10-03 14:40:49,606][starVLA.training.train_starvla][INFO] - Step 266, Loss: {'train/loss': 0.5787197351455688, 'train/grad_norm_pre_clip': 57.682289123535156, 'timing/data': 0.0003873800014844164, 'timing/model': 0.37062675699417014, 'learning_rate/qwen_vl_interface': 9.97312321046338e-06, 'learning_rate/action_model': 9.97312321046338e-05, 'epoch': 0.02}) +[2026-10-03 14:40:50,267][starVLA.training.train_starvla][INFO] - Step 267, Loss: {'train/loss': 0.32689929008483887, 'train/grad_norm_pre_clip': 55.402496337890625, 'timing/data': 0.00024960999144241214, 'timing/model': 0.37135443699662574, 'learning_rate/qwen_vl_interface': 9.972798728715343e-06, 'learning_rate/action_model': 9.972798728715344e-05, 'epoch': 0.02}) +[2026-10-03 14:40:50,929][starVLA.training.train_starvla][INFO] - Step 268, Loss: {'train/loss': 0.5290351510047913, 'train/grad_norm_pre_clip': 91.55545043945312, 'timing/data': 0.00034505000803619623, 'timing/model': 0.3724940459942445, 'learning_rate/qwen_vl_interface': 9.972472305603145e-06, 'learning_rate/action_model': 9.972472305603144e-05, 'epoch': 0.02}) +[2026-10-03 14:40:51,590][starVLA.training.train_starvla][INFO] - Step 269, Loss: {'train/loss': 0.6187663674354553, 'train/grad_norm_pre_clip': 47.3998908996582, 'timing/data': 0.0002170999941881746, 'timing/model': 0.3710998470050981, 'learning_rate/qwen_vl_interface': 9.972143941260965e-06, 'learning_rate/action_model': 9.972143941260965e-05, 'epoch': 0.02}) +[2026-10-03 14:40:52,249][starVLA.training.train_starvla][INFO] - Step 270, Loss: {'train/loss': 0.9346380233764648, 'train/grad_norm_pre_clip': 41.18217086791992, 'timing/data': 0.00021136000577826053, 'timing/model': 0.3705596170038916, 'learning_rate/qwen_vl_interface': 9.971813635823781e-06, 'learning_rate/action_model': 9.971813635823781e-05, 'epoch': 0.02}) +[2026-10-03 14:40:53,116][starVLA.training.train_starvla][INFO] - Step 271, Loss: {'train/loss': 0.3372708857059479, 'train/grad_norm_pre_clip': 31.11144256591797, 'timing/data': 0.0003609100094763562, 'timing/model': 0.5774832530005369, 'learning_rate/qwen_vl_interface': 9.971481389427372e-06, 'learning_rate/action_model': 9.971481389427371e-05, 'epoch': 0.02}) +[2026-10-03 14:40:53,777][starVLA.training.train_starvla][INFO] - Step 272, Loss: {'train/loss': 0.21943846344947815, 'train/grad_norm_pre_clip': 13.802621841430664, 'timing/data': 0.00021661999926436692, 'timing/model': 0.3714519269997254, 'learning_rate/qwen_vl_interface': 9.971147202208308e-06, 'learning_rate/action_model': 9.971147202208308e-05, 'epoch': 0.02}) +[2026-10-03 14:40:54,439][starVLA.training.train_starvla][INFO] - Step 273, Loss: {'train/loss': 0.16063977777957916, 'train/grad_norm_pre_clip': 9.561921119689941, 'timing/data': 0.00036620999162551016, 'timing/model': 0.37174287700327113, 'learning_rate/qwen_vl_interface': 9.970811074303962e-06, 'learning_rate/action_model': 9.970811074303962e-05, 'epoch': 0.02}) +[2026-10-03 14:40:55,098][starVLA.training.train_starvla][INFO] - Step 274, Loss: {'train/loss': 0.7845795154571533, 'train/grad_norm_pre_clip': 52.45133590698242, 'timing/data': 0.0002529199991840869, 'timing/model': 0.3703830470039975, 'learning_rate/qwen_vl_interface': 9.970473005852505e-06, 'learning_rate/action_model': 9.970473005852505e-05, 'epoch': 0.02}) +[2026-10-03 14:40:55,735][starVLA.training.train_starvla][INFO] - Step 275, Loss: {'train/loss': 0.8673148155212402, 'train/grad_norm_pre_clip': 106.46015167236328, 'timing/data': 0.00027130999660585076, 'timing/model': 0.370469868008513, 'learning_rate/qwen_vl_interface': 9.970132996992904e-06, 'learning_rate/action_model': 9.970132996992904e-05, 'epoch': 0.02}) +[2026-10-03 14:40:56,396][starVLA.training.train_starvla][INFO] - Step 276, Loss: {'train/loss': 0.8781269788742065, 'train/grad_norm_pre_clip': 224.23422241210938, 'timing/data': 0.00033409999741706997, 'timing/model': 0.3712790169956861, 'learning_rate/qwen_vl_interface': 9.969791047864921e-06, 'learning_rate/action_model': 9.969791047864921e-05, 'epoch': 0.02}) +[2026-10-03 14:40:57,250][starVLA.training.train_starvla][INFO] - Step 277, Loss: {'train/loss': 1.5618786811828613, 'train/grad_norm_pre_clip': 96.99415588378906, 'timing/data': 0.0003635300090536475, 'timing/model': 0.3711982360109687, 'learning_rate/qwen_vl_interface': 9.969447158609122e-06, 'learning_rate/action_model': 9.96944715860912e-05, 'epoch': 0.02}) +[2026-10-03 14:40:57,910][starVLA.training.train_starvla][INFO] - Step 278, Loss: {'train/loss': 1.0399951934814453, 'train/grad_norm_pre_clip': 86.37161254882812, 'timing/data': 0.0003036599955521524, 'timing/model': 0.3706195770064369, 'learning_rate/qwen_vl_interface': 9.969101329366864e-06, 'learning_rate/action_model': 9.969101329366863e-05, 'epoch': 0.02}) +[2026-10-03 14:40:58,571][starVLA.training.train_starvla][INFO] - Step 279, Loss: {'train/loss': 0.6338579654693604, 'train/grad_norm_pre_clip': 45.18242645263672, 'timing/data': 0.00033268998959101737, 'timing/model': 0.3713502070022514, 'learning_rate/qwen_vl_interface': 9.968753560280305e-06, 'learning_rate/action_model': 9.968753560280304e-05, 'epoch': 0.02}) +[2026-10-03 14:40:59,208][starVLA.training.train_starvla][INFO] - Step 280, Loss: {'train/loss': 0.6572560667991638, 'train/grad_norm_pre_clip': 30.966388702392578, 'timing/data': 0.0002242600021418184, 'timing/model': 0.3707643070083577, 'learning_rate/qwen_vl_interface': 9.968403851492402e-06, 'learning_rate/action_model': 9.968403851492401e-05, 'epoch': 0.02}) +[2026-10-03 14:40:59,870][starVLA.training.train_starvla][INFO] - Step 281, Loss: {'train/loss': 0.661227285861969, 'train/grad_norm_pre_clip': 36.68641662597656, 'timing/data': 0.00021706000552512705, 'timing/model': 0.3710143270000117, 'learning_rate/qwen_vl_interface': 9.968052203146904e-06, 'learning_rate/action_model': 9.968052203146903e-05, 'epoch': 0.02}) +[2026-10-03 14:41:00,726][starVLA.training.train_starvla][INFO] - Step 282, Loss: {'train/loss': 0.3820602297782898, 'train/grad_norm_pre_clip': 36.211090087890625, 'timing/data': 0.00029819899646099657, 'timing/model': 0.5664508850022685, 'learning_rate/qwen_vl_interface': 9.967698615388363e-06, 'learning_rate/action_model': 9.967698615388362e-05, 'epoch': 0.02}) +[2026-10-03 14:41:01,385][starVLA.training.train_starvla][INFO] - Step 283, Loss: {'train/loss': 0.5614218711853027, 'train/grad_norm_pre_clip': 27.389421463012695, 'timing/data': 0.000315440003760159, 'timing/model': 0.370513927002321, 'learning_rate/qwen_vl_interface': 9.967343088362123e-06, 'learning_rate/action_model': 9.967343088362122e-05, 'epoch': 0.02}) +[2026-10-03 14:41:02,047][starVLA.training.train_starvla][INFO] - Step 284, Loss: {'train/loss': 0.26213109493255615, 'train/grad_norm_pre_clip': 36.2867431640625, 'timing/data': 0.0002608899958431721, 'timing/model': 0.37179831600224134, 'learning_rate/qwen_vl_interface': 9.966985622214328e-06, 'learning_rate/action_model': 9.966985622214329e-05, 'epoch': 0.02}) +[2026-10-03 14:41:02,708][starVLA.training.train_starvla][INFO] - Step 285, Loss: {'train/loss': 0.5835426449775696, 'train/grad_norm_pre_clip': 87.42507934570312, 'timing/data': 0.0003511700051603839, 'timing/model': 0.37097012699814513, 'learning_rate/qwen_vl_interface': 9.966626217091922e-06, 'learning_rate/action_model': 9.966626217091922e-05, 'epoch': 0.02}) +[2026-10-03 14:41:03,368][starVLA.training.train_starvla][INFO] - Step 286, Loss: {'train/loss': 0.454420268535614, 'train/grad_norm_pre_clip': 69.4471664428711, 'timing/data': 0.0002348300040466711, 'timing/model': 0.3706026559957536, 'learning_rate/qwen_vl_interface': 9.966264873142641e-06, 'learning_rate/action_model': 9.96626487314264e-05, 'epoch': 0.02}) +[2026-10-03 14:41:04,191][starVLA.training.train_starvla][INFO] - Step 287, Loss: {'train/loss': 0.7358395457267761, 'train/grad_norm_pre_clip': 54.04389190673828, 'timing/data': 0.0003432399971643463, 'timing/model': 0.5339728780090809, 'learning_rate/qwen_vl_interface': 9.965901590515018e-06, 'learning_rate/action_model': 9.965901590515018e-05, 'epoch': 0.02}) +[2026-10-03 14:41:04,853][starVLA.training.train_starvla][INFO] - Step 288, Loss: {'train/loss': 0.21868032217025757, 'train/grad_norm_pre_clip': 32.09323501586914, 'timing/data': 0.000311290001263842, 'timing/model': 0.37132202700013295, 'learning_rate/qwen_vl_interface': 9.965536369358386e-06, 'learning_rate/action_model': 9.965536369358386e-05, 'epoch': 0.02}) +[2026-10-03 14:41:05,513][starVLA.training.train_starvla][INFO] - Step 289, Loss: {'train/loss': 0.387648344039917, 'train/grad_norm_pre_clip': 25.00216293334961, 'timing/data': 0.000315849989419803, 'timing/model': 0.37083741699461825, 'learning_rate/qwen_vl_interface': 9.965169209822874e-06, 'learning_rate/action_model': 9.965169209822874e-05, 'epoch': 0.02}) +[2026-10-03 14:41:06,174][starVLA.training.train_starvla][INFO] - Step 290, Loss: {'train/loss': 0.13082613050937653, 'train/grad_norm_pre_clip': 7.463263988494873, 'timing/data': 0.0002697690069908276, 'timing/model': 0.37112883699592203, 'learning_rate/qwen_vl_interface': 9.96480011205941e-06, 'learning_rate/action_model': 9.964800112059409e-05, 'epoch': 0.02}) +[2026-10-03 14:41:06,812][starVLA.training.train_starvla][INFO] - Step 291, Loss: {'train/loss': 0.3243299126625061, 'train/grad_norm_pre_clip': 25.517440795898438, 'timing/data': 0.00033942000300157815, 'timing/model': 0.34836083999834955, 'learning_rate/qwen_vl_interface': 9.964429076219711e-06, 'learning_rate/action_model': 9.964429076219711e-05, 'epoch': 0.02}) +[2026-10-03 14:41:07,472][starVLA.training.train_starvla][INFO] - Step 292, Loss: {'train/loss': 0.334242045879364, 'train/grad_norm_pre_clip': 17.92633819580078, 'timing/data': 0.0003387100005056709, 'timing/model': 0.3708773979888065, 'learning_rate/qwen_vl_interface': 9.964056102456299e-06, 'learning_rate/action_model': 9.964056102456299e-05, 'epoch': 0.02}) +[2026-10-03 14:41:08,330][starVLA.training.train_starvla][INFO] - Step 293, Loss: {'train/loss': 0.1232578456401825, 'train/grad_norm_pre_clip': 33.596519470214844, 'timing/data': 0.00032873000600375235, 'timing/model': 0.37180151700158603, 'learning_rate/qwen_vl_interface': 9.963681190922489e-06, 'learning_rate/action_model': 9.963681190922488e-05, 'epoch': 0.02}) +[2026-10-03 14:41:08,966][starVLA.training.train_starvla][INFO] - Step 294, Loss: {'train/loss': 0.19773045182228088, 'train/grad_norm_pre_clip': 43.17761993408203, 'timing/data': 0.00026544999855104834, 'timing/model': 0.34702487000322435, 'learning_rate/qwen_vl_interface': 9.963304341772391e-06, 'learning_rate/action_model': 9.963304341772391e-05, 'epoch': 0.02}) +[2026-10-03 14:41:09,626][starVLA.training.train_starvla][INFO] - Step 295, Loss: {'train/loss': 0.3991280794143677, 'train/grad_norm_pre_clip': 38.19279479980469, 'timing/data': 0.0002510899939807132, 'timing/model': 0.3709132270014379, 'learning_rate/qwen_vl_interface': 9.962925555160919e-06, 'learning_rate/action_model': 9.962925555160918e-05, 'epoch': 0.02}) +[2026-10-03 14:41:10,288][starVLA.training.train_starvla][INFO] - Step 296, Loss: {'train/loss': 0.29911959171295166, 'train/grad_norm_pre_clip': 22.362693786621094, 'timing/data': 0.00032084000122267753, 'timing/model': 0.3710231369914254, 'learning_rate/qwen_vl_interface': 9.96254483124377e-06, 'learning_rate/action_model': 9.96254483124377e-05, 'epoch': 0.02}) +[2026-10-03 14:41:10,949][starVLA.training.train_starvla][INFO] - Step 297, Loss: {'train/loss': 0.13051526248455048, 'train/grad_norm_pre_clip': 7.9805216789245605, 'timing/data': 0.00025440000172238797, 'timing/model': 0.3721824870008277, 'learning_rate/qwen_vl_interface': 9.96216217017745e-06, 'learning_rate/action_model': 9.962162170177451e-05, 'epoch': 0.02}) +[2026-10-03 14:41:11,810][starVLA.training.train_starvla][INFO] - Step 298, Loss: {'train/loss': 0.32004791498184204, 'train/grad_norm_pre_clip': 38.58296585083008, 'timing/data': 0.0003604499943321571, 'timing/model': 0.37079772700963076, 'learning_rate/qwen_vl_interface': 9.961777572119258e-06, 'learning_rate/action_model': 9.961777572119258e-05, 'epoch': 0.02}) +[2026-10-03 14:41:12,471][starVLA.training.train_starvla][INFO] - Step 299, Loss: {'train/loss': 0.4531422257423401, 'train/grad_norm_pre_clip': 14.9446439743042, 'timing/data': 0.0002663000050233677, 'timing/model': 0.3713527670042822, 'learning_rate/qwen_vl_interface': 9.961391037227286e-06, 'learning_rate/action_model': 9.961391037227284e-05, 'epoch': 0.02}) +[2026-10-03 14:41:13,132][starVLA.training.train_starvla][INFO] - Step 300, Loss: {'train/loss': 0.511236310005188, 'train/grad_norm_pre_clip': 41.34547424316406, 'timing/data': 0.0003304800047772005, 'timing/model': 0.3710529070085613, 'learning_rate/qwen_vl_interface': 9.961002565660419e-06, 'learning_rate/action_model': 9.961002565660419e-05, 'epoch': 0.02}) +[2026-10-03 14:41:13,793][starVLA.training.train_starvla][INFO] - Step 301, Loss: {'train/loss': 0.8424093723297119, 'train/grad_norm_pre_clip': 63.232357025146484, 'timing/data': 0.000215559994103387, 'timing/model': 0.3712051159964176, 'learning_rate/qwen_vl_interface': 9.96061215757835e-06, 'learning_rate/action_model': 9.960612157578349e-05, 'epoch': 0.02}) +[2026-10-03 14:41:14,454][starVLA.training.train_starvla][INFO] - Step 302, Loss: {'train/loss': 0.11369486153125763, 'train/grad_norm_pre_clip': 6.878000259399414, 'timing/data': 0.000294829995254986, 'timing/model': 0.37107744699460454, 'learning_rate/qwen_vl_interface': 9.96021981314156e-06, 'learning_rate/action_model': 9.96021981314156e-05, 'epoch': 0.02}) +[2026-10-03 14:41:15,291][starVLA.training.train_starvla][INFO] - Step 303, Loss: {'train/loss': 1.214430809020996, 'train/grad_norm_pre_clip': 208.89651489257812, 'timing/data': 0.0002620599989313632, 'timing/model': 0.5705364040040877, 'learning_rate/qwen_vl_interface': 9.959825532511322e-06, 'learning_rate/action_model': 9.959825532511322e-05, 'epoch': 0.02}) +[2026-10-03 14:41:15,953][starVLA.training.train_starvla][INFO] - Step 304, Loss: {'train/loss': 0.5027456879615784, 'train/grad_norm_pre_clip': 62.263404846191406, 'timing/data': 0.0002538699918659404, 'timing/model': 0.37203724699793383, 'learning_rate/qwen_vl_interface': 9.959429315849715e-06, 'learning_rate/action_model': 9.959429315849715e-05, 'epoch': 0.02}) +[2026-10-03 14:41:16,615][starVLA.training.train_starvla][INFO] - Step 305, Loss: {'train/loss': 0.5447955131530762, 'train/grad_norm_pre_clip': 11.377476692199707, 'timing/data': 0.00021868999465368688, 'timing/model': 0.371978957002284, 'learning_rate/qwen_vl_interface': 9.959031163319606e-06, 'learning_rate/action_model': 9.959031163319607e-05, 'epoch': 0.02}) +[2026-10-03 14:41:17,275][starVLA.training.train_starvla][INFO] - Step 306, Loss: {'train/loss': 0.7629846930503845, 'train/grad_norm_pre_clip': 54.317481994628906, 'timing/data': 0.00034002999018412083, 'timing/model': 0.3708946270053275, 'learning_rate/qwen_vl_interface': 9.958631075084664e-06, 'learning_rate/action_model': 9.958631075084664e-05, 'epoch': 0.02}) +[2026-10-03 14:41:17,936][starVLA.training.train_starvla][INFO] - Step 307, Loss: {'train/loss': 0.1676511913537979, 'train/grad_norm_pre_clip': 28.069446563720703, 'timing/data': 0.00026119999529328197, 'timing/model': 0.37137169699417427, 'learning_rate/qwen_vl_interface': 9.958229051309346e-06, 'learning_rate/action_model': 9.958229051309346e-05, 'epoch': 0.02}) +[2026-10-03 14:41:18,598][starVLA.training.train_starvla][INFO] - Step 308, Loss: {'train/loss': 0.4792187809944153, 'train/grad_norm_pre_clip': 70.97738647460938, 'timing/data': 0.00021112000104039907, 'timing/model': 0.3720547870034352, 'learning_rate/qwen_vl_interface': 9.95782509215891e-06, 'learning_rate/action_model': 9.95782509215891e-05, 'epoch': 0.02}) +[2026-10-03 14:41:19,455][starVLA.training.train_starvla][INFO] - Step 309, Loss: {'train/loss': 0.26893511414527893, 'train/grad_norm_pre_clip': 21.58352279663086, 'timing/data': 0.00043884001206606627, 'timing/model': 0.371235277998494, 'learning_rate/qwen_vl_interface': 9.95741919779941e-06, 'learning_rate/action_model': 9.95741919779941e-05, 'epoch': 0.02}) +[2026-10-03 14:41:20,116][starVLA.training.train_starvla][INFO] - Step 310, Loss: {'train/loss': 0.33835792541503906, 'train/grad_norm_pre_clip': 104.83773803710938, 'timing/data': 0.00030929999775253236, 'timing/model': 0.3709448969893856, 'learning_rate/qwen_vl_interface': 9.957011368397692e-06, 'learning_rate/action_model': 9.95701136839769e-05, 'epoch': 0.02}) +[2026-10-03 14:41:20,778][starVLA.training.train_starvla][INFO] - Step 311, Loss: {'train/loss': 0.6014848947525024, 'train/grad_norm_pre_clip': 32.33149337768555, 'timing/data': 0.00032468000426888466, 'timing/model': 0.37168771799770184, 'learning_rate/qwen_vl_interface': 9.956601604121398e-06, 'learning_rate/action_model': 9.956601604121397e-05, 'epoch': 0.02}) +[2026-10-03 14:41:21,416][starVLA.training.train_starvla][INFO] - Step 312, Loss: {'train/loss': 0.7279342412948608, 'train/grad_norm_pre_clip': 116.99189758300781, 'timing/data': 0.0003013199893757701, 'timing/model': 0.3484947690012632, 'learning_rate/qwen_vl_interface': 9.95618990513897e-06, 'learning_rate/action_model': 9.95618990513897e-05, 'epoch': 0.02}) +[2026-10-03 14:41:22,077][starVLA.training.train_starvla][INFO] - Step 313, Loss: {'train/loss': 1.1572239398956299, 'train/grad_norm_pre_clip': 115.9422836303711, 'timing/data': 0.0002699899923754856, 'timing/model': 0.37131556699750945, 'learning_rate/qwen_vl_interface': 9.95577627161964e-06, 'learning_rate/action_model': 9.95577627161964e-05, 'epoch': 0.02}) +[2026-10-03 14:41:22,935][starVLA.training.train_starvla][INFO] - Step 314, Loss: {'train/loss': 0.19942869246006012, 'train/grad_norm_pre_clip': 46.37627410888672, 'timing/data': 0.0002899699902627617, 'timing/model': 0.5685767139948439, 'learning_rate/qwen_vl_interface': 9.955360703733438e-06, 'learning_rate/action_model': 9.955360703733437e-05, 'epoch': 0.02}) +[2026-10-03 14:41:23,595][starVLA.training.train_starvla][INFO] - Step 315, Loss: {'train/loss': 0.7185500860214233, 'train/grad_norm_pre_clip': 34.166229248046875, 'timing/data': 0.00025102999643422663, 'timing/model': 0.3710392059874721, 'learning_rate/qwen_vl_interface': 9.954943201651187e-06, 'learning_rate/action_model': 9.954943201651187e-05, 'epoch': 0.02}) +[2026-10-03 14:41:24,258][starVLA.training.train_starvla][INFO] - Step 316, Loss: {'train/loss': 0.8296049237251282, 'train/grad_norm_pre_clip': 87.43614196777344, 'timing/data': 0.00027227000100538135, 'timing/model': 0.3723288970068097, 'learning_rate/qwen_vl_interface': 9.954523765544507e-06, 'learning_rate/action_model': 9.954523765544508e-05, 'epoch': 0.02}) +[2026-10-03 14:41:24,923][starVLA.training.train_starvla][INFO] - Step 317, Loss: {'train/loss': 0.31310516595840454, 'train/grad_norm_pre_clip': 33.27657699584961, 'timing/data': 0.00037306000012904406, 'timing/model': 0.3737404470011825, 'learning_rate/qwen_vl_interface': 9.954102395585814e-06, 'learning_rate/action_model': 9.954102395585813e-05, 'epoch': 0.02}) +[2026-10-03 14:41:25,583][starVLA.training.train_starvla][INFO] - Step 318, Loss: {'train/loss': 0.7818573713302612, 'train/grad_norm_pre_clip': 73.26923370361328, 'timing/data': 0.0003191800060449168, 'timing/model': 0.3706312669964973, 'learning_rate/qwen_vl_interface': 9.953679091948313e-06, 'learning_rate/action_model': 9.953679091948314e-05, 'epoch': 0.02}) +[2026-10-03 14:41:26,244][starVLA.training.train_starvla][INFO] - Step 319, Loss: {'train/loss': 0.2318497896194458, 'train/grad_norm_pre_clip': 38.3856201171875, 'timing/data': 0.0002656599972397089, 'timing/model': 0.37162778701167554, 'learning_rate/qwen_vl_interface': 9.953253854806013e-06, 'learning_rate/action_model': 9.953253854806012e-05, 'epoch': 0.02}) +[2026-10-03 14:41:27,108][starVLA.training.train_starvla][INFO] - Step 320, Loss: {'train/loss': 0.3063639998435974, 'train/grad_norm_pre_clip': 23.55976104736328, 'timing/data': 0.00022662999981548637, 'timing/model': 0.3707985969958827, 'learning_rate/qwen_vl_interface': 9.95282668433371e-06, 'learning_rate/action_model': 9.952826684333709e-05, 'epoch': 0.02}) +[2026-10-03 14:41:27,772][starVLA.training.train_starvla][INFO] - Step 321, Loss: {'train/loss': 0.46708276867866516, 'train/grad_norm_pre_clip': 33.26036834716797, 'timing/data': 0.00033229000109713525, 'timing/model': 0.3734449869953096, 'learning_rate/qwen_vl_interface': 9.952397580706998e-06, 'learning_rate/action_model': 9.952397580706998e-05, 'epoch': 0.02}) +[2026-10-03 14:41:28,432][starVLA.training.train_starvla][INFO] - Step 322, Loss: {'train/loss': 0.10221868753433228, 'train/grad_norm_pre_clip': 24.2477970123291, 'timing/data': 0.0002172800013795495, 'timing/model': 0.3709730060072616, 'learning_rate/qwen_vl_interface': 9.951966544102267e-06, 'learning_rate/action_model': 9.951966544102266e-05, 'epoch': 0.02}) +[2026-10-03 14:41:29,092][starVLA.training.train_starvla][INFO] - Step 323, Loss: {'train/loss': 0.29443350434303284, 'train/grad_norm_pre_clip': 49.521392822265625, 'timing/data': 0.0002196600107708946, 'timing/model': 0.37128788699919824, 'learning_rate/qwen_vl_interface': 9.951533574696696e-06, 'learning_rate/action_model': 9.951533574696697e-05, 'epoch': 0.02}) +[2026-10-03 14:41:29,754][starVLA.training.train_starvla][INFO] - Step 324, Loss: {'train/loss': 0.704282283782959, 'train/grad_norm_pre_clip': 71.7889175415039, 'timing/data': 0.0003483300097286701, 'timing/model': 0.3716000959975645, 'learning_rate/qwen_vl_interface': 9.951098672668266e-06, 'learning_rate/action_model': 9.951098672668265e-05, 'epoch': 0.02}) +[2026-10-03 14:41:30,620][starVLA.training.train_starvla][INFO] - Step 325, Loss: {'train/loss': 0.21241256594657898, 'train/grad_norm_pre_clip': 35.02281188964844, 'timing/data': 0.00023945000430103391, 'timing/model': 0.3737950369977625, 'learning_rate/qwen_vl_interface': 9.95066183819575e-06, 'learning_rate/action_model': 9.950661838195749e-05, 'epoch': 0.02}) +[2026-10-03 14:41:31,233][starVLA.training.train_starvla][INFO] - Step 326, Loss: {'train/loss': 0.2661222815513611, 'train/grad_norm_pre_clip': 39.286556243896484, 'timing/data': 0.00023812000290490687, 'timing/model': 0.34698724999907427, 'learning_rate/qwen_vl_interface': 9.95022307145871e-06, 'learning_rate/action_model': 9.950223071458708e-05, 'epoch': 0.02}) +[2026-10-03 14:41:31,893][starVLA.training.train_starvla][INFO] - Step 327, Loss: {'train/loss': 0.23412954807281494, 'train/grad_norm_pre_clip': 41.34227752685547, 'timing/data': 0.00022208999143913388, 'timing/model': 0.37092115700943395, 'learning_rate/qwen_vl_interface': 9.949782372637507e-06, 'learning_rate/action_model': 9.949782372637506e-05, 'epoch': 0.02}) +[2026-10-03 14:41:32,555][starVLA.training.train_starvla][INFO] - Step 328, Loss: {'train/loss': 0.4357680678367615, 'train/grad_norm_pre_clip': 44.725257873535156, 'timing/data': 0.00026884999533649534, 'timing/model': 0.3715950559999328, 'learning_rate/qwen_vl_interface': 9.949339741913297e-06, 'learning_rate/action_model': 9.949339741913297e-05, 'epoch': 0.02}) +[2026-10-03 14:41:33,219][starVLA.training.train_starvla][INFO] - Step 329, Loss: {'train/loss': 0.25251781940460205, 'train/grad_norm_pre_clip': 30.27629280090332, 'timing/data': 0.00036311999429017305, 'timing/model': 0.37400426699605305, 'learning_rate/qwen_vl_interface': 9.948895179468032e-06, 'learning_rate/action_model': 9.948895179468031e-05, 'epoch': 0.02}) +[2026-10-03 14:41:34,079][starVLA.training.train_starvla][INFO] - Step 330, Loss: {'train/loss': 0.2237306535243988, 'train/grad_norm_pre_clip': 36.59273147583008, 'timing/data': 0.00022520001220982522, 'timing/model': 0.5663610240007984, 'learning_rate/qwen_vl_interface': 9.948448685484448e-06, 'learning_rate/action_model': 9.948448685484447e-05, 'epoch': 0.02}) +[2026-10-03 14:41:34,736][starVLA.training.train_starvla][INFO] - Step 331, Loss: {'train/loss': 0.4322592318058014, 'train/grad_norm_pre_clip': 15.585450172424316, 'timing/data': 0.00037898001028224826, 'timing/model': 0.3700898370007053, 'learning_rate/qwen_vl_interface': 9.948000260146087e-06, 'learning_rate/action_model': 9.948000260146087e-05, 'epoch': 0.02}) +[2026-10-03 14:41:35,394][starVLA.training.train_starvla][INFO] - Step 332, Loss: {'train/loss': 0.8061368465423584, 'train/grad_norm_pre_clip': 120.48966979980469, 'timing/data': 0.0003367400058778003, 'timing/model': 0.371205186995212, 'learning_rate/qwen_vl_interface': 9.947549903637277e-06, 'learning_rate/action_model': 9.947549903637278e-05, 'epoch': 0.02}) +[2026-10-03 14:41:36,052][starVLA.training.train_starvla][INFO] - Step 333, Loss: {'train/loss': 0.7091614007949829, 'train/grad_norm_pre_clip': 77.34306335449219, 'timing/data': 0.00035275999107398093, 'timing/model': 0.3691188179946039, 'learning_rate/qwen_vl_interface': 9.947097616143145e-06, 'learning_rate/action_model': 9.947097616143144e-05, 'epoch': 0.02}) +[2026-10-03 14:41:36,709][starVLA.training.train_starvla][INFO] - Step 334, Loss: {'train/loss': 0.2358684539794922, 'train/grad_norm_pre_clip': 43.49164962768555, 'timing/data': 0.000255559993092902, 'timing/model': 0.37075066700344905, 'learning_rate/qwen_vl_interface': 9.946643397849607e-06, 'learning_rate/action_model': 9.946643397849607e-05, 'epoch': 0.02}) +[2026-10-03 14:41:37,366][starVLA.training.train_starvla][INFO] - Step 335, Loss: {'train/loss': 0.6680563688278198, 'train/grad_norm_pre_clip': 56.20586013793945, 'timing/data': 0.0002439399977447465, 'timing/model': 0.3700330769934226, 'learning_rate/qwen_vl_interface': 9.946187248943375e-06, 'learning_rate/action_model': 9.946187248943374e-05, 'epoch': 0.02}) +[2026-10-03 14:41:38,213][starVLA.training.train_starvla][INFO] - Step 336, Loss: {'train/loss': 0.24471455812454224, 'train/grad_norm_pre_clip': 30.865188598632812, 'timing/data': 0.00028088998806197196, 'timing/model': 0.3706602070014924, 'learning_rate/qwen_vl_interface': 9.945729169611957e-06, 'learning_rate/action_model': 9.945729169611958e-05, 'epoch': 0.02}) +[2026-10-03 14:41:38,868][starVLA.training.train_starvla][INFO] - Step 337, Loss: {'train/loss': 0.5841405987739563, 'train/grad_norm_pre_clip': 37.29802322387695, 'timing/data': 0.00024446999304927886, 'timing/model': 0.3683103669900447, 'learning_rate/qwen_vl_interface': 9.94526916004365e-06, 'learning_rate/action_model': 9.94526916004365e-05, 'epoch': 0.02}) +[2026-10-03 14:41:39,527][starVLA.training.train_starvla][INFO] - Step 338, Loss: {'train/loss': 0.30964046716690063, 'train/grad_norm_pre_clip': 18.9864444732666, 'timing/data': 0.00020307999511715025, 'timing/model': 0.3717386770003941, 'learning_rate/qwen_vl_interface': 9.944807220427549e-06, 'learning_rate/action_model': 9.944807220427548e-05, 'epoch': 0.02}) +[2026-10-03 14:41:40,184][starVLA.training.train_starvla][INFO] - Step 339, Loss: {'train/loss': 0.4359034299850464, 'train/grad_norm_pre_clip': 34.31871032714844, 'timing/data': 0.000246069990680553, 'timing/model': 0.3701446670020232, 'learning_rate/qwen_vl_interface': 9.944343350953534e-06, 'learning_rate/action_model': 9.944343350953533e-05, 'epoch': 0.02}) +[2026-10-03 14:41:40,842][starVLA.training.train_starvla][INFO] - Step 340, Loss: {'train/loss': 0.35742413997650146, 'train/grad_norm_pre_clip': 33.112728118896484, 'timing/data': 0.00028202000248711556, 'timing/model': 0.3706986559991492, 'learning_rate/qwen_vl_interface': 9.94387755181229e-06, 'learning_rate/action_model': 9.94387755181229e-05, 'epoch': 0.02}) +[2026-10-03 14:41:41,693][starVLA.training.train_starvla][INFO] - Step 341, Loss: {'train/loss': 0.3733079135417938, 'train/grad_norm_pre_clip': 43.949947357177734, 'timing/data': 0.0003687699936563149, 'timing/model': 0.36908162699546665, 'learning_rate/qwen_vl_interface': 9.943409823195289e-06, 'learning_rate/action_model': 9.943409823195288e-05, 'epoch': 0.02}) +[2026-10-03 14:41:42,350][starVLA.training.train_starvla][INFO] - Step 342, Loss: {'train/loss': 0.38385963439941406, 'train/grad_norm_pre_clip': 42.76972198486328, 'timing/data': 0.000244909999310039, 'timing/model': 0.37058171699754894, 'learning_rate/qwen_vl_interface': 9.942940165294794e-06, 'learning_rate/action_model': 9.942940165294793e-05, 'epoch': 0.03}) +[2026-10-03 14:41:43,008][starVLA.training.train_starvla][INFO] - Step 343, Loss: {'train/loss': 0.4176886975765228, 'train/grad_norm_pre_clip': 36.07356262207031, 'timing/data': 0.000248609998379834, 'timing/model': 0.37031226699764375, 'learning_rate/qwen_vl_interface': 9.942468578303865e-06, 'learning_rate/action_model': 9.942468578303865e-05, 'epoch': 0.03}) +[2026-10-03 14:41:43,665][starVLA.training.train_starvla][INFO] - Step 344, Loss: {'train/loss': 0.3400432765483856, 'train/grad_norm_pre_clip': 22.7187442779541, 'timing/data': 0.00025530999118927866, 'timing/model': 0.37057746699429117, 'learning_rate/qwen_vl_interface': 9.941995062416354e-06, 'learning_rate/action_model': 9.941995062416354e-05, 'epoch': 0.03}) +[2026-10-03 14:41:44,298][starVLA.training.train_starvla][INFO] - Step 345, Loss: {'train/loss': 0.26083895564079285, 'train/grad_norm_pre_clip': 27.465425491333008, 'timing/data': 0.0002115399984177202, 'timing/model': 0.3684333270066418, 'learning_rate/qwen_vl_interface': 9.941519617826906e-06, 'learning_rate/action_model': 9.941519617826906e-05, 'epoch': 0.03}) +[2026-10-03 14:41:45,147][starVLA.training.train_starvla][INFO] - Step 346, Loss: {'train/loss': 0.24197952449321747, 'train/grad_norm_pre_clip': 16.504793167114258, 'timing/data': 0.0002438200026517734, 'timing/model': 0.5624480449914699, 'learning_rate/qwen_vl_interface': 9.941042244730954e-06, 'learning_rate/action_model': 9.941042244730955e-05, 'epoch': 0.03}) +[2026-10-03 14:41:45,805][starVLA.training.train_starvla][INFO] - Step 347, Loss: {'train/loss': 0.7106620669364929, 'train/grad_norm_pre_clip': 46.73457717895508, 'timing/data': 0.0002716899907682091, 'timing/model': 0.37063037599727977, 'learning_rate/qwen_vl_interface': 9.940562943324736e-06, 'learning_rate/action_model': 9.940562943324735e-05, 'epoch': 0.03}) +[2026-10-03 14:41:46,441][starVLA.training.train_starvla][INFO] - Step 348, Loss: {'train/loss': 0.27073854207992554, 'train/grad_norm_pre_clip': 22.417116165161133, 'timing/data': 0.00032256000849884003, 'timing/model': 0.3473137100081658, 'learning_rate/qwen_vl_interface': 9.940081713805266e-06, 'learning_rate/action_model': 9.940081713805266e-05, 'epoch': 0.03}) +[2026-10-03 14:41:47,074][starVLA.training.train_starvla][INFO] - Step 349, Loss: {'train/loss': 0.145699143409729, 'train/grad_norm_pre_clip': 20.70162010192871, 'timing/data': 0.0003785599983530119, 'timing/model': 0.36876602699339855, 'learning_rate/qwen_vl_interface': 9.939598556370366e-06, 'learning_rate/action_model': 9.939598556370365e-05, 'epoch': 0.03}) +[2026-10-03 14:41:47,732][starVLA.training.train_starvla][INFO] - Step 350, Loss: {'train/loss': 0.3704301714897156, 'train/grad_norm_pre_clip': 78.61494445800781, 'timing/data': 0.0003013200039276853, 'timing/model': 0.37073931698978413, 'learning_rate/qwen_vl_interface': 9.93911347121864e-06, 'learning_rate/action_model': 9.939113471218639e-05, 'epoch': 0.03}) +[2026-10-03 14:41:48,585][starVLA.training.train_starvla][INFO] - Step 351, Loss: {'train/loss': 0.27037131786346436, 'train/grad_norm_pre_clip': 38.381996154785156, 'timing/data': 0.0002025800058618188, 'timing/model': 0.5654587250028271, 'learning_rate/qwen_vl_interface': 9.938626458549488e-06, 'learning_rate/action_model': 9.938626458549488e-05, 'epoch': 0.03}) +[2026-10-03 14:41:49,242][starVLA.training.train_starvla][INFO] - Step 352, Loss: {'train/loss': 0.4104023575782776, 'train/grad_norm_pre_clip': 54.5626335144043, 'timing/data': 0.00022883000201545656, 'timing/model': 0.37039736700535286, 'learning_rate/qwen_vl_interface': 9.938137518563106e-06, 'learning_rate/action_model': 9.938137518563106e-05, 'epoch': 0.03}) +[2026-10-03 14:41:49,899][starVLA.training.train_starvla][INFO] - Step 353, Loss: {'train/loss': 0.942839503288269, 'train/grad_norm_pre_clip': 61.00687789916992, 'timing/data': 0.00022996000188868493, 'timing/model': 0.36917632799304556, 'learning_rate/qwen_vl_interface': 9.937646651460475e-06, 'learning_rate/action_model': 9.937646651460474e-05, 'epoch': 0.03}) +[2026-10-03 14:41:50,557][starVLA.training.train_starvla][INFO] - Step 354, Loss: {'train/loss': 0.5046491026878357, 'train/grad_norm_pre_clip': 54.81663131713867, 'timing/data': 0.0002685100043890998, 'timing/model': 0.37144047699985094, 'learning_rate/qwen_vl_interface': 9.937153857443375e-06, 'learning_rate/action_model': 9.937153857443374e-05, 'epoch': 0.03}) +[2026-10-03 14:41:51,215][starVLA.training.train_starvla][INFO] - Step 355, Loss: {'train/loss': 0.15077492594718933, 'train/grad_norm_pre_clip': 50.81932830810547, 'timing/data': 0.0002547899930505082, 'timing/model': 0.3709326170064742, 'learning_rate/qwen_vl_interface': 9.936659136714372e-06, 'learning_rate/action_model': 9.936659136714371e-05, 'epoch': 0.03}) +[2026-10-03 14:41:51,874][starVLA.training.train_starvla][INFO] - Step 356, Loss: {'train/loss': 0.22047372162342072, 'train/grad_norm_pre_clip': 47.18801498413086, 'timing/data': 0.00031616000342182815, 'timing/model': 0.3708629869943252, 'learning_rate/qwen_vl_interface': 9.936162489476827e-06, 'learning_rate/action_model': 9.936162489476828e-05, 'epoch': 0.03}) +[2026-10-03 14:41:52,726][starVLA.training.train_starvla][INFO] - Step 357, Loss: {'train/loss': 0.7430737018585205, 'train/grad_norm_pre_clip': 27.47723388671875, 'timing/data': 0.0003394899977138266, 'timing/model': 0.36919357799342833, 'learning_rate/qwen_vl_interface': 9.935663915934899e-06, 'learning_rate/action_model': 9.935663915934898e-05, 'epoch': 0.03}) +[2026-10-03 14:41:53,384][starVLA.training.train_starvla][INFO] - Step 358, Loss: {'train/loss': 0.050578512251377106, 'train/grad_norm_pre_clip': 34.364898681640625, 'timing/data': 0.00023659999715164304, 'timing/model': 0.37169437599368393, 'learning_rate/qwen_vl_interface': 9.935163416293525e-06, 'learning_rate/action_model': 9.935163416293525e-05, 'epoch': 0.03}) +[2026-10-03 14:41:54,020][starVLA.training.train_starvla][INFO] - Step 359, Loss: {'train/loss': 0.7159076929092407, 'train/grad_norm_pre_clip': 45.773128509521484, 'timing/data': 0.00025131000438705087, 'timing/model': 0.34740310000779573, 'learning_rate/qwen_vl_interface': 9.934660990758445e-06, 'learning_rate/action_model': 9.934660990758445e-05, 'epoch': 0.03}) +[2026-10-03 14:41:54,678][starVLA.training.train_starvla][INFO] - Step 360, Loss: {'train/loss': 0.6582701802253723, 'train/grad_norm_pre_clip': 131.69261169433594, 'timing/data': 0.0002988599881064147, 'timing/model': 0.3707696570054395, 'learning_rate/qwen_vl_interface': 9.934156639536186e-06, 'learning_rate/action_model': 9.934156639536187e-05, 'epoch': 0.03}) +[2026-10-03 14:41:55,334][starVLA.training.train_starvla][INFO] - Step 361, Loss: {'train/loss': 0.023402798920869827, 'train/grad_norm_pre_clip': 3.088324785232544, 'timing/data': 0.00025167000421788543, 'timing/model': 0.3689798970008269, 'learning_rate/qwen_vl_interface': 9.93365036283407e-06, 'learning_rate/action_model': 9.93365036283407e-05, 'epoch': 0.03}) +[2026-10-03 14:41:56,158][starVLA.training.train_starvla][INFO] - Step 362, Loss: {'train/loss': 1.0172322988510132, 'train/grad_norm_pre_clip': 64.71083068847656, 'timing/data': 0.00029892000020481646, 'timing/model': 0.5599105550063541, 'learning_rate/qwen_vl_interface': 9.933142160860208e-06, 'learning_rate/action_model': 9.933142160860208e-05, 'epoch': 0.03}) +[2026-10-03 14:41:56,816][starVLA.training.train_starvla][INFO] - Step 363, Loss: {'train/loss': 0.06064319610595703, 'train/grad_norm_pre_clip': 14.224609375, 'timing/data': 0.0003112600097665563, 'timing/model': 0.3708250780036906, 'learning_rate/qwen_vl_interface': 9.932632033823499e-06, 'learning_rate/action_model': 9.932632033823499e-05, 'epoch': 0.03}) +[2026-10-03 14:41:57,474][starVLA.training.train_starvla][INFO] - Step 364, Loss: {'train/loss': 0.2563905119895935, 'train/grad_norm_pre_clip': 72.85272979736328, 'timing/data': 0.00024476999533362687, 'timing/model': 0.37091637699631974, 'learning_rate/qwen_vl_interface': 9.93211998193364e-06, 'learning_rate/action_model': 9.932119981933639e-05, 'epoch': 0.03}) +[2026-10-03 14:41:58,132][starVLA.training.train_starvla][INFO] - Step 365, Loss: {'train/loss': 0.6967508792877197, 'train/grad_norm_pre_clip': 25.83563232421875, 'timing/data': 0.00024762000248301774, 'timing/model': 0.369187887001317, 'learning_rate/qwen_vl_interface': 9.931606005401117e-06, 'learning_rate/action_model': 9.931606005401116e-05, 'epoch': 0.03}) +[2026-10-03 14:41:58,789][starVLA.training.train_starvla][INFO] - Step 366, Loss: {'train/loss': 0.28912976384162903, 'train/grad_norm_pre_clip': 27.25821876525879, 'timing/data': 0.0001997600047616288, 'timing/model': 0.37091133699868806, 'learning_rate/qwen_vl_interface': 9.931090104437202e-06, 'learning_rate/action_model': 9.931090104437202e-05, 'epoch': 0.03}) +[2026-10-03 14:41:59,643][starVLA.training.train_starvla][INFO] - Step 367, Loss: {'train/loss': 0.6069158911705017, 'train/grad_norm_pre_clip': 59.63036346435547, 'timing/data': 0.00027502999000716954, 'timing/model': 0.5663712450041203, 'learning_rate/qwen_vl_interface': 9.930572279253967e-06, 'learning_rate/action_model': 9.930572279253967e-05, 'epoch': 0.03}) +[2026-10-03 14:42:00,302][starVLA.training.train_starvla][INFO] - Step 368, Loss: {'train/loss': 0.5408772826194763, 'train/grad_norm_pre_clip': 67.0318832397461, 'timing/data': 0.00024149999080691487, 'timing/model': 0.3714898869948229, 'learning_rate/qwen_vl_interface': 9.930052530064269e-06, 'learning_rate/action_model': 9.93005253006427e-05, 'epoch': 0.03}) +[2026-10-03 14:42:00,958][starVLA.training.train_starvla][INFO] - Step 369, Loss: {'train/loss': 0.34340280294418335, 'train/grad_norm_pre_clip': 31.925540924072266, 'timing/data': 0.00026293999690096825, 'timing/model': 0.3688711470022099, 'learning_rate/qwen_vl_interface': 9.929530857081757e-06, 'learning_rate/action_model': 9.929530857081757e-05, 'epoch': 0.03}) +[2026-10-03 14:42:01,617][starVLA.training.train_starvla][INFO] - Step 370, Loss: {'train/loss': 0.14960846304893494, 'train/grad_norm_pre_clip': 34.90095138549805, 'timing/data': 0.00020493999181780964, 'timing/model': 0.37138089700601995, 'learning_rate/qwen_vl_interface': 9.92900726052087e-06, 'learning_rate/action_model': 9.929007260520869e-05, 'epoch': 0.03}) +[2026-10-03 14:42:02,275][starVLA.training.train_starvla][INFO] - Step 371, Loss: {'train/loss': 0.4426009953022003, 'train/grad_norm_pre_clip': 54.89262390136719, 'timing/data': 0.0002508200122974813, 'timing/model': 0.3705493879970163, 'learning_rate/qwen_vl_interface': 9.92848174059684e-06, 'learning_rate/action_model': 9.92848174059684e-05, 'epoch': 0.03}) +[2026-10-03 14:42:03,129][starVLA.training.train_starvla][INFO] - Step 372, Loss: {'train/loss': 0.5087581276893616, 'train/grad_norm_pre_clip': 59.10880661010742, 'timing/data': 0.0002723499928833917, 'timing/model': 0.5661776240012841, 'learning_rate/qwen_vl_interface': 9.92795429752569e-06, 'learning_rate/action_model': 9.927954297525689e-05, 'epoch': 0.03}) +[2026-10-03 14:42:03,785][starVLA.training.train_starvla][INFO] - Step 373, Loss: {'train/loss': 0.5780847072601318, 'train/grad_norm_pre_clip': 36.51555252075195, 'timing/data': 0.0001994399935938418, 'timing/model': 0.3688237470050808, 'learning_rate/qwen_vl_interface': 9.927424931524228e-06, 'learning_rate/action_model': 9.927424931524228e-05, 'epoch': 0.03}) +[2026-10-03 14:42:04,442][starVLA.training.train_starvla][INFO] - Step 374, Loss: {'train/loss': 0.6272035837173462, 'train/grad_norm_pre_clip': 45.59873580932617, 'timing/data': 0.00024295000184793025, 'timing/model': 0.3707323069975246, 'learning_rate/qwen_vl_interface': 9.926893642810058e-06, 'learning_rate/action_model': 9.926893642810059e-05, 'epoch': 0.03}) +[2026-10-03 14:42:05,100][starVLA.training.train_starvla][INFO] - Step 375, Loss: {'train/loss': 0.595710277557373, 'train/grad_norm_pre_clip': 98.6710433959961, 'timing/data': 0.00030251000134740025, 'timing/model': 0.37018256800365634, 'learning_rate/qwen_vl_interface': 9.926360431601575e-06, 'learning_rate/action_model': 9.926360431601574e-05, 'epoch': 0.03}) +[2026-10-03 14:42:05,759][starVLA.training.train_starvla][INFO] - Step 376, Loss: {'train/loss': 0.5260775089263916, 'train/grad_norm_pre_clip': 58.51103591918945, 'timing/data': 0.00021502000163309276, 'timing/model': 0.37181674600287806, 'learning_rate/qwen_vl_interface': 9.92582529811796e-06, 'learning_rate/action_model': 9.925825298117959e-05, 'epoch': 0.03}) +[2026-10-03 14:42:06,415][starVLA.training.train_starvla][INFO] - Step 377, Loss: {'train/loss': 0.14183738827705383, 'train/grad_norm_pre_clip': 13.832243919372559, 'timing/data': 0.0002768600097624585, 'timing/model': 0.36866015700798016, 'learning_rate/qwen_vl_interface': 9.925288242579184e-06, 'learning_rate/action_model': 9.925288242579184e-05, 'epoch': 0.03}) +[2026-10-03 14:42:07,272][starVLA.training.train_starvla][INFO] - Step 378, Loss: {'train/loss': 0.44687145948410034, 'train/grad_norm_pre_clip': 33.60283279418945, 'timing/data': 0.00033804000122472644, 'timing/model': 0.3708566869900096, 'learning_rate/qwen_vl_interface': 9.924749265206015e-06, 'learning_rate/action_model': 9.924749265206015e-05, 'epoch': 0.03}) +[2026-10-03 14:42:07,927][starVLA.training.train_starvla][INFO] - Step 379, Loss: {'train/loss': 0.11818496882915497, 'train/grad_norm_pre_clip': 8.041208267211914, 'timing/data': 0.00025183000252582133, 'timing/model': 0.3671053479920374, 'learning_rate/qwen_vl_interface': 9.924208366220003e-06, 'learning_rate/action_model': 9.924208366220002e-05, 'epoch': 0.03}) +[2026-10-03 14:42:08,580][starVLA.training.train_starvla][INFO] - Step 380, Loss: {'train/loss': 0.43406355381011963, 'train/grad_norm_pre_clip': 30.04363250732422, 'timing/data': 0.0003885289916070178, 'timing/model': 0.3674832679971587, 'learning_rate/qwen_vl_interface': 9.923665545843492e-06, 'learning_rate/action_model': 9.923665545843491e-05, 'epoch': 0.03}) +[2026-10-03 14:42:09,232][starVLA.training.train_starvla][INFO] - Step 381, Loss: {'train/loss': 0.17391899228096008, 'train/grad_norm_pre_clip': 35.29423141479492, 'timing/data': 0.00032638999982737005, 'timing/model': 0.3666321170021547, 'learning_rate/qwen_vl_interface': 9.923120804299615e-06, 'learning_rate/action_model': 9.923120804299613e-05, 'epoch': 0.03}) +[2026-10-03 14:42:09,860][starVLA.training.train_starvla][INFO] - Step 382, Loss: {'train/loss': 0.2739259600639343, 'train/grad_norm_pre_clip': 83.92838287353516, 'timing/data': 0.00026246999914292246, 'timing/model': 0.36651814699871466, 'learning_rate/qwen_vl_interface': 9.922574141812295e-06, 'learning_rate/action_model': 9.922574141812294e-05, 'epoch': 0.03}) +[2026-10-03 14:42:10,712][starVLA.training.train_starvla][INFO] - Step 383, Loss: {'train/loss': 0.5921667218208313, 'train/grad_norm_pre_clip': 81.38385009765625, 'timing/data': 0.00030325999250635505, 'timing/model': 0.5668353750079405, 'learning_rate/qwen_vl_interface': 9.922025558606242e-06, 'learning_rate/action_model': 9.922025558606242e-05, 'epoch': 0.03}) +[2026-10-03 14:42:11,341][starVLA.training.train_starvla][INFO] - Step 384, Loss: {'train/loss': 0.26186269521713257, 'train/grad_norm_pre_clip': 40.46132278442383, 'timing/data': 0.00033831999462563545, 'timing/model': 0.34386722098861355, 'learning_rate/qwen_vl_interface': 9.92147505490696e-06, 'learning_rate/action_model': 9.92147505490696e-05, 'epoch': 0.03}) +[2026-10-03 14:42:11,969][starVLA.training.train_starvla][INFO] - Step 385, Loss: {'train/loss': 0.22885924577713013, 'train/grad_norm_pre_clip': 58.932579040527344, 'timing/data': 0.00029290000384207815, 'timing/model': 0.34362696000607684, 'learning_rate/qwen_vl_interface': 9.920922630940741e-06, 'learning_rate/action_model': 9.920922630940742e-05, 'epoch': 0.03}) +[2026-10-03 14:42:12,597][starVLA.training.train_starvla][INFO] - Step 386, Loss: {'train/loss': 0.4923517405986786, 'train/grad_norm_pre_clip': 54.6190299987793, 'timing/data': 0.00028404999466147274, 'timing/model': 0.3663369679998141, 'learning_rate/qwen_vl_interface': 9.920368286934664e-06, 'learning_rate/action_model': 9.920368286934663e-05, 'epoch': 0.03}) +[2026-10-03 14:42:13,248][starVLA.training.train_starvla][INFO] - Step 387, Loss: {'train/loss': 0.5373445749282837, 'train/grad_norm_pre_clip': 70.77501678466797, 'timing/data': 0.0002640200109453872, 'timing/model': 0.3666283780039521, 'learning_rate/qwen_vl_interface': 9.9198120231166e-06, 'learning_rate/action_model': 9.9198120231166e-05, 'epoch': 0.03}) +[2026-10-03 14:42:13,900][starVLA.training.train_starvla][INFO] - Step 388, Loss: {'train/loss': 0.16432693600654602, 'train/grad_norm_pre_clip': 45.277523040771484, 'timing/data': 0.0002477699890732765, 'timing/model': 0.36729708799975924, 'learning_rate/qwen_vl_interface': 9.919253839715208e-06, 'learning_rate/action_model': 9.919253839715208e-05, 'epoch': 0.03}) +[2026-10-03 14:42:14,743][starVLA.training.train_starvla][INFO] - Step 389, Loss: {'train/loss': 0.34636449813842773, 'train/grad_norm_pre_clip': 44.02583694458008, 'timing/data': 0.0003851389919873327, 'timing/model': 0.3663093280047178, 'learning_rate/qwen_vl_interface': 9.918693736959935e-06, 'learning_rate/action_model': 9.918693736959934e-05, 'epoch': 0.03}) +[2026-10-03 14:42:15,395][starVLA.training.train_starvla][INFO] - Step 390, Loss: {'train/loss': 0.8532751798629761, 'train/grad_norm_pre_clip': 48.832969665527344, 'timing/data': 0.0002069999900413677, 'timing/model': 0.3667515870038187, 'learning_rate/qwen_vl_interface': 9.918131715081019e-06, 'learning_rate/action_model': 9.918131715081018e-05, 'epoch': 0.03}) +[2026-10-03 14:42:16,047][starVLA.training.train_starvla][INFO] - Step 391, Loss: {'train/loss': 0.06144510954618454, 'train/grad_norm_pre_clip': 13.205744743347168, 'timing/data': 0.00037028000224381685, 'timing/model': 0.3668241880077403, 'learning_rate/qwen_vl_interface': 9.917567774309486e-06, 'learning_rate/action_model': 9.917567774309485e-05, 'epoch': 0.03}) +[2026-10-03 14:42:16,678][starVLA.training.train_starvla][INFO] - Step 392, Loss: {'train/loss': 0.14753058552742004, 'train/grad_norm_pre_clip': 50.35417938232422, 'timing/data': 0.00019712999346666038, 'timing/model': 0.3667505479970714, 'learning_rate/qwen_vl_interface': 9.91700191487715e-06, 'learning_rate/action_model': 9.91700191487715e-05, 'epoch': 0.03}) +[2026-10-03 14:42:17,332][starVLA.training.train_starvla][INFO] - Step 393, Loss: {'train/loss': 0.45822542905807495, 'train/grad_norm_pre_clip': 80.98030090332031, 'timing/data': 0.00020742999913636595, 'timing/model': 0.3664235570031451, 'learning_rate/qwen_vl_interface': 9.916434137016614e-06, 'learning_rate/action_model': 9.916434137016613e-05, 'epoch': 0.03}) +[2026-10-03 14:42:18,152][starVLA.training.train_starvla][INFO] - Step 394, Loss: {'train/loss': 0.31698477268218994, 'train/grad_norm_pre_clip': 62.06095886230469, 'timing/data': 0.00041309000516775995, 'timing/model': 0.34359756000048947, 'learning_rate/qwen_vl_interface': 9.915864440961272e-06, 'learning_rate/action_model': 9.915864440961272e-05, 'epoch': 0.03}) +[2026-10-03 14:42:18,804][starVLA.training.train_starvla][INFO] - Step 395, Loss: {'train/loss': 0.05867604911327362, 'train/grad_norm_pre_clip': 18.4974422454834, 'timing/data': 0.0003105100040556863, 'timing/model': 0.36718291700526606, 'learning_rate/qwen_vl_interface': 9.915292826945303e-06, 'learning_rate/action_model': 9.915292826945303e-05, 'epoch': 0.03}) +[2026-10-03 14:42:19,457][starVLA.training.train_starvla][INFO] - Step 396, Loss: {'train/loss': 0.29730188846588135, 'train/grad_norm_pre_clip': 71.9268569946289, 'timing/data': 0.0002055000077234581, 'timing/model': 0.3678509680030402, 'learning_rate/qwen_vl_interface': 9.914719295203676e-06, 'learning_rate/action_model': 9.914719295203676e-05, 'epoch': 0.03}) +[2026-10-03 14:42:20,086][starVLA.training.train_starvla][INFO] - Step 397, Loss: {'train/loss': 0.5899510383605957, 'train/grad_norm_pre_clip': 35.20531463623047, 'timing/data': 0.00020268000662326813, 'timing/model': 0.3437945000041509, 'learning_rate/qwen_vl_interface': 9.91414384597215e-06, 'learning_rate/action_model': 9.91414384597215e-05, 'epoch': 0.03}) +[2026-10-03 14:42:20,737][starVLA.training.train_starvla][INFO] - Step 398, Loss: {'train/loss': 0.40003687143325806, 'train/grad_norm_pre_clip': 63.81954574584961, 'timing/data': 0.0004592199984472245, 'timing/model': 0.3670046380138956, 'learning_rate/qwen_vl_interface': 9.913566479487267e-06, 'learning_rate/action_model': 9.913566479487267e-05, 'epoch': 0.03}) +[2026-10-03 14:42:21,580][starVLA.training.train_starvla][INFO] - Step 399, Loss: {'train/loss': 0.43093326687812805, 'train/grad_norm_pre_clip': 102.22265625, 'timing/data': 0.00033165000786539167, 'timing/model': 0.5582705949927913, 'learning_rate/qwen_vl_interface': 9.912987195986366e-06, 'learning_rate/action_model': 9.912987195986365e-05, 'epoch': 0.03}) +[2026-10-03 14:42:22,210][starVLA.training.train_starvla][INFO] - Step 400, Loss: {'train/loss': 0.17262576520442963, 'train/grad_norm_pre_clip': 70.82576751708984, 'timing/data': 0.00031132998992688954, 'timing/model': 0.34411284000088926, 'learning_rate/qwen_vl_interface': 9.912405995707562e-06, 'learning_rate/action_model': 9.912405995707561e-05, 'epoch': 0.03}) +[2026-10-03 14:42:22,861][starVLA.training.train_starvla][INFO] - Step 401, Loss: {'train/loss': 0.595360279083252, 'train/grad_norm_pre_clip': 64.59151458740234, 'timing/data': 0.0003296999930171296, 'timing/model': 0.366265367003507, 'learning_rate/qwen_vl_interface': 9.911822878889767e-06, 'learning_rate/action_model': 9.911822878889767e-05, 'epoch': 0.03}) +[2026-10-03 14:42:23,512][starVLA.training.train_starvla][INFO] - Step 402, Loss: {'train/loss': 0.14720499515533447, 'train/grad_norm_pre_clip': 75.02381134033203, 'timing/data': 0.0002543800073908642, 'timing/model': 0.36688238799979445, 'learning_rate/qwen_vl_interface': 9.911237845772681e-06, 'learning_rate/action_model': 9.911237845772681e-05, 'epoch': 0.03}) +[2026-10-03 14:42:24,141][starVLA.training.train_starvla][INFO] - Step 403, Loss: {'train/loss': 0.20131485164165497, 'train/grad_norm_pre_clip': 89.64248657226562, 'timing/data': 0.00033348999568261206, 'timing/model': 0.3433988100005081, 'learning_rate/qwen_vl_interface': 9.910650896596787e-06, 'learning_rate/action_model': 9.910650896596786e-05, 'epoch': 0.03}) +[2026-10-03 14:42:24,981][starVLA.training.train_starvla][INFO] - Step 404, Loss: {'train/loss': 0.28054848313331604, 'train/grad_norm_pre_clip': 45.37934112548828, 'timing/data': 0.000304058994515799, 'timing/model': 0.5550710359966615, 'learning_rate/qwen_vl_interface': 9.910062031603356e-06, 'learning_rate/action_model': 9.910062031603356e-05, 'epoch': 0.03}) +[2026-10-03 14:42:25,633][starVLA.training.train_starvla][INFO] - Step 405, Loss: {'train/loss': 0.1964101642370224, 'train/grad_norm_pre_clip': 53.211219787597656, 'timing/data': 0.00028577999910339713, 'timing/model': 0.366408998001134, 'learning_rate/qwen_vl_interface': 9.909471251034451e-06, 'learning_rate/action_model': 9.90947125103445e-05, 'epoch': 0.03}) +[2026-10-03 14:42:26,285][starVLA.training.train_starvla][INFO] - Step 406, Loss: {'train/loss': 0.09787289798259735, 'train/grad_norm_pre_clip': 26.56349754333496, 'timing/data': 0.00026439000794198364, 'timing/model': 0.36731103700003587, 'learning_rate/qwen_vl_interface': 9.908878555132918e-06, 'learning_rate/action_model': 9.908878555132918e-05, 'epoch': 0.03}) +[2026-10-03 14:42:26,936][starVLA.training.train_starvla][INFO] - Step 407, Loss: {'train/loss': 0.08349089324474335, 'train/grad_norm_pre_clip': 22.560894012451172, 'timing/data': 0.000370979993022047, 'timing/model': 0.36674762700567953, 'learning_rate/qwen_vl_interface': 9.908283944142391e-06, 'learning_rate/action_model': 9.908283944142391e-05, 'epoch': 0.03}) +[2026-10-03 14:42:27,588][starVLA.training.train_starvla][INFO] - Step 408, Loss: {'train/loss': 0.20141734182834625, 'train/grad_norm_pre_clip': 72.42967224121094, 'timing/data': 0.00024095000117085874, 'timing/model': 0.36674986699654255, 'learning_rate/qwen_vl_interface': 9.907687418307297e-06, 'learning_rate/action_model': 9.907687418307296e-05, 'epoch': 0.03}) +[2026-10-03 14:42:28,239][starVLA.training.train_starvla][INFO] - Step 409, Loss: {'train/loss': 0.056139200925827026, 'train/grad_norm_pre_clip': 26.56763458251953, 'timing/data': 0.0003251890011597425, 'timing/model': 0.36644403799436986, 'learning_rate/qwen_vl_interface': 9.907088977872837e-06, 'learning_rate/action_model': 9.907088977872837e-05, 'epoch': 0.03}) +[2026-10-03 14:42:29,081][starVLA.training.train_starvla][INFO] - Step 410, Loss: {'train/loss': 0.3735446333885193, 'train/grad_norm_pre_clip': 49.12153244018555, 'timing/data': 0.00023175000387709588, 'timing/model': 0.36668463799287565, 'learning_rate/qwen_vl_interface': 9.906488623085015e-06, 'learning_rate/action_model': 9.906488623085015e-05, 'epoch': 0.03}) +[2026-10-03 14:42:29,733][starVLA.training.train_starvla][INFO] - Step 411, Loss: {'train/loss': 0.05994519591331482, 'train/grad_norm_pre_clip': 27.632856369018555, 'timing/data': 0.00020771000708919019, 'timing/model': 0.3673769080050988, 'learning_rate/qwen_vl_interface': 9.905886354190611e-06, 'learning_rate/action_model': 9.90588635419061e-05, 'epoch': 0.03}) +[2026-10-03 14:42:30,390][starVLA.training.train_starvla][INFO] - Step 412, Loss: {'train/loss': 0.31092363595962524, 'train/grad_norm_pre_clip': 38.81443786621094, 'timing/data': 0.00030188000528141856, 'timing/model': 0.3668047370010754, 'learning_rate/qwen_vl_interface': 9.905282171437192e-06, 'learning_rate/action_model': 9.905282171437192e-05, 'epoch': 0.03}) +[2026-10-03 14:42:31,019][starVLA.training.train_starvla][INFO] - Step 413, Loss: {'train/loss': 0.06394004076719284, 'train/grad_norm_pre_clip': 33.41965866088867, 'timing/data': 0.00032928999280557036, 'timing/model': 0.3436013600003207, 'learning_rate/qwen_vl_interface': 9.904676075073121e-06, 'learning_rate/action_model': 9.904676075073121e-05, 'epoch': 0.03}) +[2026-10-03 14:42:31,671][starVLA.training.train_starvla][INFO] - Step 414, Loss: {'train/loss': 0.07955930382013321, 'train/grad_norm_pre_clip': 18.361083984375, 'timing/data': 0.0003005600010510534, 'timing/model': 0.36681217700242996, 'learning_rate/qwen_vl_interface': 9.904068065347536e-06, 'learning_rate/action_model': 9.904068065347536e-05, 'epoch': 0.03}) +[2026-10-03 14:42:32,511][starVLA.training.train_starvla][INFO] - Step 415, Loss: {'train/loss': 0.4499626159667969, 'train/grad_norm_pre_clip': 155.82986450195312, 'timing/data': 0.00025999000354204327, 'timing/model': 0.5553204150055535, 'learning_rate/qwen_vl_interface': 9.90345814251037e-06, 'learning_rate/action_model': 9.90345814251037e-05, 'epoch': 0.03}) +[2026-10-03 14:42:33,163][starVLA.training.train_starvla][INFO] - Step 416, Loss: {'train/loss': 0.16592003405094147, 'train/grad_norm_pre_clip': 41.94698715209961, 'timing/data': 0.0002968500048154965, 'timing/model': 0.3671550469880458, 'learning_rate/qwen_vl_interface': 9.902846306812339e-06, 'learning_rate/action_model': 9.902846306812338e-05, 'epoch': 0.03}) +[2026-10-03 14:42:33,817][starVLA.training.train_starvla][INFO] - Step 417, Loss: {'train/loss': 0.04227498546242714, 'train/grad_norm_pre_clip': 22.762601852416992, 'timing/data': 0.0004155899951001629, 'timing/model': 0.36832521700125653, 'learning_rate/qwen_vl_interface': 9.902232558504942e-06, 'learning_rate/action_model': 9.902232558504941e-05, 'epoch': 0.03}) +[2026-10-03 14:42:34,469][starVLA.training.train_starvla][INFO] - Step 418, Loss: {'train/loss': 0.06246402859687805, 'train/grad_norm_pre_clip': 46.78661346435547, 'timing/data': 0.0002033400087384507, 'timing/model': 0.36749629699625075, 'learning_rate/qwen_vl_interface': 9.901616897840471e-06, 'learning_rate/action_model': 9.90161689784047e-05, 'epoch': 0.03}) +[2026-10-03 14:42:35,121][starVLA.training.train_starvla][INFO] - Step 419, Loss: {'train/loss': 0.010857904329895973, 'train/grad_norm_pre_clip': 2.451842784881592, 'timing/data': 0.0002930899936472997, 'timing/model': 0.36691247700946406, 'learning_rate/qwen_vl_interface': 9.900999325072e-06, 'learning_rate/action_model': 9.900999325072001e-05, 'epoch': 0.03}) +[2026-10-03 14:42:35,971][starVLA.training.train_starvla][INFO] - Step 420, Loss: {'train/loss': 0.01913023553788662, 'train/grad_norm_pre_clip': 4.366938591003418, 'timing/data': 0.0003304899873910472, 'timing/model': 0.5641999950021273, 'learning_rate/qwen_vl_interface': 9.900379840453392e-06, 'learning_rate/action_model': 9.900379840453392e-05, 'epoch': 0.03}) +[2026-10-03 14:42:36,623][starVLA.training.train_starvla][INFO] - Step 421, Loss: {'train/loss': 0.5104599595069885, 'train/grad_norm_pre_clip': 73.18182373046875, 'timing/data': 0.0002501599956303835, 'timing/model': 0.36674135700741317, 'learning_rate/qwen_vl_interface': 9.89975844423929e-06, 'learning_rate/action_model': 9.899758444239289e-05, 'epoch': 0.03}) +[2026-10-03 14:42:37,275][starVLA.training.train_starvla][INFO] - Step 422, Loss: {'train/loss': 0.9056670069694519, 'train/grad_norm_pre_clip': 55.484004974365234, 'timing/data': 0.0002605600020615384, 'timing/model': 0.3669110969931353, 'learning_rate/qwen_vl_interface': 9.899135136685131e-06, 'learning_rate/action_model': 9.89913513668513e-05, 'epoch': 0.03}) +[2026-10-03 14:42:37,927][starVLA.training.train_starvla][INFO] - Step 423, Loss: {'train/loss': 0.21809415519237518, 'train/grad_norm_pre_clip': 109.12649536132812, 'timing/data': 0.00021152000408619642, 'timing/model': 0.3672186969924951, 'learning_rate/qwen_vl_interface': 9.898509918047127e-06, 'learning_rate/action_model': 9.898509918047126e-05, 'epoch': 0.03}) +[2026-10-03 14:42:38,578][starVLA.training.train_starvla][INFO] - Step 424, Loss: {'train/loss': 0.03837815299630165, 'train/grad_norm_pre_clip': 23.681137084960938, 'timing/data': 0.0003742199914995581, 'timing/model': 0.3666852280002786, 'learning_rate/qwen_vl_interface': 9.897882788582286e-06, 'learning_rate/action_model': 9.897882788582287e-05, 'epoch': 0.03}) +[2026-10-03 14:42:39,230][starVLA.training.train_starvla][INFO] - Step 425, Loss: {'train/loss': 0.7926517724990845, 'train/grad_norm_pre_clip': 162.5557403564453, 'timing/data': 0.00029414000164251775, 'timing/model': 0.36667787699843757, 'learning_rate/qwen_vl_interface': 9.897253748548398e-06, 'learning_rate/action_model': 9.897253748548399e-05, 'epoch': 0.03}) +[2026-10-03 14:42:40,050][starVLA.training.train_starvla][INFO] - Step 426, Loss: {'train/loss': 0.0023610182106494904, 'train/grad_norm_pre_clip': 0.529380202293396, 'timing/data': 0.0002764200035016984, 'timing/model': 0.3435823899926618, 'learning_rate/qwen_vl_interface': 9.896622798204035e-06, 'learning_rate/action_model': 9.896622798204035e-05, 'epoch': 0.03}) +[2026-10-03 14:42:40,702][starVLA.training.train_starvla][INFO] - Step 427, Loss: {'train/loss': 0.12517084181308746, 'train/grad_norm_pre_clip': 131.58689880371094, 'timing/data': 0.0002445699938107282, 'timing/model': 0.36717899701034185, 'learning_rate/qwen_vl_interface': 9.89598993780856e-06, 'learning_rate/action_model': 9.89598993780856e-05, 'epoch': 0.03}) +[2026-10-03 14:42:41,330][starVLA.training.train_starvla][INFO] - Step 428, Loss: {'train/loss': 0.542444109916687, 'train/grad_norm_pre_clip': 174.20997619628906, 'timing/data': 0.0002427099971100688, 'timing/model': 0.3437118699948769, 'learning_rate/qwen_vl_interface': 9.895355167622113e-06, 'learning_rate/action_model': 9.895355167622113e-05, 'epoch': 0.03}) +[2026-10-03 14:42:41,981][starVLA.training.train_starvla][INFO] - Step 429, Loss: {'train/loss': 0.0015541489701718092, 'train/grad_norm_pre_clip': 0.6482000350952148, 'timing/data': 0.0002978699922095984, 'timing/model': 0.3657282780040987, 'learning_rate/qwen_vl_interface': 9.894718487905629e-06, 'learning_rate/action_model': 9.894718487905629e-05, 'epoch': 0.03}) +[2026-10-03 14:42:42,632][starVLA.training.train_starvla][INFO] - Step 430, Loss: {'train/loss': 1.3071885108947754, 'train/grad_norm_pre_clip': 119.70787811279297, 'timing/data': 0.00030870000773575157, 'timing/model': 0.36685492799733765, 'learning_rate/qwen_vl_interface': 9.89407989892082e-06, 'learning_rate/action_model': 9.894079898920821e-05, 'epoch': 0.03}) +[2026-10-03 14:42:43,481][starVLA.training.train_starvla][INFO] - Step 431, Loss: {'train/loss': 0.6342116594314575, 'train/grad_norm_pre_clip': 149.25506591796875, 'timing/data': 0.00026015000184997916, 'timing/model': 0.36681156800477766, 'learning_rate/qwen_vl_interface': 9.893439400930187e-06, 'learning_rate/action_model': 9.893439400930187e-05, 'epoch': 0.03}) +[2026-10-03 14:42:44,132][starVLA.training.train_starvla][INFO] - Step 432, Loss: {'train/loss': 0.46267786622047424, 'train/grad_norm_pre_clip': 55.86607360839844, 'timing/data': 0.00020296000002417713, 'timing/model': 0.3666655380075099, 'learning_rate/qwen_vl_interface': 9.892796994197014e-06, 'learning_rate/action_model': 9.892796994197014e-05, 'epoch': 0.03}) +[2026-10-03 14:42:44,783][starVLA.training.train_starvla][INFO] - Step 433, Loss: {'train/loss': 0.003133677178993821, 'train/grad_norm_pre_clip': 1.2207789421081543, 'timing/data': 0.000205939999432303, 'timing/model': 0.366486878003343, 'learning_rate/qwen_vl_interface': 9.892152678985372e-06, 'learning_rate/action_model': 9.892152678985372e-05, 'epoch': 0.03}) +[2026-10-03 14:42:45,435][starVLA.training.train_starvla][INFO] - Step 434, Loss: {'train/loss': 0.14931988716125488, 'train/grad_norm_pre_clip': 163.48159790039062, 'timing/data': 0.00020434000180102885, 'timing/model': 0.3673923280002782, 'learning_rate/qwen_vl_interface': 9.891506455560111e-06, 'learning_rate/action_model': 9.891506455560111e-05, 'epoch': 0.03}) +[2026-10-03 14:42:46,087][starVLA.training.train_starvla][INFO] - Step 435, Loss: {'train/loss': 0.0779726579785347, 'train/grad_norm_pre_clip': 136.04615783691406, 'timing/data': 0.000308190006762743, 'timing/model': 0.3668554070027312, 'learning_rate/qwen_vl_interface': 9.890858324186872e-06, 'learning_rate/action_model': 9.890858324186872e-05, 'epoch': 0.03}) +[2026-10-03 14:42:46,937][starVLA.training.train_starvla][INFO] - Step 436, Loss: {'train/loss': 0.1148419976234436, 'train/grad_norm_pre_clip': 114.33980560302734, 'timing/data': 0.00024449999909847975, 'timing/model': 0.5650495040026726, 'learning_rate/qwen_vl_interface': 9.890208285132077e-06, 'learning_rate/action_model': 9.890208285132077e-05, 'epoch': 0.03}) +[2026-10-03 14:42:47,588][starVLA.training.train_starvla][INFO] - Step 437, Loss: {'train/loss': 0.5423455834388733, 'train/grad_norm_pre_clip': 77.15259552001953, 'timing/data': 0.00038409999979194254, 'timing/model': 0.3664048179925885, 'learning_rate/qwen_vl_interface': 9.889556338662933e-06, 'learning_rate/action_model': 9.889556338662932e-05, 'epoch': 0.03}) +[2026-10-03 14:42:48,239][starVLA.training.train_starvla][INFO] - Step 438, Loss: {'train/loss': 0.46769919991493225, 'train/grad_norm_pre_clip': 38.26657485961914, 'timing/data': 0.0003630599967436865, 'timing/model': 0.3666302580095362, 'learning_rate/qwen_vl_interface': 9.888902485047427e-06, 'learning_rate/action_model': 9.888902485047427e-05, 'epoch': 0.03}) +[2026-10-03 14:42:48,892][starVLA.training.train_starvla][INFO] - Step 439, Loss: {'train/loss': 0.28909143805503845, 'train/grad_norm_pre_clip': 25.16301918029785, 'timing/data': 0.00020696999854408205, 'timing/model': 0.3675453970063245, 'learning_rate/qwen_vl_interface': 9.888246724554337e-06, 'learning_rate/action_model': 9.888246724554337e-05, 'epoch': 0.03}) +[2026-10-03 14:42:49,544][starVLA.training.train_starvla][INFO] - Step 440, Loss: {'train/loss': 0.0032116647344082594, 'train/grad_norm_pre_clip': 1.501247525215149, 'timing/data': 0.00026589000481180847, 'timing/model': 0.3670989370002644, 'learning_rate/qwen_vl_interface': 9.887589057453222e-06, 'learning_rate/action_model': 9.88758905745322e-05, 'epoch': 0.03}) +[2026-10-03 14:42:50,195][starVLA.training.train_starvla][INFO] - Step 441, Loss: {'train/loss': 0.7867562770843506, 'train/grad_norm_pre_clip': 80.24986267089844, 'timing/data': 0.0002015899954130873, 'timing/model': 0.3665625470021041, 'learning_rate/qwen_vl_interface': 9.88692948401442e-06, 'learning_rate/action_model': 9.88692948401442e-05, 'epoch': 0.03}) +[2026-10-03 14:42:51,039][starVLA.training.train_starvla][INFO] - Step 442, Loss: {'train/loss': 0.0013234344078227878, 'train/grad_norm_pre_clip': 0.5029686689376831, 'timing/data': 0.0003057100111618638, 'timing/model': 0.3672101870033657, 'learning_rate/qwen_vl_interface': 9.88626800450906e-06, 'learning_rate/action_model': 9.88626800450906e-05, 'epoch': 0.03}) +[2026-10-03 14:42:51,691][starVLA.training.train_starvla][INFO] - Step 443, Loss: {'train/loss': 0.15262557566165924, 'train/grad_norm_pre_clip': 68.86856079101562, 'timing/data': 0.000344169995514676, 'timing/model': 0.36658563699165825, 'learning_rate/qwen_vl_interface': 9.885604619209052e-06, 'learning_rate/action_model': 9.885604619209051e-05, 'epoch': 0.03}) +[2026-10-03 14:42:52,342][starVLA.training.train_starvla][INFO] - Step 444, Loss: {'train/loss': 1.008123517036438, 'train/grad_norm_pre_clip': 71.74352264404297, 'timing/data': 0.0003398099943296984, 'timing/model': 0.36675400800595526, 'learning_rate/qwen_vl_interface': 9.884939328387085e-06, 'learning_rate/action_model': 9.884939328387084e-05, 'epoch': 0.03}) +[2026-10-03 14:42:52,993][starVLA.training.train_starvla][INFO] - Step 445, Loss: {'train/loss': 1.3096370697021484, 'train/grad_norm_pre_clip': 112.46293640136719, 'timing/data': 0.00028513999131973833, 'timing/model': 0.3661388870095834, 'learning_rate/qwen_vl_interface': 9.884272132316638e-06, 'learning_rate/action_model': 9.884272132316637e-05, 'epoch': 0.03}) +[2026-10-03 14:42:53,644][starVLA.training.train_starvla][INFO] - Step 446, Loss: {'train/loss': 0.9402968883514404, 'train/grad_norm_pre_clip': 229.26156616210938, 'timing/data': 0.00020491999748628587, 'timing/model': 0.36686393700074404, 'learning_rate/qwen_vl_interface': 9.883603031271968e-06, 'learning_rate/action_model': 9.883603031271969e-05, 'epoch': 0.03}) +[2026-10-03 14:42:54,486][starVLA.training.train_starvla][INFO] - Step 447, Loss: {'train/loss': 0.7747747898101807, 'train/grad_norm_pre_clip': 128.46343994140625, 'timing/data': 0.0002535999956307933, 'timing/model': 0.5572024849971058, 'learning_rate/qwen_vl_interface': 9.882932025528122e-06, 'learning_rate/action_model': 9.882932025528121e-05, 'epoch': 0.03}) +[2026-10-03 14:42:55,138][starVLA.training.train_starvla][INFO] - Step 448, Loss: {'train/loss': 0.5032206177711487, 'train/grad_norm_pre_clip': 38.78126907348633, 'timing/data': 0.0001960789959412068, 'timing/model': 0.36711049800214823, 'learning_rate/qwen_vl_interface': 9.882259115360918e-06, 'learning_rate/action_model': 9.882259115360917e-05, 'epoch': 0.03}) +[2026-10-03 14:42:55,767][starVLA.training.train_starvla][INFO] - Step 449, Loss: {'train/loss': 0.5958857536315918, 'train/grad_norm_pre_clip': 61.71734619140625, 'timing/data': 0.00020537000091280788, 'timing/model': 0.36691426699690055, 'learning_rate/qwen_vl_interface': 9.881584301046968e-06, 'learning_rate/action_model': 9.881584301046968e-05, 'epoch': 0.03}) +[2026-10-03 14:42:56,419][starVLA.training.train_starvla][INFO] - Step 450, Loss: {'train/loss': 0.09066590666770935, 'train/grad_norm_pre_clip': 44.009185791015625, 'timing/data': 0.0003266499988967553, 'timing/model': 0.366389746995992, 'learning_rate/qwen_vl_interface': 9.880907582863662e-06, 'learning_rate/action_model': 9.880907582863662e-05, 'epoch': 0.03}) +[2026-10-03 14:42:57,047][starVLA.training.train_starvla][INFO] - Step 451, Loss: {'train/loss': 0.5001054406166077, 'train/grad_norm_pre_clip': 43.979408264160156, 'timing/data': 0.0002042299893219024, 'timing/model': 0.34412329000770114, 'learning_rate/qwen_vl_interface': 9.880228961089175e-06, 'learning_rate/action_model': 9.880228961089175e-05, 'epoch': 0.03}) +[2026-10-03 14:42:57,872][starVLA.training.train_starvla][INFO] - Step 452, Loss: {'train/loss': 0.5635961890220642, 'train/grad_norm_pre_clip': 61.25782012939453, 'timing/data': 0.0002038200036622584, 'timing/model': 0.5623160949908197, 'learning_rate/qwen_vl_interface': 9.87954843600246e-06, 'learning_rate/action_model': 9.879548436002459e-05, 'epoch': 0.03}) +[2026-10-03 14:42:58,523][starVLA.training.train_starvla][INFO] - Step 453, Loss: {'train/loss': 0.714470386505127, 'train/grad_norm_pre_clip': 47.248809814453125, 'timing/data': 0.00035013000888284296, 'timing/model': 0.36639696700149216, 'learning_rate/qwen_vl_interface': 9.878866007883256e-06, 'learning_rate/action_model': 9.878866007883257e-05, 'epoch': 0.03}) +[2026-10-03 14:42:59,175][starVLA.training.train_starvla][INFO] - Step 454, Loss: {'train/loss': 0.16642379760742188, 'train/grad_norm_pre_clip': 19.329421997070312, 'timing/data': 0.00020482999389059842, 'timing/model': 0.36698520700156223, 'learning_rate/qwen_vl_interface': 9.878181677012084e-06, 'learning_rate/action_model': 9.878181677012085e-05, 'epoch': 0.03}) +[2026-10-03 14:42:59,804][starVLA.training.train_starvla][INFO] - Step 455, Loss: {'train/loss': 0.26586198806762695, 'train/grad_norm_pre_clip': 25.073678970336914, 'timing/data': 0.0003269599983468652, 'timing/model': 0.36689375700370874, 'learning_rate/qwen_vl_interface': 9.877495443670247e-06, 'learning_rate/action_model': 9.877495443670246e-05, 'epoch': 0.03}) +[2026-10-03 14:43:00,456][starVLA.training.train_starvla][INFO] - Step 456, Loss: {'train/loss': 0.178708016872406, 'train/grad_norm_pre_clip': 33.0592041015625, 'timing/data': 0.0003105590003542602, 'timing/model': 0.3664777780068107, 'learning_rate/qwen_vl_interface': 9.876807308139828e-06, 'learning_rate/action_model': 9.876807308139829e-05, 'epoch': 0.03}) +[2026-10-03 14:43:01,106][starVLA.training.train_starvla][INFO] - Step 457, Loss: {'train/loss': 0.061232082545757294, 'train/grad_norm_pre_clip': 15.963708877563477, 'timing/data': 0.0002047099987976253, 'timing/model': 0.36634351800603326, 'learning_rate/qwen_vl_interface': 9.876117270703695e-06, 'learning_rate/action_model': 9.876117270703694e-05, 'epoch': 0.03}) +[2026-10-03 14:43:01,956][starVLA.training.train_starvla][INFO] - Step 458, Loss: {'train/loss': 0.10773938149213791, 'train/grad_norm_pre_clip': 85.93840026855469, 'timing/data': 0.00022409998928196728, 'timing/model': 0.366698437006562, 'learning_rate/qwen_vl_interface': 9.875425331645495e-06, 'learning_rate/action_model': 9.875425331645495e-05, 'epoch': 0.03}) +[2026-10-03 14:43:02,608][starVLA.training.train_starvla][INFO] - Step 459, Loss: {'train/loss': 0.6158266067504883, 'train/grad_norm_pre_clip': 119.12024688720703, 'timing/data': 0.00027831000625155866, 'timing/model': 0.36664770801144186, 'learning_rate/qwen_vl_interface': 9.874731491249658e-06, 'learning_rate/action_model': 9.874731491249657e-05, 'epoch': 0.03}) +[2026-10-03 14:43:03,260][starVLA.training.train_starvla][INFO] - Step 460, Loss: {'train/loss': 0.18808196485042572, 'train/grad_norm_pre_clip': 83.23948669433594, 'timing/data': 0.0003103600029135123, 'timing/model': 0.36674950701126363, 'learning_rate/qwen_vl_interface': 9.874035749801398e-06, 'learning_rate/action_model': 9.874035749801399e-05, 'epoch': 0.03}) +[2026-10-03 14:43:03,912][starVLA.training.train_starvla][INFO] - Step 461, Loss: {'train/loss': 0.45814239978790283, 'train/grad_norm_pre_clip': 22.950191497802734, 'timing/data': 0.0003732799959834665, 'timing/model': 0.3667500169976847, 'learning_rate/qwen_vl_interface': 9.873338107586704e-06, 'learning_rate/action_model': 9.873338107586705e-05, 'epoch': 0.03}) +[2026-10-03 14:43:04,540][starVLA.training.train_starvla][INFO] - Step 462, Loss: {'train/loss': 1.1572792530059814, 'train/grad_norm_pre_clip': 94.71497344970703, 'timing/data': 0.0002020599931711331, 'timing/model': 0.36635977699188516, 'learning_rate/qwen_vl_interface': 9.872638564892358e-06, 'learning_rate/action_model': 9.872638564892357e-05, 'epoch': 0.03}) +[2026-10-03 14:43:05,394][starVLA.training.train_starvla][INFO] - Step 463, Loss: {'train/loss': 0.6540148258209229, 'train/grad_norm_pre_clip': 96.85028076171875, 'timing/data': 0.00020959899120498449, 'timing/model': 0.367381967997062, 'learning_rate/qwen_vl_interface': 9.871937122005905e-06, 'learning_rate/action_model': 9.871937122005905e-05, 'epoch': 0.03}) +[2026-10-03 14:43:06,046][starVLA.training.train_starvla][INFO] - Step 464, Loss: {'train/loss': 0.36383792757987976, 'train/grad_norm_pre_clip': 58.274959564208984, 'timing/data': 0.00032669000211171806, 'timing/model': 0.36678612799732946, 'learning_rate/qwen_vl_interface': 9.87123377921569e-06, 'learning_rate/action_model': 9.871233779215689e-05, 'epoch': 0.03}) +[2026-10-03 14:43:06,698][starVLA.training.train_starvla][INFO] - Step 465, Loss: {'train/loss': 0.783013105392456, 'train/grad_norm_pre_clip': 45.415767669677734, 'timing/data': 0.0002576199913164601, 'timing/model': 0.36618173800525256, 'learning_rate/qwen_vl_interface': 9.870528536810825e-06, 'learning_rate/action_model': 9.870528536810825e-05, 'epoch': 0.03}) +[2026-10-03 14:43:07,349][starVLA.training.train_starvla][INFO] - Step 466, Loss: {'train/loss': 0.055496275424957275, 'train/grad_norm_pre_clip': 87.36066436767578, 'timing/data': 0.0002030100004049018, 'timing/model': 0.3665247669996461, 'learning_rate/qwen_vl_interface': 9.869821395081214e-06, 'learning_rate/action_model': 9.869821395081214e-05, 'epoch': 0.03}) +[2026-10-03 14:43:08,001][starVLA.training.train_starvla][INFO] - Step 467, Loss: {'train/loss': 0.4714050889015198, 'train/grad_norm_pre_clip': 54.2667236328125, 'timing/data': 0.00029889999132137746, 'timing/model': 0.3670166770025389, 'learning_rate/qwen_vl_interface': 9.869112354317532e-06, 'learning_rate/action_model': 9.869112354317532e-05, 'epoch': 0.03}) +[2026-10-03 14:43:08,850][starVLA.training.train_starvla][INFO] - Step 468, Loss: {'train/loss': 0.2810540795326233, 'train/grad_norm_pre_clip': 27.492942810058594, 'timing/data': 0.0003089300007559359, 'timing/model': 0.5633180540025933, 'learning_rate/qwen_vl_interface': 9.86840141481124e-06, 'learning_rate/action_model': 9.868401414811239e-05, 'epoch': 0.03}) +[2026-10-03 14:43:09,501][starVLA.training.train_starvla][INFO] - Step 469, Loss: {'train/loss': 0.28907084465026855, 'train/grad_norm_pre_clip': 39.4588737487793, 'timing/data': 0.0003199299972038716, 'timing/model': 0.3664202479994856, 'learning_rate/qwen_vl_interface': 9.867688576854578e-06, 'learning_rate/action_model': 9.867688576854578e-05, 'epoch': 0.03}) +[2026-10-03 14:43:10,152][starVLA.training.train_starvla][INFO] - Step 470, Loss: {'train/loss': 0.3264826834201813, 'train/grad_norm_pre_clip': 60.6090202331543, 'timing/data': 0.00024339000810869038, 'timing/model': 0.366625868002302, 'learning_rate/qwen_vl_interface': 9.86697384074057e-06, 'learning_rate/action_model': 9.866973840740568e-05, 'epoch': 0.03}) +[2026-10-03 14:43:10,804][starVLA.training.train_starvla][INFO] - Step 471, Loss: {'train/loss': 0.21820886433124542, 'train/grad_norm_pre_clip': 47.77943420410156, 'timing/data': 0.0002639699960127473, 'timing/model': 0.36742136799148284, 'learning_rate/qwen_vl_interface': 9.86625720676301e-06, 'learning_rate/action_model': 9.86625720676301e-05, 'epoch': 0.03}) +[2026-10-03 14:43:11,457][starVLA.training.train_starvla][INFO] - Step 472, Loss: {'train/loss': 0.1001637652516365, 'train/grad_norm_pre_clip': 31.617883682250977, 'timing/data': 0.00020766000670846552, 'timing/model': 0.36709613699349575, 'learning_rate/qwen_vl_interface': 9.865538675216488e-06, 'learning_rate/action_model': 9.865538675216486e-05, 'epoch': 0.03}) +[2026-10-03 14:43:12,300][starVLA.training.train_starvla][INFO] - Step 473, Loss: {'train/loss': 0.2696620225906372, 'train/grad_norm_pre_clip': 61.23011779785156, 'timing/data': 0.00024464000307489187, 'timing/model': 0.5591835750092287, 'learning_rate/qwen_vl_interface': 9.864818246396358e-06, 'learning_rate/action_model': 9.864818246396358e-05, 'epoch': 0.03}) +[2026-10-03 14:43:12,952][starVLA.training.train_starvla][INFO] - Step 474, Loss: {'train/loss': 0.8026860356330872, 'train/grad_norm_pre_clip': 79.69853210449219, 'timing/data': 0.000295409990940243, 'timing/model': 0.36681908700848, 'learning_rate/qwen_vl_interface': 9.864095920598764e-06, 'learning_rate/action_model': 9.864095920598763e-05, 'epoch': 0.03}) +[2026-10-03 14:43:13,580][starVLA.training.train_starvla][INFO] - Step 475, Loss: {'train/loss': 0.07127504050731659, 'train/grad_norm_pre_clip': 16.309669494628906, 'timing/data': 0.00024793000193312764, 'timing/model': 0.34401470999000594, 'learning_rate/qwen_vl_interface': 9.86337169812063e-06, 'learning_rate/action_model': 9.863371698120629e-05, 'epoch': 0.03}) +[2026-10-03 14:43:14,232][starVLA.training.train_starvla][INFO] - Step 476, Loss: {'train/loss': 0.1730402708053589, 'train/grad_norm_pre_clip': 44.29210662841797, 'timing/data': 0.00020229999790899456, 'timing/model': 0.36675090801145416, 'learning_rate/qwen_vl_interface': 9.862645579259648e-06, 'learning_rate/action_model': 9.862645579259648e-05, 'epoch': 0.03}) +[2026-10-03 14:43:14,861][starVLA.training.train_starvla][INFO] - Step 477, Loss: {'train/loss': 0.2699911594390869, 'train/grad_norm_pre_clip': 46.325191497802734, 'timing/data': 0.0002927799941971898, 'timing/model': 0.3436659709986998, 'learning_rate/qwen_vl_interface': 9.861917564314308e-06, 'learning_rate/action_model': 9.861917564314308e-05, 'epoch': 0.03}) +[2026-10-03 14:43:15,513][starVLA.training.train_starvla][INFO] - Step 478, Loss: {'train/loss': 0.21181897819042206, 'train/grad_norm_pre_clip': 29.212635040283203, 'timing/data': 0.0002676500007510185, 'timing/model': 0.3669227369973669, 'learning_rate/qwen_vl_interface': 9.861187653583862e-06, 'learning_rate/action_model': 9.861187653583862e-05, 'epoch': 0.03}) +[2026-10-03 14:43:16,355][starVLA.training.train_starvla][INFO] - Step 479, Loss: {'train/loss': 0.12518930435180664, 'train/grad_norm_pre_clip': 20.34508514404297, 'timing/data': 0.00025711000489536673, 'timing/model': 0.3667657379992306, 'learning_rate/qwen_vl_interface': 9.860455847368352e-06, 'learning_rate/action_model': 9.860455847368351e-05, 'epoch': 0.04}) +[2026-10-03 14:43:17,006][starVLA.training.train_starvla][INFO] - Step 480, Loss: {'train/loss': 0.37960705161094666, 'train/grad_norm_pre_clip': 55.068572998046875, 'timing/data': 0.0003292199980933219, 'timing/model': 0.3667342669941718, 'learning_rate/qwen_vl_interface': 9.859722145968595e-06, 'learning_rate/action_model': 9.859722145968595e-05, 'epoch': 0.04}) +[2026-10-03 14:43:17,657][starVLA.training.train_starvla][INFO] - Step 481, Loss: {'train/loss': 0.0750008076429367, 'train/grad_norm_pre_clip': 22.381366729736328, 'timing/data': 0.0003393299994058907, 'timing/model': 0.3661938069999451, 'learning_rate/qwen_vl_interface': 9.858986549686188e-06, 'learning_rate/action_model': 9.858986549686188e-05, 'epoch': 0.04}) +[2026-10-03 14:43:18,308][starVLA.training.train_starvla][INFO] - Step 482, Loss: {'train/loss': 0.3815534710884094, 'train/grad_norm_pre_clip': 42.91883850097656, 'timing/data': 0.00028639999800361693, 'timing/model': 0.36680171800253447, 'learning_rate/qwen_vl_interface': 9.85824905882351e-06, 'learning_rate/action_model': 9.858249058823509e-05, 'epoch': 0.04}) +[2026-10-03 14:43:18,959][starVLA.training.train_starvla][INFO] - Step 483, Loss: {'train/loss': 0.23361912369728088, 'train/grad_norm_pre_clip': 35.914737701416016, 'timing/data': 0.0002814600011333823, 'timing/model': 0.3667821270064451, 'learning_rate/qwen_vl_interface': 9.857509673683709e-06, 'learning_rate/action_model': 9.857509673683708e-05, 'epoch': 0.04}) +[2026-10-03 14:43:19,810][starVLA.training.train_starvla][INFO] - Step 484, Loss: {'train/loss': 0.1111026257276535, 'train/grad_norm_pre_clip': 42.55402374267578, 'timing/data': 0.0003161999920848757, 'timing/model': 0.3666630069928942, 'learning_rate/qwen_vl_interface': 9.856768394570725e-06, 'learning_rate/action_model': 9.856768394570723e-05, 'epoch': 0.04}) +[2026-10-03 14:43:20,439][starVLA.training.train_starvla][INFO] - Step 485, Loss: {'train/loss': 0.009208779782056808, 'train/grad_norm_pre_clip': 1.632431149482727, 'timing/data': 0.00025228998856619, 'timing/model': 0.3660327270044945, 'learning_rate/qwen_vl_interface': 9.856025221789264e-06, 'learning_rate/action_model': 9.856025221789264e-05, 'epoch': 0.04}) +[2026-10-03 14:43:21,091][starVLA.training.train_starvla][INFO] - Step 486, Loss: {'train/loss': 0.007900884374976158, 'train/grad_norm_pre_clip': 1.4519994258880615, 'timing/data': 0.00030934999813325703, 'timing/model': 0.36717571700864937, 'learning_rate/qwen_vl_interface': 9.85528015564482e-06, 'learning_rate/action_model': 9.85528015564482e-05, 'epoch': 0.04}) +[2026-10-03 14:43:21,721][starVLA.training.train_starvla][INFO] - Step 487, Loss: {'train/loss': 0.35126087069511414, 'train/grad_norm_pre_clip': 43.339447021484375, 'timing/data': 0.0002512700011720881, 'timing/model': 0.3439993400097592, 'learning_rate/qwen_vl_interface': 9.854533196443661e-06, 'learning_rate/action_model': 9.854533196443661e-05, 'epoch': 0.04}) +[2026-10-03 14:43:22,373][starVLA.training.train_starvla][INFO] - Step 488, Loss: {'train/loss': 0.015613887459039688, 'train/grad_norm_pre_clip': 8.457715034484863, 'timing/data': 0.0002861500106519088, 'timing/model': 0.36654344700218644, 'learning_rate/qwen_vl_interface': 9.853784344492833e-06, 'learning_rate/action_model': 9.853784344492833e-05, 'epoch': 0.04}) +[2026-10-03 14:43:23,223][starVLA.training.train_starvla][INFO] - Step 489, Loss: {'train/loss': 0.004415314644575119, 'train/grad_norm_pre_clip': 1.7247062921524048, 'timing/data': 0.00034545999369584024, 'timing/model': 0.5654152939969208, 'learning_rate/qwen_vl_interface': 9.853033600100162e-06, 'learning_rate/action_model': 9.853033600100162e-05, 'epoch': 0.04}) +[2026-10-03 14:43:23,874][starVLA.training.train_starvla][INFO] - Step 490, Loss: {'train/loss': 0.009652998298406601, 'train/grad_norm_pre_clip': 7.51836633682251, 'timing/data': 0.00029579999682027847, 'timing/model': 0.3666783270018641, 'learning_rate/qwen_vl_interface': 9.85228096357425e-06, 'learning_rate/action_model': 9.852280963574249e-05, 'epoch': 0.04}) +[2026-10-03 14:43:24,524][starVLA.training.train_starvla][INFO] - Step 491, Loss: {'train/loss': 0.41517260670661926, 'train/grad_norm_pre_clip': 31.54413604736328, 'timing/data': 0.00035010999999940395, 'timing/model': 0.3666767779941438, 'learning_rate/qwen_vl_interface': 9.851526435224475e-06, 'learning_rate/action_model': 9.851526435224474e-05, 'epoch': 0.04}) +[2026-10-03 14:43:25,176][starVLA.training.train_starvla][INFO] - Step 492, Loss: {'train/loss': 0.0013925159582868218, 'train/grad_norm_pre_clip': 0.29480183124542236, 'timing/data': 0.0003132800047751516, 'timing/model': 0.36642814699735027, 'learning_rate/qwen_vl_interface': 9.850770015360999e-06, 'learning_rate/action_model': 9.850770015360999e-05, 'epoch': 0.04}) +[2026-10-03 14:43:25,827][starVLA.training.train_starvla][INFO] - Step 493, Loss: {'train/loss': 1.0613142251968384, 'train/grad_norm_pre_clip': 151.2557830810547, 'timing/data': 0.0003467900096438825, 'timing/model': 0.3660704779904336, 'learning_rate/qwen_vl_interface': 9.850011704294755e-06, 'learning_rate/action_model': 9.850011704294755e-05, 'epoch': 0.04}) +[2026-10-03 14:43:26,677][starVLA.training.train_starvla][INFO] - Step 494, Loss: {'train/loss': 0.16921478509902954, 'train/grad_norm_pre_clip': 176.15460205078125, 'timing/data': 0.0003602699871407822, 'timing/model': 0.565224534002482, 'learning_rate/qwen_vl_interface': 9.849251502337458e-06, 'learning_rate/action_model': 9.849251502337458e-05, 'epoch': 0.04}) +[2026-10-03 14:43:27,328][starVLA.training.train_starvla][INFO] - Step 495, Loss: {'train/loss': 0.013650055974721909, 'train/grad_norm_pre_clip': 25.114330291748047, 'timing/data': 0.0002744899975368753, 'timing/model': 0.36711578800168354, 'learning_rate/qwen_vl_interface': 9.848489409801598e-06, 'learning_rate/action_model': 9.848489409801598e-05, 'epoch': 0.04}) +[2026-10-03 14:43:27,979][starVLA.training.train_starvla][INFO] - Step 496, Loss: {'train/loss': 0.004865587688982487, 'train/grad_norm_pre_clip': 4.53016996383667, 'timing/data': 0.00033734999306034297, 'timing/model': 0.3664828270120779, 'learning_rate/qwen_vl_interface': 9.84772542700044e-06, 'learning_rate/action_model': 9.84772542700044e-05, 'epoch': 0.04}) +[2026-10-03 14:43:28,630][starVLA.training.train_starvla][INFO] - Step 497, Loss: {'train/loss': 0.0003279416705481708, 'train/grad_norm_pre_clip': 0.1710805743932724, 'timing/data': 0.00024482001026626676, 'timing/model': 0.3664727569994284, 'learning_rate/qwen_vl_interface': 9.846959554248032e-06, 'learning_rate/action_model': 9.84695955424803e-05, 'epoch': 0.04}) +[2026-10-03 14:43:29,282][starVLA.training.train_starvla][INFO] - Step 498, Loss: {'train/loss': 1.182311773300171, 'train/grad_norm_pre_clip': 285.28240966796875, 'timing/data': 0.0003266400017309934, 'timing/model': 0.36639941700559575, 'learning_rate/qwen_vl_interface': 9.846191791859192e-06, 'learning_rate/action_model': 9.846191791859192e-05, 'epoch': 0.04}) +[2026-10-03 14:43:29,934][starVLA.training.train_starvla][INFO] - Step 499, Loss: {'train/loss': 1.2114777565002441, 'train/grad_norm_pre_clip': 55.446537017822266, 'timing/data': 0.0003120299952570349, 'timing/model': 0.367122556999675, 'learning_rate/qwen_vl_interface': 9.845422140149518e-06, 'learning_rate/action_model': 9.845422140149519e-05, 'epoch': 0.04}) +[2026-10-03 14:43:30,785][starVLA.training.train_starvla][INFO] - Step 500, Loss: {'train/loss': 0.29336613416671753, 'train/grad_norm_pre_clip': 43.24091720581055, 'timing/data': 0.0003068999940296635, 'timing/model': 0.36615278699900955, 'learning_rate/qwen_vl_interface': 9.84465059943539e-06, 'learning_rate/action_model': 9.84465059943539e-05, 'epoch': 0.04}) +[2026-10-03 14:43:31,436][starVLA.training.train_starvla][INFO] - Step 501, Loss: {'train/loss': 0.7666529417037964, 'train/grad_norm_pre_clip': 327.7936706542969, 'timing/data': 0.0002388499997323379, 'timing/model': 0.36659679700096603, 'learning_rate/qwen_vl_interface': 9.843877170033956e-06, 'learning_rate/action_model': 9.843877170033954e-05, 'epoch': 0.04}) +[2026-10-03 14:43:32,087][starVLA.training.train_starvla][INFO] - Step 502, Loss: {'train/loss': 1.2101280689239502, 'train/grad_norm_pre_clip': 63.983577728271484, 'timing/data': 0.0002186700003221631, 'timing/model': 0.36657505700713955, 'learning_rate/qwen_vl_interface': 9.84310185226314e-06, 'learning_rate/action_model': 9.84310185226314e-05, 'epoch': 0.04}) +[2026-10-03 14:43:32,716][starVLA.training.train_starvla][INFO] - Step 503, Loss: {'train/loss': 1.250800609588623, 'train/grad_norm_pre_clip': 176.29586791992188, 'timing/data': 0.00027299999783281237, 'timing/model': 0.3440400399995269, 'learning_rate/qwen_vl_interface': 9.842324646441652e-06, 'learning_rate/action_model': 9.842324646441652e-05, 'epoch': 0.04}) +[2026-10-03 14:43:33,367][starVLA.training.train_starvla][INFO] - Step 504, Loss: {'train/loss': 0.36125943064689636, 'train/grad_norm_pre_clip': 50.50596618652344, 'timing/data': 0.00024188999668695033, 'timing/model': 0.3669223670003703, 'learning_rate/qwen_vl_interface': 9.841545552888968e-06, 'learning_rate/action_model': 9.841545552888967e-05, 'epoch': 0.04}) +[2026-10-03 14:43:34,198][starVLA.training.train_starvla][INFO] - Step 505, Loss: {'train/loss': 0.24995316565036774, 'train/grad_norm_pre_clip': 121.99102020263672, 'timing/data': 0.00024962000316008925, 'timing/model': 0.5459995470009744, 'learning_rate/qwen_vl_interface': 9.840764571925346e-06, 'learning_rate/action_model': 9.840764571925345e-05, 'epoch': 0.04}) +[2026-10-03 14:43:34,804][starVLA.training.train_starvla][INFO] - Step 506, Loss: {'train/loss': 0.1305609494447708, 'train/grad_norm_pre_clip': 32.91838073730469, 'timing/data': 0.0002555999963078648, 'timing/model': 0.3434144600032596, 'learning_rate/qwen_vl_interface': 9.839981703871815e-06, 'learning_rate/action_model': 9.839981703871814e-05, 'epoch': 0.04}) +[2026-10-03 14:43:35,456][starVLA.training.train_starvla][INFO] - Step 507, Loss: {'train/loss': 1.171130895614624, 'train/grad_norm_pre_clip': 452.8472900390625, 'timing/data': 0.0003098500019405037, 'timing/model': 0.36704223799461033, 'learning_rate/qwen_vl_interface': 9.839196949050185e-06, 'learning_rate/action_model': 9.839196949050185e-05, 'epoch': 0.04}) +[2026-10-03 14:43:36,109][starVLA.training.train_starvla][INFO] - Step 508, Loss: {'train/loss': 0.18256907165050507, 'train/grad_norm_pre_clip': 148.39447021484375, 'timing/data': 0.00024711000151000917, 'timing/model': 0.3669398269994417, 'learning_rate/qwen_vl_interface': 9.838410307783038e-06, 'learning_rate/action_model': 9.838410307783038e-05, 'epoch': 0.04}) +[2026-10-03 14:43:36,762][starVLA.training.train_starvla][INFO] - Step 509, Loss: {'train/loss': 0.014237855561077595, 'train/grad_norm_pre_clip': 28.851884841918945, 'timing/data': 0.0003368600009707734, 'timing/model': 0.3673834369983524, 'learning_rate/qwen_vl_interface': 9.837621780393734e-06, 'learning_rate/action_model': 9.837621780393733e-05, 'epoch': 0.04}) +[2026-10-03 14:43:37,391][starVLA.training.train_starvla][INFO] - Step 510, Loss: {'train/loss': 0.6284124851226807, 'train/grad_norm_pre_clip': 117.95974731445312, 'timing/data': 0.00030841999978292733, 'timing/model': 0.3443703499942785, 'learning_rate/qwen_vl_interface': 9.836831367206406e-06, 'learning_rate/action_model': 9.836831367206406e-05, 'epoch': 0.04}) +[2026-10-03 14:43:38,209][starVLA.training.train_starvla][INFO] - Step 511, Loss: {'train/loss': 0.006834153085947037, 'train/grad_norm_pre_clip': 5.446167945861816, 'timing/data': 0.0002443900011712685, 'timing/model': 0.555629994996707, 'learning_rate/qwen_vl_interface': 9.836039068545962e-06, 'learning_rate/action_model': 9.836039068545961e-05, 'epoch': 0.04}) +[2026-10-03 14:43:38,860][starVLA.training.train_starvla][INFO] - Step 512, Loss: {'train/loss': 0.46159905195236206, 'train/grad_norm_pre_clip': 117.62631225585938, 'timing/data': 0.00019613999756984413, 'timing/model': 0.36663067700283136, 'learning_rate/qwen_vl_interface': 9.835244884738086e-06, 'learning_rate/action_model': 9.835244884738085e-05, 'epoch': 0.04}) +[2026-10-03 14:43:39,511][starVLA.training.train_starvla][INFO] - Step 513, Loss: {'train/loss': 0.4994077682495117, 'train/grad_norm_pre_clip': 269.0764465332031, 'timing/data': 0.0002028299932135269, 'timing/model': 0.36645101699105, 'learning_rate/qwen_vl_interface': 9.83444881610924e-06, 'learning_rate/action_model': 9.834448816109238e-05, 'epoch': 0.04}) +[2026-10-03 14:43:40,163][starVLA.training.train_starvla][INFO] - Step 514, Loss: {'train/loss': 0.001081367488950491, 'train/grad_norm_pre_clip': 0.4000810980796814, 'timing/data': 0.0003213900054106489, 'timing/model': 0.36663349700393155, 'learning_rate/qwen_vl_interface': 9.833650862986652e-06, 'learning_rate/action_model': 9.833650862986653e-05, 'epoch': 0.04}) +[2026-10-03 14:43:40,814][starVLA.training.train_starvla][INFO] - Step 515, Loss: {'train/loss': 0.19430528581142426, 'train/grad_norm_pre_clip': 181.7600555419922, 'timing/data': 0.00029896000341977924, 'timing/model': 0.36689249699702486, 'learning_rate/qwen_vl_interface': 9.832851025698335e-06, 'learning_rate/action_model': 9.832851025698335e-05, 'epoch': 0.04}) +[2026-10-03 14:43:41,665][starVLA.training.train_starvla][INFO] - Step 516, Loss: {'train/loss': 0.04627572000026703, 'train/grad_norm_pre_clip': 59.990318298339844, 'timing/data': 0.00029429000278469175, 'timing/model': 0.5655914539966034, 'learning_rate/qwen_vl_interface': 9.832049304573072e-06, 'learning_rate/action_model': 9.832049304573072e-05, 'epoch': 0.04}) +[2026-10-03 14:43:42,316][starVLA.training.train_starvla][INFO] - Step 517, Loss: {'train/loss': 0.4253724217414856, 'train/grad_norm_pre_clip': 44.92701721191406, 'timing/data': 0.00023048999719321728, 'timing/model': 0.36619100700772833, 'learning_rate/qwen_vl_interface': 9.831245699940419e-06, 'learning_rate/action_model': 9.831245699940418e-05, 'epoch': 0.04}) +[2026-10-03 14:43:42,967][starVLA.training.train_starvla][INFO] - Step 518, Loss: {'train/loss': 0.3380224108695984, 'train/grad_norm_pre_clip': 16.198354721069336, 'timing/data': 0.00029630000062752515, 'timing/model': 0.3666643379983725, 'learning_rate/qwen_vl_interface': 9.830440212130707e-06, 'learning_rate/action_model': 9.830440212130708e-05, 'epoch': 0.04}) +[2026-10-03 14:43:43,619][starVLA.training.train_starvla][INFO] - Step 519, Loss: {'train/loss': 0.42166781425476074, 'train/grad_norm_pre_clip': 58.2021598815918, 'timing/data': 0.00030775000050198287, 'timing/model': 0.367112346997601, 'learning_rate/qwen_vl_interface': 9.829632841475041e-06, 'learning_rate/action_model': 9.829632841475041e-05, 'epoch': 0.04}) +[2026-10-03 14:43:44,270][starVLA.training.train_starvla][INFO] - Step 520, Loss: {'train/loss': 0.48596516251564026, 'train/grad_norm_pre_clip': 81.35397338867188, 'timing/data': 0.00032543999259360135, 'timing/model': 0.3668091569998069, 'learning_rate/qwen_vl_interface': 9.828823588305305e-06, 'learning_rate/action_model': 9.828823588305303e-05, 'epoch': 0.04}) +[2026-10-03 14:43:45,120][starVLA.training.train_starvla][INFO] - Step 521, Loss: {'train/loss': 0.20620323717594147, 'train/grad_norm_pre_clip': 58.019248962402344, 'timing/data': 0.00023982999846339226, 'timing/model': 0.5649397550005233, 'learning_rate/qwen_vl_interface': 9.828012452954146e-06, 'learning_rate/action_model': 9.828012452954145e-05, 'epoch': 0.04}) +[2026-10-03 14:43:45,771][starVLA.training.train_starvla][INFO] - Step 522, Loss: {'train/loss': 0.574013352394104, 'train/grad_norm_pre_clip': 104.18023681640625, 'timing/data': 0.0003576099989004433, 'timing/model': 0.366171817004215, 'learning_rate/qwen_vl_interface': 9.827199435754995e-06, 'learning_rate/action_model': 9.827199435754995e-05, 'epoch': 0.04}) +[2026-10-03 14:43:46,422][starVLA.training.train_starvla][INFO] - Step 523, Loss: {'train/loss': 0.5696223974227905, 'train/grad_norm_pre_clip': 153.27755737304688, 'timing/data': 0.0003145900118397549, 'timing/model': 0.36644150799838826, 'learning_rate/qwen_vl_interface': 9.826384537042052e-06, 'learning_rate/action_model': 9.826384537042052e-05, 'epoch': 0.04}) +[2026-10-03 14:43:47,073][starVLA.training.train_starvla][INFO] - Step 524, Loss: {'train/loss': 0.30168581008911133, 'train/grad_norm_pre_clip': 73.02584838867188, 'timing/data': 0.00030610000248998404, 'timing/model': 0.36659707699436694, 'learning_rate/qwen_vl_interface': 9.825567757150291e-06, 'learning_rate/action_model': 9.82556775715029e-05, 'epoch': 0.04}) +[2026-10-03 14:43:47,724][starVLA.training.train_starvla][INFO] - Step 525, Loss: {'train/loss': 0.19654610753059387, 'train/grad_norm_pre_clip': 79.73650360107422, 'timing/data': 0.00019924899970646948, 'timing/model': 0.36646036800812, 'learning_rate/qwen_vl_interface': 9.82474909641546e-06, 'learning_rate/action_model': 9.82474909641546e-05, 'epoch': 0.04}) +[2026-10-03 14:43:48,375][starVLA.training.train_starvla][INFO] - Step 526, Loss: {'train/loss': 0.6359623670578003, 'train/grad_norm_pre_clip': 68.14297485351562, 'timing/data': 0.00033865000295918435, 'timing/model': 0.3664601170021342, 'learning_rate/qwen_vl_interface': 9.823928555174078e-06, 'learning_rate/action_model': 9.823928555174077e-05, 'epoch': 0.04}) +[2026-10-03 14:43:49,233][starVLA.training.train_starvla][INFO] - Step 527, Loss: {'train/loss': 0.2010602504014969, 'train/grad_norm_pre_clip': 55.16626739501953, 'timing/data': 0.00020179999410174787, 'timing/model': 0.36688587699609343, 'learning_rate/qwen_vl_interface': 9.82310613376344e-06, 'learning_rate/action_model': 9.82310613376344e-05, 'epoch': 0.04}) +[2026-10-03 14:43:49,884][starVLA.training.train_starvla][INFO] - Step 528, Loss: {'train/loss': 0.805854320526123, 'train/grad_norm_pre_clip': 66.1533432006836, 'timing/data': 0.00019511999562382698, 'timing/model': 0.3674405670026317, 'learning_rate/qwen_vl_interface': 9.822281832521613e-06, 'learning_rate/action_model': 9.822281832521613e-05, 'epoch': 0.04}) +[2026-10-03 14:43:50,536][starVLA.training.train_starvla][INFO] - Step 529, Loss: {'train/loss': 0.16396187245845795, 'train/grad_norm_pre_clip': 44.97144317626953, 'timing/data': 0.00035774000571109354, 'timing/model': 0.3668203280103626, 'learning_rate/qwen_vl_interface': 9.821455651787432e-06, 'learning_rate/action_model': 9.821455651787432e-05, 'epoch': 0.04}) +[2026-10-03 14:43:51,189][starVLA.training.train_starvla][INFO] - Step 530, Loss: {'train/loss': 0.2700914740562439, 'train/grad_norm_pre_clip': 32.075740814208984, 'timing/data': 0.000324180000461638, 'timing/model': 0.3667701770027634, 'learning_rate/qwen_vl_interface': 9.820627591900514e-06, 'learning_rate/action_model': 9.820627591900513e-05, 'epoch': 0.04}) +[2026-10-03 14:43:51,817][starVLA.training.train_starvla][INFO] - Step 531, Loss: {'train/loss': 0.4225391447544098, 'train/grad_norm_pre_clip': 30.309289932250977, 'timing/data': 0.0003027600032510236, 'timing/model': 0.3430173899978399, 'learning_rate/qwen_vl_interface': 9.819797653201238e-06, 'learning_rate/action_model': 9.819797653201238e-05, 'epoch': 0.04}) +[2026-10-03 14:43:52,663][starVLA.training.train_starvla][INFO] - Step 532, Loss: {'train/loss': 0.46232572197914124, 'train/grad_norm_pre_clip': 97.72285461425781, 'timing/data': 0.00023548000899609178, 'timing/model': 0.5599521249969257, 'learning_rate/qwen_vl_interface': 9.818965836030765e-06, 'learning_rate/action_model': 9.818965836030764e-05, 'epoch': 0.04}) +[2026-10-03 14:43:53,314][starVLA.training.train_starvla][INFO] - Step 533, Loss: {'train/loss': 0.30091798305511475, 'train/grad_norm_pre_clip': 20.377201080322266, 'timing/data': 0.00019492000865284353, 'timing/model': 0.36639022699091583, 'learning_rate/qwen_vl_interface': 9.81813214073102e-06, 'learning_rate/action_model': 9.81813214073102e-05, 'epoch': 0.04}) +[2026-10-03 14:43:53,965][starVLA.training.train_starvla][INFO] - Step 534, Loss: {'train/loss': 0.1254390925168991, 'train/grad_norm_pre_clip': 27.6384334564209, 'timing/data': 0.0003831790090771392, 'timing/model': 0.36680544799310155, 'learning_rate/qwen_vl_interface': 9.817296567644709e-06, 'learning_rate/action_model': 9.817296567644709e-05, 'epoch': 0.04}) +[2026-10-03 14:43:54,594][starVLA.training.train_starvla][INFO] - Step 535, Loss: {'train/loss': 0.08757898211479187, 'train/grad_norm_pre_clip': 15.21627426147461, 'timing/data': 0.0001972600002773106, 'timing/model': 0.3445495089981705, 'learning_rate/qwen_vl_interface': 9.816459117115296e-06, 'learning_rate/action_model': 9.816459117115297e-05, 'epoch': 0.04}) +[2026-10-03 14:43:55,247][starVLA.training.train_starvla][INFO] - Step 536, Loss: {'train/loss': 0.07641062885522842, 'train/grad_norm_pre_clip': 11.043810844421387, 'timing/data': 0.00022317998809739947, 'timing/model': 0.3673719969956437, 'learning_rate/qwen_vl_interface': 9.815619789487035e-06, 'learning_rate/action_model': 9.815619789487034e-05, 'epoch': 0.04}) +[2026-10-03 14:43:56,067][starVLA.training.train_starvla][INFO] - Step 537, Loss: {'train/loss': 0.05678142234683037, 'train/grad_norm_pre_clip': 15.301898956298828, 'timing/data': 0.0003511200047796592, 'timing/model': 0.5350484179944033, 'learning_rate/qwen_vl_interface': 9.814778585104933e-06, 'learning_rate/action_model': 9.814778585104933e-05, 'epoch': 0.04}) +[2026-10-03 14:43:56,718][starVLA.training.train_starvla][INFO] - Step 538, Loss: {'train/loss': 0.013470945879817009, 'train/grad_norm_pre_clip': 1.459354281425476, 'timing/data': 0.00032269000075757504, 'timing/model': 0.36656922700058203, 'learning_rate/qwen_vl_interface': 9.813935504314784e-06, 'learning_rate/action_model': 9.813935504314784e-05, 'epoch': 0.04}) +[2026-10-03 14:43:57,370][starVLA.training.train_starvla][INFO] - Step 539, Loss: {'train/loss': 0.06131131947040558, 'train/grad_norm_pre_clip': 51.7503776550293, 'timing/data': 0.00024003999715205282, 'timing/model': 0.36685236799530685, 'learning_rate/qwen_vl_interface': 9.813090547463144e-06, 'learning_rate/action_model': 9.813090547463143e-05, 'epoch': 0.04}) +[2026-10-03 14:43:58,022][starVLA.training.train_starvla][INFO] - Step 540, Loss: {'train/loss': 0.504357099533081, 'train/grad_norm_pre_clip': 41.6037483215332, 'timing/data': 0.0002831799938576296, 'timing/model': 0.36723179700493347, 'learning_rate/qwen_vl_interface': 9.812243714897342e-06, 'learning_rate/action_model': 9.812243714897343e-05, 'epoch': 0.04}) +[2026-10-03 14:43:58,673][starVLA.training.train_starvla][INFO] - Step 541, Loss: {'train/loss': 0.2198343724012375, 'train/grad_norm_pre_clip': 13.410099029541016, 'timing/data': 0.0003174699959345162, 'timing/model': 0.36619361699558794, 'learning_rate/qwen_vl_interface': 9.811395006965481e-06, 'learning_rate/action_model': 9.81139500696548e-05, 'epoch': 0.04}) +[2026-10-03 14:43:59,522][starVLA.training.train_starvla][INFO] - Step 542, Loss: {'train/loss': 0.5820907354354858, 'train/grad_norm_pre_clip': 34.72465896606445, 'timing/data': 0.0002671190013643354, 'timing/model': 0.5645408650016179, 'learning_rate/qwen_vl_interface': 9.81054442401643e-06, 'learning_rate/action_model': 9.81054442401643e-05, 'epoch': 0.04}) +[2026-10-03 14:44:00,174][starVLA.training.train_starvla][INFO] - Step 543, Loss: {'train/loss': 0.4889191687107086, 'train/grad_norm_pre_clip': 112.51211547851562, 'timing/data': 0.00019855999562423676, 'timing/model': 0.36722946699592285, 'learning_rate/qwen_vl_interface': 9.809691966399834e-06, 'learning_rate/action_model': 9.809691966399835e-05, 'epoch': 0.04}) +[2026-10-03 14:44:00,827][starVLA.training.train_starvla][INFO] - Step 544, Loss: {'train/loss': 0.2827123999595642, 'train/grad_norm_pre_clip': 64.59098052978516, 'timing/data': 0.00029633000667672604, 'timing/model': 0.3669844469986856, 'learning_rate/qwen_vl_interface': 9.808837634466104e-06, 'learning_rate/action_model': 9.808837634466104e-05, 'epoch': 0.04}) +[2026-10-03 14:44:01,478][starVLA.training.train_starvla][INFO] - Step 545, Loss: {'train/loss': 0.4938996136188507, 'train/grad_norm_pre_clip': 29.64715003967285, 'timing/data': 0.0002597400016384199, 'timing/model': 0.36674715699336957, 'learning_rate/qwen_vl_interface': 9.807981428566428e-06, 'learning_rate/action_model': 9.807981428566427e-05, 'epoch': 0.04}) +[2026-10-03 14:44:02,129][starVLA.training.train_starvla][INFO] - Step 546, Loss: {'train/loss': 0.02131902426481247, 'train/grad_norm_pre_clip': 11.829398155212402, 'timing/data': 0.00028843000472988933, 'timing/model': 0.3664209470007336, 'learning_rate/qwen_vl_interface': 9.807123349052754e-06, 'learning_rate/action_model': 9.807123349052753e-05, 'epoch': 0.04}) +[2026-10-03 14:44:02,781][starVLA.training.train_starvla][INFO] - Step 547, Loss: {'train/loss': 0.35137102007865906, 'train/grad_norm_pre_clip': 54.54116439819336, 'timing/data': 0.0002697300078580156, 'timing/model': 0.36739175699767657, 'learning_rate/qwen_vl_interface': 9.80626339627781e-06, 'learning_rate/action_model': 9.806263396277809e-05, 'epoch': 0.04}) +[2026-10-03 14:44:03,632][starVLA.training.train_starvla][INFO] - Step 548, Loss: {'train/loss': 0.14404316246509552, 'train/grad_norm_pre_clip': 41.34051513671875, 'timing/data': 0.00019947999680880457, 'timing/model': 0.3673498669959372, 'learning_rate/qwen_vl_interface': 9.805401570595087e-06, 'learning_rate/action_model': 9.805401570595087e-05, 'epoch': 0.04}) +[2026-10-03 14:44:04,260][starVLA.training.train_starvla][INFO] - Step 549, Loss: {'train/loss': 0.006331909913569689, 'train/grad_norm_pre_clip': 0.9786121249198914, 'timing/data': 0.00019798999710474163, 'timing/model': 0.343787899997551, 'learning_rate/qwen_vl_interface': 9.804537872358853e-06, 'learning_rate/action_model': 9.804537872358853e-05, 'epoch': 0.04}) +[2026-10-03 14:44:04,912][starVLA.training.train_starvla][INFO] - Step 550, Loss: {'train/loss': 0.04453588277101517, 'train/grad_norm_pre_clip': 57.61848068237305, 'timing/data': 0.0002011399919865653, 'timing/model': 0.36677927800337784, 'learning_rate/qwen_vl_interface': 9.80367230192414e-06, 'learning_rate/action_model': 9.803672301924138e-05, 'epoch': 0.04}) +[2026-10-03 14:44:05,565][starVLA.training.train_starvla][INFO] - Step 551, Loss: {'train/loss': 0.30332911014556885, 'train/grad_norm_pre_clip': 61.46824264526367, 'timing/data': 0.00019171999883837998, 'timing/model': 0.3682416570081841, 'learning_rate/qwen_vl_interface': 9.80280485964675e-06, 'learning_rate/action_model': 9.80280485964675e-05, 'epoch': 0.04}) +[2026-10-03 14:44:06,218][starVLA.training.train_starvla][INFO] - Step 552, Loss: {'train/loss': 0.12746620178222656, 'train/grad_norm_pre_clip': 79.56966400146484, 'timing/data': 0.00033958999847527593, 'timing/model': 0.36669827699370217, 'learning_rate/qwen_vl_interface': 9.801935545883255e-06, 'learning_rate/action_model': 9.801935545883255e-05, 'epoch': 0.04}) +[2026-10-03 14:44:07,074][starVLA.training.train_starvla][INFO] - Step 553, Loss: {'train/loss': 0.5473675727844238, 'train/grad_norm_pre_clip': 54.71067810058594, 'timing/data': 0.00031555999885313213, 'timing/model': 0.5716175840061624, 'learning_rate/qwen_vl_interface': 9.801064360991e-06, 'learning_rate/action_model': 9.801064360991e-05, 'epoch': 0.04}) +[2026-10-03 14:44:07,725][starVLA.training.train_starvla][INFO] - Step 554, Loss: {'train/loss': 0.2645895779132843, 'train/grad_norm_pre_clip': 31.45313262939453, 'timing/data': 0.00023494000197388232, 'timing/model': 0.3664278869982809, 'learning_rate/qwen_vl_interface': 9.800191305328095e-06, 'learning_rate/action_model': 9.800191305328094e-05, 'epoch': 0.04}) +[2026-10-03 14:44:08,354][starVLA.training.train_starvla][INFO] - Step 555, Loss: {'train/loss': 0.14345236122608185, 'train/grad_norm_pre_clip': 46.49858093261719, 'timing/data': 0.00019558999338187277, 'timing/model': 0.34396193000429776, 'learning_rate/qwen_vl_interface': 9.79931637925342e-06, 'learning_rate/action_model': 9.79931637925342e-05, 'epoch': 0.04}) +[2026-10-03 14:44:09,006][starVLA.training.train_starvla][INFO] - Step 556, Loss: {'train/loss': 0.24323952198028564, 'train/grad_norm_pre_clip': 76.68973541259766, 'timing/data': 0.0003258799988543615, 'timing/model': 0.3669438069919124, 'learning_rate/qwen_vl_interface': 9.798439583126625e-06, 'learning_rate/action_model': 9.798439583126625e-05, 'epoch': 0.04}) +[2026-10-03 14:44:09,658][starVLA.training.train_starvla][INFO] - Step 557, Loss: {'train/loss': 0.22856438159942627, 'train/grad_norm_pre_clip': 78.77752685546875, 'timing/data': 0.0002521999995224178, 'timing/model': 0.3665061270003207, 'learning_rate/qwen_vl_interface': 9.797560917308126e-06, 'learning_rate/action_model': 9.797560917308126e-05, 'epoch': 0.04}) +[2026-10-03 14:44:10,512][starVLA.training.train_starvla][INFO] - Step 558, Loss: {'train/loss': 0.7168130278587341, 'train/grad_norm_pre_clip': 29.575063705444336, 'timing/data': 0.0002602099993964657, 'timing/model': 0.569944452989148, 'learning_rate/qwen_vl_interface': 9.796680382159112e-06, 'learning_rate/action_model': 9.79668038215911e-05, 'epoch': 0.04}) +[2026-10-03 14:44:11,164][starVLA.training.train_starvla][INFO] - Step 559, Loss: {'train/loss': 0.07919004559516907, 'train/grad_norm_pre_clip': 14.225601196289062, 'timing/data': 0.00023672000679653138, 'timing/model': 0.36695910700655077, 'learning_rate/qwen_vl_interface': 9.795797978041533e-06, 'learning_rate/action_model': 9.795797978041533e-05, 'epoch': 0.04}) +[2026-10-03 14:44:11,817][starVLA.training.train_starvla][INFO] - Step 560, Loss: {'train/loss': 0.8468048572540283, 'train/grad_norm_pre_clip': 78.16162872314453, 'timing/data': 0.00026115000946447253, 'timing/model': 0.3672678370057838, 'learning_rate/qwen_vl_interface': 9.794913705318117e-06, 'learning_rate/action_model': 9.794913705318118e-05, 'epoch': 0.04}) +[2026-10-03 14:44:12,468][starVLA.training.train_starvla][INFO] - Step 561, Loss: {'train/loss': 1.0332750082015991, 'train/grad_norm_pre_clip': 39.73060989379883, 'timing/data': 0.0002859999949578196, 'timing/model': 0.366281746988534, 'learning_rate/qwen_vl_interface': 9.794027564352352e-06, 'learning_rate/action_model': 9.794027564352352e-05, 'epoch': 0.04}) +[2026-10-03 14:44:13,119][starVLA.training.train_starvla][INFO] - Step 562, Loss: {'train/loss': 0.0692015215754509, 'train/grad_norm_pre_clip': 25.644084930419922, 'timing/data': 0.00019754000823013484, 'timing/model': 0.3668651069892803, 'learning_rate/qwen_vl_interface': 9.793139555508501e-06, 'learning_rate/action_model': 9.7931395555085e-05, 'epoch': 0.04}) +[2026-10-03 14:44:13,725][starVLA.training.train_starvla][INFO] - Step 563, Loss: {'train/loss': 0.13210363686084747, 'train/grad_norm_pre_clip': 29.162145614624023, 'timing/data': 0.00027287998818792403, 'timing/model': 0.34381166999810375, 'learning_rate/qwen_vl_interface': 9.792249679151585e-06, 'learning_rate/action_model': 9.792249679151585e-05, 'epoch': 0.04}) +[2026-10-03 14:44:14,574][starVLA.training.train_starvla][INFO] - Step 564, Loss: {'train/loss': 0.19890320301055908, 'train/grad_norm_pre_clip': 88.73171997070312, 'timing/data': 0.00029665000329259783, 'timing/model': 0.3669917670049472, 'learning_rate/qwen_vl_interface': 9.791357935647403e-06, 'learning_rate/action_model': 9.791357935647403e-05, 'epoch': 0.04}) +[2026-10-03 14:44:15,224][starVLA.training.train_starvla][INFO] - Step 565, Loss: {'train/loss': 0.007482157088816166, 'train/grad_norm_pre_clip': 1.5298874378204346, 'timing/data': 0.00030564000189770013, 'timing/model': 0.3659223569993628, 'learning_rate/qwen_vl_interface': 9.790464325362516e-06, 'learning_rate/action_model': 9.790464325362516e-05, 'epoch': 0.04}) +[2026-10-03 14:44:15,853][starVLA.training.train_starvla][INFO] - Step 566, Loss: {'train/loss': 0.04726084694266319, 'train/grad_norm_pre_clip': 14.500646591186523, 'timing/data': 0.0002993300004163757, 'timing/model': 0.3662940280046314, 'learning_rate/qwen_vl_interface': 9.789568848664254e-06, 'learning_rate/action_model': 9.789568848664253e-05, 'epoch': 0.04}) +[2026-10-03 14:44:16,504][starVLA.training.train_starvla][INFO] - Step 567, Loss: {'train/loss': 0.04723731055855751, 'train/grad_norm_pre_clip': 12.040533065795898, 'timing/data': 0.0003851100045721978, 'timing/model': 0.36631146700528916, 'learning_rate/qwen_vl_interface': 9.788671505920712e-06, 'learning_rate/action_model': 9.788671505920711e-05, 'epoch': 0.04}) +[2026-10-03 14:44:17,157][starVLA.training.train_starvla][INFO] - Step 568, Loss: {'train/loss': 0.7191352248191833, 'train/grad_norm_pre_clip': 107.45071411132812, 'timing/data': 0.0002937499957624823, 'timing/model': 0.36750659700192045, 'learning_rate/qwen_vl_interface': 9.787772297500754e-06, 'learning_rate/action_model': 9.787772297500754e-05, 'epoch': 0.04}) +[2026-10-03 14:44:17,998][starVLA.training.train_starvla][INFO] - Step 569, Loss: {'train/loss': 0.12297327816486359, 'train/grad_norm_pre_clip': 47.28368377685547, 'timing/data': 0.0001952500024344772, 'timing/model': 0.5558904960053042, 'learning_rate/qwen_vl_interface': 9.78687122377401e-06, 'learning_rate/action_model': 9.78687122377401e-05, 'epoch': 0.04}) +[2026-10-03 14:44:18,649][starVLA.training.train_starvla][INFO] - Step 570, Loss: {'train/loss': 0.493460476398468, 'train/grad_norm_pre_clip': 28.859718322753906, 'timing/data': 0.00019905000226572156, 'timing/model': 0.3667309869924793, 'learning_rate/qwen_vl_interface': 9.78596828511088e-06, 'learning_rate/action_model': 9.78596828511088e-05, 'epoch': 0.04}) +[2026-10-03 14:44:19,301][starVLA.training.train_starvla][INFO] - Step 571, Loss: {'train/loss': 0.34252679347991943, 'train/grad_norm_pre_clip': 61.97998809814453, 'timing/data': 0.00020323999342508614, 'timing/model': 0.3673843979922822, 'learning_rate/qwen_vl_interface': 9.785063481882527e-06, 'learning_rate/action_model': 9.785063481882527e-05, 'epoch': 0.04}) +[2026-10-03 14:44:19,953][starVLA.training.train_starvla][INFO] - Step 572, Loss: {'train/loss': 0.31296321749687195, 'train/grad_norm_pre_clip': 53.63330078125, 'timing/data': 0.0002664299972821027, 'timing/model': 0.3670234269957291, 'learning_rate/qwen_vl_interface': 9.78415681446088e-06, 'learning_rate/action_model': 9.784156814460879e-05, 'epoch': 0.04}) +[2026-10-03 14:44:20,605][starVLA.training.train_starvla][INFO] - Step 573, Loss: {'train/loss': 0.028872769325971603, 'train/grad_norm_pre_clip': 5.214104652404785, 'timing/data': 0.0002572800003690645, 'timing/model': 0.3667576570005622, 'learning_rate/qwen_vl_interface': 9.783248283218638e-06, 'learning_rate/action_model': 9.783248283218638e-05, 'epoch': 0.04}) +[2026-10-03 14:44:21,425][starVLA.training.train_starvla][INFO] - Step 574, Loss: {'train/loss': 0.28849461674690247, 'train/grad_norm_pre_clip': 63.0698356628418, 'timing/data': 0.00019813000108115375, 'timing/model': 0.533618226996623, 'learning_rate/qwen_vl_interface': 9.782337888529261e-06, 'learning_rate/action_model': 9.782337888529261e-05, 'epoch': 0.04}) +[2026-10-03 14:44:22,077][starVLA.training.train_starvla][INFO] - Step 575, Loss: {'train/loss': 0.12539741396903992, 'train/grad_norm_pre_clip': 92.29753875732422, 'timing/data': 0.00029690000519622117, 'timing/model': 0.36707551700237673, 'learning_rate/qwen_vl_interface': 9.781425630766978e-06, 'learning_rate/action_model': 9.781425630766979e-05, 'epoch': 0.04}) +[2026-10-03 14:44:22,730][starVLA.training.train_starvla][INFO] - Step 576, Loss: {'train/loss': 0.055019259452819824, 'train/grad_norm_pre_clip': 35.96095275878906, 'timing/data': 0.00030672000139020383, 'timing/model': 0.3668156670028111, 'learning_rate/qwen_vl_interface': 9.780511510306789e-06, 'learning_rate/action_model': 9.780511510306788e-05, 'epoch': 0.04}) +[2026-10-03 14:44:23,381][starVLA.training.train_starvla][INFO] - Step 577, Loss: {'train/loss': 0.11820250749588013, 'train/grad_norm_pre_clip': 72.10320281982422, 'timing/data': 0.00031957999453879893, 'timing/model': 0.3664034879911924, 'learning_rate/qwen_vl_interface': 9.779595527524447e-06, 'learning_rate/action_model': 9.779595527524447e-05, 'epoch': 0.04}) +[2026-10-03 14:44:24,032][starVLA.training.train_starvla][INFO] - Step 578, Loss: {'train/loss': 0.008948097936809063, 'train/grad_norm_pre_clip': 8.87746810913086, 'timing/data': 0.00033703999361023307, 'timing/model': 0.3659646969899768, 'learning_rate/qwen_vl_interface': 9.778677682796483e-06, 'learning_rate/action_model': 9.778677682796483e-05, 'epoch': 0.04}) +[2026-10-03 14:44:24,880][starVLA.training.train_starvla][INFO] - Step 579, Loss: {'train/loss': 0.2584588825702667, 'train/grad_norm_pre_clip': 83.08724212646484, 'timing/data': 0.00024901999859139323, 'timing/model': 0.564672813998186, 'learning_rate/qwen_vl_interface': 9.777757976500185e-06, 'learning_rate/action_model': 9.777757976500186e-05, 'epoch': 0.04}) +[2026-10-03 14:44:25,508][starVLA.training.train_starvla][INFO] - Step 580, Loss: {'train/loss': 0.6064598560333252, 'train/grad_norm_pre_clip': 176.50914001464844, 'timing/data': 0.00036261999048292637, 'timing/model': 0.3663680580066284, 'learning_rate/qwen_vl_interface': 9.776836409013614e-06, 'learning_rate/action_model': 9.776836409013615e-05, 'epoch': 0.04}) +[2026-10-03 14:44:26,159][starVLA.training.train_starvla][INFO] - Step 581, Loss: {'train/loss': 0.5271447896957397, 'train/grad_norm_pre_clip': 97.52049255371094, 'timing/data': 0.00032005000684875995, 'timing/model': 0.3661674269969808, 'learning_rate/qwen_vl_interface': 9.775912980715589e-06, 'learning_rate/action_model': 9.775912980715588e-05, 'epoch': 0.04}) +[2026-10-03 14:44:26,809][starVLA.training.train_starvla][INFO] - Step 582, Loss: {'train/loss': 0.015842493623495102, 'train/grad_norm_pre_clip': 7.169566631317139, 'timing/data': 0.00025872999685816467, 'timing/model': 0.3658648370037554, 'learning_rate/qwen_vl_interface': 9.774987691985694e-06, 'learning_rate/action_model': 9.774987691985694e-05, 'epoch': 0.04}) +[2026-10-03 14:44:27,460][starVLA.training.train_starvla][INFO] - Step 583, Loss: {'train/loss': 0.032423235476017, 'train/grad_norm_pre_clip': 36.25408172607422, 'timing/data': 0.00024497001140844077, 'timing/model': 0.3668494169978658, 'learning_rate/qwen_vl_interface': 9.774060543204283e-06, 'learning_rate/action_model': 9.774060543204283e-05, 'epoch': 0.04}) +[2026-10-03 14:44:28,111][starVLA.training.train_starvla][INFO] - Step 584, Loss: {'train/loss': 0.001335964072495699, 'train/grad_norm_pre_clip': 0.3031199872493744, 'timing/data': 0.0002954399969894439, 'timing/model': 0.36629280699708033, 'learning_rate/qwen_vl_interface': 9.773131534752472e-06, 'learning_rate/action_model': 9.773131534752472e-05, 'epoch': 0.04}) +[2026-10-03 14:44:28,968][starVLA.training.train_starvla][INFO] - Step 585, Loss: {'train/loss': 0.015890834853053093, 'train/grad_norm_pre_clip': 8.909357070922852, 'timing/data': 0.00031398001010529697, 'timing/model': 0.3662280179996742, 'learning_rate/qwen_vl_interface': 9.772200667012142e-06, 'learning_rate/action_model': 9.772200667012142e-05, 'epoch': 0.04}) +[2026-10-03 14:44:29,619][starVLA.training.train_starvla][INFO] - Step 586, Loss: {'train/loss': 0.13284936547279358, 'train/grad_norm_pre_clip': 78.3419189453125, 'timing/data': 0.000309049995848909, 'timing/model': 0.3667724170081783, 'learning_rate/qwen_vl_interface': 9.771267940365934e-06, 'learning_rate/action_model': 9.771267940365935e-05, 'epoch': 0.04}) +[2026-10-03 14:44:30,271][starVLA.training.train_starvla][INFO] - Step 587, Loss: {'train/loss': 0.4739576280117035, 'train/grad_norm_pre_clip': 28.303865432739258, 'timing/data': 0.00024195999139919877, 'timing/model': 0.36687809698923957, 'learning_rate/qwen_vl_interface': 9.770333355197262e-06, 'learning_rate/action_model': 9.77033335519726e-05, 'epoch': 0.04}) +[2026-10-03 14:44:30,877][starVLA.training.train_starvla][INFO] - Step 588, Loss: {'train/loss': 0.5835956931114197, 'train/grad_norm_pre_clip': 193.479736328125, 'timing/data': 0.00024079000286292285, 'timing/model': 0.34390975900168996, 'learning_rate/qwen_vl_interface': 9.769396911890291e-06, 'learning_rate/action_model': 9.769396911890291e-05, 'epoch': 0.04}) +[2026-10-03 14:44:31,528][starVLA.training.train_starvla][INFO] - Step 589, Loss: {'train/loss': 0.007858479395508766, 'train/grad_norm_pre_clip': 2.619387626647949, 'timing/data': 0.00020031999156344682, 'timing/model': 0.36612546700052917, 'learning_rate/qwen_vl_interface': 9.768458610829965e-06, 'learning_rate/action_model': 9.768458610829965e-05, 'epoch': 0.04}) +[2026-10-03 14:44:32,373][starVLA.training.train_starvla][INFO] - Step 590, Loss: {'train/loss': 0.5475621223449707, 'train/grad_norm_pre_clip': 116.11891174316406, 'timing/data': 0.00023553000937681645, 'timing/model': 0.560104024989414, 'learning_rate/qwen_vl_interface': 9.767518452401981e-06, 'learning_rate/action_model': 9.76751845240198e-05, 'epoch': 0.04}) +[2026-10-03 14:44:33,024][starVLA.training.train_starvla][INFO] - Step 591, Loss: {'train/loss': 0.312713623046875, 'train/grad_norm_pre_clip': 90.45653533935547, 'timing/data': 0.00024573999689891934, 'timing/model': 0.3665209169994341, 'learning_rate/qwen_vl_interface': 9.766576436992802e-06, 'learning_rate/action_model': 9.766576436992801e-05, 'epoch': 0.04}) +[2026-10-03 14:44:33,677][starVLA.training.train_starvla][INFO] - Step 592, Loss: {'train/loss': 0.3824658989906311, 'train/grad_norm_pre_clip': 14.450448036193848, 'timing/data': 0.0002944699954241514, 'timing/model': 0.36683043600351084, 'learning_rate/qwen_vl_interface': 9.765632564989657e-06, 'learning_rate/action_model': 9.765632564989656e-05, 'epoch': 0.04}) +[2026-10-03 14:44:34,330][starVLA.training.train_starvla][INFO] - Step 593, Loss: {'train/loss': 0.16148358583450317, 'train/grad_norm_pre_clip': 132.8454132080078, 'timing/data': 0.00033121999877039343, 'timing/model': 0.3652039869921282, 'learning_rate/qwen_vl_interface': 9.764686836780533e-06, 'learning_rate/action_model': 9.764686836780532e-05, 'epoch': 0.04}) +[2026-10-03 14:44:34,958][starVLA.training.train_starvla][INFO] - Step 594, Loss: {'train/loss': 0.4507833421230316, 'train/grad_norm_pre_clip': 20.757959365844727, 'timing/data': 0.0002482100098859519, 'timing/model': 0.3658747669978766, 'learning_rate/qwen_vl_interface': 9.763739252754187e-06, 'learning_rate/action_model': 9.763739252754186e-05, 'epoch': 0.04}) +[2026-10-03 14:44:35,610][starVLA.training.train_starvla][INFO] - Step 595, Loss: {'train/loss': 0.7591402530670166, 'train/grad_norm_pre_clip': 78.685791015625, 'timing/data': 0.0002639500016812235, 'timing/model': 0.36802853700646665, 'learning_rate/qwen_vl_interface': 9.76278981330013e-06, 'learning_rate/action_model': 9.762789813300131e-05, 'epoch': 0.04}) +[2026-10-03 14:44:36,441][starVLA.training.train_starvla][INFO] - Step 596, Loss: {'train/loss': 0.2319691777229309, 'train/grad_norm_pre_clip': 74.74848175048828, 'timing/data': 0.00029973000346217304, 'timing/model': 0.36636274700867943, 'learning_rate/qwen_vl_interface': 9.761838518808647e-06, 'learning_rate/action_model': 9.761838518808646e-05, 'epoch': 0.04}) +[2026-10-03 14:44:37,091][starVLA.training.train_starvla][INFO] - Step 597, Loss: {'train/loss': 0.015348829329013824, 'train/grad_norm_pre_clip': 3.910775661468506, 'timing/data': 0.0004137700016144663, 'timing/model': 0.36573142700945027, 'learning_rate/qwen_vl_interface': 9.760885369670776e-06, 'learning_rate/action_model': 9.760885369670776e-05, 'epoch': 0.04}) +[2026-10-03 14:44:37,742][starVLA.training.train_starvla][INFO] - Step 598, Loss: {'train/loss': 0.03387404978275299, 'train/grad_norm_pre_clip': 13.196184158325195, 'timing/data': 0.00025676999939605594, 'timing/model': 0.3663784570089774, 'learning_rate/qwen_vl_interface': 9.759930366278319e-06, 'learning_rate/action_model': 9.759930366278318e-05, 'epoch': 0.04}) +[2026-10-03 14:44:38,370][starVLA.training.train_starvla][INFO] - Step 599, Loss: {'train/loss': 0.6847507953643799, 'train/grad_norm_pre_clip': 114.03971862792969, 'timing/data': 0.000387890002457425, 'timing/model': 0.36606695799855515, 'learning_rate/qwen_vl_interface': 9.758973509023842e-06, 'learning_rate/action_model': 9.758973509023843e-05, 'epoch': 0.04}) +[2026-10-03 14:44:38,999][starVLA.training.train_starvla][INFO] - Step 600, Loss: {'train/loss': 0.6273152232170105, 'train/grad_norm_pre_clip': 41.945430755615234, 'timing/data': 0.00024056999245658517, 'timing/model': 0.36653185699833557, 'learning_rate/qwen_vl_interface': 9.758014798300677e-06, 'learning_rate/action_model': 9.758014798300676e-05, 'epoch': 0.04}) +[2026-10-03 14:44:39,856][starVLA.training.train_starvla][INFO] - Step 601, Loss: {'train/loss': 0.02515421435236931, 'train/grad_norm_pre_clip': 4.027650356292725, 'timing/data': 0.0003137099993182346, 'timing/model': 0.3661245770053938, 'learning_rate/qwen_vl_interface': 9.757054234502909e-06, 'learning_rate/action_model': 9.757054234502909e-05, 'epoch': 0.04}) +[2026-10-03 14:44:40,506][starVLA.training.train_starvla][INFO] - Step 602, Loss: {'train/loss': 0.03265320137143135, 'train/grad_norm_pre_clip': 14.877497673034668, 'timing/data': 0.0003084299969486892, 'timing/model': 0.3661981870100135, 'learning_rate/qwen_vl_interface': 9.756091818025394e-06, 'learning_rate/action_model': 9.756091818025394e-05, 'epoch': 0.04}) +[2026-10-03 14:44:41,157][starVLA.training.train_starvla][INFO] - Step 603, Loss: {'train/loss': 0.18235382437705994, 'train/grad_norm_pre_clip': 25.01503562927246, 'timing/data': 0.00023709000379312783, 'timing/model': 0.36655211700417567, 'learning_rate/qwen_vl_interface': 9.75512754926374e-06, 'learning_rate/action_model': 9.755127549263741e-05, 'epoch': 0.04}) +[2026-10-03 14:44:41,808][starVLA.training.train_starvla][INFO] - Step 604, Loss: {'train/loss': 0.014422634616494179, 'train/grad_norm_pre_clip': 2.1983156204223633, 'timing/data': 0.0002003900008276105, 'timing/model': 0.3666368470003363, 'learning_rate/qwen_vl_interface': 9.754161428614325e-06, 'learning_rate/action_model': 9.754161428614324e-05, 'epoch': 0.04}) +[2026-10-03 14:44:42,458][starVLA.training.train_starvla][INFO] - Step 605, Loss: {'train/loss': 0.03396235778927803, 'train/grad_norm_pre_clip': 77.95804595947266, 'timing/data': 0.00026281000464223325, 'timing/model': 0.3652980769984424, 'learning_rate/qwen_vl_interface': 9.753193456474284e-06, 'learning_rate/action_model': 9.753193456474284e-05, 'epoch': 0.04}) +[2026-10-03 14:44:43,307][starVLA.training.train_starvla][INFO] - Step 606, Loss: {'train/loss': 0.6896643042564392, 'train/grad_norm_pre_clip': 49.68811798095703, 'timing/data': 0.00031127998954616487, 'timing/model': 0.5654652240045834, 'learning_rate/qwen_vl_interface': 9.752223633241514e-06, 'learning_rate/action_model': 9.752223633241514e-05, 'epoch': 0.04}) +[2026-10-03 14:44:43,956][starVLA.training.train_starvla][INFO] - Step 607, Loss: {'train/loss': 0.9730320572853088, 'train/grad_norm_pre_clip': 41.19021987915039, 'timing/data': 0.0003239799989387393, 'timing/model': 0.3654363670066232, 'learning_rate/qwen_vl_interface': 9.751251959314671e-06, 'learning_rate/action_model': 9.751251959314672e-05, 'epoch': 0.04}) +[2026-10-03 14:44:44,605][starVLA.training.train_starvla][INFO] - Step 608, Loss: {'train/loss': 0.06409642100334167, 'train/grad_norm_pre_clip': 109.60023498535156, 'timing/data': 0.0002385300031164661, 'timing/model': 0.36556051699153613, 'learning_rate/qwen_vl_interface': 9.750278435093176e-06, 'learning_rate/action_model': 9.750278435093177e-05, 'epoch': 0.04}) +[2026-10-03 14:44:45,253][starVLA.training.train_starvla][INFO] - Step 609, Loss: {'train/loss': 0.029283929616212845, 'train/grad_norm_pre_clip': 11.713047981262207, 'timing/data': 0.0003892700042342767, 'timing/model': 0.364593567996053, 'learning_rate/qwen_vl_interface': 9.749303060977205e-06, 'learning_rate/action_model': 9.749303060977205e-05, 'epoch': 0.04}) +[2026-10-03 14:44:45,903][starVLA.training.train_starvla][INFO] - Step 610, Loss: {'train/loss': 0.09977616369724274, 'train/grad_norm_pre_clip': 106.40150451660156, 'timing/data': 0.000315440003760159, 'timing/model': 0.3665721479919739, 'learning_rate/qwen_vl_interface': 9.748325837367701e-06, 'learning_rate/action_model': 9.748325837367702e-05, 'epoch': 0.04}) +[2026-10-03 14:44:46,742][starVLA.training.train_starvla][INFO] - Step 611, Loss: {'train/loss': 0.002866648370400071, 'train/grad_norm_pre_clip': 3.4344780445098877, 'timing/data': 0.00033853000786621124, 'timing/model': 0.5776174920029007, 'learning_rate/qwen_vl_interface': 9.747346764666362e-06, 'learning_rate/action_model': 9.747346764666361e-05, 'epoch': 0.04}) +[2026-10-03 14:44:47,391][starVLA.training.train_starvla][INFO] - Step 612, Loss: {'train/loss': 0.03167898207902908, 'train/grad_norm_pre_clip': 20.551876068115234, 'timing/data': 0.00033850000181701034, 'timing/model': 0.36543156699917745, 'learning_rate/qwen_vl_interface': 9.746365843275648e-06, 'learning_rate/action_model': 9.746365843275646e-05, 'epoch': 0.04}) +[2026-10-03 14:44:48,039][starVLA.training.train_starvla][INFO] - Step 613, Loss: {'train/loss': 0.26335328817367554, 'train/grad_norm_pre_clip': 120.65387725830078, 'timing/data': 0.000248780008405447, 'timing/model': 0.3646060570026748, 'learning_rate/qwen_vl_interface': 9.745383073598776e-06, 'learning_rate/action_model': 9.745383073598775e-05, 'epoch': 0.04}) +[2026-10-03 14:44:48,664][starVLA.training.train_starvla][INFO] - Step 614, Loss: {'train/loss': 0.004296516999602318, 'train/grad_norm_pre_clip': 2.061192750930786, 'timing/data': 0.00030813000921625644, 'timing/model': 0.36471377800626215, 'learning_rate/qwen_vl_interface': 9.744398456039729e-06, 'learning_rate/action_model': 9.744398456039729e-05, 'epoch': 0.04}) +[2026-10-03 14:44:49,313][starVLA.training.train_starvla][INFO] - Step 615, Loss: {'train/loss': 0.31334877014160156, 'train/grad_norm_pre_clip': 32.198204040527344, 'timing/data': 0.00037552999856416136, 'timing/model': 0.3653920270007802, 'learning_rate/qwen_vl_interface': 9.743411991003246e-06, 'learning_rate/action_model': 9.743411991003247e-05, 'epoch': 0.05}) +[2026-10-03 14:44:49,962][starVLA.training.train_starvla][INFO] - Step 616, Loss: {'train/loss': 0.6397958993911743, 'train/grad_norm_pre_clip': 152.69830322265625, 'timing/data': 0.00031336999381892383, 'timing/model': 0.36577299700002186, 'learning_rate/qwen_vl_interface': 9.742423678894823e-06, 'learning_rate/action_model': 9.742423678894823e-05, 'epoch': 0.05}) +[2026-10-03 14:44:50,800][starVLA.training.train_starvla][INFO] - Step 617, Loss: {'train/loss': 0.05977581441402435, 'train/grad_norm_pre_clip': 93.1495590209961, 'timing/data': 0.0002582400047685951, 'timing/model': 0.5548085149930557, 'learning_rate/qwen_vl_interface': 9.741433520120722e-06, 'learning_rate/action_model': 9.74143352012072e-05, 'epoch': 0.05}) +[2026-10-03 14:44:51,449][starVLA.training.train_starvla][INFO] - Step 618, Loss: {'train/loss': 0.36969152092933655, 'train/grad_norm_pre_clip': 17.648427963256836, 'timing/data': 0.0002406900021014735, 'timing/model': 0.36508049700933043, 'learning_rate/qwen_vl_interface': 9.740441515087953e-06, 'learning_rate/action_model': 9.740441515087953e-05, 'epoch': 0.05}) +[2026-10-03 14:44:52,096][starVLA.training.train_starvla][INFO] - Step 619, Loss: {'train/loss': 0.0716795101761818, 'train/grad_norm_pre_clip': 104.31196594238281, 'timing/data': 0.0002549899945734069, 'timing/model': 0.3648002380068647, 'learning_rate/qwen_vl_interface': 9.739447664204298e-06, 'learning_rate/action_model': 9.739447664204297e-05, 'epoch': 0.05}) +[2026-10-03 14:44:52,744][starVLA.training.train_starvla][INFO] - Step 620, Loss: {'train/loss': 0.7129137516021729, 'train/grad_norm_pre_clip': 80.2188491821289, 'timing/data': 0.0001980099914362654, 'timing/model': 0.36499514800379984, 'learning_rate/qwen_vl_interface': 9.73845196787829e-06, 'learning_rate/action_model': 9.738451967878289e-05, 'epoch': 0.05}) +[2026-10-03 14:44:53,392][starVLA.training.train_starvla][INFO] - Step 621, Loss: {'train/loss': 0.004522849339991808, 'train/grad_norm_pre_clip': 1.0247198343276978, 'timing/data': 0.0003626100078690797, 'timing/model': 0.36464557799627073, 'learning_rate/qwen_vl_interface': 9.737454426519221e-06, 'learning_rate/action_model': 9.737454426519221e-05, 'epoch': 0.05}) +[2026-10-03 14:44:54,243][starVLA.training.train_starvla][INFO] - Step 622, Loss: {'train/loss': 0.18786053359508514, 'train/grad_norm_pre_clip': 88.58772277832031, 'timing/data': 0.0001963099930435419, 'timing/model': 0.56815246300539, 'learning_rate/qwen_vl_interface': 9.736455040537145e-06, 'learning_rate/action_model': 9.736455040537144e-05, 'epoch': 0.05}) +[2026-10-03 14:44:54,867][starVLA.training.train_starvla][INFO] - Step 623, Loss: {'train/loss': 0.242649644613266, 'train/grad_norm_pre_clip': 38.88898849487305, 'timing/data': 0.00019701999553944916, 'timing/model': 0.3646902280015638, 'learning_rate/qwen_vl_interface': 9.735453810342868e-06, 'learning_rate/action_model': 9.735453810342867e-05, 'epoch': 0.05}) +[2026-10-03 14:44:55,515][starVLA.training.train_starvla][INFO] - Step 624, Loss: {'train/loss': 0.12325181066989899, 'train/grad_norm_pre_clip': 23.152114868164062, 'timing/data': 0.00040412000089418143, 'timing/model': 0.3644320280000102, 'learning_rate/qwen_vl_interface': 9.734450736347958e-06, 'learning_rate/action_model': 9.73445073634796e-05, 'epoch': 0.05}) +[2026-10-03 14:44:56,161][starVLA.training.train_starvla][INFO] - Step 625, Loss: {'train/loss': 0.5521771311759949, 'train/grad_norm_pre_clip': 35.41226577758789, 'timing/data': 0.0002009699965128675, 'timing/model': 0.3642452669882914, 'learning_rate/qwen_vl_interface': 9.733445818964746e-06, 'learning_rate/action_model': 9.733445818964746e-05, 'epoch': 0.05}) +[2026-10-03 14:44:56,809][starVLA.training.train_starvla][INFO] - Step 626, Loss: {'train/loss': 0.2909957468509674, 'train/grad_norm_pre_clip': 75.69583129882812, 'timing/data': 0.000232020000112243, 'timing/model': 0.3645149579970166, 'learning_rate/qwen_vl_interface': 9.732439058606311e-06, 'learning_rate/action_model': 9.732439058606312e-05, 'epoch': 0.05}) +[2026-10-03 14:44:57,656][starVLA.training.train_starvla][INFO] - Step 627, Loss: {'train/loss': 0.3038013279438019, 'train/grad_norm_pre_clip': 133.9118194580078, 'timing/data': 0.0001984999980777502, 'timing/model': 0.5656930940021994, 'learning_rate/qwen_vl_interface': 9.731430455686499e-06, 'learning_rate/action_model': 9.731430455686498e-05, 'epoch': 0.05}) +[2026-10-03 14:44:58,303][starVLA.training.train_starvla][INFO] - Step 628, Loss: {'train/loss': 1.5739717483520508, 'train/grad_norm_pre_clip': 99.05208587646484, 'timing/data': 0.00019470999541226774, 'timing/model': 0.3644445679965429, 'learning_rate/qwen_vl_interface': 9.730420010619901e-06, 'learning_rate/action_model': 9.7304200106199e-05, 'epoch': 0.05}) +[2026-10-03 14:44:58,950][starVLA.training.train_starvla][INFO] - Step 629, Loss: {'train/loss': 0.2702637314796448, 'train/grad_norm_pre_clip': 121.2109146118164, 'timing/data': 0.00019929998961742967, 'timing/model': 0.3640970370033756, 'learning_rate/qwen_vl_interface': 9.72940772382188e-06, 'learning_rate/action_model': 9.729407723821878e-05, 'epoch': 0.05}) +[2026-10-03 14:44:59,598][starVLA.training.train_starvla][INFO] - Step 630, Loss: {'train/loss': 0.04974270239472389, 'train/grad_norm_pre_clip': 10.619278907775879, 'timing/data': 0.00019932999566663057, 'timing/model': 0.3650252169900341, 'learning_rate/qwen_vl_interface': 9.728393595708542e-06, 'learning_rate/action_model': 9.728393595708542e-05, 'epoch': 0.05}) +[2026-10-03 14:45:00,245][starVLA.training.train_starvla][INFO] - Step 631, Loss: {'train/loss': 0.7458975315093994, 'train/grad_norm_pre_clip': 178.20774841308594, 'timing/data': 0.000242260008235462, 'timing/model': 0.3646428280044347, 'learning_rate/qwen_vl_interface': 9.727377626696762e-06, 'learning_rate/action_model': 9.727377626696762e-05, 'epoch': 0.05}) +[2026-10-03 14:45:00,894][starVLA.training.train_starvla][INFO] - Step 632, Loss: {'train/loss': 0.024240897968411446, 'train/grad_norm_pre_clip': 7.83634614944458, 'timing/data': 0.00019271000928711146, 'timing/model': 0.36589754799206275, 'learning_rate/qwen_vl_interface': 9.726359817204165e-06, 'learning_rate/action_model': 9.726359817204166e-05, 'epoch': 0.05}) +[2026-10-03 14:45:01,738][starVLA.training.train_starvla][INFO] - Step 633, Loss: {'train/loss': 0.006883899215608835, 'train/grad_norm_pre_clip': 1.3684308528900146, 'timing/data': 0.0002005300048040226, 'timing/model': 0.36433048799517564, 'learning_rate/qwen_vl_interface': 9.725340167649136e-06, 'learning_rate/action_model': 9.725340167649136e-05, 'epoch': 0.05}) +[2026-10-03 14:45:02,385][starVLA.training.train_starvla][INFO] - Step 634, Loss: {'train/loss': 0.004739239811897278, 'train/grad_norm_pre_clip': 4.348404884338379, 'timing/data': 0.00019832000543829054, 'timing/model': 0.364684756990755, 'learning_rate/qwen_vl_interface': 9.72431867845081e-06, 'learning_rate/action_model': 9.72431867845081e-05, 'epoch': 0.05}) +[2026-10-03 14:45:03,032][starVLA.training.train_starvla][INFO] - Step 635, Loss: {'train/loss': 0.012907743453979492, 'train/grad_norm_pre_clip': 28.00319480895996, 'timing/data': 0.00019567999697756022, 'timing/model': 0.36489513800188433, 'learning_rate/qwen_vl_interface': 9.723295350029086e-06, 'learning_rate/action_model': 9.723295350029086e-05, 'epoch': 0.05}) +[2026-10-03 14:45:03,679][starVLA.training.train_starvla][INFO] - Step 636, Loss: {'train/loss': 0.6009472012519836, 'train/grad_norm_pre_clip': 24.471351623535156, 'timing/data': 0.00025443000777158886, 'timing/model': 0.3649070679966826, 'learning_rate/qwen_vl_interface': 9.722270182804616e-06, 'learning_rate/action_model': 9.722270182804614e-05, 'epoch': 0.05}) +[2026-10-03 14:45:04,326][starVLA.training.train_starvla][INFO] - Step 637, Loss: {'train/loss': 0.24723441898822784, 'train/grad_norm_pre_clip': 49.984249114990234, 'timing/data': 0.00025864000781439245, 'timing/model': 0.36450446699745953, 'learning_rate/qwen_vl_interface': 9.721243177198805e-06, 'learning_rate/action_model': 9.721243177198804e-05, 'epoch': 0.05}) +[2026-10-03 14:45:05,160][starVLA.training.train_starvla][INFO] - Step 638, Loss: {'train/loss': 0.6607577800750732, 'train/grad_norm_pre_clip': 66.68977355957031, 'timing/data': 0.00024323999241460115, 'timing/model': 0.5506616759957979, 'learning_rate/qwen_vl_interface': 9.720214333633817e-06, 'learning_rate/action_model': 9.720214333633817e-05, 'epoch': 0.05}) +[2026-10-03 14:45:05,785][starVLA.training.train_starvla][INFO] - Step 639, Loss: {'train/loss': 1.391505241394043, 'train/grad_norm_pre_clip': 108.95310974121094, 'timing/data': 0.0002947899920400232, 'timing/model': 0.3658089679956902, 'learning_rate/qwen_vl_interface': 9.719183652532572e-06, 'learning_rate/action_model': 9.719183652532572e-05, 'epoch': 0.05}) +[2026-10-03 14:45:06,436][starVLA.training.train_starvla][INFO] - Step 640, Loss: {'train/loss': 0.4791368246078491, 'train/grad_norm_pre_clip': 42.344600677490234, 'timing/data': 0.00021046999609097838, 'timing/model': 0.3664166969974758, 'learning_rate/qwen_vl_interface': 9.718151134318743e-06, 'learning_rate/action_model': 9.718151134318743e-05, 'epoch': 0.05}) +[2026-10-03 14:45:07,085][starVLA.training.train_starvla][INFO] - Step 641, Loss: {'train/loss': 0.00021952031238470227, 'train/grad_norm_pre_clip': 0.037457507103681564, 'timing/data': 0.00023954999051056802, 'timing/model': 0.3655935970018618, 'learning_rate/qwen_vl_interface': 9.71711677941676e-06, 'learning_rate/action_model': 9.717116779416758e-05, 'epoch': 0.05}) +[2026-10-03 14:45:07,735][starVLA.training.train_starvla][INFO] - Step 642, Loss: {'train/loss': 0.5693516731262207, 'train/grad_norm_pre_clip': 61.043670654296875, 'timing/data': 0.0002377700002398342, 'timing/model': 0.36612573800084647, 'learning_rate/qwen_vl_interface': 9.716080588251805e-06, 'learning_rate/action_model': 9.716080588251805e-05, 'epoch': 0.05}) +[2026-10-03 14:45:08,573][starVLA.training.train_starvla][INFO] - Step 643, Loss: {'train/loss': 0.7477976679801941, 'train/grad_norm_pre_clip': 190.38897705078125, 'timing/data': 0.00024437000683974475, 'timing/model': 0.5541094660002273, 'learning_rate/qwen_vl_interface': 9.71504256124982e-06, 'learning_rate/action_model': 9.71504256124982e-05, 'epoch': 0.05}) +[2026-10-03 14:45:09,223][starVLA.training.train_starvla][INFO] - Step 644, Loss: {'train/loss': 0.09218889474868774, 'train/grad_norm_pre_clip': 61.29332733154297, 'timing/data': 0.00024044999736361206, 'timing/model': 0.36643167700094637, 'learning_rate/qwen_vl_interface': 9.714002698837499e-06, 'learning_rate/action_model': 9.714002698837498e-05, 'epoch': 0.05}) +[2026-10-03 14:45:09,872][starVLA.training.train_starvla][INFO] - Step 645, Loss: {'train/loss': 0.3856973350048065, 'train/grad_norm_pre_clip': 69.46207427978516, 'timing/data': 0.00023465999402105808, 'timing/model': 0.3652866969932802, 'learning_rate/qwen_vl_interface': 9.712961001442285e-06, 'learning_rate/action_model': 9.712961001442286e-05, 'epoch': 0.05}) +[2026-10-03 14:45:10,522][starVLA.training.train_starvla][INFO] - Step 646, Loss: {'train/loss': 0.0020326459780335426, 'train/grad_norm_pre_clip': 0.7568768262863159, 'timing/data': 0.00024010900233406574, 'timing/model': 0.3659635570074897, 'learning_rate/qwen_vl_interface': 9.711917469492388e-06, 'learning_rate/action_model': 9.711917469492388e-05, 'epoch': 0.05}) +[2026-10-03 14:45:11,172][starVLA.training.train_starvla][INFO] - Step 647, Loss: {'train/loss': 0.031847912818193436, 'train/grad_norm_pre_clip': 7.645709037780762, 'timing/data': 0.00021731999004259706, 'timing/model': 0.3664553779963171, 'learning_rate/qwen_vl_interface': 9.71087210341676e-06, 'learning_rate/action_model': 9.71087210341676e-05, 'epoch': 0.05}) +[2026-10-03 14:45:12,017][starVLA.training.train_starvla][INFO] - Step 648, Loss: {'train/loss': 0.09950141608715057, 'train/grad_norm_pre_clip': 72.78984069824219, 'timing/data': 0.00024110999947879463, 'timing/model': 0.5604099049960496, 'learning_rate/qwen_vl_interface': 9.70982490364511e-06, 'learning_rate/action_model': 9.70982490364511e-05, 'epoch': 0.05}) +[2026-10-03 14:45:12,644][starVLA.training.train_starvla][INFO] - Step 649, Loss: {'train/loss': 0.1749352514743805, 'train/grad_norm_pre_clip': 83.00244903564453, 'timing/data': 0.00024071999359875917, 'timing/model': 0.36559989700617734, 'learning_rate/qwen_vl_interface': 9.708775870607905e-06, 'learning_rate/action_model': 9.708775870607904e-05, 'epoch': 0.05}) +[2026-10-03 14:45:13,293][starVLA.training.train_starvla][INFO] - Step 650, Loss: {'train/loss': 0.03412996232509613, 'train/grad_norm_pre_clip': 30.472705841064453, 'timing/data': 0.0002416200004518032, 'timing/model': 0.3659707569895545, 'learning_rate/qwen_vl_interface': 9.707725004736363e-06, 'learning_rate/action_model': 9.707725004736363e-05, 'epoch': 0.05}) +[2026-10-03 14:45:13,943][starVLA.training.train_starvla][INFO] - Step 651, Loss: {'train/loss': 0.5392553806304932, 'train/grad_norm_pre_clip': 58.7563362121582, 'timing/data': 0.00029945001006126404, 'timing/model': 0.36589594700490125, 'learning_rate/qwen_vl_interface': 9.706672306462454e-06, 'learning_rate/action_model': 9.706672306462454e-05, 'epoch': 0.05}) +[2026-10-03 14:45:14,594][starVLA.training.train_starvla][INFO] - Step 652, Loss: {'train/loss': 0.28193795680999756, 'train/grad_norm_pre_clip': 74.72106170654297, 'timing/data': 0.00024213000142481178, 'timing/model': 0.36624369700439274, 'learning_rate/qwen_vl_interface': 9.705617776218905e-06, 'learning_rate/action_model': 9.705617776218904e-05, 'epoch': 0.05}) +[2026-10-03 14:45:15,242][starVLA.training.train_starvla][INFO] - Step 653, Loss: {'train/loss': 0.5841569304466248, 'train/grad_norm_pre_clip': 220.9934539794922, 'timing/data': 0.0002706399973249063, 'timing/model': 0.3652787070022896, 'learning_rate/qwen_vl_interface': 9.70456141443919e-06, 'learning_rate/action_model': 9.704561414439188e-05, 'epoch': 0.05}) +[2026-10-03 14:45:16,058][starVLA.training.train_starvla][INFO] - Step 654, Loss: {'train/loss': 0.24612665176391602, 'train/grad_norm_pre_clip': 46.37116241455078, 'timing/data': 0.00024262999068014324, 'timing/model': 0.3430305700021563, 'learning_rate/qwen_vl_interface': 9.703503221557538e-06, 'learning_rate/action_model': 9.703503221557538e-05, 'epoch': 0.05}) +[2026-10-03 14:45:16,708][starVLA.training.train_starvla][INFO] - Step 655, Loss: {'train/loss': 0.22053393721580505, 'train/grad_norm_pre_clip': 58.87565612792969, 'timing/data': 0.00026020000223070383, 'timing/model': 0.3663953369978117, 'learning_rate/qwen_vl_interface': 9.70244319800894e-06, 'learning_rate/action_model': 9.70244319800894e-05, 'epoch': 0.05}) +[2026-10-03 14:45:17,359][starVLA.training.train_starvla][INFO] - Step 656, Loss: {'train/loss': 0.0246737040579319, 'train/grad_norm_pre_clip': 13.216497421264648, 'timing/data': 0.00023771999985910952, 'timing/model': 0.3665680470003281, 'learning_rate/qwen_vl_interface': 9.701381344229124e-06, 'learning_rate/action_model': 9.701381344229123e-05, 'epoch': 0.05}) +[2026-10-03 14:45:18,008][starVLA.training.train_starvla][INFO] - Step 657, Loss: {'train/loss': 0.010701840743422508, 'train/grad_norm_pre_clip': 2.5306899547576904, 'timing/data': 0.0003501799947116524, 'timing/model': 0.36527897699852474, 'learning_rate/qwen_vl_interface': 9.700317660654581e-06, 'learning_rate/action_model': 9.700317660654581e-05, 'epoch': 0.05}) +[2026-10-03 14:45:18,657][starVLA.training.train_starvla][INFO] - Step 658, Loss: {'train/loss': 0.19602204859256744, 'train/grad_norm_pre_clip': 37.79589080810547, 'timing/data': 0.0003313100023660809, 'timing/model': 0.365882447003969, 'learning_rate/qwen_vl_interface': 9.699252147722554e-06, 'learning_rate/action_model': 9.699252147722553e-05, 'epoch': 0.05}) +[2026-10-03 14:45:19,504][starVLA.training.train_starvla][INFO] - Step 659, Loss: {'train/loss': 0.442491352558136, 'train/grad_norm_pre_clip': 9.89272689819336, 'timing/data': 0.00031291000777855515, 'timing/model': 0.36661007700604387, 'learning_rate/qwen_vl_interface': 9.698184805871028e-06, 'learning_rate/action_model': 9.698184805871028e-05, 'epoch': 0.05}) +[2026-10-03 14:45:20,156][starVLA.training.train_starvla][INFO] - Step 660, Loss: {'train/loss': 0.1728244572877884, 'train/grad_norm_pre_clip': 335.06121826171875, 'timing/data': 0.00029457999335136265, 'timing/model': 0.36679767700843513, 'learning_rate/qwen_vl_interface': 9.697115635538753e-06, 'learning_rate/action_model': 9.697115635538753e-05, 'epoch': 0.05}) +[2026-10-03 14:45:20,806][starVLA.training.train_starvla][INFO] - Step 661, Loss: {'train/loss': 0.30421310663223267, 'train/grad_norm_pre_clip': 61.584144592285156, 'timing/data': 0.00026405999960843474, 'timing/model': 0.3665543070092099, 'learning_rate/qwen_vl_interface': 9.69604463716522e-06, 'learning_rate/action_model': 9.69604463716522e-05, 'epoch': 0.05}) +[2026-10-03 14:45:21,457][starVLA.training.train_starvla][INFO] - Step 662, Loss: {'train/loss': 0.3646191954612732, 'train/grad_norm_pre_clip': 30.71782875061035, 'timing/data': 0.00022965999960433692, 'timing/model': 0.3660631380043924, 'learning_rate/qwen_vl_interface': 9.69497181119068e-06, 'learning_rate/action_model': 9.694971811190681e-05, 'epoch': 0.05}) +[2026-10-03 14:45:22,107][starVLA.training.train_starvla][INFO] - Step 663, Loss: {'train/loss': 0.3027735948562622, 'train/grad_norm_pre_clip': 56.04571533203125, 'timing/data': 0.00024089000362437218, 'timing/model': 0.3661082970065763, 'learning_rate/qwen_vl_interface': 9.693897158056132e-06, 'learning_rate/action_model': 9.693897158056132e-05, 'epoch': 0.05}) +[2026-10-03 14:45:22,924][starVLA.training.train_starvla][INFO] - Step 664, Loss: {'train/loss': 0.41459763050079346, 'train/grad_norm_pre_clip': 58.05123519897461, 'timing/data': 0.0002469200117047876, 'timing/model': 0.36567087699950207, 'learning_rate/qwen_vl_interface': 9.692820678203322e-06, 'learning_rate/action_model': 9.692820678203322e-05, 'epoch': 0.05}) +[2026-10-03 14:45:23,573][starVLA.training.train_starvla][INFO] - Step 665, Loss: {'train/loss': 0.16255582869052887, 'train/grad_norm_pre_clip': 71.09452056884766, 'timing/data': 0.000321789993904531, 'timing/model': 0.36561352699936833, 'learning_rate/qwen_vl_interface': 9.691742372074749e-06, 'learning_rate/action_model': 9.691742372074749e-05, 'epoch': 0.05}) +[2026-10-03 14:45:24,200][starVLA.training.train_starvla][INFO] - Step 666, Loss: {'train/loss': 0.49624186754226685, 'train/grad_norm_pre_clip': 137.2202911376953, 'timing/data': 0.0004233799991197884, 'timing/model': 0.3424489899916807, 'learning_rate/qwen_vl_interface': 9.690662240113669e-06, 'learning_rate/action_model': 9.690662240113668e-05, 'epoch': 0.05}) +[2026-10-03 14:45:24,850][starVLA.training.train_starvla][INFO] - Step 667, Loss: {'train/loss': 0.2720229923725128, 'train/grad_norm_pre_clip': 52.43126678466797, 'timing/data': 0.0002586600021459162, 'timing/model': 0.3656859469920164, 'learning_rate/qwen_vl_interface': 9.689580282764077e-06, 'learning_rate/action_model': 9.689580282764077e-05, 'epoch': 0.05}) +[2026-10-03 14:45:25,500][starVLA.training.train_starvla][INFO] - Step 668, Loss: {'train/loss': 0.4084685742855072, 'train/grad_norm_pre_clip': 32.74266815185547, 'timing/data': 0.00025313999503850937, 'timing/model': 0.3667186969978502, 'learning_rate/qwen_vl_interface': 9.688496500470731e-06, 'learning_rate/action_model': 9.68849650047073e-05, 'epoch': 0.05}) +[2026-10-03 14:45:26,150][starVLA.training.train_starvla][INFO] - Step 669, Loss: {'train/loss': 0.5252193212509155, 'train/grad_norm_pre_clip': 42.86150360107422, 'timing/data': 0.0003847600019071251, 'timing/model': 0.3653271869989112, 'learning_rate/qwen_vl_interface': 9.68741089367913e-06, 'learning_rate/action_model': 9.687410893679129e-05, 'epoch': 0.05}) +[2026-10-03 14:45:26,992][starVLA.training.train_starvla][INFO] - Step 670, Loss: {'train/loss': 0.060775771737098694, 'train/grad_norm_pre_clip': 9.564766883850098, 'timing/data': 0.0003133300051558763, 'timing/model': 0.36527491599554196, 'learning_rate/qwen_vl_interface': 9.686323462835527e-06, 'learning_rate/action_model': 9.686323462835526e-05, 'epoch': 0.05}) +[2026-10-03 14:45:27,641][starVLA.training.train_starvla][INFO] - Step 671, Loss: {'train/loss': 0.2043417990207672, 'train/grad_norm_pre_clip': 21.158641815185547, 'timing/data': 0.0003131400007987395, 'timing/model': 0.36612127600528765, 'learning_rate/qwen_vl_interface': 9.685234208386922e-06, 'learning_rate/action_model': 9.685234208386921e-05, 'epoch': 0.05}) +[2026-10-03 14:45:28,291][starVLA.training.train_starvla][INFO] - Step 672, Loss: {'train/loss': 0.3716829717159271, 'train/grad_norm_pre_clip': 31.21204376220703, 'timing/data': 0.0002825299889082089, 'timing/model': 0.3663206969940802, 'learning_rate/qwen_vl_interface': 9.684143130781066e-06, 'learning_rate/action_model': 9.684143130781066e-05, 'epoch': 0.05}) +[2026-10-03 14:45:28,919][starVLA.training.train_starvla][INFO] - Step 673, Loss: {'train/loss': 0.044640012085437775, 'train/grad_norm_pre_clip': 13.477466583251953, 'timing/data': 0.00042016999213956296, 'timing/model': 0.36547854699892923, 'learning_rate/qwen_vl_interface': 9.683050230466462e-06, 'learning_rate/action_model': 9.683050230466462e-05, 'epoch': 0.05}) +[2026-10-03 14:45:29,568][starVLA.training.train_starvla][INFO] - Step 674, Loss: {'train/loss': 0.5357527136802673, 'train/grad_norm_pre_clip': 19.00342559814453, 'timing/data': 0.0002366800035815686, 'timing/model': 0.36606077699980233, 'learning_rate/qwen_vl_interface': 9.681955507892361e-06, 'learning_rate/action_model': 9.68195550789236e-05, 'epoch': 0.05}) +[2026-10-03 14:45:30,410][starVLA.training.train_starvla][INFO] - Step 675, Loss: {'train/loss': 0.5866393446922302, 'train/grad_norm_pre_clip': 44.39293670654297, 'timing/data': 0.00020387998665682971, 'timing/model': 0.36460646700288635, 'learning_rate/qwen_vl_interface': 9.680858963508759e-06, 'learning_rate/action_model': 9.68085896350876e-05, 'epoch': 0.05}) +[2026-10-03 14:45:31,057][starVLA.training.train_starvla][INFO] - Step 676, Loss: {'train/loss': 0.018931984901428223, 'train/grad_norm_pre_clip': 1.8271687030792236, 'timing/data': 0.00031070999102666974, 'timing/model': 0.36483404799946584, 'learning_rate/qwen_vl_interface': 9.679760597766405e-06, 'learning_rate/action_model': 9.679760597766405e-05, 'epoch': 0.05}) +[2026-10-03 14:45:31,703][starVLA.training.train_starvla][INFO] - Step 677, Loss: {'train/loss': 0.5033807754516602, 'train/grad_norm_pre_clip': 20.275890350341797, 'timing/data': 0.00019619001250248402, 'timing/model': 0.3642445480072638, 'learning_rate/qwen_vl_interface': 9.678660411116796e-06, 'learning_rate/action_model': 9.678660411116796e-05, 'epoch': 0.05}) +[2026-10-03 14:45:32,350][starVLA.training.train_starvla][INFO] - Step 678, Loss: {'train/loss': 0.16797153651714325, 'train/grad_norm_pre_clip': 38.30036926269531, 'timing/data': 0.00043420000292826444, 'timing/model': 0.36457240699382965, 'learning_rate/qwen_vl_interface': 9.677558404012178e-06, 'learning_rate/action_model': 9.677558404012177e-05, 'epoch': 0.05}) +[2026-10-03 14:45:32,998][starVLA.training.train_starvla][INFO] - Step 679, Loss: {'train/loss': 0.23948699235916138, 'train/grad_norm_pre_clip': 20.37281608581543, 'timing/data': 0.00019586000416893512, 'timing/model': 0.3654025180003373, 'learning_rate/qwen_vl_interface': 9.676454576905542e-06, 'learning_rate/action_model': 9.676454576905542e-05, 'epoch': 0.05}) +[2026-10-03 14:45:33,847][starVLA.training.train_starvla][INFO] - Step 680, Loss: {'train/loss': 0.23396393656730652, 'train/grad_norm_pre_clip': 20.529714584350586, 'timing/data': 0.0002137099945684895, 'timing/model': 0.36472741700708866, 'learning_rate/qwen_vl_interface': 9.675348930250633e-06, 'learning_rate/action_model': 9.675348930250632e-05, 'epoch': 0.05}) +[2026-10-03 14:45:34,471][starVLA.training.train_starvla][INFO] - Step 681, Loss: {'train/loss': 0.09946002811193466, 'train/grad_norm_pre_clip': 32.68662643432617, 'timing/data': 0.00019885999790858477, 'timing/model': 0.34157271998992655, 'learning_rate/qwen_vl_interface': 9.674241464501938e-06, 'learning_rate/action_model': 9.674241464501937e-05, 'epoch': 0.05}) +[2026-10-03 14:45:35,118][starVLA.training.train_starvla][INFO] - Step 682, Loss: {'train/loss': 0.07985004037618637, 'train/grad_norm_pre_clip': 22.47559356689453, 'timing/data': 0.00020345000666566193, 'timing/model': 0.36495347700838465, 'learning_rate/qwen_vl_interface': 9.673132180114695e-06, 'learning_rate/action_model': 9.673132180114695e-05, 'epoch': 0.05}) +[2026-10-03 14:45:35,764][starVLA.training.train_starvla][INFO] - Step 683, Loss: {'train/loss': 0.38168656826019287, 'train/grad_norm_pre_clip': 36.63973617553711, 'timing/data': 0.00030860000697430223, 'timing/model': 0.3643552669964265, 'learning_rate/qwen_vl_interface': 9.67202107754489e-06, 'learning_rate/action_model': 9.672021077544889e-05, 'epoch': 0.05}) +[2026-10-03 14:45:36,390][starVLA.training.train_starvla][INFO] - Step 684, Loss: {'train/loss': 0.06767424941062927, 'train/grad_norm_pre_clip': 21.707347869873047, 'timing/data': 0.00019567999697756022, 'timing/model': 0.3427713900018716, 'learning_rate/qwen_vl_interface': 9.670908157249254e-06, 'learning_rate/action_model': 9.670908157249253e-05, 'epoch': 0.05}) +[2026-10-03 14:45:37,234][starVLA.training.train_starvla][INFO] - Step 685, Loss: {'train/loss': 0.0237744003534317, 'train/grad_norm_pre_clip': 3.5810444355010986, 'timing/data': 0.00019790999067481607, 'timing/model': 0.5610343440057477, 'learning_rate/qwen_vl_interface': 9.669793419685267e-06, 'learning_rate/action_model': 9.669793419685266e-05, 'epoch': 0.05}) +[2026-10-03 14:45:37,858][starVLA.training.train_starvla][INFO] - Step 686, Loss: {'train/loss': 0.5603078603744507, 'train/grad_norm_pre_clip': 43.85650634765625, 'timing/data': 0.00019874999998137355, 'timing/model': 0.34168514001066796, 'learning_rate/qwen_vl_interface': 9.668676865311155e-06, 'learning_rate/action_model': 9.668676865311154e-05, 'epoch': 0.05}) +[2026-10-03 14:45:38,504][starVLA.training.train_starvla][INFO] - Step 687, Loss: {'train/loss': 0.08426114916801453, 'train/grad_norm_pre_clip': 57.49123001098633, 'timing/data': 0.00019423000048846006, 'timing/model': 0.3644586770096794, 'learning_rate/qwen_vl_interface': 9.667558494585891e-06, 'learning_rate/action_model': 9.66755849458589e-05, 'epoch': 0.05}) +[2026-10-03 14:45:39,152][starVLA.training.train_starvla][INFO] - Step 688, Loss: {'train/loss': 0.11334560066461563, 'train/grad_norm_pre_clip': 71.89177703857422, 'timing/data': 0.00029424999956972897, 'timing/model': 0.3646412169910036, 'learning_rate/qwen_vl_interface': 9.666438307969194e-06, 'learning_rate/action_model': 9.666438307969195e-05, 'epoch': 0.05}) +[2026-10-03 14:45:39,798][starVLA.training.train_starvla][INFO] - Step 689, Loss: {'train/loss': 0.017842400819063187, 'train/grad_norm_pre_clip': 5.4094085693359375, 'timing/data': 0.00019921000057365745, 'timing/model': 0.36371970699110534, 'learning_rate/qwen_vl_interface': 9.665316305921535e-06, 'learning_rate/action_model': 9.665316305921535e-05, 'epoch': 0.05}) +[2026-10-03 14:45:40,443][starVLA.training.train_starvla][INFO] - Step 690, Loss: {'train/loss': 0.10610570758581161, 'train/grad_norm_pre_clip': 32.84357833862305, 'timing/data': 0.0003102000046055764, 'timing/model': 0.3638825980015099, 'learning_rate/qwen_vl_interface': 9.664192488904123e-06, 'learning_rate/action_model': 9.664192488904122e-05, 'epoch': 0.05}) +[2026-10-03 14:45:41,281][starVLA.training.train_starvla][INFO] - Step 691, Loss: {'train/loss': 0.05441685765981674, 'train/grad_norm_pre_clip': 16.20979881286621, 'timing/data': 0.00020173999655526131, 'timing/model': 0.5547357359901071, 'learning_rate/qwen_vl_interface': 9.663066857378917e-06, 'learning_rate/action_model': 9.663066857378917e-05, 'epoch': 0.05}) +[2026-10-03 14:45:41,927][starVLA.training.train_starvla][INFO] - Step 692, Loss: {'train/loss': 0.3099982440471649, 'train/grad_norm_pre_clip': 26.991878509521484, 'timing/data': 0.00031777001277077943, 'timing/model': 0.3640487080119783, 'learning_rate/qwen_vl_interface': 9.661939411808623e-06, 'learning_rate/action_model': 9.661939411808622e-05, 'epoch': 0.05}) +[2026-10-03 14:45:42,574][starVLA.training.train_starvla][INFO] - Step 693, Loss: {'train/loss': 0.02344752848148346, 'train/grad_norm_pre_clip': 5.883421897888184, 'timing/data': 0.00019614001212175936, 'timing/model': 0.3638637679978274, 'learning_rate/qwen_vl_interface': 9.66081015265669e-06, 'learning_rate/action_model': 9.660810152656689e-05, 'epoch': 0.05}) +[2026-10-03 14:45:43,220][starVLA.training.train_starvla][INFO] - Step 694, Loss: {'train/loss': 0.43633872270584106, 'train/grad_norm_pre_clip': 78.63496398925781, 'timing/data': 0.00027501999284140766, 'timing/model': 0.3643730569892796, 'learning_rate/qwen_vl_interface': 9.659679080387314e-06, 'learning_rate/action_model': 9.659679080387313e-05, 'epoch': 0.05}) +[2026-10-03 14:45:43,866][starVLA.training.train_starvla][INFO] - Step 695, Loss: {'train/loss': 0.513823390007019, 'train/grad_norm_pre_clip': 67.09066009521484, 'timing/data': 0.00020012998720631003, 'timing/model': 0.3644686569896294, 'learning_rate/qwen_vl_interface': 9.658546195465436e-06, 'learning_rate/action_model': 9.658546195465435e-05, 'epoch': 0.05}) +[2026-10-03 14:45:44,687][starVLA.training.train_starvla][INFO] - Step 696, Loss: {'train/loss': 0.11970854550600052, 'train/grad_norm_pre_clip': 18.3523006439209, 'timing/data': 0.00032935000490397215, 'timing/model': 0.5603961040033028, 'learning_rate/qwen_vl_interface': 9.65741149835674e-06, 'learning_rate/action_model': 9.657411498356741e-05, 'epoch': 0.05}) +[2026-10-03 14:45:45,333][starVLA.training.train_starvla][INFO] - Step 697, Loss: {'train/loss': 0.10766389966011047, 'train/grad_norm_pre_clip': 123.17244720458984, 'timing/data': 0.00032993000058922917, 'timing/model': 0.36414142699504737, 'learning_rate/qwen_vl_interface': 9.656274989527663e-06, 'learning_rate/action_model': 9.656274989527664e-05, 'epoch': 0.05}) +[2026-10-03 14:45:45,979][starVLA.training.train_starvla][INFO] - Step 698, Loss: {'train/loss': 0.022020677104592323, 'train/grad_norm_pre_clip': 12.416147232055664, 'timing/data': 0.00023893000616226345, 'timing/model': 0.3642746480036294, 'learning_rate/qwen_vl_interface': 9.655136669445378e-06, 'learning_rate/action_model': 9.655136669445378e-05, 'epoch': 0.05}) +[2026-10-03 14:45:46,626][starVLA.training.train_starvla][INFO] - Step 699, Loss: {'train/loss': 0.3110184967517853, 'train/grad_norm_pre_clip': 9.969895362854004, 'timing/data': 0.00029369899129960686, 'timing/model': 0.3648295279999729, 'learning_rate/qwen_vl_interface': 9.653996538577802e-06, 'learning_rate/action_model': 9.653996538577802e-05, 'epoch': 0.05}) +[2026-10-03 14:45:47,274][starVLA.training.train_starvla][INFO] - Step 700, Loss: {'train/loss': 0.01341240294277668, 'train/grad_norm_pre_clip': 2.9885599613189697, 'timing/data': 0.0001942100061569363, 'timing/model': 0.3650460969947744, 'learning_rate/qwen_vl_interface': 9.652854597393604e-06, 'learning_rate/action_model': 9.652854597393603e-05, 'epoch': 0.05}) +[2026-10-03 14:45:47,898][starVLA.training.train_starvla][INFO] - Step 701, Loss: {'train/loss': 0.007758507505059242, 'train/grad_norm_pre_clip': 8.640767097473145, 'timing/data': 0.0001979100052267313, 'timing/model': 0.3640307869936805, 'learning_rate/qwen_vl_interface': 9.651710846362192e-06, 'learning_rate/action_model': 9.651710846362192e-05, 'epoch': 0.05}) +[2026-10-03 14:45:48,744][starVLA.training.train_starvla][INFO] - Step 702, Loss: {'train/loss': 0.4640156924724579, 'train/grad_norm_pre_clip': 32.027645111083984, 'timing/data': 0.00020231999224051833, 'timing/model': 0.3643566069949884, 'learning_rate/qwen_vl_interface': 9.650565285953716e-06, 'learning_rate/action_model': 9.650565285953715e-05, 'epoch': 0.05}) +[2026-10-03 14:45:49,391][starVLA.training.train_starvla][INFO] - Step 703, Loss: {'train/loss': 0.034980058670043945, 'train/grad_norm_pre_clip': 20.99765968322754, 'timing/data': 0.0001965100091183558, 'timing/model': 0.3647460969950771, 'learning_rate/qwen_vl_interface': 9.649417916639077e-06, 'learning_rate/action_model': 9.649417916639076e-05, 'epoch': 0.05}) +[2026-10-03 14:45:50,039][starVLA.training.train_starvla][INFO] - Step 704, Loss: {'train/loss': 0.3907741606235504, 'train/grad_norm_pre_clip': 128.0163116455078, 'timing/data': 0.0001946300035342574, 'timing/model': 0.3648800270020729, 'learning_rate/qwen_vl_interface': 9.648268738889911e-06, 'learning_rate/action_model': 9.648268738889911e-05, 'epoch': 0.05}) +[2026-10-03 14:45:50,687][starVLA.training.train_starvla][INFO] - Step 705, Loss: {'train/loss': 0.31654706597328186, 'train/grad_norm_pre_clip': 45.3626708984375, 'timing/data': 0.00020009999570902437, 'timing/model': 0.36559649799892213, 'learning_rate/qwen_vl_interface': 9.647117753178604e-06, 'learning_rate/action_model': 9.647117753178604e-05, 'epoch': 0.05}) +[2026-10-03 14:45:51,333][starVLA.training.train_starvla][INFO] - Step 706, Loss: {'train/loss': 0.04957594349980354, 'train/grad_norm_pre_clip': 43.10078048706055, 'timing/data': 0.0001958599896170199, 'timing/model': 0.3643048880039714, 'learning_rate/qwen_vl_interface': 9.645964959978284e-06, 'learning_rate/action_model': 9.645964959978284e-05, 'epoch': 0.05}) +[2026-10-03 14:45:52,179][starVLA.training.train_starvla][INFO] - Step 707, Loss: {'train/loss': 0.15120971202850342, 'train/grad_norm_pre_clip': 110.39724731445312, 'timing/data': 0.00021264998940750957, 'timing/model': 0.36481897700286936, 'learning_rate/qwen_vl_interface': 9.644810359762818e-06, 'learning_rate/action_model': 9.644810359762817e-05, 'epoch': 0.05}) +[2026-10-03 14:45:52,805][starVLA.training.train_starvla][INFO] - Step 708, Loss: {'train/loss': 0.38295257091522217, 'train/grad_norm_pre_clip': 19.423032760620117, 'timing/data': 0.00019612000323832035, 'timing/model': 0.3646832669910509, 'learning_rate/qwen_vl_interface': 9.64365395300682e-06, 'learning_rate/action_model': 9.643653953006819e-05, 'epoch': 0.05}) +[2026-10-03 14:45:53,452][starVLA.training.train_starvla][INFO] - Step 709, Loss: {'train/loss': 0.11158883571624756, 'train/grad_norm_pre_clip': 71.1383285522461, 'timing/data': 0.00019904000509995967, 'timing/model': 0.3642643479979597, 'learning_rate/qwen_vl_interface': 9.642495740185644e-06, 'learning_rate/action_model': 9.642495740185643e-05, 'epoch': 0.05}) +[2026-10-03 14:45:54,076][starVLA.training.train_starvla][INFO] - Step 710, Loss: {'train/loss': 1.765893578529358, 'train/grad_norm_pre_clip': 209.87570190429688, 'timing/data': 0.00019778999558184296, 'timing/model': 0.3416170700074872, 'learning_rate/qwen_vl_interface': 9.641335721775388e-06, 'learning_rate/action_model': 9.641335721775387e-05, 'epoch': 0.05}) +[2026-10-03 14:45:54,722][starVLA.training.train_starvla][INFO] - Step 711, Loss: {'train/loss': 0.3096333146095276, 'train/grad_norm_pre_clip': 7.759307384490967, 'timing/data': 0.00036970000655855983, 'timing/model': 0.36431552800058853, 'learning_rate/qwen_vl_interface': 9.640173898252893e-06, 'learning_rate/action_model': 9.640173898252892e-05, 'epoch': 0.05}) +[2026-10-03 14:45:55,567][starVLA.training.train_starvla][INFO] - Step 712, Loss: {'train/loss': 0.04916517063975334, 'train/grad_norm_pre_clip': 83.21691131591797, 'timing/data': 0.0002603600005386397, 'timing/model': 0.3643179280043114, 'learning_rate/qwen_vl_interface': 9.639010270095738e-06, 'learning_rate/action_model': 9.639010270095739e-05, 'epoch': 0.05}) +[2026-10-03 14:45:56,191][starVLA.training.train_starvla][INFO] - Step 713, Loss: {'train/loss': 0.15419459342956543, 'train/grad_norm_pre_clip': 7.384910583496094, 'timing/data': 0.00019793999672401696, 'timing/model': 0.3420507499977248, 'learning_rate/qwen_vl_interface': 9.637844837782251e-06, 'learning_rate/action_model': 9.63784483778225e-05, 'epoch': 0.05}) +[2026-10-03 14:45:56,837][starVLA.training.train_starvla][INFO] - Step 714, Loss: {'train/loss': 0.24260738492012024, 'train/grad_norm_pre_clip': 81.10419464111328, 'timing/data': 0.00019956000323873013, 'timing/model': 0.3643834480026271, 'learning_rate/qwen_vl_interface': 9.636677601791494e-06, 'learning_rate/action_model': 9.636677601791493e-05, 'epoch': 0.05}) +[2026-10-03 14:45:57,484][starVLA.training.train_starvla][INFO] - Step 715, Loss: {'train/loss': 0.0316978357732296, 'train/grad_norm_pre_clip': 4.428421974182129, 'timing/data': 0.000302529995678924, 'timing/model': 0.3647951379971346, 'learning_rate/qwen_vl_interface': 9.635508562603273e-06, 'learning_rate/action_model': 9.635508562603272e-05, 'epoch': 0.05}) +[2026-10-03 14:45:58,131][starVLA.training.train_starvla][INFO] - Step 716, Loss: {'train/loss': 0.7475342750549316, 'train/grad_norm_pre_clip': 126.57278442382812, 'timing/data': 0.00019488000543788075, 'timing/model': 0.36434586699760985, 'learning_rate/qwen_vl_interface': 9.634337720698139e-06, 'learning_rate/action_model': 9.634337720698137e-05, 'epoch': 0.05}) +[2026-10-03 14:45:58,973][starVLA.training.train_starvla][INFO] - Step 717, Loss: {'train/loss': 0.1741620898246765, 'train/grad_norm_pre_clip': 32.35855484008789, 'timing/data': 0.00019783999596256763, 'timing/model': 0.5606265740061644, 'learning_rate/qwen_vl_interface': 9.633165076557376e-06, 'learning_rate/action_model': 9.633165076557375e-05, 'epoch': 0.05}) +[2026-10-03 14:45:59,620][starVLA.training.train_starvla][INFO] - Step 718, Loss: {'train/loss': 0.2518910765647888, 'train/grad_norm_pre_clip': 560.8128662109375, 'timing/data': 0.00019901999621652067, 'timing/model': 0.36494336801115423, 'learning_rate/qwen_vl_interface': 9.631990630663019e-06, 'learning_rate/action_model': 9.631990630663018e-05, 'epoch': 0.05}) +[2026-10-03 14:46:00,266][starVLA.training.train_starvla][INFO] - Step 719, Loss: {'train/loss': 0.20425082743167877, 'train/grad_norm_pre_clip': 25.878421783447266, 'timing/data': 0.00020418000349309295, 'timing/model': 0.36463472800096497, 'learning_rate/qwen_vl_interface': 9.630814383497836e-06, 'learning_rate/action_model': 9.630814383497835e-05, 'epoch': 0.05}) +[2026-10-03 14:46:00,915][starVLA.training.train_starvla][INFO] - Step 720, Loss: {'train/loss': 0.02332598902285099, 'train/grad_norm_pre_clip': 12.618072509765625, 'timing/data': 0.00021853999351151288, 'timing/model': 0.3650248379999539, 'learning_rate/qwen_vl_interface': 9.629636335545336e-06, 'learning_rate/action_model': 9.629636335545335e-05, 'epoch': 0.05}) +[2026-10-03 14:46:01,563][starVLA.training.train_starvla][INFO] - Step 721, Loss: {'train/loss': 0.5103883147239685, 'train/grad_norm_pre_clip': 321.4248962402344, 'timing/data': 0.00019752999651245773, 'timing/model': 0.36458766799478326, 'learning_rate/qwen_vl_interface': 9.628456487289773e-06, 'learning_rate/action_model': 9.628456487289773e-05, 'epoch': 0.05}) +[2026-10-03 14:46:02,210][starVLA.training.train_starvla][INFO] - Step 722, Loss: {'train/loss': 0.0239969901740551, 'train/grad_norm_pre_clip': 5.0934343338012695, 'timing/data': 0.00020046999270562083, 'timing/model': 0.3644989080057712, 'learning_rate/qwen_vl_interface': 9.627274839216135e-06, 'learning_rate/action_model': 9.627274839216136e-05, 'epoch': 0.05}) +[2026-10-03 14:46:03,051][starVLA.training.train_starvla][INFO] - Step 723, Loss: {'train/loss': 0.34351423382759094, 'train/grad_norm_pre_clip': 91.26560974121094, 'timing/data': 0.00020062000839971006, 'timing/model': 0.3649576969910413, 'learning_rate/qwen_vl_interface': 9.626091391810158e-06, 'learning_rate/action_model': 9.626091391810157e-05, 'epoch': 0.05}) +[2026-10-03 14:46:03,698][starVLA.training.train_starvla][INFO] - Step 724, Loss: {'train/loss': 0.100709468126297, 'train/grad_norm_pre_clip': 19.635578155517578, 'timing/data': 0.00019524000526871532, 'timing/model': 0.36446370799967553, 'learning_rate/qwen_vl_interface': 9.624906145558309e-06, 'learning_rate/action_model': 9.624906145558307e-05, 'epoch': 0.05}) +[2026-10-03 14:46:04,322][starVLA.training.train_starvla][INFO] - Step 725, Loss: {'train/loss': 0.041437000036239624, 'train/grad_norm_pre_clip': 7.48345947265625, 'timing/data': 0.0001969299919437617, 'timing/model': 0.36440144799416885, 'learning_rate/qwen_vl_interface': 9.623719100947795e-06, 'learning_rate/action_model': 9.623719100947795e-05, 'epoch': 0.05}) +[2026-10-03 14:46:04,970][starVLA.training.train_starvla][INFO] - Step 726, Loss: {'train/loss': 0.8730414509773254, 'train/grad_norm_pre_clip': 42.018131256103516, 'timing/data': 0.00020136000239290297, 'timing/model': 0.36484541700338013, 'learning_rate/qwen_vl_interface': 9.622530258466571e-06, 'learning_rate/action_model': 9.62253025846657e-05, 'epoch': 0.05}) +[2026-10-03 14:46:05,617][starVLA.training.train_starvla][INFO] - Step 727, Loss: {'train/loss': 0.07918599992990494, 'train/grad_norm_pre_clip': 133.0796356201172, 'timing/data': 0.00019842998881358653, 'timing/model': 0.3652929480012972, 'learning_rate/qwen_vl_interface': 9.62133961860332e-06, 'learning_rate/action_model': 9.62133961860332e-05, 'epoch': 0.05}) +[2026-10-03 14:46:06,464][starVLA.training.train_starvla][INFO] - Step 728, Loss: {'train/loss': 0.371751070022583, 'train/grad_norm_pre_clip': 136.49020385742188, 'timing/data': 0.00020454000332392752, 'timing/model': 0.36568388699379284, 'learning_rate/qwen_vl_interface': 9.620147181847474e-06, 'learning_rate/action_model': 9.620147181847475e-05, 'epoch': 0.05}) +[2026-10-03 14:46:07,111][starVLA.training.train_starvla][INFO] - Step 729, Loss: {'train/loss': 0.03162193298339844, 'train/grad_norm_pre_clip': 7.90088415145874, 'timing/data': 0.00019814999541267753, 'timing/model': 0.364807916994323, 'learning_rate/qwen_vl_interface': 9.618952948689196e-06, 'learning_rate/action_model': 9.618952948689196e-05, 'epoch': 0.05}) +[2026-10-03 14:46:07,759][starVLA.training.train_starvla][INFO] - Step 730, Loss: {'train/loss': 0.520368218421936, 'train/grad_norm_pre_clip': 61.60950469970703, 'timing/data': 0.00019453999993856996, 'timing/model': 0.36506993700459134, 'learning_rate/qwen_vl_interface': 9.617756919619393e-06, 'learning_rate/action_model': 9.617756919619392e-05, 'epoch': 0.05}) +[2026-10-03 14:46:08,407][starVLA.training.train_starvla][INFO] - Step 731, Loss: {'train/loss': 0.02149488776922226, 'train/grad_norm_pre_clip': 13.572430610656738, 'timing/data': 0.00028101001225877553, 'timing/model': 0.36527734699484427, 'learning_rate/qwen_vl_interface': 9.616559095129701e-06, 'learning_rate/action_model': 9.6165590951297e-05, 'epoch': 0.05}) +[2026-10-03 14:46:09,056][starVLA.training.train_starvla][INFO] - Step 732, Loss: {'train/loss': 0.12022758275270462, 'train/grad_norm_pre_clip': 31.90403175354004, 'timing/data': 0.00020036000933032483, 'timing/model': 0.3650612670026021, 'learning_rate/qwen_vl_interface': 9.615359475712506e-06, 'learning_rate/action_model': 9.615359475712506e-05, 'epoch': 0.05}) +[2026-10-03 14:46:09,898][starVLA.training.train_starvla][INFO] - Step 733, Loss: {'train/loss': 0.5294135212898254, 'train/grad_norm_pre_clip': 124.38177490234375, 'timing/data': 0.00020930000755470246, 'timing/model': 0.5583277340047061, 'learning_rate/qwen_vl_interface': 9.614158061860924e-06, 'learning_rate/action_model': 9.614158061860924e-05, 'epoch': 0.05}) +[2026-10-03 14:46:10,523][starVLA.training.train_starvla][INFO] - Step 734, Loss: {'train/loss': 0.5473702549934387, 'train/grad_norm_pre_clip': 90.31498718261719, 'timing/data': 0.00026450000586919487, 'timing/model': 0.342230340000242, 'learning_rate/qwen_vl_interface': 9.612954854068813e-06, 'learning_rate/action_model': 9.612954854068812e-05, 'epoch': 0.05}) +[2026-10-03 14:46:11,171][starVLA.training.train_starvla][INFO] - Step 735, Loss: {'train/loss': 0.009084413759410381, 'train/grad_norm_pre_clip': 3.0326738357543945, 'timing/data': 0.00020091999613214284, 'timing/model': 0.36521338699094485, 'learning_rate/qwen_vl_interface': 9.611749852830763e-06, 'learning_rate/action_model': 9.611749852830763e-05, 'epoch': 0.05}) +[2026-10-03 14:46:11,819][starVLA.training.train_starvla][INFO] - Step 736, Loss: {'train/loss': 0.0072048502042889595, 'train/grad_norm_pre_clip': 1.7183483839035034, 'timing/data': 0.00019993999740108848, 'timing/model': 0.3651839569938602, 'learning_rate/qwen_vl_interface': 9.610543058642107e-06, 'learning_rate/action_model': 9.610543058642107e-05, 'epoch': 0.05}) +[2026-10-03 14:46:12,467][starVLA.training.train_starvla][INFO] - Step 737, Loss: {'train/loss': 0.4751300513744354, 'train/grad_norm_pre_clip': 23.890352249145508, 'timing/data': 0.0002467800077283755, 'timing/model': 0.3650105170090683, 'learning_rate/qwen_vl_interface': 9.609334471998911e-06, 'learning_rate/action_model': 9.609334471998912e-05, 'epoch': 0.05}) +[2026-10-03 14:46:13,312][starVLA.training.train_starvla][INFO] - Step 738, Loss: {'train/loss': 0.034840650856494904, 'train/grad_norm_pre_clip': 72.25907897949219, 'timing/data': 0.0003121399931842461, 'timing/model': 0.5616158939956222, 'learning_rate/qwen_vl_interface': 9.60812409339798e-06, 'learning_rate/action_model': 9.60812409339798e-05, 'epoch': 0.05}) +[2026-10-03 14:46:13,960][starVLA.training.train_starvla][INFO] - Step 739, Loss: {'train/loss': 0.4286612868309021, 'train/grad_norm_pre_clip': 30.62786102294922, 'timing/data': 0.00019802000315394253, 'timing/model': 0.36526975700689945, 'learning_rate/qwen_vl_interface': 9.606911923336856e-06, 'learning_rate/action_model': 9.606911923336856e-05, 'epoch': 0.05}) +[2026-10-03 14:46:14,609][starVLA.training.train_starvla][INFO] - Step 740, Loss: {'train/loss': 0.04437562823295593, 'train/grad_norm_pre_clip': 110.64747619628906, 'timing/data': 0.0002458200033288449, 'timing/model': 0.36502230699989013, 'learning_rate/qwen_vl_interface': 9.605697962313817e-06, 'learning_rate/action_model': 9.605697962313816e-05, 'epoch': 0.05}) +[2026-10-03 14:46:15,256][starVLA.training.train_starvla][INFO] - Step 741, Loss: {'train/loss': 0.3085363507270813, 'train/grad_norm_pre_clip': 40.52898025512695, 'timing/data': 0.00034030999813694507, 'timing/model': 0.36471549700945616, 'learning_rate/qwen_vl_interface': 9.604482210827871e-06, 'learning_rate/action_model': 9.604482210827871e-05, 'epoch': 0.05}) +[2026-10-03 14:46:15,904][starVLA.training.train_starvla][INFO] - Step 742, Loss: {'train/loss': 0.42640453577041626, 'train/grad_norm_pre_clip': 9.360907554626465, 'timing/data': 0.00019885000074282289, 'timing/model': 0.3650017270119861, 'learning_rate/qwen_vl_interface': 9.603264669378772e-06, 'learning_rate/action_model': 9.603264669378772e-05, 'epoch': 0.05}) +[2026-10-03 14:46:16,552][starVLA.training.train_starvla][INFO] - Step 743, Loss: {'train/loss': 0.09458238631486893, 'train/grad_norm_pre_clip': 45.36985778808594, 'timing/data': 0.00019712001085281372, 'timing/model': 0.3653618069947697, 'learning_rate/qwen_vl_interface': 9.602045338467007e-06, 'learning_rate/action_model': 9.602045338467006e-05, 'epoch': 0.05}) +[2026-10-03 14:46:17,368][starVLA.training.train_starvla][INFO] - Step 744, Loss: {'train/loss': 0.11530843377113342, 'train/grad_norm_pre_clip': 128.75083923339844, 'timing/data': 0.00020441999367903918, 'timing/model': 0.3655163469957188, 'learning_rate/qwen_vl_interface': 9.600824218593791e-06, 'learning_rate/action_model': 9.600824218593791e-05, 'epoch': 0.05}) +[2026-10-03 14:46:18,017][starVLA.training.train_starvla][INFO] - Step 745, Loss: {'train/loss': 0.009862301871180534, 'train/grad_norm_pre_clip': 2.002250909805298, 'timing/data': 0.0001996000064536929, 'timing/model': 0.3650406369997654, 'learning_rate/qwen_vl_interface': 9.599601310261087e-06, 'learning_rate/action_model': 9.599601310261085e-05, 'epoch': 0.05}) +[2026-10-03 14:46:18,664][starVLA.training.train_starvla][INFO] - Step 746, Loss: {'train/loss': 0.22116483747959137, 'train/grad_norm_pre_clip': 27.37984848022461, 'timing/data': 0.00020028000290039927, 'timing/model': 0.36484427700634114, 'learning_rate/qwen_vl_interface': 9.598376613971582e-06, 'learning_rate/action_model': 9.598376613971582e-05, 'epoch': 0.05}) +[2026-10-03 14:46:19,312][starVLA.training.train_starvla][INFO] - Step 747, Loss: {'train/loss': 0.5003035068511963, 'train/grad_norm_pre_clip': 100.03884887695312, 'timing/data': 0.0002386800042586401, 'timing/model': 0.36489548700046726, 'learning_rate/qwen_vl_interface': 9.597150130228703e-06, 'learning_rate/action_model': 9.597150130228703e-05, 'epoch': 0.05}) +[2026-10-03 14:46:19,960][starVLA.training.train_starvla][INFO] - Step 748, Loss: {'train/loss': 0.00658594211563468, 'train/grad_norm_pre_clip': 0.7346428632736206, 'timing/data': 0.00019704000442288816, 'timing/model': 0.3649199169885833, 'learning_rate/qwen_vl_interface': 9.595921859536612e-06, 'learning_rate/action_model': 9.595921859536611e-05, 'epoch': 0.05}) +[2026-10-03 14:46:20,800][starVLA.training.train_starvla][INFO] - Step 749, Loss: {'train/loss': 0.35606649518013, 'train/grad_norm_pre_clip': 34.51319885253906, 'timing/data': 0.0002003800036618486, 'timing/model': 0.5563001549890032, 'learning_rate/qwen_vl_interface': 9.594691802400203e-06, 'learning_rate/action_model': 9.594691802400202e-05, 'epoch': 0.05}) +[2026-10-03 14:46:21,448][starVLA.training.train_starvla][INFO] - Step 750, Loss: {'train/loss': 0.18657369911670685, 'train/grad_norm_pre_clip': 25.442644119262695, 'timing/data': 0.0001983800029847771, 'timing/model': 0.3653084170073271, 'learning_rate/qwen_vl_interface': 9.59345995932511e-06, 'learning_rate/action_model': 9.59345995932511e-05, 'epoch': 0.05}) +[2026-10-03 14:46:22,096][starVLA.training.train_starvla][INFO] - Step 751, Loss: {'train/loss': 0.4140232503414154, 'train/grad_norm_pre_clip': 59.37317657470703, 'timing/data': 0.00019927001267205924, 'timing/model': 0.3658139760082122, 'learning_rate/qwen_vl_interface': 9.592226330817693e-06, 'learning_rate/action_model': 9.592226330817693e-05, 'epoch': 0.05}) +[2026-10-03 14:46:22,745][starVLA.training.train_starvla][INFO] - Step 752, Loss: {'train/loss': 0.355063796043396, 'train/grad_norm_pre_clip': 135.0386962890625, 'timing/data': 0.00019542999507393688, 'timing/model': 0.365352997003356, 'learning_rate/qwen_vl_interface': 9.590990917385052e-06, 'learning_rate/action_model': 9.590990917385052e-05, 'epoch': 0.06}) +[2026-10-03 14:46:23,370][starVLA.training.train_starvla][INFO] - Step 753, Loss: {'train/loss': 0.7116759419441223, 'train/grad_norm_pre_clip': 78.79971313476562, 'timing/data': 0.00019762000010814518, 'timing/model': 0.3421639110019896, 'learning_rate/qwen_vl_interface': 9.58975371953502e-06, 'learning_rate/action_model': 9.589753719535018e-05, 'epoch': 0.06}) +[2026-10-03 14:46:24,209][starVLA.training.train_starvla][INFO] - Step 754, Loss: {'train/loss': 0.276985764503479, 'train/grad_norm_pre_clip': 37.75891876220703, 'timing/data': 0.00019963999511674047, 'timing/model': 0.5546804949990474, 'learning_rate/qwen_vl_interface': 9.588514737776158e-06, 'learning_rate/action_model': 9.588514737776157e-05, 'epoch': 0.06}) +[2026-10-03 14:46:24,856][starVLA.training.train_starvla][INFO] - Step 755, Loss: {'train/loss': 0.013159682974219322, 'train/grad_norm_pre_clip': 3.8855783939361572, 'timing/data': 0.0002180599985877052, 'timing/model': 0.3654158170102164, 'learning_rate/qwen_vl_interface': 9.587273972617769e-06, 'learning_rate/action_model': 9.587273972617768e-05, 'epoch': 0.06}) +[2026-10-03 14:46:25,505][starVLA.training.train_starvla][INFO] - Step 756, Loss: {'train/loss': 0.012633511796593666, 'train/grad_norm_pre_clip': 3.2922093868255615, 'timing/data': 0.0001939000067068264, 'timing/model': 0.3657382579986006, 'learning_rate/qwen_vl_interface': 9.586031424569882e-06, 'learning_rate/action_model': 9.58603142456988e-05, 'epoch': 0.06}) +[2026-10-03 14:46:26,154][starVLA.training.train_starvla][INFO] - Step 757, Loss: {'train/loss': 0.04928407073020935, 'train/grad_norm_pre_clip': 34.404296875, 'timing/data': 0.0001943399984156713, 'timing/model': 0.36521502799587324, 'learning_rate/qwen_vl_interface': 9.584787094143265e-06, 'learning_rate/action_model': 9.584787094143264e-05, 'epoch': 0.06}) +[2026-10-03 14:46:26,779][starVLA.training.train_starvla][INFO] - Step 758, Loss: {'train/loss': 0.024014372378587723, 'train/grad_norm_pre_clip': 22.472997665405273, 'timing/data': 0.000281209999229759, 'timing/model': 0.34219183000095654, 'learning_rate/qwen_vl_interface': 9.58354098184941e-06, 'learning_rate/action_model': 9.583540981849409e-05, 'epoch': 0.06}) +[2026-10-03 14:46:27,626][starVLA.training.train_starvla][INFO] - Step 759, Loss: {'train/loss': 0.535548985004425, 'train/grad_norm_pre_clip': 142.5154266357422, 'timing/data': 0.00033101999724749476, 'timing/model': 0.5623818839958403, 'learning_rate/qwen_vl_interface': 9.582293088200549e-06, 'learning_rate/action_model': 9.582293088200549e-05, 'epoch': 0.06}) +[2026-10-03 14:46:28,275][starVLA.training.train_starvla][INFO] - Step 760, Loss: {'train/loss': 0.06748870015144348, 'train/grad_norm_pre_clip': 14.85818099975586, 'timing/data': 0.00026681998861022294, 'timing/model': 0.36535289700259455, 'learning_rate/qwen_vl_interface': 9.581043413709646e-06, 'learning_rate/action_model': 9.581043413709646e-05, 'epoch': 0.06}) +[2026-10-03 14:46:28,923][starVLA.training.train_starvla][INFO] - Step 761, Loss: {'train/loss': 0.20901551842689514, 'train/grad_norm_pre_clip': 79.14727783203125, 'timing/data': 0.00036901999555993825, 'timing/model': 0.36503402699599974, 'learning_rate/qwen_vl_interface': 9.579791958890393e-06, 'learning_rate/action_model': 9.579791958890394e-05, 'epoch': 0.06}) +[2026-10-03 14:46:29,571][starVLA.training.train_starvla][INFO] - Step 762, Loss: {'train/loss': 0.19772712886333466, 'train/grad_norm_pre_clip': 53.84158706665039, 'timing/data': 0.00020697999570984393, 'timing/model': 0.3649324669968337, 'learning_rate/qwen_vl_interface': 9.578538724257218e-06, 'learning_rate/action_model': 9.578538724257216e-05, 'epoch': 0.06}) +[2026-10-03 14:46:30,220][starVLA.training.train_starvla][INFO] - Step 763, Loss: {'train/loss': 0.05365590378642082, 'train/grad_norm_pre_clip': 68.923583984375, 'timing/data': 0.00019957000040449202, 'timing/model': 0.36530879700148944, 'learning_rate/qwen_vl_interface': 9.577283710325274e-06, 'learning_rate/action_model': 9.577283710325274e-05, 'epoch': 0.06}) +[2026-10-03 14:46:31,068][starVLA.training.train_starvla][INFO] - Step 764, Loss: {'train/loss': 1.0593554973602295, 'train/grad_norm_pre_clip': 52.300498962402344, 'timing/data': 0.00026123000134248286, 'timing/model': 0.5647975539905019, 'learning_rate/qwen_vl_interface': 9.576026917610455e-06, 'learning_rate/action_model': 9.576026917610455e-05, 'epoch': 0.06}) +[2026-10-03 14:46:31,719][starVLA.training.train_starvla][INFO] - Step 765, Loss: {'train/loss': 0.009928584098815918, 'train/grad_norm_pre_clip': 5.719570636749268, 'timing/data': 0.0002360900107305497, 'timing/model': 0.3660824760008836, 'learning_rate/qwen_vl_interface': 9.574768346629376e-06, 'learning_rate/action_model': 9.574768346629376e-05, 'epoch': 0.06}) +[2026-10-03 14:46:32,370][starVLA.training.train_starvla][INFO] - Step 766, Loss: {'train/loss': 0.006514124572277069, 'train/grad_norm_pre_clip': 1.34357488155365, 'timing/data': 0.0002663199993548915, 'timing/model': 0.3667003659938928, 'learning_rate/qwen_vl_interface': 9.573507997899391e-06, 'learning_rate/action_model': 9.573507997899392e-05, 'epoch': 0.06}) +[2026-10-03 14:46:33,020][starVLA.training.train_starvla][INFO] - Step 767, Loss: {'train/loss': 0.4528142213821411, 'train/grad_norm_pre_clip': 110.5382080078125, 'timing/data': 0.0002990399952977896, 'timing/model': 0.3661361570120789, 'learning_rate/qwen_vl_interface': 9.572245871938582e-06, 'learning_rate/action_model': 9.572245871938582e-05, 'epoch': 0.06}) +[2026-10-03 14:46:33,672][starVLA.training.train_starvla][INFO] - Step 768, Loss: {'train/loss': 0.10485641658306122, 'train/grad_norm_pre_clip': 23.27536964416504, 'timing/data': 0.0002894500066759065, 'timing/model': 0.36711381599889137, 'learning_rate/qwen_vl_interface': 9.570981969265761e-06, 'learning_rate/action_model': 9.570981969265761e-05, 'epoch': 0.06}) +[2026-10-03 14:46:34,323][starVLA.training.train_starvla][INFO] - Step 769, Loss: {'train/loss': 0.10056788474321365, 'train/grad_norm_pre_clip': 67.31306457519531, 'timing/data': 0.00035614999069366604, 'timing/model': 0.36595981699065305, 'learning_rate/qwen_vl_interface': 9.56971629040047e-06, 'learning_rate/action_model': 9.56971629040047e-05, 'epoch': 0.06}) +[2026-10-03 14:46:35,167][starVLA.training.train_starvla][INFO] - Step 770, Loss: {'train/loss': 0.29600274562835693, 'train/grad_norm_pre_clip': 25.32308578491211, 'timing/data': 0.00020198000129312277, 'timing/model': 0.36586320800415706, 'learning_rate/qwen_vl_interface': 9.56844883586298e-06, 'learning_rate/action_model': 9.568448835862981e-05, 'epoch': 0.06}) +[2026-10-03 14:46:35,819][starVLA.training.train_starvla][INFO] - Step 771, Loss: {'train/loss': 0.019947592169046402, 'train/grad_norm_pre_clip': 19.31336212158203, 'timing/data': 0.0002456299989717081, 'timing/model': 0.3671935169986682, 'learning_rate/qwen_vl_interface': 9.567179606174299e-06, 'learning_rate/action_model': 9.567179606174299e-05, 'epoch': 0.06}) +[2026-10-03 14:46:36,471][starVLA.training.train_starvla][INFO] - Step 772, Loss: {'train/loss': 0.09628864377737045, 'train/grad_norm_pre_clip': 61.11882019042969, 'timing/data': 0.00028993000159971416, 'timing/model': 0.36711450600705575, 'learning_rate/qwen_vl_interface': 9.565908601856155e-06, 'learning_rate/action_model': 9.565908601856154e-05, 'epoch': 0.06}) +[2026-10-03 14:46:37,121][starVLA.training.train_starvla][INFO] - Step 773, Loss: {'train/loss': 0.009633667767047882, 'train/grad_norm_pre_clip': 12.97326946258545, 'timing/data': 0.00028897999436594546, 'timing/model': 0.3663764770026319, 'learning_rate/qwen_vl_interface': 9.56463582343101e-06, 'learning_rate/action_model': 9.56463582343101e-05, 'epoch': 0.06}) +[2026-10-03 14:46:37,772][starVLA.training.train_starvla][INFO] - Step 774, Loss: {'train/loss': 0.15348036587238312, 'train/grad_norm_pre_clip': 34.749507904052734, 'timing/data': 0.00023705999774392694, 'timing/model': 0.36643469698901754, 'learning_rate/qwen_vl_interface': 9.563361271422056e-06, 'learning_rate/action_model': 9.563361271422056e-05, 'epoch': 0.06}) +[2026-10-03 14:46:38,612][starVLA.training.train_starvla][INFO] - Step 775, Loss: {'train/loss': 0.5027526021003723, 'train/grad_norm_pre_clip': 30.258747100830078, 'timing/data': 0.0002430700114928186, 'timing/model': 0.5559783860080643, 'learning_rate/qwen_vl_interface': 9.562084946353212e-06, 'learning_rate/action_model': 9.562084946353213e-05, 'epoch': 0.06}) +[2026-10-03 14:46:39,262][starVLA.training.train_starvla][INFO] - Step 776, Loss: {'train/loss': 0.009759119711816311, 'train/grad_norm_pre_clip': 5.180988311767578, 'timing/data': 0.00034726900048553944, 'timing/model': 0.36551947800035123, 'learning_rate/qwen_vl_interface': 9.560806848749132e-06, 'learning_rate/action_model': 9.560806848749132e-05, 'epoch': 0.06}) +[2026-10-03 14:46:39,910][starVLA.training.train_starvla][INFO] - Step 777, Loss: {'train/loss': 0.09625899791717529, 'train/grad_norm_pre_clip': 33.45228958129883, 'timing/data': 0.0002604189940029755, 'timing/model': 0.3645042769931024, 'learning_rate/qwen_vl_interface': 9.559526979135186e-06, 'learning_rate/action_model': 9.559526979135187e-05, 'epoch': 0.06}) +[2026-10-03 14:46:40,558][starVLA.training.train_starvla][INFO] - Step 778, Loss: {'train/loss': 0.14944222569465637, 'train/grad_norm_pre_clip': 39.305763244628906, 'timing/data': 0.0002035100042121485, 'timing/model': 0.3654685259971302, 'learning_rate/qwen_vl_interface': 9.558245338037486e-06, 'learning_rate/action_model': 9.558245338037486e-05, 'epoch': 0.06}) +[2026-10-03 14:46:41,207][starVLA.training.train_starvla][INFO] - Step 779, Loss: {'train/loss': 0.024518249556422234, 'train/grad_norm_pre_clip': 5.863524436950684, 'timing/data': 0.00019874999998137355, 'timing/model': 0.36601305700605735, 'learning_rate/qwen_vl_interface': 9.556961925982864e-06, 'learning_rate/action_model': 9.556961925982863e-05, 'epoch': 0.06}) +[2026-10-03 14:46:42,054][starVLA.training.train_starvla][INFO] - Step 780, Loss: {'train/loss': 0.3665207624435425, 'train/grad_norm_pre_clip': 37.751014709472656, 'timing/data': 0.0001963199902093038, 'timing/model': 0.5633736539894016, 'learning_rate/qwen_vl_interface': 9.555676743498882e-06, 'learning_rate/action_model': 9.555676743498881e-05, 'epoch': 0.06}) +[2026-10-03 14:46:42,702][starVLA.training.train_starvla][INFO] - Step 781, Loss: {'train/loss': 0.3374384641647339, 'train/grad_norm_pre_clip': 33.87385940551758, 'timing/data': 0.0001930699945660308, 'timing/model': 0.36502426701190416, 'learning_rate/qwen_vl_interface': 9.55438979111383e-06, 'learning_rate/action_model': 9.554389791113831e-05, 'epoch': 0.06}) +[2026-10-03 14:46:43,350][starVLA.training.train_starvla][INFO] - Step 782, Loss: {'train/loss': 0.09416405111551285, 'train/grad_norm_pre_clip': 115.09182739257812, 'timing/data': 0.00019795999105554074, 'timing/model': 0.36518996699305717, 'learning_rate/qwen_vl_interface': 9.553101069356728e-06, 'learning_rate/action_model': 9.553101069356727e-05, 'epoch': 0.06}) +[2026-10-03 14:46:43,998][starVLA.training.train_starvla][INFO] - Step 783, Loss: {'train/loss': 0.011010357178747654, 'train/grad_norm_pre_clip': 3.7975080013275146, 'timing/data': 0.0001975699997274205, 'timing/model': 0.3652307870070217, 'learning_rate/qwen_vl_interface': 9.551810578757316e-06, 'learning_rate/action_model': 9.551810578757315e-05, 'epoch': 0.06}) +[2026-10-03 14:46:44,623][starVLA.training.train_starvla][INFO] - Step 784, Loss: {'train/loss': 0.6839351654052734, 'train/grad_norm_pre_clip': 50.34176254272461, 'timing/data': 0.00019700000120792538, 'timing/model': 0.34264583999174647, 'learning_rate/qwen_vl_interface': 9.550518319846069e-06, 'learning_rate/action_model': 9.550518319846069e-05, 'epoch': 0.06}) +[2026-10-03 14:46:45,471][starVLA.training.train_starvla][INFO] - Step 785, Loss: {'train/loss': 0.08200129866600037, 'train/grad_norm_pre_clip': 47.61308670043945, 'timing/data': 0.00037511000118684024, 'timing/model': 0.5656463949999306, 'learning_rate/qwen_vl_interface': 9.549224293154187e-06, 'learning_rate/action_model': 9.549224293154187e-05, 'epoch': 0.06}) +[2026-10-03 14:46:46,118][starVLA.training.train_starvla][INFO] - Step 786, Loss: {'train/loss': 0.3863503336906433, 'train/grad_norm_pre_clip': 328.36383056640625, 'timing/data': 0.0002099999983329326, 'timing/model': 0.3647988680022536, 'learning_rate/qwen_vl_interface': 9.547928499213593e-06, 'learning_rate/action_model': 9.547928499213593e-05, 'epoch': 0.06}) +[2026-10-03 14:46:46,765][starVLA.training.train_starvla][INFO] - Step 787, Loss: {'train/loss': 0.4299538731575012, 'train/grad_norm_pre_clip': 12.775209426879883, 'timing/data': 0.00024001000565476716, 'timing/model': 0.36430164700141177, 'learning_rate/qwen_vl_interface': 9.54663093855694e-06, 'learning_rate/action_model': 9.54663093855694e-05, 'epoch': 0.06}) +[2026-10-03 14:46:47,415][starVLA.training.train_starvla][INFO] - Step 788, Loss: {'train/loss': 0.17368869483470917, 'train/grad_norm_pre_clip': 49.03972244262695, 'timing/data': 0.0002447400038363412, 'timing/model': 0.3660714160068892, 'learning_rate/qwen_vl_interface': 9.545331611717607e-06, 'learning_rate/action_model': 9.545331611717608e-05, 'epoch': 0.06}) +[2026-10-03 14:46:48,065][starVLA.training.train_starvla][INFO] - Step 789, Loss: {'train/loss': 0.03453512489795685, 'train/grad_norm_pre_clip': 55.27741241455078, 'timing/data': 0.00023874000180512667, 'timing/model': 0.3655912079993868, 'learning_rate/qwen_vl_interface': 9.544030519229697e-06, 'learning_rate/action_model': 9.544030519229697e-05, 'epoch': 0.06}) +[2026-10-03 14:46:48,714][starVLA.training.train_starvla][INFO] - Step 790, Loss: {'train/loss': 0.0032634236849844456, 'train/grad_norm_pre_clip': 0.42412444949150085, 'timing/data': 0.0003242199891246855, 'timing/model': 0.365581028003362, 'learning_rate/qwen_vl_interface': 9.54272766162804e-06, 'learning_rate/action_model': 9.54272766162804e-05, 'epoch': 0.06}) +[2026-10-03 14:46:49,563][starVLA.training.train_starvla][INFO] - Step 791, Loss: {'train/loss': 0.4648064970970154, 'train/grad_norm_pre_clip': 28.461673736572266, 'timing/data': 0.0002434499911032617, 'timing/model': 0.36603287600155454, 'learning_rate/qwen_vl_interface': 9.54142303944819e-06, 'learning_rate/action_model': 9.541423039448191e-05, 'epoch': 0.06}) +[2026-10-03 14:46:50,219][starVLA.training.train_starvla][INFO] - Step 792, Loss: {'train/loss': 0.1849418431520462, 'train/grad_norm_pre_clip': 61.89747619628906, 'timing/data': 0.00019993999740108848, 'timing/model': 0.3663225679920288, 'learning_rate/qwen_vl_interface': 9.540116653226433e-06, 'learning_rate/action_model': 9.540116653226431e-05, 'epoch': 0.06}) +[2026-10-03 14:46:50,871][starVLA.training.train_starvla][INFO] - Step 793, Loss: {'train/loss': 0.0540001355111599, 'train/grad_norm_pre_clip': 54.313499450683594, 'timing/data': 0.0003543799975886941, 'timing/model': 0.36709272699954454, 'learning_rate/qwen_vl_interface': 9.53880850349977e-06, 'learning_rate/action_model': 9.538808503499768e-05, 'epoch': 0.06}) +[2026-10-03 14:46:51,521][starVLA.training.train_starvla][INFO] - Step 794, Loss: {'train/loss': 0.44119536876678467, 'train/grad_norm_pre_clip': 62.72871780395508, 'timing/data': 0.00019659000099636614, 'timing/model': 0.36602788799791597, 'learning_rate/qwen_vl_interface': 9.537498590805934e-06, 'learning_rate/action_model': 9.537498590805933e-05, 'epoch': 0.06}) +[2026-10-03 14:46:52,148][starVLA.training.train_starvla][INFO] - Step 795, Loss: {'train/loss': 0.42354416847229004, 'train/grad_norm_pre_clip': 12.287079811096191, 'timing/data': 0.0003016400005435571, 'timing/model': 0.36591495700122323, 'learning_rate/qwen_vl_interface': 9.536186915683379e-06, 'learning_rate/action_model': 9.53618691568338e-05, 'epoch': 0.06}) +[2026-10-03 14:46:52,990][starVLA.training.train_starvla][INFO] - Step 796, Loss: {'train/loss': 0.047075800597667694, 'train/grad_norm_pre_clip': 67.89171600341797, 'timing/data': 0.00019827000505756587, 'timing/model': 0.5578792839951348, 'learning_rate/qwen_vl_interface': 9.534873478671289e-06, 'learning_rate/action_model': 9.534873478671288e-05, 'epoch': 0.06}) +[2026-10-03 14:46:53,640][starVLA.training.train_starvla][INFO] - Step 797, Loss: {'train/loss': 0.009664238430559635, 'train/grad_norm_pre_clip': 2.371211051940918, 'timing/data': 0.00037588999839499593, 'timing/model': 0.36579518699727487, 'learning_rate/qwen_vl_interface': 9.533558280309563e-06, 'learning_rate/action_model': 9.533558280309563e-05, 'epoch': 0.06}) +[2026-10-03 14:46:54,291][starVLA.training.train_starvla][INFO] - Step 798, Loss: {'train/loss': 0.0060297641903162, 'train/grad_norm_pre_clip': 1.0032438039779663, 'timing/data': 0.00020226999185979366, 'timing/model': 0.3662473870062968, 'learning_rate/qwen_vl_interface': 9.532241321138832e-06, 'learning_rate/action_model': 9.532241321138833e-05, 'epoch': 0.06}) +[2026-10-03 14:46:54,918][starVLA.training.train_starvla][INFO] - Step 799, Loss: {'train/loss': 0.0069246673956513405, 'train/grad_norm_pre_clip': 0.8082353472709656, 'timing/data': 0.0002867400035029277, 'timing/model': 0.3658015169930877, 'learning_rate/qwen_vl_interface': 9.530922601700451e-06, 'learning_rate/action_model': 9.530922601700452e-05, 'epoch': 0.06}) +[2026-10-03 14:46:55,568][starVLA.training.train_starvla][INFO] - Step 800, Loss: {'train/loss': 0.006923758424818516, 'train/grad_norm_pre_clip': 1.1467300653457642, 'timing/data': 0.00026046000130008906, 'timing/model': 0.36655028599489015, 'learning_rate/qwen_vl_interface': 9.52960212253649e-06, 'learning_rate/action_model': 9.529602122536491e-05, 'epoch': 0.06}) +[2026-10-03 14:46:56,422][starVLA.training.train_starvla][INFO] - Step 801, Loss: {'train/loss': 0.011793906800448895, 'train/grad_norm_pre_clip': 10.01273250579834, 'timing/data': 0.00024139999004546553, 'timing/model': 0.5691140920098405, 'learning_rate/qwen_vl_interface': 9.528279884189754e-06, 'learning_rate/action_model': 9.528279884189754e-05, 'epoch': 0.06}) +[2026-10-03 14:46:57,071][starVLA.training.train_starvla][INFO] - Step 802, Loss: {'train/loss': 0.0031008392106741667, 'train/grad_norm_pre_clip': 0.8791479468345642, 'timing/data': 0.00024231000861618668, 'timing/model': 0.365874797993456, 'learning_rate/qwen_vl_interface': 9.526955887203762e-06, 'learning_rate/action_model': 9.526955887203761e-05, 'epoch': 0.06}) +[2026-10-03 14:46:57,720][starVLA.training.train_starvla][INFO] - Step 803, Loss: {'train/loss': 0.23533660173416138, 'train/grad_norm_pre_clip': 57.52690505981445, 'timing/data': 0.000294529992970638, 'timing/model': 0.3659635259973584, 'learning_rate/qwen_vl_interface': 9.525630132122758e-06, 'learning_rate/action_model': 9.525630132122758e-05, 'epoch': 0.06}) +[2026-10-03 14:46:58,369][starVLA.training.train_starvla][INFO] - Step 804, Loss: {'train/loss': 0.11638425290584564, 'train/grad_norm_pre_clip': 78.3705062866211, 'timing/data': 0.00019643999985419214, 'timing/model': 0.3658268169965595, 'learning_rate/qwen_vl_interface': 9.524302619491712e-06, 'learning_rate/action_model': 9.524302619491712e-05, 'epoch': 0.06}) +[2026-10-03 14:46:59,018][starVLA.training.train_starvla][INFO] - Step 805, Loss: {'train/loss': 0.5357591509819031, 'train/grad_norm_pre_clip': 36.21294021606445, 'timing/data': 0.0002905600122176111, 'timing/model': 0.36509271699469537, 'learning_rate/qwen_vl_interface': 9.522973349856315e-06, 'learning_rate/action_model': 9.522973349856315e-05, 'epoch': 0.06}) +[2026-10-03 14:46:59,667][starVLA.training.train_starvla][INFO] - Step 806, Loss: {'train/loss': 0.0005915568908676505, 'train/grad_norm_pre_clip': 0.12935030460357666, 'timing/data': 0.000360630001523532, 'timing/model': 0.36552179700811394, 'learning_rate/qwen_vl_interface': 9.521642323762976e-06, 'learning_rate/action_model': 9.521642323762976e-05, 'epoch': 0.06}) +[2026-10-03 14:47:00,515][starVLA.training.train_starvla][INFO] - Step 807, Loss: {'train/loss': 1.040266513824463, 'train/grad_norm_pre_clip': 72.55941772460938, 'timing/data': 0.00031342000875156373, 'timing/model': 0.36574615599238314, 'learning_rate/qwen_vl_interface': 9.520309541758833e-06, 'learning_rate/action_model': 9.520309541758833e-05, 'epoch': 0.06}) +[2026-10-03 14:47:01,165][starVLA.training.train_starvla][INFO] - Step 808, Loss: {'train/loss': 0.0004491739091463387, 'train/grad_norm_pre_clip': 0.33216333389282227, 'timing/data': 0.0002340200007893145, 'timing/model': 0.365775466998457, 'learning_rate/qwen_vl_interface': 9.51897500439174e-06, 'learning_rate/action_model': 9.51897500439174e-05, 'epoch': 0.06}) +[2026-10-03 14:47:01,814][starVLA.training.train_starvla][INFO] - Step 809, Loss: {'train/loss': 0.0008727884851396084, 'train/grad_norm_pre_clip': 1.210956335067749, 'timing/data': 0.00031137898622546345, 'timing/model': 0.36517402798926923, 'learning_rate/qwen_vl_interface': 9.517638712210277e-06, 'learning_rate/action_model': 9.517638712210276e-05, 'epoch': 0.06}) +[2026-10-03 14:47:02,441][starVLA.training.train_starvla][INFO] - Step 810, Loss: {'train/loss': 0.45087558031082153, 'train/grad_norm_pre_clip': 12.44638442993164, 'timing/data': 0.00029936998907942325, 'timing/model': 0.36541825700260233, 'learning_rate/qwen_vl_interface': 9.51630066576374e-06, 'learning_rate/action_model': 9.51630066576374e-05, 'epoch': 0.06}) +[2026-10-03 14:47:03,088][starVLA.training.train_starvla][INFO] - Step 811, Loss: {'train/loss': 0.5988184213638306, 'train/grad_norm_pre_clip': 13.117082595825195, 'timing/data': 0.0002469099999871105, 'timing/model': 0.3644931180024287, 'learning_rate/qwen_vl_interface': 9.514960865602154e-06, 'learning_rate/action_model': 9.514960865602153e-05, 'epoch': 0.06}) +[2026-10-03 14:47:03,933][starVLA.training.train_starvla][INFO] - Step 812, Loss: {'train/loss': 0.004032261669635773, 'train/grad_norm_pre_clip': 11.062405586242676, 'timing/data': 0.0002584300091257319, 'timing/model': 0.5619920039898716, 'learning_rate/qwen_vl_interface': 9.513619312276259e-06, 'learning_rate/action_model': 9.513619312276258e-05, 'epoch': 0.06}) +[2026-10-03 14:47:04,580][starVLA.training.train_starvla][INFO] - Step 813, Loss: {'train/loss': 0.0005108960322104394, 'train/grad_norm_pre_clip': 0.07040036469697952, 'timing/data': 0.00029429000278469175, 'timing/model': 0.3642572270036908, 'learning_rate/qwen_vl_interface': 9.512276006337514e-06, 'learning_rate/action_model': 9.512276006337513e-05, 'epoch': 0.06}) +[2026-10-03 14:47:05,226][starVLA.training.train_starvla][INFO] - Step 814, Loss: {'train/loss': 0.430859237909317, 'train/grad_norm_pre_clip': 21.041780471801758, 'timing/data': 0.00019967000116594136, 'timing/model': 0.3645649170066463, 'learning_rate/qwen_vl_interface': 9.510930948338102e-06, 'learning_rate/action_model': 9.510930948338102e-05, 'epoch': 0.06}) +[2026-10-03 14:47:05,875][starVLA.training.train_starvla][INFO] - Step 815, Loss: {'train/loss': 0.013876066543161869, 'train/grad_norm_pre_clip': 54.21328353881836, 'timing/data': 0.00023623000015504658, 'timing/model': 0.3652396780089475, 'learning_rate/qwen_vl_interface': 9.509584138830929e-06, 'learning_rate/action_model': 9.509584138830928e-05, 'epoch': 0.06}) +[2026-10-03 14:47:06,523][starVLA.training.train_starvla][INFO] - Step 816, Loss: {'train/loss': 0.0021992933470755816, 'train/grad_norm_pre_clip': 0.4652022421360016, 'timing/data': 0.00019619999511633068, 'timing/model': 0.3651403179974295, 'learning_rate/qwen_vl_interface': 9.508235578369613e-06, 'learning_rate/action_model': 9.508235578369612e-05, 'epoch': 0.06}) +[2026-10-03 14:47:07,377][starVLA.training.train_starvla][INFO] - Step 817, Loss: {'train/loss': 0.3704974055290222, 'train/grad_norm_pre_clip': 13.403406143188477, 'timing/data': 0.0003959399909945205, 'timing/model': 0.5712724629993318, 'learning_rate/qwen_vl_interface': 9.506885267508498e-06, 'learning_rate/action_model': 9.506885267508498e-05, 'epoch': 0.06}) +[2026-10-03 14:47:08,023][starVLA.training.train_starvla][INFO] - Step 818, Loss: {'train/loss': 0.002868889132514596, 'train/grad_norm_pre_clip': 0.3908774256706238, 'timing/data': 0.00019520000205375254, 'timing/model': 0.3643014670087723, 'learning_rate/qwen_vl_interface': 9.505533206802648e-06, 'learning_rate/action_model': 9.505533206802647e-05, 'epoch': 0.06}) +[2026-10-03 14:47:08,671][starVLA.training.train_starvla][INFO] - Step 819, Loss: {'train/loss': 0.644860565662384, 'train/grad_norm_pre_clip': 25.504039764404297, 'timing/data': 0.00019865001377183944, 'timing/model': 0.36509438800567295, 'learning_rate/qwen_vl_interface': 9.50417939680784e-06, 'learning_rate/action_model': 9.504179396807839e-05, 'epoch': 0.06}) +[2026-10-03 14:47:09,319][starVLA.training.train_starvla][INFO] - Step 820, Loss: {'train/loss': 0.007185168098658323, 'train/grad_norm_pre_clip': 1.4783592224121094, 'timing/data': 0.00028594001196324825, 'timing/model': 0.36473764799302444, 'learning_rate/qwen_vl_interface': 9.502823838080578e-06, 'learning_rate/action_model': 9.502823838080578e-05, 'epoch': 0.06}) +[2026-10-03 14:47:09,966][starVLA.training.train_starvla][INFO] - Step 821, Loss: {'train/loss': 0.40284910798072815, 'train/grad_norm_pre_clip': 295.90667724609375, 'timing/data': 0.0003729599993675947, 'timing/model': 0.36421467798936646, 'learning_rate/qwen_vl_interface': 9.501466531178079e-06, 'learning_rate/action_model': 9.501466531178079e-05, 'epoch': 0.06}) +[2026-10-03 14:47:10,613][starVLA.training.train_starvla][INFO] - Step 822, Loss: {'train/loss': 0.14837615191936493, 'train/grad_norm_pre_clip': 41.86236572265625, 'timing/data': 0.0002450400061206892, 'timing/model': 0.3649944869976025, 'learning_rate/qwen_vl_interface': 9.50010747665828e-06, 'learning_rate/action_model': 9.500107476658281e-05, 'epoch': 0.06}) +[2026-10-03 14:47:11,460][starVLA.training.train_starvla][INFO] - Step 823, Loss: {'train/loss': 0.019736118614673615, 'train/grad_norm_pre_clip': 23.838720321655273, 'timing/data': 0.00027845999284181744, 'timing/model': 0.3647576270013815, 'learning_rate/qwen_vl_interface': 9.498746675079842e-06, 'learning_rate/action_model': 9.49874667507984e-05, 'epoch': 0.06}) +[2026-10-03 14:47:12,108][starVLA.training.train_starvla][INFO] - Step 824, Loss: {'train/loss': 0.00885260384529829, 'train/grad_norm_pre_clip': 3.00504207611084, 'timing/data': 0.00019931999850086868, 'timing/model': 0.3650596669904189, 'learning_rate/qwen_vl_interface': 9.49738412700213e-06, 'learning_rate/action_model': 9.49738412700213e-05, 'epoch': 0.06}) +[2026-10-03 14:47:12,756][starVLA.training.train_starvla][INFO] - Step 825, Loss: {'train/loss': 0.011658024042844772, 'train/grad_norm_pre_clip': 8.487895011901855, 'timing/data': 0.0001997600047616288, 'timing/model': 0.3646994279988576, 'learning_rate/qwen_vl_interface': 9.496019832985243e-06, 'learning_rate/action_model': 9.496019832985243e-05, 'epoch': 0.06}) +[2026-10-03 14:47:13,402][starVLA.training.train_starvla][INFO] - Step 826, Loss: {'train/loss': 0.022418254986405373, 'train/grad_norm_pre_clip': 9.811163902282715, 'timing/data': 0.00042830999882426113, 'timing/model': 0.3640709189930931, 'learning_rate/qwen_vl_interface': 9.49465379358999e-06, 'learning_rate/action_model': 9.49465379358999e-05, 'epoch': 0.06}) +[2026-10-03 14:47:14,049][starVLA.training.train_starvla][INFO] - Step 827, Loss: {'train/loss': 0.4240717589855194, 'train/grad_norm_pre_clip': 129.73077392578125, 'timing/data': 0.00019774000975303352, 'timing/model': 0.36475911700108554, 'learning_rate/qwen_vl_interface': 9.493286009377896e-06, 'learning_rate/action_model': 9.493286009377896e-05, 'epoch': 0.06}) +[2026-10-03 14:47:14,886][starVLA.training.train_starvla][INFO] - Step 828, Loss: {'train/loss': 0.002431547734886408, 'train/grad_norm_pre_clip': 0.2611757218837738, 'timing/data': 0.00019658000383060426, 'timing/model': 0.5540692450013012, 'learning_rate/qwen_vl_interface': 9.491916480911208e-06, 'learning_rate/action_model': 9.491916480911207e-05, 'epoch': 0.06}) +[2026-10-03 14:47:15,511][starVLA.training.train_starvla][INFO] - Step 829, Loss: {'train/loss': 0.0023582454305142164, 'train/grad_norm_pre_clip': 1.0657228231430054, 'timing/data': 0.0002000499953282997, 'timing/model': 0.3644345279899426, 'learning_rate/qwen_vl_interface': 9.490545208752884e-06, 'learning_rate/action_model': 9.490545208752883e-05, 'epoch': 0.06}) +[2026-10-03 14:47:16,158][starVLA.training.train_starvla][INFO] - Step 830, Loss: {'train/loss': 0.21884894371032715, 'train/grad_norm_pre_clip': 51.37520980834961, 'timing/data': 0.000242390000494197, 'timing/model': 0.36476840599789284, 'learning_rate/qwen_vl_interface': 9.489172193466603e-06, 'learning_rate/action_model': 9.489172193466603e-05, 'epoch': 0.06}) +[2026-10-03 14:47:16,805][starVLA.training.train_starvla][INFO] - Step 831, Loss: {'train/loss': 0.001194783253595233, 'train/grad_norm_pre_clip': 0.1299963891506195, 'timing/data': 0.00019956000323873013, 'timing/model': 0.3646911270043347, 'learning_rate/qwen_vl_interface': 9.487797435616762e-06, 'learning_rate/action_model': 9.487797435616761e-05, 'epoch': 0.06}) +[2026-10-03 14:47:17,431][starVLA.training.train_starvla][INFO] - Step 832, Loss: {'train/loss': 0.6961491107940674, 'train/grad_norm_pre_clip': 29.329160690307617, 'timing/data': 0.00023571000201627612, 'timing/model': 0.36471375699329656, 'learning_rate/qwen_vl_interface': 9.486420935768472e-06, 'learning_rate/action_model': 9.486420935768472e-05, 'epoch': 0.06}) +[2026-10-03 14:47:18,271][starVLA.training.train_starvla][INFO] - Step 833, Loss: {'train/loss': 0.001116044819355011, 'train/grad_norm_pre_clip': 0.1237834095954895, 'timing/data': 0.00027816000510938466, 'timing/model': 0.557015013997443, 'learning_rate/qwen_vl_interface': 9.485042694487558e-06, 'learning_rate/action_model': 9.485042694487559e-05, 'epoch': 0.06}) +[2026-10-03 14:47:18,918][starVLA.training.train_starvla][INFO] - Step 834, Loss: {'train/loss': 0.000844296533614397, 'train/grad_norm_pre_clip': 0.09710324555635452, 'timing/data': 0.00029074000485707074, 'timing/model': 0.36468000599415973, 'learning_rate/qwen_vl_interface': 9.483662712340565e-06, 'learning_rate/action_model': 9.483662712340564e-05, 'epoch': 0.06}) +[2026-10-03 14:47:19,566][starVLA.training.train_starvla][INFO] - Step 835, Loss: {'train/loss': 0.0009959901217371225, 'train/grad_norm_pre_clip': 0.10468113422393799, 'timing/data': 0.00030277999758254737, 'timing/model': 0.36463397700572386, 'learning_rate/qwen_vl_interface': 9.482280989894748e-06, 'learning_rate/action_model': 9.482280989894748e-05, 'epoch': 0.06}) +[2026-10-03 14:47:20,213][starVLA.training.train_starvla][INFO] - Step 836, Loss: {'train/loss': 0.030861245468258858, 'train/grad_norm_pre_clip': 41.85261917114258, 'timing/data': 0.00024884000595193356, 'timing/model': 0.3646203870011959, 'learning_rate/qwen_vl_interface': 9.480897527718083e-06, 'learning_rate/action_model': 9.480897527718084e-05, 'epoch': 0.06}) +[2026-10-03 14:47:20,860][starVLA.training.train_starvla][INFO] - Step 837, Loss: {'train/loss': 0.035947252064943314, 'train/grad_norm_pre_clip': 61.246978759765625, 'timing/data': 0.00025661999825388193, 'timing/model': 0.36420296800497454, 'learning_rate/qwen_vl_interface': 9.479512326379263e-06, 'learning_rate/action_model': 9.479512326379264e-05, 'epoch': 0.06}) +[2026-10-03 14:47:21,485][starVLA.training.train_starvla][INFO] - Step 838, Loss: {'train/loss': 0.0005851808236911893, 'train/grad_norm_pre_clip': 0.06028475612401962, 'timing/data': 0.0003465799964033067, 'timing/model': 0.36428750700724777, 'learning_rate/qwen_vl_interface': 9.478125386447686e-06, 'learning_rate/action_model': 9.478125386447686e-05, 'epoch': 0.06}) +[2026-10-03 14:47:22,338][starVLA.training.train_starvla][INFO] - Step 839, Loss: {'train/loss': 0.7528854608535767, 'train/grad_norm_pre_clip': 30.809974670410156, 'timing/data': 0.0003672699967864901, 'timing/model': 0.3645525070023723, 'learning_rate/qwen_vl_interface': 9.476736708493472e-06, 'learning_rate/action_model': 9.476736708493472e-05, 'epoch': 0.06}) +[2026-10-03 14:47:22,985][starVLA.training.train_starvla][INFO] - Step 840, Loss: {'train/loss': 0.0014463231200352311, 'train/grad_norm_pre_clip': 0.24743176996707916, 'timing/data': 0.00023965998843777925, 'timing/model': 0.3649674569896888, 'learning_rate/qwen_vl_interface': 9.475346293087457e-06, 'learning_rate/action_model': 9.475346293087458e-05, 'epoch': 0.06}) +[2026-10-03 14:47:23,631][starVLA.training.train_starvla][INFO] - Step 841, Loss: {'train/loss': 0.0032719222363084555, 'train/grad_norm_pre_clip': 4.383364677429199, 'timing/data': 0.0002368099958403036, 'timing/model': 0.36440701699757483, 'learning_rate/qwen_vl_interface': 9.473954140801184e-06, 'learning_rate/action_model': 9.473954140801185e-05, 'epoch': 0.06}) +[2026-10-03 14:47:24,257][starVLA.training.train_starvla][INFO] - Step 842, Loss: {'train/loss': 0.007541160099208355, 'train/grad_norm_pre_clip': 17.690061569213867, 'timing/data': 0.00027362999389879405, 'timing/model': 0.364130717003718, 'learning_rate/qwen_vl_interface': 9.472560252206918e-06, 'learning_rate/action_model': 9.472560252206918e-05, 'epoch': 0.06}) +[2026-10-03 14:47:24,905][starVLA.training.train_starvla][INFO] - Step 843, Loss: {'train/loss': 0.015289629809558392, 'train/grad_norm_pre_clip': 29.96164894104004, 'timing/data': 0.00019500000053085387, 'timing/model': 0.36534774699248374, 'learning_rate/qwen_vl_interface': 9.471164627877633e-06, 'learning_rate/action_model': 9.471164627877633e-05, 'epoch': 0.06}) +[2026-10-03 14:47:25,753][starVLA.training.train_starvla][INFO] - Step 844, Loss: {'train/loss': 0.03629445284605026, 'train/grad_norm_pre_clip': 164.64854431152344, 'timing/data': 0.00030902000435162336, 'timing/model': 0.3652472069952637, 'learning_rate/qwen_vl_interface': 9.469767268387016e-06, 'learning_rate/action_model': 9.469767268387016e-05, 'epoch': 0.06}) +[2026-10-03 14:47:26,400][starVLA.training.train_starvla][INFO] - Step 845, Loss: {'train/loss': 1.488964319229126, 'train/grad_norm_pre_clip': 207.07061767578125, 'timing/data': 0.00031755000236444175, 'timing/model': 0.36422620600205846, 'learning_rate/qwen_vl_interface': 9.468368174309468e-06, 'learning_rate/action_model': 9.468368174309469e-05, 'epoch': 0.06}) +[2026-10-03 14:47:27,048][starVLA.training.train_starvla][INFO] - Step 846, Loss: {'train/loss': 0.9692286849021912, 'train/grad_norm_pre_clip': 162.2580108642578, 'timing/data': 0.0003146599920000881, 'timing/model': 0.36495463699975517, 'learning_rate/qwen_vl_interface': 9.466967346220107e-06, 'learning_rate/action_model': 9.466967346220107e-05, 'epoch': 0.06}) +[2026-10-03 14:47:27,696][starVLA.training.train_starvla][INFO] - Step 847, Loss: {'train/loss': 0.5332226753234863, 'train/grad_norm_pre_clip': 15.178141593933105, 'timing/data': 0.0002476499939803034, 'timing/model': 0.36525169799278956, 'learning_rate/qwen_vl_interface': 9.465564784694759e-06, 'learning_rate/action_model': 9.465564784694759e-05, 'epoch': 0.06}) +[2026-10-03 14:47:28,344][starVLA.training.train_starvla][INFO] - Step 848, Loss: {'train/loss': 0.002337796613574028, 'train/grad_norm_pre_clip': 0.9768239259719849, 'timing/data': 0.00031537999166175723, 'timing/model': 0.3651480570115382, 'learning_rate/qwen_vl_interface': 9.464160490309962e-06, 'learning_rate/action_model': 9.464160490309962e-05, 'epoch': 0.06}) +[2026-10-03 14:47:29,189][starVLA.training.train_starvla][INFO] - Step 849, Loss: {'train/loss': 1.4895399808883667, 'train/grad_norm_pre_clip': 52.07878494262695, 'timing/data': 0.00019555000471882522, 'timing/model': 0.5615586339990841, 'learning_rate/qwen_vl_interface': 9.462754463642972e-06, 'learning_rate/action_model': 9.46275446364297e-05, 'epoch': 0.06}) +[2026-10-03 14:47:29,837][starVLA.training.train_starvla][INFO] - Step 850, Loss: {'train/loss': 0.0005090279155410826, 'train/grad_norm_pre_clip': 0.06415557116270065, 'timing/data': 0.00020034000044688582, 'timing/model': 0.36535091800033115, 'learning_rate/qwen_vl_interface': 9.461346705271753e-06, 'learning_rate/action_model': 9.461346705271752e-05, 'epoch': 0.06}) +[2026-10-03 14:47:30,485][starVLA.training.train_starvla][INFO] - Step 851, Loss: {'train/loss': 0.0034247967414557934, 'train/grad_norm_pre_clip': 6.913843631744385, 'timing/data': 0.00019718000839930028, 'timing/model': 0.36529241799144074, 'learning_rate/qwen_vl_interface': 9.459937215774976e-06, 'learning_rate/action_model': 9.459937215774976e-05, 'epoch': 0.06}) +[2026-10-03 14:47:31,133][starVLA.training.train_starvla][INFO] - Step 852, Loss: {'train/loss': 0.7443920373916626, 'train/grad_norm_pre_clip': 63.096031188964844, 'timing/data': 0.0001968300057342276, 'timing/model': 0.3649366580066271, 'learning_rate/qwen_vl_interface': 9.458525995732035e-06, 'learning_rate/action_model': 9.458525995732035e-05, 'epoch': 0.06}) +[2026-10-03 14:47:31,781][starVLA.training.train_starvla][INFO] - Step 853, Loss: {'train/loss': 0.5729342699050903, 'train/grad_norm_pre_clip': 38.59847640991211, 'timing/data': 0.0002380399964749813, 'timing/model': 0.36452588700922206, 'learning_rate/qwen_vl_interface': 9.457113045723027e-06, 'learning_rate/action_model': 9.457113045723026e-05, 'epoch': 0.06}) +[2026-10-03 14:47:32,629][starVLA.training.train_starvla][INFO] - Step 854, Loss: {'train/loss': 0.25948476791381836, 'train/grad_norm_pre_clip': 24.25652503967285, 'timing/data': 0.00020163999579381198, 'timing/model': 0.5649441839923384, 'learning_rate/qwen_vl_interface': 9.455698366328762e-06, 'learning_rate/action_model': 9.455698366328762e-05, 'epoch': 0.06}) +[2026-10-03 14:47:33,275][starVLA.training.train_starvla][INFO] - Step 855, Loss: {'train/loss': 0.0040140580385923386, 'train/grad_norm_pre_clip': 0.6705502867698669, 'timing/data': 0.00030096000409685075, 'timing/model': 0.3645783469983144, 'learning_rate/qwen_vl_interface': 9.454281958130761e-06, 'learning_rate/action_model': 9.454281958130761e-05, 'epoch': 0.06}) +[2026-10-03 14:47:33,923][starVLA.training.train_starvla][INFO] - Step 856, Loss: {'train/loss': 0.004156458657234907, 'train/grad_norm_pre_clip': 0.5954177379608154, 'timing/data': 0.0001999200030695647, 'timing/model': 0.3644265970069682, 'learning_rate/qwen_vl_interface': 9.452863821711257e-06, 'learning_rate/action_model': 9.452863821711258e-05, 'epoch': 0.06}) +[2026-10-03 14:47:34,570][starVLA.training.train_starvla][INFO] - Step 857, Loss: {'train/loss': 0.4462089538574219, 'train/grad_norm_pre_clip': 51.727027893066406, 'timing/data': 0.00041978899389505386, 'timing/model': 0.36398148800071795, 'learning_rate/qwen_vl_interface': 9.451443957653195e-06, 'learning_rate/action_model': 9.451443957653195e-05, 'epoch': 0.06}) +[2026-10-03 14:47:35,217][starVLA.training.train_starvla][INFO] - Step 858, Loss: {'train/loss': 0.005609662272036076, 'train/grad_norm_pre_clip': 2.5138556957244873, 'timing/data': 0.00031308000325225294, 'timing/model': 0.36470945700421, 'learning_rate/qwen_vl_interface': 9.450022366540221e-06, 'learning_rate/action_model': 9.45002236654022e-05, 'epoch': 0.06}) +[2026-10-03 14:47:35,865][starVLA.training.train_starvla][INFO] - Step 859, Loss: {'train/loss': 0.5800915956497192, 'train/grad_norm_pre_clip': 33.0374755859375, 'timing/data': 0.00025271000049542636, 'timing/model': 0.36525794700719416, 'learning_rate/qwen_vl_interface': 9.448599048956702e-06, 'learning_rate/action_model': 9.448599048956702e-05, 'epoch': 0.06}) +[2026-10-03 14:47:36,717][starVLA.training.train_starvla][INFO] - Step 860, Loss: {'train/loss': 0.0020365959499031305, 'train/grad_norm_pre_clip': 1.1262620687484741, 'timing/data': 0.0002906000008806586, 'timing/model': 0.3645731260039611, 'learning_rate/qwen_vl_interface': 9.447174005487708e-06, 'learning_rate/action_model': 9.447174005487708e-05, 'epoch': 0.06}) +[2026-10-03 14:47:37,364][starVLA.training.train_starvla][INFO] - Step 861, Loss: {'train/loss': 0.0006907952483743429, 'train/grad_norm_pre_clip': 0.123216912150383, 'timing/data': 0.0002399500081082806, 'timing/model': 0.3644350169925019, 'learning_rate/qwen_vl_interface': 9.445747236719022e-06, 'learning_rate/action_model': 9.445747236719022e-05, 'epoch': 0.06}) +[2026-10-03 14:47:38,011][starVLA.training.train_starvla][INFO] - Step 862, Loss: {'train/loss': 0.3623647391796112, 'train/grad_norm_pre_clip': 102.41654205322266, 'timing/data': 0.0003112700069323182, 'timing/model': 0.36443051700189244, 'learning_rate/qwen_vl_interface': 9.444318743237135e-06, 'learning_rate/action_model': 9.444318743237134e-05, 'epoch': 0.06}) +[2026-10-03 14:47:38,658][starVLA.training.train_starvla][INFO] - Step 863, Loss: {'train/loss': 0.586999237537384, 'train/grad_norm_pre_clip': 125.11366271972656, 'timing/data': 0.00028332001238595694, 'timing/model': 0.3645091579965083, 'learning_rate/qwen_vl_interface': 9.442888525629244e-06, 'learning_rate/action_model': 9.442888525629245e-05, 'epoch': 0.06}) +[2026-10-03 14:47:39,306][starVLA.training.train_starvla][INFO] - Step 864, Loss: {'train/loss': 0.29126283526420593, 'train/grad_norm_pre_clip': 21.030868530273438, 'timing/data': 0.0002754699962679297, 'timing/model': 0.3648485379962949, 'learning_rate/qwen_vl_interface': 9.441456584483261e-06, 'learning_rate/action_model': 9.441456584483261e-05, 'epoch': 0.06}) +[2026-10-03 14:47:40,140][starVLA.training.train_starvla][INFO] - Step 865, Loss: {'train/loss': 0.6023144721984863, 'train/grad_norm_pre_clip': 51.53408432006836, 'timing/data': 0.0004045200039399788, 'timing/model': 0.5514763750106795, 'learning_rate/qwen_vl_interface': 9.440022920387801e-06, 'learning_rate/action_model': 9.440022920387801e-05, 'epoch': 0.06}) +[2026-10-03 14:47:40,788][starVLA.training.train_starvla][INFO] - Step 866, Loss: {'train/loss': 0.28707316517829895, 'train/grad_norm_pre_clip': 24.090259552001953, 'timing/data': 0.00023159000556915998, 'timing/model': 0.3648236970038852, 'learning_rate/qwen_vl_interface': 9.438587533932191e-06, 'learning_rate/action_model': 9.43858753393219e-05, 'epoch': 0.06}) +[2026-10-03 14:47:41,435][starVLA.training.train_starvla][INFO] - Step 867, Loss: {'train/loss': 0.10610109567642212, 'train/grad_norm_pre_clip': 17.540834426879883, 'timing/data': 0.0002903099957620725, 'timing/model': 0.36468424800841603, 'learning_rate/qwen_vl_interface': 9.43715042570646e-06, 'learning_rate/action_model': 9.43715042570646e-05, 'epoch': 0.06}) +[2026-10-03 14:47:42,082][starVLA.training.train_starvla][INFO] - Step 868, Loss: {'train/loss': 0.17795400321483612, 'train/grad_norm_pre_clip': 43.414642333984375, 'timing/data': 0.00032654999813530594, 'timing/model': 0.36489233700558543, 'learning_rate/qwen_vl_interface': 9.435711596301353e-06, 'learning_rate/action_model': 9.435711596301352e-05, 'epoch': 0.06}) +[2026-10-03 14:47:42,728][starVLA.training.train_starvla][INFO] - Step 869, Loss: {'train/loss': 0.15857097506523132, 'train/grad_norm_pre_clip': 15.335685729980469, 'timing/data': 0.00024699000641703606, 'timing/model': 0.36347353699966334, 'learning_rate/qwen_vl_interface': 9.434271046308317e-06, 'learning_rate/action_model': 9.434271046308317e-05, 'epoch': 0.06}) +[2026-10-03 14:47:43,544][starVLA.training.train_starvla][INFO] - Step 870, Loss: {'train/loss': 0.08744101226329803, 'train/grad_norm_pre_clip': 25.345256805419922, 'timing/data': 0.00036251000710763037, 'timing/model': 0.5555730939959176, 'learning_rate/qwen_vl_interface': 9.432828776319507e-06, 'learning_rate/action_model': 9.432828776319505e-05, 'epoch': 0.06}) +[2026-10-03 14:47:44,190][starVLA.training.train_starvla][INFO] - Step 871, Loss: {'train/loss': 0.11404518783092499, 'train/grad_norm_pre_clip': 129.6183624267578, 'timing/data': 0.00020299000607337803, 'timing/model': 0.36442686700320337, 'learning_rate/qwen_vl_interface': 9.431384786927785e-06, 'learning_rate/action_model': 9.431384786927786e-05, 'epoch': 0.06}) +[2026-10-03 14:47:44,838][starVLA.training.train_starvla][INFO] - Step 872, Loss: {'train/loss': 0.005888528190553188, 'train/grad_norm_pre_clip': 1.9294620752334595, 'timing/data': 0.0003290799941169098, 'timing/model': 0.3648105259926524, 'learning_rate/qwen_vl_interface': 9.429939078726723e-06, 'learning_rate/action_model': 9.429939078726722e-05, 'epoch': 0.06}) +[2026-10-03 14:47:45,485][starVLA.training.train_starvla][INFO] - Step 873, Loss: {'train/loss': 0.10259848088026047, 'train/grad_norm_pre_clip': 24.459381103515625, 'timing/data': 0.00028623000252991915, 'timing/model': 0.3646913470001891, 'learning_rate/qwen_vl_interface': 9.428491652310598e-06, 'learning_rate/action_model': 9.428491652310597e-05, 'epoch': 0.06}) +[2026-10-03 14:47:46,132][starVLA.training.train_starvla][INFO] - Step 874, Loss: {'train/loss': 0.5440885424613953, 'train/grad_norm_pre_clip': 62.74118423461914, 'timing/data': 0.0003016900009242818, 'timing/model': 0.36468032700940967, 'learning_rate/qwen_vl_interface': 9.427042508274386e-06, 'learning_rate/action_model': 9.427042508274386e-05, 'epoch': 0.06}) +[2026-10-03 14:47:46,779][starVLA.training.train_starvla][INFO] - Step 875, Loss: {'train/loss': 0.0030461130663752556, 'train/grad_norm_pre_clip': 0.7926054000854492, 'timing/data': 0.00019736999820452183, 'timing/model': 0.3650020659988513, 'learning_rate/qwen_vl_interface': 9.425591647213783e-06, 'learning_rate/action_model': 9.425591647213782e-05, 'epoch': 0.06}) +[2026-10-03 14:47:47,628][starVLA.training.train_starvla][INFO] - Step 876, Loss: {'train/loss': 0.6935167908668518, 'train/grad_norm_pre_clip': 13.306660652160645, 'timing/data': 0.00023530999897047877, 'timing/model': 0.3649213169992436, 'learning_rate/qwen_vl_interface': 9.424139069725177e-06, 'learning_rate/action_model': 9.424139069725178e-05, 'epoch': 0.06}) +[2026-10-03 14:47:48,275][starVLA.training.train_starvla][INFO] - Step 877, Loss: {'train/loss': 0.04950939863920212, 'train/grad_norm_pre_clip': 48.65880584716797, 'timing/data': 0.00020080000103916973, 'timing/model': 0.3643064960051561, 'learning_rate/qwen_vl_interface': 9.422684776405673e-06, 'learning_rate/action_model': 9.422684776405672e-05, 'epoch': 0.06}) +[2026-10-03 14:47:48,922][starVLA.training.train_starvla][INFO] - Step 878, Loss: {'train/loss': 0.019751060754060745, 'train/grad_norm_pre_clip': 7.23075008392334, 'timing/data': 0.0003083299961872399, 'timing/model': 0.3643467969959602, 'learning_rate/qwen_vl_interface': 9.421228767853072e-06, 'learning_rate/action_model': 9.421228767853072e-05, 'epoch': 0.06}) +[2026-10-03 14:47:49,569][starVLA.training.train_starvla][INFO] - Step 879, Loss: {'train/loss': 0.007981985807418823, 'train/grad_norm_pre_clip': 6.124813556671143, 'timing/data': 0.00024460999702569097, 'timing/model': 0.3649642069940455, 'learning_rate/qwen_vl_interface': 9.419771044665887e-06, 'learning_rate/action_model': 9.419771044665887e-05, 'epoch': 0.06}) +[2026-10-03 14:47:50,216][starVLA.training.train_starvla][INFO] - Step 880, Loss: {'train/loss': 0.11279242485761642, 'train/grad_norm_pre_clip': 42.67588424682617, 'timing/data': 0.00036081999132875353, 'timing/model': 0.36431422700115945, 'learning_rate/qwen_vl_interface': 9.41831160744333e-06, 'learning_rate/action_model': 9.41831160744333e-05, 'epoch': 0.06}) +[2026-10-03 14:47:51,063][starVLA.training.train_starvla][INFO] - Step 881, Loss: {'train/loss': 0.07125259190797806, 'train/grad_norm_pre_clip': 60.867576599121094, 'timing/data': 0.0004902000073343515, 'timing/model': 0.36441621799895074, 'learning_rate/qwen_vl_interface': 9.416850456785322e-06, 'learning_rate/action_model': 9.416850456785321e-05, 'epoch': 0.06}) +[2026-10-03 14:47:51,709][starVLA.training.train_starvla][INFO] - Step 882, Loss: {'train/loss': 0.34136155247688293, 'train/grad_norm_pre_clip': 66.82772064208984, 'timing/data': 0.0001993689947994426, 'timing/model': 0.36427199801255483, 'learning_rate/qwen_vl_interface': 9.415387593292489e-06, 'learning_rate/action_model': 9.415387593292488e-05, 'epoch': 0.06}) +[2026-10-03 14:47:52,356][starVLA.training.train_starvla][INFO] - Step 883, Loss: {'train/loss': 0.2618807852268219, 'train/grad_norm_pre_clip': 17.841310501098633, 'timing/data': 0.0003399800043553114, 'timing/model': 0.36469126598967705, 'learning_rate/qwen_vl_interface': 9.413923017566155e-06, 'learning_rate/action_model': 9.413923017566155e-05, 'epoch': 0.06}) +[2026-10-03 14:47:53,004][starVLA.training.train_starvla][INFO] - Step 884, Loss: {'train/loss': 0.6008263826370239, 'train/grad_norm_pre_clip': 44.72724533081055, 'timing/data': 0.0003156489983666688, 'timing/model': 0.364333017001627, 'learning_rate/qwen_vl_interface': 9.412456730208352e-06, 'learning_rate/action_model': 9.412456730208352e-05, 'epoch': 0.06}) +[2026-10-03 14:47:53,650][starVLA.training.train_starvla][INFO] - Step 885, Loss: {'train/loss': 0.10264379531145096, 'train/grad_norm_pre_clip': 69.21337890625, 'timing/data': 0.00021215999731794, 'timing/model': 0.3643532569985837, 'learning_rate/qwen_vl_interface': 9.41098873182182e-06, 'learning_rate/action_model': 9.41098873182182e-05, 'epoch': 0.06}) +[2026-10-03 14:47:54,490][starVLA.training.train_starvla][INFO] - Step 886, Loss: {'train/loss': 0.07219535112380981, 'train/grad_norm_pre_clip': 47.142765045166016, 'timing/data': 0.00044180000259075314, 'timing/model': 0.5564420739974594, 'learning_rate/qwen_vl_interface': 9.409519023009993e-06, 'learning_rate/action_model': 9.409519023009993e-05, 'epoch': 0.06}) +[2026-10-03 14:47:55,138][starVLA.training.train_starvla][INFO] - Step 887, Loss: {'train/loss': 0.19428962469100952, 'train/grad_norm_pre_clip': 104.50347900390625, 'timing/data': 0.00020122001296840608, 'timing/model': 0.36517119700147305, 'learning_rate/qwen_vl_interface': 9.408047604377014e-06, 'learning_rate/action_model': 9.408047604377014e-05, 'epoch': 0.06}) +[2026-10-03 14:47:55,764][starVLA.training.train_starvla][INFO] - Step 888, Loss: {'train/loss': 0.009334997273981571, 'train/grad_norm_pre_clip': 4.8723883628845215, 'timing/data': 0.0003283400001237169, 'timing/model': 0.34256887999072205, 'learning_rate/qwen_vl_interface': 9.40657447652773e-06, 'learning_rate/action_model': 9.406574476527731e-05, 'epoch': 0.06}) +[2026-10-03 14:47:56,411][starVLA.training.train_starvla][INFO] - Step 889, Loss: {'train/loss': 0.00917457789182663, 'train/grad_norm_pre_clip': 8.285632133483887, 'timing/data': 0.0001974299957510084, 'timing/model': 0.36516192799899727, 'learning_rate/qwen_vl_interface': 9.405099640067685e-06, 'learning_rate/action_model': 9.405099640067685e-05, 'epoch': 0.07}) +[2026-10-03 14:47:57,058][starVLA.training.train_starvla][INFO] - Step 890, Loss: {'train/loss': 0.018314208835363388, 'train/grad_norm_pre_clip': 13.23266887664795, 'timing/data': 0.0003467799979262054, 'timing/model': 0.3649486859940225, 'learning_rate/qwen_vl_interface': 9.40362309560313e-06, 'learning_rate/action_model': 9.40362309560313e-05, 'epoch': 0.07}) +[2026-10-03 14:47:57,876][starVLA.training.train_starvla][INFO] - Step 891, Loss: {'train/loss': 0.0037441791500896215, 'train/grad_norm_pre_clip': 0.7623031139373779, 'timing/data': 0.000301220003166236, 'timing/model': 0.5570353340008296, 'learning_rate/qwen_vl_interface': 9.402144843741016e-06, 'learning_rate/action_model': 9.402144843741017e-05, 'epoch': 0.07}) +[2026-10-03 14:47:58,525][starVLA.training.train_starvla][INFO] - Step 892, Loss: {'train/loss': 0.0018231747671961784, 'train/grad_norm_pre_clip': 0.19804903864860535, 'timing/data': 0.00019407000218052417, 'timing/model': 0.36526302600395866, 'learning_rate/qwen_vl_interface': 9.400664885088999e-06, 'learning_rate/action_model': 9.400664885088999e-05, 'epoch': 0.07}) +[2026-10-03 14:47:59,173][starVLA.training.train_starvla][INFO] - Step 893, Loss: {'train/loss': 0.7570635080337524, 'train/grad_norm_pre_clip': 26.80179214477539, 'timing/data': 0.00032002999796532094, 'timing/model': 0.3645876669907011, 'learning_rate/qwen_vl_interface': 9.39918322025543e-06, 'learning_rate/action_model': 9.39918322025543e-05, 'epoch': 0.07}) +[2026-10-03 14:47:59,798][starVLA.training.train_starvla][INFO] - Step 894, Loss: {'train/loss': 0.0011748264078050852, 'train/grad_norm_pre_clip': 0.1180139109492302, 'timing/data': 0.0003327300073578954, 'timing/model': 0.34203379999962635, 'learning_rate/qwen_vl_interface': 9.397699849849368e-06, 'learning_rate/action_model': 9.397699849849367e-05, 'epoch': 0.07}) +[2026-10-03 14:48:00,446][starVLA.training.train_starvla][INFO] - Step 895, Loss: {'train/loss': 0.5872862935066223, 'train/grad_norm_pre_clip': 11.91258716583252, 'timing/data': 0.0001971400051843375, 'timing/model': 0.36565574699488934, 'learning_rate/qwen_vl_interface': 9.39621477448057e-06, 'learning_rate/action_model': 9.39621477448057e-05, 'epoch': 0.07}) +[2026-10-03 14:48:01,096][starVLA.training.train_starvla][INFO] - Step 896, Loss: {'train/loss': 0.678816556930542, 'train/grad_norm_pre_clip': 35.268829345703125, 'timing/data': 0.00021853999351151288, 'timing/model': 0.36561045798589475, 'learning_rate/qwen_vl_interface': 9.394727994759495e-06, 'learning_rate/action_model': 9.394727994759495e-05, 'epoch': 0.07}) +[2026-10-03 14:48:01,940][starVLA.training.train_starvla][INFO] - Step 897, Loss: {'train/loss': 0.28697678446769714, 'train/grad_norm_pre_clip': 103.35176849365234, 'timing/data': 0.000206069991691038, 'timing/model': 0.36458669700368773, 'learning_rate/qwen_vl_interface': 9.3932395112973e-06, 'learning_rate/action_model': 9.3932395112973e-05, 'epoch': 0.07}) +[2026-10-03 14:48:02,588][starVLA.training.train_starvla][INFO] - Step 898, Loss: {'train/loss': 0.1357761174440384, 'train/grad_norm_pre_clip': 240.20123291015625, 'timing/data': 0.00030375999631360173, 'timing/model': 0.3650217270042049, 'learning_rate/qwen_vl_interface': 9.391749324705846e-06, 'learning_rate/action_model': 9.391749324705846e-05, 'epoch': 0.07}) +[2026-10-03 14:48:03,236][starVLA.training.train_starvla][INFO] - Step 899, Loss: {'train/loss': 0.2161651849746704, 'train/grad_norm_pre_clip': 70.29589080810547, 'timing/data': 0.00031687998853158206, 'timing/model': 0.36490401600894984, 'learning_rate/qwen_vl_interface': 9.390257435597692e-06, 'learning_rate/action_model': 9.390257435597692e-05, 'epoch': 0.07}) +[2026-10-03 14:48:03,884][starVLA.training.train_starvla][INFO] - Step 900, Loss: {'train/loss': 0.00335371564142406, 'train/grad_norm_pre_clip': 0.4032644033432007, 'timing/data': 0.0003279899974586442, 'timing/model': 0.36511596699710935, 'learning_rate/qwen_vl_interface': 9.3887638445861e-06, 'learning_rate/action_model': 9.3887638445861e-05, 'epoch': 0.07}) +[2026-10-03 14:48:04,531][starVLA.training.train_starvla][INFO] - Step 901, Loss: {'train/loss': 0.005368705838918686, 'train/grad_norm_pre_clip': 2.863166332244873, 'timing/data': 0.0001965300034498796, 'timing/model': 0.3644914769975003, 'learning_rate/qwen_vl_interface': 9.387268552285024e-06, 'learning_rate/action_model': 9.387268552285025e-05, 'epoch': 0.07}) +[2026-10-03 14:48:05,368][starVLA.training.train_starvla][INFO] - Step 902, Loss: {'train/loss': 0.005628710146993399, 'train/grad_norm_pre_clip': 1.0916861295700073, 'timing/data': 0.0001973000034922734, 'timing/model': 0.5550427349953679, 'learning_rate/qwen_vl_interface': 9.385771559309125e-06, 'learning_rate/action_model': 9.385771559309125e-05, 'epoch': 0.07}) +[2026-10-03 14:48:06,016][starVLA.training.train_starvla][INFO] - Step 903, Loss: {'train/loss': 0.14390835165977478, 'train/grad_norm_pre_clip': 158.04049682617188, 'timing/data': 0.0002008500014198944, 'timing/model': 0.3650879179913318, 'learning_rate/qwen_vl_interface': 9.38427286627376e-06, 'learning_rate/action_model': 9.38427286627376e-05, 'epoch': 0.07}) +[2026-10-03 14:48:06,663][starVLA.training.train_starvla][INFO] - Step 904, Loss: {'train/loss': 0.002103186910971999, 'train/grad_norm_pre_clip': 0.22046327590942383, 'timing/data': 0.00020383000082802027, 'timing/model': 0.3646876379934838, 'learning_rate/qwen_vl_interface': 9.382772473794985e-06, 'learning_rate/action_model': 9.382772473794985e-05, 'epoch': 0.07}) +[2026-10-03 14:48:07,309][starVLA.training.train_starvla][INFO] - Step 905, Loss: {'train/loss': 0.04576056823134422, 'train/grad_norm_pre_clip': 49.76847839355469, 'timing/data': 0.0002516899985494092, 'timing/model': 0.36402074599754997, 'learning_rate/qwen_vl_interface': 9.381270382489554e-06, 'learning_rate/action_model': 9.381270382489554e-05, 'epoch': 0.07}) +[2026-10-03 14:48:07,955][starVLA.training.train_starvla][INFO] - Step 906, Loss: {'train/loss': 0.09303493797779083, 'train/grad_norm_pre_clip': 66.00017547607422, 'timing/data': 0.00029397000616881996, 'timing/model': 0.3639284380042227, 'learning_rate/qwen_vl_interface': 9.37976659297492e-06, 'learning_rate/action_model': 9.379766592974921e-05, 'epoch': 0.07}) +[2026-10-03 14:48:08,771][starVLA.training.train_starvla][INFO] - Step 907, Loss: {'train/loss': 0.5904206037521362, 'train/grad_norm_pre_clip': 70.94515228271484, 'timing/data': 0.00020086999575141817, 'timing/model': 0.5327441569970688, 'learning_rate/qwen_vl_interface': 9.378261105869237e-06, 'learning_rate/action_model': 9.378261105869237e-05, 'epoch': 0.07}) +[2026-10-03 14:48:09,419][starVLA.training.train_starvla][INFO] - Step 908, Loss: {'train/loss': 0.0014139680424705148, 'train/grad_norm_pre_clip': 0.1453651636838913, 'timing/data': 0.00019008999515790492, 'timing/model': 0.3648387479915982, 'learning_rate/qwen_vl_interface': 9.376753921791348e-06, 'learning_rate/action_model': 9.376753921791348e-05, 'epoch': 0.07}) +[2026-10-03 14:48:10,065][starVLA.training.train_starvla][INFO] - Step 909, Loss: {'train/loss': 0.04693064093589783, 'train/grad_norm_pre_clip': 18.526453018188477, 'timing/data': 0.0002839090011548251, 'timing/model': 0.3636138080037199, 'learning_rate/qwen_vl_interface': 9.375245041360805e-06, 'learning_rate/action_model': 9.375245041360804e-05, 'epoch': 0.07}) +[2026-10-03 14:48:10,688][starVLA.training.train_starvla][INFO] - Step 910, Loss: {'train/loss': 0.133357971906662, 'train/grad_norm_pre_clip': 156.72415161132812, 'timing/data': 0.0003352499916218221, 'timing/model': 0.3414604099962162, 'learning_rate/qwen_vl_interface': 9.37373446519785e-06, 'learning_rate/action_model': 9.373734465197849e-05, 'epoch': 0.07}) +[2026-10-03 14:48:11,335][starVLA.training.train_starvla][INFO] - Step 911, Loss: {'train/loss': 0.0010239656548947096, 'train/grad_norm_pre_clip': 0.2654534578323364, 'timing/data': 0.00028464999923016876, 'timing/model': 0.36454925600264687, 'learning_rate/qwen_vl_interface': 9.372222193923421e-06, 'learning_rate/action_model': 9.372222193923421e-05, 'epoch': 0.07}) +[2026-10-03 14:48:12,182][starVLA.training.train_starvla][INFO] - Step 912, Loss: {'train/loss': 0.2182723581790924, 'train/grad_norm_pre_clip': 124.22175598144531, 'timing/data': 0.00019769000937230885, 'timing/model': 0.5644697740062838, 'learning_rate/qwen_vl_interface': 9.370708228159158e-06, 'learning_rate/action_model': 9.370708228159159e-05, 'epoch': 0.07}) +[2026-10-03 14:48:12,805][starVLA.training.train_starvla][INFO] - Step 913, Loss: {'train/loss': 0.44422659277915955, 'train/grad_norm_pre_clip': 13.480304718017578, 'timing/data': 0.00038457001210190356, 'timing/model': 0.3633306459960295, 'learning_rate/qwen_vl_interface': 9.369192568527398e-06, 'learning_rate/action_model': 9.369192568527398e-05, 'epoch': 0.07}) +[2026-10-03 14:48:13,451][starVLA.training.train_starvla][INFO] - Step 914, Loss: {'train/loss': 0.4273485541343689, 'train/grad_norm_pre_clip': 48.637481689453125, 'timing/data': 0.0002735900052357465, 'timing/model': 0.3643578369956231, 'learning_rate/qwen_vl_interface': 9.367675215651168e-06, 'learning_rate/action_model': 9.367675215651168e-05, 'epoch': 0.07}) +[2026-10-03 14:48:14,097][starVLA.training.train_starvla][INFO] - Step 915, Loss: {'train/loss': 0.4162180721759796, 'train/grad_norm_pre_clip': 194.3145751953125, 'timing/data': 0.00033628998789936304, 'timing/model': 0.36399845700361766, 'learning_rate/qwen_vl_interface': 9.366156170154194e-06, 'learning_rate/action_model': 9.366156170154194e-05, 'epoch': 0.07}) +[2026-10-03 14:48:14,744][starVLA.training.train_starvla][INFO] - Step 916, Loss: {'train/loss': 0.001494244672358036, 'train/grad_norm_pre_clip': 0.21606405079364777, 'timing/data': 0.00024480999854858965, 'timing/model': 0.36429645800672006, 'learning_rate/qwen_vl_interface': 9.3646354326609e-06, 'learning_rate/action_model': 9.364635432660899e-05, 'epoch': 0.07}) +[2026-10-03 14:48:15,391][starVLA.training.train_starvla][INFO] - Step 917, Loss: {'train/loss': 0.007239573169499636, 'train/grad_norm_pre_clip': 5.370279312133789, 'timing/data': 0.00024449999909847975, 'timing/model': 0.36420333599380683, 'learning_rate/qwen_vl_interface': 9.3631130037964e-06, 'learning_rate/action_model': 9.363113003796401e-05, 'epoch': 0.07}) +[2026-10-03 14:48:16,184][starVLA.training.train_starvla][INFO] - Step 918, Loss: {'train/loss': 0.0019405530765652657, 'train/grad_norm_pre_clip': 0.25755074620246887, 'timing/data': 0.00022935000015422702, 'timing/model': 0.3410546699888073, 'learning_rate/qwen_vl_interface': 9.361588884186513e-06, 'learning_rate/action_model': 9.361588884186513e-05, 'epoch': 0.07}) +[2026-10-03 14:48:16,831][starVLA.training.train_starvla][INFO] - Step 919, Loss: {'train/loss': 0.9985647797584534, 'train/grad_norm_pre_clip': 31.166259765625, 'timing/data': 0.0002456900110701099, 'timing/model': 0.36414466598944273, 'learning_rate/qwen_vl_interface': 9.360063074457744e-06, 'learning_rate/action_model': 9.360063074457743e-05, 'epoch': 0.07}) +[2026-10-03 14:48:17,478][starVLA.training.train_starvla][INFO] - Step 920, Loss: {'train/loss': 0.14341475069522858, 'train/grad_norm_pre_clip': 25.71387481689453, 'timing/data': 0.0002582400047685951, 'timing/model': 0.3642088769993279, 'learning_rate/qwen_vl_interface': 9.358535575237296e-06, 'learning_rate/action_model': 9.358535575237295e-05, 'epoch': 0.07}) +[2026-10-03 14:48:18,123][starVLA.training.train_starvla][INFO] - Step 921, Loss: {'train/loss': 0.837803840637207, 'train/grad_norm_pre_clip': 36.646583557128906, 'timing/data': 0.00032531999750062823, 'timing/model': 0.36367031700501684, 'learning_rate/qwen_vl_interface': 9.357006387153065e-06, 'learning_rate/action_model': 9.357006387153065e-05, 'epoch': 0.07}) +[2026-10-03 14:48:18,747][starVLA.training.train_starvla][INFO] - Step 922, Loss: {'train/loss': 0.719007134437561, 'train/grad_norm_pre_clip': 40.08861541748047, 'timing/data': 0.0002367100096307695, 'timing/model': 0.36414563699509017, 'learning_rate/qwen_vl_interface': 9.355475510833643e-06, 'learning_rate/action_model': 9.355475510833643e-05, 'epoch': 0.07}) +[2026-10-03 14:48:19,584][starVLA.training.train_starvla][INFO] - Step 923, Loss: {'train/loss': 0.0022205482237040997, 'train/grad_norm_pre_clip': 0.5264411568641663, 'timing/data': 0.00038951000897213817, 'timing/model': 0.5547232250100933, 'learning_rate/qwen_vl_interface': 9.353942946908318e-06, 'learning_rate/action_model': 9.353942946908317e-05, 'epoch': 0.07}) +[2026-10-03 14:48:20,231][starVLA.training.train_starvla][INFO] - Step 924, Loss: {'train/loss': 0.0028379380237311125, 'train/grad_norm_pre_clip': 0.3826095759868622, 'timing/data': 0.00031582999508827925, 'timing/model': 0.3644140480028, 'learning_rate/qwen_vl_interface': 9.352408696007065e-06, 'learning_rate/action_model': 9.352408696007065e-05, 'epoch': 0.07}) +[2026-10-03 14:48:20,877][starVLA.training.train_starvla][INFO] - Step 925, Loss: {'train/loss': 0.03407270833849907, 'train/grad_norm_pre_clip': 51.12167739868164, 'timing/data': 0.0004079200007254258, 'timing/model': 0.36358595700585283, 'learning_rate/qwen_vl_interface': 9.350872758760558e-06, 'learning_rate/action_model': 9.350872758760559e-05, 'epoch': 0.07}) +[2026-10-03 14:48:21,501][starVLA.training.train_starvla][INFO] - Step 926, Loss: {'train/loss': 0.7897387742996216, 'train/grad_norm_pre_clip': 86.68312072753906, 'timing/data': 0.00029659998835995793, 'timing/model': 0.36416660599934403, 'learning_rate/qwen_vl_interface': 9.349335135800164e-06, 'learning_rate/action_model': 9.349335135800163e-05, 'epoch': 0.07}) +[2026-10-03 14:48:22,147][starVLA.training.train_starvla][INFO] - Step 927, Loss: {'train/loss': 0.4839785099029541, 'train/grad_norm_pre_clip': 108.28429412841797, 'timing/data': 0.0002956200041808188, 'timing/model': 0.3642279279883951, 'learning_rate/qwen_vl_interface': 9.347795827757942e-06, 'learning_rate/action_model': 9.347795827757941e-05, 'epoch': 0.07}) +[2026-10-03 14:48:22,998][starVLA.training.train_starvla][INFO] - Step 928, Loss: {'train/loss': 0.9198353290557861, 'train/grad_norm_pre_clip': 184.15408325195312, 'timing/data': 0.0002763700031209737, 'timing/model': 0.5681553229951533, 'learning_rate/qwen_vl_interface': 9.346254835266643e-06, 'learning_rate/action_model': 9.346254835266642e-05, 'epoch': 0.07}) +[2026-10-03 14:48:23,644][starVLA.training.train_starvla][INFO] - Step 929, Loss: {'train/loss': 0.028116339817643166, 'train/grad_norm_pre_clip': 4.09796667098999, 'timing/data': 0.00020383000082802027, 'timing/model': 0.36381674800941255, 'learning_rate/qwen_vl_interface': 9.344712158959709e-06, 'learning_rate/action_model': 9.344712158959709e-05, 'epoch': 0.07}) +[2026-10-03 14:48:24,291][starVLA.training.train_starvla][INFO] - Step 930, Loss: {'train/loss': 0.05406695976853371, 'train/grad_norm_pre_clip': 28.030899047851562, 'timing/data': 0.00024480999854858965, 'timing/model': 0.3645234379946487, 'learning_rate/qwen_vl_interface': 9.343167799471278e-06, 'learning_rate/action_model': 9.343167799471277e-05, 'epoch': 0.07}) +[2026-10-03 14:48:24,937][starVLA.training.train_starvla][INFO] - Step 931, Loss: {'train/loss': 0.02904641069471836, 'train/grad_norm_pre_clip': 12.486445426940918, 'timing/data': 0.000291559990728274, 'timing/model': 0.3643496860022424, 'learning_rate/qwen_vl_interface': 9.341621757436177e-06, 'learning_rate/action_model': 9.341621757436177e-05, 'epoch': 0.07}) +[2026-10-03 14:48:25,586][starVLA.training.train_starvla][INFO] - Step 932, Loss: {'train/loss': 0.004366298206150532, 'train/grad_norm_pre_clip': 0.6277959942817688, 'timing/data': 0.000322369989589788, 'timing/model': 0.3656830879917834, 'learning_rate/qwen_vl_interface': 9.340074033489924e-06, 'learning_rate/action_model': 9.340074033489923e-05, 'epoch': 0.07}) +[2026-10-03 14:48:26,211][starVLA.training.train_starvla][INFO] - Step 933, Loss: {'train/loss': 0.00239439494907856, 'train/grad_norm_pre_clip': 0.7186001539230347, 'timing/data': 0.0003100199974142015, 'timing/model': 0.34193076999508776, 'learning_rate/qwen_vl_interface': 9.338524628268734e-06, 'learning_rate/action_model': 9.338524628268733e-05, 'epoch': 0.07}) +[2026-10-03 14:48:27,051][starVLA.training.train_starvla][INFO] - Step 934, Loss: {'train/loss': 0.02361380308866501, 'train/grad_norm_pre_clip': 4.360322952270508, 'timing/data': 0.0002847900032065809, 'timing/model': 0.5570032840041677, 'learning_rate/qwen_vl_interface': 9.336973542409505e-06, 'learning_rate/action_model': 9.336973542409504e-05, 'epoch': 0.07}) +[2026-10-03 14:48:27,675][starVLA.training.train_starvla][INFO] - Step 935, Loss: {'train/loss': 0.2969185709953308, 'train/grad_norm_pre_clip': 112.70548248291016, 'timing/data': 0.00023975000658538193, 'timing/model': 0.34228201999212615, 'learning_rate/qwen_vl_interface': 9.33542077654983e-06, 'learning_rate/action_model': 9.33542077654983e-05, 'epoch': 0.07}) +[2026-10-03 14:48:28,322][starVLA.training.train_starvla][INFO] - Step 936, Loss: {'train/loss': 0.007299022749066353, 'train/grad_norm_pre_clip': 3.8935511112213135, 'timing/data': 0.0002931400085799396, 'timing/model': 0.3643762060091831, 'learning_rate/qwen_vl_interface': 9.333866331327995e-06, 'learning_rate/action_model': 9.333866331327995e-05, 'epoch': 0.07}) +[2026-10-03 14:48:28,947][starVLA.training.train_starvla][INFO] - Step 937, Loss: {'train/loss': 0.4196588397026062, 'train/grad_norm_pre_clip': 33.589847564697266, 'timing/data': 0.00024682999355718493, 'timing/model': 0.34171149000758305, 'learning_rate/qwen_vl_interface': 9.332310207382972e-06, 'learning_rate/action_model': 9.332310207382972e-05, 'epoch': 0.07}) +[2026-10-03 14:48:29,594][starVLA.training.train_starvla][INFO] - Step 938, Loss: {'train/loss': 0.9689321517944336, 'train/grad_norm_pre_clip': 29.404024124145508, 'timing/data': 0.00030900999263394624, 'timing/model': 0.36450911799329333, 'learning_rate/qwen_vl_interface': 9.330752405354426e-06, 'learning_rate/action_model': 9.330752405354426e-05, 'epoch': 0.07}) +[2026-10-03 14:48:30,437][starVLA.training.train_starvla][INFO] - Step 939, Loss: {'train/loss': 0.524052619934082, 'train/grad_norm_pre_clip': 58.764686584472656, 'timing/data': 0.00033726000401657075, 'timing/model': 0.5599844039970776, 'learning_rate/qwen_vl_interface': 9.32919292588271e-06, 'learning_rate/action_model': 9.32919292588271e-05, 'epoch': 0.07}) +[2026-10-03 14:48:31,085][starVLA.training.train_starvla][INFO] - Step 940, Loss: {'train/loss': 0.0016036813613027334, 'train/grad_norm_pre_clip': 0.37666967511177063, 'timing/data': 0.0003089500096393749, 'timing/model': 0.36488013800408226, 'learning_rate/qwen_vl_interface': 9.327631769608864e-06, 'learning_rate/action_model': 9.327631769608864e-05, 'epoch': 0.07}) +[2026-10-03 14:48:31,732][starVLA.training.train_starvla][INFO] - Step 941, Loss: {'train/loss': 0.001727847382426262, 'train/grad_norm_pre_clip': 0.41316190361976624, 'timing/data': 0.00028993999876547605, 'timing/model': 0.3639281479991041, 'learning_rate/qwen_vl_interface': 9.326068937174627e-06, 'learning_rate/action_model': 9.326068937174626e-05, 'epoch': 0.07}) +[2026-10-03 14:48:32,378][starVLA.training.train_starvla][INFO] - Step 942, Loss: {'train/loss': 0.23713800311088562, 'train/grad_norm_pre_clip': 29.46596336364746, 'timing/data': 0.00031991000287234783, 'timing/model': 0.3644722559984075, 'learning_rate/qwen_vl_interface': 9.324504429222412e-06, 'learning_rate/action_model': 9.324504429222412e-05, 'epoch': 0.07}) +[2026-10-03 14:48:33,026][starVLA.training.train_starvla][INFO] - Step 943, Loss: {'train/loss': 0.16649019718170166, 'train/grad_norm_pre_clip': 161.45944213867188, 'timing/data': 0.0002470600011292845, 'timing/model': 0.3648308479896514, 'learning_rate/qwen_vl_interface': 9.322938246395336e-06, 'learning_rate/action_model': 9.322938246395337e-05, 'epoch': 0.07}) +[2026-10-03 14:48:33,873][starVLA.training.train_starvla][INFO] - Step 944, Loss: {'train/loss': 0.005469006020575762, 'train/grad_norm_pre_clip': 1.6147868633270264, 'timing/data': 0.00028176000341773033, 'timing/model': 0.5644850940007018, 'learning_rate/qwen_vl_interface': 9.321370389337196e-06, 'learning_rate/action_model': 9.321370389337196e-05, 'epoch': 0.07}) +[2026-10-03 14:48:34,520][starVLA.training.train_starvla][INFO] - Step 945, Loss: {'train/loss': 0.00418563187122345, 'train/grad_norm_pre_clip': 1.0245399475097656, 'timing/data': 0.0003169599949615076, 'timing/model': 0.36464069699286483, 'learning_rate/qwen_vl_interface': 9.319800858692478e-06, 'learning_rate/action_model': 9.319800858692476e-05, 'epoch': 0.07}) +[2026-10-03 14:48:35,167][starVLA.training.train_starvla][INFO] - Step 946, Loss: {'train/loss': 0.3302239775657654, 'train/grad_norm_pre_clip': 65.11872100830078, 'timing/data': 0.00029938999796286225, 'timing/model': 0.36440860699804034, 'learning_rate/qwen_vl_interface': 9.318229655106357e-06, 'learning_rate/action_model': 9.318229655106356e-05, 'epoch': 0.07}) +[2026-10-03 14:48:35,816][starVLA.training.train_starvla][INFO] - Step 947, Loss: {'train/loss': 0.12016621977090836, 'train/grad_norm_pre_clip': 86.04994201660156, 'timing/data': 0.00024280000070575625, 'timing/model': 0.36519792600302026, 'learning_rate/qwen_vl_interface': 9.316656779224693e-06, 'learning_rate/action_model': 9.316656779224693e-05, 'epoch': 0.07}) +[2026-10-03 14:48:36,442][starVLA.training.train_starvla][INFO] - Step 948, Loss: {'train/loss': 0.0010048651602119207, 'train/grad_norm_pre_clip': 0.14242620766162872, 'timing/data': 0.00030247999529819936, 'timing/model': 0.3428058389981743, 'learning_rate/qwen_vl_interface': 9.31508223169404e-06, 'learning_rate/action_model': 9.31508223169404e-05, 'epoch': 0.07}) +[2026-10-03 14:48:37,089][starVLA.training.train_starvla][INFO] - Step 949, Loss: {'train/loss': 0.40994542837142944, 'train/grad_norm_pre_clip': 94.44820404052734, 'timing/data': 0.00025182000536005944, 'timing/model': 0.36422580800717697, 'learning_rate/qwen_vl_interface': 9.313506013161633e-06, 'learning_rate/action_model': 9.313506013161633e-05, 'epoch': 0.07}) +[2026-10-03 14:48:37,943][starVLA.training.train_starvla][INFO] - Step 950, Loss: {'train/loss': 0.0011165594914928079, 'train/grad_norm_pre_clip': 0.20032818615436554, 'timing/data': 0.0002543989976402372, 'timing/model': 0.3643599179922603, 'learning_rate/qwen_vl_interface': 9.311928124275396e-06, 'learning_rate/action_model': 9.311928124275395e-05, 'epoch': 0.07}) +[2026-10-03 14:48:38,569][starVLA.training.train_starvla][INFO] - Step 951, Loss: {'train/loss': 0.000706614344380796, 'train/grad_norm_pre_clip': 0.09546170383691788, 'timing/data': 0.00031262000265996903, 'timing/model': 0.36432485800469294, 'learning_rate/qwen_vl_interface': 9.310348565683938e-06, 'learning_rate/action_model': 9.310348565683939e-05, 'epoch': 0.07}) +[2026-10-03 14:48:39,217][starVLA.training.train_starvla][INFO] - Step 952, Loss: {'train/loss': 0.011434155516326427, 'train/grad_norm_pre_clip': 10.915838241577148, 'timing/data': 0.0002970299974549562, 'timing/model': 0.3646864659967832, 'learning_rate/qwen_vl_interface': 9.30876733803656e-06, 'learning_rate/action_model': 9.308767338036559e-05, 'epoch': 0.07}) +[2026-10-03 14:48:39,864][starVLA.training.train_starvla][INFO] - Step 953, Loss: {'train/loss': 0.04916414991021156, 'train/grad_norm_pre_clip': 70.95599365234375, 'timing/data': 0.00023824999516364187, 'timing/model': 0.36438635700324085, 'learning_rate/qwen_vl_interface': 9.30718444198324e-06, 'learning_rate/action_model': 9.307184441983239e-05, 'epoch': 0.07}) +[2026-10-03 14:48:40,488][starVLA.training.train_starvla][INFO] - Step 954, Loss: {'train/loss': 0.017489543184638023, 'train/grad_norm_pre_clip': 53.1141357421875, 'timing/data': 0.00024542999744880944, 'timing/model': 0.34222449001390487, 'learning_rate/qwen_vl_interface': 9.30559987817465e-06, 'learning_rate/action_model': 9.30559987817465e-05, 'epoch': 0.07}) +[2026-10-03 14:48:41,113][starVLA.training.train_starvla][INFO] - Step 955, Loss: {'train/loss': 0.004233065992593765, 'train/grad_norm_pre_clip': 4.835815906524658, 'timing/data': 0.000241660003666766, 'timing/model': 0.34240651001164224, 'learning_rate/qwen_vl_interface': 9.304013647262143e-06, 'learning_rate/action_model': 9.304013647262143e-05, 'epoch': 0.07}) +[2026-10-03 14:48:41,954][starVLA.training.train_starvla][INFO] - Step 956, Loss: {'train/loss': 0.7824906706809998, 'train/grad_norm_pre_clip': 57.14323425292969, 'timing/data': 0.0002857200015569106, 'timing/model': 0.3648268880060641, 'learning_rate/qwen_vl_interface': 9.302425749897761e-06, 'learning_rate/action_model': 9.302425749897761e-05, 'epoch': 0.07}) +[2026-10-03 14:48:42,578][starVLA.training.train_starvla][INFO] - Step 957, Loss: {'train/loss': 0.0012856351677328348, 'train/grad_norm_pre_clip': 1.0908564329147339, 'timing/data': 0.00028858998848591, 'timing/model': 0.34155642898986116, 'learning_rate/qwen_vl_interface': 9.300836186734225e-06, 'learning_rate/action_model': 9.300836186734226e-05, 'epoch': 0.07}) +[2026-10-03 14:48:43,203][starVLA.training.train_starvla][INFO] - Step 958, Loss: {'train/loss': 0.357731431722641, 'train/grad_norm_pre_clip': 46.49420166015625, 'timing/data': 0.0003843099984806031, 'timing/model': 0.341753319997224, 'learning_rate/qwen_vl_interface': 9.299244958424947e-06, 'learning_rate/action_model': 9.299244958424946e-05, 'epoch': 0.07}) +[2026-10-03 14:48:43,851][starVLA.training.train_starvla][INFO] - Step 959, Loss: {'train/loss': 0.000615482684224844, 'train/grad_norm_pre_clip': 0.5216050744056702, 'timing/data': 0.0002533899969421327, 'timing/model': 0.36530412700085435, 'learning_rate/qwen_vl_interface': 9.29765206562402e-06, 'learning_rate/action_model': 9.297652065624018e-05, 'epoch': 0.07}) +[2026-10-03 14:48:44,499][starVLA.training.train_starvla][INFO] - Step 960, Loss: {'train/loss': 0.0010703925509005785, 'train/grad_norm_pre_clip': 0.6852293610572815, 'timing/data': 0.00023576999956276268, 'timing/model': 0.3649046670034295, 'learning_rate/qwen_vl_interface': 9.296057508986221e-06, 'learning_rate/action_model': 9.296057508986221e-05, 'epoch': 0.07}) +[2026-10-03 14:48:45,341][starVLA.training.train_starvla][INFO] - Step 961, Loss: {'train/loss': 0.000783524417784065, 'train/grad_norm_pre_clip': 0.30931201577186584, 'timing/data': 0.0002499900001566857, 'timing/model': 0.3641019179922296, 'learning_rate/qwen_vl_interface': 9.294461289167013e-06, 'learning_rate/action_model': 9.294461289167014e-05, 'epoch': 0.07}) +[2026-10-03 14:48:45,987][starVLA.training.train_starvla][INFO] - Step 962, Loss: {'train/loss': 1.575913667678833, 'train/grad_norm_pre_clip': 196.35397338867188, 'timing/data': 0.0003099590103374794, 'timing/model': 0.36427074699895456, 'learning_rate/qwen_vl_interface': 9.292863406822543e-06, 'learning_rate/action_model': 9.292863406822542e-05, 'epoch': 0.07}) +[2026-10-03 14:48:46,635][starVLA.training.train_starvla][INFO] - Step 963, Loss: {'train/loss': 1.2956509590148926, 'train/grad_norm_pre_clip': 186.2753448486328, 'timing/data': 0.0002453799970680848, 'timing/model': 0.36484126799041405, 'learning_rate/qwen_vl_interface': 9.291263862609638e-06, 'learning_rate/action_model': 9.291263862609638e-05, 'epoch': 0.07}) +[2026-10-03 14:48:47,261][starVLA.training.train_starvla][INFO] - Step 964, Loss: {'train/loss': 0.14506204426288605, 'train/grad_norm_pre_clip': 355.6822814941406, 'timing/data': 0.0003645999968284741, 'timing/model': 0.36456369698862545, 'learning_rate/qwen_vl_interface': 9.28966265718581e-06, 'learning_rate/action_model': 9.28966265718581e-05, 'epoch': 0.07}) +[2026-10-03 14:48:47,908][starVLA.training.train_starvla][INFO] - Step 965, Loss: {'train/loss': 0.804912805557251, 'train/grad_norm_pre_clip': 218.6978302001953, 'timing/data': 0.0002890699979616329, 'timing/model': 0.364289856996038, 'learning_rate/qwen_vl_interface': 9.288059791209256e-06, 'learning_rate/action_model': 9.288059791209256e-05, 'epoch': 0.07}) +[2026-10-03 14:48:48,757][starVLA.training.train_starvla][INFO] - Step 966, Loss: {'train/loss': 0.035470787435770035, 'train/grad_norm_pre_clip': 108.3056869506836, 'timing/data': 0.0002900700055761263, 'timing/model': 0.5658341929956805, 'learning_rate/qwen_vl_interface': 9.286455265338853e-06, 'learning_rate/action_model': 9.286455265338853e-05, 'epoch': 0.07}) +[2026-10-03 14:48:49,405][starVLA.training.train_starvla][INFO] - Step 967, Loss: {'train/loss': 0.0002426888677291572, 'train/grad_norm_pre_clip': 0.05663171410560608, 'timing/data': 0.0002982800069730729, 'timing/model': 0.3645955359970685, 'learning_rate/qwen_vl_interface': 9.284849080234161e-06, 'learning_rate/action_model': 9.28484908023416e-05, 'epoch': 0.07}) +[2026-10-03 14:48:50,052][starVLA.training.train_starvla][INFO] - Step 968, Loss: {'train/loss': 0.00030440723639912903, 'train/grad_norm_pre_clip': 0.11282608658075333, 'timing/data': 0.0002906500012613833, 'timing/model': 0.36480070700054057, 'learning_rate/qwen_vl_interface': 9.283241236555421e-06, 'learning_rate/action_model': 9.283241236555421e-05, 'epoch': 0.07}) +[2026-10-03 14:48:50,700][starVLA.training.train_starvla][INFO] - Step 969, Loss: {'train/loss': 0.4578540027141571, 'train/grad_norm_pre_clip': 28.847274780273438, 'timing/data': 0.0003319500101497397, 'timing/model': 0.36488051699416246, 'learning_rate/qwen_vl_interface': 9.281631734963557e-06, 'learning_rate/action_model': 9.281631734963557e-05, 'epoch': 0.07}) +[2026-10-03 14:48:51,347][starVLA.training.train_starvla][INFO] - Step 970, Loss: {'train/loss': 0.0006230080616660416, 'train/grad_norm_pre_clip': 0.5447362065315247, 'timing/data': 0.00019764999160543084, 'timing/model': 0.36499161600659136, 'learning_rate/qwen_vl_interface': 9.280020576120176e-06, 'learning_rate/action_model': 9.280020576120176e-05, 'epoch': 0.07}) +[2026-10-03 14:48:52,197][starVLA.training.train_starvla][INFO] - Step 971, Loss: {'train/loss': 0.6638915538787842, 'train/grad_norm_pre_clip': 25.017728805541992, 'timing/data': 0.00031960000342223793, 'timing/model': 0.5671345230075531, 'learning_rate/qwen_vl_interface': 9.278407760687562e-06, 'learning_rate/action_model': 9.278407760687562e-05, 'epoch': 0.07}) +[2026-10-03 14:48:52,845][starVLA.training.train_starvla][INFO] - Step 972, Loss: {'train/loss': 0.00034372523077763617, 'train/grad_norm_pre_clip': 0.055859919637441635, 'timing/data': 0.0003002099983859807, 'timing/model': 0.3649146770039806, 'learning_rate/qwen_vl_interface': 9.276793289328686e-06, 'learning_rate/action_model': 9.276793289328685e-05, 'epoch': 0.07}) +[2026-10-03 14:48:53,492][starVLA.training.train_starvla][INFO] - Step 973, Loss: {'train/loss': 0.04656665772199631, 'train/grad_norm_pre_clip': 138.1688232421875, 'timing/data': 0.0002460100076859817, 'timing/model': 0.3640349170018453, 'learning_rate/qwen_vl_interface': 9.27517716270719e-06, 'learning_rate/action_model': 9.275177162707191e-05, 'epoch': 0.07}) +[2026-10-03 14:48:54,139][starVLA.training.train_starvla][INFO] - Step 974, Loss: {'train/loss': 0.00089081353507936, 'train/grad_norm_pre_clip': 0.25781044363975525, 'timing/data': 0.00033815999631769955, 'timing/model': 0.36426055699121207, 'learning_rate/qwen_vl_interface': 9.273559381487409e-06, 'learning_rate/action_model': 9.273559381487409e-05, 'epoch': 0.07}) +[2026-10-03 14:48:54,787][starVLA.training.train_starvla][INFO] - Step 975, Loss: {'train/loss': 0.0005372806335799396, 'train/grad_norm_pre_clip': 0.09572714567184448, 'timing/data': 0.0003230599977541715, 'timing/model': 0.3647402560018236, 'learning_rate/qwen_vl_interface': 9.27193994633435e-06, 'learning_rate/action_model': 9.27193994633435e-05, 'epoch': 0.07}) +[2026-10-03 14:48:55,435][starVLA.training.train_starvla][INFO] - Step 976, Loss: {'train/loss': 0.04842101037502289, 'train/grad_norm_pre_clip': 122.80020141601562, 'timing/data': 0.00019438000163063407, 'timing/model': 0.3649164469970856, 'learning_rate/qwen_vl_interface': 9.2703188579137e-06, 'learning_rate/action_model': 9.2703188579137e-05, 'epoch': 0.07}) +[2026-10-03 14:48:56,275][starVLA.training.train_starvla][INFO] - Step 977, Loss: {'train/loss': 0.0007967092096805573, 'train/grad_norm_pre_clip': 0.15159614384174347, 'timing/data': 0.0002551600045990199, 'timing/model': 0.3640841469896259, 'learning_rate/qwen_vl_interface': 9.26869611689183e-06, 'learning_rate/action_model': 9.26869611689183e-05, 'epoch': 0.07}) +[2026-10-03 14:48:56,922][starVLA.training.train_starvla][INFO] - Step 978, Loss: {'train/loss': 0.009411836974322796, 'train/grad_norm_pre_clip': 11.203180313110352, 'timing/data': 0.0003583999932743609, 'timing/model': 0.3644048170099268, 'learning_rate/qwen_vl_interface': 9.267071723935784e-06, 'learning_rate/action_model': 9.267071723935784e-05, 'epoch': 0.07}) +[2026-10-03 14:48:57,569][starVLA.training.train_starvla][INFO] - Step 979, Loss: {'train/loss': 0.0012467284686863422, 'train/grad_norm_pre_clip': 0.298504501581192, 'timing/data': 0.0003180199855705723, 'timing/model': 0.3643403359892545, 'learning_rate/qwen_vl_interface': 9.26544567971329e-06, 'learning_rate/action_model': 9.26544567971329e-05, 'epoch': 0.07}) +[2026-10-03 14:48:58,217][starVLA.training.train_starvla][INFO] - Step 980, Loss: {'train/loss': 0.0009716093190945685, 'train/grad_norm_pre_clip': 0.20000521838665009, 'timing/data': 0.00031712000782135874, 'timing/model': 0.3651181469904259, 'learning_rate/qwen_vl_interface': 9.263817984892756e-06, 'learning_rate/action_model': 9.263817984892757e-05, 'epoch': 0.07}) +[2026-10-03 14:48:58,865][starVLA.training.train_starvla][INFO] - Step 981, Loss: {'train/loss': 0.10477268695831299, 'train/grad_norm_pre_clip': 137.45777893066406, 'timing/data': 0.00024285000108648092, 'timing/model': 0.364240917013376, 'learning_rate/qwen_vl_interface': 9.262188640143262e-06, 'learning_rate/action_model': 9.262188640143262e-05, 'epoch': 0.07}) +[2026-10-03 14:48:59,708][starVLA.training.train_starvla][INFO] - Step 982, Loss: {'train/loss': 0.3944583237171173, 'train/grad_norm_pre_clip': 26.69839859008789, 'timing/data': 0.00020782998763024807, 'timing/model': 0.36505889700492844, 'learning_rate/qwen_vl_interface': 9.260557646134572e-06, 'learning_rate/action_model': 9.260557646134572e-05, 'epoch': 0.07}) +[2026-10-03 14:49:00,355][starVLA.training.train_starvla][INFO] - Step 983, Loss: {'train/loss': 0.5995628237724304, 'train/grad_norm_pre_clip': 27.23029327392578, 'timing/data': 0.0002031399926636368, 'timing/model': 0.3652646060072584, 'learning_rate/qwen_vl_interface': 9.258925003537127e-06, 'learning_rate/action_model': 9.258925003537126e-05, 'epoch': 0.07}) +[2026-10-03 14:49:01,004][starVLA.training.train_starvla][INFO] - Step 984, Loss: {'train/loss': 0.0015694680623710155, 'train/grad_norm_pre_clip': 1.688518762588501, 'timing/data': 0.0003142100031254813, 'timing/model': 0.36498145799851045, 'learning_rate/qwen_vl_interface': 9.257290713022041e-06, 'learning_rate/action_model': 9.257290713022041e-05, 'epoch': 0.07}) +[2026-10-03 14:49:01,651][starVLA.training.train_starvla][INFO] - Step 985, Loss: {'train/loss': 0.0005175253027118742, 'train/grad_norm_pre_clip': 0.08262943476438522, 'timing/data': 0.00032813000143505633, 'timing/model': 0.3641409780102549, 'learning_rate/qwen_vl_interface': 9.255654775261112e-06, 'learning_rate/action_model': 9.255654775261112e-05, 'epoch': 0.07}) +[2026-10-03 14:49:02,298][starVLA.training.train_starvla][INFO] - Step 986, Loss: {'train/loss': 0.4890729486942291, 'train/grad_norm_pre_clip': 121.8162841796875, 'timing/data': 0.00020060999668203294, 'timing/model': 0.36458667599072214, 'learning_rate/qwen_vl_interface': 9.254017190926813e-06, 'learning_rate/action_model': 9.254017190926813e-05, 'epoch': 0.07}) +[2026-10-03 14:49:03,140][starVLA.training.train_starvla][INFO] - Step 987, Loss: {'train/loss': 0.0006572986021637917, 'train/grad_norm_pre_clip': 0.13454148173332214, 'timing/data': 0.00023968999448698014, 'timing/model': 0.5598427639924921, 'learning_rate/qwen_vl_interface': 9.25237796069229e-06, 'learning_rate/action_model': 9.25237796069229e-05, 'epoch': 0.07}) +[2026-10-03 14:49:03,788][starVLA.training.train_starvla][INFO] - Step 988, Loss: {'train/loss': 0.0005849923472851515, 'train/grad_norm_pre_clip': 0.1103648766875267, 'timing/data': 0.0001960900117410347, 'timing/model': 0.36541699701047037, 'learning_rate/qwen_vl_interface': 9.250737085231371e-06, 'learning_rate/action_model': 9.250737085231372e-05, 'epoch': 0.07}) +[2026-10-03 14:49:04,436][starVLA.training.train_starvla][INFO] - Step 989, Loss: {'train/loss': 0.050229862332344055, 'train/grad_norm_pre_clip': 41.21651840209961, 'timing/data': 0.00032125000143423676, 'timing/model': 0.3646328679897124, 'learning_rate/qwen_vl_interface': 9.24909456521856e-06, 'learning_rate/action_model': 9.249094565218559e-05, 'epoch': 0.07}) +[2026-10-03 14:49:05,083][starVLA.training.train_starvla][INFO] - Step 990, Loss: {'train/loss': 0.0006458169664256275, 'train/grad_norm_pre_clip': 0.14228519797325134, 'timing/data': 0.0001964900002349168, 'timing/model': 0.3649645179975778, 'learning_rate/qwen_vl_interface': 9.247450401329027e-06, 'learning_rate/action_model': 9.247450401329027e-05, 'epoch': 0.07}) +[2026-10-03 14:49:05,732][starVLA.training.train_starvla][INFO] - Step 991, Loss: {'train/loss': 0.0006368356407620013, 'train/grad_norm_pre_clip': 0.15569786727428436, 'timing/data': 0.0002969699999084696, 'timing/model': 0.36579363599594217, 'learning_rate/qwen_vl_interface': 9.245804594238637e-06, 'learning_rate/action_model': 9.245804594238636e-05, 'epoch': 0.07}) +[2026-10-03 14:49:06,591][starVLA.training.train_starvla][INFO] - Step 992, Loss: {'train/loss': 0.2873727083206177, 'train/grad_norm_pre_clip': 111.03762817382812, 'timing/data': 0.00019958001212216914, 'timing/model': 0.5714995619928231, 'learning_rate/qwen_vl_interface': 9.244157144623909e-06, 'learning_rate/action_model': 9.244157144623909e-05, 'epoch': 0.07}) +[2026-10-03 14:49:07,237][starVLA.training.train_starvla][INFO] - Step 993, Loss: {'train/loss': 0.41388410329818726, 'train/grad_norm_pre_clip': 61.51397705078125, 'timing/data': 0.00032577000092715025, 'timing/model': 0.36434269799792673, 'learning_rate/qwen_vl_interface': 9.242508053162051e-06, 'learning_rate/action_model': 9.242508053162051e-05, 'epoch': 0.07}) +[2026-10-03 14:49:07,861][starVLA.training.train_starvla][INFO] - Step 994, Loss: {'train/loss': 0.0033388754818588495, 'train/grad_norm_pre_clip': 1.8932433128356934, 'timing/data': 0.00023272000544238836, 'timing/model': 0.34187395000481047, 'learning_rate/qwen_vl_interface': 9.240857320530944e-06, 'learning_rate/action_model': 9.240857320530943e-05, 'epoch': 0.07}) +[2026-10-03 14:49:08,509][starVLA.training.train_starvla][INFO] - Step 995, Loss: {'train/loss': 0.004485331010073423, 'train/grad_norm_pre_clip': 3.339925527572632, 'timing/data': 0.00023632000375073403, 'timing/model': 0.3650071359879803, 'learning_rate/qwen_vl_interface': 9.23920494740914e-06, 'learning_rate/action_model': 9.23920494740914e-05, 'epoch': 0.07}) +[2026-10-03 14:49:09,157][starVLA.training.train_starvla][INFO] - Step 996, Loss: {'train/loss': 0.0014926978619769216, 'train/grad_norm_pre_clip': 0.6257001161575317, 'timing/data': 0.00023426000552717596, 'timing/model': 0.36520223700790666, 'learning_rate/qwen_vl_interface': 9.237550934475868e-06, 'learning_rate/action_model': 9.237550934475867e-05, 'epoch': 0.07}) +[2026-10-03 14:49:09,804][starVLA.training.train_starvla][INFO] - Step 997, Loss: {'train/loss': 0.2477378398180008, 'train/grad_norm_pre_clip': 54.44182205200195, 'timing/data': 0.0004782799951499328, 'timing/model': 0.3642710570129566, 'learning_rate/qwen_vl_interface': 9.235895282411029e-06, 'learning_rate/action_model': 9.235895282411029e-05, 'epoch': 0.07}) +[2026-10-03 14:49:10,635][starVLA.training.train_starvla][INFO] - Step 998, Loss: {'train/loss': 0.0005576069233939052, 'train/grad_norm_pre_clip': 0.11349696666002274, 'timing/data': 0.0003160499909427017, 'timing/model': 0.3421287300006952, 'learning_rate/qwen_vl_interface': 9.2342379918952e-06, 'learning_rate/action_model': 9.2342379918952e-05, 'epoch': 0.07}) +[2026-10-03 14:49:11,282][starVLA.training.train_starvla][INFO] - Step 999, Loss: {'train/loss': 0.10214174538850784, 'train/grad_norm_pre_clip': 58.21605682373047, 'timing/data': 0.00024137999571394175, 'timing/model': 0.36510613599966746, 'learning_rate/qwen_vl_interface': 9.232579063609631e-06, 'learning_rate/action_model': 9.232579063609631e-05, 'epoch': 0.07}) +[2026-10-03 14:49:11,932][starVLA.training.train_starvla][INFO] - Step 1000, Loss: {'train/loss': 0.46413454413414, 'train/grad_norm_pre_clip': 32.13639831542969, 'timing/data': 0.0002977599942823872, 'timing/model': 0.3668908370018471, 'learning_rate/qwen_vl_interface': 9.230918498236243e-06, 'learning_rate/action_model': 9.230918498236244e-05, 'epoch': 0.07}) +[2026-10-03 14:49:12,583][starVLA.training.train_starvla][INFO] - Step 1001, Loss: {'train/loss': 0.0007426568772643805, 'train/grad_norm_pre_clip': 1.315412998199463, 'timing/data': 0.00023026899725664407, 'timing/model': 0.3662372580001829, 'learning_rate/qwen_vl_interface': 9.229256296457635e-06, 'learning_rate/action_model': 9.229256296457634e-05, 'epoch': 0.07}) +[2026-10-03 14:49:13,234][starVLA.training.train_starvla][INFO] - Step 1002, Loss: {'train/loss': 0.00015185224765446037, 'train/grad_norm_pre_clip': 0.01969868503510952, 'timing/data': 0.00019943001097999513, 'timing/model': 0.3665578660002211, 'learning_rate/qwen_vl_interface': 9.22759245895707e-06, 'learning_rate/action_model': 9.227592458957071e-05, 'epoch': 0.07}) +[2026-10-03 14:49:14,074][starVLA.training.train_starvla][INFO] - Step 1003, Loss: {'train/loss': 0.7218962907791138, 'train/grad_norm_pre_clip': 153.48178100585938, 'timing/data': 0.0002812799939420074, 'timing/model': 0.5558548639965011, 'learning_rate/qwen_vl_interface': 9.225926986418494e-06, 'learning_rate/action_model': 9.225926986418494e-05, 'epoch': 0.07}) +[2026-10-03 14:49:14,725][starVLA.training.train_starvla][INFO] - Step 1004, Loss: {'train/loss': 0.0002301196800544858, 'train/grad_norm_pre_clip': 0.053589317947626114, 'timing/data': 0.0002834200131474063, 'timing/model': 0.36687872699985746, 'learning_rate/qwen_vl_interface': 9.224259879526519e-06, 'learning_rate/action_model': 9.224259879526519e-05, 'epoch': 0.07}) +[2026-10-03 14:49:15,376][starVLA.training.train_starvla][INFO] - Step 1005, Loss: {'train/loss': 0.005096188746392727, 'train/grad_norm_pre_clip': 10.973888397216797, 'timing/data': 0.0003667800046969205, 'timing/model': 0.3659450779960025, 'learning_rate/qwen_vl_interface': 9.222591138966427e-06, 'learning_rate/action_model': 9.222591138966426e-05, 'epoch': 0.07}) +[2026-10-03 14:49:16,026][starVLA.training.train_starvla][INFO] - Step 1006, Loss: {'train/loss': 0.00034905943903140724, 'train/grad_norm_pre_clip': 0.19914835691452026, 'timing/data': 0.00029334999271668494, 'timing/model': 0.3658163060026709, 'learning_rate/qwen_vl_interface': 9.220920765424175e-06, 'learning_rate/action_model': 9.220920765424174e-05, 'epoch': 0.07}) +[2026-10-03 14:49:16,677][starVLA.training.train_starvla][INFO] - Step 1007, Loss: {'train/loss': 0.40058207511901855, 'train/grad_norm_pre_clip': 162.22329711914062, 'timing/data': 0.00027529000362847, 'timing/model': 0.3664689869910944, 'learning_rate/qwen_vl_interface': 9.219248759586392e-06, 'learning_rate/action_model': 9.219248759586391e-05, 'epoch': 0.07}) +[2026-10-03 14:49:17,328][starVLA.training.train_starvla][INFO] - Step 1008, Loss: {'train/loss': 0.7344551086425781, 'train/grad_norm_pre_clip': 134.54702758789062, 'timing/data': 0.0001952500024344772, 'timing/model': 0.3666155269893352, 'learning_rate/qwen_vl_interface': 9.217575122140375e-06, 'learning_rate/action_model': 9.217575122140375e-05, 'epoch': 0.07}) +[2026-10-03 14:49:18,143][starVLA.training.train_starvla][INFO] - Step 1009, Loss: {'train/loss': 0.00048429632443003356, 'train/grad_norm_pre_clip': 0.603882908821106, 'timing/data': 0.0003742800035979599, 'timing/model': 0.36527370699332096, 'learning_rate/qwen_vl_interface': 9.215899853774094e-06, 'learning_rate/action_model': 9.215899853774093e-05, 'epoch': 0.07}) +[2026-10-03 14:49:18,770][starVLA.training.train_starvla][INFO] - Step 1010, Loss: {'train/loss': 0.5335343480110168, 'train/grad_norm_pre_clip': 45.96559524536133, 'timing/data': 0.00023577999672852457, 'timing/model': 0.3430502380069811, 'learning_rate/qwen_vl_interface': 9.214222955176189e-06, 'learning_rate/action_model': 9.214222955176189e-05, 'epoch': 0.07}) +[2026-10-03 14:49:19,421][starVLA.training.train_starvla][INFO] - Step 1011, Loss: {'train/loss': 0.00584312854334712, 'train/grad_norm_pre_clip': 11.708998680114746, 'timing/data': 0.00024959999427665025, 'timing/model': 0.36655863698979374, 'learning_rate/qwen_vl_interface': 9.212544427035967e-06, 'learning_rate/action_model': 9.212544427035967e-05, 'epoch': 0.07}) +[2026-10-03 14:49:20,072][starVLA.training.train_starvla][INFO] - Step 1012, Loss: {'train/loss': 0.00020934682106599212, 'train/grad_norm_pre_clip': 0.041893016546964645, 'timing/data': 0.0002296499878866598, 'timing/model': 0.367390836996492, 'learning_rate/qwen_vl_interface': 9.21086427004341e-06, 'learning_rate/action_model': 9.210864270043411e-05, 'epoch': 0.07}) +[2026-10-03 14:49:20,701][starVLA.training.train_starvla][INFO] - Step 1013, Loss: {'train/loss': 0.0013578049838542938, 'train/grad_norm_pre_clip': 2.6469783782958984, 'timing/data': 0.00033518999407533556, 'timing/model': 0.36626990599324927, 'learning_rate/qwen_vl_interface': 9.209182484889168e-06, 'learning_rate/action_model': 9.209182484889167e-05, 'epoch': 0.07}) +[2026-10-03 14:49:21,548][starVLA.training.train_starvla][INFO] - Step 1014, Loss: {'train/loss': 0.4259748160839081, 'train/grad_norm_pre_clip': 128.91595458984375, 'timing/data': 0.000400630000513047, 'timing/model': 0.3664410960045643, 'learning_rate/qwen_vl_interface': 9.207499072264557e-06, 'learning_rate/action_model': 9.207499072264558e-05, 'epoch': 0.07}) +[2026-10-03 14:49:22,198][starVLA.training.train_starvla][INFO] - Step 1015, Loss: {'train/loss': 0.6736224293708801, 'train/grad_norm_pre_clip': 67.42932891845703, 'timing/data': 0.00019653899653349072, 'timing/model': 0.3662876480084378, 'learning_rate/qwen_vl_interface': 9.205814032861566e-06, 'learning_rate/action_model': 9.205814032861566e-05, 'epoch': 0.07}) +[2026-10-03 14:49:22,828][starVLA.training.train_starvla][INFO] - Step 1016, Loss: {'train/loss': 0.00044744822662323713, 'train/grad_norm_pre_clip': 0.29874637722969055, 'timing/data': 0.00029398000333458185, 'timing/model': 0.3667874469974777, 'learning_rate/qwen_vl_interface': 9.204127367372849e-06, 'learning_rate/action_model': 9.204127367372849e-05, 'epoch': 0.07}) +[2026-10-03 14:49:23,455][starVLA.training.train_starvla][INFO] - Step 1017, Loss: {'train/loss': 0.015691909939050674, 'train/grad_norm_pre_clip': 38.25628662109375, 'timing/data': 0.0003283200057921931, 'timing/model': 0.3657615859992802, 'learning_rate/qwen_vl_interface': 9.202439076491732e-06, 'learning_rate/action_model': 9.202439076491732e-05, 'epoch': 0.07}) +[2026-10-03 14:49:24,104][starVLA.training.train_starvla][INFO] - Step 1018, Loss: {'train/loss': 0.18811668455600739, 'train/grad_norm_pre_clip': 78.95691680908203, 'timing/data': 0.00025821900635492057, 'timing/model': 0.3653002280043438, 'learning_rate/qwen_vl_interface': 9.200749160912208e-06, 'learning_rate/action_model': 9.200749160912208e-05, 'epoch': 0.07}) +[2026-10-03 14:49:24,953][starVLA.training.train_starvla][INFO] - Step 1019, Loss: {'train/loss': 0.0015184428775683045, 'train/grad_norm_pre_clip': 1.9695756435394287, 'timing/data': 0.00020144000882282853, 'timing/model': 0.5639119539991952, 'learning_rate/qwen_vl_interface': 9.199057621328937e-06, 'learning_rate/action_model': 9.199057621328935e-05, 'epoch': 0.07}) +[2026-10-03 14:49:25,605][starVLA.training.train_starvla][INFO] - Step 1020, Loss: {'train/loss': 0.006198110058903694, 'train/grad_norm_pre_clip': 14.73672103881836, 'timing/data': 0.00020261999452486634, 'timing/model': 0.36688137700548396, 'learning_rate/qwen_vl_interface': 9.197364458437244e-06, 'learning_rate/action_model': 9.197364458437245e-05, 'epoch': 0.07}) +[2026-10-03 14:49:26,256][starVLA.training.train_starvla][INFO] - Step 1021, Loss: {'train/loss': 0.009682547301054, 'train/grad_norm_pre_clip': 22.200828552246094, 'timing/data': 0.00021379999816417694, 'timing/model': 0.3664251460140804, 'learning_rate/qwen_vl_interface': 9.19566967293313e-06, 'learning_rate/action_model': 9.19566967293313e-05, 'epoch': 0.07}) +[2026-10-03 14:49:26,907][starVLA.training.train_starvla][INFO] - Step 1022, Loss: {'train/loss': 0.0002813065075315535, 'train/grad_norm_pre_clip': 0.07146720588207245, 'timing/data': 0.00028961000498384237, 'timing/model': 0.36653782799839973, 'learning_rate/qwen_vl_interface': 9.193973265513253e-06, 'learning_rate/action_model': 9.193973265513253e-05, 'epoch': 0.07}) +[2026-10-03 14:49:27,558][starVLA.training.train_starvla][INFO] - Step 1023, Loss: {'train/loss': 0.26037856936454773, 'train/grad_norm_pre_clip': 67.74990844726562, 'timing/data': 0.0002030400064541027, 'timing/model': 0.3670489770011045, 'learning_rate/qwen_vl_interface': 9.192275236874946e-06, 'learning_rate/action_model': 9.192275236874946e-05, 'epoch': 0.07}) +[2026-10-03 14:49:28,188][starVLA.training.train_starvla][INFO] - Step 1024, Loss: {'train/loss': 0.0005209548980928957, 'train/grad_norm_pre_clip': 0.12950803339481354, 'timing/data': 0.0003022599994437769, 'timing/model': 0.36684289599361364, 'learning_rate/qwen_vl_interface': 9.190575587716199e-06, 'learning_rate/action_model': 9.190575587716199e-05, 'epoch': 0.07}) +[2026-10-03 14:49:29,036][starVLA.training.train_starvla][INFO] - Step 1025, Loss: {'train/loss': 0.007755660451948643, 'train/grad_norm_pre_clip': 4.299933910369873, 'timing/data': 0.000310860006720759, 'timing/model': 0.3659629560133908, 'learning_rate/qwen_vl_interface': 9.188874318735677e-06, 'learning_rate/action_model': 9.188874318735676e-05, 'epoch': 0.08}) +[2026-10-03 14:49:29,687][starVLA.training.train_starvla][INFO] - Step 1026, Loss: {'train/loss': 0.043158624321222305, 'train/grad_norm_pre_clip': 35.74431610107422, 'timing/data': 0.00020080999820493162, 'timing/model': 0.36673936700390186, 'learning_rate/qwen_vl_interface': 9.187171430632707e-06, 'learning_rate/action_model': 9.187171430632707e-05, 'epoch': 0.08}) +[2026-10-03 14:49:30,338][starVLA.training.train_starvla][INFO] - Step 1027, Loss: {'train/loss': 0.0005628734943456948, 'train/grad_norm_pre_clip': 0.25649479031562805, 'timing/data': 0.00030053000955376774, 'timing/model': 0.3668676980014425, 'learning_rate/qwen_vl_interface': 9.185466924107283e-06, 'learning_rate/action_model': 9.185466924107282e-05, 'epoch': 0.08}) +[2026-10-03 14:49:30,990][starVLA.training.train_starvla][INFO] - Step 1028, Loss: {'train/loss': 0.8138591051101685, 'train/grad_norm_pre_clip': 160.88270568847656, 'timing/data': 0.00029210999491624534, 'timing/model': 0.3668886060040677, 'learning_rate/qwen_vl_interface': 9.18376079986006e-06, 'learning_rate/action_model': 9.183760799860061e-05, 'epoch': 0.08}) +[2026-10-03 14:49:31,641][starVLA.training.train_starvla][INFO] - Step 1029, Loss: {'train/loss': 0.0005828134599141777, 'train/grad_norm_pre_clip': 0.2539844512939453, 'timing/data': 0.00031947001116350293, 'timing/model': 0.3661525779898511, 'learning_rate/qwen_vl_interface': 9.182053058592365e-06, 'learning_rate/action_model': 9.182053058592365e-05, 'epoch': 0.08}) +[2026-10-03 14:49:32,488][starVLA.training.train_starvla][INFO] - Step 1030, Loss: {'train/loss': 0.045918870717287064, 'train/grad_norm_pre_clip': 85.35090637207031, 'timing/data': 0.00019744999008253217, 'timing/model': 0.5636104840086773, 'learning_rate/qwen_vl_interface': 9.180343701006183e-06, 'learning_rate/action_model': 9.180343701006183e-05, 'epoch': 0.08}) +[2026-10-03 14:49:33,116][starVLA.training.train_starvla][INFO] - Step 1031, Loss: {'train/loss': 0.000209222620469518, 'train/grad_norm_pre_clip': 0.06061391159892082, 'timing/data': 0.0002370100119151175, 'timing/model': 0.3665114659961546, 'learning_rate/qwen_vl_interface': 9.17863272780417e-06, 'learning_rate/action_model': 9.178632727804168e-05, 'epoch': 0.08}) +[2026-10-03 14:49:33,745][starVLA.training.train_starvla][INFO] - Step 1032, Loss: {'train/loss': 0.5223486423492432, 'train/grad_norm_pre_clip': 201.2085418701172, 'timing/data': 0.00024272999144159257, 'timing/model': 0.34363483999914024, 'learning_rate/qwen_vl_interface': 9.176920139689638e-06, 'learning_rate/action_model': 9.176920139689637e-05, 'epoch': 0.08}) +[2026-10-03 14:49:34,401][starVLA.training.train_starvla][INFO] - Step 1033, Loss: {'train/loss': 0.20769421756267548, 'train/grad_norm_pre_clip': 70.69249725341797, 'timing/data': 0.00019746000180020928, 'timing/model': 0.37185009599488694, 'learning_rate/qwen_vl_interface': 9.175205937366568e-06, 'learning_rate/action_model': 9.175205937366569e-05, 'epoch': 0.08}) +[2026-10-03 14:49:35,035][starVLA.training.train_starvla][INFO] - Step 1034, Loss: {'train/loss': 0.00663928035646677, 'train/grad_norm_pre_clip': 9.224503517150879, 'timing/data': 0.00019828999938908964, 'timing/model': 0.36999165499582887, 'learning_rate/qwen_vl_interface': 9.173490121539608e-06, 'learning_rate/action_model': 9.173490121539608e-05, 'epoch': 0.08}) +[2026-10-03 14:49:35,692][starVLA.training.train_starvla][INFO] - Step 1035, Loss: {'train/loss': 0.014868024736642838, 'train/grad_norm_pre_clip': 46.04323196411133, 'timing/data': 0.000295300007564947, 'timing/model': 0.37081039599434007, 'learning_rate/qwen_vl_interface': 9.171772692914062e-06, 'learning_rate/action_model': 9.171772692914061e-05, 'epoch': 0.08}) +[2026-10-03 14:49:36,546][starVLA.training.train_starvla][INFO] - Step 1036, Loss: {'train/loss': 0.0002236741129308939, 'train/grad_norm_pre_clip': 0.06609740108251572, 'timing/data': 0.00026960000104736537, 'timing/model': 0.3708625869912794, 'learning_rate/qwen_vl_interface': 9.1700536521959e-06, 'learning_rate/action_model': 9.170053652195899e-05, 'epoch': 0.08}) +[2026-10-03 14:49:37,202][starVLA.training.train_starvla][INFO] - Step 1037, Loss: {'train/loss': 0.014641190879046917, 'train/grad_norm_pre_clip': 28.32918357849121, 'timing/data': 0.00021643000945914537, 'timing/model': 0.36912808699707966, 'learning_rate/qwen_vl_interface': 9.168333000091754e-06, 'learning_rate/action_model': 9.168333000091754e-05, 'epoch': 0.08}) +[2026-10-03 14:49:37,858][starVLA.training.train_starvla][INFO] - Step 1038, Loss: {'train/loss': 0.00014892866602167487, 'train/grad_norm_pre_clip': 0.08469745516777039, 'timing/data': 0.00023607999901287258, 'timing/model': 0.3700461260013981, 'learning_rate/qwen_vl_interface': 9.16661073730892e-06, 'learning_rate/action_model': 9.166610737308919e-05, 'epoch': 0.08}) +[2026-10-03 14:49:38,516][starVLA.training.train_starvla][INFO] - Step 1039, Loss: {'train/loss': 0.0004897488397546113, 'train/grad_norm_pre_clip': 0.5046167373657227, 'timing/data': 0.0003053899999940768, 'timing/model': 0.37056811701040715, 'learning_rate/qwen_vl_interface': 9.164886864555355e-06, 'learning_rate/action_model': 9.164886864555355e-05, 'epoch': 0.08}) +[2026-10-03 14:49:39,174][starVLA.training.train_starvla][INFO] - Step 1040, Loss: {'train/loss': 0.382830947637558, 'train/grad_norm_pre_clip': 110.15870666503906, 'timing/data': 0.00023586000315845013, 'timing/model': 0.3707925350026926, 'learning_rate/qwen_vl_interface': 9.163161382539678e-06, 'learning_rate/action_model': 9.163161382539678e-05, 'epoch': 0.08}) +[2026-10-03 14:49:40,026][starVLA.training.train_starvla][INFO] - Step 1041, Loss: {'train/loss': 0.3218913674354553, 'train/grad_norm_pre_clip': 169.08827209472656, 'timing/data': 0.0003347300080349669, 'timing/model': 0.5641626130091026, 'learning_rate/qwen_vl_interface': 9.16143429197117e-06, 'learning_rate/action_model': 9.16143429197117e-05, 'epoch': 0.08}) +[2026-10-03 14:49:40,684][starVLA.training.train_starvla][INFO] - Step 1042, Loss: {'train/loss': 1.9713699657586403e-05, 'train/grad_norm_pre_clip': 0.00469545740634203, 'timing/data': 0.00023840999347157776, 'timing/model': 0.3707766869920306, 'learning_rate/qwen_vl_interface': 9.159705593559773e-06, 'learning_rate/action_model': 9.159705593559773e-05, 'epoch': 0.08}) +[2026-10-03 14:49:41,343][starVLA.training.train_starvla][INFO] - Step 1043, Loss: {'train/loss': 0.627991259098053, 'train/grad_norm_pre_clip': 109.26544189453125, 'timing/data': 0.0002432300098007545, 'timing/model': 0.3705410649999976, 'learning_rate/qwen_vl_interface': 9.157975288016089e-06, 'learning_rate/action_model': 9.157975288016088e-05, 'epoch': 0.08}) +[2026-10-03 14:49:42,000][starVLA.training.train_starvla][INFO] - Step 1044, Loss: {'train/loss': 0.00015879613056313246, 'train/grad_norm_pre_clip': 0.060677774250507355, 'timing/data': 0.0003547999949660152, 'timing/model': 0.37038190600287635, 'learning_rate/qwen_vl_interface': 9.156243376051381e-06, 'learning_rate/action_model': 9.156243376051381e-05, 'epoch': 0.08}) +[2026-10-03 14:49:42,656][starVLA.training.train_starvla][INFO] - Step 1045, Loss: {'train/loss': 0.313400000333786, 'train/grad_norm_pre_clip': 153.58677673339844, 'timing/data': 0.00019706999592017382, 'timing/model': 0.3686863469920354, 'learning_rate/qwen_vl_interface': 9.154509858377574e-06, 'learning_rate/action_model': 9.154509858377574e-05, 'epoch': 0.08}) +[2026-10-03 14:49:43,478][starVLA.training.train_starvla][INFO] - Step 1046, Loss: {'train/loss': 0.00010906732495641336, 'train/grad_norm_pre_clip': 0.04235239699482918, 'timing/data': 0.0001952399907168001, 'timing/model': 0.53516329699778, 'learning_rate/qwen_vl_interface': 9.152774735707248e-06, 'learning_rate/action_model': 9.152774735707247e-05, 'epoch': 0.08}) +[2026-10-03 14:49:44,136][starVLA.training.train_starvla][INFO] - Step 1047, Loss: {'train/loss': 0.02333913743495941, 'train/grad_norm_pre_clip': 101.18472290039062, 'timing/data': 0.00023500999668613076, 'timing/model': 0.37085671500244644, 'learning_rate/qwen_vl_interface': 9.15103800875365e-06, 'learning_rate/action_model': 9.15103800875365e-05, 'epoch': 0.08}) +[2026-10-03 14:49:44,771][starVLA.training.train_starvla][INFO] - Step 1048, Loss: {'train/loss': 1.884718656539917, 'train/grad_norm_pre_clip': 94.17450714111328, 'timing/data': 0.00023944000713527203, 'timing/model': 0.3708173369959695, 'learning_rate/qwen_vl_interface': 9.149299678230681e-06, 'learning_rate/action_model': 9.149299678230681e-05, 'epoch': 0.08}) +[2026-10-03 14:49:45,428][starVLA.training.train_starvla][INFO] - Step 1049, Loss: {'train/loss': 0.3793070316314697, 'train/grad_norm_pre_clip': 44.3088264465332, 'timing/data': 0.00032835999445524067, 'timing/model': 0.3687748369993642, 'learning_rate/qwen_vl_interface': 9.147559744852905e-06, 'learning_rate/action_model': 9.147559744852905e-05, 'epoch': 0.08}) +[2026-10-03 14:49:46,085][starVLA.training.train_starvla][INFO] - Step 1050, Loss: {'train/loss': 0.6224389672279358, 'train/grad_norm_pre_clip': 134.212158203125, 'timing/data': 0.00027501999284140766, 'timing/model': 0.3698199369973736, 'learning_rate/qwen_vl_interface': 9.14581820933554e-06, 'learning_rate/action_model': 9.14581820933554e-05, 'epoch': 0.08}) +[2026-10-03 14:49:46,720][starVLA.training.train_starvla][INFO] - Step 1051, Loss: {'train/loss': 6.484514597104862e-05, 'train/grad_norm_pre_clip': 0.009886637330055237, 'timing/data': 0.0002450299944030121, 'timing/model': 0.37064734699379187, 'learning_rate/qwen_vl_interface': 9.14407507239447e-06, 'learning_rate/action_model': 9.14407507239447e-05, 'epoch': 0.08}) +[2026-10-03 14:49:47,573][starVLA.training.train_starvla][INFO] - Step 1052, Loss: {'train/loss': 0.479129821062088, 'train/grad_norm_pre_clip': 137.43983459472656, 'timing/data': 0.00022093999723438174, 'timing/model': 0.37082859700603876, 'learning_rate/qwen_vl_interface': 9.142330334746227e-06, 'learning_rate/action_model': 9.142330334746227e-05, 'epoch': 0.08}) +[2026-10-03 14:49:48,230][starVLA.training.train_starvla][INFO] - Step 1053, Loss: {'train/loss': 0.5480715036392212, 'train/grad_norm_pre_clip': 23.661088943481445, 'timing/data': 0.0002502699935575947, 'timing/model': 0.36924569500843063, 'learning_rate/qwen_vl_interface': 9.14058399710801e-06, 'learning_rate/action_model': 9.14058399710801e-05, 'epoch': 0.08}) +[2026-10-03 14:49:48,887][starVLA.training.train_starvla][INFO] - Step 1054, Loss: {'train/loss': 0.006988514680415392, 'train/grad_norm_pre_clip': 6.043398380279541, 'timing/data': 0.000288889990770258, 'timing/model': 0.3701743260025978, 'learning_rate/qwen_vl_interface': 9.138836060197673e-06, 'learning_rate/action_model': 9.138836060197672e-05, 'epoch': 0.08}) +[2026-10-03 14:49:49,522][starVLA.training.train_starvla][INFO] - Step 1055, Loss: {'train/loss': 0.004361896310001612, 'train/grad_norm_pre_clip': 6.5375261306762695, 'timing/data': 0.00023918999067973346, 'timing/model': 0.34681766999710817, 'learning_rate/qwen_vl_interface': 9.137086524733725e-06, 'learning_rate/action_model': 9.137086524733724e-05, 'epoch': 0.08}) +[2026-10-03 14:49:50,158][starVLA.training.train_starvla][INFO] - Step 1056, Loss: {'train/loss': 0.2834131419658661, 'train/grad_norm_pre_clip': 46.21072006225586, 'timing/data': 0.0003050399973290041, 'timing/model': 0.3473311289999401, 'learning_rate/qwen_vl_interface': 9.135335391435336e-06, 'learning_rate/action_model': 9.135335391435335e-05, 'epoch': 0.08}) +[2026-10-03 14:49:51,001][starVLA.training.train_starvla][INFO] - Step 1057, Loss: {'train/loss': 0.5283669233322144, 'train/grad_norm_pre_clip': 162.9891357421875, 'timing/data': 0.00030189999961294234, 'timing/model': 0.555098384997109, 'learning_rate/qwen_vl_interface': 9.13358266102233e-06, 'learning_rate/action_model': 9.133582661022328e-05, 'epoch': 0.08}) +[2026-10-03 14:49:51,658][starVLA.training.train_starvla][INFO] - Step 1058, Loss: {'train/loss': 0.15570884943008423, 'train/grad_norm_pre_clip': 58.52938461303711, 'timing/data': 0.00028888000815641135, 'timing/model': 0.3695721670083003, 'learning_rate/qwen_vl_interface': 9.131828334215185e-06, 'learning_rate/action_model': 9.131828334215185e-05, 'epoch': 0.08}) +[2026-10-03 14:49:52,315][starVLA.training.train_starvla][INFO] - Step 1059, Loss: {'train/loss': 0.0010591543978080153, 'train/grad_norm_pre_clip': 0.15992602705955505, 'timing/data': 0.000304619999951683, 'timing/model': 0.36993691499810666, 'learning_rate/qwen_vl_interface': 9.130072411735043e-06, 'learning_rate/action_model': 9.130072411735043e-05, 'epoch': 0.08}) +[2026-10-03 14:49:52,973][starVLA.training.train_starvla][INFO] - Step 1060, Loss: {'train/loss': 0.4239218831062317, 'train/grad_norm_pre_clip': 53.0905647277832, 'timing/data': 0.00022431000252254307, 'timing/model': 0.37082718599413056, 'learning_rate/qwen_vl_interface': 9.128314894303693e-06, 'learning_rate/action_model': 9.128314894303693e-05, 'epoch': 0.08}) +[2026-10-03 14:49:53,630][starVLA.training.train_starvla][INFO] - Step 1061, Loss: {'train/loss': 0.2680538594722748, 'train/grad_norm_pre_clip': 93.2176742553711, 'timing/data': 0.00032202999864239246, 'timing/model': 0.3695622860104777, 'learning_rate/qwen_vl_interface': 9.126555782643592e-06, 'learning_rate/action_model': 9.126555782643591e-05, 'epoch': 0.08}) +[2026-10-03 14:49:54,483][starVLA.training.train_starvla][INFO] - Step 1062, Loss: {'train/loss': 0.3983398675918579, 'train/grad_norm_pre_clip': 73.50696563720703, 'timing/data': 0.0002471799962222576, 'timing/model': 0.5651295139978174, 'learning_rate/qwen_vl_interface': 9.124795077477835e-06, 'learning_rate/action_model': 9.124795077477834e-05, 'epoch': 0.08}) +[2026-10-03 14:49:55,140][starVLA.training.train_starvla][INFO] - Step 1063, Loss: {'train/loss': 0.0020851970184594393, 'train/grad_norm_pre_clip': 0.42061877250671387, 'timing/data': 0.0002861200046027079, 'timing/model': 0.37021467699378263, 'learning_rate/qwen_vl_interface': 9.123032779530186e-06, 'learning_rate/action_model': 9.123032779530187e-05, 'epoch': 0.08}) +[2026-10-03 14:49:55,799][starVLA.training.train_starvla][INFO] - Step 1064, Loss: {'train/loss': 0.002066501881927252, 'train/grad_norm_pre_clip': 0.32182738184928894, 'timing/data': 0.0002653900010045618, 'timing/model': 0.3709678949962836, 'learning_rate/qwen_vl_interface': 9.12126888952506e-06, 'learning_rate/action_model': 9.12126888952506e-05, 'epoch': 0.08}) +[2026-10-03 14:49:56,455][starVLA.training.train_starvla][INFO] - Step 1065, Loss: {'train/loss': 0.011485819704830647, 'train/grad_norm_pre_clip': 14.544840812683105, 'timing/data': 0.0003348399914102629, 'timing/model': 0.3683733170037158, 'learning_rate/qwen_vl_interface': 9.119503408187521e-06, 'learning_rate/action_model': 9.119503408187522e-05, 'epoch': 0.08}) +[2026-10-03 14:49:57,112][starVLA.training.train_starvla][INFO] - Step 1066, Loss: {'train/loss': 0.10992997884750366, 'train/grad_norm_pre_clip': 52.767127990722656, 'timing/data': 0.0002509099867893383, 'timing/model': 0.3700294669979485, 'learning_rate/qwen_vl_interface': 9.117736336243299e-06, 'learning_rate/action_model': 9.117736336243298e-05, 'epoch': 0.08}) +[2026-10-03 14:49:57,771][starVLA.training.train_starvla][INFO] - Step 1067, Loss: {'train/loss': 0.4698653817176819, 'train/grad_norm_pre_clip': 50.19789123535156, 'timing/data': 0.00024342999677173793, 'timing/model': 0.3705554860061966, 'learning_rate/qwen_vl_interface': 9.115967674418765e-06, 'learning_rate/action_model': 9.115967674418765e-05, 'epoch': 0.08}) +[2026-10-03 14:49:58,618][starVLA.training.train_starvla][INFO] - Step 1068, Loss: {'train/loss': 0.01718207076191902, 'train/grad_norm_pre_clip': 15.066251754760742, 'timing/data': 0.00026434000756125897, 'timing/model': 0.3712522470013937, 'learning_rate/qwen_vl_interface': 9.114197423440949e-06, 'learning_rate/action_model': 9.11419742344095e-05, 'epoch': 0.08}) +[2026-10-03 14:49:59,275][starVLA.training.train_starvla][INFO] - Step 1069, Loss: {'train/loss': 0.1741970330476761, 'train/grad_norm_pre_clip': 79.63282775878906, 'timing/data': 0.00019643999985419214, 'timing/model': 0.368905467010336, 'learning_rate/qwen_vl_interface': 9.112425584037536e-06, 'learning_rate/action_model': 9.112425584037536e-05, 'epoch': 0.08}) +[2026-10-03 14:49:59,932][starVLA.training.train_starvla][INFO] - Step 1070, Loss: {'train/loss': 0.0032990481704473495, 'train/grad_norm_pre_clip': 1.6363059282302856, 'timing/data': 0.0002940900012617931, 'timing/model': 0.3701563759968849, 'learning_rate/qwen_vl_interface': 9.110652156936863e-06, 'learning_rate/action_model': 9.110652156936863e-05, 'epoch': 0.08}) +[2026-10-03 14:50:00,590][starVLA.training.train_starvla][INFO] - Step 1071, Loss: {'train/loss': 0.1853165328502655, 'train/grad_norm_pre_clip': 109.30706787109375, 'timing/data': 0.0002476800000295043, 'timing/model': 0.37066977701033466, 'learning_rate/qwen_vl_interface': 9.108877142867918e-06, 'learning_rate/action_model': 9.108877142867918e-05, 'epoch': 0.08}) +[2026-10-03 14:50:01,249][starVLA.training.train_starvla][INFO] - Step 1072, Loss: {'train/loss': 0.0036164596676826477, 'train/grad_norm_pre_clip': 2.1796419620513916, 'timing/data': 0.00019551000150386244, 'timing/model': 0.3710713949985802, 'learning_rate/qwen_vl_interface': 9.107100542560342e-06, 'learning_rate/action_model': 9.107100542560342e-05, 'epoch': 0.08}) +[2026-10-03 14:50:02,103][starVLA.training.train_starvla][INFO] - Step 1073, Loss: {'train/loss': 0.5372785329818726, 'train/grad_norm_pre_clip': 74.63162231445312, 'timing/data': 0.000328759997501038, 'timing/model': 0.36858386600215454, 'learning_rate/qwen_vl_interface': 9.105322356744429e-06, 'learning_rate/action_model': 9.105322356744429e-05, 'epoch': 0.08}) +[2026-10-03 14:50:02,761][starVLA.training.train_starvla][INFO] - Step 1074, Loss: {'train/loss': 0.444642573595047, 'train/grad_norm_pre_clip': 52.812461853027344, 'timing/data': 0.0002202700125053525, 'timing/model': 0.3708026460080873, 'learning_rate/qwen_vl_interface': 9.103542586151121e-06, 'learning_rate/action_model': 9.10354258615112e-05, 'epoch': 0.08}) +[2026-10-03 14:50:03,419][starVLA.training.train_starvla][INFO] - Step 1075, Loss: {'train/loss': 0.0006681567174382508, 'train/grad_norm_pre_clip': 0.30393195152282715, 'timing/data': 0.00029507999715860933, 'timing/model': 0.3703406550048385, 'learning_rate/qwen_vl_interface': 9.101761231512017e-06, 'learning_rate/action_model': 9.101761231512017e-05, 'epoch': 0.08}) +[2026-10-03 14:50:04,076][starVLA.training.train_starvla][INFO] - Step 1076, Loss: {'train/loss': 0.16132554411888123, 'train/grad_norm_pre_clip': 85.08712005615234, 'timing/data': 0.00020002000383101404, 'timing/model': 0.37054699700092897, 'learning_rate/qwen_vl_interface': 9.099978293559365e-06, 'learning_rate/action_model': 9.099978293559366e-05, 'epoch': 0.08}) +[2026-10-03 14:50:04,733][starVLA.training.train_starvla][INFO] - Step 1077, Loss: {'train/loss': 0.31566599011421204, 'train/grad_norm_pre_clip': 46.01083755493164, 'timing/data': 0.0003090800018981099, 'timing/model': 0.3688209660031134, 'learning_rate/qwen_vl_interface': 9.098193773026065e-06, 'learning_rate/action_model': 9.098193773026064e-05, 'epoch': 0.08}) +[2026-10-03 14:50:05,590][starVLA.training.train_starvla][INFO] - Step 1078, Loss: {'train/loss': 0.35282719135284424, 'train/grad_norm_pre_clip': 61.25851058959961, 'timing/data': 0.00024371000472456217, 'timing/model': 0.570156683010282, 'learning_rate/qwen_vl_interface': 9.096407670645663e-06, 'learning_rate/action_model': 9.096407670645662e-05, 'epoch': 0.08}) +[2026-10-03 14:50:06,248][starVLA.training.train_starvla][INFO] - Step 1079, Loss: {'train/loss': 1.0517675876617432, 'train/grad_norm_pre_clip': 231.44569396972656, 'timing/data': 0.00026051000168081373, 'timing/model': 0.3707674470060738, 'learning_rate/qwen_vl_interface': 9.094619987152357e-06, 'learning_rate/action_model': 9.094619987152358e-05, 'epoch': 0.08}) +[2026-10-03 14:50:06,907][starVLA.training.train_starvla][INFO] - Step 1080, Loss: {'train/loss': 0.0006517563597299159, 'train/grad_norm_pre_clip': 0.13778430223464966, 'timing/data': 0.0002840599918272346, 'timing/model': 0.371015124997939, 'learning_rate/qwen_vl_interface': 9.092830723281e-06, 'learning_rate/action_model': 9.092830723281e-05, 'epoch': 0.08}) +[2026-10-03 14:50:07,562][starVLA.training.train_starvla][INFO] - Step 1081, Loss: {'train/loss': 0.015141311101615429, 'train/grad_norm_pre_clip': 22.21809959411621, 'timing/data': 0.00020218000281602144, 'timing/model': 0.3687777070008451, 'learning_rate/qwen_vl_interface': 9.09103987976709e-06, 'learning_rate/action_model': 9.09103987976709e-05, 'epoch': 0.08}) +[2026-10-03 14:50:08,219][starVLA.training.train_starvla][INFO] - Step 1082, Loss: {'train/loss': 0.0022140583023428917, 'train/grad_norm_pre_clip': 0.9305716156959534, 'timing/data': 0.00026471000455785543, 'timing/model': 0.36998508698889054, 'learning_rate/qwen_vl_interface': 9.089247457346774e-06, 'learning_rate/action_model': 9.089247457346773e-05, 'epoch': 0.08}) +[2026-10-03 14:50:08,877][starVLA.training.train_starvla][INFO] - Step 1083, Loss: {'train/loss': 0.005722734145820141, 'train/grad_norm_pre_clip': 6.263409614562988, 'timing/data': 0.0002877699880627915, 'timing/model': 0.37011064699618146, 'learning_rate/qwen_vl_interface': 9.087453456756849e-06, 'learning_rate/action_model': 9.087453456756849e-05, 'epoch': 0.08}) +[2026-10-03 14:50:09,720][starVLA.training.train_starvla][INFO] - Step 1084, Loss: {'train/loss': 0.00020127468451391906, 'train/grad_norm_pre_clip': 0.0434875525534153, 'timing/data': 0.0002682699996512383, 'timing/model': 0.3708799070009263, 'learning_rate/qwen_vl_interface': 9.085657878734759e-06, 'learning_rate/action_model': 9.085657878734758e-05, 'epoch': 0.08}) +[2026-10-03 14:50:10,376][starVLA.training.train_starvla][INFO] - Step 1085, Loss: {'train/loss': 0.11198399215936661, 'train/grad_norm_pre_clip': 72.78060150146484, 'timing/data': 0.00025353000091854483, 'timing/model': 0.3686933270073496, 'learning_rate/qwen_vl_interface': 9.083860724018603e-06, 'learning_rate/action_model': 9.083860724018602e-05, 'epoch': 0.08}) +[2026-10-03 14:50:11,033][starVLA.training.train_starvla][INFO] - Step 1086, Loss: {'train/loss': 0.8741498589515686, 'train/grad_norm_pre_clip': 148.98757934570312, 'timing/data': 0.00019647000590339303, 'timing/model': 0.3708413269923767, 'learning_rate/qwen_vl_interface': 9.08206199334712e-06, 'learning_rate/action_model': 9.082061993347119e-05, 'epoch': 0.08}) +[2026-10-03 14:50:11,691][starVLA.training.train_starvla][INFO] - Step 1087, Loss: {'train/loss': 0.561935544013977, 'train/grad_norm_pre_clip': 78.66519165039062, 'timing/data': 0.00023892000899650156, 'timing/model': 0.370219296994037, 'learning_rate/qwen_vl_interface': 9.0802616874597e-06, 'learning_rate/action_model': 9.0802616874597e-05, 'epoch': 0.08}) +[2026-10-03 14:50:12,349][starVLA.training.train_starvla][INFO] - Step 1088, Loss: {'train/loss': 0.002094608498737216, 'train/grad_norm_pre_clip': 1.4434070587158203, 'timing/data': 0.0002501800045138225, 'timing/model': 0.3704919050069293, 'learning_rate/qwen_vl_interface': 9.078459807096385e-06, 'learning_rate/action_model': 9.078459807096385e-05, 'epoch': 0.08}) +[2026-10-03 14:50:13,205][starVLA.training.train_starvla][INFO] - Step 1089, Loss: {'train/loss': 0.2020207941532135, 'train/grad_norm_pre_clip': 111.6681900024414, 'timing/data': 0.0002735500020207837, 'timing/model': 0.36893961700843647, 'learning_rate/qwen_vl_interface': 9.076656352997854e-06, 'learning_rate/action_model': 9.076656352997853e-05, 'epoch': 0.08}) +[2026-10-03 14:50:13,863][starVLA.training.train_starvla][INFO] - Step 1090, Loss: {'train/loss': 0.0006265393458306789, 'train/grad_norm_pre_clip': 0.1256108582019806, 'timing/data': 0.00023737999435979873, 'timing/model': 0.3703151660010917, 'learning_rate/qwen_vl_interface': 9.074851325905443e-06, 'learning_rate/action_model': 9.074851325905442e-05, 'epoch': 0.08}) +[2026-10-03 14:50:14,520][starVLA.training.train_starvla][INFO] - Step 1091, Loss: {'train/loss': 0.0001511359732830897, 'train/grad_norm_pre_clip': 0.03436015173792839, 'timing/data': 0.00024678999034222215, 'timing/model': 0.37047409600927494, 'learning_rate/qwen_vl_interface': 9.073044726561128e-06, 'learning_rate/action_model': 9.073044726561129e-05, 'epoch': 0.08}) +[2026-10-03 14:50:15,179][starVLA.training.train_starvla][INFO] - Step 1092, Loss: {'train/loss': 0.9352523684501648, 'train/grad_norm_pre_clip': 132.843505859375, 'timing/data': 0.00020168999617453665, 'timing/model': 0.371323706000112, 'learning_rate/qwen_vl_interface': 9.071236555707536e-06, 'learning_rate/action_model': 9.071236555707537e-05, 'epoch': 0.08}) +[2026-10-03 14:50:15,812][starVLA.training.train_starvla][INFO] - Step 1093, Loss: {'train/loss': 0.0042516859248280525, 'train/grad_norm_pre_clip': 6.158367156982422, 'timing/data': 0.00033106000046245754, 'timing/model': 0.3458475080115022, 'learning_rate/qwen_vl_interface': 9.069426814087937e-06, 'learning_rate/action_model': 9.069426814087937e-05, 'epoch': 0.08}) +[2026-10-03 14:50:16,632][starVLA.training.train_starvla][INFO] - Step 1094, Loss: {'train/loss': 0.003116415347903967, 'train/grad_norm_pre_clip': 3.118276357650757, 'timing/data': 0.00024055999529082328, 'timing/model': 0.5317465269909007, 'learning_rate/qwen_vl_interface': 9.067615502446247e-06, 'learning_rate/action_model': 9.067615502446246e-05, 'epoch': 0.08}) +[2026-10-03 14:50:17,289][starVLA.training.train_starvla][INFO] - Step 1095, Loss: {'train/loss': 0.00565440533682704, 'train/grad_norm_pre_clip': 5.404173851013184, 'timing/data': 0.00021333999757189304, 'timing/model': 0.37049834498611744, 'learning_rate/qwen_vl_interface': 9.06580262152703e-06, 'learning_rate/action_model': 9.065802621527028e-05, 'epoch': 0.08}) +[2026-10-03 14:50:17,948][starVLA.training.train_starvla][INFO] - Step 1096, Loss: {'train/loss': 0.000384200771804899, 'train/grad_norm_pre_clip': 0.18813954293727875, 'timing/data': 0.00020696999854408205, 'timing/model': 0.3712324950029142, 'learning_rate/qwen_vl_interface': 9.063988172075486e-06, 'learning_rate/action_model': 9.063988172075486e-05, 'epoch': 0.08}) +[2026-10-03 14:50:18,604][starVLA.training.train_starvla][INFO] - Step 1097, Loss: {'train/loss': 0.012037561275064945, 'train/grad_norm_pre_clip': 40.282508850097656, 'timing/data': 0.00020092999329790473, 'timing/model': 0.3690013059967896, 'learning_rate/qwen_vl_interface': 9.062172154837475e-06, 'learning_rate/action_model': 9.062172154837475e-05, 'epoch': 0.08}) +[2026-10-03 14:50:19,260][starVLA.training.train_starvla][INFO] - Step 1098, Loss: {'train/loss': 1.0692569017410278, 'train/grad_norm_pre_clip': 169.1259307861328, 'timing/data': 0.00024357999791391194, 'timing/model': 0.36964487499790266, 'learning_rate/qwen_vl_interface': 9.060354570559487e-06, 'learning_rate/action_model': 9.060354570559487e-05, 'epoch': 0.08}) +[2026-10-03 14:50:19,895][starVLA.training.train_starvla][INFO] - Step 1099, Loss: {'train/loss': 1.2324860095977783, 'train/grad_norm_pre_clip': 59.316524505615234, 'timing/data': 0.00028716999804601073, 'timing/model': 0.3468258490029257, 'learning_rate/qwen_vl_interface': 9.058535419988667e-06, 'learning_rate/action_model': 9.058535419988667e-05, 'epoch': 0.08}) +[2026-10-03 14:50:20,748][starVLA.training.train_starvla][INFO] - Step 1100, Loss: {'train/loss': 1.5689773559570312, 'train/grad_norm_pre_clip': 52.40778350830078, 'timing/data': 0.0002639199956320226, 'timing/model': 0.3714072659931844, 'learning_rate/qwen_vl_interface': 9.056714703872794e-06, 'learning_rate/action_model': 9.056714703872794e-05, 'epoch': 0.08}) +[2026-10-03 14:50:21,404][starVLA.training.train_starvla][INFO] - Step 1101, Loss: {'train/loss': 1.0695788860321045, 'train/grad_norm_pre_clip': 190.4739227294922, 'timing/data': 0.00032560000545345247, 'timing/model': 0.3684470570005942, 'learning_rate/qwen_vl_interface': 9.0548924229603e-06, 'learning_rate/action_model': 9.054892422960301e-05, 'epoch': 0.08}) +[2026-10-03 14:50:22,038][starVLA.training.train_starvla][INFO] - Step 1102, Loss: {'train/loss': 0.6934989094734192, 'train/grad_norm_pre_clip': 89.962890625, 'timing/data': 0.00025004999770317227, 'timing/model': 0.3467761689971667, 'learning_rate/qwen_vl_interface': 9.053068578000256e-06, 'learning_rate/action_model': 9.053068578000256e-05, 'epoch': 0.08}) +[2026-10-03 14:50:22,695][starVLA.training.train_starvla][INFO] - Step 1103, Loss: {'train/loss': 0.011180443689227104, 'train/grad_norm_pre_clip': 20.193857192993164, 'timing/data': 0.00033737000194378197, 'timing/model': 0.3703162759920815, 'learning_rate/qwen_vl_interface': 9.051243169742372e-06, 'learning_rate/action_model': 9.051243169742372e-05, 'epoch': 0.08}) +[2026-10-03 14:50:23,353][starVLA.training.train_starvla][INFO] - Step 1104, Loss: {'train/loss': 0.008243158459663391, 'train/grad_norm_pre_clip': 14.488546371459961, 'timing/data': 0.00029642999288626015, 'timing/model': 0.3709729660040466, 'learning_rate/qwen_vl_interface': 9.049416198937008e-06, 'learning_rate/action_model': 9.049416198937007e-05, 'epoch': 0.08}) +[2026-10-03 14:50:24,206][starVLA.training.train_starvla][INFO] - Step 1105, Loss: {'train/loss': 0.9973068833351135, 'train/grad_norm_pre_clip': 91.1435775756836, 'timing/data': 0.00039740999636705965, 'timing/model': 0.36930841699359007, 'learning_rate/qwen_vl_interface': 9.047587666335159e-06, 'learning_rate/action_model': 9.047587666335159e-05, 'epoch': 0.08}) +[2026-10-03 14:50:24,862][starVLA.training.train_starvla][INFO] - Step 1106, Loss: {'train/loss': 0.4925830662250519, 'train/grad_norm_pre_clip': 85.050537109375, 'timing/data': 0.0002382300008321181, 'timing/model': 0.370333307000692, 'learning_rate/qwen_vl_interface': 9.045757572688469e-06, 'learning_rate/action_model': 9.04575757268847e-05, 'epoch': 0.08}) +[2026-10-03 14:50:25,520][starVLA.training.train_starvla][INFO] - Step 1107, Loss: {'train/loss': 0.0609978623688221, 'train/grad_norm_pre_clip': 46.92506790161133, 'timing/data': 0.00025329999334644526, 'timing/model': 0.37084699601109605, 'learning_rate/qwen_vl_interface': 9.043925918749218e-06, 'learning_rate/action_model': 9.043925918749218e-05, 'epoch': 0.08}) +[2026-10-03 14:50:26,180][starVLA.training.train_starvla][INFO] - Step 1108, Loss: {'train/loss': 0.0012963726185262203, 'train/grad_norm_pre_clip': 0.24476824700832367, 'timing/data': 0.0002515700034564361, 'timing/model': 0.3713264359976165, 'learning_rate/qwen_vl_interface': 9.042092705270333e-06, 'learning_rate/action_model': 9.042092705270333e-05, 'epoch': 0.08}) +[2026-10-03 14:50:26,837][starVLA.training.train_starvla][INFO] - Step 1109, Loss: {'train/loss': 0.056619927287101746, 'train/grad_norm_pre_clip': 18.104875564575195, 'timing/data': 0.00024588999804109335, 'timing/model': 0.3688629269890953, 'learning_rate/qwen_vl_interface': 9.040257933005375e-06, 'learning_rate/action_model': 9.040257933005374e-05, 'epoch': 0.08}) +[2026-10-03 14:50:27,689][starVLA.training.train_starvla][INFO] - Step 1110, Loss: {'train/loss': 0.0778011679649353, 'train/grad_norm_pre_clip': 66.05443572998047, 'timing/data': 0.00020886999845970422, 'timing/model': 0.5645505230058916, 'learning_rate/qwen_vl_interface': 9.038421602708552e-06, 'learning_rate/action_model': 9.038421602708552e-05, 'epoch': 0.08}) +[2026-10-03 14:50:28,349][starVLA.training.train_starvla][INFO] - Step 1111, Loss: {'train/loss': 0.5737636089324951, 'train/grad_norm_pre_clip': 83.1611328125, 'timing/data': 0.00024331000167876482, 'timing/model': 0.37278100599360187, 'learning_rate/qwen_vl_interface': 9.036583715134709e-06, 'learning_rate/action_model': 9.036583715134708e-05, 'epoch': 0.08}) +[2026-10-03 14:50:29,008][starVLA.training.train_starvla][INFO] - Step 1112, Loss: {'train/loss': 0.001214767573401332, 'train/grad_norm_pre_clip': 0.24939368665218353, 'timing/data': 0.0002503400028217584, 'timing/model': 0.3712076070078183, 'learning_rate/qwen_vl_interface': 9.034744271039332e-06, 'learning_rate/action_model': 9.034744271039332e-05, 'epoch': 0.08}) +[2026-10-03 14:50:29,665][starVLA.training.train_starvla][INFO] - Step 1113, Loss: {'train/loss': 0.007004124112427235, 'train/grad_norm_pre_clip': 5.3543901443481445, 'timing/data': 0.0003235299955122173, 'timing/model': 0.36855465700500645, 'learning_rate/qwen_vl_interface': 9.03290327117855e-06, 'learning_rate/action_model': 9.032903271178549e-05, 'epoch': 0.08}) +[2026-10-03 14:50:30,324][starVLA.training.train_starvla][INFO] - Step 1114, Loss: {'train/loss': 0.0011857545468956232, 'train/grad_norm_pre_clip': 0.31022611260414124, 'timing/data': 0.00023617899569217116, 'timing/model': 0.37100107698643114, 'learning_rate/qwen_vl_interface': 9.031060716309122e-06, 'learning_rate/action_model': 9.031060716309121e-05, 'epoch': 0.08}) +[2026-10-03 14:50:31,179][starVLA.training.train_starvla][INFO] - Step 1115, Loss: {'train/loss': 0.011964649893343449, 'train/grad_norm_pre_clip': 5.440002918243408, 'timing/data': 0.0003151499986415729, 'timing/model': 0.5671887529897504, 'learning_rate/qwen_vl_interface': 9.02921660718846e-06, 'learning_rate/action_model': 9.029216607188459e-05, 'epoch': 0.08}) +[2026-10-03 14:50:31,837][starVLA.training.train_starvla][INFO] - Step 1116, Loss: {'train/loss': 0.0023836069740355015, 'train/grad_norm_pre_clip': 0.5440636277198792, 'timing/data': 0.00023403999512083828, 'timing/model': 0.3710384160076501, 'learning_rate/qwen_vl_interface': 9.027370944574604e-06, 'learning_rate/action_model': 9.027370944574604e-05, 'epoch': 0.08}) +[2026-10-03 14:50:32,493][starVLA.training.train_starvla][INFO] - Step 1117, Loss: {'train/loss': 0.0012331566540524364, 'train/grad_norm_pre_clip': 0.2357935607433319, 'timing/data': 0.00024091999512165785, 'timing/model': 0.3687915759946918, 'learning_rate/qwen_vl_interface': 9.025523729226236e-06, 'learning_rate/action_model': 9.025523729226236e-05, 'epoch': 0.08}) +[2026-10-03 14:50:33,152][starVLA.training.train_starvla][INFO] - Step 1118, Loss: {'train/loss': 0.25811028480529785, 'train/grad_norm_pre_clip': 47.93315124511719, 'timing/data': 0.00029676000121980906, 'timing/model': 0.3709251359978225, 'learning_rate/qwen_vl_interface': 9.023674961902677e-06, 'learning_rate/action_model': 9.023674961902676e-05, 'epoch': 0.08}) +[2026-10-03 14:50:33,810][starVLA.training.train_starvla][INFO] - Step 1119, Loss: {'train/loss': 0.0007478785119019449, 'train/grad_norm_pre_clip': 0.16918784379959106, 'timing/data': 0.0002802399976644665, 'timing/model': 0.37086434599768836, 'learning_rate/qwen_vl_interface': 9.021824643363883e-06, 'learning_rate/action_model': 9.021824643363883e-05, 'epoch': 0.08}) +[2026-10-03 14:50:34,445][starVLA.training.train_starvla][INFO] - Step 1120, Loss: {'train/loss': 0.0009718272485770285, 'train/grad_norm_pre_clip': 0.3029545843601227, 'timing/data': 0.00022120999346952885, 'timing/model': 0.3476314290019218, 'learning_rate/qwen_vl_interface': 9.019972774370456e-06, 'learning_rate/action_model': 9.019972774370455e-05, 'epoch': 0.08}) +[2026-10-03 14:50:35,295][starVLA.training.train_starvla][INFO] - Step 1121, Loss: {'train/loss': 0.05631791800260544, 'train/grad_norm_pre_clip': 79.17169189453125, 'timing/data': 0.00034466999932192266, 'timing/model': 0.3690646870090859, 'learning_rate/qwen_vl_interface': 9.018119355683623e-06, 'learning_rate/action_model': 9.018119355683623e-05, 'epoch': 0.08}) +[2026-10-03 14:50:35,952][starVLA.training.train_starvla][INFO] - Step 1122, Loss: {'train/loss': 0.15873895585536957, 'train/grad_norm_pre_clip': 58.800559997558594, 'timing/data': 0.00020152000070083886, 'timing/model': 0.3704411560029257, 'learning_rate/qwen_vl_interface': 9.016264388065258e-06, 'learning_rate/action_model': 9.016264388065258e-05, 'epoch': 0.08}) +[2026-10-03 14:50:36,610][starVLA.training.train_starvla][INFO] - Step 1123, Loss: {'train/loss': 0.0015790470642969012, 'train/grad_norm_pre_clip': 1.7340635061264038, 'timing/data': 0.00029332999838516116, 'timing/model': 0.37055266699462663, 'learning_rate/qwen_vl_interface': 9.014407872277869e-06, 'learning_rate/action_model': 9.014407872277869e-05, 'epoch': 0.08}) +[2026-10-03 14:50:37,268][starVLA.training.train_starvla][INFO] - Step 1124, Loss: {'train/loss': 0.0007541406666859984, 'train/grad_norm_pre_clip': 0.5533018112182617, 'timing/data': 0.00027193999267183244, 'timing/model': 0.37042906699934974, 'learning_rate/qwen_vl_interface': 9.012549809084595e-06, 'learning_rate/action_model': 9.012549809084595e-05, 'epoch': 0.08}) +[2026-10-03 14:50:37,925][starVLA.training.train_starvla][INFO] - Step 1125, Loss: {'train/loss': 0.6033138036727905, 'train/grad_norm_pre_clip': 144.44393920898438, 'timing/data': 0.00023790000705048442, 'timing/model': 0.3690086270071333, 'learning_rate/qwen_vl_interface': 9.01069019924922e-06, 'learning_rate/action_model': 9.010690199249219e-05, 'epoch': 0.08}) +[2026-10-03 14:50:38,766][starVLA.training.train_starvla][INFO] - Step 1126, Loss: {'train/loss': 0.36125364899635315, 'train/grad_norm_pre_clip': 286.3260803222656, 'timing/data': 0.00019485999655444175, 'timing/model': 0.5554344049887732, 'learning_rate/qwen_vl_interface': 9.008829043536157e-06, 'learning_rate/action_model': 9.008829043536156e-05, 'epoch': 0.08}) +[2026-10-03 14:50:39,400][starVLA.training.train_starvla][INFO] - Step 1127, Loss: {'train/loss': 0.4198552072048187, 'train/grad_norm_pre_clip': 79.0436019897461, 'timing/data': 0.0002877100050682202, 'timing/model': 0.3473625990009168, 'learning_rate/qwen_vl_interface': 9.006966342710458e-06, 'learning_rate/action_model': 9.006966342710457e-05, 'epoch': 0.08}) +[2026-10-03 14:50:40,059][starVLA.training.train_starvla][INFO] - Step 1128, Loss: {'train/loss': 0.7611222267150879, 'train/grad_norm_pre_clip': 57.42634963989258, 'timing/data': 0.0002874199999496341, 'timing/model': 0.3711097869963851, 'learning_rate/qwen_vl_interface': 9.005102097537809e-06, 'learning_rate/action_model': 9.005102097537809e-05, 'epoch': 0.08}) +[2026-10-03 14:50:40,715][starVLA.training.train_starvla][INFO] - Step 1129, Loss: {'train/loss': 0.19132204353809357, 'train/grad_norm_pre_clip': 70.26161193847656, 'timing/data': 0.00024266999389510602, 'timing/model': 0.36870935700426344, 'learning_rate/qwen_vl_interface': 9.00323630878453e-06, 'learning_rate/action_model': 9.00323630878453e-05, 'epoch': 0.08}) +[2026-10-03 14:50:41,373][starVLA.training.train_starvla][INFO] - Step 1130, Loss: {'train/loss': 0.039476167410612106, 'train/grad_norm_pre_clip': 32.7893180847168, 'timing/data': 0.0002317100006621331, 'timing/model': 0.37011294600961264, 'learning_rate/qwen_vl_interface': 9.001368977217577e-06, 'learning_rate/action_model': 9.001368977217577e-05, 'epoch': 0.08}) +[2026-10-03 14:50:42,220][starVLA.training.train_starvla][INFO] - Step 1131, Loss: {'train/loss': 0.3107828199863434, 'train/grad_norm_pre_clip': 29.166555404663086, 'timing/data': 0.00029178000113461167, 'timing/model': 0.5597050439973827, 'learning_rate/qwen_vl_interface': 8.999500103604538e-06, 'learning_rate/action_model': 8.999500103604536e-05, 'epoch': 0.08}) +[2026-10-03 14:50:42,878][starVLA.training.train_starvla][INFO] - Step 1132, Loss: {'train/loss': 0.05075409635901451, 'train/grad_norm_pre_clip': 32.78403091430664, 'timing/data': 0.00029938000079710037, 'timing/model': 0.3705763160105562, 'learning_rate/qwen_vl_interface': 8.997629688713639e-06, 'learning_rate/action_model': 8.997629688713638e-05, 'epoch': 0.08}) +[2026-10-03 14:50:43,534][starVLA.training.train_starvla][INFO] - Step 1133, Loss: {'train/loss': 0.015884114429354668, 'train/grad_norm_pre_clip': 11.030322074890137, 'timing/data': 0.00021312999888323247, 'timing/model': 0.36944272700930014, 'learning_rate/qwen_vl_interface': 8.995757733313733e-06, 'learning_rate/action_model': 8.995757733313732e-05, 'epoch': 0.08}) +[2026-10-03 14:50:44,169][starVLA.training.train_starvla][INFO] - Step 1134, Loss: {'train/loss': 0.0121703390032053, 'train/grad_norm_pre_clip': 10.485518455505371, 'timing/data': 0.0002497200039215386, 'timing/model': 0.37056722700071987, 'learning_rate/qwen_vl_interface': 8.993884238174314e-06, 'learning_rate/action_model': 8.993884238174314e-05, 'epoch': 0.08}) +[2026-10-03 14:50:44,826][starVLA.training.train_starvla][INFO] - Step 1135, Loss: {'train/loss': 0.0015526596689596772, 'train/grad_norm_pre_clip': 0.4644821584224701, 'timing/data': 0.0002873399935197085, 'timing/model': 0.37035294600354973, 'learning_rate/qwen_vl_interface': 8.992009204065505e-06, 'learning_rate/action_model': 8.992009204065505e-05, 'epoch': 0.08}) +[2026-10-03 14:50:45,680][starVLA.training.train_starvla][INFO] - Step 1136, Loss: {'train/loss': 0.003181064035743475, 'train/grad_norm_pre_clip': 0.8614673614501953, 'timing/data': 0.00029727000219281763, 'timing/model': 0.5661246130039217, 'learning_rate/qwen_vl_interface': 8.990132631758058e-06, 'learning_rate/action_model': 8.990132631758059e-05, 'epoch': 0.08}) +[2026-10-03 14:50:46,335][starVLA.training.train_starvla][INFO] - Step 1137, Loss: {'train/loss': 0.1813001185655594, 'train/grad_norm_pre_clip': 133.30691528320312, 'timing/data': 0.0002196300047216937, 'timing/model': 0.36861783600761555, 'learning_rate/qwen_vl_interface': 8.988254522023364e-06, 'learning_rate/action_model': 8.988254522023363e-05, 'epoch': 0.08}) +[2026-10-03 14:50:46,993][starVLA.training.train_starvla][INFO] - Step 1138, Loss: {'train/loss': 0.022719833999872208, 'train/grad_norm_pre_clip': 20.81949234008789, 'timing/data': 0.00028033000126015395, 'timing/model': 0.37028886599000543, 'learning_rate/qwen_vl_interface': 8.986374875633441e-06, 'learning_rate/action_model': 8.98637487563344e-05, 'epoch': 0.08}) +[2026-10-03 14:50:47,650][starVLA.training.train_starvla][INFO] - Step 1139, Loss: {'train/loss': 0.5695463418960571, 'train/grad_norm_pre_clip': 84.35345458984375, 'timing/data': 0.00020525000581983477, 'timing/model': 0.37003467699105386, 'learning_rate/qwen_vl_interface': 8.984493693360944e-06, 'learning_rate/action_model': 8.984493693360943e-05, 'epoch': 0.08}) +[2026-10-03 14:50:48,307][starVLA.training.train_starvla][INFO] - Step 1140, Loss: {'train/loss': 0.09798753261566162, 'train/grad_norm_pre_clip': 58.860984802246094, 'timing/data': 0.00021673999435734004, 'timing/model': 0.37091197600238957, 'learning_rate/qwen_vl_interface': 8.982610975979153e-06, 'learning_rate/action_model': 8.982610975979152e-05, 'epoch': 0.08}) +[2026-10-03 14:50:48,943][starVLA.training.train_starvla][INFO] - Step 1141, Loss: {'train/loss': 0.058455467224121094, 'train/grad_norm_pre_clip': 49.31856918334961, 'timing/data': 0.0003209100104868412, 'timing/model': 0.34667519900540356, 'learning_rate/qwen_vl_interface': 8.98072672426198e-06, 'learning_rate/action_model': 8.980726724261981e-05, 'epoch': 0.08}) +[2026-10-03 14:50:49,798][starVLA.training.train_starvla][INFO] - Step 1142, Loss: {'train/loss': 0.010735983029007912, 'train/grad_norm_pre_clip': 17.889495849609375, 'timing/data': 0.00023879999935161322, 'timing/model': 0.5679537230025744, 'learning_rate/qwen_vl_interface': 8.978840938983977e-06, 'learning_rate/action_model': 8.978840938983977e-05, 'epoch': 0.08}) +[2026-10-03 14:50:50,457][starVLA.training.train_starvla][INFO] - Step 1143, Loss: {'train/loss': 0.010712525807321072, 'train/grad_norm_pre_clip': 14.935362815856934, 'timing/data': 0.00019817000429611653, 'timing/model': 0.3709916960069677, 'learning_rate/qwen_vl_interface': 8.976953620920313e-06, 'learning_rate/action_model': 8.976953620920312e-05, 'epoch': 0.08}) +[2026-10-03 14:50:51,117][starVLA.training.train_starvla][INFO] - Step 1144, Loss: {'train/loss': 0.06672536581754684, 'train/grad_norm_pre_clip': 69.59648132324219, 'timing/data': 0.00019521999638527632, 'timing/model': 0.3715957660024287, 'learning_rate/qwen_vl_interface': 8.975064770846794e-06, 'learning_rate/action_model': 8.975064770846794e-05, 'epoch': 0.08}) +[2026-10-03 14:50:51,775][starVLA.training.train_starvla][INFO] - Step 1145, Loss: {'train/loss': 0.040253717452287674, 'train/grad_norm_pre_clip': 119.50695037841797, 'timing/data': 0.0002834599872585386, 'timing/model': 0.3695716359943617, 'learning_rate/qwen_vl_interface': 8.973174389539856e-06, 'learning_rate/action_model': 8.973174389539855e-05, 'epoch': 0.08}) +[2026-10-03 14:50:52,434][starVLA.training.train_starvla][INFO] - Step 1146, Loss: {'train/loss': 0.534411609172821, 'train/grad_norm_pre_clip': 17.2572021484375, 'timing/data': 0.0002132900117430836, 'timing/model': 0.3709168060013326, 'learning_rate/qwen_vl_interface': 8.971282477776563e-06, 'learning_rate/action_model': 8.971282477776563e-05, 'epoch': 0.08}) +[2026-10-03 14:50:53,283][starVLA.training.train_starvla][INFO] - Step 1147, Loss: {'train/loss': 0.33209654688835144, 'train/grad_norm_pre_clip': 45.70415496826172, 'timing/data': 0.00019844000053126365, 'timing/model': 0.5610402240126859, 'learning_rate/qwen_vl_interface': 8.969389036334609e-06, 'learning_rate/action_model': 8.969389036334609e-05, 'epoch': 0.08}) +[2026-10-03 14:50:53,919][starVLA.training.train_starvla][INFO] - Step 1148, Loss: {'train/loss': 0.6579644083976746, 'train/grad_norm_pre_clip': 75.28055572509766, 'timing/data': 0.0002745500096352771, 'timing/model': 0.371134525994421, 'learning_rate/qwen_vl_interface': 8.967494065992315e-06, 'learning_rate/action_model': 8.967494065992315e-05, 'epoch': 0.08}) +[2026-10-03 14:50:54,576][starVLA.training.train_starvla][INFO] - Step 1149, Loss: {'train/loss': 0.6177244186401367, 'train/grad_norm_pre_clip': 102.67346954345703, 'timing/data': 0.00031541999487672, 'timing/model': 0.3687364869983867, 'learning_rate/qwen_vl_interface': 8.965597567528634e-06, 'learning_rate/action_model': 8.965597567528633e-05, 'epoch': 0.08}) +[2026-10-03 14:50:55,234][starVLA.training.train_starvla][INFO] - Step 1150, Loss: {'train/loss': 0.31229791045188904, 'train/grad_norm_pre_clip': 157.42091369628906, 'timing/data': 0.0002847600117092952, 'timing/model': 0.37086770599125884, 'learning_rate/qwen_vl_interface': 8.963699541723143e-06, 'learning_rate/action_model': 8.963699541723142e-05, 'epoch': 0.08}) +[2026-10-03 14:50:55,892][starVLA.training.train_starvla][INFO] - Step 1151, Loss: {'train/loss': 0.00020303597557358444, 'train/grad_norm_pre_clip': 0.07506970316171646, 'timing/data': 0.00024105000193230808, 'timing/model': 0.36968620600237045, 'learning_rate/qwen_vl_interface': 8.961799989356048e-06, 'learning_rate/action_model': 8.961799989356048e-05, 'epoch': 0.08}) +[2026-10-03 14:50:56,551][starVLA.training.train_starvla][INFO] - Step 1152, Loss: {'train/loss': 0.44216030836105347, 'train/grad_norm_pre_clip': 59.22420883178711, 'timing/data': 0.0003453800018178299, 'timing/model': 0.3711651659978088, 'learning_rate/qwen_vl_interface': 8.959898911208185e-06, 'learning_rate/action_model': 8.959898911208185e-05, 'epoch': 0.08}) +[2026-10-03 14:50:57,401][starVLA.training.train_starvla][INFO] - Step 1153, Loss: {'train/loss': 0.3926367163658142, 'train/grad_norm_pre_clip': 415.5056457519531, 'timing/data': 0.00038488001155201346, 'timing/model': 0.36929431700264104, 'learning_rate/qwen_vl_interface': 8.957996308061015e-06, 'learning_rate/action_model': 8.957996308061014e-05, 'epoch': 0.08}) +[2026-10-03 14:50:58,057][starVLA.training.train_starvla][INFO] - Step 1154, Loss: {'train/loss': 0.195945605635643, 'train/grad_norm_pre_clip': 23.702266693115234, 'timing/data': 0.00028651999309659004, 'timing/model': 0.3701181259966688, 'learning_rate/qwen_vl_interface': 8.956092180696626e-06, 'learning_rate/action_model': 8.956092180696625e-05, 'epoch': 0.08}) +[2026-10-03 14:50:58,715][starVLA.training.train_starvla][INFO] - Step 1155, Loss: {'train/loss': 0.018719621002674103, 'train/grad_norm_pre_clip': 10.68658447265625, 'timing/data': 0.0002617899881443009, 'timing/model': 0.3703287160024047, 'learning_rate/qwen_vl_interface': 8.95418652989773e-06, 'learning_rate/action_model': 8.95418652989773e-05, 'epoch': 0.08}) +[2026-10-03 14:50:59,373][starVLA.training.train_starvla][INFO] - Step 1156, Loss: {'train/loss': 0.546259880065918, 'train/grad_norm_pre_clip': 27.28406524658203, 'timing/data': 0.00030468999466393143, 'timing/model': 0.37090464601351414, 'learning_rate/qwen_vl_interface': 8.952279356447675e-06, 'learning_rate/action_model': 8.952279356447675e-05, 'epoch': 0.08}) +[2026-10-03 14:51:00,029][starVLA.training.train_starvla][INFO] - Step 1157, Loss: {'train/loss': 0.060191452503204346, 'train/grad_norm_pre_clip': 31.709733963012695, 'timing/data': 0.0002981099969474599, 'timing/model': 0.3692053460108582, 'learning_rate/qwen_vl_interface': 8.950370661130421e-06, 'learning_rate/action_model': 8.950370661130422e-05, 'epoch': 0.08}) +[2026-10-03 14:51:00,875][starVLA.training.train_starvla][INFO] - Step 1158, Loss: {'train/loss': 0.05602472275495529, 'train/grad_norm_pre_clip': 15.631169319152832, 'timing/data': 0.00035663000016938895, 'timing/model': 0.37000848600291647, 'learning_rate/qwen_vl_interface': 8.948460444730566e-06, 'learning_rate/action_model': 8.948460444730566e-05, 'epoch': 0.08}) +[2026-10-03 14:51:01,532][starVLA.training.train_starvla][INFO] - Step 1159, Loss: {'train/loss': 0.00546478945761919, 'train/grad_norm_pre_clip': 1.0607879161834717, 'timing/data': 0.00023050999152474105, 'timing/model': 0.3697032960044453, 'learning_rate/qwen_vl_interface': 8.946548708033324e-06, 'learning_rate/action_model': 8.946548708033323e-05, 'epoch': 0.08}) +[2026-10-03 14:51:02,189][starVLA.training.train_starvla][INFO] - Step 1160, Loss: {'train/loss': 0.028175724670290947, 'train/grad_norm_pre_clip': 21.863574981689453, 'timing/data': 0.0003014499961864203, 'timing/model': 0.3704475059930701, 'learning_rate/qwen_vl_interface': 8.944635451824538e-06, 'learning_rate/action_model': 8.944635451824538e-05, 'epoch': 0.08}) +[2026-10-03 14:51:02,845][starVLA.training.train_starvla][INFO] - Step 1161, Loss: {'train/loss': 0.12197594344615936, 'train/grad_norm_pre_clip': 35.64457702636719, 'timing/data': 0.0003207300032954663, 'timing/model': 0.3687955470086308, 'learning_rate/qwen_vl_interface': 8.942720676890677e-06, 'learning_rate/action_model': 8.942720676890677e-05, 'epoch': 0.08}) +[2026-10-03 14:51:03,503][starVLA.training.train_starvla][INFO] - Step 1162, Loss: {'train/loss': 0.15004831552505493, 'train/grad_norm_pre_clip': 70.2494125366211, 'timing/data': 0.00019953999435529113, 'timing/model': 0.3701877860003151, 'learning_rate/qwen_vl_interface': 8.940804384018833e-06, 'learning_rate/action_model': 8.940804384018832e-05, 'epoch': 0.09}) +[2026-10-03 14:51:04,349][starVLA.training.train_starvla][INFO] - Step 1163, Loss: {'train/loss': 0.0046083154156804085, 'train/grad_norm_pre_clip': 1.331345558166504, 'timing/data': 0.00020219999714754522, 'timing/model': 0.5596425940020708, 'learning_rate/qwen_vl_interface': 8.93888657399672e-06, 'learning_rate/action_model': 8.93888657399672e-05, 'epoch': 0.09}) +[2026-10-03 14:51:05,006][starVLA.training.train_starvla][INFO] - Step 1164, Loss: {'train/loss': 0.3572171926498413, 'train/grad_norm_pre_clip': 25.843585968017578, 'timing/data': 0.00021584000205621123, 'timing/model': 0.37001383599999826, 'learning_rate/qwen_vl_interface': 8.936967247612678e-06, 'learning_rate/action_model': 8.936967247612678e-05, 'epoch': 0.09}) +[2026-10-03 14:51:05,663][starVLA.training.train_starvla][INFO] - Step 1165, Loss: {'train/loss': 0.17342454195022583, 'train/grad_norm_pre_clip': 48.13882827758789, 'timing/data': 0.00032423000084236264, 'timing/model': 0.36865732599108014, 'learning_rate/qwen_vl_interface': 8.93504640565567e-06, 'learning_rate/action_model': 8.93504640565567e-05, 'epoch': 0.09}) +[2026-10-03 14:51:06,322][starVLA.training.train_starvla][INFO] - Step 1166, Loss: {'train/loss': 0.0007552430615760386, 'train/grad_norm_pre_clip': 0.10069059580564499, 'timing/data': 0.00025167899730149657, 'timing/model': 0.3709069469914539, 'learning_rate/qwen_vl_interface': 8.933124048915278e-06, 'learning_rate/action_model': 8.933124048915278e-05, 'epoch': 0.09}) +[2026-10-03 14:51:06,979][starVLA.training.train_starvla][INFO] - Step 1167, Loss: {'train/loss': 0.005076861009001732, 'train/grad_norm_pre_clip': 2.5726213455200195, 'timing/data': 0.000269349999143742, 'timing/model': 0.36962187600147445, 'learning_rate/qwen_vl_interface': 8.931200178181715e-06, 'learning_rate/action_model': 8.931200178181715e-05, 'epoch': 0.09}) +[2026-10-03 14:51:07,808][starVLA.training.train_starvla][INFO] - Step 1168, Loss: {'train/loss': 0.0028014336712658405, 'train/grad_norm_pre_clip': 1.1667152643203735, 'timing/data': 0.00020695000421255827, 'timing/model': 0.5641856430011103, 'learning_rate/qwen_vl_interface': 8.929274794245812e-06, 'learning_rate/action_model': 8.929274794245812e-05, 'epoch': 0.09}) +[2026-10-03 14:51:08,441][starVLA.training.train_starvla][INFO] - Step 1169, Loss: {'train/loss': 0.15241500735282898, 'train/grad_norm_pre_clip': 29.671443939208984, 'timing/data': 0.00032730000384617597, 'timing/model': 0.3690744060004363, 'learning_rate/qwen_vl_interface': 8.927347897899018e-06, 'learning_rate/action_model': 8.927347897899016e-05, 'epoch': 0.09}) +[2026-10-03 14:51:09,097][starVLA.training.train_starvla][INFO] - Step 1170, Loss: {'train/loss': 0.005534653551876545, 'train/grad_norm_pre_clip': 2.881479263305664, 'timing/data': 0.0001989599986700341, 'timing/model': 0.36998535599559546, 'learning_rate/qwen_vl_interface': 8.925419489933407e-06, 'learning_rate/action_model': 8.925419489933408e-05, 'epoch': 0.09}) +[2026-10-03 14:51:09,755][starVLA.training.train_starvla][INFO] - Step 1171, Loss: {'train/loss': 0.03563644737005234, 'train/grad_norm_pre_clip': 53.17621994018555, 'timing/data': 0.000275399987003766, 'timing/model': 0.37020845599181484, 'learning_rate/qwen_vl_interface': 8.923489571141679e-06, 'learning_rate/action_model': 8.923489571141679e-05, 'epoch': 0.09}) +[2026-10-03 14:51:10,413][starVLA.training.train_starvla][INFO] - Step 1172, Loss: {'train/loss': 0.06149901822209358, 'train/grad_norm_pre_clip': 49.10011672973633, 'timing/data': 0.00026498000079300255, 'timing/model': 0.3705153860064456, 'learning_rate/qwen_vl_interface': 8.921558142317147e-06, 'learning_rate/action_model': 8.921558142317147e-05, 'epoch': 0.09}) +[2026-10-03 14:51:11,268][starVLA.training.train_starvla][INFO] - Step 1173, Loss: {'train/loss': 0.00861001294106245, 'train/grad_norm_pre_clip': 10.764451026916504, 'timing/data': 0.00023515999782830477, 'timing/model': 0.5671048939984757, 'learning_rate/qwen_vl_interface': 8.91962520425375e-06, 'learning_rate/action_model': 8.91962520425375e-05, 'epoch': 0.09}) +[2026-10-03 14:51:11,925][starVLA.training.train_starvla][INFO] - Step 1174, Loss: {'train/loss': 0.0001998566440306604, 'train/grad_norm_pre_clip': 0.02795291505753994, 'timing/data': 0.00028207000286784023, 'timing/model': 0.3703310399869224, 'learning_rate/qwen_vl_interface': 8.917690757746044e-06, 'learning_rate/action_model': 8.917690757746044e-05, 'epoch': 0.09}) +[2026-10-03 14:51:12,583][starVLA.training.train_starvla][INFO] - Step 1175, Loss: {'train/loss': 0.0007408742094412446, 'train/grad_norm_pre_clip': 0.3001341223716736, 'timing/data': 0.00023612000222783536, 'timing/model': 0.3704688809957588, 'learning_rate/qwen_vl_interface': 8.915754803589208e-06, 'learning_rate/action_model': 8.915754803589209e-05, 'epoch': 0.09}) +[2026-10-03 14:51:13,241][starVLA.training.train_starvla][INFO] - Step 1176, Loss: {'train/loss': 1.0060348510742188, 'train/grad_norm_pre_clip': 85.6083984375, 'timing/data': 0.00023581999994348735, 'timing/model': 0.37062268098816276, 'learning_rate/qwen_vl_interface': 8.913817342579042e-06, 'learning_rate/action_model': 8.913817342579042e-05, 'epoch': 0.09}) +[2026-10-03 14:51:13,873][starVLA.training.train_starvla][INFO] - Step 1177, Loss: {'train/loss': 0.6583911180496216, 'train/grad_norm_pre_clip': 10.860798835754395, 'timing/data': 0.00021053000818938017, 'timing/model': 0.36810317100025713, 'learning_rate/qwen_vl_interface': 8.91187837551196e-06, 'learning_rate/action_model': 8.911878375511961e-05, 'epoch': 0.09}) +[2026-10-03 14:51:14,530][starVLA.training.train_starvla][INFO] - Step 1178, Loss: {'train/loss': 0.22589588165283203, 'train/grad_norm_pre_clip': 76.9469223022461, 'timing/data': 0.0002987199986819178, 'timing/model': 0.37048302999755833, 'learning_rate/qwen_vl_interface': 8.909937903185e-06, 'learning_rate/action_model': 8.909937903185e-05, 'epoch': 0.09}) +[2026-10-03 14:51:15,378][starVLA.training.train_starvla][INFO] - Step 1179, Loss: {'train/loss': 0.044021595269441605, 'train/grad_norm_pre_clip': 37.99898147583008, 'timing/data': 0.00031015000422485173, 'timing/model': 0.37101178999000695, 'learning_rate/qwen_vl_interface': 8.907995926395814e-06, 'learning_rate/action_model': 8.907995926395815e-05, 'epoch': 0.09}) +[2026-10-03 14:51:16,036][starVLA.training.train_starvla][INFO] - Step 1180, Loss: {'train/loss': 0.19415231049060822, 'train/grad_norm_pre_clip': 33.397247314453125, 'timing/data': 0.0002889199968194589, 'timing/model': 0.3709219000011217, 'learning_rate/qwen_vl_interface': 8.90605244594268e-06, 'learning_rate/action_model': 8.906052445942679e-05, 'epoch': 0.09}) +[2026-10-03 14:51:16,691][starVLA.training.train_starvla][INFO] - Step 1181, Loss: {'train/loss': 0.5102121829986572, 'train/grad_norm_pre_clip': 192.9248809814453, 'timing/data': 0.0003355300141265616, 'timing/model': 0.3683792100055143, 'learning_rate/qwen_vl_interface': 8.904107462624484e-06, 'learning_rate/action_model': 8.904107462624483e-05, 'epoch': 0.09}) +[2026-10-03 14:51:17,349][starVLA.training.train_starvla][INFO] - Step 1182, Loss: {'train/loss': 0.01333747711032629, 'train/grad_norm_pre_clip': 10.22640323638916, 'timing/data': 0.0002832999889506027, 'timing/model': 0.37041479999606963, 'learning_rate/qwen_vl_interface': 8.902160977240739e-06, 'learning_rate/action_model': 8.902160977240739e-05, 'epoch': 0.09}) +[2026-10-03 14:51:18,006][starVLA.training.train_starvla][INFO] - Step 1183, Loss: {'train/loss': 0.4701092541217804, 'train/grad_norm_pre_clip': 24.893461227416992, 'timing/data': 0.0002454499917803332, 'timing/model': 0.37052115000551566, 'learning_rate/qwen_vl_interface': 8.90021299059157e-06, 'learning_rate/action_model': 8.90021299059157e-05, 'epoch': 0.09}) +[2026-10-03 14:51:18,866][starVLA.training.train_starvla][INFO] - Step 1184, Loss: {'train/loss': 0.00406369986012578, 'train/grad_norm_pre_clip': 8.530158042907715, 'timing/data': 0.0002630399976624176, 'timing/model': 0.3707580709888134, 'learning_rate/qwen_vl_interface': 8.898263503477722e-06, 'learning_rate/action_model': 8.898263503477721e-05, 'epoch': 0.09}) +[2026-10-03 14:51:19,522][starVLA.training.train_starvla][INFO] - Step 1185, Loss: {'train/loss': 0.004635263234376907, 'train/grad_norm_pre_clip': 0.7402451634407043, 'timing/data': 0.0002451699983794242, 'timing/model': 0.36896323099790607, 'learning_rate/qwen_vl_interface': 8.896312516700553e-06, 'learning_rate/action_model': 8.896312516700554e-05, 'epoch': 0.09}) +[2026-10-03 14:51:20,178][starVLA.training.train_starvla][INFO] - Step 1186, Loss: {'train/loss': 0.029007231816649437, 'train/grad_norm_pre_clip': 23.629989624023438, 'timing/data': 0.0002819200017256662, 'timing/model': 0.37006868100434076, 'learning_rate/qwen_vl_interface': 8.894360031062043e-06, 'learning_rate/action_model': 8.894360031062043e-05, 'epoch': 0.09}) +[2026-10-03 14:51:20,835][starVLA.training.train_starvla][INFO] - Step 1187, Loss: {'train/loss': 0.003180419560521841, 'train/grad_norm_pre_clip': 0.5485919713973999, 'timing/data': 0.00024837000819388777, 'timing/model': 0.3697549710050225, 'learning_rate/qwen_vl_interface': 8.892406047364783e-06, 'learning_rate/action_model': 8.892406047364784e-05, 'epoch': 0.09}) +[2026-10-03 14:51:21,493][starVLA.training.train_starvla][INFO] - Step 1188, Loss: {'train/loss': 0.0029232578817754984, 'train/grad_norm_pre_clip': 0.8468037247657776, 'timing/data': 0.00019408999651204795, 'timing/model': 0.3699683110025944, 'learning_rate/qwen_vl_interface': 8.890450566411984e-06, 'learning_rate/action_model': 8.890450566411984e-05, 'epoch': 0.09}) +[2026-10-03 14:51:22,342][starVLA.training.train_starvla][INFO] - Step 1189, Loss: {'train/loss': 0.002330583054572344, 'train/grad_norm_pre_clip': 0.3204452395439148, 'timing/data': 0.00025203000404872, 'timing/model': 0.5620257200062042, 'learning_rate/qwen_vl_interface': 8.888493589007467e-06, 'learning_rate/action_model': 8.888493589007468e-05, 'epoch': 0.09}) +[2026-10-03 14:51:22,998][starVLA.training.train_starvla][INFO] - Step 1190, Loss: {'train/loss': 0.11332499980926514, 'train/grad_norm_pre_clip': 73.9374771118164, 'timing/data': 0.00023160000273492187, 'timing/model': 0.36931724099849816, 'learning_rate/qwen_vl_interface': 8.886535115955676e-06, 'learning_rate/action_model': 8.886535115955675e-05, 'epoch': 0.09}) +[2026-10-03 14:51:23,660][starVLA.training.train_starvla][INFO] - Step 1191, Loss: {'train/loss': 0.00036563165485858917, 'train/grad_norm_pre_clip': 0.034741513431072235, 'timing/data': 0.00025852899125311524, 'timing/model': 0.36941398099588696, 'learning_rate/qwen_vl_interface': 8.884575148061662e-06, 'learning_rate/action_model': 8.88457514806166e-05, 'epoch': 0.09}) +[2026-10-03 14:51:24,315][starVLA.training.train_starvla][INFO] - Step 1192, Loss: {'train/loss': 0.0008432167232967913, 'train/grad_norm_pre_clip': 0.2656140923500061, 'timing/data': 0.0002561900037107989, 'timing/model': 0.36901869099528994, 'learning_rate/qwen_vl_interface': 8.882613686131095e-06, 'learning_rate/action_model': 8.882613686131093e-05, 'epoch': 0.09}) +[2026-10-03 14:51:24,970][starVLA.training.train_starvla][INFO] - Step 1193, Loss: {'train/loss': 0.0024123156908899546, 'train/grad_norm_pre_clip': 1.9473917484283447, 'timing/data': 0.0002406699932180345, 'timing/model': 0.36810750199947506, 'learning_rate/qwen_vl_interface': 8.880650730970254e-06, 'learning_rate/action_model': 8.880650730970254e-05, 'epoch': 0.09}) +[2026-10-03 14:51:25,790][starVLA.training.train_starvla][INFO] - Step 1194, Loss: {'train/loss': 0.0748857632279396, 'train/grad_norm_pre_clip': 63.559547424316406, 'timing/data': 0.00023628000053577125, 'timing/model': 0.5339399060030701, 'learning_rate/qwen_vl_interface': 8.878686283386042e-06, 'learning_rate/action_model': 8.878686283386042e-05, 'epoch': 0.09}) +[2026-10-03 14:51:26,445][starVLA.training.train_starvla][INFO] - Step 1195, Loss: {'train/loss': 0.00028517539612948895, 'train/grad_norm_pre_clip': 0.024636059999465942, 'timing/data': 0.00023846000840421766, 'timing/model': 0.36925478199555073, 'learning_rate/qwen_vl_interface': 8.876720344185967e-06, 'learning_rate/action_model': 8.876720344185967e-05, 'epoch': 0.09}) +[2026-10-03 14:51:27,101][starVLA.training.train_starvla][INFO] - Step 1196, Loss: {'train/loss': 0.00021249998826533556, 'train/grad_norm_pre_clip': 0.019764041528105736, 'timing/data': 0.00024264999956358224, 'timing/model': 0.3694768510031281, 'learning_rate/qwen_vl_interface': 8.874752914178149e-06, 'learning_rate/action_model': 8.87475291417815e-05, 'epoch': 0.09}) +[2026-10-03 14:51:27,755][starVLA.training.train_starvla][INFO] - Step 1197, Loss: {'train/loss': 0.9993294477462769, 'train/grad_norm_pre_clip': 34.77068328857422, 'timing/data': 0.00023626998881809413, 'timing/model': 0.3682738019997487, 'learning_rate/qwen_vl_interface': 8.87278399417133e-06, 'learning_rate/action_model': 8.872783994171329e-05, 'epoch': 0.09}) +[2026-10-03 14:51:28,387][starVLA.training.train_starvla][INFO] - Step 1198, Loss: {'train/loss': 0.04646647721529007, 'train/grad_norm_pre_clip': 17.844846725463867, 'timing/data': 0.00020244000188540667, 'timing/model': 0.34662417600338813, 'learning_rate/qwen_vl_interface': 8.870813584974852e-06, 'learning_rate/action_model': 8.870813584974852e-05, 'epoch': 0.09}) +[2026-10-03 14:51:29,043][starVLA.training.train_starvla][INFO] - Step 1199, Loss: {'train/loss': 0.000140322808874771, 'train/grad_norm_pre_clip': 0.014808372594416142, 'timing/data': 0.00024927999766077846, 'timing/model': 0.3691301810031291, 'learning_rate/qwen_vl_interface': 8.868841687398683e-06, 'learning_rate/action_model': 8.868841687398683e-05, 'epoch': 0.09}) +[2026-10-03 14:51:29,892][starVLA.training.train_starvla][INFO] - Step 1200, Loss: {'train/loss': 0.0004369376110844314, 'train/grad_norm_pre_clip': 0.31121933460235596, 'timing/data': 0.0003377800021553412, 'timing/model': 0.36885123200772796, 'learning_rate/qwen_vl_interface': 8.86686830225339e-06, 'learning_rate/action_model': 8.866868302253389e-05, 'epoch': 0.09}) +[2026-10-03 14:51:30,547][starVLA.training.train_starvla][INFO] - Step 1201, Loss: {'train/loss': 0.03462493047118187, 'train/grad_norm_pre_clip': 169.1032257080078, 'timing/data': 0.0002484299911884591, 'timing/model': 0.36759077200258616, 'learning_rate/qwen_vl_interface': 8.864893430350159e-06, 'learning_rate/action_model': 8.864893430350159e-05, 'epoch': 0.09}) +[2026-10-03 14:51:31,202][starVLA.training.train_starvla][INFO] - Step 1202, Loss: {'train/loss': 0.5811464190483093, 'train/grad_norm_pre_clip': 9.111241340637207, 'timing/data': 0.00032994999492075294, 'timing/model': 0.3698821809957735, 'learning_rate/qwen_vl_interface': 8.862917072500785e-06, 'learning_rate/action_model': 8.862917072500784e-05, 'epoch': 0.09}) +[2026-10-03 14:51:31,858][starVLA.training.train_starvla][INFO] - Step 1203, Loss: {'train/loss': 1.0827394723892212, 'train/grad_norm_pre_clip': 121.79662322998047, 'timing/data': 0.0002958999975817278, 'timing/model': 0.3690737819997594, 'learning_rate/qwen_vl_interface': 8.860939229517674e-06, 'learning_rate/action_model': 8.860939229517674e-05, 'epoch': 0.09}) +[2026-10-03 14:51:32,514][starVLA.training.train_starvla][INFO] - Step 1204, Loss: {'train/loss': 0.00040500209433957934, 'train/grad_norm_pre_clip': 0.10612618178129196, 'timing/data': 0.0002411699970252812, 'timing/model': 0.3697210410027765, 'learning_rate/qwen_vl_interface': 8.858959902213844e-06, 'learning_rate/action_model': 8.858959902213845e-05, 'epoch': 0.09}) +[2026-10-03 14:51:33,359][starVLA.training.train_starvla][INFO] - Step 1205, Loss: {'train/loss': 0.0003145118826068938, 'train/grad_norm_pre_clip': 0.032033395022153854, 'timing/data': 0.000287600007141009, 'timing/model': 0.5571312020038022, 'learning_rate/qwen_vl_interface': 8.856979091402921e-06, 'learning_rate/action_model': 8.85697909140292e-05, 'epoch': 0.09}) +[2026-10-03 14:51:34,014][starVLA.training.train_starvla][INFO] - Step 1206, Loss: {'train/loss': 0.28904950618743896, 'train/grad_norm_pre_clip': 145.19329833984375, 'timing/data': 0.0002016200014622882, 'timing/model': 0.36938633100362495, 'learning_rate/qwen_vl_interface': 8.854996797899139e-06, 'learning_rate/action_model': 8.854996797899139e-05, 'epoch': 0.09}) +[2026-10-03 14:51:34,670][starVLA.training.train_starvla][INFO] - Step 1207, Loss: {'train/loss': 0.0006794530781917274, 'train/grad_norm_pre_clip': 0.11120728403329849, 'timing/data': 0.00024665999808348715, 'timing/model': 0.3692758410034003, 'learning_rate/qwen_vl_interface': 8.853013022517348e-06, 'learning_rate/action_model': 8.853013022517348e-05, 'epoch': 0.09}) +[2026-10-03 14:51:35,327][starVLA.training.train_starvla][INFO] - Step 1208, Loss: {'train/loss': 0.007896294817328453, 'train/grad_norm_pre_clip': 8.75715160369873, 'timing/data': 0.0002464400022290647, 'timing/model': 0.37023470099666156, 'learning_rate/qwen_vl_interface': 8.851027766073002e-06, 'learning_rate/action_model': 8.851027766073001e-05, 'epoch': 0.09}) +[2026-10-03 14:51:35,982][starVLA.training.train_starvla][INFO] - Step 1209, Loss: {'train/loss': 0.0010277724359184504, 'train/grad_norm_pre_clip': 0.1463879644870758, 'timing/data': 0.000393959999200888, 'timing/model': 0.36830993100011256, 'learning_rate/qwen_vl_interface': 8.84904102938216e-06, 'learning_rate/action_model': 8.84904102938216e-05, 'epoch': 0.09}) +[2026-10-03 14:51:36,838][starVLA.training.train_starvla][INFO] - Step 1210, Loss: {'train/loss': 0.7735018134117126, 'train/grad_norm_pre_clip': 30.647136688232422, 'timing/data': 0.0003202099906047806, 'timing/model': 0.5696100799978012, 'learning_rate/qwen_vl_interface': 8.8470528132615e-06, 'learning_rate/action_model': 8.8470528132615e-05, 'epoch': 0.09}) +[2026-10-03 14:51:37,493][starVLA.training.train_starvla][INFO] - Step 1211, Loss: {'train/loss': 0.000821890716906637, 'train/grad_norm_pre_clip': 0.20945611596107483, 'timing/data': 0.00023314000281970948, 'timing/model': 0.36978471199108753, 'learning_rate/qwen_vl_interface': 8.845063118528303e-06, 'learning_rate/action_model': 8.845063118528304e-05, 'epoch': 0.09}) +[2026-10-03 14:51:38,150][starVLA.training.train_starvla][INFO] - Step 1212, Loss: {'train/loss': 0.0009292555623687804, 'train/grad_norm_pre_clip': 0.17286136746406555, 'timing/data': 0.00031382999441120774, 'timing/model': 0.3696523620019434, 'learning_rate/qwen_vl_interface': 8.843071946000455e-06, 'learning_rate/action_model': 8.843071946000456e-05, 'epoch': 0.09}) +[2026-10-03 14:51:38,804][starVLA.training.train_starvla][INFO] - Step 1213, Loss: {'train/loss': 0.002238406566902995, 'train/grad_norm_pre_clip': 2.181997776031494, 'timing/data': 0.00022409000666812062, 'timing/model': 0.36810770200099796, 'learning_rate/qwen_vl_interface': 8.841079296496451e-06, 'learning_rate/action_model': 8.84107929649645e-05, 'epoch': 0.09}) +[2026-10-03 14:51:39,461][starVLA.training.train_starvla][INFO] - Step 1214, Loss: {'train/loss': 0.08967065811157227, 'train/grad_norm_pre_clip': 118.8294906616211, 'timing/data': 0.00032315000134985894, 'timing/model': 0.36977785199997015, 'learning_rate/qwen_vl_interface': 8.839085170835396e-06, 'learning_rate/action_model': 8.839085170835395e-05, 'epoch': 0.09}) +[2026-10-03 14:51:40,117][starVLA.training.train_starvla][INFO] - Step 1215, Loss: {'train/loss': 0.001527274725958705, 'train/grad_norm_pre_clip': 0.19182239472866058, 'timing/data': 0.00020546000450849533, 'timing/model': 0.36980734100507107, 'learning_rate/qwen_vl_interface': 8.837089569837e-06, 'learning_rate/action_model': 8.837089569836998e-05, 'epoch': 0.09}) +[2026-10-03 14:51:40,972][starVLA.training.train_starvla][INFO] - Step 1216, Loss: {'train/loss': 0.7053810954093933, 'train/grad_norm_pre_clip': 99.99006652832031, 'timing/data': 0.00034220000088680536, 'timing/model': 0.37008642200089525, 'learning_rate/qwen_vl_interface': 8.835092494321575e-06, 'learning_rate/action_model': 8.835092494321575e-05, 'epoch': 0.09}) +[2026-10-03 14:51:41,628][starVLA.training.train_starvla][INFO] - Step 1217, Loss: {'train/loss': 0.28364887833595276, 'train/grad_norm_pre_clip': 41.0464973449707, 'timing/data': 0.0002040999970631674, 'timing/model': 0.36835196299944073, 'learning_rate/qwen_vl_interface': 8.833093945110048e-06, 'learning_rate/action_model': 8.833093945110047e-05, 'epoch': 0.09}) +[2026-10-03 14:51:42,283][starVLA.training.train_starvla][INFO] - Step 1218, Loss: {'train/loss': 0.0036919964477419853, 'train/grad_norm_pre_clip': 2.3741321563720703, 'timing/data': 0.00039154000114649534, 'timing/model': 0.3695857920101844, 'learning_rate/qwen_vl_interface': 8.831093923023946e-06, 'learning_rate/action_model': 8.831093923023945e-05, 'epoch': 0.09}) +[2026-10-03 14:51:42,939][starVLA.training.train_starvla][INFO] - Step 1219, Loss: {'train/loss': 0.0008843217510730028, 'train/grad_norm_pre_clip': 0.10373737663030624, 'timing/data': 0.00019870999676641077, 'timing/model': 0.3693019829952391, 'learning_rate/qwen_vl_interface': 8.8290924288854e-06, 'learning_rate/action_model': 8.8290924288854e-05, 'epoch': 0.09}) +[2026-10-03 14:51:43,596][starVLA.training.train_starvla][INFO] - Step 1220, Loss: {'train/loss': 0.3939146399497986, 'train/grad_norm_pre_clip': 105.91709899902344, 'timing/data': 0.00021003998699598014, 'timing/model': 0.36965270299697295, 'learning_rate/qwen_vl_interface': 8.827089463517153e-06, 'learning_rate/action_model': 8.827089463517153e-05, 'epoch': 0.09}) +[2026-10-03 14:51:44,442][starVLA.training.train_starvla][INFO] - Step 1221, Loss: {'train/loss': 0.18338243663311005, 'train/grad_norm_pre_clip': 108.52037811279297, 'timing/data': 0.000198540001292713, 'timing/model': 0.5593759530020179, 'learning_rate/qwen_vl_interface': 8.825085027742545e-06, 'learning_rate/action_model': 8.825085027742544e-05, 'epoch': 0.09}) +[2026-10-03 14:51:45,097][starVLA.training.train_starvla][INFO] - Step 1222, Loss: {'train/loss': 0.7371973991394043, 'train/grad_norm_pre_clip': 18.26914405822754, 'timing/data': 0.0002966699976241216, 'timing/model': 0.3693852330034133, 'learning_rate/qwen_vl_interface': 8.823079122385523e-06, 'learning_rate/action_model': 8.823079122385523e-05, 'epoch': 0.09}) +[2026-10-03 14:51:45,754][starVLA.training.train_starvla][INFO] - Step 1223, Loss: {'train/loss': 0.0008618168067187071, 'train/grad_norm_pre_clip': 0.11451460421085358, 'timing/data': 0.00025149001157842577, 'timing/model': 0.36990868298744317, 'learning_rate/qwen_vl_interface': 8.821071748270644e-06, 'learning_rate/action_model': 8.821071748270644e-05, 'epoch': 0.09}) +[2026-10-03 14:51:46,410][starVLA.training.train_starvla][INFO] - Step 1224, Loss: {'train/loss': 0.0008760517230257392, 'train/grad_norm_pre_clip': 0.1627916395664215, 'timing/data': 0.00025364999601151794, 'timing/model': 0.369556243007537, 'learning_rate/qwen_vl_interface': 8.81906290622306e-06, 'learning_rate/action_model': 8.81906290622306e-05, 'epoch': 0.09}) +[2026-10-03 14:51:47,041][starVLA.training.train_starvla][INFO] - Step 1225, Loss: {'train/loss': 0.19141125679016113, 'train/grad_norm_pre_clip': 84.47586822509766, 'timing/data': 0.0003104999923380092, 'timing/model': 0.34469853799964767, 'learning_rate/qwen_vl_interface': 8.81705259706853e-06, 'learning_rate/action_model': 8.817052597068529e-05, 'epoch': 0.09}) +[2026-10-03 14:51:47,895][starVLA.training.train_starvla][INFO] - Step 1226, Loss: {'train/loss': 0.7708978056907654, 'train/grad_norm_pre_clip': 16.717679977416992, 'timing/data': 0.0003104799980064854, 'timing/model': 0.5681115509942174, 'learning_rate/qwen_vl_interface': 8.815040821633416e-06, 'learning_rate/action_model': 8.815040821633416e-05, 'epoch': 0.09}) +[2026-10-03 14:51:48,552][starVLA.training.train_starvla][INFO] - Step 1227, Loss: {'train/loss': 0.26941922307014465, 'train/grad_norm_pre_clip': 61.44990539550781, 'timing/data': 0.0002516590029699728, 'timing/model': 0.36944977298844606, 'learning_rate/qwen_vl_interface': 8.813027580744685e-06, 'learning_rate/action_model': 8.813027580744684e-05, 'epoch': 0.09}) +[2026-10-03 14:51:49,185][starVLA.training.train_starvla][INFO] - Step 1228, Loss: {'train/loss': 0.8363901972770691, 'train/grad_norm_pre_clip': 89.19869995117188, 'timing/data': 0.00032147001184057444, 'timing/model': 0.34672857700206805, 'learning_rate/qwen_vl_interface': 8.8110128752299e-06, 'learning_rate/action_model': 8.8110128752299e-05, 'epoch': 0.09}) +[2026-10-03 14:51:49,840][starVLA.training.train_starvla][INFO] - Step 1229, Loss: {'train/loss': 0.0014669712400063872, 'train/grad_norm_pre_clip': 0.8007796406745911, 'timing/data': 0.00020179999410174787, 'timing/model': 0.3683286129962653, 'learning_rate/qwen_vl_interface': 8.808996705917235e-06, 'learning_rate/action_model': 8.808996705917235e-05, 'epoch': 0.09}) +[2026-10-03 14:51:50,496][starVLA.training.train_starvla][INFO] - Step 1230, Loss: {'train/loss': 0.006674251519143581, 'train/grad_norm_pre_clip': 10.404399871826172, 'timing/data': 0.00025102999643422663, 'timing/model': 0.36954316298943013, 'learning_rate/qwen_vl_interface': 8.80697907363546e-06, 'learning_rate/action_model': 8.80697907363546e-05, 'epoch': 0.09}) +[2026-10-03 14:51:51,151][starVLA.training.train_starvla][INFO] - Step 1231, Loss: {'train/loss': 0.2907613515853882, 'train/grad_norm_pre_clip': 37.902950286865234, 'timing/data': 0.00025548999838065356, 'timing/model': 0.36904047199641354, 'learning_rate/qwen_vl_interface': 8.804959979213945e-06, 'learning_rate/action_model': 8.804959979213945e-05, 'epoch': 0.09}) +[2026-10-03 14:51:52,003][starVLA.training.train_starvla][INFO] - Step 1232, Loss: {'train/loss': 0.278635710477829, 'train/grad_norm_pre_clip': 141.35011291503906, 'timing/data': 0.000317530008032918, 'timing/model': 0.36918265199346934, 'learning_rate/qwen_vl_interface': 8.802939423482667e-06, 'learning_rate/action_model': 8.802939423482666e-05, 'epoch': 0.09}) +[2026-10-03 14:51:52,657][starVLA.training.train_starvla][INFO] - Step 1233, Loss: {'train/loss': 0.449871689081192, 'train/grad_norm_pre_clip': 89.69293212890625, 'timing/data': 0.0002406900021014735, 'timing/model': 0.3674159930087626, 'learning_rate/qwen_vl_interface': 8.800917407272198e-06, 'learning_rate/action_model': 8.800917407272198e-05, 'epoch': 0.09}) +[2026-10-03 14:51:53,288][starVLA.training.train_starvla][INFO] - Step 1234, Loss: {'train/loss': 0.38601940870285034, 'train/grad_norm_pre_clip': 13.370768547058105, 'timing/data': 0.00023859000066295266, 'timing/model': 0.34645301700220443, 'learning_rate/qwen_vl_interface': 8.798893931413711e-06, 'learning_rate/action_model': 8.79889393141371e-05, 'epoch': 0.09}) +[2026-10-03 14:51:53,944][starVLA.training.train_starvla][INFO] - Step 1235, Loss: {'train/loss': 0.0747874528169632, 'train/grad_norm_pre_clip': 53.60426330566406, 'timing/data': 0.00023764999059494585, 'timing/model': 0.3693838430044707, 'learning_rate/qwen_vl_interface': 8.796868996738985e-06, 'learning_rate/action_model': 8.796868996738984e-05, 'epoch': 0.09}) +[2026-10-03 14:51:54,600][starVLA.training.train_starvla][INFO] - Step 1236, Loss: {'train/loss': 0.003157305531203747, 'train/grad_norm_pre_clip': 0.6946765780448914, 'timing/data': 0.000245789997279644, 'timing/model': 0.36967226199340075, 'learning_rate/qwen_vl_interface': 8.794842604080389e-06, 'learning_rate/action_model': 8.794842604080389e-05, 'epoch': 0.09}) +[2026-10-03 14:51:55,438][starVLA.training.train_starvla][INFO] - Step 1237, Loss: {'train/loss': 0.0032499153167009354, 'train/grad_norm_pre_clip': 0.7436051964759827, 'timing/data': 0.00033454899676144123, 'timing/model': 0.3453290379984537, 'learning_rate/qwen_vl_interface': 8.7928147542709e-06, 'learning_rate/action_model': 8.7928147542709e-05, 'epoch': 0.09}) +[2026-10-03 14:51:56,070][starVLA.training.train_starvla][INFO] - Step 1238, Loss: {'train/loss': 0.0023885369300842285, 'train/grad_norm_pre_clip': 0.3417980372905731, 'timing/data': 0.00023573999351356179, 'timing/model': 0.36859315299079753, 'learning_rate/qwen_vl_interface': 8.790785448144088e-06, 'learning_rate/action_model': 8.790785448144088e-05, 'epoch': 0.09}) +[2026-10-03 14:51:56,725][starVLA.training.train_starvla][INFO] - Step 1239, Loss: {'train/loss': 0.002867886098101735, 'train/grad_norm_pre_clip': 0.9211944937705994, 'timing/data': 0.00023762999626342207, 'timing/model': 0.36869424299220555, 'learning_rate/qwen_vl_interface': 8.788754686534126e-06, 'learning_rate/action_model': 8.788754686534127e-05, 'epoch': 0.09}) +[2026-10-03 14:51:57,357][starVLA.training.train_starvla][INFO] - Step 1240, Loss: {'train/loss': 0.27897822856903076, 'train/grad_norm_pre_clip': 53.888492584228516, 'timing/data': 0.0002899999963119626, 'timing/model': 0.3692682029941352, 'learning_rate/qwen_vl_interface': 8.786722470275784e-06, 'learning_rate/action_model': 8.786722470275784e-05, 'epoch': 0.09}) +[2026-10-03 14:51:57,989][starVLA.training.train_starvla][INFO] - Step 1241, Loss: {'train/loss': 0.0043630353175103664, 'train/grad_norm_pre_clip': 2.6576297283172607, 'timing/data': 0.0002126199979102239, 'timing/model': 0.3674104039964732, 'learning_rate/qwen_vl_interface': 8.784688800204427e-06, 'learning_rate/action_model': 8.784688800204426e-05, 'epoch': 0.09}) +[2026-10-03 14:51:58,833][starVLA.training.train_starvla][INFO] - Step 1242, Loss: {'train/loss': 0.004139117896556854, 'train/grad_norm_pre_clip': 1.5126310586929321, 'timing/data': 0.00023836000764276832, 'timing/model': 0.5582692039897665, 'learning_rate/qwen_vl_interface': 8.782653677156022e-06, 'learning_rate/action_model': 8.782653677156022e-05, 'epoch': 0.09}) +[2026-10-03 14:51:59,488][starVLA.training.train_starvla][INFO] - Step 1243, Loss: {'train/loss': 0.0009904478210955858, 'train/grad_norm_pre_clip': 0.12042027711868286, 'timing/data': 0.00023816000611986965, 'timing/model': 0.3693473930034088, 'learning_rate/qwen_vl_interface': 8.780617101967129e-06, 'learning_rate/action_model': 8.780617101967129e-05, 'epoch': 0.09}) +[2026-10-03 14:52:00,143][starVLA.training.train_starvla][INFO] - Step 1244, Loss: {'train/loss': 0.0008145578904077411, 'train/grad_norm_pre_clip': 0.08142194896936417, 'timing/data': 0.00023288000375032425, 'timing/model': 0.36856759400689043, 'learning_rate/qwen_vl_interface': 8.77857907547491e-06, 'learning_rate/action_model': 8.77857907547491e-05, 'epoch': 0.09}) +[2026-10-03 14:52:00,796][starVLA.training.train_starvla][INFO] - Step 1245, Loss: {'train/loss': 0.27869126200675964, 'train/grad_norm_pre_clip': 20.433753967285156, 'timing/data': 0.0002484099968569353, 'timing/model': 0.3672822240041569, 'learning_rate/qwen_vl_interface': 8.776539598517119e-06, 'learning_rate/action_model': 8.776539598517118e-05, 'epoch': 0.09}) +[2026-10-03 14:52:01,450][starVLA.training.train_starvla][INFO] - Step 1246, Loss: {'train/loss': 0.015620420686900616, 'train/grad_norm_pre_clip': 5.065249919891357, 'timing/data': 0.0002263999922433868, 'timing/model': 0.3691108630009694, 'learning_rate/qwen_vl_interface': 8.774498671932108e-06, 'learning_rate/action_model': 8.774498671932107e-05, 'epoch': 0.09}) +[2026-10-03 14:52:02,105][starVLA.training.train_starvla][INFO] - Step 1247, Loss: {'train/loss': 0.0027754991315305233, 'train/grad_norm_pre_clip': 1.9077848196029663, 'timing/data': 0.000248480006121099, 'timing/model': 0.3691458739922382, 'learning_rate/qwen_vl_interface': 8.772456296558824e-06, 'learning_rate/action_model': 8.772456296558824e-05, 'epoch': 0.09}) +[2026-10-03 14:52:02,964][starVLA.training.train_starvla][INFO] - Step 1248, Loss: {'train/loss': 0.7023493647575378, 'train/grad_norm_pre_clip': 38.050926208496094, 'timing/data': 0.00028122999356128275, 'timing/model': 0.36926226400828455, 'learning_rate/qwen_vl_interface': 8.77041247323681e-06, 'learning_rate/action_model': 8.770412473236811e-05, 'epoch': 0.09}) +[2026-10-03 14:52:03,618][starVLA.training.train_starvla][INFO] - Step 1249, Loss: {'train/loss': 0.2315221130847931, 'train/grad_norm_pre_clip': 24.96272087097168, 'timing/data': 0.00032513999030925333, 'timing/model': 0.3671864439966157, 'learning_rate/qwen_vl_interface': 8.76836720280621e-06, 'learning_rate/action_model': 8.768367202806209e-05, 'epoch': 0.09}) +[2026-10-03 14:52:04,272][starVLA.training.train_starvla][INFO] - Step 1250, Loss: {'train/loss': 0.28138238191604614, 'train/grad_norm_pre_clip': 68.33226776123047, 'timing/data': 0.0002487199963070452, 'timing/model': 0.36860498299938627, 'learning_rate/qwen_vl_interface': 8.766320486107749e-06, 'learning_rate/action_model': 8.766320486107749e-05, 'epoch': 0.09}) +[2026-10-03 14:52:04,929][starVLA.training.train_starvla][INFO] - Step 1251, Loss: {'train/loss': 0.0817594826221466, 'train/grad_norm_pre_clip': 29.719287872314453, 'timing/data': 0.00023593999503646046, 'timing/model': 0.3696926730044652, 'learning_rate/qwen_vl_interface': 8.76427232398276e-06, 'learning_rate/action_model': 8.76427232398276e-05, 'epoch': 0.09}) +[2026-10-03 14:52:05,586][starVLA.training.train_starvla][INFO] - Step 1252, Loss: {'train/loss': 0.2713456451892853, 'train/grad_norm_pre_clip': 32.13526153564453, 'timing/data': 0.00020025999401696026, 'timing/model': 0.37032424300559796, 'learning_rate/qwen_vl_interface': 8.762222717273167e-06, 'learning_rate/action_model': 8.762222717273167e-05, 'epoch': 0.09}) +[2026-10-03 14:52:06,438][starVLA.training.train_starvla][INFO] - Step 1253, Loss: {'train/loss': 0.8332873582839966, 'train/grad_norm_pre_clip': 26.66098403930664, 'timing/data': 0.00039372000901494175, 'timing/model': 0.36866283300332725, 'learning_rate/qwen_vl_interface': 8.76017166682148e-06, 'learning_rate/action_model': 8.760171666821481e-05, 'epoch': 0.09}) +[2026-10-03 14:52:07,093][starVLA.training.train_starvla][INFO] - Step 1254, Loss: {'train/loss': 0.028282934799790382, 'train/grad_norm_pre_clip': 9.872709274291992, 'timing/data': 0.00023332001001108438, 'timing/model': 0.3689185730036115, 'learning_rate/qwen_vl_interface': 8.758119173470815e-06, 'learning_rate/action_model': 8.758119173470815e-05, 'epoch': 0.09}) +[2026-10-03 14:52:07,748][starVLA.training.train_starvla][INFO] - Step 1255, Loss: {'train/loss': 0.0006954582640901208, 'train/grad_norm_pre_clip': 0.19160223007202148, 'timing/data': 0.000289039991912432, 'timing/model': 0.3687873640010366, 'learning_rate/qwen_vl_interface': 8.75606523806487e-06, 'learning_rate/action_model': 8.756065238064868e-05, 'epoch': 0.09}) +[2026-10-03 14:52:08,404][starVLA.training.train_starvla][INFO] - Step 1256, Loss: {'train/loss': 0.00018745554552879184, 'train/grad_norm_pre_clip': 0.06007090210914612, 'timing/data': 0.00024087999190669507, 'timing/model': 0.36964676399657037, 'learning_rate/qwen_vl_interface': 8.75400986144794e-06, 'learning_rate/action_model': 8.75400986144794e-05, 'epoch': 0.09}) +[2026-10-03 14:52:09,058][starVLA.training.train_starvla][INFO] - Step 1257, Loss: {'train/loss': 0.0061606233939528465, 'train/grad_norm_pre_clip': 9.860932350158691, 'timing/data': 0.0002610299998195842, 'timing/model': 0.36761114399996586, 'learning_rate/qwen_vl_interface': 8.751953044464913e-06, 'learning_rate/action_model': 8.751953044464913e-05, 'epoch': 0.09}) +[2026-10-03 14:52:09,915][starVLA.training.train_starvla][INFO] - Step 1258, Loss: {'train/loss': 0.01996573619544506, 'train/grad_norm_pre_clip': 13.778800010681152, 'timing/data': 0.0002397900098003447, 'timing/model': 0.570976122005959, 'learning_rate/qwen_vl_interface': 8.749894787961272e-06, 'learning_rate/action_model': 8.749894787961271e-05, 'epoch': 0.09}) +[2026-10-03 14:52:10,571][starVLA.training.train_starvla][INFO] - Step 1259, Loss: {'train/loss': 0.005326612386852503, 'train/grad_norm_pre_clip': 2.223435163497925, 'timing/data': 0.00022046999947633594, 'timing/model': 0.369526042995858, 'learning_rate/qwen_vl_interface': 8.747835092783086e-06, 'learning_rate/action_model': 8.747835092783085e-05, 'epoch': 0.09}) +[2026-10-03 14:52:11,227][starVLA.training.train_starvla][INFO] - Step 1260, Loss: {'train/loss': 0.44617462158203125, 'train/grad_norm_pre_clip': 69.72547912597656, 'timing/data': 0.00028494998696260154, 'timing/model': 0.36971668399928603, 'learning_rate/qwen_vl_interface': 8.745773959777018e-06, 'learning_rate/action_model': 8.745773959777017e-05, 'epoch': 0.09}) +[2026-10-03 14:52:11,881][starVLA.training.train_starvla][INFO] - Step 1261, Loss: {'train/loss': 0.0006193279405124485, 'train/grad_norm_pre_clip': 0.11818460375070572, 'timing/data': 0.00021791999461129308, 'timing/model': 0.36798405399895273, 'learning_rate/qwen_vl_interface': 8.743711389790321e-06, 'learning_rate/action_model': 8.74371138979032e-05, 'epoch': 0.09}) +[2026-10-03 14:52:12,537][starVLA.training.train_starvla][INFO] - Step 1262, Loss: {'train/loss': 0.22899293899536133, 'train/grad_norm_pre_clip': 59.338623046875, 'timing/data': 0.0002281600027345121, 'timing/model': 0.3697073239891324, 'learning_rate/qwen_vl_interface': 8.741647383670839e-06, 'learning_rate/action_model': 8.741647383670839e-05, 'epoch': 0.09}) +[2026-10-03 14:52:13,361][starVLA.training.train_starvla][INFO] - Step 1263, Loss: {'train/loss': 0.2765688896179199, 'train/grad_norm_pre_clip': 23.81048583984375, 'timing/data': 0.00020583000150509179, 'timing/model': 0.5381678990088403, 'learning_rate/qwen_vl_interface': 8.739581942267011e-06, 'learning_rate/action_model': 8.739581942267011e-05, 'epoch': 0.09}) +[2026-10-03 14:52:14,018][starVLA.training.train_starvla][INFO] - Step 1264, Loss: {'train/loss': 0.0017329108668491244, 'train/grad_norm_pre_clip': 0.6350512504577637, 'timing/data': 0.00024511000083293766, 'timing/model': 0.3705561040114844, 'learning_rate/qwen_vl_interface': 8.737515066427861e-06, 'learning_rate/action_model': 8.73751506642786e-05, 'epoch': 0.09}) +[2026-10-03 14:52:14,673][starVLA.training.train_starvla][INFO] - Step 1265, Loss: {'train/loss': 0.08415932953357697, 'train/grad_norm_pre_clip': 64.31892395019531, 'timing/data': 0.0003228800051147118, 'timing/model': 0.3678570839983877, 'learning_rate/qwen_vl_interface': 8.735446757003e-06, 'learning_rate/action_model': 8.735446757003e-05, 'epoch': 0.09}) +[2026-10-03 14:52:15,327][starVLA.training.train_starvla][INFO] - Step 1266, Loss: {'train/loss': 0.02344721555709839, 'train/grad_norm_pre_clip': 95.07183837890625, 'timing/data': 0.00023859999782871455, 'timing/model': 0.36937955500616226, 'learning_rate/qwen_vl_interface': 8.733377014842635e-06, 'learning_rate/action_model': 8.733377014842635e-05, 'epoch': 0.09}) +[2026-10-03 14:52:15,984][starVLA.training.train_starvla][INFO] - Step 1267, Loss: {'train/loss': 0.0007904272642917931, 'train/grad_norm_pre_clip': 0.147177055478096, 'timing/data': 0.0002401000092504546, 'timing/model': 0.3694459639955312, 'learning_rate/qwen_vl_interface': 8.731305840797557e-06, 'learning_rate/action_model': 8.731305840797558e-05, 'epoch': 0.09}) +[2026-10-03 14:52:16,640][starVLA.training.train_starvla][INFO] - Step 1268, Loss: {'train/loss': 0.6505775451660156, 'train/grad_norm_pre_clip': 20.2314395904541, 'timing/data': 0.00024110000231303275, 'timing/model': 0.36966028400638606, 'learning_rate/qwen_vl_interface': 8.72923323571915e-06, 'learning_rate/action_model': 8.72923323571915e-05, 'epoch': 0.09}) +[2026-10-03 14:52:17,493][starVLA.training.train_starvla][INFO] - Step 1269, Loss: {'train/loss': 0.0018517617136240005, 'train/grad_norm_pre_clip': 1.2292909622192383, 'timing/data': 0.00030898999830242246, 'timing/model': 0.3691213650017744, 'learning_rate/qwen_vl_interface': 8.727159200459381e-06, 'learning_rate/action_model': 8.727159200459381e-05, 'epoch': 0.09}) +[2026-10-03 14:52:18,149][starVLA.training.train_starvla][INFO] - Step 1270, Loss: {'train/loss': 0.017122279852628708, 'train/grad_norm_pre_clip': 33.252410888671875, 'timing/data': 0.00028403000032994896, 'timing/model': 0.36985970399109647, 'learning_rate/qwen_vl_interface': 8.725083735870807e-06, 'learning_rate/action_model': 8.725083735870807e-05, 'epoch': 0.09}) +[2026-10-03 14:52:18,806][starVLA.training.train_starvla][INFO] - Step 1271, Loss: {'train/loss': 0.00977400317788124, 'train/grad_norm_pre_clip': 9.588011741638184, 'timing/data': 0.0002496899978723377, 'timing/model': 0.3698905439960072, 'learning_rate/qwen_vl_interface': 8.723006842806575e-06, 'learning_rate/action_model': 8.723006842806576e-05, 'epoch': 0.09}) +[2026-10-03 14:52:19,463][starVLA.training.train_starvla][INFO] - Step 1272, Loss: {'train/loss': 0.160253643989563, 'train/grad_norm_pre_clip': 121.53585815429688, 'timing/data': 0.00028651999309659004, 'timing/model': 0.3706293439900037, 'learning_rate/qwen_vl_interface': 8.720928522120419e-06, 'learning_rate/action_model': 8.720928522120418e-05, 'epoch': 0.09}) +[2026-10-03 14:52:20,118][starVLA.training.train_starvla][INFO] - Step 1273, Loss: {'train/loss': 0.0011703629279509187, 'train/grad_norm_pre_clip': 0.704422652721405, 'timing/data': 0.0003306800063000992, 'timing/model': 0.36775759400916286, 'learning_rate/qwen_vl_interface': 8.718848774666655e-06, 'learning_rate/action_model': 8.718848774666655e-05, 'epoch': 0.09}) +[2026-10-03 14:52:20,967][starVLA.training.train_starvla][INFO] - Step 1274, Loss: {'train/loss': 0.14521493017673492, 'train/grad_norm_pre_clip': 60.338890075683594, 'timing/data': 0.0003299200034234673, 'timing/model': 0.37008845499076415, 'learning_rate/qwen_vl_interface': 8.716767601300189e-06, 'learning_rate/action_model': 8.716767601300188e-05, 'epoch': 0.09}) +[2026-10-03 14:52:21,623][starVLA.training.train_starvla][INFO] - Step 1275, Loss: {'train/loss': 0.000666359846945852, 'train/grad_norm_pre_clip': 0.32630905508995056, 'timing/data': 0.00022028000967111439, 'timing/model': 0.36985775400535204, 'learning_rate/qwen_vl_interface': 8.714685002876514e-06, 'learning_rate/action_model': 8.714685002876513e-05, 'epoch': 0.09}) +[2026-10-03 14:52:22,279][starVLA.training.train_starvla][INFO] - Step 1276, Loss: {'train/loss': 1.0167678594589233, 'train/grad_norm_pre_clip': 44.8702507019043, 'timing/data': 0.00022669999452773482, 'timing/model': 0.36990240500017535, 'learning_rate/qwen_vl_interface': 8.712600980251706e-06, 'learning_rate/action_model': 8.712600980251705e-05, 'epoch': 0.09}) +[2026-10-03 14:52:22,934][starVLA.training.train_starvla][INFO] - Step 1277, Loss: {'train/loss': 0.057428989559412, 'train/grad_norm_pre_clip': 20.101417541503906, 'timing/data': 0.0003357199893798679, 'timing/model': 0.36764553500688635, 'learning_rate/qwen_vl_interface': 8.710515534282428e-06, 'learning_rate/action_model': 8.710515534282428e-05, 'epoch': 0.09}) +[2026-10-03 14:52:23,590][starVLA.training.train_starvla][INFO] - Step 1278, Loss: {'train/loss': 0.0008147053886204958, 'train/grad_norm_pre_clip': 0.27928733825683594, 'timing/data': 0.00028824000037275255, 'timing/model': 0.3698941049951827, 'learning_rate/qwen_vl_interface': 8.708428665825931e-06, 'learning_rate/action_model': 8.708428665825931e-05, 'epoch': 0.09}) +[2026-10-03 14:52:24,444][starVLA.training.train_starvla][INFO] - Step 1279, Loss: {'train/loss': 0.45524656772613525, 'train/grad_norm_pre_clip': 26.829065322875977, 'timing/data': 0.00030417999369092286, 'timing/model': 0.5678113040048629, 'learning_rate/qwen_vl_interface': 8.706340375740045e-06, 'learning_rate/action_model': 8.706340375740045e-05, 'epoch': 0.09}) +[2026-10-03 14:52:25,101][starVLA.training.train_starvla][INFO] - Step 1280, Loss: {'train/loss': 0.532613217830658, 'train/grad_norm_pre_clip': 29.949201583862305, 'timing/data': 0.0001945489930221811, 'timing/model': 0.3701827249897178, 'learning_rate/qwen_vl_interface': 8.704250664883188e-06, 'learning_rate/action_model': 8.704250664883188e-05, 'epoch': 0.09}) +[2026-10-03 14:52:25,756][starVLA.training.train_starvla][INFO] - Step 1281, Loss: {'train/loss': 0.0047444067895412445, 'train/grad_norm_pre_clip': 2.3745381832122803, 'timing/data': 0.0002065599983325228, 'timing/model': 0.36801934499817435, 'learning_rate/qwen_vl_interface': 8.70215953411436e-06, 'learning_rate/action_model': 8.70215953411436e-05, 'epoch': 0.09}) +[2026-10-03 14:52:26,411][starVLA.training.train_starvla][INFO] - Step 1282, Loss: {'train/loss': 0.5957881212234497, 'train/grad_norm_pre_clip': 14.84969425201416, 'timing/data': 0.0002584790054243058, 'timing/model': 0.3696912149898708, 'learning_rate/qwen_vl_interface': 8.70006698429315e-06, 'learning_rate/action_model': 8.700066984293149e-05, 'epoch': 0.09}) +[2026-10-03 14:52:27,068][starVLA.training.train_starvla][INFO] - Step 1283, Loss: {'train/loss': 0.585375964641571, 'train/grad_norm_pre_clip': 160.513671875, 'timing/data': 0.00023932999465614557, 'timing/model': 0.3702347340004053, 'learning_rate/qwen_vl_interface': 8.697973016279722e-06, 'learning_rate/action_model': 8.697973016279722e-05, 'epoch': 0.09}) +[2026-10-03 14:52:27,928][starVLA.training.train_starvla][INFO] - Step 1284, Loss: {'train/loss': 0.0011959755793213844, 'train/grad_norm_pre_clip': 0.21151910722255707, 'timing/data': 0.00024231000861618668, 'timing/model': 0.5735214529995574, 'learning_rate/qwen_vl_interface': 8.695877630934827e-06, 'learning_rate/action_model': 8.695877630934826e-05, 'epoch': 0.09}) +[2026-10-03 14:52:28,582][starVLA.training.train_starvla][INFO] - Step 1285, Loss: {'train/loss': 0.0008050320902839303, 'train/grad_norm_pre_clip': 0.14065499603748322, 'timing/data': 0.00019329000497236848, 'timing/model': 0.3679402459965786, 'learning_rate/qwen_vl_interface': 8.693780829119801e-06, 'learning_rate/action_model': 8.693780829119801e-05, 'epoch': 0.09}) +[2026-10-03 14:52:29,238][starVLA.training.train_starvla][INFO] - Step 1286, Loss: {'train/loss': 0.14939764142036438, 'train/grad_norm_pre_clip': 78.66905975341797, 'timing/data': 0.00021552000544033945, 'timing/model': 0.369721915005357, 'learning_rate/qwen_vl_interface': 8.69168261169656e-06, 'learning_rate/action_model': 8.691682611696559e-05, 'epoch': 0.09}) +[2026-10-03 14:52:29,894][starVLA.training.train_starvla][INFO] - Step 1287, Loss: {'train/loss': 0.5712587833404541, 'train/grad_norm_pre_clip': 98.82838439941406, 'timing/data': 0.000274819991318509, 'timing/model': 0.3697531939978944, 'learning_rate/qwen_vl_interface': 8.689582979527599e-06, 'learning_rate/action_model': 8.689582979527599e-05, 'epoch': 0.09}) +[2026-10-03 14:52:30,550][starVLA.training.train_starvla][INFO] - Step 1288, Loss: {'train/loss': 0.009617309086024761, 'train/grad_norm_pre_clip': 13.101305961608887, 'timing/data': 0.00023586000315845013, 'timing/model': 0.3699376149888849, 'learning_rate/qwen_vl_interface': 8.687481933476e-06, 'learning_rate/action_model': 8.687481933476e-05, 'epoch': 0.09}) +[2026-10-03 14:52:31,204][starVLA.training.train_starvla][INFO] - Step 1289, Loss: {'train/loss': 0.1416713297367096, 'train/grad_norm_pre_clip': 76.7997817993164, 'timing/data': 0.0003234799951314926, 'timing/model': 0.36783067499345634, 'learning_rate/qwen_vl_interface': 8.685379474405423e-06, 'learning_rate/action_model': 8.685379474405423e-05, 'epoch': 0.09}) +[2026-10-03 14:52:32,050][starVLA.training.train_starvla][INFO] - Step 1290, Loss: {'train/loss': 0.0008901397814042866, 'train/grad_norm_pre_clip': 0.1116785854101181, 'timing/data': 0.0002154200046788901, 'timing/model': 0.3696307450009044, 'learning_rate/qwen_vl_interface': 8.68327560318011e-06, 'learning_rate/action_model': 8.68327560318011e-05, 'epoch': 0.09}) +[2026-10-03 14:52:32,707][starVLA.training.train_starvla][INFO] - Step 1291, Loss: {'train/loss': 0.10134314745664597, 'train/grad_norm_pre_clip': 86.87618255615234, 'timing/data': 0.00020257999130990356, 'timing/model': 0.37034258400672115, 'learning_rate/qwen_vl_interface': 8.681170320664883e-06, 'learning_rate/action_model': 8.681170320664883e-05, 'epoch': 0.09}) +[2026-10-03 14:52:33,364][starVLA.training.train_starvla][INFO] - Step 1292, Loss: {'train/loss': 0.28338807821273804, 'train/grad_norm_pre_clip': 26.096094131469727, 'timing/data': 0.00024865999876055866, 'timing/model': 0.3697573350073071, 'learning_rate/qwen_vl_interface': 8.679063627725143e-06, 'learning_rate/action_model': 8.679063627725144e-05, 'epoch': 0.09}) +[2026-10-03 14:52:34,020][starVLA.training.train_starvla][INFO] - Step 1293, Loss: {'train/loss': 0.004053268115967512, 'train/grad_norm_pre_clip': 0.6952046155929565, 'timing/data': 0.00033660000190138817, 'timing/model': 0.3681605150050018, 'learning_rate/qwen_vl_interface': 8.676955525226876e-06, 'learning_rate/action_model': 8.676955525226875e-05, 'epoch': 0.09}) +[2026-10-03 14:52:34,675][starVLA.training.train_starvla][INFO] - Step 1294, Loss: {'train/loss': 0.06810805201530457, 'train/grad_norm_pre_clip': 23.00602912902832, 'timing/data': 0.00024015000963117927, 'timing/model': 0.36948500499420334, 'learning_rate/qwen_vl_interface': 8.674846014036636e-06, 'learning_rate/action_model': 8.674846014036636e-05, 'epoch': 0.09}) +[2026-10-03 14:52:35,523][starVLA.training.train_starvla][INFO] - Step 1295, Loss: {'train/loss': 0.005747136659920216, 'train/grad_norm_pre_clip': 1.4933955669403076, 'timing/data': 0.0002663000050233677, 'timing/model': 0.36893307500577066, 'learning_rate/qwen_vl_interface': 8.672735095021569e-06, 'learning_rate/action_model': 8.672735095021569e-05, 'epoch': 0.09}) +[2026-10-03 14:52:36,179][starVLA.training.train_starvla][INFO] - Step 1296, Loss: {'train/loss': 0.012865167111158371, 'train/grad_norm_pre_clip': 5.232937335968018, 'timing/data': 0.0002757600013865158, 'timing/model': 0.3687584250001237, 'learning_rate/qwen_vl_interface': 8.670622769049393e-06, 'learning_rate/action_model': 8.670622769049392e-05, 'epoch': 0.09}) +[2026-10-03 14:52:36,832][starVLA.training.train_starvla][INFO] - Step 1297, Loss: {'train/loss': 0.07304241508245468, 'train/grad_norm_pre_clip': 69.81434631347656, 'timing/data': 0.00024275000032503158, 'timing/model': 0.36741825500212144, 'learning_rate/qwen_vl_interface': 8.668509036988406e-06, 'learning_rate/action_model': 8.668509036988405e-05, 'epoch': 0.09}) +[2026-10-03 14:52:37,488][starVLA.training.train_starvla][INFO] - Step 1298, Loss: {'train/loss': 0.00652724364772439, 'train/grad_norm_pre_clip': 5.70702600479126, 'timing/data': 0.0002476099907653406, 'timing/model': 0.37040140500175767, 'learning_rate/qwen_vl_interface': 8.66639389970748e-06, 'learning_rate/action_model': 8.666393899707479e-05, 'epoch': 0.1}) +[2026-10-03 14:52:38,122][starVLA.training.train_starvla][INFO] - Step 1299, Loss: {'train/loss': 0.04111772030591965, 'train/grad_norm_pre_clip': 34.6291389465332, 'timing/data': 0.00019891999545507133, 'timing/model': 0.34677521999401506, 'learning_rate/qwen_vl_interface': 8.66427735807607e-06, 'learning_rate/action_model': 8.664277358076069e-05, 'epoch': 0.1}) +[2026-10-03 14:52:38,974][starVLA.training.train_starvla][INFO] - Step 1300, Loss: {'train/loss': 0.42267853021621704, 'train/grad_norm_pre_clip': 43.50646209716797, 'timing/data': 0.00027905000024475157, 'timing/model': 0.5649338649964193, 'learning_rate/qwen_vl_interface': 8.662159412964208e-06, 'learning_rate/action_model': 8.662159412964207e-05, 'epoch': 0.1}) +[2026-10-03 14:52:39,631][starVLA.training.train_starvla][INFO] - Step 1301, Loss: {'train/loss': 0.09956598281860352, 'train/grad_norm_pre_clip': 25.292713165283203, 'timing/data': 0.0002547899930505082, 'timing/model': 0.3693734959961148, 'learning_rate/qwen_vl_interface': 8.660040065242499e-06, 'learning_rate/action_model': 8.660040065242498e-05, 'epoch': 0.1}) +[2026-10-03 14:52:40,289][starVLA.training.train_starvla][INFO] - Step 1302, Loss: {'train/loss': 0.0029484254773706198, 'train/grad_norm_pre_clip': 0.8350183367729187, 'timing/data': 0.00033012000494636595, 'timing/model': 0.3698716159997275, 'learning_rate/qwen_vl_interface': 8.657919315782127e-06, 'learning_rate/action_model': 8.657919315782127e-05, 'epoch': 0.1}) +[2026-10-03 14:52:40,947][starVLA.training.train_starvla][INFO] - Step 1303, Loss: {'train/loss': 0.005141208879649639, 'train/grad_norm_pre_clip': 2.2180283069610596, 'timing/data': 0.0002745300007518381, 'timing/model': 0.370238445000723, 'learning_rate/qwen_vl_interface': 8.655797165454853e-06, 'learning_rate/action_model': 8.655797165454853e-05, 'epoch': 0.1}) +[2026-10-03 14:52:41,607][starVLA.training.train_starvla][INFO] - Step 1304, Loss: {'train/loss': 0.10753370821475983, 'train/grad_norm_pre_clip': 156.8333740234375, 'timing/data': 0.00026436999905854464, 'timing/model': 0.3722393350035418, 'learning_rate/qwen_vl_interface': 8.653673615133013e-06, 'learning_rate/action_model': 8.653673615133013e-05, 'epoch': 0.1}) +[2026-10-03 14:52:42,457][starVLA.training.train_starvla][INFO] - Step 1305, Loss: {'train/loss': 0.30148470401763916, 'train/grad_norm_pre_clip': 142.42327880859375, 'timing/data': 0.0003311800101073459, 'timing/model': 0.5626632560015423, 'learning_rate/qwen_vl_interface': 8.651548665689519e-06, 'learning_rate/action_model': 8.651548665689518e-05, 'epoch': 0.1}) +[2026-10-03 14:52:43,090][starVLA.training.train_starvla][INFO] - Step 1306, Loss: {'train/loss': 0.6220130920410156, 'train/grad_norm_pre_clip': 36.9837646484375, 'timing/data': 0.0002414900081930682, 'timing/model': 0.3461645000061253, 'learning_rate/qwen_vl_interface': 8.649422317997855e-06, 'learning_rate/action_model': 8.649422317997856e-05, 'epoch': 0.1}) +[2026-10-03 14:52:43,747][starVLA.training.train_starvla][INFO] - Step 1307, Loss: {'train/loss': 0.019241612404584885, 'train/grad_norm_pre_clip': 5.300920009613037, 'timing/data': 0.00024958900758065283, 'timing/model': 0.36986949600395747, 'learning_rate/qwen_vl_interface': 8.647294572932082e-06, 'learning_rate/action_model': 8.647294572932082e-05, 'epoch': 0.1}) +[2026-10-03 14:52:44,405][starVLA.training.train_starvla][INFO] - Step 1308, Loss: {'train/loss': 0.03519264981150627, 'train/grad_norm_pre_clip': 46.71289825439453, 'timing/data': 0.0002442099939798936, 'timing/model': 0.3697246149968123, 'learning_rate/qwen_vl_interface': 8.64516543136684e-06, 'learning_rate/action_model': 8.64516543136684e-05, 'epoch': 0.1}) +[2026-10-03 14:52:45,061][starVLA.training.train_starvla][INFO] - Step 1309, Loss: {'train/loss': 0.02040775679051876, 'train/grad_norm_pre_clip': 29.790000915527344, 'timing/data': 0.0001975499908439815, 'timing/model': 0.36906109600386117, 'learning_rate/qwen_vl_interface': 8.643034894177335e-06, 'learning_rate/action_model': 8.643034894177335e-05, 'epoch': 0.1}) +[2026-10-03 14:52:45,719][starVLA.training.train_starvla][INFO] - Step 1310, Loss: {'train/loss': 0.01394932996481657, 'train/grad_norm_pre_clip': 11.07494831085205, 'timing/data': 0.00029258000722620636, 'timing/model': 0.37075784600165207, 'learning_rate/qwen_vl_interface': 8.64090296223935e-06, 'learning_rate/action_model': 8.64090296223935e-05, 'epoch': 0.1}) +[2026-10-03 14:52:46,548][starVLA.training.train_starvla][INFO] - Step 1311, Loss: {'train/loss': 0.000268651288934052, 'train/grad_norm_pre_clip': 0.05253402888774872, 'timing/data': 0.0003347099991515279, 'timing/model': 0.34752029000082985, 'learning_rate/qwen_vl_interface': 8.638769636429245e-06, 'learning_rate/action_model': 8.638769636429245e-05, 'epoch': 0.1}) +[2026-10-03 14:52:47,209][starVLA.training.train_starvla][INFO] - Step 1312, Loss: {'train/loss': 0.06015452370047569, 'train/grad_norm_pre_clip': 33.8939208984375, 'timing/data': 0.00021505000768229365, 'timing/model': 0.3701655959885102, 'learning_rate/qwen_vl_interface': 8.636634917623946e-06, 'learning_rate/action_model': 8.636634917623944e-05, 'epoch': 0.1}) +[2026-10-03 14:52:47,868][starVLA.training.train_starvla][INFO] - Step 1313, Loss: {'train/loss': 0.001870021689683199, 'train/grad_norm_pre_clip': 0.7857998013496399, 'timing/data': 0.0002720600023167208, 'timing/model': 0.3714271150092827, 'learning_rate/qwen_vl_interface': 8.634498806700955e-06, 'learning_rate/action_model': 8.634498806700954e-05, 'epoch': 0.1}) +[2026-10-03 14:52:48,530][starVLA.training.train_starvla][INFO] - Step 1314, Loss: {'train/loss': 0.3290449380874634, 'train/grad_norm_pre_clip': 94.31261444091797, 'timing/data': 0.0004154899943387136, 'timing/model': 0.37109900498762727, 'learning_rate/qwen_vl_interface': 8.63236130453835e-06, 'learning_rate/action_model': 8.63236130453835e-05, 'epoch': 0.1}) +[2026-10-03 14:52:49,190][starVLA.training.train_starvla][INFO] - Step 1315, Loss: {'train/loss': 0.00019910857372451574, 'train/grad_norm_pre_clip': 0.03997758775949478, 'timing/data': 0.0003794799995375797, 'timing/model': 0.3712339959893143, 'learning_rate/qwen_vl_interface': 8.630222412014777e-06, 'learning_rate/action_model': 8.630222412014776e-05, 'epoch': 0.1}) +[2026-10-03 14:52:50,037][starVLA.training.train_starvla][INFO] - Step 1316, Loss: {'train/loss': 0.6266326308250427, 'train/grad_norm_pre_clip': 33.52256774902344, 'timing/data': 0.0002530000056140125, 'timing/model': 0.5578608880023239, 'learning_rate/qwen_vl_interface': 8.628082130009451e-06, 'learning_rate/action_model': 8.628082130009451e-05, 'epoch': 0.1}) +[2026-10-03 14:52:50,697][starVLA.training.train_starvla][INFO] - Step 1317, Loss: {'train/loss': 0.26247015595436096, 'train/grad_norm_pre_clip': 147.95872497558594, 'timing/data': 0.0002497100067557767, 'timing/model': 0.3708238649996929, 'learning_rate/qwen_vl_interface': 8.62594045940216e-06, 'learning_rate/action_model': 8.62594045940216e-05, 'epoch': 0.1}) +[2026-10-03 14:52:51,359][starVLA.training.train_starvla][INFO] - Step 1318, Loss: {'train/loss': 0.00034011053503490984, 'train/grad_norm_pre_clip': 0.09583389014005661, 'timing/data': 0.00035909999860450625, 'timing/model': 0.3706443850096548, 'learning_rate/qwen_vl_interface': 8.623797401073272e-06, 'learning_rate/action_model': 8.623797401073271e-05, 'epoch': 0.1}) +[2026-10-03 14:52:52,018][starVLA.training.train_starvla][INFO] - Step 1319, Loss: {'train/loss': 0.0085682924836874, 'train/grad_norm_pre_clip': 6.731700897216797, 'timing/data': 0.0002525890013203025, 'timing/model': 0.37041643699922133, 'learning_rate/qwen_vl_interface': 8.62165295590371e-06, 'learning_rate/action_model': 8.621652955903709e-05, 'epoch': 0.1}) +[2026-10-03 14:52:52,655][starVLA.training.train_starvla][INFO] - Step 1320, Loss: {'train/loss': 0.0008111890638247132, 'train/grad_norm_pre_clip': 0.38393935561180115, 'timing/data': 0.0002555899991421029, 'timing/model': 0.37093160599761177, 'learning_rate/qwen_vl_interface': 8.619507124774975e-06, 'learning_rate/action_model': 8.619507124774975e-05, 'epoch': 0.1}) +[2026-10-03 14:52:53,502][starVLA.training.train_starvla][INFO] - Step 1321, Loss: {'train/loss': 0.15368439257144928, 'train/grad_norm_pre_clip': 114.87493133544922, 'timing/data': 0.00030725999386049807, 'timing/model': 0.5581180779990973, 'learning_rate/qwen_vl_interface': 8.61735990856914e-06, 'learning_rate/action_model': 8.617359908569139e-05, 'epoch': 0.1}) +[2026-10-03 14:52:54,160][starVLA.training.train_starvla][INFO] - Step 1322, Loss: {'train/loss': 0.00015464643365703523, 'train/grad_norm_pre_clip': 0.022226568311452866, 'timing/data': 0.00035180999839212745, 'timing/model': 0.36765417699643876, 'learning_rate/qwen_vl_interface': 8.615211308168842e-06, 'learning_rate/action_model': 8.61521130816884e-05, 'epoch': 0.1}) +[2026-10-03 14:52:54,811][starVLA.training.train_starvla][INFO] - Step 1323, Loss: {'train/loss': 0.21614424884319305, 'train/grad_norm_pre_clip': 66.0232925415039, 'timing/data': 0.0003028799983439967, 'timing/model': 0.3674633159971563, 'learning_rate/qwen_vl_interface': 8.613061324457289e-06, 'learning_rate/action_model': 8.613061324457288e-05, 'epoch': 0.1}) +[2026-10-03 14:52:55,461][starVLA.training.train_starvla][INFO] - Step 1324, Loss: {'train/loss': 0.14623674750328064, 'train/grad_norm_pre_clip': 131.74790954589844, 'timing/data': 0.00037449000228662044, 'timing/model': 0.3659670869965339, 'learning_rate/qwen_vl_interface': 8.610909958318262e-06, 'learning_rate/action_model': 8.610909958318261e-05, 'epoch': 0.1}) +[2026-10-03 14:52:56,092][starVLA.training.train_starvla][INFO] - Step 1325, Loss: {'train/loss': 0.6003695726394653, 'train/grad_norm_pre_clip': 96.67949676513672, 'timing/data': 0.00034548000257927924, 'timing/model': 0.36849227600032464, 'learning_rate/qwen_vl_interface': 8.608757210636101e-06, 'learning_rate/action_model': 8.608757210636101e-05, 'epoch': 0.1}) +[2026-10-03 14:52:56,941][starVLA.training.train_starvla][INFO] - Step 1326, Loss: {'train/loss': 0.9022085666656494, 'train/grad_norm_pre_clip': 174.16510009765625, 'timing/data': 0.0003517900040606037, 'timing/model': 0.5652101470041089, 'learning_rate/qwen_vl_interface': 8.606603082295724e-06, 'learning_rate/action_model': 8.606603082295724e-05, 'epoch': 0.1}) +[2026-10-03 14:52:57,593][starVLA.training.train_starvla][INFO] - Step 1327, Loss: {'train/loss': 0.1122230812907219, 'train/grad_norm_pre_clip': 162.1686248779297, 'timing/data': 0.0003258200013078749, 'timing/model': 0.36729417598689906, 'learning_rate/qwen_vl_interface': 8.60444757418261e-06, 'learning_rate/action_model': 8.604447574182609e-05, 'epoch': 0.1}) +[2026-10-03 14:52:58,220][starVLA.training.train_starvla][INFO] - Step 1328, Loss: {'train/loss': 0.0022513153962790966, 'train/grad_norm_pre_clip': 1.0337742567062378, 'timing/data': 0.0002769800048554316, 'timing/model': 0.36622562700358685, 'learning_rate/qwen_vl_interface': 8.602290687182805e-06, 'learning_rate/action_model': 8.602290687182806e-05, 'epoch': 0.1}) +[2026-10-03 14:52:58,874][starVLA.training.train_starvla][INFO] - Step 1329, Loss: {'train/loss': 0.0008851119200699031, 'train/grad_norm_pre_clip': 0.3235228955745697, 'timing/data': 0.0002266000083182007, 'timing/model': 0.3684805060038343, 'learning_rate/qwen_vl_interface': 8.600132422182929e-06, 'learning_rate/action_model': 8.600132422182928e-05, 'epoch': 0.1}) +[2026-10-03 14:52:59,524][starVLA.training.train_starvla][INFO] - Step 1330, Loss: {'train/loss': 0.593916654586792, 'train/grad_norm_pre_clip': 72.12983703613281, 'timing/data': 0.0002620599989313632, 'timing/model': 0.366554067004472, 'learning_rate/qwen_vl_interface': 8.597972780070156e-06, 'learning_rate/action_model': 8.597972780070155e-05, 'epoch': 0.1}) +[2026-10-03 14:53:00,374][starVLA.training.train_starvla][INFO] - Step 1331, Loss: {'train/loss': 0.0017296315636485815, 'train/grad_norm_pre_clip': 2.7601287364959717, 'timing/data': 0.0003369799960637465, 'timing/model': 0.5651470269949641, 'learning_rate/qwen_vl_interface': 8.595811761732239e-06, 'learning_rate/action_model': 8.595811761732238e-05, 'epoch': 0.1}) +[2026-10-03 14:53:01,023][starVLA.training.train_starvla][INFO] - Step 1332, Loss: {'train/loss': 0.0152180977165699, 'train/grad_norm_pre_clip': 45.643157958984375, 'timing/data': 0.00024649999977555126, 'timing/model': 0.36543662699114066, 'learning_rate/qwen_vl_interface': 8.593649368057489e-06, 'learning_rate/action_model': 8.59364936805749e-05, 'epoch': 0.1}) +[2026-10-03 14:53:01,675][starVLA.training.train_starvla][INFO] - Step 1333, Loss: {'train/loss': 0.0004758876748383045, 'train/grad_norm_pre_clip': 0.19064289331436157, 'timing/data': 0.0002560699940659106, 'timing/model': 0.36800210700312164, 'learning_rate/qwen_vl_interface': 8.591485599934786e-06, 'learning_rate/action_model': 8.591485599934785e-05, 'epoch': 0.1}) +[2026-10-03 14:53:02,324][starVLA.training.train_starvla][INFO] - Step 1334, Loss: {'train/loss': 0.0002760633360594511, 'train/grad_norm_pre_clip': 0.06946359574794769, 'timing/data': 0.00034633999166544527, 'timing/model': 0.36526195700571407, 'learning_rate/qwen_vl_interface': 8.58932045825357e-06, 'learning_rate/action_model': 8.589320458253569e-05, 'epoch': 0.1}) +[2026-10-03 14:53:02,972][starVLA.training.train_starvla][INFO] - Step 1335, Loss: {'train/loss': 0.0002681428741198033, 'train/grad_norm_pre_clip': 0.12232611328363419, 'timing/data': 0.00020385000971145928, 'timing/model': 0.3651640669995686, 'learning_rate/qwen_vl_interface': 8.58715394390385e-06, 'learning_rate/action_model': 8.58715394390385e-05, 'epoch': 0.1}) +[2026-10-03 14:53:03,621][starVLA.training.train_starvla][INFO] - Step 1336, Loss: {'train/loss': 0.00025555145111866295, 'train/grad_norm_pre_clip': 0.0311402827501297, 'timing/data': 0.0002594599936855957, 'timing/model': 0.3653248870105017, 'learning_rate/qwen_vl_interface': 8.5849860577762e-06, 'learning_rate/action_model': 8.5849860577762e-05, 'epoch': 0.1}) +[2026-10-03 14:53:04,475][starVLA.training.train_starvla][INFO] - Step 1337, Loss: {'train/loss': 0.0003264394181314856, 'train/grad_norm_pre_clip': 0.12850821018218994, 'timing/data': 0.0003114499995717779, 'timing/model': 0.3670373470085906, 'learning_rate/qwen_vl_interface': 8.582816800761756e-06, 'learning_rate/action_model': 8.582816800761755e-05, 'epoch': 0.1}) +[2026-10-03 14:53:05,123][starVLA.training.train_starvla][INFO] - Step 1338, Loss: {'train/loss': 1.1611698865890503, 'train/grad_norm_pre_clip': 60.53227615356445, 'timing/data': 0.000302400003420189, 'timing/model': 0.3646461479947902, 'learning_rate/qwen_vl_interface': 8.580646173752214e-06, 'learning_rate/action_model': 8.580646173752213e-05, 'epoch': 0.1}) +[2026-10-03 14:53:05,772][starVLA.training.train_starvla][INFO] - Step 1339, Loss: {'train/loss': 0.011720209382474422, 'train/grad_norm_pre_clip': 13.728755950927734, 'timing/data': 0.00025346998882014304, 'timing/model': 0.36609274699003436, 'learning_rate/qwen_vl_interface': 8.578474177639841e-06, 'learning_rate/action_model': 8.578474177639842e-05, 'epoch': 0.1}) +[2026-10-03 14:53:06,420][starVLA.training.train_starvla][INFO] - Step 1340, Loss: {'train/loss': 0.001280102296732366, 'train/grad_norm_pre_clip': 1.0916818380355835, 'timing/data': 0.0002650300011737272, 'timing/model': 0.3651978970010532, 'learning_rate/qwen_vl_interface': 8.576300813317463e-06, 'learning_rate/action_model': 8.576300813317462e-05, 'epoch': 0.1}) +[2026-10-03 14:53:07,072][starVLA.training.train_starvla][INFO] - Step 1341, Loss: {'train/loss': 0.00016016341396607459, 'train/grad_norm_pre_clip': 0.02673918940126896, 'timing/data': 0.00028080999618396163, 'timing/model': 0.36773675700533204, 'learning_rate/qwen_vl_interface': 8.574126081678463e-06, 'learning_rate/action_model': 8.574126081678462e-05, 'epoch': 0.1}) +[2026-10-03 14:53:07,915][starVLA.training.train_starvla][INFO] - Step 1342, Loss: {'train/loss': 0.0008886380819603801, 'train/grad_norm_pre_clip': 0.789421558380127, 'timing/data': 0.00023700999736320227, 'timing/model': 0.5596985589945689, 'learning_rate/qwen_vl_interface': 8.571949983616796e-06, 'learning_rate/action_model': 8.571949983616795e-05, 'epoch': 0.1}) +[2026-10-03 14:53:08,564][starVLA.training.train_starvla][INFO] - Step 1343, Loss: {'train/loss': 0.0002831443562172353, 'train/grad_norm_pre_clip': 0.044127315282821655, 'timing/data': 0.00031177999335341156, 'timing/model': 0.36576561800029594, 'learning_rate/qwen_vl_interface': 8.569772520026971e-06, 'learning_rate/action_model': 8.569772520026971e-05, 'epoch': 0.1}) +[2026-10-03 14:53:09,212][starVLA.training.train_starvla][INFO] - Step 1344, Loss: {'train/loss': 0.0029489086009562016, 'train/grad_norm_pre_clip': 1.7803996801376343, 'timing/data': 0.0002818500070134178, 'timing/model': 0.3652707669971278, 'learning_rate/qwen_vl_interface': 8.567593691804061e-06, 'learning_rate/action_model': 8.567593691804061e-05, 'epoch': 0.1}) +[2026-10-03 14:53:09,864][starVLA.training.train_starvla][INFO] - Step 1345, Loss: {'train/loss': 0.7240279912948608, 'train/grad_norm_pre_clip': 14.550765991210938, 'timing/data': 0.00029308001103345305, 'timing/model': 0.3677549669955624, 'learning_rate/qwen_vl_interface': 8.565413499843703e-06, 'learning_rate/action_model': 8.565413499843703e-05, 'epoch': 0.1}) +[2026-10-03 14:53:10,514][starVLA.training.train_starvla][INFO] - Step 1346, Loss: {'train/loss': 0.0003303055127616972, 'train/grad_norm_pre_clip': 0.13834817707538605, 'timing/data': 0.00029940000968053937, 'timing/model': 0.36585714800457936, 'learning_rate/qwen_vl_interface': 8.563231945042087e-06, 'learning_rate/action_model': 8.563231945042088e-05, 'epoch': 0.1}) +[2026-10-03 14:53:11,366][starVLA.training.train_starvla][INFO] - Step 1347, Loss: {'train/loss': 0.4878266453742981, 'train/grad_norm_pre_clip': 39.36823272705078, 'timing/data': 0.0002190100058214739, 'timing/model': 0.5685534970107256, 'learning_rate/qwen_vl_interface': 8.56104902829597e-06, 'learning_rate/action_model': 8.56104902829597e-05, 'epoch': 0.1}) +[2026-10-03 14:53:12,015][starVLA.training.train_starvla][INFO] - Step 1348, Loss: {'train/loss': 0.00012593503925018013, 'train/grad_norm_pre_clip': 0.014433645643293858, 'timing/data': 0.0002436299982946366, 'timing/model': 0.3656827980012167, 'learning_rate/qwen_vl_interface': 8.558864750502669e-06, 'learning_rate/action_model': 8.558864750502668e-05, 'epoch': 0.1}) +[2026-10-03 14:53:12,644][starVLA.training.train_starvla][INFO] - Step 1349, Loss: {'train/loss': 0.00020946242148056626, 'train/grad_norm_pre_clip': 0.02861958183348179, 'timing/data': 0.0002828599972417578, 'timing/model': 0.36792379699181765, 'learning_rate/qwen_vl_interface': 8.556679112560055e-06, 'learning_rate/action_model': 8.556679112560054e-05, 'epoch': 0.1}) +[2026-10-03 14:53:13,293][starVLA.training.train_starvla][INFO] - Step 1350, Loss: {'train/loss': 0.000935006479267031, 'train/grad_norm_pre_clip': 0.427508682012558, 'timing/data': 0.0002369599969824776, 'timing/model': 0.36541641798976343, 'learning_rate/qwen_vl_interface': 8.554492115366562e-06, 'learning_rate/action_model': 8.554492115366561e-05, 'epoch': 0.1}) +[2026-10-03 14:53:13,943][starVLA.training.train_starvla][INFO] - Step 1351, Loss: {'train/loss': 0.001033521955832839, 'train/grad_norm_pre_clip': 0.48350003361701965, 'timing/data': 0.00026590000197757035, 'timing/model': 0.3653905079991091, 'learning_rate/qwen_vl_interface': 8.55230375982118e-06, 'learning_rate/action_model': 8.55230375982118e-05, 'epoch': 0.1}) +[2026-10-03 14:53:14,593][starVLA.training.train_starvla][INFO] - Step 1352, Loss: {'train/loss': 0.000671044981572777, 'train/grad_norm_pre_clip': 0.19083379209041595, 'timing/data': 0.00023567001335322857, 'timing/model': 0.3659125879930798, 'learning_rate/qwen_vl_interface': 8.550114046823463e-06, 'learning_rate/action_model': 8.550114046823463e-05, 'epoch': 0.1}) +[2026-10-03 14:53:15,445][starVLA.training.train_starvla][INFO] - Step 1353, Loss: {'train/loss': 0.0001807891676435247, 'train/grad_norm_pre_clip': 0.019887618720531464, 'timing/data': 0.0002523800067137927, 'timing/model': 0.3682863670110237, 'learning_rate/qwen_vl_interface': 8.547922977273518e-06, 'learning_rate/action_model': 8.547922977273518e-05, 'epoch': 0.1}) +[2026-10-03 14:53:16,095][starVLA.training.train_starvla][INFO] - Step 1354, Loss: {'train/loss': 0.001998346531763673, 'train/grad_norm_pre_clip': 1.1284562349319458, 'timing/data': 0.00020120000408496708, 'timing/model': 0.3653582980041392, 'learning_rate/qwen_vl_interface': 8.54573055207201e-06, 'learning_rate/action_model': 8.545730552072008e-05, 'epoch': 0.1}) +[2026-10-03 14:53:16,721][starVLA.training.train_starvla][INFO] - Step 1355, Loss: {'train/loss': 0.0005272527923807502, 'train/grad_norm_pre_clip': 0.08880161494016647, 'timing/data': 0.00022401000023819506, 'timing/model': 0.3425863330048742, 'learning_rate/qwen_vl_interface': 8.543536772120161e-06, 'learning_rate/action_model': 8.54353677212016e-05, 'epoch': 0.1}) +[2026-10-03 14:53:17,370][starVLA.training.train_starvla][INFO] - Step 1356, Loss: {'train/loss': 0.0006960771861486137, 'train/grad_norm_pre_clip': 0.7380182147026062, 'timing/data': 0.0003164100053254515, 'timing/model': 0.3657392780005466, 'learning_rate/qwen_vl_interface': 8.541341638319754e-06, 'learning_rate/action_model': 8.541341638319754e-05, 'epoch': 0.1}) +[2026-10-03 14:53:18,022][starVLA.training.train_starvla][INFO] - Step 1357, Loss: {'train/loss': 0.00021450602798722684, 'train/grad_norm_pre_clip': 0.022562550380825996, 'timing/data': 0.00023430000874213874, 'timing/model': 0.3673711969895521, 'learning_rate/qwen_vl_interface': 8.539145151573126e-06, 'learning_rate/action_model': 8.539145151573125e-05, 'epoch': 0.1}) +[2026-10-03 14:53:18,843][starVLA.training.train_starvla][INFO] - Step 1358, Loss: {'train/loss': 0.00046652695164084435, 'train/grad_norm_pre_clip': 0.17679445445537567, 'timing/data': 0.00025633000768721104, 'timing/model': 0.5370988040085649, 'learning_rate/qwen_vl_interface': 8.536947312783167e-06, 'learning_rate/action_model': 8.536947312783166e-05, 'epoch': 0.1}) +[2026-10-03 14:53:19,491][starVLA.training.train_starvla][INFO] - Step 1359, Loss: {'train/loss': 0.00017627709894441068, 'train/grad_norm_pre_clip': 0.028456805273890495, 'timing/data': 0.0003028899955097586, 'timing/model': 0.36532487800286617, 'learning_rate/qwen_vl_interface': 8.534748122853326e-06, 'learning_rate/action_model': 8.534748122853325e-05, 'epoch': 0.1}) +[2026-10-03 14:53:20,140][starVLA.training.train_starvla][INFO] - Step 1360, Loss: {'train/loss': 0.0003199189668521285, 'train/grad_norm_pre_clip': 0.07925469428300858, 'timing/data': 0.00020073000632692128, 'timing/model': 0.3653366390062729, 'learning_rate/qwen_vl_interface': 8.532547582687609e-06, 'learning_rate/action_model': 8.532547582687608e-05, 'epoch': 0.1}) +[2026-10-03 14:53:20,793][starVLA.training.train_starvla][INFO] - Step 1361, Loss: {'train/loss': 0.012938189320266247, 'train/grad_norm_pre_clip': 15.37630558013916, 'timing/data': 0.00020419999782461673, 'timing/model': 0.36844599699543323, 'learning_rate/qwen_vl_interface': 8.530345693190573e-06, 'learning_rate/action_model': 8.530345693190572e-05, 'epoch': 0.1}) +[2026-10-03 14:53:21,443][starVLA.training.train_starvla][INFO] - Step 1362, Loss: {'train/loss': 0.00013593872427009046, 'train/grad_norm_pre_clip': 0.013709530234336853, 'timing/data': 0.0002407300053164363, 'timing/model': 0.36590185899694916, 'learning_rate/qwen_vl_interface': 8.528142455267328e-06, 'learning_rate/action_model': 8.528142455267329e-05, 'epoch': 0.1}) +[2026-10-03 14:53:22,293][starVLA.training.train_starvla][INFO] - Step 1363, Loss: {'train/loss': 0.00016809703083708882, 'train/grad_norm_pre_clip': 0.022424595430493355, 'timing/data': 0.0003426100010983646, 'timing/model': 0.5660428190021776, 'learning_rate/qwen_vl_interface': 8.52593786982355e-06, 'learning_rate/action_model': 8.525937869823551e-05, 'epoch': 0.1}) +[2026-10-03 14:53:22,942][starVLA.training.train_starvla][INFO] - Step 1364, Loss: {'train/loss': 0.0003449915675446391, 'train/grad_norm_pre_clip': 0.1683865487575531, 'timing/data': 0.0002653600095072761, 'timing/model': 0.3661943479964975, 'learning_rate/qwen_vl_interface': 8.523731937765456e-06, 'learning_rate/action_model': 8.523731937765456e-05, 'epoch': 0.1}) +[2026-10-03 14:53:23,594][starVLA.training.train_starvla][INFO] - Step 1365, Loss: {'train/loss': 8.878855442162603e-05, 'train/grad_norm_pre_clip': 0.008799136616289616, 'timing/data': 0.0002499090041965246, 'timing/model': 0.3673456079995958, 'learning_rate/qwen_vl_interface': 8.521524659999823e-06, 'learning_rate/action_model': 8.521524659999822e-05, 'epoch': 0.1}) +[2026-10-03 14:53:24,256][starVLA.training.train_starvla][INFO] - Step 1366, Loss: {'train/loss': 0.00011592377995839342, 'train/grad_norm_pre_clip': 0.014223220758140087, 'timing/data': 0.00027612999838311225, 'timing/model': 0.3654391680029221, 'learning_rate/qwen_vl_interface': 8.519316037433976e-06, 'learning_rate/action_model': 8.519316037433977e-05, 'epoch': 0.1}) +[2026-10-03 14:53:24,905][starVLA.training.train_starvla][INFO] - Step 1367, Loss: {'train/loss': 9.349710308015347e-05, 'train/grad_norm_pre_clip': 0.019816450774669647, 'timing/data': 0.0004192500055069104, 'timing/model': 0.36484396799642127, 'learning_rate/qwen_vl_interface': 8.517106070975801e-06, 'learning_rate/action_model': 8.5171060709758e-05, 'epoch': 0.1}) +[2026-10-03 14:53:25,555][starVLA.training.train_starvla][INFO] - Step 1368, Loss: {'train/loss': 0.00022548824199475348, 'train/grad_norm_pre_clip': 0.3473585546016693, 'timing/data': 0.00026119999529328197, 'timing/model': 0.36563985800603405, 'learning_rate/qwen_vl_interface': 8.514894761533729e-06, 'learning_rate/action_model': 8.514894761533728e-05, 'epoch': 0.1}) +[2026-10-03 14:53:26,402][starVLA.training.train_starvla][INFO] - Step 1369, Loss: {'train/loss': 0.0002484864671714604, 'train/grad_norm_pre_clip': 0.07723454385995865, 'timing/data': 0.00027288999990560114, 'timing/model': 0.36881230799190234, 'learning_rate/qwen_vl_interface': 8.512682110016747e-06, 'learning_rate/action_model': 8.512682110016747e-05, 'epoch': 0.1}) +[2026-10-03 14:53:27,055][starVLA.training.train_starvla][INFO] - Step 1370, Loss: {'train/loss': 0.4504218101501465, 'train/grad_norm_pre_clip': 104.90077209472656, 'timing/data': 0.0003158899926347658, 'timing/model': 0.36717006799881347, 'learning_rate/qwen_vl_interface': 8.51046811733439e-06, 'learning_rate/action_model': 8.51046811733439e-05, 'epoch': 0.1}) +[2026-10-03 14:53:27,684][starVLA.training.train_starvla][INFO] - Step 1371, Loss: {'train/loss': 0.84380704164505, 'train/grad_norm_pre_clip': 149.01426696777344, 'timing/data': 0.00026716000866144896, 'timing/model': 0.3655645480030216, 'learning_rate/qwen_vl_interface': 8.50825278439675e-06, 'learning_rate/action_model': 8.508252784396749e-05, 'epoch': 0.1}) +[2026-10-03 14:53:28,334][starVLA.training.train_starvla][INFO] - Step 1372, Loss: {'train/loss': 0.5861647129058838, 'train/grad_norm_pre_clip': 27.12210464477539, 'timing/data': 0.00020042998949065804, 'timing/model': 0.3655886690103216, 'learning_rate/qwen_vl_interface': 8.506036112114465e-06, 'learning_rate/action_model': 8.506036112114464e-05, 'epoch': 0.1}) +[2026-10-03 14:53:28,989][starVLA.training.train_starvla][INFO] - Step 1373, Loss: {'train/loss': 7.474640005966648e-05, 'train/grad_norm_pre_clip': 0.00890283565968275, 'timing/data': 0.00037432000681292266, 'timing/model': 0.3693348079978023, 'learning_rate/qwen_vl_interface': 8.503818101398725e-06, 'learning_rate/action_model': 8.503818101398725e-05, 'epoch': 0.1}) +[2026-10-03 14:53:29,839][starVLA.training.train_starvla][INFO] - Step 1374, Loss: {'train/loss': 5.498257087310776e-05, 'train/grad_norm_pre_clip': 0.009992695413529873, 'timing/data': 0.00023818000045139343, 'timing/model': 0.5656514900038019, 'learning_rate/qwen_vl_interface': 8.501598753161273e-06, 'learning_rate/action_model': 8.501598753161272e-05, 'epoch': 0.1}) +[2026-10-03 14:53:30,491][starVLA.training.train_starvla][INFO] - Step 1375, Loss: {'train/loss': 0.00012386060552671552, 'train/grad_norm_pre_clip': 0.019312074407935143, 'timing/data': 0.0003584000078262761, 'timing/model': 0.36679062800249085, 'learning_rate/qwen_vl_interface': 8.499378068314395e-06, 'learning_rate/action_model': 8.499378068314395e-05, 'epoch': 0.1}) +[2026-10-03 14:53:31,143][starVLA.training.train_starvla][INFO] - Step 1376, Loss: {'train/loss': 0.06370487809181213, 'train/grad_norm_pre_clip': 168.10018920898438, 'timing/data': 0.00028760998975485563, 'timing/model': 0.3675777280004695, 'learning_rate/qwen_vl_interface': 8.497156047770932e-06, 'learning_rate/action_model': 8.497156047770933e-05, 'epoch': 0.1}) +[2026-10-03 14:53:31,797][starVLA.training.train_starvla][INFO] - Step 1377, Loss: {'train/loss': 8.006696589291096e-05, 'train/grad_norm_pre_clip': 0.009062494151294231, 'timing/data': 0.0002036200021393597, 'timing/model': 0.36834359799104277, 'learning_rate/qwen_vl_interface': 8.494932692444276e-06, 'learning_rate/action_model': 8.494932692444276e-05, 'epoch': 0.1}) +[2026-10-03 14:53:32,446][starVLA.training.train_starvla][INFO] - Step 1378, Loss: {'train/loss': 1.2637357711791992, 'train/grad_norm_pre_clip': 45.47425842285156, 'timing/data': 0.0002608299982966855, 'timing/model': 0.3658122579945484, 'learning_rate/qwen_vl_interface': 8.492708003248362e-06, 'learning_rate/action_model': 8.492708003248362e-05, 'epoch': 0.1}) +[2026-10-03 14:53:33,096][starVLA.training.train_starvla][INFO] - Step 1379, Loss: {'train/loss': 0.0018317806534469128, 'train/grad_norm_pre_clip': 6.882647514343262, 'timing/data': 0.00026169999910052866, 'timing/model': 0.365740998997353, 'learning_rate/qwen_vl_interface': 8.490481981097674e-06, 'learning_rate/action_model': 8.490481981097674e-05, 'epoch': 0.1}) +[2026-10-03 14:53:33,941][starVLA.training.train_starvla][INFO] - Step 1380, Loss: {'train/loss': 0.0001445548259653151, 'train/grad_norm_pre_clip': 0.017570024356245995, 'timing/data': 0.0003694000042742118, 'timing/model': 0.3655963189958129, 'learning_rate/qwen_vl_interface': 8.488254626907249e-06, 'learning_rate/action_model': 8.488254626907249e-05, 'epoch': 0.1}) +[2026-10-03 14:53:34,594][starVLA.training.train_starvla][INFO] - Step 1381, Loss: {'train/loss': 1.1767373085021973, 'train/grad_norm_pre_clip': 27.65150260925293, 'timing/data': 0.00032341000041924417, 'timing/model': 0.3688018179964274, 'learning_rate/qwen_vl_interface': 8.486025941592664e-06, 'learning_rate/action_model': 8.486025941592664e-05, 'epoch': 0.1}) +[2026-10-03 14:53:35,243][starVLA.training.train_starvla][INFO] - Step 1382, Loss: {'train/loss': 0.00012411439092829823, 'train/grad_norm_pre_clip': 0.015013755299150944, 'timing/data': 0.00025489000836387277, 'timing/model': 0.3654737680044491, 'learning_rate/qwen_vl_interface': 8.483795926070054e-06, 'learning_rate/action_model': 8.483795926070052e-05, 'epoch': 0.1}) +[2026-10-03 14:53:35,894][starVLA.training.train_starvla][INFO] - Step 1383, Loss: {'train/loss': 0.3947162628173828, 'train/grad_norm_pre_clip': 40.090675354003906, 'timing/data': 0.0003233199968235567, 'timing/model': 0.365822178995586, 'learning_rate/qwen_vl_interface': 8.481564581256087e-06, 'learning_rate/action_model': 8.481564581256087e-05, 'epoch': 0.1}) +[2026-10-03 14:53:36,542][starVLA.training.train_starvla][INFO] - Step 1384, Loss: {'train/loss': 0.2570632994174957, 'train/grad_norm_pre_clip': 90.24063110351562, 'timing/data': 0.00024931000370997936, 'timing/model': 0.3650167190062348, 'learning_rate/qwen_vl_interface': 8.479331908067987e-06, 'learning_rate/action_model': 8.479331908067987e-05, 'epoch': 0.1}) +[2026-10-03 14:53:37,388][starVLA.training.train_starvla][INFO] - Step 1385, Loss: {'train/loss': 0.0030142460018396378, 'train/grad_norm_pre_clip': 2.6035783290863037, 'timing/data': 0.0002621500025270507, 'timing/model': 0.36762827799248043, 'learning_rate/qwen_vl_interface': 8.477097907423524e-06, 'learning_rate/action_model': 8.477097907423524e-05, 'epoch': 0.1}) +[2026-10-03 14:53:38,013][starVLA.training.train_starvla][INFO] - Step 1386, Loss: {'train/loss': 0.5779454112052917, 'train/grad_norm_pre_clip': 40.90806198120117, 'timing/data': 0.00026667000201996416, 'timing/model': 0.3644304089975776, 'learning_rate/qwen_vl_interface': 8.474862580241006e-06, 'learning_rate/action_model': 8.474862580241006e-05, 'epoch': 0.1}) +[2026-10-03 14:53:38,617][starVLA.training.train_starvla][INFO] - Step 1387, Loss: {'train/loss': 0.011343037709593773, 'train/grad_norm_pre_clip': 9.205013275146484, 'timing/data': 0.000325189990689978, 'timing/model': 0.3427142529981211, 'learning_rate/qwen_vl_interface': 8.472625927439299e-06, 'learning_rate/action_model': 8.472625927439299e-05, 'epoch': 0.1}) +[2026-10-03 14:53:39,265][starVLA.training.train_starvla][INFO] - Step 1388, Loss: {'train/loss': 0.3926289677619934, 'train/grad_norm_pre_clip': 117.65632629394531, 'timing/data': 0.0002889500028686598, 'timing/model': 0.3650989989982918, 'learning_rate/qwen_vl_interface': 8.4703879499378e-06, 'learning_rate/action_model': 8.4703879499378e-05, 'epoch': 0.1}) +[2026-10-03 14:53:39,917][starVLA.training.train_starvla][INFO] - Step 1389, Loss: {'train/loss': 0.008036768063902855, 'train/grad_norm_pre_clip': 3.612619400024414, 'timing/data': 0.00028202999965287745, 'timing/model': 0.3674907589884242, 'learning_rate/qwen_vl_interface': 8.46814864865646e-06, 'learning_rate/action_model': 8.46814864865646e-05, 'epoch': 0.1}) +[2026-10-03 14:53:40,765][starVLA.training.train_starvla][INFO] - Step 1390, Loss: {'train/loss': 0.002853754907846451, 'train/grad_norm_pre_clip': 0.3651416599750519, 'timing/data': 0.00025863001064863056, 'timing/model': 0.5643206099921372, 'learning_rate/qwen_vl_interface': 8.465908024515771e-06, 'learning_rate/action_model': 8.46590802451577e-05, 'epoch': 0.1}) +[2026-10-03 14:53:41,419][starVLA.training.train_starvla][INFO] - Step 1391, Loss: {'train/loss': 0.1033094972372055, 'train/grad_norm_pre_clip': 183.1060028076172, 'timing/data': 0.00033907999750226736, 'timing/model': 0.36617357800423633, 'learning_rate/qwen_vl_interface': 8.463666078436769e-06, 'learning_rate/action_model': 8.463666078436769e-05, 'epoch': 0.1}) +[2026-10-03 14:53:42,067][starVLA.training.train_starvla][INFO] - Step 1392, Loss: {'train/loss': 0.0024986634962260723, 'train/grad_norm_pre_clip': 0.34897395968437195, 'timing/data': 0.0002061300037894398, 'timing/model': 0.36537217001023237, 'learning_rate/qwen_vl_interface': 8.46142281134103e-06, 'learning_rate/action_model': 8.461422811341031e-05, 'epoch': 0.1}) +[2026-10-03 14:53:42,719][starVLA.training.train_starvla][INFO] - Step 1393, Loss: {'train/loss': 0.004696377087384462, 'train/grad_norm_pre_clip': 0.6381997466087341, 'timing/data': 0.00030137000430841, 'timing/model': 0.3677462579944404, 'learning_rate/qwen_vl_interface': 8.459178224150683e-06, 'learning_rate/action_model': 8.459178224150683e-05, 'epoch': 0.1}) +[2026-10-03 14:53:43,367][starVLA.training.train_starvla][INFO] - Step 1394, Loss: {'train/loss': 0.35175076127052307, 'train/grad_norm_pre_clip': 115.2396469116211, 'timing/data': 0.0002148400089936331, 'timing/model': 0.36527856899192557, 'learning_rate/qwen_vl_interface': 8.456932317788389e-06, 'learning_rate/action_model': 8.456932317788389e-05, 'epoch': 0.1}) +[2026-10-03 14:53:44,212][starVLA.training.train_starvla][INFO] - Step 1395, Loss: {'train/loss': 0.0029437262564897537, 'train/grad_norm_pre_clip': 2.4161109924316406, 'timing/data': 0.00020562000281643122, 'timing/model': 0.5612239809997845, 'learning_rate/qwen_vl_interface': 8.454685093177356e-06, 'learning_rate/action_model': 8.454685093177354e-05, 'epoch': 0.1}) +[2026-10-03 14:53:44,861][starVLA.training.train_starvla][INFO] - Step 1396, Loss: {'train/loss': 0.000773310661315918, 'train/grad_norm_pre_clip': 0.09775151312351227, 'timing/data': 0.0003045899939024821, 'timing/model': 0.36530158900131937, 'learning_rate/qwen_vl_interface': 8.452436551241328e-06, 'learning_rate/action_model': 8.452436551241328e-05, 'epoch': 0.1}) +[2026-10-03 14:53:45,512][starVLA.training.train_starvla][INFO] - Step 1397, Loss: {'train/loss': 0.08541814982891083, 'train/grad_norm_pre_clip': 68.83663940429688, 'timing/data': 0.00019550000433810055, 'timing/model': 0.36821293900720775, 'learning_rate/qwen_vl_interface': 8.450186692904605e-06, 'learning_rate/action_model': 8.450186692904606e-05, 'epoch': 0.1}) +[2026-10-03 14:53:46,161][starVLA.training.train_starvla][INFO] - Step 1398, Loss: {'train/loss': 0.5651479363441467, 'train/grad_norm_pre_clip': 75.48002624511719, 'timing/data': 0.00019859999883919954, 'timing/model': 0.36546912899939343, 'learning_rate/qwen_vl_interface': 8.447935519092014e-06, 'learning_rate/action_model': 8.447935519092014e-05, 'epoch': 0.1}) +[2026-10-03 14:53:46,810][starVLA.training.train_starvla][INFO] - Step 1399, Loss: {'train/loss': 0.0003050318919122219, 'train/grad_norm_pre_clip': 0.03347199037671089, 'timing/data': 0.00020283000776544213, 'timing/model': 0.3658424999885028, 'learning_rate/qwen_vl_interface': 8.445683030728925e-06, 'learning_rate/action_model': 8.445683030728925e-05, 'epoch': 0.1}) +[2026-10-03 14:53:47,457][starVLA.training.train_starvla][INFO] - Step 1400, Loss: {'train/loss': 0.00021883807494305074, 'train/grad_norm_pre_clip': 0.029032288119196892, 'timing/data': 0.00019597000209614635, 'timing/model': 0.36495782900601625, 'learning_rate/qwen_vl_interface': 8.443429228741256e-06, 'learning_rate/action_model': 8.443429228741255e-05, 'epoch': 0.1}) +[2026-10-03 14:53:48,313][starVLA.training.train_starvla][INFO] - Step 1401, Loss: {'train/loss': 0.13024888932704926, 'train/grad_norm_pre_clip': 125.40191650390625, 'timing/data': 0.00019595999037846923, 'timing/model': 0.368845778008108, 'learning_rate/qwen_vl_interface': 8.441174114055457e-06, 'learning_rate/action_model': 8.441174114055456e-05, 'epoch': 0.1}) +[2026-10-03 14:53:48,964][starVLA.training.train_starvla][INFO] - Step 1402, Loss: {'train/loss': 0.0009345525177195668, 'train/grad_norm_pre_clip': 0.33792901039123535, 'timing/data': 0.00021826000011060387, 'timing/model': 0.3669093589996919, 'learning_rate/qwen_vl_interface': 8.43891768759852e-06, 'learning_rate/action_model': 8.43891768759852e-05, 'epoch': 0.1}) +[2026-10-03 14:53:49,616][starVLA.training.train_starvla][INFO] - Step 1403, Loss: {'train/loss': 0.00016516726464033127, 'train/grad_norm_pre_clip': 0.020153148099780083, 'timing/data': 0.0002001590037252754, 'timing/model': 0.36710495001170784, 'learning_rate/qwen_vl_interface': 8.436659950297978e-06, 'learning_rate/action_model': 8.436659950297977e-05, 'epoch': 0.1}) +[2026-10-03 14:53:50,263][starVLA.training.train_starvla][INFO] - Step 1404, Loss: {'train/loss': 0.00018303316028323025, 'train/grad_norm_pre_clip': 0.02057323418557644, 'timing/data': 0.0003882400051224977, 'timing/model': 0.36454126999888103, 'learning_rate/qwen_vl_interface': 8.4344009030819e-06, 'learning_rate/action_model': 8.4344009030819e-05, 'epoch': 0.1}) +[2026-10-03 14:53:50,915][starVLA.training.train_starvla][INFO] - Step 1405, Loss: {'train/loss': 0.388828843832016, 'train/grad_norm_pre_clip': 57.43616485595703, 'timing/data': 0.00030393000633921474, 'timing/model': 0.3673245690006297, 'learning_rate/qwen_vl_interface': 8.432140546878899e-06, 'learning_rate/action_model': 8.432140546878898e-05, 'epoch': 0.1}) +[2026-10-03 14:53:51,757][starVLA.training.train_starvla][INFO] - Step 1406, Loss: {'train/loss': 9.280353697249666e-05, 'train/grad_norm_pre_clip': 0.010864700190722942, 'timing/data': 0.00020071900507900864, 'timing/model': 0.5579575029987609, 'learning_rate/qwen_vl_interface': 8.429878882618116e-06, 'learning_rate/action_model': 8.429878882618116e-05, 'epoch': 0.1}) +[2026-10-03 14:53:52,404][starVLA.training.train_starvla][INFO] - Step 1407, Loss: {'train/loss': 0.12229112535715103, 'train/grad_norm_pre_clip': 53.665191650390625, 'timing/data': 0.0002740500058280304, 'timing/model': 0.36521523899864405, 'learning_rate/qwen_vl_interface': 8.42761591122924e-06, 'learning_rate/action_model': 8.427615911229241e-05, 'epoch': 0.1}) +[2026-10-03 14:53:53,052][starVLA.training.train_starvla][INFO] - Step 1408, Loss: {'train/loss': 0.0002499638358131051, 'train/grad_norm_pre_clip': 0.06816983222961426, 'timing/data': 0.0003341699921293184, 'timing/model': 0.36484275000111666, 'learning_rate/qwen_vl_interface': 8.425351633642491e-06, 'learning_rate/action_model': 8.425351633642492e-05, 'epoch': 0.1}) +[2026-10-03 14:53:53,703][starVLA.training.train_starvla][INFO] - Step 1409, Loss: {'train/loss': 0.30711719393730164, 'train/grad_norm_pre_clip': 103.00983428955078, 'timing/data': 0.00023768999380990863, 'timing/model': 0.36675199899764266, 'learning_rate/qwen_vl_interface': 8.423086050788632e-06, 'learning_rate/action_model': 8.423086050788631e-05, 'epoch': 0.1}) +[2026-10-03 14:53:54,350][starVLA.training.train_starvla][INFO] - Step 1410, Loss: {'train/loss': 0.00020195388060528785, 'train/grad_norm_pre_clip': 0.1279161125421524, 'timing/data': 0.0002767999976640567, 'timing/model': 0.36457035000785254, 'learning_rate/qwen_vl_interface': 8.420819163598953e-06, 'learning_rate/action_model': 8.420819163598952e-05, 'epoch': 0.1}) +[2026-10-03 14:53:55,197][starVLA.training.train_starvla][INFO] - Step 1411, Loss: {'train/loss': 0.46333640813827515, 'train/grad_norm_pre_clip': 33.279052734375, 'timing/data': 0.0002051699993899092, 'timing/model': 0.5641124719986692, 'learning_rate/qwen_vl_interface': 8.418550973005289e-06, 'learning_rate/action_model': 8.418550973005289e-05, 'epoch': 0.1}) +[2026-10-03 14:53:55,846][starVLA.training.train_starvla][INFO] - Step 1412, Loss: {'train/loss': 0.011154389008879662, 'train/grad_norm_pre_clip': 21.73942756652832, 'timing/data': 0.00031210000452119857, 'timing/model': 0.36517215000640135, 'learning_rate/qwen_vl_interface': 8.416281479940005e-06, 'learning_rate/action_model': 8.416281479940005e-05, 'epoch': 0.1}) +[2026-10-03 14:53:56,497][starVLA.training.train_starvla][INFO] - Step 1413, Loss: {'train/loss': 5.269416578812525e-05, 'train/grad_norm_pre_clip': 0.007073680404573679, 'timing/data': 0.00026294999406673014, 'timing/model': 0.36724733900337014, 'learning_rate/qwen_vl_interface': 8.414010685336008e-06, 'learning_rate/action_model': 8.414010685336008e-05, 'epoch': 0.1}) +[2026-10-03 14:53:57,145][starVLA.training.train_starvla][INFO] - Step 1414, Loss: {'train/loss': 8.068306487984955e-05, 'train/grad_norm_pre_clip': 0.008864183910191059, 'timing/data': 0.0002625899942358956, 'timing/model': 0.36496833999990486, 'learning_rate/qwen_vl_interface': 8.411738590126733e-06, 'learning_rate/action_model': 8.411738590126732e-05, 'epoch': 0.1}) +[2026-10-03 14:53:57,794][starVLA.training.train_starvla][INFO] - Step 1415, Loss: {'train/loss': 7.531732262577862e-05, 'train/grad_norm_pre_clip': 0.008273632265627384, 'timing/data': 0.0002982599980896339, 'timing/model': 0.3657367089908803, 'learning_rate/qwen_vl_interface': 8.409465195246152e-06, 'learning_rate/action_model': 8.409465195246152e-05, 'epoch': 0.1}) +[2026-10-03 14:53:58,442][starVLA.training.train_starvla][INFO] - Step 1416, Loss: {'train/loss': 0.0030618628952652216, 'train/grad_norm_pre_clip': 1.6830946207046509, 'timing/data': 0.00025968000409193337, 'timing/model': 0.36547593999421224, 'learning_rate/qwen_vl_interface': 8.407190501628772e-06, 'learning_rate/action_model': 8.407190501628772e-05, 'epoch': 0.1}) +[2026-10-03 14:53:59,296][starVLA.training.train_starvla][INFO] - Step 1417, Loss: {'train/loss': 0.0020159794948995113, 'train/grad_norm_pre_clip': 1.0045405626296997, 'timing/data': 0.00025419901066925377, 'timing/model': 0.367494699996314, 'learning_rate/qwen_vl_interface': 8.404914510209635e-06, 'learning_rate/action_model': 8.404914510209635e-05, 'epoch': 0.1}) +[2026-10-03 14:53:59,943][starVLA.training.train_starvla][INFO] - Step 1418, Loss: {'train/loss': 0.0016359547153115273, 'train/grad_norm_pre_clip': 1.0232493877410889, 'timing/data': 0.00033290000283159316, 'timing/model': 0.3642719199997373, 'learning_rate/qwen_vl_interface': 8.402637221924315e-06, 'learning_rate/action_model': 8.402637221924314e-05, 'epoch': 0.1}) +[2026-10-03 14:54:00,592][starVLA.training.train_starvla][INFO] - Step 1419, Loss: {'train/loss': 0.3955455422401428, 'train/grad_norm_pre_clip': 51.72596740722656, 'timing/data': 0.00028684000426437706, 'timing/model': 0.3651607110077748, 'learning_rate/qwen_vl_interface': 8.400358637708919e-06, 'learning_rate/action_model': 8.400358637708918e-05, 'epoch': 0.1}) +[2026-10-03 14:54:01,241][starVLA.training.train_starvla][INFO] - Step 1420, Loss: {'train/loss': 6.575229053851217e-05, 'train/grad_norm_pre_clip': 0.007914070039987564, 'timing/data': 0.00023417000193148851, 'timing/model': 0.36587531901022885, 'learning_rate/qwen_vl_interface': 8.398078758500086e-06, 'learning_rate/action_model': 8.398078758500086e-05, 'epoch': 0.1}) +[2026-10-03 14:54:01,895][starVLA.training.train_starvla][INFO] - Step 1421, Loss: {'train/loss': 0.5925776362419128, 'train/grad_norm_pre_clip': 15.173718452453613, 'timing/data': 0.00025996999465860426, 'timing/model': 0.36881610899581574, 'learning_rate/qwen_vl_interface': 8.395797585234987e-06, 'learning_rate/action_model': 8.395797585234988e-05, 'epoch': 0.1}) +[2026-10-03 14:54:02,749][starVLA.training.train_starvla][INFO] - Step 1422, Loss: {'train/loss': 3.529978130245581e-05, 'train/grad_norm_pre_clip': 0.003747532842680812, 'timing/data': 0.0002732599969021976, 'timing/model': 0.3665206799923908, 'learning_rate/qwen_vl_interface': 8.39351511885133e-06, 'learning_rate/action_model': 8.393515118851329e-05, 'epoch': 0.1}) +[2026-10-03 14:54:03,401][starVLA.training.train_starvla][INFO] - Step 1423, Loss: {'train/loss': 0.0006650816649198532, 'train/grad_norm_pre_clip': 0.7042314410209656, 'timing/data': 0.00032585998997092247, 'timing/model': 0.36658726999303326, 'learning_rate/qwen_vl_interface': 8.391231360287348e-06, 'learning_rate/action_model': 8.391231360287347e-05, 'epoch': 0.1}) +[2026-10-03 14:54:04,053][starVLA.training.train_starvla][INFO] - Step 1424, Loss: {'train/loss': 0.0733390748500824, 'train/grad_norm_pre_clip': 153.22247314453125, 'timing/data': 0.0003033800021512434, 'timing/model': 0.3663787100085756, 'learning_rate/qwen_vl_interface': 8.388946310481807e-06, 'learning_rate/action_model': 8.388946310481807e-05, 'epoch': 0.1}) +[2026-10-03 14:54:04,707][starVLA.training.train_starvla][INFO] - Step 1425, Loss: {'train/loss': 7.149760494939983e-05, 'train/grad_norm_pre_clip': 0.008731469511985779, 'timing/data': 0.0002962400030810386, 'timing/model': 0.3687168489996111, 'learning_rate/qwen_vl_interface': 8.386659970374007e-06, 'learning_rate/action_model': 8.386659970374006e-05, 'epoch': 0.1}) +[2026-10-03 14:54:05,358][starVLA.training.train_starvla][INFO] - Step 1426, Loss: {'train/loss': 0.5799207091331482, 'train/grad_norm_pre_clip': 23.310110092163086, 'timing/data': 0.00024013999791350216, 'timing/model': 0.36630187000264414, 'learning_rate/qwen_vl_interface': 8.384372340903775e-06, 'learning_rate/action_model': 8.384372340903775e-05, 'epoch': 0.1}) +[2026-10-03 14:54:06,213][starVLA.training.train_starvla][INFO] - Step 1427, Loss: {'train/loss': 0.34747838973999023, 'train/grad_norm_pre_clip': 58.40425109863281, 'timing/data': 0.00026818999322131276, 'timing/model': 0.568900112004485, 'learning_rate/qwen_vl_interface': 8.382083423011472e-06, 'learning_rate/action_model': 8.382083423011471e-05, 'epoch': 0.1}) +[2026-10-03 14:54:06,819][starVLA.training.train_starvla][INFO] - Step 1428, Loss: {'train/loss': 0.000644062296487391, 'train/grad_norm_pre_clip': 0.23481260240077972, 'timing/data': 0.0002501800045138225, 'timing/model': 0.3436529240134405, 'learning_rate/qwen_vl_interface': 8.379793217637979e-06, 'learning_rate/action_model': 8.379793217637979e-05, 'epoch': 0.1}) +[2026-10-03 14:54:07,473][starVLA.training.train_starvla][INFO] - Step 1429, Loss: {'train/loss': 0.0011742800706997514, 'train/grad_norm_pre_clip': 1.4702285528182983, 'timing/data': 0.0002442999975755811, 'timing/model': 0.3691423290001694, 'learning_rate/qwen_vl_interface': 8.377501725724719e-06, 'learning_rate/action_model': 8.377501725724719e-05, 'epoch': 0.1}) +[2026-10-03 14:54:08,124][starVLA.training.train_starvla][INFO] - Step 1430, Loss: {'train/loss': 0.711294949054718, 'train/grad_norm_pre_clip': 307.6275329589844, 'timing/data': 0.00020015001064166427, 'timing/model': 0.36674381999182515, 'learning_rate/qwen_vl_interface': 8.375208948213636e-06, 'learning_rate/action_model': 8.375208948213635e-05, 'epoch': 0.1}) +[2026-10-03 14:54:08,775][starVLA.training.train_starvla][INFO] - Step 1431, Loss: {'train/loss': 0.6552658677101135, 'train/grad_norm_pre_clip': 61.78565216064453, 'timing/data': 0.00020106000010855496, 'timing/model': 0.36697970000386704, 'learning_rate/qwen_vl_interface': 8.372914886047203e-06, 'learning_rate/action_model': 8.372914886047203e-05, 'epoch': 0.1}) +[2026-10-03 14:54:09,426][starVLA.training.train_starvla][INFO] - Step 1432, Loss: {'train/loss': 0.3067846894264221, 'train/grad_norm_pre_clip': 119.47130584716797, 'timing/data': 0.00029814000299666077, 'timing/model': 0.36602033000963274, 'learning_rate/qwen_vl_interface': 8.370619540168422e-06, 'learning_rate/action_model': 8.370619540168423e-05, 'epoch': 0.1}) +[2026-10-03 14:54:10,274][starVLA.training.train_starvla][INFO] - Step 1433, Loss: {'train/loss': 0.002991176676005125, 'train/grad_norm_pre_clip': 2.568448781967163, 'timing/data': 0.00024002000282052904, 'timing/model': 0.3690377499879105, 'learning_rate/qwen_vl_interface': 8.368322911520826e-06, 'learning_rate/action_model': 8.368322911520826e-05, 'epoch': 0.1}) +[2026-10-03 14:54:10,924][starVLA.training.train_starvla][INFO] - Step 1434, Loss: {'train/loss': 0.007219741120934486, 'train/grad_norm_pre_clip': 13.159186363220215, 'timing/data': 0.000258979998761788, 'timing/model': 0.3661254399921745, 'learning_rate/qwen_vl_interface': 8.36602500104847e-06, 'learning_rate/action_model': 8.36602500104847e-05, 'epoch': 0.1}) +[2026-10-03 14:54:11,553][starVLA.training.train_starvla][INFO] - Step 1435, Loss: {'train/loss': 0.3386573791503906, 'train/grad_norm_pre_clip': 39.85346221923828, 'timing/data': 0.0003028799983439967, 'timing/model': 0.36638822000531945, 'learning_rate/qwen_vl_interface': 8.36372580969594e-06, 'learning_rate/action_model': 8.36372580969594e-05, 'epoch': 0.11}) +[2026-10-03 14:54:12,203][starVLA.training.train_starvla][INFO] - Step 1436, Loss: {'train/loss': 0.001746860216371715, 'train/grad_norm_pre_clip': 1.1145403385162354, 'timing/data': 0.00024897999537643045, 'timing/model': 0.3660195699922042, 'learning_rate/qwen_vl_interface': 8.361425338408344e-06, 'learning_rate/action_model': 8.361425338408343e-05, 'epoch': 0.11}) +[2026-10-03 14:54:12,856][starVLA.training.train_starvla][INFO] - Step 1437, Loss: {'train/loss': 0.3196767568588257, 'train/grad_norm_pre_clip': 15.627431869506836, 'timing/data': 0.0002502200077287853, 'timing/model': 0.3685801800020272, 'learning_rate/qwen_vl_interface': 8.359123588131319e-06, 'learning_rate/action_model': 8.359123588131319e-05, 'epoch': 0.11}) +[2026-10-03 14:54:13,685][starVLA.training.train_starvla][INFO] - Step 1438, Loss: {'train/loss': 0.04161832481622696, 'train/grad_norm_pre_clip': 26.533479690551758, 'timing/data': 0.0002035999932559207, 'timing/model': 0.5662167620030232, 'learning_rate/qwen_vl_interface': 8.35682055981103e-06, 'learning_rate/action_model': 8.35682055981103e-05, 'epoch': 0.11}) +[2026-10-03 14:54:14,335][starVLA.training.train_starvla][INFO] - Step 1439, Loss: {'train/loss': 0.005691304802894592, 'train/grad_norm_pre_clip': 13.200074195861816, 'timing/data': 0.0003478599974187091, 'timing/model': 0.3661219500063453, 'learning_rate/qwen_vl_interface': 8.354516254394163e-06, 'learning_rate/action_model': 8.354516254394162e-05, 'epoch': 0.11}) +[2026-10-03 14:54:14,984][starVLA.training.train_starvla][INFO] - Step 1440, Loss: {'train/loss': 0.0004801823233719915, 'train/grad_norm_pre_clip': 0.06639532744884491, 'timing/data': 0.00032684000325389206, 'timing/model': 0.36516484098683577, 'learning_rate/qwen_vl_interface': 8.352210672827932e-06, 'learning_rate/action_model': 8.352210672827931e-05, 'epoch': 0.11}) +[2026-10-03 14:54:15,636][starVLA.training.train_starvla][INFO] - Step 1441, Loss: {'train/loss': 0.0013318858109414577, 'train/grad_norm_pre_clip': 0.22132807970046997, 'timing/data': 0.00029320000612642616, 'timing/model': 0.36801893000665586, 'learning_rate/qwen_vl_interface': 8.349903816060074e-06, 'learning_rate/action_model': 8.349903816060075e-05, 'epoch': 0.11}) +[2026-10-03 14:54:16,286][starVLA.training.train_starvla][INFO] - Step 1442, Loss: {'train/loss': 0.0020631328225135803, 'train/grad_norm_pre_clip': 1.069520115852356, 'timing/data': 0.00031663000117987394, 'timing/model': 0.365959691000171, 'learning_rate/qwen_vl_interface': 8.34759568503885e-06, 'learning_rate/action_model': 8.347595685038851e-05, 'epoch': 0.11}) +[2026-10-03 14:54:17,135][starVLA.training.train_starvla][INFO] - Step 1443, Loss: {'train/loss': 0.1458498239517212, 'train/grad_norm_pre_clip': 94.58924102783203, 'timing/data': 0.00031103000219445676, 'timing/model': 0.5645193130039843, 'learning_rate/qwen_vl_interface': 8.345286280713049e-06, 'learning_rate/action_model': 8.345286280713048e-05, 'epoch': 0.11}) +[2026-10-03 14:54:17,784][starVLA.training.train_starvla][INFO] - Step 1444, Loss: {'train/loss': 0.0037161805666983128, 'train/grad_norm_pre_clip': 2.0659091472625732, 'timing/data': 0.00024275000032503158, 'timing/model': 0.36579323999467306, 'learning_rate/qwen_vl_interface': 8.342975604031974e-06, 'learning_rate/action_model': 8.342975604031974e-05, 'epoch': 0.11}) +[2026-10-03 14:54:18,412][starVLA.training.train_starvla][INFO] - Step 1445, Loss: {'train/loss': 0.004788434598594904, 'train/grad_norm_pre_clip': 3.217985153198242, 'timing/data': 0.00034793000668287277, 'timing/model': 0.34433464499306865, 'learning_rate/qwen_vl_interface': 8.340663655945462e-06, 'learning_rate/action_model': 8.340663655945462e-05, 'epoch': 0.11}) +[2026-10-03 14:54:19,069][starVLA.training.train_starvla][INFO] - Step 1446, Loss: {'train/loss': 0.0006909518851898611, 'train/grad_norm_pre_clip': 0.34711483120918274, 'timing/data': 0.00034654000774025917, 'timing/model': 0.3693608300091, 'learning_rate/qwen_vl_interface': 8.338350437403866e-06, 'learning_rate/action_model': 8.338350437403865e-05, 'epoch': 0.11}) +[2026-10-03 14:54:19,726][starVLA.training.train_starvla][INFO] - Step 1447, Loss: {'train/loss': 0.7648534774780273, 'train/grad_norm_pre_clip': 99.36593627929688, 'timing/data': 0.00021321000531315804, 'timing/model': 0.36883506999583915, 'learning_rate/qwen_vl_interface': 8.336035949358063e-06, 'learning_rate/action_model': 8.336035949358062e-05, 'epoch': 0.11}) +[2026-10-03 14:54:20,578][starVLA.training.train_starvla][INFO] - Step 1448, Loss: {'train/loss': 0.044773198664188385, 'train/grad_norm_pre_clip': 42.98476028442383, 'timing/data': 0.0003404599992791191, 'timing/model': 0.564098433998879, 'learning_rate/qwen_vl_interface': 8.333720192759448e-06, 'learning_rate/action_model': 8.333720192759448e-05, 'epoch': 0.11}) +[2026-10-03 14:54:21,236][starVLA.training.train_starvla][INFO] - Step 1449, Loss: {'train/loss': 0.0026465982664376497, 'train/grad_norm_pre_clip': 1.9857125282287598, 'timing/data': 0.00031690999458078295, 'timing/model': 0.37007715000072494, 'learning_rate/qwen_vl_interface': 8.331403168559946e-06, 'learning_rate/action_model': 8.331403168559945e-05, 'epoch': 0.11}) +[2026-10-03 14:54:21,892][starVLA.training.train_starvla][INFO] - Step 1450, Loss: {'train/loss': 0.007265686057507992, 'train/grad_norm_pre_clip': 15.324518203735352, 'timing/data': 0.0003554299910319969, 'timing/model': 0.3684671810042346, 'learning_rate/qwen_vl_interface': 8.329084877711995e-06, 'learning_rate/action_model': 8.329084877711996e-05, 'epoch': 0.11}) +[2026-10-03 14:54:22,549][starVLA.training.train_starvla][INFO] - Step 1451, Loss: {'train/loss': 0.297932893037796, 'train/grad_norm_pre_clip': 22.944154739379883, 'timing/data': 0.0003341900010127574, 'timing/model': 0.36889557899849024, 'learning_rate/qwen_vl_interface': 8.32676532116856e-06, 'learning_rate/action_model': 8.326765321168559e-05, 'epoch': 0.11}) +[2026-10-03 14:54:23,205][starVLA.training.train_starvla][INFO] - Step 1452, Loss: {'train/loss': 0.0021025061141699553, 'train/grad_norm_pre_clip': 6.0537028312683105, 'timing/data': 0.00032590999035164714, 'timing/model': 0.3691769909928553, 'learning_rate/qwen_vl_interface': 8.324444499883119e-06, 'learning_rate/action_model': 8.324444499883119e-05, 'epoch': 0.11}) +[2026-10-03 14:54:23,864][starVLA.training.train_starvla][INFO] - Step 1453, Loss: {'train/loss': 0.04238036647439003, 'train/grad_norm_pre_clip': 51.97482681274414, 'timing/data': 0.0003160899941576645, 'timing/model': 0.3707057099964004, 'learning_rate/qwen_vl_interface': 8.322122414809678e-06, 'learning_rate/action_model': 8.322122414809677e-05, 'epoch': 0.11}) +[2026-10-03 14:54:24,717][starVLA.training.train_starvla][INFO] - Step 1454, Loss: {'train/loss': 0.005451191682368517, 'train/grad_norm_pre_clip': 6.060514450073242, 'timing/data': 0.000348770001437515, 'timing/model': 0.36907031999726314, 'learning_rate/qwen_vl_interface': 8.319799066902755e-06, 'learning_rate/action_model': 8.319799066902755e-05, 'epoch': 0.11}) +[2026-10-03 14:54:25,374][starVLA.training.train_starvla][INFO] - Step 1455, Loss: {'train/loss': 0.000491200597025454, 'train/grad_norm_pre_clip': 0.08148711174726486, 'timing/data': 0.00020348999532870948, 'timing/model': 0.3692755600059172, 'learning_rate/qwen_vl_interface': 8.317474457117393e-06, 'learning_rate/action_model': 8.317474457117392e-05, 'epoch': 0.11}) +[2026-10-03 14:54:26,031][starVLA.training.train_starvla][INFO] - Step 1456, Loss: {'train/loss': 0.40605220198631287, 'train/grad_norm_pre_clip': 19.25811195373535, 'timing/data': 0.0003051000094274059, 'timing/model': 0.3693097600043984, 'learning_rate/qwen_vl_interface': 8.315148586409148e-06, 'learning_rate/action_model': 8.315148586409148e-05, 'epoch': 0.11}) +[2026-10-03 14:54:26,689][starVLA.training.train_starvla][INFO] - Step 1457, Loss: {'train/loss': 0.0007655879016965628, 'train/grad_norm_pre_clip': 0.20319703221321106, 'timing/data': 0.000268019997747615, 'timing/model': 0.37028224000823684, 'learning_rate/qwen_vl_interface': 8.312821455734104e-06, 'learning_rate/action_model': 8.312821455734103e-05, 'epoch': 0.11}) +[2026-10-03 14:54:27,346][starVLA.training.train_starvla][INFO] - Step 1458, Loss: {'train/loss': 0.5017954707145691, 'train/grad_norm_pre_clip': 56.36510467529297, 'timing/data': 0.0002610200026538223, 'timing/model': 0.36890833999495953, 'learning_rate/qwen_vl_interface': 8.31049306604885e-06, 'learning_rate/action_model': 8.310493066048849e-05, 'epoch': 0.11}) +[2026-10-03 14:54:28,204][starVLA.training.train_starvla][INFO] - Step 1459, Loss: {'train/loss': 0.06014164164662361, 'train/grad_norm_pre_clip': 67.64253997802734, 'timing/data': 0.0003819800040218979, 'timing/model': 0.3695814709935803, 'learning_rate/qwen_vl_interface': 8.3081634183105e-06, 'learning_rate/action_model': 8.308163418310501e-05, 'epoch': 0.11}) +[2026-10-03 14:54:28,861][starVLA.training.train_starvla][INFO] - Step 1460, Loss: {'train/loss': 0.9345629215240479, 'train/grad_norm_pre_clip': 145.406494140625, 'timing/data': 0.000261370005318895, 'timing/model': 0.3695561409986112, 'learning_rate/qwen_vl_interface': 8.305832513476689e-06, 'learning_rate/action_model': 8.305832513476688e-05, 'epoch': 0.11}) +[2026-10-03 14:54:29,519][starVLA.training.train_starvla][INFO] - Step 1461, Loss: {'train/loss': 0.031787171959877014, 'train/grad_norm_pre_clip': 92.27787780761719, 'timing/data': 0.00026821000210475177, 'timing/model': 0.3703598299907753, 'learning_rate/qwen_vl_interface': 8.303500352505557e-06, 'learning_rate/action_model': 8.303500352505557e-05, 'epoch': 0.11}) +[2026-10-03 14:54:30,176][starVLA.training.train_starvla][INFO] - Step 1462, Loss: {'train/loss': 0.0010824783239513636, 'train/grad_norm_pre_clip': 0.205653578042984, 'timing/data': 0.00047821000043768436, 'timing/model': 0.3689972900028806, 'learning_rate/qwen_vl_interface': 8.301166936355773e-06, 'learning_rate/action_model': 8.301166936355773e-05, 'epoch': 0.11}) +[2026-10-03 14:54:30,835][starVLA.training.train_starvla][INFO] - Step 1463, Loss: {'train/loss': 0.5386735200881958, 'train/grad_norm_pre_clip': 85.17041778564453, 'timing/data': 0.00030249998962972313, 'timing/model': 0.36988565100182313, 'learning_rate/qwen_vl_interface': 8.298832265986514e-06, 'learning_rate/action_model': 8.298832265986513e-05, 'epoch': 0.11}) +[2026-10-03 14:54:31,684][starVLA.training.train_starvla][INFO] - Step 1464, Loss: {'train/loss': 0.3622298240661621, 'train/grad_norm_pre_clip': 78.4907455444336, 'timing/data': 0.0002867199946194887, 'timing/model': 0.5617064949910855, 'learning_rate/qwen_vl_interface': 8.296496342357475e-06, 'learning_rate/action_model': 8.296496342357474e-05, 'epoch': 0.11}) +[2026-10-03 14:54:32,320][starVLA.training.train_starvla][INFO] - Step 1465, Loss: {'train/loss': 0.009692694991827011, 'train/grad_norm_pre_clip': 6.367305755615234, 'timing/data': 0.00028745000599883497, 'timing/model': 0.37039050000021234, 'learning_rate/qwen_vl_interface': 8.294159166428864e-06, 'learning_rate/action_model': 8.294159166428864e-05, 'epoch': 0.11}) +[2026-10-03 14:54:32,955][starVLA.training.train_starvla][INFO] - Step 1466, Loss: {'train/loss': 0.0027065128087997437, 'train/grad_norm_pre_clip': 0.862558901309967, 'timing/data': 0.00035427999682724476, 'timing/model': 0.34689720399910584, 'learning_rate/qwen_vl_interface': 8.29182073916141e-06, 'learning_rate/action_model': 8.29182073916141e-05, 'epoch': 0.11}) +[2026-10-03 14:54:33,613][starVLA.training.train_starvla][INFO] - Step 1467, Loss: {'train/loss': 0.0009434617240913212, 'train/grad_norm_pre_clip': 0.18559864163398743, 'timing/data': 0.0002543800073908642, 'timing/model': 0.369077520997962, 'learning_rate/qwen_vl_interface': 8.28948106151635e-06, 'learning_rate/action_model': 8.28948106151635e-05, 'epoch': 0.11}) +[2026-10-03 14:54:34,270][starVLA.training.train_starvla][INFO] - Step 1468, Loss: {'train/loss': 0.3654259741306305, 'train/grad_norm_pre_clip': 17.785600662231445, 'timing/data': 0.00026435000472702086, 'timing/model': 0.3692924299975857, 'learning_rate/qwen_vl_interface': 8.287140134455435e-06, 'learning_rate/action_model': 8.287140134455435e-05, 'epoch': 0.11}) +[2026-10-03 14:54:34,928][starVLA.training.train_starvla][INFO] - Step 1469, Loss: {'train/loss': 0.0009135232539847493, 'train/grad_norm_pre_clip': 0.1584567278623581, 'timing/data': 0.0003163599903928116, 'timing/model': 0.3704996110027423, 'learning_rate/qwen_vl_interface': 8.284797958940934e-06, 'learning_rate/action_model': 8.284797958940935e-05, 'epoch': 0.11}) +[2026-10-03 14:54:35,780][starVLA.training.train_starvla][INFO] - Step 1470, Loss: {'train/loss': 0.028178798034787178, 'train/grad_norm_pre_clip': 36.58536148071289, 'timing/data': 0.00033213000278919935, 'timing/model': 0.37007207999704406, 'learning_rate/qwen_vl_interface': 8.282454535935626e-06, 'learning_rate/action_model': 8.282454535935626e-05, 'epoch': 0.11}) +[2026-10-03 14:54:36,438][starVLA.training.train_starvla][INFO] - Step 1471, Loss: {'train/loss': 0.016567770391702652, 'train/grad_norm_pre_clip': 17.60289192199707, 'timing/data': 0.0003327000013086945, 'timing/model': 0.3696732609969331, 'learning_rate/qwen_vl_interface': 8.280109866402805e-06, 'learning_rate/action_model': 8.280109866402806e-05, 'epoch': 0.11}) +[2026-10-03 14:54:37,095][starVLA.training.train_starvla][INFO] - Step 1472, Loss: {'train/loss': 0.9362367391586304, 'train/grad_norm_pre_clip': 73.04218292236328, 'timing/data': 0.0002179099974455312, 'timing/model': 0.3698067809891654, 'learning_rate/qwen_vl_interface': 8.277763951306277e-06, 'learning_rate/action_model': 8.277763951306277e-05, 'epoch': 0.11}) +[2026-10-03 14:54:37,756][starVLA.training.train_starvla][INFO] - Step 1473, Loss: {'train/loss': 0.0020755724981427193, 'train/grad_norm_pre_clip': 0.5827773809432983, 'timing/data': 0.00031257000227924436, 'timing/model': 0.3709088710020296, 'learning_rate/qwen_vl_interface': 8.275416791610356e-06, 'learning_rate/action_model': 8.275416791610355e-05, 'epoch': 0.11}) +[2026-10-03 14:54:38,413][starVLA.training.train_starvla][INFO] - Step 1474, Loss: {'train/loss': 0.31123408675193787, 'train/grad_norm_pre_clip': 220.20016479492188, 'timing/data': 0.00029106000147294253, 'timing/model': 0.36896776099456474, 'learning_rate/qwen_vl_interface': 8.27306838827987e-06, 'learning_rate/action_model': 8.273068388279871e-05, 'epoch': 0.11}) +[2026-10-03 14:54:39,266][starVLA.training.train_starvla][INFO] - Step 1475, Loss: {'train/loss': 0.0032724952325224876, 'train/grad_norm_pre_clip': 1.1632338762283325, 'timing/data': 0.0003332699998281896, 'timing/model': 0.3691976510017412, 'learning_rate/qwen_vl_interface': 8.270718742280162e-06, 'learning_rate/action_model': 8.270718742280162e-05, 'epoch': 0.11}) +[2026-10-03 14:54:39,923][starVLA.training.train_starvla][INFO] - Step 1476, Loss: {'train/loss': 0.0004918782506138086, 'train/grad_norm_pre_clip': 0.07424508780241013, 'timing/data': 0.0002665799984242767, 'timing/model': 0.36976097100705374, 'learning_rate/qwen_vl_interface': 8.268367854577084e-06, 'learning_rate/action_model': 8.268367854577082e-05, 'epoch': 0.11}) +[2026-10-03 14:54:40,582][starVLA.training.train_starvla][INFO] - Step 1477, Loss: {'train/loss': 1.1351357698440552, 'train/grad_norm_pre_clip': 72.42356872558594, 'timing/data': 0.0002639199956320226, 'timing/model': 0.3713438109989511, 'learning_rate/qwen_vl_interface': 8.266015726136991e-06, 'learning_rate/action_model': 8.266015726136991e-05, 'epoch': 0.11}) +[2026-10-03 14:54:41,217][starVLA.training.train_starvla][INFO] - Step 1478, Loss: {'train/loss': 0.18622837960720062, 'train/grad_norm_pre_clip': 135.2342987060547, 'timing/data': 0.00022495001030620188, 'timing/model': 0.3695703300036257, 'learning_rate/qwen_vl_interface': 8.26366235792676e-06, 'learning_rate/action_model': 8.26366235792676e-05, 'epoch': 0.11}) +[2026-10-03 14:54:41,875][starVLA.training.train_starvla][INFO] - Step 1479, Loss: {'train/loss': 0.00030656164744868875, 'train/grad_norm_pre_clip': 0.13237465918064117, 'timing/data': 0.00027011000202037394, 'timing/model': 0.3692507300002035, 'learning_rate/qwen_vl_interface': 8.261307750913769e-06, 'learning_rate/action_model': 8.261307750913769e-05, 'epoch': 0.11}) +[2026-10-03 14:54:42,725][starVLA.training.train_starvla][INFO] - Step 1480, Loss: {'train/loss': 0.2353857308626175, 'train/grad_norm_pre_clip': 182.71319580078125, 'timing/data': 0.00021727000421378762, 'timing/model': 0.5630648039950756, 'learning_rate/qwen_vl_interface': 8.258951906065908e-06, 'learning_rate/action_model': 8.258951906065908e-05, 'epoch': 0.11}) +[2026-10-03 14:54:43,386][starVLA.training.train_starvla][INFO] - Step 1481, Loss: {'train/loss': 0.1921156495809555, 'train/grad_norm_pre_clip': 218.82508850097656, 'timing/data': 0.00026996000087819993, 'timing/model': 0.37201172999630217, 'learning_rate/qwen_vl_interface': 8.256594824351577e-06, 'learning_rate/action_model': 8.256594824351576e-05, 'epoch': 0.11}) +[2026-10-03 14:54:44,043][starVLA.training.train_starvla][INFO] - Step 1482, Loss: {'train/loss': 0.0010603625560179353, 'train/grad_norm_pre_clip': 2.4869565963745117, 'timing/data': 0.000272019999101758, 'timing/model': 0.3692473210103344, 'learning_rate/qwen_vl_interface': 8.254236506739683e-06, 'learning_rate/action_model': 8.254236506739681e-05, 'epoch': 0.11}) +[2026-10-03 14:54:44,700][starVLA.training.train_starvla][INFO] - Step 1483, Loss: {'train/loss': 0.00019986910047009587, 'train/grad_norm_pre_clip': 0.02931523509323597, 'timing/data': 0.00032577000092715025, 'timing/model': 0.36923086100432556, 'learning_rate/qwen_vl_interface': 8.25187695419964e-06, 'learning_rate/action_model': 8.25187695419964e-05, 'epoch': 0.11}) +[2026-10-03 14:54:45,355][starVLA.training.train_starvla][INFO] - Step 1484, Loss: {'train/loss': 0.36359068751335144, 'train/grad_norm_pre_clip': 93.10005187988281, 'timing/data': 0.00021116000425536186, 'timing/model': 0.36873959099466447, 'learning_rate/qwen_vl_interface': 8.249516167701373e-06, 'learning_rate/action_model': 8.249516167701374e-05, 'epoch': 0.11}) +[2026-10-03 14:54:46,208][starVLA.training.train_starvla][INFO] - Step 1485, Loss: {'train/loss': 0.23857353627681732, 'train/grad_norm_pre_clip': 159.68063354492188, 'timing/data': 0.0002623000036692247, 'timing/model': 0.5659990650019608, 'learning_rate/qwen_vl_interface': 8.247154148215311e-06, 'learning_rate/action_model': 8.24715414821531e-05, 'epoch': 0.11}) +[2026-10-03 14:54:46,863][starVLA.training.train_starvla][INFO] - Step 1486, Loss: {'train/loss': 8.345628157258034e-05, 'train/grad_norm_pre_clip': 0.011368119157850742, 'timing/data': 0.000261220004176721, 'timing/model': 0.3681380220077699, 'learning_rate/qwen_vl_interface': 8.24479089671239e-06, 'learning_rate/action_model': 8.24479089671239e-05, 'epoch': 0.11}) +[2026-10-03 14:54:47,519][starVLA.training.train_starvla][INFO] - Step 1487, Loss: {'train/loss': 0.00012789128231815994, 'train/grad_norm_pre_clip': 0.02252533659338951, 'timing/data': 0.00020948000019416213, 'timing/model': 0.3687541120016249, 'learning_rate/qwen_vl_interface': 8.242426414164056e-06, 'learning_rate/action_model': 8.242426414164056e-05, 'epoch': 0.11}) +[2026-10-03 14:54:48,175][starVLA.training.train_starvla][INFO] - Step 1488, Loss: {'train/loss': 0.00015849203919060528, 'train/grad_norm_pre_clip': 0.0703212320804596, 'timing/data': 0.00021490899962373078, 'timing/model': 0.36864065199915785, 'learning_rate/qwen_vl_interface': 8.240060701542257e-06, 'learning_rate/action_model': 8.240060701542257e-05, 'epoch': 0.11}) +[2026-10-03 14:54:48,833][starVLA.training.train_starvla][INFO] - Step 1489, Loss: {'train/loss': 0.011612597852945328, 'train/grad_norm_pre_clip': 15.571968078613281, 'timing/data': 0.00025021999317687005, 'timing/model': 0.37058925100427587, 'learning_rate/qwen_vl_interface': 8.237693759819447e-06, 'learning_rate/action_model': 8.237693759819448e-05, 'epoch': 0.11}) +[2026-10-03 14:54:49,489][starVLA.training.train_starvla][INFO] - Step 1490, Loss: {'train/loss': 0.3407834768295288, 'train/grad_norm_pre_clip': 92.74440002441406, 'timing/data': 0.0002774500026134774, 'timing/model': 0.36901097098598257, 'learning_rate/qwen_vl_interface': 8.235325589968587e-06, 'learning_rate/action_model': 8.235325589968587e-05, 'epoch': 0.11}) +[2026-10-03 14:54:50,336][starVLA.training.train_starvla][INFO] - Step 1491, Loss: {'train/loss': 0.0058279382064938545, 'train/grad_norm_pre_clip': 16.433717727661133, 'timing/data': 0.0002141599979950115, 'timing/model': 0.36926855199271813, 'learning_rate/qwen_vl_interface': 8.232956192963143e-06, 'learning_rate/action_model': 8.232956192963144e-05, 'epoch': 0.11}) +[2026-10-03 14:54:50,992][starVLA.training.train_starvla][INFO] - Step 1492, Loss: {'train/loss': 6.30952199571766e-05, 'train/grad_norm_pre_clip': 0.008353042416274548, 'timing/data': 0.00021831999765709043, 'timing/model': 0.3690928720025113, 'learning_rate/qwen_vl_interface': 8.230585569777086e-06, 'learning_rate/action_model': 8.230585569777085e-05, 'epoch': 0.11}) +[2026-10-03 14:54:51,650][starVLA.training.train_starvla][INFO] - Step 1493, Loss: {'train/loss': 0.00012555475404951721, 'train/grad_norm_pre_clip': 0.02418757602572441, 'timing/data': 0.00024746000417508185, 'timing/model': 0.37046974200347904, 'learning_rate/qwen_vl_interface': 8.228213721384885e-06, 'learning_rate/action_model': 8.228213721384885e-05, 'epoch': 0.11}) +[2026-10-03 14:54:52,283][starVLA.training.train_starvla][INFO] - Step 1494, Loss: {'train/loss': 0.017369501292705536, 'train/grad_norm_pre_clip': 12.752889633178711, 'timing/data': 0.00025522999931126833, 'timing/model': 0.36853672200231813, 'learning_rate/qwen_vl_interface': 8.225840648761523e-06, 'learning_rate/action_model': 8.225840648761523e-05, 'epoch': 0.11}) +[2026-10-03 14:54:52,941][starVLA.training.train_starvla][INFO] - Step 1495, Loss: {'train/loss': 8.973994408734143e-05, 'train/grad_norm_pre_clip': 0.017819955945014954, 'timing/data': 0.0003511499962769449, 'timing/model': 0.36908051099453587, 'learning_rate/qwen_vl_interface': 8.223466352882476e-06, 'learning_rate/action_model': 8.223466352882475e-05, 'epoch': 0.11}) +[2026-10-03 14:54:53,800][starVLA.training.train_starvla][INFO] - Step 1496, Loss: {'train/loss': 0.4081015884876251, 'train/grad_norm_pre_clip': 64.67607879638672, 'timing/data': 0.00026323999918531626, 'timing/model': 0.37034307200519834, 'learning_rate/qwen_vl_interface': 8.22109083472373e-06, 'learning_rate/action_model': 8.22109083472373e-05, 'epoch': 0.11}) +[2026-10-03 14:54:54,459][starVLA.training.train_starvla][INFO] - Step 1497, Loss: {'train/loss': 0.0004912593285553157, 'train/grad_norm_pre_clip': 0.452282577753067, 'timing/data': 0.00029088000883348286, 'timing/model': 0.37033372100268025, 'learning_rate/qwen_vl_interface': 8.218714095261769e-06, 'learning_rate/action_model': 8.218714095261768e-05, 'epoch': 0.11}) +[2026-10-03 14:54:55,115][starVLA.training.train_starvla][INFO] - Step 1498, Loss: {'train/loss': 0.4819485545158386, 'train/grad_norm_pre_clip': 19.97831153869629, 'timing/data': 0.0003596400056267157, 'timing/model': 0.3684747810038971, 'learning_rate/qwen_vl_interface': 8.21633613547358e-06, 'learning_rate/action_model': 8.21633613547358e-05, 'epoch': 0.11}) +[2026-10-03 14:54:55,772][starVLA.training.train_starvla][INFO] - Step 1499, Loss: {'train/loss': 0.00014380141510628164, 'train/grad_norm_pre_clip': 0.06601932644844055, 'timing/data': 0.0003022699966095388, 'timing/model': 0.3689556419994915, 'learning_rate/qwen_vl_interface': 8.213956956336656e-06, 'learning_rate/action_model': 8.213956956336656e-05, 'epoch': 0.11}) +[2026-10-03 14:54:56,406][starVLA.training.train_starvla][INFO] - Step 1500, Loss: {'train/loss': 0.506283700466156, 'train/grad_norm_pre_clip': 33.93620681762695, 'timing/data': 0.00029779999749735, 'timing/model': 0.3689050419925479, 'learning_rate/qwen_vl_interface': 8.211576558828985e-06, 'learning_rate/action_model': 8.211576558828985e-05, 'epoch': 0.11}) +[2026-10-03 14:54:57,263][starVLA.training.train_starvla][INFO] - Step 1501, Loss: {'train/loss': 0.00021394321811385453, 'train/grad_norm_pre_clip': 0.03823019191622734, 'timing/data': 0.00021754999761469662, 'timing/model': 0.5692886249889852, 'learning_rate/qwen_vl_interface': 8.209194943929059e-06, 'learning_rate/action_model': 8.20919494392906e-05, 'epoch': 0.11}) +[2026-10-03 14:54:57,919][starVLA.training.train_starvla][INFO] - Step 1502, Loss: {'train/loss': 0.3084949851036072, 'train/grad_norm_pre_clip': 29.883769989013672, 'timing/data': 0.00026744000206235796, 'timing/model': 0.3689962310018018, 'learning_rate/qwen_vl_interface': 8.20681211261587e-06, 'learning_rate/action_model': 8.20681211261587e-05, 'epoch': 0.11}) +[2026-10-03 14:54:58,577][starVLA.training.train_starvla][INFO] - Step 1503, Loss: {'train/loss': 0.8936064839363098, 'train/grad_norm_pre_clip': 103.69843292236328, 'timing/data': 0.0002477999951224774, 'timing/model': 0.3692887010111008, 'learning_rate/qwen_vl_interface': 8.204428065868911e-06, 'learning_rate/action_model': 8.204428065868911e-05, 'epoch': 0.11}) +[2026-10-03 14:54:59,233][starVLA.training.train_starvla][INFO] - Step 1504, Loss: {'train/loss': 0.4359115958213806, 'train/grad_norm_pre_clip': 69.05458068847656, 'timing/data': 0.00020989000040572137, 'timing/model': 0.36888551199808717, 'learning_rate/qwen_vl_interface': 8.202042804668173e-06, 'learning_rate/action_model': 8.202042804668172e-05, 'epoch': 0.11}) +[2026-10-03 14:54:59,891][starVLA.training.train_starvla][INFO] - Step 1505, Loss: {'train/loss': 0.0015402232529595494, 'train/grad_norm_pre_clip': 0.6703803539276123, 'timing/data': 0.0003318700037198141, 'timing/model': 0.3705812019907171, 'learning_rate/qwen_vl_interface': 8.199656329994145e-06, 'learning_rate/action_model': 8.199656329994145e-05, 'epoch': 0.11}) +[2026-10-03 14:55:00,547][starVLA.training.train_starvla][INFO] - Step 1506, Loss: {'train/loss': 0.024305446073412895, 'train/grad_norm_pre_clip': 12.590161323547363, 'timing/data': 0.0002968199987662956, 'timing/model': 0.36953438199998345, 'learning_rate/qwen_vl_interface': 8.19726864282782e-06, 'learning_rate/action_model': 8.197268642827818e-05, 'epoch': 0.11}) +[2026-10-03 14:55:01,399][starVLA.training.train_starvla][INFO] - Step 1507, Loss: {'train/loss': 0.12433426082134247, 'train/grad_norm_pre_clip': 89.57328033447266, 'timing/data': 0.0003377899993211031, 'timing/model': 0.3699323119944893, 'learning_rate/qwen_vl_interface': 8.194879744150682e-06, 'learning_rate/action_model': 8.194879744150681e-05, 'epoch': 0.11}) +[2026-10-03 14:55:02,055][starVLA.training.train_starvla][INFO] - Step 1508, Loss: {'train/loss': 0.004040582571178675, 'train/grad_norm_pre_clip': 1.1138583421707153, 'timing/data': 0.0003499799931887537, 'timing/model': 0.3685989020013949, 'learning_rate/qwen_vl_interface': 8.192489634944718e-06, 'learning_rate/action_model': 8.192489634944718e-05, 'epoch': 0.11}) +[2026-10-03 14:55:02,713][starVLA.training.train_starvla][INFO] - Step 1509, Loss: {'train/loss': 0.012185582891106606, 'train/grad_norm_pre_clip': 15.258794784545898, 'timing/data': 0.00029476000054273754, 'timing/model': 0.37000998099392746, 'learning_rate/qwen_vl_interface': 8.190098316192416e-06, 'learning_rate/action_model': 8.190098316192416e-05, 'epoch': 0.11}) +[2026-10-03 14:55:03,369][starVLA.training.train_starvla][INFO] - Step 1510, Loss: {'train/loss': 0.0035296715795993805, 'train/grad_norm_pre_clip': 1.4744600057601929, 'timing/data': 0.00036627000372391194, 'timing/model': 0.36822302200016566, 'learning_rate/qwen_vl_interface': 8.18770578887675e-06, 'learning_rate/action_model': 8.18770578887675e-05, 'epoch': 0.11}) +[2026-10-03 14:55:04,027][starVLA.training.train_starvla][INFO] - Step 1511, Loss: {'train/loss': 0.0037313150241971016, 'train/grad_norm_pre_clip': 2.2064144611358643, 'timing/data': 0.0002836299972841516, 'timing/model': 0.3696326119970763, 'learning_rate/qwen_vl_interface': 8.185312053981204e-06, 'learning_rate/action_model': 8.185312053981204e-05, 'epoch': 0.11}) +[2026-10-03 14:55:04,884][starVLA.training.train_starvla][INFO] - Step 1512, Loss: {'train/loss': 0.42680302262306213, 'train/grad_norm_pre_clip': 73.59736633300781, 'timing/data': 0.0003743699926417321, 'timing/model': 0.3695546320086578, 'learning_rate/qwen_vl_interface': 8.182917112489747e-06, 'learning_rate/action_model': 8.182917112489746e-05, 'epoch': 0.11}) +[2026-10-03 14:55:05,543][starVLA.training.train_starvla][INFO] - Step 1513, Loss: {'train/loss': 0.0026217864360660315, 'train/grad_norm_pre_clip': 2.040085792541504, 'timing/data': 0.00029816999449394643, 'timing/model': 0.3712628009961918, 'learning_rate/qwen_vl_interface': 8.18052096538685e-06, 'learning_rate/action_model': 8.18052096538685e-05, 'epoch': 0.11}) +[2026-10-03 14:55:06,199][starVLA.training.train_starvla][INFO] - Step 1514, Loss: {'train/loss': 0.00019939184130635113, 'train/grad_norm_pre_clip': 0.029446342960000038, 'timing/data': 0.00021384000137913972, 'timing/model': 0.36871315199823584, 'learning_rate/qwen_vl_interface': 8.178123613657483e-06, 'learning_rate/action_model': 8.178123613657482e-05, 'epoch': 0.11}) +[2026-10-03 14:55:06,857][starVLA.training.train_starvla][INFO] - Step 1515, Loss: {'train/loss': 0.3110573887825012, 'train/grad_norm_pre_clip': 33.56487274169922, 'timing/data': 0.00025703999563120306, 'timing/model': 0.36923143098829314, 'learning_rate/qwen_vl_interface': 8.175725058287102e-06, 'learning_rate/action_model': 8.175725058287102e-05, 'epoch': 0.11}) +[2026-10-03 14:55:07,513][starVLA.training.train_starvla][INFO] - Step 1516, Loss: {'train/loss': 0.47345513105392456, 'train/grad_norm_pre_clip': 68.5793228149414, 'timing/data': 0.0002109199995175004, 'timing/model': 0.36880271199333947, 'learning_rate/qwen_vl_interface': 8.173325300261664e-06, 'learning_rate/action_model': 8.173325300261663e-05, 'epoch': 0.11}) +[2026-10-03 14:55:08,368][starVLA.training.train_starvla][INFO] - Step 1517, Loss: {'train/loss': 0.0003619990311563015, 'train/grad_norm_pre_clip': 0.0806717723608017, 'timing/data': 0.0002543700102251023, 'timing/model': 0.5658874759974424, 'learning_rate/qwen_vl_interface': 8.170924340567617e-06, 'learning_rate/action_model': 8.170924340567616e-05, 'epoch': 0.11}) +[2026-10-03 14:55:09,025][starVLA.training.train_starvla][INFO] - Step 1518, Loss: {'train/loss': 0.0006153016001917422, 'train/grad_norm_pre_clip': 0.1197018176317215, 'timing/data': 0.00021256000036373734, 'timing/model': 0.36931283300509676, 'learning_rate/qwen_vl_interface': 8.168522180191909e-06, 'learning_rate/action_model': 8.168522180191909e-05, 'epoch': 0.11}) +[2026-10-03 14:55:09,681][starVLA.training.train_starvla][INFO] - Step 1519, Loss: {'train/loss': 0.7772822380065918, 'train/grad_norm_pre_clip': 67.45880889892578, 'timing/data': 0.0002886199945351109, 'timing/model': 0.36853418199461885, 'learning_rate/qwen_vl_interface': 8.166118820121972e-06, 'learning_rate/action_model': 8.166118820121973e-05, 'epoch': 0.11}) +[2026-10-03 14:55:10,314][starVLA.training.train_starvla][INFO] - Step 1520, Loss: {'train/loss': 0.34408077597618103, 'train/grad_norm_pre_clip': 29.653823852539062, 'timing/data': 0.00032952000037766993, 'timing/model': 0.36822301299253013, 'learning_rate/qwen_vl_interface': 8.163714261345744e-06, 'learning_rate/action_model': 8.163714261345742e-05, 'epoch': 0.11}) +[2026-10-03 14:55:10,972][starVLA.training.train_starvla][INFO] - Step 1521, Loss: {'train/loss': 0.00013937940821051598, 'train/grad_norm_pre_clip': 0.015450800769031048, 'timing/data': 0.0002929500042228028, 'timing/model': 0.37010762200225145, 'learning_rate/qwen_vl_interface': 8.16130850485164e-06, 'learning_rate/action_model': 8.161308504851639e-05, 'epoch': 0.11}) +[2026-10-03 14:55:11,629][starVLA.training.train_starvla][INFO] - Step 1522, Loss: {'train/loss': 0.2771964371204376, 'train/grad_norm_pre_clip': 69.9169692993164, 'timing/data': 0.00033618000452406704, 'timing/model': 0.36919072199088987, 'learning_rate/qwen_vl_interface': 8.158901551628579e-06, 'learning_rate/action_model': 8.158901551628579e-05, 'epoch': 0.11}) +[2026-10-03 14:55:12,474][starVLA.training.train_starvla][INFO] - Step 1523, Loss: {'train/loss': 1.9650224447250366, 'train/grad_norm_pre_clip': 70.19461059570312, 'timing/data': 0.000338379992172122, 'timing/model': 0.3688309819990536, 'learning_rate/qwen_vl_interface': 8.156493402665971e-06, 'learning_rate/action_model': 8.15649340266597e-05, 'epoch': 0.11}) +[2026-10-03 14:55:13,107][starVLA.training.train_starvla][INFO] - Step 1524, Loss: {'train/loss': 0.00608938280493021, 'train/grad_norm_pre_clip': 5.356515884399414, 'timing/data': 0.00021117000142112374, 'timing/model': 0.3683804419997614, 'learning_rate/qwen_vl_interface': 8.154084058953713e-06, 'learning_rate/action_model': 8.154084058953713e-05, 'epoch': 0.11}) +[2026-10-03 14:55:13,765][starVLA.training.train_starvla][INFO] - Step 1525, Loss: {'train/loss': 0.013439045287668705, 'train/grad_norm_pre_clip': 8.92819595336914, 'timing/data': 0.0002134200040018186, 'timing/model': 0.3706886720028706, 'learning_rate/qwen_vl_interface': 8.151673521482193e-06, 'learning_rate/action_model': 8.151673521482194e-05, 'epoch': 0.11}) +[2026-10-03 14:55:14,398][starVLA.training.train_starvla][INFO] - Step 1526, Loss: {'train/loss': 0.0023791478015482426, 'train/grad_norm_pre_clip': 0.6212197542190552, 'timing/data': 0.00025815000117290765, 'timing/model': 0.3455795960035175, 'learning_rate/qwen_vl_interface': 8.149261791242297e-06, 'learning_rate/action_model': 8.149261791242297e-05, 'epoch': 0.11}) +[2026-10-03 14:55:15,055][starVLA.training.train_starvla][INFO] - Step 1527, Loss: {'train/loss': 0.002123417565599084, 'train/grad_norm_pre_clip': 0.4060266613960266, 'timing/data': 0.00033175900171045214, 'timing/model': 0.368900452987873, 'learning_rate/qwen_vl_interface': 8.146848869225391e-06, 'learning_rate/action_model': 8.146848869225391e-05, 'epoch': 0.11}) +[2026-10-03 14:55:15,911][starVLA.training.train_starvla][INFO] - Step 1528, Loss: {'train/loss': 0.0071417358703911304, 'train/grad_norm_pre_clip': 3.3059775829315186, 'timing/data': 0.00035453999589663, 'timing/model': 0.3687048019928625, 'learning_rate/qwen_vl_interface': 8.14443475642334e-06, 'learning_rate/action_model': 8.144434756423339e-05, 'epoch': 0.11}) +[2026-10-03 14:55:16,568][starVLA.training.train_starvla][INFO] - Step 1529, Loss: {'train/loss': 0.6503329277038574, 'train/grad_norm_pre_clip': 19.513134002685547, 'timing/data': 0.0003460100124357268, 'timing/model': 0.3694021630071802, 'learning_rate/qwen_vl_interface': 8.142019453828493e-06, 'learning_rate/action_model': 8.142019453828493e-05, 'epoch': 0.11}) +[2026-10-03 14:55:17,200][starVLA.training.train_starvla][INFO] - Step 1530, Loss: {'train/loss': 0.0020532463677227497, 'train/grad_norm_pre_clip': 2.684540033340454, 'timing/data': 0.00021023000590503216, 'timing/model': 0.3454633169894805, 'learning_rate/qwen_vl_interface': 8.139602962433688e-06, 'learning_rate/action_model': 8.139602962433688e-05, 'epoch': 0.11}) +[2026-10-03 14:55:17,856][starVLA.training.train_starvla][INFO] - Step 1531, Loss: {'train/loss': 0.0008729916880838573, 'train/grad_norm_pre_clip': 0.20581892132759094, 'timing/data': 0.00026326999068260193, 'timing/model': 0.3685584030026803, 'learning_rate/qwen_vl_interface': 8.137185283232255e-06, 'learning_rate/action_model': 8.137185283232256e-05, 'epoch': 0.11}) +[2026-10-03 14:55:18,511][starVLA.training.train_starvla][INFO] - Step 1532, Loss: {'train/loss': 0.001299084397032857, 'train/grad_norm_pre_clip': 1.363957166671753, 'timing/data': 0.00020929999300278723, 'timing/model': 0.3687003120139707, 'learning_rate/qwen_vl_interface': 8.134766417218014e-06, 'learning_rate/action_model': 8.134766417218012e-05, 'epoch': 0.11}) +[2026-10-03 14:55:19,168][starVLA.training.train_starvla][INFO] - Step 1533, Loss: {'train/loss': 0.0011831304291263223, 'train/grad_norm_pre_clip': 0.27039337158203125, 'timing/data': 0.0002940499980468303, 'timing/model': 0.369917791991611, 'learning_rate/qwen_vl_interface': 8.132346365385262e-06, 'learning_rate/action_model': 8.132346365385261e-05, 'epoch': 0.11}) +[2026-10-03 14:55:20,021][starVLA.training.train_starvla][INFO] - Step 1534, Loss: {'train/loss': 0.0017911982722580433, 'train/grad_norm_pre_clip': 0.5832886099815369, 'timing/data': 0.00030128000071272254, 'timing/model': 0.36817962299392093, 'learning_rate/qwen_vl_interface': 8.129925128728796e-06, 'learning_rate/action_model': 8.129925128728796e-05, 'epoch': 0.11}) +[2026-10-03 14:55:20,677][starVLA.training.train_starvla][INFO] - Step 1535, Loss: {'train/loss': 0.29533687233924866, 'train/grad_norm_pre_clip': 77.39813995361328, 'timing/data': 0.00025242000992875546, 'timing/model': 0.3683883630001219, 'learning_rate/qwen_vl_interface': 8.127502708243895e-06, 'learning_rate/action_model': 8.127502708243895e-05, 'epoch': 0.11}) +[2026-10-03 14:55:21,332][starVLA.training.train_starvla][INFO] - Step 1536, Loss: {'train/loss': 0.14363697171211243, 'train/grad_norm_pre_clip': 65.99417877197266, 'timing/data': 0.0003311300097266212, 'timing/model': 0.36820976200397126, 'learning_rate/qwen_vl_interface': 8.125079104926324e-06, 'learning_rate/action_model': 8.125079104926323e-05, 'epoch': 0.11}) +[2026-10-03 14:55:21,990][starVLA.training.train_starvla][INFO] - Step 1537, Loss: {'train/loss': 0.0003238810459151864, 'train/grad_norm_pre_clip': 0.053861431777477264, 'timing/data': 0.0002633700059959665, 'timing/model': 0.37073426198912784, 'learning_rate/qwen_vl_interface': 8.122654319772332e-06, 'learning_rate/action_model': 8.122654319772332e-05, 'epoch': 0.11}) +[2026-10-03 14:55:22,644][starVLA.training.train_starvla][INFO] - Step 1538, Loss: {'train/loss': 0.0010338572319597006, 'train/grad_norm_pre_clip': 1.0993372201919556, 'timing/data': 0.0003284900012658909, 'timing/model': 0.3681395929888822, 'learning_rate/qwen_vl_interface': 8.120228353778661e-06, 'learning_rate/action_model': 8.12022835377866e-05, 'epoch': 0.11}) +[2026-10-03 14:55:23,499][starVLA.training.train_starvla][INFO] - Step 1539, Loss: {'train/loss': 0.00017933369963429868, 'train/grad_norm_pre_clip': 0.04944339022040367, 'timing/data': 0.0004801800096174702, 'timing/model': 0.36889490199973807, 'learning_rate/qwen_vl_interface': 8.117801207942528e-06, 'learning_rate/action_model': 8.117801207942528e-05, 'epoch': 0.11}) +[2026-10-03 14:55:24,131][starVLA.training.train_starvla][INFO] - Step 1540, Loss: {'train/loss': 0.0008350385469384491, 'train/grad_norm_pre_clip': 0.2166161835193634, 'timing/data': 0.0004232100036460906, 'timing/model': 0.3681071129976772, 'learning_rate/qwen_vl_interface': 8.115372883261648e-06, 'learning_rate/action_model': 8.115372883261647e-05, 'epoch': 0.11}) +[2026-10-03 14:55:24,788][starVLA.training.train_starvla][INFO] - Step 1541, Loss: {'train/loss': 0.4588671922683716, 'train/grad_norm_pre_clip': 17.356435775756836, 'timing/data': 0.0002544499875511974, 'timing/model': 0.37024109199410304, 'learning_rate/qwen_vl_interface': 8.11294338073421e-06, 'learning_rate/action_model': 8.11294338073421e-05, 'epoch': 0.11}) +[2026-10-03 14:55:25,444][starVLA.training.train_starvla][INFO] - Step 1542, Loss: {'train/loss': 0.0010238670511171222, 'train/grad_norm_pre_clip': 0.27292269468307495, 'timing/data': 0.0003360900009283796, 'timing/model': 0.3684470529988175, 'learning_rate/qwen_vl_interface': 8.110512701358892e-06, 'learning_rate/action_model': 8.110512701358891e-05, 'epoch': 0.11}) +[2026-10-03 14:55:26,078][starVLA.training.train_starvla][INFO] - Step 1543, Loss: {'train/loss': 0.34629231691360474, 'train/grad_norm_pre_clip': 45.07528305053711, 'timing/data': 0.0002744800003711134, 'timing/model': 0.34553287700691726, 'learning_rate/qwen_vl_interface': 8.108080846134852e-06, 'learning_rate/action_model': 8.108080846134852e-05, 'epoch': 0.11}) +[2026-10-03 14:55:26,901][starVLA.training.train_starvla][INFO] - Step 1544, Loss: {'train/loss': 0.3094950020313263, 'train/grad_norm_pre_clip': 51.42971420288086, 'timing/data': 0.00030035999952815473, 'timing/model': 0.5361701319925487, 'learning_rate/qwen_vl_interface': 8.105647816061734e-06, 'learning_rate/action_model': 8.105647816061735e-05, 'epoch': 0.11}) +[2026-10-03 14:55:27,558][starVLA.training.train_starvla][INFO] - Step 1545, Loss: {'train/loss': 0.0005639268783852458, 'train/grad_norm_pre_clip': 0.10548540949821472, 'timing/data': 0.0003210499999113381, 'timing/model': 0.3694835030037211, 'learning_rate/qwen_vl_interface': 8.103213612139667e-06, 'learning_rate/action_model': 8.103213612139667e-05, 'epoch': 0.11}) +[2026-10-03 14:55:28,215][starVLA.training.train_starvla][INFO] - Step 1546, Loss: {'train/loss': 0.00048823156976141036, 'train/grad_norm_pre_clip': 0.09839661419391632, 'timing/data': 0.00034344999585300684, 'timing/model': 0.3691302819934208, 'learning_rate/qwen_vl_interface': 8.100778235369262e-06, 'learning_rate/action_model': 8.10077823536926e-05, 'epoch': 0.11}) +[2026-10-03 14:55:28,872][starVLA.training.train_starvla][INFO] - Step 1547, Loss: {'train/loss': 0.00037201103987172246, 'train/grad_norm_pre_clip': 0.056270238012075424, 'timing/data': 0.0003592799912439659, 'timing/model': 0.3693565830035368, 'learning_rate/qwen_vl_interface': 8.098341686751605e-06, 'learning_rate/action_model': 8.098341686751605e-05, 'epoch': 0.11}) +[2026-10-03 14:55:29,527][starVLA.training.train_starvla][INFO] - Step 1548, Loss: {'train/loss': 0.00018983683548867702, 'train/grad_norm_pre_clip': 0.029625270515680313, 'timing/data': 0.00034487000084482133, 'timing/model': 0.368609352997737, 'learning_rate/qwen_vl_interface': 8.095903967288273e-06, 'learning_rate/action_model': 8.095903967288272e-05, 'epoch': 0.11}) +[2026-10-03 14:55:30,184][starVLA.training.train_starvla][INFO] - Step 1549, Loss: {'train/loss': 9.885339386528358e-05, 'train/grad_norm_pre_clip': 0.012992230243980885, 'timing/data': 0.0003207099944120273, 'timing/model': 0.36981900299724657, 'learning_rate/qwen_vl_interface': 8.093465077981318e-06, 'learning_rate/action_model': 8.093465077981318e-05, 'epoch': 0.11}) +[2026-10-03 14:55:31,037][starVLA.training.train_starvla][INFO] - Step 1550, Loss: {'train/loss': 0.6022149920463562, 'train/grad_norm_pre_clip': 86.4692611694336, 'timing/data': 0.00041847999091260135, 'timing/model': 0.368236053007422, 'learning_rate/qwen_vl_interface': 8.091025019833276e-06, 'learning_rate/action_model': 8.091025019833276e-05, 'epoch': 0.11}) +[2026-10-03 14:55:31,674][starVLA.training.train_starvla][INFO] - Step 1551, Loss: {'train/loss': 0.9403625130653381, 'train/grad_norm_pre_clip': 199.37347412109375, 'timing/data': 0.00029075000202283263, 'timing/model': 0.36968792299740016, 'learning_rate/qwen_vl_interface': 8.088583793847165e-06, 'learning_rate/action_model': 8.088583793847164e-05, 'epoch': 0.11}) +[2026-10-03 14:55:32,310][starVLA.training.train_starvla][INFO] - Step 1552, Loss: {'train/loss': 0.00028393944376148283, 'train/grad_norm_pre_clip': 0.04354608803987503, 'timing/data': 0.0003373200015630573, 'timing/model': 0.34725725700263865, 'learning_rate/qwen_vl_interface': 8.086141401026478e-06, 'learning_rate/action_model': 8.086141401026478e-05, 'epoch': 0.11}) +[2026-10-03 14:55:32,970][starVLA.training.train_starvla][INFO] - Step 1553, Loss: {'train/loss': 0.00020870838488917798, 'train/grad_norm_pre_clip': 0.04746084287762642, 'timing/data': 0.00033773000177461654, 'timing/model': 0.3715022420074092, 'learning_rate/qwen_vl_interface': 8.083697842375193e-06, 'learning_rate/action_model': 8.083697842375193e-05, 'epoch': 0.11}) +[2026-10-03 14:55:33,629][starVLA.training.train_starvla][INFO] - Step 1554, Loss: {'train/loss': 0.00019043628708459437, 'train/grad_norm_pre_clip': 0.02980334311723709, 'timing/data': 0.0002611800009617582, 'timing/model': 0.36979417300608475, 'learning_rate/qwen_vl_interface': 8.081253118897762e-06, 'learning_rate/action_model': 8.081253118897761e-05, 'epoch': 0.11}) +[2026-10-03 14:55:34,478][starVLA.training.train_starvla][INFO] - Step 1555, Loss: {'train/loss': 0.18799902498722076, 'train/grad_norm_pre_clip': 64.05072784423828, 'timing/data': 0.00025074000586755574, 'timing/model': 0.5607801279984415, 'learning_rate/qwen_vl_interface': 8.07880723159912e-06, 'learning_rate/action_model': 8.07880723159912e-05, 'epoch': 0.11}) +[2026-10-03 14:55:35,134][starVLA.training.train_starvla][INFO] - Step 1556, Loss: {'train/loss': 0.00018930106307379901, 'train/grad_norm_pre_clip': 0.03251073881983757, 'timing/data': 0.0002087900065816939, 'timing/model': 0.36846121300186496, 'learning_rate/qwen_vl_interface': 8.076360181484676e-06, 'learning_rate/action_model': 8.076360181484676e-05, 'epoch': 0.11}) +[2026-10-03 14:55:35,792][starVLA.training.train_starvla][INFO] - Step 1557, Loss: {'train/loss': 0.009869913570582867, 'train/grad_norm_pre_clip': 6.4676642417907715, 'timing/data': 0.00029929001175332814, 'timing/model': 0.3710924330080161, 'learning_rate/qwen_vl_interface': 8.073911969560326e-06, 'learning_rate/action_model': 8.073911969560326e-05, 'epoch': 0.11}) +[2026-10-03 14:55:36,448][starVLA.training.train_starvla][INFO] - Step 1558, Loss: {'train/loss': 0.00026288526714779437, 'train/grad_norm_pre_clip': 0.043099720031023026, 'timing/data': 0.0003403000009711832, 'timing/model': 0.3684387829998741, 'learning_rate/qwen_vl_interface': 8.071462596832432e-06, 'learning_rate/action_model': 8.071462596832432e-05, 'epoch': 0.11}) +[2026-10-03 14:55:37,084][starVLA.training.train_starvla][INFO] - Step 1559, Loss: {'train/loss': 0.0031501660123467445, 'train/grad_norm_pre_clip': 4.191587448120117, 'timing/data': 0.0003366900054970756, 'timing/model': 0.3464205869968282, 'learning_rate/qwen_vl_interface': 8.06901206430784e-06, 'learning_rate/action_model': 8.06901206430784e-05, 'epoch': 0.11}) +[2026-10-03 14:55:37,718][starVLA.training.train_starvla][INFO] - Step 1560, Loss: {'train/loss': 0.00025295180967077613, 'train/grad_norm_pre_clip': 0.08663598448038101, 'timing/data': 0.000208480007131584, 'timing/model': 0.3693094829941401, 'learning_rate/qwen_vl_interface': 8.066560372993873e-06, 'learning_rate/action_model': 8.066560372993874e-05, 'epoch': 0.11}) +[2026-10-03 14:55:38,563][starVLA.training.train_starvla][INFO] - Step 1561, Loss: {'train/loss': 1.1299813985824585, 'train/grad_norm_pre_clip': 182.97959899902344, 'timing/data': 0.00035249000939074904, 'timing/model': 0.3701113229908515, 'learning_rate/qwen_vl_interface': 8.064107523898327e-06, 'learning_rate/action_model': 8.064107523898326e-05, 'epoch': 0.11}) +[2026-10-03 14:55:39,219][starVLA.training.train_starvla][INFO] - Step 1562, Loss: {'train/loss': 0.2772098481655121, 'train/grad_norm_pre_clip': 65.08415222167969, 'timing/data': 0.0003123700007563457, 'timing/model': 0.3685332640015986, 'learning_rate/qwen_vl_interface': 8.061653518029476e-06, 'learning_rate/action_model': 8.061653518029476e-05, 'epoch': 0.11}) +[2026-10-03 14:55:39,875][starVLA.training.train_starvla][INFO] - Step 1563, Loss: {'train/loss': 0.21003380417823792, 'train/grad_norm_pre_clip': 39.514530181884766, 'timing/data': 0.00026359000185038894, 'timing/model': 0.36838763298874255, 'learning_rate/qwen_vl_interface': 8.059198356396069e-06, 'learning_rate/action_model': 8.059198356396068e-05, 'epoch': 0.11}) +[2026-10-03 14:55:40,507][starVLA.training.train_starvla][INFO] - Step 1564, Loss: {'train/loss': 0.0006190520361997187, 'train/grad_norm_pre_clip': 0.33271142840385437, 'timing/data': 0.0002695299917832017, 'timing/model': 0.3456042769976193, 'learning_rate/qwen_vl_interface': 8.05674204000733e-06, 'learning_rate/action_model': 8.05674204000733e-05, 'epoch': 0.11}) +[2026-10-03 14:55:41,166][starVLA.training.train_starvla][INFO] - Step 1565, Loss: {'train/loss': 0.0019174432381987572, 'train/grad_norm_pre_clip': 1.525373935699463, 'timing/data': 0.000354260002495721, 'timing/model': 0.3713640030036913, 'learning_rate/qwen_vl_interface': 8.05428456987296e-06, 'learning_rate/action_model': 8.05428456987296e-05, 'epoch': 0.11}) +[2026-10-03 14:55:42,013][starVLA.training.train_starvla][INFO] - Step 1566, Loss: {'train/loss': 0.00169029098469764, 'train/grad_norm_pre_clip': 0.7903524041175842, 'timing/data': 0.00026521900144871324, 'timing/model': 0.5604517090105219, 'learning_rate/qwen_vl_interface': 8.05182594700313e-06, 'learning_rate/action_model': 8.05182594700313e-05, 'epoch': 0.11}) +[2026-10-03 14:55:42,669][starVLA.training.train_starvla][INFO] - Step 1567, Loss: {'train/loss': 0.3803805112838745, 'train/grad_norm_pre_clip': 16.305248260498047, 'timing/data': 0.00024826001026667655, 'timing/model': 0.3684979130048305, 'learning_rate/qwen_vl_interface': 8.049366172408484e-06, 'learning_rate/action_model': 8.049366172408485e-05, 'epoch': 0.11}) +[2026-10-03 14:55:43,325][starVLA.training.train_starvla][INFO] - Step 1568, Loss: {'train/loss': 0.12143383920192719, 'train/grad_norm_pre_clip': 104.56795501708984, 'timing/data': 0.00026837000041268766, 'timing/model': 0.3684800929913763, 'learning_rate/qwen_vl_interface': 8.04690524710015e-06, 'learning_rate/action_model': 8.046905247100151e-05, 'epoch': 0.11}) +[2026-10-03 14:55:43,983][starVLA.training.train_starvla][INFO] - Step 1569, Loss: {'train/loss': 0.0001907588739413768, 'train/grad_norm_pre_clip': 0.03353024274110794, 'timing/data': 0.0002720299962675199, 'timing/model': 0.3701932440017117, 'learning_rate/qwen_vl_interface': 8.044443172089715e-06, 'learning_rate/action_model': 8.044443172089714e-05, 'epoch': 0.11}) +[2026-10-03 14:55:44,638][starVLA.training.train_starvla][INFO] - Step 1570, Loss: {'train/loss': 0.00010893921717070043, 'train/grad_norm_pre_clip': 0.01856442354619503, 'timing/data': 0.0002556300023570657, 'timing/model': 0.36855772299168166, 'learning_rate/qwen_vl_interface': 8.041979948389248e-06, 'learning_rate/action_model': 8.041979948389248e-05, 'epoch': 0.11}) +[2026-10-03 14:55:45,482][starVLA.training.train_starvla][INFO] - Step 1571, Loss: {'train/loss': 0.0006534865824505687, 'train/grad_norm_pre_clip': 0.3604899048805237, 'timing/data': 0.0003101799957221374, 'timing/model': 0.5558481509942794, 'learning_rate/qwen_vl_interface': 8.039515577011287e-06, 'learning_rate/action_model': 8.039515577011288e-05, 'epoch': 0.11}) +[2026-10-03 14:55:46,139][starVLA.training.train_starvla][INFO] - Step 1572, Loss: {'train/loss': 0.16479012370109558, 'train/grad_norm_pre_clip': 194.3116455078125, 'timing/data': 0.00028589999419637024, 'timing/model': 0.36862029301119037, 'learning_rate/qwen_vl_interface': 8.037050058968841e-06, 'learning_rate/action_model': 8.03705005896884e-05, 'epoch': 0.12}) +[2026-10-03 14:55:46,797][starVLA.training.train_starvla][INFO] - Step 1573, Loss: {'train/loss': 0.0021865209564566612, 'train/grad_norm_pre_clip': 1.0259299278259277, 'timing/data': 0.00032233000092674047, 'timing/model': 0.3701936440047575, 'learning_rate/qwen_vl_interface': 8.034583395275392e-06, 'learning_rate/action_model': 8.034583395275392e-05, 'epoch': 0.12}) +[2026-10-03 14:55:47,453][starVLA.training.train_starvla][INFO] - Step 1574, Loss: {'train/loss': 0.8888007402420044, 'train/grad_norm_pre_clip': 58.51990509033203, 'timing/data': 0.0003352899948367849, 'timing/model': 0.36877819399524014, 'learning_rate/qwen_vl_interface': 8.032115586944893e-06, 'learning_rate/action_model': 8.032115586944892e-05, 'epoch': 0.12}) +[2026-10-03 14:55:48,110][starVLA.training.train_starvla][INFO] - Step 1575, Loss: {'train/loss': 3.7891644751653075e-05, 'train/grad_norm_pre_clip': 0.007206655573099852, 'timing/data': 0.0002158900024369359, 'timing/model': 0.36909630299487617, 'learning_rate/qwen_vl_interface': 8.029646634991762e-06, 'learning_rate/action_model': 8.029646634991762e-05, 'epoch': 0.12}) +[2026-10-03 14:55:48,744][starVLA.training.train_starvla][INFO] - Step 1576, Loss: {'train/loss': 0.4220220446586609, 'train/grad_norm_pre_clip': 332.2398986816406, 'timing/data': 0.0002698900061659515, 'timing/model': 0.36863563299993984, 'learning_rate/qwen_vl_interface': 8.0271765404309e-06, 'learning_rate/action_model': 8.0271765404309e-05, 'epoch': 0.12}) +[2026-10-03 14:55:49,604][starVLA.training.train_starvla][INFO] - Step 1577, Loss: {'train/loss': 0.5421153903007507, 'train/grad_norm_pre_clip': 60.48286437988281, 'timing/data': 0.00032140999974217266, 'timing/model': 0.37160886300262064, 'learning_rate/qwen_vl_interface': 8.024705304277664e-06, 'learning_rate/action_model': 8.024705304277664e-05, 'epoch': 0.12}) +[2026-10-03 14:55:50,237][starVLA.training.train_starvla][INFO] - Step 1578, Loss: {'train/loss': 0.42589202523231506, 'train/grad_norm_pre_clip': 36.70445251464844, 'timing/data': 0.00021158000163268298, 'timing/model': 0.36854345399478916, 'learning_rate/qwen_vl_interface': 8.022232927547888e-06, 'learning_rate/action_model': 8.022232927547886e-05, 'epoch': 0.12}) +[2026-10-03 14:55:50,894][starVLA.training.train_starvla][INFO] - Step 1579, Loss: {'train/loss': 0.11884775757789612, 'train/grad_norm_pre_clip': 38.19065856933594, 'timing/data': 0.00035285999183543026, 'timing/model': 0.3689810140058398, 'learning_rate/qwen_vl_interface': 8.01975941125787e-06, 'learning_rate/action_model': 8.01975941125787e-05, 'epoch': 0.12}) +[2026-10-03 14:55:51,551][starVLA.training.train_starvla][INFO] - Step 1580, Loss: {'train/loss': 0.02563527226448059, 'train/grad_norm_pre_clip': 53.31945037841797, 'timing/data': 0.00033086999610532075, 'timing/model': 0.3685868339962326, 'learning_rate/qwen_vl_interface': 8.017284756424381e-06, 'learning_rate/action_model': 8.017284756424381e-05, 'epoch': 0.12}) +[2026-10-03 14:55:52,209][starVLA.training.train_starvla][INFO] - Step 1581, Loss: {'train/loss': 0.0007302460726350546, 'train/grad_norm_pre_clip': 0.49618202447891235, 'timing/data': 0.00026290999085176736, 'timing/model': 0.37095673399744555, 'learning_rate/qwen_vl_interface': 8.01480896406466e-06, 'learning_rate/action_model': 8.01480896406466e-05, 'epoch': 0.12}) +[2026-10-03 14:55:53,034][starVLA.training.train_starvla][INFO] - Step 1582, Loss: {'train/loss': 0.001582691096700728, 'train/grad_norm_pre_clip': 0.7509132027626038, 'timing/data': 0.00033302999509032816, 'timing/model': 0.3686394739925163, 'learning_rate/qwen_vl_interface': 8.01233203519641e-06, 'learning_rate/action_model': 8.012332035196409e-05, 'epoch': 0.12}) +[2026-10-03 14:55:53,691][starVLA.training.train_starvla][INFO] - Step 1583, Loss: {'train/loss': 0.3163830041885376, 'train/grad_norm_pre_clip': 68.96717071533203, 'timing/data': 0.00021417000971268862, 'timing/model': 0.3691127539932495, 'learning_rate/qwen_vl_interface': 8.009853970837801e-06, 'learning_rate/action_model': 8.009853970837801e-05, 'epoch': 0.12}) +[2026-10-03 14:55:54,346][starVLA.training.train_starvla][INFO] - Step 1584, Loss: {'train/loss': 0.0006803517462685704, 'train/grad_norm_pre_clip': 0.8716027736663818, 'timing/data': 0.000211919003049843, 'timing/model': 0.3688336039922433, 'learning_rate/qwen_vl_interface': 8.007374772007474e-06, 'learning_rate/action_model': 8.007374772007474e-05, 'epoch': 0.12}) +[2026-10-03 14:55:55,004][starVLA.training.train_starvla][INFO] - Step 1585, Loss: {'train/loss': 0.0002734874142333865, 'train/grad_norm_pre_clip': 0.07255525887012482, 'timing/data': 0.0003175799938617274, 'timing/model': 0.37039139299304225, 'learning_rate/qwen_vl_interface': 8.004894439724533e-06, 'learning_rate/action_model': 8.004894439724533e-05, 'epoch': 0.12}) +[2026-10-03 14:55:55,659][starVLA.training.train_starvla][INFO] - Step 1586, Loss: {'train/loss': 0.0026001171208918095, 'train/grad_norm_pre_clip': 2.4268014430999756, 'timing/data': 0.00021255998581182212, 'timing/model': 0.3686179140058812, 'learning_rate/qwen_vl_interface': 8.002412975008552e-06, 'learning_rate/action_model': 8.002412975008552e-05, 'epoch': 0.12}) +[2026-10-03 14:55:56,518][starVLA.training.train_starvla][INFO] - Step 1587, Loss: {'train/loss': 0.0711403489112854, 'train/grad_norm_pre_clip': 82.45368957519531, 'timing/data': 0.00021210999693721533, 'timing/model': 0.5707635279977694, 'learning_rate/qwen_vl_interface': 7.999930378879562e-06, 'learning_rate/action_model': 7.999930378879561e-05, 'epoch': 0.12}) +[2026-10-03 14:55:57,173][starVLA.training.train_starvla][INFO] - Step 1588, Loss: {'train/loss': 0.399691104888916, 'train/grad_norm_pre_clip': 46.102787017822266, 'timing/data': 0.0002498499961802736, 'timing/model': 0.3684850429999642, 'learning_rate/qwen_vl_interface': 7.997446652358067e-06, 'learning_rate/action_model': 7.997446652358068e-05, 'epoch': 0.12}) +[2026-10-03 14:55:57,832][starVLA.training.train_starvla][INFO] - Step 1589, Loss: {'train/loss': 0.003902970813214779, 'train/grad_norm_pre_clip': 1.5384502410888672, 'timing/data': 0.00029085000278428197, 'timing/model': 0.37149050399602856, 'learning_rate/qwen_vl_interface': 7.994961796465035e-06, 'learning_rate/action_model': 7.994961796465035e-05, 'epoch': 0.12}) +[2026-10-03 14:55:58,489][starVLA.training.train_starvla][INFO] - Step 1590, Loss: {'train/loss': 0.0006439905264414847, 'train/grad_norm_pre_clip': 0.32943034172058105, 'timing/data': 0.00020620999566745013, 'timing/model': 0.36862629299866967, 'learning_rate/qwen_vl_interface': 7.992475812221894e-06, 'learning_rate/action_model': 7.992475812221894e-05, 'epoch': 0.12}) +[2026-10-03 14:55:59,149][starVLA.training.train_starvla][INFO] - Step 1591, Loss: {'train/loss': 0.2724061608314514, 'train/grad_norm_pre_clip': 104.07781982421875, 'timing/data': 0.00024340000527445227, 'timing/model': 0.36818489398865495, 'learning_rate/qwen_vl_interface': 7.989988700650538e-06, 'learning_rate/action_model': 7.989988700650538e-05, 'epoch': 0.12}) +[2026-10-03 14:55:59,804][starVLA.training.train_starvla][INFO] - Step 1592, Loss: {'train/loss': 0.0009793790522962809, 'train/grad_norm_pre_clip': 0.7879907488822937, 'timing/data': 0.00030449999030679464, 'timing/model': 0.3681172139913542, 'learning_rate/qwen_vl_interface': 7.987500462773326e-06, 'learning_rate/action_model': 7.987500462773326e-05, 'epoch': 0.12}) +[2026-10-03 14:56:00,656][starVLA.training.train_starvla][INFO] - Step 1593, Loss: {'train/loss': 0.0003227332781534642, 'train/grad_norm_pre_clip': 0.059892091900110245, 'timing/data': 0.0002934700023615733, 'timing/model': 0.3707555539876921, 'learning_rate/qwen_vl_interface': 7.985011099613078e-06, 'learning_rate/action_model': 7.985011099613077e-05, 'epoch': 0.12}) +[2026-10-03 14:56:01,312][starVLA.training.train_starvla][INFO] - Step 1594, Loss: {'train/loss': 0.00047271043877117336, 'train/grad_norm_pre_clip': 0.12563194334506989, 'timing/data': 0.00021415000082924962, 'timing/model': 0.368488915002672, 'learning_rate/qwen_vl_interface': 7.98252061219308e-06, 'learning_rate/action_model': 7.982520612193078e-05, 'epoch': 0.12}) +[2026-10-03 14:56:01,968][starVLA.training.train_starvla][INFO] - Step 1595, Loss: {'train/loss': 0.0006189050618559122, 'train/grad_norm_pre_clip': 0.2011607587337494, 'timing/data': 0.0002082900027744472, 'timing/model': 0.3681483139953343, 'learning_rate/qwen_vl_interface': 7.980029001537073e-06, 'learning_rate/action_model': 7.980029001537072e-05, 'epoch': 0.12}) +[2026-10-03 14:56:02,624][starVLA.training.train_starvla][INFO] - Step 1596, Loss: {'train/loss': 0.0003454145626164973, 'train/grad_norm_pre_clip': 0.13915568590164185, 'timing/data': 0.0002129299973603338, 'timing/model': 0.3692399440042209, 'learning_rate/qwen_vl_interface': 7.977536268669264e-06, 'learning_rate/action_model': 7.977536268669264e-05, 'epoch': 0.12}) +[2026-10-03 14:56:03,283][starVLA.training.train_starvla][INFO] - Step 1597, Loss: {'train/loss': 0.0006854555103927851, 'train/grad_norm_pre_clip': 0.43956536054611206, 'timing/data': 0.0002531600039219484, 'timing/model': 0.3706654540001182, 'learning_rate/qwen_vl_interface': 7.975042414614324e-06, 'learning_rate/action_model': 7.975042414614324e-05, 'epoch': 0.12}) +[2026-10-03 14:56:04,141][starVLA.training.train_starvla][INFO] - Step 1598, Loss: {'train/loss': 0.00020940710965078324, 'train/grad_norm_pre_clip': 0.1039772555232048, 'timing/data': 0.0002906600129790604, 'timing/model': 0.3692835250112694, 'learning_rate/qwen_vl_interface': 7.972547440397382e-06, 'learning_rate/action_model': 7.972547440397383e-05, 'epoch': 0.12}) +[2026-10-03 14:56:04,775][starVLA.training.train_starvla][INFO] - Step 1599, Loss: {'train/loss': 0.022609030827879906, 'train/grad_norm_pre_clip': 36.04244613647461, 'timing/data': 0.00025587000709492713, 'timing/model': 0.3457182690035552, 'learning_rate/qwen_vl_interface': 7.970051347044026e-06, 'learning_rate/action_model': 7.970051347044025e-05, 'epoch': 0.12}) +[2026-10-03 14:56:05,432][starVLA.training.train_starvla][INFO] - Step 1600, Loss: {'train/loss': 8.291396079584956e-05, 'train/grad_norm_pre_clip': 0.025033174082636833, 'timing/data': 0.0003306500002508983, 'timing/model': 0.36885040400375146, 'learning_rate/qwen_vl_interface': 7.967554135580307e-06, 'learning_rate/action_model': 7.967554135580307e-05, 'epoch': 0.12}) +[2026-10-03 14:56:06,092][starVLA.training.train_starvla][INFO] - Step 1601, Loss: {'train/loss': 2.5987204935518093e-05, 'train/grad_norm_pre_clip': 0.0024870068300515413, 'timing/data': 0.0003151099954266101, 'timing/model': 0.3712618239951553, 'learning_rate/qwen_vl_interface': 7.965055807032734e-06, 'learning_rate/action_model': 7.965055807032735e-05, 'epoch': 0.12}) +[2026-10-03 14:56:06,749][starVLA.training.train_starvla][INFO] - Step 1602, Loss: {'train/loss': 0.458173543214798, 'train/grad_norm_pre_clip': 42.04468536376953, 'timing/data': 0.00026200999855063856, 'timing/model': 0.36924556399753783, 'learning_rate/qwen_vl_interface': 7.962556362428275e-06, 'learning_rate/action_model': 7.962556362428274e-05, 'epoch': 0.12}) +[2026-10-03 14:56:07,612][starVLA.training.train_starvla][INFO] - Step 1603, Loss: {'train/loss': 0.04839503765106201, 'train/grad_norm_pre_clip': 59.891136169433594, 'timing/data': 0.0003514499985612929, 'timing/model': 0.36934451399429236, 'learning_rate/qwen_vl_interface': 7.960055802794358e-06, 'learning_rate/action_model': 7.960055802794357e-05, 'epoch': 0.12}) +[2026-10-03 14:56:08,269][starVLA.training.train_starvla][INFO] - Step 1604, Loss: {'train/loss': 0.0001241263817064464, 'train/grad_norm_pre_clip': 0.029348298907279968, 'timing/data': 0.00026984899886883795, 'timing/model': 0.36933701499947347, 'learning_rate/qwen_vl_interface': 7.957554129158866e-06, 'learning_rate/action_model': 7.957554129158866e-05, 'epoch': 0.12}) +[2026-10-03 14:56:08,928][starVLA.training.train_starvla][INFO] - Step 1605, Loss: {'train/loss': 0.0002651451213750988, 'train/grad_norm_pre_clip': 0.1645294576883316, 'timing/data': 0.0002927900059148669, 'timing/model': 0.3708088240091456, 'learning_rate/qwen_vl_interface': 7.955051342550148e-06, 'learning_rate/action_model': 7.955051342550148e-05, 'epoch': 0.12}) +[2026-10-03 14:56:09,562][starVLA.training.train_starvla][INFO] - Step 1606, Loss: {'train/loss': 0.6071436405181885, 'train/grad_norm_pre_clip': 71.21076965332031, 'timing/data': 0.00029260000155773014, 'timing/model': 0.3686963240033947, 'learning_rate/qwen_vl_interface': 7.952547443997e-06, 'learning_rate/action_model': 7.952547443997e-05, 'epoch': 0.12}) +[2026-10-03 14:56:10,220][starVLA.training.train_starvla][INFO] - Step 1607, Loss: {'train/loss': 2.754407250904478e-05, 'train/grad_norm_pre_clip': 0.004967035260051489, 'timing/data': 0.00024801999097689986, 'timing/model': 0.36907658400014043, 'learning_rate/qwen_vl_interface': 7.950042434528681e-06, 'learning_rate/action_model': 7.950042434528682e-05, 'epoch': 0.12}) +[2026-10-03 14:56:11,079][starVLA.training.train_starvla][INFO] - Step 1608, Loss: {'train/loss': 0.0003136568411719054, 'train/grad_norm_pre_clip': 0.15683378279209137, 'timing/data': 0.0003713100013555959, 'timing/model': 0.5719570780056529, 'learning_rate/qwen_vl_interface': 7.94753631517491e-06, 'learning_rate/action_model': 7.947536315174909e-05, 'epoch': 0.12}) +[2026-10-03 14:56:11,738][starVLA.training.train_starvla][INFO] - Step 1609, Loss: {'train/loss': 0.005403356160968542, 'train/grad_norm_pre_clip': 7.669480323791504, 'timing/data': 0.0002991799992742017, 'timing/model': 0.3708236739912536, 'learning_rate/qwen_vl_interface': 7.945029086965854e-06, 'learning_rate/action_model': 7.945029086965853e-05, 'epoch': 0.12}) +[2026-10-03 14:56:12,395][starVLA.training.train_starvla][INFO] - Step 1610, Loss: {'train/loss': 0.0004657465033233166, 'train/grad_norm_pre_clip': 0.8916107416152954, 'timing/data': 0.0003460199950495735, 'timing/model': 0.36906003500916995, 'learning_rate/qwen_vl_interface': 7.942520750932138e-06, 'learning_rate/action_model': 7.942520750932138e-05, 'epoch': 0.12}) +[2026-10-03 14:56:13,052][starVLA.training.train_starvla][INFO] - Step 1611, Loss: {'train/loss': 0.0005225600907579064, 'train/grad_norm_pre_clip': 0.4509492814540863, 'timing/data': 0.0002140400029020384, 'timing/model': 0.36944696500722785, 'learning_rate/qwen_vl_interface': 7.940011308104848e-06, 'learning_rate/action_model': 7.940011308104848e-05, 'epoch': 0.12}) +[2026-10-03 14:56:13,708][starVLA.training.train_starvla][INFO] - Step 1612, Loss: {'train/loss': 0.34850451350212097, 'train/grad_norm_pre_clip': 39.80945587158203, 'timing/data': 0.00024099998699966818, 'timing/model': 0.36881713499315083, 'learning_rate/qwen_vl_interface': 7.93750075951552e-06, 'learning_rate/action_model': 7.93750075951552e-05, 'epoch': 0.12}) +[2026-10-03 14:56:14,367][starVLA.training.train_starvla][INFO] - Step 1613, Loss: {'train/loss': 4.650398113881238e-05, 'train/grad_norm_pre_clip': 0.006242795381695032, 'timing/data': 0.00025907000235747546, 'timing/model': 0.3709713239950361, 'learning_rate/qwen_vl_interface': 7.934989106196145e-06, 'learning_rate/action_model': 7.934989106196146e-05, 'epoch': 0.12}) +[2026-10-03 14:56:15,228][starVLA.training.train_starvla][INFO] - Step 1614, Loss: {'train/loss': 0.029436394572257996, 'train/grad_norm_pre_clip': 33.464298248291016, 'timing/data': 0.00037885000347159803, 'timing/model': 0.3691925549937878, 'learning_rate/qwen_vl_interface': 7.93247634917917e-06, 'learning_rate/action_model': 7.93247634917917e-05, 'epoch': 0.12}) +[2026-10-03 14:56:15,863][starVLA.training.train_starvla][INFO] - Step 1615, Loss: {'train/loss': 0.6122617125511169, 'train/grad_norm_pre_clip': 30.245830535888672, 'timing/data': 0.00032438000198453665, 'timing/model': 0.37013331399066374, 'learning_rate/qwen_vl_interface': 7.929962489497494e-06, 'learning_rate/action_model': 7.929962489497493e-05, 'epoch': 0.12}) +[2026-10-03 14:56:16,521][starVLA.training.train_starvla][INFO] - Step 1616, Loss: {'train/loss': 8.256637374870479e-05, 'train/grad_norm_pre_clip': 0.030999798327684402, 'timing/data': 0.0003549000102793798, 'timing/model': 0.36986973499006126, 'learning_rate/qwen_vl_interface': 7.927447528184469e-06, 'learning_rate/action_model': 7.927447528184468e-05, 'epoch': 0.12}) +[2026-10-03 14:56:17,181][starVLA.training.train_starvla][INFO] - Step 1617, Loss: {'train/loss': 1.1772069931030273, 'train/grad_norm_pre_clip': 68.75963592529297, 'timing/data': 0.00033195999276358634, 'timing/model': 0.37126357400848065, 'learning_rate/qwen_vl_interface': 7.924931466273901e-06, 'learning_rate/action_model': 7.924931466273902e-05, 'epoch': 0.12}) +[2026-10-03 14:56:17,839][starVLA.training.train_starvla][INFO] - Step 1618, Loss: {'train/loss': 0.00011530353367561474, 'train/grad_norm_pre_clip': 0.027520282194018364, 'timing/data': 0.0004571399913402274, 'timing/model': 0.3703137849952327, 'learning_rate/qwen_vl_interface': 7.92241430480005e-06, 'learning_rate/action_model': 7.922414304800049e-05, 'epoch': 0.12}) +[2026-10-03 14:56:18,692][starVLA.training.train_starvla][INFO] - Step 1619, Loss: {'train/loss': 6.087429210310802e-05, 'train/grad_norm_pre_clip': 0.01111882459372282, 'timing/data': 0.0002669099922059104, 'timing/model': 0.5636271900002612, 'learning_rate/qwen_vl_interface': 7.919896044797622e-06, 'learning_rate/action_model': 7.91989604479762e-05, 'epoch': 0.12}) +[2026-10-03 14:56:19,349][starVLA.training.train_starvla][INFO] - Step 1620, Loss: {'train/loss': 0.013450264930725098, 'train/grad_norm_pre_clip': 40.52776336669922, 'timing/data': 0.00021348000154830515, 'timing/model': 0.3694104949972825, 'learning_rate/qwen_vl_interface': 7.91737668730178e-06, 'learning_rate/action_model': 7.91737668730178e-05, 'epoch': 0.12}) +[2026-10-03 14:56:20,009][starVLA.training.train_starvla][INFO] - Step 1621, Loss: {'train/loss': 4.2204977944493294e-05, 'train/grad_norm_pre_clip': 0.008525022305548191, 'timing/data': 0.000348169996868819, 'timing/model': 0.37154479399032425, 'learning_rate/qwen_vl_interface': 7.91485623334814e-06, 'learning_rate/action_model': 7.91485623334814e-05, 'epoch': 0.12}) +[2026-10-03 14:56:20,668][starVLA.training.train_starvla][INFO] - Step 1622, Loss: {'train/loss': 0.0002519094559829682, 'train/grad_norm_pre_clip': 0.08012541383504868, 'timing/data': 0.00021471999934874475, 'timing/model': 0.3702921350050019, 'learning_rate/qwen_vl_interface': 7.91233468397276e-06, 'learning_rate/action_model': 7.912334683972761e-05, 'epoch': 0.12}) +[2026-10-03 14:56:21,326][starVLA.training.train_starvla][INFO] - Step 1623, Loss: {'train/loss': 0.00033705084933899343, 'train/grad_norm_pre_clip': 0.704713761806488, 'timing/data': 0.00021845000446774065, 'timing/model': 0.3697688050015131, 'learning_rate/qwen_vl_interface': 7.90981204021216e-06, 'learning_rate/action_model': 7.90981204021216e-05, 'epoch': 0.12}) +[2026-10-03 14:56:21,960][starVLA.training.train_starvla][INFO] - Step 1624, Loss: {'train/loss': 0.0006243341485969722, 'train/grad_norm_pre_clip': 1.3045138120651245, 'timing/data': 0.0002936400123871863, 'timing/model': 0.37028562498744577, 'learning_rate/qwen_vl_interface': 7.907288303103298e-06, 'learning_rate/action_model': 7.907288303103298e-05, 'epoch': 0.12}) +[2026-10-03 14:56:22,821][starVLA.training.train_starvla][INFO] - Step 1625, Loss: {'train/loss': 0.0004921411164104939, 'train/grad_norm_pre_clip': 0.1471523940563202, 'timing/data': 0.0003960390022257343, 'timing/model': 0.3716349150054157, 'learning_rate/qwen_vl_interface': 7.90476347368359e-06, 'learning_rate/action_model': 7.90476347368359e-05, 'epoch': 0.12}) +[2026-10-03 14:56:23,481][starVLA.training.train_starvla][INFO] - Step 1626, Loss: {'train/loss': 0.2779651880264282, 'train/grad_norm_pre_clip': 120.43820190429688, 'timing/data': 0.0003495699929771945, 'timing/model': 0.3701152239955263, 'learning_rate/qwen_vl_interface': 7.902237552990897e-06, 'learning_rate/action_model': 7.902237552990897e-05, 'epoch': 0.12}) +[2026-10-03 14:56:24,142][starVLA.training.train_starvla][INFO] - Step 1627, Loss: {'train/loss': 0.851083517074585, 'train/grad_norm_pre_clip': 362.3135986328125, 'timing/data': 0.00046599000052083284, 'timing/model': 0.37082883399853017, 'learning_rate/qwen_vl_interface': 7.89971054206353e-06, 'learning_rate/action_model': 7.89971054206353e-05, 'epoch': 0.12}) +[2026-10-03 14:56:24,804][starVLA.training.train_starvla][INFO] - Step 1628, Loss: {'train/loss': 0.7603865265846252, 'train/grad_norm_pre_clip': 96.67405700683594, 'timing/data': 0.00021282999659888446, 'timing/model': 0.3732028140075272, 'learning_rate/qwen_vl_interface': 7.897182441940247e-06, 'learning_rate/action_model': 7.897182441940247e-05, 'epoch': 0.12}) +[2026-10-03 14:56:25,441][starVLA.training.train_starvla][INFO] - Step 1629, Loss: {'train/loss': 0.00046156556345522404, 'train/grad_norm_pre_clip': 0.14886043965816498, 'timing/data': 0.00021168000239413232, 'timing/model': 0.3487483579956461, 'learning_rate/qwen_vl_interface': 7.894653253660256e-06, 'learning_rate/action_model': 7.894653253660256e-05, 'epoch': 0.12}) +[2026-10-03 14:56:26,298][starVLA.training.train_starvla][INFO] - Step 1630, Loss: {'train/loss': 0.0010961112566292286, 'train/grad_norm_pre_clip': 0.4618791937828064, 'timing/data': 0.00021066000044811517, 'timing/model': 0.568580869003199, 'learning_rate/qwen_vl_interface': 7.89212297826321e-06, 'learning_rate/action_model': 7.892122978263209e-05, 'epoch': 0.12}) +[2026-10-03 14:56:26,955][starVLA.training.train_starvla][INFO] - Step 1631, Loss: {'train/loss': 0.00037134357262402773, 'train/grad_norm_pre_clip': 0.13244301080703735, 'timing/data': 0.00031391899392474443, 'timing/model': 0.36894420500902925, 'learning_rate/qwen_vl_interface': 7.889591616789206e-06, 'learning_rate/action_model': 7.889591616789206e-05, 'epoch': 0.12}) +[2026-10-03 14:56:27,610][starVLA.training.train_starvla][INFO] - Step 1632, Loss: {'train/loss': 0.7480552196502686, 'train/grad_norm_pre_clip': 22.331645965576172, 'timing/data': 0.00032825999369379133, 'timing/model': 0.3682129149965476, 'learning_rate/qwen_vl_interface': 7.887059170278799e-06, 'learning_rate/action_model': 7.887059170278797e-05, 'epoch': 0.12}) +[2026-10-03 14:56:28,269][starVLA.training.train_starvla][INFO] - Step 1633, Loss: {'train/loss': 0.2827945053577423, 'train/grad_norm_pre_clip': 95.15992736816406, 'timing/data': 0.00020715000573545694, 'timing/model': 0.3718541849957546, 'learning_rate/qwen_vl_interface': 7.884525639772974e-06, 'learning_rate/action_model': 7.884525639772974e-05, 'epoch': 0.12}) +[2026-10-03 14:56:28,927][starVLA.training.train_starvla][INFO] - Step 1634, Loss: {'train/loss': 0.0012611713027581573, 'train/grad_norm_pre_clip': 4.812566757202148, 'timing/data': 0.00033767000422813, 'timing/model': 0.3697484349977458, 'learning_rate/qwen_vl_interface': 7.881991026313176e-06, 'learning_rate/action_model': 7.881991026313175e-05, 'epoch': 0.12}) +[2026-10-03 14:56:29,758][starVLA.training.train_starvla][INFO] - Step 1635, Loss: {'train/loss': 0.002966409781947732, 'train/grad_norm_pre_clip': 4.167466163635254, 'timing/data': 0.00026589000481180847, 'timing/model': 0.5645846910047112, 'learning_rate/qwen_vl_interface': 7.879455330941283e-06, 'learning_rate/action_model': 7.879455330941282e-05, 'epoch': 0.12}) +[2026-10-03 14:56:30,414][starVLA.training.train_starvla][INFO] - Step 1636, Loss: {'train/loss': 0.00017962799756787717, 'train/grad_norm_pre_clip': 0.0387609601020813, 'timing/data': 0.0003249000001233071, 'timing/model': 0.3688103949971264, 'learning_rate/qwen_vl_interface': 7.876918554699627e-06, 'learning_rate/action_model': 7.876918554699627e-05, 'epoch': 0.12}) +[2026-10-03 14:56:31,073][starVLA.training.train_starvla][INFO] - Step 1637, Loss: {'train/loss': 0.6511213779449463, 'train/grad_norm_pre_clip': 42.262027740478516, 'timing/data': 0.00026792999415192753, 'timing/model': 0.37089099500735756, 'learning_rate/qwen_vl_interface': 7.874380698630982e-06, 'learning_rate/action_model': 7.874380698630982e-05, 'epoch': 0.12}) +[2026-10-03 14:56:31,730][starVLA.training.train_starvla][INFO] - Step 1638, Loss: {'train/loss': 0.044660910964012146, 'train/grad_norm_pre_clip': 48.783443450927734, 'timing/data': 0.0002794299944071099, 'timing/model': 0.36904037499334663, 'learning_rate/qwen_vl_interface': 7.871841763778562e-06, 'learning_rate/action_model': 7.871841763778563e-05, 'epoch': 0.12}) +[2026-10-03 14:56:32,388][starVLA.training.train_starvla][INFO] - Step 1639, Loss: {'train/loss': 0.24503971636295319, 'train/grad_norm_pre_clip': 101.7408447265625, 'timing/data': 0.0002108000044245273, 'timing/model': 0.36877644500054885, 'learning_rate/qwen_vl_interface': 7.86930175118603e-06, 'learning_rate/action_model': 7.86930175118603e-05, 'epoch': 0.12}) +[2026-10-03 14:56:33,045][starVLA.training.train_starvla][INFO] - Step 1640, Loss: {'train/loss': 0.018040956929326057, 'train/grad_norm_pre_clip': 13.034104347229004, 'timing/data': 0.00021706000552512705, 'timing/model': 0.3691945449972991, 'learning_rate/qwen_vl_interface': 7.866760661897487e-06, 'learning_rate/action_model': 7.866760661897487e-05, 'epoch': 0.12}) +[2026-10-03 14:56:33,897][starVLA.training.train_starvla][INFO] - Step 1641, Loss: {'train/loss': 0.0004607529263012111, 'train/grad_norm_pre_clip': 0.13571473956108093, 'timing/data': 0.0003540699981385842, 'timing/model': 0.3700863450067118, 'learning_rate/qwen_vl_interface': 7.86421849695748e-06, 'learning_rate/action_model': 7.864218496957481e-05, 'epoch': 0.12}) +[2026-10-03 14:56:34,553][starVLA.training.train_starvla][INFO] - Step 1642, Loss: {'train/loss': 0.4066237211227417, 'train/grad_norm_pre_clip': 34.29891586303711, 'timing/data': 0.00025705998996272683, 'timing/model': 0.3689277749945177, 'learning_rate/qwen_vl_interface': 7.861675257410998e-06, 'learning_rate/action_model': 7.861675257410997e-05, 'epoch': 0.12}) +[2026-10-03 14:56:35,209][starVLA.training.train_starvla][INFO] - Step 1643, Loss: {'train/loss': 0.0005015892093069851, 'train/grad_norm_pre_clip': 0.3569050431251526, 'timing/data': 0.00025306000316049904, 'timing/model': 0.36882811499526724, 'learning_rate/qwen_vl_interface': 7.859130944303468e-06, 'learning_rate/action_model': 7.859130944303467e-05, 'epoch': 0.12}) +[2026-10-03 14:56:35,866][starVLA.training.train_starvla][INFO] - Step 1644, Loss: {'train/loss': 0.5483667254447937, 'train/grad_norm_pre_clip': 34.86095428466797, 'timing/data': 0.0003667800046969205, 'timing/model': 0.3692244150006445, 'learning_rate/qwen_vl_interface': 7.856585558680761e-06, 'learning_rate/action_model': 7.856585558680761e-05, 'epoch': 0.12}) +[2026-10-03 14:56:36,525][starVLA.training.train_starvla][INFO] - Step 1645, Loss: {'train/loss': 0.023962466046214104, 'train/grad_norm_pre_clip': 37.12628936767578, 'timing/data': 0.00026142000569961965, 'timing/model': 0.3700844550039619, 'learning_rate/qwen_vl_interface': 7.854039101589191e-06, 'learning_rate/action_model': 7.854039101589191e-05, 'epoch': 0.12}) +[2026-10-03 14:56:37,380][starVLA.training.train_starvla][INFO] - Step 1646, Loss: {'train/loss': 0.0005526365712285042, 'train/grad_norm_pre_clip': 0.26603829860687256, 'timing/data': 0.00022231000184547156, 'timing/model': 0.36897371501254383, 'learning_rate/qwen_vl_interface': 7.851491574075509e-06, 'learning_rate/action_model': 7.851491574075509e-05, 'epoch': 0.12}) +[2026-10-03 14:56:38,037][starVLA.training.train_starvla][INFO] - Step 1647, Loss: {'train/loss': 0.4981091320514679, 'train/grad_norm_pre_clip': 27.169511795043945, 'timing/data': 0.00029506999999284744, 'timing/model': 0.36883559499983676, 'learning_rate/qwen_vl_interface': 7.848942977186909e-06, 'learning_rate/action_model': 7.848942977186908e-05, 'epoch': 0.12}) +[2026-10-03 14:56:38,693][starVLA.training.train_starvla][INFO] - Step 1648, Loss: {'train/loss': 0.17382241785526276, 'train/grad_norm_pre_clip': 78.86666870117188, 'timing/data': 0.0003342500131111592, 'timing/model': 0.3683179060026305, 'learning_rate/qwen_vl_interface': 7.84639331197102e-06, 'learning_rate/action_model': 7.84639331197102e-05, 'epoch': 0.12}) +[2026-10-03 14:56:39,327][starVLA.training.train_starvla][INFO] - Step 1649, Loss: {'train/loss': 0.106043741106987, 'train/grad_norm_pre_clip': 60.60501480102539, 'timing/data': 0.00030725900433026254, 'timing/model': 0.3697433650086168, 'learning_rate/qwen_vl_interface': 7.843842579475918e-06, 'learning_rate/action_model': 7.843842579475918e-05, 'epoch': 0.12}) +[2026-10-03 14:56:39,961][starVLA.training.train_starvla][INFO] - Step 1650, Loss: {'train/loss': 0.5438579320907593, 'train/grad_norm_pre_clip': 65.33546447753906, 'timing/data': 0.0003282999969087541, 'timing/model': 0.36826463499164674, 'learning_rate/qwen_vl_interface': 7.841290780750107e-06, 'learning_rate/action_model': 7.841290780750107e-05, 'epoch': 0.12}) +[2026-10-03 14:56:40,790][starVLA.training.train_starvla][INFO] - Step 1651, Loss: {'train/loss': 0.00048444076674059033, 'train/grad_norm_pre_clip': 0.0921417623758316, 'timing/data': 0.00032407000253442675, 'timing/model': 0.540482335985871, 'learning_rate/qwen_vl_interface': 7.838737916842538e-06, 'learning_rate/action_model': 7.838737916842538e-05, 'epoch': 0.12}) +[2026-10-03 14:56:41,447][starVLA.training.train_starvla][INFO] - Step 1652, Loss: {'train/loss': 0.0004686622123699635, 'train/grad_norm_pre_clip': 0.07719628512859344, 'timing/data': 0.0003325199941173196, 'timing/model': 0.36869454599218443, 'learning_rate/qwen_vl_interface': 7.836183988802596e-06, 'learning_rate/action_model': 7.836183988802595e-05, 'epoch': 0.12}) +[2026-10-03 14:56:42,105][starVLA.training.train_starvla][INFO] - Step 1653, Loss: {'train/loss': 0.0013010087423026562, 'train/grad_norm_pre_clip': 0.21880953013896942, 'timing/data': 0.0002639900048961863, 'timing/model': 0.37048770500405226, 'learning_rate/qwen_vl_interface': 7.833628997680107e-06, 'learning_rate/action_model': 7.833628997680107e-05, 'epoch': 0.12}) +[2026-10-03 14:56:42,762][starVLA.training.train_starvla][INFO] - Step 1654, Loss: {'train/loss': 0.012752695940434933, 'train/grad_norm_pre_clip': 8.03948974609375, 'timing/data': 0.0003122100024484098, 'timing/model': 0.36901221600419376, 'learning_rate/qwen_vl_interface': 7.831072944525329e-06, 'learning_rate/action_model': 7.831072944525328e-05, 'epoch': 0.12}) +[2026-10-03 14:56:43,417][starVLA.training.train_starvla][INFO] - Step 1655, Loss: {'train/loss': 0.3999709188938141, 'train/grad_norm_pre_clip': 64.13882446289062, 'timing/data': 0.0003364099975442514, 'timing/model': 0.3682672159920912, 'learning_rate/qwen_vl_interface': 7.828515830388956e-06, 'learning_rate/action_model': 7.828515830388956e-05, 'epoch': 0.12}) +[2026-10-03 14:56:44,275][starVLA.training.train_starvla][INFO] - Step 1656, Loss: {'train/loss': 0.7420964241027832, 'train/grad_norm_pre_clip': 212.81698608398438, 'timing/data': 0.00031085999216884375, 'timing/model': 0.5709933000034653, 'learning_rate/qwen_vl_interface': 7.825957656322129e-06, 'learning_rate/action_model': 7.825957656322128e-05, 'epoch': 0.12}) +[2026-10-03 14:56:44,933][starVLA.training.train_starvla][INFO] - Step 1657, Loss: {'train/loss': 0.0011248666560277343, 'train/grad_norm_pre_clip': 0.2519858777523041, 'timing/data': 0.00031447000219486654, 'timing/model': 0.3698591749998741, 'learning_rate/qwen_vl_interface': 7.823398423376406e-06, 'learning_rate/action_model': 7.823398423376406e-05, 'epoch': 0.12}) +[2026-10-03 14:56:45,588][starVLA.training.train_starvla][INFO] - Step 1658, Loss: {'train/loss': 0.09293978661298752, 'train/grad_norm_pre_clip': 75.52104187011719, 'timing/data': 0.00033580999297555536, 'timing/model': 0.3683282650017645, 'learning_rate/qwen_vl_interface': 7.820838132603801e-06, 'learning_rate/action_model': 7.8208381326038e-05, 'epoch': 0.12}) +[2026-10-03 14:56:46,247][starVLA.training.train_starvla][INFO] - Step 1659, Loss: {'train/loss': 0.27400681376457214, 'train/grad_norm_pre_clip': 84.64601135253906, 'timing/data': 0.00020907999714836478, 'timing/model': 0.369658655996318, 'learning_rate/qwen_vl_interface': 7.818276785056748e-06, 'learning_rate/action_model': 7.818276785056749e-05, 'epoch': 0.12}) +[2026-10-03 14:56:46,903][starVLA.training.train_starvla][INFO] - Step 1660, Loss: {'train/loss': 0.10855931043624878, 'train/grad_norm_pre_clip': 134.78038024902344, 'timing/data': 0.00021877000108361244, 'timing/model': 0.3688687850080896, 'learning_rate/qwen_vl_interface': 7.815714381788123e-06, 'learning_rate/action_model': 7.815714381788122e-05, 'epoch': 0.12}) +[2026-10-03 14:56:47,560][starVLA.training.train_starvla][INFO] - Step 1661, Loss: {'train/loss': 0.00038857580511830747, 'train/grad_norm_pre_clip': 0.04578687623143196, 'timing/data': 0.00026513000193517655, 'timing/model': 0.3704566249944037, 'learning_rate/qwen_vl_interface': 7.813150923851227e-06, 'learning_rate/action_model': 7.813150923851227e-05, 'epoch': 0.12}) +[2026-10-03 14:56:48,412][starVLA.training.train_starvla][INFO] - Step 1662, Loss: {'train/loss': 0.00104415079113096, 'train/grad_norm_pre_clip': 0.16173414885997772, 'timing/data': 0.00022443999478127807, 'timing/model': 0.36869010599912144, 'learning_rate/qwen_vl_interface': 7.810586412299807e-06, 'learning_rate/action_model': 7.810586412299806e-05, 'epoch': 0.12}) +[2026-10-03 14:56:49,069][starVLA.training.train_starvla][INFO] - Step 1663, Loss: {'train/loss': 0.34853583574295044, 'train/grad_norm_pre_clip': 14.047554016113281, 'timing/data': 0.0003094790008617565, 'timing/model': 0.36912145699898247, 'learning_rate/qwen_vl_interface': 7.808020848188034e-06, 'learning_rate/action_model': 7.808020848188034e-05, 'epoch': 0.12}) +[2026-10-03 14:56:49,726][starVLA.training.train_starvla][INFO] - Step 1664, Loss: {'train/loss': 0.3207090198993683, 'train/grad_norm_pre_clip': 27.061704635620117, 'timing/data': 0.00025426999491173774, 'timing/model': 0.36904584600415546, 'learning_rate/qwen_vl_interface': 7.805454232570517e-06, 'learning_rate/action_model': 7.805454232570517e-05, 'epoch': 0.12}) +[2026-10-03 14:56:50,384][starVLA.training.train_starvla][INFO] - Step 1665, Loss: {'train/loss': 0.39239799976348877, 'train/grad_norm_pre_clip': 32.4389762878418, 'timing/data': 0.00025783000455703586, 'timing/model': 0.3711464249936398, 'learning_rate/qwen_vl_interface': 7.80288656650229e-06, 'learning_rate/action_model': 7.80288656650229e-05, 'epoch': 0.12}) +[2026-10-03 14:56:51,041][starVLA.training.train_starvla][INFO] - Step 1666, Loss: {'train/loss': 0.22525858879089355, 'train/grad_norm_pre_clip': 106.05750274658203, 'timing/data': 0.00021718000061810017, 'timing/model': 0.3694780060031917, 'learning_rate/qwen_vl_interface': 7.800317851038829e-06, 'learning_rate/action_model': 7.800317851038828e-05, 'epoch': 0.12}) +[2026-10-03 14:56:51,905][starVLA.training.train_starvla][INFO] - Step 1667, Loss: {'train/loss': 0.011824777349829674, 'train/grad_norm_pre_clip': 2.8351569175720215, 'timing/data': 0.0003034800029126927, 'timing/model': 0.5742553399904864, 'learning_rate/qwen_vl_interface': 7.797748087236033e-06, 'learning_rate/action_model': 7.797748087236033e-05, 'epoch': 0.12}) +[2026-10-03 14:56:52,561][starVLA.training.train_starvla][INFO] - Step 1668, Loss: {'train/loss': 0.002031284850090742, 'train/grad_norm_pre_clip': 0.5180618762969971, 'timing/data': 0.0002562000008765608, 'timing/model': 0.3687657460104674, 'learning_rate/qwen_vl_interface': 7.795177276150234e-06, 'learning_rate/action_model': 7.795177276150234e-05, 'epoch': 0.12}) +[2026-10-03 14:56:53,219][starVLA.training.train_starvla][INFO] - Step 1669, Loss: {'train/loss': 0.30428144335746765, 'train/grad_norm_pre_clip': 14.982789039611816, 'timing/data': 0.00025081999774556607, 'timing/model': 0.3705430050031282, 'learning_rate/qwen_vl_interface': 7.792605418838196e-06, 'learning_rate/action_model': 7.792605418838197e-05, 'epoch': 0.12}) +[2026-10-03 14:56:53,876][starVLA.training.train_starvla][INFO] - Step 1670, Loss: {'train/loss': 0.0011745328083634377, 'train/grad_norm_pre_clip': 0.21072699129581451, 'timing/data': 0.0002519200061215088, 'timing/model': 0.36881454499962274, 'learning_rate/qwen_vl_interface': 7.790032516357116e-06, 'learning_rate/action_model': 7.790032516357117e-05, 'epoch': 0.12}) +[2026-10-03 14:56:54,533][starVLA.training.train_starvla][INFO] - Step 1671, Loss: {'train/loss': 0.0029290877282619476, 'train/grad_norm_pre_clip': 0.5040503740310669, 'timing/data': 0.00032079000084195286, 'timing/model': 0.3686652260075789, 'learning_rate/qwen_vl_interface': 7.787458569764613e-06, 'learning_rate/action_model': 7.787458569764612e-05, 'epoch': 0.12}) +[2026-10-03 14:56:55,166][starVLA.training.train_starvla][INFO] - Step 1672, Loss: {'train/loss': 0.4938782751560211, 'train/grad_norm_pre_clip': 20.984695434570312, 'timing/data': 0.0003255600022384897, 'timing/model': 0.34568382998986635, 'learning_rate/qwen_vl_interface': 7.784883580118738e-06, 'learning_rate/action_model': 7.784883580118738e-05, 'epoch': 0.12}) +[2026-10-03 14:56:56,001][starVLA.training.train_starvla][INFO] - Step 1673, Loss: {'train/loss': 0.020803475752472878, 'train/grad_norm_pre_clip': 23.382247924804688, 'timing/data': 0.0003077300061704591, 'timing/model': 0.348088919010479, 'learning_rate/qwen_vl_interface': 7.782307548477975e-06, 'learning_rate/action_model': 7.782307548477975e-05, 'epoch': 0.12}) +[2026-10-03 14:56:56,656][starVLA.training.train_starvla][INFO] - Step 1674, Loss: {'train/loss': 0.0013530404539778829, 'train/grad_norm_pre_clip': 0.2001495063304901, 'timing/data': 0.00020742999913636595, 'timing/model': 0.36860514599538874, 'learning_rate/qwen_vl_interface': 7.779730475901233e-06, 'learning_rate/action_model': 7.779730475901232e-05, 'epoch': 0.12}) +[2026-10-03 14:56:57,312][starVLA.training.train_starvla][INFO] - Step 1675, Loss: {'train/loss': 0.5860545039176941, 'train/grad_norm_pre_clip': 38.4520263671875, 'timing/data': 0.0003197600017301738, 'timing/model': 0.36849676600832026, 'learning_rate/qwen_vl_interface': 7.777152363447849e-06, 'learning_rate/action_model': 7.77715236344785e-05, 'epoch': 0.12}) +[2026-10-03 14:56:57,968][starVLA.training.train_starvla][INFO] - Step 1676, Loss: {'train/loss': 0.004442611709237099, 'train/grad_norm_pre_clip': 1.5689252614974976, 'timing/data': 0.00031459999445360154, 'timing/model': 0.3689542260108283, 'learning_rate/qwen_vl_interface': 7.774573212177587e-06, 'learning_rate/action_model': 7.774573212177588e-05, 'epoch': 0.12}) +[2026-10-03 14:56:58,626][starVLA.training.train_starvla][INFO] - Step 1677, Loss: {'train/loss': 0.028565097600221634, 'train/grad_norm_pre_clip': 9.079319953918457, 'timing/data': 0.000298379993182607, 'timing/model': 0.37012040500121657, 'learning_rate/qwen_vl_interface': 7.77199302315064e-06, 'learning_rate/action_model': 7.77199302315064e-05, 'epoch': 0.12}) +[2026-10-03 14:56:59,474][starVLA.training.train_starvla][INFO] - Step 1678, Loss: {'train/loss': 0.06908364593982697, 'train/grad_norm_pre_clip': 32.75307846069336, 'timing/data': 0.0002564200112828985, 'timing/model': 0.5614368929964257, 'learning_rate/qwen_vl_interface': 7.769411797427622e-06, 'learning_rate/action_model': 7.769411797427623e-05, 'epoch': 0.12}) +[2026-10-03 14:57:00,130][starVLA.training.train_starvla][INFO] - Step 1679, Loss: {'train/loss': 0.05417802184820175, 'train/grad_norm_pre_clip': 86.88734436035156, 'timing/data': 0.00024580000899732113, 'timing/model': 0.3685195469879545, 'learning_rate/qwen_vl_interface': 7.766829536069584e-06, 'learning_rate/action_model': 7.766829536069584e-05, 'epoch': 0.12}) +[2026-10-03 14:57:00,786][starVLA.training.train_starvla][INFO] - Step 1680, Loss: {'train/loss': 0.002986969891935587, 'train/grad_norm_pre_clip': 1.2406874895095825, 'timing/data': 0.00031478000164497644, 'timing/model': 0.36838491600065026, 'learning_rate/qwen_vl_interface': 7.764246240137991e-06, 'learning_rate/action_model': 7.764246240137991e-05, 'epoch': 0.12}) +[2026-10-03 14:57:01,444][starVLA.training.train_starvla][INFO] - Step 1681, Loss: {'train/loss': 0.0046348185278475285, 'train/grad_norm_pre_clip': 5.2498884201049805, 'timing/data': 0.0003067599900532514, 'timing/model': 0.3700154760008445, 'learning_rate/qwen_vl_interface': 7.761661910694741e-06, 'learning_rate/action_model': 7.761661910694741e-05, 'epoch': 0.12}) +[2026-10-03 14:57:02,100][starVLA.training.train_starvla][INFO] - Step 1682, Loss: {'train/loss': 0.28943467140197754, 'train/grad_norm_pre_clip': 26.375661849975586, 'timing/data': 0.00024213000142481178, 'timing/model': 0.36893032600346487, 'learning_rate/qwen_vl_interface': 7.759076548802153e-06, 'learning_rate/action_model': 7.759076548802152e-05, 'epoch': 0.12}) +[2026-10-03 14:57:02,755][starVLA.training.train_starvla][INFO] - Step 1683, Loss: {'train/loss': 0.0036561163142323494, 'train/grad_norm_pre_clip': 2.1173653602600098, 'timing/data': 0.0002502999996067956, 'timing/model': 0.36804551699606236, 'learning_rate/qwen_vl_interface': 7.75649015552297e-06, 'learning_rate/action_model': 7.75649015552297e-05, 'epoch': 0.12}) +[2026-10-03 14:57:03,604][starVLA.training.train_starvla][INFO] - Step 1684, Loss: {'train/loss': 0.0059809437952935696, 'train/grad_norm_pre_clip': 3.3260610103607178, 'timing/data': 0.0002611899981275201, 'timing/model': 0.36827140700188465, 'learning_rate/qwen_vl_interface': 7.753902731920364e-06, 'learning_rate/action_model': 7.753902731920364e-05, 'epoch': 0.12}) +[2026-10-03 14:57:04,260][starVLA.training.train_starvla][INFO] - Step 1685, Loss: {'train/loss': 0.0006734454655088484, 'train/grad_norm_pre_clip': 0.9826749563217163, 'timing/data': 0.00030976999551057816, 'timing/model': 0.36948265600949526, 'learning_rate/qwen_vl_interface': 7.751314279057925e-06, 'learning_rate/action_model': 7.751314279057923e-05, 'epoch': 0.12}) +[2026-10-03 14:57:04,916][starVLA.training.train_starvla][INFO] - Step 1686, Loss: {'train/loss': 0.0001421256165485829, 'train/grad_norm_pre_clip': 0.023694917559623718, 'timing/data': 0.0002443900011712685, 'timing/model': 0.3685259460035013, 'learning_rate/qwen_vl_interface': 7.74872479799967e-06, 'learning_rate/action_model': 7.74872479799967e-05, 'epoch': 0.12}) +[2026-10-03 14:57:05,573][starVLA.training.train_starvla][INFO] - Step 1687, Loss: {'train/loss': 0.00013870783732272685, 'train/grad_norm_pre_clip': 0.018263043835759163, 'timing/data': 0.00026713000261224806, 'timing/model': 0.3691014060023008, 'learning_rate/qwen_vl_interface': 7.746134289810034e-06, 'learning_rate/action_model': 7.746134289810033e-05, 'epoch': 0.12}) +[2026-10-03 14:57:06,229][starVLA.training.train_starvla][INFO] - Step 1688, Loss: {'train/loss': 9.811304335016757e-05, 'train/grad_norm_pre_clip': 0.0150356600061059, 'timing/data': 0.00027534000400919467, 'timing/model': 0.3688901760033332, 'learning_rate/qwen_vl_interface': 7.74354275555388e-06, 'learning_rate/action_model': 7.74354275555388e-05, 'epoch': 0.12}) +[2026-10-03 14:57:07,084][starVLA.training.train_starvla][INFO] - Step 1689, Loss: {'train/loss': 0.47530099749565125, 'train/grad_norm_pre_clip': 31.085777282714844, 'timing/data': 0.00028095999732613564, 'timing/model': 0.3715435560006881, 'learning_rate/qwen_vl_interface': 7.740950196296488e-06, 'learning_rate/action_model': 7.740950196296488e-05, 'epoch': 0.12}) +[2026-10-03 14:57:07,744][starVLA.training.train_starvla][INFO] - Step 1690, Loss: {'train/loss': 0.3212450444698334, 'train/grad_norm_pre_clip': 89.37864685058594, 'timing/data': 0.00026189000345766544, 'timing/model': 0.3703997859993251, 'learning_rate/qwen_vl_interface': 7.738356613103562e-06, 'learning_rate/action_model': 7.73835661310356e-05, 'epoch': 0.12}) +[2026-10-03 14:57:08,404][starVLA.training.train_starvla][INFO] - Step 1691, Loss: {'train/loss': 0.29992592334747314, 'train/grad_norm_pre_clip': 106.68224334716797, 'timing/data': 0.00022672000341117382, 'timing/model': 0.3708366159989964, 'learning_rate/qwen_vl_interface': 7.735762007041226e-06, 'learning_rate/action_model': 7.735762007041226e-05, 'epoch': 0.12}) +[2026-10-03 14:57:09,041][starVLA.training.train_starvla][INFO] - Step 1692, Loss: {'train/loss': 0.4345647096633911, 'train/grad_norm_pre_clip': 49.96231460571289, 'timing/data': 0.00023131999478209764, 'timing/model': 0.347604840004351, 'learning_rate/qwen_vl_interface': 7.733166379176023e-06, 'learning_rate/action_model': 7.733166379176023e-05, 'epoch': 0.12}) +[2026-10-03 14:57:09,702][starVLA.training.train_starvla][INFO] - Step 1693, Loss: {'train/loss': 0.10055595636367798, 'train/grad_norm_pre_clip': 195.1622772216797, 'timing/data': 0.00035182999272365123, 'timing/model': 0.37243499601026997, 'learning_rate/qwen_vl_interface': 7.730569730574921e-06, 'learning_rate/action_model': 7.73056973057492e-05, 'epoch': 0.12}) +[2026-10-03 14:57:10,561][starVLA.training.train_starvla][INFO] - Step 1694, Loss: {'train/loss': 0.0001513971947133541, 'train/grad_norm_pre_clip': 0.03241918608546257, 'timing/data': 0.00022731999342795461, 'timing/model': 0.37122606599587016, 'learning_rate/qwen_vl_interface': 7.727972062305299e-06, 'learning_rate/action_model': 7.727972062305299e-05, 'epoch': 0.12}) +[2026-10-03 14:57:11,199][starVLA.training.train_starvla][INFO] - Step 1695, Loss: {'train/loss': 8.214701665565372e-05, 'train/grad_norm_pre_clip': 0.009870501235127449, 'timing/data': 0.00021085000480525196, 'timing/model': 0.3483947200002149, 'learning_rate/qwen_vl_interface': 7.725373375434965e-06, 'learning_rate/action_model': 7.725373375434965e-05, 'epoch': 0.12}) +[2026-10-03 14:57:11,859][starVLA.training.train_starvla][INFO] - Step 1696, Loss: {'train/loss': 0.00017123273573815823, 'train/grad_norm_pre_clip': 0.06387580931186676, 'timing/data': 0.00021295998885761946, 'timing/model': 0.37105427599453833, 'learning_rate/qwen_vl_interface': 7.722773671032141e-06, 'learning_rate/action_model': 7.722773671032142e-05, 'epoch': 0.12}) +[2026-10-03 14:57:12,496][starVLA.training.train_starvla][INFO] - Step 1697, Loss: {'train/loss': 0.00020005965780001134, 'train/grad_norm_pre_clip': 0.05501818284392357, 'timing/data': 0.00021610999829135835, 'timing/model': 0.37117803598812316, 'learning_rate/qwen_vl_interface': 7.720172950165466e-06, 'learning_rate/action_model': 7.720172950165465e-05, 'epoch': 0.12}) +[2026-10-03 14:57:13,130][starVLA.training.train_starvla][INFO] - Step 1698, Loss: {'train/loss': 8.97627032827586e-05, 'train/grad_norm_pre_clip': 0.01907602697610855, 'timing/data': 0.00020889000734314322, 'timing/model': 0.3459971400006907, 'learning_rate/qwen_vl_interface': 7.717571213903996e-06, 'learning_rate/action_model': 7.717571213903996e-05, 'epoch': 0.12}) +[2026-10-03 14:57:13,961][starVLA.training.train_starvla][INFO] - Step 1699, Loss: {'train/loss': 0.0027257513720542192, 'train/grad_norm_pre_clip': 0.7432913184165955, 'timing/data': 0.0003130900004180148, 'timing/model': 0.5424647369945887, 'learning_rate/qwen_vl_interface': 7.714968463317207e-06, 'learning_rate/action_model': 7.714968463317207e-05, 'epoch': 0.12}) +[2026-10-03 14:57:14,619][starVLA.training.train_starvla][INFO] - Step 1700, Loss: {'train/loss': 0.5775768756866455, 'train/grad_norm_pre_clip': 31.9842472076416, 'timing/data': 0.0003828300104942173, 'timing/model': 0.3697567060007714, 'learning_rate/qwen_vl_interface': 7.712364699474996e-06, 'learning_rate/action_model': 7.712364699474996e-05, 'epoch': 0.12}) +[2026-10-03 14:57:15,278][starVLA.training.train_starvla][INFO] - Step 1701, Loss: {'train/loss': 0.0005309795960783958, 'train/grad_norm_pre_clip': 0.3008453845977783, 'timing/data': 0.00031399000727105886, 'timing/model': 0.3707640060019912, 'learning_rate/qwen_vl_interface': 7.709759923447667e-06, 'learning_rate/action_model': 7.709759923447665e-05, 'epoch': 0.12}) +[2026-10-03 14:57:15,916][starVLA.training.train_starvla][INFO] - Step 1702, Loss: {'train/loss': 0.0020792055875062943, 'train/grad_norm_pre_clip': 1.470956563949585, 'timing/data': 0.0002233300037914887, 'timing/model': 0.34779872999934014, 'learning_rate/qwen_vl_interface': 7.707154136305945e-06, 'learning_rate/action_model': 7.707154136305945e-05, 'epoch': 0.12}) +[2026-10-03 14:57:16,576][starVLA.training.train_starvla][INFO] - Step 1703, Loss: {'train/loss': 0.12097828835248947, 'train/grad_norm_pre_clip': 324.86077880859375, 'timing/data': 0.00030683999648317695, 'timing/model': 0.3708274469972821, 'learning_rate/qwen_vl_interface': 7.704547339120974e-06, 'learning_rate/action_model': 7.704547339120974e-05, 'epoch': 0.12}) +[2026-10-03 14:57:17,212][starVLA.training.train_starvla][INFO] - Step 1704, Loss: {'train/loss': 0.05654875189065933, 'train/grad_norm_pre_clip': 155.0659637451172, 'timing/data': 0.0002924100117525086, 'timing/model': 0.37004271599289495, 'learning_rate/qwen_vl_interface': 7.70193953296431e-06, 'learning_rate/action_model': 7.70193953296431e-05, 'epoch': 0.12}) +[2026-10-03 14:57:18,071][starVLA.training.train_starvla][INFO] - Step 1705, Loss: {'train/loss': 5.559972851187922e-05, 'train/grad_norm_pre_clip': 0.009377610869705677, 'timing/data': 0.00034910900285467505, 'timing/model': 0.37107004699646495, 'learning_rate/qwen_vl_interface': 7.69933071890792e-06, 'learning_rate/action_model': 7.699330718907919e-05, 'epoch': 0.12}) +[2026-10-03 14:57:18,728][starVLA.training.train_starvla][INFO] - Step 1706, Loss: {'train/loss': 0.3076256811618805, 'train/grad_norm_pre_clip': 53.63338851928711, 'timing/data': 0.00027956000121776015, 'timing/model': 0.368682906992035, 'learning_rate/qwen_vl_interface': 7.696720898024188e-06, 'learning_rate/action_model': 7.69672089802419e-05, 'epoch': 0.12}) +[2026-10-03 14:57:19,385][starVLA.training.train_starvla][INFO] - Step 1707, Loss: {'train/loss': 0.009124692529439926, 'train/grad_norm_pre_clip': 65.87836456298828, 'timing/data': 0.0002104599989252165, 'timing/model': 0.3691267670074012, 'learning_rate/qwen_vl_interface': 7.69411007138592e-06, 'learning_rate/action_model': 7.694110071385919e-05, 'epoch': 0.12}) +[2026-10-03 14:57:20,042][starVLA.training.train_starvla][INFO] - Step 1708, Loss: {'train/loss': 0.4044124484062195, 'train/grad_norm_pre_clip': 69.11973571777344, 'timing/data': 0.00026351999258622527, 'timing/model': 0.3699221669958206, 'learning_rate/qwen_vl_interface': 7.691498240066322e-06, 'learning_rate/action_model': 7.691498240066322e-05, 'epoch': 0.13}) +[2026-10-03 14:57:20,700][starVLA.training.train_starvla][INFO] - Step 1709, Loss: {'train/loss': 0.00014691510295961052, 'train/grad_norm_pre_clip': 0.035367295145988464, 'timing/data': 0.00025582000671420246, 'timing/model': 0.37045237600978, 'learning_rate/qwen_vl_interface': 7.688885405139022e-06, 'learning_rate/action_model': 7.688885405139021e-05, 'epoch': 0.13}) +[2026-10-03 14:57:21,556][starVLA.training.train_starvla][INFO] - Step 1710, Loss: {'train/loss': 0.0021251975558698177, 'train/grad_norm_pre_clip': 3.094860076904297, 'timing/data': 0.00033427999005652964, 'timing/model': 0.5666743430047063, 'learning_rate/qwen_vl_interface': 7.686271567678056e-06, 'learning_rate/action_model': 7.686271567678054e-05, 'epoch': 0.13}) +[2026-10-03 14:57:22,214][starVLA.training.train_starvla][INFO] - Step 1711, Loss: {'train/loss': 0.0005963257281109691, 'train/grad_norm_pre_clip': 0.29736781120300293, 'timing/data': 0.00024331000167876482, 'timing/model': 0.3698682260001078, 'learning_rate/qwen_vl_interface': 7.683656728757873e-06, 'learning_rate/action_model': 7.683656728757872e-05, 'epoch': 0.13}) +[2026-10-03 14:57:22,872][starVLA.training.train_starvla][INFO] - Step 1712, Loss: {'train/loss': 0.10915791988372803, 'train/grad_norm_pre_clip': 104.30648803710938, 'timing/data': 0.0002898899983847514, 'timing/model': 0.3697982159937965, 'learning_rate/qwen_vl_interface': 7.681040889453337e-06, 'learning_rate/action_model': 7.681040889453336e-05, 'epoch': 0.13}) +[2026-10-03 14:57:23,507][starVLA.training.train_starvla][INFO] - Step 1713, Loss: {'train/loss': 0.00015610537957400084, 'train/grad_norm_pre_clip': 0.022264912724494934, 'timing/data': 0.00039260000630747527, 'timing/model': 0.37042058700171765, 'learning_rate/qwen_vl_interface': 7.678424050839722e-06, 'learning_rate/action_model': 7.67842405083972e-05, 'epoch': 0.13}) +[2026-10-03 14:57:24,165][starVLA.training.train_starvla][INFO] - Step 1714, Loss: {'train/loss': 0.12521688640117645, 'train/grad_norm_pre_clip': 144.0159149169922, 'timing/data': 0.000375259987777099, 'timing/model': 0.37037440699350554, 'learning_rate/qwen_vl_interface': 7.675806213992706e-06, 'learning_rate/action_model': 7.675806213992707e-05, 'epoch': 0.13}) +[2026-10-03 14:57:24,824][starVLA.training.train_starvla][INFO] - Step 1715, Loss: {'train/loss': 0.0001604545395821333, 'train/grad_norm_pre_clip': 0.06629344075918198, 'timing/data': 0.0002190100058214739, 'timing/model': 0.3697700159973465, 'learning_rate/qwen_vl_interface': 7.673187379988388e-06, 'learning_rate/action_model': 7.673187379988388e-05, 'epoch': 0.13}) +[2026-10-03 14:57:25,678][starVLA.training.train_starvla][INFO] - Step 1716, Loss: {'train/loss': 0.00012031344522256404, 'train/grad_norm_pre_clip': 0.04060424119234085, 'timing/data': 0.00022597999486606568, 'timing/model': 0.36997725699620787, 'learning_rate/qwen_vl_interface': 7.67056754990327e-06, 'learning_rate/action_model': 7.670567549903269e-05, 'epoch': 0.13}) +[2026-10-03 14:57:26,337][starVLA.training.train_starvla][INFO] - Step 1717, Loss: {'train/loss': 0.5260235667228699, 'train/grad_norm_pre_clip': 39.319580078125, 'timing/data': 0.00021236999600660056, 'timing/model': 0.37114302700501867, 'learning_rate/qwen_vl_interface': 7.667946724814264e-06, 'learning_rate/action_model': 7.667946724814265e-05, 'epoch': 0.13}) +[2026-10-03 14:57:26,995][starVLA.training.train_starvla][INFO] - Step 1718, Loss: {'train/loss': 0.491772323846817, 'train/grad_norm_pre_clip': 24.527883529663086, 'timing/data': 0.00022606000129599124, 'timing/model': 0.37007396599801723, 'learning_rate/qwen_vl_interface': 7.665324905798694e-06, 'learning_rate/action_model': 7.665324905798695e-05, 'epoch': 0.13}) +[2026-10-03 14:57:27,631][starVLA.training.train_starvla][INFO] - Step 1719, Loss: {'train/loss': 0.00017656164709478617, 'train/grad_norm_pre_clip': 0.04444381967186928, 'timing/data': 0.000215580002986826, 'timing/model': 0.34644400999241043, 'learning_rate/qwen_vl_interface': 7.66270209393429e-06, 'learning_rate/action_model': 7.66270209393429e-05, 'epoch': 0.13}) +[2026-10-03 14:57:28,290][starVLA.training.train_starvla][INFO] - Step 1720, Loss: {'train/loss': 0.000635225442238152, 'train/grad_norm_pre_clip': 0.5220257639884949, 'timing/data': 0.0002889400057028979, 'timing/model': 0.3699415259907255, 'learning_rate/qwen_vl_interface': 7.66007829029919e-06, 'learning_rate/action_model': 7.660078290299191e-05, 'epoch': 0.13}) +[2026-10-03 14:57:29,145][starVLA.training.train_starvla][INFO] - Step 1721, Loss: {'train/loss': 0.0004189118044450879, 'train/grad_norm_pre_clip': 0.5521930456161499, 'timing/data': 0.00037116999737918377, 'timing/model': 0.37063588699675165, 'learning_rate/qwen_vl_interface': 7.657453495971942e-06, 'learning_rate/action_model': 7.657453495971943e-05, 'epoch': 0.13}) +[2026-10-03 14:57:29,803][starVLA.training.train_starvla][INFO] - Step 1722, Loss: {'train/loss': 0.0003139124601148069, 'train/grad_norm_pre_clip': 0.09512551128864288, 'timing/data': 0.00021704001119360328, 'timing/model': 0.3701527770026587, 'learning_rate/qwen_vl_interface': 7.6548277120315e-06, 'learning_rate/action_model': 7.6548277120315e-05, 'epoch': 0.13}) +[2026-10-03 14:57:30,438][starVLA.training.train_starvla][INFO] - Step 1723, Loss: {'train/loss': 0.00020992668578401208, 'train/grad_norm_pre_clip': 0.03385334089398384, 'timing/data': 0.00021284000831656158, 'timing/model': 0.3695278170052916, 'learning_rate/qwen_vl_interface': 7.652200939557225e-06, 'learning_rate/action_model': 7.652200939557225e-05, 'epoch': 0.13}) +[2026-10-03 14:57:31,096][starVLA.training.train_starvla][INFO] - Step 1724, Loss: {'train/loss': 0.041849534958601, 'train/grad_norm_pre_clip': 126.45799255371094, 'timing/data': 0.0002198799920734018, 'timing/model': 0.3696861570060719, 'learning_rate/qwen_vl_interface': 7.649573179628881e-06, 'learning_rate/action_model': 7.649573179628881e-05, 'epoch': 0.13}) +[2026-10-03 14:57:31,755][starVLA.training.train_starvla][INFO] - Step 1725, Loss: {'train/loss': 0.00029545629513449967, 'train/grad_norm_pre_clip': 0.12479496002197266, 'timing/data': 0.0002632000105222687, 'timing/model': 0.3707945960049983, 'learning_rate/qwen_vl_interface': 7.646944433326645e-06, 'learning_rate/action_model': 7.646944433326643e-05, 'epoch': 0.13}) +[2026-10-03 14:57:32,617][starVLA.training.train_starvla][INFO] - Step 1726, Loss: {'train/loss': 0.3566342890262604, 'train/grad_norm_pre_clip': 76.80828094482422, 'timing/data': 0.00021112000104039907, 'timing/model': 0.572283432004042, 'learning_rate/qwen_vl_interface': 7.64431470173109e-06, 'learning_rate/action_model': 7.64431470173109e-05, 'epoch': 0.13}) +[2026-10-03 14:57:33,275][starVLA.training.train_starvla][INFO] - Step 1727, Loss: {'train/loss': 0.01806424930691719, 'train/grad_norm_pre_clip': 42.375999450683594, 'timing/data': 0.00035057000059168786, 'timing/model': 0.36942507699131966, 'learning_rate/qwen_vl_interface': 7.641683985923203e-06, 'learning_rate/action_model': 7.641683985923203e-05, 'epoch': 0.13}) +[2026-10-03 14:57:33,933][starVLA.training.train_starvla][INFO] - Step 1728, Loss: {'train/loss': 0.0003905220946762711, 'train/grad_norm_pre_clip': 0.07278568297624588, 'timing/data': 0.0003120299952570349, 'timing/model': 0.3695987169921864, 'learning_rate/qwen_vl_interface': 7.63905228698437e-06, 'learning_rate/action_model': 7.63905228698437e-05, 'epoch': 0.13}) +[2026-10-03 14:57:34,592][starVLA.training.train_starvla][INFO] - Step 1729, Loss: {'train/loss': 0.2833191156387329, 'train/grad_norm_pre_clip': 43.98036575317383, 'timing/data': 0.00026437001361045986, 'timing/model': 0.37162027599697467, 'learning_rate/qwen_vl_interface': 7.636419605996385e-06, 'learning_rate/action_model': 7.636419605996384e-05, 'epoch': 0.13}) +[2026-10-03 14:57:35,250][starVLA.training.train_starvla][INFO] - Step 1730, Loss: {'train/loss': 0.00040276808431372046, 'train/grad_norm_pre_clip': 0.09157387167215347, 'timing/data': 0.0002146299957530573, 'timing/model': 0.3691263470100239, 'learning_rate/qwen_vl_interface': 7.633785944041444e-06, 'learning_rate/action_model': 7.633785944041444e-05, 'epoch': 0.13}) +[2026-10-03 14:57:36,103][starVLA.training.train_starvla][INFO] - Step 1731, Loss: {'train/loss': 0.2877954840660095, 'train/grad_norm_pre_clip': 60.31313705444336, 'timing/data': 0.00039633999404031783, 'timing/model': 0.5642767440003809, 'learning_rate/qwen_vl_interface': 7.631151302202145e-06, 'learning_rate/action_model': 7.631151302202144e-05, 'epoch': 0.13}) +[2026-10-03 14:57:36,761][starVLA.training.train_starvla][INFO] - Step 1732, Loss: {'train/loss': 0.00016832217806950212, 'train/grad_norm_pre_clip': 0.03627578169107437, 'timing/data': 0.0002095399977406487, 'timing/model': 0.36987259599845856, 'learning_rate/qwen_vl_interface': 7.628515681561488e-06, 'learning_rate/action_model': 7.628515681561487e-05, 'epoch': 0.13}) +[2026-10-03 14:57:37,420][starVLA.training.train_starvla][INFO] - Step 1733, Loss: {'train/loss': 0.0017991343047469854, 'train/grad_norm_pre_clip': 0.6065610647201538, 'timing/data': 0.0002785599936032668, 'timing/model': 0.3707999760081293, 'learning_rate/qwen_vl_interface': 7.62587908320288e-06, 'learning_rate/action_model': 7.62587908320288e-05, 'epoch': 0.13}) +[2026-10-03 14:57:38,080][starVLA.training.train_starvla][INFO] - Step 1734, Loss: {'train/loss': 0.0006044590845704079, 'train/grad_norm_pre_clip': 0.12522920966148376, 'timing/data': 0.0003159000043524429, 'timing/model': 0.36980506699183024, 'learning_rate/qwen_vl_interface': 7.6232415082101255e-06, 'learning_rate/action_model': 7.623241508210124e-05, 'epoch': 0.13}) +[2026-10-03 14:57:38,739][starVLA.training.train_starvla][INFO] - Step 1735, Loss: {'train/loss': 0.03161546587944031, 'train/grad_norm_pre_clip': 94.18206024169922, 'timing/data': 0.0002862799883587286, 'timing/model': 0.37000298699422274, 'learning_rate/qwen_vl_interface': 7.620602957667432e-06, 'learning_rate/action_model': 7.620602957667432e-05, 'epoch': 0.13}) +[2026-10-03 14:57:39,397][starVLA.training.train_starvla][INFO] - Step 1736, Loss: {'train/loss': 0.012472804635763168, 'train/grad_norm_pre_clip': 5.653670310974121, 'timing/data': 0.00033192998671438545, 'timing/model': 0.36954113699903246, 'learning_rate/qwen_vl_interface': 7.617963432659408e-06, 'learning_rate/action_model': 7.617963432659408e-05, 'epoch': 0.13}) +[2026-10-03 14:57:40,261][starVLA.training.train_starvla][INFO] - Step 1737, Loss: {'train/loss': 0.09271395951509476, 'train/grad_norm_pre_clip': 23.4355411529541, 'timing/data': 0.00022873000125400722, 'timing/model': 0.3712791069992818, 'learning_rate/qwen_vl_interface': 7.615322934271063e-06, 'learning_rate/action_model': 7.615322934271064e-05, 'epoch': 0.13}) +[2026-10-03 14:57:40,919][starVLA.training.train_starvla][INFO] - Step 1738, Loss: {'train/loss': 0.27023234963417053, 'train/grad_norm_pre_clip': 99.20374298095703, 'timing/data': 0.00021652999566867948, 'timing/model': 0.3699482869997155, 'learning_rate/qwen_vl_interface': 7.6126814635878055e-06, 'learning_rate/action_model': 7.612681463587805e-05, 'epoch': 0.13}) +[2026-10-03 14:57:41,577][starVLA.training.train_starvla][INFO] - Step 1739, Loss: {'train/loss': 0.011855488643050194, 'train/grad_norm_pre_clip': 5.9225945472717285, 'timing/data': 0.0002111000067088753, 'timing/model': 0.3704050969972741, 'learning_rate/qwen_vl_interface': 7.6100390216954445e-06, 'learning_rate/action_model': 7.610039021695445e-05, 'epoch': 0.13}) +[2026-10-03 14:57:42,235][starVLA.training.train_starvla][INFO] - Step 1740, Loss: {'train/loss': 0.00017481982649769634, 'train/grad_norm_pre_clip': 0.04912574961781502, 'timing/data': 0.0002906900044763461, 'timing/model': 0.36940706700261217, 'learning_rate/qwen_vl_interface': 7.607395609680189e-06, 'learning_rate/action_model': 7.60739560968019e-05, 'epoch': 0.13}) +[2026-10-03 14:57:42,896][starVLA.training.train_starvla][INFO] - Step 1741, Loss: {'train/loss': 0.0030631585977971554, 'train/grad_norm_pre_clip': 1.5760993957519531, 'timing/data': 0.0003239700017729774, 'timing/model': 0.37145836699346546, 'learning_rate/qwen_vl_interface': 7.604751228628644e-06, 'learning_rate/action_model': 7.604751228628644e-05, 'epoch': 0.13}) +[2026-10-03 14:57:43,751][starVLA.training.train_starvla][INFO] - Step 1742, Loss: {'train/loss': 0.35151493549346924, 'train/grad_norm_pre_clip': 33.637447357177734, 'timing/data': 0.00023936999787110835, 'timing/model': 0.3693620970007032, 'learning_rate/qwen_vl_interface': 7.602105879627814e-06, 'learning_rate/action_model': 7.602105879627814e-05, 'epoch': 0.13}) +[2026-10-03 14:57:44,408][starVLA.training.train_starvla][INFO] - Step 1743, Loss: {'train/loss': 0.0003507972869556397, 'train/grad_norm_pre_clip': 0.23403601348400116, 'timing/data': 0.00021733000176027417, 'timing/model': 0.36896772700129077, 'learning_rate/qwen_vl_interface': 7.599459563765107e-06, 'learning_rate/action_model': 7.599459563765106e-05, 'epoch': 0.13}) +[2026-10-03 14:57:45,064][starVLA.training.train_starvla][INFO] - Step 1744, Loss: {'train/loss': 0.030505729839205742, 'train/grad_norm_pre_clip': 60.13220977783203, 'timing/data': 0.00021297999774105847, 'timing/model': 0.3687951379979495, 'learning_rate/qwen_vl_interface': 7.596812282128319e-06, 'learning_rate/action_model': 7.596812282128319e-05, 'epoch': 0.13}) +[2026-10-03 14:57:45,723][starVLA.training.train_starvla][INFO] - Step 1745, Loss: {'train/loss': 0.00018778929370455444, 'train/grad_norm_pre_clip': 0.04369721561670303, 'timing/data': 0.0002608700015116483, 'timing/model': 0.3710408769984497, 'learning_rate/qwen_vl_interface': 7.594164035805648e-06, 'learning_rate/action_model': 7.594164035805648e-05, 'epoch': 0.13}) +[2026-10-03 14:57:46,380][starVLA.training.train_starvla][INFO] - Step 1746, Loss: {'train/loss': 0.0007586335996165872, 'train/grad_norm_pre_clip': 0.5281668901443481, 'timing/data': 0.00031478000164497644, 'timing/model': 0.3695300670078723, 'learning_rate/qwen_vl_interface': 7.59151482588569e-06, 'learning_rate/action_model': 7.59151482588569e-05, 'epoch': 0.13}) +[2026-10-03 14:57:47,236][starVLA.training.train_starvla][INFO] - Step 1747, Loss: {'train/loss': 0.03198585659265518, 'train/grad_norm_pre_clip': 208.2840576171875, 'timing/data': 0.00021761999232694507, 'timing/model': 0.5672044629900483, 'learning_rate/qwen_vl_interface': 7.588864653457433e-06, 'learning_rate/action_model': 7.588864653457432e-05, 'epoch': 0.13}) +[2026-10-03 14:57:47,892][starVLA.training.train_starvla][INFO] - Step 1748, Loss: {'train/loss': 0.00024482287699356675, 'train/grad_norm_pre_clip': 0.06832189112901688, 'timing/data': 0.00021769000159110874, 'timing/model': 0.36926071700872853, 'learning_rate/qwen_vl_interface': 7.5862135196102634e-06, 'learning_rate/action_model': 7.586213519610263e-05, 'epoch': 0.13}) +[2026-10-03 14:57:48,552][starVLA.training.train_starvla][INFO] - Step 1749, Loss: {'train/loss': 0.5193050503730774, 'train/grad_norm_pre_clip': 66.85057830810547, 'timing/data': 0.00035304999619256705, 'timing/model': 0.3714961069927085, 'learning_rate/qwen_vl_interface': 7.583561425433962e-06, 'learning_rate/action_model': 7.583561425433962e-05, 'epoch': 0.13}) +[2026-10-03 14:57:49,210][starVLA.training.train_starvla][INFO] - Step 1750, Loss: {'train/loss': 0.4655241072177887, 'train/grad_norm_pre_clip': 29.124271392822266, 'timing/data': 0.00021036999532952905, 'timing/model': 0.36972131799848285, 'learning_rate/qwen_vl_interface': 7.580908372018706e-06, 'learning_rate/action_model': 7.580908372018705e-05, 'epoch': 0.13}) +[2026-10-03 14:57:49,869][starVLA.training.train_starvla][INFO] - Step 1751, Loss: {'train/loss': 0.278347373008728, 'train/grad_norm_pre_clip': 88.88310241699219, 'timing/data': 0.0002524699957575649, 'timing/model': 0.3705725070030894, 'learning_rate/qwen_vl_interface': 7.5782543604550655e-06, 'learning_rate/action_model': 7.578254360455064e-05, 'epoch': 0.13}) +[2026-10-03 14:57:50,726][starVLA.training.train_starvla][INFO] - Step 1752, Loss: {'train/loss': 0.5895708203315735, 'train/grad_norm_pre_clip': 116.73615264892578, 'timing/data': 0.00021315999038051814, 'timing/model': 0.5695385230064858, 'learning_rate/qwen_vl_interface': 7.575599391834e-06, 'learning_rate/action_model': 7.575599391834e-05, 'epoch': 0.13}) +[2026-10-03 14:57:51,389][starVLA.training.train_starvla][INFO] - Step 1753, Loss: {'train/loss': 7.887196261435747e-05, 'train/grad_norm_pre_clip': 0.01604011468589306, 'timing/data': 0.0002778999914880842, 'timing/model': 0.3734760569932405, 'learning_rate/qwen_vl_interface': 7.572943467246871e-06, 'learning_rate/action_model': 7.57294346724687e-05, 'epoch': 0.13}) +[2026-10-03 14:57:52,049][starVLA.training.train_starvla][INFO] - Step 1754, Loss: {'train/loss': 0.17940568923950195, 'train/grad_norm_pre_clip': 86.137451171875, 'timing/data': 0.00021580999600701034, 'timing/model': 0.37034945700725075, 'learning_rate/qwen_vl_interface': 7.570286587785428e-06, 'learning_rate/action_model': 7.570286587785428e-05, 'epoch': 0.13}) +[2026-10-03 14:57:52,710][starVLA.training.train_starvla][INFO] - Step 1755, Loss: {'train/loss': 0.2807031273841858, 'train/grad_norm_pre_clip': 286.3025207519531, 'timing/data': 0.00021380001271609217, 'timing/model': 0.3712264179921476, 'learning_rate/qwen_vl_interface': 7.567628754541815e-06, 'learning_rate/action_model': 7.567628754541814e-05, 'epoch': 0.13}) +[2026-10-03 14:57:53,371][starVLA.training.train_starvla][INFO] - Step 1756, Loss: {'train/loss': 0.15026798844337463, 'train/grad_norm_pre_clip': 160.9879150390625, 'timing/data': 0.00021569999807979912, 'timing/model': 0.3713920070003951, 'learning_rate/qwen_vl_interface': 7.564969968608565e-06, 'learning_rate/action_model': 7.564969968608565e-05, 'epoch': 0.13}) +[2026-10-03 14:57:54,032][starVLA.training.train_starvla][INFO] - Step 1757, Loss: {'train/loss': 0.5799598097801208, 'train/grad_norm_pre_clip': 22.17996597290039, 'timing/data': 0.0002701500052353367, 'timing/model': 0.3719314469926758, 'learning_rate/qwen_vl_interface': 7.562310231078605e-06, 'learning_rate/action_model': 7.562310231078604e-05, 'epoch': 0.13}) +[2026-10-03 14:57:54,892][starVLA.training.train_starvla][INFO] - Step 1758, Loss: {'train/loss': 0.025153866037726402, 'train/grad_norm_pre_clip': 36.9298095703125, 'timing/data': 0.00036794999323319644, 'timing/model': 0.37080736699863337, 'learning_rate/qwen_vl_interface': 7.559649543045252e-06, 'learning_rate/action_model': 7.559649543045252e-05, 'epoch': 0.13}) +[2026-10-03 14:57:55,553][starVLA.training.train_starvla][INFO] - Step 1759, Loss: {'train/loss': 0.0001108415235648863, 'train/grad_norm_pre_clip': 0.014264113269746304, 'timing/data': 0.0002988399937748909, 'timing/model': 0.3714238380052848, 'learning_rate/qwen_vl_interface': 7.556987905602215e-06, 'learning_rate/action_model': 7.556987905602215e-05, 'epoch': 0.13}) +[2026-10-03 14:57:56,191][starVLA.training.train_starvla][INFO] - Step 1760, Loss: {'train/loss': 0.006995454430580139, 'train/grad_norm_pre_clip': 12.52707290649414, 'timing/data': 0.00022109001292847097, 'timing/model': 0.34786006099602673, 'learning_rate/qwen_vl_interface': 7.554325319843593e-06, 'learning_rate/action_model': 7.554325319843592e-05, 'epoch': 0.13}) +[2026-10-03 14:57:56,830][starVLA.training.train_starvla][INFO] - Step 1761, Loss: {'train/loss': 0.000795636442489922, 'train/grad_norm_pre_clip': 0.23970435559749603, 'timing/data': 0.00023053999757394195, 'timing/model': 0.37245697699836455, 'learning_rate/qwen_vl_interface': 7.551661786863874e-06, 'learning_rate/action_model': 7.551661786863874e-05, 'epoch': 0.13}) +[2026-10-03 14:57:57,491][starVLA.training.train_starvla][INFO] - Step 1762, Loss: {'train/loss': 0.4697108566761017, 'train/grad_norm_pre_clip': 27.59670066833496, 'timing/data': 0.00035675999242812395, 'timing/model': 0.37088363799557555, 'learning_rate/qwen_vl_interface': 7.548997307757936e-06, 'learning_rate/action_model': 7.548997307757935e-05, 'epoch': 0.13}) +[2026-10-03 14:57:58,322][starVLA.training.train_starvla][INFO] - Step 1763, Loss: {'train/loss': 0.3207574486732483, 'train/grad_norm_pre_clip': 25.840717315673828, 'timing/data': 0.00021392000780906528, 'timing/model': 0.5399511979921954, 'learning_rate/qwen_vl_interface': 7.546331883621044e-06, 'learning_rate/action_model': 7.546331883621044e-05, 'epoch': 0.13}) +[2026-10-03 14:57:58,958][starVLA.training.train_starvla][INFO] - Step 1764, Loss: {'train/loss': 0.41654688119888306, 'train/grad_norm_pre_clip': 25.61224365234375, 'timing/data': 0.0003213700110791251, 'timing/model': 0.34785257100884337, 'learning_rate/qwen_vl_interface': 7.543665515548857e-06, 'learning_rate/action_model': 7.543665515548856e-05, 'epoch': 0.13}) +[2026-10-03 14:57:59,620][starVLA.training.train_starvla][INFO] - Step 1765, Loss: {'train/loss': 0.17660203576087952, 'train/grad_norm_pre_clip': 113.417236328125, 'timing/data': 0.0002837000065483153, 'timing/model': 0.3716732869943371, 'learning_rate/qwen_vl_interface': 7.540998204637413e-06, 'learning_rate/action_model': 7.540998204637413e-05, 'epoch': 0.13}) +[2026-10-03 14:58:00,279][starVLA.training.train_starvla][INFO] - Step 1766, Loss: {'train/loss': 0.49546244740486145, 'train/grad_norm_pre_clip': 285.3501892089844, 'timing/data': 0.0003495699929771945, 'timing/model': 0.3704973279964179, 'learning_rate/qwen_vl_interface': 7.538329951983148e-06, 'learning_rate/action_model': 7.538329951983147e-05, 'epoch': 0.13}) +[2026-10-03 14:58:00,940][starVLA.training.train_starvla][INFO] - Step 1767, Loss: {'train/loss': 0.001097200671210885, 'train/grad_norm_pre_clip': 0.15480440855026245, 'timing/data': 0.00021640000340994447, 'timing/model': 0.3709386069967877, 'learning_rate/qwen_vl_interface': 7.535660758682877e-06, 'learning_rate/action_model': 7.535660758682876e-05, 'epoch': 0.13}) +[2026-10-03 14:58:01,796][starVLA.training.train_starvla][INFO] - Step 1768, Loss: {'train/loss': 0.6897033452987671, 'train/grad_norm_pre_clip': 69.0270767211914, 'timing/data': 0.00026494999474380165, 'timing/model': 0.370639996996033, 'learning_rate/qwen_vl_interface': 7.532990625833806e-06, 'learning_rate/action_model': 7.532990625833805e-05, 'epoch': 0.13}) +[2026-10-03 14:58:02,458][starVLA.training.train_starvla][INFO] - Step 1769, Loss: {'train/loss': 0.002386243548244238, 'train/grad_norm_pre_clip': 0.4457191228866577, 'timing/data': 0.00036048000038135797, 'timing/model': 0.37247200700221583, 'learning_rate/qwen_vl_interface': 7.5303195545335225e-06, 'learning_rate/action_model': 7.530319554533522e-05, 'epoch': 0.13}) +[2026-10-03 14:58:03,119][starVLA.training.train_starvla][INFO] - Step 1770, Loss: {'train/loss': 0.002804865362122655, 'train/grad_norm_pre_clip': 0.4663516581058502, 'timing/data': 0.0003060489980271086, 'timing/model': 0.370922578003956, 'learning_rate/qwen_vl_interface': 7.527647545880007e-06, 'learning_rate/action_model': 7.527647545880007e-05, 'epoch': 0.13}) +[2026-10-03 14:58:03,780][starVLA.training.train_starvla][INFO] - Step 1771, Loss: {'train/loss': 0.28108930587768555, 'train/grad_norm_pre_clip': 65.25900268554688, 'timing/data': 0.00021297000057529658, 'timing/model': 0.3709892370097805, 'learning_rate/qwen_vl_interface': 7.52497460097162e-06, 'learning_rate/action_model': 7.52497460097162e-05, 'epoch': 0.13}) +[2026-10-03 14:58:04,440][starVLA.training.train_starvla][INFO] - Step 1772, Loss: {'train/loss': 0.26685890555381775, 'train/grad_norm_pre_clip': 31.647266387939453, 'timing/data': 0.0002101499994751066, 'timing/model': 0.3712231480021728, 'learning_rate/qwen_vl_interface': 7.522300720907105e-06, 'learning_rate/action_model': 7.522300720907105e-05, 'epoch': 0.13}) +[2026-10-03 14:58:05,303][starVLA.training.train_starvla][INFO] - Step 1773, Loss: {'train/loss': 0.001137948827818036, 'train/grad_norm_pre_clip': 0.1925647109746933, 'timing/data': 0.00032658000418450683, 'timing/model': 0.5731756840104936, 'learning_rate/qwen_vl_interface': 7.519625906785601e-06, 'learning_rate/action_model': 7.5196259067856e-05, 'epoch': 0.13}) +[2026-10-03 14:58:05,964][starVLA.training.train_starvla][INFO] - Step 1774, Loss: {'train/loss': 0.0012037861160933971, 'train/grad_norm_pre_clip': 0.17243452370166779, 'timing/data': 0.0003371600032551214, 'timing/model': 0.3713134680001531, 'learning_rate/qwen_vl_interface': 7.516950159706614e-06, 'learning_rate/action_model': 7.516950159706614e-05, 'epoch': 0.13}) +[2026-10-03 14:58:06,625][starVLA.training.train_starvla][INFO] - Step 1775, Loss: {'train/loss': 0.0009083416080102324, 'train/grad_norm_pre_clip': 0.1670377552509308, 'timing/data': 0.00031581999792251736, 'timing/model': 0.3713821370038204, 'learning_rate/qwen_vl_interface': 7.514273480770047e-06, 'learning_rate/action_model': 7.514273480770047e-05, 'epoch': 0.13}) +[2026-10-03 14:58:07,285][starVLA.training.train_starvla][INFO] - Step 1776, Loss: {'train/loss': 0.3970259726047516, 'train/grad_norm_pre_clip': 138.53045654296875, 'timing/data': 0.0002144900063285604, 'timing/model': 0.37051975799840875, 'learning_rate/qwen_vl_interface': 7.511595871076181e-06, 'learning_rate/action_model': 7.511595871076181e-05, 'epoch': 0.13}) +[2026-10-03 14:58:07,947][starVLA.training.train_starvla][INFO] - Step 1777, Loss: {'train/loss': 0.0006992386770434678, 'train/grad_norm_pre_clip': 0.17362837493419647, 'timing/data': 0.00026760000037029386, 'timing/model': 0.37248443700082134, 'learning_rate/qwen_vl_interface': 7.50891733172568e-06, 'learning_rate/action_model': 7.50891733172568e-05, 'epoch': 0.13}) +[2026-10-03 14:58:08,609][starVLA.training.train_starvla][INFO] - Step 1778, Loss: {'train/loss': 0.0006267131539061666, 'train/grad_norm_pre_clip': 0.12914900481700897, 'timing/data': 0.00032199900306295604, 'timing/model': 0.3710690580046503, 'learning_rate/qwen_vl_interface': 7.506237863819593e-06, 'learning_rate/action_model': 7.506237863819593e-05, 'epoch': 0.13}) +[2026-10-03 14:58:09,465][starVLA.training.train_starvla][INFO] - Step 1779, Loss: {'train/loss': 0.31653398275375366, 'train/grad_norm_pre_clip': 19.808090209960938, 'timing/data': 0.00030063999292906374, 'timing/model': 0.3708147779980209, 'learning_rate/qwen_vl_interface': 7.503557468459345e-06, 'learning_rate/action_model': 7.503557468459344e-05, 'epoch': 0.13}) +[2026-10-03 14:58:10,124][starVLA.training.train_starvla][INFO] - Step 1780, Loss: {'train/loss': 0.016854925081133842, 'train/grad_norm_pre_clip': 22.645381927490234, 'timing/data': 0.00023921999672893435, 'timing/model': 0.37077992799459025, 'learning_rate/qwen_vl_interface': 7.5008761467467435e-06, 'learning_rate/action_model': 7.500876146746743e-05, 'epoch': 0.13}) +[2026-10-03 14:58:10,786][starVLA.training.train_starvla][INFO] - Step 1781, Loss: {'train/loss': 0.0026990296319127083, 'train/grad_norm_pre_clip': 0.8674134016036987, 'timing/data': 0.0002504700096324086, 'timing/model': 0.37272450800810475, 'learning_rate/qwen_vl_interface': 7.498193899783983e-06, 'learning_rate/action_model': 7.498193899783983e-05, 'epoch': 0.13}) +[2026-10-03 14:58:11,423][starVLA.training.train_starvla][INFO] - Step 1782, Loss: {'train/loss': 0.001846505794674158, 'train/grad_norm_pre_clip': 0.4429687559604645, 'timing/data': 0.00033564999466761947, 'timing/model': 0.3705172979971394, 'learning_rate/qwen_vl_interface': 7.495510728673631e-06, 'learning_rate/action_model': 7.495510728673632e-05, 'epoch': 0.13}) +[2026-10-03 14:58:12,084][starVLA.training.train_starvla][INFO] - Step 1783, Loss: {'train/loss': 0.13287489116191864, 'train/grad_norm_pre_clip': 145.41897583007812, 'timing/data': 0.00034868999500758946, 'timing/model': 0.3704805680026766, 'learning_rate/qwen_vl_interface': 7.492826634518642e-06, 'learning_rate/action_model': 7.492826634518641e-05, 'epoch': 0.13}) +[2026-10-03 14:58:12,917][starVLA.training.train_starvla][INFO] - Step 1784, Loss: {'train/loss': 0.00022546399850398302, 'train/grad_norm_pre_clip': 0.04373021423816681, 'timing/data': 0.00023052000324241817, 'timing/model': 0.3708776170096826, 'learning_rate/qwen_vl_interface': 7.4901416184223404e-06, 'learning_rate/action_model': 7.49014161842234e-05, 'epoch': 0.13}) +[2026-10-03 14:58:13,556][starVLA.training.train_starvla][INFO] - Step 1785, Loss: {'train/loss': 0.09256429970264435, 'train/grad_norm_pre_clip': 89.20333862304688, 'timing/data': 0.0002531299978727475, 'timing/model': 0.37224025699833874, 'learning_rate/qwen_vl_interface': 7.487455681488437e-06, 'learning_rate/action_model': 7.487455681488436e-05, 'epoch': 0.13}) +[2026-10-03 14:58:14,216][starVLA.training.train_starvla][INFO] - Step 1786, Loss: {'train/loss': 0.0004096459597349167, 'train/grad_norm_pre_clip': 0.08590887486934662, 'timing/data': 0.00022036000154912472, 'timing/model': 0.37081105799006764, 'learning_rate/qwen_vl_interface': 7.484768824821018e-06, 'learning_rate/action_model': 7.484768824821018e-05, 'epoch': 0.13}) +[2026-10-03 14:58:14,874][starVLA.training.train_starvla][INFO] - Step 1787, Loss: {'train/loss': 0.0002587391354609281, 'train/grad_norm_pre_clip': 0.12390033900737762, 'timing/data': 0.0003013199893757701, 'timing/model': 0.36961142800282687, 'learning_rate/qwen_vl_interface': 7.482081049524551e-06, 'learning_rate/action_model': 7.482081049524551e-05, 'epoch': 0.13}) +[2026-10-03 14:58:15,532][starVLA.training.train_starvla][INFO] - Step 1788, Loss: {'train/loss': 0.00044050943688489497, 'train/grad_norm_pre_clip': 0.09842150658369064, 'timing/data': 0.00039180999738164246, 'timing/model': 0.36910303799959365, 'learning_rate/qwen_vl_interface': 7.479392356703878e-06, 'learning_rate/action_model': 7.479392356703878e-05, 'epoch': 0.13}) +[2026-10-03 14:58:16,397][starVLA.training.train_starvla][INFO] - Step 1789, Loss: {'train/loss': 0.00035757641308009624, 'train/grad_norm_pre_clip': 0.11188887804746628, 'timing/data': 0.00027053999656345695, 'timing/model': 0.3709061980043771, 'learning_rate/qwen_vl_interface': 7.476702747464219e-06, 'learning_rate/action_model': 7.476702747464219e-05, 'epoch': 0.13}) +[2026-10-03 14:58:17,058][starVLA.training.train_starvla][INFO] - Step 1790, Loss: {'train/loss': 0.00010812012624228373, 'train/grad_norm_pre_clip': 0.017361994832754135, 'timing/data': 0.0004976599884685129, 'timing/model': 0.3684615680103889, 'learning_rate/qwen_vl_interface': 7.474012222911169e-06, 'learning_rate/action_model': 7.474012222911169e-05, 'epoch': 0.13}) +[2026-10-03 14:58:17,714][starVLA.training.train_starvla][INFO] - Step 1791, Loss: {'train/loss': 0.013655630871653557, 'train/grad_norm_pre_clip': 21.28121566772461, 'timing/data': 0.00022863000049255788, 'timing/model': 0.3684161679993849, 'learning_rate/qwen_vl_interface': 7.471320784150703e-06, 'learning_rate/action_model': 7.471320784150703e-05, 'epoch': 0.13}) +[2026-10-03 14:58:18,370][starVLA.training.train_starvla][INFO] - Step 1792, Loss: {'train/loss': 0.0003877212875522673, 'train/grad_norm_pre_clip': 0.1227714791893959, 'timing/data': 0.00030798000807408243, 'timing/model': 0.3688225289952243, 'learning_rate/qwen_vl_interface': 7.468628432289169e-06, 'learning_rate/action_model': 7.468628432289168e-05, 'epoch': 0.13}) +[2026-10-03 14:58:19,028][starVLA.training.train_starvla][INFO] - Step 1793, Loss: {'train/loss': 0.00030064769089221954, 'train/grad_norm_pre_clip': 0.10775336623191833, 'timing/data': 0.00025490000552963465, 'timing/model': 0.37048232700908557, 'learning_rate/qwen_vl_interface': 7.465935168433292e-06, 'learning_rate/action_model': 7.465935168433292e-05, 'epoch': 0.13}) +[2026-10-03 14:58:19,891][starVLA.training.train_starvla][INFO] - Step 1794, Loss: {'train/loss': 0.1642216444015503, 'train/grad_norm_pre_clip': 102.1424789428711, 'timing/data': 0.0002132600056938827, 'timing/model': 0.5757694729982177, 'learning_rate/qwen_vl_interface': 7.463240993690168e-06, 'learning_rate/action_model': 7.463240993690168e-05, 'epoch': 0.13}) +[2026-10-03 14:58:20,548][starVLA.training.train_starvla][INFO] - Step 1795, Loss: {'train/loss': 8.915751095628366e-05, 'train/grad_norm_pre_clip': 0.0185391902923584, 'timing/data': 0.00020800999482162297, 'timing/model': 0.3687732390098972, 'learning_rate/qwen_vl_interface': 7.460545909167276e-06, 'learning_rate/action_model': 7.460545909167275e-05, 'epoch': 0.13}) +[2026-10-03 14:58:21,182][starVLA.training.train_starvla][INFO] - Step 1796, Loss: {'train/loss': 9.591742855263874e-05, 'train/grad_norm_pre_clip': 0.027507074177265167, 'timing/data': 0.00021437999384943396, 'timing/model': 0.36947215799591504, 'learning_rate/qwen_vl_interface': 7.457849915972459e-06, 'learning_rate/action_model': 7.457849915972458e-05, 'epoch': 0.13}) +[2026-10-03 14:58:21,845][starVLA.training.train_starvla][INFO] - Step 1797, Loss: {'train/loss': 0.00011726833326974884, 'train/grad_norm_pre_clip': 0.03698676824569702, 'timing/data': 0.00026385999808553606, 'timing/model': 0.37246436800342053, 'learning_rate/qwen_vl_interface': 7.455153015213941e-06, 'learning_rate/action_model': 7.45515301521394e-05, 'epoch': 0.13}) +[2026-10-03 14:58:22,505][starVLA.training.train_starvla][INFO] - Step 1798, Loss: {'train/loss': 4.780676681548357e-05, 'train/grad_norm_pre_clip': 0.008939829654991627, 'timing/data': 0.00033229999826289713, 'timing/model': 0.37079873800394125, 'learning_rate/qwen_vl_interface': 7.452455208000311e-06, 'learning_rate/action_model': 7.452455208000311e-05, 'epoch': 0.13}) +[2026-10-03 14:58:23,360][starVLA.training.train_starvla][INFO] - Step 1799, Loss: {'train/loss': 0.5955315828323364, 'train/grad_norm_pre_clip': 19.44757843017578, 'timing/data': 0.00031447000219486654, 'timing/model': 0.5645197959966026, 'learning_rate/qwen_vl_interface': 7.449756495440541e-06, 'learning_rate/action_model': 7.44975649544054e-05, 'epoch': 0.13}) +[2026-10-03 14:58:24,024][starVLA.training.train_starvla][INFO] - Step 1800, Loss: {'train/loss': 0.3783254027366638, 'train/grad_norm_pre_clip': 86.88776397705078, 'timing/data': 0.00026012999296654016, 'timing/model': 0.37319802699494176, 'learning_rate/qwen_vl_interface': 7.447056878643969e-06, 'learning_rate/action_model': 7.447056878643968e-05, 'epoch': 0.13}) +[2026-10-03 14:58:24,686][starVLA.training.train_starvla][INFO] - Step 1801, Loss: {'train/loss': 0.19998033344745636, 'train/grad_norm_pre_clip': 24.497724533081055, 'timing/data': 0.000318280013743788, 'timing/model': 0.3723792480013799, 'learning_rate/qwen_vl_interface': 7.444356358720303e-06, 'learning_rate/action_model': 7.444356358720302e-05, 'epoch': 0.13}) +[2026-10-03 14:58:25,346][starVLA.training.train_starvla][INFO] - Step 1802, Loss: {'train/loss': 0.2588326036930084, 'train/grad_norm_pre_clip': 208.02561950683594, 'timing/data': 0.00034890000824816525, 'timing/model': 0.3709466779982904, 'learning_rate/qwen_vl_interface': 7.4416549367796255e-06, 'learning_rate/action_model': 7.441654936779625e-05, 'epoch': 0.13}) +[2026-10-03 14:58:26,006][starVLA.training.train_starvla][INFO] - Step 1803, Loss: {'train/loss': 0.028231393545866013, 'train/grad_norm_pre_clip': 70.69631958007812, 'timing/data': 0.00021584999922197312, 'timing/model': 0.3695262190012727, 'learning_rate/qwen_vl_interface': 7.4389526139323895e-06, 'learning_rate/action_model': 7.438952613932389e-05, 'epoch': 0.13}) +[2026-10-03 14:58:26,663][starVLA.training.train_starvla][INFO] - Step 1804, Loss: {'train/loss': 0.021179646253585815, 'train/grad_norm_pre_clip': 10.74730396270752, 'timing/data': 0.0003523799969116226, 'timing/model': 0.36953284799528774, 'learning_rate/qwen_vl_interface': 7.436249391289419e-06, 'learning_rate/action_model': 7.436249391289418e-05, 'epoch': 0.13}) +[2026-10-03 14:58:27,523][starVLA.training.train_starvla][INFO] - Step 1805, Loss: {'train/loss': 9.642499207984656e-05, 'train/grad_norm_pre_clip': 0.034489065408706665, 'timing/data': 0.0003529499954311177, 'timing/model': 0.3707486179919215, 'learning_rate/qwen_vl_interface': 7.433545269961906e-06, 'learning_rate/action_model': 7.433545269961906e-05, 'epoch': 0.13}) +[2026-10-03 14:58:28,180][starVLA.training.train_starvla][INFO] - Step 1806, Loss: {'train/loss': 0.2265465259552002, 'train/grad_norm_pre_clip': 69.26409912109375, 'timing/data': 0.0002585900074336678, 'timing/model': 0.36965949898876715, 'learning_rate/qwen_vl_interface': 7.430840251061411e-06, 'learning_rate/action_model': 7.430840251061411e-05, 'epoch': 0.13}) +[2026-10-03 14:58:28,839][starVLA.training.train_starvla][INFO] - Step 1807, Loss: {'train/loss': 0.0004081063671037555, 'train/grad_norm_pre_clip': 0.10458715260028839, 'timing/data': 0.00021806999575346708, 'timing/model': 0.3700648679950973, 'learning_rate/qwen_vl_interface': 7.428134335699866e-06, 'learning_rate/action_model': 7.428134335699865e-05, 'epoch': 0.13}) +[2026-10-03 14:58:29,497][starVLA.training.train_starvla][INFO] - Step 1808, Loss: {'train/loss': 0.00040439958684146404, 'train/grad_norm_pre_clip': 0.12081380188465118, 'timing/data': 0.0002662600018084049, 'timing/model': 0.3693759979942115, 'learning_rate/qwen_vl_interface': 7.425427524989572e-06, 'learning_rate/action_model': 7.425427524989572e-05, 'epoch': 0.13}) +[2026-10-03 14:58:30,156][starVLA.training.train_starvla][INFO] - Step 1809, Loss: {'train/loss': 0.00019198721565771848, 'train/grad_norm_pre_clip': 0.047105822712183, 'timing/data': 0.0003525399952195585, 'timing/model': 0.3707814689987572, 'learning_rate/qwen_vl_interface': 7.422719820043194e-06, 'learning_rate/action_model': 7.422719820043193e-05, 'epoch': 0.13}) +[2026-10-03 14:58:31,014][starVLA.training.train_starvla][INFO] - Step 1810, Loss: {'train/loss': 0.4631028473377228, 'train/grad_norm_pre_clip': 59.11904525756836, 'timing/data': 0.00028136999753769487, 'timing/model': 0.3695834889949765, 'learning_rate/qwen_vl_interface': 7.420011221973771e-06, 'learning_rate/action_model': 7.420011221973771e-05, 'epoch': 0.13}) +[2026-10-03 14:58:31,672][starVLA.training.train_starvla][INFO] - Step 1811, Loss: {'train/loss': 0.0002021005202550441, 'train/grad_norm_pre_clip': 0.09920289367437363, 'timing/data': 0.00032277998980134726, 'timing/model': 0.36956114901113324, 'learning_rate/qwen_vl_interface': 7.417301731894702e-06, 'learning_rate/action_model': 7.417301731894701e-05, 'epoch': 0.13}) +[2026-10-03 14:58:32,329][starVLA.training.train_starvla][INFO] - Step 1812, Loss: {'train/loss': 0.0004995653289370239, 'train/grad_norm_pre_clip': 0.11996227502822876, 'timing/data': 0.0002130799985025078, 'timing/model': 0.36935296800220385, 'learning_rate/qwen_vl_interface': 7.414591350919756e-06, 'learning_rate/action_model': 7.414591350919756e-05, 'epoch': 0.13}) +[2026-10-03 14:58:32,987][starVLA.training.train_starvla][INFO] - Step 1813, Loss: {'train/loss': 0.018441423773765564, 'train/grad_norm_pre_clip': 63.308162689208984, 'timing/data': 0.00030487000185530633, 'timing/model': 0.37077145800867584, 'learning_rate/qwen_vl_interface': 7.4118800801630715e-06, 'learning_rate/action_model': 7.411880080163072e-05, 'epoch': 0.13}) +[2026-10-03 14:58:33,645][starVLA.training.train_starvla][INFO] - Step 1814, Loss: {'train/loss': 0.07499039173126221, 'train/grad_norm_pre_clip': 25.881948471069336, 'timing/data': 0.00025769999774638563, 'timing/model': 0.3694837289949646, 'learning_rate/qwen_vl_interface': 7.409167920739146e-06, 'learning_rate/action_model': 7.409167920739147e-05, 'epoch': 0.13}) +[2026-10-03 14:58:34,497][starVLA.training.train_starvla][INFO] - Step 1815, Loss: {'train/loss': 0.0016630012542009354, 'train/grad_norm_pre_clip': 0.7021277546882629, 'timing/data': 0.00032859899511095136, 'timing/model': 0.5632827770023141, 'learning_rate/qwen_vl_interface': 7.406454873762848e-06, 'learning_rate/action_model': 7.406454873762848e-05, 'epoch': 0.13}) +[2026-10-03 14:58:35,154][starVLA.training.train_starvla][INFO] - Step 1816, Loss: {'train/loss': 0.00036550400545820594, 'train/grad_norm_pre_clip': 0.13003718852996826, 'timing/data': 0.00021532000391744077, 'timing/model': 0.36938781800563447, 'learning_rate/qwen_vl_interface': 7.403740940349408e-06, 'learning_rate/action_model': 7.403740940349408e-05, 'epoch': 0.13}) +[2026-10-03 14:58:35,813][starVLA.training.train_starvla][INFO] - Step 1817, Loss: {'train/loss': 0.0030430504120886326, 'train/grad_norm_pre_clip': 10.724976539611816, 'timing/data': 0.00035157000820618123, 'timing/model': 0.37127657800738234, 'learning_rate/qwen_vl_interface': 7.401026121614421e-06, 'learning_rate/action_model': 7.40102612161442e-05, 'epoch': 0.13}) +[2026-10-03 14:58:36,470][starVLA.training.train_starvla][INFO] - Step 1818, Loss: {'train/loss': 0.006778105162084103, 'train/grad_norm_pre_clip': 8.127081871032715, 'timing/data': 0.00023519000387750566, 'timing/model': 0.36938060900138225, 'learning_rate/qwen_vl_interface': 7.398310418673846e-06, 'learning_rate/action_model': 7.398310418673846e-05, 'epoch': 0.13}) +[2026-10-03 14:58:37,128][starVLA.training.train_starvla][INFO] - Step 1819, Loss: {'train/loss': 0.00020718941232189536, 'train/grad_norm_pre_clip': 0.08681363612413406, 'timing/data': 0.0002625600027386099, 'timing/model': 0.36974957899656147, 'learning_rate/qwen_vl_interface': 7.395593832644007e-06, 'learning_rate/action_model': 7.395593832644007e-05, 'epoch': 0.13}) +[2026-10-03 14:58:37,992][starVLA.training.train_starvla][INFO] - Step 1820, Loss: {'train/loss': 0.9132356643676758, 'train/grad_norm_pre_clip': 63.26620864868164, 'timing/data': 0.0002605400077300146, 'timing/model': 0.5758671140065417, 'learning_rate/qwen_vl_interface': 7.39287636464159e-06, 'learning_rate/action_model': 7.39287636464159e-05, 'epoch': 0.13}) +[2026-10-03 14:58:38,650][starVLA.training.train_starvla][INFO] - Step 1821, Loss: {'train/loss': 0.5223396420478821, 'train/grad_norm_pre_clip': 62.0997314453125, 'timing/data': 0.0003552100097294897, 'timing/model': 0.3708684379962506, 'learning_rate/qwen_vl_interface': 7.3901580157836425e-06, 'learning_rate/action_model': 7.390158015783643e-05, 'epoch': 0.13}) +[2026-10-03 14:58:39,307][starVLA.training.train_starvla][INFO] - Step 1822, Loss: {'train/loss': 0.00018762185936793685, 'train/grad_norm_pre_clip': 0.04078451544046402, 'timing/data': 0.0002194299886468798, 'timing/model': 0.3695669890003046, 'learning_rate/qwen_vl_interface': 7.387438787187576e-06, 'learning_rate/action_model': 7.387438787187576e-05, 'epoch': 0.13}) +[2026-10-03 14:58:39,943][starVLA.training.train_starvla][INFO] - Step 1823, Loss: {'train/loss': 0.029245998710393906, 'train/grad_norm_pre_clip': 23.21605110168457, 'timing/data': 0.00022548998822458088, 'timing/model': 0.347307522999472, 'learning_rate/qwen_vl_interface': 7.3847186799711625e-06, 'learning_rate/action_model': 7.384718679971162e-05, 'epoch': 0.13}) +[2026-10-03 14:58:40,601][starVLA.training.train_starvla][INFO] - Step 1824, Loss: {'train/loss': 0.0009097656584344804, 'train/grad_norm_pre_clip': 0.653235137462616, 'timing/data': 0.0003765699948417023, 'timing/model': 0.3695352890063077, 'learning_rate/qwen_vl_interface': 7.381997695252535e-06, 'learning_rate/action_model': 7.381997695252535e-05, 'epoch': 0.13}) +[2026-10-03 14:58:41,432][starVLA.training.train_starvla][INFO] - Step 1825, Loss: {'train/loss': 0.034430284053087234, 'train/grad_norm_pre_clip': 86.17366027832031, 'timing/data': 0.0002602800086606294, 'timing/model': 0.5434714899893152, 'learning_rate/qwen_vl_interface': 7.379275834150186e-06, 'learning_rate/action_model': 7.379275834150186e-05, 'epoch': 0.13}) +[2026-10-03 14:58:42,090][starVLA.training.train_starvla][INFO] - Step 1826, Loss: {'train/loss': 0.1448126882314682, 'train/grad_norm_pre_clip': 95.68888092041016, 'timing/data': 0.00022195999918039888, 'timing/model': 0.3696378200111212, 'learning_rate/qwen_vl_interface': 7.376553097782972e-06, 'learning_rate/action_model': 7.376553097782972e-05, 'epoch': 0.13}) +[2026-10-03 14:58:42,750][starVLA.training.train_starvla][INFO] - Step 1827, Loss: {'train/loss': 0.8198556303977966, 'train/grad_norm_pre_clip': 20.46551513671875, 'timing/data': 0.00021923000167589635, 'timing/model': 0.3705227789905621, 'learning_rate/qwen_vl_interface': 7.373829487270109e-06, 'learning_rate/action_model': 7.373829487270109e-05, 'epoch': 0.13}) +[2026-10-03 14:58:43,408][starVLA.training.train_starvla][INFO] - Step 1828, Loss: {'train/loss': 8.481954864691943e-05, 'train/grad_norm_pre_clip': 0.01323709636926651, 'timing/data': 0.0003411899961065501, 'timing/model': 0.3698557080060709, 'learning_rate/qwen_vl_interface': 7.371105003731165e-06, 'learning_rate/action_model': 7.371105003731164e-05, 'epoch': 0.13}) +[2026-10-03 14:58:44,068][starVLA.training.train_starvla][INFO] - Step 1829, Loss: {'train/loss': 0.00573284737765789, 'train/grad_norm_pre_clip': 12.74691390991211, 'timing/data': 0.0002675200084922835, 'timing/model': 0.37145156900805887, 'learning_rate/qwen_vl_interface': 7.3683796482860755e-06, 'learning_rate/action_model': 7.368379648286075e-05, 'epoch': 0.13}) +[2026-10-03 14:58:44,725][starVLA.training.train_starvla][INFO] - Step 1830, Loss: {'train/loss': 0.14932581782341003, 'train/grad_norm_pre_clip': 82.59835052490234, 'timing/data': 0.00021381000988185406, 'timing/model': 0.3696282879973296, 'learning_rate/qwen_vl_interface': 7.365653422055131e-06, 'learning_rate/action_model': 7.365653422055131e-05, 'epoch': 0.13}) +[2026-10-03 14:58:45,581][starVLA.training.train_starvla][INFO] - Step 1831, Loss: {'train/loss': 0.0015637761680409312, 'train/grad_norm_pre_clip': 0.8097149133682251, 'timing/data': 0.00033961000735871494, 'timing/model': 0.37021210900275037, 'learning_rate/qwen_vl_interface': 7.3629263261589816e-06, 'learning_rate/action_model': 7.362926326158981e-05, 'epoch': 0.13}) +[2026-10-03 14:58:46,239][starVLA.training.train_starvla][INFO] - Step 1832, Loss: {'train/loss': 0.20003917813301086, 'train/grad_norm_pre_clip': 150.9034423828125, 'timing/data': 0.00022498000180348754, 'timing/model': 0.36968710900691804, 'learning_rate/qwen_vl_interface': 7.36019836171863e-06, 'learning_rate/action_model': 7.360198361718629e-05, 'epoch': 0.13}) +[2026-10-03 14:58:46,900][starVLA.training.train_starvla][INFO] - Step 1833, Loss: {'train/loss': 0.10059192031621933, 'train/grad_norm_pre_clip': 90.27000427246094, 'timing/data': 0.0002262799971504137, 'timing/model': 0.37141608900856227, 'learning_rate/qwen_vl_interface': 7.357469529855441e-06, 'learning_rate/action_model': 7.35746952985544e-05, 'epoch': 0.13}) +[2026-10-03 14:58:47,558][starVLA.training.train_starvla][INFO] - Step 1834, Loss: {'train/loss': 0.0012118940940126777, 'train/grad_norm_pre_clip': 1.4071874618530273, 'timing/data': 0.0002239199966425076, 'timing/model': 0.36958882900944445, 'learning_rate/qwen_vl_interface': 7.354739831691134e-06, 'learning_rate/action_model': 7.354739831691135e-05, 'epoch': 0.13}) +[2026-10-03 14:58:48,217][starVLA.training.train_starvla][INFO] - Step 1835, Loss: {'train/loss': 0.25122207403182983, 'train/grad_norm_pre_clip': 72.19137573242188, 'timing/data': 0.00035960999957751483, 'timing/model': 0.3703097679972416, 'learning_rate/qwen_vl_interface': 7.352009268347786e-06, 'learning_rate/action_model': 7.352009268347786e-05, 'epoch': 0.13}) +[2026-10-03 14:58:49,053][starVLA.training.train_starvla][INFO] - Step 1836, Loss: {'train/loss': 0.048245105892419815, 'train/grad_norm_pre_clip': 52.00086975097656, 'timing/data': 0.0003502199979266152, 'timing/model': 0.5701817060034955, 'learning_rate/qwen_vl_interface': 7.349277840947828e-06, 'learning_rate/action_model': 7.349277840947828e-05, 'epoch': 0.13}) +[2026-10-03 14:58:49,713][starVLA.training.train_starvla][INFO] - Step 1837, Loss: {'train/loss': 0.0005405094707384706, 'train/grad_norm_pre_clip': 0.2127196490764618, 'timing/data': 0.0002558299893280491, 'timing/model': 0.3721644879988162, 'learning_rate/qwen_vl_interface': 7.346545550614046e-06, 'learning_rate/action_model': 7.346545550614046e-05, 'epoch': 0.13}) +[2026-10-03 14:58:50,372][starVLA.training.train_starvla][INFO] - Step 1838, Loss: {'train/loss': 0.6324058771133423, 'train/grad_norm_pre_clip': 24.78495216369629, 'timing/data': 0.00028008001390844584, 'timing/model': 0.36953513900516555, 'learning_rate/qwen_vl_interface': 7.343812398469582e-06, 'learning_rate/action_model': 7.343812398469581e-05, 'epoch': 0.13}) +[2026-10-03 14:58:51,033][starVLA.training.train_starvla][INFO] - Step 1839, Loss: {'train/loss': 0.07259551435709, 'train/grad_norm_pre_clip': 29.92206382751465, 'timing/data': 0.00035458999627735466, 'timing/model': 0.3704271389869973, 'learning_rate/qwen_vl_interface': 7.34107838563793e-06, 'learning_rate/action_model': 7.341078385637929e-05, 'epoch': 0.13}) +[2026-10-03 14:58:51,692][starVLA.training.train_starvla][INFO] - Step 1840, Loss: {'train/loss': 0.5001487135887146, 'train/grad_norm_pre_clip': 36.489864349365234, 'timing/data': 0.0003488100046524778, 'timing/model': 0.37018838900257833, 'learning_rate/qwen_vl_interface': 7.338343513242941e-06, 'learning_rate/action_model': 7.338343513242941e-05, 'epoch': 0.13}) +[2026-10-03 14:58:52,556][starVLA.training.train_starvla][INFO] - Step 1841, Loss: {'train/loss': 0.269086092710495, 'train/grad_norm_pre_clip': 20.50967788696289, 'timing/data': 0.000440629999502562, 'timing/model': 0.5754920659965137, 'learning_rate/qwen_vl_interface': 7.335607782408818e-06, 'learning_rate/action_model': 7.335607782408818e-05, 'epoch': 0.13}) +[2026-10-03 14:58:53,214][starVLA.training.train_starvla][INFO] - Step 1842, Loss: {'train/loss': 0.0002502642455510795, 'train/grad_norm_pre_clip': 0.060088999569416046, 'timing/data': 0.00035597999521996826, 'timing/model': 0.3690739789890358, 'learning_rate/qwen_vl_interface': 7.3328711942601145e-06, 'learning_rate/action_model': 7.332871194260115e-05, 'epoch': 0.13}) +[2026-10-03 14:58:53,874][starVLA.training.train_starvla][INFO] - Step 1843, Loss: {'train/loss': 0.040045399218797684, 'train/grad_norm_pre_clip': 43.09169387817383, 'timing/data': 0.0003884489997290075, 'timing/model': 0.37043317999632563, 'learning_rate/qwen_vl_interface': 7.33013374992174e-06, 'learning_rate/action_model': 7.33013374992174e-05, 'epoch': 0.13}) +[2026-10-03 14:58:54,532][starVLA.training.train_starvla][INFO] - Step 1844, Loss: {'train/loss': 0.08317610621452332, 'train/grad_norm_pre_clip': 62.89384460449219, 'timing/data': 0.0003404099988983944, 'timing/model': 0.36986817899742164, 'learning_rate/qwen_vl_interface': 7.327395450518956e-06, 'learning_rate/action_model': 7.327395450518956e-05, 'epoch': 0.13}) +[2026-10-03 14:58:55,193][starVLA.training.train_starvla][INFO] - Step 1845, Loss: {'train/loss': 0.0001744020846672356, 'train/grad_norm_pre_clip': 0.0335090234875679, 'timing/data': 0.0003875099937431514, 'timing/model': 0.37139309899066575, 'learning_rate/qwen_vl_interface': 7.324656297177372e-06, 'learning_rate/action_model': 7.324656297177371e-05, 'epoch': 0.14}) +[2026-10-03 14:58:55,851][starVLA.training.train_starvla][INFO] - Step 1846, Loss: {'train/loss': 0.001661364920437336, 'train/grad_norm_pre_clip': 0.6131846308708191, 'timing/data': 0.00032519000524189323, 'timing/model': 0.37051791900012176, 'learning_rate/qwen_vl_interface': 7.32191629102295e-06, 'learning_rate/action_model': 7.32191629102295e-05, 'epoch': 0.14}) +[2026-10-03 14:58:56,719][starVLA.training.train_starvla][INFO] - Step 1847, Loss: {'train/loss': 0.0013627365697175264, 'train/grad_norm_pre_clip': 0.7123888731002808, 'timing/data': 0.00022572999296244234, 'timing/model': 0.3697069290064974, 'learning_rate/qwen_vl_interface': 7.319175433182004e-06, 'learning_rate/action_model': 7.319175433182004e-05, 'epoch': 0.14}) +[2026-10-03 14:58:57,377][starVLA.training.train_starvla][INFO] - Step 1848, Loss: {'train/loss': 0.39637625217437744, 'train/grad_norm_pre_clip': 66.3143081665039, 'timing/data': 0.0003192799922544509, 'timing/model': 0.3700643190095434, 'learning_rate/qwen_vl_interface': 7.316433724781198e-06, 'learning_rate/action_model': 7.316433724781197e-05, 'epoch': 0.14}) +[2026-10-03 14:58:58,036][starVLA.training.train_starvla][INFO] - Step 1849, Loss: {'train/loss': 0.00033819081727415323, 'train/grad_norm_pre_clip': 0.10755764693021774, 'timing/data': 0.00027457000396680087, 'timing/model': 0.3712417789938627, 'learning_rate/qwen_vl_interface': 7.313691166947543e-06, 'learning_rate/action_model': 7.313691166947543e-05, 'epoch': 0.14}) +[2026-10-03 14:58:58,693][starVLA.training.train_starvla][INFO] - Step 1850, Loss: {'train/loss': 0.00018210563575848937, 'train/grad_norm_pre_clip': 0.028362566605210304, 'timing/data': 0.00025966999237425625, 'timing/model': 0.3691585900087375, 'learning_rate/qwen_vl_interface': 7.310947760808401e-06, 'learning_rate/action_model': 7.3109477608084e-05, 'epoch': 0.14}) +[2026-10-03 14:58:59,351][starVLA.training.train_starvla][INFO] - Step 1851, Loss: {'train/loss': 0.01513504609465599, 'train/grad_norm_pre_clip': 32.7916259765625, 'timing/data': 0.0002563499874668196, 'timing/model': 0.36897709900222253, 'learning_rate/qwen_vl_interface': 7.308203507491485e-06, 'learning_rate/action_model': 7.308203507491484e-05, 'epoch': 0.14}) +[2026-10-03 14:59:00,196][starVLA.training.train_starvla][INFO] - Step 1852, Loss: {'train/loss': 0.00022380004520528018, 'train/grad_norm_pre_clip': 0.03791723772883415, 'timing/data': 0.00031956000020727515, 'timing/model': 0.5570273880002787, 'learning_rate/qwen_vl_interface': 7.3054584081248515e-06, 'learning_rate/action_model': 7.305458408124852e-05, 'epoch': 0.14}) +[2026-10-03 14:59:00,856][starVLA.training.train_starvla][INFO] - Step 1853, Loss: {'train/loss': 0.0004328081849962473, 'train/grad_norm_pre_clip': 0.4350633919239044, 'timing/data': 0.00026929999876301736, 'timing/model': 0.3709347390104085, 'learning_rate/qwen_vl_interface': 7.302712463836907e-06, 'learning_rate/action_model': 7.302712463836907e-05, 'epoch': 0.14}) +[2026-10-03 14:59:01,514][starVLA.training.train_starvla][INFO] - Step 1854, Loss: {'train/loss': 0.0009130599210038781, 'train/grad_norm_pre_clip': 0.33252015709877014, 'timing/data': 0.00021405999723356217, 'timing/model': 0.37007834899122827, 'learning_rate/qwen_vl_interface': 7.2999656757564086e-06, 'learning_rate/action_model': 7.299965675756409e-05, 'epoch': 0.14}) +[2026-10-03 14:59:02,172][starVLA.training.train_starvla][INFO] - Step 1855, Loss: {'train/loss': 0.31417733430862427, 'train/grad_norm_pre_clip': 43.4028205871582, 'timing/data': 0.00026914999762084335, 'timing/model': 0.36951053999655414, 'learning_rate/qwen_vl_interface': 7.297218045012453e-06, 'learning_rate/action_model': 7.297218045012453e-05, 'epoch': 0.14}) +[2026-10-03 14:59:02,829][starVLA.training.train_starvla][INFO] - Step 1856, Loss: {'train/loss': 0.00011190256918780506, 'train/grad_norm_pre_clip': 0.019892267882823944, 'timing/data': 0.0003429400094319135, 'timing/model': 0.3693080689990893, 'learning_rate/qwen_vl_interface': 7.29446957273449e-06, 'learning_rate/action_model': 7.294469572734489e-05, 'epoch': 0.14}) +[2026-10-03 14:59:03,687][starVLA.training.train_starvla][INFO] - Step 1857, Loss: {'train/loss': 0.00011413699394324794, 'train/grad_norm_pre_clip': 0.02108578197658062, 'timing/data': 0.00028002999897580594, 'timing/model': 0.569953625992639, 'learning_rate/qwen_vl_interface': 7.2917202600523104e-06, 'learning_rate/action_model': 7.29172026005231e-05, 'epoch': 0.14}) +[2026-10-03 14:59:04,343][starVLA.training.train_starvla][INFO] - Step 1858, Loss: {'train/loss': 0.2447012960910797, 'train/grad_norm_pre_clip': 125.78775787353516, 'timing/data': 0.0003865400067297742, 'timing/model': 0.36845462999190204, 'learning_rate/qwen_vl_interface': 7.2889701080960565e-06, 'learning_rate/action_model': 7.288970108096056e-05, 'epoch': 0.14}) +[2026-10-03 14:59:05,001][starVLA.training.train_starvla][INFO] - Step 1859, Loss: {'train/loss': 0.13752642273902893, 'train/grad_norm_pre_clip': 162.25965881347656, 'timing/data': 0.00035339000169187784, 'timing/model': 0.369477429005201, 'learning_rate/qwen_vl_interface': 7.286219117996209e-06, 'learning_rate/action_model': 7.286219117996207e-05, 'epoch': 0.14}) +[2026-10-03 14:59:05,658][starVLA.training.train_starvla][INFO] - Step 1860, Loss: {'train/loss': 0.08305178582668304, 'train/grad_norm_pre_clip': 26.872867584228516, 'timing/data': 0.000332009993144311, 'timing/model': 0.3690446289983811, 'learning_rate/qwen_vl_interface': 7.283467290883598e-06, 'learning_rate/action_model': 7.283467290883598e-05, 'epoch': 0.14}) +[2026-10-03 14:59:06,316][starVLA.training.train_starvla][INFO] - Step 1861, Loss: {'train/loss': 0.0016043868381530046, 'train/grad_norm_pre_clip': 1.1770973205566406, 'timing/data': 0.00021736000780947506, 'timing/model': 0.3704924189951271, 'learning_rate/qwen_vl_interface': 7.2807146278893925e-06, 'learning_rate/action_model': 7.280714627889393e-05, 'epoch': 0.14}) +[2026-10-03 14:59:06,927][starVLA.training.train_starvla][INFO] - Step 1862, Loss: {'train/loss': 4.9169855628861114e-05, 'train/grad_norm_pre_clip': 0.007820289582014084, 'timing/data': 0.00021905000903643668, 'timing/model': 0.34625828299613204, 'learning_rate/qwen_vl_interface': 7.277961130145111e-06, 'learning_rate/action_model': 7.277961130145111e-05, 'epoch': 0.14}) +[2026-10-03 14:59:07,790][starVLA.training.train_starvla][INFO] - Step 1863, Loss: {'train/loss': 1.1597435474395752, 'train/grad_norm_pre_clip': 50.47203063964844, 'timing/data': 0.0003352189960423857, 'timing/model': 0.37011246000474785, 'learning_rate/qwen_vl_interface': 7.2752067987826116e-06, 'learning_rate/action_model': 7.275206798782611e-05, 'epoch': 0.14}) +[2026-10-03 14:59:08,447][starVLA.training.train_starvla][INFO] - Step 1864, Loss: {'train/loss': 0.00021297033526934683, 'train/grad_norm_pre_clip': 0.06036090850830078, 'timing/data': 0.00034927998785860837, 'timing/model': 0.36980770899390336, 'learning_rate/qwen_vl_interface': 7.272451634934096e-06, 'learning_rate/action_model': 7.272451634934096e-05, 'epoch': 0.14}) +[2026-10-03 14:59:09,106][starVLA.training.train_starvla][INFO] - Step 1865, Loss: {'train/loss': 0.10429882258176804, 'train/grad_norm_pre_clip': 201.4456024169922, 'timing/data': 0.0002587199996924028, 'timing/model': 0.3708621289988514, 'learning_rate/qwen_vl_interface': 7.269695639732108e-06, 'learning_rate/action_model': 7.269695639732108e-05, 'epoch': 0.14}) +[2026-10-03 14:59:09,763][starVLA.training.train_starvla][INFO] - Step 1866, Loss: {'train/loss': 0.5477119088172913, 'train/grad_norm_pre_clip': 47.97016143798828, 'timing/data': 0.0002893700002459809, 'timing/model': 0.3688665499939816, 'learning_rate/qwen_vl_interface': 7.266938814309533e-06, 'learning_rate/action_model': 7.266938814309533e-05, 'epoch': 0.14}) +[2026-10-03 14:59:10,420][starVLA.training.train_starvla][INFO] - Step 1867, Loss: {'train/loss': 0.000421107659349218, 'train/grad_norm_pre_clip': 0.1349043995141983, 'timing/data': 0.0003201789950253442, 'timing/model': 0.3684865899995202, 'learning_rate/qwen_vl_interface': 7.264181159799598e-06, 'learning_rate/action_model': 7.264181159799597e-05, 'epoch': 0.14}) +[2026-10-03 14:59:11,280][starVLA.training.train_starvla][INFO] - Step 1868, Loss: {'train/loss': 0.0002100747951772064, 'train/grad_norm_pre_clip': 0.0354444682598114, 'timing/data': 0.00040581999928690493, 'timing/model': 0.3697111500077881, 'learning_rate/qwen_vl_interface': 7.261422677335871e-06, 'learning_rate/action_model': 7.261422677335871e-05, 'epoch': 0.14}) +[2026-10-03 14:59:11,916][starVLA.training.train_starvla][INFO] - Step 1869, Loss: {'train/loss': 0.00019290779891889542, 'train/grad_norm_pre_clip': 0.042165789753198624, 'timing/data': 0.00029632000951096416, 'timing/model': 0.370229599007871, 'learning_rate/qwen_vl_interface': 7.258663368052259e-06, 'learning_rate/action_model': 7.258663368052259e-05, 'epoch': 0.14}) +[2026-10-03 14:59:12,572][starVLA.training.train_starvla][INFO] - Step 1870, Loss: {'train/loss': 0.00021134145208634436, 'train/grad_norm_pre_clip': 0.05191447585821152, 'timing/data': 0.0003265999985160306, 'timing/model': 0.3687692390085431, 'learning_rate/qwen_vl_interface': 7.2559032330830105e-06, 'learning_rate/action_model': 7.25590323308301e-05, 'epoch': 0.14}) +[2026-10-03 14:59:13,229][starVLA.training.train_starvla][INFO] - Step 1871, Loss: {'train/loss': 0.000451442610938102, 'train/grad_norm_pre_clip': 0.14889900386333466, 'timing/data': 0.00021486000332515687, 'timing/model': 0.3690707299974747, 'learning_rate/qwen_vl_interface': 7.253142273562712e-06, 'learning_rate/action_model': 7.253142273562712e-05, 'epoch': 0.14}) +[2026-10-03 14:59:13,885][starVLA.training.train_starvla][INFO] - Step 1872, Loss: {'train/loss': 0.0005221716710366309, 'train/grad_norm_pre_clip': 0.9871389865875244, 'timing/data': 0.00025293001090176404, 'timing/model': 0.3687604799924884, 'learning_rate/qwen_vl_interface': 7.25038049062629e-06, 'learning_rate/action_model': 7.250380490626289e-05, 'epoch': 0.14}) +[2026-10-03 14:59:14,743][starVLA.training.train_starvla][INFO] - Step 1873, Loss: {'train/loss': 0.019877523183822632, 'train/grad_norm_pre_clip': 55.52286148071289, 'timing/data': 0.00026452000020071864, 'timing/model': 0.5701127169886604, 'learning_rate/qwen_vl_interface': 7.24761788540901e-06, 'learning_rate/action_model': 7.24761788540901e-05, 'epoch': 0.14}) +[2026-10-03 14:59:15,400][starVLA.training.train_starvla][INFO] - Step 1874, Loss: {'train/loss': 0.00030494233942590654, 'train/grad_norm_pre_clip': 0.0638812929391861, 'timing/data': 0.00036827000440098345, 'timing/model': 0.3691721189970849, 'learning_rate/qwen_vl_interface': 7.244854459046473e-06, 'learning_rate/action_model': 7.244854459046473e-05, 'epoch': 0.14}) +[2026-10-03 14:59:16,059][starVLA.training.train_starvla][INFO] - Step 1875, Loss: {'train/loss': 0.3253065347671509, 'train/grad_norm_pre_clip': 61.77906799316406, 'timing/data': 0.00025772000662982464, 'timing/model': 0.3696016700123437, 'learning_rate/qwen_vl_interface': 7.24209021267462e-06, 'learning_rate/action_model': 7.24209021267462e-05, 'epoch': 0.14}) +[2026-10-03 14:59:16,692][starVLA.training.train_starvla][INFO] - Step 1876, Loss: {'train/loss': 0.00010744984319899231, 'train/grad_norm_pre_clip': 0.011915918439626694, 'timing/data': 0.00021844998991582543, 'timing/model': 0.3688489699998172, 'learning_rate/qwen_vl_interface': 7.23932514742973e-06, 'learning_rate/action_model': 7.23932514742973e-05, 'epoch': 0.14}) +[2026-10-03 14:59:17,351][starVLA.training.train_starvla][INFO] - Step 1877, Loss: {'train/loss': 0.00014946061128284782, 'train/grad_norm_pre_clip': 0.025746585801243782, 'timing/data': 0.00021363999985624105, 'timing/model': 0.3708987300051376, 'learning_rate/qwen_vl_interface': 7.236559264448413e-06, 'learning_rate/action_model': 7.236559264448412e-05, 'epoch': 0.14}) +[2026-10-03 14:59:18,210][starVLA.training.train_starvla][INFO] - Step 1878, Loss: {'train/loss': 0.023387471213936806, 'train/grad_norm_pre_clip': 61.450740814208984, 'timing/data': 0.0003461599990259856, 'timing/model': 0.5711830569925951, 'learning_rate/qwen_vl_interface': 7.233792564867623e-06, 'learning_rate/action_model': 7.233792564867623e-05, 'epoch': 0.14}) +[2026-10-03 14:59:18,868][starVLA.training.train_starvla][INFO] - Step 1879, Loss: {'train/loss': 0.0001548978325445205, 'train/grad_norm_pre_clip': 0.043520815670490265, 'timing/data': 0.0003215000033378601, 'timing/model': 0.3693303590116557, 'learning_rate/qwen_vl_interface': 7.231025049824644e-06, 'learning_rate/action_model': 7.231025049824643e-05, 'epoch': 0.14}) +[2026-10-03 14:59:19,524][starVLA.training.train_starvla][INFO] - Step 1880, Loss: {'train/loss': 0.1752815991640091, 'train/grad_norm_pre_clip': 108.12406158447266, 'timing/data': 0.0004227089957566932, 'timing/model': 0.36909944999206346, 'learning_rate/qwen_vl_interface': 7.228256720457097e-06, 'learning_rate/action_model': 7.228256720457096e-05, 'epoch': 0.14}) +[2026-10-03 14:59:20,183][starVLA.training.train_starvla][INFO] - Step 1881, Loss: {'train/loss': 0.00014724585344083607, 'train/grad_norm_pre_clip': 0.02667883224785328, 'timing/data': 0.000301688996842131, 'timing/model': 0.37021475000074133, 'learning_rate/qwen_vl_interface': 7.225487577902937e-06, 'learning_rate/action_model': 7.225487577902936e-05, 'epoch': 0.14}) +[2026-10-03 14:59:20,817][starVLA.training.train_starvla][INFO] - Step 1882, Loss: {'train/loss': 0.0002994398819282651, 'train/grad_norm_pre_clip': 0.07935187220573425, 'timing/data': 0.0003285999991931021, 'timing/model': 0.36890525999478996, 'learning_rate/qwen_vl_interface': 7.222717623300455e-06, 'learning_rate/action_model': 7.222717623300455e-05, 'epoch': 0.14}) +[2026-10-03 14:59:21,475][starVLA.training.train_starvla][INFO] - Step 1883, Loss: {'train/loss': 0.0006191448774188757, 'train/grad_norm_pre_clip': 0.44558289647102356, 'timing/data': 0.0002149099891539663, 'timing/model': 0.36935762999928556, 'learning_rate/qwen_vl_interface': 7.219946857788273e-06, 'learning_rate/action_model': 7.219946857788272e-05, 'epoch': 0.14}) +[2026-10-03 14:59:22,326][starVLA.training.train_starvla][INFO] - Step 1884, Loss: {'train/loss': 0.2773328125476837, 'train/grad_norm_pre_clip': 51.48780822753906, 'timing/data': 0.000275440004770644, 'timing/model': 0.3696630600024946, 'learning_rate/qwen_vl_interface': 7.21717528250535e-06, 'learning_rate/action_model': 7.21717528250535e-05, 'epoch': 0.14}) +[2026-10-03 14:59:22,984][starVLA.training.train_starvla][INFO] - Step 1885, Loss: {'train/loss': 0.0034763235598802567, 'train/grad_norm_pre_clip': 3.687720775604248, 'timing/data': 0.00027145999774802476, 'timing/model': 0.3708584800042445, 'learning_rate/qwen_vl_interface': 7.214402898590974e-06, 'learning_rate/action_model': 7.214402898590973e-05, 'epoch': 0.14}) +[2026-10-03 14:59:23,641][starVLA.training.train_starvla][INFO] - Step 1886, Loss: {'train/loss': 0.444070428609848, 'train/grad_norm_pre_clip': 107.3156967163086, 'timing/data': 0.0002655300049809739, 'timing/model': 0.36942773999180645, 'learning_rate/qwen_vl_interface': 7.211629707184768e-06, 'learning_rate/action_model': 7.211629707184768e-05, 'epoch': 0.14}) +[2026-10-03 14:59:24,303][starVLA.training.train_starvla][INFO] - Step 1887, Loss: {'train/loss': 0.0004945715772919357, 'train/grad_norm_pre_clip': 0.14538076519966125, 'timing/data': 0.00035051998565904796, 'timing/model': 0.3701415200048359, 'learning_rate/qwen_vl_interface': 7.2088557094266885e-06, 'learning_rate/action_model': 7.208855709426688e-05, 'epoch': 0.14}) +[2026-10-03 14:59:24,961][starVLA.training.train_starvla][INFO] - Step 1888, Loss: {'train/loss': 0.6121873259544373, 'train/grad_norm_pre_clip': 27.036426544189453, 'timing/data': 0.0003495900018606335, 'timing/model': 0.36902579000161495, 'learning_rate/qwen_vl_interface': 7.2060809064570195e-06, 'learning_rate/action_model': 7.20608090645702e-05, 'epoch': 0.14}) +[2026-10-03 14:59:25,817][starVLA.training.train_starvla][INFO] - Step 1889, Loss: {'train/loss': 0.00020929799939040095, 'train/grad_norm_pre_clip': 0.082125224173069, 'timing/data': 0.00026362000789958984, 'timing/model': 0.5679477880039485, 'learning_rate/qwen_vl_interface': 7.203305299416375e-06, 'learning_rate/action_model': 7.203305299416375e-05, 'epoch': 0.14}) +[2026-10-03 14:59:26,475][starVLA.training.train_starvla][INFO] - Step 1890, Loss: {'train/loss': 0.0020324070937931538, 'train/grad_norm_pre_clip': 2.2058680057525635, 'timing/data': 0.00034620999940671027, 'timing/model': 0.3696473599993624, 'learning_rate/qwen_vl_interface': 7.200528889445707e-06, 'learning_rate/action_model': 7.200528889445706e-05, 'epoch': 0.14}) +[2026-10-03 14:59:27,133][starVLA.training.train_starvla][INFO] - Step 1891, Loss: {'train/loss': 0.000656821473967284, 'train/grad_norm_pre_clip': 0.3514777421951294, 'timing/data': 0.00026004000392276794, 'timing/model': 0.3693767699878663, 'learning_rate/qwen_vl_interface': 7.19775167768629e-06, 'learning_rate/action_model': 7.19775167768629e-05, 'epoch': 0.14}) +[2026-10-03 14:59:27,768][starVLA.training.train_starvla][INFO] - Step 1892, Loss: {'train/loss': 0.48186373710632324, 'train/grad_norm_pre_clip': 45.025054931640625, 'timing/data': 0.0002188200014643371, 'timing/model': 0.347006103998865, 'learning_rate/qwen_vl_interface': 7.194973665279731e-06, 'learning_rate/action_model': 7.19497366527973e-05, 'epoch': 0.14}) +[2026-10-03 14:59:28,427][starVLA.training.train_starvla][INFO] - Step 1893, Loss: {'train/loss': 0.545275092124939, 'train/grad_norm_pre_clip': 240.16688537597656, 'timing/data': 0.00032084000122267753, 'timing/model': 0.3713115900027333, 'learning_rate/qwen_vl_interface': 7.1921948533679666e-06, 'learning_rate/action_model': 7.192194853367967e-05, 'epoch': 0.14}) +[2026-10-03 14:59:29,085][starVLA.training.train_starvla][INFO] - Step 1894, Loss: {'train/loss': 0.0008406371343880892, 'train/grad_norm_pre_clip': 0.24381758272647858, 'timing/data': 0.0003722199908224866, 'timing/model': 0.36898354000004474, 'learning_rate/qwen_vl_interface': 7.1894152430932605e-06, 'learning_rate/action_model': 7.189415243093261e-05, 'epoch': 0.14}) +[2026-10-03 14:59:29,949][starVLA.training.train_starvla][INFO] - Step 1895, Loss: {'train/loss': 0.003599998541176319, 'train/grad_norm_pre_clip': 3.383941173553467, 'timing/data': 0.00034062999475281686, 'timing/model': 0.3695765489974292, 'learning_rate/qwen_vl_interface': 7.186634835598205e-06, 'learning_rate/action_model': 7.186634835598205e-05, 'epoch': 0.14}) +[2026-10-03 14:59:30,606][starVLA.training.train_starvla][INFO] - Step 1896, Loss: {'train/loss': 0.00039259486948139966, 'train/grad_norm_pre_clip': 0.15101419389247894, 'timing/data': 0.0003465100016910583, 'timing/model': 0.36928045899549033, 'learning_rate/qwen_vl_interface': 7.183853632025724e-06, 'learning_rate/action_model': 7.183853632025723e-05, 'epoch': 0.14}) +[2026-10-03 14:59:31,240][starVLA.training.train_starvla][INFO] - Step 1897, Loss: {'train/loss': 0.6034115552902222, 'train/grad_norm_pre_clip': 23.38608741760254, 'timing/data': 0.00026543000421952456, 'timing/model': 0.37000113000976853, 'learning_rate/qwen_vl_interface': 7.181071633519062e-06, 'learning_rate/action_model': 7.181071633519062e-05, 'epoch': 0.14}) +[2026-10-03 14:59:31,898][starVLA.training.train_starvla][INFO] - Step 1898, Loss: {'train/loss': 0.00025109644047915936, 'train/grad_norm_pre_clip': 0.08397882431745529, 'timing/data': 0.000261649998719804, 'timing/model': 0.3692449699883582, 'learning_rate/qwen_vl_interface': 7.178288841221795e-06, 'learning_rate/action_model': 7.178288841221794e-05, 'epoch': 0.14}) +[2026-10-03 14:59:32,556][starVLA.training.train_starvla][INFO] - Step 1899, Loss: {'train/loss': 0.001750760362483561, 'train/grad_norm_pre_clip': 0.6235718727111816, 'timing/data': 0.00032539998937863857, 'timing/model': 0.3685732900048606, 'learning_rate/qwen_vl_interface': 7.1755052562778215e-06, 'learning_rate/action_model': 7.175505256277822e-05, 'epoch': 0.14}) +[2026-10-03 14:59:33,387][starVLA.training.train_starvla][INFO] - Step 1900, Loss: {'train/loss': 0.08447862416505814, 'train/grad_norm_pre_clip': 115.74708557128906, 'timing/data': 0.0003466890048002824, 'timing/model': 0.3690528709994396, 'learning_rate/qwen_vl_interface': 7.1727208798313704e-06, 'learning_rate/action_model': 7.172720879831371e-05, 'epoch': 0.14}) +[2026-10-03 14:59:34,045][starVLA.training.train_starvla][INFO] - Step 1901, Loss: {'train/loss': 0.39962896704673767, 'train/grad_norm_pre_clip': 38.43839645385742, 'timing/data': 0.00025131000438705087, 'timing/model': 0.3703766300022835, 'learning_rate/qwen_vl_interface': 7.169935713026994e-06, 'learning_rate/action_model': 7.169935713026994e-05, 'epoch': 0.14}) +[2026-10-03 14:59:34,702][starVLA.training.train_starvla][INFO] - Step 1902, Loss: {'train/loss': 0.00015758510562591255, 'train/grad_norm_pre_clip': 0.028999503701925278, 'timing/data': 0.00030870999034959823, 'timing/model': 0.36948067000776064, 'learning_rate/qwen_vl_interface': 7.1671497570095684e-06, 'learning_rate/action_model': 7.167149757009568e-05, 'epoch': 0.14}) +[2026-10-03 14:59:35,360][starVLA.training.train_starvla][INFO] - Step 1903, Loss: {'train/loss': 0.05996902287006378, 'train/grad_norm_pre_clip': 137.63949584960938, 'timing/data': 0.0003447399940341711, 'timing/model': 0.3692905000061728, 'learning_rate/qwen_vl_interface': 7.164363012924295e-06, 'learning_rate/action_model': 7.164363012924294e-05, 'epoch': 0.14}) +[2026-10-03 14:59:36,017][starVLA.training.train_starvla][INFO] - Step 1904, Loss: {'train/loss': 0.02376827783882618, 'train/grad_norm_pre_clip': 26.95624351501465, 'timing/data': 0.0002681799960555509, 'timing/model': 0.3695704199926695, 'learning_rate/qwen_vl_interface': 7.161575481916699e-06, 'learning_rate/action_model': 7.1615754819167e-05, 'epoch': 0.14}) +[2026-10-03 14:59:36,875][starVLA.training.train_starvla][INFO] - Step 1905, Loss: {'train/loss': 0.00017045318963937461, 'train/grad_norm_pre_clip': 0.029527388513088226, 'timing/data': 0.00026378000620752573, 'timing/model': 0.5694224269973347, 'learning_rate/qwen_vl_interface': 7.158787165132627e-06, 'learning_rate/action_model': 7.158787165132626e-05, 'epoch': 0.14}) +[2026-10-03 14:59:37,532][starVLA.training.train_starvla][INFO] - Step 1906, Loss: {'train/loss': 0.0001922147930599749, 'train/grad_norm_pre_clip': 0.022513698786497116, 'timing/data': 0.0003356099914526567, 'timing/model': 0.36919980999664403, 'learning_rate/qwen_vl_interface': 7.1559980637182535e-06, 'learning_rate/action_model': 7.155998063718254e-05, 'epoch': 0.14}) +[2026-10-03 14:59:38,191][starVLA.training.train_starvla][INFO] - Step 1907, Loss: {'train/loss': 0.01669568568468094, 'train/grad_norm_pre_clip': 98.8290786743164, 'timing/data': 0.00022238001110963523, 'timing/model': 0.36933378998946864, 'learning_rate/qwen_vl_interface': 7.153208178820071e-06, 'learning_rate/action_model': 7.153208178820072e-05, 'epoch': 0.14}) +[2026-10-03 14:59:38,847][starVLA.training.train_starvla][INFO] - Step 1908, Loss: {'train/loss': 0.00018525260384194553, 'train/grad_norm_pre_clip': 0.02283739298582077, 'timing/data': 0.0003407600015634671, 'timing/model': 0.36908045000745915, 'learning_rate/qwen_vl_interface': 7.150417511584898e-06, 'learning_rate/action_model': 7.150417511584898e-05, 'epoch': 0.14}) +[2026-10-03 14:59:39,507][starVLA.training.train_starvla][INFO] - Step 1909, Loss: {'train/loss': 0.08608204871416092, 'train/grad_norm_pre_clip': 37.540245056152344, 'timing/data': 0.00026343000354245305, 'timing/model': 0.37085072000627406, 'learning_rate/qwen_vl_interface': 7.147626063159872e-06, 'learning_rate/action_model': 7.147626063159872e-05, 'epoch': 0.14}) +[2026-10-03 14:59:40,357][starVLA.training.train_starvla][INFO] - Step 1910, Loss: {'train/loss': 0.0002862586115952581, 'train/grad_norm_pre_clip': 0.0700736939907074, 'timing/data': 0.00021195999579504132, 'timing/model': 0.5628200889914297, 'learning_rate/qwen_vl_interface': 7.14483383469245e-06, 'learning_rate/action_model': 7.144833834692449e-05, 'epoch': 0.14}) +[2026-10-03 14:59:41,016][starVLA.training.train_starvla][INFO] - Step 1911, Loss: {'train/loss': 0.00024251714057754725, 'train/grad_norm_pre_clip': 0.044503353536129, 'timing/data': 0.0002908300084527582, 'timing/model': 0.3699827599921264, 'learning_rate/qwen_vl_interface': 7.142040827330411e-06, 'learning_rate/action_model': 7.142040827330411e-05, 'epoch': 0.14}) +[2026-10-03 14:59:41,674][starVLA.training.train_starvla][INFO] - Step 1912, Loss: {'train/loss': 0.0003900897572748363, 'train/grad_norm_pre_clip': 0.06785544008016586, 'timing/data': 0.0002639199956320226, 'timing/model': 0.37052194999705534, 'learning_rate/qwen_vl_interface': 7.139247042221859e-06, 'learning_rate/action_model': 7.139247042221858e-05, 'epoch': 0.14}) +[2026-10-03 14:59:42,332][starVLA.training.train_starvla][INFO] - Step 1913, Loss: {'train/loss': 0.0005720501649193466, 'train/grad_norm_pre_clip': 0.11665327101945877, 'timing/data': 0.00023576901003252715, 'timing/model': 0.3706052110064775, 'learning_rate/qwen_vl_interface': 7.136452480515208e-06, 'learning_rate/action_model': 7.136452480515208e-05, 'epoch': 0.14}) +[2026-10-03 14:59:42,989][starVLA.training.train_starvla][INFO] - Step 1914, Loss: {'train/loss': 0.00034686806611716747, 'train/grad_norm_pre_clip': 0.09528903663158417, 'timing/data': 0.0003087999939452857, 'timing/model': 0.36944050100282766, 'learning_rate/qwen_vl_interface': 7.133657143359203e-06, 'learning_rate/action_model': 7.133657143359202e-05, 'epoch': 0.14}) +[2026-10-03 14:59:43,647][starVLA.training.train_starvla][INFO] - Step 1915, Loss: {'train/loss': 0.025839118286967278, 'train/grad_norm_pre_clip': 13.421911239624023, 'timing/data': 0.00021114999253768474, 'timing/model': 0.36952316100359894, 'learning_rate/qwen_vl_interface': 7.130861031902897e-06, 'learning_rate/action_model': 7.130861031902897e-05, 'epoch': 0.14}) +[2026-10-03 14:59:44,503][starVLA.training.train_starvla][INFO] - Step 1916, Loss: {'train/loss': 0.0005142742302268744, 'train/grad_norm_pre_clip': 0.19454263150691986, 'timing/data': 0.00030170999525580555, 'timing/model': 0.3701606999966316, 'learning_rate/qwen_vl_interface': 7.128064147295667e-06, 'learning_rate/action_model': 7.128064147295667e-05, 'epoch': 0.14}) +[2026-10-03 14:59:45,163][starVLA.training.train_starvla][INFO] - Step 1917, Loss: {'train/loss': 0.0031357784755527973, 'train/grad_norm_pre_clip': 1.3560774326324463, 'timing/data': 0.00031906999356579036, 'timing/model': 0.3715472089970717, 'learning_rate/qwen_vl_interface': 7.125266490687206e-06, 'learning_rate/action_model': 7.125266490687206e-05, 'epoch': 0.14}) +[2026-10-03 14:59:45,796][starVLA.training.train_starvla][INFO] - Step 1918, Loss: {'train/loss': 9.794544894248247e-05, 'train/grad_norm_pre_clip': 0.02227644808590412, 'timing/data': 0.00021596999431494623, 'timing/model': 0.3465867740014801, 'learning_rate/qwen_vl_interface': 7.1224680632275265e-06, 'learning_rate/action_model': 7.122468063227527e-05, 'epoch': 0.14}) +[2026-10-03 14:59:46,454][starVLA.training.train_starvla][INFO] - Step 1919, Loss: {'train/loss': 0.00017203643801622093, 'train/grad_norm_pre_clip': 0.02599545754492283, 'timing/data': 0.00021425000159069896, 'timing/model': 0.3693628500041086, 'learning_rate/qwen_vl_interface': 7.1196688660669555e-06, 'learning_rate/action_model': 7.119668866066955e-05, 'epoch': 0.14}) +[2026-10-03 14:59:47,113][starVLA.training.train_starvla][INFO] - Step 1920, Loss: {'train/loss': 0.0010129543952643871, 'train/grad_norm_pre_clip': 0.4325919449329376, 'timing/data': 0.0003590500127756968, 'timing/model': 0.37050085999362636, 'learning_rate/qwen_vl_interface': 7.116868900356138e-06, 'learning_rate/action_model': 7.116868900356138e-05, 'epoch': 0.14}) +[2026-10-03 14:59:47,971][starVLA.training.train_starvla][INFO] - Step 1921, Loss: {'train/loss': 0.003694199724122882, 'train/grad_norm_pre_clip': 6.227766990661621, 'timing/data': 0.0003401300054974854, 'timing/model': 0.5687842889892636, 'learning_rate/qwen_vl_interface': 7.114068167246033e-06, 'learning_rate/action_model': 7.114068167246033e-05, 'epoch': 0.14}) +[2026-10-03 14:59:48,628][starVLA.training.train_starvla][INFO] - Step 1922, Loss: {'train/loss': 0.00041406438685953617, 'train/grad_norm_pre_clip': 0.0925382673740387, 'timing/data': 0.0002448800078127533, 'timing/model': 0.3691866910085082, 'learning_rate/qwen_vl_interface': 7.111266667887916e-06, 'learning_rate/action_model': 7.111266667887915e-05, 'epoch': 0.14}) +[2026-10-03 14:59:49,262][starVLA.training.train_starvla][INFO] - Step 1923, Loss: {'train/loss': 0.30652734637260437, 'train/grad_norm_pre_clip': 54.30305099487305, 'timing/data': 0.0002703200007090345, 'timing/model': 0.36887217000185046, 'learning_rate/qwen_vl_interface': 7.108464403433378e-06, 'learning_rate/action_model': 7.108464403433378e-05, 'epoch': 0.14}) +[2026-10-03 14:59:49,920][starVLA.training.train_starvla][INFO] - Step 1924, Loss: {'train/loss': 0.045048788189888, 'train/grad_norm_pre_clip': 167.2637939453125, 'timing/data': 0.000358710007276386, 'timing/model': 0.3704532999981893, 'learning_rate/qwen_vl_interface': 7.105661375034324e-06, 'learning_rate/action_model': 7.105661375034324e-05, 'epoch': 0.14}) +[2026-10-03 14:59:50,580][starVLA.training.train_starvla][INFO] - Step 1925, Loss: {'train/loss': 0.00013409974053502083, 'train/grad_norm_pre_clip': 0.020147977396845818, 'timing/data': 0.0002616199926706031, 'timing/model': 0.37138096999842674, 'learning_rate/qwen_vl_interface': 7.102857583842973e-06, 'learning_rate/action_model': 7.102857583842974e-05, 'epoch': 0.14}) +[2026-10-03 14:59:51,435][starVLA.training.train_starvla][INFO] - Step 1926, Loss: {'train/loss': 0.00020885128469672054, 'train/grad_norm_pre_clip': 0.03036196157336235, 'timing/data': 0.00021643999207299203, 'timing/model': 0.5669390989933163, 'learning_rate/qwen_vl_interface': 7.100053031011859e-06, 'learning_rate/action_model': 7.100053031011859e-05, 'epoch': 0.14}) +[2026-10-03 14:59:52,092][starVLA.training.train_starvla][INFO] - Step 1927, Loss: {'train/loss': 0.4229162931442261, 'train/grad_norm_pre_clip': 88.72930908203125, 'timing/data': 0.0003492399991955608, 'timing/model': 0.3690156910015503, 'learning_rate/qwen_vl_interface': 7.097247717693827e-06, 'learning_rate/action_model': 7.097247717693827e-05, 'epoch': 0.14}) +[2026-10-03 14:59:52,751][starVLA.training.train_starvla][INFO] - Step 1928, Loss: {'train/loss': 0.00010821795149240643, 'train/grad_norm_pre_clip': 0.0268058143556118, 'timing/data': 0.00021754000044893473, 'timing/model': 0.370554599998286, 'learning_rate/qwen_vl_interface': 7.094441645042033e-06, 'learning_rate/action_model': 7.094441645042032e-05, 'epoch': 0.14}) +[2026-10-03 14:59:53,411][starVLA.training.train_starvla][INFO] - Step 1929, Loss: {'train/loss': 0.0001930681464727968, 'train/grad_norm_pre_clip': 0.06575045734643936, 'timing/data': 0.00037630999577231705, 'timing/model': 0.37091912000323646, 'learning_rate/qwen_vl_interface': 7.09163481420995e-06, 'learning_rate/action_model': 7.09163481420995e-05, 'epoch': 0.14}) +[2026-10-03 14:59:54,068][starVLA.training.train_starvla][INFO] - Step 1930, Loss: {'train/loss': 0.15291690826416016, 'train/grad_norm_pre_clip': 73.03413391113281, 'timing/data': 0.00022418999287765473, 'timing/model': 0.36952276100055315, 'learning_rate/qwen_vl_interface': 7.0888272263513595e-06, 'learning_rate/action_model': 7.088827226351359e-05, 'epoch': 0.14}) +[2026-10-03 14:59:54,726][starVLA.training.train_starvla][INFO] - Step 1931, Loss: {'train/loss': 7.613989873789251e-05, 'train/grad_norm_pre_clip': 0.012718812562525272, 'timing/data': 0.00021394000214058906, 'timing/model': 0.36928757099667564, 'learning_rate/qwen_vl_interface': 7.086018882620354e-06, 'learning_rate/action_model': 7.086018882620354e-05, 'epoch': 0.14}) +[2026-10-03 14:59:55,582][starVLA.training.train_starvla][INFO] - Step 1932, Loss: {'train/loss': 5.449101809062995e-05, 'train/grad_norm_pre_clip': 0.0076736025512218475, 'timing/data': 0.0002662799961399287, 'timing/model': 0.3709648100048071, 'learning_rate/qwen_vl_interface': 7.083209784171341e-06, 'learning_rate/action_model': 7.083209784171341e-05, 'epoch': 0.14}) +[2026-10-03 14:59:56,241][starVLA.training.train_starvla][INFO] - Step 1933, Loss: {'train/loss': 0.00011447109864093363, 'train/grad_norm_pre_clip': 0.029927829280495644, 'timing/data': 0.00021167000522837043, 'timing/model': 0.3706740399939008, 'learning_rate/qwen_vl_interface': 7.080399932159029e-06, 'learning_rate/action_model': 7.080399932159029e-05, 'epoch': 0.14}) +[2026-10-03 14:59:56,898][starVLA.training.train_starvla][INFO] - Step 1934, Loss: {'train/loss': 0.009684749878942966, 'train/grad_norm_pre_clip': 22.62002182006836, 'timing/data': 0.000298099999781698, 'timing/model': 0.36899061099393293, 'learning_rate/qwen_vl_interface': 7.077589327738448e-06, 'learning_rate/action_model': 7.077589327738447e-05, 'epoch': 0.14}) +[2026-10-03 14:59:57,556][starVLA.training.train_starvla][INFO] - Step 1935, Loss: {'train/loss': 0.43420761823654175, 'train/grad_norm_pre_clip': 290.2238464355469, 'timing/data': 0.000209039993933402, 'timing/model': 0.3688049909978872, 'learning_rate/qwen_vl_interface': 7.0747779720649265e-06, 'learning_rate/action_model': 7.074777972064927e-05, 'epoch': 0.14}) +[2026-10-03 14:59:58,213][starVLA.training.train_starvla][INFO] - Step 1936, Loss: {'train/loss': 0.0011293417774140835, 'train/grad_norm_pre_clip': 3.2992563247680664, 'timing/data': 0.0002654900017660111, 'timing/model': 0.37027107999892905, 'learning_rate/qwen_vl_interface': 7.071965866294108e-06, 'learning_rate/action_model': 7.071965866294108e-05, 'epoch': 0.14}) +[2026-10-03 14:59:59,074][starVLA.training.train_starvla][INFO] - Step 1937, Loss: {'train/loss': 0.0008989953785203397, 'train/grad_norm_pre_clip': 1.7332466840744019, 'timing/data': 0.0002740600029937923, 'timing/model': 0.3708077699993737, 'learning_rate/qwen_vl_interface': 7.0691530115819454e-06, 'learning_rate/action_model': 7.069153011581945e-05, 'epoch': 0.14}) +[2026-10-03 14:59:59,730][starVLA.training.train_starvla][INFO] - Step 1938, Loss: {'train/loss': 5.4275093134492636e-05, 'train/grad_norm_pre_clip': 0.009286419488489628, 'timing/data': 0.00021090000518597662, 'timing/model': 0.36871840999810956, 'learning_rate/qwen_vl_interface': 7.066339409084694e-06, 'learning_rate/action_model': 7.066339409084694e-05, 'epoch': 0.14}) +[2026-10-03 15:00:00,363][starVLA.training.train_starvla][INFO] - Step 1939, Loss: {'train/loss': 0.00021826406009495258, 'train/grad_norm_pre_clip': 0.09626167267560959, 'timing/data': 0.00021165001089684665, 'timing/model': 0.3684196310059633, 'learning_rate/qwen_vl_interface': 7.0635250599589215e-06, 'learning_rate/action_model': 7.06352505995892e-05, 'epoch': 0.14}) +[2026-10-03 15:00:01,021][starVLA.training.train_starvla][INFO] - Step 1940, Loss: {'train/loss': 3.46816741512157e-05, 'train/grad_norm_pre_clip': 0.0046515692956745625, 'timing/data': 0.0003193300071870908, 'timing/model': 0.3702348609949695, 'learning_rate/qwen_vl_interface': 7.060709965361499e-06, 'learning_rate/action_model': 7.060709965361499e-05, 'epoch': 0.14}) +[2026-10-03 15:00:01,680][starVLA.training.train_starvla][INFO] - Step 1941, Loss: {'train/loss': 0.4916946589946747, 'train/grad_norm_pre_clip': 45.214378356933594, 'timing/data': 0.0002665089996298775, 'timing/model': 0.3704221009975299, 'learning_rate/qwen_vl_interface': 7.057894126449606e-06, 'learning_rate/action_model': 7.057894126449606e-05, 'epoch': 0.14}) +[2026-10-03 15:00:02,538][starVLA.training.train_starvla][INFO] - Step 1942, Loss: {'train/loss': 0.31304627656936646, 'train/grad_norm_pre_clip': 64.62451171875, 'timing/data': 0.00021317000209819525, 'timing/model': 0.5707579580048332, 'learning_rate/qwen_vl_interface': 7.055077544380728e-06, 'learning_rate/action_model': 7.055077544380728e-05, 'epoch': 0.14}) +[2026-10-03 15:00:03,196][starVLA.training.train_starvla][INFO] - Step 1943, Loss: {'train/loss': 0.00013441607006825507, 'train/grad_norm_pre_clip': 0.02875955030322075, 'timing/data': 0.00042856000072788447, 'timing/model': 0.3693247409973992, 'learning_rate/qwen_vl_interface': 7.0522602203126554e-06, 'learning_rate/action_model': 7.052260220312656e-05, 'epoch': 0.14}) +[2026-10-03 15:00:03,854][starVLA.training.train_starvla][INFO] - Step 1944, Loss: {'train/loss': 0.1175009086728096, 'train/grad_norm_pre_clip': 147.68943786621094, 'timing/data': 0.00032649999775458127, 'timing/model': 0.3704756399965845, 'learning_rate/qwen_vl_interface': 7.0494421554034845e-06, 'learning_rate/action_model': 7.049442155403484e-05, 'epoch': 0.14}) +[2026-10-03 15:00:04,513][starVLA.training.train_starvla][INFO] - Step 1945, Loss: {'train/loss': 0.44018933176994324, 'train/grad_norm_pre_clip': 95.68631744384766, 'timing/data': 0.0003072499966947362, 'timing/model': 0.37026464099471923, 'learning_rate/qwen_vl_interface': 7.046623350811612e-06, 'learning_rate/action_model': 7.046623350811612e-05, 'epoch': 0.14}) +[2026-10-03 15:00:05,170][starVLA.training.train_starvla][INFO] - Step 1946, Loss: {'train/loss': 0.00030048604821786284, 'train/grad_norm_pre_clip': 0.3831654489040375, 'timing/data': 0.00029290000384207815, 'timing/model': 0.3693400810007006, 'learning_rate/qwen_vl_interface': 7.043803807695747e-06, 'learning_rate/action_model': 7.043803807695747e-05, 'epoch': 0.14}) +[2026-10-03 15:00:06,035][starVLA.training.train_starvla][INFO] - Step 1947, Loss: {'train/loss': 0.007929999381303787, 'train/grad_norm_pre_clip': 6.952559471130371, 'timing/data': 0.00032032000308390707, 'timing/model': 0.5767259479907807, 'learning_rate/qwen_vl_interface': 7.040983527214893e-06, 'learning_rate/action_model': 7.040983527214893e-05, 'epoch': 0.14}) +[2026-10-03 15:00:06,693][starVLA.training.train_starvla][INFO] - Step 1948, Loss: {'train/loss': 0.001955535029992461, 'train/grad_norm_pre_clip': 1.3991954326629639, 'timing/data': 0.00021116998686920851, 'timing/model': 0.37015403000987135, 'learning_rate/qwen_vl_interface': 7.038162510528363e-06, 'learning_rate/action_model': 7.038162510528362e-05, 'epoch': 0.14}) +[2026-10-03 15:00:07,351][starVLA.training.train_starvla][INFO] - Step 1949, Loss: {'train/loss': 0.04881752282381058, 'train/grad_norm_pre_clip': 38.68801498413086, 'timing/data': 0.0002870500029530376, 'timing/model': 0.3702300209988607, 'learning_rate/qwen_vl_interface': 7.03534075879577e-06, 'learning_rate/action_model': 7.03534075879577e-05, 'epoch': 0.14}) +[2026-10-03 15:00:08,008][starVLA.training.train_starvla][INFO] - Step 1950, Loss: {'train/loss': 8.26218820293434e-05, 'train/grad_norm_pre_clip': 0.021200472488999367, 'timing/data': 0.0002166399935958907, 'timing/model': 0.3696056310000131, 'learning_rate/qwen_vl_interface': 7.032518273177028e-06, 'learning_rate/action_model': 7.032518273177028e-05, 'epoch': 0.14}) +[2026-10-03 15:00:08,667][starVLA.training.train_starvla][INFO] - Step 1951, Loss: {'train/loss': 0.15939442813396454, 'train/grad_norm_pre_clip': 77.20166778564453, 'timing/data': 0.00021001999266445637, 'timing/model': 0.37005639100971166, 'learning_rate/qwen_vl_interface': 7.029695054832357e-06, 'learning_rate/action_model': 7.029695054832357e-05, 'epoch': 0.14}) +[2026-10-03 15:00:09,302][starVLA.training.train_starvla][INFO] - Step 1952, Loss: {'train/loss': 0.3453081548213959, 'train/grad_norm_pre_clip': 101.1419448852539, 'timing/data': 0.0003448899951763451, 'timing/model': 0.3474511239910498, 'learning_rate/qwen_vl_interface': 7.026871104922273e-06, 'learning_rate/action_model': 7.026871104922273e-05, 'epoch': 0.14}) +[2026-10-03 15:00:10,159][starVLA.training.train_starvla][INFO] - Step 1953, Loss: {'train/loss': 0.48181384801864624, 'train/grad_norm_pre_clip': 25.209089279174805, 'timing/data': 0.00022343000455293804, 'timing/model': 0.3711574909975752, 'learning_rate/qwen_vl_interface': 7.024046424607596e-06, 'learning_rate/action_model': 7.024046424607595e-05, 'epoch': 0.14}) +[2026-10-03 15:00:10,818][starVLA.training.train_starvla][INFO] - Step 1954, Loss: {'train/loss': 0.0037718622479587793, 'train/grad_norm_pre_clip': 8.385114669799805, 'timing/data': 0.00022868000087328255, 'timing/model': 0.3704047899955185, 'learning_rate/qwen_vl_interface': 7.021221015049445e-06, 'learning_rate/action_model': 7.021221015049446e-05, 'epoch': 0.14}) +[2026-10-03 15:00:11,476][starVLA.training.train_starvla][INFO] - Step 1955, Loss: {'train/loss': 8.976243407232687e-05, 'train/grad_norm_pre_clip': 0.0313560888171196, 'timing/data': 0.00021336000645533204, 'timing/model': 0.36951846099691465, 'learning_rate/qwen_vl_interface': 7.01839487740924e-06, 'learning_rate/action_model': 7.01839487740924e-05, 'epoch': 0.14}) +[2026-10-03 15:00:12,135][starVLA.training.train_starvla][INFO] - Step 1956, Loss: {'train/loss': 0.23549111187458038, 'train/grad_norm_pre_clip': 102.82516479492188, 'timing/data': 0.00026111000624950975, 'timing/model': 0.3708543609973276, 'learning_rate/qwen_vl_interface': 7.015568012848699e-06, 'learning_rate/action_model': 7.015568012848699e-05, 'epoch': 0.14}) +[2026-10-03 15:00:12,796][starVLA.training.train_starvla][INFO] - Step 1957, Loss: {'train/loss': 0.0017590899951756, 'train/grad_norm_pre_clip': 1.5685086250305176, 'timing/data': 0.0003291500033810735, 'timing/model': 0.37160607099940535, 'learning_rate/qwen_vl_interface': 7.01274042252984e-06, 'learning_rate/action_model': 7.01274042252984e-05, 'epoch': 0.14}) +[2026-10-03 15:00:13,649][starVLA.training.train_starvla][INFO] - Step 1958, Loss: {'train/loss': 0.00027498588315211236, 'train/grad_norm_pre_clip': 0.053522996604442596, 'timing/data': 0.0002128490014001727, 'timing/model': 0.5655871100025252, 'learning_rate/qwen_vl_interface': 7.009912107614978e-06, 'learning_rate/action_model': 7.009912107614978e-05, 'epoch': 0.14}) +[2026-10-03 15:00:14,308][starVLA.training.train_starvla][INFO] - Step 1959, Loss: {'train/loss': 0.0013099168427288532, 'train/grad_norm_pre_clip': 2.5838334560394287, 'timing/data': 0.00021564999769907445, 'timing/model': 0.3696887609985424, 'learning_rate/qwen_vl_interface': 7.0070830692667244e-06, 'learning_rate/action_model': 7.007083069266724e-05, 'epoch': 0.14}) +[2026-10-03 15:00:14,966][starVLA.training.train_starvla][INFO] - Step 1960, Loss: {'train/loss': 0.0007938435301184654, 'train/grad_norm_pre_clip': 2.360734462738037, 'timing/data': 0.00036489000194706023, 'timing/model': 0.37075775100674946, 'learning_rate/qwen_vl_interface': 7.004253308647993e-06, 'learning_rate/action_model': 7.004253308647994e-05, 'epoch': 0.14}) +[2026-10-03 15:00:15,625][starVLA.training.train_starvla][INFO] - Step 1961, Loss: {'train/loss': 0.00016129808500409126, 'train/grad_norm_pre_clip': 0.06376569718122482, 'timing/data': 0.0003723700065165758, 'timing/model': 0.370466121006757, 'learning_rate/qwen_vl_interface': 7.001422826921991e-06, 'learning_rate/action_model': 7.00142282692199e-05, 'epoch': 0.14}) +[2026-10-03 15:00:16,282][starVLA.training.train_starvla][INFO] - Step 1962, Loss: {'train/loss': 0.47911590337753296, 'train/grad_norm_pre_clip': 52.50507736206055, 'timing/data': 0.00022415998682845384, 'timing/model': 0.3689826209883904, 'learning_rate/qwen_vl_interface': 6.9985916252522225e-06, 'learning_rate/action_model': 6.998591625252222e-05, 'epoch': 0.14}) +[2026-10-03 15:00:16,940][starVLA.training.train_starvla][INFO] - Step 1963, Loss: {'train/loss': 0.0012206172104924917, 'train/grad_norm_pre_clip': 0.31960874795913696, 'timing/data': 0.00022926999372430146, 'timing/model': 0.36903989099664614, 'learning_rate/qwen_vl_interface': 6.995759704802485e-06, 'learning_rate/action_model': 6.995759704802485e-05, 'epoch': 0.14}) +[2026-10-03 15:00:17,795][starVLA.training.train_starvla][INFO] - Step 1964, Loss: {'train/loss': 0.00011498195817694068, 'train/grad_norm_pre_clip': 0.030652662739157677, 'timing/data': 0.00027478000265546143, 'timing/model': 0.37123419099953026, 'learning_rate/qwen_vl_interface': 6.992927066736876e-06, 'learning_rate/action_model': 6.992927066736875e-05, 'epoch': 0.14}) +[2026-10-03 15:00:18,454][starVLA.training.train_starvla][INFO] - Step 1965, Loss: {'train/loss': 7.954025932122022e-05, 'train/grad_norm_pre_clip': 0.019545817747712135, 'timing/data': 0.00026298000011593103, 'timing/model': 0.37094246000924613, 'learning_rate/qwen_vl_interface': 6.990093712219787e-06, 'learning_rate/action_model': 6.990093712219787e-05, 'epoch': 0.14}) +[2026-10-03 15:00:19,112][starVLA.training.train_starvla][INFO] - Step 1966, Loss: {'train/loss': 0.0003956672444473952, 'train/grad_norm_pre_clip': 0.09147066622972488, 'timing/data': 0.00026477999927010387, 'timing/model': 0.36943082000652794, 'learning_rate/qwen_vl_interface': 6.9872596424159e-06, 'learning_rate/action_model': 6.9872596424159e-05, 'epoch': 0.14}) +[2026-10-03 15:00:19,770][starVLA.training.train_starvla][INFO] - Step 1967, Loss: {'train/loss': 2.7349631636752747e-05, 'train/grad_norm_pre_clip': 0.0042672972194850445, 'timing/data': 0.00021236900647636503, 'timing/model': 0.36971107200952247, 'learning_rate/qwen_vl_interface': 6.9844248584902e-06, 'learning_rate/action_model': 6.984424858490199e-05, 'epoch': 0.14}) +[2026-10-03 15:00:20,428][starVLA.training.train_starvla][INFO] - Step 1968, Loss: {'train/loss': 0.002748732455074787, 'train/grad_norm_pre_clip': 2.8197643756866455, 'timing/data': 0.00033352999889757484, 'timing/model': 0.37036083100247197, 'learning_rate/qwen_vl_interface': 6.981589361607952e-06, 'learning_rate/action_model': 6.981589361607953e-05, 'epoch': 0.14}) +[2026-10-03 15:00:21,283][starVLA.training.train_starvla][INFO] - Step 1969, Loss: {'train/loss': 0.0001501395890954882, 'train/grad_norm_pre_clip': 0.35696226358413696, 'timing/data': 0.00022143000387586653, 'timing/model': 0.3705336209968664, 'learning_rate/qwen_vl_interface': 6.9787531529347274e-06, 'learning_rate/action_model': 6.978753152934727e-05, 'epoch': 0.14}) +[2026-10-03 15:00:21,918][starVLA.training.train_starvla][INFO] - Step 1970, Loss: {'train/loss': 0.00021678005577996373, 'train/grad_norm_pre_clip': 0.07259348034858704, 'timing/data': 0.0002236400032415986, 'timing/model': 0.34631391399307176, 'learning_rate/qwen_vl_interface': 6.97591623363638e-06, 'learning_rate/action_model': 6.97591623363638e-05, 'epoch': 0.14}) +[2026-10-03 15:00:22,576][starVLA.training.train_starvla][INFO] - Step 1971, Loss: {'train/loss': 4.484935197979212e-05, 'train/grad_norm_pre_clip': 0.007810917217284441, 'timing/data': 0.000214998988667503, 'timing/model': 0.3691848519956693, 'learning_rate/qwen_vl_interface': 6.9730786048790644e-06, 'learning_rate/action_model': 6.973078604879064e-05, 'epoch': 0.14}) +[2026-10-03 15:00:23,235][starVLA.training.train_starvla][INFO] - Step 1972, Loss: {'train/loss': 0.00021928161731921136, 'train/grad_norm_pre_clip': 0.08961119502782822, 'timing/data': 0.0002203900075983256, 'timing/model': 0.3714450709958328, 'learning_rate/qwen_vl_interface': 6.970240267829221e-06, 'learning_rate/action_model': 6.970240267829222e-05, 'epoch': 0.14}) +[2026-10-03 15:00:23,894][starVLA.training.train_starvla][INFO] - Step 1973, Loss: {'train/loss': 3.724260022863746e-05, 'train/grad_norm_pre_clip': 0.011192576959729195, 'timing/data': 0.00026374000299256295, 'timing/model': 0.37092548099462874, 'learning_rate/qwen_vl_interface': 6.967401223653582e-06, 'learning_rate/action_model': 6.967401223653581e-05, 'epoch': 0.14}) +[2026-10-03 15:00:24,750][starVLA.training.train_starvla][INFO] - Step 1974, Loss: {'train/loss': 0.0002969157649204135, 'train/grad_norm_pre_clip': 0.09035190939903259, 'timing/data': 0.00030425000295508653, 'timing/model': 0.5665887999930419, 'learning_rate/qwen_vl_interface': 6.9645614735191724e-06, 'learning_rate/action_model': 6.964561473519172e-05, 'epoch': 0.14}) +[2026-10-03 15:00:25,408][starVLA.training.train_starvla][INFO] - Step 1975, Loss: {'train/loss': 0.0003391356731299311, 'train/grad_norm_pre_clip': 0.15933850407600403, 'timing/data': 0.00021065000328235328, 'timing/model': 0.3693868310074322, 'learning_rate/qwen_vl_interface': 6.961721018593305e-06, 'learning_rate/action_model': 6.961721018593305e-05, 'epoch': 0.14}) +[2026-10-03 15:00:26,042][starVLA.training.train_starvla][INFO] - Step 1976, Loss: {'train/loss': 0.6337882876396179, 'train/grad_norm_pre_clip': 50.67327880859375, 'timing/data': 0.00045028999738860875, 'timing/model': 0.34744003500964027, 'learning_rate/qwen_vl_interface': 6.958879860043585e-06, 'learning_rate/action_model': 6.958879860043585e-05, 'epoch': 0.14}) +[2026-10-03 15:00:26,704][starVLA.training.train_starvla][INFO] - Step 1977, Loss: {'train/loss': 0.6538445353507996, 'train/grad_norm_pre_clip': 28.7152156829834, 'timing/data': 0.00021581999317277223, 'timing/model': 0.3726330610079458, 'learning_rate/qwen_vl_interface': 6.9560379990379025e-06, 'learning_rate/action_model': 6.956037999037902e-05, 'epoch': 0.14}) +[2026-10-03 15:00:27,366][starVLA.training.train_starvla][INFO] - Step 1978, Loss: {'train/loss': 0.011613136157393456, 'train/grad_norm_pre_clip': 28.805204391479492, 'timing/data': 0.000222250004298985, 'timing/model': 0.37172349100001156, 'learning_rate/qwen_vl_interface': 6.953195436744442e-06, 'learning_rate/action_model': 6.953195436744442e-05, 'epoch': 0.14}) +[2026-10-03 15:00:28,220][starVLA.training.train_starvla][INFO] - Step 1979, Loss: {'train/loss': 0.6017383933067322, 'train/grad_norm_pre_clip': 47.099212646484375, 'timing/data': 0.00020978999964427203, 'timing/model': 0.5636122800060548, 'learning_rate/qwen_vl_interface': 6.950352174331676e-06, 'learning_rate/action_model': 6.950352174331675e-05, 'epoch': 0.14}) +[2026-10-03 15:00:28,878][starVLA.training.train_starvla][INFO] - Step 1980, Loss: {'train/loss': 9.182705980492756e-05, 'train/grad_norm_pre_clip': 0.029726235195994377, 'timing/data': 0.00036981000448577106, 'timing/model': 0.3712846400012495, 'learning_rate/qwen_vl_interface': 6.947508212968356e-06, 'learning_rate/action_model': 6.947508212968356e-05, 'epoch': 0.14}) +[2026-10-03 15:00:29,539][starVLA.training.train_starvla][INFO] - Step 1981, Loss: {'train/loss': 5.5525157222291455e-05, 'train/grad_norm_pre_clip': 0.009615673683583736, 'timing/data': 0.00021346000721678138, 'timing/model': 0.37197115100570954, 'learning_rate/qwen_vl_interface': 6.9446635538235295e-06, 'learning_rate/action_model': 6.944663553823529e-05, 'epoch': 0.14}) +[2026-10-03 15:00:30,199][starVLA.training.train_starvla][INFO] - Step 1982, Loss: {'train/loss': 0.00038855071761645377, 'train/grad_norm_pre_clip': 0.22543172538280487, 'timing/data': 0.00043419000576250255, 'timing/model': 0.36965450100251473, 'learning_rate/qwen_vl_interface': 6.94181819806653e-06, 'learning_rate/action_model': 6.94181819806653e-05, 'epoch': 0.15}) +[2026-10-03 15:00:30,861][starVLA.training.train_starvla][INFO] - Step 1983, Loss: {'train/loss': 9.357989620184526e-05, 'train/grad_norm_pre_clip': 0.03080524504184723, 'timing/data': 0.00022128999989945441, 'timing/model': 0.37193256099999417, 'learning_rate/qwen_vl_interface': 6.938972146866975e-06, 'learning_rate/action_model': 6.938972146866974e-05, 'epoch': 0.15}) +[2026-10-03 15:00:31,726][starVLA.training.train_starvla][INFO] - Step 1984, Loss: {'train/loss': 0.0002712535497266799, 'train/grad_norm_pre_clip': 0.0973573699593544, 'timing/data': 0.0002186399942729622, 'timing/model': 0.5750946390035097, 'learning_rate/qwen_vl_interface': 6.936125401394769e-06, 'learning_rate/action_model': 6.936125401394768e-05, 'epoch': 0.15}) +[2026-10-03 15:00:32,384][starVLA.training.train_starvla][INFO] - Step 1985, Loss: {'train/loss': 3.574623406166211e-05, 'train/grad_norm_pre_clip': 0.0065461378544569016, 'timing/data': 0.0002107499894918874, 'timing/model': 0.36995422100881115, 'learning_rate/qwen_vl_interface': 6.933277962820101e-06, 'learning_rate/action_model': 6.9332779628201e-05, 'epoch': 0.15}) +[2026-10-03 15:00:33,041][starVLA.training.train_starvla][INFO] - Step 1986, Loss: {'train/loss': 4.652679854189046e-05, 'train/grad_norm_pre_clip': 0.00922547746449709, 'timing/data': 0.00038449998828582466, 'timing/model': 0.36910400200576987, 'learning_rate/qwen_vl_interface': 6.930429832313444e-06, 'learning_rate/action_model': 6.930429832313444e-05, 'epoch': 0.15}) +[2026-10-03 15:00:33,699][starVLA.training.train_starvla][INFO] - Step 1987, Loss: {'train/loss': 8.056801743805408e-05, 'train/grad_norm_pre_clip': 0.01474852580577135, 'timing/data': 0.00021035000099800527, 'timing/model': 0.3689544919907348, 'learning_rate/qwen_vl_interface': 6.92758101104556e-06, 'learning_rate/action_model': 6.92758101104556e-05, 'epoch': 0.15}) +[2026-10-03 15:00:34,356][starVLA.training.train_starvla][INFO] - Step 1988, Loss: {'train/loss': 0.0007459975313395262, 'train/grad_norm_pre_clip': 0.5380946397781372, 'timing/data': 0.0003964200004702434, 'timing/model': 0.37010321200068574, 'learning_rate/qwen_vl_interface': 6.9247315001874885e-06, 'learning_rate/action_model': 6.924731500187488e-05, 'epoch': 0.15}) +[2026-10-03 15:00:35,016][starVLA.training.train_starvla][INFO] - Step 1989, Loss: {'train/loss': 0.10050494223833084, 'train/grad_norm_pre_clip': 185.75123596191406, 'timing/data': 0.00035849001142196357, 'timing/model': 0.37071448100323323, 'learning_rate/qwen_vl_interface': 6.9218813009105605e-06, 'learning_rate/action_model': 6.92188130091056e-05, 'epoch': 0.15}) +[2026-10-03 15:00:35,880][starVLA.training.train_starvla][INFO] - Step 1990, Loss: {'train/loss': 3.246100459364243e-05, 'train/grad_norm_pre_clip': 0.006656613200902939, 'timing/data': 0.0002256599982501939, 'timing/model': 0.37119120100396685, 'learning_rate/qwen_vl_interface': 6.919030414386378e-06, 'learning_rate/action_model': 6.919030414386379e-05, 'epoch': 0.15}) +[2026-10-03 15:00:36,541][starVLA.training.train_starvla][INFO] - Step 1991, Loss: {'train/loss': 0.410498708486557, 'train/grad_norm_pre_clip': 38.72602081298828, 'timing/data': 0.0002133899979526177, 'timing/model': 0.37053328099136706, 'learning_rate/qwen_vl_interface': 6.916178841786839e-06, 'learning_rate/action_model': 6.916178841786838e-05, 'epoch': 0.15}) +[2026-10-03 15:00:37,201][starVLA.training.train_starvla][INFO] - Step 1992, Loss: {'train/loss': 0.3662942945957184, 'train/grad_norm_pre_clip': 39.56359100341797, 'timing/data': 0.0003197399928467348, 'timing/model': 0.3716493120009545, 'learning_rate/qwen_vl_interface': 6.913326584284115e-06, 'learning_rate/action_model': 6.913326584284115e-05, 'epoch': 0.15}) +[2026-10-03 15:00:37,863][starVLA.training.train_starvla][INFO] - Step 1993, Loss: {'train/loss': 0.00025068738614208996, 'train/grad_norm_pre_clip': 0.12103714048862457, 'timing/data': 0.00026722998882178217, 'timing/model': 0.3722119909944013, 'learning_rate/qwen_vl_interface': 6.910473643050662e-06, 'learning_rate/action_model': 6.910473643050662e-05, 'epoch': 0.15}) +[2026-10-03 15:00:38,524][starVLA.training.train_starvla][INFO] - Step 1994, Loss: {'train/loss': 0.00012630138371605426, 'train/grad_norm_pre_clip': 0.0955733209848404, 'timing/data': 0.0002230500103905797, 'timing/model': 0.3713344510033494, 'learning_rate/qwen_vl_interface': 6.907620019259215e-06, 'learning_rate/action_model': 6.907620019259216e-05, 'epoch': 0.15}) +[2026-10-03 15:00:39,386][starVLA.training.train_starvla][INFO] - Step 1995, Loss: {'train/loss': 0.0002092289796564728, 'train/grad_norm_pre_clip': 0.11521916091442108, 'timing/data': 0.0003748699964489788, 'timing/model': 0.37115781199827325, 'learning_rate/qwen_vl_interface': 6.9047657140827935e-06, 'learning_rate/action_model': 6.904765714082794e-05, 'epoch': 0.15}) +[2026-10-03 15:00:40,048][starVLA.training.train_starvla][INFO] - Step 1996, Loss: {'train/loss': 0.0002040917461272329, 'train/grad_norm_pre_clip': 0.06429837644100189, 'timing/data': 0.000358239994966425, 'timing/model': 0.3736311210086569, 'learning_rate/qwen_vl_interface': 6.901910728694694e-06, 'learning_rate/action_model': 6.901910728694694e-05, 'epoch': 0.15}) +[2026-10-03 15:00:40,686][starVLA.training.train_starvla][INFO] - Step 1997, Loss: {'train/loss': 0.0019029908580705523, 'train/grad_norm_pre_clip': 3.7354533672332764, 'timing/data': 0.00035342000774107873, 'timing/model': 0.34903384500648826, 'learning_rate/qwen_vl_interface': 6.899055064268489e-06, 'learning_rate/action_model': 6.899055064268489e-05, 'epoch': 0.15}) +[2026-10-03 15:00:41,347][starVLA.training.train_starvla][INFO] - Step 1998, Loss: {'train/loss': 0.592175304889679, 'train/grad_norm_pre_clip': 18.255008697509766, 'timing/data': 0.0002968600019812584, 'timing/model': 0.37137762100610416, 'learning_rate/qwen_vl_interface': 6.896198721978038e-06, 'learning_rate/action_model': 6.896198721978038e-05, 'epoch': 0.15}) +[2026-10-03 15:00:42,010][starVLA.training.train_starvla][INFO] - Step 1999, Loss: {'train/loss': 0.0003174595767632127, 'train/grad_norm_pre_clip': 0.227036714553833, 'timing/data': 0.00022452899429481477, 'timing/model': 0.372111112010316, 'learning_rate/qwen_vl_interface': 6.893341702997475e-06, 'learning_rate/action_model': 6.893341702997475e-05, 'epoch': 0.15}) +[2026-10-03 15:00:42,870][starVLA.training.train_starvla][INFO] - Step 2000, Loss: {'train/loss': 0.056965555995702744, 'train/grad_norm_pre_clip': 27.39312171936035, 'timing/data': 0.00034347000473644584, 'timing/model': 0.5696370700025, 'learning_rate/qwen_vl_interface': 6.890484008501211e-06, 'learning_rate/action_model': 6.890484008501211e-05, 'epoch': 0.15}) +[2026-10-03 15:00:43,533][starVLA.training.train_starvla][INFO] - Step 2001, Loss: {'train/loss': 0.005898703821003437, 'train/grad_norm_pre_clip': 6.781313419342041, 'timing/data': 0.00036484000156633556, 'timing/model': 0.37314399098977447, 'learning_rate/qwen_vl_interface': 6.887625639663941e-06, 'learning_rate/action_model': 6.88762563966394e-05, 'epoch': 0.15}) +[2026-10-03 15:00:44,195][starVLA.training.train_starvla][INFO] - Step 2002, Loss: {'train/loss': 0.0007154685445129871, 'train/grad_norm_pre_clip': 0.29752281308174133, 'timing/data': 0.00022871000692248344, 'timing/model': 0.371311491006054, 'learning_rate/qwen_vl_interface': 6.884766597660626e-06, 'learning_rate/action_model': 6.884766597660626e-05, 'epoch': 0.15}) +[2026-10-03 15:00:44,834][starVLA.training.train_starvla][INFO] - Step 2003, Loss: {'train/loss': 9.305265848524868e-05, 'train/grad_norm_pre_clip': 0.015574011951684952, 'timing/data': 0.00023440000950358808, 'timing/model': 0.37134254099510144, 'learning_rate/qwen_vl_interface': 6.881906883666515e-06, 'learning_rate/action_model': 6.881906883666515e-05, 'epoch': 0.15}) +[2026-10-03 15:00:45,508][starVLA.training.train_starvla][INFO] - Step 2004, Loss: {'train/loss': 0.0001540172379463911, 'train/grad_norm_pre_clip': 0.028917118906974792, 'timing/data': 0.00034378000418655574, 'timing/model': 0.38409130000218283, 'learning_rate/qwen_vl_interface': 6.879046498857125e-06, 'learning_rate/action_model': 6.879046498857124e-05, 'epoch': 0.15}) +[2026-10-03 15:00:46,364][starVLA.training.train_starvla][INFO] - Step 2005, Loss: {'train/loss': 0.0012397097889333963, 'train/grad_norm_pre_clip': 0.8567128777503967, 'timing/data': 0.0003464900073595345, 'timing/model': 0.5672960609954316, 'learning_rate/qwen_vl_interface': 6.876185444408255e-06, 'learning_rate/action_model': 6.876185444408255e-05, 'epoch': 0.15}) +[2026-10-03 15:00:46,999][starVLA.training.train_starvla][INFO] - Step 2006, Loss: {'train/loss': 0.0002070441551040858, 'train/grad_norm_pre_clip': 0.07914555817842484, 'timing/data': 0.00023006000265013427, 'timing/model': 0.3465752850024728, 'learning_rate/qwen_vl_interface': 6.873323721495976e-06, 'learning_rate/action_model': 6.873323721495976e-05, 'epoch': 0.15}) +[2026-10-03 15:00:47,634][starVLA.training.train_starvla][INFO] - Step 2007, Loss: {'train/loss': 0.00029755907598882914, 'train/grad_norm_pre_clip': 0.38717442750930786, 'timing/data': 0.00021317999926395714, 'timing/model': 0.36950614101078827, 'learning_rate/qwen_vl_interface': 6.870461331296632e-06, 'learning_rate/action_model': 6.870461331296632e-05, 'epoch': 0.15}) +[2026-10-03 15:00:48,271][starVLA.training.train_starvla][INFO] - Step 2008, Loss: {'train/loss': 0.07340461015701294, 'train/grad_norm_pre_clip': 74.76914978027344, 'timing/data': 0.00022493000142276287, 'timing/model': 0.37098519199935254, 'learning_rate/qwen_vl_interface': 6.8675982749868455e-06, 'learning_rate/action_model': 6.867598274986845e-05, 'epoch': 0.15}) +[2026-10-03 15:00:48,929][starVLA.training.train_starvla][INFO] - Step 2009, Loss: {'train/loss': 0.00019073164730798453, 'train/grad_norm_pre_clip': 0.05203831568360329, 'timing/data': 0.00031737999233882874, 'timing/model': 0.3707787419989472, 'learning_rate/qwen_vl_interface': 6.864734553743509e-06, 'learning_rate/action_model': 6.86473455374351e-05, 'epoch': 0.15}) +[2026-10-03 15:00:49,589][starVLA.training.train_starvla][INFO] - Step 2010, Loss: {'train/loss': 6.281746027525514e-05, 'train/grad_norm_pre_clip': 0.02079182118177414, 'timing/data': 0.00022243999410420656, 'timing/model': 0.3713746319990605, 'learning_rate/qwen_vl_interface': 6.861870168743793e-06, 'learning_rate/action_model': 6.861870168743793e-05, 'epoch': 0.15}) +[2026-10-03 15:00:50,450][starVLA.training.train_starvla][INFO] - Step 2011, Loss: {'train/loss': 0.0013754558749496937, 'train/grad_norm_pre_clip': 1.7724577188491821, 'timing/data': 0.00021903999731875956, 'timing/model': 0.36970032200042624, 'learning_rate/qwen_vl_interface': 6.859005121165135e-06, 'learning_rate/action_model': 6.859005121165135e-05, 'epoch': 0.15}) +[2026-10-03 15:00:51,109][starVLA.training.train_starvla][INFO] - Step 2012, Loss: {'train/loss': 4.3008491047658026e-05, 'train/grad_norm_pre_clip': 0.006057374179363251, 'timing/data': 0.00031936999585013837, 'timing/model': 0.37052282100194134, 'learning_rate/qwen_vl_interface': 6.85613941218525e-06, 'learning_rate/action_model': 6.85613941218525e-05, 'epoch': 0.15}) +[2026-10-03 15:00:51,770][starVLA.training.train_starvla][INFO] - Step 2013, Loss: {'train/loss': 0.00036151381209492683, 'train/grad_norm_pre_clip': 0.11900730431079865, 'timing/data': 0.00022007999359630048, 'timing/model': 0.37264417100232095, 'learning_rate/qwen_vl_interface': 6.8532730429821215e-06, 'learning_rate/action_model': 6.853273042982122e-05, 'epoch': 0.15}) +[2026-10-03 15:00:52,428][starVLA.training.train_starvla][INFO] - Step 2014, Loss: {'train/loss': 0.0002936512464657426, 'train/grad_norm_pre_clip': 0.10327839106321335, 'timing/data': 0.00022898000315763056, 'timing/model': 0.3699258410051698, 'learning_rate/qwen_vl_interface': 6.850406014734005e-06, 'learning_rate/action_model': 6.850406014734005e-05, 'epoch': 0.15}) +[2026-10-03 15:00:53,087][starVLA.training.train_starvla][INFO] - Step 2015, Loss: {'train/loss': 0.02840007282793522, 'train/grad_norm_pre_clip': 21.62594985961914, 'timing/data': 0.00025054000434465706, 'timing/model': 0.36938663199543953, 'learning_rate/qwen_vl_interface': 6.8475383286194295e-06, 'learning_rate/action_model': 6.84753832861943e-05, 'epoch': 0.15}) +[2026-10-03 15:00:53,946][starVLA.training.train_starvla][INFO] - Step 2016, Loss: {'train/loss': 0.1778133660554886, 'train/grad_norm_pre_clip': 300.3623352050781, 'timing/data': 0.0002603600005386397, 'timing/model': 0.5714015699923038, 'learning_rate/qwen_vl_interface': 6.844669985817192e-06, 'learning_rate/action_model': 6.844669985817192e-05, 'epoch': 0.15}) +[2026-10-03 15:00:54,605][starVLA.training.train_starvla][INFO] - Step 2017, Loss: {'train/loss': 3.2832900615176186e-05, 'train/grad_norm_pre_clip': 0.00891165528446436, 'timing/data': 0.00021162998746149242, 'timing/model': 0.370550962004927, 'learning_rate/qwen_vl_interface': 6.8418009875063594e-06, 'learning_rate/action_model': 6.84180098750636e-05, 'epoch': 0.15}) +[2026-10-03 15:00:55,264][starVLA.training.train_starvla][INFO] - Step 2018, Loss: {'train/loss': 0.24479281902313232, 'train/grad_norm_pre_clip': 67.76261138916016, 'timing/data': 0.00022577001072932035, 'timing/model': 0.36985954199917614, 'learning_rate/qwen_vl_interface': 6.838931334866267e-06, 'learning_rate/action_model': 6.838931334866266e-05, 'epoch': 0.15}) +[2026-10-03 15:00:55,923][starVLA.training.train_starvla][INFO] - Step 2019, Loss: {'train/loss': 0.00012468511704355478, 'train/grad_norm_pre_clip': 0.025616219267249107, 'timing/data': 0.00022200999956112355, 'timing/model': 0.3699998920055805, 'learning_rate/qwen_vl_interface': 6.836061029076523e-06, 'learning_rate/action_model': 6.836061029076522e-05, 'epoch': 0.15}) +[2026-10-03 15:00:56,581][starVLA.training.train_starvla][INFO] - Step 2020, Loss: {'train/loss': 6.67011336190626e-05, 'train/grad_norm_pre_clip': 0.014536157250404358, 'timing/data': 0.000215430001844652, 'timing/model': 0.37070105101156514, 'learning_rate/qwen_vl_interface': 6.833190071317e-06, 'learning_rate/action_model': 6.833190071316999e-05, 'epoch': 0.15}) +[2026-10-03 15:00:57,434][starVLA.training.train_starvla][INFO] - Step 2021, Loss: {'train/loss': 2.9225826438050717e-05, 'train/grad_norm_pre_clip': 0.0117563521489501, 'timing/data': 0.00022074999287724495, 'timing/model': 0.5658415799989598, 'learning_rate/qwen_vl_interface': 6.830318462767841e-06, 'learning_rate/action_model': 6.830318462767841e-05, 'epoch': 0.15}) +[2026-10-03 15:00:58,095][starVLA.training.train_starvla][INFO] - Step 2022, Loss: {'train/loss': 7.252497744048014e-05, 'train/grad_norm_pre_clip': 0.01487506553530693, 'timing/data': 0.0003711900062626228, 'timing/model': 0.3695890519884415, 'learning_rate/qwen_vl_interface': 6.827446204609456e-06, 'learning_rate/action_model': 6.827446204609456e-05, 'epoch': 0.15}) +[2026-10-03 15:00:58,751][starVLA.training.train_starvla][INFO] - Step 2023, Loss: {'train/loss': 4.142528268857859e-05, 'train/grad_norm_pre_clip': 0.006634765770286322, 'timing/data': 0.0002099900011671707, 'timing/model': 0.36925170199538115, 'learning_rate/qwen_vl_interface': 6.824573298022522e-06, 'learning_rate/action_model': 6.824573298022522e-05, 'epoch': 0.15}) +[2026-10-03 15:00:59,409][starVLA.training.train_starvla][INFO] - Step 2024, Loss: {'train/loss': 0.0011171880178153515, 'train/grad_norm_pre_clip': 1.1424282789230347, 'timing/data': 0.0003128200041828677, 'timing/model': 0.3699899019993609, 'learning_rate/qwen_vl_interface': 6.8216997441879805e-06, 'learning_rate/action_model': 6.821699744187981e-05, 'epoch': 0.15}) +[2026-10-03 15:01:00,068][starVLA.training.train_starvla][INFO] - Step 2025, Loss: {'train/loss': 8.240262104663998e-06, 'train/grad_norm_pre_clip': 0.0019669735338538885, 'timing/data': 0.00027083999884780496, 'timing/model': 0.37104684200312477, 'learning_rate/qwen_vl_interface': 6.818825544287045e-06, 'learning_rate/action_model': 6.818825544287045e-05, 'epoch': 0.15}) +[2026-10-03 15:01:00,727][starVLA.training.train_starvla][INFO] - Step 2026, Loss: {'train/loss': 2.1679545170627534e-05, 'train/grad_norm_pre_clip': 0.003400239860638976, 'timing/data': 0.00035658999695442617, 'timing/model': 0.36952539200137835, 'learning_rate/qwen_vl_interface': 6.815950699501186e-06, 'learning_rate/action_model': 6.815950699501186e-05, 'epoch': 0.15}) +[2026-10-03 15:01:01,587][starVLA.training.train_starvla][INFO] - Step 2027, Loss: {'train/loss': 0.2920323312282562, 'train/grad_norm_pre_clip': 54.24049758911133, 'timing/data': 0.000222079994273372, 'timing/model': 0.37029791199893225, 'learning_rate/qwen_vl_interface': 6.813075211012148e-06, 'learning_rate/action_model': 6.813075211012147e-05, 'epoch': 0.15}) +[2026-10-03 15:01:02,245][starVLA.training.train_starvla][INFO] - Step 2028, Loss: {'train/loss': 6.706932617817074e-05, 'train/grad_norm_pre_clip': 0.014823016710579395, 'timing/data': 0.00032129000464919955, 'timing/model': 0.37129923100292217, 'learning_rate/qwen_vl_interface': 6.81019908000193e-06, 'learning_rate/action_model': 6.81019908000193e-05, 'epoch': 0.15}) +[2026-10-03 15:01:02,904][starVLA.training.train_starvla][INFO] - Step 2029, Loss: {'train/loss': 0.38243091106414795, 'train/grad_norm_pre_clip': 62.65940475463867, 'timing/data': 0.0003175799938617274, 'timing/model': 0.3707001919974573, 'learning_rate/qwen_vl_interface': 6.807322307652806e-06, 'learning_rate/action_model': 6.807322307652806e-05, 'epoch': 0.15}) +[2026-10-03 15:01:03,562][starVLA.training.train_starvla][INFO] - Step 2030, Loss: {'train/loss': 0.0010584238916635513, 'train/grad_norm_pre_clip': 0.56940758228302, 'timing/data': 0.0003492600080789998, 'timing/model': 0.3693015830067452, 'learning_rate/qwen_vl_interface': 6.804444895147306e-06, 'learning_rate/action_model': 6.804444895147306e-05, 'epoch': 0.15}) +[2026-10-03 15:01:04,221][starVLA.training.train_starvla][INFO] - Step 2031, Loss: {'train/loss': 0.0008021736866794527, 'train/grad_norm_pre_clip': 1.0146788358688354, 'timing/data': 0.0004177800001343712, 'timing/model': 0.3692559820046881, 'learning_rate/qwen_vl_interface': 6.8015668436682255e-06, 'learning_rate/action_model': 6.801566843668226e-05, 'epoch': 0.15}) +[2026-10-03 15:01:05,084][starVLA.training.train_starvla][INFO] - Step 2032, Loss: {'train/loss': 3.6019584513269365e-05, 'train/grad_norm_pre_clip': 0.007155553437769413, 'timing/data': 0.00022996000188868493, 'timing/model': 0.37161988200386986, 'learning_rate/qwen_vl_interface': 6.7986881543986245e-06, 'learning_rate/action_model': 6.798688154398624e-05, 'epoch': 0.15}) +[2026-10-03 15:01:05,744][starVLA.training.train_starvla][INFO] - Step 2033, Loss: {'train/loss': 0.00015663437079638243, 'train/grad_norm_pre_clip': 0.10223124921321869, 'timing/data': 0.0003323199925944209, 'timing/model': 0.3709647310024593, 'learning_rate/qwen_vl_interface': 6.795808828521822e-06, 'learning_rate/action_model': 6.795808828521821e-05, 'epoch': 0.15}) +[2026-10-03 15:01:06,401][starVLA.training.train_starvla][INFO] - Step 2034, Loss: {'train/loss': 0.29096439480781555, 'train/grad_norm_pre_clip': 24.0909481048584, 'timing/data': 0.00022506999084725976, 'timing/model': 0.36888956199982204, 'learning_rate/qwen_vl_interface': 6.792928867221401e-06, 'learning_rate/action_model': 6.792928867221401e-05, 'epoch': 0.15}) +[2026-10-03 15:01:07,058][starVLA.training.train_starvla][INFO] - Step 2035, Loss: {'train/loss': 0.0003283941186964512, 'train/grad_norm_pre_clip': 0.19080016016960144, 'timing/data': 0.00021367000590544194, 'timing/model': 0.36931615199137013, 'learning_rate/qwen_vl_interface': 6.790048271681202e-06, 'learning_rate/action_model': 6.790048271681202e-05, 'epoch': 0.15}) +[2026-10-03 15:01:07,716][starVLA.training.train_starvla][INFO] - Step 2036, Loss: {'train/loss': 2.8928428946528584e-05, 'train/grad_norm_pre_clip': 0.008566716685891151, 'timing/data': 0.0002803800016408786, 'timing/model': 0.37052169199159835, 'learning_rate/qwen_vl_interface': 6.7871670430853335e-06, 'learning_rate/action_model': 6.787167043085333e-05, 'epoch': 0.15}) +[2026-10-03 15:01:08,576][starVLA.training.train_starvla][INFO] - Step 2037, Loss: {'train/loss': 0.01441324781626463, 'train/grad_norm_pre_clip': 27.160343170166016, 'timing/data': 0.0002613199903862551, 'timing/model': 0.5725246500078356, 'learning_rate/qwen_vl_interface': 6.784285182618156e-06, 'learning_rate/action_model': 6.784285182618156e-05, 'epoch': 0.15}) +[2026-10-03 15:01:09,234][starVLA.training.train_starvla][INFO] - Step 2038, Loss: {'train/loss': 0.1161041259765625, 'train/grad_norm_pre_clip': 52.15652084350586, 'timing/data': 0.00035895999462809414, 'timing/model': 0.36945281199587043, 'learning_rate/qwen_vl_interface': 6.781402691464297e-06, 'learning_rate/action_model': 6.781402691464297e-05, 'epoch': 0.15}) +[2026-10-03 15:01:09,892][starVLA.training.train_starvla][INFO] - Step 2039, Loss: {'train/loss': 0.1904439628124237, 'train/grad_norm_pre_clip': 52.737770080566406, 'timing/data': 0.00022508999973069876, 'timing/model': 0.36926817199855577, 'learning_rate/qwen_vl_interface': 6.778519570808636e-06, 'learning_rate/action_model': 6.778519570808636e-05, 'epoch': 0.15}) +[2026-10-03 15:01:10,527][starVLA.training.train_starvla][INFO] - Step 2040, Loss: {'train/loss': 0.784485399723053, 'train/grad_norm_pre_clip': 26.03663444519043, 'timing/data': 0.0003325790021335706, 'timing/model': 0.37057793299027253, 'learning_rate/qwen_vl_interface': 6.775635821836317e-06, 'learning_rate/action_model': 6.775635821836317e-05, 'epoch': 0.15}) +[2026-10-03 15:01:11,185][starVLA.training.train_starvla][INFO] - Step 2041, Loss: {'train/loss': 5.913237691856921e-05, 'train/grad_norm_pre_clip': 0.019315313547849655, 'timing/data': 0.00026953899941872805, 'timing/model': 0.37030801198852714, 'learning_rate/qwen_vl_interface': 6.77275144573274e-06, 'learning_rate/action_model': 6.77275144573274e-05, 'epoch': 0.15}) +[2026-10-03 15:01:12,048][starVLA.training.train_starvla][INFO] - Step 2042, Loss: {'train/loss': 0.002064821310341358, 'train/grad_norm_pre_clip': 1.2799742221832275, 'timing/data': 0.00033801999234128743, 'timing/model': 0.5723295400093775, 'learning_rate/qwen_vl_interface': 6.769866443683564e-06, 'learning_rate/action_model': 6.769866443683563e-05, 'epoch': 0.15}) +[2026-10-03 15:01:12,706][starVLA.training.train_starvla][INFO] - Step 2043, Loss: {'train/loss': 0.43213188648223877, 'train/grad_norm_pre_clip': 63.81997299194336, 'timing/data': 0.00020858900097664446, 'timing/model': 0.3696094130136771, 'learning_rate/qwen_vl_interface': 6.766980816874701e-06, 'learning_rate/action_model': 6.7669808168747e-05, 'epoch': 0.15}) +[2026-10-03 15:01:13,365][starVLA.training.train_starvla][INFO] - Step 2044, Loss: {'train/loss': 0.21975009143352509, 'train/grad_norm_pre_clip': 40.00326156616211, 'timing/data': 0.00030564000189770013, 'timing/model': 0.3705325519986218, 'learning_rate/qwen_vl_interface': 6.764094566492327e-06, 'learning_rate/action_model': 6.764094566492328e-05, 'epoch': 0.15}) +[2026-10-03 15:01:14,024][starVLA.training.train_starvla][INFO] - Step 2045, Loss: {'train/loss': 0.3731074929237366, 'train/grad_norm_pre_clip': 184.09478759765625, 'timing/data': 0.00022852999973110855, 'timing/model': 0.3713961119938176, 'learning_rate/qwen_vl_interface': 6.7612076937228685e-06, 'learning_rate/action_model': 6.761207693722869e-05, 'epoch': 0.15}) +[2026-10-03 15:01:14,683][starVLA.training.train_starvla][INFO] - Step 2046, Loss: {'train/loss': 0.01263117603957653, 'train/grad_norm_pre_clip': 12.076108932495117, 'timing/data': 0.00022699999681208283, 'timing/model': 0.3696603230055189, 'learning_rate/qwen_vl_interface': 6.7583201997530126e-06, 'learning_rate/action_model': 6.758320199753012e-05, 'epoch': 0.15}) +[2026-10-03 15:01:15,341][starVLA.training.train_starvla][INFO] - Step 2047, Loss: {'train/loss': 0.001074143685400486, 'train/grad_norm_pre_clip': 0.5740177035331726, 'timing/data': 0.00033275999885518104, 'timing/model': 0.36969022199627943, 'learning_rate/qwen_vl_interface': 6.755432085769693e-06, 'learning_rate/action_model': 6.755432085769694e-05, 'epoch': 0.15}) +[2026-10-03 15:01:16,203][starVLA.training.train_starvla][INFO] - Step 2048, Loss: {'train/loss': 0.566543698310852, 'train/grad_norm_pre_clip': 28.815988540649414, 'timing/data': 0.0003498900041449815, 'timing/model': 0.3713707720016828, 'learning_rate/qwen_vl_interface': 6.75254335296011e-06, 'learning_rate/action_model': 6.75254335296011e-05, 'epoch': 0.15}) +[2026-10-03 15:01:16,863][starVLA.training.train_starvla][INFO] - Step 2049, Loss: {'train/loss': 0.004042617045342922, 'train/grad_norm_pre_clip': 5.868986129760742, 'timing/data': 0.00030727998819202185, 'timing/model': 0.3713893820095109, 'learning_rate/qwen_vl_interface': 6.749654002511712e-06, 'learning_rate/action_model': 6.749654002511712e-05, 'epoch': 0.15}) +[2026-10-03 15:01:17,522][starVLA.training.train_starvla][INFO] - Step 2050, Loss: {'train/loss': 0.00857728160917759, 'train/grad_norm_pre_clip': 5.16742467880249, 'timing/data': 0.0003023000026587397, 'timing/model': 0.37014424199878704, 'learning_rate/qwen_vl_interface': 6.7467640356122e-06, 'learning_rate/action_model': 6.746764035612199e-05, 'epoch': 0.15}) +[2026-10-03 15:01:18,181][starVLA.training.train_starvla][INFO] - Step 2051, Loss: {'train/loss': 0.00013953150482848287, 'train/grad_norm_pre_clip': 0.030727479606866837, 'timing/data': 0.00033552000240888447, 'timing/model': 0.3694739529892104, 'learning_rate/qwen_vl_interface': 6.74387345344953e-06, 'learning_rate/action_model': 6.74387345344953e-05, 'epoch': 0.15}) +[2026-10-03 15:01:18,841][starVLA.training.train_starvla][INFO] - Step 2052, Loss: {'train/loss': 0.003324022050946951, 'train/grad_norm_pre_clip': 0.9197923541069031, 'timing/data': 0.00023299999884329736, 'timing/model': 0.3709295430016937, 'learning_rate/qwen_vl_interface': 6.740982257211911e-06, 'learning_rate/action_model': 6.740982257211911e-05, 'epoch': 0.15}) +[2026-10-03 15:01:19,697][starVLA.training.train_starvla][INFO] - Step 2053, Loss: {'train/loss': 0.00260633765719831, 'train/grad_norm_pre_clip': 1.1936590671539307, 'timing/data': 0.00022204000561032444, 'timing/model': 0.3710998119931901, 'learning_rate/qwen_vl_interface': 6.738090448087806e-06, 'learning_rate/action_model': 6.738090448087806e-05, 'epoch': 0.15}) +[2026-10-03 15:01:20,356][starVLA.training.train_starvla][INFO] - Step 2054, Loss: {'train/loss': 0.00032122747506946325, 'train/grad_norm_pre_clip': 0.15471237897872925, 'timing/data': 0.00022783999156672508, 'timing/model': 0.36994037199474405, 'learning_rate/qwen_vl_interface': 6.735198027265928e-06, 'learning_rate/action_model': 6.735198027265928e-05, 'epoch': 0.15}) +[2026-10-03 15:01:21,015][starVLA.training.train_starvla][INFO] - Step 2055, Loss: {'train/loss': 0.22838066518306732, 'train/grad_norm_pre_clip': 83.25752258300781, 'timing/data': 0.0002506299933884293, 'timing/model': 0.3699880420026602, 'learning_rate/qwen_vl_interface': 6.7323049959352424e-06, 'learning_rate/action_model': 6.732304995935243e-05, 'epoch': 0.15}) +[2026-10-03 15:01:21,651][starVLA.training.train_starvla][INFO] - Step 2056, Loss: {'train/loss': 0.0003944848431274295, 'train/grad_norm_pre_clip': 0.11574702709913254, 'timing/data': 0.0004064999957336113, 'timing/model': 0.3700424819980981, 'learning_rate/qwen_vl_interface': 6.729411355284962e-06, 'learning_rate/action_model': 6.729411355284963e-05, 'epoch': 0.15}) +[2026-10-03 15:01:22,309][starVLA.training.train_starvla][INFO] - Step 2057, Loss: {'train/loss': 0.016020873561501503, 'train/grad_norm_pre_clip': 12.673038482666016, 'timing/data': 0.0002254699938930571, 'timing/model': 0.37071927200304344, 'learning_rate/qwen_vl_interface': 6.726517106504559e-06, 'learning_rate/action_model': 6.726517106504558e-05, 'epoch': 0.15}) +[2026-10-03 15:01:23,165][starVLA.training.train_starvla][INFO] - Step 2058, Loss: {'train/loss': 0.4618472754955292, 'train/grad_norm_pre_clip': 32.670936584472656, 'timing/data': 0.0003697299980558455, 'timing/model': 0.5669895920000272, 'learning_rate/qwen_vl_interface': 6.723622250783745e-06, 'learning_rate/action_model': 6.723622250783744e-05, 'epoch': 0.15}) +[2026-10-03 15:01:23,824][starVLA.training.train_starvla][INFO] - Step 2059, Loss: {'train/loss': 0.2636491358280182, 'train/grad_norm_pre_clip': 136.41236877441406, 'timing/data': 0.0002612000098451972, 'timing/model': 0.3698946119984612, 'learning_rate/qwen_vl_interface': 6.720726789312487e-06, 'learning_rate/action_model': 6.720726789312487e-05, 'epoch': 0.15}) +[2026-10-03 15:01:24,459][starVLA.training.train_starvla][INFO] - Step 2060, Loss: {'train/loss': 7.025263039395213e-05, 'train/grad_norm_pre_clip': 0.01814156211912632, 'timing/data': 0.0002590500080259517, 'timing/model': 0.37021691200789064, 'learning_rate/qwen_vl_interface': 6.7178307232810015e-06, 'learning_rate/action_model': 6.717830723281e-05, 'epoch': 0.15}) +[2026-10-03 15:01:25,119][starVLA.training.train_starvla][INFO] - Step 2061, Loss: {'train/loss': 0.00013391688116826117, 'train/grad_norm_pre_clip': 0.03696278855204582, 'timing/data': 0.0002640999882714823, 'timing/model': 0.3705504030076554, 'learning_rate/qwen_vl_interface': 6.714934053879749e-06, 'learning_rate/action_model': 6.71493405387975e-05, 'epoch': 0.15}) +[2026-10-03 15:01:25,777][starVLA.training.train_starvla][INFO] - Step 2062, Loss: {'train/loss': 0.46490535140037537, 'train/grad_norm_pre_clip': 14.051193237304688, 'timing/data': 0.0003633800079114735, 'timing/model': 0.36960064199229237, 'learning_rate/qwen_vl_interface': 6.712036782299442e-06, 'learning_rate/action_model': 6.712036782299443e-05, 'epoch': 0.15}) +[2026-10-03 15:01:26,435][starVLA.training.train_starvla][INFO] - Step 2063, Loss: {'train/loss': 0.0001768628862919286, 'train/grad_norm_pre_clip': 0.058996908366680145, 'timing/data': 0.0002138900017598644, 'timing/model': 0.36927663200185634, 'learning_rate/qwen_vl_interface': 6.709138909731041e-06, 'learning_rate/action_model': 6.709138909731041e-05, 'epoch': 0.15}) +[2026-10-03 15:01:27,291][starVLA.training.train_starvla][INFO] - Step 2064, Loss: {'train/loss': 0.0011888379231095314, 'train/grad_norm_pre_clip': 0.32994553446769714, 'timing/data': 0.0003617100010160357, 'timing/model': 0.37097235200053547, 'learning_rate/qwen_vl_interface': 6.706240437365751e-06, 'learning_rate/action_model': 6.70624043736575e-05, 'epoch': 0.15}) +[2026-10-03 15:01:27,949][starVLA.training.train_starvla][INFO] - Step 2065, Loss: {'train/loss': 0.0001316135749220848, 'train/grad_norm_pre_clip': 0.025689205154776573, 'timing/data': 0.0003216199984308332, 'timing/model': 0.3702432630088879, 'learning_rate/qwen_vl_interface': 6.703341366395023e-06, 'learning_rate/action_model': 6.703341366395023e-05, 'epoch': 0.15}) +[2026-10-03 15:01:28,608][starVLA.training.train_starvla][INFO] - Step 2066, Loss: {'train/loss': 6.420422869268805e-05, 'train/grad_norm_pre_clip': 0.01008305698633194, 'timing/data': 0.00022360000002663583, 'timing/model': 0.3697832120087696, 'learning_rate/qwen_vl_interface': 6.700441698010558e-06, 'learning_rate/action_model': 6.700441698010558e-05, 'epoch': 0.15}) +[2026-10-03 15:01:29,267][starVLA.training.train_starvla][INFO] - Step 2067, Loss: {'train/loss': 0.00013766842312179506, 'train/grad_norm_pre_clip': 0.048118606209754944, 'timing/data': 0.00021096000273246318, 'timing/model': 0.3696749020018615, 'learning_rate/qwen_vl_interface': 6.697541433404299e-06, 'learning_rate/action_model': 6.697541433404299e-05, 'epoch': 0.15}) +[2026-10-03 15:01:30,134][starVLA.training.train_starvla][INFO] - Step 2068, Loss: {'train/loss': 0.022798430174589157, 'train/grad_norm_pre_clip': 143.45880126953125, 'timing/data': 0.00026522998814471066, 'timing/model': 0.5792504889977863, 'learning_rate/qwen_vl_interface': 6.6946405737684325e-06, 'learning_rate/action_model': 6.694640573768433e-05, 'epoch': 0.15}) +[2026-10-03 15:01:30,770][starVLA.training.train_starvla][INFO] - Step 2069, Loss: {'train/loss': 0.003487437032163143, 'train/grad_norm_pre_clip': 6.328918933868408, 'timing/data': 0.0002631799870869145, 'timing/model': 0.34780138499627355, 'learning_rate/qwen_vl_interface': 6.691739120295397e-06, 'learning_rate/action_model': 6.691739120295397e-05, 'epoch': 0.15}) +[2026-10-03 15:01:31,428][starVLA.training.train_starvla][INFO] - Step 2070, Loss: {'train/loss': 0.0063802264630794525, 'train/grad_norm_pre_clip': 3.939836025238037, 'timing/data': 0.00022857901058159769, 'timing/model': 0.3695400130091002, 'learning_rate/qwen_vl_interface': 6.688837074177868e-06, 'learning_rate/action_model': 6.688837074177868e-05, 'epoch': 0.15}) +[2026-10-03 15:01:32,086][starVLA.training.train_starvla][INFO] - Step 2071, Loss: {'train/loss': 0.3225012421607971, 'train/grad_norm_pre_clip': 44.59381103515625, 'timing/data': 0.00038887000118847936, 'timing/model': 0.3686862829927122, 'learning_rate/qwen_vl_interface': 6.68593443660877e-06, 'learning_rate/action_model': 6.685934436608769e-05, 'epoch': 0.15}) +[2026-10-03 15:01:32,743][starVLA.training.train_starvla][INFO] - Step 2072, Loss: {'train/loss': 0.00012177391909062862, 'train/grad_norm_pre_clip': 0.03345263749361038, 'timing/data': 0.0002593399985926226, 'timing/model': 0.37044047299423255, 'learning_rate/qwen_vl_interface': 6.6830312087812625e-06, 'learning_rate/action_model': 6.683031208781262e-05, 'epoch': 0.15}) +[2026-10-03 15:01:33,400][starVLA.training.train_starvla][INFO] - Step 2073, Loss: {'train/loss': 0.21040308475494385, 'train/grad_norm_pre_clip': 182.28358459472656, 'timing/data': 0.0003130799887003377, 'timing/model': 0.37013648300489876, 'learning_rate/qwen_vl_interface': 6.680127391888754e-06, 'learning_rate/action_model': 6.680127391888754e-05, 'epoch': 0.15}) +[2026-10-03 15:01:34,257][starVLA.training.train_starvla][INFO] - Step 2074, Loss: {'train/loss': 0.0003001520235557109, 'train/grad_norm_pre_clip': 0.05472349002957344, 'timing/data': 0.0002189699880545959, 'timing/model': 0.3694695330050308, 'learning_rate/qwen_vl_interface': 6.677222987124897e-06, 'learning_rate/action_model': 6.677222987124897e-05, 'epoch': 0.15}) +[2026-10-03 15:01:34,892][starVLA.training.train_starvla][INFO] - Step 2075, Loss: {'train/loss': 0.00015316282224375755, 'train/grad_norm_pre_clip': 0.04216335713863373, 'timing/data': 0.0002718899922911078, 'timing/model': 0.3687683129974175, 'learning_rate/qwen_vl_interface': 6.67431799568358e-06, 'learning_rate/action_model': 6.67431799568358e-05, 'epoch': 0.15}) +[2026-10-03 15:01:35,551][starVLA.training.train_starvla][INFO] - Step 2076, Loss: {'train/loss': 0.00019601703388616443, 'train/grad_norm_pre_clip': 0.041139110922813416, 'timing/data': 0.00036778999492526054, 'timing/model': 0.37067880199174397, 'learning_rate/qwen_vl_interface': 6.671412418758939e-06, 'learning_rate/action_model': 6.671412418758938e-05, 'epoch': 0.15}) +[2026-10-03 15:01:36,211][starVLA.training.train_starvla][INFO] - Step 2077, Loss: {'train/loss': 0.00018884870223701, 'train/grad_norm_pre_clip': 0.04125532880425453, 'timing/data': 0.00032148000900633633, 'timing/model': 0.37122041199472733, 'learning_rate/qwen_vl_interface': 6.668506257545341e-06, 'learning_rate/action_model': 6.668506257545341e-05, 'epoch': 0.15}) +[2026-10-03 15:01:36,870][starVLA.training.train_starvla][INFO] - Step 2078, Loss: {'train/loss': 0.00027884170413017273, 'train/grad_norm_pre_clip': 0.08097559958696365, 'timing/data': 0.0002967000036733225, 'timing/model': 0.3695403229939984, 'learning_rate/qwen_vl_interface': 6.665599513237403e-06, 'learning_rate/action_model': 6.665599513237402e-05, 'epoch': 0.15}) +[2026-10-03 15:01:37,725][starVLA.training.train_starvla][INFO] - Step 2079, Loss: {'train/loss': 0.0026746164076030254, 'train/grad_norm_pre_clip': 5.926705837249756, 'timing/data': 0.00045873000635765493, 'timing/model': 0.5669952320022276, 'learning_rate/qwen_vl_interface': 6.662692187029978e-06, 'learning_rate/action_model': 6.662692187029978e-05, 'epoch': 0.15}) +[2026-10-03 15:01:38,383][starVLA.training.train_starvla][INFO] - Step 2080, Loss: {'train/loss': 0.00011414830078138039, 'train/grad_norm_pre_clip': 0.029064558446407318, 'timing/data': 0.0003076899884035811, 'timing/model': 0.37022312299814075, 'learning_rate/qwen_vl_interface': 6.659784280118156e-06, 'learning_rate/action_model': 6.659784280118156e-05, 'epoch': 0.15}) +[2026-10-03 15:01:39,042][starVLA.training.train_starvla][INFO] - Step 2081, Loss: {'train/loss': 0.7511450052261353, 'train/grad_norm_pre_clip': 154.0740966796875, 'timing/data': 0.0003599499905249104, 'timing/model': 0.3704873620008584, 'learning_rate/qwen_vl_interface': 6.65687579369727e-06, 'learning_rate/action_model': 6.656875793697271e-05, 'epoch': 0.15}) +[2026-10-03 15:01:39,700][starVLA.training.train_starvla][INFO] - Step 2082, Loss: {'train/loss': 0.0006477124989032745, 'train/grad_norm_pre_clip': 0.2878951132297516, 'timing/data': 0.00030808999144937843, 'timing/model': 0.36950208198686596, 'learning_rate/qwen_vl_interface': 6.653966728962885e-06, 'learning_rate/action_model': 6.653966728962885e-05, 'epoch': 0.15}) +[2026-10-03 15:01:40,358][starVLA.training.train_starvla][INFO] - Step 2083, Loss: {'train/loss': 0.0009701132075861096, 'train/grad_norm_pre_clip': 0.4584348797798157, 'timing/data': 0.00026556999364402145, 'timing/model': 0.3691636220028158, 'learning_rate/qwen_vl_interface': 6.651057087110813e-06, 'learning_rate/action_model': 6.651057087110813e-05, 'epoch': 0.15}) +[2026-10-03 15:01:41,016][starVLA.training.train_starvla][INFO] - Step 2084, Loss: {'train/loss': 0.00024632044369354844, 'train/grad_norm_pre_clip': 0.07596130669116974, 'timing/data': 0.00034037001023534685, 'timing/model': 0.3701513720006915, 'learning_rate/qwen_vl_interface': 6.6481468693370924e-06, 'learning_rate/action_model': 6.648146869337091e-05, 'epoch': 0.15}) +[2026-10-03 15:01:41,874][starVLA.training.train_starvla][INFO] - Step 2085, Loss: {'train/loss': 0.0007083032396622002, 'train/grad_norm_pre_clip': 0.4657362401485443, 'timing/data': 0.00026811000134330243, 'timing/model': 0.3707867219927721, 'learning_rate/qwen_vl_interface': 6.645236076838009e-06, 'learning_rate/action_model': 6.645236076838009e-05, 'epoch': 0.15}) +[2026-10-03 15:01:42,510][starVLA.training.train_starvla][INFO] - Step 2086, Loss: {'train/loss': 0.09495081752538681, 'train/grad_norm_pre_clip': 118.05789947509766, 'timing/data': 0.00034933999995701015, 'timing/model': 0.36992536300385837, 'learning_rate/qwen_vl_interface': 6.642324710810075e-06, 'learning_rate/action_model': 6.642324710810076e-05, 'epoch': 0.15}) +[2026-10-03 15:01:43,168][starVLA.training.train_starvla][INFO] - Step 2087, Loss: {'train/loss': 3.9247432141564786e-05, 'train/grad_norm_pre_clip': 0.007540257181972265, 'timing/data': 0.0003339600079925731, 'timing/model': 0.36990813301235903, 'learning_rate/qwen_vl_interface': 6.639412772450047e-06, 'learning_rate/action_model': 6.639412772450047e-05, 'epoch': 0.15}) +[2026-10-03 15:01:43,827][starVLA.training.train_starvla][INFO] - Step 2088, Loss: {'train/loss': 1.7277634469792247e-05, 'train/grad_norm_pre_clip': 0.003560052253305912, 'timing/data': 0.00022363000607583672, 'timing/model': 0.3707830329949502, 'learning_rate/qwen_vl_interface': 6.6365002629549104e-06, 'learning_rate/action_model': 6.63650026295491e-05, 'epoch': 0.15}) +[2026-10-03 15:01:44,486][starVLA.training.train_starvla][INFO] - Step 2089, Loss: {'train/loss': 0.00022519382764585316, 'train/grad_norm_pre_clip': 0.11319069564342499, 'timing/data': 0.0002946599997812882, 'timing/model': 0.37085204300819896, 'learning_rate/qwen_vl_interface': 6.633587183521887e-06, 'learning_rate/action_model': 6.633587183521887e-05, 'epoch': 0.15}) +[2026-10-03 15:01:45,336][starVLA.training.train_starvla][INFO] - Step 2090, Loss: {'train/loss': 3.2468225981574506e-05, 'train/grad_norm_pre_clip': 0.0045399172231554985, 'timing/data': 0.00022098000044934452, 'timing/model': 0.5615750329889124, 'learning_rate/qwen_vl_interface': 6.6306735353484365e-06, 'learning_rate/action_model': 6.630673535348436e-05, 'epoch': 0.15}) +[2026-10-03 15:01:45,993][starVLA.training.train_starvla][INFO] - Step 2091, Loss: {'train/loss': 0.00013182676048018038, 'train/grad_norm_pre_clip': 0.14038647711277008, 'timing/data': 0.0003544400096870959, 'timing/model': 0.36896884300222155, 'learning_rate/qwen_vl_interface': 6.627759319632246e-06, 'learning_rate/action_model': 6.627759319632247e-05, 'epoch': 0.15}) +[2026-10-03 15:01:46,651][starVLA.training.train_starvla][INFO] - Step 2092, Loss: {'train/loss': 0.030389852821826935, 'train/grad_norm_pre_clip': 46.43883514404297, 'timing/data': 0.00025594999897293746, 'timing/model': 0.3704403320007259, 'learning_rate/qwen_vl_interface': 6.624844537571245e-06, 'learning_rate/action_model': 6.624844537571245e-05, 'epoch': 0.15}) +[2026-10-03 15:01:47,309][starVLA.training.train_starvla][INFO] - Step 2093, Loss: {'train/loss': 4.648599497159012e-05, 'train/grad_norm_pre_clip': 0.01858566142618656, 'timing/data': 0.00026174000231549144, 'timing/model': 0.37108344299485907, 'learning_rate/qwen_vl_interface': 6.621929190363585e-06, 'learning_rate/action_model': 6.621929190363585e-05, 'epoch': 0.15}) +[2026-10-03 15:01:47,967][starVLA.training.train_starvla][INFO] - Step 2094, Loss: {'train/loss': 3.241565354983322e-05, 'train/grad_norm_pre_clip': 0.009202747605741024, 'timing/data': 0.00030875000811647624, 'timing/model': 0.36936049300129525, 'learning_rate/qwen_vl_interface': 6.619013279207657e-06, 'learning_rate/action_model': 6.619013279207656e-05, 'epoch': 0.15}) +[2026-10-03 15:01:48,823][starVLA.training.train_starvla][INFO] - Step 2095, Loss: {'train/loss': 0.5103344917297363, 'train/grad_norm_pre_clip': 24.761924743652344, 'timing/data': 0.00038158000097610056, 'timing/model': 0.5673208619991783, 'learning_rate/qwen_vl_interface': 6.616096805302083e-06, 'learning_rate/action_model': 6.616096805302082e-05, 'epoch': 0.15}) +[2026-10-03 15:01:49,458][starVLA.training.train_starvla][INFO] - Step 2096, Loss: {'train/loss': 8.142315346049145e-05, 'train/grad_norm_pre_clip': 0.05906269699335098, 'timing/data': 0.00030108999635558575, 'timing/model': 0.3701951629918767, 'learning_rate/qwen_vl_interface': 6.6131797698457135e-06, 'learning_rate/action_model': 6.613179769845713e-05, 'epoch': 0.15}) +[2026-10-03 15:01:50,092][starVLA.training.train_starvla][INFO] - Step 2097, Loss: {'train/loss': 0.022086776793003082, 'train/grad_norm_pre_clip': 95.01361846923828, 'timing/data': 0.00030813000921625644, 'timing/model': 0.36909096300951205, 'learning_rate/qwen_vl_interface': 6.610262174037633e-06, 'learning_rate/action_model': 6.610262174037633e-05, 'epoch': 0.15}) +[2026-10-03 15:01:50,750][starVLA.training.train_starvla][INFO] - Step 2098, Loss: {'train/loss': 2.571859658928588e-05, 'train/grad_norm_pre_clip': 0.004077887162566185, 'timing/data': 0.0003552999987732619, 'timing/model': 0.3695875840057852, 'learning_rate/qwen_vl_interface': 6.607344019077155e-06, 'learning_rate/action_model': 6.607344019077156e-05, 'epoch': 0.15}) +[2026-10-03 15:01:51,389][starVLA.training.train_starvla][INFO] - Step 2099, Loss: {'train/loss': 4.905891910311766e-05, 'train/grad_norm_pre_clip': 0.011191649362444878, 'timing/data': 0.0002115700044669211, 'timing/model': 0.3711009219987318, 'learning_rate/qwen_vl_interface': 6.604425306163822e-06, 'learning_rate/action_model': 6.604425306163822e-05, 'epoch': 0.15}) +[2026-10-03 15:01:52,051][starVLA.training.train_starvla][INFO] - Step 2100, Loss: {'train/loss': 0.004227018915116787, 'train/grad_norm_pre_clip': 2.476508617401123, 'timing/data': 0.00031656998908147216, 'timing/model': 0.37208028200257104, 'learning_rate/qwen_vl_interface': 6.6015060364974045e-06, 'learning_rate/action_model': 6.601506036497404e-05, 'epoch': 0.15}) +[2026-10-03 15:01:52,919][starVLA.training.train_starvla][INFO] - Step 2101, Loss: {'train/loss': 0.00010922152432613075, 'train/grad_norm_pre_clip': 0.029591159895062447, 'timing/data': 0.00037213000177871436, 'timing/model': 0.37282759200024884, 'learning_rate/qwen_vl_interface': 6.598586211277907e-06, 'learning_rate/action_model': 6.598586211277907e-05, 'epoch': 0.15}) +[2026-10-03 15:01:53,581][starVLA.training.train_starvla][INFO] - Step 2102, Loss: {'train/loss': 0.5883994698524475, 'train/grad_norm_pre_clip': 29.97591209411621, 'timing/data': 0.0002233400009572506, 'timing/model': 0.3713871630025096, 'learning_rate/qwen_vl_interface': 6.59566583170556e-06, 'learning_rate/action_model': 6.59566583170556e-05, 'epoch': 0.15}) +[2026-10-03 15:01:54,243][starVLA.training.train_starvla][INFO] - Step 2103, Loss: {'train/loss': 0.00014502753037959337, 'train/grad_norm_pre_clip': 0.04291807487607002, 'timing/data': 0.0002164899924537167, 'timing/model': 0.37192701299500186, 'learning_rate/qwen_vl_interface': 6.59274489898082e-06, 'learning_rate/action_model': 6.592744898980819e-05, 'epoch': 0.15}) +[2026-10-03 15:01:54,905][starVLA.training.train_starvla][INFO] - Step 2104, Loss: {'train/loss': 9.243840759154409e-05, 'train/grad_norm_pre_clip': 0.01965688355267048, 'timing/data': 0.0003153400029987097, 'timing/model': 0.3725408019963652, 'learning_rate/qwen_vl_interface': 6.5898234143043716e-06, 'learning_rate/action_model': 6.589823414304371e-05, 'epoch': 0.15}) +[2026-10-03 15:01:55,567][starVLA.training.train_starvla][INFO] - Step 2105, Loss: {'train/loss': 5.551458889385685e-05, 'train/grad_norm_pre_clip': 0.016811825335025787, 'timing/data': 0.00025625999842304736, 'timing/model': 0.37291844200808555, 'learning_rate/qwen_vl_interface': 6.5869013788771264e-06, 'learning_rate/action_model': 6.586901378877126e-05, 'epoch': 0.15}) +[2026-10-03 15:01:56,423][starVLA.training.train_starvla][INFO] - Step 2106, Loss: {'train/loss': 0.0001413899881299585, 'train/grad_norm_pre_clip': 0.04685124009847641, 'timing/data': 0.00022447999799624085, 'timing/model': 0.56466777199239, 'learning_rate/qwen_vl_interface': 6.583978793900223e-06, 'learning_rate/action_model': 6.583978793900223e-05, 'epoch': 0.15}) +[2026-10-03 15:01:57,084][starVLA.training.train_starvla][INFO] - Step 2107, Loss: {'train/loss': 5.1526338211260736e-05, 'train/grad_norm_pre_clip': 0.01169031672179699, 'timing/data': 0.0004576899955281988, 'timing/model': 0.37075055300374515, 'learning_rate/qwen_vl_interface': 6.581055660575028e-06, 'learning_rate/action_model': 6.581055660575027e-05, 'epoch': 0.15}) +[2026-10-03 15:01:57,746][starVLA.training.train_starvla][INFO] - Step 2108, Loss: {'train/loss': 0.2353055477142334, 'train/grad_norm_pre_clip': 50.68622970581055, 'timing/data': 0.00027565000345930457, 'timing/model': 0.37251487201137934, 'learning_rate/qwen_vl_interface': 6.578131980103128e-06, 'learning_rate/action_model': 6.578131980103128e-05, 'epoch': 0.15}) +[2026-10-03 15:01:58,409][starVLA.training.train_starvla][INFO] - Step 2109, Loss: {'train/loss': 6.19914208073169e-05, 'train/grad_norm_pre_clip': 0.012201186269521713, 'timing/data': 0.0002833700127666816, 'timing/model': 0.37285806200816296, 'learning_rate/qwen_vl_interface': 6.5752077536863404e-06, 'learning_rate/action_model': 6.57520775368634e-05, 'epoch': 0.15}) +[2026-10-03 15:01:59,071][starVLA.training.train_starvla][INFO] - Step 2110, Loss: {'train/loss': 8.062094275373966e-05, 'train/grad_norm_pre_clip': 0.019271790981292725, 'timing/data': 0.00040271000762004405, 'timing/model': 0.3713687629933702, 'learning_rate/qwen_vl_interface': 6.572282982526702e-06, 'learning_rate/action_model': 6.572282982526702e-05, 'epoch': 0.15}) +[2026-10-03 15:01:59,932][starVLA.training.train_starvla][INFO] - Step 2111, Loss: {'train/loss': 0.320148766040802, 'train/grad_norm_pre_clip': 79.46659851074219, 'timing/data': 0.00021331000607460737, 'timing/model': 0.5708460520108929, 'learning_rate/qwen_vl_interface': 6.5693576678264755e-06, 'learning_rate/action_model': 6.569357667826476e-05, 'epoch': 0.15}) +[2026-10-03 15:02:00,593][starVLA.training.train_starvla][INFO] - Step 2112, Loss: {'train/loss': 0.043730393052101135, 'train/grad_norm_pre_clip': 64.44800567626953, 'timing/data': 0.00025174999609589577, 'timing/model': 0.37202490199706517, 'learning_rate/qwen_vl_interface': 6.566431810788148e-06, 'learning_rate/action_model': 6.566431810788148e-05, 'epoch': 0.15}) +[2026-10-03 15:02:01,255][starVLA.training.train_starvla][INFO] - Step 2113, Loss: {'train/loss': 4.8493944632355124e-05, 'train/grad_norm_pre_clip': 0.008001611568033695, 'timing/data': 0.00026028999127447605, 'timing/model': 0.37264165200758725, 'learning_rate/qwen_vl_interface': 6.5635054126144306e-06, 'learning_rate/action_model': 6.56350541261443e-05, 'epoch': 0.15}) +[2026-10-03 15:02:01,917][starVLA.training.train_starvla][INFO] - Step 2114, Loss: {'train/loss': 0.12562592327594757, 'train/grad_norm_pre_clip': 123.33097839355469, 'timing/data': 0.00037957000313326716, 'timing/model': 0.3712024019914679, 'learning_rate/qwen_vl_interface': 6.560578474508254e-06, 'learning_rate/action_model': 6.560578474508253e-05, 'epoch': 0.15}) +[2026-10-03 15:02:02,555][starVLA.training.train_starvla][INFO] - Step 2115, Loss: {'train/loss': 0.007162594702094793, 'train/grad_norm_pre_clip': 37.50278854370117, 'timing/data': 0.00021182000637054443, 'timing/model': 0.3477463059971342, 'learning_rate/qwen_vl_interface': 6.5576509976727685e-06, 'learning_rate/action_model': 6.557650997672769e-05, 'epoch': 0.15}) +[2026-10-03 15:02:03,216][starVLA.training.train_starvla][INFO] - Step 2116, Loss: {'train/loss': 0.2060403972864151, 'train/grad_norm_pre_clip': 79.81722259521484, 'timing/data': 0.00021143000049050897, 'timing/model': 0.3716811429912923, 'learning_rate/qwen_vl_interface': 6.554722983311354e-06, 'learning_rate/action_model': 6.554722983311355e-05, 'epoch': 0.15}) +[2026-10-03 15:02:04,078][starVLA.training.train_starvla][INFO] - Step 2117, Loss: {'train/loss': 0.0599047876894474, 'train/grad_norm_pre_clip': 37.23849105834961, 'timing/data': 0.00034154999593738467, 'timing/model': 0.372486571999616, 'learning_rate/qwen_vl_interface': 6.551794432627605e-06, 'learning_rate/action_model': 6.551794432627605e-05, 'epoch': 0.15}) +[2026-10-03 15:02:04,740][starVLA.training.train_starvla][INFO] - Step 2118, Loss: {'train/loss': 0.00020593786030076444, 'train/grad_norm_pre_clip': 0.06813264638185501, 'timing/data': 0.0002230099926237017, 'timing/model': 0.37128596199909225, 'learning_rate/qwen_vl_interface': 6.548865346825338e-06, 'learning_rate/action_model': 6.548865346825338e-05, 'epoch': 0.16}) +[2026-10-03 15:02:05,401][starVLA.training.train_starvla][INFO] - Step 2119, Loss: {'train/loss': 0.00014240399468690157, 'train/grad_norm_pre_clip': 0.07515928149223328, 'timing/data': 0.00022016000002622604, 'timing/model': 0.3705864429939538, 'learning_rate/qwen_vl_interface': 6.545935727108589e-06, 'learning_rate/action_model': 6.545935727108588e-05, 'epoch': 0.16}) +[2026-10-03 15:02:06,063][starVLA.training.train_starvla][INFO] - Step 2120, Loss: {'train/loss': 0.5646709203720093, 'train/grad_norm_pre_clip': 29.451108932495117, 'timing/data': 0.0002506299933884293, 'timing/model': 0.3719088830112014, 'learning_rate/qwen_vl_interface': 6.543005574681613e-06, 'learning_rate/action_model': 6.543005574681613e-05, 'epoch': 0.16}) +[2026-10-03 15:02:06,726][starVLA.training.train_starvla][INFO] - Step 2121, Loss: {'train/loss': 0.0003049414081033319, 'train/grad_norm_pre_clip': 0.06512514501810074, 'timing/data': 0.0002469989995006472, 'timing/model': 0.37281921299290843, 'learning_rate/qwen_vl_interface': 6.5400748907488886e-06, 'learning_rate/action_model': 6.540074890748888e-05, 'epoch': 0.16}) +[2026-10-03 15:02:07,584][starVLA.training.train_starvla][INFO] - Step 2122, Loss: {'train/loss': 0.00013067215331830084, 'train/grad_norm_pre_clip': 0.03724837675690651, 'timing/data': 0.00035385000228416175, 'timing/model': 0.37160706300346646, 'learning_rate/qwen_vl_interface': 6.537143676515105e-06, 'learning_rate/action_model': 6.537143676515105e-05, 'epoch': 0.16}) +[2026-10-03 15:02:08,245][starVLA.training.train_starvla][INFO] - Step 2123, Loss: {'train/loss': 0.47652605175971985, 'train/grad_norm_pre_clip': 571.1779174804688, 'timing/data': 0.0002822699898388237, 'timing/model': 0.3704390839993721, 'learning_rate/qwen_vl_interface': 6.534211933185176e-06, 'learning_rate/action_model': 6.534211933185176e-05, 'epoch': 0.16}) +[2026-10-03 15:02:08,906][starVLA.training.train_starvla][INFO] - Step 2124, Loss: {'train/loss': 0.003512427443638444, 'train/grad_norm_pre_clip': 3.913027286529541, 'timing/data': 0.0002773000014713034, 'timing/model': 0.37223305299994536, 'learning_rate/qwen_vl_interface': 6.531279661964231e-06, 'learning_rate/action_model': 6.531279661964231e-05, 'epoch': 0.16}) +[2026-10-03 15:02:09,545][starVLA.training.train_starvla][INFO] - Step 2125, Loss: {'train/loss': 0.0003928307560272515, 'train/grad_norm_pre_clip': 0.0945902094244957, 'timing/data': 0.0002501499984646216, 'timing/model': 0.3718098329991335, 'learning_rate/qwen_vl_interface': 6.528346864057616e-06, 'learning_rate/action_model': 6.528346864057615e-05, 'epoch': 0.16}) +[2026-10-03 15:02:10,207][starVLA.training.train_starvla][INFO] - Step 2126, Loss: {'train/loss': 0.0002747558173723519, 'train/grad_norm_pre_clip': 0.06022273004055023, 'timing/data': 0.00035848999687004834, 'timing/model': 0.37124802298785653, 'learning_rate/qwen_vl_interface': 6.5254135406708906e-06, 'learning_rate/action_model': 6.52541354067089e-05, 'epoch': 0.16}) +[2026-10-03 15:02:11,069][starVLA.training.train_starvla][INFO] - Step 2127, Loss: {'train/loss': 0.0018887603655457497, 'train/grad_norm_pre_clip': 1.3474032878875732, 'timing/data': 0.0003221299994038418, 'timing/model': 0.5725304520019563, 'learning_rate/qwen_vl_interface': 6.522479693009836e-06, 'learning_rate/action_model': 6.522479693009837e-05, 'epoch': 0.16}) +[2026-10-03 15:02:11,731][starVLA.training.train_starvla][INFO] - Step 2128, Loss: {'train/loss': 0.005795512348413467, 'train/grad_norm_pre_clip': 2.6313064098358154, 'timing/data': 0.0002462299889884889, 'timing/model': 0.37246397299168166, 'learning_rate/qwen_vl_interface': 6.519545322280446e-06, 'learning_rate/action_model': 6.519545322280446e-05, 'epoch': 0.16}) +[2026-10-03 15:02:12,346][starVLA.training.train_starvla][INFO] - Step 2129, Loss: {'train/loss': 0.17079049348831177, 'train/grad_norm_pre_clip': 125.22480010986328, 'timing/data': 0.00032172999635804445, 'timing/model': 0.3487076059973333, 'learning_rate/qwen_vl_interface': 6.5166104296889275e-06, 'learning_rate/action_model': 6.516610429688927e-05, 'epoch': 0.16}) +[2026-10-03 15:02:13,008][starVLA.training.train_starvla][INFO] - Step 2130, Loss: {'train/loss': 0.0021780768875032663, 'train/grad_norm_pre_clip': 0.6678982973098755, 'timing/data': 0.00033335000625811517, 'timing/model': 0.371242032997543, 'learning_rate/qwen_vl_interface': 6.513675016441707e-06, 'learning_rate/action_model': 6.513675016441707e-05, 'epoch': 0.16}) +[2026-10-03 15:02:13,669][starVLA.training.train_starvla][INFO] - Step 2131, Loss: {'train/loss': 0.0002481220872141421, 'train/grad_norm_pre_clip': 0.09522239118814468, 'timing/data': 0.00029874999017920345, 'timing/model': 0.3709325529925991, 'learning_rate/qwen_vl_interface': 6.510739083745422e-06, 'learning_rate/action_model': 6.510739083745422e-05, 'epoch': 0.16}) +[2026-10-03 15:02:14,524][starVLA.training.train_starvla][INFO] - Step 2132, Loss: {'train/loss': 0.005934057757258415, 'train/grad_norm_pre_clip': 4.900590896606445, 'timing/data': 0.00026061999960802495, 'timing/model': 0.5666275929979747, 'learning_rate/qwen_vl_interface': 6.507802632806919e-06, 'learning_rate/action_model': 6.507802632806919e-05, 'epoch': 0.16}) +[2026-10-03 15:02:15,174][starVLA.training.train_starvla][INFO] - Step 2133, Loss: {'train/loss': 0.9053012132644653, 'train/grad_norm_pre_clip': 40.35964584350586, 'timing/data': 0.0001985300041269511, 'timing/model': 0.36544347400194965, 'learning_rate/qwen_vl_interface': 6.5048656648332646e-06, 'learning_rate/action_model': 6.504865664833264e-05, 'epoch': 0.16}) +[2026-10-03 15:02:15,822][starVLA.training.train_starvla][INFO] - Step 2134, Loss: {'train/loss': 0.000465972232632339, 'train/grad_norm_pre_clip': 0.34325042366981506, 'timing/data': 0.0003601900098146871, 'timing/model': 0.3648256940068677, 'learning_rate/qwen_vl_interface': 6.501928181031737e-06, 'learning_rate/action_model': 6.501928181031737e-05, 'epoch': 0.16}) +[2026-10-03 15:02:16,473][starVLA.training.train_starvla][INFO] - Step 2135, Loss: {'train/loss': 0.00010586934513412416, 'train/grad_norm_pre_clip': 0.05055695027112961, 'timing/data': 0.00020311999833211303, 'timing/model': 0.3655372439970961, 'learning_rate/qwen_vl_interface': 6.498990182609823e-06, 'learning_rate/action_model': 6.498990182609823e-05, 'epoch': 0.16}) +[2026-10-03 15:02:17,121][starVLA.training.train_starvla][INFO] - Step 2136, Loss: {'train/loss': 0.08158544450998306, 'train/grad_norm_pre_clip': 76.01701354980469, 'timing/data': 0.00023027999850455672, 'timing/model': 0.364896653991309, 'learning_rate/qwen_vl_interface': 6.4960516707752255e-06, 'learning_rate/action_model': 6.496051670775226e-05, 'epoch': 0.16}) +[2026-10-03 15:02:17,964][starVLA.training.train_starvla][INFO] - Step 2137, Loss: {'train/loss': 0.00017426397243980318, 'train/grad_norm_pre_clip': 0.39940956234931946, 'timing/data': 0.00019227000302635133, 'timing/model': 0.559750473999884, 'learning_rate/qwen_vl_interface': 6.493112646735851e-06, 'learning_rate/action_model': 6.493112646735851e-05, 'epoch': 0.16}) +[2026-10-03 15:02:18,612][starVLA.training.train_starvla][INFO] - Step 2138, Loss: {'train/loss': 3.440520595177077e-05, 'train/grad_norm_pre_clip': 0.007231310475617647, 'timing/data': 0.0002712599962251261, 'timing/model': 0.36498456400295254, 'learning_rate/qwen_vl_interface': 6.490173111699824e-06, 'learning_rate/action_model': 6.490173111699825e-05, 'epoch': 0.16}) +[2026-10-03 15:02:19,263][starVLA.training.train_starvla][INFO] - Step 2139, Loss: {'train/loss': 0.49065104126930237, 'train/grad_norm_pre_clip': 20.129060745239258, 'timing/data': 0.0002651100076036528, 'timing/model': 0.36545790400123224, 'learning_rate/qwen_vl_interface': 6.487233066875479e-06, 'learning_rate/action_model': 6.487233066875479e-05, 'epoch': 0.16}) +[2026-10-03 15:02:19,889][starVLA.training.train_starvla][INFO] - Step 2140, Loss: {'train/loss': 0.00032373916474170983, 'train/grad_norm_pre_clip': 0.13816702365875244, 'timing/data': 0.0003102999908151105, 'timing/model': 0.3424473869963549, 'learning_rate/qwen_vl_interface': 6.484292513471352e-06, 'learning_rate/action_model': 6.484292513471352e-05, 'epoch': 0.16}) +[2026-10-03 15:02:20,537][starVLA.training.train_starvla][INFO] - Step 2141, Loss: {'train/loss': 0.0003875504480674863, 'train/grad_norm_pre_clip': 0.17800815403461456, 'timing/data': 0.000310860006720759, 'timing/model': 0.36450487399997655, 'learning_rate/qwen_vl_interface': 6.4813514526961975e-06, 'learning_rate/action_model': 6.481351452696198e-05, 'epoch': 0.16}) +[2026-10-03 15:02:21,184][starVLA.training.train_starvla][INFO] - Step 2142, Loss: {'train/loss': 8.066747250268236e-05, 'train/grad_norm_pre_clip': 0.06542814522981644, 'timing/data': 0.0003050899977097288, 'timing/model': 0.365051313987351, 'learning_rate/qwen_vl_interface': 6.478409885758972e-06, 'learning_rate/action_model': 6.478409885758971e-05, 'epoch': 0.16}) +[2026-10-03 15:02:22,038][starVLA.training.train_starvla][INFO] - Step 2143, Loss: {'train/loss': 0.00011946324957534671, 'train/grad_norm_pre_clip': 0.037576332688331604, 'timing/data': 0.00024230999406427145, 'timing/model': 0.36497531400527805, 'learning_rate/qwen_vl_interface': 6.475467813868844e-06, 'learning_rate/action_model': 6.475467813868844e-05, 'epoch': 0.16}) +[2026-10-03 15:02:22,685][starVLA.training.train_starvla][INFO] - Step 2144, Loss: {'train/loss': 0.0006468796636909246, 'train/grad_norm_pre_clip': 0.5098770260810852, 'timing/data': 0.0003221200022380799, 'timing/model': 0.364940323997871, 'learning_rate/qwen_vl_interface': 6.472525238235187e-06, 'learning_rate/action_model': 6.472525238235186e-05, 'epoch': 0.16}) +[2026-10-03 15:02:23,332][starVLA.training.train_starvla][INFO] - Step 2145, Loss: {'train/loss': 0.35353460907936096, 'train/grad_norm_pre_clip': 36.247108459472656, 'timing/data': 0.0002869199961423874, 'timing/model': 0.3645413040067069, 'learning_rate/qwen_vl_interface': 6.4695821600675834e-06, 'learning_rate/action_model': 6.469582160067583e-05, 'epoch': 0.16}) +[2026-10-03 15:02:23,980][starVLA.training.train_starvla][INFO] - Step 2146, Loss: {'train/loss': 2.492908970452845e-05, 'train/grad_norm_pre_clip': 0.003969760611653328, 'timing/data': 0.00019393999536987394, 'timing/model': 0.36459976399783045, 'learning_rate/qwen_vl_interface': 6.466638580575824e-06, 'learning_rate/action_model': 6.466638580575824e-05, 'epoch': 0.16}) +[2026-10-03 15:02:24,628][starVLA.training.train_starvla][INFO] - Step 2147, Loss: {'train/loss': 6.903689791215584e-05, 'train/grad_norm_pre_clip': 0.02137874998152256, 'timing/data': 0.00028835999546572566, 'timing/model': 0.3647814740106696, 'learning_rate/qwen_vl_interface': 6.463694500969898e-06, 'learning_rate/action_model': 6.463694500969898e-05, 'epoch': 0.16}) +[2026-10-03 15:02:25,469][starVLA.training.train_starvla][INFO] - Step 2148, Loss: {'train/loss': 5.964513911749236e-05, 'train/grad_norm_pre_clip': 0.013930361717939377, 'timing/data': 0.0001941199880093336, 'timing/model': 0.5575806249980815, 'learning_rate/qwen_vl_interface': 6.460749922460008e-06, 'learning_rate/action_model': 6.460749922460008e-05, 'epoch': 0.16}) +[2026-10-03 15:02:26,117][starVLA.training.train_starvla][INFO] - Step 2149, Loss: {'train/loss': 0.00011874439951498061, 'train/grad_norm_pre_clip': 0.02670331485569477, 'timing/data': 0.00019612000323832035, 'timing/model': 0.3647016139875632, 'learning_rate/qwen_vl_interface': 6.457804846256559e-06, 'learning_rate/action_model': 6.457804846256559e-05, 'epoch': 0.16}) +[2026-10-03 15:02:26,764][starVLA.training.train_starvla][INFO] - Step 2150, Loss: {'train/loss': 0.008307685144245625, 'train/grad_norm_pre_clip': 32.601356506347656, 'timing/data': 0.00023366899404209107, 'timing/model': 0.3645413550111698, 'learning_rate/qwen_vl_interface': 6.45485927357016e-06, 'learning_rate/action_model': 6.454859273570159e-05, 'epoch': 0.16}) +[2026-10-03 15:02:27,414][starVLA.training.train_starvla][INFO] - Step 2151, Loss: {'train/loss': 0.03417506814002991, 'train/grad_norm_pre_clip': 98.92855834960938, 'timing/data': 0.00019895000150427222, 'timing/model': 0.3647466350084869, 'learning_rate/qwen_vl_interface': 6.451913205611626e-06, 'learning_rate/action_model': 6.451913205611626e-05, 'epoch': 0.16}) +[2026-10-03 15:02:28,062][starVLA.training.train_starvla][INFO] - Step 2152, Loss: {'train/loss': 0.5435673594474792, 'train/grad_norm_pre_clip': 33.670528411865234, 'timing/data': 0.0003538699966156855, 'timing/model': 0.3647672649967717, 'learning_rate/qwen_vl_interface': 6.448966643591974e-06, 'learning_rate/action_model': 6.448966643591974e-05, 'epoch': 0.16}) +[2026-10-03 15:02:28,902][starVLA.training.train_starvla][INFO] - Step 2153, Loss: {'train/loss': 0.027350934222340584, 'train/grad_norm_pre_clip': 52.37928009033203, 'timing/data': 0.0001943500101333484, 'timing/model': 0.5569885050063021, 'learning_rate/qwen_vl_interface': 6.446019588722423e-06, 'learning_rate/action_model': 6.446019588722422e-05, 'epoch': 0.16}) +[2026-10-03 15:02:29,549][starVLA.training.train_starvla][INFO] - Step 2154, Loss: {'train/loss': 0.018711159005761147, 'train/grad_norm_pre_clip': 17.657745361328125, 'timing/data': 0.00030853999487590045, 'timing/model': 0.364580143999774, 'learning_rate/qwen_vl_interface': 6.443072042214398e-06, 'learning_rate/action_model': 6.443072042214397e-05, 'epoch': 0.16}) +[2026-10-03 15:02:30,199][starVLA.training.train_starvla][INFO] - Step 2155, Loss: {'train/loss': 0.554317057132721, 'train/grad_norm_pre_clip': 20.965980529785156, 'timing/data': 0.0002606000052765012, 'timing/model': 0.36540164399775676, 'learning_rate/qwen_vl_interface': 6.440124005279521e-06, 'learning_rate/action_model': 6.440124005279521e-05, 'epoch': 0.16}) +[2026-10-03 15:02:30,848][starVLA.training.train_starvla][INFO] - Step 2156, Loss: {'train/loss': 0.00765227572992444, 'train/grad_norm_pre_clip': 17.805936813354492, 'timing/data': 0.00019685999723151326, 'timing/model': 0.3652408839989221, 'learning_rate/qwen_vl_interface': 6.43717547912962e-06, 'learning_rate/action_model': 6.437175479129621e-05, 'epoch': 0.16}) +[2026-10-03 15:02:31,494][starVLA.training.train_starvla][INFO] - Step 2157, Loss: {'train/loss': 0.45819079875946045, 'train/grad_norm_pre_clip': 15.30795669555664, 'timing/data': 0.0002644500054884702, 'timing/model': 0.36462879400642123, 'learning_rate/qwen_vl_interface': 6.434226464976725e-06, 'learning_rate/action_model': 6.434226464976724e-05, 'epoch': 0.16}) +[2026-10-03 15:02:32,142][starVLA.training.train_starvla][INFO] - Step 2158, Loss: {'train/loss': 0.00016350953956134617, 'train/grad_norm_pre_clip': 0.05128520354628563, 'timing/data': 0.00023926999710965902, 'timing/model': 0.3645669140096288, 'learning_rate/qwen_vl_interface': 6.431276964033062e-06, 'learning_rate/action_model': 6.431276964033062e-05, 'epoch': 0.16}) +[2026-10-03 15:02:32,989][starVLA.training.train_starvla][INFO] - Step 2159, Loss: {'train/loss': 9.074076660908759e-05, 'train/grad_norm_pre_clip': 0.015793366357684135, 'timing/data': 0.000297670005238615, 'timing/model': 0.3651288639957784, 'learning_rate/qwen_vl_interface': 6.428326977511059e-06, 'learning_rate/action_model': 6.428326977511058e-05, 'epoch': 0.16}) +[2026-10-03 15:02:33,637][starVLA.training.train_starvla][INFO] - Step 2160, Loss: {'train/loss': 0.056440334767103195, 'train/grad_norm_pre_clip': 78.20733642578125, 'timing/data': 0.00023438999778591096, 'timing/model': 0.3649385939934291, 'learning_rate/qwen_vl_interface': 6.425376506623344e-06, 'learning_rate/action_model': 6.425376506623344e-05, 'epoch': 0.16}) +[2026-10-03 15:02:34,285][starVLA.training.train_starvla][INFO] - Step 2161, Loss: {'train/loss': 0.0015065472107380629, 'train/grad_norm_pre_clip': 0.6840689182281494, 'timing/data': 0.00022323000303003937, 'timing/model': 0.36454433399194386, 'learning_rate/qwen_vl_interface': 6.422425552582747e-06, 'learning_rate/action_model': 6.422425552582747e-05, 'epoch': 0.16}) +[2026-10-03 15:02:34,932][starVLA.training.train_starvla][INFO] - Step 2162, Loss: {'train/loss': 0.0061814310029149055, 'train/grad_norm_pre_clip': 13.430424690246582, 'timing/data': 0.0003244700055802241, 'timing/model': 0.3643571149877971, 'learning_rate/qwen_vl_interface': 6.41947411660229e-06, 'learning_rate/action_model': 6.41947411660229e-05, 'epoch': 0.16}) +[2026-10-03 15:02:35,583][starVLA.training.train_starvla][INFO] - Step 2163, Loss: {'train/loss': 0.00027821457479149103, 'train/grad_norm_pre_clip': 0.0713602751493454, 'timing/data': 0.00036002999695483595, 'timing/model': 0.3653880140045658, 'learning_rate/qwen_vl_interface': 6.4165221998951995e-06, 'learning_rate/action_model': 6.4165221998952e-05, 'epoch': 0.16}) +[2026-10-03 15:02:36,417][starVLA.training.train_starvla][INFO] - Step 2164, Loss: {'train/loss': 0.00039050853229127824, 'train/grad_norm_pre_clip': 0.17502863705158234, 'timing/data': 0.0004328200011514127, 'timing/model': 0.5514045960007934, 'learning_rate/qwen_vl_interface': 6.413569803674896e-06, 'learning_rate/action_model': 6.413569803674896e-05, 'epoch': 0.16}) +[2026-10-03 15:02:37,063][starVLA.training.train_starvla][INFO] - Step 2165, Loss: {'train/loss': 0.25449806451797485, 'train/grad_norm_pre_clip': 24.080596923828125, 'timing/data': 0.00024542999744880944, 'timing/model': 0.3638619350094814, 'learning_rate/qwen_vl_interface': 6.410616929154997e-06, 'learning_rate/action_model': 6.410616929154996e-05, 'epoch': 0.16}) +[2026-10-03 15:02:37,709][starVLA.training.train_starvla][INFO] - Step 2166, Loss: {'train/loss': 6.214859604369849e-05, 'train/grad_norm_pre_clip': 0.010977465659379959, 'timing/data': 0.00024087999190669507, 'timing/model': 0.36407853399578016, 'learning_rate/qwen_vl_interface': 6.40766357754932e-06, 'learning_rate/action_model': 6.407663577549319e-05, 'epoch': 0.16}) +[2026-10-03 15:02:38,359][starVLA.training.train_starvla][INFO] - Step 2167, Loss: {'train/loss': 0.0005819592624902725, 'train/grad_norm_pre_clip': 0.15737131237983704, 'timing/data': 0.00027903998852707446, 'timing/model': 0.36500147500191815, 'learning_rate/qwen_vl_interface': 6.404709750071875e-06, 'learning_rate/action_model': 6.404709750071874e-05, 'epoch': 0.16}) +[2026-10-03 15:02:39,006][starVLA.training.train_starvla][INFO] - Step 2168, Loss: {'train/loss': 0.007382575422525406, 'train/grad_norm_pre_clip': 5.350575923919678, 'timing/data': 0.00029883001116104424, 'timing/model': 0.3645181150059216, 'learning_rate/qwen_vl_interface': 6.401755447936869e-06, 'learning_rate/action_model': 6.401755447936869e-05, 'epoch': 0.16}) +[2026-10-03 15:02:39,848][starVLA.training.train_starvla][INFO] - Step 2169, Loss: {'train/loss': 6.513665721286088e-05, 'train/grad_norm_pre_clip': 0.01359565183520317, 'timing/data': 0.0003086600045207888, 'timing/model': 0.5590146140020806, 'learning_rate/qwen_vl_interface': 6.3988006723587036e-06, 'learning_rate/action_model': 6.398800672358703e-05, 'epoch': 0.16}) +[2026-10-03 15:02:40,494][starVLA.training.train_starvla][INFO] - Step 2170, Loss: {'train/loss': 0.00019297076505608857, 'train/grad_norm_pre_clip': 0.04738068953156471, 'timing/data': 0.00019194000924471766, 'timing/model': 0.3638572649942944, 'learning_rate/qwen_vl_interface': 6.395845424551975e-06, 'learning_rate/action_model': 6.395845424551975e-05, 'epoch': 0.16}) +[2026-10-03 15:02:41,121][starVLA.training.train_starvla][INFO] - Step 2171, Loss: {'train/loss': 0.00012418802361935377, 'train/grad_norm_pre_clip': 0.02722904272377491, 'timing/data': 0.00028943999495822936, 'timing/model': 0.36445052399358246, 'learning_rate/qwen_vl_interface': 6.392889705731475e-06, 'learning_rate/action_model': 6.392889705731476e-05, 'epoch': 0.16}) +[2026-10-03 15:02:41,769][starVLA.training.train_starvla][INFO] - Step 2172, Loss: {'train/loss': 0.0027554496191442013, 'train/grad_norm_pre_clip': 2.0662147998809814, 'timing/data': 0.00031567999394610524, 'timing/model': 0.3648711749992799, 'learning_rate/qwen_vl_interface': 6.389933517112188e-06, 'learning_rate/action_model': 6.389933517112187e-05, 'epoch': 0.16}) +[2026-10-03 15:02:42,416][starVLA.training.train_starvla][INFO] - Step 2173, Loss: {'train/loss': 0.3511222302913666, 'train/grad_norm_pre_clip': 58.29799270629883, 'timing/data': 0.00028033000126015395, 'timing/model': 0.3642402139957994, 'learning_rate/qwen_vl_interface': 6.386976859909293e-06, 'learning_rate/action_model': 6.386976859909292e-05, 'epoch': 0.16}) +[2026-10-03 15:02:43,062][starVLA.training.train_starvla][INFO] - Step 2174, Loss: {'train/loss': 0.2173621952533722, 'train/grad_norm_pre_clip': 16.139102935791016, 'timing/data': 0.00025160999211948365, 'timing/model': 0.363925694997306, 'learning_rate/qwen_vl_interface': 6.384019735338156e-06, 'learning_rate/action_model': 6.384019735338156e-05, 'epoch': 0.16}) +[2026-10-03 15:02:43,902][starVLA.training.train_starvla][INFO] - Step 2175, Loss: {'train/loss': 0.00025792152155190706, 'train/grad_norm_pre_clip': 0.06275907158851624, 'timing/data': 0.0002873699995689094, 'timing/model': 0.36481971500325017, 'learning_rate/qwen_vl_interface': 6.381062144614342e-06, 'learning_rate/action_model': 6.381062144614342e-05, 'epoch': 0.16}) +[2026-10-03 15:02:44,548][starVLA.training.train_starvla][INFO] - Step 2176, Loss: {'train/loss': 0.012607041746377945, 'train/grad_norm_pre_clip': 5.458329677581787, 'timing/data': 0.00023813000007066876, 'timing/model': 0.3642398149968358, 'learning_rate/qwen_vl_interface': 6.378104088953605e-06, 'learning_rate/action_model': 6.378104088953605e-05, 'epoch': 0.16}) +[2026-10-03 15:02:45,194][starVLA.training.train_starvla][INFO] - Step 2177, Loss: {'train/loss': 0.0018739545485004783, 'train/grad_norm_pre_clip': 2.939892292022705, 'timing/data': 0.0002935800002887845, 'timing/model': 0.3640302939893445, 'learning_rate/qwen_vl_interface': 6.375145569571889e-06, 'learning_rate/action_model': 6.375145569571889e-05, 'epoch': 0.16}) +[2026-10-03 15:02:45,840][starVLA.training.train_starvla][INFO] - Step 2178, Loss: {'train/loss': 5.2486495405901223e-05, 'train/grad_norm_pre_clip': 0.009199941530823708, 'timing/data': 0.0003456900012679398, 'timing/model': 0.3638981949916342, 'learning_rate/qwen_vl_interface': 6.372186587685329e-06, 'learning_rate/action_model': 6.372186587685329e-05, 'epoch': 0.16}) +[2026-10-03 15:02:46,468][starVLA.training.train_starvla][INFO] - Step 2179, Loss: {'train/loss': 8.593299571657553e-05, 'train/grad_norm_pre_clip': 0.020422345027327538, 'timing/data': 0.0002858399966498837, 'timing/model': 0.3655695149936946, 'learning_rate/qwen_vl_interface': 6.369227144510252e-06, 'learning_rate/action_model': 6.369227144510252e-05, 'epoch': 0.16}) +[2026-10-03 15:02:47,318][starVLA.training.train_starvla][INFO] - Step 2180, Loss: {'train/loss': 0.36540892720222473, 'train/grad_norm_pre_clip': 32.396522521972656, 'timing/data': 0.00030825000430922955, 'timing/model': 0.3643363839946687, 'learning_rate/qwen_vl_interface': 6.3662672412631736e-06, 'learning_rate/action_model': 6.366267241263173e-05, 'epoch': 0.16}) +[2026-10-03 15:02:47,965][starVLA.training.train_starvla][INFO] - Step 2181, Loss: {'train/loss': 0.0005753912264481187, 'train/grad_norm_pre_clip': 0.36622095108032227, 'timing/data': 0.0002430789900245145, 'timing/model': 0.3643173349992139, 'learning_rate/qwen_vl_interface': 6.3633068791607954e-06, 'learning_rate/action_model': 6.363306879160795e-05, 'epoch': 0.16}) +[2026-10-03 15:02:48,611][starVLA.training.train_starvla][INFO] - Step 2182, Loss: {'train/loss': 0.0016600028611719608, 'train/grad_norm_pre_clip': 0.4365185797214508, 'timing/data': 0.0002227100048912689, 'timing/model': 0.36401043500518426, 'learning_rate/qwen_vl_interface': 6.360346059420014e-06, 'learning_rate/action_model': 6.360346059420014e-05, 'epoch': 0.16}) +[2026-10-03 15:02:49,262][starVLA.training.train_starvla][INFO] - Step 2183, Loss: {'train/loss': 0.0008711197879165411, 'train/grad_norm_pre_clip': 1.9674665927886963, 'timing/data': 0.00027789000887423754, 'timing/model': 0.36509762500645593, 'learning_rate/qwen_vl_interface': 6.357384783257909e-06, 'learning_rate/action_model': 6.357384783257908e-05, 'epoch': 0.16}) +[2026-10-03 15:02:49,912][starVLA.training.train_starvla][INFO] - Step 2184, Loss: {'train/loss': 0.0016462974017485976, 'train/grad_norm_pre_clip': 4.529562473297119, 'timing/data': 0.0003037999995285645, 'timing/model': 0.3663791840081103, 'learning_rate/qwen_vl_interface': 6.354423051891748e-06, 'learning_rate/action_model': 6.354423051891749e-05, 'epoch': 0.16}) +[2026-10-03 15:02:50,763][starVLA.training.train_starvla][INFO] - Step 2185, Loss: {'train/loss': 0.00348273990675807, 'train/grad_norm_pre_clip': 8.21141242980957, 'timing/data': 0.0001946499978657812, 'timing/model': 0.5682328339898959, 'learning_rate/qwen_vl_interface': 6.35146086653899e-06, 'learning_rate/action_model': 6.35146086653899e-05, 'epoch': 0.16}) +[2026-10-03 15:02:51,410][starVLA.training.train_starvla][INFO] - Step 2186, Loss: {'train/loss': 0.00057319545885548, 'train/grad_norm_pre_clip': 0.7741656303405762, 'timing/data': 0.00023317000886891037, 'timing/model': 0.36450483500084374, 'learning_rate/qwen_vl_interface': 6.348498228417273e-06, 'learning_rate/action_model': 6.348498228417273e-05, 'epoch': 0.16}) +[2026-10-03 15:02:52,037][starVLA.training.train_starvla][INFO] - Step 2187, Loss: {'train/loss': 0.31721949577331543, 'train/grad_norm_pre_clip': 24.035404205322266, 'timing/data': 0.0003188699920428917, 'timing/model': 0.34197030900395475, 'learning_rate/qwen_vl_interface': 6.34553513874443e-06, 'learning_rate/action_model': 6.34553513874443e-05, 'epoch': 0.16}) +[2026-10-03 15:02:52,685][starVLA.training.train_starvla][INFO] - Step 2188, Loss: {'train/loss': 0.00018222203652840108, 'train/grad_norm_pre_clip': 0.10730879753828049, 'timing/data': 0.00031335999665316194, 'timing/model': 0.36377557400555816, 'learning_rate/qwen_vl_interface': 6.342571598738473e-06, 'learning_rate/action_model': 6.342571598738472e-05, 'epoch': 0.16}) +[2026-10-03 15:02:53,313][starVLA.training.train_starvla][INFO] - Step 2189, Loss: {'train/loss': 0.001235931646078825, 'train/grad_norm_pre_clip': 0.33956825733184814, 'timing/data': 0.0003089300007559359, 'timing/model': 0.3635818150069099, 'learning_rate/qwen_vl_interface': 6.339607609617603e-06, 'learning_rate/action_model': 6.339607609617603e-05, 'epoch': 0.16}) +[2026-10-03 15:02:54,160][starVLA.training.train_starvla][INFO] - Step 2190, Loss: {'train/loss': 5.2113700803602114e-05, 'train/grad_norm_pre_clip': 0.007882189005613327, 'timing/data': 0.00020398999913595617, 'timing/model': 0.5647601249947911, 'learning_rate/qwen_vl_interface': 6.336643172600202e-06, 'learning_rate/action_model': 6.336643172600202e-05, 'epoch': 0.16}) +[2026-10-03 15:02:54,809][starVLA.training.train_starvla][INFO] - Step 2191, Loss: {'train/loss': 0.0024408132303506136, 'train/grad_norm_pre_clip': 0.47831064462661743, 'timing/data': 0.00035427900729700923, 'timing/model': 0.3646531949925702, 'learning_rate/qwen_vl_interface': 6.333678288904838e-06, 'learning_rate/action_model': 6.333678288904837e-05, 'epoch': 0.16}) +[2026-10-03 15:02:55,456][starVLA.training.train_starvla][INFO] - Step 2192, Loss: {'train/loss': 2.4481898435624316e-05, 'train/grad_norm_pre_clip': 0.006055192556232214, 'timing/data': 0.0002724000078160316, 'timing/model': 0.3641843049990712, 'learning_rate/qwen_vl_interface': 6.330712959750264e-06, 'learning_rate/action_model': 6.330712959750263e-05, 'epoch': 0.16}) +[2026-10-03 15:02:56,106][starVLA.training.train_starvla][INFO] - Step 2193, Loss: {'train/loss': 1.028227686882019, 'train/grad_norm_pre_clip': 31.437225341796875, 'timing/data': 0.00019758999405894428, 'timing/model': 0.36619405400415417, 'learning_rate/qwen_vl_interface': 6.3277471863554145e-06, 'learning_rate/action_model': 6.327747186355414e-05, 'epoch': 0.16}) +[2026-10-03 15:02:56,756][starVLA.training.train_starvla][INFO] - Step 2194, Loss: {'train/loss': 7.802362961228937e-05, 'train/grad_norm_pre_clip': 0.049093399196863174, 'timing/data': 0.0003295700007583946, 'timing/model': 0.36555269500240684, 'learning_rate/qwen_vl_interface': 6.324780969939406e-06, 'learning_rate/action_model': 6.324780969939405e-05, 'epoch': 0.16}) +[2026-10-03 15:02:57,409][starVLA.training.train_starvla][INFO] - Step 2195, Loss: {'train/loss': 0.0002080486447084695, 'train/grad_norm_pre_clip': 0.07298539578914642, 'timing/data': 0.0002916899975389242, 'timing/model': 0.3663540150009794, 'learning_rate/qwen_vl_interface': 6.32181431172154e-06, 'learning_rate/action_model': 6.32181431172154e-05, 'epoch': 0.16}) +[2026-10-03 15:02:58,259][starVLA.training.train_starvla][INFO] - Step 2196, Loss: {'train/loss': 8.474823698634282e-05, 'train/grad_norm_pre_clip': 0.0177226010710001, 'timing/data': 0.00030420999974012375, 'timing/model': 0.3656905550014926, 'learning_rate/qwen_vl_interface': 6.318847212921295e-06, 'learning_rate/action_model': 6.318847212921295e-05, 'epoch': 0.16}) +[2026-10-03 15:02:58,910][starVLA.training.train_starvla][INFO] - Step 2197, Loss: {'train/loss': 0.015555919148027897, 'train/grad_norm_pre_clip': 62.552181243896484, 'timing/data': 0.00030494999373331666, 'timing/model': 0.3659108050051145, 'learning_rate/qwen_vl_interface': 6.3158796747583364e-06, 'learning_rate/action_model': 6.315879674758336e-05, 'epoch': 0.16}) +[2026-10-03 15:02:59,558][starVLA.training.train_starvla][INFO] - Step 2198, Loss: {'train/loss': 0.0002457860391587019, 'train/grad_norm_pre_clip': 0.23357151448726654, 'timing/data': 0.00031632000172976404, 'timing/model': 0.36537462400156073, 'learning_rate/qwen_vl_interface': 6.312911698452503e-06, 'learning_rate/action_model': 6.312911698452502e-05, 'epoch': 0.16}) +[2026-10-03 15:03:00,211][starVLA.training.train_starvla][INFO] - Step 2199, Loss: {'train/loss': 0.00014321577327791601, 'train/grad_norm_pre_clip': 0.1475156843662262, 'timing/data': 0.000358819990651682, 'timing/model': 0.3663746340025682, 'learning_rate/qwen_vl_interface': 6.309943285223824e-06, 'learning_rate/action_model': 6.309943285223824e-05, 'epoch': 0.16}) +[2026-10-03 15:03:00,862][starVLA.training.train_starvla][INFO] - Step 2200, Loss: {'train/loss': 0.00019466193043626845, 'train/grad_norm_pre_clip': 0.27067747712135315, 'timing/data': 0.00031955000304151326, 'timing/model': 0.36614718500641175, 'learning_rate/qwen_vl_interface': 6.306974436292493e-06, 'learning_rate/action_model': 6.306974436292494e-05, 'epoch': 0.16}) +[2026-10-03 15:03:01,684][starVLA.training.train_starvla][INFO] - Step 2201, Loss: {'train/loss': 0.00047865105443634093, 'train/grad_norm_pre_clip': 0.13971160352230072, 'timing/data': 0.00029989000177010894, 'timing/model': 0.5602625150058884, 'learning_rate/qwen_vl_interface': 6.304005152878902e-06, 'learning_rate/action_model': 6.304005152878901e-05, 'epoch': 0.16}) +[2026-10-03 15:03:02,333][starVLA.training.train_starvla][INFO] - Step 2202, Loss: {'train/loss': 0.07133550196886063, 'train/grad_norm_pre_clip': 175.75912475585938, 'timing/data': 0.0002805800031637773, 'timing/model': 0.365240265004104, 'learning_rate/qwen_vl_interface': 6.301035436203602e-06, 'learning_rate/action_model': 6.301035436203602e-05, 'epoch': 0.16}) +[2026-10-03 15:03:02,985][starVLA.training.train_starvla][INFO] - Step 2203, Loss: {'train/loss': 0.3119058609008789, 'train/grad_norm_pre_clip': 29.220796585083008, 'timing/data': 0.00032483000541105866, 'timing/model': 0.3658091040124418, 'learning_rate/qwen_vl_interface': 6.298065287487337e-06, 'learning_rate/action_model': 6.298065287487337e-05, 'epoch': 0.16}) +[2026-10-03 15:03:03,636][starVLA.training.train_starvla][INFO] - Step 2204, Loss: {'train/loss': 5.952394712949172e-05, 'train/grad_norm_pre_clip': 0.02086465433239937, 'timing/data': 0.00031149999995250255, 'timing/model': 0.3662346650089603, 'learning_rate/qwen_vl_interface': 6.295094707951019e-06, 'learning_rate/action_model': 6.295094707951019e-05, 'epoch': 0.16}) +[2026-10-03 15:03:04,285][starVLA.training.train_starvla][INFO] - Step 2205, Loss: {'train/loss': 0.465915322303772, 'train/grad_norm_pre_clip': 62.10475540161133, 'timing/data': 0.0003319200041005388, 'timing/model': 0.36555711401160806, 'learning_rate/qwen_vl_interface': 6.292123698815745e-06, 'learning_rate/action_model': 6.292123698815745e-05, 'epoch': 0.16}) +[2026-10-03 15:03:04,935][starVLA.training.train_starvla][INFO] - Step 2206, Loss: {'train/loss': 0.0011450557503849268, 'train/grad_norm_pre_clip': 1.4755396842956543, 'timing/data': 0.000341649996698834, 'timing/model': 0.36595580499852076, 'learning_rate/qwen_vl_interface': 6.289152261302783e-06, 'learning_rate/action_model': 6.289152261302784e-05, 'epoch': 0.16}) +[2026-10-03 15:03:05,787][starVLA.training.train_starvla][INFO] - Step 2207, Loss: {'train/loss': 1.0114176273345947, 'train/grad_norm_pre_clip': 179.50405883789062, 'timing/data': 0.00023499000235460699, 'timing/model': 0.36719878400617745, 'learning_rate/qwen_vl_interface': 6.286180396633579e-06, 'learning_rate/action_model': 6.286180396633579e-05, 'epoch': 0.16}) +[2026-10-03 15:03:06,437][starVLA.training.train_starvla][INFO] - Step 2208, Loss: {'train/loss': 0.0003230470756534487, 'train/grad_norm_pre_clip': 0.16925430297851562, 'timing/data': 0.00031705999572295696, 'timing/model': 0.36593506499775685, 'learning_rate/qwen_vl_interface': 6.283208106029753e-06, 'learning_rate/action_model': 6.283208106029752e-05, 'epoch': 0.16}) +[2026-10-03 15:03:07,087][starVLA.training.train_starvla][INFO] - Step 2209, Loss: {'train/loss': 0.6308073997497559, 'train/grad_norm_pre_clip': 34.14139175415039, 'timing/data': 0.0003080100141232833, 'timing/model': 0.36588020500494167, 'learning_rate/qwen_vl_interface': 6.280235390713102e-06, 'learning_rate/action_model': 6.280235390713102e-05, 'epoch': 0.16}) +[2026-10-03 15:03:07,713][starVLA.training.train_starvla][INFO] - Step 2210, Loss: {'train/loss': 0.00016889936523512006, 'train/grad_norm_pre_clip': 0.0993756353855133, 'timing/data': 0.0003182700020261109, 'timing/model': 0.34262504801154137, 'learning_rate/qwen_vl_interface': 6.277262251905597e-06, 'learning_rate/action_model': 6.277262251905597e-05, 'epoch': 0.16}) +[2026-10-03 15:03:08,365][starVLA.training.train_starvla][INFO] - Step 2211, Loss: {'train/loss': 0.0012003635056316853, 'train/grad_norm_pre_clip': 2.106397867202759, 'timing/data': 0.00019804001203738153, 'timing/model': 0.36637675401289016, 'learning_rate/qwen_vl_interface': 6.274288690829385e-06, 'learning_rate/action_model': 6.274288690829384e-05, 'epoch': 0.16}) +[2026-10-03 15:03:09,221][starVLA.training.train_starvla][INFO] - Step 2212, Loss: {'train/loss': 0.2889546751976013, 'train/grad_norm_pre_clip': 17.689373016357422, 'timing/data': 0.0003365699958521873, 'timing/model': 0.3667406439926708, 'learning_rate/qwen_vl_interface': 6.271314708706781e-06, 'learning_rate/action_model': 6.271314708706782e-05, 'epoch': 0.16}) +[2026-10-03 15:03:09,871][starVLA.training.train_starvla][INFO] - Step 2213, Loss: {'train/loss': 0.47599753737449646, 'train/grad_norm_pre_clip': 29.632591247558594, 'timing/data': 0.00022555000032298267, 'timing/model': 0.36558996498933993, 'learning_rate/qwen_vl_interface': 6.26834030676028e-06, 'learning_rate/action_model': 6.26834030676028e-05, 'epoch': 0.16}) +[2026-10-03 15:03:10,520][starVLA.training.train_starvla][INFO] - Step 2214, Loss: {'train/loss': 0.5275378227233887, 'train/grad_norm_pre_clip': 33.444828033447266, 'timing/data': 0.00031875999411568046, 'timing/model': 0.36550813500070944, 'learning_rate/qwen_vl_interface': 6.265365486212545e-06, 'learning_rate/action_model': 6.265365486212545e-05, 'epoch': 0.16}) +[2026-10-03 15:03:11,172][starVLA.training.train_starvla][INFO] - Step 2215, Loss: {'train/loss': 0.014999163337051868, 'train/grad_norm_pre_clip': 5.739858150482178, 'timing/data': 0.0003409000055398792, 'timing/model': 0.36569021400646307, 'learning_rate/qwen_vl_interface': 6.26239024828641e-06, 'learning_rate/action_model': 6.26239024828641e-05, 'epoch': 0.16}) +[2026-10-03 15:03:11,823][starVLA.training.train_starvla][INFO] - Step 2216, Loss: {'train/loss': 0.000429314561188221, 'train/grad_norm_pre_clip': 0.11815470457077026, 'timing/data': 0.00023423999664373696, 'timing/model': 0.36604707399965264, 'learning_rate/qwen_vl_interface': 6.259414594204888e-06, 'learning_rate/action_model': 6.259414594204888e-05, 'epoch': 0.16}) +[2026-10-03 15:03:12,670][starVLA.training.train_starvla][INFO] - Step 2217, Loss: {'train/loss': 0.14973358809947968, 'train/grad_norm_pre_clip': 193.77069091796875, 'timing/data': 0.00023755000438541174, 'timing/model': 0.5631321550026769, 'learning_rate/qwen_vl_interface': 6.256438525191155e-06, 'learning_rate/action_model': 6.256438525191156e-05, 'epoch': 0.16}) +[2026-10-03 15:03:13,318][starVLA.training.train_starvla][INFO] - Step 2218, Loss: {'train/loss': 0.2900882661342621, 'train/grad_norm_pre_clip': 69.70355987548828, 'timing/data': 0.0002422700054012239, 'timing/model': 0.3650075249897782, 'learning_rate/qwen_vl_interface': 6.253462042468561e-06, 'learning_rate/action_model': 6.25346204246856e-05, 'epoch': 0.16}) +[2026-10-03 15:03:13,971][starVLA.training.train_starvla][INFO] - Step 2219, Loss: {'train/loss': 0.004512591287493706, 'train/grad_norm_pre_clip': 1.3565481901168823, 'timing/data': 0.0003628999984357506, 'timing/model': 0.36616684499313124, 'learning_rate/qwen_vl_interface': 6.2504851472606246e-06, 'learning_rate/action_model': 6.250485147260625e-05, 'epoch': 0.16}) +[2026-10-03 15:03:14,620][starVLA.training.train_starvla][INFO] - Step 2220, Loss: {'train/loss': 0.00041230564238503575, 'train/grad_norm_pre_clip': 0.12377793341875076, 'timing/data': 0.0002371099981246516, 'timing/model': 0.36535849499341566, 'learning_rate/qwen_vl_interface': 6.247507840791035e-06, 'learning_rate/action_model': 6.247507840791035e-05, 'epoch': 0.16}) +[2026-10-03 15:03:15,269][starVLA.training.train_starvla][INFO] - Step 2221, Loss: {'train/loss': 0.004736906383186579, 'train/grad_norm_pre_clip': 5.932778358459473, 'timing/data': 0.00019280999549664557, 'timing/model': 0.36562834499636665, 'learning_rate/qwen_vl_interface': 6.244530124283651e-06, 'learning_rate/action_model': 6.24453012428365e-05, 'epoch': 0.16}) +[2026-10-03 15:03:16,096][starVLA.training.train_starvla][INFO] - Step 2222, Loss: {'train/loss': 0.00011697055742843077, 'train/grad_norm_pre_clip': 0.03320654109120369, 'timing/data': 0.00023227999918162823, 'timing/model': 0.5654823039949406, 'learning_rate/qwen_vl_interface': 6.2415519989625016e-06, 'learning_rate/action_model': 6.241551998962501e-05, 'epoch': 0.16}) +[2026-10-03 15:03:16,748][starVLA.training.train_starvla][INFO] - Step 2223, Loss: {'train/loss': 0.005052762106060982, 'train/grad_norm_pre_clip': 5.346744060516357, 'timing/data': 0.00020172000222373754, 'timing/model': 0.3658921449969057, 'learning_rate/qwen_vl_interface': 6.238573466051781e-06, 'learning_rate/action_model': 6.23857346605178e-05, 'epoch': 0.16}) +[2026-10-03 15:03:17,377][starVLA.training.train_starvla][INFO] - Step 2224, Loss: {'train/loss': 0.026308847591280937, 'train/grad_norm_pre_clip': 38.9439811706543, 'timing/data': 0.0001971689925994724, 'timing/model': 0.36615452500700485, 'learning_rate/qwen_vl_interface': 6.2355945267758465e-06, 'learning_rate/action_model': 6.235594526775846e-05, 'epoch': 0.16}) +[2026-10-03 15:03:18,026][starVLA.training.train_starvla][INFO] - Step 2225, Loss: {'train/loss': 0.21505799889564514, 'train/grad_norm_pre_clip': 64.50859069824219, 'timing/data': 0.0001994900085264817, 'timing/model': 0.36556076499982737, 'learning_rate/qwen_vl_interface': 6.232615182359231e-06, 'learning_rate/action_model': 6.232615182359231e-05, 'epoch': 0.16}) +[2026-10-03 15:03:18,675][starVLA.training.train_starvla][INFO] - Step 2226, Loss: {'train/loss': 0.000286454102024436, 'train/grad_norm_pre_clip': 0.08636331558227539, 'timing/data': 0.00022652999905403703, 'timing/model': 0.3654981540021254, 'learning_rate/qwen_vl_interface': 6.229635434026634e-06, 'learning_rate/action_model': 6.229635434026634e-05, 'epoch': 0.16}) +[2026-10-03 15:03:19,337][starVLA.training.train_starvla][INFO] - Step 2227, Loss: {'train/loss': 0.0006971398834139109, 'train/grad_norm_pre_clip': 0.32214879989624023, 'timing/data': 0.0002876700018532574, 'timing/model': 0.3645841149991611, 'learning_rate/qwen_vl_interface': 6.226655283002912e-06, 'learning_rate/action_model': 6.226655283002911e-05, 'epoch': 0.16}) +[2026-10-03 15:03:20,188][starVLA.training.train_starvla][INFO] - Step 2228, Loss: {'train/loss': 0.25304296612739563, 'train/grad_norm_pre_clip': 34.97714614868164, 'timing/data': 0.0003099200112046674, 'timing/model': 0.36445099499542266, 'learning_rate/qwen_vl_interface': 6.223674730513096e-06, 'learning_rate/action_model': 6.223674730513096e-05, 'epoch': 0.16}) +[2026-10-03 15:03:20,814][starVLA.training.train_starvla][INFO] - Step 2229, Loss: {'train/loss': 0.0005693170242011547, 'train/grad_norm_pre_clip': 0.19280770421028137, 'timing/data': 0.00025414000265300274, 'timing/model': 0.3647397160093533, 'learning_rate/qwen_vl_interface': 6.220693777782375e-06, 'learning_rate/action_model': 6.220693777782375e-05, 'epoch': 0.16}) +[2026-10-03 15:03:21,439][starVLA.training.train_starvla][INFO] - Step 2230, Loss: {'train/loss': 0.0004251793725416064, 'train/grad_norm_pre_clip': 0.09775185585021973, 'timing/data': 0.0003074999985983595, 'timing/model': 0.34254314900317695, 'learning_rate/qwen_vl_interface': 6.21771242603611e-06, 'learning_rate/action_model': 6.21771242603611e-05, 'epoch': 0.16}) +[2026-10-03 15:03:22,101][starVLA.training.train_starvla][INFO] - Step 2231, Loss: {'train/loss': 0.0003803892177529633, 'train/grad_norm_pre_clip': 0.21518631279468536, 'timing/data': 0.0002609000075608492, 'timing/model': 0.36425786500331014, 'learning_rate/qwen_vl_interface': 6.21473067649982e-06, 'learning_rate/action_model': 6.21473067649982e-05, 'epoch': 0.16}) +[2026-10-03 15:03:22,748][starVLA.training.train_starvla][INFO] - Step 2232, Loss: {'train/loss': 0.4117005169391632, 'train/grad_norm_pre_clip': 26.941646575927734, 'timing/data': 0.00031583999225404114, 'timing/model': 0.36473825499706436, 'learning_rate/qwen_vl_interface': 6.211748530399189e-06, 'learning_rate/action_model': 6.211748530399189e-05, 'epoch': 0.16}) +[2026-10-03 15:03:23,596][starVLA.training.train_starvla][INFO] - Step 2233, Loss: {'train/loss': 0.000482297531561926, 'train/grad_norm_pre_clip': 0.11792152374982834, 'timing/data': 0.0003340700059197843, 'timing/model': 0.5653108849946875, 'learning_rate/qwen_vl_interface': 6.208765988960067e-06, 'learning_rate/action_model': 6.208765988960066e-05, 'epoch': 0.16}) +[2026-10-03 15:03:24,219][starVLA.training.train_starvla][INFO] - Step 2234, Loss: {'train/loss': 0.5617286562919617, 'train/grad_norm_pre_clip': 87.35165405273438, 'timing/data': 0.00032725899654906243, 'timing/model': 0.34101870899030473, 'learning_rate/qwen_vl_interface': 6.2057830534084615e-06, 'learning_rate/action_model': 6.205783053408461e-05, 'epoch': 0.16}) +[2026-10-03 15:03:24,881][starVLA.training.train_starvla][INFO] - Step 2235, Loss: {'train/loss': 0.23518407344818115, 'train/grad_norm_pre_clip': 33.688358306884766, 'timing/data': 0.0002310599957127124, 'timing/model': 0.3646107249951456, 'learning_rate/qwen_vl_interface': 6.202799724970546e-06, 'learning_rate/action_model': 6.202799724970546e-05, 'epoch': 0.16}) +[2026-10-03 15:03:25,528][starVLA.training.train_starvla][INFO] - Step 2236, Loss: {'train/loss': 0.005461178254336119, 'train/grad_norm_pre_clip': 1.3306769132614136, 'timing/data': 0.0003396399988560006, 'timing/model': 0.36443918499571737, 'learning_rate/qwen_vl_interface': 6.199816004872655e-06, 'learning_rate/action_model': 6.199816004872655e-05, 'epoch': 0.16}) +[2026-10-03 15:03:26,176][starVLA.training.train_starvla][INFO] - Step 2237, Loss: {'train/loss': 0.0005344115197658539, 'train/grad_norm_pre_clip': 0.26026666164398193, 'timing/data': 0.00028634999762289226, 'timing/model': 0.364540644994122, 'learning_rate/qwen_vl_interface': 6.196831894341281e-06, 'learning_rate/action_model': 6.19683189434128e-05, 'epoch': 0.16}) +[2026-10-03 15:03:26,822][starVLA.training.train_starvla][INFO] - Step 2238, Loss: {'train/loss': 0.2142127901315689, 'train/grad_norm_pre_clip': 11.491776466369629, 'timing/data': 0.00036751000152435154, 'timing/model': 0.36430116501287557, 'learning_rate/qwen_vl_interface': 6.1938473946030795e-06, 'learning_rate/action_model': 6.193847394603079e-05, 'epoch': 0.16}) +[2026-10-03 15:03:27,677][starVLA.training.train_starvla][INFO] - Step 2239, Loss: {'train/loss': 0.20292295515537262, 'train/grad_norm_pre_clip': 41.21123123168945, 'timing/data': 0.00032869000278878957, 'timing/model': 0.5608700860029785, 'learning_rate/qwen_vl_interface': 6.1908625068848695e-06, 'learning_rate/action_model': 6.19086250688487e-05, 'epoch': 0.16}) +[2026-10-03 15:03:28,323][starVLA.training.train_starvla][INFO] - Step 2240, Loss: {'train/loss': 0.12914472818374634, 'train/grad_norm_pre_clip': 245.4383544921875, 'timing/data': 0.0002544599992688745, 'timing/model': 0.3644728849903913, 'learning_rate/qwen_vl_interface': 6.187877232413619e-06, 'learning_rate/action_model': 6.18787723241362e-05, 'epoch': 0.16}) +[2026-10-03 15:03:28,970][starVLA.training.train_starvla][INFO] - Step 2241, Loss: {'train/loss': 0.0024594622664153576, 'train/grad_norm_pre_clip': 0.5229252576828003, 'timing/data': 0.0002857400104403496, 'timing/model': 0.36399833600444254, 'learning_rate/qwen_vl_interface': 6.184891572416467e-06, 'learning_rate/action_model': 6.184891572416466e-05, 'epoch': 0.16}) +[2026-10-03 15:03:29,617][starVLA.training.train_starvla][INFO] - Step 2242, Loss: {'train/loss': 0.24691402912139893, 'train/grad_norm_pre_clip': 23.39330291748047, 'timing/data': 0.0003017900016857311, 'timing/model': 0.36427339499641676, 'learning_rate/qwen_vl_interface': 6.181905528120703e-06, 'learning_rate/action_model': 6.181905528120703e-05, 'epoch': 0.16}) +[2026-10-03 15:03:30,266][starVLA.training.train_starvla][INFO] - Step 2243, Loss: {'train/loss': 0.005040089134126902, 'train/grad_norm_pre_clip': 0.9568135738372803, 'timing/data': 0.0002891000040108338, 'timing/model': 0.36487410600238945, 'learning_rate/qwen_vl_interface': 6.178919100753778e-06, 'learning_rate/action_model': 6.178919100753778e-05, 'epoch': 0.16}) +[2026-10-03 15:03:31,109][starVLA.training.train_starvla][INFO] - Step 2244, Loss: {'train/loss': 0.0006264318362809718, 'train/grad_norm_pre_clip': 0.13837607204914093, 'timing/data': 0.00036300999636296183, 'timing/model': 0.5603584159980528, 'learning_rate/qwen_vl_interface': 6.175932291543299e-06, 'learning_rate/action_model': 6.175932291543298e-05, 'epoch': 0.16}) +[2026-10-03 15:03:31,755][starVLA.training.train_starvla][INFO] - Step 2245, Loss: {'train/loss': 0.4833861291408539, 'train/grad_norm_pre_clip': 145.3201904296875, 'timing/data': 0.0001977600040845573, 'timing/model': 0.364520574992639, 'learning_rate/qwen_vl_interface': 6.172945101717029e-06, 'learning_rate/action_model': 6.17294510171703e-05, 'epoch': 0.16}) +[2026-10-03 15:03:32,402][starVLA.training.train_starvla][INFO] - Step 2246, Loss: {'train/loss': 0.0007398223388008773, 'train/grad_norm_pre_clip': 0.28190863132476807, 'timing/data': 0.0002679799945326522, 'timing/model': 0.3643161250074627, 'learning_rate/qwen_vl_interface': 6.169957532502892e-06, 'learning_rate/action_model': 6.169957532502892e-05, 'epoch': 0.16}) +[2026-10-03 15:03:33,052][starVLA.training.train_starvla][INFO] - Step 2247, Loss: {'train/loss': 0.01163129135966301, 'train/grad_norm_pre_clip': 2.2392444610595703, 'timing/data': 0.0001969699951587245, 'timing/model': 0.3648157249990618, 'learning_rate/qwen_vl_interface': 6.166969585128962e-06, 'learning_rate/action_model': 6.166969585128962e-05, 'epoch': 0.16}) +[2026-10-03 15:03:33,699][starVLA.training.train_starvla][INFO] - Step 2248, Loss: {'train/loss': 0.004438157193362713, 'train/grad_norm_pre_clip': 4.839122295379639, 'timing/data': 0.00020193999807816, 'timing/model': 0.3645015750080347, 'learning_rate/qwen_vl_interface': 6.163981260823473e-06, 'learning_rate/action_model': 6.163981260823473e-05, 'epoch': 0.16}) +[2026-10-03 15:03:34,546][starVLA.training.train_starvla][INFO] - Step 2249, Loss: {'train/loss': 0.21989724040031433, 'train/grad_norm_pre_clip': 29.392669677734375, 'timing/data': 0.0003195000026607886, 'timing/model': 0.5645305549987825, 'learning_rate/qwen_vl_interface': 6.16099256081481e-06, 'learning_rate/action_model': 6.16099256081481e-05, 'epoch': 0.16}) +[2026-10-03 15:03:35,193][starVLA.training.train_starvla][INFO] - Step 2250, Loss: {'train/loss': 0.0003715496277436614, 'train/grad_norm_pre_clip': 0.08316302299499512, 'timing/data': 0.00026174000231549144, 'timing/model': 0.3643032549880445, 'learning_rate/qwen_vl_interface': 6.15800348633152e-06, 'learning_rate/action_model': 6.15800348633152e-05, 'epoch': 0.16}) +[2026-10-03 15:03:35,844][starVLA.training.train_starvla][INFO] - Step 2251, Loss: {'train/loss': 0.006266840733587742, 'train/grad_norm_pre_clip': 3.155301809310913, 'timing/data': 0.0002993700036313385, 'timing/model': 0.3656964260007953, 'learning_rate/qwen_vl_interface': 6.155014038602293e-06, 'learning_rate/action_model': 6.155014038602292e-05, 'epoch': 0.16}) +[2026-10-03 15:03:36,491][starVLA.training.train_starvla][INFO] - Step 2252, Loss: {'train/loss': 0.01649489440023899, 'train/grad_norm_pre_clip': 13.736614227294922, 'timing/data': 0.00022645998978987336, 'timing/model': 0.36448633500549477, 'learning_rate/qwen_vl_interface': 6.152024218855979e-06, 'learning_rate/action_model': 6.15202421885598e-05, 'epoch': 0.16}) +[2026-10-03 15:03:37,117][starVLA.training.train_starvla][INFO] - Step 2253, Loss: {'train/loss': 0.0013821425382047892, 'train/grad_norm_pre_clip': 8.806182861328125, 'timing/data': 0.00019778999558184296, 'timing/model': 0.3651603149919538, 'learning_rate/qwen_vl_interface': 6.149034028321583e-06, 'learning_rate/action_model': 6.149034028321583e-05, 'epoch': 0.16}) +[2026-10-03 15:03:37,765][starVLA.training.train_starvla][INFO] - Step 2254, Loss: {'train/loss': 0.005568147171288729, 'train/grad_norm_pre_clip': 1.2644104957580566, 'timing/data': 0.00023739000607747585, 'timing/model': 0.3645525949978037, 'learning_rate/qwen_vl_interface': 6.1460434682282566e-06, 'learning_rate/action_model': 6.146043468228256e-05, 'epoch': 0.16}) +[2026-10-03 15:03:38,607][starVLA.training.train_starvla][INFO] - Step 2255, Loss: {'train/loss': 0.3335455060005188, 'train/grad_norm_pre_clip': 81.26727294921875, 'timing/data': 0.00024530000519007444, 'timing/model': 0.36543657499714755, 'learning_rate/qwen_vl_interface': 6.143052539805308e-06, 'learning_rate/action_model': 6.143052539805308e-05, 'epoch': 0.17}) +[2026-10-03 15:03:39,255][starVLA.training.train_starvla][INFO] - Step 2256, Loss: {'train/loss': 0.0011985014425590634, 'train/grad_norm_pre_clip': 0.30389341711997986, 'timing/data': 0.0002486199955455959, 'timing/model': 0.3646582950023003, 'learning_rate/qwen_vl_interface': 6.140061244282194e-06, 'learning_rate/action_model': 6.140061244282194e-05, 'epoch': 0.17}) +[2026-10-03 15:03:39,902][starVLA.training.train_starvla][INFO] - Step 2257, Loss: {'train/loss': 0.0027855639345943928, 'train/grad_norm_pre_clip': 1.8527195453643799, 'timing/data': 0.0003253900067647919, 'timing/model': 0.3643173860036768, 'learning_rate/qwen_vl_interface': 6.137069582888521e-06, 'learning_rate/action_model': 6.137069582888521e-05, 'epoch': 0.17}) +[2026-10-03 15:03:40,550][starVLA.training.train_starvla][INFO] - Step 2258, Loss: {'train/loss': 0.25077947974205017, 'train/grad_norm_pre_clip': 27.16257095336914, 'timing/data': 0.0002576300030341372, 'timing/model': 0.3647928860009415, 'learning_rate/qwen_vl_interface': 6.13407755685405e-06, 'learning_rate/action_model': 6.13407755685405e-05, 'epoch': 0.17}) +[2026-10-03 15:03:41,177][starVLA.training.train_starvla][INFO] - Step 2259, Loss: {'train/loss': 0.0006447173655033112, 'train/grad_norm_pre_clip': 0.6758893728256226, 'timing/data': 0.00024879998818505555, 'timing/model': 0.3422816289967159, 'learning_rate/qwen_vl_interface': 6.131085167408692e-06, 'learning_rate/action_model': 6.131085167408692e-05, 'epoch': 0.17}) +[2026-10-03 15:03:42,026][starVLA.training.train_starvla][INFO] - Step 2260, Loss: {'train/loss': 0.25009670853614807, 'train/grad_norm_pre_clip': 65.96241760253906, 'timing/data': 0.0002439900126773864, 'timing/model': 0.5644945759995608, 'learning_rate/qwen_vl_interface': 6.128092415782502e-06, 'learning_rate/action_model': 6.128092415782501e-05, 'epoch': 0.17}) +[2026-10-03 15:03:42,672][starVLA.training.train_starvla][INFO] - Step 2261, Loss: {'train/loss': 0.0003899865841958672, 'train/grad_norm_pre_clip': 0.14941024780273438, 'timing/data': 0.0002467099984642118, 'timing/model': 0.3642813449987443, 'learning_rate/qwen_vl_interface': 6.125099303205688e-06, 'learning_rate/action_model': 6.125099303205688e-05, 'epoch': 0.17}) +[2026-10-03 15:03:43,297][starVLA.training.train_starvla][INFO] - Step 2262, Loss: {'train/loss': 0.08270032703876495, 'train/grad_norm_pre_clip': 132.03700256347656, 'timing/data': 0.00030936000985093415, 'timing/model': 0.3639003760035848, 'learning_rate/qwen_vl_interface': 6.122105830908607e-06, 'learning_rate/action_model': 6.122105830908606e-05, 'epoch': 0.17}) +[2026-10-03 15:03:43,947][starVLA.training.train_starvla][INFO] - Step 2263, Loss: {'train/loss': 0.00010397963342256844, 'train/grad_norm_pre_clip': 0.02423202618956566, 'timing/data': 0.0003406000032555312, 'timing/model': 0.36511437500303145, 'learning_rate/qwen_vl_interface': 6.11911200012176e-06, 'learning_rate/action_model': 6.11911200012176e-05, 'epoch': 0.17}) +[2026-10-03 15:03:44,594][starVLA.training.train_starvla][INFO] - Step 2264, Loss: {'train/loss': 0.23832878470420837, 'train/grad_norm_pre_clip': 67.36530303955078, 'timing/data': 0.00026028999127447605, 'timing/model': 0.36450440500630066, 'learning_rate/qwen_vl_interface': 6.116117812075799e-06, 'learning_rate/action_model': 6.116117812075799e-05, 'epoch': 0.17}) +[2026-10-03 15:03:45,433][starVLA.training.train_starvla][INFO] - Step 2265, Loss: {'train/loss': 0.2725837826728821, 'train/grad_norm_pre_clip': 126.67133331298828, 'timing/data': 0.00024307999410666525, 'timing/model': 0.5558138969936408, 'learning_rate/qwen_vl_interface': 6.113123268001524e-06, 'learning_rate/action_model': 6.113123268001524e-05, 'epoch': 0.17}) +[2026-10-03 15:03:46,080][starVLA.training.train_starvla][INFO] - Step 2266, Loss: {'train/loss': 0.00010792507964652032, 'train/grad_norm_pre_clip': 0.031191235408186913, 'timing/data': 0.0002629200025694445, 'timing/model': 0.36445820500375703, 'learning_rate/qwen_vl_interface': 6.110128369129873e-06, 'learning_rate/action_model': 6.110128369129873e-05, 'epoch': 0.17}) +[2026-10-03 15:03:46,729][starVLA.training.train_starvla][INFO] - Step 2267, Loss: {'train/loss': 0.00017854259931482375, 'train/grad_norm_pre_clip': 0.05124146118760109, 'timing/data': 0.00025004999770317227, 'timing/model': 0.3642992450040765, 'learning_rate/qwen_vl_interface': 6.107133116691943e-06, 'learning_rate/action_model': 6.107133116691942e-05, 'epoch': 0.17}) +[2026-10-03 15:03:47,377][starVLA.training.train_starvla][INFO] - Step 2268, Loss: {'train/loss': 0.30114611983299255, 'train/grad_norm_pre_clip': 81.10243225097656, 'timing/data': 0.0003331699990667403, 'timing/model': 0.36466332599229645, 'learning_rate/qwen_vl_interface': 6.104137511918963e-06, 'learning_rate/action_model': 6.104137511918963e-05, 'epoch': 0.17}) +[2026-10-03 15:03:48,025][starVLA.training.train_starvla][INFO] - Step 2269, Loss: {'train/loss': 0.00014432409079745412, 'train/grad_norm_pre_clip': 0.055969394743442535, 'timing/data': 0.00024291999579872936, 'timing/model': 0.3646669660083717, 'learning_rate/qwen_vl_interface': 6.1011415560423155e-06, 'learning_rate/action_model': 6.101141556042316e-05, 'epoch': 0.17}) +[2026-10-03 15:03:48,875][starVLA.training.train_starvla][INFO] - Step 2270, Loss: {'train/loss': 7.530723087256774e-05, 'train/grad_norm_pre_clip': 0.020616022869944572, 'timing/data': 0.00020198999845888466, 'timing/model': 0.5673969649942592, 'learning_rate/qwen_vl_interface': 6.098145250293525e-06, 'learning_rate/action_model': 6.0981452502935246e-05, 'epoch': 0.17}) +[2026-10-03 15:03:49,525][starVLA.training.train_starvla][INFO] - Step 2271, Loss: {'train/loss': 0.42560142278671265, 'train/grad_norm_pre_clip': 62.14378356933594, 'timing/data': 0.00024467000912409276, 'timing/model': 0.36493530600273516, 'learning_rate/qwen_vl_interface': 6.09514859590426e-06, 'learning_rate/action_model': 6.09514859590426e-05, 'epoch': 0.17}) +[2026-10-03 15:03:50,173][starVLA.training.train_starvla][INFO] - Step 2272, Loss: {'train/loss': 0.00012046944902976975, 'train/grad_norm_pre_clip': 0.031073827296495438, 'timing/data': 0.0003131800040137023, 'timing/model': 0.3645840850076638, 'learning_rate/qwen_vl_interface': 6.09215159410633e-06, 'learning_rate/action_model': 6.09215159410633e-05, 'epoch': 0.17}) +[2026-10-03 15:03:50,822][starVLA.training.train_starvla][INFO] - Step 2273, Loss: {'train/loss': 0.0003457383718341589, 'train/grad_norm_pre_clip': 0.17302517592906952, 'timing/data': 0.00042378999933134764, 'timing/model': 0.3649581359932199, 'learning_rate/qwen_vl_interface': 6.089154246131692e-06, 'learning_rate/action_model': 6.089154246131692e-05, 'epoch': 0.17}) +[2026-10-03 15:03:51,470][starVLA.training.train_starvla][INFO] - Step 2274, Loss: {'train/loss': 0.26322442293167114, 'train/grad_norm_pre_clip': 353.140380859375, 'timing/data': 0.00026325999351684004, 'timing/model': 0.3654069360054564, 'learning_rate/qwen_vl_interface': 6.08615655321244e-06, 'learning_rate/action_model': 6.08615655321244e-05, 'epoch': 0.17}) +[2026-10-03 15:03:52,120][starVLA.training.train_starvla][INFO] - Step 2275, Loss: {'train/loss': 0.0003469028160907328, 'train/grad_norm_pre_clip': 0.14190377295017242, 'timing/data': 0.0003058199945371598, 'timing/model': 0.3645373559993459, 'learning_rate/qwen_vl_interface': 6.083158516580814e-06, 'learning_rate/action_model': 6.083158516580814e-05, 'epoch': 0.17}) +[2026-10-03 15:03:52,958][starVLA.training.train_starvla][INFO] - Step 2276, Loss: {'train/loss': 5.419155058916658e-05, 'train/grad_norm_pre_clip': 0.010233751498162746, 'timing/data': 0.00020102999405935407, 'timing/model': 0.3656731660012156, 'learning_rate/qwen_vl_interface': 6.080160137469192e-06, 'learning_rate/action_model': 6.0801601374691916e-05, 'epoch': 0.17}) +[2026-10-03 15:03:53,607][starVLA.training.train_starvla][INFO] - Step 2277, Loss: {'train/loss': 0.0041956715285778046, 'train/grad_norm_pre_clip': 13.147635459899902, 'timing/data': 0.00019202999828848988, 'timing/model': 0.36586266598897055, 'learning_rate/qwen_vl_interface': 6.077161417110097e-06, 'learning_rate/action_model': 6.0771614171100965e-05, 'epoch': 0.17}) +[2026-10-03 15:03:54,255][starVLA.training.train_starvla][INFO] - Step 2278, Loss: {'train/loss': 0.00042303497320972383, 'train/grad_norm_pre_clip': 0.16521263122558594, 'timing/data': 0.00026263999461662024, 'timing/model': 0.3651048550091218, 'learning_rate/qwen_vl_interface': 6.0741623567361864e-06, 'learning_rate/action_model': 6.074162356736186e-05, 'epoch': 0.17}) +[2026-10-03 15:03:54,906][starVLA.training.train_starvla][INFO] - Step 2279, Loss: {'train/loss': 0.00023449855507351458, 'train/grad_norm_pre_clip': 0.058807868510484695, 'timing/data': 0.00022183998953551054, 'timing/model': 0.3652025759947719, 'learning_rate/qwen_vl_interface': 6.071162957580262e-06, 'learning_rate/action_model': 6.071162957580262e-05, 'epoch': 0.17}) +[2026-10-03 15:03:55,555][starVLA.training.train_starvla][INFO] - Step 2280, Loss: {'train/loss': 0.00015521464229095727, 'train/grad_norm_pre_clip': 0.0786052867770195, 'timing/data': 0.00028121999639552087, 'timing/model': 0.3656342359900009, 'learning_rate/qwen_vl_interface': 6.068163220875266e-06, 'learning_rate/action_model': 6.068163220875266e-05, 'epoch': 0.17}) +[2026-10-03 15:03:56,409][starVLA.training.train_starvla][INFO] - Step 2281, Loss: {'train/loss': 0.00032693971297703683, 'train/grad_norm_pre_clip': 0.09958010911941528, 'timing/data': 0.0002070900081889704, 'timing/model': 0.3657045959989773, 'learning_rate/qwen_vl_interface': 6.065163147854273e-06, 'learning_rate/action_model': 6.065163147854273e-05, 'epoch': 0.17}) +[2026-10-03 15:03:57,058][starVLA.training.train_starvla][INFO] - Step 2282, Loss: {'train/loss': 0.000226593590923585, 'train/grad_norm_pre_clip': 0.06407101452350616, 'timing/data': 0.00019422000332269818, 'timing/model': 0.36527639599808026, 'learning_rate/qwen_vl_interface': 6.062162739750502e-06, 'learning_rate/action_model': 6.062162739750502e-05, 'epoch': 0.17}) +[2026-10-03 15:03:57,709][starVLA.training.train_starvla][INFO] - Step 2283, Loss: {'train/loss': 0.000305620429571718, 'train/grad_norm_pre_clip': 0.08594504743814468, 'timing/data': 0.0001941499940585345, 'timing/model': 0.3662724649911979, 'learning_rate/qwen_vl_interface': 6.059161997797306e-06, 'learning_rate/action_model': 6.0591619977973056e-05, 'epoch': 0.17}) +[2026-10-03 15:03:58,359][starVLA.training.train_starvla][INFO] - Step 2284, Loss: {'train/loss': 0.761148989200592, 'train/grad_norm_pre_clip': 26.72593879699707, 'timing/data': 0.0001975699997274205, 'timing/model': 0.3658455459953984, 'learning_rate/qwen_vl_interface': 6.056160923228178e-06, 'learning_rate/action_model': 6.056160923228178e-05, 'epoch': 0.17}) +[2026-10-03 15:03:59,008][starVLA.training.train_starvla][INFO] - Step 2285, Loss: {'train/loss': 0.5582135915756226, 'train/grad_norm_pre_clip': 66.92695617675781, 'timing/data': 0.0002954700030386448, 'timing/model': 0.3652022360038245, 'learning_rate/qwen_vl_interface': 6.0531595172767455e-06, 'learning_rate/action_model': 6.0531595172767455e-05, 'epoch': 0.17}) +[2026-10-03 15:03:59,854][starVLA.training.train_starvla][INFO] - Step 2286, Loss: {'train/loss': 0.00284438650123775, 'train/grad_norm_pre_clip': 2.317326545715332, 'timing/data': 0.00020052000763826072, 'timing/model': 0.5632279459969141, 'learning_rate/qwen_vl_interface': 6.050157781176774e-06, 'learning_rate/action_model': 6.050157781176774e-05, 'epoch': 0.17}) +[2026-10-03 15:04:00,506][starVLA.training.train_starvla][INFO] - Step 2287, Loss: {'train/loss': 0.264690637588501, 'train/grad_norm_pre_clip': 25.1779842376709, 'timing/data': 0.00029754999559372663, 'timing/model': 0.3656721449951874, 'learning_rate/qwen_vl_interface': 6.047155716162164e-06, 'learning_rate/action_model': 6.047155716162164e-05, 'epoch': 0.17}) +[2026-10-03 15:04:01,155][starVLA.training.train_starvla][INFO] - Step 2288, Loss: {'train/loss': 0.002792464103549719, 'train/grad_norm_pre_clip': 1.3308002948760986, 'timing/data': 0.0002920199913205579, 'timing/model': 0.36577999500150327, 'learning_rate/qwen_vl_interface': 6.044153323466948e-06, 'learning_rate/action_model': 6.044153323466948e-05, 'epoch': 0.17}) +[2026-10-03 15:04:01,807][starVLA.training.train_starvla][INFO] - Step 2289, Loss: {'train/loss': 0.7151415944099426, 'train/grad_norm_pre_clip': 27.244422912597656, 'timing/data': 0.0002445900026941672, 'timing/model': 0.3669549559999723, 'learning_rate/qwen_vl_interface': 6.0411506043253e-06, 'learning_rate/action_model': 6.0411506043253e-05, 'epoch': 0.17}) +[2026-10-03 15:04:02,456][starVLA.training.train_starvla][INFO] - Step 2290, Loss: {'train/loss': 2.60540000454057e-05, 'train/grad_norm_pre_clip': 0.006574665661901236, 'timing/data': 0.0002047700108960271, 'timing/model': 0.36614360599196516, 'learning_rate/qwen_vl_interface': 6.038147559971523e-06, 'learning_rate/action_model': 6.038147559971523e-05, 'epoch': 0.17}) +[2026-10-03 15:04:03,108][starVLA.training.train_starvla][INFO] - Step 2291, Loss: {'train/loss': 0.35495632886886597, 'train/grad_norm_pre_clip': 32.526611328125, 'timing/data': 0.0003969699901062995, 'timing/model': 0.3653422559873434, 'learning_rate/qwen_vl_interface': 6.035144191640055e-06, 'learning_rate/action_model': 6.035144191640055e-05, 'epoch': 0.17}) +[2026-10-03 15:04:03,947][starVLA.training.train_starvla][INFO] - Step 2292, Loss: {'train/loss': 8.879721281118691e-05, 'train/grad_norm_pre_clip': 0.023902935907244682, 'timing/data': 0.00022867000370752066, 'timing/model': 0.3648916759993881, 'learning_rate/qwen_vl_interface': 6.032140500565467e-06, 'learning_rate/action_model': 6.0321405005654665e-05, 'epoch': 0.17}) +[2026-10-03 15:04:04,594][starVLA.training.train_starvla][INFO] - Step 2293, Loss: {'train/loss': 0.00014958565589040518, 'train/grad_norm_pre_clip': 0.0810830146074295, 'timing/data': 0.00022131000878289342, 'timing/model': 0.36436045600567013, 'learning_rate/qwen_vl_interface': 6.029136487982463e-06, 'learning_rate/action_model': 6.029136487982463e-05, 'epoch': 0.17}) +[2026-10-03 15:04:05,242][starVLA.training.train_starvla][INFO] - Step 2294, Loss: {'train/loss': 0.0003506345092318952, 'train/grad_norm_pre_clip': 0.05844799429178238, 'timing/data': 0.00025298999389633536, 'timing/model': 0.3646884859917918, 'learning_rate/qwen_vl_interface': 6.02613215512588e-06, 'learning_rate/action_model': 6.0261321551258796e-05, 'epoch': 0.17}) +[2026-10-03 15:04:05,893][starVLA.training.train_starvla][INFO] - Step 2295, Loss: {'train/loss': 3.854732131003402e-05, 'train/grad_norm_pre_clip': 0.008682132698595524, 'timing/data': 0.00031088999821804464, 'timing/model': 0.3652204160025576, 'learning_rate/qwen_vl_interface': 6.023127503230684e-06, 'learning_rate/action_model': 6.023127503230683e-05, 'epoch': 0.17}) +[2026-10-03 15:04:06,542][starVLA.training.train_starvla][INFO] - Step 2296, Loss: {'train/loss': 4.65342600364238e-05, 'train/grad_norm_pre_clip': 0.00918351300060749, 'timing/data': 0.0003086600045207888, 'timing/model': 0.365078654998797, 'learning_rate/qwen_vl_interface': 6.020122533531976e-06, 'learning_rate/action_model': 6.0201225335319755e-05, 'epoch': 0.17}) +[2026-10-03 15:04:07,380][starVLA.training.train_starvla][INFO] - Step 2297, Loss: {'train/loss': 0.10191425681114197, 'train/grad_norm_pre_clip': 62.77193832397461, 'timing/data': 0.00033368000003974885, 'timing/model': 0.3647273149981629, 'learning_rate/qwen_vl_interface': 6.017117247264983e-06, 'learning_rate/action_model': 6.017117247264983e-05, 'epoch': 0.17}) +[2026-10-03 15:04:08,005][starVLA.training.train_starvla][INFO] - Step 2298, Loss: {'train/loss': 0.008254836313426495, 'train/grad_norm_pre_clip': 29.848472595214844, 'timing/data': 0.00033889900078065693, 'timing/model': 0.341719469986856, 'learning_rate/qwen_vl_interface': 6.014111645665069e-06, 'learning_rate/action_model': 6.0141116456650684e-05, 'epoch': 0.17}) +[2026-10-03 15:04:08,655][starVLA.training.train_starvla][INFO] - Step 2299, Loss: {'train/loss': 0.37870633602142334, 'train/grad_norm_pre_clip': 47.70650100708008, 'timing/data': 0.00033907999750226736, 'timing/model': 0.3654205860075308, 'learning_rate/qwen_vl_interface': 6.011105729967719e-06, 'learning_rate/action_model': 6.0111057299677184e-05, 'epoch': 0.17}) +[2026-10-03 15:04:09,303][starVLA.training.train_starvla][INFO] - Step 2300, Loss: {'train/loss': 0.0009683287935331464, 'train/grad_norm_pre_clip': 0.5753851532936096, 'timing/data': 0.00023348999093286693, 'timing/model': 0.3650130349997198, 'learning_rate/qwen_vl_interface': 6.008099501408552e-06, 'learning_rate/action_model': 6.008099501408552e-05, 'epoch': 0.17}) +[2026-10-03 15:04:09,951][starVLA.training.train_starvla][INFO] - Step 2301, Loss: {'train/loss': 9.203907393384725e-05, 'train/grad_norm_pre_clip': 0.02269919216632843, 'timing/data': 0.0003322300035506487, 'timing/model': 0.3645984359900467, 'learning_rate/qwen_vl_interface': 6.0050929612233165e-06, 'learning_rate/action_model': 6.005092961223316e-05, 'epoch': 0.17}) +[2026-10-03 15:04:10,768][starVLA.training.train_starvla][INFO] - Step 2302, Loss: {'train/loss': 0.0003312827320769429, 'train/grad_norm_pre_clip': 0.09665874391794205, 'timing/data': 0.00024900000425986946, 'timing/model': 0.5569613370025763, 'learning_rate/qwen_vl_interface': 6.002086110647888e-06, 'learning_rate/action_model': 6.002086110647888e-05, 'epoch': 0.17}) +[2026-10-03 15:04:11,418][starVLA.training.train_starvla][INFO] - Step 2303, Loss: {'train/loss': 0.0006237470079213381, 'train/grad_norm_pre_clip': 0.13083204627037048, 'timing/data': 0.0002514999941922724, 'timing/model': 0.3656021960050566, 'learning_rate/qwen_vl_interface': 5.999078950918267e-06, 'learning_rate/action_model': 5.999078950918267e-05, 'epoch': 0.17}) +[2026-10-03 15:04:12,066][starVLA.training.train_starvla][INFO] - Step 2304, Loss: {'train/loss': 0.00013530231080949306, 'train/grad_norm_pre_clip': 0.02844073250889778, 'timing/data': 0.0003484899934846908, 'timing/model': 0.36447516600310337, 'learning_rate/qwen_vl_interface': 5.996071483270585e-06, 'learning_rate/action_model': 5.996071483270585e-05, 'epoch': 0.17}) +[2026-10-03 15:04:12,713][starVLA.training.train_starvla][INFO] - Step 2305, Loss: {'train/loss': 0.0005027222796343267, 'train/grad_norm_pre_clip': 0.1373021900653839, 'timing/data': 0.0002451600012136623, 'timing/model': 0.3643962860078318, 'learning_rate/qwen_vl_interface': 5.993063708941098e-06, 'learning_rate/action_model': 5.993063708941098e-05, 'epoch': 0.17}) +[2026-10-03 15:04:13,361][starVLA.training.train_starvla][INFO] - Step 2306, Loss: {'train/loss': 0.00013419263996183872, 'train/grad_norm_pre_clip': 0.03808978572487831, 'timing/data': 0.0002410000015515834, 'timing/model': 0.36420088600425515, 'learning_rate/qwen_vl_interface': 5.990055629166187e-06, 'learning_rate/action_model': 5.990055629166187e-05, 'epoch': 0.17}) +[2026-10-03 15:04:14,211][starVLA.training.train_starvla][INFO] - Step 2307, Loss: {'train/loss': 0.00015616924792993814, 'train/grad_norm_pre_clip': 0.0418989360332489, 'timing/data': 0.00029749999521300197, 'timing/model': 0.5655096059927018, 'learning_rate/qwen_vl_interface': 5.987047245182363e-06, 'learning_rate/action_model': 5.987047245182362e-05, 'epoch': 0.17}) +[2026-10-03 15:04:14,859][starVLA.training.train_starvla][INFO] - Step 2308, Loss: {'train/loss': 0.003533479757606983, 'train/grad_norm_pre_clip': 3.3119728565216064, 'timing/data': 0.0002388700086157769, 'timing/model': 0.3648652370029595, 'learning_rate/qwen_vl_interface': 5.984038558226255e-06, 'learning_rate/action_model': 5.9840385582262546e-05, 'epoch': 0.17}) +[2026-10-03 15:04:15,507][starVLA.training.train_starvla][INFO] - Step 2309, Loss: {'train/loss': 0.0907762348651886, 'train/grad_norm_pre_clip': 162.5433349609375, 'timing/data': 0.0002751800057012588, 'timing/model': 0.36527049500728026, 'learning_rate/qwen_vl_interface': 5.981029569534625e-06, 'learning_rate/action_model': 5.981029569534625e-05, 'epoch': 0.17}) +[2026-10-03 15:04:16,155][starVLA.training.train_starvla][INFO] - Step 2310, Loss: {'train/loss': 0.00025076582096517086, 'train/grad_norm_pre_clip': 0.13519807159900665, 'timing/data': 0.00031835000845603645, 'timing/model': 0.36465475599106867, 'learning_rate/qwen_vl_interface': 5.978020280344351e-06, 'learning_rate/action_model': 5.9780202803443506e-05, 'epoch': 0.17}) +[2026-10-03 15:04:16,805][starVLA.training.train_starvla][INFO] - Step 2311, Loss: {'train/loss': 0.5371901392936707, 'train/grad_norm_pre_clip': 122.11034393310547, 'timing/data': 0.0003024600009666756, 'timing/model': 0.3647639560076641, 'learning_rate/qwen_vl_interface': 5.97501069189244e-06, 'learning_rate/action_model': 5.97501069189244e-05, 'epoch': 0.17}) +[2026-10-03 15:04:17,653][starVLA.training.train_starvla][INFO] - Step 2312, Loss: {'train/loss': 0.012899821624159813, 'train/grad_norm_pre_clip': 20.643230438232422, 'timing/data': 0.00032360000477638096, 'timing/model': 0.5643512659880798, 'learning_rate/qwen_vl_interface': 5.97200080541602e-06, 'learning_rate/action_model': 5.97200080541602e-05, 'epoch': 0.17}) +[2026-10-03 15:04:18,301][starVLA.training.train_starvla][INFO] - Step 2313, Loss: {'train/loss': 0.00018067468772642314, 'train/grad_norm_pre_clip': 0.0803632140159607, 'timing/data': 0.0003331399930175394, 'timing/model': 0.364992645001621, 'learning_rate/qwen_vl_interface': 5.968990622152341e-06, 'learning_rate/action_model': 5.9689906221523404e-05, 'epoch': 0.17}) +[2026-10-03 15:04:18,949][starVLA.training.train_starvla][INFO] - Step 2314, Loss: {'train/loss': 0.29702916741371155, 'train/grad_norm_pre_clip': 25.017282485961914, 'timing/data': 0.00031707999005448073, 'timing/model': 0.3644793359999312, 'learning_rate/qwen_vl_interface': 5.965980143338774e-06, 'learning_rate/action_model': 5.965980143338774e-05, 'epoch': 0.17}) +[2026-10-03 15:04:19,600][starVLA.training.train_starvla][INFO] - Step 2315, Loss: {'train/loss': 0.00016067721298895776, 'train/grad_norm_pre_clip': 0.08318532258272171, 'timing/data': 0.00026322998746763915, 'timing/model': 0.3648758259951137, 'learning_rate/qwen_vl_interface': 5.962969370212819e-06, 'learning_rate/action_model': 5.9629693702128185e-05, 'epoch': 0.17}) +[2026-10-03 15:04:20,248][starVLA.training.train_starvla][INFO] - Step 2316, Loss: {'train/loss': 8.427734428551048e-05, 'train/grad_norm_pre_clip': 0.02099236659705639, 'timing/data': 0.00024259999918285757, 'timing/model': 0.36523776600370184, 'learning_rate/qwen_vl_interface': 5.959958304012085e-06, 'learning_rate/action_model': 5.9599583040120845e-05, 'epoch': 0.17}) +[2026-10-03 15:04:20,897][starVLA.training.train_starvla][INFO] - Step 2317, Loss: {'train/loss': 0.0004254758241586387, 'train/grad_norm_pre_clip': 0.12041013687849045, 'timing/data': 0.00020002000383101404, 'timing/model': 0.36505455600854475, 'learning_rate/qwen_vl_interface': 5.956946945974309e-06, 'learning_rate/action_model': 5.9569469459743085e-05, 'epoch': 0.17}) +[2026-10-03 15:04:21,738][starVLA.training.train_starvla][INFO] - Step 2318, Loss: {'train/loss': 0.0001409201795468107, 'train/grad_norm_pre_clip': 0.03548267111182213, 'timing/data': 0.00020167999900877476, 'timing/model': 0.3664703460090095, 'learning_rate/qwen_vl_interface': 5.953935297337347e-06, 'learning_rate/action_model': 5.953935297337347e-05, 'epoch': 0.17}) +[2026-10-03 15:04:22,366][starVLA.training.train_starvla][INFO] - Step 2319, Loss: {'train/loss': 0.0002270155819132924, 'train/grad_norm_pre_clip': 0.14598724246025085, 'timing/data': 0.00023542999406345189, 'timing/model': 0.3653477159969043, 'learning_rate/qwen_vl_interface': 5.950923359339175e-06, 'learning_rate/action_model': 5.950923359339174e-05, 'epoch': 0.17}) +[2026-10-03 15:04:23,015][starVLA.training.train_starvla][INFO] - Step 2320, Loss: {'train/loss': 0.00023163466539699584, 'train/grad_norm_pre_clip': 0.0629914179444313, 'timing/data': 0.00025689999165479094, 'timing/model': 0.3651008059969172, 'learning_rate/qwen_vl_interface': 5.947911133217886e-06, 'learning_rate/action_model': 5.947911133217886e-05, 'epoch': 0.17}) +[2026-10-03 15:04:23,663][starVLA.training.train_starvla][INFO] - Step 2321, Loss: {'train/loss': 0.0002015366335399449, 'train/grad_norm_pre_clip': 0.04883567616343498, 'timing/data': 0.0002481700066709891, 'timing/model': 0.3653337959985947, 'learning_rate/qwen_vl_interface': 5.944898620211693e-06, 'learning_rate/action_model': 5.944898620211693e-05, 'epoch': 0.17}) +[2026-10-03 15:04:24,311][starVLA.training.train_starvla][INFO] - Step 2322, Loss: {'train/loss': 0.00011684455967042595, 'train/grad_norm_pre_clip': 0.026214633136987686, 'timing/data': 0.000328759997501038, 'timing/model': 0.36482399600208737, 'learning_rate/qwen_vl_interface': 5.941885821558925e-06, 'learning_rate/action_model': 5.9418858215589245e-05, 'epoch': 0.17}) +[2026-10-03 15:04:25,161][starVLA.training.train_starvla][INFO] - Step 2323, Loss: {'train/loss': 0.0005338148912414908, 'train/grad_norm_pre_clip': 0.126450315117836, 'timing/data': 0.00030497000261675566, 'timing/model': 0.36530339599994477, 'learning_rate/qwen_vl_interface': 5.938872738498031e-06, 'learning_rate/action_model': 5.93887273849803e-05, 'epoch': 0.17}) +[2026-10-03 15:04:25,810][starVLA.training.train_starvla][INFO] - Step 2324, Loss: {'train/loss': 8.59066640259698e-05, 'train/grad_norm_pre_clip': 0.022891908884048462, 'timing/data': 0.0002261100016767159, 'timing/model': 0.3655923159967642, 'learning_rate/qwen_vl_interface': 5.935859372267575e-06, 'learning_rate/action_model': 5.935859372267575e-05, 'epoch': 0.17}) +[2026-10-03 15:04:26,458][starVLA.training.train_starvla][INFO] - Step 2325, Loss: {'train/loss': 0.0004556066996883601, 'train/grad_norm_pre_clip': 0.21664215624332428, 'timing/data': 0.00024176998704206198, 'timing/model': 0.36519311599840876, 'learning_rate/qwen_vl_interface': 5.9328457241062396e-06, 'learning_rate/action_model': 5.9328457241062394e-05, 'epoch': 0.17}) +[2026-10-03 15:04:27,106][starVLA.training.train_starvla][INFO] - Step 2326, Loss: {'train/loss': 0.0035151115152984858, 'train/grad_norm_pre_clip': 8.97152042388916, 'timing/data': 0.00022602999524679035, 'timing/model': 0.36471212600008585, 'learning_rate/qwen_vl_interface': 5.929831795252819e-06, 'learning_rate/action_model': 5.9298317952528184e-05, 'epoch': 0.17}) +[2026-10-03 15:04:27,757][starVLA.training.train_starvla][INFO] - Step 2327, Loss: {'train/loss': 0.0001350257807644084, 'train/grad_norm_pre_clip': 0.0569380521774292, 'timing/data': 0.0002587890048744157, 'timing/model': 0.3655530769901816, 'learning_rate/qwen_vl_interface': 5.926817586946227e-06, 'learning_rate/action_model': 5.926817586946227e-05, 'epoch': 0.17}) +[2026-10-03 15:04:28,600][starVLA.training.train_starvla][INFO] - Step 2328, Loss: {'train/loss': 0.046257585287094116, 'train/grad_norm_pre_clip': 18.753826141357422, 'timing/data': 0.00023761999909766018, 'timing/model': 0.5583817379956599, 'learning_rate/qwen_vl_interface': 5.923803100425491e-06, 'learning_rate/action_model': 5.92380310042549e-05, 'epoch': 0.17}) +[2026-10-03 15:04:29,225][starVLA.training.train_starvla][INFO] - Step 2329, Loss: {'train/loss': 0.24669678509235382, 'train/grad_norm_pre_clip': 78.52474212646484, 'timing/data': 0.00033241999335587025, 'timing/model': 0.3422603189974325, 'learning_rate/qwen_vl_interface': 5.92078833692975e-06, 'learning_rate/action_model': 5.92078833692975e-05, 'epoch': 0.17}) +[2026-10-03 15:04:29,874][starVLA.training.train_starvla][INFO] - Step 2330, Loss: {'train/loss': 0.00042462232522666454, 'train/grad_norm_pre_clip': 0.13830526173114777, 'timing/data': 0.0004239599948050454, 'timing/model': 0.36426613600633573, 'learning_rate/qwen_vl_interface': 5.917773297698264e-06, 'learning_rate/action_model': 5.9177732976982636e-05, 'epoch': 0.17}) +[2026-10-03 15:04:30,525][starVLA.training.train_starvla][INFO] - Step 2331, Loss: {'train/loss': 0.00014624654431827366, 'train/grad_norm_pre_clip': 0.03718622028827667, 'timing/data': 0.00028819999715778977, 'timing/model': 0.36514850599633064, 'learning_rate/qwen_vl_interface': 5.9147579839704e-06, 'learning_rate/action_model': 5.9147579839704e-05, 'epoch': 0.17}) +[2026-10-03 15:04:31,150][starVLA.training.train_starvla][INFO] - Step 2332, Loss: {'train/loss': 0.00014173146337270737, 'train/grad_norm_pre_clip': 0.08913042396306992, 'timing/data': 0.0001933699968503788, 'timing/model': 0.34258640999905765, 'learning_rate/qwen_vl_interface': 5.911742396985637e-06, 'learning_rate/action_model': 5.9117423969856364e-05, 'epoch': 0.17}) +[2026-10-03 15:04:31,982][starVLA.training.train_starvla][INFO] - Step 2333, Loss: {'train/loss': 5.597611016128212e-05, 'train/grad_norm_pre_clip': 0.016468089073896408, 'timing/data': 0.00024248000408988446, 'timing/model': 0.5493545089993859, 'learning_rate/qwen_vl_interface': 5.908726537983571e-06, 'learning_rate/action_model': 5.908726537983571e-05, 'epoch': 0.17}) +[2026-10-03 15:04:32,631][starVLA.training.train_starvla][INFO] - Step 2334, Loss: {'train/loss': 1.1223058700561523, 'train/grad_norm_pre_clip': 95.25594329833984, 'timing/data': 0.00030140999297145754, 'timing/model': 0.36467266600811854, 'learning_rate/qwen_vl_interface': 5.9057104082039085e-06, 'learning_rate/action_model': 5.905710408203908e-05, 'epoch': 0.17}) +[2026-10-03 15:04:33,281][starVLA.training.train_starvla][INFO] - Step 2335, Loss: {'train/loss': 6.983179628150538e-05, 'train/grad_norm_pre_clip': 0.014118017628788948, 'timing/data': 0.00028030999237671494, 'timing/model': 0.3648641559993848, 'learning_rate/qwen_vl_interface': 5.902694008886466e-06, 'learning_rate/action_model': 5.902694008886466e-05, 'epoch': 0.17}) +[2026-10-03 15:04:33,930][starVLA.training.train_starvla][INFO] - Step 2336, Loss: {'train/loss': 0.00023575875093229115, 'train/grad_norm_pre_clip': 0.09273774176836014, 'timing/data': 0.000252760000876151, 'timing/model': 0.36529813600645866, 'learning_rate/qwen_vl_interface': 5.8996773412711694e-06, 'learning_rate/action_model': 5.899677341271169e-05, 'epoch': 0.17}) +[2026-10-03 15:04:34,578][starVLA.training.train_starvla][INFO] - Step 2337, Loss: {'train/loss': 0.2612975239753723, 'train/grad_norm_pre_clip': 108.19840240478516, 'timing/data': 0.00031416000274475664, 'timing/model': 0.365512864998891, 'learning_rate/qwen_vl_interface': 5.896660406598059e-06, 'learning_rate/action_model': 5.896660406598059e-05, 'epoch': 0.17}) +[2026-10-03 15:04:35,423][starVLA.training.train_starvla][INFO] - Step 2338, Loss: {'train/loss': 4.9504909839015454e-05, 'train/grad_norm_pre_clip': 0.010194863192737103, 'timing/data': 0.00024873000802472234, 'timing/model': 0.5616669459996046, 'learning_rate/qwen_vl_interface': 5.893643206107283e-06, 'learning_rate/action_model': 5.893643206107282e-05, 'epoch': 0.17}) +[2026-10-03 15:04:36,076][starVLA.training.train_starvla][INFO] - Step 2339, Loss: {'train/loss': 0.4239082634449005, 'train/grad_norm_pre_clip': 37.461063385009766, 'timing/data': 0.00031936999585013837, 'timing/model': 0.3651698660105467, 'learning_rate/qwen_vl_interface': 5.890625741039096e-06, 'learning_rate/action_model': 5.890625741039095e-05, 'epoch': 0.17}) +[2026-10-03 15:04:36,726][starVLA.training.train_starvla][INFO] - Step 2340, Loss: {'train/loss': 7.864637882448733e-05, 'train/grad_norm_pre_clip': 0.04894861951470375, 'timing/data': 0.0002200499875470996, 'timing/model': 0.36565604600764345, 'learning_rate/qwen_vl_interface': 5.887608012633864e-06, 'learning_rate/action_model': 5.887608012633864e-05, 'epoch': 0.17}) +[2026-10-03 15:04:37,351][starVLA.training.train_starvla][INFO] - Step 2341, Loss: {'train/loss': 4.7194356739055365e-05, 'train/grad_norm_pre_clip': 0.02831052616238594, 'timing/data': 0.00025082999491132796, 'timing/model': 0.34247158899961505, 'learning_rate/qwen_vl_interface': 5.884590022132064e-06, 'learning_rate/action_model': 5.884590022132064e-05, 'epoch': 0.17}) +[2026-10-03 15:04:37,977][starVLA.training.train_starvla][INFO] - Step 2342, Loss: {'train/loss': 5.2363931899890304e-05, 'train/grad_norm_pre_clip': 0.015907736495137215, 'timing/data': 0.0003466099879005924, 'timing/model': 0.36409310599265154, 'learning_rate/qwen_vl_interface': 5.881571770774278e-06, 'learning_rate/action_model': 5.881571770774278e-05, 'epoch': 0.17}) +[2026-10-03 15:04:38,629][starVLA.training.train_starvla][INFO] - Step 2343, Loss: {'train/loss': 0.008584631606936455, 'train/grad_norm_pre_clip': 7.880939960479736, 'timing/data': 0.00029067999275866896, 'timing/model': 0.36648902599699795, 'learning_rate/qwen_vl_interface': 5.878553259801192e-06, 'learning_rate/action_model': 5.878553259801192e-05, 'epoch': 0.17}) +[2026-10-03 15:04:39,473][starVLA.training.train_starvla][INFO] - Step 2344, Loss: {'train/loss': 0.0015917349373921752, 'train/grad_norm_pre_clip': 2.671891212463379, 'timing/data': 0.00024268000561278313, 'timing/model': 0.5611428470001556, 'learning_rate/qwen_vl_interface': 5.8755344904536025e-06, 'learning_rate/action_model': 5.8755344904536025e-05, 'epoch': 0.17}) +[2026-10-03 15:04:40,122][starVLA.training.train_starvla][INFO] - Step 2345, Loss: {'train/loss': 0.003675401210784912, 'train/grad_norm_pre_clip': 2.3601834774017334, 'timing/data': 0.00024465999740641564, 'timing/model': 0.36523938599566463, 'learning_rate/qwen_vl_interface': 5.872515463972415e-06, 'learning_rate/action_model': 5.8725154639724146e-05, 'epoch': 0.17}) +[2026-10-03 15:04:40,770][starVLA.training.train_starvla][INFO] - Step 2346, Loss: {'train/loss': 0.2772556245326996, 'train/grad_norm_pre_clip': 75.94479370117188, 'timing/data': 0.00030658001196570694, 'timing/model': 0.3651108850026503, 'learning_rate/qwen_vl_interface': 5.869496181598633e-06, 'learning_rate/action_model': 5.869496181598633e-05, 'epoch': 0.17}) +[2026-10-03 15:04:41,421][starVLA.training.train_starvla][INFO] - Step 2347, Loss: {'train/loss': 0.00015397954848594964, 'train/grad_norm_pre_clip': 0.07504050433635712, 'timing/data': 0.0003037799906451255, 'timing/model': 0.36501855599635746, 'learning_rate/qwen_vl_interface': 5.866476644573371e-06, 'learning_rate/action_model': 5.866476644573371e-05, 'epoch': 0.17}) +[2026-10-03 15:04:42,072][starVLA.training.train_starvla][INFO] - Step 2348, Loss: {'train/loss': 9.451160440221429e-05, 'train/grad_norm_pre_clip': 0.05371949449181557, 'timing/data': 0.00035913000465370715, 'timing/model': 0.36571493599331006, 'learning_rate/qwen_vl_interface': 5.863456854137848e-06, 'learning_rate/action_model': 5.8634568541378474e-05, 'epoch': 0.17}) +[2026-10-03 15:04:42,912][starVLA.training.train_starvla][INFO] - Step 2349, Loss: {'train/loss': 0.2050182968378067, 'train/grad_norm_pre_clip': 82.100341796875, 'timing/data': 0.0002543099981267005, 'timing/model': 0.5558821279992117, 'learning_rate/qwen_vl_interface': 5.860436811533383e-06, 'learning_rate/action_model': 5.860436811533383e-05, 'epoch': 0.17}) +[2026-10-03 15:04:43,561][starVLA.training.train_starvla][INFO] - Step 2350, Loss: {'train/loss': 0.03414842486381531, 'train/grad_norm_pre_clip': 45.79191589355469, 'timing/data': 0.0002448899904266, 'timing/model': 0.3654987159970915, 'learning_rate/qwen_vl_interface': 5.857416518001404e-06, 'learning_rate/action_model': 5.857416518001404e-05, 'epoch': 0.17}) +[2026-10-03 15:04:44,211][starVLA.training.train_starvla][INFO] - Step 2351, Loss: {'train/loss': 0.00013202654372435063, 'train/grad_norm_pre_clip': 0.08269640058279037, 'timing/data': 0.00019727999460883439, 'timing/model': 0.36511235700163525, 'learning_rate/qwen_vl_interface': 5.854395974783439e-06, 'learning_rate/action_model': 5.8543959747834386e-05, 'epoch': 0.17}) +[2026-10-03 15:04:44,861][starVLA.training.train_starvla][INFO] - Step 2352, Loss: {'train/loss': 6.402264989446849e-05, 'train/grad_norm_pre_clip': 0.029729602858424187, 'timing/data': 0.0003402799920877442, 'timing/model': 0.36541780599509366, 'learning_rate/qwen_vl_interface': 5.851375183121119e-06, 'learning_rate/action_model': 5.8513751831211186e-05, 'epoch': 0.17}) +[2026-10-03 15:04:45,510][starVLA.training.train_starvla][INFO] - Step 2353, Loss: {'train/loss': 0.8103465437889099, 'train/grad_norm_pre_clip': 122.42471313476562, 'timing/data': 0.00024256001051981002, 'timing/model': 0.365477095998358, 'learning_rate/qwen_vl_interface': 5.848354144256175e-06, 'learning_rate/action_model': 5.8483541442561754e-05, 'epoch': 0.17}) +[2026-10-03 15:04:46,158][starVLA.training.train_starvla][INFO] - Step 2354, Loss: {'train/loss': 7.748532880214043e-06, 'train/grad_norm_pre_clip': 0.0012957638828083873, 'timing/data': 0.00025553999876137823, 'timing/model': 0.36483813500672113, 'learning_rate/qwen_vl_interface': 5.845332859430446e-06, 'learning_rate/action_model': 5.845332859430446e-05, 'epoch': 0.17}) +[2026-10-03 15:04:47,003][starVLA.training.train_starvla][INFO] - Step 2355, Loss: {'train/loss': 0.0010888290125876665, 'train/grad_norm_pre_clip': 0.28791776299476624, 'timing/data': 0.00030824000714346766, 'timing/model': 0.36486958600289654, 'learning_rate/qwen_vl_interface': 5.842311329885866e-06, 'learning_rate/action_model': 5.842311329885866e-05, 'epoch': 0.17}) +[2026-10-03 15:04:47,651][starVLA.training.train_starvla][INFO] - Step 2356, Loss: {'train/loss': 0.06282884627580643, 'train/grad_norm_pre_clip': 115.42430877685547, 'timing/data': 0.0002638000005390495, 'timing/model': 0.36502657701203134, 'learning_rate/qwen_vl_interface': 5.839289556864472e-06, 'learning_rate/action_model': 5.8392895568644714e-05, 'epoch': 0.17}) +[2026-10-03 15:04:48,298][starVLA.training.train_starvla][INFO] - Step 2357, Loss: {'train/loss': 2.2320506104733795e-05, 'train/grad_norm_pre_clip': 0.007095027714967728, 'timing/data': 0.00031543000659439713, 'timing/model': 0.3648847359872889, 'learning_rate/qwen_vl_interface': 5.836267541608398e-06, 'learning_rate/action_model': 5.836267541608398e-05, 'epoch': 0.17}) +[2026-10-03 15:04:48,947][starVLA.training.train_starvla][INFO] - Step 2358, Loss: {'train/loss': 0.00042427369044162333, 'train/grad_norm_pre_clip': 0.08431190252304077, 'timing/data': 0.00032000998908188194, 'timing/model': 0.3649017760035349, 'learning_rate/qwen_vl_interface': 5.833245285359887e-06, 'learning_rate/action_model': 5.833245285359886e-05, 'epoch': 0.17}) +[2026-10-03 15:04:49,598][starVLA.training.train_starvla][INFO] - Step 2359, Loss: {'train/loss': 0.0007181093096733093, 'train/grad_norm_pre_clip': 0.3069220781326294, 'timing/data': 0.00028842000756412745, 'timing/model': 0.3657299670012435, 'learning_rate/qwen_vl_interface': 5.830222789361268e-06, 'learning_rate/action_model': 5.8302227893612684e-05, 'epoch': 0.17}) +[2026-10-03 15:04:50,453][starVLA.training.train_starvla][INFO] - Step 2360, Loss: {'train/loss': 1.9869505194947124e-05, 'train/grad_norm_pre_clip': 0.004405766259878874, 'timing/data': 0.00024624000070616603, 'timing/model': 0.36528793499746826, 'learning_rate/qwen_vl_interface': 5.8272000548549776e-06, 'learning_rate/action_model': 5.827200054854978e-05, 'epoch': 0.17}) +[2026-10-03 15:04:51,102][starVLA.training.train_starvla][INFO] - Step 2361, Loss: {'train/loss': 4.4579697714652866e-05, 'train/grad_norm_pre_clip': 0.01839526742696762, 'timing/data': 0.0003158300096401945, 'timing/model': 0.3650242059957236, 'learning_rate/qwen_vl_interface': 5.824177083083547e-06, 'learning_rate/action_model': 5.8241770830835475e-05, 'epoch': 0.17}) +[2026-10-03 15:04:51,751][starVLA.training.train_starvla][INFO] - Step 2362, Loss: {'train/loss': 0.1609642207622528, 'train/grad_norm_pre_clip': 95.27490234375, 'timing/data': 0.00021411999478004873, 'timing/model': 0.3658186569955433, 'learning_rate/qwen_vl_interface': 5.8211538752896056e-06, 'learning_rate/action_model': 5.821153875289606e-05, 'epoch': 0.17}) +[2026-10-03 15:04:52,401][starVLA.training.train_starvla][INFO] - Step 2363, Loss: {'train/loss': 0.0018467857735231519, 'train/grad_norm_pre_clip': 2.0852177143096924, 'timing/data': 0.00030760999652557075, 'timing/model': 0.3652003960014554, 'learning_rate/qwen_vl_interface': 5.8181304327158814e-06, 'learning_rate/action_model': 5.8181304327158814e-05, 'epoch': 0.17}) +[2026-10-03 15:04:53,050][starVLA.training.train_starvla][INFO] - Step 2364, Loss: {'train/loss': 0.3271504044532776, 'train/grad_norm_pre_clip': 21.45872688293457, 'timing/data': 0.00019890999828930944, 'timing/model': 0.36503002600511536, 'learning_rate/qwen_vl_interface': 5.8151067566051965e-06, 'learning_rate/action_model': 5.815106756605196e-05, 'epoch': 0.17}) +[2026-10-03 15:04:53,895][starVLA.training.train_starvla][INFO] - Step 2365, Loss: {'train/loss': 0.5042235851287842, 'train/grad_norm_pre_clip': 512.5599975585938, 'timing/data': 0.00023957999655976892, 'timing/model': 0.5625689980079187, 'learning_rate/qwen_vl_interface': 5.81208284820047e-06, 'learning_rate/action_model': 5.81208284820047e-05, 'epoch': 0.17}) +[2026-10-03 15:04:54,544][starVLA.training.train_starvla][INFO] - Step 2366, Loss: {'train/loss': 5.684049392584711e-05, 'train/grad_norm_pre_clip': 0.04384702071547508, 'timing/data': 0.00032623999868519604, 'timing/model': 0.3655490460077999, 'learning_rate/qwen_vl_interface': 5.809058708744716e-06, 'learning_rate/action_model': 5.809058708744716e-05, 'epoch': 0.17}) +[2026-10-03 15:04:55,194][starVLA.training.train_starvla][INFO] - Step 2367, Loss: {'train/loss': 0.0005858835647813976, 'train/grad_norm_pre_clip': 0.23384074866771698, 'timing/data': 0.00020018999930471182, 'timing/model': 0.3649270159949083, 'learning_rate/qwen_vl_interface': 5.8060343394810445e-06, 'learning_rate/action_model': 5.806034339481045e-05, 'epoch': 0.17}) +[2026-10-03 15:04:55,842][starVLA.training.train_starvla][INFO] - Step 2368, Loss: {'train/loss': 0.0013795809354633093, 'train/grad_norm_pre_clip': 0.5216519832611084, 'timing/data': 0.00022378998983185738, 'timing/model': 0.36527050599397626, 'learning_rate/qwen_vl_interface': 5.803009741652659e-06, 'learning_rate/action_model': 5.8030097416526585e-05, 'epoch': 0.17}) +[2026-10-03 15:04:56,490][starVLA.training.train_starvla][INFO] - Step 2369, Loss: {'train/loss': 2.161984593840316e-05, 'train/grad_norm_pre_clip': 0.007173551712185144, 'timing/data': 0.00019334000535309315, 'timing/model': 0.36467215699667577, 'learning_rate/qwen_vl_interface': 5.799984916502859e-06, 'learning_rate/action_model': 5.799984916502859e-05, 'epoch': 0.17}) +[2026-10-03 15:04:57,333][starVLA.training.train_starvla][INFO] - Step 2370, Loss: {'train/loss': 0.0019116371404379606, 'train/grad_norm_pre_clip': 0.5303551554679871, 'timing/data': 0.00038504000985994935, 'timing/model': 0.5593618379934924, 'learning_rate/qwen_vl_interface': 5.7969598652750334e-06, 'learning_rate/action_model': 5.796959865275033e-05, 'epoch': 0.17}) +[2026-10-03 15:04:57,984][starVLA.training.train_starvla][INFO] - Step 2371, Loss: {'train/loss': 0.21301966905593872, 'train/grad_norm_pre_clip': 64.19454193115234, 'timing/data': 0.00019797000277321786, 'timing/model': 0.3655477269930998, 'learning_rate/qwen_vl_interface': 5.793934589212666e-06, 'learning_rate/action_model': 5.793934589212666e-05, 'epoch': 0.17}) +[2026-10-03 15:04:58,631][starVLA.training.train_starvla][INFO] - Step 2372, Loss: {'train/loss': 0.9459322690963745, 'train/grad_norm_pre_clip': 43.32292175292969, 'timing/data': 0.0002368400018895045, 'timing/model': 0.3647127959993668, 'learning_rate/qwen_vl_interface': 5.7909090895593375e-06, 'learning_rate/action_model': 5.790909089559338e-05, 'epoch': 0.17}) +[2026-10-03 15:04:59,279][starVLA.training.train_starvla][INFO] - Step 2373, Loss: {'train/loss': 0.2142064869403839, 'train/grad_norm_pre_clip': 149.3445281982422, 'timing/data': 0.0003055400011362508, 'timing/model': 0.36476732599840034, 'learning_rate/qwen_vl_interface': 5.7878833675587135e-06, 'learning_rate/action_model': 5.787883367558713e-05, 'epoch': 0.17}) +[2026-10-03 15:04:59,928][starVLA.training.train_starvla][INFO] - Step 2374, Loss: {'train/loss': 0.17756235599517822, 'train/grad_norm_pre_clip': 121.57386779785156, 'timing/data': 0.0003115299914497882, 'timing/model': 0.36504298698855564, 'learning_rate/qwen_vl_interface': 5.784857424454556e-06, 'learning_rate/action_model': 5.7848574244545563e-05, 'epoch': 0.17}) +[2026-10-03 15:05:00,579][starVLA.training.train_starvla][INFO] - Step 2375, Loss: {'train/loss': 1.9072616851190105e-05, 'train/grad_norm_pre_clip': 0.007545695640146732, 'timing/data': 0.00022714000078849494, 'timing/model': 0.3652483859914355, 'learning_rate/qwen_vl_interface': 5.7818312614907125e-06, 'learning_rate/action_model': 5.781831261490712e-05, 'epoch': 0.17}) +[2026-10-03 15:05:01,425][starVLA.training.train_starvla][INFO] - Step 2376, Loss: {'train/loss': 0.0003770571784116328, 'train/grad_norm_pre_clip': 0.23823630809783936, 'timing/data': 0.00032719000591896474, 'timing/model': 0.36491536600806285, 'learning_rate/qwen_vl_interface': 5.778804879911129e-06, 'learning_rate/action_model': 5.778804879911128e-05, 'epoch': 0.17}) +[2026-10-03 15:05:02,073][starVLA.training.train_starvla][INFO] - Step 2377, Loss: {'train/loss': 0.00015640771016478539, 'train/grad_norm_pre_clip': 0.044880304485559464, 'timing/data': 0.00029310899844858795, 'timing/model': 0.3641272469976684, 'learning_rate/qwen_vl_interface': 5.775778280959832e-06, 'learning_rate/action_model': 5.7757782809598315e-05, 'epoch': 0.17}) +[2026-10-03 15:05:02,720][starVLA.training.train_starvla][INFO] - Step 2378, Loss: {'train/loss': 1.711325785436202e-05, 'train/grad_norm_pre_clip': 0.009760024026036263, 'timing/data': 0.0003297689981991425, 'timing/model': 0.3643016069981968, 'learning_rate/qwen_vl_interface': 5.7727514658809435e-06, 'learning_rate/action_model': 5.772751465880944e-05, 'epoch': 0.17}) +[2026-10-03 15:05:03,371][starVLA.training.train_starvla][INFO] - Step 2379, Loss: {'train/loss': 0.0004841409681830555, 'train/grad_norm_pre_clip': 0.1448894739151001, 'timing/data': 0.00019833000260405242, 'timing/model': 0.3650571670004865, 'learning_rate/qwen_vl_interface': 5.769724435918676e-06, 'learning_rate/action_model': 5.7697244359186755e-05, 'epoch': 0.17}) +[2026-10-03 15:05:04,019][starVLA.training.train_starvla][INFO] - Step 2380, Loss: {'train/loss': 7.85162192187272e-05, 'train/grad_norm_pre_clip': 0.023911310359835625, 'timing/data': 0.00019196000357624143, 'timing/model': 0.3652806369937025, 'learning_rate/qwen_vl_interface': 5.766697192317323e-06, 'learning_rate/action_model': 5.766697192317323e-05, 'epoch': 0.17}) +[2026-10-03 15:05:04,861][starVLA.training.train_starvla][INFO] - Step 2381, Loss: {'train/loss': 4.31037406087853e-05, 'train/grad_norm_pre_clip': 0.013767036609351635, 'timing/data': 0.00026881998928729445, 'timing/model': 0.364227607002249, 'learning_rate/qwen_vl_interface': 5.763669736321271e-06, 'learning_rate/action_model': 5.763669736321271e-05, 'epoch': 0.17}) +[2026-10-03 15:05:05,511][starVLA.training.train_starvla][INFO] - Step 2382, Loss: {'train/loss': 8.649672963656485e-05, 'train/grad_norm_pre_clip': 0.029965151101350784, 'timing/data': 0.000291279997327365, 'timing/model': 0.3660144759924151, 'learning_rate/qwen_vl_interface': 5.760642069174994e-06, 'learning_rate/action_model': 5.760642069174994e-05, 'epoch': 0.17}) +[2026-10-03 15:05:06,164][starVLA.training.train_starvla][INFO] - Step 2383, Loss: {'train/loss': 9.142784983851016e-05, 'train/grad_norm_pre_clip': 0.038890291005373, 'timing/data': 0.00019967999833170325, 'timing/model': 0.3665138069918612, 'learning_rate/qwen_vl_interface': 5.757614192123051e-06, 'learning_rate/action_model': 5.757614192123051e-05, 'epoch': 0.17}) +[2026-10-03 15:05:06,815][starVLA.training.train_starvla][INFO] - Step 2384, Loss: {'train/loss': 0.0009888792410492897, 'train/grad_norm_pre_clip': 0.7743114233016968, 'timing/data': 0.00024638000468257815, 'timing/model': 0.36614139600715134, 'learning_rate/qwen_vl_interface': 5.754586106410089e-06, 'learning_rate/action_model': 5.754586106410088e-05, 'epoch': 0.17}) +[2026-10-03 15:05:07,465][starVLA.training.train_starvla][INFO] - Step 2385, Loss: {'train/loss': 0.8242477774620056, 'train/grad_norm_pre_clip': 41.780609130859375, 'timing/data': 0.0002253199927508831, 'timing/model': 0.3659839260071749, 'learning_rate/qwen_vl_interface': 5.751557813280838e-06, 'learning_rate/action_model': 5.751557813280838e-05, 'epoch': 0.17}) +[2026-10-03 15:05:08,306][starVLA.training.train_starvla][INFO] - Step 2386, Loss: {'train/loss': 0.0003632883890531957, 'train/grad_norm_pre_clip': 0.3404616415500641, 'timing/data': 0.00023604999296367168, 'timing/model': 0.5570766180026112, 'learning_rate/qwen_vl_interface': 5.7485293139801146e-06, 'learning_rate/action_model': 5.7485293139801146e-05, 'epoch': 0.17}) +[2026-10-03 15:05:08,936][starVLA.training.train_starvla][INFO] - Step 2387, Loss: {'train/loss': 8.888273441698402e-06, 'train/grad_norm_pre_clip': 0.0051765888929367065, 'timing/data': 0.0003105200012214482, 'timing/model': 0.3661715360067319, 'learning_rate/qwen_vl_interface': 5.745500609752821e-06, 'learning_rate/action_model': 5.745500609752821e-05, 'epoch': 0.17}) +[2026-10-03 15:05:09,588][starVLA.training.train_starvla][INFO] - Step 2388, Loss: {'train/loss': 4.38090501120314e-06, 'train/grad_norm_pre_clip': 0.0008802280644886196, 'timing/data': 0.0002400499943178147, 'timing/model': 0.3662122559908312, 'learning_rate/qwen_vl_interface': 5.742471701843943e-06, 'learning_rate/action_model': 5.742471701843943e-05, 'epoch': 0.17}) +[2026-10-03 15:05:10,243][starVLA.training.train_starvla][INFO] - Step 2389, Loss: {'train/loss': 0.4091346263885498, 'train/grad_norm_pre_clip': 52.47050476074219, 'timing/data': 0.0002011299948208034, 'timing/model': 0.3661694360052934, 'learning_rate/qwen_vl_interface': 5.739442591498553e-06, 'learning_rate/action_model': 5.739442591498553e-05, 'epoch': 0.17}) +[2026-10-03 15:05:10,893][starVLA.training.train_starvla][INFO] - Step 2390, Loss: {'train/loss': 2.721527562243864e-05, 'train/grad_norm_pre_clip': 0.005698010325431824, 'timing/data': 0.00032068000291474164, 'timing/model': 0.365696957000182, 'learning_rate/qwen_vl_interface': 5.7364132799617985e-06, 'learning_rate/action_model': 5.7364132799617984e-05, 'epoch': 0.17}) +[2026-10-03 15:05:11,693][starVLA.training.train_starvla][INFO] - Step 2391, Loss: {'train/loss': 0.3583217263221741, 'train/grad_norm_pre_clip': 75.81462097167969, 'timing/data': 0.00020016999042127281, 'timing/model': 0.5357394520106027, 'learning_rate/qwen_vl_interface': 5.733383768478923e-06, 'learning_rate/action_model': 5.7333837684789225e-05, 'epoch': 0.18}) +[2026-10-03 15:05:12,343][starVLA.training.train_starvla][INFO] - Step 2392, Loss: {'train/loss': 0.00012687226990237832, 'train/grad_norm_pre_clip': 0.14209693670272827, 'timing/data': 0.00029594999796245247, 'timing/model': 0.365652726002736, 'learning_rate/qwen_vl_interface': 5.730354058295234e-06, 'learning_rate/action_model': 5.730354058295234e-05, 'epoch': 0.18}) +[2026-10-03 15:05:12,994][starVLA.training.train_starvla][INFO] - Step 2393, Loss: {'train/loss': 8.624198380857706e-05, 'train/grad_norm_pre_clip': 0.04001488909125328, 'timing/data': 0.0002875500067602843, 'timing/model': 0.365843566993135, 'learning_rate/qwen_vl_interface': 5.72732415065614e-06, 'learning_rate/action_model': 5.727324150656139e-05, 'epoch': 0.18}) +[2026-10-03 15:05:13,644][starVLA.training.train_starvla][INFO] - Step 2394, Loss: {'train/loss': 0.402458518743515, 'train/grad_norm_pre_clip': 20.431169509887695, 'timing/data': 0.0002907799935201183, 'timing/model': 0.36618055700091645, 'learning_rate/qwen_vl_interface': 5.724294046807115e-06, 'learning_rate/action_model': 5.724294046807115e-05, 'epoch': 0.18}) +[2026-10-03 15:05:14,297][starVLA.training.train_starvla][INFO] - Step 2395, Loss: {'train/loss': 0.0006768872262910008, 'train/grad_norm_pre_clip': 0.5898182392120361, 'timing/data': 0.000374399998690933, 'timing/model': 0.3658727860020008, 'learning_rate/qwen_vl_interface': 5.7212637479937234e-06, 'learning_rate/action_model': 5.721263747993724e-05, 'epoch': 0.18}) +[2026-10-03 15:05:14,947][starVLA.training.train_starvla][INFO] - Step 2396, Loss: {'train/loss': 0.00027251793653704226, 'train/grad_norm_pre_clip': 0.18753187358379364, 'timing/data': 0.0001935600012075156, 'timing/model': 0.36659046600107104, 'learning_rate/qwen_vl_interface': 5.718233255461607e-06, 'learning_rate/action_model': 5.718233255461607e-05, 'epoch': 0.18}) +[2026-10-03 15:05:15,771][starVLA.training.train_starvla][INFO] - Step 2397, Loss: {'train/loss': 2.4086924895527773e-05, 'train/grad_norm_pre_clip': 0.006764256861060858, 'timing/data': 0.00031899000168778, 'timing/model': 0.3430215200060047, 'learning_rate/qwen_vl_interface': 5.715202570456486e-06, 'learning_rate/action_model': 5.715202570456486e-05, 'epoch': 0.18}) +[2026-10-03 15:05:16,420][starVLA.training.train_starvla][INFO] - Step 2398, Loss: {'train/loss': 1.0861748456954956, 'train/grad_norm_pre_clip': 52.6805305480957, 'timing/data': 0.00031817999843042344, 'timing/model': 0.365393056999892, 'learning_rate/qwen_vl_interface': 5.7121716942241615e-06, 'learning_rate/action_model': 5.712171694224161e-05, 'epoch': 0.18}) +[2026-10-03 15:05:17,074][starVLA.training.train_starvla][INFO] - Step 2399, Loss: {'train/loss': 0.0003770244074985385, 'train/grad_norm_pre_clip': 0.2344011664390564, 'timing/data': 0.00022940999770071357, 'timing/model': 0.36750999599462375, 'learning_rate/qwen_vl_interface': 5.709140628010509e-06, 'learning_rate/action_model': 5.7091406280105085e-05, 'epoch': 0.18}) +[2026-10-03 15:05:17,725][starVLA.training.train_starvla][INFO] - Step 2400, Loss: {'train/loss': 0.002947887871414423, 'train/grad_norm_pre_clip': 2.4172441959381104, 'timing/data': 0.00030555999546777457, 'timing/model': 0.3662864770012675, 'learning_rate/qwen_vl_interface': 5.706109373061489e-06, 'learning_rate/action_model': 5.706109373061489e-05, 'epoch': 0.18}) +[2026-10-03 15:05:18,375][starVLA.training.train_starvla][INFO] - Step 2401, Loss: {'train/loss': 1.7024231055984274e-05, 'train/grad_norm_pre_clip': 0.004317350685596466, 'timing/data': 0.00019468000391498208, 'timing/model': 0.366740657002083, 'learning_rate/qwen_vl_interface': 5.703077930623134e-06, 'learning_rate/action_model': 5.7030779306231344e-05, 'epoch': 0.18}) +[2026-10-03 15:05:19,221][starVLA.training.train_starvla][INFO] - Step 2402, Loss: {'train/loss': 0.2032291144132614, 'train/grad_norm_pre_clip': 13.71624755859375, 'timing/data': 0.00030932000663597137, 'timing/model': 0.36564382699725684, 'learning_rate/qwen_vl_interface': 5.7000463019415565e-06, 'learning_rate/action_model': 5.7000463019415563e-05, 'epoch': 0.18}) +[2026-10-03 15:05:19,874][starVLA.training.train_starvla][INFO] - Step 2403, Loss: {'train/loss': 3.2967160223051906e-05, 'train/grad_norm_pre_clip': 0.006622062996029854, 'timing/data': 0.00024762999964877963, 'timing/model': 0.36622612600331195, 'learning_rate/qwen_vl_interface': 5.697014488262945e-06, 'learning_rate/action_model': 5.697014488262945e-05, 'epoch': 0.18}) +[2026-10-03 15:05:20,525][starVLA.training.train_starvla][INFO] - Step 2404, Loss: {'train/loss': 0.00300047779455781, 'train/grad_norm_pre_clip': 3.7666521072387695, 'timing/data': 0.00028592000307980925, 'timing/model': 0.36611635600274894, 'learning_rate/qwen_vl_interface': 5.693982490833561e-06, 'learning_rate/action_model': 5.693982490833561e-05, 'epoch': 0.18}) +[2026-10-03 15:05:21,152][starVLA.training.train_starvla][INFO] - Step 2405, Loss: {'train/loss': 0.0011868332512676716, 'train/grad_norm_pre_clip': 2.1238715648651123, 'timing/data': 0.00019181000243406743, 'timing/model': 0.3648463860008633, 'learning_rate/qwen_vl_interface': 5.690950310899745e-06, 'learning_rate/action_model': 5.6909503108997456e-05, 'epoch': 0.18}) +[2026-10-03 15:05:21,800][starVLA.training.train_starvla][INFO] - Step 2406, Loss: {'train/loss': 0.0004916946054436266, 'train/grad_norm_pre_clip': 0.7061918377876282, 'timing/data': 0.0002197800058638677, 'timing/model': 0.3655504869966535, 'learning_rate/qwen_vl_interface': 5.687917949707913e-06, 'learning_rate/action_model': 5.687917949707913e-05, 'epoch': 0.18}) +[2026-10-03 15:05:22,625][starVLA.training.train_starvla][INFO] - Step 2407, Loss: {'train/loss': 0.3278493285179138, 'train/grad_norm_pre_clip': 16.044153213500977, 'timing/data': 0.000268320000031963, 'timing/model': 0.5618565680051688, 'learning_rate/qwen_vl_interface': 5.684885408504555e-06, 'learning_rate/action_model': 5.684885408504554e-05, 'epoch': 0.18}) +[2026-10-03 15:05:23,271][starVLA.training.train_starvla][INFO] - Step 2408, Loss: {'train/loss': 0.00013310795475263149, 'train/grad_norm_pre_clip': 0.05404062569141388, 'timing/data': 0.00035193999065086246, 'timing/model': 0.36387416601064615, 'learning_rate/qwen_vl_interface': 5.681852688536231e-06, 'learning_rate/action_model': 5.681852688536231e-05, 'epoch': 0.18}) +[2026-10-03 15:05:23,896][starVLA.training.train_starvla][INFO] - Step 2409, Loss: {'train/loss': 0.23330117762088776, 'train/grad_norm_pre_clip': 103.39497375488281, 'timing/data': 0.00021063999156467617, 'timing/model': 0.34200283999962267, 'learning_rate/qwen_vl_interface': 5.67881979104958e-06, 'learning_rate/action_model': 5.67881979104958e-05, 'epoch': 0.18}) +[2026-10-03 15:05:24,542][starVLA.training.train_starvla][INFO] - Step 2410, Loss: {'train/loss': 5.3020867198938504e-05, 'train/grad_norm_pre_clip': 0.017656052485108376, 'timing/data': 0.00024480000138282776, 'timing/model': 0.36399129599158186, 'learning_rate/qwen_vl_interface': 5.67578671729131e-06, 'learning_rate/action_model': 5.675786717291309e-05, 'epoch': 0.18}) +[2026-10-03 15:05:25,191][starVLA.training.train_starvla][INFO] - Step 2411, Loss: {'train/loss': 4.201976116746664e-05, 'train/grad_norm_pre_clip': 0.007718953769654036, 'timing/data': 0.00028854000265710056, 'timing/model': 0.36430119701253716, 'learning_rate/qwen_vl_interface': 5.672753468508204e-06, 'learning_rate/action_model': 5.672753468508204e-05, 'epoch': 0.18}) +[2026-10-03 15:05:25,839][starVLA.training.train_starvla][INFO] - Step 2412, Loss: {'train/loss': 0.3845161497592926, 'train/grad_norm_pre_clip': 59.30312728881836, 'timing/data': 0.00022464999346993864, 'timing/model': 0.3642948970082216, 'learning_rate/qwen_vl_interface': 5.6697200459471165e-06, 'learning_rate/action_model': 5.669720045947117e-05, 'epoch': 0.18}) +[2026-10-03 15:05:26,679][starVLA.training.train_starvla][INFO] - Step 2413, Loss: {'train/loss': 4.240660200593993e-05, 'train/grad_norm_pre_clip': 0.009348385967314243, 'timing/data': 0.000334889002260752, 'timing/model': 0.3638748170051258, 'learning_rate/qwen_vl_interface': 5.666686450854973e-06, 'learning_rate/action_model': 5.666686450854973e-05, 'epoch': 0.18}) +[2026-10-03 15:05:27,324][starVLA.training.train_starvla][INFO] - Step 2414, Loss: {'train/loss': 0.0002712639980018139, 'train/grad_norm_pre_clip': 0.15428371727466583, 'timing/data': 0.0003138899919576943, 'timing/model': 0.3634185470000375, 'learning_rate/qwen_vl_interface': 5.663652684478768e-06, 'learning_rate/action_model': 5.6636526844787676e-05, 'epoch': 0.18}) +[2026-10-03 15:05:27,974][starVLA.training.train_starvla][INFO] - Step 2415, Loss: {'train/loss': 0.0008575120009481907, 'train/grad_norm_pre_clip': 0.8669489622116089, 'timing/data': 0.00028443998598959297, 'timing/model': 0.3644064870022703, 'learning_rate/qwen_vl_interface': 5.660618748065569e-06, 'learning_rate/action_model': 5.6606187480655695e-05, 'epoch': 0.18}) +[2026-10-03 15:05:28,621][starVLA.training.train_starvla][INFO] - Step 2416, Loss: {'train/loss': 0.09912103414535522, 'train/grad_norm_pre_clip': 152.219482421875, 'timing/data': 0.00024279000353999436, 'timing/model': 0.3641116479993798, 'learning_rate/qwen_vl_interface': 5.657584642862514e-06, 'learning_rate/action_model': 5.657584642862513e-05, 'epoch': 0.18}) +[2026-10-03 15:05:29,270][starVLA.training.train_starvla][INFO] - Step 2417, Loss: {'train/loss': 7.83629366196692e-05, 'train/grad_norm_pre_clip': 0.02413237653672695, 'timing/data': 0.0003094400017289445, 'timing/model': 0.36493247699399944, 'learning_rate/qwen_vl_interface': 5.654550370116808e-06, 'learning_rate/action_model': 5.654550370116808e-05, 'epoch': 0.18}) +[2026-10-03 15:05:30,110][starVLA.training.train_starvla][INFO] - Step 2418, Loss: {'train/loss': 0.00014385790564119816, 'train/grad_norm_pre_clip': 0.04965498670935631, 'timing/data': 0.0002118199918186292, 'timing/model': 0.5578097090037772, 'learning_rate/qwen_vl_interface': 5.651515931075726e-06, 'learning_rate/action_model': 5.651515931075726e-05, 'epoch': 0.18}) +[2026-10-03 15:05:30,761][starVLA.training.train_starvla][INFO] - Step 2419, Loss: {'train/loss': 0.00011376102338545024, 'train/grad_norm_pre_clip': 0.05722212791442871, 'timing/data': 0.0002875299978768453, 'timing/model': 0.3642739770002663, 'learning_rate/qwen_vl_interface': 5.6484813269866126e-06, 'learning_rate/action_model': 5.648481326986612e-05, 'epoch': 0.18}) +[2026-10-03 15:05:31,408][starVLA.training.train_starvla][INFO] - Step 2420, Loss: {'train/loss': 0.000102511745353695, 'train/grad_norm_pre_clip': 0.03643125668168068, 'timing/data': 0.00023716999567113817, 'timing/model': 0.36404881700582337, 'learning_rate/qwen_vl_interface': 5.645446559096875e-06, 'learning_rate/action_model': 5.645446559096875e-05, 'epoch': 0.18}) +[2026-10-03 15:05:32,054][starVLA.training.train_starvla][INFO] - Step 2421, Loss: {'train/loss': 4.516874469118193e-05, 'train/grad_norm_pre_clip': 0.01094140112400055, 'timing/data': 0.0002306600072188303, 'timing/model': 0.3637779970013071, 'learning_rate/qwen_vl_interface': 5.6424116286539965e-06, 'learning_rate/action_model': 5.642411628653996e-05, 'epoch': 0.18}) +[2026-10-03 15:05:32,700][starVLA.training.train_starvla][INFO] - Step 2422, Loss: {'train/loss': 0.21667587757110596, 'train/grad_norm_pre_clip': 92.33594512939453, 'timing/data': 0.00021757000649813563, 'timing/model': 0.363873787006014, 'learning_rate/qwen_vl_interface': 5.639376536905517e-06, 'learning_rate/action_model': 5.6393765369055175e-05, 'epoch': 0.18}) +[2026-10-03 15:05:33,542][starVLA.training.train_starvla][INFO] - Step 2423, Loss: {'train/loss': 0.06038226932287216, 'train/grad_norm_pre_clip': 28.314678192138672, 'timing/data': 0.00025363999884575605, 'timing/model': 0.5553453589964192, 'learning_rate/qwen_vl_interface': 5.636341285099055e-06, 'learning_rate/action_model': 5.636341285099055e-05, 'epoch': 0.18}) +[2026-10-03 15:05:34,193][starVLA.training.train_starvla][INFO] - Step 2424, Loss: {'train/loss': 0.00020442788081709296, 'train/grad_norm_pre_clip': 0.10572504997253418, 'timing/data': 0.0002292099961778149, 'timing/model': 0.36583394699846394, 'learning_rate/qwen_vl_interface': 5.6333058744822835e-06, 'learning_rate/action_model': 5.633305874482283e-05, 'epoch': 0.18}) +[2026-10-03 15:05:34,843][starVLA.training.train_starvla][INFO] - Step 2425, Loss: {'train/loss': 0.003044044366106391, 'train/grad_norm_pre_clip': 0.9590603113174438, 'timing/data': 0.00020260999735910445, 'timing/model': 0.3651365169935161, 'learning_rate/qwen_vl_interface': 5.630270306302944e-06, 'learning_rate/action_model': 5.6302703063029436e-05, 'epoch': 0.18}) +[2026-10-03 15:05:35,492][starVLA.training.train_starvla][INFO] - Step 2426, Loss: {'train/loss': 0.0018636975437402725, 'train/grad_norm_pre_clip': 0.8316870331764221, 'timing/data': 0.00029085000278428197, 'timing/model': 0.36607894700136967, 'learning_rate/qwen_vl_interface': 5.627234581808848e-06, 'learning_rate/action_model': 5.627234581808848e-05, 'epoch': 0.18}) +[2026-10-03 15:05:36,124][starVLA.training.train_starvla][INFO] - Step 2427, Loss: {'train/loss': 3.812287104665302e-05, 'train/grad_norm_pre_clip': 0.007202330976724625, 'timing/data': 0.00019863000488840044, 'timing/model': 0.3442450799921062, 'learning_rate/qwen_vl_interface': 5.624198702247863e-06, 'learning_rate/action_model': 5.624198702247863e-05, 'epoch': 0.18}) +[2026-10-03 15:05:36,774][starVLA.training.train_starvla][INFO] - Step 2428, Loss: {'train/loss': 0.4264998733997345, 'train/grad_norm_pre_clip': 63.2632942199707, 'timing/data': 0.00024361000396311283, 'timing/model': 0.3651866270083701, 'learning_rate/qwen_vl_interface': 5.621162668867925e-06, 'learning_rate/action_model': 5.621162668867925e-05, 'epoch': 0.18}) +[2026-10-03 15:05:37,575][starVLA.training.train_starvla][INFO] - Step 2429, Loss: {'train/loss': 5.956751556368545e-05, 'train/grad_norm_pre_clip': 0.024566983804106712, 'timing/data': 0.0002391500020166859, 'timing/model': 0.3425722600077279, 'learning_rate/qwen_vl_interface': 5.618126482917037e-06, 'learning_rate/action_model': 5.618126482917036e-05, 'epoch': 0.18}) +[2026-10-03 15:05:38,224][starVLA.training.train_starvla][INFO] - Step 2430, Loss: {'train/loss': 7.775994890835136e-05, 'train/grad_norm_pre_clip': 0.014875396154820919, 'timing/data': 0.0003038989962078631, 'timing/model': 0.365248097004951, 'learning_rate/qwen_vl_interface': 5.6150901456432546e-06, 'learning_rate/action_model': 5.615090145643255e-05, 'epoch': 0.18}) +[2026-10-03 15:05:38,875][starVLA.training.train_starvla][INFO] - Step 2431, Loss: {'train/loss': 0.00014789853594265878, 'train/grad_norm_pre_clip': 0.12664377689361572, 'timing/data': 0.0003006089973496273, 'timing/model': 0.36540403800609056, 'learning_rate/qwen_vl_interface': 5.612053658294705e-06, 'learning_rate/action_model': 5.6120536582947045e-05, 'epoch': 0.18}) +[2026-10-03 15:05:39,504][starVLA.training.train_starvla][INFO] - Step 2432, Loss: {'train/loss': 0.07203741371631622, 'train/grad_norm_pre_clip': 7.143466949462891, 'timing/data': 0.0003476699930615723, 'timing/model': 0.36635462699632626, 'learning_rate/qwen_vl_interface': 5.609017022119572e-06, 'learning_rate/action_model': 5.609017022119573e-05, 'epoch': 0.18}) +[2026-10-03 15:05:40,155][starVLA.training.train_starvla][INFO] - Step 2433, Loss: {'train/loss': 0.0011884202249348164, 'train/grad_norm_pre_clip': 4.249905586242676, 'timing/data': 0.0002988499909406528, 'timing/model': 0.36548130700248294, 'learning_rate/qwen_vl_interface': 5.605980238366104e-06, 'learning_rate/action_model': 5.605980238366103e-05, 'epoch': 0.18}) +[2026-10-03 15:05:40,997][starVLA.training.train_starvla][INFO] - Step 2434, Loss: {'train/loss': 0.03546077013015747, 'train/grad_norm_pre_clip': 30.501638412475586, 'timing/data': 0.00024901999859139323, 'timing/model': 0.36507536700810306, 'learning_rate/qwen_vl_interface': 5.602943308282606e-06, 'learning_rate/action_model': 5.602943308282605e-05, 'epoch': 0.18}) +[2026-10-03 15:05:41,650][starVLA.training.train_starvla][INFO] - Step 2435, Loss: {'train/loss': 0.21297071874141693, 'train/grad_norm_pre_clip': 191.201416015625, 'timing/data': 0.0002654900017660111, 'timing/model': 0.36620493700320367, 'learning_rate/qwen_vl_interface': 5.599906233117445e-06, 'learning_rate/action_model': 5.5999062331174446e-05, 'epoch': 0.18}) +[2026-10-03 15:05:42,300][starVLA.training.train_starvla][INFO] - Step 2436, Loss: {'train/loss': 4.120666199014522e-05, 'train/grad_norm_pre_clip': 0.006919089704751968, 'timing/data': 0.00033423998684156686, 'timing/model': 0.36550825599988457, 'learning_rate/qwen_vl_interface': 5.596869014119048e-06, 'learning_rate/action_model': 5.596869014119048e-05, 'epoch': 0.18}) +[2026-10-03 15:05:42,950][starVLA.training.train_starvla][INFO] - Step 2437, Loss: {'train/loss': 7.838028250262141e-05, 'train/grad_norm_pre_clip': 0.019178083166480064, 'timing/data': 0.00030860000697430223, 'timing/model': 0.3654202459874796, 'learning_rate/qwen_vl_interface': 5.593831652535902e-06, 'learning_rate/action_model': 5.593831652535902e-05, 'epoch': 0.18}) +[2026-10-03 15:05:43,577][starVLA.training.train_starvla][INFO] - Step 2438, Loss: {'train/loss': 0.009212231263518333, 'train/grad_norm_pre_clip': 20.75285530090332, 'timing/data': 0.00029711000388488173, 'timing/model': 0.3433266799984267, 'learning_rate/qwen_vl_interface': 5.590794149616552e-06, 'learning_rate/action_model': 5.590794149616552e-05, 'epoch': 0.18}) +[2026-10-03 15:05:44,425][starVLA.training.train_starvla][INFO] - Step 2439, Loss: {'train/loss': 0.11159995943307877, 'train/grad_norm_pre_clip': 18.859142303466797, 'timing/data': 0.0003167099930578843, 'timing/model': 0.3657012269977713, 'learning_rate/qwen_vl_interface': 5.587756506609601e-06, 'learning_rate/action_model': 5.5877565066096006e-05, 'epoch': 0.18}) +[2026-10-03 15:05:45,052][starVLA.training.train_starvla][INFO] - Step 2440, Loss: {'train/loss': 3.380931957508437e-05, 'train/grad_norm_pre_clip': 0.007729765027761459, 'timing/data': 0.00020352000137791038, 'timing/model': 0.34269824999501, 'learning_rate/qwen_vl_interface': 5.584718724763707e-06, 'learning_rate/action_model': 5.584718724763707e-05, 'epoch': 0.18}) +[2026-10-03 15:05:45,702][starVLA.training.train_starvla][INFO] - Step 2441, Loss: {'train/loss': 0.001518616802059114, 'train/grad_norm_pre_clip': 2.078171968460083, 'timing/data': 0.00023598999541718513, 'timing/model': 0.3649585070088506, 'learning_rate/qwen_vl_interface': 5.581680805327588e-06, 'learning_rate/action_model': 5.5816808053275885e-05, 'epoch': 0.18}) +[2026-10-03 15:05:46,351][starVLA.training.train_starvla][INFO] - Step 2442, Loss: {'train/loss': 0.031101422384381294, 'train/grad_norm_pre_clip': 13.530348777770996, 'timing/data': 0.00027751000015996397, 'timing/model': 0.3656058269989444, 'learning_rate/qwen_vl_interface': 5.578642749550019e-06, 'learning_rate/action_model': 5.578642749550019e-05, 'epoch': 0.18}) +[2026-10-03 15:05:46,982][starVLA.training.train_starvla][INFO] - Step 2443, Loss: {'train/loss': 9.505565685685724e-05, 'train/grad_norm_pre_clip': 0.02891278639435768, 'timing/data': 0.00023474999761674553, 'timing/model': 0.3667895969992969, 'learning_rate/qwen_vl_interface': 5.575604558679831e-06, 'learning_rate/action_model': 5.57560455867983e-05, 'epoch': 0.18}) +[2026-10-03 15:05:47,826][starVLA.training.train_starvla][INFO] - Step 2444, Loss: {'train/loss': 0.09721839427947998, 'train/grad_norm_pre_clip': 78.12474822998047, 'timing/data': 0.00022319999698083848, 'timing/model': 0.5601495580049232, 'learning_rate/qwen_vl_interface': 5.5725662339659085e-06, 'learning_rate/action_model': 5.572566233965908e-05, 'epoch': 0.18}) +[2026-10-03 15:05:48,477][starVLA.training.train_starvla][INFO] - Step 2445, Loss: {'train/loss': 0.008000572212040424, 'train/grad_norm_pre_clip': 33.32403564453125, 'timing/data': 0.00029630000062752515, 'timing/model': 0.36571994599944446, 'learning_rate/qwen_vl_interface': 5.56952777665719e-06, 'learning_rate/action_model': 5.56952777665719e-05, 'epoch': 0.18}) +[2026-10-03 15:05:49,127][starVLA.training.train_starvla][INFO] - Step 2446, Loss: {'train/loss': 7.318711868720129e-05, 'train/grad_norm_pre_clip': 0.021347124129533768, 'timing/data': 0.00031877998844720423, 'timing/model': 0.36577251699054614, 'learning_rate/qwen_vl_interface': 5.566489188002671e-06, 'learning_rate/action_model': 5.56648918800267e-05, 'epoch': 0.18}) +[2026-10-03 15:05:49,780][starVLA.training.train_starvla][INFO] - Step 2447, Loss: {'train/loss': 0.00012754552881233394, 'train/grad_norm_pre_clip': 0.0636512041091919, 'timing/data': 0.00028947000100743026, 'timing/model': 0.3660883569973521, 'learning_rate/qwen_vl_interface': 5.563450469251403e-06, 'learning_rate/action_model': 5.563450469251402e-05, 'epoch': 0.18}) +[2026-10-03 15:05:50,429][starVLA.training.train_starvla][INFO] - Step 2448, Loss: {'train/loss': 5.08765151607804e-05, 'train/grad_norm_pre_clip': 0.021733565255999565, 'timing/data': 0.00023337001039180905, 'timing/model': 0.3656782170000952, 'learning_rate/qwen_vl_interface': 5.560411621652488e-06, 'learning_rate/action_model': 5.560411621652488e-05, 'epoch': 0.18}) +[2026-10-03 15:05:51,080][starVLA.training.train_starvla][INFO] - Step 2449, Loss: {'train/loss': 0.03529784083366394, 'train/grad_norm_pre_clip': 6.466352462768555, 'timing/data': 0.00019874999998137355, 'timing/model': 0.36653296698932536, 'learning_rate/qwen_vl_interface': 5.557372646455081e-06, 'learning_rate/action_model': 5.55737264645508e-05, 'epoch': 0.18}) +[2026-10-03 15:05:51,897][starVLA.training.train_starvla][INFO] - Step 2450, Loss: {'train/loss': 0.5545604825019836, 'train/grad_norm_pre_clip': 81.93463134765625, 'timing/data': 0.0003016400005435571, 'timing/model': 0.36448423699766863, 'learning_rate/qwen_vl_interface': 5.554333544908388e-06, 'learning_rate/action_model': 5.5543335449083875e-05, 'epoch': 0.18}) +[2026-10-03 15:05:52,550][starVLA.training.train_starvla][INFO] - Step 2451, Loss: {'train/loss': 0.786604642868042, 'train/grad_norm_pre_clip': 28.355756759643555, 'timing/data': 0.00027874999796040356, 'timing/model': 0.36583812700700946, 'learning_rate/qwen_vl_interface': 5.551294318261671e-06, 'learning_rate/action_model': 5.5512943182616705e-05, 'epoch': 0.18}) +[2026-10-03 15:05:53,199][starVLA.training.train_starvla][INFO] - Step 2452, Loss: {'train/loss': 0.019390111789107323, 'train/grad_norm_pre_clip': 26.93022918701172, 'timing/data': 0.00031802999728824943, 'timing/model': 0.36487582699919585, 'learning_rate/qwen_vl_interface': 5.548254967764239e-06, 'learning_rate/action_model': 5.5482549677642394e-05, 'epoch': 0.18}) +[2026-10-03 15:05:53,848][starVLA.training.train_starvla][INFO] - Step 2453, Loss: {'train/loss': 0.00019107552361674607, 'train/grad_norm_pre_clip': 0.05618683621287346, 'timing/data': 0.0002943699946627021, 'timing/model': 0.3652786559978267, 'learning_rate/qwen_vl_interface': 5.545215494665458e-06, 'learning_rate/action_model': 5.545215494665458e-05, 'epoch': 0.18}) +[2026-10-03 15:05:54,496][starVLA.training.train_starvla][INFO] - Step 2454, Loss: {'train/loss': 0.3618208169937134, 'train/grad_norm_pre_clip': 24.04341697692871, 'timing/data': 0.0002651899994816631, 'timing/model': 0.36452656700566877, 'learning_rate/qwen_vl_interface': 5.5421759002147376e-06, 'learning_rate/action_model': 5.542175900214738e-05, 'epoch': 0.18}) +[2026-10-03 15:05:55,335][starVLA.training.train_starvla][INFO] - Step 2455, Loss: {'train/loss': 6.366957677528262e-05, 'train/grad_norm_pre_clip': 0.012883175164461136, 'timing/data': 0.0003055499983020127, 'timing/model': 0.5522191390045919, 'learning_rate/qwen_vl_interface': 5.53913618566154e-06, 'learning_rate/action_model': 5.53913618566154e-05, 'epoch': 0.18}) +[2026-10-03 15:05:55,984][starVLA.training.train_starvla][INFO] - Step 2456, Loss: {'train/loss': 0.007107163313776255, 'train/grad_norm_pre_clip': 1.5968621969223022, 'timing/data': 0.00018668000120669603, 'timing/model': 0.3650525970006129, 'learning_rate/qwen_vl_interface': 5.536096352255381e-06, 'learning_rate/action_model': 5.536096352255381e-05, 'epoch': 0.18}) +[2026-10-03 15:05:56,634][starVLA.training.train_starvla][INFO] - Step 2457, Loss: {'train/loss': 0.00011651913519017398, 'train/grad_norm_pre_clip': 0.023701054975390434, 'timing/data': 0.00019629999587778002, 'timing/model': 0.3658028570062015, 'learning_rate/qwen_vl_interface': 5.533056401245817e-06, 'learning_rate/action_model': 5.5330564012458164e-05, 'epoch': 0.18}) +[2026-10-03 15:05:57,282][starVLA.training.train_starvla][INFO] - Step 2458, Loss: {'train/loss': 0.5088520050048828, 'train/grad_norm_pre_clip': 33.28902053833008, 'timing/data': 0.0002030100004049018, 'timing/model': 0.36481345701031387, 'learning_rate/qwen_vl_interface': 5.530016333882456e-06, 'learning_rate/action_model': 5.5300163338824565e-05, 'epoch': 0.18}) +[2026-10-03 15:05:57,935][starVLA.training.train_starvla][INFO] - Step 2459, Loss: {'train/loss': 0.0018995009595528245, 'train/grad_norm_pre_clip': 4.628472805023193, 'timing/data': 0.0002817700005834922, 'timing/model': 0.3654929769982118, 'learning_rate/qwen_vl_interface': 5.52697615141496e-06, 'learning_rate/action_model': 5.52697615141496e-05, 'epoch': 0.18}) +[2026-10-03 15:05:58,776][starVLA.training.train_starvla][INFO] - Step 2460, Loss: {'train/loss': 0.0004915676545351744, 'train/grad_norm_pre_clip': 0.33446836471557617, 'timing/data': 0.0003485400084173307, 'timing/model': 0.5568208490003599, 'learning_rate/qwen_vl_interface': 5.52393585509303e-06, 'learning_rate/action_model': 5.52393585509303e-05, 'epoch': 0.18}) +[2026-10-03 15:05:59,426][starVLA.training.train_starvla][INFO] - Step 2461, Loss: {'train/loss': 0.1516403704881668, 'train/grad_norm_pre_clip': 33.6043815612793, 'timing/data': 0.00030064000748097897, 'timing/model': 0.366081086001941, 'learning_rate/qwen_vl_interface': 5.520895446166418e-06, 'learning_rate/action_model': 5.5208954461664175e-05, 'epoch': 0.18}) +[2026-10-03 15:06:00,075][starVLA.training.train_starvla][INFO] - Step 2462, Loss: {'train/loss': 0.00012100282765459269, 'train/grad_norm_pre_clip': 0.05147120729088783, 'timing/data': 0.00024947999918367714, 'timing/model': 0.36470229699625634, 'learning_rate/qwen_vl_interface': 5.517854925884919e-06, 'learning_rate/action_model': 5.517854925884919e-05, 'epoch': 0.18}) +[2026-10-03 15:06:00,727][starVLA.training.train_starvla][INFO] - Step 2463, Loss: {'train/loss': 8.761562639847398e-05, 'train/grad_norm_pre_clip': 0.026390764862298965, 'timing/data': 0.00022899000032339245, 'timing/model': 0.36524943700351287, 'learning_rate/qwen_vl_interface': 5.514814295498379e-06, 'learning_rate/action_model': 5.514814295498378e-05, 'epoch': 0.18}) +[2026-10-03 15:06:01,376][starVLA.training.train_starvla][INFO] - Step 2464, Loss: {'train/loss': 0.4003117084503174, 'train/grad_norm_pre_clip': 32.8956298828125, 'timing/data': 0.00023912001051940024, 'timing/model': 0.36548042700451333, 'learning_rate/qwen_vl_interface': 5.5117735562566846e-06, 'learning_rate/action_model': 5.5117735562566844e-05, 'epoch': 0.18}) +[2026-10-03 15:06:02,026][starVLA.training.train_starvla][INFO] - Step 2465, Loss: {'train/loss': 0.016956256702542305, 'train/grad_norm_pre_clip': 25.28325653076172, 'timing/data': 0.00022595000336878002, 'timing/model': 0.3656088170100702, 'learning_rate/qwen_vl_interface': 5.5087327094097685e-06, 'learning_rate/action_model': 5.508732709409769e-05, 'epoch': 0.18}) +[2026-10-03 15:06:02,866][starVLA.training.train_starvla][INFO] - Step 2466, Loss: {'train/loss': 0.002711344975978136, 'train/grad_norm_pre_clip': 0.760971188545227, 'timing/data': 0.00024020900309551507, 'timing/model': 0.36501200699422043, 'learning_rate/qwen_vl_interface': 5.505691756207609e-06, 'learning_rate/action_model': 5.505691756207609e-05, 'epoch': 0.18}) +[2026-10-03 15:06:03,518][starVLA.training.train_starvla][INFO] - Step 2467, Loss: {'train/loss': 0.001183818676508963, 'train/grad_norm_pre_clip': 0.2634611427783966, 'timing/data': 0.00034439000592101365, 'timing/model': 0.36562331700406503, 'learning_rate/qwen_vl_interface': 5.502650697900228e-06, 'learning_rate/action_model': 5.502650697900228e-05, 'epoch': 0.18}) +[2026-10-03 15:06:04,167][starVLA.training.train_starvla][INFO] - Step 2468, Loss: {'train/loss': 0.00012487260391935706, 'train/grad_norm_pre_clip': 0.03771764785051346, 'timing/data': 0.00022765000176150352, 'timing/model': 0.36501510700327344, 'learning_rate/qwen_vl_interface': 5.499609535737688e-06, 'learning_rate/action_model': 5.4996095357376875e-05, 'epoch': 0.18}) +[2026-10-03 15:06:04,816][starVLA.training.train_starvla][INFO] - Step 2469, Loss: {'train/loss': 0.004093295894563198, 'train/grad_norm_pre_clip': 5.891446113586426, 'timing/data': 0.00030482000147458166, 'timing/model': 0.36486910699750297, 'learning_rate/qwen_vl_interface': 5.496568270970096e-06, 'learning_rate/action_model': 5.496568270970096e-05, 'epoch': 0.18}) +[2026-10-03 15:06:05,464][starVLA.training.train_starvla][INFO] - Step 2470, Loss: {'train/loss': 0.13633817434310913, 'train/grad_norm_pre_clip': 108.96575927734375, 'timing/data': 0.00021247999393381178, 'timing/model': 0.3650596469960874, 'learning_rate/qwen_vl_interface': 5.4935269048476015e-06, 'learning_rate/action_model': 5.493526904847601e-05, 'epoch': 0.18}) +[2026-10-03 15:06:06,312][starVLA.training.train_starvla][INFO] - Step 2471, Loss: {'train/loss': 0.00020427085109986365, 'train/grad_norm_pre_clip': 0.0658784732222557, 'timing/data': 0.0002022300031967461, 'timing/model': 0.3656682269938756, 'learning_rate/qwen_vl_interface': 5.490485438620397e-06, 'learning_rate/action_model': 5.4904854386203965e-05, 'epoch': 0.18}) +[2026-10-03 15:06:06,962][starVLA.training.train_starvla][INFO] - Step 2472, Loss: {'train/loss': 0.000294381141429767, 'train/grad_norm_pre_clip': 0.10518442094326019, 'timing/data': 0.00020122999558225274, 'timing/model': 0.36577765698893927, 'learning_rate/qwen_vl_interface': 5.487443873538712e-06, 'learning_rate/action_model': 5.4874438735387115e-05, 'epoch': 0.18}) +[2026-10-03 15:06:07,612][starVLA.training.train_starvla][INFO] - Step 2473, Loss: {'train/loss': 0.5317791700363159, 'train/grad_norm_pre_clip': 30.244384765625, 'timing/data': 0.00031786000181455165, 'timing/model': 0.365538877013023, 'learning_rate/qwen_vl_interface': 5.484402210852819e-06, 'learning_rate/action_model': 5.484402210852818e-05, 'epoch': 0.18}) +[2026-10-03 15:06:08,260][starVLA.training.train_starvla][INFO] - Step 2474, Loss: {'train/loss': 0.0003514538984745741, 'train/grad_norm_pre_clip': 0.22635865211486816, 'timing/data': 0.0002312199940206483, 'timing/model': 0.36469941699760966, 'learning_rate/qwen_vl_interface': 5.48136045181303e-06, 'learning_rate/action_model': 5.48136045181303e-05, 'epoch': 0.18}) +[2026-10-03 15:06:08,913][starVLA.training.train_starvla][INFO] - Step 2475, Loss: {'train/loss': 0.0002914803335443139, 'train/grad_norm_pre_clip': 0.1308029294013977, 'timing/data': 0.0002521299902582541, 'timing/model': 0.365582237005583, 'learning_rate/qwen_vl_interface': 5.478318597669699e-06, 'learning_rate/action_model': 5.4783185976696995e-05, 'epoch': 0.18}) +[2026-10-03 15:06:09,760][starVLA.training.train_starvla][INFO] - Step 2476, Loss: {'train/loss': 0.00015248889394570142, 'train/grad_norm_pre_clip': 0.043840110301971436, 'timing/data': 0.00036438999813981354, 'timing/model': 0.5625343580031767, 'learning_rate/qwen_vl_interface': 5.475276649673218e-06, 'learning_rate/action_model': 5.4752766496732174e-05, 'epoch': 0.18}) +[2026-10-03 15:06:10,408][starVLA.training.train_starvla][INFO] - Step 2477, Loss: {'train/loss': 0.00011028200970031321, 'train/grad_norm_pre_clip': 0.025216495618224144, 'timing/data': 0.0003143399953842163, 'timing/model': 0.36518086699652486, 'learning_rate/qwen_vl_interface': 5.472234609074014e-06, 'learning_rate/action_model': 5.472234609074014e-05, 'epoch': 0.18}) +[2026-10-03 15:06:11,058][starVLA.training.train_starvla][INFO] - Step 2478, Loss: {'train/loss': 0.003992870915681124, 'train/grad_norm_pre_clip': 4.545530796051025, 'timing/data': 0.0001946600095834583, 'timing/model': 0.3654178180004237, 'learning_rate/qwen_vl_interface': 5.46919247712256e-06, 'learning_rate/action_model': 5.46919247712256e-05, 'epoch': 0.18}) +[2026-10-03 15:06:11,710][starVLA.training.train_starvla][INFO] - Step 2479, Loss: {'train/loss': 0.03389422595500946, 'train/grad_norm_pre_clip': 83.83123779296875, 'timing/data': 0.00029242999153211713, 'timing/model': 0.36538865699549206, 'learning_rate/qwen_vl_interface': 5.466150255069356e-06, 'learning_rate/action_model': 5.4661502550693555e-05, 'epoch': 0.18}) +[2026-10-03 15:06:12,359][starVLA.training.train_starvla][INFO] - Step 2480, Loss: {'train/loss': 0.0004132186295464635, 'train/grad_norm_pre_clip': 0.13989365100860596, 'timing/data': 0.00026963000709656626, 'timing/model': 0.3653345070051728, 'learning_rate/qwen_vl_interface': 5.463107944164947e-06, 'learning_rate/action_model': 5.4631079441649466e-05, 'epoch': 0.18}) +[2026-10-03 15:06:13,008][starVLA.training.train_starvla][INFO] - Step 2481, Loss: {'train/loss': 0.165821373462677, 'train/grad_norm_pre_clip': 177.34654235839844, 'timing/data': 0.00028819999715778977, 'timing/model': 0.3653154669882497, 'learning_rate/qwen_vl_interface': 5.46006554565991e-06, 'learning_rate/action_model': 5.4600655456599105e-05, 'epoch': 0.18}) +[2026-10-03 15:06:13,849][starVLA.training.train_starvla][INFO] - Step 2482, Loss: {'train/loss': 0.0011127409525215626, 'train/grad_norm_pre_clip': 2.6963677406311035, 'timing/data': 0.0003634000022429973, 'timing/model': 0.36501416700775735, 'learning_rate/qwen_vl_interface': 5.457023060804865e-06, 'learning_rate/action_model': 5.457023060804864e-05, 'epoch': 0.18}) +[2026-10-03 15:06:14,506][starVLA.training.train_starvla][INFO] - Step 2483, Loss: {'train/loss': 0.0005870318273082376, 'train/grad_norm_pre_clip': 0.9914833307266235, 'timing/data': 0.00028531000134535134, 'timing/model': 0.3674639870005194, 'learning_rate/qwen_vl_interface': 5.453980490850459e-06, 'learning_rate/action_model': 5.453980490850459e-05, 'epoch': 0.18}) +[2026-10-03 15:06:15,132][starVLA.training.train_starvla][INFO] - Step 2484, Loss: {'train/loss': 0.00023362436331808567, 'train/grad_norm_pre_clip': 0.06208867207169533, 'timing/data': 0.000248910000664182, 'timing/model': 0.3427735899895197, 'learning_rate/qwen_vl_interface': 5.4509378370473785e-06, 'learning_rate/action_model': 5.450937837047378e-05, 'epoch': 0.18}) +[2026-10-03 15:06:15,781][starVLA.training.train_starvla][INFO] - Step 2485, Loss: {'train/loss': 0.0001305418845731765, 'train/grad_norm_pre_clip': 0.04826463758945465, 'timing/data': 0.00032808000105433166, 'timing/model': 0.3649470670061419, 'learning_rate/qwen_vl_interface': 5.4478951006463424e-06, 'learning_rate/action_model': 5.447895100646342e-05, 'epoch': 0.18}) +[2026-10-03 15:06:16,430][starVLA.training.train_starvla][INFO] - Step 2486, Loss: {'train/loss': 0.00014820316573604941, 'train/grad_norm_pre_clip': 0.05404253676533699, 'timing/data': 0.0003281699900981039, 'timing/model': 0.3652390169881983, 'learning_rate/qwen_vl_interface': 5.444852282898106e-06, 'learning_rate/action_model': 5.444852282898106e-05, 'epoch': 0.18}) +[2026-10-03 15:06:17,284][starVLA.training.train_starvla][INFO] - Step 2487, Loss: {'train/loss': 6.69164364808239e-05, 'train/grad_norm_pre_clip': 0.0121870506554842, 'timing/data': 0.00036318000638857484, 'timing/model': 0.36518552800407633, 'learning_rate/qwen_vl_interface': 5.441809385053456e-06, 'learning_rate/action_model': 5.4418093850534555e-05, 'epoch': 0.18}) +[2026-10-03 15:06:17,933][starVLA.training.train_starvla][INFO] - Step 2488, Loss: {'train/loss': 0.5639731287956238, 'train/grad_norm_pre_clip': 77.07979583740234, 'timing/data': 0.00023113000497687608, 'timing/model': 0.3660672469995916, 'learning_rate/qwen_vl_interface': 5.438766408363214e-06, 'learning_rate/action_model': 5.438766408363214e-05, 'epoch': 0.18}) +[2026-10-03 15:06:18,581][starVLA.training.train_starvla][INFO] - Step 2489, Loss: {'train/loss': 0.0001662158902036026, 'train/grad_norm_pre_clip': 0.09413126856088638, 'timing/data': 0.00027910999779123813, 'timing/model': 0.36462751700310037, 'learning_rate/qwen_vl_interface': 5.43572335407823e-06, 'learning_rate/action_model': 5.43572335407823e-05, 'epoch': 0.18}) +[2026-10-03 15:06:19,205][starVLA.training.train_starvla][INFO] - Step 2490, Loss: {'train/loss': 0.000152421576785855, 'train/grad_norm_pre_clip': 0.02945471927523613, 'timing/data': 0.0002247600059490651, 'timing/model': 0.36419193701294716, 'learning_rate/qwen_vl_interface': 5.432680223449392e-06, 'learning_rate/action_model': 5.4326802234493924e-05, 'epoch': 0.18}) +[2026-10-03 15:06:19,832][starVLA.training.train_starvla][INFO] - Step 2491, Loss: {'train/loss': 0.0006859349668957293, 'train/grad_norm_pre_clip': 0.9558663368225098, 'timing/data': 0.00028512999415397644, 'timing/model': 0.34250523000082467, 'learning_rate/qwen_vl_interface': 5.429637017727614e-06, 'learning_rate/action_model': 5.4296370177276144e-05, 'epoch': 0.18}) +[2026-10-03 15:06:20,678][starVLA.training.train_starvla][INFO] - Step 2492, Loss: {'train/loss': 0.0006144960061646998, 'train/grad_norm_pre_clip': 0.10743480175733566, 'timing/data': 0.00033862999407574534, 'timing/model': 0.3653705769975204, 'learning_rate/qwen_vl_interface': 5.426593738163845e-06, 'learning_rate/action_model': 5.4265937381638446e-05, 'epoch': 0.18}) +[2026-10-03 15:06:21,325][starVLA.training.train_starvla][INFO] - Step 2493, Loss: {'train/loss': 0.0002055532968370244, 'train/grad_norm_pre_clip': 0.1578151434659958, 'timing/data': 0.00023870999575592577, 'timing/model': 0.3646149370033527, 'learning_rate/qwen_vl_interface': 5.42355038600906e-06, 'learning_rate/action_model': 5.42355038600906e-05, 'epoch': 0.18}) +[2026-10-03 15:06:21,971][starVLA.training.train_starvla][INFO] - Step 2494, Loss: {'train/loss': 0.3856700360774994, 'train/grad_norm_pre_clip': 64.28790283203125, 'timing/data': 0.0002943599974969402, 'timing/model': 0.3638680170115549, 'learning_rate/qwen_vl_interface': 5.420506962514267e-06, 'learning_rate/action_model': 5.420506962514267e-05, 'epoch': 0.18}) +[2026-10-03 15:06:22,620][starVLA.training.train_starvla][INFO] - Step 2495, Loss: {'train/loss': 3.714019840117544e-05, 'train/grad_norm_pre_clip': 0.005728641990572214, 'timing/data': 0.00032300000020768493, 'timing/model': 0.36418399700778536, 'learning_rate/qwen_vl_interface': 5.4174634689305015e-06, 'learning_rate/action_model': 5.417463468930501e-05, 'epoch': 0.18}) +[2026-10-03 15:06:23,268][starVLA.training.train_starvla][INFO] - Step 2496, Loss: {'train/loss': 0.00011733781138900667, 'train/grad_norm_pre_clip': 0.030346771702170372, 'timing/data': 0.00019941999926231802, 'timing/model': 0.3643265379942022, 'learning_rate/qwen_vl_interface': 5.41441990650883e-06, 'learning_rate/action_model': 5.4144199065088295e-05, 'epoch': 0.18}) +[2026-10-03 15:06:24,113][starVLA.training.train_starvla][INFO] - Step 2497, Loss: {'train/loss': 0.28482452034950256, 'train/grad_norm_pre_clip': 12.51772403717041, 'timing/data': 0.00024287000996991992, 'timing/model': 0.5641089290002128, 'learning_rate/qwen_vl_interface': 5.411376276500345e-06, 'learning_rate/action_model': 5.411376276500345e-05, 'epoch': 0.18}) +[2026-10-03 15:06:24,759][starVLA.training.train_starvla][INFO] - Step 2498, Loss: {'train/loss': 0.0025400109589099884, 'train/grad_norm_pre_clip': 2.0574123859405518, 'timing/data': 0.00022928000544197857, 'timing/model': 0.3639941879955586, 'learning_rate/qwen_vl_interface': 5.408332580156168e-06, 'learning_rate/action_model': 5.408332580156168e-05, 'epoch': 0.18}) +[2026-10-03 15:06:25,413][starVLA.training.train_starvla][INFO] - Step 2499, Loss: {'train/loss': 5.051281186752021e-05, 'train/grad_norm_pre_clip': 0.008931194432079792, 'timing/data': 0.0003522100014379248, 'timing/model': 0.36898783600190654, 'learning_rate/qwen_vl_interface': 5.40528881872745e-06, 'learning_rate/action_model': 5.40528881872745e-05, 'epoch': 0.18}) +[2026-10-03 15:06:26,074][starVLA.training.train_starvla][INFO] - Step 2500, Loss: {'train/loss': 7.571080641355366e-05, 'train/grad_norm_pre_clip': 0.025001512840390205, 'timing/data': 0.00032954900234472007, 'timing/model': 0.3718011769960867, 'learning_rate/qwen_vl_interface': 5.402244993465364e-06, 'learning_rate/action_model': 5.402244993465364e-05, 'epoch': 0.18}) +[2026-10-03 15:06:26,736][starVLA.training.train_starvla][INFO] - Step 2501, Loss: {'train/loss': 0.0017134032677859068, 'train/grad_norm_pre_clip': 1.18121337890625, 'timing/data': 0.00026787999377120286, 'timing/model': 0.3724486059945775, 'learning_rate/qwen_vl_interface': 5.399201105621111e-06, 'learning_rate/action_model': 5.39920110562111e-05, 'epoch': 0.18}) +[2026-10-03 15:06:27,604][starVLA.training.train_starvla][INFO] - Step 2502, Loss: {'train/loss': 0.00018553333939053118, 'train/grad_norm_pre_clip': 0.08804049342870712, 'timing/data': 0.00033830000029411167, 'timing/model': 0.578234626998892, 'learning_rate/qwen_vl_interface': 5.396157156445918e-06, 'learning_rate/action_model': 5.396157156445918e-05, 'epoch': 0.18}) +[2026-10-03 15:06:28,263][starVLA.training.train_starvla][INFO] - Step 2503, Loss: {'train/loss': 0.45463645458221436, 'train/grad_norm_pre_clip': 64.50665283203125, 'timing/data': 0.00025768000341486186, 'timing/model': 0.37009437699452974, 'learning_rate/qwen_vl_interface': 5.39311314719104e-06, 'learning_rate/action_model': 5.39311314719104e-05, 'epoch': 0.18}) +[2026-10-03 15:06:28,923][starVLA.training.train_starvla][INFO] - Step 2504, Loss: {'train/loss': 5.850966044818051e-05, 'train/grad_norm_pre_clip': 0.012099431827664375, 'timing/data': 0.0002912700001616031, 'timing/model': 0.3715420969965635, 'learning_rate/qwen_vl_interface': 5.390069079107755e-06, 'learning_rate/action_model': 5.390069079107755e-05, 'epoch': 0.18}) +[2026-10-03 15:06:29,586][starVLA.training.train_starvla][INFO] - Step 2505, Loss: {'train/loss': 0.22860202193260193, 'train/grad_norm_pre_clip': 11.885066986083984, 'timing/data': 0.0002634100092109293, 'timing/model': 0.37292114600131754, 'learning_rate/qwen_vl_interface': 5.387024953447361e-06, 'learning_rate/action_model': 5.387024953447361e-05, 'epoch': 0.18}) +[2026-10-03 15:06:30,247][starVLA.training.train_starvla][INFO] - Step 2506, Loss: {'train/loss': 0.32150787115097046, 'train/grad_norm_pre_clip': 41.3072395324707, 'timing/data': 0.00034512999991420656, 'timing/model': 0.3710750070022186, 'learning_rate/qwen_vl_interface': 5.383980771461184e-06, 'learning_rate/action_model': 5.3839807714611846e-05, 'epoch': 0.18}) +[2026-10-03 15:06:30,908][starVLA.training.train_starvla][INFO] - Step 2507, Loss: {'train/loss': 0.16582781076431274, 'train/grad_norm_pre_clip': 78.9541244506836, 'timing/data': 0.00026563000574242324, 'timing/model': 0.3705724859901238, 'learning_rate/qwen_vl_interface': 5.380936534400572e-06, 'learning_rate/action_model': 5.3809365344005726e-05, 'epoch': 0.18}) +[2026-10-03 15:06:31,771][starVLA.training.train_starvla][INFO] - Step 2508, Loss: {'train/loss': 0.0005055232322774827, 'train/grad_norm_pre_clip': 0.14156004786491394, 'timing/data': 0.0003207200061297044, 'timing/model': 0.37281923700356856, 'learning_rate/qwen_vl_interface': 5.377892243516899e-06, 'learning_rate/action_model': 5.3778922435168985e-05, 'epoch': 0.18}) +[2026-10-03 15:06:32,433][starVLA.training.train_starvla][INFO] - Step 2509, Loss: {'train/loss': 6.407078763004392e-05, 'train/grad_norm_pre_clip': 0.010677103884518147, 'timing/data': 0.0003351799969095737, 'timing/model': 0.3722894870006712, 'learning_rate/qwen_vl_interface': 5.3748479000615525e-06, 'learning_rate/action_model': 5.374847900061553e-05, 'epoch': 0.18}) +[2026-10-03 15:06:33,094][starVLA.training.train_starvla][INFO] - Step 2510, Loss: {'train/loss': 0.21286867558956146, 'train/grad_norm_pre_clip': 275.4687805175781, 'timing/data': 0.0002227500081062317, 'timing/model': 0.3708198359963717, 'learning_rate/qwen_vl_interface': 5.371803505285954e-06, 'learning_rate/action_model': 5.371803505285954e-05, 'epoch': 0.18}) +[2026-10-03 15:06:33,756][starVLA.training.train_starvla][INFO] - Step 2511, Loss: {'train/loss': 0.0003108790551777929, 'train/grad_norm_pre_clip': 0.12536028027534485, 'timing/data': 0.0003223000094294548, 'timing/model': 0.37098696699831635, 'learning_rate/qwen_vl_interface': 5.368759060441534e-06, 'learning_rate/action_model': 5.368759060441534e-05, 'epoch': 0.18}) +[2026-10-03 15:06:34,394][starVLA.training.train_starvla][INFO] - Step 2512, Loss: {'train/loss': 0.00011395435285521671, 'train/grad_norm_pre_clip': 0.03724295645952225, 'timing/data': 0.0002723500074353069, 'timing/model': 0.3488212390075205, 'learning_rate/qwen_vl_interface': 5.365714566779749e-06, 'learning_rate/action_model': 5.3657145667797484e-05, 'epoch': 0.18}) +[2026-10-03 15:06:35,261][starVLA.training.train_starvla][INFO] - Step 2513, Loss: {'train/loss': 0.15020938217639923, 'train/grad_norm_pre_clip': 83.72671508789062, 'timing/data': 0.00022832000104244798, 'timing/model': 0.37305709600332193, 'learning_rate/qwen_vl_interface': 5.362670025552077e-06, 'learning_rate/action_model': 5.362670025552077e-05, 'epoch': 0.18}) +[2026-10-03 15:06:35,923][starVLA.training.train_starvla][INFO] - Step 2514, Loss: {'train/loss': 0.4059171676635742, 'train/grad_norm_pre_clip': 47.014034271240234, 'timing/data': 0.0002312700089532882, 'timing/model': 0.3719827259919839, 'learning_rate/qwen_vl_interface': 5.359625438010013e-06, 'learning_rate/action_model': 5.359625438010013e-05, 'epoch': 0.18}) +[2026-10-03 15:06:36,562][starVLA.training.train_starvla][INFO] - Step 2515, Loss: {'train/loss': 0.4559045732021332, 'train/grad_norm_pre_clip': 19.349523544311523, 'timing/data': 0.00032448999991174787, 'timing/model': 0.34835174099134747, 'learning_rate/qwen_vl_interface': 5.356580805405075e-06, 'learning_rate/action_model': 5.356580805405075e-05, 'epoch': 0.18}) +[2026-10-03 15:06:37,223][starVLA.training.train_starvla][INFO] - Step 2516, Loss: {'train/loss': 0.0001444996742065996, 'train/grad_norm_pre_clip': 0.0337800495326519, 'timing/data': 0.0002757699985522777, 'timing/model': 0.37210366599902045, 'learning_rate/qwen_vl_interface': 5.353536128988791e-06, 'learning_rate/action_model': 5.353536128988791e-05, 'epoch': 0.18}) +[2026-10-03 15:06:37,885][starVLA.training.train_starvla][INFO] - Step 2517, Loss: {'train/loss': 0.000584045541472733, 'train/grad_norm_pre_clip': 0.20230679214000702, 'timing/data': 0.00022589000582229346, 'timing/model': 0.3723656369984383, 'learning_rate/qwen_vl_interface': 5.350491410012717e-06, 'learning_rate/action_model': 5.350491410012717e-05, 'epoch': 0.18}) +[2026-10-03 15:06:38,745][starVLA.training.train_starvla][INFO] - Step 2518, Loss: {'train/loss': 0.002699455013498664, 'train/grad_norm_pre_clip': 2.1043334007263184, 'timing/data': 0.0002228800003649667, 'timing/model': 0.5692038080014754, 'learning_rate/qwen_vl_interface': 5.347446649728419e-06, 'learning_rate/action_model': 5.347446649728418e-05, 'epoch': 0.18}) +[2026-10-03 15:06:39,405][starVLA.training.train_starvla][INFO] - Step 2519, Loss: {'train/loss': 0.13910289108753204, 'train/grad_norm_pre_clip': 14.105972290039062, 'timing/data': 0.00020954999490641057, 'timing/model': 0.3712582269945415, 'learning_rate/qwen_vl_interface': 5.344401849387484e-06, 'learning_rate/action_model': 5.344401849387484e-05, 'epoch': 0.18}) +[2026-10-03 15:06:40,066][starVLA.training.train_starvla][INFO] - Step 2520, Loss: {'train/loss': 0.0015186518430709839, 'train/grad_norm_pre_clip': 1.7939436435699463, 'timing/data': 0.0003573099966160953, 'timing/model': 0.37067438699887134, 'learning_rate/qwen_vl_interface': 5.3413570102415144e-06, 'learning_rate/action_model': 5.3413570102415144e-05, 'epoch': 0.18}) +[2026-10-03 15:06:40,724][starVLA.training.train_starvla][INFO] - Step 2521, Loss: {'train/loss': 0.0003014832327608019, 'train/grad_norm_pre_clip': 0.09890490770339966, 'timing/data': 0.0003436400002101436, 'timing/model': 0.36965947700082324, 'learning_rate/qwen_vl_interface': 5.338312133542131e-06, 'learning_rate/action_model': 5.3383121335421305e-05, 'epoch': 0.18}) +[2026-10-03 15:06:41,379][starVLA.training.train_starvla][INFO] - Step 2522, Loss: {'train/loss': 0.00026640083524398506, 'train/grad_norm_pre_clip': 0.1804926097393036, 'timing/data': 0.0003182700020261109, 'timing/model': 0.3660248980013421, 'learning_rate/qwen_vl_interface': 5.335267220540963e-06, 'learning_rate/action_model': 5.335267220540963e-05, 'epoch': 0.18}) +[2026-10-03 15:06:42,031][starVLA.training.train_starvla][INFO] - Step 2523, Loss: {'train/loss': 0.00011967765749432147, 'train/grad_norm_pre_clip': 0.026067465543746948, 'timing/data': 0.0003084600029978901, 'timing/model': 0.3652050669916207, 'learning_rate/qwen_vl_interface': 5.332222272489662e-06, 'learning_rate/action_model': 5.332222272489662e-05, 'epoch': 0.18}) +[2026-10-03 15:06:42,870][starVLA.training.train_starvla][INFO] - Step 2524, Loss: {'train/loss': 0.0009964279597625136, 'train/grad_norm_pre_clip': 0.53335040807724, 'timing/data': 0.0002072100032819435, 'timing/model': 0.36415805701108184, 'learning_rate/qwen_vl_interface': 5.3291772906398895e-06, 'learning_rate/action_model': 5.32917729063989e-05, 'epoch': 0.18}) +[2026-10-03 15:06:43,516][starVLA.training.train_starvla][INFO] - Step 2525, Loss: {'train/loss': 0.0001461543724872172, 'train/grad_norm_pre_clip': 0.02847328968346119, 'timing/data': 0.00022327000624500215, 'timing/model': 0.3637038879969623, 'learning_rate/qwen_vl_interface': 5.326132276243327e-06, 'learning_rate/action_model': 5.3261322762433266e-05, 'epoch': 0.18}) +[2026-10-03 15:06:44,140][starVLA.training.train_starvla][INFO] - Step 2526, Loss: {'train/loss': 0.0003592294524423778, 'train/grad_norm_pre_clip': 0.10142099857330322, 'timing/data': 0.00022234900097828358, 'timing/model': 0.36349673799122684, 'learning_rate/qwen_vl_interface': 5.323087230551659e-06, 'learning_rate/action_model': 5.323087230551659e-05, 'epoch': 0.18}) +[2026-10-03 15:06:44,790][starVLA.training.train_starvla][INFO] - Step 2527, Loss: {'train/loss': 0.002232427941635251, 'train/grad_norm_pre_clip': 1.4686881303787231, 'timing/data': 0.00019935000455006957, 'timing/model': 0.3652287879958749, 'learning_rate/qwen_vl_interface': 5.3200421548165916e-06, 'learning_rate/action_model': 5.3200421548165914e-05, 'epoch': 0.18}) +[2026-10-03 15:06:45,439][starVLA.training.train_starvla][INFO] - Step 2528, Loss: {'train/loss': 0.5090991854667664, 'train/grad_norm_pre_clip': 55.39838790893555, 'timing/data': 0.0003528499946696684, 'timing/model': 0.36485123699822, 'learning_rate/qwen_vl_interface': 5.3169970502898405e-06, 'learning_rate/action_model': 5.31699705028984e-05, 'epoch': 0.19}) +[2026-10-03 15:06:46,284][starVLA.training.train_starvla][INFO] - Step 2529, Loss: {'train/loss': 0.2653873562812805, 'train/grad_norm_pre_clip': 38.821285247802734, 'timing/data': 0.0002494689979357645, 'timing/model': 0.36435407800308894, 'learning_rate/qwen_vl_interface': 5.31395191822313e-06, 'learning_rate/action_model': 5.31395191822313e-05, 'epoch': 0.19}) +[2026-10-03 15:06:46,930][starVLA.training.train_starvla][INFO] - Step 2530, Loss: {'train/loss': 0.00022765633184462786, 'train/grad_norm_pre_clip': 0.05698738992214203, 'timing/data': 0.0003318700037198141, 'timing/model': 0.3636897069955012, 'learning_rate/qwen_vl_interface': 5.310906759868205e-06, 'learning_rate/action_model': 5.310906759868205e-05, 'epoch': 0.19}) +[2026-10-03 15:06:47,580][starVLA.training.train_starvla][INFO] - Step 2531, Loss: {'train/loss': 0.0002838868531398475, 'train/grad_norm_pre_clip': 0.05177517980337143, 'timing/data': 0.00030436999804805964, 'timing/model': 0.3644353580020834, 'learning_rate/qwen_vl_interface': 5.30786157647681e-06, 'learning_rate/action_model': 5.30786157647681e-05, 'epoch': 0.19}) +[2026-10-03 15:06:48,227][starVLA.training.train_starvla][INFO] - Step 2532, Loss: {'train/loss': 0.42280492186546326, 'train/grad_norm_pre_clip': 153.51214599609375, 'timing/data': 0.00019851000979542732, 'timing/model': 0.364342647997546, 'learning_rate/qwen_vl_interface': 5.304816369300709e-06, 'learning_rate/action_model': 5.304816369300709e-05, 'epoch': 0.19}) +[2026-10-03 15:06:48,875][starVLA.training.train_starvla][INFO] - Step 2533, Loss: {'train/loss': 0.042944055050611496, 'train/grad_norm_pre_clip': 97.8616943359375, 'timing/data': 0.0002442599943606183, 'timing/model': 0.36488674698921386, 'learning_rate/qwen_vl_interface': 5.301771139591666e-06, 'learning_rate/action_model': 5.301771139591666e-05, 'epoch': 0.19}) +[2026-10-03 15:06:49,716][starVLA.training.train_starvla][INFO] - Step 2534, Loss: {'train/loss': 0.07747296243906021, 'train/grad_norm_pre_clip': 41.92607116699219, 'timing/data': 0.00023890999727882445, 'timing/model': 0.5580254389933543, 'learning_rate/qwen_vl_interface': 5.298725888601465e-06, 'learning_rate/action_model': 5.298725888601465e-05, 'epoch': 0.19}) +[2026-10-03 15:06:50,368][starVLA.training.train_starvla][INFO] - Step 2535, Loss: {'train/loss': 0.3422822058200836, 'train/grad_norm_pre_clip': 97.12089538574219, 'timing/data': 0.0002997999981744215, 'timing/model': 0.36626588700164575, 'learning_rate/qwen_vl_interface': 5.295680617581891e-06, 'learning_rate/action_model': 5.2956806175818904e-05, 'epoch': 0.19}) +[2026-10-03 15:06:51,017][starVLA.training.train_starvla][INFO] - Step 2536, Loss: {'train/loss': 0.00010015735460910946, 'train/grad_norm_pre_clip': 0.022738011553883553, 'timing/data': 0.00019968999549746513, 'timing/model': 0.36533139800303616, 'learning_rate/qwen_vl_interface': 5.2926353277847395e-06, 'learning_rate/action_model': 5.292635327784739e-05, 'epoch': 0.19}) +[2026-10-03 15:06:51,666][starVLA.training.train_starvla][INFO] - Step 2537, Loss: {'train/loss': 0.0006925425259396434, 'train/grad_norm_pre_clip': 0.44069144129753113, 'timing/data': 0.00029395999445114285, 'timing/model': 0.36536225699819624, 'learning_rate/qwen_vl_interface': 5.2895900204618175e-06, 'learning_rate/action_model': 5.289590020461818e-05, 'epoch': 0.19}) +[2026-10-03 15:06:52,314][starVLA.training.train_starvla][INFO] - Step 2538, Loss: {'train/loss': 0.7090128660202026, 'train/grad_norm_pre_clip': 38.9944953918457, 'timing/data': 0.00030929999775253236, 'timing/model': 0.36459967799601145, 'learning_rate/qwen_vl_interface': 5.286544696864933e-06, 'learning_rate/action_model': 5.2865446968649324e-05, 'epoch': 0.19}) +[2026-10-03 15:06:53,162][starVLA.training.train_starvla][INFO] - Step 2539, Loss: {'train/loss': 0.001294230343773961, 'train/grad_norm_pre_clip': 0.5410853028297424, 'timing/data': 0.00030327000422403216, 'timing/model': 0.5628559600008884, 'learning_rate/qwen_vl_interface': 5.283499358245904e-06, 'learning_rate/action_model': 5.2834993582459034e-05, 'epoch': 0.19}) +[2026-10-03 15:06:53,811][starVLA.training.train_starvla][INFO] - Step 2540, Loss: {'train/loss': 0.21860745549201965, 'train/grad_norm_pre_clip': 97.93850708007812, 'timing/data': 0.00023401000362355262, 'timing/model': 0.36507482698652893, 'learning_rate/qwen_vl_interface': 5.280454005856553e-06, 'learning_rate/action_model': 5.2804540058565534e-05, 'epoch': 0.19}) +[2026-10-03 15:06:54,460][starVLA.training.train_starvla][INFO] - Step 2541, Loss: {'train/loss': 0.04181652516126633, 'train/grad_norm_pre_clip': 8.261224746704102, 'timing/data': 0.0003249000001233071, 'timing/model': 0.3651521769934334, 'learning_rate/qwen_vl_interface': 5.277408640948711e-06, 'learning_rate/action_model': 5.277408640948711e-05, 'epoch': 0.19}) +[2026-10-03 15:06:55,108][starVLA.training.train_starvla][INFO] - Step 2542, Loss: {'train/loss': 0.0003247746790293604, 'train/grad_norm_pre_clip': 0.13837040960788727, 'timing/data': 0.00024408000172115862, 'timing/model': 0.3648735479946481, 'learning_rate/qwen_vl_interface': 5.274363264774213e-06, 'learning_rate/action_model': 5.274363264774212e-05, 'epoch': 0.19}) +[2026-10-03 15:06:55,759][starVLA.training.train_starvla][INFO] - Step 2543, Loss: {'train/loss': 0.1189965307712555, 'train/grad_norm_pre_clip': 107.05355834960938, 'timing/data': 0.0002664499916136265, 'timing/model': 0.36526202798995655, 'learning_rate/qwen_vl_interface': 5.271317878584897e-06, 'learning_rate/action_model': 5.2713178785848974e-05, 'epoch': 0.19}) +[2026-10-03 15:06:56,407][starVLA.training.train_starvla][INFO] - Step 2544, Loss: {'train/loss': 0.00013733608648180962, 'train/grad_norm_pre_clip': 0.039411503821611404, 'timing/data': 0.0003360499977134168, 'timing/model': 0.3655091880063992, 'learning_rate/qwen_vl_interface': 5.268272483632606e-06, 'learning_rate/action_model': 5.268272483632606e-05, 'epoch': 0.19}) +[2026-10-03 15:06:57,249][starVLA.training.train_starvla][INFO] - Step 2545, Loss: {'train/loss': 0.003298745257779956, 'train/grad_norm_pre_clip': 4.734005451202393, 'timing/data': 0.0003733500052476302, 'timing/model': 0.3647661080030957, 'learning_rate/qwen_vl_interface': 5.265227081169187e-06, 'learning_rate/action_model': 5.265227081169186e-05, 'epoch': 0.19}) +[2026-10-03 15:06:57,897][starVLA.training.train_starvla][INFO] - Step 2546, Loss: {'train/loss': 0.00025575837935321033, 'train/grad_norm_pre_clip': 0.051212336868047714, 'timing/data': 0.0002028599992627278, 'timing/model': 0.3644121380057186, 'learning_rate/qwen_vl_interface': 5.2621816724464895e-06, 'learning_rate/action_model': 5.262181672446489e-05, 'epoch': 0.19}) +[2026-10-03 15:06:58,548][starVLA.training.train_starvla][INFO] - Step 2547, Loss: {'train/loss': 0.00027919444255530834, 'train/grad_norm_pre_clip': 0.07025128602981567, 'timing/data': 0.00019876999431289732, 'timing/model': 0.3661888770002406, 'learning_rate/qwen_vl_interface': 5.259136258716367e-06, 'learning_rate/action_model': 5.2591362587163675e-05, 'epoch': 0.19}) +[2026-10-03 15:06:59,197][starVLA.training.train_starvla][INFO] - Step 2548, Loss: {'train/loss': 0.015215923078358173, 'train/grad_norm_pre_clip': 3.5395398139953613, 'timing/data': 0.0003397599939489737, 'timing/model': 0.36506571699283086, 'learning_rate/qwen_vl_interface': 5.256090841230675e-06, 'learning_rate/action_model': 5.256090841230674e-05, 'epoch': 0.19}) +[2026-10-03 15:06:59,846][starVLA.training.train_starvla][INFO] - Step 2549, Loss: {'train/loss': 0.5989774465560913, 'train/grad_norm_pre_clip': 21.98331642150879, 'timing/data': 0.00019956000323873013, 'timing/model': 0.3653476769977715, 'learning_rate/qwen_vl_interface': 5.253045421241266e-06, 'learning_rate/action_model': 5.253045421241266e-05, 'epoch': 0.19}) +[2026-10-03 15:07:00,657][starVLA.training.train_starvla][INFO] - Step 2550, Loss: {'train/loss': 0.0011023059487342834, 'train/grad_norm_pre_clip': 0.4818664789199829, 'timing/data': 0.000262100002146326, 'timing/model': 0.3645951479993528, 'learning_rate/qwen_vl_interface': 5.2500000000000006e-06, 'learning_rate/action_model': 5.25e-05, 'epoch': 0.19}) +[2026-10-03 15:07:01,309][starVLA.training.train_starvla][INFO] - Step 2551, Loss: {'train/loss': 0.42714497447013855, 'train/grad_norm_pre_clip': 48.6093635559082, 'timing/data': 0.00031419000879395753, 'timing/model': 0.36664958800247405, 'learning_rate/qwen_vl_interface': 5.246954578758735e-06, 'learning_rate/action_model': 5.246954578758735e-05, 'epoch': 0.19}) +[2026-10-03 15:07:01,958][starVLA.training.train_starvla][INFO] - Step 2552, Loss: {'train/loss': 0.0009040232980623841, 'train/grad_norm_pre_clip': 0.5749202966690063, 'timing/data': 0.00024038899573497474, 'timing/model': 0.3647457680053776, 'learning_rate/qwen_vl_interface': 5.243909158769327e-06, 'learning_rate/action_model': 5.2439091587693264e-05, 'epoch': 0.19}) +[2026-10-03 15:07:02,606][starVLA.training.train_starvla][INFO] - Step 2553, Loss: {'train/loss': 0.19827288389205933, 'train/grad_norm_pre_clip': 27.419078826904297, 'timing/data': 0.00024402000417467207, 'timing/model': 0.3652958970051259, 'learning_rate/qwen_vl_interface': 5.240863741283634e-06, 'learning_rate/action_model': 5.2408637412836336e-05, 'epoch': 0.19}) +[2026-10-03 15:07:03,254][starVLA.training.train_starvla][INFO] - Step 2554, Loss: {'train/loss': 0.0003481586172711104, 'train/grad_norm_pre_clip': 0.08057377487421036, 'timing/data': 0.0003257000062149018, 'timing/model': 0.36494168800709303, 'learning_rate/qwen_vl_interface': 5.2378183275535125e-06, 'learning_rate/action_model': 5.237818327553512e-05, 'epoch': 0.19}) +[2026-10-03 15:07:04,105][starVLA.training.train_starvla][INFO] - Step 2555, Loss: {'train/loss': 0.06552828848361969, 'train/grad_norm_pre_clip': 117.61701202392578, 'timing/data': 0.0002470500039635226, 'timing/model': 0.565236629001447, 'learning_rate/qwen_vl_interface': 5.234772918830815e-06, 'learning_rate/action_model': 5.234772918830815e-05, 'epoch': 0.19}) +[2026-10-03 15:07:04,754][starVLA.training.train_starvla][INFO] - Step 2556, Loss: {'train/loss': 0.00014685631322208792, 'train/grad_norm_pre_clip': 0.08524653315544128, 'timing/data': 0.00025429000379517674, 'timing/model': 0.36489823799638543, 'learning_rate/qwen_vl_interface': 5.231727516367395e-06, 'learning_rate/action_model': 5.231727516367395e-05, 'epoch': 0.19}) +[2026-10-03 15:07:05,401][starVLA.training.train_starvla][INFO] - Step 2557, Loss: {'train/loss': 0.00010800959717016667, 'train/grad_norm_pre_clip': 0.025047654286026955, 'timing/data': 0.0003124599898001179, 'timing/model': 0.3645504180021817, 'learning_rate/qwen_vl_interface': 5.228682121415104e-06, 'learning_rate/action_model': 5.228682121415104e-05, 'epoch': 0.19}) +[2026-10-03 15:07:06,049][starVLA.training.train_starvla][INFO] - Step 2558, Loss: {'train/loss': 6.641430809395388e-05, 'train/grad_norm_pre_clip': 0.013321205042302608, 'timing/data': 0.00020435999613255262, 'timing/model': 0.3644817479944322, 'learning_rate/qwen_vl_interface': 5.225636735225789e-06, 'learning_rate/action_model': 5.225636735225788e-05, 'epoch': 0.19}) +[2026-10-03 15:07:06,698][starVLA.training.train_starvla][INFO] - Step 2559, Loss: {'train/loss': 0.0021416661329567432, 'train/grad_norm_pre_clip': 5.875970840454102, 'timing/data': 0.0002857999934349209, 'timing/model': 0.3643399180000415, 'learning_rate/qwen_vl_interface': 5.22259135905129e-06, 'learning_rate/action_model': 5.2225913590512895e-05, 'epoch': 0.19}) +[2026-10-03 15:07:07,550][starVLA.training.train_starvla][INFO] - Step 2560, Loss: {'train/loss': 0.00038054276956245303, 'train/grad_norm_pre_clip': 0.4172254204750061, 'timing/data': 0.0003799090045504272, 'timing/model': 0.5678153689950705, 'learning_rate/qwen_vl_interface': 5.219545994143448e-06, 'learning_rate/action_model': 5.219545994143448e-05, 'epoch': 0.19}) +[2026-10-03 15:07:08,197][starVLA.training.train_starvla][INFO] - Step 2561, Loss: {'train/loss': 0.0005432490725070238, 'train/grad_norm_pre_clip': 0.19860048592090607, 'timing/data': 0.00024452000798191875, 'timing/model': 0.36438810800609645, 'learning_rate/qwen_vl_interface': 5.216500641754097e-06, 'learning_rate/action_model': 5.216500641754097e-05, 'epoch': 0.19}) +[2026-10-03 15:07:08,844][starVLA.training.train_starvla][INFO] - Step 2562, Loss: {'train/loss': 0.0006541848997585475, 'train/grad_norm_pre_clip': 0.8067309260368347, 'timing/data': 0.0003168599942000583, 'timing/model': 0.3642899080005009, 'learning_rate/qwen_vl_interface': 5.213455303135067e-06, 'learning_rate/action_model': 5.2134553031350673e-05, 'epoch': 0.19}) +[2026-10-03 15:07:09,494][starVLA.training.train_starvla][INFO] - Step 2563, Loss: {'train/loss': 0.004250084515661001, 'train/grad_norm_pre_clip': 5.804263591766357, 'timing/data': 0.0001978799991775304, 'timing/model': 0.3648503680014983, 'learning_rate/qwen_vl_interface': 5.2104099795381836e-06, 'learning_rate/action_model': 5.210409979538183e-05, 'epoch': 0.19}) +[2026-10-03 15:07:10,142][starVLA.training.train_starvla][INFO] - Step 2564, Loss: {'train/loss': 0.00028540630592033267, 'train/grad_norm_pre_clip': 0.13150568306446075, 'timing/data': 0.00033987000642810017, 'timing/model': 0.36473903799196705, 'learning_rate/qwen_vl_interface': 5.207364672215261e-06, 'learning_rate/action_model': 5.2073646722152604e-05, 'epoch': 0.19}) +[2026-10-03 15:07:10,789][starVLA.training.train_starvla][INFO] - Step 2565, Loss: {'train/loss': 6.504809425678104e-05, 'train/grad_norm_pre_clip': 0.014175459742546082, 'timing/data': 0.00024873000802472234, 'timing/model': 0.36451778799528256, 'learning_rate/qwen_vl_interface': 5.2043193824181095e-06, 'learning_rate/action_model': 5.2043193824181094e-05, 'epoch': 0.19}) +[2026-10-03 15:07:11,602][starVLA.training.train_starvla][INFO] - Step 2566, Loss: {'train/loss': 0.21209406852722168, 'train/grad_norm_pre_clip': 156.822021484375, 'timing/data': 0.0003207400004612282, 'timing/model': 0.3651364279940026, 'learning_rate/qwen_vl_interface': 5.201274111398535e-06, 'learning_rate/action_model': 5.2012741113985356e-05, 'epoch': 0.19}) +[2026-10-03 15:07:12,252][starVLA.training.train_starvla][INFO] - Step 2567, Loss: {'train/loss': 0.0002651028335094452, 'train/grad_norm_pre_clip': 0.345691978931427, 'timing/data': 0.00026061000244226307, 'timing/model': 0.3649502080079401, 'learning_rate/qwen_vl_interface': 5.198228860408334e-06, 'learning_rate/action_model': 5.198228860408334e-05, 'epoch': 0.19}) +[2026-10-03 15:07:12,899][starVLA.training.train_starvla][INFO] - Step 2568, Loss: {'train/loss': 0.005927456542849541, 'train/grad_norm_pre_clip': 1.3257122039794922, 'timing/data': 0.00032905000261962414, 'timing/model': 0.3640630379959475, 'learning_rate/qwen_vl_interface': 5.195183630699292e-06, 'learning_rate/action_model': 5.195183630699292e-05, 'epoch': 0.19}) +[2026-10-03 15:07:13,545][starVLA.training.train_starvla][INFO] - Step 2569, Loss: {'train/loss': 0.00026738489395938814, 'train/grad_norm_pre_clip': 0.11188506335020065, 'timing/data': 0.000244759998167865, 'timing/model': 0.3640827880008146, 'learning_rate/qwen_vl_interface': 5.192138423523189e-06, 'learning_rate/action_model': 5.1921384235231895e-05, 'epoch': 0.19}) +[2026-10-03 15:07:14,193][starVLA.training.train_starvla][INFO] - Step 2570, Loss: {'train/loss': 6.195095920702443e-05, 'train/grad_norm_pre_clip': 0.02041228488087654, 'timing/data': 0.00025848000950645655, 'timing/model': 0.36452861700672656, 'learning_rate/qwen_vl_interface': 5.189093240131795e-06, 'learning_rate/action_model': 5.189093240131795e-05, 'epoch': 0.19}) +[2026-10-03 15:07:15,037][starVLA.training.train_starvla][INFO] - Step 2571, Loss: {'train/loss': 0.0006267299759201705, 'train/grad_norm_pre_clip': 0.6058207154273987, 'timing/data': 0.00029610999627038836, 'timing/model': 0.36467748900759034, 'learning_rate/qwen_vl_interface': 5.186048081776869e-06, 'learning_rate/action_model': 5.1860480817768685e-05, 'epoch': 0.19}) +[2026-10-03 15:07:15,684][starVLA.training.train_starvla][INFO] - Step 2572, Loss: {'train/loss': 0.005912140477448702, 'train/grad_norm_pre_clip': 4.520036697387695, 'timing/data': 0.0003323900018585846, 'timing/model': 0.3646036779973656, 'learning_rate/qwen_vl_interface': 5.183002949710161e-06, 'learning_rate/action_model': 5.18300294971016e-05, 'epoch': 0.19}) +[2026-10-03 15:07:16,332][starVLA.training.train_starvla][INFO] - Step 2573, Loss: {'train/loss': 2.921231134678237e-05, 'train/grad_norm_pre_clip': 0.00973929837346077, 'timing/data': 0.00028634999762289226, 'timing/model': 0.3648080480052158, 'learning_rate/qwen_vl_interface': 5.179957845183411e-06, 'learning_rate/action_model': 5.179957845183411e-05, 'epoch': 0.19}) +[2026-10-03 15:07:16,978][starVLA.training.train_starvla][INFO] - Step 2574, Loss: {'train/loss': 1.5035219192504883, 'train/grad_norm_pre_clip': 116.67532348632812, 'timing/data': 0.00019356999837327749, 'timing/model': 0.3644755580025958, 'learning_rate/qwen_vl_interface': 5.176912769448343e-06, 'learning_rate/action_model': 5.176912769448343e-05, 'epoch': 0.19}) +[2026-10-03 15:07:17,628][starVLA.training.train_starvla][INFO] - Step 2575, Loss: {'train/loss': 0.00018084849580191076, 'train/grad_norm_pre_clip': 0.08214215189218521, 'timing/data': 0.00033570000960025936, 'timing/model': 0.364057008002419, 'learning_rate/qwen_vl_interface': 5.173867723756676e-06, 'learning_rate/action_model': 5.173867723756676e-05, 'epoch': 0.19}) +[2026-10-03 15:07:18,467][starVLA.training.train_starvla][INFO] - Step 2576, Loss: {'train/loss': 0.002142852172255516, 'train/grad_norm_pre_clip': 0.44877272844314575, 'timing/data': 0.0003641600051196292, 'timing/model': 0.5561553699953947, 'learning_rate/qwen_vl_interface': 5.170822709360112e-06, 'learning_rate/action_model': 5.170822709360111e-05, 'epoch': 0.19}) +[2026-10-03 15:07:19,113][starVLA.training.train_starvla][INFO] - Step 2577, Loss: {'train/loss': 2.121138277288992e-05, 'train/grad_norm_pre_clip': 0.004781165160238743, 'timing/data': 0.00031813999521546066, 'timing/model': 0.36394073799601756, 'learning_rate/qwen_vl_interface': 5.16777772751034e-06, 'learning_rate/action_model': 5.16777772751034e-05, 'epoch': 0.19}) +[2026-10-03 15:07:19,758][starVLA.training.train_starvla][INFO] - Step 2578, Loss: {'train/loss': 1.2918509244918823, 'train/grad_norm_pre_clip': 48.488826751708984, 'timing/data': 0.0002729600091697648, 'timing/model': 0.3634953279979527, 'learning_rate/qwen_vl_interface': 5.16473277945904e-06, 'learning_rate/action_model': 5.1647327794590396e-05, 'epoch': 0.19}) +[2026-10-03 15:07:20,408][starVLA.training.train_starvla][INFO] - Step 2579, Loss: {'train/loss': 0.8751540184020996, 'train/grad_norm_pre_clip': 40.14617156982422, 'timing/data': 0.00029129999165888876, 'timing/model': 0.3654380480002146, 'learning_rate/qwen_vl_interface': 5.161687866457872e-06, 'learning_rate/action_model': 5.161687866457871e-05, 'epoch': 0.19}) +[2026-10-03 15:07:21,054][starVLA.training.train_starvla][INFO] - Step 2580, Loss: {'train/loss': 6.355428195092827e-05, 'train/grad_norm_pre_clip': 0.021299103274941444, 'timing/data': 0.00019987999985460192, 'timing/model': 0.364186927996343, 'learning_rate/qwen_vl_interface': 5.158642989758487e-06, 'learning_rate/action_model': 5.158642989758487e-05, 'epoch': 0.19}) +[2026-10-03 15:07:21,873][starVLA.training.train_starvla][INFO] - Step 2581, Loss: {'train/loss': 0.5630314946174622, 'train/grad_norm_pre_clip': 78.96529388427734, 'timing/data': 0.00030656000308226794, 'timing/model': 0.5589552099991124, 'learning_rate/qwen_vl_interface': 5.155598150612519e-06, 'learning_rate/action_model': 5.1555981506125185e-05, 'epoch': 0.19}) +[2026-10-03 15:07:22,520][starVLA.training.train_starvla][INFO] - Step 2582, Loss: {'train/loss': 0.000331850751535967, 'train/grad_norm_pre_clip': 0.08755393326282501, 'timing/data': 0.00024052000662777573, 'timing/model': 0.3646665770065738, 'learning_rate/qwen_vl_interface': 5.152553350271584e-06, 'learning_rate/action_model': 5.152553350271584e-05, 'epoch': 0.19}) +[2026-10-03 15:07:23,170][starVLA.training.train_starvla][INFO] - Step 2583, Loss: {'train/loss': 0.48071950674057007, 'train/grad_norm_pre_clip': 30.624408721923828, 'timing/data': 0.0003855399991152808, 'timing/model': 0.3649142869981006, 'learning_rate/qwen_vl_interface': 5.149508589987286e-06, 'learning_rate/action_model': 5.1495085899872855e-05, 'epoch': 0.19}) +[2026-10-03 15:07:23,817][starVLA.training.train_starvla][INFO] - Step 2584, Loss: {'train/loss': 0.40118375420570374, 'train/grad_norm_pre_clip': 45.176753997802734, 'timing/data': 0.0003412000078242272, 'timing/model': 0.3643811980000464, 'learning_rate/qwen_vl_interface': 5.14646387101121e-06, 'learning_rate/action_model': 5.14646387101121e-05, 'epoch': 0.19}) +[2026-10-03 15:07:24,464][starVLA.training.train_starvla][INFO] - Step 2585, Loss: {'train/loss': 0.0007008202373981476, 'train/grad_norm_pre_clip': 0.2772506773471832, 'timing/data': 0.000332309995428659, 'timing/model': 0.36431947800156195, 'learning_rate/qwen_vl_interface': 5.143419194594928e-06, 'learning_rate/action_model': 5.143419194594927e-05, 'epoch': 0.19}) +[2026-10-03 15:07:25,111][starVLA.training.train_starvla][INFO] - Step 2586, Loss: {'train/loss': 0.00015174495638348162, 'train/grad_norm_pre_clip': 0.046299394220113754, 'timing/data': 0.00030412001069635153, 'timing/model': 0.36424315700423904, 'learning_rate/qwen_vl_interface': 5.140374561989988e-06, 'learning_rate/action_model': 5.140374561989988e-05, 'epoch': 0.19}) +[2026-10-03 15:07:25,960][starVLA.training.train_starvla][INFO] - Step 2587, Loss: {'train/loss': 0.00016526661056559533, 'train/grad_norm_pre_clip': 0.03076937422156334, 'timing/data': 0.00029207998886704445, 'timing/model': 0.36493266800243873, 'learning_rate/qwen_vl_interface': 5.137329974447924e-06, 'learning_rate/action_model': 5.137329974447924e-05, 'epoch': 0.19}) +[2026-10-03 15:07:26,591][starVLA.training.train_starvla][INFO] - Step 2588, Loss: {'train/loss': 0.0003828869666904211, 'train/grad_norm_pre_clip': 0.1139695942401886, 'timing/data': 0.00022146000992506742, 'timing/model': 0.3705008170072688, 'learning_rate/qwen_vl_interface': 5.1342854332202516e-06, 'learning_rate/action_model': 5.1342854332202514e-05, 'epoch': 0.19}) +[2026-10-03 15:07:27,238][starVLA.training.train_starvla][INFO] - Step 2589, Loss: {'train/loss': 0.0004005951341241598, 'train/grad_norm_pre_clip': 0.08687283843755722, 'timing/data': 0.00019829999655485153, 'timing/model': 0.364460587996291, 'learning_rate/qwen_vl_interface': 5.131240939558467e-06, 'learning_rate/action_model': 5.131240939558467e-05, 'epoch': 0.19}) +[2026-10-03 15:07:27,886][starVLA.training.train_starvla][INFO] - Step 2590, Loss: {'train/loss': 0.00017901592946145684, 'train/grad_norm_pre_clip': 0.046092819422483444, 'timing/data': 0.0002977900003315881, 'timing/model': 0.36430144699988887, 'learning_rate/qwen_vl_interface': 5.128196494714048e-06, 'learning_rate/action_model': 5.128196494714047e-05, 'epoch': 0.19}) +[2026-10-03 15:07:28,536][starVLA.training.train_starvla][INFO] - Step 2591, Loss: {'train/loss': 0.0001125735288951546, 'train/grad_norm_pre_clip': 0.02465064264833927, 'timing/data': 0.0003370400081621483, 'timing/model': 0.3654426770081045, 'learning_rate/qwen_vl_interface': 5.125152099938448e-06, 'learning_rate/action_model': 5.125152099938447e-05, 'epoch': 0.19}) +[2026-10-03 15:07:29,370][starVLA.training.train_starvla][INFO] - Step 2592, Loss: {'train/loss': 0.00035222736187279224, 'train/grad_norm_pre_clip': 0.06982088834047318, 'timing/data': 0.00021209000260569155, 'timing/model': 0.5507422509981552, 'learning_rate/qwen_vl_interface': 5.122107756483104e-06, 'learning_rate/action_model': 5.122107756483103e-05, 'epoch': 0.19}) +[2026-10-03 15:07:30,017][starVLA.training.train_starvla][INFO] - Step 2593, Loss: {'train/loss': 0.00032009146525524557, 'train/grad_norm_pre_clip': 0.07588351517915726, 'timing/data': 0.000314519988023676, 'timing/model': 0.36427729899878614, 'learning_rate/qwen_vl_interface': 5.119063465599429e-06, 'learning_rate/action_model': 5.1190634655994285e-05, 'epoch': 0.19}) +[2026-10-03 15:07:30,665][starVLA.training.train_starvla][INFO] - Step 2594, Loss: {'train/loss': 0.0005028939922340214, 'train/grad_norm_pre_clip': 0.13373108208179474, 'timing/data': 0.00022048001119401306, 'timing/model': 0.3646246779971989, 'learning_rate/qwen_vl_interface': 5.116019228538818e-06, 'learning_rate/action_model': 5.116019228538818e-05, 'epoch': 0.19}) +[2026-10-03 15:07:31,292][starVLA.training.train_starvla][INFO] - Step 2595, Loss: {'train/loss': 0.48723942041397095, 'train/grad_norm_pre_clip': 41.189300537109375, 'timing/data': 0.00020167999900877476, 'timing/model': 0.34312583100108895, 'learning_rate/qwen_vl_interface': 5.112975046552641e-06, 'learning_rate/action_model': 5.1129750465526404e-05, 'epoch': 0.19}) +[2026-10-03 15:07:31,941][starVLA.training.train_starvla][INFO] - Step 2596, Loss: {'train/loss': 0.0003769596805796027, 'train/grad_norm_pre_clip': 0.10283266752958298, 'timing/data': 0.00019768001220654696, 'timing/model': 0.36530157800007146, 'learning_rate/qwen_vl_interface': 5.109930920892246e-06, 'learning_rate/action_model': 5.1099309208922464e-05, 'epoch': 0.19}) +[2026-10-03 15:07:32,784][starVLA.training.train_starvla][INFO] - Step 2597, Loss: {'train/loss': 0.07410743087530136, 'train/grad_norm_pre_clip': 162.74063110351562, 'timing/data': 0.0002753899898380041, 'timing/model': 0.5603146689973073, 'learning_rate/qwen_vl_interface': 5.10688685280896e-06, 'learning_rate/action_model': 5.10688685280896e-05, 'epoch': 0.19}) +[2026-10-03 15:07:33,431][starVLA.training.train_starvla][INFO] - Step 2598, Loss: {'train/loss': 0.4784794747829437, 'train/grad_norm_pre_clip': 32.435211181640625, 'timing/data': 0.00019808000070042908, 'timing/model': 0.3646960490004858, 'learning_rate/qwen_vl_interface': 5.103842843554082e-06, 'learning_rate/action_model': 5.103842843554082e-05, 'epoch': 0.19}) +[2026-10-03 15:07:34,080][starVLA.training.train_starvla][INFO] - Step 2599, Loss: {'train/loss': 0.23281602561473846, 'train/grad_norm_pre_clip': 90.55223083496094, 'timing/data': 0.00031535999733023345, 'timing/model': 0.36508832800609525, 'learning_rate/qwen_vl_interface': 5.100798894378891e-06, 'learning_rate/action_model': 5.100798894378891e-05, 'epoch': 0.19}) +[2026-10-03 15:07:34,706][starVLA.training.train_starvla][INFO] - Step 2600, Loss: {'train/loss': 0.0009840333368629217, 'train/grad_norm_pre_clip': 0.2739288806915283, 'timing/data': 0.00021868999465368688, 'timing/model': 0.34276822199171875, 'learning_rate/qwen_vl_interface': 5.097755006534638e-06, 'learning_rate/action_model': 5.097755006534638e-05, 'epoch': 0.19}) +[2026-10-03 15:07:35,354][starVLA.training.train_starvla][INFO] - Step 2601, Loss: {'train/loss': 0.38511234521865845, 'train/grad_norm_pre_clip': 219.53807067871094, 'timing/data': 0.0002969999914057553, 'timing/model': 0.3649137780012097, 'learning_rate/qwen_vl_interface': 5.094711181272551e-06, 'learning_rate/action_model': 5.094711181272551e-05, 'epoch': 0.19}) +[2026-10-03 15:07:36,002][starVLA.training.train_starvla][INFO] - Step 2602, Loss: {'train/loss': 0.3151564598083496, 'train/grad_norm_pre_clip': 39.653236389160156, 'timing/data': 0.0002397600037511438, 'timing/model': 0.36458852799842134, 'learning_rate/qwen_vl_interface': 5.091667419843833e-06, 'learning_rate/action_model': 5.091667419843833e-05, 'epoch': 0.19}) +[2026-10-03 15:07:36,844][starVLA.training.train_starvla][INFO] - Step 2603, Loss: {'train/loss': 0.0008367456612177193, 'train/grad_norm_pre_clip': 0.12415634095668793, 'timing/data': 0.0002968100016005337, 'timing/model': 0.3649951380066341, 'learning_rate/qwen_vl_interface': 5.088623723499657e-06, 'learning_rate/action_model': 5.088623723499657e-05, 'epoch': 0.19}) +[2026-10-03 15:07:37,469][starVLA.training.train_starvla][INFO] - Step 2604, Loss: {'train/loss': 0.0009109072270803154, 'train/grad_norm_pre_clip': 0.1099555641412735, 'timing/data': 0.00025128999550361186, 'timing/model': 0.34217745100613683, 'learning_rate/qwen_vl_interface': 5.085580093491171e-06, 'learning_rate/action_model': 5.085580093491171e-05, 'epoch': 0.19}) +[2026-10-03 15:07:38,116][starVLA.training.train_starvla][INFO] - Step 2605, Loss: {'train/loss': 0.0016607161378487945, 'train/grad_norm_pre_clip': 1.1310898065567017, 'timing/data': 0.0003601200005505234, 'timing/model': 0.3641923989925999, 'learning_rate/qwen_vl_interface': 5.0825365310695e-06, 'learning_rate/action_model': 5.082536531069499e-05, 'epoch': 0.19}) +[2026-10-03 15:07:38,741][starVLA.training.train_starvla][INFO] - Step 2606, Loss: {'train/loss': 0.0013223827118054032, 'train/grad_norm_pre_clip': 0.1659628003835678, 'timing/data': 0.00020244999905116856, 'timing/model': 0.3641109470045194, 'learning_rate/qwen_vl_interface': 5.079493037485735e-06, 'learning_rate/action_model': 5.0794930374857344e-05, 'epoch': 0.19}) +[2026-10-03 15:07:39,392][starVLA.training.train_starvla][INFO] - Step 2607, Loss: {'train/loss': 0.26997748017311096, 'train/grad_norm_pre_clip': 9.543937683105469, 'timing/data': 0.0003202299994882196, 'timing/model': 0.36553913800162263, 'learning_rate/qwen_vl_interface': 5.076449613990941e-06, 'learning_rate/action_model': 5.076449613990941e-05, 'epoch': 0.19}) +[2026-10-03 15:07:40,234][starVLA.training.train_starvla][INFO] - Step 2608, Loss: {'train/loss': 0.00039102142909541726, 'train/grad_norm_pre_clip': 0.10617898404598236, 'timing/data': 0.0003367300087120384, 'timing/model': 0.36526381799194496, 'learning_rate/qwen_vl_interface': 5.073406261836157e-06, 'learning_rate/action_model': 5.0734062618361565e-05, 'epoch': 0.19}) +[2026-10-03 15:07:40,859][starVLA.training.train_starvla][INFO] - Step 2609, Loss: {'train/loss': 0.011752987280488014, 'train/grad_norm_pre_clip': 55.26895523071289, 'timing/data': 0.0003077000001212582, 'timing/model': 0.3643673280021176, 'learning_rate/qwen_vl_interface': 5.070362982272386e-06, 'learning_rate/action_model': 5.070362982272385e-05, 'epoch': 0.19}) +[2026-10-03 15:07:41,507][starVLA.training.train_starvla][INFO] - Step 2610, Loss: {'train/loss': 0.0001629875914659351, 'train/grad_norm_pre_clip': 0.038705505430698395, 'timing/data': 0.00031493999995291233, 'timing/model': 0.36455590900732204, 'learning_rate/qwen_vl_interface': 5.067319776550607e-06, 'learning_rate/action_model': 5.067319776550607e-05, 'epoch': 0.19}) +[2026-10-03 15:07:42,135][starVLA.training.train_starvla][INFO] - Step 2611, Loss: {'train/loss': 0.40751001238822937, 'train/grad_norm_pre_clip': 69.66349792480469, 'timing/data': 0.0002565199974924326, 'timing/model': 0.3424905809952179, 'learning_rate/qwen_vl_interface': 5.064276645921768e-06, 'learning_rate/action_model': 5.064276645921768e-05, 'epoch': 0.19}) +[2026-10-03 15:07:42,761][starVLA.training.train_starvla][INFO] - Step 2612, Loss: {'train/loss': 0.00019617157522588968, 'train/grad_norm_pre_clip': 0.04127981513738632, 'timing/data': 0.00033040999551303685, 'timing/model': 0.3423249309998937, 'learning_rate/qwen_vl_interface': 5.061233591636787e-06, 'learning_rate/action_model': 5.061233591636787e-05, 'epoch': 0.19}) +[2026-10-03 15:07:43,600][starVLA.training.train_starvla][INFO] - Step 2613, Loss: {'train/loss': 0.0015165802324190736, 'train/grad_norm_pre_clip': 0.2991311252117157, 'timing/data': 0.0002540700079407543, 'timing/model': 0.5556810200068867, 'learning_rate/qwen_vl_interface': 5.058190614946544e-06, 'learning_rate/action_model': 5.058190614946544e-05, 'epoch': 0.19}) +[2026-10-03 15:07:44,248][starVLA.training.train_starvla][INFO] - Step 2614, Loss: {'train/loss': 0.4757012128829956, 'train/grad_norm_pre_clip': 33.890995025634766, 'timing/data': 0.00024924999161157757, 'timing/model': 0.36470688799454365, 'learning_rate/qwen_vl_interface': 5.055147717101895e-06, 'learning_rate/action_model': 5.055147717101894e-05, 'epoch': 0.19}) +[2026-10-03 15:07:44,899][starVLA.training.train_starvla][INFO] - Step 2615, Loss: {'train/loss': 0.0022379804868251085, 'train/grad_norm_pre_clip': 0.7644831538200378, 'timing/data': 0.0003197799960616976, 'timing/model': 0.36551764799514785, 'learning_rate/qwen_vl_interface': 5.052104899353658e-06, 'learning_rate/action_model': 5.0521048993536577e-05, 'epoch': 0.19}) +[2026-10-03 15:07:45,547][starVLA.training.train_starvla][INFO] - Step 2616, Loss: {'train/loss': 0.3635152578353882, 'train/grad_norm_pre_clip': 42.10692596435547, 'timing/data': 0.00024340000527445227, 'timing/model': 0.36476361700624693, 'learning_rate/qwen_vl_interface': 5.049062162952622e-06, 'learning_rate/action_model': 5.0490621629526214e-05, 'epoch': 0.19}) +[2026-10-03 15:07:46,194][starVLA.training.train_starvla][INFO] - Step 2617, Loss: {'train/loss': 0.0016144653782248497, 'train/grad_norm_pre_clip': 0.35745325684547424, 'timing/data': 0.000357939992682077, 'timing/model': 0.3643412170058582, 'learning_rate/qwen_vl_interface': 5.046019509149541e-06, 'learning_rate/action_model': 5.046019509149541e-05, 'epoch': 0.19}) +[2026-10-03 15:07:47,018][starVLA.training.train_starvla][INFO] - Step 2618, Loss: {'train/loss': 0.0002098659606417641, 'train/grad_norm_pre_clip': 0.05771060660481453, 'timing/data': 0.00030402999254874885, 'timing/model': 0.5632038299954729, 'learning_rate/qwen_vl_interface': 5.042976939195135e-06, 'learning_rate/action_model': 5.0429769391951355e-05, 'epoch': 0.19}) +[2026-10-03 15:07:47,669][starVLA.training.train_starvla][INFO] - Step 2619, Loss: {'train/loss': 0.0030701658688485622, 'train/grad_norm_pre_clip': 2.31854510307312, 'timing/data': 0.0002940400008810684, 'timing/model': 0.3657924680010183, 'learning_rate/qwen_vl_interface': 5.039934454340091e-06, 'learning_rate/action_model': 5.0399344543400907e-05, 'epoch': 0.19}) +[2026-10-03 15:07:48,317][starVLA.training.train_starvla][INFO] - Step 2620, Loss: {'train/loss': 0.00042231567203998566, 'train/grad_norm_pre_clip': 0.12344633787870407, 'timing/data': 0.00024027000472415239, 'timing/model': 0.3650776489957934, 'learning_rate/qwen_vl_interface': 5.036892055835056e-06, 'learning_rate/action_model': 5.0368920558350566e-05, 'epoch': 0.19}) +[2026-10-03 15:07:48,966][starVLA.training.train_starvla][INFO] - Step 2621, Loss: {'train/loss': 0.00043578469194471836, 'train/grad_norm_pre_clip': 0.13041649758815765, 'timing/data': 0.00023784999211784452, 'timing/model': 0.3651769480056828, 'learning_rate/qwen_vl_interface': 5.033849744930647e-06, 'learning_rate/action_model': 5.033849744930646e-05, 'epoch': 0.19}) +[2026-10-03 15:07:49,614][starVLA.training.train_starvla][INFO] - Step 2622, Loss: {'train/loss': 0.09937895834445953, 'train/grad_norm_pre_clip': 121.03645324707031, 'timing/data': 0.00030965999758336693, 'timing/model': 0.36477750798803754, 'learning_rate/qwen_vl_interface': 5.030807522877443e-06, 'learning_rate/action_model': 5.030807522877443e-05, 'epoch': 0.19}) +[2026-10-03 15:07:50,242][starVLA.training.train_starvla][INFO] - Step 2623, Loss: {'train/loss': 0.33268219232559204, 'train/grad_norm_pre_clip': 29.20595932006836, 'timing/data': 0.0002541499998187646, 'timing/model': 0.34268267200968694, 'learning_rate/qwen_vl_interface': 5.027765390925987e-06, 'learning_rate/action_model': 5.027765390925987e-05, 'epoch': 0.19}) +[2026-10-03 15:07:51,091][starVLA.training.train_starvla][INFO] - Step 2624, Loss: {'train/loss': 0.004082518629729748, 'train/grad_norm_pre_clip': 4.852646350860596, 'timing/data': 0.0002417700015939772, 'timing/model': 0.36491587800264824, 'learning_rate/qwen_vl_interface': 5.024723350326783e-06, 'learning_rate/action_model': 5.024723350326783e-05, 'epoch': 0.19}) +[2026-10-03 15:07:51,739][starVLA.training.train_starvla][INFO] - Step 2625, Loss: {'train/loss': 0.2395423799753189, 'train/grad_norm_pre_clip': 128.18421936035156, 'timing/data': 0.0002951700007542968, 'timing/model': 0.3648936380050145, 'learning_rate/qwen_vl_interface': 5.021681402330303e-06, 'learning_rate/action_model': 5.021681402330303e-05, 'epoch': 0.19}) +[2026-10-03 15:07:52,387][starVLA.training.train_starvla][INFO] - Step 2626, Loss: {'train/loss': 0.0007255522068589926, 'train/grad_norm_pre_clip': 0.39127519726753235, 'timing/data': 0.00030734000029042363, 'timing/model': 0.3651952879881719, 'learning_rate/qwen_vl_interface': 5.018639548186972e-06, 'learning_rate/action_model': 5.0186395481869716e-05, 'epoch': 0.19}) +[2026-10-03 15:07:53,038][starVLA.training.train_starvla][INFO] - Step 2627, Loss: {'train/loss': 0.7207974195480347, 'train/grad_norm_pre_clip': 16.106101989746094, 'timing/data': 0.00029306000215001404, 'timing/model': 0.36542022800131235, 'learning_rate/qwen_vl_interface': 5.015597789147184e-06, 'learning_rate/action_model': 5.015597789147184e-05, 'epoch': 0.19}) +[2026-10-03 15:07:53,686][starVLA.training.train_starvla][INFO] - Step 2628, Loss: {'train/loss': 0.6141852140426636, 'train/grad_norm_pre_clip': 29.779760360717773, 'timing/data': 0.0002401999954599887, 'timing/model': 0.3651407779980218, 'learning_rate/qwen_vl_interface': 5.012556126461291e-06, 'learning_rate/action_model': 5.012556126461291e-05, 'epoch': 0.19}) +[2026-10-03 15:07:54,530][starVLA.training.train_starvla][INFO] - Step 2629, Loss: {'train/loss': 0.0015709373401477933, 'train/grad_norm_pre_clip': 0.601771354675293, 'timing/data': 0.0001962200039997697, 'timing/model': 0.5598269210022409, 'learning_rate/qwen_vl_interface': 5.0095145613796044e-06, 'learning_rate/action_model': 5.0095145613796046e-05, 'epoch': 0.19}) +[2026-10-03 15:07:55,177][starVLA.training.train_starvla][INFO] - Step 2630, Loss: {'train/loss': 0.00053928152192384, 'train/grad_norm_pre_clip': 0.16643013060092926, 'timing/data': 0.00032388999534305185, 'timing/model': 0.36438988900044933, 'learning_rate/qwen_vl_interface': 5.0064730951524e-06, 'learning_rate/action_model': 5.0064730951524e-05, 'epoch': 0.19}) +[2026-10-03 15:07:55,827][starVLA.training.train_starvla][INFO] - Step 2631, Loss: {'train/loss': 0.00027833355125039816, 'train/grad_norm_pre_clip': 0.0797504261136055, 'timing/data': 0.00026292999973520637, 'timing/model': 0.3651631379907485, 'learning_rate/qwen_vl_interface': 5.003431729029905e-06, 'learning_rate/action_model': 5.003431729029905e-05, 'epoch': 0.19}) +[2026-10-03 15:07:56,474][starVLA.training.train_starvla][INFO] - Step 2632, Loss: {'train/loss': 0.39147868752479553, 'train/grad_norm_pre_clip': 56.40558624267578, 'timing/data': 0.0002481400006217882, 'timing/model': 0.36444272799417377, 'learning_rate/qwen_vl_interface': 5.000390464262314e-06, 'learning_rate/action_model': 5.0003904642623136e-05, 'epoch': 0.19}) +[2026-10-03 15:07:57,122][starVLA.training.train_starvla][INFO] - Step 2633, Loss: {'train/loss': 0.006062635686248541, 'train/grad_norm_pre_clip': 4.135229587554932, 'timing/data': 0.00024002000282052904, 'timing/model': 0.3644974979979452, 'learning_rate/qwen_vl_interface': 4.997349302099774e-06, 'learning_rate/action_model': 4.997349302099773e-05, 'epoch': 0.19}) +[2026-10-03 15:07:57,768][starVLA.training.train_starvla][INFO] - Step 2634, Loss: {'train/loss': 0.0013260925188660622, 'train/grad_norm_pre_clip': 0.6361791491508484, 'timing/data': 0.00024499000573996454, 'timing/model': 0.3642890290066134, 'learning_rate/qwen_vl_interface': 4.994308243792391e-06, 'learning_rate/action_model': 4.994308243792391e-05, 'epoch': 0.19}) +[2026-10-03 15:07:58,604][starVLA.training.train_starvla][INFO] - Step 2635, Loss: {'train/loss': 0.5365779995918274, 'train/grad_norm_pre_clip': 18.309911727905273, 'timing/data': 0.0002041200059466064, 'timing/model': 0.3653263580054045, 'learning_rate/qwen_vl_interface': 4.991267290590233e-06, 'learning_rate/action_model': 4.991267290590233e-05, 'epoch': 0.19}) +[2026-10-03 15:07:59,252][starVLA.training.train_starvla][INFO] - Step 2636, Loss: {'train/loss': 0.0005149483331479132, 'train/grad_norm_pre_clip': 0.1008690595626831, 'timing/data': 0.00029726000502705574, 'timing/model': 0.36468108800181653, 'learning_rate/qwen_vl_interface': 4.988226443743317e-06, 'learning_rate/action_model': 4.9882264437433174e-05, 'epoch': 0.19}) +[2026-10-03 15:07:59,900][starVLA.training.train_starvla][INFO] - Step 2637, Loss: {'train/loss': 0.0005161743611097336, 'train/grad_norm_pre_clip': 0.09569256752729416, 'timing/data': 0.0002506199962226674, 'timing/model': 0.3645381480018841, 'learning_rate/qwen_vl_interface': 4.985185704501623e-06, 'learning_rate/action_model': 4.985185704501623e-05, 'epoch': 0.19}) +[2026-10-03 15:08:00,547][starVLA.training.train_starvla][INFO] - Step 2638, Loss: {'train/loss': 0.0020570147316902876, 'train/grad_norm_pre_clip': 1.8684720993041992, 'timing/data': 0.0002501299895811826, 'timing/model': 0.3650050680007553, 'learning_rate/qwen_vl_interface': 4.982145074115082e-06, 'learning_rate/action_model': 4.982145074115082e-05, 'epoch': 0.19}) +[2026-10-03 15:08:01,198][starVLA.training.train_starvla][INFO] - Step 2639, Loss: {'train/loss': 0.002858372638002038, 'train/grad_norm_pre_clip': 4.043507099151611, 'timing/data': 0.00032681999437045306, 'timing/model': 0.36562111899547745, 'learning_rate/qwen_vl_interface': 4.979104553833584e-06, 'learning_rate/action_model': 4.979104553833584e-05, 'epoch': 0.19}) +[2026-10-03 15:08:02,035][starVLA.training.train_starvla][INFO] - Step 2640, Loss: {'train/loss': 0.000429179344791919, 'train/grad_norm_pre_clip': 0.0865241140127182, 'timing/data': 0.000205789998290129, 'timing/model': 0.3649948980018962, 'learning_rate/qwen_vl_interface': 4.9760641449069724e-06, 'learning_rate/action_model': 4.976064144906972e-05, 'epoch': 0.19}) +[2026-10-03 15:08:02,684][starVLA.training.train_starvla][INFO] - Step 2641, Loss: {'train/loss': 0.02252071350812912, 'train/grad_norm_pre_clip': 32.842124938964844, 'timing/data': 0.00019741000141948462, 'timing/model': 0.36508818800211884, 'learning_rate/qwen_vl_interface': 4.973023848585042e-06, 'learning_rate/action_model': 4.973023848585042e-05, 'epoch': 0.19}) +[2026-10-03 15:08:03,332][starVLA.training.train_starvla][INFO] - Step 2642, Loss: {'train/loss': 0.0003174354205839336, 'train/grad_norm_pre_clip': 0.053786516189575195, 'timing/data': 0.00019708000763785094, 'timing/model': 0.36501709900039714, 'learning_rate/qwen_vl_interface': 4.969983666117545e-06, 'learning_rate/action_model': 4.9699836661175446e-05, 'epoch': 0.19}) +[2026-10-03 15:08:03,960][starVLA.training.train_starvla][INFO] - Step 2643, Loss: {'train/loss': 0.0005199555889703333, 'train/grad_norm_pre_clip': 0.09781504422426224, 'timing/data': 0.0002745399979176, 'timing/model': 0.3429551119916141, 'learning_rate/qwen_vl_interface': 4.966943598754185e-06, 'learning_rate/action_model': 4.9669435987541854e-05, 'epoch': 0.19}) +[2026-10-03 15:08:04,609][starVLA.training.train_starvla][INFO] - Step 2644, Loss: {'train/loss': 0.0007439792971126735, 'train/grad_norm_pre_clip': 0.11709591001272202, 'timing/data': 0.0001960399968083948, 'timing/model': 0.3657772180013126, 'learning_rate/qwen_vl_interface': 4.963903647744621e-06, 'learning_rate/action_model': 4.96390364774462e-05, 'epoch': 0.19}) +[2026-10-03 15:08:05,453][starVLA.training.train_starvla][INFO] - Step 2645, Loss: {'train/loss': 0.00028954976005479693, 'train/grad_norm_pre_clip': 0.047582291066646576, 'timing/data': 0.00033491999784018844, 'timing/model': 0.3654035479994491, 'learning_rate/qwen_vl_interface': 4.9608638143384604e-06, 'learning_rate/action_model': 4.960863814338461e-05, 'epoch': 0.19}) +[2026-10-03 15:08:06,101][starVLA.training.train_starvla][INFO] - Step 2646, Loss: {'train/loss': 0.0012441194849088788, 'train/grad_norm_pre_clip': 0.5231923460960388, 'timing/data': 0.0002176100097130984, 'timing/model': 0.36467129799711984, 'learning_rate/qwen_vl_interface': 4.9578240997852636e-06, 'learning_rate/action_model': 4.957824099785264e-05, 'epoch': 0.19}) +[2026-10-03 15:08:06,751][starVLA.training.train_starvla][INFO] - Step 2647, Loss: {'train/loss': 0.028160635381937027, 'train/grad_norm_pre_clip': 49.528995513916016, 'timing/data': 0.00022077999892644584, 'timing/model': 0.3651541989966063, 'learning_rate/qwen_vl_interface': 4.954784505334544e-06, 'learning_rate/action_model': 4.9547845053345434e-05, 'epoch': 0.19}) +[2026-10-03 15:08:07,399][starVLA.training.train_starvla][INFO] - Step 2648, Loss: {'train/loss': 0.006620714906603098, 'train/grad_norm_pre_clip': 8.81078052520752, 'timing/data': 0.00031645999115426093, 'timing/model': 0.3651189990050625, 'learning_rate/qwen_vl_interface': 4.951745032235762e-06, 'learning_rate/action_model': 4.951745032235762e-05, 'epoch': 0.19}) +[2026-10-03 15:08:08,047][starVLA.training.train_starvla][INFO] - Step 2649, Loss: {'train/loss': 0.00041819358011707664, 'train/grad_norm_pre_clip': 0.06948202848434448, 'timing/data': 0.0003040200099349022, 'timing/model': 0.36448341800132766, 'learning_rate/qwen_vl_interface': 4.94870568173833e-06, 'learning_rate/action_model': 4.9487056817383307e-05, 'epoch': 0.19}) +[2026-10-03 15:08:08,888][starVLA.training.train_starvla][INFO] - Step 2650, Loss: {'train/loss': 0.006131310481578112, 'train/grad_norm_pre_clip': 7.168865203857422, 'timing/data': 0.00020638000569306314, 'timing/model': 0.5585354309878312, 'learning_rate/qwen_vl_interface': 4.945666455091614e-06, 'learning_rate/action_model': 4.9456664550916136e-05, 'epoch': 0.19}) +[2026-10-03 15:08:09,538][starVLA.training.train_starvla][INFO] - Step 2651, Loss: {'train/loss': 0.17858460545539856, 'train/grad_norm_pre_clip': 84.3000259399414, 'timing/data': 0.00036345000262372196, 'timing/model': 0.36553515800915193, 'learning_rate/qwen_vl_interface': 4.942627353544921e-06, 'learning_rate/action_model': 4.942627353544921e-05, 'epoch': 0.19}) +[2026-10-03 15:08:10,186][starVLA.training.train_starvla][INFO] - Step 2652, Loss: {'train/loss': 0.0020862710662186146, 'train/grad_norm_pre_clip': 1.6798871755599976, 'timing/data': 0.00025086000096052885, 'timing/model': 0.3649984189978568, 'learning_rate/qwen_vl_interface': 4.939588378347513e-06, 'learning_rate/action_model': 4.9395883783475125e-05, 'epoch': 0.19}) +[2026-10-03 15:08:10,835][starVLA.training.train_starvla][INFO] - Step 2653, Loss: {'train/loss': 0.00034848583163693547, 'train/grad_norm_pre_clip': 0.1835898458957672, 'timing/data': 0.0003335000074002892, 'timing/model': 0.3649674390035216, 'learning_rate/qwen_vl_interface': 4.936549530748597e-06, 'learning_rate/action_model': 4.9365495307485963e-05, 'epoch': 0.19}) +[2026-10-03 15:08:11,484][starVLA.training.train_starvla][INFO] - Step 2654, Loss: {'train/loss': 0.0006038059946149588, 'train/grad_norm_pre_clip': 0.16389256715774536, 'timing/data': 0.00032071000896394253, 'timing/model': 0.36489035800332204, 'learning_rate/qwen_vl_interface': 4.9335108119973286e-06, 'learning_rate/action_model': 4.933510811997328e-05, 'epoch': 0.19}) +[2026-10-03 15:08:12,133][starVLA.training.train_starvla][INFO] - Step 2655, Loss: {'train/loss': 0.0005616773851215839, 'train/grad_norm_pre_clip': 0.09883944690227509, 'timing/data': 0.00020173999655526131, 'timing/model': 0.36555125800077803, 'learning_rate/qwen_vl_interface': 4.930472223342812e-06, 'learning_rate/action_model': 4.930472223342812e-05, 'epoch': 0.19}) +[2026-10-03 15:08:12,968][starVLA.training.train_starvla][INFO] - Step 2656, Loss: {'train/loss': 0.0004271166690159589, 'train/grad_norm_pre_clip': 0.10401622205972672, 'timing/data': 0.00032807000388856977, 'timing/model': 0.36497623901232146, 'learning_rate/qwen_vl_interface': 4.927433766034093e-06, 'learning_rate/action_model': 4.9274337660340934e-05, 'epoch': 0.19}) +[2026-10-03 15:08:13,617][starVLA.training.train_starvla][INFO] - Step 2657, Loss: {'train/loss': 0.0004131317837163806, 'train/grad_norm_pre_clip': 0.07749896496534348, 'timing/data': 0.0002404890110483393, 'timing/model': 0.3650310490047559, 'learning_rate/qwen_vl_interface': 4.92439544132017e-06, 'learning_rate/action_model': 4.92439544132017e-05, 'epoch': 0.19}) +[2026-10-03 15:08:14,265][starVLA.training.train_starvla][INFO] - Step 2658, Loss: {'train/loss': 0.12955307960510254, 'train/grad_norm_pre_clip': 136.12191772460938, 'timing/data': 0.0002218899899162352, 'timing/model': 0.3650701680016937, 'learning_rate/qwen_vl_interface': 4.92135725044998e-06, 'learning_rate/action_model': 4.92135725044998e-05, 'epoch': 0.19}) +[2026-10-03 15:08:14,914][starVLA.training.train_starvla][INFO] - Step 2659, Loss: {'train/loss': 0.0003980739274993539, 'train/grad_norm_pre_clip': 0.07828214764595032, 'timing/data': 0.00019931999850086868, 'timing/model': 0.3652573580038734, 'learning_rate/qwen_vl_interface': 4.918319194672412e-06, 'learning_rate/action_model': 4.918319194672412e-05, 'epoch': 0.19}) +[2026-10-03 15:08:15,562][starVLA.training.train_starvla][INFO] - Step 2660, Loss: {'train/loss': 0.20652064681053162, 'train/grad_norm_pre_clip': 108.49588775634766, 'timing/data': 0.00024383999698329717, 'timing/model': 0.36506715799623635, 'learning_rate/qwen_vl_interface': 4.9152812752362935e-06, 'learning_rate/action_model': 4.915281275236293e-05, 'epoch': 0.19}) +[2026-10-03 15:08:16,380][starVLA.training.train_starvla][INFO] - Step 2661, Loss: {'train/loss': 0.0015699504874646664, 'train/grad_norm_pre_clip': 1.9539432525634766, 'timing/data': 0.00031799999123904854, 'timing/model': 0.34188601200003177, 'learning_rate/qwen_vl_interface': 4.9122434933904e-06, 'learning_rate/action_model': 4.9122434933904e-05, 'epoch': 0.19}) +[2026-10-03 15:08:17,028][starVLA.training.train_starvla][INFO] - Step 2662, Loss: {'train/loss': 0.0010000865440815687, 'train/grad_norm_pre_clip': 6.453472137451172, 'timing/data': 0.00019409999367780983, 'timing/model': 0.36505331900843885, 'learning_rate/qwen_vl_interface': 4.909205850383449e-06, 'learning_rate/action_model': 4.9092058503834485e-05, 'epoch': 0.19}) +[2026-10-03 15:08:17,678][starVLA.training.train_starvla][INFO] - Step 2663, Loss: {'train/loss': 0.0009868493070825934, 'train/grad_norm_pre_clip': 0.2506556808948517, 'timing/data': 0.0002673400013009086, 'timing/model': 0.3648698479955783, 'learning_rate/qwen_vl_interface': 4.906168347464098e-06, 'learning_rate/action_model': 4.906168347464098e-05, 'epoch': 0.19}) +[2026-10-03 15:08:18,325][starVLA.training.train_starvla][INFO] - Step 2664, Loss: {'train/loss': 0.0002898763632401824, 'train/grad_norm_pre_clip': 0.07177074253559113, 'timing/data': 0.0003264699917053804, 'timing/model': 0.36434033801197074, 'learning_rate/qwen_vl_interface': 4.903130985880955e-06, 'learning_rate/action_model': 4.903130985880954e-05, 'epoch': 0.19}) +[2026-10-03 15:08:18,974][starVLA.training.train_starvla][INFO] - Step 2665, Loss: {'train/loss': 0.5491518974304199, 'train/grad_norm_pre_clip': 125.92857360839844, 'timing/data': 0.00030110999068710953, 'timing/model': 0.3651876380026806, 'learning_rate/qwen_vl_interface': 4.900093766882558e-06, 'learning_rate/action_model': 4.900093766882558e-05, 'epoch': 0.2}) +[2026-10-03 15:08:19,826][starVLA.training.train_starvla][INFO] - Step 2666, Loss: {'train/loss': 0.5921549201011658, 'train/grad_norm_pre_clip': 30.119762420654297, 'timing/data': 0.0003301899996586144, 'timing/model': 0.3650039289932465, 'learning_rate/qwen_vl_interface': 4.897056691717397e-06, 'learning_rate/action_model': 4.8970566917173966e-05, 'epoch': 0.2}) +[2026-10-03 15:08:20,475][starVLA.training.train_starvla][INFO] - Step 2667, Loss: {'train/loss': 0.0006500339368358254, 'train/grad_norm_pre_clip': 1.4140297174453735, 'timing/data': 0.0002948900073533878, 'timing/model': 0.3649942180054495, 'learning_rate/qwen_vl_interface': 4.894019761633898e-06, 'learning_rate/action_model': 4.894019761633898e-05, 'epoch': 0.2}) +[2026-10-03 15:08:21,123][starVLA.training.train_starvla][INFO] - Step 2668, Loss: {'train/loss': 0.00019983778474852443, 'train/grad_norm_pre_clip': 0.04226738214492798, 'timing/data': 0.00031165999826043844, 'timing/model': 0.364992038012133, 'learning_rate/qwen_vl_interface': 4.8909829778804304e-06, 'learning_rate/action_model': 4.89098297788043e-05, 'epoch': 0.2}) +[2026-10-03 15:08:21,771][starVLA.training.train_starvla][INFO] - Step 2669, Loss: {'train/loss': 5.363432865124196e-05, 'train/grad_norm_pre_clip': 0.008964519947767258, 'timing/data': 0.0002797799970721826, 'timing/model': 0.3645562789897667, 'learning_rate/qwen_vl_interface': 4.887946341705298e-06, 'learning_rate/action_model': 4.887946341705297e-05, 'epoch': 0.2}) +[2026-10-03 15:08:22,418][starVLA.training.train_starvla][INFO] - Step 2670, Loss: {'train/loss': 0.47731977701187134, 'train/grad_norm_pre_clip': 17.681411743164062, 'timing/data': 0.00030015000083949417, 'timing/model': 0.36428882800100837, 'learning_rate/qwen_vl_interface': 4.8849098543567465e-06, 'learning_rate/action_model': 4.884909854356747e-05, 'epoch': 0.2}) +[2026-10-03 15:08:23,261][starVLA.training.train_starvla][INFO] - Step 2671, Loss: {'train/loss': 0.0003494929987937212, 'train/grad_norm_pre_clip': 0.13524846732616425, 'timing/data': 0.00020244000188540667, 'timing/model': 0.5585104409983614, 'learning_rate/qwen_vl_interface': 4.881873517082965e-06, 'learning_rate/action_model': 4.8818735170829655e-05, 'epoch': 0.2}) +[2026-10-03 15:08:23,909][starVLA.training.train_starvla][INFO] - Step 2672, Loss: {'train/loss': 0.0013891473645344377, 'train/grad_norm_pre_clip': 0.2814914882183075, 'timing/data': 0.00020120000408496708, 'timing/model': 0.36566731800849084, 'learning_rate/qwen_vl_interface': 4.878837331132076e-06, 'learning_rate/action_model': 4.8788373311320754e-05, 'epoch': 0.2}) +[2026-10-03 15:08:24,556][starVLA.training.train_starvla][INFO] - Step 2673, Loss: {'train/loss': 0.00016906904056668282, 'train/grad_norm_pre_clip': 0.052480049431324005, 'timing/data': 0.0001966499985428527, 'timing/model': 0.3646964489889797, 'learning_rate/qwen_vl_interface': 4.87580129775214e-06, 'learning_rate/action_model': 4.87580129775214e-05, 'epoch': 0.2}) +[2026-10-03 15:08:25,204][starVLA.training.train_starvla][INFO] - Step 2674, Loss: {'train/loss': 0.0001641125709284097, 'train/grad_norm_pre_clip': 0.08466072380542755, 'timing/data': 0.00025004999770317227, 'timing/model': 0.3646041080064606, 'learning_rate/qwen_vl_interface': 4.872765418191156e-06, 'learning_rate/action_model': 4.872765418191155e-05, 'epoch': 0.2}) +[2026-10-03 15:08:25,854][starVLA.training.train_starvla][INFO] - Step 2675, Loss: {'train/loss': 0.4732000231742859, 'train/grad_norm_pre_clip': 44.12490463256836, 'timing/data': 0.00020342000061646104, 'timing/model': 0.36497334799787495, 'learning_rate/qwen_vl_interface': 4.8697296936970575e-06, 'learning_rate/action_model': 4.869729693697057e-05, 'epoch': 0.2}) +[2026-10-03 15:08:26,502][starVLA.training.train_starvla][INFO] - Step 2676, Loss: {'train/loss': 0.4593724310398102, 'train/grad_norm_pre_clip': 56.989139556884766, 'timing/data': 0.0001963499962585047, 'timing/model': 0.36540012899786234, 'learning_rate/qwen_vl_interface': 4.8666941255177185e-06, 'learning_rate/action_model': 4.8666941255177185e-05, 'epoch': 0.2}) +[2026-10-03 15:08:27,337][starVLA.training.train_starvla][INFO] - Step 2677, Loss: {'train/loss': 0.00016507756663486362, 'train/grad_norm_pre_clip': 0.051610950380563736, 'timing/data': 0.00025647001166362315, 'timing/model': 0.36483905899513047, 'learning_rate/qwen_vl_interface': 4.863658714900945e-06, 'learning_rate/action_model': 4.863658714900945e-05, 'epoch': 0.2}) +[2026-10-03 15:08:27,984][starVLA.training.train_starvla][INFO] - Step 2678, Loss: {'train/loss': 0.0001384367496939376, 'train/grad_norm_pre_clip': 0.05640242621302605, 'timing/data': 0.00024994999694172293, 'timing/model': 0.3639168790105032, 'learning_rate/qwen_vl_interface': 4.860623463094483e-06, 'learning_rate/action_model': 4.860623463094483e-05, 'epoch': 0.2}) +[2026-10-03 15:08:28,634][starVLA.training.train_starvla][INFO] - Step 2679, Loss: {'train/loss': 0.001067065168172121, 'train/grad_norm_pre_clip': 0.4507145285606384, 'timing/data': 0.00020431999291758984, 'timing/model': 0.3646324779983843, 'learning_rate/qwen_vl_interface': 4.857588371346006e-06, 'learning_rate/action_model': 4.8575883713460057e-05, 'epoch': 0.2}) +[2026-10-03 15:08:29,282][starVLA.training.train_starvla][INFO] - Step 2680, Loss: {'train/loss': 0.000185824726941064, 'train/grad_norm_pre_clip': 0.0451374426484108, 'timing/data': 0.0002422499965177849, 'timing/model': 0.36493001801136415, 'learning_rate/qwen_vl_interface': 4.854553440903127e-06, 'learning_rate/action_model': 4.854553440903126e-05, 'epoch': 0.2}) +[2026-10-03 15:08:29,928][starVLA.training.train_starvla][INFO] - Step 2681, Loss: {'train/loss': 7.468772673746571e-05, 'train/grad_norm_pre_clip': 0.012278084643185139, 'timing/data': 0.00037058000452816486, 'timing/model': 0.3644023489905521, 'learning_rate/qwen_vl_interface': 4.85151867301339e-06, 'learning_rate/action_model': 4.85151867301339e-05, 'epoch': 0.2}) +[2026-10-03 15:08:30,767][starVLA.training.train_starvla][INFO] - Step 2682, Loss: {'train/loss': 0.40965625643730164, 'train/grad_norm_pre_clip': 21.995668411254883, 'timing/data': 0.00045983899326529354, 'timing/model': 0.3649632490123622, 'learning_rate/qwen_vl_interface': 4.848484068924274e-06, 'learning_rate/action_model': 4.848484068924275e-05, 'epoch': 0.2}) +[2026-10-03 15:08:31,417][starVLA.training.train_starvla][INFO] - Step 2683, Loss: {'train/loss': 0.0007698754779994488, 'train/grad_norm_pre_clip': 0.27718639373779297, 'timing/data': 0.00032526999711990356, 'timing/model': 0.3652844080061186, 'learning_rate/qwen_vl_interface': 4.845449629883193e-06, 'learning_rate/action_model': 4.8454496298831934e-05, 'epoch': 0.2}) +[2026-10-03 15:08:32,066][starVLA.training.train_starvla][INFO] - Step 2684, Loss: {'train/loss': 0.44117146730422974, 'train/grad_norm_pre_clip': 70.72428894042969, 'timing/data': 0.0001953799946932122, 'timing/model': 0.3650479789939709, 'learning_rate/qwen_vl_interface': 4.842415357137487e-06, 'learning_rate/action_model': 4.842415357137487e-05, 'epoch': 0.2}) +[2026-10-03 15:08:32,714][starVLA.training.train_starvla][INFO] - Step 2685, Loss: {'train/loss': 0.00023762152704875916, 'train/grad_norm_pre_clip': 0.06496112048625946, 'timing/data': 0.0003137800085823983, 'timing/model': 0.3649915390124079, 'learning_rate/qwen_vl_interface': 4.839381251934432e-06, 'learning_rate/action_model': 4.8393812519344316e-05, 'epoch': 0.2}) +[2026-10-03 15:08:33,363][starVLA.training.train_starvla][INFO] - Step 2686, Loss: {'train/loss': 0.00016814250557217747, 'train/grad_norm_pre_clip': 0.03559339419007301, 'timing/data': 0.00034656000207178295, 'timing/model': 0.3650207379978383, 'learning_rate/qwen_vl_interface': 4.836347315521233e-06, 'learning_rate/action_model': 4.836347315521233e-05, 'epoch': 0.2}) +[2026-10-03 15:08:34,178][starVLA.training.train_starvla][INFO] - Step 2687, Loss: {'train/loss': 0.00040692143375054, 'train/grad_norm_pre_clip': 0.09382446855306625, 'timing/data': 0.00030139999580569565, 'timing/model': 0.5525508819991956, 'learning_rate/qwen_vl_interface': 4.833313549145029e-06, 'learning_rate/action_model': 4.833313549145029e-05, 'epoch': 0.2}) +[2026-10-03 15:08:34,827][starVLA.training.train_starvla][INFO] - Step 2688, Loss: {'train/loss': 7.378305599559098e-05, 'train/grad_norm_pre_clip': 0.013115406036376953, 'timing/data': 0.00023549000616185367, 'timing/model': 0.36521364899817854, 'learning_rate/qwen_vl_interface': 4.830279954052884e-06, 'learning_rate/action_model': 4.8302799540528836e-05, 'epoch': 0.2}) +[2026-10-03 15:08:35,475][starVLA.training.train_starvla][INFO] - Step 2689, Loss: {'train/loss': 0.000458118156529963, 'train/grad_norm_pre_clip': 0.2771836519241333, 'timing/data': 0.00021098999422974885, 'timing/model': 0.3652603679947788, 'learning_rate/qwen_vl_interface': 4.827246531491797e-06, 'learning_rate/action_model': 4.827246531491797e-05, 'epoch': 0.2}) +[2026-10-03 15:08:36,123][starVLA.training.train_starvla][INFO] - Step 2690, Loss: {'train/loss': 0.2871064245700836, 'train/grad_norm_pre_clip': 28.497190475463867, 'timing/data': 0.00033270999847445637, 'timing/model': 0.36502116901101544, 'learning_rate/qwen_vl_interface': 4.824213282708691e-06, 'learning_rate/action_model': 4.824213282708691e-05, 'epoch': 0.2}) +[2026-10-03 15:08:36,774][starVLA.training.train_starvla][INFO] - Step 2691, Loss: {'train/loss': 0.26197177171707153, 'train/grad_norm_pre_clip': 41.343013763427734, 'timing/data': 0.00020542000129353255, 'timing/model': 0.3654581289883936, 'learning_rate/qwen_vl_interface': 4.821180208950422e-06, 'learning_rate/action_model': 4.821180208950421e-05, 'epoch': 0.2}) +[2026-10-03 15:08:37,618][starVLA.training.train_starvla][INFO] - Step 2692, Loss: {'train/loss': 0.3417188227176666, 'train/grad_norm_pre_clip': 22.861997604370117, 'timing/data': 0.00023397000040858984, 'timing/model': 0.5609880209958646, 'learning_rate/qwen_vl_interface': 4.81814731146377e-06, 'learning_rate/action_model': 4.8181473114637695e-05, 'epoch': 0.2}) +[2026-10-03 15:08:38,244][starVLA.training.train_starvla][INFO] - Step 2693, Loss: {'train/loss': 0.3102250099182129, 'train/grad_norm_pre_clip': 16.587636947631836, 'timing/data': 0.00033078000706154853, 'timing/model': 0.3653391779953381, 'learning_rate/qwen_vl_interface': 4.8151145914954465e-06, 'learning_rate/action_model': 4.815114591495446e-05, 'epoch': 0.2}) +[2026-10-03 15:08:38,892][starVLA.training.train_starvla][INFO] - Step 2694, Loss: {'train/loss': 0.00013977425987832248, 'train/grad_norm_pre_clip': 0.02441651001572609, 'timing/data': 0.00019460001203697175, 'timing/model': 0.36504411800706293, 'learning_rate/qwen_vl_interface': 4.812082050292088e-06, 'learning_rate/action_model': 4.8120820502920876e-05, 'epoch': 0.2}) +[2026-10-03 15:08:39,543][starVLA.training.train_starvla][INFO] - Step 2695, Loss: {'train/loss': 0.16204877197742462, 'train/grad_norm_pre_clip': 230.5441436767578, 'timing/data': 0.00034343999868724495, 'timing/model': 0.3651676680019591, 'learning_rate/qwen_vl_interface': 4.809049689100256e-06, 'learning_rate/action_model': 4.8090496891002555e-05, 'epoch': 0.2}) +[2026-10-03 15:08:40,192][starVLA.training.train_starvla][INFO] - Step 2696, Loss: {'train/loss': 0.00021643727086484432, 'train/grad_norm_pre_clip': 0.07098577171564102, 'timing/data': 0.00023229001089930534, 'timing/model': 0.36516657899483107, 'learning_rate/qwen_vl_interface': 4.80601750916644e-06, 'learning_rate/action_model': 4.8060175091664404e-05, 'epoch': 0.2}) +[2026-10-03 15:08:41,037][starVLA.training.train_starvla][INFO] - Step 2697, Loss: {'train/loss': 7.774040568619967e-05, 'train/grad_norm_pre_clip': 0.01454723346978426, 'timing/data': 0.00019432899716775864, 'timing/model': 0.5610121310019167, 'learning_rate/qwen_vl_interface': 4.802985511737057e-06, 'learning_rate/action_model': 4.802985511737057e-05, 'epoch': 0.2}) +[2026-10-03 15:08:41,686][starVLA.training.train_starvla][INFO] - Step 2698, Loss: {'train/loss': 0.0003318284288980067, 'train/grad_norm_pre_clip': 0.06428729742765427, 'timing/data': 0.00032623999868519604, 'timing/model': 0.3651469390024431, 'learning_rate/qwen_vl_interface': 4.799953698058444e-06, 'learning_rate/action_model': 4.7999536980584434e-05, 'epoch': 0.2}) +[2026-10-03 15:08:42,337][starVLA.training.train_starvla][INFO] - Step 2699, Loss: {'train/loss': 0.35399267077445984, 'train/grad_norm_pre_clip': 27.099178314208984, 'timing/data': 0.00032064999686554074, 'timing/model': 0.36538244799885433, 'learning_rate/qwen_vl_interface': 4.796922069376867e-06, 'learning_rate/action_model': 4.796922069376867e-05, 'epoch': 0.2}) +[2026-10-03 15:08:42,985][starVLA.training.train_starvla][INFO] - Step 2700, Loss: {'train/loss': 0.42824235558509827, 'train/grad_norm_pre_clip': 49.194976806640625, 'timing/data': 0.0002000800013775006, 'timing/model': 0.3652873789978912, 'learning_rate/qwen_vl_interface': 4.793890626938511e-06, 'learning_rate/action_model': 4.793890626938511e-05, 'epoch': 0.2}) +[2026-10-03 15:08:43,634][starVLA.training.train_starvla][INFO] - Step 2701, Loss: {'train/loss': 0.00020617959671653807, 'train/grad_norm_pre_clip': 0.09621680527925491, 'timing/data': 0.00024558999575674534, 'timing/model': 0.36530302799656056, 'learning_rate/qwen_vl_interface': 4.7908593719894914e-06, 'learning_rate/action_model': 4.790859371989491e-05, 'epoch': 0.2}) +[2026-10-03 15:08:44,281][starVLA.training.train_starvla][INFO] - Step 2702, Loss: {'train/loss': 0.00035283193574287, 'train/grad_norm_pre_clip': 0.07889285683631897, 'timing/data': 0.00024002000282052904, 'timing/model': 0.3644978989905212, 'learning_rate/qwen_vl_interface': 4.78782830577584e-06, 'learning_rate/action_model': 4.7878283057758396e-05, 'epoch': 0.2}) +[2026-10-03 15:08:45,096][starVLA.training.train_starvla][INFO] - Step 2703, Loss: {'train/loss': 0.0041734203696250916, 'train/grad_norm_pre_clip': 1.4321634769439697, 'timing/data': 0.00033165999047923833, 'timing/model': 0.3653180380060803, 'learning_rate/qwen_vl_interface': 4.784797429543513e-06, 'learning_rate/action_model': 4.784797429543513e-05, 'epoch': 0.2}) +[2026-10-03 15:08:45,745][starVLA.training.train_starvla][INFO] - Step 2704, Loss: {'train/loss': 0.00607200525701046, 'train/grad_norm_pre_clip': 3.410917043685913, 'timing/data': 0.0002555999963078648, 'timing/model': 0.3650499690120341, 'learning_rate/qwen_vl_interface': 4.781766744538393e-06, 'learning_rate/action_model': 4.781766744538393e-05, 'epoch': 0.2}) +[2026-10-03 15:08:46,393][starVLA.training.train_starvla][INFO] - Step 2705, Loss: {'train/loss': 0.00014930893667042255, 'train/grad_norm_pre_clip': 0.03881048411130905, 'timing/data': 0.00020441999367903918, 'timing/model': 0.36480019900773186, 'learning_rate/qwen_vl_interface': 4.778736252006277e-06, 'learning_rate/action_model': 4.7787362520062765e-05, 'epoch': 0.2}) +[2026-10-03 15:08:47,040][starVLA.training.train_starvla][INFO] - Step 2706, Loss: {'train/loss': 0.0001917381741805002, 'train/grad_norm_pre_clip': 0.05674923211336136, 'timing/data': 0.00019320000137668103, 'timing/model': 0.3645006580045447, 'learning_rate/qwen_vl_interface': 4.775705953192885e-06, 'learning_rate/action_model': 4.7757059531928856e-05, 'epoch': 0.2}) +[2026-10-03 15:08:47,689][starVLA.training.train_starvla][INFO] - Step 2707, Loss: {'train/loss': 0.00015294042532332242, 'train/grad_norm_pre_clip': 0.045625969767570496, 'timing/data': 0.00038598899845965207, 'timing/model': 0.3651410290040076, 'learning_rate/qwen_vl_interface': 4.7726758493438615e-06, 'learning_rate/action_model': 4.7726758493438613e-05, 'epoch': 0.2}) +[2026-10-03 15:08:48,533][starVLA.training.train_starvla][INFO] - Step 2708, Loss: {'train/loss': 0.0013294139644131064, 'train/grad_norm_pre_clip': 0.34011638164520264, 'timing/data': 0.00024104998738039285, 'timing/model': 0.36485192899999674, 'learning_rate/qwen_vl_interface': 4.769645941704766e-06, 'learning_rate/action_model': 4.769645941704766e-05, 'epoch': 0.2}) +[2026-10-03 15:08:49,181][starVLA.training.train_starvla][INFO] - Step 2709, Loss: {'train/loss': 0.00024097785353660583, 'train/grad_norm_pre_clip': 0.09308763593435287, 'timing/data': 0.0002582099987193942, 'timing/model': 0.3651922690041829, 'learning_rate/qwen_vl_interface': 4.766616231521081e-06, 'learning_rate/action_model': 4.7666162315210806e-05, 'epoch': 0.2}) +[2026-10-03 15:08:49,829][starVLA.training.train_starvla][INFO] - Step 2710, Loss: {'train/loss': 9.451900405110791e-05, 'train/grad_norm_pre_clip': 0.015864448621869087, 'timing/data': 0.0002748200058704242, 'timing/model': 0.36460109800100327, 'learning_rate/qwen_vl_interface': 4.763586720038202e-06, 'learning_rate/action_model': 4.7635867200382014e-05, 'epoch': 0.2}) +[2026-10-03 15:08:50,480][starVLA.training.train_starvla][INFO] - Step 2711, Loss: {'train/loss': 0.07970359921455383, 'train/grad_norm_pre_clip': 94.96695709228516, 'timing/data': 0.00030595000134781003, 'timing/model': 0.3654406889982056, 'learning_rate/qwen_vl_interface': 4.76055740850145e-06, 'learning_rate/action_model': 4.76055740850145e-05, 'epoch': 0.2}) +[2026-10-03 15:08:51,128][starVLA.training.train_starvla][INFO] - Step 2712, Loss: {'train/loss': 0.0009567224187776446, 'train/grad_norm_pre_clip': 0.26790937781333923, 'timing/data': 0.0002960999991046265, 'timing/model': 0.36466064798878506, 'learning_rate/qwen_vl_interface': 4.757528298156058e-06, 'learning_rate/action_model': 4.757528298156058e-05, 'epoch': 0.2}) +[2026-10-03 15:08:51,970][starVLA.training.train_starvla][INFO] - Step 2713, Loss: {'train/loss': 0.00047622149577364326, 'train/grad_norm_pre_clip': 0.13272570073604584, 'timing/data': 0.00027153000701218843, 'timing/model': 0.5591550920071313, 'learning_rate/qwen_vl_interface': 4.754499390247181e-06, 'learning_rate/action_model': 4.754499390247181e-05, 'epoch': 0.2}) +[2026-10-03 15:08:52,617][starVLA.training.train_starvla][INFO] - Step 2714, Loss: {'train/loss': 0.0006592384888790548, 'train/grad_norm_pre_clip': 0.209809809923172, 'timing/data': 0.00024049999774433672, 'timing/model': 0.36430412900517695, 'learning_rate/qwen_vl_interface': 4.751470686019887e-06, 'learning_rate/action_model': 4.751470686019888e-05, 'epoch': 0.2}) +[2026-10-03 15:08:53,266][starVLA.training.train_starvla][INFO] - Step 2715, Loss: {'train/loss': 0.004854983650147915, 'train/grad_norm_pre_clip': 2.466330051422119, 'timing/data': 0.00025845000345725566, 'timing/model': 0.36505995799961966, 'learning_rate/qwen_vl_interface': 4.7484421867191645e-06, 'learning_rate/action_model': 4.748442186719165e-05, 'epoch': 0.2}) +[2026-10-03 15:08:53,914][starVLA.training.train_starvla][INFO] - Step 2716, Loss: {'train/loss': 0.08247239887714386, 'train/grad_norm_pre_clip': 104.79305267333984, 'timing/data': 0.00024608000239823014, 'timing/model': 0.3648220780014526, 'learning_rate/qwen_vl_interface': 4.745413893589913e-06, 'learning_rate/action_model': 4.7454138935899125e-05, 'epoch': 0.2}) +[2026-10-03 15:08:54,562][starVLA.training.train_starvla][INFO] - Step 2717, Loss: {'train/loss': 0.00019371356756892055, 'train/grad_norm_pre_clip': 0.10872106999158859, 'timing/data': 0.00022604000696446747, 'timing/model': 0.36496491900470573, 'learning_rate/qwen_vl_interface': 4.7423858078769506e-06, 'learning_rate/action_model': 4.742385807876951e-05, 'epoch': 0.2}) +[2026-10-03 15:08:55,407][starVLA.training.train_starvla][INFO] - Step 2718, Loss: {'train/loss': 0.0025943529326468706, 'train/grad_norm_pre_clip': 0.8959831595420837, 'timing/data': 0.0004479300114326179, 'timing/model': 0.562414821004495, 'learning_rate/qwen_vl_interface': 4.739357930825008e-06, 'learning_rate/action_model': 4.739357930825008e-05, 'epoch': 0.2}) +[2026-10-03 15:08:56,057][starVLA.training.train_starvla][INFO] - Step 2719, Loss: {'train/loss': 0.00019226246513426304, 'train/grad_norm_pre_clip': 0.08319087326526642, 'timing/data': 0.00043988999095745385, 'timing/model': 0.3645431990007637, 'learning_rate/qwen_vl_interface': 4.736330263678731e-06, 'learning_rate/action_model': 4.736330263678731e-05, 'epoch': 0.2}) +[2026-10-03 15:08:56,705][starVLA.training.train_starvla][INFO] - Step 2720, Loss: {'train/loss': 5.94148768868763e-05, 'train/grad_norm_pre_clip': 0.01163024827837944, 'timing/data': 0.0002411900059087202, 'timing/model': 0.3650874080049107, 'learning_rate/qwen_vl_interface': 4.733302807682679e-06, 'learning_rate/action_model': 4.733302807682679e-05, 'epoch': 0.2}) +[2026-10-03 15:08:57,353][starVLA.training.train_starvla][INFO] - Step 2721, Loss: {'train/loss': 7.29317544028163e-05, 'train/grad_norm_pre_clip': 0.03304256498813629, 'timing/data': 0.0002410000015515834, 'timing/model': 0.36443024799518753, 'learning_rate/qwen_vl_interface': 4.7302755640813255e-06, 'learning_rate/action_model': 4.7302755640813256e-05, 'epoch': 0.2}) +[2026-10-03 15:08:58,000][starVLA.training.train_starvla][INFO] - Step 2722, Loss: {'train/loss': 0.0002887402370106429, 'train/grad_norm_pre_clip': 0.07343801856040955, 'timing/data': 0.00024166998628061265, 'timing/model': 0.36476344800030347, 'learning_rate/qwen_vl_interface': 4.727248534119058e-06, 'learning_rate/action_model': 4.7272485341190574e-05, 'epoch': 0.2}) +[2026-10-03 15:08:58,649][starVLA.training.train_starvla][INFO] - Step 2723, Loss: {'train/loss': 0.0024216482415795326, 'train/grad_norm_pre_clip': 0.48197141289711, 'timing/data': 0.0002892499906010926, 'timing/model': 0.3652363490109565, 'learning_rate/qwen_vl_interface': 4.72422171904017e-06, 'learning_rate/action_model': 4.7242217190401697e-05, 'epoch': 0.2}) +[2026-10-03 15:08:59,486][starVLA.training.train_starvla][INFO] - Step 2724, Loss: {'train/loss': 0.0005911615444347262, 'train/grad_norm_pre_clip': 0.43314242362976074, 'timing/data': 0.0003237799974158406, 'timing/model': 0.36513514899706934, 'learning_rate/qwen_vl_interface': 4.721195120088873e-06, 'learning_rate/action_model': 4.721195120088873e-05, 'epoch': 0.2}) +[2026-10-03 15:09:00,134][starVLA.training.train_starvla][INFO] - Step 2725, Loss: {'train/loss': 7.365408964687958e-05, 'train/grad_norm_pre_clip': 0.018672123551368713, 'timing/data': 0.00030991999665275216, 'timing/model': 0.3646726980077801, 'learning_rate/qwen_vl_interface': 4.718168738509288e-06, 'learning_rate/action_model': 4.718168738509288e-05, 'epoch': 0.2}) +[2026-10-03 15:09:00,781][starVLA.training.train_starvla][INFO] - Step 2726, Loss: {'train/loss': 5.455442442325875e-05, 'train/grad_norm_pre_clip': 0.01631913147866726, 'timing/data': 0.0003481599997030571, 'timing/model': 0.3646364390006056, 'learning_rate/qwen_vl_interface': 4.715142575545445e-06, 'learning_rate/action_model': 4.7151425755454454e-05, 'epoch': 0.2}) +[2026-10-03 15:09:01,431][starVLA.training.train_starvla][INFO] - Step 2727, Loss: {'train/loss': 4.233969593769871e-05, 'train/grad_norm_pre_clip': 0.01253924798220396, 'timing/data': 0.00020880000374745578, 'timing/model': 0.36514637900108937, 'learning_rate/qwen_vl_interface': 4.712116632441287e-06, 'learning_rate/action_model': 4.7121166324412866e-05, 'epoch': 0.2}) +[2026-10-03 15:09:02,079][starVLA.training.train_starvla][INFO] - Step 2728, Loss: {'train/loss': 0.0005085932789370418, 'train/grad_norm_pre_clip': 0.28797656297683716, 'timing/data': 0.0003293699992354959, 'timing/model': 0.3645879789983155, 'learning_rate/qwen_vl_interface': 4.709090910440664e-06, 'learning_rate/action_model': 4.7090909104406634e-05, 'epoch': 0.2}) +[2026-10-03 15:09:02,929][starVLA.training.train_starvla][INFO] - Step 2729, Loss: {'train/loss': 0.00019174665794707835, 'train/grad_norm_pre_clip': 0.11400190740823746, 'timing/data': 0.00032792999991215765, 'timing/model': 0.36482459800026845, 'learning_rate/qwen_vl_interface': 4.7060654107873335e-06, 'learning_rate/action_model': 4.7060654107873335e-05, 'epoch': 0.2}) +[2026-10-03 15:09:03,575][starVLA.training.train_starvla][INFO] - Step 2730, Loss: {'train/loss': 0.2370641827583313, 'train/grad_norm_pre_clip': 112.81362915039062, 'timing/data': 0.0002392799942754209, 'timing/model': 0.3637879489979241, 'learning_rate/qwen_vl_interface': 4.7030401347249685e-06, 'learning_rate/action_model': 4.703040134724968e-05, 'epoch': 0.2}) +[2026-10-03 15:09:04,225][starVLA.training.train_starvla][INFO] - Step 2731, Loss: {'train/loss': 3.1133702577790245e-05, 'train/grad_norm_pre_clip': 0.007517607882618904, 'timing/data': 0.00024619999749120325, 'timing/model': 0.36576497899659444, 'learning_rate/qwen_vl_interface': 4.700015083497143e-06, 'learning_rate/action_model': 4.700015083497142e-05, 'epoch': 0.2}) +[2026-10-03 15:09:04,872][starVLA.training.train_starvla][INFO] - Step 2732, Loss: {'train/loss': 2.0369123376440257e-05, 'train/grad_norm_pre_clip': 0.00505044125020504, 'timing/data': 0.00034721000702120364, 'timing/model': 0.3645636290020775, 'learning_rate/qwen_vl_interface': 4.696990258347342e-06, 'learning_rate/action_model': 4.696990258347342e-05, 'epoch': 0.2}) +[2026-10-03 15:09:05,520][starVLA.training.train_starvla][INFO] - Step 2733, Loss: {'train/loss': 0.01180039532482624, 'train/grad_norm_pre_clip': 18.505767822265625, 'timing/data': 0.0002740199997788295, 'timing/model': 0.36521645900211297, 'learning_rate/qwen_vl_interface': 4.693965660518957e-06, 'learning_rate/action_model': 4.693965660518957e-05, 'epoch': 0.2}) +[2026-10-03 15:09:06,355][starVLA.training.train_starvla][INFO] - Step 2734, Loss: {'train/loss': 5.3097945055924356e-05, 'train/grad_norm_pre_clip': 0.015465249307453632, 'timing/data': 0.00032939000811893493, 'timing/model': 0.5518911419931101, 'learning_rate/qwen_vl_interface': 4.690941291255285e-06, 'learning_rate/action_model': 4.6909412912552844e-05, 'epoch': 0.2}) +[2026-10-03 15:09:07,005][starVLA.training.train_starvla][INFO] - Step 2735, Loss: {'train/loss': 0.0005343856173567474, 'train/grad_norm_pre_clip': 0.10605882108211517, 'timing/data': 0.0002482699928805232, 'timing/model': 0.3655818690021988, 'learning_rate/qwen_vl_interface': 4.68791715179953e-06, 'learning_rate/action_model': 4.68791715179953e-05, 'epoch': 0.2}) +[2026-10-03 15:09:07,653][starVLA.training.train_starvla][INFO] - Step 2736, Loss: {'train/loss': 0.08956514298915863, 'train/grad_norm_pre_clip': 183.2653350830078, 'timing/data': 0.0003289399901404977, 'timing/model': 0.36451941799896304, 'learning_rate/qwen_vl_interface': 4.684893243394804e-06, 'learning_rate/action_model': 4.6848932433948036e-05, 'epoch': 0.2}) +[2026-10-03 15:09:08,302][starVLA.training.train_starvla][INFO] - Step 2737, Loss: {'train/loss': 4.0124978113453835e-05, 'train/grad_norm_pre_clip': 0.016227953135967255, 'timing/data': 0.00030595000134781003, 'timing/model': 0.36476992900134064, 'learning_rate/qwen_vl_interface': 4.68186956728412e-06, 'learning_rate/action_model': 4.68186956728412e-05, 'epoch': 0.2}) +[2026-10-03 15:09:08,950][starVLA.training.train_starvla][INFO] - Step 2738, Loss: {'train/loss': 4.3604046368272975e-05, 'train/grad_norm_pre_clip': 0.020213676616549492, 'timing/data': 0.0002293599973199889, 'timing/model': 0.3646579789929092, 'learning_rate/qwen_vl_interface': 4.678846124710395e-06, 'learning_rate/action_model': 4.678846124710394e-05, 'epoch': 0.2}) +[2026-10-03 15:09:09,600][starVLA.training.train_starvla][INFO] - Step 2739, Loss: {'train/loss': 0.134684756398201, 'train/grad_norm_pre_clip': 366.6327819824219, 'timing/data': 0.0003224300016881898, 'timing/model': 0.3655354289949173, 'learning_rate/qwen_vl_interface': 4.675822916916454e-06, 'learning_rate/action_model': 4.6758229169164536e-05, 'epoch': 0.2}) +[2026-10-03 15:09:10,412][starVLA.training.train_starvla][INFO] - Step 2740, Loss: {'train/loss': 0.00014520258991979063, 'train/grad_norm_pre_clip': 0.2586909830570221, 'timing/data': 0.0003362499992363155, 'timing/model': 0.36435243900632486, 'learning_rate/qwen_vl_interface': 4.6727999451450236e-06, 'learning_rate/action_model': 4.672799945145023e-05, 'epoch': 0.2}) +[2026-10-03 15:09:11,060][starVLA.training.train_starvla][INFO] - Step 2741, Loss: {'train/loss': 1.1808967428805772e-05, 'train/grad_norm_pre_clip': 0.003911760635674, 'timing/data': 0.00033526000333949924, 'timing/model': 0.3648477090027882, 'learning_rate/qwen_vl_interface': 4.669777210638733e-06, 'learning_rate/action_model': 4.669777210638733e-05, 'epoch': 0.2}) +[2026-10-03 15:09:11,708][starVLA.training.train_starvla][INFO] - Step 2742, Loss: {'train/loss': 2.2029897081665695e-05, 'train/grad_norm_pre_clip': 0.008957060053944588, 'timing/data': 0.0002013199991779402, 'timing/model': 0.3649603689991636, 'learning_rate/qwen_vl_interface': 4.666754714640115e-06, 'learning_rate/action_model': 4.666754714640115e-05, 'epoch': 0.2}) +[2026-10-03 15:09:12,359][starVLA.training.train_starvla][INFO] - Step 2743, Loss: {'train/loss': 2.2879783500684425e-05, 'train/grad_norm_pre_clip': 0.006144796498119831, 'timing/data': 0.00022531001013703644, 'timing/model': 0.36578576800820883, 'learning_rate/qwen_vl_interface': 4.663732458391602e-06, 'learning_rate/action_model': 4.663732458391601e-05, 'epoch': 0.2}) +[2026-10-03 15:09:13,008][starVLA.training.train_starvla][INFO] - Step 2744, Loss: {'train/loss': 0.3400644361972809, 'train/grad_norm_pre_clip': 10.912497520446777, 'timing/data': 0.0003224200045224279, 'timing/model': 0.3650204779987689, 'learning_rate/qwen_vl_interface': 4.6607104431355295e-06, 'learning_rate/action_model': 4.66071044313553e-05, 'epoch': 0.2}) +[2026-10-03 15:09:13,847][starVLA.training.train_starvla][INFO] - Step 2745, Loss: {'train/loss': 0.18583643436431885, 'train/grad_norm_pre_clip': 183.862060546875, 'timing/data': 0.0003258600045228377, 'timing/model': 0.3651167589996476, 'learning_rate/qwen_vl_interface': 4.657688670114135e-06, 'learning_rate/action_model': 4.657688670114135e-05, 'epoch': 0.2}) +[2026-10-03 15:09:14,495][starVLA.training.train_starvla][INFO] - Step 2746, Loss: {'train/loss': 0.00011685895879054442, 'train/grad_norm_pre_clip': 0.07757347077131271, 'timing/data': 0.0003218500060029328, 'timing/model': 0.3646535689913435, 'learning_rate/qwen_vl_interface': 4.654667140569555e-06, 'learning_rate/action_model': 4.6546671405695544e-05, 'epoch': 0.2}) +[2026-10-03 15:09:15,124][starVLA.training.train_starvla][INFO] - Step 2747, Loss: {'train/loss': 0.0018827521707862616, 'train/grad_norm_pre_clip': 6.787985324859619, 'timing/data': 0.000302270011161454, 'timing/model': 0.36590025800978765, 'learning_rate/qwen_vl_interface': 4.651645855743825e-06, 'learning_rate/action_model': 4.651645855743825e-05, 'epoch': 0.2}) +[2026-10-03 15:09:15,772][starVLA.training.train_starvla][INFO] - Step 2748, Loss: {'train/loss': 0.44367218017578125, 'train/grad_norm_pre_clip': 32.38632583618164, 'timing/data': 0.00032266999187413603, 'timing/model': 0.36473260998900514, 'learning_rate/qwen_vl_interface': 4.648624816878883e-06, 'learning_rate/action_model': 4.6486248168788825e-05, 'epoch': 0.2}) +[2026-10-03 15:09:16,421][starVLA.training.train_starvla][INFO] - Step 2749, Loss: {'train/loss': 0.00013858806050848216, 'train/grad_norm_pre_clip': 0.07179775834083557, 'timing/data': 0.00024095000117085874, 'timing/model': 0.36527410900453106, 'learning_rate/qwen_vl_interface': 4.6456040252165615e-06, 'learning_rate/action_model': 4.645604025216562e-05, 'epoch': 0.2}) +[2026-10-03 15:09:17,245][starVLA.training.train_starvla][INFO] - Step 2750, Loss: {'train/loss': 0.21049864590168, 'train/grad_norm_pre_clip': 97.21041870117188, 'timing/data': 0.00020037899957969785, 'timing/model': 0.34209006300079636, 'learning_rate/qwen_vl_interface': 4.642583481998596e-06, 'learning_rate/action_model': 4.642583481998596e-05, 'epoch': 0.2}) +[2026-10-03 15:09:17,896][starVLA.training.train_starvla][INFO] - Step 2751, Loss: {'train/loss': 0.00024081776791717857, 'train/grad_norm_pre_clip': 0.24252252280712128, 'timing/data': 0.00031912000849843025, 'timing/model': 0.3654042090056464, 'learning_rate/qwen_vl_interface': 4.639563188466616e-06, 'learning_rate/action_model': 4.6395631884666165e-05, 'epoch': 0.2}) +[2026-10-03 15:09:18,544][starVLA.training.train_starvla][INFO] - Step 2752, Loss: {'train/loss': 0.5311660170555115, 'train/grad_norm_pre_clip': 38.90137481689453, 'timing/data': 0.00033031000930350274, 'timing/model': 0.3650352780096, 'learning_rate/qwen_vl_interface': 4.636543145862152e-06, 'learning_rate/action_model': 4.6365431458621523e-05, 'epoch': 0.2}) +[2026-10-03 15:09:19,192][starVLA.training.train_starvla][INFO] - Step 2753, Loss: {'train/loss': 0.8520645499229431, 'train/grad_norm_pre_clip': 26.410078048706055, 'timing/data': 0.0003162600041832775, 'timing/model': 0.36478396799066104, 'learning_rate/qwen_vl_interface': 4.633523355426629e-06, 'learning_rate/action_model': 4.6335233554266295e-05, 'epoch': 0.2}) +[2026-10-03 15:09:19,841][starVLA.training.train_starvla][INFO] - Step 2754, Loss: {'train/loss': 0.025721147656440735, 'train/grad_norm_pre_clip': 186.54164123535156, 'timing/data': 0.0002764100063359365, 'timing/model': 0.3652479789889185, 'learning_rate/qwen_vl_interface': 4.6305038184013675e-06, 'learning_rate/action_model': 4.6305038184013675e-05, 'epoch': 0.2}) +[2026-10-03 15:09:20,664][starVLA.training.train_starvla][INFO] - Step 2755, Loss: {'train/loss': 0.014844496734440327, 'train/grad_norm_pre_clip': 17.330469131469727, 'timing/data': 0.0004550099984044209, 'timing/model': 0.5385778139898321, 'learning_rate/qwen_vl_interface': 4.627484536027589e-06, 'learning_rate/action_model': 4.627484536027588e-05, 'epoch': 0.2}) +[2026-10-03 15:09:21,313][starVLA.training.train_starvla][INFO] - Step 2756, Loss: {'train/loss': 0.00012244748359080404, 'train/grad_norm_pre_clip': 0.04388832300901413, 'timing/data': 0.00019612000323832035, 'timing/model': 0.3649868479988072, 'learning_rate/qwen_vl_interface': 4.624465509546399e-06, 'learning_rate/action_model': 4.6244655095463986e-05, 'epoch': 0.2}) +[2026-10-03 15:09:21,939][starVLA.training.train_starvla][INFO] - Step 2757, Loss: {'train/loss': 6.380565173458308e-05, 'train/grad_norm_pre_clip': 0.015268055722117424, 'timing/data': 0.00019265999435447156, 'timing/model': 0.36495149800612126, 'learning_rate/qwen_vl_interface': 4.6214467401988106e-06, 'learning_rate/action_model': 4.621446740198811e-05, 'epoch': 0.2}) +[2026-10-03 15:09:22,587][starVLA.training.train_starvla][INFO] - Step 2758, Loss: {'train/loss': 0.0001697847037576139, 'train/grad_norm_pre_clip': 0.3442437946796417, 'timing/data': 0.0001977299980353564, 'timing/model': 0.36483314899669494, 'learning_rate/qwen_vl_interface': 4.6184282292257245e-06, 'learning_rate/action_model': 4.618428229225725e-05, 'epoch': 0.2}) +[2026-10-03 15:09:23,236][starVLA.training.train_starvla][INFO] - Step 2759, Loss: {'train/loss': 4.1063591197598726e-05, 'train/grad_norm_pre_clip': 0.011635556817054749, 'timing/data': 0.00021630000264849514, 'timing/model': 0.3645632090047002, 'learning_rate/qwen_vl_interface': 4.615409977867938e-06, 'learning_rate/action_model': 4.615409977867937e-05, 'epoch': 0.2}) +[2026-10-03 15:09:23,884][starVLA.training.train_starvla][INFO] - Step 2760, Loss: {'train/loss': 0.00010946957627311349, 'train/grad_norm_pre_clip': 0.039604995399713516, 'timing/data': 0.00025052999262697995, 'timing/model': 0.3652069989911979, 'learning_rate/qwen_vl_interface': 4.612391987366138e-06, 'learning_rate/action_model': 4.612391987366138e-05, 'epoch': 0.2}) +[2026-10-03 15:09:24,726][starVLA.training.train_starvla][INFO] - Step 2761, Loss: {'train/loss': 7.467174873454496e-05, 'train/grad_norm_pre_clip': 0.03602223098278046, 'timing/data': 0.0003511700051603839, 'timing/model': 0.3646076489967527, 'learning_rate/qwen_vl_interface': 4.609374258960908e-06, 'learning_rate/action_model': 4.6093742589609074e-05, 'epoch': 0.2}) +[2026-10-03 15:09:25,372][starVLA.training.train_starvla][INFO] - Step 2762, Loss: {'train/loss': 0.00034124244120903313, 'train/grad_norm_pre_clip': 0.24797841906547546, 'timing/data': 0.00034062999475281686, 'timing/model': 0.36398419000033755, 'learning_rate/qwen_vl_interface': 4.60635679389272e-06, 'learning_rate/action_model': 4.60635679389272e-05, 'epoch': 0.2}) +[2026-10-03 15:09:26,021][starVLA.training.train_starvla][INFO] - Step 2763, Loss: {'train/loss': 0.00020451155432965606, 'train/grad_norm_pre_clip': 0.07095997780561447, 'timing/data': 0.000325919987517409, 'timing/model': 0.36494300900085364, 'learning_rate/qwen_vl_interface': 4.603339593401943e-06, 'learning_rate/action_model': 4.603339593401942e-05, 'epoch': 0.2}) +[2026-10-03 15:09:26,670][starVLA.training.train_starvla][INFO] - Step 2764, Loss: {'train/loss': 0.6312095522880554, 'train/grad_norm_pre_clip': 28.690889358520508, 'timing/data': 0.0003307399892946705, 'timing/model': 0.3655591389979236, 'learning_rate/qwen_vl_interface': 4.600322658728832e-06, 'learning_rate/action_model': 4.600322658728832e-05, 'epoch': 0.2}) +[2026-10-03 15:09:27,318][starVLA.training.train_starvla][INFO] - Step 2765, Loss: {'train/loss': 0.0005879492382518947, 'train/grad_norm_pre_clip': 0.840663492679596, 'timing/data': 0.0002443699922878295, 'timing/model': 0.36466968899185304, 'learning_rate/qwen_vl_interface': 4.597305991113537e-06, 'learning_rate/action_model': 4.597305991113537e-05, 'epoch': 0.2}) +[2026-10-03 15:09:28,158][starVLA.training.train_starvla][INFO] - Step 2766, Loss: {'train/loss': 0.0013473256258293986, 'train/grad_norm_pre_clip': 0.4347771406173706, 'timing/data': 0.000338979996740818, 'timing/model': 0.3646824989991728, 'learning_rate/qwen_vl_interface': 4.5942895917960935e-06, 'learning_rate/action_model': 4.5942895917960935e-05, 'epoch': 0.2}) +[2026-10-03 15:09:28,808][starVLA.training.train_starvla][INFO] - Step 2767, Loss: {'train/loss': 0.00027011800557374954, 'train/grad_norm_pre_clip': 0.24394699931144714, 'timing/data': 0.00025722999998833984, 'timing/model': 0.36512958898674697, 'learning_rate/qwen_vl_interface': 4.59127346201643e-06, 'learning_rate/action_model': 4.59127346201643e-05, 'epoch': 0.2}) +[2026-10-03 15:09:29,456][starVLA.training.train_starvla][INFO] - Step 2768, Loss: {'train/loss': 0.00010539322101976722, 'train/grad_norm_pre_clip': 0.03874517232179642, 'timing/data': 0.0002503100113244727, 'timing/model': 0.3646774689987069, 'learning_rate/qwen_vl_interface': 4.588257603014364e-06, 'learning_rate/action_model': 4.588257603014364e-05, 'epoch': 0.2}) +[2026-10-03 15:09:30,082][starVLA.training.train_starvla][INFO] - Step 2769, Loss: {'train/loss': 0.0004887878312729299, 'train/grad_norm_pre_clip': 0.3286067545413971, 'timing/data': 0.00033907000033650547, 'timing/model': 0.3647114889899967, 'learning_rate/qwen_vl_interface': 4.585242016029602e-06, 'learning_rate/action_model': 4.5852420160296015e-05, 'epoch': 0.2}) +[2026-10-03 15:09:30,729][starVLA.training.train_starvla][INFO] - Step 2770, Loss: {'train/loss': 0.4423750936985016, 'train/grad_norm_pre_clip': 62.225242614746094, 'timing/data': 0.0003197799960616976, 'timing/model': 0.3645989689975977, 'learning_rate/qwen_vl_interface': 4.582226702301737e-06, 'learning_rate/action_model': 4.5822267023017375e-05, 'epoch': 0.2}) +[2026-10-03 15:09:31,574][starVLA.training.train_starvla][INFO] - Step 2771, Loss: {'train/loss': 2.9272636311361566e-05, 'train/grad_norm_pre_clip': 0.00540492357686162, 'timing/data': 0.00032871999428607523, 'timing/model': 0.5587413119937992, 'learning_rate/qwen_vl_interface': 4.57921166307025e-06, 'learning_rate/action_model': 4.57921166307025e-05, 'epoch': 0.2}) +[2026-10-03 15:09:32,223][starVLA.training.train_starvla][INFO] - Step 2772, Loss: {'train/loss': 9.67765663517639e-05, 'train/grad_norm_pre_clip': 0.025823155418038368, 'timing/data': 0.00025872999685816467, 'timing/model': 0.36554551898734644, 'learning_rate/qwen_vl_interface': 4.576196899574511e-06, 'learning_rate/action_model': 4.576196899574511e-05, 'epoch': 0.2}) +[2026-10-03 15:09:32,871][starVLA.training.train_starvla][INFO] - Step 2773, Loss: {'train/loss': 0.2705962657928467, 'train/grad_norm_pre_clip': 13.072038650512695, 'timing/data': 0.00034294999204576015, 'timing/model': 0.3650028590054717, 'learning_rate/qwen_vl_interface': 4.573182413053775e-06, 'learning_rate/action_model': 4.5731824130537746e-05, 'epoch': 0.2}) +[2026-10-03 15:09:33,518][starVLA.training.train_starvla][INFO] - Step 2774, Loss: {'train/loss': 0.0007962436648085713, 'train/grad_norm_pre_clip': 0.3969274163246155, 'timing/data': 0.00025456000003032386, 'timing/model': 0.36444490899157245, 'learning_rate/qwen_vl_interface': 4.570168204747183e-06, 'learning_rate/action_model': 4.570168204747183e-05, 'epoch': 0.2}) +[2026-10-03 15:09:34,168][starVLA.training.train_starvla][INFO] - Step 2775, Loss: {'train/loss': 0.3736124038696289, 'train/grad_norm_pre_clip': 95.71817016601562, 'timing/data': 0.00033915000676643103, 'timing/model': 0.3652359490079107, 'learning_rate/qwen_vl_interface': 4.567154275893762e-06, 'learning_rate/action_model': 4.5671542758937624e-05, 'epoch': 0.2}) +[2026-10-03 15:09:35,001][starVLA.training.train_starvla][INFO] - Step 2776, Loss: {'train/loss': 0.6474484205245972, 'train/grad_norm_pre_clip': 27.716915130615234, 'timing/data': 0.00019719998817890882, 'timing/model': 0.5490466729970649, 'learning_rate/qwen_vl_interface': 4.564140627732426e-06, 'learning_rate/action_model': 4.564140627732426e-05, 'epoch': 0.2}) +[2026-10-03 15:09:35,650][starVLA.training.train_starvla][INFO] - Step 2777, Loss: {'train/loss': 0.2679847776889801, 'train/grad_norm_pre_clip': 40.580806732177734, 'timing/data': 0.00019762000010814518, 'timing/model': 0.3657938579999609, 'learning_rate/qwen_vl_interface': 4.56112726150197e-06, 'learning_rate/action_model': 4.56112726150197e-05, 'epoch': 0.2}) +[2026-10-03 15:09:36,298][starVLA.training.train_starvla][INFO] - Step 2778, Loss: {'train/loss': 0.00010636254592100158, 'train/grad_norm_pre_clip': 0.022897038608789444, 'timing/data': 0.0003170400013914332, 'timing/model': 0.36484057900088374, 'learning_rate/qwen_vl_interface': 4.558114178441076e-06, 'learning_rate/action_model': 4.558114178441076e-05, 'epoch': 0.2}) +[2026-10-03 15:09:36,949][starVLA.training.train_starvla][INFO] - Step 2779, Loss: {'train/loss': 0.1443023383617401, 'train/grad_norm_pre_clip': 153.06529235839844, 'timing/data': 0.00024916000256780535, 'timing/model': 0.36549011900206096, 'learning_rate/qwen_vl_interface': 4.555101379788308e-06, 'learning_rate/action_model': 4.555101379788308e-05, 'epoch': 0.2}) +[2026-10-03 15:09:37,599][starVLA.training.train_starvla][INFO] - Step 2780, Loss: {'train/loss': 0.33711278438568115, 'train/grad_norm_pre_clip': 40.874855041503906, 'timing/data': 0.000331729999743402, 'timing/model': 0.3666295590082882, 'learning_rate/qwen_vl_interface': 4.552088866782115e-06, 'learning_rate/action_model': 4.552088866782115e-05, 'epoch': 0.2}) +[2026-10-03 15:09:38,249][starVLA.training.train_starvla][INFO] - Step 2781, Loss: {'train/loss': 0.0007251162896864116, 'train/grad_norm_pre_clip': 0.23082490265369415, 'timing/data': 0.00030048000917304307, 'timing/model': 0.36536581900145393, 'learning_rate/qwen_vl_interface': 4.5490766406608265e-06, 'learning_rate/action_model': 4.549076640660827e-05, 'epoch': 0.2}) +[2026-10-03 15:09:39,092][starVLA.training.train_starvla][INFO] - Step 2782, Loss: {'train/loss': 0.21261267364025116, 'train/grad_norm_pre_clip': 41.46632766723633, 'timing/data': 0.0001992799952859059, 'timing/model': 0.3652405790053308, 'learning_rate/qwen_vl_interface': 4.546064702662655e-06, 'learning_rate/action_model': 4.546064702662655e-05, 'epoch': 0.2}) +[2026-10-03 15:09:39,743][starVLA.training.train_starvla][INFO] - Step 2783, Loss: {'train/loss': 0.19059565663337708, 'train/grad_norm_pre_clip': 197.81179809570312, 'timing/data': 0.00033429000177420676, 'timing/model': 0.36575165900285356, 'learning_rate/qwen_vl_interface': 4.543053054025693e-06, 'learning_rate/action_model': 4.5430530540256926e-05, 'epoch': 0.2}) +[2026-10-03 15:09:40,391][starVLA.training.train_starvla][INFO] - Step 2784, Loss: {'train/loss': 0.00011314554285490885, 'train/grad_norm_pre_clip': 0.04539162293076515, 'timing/data': 0.0003315289941383526, 'timing/model': 0.36530691900406964, 'learning_rate/qwen_vl_interface': 4.540041695987917e-06, 'learning_rate/action_model': 4.540041695987917e-05, 'epoch': 0.2}) +[2026-10-03 15:09:41,040][starVLA.training.train_starvla][INFO] - Step 2785, Loss: {'train/loss': 0.00029728744993917644, 'train/grad_norm_pre_clip': 0.11083408445119858, 'timing/data': 0.00032376000308431685, 'timing/model': 0.36522124899784103, 'learning_rate/qwen_vl_interface': 4.537030629787182e-06, 'learning_rate/action_model': 4.5370306297871826e-05, 'epoch': 0.2}) +[2026-10-03 15:09:41,687][starVLA.training.train_starvla][INFO] - Step 2786, Loss: {'train/loss': 0.002779344329610467, 'train/grad_norm_pre_clip': 0.37237972021102905, 'timing/data': 0.0002959200064651668, 'timing/model': 0.3645972390077077, 'learning_rate/qwen_vl_interface': 4.534019856661226e-06, 'learning_rate/action_model': 4.534019856661226e-05, 'epoch': 0.2}) +[2026-10-03 15:09:42,509][starVLA.training.train_starvla][INFO] - Step 2787, Loss: {'train/loss': 0.004175836220383644, 'train/grad_norm_pre_clip': 0.893170952796936, 'timing/data': 0.0003325400030007586, 'timing/model': 0.34246289200382307, 'learning_rate/qwen_vl_interface': 4.531009377847661e-06, 'learning_rate/action_model': 4.531009377847661e-05, 'epoch': 0.2}) +[2026-10-03 15:09:43,160][starVLA.training.train_starvla][INFO] - Step 2788, Loss: {'train/loss': 0.002021010732278228, 'train/grad_norm_pre_clip': 2.60956072807312, 'timing/data': 0.00034047999361064285, 'timing/model': 0.3644061189988861, 'learning_rate/qwen_vl_interface': 4.527999194583982e-06, 'learning_rate/action_model': 4.527999194583981e-05, 'epoch': 0.2}) +[2026-10-03 15:09:43,808][starVLA.training.train_starvla][INFO] - Step 2789, Loss: {'train/loss': 0.0038940259255468845, 'train/grad_norm_pre_clip': 3.496920347213745, 'timing/data': 0.00019674999930430204, 'timing/model': 0.36510949800140224, 'learning_rate/qwen_vl_interface': 4.52498930810756e-06, 'learning_rate/action_model': 4.52498930810756e-05, 'epoch': 0.2}) +[2026-10-03 15:09:44,455][starVLA.training.train_starvla][INFO] - Step 2790, Loss: {'train/loss': 0.00034593266900628805, 'train/grad_norm_pre_clip': 0.07561991363763809, 'timing/data': 0.0002486500015947968, 'timing/model': 0.36481166898738593, 'learning_rate/qwen_vl_interface': 4.521979719655648e-06, 'learning_rate/action_model': 4.5219797196556484e-05, 'epoch': 0.2}) +[2026-10-03 15:09:45,105][starVLA.training.train_starvla][INFO] - Step 2791, Loss: {'train/loss': 0.4689065217971802, 'train/grad_norm_pre_clip': 16.438983917236328, 'timing/data': 0.00025384000036865473, 'timing/model': 0.365468988995417, 'learning_rate/qwen_vl_interface': 4.518970430465377e-06, 'learning_rate/action_model': 4.518970430465376e-05, 'epoch': 0.2}) +[2026-10-03 15:09:45,940][starVLA.training.train_starvla][INFO] - Step 2792, Loss: {'train/loss': 0.0006551046390086412, 'train/grad_norm_pre_clip': 0.22979149222373962, 'timing/data': 0.0003077099972870201, 'timing/model': 0.5506154939939734, 'learning_rate/qwen_vl_interface': 4.515961441773745e-06, 'learning_rate/action_model': 4.515961441773745e-05, 'epoch': 0.2}) +[2026-10-03 15:09:46,588][starVLA.training.train_starvla][INFO] - Step 2793, Loss: {'train/loss': 0.07004240900278091, 'train/grad_norm_pre_clip': 48.43505096435547, 'timing/data': 0.0002338900085305795, 'timing/model': 0.36523426898929756, 'learning_rate/qwen_vl_interface': 4.5129527548176385e-06, 'learning_rate/action_model': 4.512952754817638e-05, 'epoch': 0.2}) +[2026-10-03 15:09:47,235][starVLA.training.train_starvla][INFO] - Step 2794, Loss: {'train/loss': 0.0015227016992866993, 'train/grad_norm_pre_clip': 0.21431143581867218, 'timing/data': 0.00019625001004897058, 'timing/model': 0.36479450900515076, 'learning_rate/qwen_vl_interface': 4.509944370833813e-06, 'learning_rate/action_model': 4.509944370833813e-05, 'epoch': 0.2}) +[2026-10-03 15:09:47,885][starVLA.training.train_starvla][INFO] - Step 2795, Loss: {'train/loss': 0.0013385065831243992, 'train/grad_norm_pre_clip': 0.47352585196495056, 'timing/data': 0.0002253400016343221, 'timing/model': 0.36528801900567487, 'learning_rate/qwen_vl_interface': 4.506936291058902e-06, 'learning_rate/action_model': 4.506936291058902e-05, 'epoch': 0.2}) +[2026-10-03 15:09:48,534][starVLA.training.train_starvla][INFO] - Step 2796, Loss: {'train/loss': 0.023066911846399307, 'train/grad_norm_pre_clip': 21.63747787475586, 'timing/data': 0.0002531299978727475, 'timing/model': 0.36495981899497565, 'learning_rate/qwen_vl_interface': 4.503928516729416e-06, 'learning_rate/action_model': 4.503928516729416e-05, 'epoch': 0.2}) +[2026-10-03 15:09:49,367][starVLA.training.train_starvla][INFO] - Step 2797, Loss: {'train/loss': 0.0008996255346573889, 'train/grad_norm_pre_clip': 0.4127850830554962, 'timing/data': 0.00023761999909766018, 'timing/model': 0.5511164329946041, 'learning_rate/qwen_vl_interface': 4.500921049081734e-06, 'learning_rate/action_model': 4.500921049081734e-05, 'epoch': 0.2}) +[2026-10-03 15:09:50,015][starVLA.training.train_starvla][INFO] - Step 2798, Loss: {'train/loss': 0.0007706530159339309, 'train/grad_norm_pre_clip': 0.46379709243774414, 'timing/data': 0.00020516000222414732, 'timing/model': 0.365773117999197, 'learning_rate/qwen_vl_interface': 4.497913889352113e-06, 'learning_rate/action_model': 4.4979138893521126e-05, 'epoch': 0.2}) +[2026-10-03 15:09:50,666][starVLA.training.train_starvla][INFO] - Step 2799, Loss: {'train/loss': 0.00016168624279089272, 'train/grad_norm_pre_clip': 0.04144391790032387, 'timing/data': 0.0002945099986391142, 'timing/model': 0.3650726290070452, 'learning_rate/qwen_vl_interface': 4.494907038776684e-06, 'learning_rate/action_model': 4.4949070387766836e-05, 'epoch': 0.2}) +[2026-10-03 15:09:51,312][starVLA.training.train_starvla][INFO] - Step 2800, Loss: {'train/loss': 0.00428814860060811, 'train/grad_norm_pre_clip': 10.779651641845703, 'timing/data': 0.0003403799928491935, 'timing/model': 0.36423296900466084, 'learning_rate/qwen_vl_interface': 4.49190049859145e-06, 'learning_rate/action_model': 4.49190049859145e-05, 'epoch': 0.2}) +[2026-10-03 15:09:51,960][starVLA.training.train_starvla][INFO] - Step 2801, Loss: {'train/loss': 0.00013863149797543883, 'train/grad_norm_pre_clip': 0.024971844628453255, 'timing/data': 0.00019893000717274845, 'timing/model': 0.36512746900552884, 'learning_rate/qwen_vl_interface': 4.488894270032284e-06, 'learning_rate/action_model': 4.4888942700322834e-05, 'epoch': 0.21}) +[2026-10-03 15:09:52,802][starVLA.training.train_starvla][INFO] - Step 2802, Loss: {'train/loss': 0.00011047301813960075, 'train/grad_norm_pre_clip': 0.02525639906525612, 'timing/data': 0.00027940000290982425, 'timing/model': 0.5596085519937333, 'learning_rate/qwen_vl_interface': 4.485888354334933e-06, 'learning_rate/action_model': 4.4858883543349334e-05, 'epoch': 0.21}) +[2026-10-03 15:09:53,450][starVLA.training.train_starvla][INFO] - Step 2803, Loss: {'train/loss': 0.0539570115506649, 'train/grad_norm_pre_clip': 207.44174194335938, 'timing/data': 0.0003907899954356253, 'timing/model': 0.36449324899876956, 'learning_rate/qwen_vl_interface': 4.482882752735017e-06, 'learning_rate/action_model': 4.482882752735017e-05, 'epoch': 0.21}) +[2026-10-03 15:09:54,097][starVLA.training.train_starvla][INFO] - Step 2804, Loss: {'train/loss': 0.0010701020946726203, 'train/grad_norm_pre_clip': 0.646786093711853, 'timing/data': 0.00033290999999735504, 'timing/model': 0.3644306189962663, 'learning_rate/qwen_vl_interface': 4.479877466468027e-06, 'learning_rate/action_model': 4.479877466468026e-05, 'epoch': 0.21}) +[2026-10-03 15:09:54,722][starVLA.training.train_starvla][INFO] - Step 2805, Loss: {'train/loss': 0.0009724153205752373, 'train/grad_norm_pre_clip': 0.31273481249809265, 'timing/data': 0.00019952000002376735, 'timing/model': 0.3645379390072776, 'learning_rate/qwen_vl_interface': 4.476872496769318e-06, 'learning_rate/action_model': 4.476872496769318e-05, 'epoch': 0.21}) +[2026-10-03 15:09:55,369][starVLA.training.train_starvla][INFO] - Step 2806, Loss: {'train/loss': 9.436463005840778e-05, 'train/grad_norm_pre_clip': 0.01942860707640648, 'timing/data': 0.000321980012813583, 'timing/model': 0.36378978900029324, 'learning_rate/qwen_vl_interface': 4.473867844874122e-06, 'learning_rate/action_model': 4.473867844874122e-05, 'epoch': 0.21}) +[2026-10-03 15:09:56,017][starVLA.training.train_starvla][INFO] - Step 2807, Loss: {'train/loss': 0.001155621139332652, 'train/grad_norm_pre_clip': 0.3905489146709442, 'timing/data': 0.0002523400034988299, 'timing/model': 0.36434039899904747, 'learning_rate/qwen_vl_interface': 4.470863512017539e-06, 'learning_rate/action_model': 4.470863512017539e-05, 'epoch': 0.21}) +[2026-10-03 15:09:56,834][starVLA.training.train_starvla][INFO] - Step 2808, Loss: {'train/loss': 0.0006376043311320245, 'train/grad_norm_pre_clip': 0.2015783190727234, 'timing/data': 0.0002035100042121485, 'timing/model': 0.36388320999685675, 'learning_rate/qwen_vl_interface': 4.4678594994345344e-06, 'learning_rate/action_model': 4.467859499434534e-05, 'epoch': 0.21}) +[2026-10-03 15:09:57,479][starVLA.training.train_starvla][INFO] - Step 2809, Loss: {'train/loss': 0.00013472263526637107, 'train/grad_norm_pre_clip': 0.04335023835301399, 'timing/data': 0.0003105500072706491, 'timing/model': 0.3634178390057059, 'learning_rate/qwen_vl_interface': 4.4648558083599475e-06, 'learning_rate/action_model': 4.464855808359947e-05, 'epoch': 0.21}) +[2026-10-03 15:09:58,125][starVLA.training.train_starvla][INFO] - Step 2810, Loss: {'train/loss': 0.728762149810791, 'train/grad_norm_pre_clip': 98.72029876708984, 'timing/data': 0.00024744999245740473, 'timing/model': 0.363699869994889, 'learning_rate/qwen_vl_interface': 4.461852440028479e-06, 'learning_rate/action_model': 4.461852440028479e-05, 'epoch': 0.21}) +[2026-10-03 15:09:58,773][starVLA.training.train_starvla][INFO] - Step 2811, Loss: {'train/loss': 0.00015980898751877248, 'train/grad_norm_pre_clip': 0.06213589385151863, 'timing/data': 0.00020216999109834433, 'timing/model': 0.36461209900153335, 'learning_rate/qwen_vl_interface': 4.458849395674702e-06, 'learning_rate/action_model': 4.458849395674701e-05, 'epoch': 0.21}) +[2026-10-03 15:09:59,420][starVLA.training.train_starvla][INFO] - Step 2812, Loss: {'train/loss': 0.0001851208508014679, 'train/grad_norm_pre_clip': 0.05460740253329277, 'timing/data': 0.00019458000315353274, 'timing/model': 0.36417785900994204, 'learning_rate/qwen_vl_interface': 4.455846676533052e-06, 'learning_rate/action_model': 4.4558466765330525e-05, 'epoch': 0.21}) +[2026-10-03 15:10:00,256][starVLA.training.train_starvla][INFO] - Step 2813, Loss: {'train/loss': 0.0007106245029717684, 'train/grad_norm_pre_clip': 0.1036774218082428, 'timing/data': 0.0002443900011712685, 'timing/model': 0.5532436719950056, 'learning_rate/qwen_vl_interface': 4.452844283837838e-06, 'learning_rate/action_model': 4.452844283837838e-05, 'epoch': 0.21}) +[2026-10-03 15:10:00,902][starVLA.training.train_starvla][INFO] - Step 2814, Loss: {'train/loss': 0.0001646737800911069, 'train/grad_norm_pre_clip': 0.06930558383464813, 'timing/data': 0.00019928999245166779, 'timing/model': 0.36383368899987545, 'learning_rate/qwen_vl_interface': 4.449842218823228e-06, 'learning_rate/action_model': 4.449842218823228e-05, 'epoch': 0.21}) +[2026-10-03 15:10:01,551][starVLA.training.train_starvla][INFO] - Step 2815, Loss: {'train/loss': 5.350603169063106e-05, 'train/grad_norm_pre_clip': 0.01948253996670246, 'timing/data': 0.00020463000691961497, 'timing/model': 0.3648904000001494, 'learning_rate/qwen_vl_interface': 4.446840482723256e-06, 'learning_rate/action_model': 4.446840482723255e-05, 'epoch': 0.21}) +[2026-10-03 15:10:02,176][starVLA.training.train_starvla][INFO] - Step 2816, Loss: {'train/loss': 0.00021811030455864966, 'train/grad_norm_pre_clip': 0.0648283064365387, 'timing/data': 0.0001988300064112991, 'timing/model': 0.3420892630092567, 'learning_rate/qwen_vl_interface': 4.443839076771823e-06, 'learning_rate/action_model': 4.443839076771823e-05, 'epoch': 0.21}) +[2026-10-03 15:10:02,801][starVLA.training.train_starvla][INFO] - Step 2817, Loss: {'train/loss': 0.00010889561963267624, 'train/grad_norm_pre_clip': 0.034815892577171326, 'timing/data': 0.00031197899079415947, 'timing/model': 0.342608923005173, 'learning_rate/qwen_vl_interface': 4.4408380022026945e-06, 'learning_rate/action_model': 4.440838002202694e-05, 'epoch': 0.21}) +[2026-10-03 15:10:03,633][starVLA.training.train_starvla][INFO] - Step 2818, Loss: {'train/loss': 0.0008303801296278834, 'train/grad_norm_pre_clip': 1.4659284353256226, 'timing/data': 0.0002996600087499246, 'timing/model': 0.5494415640132502, 'learning_rate/qwen_vl_interface': 4.437837260249499e-06, 'learning_rate/action_model': 4.437837260249499e-05, 'epoch': 0.21}) +[2026-10-03 15:10:04,280][starVLA.training.train_starvla][INFO] - Step 2819, Loss: {'train/loss': 0.0003571051638573408, 'train/grad_norm_pre_clip': 0.3171337842941284, 'timing/data': 0.00020123999274801463, 'timing/model': 0.3639928800112102, 'learning_rate/qwen_vl_interface': 4.434836852145728e-06, 'learning_rate/action_model': 4.434836852145728e-05, 'epoch': 0.21}) +[2026-10-03 15:10:04,926][starVLA.training.train_starvla][INFO] - Step 2820, Loss: {'train/loss': 7.343134348047897e-05, 'train/grad_norm_pre_clip': 0.018654286861419678, 'timing/data': 0.00033310000435449183, 'timing/model': 0.36421252899162937, 'learning_rate/qwen_vl_interface': 4.431836779124736e-06, 'learning_rate/action_model': 4.431836779124735e-05, 'epoch': 0.21}) +[2026-10-03 15:10:05,550][starVLA.training.train_starvla][INFO] - Step 2821, Loss: {'train/loss': 0.00022239149257075042, 'train/grad_norm_pre_clip': 0.18295159935951233, 'timing/data': 0.00030032999347895384, 'timing/model': 0.34193152200896293, 'learning_rate/qwen_vl_interface': 4.428837042419738e-06, 'learning_rate/action_model': 4.428837042419738e-05, 'epoch': 0.21}) +[2026-10-03 15:10:06,196][starVLA.training.train_starvla][INFO] - Step 2822, Loss: {'train/loss': 0.0003504467604216188, 'train/grad_norm_pre_clip': 0.04880969971418381, 'timing/data': 0.0003264400002080947, 'timing/model': 0.36330682899279054, 'learning_rate/qwen_vl_interface': 4.425837643263815e-06, 'learning_rate/action_model': 4.425837643263814e-05, 'epoch': 0.21}) +[2026-10-03 15:10:07,042][starVLA.training.train_starvla][INFO] - Step 2823, Loss: {'train/loss': 0.0007145777344703674, 'train/grad_norm_pre_clip': 0.22162273526191711, 'timing/data': 0.00022298999829217792, 'timing/model': 0.5616567820106866, 'learning_rate/qwen_vl_interface': 4.4228385828899045e-06, 'learning_rate/action_model': 4.4228385828899046e-05, 'epoch': 0.21}) +[2026-10-03 15:10:07,688][starVLA.training.train_starvla][INFO] - Step 2824, Loss: {'train/loss': 0.10491908341646194, 'train/grad_norm_pre_clip': 143.01678466796875, 'timing/data': 0.0003303400008007884, 'timing/model': 0.36442490899935365, 'learning_rate/qwen_vl_interface': 4.419839862530809e-06, 'learning_rate/action_model': 4.419839862530809e-05, 'epoch': 0.21}) +[2026-10-03 15:10:08,335][starVLA.training.train_starvla][INFO] - Step 2825, Loss: {'train/loss': 0.0004122497630305588, 'train/grad_norm_pre_clip': 0.07028714567422867, 'timing/data': 0.000377670003217645, 'timing/model': 0.36448525899322703, 'learning_rate/qwen_vl_interface': 4.416841483419188e-06, 'learning_rate/action_model': 4.416841483419188e-05, 'epoch': 0.21}) +[2026-10-03 15:10:08,981][starVLA.training.train_starvla][INFO] - Step 2826, Loss: {'train/loss': 1.681570392975118e-05, 'train/grad_norm_pre_clip': 0.0031114513985812664, 'timing/data': 0.00020511000184342265, 'timing/model': 0.3638231990044005, 'learning_rate/qwen_vl_interface': 4.413843446787561e-06, 'learning_rate/action_model': 4.413843446787561e-05, 'epoch': 0.21}) +[2026-10-03 15:10:09,630][starVLA.training.train_starvla][INFO] - Step 2827, Loss: {'train/loss': 0.0010165916755795479, 'train/grad_norm_pre_clip': 0.997622013092041, 'timing/data': 0.0002079699916066602, 'timing/model': 0.36499743899912573, 'learning_rate/qwen_vl_interface': 4.41084575386831e-06, 'learning_rate/action_model': 4.4108457538683093e-05, 'epoch': 0.21}) +[2026-10-03 15:10:10,277][starVLA.training.train_starvla][INFO] - Step 2828, Loss: {'train/loss': 0.4714292883872986, 'train/grad_norm_pre_clip': 131.38790893554688, 'timing/data': 0.00024096999550238252, 'timing/model': 0.36444421899796, 'learning_rate/qwen_vl_interface': 4.40784840589367e-06, 'learning_rate/action_model': 4.4078484058936706e-05, 'epoch': 0.21}) +[2026-10-03 15:10:11,117][starVLA.training.train_starvla][INFO] - Step 2829, Loss: {'train/loss': 6.289944576565176e-05, 'train/grad_norm_pre_clip': 0.0189052764326334, 'timing/data': 0.0003536589938448742, 'timing/model': 0.3645935800013831, 'learning_rate/qwen_vl_interface': 4.404851404095742e-06, 'learning_rate/action_model': 4.404851404095742e-05, 'epoch': 0.21}) +[2026-10-03 15:10:11,765][starVLA.training.train_starvla][INFO] - Step 2830, Loss: {'train/loss': 0.1361585259437561, 'train/grad_norm_pre_clip': 209.63998413085938, 'timing/data': 0.00031154000316746533, 'timing/model': 0.3646154900052352, 'learning_rate/qwen_vl_interface': 4.4018547497064765e-06, 'learning_rate/action_model': 4.401854749706476e-05, 'epoch': 0.21}) +[2026-10-03 15:10:12,415][starVLA.training.train_starvla][INFO] - Step 2831, Loss: {'train/loss': 0.017315248027443886, 'train/grad_norm_pre_clip': 154.67977905273438, 'timing/data': 0.00036191999970469624, 'timing/model': 0.36503845900006127, 'learning_rate/qwen_vl_interface': 4.398858443957686e-06, 'learning_rate/action_model': 4.398858443957686e-05, 'epoch': 0.21}) +[2026-10-03 15:10:13,062][starVLA.training.train_starvla][INFO] - Step 2832, Loss: {'train/loss': 0.0024727052077651024, 'train/grad_norm_pre_clip': 9.685461044311523, 'timing/data': 0.00027490001230034977, 'timing/model': 0.3646404790051747, 'learning_rate/qwen_vl_interface': 4.395862488081038e-06, 'learning_rate/action_model': 4.395862488081038e-05, 'epoch': 0.21}) +[2026-10-03 15:10:13,710][starVLA.training.train_starvla][INFO] - Step 2833, Loss: {'train/loss': 1.0854133367538452, 'train/grad_norm_pre_clip': 44.52206802368164, 'timing/data': 0.00019778999558184296, 'timing/model': 0.3648247490054928, 'learning_rate/qwen_vl_interface': 4.392866883308059e-06, 'learning_rate/action_model': 4.392866883308059e-05, 'epoch': 0.21}) +[2026-10-03 15:10:14,552][starVLA.training.train_starvla][INFO] - Step 2834, Loss: {'train/loss': 0.3958781659603119, 'train/grad_norm_pre_clip': 43.47956466674805, 'timing/data': 0.0002483400021446869, 'timing/model': 0.3640744690055726, 'learning_rate/qwen_vl_interface': 4.389871630870127e-06, 'learning_rate/action_model': 4.389871630870127e-05, 'epoch': 0.21}) +[2026-10-03 15:10:15,202][starVLA.training.train_starvla][INFO] - Step 2835, Loss: {'train/loss': 7.324366742977872e-05, 'train/grad_norm_pre_clip': 0.025278225541114807, 'timing/data': 0.00020558999676723033, 'timing/model': 0.3649112389975926, 'learning_rate/qwen_vl_interface': 4.386876731998478e-06, 'learning_rate/action_model': 4.386876731998478e-05, 'epoch': 0.21}) +[2026-10-03 15:10:15,850][starVLA.training.train_starvla][INFO] - Step 2836, Loss: {'train/loss': 2.2202199033927172e-05, 'train/grad_norm_pre_clip': 0.0037593564484268427, 'timing/data': 0.000248309996095486, 'timing/model': 0.36510128900408745, 'learning_rate/qwen_vl_interface': 4.3838821879242006e-06, 'learning_rate/action_model': 4.3838821879242e-05, 'epoch': 0.21}) +[2026-10-03 15:10:16,497][starVLA.training.train_starvla][INFO] - Step 2837, Loss: {'train/loss': 1.7702303011901677e-05, 'train/grad_norm_pre_clip': 0.004100701306015253, 'timing/data': 0.00024308999127242714, 'timing/model': 0.36470181900949683, 'learning_rate/qwen_vl_interface': 4.38088799987824e-06, 'learning_rate/action_model': 4.38088799987824e-05, 'epoch': 0.21}) +[2026-10-03 15:10:17,143][starVLA.training.train_starvla][INFO] - Step 2838, Loss: {'train/loss': 0.0013767877826467156, 'train/grad_norm_pre_clip': 2.2831852436065674, 'timing/data': 0.0002436199865769595, 'timing/model': 0.36392532898753416, 'learning_rate/qwen_vl_interface': 4.3778941690913935e-06, 'learning_rate/action_model': 4.3778941690913934e-05, 'epoch': 0.21}) +[2026-10-03 15:10:17,982][starVLA.training.train_starvla][INFO] - Step 2839, Loss: {'train/loss': 0.00020406472322065383, 'train/grad_norm_pre_clip': 0.2767280340194702, 'timing/data': 0.0003334900102345273, 'timing/model': 0.5543812529940624, 'learning_rate/qwen_vl_interface': 4.374900696794311e-06, 'learning_rate/action_model': 4.374900696794311e-05, 'epoch': 0.21}) +[2026-10-03 15:10:18,608][starVLA.training.train_starvla][INFO] - Step 2840, Loss: {'train/loss': 0.00026733402046374977, 'train/grad_norm_pre_clip': 0.16648896038532257, 'timing/data': 0.00033260999771300703, 'timing/model': 0.36468678900564555, 'learning_rate/qwen_vl_interface': 4.371907584217499e-06, 'learning_rate/action_model': 4.371907584217498e-05, 'epoch': 0.21}) +[2026-10-03 15:10:19,255][starVLA.training.train_starvla][INFO] - Step 2841, Loss: {'train/loss': 1.8246075342176482e-05, 'train/grad_norm_pre_clip': 0.003764705266803503, 'timing/data': 0.00024247000692412257, 'timing/model': 0.36479339899960905, 'learning_rate/qwen_vl_interface': 4.368914832591309e-06, 'learning_rate/action_model': 4.368914832591309e-05, 'epoch': 0.21}) +[2026-10-03 15:10:19,902][starVLA.training.train_starvla][INFO] - Step 2842, Loss: {'train/loss': 0.3615684509277344, 'train/grad_norm_pre_clip': 29.174362182617188, 'timing/data': 0.00021268001000862569, 'timing/model': 0.36430517899862025, 'learning_rate/qwen_vl_interface': 4.365922443145949e-06, 'learning_rate/action_model': 4.365922443145949e-05, 'epoch': 0.21}) +[2026-10-03 15:10:20,551][starVLA.training.train_starvla][INFO] - Step 2843, Loss: {'train/loss': 0.49241212010383606, 'train/grad_norm_pre_clip': 46.963134765625, 'timing/data': 0.0003306699945824221, 'timing/model': 0.36476769899309147, 'learning_rate/qwen_vl_interface': 4.362930417111479e-06, 'learning_rate/action_model': 4.362930417111479e-05, 'epoch': 0.21}) +[2026-10-03 15:10:21,395][starVLA.training.train_starvla][INFO] - Step 2844, Loss: {'train/loss': 4.5600147132063285e-05, 'train/grad_norm_pre_clip': 0.012225217185914516, 'timing/data': 0.0003053199907299131, 'timing/model': 0.5607231819885783, 'learning_rate/qwen_vl_interface': 4.359938755717807e-06, 'learning_rate/action_model': 4.359938755717807e-05, 'epoch': 0.21}) +[2026-10-03 15:10:22,043][starVLA.training.train_starvla][INFO] - Step 2845, Loss: {'train/loss': 0.0005235910648480058, 'train/grad_norm_pre_clip': 0.18861187994480133, 'timing/data': 0.00019735000387299806, 'timing/model': 0.3642935889947694, 'learning_rate/qwen_vl_interface': 4.3569474601946934e-06, 'learning_rate/action_model': 4.356947460194694e-05, 'epoch': 0.21}) +[2026-10-03 15:10:22,689][starVLA.training.train_starvla][INFO] - Step 2846, Loss: {'train/loss': 0.0003193407319486141, 'train/grad_norm_pre_clip': 0.5285038948059082, 'timing/data': 0.0003125299990642816, 'timing/model': 0.3644215499953134, 'learning_rate/qwen_vl_interface': 4.3539565317717446e-06, 'learning_rate/action_model': 4.353956531771745e-05, 'epoch': 0.21}) +[2026-10-03 15:10:23,338][starVLA.training.train_starvla][INFO] - Step 2847, Loss: {'train/loss': 0.00011801051732618362, 'train/grad_norm_pre_clip': 0.022170523181557655, 'timing/data': 0.0003161100030411035, 'timing/model': 0.36453589900338557, 'learning_rate/qwen_vl_interface': 4.350965971678419e-06, 'learning_rate/action_model': 4.3509659716784185e-05, 'epoch': 0.21}) +[2026-10-03 15:10:23,986][starVLA.training.train_starvla][INFO] - Step 2848, Loss: {'train/loss': 2.7886715542990714e-05, 'train/grad_norm_pre_clip': 0.005809517577290535, 'timing/data': 0.000247880001552403, 'timing/model': 0.364864408998983, 'learning_rate/qwen_vl_interface': 4.3479757811440225e-06, 'learning_rate/action_model': 4.347975781144022e-05, 'epoch': 0.21}) +[2026-10-03 15:10:24,632][starVLA.training.train_starvla][INFO] - Step 2849, Loss: {'train/loss': 7.363149779848754e-05, 'train/grad_norm_pre_clip': 0.01567617803812027, 'timing/data': 0.00035215000389143825, 'timing/model': 0.364053248995333, 'learning_rate/qwen_vl_interface': 4.34498596139771e-06, 'learning_rate/action_model': 4.34498596139771e-05, 'epoch': 0.21}) +[2026-10-03 15:10:25,467][starVLA.training.train_starvla][INFO] - Step 2850, Loss: {'train/loss': 5.4152857046574354e-05, 'train/grad_norm_pre_clip': 0.015592385083436966, 'timing/data': 0.0003466900088824332, 'timing/model': 0.364128109999001, 'learning_rate/qwen_vl_interface': 4.341996513668483e-06, 'learning_rate/action_model': 4.3419965136684824e-05, 'epoch': 0.21}) +[2026-10-03 15:10:26,115][starVLA.training.train_starvla][INFO] - Step 2851, Loss: {'train/loss': 6.6679174778983e-05, 'train/grad_norm_pre_clip': 0.03249788656830788, 'timing/data': 0.00028962999931536615, 'timing/model': 0.3642838790110545, 'learning_rate/qwen_vl_interface': 4.33900743918519e-06, 'learning_rate/action_model': 4.33900743918519e-05, 'epoch': 0.21}) +[2026-10-03 15:10:26,762][starVLA.training.train_starvla][INFO] - Step 2852, Loss: {'train/loss': 9.942044562194496e-05, 'train/grad_norm_pre_clip': 0.07435750216245651, 'timing/data': 0.00025011000980157405, 'timing/model': 0.3645442790002562, 'learning_rate/qwen_vl_interface': 4.336018739176529e-06, 'learning_rate/action_model': 4.33601873917653e-05, 'epoch': 0.21}) +[2026-10-03 15:10:27,386][starVLA.training.train_starvla][INFO] - Step 2853, Loss: {'train/loss': 4.7648220061091706e-05, 'train/grad_norm_pre_clip': 0.020666588097810745, 'timing/data': 0.0002380799996899441, 'timing/model': 0.3420055530004902, 'learning_rate/qwen_vl_interface': 4.333030414871041e-06, 'learning_rate/action_model': 4.33303041487104e-05, 'epoch': 0.21}) +[2026-10-03 15:10:28,032][starVLA.training.train_starvla][INFO] - Step 2854, Loss: {'train/loss': 0.00017566360475029796, 'train/grad_norm_pre_clip': 0.04305662959814072, 'timing/data': 0.00032774900319054723, 'timing/model': 0.36353324000083376, 'learning_rate/qwen_vl_interface': 4.330042467497111e-06, 'learning_rate/action_model': 4.330042467497111e-05, 'epoch': 0.21}) +[2026-10-03 15:10:28,883][starVLA.training.train_starvla][INFO] - Step 2855, Loss: {'train/loss': 4.8380410589743406e-05, 'train/grad_norm_pre_clip': 0.011800424195826054, 'timing/data': 0.00023636998957954347, 'timing/model': 0.36726281899609603, 'learning_rate/qwen_vl_interface': 4.327054898282973e-06, 'learning_rate/action_model': 4.3270548982829727e-05, 'epoch': 0.21}) +[2026-10-03 15:10:29,534][starVLA.training.train_starvla][INFO] - Step 2856, Loss: {'train/loss': 0.00011488658492453396, 'train/grad_norm_pre_clip': 0.03350275009870529, 'timing/data': 0.0002001999964704737, 'timing/model': 0.36630680999951437, 'learning_rate/qwen_vl_interface': 4.324067708456703e-06, 'learning_rate/action_model': 4.3240677084567036e-05, 'epoch': 0.21}) +[2026-10-03 15:10:30,184][starVLA.training.train_starvla][INFO] - Step 2857, Loss: {'train/loss': 0.0031020399183034897, 'train/grad_norm_pre_clip': 1.8890055418014526, 'timing/data': 0.00032916000054683536, 'timing/model': 0.36601988899928983, 'learning_rate/qwen_vl_interface': 4.321080899246225e-06, 'learning_rate/action_model': 4.321080899246225e-05, 'epoch': 0.21}) +[2026-10-03 15:10:30,835][starVLA.training.train_starvla][INFO] - Step 2858, Loss: {'train/loss': 1.8521892343414947e-05, 'train/grad_norm_pre_clip': 0.002716122893616557, 'timing/data': 0.00023960000544320792, 'timing/model': 0.36632942999131046, 'learning_rate/qwen_vl_interface': 4.318094471879299e-06, 'learning_rate/action_model': 4.3180944718792985e-05, 'epoch': 0.21}) +[2026-10-03 15:10:31,487][starVLA.training.train_starvla][INFO] - Step 2859, Loss: {'train/loss': 0.514141321182251, 'train/grad_norm_pre_clip': 44.31010437011719, 'timing/data': 0.00023765998776070774, 'timing/model': 0.36680685900500976, 'learning_rate/qwen_vl_interface': 4.315108427583535e-06, 'learning_rate/action_model': 4.315108427583535e-05, 'epoch': 0.21}) +[2026-10-03 15:10:32,329][starVLA.training.train_starvla][INFO] - Step 2860, Loss: {'train/loss': 0.009176154620945454, 'train/grad_norm_pre_clip': 32.220947265625, 'timing/data': 0.00027222999779041857, 'timing/model': 0.5567694330093218, 'learning_rate/qwen_vl_interface': 4.312122767586381e-06, 'learning_rate/action_model': 4.3121227675863815e-05, 'epoch': 0.21}) +[2026-10-03 15:10:32,980][starVLA.training.train_starvla][INFO] - Step 2861, Loss: {'train/loss': 0.0026478315703570843, 'train/grad_norm_pre_clip': 9.054372787475586, 'timing/data': 0.0002471300103934482, 'timing/model': 0.36583963900920935, 'learning_rate/qwen_vl_interface': 4.3091374931151324e-06, 'learning_rate/action_model': 4.309137493115132e-05, 'epoch': 0.21}) +[2026-10-03 15:10:33,630][starVLA.training.train_starvla][INFO] - Step 2862, Loss: {'train/loss': 5.75194826524239e-05, 'train/grad_norm_pre_clip': 0.021375838667154312, 'timing/data': 0.0003190200077369809, 'timing/model': 0.36618225899292156, 'learning_rate/qwen_vl_interface': 4.306152605396921e-06, 'learning_rate/action_model': 4.306152605396921e-05, 'epoch': 0.21}) +[2026-10-03 15:10:34,283][starVLA.training.train_starvla][INFO] - Step 2863, Loss: {'train/loss': 2.2358775822794996e-05, 'train/grad_norm_pre_clip': 0.004258779808878899, 'timing/data': 0.00028577000193763524, 'timing/model': 0.3666450780001469, 'learning_rate/qwen_vl_interface': 4.30316810565872e-06, 'learning_rate/action_model': 4.3031681056587206e-05, 'epoch': 0.21}) +[2026-10-03 15:10:34,911][starVLA.training.train_starvla][INFO] - Step 2864, Loss: {'train/loss': 2.5450422981521115e-05, 'train/grad_norm_pre_clip': 0.0055902316235005856, 'timing/data': 0.0002599700092105195, 'timing/model': 0.34377417199721094, 'learning_rate/qwen_vl_interface': 4.300183995127347e-06, 'learning_rate/action_model': 4.300183995127346e-05, 'epoch': 0.21}) +[2026-10-03 15:10:35,753][starVLA.training.train_starvla][INFO] - Step 2865, Loss: {'train/loss': 0.07320499420166016, 'train/grad_norm_pre_clip': 86.16999053955078, 'timing/data': 0.00036996000562794507, 'timing/model': 0.5571798129967647, 'learning_rate/qwen_vl_interface': 4.297200275029454e-06, 'learning_rate/action_model': 4.297200275029454e-05, 'epoch': 0.21}) +[2026-10-03 15:10:36,405][starVLA.training.train_starvla][INFO] - Step 2866, Loss: {'train/loss': 0.23634515702724457, 'train/grad_norm_pre_clip': 134.73097229003906, 'timing/data': 0.00019307000911794603, 'timing/model': 0.36706842899729963, 'learning_rate/qwen_vl_interface': 4.29421694659154e-06, 'learning_rate/action_model': 4.294216946591539e-05, 'epoch': 0.21}) +[2026-10-03 15:10:37,035][starVLA.training.train_starvla][INFO] - Step 2867, Loss: {'train/loss': 0.00032675807597115636, 'train/grad_norm_pre_clip': 0.0695677325129509, 'timing/data': 0.00021270000434014946, 'timing/model': 0.36696590899373405, 'learning_rate/qwen_vl_interface': 4.2912340110399345e-06, 'learning_rate/action_model': 4.291234011039934e-05, 'epoch': 0.21}) +[2026-10-03 15:10:37,686][starVLA.training.train_starvla][INFO] - Step 2868, Loss: {'train/loss': 2.3014195903670043e-05, 'train/grad_norm_pre_clip': 0.004316600505262613, 'timing/data': 0.00024336999922525138, 'timing/model': 0.36656956899969373, 'learning_rate/qwen_vl_interface': 4.2882514696008125e-06, 'learning_rate/action_model': 4.2882514696008125e-05, 'epoch': 0.21}) +[2026-10-03 15:10:38,337][starVLA.training.train_starvla][INFO] - Step 2869, Loss: {'train/loss': 0.0001570344320498407, 'train/grad_norm_pre_clip': 0.07708930969238281, 'timing/data': 0.000327600006130524, 'timing/model': 0.3662003189965617, 'learning_rate/qwen_vl_interface': 4.2852693235001815e-06, 'learning_rate/action_model': 4.2852693235001815e-05, 'epoch': 0.21}) +[2026-10-03 15:10:38,988][starVLA.training.train_starvla][INFO] - Step 2870, Loss: {'train/loss': 0.5877924561500549, 'train/grad_norm_pre_clip': 61.57829284667969, 'timing/data': 0.00019747999613173306, 'timing/model': 0.3666943790012738, 'learning_rate/qwen_vl_interface': 4.2822875739638915e-06, 'learning_rate/action_model': 4.282287573963891e-05, 'epoch': 0.21}) +[2026-10-03 15:10:39,811][starVLA.training.train_starvla][INFO] - Step 2871, Loss: {'train/loss': 4.7218811232596636e-05, 'train/grad_norm_pre_clip': 0.010101446881890297, 'timing/data': 0.0003796090022660792, 'timing/model': 0.36732946999836713, 'learning_rate/qwen_vl_interface': 4.279306222217625e-06, 'learning_rate/action_model': 4.279306222217626e-05, 'epoch': 0.21}) +[2026-10-03 15:10:40,462][starVLA.training.train_starvla][INFO] - Step 2872, Loss: {'train/loss': 0.04726061597466469, 'train/grad_norm_pre_clip': 146.31130981445312, 'timing/data': 0.0002633599942782894, 'timing/model': 0.3666241790051572, 'learning_rate/qwen_vl_interface': 4.276325269486906e-06, 'learning_rate/action_model': 4.276325269486906e-05, 'epoch': 0.21}) +[2026-10-03 15:10:41,113][starVLA.training.train_starvla][INFO] - Step 2873, Loss: {'train/loss': 0.00013814984413329512, 'train/grad_norm_pre_clip': 0.034178946167230606, 'timing/data': 0.00030451000202447176, 'timing/model': 0.36659969900210854, 'learning_rate/qwen_vl_interface': 4.273344716997089e-06, 'learning_rate/action_model': 4.273344716997089e-05, 'epoch': 0.21}) +[2026-10-03 15:10:41,763][starVLA.training.train_starvla][INFO] - Step 2874, Loss: {'train/loss': 0.09195782989263535, 'train/grad_norm_pre_clip': 119.75719451904297, 'timing/data': 0.0002346499968552962, 'timing/model': 0.36549667999497615, 'learning_rate/qwen_vl_interface': 4.270364565973368e-06, 'learning_rate/action_model': 4.270364565973367e-05, 'epoch': 0.21}) +[2026-10-03 15:10:42,393][starVLA.training.train_starvla][INFO] - Step 2875, Loss: {'train/loss': 5.232911280472763e-05, 'train/grad_norm_pre_clip': 0.012362816371023655, 'timing/data': 0.0002805599942803383, 'timing/model': 0.34407682200253475, 'learning_rate/qwen_vl_interface': 4.267384817640768e-06, 'learning_rate/action_model': 4.2673848176407685e-05, 'epoch': 0.21}) +[2026-10-03 15:10:43,246][starVLA.training.train_starvla][INFO] - Step 2876, Loss: {'train/loss': 0.08595897257328033, 'train/grad_norm_pre_clip': 120.17720031738281, 'timing/data': 0.0002048599999397993, 'timing/model': 0.3653109089937061, 'learning_rate/qwen_vl_interface': 4.264405473224155e-06, 'learning_rate/action_model': 4.2644054732241546e-05, 'epoch': 0.21}) +[2026-10-03 15:10:43,894][starVLA.training.train_starvla][INFO] - Step 2877, Loss: {'train/loss': 0.00019409590458963066, 'train/grad_norm_pre_clip': 0.10481517761945724, 'timing/data': 0.00019749000784941018, 'timing/model': 0.36479390900058206, 'learning_rate/qwen_vl_interface': 4.261426533948221e-06, 'learning_rate/action_model': 4.2614265339482213e-05, 'epoch': 0.21}) +[2026-10-03 15:10:44,541][starVLA.training.train_starvla][INFO] - Step 2878, Loss: {'train/loss': 8.319649350596592e-05, 'train/grad_norm_pre_clip': 0.019783534109592438, 'timing/data': 0.00031411999952979386, 'timing/model': 0.36440095900616143, 'learning_rate/qwen_vl_interface': 4.258448001037499e-06, 'learning_rate/action_model': 4.2584480010374985e-05, 'epoch': 0.21}) +[2026-10-03 15:10:45,158][starVLA.training.train_starvla][INFO] - Step 2879, Loss: {'train/loss': 0.00015610229456797242, 'train/grad_norm_pre_clip': 0.04367278516292572, 'timing/data': 0.0002022700064117089, 'timing/model': 0.34200709300057497, 'learning_rate/qwen_vl_interface': 4.255469875716348e-06, 'learning_rate/action_model': 4.255469875716348e-05, 'epoch': 0.21}) +[2026-10-03 15:10:45,805][starVLA.training.train_starvla][INFO] - Step 2880, Loss: {'train/loss': 0.0004716848488897085, 'train/grad_norm_pre_clip': 0.0816766545176506, 'timing/data': 0.0001944800023920834, 'timing/model': 0.36504119899473153, 'learning_rate/qwen_vl_interface': 4.252492159208966e-06, 'learning_rate/action_model': 4.252492159208966e-05, 'epoch': 0.21}) +[2026-10-03 15:10:46,652][starVLA.training.train_starvla][INFO] - Step 2881, Loss: {'train/loss': 0.0003644528624136001, 'train/grad_norm_pre_clip': 0.11333682388067245, 'timing/data': 0.00029675000405404717, 'timing/model': 0.5643270220025443, 'learning_rate/qwen_vl_interface': 4.2495148527393766e-06, 'learning_rate/action_model': 4.2495148527393766e-05, 'epoch': 0.21}) +[2026-10-03 15:10:47,298][starVLA.training.train_starvla][INFO] - Step 2882, Loss: {'train/loss': 3.861503864754923e-05, 'train/grad_norm_pre_clip': 0.00796055793762207, 'timing/data': 0.00019903999054804444, 'timing/model': 0.3641881989897229, 'learning_rate/qwen_vl_interface': 4.246537957531441e-06, 'learning_rate/action_model': 4.2465379575314404e-05, 'epoch': 0.21}) +[2026-10-03 15:10:47,948][starVLA.training.train_starvla][INFO] - Step 2883, Loss: {'train/loss': 0.4061042070388794, 'train/grad_norm_pre_clip': 160.342529296875, 'timing/data': 0.00027299999783281237, 'timing/model': 0.3649208290007664, 'learning_rate/qwen_vl_interface': 4.243561474808846e-06, 'learning_rate/action_model': 4.2435614748088456e-05, 'epoch': 0.21}) +[2026-10-03 15:10:48,595][starVLA.training.train_starvla][INFO] - Step 2884, Loss: {'train/loss': 2.5614584956201725e-05, 'train/grad_norm_pre_clip': 0.004521614871919155, 'timing/data': 0.00019660999532788992, 'timing/model': 0.364861818990903, 'learning_rate/qwen_vl_interface': 4.240585405795112e-06, 'learning_rate/action_model': 4.240585405795112e-05, 'epoch': 0.21}) +[2026-10-03 15:10:49,243][starVLA.training.train_starvla][INFO] - Step 2885, Loss: {'train/loss': 7.771380478516221e-05, 'train/grad_norm_pre_clip': 0.05221119895577431, 'timing/data': 0.00029723999614361674, 'timing/model': 0.36493551000603475, 'learning_rate/qwen_vl_interface': 4.237609751713589e-06, 'learning_rate/action_model': 4.2376097517135896e-05, 'epoch': 0.21}) +[2026-10-03 15:10:49,890][starVLA.training.train_starvla][INFO] - Step 2886, Loss: {'train/loss': 0.04854952543973923, 'train/grad_norm_pre_clip': 119.07079315185547, 'timing/data': 0.0003851200017379597, 'timing/model': 0.36407605999556836, 'learning_rate/qwen_vl_interface': 4.234634513787455e-06, 'learning_rate/action_model': 4.2346345137874556e-05, 'epoch': 0.21}) +[2026-10-03 15:10:50,741][starVLA.training.train_starvla][INFO] - Step 2887, Loss: {'train/loss': 2.8594713512575254e-05, 'train/grad_norm_pre_clip': 0.006463042460381985, 'timing/data': 0.0002029899915214628, 'timing/model': 0.3665844290080713, 'learning_rate/qwen_vl_interface': 4.2316596932397195e-06, 'learning_rate/action_model': 4.2316596932397195e-05, 'epoch': 0.21}) +[2026-10-03 15:10:51,391][starVLA.training.train_starvla][INFO] - Step 2888, Loss: {'train/loss': 6.315470818663016e-05, 'train/grad_norm_pre_clip': 0.01683049276471138, 'timing/data': 0.00031628999568056315, 'timing/model': 0.36634960999072064, 'learning_rate/qwen_vl_interface': 4.22868529129322e-06, 'learning_rate/action_model': 4.22868529129322e-05, 'epoch': 0.21}) +[2026-10-03 15:10:52,043][starVLA.training.train_starvla][INFO] - Step 2889, Loss: {'train/loss': 1.5481988157262094e-05, 'train/grad_norm_pre_clip': 0.0035174530930817127, 'timing/data': 0.0001979200023924932, 'timing/model': 0.3662067189870868, 'learning_rate/qwen_vl_interface': 4.2257113091706165e-06, 'learning_rate/action_model': 4.2257113091706165e-05, 'epoch': 0.21}) +[2026-10-03 15:10:52,693][starVLA.training.train_starvla][INFO] - Step 2890, Loss: {'train/loss': 0.003173914272338152, 'train/grad_norm_pre_clip': 4.3715996742248535, 'timing/data': 0.00029196000832598656, 'timing/model': 0.3658636799955275, 'learning_rate/qwen_vl_interface': 4.222737748094404e-06, 'learning_rate/action_model': 4.222737748094404e-05, 'epoch': 0.21}) +[2026-10-03 15:10:53,322][starVLA.training.train_starvla][INFO] - Step 2891, Loss: {'train/loss': 0.0007257991819642484, 'train/grad_norm_pre_clip': 0.13431581854820251, 'timing/data': 0.0002472000051056966, 'timing/model': 0.34367291200032923, 'learning_rate/qwen_vl_interface': 4.219764609286901e-06, 'learning_rate/action_model': 4.219764609286901e-05, 'epoch': 0.21}) +[2026-10-03 15:10:54,165][starVLA.training.train_starvla][INFO] - Step 2892, Loss: {'train/loss': 2.0950392354279757e-05, 'train/grad_norm_pre_clip': 0.005171899683773518, 'timing/data': 0.00024480000138282776, 'timing/model': 0.5587124230078189, 'learning_rate/qwen_vl_interface': 4.21679189397025e-06, 'learning_rate/action_model': 4.21679189397025e-05, 'epoch': 0.21}) +[2026-10-03 15:10:54,814][starVLA.training.train_starvla][INFO] - Step 2893, Loss: {'train/loss': 0.13895517587661743, 'train/grad_norm_pre_clip': 269.22711181640625, 'timing/data': 0.0003161799977533519, 'timing/model': 0.3653273499949137, 'learning_rate/qwen_vl_interface': 4.213819603366424e-06, 'learning_rate/action_model': 4.213819603366424e-05, 'epoch': 0.21}) +[2026-10-03 15:10:55,464][starVLA.training.train_starvla][INFO] - Step 2894, Loss: {'train/loss': 0.001182842766866088, 'train/grad_norm_pre_clip': 2.5175421237945557, 'timing/data': 0.0002596589911263436, 'timing/model': 0.3655685400008224, 'learning_rate/qwen_vl_interface': 4.210847738697219e-06, 'learning_rate/action_model': 4.210847738697218e-05, 'epoch': 0.21}) +[2026-10-03 15:10:56,116][starVLA.training.train_starvla][INFO] - Step 2895, Loss: {'train/loss': 0.11262442916631699, 'train/grad_norm_pre_clip': 190.0319061279297, 'timing/data': 0.00033712000004015863, 'timing/model': 0.36610689999361057, 'learning_rate/qwen_vl_interface': 4.207876301184256e-06, 'learning_rate/action_model': 4.207876301184256e-05, 'epoch': 0.21}) +[2026-10-03 15:10:56,765][starVLA.training.train_starvla][INFO] - Step 2896, Loss: {'train/loss': 2.721564487728756e-05, 'train/grad_norm_pre_clip': 0.006114909891039133, 'timing/data': 0.00036363000981509686, 'timing/model': 0.3649845390027622, 'learning_rate/qwen_vl_interface': 4.2049052920489825e-06, 'learning_rate/action_model': 4.2049052920489825e-05, 'epoch': 0.21}) +[2026-10-03 15:10:57,414][starVLA.training.train_starvla][INFO] - Step 2897, Loss: {'train/loss': 0.0012407267931848764, 'train/grad_norm_pre_clip': 1.4290636777877808, 'timing/data': 0.00033075000101234764, 'timing/model': 0.3654215499991551, 'learning_rate/qwen_vl_interface': 4.201934712512666e-06, 'learning_rate/action_model': 4.2019347125126656e-05, 'epoch': 0.21}) +[2026-10-03 15:10:58,262][starVLA.training.train_starvla][INFO] - Step 2898, Loss: {'train/loss': 7.611071487190202e-05, 'train/grad_norm_pre_clip': 0.021593932062387466, 'timing/data': 0.00020261999452486634, 'timing/model': 0.36535422000451945, 'learning_rate/qwen_vl_interface': 4.1989645637964e-06, 'learning_rate/action_model': 4.1989645637964e-05, 'epoch': 0.21}) +[2026-10-03 15:10:58,914][starVLA.training.train_starvla][INFO] - Step 2899, Loss: {'train/loss': 0.32183998823165894, 'train/grad_norm_pre_clip': 125.77069854736328, 'timing/data': 0.00031740999838802963, 'timing/model': 0.3661657789925812, 'learning_rate/qwen_vl_interface': 4.195994847121101e-06, 'learning_rate/action_model': 4.195994847121101e-05, 'epoch': 0.21}) +[2026-10-03 15:10:59,563][starVLA.training.train_starvla][INFO] - Step 2900, Loss: {'train/loss': 0.012546704150736332, 'train/grad_norm_pre_clip': 40.29672622680664, 'timing/data': 0.0002496700035408139, 'timing/model': 0.36617759900400415, 'learning_rate/qwen_vl_interface': 4.193025563707507e-06, 'learning_rate/action_model': 4.1930255637075066e-05, 'epoch': 0.21}) +[2026-10-03 15:11:00,212][starVLA.training.train_starvla][INFO] - Step 2901, Loss: {'train/loss': 0.5044158101081848, 'train/grad_norm_pre_clip': 83.93373107910156, 'timing/data': 0.0003079699963564053, 'timing/model': 0.365479789004894, 'learning_rate/qwen_vl_interface': 4.19005671477618e-06, 'learning_rate/action_model': 4.1900567147761794e-05, 'epoch': 0.21}) +[2026-10-03 15:11:00,862][starVLA.training.train_starvla][INFO] - Step 2902, Loss: {'train/loss': 0.00021951193048153073, 'train/grad_norm_pre_clip': 0.16070309281349182, 'timing/data': 0.00020310000400058925, 'timing/model': 0.36567300898605026, 'learning_rate/qwen_vl_interface': 4.1870883015474975e-06, 'learning_rate/action_model': 4.187088301547497e-05, 'epoch': 0.21}) +[2026-10-03 15:11:01,712][starVLA.training.train_starvla][INFO] - Step 2903, Loss: {'train/loss': 2.1851226847502403e-05, 'train/grad_norm_pre_clip': 0.005621600430458784, 'timing/data': 0.0002265599905513227, 'timing/model': 0.3662956199987093, 'learning_rate/qwen_vl_interface': 4.1841203252416655e-06, 'learning_rate/action_model': 4.1841203252416655e-05, 'epoch': 0.21}) +[2026-10-03 15:11:02,361][starVLA.training.train_starvla][INFO] - Step 2904, Loss: {'train/loss': 0.008804620243608952, 'train/grad_norm_pre_clip': 37.651493072509766, 'timing/data': 0.00026682000316213816, 'timing/model': 0.36566963900986593, 'learning_rate/qwen_vl_interface': 4.181152787078705e-06, 'learning_rate/action_model': 4.1811527870787056e-05, 'epoch': 0.21}) +[2026-10-03 15:11:03,009][starVLA.training.train_starvla][INFO] - Step 2905, Loss: {'train/loss': 0.0005430279998108745, 'train/grad_norm_pre_clip': 0.20407092571258545, 'timing/data': 0.00024355000641662627, 'timing/model': 0.3647090290032793, 'learning_rate/qwen_vl_interface': 4.178185688278462e-06, 'learning_rate/action_model': 4.1781856882784616e-05, 'epoch': 0.21}) +[2026-10-03 15:11:03,656][starVLA.training.train_starvla][INFO] - Step 2906, Loss: {'train/loss': 0.0013219988904893398, 'train/grad_norm_pre_clip': 0.7847811579704285, 'timing/data': 0.00037152000004425645, 'timing/model': 0.3644473300082609, 'learning_rate/qwen_vl_interface': 4.175219030060595e-06, 'learning_rate/action_model': 4.175219030060595e-05, 'epoch': 0.21}) +[2026-10-03 15:11:04,307][starVLA.training.train_starvla][INFO] - Step 2907, Loss: {'train/loss': 2.771479921648279e-05, 'train/grad_norm_pre_clip': 0.007128862664103508, 'timing/data': 0.0002690799883566797, 'timing/model': 0.36516524000035133, 'learning_rate/qwen_vl_interface': 4.1722528136445875e-06, 'learning_rate/action_model': 4.172252813644588e-05, 'epoch': 0.21}) +[2026-10-03 15:11:05,158][starVLA.training.train_starvla][INFO] - Step 2908, Loss: {'train/loss': 1.813403287087567e-05, 'train/grad_norm_pre_clip': 0.005336361937224865, 'timing/data': 0.00024190000840462744, 'timing/model': 0.5679381619993364, 'learning_rate/qwen_vl_interface': 4.169287040249738e-06, 'learning_rate/action_model': 4.169287040249738e-05, 'epoch': 0.21}) +[2026-10-03 15:11:05,807][starVLA.training.train_starvla][INFO] - Step 2909, Loss: {'train/loss': 4.2061787098646164e-05, 'train/grad_norm_pre_clip': 0.014634679071605206, 'timing/data': 0.0003188800037605688, 'timing/model': 0.3651457689993549, 'learning_rate/qwen_vl_interface': 4.166321711095164e-06, 'learning_rate/action_model': 4.166321711095164e-05, 'epoch': 0.21}) +[2026-10-03 15:11:06,455][starVLA.training.train_starvla][INFO] - Step 2910, Loss: {'train/loss': 0.09576345235109329, 'train/grad_norm_pre_clip': 231.54507446289062, 'timing/data': 0.0002935399970738217, 'timing/model': 0.3651675999863073, 'learning_rate/qwen_vl_interface': 4.1633568273998e-06, 'learning_rate/action_model': 4.1633568273998e-05, 'epoch': 0.21}) +[2026-10-03 15:11:07,107][starVLA.training.train_starvla][INFO] - Step 2911, Loss: {'train/loss': 3.635551911429502e-05, 'train/grad_norm_pre_clip': 0.00995020940899849, 'timing/data': 0.00024792000476736575, 'timing/model': 0.3661140790063655, 'learning_rate/qwen_vl_interface': 4.1603923903823995e-06, 'learning_rate/action_model': 4.160392390382399e-05, 'epoch': 0.21}) +[2026-10-03 15:11:07,756][starVLA.training.train_starvla][INFO] - Step 2912, Loss: {'train/loss': 0.00019772002997342497, 'train/grad_norm_pre_clip': 0.2764689028263092, 'timing/data': 0.0003185599925927818, 'timing/model': 0.36586281000927556, 'learning_rate/qwen_vl_interface': 4.1574284012615276e-06, 'learning_rate/action_model': 4.157428401261528e-05, 'epoch': 0.21}) +[2026-10-03 15:11:08,565][starVLA.training.train_starvla][INFO] - Step 2913, Loss: {'train/loss': 1.5191273632808588e-05, 'train/grad_norm_pre_clip': 0.004814896732568741, 'timing/data': 0.0003743299894267693, 'timing/model': 0.5246001979976427, 'learning_rate/qwen_vl_interface': 4.154464861255571e-06, 'learning_rate/action_model': 4.154464861255571e-05, 'epoch': 0.21}) +[2026-10-03 15:11:09,214][starVLA.training.train_starvla][INFO] - Step 2914, Loss: {'train/loss': 0.30129191279411316, 'train/grad_norm_pre_clip': 93.6244125366211, 'timing/data': 0.0002360399957979098, 'timing/model': 0.3656606189906597, 'learning_rate/qwen_vl_interface': 4.151501771582728e-06, 'learning_rate/action_model': 4.151501771582728e-05, 'epoch': 0.21}) +[2026-10-03 15:11:09,866][starVLA.training.train_starvla][INFO] - Step 2915, Loss: {'train/loss': 0.5562054514884949, 'train/grad_norm_pre_clip': 22.6757755279541, 'timing/data': 0.0002928100002463907, 'timing/model': 0.3661983700003475, 'learning_rate/qwen_vl_interface': 4.148539133461012e-06, 'learning_rate/action_model': 4.1485391334610116e-05, 'epoch': 0.21}) +[2026-10-03 15:11:10,516][starVLA.training.train_starvla][INFO] - Step 2916, Loss: {'train/loss': 8.284932846436277e-06, 'train/grad_norm_pre_clip': 0.001844945247285068, 'timing/data': 0.0003157500032102689, 'timing/model': 0.36565836900263093, 'learning_rate/qwen_vl_interface': 4.145576948108254e-06, 'learning_rate/action_model': 4.145576948108253e-05, 'epoch': 0.21}) +[2026-10-03 15:11:11,143][starVLA.training.train_starvla][INFO] - Step 2917, Loss: {'train/loss': 0.00012362920097075403, 'train/grad_norm_pre_clip': 0.06869859993457794, 'timing/data': 0.00019909000548068434, 'timing/model': 0.3433680020098109, 'learning_rate/qwen_vl_interface': 4.142615216742093e-06, 'learning_rate/action_model': 4.142615216742093e-05, 'epoch': 0.21}) +[2026-10-03 15:11:11,792][starVLA.training.train_starvla][INFO] - Step 2918, Loss: {'train/loss': 3.0247676477301866e-05, 'train/grad_norm_pre_clip': 0.017487965524196625, 'timing/data': 0.00030699999479111284, 'timing/model': 0.3655082799959928, 'learning_rate/qwen_vl_interface': 4.139653940579988e-06, 'learning_rate/action_model': 4.1396539405799874e-05, 'epoch': 0.21}) +[2026-10-03 15:11:12,638][starVLA.training.train_starvla][INFO] - Step 2919, Loss: {'train/loss': 5.717097519664094e-05, 'train/grad_norm_pre_clip': 0.013593764975667, 'timing/data': 0.00024942999880295247, 'timing/model': 0.36618129900307395, 'learning_rate/qwen_vl_interface': 4.136693120839206e-06, 'learning_rate/action_model': 4.136693120839205e-05, 'epoch': 0.21}) +[2026-10-03 15:11:13,287][starVLA.training.train_starvla][INFO] - Step 2920, Loss: {'train/loss': 0.0001867323589976877, 'train/grad_norm_pre_clip': 0.10104869306087494, 'timing/data': 0.00033542000164743513, 'timing/model': 0.3653700690047117, 'learning_rate/qwen_vl_interface': 4.133732758736828e-06, 'learning_rate/action_model': 4.133732758736828e-05, 'epoch': 0.21}) +[2026-10-03 15:11:13,937][starVLA.training.train_starvla][INFO] - Step 2921, Loss: {'train/loss': 0.0001662330178078264, 'train/grad_norm_pre_clip': 0.06221235916018486, 'timing/data': 0.0003967199882026762, 'timing/model': 0.36618118900514673, 'learning_rate/qwen_vl_interface': 4.13077285548975e-06, 'learning_rate/action_model': 4.13077285548975e-05, 'epoch': 0.21}) +[2026-10-03 15:11:14,586][starVLA.training.train_starvla][INFO] - Step 2922, Loss: {'train/loss': 5.6220094847958535e-05, 'train/grad_norm_pre_clip': 0.026636401191353798, 'timing/data': 0.00024333999317605048, 'timing/model': 0.36573648899502587, 'learning_rate/qwen_vl_interface': 4.127813412314672e-06, 'learning_rate/action_model': 4.1278134123146725e-05, 'epoch': 0.21}) +[2026-10-03 15:11:15,238][starVLA.training.train_starvla][INFO] - Step 2923, Loss: {'train/loss': 0.5710768103599548, 'train/grad_norm_pre_clip': 28.62422752380371, 'timing/data': 0.0002518299879739061, 'timing/model': 0.36621789999480825, 'learning_rate/qwen_vl_interface': 4.124854430428113e-06, 'learning_rate/action_model': 4.1248544304281126e-05, 'epoch': 0.21}) +[2026-10-03 15:11:16,083][starVLA.training.train_starvla][INFO] - Step 2924, Loss: {'train/loss': 0.00010158099757973105, 'train/grad_norm_pre_clip': 0.04552074521780014, 'timing/data': 0.00023685999622102827, 'timing/model': 0.5610830730001908, 'learning_rate/qwen_vl_interface': 4.121895911046397e-06, 'learning_rate/action_model': 4.121895911046396e-05, 'epoch': 0.21}) +[2026-10-03 15:11:16,732][starVLA.training.train_starvla][INFO] - Step 2925, Loss: {'train/loss': 0.06030545383691788, 'train/grad_norm_pre_clip': 90.04714965820312, 'timing/data': 0.0003199700004188344, 'timing/model': 0.3653032800066285, 'learning_rate/qwen_vl_interface': 4.118937855385659e-06, 'learning_rate/action_model': 4.1189378553856584e-05, 'epoch': 0.21}) +[2026-10-03 15:11:17,382][starVLA.training.train_starvla][INFO] - Step 2926, Loss: {'train/loss': 0.004482431337237358, 'train/grad_norm_pre_clip': 5.535989284515381, 'timing/data': 0.0002475400106050074, 'timing/model': 0.3659032690047752, 'learning_rate/qwen_vl_interface': 4.115980264661844e-06, 'learning_rate/action_model': 4.115980264661844e-05, 'epoch': 0.21}) +[2026-10-03 15:11:18,033][starVLA.training.train_starvla][INFO] - Step 2927, Loss: {'train/loss': 0.0002281113265780732, 'train/grad_norm_pre_clip': 0.04152662307024002, 'timing/data': 0.00024820999533403665, 'timing/model': 0.36565605898795184, 'learning_rate/qwen_vl_interface': 4.113023140090709e-06, 'learning_rate/action_model': 4.113023140090709e-05, 'epoch': 0.21}) +[2026-10-03 15:11:18,683][starVLA.training.train_starvla][INFO] - Step 2928, Loss: {'train/loss': 0.0023513201158493757, 'train/grad_norm_pre_clip': 3.297889232635498, 'timing/data': 0.0002487499878043309, 'timing/model': 0.36604380000790115, 'learning_rate/qwen_vl_interface': 4.110066482887812e-06, 'learning_rate/action_model': 4.110066482887812e-05, 'epoch': 0.21}) +[2026-10-03 15:11:19,532][starVLA.training.train_starvla][INFO] - Step 2929, Loss: {'train/loss': 0.7596378922462463, 'train/grad_norm_pre_clip': 25.7182559967041, 'timing/data': 0.0003198300109943375, 'timing/model': 0.5651312609988963, 'learning_rate/qwen_vl_interface': 4.107110294268525e-06, 'learning_rate/action_model': 4.107110294268525e-05, 'epoch': 0.21}) +[2026-10-03 15:11:20,181][starVLA.training.train_starvla][INFO] - Step 2930, Loss: {'train/loss': 3.522507176967338e-05, 'train/grad_norm_pre_clip': 0.0064130863174796104, 'timing/data': 0.00024421000853180885, 'timing/model': 0.3654764100065222, 'learning_rate/qwen_vl_interface': 4.104154575448026e-06, 'learning_rate/action_model': 4.104154575448025e-05, 'epoch': 0.21}) +[2026-10-03 15:11:20,834][starVLA.training.train_starvla][INFO] - Step 2931, Loss: {'train/loss': 6.28449342912063e-05, 'train/grad_norm_pre_clip': 0.016377514228224754, 'timing/data': 0.00024931000370997936, 'timing/model': 0.36636098999588285, 'learning_rate/qwen_vl_interface': 4.101199327641297e-06, 'learning_rate/action_model': 4.101199327641297e-05, 'epoch': 0.21}) +[2026-10-03 15:11:21,460][starVLA.training.train_starvla][INFO] - Step 2932, Loss: {'train/loss': 0.3816281259059906, 'train/grad_norm_pre_clip': 63.24054718017578, 'timing/data': 0.00024701000074855983, 'timing/model': 0.3431963920011185, 'learning_rate/qwen_vl_interface': 4.098244552063133e-06, 'learning_rate/action_model': 4.098244552063132e-05, 'epoch': 0.21}) +[2026-10-03 15:11:22,110][starVLA.training.train_starvla][INFO] - Step 2933, Loss: {'train/loss': 0.35936275124549866, 'train/grad_norm_pre_clip': 39.33719253540039, 'timing/data': 0.00030996999703347683, 'timing/model': 0.3654357990017161, 'learning_rate/qwen_vl_interface': 4.0952902499281254e-06, 'learning_rate/action_model': 4.095290249928125e-05, 'epoch': 0.21}) +[2026-10-03 15:11:22,759][starVLA.training.train_starvla][INFO] - Step 2934, Loss: {'train/loss': 0.0002016987418755889, 'train/grad_norm_pre_clip': 0.045299094170331955, 'timing/data': 0.00024539999139960855, 'timing/model': 0.3657704789948184, 'learning_rate/qwen_vl_interface': 4.092336422450681e-06, 'learning_rate/action_model': 4.0923364224506804e-05, 'epoch': 0.21}) +[2026-10-03 15:11:23,586][starVLA.training.train_starvla][INFO] - Step 2935, Loss: {'train/loss': 0.00035179295809939504, 'train/grad_norm_pre_clip': 0.25970280170440674, 'timing/data': 0.00024650999694131315, 'timing/model': 0.3433817529876251, 'learning_rate/qwen_vl_interface': 4.089383070845003e-06, 'learning_rate/action_model': 4.089383070845003e-05, 'epoch': 0.21}) +[2026-10-03 15:11:24,235][starVLA.training.train_starvla][INFO] - Step 2936, Loss: {'train/loss': 4.2219326132908463e-05, 'train/grad_norm_pre_clip': 0.014250978827476501, 'timing/data': 0.00026621999859344214, 'timing/model': 0.3645515889948001, 'learning_rate/qwen_vl_interface': 4.086430196325106e-06, 'learning_rate/action_model': 4.086430196325106e-05, 'epoch': 0.21}) +[2026-10-03 15:11:24,884][starVLA.training.train_starvla][INFO] - Step 2937, Loss: {'train/loss': 0.7435664534568787, 'train/grad_norm_pre_clip': 174.5987548828125, 'timing/data': 0.0003135299921268597, 'timing/model': 0.36502923000080045, 'learning_rate/qwen_vl_interface': 4.083477800104802e-06, 'learning_rate/action_model': 4.083477800104802e-05, 'epoch': 0.21}) +[2026-10-03 15:11:25,533][starVLA.training.train_starvla][INFO] - Step 2938, Loss: {'train/loss': 0.5790340304374695, 'train/grad_norm_pre_clip': 50.85342788696289, 'timing/data': 0.00034146000689361244, 'timing/model': 0.36547068000072613, 'learning_rate/qwen_vl_interface': 4.080525883397711e-06, 'learning_rate/action_model': 4.080525883397711e-05, 'epoch': 0.22}) +[2026-10-03 15:11:26,186][starVLA.training.train_starvla][INFO] - Step 2939, Loss: {'train/loss': 0.0010305088944733143, 'train/grad_norm_pre_clip': 1.3120251893997192, 'timing/data': 0.00024990999372676015, 'timing/model': 0.36635700899933, 'learning_rate/qwen_vl_interface': 4.077574447417256e-06, 'learning_rate/action_model': 4.077574447417256e-05, 'epoch': 0.22}) +[2026-10-03 15:11:27,029][starVLA.training.train_starvla][INFO] - Step 2940, Loss: {'train/loss': 4.121673555346206e-05, 'train/grad_norm_pre_clip': 0.01453928742557764, 'timing/data': 0.00021490000654011965, 'timing/model': 0.5589821419998771, 'learning_rate/qwen_vl_interface': 4.074623493376657e-06, 'learning_rate/action_model': 4.074623493376657e-05, 'epoch': 0.22}) +[2026-10-03 15:11:27,678][starVLA.training.train_starvla][INFO] - Step 2941, Loss: {'train/loss': 0.0001011489875963889, 'train/grad_norm_pre_clip': 0.0299372561275959, 'timing/data': 0.00024593000125605613, 'timing/model': 0.36558163999870885, 'learning_rate/qwen_vl_interface': 4.071673022488944e-06, 'learning_rate/action_model': 4.071673022488944e-05, 'epoch': 0.22}) +[2026-10-03 15:11:28,328][starVLA.training.train_starvla][INFO] - Step 2942, Loss: {'train/loss': 0.0038456465117633343, 'train/grad_norm_pre_clip': 6.574202060699463, 'timing/data': 0.00032329000532627106, 'timing/model': 0.36565935901307967, 'learning_rate/qwen_vl_interface': 4.06872303596694e-06, 'learning_rate/action_model': 4.06872303596694e-05, 'epoch': 0.22}) +[2026-10-03 15:11:28,979][starVLA.training.train_starvla][INFO] - Step 2943, Loss: {'train/loss': 0.1877489686012268, 'train/grad_norm_pre_clip': 115.18195343017578, 'timing/data': 0.00023908998991828412, 'timing/model': 0.3664721889945213, 'learning_rate/qwen_vl_interface': 4.065773535023277e-06, 'learning_rate/action_model': 4.0657735350232766e-05, 'epoch': 0.22}) +[2026-10-03 15:11:29,629][starVLA.training.train_starvla][INFO] - Step 2944, Loss: {'train/loss': 0.00022544883540831506, 'train/grad_norm_pre_clip': 0.04833774268627167, 'timing/data': 0.0003271000023232773, 'timing/model': 0.36569325999880675, 'learning_rate/qwen_vl_interface': 4.062824520870382e-06, 'learning_rate/action_model': 4.062824520870381e-05, 'epoch': 0.22}) +[2026-10-03 15:11:30,279][starVLA.training.train_starvla][INFO] - Step 2945, Loss: {'train/loss': 0.00011893070040969178, 'train/grad_norm_pre_clip': 0.042648669332265854, 'timing/data': 0.00025125998945441097, 'timing/model': 0.36565377000079025, 'learning_rate/qwen_vl_interface': 4.059875994720481e-06, 'learning_rate/action_model': 4.059875994720481e-05, 'epoch': 0.22}) +[2026-10-03 15:11:31,115][starVLA.training.train_starvla][INFO] - Step 2946, Loss: {'train/loss': 9.163231879938394e-05, 'train/grad_norm_pre_clip': 0.022150926291942596, 'timing/data': 0.00030220999906305224, 'timing/model': 0.3652371200005291, 'learning_rate/qwen_vl_interface': 4.056927957785605e-06, 'learning_rate/action_model': 4.056927957785605e-05, 'epoch': 0.22}) +[2026-10-03 15:11:31,767][starVLA.training.train_starvla][INFO] - Step 2947, Loss: {'train/loss': 8.897577208699659e-05, 'train/grad_norm_pre_clip': 0.029235469177365303, 'timing/data': 0.000255279999691993, 'timing/model': 0.3665066300018225, 'learning_rate/qwen_vl_interface': 4.0539804112775785e-06, 'learning_rate/action_model': 4.053980411277578e-05, 'epoch': 0.22}) +[2026-10-03 15:11:32,416][starVLA.training.train_starvla][INFO] - Step 2948, Loss: {'train/loss': 2.598711034806911e-05, 'train/grad_norm_pre_clip': 0.005836733151227236, 'timing/data': 0.0003423400048632175, 'timing/model': 0.36555142000725027, 'learning_rate/qwen_vl_interface': 4.051033356408027e-06, 'learning_rate/action_model': 4.0510333564080264e-05, 'epoch': 0.22}) +[2026-10-03 15:11:33,043][starVLA.training.train_starvla][INFO] - Step 2949, Loss: {'train/loss': 0.351542204618454, 'train/grad_norm_pre_clip': 314.6462097167969, 'timing/data': 0.0003100589965470135, 'timing/model': 0.36509705999924336, 'learning_rate/qwen_vl_interface': 4.048086794388375e-06, 'learning_rate/action_model': 4.048086794388375e-05, 'epoch': 0.22}) +[2026-10-03 15:11:33,692][starVLA.training.train_starvla][INFO] - Step 2950, Loss: {'train/loss': 3.2155927328858525e-05, 'train/grad_norm_pre_clip': 0.005667244549840689, 'timing/data': 0.0002965900057461113, 'timing/model': 0.36542773999099154, 'learning_rate/qwen_vl_interface': 4.045140726429841e-06, 'learning_rate/action_model': 4.0451407264298406e-05, 'epoch': 0.22}) +[2026-10-03 15:11:34,537][starVLA.training.train_starvla][INFO] - Step 2951, Loss: {'train/loss': 0.33444252610206604, 'train/grad_norm_pre_clip': 116.21334075927734, 'timing/data': 0.000268358999164775, 'timing/model': 0.36620130999654066, 'learning_rate/qwen_vl_interface': 4.042195153743443e-06, 'learning_rate/action_model': 4.042195153743443e-05, 'epoch': 0.22}) +[2026-10-03 15:11:35,188][starVLA.training.train_starvla][INFO] - Step 2952, Loss: {'train/loss': 4.888081457465887e-05, 'train/grad_norm_pre_clip': 0.007917559705674648, 'timing/data': 0.000254829996265471, 'timing/model': 0.3659628699970199, 'learning_rate/qwen_vl_interface': 4.039250077539993e-06, 'learning_rate/action_model': 4.039250077539993e-05, 'epoch': 0.22}) +[2026-10-03 15:11:35,815][starVLA.training.train_starvla][INFO] - Step 2953, Loss: {'train/loss': 6.62044039927423e-05, 'train/grad_norm_pre_clip': 0.01683649979531765, 'timing/data': 0.0003301699907751754, 'timing/model': 0.34305897299782373, 'learning_rate/qwen_vl_interface': 4.036305499030104e-06, 'learning_rate/action_model': 4.0363054990301036e-05, 'epoch': 0.22}) +[2026-10-03 15:11:36,464][starVLA.training.train_starvla][INFO] - Step 2954, Loss: {'train/loss': 0.00035687422496266663, 'train/grad_norm_pre_clip': 0.05513551086187363, 'timing/data': 0.0003014800022356212, 'timing/model': 0.36541545900399797, 'learning_rate/qwen_vl_interface': 4.033361419424178e-06, 'learning_rate/action_model': 4.033361419424178e-05, 'epoch': 0.22}) +[2026-10-03 15:11:37,115][starVLA.training.train_starvla][INFO] - Step 2955, Loss: {'train/loss': 4.4357177102938294e-05, 'train/grad_norm_pre_clip': 0.012427552603185177, 'timing/data': 0.0002459899988025427, 'timing/model': 0.3657495499937795, 'learning_rate/qwen_vl_interface': 4.030417839932418e-06, 'learning_rate/action_model': 4.030417839932417e-05, 'epoch': 0.22}) +[2026-10-03 15:11:37,965][starVLA.training.train_starvla][INFO] - Step 2956, Loss: {'train/loss': 0.005898382980376482, 'train/grad_norm_pre_clip': 12.41538143157959, 'timing/data': 0.000249360004090704, 'timing/model': 0.3651035189977847, 'learning_rate/qwen_vl_interface': 4.027474761764815e-06, 'learning_rate/action_model': 4.027474761764814e-05, 'epoch': 0.22}) +[2026-10-03 15:11:38,614][starVLA.training.train_starvla][INFO] - Step 2957, Loss: {'train/loss': 0.5422273874282837, 'train/grad_norm_pre_clip': 140.9532928466797, 'timing/data': 0.0003162700013490394, 'timing/model': 0.36519987900101114, 'learning_rate/qwen_vl_interface': 4.024532186131158e-06, 'learning_rate/action_model': 4.0245321861311575e-05, 'epoch': 0.22}) +[2026-10-03 15:11:39,262][starVLA.training.train_starvla][INFO] - Step 2958, Loss: {'train/loss': 5.8241526858182624e-05, 'train/grad_norm_pre_clip': 0.020163528621196747, 'timing/data': 0.00024587998632341623, 'timing/model': 0.3649455999984639, 'learning_rate/qwen_vl_interface': 4.021590114241029e-06, 'learning_rate/action_model': 4.021590114241029e-05, 'epoch': 0.22}) +[2026-10-03 15:11:39,913][starVLA.training.train_starvla][INFO] - Step 2959, Loss: {'train/loss': 5.779611456091516e-05, 'train/grad_norm_pre_clip': 0.013542914763092995, 'timing/data': 0.00026802999491337687, 'timing/model': 0.36485651999828406, 'learning_rate/qwen_vl_interface': 4.018648547303804e-06, 'learning_rate/action_model': 4.0186485473038036e-05, 'epoch': 0.22}) +[2026-10-03 15:11:40,560][starVLA.training.train_starvla][INFO] - Step 2960, Loss: {'train/loss': 2.4616161681478843e-05, 'train/grad_norm_pre_clip': 0.0035771650727838278, 'timing/data': 0.00019751000218093395, 'timing/model': 0.36478823999641463, 'learning_rate/qwen_vl_interface': 4.015707486528649e-06, 'learning_rate/action_model': 4.015707486528649e-05, 'epoch': 0.22}) +[2026-10-03 15:11:41,405][starVLA.training.train_starvla][INFO] - Step 2961, Loss: {'train/loss': 0.0006464679609052837, 'train/grad_norm_pre_clip': 0.33912262320518494, 'timing/data': 0.0003308199957245961, 'timing/model': 0.5621958020055899, 'learning_rate/qwen_vl_interface': 4.012766933124523e-06, 'learning_rate/action_model': 4.012766933124523e-05, 'epoch': 0.22}) +[2026-10-03 15:11:42,007][starVLA.training.train_starvla][INFO] - Step 2962, Loss: {'train/loss': 0.00016530099674127996, 'train/grad_norm_pre_clip': 0.07630257308483124, 'timing/data': 0.0002797899942379445, 'timing/model': 0.340827473992249, 'learning_rate/qwen_vl_interface': 4.009826888300176e-06, 'learning_rate/action_model': 4.0098268883001755e-05, 'epoch': 0.22}) +[2026-10-03 15:11:42,656][starVLA.training.train_starvla][INFO] - Step 2963, Loss: {'train/loss': 0.000149795479956083, 'train/grad_norm_pre_clip': 0.06923074275255203, 'timing/data': 0.00030430998594965786, 'timing/model': 0.36440731899347156, 'learning_rate/qwen_vl_interface': 4.00688735326415e-06, 'learning_rate/action_model': 4.0068873532641495e-05, 'epoch': 0.22}) +[2026-10-03 15:11:43,302][starVLA.training.train_starvla][INFO] - Step 2964, Loss: {'train/loss': 9.022327867569402e-05, 'train/grad_norm_pre_clip': 0.02684696950018406, 'timing/data': 0.00024649999977555126, 'timing/model': 0.36437630000000354, 'learning_rate/qwen_vl_interface': 4.0039483292247765e-06, 'learning_rate/action_model': 4.003948329224776e-05, 'epoch': 0.22}) +[2026-10-03 15:11:43,949][starVLA.training.train_starvla][INFO] - Step 2965, Loss: {'train/loss': 0.00020577898249030113, 'train/grad_norm_pre_clip': 0.2697352170944214, 'timing/data': 0.000245339993853122, 'timing/model': 0.36455322000256274, 'learning_rate/qwen_vl_interface': 4.001009817390178e-06, 'learning_rate/action_model': 4.001009817390178e-05, 'epoch': 0.22}) +[2026-10-03 15:11:44,595][starVLA.training.train_starvla][INFO] - Step 2966, Loss: {'train/loss': 0.036116618663072586, 'train/grad_norm_pre_clip': 120.8888168334961, 'timing/data': 0.00019746999896597117, 'timing/model': 0.3640638399956515, 'learning_rate/qwen_vl_interface': 3.998071818968264e-06, 'learning_rate/action_model': 3.998071818968264e-05, 'epoch': 0.22}) +[2026-10-03 15:11:45,434][starVLA.training.train_starvla][INFO] - Step 2967, Loss: {'train/loss': 0.0003128465614281595, 'train/grad_norm_pre_clip': 0.07582475990056992, 'timing/data': 0.0002214700070908293, 'timing/model': 0.3649252989998786, 'learning_rate/qwen_vl_interface': 3.995134335166736e-06, 'learning_rate/action_model': 3.995134335166736e-05, 'epoch': 0.22}) +[2026-10-03 15:11:46,080][starVLA.training.train_starvla][INFO] - Step 2968, Loss: {'train/loss': 0.000542520429007709, 'train/grad_norm_pre_clip': 0.7851083278656006, 'timing/data': 0.00027504999889060855, 'timing/model': 0.36396429999149404, 'learning_rate/qwen_vl_interface': 3.992197367193083e-06, 'learning_rate/action_model': 3.992197367193082e-05, 'epoch': 0.22}) +[2026-10-03 15:11:46,727][starVLA.training.train_starvla][INFO] - Step 2969, Loss: {'train/loss': 0.2930905520915985, 'train/grad_norm_pre_clip': 43.04420852661133, 'timing/data': 0.00019937999604735523, 'timing/model': 0.36415423899597954, 'learning_rate/qwen_vl_interface': 3.98926091625458e-06, 'learning_rate/action_model': 3.9892609162545795e-05, 'epoch': 0.22}) +[2026-10-03 15:11:47,351][starVLA.training.train_starvla][INFO] - Step 2970, Loss: {'train/loss': 0.36635980010032654, 'train/grad_norm_pre_clip': 14.072447776794434, 'timing/data': 0.00030328999855555594, 'timing/model': 0.36406628899567295, 'learning_rate/qwen_vl_interface': 3.986324983558294e-06, 'learning_rate/action_model': 3.9863249835582934e-05, 'epoch': 0.22}) +[2026-10-03 15:11:48,000][starVLA.training.train_starvla][INFO] - Step 2971, Loss: {'train/loss': 3.91732100979425e-05, 'train/grad_norm_pre_clip': 0.007021201308816671, 'timing/data': 0.00029553999775089324, 'timing/model': 0.36470365899731405, 'learning_rate/qwen_vl_interface': 3.983389570311073e-06, 'learning_rate/action_model': 3.9833895703110724e-05, 'epoch': 0.22}) +[2026-10-03 15:11:48,845][starVLA.training.train_starvla][INFO] - Step 2972, Loss: {'train/loss': 0.0374540239572525, 'train/grad_norm_pre_clip': 90.64098358154297, 'timing/data': 0.00020296999718993902, 'timing/model': 0.36438964000262786, 'learning_rate/qwen_vl_interface': 3.9804546777195546e-06, 'learning_rate/action_model': 3.980454677719554e-05, 'epoch': 0.22}) +[2026-10-03 15:11:49,492][starVLA.training.train_starvla][INFO] - Step 2973, Loss: {'train/loss': 0.397026926279068, 'train/grad_norm_pre_clip': 14.959068298339844, 'timing/data': 0.00020004001271445304, 'timing/model': 0.36470511999505106, 'learning_rate/qwen_vl_interface': 3.977520306990164e-06, 'learning_rate/action_model': 3.977520306990163e-05, 'epoch': 0.22}) +[2026-10-03 15:11:50,116][starVLA.training.train_starvla][INFO] - Step 2974, Loss: {'train/loss': 2.9153430659789592e-05, 'train/grad_norm_pre_clip': 0.005353278946131468, 'timing/data': 0.00024375000793952495, 'timing/model': 0.3419752729969332, 'learning_rate/qwen_vl_interface': 3.974586459329109e-06, 'learning_rate/action_model': 3.9745864593291085e-05, 'epoch': 0.22}) +[2026-10-03 15:11:50,765][starVLA.training.train_starvla][INFO] - Step 2975, Loss: {'train/loss': 7.684111187700182e-05, 'train/grad_norm_pre_clip': 0.023572027683258057, 'timing/data': 0.0002575899998191744, 'timing/model': 0.3652807699982077, 'learning_rate/qwen_vl_interface': 3.9716531359423855e-06, 'learning_rate/action_model': 3.971653135942385e-05, 'epoch': 0.22}) +[2026-10-03 15:11:51,412][starVLA.training.train_starvla][INFO] - Step 2976, Loss: {'train/loss': 0.00017683155601844192, 'train/grad_norm_pre_clip': 0.12602019309997559, 'timing/data': 0.00030841000261716545, 'timing/model': 0.3648765700054355, 'learning_rate/qwen_vl_interface': 3.968720338035769e-06, 'learning_rate/action_model': 3.968720338035769e-05, 'epoch': 0.22}) +[2026-10-03 15:11:52,257][starVLA.training.train_starvla][INFO] - Step 2977, Loss: {'train/loss': 4.644082218874246e-05, 'train/grad_norm_pre_clip': 0.014040198177099228, 'timing/data': 0.00019828999938908964, 'timing/model': 0.5625216129992623, 'learning_rate/qwen_vl_interface': 3.965788066814824e-06, 'learning_rate/action_model': 3.965788066814823e-05, 'epoch': 0.22}) +[2026-10-03 15:11:52,903][starVLA.training.train_starvla][INFO] - Step 2978, Loss: {'train/loss': 5.532935756491497e-05, 'train/grad_norm_pre_clip': 0.0168380755931139, 'timing/data': 0.00024155000573955476, 'timing/model': 0.3643154689925723, 'learning_rate/qwen_vl_interface': 3.962856323484895e-06, 'learning_rate/action_model': 3.9628563234848945e-05, 'epoch': 0.22}) +[2026-10-03 15:11:53,552][starVLA.training.train_starvla][INFO] - Step 2979, Loss: {'train/loss': 8.614030957687646e-05, 'train/grad_norm_pre_clip': 0.028667647391557693, 'timing/data': 0.0002486500015947968, 'timing/model': 0.36478477000491694, 'learning_rate/qwen_vl_interface': 3.959925109251112e-06, 'learning_rate/action_model': 3.959925109251112e-05, 'epoch': 0.22}) +[2026-10-03 15:11:54,199][starVLA.training.train_starvla][INFO] - Step 2980, Loss: {'train/loss': 0.00011004378029610962, 'train/grad_norm_pre_clip': 0.06506340950727463, 'timing/data': 0.00032809999538585544, 'timing/model': 0.3641466599947307, 'learning_rate/qwen_vl_interface': 3.956994425318387e-06, 'learning_rate/action_model': 3.956994425318387e-05, 'epoch': 0.22}) +[2026-10-03 15:11:54,846][starVLA.training.train_starvla][INFO] - Step 2981, Loss: {'train/loss': 0.24954922497272491, 'train/grad_norm_pre_clip': 51.46879577636719, 'timing/data': 0.00024530000519007444, 'timing/model': 0.3645606689970009, 'learning_rate/qwen_vl_interface': 3.9540642728914115e-06, 'learning_rate/action_model': 3.9540642728914115e-05, 'epoch': 0.22}) +[2026-10-03 15:11:55,492][starVLA.training.train_starvla][INFO] - Step 2982, Loss: {'train/loss': 8.581207657698542e-05, 'train/grad_norm_pre_clip': 0.03831278160214424, 'timing/data': 0.0004399899917189032, 'timing/model': 0.36393232000409625, 'learning_rate/qwen_vl_interface': 3.951134653174664e-06, 'learning_rate/action_model': 3.951134653174663e-05, 'epoch': 0.22}) +[2026-10-03 15:11:56,316][starVLA.training.train_starvla][INFO] - Step 2983, Loss: {'train/loss': 9.445710747968405e-05, 'train/grad_norm_pre_clip': 0.06243162229657173, 'timing/data': 0.00034559999767225236, 'timing/model': 0.3414208429894643, 'learning_rate/qwen_vl_interface': 3.948205567372397e-06, 'learning_rate/action_model': 3.948205567372397e-05, 'epoch': 0.22}) +[2026-10-03 15:11:56,963][starVLA.training.train_starvla][INFO] - Step 2984, Loss: {'train/loss': 0.00031910472898744047, 'train/grad_norm_pre_clip': 0.10573028773069382, 'timing/data': 0.00025239000387955457, 'timing/model': 0.36473744999966584, 'learning_rate/qwen_vl_interface': 3.945277016688647e-06, 'learning_rate/action_model': 3.945277016688647e-05, 'epoch': 0.22}) +[2026-10-03 15:11:57,588][starVLA.training.train_starvla][INFO] - Step 2985, Loss: {'train/loss': 0.4062487781047821, 'train/grad_norm_pre_clip': 30.09549331665039, 'timing/data': 0.0002425599959678948, 'timing/model': 0.3635926400020253, 'learning_rate/qwen_vl_interface': 3.9423490023272326e-06, 'learning_rate/action_model': 3.9423490023272324e-05, 'epoch': 0.22}) +[2026-10-03 15:11:58,234][starVLA.training.train_starvla][INFO] - Step 2986, Loss: {'train/loss': 0.0005461899563670158, 'train/grad_norm_pre_clip': 0.33740749955177307, 'timing/data': 0.0002194299886468798, 'timing/model': 0.3642008889873978, 'learning_rate/qwen_vl_interface': 3.939421525491748e-06, 'learning_rate/action_model': 3.9394215254917483e-05, 'epoch': 0.22}) +[2026-10-03 15:11:58,885][starVLA.training.train_starvla][INFO] - Step 2987, Loss: {'train/loss': 0.0009764391579665244, 'train/grad_norm_pre_clip': 0.7300037145614624, 'timing/data': 0.0003311400068923831, 'timing/model': 0.36489338999672327, 'learning_rate/qwen_vl_interface': 3.9364945873855705e-06, 'learning_rate/action_model': 3.93649458738557e-05, 'epoch': 0.22}) +[2026-10-03 15:11:59,732][starVLA.training.train_starvla][INFO] - Step 2988, Loss: {'train/loss': 0.003441933775320649, 'train/grad_norm_pre_clip': 6.478848457336426, 'timing/data': 0.00019937999604735523, 'timing/model': 0.5610674920026213, 'learning_rate/qwen_vl_interface': 3.9335681892118536e-06, 'learning_rate/action_model': 3.933568189211854e-05, 'epoch': 0.22}) +[2026-10-03 15:12:00,380][starVLA.training.train_starvla][INFO] - Step 2989, Loss: {'train/loss': 0.2531103789806366, 'train/grad_norm_pre_clip': 66.50279235839844, 'timing/data': 0.00019366999913472682, 'timing/model': 0.36449736000213306, 'learning_rate/qwen_vl_interface': 3.930642332173527e-06, 'learning_rate/action_model': 3.930642332173527e-05, 'epoch': 0.22}) +[2026-10-03 15:12:01,027][starVLA.training.train_starvla][INFO] - Step 2990, Loss: {'train/loss': 0.0008274709689430892, 'train/grad_norm_pre_clip': 1.2932533025741577, 'timing/data': 0.00024254000163637102, 'timing/model': 0.3646744190045865, 'learning_rate/qwen_vl_interface': 3.927717017473301e-06, 'learning_rate/action_model': 3.9277170174733006e-05, 'epoch': 0.22}) +[2026-10-03 15:12:01,676][starVLA.training.train_starvla][INFO] - Step 2991, Loss: {'train/loss': 0.001700264518149197, 'train/grad_norm_pre_clip': 0.30556419491767883, 'timing/data': 0.00032361000194214284, 'timing/model': 0.3646774799999548, 'learning_rate/qwen_vl_interface': 3.9247922463136615e-06, 'learning_rate/action_model': 3.924792246313662e-05, 'epoch': 0.22}) +[2026-10-03 15:12:02,301][starVLA.training.train_starvla][INFO] - Step 2992, Loss: {'train/loss': 0.018562406301498413, 'train/grad_norm_pre_clip': 14.498929977416992, 'timing/data': 0.0001976100029423833, 'timing/model': 0.34210116301255766, 'learning_rate/qwen_vl_interface': 3.921868019896873e-06, 'learning_rate/action_model': 3.921868019896873e-05, 'epoch': 0.22}) +[2026-10-03 15:12:02,949][starVLA.training.train_starvla][INFO] - Step 2993, Loss: {'train/loss': 7.950349390739575e-05, 'train/grad_norm_pre_clip': 0.04175128415226936, 'timing/data': 0.00035324000054970384, 'timing/model': 0.3645053799991729, 'learning_rate/qwen_vl_interface': 3.918944339424974e-06, 'learning_rate/action_model': 3.9189443394249736e-05, 'epoch': 0.22}) +[2026-10-03 15:12:03,789][starVLA.training.train_starvla][INFO] - Step 2994, Loss: {'train/loss': 0.006272651720792055, 'train/grad_norm_pre_clip': 16.270252227783203, 'timing/data': 0.00020588000188581645, 'timing/model': 0.3647512899915455, 'learning_rate/qwen_vl_interface': 3.9160212060997774e-06, 'learning_rate/action_model': 3.9160212060997774e-05, 'epoch': 0.22}) +[2026-10-03 15:12:04,439][starVLA.training.train_starvla][INFO] - Step 2995, Loss: {'train/loss': 0.00023086539295036346, 'train/grad_norm_pre_clip': 0.06590044498443604, 'timing/data': 0.00022384000476449728, 'timing/model': 0.3654584799951408, 'learning_rate/qwen_vl_interface': 3.913098621122875e-06, 'learning_rate/action_model': 3.913098621122875e-05, 'epoch': 0.22}) +[2026-10-03 15:12:05,086][starVLA.training.train_starvla][INFO] - Step 2996, Loss: {'train/loss': 0.26743459701538086, 'train/grad_norm_pre_clip': 97.10152435302734, 'timing/data': 0.00029723999614361674, 'timing/model': 0.36463437900238205, 'learning_rate/qwen_vl_interface': 3.91017658569563e-06, 'learning_rate/action_model': 3.91017658569563e-05, 'epoch': 0.22}) +[2026-10-03 15:12:05,734][starVLA.training.train_starvla][INFO] - Step 2997, Loss: {'train/loss': 0.3484877049922943, 'train/grad_norm_pre_clip': 30.296785354614258, 'timing/data': 0.00031999999191612005, 'timing/model': 0.3650704289902933, 'learning_rate/qwen_vl_interface': 3.907255101019181e-06, 'learning_rate/action_model': 3.907255101019181e-05, 'epoch': 0.22}) +[2026-10-03 15:12:06,358][starVLA.training.train_starvla][INFO] - Step 2998, Loss: {'train/loss': 0.44729864597320557, 'train/grad_norm_pre_clip': 29.547405242919922, 'timing/data': 0.0002451489999657497, 'timing/model': 0.36405278100573923, 'learning_rate/qwen_vl_interface': 3.904334168294442e-06, 'learning_rate/action_model': 3.9043341682944415e-05, 'epoch': 0.22}) +[2026-10-03 15:12:07,209][starVLA.training.train_starvla][INFO] - Step 2999, Loss: {'train/loss': 0.0005383974057622254, 'train/grad_norm_pre_clip': 0.24884305894374847, 'timing/data': 0.00025178999931085855, 'timing/model': 0.3647926100093173, 'learning_rate/qwen_vl_interface': 3.901413788722093e-06, 'learning_rate/action_model': 3.901413788722093e-05, 'epoch': 0.22}) +[2026-10-03 15:12:07,835][starVLA.training.train_starvla][INFO] - Step 3000, Loss: {'train/loss': 0.0017898159567266703, 'train/grad_norm_pre_clip': 0.621915340423584, 'timing/data': 0.00019813000108115375, 'timing/model': 0.3647591999906581, 'learning_rate/qwen_vl_interface': 3.8984939635025966e-06, 'learning_rate/action_model': 3.898493963502596e-05, 'epoch': 0.22}) +[2026-10-03 15:12:08,483][starVLA.training.train_starvla][INFO] - Step 3001, Loss: {'train/loss': 0.0010986393317580223, 'train/grad_norm_pre_clip': 0.16801270842552185, 'timing/data': 0.0002110499917762354, 'timing/model': 0.36497552000219, 'learning_rate/qwen_vl_interface': 3.895574693836181e-06, 'learning_rate/action_model': 3.8955746938361804e-05, 'epoch': 0.22}) +[2026-10-03 15:12:09,130][starVLA.training.train_starvla][INFO] - Step 3002, Loss: {'train/loss': 0.001689153490588069, 'train/grad_norm_pre_clip': 2.394453287124634, 'timing/data': 0.0002500400005374104, 'timing/model': 0.36461027000041213, 'learning_rate/qwen_vl_interface': 3.892655980922847e-06, 'learning_rate/action_model': 3.8926559809228466e-05, 'epoch': 0.22}) +[2026-10-03 15:12:09,780][starVLA.training.train_starvla][INFO] - Step 3003, Loss: {'train/loss': 0.2765459716320038, 'train/grad_norm_pre_clip': 153.5543670654297, 'timing/data': 0.00033931000507436693, 'timing/model': 0.36527939900406636, 'learning_rate/qwen_vl_interface': 3.889737825962368e-06, 'learning_rate/action_model': 3.889737825962368e-05, 'epoch': 0.22}) +[2026-10-03 15:12:10,629][starVLA.training.train_starvla][INFO] - Step 3004, Loss: {'train/loss': 8.381587394978851e-05, 'train/grad_norm_pre_clip': 0.024770468473434448, 'timing/data': 0.00031713000498712063, 'timing/model': 0.5655435020016739, 'learning_rate/qwen_vl_interface': 3.886820230154288e-06, 'learning_rate/action_model': 3.886820230154287e-05, 'epoch': 0.22}) +[2026-10-03 15:12:11,277][starVLA.training.train_starvla][INFO] - Step 3005, Loss: {'train/loss': 0.0006518309819512069, 'train/grad_norm_pre_clip': 0.3172360360622406, 'timing/data': 0.00031481000769417733, 'timing/model': 0.3644618700054707, 'learning_rate/qwen_vl_interface': 3.8839031946979184e-06, 'learning_rate/action_model': 3.883903194697918e-05, 'epoch': 0.22}) +[2026-10-03 15:12:11,924][starVLA.training.train_starvla][INFO] - Step 3006, Loss: {'train/loss': 0.0009012625087052584, 'train/grad_norm_pre_clip': 1.0830937623977661, 'timing/data': 0.00032402999931946397, 'timing/model': 0.3642346800043015, 'learning_rate/qwen_vl_interface': 3.880986720792344e-06, 'learning_rate/action_model': 3.880986720792344e-05, 'epoch': 0.22}) +[2026-10-03 15:12:12,574][starVLA.training.train_starvla][INFO] - Step 3007, Loss: {'train/loss': 0.4049680531024933, 'train/grad_norm_pre_clip': 31.741573333740234, 'timing/data': 0.00029317999724298716, 'timing/model': 0.3651876100047957, 'learning_rate/qwen_vl_interface': 3.878070809636416e-06, 'learning_rate/action_model': 3.878070809636416e-05, 'epoch': 0.22}) +[2026-10-03 15:12:13,222][starVLA.training.train_starvla][INFO] - Step 3008, Loss: {'train/loss': 0.00014836217451374978, 'train/grad_norm_pre_clip': 0.03596924617886543, 'timing/data': 0.00033629999961704016, 'timing/model': 0.3643028700025752, 'learning_rate/qwen_vl_interface': 3.875155462428758e-06, 'learning_rate/action_model': 3.875155462428758e-05, 'epoch': 0.22}) +[2026-10-03 15:12:13,870][starVLA.training.train_starvla][INFO] - Step 3009, Loss: {'train/loss': 9.716353088151664e-05, 'train/grad_norm_pre_clip': 0.022430207580327988, 'timing/data': 0.0003259700024500489, 'timing/model': 0.36464016999525484, 'learning_rate/qwen_vl_interface': 3.872240680367754e-06, 'learning_rate/action_model': 3.872240680367754e-05, 'epoch': 0.22}) +[2026-10-03 15:12:14,712][starVLA.training.train_starvla][INFO] - Step 3010, Loss: {'train/loss': 0.47201865911483765, 'train/grad_norm_pre_clip': 19.86410140991211, 'timing/data': 0.00029587000608444214, 'timing/model': 0.36597680900013074, 'learning_rate/qwen_vl_interface': 3.869326464651566e-06, 'learning_rate/action_model': 3.869326464651566e-05, 'epoch': 0.22}) +[2026-10-03 15:12:15,364][starVLA.training.train_starvla][INFO] - Step 3011, Loss: {'train/loss': 0.0018996403086930513, 'train/grad_norm_pre_clip': 1.3010205030441284, 'timing/data': 0.0003066200006287545, 'timing/model': 0.366457910000463, 'learning_rate/qwen_vl_interface': 3.866412816478114e-06, 'learning_rate/action_model': 3.8664128164781144e-05, 'epoch': 0.22}) +[2026-10-03 15:12:16,015][starVLA.training.train_starvla][INFO] - Step 3012, Loss: {'train/loss': 0.2936374545097351, 'train/grad_norm_pre_clip': 299.17828369140625, 'timing/data': 0.00046147999819368124, 'timing/model': 0.3659294990065973, 'learning_rate/qwen_vl_interface': 3.8634997370450915e-06, 'learning_rate/action_model': 3.8634997370450914e-05, 'epoch': 0.22}) +[2026-10-03 15:12:16,643][starVLA.training.train_starvla][INFO] - Step 3013, Loss: {'train/loss': 0.009165571071207523, 'train/grad_norm_pre_clip': 25.976457595825195, 'timing/data': 0.000334120006300509, 'timing/model': 0.3655392789951293, 'learning_rate/qwen_vl_interface': 3.860587227549956e-06, 'learning_rate/action_model': 3.860587227549955e-05, 'epoch': 0.22}) +[2026-10-03 15:12:17,293][starVLA.training.train_starvla][INFO] - Step 3014, Loss: {'train/loss': 0.00047518464270979166, 'train/grad_norm_pre_clip': 0.20007845759391785, 'timing/data': 0.00023887999122962356, 'timing/model': 0.3655405699973926, 'learning_rate/qwen_vl_interface': 3.857675289189926e-06, 'learning_rate/action_model': 3.857675289189926e-05, 'epoch': 0.22}) +[2026-10-03 15:12:18,108][starVLA.training.train_starvla][INFO] - Step 3015, Loss: {'train/loss': 0.00018931154045276344, 'train/grad_norm_pre_clip': 0.09019151329994202, 'timing/data': 0.00025490000552963465, 'timing/model': 0.36673083899950143, 'learning_rate/qwen_vl_interface': 3.854763923161993e-06, 'learning_rate/action_model': 3.854763923161992e-05, 'epoch': 0.22}) +[2026-10-03 15:12:18,759][starVLA.training.train_starvla][INFO] - Step 3016, Loss: {'train/loss': 0.00010799428127938882, 'train/grad_norm_pre_clip': 0.024593669921159744, 'timing/data': 0.00020152999786660075, 'timing/model': 0.3665053199947579, 'learning_rate/qwen_vl_interface': 3.851853130662908e-06, 'learning_rate/action_model': 3.8518531306629077e-05, 'epoch': 0.22}) +[2026-10-03 15:12:19,409][starVLA.training.train_starvla][INFO] - Step 3017, Loss: {'train/loss': 0.0005268290406093001, 'train/grad_norm_pre_clip': 0.1334715336561203, 'timing/data': 0.00023924000561237335, 'timing/model': 0.3660731399868382, 'learning_rate/qwen_vl_interface': 3.8489429128891884e-06, 'learning_rate/action_model': 3.848942912889188e-05, 'epoch': 0.22}) +[2026-10-03 15:12:20,059][starVLA.training.train_starvla][INFO] - Step 3018, Loss: {'train/loss': 3.2058895158115774e-05, 'train/grad_norm_pre_clip': 0.0063142115250229836, 'timing/data': 0.00030471000354737043, 'timing/model': 0.36592296000162605, 'learning_rate/qwen_vl_interface': 3.846033271037114e-06, 'learning_rate/action_model': 3.8460332710371136e-05, 'epoch': 0.22}) +[2026-10-03 15:12:20,713][starVLA.training.train_starvla][INFO] - Step 3019, Loss: {'train/loss': 8.746593084651977e-05, 'train/grad_norm_pre_clip': 0.027416404336690903, 'timing/data': 0.00025461000041104853, 'timing/model': 0.3669034600025043, 'learning_rate/qwen_vl_interface': 3.843124206302732e-06, 'learning_rate/action_model': 3.8431242063027316e-05, 'epoch': 0.22}) +[2026-10-03 15:12:21,533][starVLA.training.train_starvla][INFO] - Step 3020, Loss: {'train/loss': 0.0003356463566888124, 'train/grad_norm_pre_clip': 0.09574146568775177, 'timing/data': 0.00019582000095397234, 'timing/model': 0.5368125559907639, 'learning_rate/qwen_vl_interface': 3.840215719881844e-06, 'learning_rate/action_model': 3.8402157198818445e-05, 'epoch': 0.22}) +[2026-10-03 15:12:22,183][starVLA.training.train_starvla][INFO] - Step 3021, Loss: {'train/loss': 0.00012294037151150405, 'train/grad_norm_pre_clip': 0.12676902115345, 'timing/data': 0.0002143299934687093, 'timing/model': 0.36594447000243235, 'learning_rate/qwen_vl_interface': 3.837307812970022e-06, 'learning_rate/action_model': 3.837307812970022e-05, 'epoch': 0.22}) +[2026-10-03 15:12:22,833][starVLA.training.train_starvla][INFO] - Step 3022, Loss: {'train/loss': 0.00013389372907113284, 'train/grad_norm_pre_clip': 0.056582506746053696, 'timing/data': 0.0003115500003332272, 'timing/model': 0.3654733900038991, 'learning_rate/qwen_vl_interface': 3.834400486762596e-06, 'learning_rate/action_model': 3.834400486762596e-05, 'epoch': 0.22}) +[2026-10-03 15:12:23,485][starVLA.training.train_starvla][INFO] - Step 3023, Loss: {'train/loss': 0.46757104992866516, 'train/grad_norm_pre_clip': 18.983354568481445, 'timing/data': 0.00020454000332392752, 'timing/model': 0.3665373690018896, 'learning_rate/qwen_vl_interface': 3.8314937424546585e-06, 'learning_rate/action_model': 3.8314937424546584e-05, 'epoch': 0.22}) +[2026-10-03 15:12:24,135][starVLA.training.train_starvla][INFO] - Step 3024, Loss: {'train/loss': 0.4298482835292816, 'train/grad_norm_pre_clip': 140.46864318847656, 'timing/data': 0.0003383699950063601, 'timing/model': 0.3656641589914216, 'learning_rate/qwen_vl_interface': 3.828587581241064e-06, 'learning_rate/action_model': 3.8285875812410635e-05, 'epoch': 0.22}) +[2026-10-03 15:12:24,785][starVLA.training.train_starvla][INFO] - Step 3025, Loss: {'train/loss': 0.0009912968380376697, 'train/grad_norm_pre_clip': 0.13724343478679657, 'timing/data': 0.0003321100084576756, 'timing/model': 0.36581646899867337, 'learning_rate/qwen_vl_interface': 3.825682004316419e-06, 'learning_rate/action_model': 3.825682004316419e-05, 'epoch': 0.22}) +[2026-10-03 15:12:25,627][starVLA.training.train_starvla][INFO] - Step 3026, Loss: {'train/loss': 0.0003182984364684671, 'train/grad_norm_pre_clip': 0.1022760197520256, 'timing/data': 0.0003024600009666756, 'timing/model': 0.36569736000092234, 'learning_rate/qwen_vl_interface': 3.822777012875104e-06, 'learning_rate/action_model': 3.822777012875103e-05, 'epoch': 0.22}) +[2026-10-03 15:12:26,257][starVLA.training.train_starvla][INFO] - Step 3027, Loss: {'train/loss': 0.0005442008259706199, 'train/grad_norm_pre_clip': 0.22021900117397308, 'timing/data': 0.00023490999592468143, 'timing/model': 0.34409959299955517, 'learning_rate/qwen_vl_interface': 3.819872608111248e-06, 'learning_rate/action_model': 3.819872608111248e-05, 'epoch': 0.22}) +[2026-10-03 15:12:26,907][starVLA.training.train_starvla][INFO] - Step 3028, Loss: {'train/loss': 0.00014974728401284665, 'train/grad_norm_pre_clip': 0.03245075047016144, 'timing/data': 0.0002674699935596436, 'timing/model': 0.3657145300094271, 'learning_rate/qwen_vl_interface': 3.816968791218741e-06, 'learning_rate/action_model': 3.816968791218741e-05, 'epoch': 0.22}) +[2026-10-03 15:12:27,558][starVLA.training.train_starvla][INFO] - Step 3029, Loss: {'train/loss': 0.00011754488514270633, 'train/grad_norm_pre_clip': 0.027482271194458008, 'timing/data': 0.0002572800003690645, 'timing/model': 0.366108059999533, 'learning_rate/qwen_vl_interface': 3.8140655633912337e-06, 'learning_rate/action_model': 3.814065563391234e-05, 'epoch': 0.22}) +[2026-10-03 15:12:28,208][starVLA.training.train_starvla][INFO] - Step 3030, Loss: {'train/loss': 0.001190120237879455, 'train/grad_norm_pre_clip': 0.5191493630409241, 'timing/data': 0.0002140099968528375, 'timing/model': 0.36646663999999873, 'learning_rate/qwen_vl_interface': 3.811162925822133e-06, 'learning_rate/action_model': 3.811162925822132e-05, 'epoch': 0.22}) +[2026-10-03 15:12:29,066][starVLA.training.train_starvla][INFO] - Step 3031, Loss: {'train/loss': 0.0004800763272214681, 'train/grad_norm_pre_clip': 0.0908367708325386, 'timing/data': 0.0003157600003760308, 'timing/model': 0.3657852710020961, 'learning_rate/qwen_vl_interface': 3.8082608797046045e-06, 'learning_rate/action_model': 3.8082608797046045e-05, 'epoch': 0.22}) +[2026-10-03 15:12:29,716][starVLA.training.train_starvla][INFO] - Step 3032, Loss: {'train/loss': 0.13013751804828644, 'train/grad_norm_pre_clip': 183.88507080078125, 'timing/data': 0.0003401700087124482, 'timing/model': 0.36577530099020805, 'learning_rate/qwen_vl_interface': 3.8053594262315687e-06, 'learning_rate/action_model': 3.805359426231568e-05, 'epoch': 0.22}) +[2026-10-03 15:12:30,344][starVLA.training.train_starvla][INFO] - Step 3033, Loss: {'train/loss': 0.00021911220392212272, 'train/grad_norm_pre_clip': 0.052670739591121674, 'timing/data': 0.0003256390045862645, 'timing/model': 0.3660158700076863, 'learning_rate/qwen_vl_interface': 3.802458566595704e-06, 'learning_rate/action_model': 3.802458566595704e-05, 'epoch': 0.22}) +[2026-10-03 15:12:30,994][starVLA.training.train_starvla][INFO] - Step 3034, Loss: {'train/loss': 4.622832784662023e-05, 'train/grad_norm_pre_clip': 0.010696320794522762, 'timing/data': 0.0003117299929726869, 'timing/model': 0.3655940499884309, 'learning_rate/qwen_vl_interface': 3.7995583019894444e-06, 'learning_rate/action_model': 3.7995583019894445e-05, 'epoch': 0.22}) +[2026-10-03 15:12:31,646][starVLA.training.train_starvla][INFO] - Step 3035, Loss: {'train/loss': 0.0001327360951108858, 'train/grad_norm_pre_clip': 0.035389095544815063, 'timing/data': 0.0003294199996162206, 'timing/model': 0.36656341000343673, 'learning_rate/qwen_vl_interface': 3.7966586336049782e-06, 'learning_rate/action_model': 3.796658633604978e-05, 'epoch': 0.22}) +[2026-10-03 15:12:32,497][starVLA.training.train_starvla][INFO] - Step 3036, Loss: {'train/loss': 0.00037214046460576355, 'train/grad_norm_pre_clip': 0.05191456153988838, 'timing/data': 0.00035319000016897917, 'timing/model': 0.566883441992104, 'learning_rate/qwen_vl_interface': 3.7937595626342516e-06, 'learning_rate/action_model': 3.793759562634252e-05, 'epoch': 0.22}) +[2026-10-03 15:12:33,126][starVLA.training.train_starvla][INFO] - Step 3037, Loss: {'train/loss': 0.28947311639785767, 'train/grad_norm_pre_clip': 126.9822769165039, 'timing/data': 0.0002473700005793944, 'timing/model': 0.36581572001159657, 'learning_rate/qwen_vl_interface': 3.79086109026896e-06, 'learning_rate/action_model': 3.7908610902689604e-05, 'epoch': 0.22}) +[2026-10-03 15:12:33,775][starVLA.training.train_starvla][INFO] - Step 3038, Loss: {'train/loss': 0.0008415278280153871, 'train/grad_norm_pre_clip': 0.22735252976417542, 'timing/data': 0.0002182000025641173, 'timing/model': 0.3658634000021266, 'learning_rate/qwen_vl_interface': 3.7879632177005584e-06, 'learning_rate/action_model': 3.787963217700558e-05, 'epoch': 0.22}) +[2026-10-03 15:12:34,428][starVLA.training.train_starvla][INFO] - Step 3039, Loss: {'train/loss': 0.0006305496208369732, 'train/grad_norm_pre_clip': 0.3523627817630768, 'timing/data': 0.0003164100053254515, 'timing/model': 0.3668597390060313, 'learning_rate/qwen_vl_interface': 3.7850659461202523e-06, 'learning_rate/action_model': 3.785065946120252e-05, 'epoch': 0.22}) +[2026-10-03 15:12:35,079][starVLA.training.train_starvla][INFO] - Step 3040, Loss: {'train/loss': 0.00017994828522205353, 'train/grad_norm_pre_clip': 0.04139191284775734, 'timing/data': 0.0003631799918366596, 'timing/model': 0.3661100199969951, 'learning_rate/qwen_vl_interface': 3.782169276719e-06, 'learning_rate/action_model': 3.782169276719e-05, 'epoch': 0.22}) +[2026-10-03 15:12:35,730][starVLA.training.train_starvla][INFO] - Step 3041, Loss: {'train/loss': 0.0006005744216963649, 'train/grad_norm_pre_clip': 0.25189223885536194, 'timing/data': 0.00035254999238532037, 'timing/model': 0.36645013000816107, 'learning_rate/qwen_vl_interface': 3.779273210687514e-06, 'learning_rate/action_model': 3.779273210687514e-05, 'epoch': 0.22}) +[2026-10-03 15:12:36,580][starVLA.training.train_starvla][INFO] - Step 3042, Loss: {'train/loss': 0.00010855171422008425, 'train/grad_norm_pre_clip': 0.03348010778427124, 'timing/data': 0.00033568000071682036, 'timing/model': 0.3663882099936018, 'learning_rate/qwen_vl_interface': 3.776377749216257e-06, 'learning_rate/action_model': 3.776377749216257e-05, 'epoch': 0.22}) +[2026-10-03 15:12:37,229][starVLA.training.train_starvla][INFO] - Step 3043, Loss: {'train/loss': 0.00019075660384260118, 'train/grad_norm_pre_clip': 0.06552200764417648, 'timing/data': 0.0003432299999985844, 'timing/model': 0.3647322799952235, 'learning_rate/qwen_vl_interface': 3.773482893495442e-06, 'learning_rate/action_model': 3.773482893495442e-05, 'epoch': 0.22}) +[2026-10-03 15:12:37,877][starVLA.training.train_starvla][INFO] - Step 3044, Loss: {'train/loss': 7.774870027787983e-05, 'train/grad_norm_pre_clip': 0.02558227814733982, 'timing/data': 0.0001971099991351366, 'timing/model': 0.3647612399945501, 'learning_rate/qwen_vl_interface': 3.7705886447150375e-06, 'learning_rate/action_model': 3.7705886447150376e-05, 'epoch': 0.22}) +[2026-10-03 15:12:38,524][starVLA.training.train_starvla][INFO] - Step 3045, Loss: {'train/loss': 0.0007341346354223788, 'train/grad_norm_pre_clip': 0.48144957423210144, 'timing/data': 0.00019811000674962997, 'timing/model': 0.36467736000486184, 'learning_rate/qwen_vl_interface': 3.767695004064759e-06, 'learning_rate/action_model': 3.767695004064759e-05, 'epoch': 0.22}) +[2026-10-03 15:12:39,171][starVLA.training.train_starvla][INFO] - Step 3046, Loss: {'train/loss': 0.00022744380112271756, 'train/grad_norm_pre_clip': 0.04673672094941139, 'timing/data': 0.00029897000058554113, 'timing/model': 0.3644184500008123, 'learning_rate/qwen_vl_interface': 3.7648019727340725e-06, 'learning_rate/action_model': 3.764801972734072e-05, 'epoch': 0.22}) +[2026-10-03 15:12:40,019][starVLA.training.train_starvla][INFO] - Step 3047, Loss: {'train/loss': 0.7560538649559021, 'train/grad_norm_pre_clip': 234.03346252441406, 'timing/data': 0.0003059900045627728, 'timing/model': 0.364447130006738, 'learning_rate/qwen_vl_interface': 3.761909551912196e-06, 'learning_rate/action_model': 3.761909551912196e-05, 'epoch': 0.22}) +[2026-10-03 15:12:40,667][starVLA.training.train_starvla][INFO] - Step 3048, Loss: {'train/loss': 0.0035053747706115246, 'train/grad_norm_pre_clip': 12.478754997253418, 'timing/data': 0.00034093999420292675, 'timing/model': 0.36468225000135135, 'learning_rate/qwen_vl_interface': 3.75901774278809e-06, 'learning_rate/action_model': 3.7590177427880904e-05, 'epoch': 0.22}) +[2026-10-03 15:12:41,314][starVLA.training.train_starvla][INFO] - Step 3049, Loss: {'train/loss': 5.677434819517657e-05, 'train/grad_norm_pre_clip': 0.014783741906285286, 'timing/data': 0.0003281399986008182, 'timing/model': 0.3644867709954269, 'learning_rate/qwen_vl_interface': 3.756126546550472e-06, 'learning_rate/action_model': 3.756126546550472e-05, 'epoch': 0.22}) +[2026-10-03 15:12:41,962][starVLA.training.train_starvla][INFO] - Step 3050, Loss: {'train/loss': 0.00011808930867118761, 'train/grad_norm_pre_clip': 0.032459404319524765, 'timing/data': 0.0002472299966029823, 'timing/model': 0.3653675999958068, 'learning_rate/qwen_vl_interface': 3.753235964387802e-06, 'learning_rate/action_model': 3.753235964387802e-05, 'epoch': 0.22}) +[2026-10-03 15:12:42,612][starVLA.training.train_starvla][INFO] - Step 3051, Loss: {'train/loss': 0.00032518102671019733, 'train/grad_norm_pre_clip': 0.07432832568883896, 'timing/data': 0.00023936999787110835, 'timing/model': 0.36527158999524545, 'learning_rate/qwen_vl_interface': 3.750345997488289e-06, 'learning_rate/action_model': 3.7503459974882886e-05, 'epoch': 0.22}) +[2026-10-03 15:12:43,460][starVLA.training.train_starvla][INFO] - Step 3052, Loss: {'train/loss': 7.831940456526354e-05, 'train/grad_norm_pre_clip': 0.020140714943408966, 'timing/data': 0.0003376300010131672, 'timing/model': 0.5651956829970004, 'learning_rate/qwen_vl_interface': 3.7474566470398903e-06, 'learning_rate/action_model': 3.74745664703989e-05, 'epoch': 0.22}) +[2026-10-03 15:12:44,106][starVLA.training.train_starvla][INFO] - Step 3053, Loss: {'train/loss': 3.3027165045496076e-05, 'train/grad_norm_pre_clip': 0.006126353517174721, 'timing/data': 0.00034458001027815044, 'timing/model': 0.36421572099789046, 'learning_rate/qwen_vl_interface': 3.744567914230307e-06, 'learning_rate/action_model': 3.744567914230307e-05, 'epoch': 0.22}) +[2026-10-03 15:12:44,753][starVLA.training.train_starvla][INFO] - Step 3054, Loss: {'train/loss': 2.7722358936443925e-05, 'train/grad_norm_pre_clip': 0.005644900258630514, 'timing/data': 0.00024448000476695597, 'timing/model': 0.36428095000155736, 'learning_rate/qwen_vl_interface': 3.7416798002469902e-06, 'learning_rate/action_model': 3.74167980024699e-05, 'epoch': 0.22}) +[2026-10-03 15:12:45,401][starVLA.training.train_starvla][INFO] - Step 3055, Loss: {'train/loss': 0.46672457456588745, 'train/grad_norm_pre_clip': 52.13890075683594, 'timing/data': 0.0003260100056650117, 'timing/model': 0.36421715999313165, 'learning_rate/qwen_vl_interface': 3.7387923062771313e-06, 'learning_rate/action_model': 3.738792306277131e-05, 'epoch': 0.22}) +[2026-10-03 15:12:46,048][starVLA.training.train_starvla][INFO] - Step 3056, Loss: {'train/loss': 0.3544149398803711, 'train/grad_norm_pre_clip': 10.864924430847168, 'timing/data': 0.00037325000448618084, 'timing/model': 0.3645209289970808, 'learning_rate/qwen_vl_interface': 3.7359054335076732e-06, 'learning_rate/action_model': 3.7359054335076734e-05, 'epoch': 0.22}) +[2026-10-03 15:12:46,674][starVLA.training.train_starvla][INFO] - Step 3057, Loss: {'train/loss': 2.5144496248685755e-05, 'train/grad_norm_pre_clip': 0.006067910231649876, 'timing/data': 0.00029984000138938427, 'timing/model': 0.36449716000061017, 'learning_rate/qwen_vl_interface': 3.7330191831253006e-06, 'learning_rate/action_model': 3.7330191831253e-05, 'epoch': 0.22}) +[2026-10-03 15:12:47,509][starVLA.training.train_starvla][INFO] - Step 3058, Loss: {'train/loss': 0.00014348301920108497, 'train/grad_norm_pre_clip': 0.037840478122234344, 'timing/data': 0.00032610000926069915, 'timing/model': 0.3642292099975748, 'learning_rate/qwen_vl_interface': 3.7301335563164382e-06, 'learning_rate/action_model': 3.730133556316438e-05, 'epoch': 0.22}) +[2026-10-03 15:12:48,158][starVLA.training.train_starvla][INFO] - Step 3059, Loss: {'train/loss': 0.00010238254617433995, 'train/grad_norm_pre_clip': 0.0781807154417038, 'timing/data': 0.0003086299984715879, 'timing/model': 0.3646232600003714, 'learning_rate/qwen_vl_interface': 3.727248554267261e-06, 'learning_rate/action_model': 3.727248554267261e-05, 'epoch': 0.22}) +[2026-10-03 15:12:48,804][starVLA.training.train_starvla][INFO] - Step 3060, Loss: {'train/loss': 7.650189945707098e-05, 'train/grad_norm_pre_clip': 0.027480674907565117, 'timing/data': 0.0003423099988140166, 'timing/model': 0.36414438900828827, 'learning_rate/qwen_vl_interface': 3.7243641781636838e-06, 'learning_rate/action_model': 3.7243641781636836e-05, 'epoch': 0.22}) +[2026-10-03 15:12:49,451][starVLA.training.train_starvla][INFO] - Step 3061, Loss: {'train/loss': 1.5102134057087824e-05, 'train/grad_norm_pre_clip': 0.002743011573329568, 'timing/data': 0.00019669000175781548, 'timing/model': 0.36462183999537956, 'learning_rate/qwen_vl_interface': 3.721480429191364e-06, 'learning_rate/action_model': 3.721480429191364e-05, 'epoch': 0.22}) +[2026-10-03 15:12:50,075][starVLA.training.train_starvla][INFO] - Step 3062, Loss: {'train/loss': 3.002403536811471e-05, 'train/grad_norm_pre_clip': 0.006207448896020651, 'timing/data': 0.0003071200044360012, 'timing/model': 0.36414744101057295, 'learning_rate/qwen_vl_interface': 3.7185973085357045e-06, 'learning_rate/action_model': 3.718597308535705e-05, 'epoch': 0.22}) +[2026-10-03 15:12:50,925][starVLA.training.train_starvla][INFO] - Step 3063, Loss: {'train/loss': 0.0560191348195076, 'train/grad_norm_pre_clip': 183.64137268066406, 'timing/data': 0.00031568000849802047, 'timing/model': 0.3647864709928399, 'learning_rate/qwen_vl_interface': 3.7157148173818443e-06, 'learning_rate/action_model': 3.715714817381844e-05, 'epoch': 0.22}) +[2026-10-03 15:12:51,572][starVLA.training.train_starvla][INFO] - Step 3064, Loss: {'train/loss': 1.538517426524777e-05, 'train/grad_norm_pre_clip': 0.002663401886820793, 'timing/data': 0.0002413100010016933, 'timing/model': 0.3645248099928722, 'learning_rate/qwen_vl_interface': 3.7128329569146672e-06, 'learning_rate/action_model': 3.7128329569146666e-05, 'epoch': 0.22}) +[2026-10-03 15:12:52,219][starVLA.training.train_starvla][INFO] - Step 3065, Loss: {'train/loss': 0.00021159989410080016, 'train/grad_norm_pre_clip': 0.13391046226024628, 'timing/data': 0.00019746999896597117, 'timing/model': 0.3645083300070837, 'learning_rate/qwen_vl_interface': 3.7099517283187977e-06, 'learning_rate/action_model': 3.7099517283187976e-05, 'epoch': 0.22}) +[2026-10-03 15:12:52,866][starVLA.training.train_starvla][INFO] - Step 3066, Loss: {'train/loss': 6.487412611022592e-05, 'train/grad_norm_pre_clip': 0.024505119770765305, 'timing/data': 0.00032253000244963914, 'timing/model': 0.3637484800128732, 'learning_rate/qwen_vl_interface': 3.7070711327786e-06, 'learning_rate/action_model': 3.7070711327785997e-05, 'epoch': 0.22}) +[2026-10-03 15:12:53,514][starVLA.training.train_starvla][INFO] - Step 3067, Loss: {'train/loss': 1.3142611351213418e-05, 'train/grad_norm_pre_clip': 0.0027756288181990385, 'timing/data': 0.00024793899501673877, 'timing/model': 0.36466204999305774, 'learning_rate/qwen_vl_interface': 3.704191171478179e-06, 'learning_rate/action_model': 3.704191171478179e-05, 'epoch': 0.22}) +[2026-10-03 15:12:54,354][starVLA.training.train_starvla][INFO] - Step 3068, Loss: {'train/loss': 0.20255695283412933, 'train/grad_norm_pre_clip': 62.58626174926758, 'timing/data': 0.0002416200004518032, 'timing/model': 0.5576056440040702, 'learning_rate/qwen_vl_interface': 3.701311845601376e-06, 'learning_rate/action_model': 3.701311845601376e-05, 'epoch': 0.22}) +[2026-10-03 15:12:55,001][starVLA.training.train_starvla][INFO] - Step 3069, Loss: {'train/loss': 0.5508129000663757, 'train/grad_norm_pre_clip': 48.63923645019531, 'timing/data': 0.00024228999973274767, 'timing/model': 0.3645728400006192, 'learning_rate/qwen_vl_interface': 3.6984331563317744e-06, 'learning_rate/action_model': 3.698433156331774e-05, 'epoch': 0.22}) +[2026-10-03 15:12:55,647][starVLA.training.train_starvla][INFO] - Step 3070, Loss: {'train/loss': 3.270590241299942e-05, 'train/grad_norm_pre_clip': 0.011593937873840332, 'timing/data': 0.00019784999312832952, 'timing/model': 0.3643464899942046, 'learning_rate/qwen_vl_interface': 3.695555104852695e-06, 'learning_rate/action_model': 3.6955551048526946e-05, 'epoch': 0.22}) +[2026-10-03 15:12:56,274][starVLA.training.train_starvla][INFO] - Step 3071, Loss: {'train/loss': 4.575095226755366e-05, 'train/grad_norm_pre_clip': 0.011606303974986076, 'timing/data': 0.00021741000819019973, 'timing/model': 0.364647939990391, 'learning_rate/qwen_vl_interface': 3.6926776923471948e-06, 'learning_rate/action_model': 3.6926776923471946e-05, 'epoch': 0.22}) +[2026-10-03 15:12:56,921][starVLA.training.train_starvla][INFO] - Step 3072, Loss: {'train/loss': 6.559888424817473e-05, 'train/grad_norm_pre_clip': 0.04434308037161827, 'timing/data': 0.00019827000505756587, 'timing/model': 0.36470861099951435, 'learning_rate/qwen_vl_interface': 3.6898009199980715e-06, 'learning_rate/action_model': 3.6898009199980714e-05, 'epoch': 0.22}) +[2026-10-03 15:12:57,568][starVLA.training.train_starvla][INFO] - Step 3073, Loss: {'train/loss': 0.049304861575365067, 'train/grad_norm_pre_clip': 55.69441604614258, 'timing/data': 0.0001983800029847771, 'timing/model': 0.3646401800069725, 'learning_rate/qwen_vl_interface': 3.686924788987855e-06, 'learning_rate/action_model': 3.686924788987855e-05, 'epoch': 0.22}) +[2026-10-03 15:12:58,407][starVLA.training.train_starvla][INFO] - Step 3074, Loss: {'train/loss': 0.0011484634596854448, 'train/grad_norm_pre_clip': 0.22484458982944489, 'timing/data': 0.0003415399987716228, 'timing/model': 0.36316491100296844, 'learning_rate/qwen_vl_interface': 3.6840493004988155e-06, 'learning_rate/action_model': 3.684049300498815e-05, 'epoch': 0.22}) +[2026-10-03 15:12:59,055][starVLA.training.train_starvla][INFO] - Step 3075, Loss: {'train/loss': 0.0004969279980286956, 'train/grad_norm_pre_clip': 1.6694376468658447, 'timing/data': 0.00031261000549420714, 'timing/model': 0.3641507699940121, 'learning_rate/qwen_vl_interface': 3.681174455712958e-06, 'learning_rate/action_model': 3.681174455712958e-05, 'epoch': 0.23}) +[2026-10-03 15:12:59,679][starVLA.training.train_starvla][INFO] - Step 3076, Loss: {'train/loss': 6.936577119631693e-05, 'train/grad_norm_pre_clip': 0.027436429634690285, 'timing/data': 0.00031737999233882874, 'timing/model': 0.3636590009991778, 'learning_rate/qwen_vl_interface': 3.6783002558120206e-06, 'learning_rate/action_model': 3.67830025581202e-05, 'epoch': 0.23}) +[2026-10-03 15:13:00,326][starVLA.training.train_starvla][INFO] - Step 3077, Loss: {'train/loss': 8.765609527472407e-05, 'train/grad_norm_pre_clip': 0.029306963086128235, 'timing/data': 0.0002530799974920228, 'timing/model': 0.36410897999303415, 'learning_rate/qwen_vl_interface': 3.6754267019774804e-06, 'learning_rate/action_model': 3.67542670197748e-05, 'epoch': 0.23}) +[2026-10-03 15:13:00,971][starVLA.training.train_starvla][INFO] - Step 3078, Loss: {'train/loss': 0.0008860538364388049, 'train/grad_norm_pre_clip': 0.18986456096172333, 'timing/data': 0.0003207299887435511, 'timing/model': 0.36375101999146864, 'learning_rate/qwen_vl_interface': 3.6725537953905453e-06, 'learning_rate/action_model': 3.672553795390545e-05, 'epoch': 0.23}) +[2026-10-03 15:13:01,814][starVLA.training.train_starvla][INFO] - Step 3079, Loss: {'train/loss': 0.00048603658797219396, 'train/grad_norm_pre_clip': 0.48482745885849, 'timing/data': 0.00028570900030899793, 'timing/model': 0.36391418099810835, 'learning_rate/qwen_vl_interface': 3.6696815372321603e-06, 'learning_rate/action_model': 3.6696815372321604e-05, 'epoch': 0.23}) +[2026-10-03 15:13:02,459][starVLA.training.train_starvla][INFO] - Step 3080, Loss: {'train/loss': 0.0004605258291121572, 'train/grad_norm_pre_clip': 0.32377541065216064, 'timing/data': 0.00024044000019785017, 'timing/model': 0.36415511000086553, 'learning_rate/qwen_vl_interface': 3.6668099286830023e-06, 'learning_rate/action_model': 3.666809928683002e-05, 'epoch': 0.23}) +[2026-10-03 15:13:03,106][starVLA.training.train_starvla][INFO] - Step 3081, Loss: {'train/loss': 0.00014885283599141985, 'train/grad_norm_pre_clip': 0.13177528977394104, 'timing/data': 0.00019849000091198832, 'timing/model': 0.3640082799975062, 'learning_rate/qwen_vl_interface': 3.66393897092348e-06, 'learning_rate/action_model': 3.66393897092348e-05, 'epoch': 0.23}) +[2026-10-03 15:13:03,750][starVLA.training.train_starvla][INFO] - Step 3082, Loss: {'train/loss': 0.0010837375884875655, 'train/grad_norm_pre_clip': 0.335440993309021, 'timing/data': 0.0003428900090511888, 'timing/model': 0.36316292099945713, 'learning_rate/qwen_vl_interface': 3.661068665133735e-06, 'learning_rate/action_model': 3.661068665133735e-05, 'epoch': 0.23}) +[2026-10-03 15:13:04,398][starVLA.training.train_starvla][INFO] - Step 3083, Loss: {'train/loss': 0.00010573276085779071, 'train/grad_norm_pre_clip': 0.03589170053601265, 'timing/data': 0.00021208998805377632, 'timing/model': 0.3641728009970393, 'learning_rate/qwen_vl_interface': 3.6581990124936434e-06, 'learning_rate/action_model': 3.6581990124936434e-05, 'epoch': 0.23}) +[2026-10-03 15:13:05,245][starVLA.training.train_starvla][INFO] - Step 3084, Loss: {'train/loss': 2.7826707082567737e-05, 'train/grad_norm_pre_clip': 0.008389163762331009, 'timing/data': 0.00036123998870607466, 'timing/model': 0.36404480000783224, 'learning_rate/qwen_vl_interface': 3.65533001418281e-06, 'learning_rate/action_model': 3.655330014182809e-05, 'epoch': 0.23}) +[2026-10-03 15:13:05,870][starVLA.training.train_starvla][INFO] - Step 3085, Loss: {'train/loss': 1.2814767615054734e-05, 'train/grad_norm_pre_clip': 0.003414493054151535, 'timing/data': 0.0002483199932612479, 'timing/model': 0.3419904340116773, 'learning_rate/qwen_vl_interface': 3.652461671380572e-06, 'learning_rate/action_model': 3.652461671380572e-05, 'epoch': 0.23}) +[2026-10-03 15:13:06,516][starVLA.training.train_starvla][INFO] - Step 3086, Loss: {'train/loss': 0.44079938530921936, 'train/grad_norm_pre_clip': 33.009063720703125, 'timing/data': 0.0003161000058753416, 'timing/model': 0.3643331499915803, 'learning_rate/qwen_vl_interface': 3.6495939852659966e-06, 'learning_rate/action_model': 3.649593985265997e-05, 'epoch': 0.23}) +[2026-10-03 15:13:07,140][starVLA.training.train_starvla][INFO] - Step 3087, Loss: {'train/loss': 1.1660012205538806e-05, 'train/grad_norm_pre_clip': 0.003445202484726906, 'timing/data': 0.00020080000103916973, 'timing/model': 0.3419635530008236, 'learning_rate/qwen_vl_interface': 3.646726957017881e-06, 'learning_rate/action_model': 3.646726957017881e-05, 'epoch': 0.23}) +[2026-10-03 15:13:07,787][starVLA.training.train_starvla][INFO] - Step 3088, Loss: {'train/loss': 0.002268990036100149, 'train/grad_norm_pre_clip': 11.18762493133545, 'timing/data': 0.00019271000928711146, 'timing/model': 0.3646637100027874, 'learning_rate/qwen_vl_interface': 3.643860587814752e-06, 'learning_rate/action_model': 3.643860587814752e-05, 'epoch': 0.23}) +[2026-10-03 15:13:08,599][starVLA.training.train_starvla][INFO] - Step 3089, Loss: {'train/loss': 0.00012192645226605237, 'train/grad_norm_pre_clip': 0.06045253947377205, 'timing/data': 0.0003267299907747656, 'timing/model': 0.529199458003859, 'learning_rate/qwen_vl_interface': 3.640994878834866e-06, 'learning_rate/action_model': 3.640994878834866e-05, 'epoch': 0.23}) +[2026-10-03 15:13:09,245][starVLA.training.train_starvla][INFO] - Step 3090, Loss: {'train/loss': 0.0022574150934815407, 'train/grad_norm_pre_clip': 3.4088103771209717, 'timing/data': 0.0002461700059939176, 'timing/model': 0.3640743699943414, 'learning_rate/qwen_vl_interface': 3.6381298312562095e-06, 'learning_rate/action_model': 3.6381298312562094e-05, 'epoch': 0.23}) +[2026-10-03 15:13:09,894][starVLA.training.train_starvla][INFO] - Step 3091, Loss: {'train/loss': 0.00021662549988832325, 'train/grad_norm_pre_clip': 0.11681145429611206, 'timing/data': 0.00024388000019825995, 'timing/model': 0.3644292599929031, 'learning_rate/qwen_vl_interface': 3.635265446256493e-06, 'learning_rate/action_model': 3.6352654462564926e-05, 'epoch': 0.23}) +[2026-10-03 15:13:10,540][starVLA.training.train_starvla][INFO] - Step 3092, Loss: {'train/loss': 0.00021117294090799987, 'train/grad_norm_pre_clip': 0.07170210033655167, 'timing/data': 0.00024894998932722956, 'timing/model': 0.3638014199968893, 'learning_rate/qwen_vl_interface': 3.6324017250131565e-06, 'learning_rate/action_model': 3.6324017250131567e-05, 'epoch': 0.23}) +[2026-10-03 15:13:11,164][starVLA.training.train_starvla][INFO] - Step 3093, Loss: {'train/loss': 4.9619768105912954e-05, 'train/grad_norm_pre_clip': 0.024379704147577286, 'timing/data': 0.00030235000303946435, 'timing/model': 0.3639978599967435, 'learning_rate/qwen_vl_interface': 3.6295386687033696e-06, 'learning_rate/action_model': 3.6295386687033695e-05, 'epoch': 0.23}) +[2026-10-03 15:13:11,810][starVLA.training.train_starvla][INFO] - Step 3094, Loss: {'train/loss': 0.0004123672842979431, 'train/grad_norm_pre_clip': 0.06517677754163742, 'timing/data': 0.000232149992370978, 'timing/model': 0.3643800809950335, 'learning_rate/qwen_vl_interface': 3.6266762785040253e-06, 'learning_rate/action_model': 3.626676278504025e-05, 'epoch': 0.23}) +[2026-10-03 15:13:12,661][starVLA.training.train_starvla][INFO] - Step 3095, Loss: {'train/loss': 0.00046582083450630307, 'train/grad_norm_pre_clip': 1.3067247867584229, 'timing/data': 0.0002599300059955567, 'timing/model': 0.3657461699913256, 'learning_rate/qwen_vl_interface': 3.6238145555917464e-06, 'learning_rate/action_model': 3.6238145555917465e-05, 'epoch': 0.23}) +[2026-10-03 15:13:13,311][starVLA.training.train_starvla][INFO] - Step 3096, Loss: {'train/loss': 8.247739060607273e-06, 'train/grad_norm_pre_clip': 0.0015371752670034766, 'timing/data': 0.0003269600128987804, 'timing/model': 0.365743129994371, 'learning_rate/qwen_vl_interface': 3.6209535011428757e-06, 'learning_rate/action_model': 3.620953501142876e-05, 'epoch': 0.23}) +[2026-10-03 15:13:13,959][starVLA.training.train_starvla][INFO] - Step 3097, Loss: {'train/loss': 0.00011367532715667039, 'train/grad_norm_pre_clip': 0.05127698928117752, 'timing/data': 0.00025892999838106334, 'timing/model': 0.3651004190032836, 'learning_rate/qwen_vl_interface': 3.6180931163334863e-06, 'learning_rate/action_model': 3.6180931163334863e-05, 'epoch': 0.23}) +[2026-10-03 15:13:14,607][starVLA.training.train_starvla][INFO] - Step 3098, Loss: {'train/loss': 0.060899633914232254, 'train/grad_norm_pre_clip': 310.7099609375, 'timing/data': 0.00020075000065844506, 'timing/model': 0.3656612099875929, 'learning_rate/qwen_vl_interface': 3.6152334023393747e-06, 'learning_rate/action_model': 3.6152334023393745e-05, 'epoch': 0.23}) +[2026-10-03 15:13:15,259][starVLA.training.train_starvla][INFO] - Step 3099, Loss: {'train/loss': 0.003468814305961132, 'train/grad_norm_pre_clip': 14.061534881591797, 'timing/data': 0.0003059699956793338, 'timing/model': 0.3659239099943079, 'learning_rate/qwen_vl_interface': 3.6123743603360613e-06, 'learning_rate/action_model': 3.612374360336061e-05, 'epoch': 0.23}) +[2026-10-03 15:13:16,100][starVLA.training.train_starvla][INFO] - Step 3100, Loss: {'train/loss': 0.18850041925907135, 'train/grad_norm_pre_clip': 143.80874633789062, 'timing/data': 0.0003608699917094782, 'timing/model': 0.5563225839869119, 'learning_rate/qwen_vl_interface': 3.6095159914987893e-06, 'learning_rate/action_model': 3.6095159914987894e-05, 'epoch': 0.23}) +[2026-10-03 15:13:16,749][starVLA.training.train_starvla][INFO] - Step 3101, Loss: {'train/loss': 0.3539867401123047, 'train/grad_norm_pre_clip': 72.63121795654297, 'timing/data': 0.00019780000729952008, 'timing/model': 0.36566070999833755, 'learning_rate/qwen_vl_interface': 3.6066582970025266e-06, 'learning_rate/action_model': 3.6066582970025264e-05, 'epoch': 0.23}) +[2026-10-03 15:13:17,397][starVLA.training.train_starvla][INFO] - Step 3102, Loss: {'train/loss': 0.4039122462272644, 'train/grad_norm_pre_clip': 70.43269348144531, 'timing/data': 0.0002001799875870347, 'timing/model': 0.3655546200025128, 'learning_rate/qwen_vl_interface': 3.603801278021964e-06, 'learning_rate/action_model': 3.6038012780219635e-05, 'epoch': 0.23}) +[2026-10-03 15:13:18,048][starVLA.training.train_starvla][INFO] - Step 3103, Loss: {'train/loss': 4.133379479753785e-05, 'train/grad_norm_pre_clip': 0.041696298867464066, 'timing/data': 0.00024172000121325254, 'timing/model': 0.36553450099017937, 'learning_rate/qwen_vl_interface': 3.6009449357315122e-06, 'learning_rate/action_model': 3.6009449357315125e-05, 'epoch': 0.23}) +[2026-10-03 15:13:18,675][starVLA.training.train_starvla][INFO] - Step 3104, Loss: {'train/loss': 0.00014293573622126132, 'train/grad_norm_pre_clip': 0.13656924664974213, 'timing/data': 0.00019499000336509198, 'timing/model': 0.3434322329994757, 'learning_rate/qwen_vl_interface': 3.5980892713053084e-06, 'learning_rate/action_model': 3.598089271305308e-05, 'epoch': 0.23}) +[2026-10-03 15:13:19,514][starVLA.training.train_starvla][INFO] - Step 3105, Loss: {'train/loss': 0.81704181432724, 'train/grad_norm_pre_clip': 206.96604919433594, 'timing/data': 0.00020058998779859394, 'timing/model': 0.5561034540005494, 'learning_rate/qwen_vl_interface': 3.5952342859172063e-06, 'learning_rate/action_model': 3.595234285917206e-05, 'epoch': 0.23}) +[2026-10-03 15:13:20,163][starVLA.training.train_starvla][INFO] - Step 3106, Loss: {'train/loss': 0.19439828395843506, 'train/grad_norm_pre_clip': 139.1326141357422, 'timing/data': 0.00019987000268884003, 'timing/model': 0.3657243099878542, 'learning_rate/qwen_vl_interface': 3.592379980740785e-06, 'learning_rate/action_model': 3.5923799807407845e-05, 'epoch': 0.23}) +[2026-10-03 15:13:20,815][starVLA.training.train_starvla][INFO] - Step 3107, Loss: {'train/loss': 0.0005947091849520802, 'train/grad_norm_pre_clip': 0.10006579756736755, 'timing/data': 0.0002069100009975955, 'timing/model': 0.36594782999600284, 'learning_rate/qwen_vl_interface': 3.5895263569493384e-06, 'learning_rate/action_model': 3.589526356949338e-05, 'epoch': 0.23}) +[2026-10-03 15:13:21,464][starVLA.training.train_starvla][INFO] - Step 3108, Loss: {'train/loss': 3.9163787732832134e-05, 'train/grad_norm_pre_clip': 0.013982253149151802, 'timing/data': 0.0003146399976685643, 'timing/model': 0.36579876000178047, 'learning_rate/qwen_vl_interface': 3.586673415715885e-06, 'learning_rate/action_model': 3.586673415715885e-05, 'epoch': 0.23}) +[2026-10-03 15:13:22,115][starVLA.training.train_starvla][INFO] - Step 3109, Loss: {'train/loss': 2.2097297915024683e-05, 'train/grad_norm_pre_clip': 0.006821570452302694, 'timing/data': 0.00030047999462112784, 'timing/model': 0.3658300200040685, 'learning_rate/qwen_vl_interface': 3.5838211582131603e-06, 'learning_rate/action_model': 3.58382115821316e-05, 'epoch': 0.23}) +[2026-10-03 15:13:22,958][starVLA.training.train_starvla][INFO] - Step 3110, Loss: {'train/loss': 0.0006411444628611207, 'train/grad_norm_pre_clip': 0.1389043927192688, 'timing/data': 0.0002399299992248416, 'timing/model': 0.559619494000799, 'learning_rate/qwen_vl_interface': 3.580969585613622e-06, 'learning_rate/action_model': 3.580969585613622e-05, 'epoch': 0.23}) +[2026-10-03 15:13:23,609][starVLA.training.train_starvla][INFO] - Step 3111, Loss: {'train/loss': 0.0006050360389053822, 'train/grad_norm_pre_clip': 0.12057583779096603, 'timing/data': 0.00031118000333663076, 'timing/model': 0.3657558499980951, 'learning_rate/qwen_vl_interface': 3.5781186990894423e-06, 'learning_rate/action_model': 3.578118699089442e-05, 'epoch': 0.23}) +[2026-10-03 15:13:24,257][starVLA.training.train_starvla][INFO] - Step 3112, Loss: {'train/loss': 2.444443089189008e-05, 'train/grad_norm_pre_clip': 0.006273861974477768, 'timing/data': 0.00019729899941012263, 'timing/model': 0.36509184099850245, 'learning_rate/qwen_vl_interface': 3.5752684998125113e-06, 'learning_rate/action_model': 3.5752684998125116e-05, 'epoch': 0.23}) +[2026-10-03 15:13:24,906][starVLA.training.train_starvla][INFO] - Step 3113, Loss: {'train/loss': 7.770811498630792e-05, 'train/grad_norm_pre_clip': 0.022783752530813217, 'timing/data': 0.00035521999234333634, 'timing/model': 0.36487228999612853, 'learning_rate/qwen_vl_interface': 3.5724189889544407e-06, 'learning_rate/action_model': 3.572418988954441e-05, 'epoch': 0.23}) +[2026-10-03 15:13:25,555][starVLA.training.train_starvla][INFO] - Step 3114, Loss: {'train/loss': 4.582168548949994e-05, 'train/grad_norm_pre_clip': 0.01937561295926571, 'timing/data': 0.00024022000434342772, 'timing/model': 0.365618911004276, 'learning_rate/qwen_vl_interface': 3.569570167686556e-06, 'learning_rate/action_model': 3.569570167686556e-05, 'epoch': 0.23}) +[2026-10-03 15:13:26,408][starVLA.training.train_starvla][INFO] - Step 3115, Loss: {'train/loss': 3.871753142448142e-05, 'train/grad_norm_pre_clip': 0.014354034326970577, 'timing/data': 0.000325209999573417, 'timing/model': 0.56669839299866, 'learning_rate/qwen_vl_interface': 3.5667220371798995e-06, 'learning_rate/action_model': 3.566722037179899e-05, 'epoch': 0.23}) +[2026-10-03 15:13:27,057][starVLA.training.train_starvla][INFO] - Step 3116, Loss: {'train/loss': 2.023531123995781e-05, 'train/grad_norm_pre_clip': 0.004782519768923521, 'timing/data': 0.0002711200068006292, 'timing/model': 0.3655311799957417, 'learning_rate/qwen_vl_interface': 3.563874598605232e-06, 'learning_rate/action_model': 3.5638745986052316e-05, 'epoch': 0.23}) +[2026-10-03 15:13:27,706][starVLA.training.train_starvla][INFO] - Step 3117, Loss: {'train/loss': 1.0058147381641902e-05, 'train/grad_norm_pre_clip': 0.002578816842287779, 'timing/data': 0.0003280200035078451, 'timing/model': 0.36561358999460936, 'learning_rate/qwen_vl_interface': 3.561027853133025e-06, 'learning_rate/action_model': 3.561027853133025e-05, 'epoch': 0.23}) +[2026-10-03 15:13:28,355][starVLA.training.train_starvla][INFO] - Step 3118, Loss: {'train/loss': 0.003952711820602417, 'train/grad_norm_pre_clip': 3.526292562484741, 'timing/data': 0.0003393200022401288, 'timing/model': 0.3653821100015193, 'learning_rate/qwen_vl_interface': 3.558181801933472e-06, 'learning_rate/action_model': 3.5581818019334714e-05, 'epoch': 0.23}) +[2026-10-03 15:13:29,007][starVLA.training.train_starvla][INFO] - Step 3119, Loss: {'train/loss': 0.0002269506803713739, 'train/grad_norm_pre_clip': 0.1807563602924347, 'timing/data': 0.00030636000155936927, 'timing/model': 0.36587145998782944, 'learning_rate/qwen_vl_interface': 3.555336446176473e-06, 'learning_rate/action_model': 3.5553364461764726e-05, 'epoch': 0.23}) +[2026-10-03 15:13:29,656][starVLA.training.train_starvla][INFO] - Step 3120, Loss: {'train/loss': 0.02648201957345009, 'train/grad_norm_pre_clip': 129.77471923828125, 'timing/data': 0.00030725999386049807, 'timing/model': 0.3650884300004691, 'learning_rate/qwen_vl_interface': 3.552491787031647e-06, 'learning_rate/action_model': 3.552491787031647e-05, 'epoch': 0.23}) +[2026-10-03 15:13:30,496][starVLA.training.train_starvla][INFO] - Step 3121, Loss: {'train/loss': 7.536751218140125e-05, 'train/grad_norm_pre_clip': 0.03751887381076813, 'timing/data': 0.0002169700019294396, 'timing/model': 0.36543170000368264, 'learning_rate/qwen_vl_interface': 3.5496478256683277e-06, 'learning_rate/action_model': 3.549647825668327e-05, 'epoch': 0.23}) +[2026-10-03 15:13:31,145][starVLA.training.train_starvla][INFO] - Step 3122, Loss: {'train/loss': 0.0003326494770590216, 'train/grad_norm_pre_clip': 0.6927018761634827, 'timing/data': 0.00020013999892398715, 'timing/model': 0.3653740599984303, 'learning_rate/qwen_vl_interface': 3.5468045632555585e-06, 'learning_rate/action_model': 3.546804563255559e-05, 'epoch': 0.23}) +[2026-10-03 15:13:31,796][starVLA.training.train_starvla][INFO] - Step 3123, Loss: {'train/loss': 6.466428749263287e-05, 'train/grad_norm_pre_clip': 0.021199693903326988, 'timing/data': 0.00020680000307038426, 'timing/model': 0.365983089999645, 'learning_rate/qwen_vl_interface': 3.5439620009620995e-06, 'learning_rate/action_model': 3.543962000962099e-05, 'epoch': 0.23}) +[2026-10-03 15:13:32,401][starVLA.training.train_starvla][INFO] - Step 3124, Loss: {'train/loss': 2.1448340703500435e-05, 'train/grad_norm_pre_clip': 0.007862822152674198, 'timing/data': 0.0003645999968284741, 'timing/model': 0.3430317530001048, 'learning_rate/qwen_vl_interface': 3.541120139956419e-06, 'learning_rate/action_model': 3.5411201399564184e-05, 'epoch': 0.23}) +[2026-10-03 15:13:33,050][starVLA.training.train_starvla][INFO] - Step 3125, Loss: {'train/loss': 0.3023582696914673, 'train/grad_norm_pre_clip': 36.718013763427734, 'timing/data': 0.00020088999008294195, 'timing/model': 0.36526459000015166, 'learning_rate/qwen_vl_interface': 3.538278981406697e-06, 'learning_rate/action_model': 3.5382789814066966e-05, 'epoch': 0.23}) +[2026-10-03 15:13:33,888][starVLA.training.train_starvla][INFO] - Step 3126, Loss: {'train/loss': 6.034929356246721e-06, 'train/grad_norm_pre_clip': 0.0012942245230078697, 'timing/data': 0.00019987999985460192, 'timing/model': 0.5542534040141618, 'learning_rate/qwen_vl_interface': 3.5354385264808295e-06, 'learning_rate/action_model': 3.535438526480829e-05, 'epoch': 0.23}) +[2026-10-03 15:13:34,539][starVLA.training.train_starvla][INFO] - Step 3127, Loss: {'train/loss': 0.00012734120537061244, 'train/grad_norm_pre_clip': 0.19682805240154266, 'timing/data': 0.0003189100098097697, 'timing/model': 0.3661510799865937, 'learning_rate/qwen_vl_interface': 3.5325987763464196e-06, 'learning_rate/action_model': 3.5325987763464194e-05, 'epoch': 0.23}) +[2026-10-03 15:13:35,188][starVLA.training.train_starvla][INFO] - Step 3128, Loss: {'train/loss': 0.0001354702835669741, 'train/grad_norm_pre_clip': 0.0800391435623169, 'timing/data': 0.0002435100032016635, 'timing/model': 0.3656422899948666, 'learning_rate/qwen_vl_interface': 3.529759732170781e-06, 'learning_rate/action_model': 3.529759732170781e-05, 'epoch': 0.23}) +[2026-10-03 15:13:35,838][starVLA.training.train_starvla][INFO] - Step 3129, Loss: {'train/loss': 2.098732511512935e-05, 'train/grad_norm_pre_clip': 0.00646293256431818, 'timing/data': 0.00020493999181780964, 'timing/model': 0.3660607509955298, 'learning_rate/qwen_vl_interface': 3.5269213951209375e-06, 'learning_rate/action_model': 3.5269213951209374e-05, 'epoch': 0.23}) +[2026-10-03 15:13:36,488][starVLA.training.train_starvla][INFO] - Step 3130, Loss: {'train/loss': 0.0001619930990273133, 'train/grad_norm_pre_clip': 0.1560303270816803, 'timing/data': 0.0003798789985012263, 'timing/model': 0.36559562099864706, 'learning_rate/qwen_vl_interface': 3.5240837663636207e-06, 'learning_rate/action_model': 3.5240837663636204e-05, 'epoch': 0.23}) +[2026-10-03 15:13:37,139][starVLA.training.train_starvla][INFO] - Step 3131, Loss: {'train/loss': 0.3494287133216858, 'train/grad_norm_pre_clip': 13.287035942077637, 'timing/data': 0.00020698001026175916, 'timing/model': 0.3664022299926728, 'learning_rate/qwen_vl_interface': 3.5212468470652745e-06, 'learning_rate/action_model': 3.521246847065274e-05, 'epoch': 0.23}) +[2026-10-03 15:13:37,980][starVLA.training.train_starvla][INFO] - Step 3132, Loss: {'train/loss': 7.340620504692197e-05, 'train/grad_norm_pre_clip': 0.024674247950315475, 'timing/data': 0.00020207000488881022, 'timing/model': 0.36575280100805685, 'learning_rate/qwen_vl_interface': 3.5184106383920487e-06, 'learning_rate/action_model': 3.5184106383920486e-05, 'epoch': 0.23}) +[2026-10-03 15:13:38,630][starVLA.training.train_starvla][INFO] - Step 3133, Loss: {'train/loss': 0.002045438624918461, 'train/grad_norm_pre_clip': 3.5344762802124023, 'timing/data': 0.0003160399937769398, 'timing/model': 0.36518632000661455, 'learning_rate/qwen_vl_interface': 3.5155751415098022e-06, 'learning_rate/action_model': 3.515575141509802e-05, 'epoch': 0.23}) +[2026-10-03 15:13:39,278][starVLA.training.train_starvla][INFO] - Step 3134, Loss: {'train/loss': 0.0002862681867554784, 'train/grad_norm_pre_clip': 0.048254240304231644, 'timing/data': 0.00021771001047454774, 'timing/model': 0.3652896200073883, 'learning_rate/qwen_vl_interface': 3.5127403575841e-06, 'learning_rate/action_model': 3.5127403575841e-05, 'epoch': 0.23}) +[2026-10-03 15:13:39,908][starVLA.training.train_starvla][INFO] - Step 3135, Loss: {'train/loss': 2.0048415535711683e-05, 'train/grad_norm_pre_clip': 0.01295547466725111, 'timing/data': 0.000271140001132153, 'timing/model': 0.3662972099991748, 'learning_rate/qwen_vl_interface': 3.509906287780215e-06, 'learning_rate/action_model': 3.5099062877802144e-05, 'epoch': 0.23}) +[2026-10-03 15:13:40,557][starVLA.training.train_starvla][INFO] - Step 3136, Loss: {'train/loss': 0.00010058380576083437, 'train/grad_norm_pre_clip': 0.039214372634887695, 'timing/data': 0.0003523499908624217, 'timing/model': 0.3656066010007635, 'learning_rate/qwen_vl_interface': 3.5070729332631247e-06, 'learning_rate/action_model': 3.507072933263125e-05, 'epoch': 0.23}) +[2026-10-03 15:13:41,396][starVLA.training.train_starvla][INFO] - Step 3137, Loss: {'train/loss': 8.011144382180646e-05, 'train/grad_norm_pre_clip': 0.05571090430021286, 'timing/data': 0.00029464000544976443, 'timing/model': 0.5548427650064696, 'learning_rate/qwen_vl_interface': 3.5042402951975167e-06, 'learning_rate/action_model': 3.5042402951975167e-05, 'epoch': 0.23}) +[2026-10-03 15:13:42,046][starVLA.training.train_starvla][INFO] - Step 3138, Loss: {'train/loss': 0.00027989278896711767, 'train/grad_norm_pre_clip': 0.1255783587694168, 'timing/data': 0.00020046999270562083, 'timing/model': 0.3654139300051611, 'learning_rate/qwen_vl_interface': 3.50140837474778e-06, 'learning_rate/action_model': 3.50140837474778e-05, 'epoch': 0.23}) +[2026-10-03 15:13:42,697][starVLA.training.train_starvla][INFO] - Step 3139, Loss: {'train/loss': 7.197207742137834e-05, 'train/grad_norm_pre_clip': 0.02708725444972515, 'timing/data': 0.00021369000023696572, 'timing/model': 0.3663506799930474, 'learning_rate/qwen_vl_interface': 3.498577173078011e-06, 'learning_rate/action_model': 3.498577173078011e-05, 'epoch': 0.23}) +[2026-10-03 15:13:43,324][starVLA.training.train_starvla][INFO] - Step 3140, Loss: {'train/loss': 0.0003747068112716079, 'train/grad_norm_pre_clip': 0.0799214094877243, 'timing/data': 0.00019687999156303704, 'timing/model': 0.3657557600090513, 'learning_rate/qwen_vl_interface': 3.495746691352008e-06, 'learning_rate/action_model': 3.495746691352008e-05, 'epoch': 0.23}) +[2026-10-03 15:13:43,974][starVLA.training.train_starvla][INFO] - Step 3141, Loss: {'train/loss': 5.019712989451364e-05, 'train/grad_norm_pre_clip': 0.018346039578318596, 'timing/data': 0.00020206000772304833, 'timing/model': 0.3651172610116191, 'learning_rate/qwen_vl_interface': 3.4929169307332763e-06, 'learning_rate/action_model': 3.492916930733276e-05, 'epoch': 0.23}) +[2026-10-03 15:13:44,818][starVLA.training.train_starvla][INFO] - Step 3142, Loss: {'train/loss': 0.3199429214000702, 'train/grad_norm_pre_clip': 13.187758445739746, 'timing/data': 0.0002483400021446869, 'timing/model': 0.5608967739972286, 'learning_rate/qwen_vl_interface': 3.4900878923850234e-06, 'learning_rate/action_model': 3.490087892385023e-05, 'epoch': 0.23}) +[2026-10-03 15:13:45,469][starVLA.training.train_starvla][INFO] - Step 3143, Loss: {'train/loss': 2.4645401936140843e-05, 'train/grad_norm_pre_clip': 0.008055515587329865, 'timing/data': 0.0002999199932673946, 'timing/model': 0.36575493999407627, 'learning_rate/qwen_vl_interface': 3.4872595774701607e-06, 'learning_rate/action_model': 3.487259577470161e-05, 'epoch': 0.23}) +[2026-10-03 15:13:46,118][starVLA.training.train_starvla][INFO] - Step 3144, Loss: {'train/loss': 3.9355298213195056e-05, 'train/grad_norm_pre_clip': 0.012855235487222672, 'timing/data': 0.0003179100021952763, 'timing/model': 0.36515754999709316, 'learning_rate/qwen_vl_interface': 3.4844319871513014e-06, 'learning_rate/action_model': 3.484431987151301e-05, 'epoch': 0.23}) +[2026-10-03 15:13:46,767][starVLA.training.train_starvla][INFO] - Step 3145, Loss: {'train/loss': 0.5639809966087341, 'train/grad_norm_pre_clip': 61.866310119628906, 'timing/data': 0.00020196998957544565, 'timing/model': 0.36497569100174587, 'learning_rate/qwen_vl_interface': 3.4816051225907606e-06, 'learning_rate/action_model': 3.48160512259076e-05, 'epoch': 0.23}) +[2026-10-03 15:13:47,416][starVLA.training.train_starvla][INFO] - Step 3146, Loss: {'train/loss': 2.4033359295572154e-05, 'train/grad_norm_pre_clip': 0.019342172890901566, 'timing/data': 0.0003079699963564053, 'timing/model': 0.3654580800066469, 'learning_rate/qwen_vl_interface': 3.478778984950556e-06, 'learning_rate/action_model': 3.478778984950556e-05, 'epoch': 0.23}) +[2026-10-03 15:13:48,067][starVLA.training.train_starvla][INFO] - Step 3147, Loss: {'train/loss': 8.352004442713223e-06, 'train/grad_norm_pre_clip': 0.003097902052104473, 'timing/data': 0.00031311000930145383, 'timing/model': 0.36572321999119595, 'learning_rate/qwen_vl_interface': 3.475953575392405e-06, 'learning_rate/action_model': 3.475953575392405e-05, 'epoch': 0.23}) +[2026-10-03 15:13:48,888][starVLA.training.train_starvla][INFO] - Step 3148, Loss: {'train/loss': 3.198971535312012e-05, 'train/grad_norm_pre_clip': 0.012354640290141106, 'timing/data': 0.0003375999949639663, 'timing/model': 0.34299389300576877, 'learning_rate/qwen_vl_interface': 3.473128895077728e-06, 'learning_rate/action_model': 3.473128895077728e-05, 'epoch': 0.23}) +[2026-10-03 15:13:49,537][starVLA.training.train_starvla][INFO] - Step 3149, Loss: {'train/loss': 1.2345251889200881e-05, 'train/grad_norm_pre_clip': 0.004217891488224268, 'timing/data': 0.00019873000564984977, 'timing/model': 0.36517453999840654, 'learning_rate/qwen_vl_interface': 3.4703049451676453e-06, 'learning_rate/action_model': 3.4703049451676454e-05, 'epoch': 0.23}) +[2026-10-03 15:13:50,185][starVLA.training.train_starvla][INFO] - Step 3150, Loss: {'train/loss': 1.697205334494356e-05, 'train/grad_norm_pre_clip': 0.004531737416982651, 'timing/data': 0.0001957299973582849, 'timing/model': 0.36518078099470586, 'learning_rate/qwen_vl_interface': 3.4674817268229732e-06, 'learning_rate/action_model': 3.467481726822973e-05, 'epoch': 0.23}) +[2026-10-03 15:13:50,838][starVLA.training.train_starvla][INFO] - Step 3151, Loss: {'train/loss': 0.1693892478942871, 'train/grad_norm_pre_clip': 166.83847045898438, 'timing/data': 0.00031256998772732913, 'timing/model': 0.36693733000720385, 'learning_rate/qwen_vl_interface': 3.464659241204232e-06, 'learning_rate/action_model': 3.4646592412042314e-05, 'epoch': 0.23}) +[2026-10-03 15:13:51,488][starVLA.training.train_starvla][INFO] - Step 3152, Loss: {'train/loss': 0.00018972877296619117, 'train/grad_norm_pre_clip': 0.078940249979496, 'timing/data': 0.0002428800071356818, 'timing/model': 0.36562441001296975, 'learning_rate/qwen_vl_interface': 3.4618374894716377e-06, 'learning_rate/action_model': 3.461837489471638e-05, 'epoch': 0.23}) +[2026-10-03 15:13:52,339][starVLA.training.train_starvla][INFO] - Step 3153, Loss: {'train/loss': 0.00022119440836831927, 'train/grad_norm_pre_clip': 0.08617186546325684, 'timing/data': 0.0002580700092948973, 'timing/model': 0.36574571099481545, 'learning_rate/qwen_vl_interface': 3.459016472785107e-06, 'learning_rate/action_model': 3.459016472785107e-05, 'epoch': 0.23}) +[2026-10-03 15:13:52,988][starVLA.training.train_starvla][INFO] - Step 3154, Loss: {'train/loss': 1.4237739378586411e-05, 'train/grad_norm_pre_clip': 0.00408405065536499, 'timing/data': 0.0002441600081510842, 'timing/model': 0.3648682209895924, 'learning_rate/qwen_vl_interface': 3.456196192304254e-06, 'learning_rate/action_model': 3.456196192304254e-05, 'epoch': 0.23}) +[2026-10-03 15:13:53,639][starVLA.training.train_starvla][INFO] - Step 3155, Loss: {'train/loss': 0.06687978655099869, 'train/grad_norm_pre_clip': 170.1306610107422, 'timing/data': 0.0003022799937753007, 'timing/model': 0.3659523709939094, 'learning_rate/qwen_vl_interface': 3.4533766491883887e-06, 'learning_rate/action_model': 3.453376649188388e-05, 'epoch': 0.23}) +[2026-10-03 15:13:54,288][starVLA.training.train_starvla][INFO] - Step 3156, Loss: {'train/loss': 0.7572153210639954, 'train/grad_norm_pre_clip': 93.7071762084961, 'timing/data': 0.00019870900723617524, 'timing/model': 0.36614254099549726, 'learning_rate/qwen_vl_interface': 3.450557844596517e-06, 'learning_rate/action_model': 3.450557844596517e-05, 'epoch': 0.23}) +[2026-10-03 15:13:54,937][starVLA.training.train_starvla][INFO] - Step 3157, Loss: {'train/loss': 6.154128641355783e-06, 'train/grad_norm_pre_clip': 0.0019161964301019907, 'timing/data': 0.00032161000126507133, 'timing/model': 0.36506641999585554, 'learning_rate/qwen_vl_interface': 3.447739779687345e-06, 'learning_rate/action_model': 3.4477397796873446e-05, 'epoch': 0.23}) +[2026-10-03 15:13:55,784][starVLA.training.train_starvla][INFO] - Step 3158, Loss: {'train/loss': 0.4754948914051056, 'train/grad_norm_pre_clip': 79.385498046875, 'timing/data': 0.00030171999242156744, 'timing/model': 0.5631493030086858, 'learning_rate/qwen_vl_interface': 3.4449224556192722e-06, 'learning_rate/action_model': 3.444922455619272e-05, 'epoch': 0.23}) +[2026-10-03 15:13:56,434][starVLA.training.train_starvla][INFO] - Step 3159, Loss: {'train/loss': 0.34601229429244995, 'train/grad_norm_pre_clip': 41.54509735107422, 'timing/data': 0.0003021700104000047, 'timing/model': 0.3657428910082672, 'learning_rate/qwen_vl_interface': 3.442105873550395e-06, 'learning_rate/action_model': 3.442105873550395e-05, 'epoch': 0.23}) +[2026-10-03 15:13:57,084][starVLA.training.train_starvla][INFO] - Step 3160, Loss: {'train/loss': 0.00021212510182522237, 'train/grad_norm_pre_clip': 0.07773223519325256, 'timing/data': 0.00033884000731632113, 'timing/model': 0.3653554799966514, 'learning_rate/qwen_vl_interface': 3.4392900346385022e-06, 'learning_rate/action_model': 3.4392900346385024e-05, 'epoch': 0.23}) +[2026-10-03 15:13:57,733][starVLA.training.train_starvla][INFO] - Step 3161, Loss: {'train/loss': 8.384964894503355e-05, 'train/grad_norm_pre_clip': 0.04064035415649414, 'timing/data': 0.0003300399985164404, 'timing/model': 0.3655601799982833, 'learning_rate/qwen_vl_interface': 3.4364749400410796e-06, 'learning_rate/action_model': 3.43647494004108e-05, 'epoch': 0.23}) +[2026-10-03 15:13:58,382][starVLA.training.train_starvla][INFO] - Step 3162, Loss: {'train/loss': 1.9444907593424432e-05, 'train/grad_norm_pre_clip': 0.0059076882898807526, 'timing/data': 0.0002565100003266707, 'timing/model': 0.3651802809908986, 'learning_rate/qwen_vl_interface': 3.4336605909153066e-06, 'learning_rate/action_model': 3.433660590915306e-05, 'epoch': 0.23}) +[2026-10-03 15:13:59,203][starVLA.training.train_starvla][INFO] - Step 3163, Loss: {'train/loss': 2.8064281650586054e-05, 'train/grad_norm_pre_clip': 0.009923634119331837, 'timing/data': 0.00020200999279040843, 'timing/model': 0.5572703040088527, 'learning_rate/qwen_vl_interface': 3.430846988418057e-06, 'learning_rate/action_model': 3.430846988418057e-05, 'epoch': 0.23}) +[2026-10-03 15:13:59,854][starVLA.training.train_starvla][INFO] - Step 3164, Loss: {'train/loss': 0.25416287779808044, 'train/grad_norm_pre_clip': 10.76140022277832, 'timing/data': 0.00021371999173425138, 'timing/model': 0.3658242609963054, 'learning_rate/qwen_vl_interface': 3.428034133705893e-06, 'learning_rate/action_model': 3.428034133705893e-05, 'epoch': 0.23}) +[2026-10-03 15:14:00,480][starVLA.training.train_starvla][INFO] - Step 3165, Loss: {'train/loss': 0.0003956010041292757, 'train/grad_norm_pre_clip': 0.1989796757698059, 'timing/data': 0.00031913000566419214, 'timing/model': 0.3651103000011062, 'learning_rate/qwen_vl_interface': 3.4252220279350763e-06, 'learning_rate/action_model': 3.425222027935076e-05, 'epoch': 0.23}) +[2026-10-03 15:14:01,130][starVLA.training.train_starvla][INFO] - Step 3166, Loss: {'train/loss': 0.0037616179324686527, 'train/grad_norm_pre_clip': 6.166271686553955, 'timing/data': 0.00031909999961499125, 'timing/model': 0.36524567000742536, 'learning_rate/qwen_vl_interface': 3.4224106722615547e-06, 'learning_rate/action_model': 3.422410672261554e-05, 'epoch': 0.23}) +[2026-10-03 15:14:01,781][starVLA.training.train_starvla][INFO] - Step 3167, Loss: {'train/loss': 0.12135445326566696, 'train/grad_norm_pre_clip': 50.007537841796875, 'timing/data': 0.000297929989756085, 'timing/model': 0.36566570099967066, 'learning_rate/qwen_vl_interface': 3.419600067840973e-06, 'learning_rate/action_model': 3.419600067840973e-05, 'epoch': 0.23}) +[2026-10-03 15:14:02,430][starVLA.training.train_starvla][INFO] - Step 3168, Loss: {'train/loss': 0.00020314492576289922, 'train/grad_norm_pre_clip': 0.10380315780639648, 'timing/data': 0.00019950000569224358, 'timing/model': 0.3657477109954925, 'learning_rate/qwen_vl_interface': 3.4167902158286627e-06, 'learning_rate/action_model': 3.4167902158286626e-05, 'epoch': 0.23}) +[2026-10-03 15:14:03,278][starVLA.training.train_starvla][INFO] - Step 3169, Loss: {'train/loss': 0.009132951498031616, 'train/grad_norm_pre_clip': 15.321181297302246, 'timing/data': 0.00034545999369584024, 'timing/model': 0.3651307100080885, 'learning_rate/qwen_vl_interface': 3.4139811173796468e-06, 'learning_rate/action_model': 3.413981117379647e-05, 'epoch': 0.23}) +[2026-10-03 15:14:03,927][starVLA.training.train_starvla][INFO] - Step 3170, Loss: {'train/loss': 0.7350030541419983, 'train/grad_norm_pre_clip': 165.87533569335938, 'timing/data': 0.00031442999897990376, 'timing/model': 0.365186619994347, 'learning_rate/qwen_vl_interface': 3.411172773648642e-06, 'learning_rate/action_model': 3.411172773648642e-05, 'epoch': 0.23}) +[2026-10-03 15:14:04,578][starVLA.training.train_starvla][INFO] - Step 3171, Loss: {'train/loss': 0.0001389526150887832, 'train/grad_norm_pre_clip': 0.05263437330722809, 'timing/data': 0.0003142000059597194, 'timing/model': 0.36518537100346293, 'learning_rate/qwen_vl_interface': 3.408365185790052e-06, 'learning_rate/action_model': 3.4083651857900515e-05, 'epoch': 0.23}) +[2026-10-03 15:14:05,204][starVLA.training.train_starvla][INFO] - Step 3172, Loss: {'train/loss': 0.11832422763109207, 'train/grad_norm_pre_clip': 79.39055633544922, 'timing/data': 0.00033830999745987356, 'timing/model': 0.34277653400204144, 'learning_rate/qwen_vl_interface': 3.40555835495797e-06, 'learning_rate/action_model': 3.40555835495797e-05, 'epoch': 0.23}) +[2026-10-03 15:14:05,854][starVLA.training.train_starvla][INFO] - Step 3173, Loss: {'train/loss': 0.0022436892613768578, 'train/grad_norm_pre_clip': 3.4484758377075195, 'timing/data': 0.0002556100080255419, 'timing/model': 0.36600715000531636, 'learning_rate/qwen_vl_interface': 3.4027522823061763e-06, 'learning_rate/action_model': 3.4027522823061765e-05, 'epoch': 0.23}) +[2026-10-03 15:14:06,676][starVLA.training.train_starvla][INFO] - Step 3174, Loss: {'train/loss': 0.27494779229164124, 'train/grad_norm_pre_clip': 75.52207946777344, 'timing/data': 0.00022495999292004853, 'timing/model': 0.34288742399076, 'learning_rate/qwen_vl_interface': 3.399946968988143e-06, 'learning_rate/action_model': 3.399946968988143e-05, 'epoch': 0.23}) +[2026-10-03 15:14:07,327][starVLA.training.train_starvla][INFO] - Step 3175, Loss: {'train/loss': 0.41235828399658203, 'train/grad_norm_pre_clip': 169.76773071289062, 'timing/data': 0.0002998099953401834, 'timing/model': 0.36591660999692976, 'learning_rate/qwen_vl_interface': 3.397142416157028e-06, 'learning_rate/action_model': 3.397142416157028e-05, 'epoch': 0.23}) +[2026-10-03 15:14:07,976][starVLA.training.train_starvla][INFO] - Step 3176, Loss: {'train/loss': 0.00011121938587166369, 'train/grad_norm_pre_clip': 0.05732043460011482, 'timing/data': 0.00034716998925432563, 'timing/model': 0.3652836900000693, 'learning_rate/qwen_vl_interface': 3.3943386249656766e-06, 'learning_rate/action_model': 3.394338624965677e-05, 'epoch': 0.23}) +[2026-10-03 15:14:08,624][starVLA.training.train_starvla][INFO] - Step 3177, Loss: {'train/loss': 0.0006600723136216402, 'train/grad_norm_pre_clip': 0.7002883553504944, 'timing/data': 0.0003098199958913028, 'timing/model': 0.36405981999996584, 'learning_rate/qwen_vl_interface': 3.391535596566624e-06, 'learning_rate/action_model': 3.391535596566624e-05, 'epoch': 0.23}) +[2026-10-03 15:14:09,270][starVLA.training.train_starvla][INFO] - Step 3178, Loss: {'train/loss': 6.809003389207646e-05, 'train/grad_norm_pre_clip': 0.035331666469573975, 'timing/data': 0.00019936000171583146, 'timing/model': 0.36368255999695975, 'learning_rate/qwen_vl_interface': 3.388733332112086e-06, 'learning_rate/action_model': 3.3887333321120856e-05, 'epoch': 0.23}) +[2026-10-03 15:14:10,114][starVLA.training.train_starvla][INFO] - Step 3179, Loss: {'train/loss': 0.2933633029460907, 'train/grad_norm_pre_clip': 82.16759490966797, 'timing/data': 0.00024148999364115298, 'timing/model': 0.5605147239984944, 'learning_rate/qwen_vl_interface': 3.3859318327539693e-06, 'learning_rate/action_model': 3.385931832753969e-05, 'epoch': 0.23}) +[2026-10-03 15:14:10,760][starVLA.training.train_starvla][INFO] - Step 3180, Loss: {'train/loss': 0.3282058835029602, 'train/grad_norm_pre_clip': 89.89322662353516, 'timing/data': 0.00019546000112313777, 'timing/model': 0.3639790309971431, 'learning_rate/qwen_vl_interface': 3.3831310996438633e-06, 'learning_rate/action_model': 3.383131099643863e-05, 'epoch': 0.23}) +[2026-10-03 15:14:11,406][starVLA.training.train_starvla][INFO] - Step 3181, Loss: {'train/loss': 0.0016314941458404064, 'train/grad_norm_pre_clip': 0.9918763637542725, 'timing/data': 0.00019570000586099923, 'timing/model': 0.3642697909963317, 'learning_rate/qwen_vl_interface': 3.3803311339330447e-06, 'learning_rate/action_model': 3.3803311339330444e-05, 'epoch': 0.23}) +[2026-10-03 15:14:12,053][starVLA.training.train_starvla][INFO] - Step 3182, Loss: {'train/loss': 0.007533465512096882, 'train/grad_norm_pre_clip': 15.806300163269043, 'timing/data': 0.00019966000400017947, 'timing/model': 0.3645673400023952, 'learning_rate/qwen_vl_interface': 3.3775319367724746e-06, 'learning_rate/action_model': 3.3775319367724745e-05, 'epoch': 0.23}) +[2026-10-03 15:14:12,702][starVLA.training.train_starvla][INFO] - Step 3183, Loss: {'train/loss': 5.565088940784335e-05, 'train/grad_norm_pre_clip': 0.01604381762444973, 'timing/data': 0.00039134999678935856, 'timing/model': 0.36440650999429636, 'learning_rate/qwen_vl_interface': 3.374733509312795e-06, 'learning_rate/action_model': 3.374733509312795e-05, 'epoch': 0.23}) +[2026-10-03 15:14:13,519][starVLA.training.train_starvla][INFO] - Step 3184, Loss: {'train/loss': 0.000554748170543462, 'train/grad_norm_pre_clip': 0.20835311710834503, 'timing/data': 0.00019895000150427222, 'timing/model': 0.5339237080042949, 'learning_rate/qwen_vl_interface': 3.3719358527043345e-06, 'learning_rate/action_model': 3.371935852704334e-05, 'epoch': 0.23}) +[2026-10-03 15:14:14,165][starVLA.training.train_starvla][INFO] - Step 3185, Loss: {'train/loss': 0.0002166575868614018, 'train/grad_norm_pre_clip': 0.12226715683937073, 'timing/data': 0.00024031000793911517, 'timing/model': 0.36422204101108946, 'learning_rate/qwen_vl_interface': 3.3691389680971047e-06, 'learning_rate/action_model': 3.3691389680971044e-05, 'epoch': 0.23}) +[2026-10-03 15:14:14,812][starVLA.training.train_starvla][INFO] - Step 3186, Loss: {'train/loss': 0.0028305326122790575, 'train/grad_norm_pre_clip': 0.7238633036613464, 'timing/data': 0.00020213999960105866, 'timing/model': 0.3644703910103999, 'learning_rate/qwen_vl_interface': 3.3663428566407985e-06, 'learning_rate/action_model': 3.366342856640798e-05, 'epoch': 0.23}) +[2026-10-03 15:14:15,465][starVLA.training.train_starvla][INFO] - Step 3187, Loss: {'train/loss': 6.903427129145712e-05, 'train/grad_norm_pre_clip': 0.026565715670585632, 'timing/data': 0.0002991900109918788, 'timing/model': 0.3683997899934184, 'learning_rate/qwen_vl_interface': 3.3635475194847927e-06, 'learning_rate/action_model': 3.363547519484793e-05, 'epoch': 0.23}) +[2026-10-03 15:14:16,112][starVLA.training.train_starvla][INFO] - Step 3188, Loss: {'train/loss': 0.0001802504702936858, 'train/grad_norm_pre_clip': 0.09205718338489532, 'timing/data': 0.00020518999372143298, 'timing/model': 0.36459589999867603, 'learning_rate/qwen_vl_interface': 3.3607529577781433e-06, 'learning_rate/action_model': 3.360752957778143e-05, 'epoch': 0.23}) +[2026-10-03 15:14:16,737][starVLA.training.train_starvla][INFO] - Step 3189, Loss: {'train/loss': 2.4012251742533408e-05, 'train/grad_norm_pre_clip': 0.007739711087197065, 'timing/data': 0.00030899999546818435, 'timing/model': 0.34184309300326277, 'learning_rate/qwen_vl_interface': 3.3579591726695906e-06, 'learning_rate/action_model': 3.35795917266959e-05, 'epoch': 0.23}) +[2026-10-03 15:14:17,587][starVLA.training.train_starvla][INFO] - Step 3190, Loss: {'train/loss': 0.0014124701265245676, 'train/grad_norm_pre_clip': 1.2185763120651245, 'timing/data': 0.00020803000370506197, 'timing/model': 0.3639647509990027, 'learning_rate/qwen_vl_interface': 3.3551661653075522e-06, 'learning_rate/action_model': 3.355166165307552e-05, 'epoch': 0.23}) +[2026-10-03 15:14:18,236][starVLA.training.train_starvla][INFO] - Step 3191, Loss: {'train/loss': 0.002998298266902566, 'train/grad_norm_pre_clip': 2.673018455505371, 'timing/data': 0.0002583490131655708, 'timing/model': 0.3644315209967317, 'learning_rate/qwen_vl_interface': 3.35237393684013e-06, 'learning_rate/action_model': 3.35237393684013e-05, 'epoch': 0.23}) +[2026-10-03 15:14:18,883][starVLA.training.train_starvla][INFO] - Step 3192, Loss: {'train/loss': 5.402208262239583e-05, 'train/grad_norm_pre_clip': 0.031565066426992416, 'timing/data': 0.0003092400002060458, 'timing/model': 0.36459091999859083, 'learning_rate/qwen_vl_interface': 3.349582488415103e-06, 'learning_rate/action_model': 3.3495824884151026e-05, 'epoch': 0.23}) +[2026-10-03 15:14:19,530][starVLA.training.train_starvla][INFO] - Step 3193, Loss: {'train/loss': 0.0002580603177193552, 'train/grad_norm_pre_clip': 0.0952957421541214, 'timing/data': 0.00020603000302799046, 'timing/model': 0.36418596099247225, 'learning_rate/qwen_vl_interface': 3.3467918211799294e-06, 'learning_rate/action_model': 3.3467918211799295e-05, 'epoch': 0.23}) +[2026-10-03 15:14:20,176][starVLA.training.train_starvla][INFO] - Step 3194, Loss: {'train/loss': 1.9773407984757796e-05, 'train/grad_norm_pre_clip': 0.005560629535466433, 'timing/data': 0.00031099999614525586, 'timing/model': 0.3638033110037213, 'learning_rate/qwen_vl_interface': 3.3440019362817476e-06, 'learning_rate/action_model': 3.344001936281747e-05, 'epoch': 0.23}) +[2026-10-03 15:14:21,015][starVLA.training.train_starvla][INFO] - Step 3195, Loss: {'train/loss': 1.4371883480635006e-05, 'train/grad_norm_pre_clip': 0.0034155806060880423, 'timing/data': 0.00020798998593818396, 'timing/model': 0.5552765949978493, 'learning_rate/qwen_vl_interface': 3.3412128348673745e-06, 'learning_rate/action_model': 3.341212834867374e-05, 'epoch': 0.23}) +[2026-10-03 15:14:21,663][starVLA.training.train_starvla][INFO] - Step 3196, Loss: {'train/loss': 5.153568781679496e-05, 'train/grad_norm_pre_clip': 0.03509129211306572, 'timing/data': 0.0001964999974006787, 'timing/model': 0.3648810400045477, 'learning_rate/qwen_vl_interface': 3.338424518083303e-06, 'learning_rate/action_model': 3.338424518083303e-05, 'epoch': 0.23}) +[2026-10-03 15:14:22,311][starVLA.training.train_starvla][INFO] - Step 3197, Loss: {'train/loss': 0.6120588183403015, 'train/grad_norm_pre_clip': 50.13736343383789, 'timing/data': 0.00024059000134002417, 'timing/model': 0.3660352699953364, 'learning_rate/qwen_vl_interface': 3.3356369870757056e-06, 'learning_rate/action_model': 3.335636987075705e-05, 'epoch': 0.23}) +[2026-10-03 15:14:22,960][starVLA.training.train_starvla][INFO] - Step 3198, Loss: {'train/loss': 8.292430720757693e-06, 'train/grad_norm_pre_clip': 0.0021403890568763018, 'timing/data': 0.00020245999621693045, 'timing/model': 0.3657547299953876, 'learning_rate/qwen_vl_interface': 3.3328502429904327e-06, 'learning_rate/action_model': 3.3328502429904325e-05, 'epoch': 0.23}) +[2026-10-03 15:14:23,610][starVLA.training.train_starvla][INFO] - Step 3199, Loss: {'train/loss': 0.0005479754181578755, 'train/grad_norm_pre_clip': 0.23623180389404297, 'timing/data': 0.00020828000560868531, 'timing/model': 0.36458356100774836, 'learning_rate/qwen_vl_interface': 3.3300642869730066e-06, 'learning_rate/action_model': 3.330064286973006e-05, 'epoch': 0.23}) +[2026-10-03 15:14:24,455][starVLA.training.train_starvla][INFO] - Step 3200, Loss: {'train/loss': 0.4349590837955475, 'train/grad_norm_pre_clip': 46.57085418701172, 'timing/data': 0.00023645999317523092, 'timing/model': 0.5618018430104712, 'learning_rate/qwen_vl_interface': 3.3272791201686294e-06, 'learning_rate/action_model': 3.327279120168629e-05, 'epoch': 0.23}) +[2026-10-03 15:14:25,100][starVLA.training.train_starvla][INFO] - Step 3201, Loss: {'train/loss': 0.00014491361798718572, 'train/grad_norm_pre_clip': 0.06551706790924072, 'timing/data': 0.00024449000193271786, 'timing/model': 0.36369752101018094, 'learning_rate/qwen_vl_interface': 3.324494743722179e-06, 'learning_rate/action_model': 3.3244947437221785e-05, 'epoch': 0.23}) +[2026-10-03 15:14:25,745][starVLA.training.train_starvla][INFO] - Step 3202, Loss: {'train/loss': 0.52894127368927, 'train/grad_norm_pre_clip': 28.720596313476562, 'timing/data': 0.00024708999262657017, 'timing/model': 0.36332515999674797, 'learning_rate/qwen_vl_interface': 3.321711158778205e-06, 'learning_rate/action_model': 3.321711158778205e-05, 'epoch': 0.23}) +[2026-10-03 15:14:26,393][starVLA.training.train_starvla][INFO] - Step 3203, Loss: {'train/loss': 0.19365708529949188, 'train/grad_norm_pre_clip': 154.6500701904297, 'timing/data': 0.0003139199980068952, 'timing/model': 0.3640940710029099, 'learning_rate/qwen_vl_interface': 3.3189283664809393e-06, 'learning_rate/action_model': 3.318928366480939e-05, 'epoch': 0.23}) +[2026-10-03 15:14:27,038][starVLA.training.train_starvla][INFO] - Step 3204, Loss: {'train/loss': 5.039244570070878e-05, 'train/grad_norm_pre_clip': 0.027737166732549667, 'timing/data': 0.00023974000941962004, 'timing/model': 0.36344201100291684, 'learning_rate/qwen_vl_interface': 3.316146367974276e-06, 'learning_rate/action_model': 3.316146367974276e-05, 'epoch': 0.23}) +[2026-10-03 15:14:27,685][starVLA.training.train_starvla][INFO] - Step 3205, Loss: {'train/loss': 8.48750933073461e-05, 'train/grad_norm_pre_clip': 0.04014861583709717, 'timing/data': 0.00031531900458503515, 'timing/model': 0.3639153609983623, 'learning_rate/qwen_vl_interface': 3.3133651644017945e-06, 'learning_rate/action_model': 3.3133651644017944e-05, 'epoch': 0.23}) +[2026-10-03 15:14:28,524][starVLA.training.train_starvla][INFO] - Step 3206, Loss: {'train/loss': 0.3998328745365143, 'train/grad_norm_pre_clip': 52.78459930419922, 'timing/data': 0.000322089996188879, 'timing/model': 0.3638755309948465, 'learning_rate/qwen_vl_interface': 3.31058475690674e-06, 'learning_rate/action_model': 3.31058475690674e-05, 'epoch': 0.23}) +[2026-10-03 15:14:29,172][starVLA.training.train_starvla][INFO] - Step 3207, Loss: {'train/loss': 0.002208961406722665, 'train/grad_norm_pre_clip': 0.5712770819664001, 'timing/data': 0.00020938999659847468, 'timing/model': 0.36417342998902313, 'learning_rate/qwen_vl_interface': 3.3078051466320337e-06, 'learning_rate/action_model': 3.307805146632034e-05, 'epoch': 0.23}) +[2026-10-03 15:14:29,818][starVLA.training.train_starvla][INFO] - Step 3208, Loss: {'train/loss': 0.00035816297167912126, 'train/grad_norm_pre_clip': 0.29925602674484253, 'timing/data': 0.00019735000387299806, 'timing/model': 0.36386243099696003, 'learning_rate/qwen_vl_interface': 3.3050263347202712e-06, 'learning_rate/action_model': 3.305026334720271e-05, 'epoch': 0.23}) +[2026-10-03 15:14:30,463][starVLA.training.train_starvla][INFO] - Step 3209, Loss: {'train/loss': 6.82784739183262e-05, 'train/grad_norm_pre_clip': 0.016226040199398994, 'timing/data': 0.00019782000163104385, 'timing/model': 0.36429563000274356, 'learning_rate/qwen_vl_interface': 3.302248322313712e-06, 'learning_rate/action_model': 3.302248322313712e-05, 'epoch': 0.23}) +[2026-10-03 15:14:31,109][starVLA.training.train_starvla][INFO] - Step 3210, Loss: {'train/loss': 6.123706407379359e-05, 'train/grad_norm_pre_clip': 0.022354573011398315, 'timing/data': 0.0002578690036898479, 'timing/model': 0.3637524409859907, 'learning_rate/qwen_vl_interface': 3.299471110554295e-06, 'learning_rate/action_model': 3.299471110554295e-05, 'epoch': 0.23}) +[2026-10-03 15:14:31,962][starVLA.training.train_starvla][INFO] - Step 3211, Loss: {'train/loss': 6.945760833332315e-05, 'train/grad_norm_pre_clip': 0.02172289416193962, 'timing/data': 0.0003128099924651906, 'timing/model': 0.3639117599959718, 'learning_rate/qwen_vl_interface': 3.296694700583626e-06, 'learning_rate/action_model': 3.296694700583626e-05, 'epoch': 0.24}) +[2026-10-03 15:14:32,586][starVLA.training.train_starvla][INFO] - Step 3212, Loss: {'train/loss': 0.2011612355709076, 'train/grad_norm_pre_clip': 94.27986907958984, 'timing/data': 0.00024417000531684607, 'timing/model': 0.34152930398704484, 'learning_rate/qwen_vl_interface': 3.2939190935429833e-06, 'learning_rate/action_model': 3.293919093542983e-05, 'epoch': 0.24}) +[2026-10-03 15:14:33,232][starVLA.training.train_starvla][INFO] - Step 3213, Loss: {'train/loss': 3.5320554161444306e-05, 'train/grad_norm_pre_clip': 0.008910534903407097, 'timing/data': 0.0003385800082469359, 'timing/model': 0.3640896610013442, 'learning_rate/qwen_vl_interface': 3.2911442905733126e-06, 'learning_rate/action_model': 3.291144290573312e-05, 'epoch': 0.24}) +[2026-10-03 15:14:33,879][starVLA.training.train_starvla][INFO] - Step 3214, Loss: {'train/loss': 2.504801523173228e-05, 'train/grad_norm_pre_clip': 0.006045666057616472, 'timing/data': 0.00020117999520152807, 'timing/model': 0.36422782999579795, 'learning_rate/qwen_vl_interface': 3.2883702928152323e-06, 'learning_rate/action_model': 3.288370292815232e-05, 'epoch': 0.24}) +[2026-10-03 15:14:34,506][starVLA.training.train_starvla][INFO] - Step 3215, Loss: {'train/loss': 0.00019023935601580888, 'train/grad_norm_pre_clip': 0.15047113597393036, 'timing/data': 0.00032852998992893845, 'timing/model': 0.3421373130113352, 'learning_rate/qwen_vl_interface': 3.2855971014090275e-06, 'learning_rate/action_model': 3.285597101409028e-05, 'epoch': 0.24}) +[2026-10-03 15:14:35,352][starVLA.training.train_starvla][INFO] - Step 3216, Loss: {'train/loss': 0.0003968187957070768, 'train/grad_norm_pre_clip': 0.1398102343082428, 'timing/data': 0.00020665999909397215, 'timing/model': 0.3646833509992575, 'learning_rate/qwen_vl_interface': 3.2828247174946528e-06, 'learning_rate/action_model': 3.2828247174946524e-05, 'epoch': 0.24}) +[2026-10-03 15:14:36,001][starVLA.training.train_starvla][INFO] - Step 3217, Loss: {'train/loss': 5.177962520974688e-05, 'train/grad_norm_pre_clip': 0.029155148193240166, 'timing/data': 0.0002511400089133531, 'timing/model': 0.36493407099624164, 'learning_rate/qwen_vl_interface': 3.280053142211729e-06, 'learning_rate/action_model': 3.280053142211729e-05, 'epoch': 0.24}) +[2026-10-03 15:14:36,649][starVLA.training.train_starvla][INFO] - Step 3218, Loss: {'train/loss': 7.611104956595227e-05, 'train/grad_norm_pre_clip': 0.047262243926525116, 'timing/data': 0.00022344000171869993, 'timing/model': 0.3651577400014503, 'learning_rate/qwen_vl_interface': 3.2772823766995475e-06, 'learning_rate/action_model': 3.277282376699548e-05, 'epoch': 0.24}) +[2026-10-03 15:14:37,298][starVLA.training.train_starvla][INFO] - Step 3219, Loss: {'train/loss': 2.8190886951051652e-05, 'train/grad_norm_pre_clip': 0.00894963089376688, 'timing/data': 0.00020337999740149826, 'timing/model': 0.36513790099706966, 'learning_rate/qwen_vl_interface': 3.2745124220970646e-06, 'learning_rate/action_model': 3.274512422097065e-05, 'epoch': 0.24}) +[2026-10-03 15:14:37,946][starVLA.training.train_starvla][INFO] - Step 3220, Loss: {'train/loss': 4.736983100883663e-05, 'train/grad_norm_pre_clip': 0.01496950164437294, 'timing/data': 0.0002025700086960569, 'timing/model': 0.36501670999859925, 'learning_rate/qwen_vl_interface': 3.2717432795429057e-06, 'learning_rate/action_model': 3.271743279542905e-05, 'epoch': 0.24}) +[2026-10-03 15:14:38,790][starVLA.training.train_starvla][INFO] - Step 3221, Loss: {'train/loss': 0.0006890830700285733, 'train/grad_norm_pre_clip': 0.5554954409599304, 'timing/data': 0.00019597000209614635, 'timing/model': 0.561142573991674, 'learning_rate/qwen_vl_interface': 3.2689749501753578e-06, 'learning_rate/action_model': 3.2689749501753574e-05, 'epoch': 0.24}) +[2026-10-03 15:14:39,437][starVLA.training.train_starvla][INFO] - Step 3222, Loss: {'train/loss': 0.00010019527690019459, 'train/grad_norm_pre_clip': 0.050498589873313904, 'timing/data': 0.00020257999130990356, 'timing/model': 0.3645013209898025, 'learning_rate/qwen_vl_interface': 3.2662074351323774e-06, 'learning_rate/action_model': 3.266207435132378e-05, 'epoch': 0.24}) +[2026-10-03 15:14:40,063][starVLA.training.train_starvla][INFO] - Step 3223, Loss: {'train/loss': 0.00010791170643642545, 'train/grad_norm_pre_clip': 0.13393746316432953, 'timing/data': 0.0003157600003760308, 'timing/model': 0.36437912000110373, 'learning_rate/qwen_vl_interface': 3.2634407355515875e-06, 'learning_rate/action_model': 3.2634407355515875e-05, 'epoch': 0.24}) +[2026-10-03 15:14:40,667][starVLA.training.train_starvla][INFO] - Step 3224, Loss: {'train/loss': 7.089397695381194e-05, 'train/grad_norm_pre_clip': 0.0235378909856081, 'timing/data': 0.00024919000861700624, 'timing/model': 0.3423424630018417, 'learning_rate/qwen_vl_interface': 3.2606748525702715e-06, 'learning_rate/action_model': 3.260674852570271e-05, 'epoch': 0.24}) +[2026-10-03 15:14:41,314][starVLA.training.train_starvla][INFO] - Step 3225, Loss: {'train/loss': 0.00013597714132629335, 'train/grad_norm_pre_clip': 0.08092253655195236, 'timing/data': 0.00022075000742916018, 'timing/model': 0.36478902099770494, 'learning_rate/qwen_vl_interface': 3.2579097873253803e-06, 'learning_rate/action_model': 3.25790978732538e-05, 'epoch': 0.24}) +[2026-10-03 15:14:41,962][starVLA.training.train_starvla][INFO] - Step 3226, Loss: {'train/loss': 0.0013885953230783343, 'train/grad_norm_pre_clip': 3.2645442485809326, 'timing/data': 0.0003069100057473406, 'timing/model': 0.36477959000330884, 'learning_rate/qwen_vl_interface': 3.2551455409535274e-06, 'learning_rate/action_model': 3.255145540953528e-05, 'epoch': 0.24}) +[2026-10-03 15:14:42,809][starVLA.training.train_starvla][INFO] - Step 3227, Loss: {'train/loss': 0.3922868072986603, 'train/grad_norm_pre_clip': 23.55525779724121, 'timing/data': 0.00019986000552307814, 'timing/model': 0.36507023000740446, 'learning_rate/qwen_vl_interface': 3.2523821145909907e-06, 'learning_rate/action_model': 3.252382114590991e-05, 'epoch': 0.24}) +[2026-10-03 15:14:43,457][starVLA.training.train_starvla][INFO] - Step 3228, Loss: {'train/loss': 0.0005853218608535826, 'train/grad_norm_pre_clip': 1.54977285861969, 'timing/data': 0.0001990899909287691, 'timing/model': 0.36502179100352805, 'learning_rate/qwen_vl_interface': 3.249619509373712e-06, 'learning_rate/action_model': 3.249619509373712e-05, 'epoch': 0.24}) +[2026-10-03 15:14:44,105][starVLA.training.train_starvla][INFO] - Step 3229, Loss: {'train/loss': 0.4230065643787384, 'train/grad_norm_pre_clip': 30.349489212036133, 'timing/data': 0.00020417000632733107, 'timing/model': 0.3650701510050567, 'learning_rate/qwen_vl_interface': 3.2468577264372905e-06, 'learning_rate/action_model': 3.24685772643729e-05, 'epoch': 0.24}) +[2026-10-03 15:14:44,752][starVLA.training.train_starvla][INFO] - Step 3230, Loss: {'train/loss': 0.704822838306427, 'train/grad_norm_pre_clip': 249.0956573486328, 'timing/data': 0.00019799001165665686, 'timing/model': 0.3645539699937217, 'learning_rate/qwen_vl_interface': 3.244096766916992e-06, 'learning_rate/action_model': 3.244096766916992e-05, 'epoch': 0.24}) +[2026-10-03 15:14:45,402][starVLA.training.train_starvla][INFO] - Step 3231, Loss: {'train/loss': 8.002496906556189e-05, 'train/grad_norm_pre_clip': 0.01955230161547661, 'timing/data': 0.000208910001674667, 'timing/model': 0.36571647999517154, 'learning_rate/qwen_vl_interface': 3.241336631947743e-06, 'learning_rate/action_model': 3.2413366319477427e-05, 'epoch': 0.24}) +[2026-10-03 15:14:46,251][starVLA.training.train_starvla][INFO] - Step 3232, Loss: {'train/loss': 8.104662992991507e-05, 'train/grad_norm_pre_clip': 0.07264884561300278, 'timing/data': 0.00032877000921871513, 'timing/model': 0.3647158310050145, 'learning_rate/qwen_vl_interface': 3.238577322664131e-06, 'learning_rate/action_model': 3.238577322664131e-05, 'epoch': 0.24}) +[2026-10-03 15:14:46,898][starVLA.training.train_starvla][INFO] - Step 3233, Loss: {'train/loss': 0.0006999185425229371, 'train/grad_norm_pre_clip': 0.6674561500549316, 'timing/data': 0.0002889500028686598, 'timing/model': 0.36457961999985855, 'learning_rate/qwen_vl_interface': 3.2358188402004026e-06, 'learning_rate/action_model': 3.2358188402004024e-05, 'epoch': 0.24}) +[2026-10-03 15:14:47,545][starVLA.training.train_starvla][INFO] - Step 3234, Loss: {'train/loss': 0.00027886469615623355, 'train/grad_norm_pre_clip': 0.11047454923391342, 'timing/data': 0.0001980399974854663, 'timing/model': 0.36463124000874814, 'learning_rate/qwen_vl_interface': 3.2330611856904675e-06, 'learning_rate/action_model': 3.233061185690468e-05, 'epoch': 0.24}) +[2026-10-03 15:14:48,195][starVLA.training.train_starvla][INFO] - Step 3235, Loss: {'train/loss': 7.797045691404492e-05, 'train/grad_norm_pre_clip': 0.02659464441239834, 'timing/data': 0.0002079800033243373, 'timing/model': 0.365492411001469, 'learning_rate/qwen_vl_interface': 3.2303043602678928e-06, 'learning_rate/action_model': 3.230304360267893e-05, 'epoch': 0.24}) +[2026-10-03 15:14:48,843][starVLA.training.train_starvla][INFO] - Step 3236, Loss: {'train/loss': 0.003039457369595766, 'train/grad_norm_pre_clip': 1.564931869506836, 'timing/data': 0.00020337999740149826, 'timing/model': 0.36523015999409836, 'learning_rate/qwen_vl_interface': 3.227548365065905e-06, 'learning_rate/action_model': 3.227548365065905e-05, 'epoch': 0.24}) +[2026-10-03 15:14:49,686][starVLA.training.train_starvla][INFO] - Step 3237, Loss: {'train/loss': 0.00428806571289897, 'train/grad_norm_pre_clip': 26.991600036621094, 'timing/data': 0.00030143000185489655, 'timing/model': 0.5588386250019539, 'learning_rate/qwen_vl_interface': 3.2247932012173887e-06, 'learning_rate/action_model': 3.2247932012173884e-05, 'epoch': 0.24}) +[2026-10-03 15:14:50,333][starVLA.training.train_starvla][INFO] - Step 3238, Loss: {'train/loss': 0.00047359458403661847, 'train/grad_norm_pre_clip': 0.18597741425037384, 'timing/data': 0.0003621299983933568, 'timing/model': 0.36475126999721397, 'learning_rate/qwen_vl_interface': 3.2220388698548897e-06, 'learning_rate/action_model': 3.22203886985489e-05, 'epoch': 0.24}) +[2026-10-03 15:14:50,983][starVLA.training.train_starvla][INFO] - Step 3239, Loss: {'train/loss': 3.532095070113428e-05, 'train/grad_norm_pre_clip': 0.006830933503806591, 'timing/data': 0.00020727000082843006, 'timing/model': 0.3654352209996432, 'learning_rate/qwen_vl_interface': 3.21928537211061e-06, 'learning_rate/action_model': 3.21928537211061e-05, 'epoch': 0.24}) +[2026-10-03 15:14:51,631][starVLA.training.train_starvla][INFO] - Step 3240, Loss: {'train/loss': 0.018806247040629387, 'train/grad_norm_pre_clip': 11.961994171142578, 'timing/data': 0.00024312001187354326, 'timing/model': 0.36500116100069135, 'learning_rate/qwen_vl_interface': 3.216532709116405e-06, 'learning_rate/action_model': 3.216532709116405e-05, 'epoch': 0.24}) +[2026-10-03 15:14:52,279][starVLA.training.train_starvla][INFO] - Step 3241, Loss: {'train/loss': 0.47774362564086914, 'train/grad_norm_pre_clip': 178.05572509765625, 'timing/data': 0.00032297000871039927, 'timing/model': 0.36444674000085797, 'learning_rate/qwen_vl_interface': 3.2137808820037925e-06, 'learning_rate/action_model': 3.2137808820037924e-05, 'epoch': 0.24}) +[2026-10-03 15:14:53,135][starVLA.training.train_starvla][INFO] - Step 3242, Loss: {'train/loss': 0.00018188596004620194, 'train/grad_norm_pre_clip': 0.07119419425725937, 'timing/data': 0.00029707999783568084, 'timing/model': 0.5737274619896198, 'learning_rate/qwen_vl_interface': 3.211029891903945e-06, 'learning_rate/action_model': 3.211029891903945e-05, 'epoch': 0.24}) +[2026-10-03 15:14:53,785][starVLA.training.train_starvla][INFO] - Step 3243, Loss: {'train/loss': 4.327626811573282e-05, 'train/grad_norm_pre_clip': 0.011228240095078945, 'timing/data': 0.00025020999601110816, 'timing/model': 0.365050031003193, 'learning_rate/qwen_vl_interface': 3.2082797399476894e-06, 'learning_rate/action_model': 3.2082797399476895e-05, 'epoch': 0.24}) +[2026-10-03 15:14:54,434][starVLA.training.train_starvla][INFO] - Step 3244, Loss: {'train/loss': 0.14553959667682648, 'train/grad_norm_pre_clip': 88.95397186279297, 'timing/data': 0.00020249999943189323, 'timing/model': 0.3650938509963453, 'learning_rate/qwen_vl_interface': 3.2055304272655114e-06, 'learning_rate/action_model': 3.205530427265511e-05, 'epoch': 0.24}) +[2026-10-03 15:14:55,082][starVLA.training.train_starvla][INFO] - Step 3245, Loss: {'train/loss': 3.628802005550824e-05, 'train/grad_norm_pre_clip': 0.012327057309448719, 'timing/data': 0.0002402199897915125, 'timing/model': 0.36523322100401856, 'learning_rate/qwen_vl_interface': 3.2027819549875477e-06, 'learning_rate/action_model': 3.202781954987548e-05, 'epoch': 0.24}) +[2026-10-03 15:14:55,730][starVLA.training.train_starvla][INFO] - Step 3246, Loss: {'train/loss': 0.00030771602177992463, 'train/grad_norm_pre_clip': 0.08119653165340424, 'timing/data': 0.00029969000024721026, 'timing/model': 0.36487299999862444, 'learning_rate/qwen_vl_interface': 3.2000343242435926e-06, 'learning_rate/action_model': 3.2000343242435924e-05, 'epoch': 0.24}) +[2026-10-03 15:14:56,379][starVLA.training.train_starvla][INFO] - Step 3247, Loss: {'train/loss': 0.00025238815578632057, 'train/grad_norm_pre_clip': 0.08338716626167297, 'timing/data': 0.0002433400077279657, 'timing/model': 0.36549388999992516, 'learning_rate/qwen_vl_interface': 3.1972875361630923e-06, 'learning_rate/action_model': 3.1972875361630923e-05, 'epoch': 0.24}) +[2026-10-03 15:14:57,225][starVLA.training.train_starvla][INFO] - Step 3248, Loss: {'train/loss': 0.00027887217584066093, 'train/grad_norm_pre_clip': 0.0806896910071373, 'timing/data': 0.0003329800092615187, 'timing/model': 0.36476508098712657, 'learning_rate/qwen_vl_interface': 3.1945415918751484e-06, 'learning_rate/action_model': 3.194541591875148e-05, 'epoch': 0.24}) +[2026-10-03 15:14:57,873][starVLA.training.train_starvla][INFO] - Step 3249, Loss: {'train/loss': 0.0005053233471699059, 'train/grad_norm_pre_clip': 0.37988489866256714, 'timing/data': 0.0003032789973076433, 'timing/model': 0.36479776100895833, 'learning_rate/qwen_vl_interface': 3.191796492508516e-06, 'learning_rate/action_model': 3.191796492508516e-05, 'epoch': 0.24}) +[2026-10-03 15:14:58,521][starVLA.training.train_starvla][INFO] - Step 3250, Loss: {'train/loss': 5.205631896387786e-05, 'train/grad_norm_pre_clip': 0.020373143255710602, 'timing/data': 0.00020663000759668648, 'timing/model': 0.3651026110019302, 'learning_rate/qwen_vl_interface': 3.189052239191599e-06, 'learning_rate/action_model': 3.189052239191599e-05, 'epoch': 0.24}) +[2026-10-03 15:14:59,170][starVLA.training.train_starvla][INFO] - Step 3251, Loss: {'train/loss': 0.00015736687055323273, 'train/grad_norm_pre_clip': 0.08939588814973831, 'timing/data': 0.0003395999956410378, 'timing/model': 0.3651900109980488, 'learning_rate/qwen_vl_interface': 3.1863088330524577e-06, 'learning_rate/action_model': 3.186308833052458e-05, 'epoch': 0.24}) +[2026-10-03 15:14:59,819][starVLA.training.train_starvla][INFO] - Step 3252, Loss: {'train/loss': 0.31250515580177307, 'train/grad_norm_pre_clip': 115.07108306884766, 'timing/data': 0.000371729998732917, 'timing/model': 0.3654609799996251, 'learning_rate/qwen_vl_interface': 3.183566275218803e-06, 'learning_rate/action_model': 3.1835662752188026e-05, 'epoch': 0.24}) +[2026-10-03 15:15:00,667][starVLA.training.train_starvla][INFO] - Step 3253, Loss: {'train/loss': 8.526066085323691e-05, 'train/grad_norm_pre_clip': 0.04092912748456001, 'timing/data': 0.00023988999600987881, 'timing/model': 0.3649391099897912, 'learning_rate/qwen_vl_interface': 3.180824566817996e-06, 'learning_rate/action_model': 3.1808245668179956e-05, 'epoch': 0.24}) +[2026-10-03 15:15:01,314][starVLA.training.train_starvla][INFO] - Step 3254, Loss: {'train/loss': 0.5715444087982178, 'train/grad_norm_pre_clip': 55.08011245727539, 'timing/data': 0.0003052600077353418, 'timing/model': 0.36436530099308584, 'learning_rate/qwen_vl_interface': 3.1780837089770524e-06, 'learning_rate/action_model': 3.1780837089770524e-05, 'epoch': 0.24}) +[2026-10-03 15:15:01,964][starVLA.training.train_starvla][INFO] - Step 3255, Loss: {'train/loss': 0.4624705910682678, 'train/grad_norm_pre_clip': 124.71745300292969, 'timing/data': 0.00026344000070821494, 'timing/model': 0.3658047609933419, 'learning_rate/qwen_vl_interface': 3.17534370282263e-06, 'learning_rate/action_model': 3.1753437028226303e-05, 'epoch': 0.24}) +[2026-10-03 15:15:02,589][starVLA.training.train_starvla][INFO] - Step 3256, Loss: {'train/loss': 8.184704347513616e-05, 'train/grad_norm_pre_clip': 0.03032979555428028, 'timing/data': 0.000251939010922797, 'timing/model': 0.34191009400819894, 'learning_rate/qwen_vl_interface': 3.172604549481046e-06, 'learning_rate/action_model': 3.1726045494810455e-05, 'epoch': 0.24}) +[2026-10-03 15:15:03,237][starVLA.training.train_starvla][INFO] - Step 3257, Loss: {'train/loss': 0.06570953875780106, 'train/grad_norm_pre_clip': 312.7525329589844, 'timing/data': 0.0003495400014799088, 'timing/model': 0.3647477710037492, 'learning_rate/qwen_vl_interface': 3.169866250078261e-06, 'learning_rate/action_model': 3.169866250078261e-05, 'epoch': 0.24}) +[2026-10-03 15:15:04,077][starVLA.training.train_starvla][INFO] - Step 3258, Loss: {'train/loss': 0.4472425878047943, 'train/grad_norm_pre_clip': 24.53349494934082, 'timing/data': 0.0002470399922458455, 'timing/model': 0.5571787140070228, 'learning_rate/qwen_vl_interface': 3.167128805739887e-06, 'learning_rate/action_model': 3.1671288057398866e-05, 'epoch': 0.24}) +[2026-10-03 15:15:04,727][starVLA.training.train_starvla][INFO] - Step 3259, Loss: {'train/loss': 0.00020185492758173496, 'train/grad_norm_pre_clip': 0.07961108535528183, 'timing/data': 0.00021518999710679054, 'timing/model': 0.3652403000014601, 'learning_rate/qwen_vl_interface': 3.1643922175911847e-06, 'learning_rate/action_model': 3.164392217591185e-05, 'epoch': 0.24}) +[2026-10-03 15:15:05,375][starVLA.training.train_starvla][INFO] - Step 3260, Loss: {'train/loss': 0.00032978897797875106, 'train/grad_norm_pre_clip': 0.2271592915058136, 'timing/data': 0.0003233500028727576, 'timing/model': 0.36477583000669256, 'learning_rate/qwen_vl_interface': 3.1616564867570603e-06, 'learning_rate/action_model': 3.16165648675706e-05, 'epoch': 0.24}) +[2026-10-03 15:15:06,022][starVLA.training.train_starvla][INFO] - Step 3261, Loss: {'train/loss': 0.29777175188064575, 'train/grad_norm_pre_clip': 41.06451416015625, 'timing/data': 0.00034077999589499086, 'timing/model': 0.36432778000016697, 'learning_rate/qwen_vl_interface': 3.1589216143620733e-06, 'learning_rate/action_model': 3.1589216143620734e-05, 'epoch': 0.24}) +[2026-10-03 15:15:06,647][starVLA.training.train_starvla][INFO] - Step 3262, Loss: {'train/loss': 9.999926260206848e-05, 'train/grad_norm_pre_clip': 0.053924210369586945, 'timing/data': 0.00033479000558145344, 'timing/model': 0.3421052940102527, 'learning_rate/qwen_vl_interface': 3.1561876015304215e-06, 'learning_rate/action_model': 3.156187601530421e-05, 'epoch': 0.24}) +[2026-10-03 15:15:07,273][starVLA.training.train_starvla][INFO] - Step 3263, Loss: {'train/loss': 0.00010850813123397529, 'train/grad_norm_pre_clip': 0.057021573185920715, 'timing/data': 0.00025100998755078763, 'timing/model': 0.3647194310033228, 'learning_rate/qwen_vl_interface': 3.1534544493859563e-06, 'learning_rate/action_model': 3.1534544493859566e-05, 'epoch': 0.24}) +[2026-10-03 15:15:08,097][starVLA.training.train_starvla][INFO] - Step 3264, Loss: {'train/loss': 0.4591335952281952, 'train/grad_norm_pre_clip': 128.63272094726562, 'timing/data': 0.00033304998942185193, 'timing/model': 0.34168837300967425, 'learning_rate/qwen_vl_interface': 3.150722159052174e-06, 'learning_rate/action_model': 3.150722159052174e-05, 'epoch': 0.24}) +[2026-10-03 15:15:08,745][starVLA.training.train_starvla][INFO] - Step 3265, Loss: {'train/loss': 0.00010570951417321339, 'train/grad_norm_pre_clip': 0.03077571839094162, 'timing/data': 0.00019596000493038446, 'timing/model': 0.36476656098966487, 'learning_rate/qwen_vl_interface': 3.147990731652215e-06, 'learning_rate/action_model': 3.147990731652215e-05, 'epoch': 0.24}) +[2026-10-03 15:15:09,392][starVLA.training.train_starvla][INFO] - Step 3266, Loss: {'train/loss': 0.000218246117583476, 'train/grad_norm_pre_clip': 0.26562774181365967, 'timing/data': 0.0003360800037626177, 'timing/model': 0.36468988000706304, 'learning_rate/qwen_vl_interface': 3.1452601683088673e-06, 'learning_rate/action_model': 3.145260168308867e-05, 'epoch': 0.24}) +[2026-10-03 15:15:10,043][starVLA.training.train_starvla][INFO] - Step 3267, Loss: {'train/loss': 0.00020898085494991392, 'train/grad_norm_pre_clip': 0.10316925495862961, 'timing/data': 0.00029595999512821436, 'timing/model': 0.3654677599988645, 'learning_rate/qwen_vl_interface': 3.1425304701445608e-06, 'learning_rate/action_model': 3.1425304701445605e-05, 'epoch': 0.24}) +[2026-10-03 15:15:10,690][starVLA.training.train_starvla][INFO] - Step 3268, Loss: {'train/loss': 0.2349635660648346, 'train/grad_norm_pre_clip': 97.56797790527344, 'timing/data': 0.00032768899109214544, 'timing/model': 0.36435048100247514, 'learning_rate/qwen_vl_interface': 3.139801638281372e-06, 'learning_rate/action_model': 3.139801638281372e-05, 'epoch': 0.24}) +[2026-10-03 15:15:11,532][starVLA.training.train_starvla][INFO] - Step 3269, Loss: {'train/loss': 3.763737186091021e-05, 'train/grad_norm_pre_clip': 0.01233722548931837, 'timing/data': 0.00033070000063162297, 'timing/model': 0.5587839939980768, 'learning_rate/qwen_vl_interface': 3.13707367384102e-06, 'learning_rate/action_model': 3.13707367384102e-05, 'epoch': 0.24}) +[2026-10-03 15:15:12,179][starVLA.training.train_starvla][INFO] - Step 3270, Loss: {'train/loss': 0.0004533258907031268, 'train/grad_norm_pre_clip': 0.10675503313541412, 'timing/data': 0.0003143900103168562, 'timing/model': 0.3641821710043587, 'learning_rate/qwen_vl_interface': 3.134346577944869e-06, 'learning_rate/action_model': 3.134346577944869e-05, 'epoch': 0.24}) +[2026-10-03 15:15:12,829][starVLA.training.train_starvla][INFO] - Step 3271, Loss: {'train/loss': 0.00041697698179632425, 'train/grad_norm_pre_clip': 0.3353941738605499, 'timing/data': 0.00029773000278510153, 'timing/model': 0.3651896010123892, 'learning_rate/qwen_vl_interface': 3.1316203517139243e-06, 'learning_rate/action_model': 3.131620351713924e-05, 'epoch': 0.24}) +[2026-10-03 15:15:13,477][starVLA.training.train_starvla][INFO] - Step 3272, Loss: {'train/loss': 0.00428549200296402, 'train/grad_norm_pre_clip': 5.201427459716797, 'timing/data': 0.0002507600001990795, 'timing/model': 0.3651651899999706, 'learning_rate/qwen_vl_interface': 3.128894996268837e-06, 'learning_rate/action_model': 3.128894996268837e-05, 'epoch': 0.24}) +[2026-10-03 15:15:14,125][starVLA.training.train_starvla][INFO] - Step 3273, Loss: {'train/loss': 0.5738775730133057, 'train/grad_norm_pre_clip': 51.15922927856445, 'timing/data': 0.000331000002915971, 'timing/model': 0.36455923100584187, 'learning_rate/qwen_vl_interface': 3.1261705127298934e-06, 'learning_rate/action_model': 3.1261705127298935e-05, 'epoch': 0.24}) +[2026-10-03 15:15:14,773][starVLA.training.train_starvla][INFO] - Step 3274, Loss: {'train/loss': 0.0005531577626243234, 'train/grad_norm_pre_clip': 0.3073991537094116, 'timing/data': 0.000245510003878735, 'timing/model': 0.3646456700080307, 'learning_rate/qwen_vl_interface': 3.123446902217028e-06, 'learning_rate/action_model': 3.1234469022170283e-05, 'epoch': 0.24}) +[2026-10-03 15:15:15,592][starVLA.training.train_starvla][INFO] - Step 3275, Loss: {'train/loss': 0.00028474448481574655, 'train/grad_norm_pre_clip': 0.08738694339990616, 'timing/data': 0.0003106800140812993, 'timing/model': 0.36489399000129197, 'learning_rate/qwen_vl_interface': 3.1207241658498145e-06, 'learning_rate/action_model': 3.1207241658498145e-05, 'epoch': 0.24}) +[2026-10-03 15:15:16,217][starVLA.training.train_starvla][INFO] - Step 3276, Loss: {'train/loss': 0.0002512433857191354, 'train/grad_norm_pre_clip': 0.14570221304893494, 'timing/data': 0.00032872000883799046, 'timing/model': 0.36413616000209004, 'learning_rate/qwen_vl_interface': 3.1180023047474667e-06, 'learning_rate/action_model': 3.1180023047474664e-05, 'epoch': 0.24}) +[2026-10-03 15:15:16,864][starVLA.training.train_starvla][INFO] - Step 3277, Loss: {'train/loss': 0.0009708115831017494, 'train/grad_norm_pre_clip': 0.6100376844406128, 'timing/data': 0.00036669000110123307, 'timing/model': 0.36404207099985797, 'learning_rate/qwen_vl_interface': 3.1152813200288386e-06, 'learning_rate/action_model': 3.115281320028838e-05, 'epoch': 0.24}) +[2026-10-03 15:15:17,512][starVLA.training.train_starvla][INFO] - Step 3278, Loss: {'train/loss': 2.269386823172681e-05, 'train/grad_norm_pre_clip': 0.003417483065277338, 'timing/data': 0.00019764999160543084, 'timing/model': 0.3650603099958971, 'learning_rate/qwen_vl_interface': 3.1125612128124244e-06, 'learning_rate/action_model': 3.112561212812424e-05, 'epoch': 0.24}) +[2026-10-03 15:15:18,162][starVLA.training.train_starvla][INFO] - Step 3279, Loss: {'train/loss': 0.04839111119508743, 'train/grad_norm_pre_clip': 81.22734832763672, 'timing/data': 0.0003074799897149205, 'timing/model': 0.365311469999142, 'learning_rate/qwen_vl_interface': 3.1098419842163573e-06, 'learning_rate/action_model': 3.109841984216357e-05, 'epoch': 0.24}) +[2026-10-03 15:15:19,009][starVLA.training.train_starvla][INFO] - Step 3280, Loss: {'train/loss': 0.00047002482460811734, 'train/grad_norm_pre_clip': 0.30462905764579773, 'timing/data': 0.0003325199941173196, 'timing/model': 0.3645685709925601, 'learning_rate/qwen_vl_interface': 3.10712363535841e-06, 'learning_rate/action_model': 3.10712363535841e-05, 'epoch': 0.24}) +[2026-10-03 15:15:19,657][starVLA.training.train_starvla][INFO] - Step 3281, Loss: {'train/loss': 0.00020112641504965723, 'train/grad_norm_pre_clip': 0.09573628008365631, 'timing/data': 0.00024913999368436635, 'timing/model': 0.3649128010001732, 'learning_rate/qwen_vl_interface': 3.1044061673559926e-06, 'learning_rate/action_model': 3.1044061673559925e-05, 'epoch': 0.24}) +[2026-10-03 15:15:20,304][starVLA.training.train_starvla][INFO] - Step 3282, Loss: {'train/loss': 2.460099130985327e-05, 'train/grad_norm_pre_clip': 0.005265029612928629, 'timing/data': 0.00022031999833416194, 'timing/model': 0.3649604110105429, 'learning_rate/qwen_vl_interface': 3.1016895813261556e-06, 'learning_rate/action_model': 3.101689581326155e-05, 'epoch': 0.24}) +[2026-10-03 15:15:20,955][starVLA.training.train_starvla][INFO] - Step 3283, Loss: {'train/loss': 0.23680603504180908, 'train/grad_norm_pre_clip': 41.732398986816406, 'timing/data': 0.00027815999055746943, 'timing/model': 0.3655125700024655, 'learning_rate/qwen_vl_interface': 3.098973878385581e-06, 'learning_rate/action_model': 3.098973878385581e-05, 'epoch': 0.24}) +[2026-10-03 15:15:21,580][starVLA.training.train_starvla][INFO] - Step 3284, Loss: {'train/loss': 7.784059562254697e-05, 'train/grad_norm_pre_clip': 0.07792767137289047, 'timing/data': 0.00020411999139469117, 'timing/model': 0.34223615399969276, 'learning_rate/qwen_vl_interface': 3.096259059650594e-06, 'learning_rate/action_model': 3.096259059650594e-05, 'epoch': 0.24}) +[2026-10-03 15:15:22,422][starVLA.training.train_starvla][INFO] - Step 3285, Loss: {'train/loss': 1.3842980479239486e-05, 'train/grad_norm_pre_clip': 0.0031708870083093643, 'timing/data': 0.000245360002736561, 'timing/model': 0.5588745639979606, 'learning_rate/qwen_vl_interface': 3.093545126237153e-06, 'learning_rate/action_model': 3.0935451262371527e-05, 'epoch': 0.24}) +[2026-10-03 15:15:23,069][starVLA.training.train_starvla][INFO] - Step 3286, Loss: {'train/loss': 8.259508467745036e-05, 'train/grad_norm_pre_clip': 0.05124653875827789, 'timing/data': 0.00024574001145083457, 'timing/model': 0.36467071100196335, 'learning_rate/qwen_vl_interface': 3.090832079260855e-06, 'learning_rate/action_model': 3.0908320792608544e-05, 'epoch': 0.24}) +[2026-10-03 15:15:23,720][starVLA.training.train_starvla][INFO] - Step 3287, Loss: {'train/loss': 0.4177578091621399, 'train/grad_norm_pre_clip': 70.90507507324219, 'timing/data': 0.000381799996830523, 'timing/model': 0.3654975710087456, 'learning_rate/qwen_vl_interface': 3.088119919836929e-06, 'learning_rate/action_model': 3.088119919836929e-05, 'epoch': 0.24}) +[2026-10-03 15:15:24,345][starVLA.training.train_starvla][INFO] - Step 3288, Loss: {'train/loss': 0.0003189757699146867, 'train/grad_norm_pre_clip': 0.1678815484046936, 'timing/data': 0.00035320999450050294, 'timing/model': 0.34199793500010855, 'learning_rate/qwen_vl_interface': 3.0854086490802433e-06, 'learning_rate/action_model': 3.085408649080243e-05, 'epoch': 0.24}) +[2026-10-03 15:15:24,992][starVLA.training.train_starvla][INFO] - Step 3289, Loss: {'train/loss': 1.4141050996840931e-05, 'train/grad_norm_pre_clip': 0.0032382758799940348, 'timing/data': 0.00030264999077189714, 'timing/model': 0.364143630999024, 'learning_rate/qwen_vl_interface': 3.082698268105299e-06, 'learning_rate/action_model': 3.082698268105299e-05, 'epoch': 0.24}) +[2026-10-03 15:15:25,841][starVLA.training.train_starvla][INFO] - Step 3290, Loss: {'train/loss': 1.7366939573548734e-05, 'train/grad_norm_pre_clip': 0.005981724243611097, 'timing/data': 0.00020030999439768493, 'timing/model': 0.5667067530011991, 'learning_rate/qwen_vl_interface': 3.0799887780262295e-06, 'learning_rate/action_model': 3.0799887780262296e-05, 'epoch': 0.24}) +[2026-10-03 15:15:26,491][starVLA.training.train_starvla][INFO] - Step 3291, Loss: {'train/loss': 0.0003229270223528147, 'train/grad_norm_pre_clip': 0.18815255165100098, 'timing/data': 0.0002668100059963763, 'timing/model': 0.36523689099703915, 'learning_rate/qwen_vl_interface': 3.0772801799568054e-06, 'learning_rate/action_model': 3.0772801799568054e-05, 'epoch': 0.24}) +[2026-10-03 15:15:27,138][starVLA.training.train_starvla][INFO] - Step 3292, Loss: {'train/loss': 0.09969781339168549, 'train/grad_norm_pre_clip': 74.84548950195312, 'timing/data': 0.00023676999262534082, 'timing/model': 0.3649277199874632, 'learning_rate/qwen_vl_interface': 3.07457247501043e-06, 'learning_rate/action_model': 3.0745724750104296e-05, 'epoch': 0.24}) +[2026-10-03 15:15:27,785][starVLA.training.train_starvla][INFO] - Step 3293, Loss: {'train/loss': 0.08681623637676239, 'train/grad_norm_pre_clip': 117.95918273925781, 'timing/data': 0.00030703000084031373, 'timing/model': 0.36448084098810796, 'learning_rate/qwen_vl_interface': 3.0718656643001353e-06, 'learning_rate/action_model': 3.071865664300135e-05, 'epoch': 0.24}) +[2026-10-03 15:15:28,433][starVLA.training.train_starvla][INFO] - Step 3294, Loss: {'train/loss': 0.028605423867702484, 'train/grad_norm_pre_clip': 54.350860595703125, 'timing/data': 0.0002366499975323677, 'timing/model': 0.36448583099991083, 'learning_rate/qwen_vl_interface': 3.069159748938591e-06, 'learning_rate/action_model': 3.069159748938591e-05, 'epoch': 0.24}) +[2026-10-03 15:15:29,083][starVLA.training.train_starvla][INFO] - Step 3295, Loss: {'train/loss': 0.175770103931427, 'train/grad_norm_pre_clip': 134.48825073242188, 'timing/data': 0.00024778000079095364, 'timing/model': 0.3647875610040501, 'learning_rate/qwen_vl_interface': 3.066454730038096e-06, 'learning_rate/action_model': 3.0664547300380956e-05, 'epoch': 0.24}) +[2026-10-03 15:15:29,938][starVLA.training.train_starvla][INFO] - Step 3296, Loss: {'train/loss': 0.0001601598778506741, 'train/grad_norm_pre_clip': 0.11197623610496521, 'timing/data': 0.00031275000947061926, 'timing/model': 0.36473536099947523, 'learning_rate/qwen_vl_interface': 3.0637506087105816e-06, 'learning_rate/action_model': 3.0637506087105815e-05, 'epoch': 0.24}) +[2026-10-03 15:15:30,586][starVLA.training.train_starvla][INFO] - Step 3297, Loss: {'train/loss': 5.3902251238469034e-05, 'train/grad_norm_pre_clip': 0.02216716855764389, 'timing/data': 0.00034109999251086265, 'timing/model': 0.3644271410012152, 'learning_rate/qwen_vl_interface': 3.0610473860676103e-06, 'learning_rate/action_model': 3.06104738606761e-05, 'epoch': 0.24}) +[2026-10-03 15:15:31,232][starVLA.training.train_starvla][INFO] - Step 3298, Loss: {'train/loss': 0.3671499490737915, 'train/grad_norm_pre_clip': 69.12723541259766, 'timing/data': 0.00025786001060623676, 'timing/model': 0.364216250993195, 'learning_rate/qwen_vl_interface': 3.0583450632203748e-06, 'learning_rate/action_model': 3.058345063220375e-05, 'epoch': 0.24}) +[2026-10-03 15:15:31,883][starVLA.training.train_starvla][INFO] - Step 3299, Loss: {'train/loss': 0.00020881701493635774, 'train/grad_norm_pre_clip': 0.13730895519256592, 'timing/data': 0.0004497199988691136, 'timing/model': 0.3653848609974375, 'learning_rate/qwen_vl_interface': 3.055643641279699e-06, 'learning_rate/action_model': 3.055643641279699e-05, 'epoch': 0.24}) +[2026-10-03 15:15:32,531][starVLA.training.train_starvla][INFO] - Step 3300, Loss: {'train/loss': 5.57032399228774e-05, 'train/grad_norm_pre_clip': 0.026818692684173584, 'timing/data': 0.00029476999770849943, 'timing/model': 0.364683480001986, 'learning_rate/qwen_vl_interface': 3.0529431213560334e-06, 'learning_rate/action_model': 3.0529431213560334e-05, 'epoch': 0.24}) +[2026-10-03 15:15:33,381][starVLA.training.train_starvla][INFO] - Step 3301, Loss: {'train/loss': 3.157434912282042e-05, 'train/grad_norm_pre_clip': 0.009425681084394455, 'timing/data': 0.00025074000586755574, 'timing/model': 0.364366000998416, 'learning_rate/qwen_vl_interface': 3.0502435045594596e-06, 'learning_rate/action_model': 3.0502435045594595e-05, 'epoch': 0.24}) +[2026-10-03 15:15:34,005][starVLA.training.train_starvla][INFO] - Step 3302, Loss: {'train/loss': 0.00041269598295912147, 'train/grad_norm_pre_clip': 0.1819615513086319, 'timing/data': 0.00024002000282052904, 'timing/model': 0.3642907409957843, 'learning_rate/qwen_vl_interface': 3.047544791999689e-06, 'learning_rate/action_model': 3.0475447919996892e-05, 'epoch': 0.24}) +[2026-10-03 15:15:34,632][starVLA.training.train_starvla][INFO] - Step 3303, Loss: {'train/loss': 0.35356467962265015, 'train/grad_norm_pre_clip': 192.1741180419922, 'timing/data': 0.00020602000586222857, 'timing/model': 0.3649918709998019, 'learning_rate/qwen_vl_interface': 3.044846984786061e-06, 'learning_rate/action_model': 3.044846984786061e-05, 'epoch': 0.24}) +[2026-10-03 15:15:35,279][starVLA.training.train_starvla][INFO] - Step 3304, Loss: {'train/loss': 0.36749255657196045, 'train/grad_norm_pre_clip': 30.646148681640625, 'timing/data': 0.0002503299911040813, 'timing/model': 0.3640498610038776, 'learning_rate/qwen_vl_interface': 3.0421500840275413e-06, 'learning_rate/action_model': 3.0421500840275413e-05, 'epoch': 0.24}) +[2026-10-03 15:15:35,926][starVLA.training.train_starvla][INFO] - Step 3305, Loss: {'train/loss': 0.22169317305088043, 'train/grad_norm_pre_clip': 50.64241027832031, 'timing/data': 0.0003201999934390187, 'timing/model': 0.36460556100064423, 'learning_rate/qwen_vl_interface': 3.039454090832727e-06, 'learning_rate/action_model': 3.0394540908327268e-05, 'epoch': 0.24}) +[2026-10-03 15:15:36,739][starVLA.training.train_starvla][INFO] - Step 3306, Loss: {'train/loss': 0.0004749442741740495, 'train/grad_norm_pre_clip': 0.3286767303943634, 'timing/data': 0.00023347999376710504, 'timing/model': 0.5308335190056823, 'learning_rate/qwen_vl_interface': 3.0367590063098333e-06, 'learning_rate/action_model': 3.0367590063098333e-05, 'epoch': 0.24}) +[2026-10-03 15:15:37,388][starVLA.training.train_starvla][INFO] - Step 3307, Loss: {'train/loss': 0.0008147405460476875, 'train/grad_norm_pre_clip': 0.36293941736221313, 'timing/data': 0.0003210300055798143, 'timing/model': 0.3643846610066248, 'learning_rate/qwen_vl_interface': 3.0340648315667115e-06, 'learning_rate/action_model': 3.0340648315667113e-05, 'epoch': 0.24}) +[2026-10-03 15:15:38,035][starVLA.training.train_starvla][INFO] - Step 3308, Loss: {'train/loss': 0.10454698652029037, 'train/grad_norm_pre_clip': 235.76065063476562, 'timing/data': 0.0003186299873050302, 'timing/model': 0.36452144099166617, 'learning_rate/qwen_vl_interface': 3.0313715677108336e-06, 'learning_rate/action_model': 3.0313715677108335e-05, 'epoch': 0.24}) +[2026-10-03 15:15:38,683][starVLA.training.train_starvla][INFO] - Step 3309, Loss: {'train/loss': 0.0006235840846784413, 'train/grad_norm_pre_clip': 0.3109916150569916, 'timing/data': 0.00024773999757599086, 'timing/model': 0.36473799000668805, 'learning_rate/qwen_vl_interface': 3.0286792158492996e-06, 'learning_rate/action_model': 3.0286792158492994e-05, 'epoch': 0.24}) +[2026-10-03 15:15:39,330][starVLA.training.train_starvla][INFO] - Step 3310, Loss: {'train/loss': 0.0002423763507977128, 'train/grad_norm_pre_clip': 0.08016126602888107, 'timing/data': 0.00024247000692412257, 'timing/model': 0.3642661910125753, 'learning_rate/qwen_vl_interface': 3.0259877770888334e-06, 'learning_rate/action_model': 3.0259877770888333e-05, 'epoch': 0.24}) +[2026-10-03 15:15:39,980][starVLA.training.train_starvla][INFO] - Step 3311, Loss: {'train/loss': 0.6572821736335754, 'train/grad_norm_pre_clip': 40.02676010131836, 'timing/data': 0.0002459499955875799, 'timing/model': 0.36523112100258004, 'learning_rate/qwen_vl_interface': 3.0232972525357836e-06, 'learning_rate/action_model': 3.0232972525357834e-05, 'epoch': 0.24}) +[2026-10-03 15:15:40,827][starVLA.training.train_starvla][INFO] - Step 3312, Loss: {'train/loss': 0.2760222554206848, 'train/grad_norm_pre_clip': 110.36396026611328, 'timing/data': 0.0004171600012341514, 'timing/model': 0.3649632899905555, 'learning_rate/qwen_vl_interface': 3.020607643296123e-06, 'learning_rate/action_model': 3.0206076432961226e-05, 'epoch': 0.24}) +[2026-10-03 15:15:41,451][starVLA.training.train_starvla][INFO] - Step 3313, Loss: {'train/loss': 9.823773143580183e-05, 'train/grad_norm_pre_clip': 0.038736939430236816, 'timing/data': 0.00023975998919922858, 'timing/model': 0.3638301409955602, 'learning_rate/qwen_vl_interface': 3.017918950475449e-06, 'learning_rate/action_model': 3.017918950475449e-05, 'epoch': 0.24}) +[2026-10-03 15:15:42,076][starVLA.training.train_starvla][INFO] - Step 3314, Loss: {'train/loss': 7.76291562942788e-05, 'train/grad_norm_pre_clip': 0.05333782732486725, 'timing/data': 0.0003023300087079406, 'timing/model': 0.364935391000472, 'learning_rate/qwen_vl_interface': 3.0152311751789815e-06, 'learning_rate/action_model': 3.0152311751789814e-05, 'epoch': 0.24}) +[2026-10-03 15:15:42,726][starVLA.training.train_starvla][INFO] - Step 3315, Loss: {'train/loss': 0.00011534098302945495, 'train/grad_norm_pre_clip': 0.044370632618665695, 'timing/data': 0.0002812790044117719, 'timing/model': 0.3652082309999969, 'learning_rate/qwen_vl_interface': 3.0125443185115655e-06, 'learning_rate/action_model': 3.0125443185115655e-05, 'epoch': 0.24}) +[2026-10-03 15:15:43,374][starVLA.training.train_starvla][INFO] - Step 3316, Loss: {'train/loss': 0.37505587935447693, 'train/grad_norm_pre_clip': 36.03470993041992, 'timing/data': 0.0002350799913983792, 'timing/model': 0.36508783099998254, 'learning_rate/qwen_vl_interface': 3.0098583815776615e-06, 'learning_rate/action_model': 3.0098583815776615e-05, 'epoch': 0.24}) +[2026-10-03 15:15:44,224][starVLA.training.train_starvla][INFO] - Step 3317, Loss: {'train/loss': 0.020536433905363083, 'train/grad_norm_pre_clip': 30.955732345581055, 'timing/data': 0.0002419500087853521, 'timing/model': 0.5668608939886326, 'learning_rate/qwen_vl_interface': 3.0071733654813595e-06, 'learning_rate/action_model': 3.0071733654813595e-05, 'epoch': 0.24}) +[2026-10-03 15:15:44,871][starVLA.training.train_starvla][INFO] - Step 3318, Loss: {'train/loss': 0.007395957596600056, 'train/grad_norm_pre_clip': 5.286248207092285, 'timing/data': 0.00030575999699067324, 'timing/model': 0.36450124099792447, 'learning_rate/qwen_vl_interface': 3.0044892713263685e-06, 'learning_rate/action_model': 3.0044892713263682e-05, 'epoch': 0.24}) +[2026-10-03 15:15:45,522][starVLA.training.train_starvla][INFO] - Step 3319, Loss: {'train/loss': 0.0005818470381200314, 'train/grad_norm_pre_clip': 0.2918577790260315, 'timing/data': 0.00029673999233637005, 'timing/model': 0.36536394100403413, 'learning_rate/qwen_vl_interface': 3.001806100216017e-06, 'learning_rate/action_model': 3.001806100216017e-05, 'epoch': 0.24}) +[2026-10-03 15:15:46,171][starVLA.training.train_starvla][INFO] - Step 3320, Loss: {'train/loss': 0.08884794265031815, 'train/grad_norm_pre_clip': 78.3472900390625, 'timing/data': 0.0003436800034251064, 'timing/model': 0.3652186609979253, 'learning_rate/qwen_vl_interface': 2.9991238532532567e-06, 'learning_rate/action_model': 2.9991238532532567e-05, 'epoch': 0.24}) +[2026-10-03 15:15:46,819][starVLA.training.train_starvla][INFO] - Step 3321, Loss: {'train/loss': 0.0007373856496997178, 'train/grad_norm_pre_clip': 0.3393503427505493, 'timing/data': 0.00028230999305378646, 'timing/model': 0.36494746000971645, 'learning_rate/qwen_vl_interface': 2.996442531540657e-06, 'learning_rate/action_model': 2.9964425315406568e-05, 'epoch': 0.24}) +[2026-10-03 15:15:47,677][starVLA.training.train_starvla][INFO] - Step 3322, Loss: {'train/loss': 0.00021395586372818798, 'train/grad_norm_pre_clip': 0.13236719369888306, 'timing/data': 0.0002266100054839626, 'timing/model': 0.5755064330005553, 'learning_rate/qwen_vl_interface': 2.993762136180408e-06, 'learning_rate/action_model': 2.993762136180408e-05, 'epoch': 0.24}) +[2026-10-03 15:15:48,327][starVLA.training.train_starvla][INFO] - Step 3323, Loss: {'train/loss': 0.47533246874809265, 'train/grad_norm_pre_clip': 194.9429473876953, 'timing/data': 0.00022632999753113836, 'timing/model': 0.36564180100685917, 'learning_rate/qwen_vl_interface': 2.9910826682743195e-06, 'learning_rate/action_model': 2.9910826682743192e-05, 'epoch': 0.24}) +[2026-10-03 15:15:48,974][starVLA.training.train_starvla][INFO] - Step 3324, Loss: {'train/loss': 0.000552377721760422, 'train/grad_norm_pre_clip': 0.7154332399368286, 'timing/data': 0.00022307000472210348, 'timing/model': 0.36469028099963907, 'learning_rate/qwen_vl_interface': 2.9884041289238192e-06, 'learning_rate/action_model': 2.988404128923819e-05, 'epoch': 0.24}) +[2026-10-03 15:15:49,622][starVLA.training.train_starvla][INFO] - Step 3325, Loss: {'train/loss': 0.00013066835526842624, 'train/grad_norm_pre_clip': 0.08405158668756485, 'timing/data': 0.0003688700089696795, 'timing/model': 0.3649078510061372, 'learning_rate/qwen_vl_interface': 2.9857265192299557e-06, 'learning_rate/action_model': 2.9857265192299556e-05, 'epoch': 0.24}) +[2026-10-03 15:15:50,271][starVLA.training.train_starvla][INFO] - Step 3326, Loss: {'train/loss': 0.00035113716148771346, 'train/grad_norm_pre_clip': 0.24351750314235687, 'timing/data': 0.00026405999960843474, 'timing/model': 0.3663672310067341, 'learning_rate/qwen_vl_interface': 2.9830498402933894e-06, 'learning_rate/action_model': 2.9830498402933894e-05, 'epoch': 0.24}) +[2026-10-03 15:15:50,922][starVLA.training.train_starvla][INFO] - Step 3327, Loss: {'train/loss': 2.681335899978876e-05, 'train/grad_norm_pre_clip': 0.00694095715880394, 'timing/data': 0.000318689999403432, 'timing/model': 0.3657928509928752, 'learning_rate/qwen_vl_interface': 2.9803740932144025e-06, 'learning_rate/action_model': 2.9803740932144024e-05, 'epoch': 0.24}) +[2026-10-03 15:15:51,766][starVLA.training.train_starvla][INFO] - Step 3328, Loss: {'train/loss': 3.151448618154973e-05, 'train/grad_norm_pre_clip': 0.010575128719210625, 'timing/data': 0.0003933400003006682, 'timing/model': 0.3651940510026179, 'learning_rate/qwen_vl_interface': 2.9776992790928953e-06, 'learning_rate/action_model': 2.9776992790928952e-05, 'epoch': 0.24}) +[2026-10-03 15:15:52,414][starVLA.training.train_starvla][INFO] - Step 3329, Loss: {'train/loss': 9.493771358393133e-05, 'train/grad_norm_pre_clip': 0.052485838532447815, 'timing/data': 0.00033966999035328627, 'timing/model': 0.36451823099923786, 'learning_rate/qwen_vl_interface': 2.975025399028382e-06, 'learning_rate/action_model': 2.975025399028382e-05, 'epoch': 0.24}) +[2026-10-03 15:15:53,060][starVLA.training.train_starvla][INFO] - Step 3330, Loss: {'train/loss': 0.011578737758100033, 'train/grad_norm_pre_clip': 14.482548713684082, 'timing/data': 0.0002746300015132874, 'timing/model': 0.36419811099767685, 'learning_rate/qwen_vl_interface': 2.9723524541199943e-06, 'learning_rate/action_model': 2.972352454119994e-05, 'epoch': 0.24}) +[2026-10-03 15:15:53,709][starVLA.training.train_starvla][INFO] - Step 3331, Loss: {'train/loss': 0.00016004244389478117, 'train/grad_norm_pre_clip': 0.0636172890663147, 'timing/data': 0.00029382000502664596, 'timing/model': 0.3642943309969269, 'learning_rate/qwen_vl_interface': 2.969680445466478e-06, 'learning_rate/action_model': 2.969680445466478e-05, 'epoch': 0.24}) +[2026-10-03 15:15:54,356][starVLA.training.train_starvla][INFO] - Step 3332, Loss: {'train/loss': 0.00015608625835739076, 'train/grad_norm_pre_clip': 0.0673706904053688, 'timing/data': 0.00035414000740274787, 'timing/model': 0.3641682109882822, 'learning_rate/qwen_vl_interface': 2.9670093741661955e-06, 'learning_rate/action_model': 2.9670093741661954e-05, 'epoch': 0.24}) +[2026-10-03 15:15:55,204][starVLA.training.train_starvla][INFO] - Step 3333, Loss: {'train/loss': 0.10748347640037537, 'train/grad_norm_pre_clip': 385.2403869628906, 'timing/data': 0.00025893001293297857, 'timing/model': 0.3644841009954689, 'learning_rate/qwen_vl_interface': 2.964339241317123e-06, 'learning_rate/action_model': 2.964339241317123e-05, 'epoch': 0.24}) +[2026-10-03 15:15:55,851][starVLA.training.train_starvla][INFO] - Step 3334, Loss: {'train/loss': 0.00014107758761383593, 'train/grad_norm_pre_clip': 0.075559601187706, 'timing/data': 0.00023422000231221318, 'timing/model': 0.36429789099202026, 'learning_rate/qwen_vl_interface': 2.961670048016852e-06, 'learning_rate/action_model': 2.9616700480168518e-05, 'epoch': 0.24}) +[2026-10-03 15:15:56,500][starVLA.training.train_starvla][INFO] - Step 3335, Loss: {'train/loss': 0.35235899686813354, 'train/grad_norm_pre_clip': 94.27149200439453, 'timing/data': 0.00037168999551795423, 'timing/model': 0.36463933099003043, 'learning_rate/qwen_vl_interface': 2.9590017953625854e-06, 'learning_rate/action_model': 2.9590017953625853e-05, 'epoch': 0.24}) +[2026-10-03 15:15:57,148][starVLA.training.train_starvla][INFO] - Step 3336, Loss: {'train/loss': 0.0034469221718609333, 'train/grad_norm_pre_clip': 7.626358509063721, 'timing/data': 0.0003293899935670197, 'timing/model': 0.36536937099299394, 'learning_rate/qwen_vl_interface': 2.9563344844511455e-06, 'learning_rate/action_model': 2.9563344844511455e-05, 'epoch': 0.24}) +[2026-10-03 15:15:57,773][starVLA.training.train_starvla][INFO] - Step 3337, Loss: {'train/loss': 5.4254203860182315e-05, 'train/grad_norm_pre_clip': 0.022482726722955704, 'timing/data': 0.00023007999698165804, 'timing/model': 0.36438026100222487, 'learning_rate/qwen_vl_interface': 2.9536681163789568e-06, 'learning_rate/action_model': 2.9536681163789564e-05, 'epoch': 0.24}) +[2026-10-03 15:15:58,420][starVLA.training.train_starvla][INFO] - Step 3338, Loss: {'train/loss': 0.00016519126074854285, 'train/grad_norm_pre_clip': 0.15127749741077423, 'timing/data': 0.00031386000046040863, 'timing/model': 0.36422278100508265, 'learning_rate/qwen_vl_interface': 2.9510026922420653e-06, 'learning_rate/action_model': 2.9510026922420653e-05, 'epoch': 0.24}) +[2026-10-03 15:15:59,262][starVLA.training.train_starvla][INFO] - Step 3339, Loss: {'train/loss': 5.745827729697339e-05, 'train/grad_norm_pre_clip': 0.032504115253686905, 'timing/data': 0.00030049000633880496, 'timing/model': 0.3649134610022884, 'learning_rate/qwen_vl_interface': 2.9483382131361273e-06, 'learning_rate/action_model': 2.948338213136127e-05, 'epoch': 0.24}) +[2026-10-03 15:15:59,910][starVLA.training.train_starvla][INFO] - Step 3340, Loss: {'train/loss': 0.00029059176449663937, 'train/grad_norm_pre_clip': 0.11250325292348862, 'timing/data': 0.00022882000484969467, 'timing/model': 0.3649183010129491, 'learning_rate/qwen_vl_interface': 2.9456746801564072e-06, 'learning_rate/action_model': 2.945674680156407e-05, 'epoch': 0.24}) +[2026-10-03 15:16:00,558][starVLA.training.train_starvla][INFO] - Step 3341, Loss: {'train/loss': 6.374704389600083e-05, 'train/grad_norm_pre_clip': 0.021914688870310783, 'timing/data': 0.0003364099975442514, 'timing/model': 0.3653973209875403, 'learning_rate/qwen_vl_interface': 2.9430120943977854e-06, 'learning_rate/action_model': 2.9430120943977853e-05, 'epoch': 0.24}) +[2026-10-03 15:16:01,204][starVLA.training.train_starvla][INFO] - Step 3342, Loss: {'train/loss': 0.36143121123313904, 'train/grad_norm_pre_clip': 69.05870056152344, 'timing/data': 0.0003073799889534712, 'timing/model': 0.3641763810010161, 'learning_rate/qwen_vl_interface': 2.940350456954748e-06, 'learning_rate/action_model': 2.940350456954748e-05, 'epoch': 0.24}) +[2026-10-03 15:16:01,857][starVLA.training.train_starvla][INFO] - Step 3343, Loss: {'train/loss': 0.00010445850057294592, 'train/grad_norm_pre_clip': 0.04638185352087021, 'timing/data': 0.00037441999302245677, 'timing/model': 0.36681096999382135, 'learning_rate/qwen_vl_interface': 2.9376897689213952e-06, 'learning_rate/action_model': 2.937689768921395e-05, 'epoch': 0.24}) +[2026-10-03 15:16:02,699][starVLA.training.train_starvla][INFO] - Step 3344, Loss: {'train/loss': 0.0003692266473080963, 'train/grad_norm_pre_clip': 0.410662978887558, 'timing/data': 0.00036441000702325255, 'timing/model': 0.3659328700014157, 'learning_rate/qwen_vl_interface': 2.9350300313914364e-06, 'learning_rate/action_model': 2.9350300313914363e-05, 'epoch': 0.24}) +[2026-10-03 15:16:03,348][starVLA.training.train_starvla][INFO] - Step 3345, Loss: {'train/loss': 0.38413628935813904, 'train/grad_norm_pre_clip': 67.57305908203125, 'timing/data': 0.0003199500060873106, 'timing/model': 0.36574350998853333, 'learning_rate/qwen_vl_interface': 2.9323712454581867e-06, 'learning_rate/action_model': 2.9323712454581866e-05, 'epoch': 0.24}) +[2026-10-03 15:16:03,997][starVLA.training.train_starvla][INFO] - Step 3346, Loss: {'train/loss': 0.0007264604791998863, 'train/grad_norm_pre_clip': 0.6786343455314636, 'timing/data': 0.0001970799930859357, 'timing/model': 0.36528700101189315, 'learning_rate/qwen_vl_interface': 2.929713412214572e-06, 'learning_rate/action_model': 2.929713412214572e-05, 'epoch': 0.24}) +[2026-10-03 15:16:04,649][starVLA.training.train_starvla][INFO] - Step 3347, Loss: {'train/loss': 1.24050766316941e-05, 'train/grad_norm_pre_clip': 0.0026887264102697372, 'timing/data': 0.0003239700017729774, 'timing/model': 0.3663752900029067, 'learning_rate/qwen_vl_interface': 2.9270565327531294e-06, 'learning_rate/action_model': 2.927056532753129e-05, 'epoch': 0.24}) +[2026-10-03 15:16:05,299][starVLA.training.train_starvla][INFO] - Step 3348, Loss: {'train/loss': 7.620754331583157e-05, 'train/grad_norm_pre_clip': 0.03326384350657463, 'timing/data': 0.00033131999953184277, 'timing/model': 0.3656177209923044, 'learning_rate/qwen_vl_interface': 2.924400608166003e-06, 'learning_rate/action_model': 2.924400608166003e-05, 'epoch': 0.25}) +[2026-10-03 15:16:06,141][starVLA.training.train_starvla][INFO] - Step 3349, Loss: {'train/loss': 0.00011584487947402522, 'train/grad_norm_pre_clip': 0.0442790724337101, 'timing/data': 0.00032037000346463174, 'timing/model': 0.5578464449936291, 'learning_rate/qwen_vl_interface': 2.921745639544939e-06, 'learning_rate/action_model': 2.921745639544939e-05, 'epoch': 0.25}) +[2026-10-03 15:16:06,790][starVLA.training.train_starvla][INFO] - Step 3350, Loss: {'train/loss': 5.739351036027074e-05, 'train/grad_norm_pre_clip': 0.018204769119620323, 'timing/data': 0.00024151000252459198, 'timing/model': 0.36561648099450395, 'learning_rate/qwen_vl_interface': 2.919091627981297e-06, 'learning_rate/action_model': 2.919091627981297e-05, 'epoch': 0.25}) +[2026-10-03 15:16:07,441][starVLA.training.train_starvla][INFO] - Step 3351, Loss: {'train/loss': 0.000117920120828785, 'train/grad_norm_pre_clip': 0.046277016401290894, 'timing/data': 0.00034641999809537083, 'timing/model': 0.3649164810049115, 'learning_rate/qwen_vl_interface': 2.9164385745660396e-06, 'learning_rate/action_model': 2.9164385745660394e-05, 'epoch': 0.25}) +[2026-10-03 15:16:08,089][starVLA.training.train_starvla][INFO] - Step 3352, Loss: {'train/loss': 0.014714519493281841, 'train/grad_norm_pre_clip': 55.55390548706055, 'timing/data': 0.0002468799939379096, 'timing/model': 0.36490306099585723, 'learning_rate/qwen_vl_interface': 2.9137864803897385e-06, 'learning_rate/action_model': 2.9137864803897387e-05, 'epoch': 0.25}) +[2026-10-03 15:16:08,737][starVLA.training.train_starvla][INFO] - Step 3353, Loss: {'train/loss': 8.57498380355537e-05, 'train/grad_norm_pre_clip': 0.032148219645023346, 'timing/data': 0.00040640999213792384, 'timing/model': 0.3644021609943593, 'learning_rate/qwen_vl_interface': 2.9111353465425696e-06, 'learning_rate/action_model': 2.9111353465425694e-05, 'epoch': 0.25}) +[2026-10-03 15:16:09,577][starVLA.training.train_starvla][INFO] - Step 3354, Loss: {'train/loss': 0.5385960936546326, 'train/grad_norm_pre_clip': 231.53663635253906, 'timing/data': 0.0003123600035905838, 'timing/model': 0.5574017860053573, 'learning_rate/qwen_vl_interface': 2.908485174114312e-06, 'learning_rate/action_model': 2.908485174114312e-05, 'epoch': 0.25}) +[2026-10-03 15:16:10,227][starVLA.training.train_starvla][INFO] - Step 3355, Loss: {'train/loss': 1.0952167940558866e-05, 'train/grad_norm_pre_clip': 0.004498994443565607, 'timing/data': 0.0003004500031238422, 'timing/model': 0.3644300409941934, 'learning_rate/qwen_vl_interface': 2.9058359641943523e-06, 'learning_rate/action_model': 2.9058359641943523e-05, 'epoch': 0.25}) +[2026-10-03 15:16:10,875][starVLA.training.train_starvla][INFO] - Step 3356, Loss: {'train/loss': 2.643364859977737e-05, 'train/grad_norm_pre_clip': 0.008041338063776493, 'timing/data': 0.00024213000142481178, 'timing/model': 0.365151830992545, 'learning_rate/qwen_vl_interface': 2.9031877178716823e-06, 'learning_rate/action_model': 2.903187717871682e-05, 'epoch': 0.25}) +[2026-10-03 15:16:11,523][starVLA.training.train_starvla][INFO] - Step 3357, Loss: {'train/loss': 2.2127169359009713e-05, 'train/grad_norm_pre_clip': 0.004795204382389784, 'timing/data': 0.0002607500064186752, 'timing/model': 0.36540565099858213, 'learning_rate/qwen_vl_interface': 2.900540436234894e-06, 'learning_rate/action_model': 2.9005404362348937e-05, 'epoch': 0.25}) +[2026-10-03 15:16:12,171][starVLA.training.train_starvla][INFO] - Step 3358, Loss: {'train/loss': 7.990456651896238e-05, 'train/grad_norm_pre_clip': 0.02667500078678131, 'timing/data': 0.00029417000769171864, 'timing/model': 0.36447882099309936, 'learning_rate/qwen_vl_interface': 2.897894120372185e-06, 'learning_rate/action_model': 2.897894120372185e-05, 'epoch': 0.25}) +[2026-10-03 15:16:13,021][starVLA.training.train_starvla][INFO] - Step 3359, Loss: {'train/loss': 2.9152575734769925e-05, 'train/grad_norm_pre_clip': 0.0099759791046381, 'timing/data': 0.00025269998877774924, 'timing/model': 0.5645363240037113, 'learning_rate/qwen_vl_interface': 2.8952487713713585e-06, 'learning_rate/action_model': 2.8952487713713582e-05, 'epoch': 0.25}) +[2026-10-03 15:16:13,668][starVLA.training.train_starvla][INFO] - Step 3360, Loss: {'train/loss': 0.047198161482810974, 'train/grad_norm_pre_clip': 136.78431701660156, 'timing/data': 0.0002506599994376302, 'timing/model': 0.3646899609884713, 'learning_rate/qwen_vl_interface': 2.8926043903198135e-06, 'learning_rate/action_model': 2.8926043903198135e-05, 'epoch': 0.25}) +[2026-10-03 15:16:14,316][starVLA.training.train_starvla][INFO] - Step 3361, Loss: {'train/loss': 6.463682802859694e-05, 'train/grad_norm_pre_clip': 0.016239741817116737, 'timing/data': 0.00031750999914947897, 'timing/model': 0.36489515099674463, 'learning_rate/qwen_vl_interface': 2.8899609783045574e-06, 'learning_rate/action_model': 2.8899609783045573e-05, 'epoch': 0.25}) +[2026-10-03 15:16:14,941][starVLA.training.train_starvla][INFO] - Step 3362, Loss: {'train/loss': 0.012772886082530022, 'train/grad_norm_pre_clip': 24.08124351501465, 'timing/data': 0.0003050899977097288, 'timing/model': 0.3421817339985864, 'learning_rate/qwen_vl_interface': 2.8873185364121965e-06, 'learning_rate/action_model': 2.8873185364121963e-05, 'epoch': 0.25}) +[2026-10-03 15:16:15,568][starVLA.training.train_starvla][INFO] - Step 3363, Loss: {'train/loss': 0.38422369956970215, 'train/grad_norm_pre_clip': 41.6580810546875, 'timing/data': 0.00026854001043830067, 'timing/model': 0.3425959730084287, 'learning_rate/qwen_vl_interface': 2.8846770657289386e-06, 'learning_rate/action_model': 2.8846770657289386e-05, 'epoch': 0.25}) +[2026-10-03 15:16:16,216][starVLA.training.train_starvla][INFO] - Step 3364, Loss: {'train/loss': 7.053280569380149e-05, 'train/grad_norm_pre_clip': 0.02607438899576664, 'timing/data': 0.00031180999940261245, 'timing/model': 0.3649335109948879, 'learning_rate/qwen_vl_interface': 2.882036567340593e-06, 'learning_rate/action_model': 2.8820365673405932e-05, 'epoch': 0.25}) +[2026-10-03 15:16:17,028][starVLA.training.train_starvla][INFO] - Step 3365, Loss: {'train/loss': 1.7396452676621266e-05, 'train/grad_norm_pre_clip': 0.0033908504992723465, 'timing/data': 0.00034293999487999827, 'timing/model': 0.3645912810025038, 'learning_rate/qwen_vl_interface': 2.8793970423325684e-06, 'learning_rate/action_model': 2.8793970423325685e-05, 'epoch': 0.25}) +[2026-10-03 15:16:17,674][starVLA.training.train_starvla][INFO] - Step 3366, Loss: {'train/loss': 0.00027256153407506645, 'train/grad_norm_pre_clip': 0.16285675764083862, 'timing/data': 0.00035756999568548054, 'timing/model': 0.3644449910061667, 'learning_rate/qwen_vl_interface': 2.8767584917898756e-06, 'learning_rate/action_model': 2.8767584917898754e-05, 'epoch': 0.25}) +[2026-10-03 15:16:18,325][starVLA.training.train_starvla][INFO] - Step 3367, Loss: {'train/loss': 0.00020265259081497788, 'train/grad_norm_pre_clip': 0.1212787926197052, 'timing/data': 0.00025444000493735075, 'timing/model': 0.36522623999917414, 'learning_rate/qwen_vl_interface': 2.874120916797121e-06, 'learning_rate/action_model': 2.8741209167971207e-05, 'epoch': 0.25}) +[2026-10-03 15:16:18,973][starVLA.training.train_starvla][INFO] - Step 3368, Loss: {'train/loss': 0.0012725789565593004, 'train/grad_norm_pre_clip': 1.5800328254699707, 'timing/data': 0.0003398300032131374, 'timing/model': 0.36527263998868875, 'learning_rate/qwen_vl_interface': 2.8714843184385124e-06, 'learning_rate/action_model': 2.871484318438512e-05, 'epoch': 0.25}) +[2026-10-03 15:16:19,599][starVLA.training.train_starvla][INFO] - Step 3369, Loss: {'train/loss': 1.5742603864055127e-05, 'train/grad_norm_pre_clip': 0.003546599531546235, 'timing/data': 0.0002503199939383194, 'timing/model': 0.3646160909993341, 'learning_rate/qwen_vl_interface': 2.8688486977978575e-06, 'learning_rate/action_model': 2.8688486977978572e-05, 'epoch': 0.25}) +[2026-10-03 15:16:20,449][starVLA.training.train_starvla][INFO] - Step 3370, Loss: {'train/loss': 0.0005447341245599091, 'train/grad_norm_pre_clip': 0.7224396467208862, 'timing/data': 0.000242239999352023, 'timing/model': 0.5675811640103348, 'learning_rate/qwen_vl_interface': 2.8662140559585584e-06, 'learning_rate/action_model': 2.866214055958558e-05, 'epoch': 0.25}) +[2026-10-03 15:16:21,098][starVLA.training.train_starvla][INFO] - Step 3371, Loss: {'train/loss': 0.0001583952980581671, 'train/grad_norm_pre_clip': 0.09225780516862869, 'timing/data': 0.00027500999567564577, 'timing/model': 0.3647753210098017, 'learning_rate/qwen_vl_interface': 2.8635803940036162e-06, 'learning_rate/action_model': 2.863580394003616e-05, 'epoch': 0.25}) +[2026-10-03 15:16:21,747][starVLA.training.train_starvla][INFO] - Step 3372, Loss: {'train/loss': 0.00027484679594635963, 'train/grad_norm_pre_clip': 0.0748908594250679, 'timing/data': 0.0002584999892860651, 'timing/model': 0.36543931100459304, 'learning_rate/qwen_vl_interface': 2.8609477130156317e-06, 'learning_rate/action_model': 2.8609477130156314e-05, 'epoch': 0.25}) +[2026-10-03 15:16:22,394][starVLA.training.train_starvla][INFO] - Step 3373, Loss: {'train/loss': 0.00013449008110910654, 'train/grad_norm_pre_clip': 0.08357483148574829, 'timing/data': 0.00019926000095438212, 'timing/model': 0.36442733100557234, 'learning_rate/qwen_vl_interface': 2.8583160140767997e-06, 'learning_rate/action_model': 2.8583160140767994e-05, 'epoch': 0.25}) +[2026-10-03 15:16:23,041][starVLA.training.train_starvla][INFO] - Step 3374, Loss: {'train/loss': 0.00022010551765561104, 'train/grad_norm_pre_clip': 0.04348534718155861, 'timing/data': 0.00020231000962667167, 'timing/model': 0.3646008810028434, 'learning_rate/qwen_vl_interface': 2.8556852982689125e-06, 'learning_rate/action_model': 2.8556852982689126e-05, 'epoch': 0.25}) +[2026-10-03 15:16:23,690][starVLA.training.train_starvla][INFO] - Step 3375, Loss: {'train/loss': 0.4073822498321533, 'train/grad_norm_pre_clip': 71.78343200683594, 'timing/data': 0.0003301700053270906, 'timing/model': 0.3641524909908185, 'learning_rate/qwen_vl_interface': 2.8530555666733584e-06, 'learning_rate/action_model': 2.8530555666733583e-05, 'epoch': 0.25}) +[2026-10-03 15:16:24,528][starVLA.training.train_starvla][INFO] - Step 3376, Loss: {'train/loss': 0.00014336107415147126, 'train/grad_norm_pre_clip': 0.08270052820444107, 'timing/data': 0.00025927000388037413, 'timing/model': 0.3635488909931155, 'learning_rate/qwen_vl_interface': 2.85042682037112e-06, 'learning_rate/action_model': 2.85042682037112e-05, 'epoch': 0.25}) +[2026-10-03 15:16:25,174][starVLA.training.train_starvla][INFO] - Step 3377, Loss: {'train/loss': 0.000252259720582515, 'train/grad_norm_pre_clip': 0.05388355255126953, 'timing/data': 0.0002623000036692247, 'timing/model': 0.3640708909952082, 'learning_rate/qwen_vl_interface': 2.8477990604427764e-06, 'learning_rate/action_model': 2.8477990604427763e-05, 'epoch': 0.25}) +[2026-10-03 15:16:25,820][starVLA.training.train_starvla][INFO] - Step 3378, Loss: {'train/loss': 0.0006488264771178365, 'train/grad_norm_pre_clip': 0.8630439043045044, 'timing/data': 0.00031850000959821045, 'timing/model': 0.36423333099810407, 'learning_rate/qwen_vl_interface': 2.8451722879684994e-06, 'learning_rate/action_model': 2.8451722879684994e-05, 'epoch': 0.25}) +[2026-10-03 15:16:26,469][starVLA.training.train_starvla][INFO] - Step 3379, Loss: {'train/loss': 0.0001626086886972189, 'train/grad_norm_pre_clip': 0.030654560774564743, 'timing/data': 0.00020463998953346163, 'timing/model': 0.36473826100700535, 'learning_rate/qwen_vl_interface': 2.842546504028059e-06, 'learning_rate/action_model': 2.8425465040280586e-05, 'epoch': 0.25}) +[2026-10-03 15:16:27,116][starVLA.training.train_starvla][INFO] - Step 3380, Loss: {'train/loss': 0.0008666070061735809, 'train/grad_norm_pre_clip': 0.5138987302780151, 'timing/data': 0.00024228999973274767, 'timing/model': 0.3643822110025212, 'learning_rate/qwen_vl_interface': 2.83992170970081e-06, 'learning_rate/action_model': 2.8399217097008102e-05, 'epoch': 0.25}) +[2026-10-03 15:16:27,959][starVLA.training.train_starvla][INFO] - Step 3381, Loss: {'train/loss': 1.6614361811662093e-05, 'train/grad_norm_pre_clip': 0.004927241709083319, 'timing/data': 0.0001961899979505688, 'timing/model': 0.5609126249910332, 'learning_rate/qwen_vl_interface': 2.837297906065711e-06, 'learning_rate/action_model': 2.8372979060657107e-05, 'epoch': 0.25}) +[2026-10-03 15:16:28,605][starVLA.training.train_starvla][INFO] - Step 3382, Loss: {'train/loss': 0.0013124521356076002, 'train/grad_norm_pre_clip': 11.596894264221191, 'timing/data': 0.00034111000422853976, 'timing/model': 0.3640180109941866, 'learning_rate/qwen_vl_interface': 2.834675094201306e-06, 'learning_rate/action_model': 2.8346750942013063e-05, 'epoch': 0.25}) +[2026-10-03 15:16:29,253][starVLA.training.train_starvla][INFO] - Step 3383, Loss: {'train/loss': 0.00017820454377215356, 'train/grad_norm_pre_clip': 0.11054686456918716, 'timing/data': 0.00043652000022120774, 'timing/model': 0.36410436099686194, 'learning_rate/qwen_vl_interface': 2.8320532751857365e-06, 'learning_rate/action_model': 2.8320532751857366e-05, 'epoch': 0.25}) +[2026-10-03 15:16:29,900][starVLA.training.train_starvla][INFO] - Step 3384, Loss: {'train/loss': 8.108615293167531e-05, 'train/grad_norm_pre_clip': 0.03211097791790962, 'timing/data': 0.0002552099904278293, 'timing/model': 0.36374358099419624, 'learning_rate/qwen_vl_interface': 2.829432450096731e-06, 'learning_rate/action_model': 2.8294324500967306e-05, 'epoch': 0.25}) +[2026-10-03 15:16:30,546][starVLA.training.train_starvla][INFO] - Step 3385, Loss: {'train/loss': 2.8137794288340956e-05, 'train/grad_norm_pre_clip': 0.009997133165597916, 'timing/data': 0.0003221700026188046, 'timing/model': 0.36410552100278437, 'learning_rate/qwen_vl_interface': 2.8268126200116118e-06, 'learning_rate/action_model': 2.8268126200116117e-05, 'epoch': 0.25}) +[2026-10-03 15:16:31,381][starVLA.training.train_starvla][INFO] - Step 3386, Loss: {'train/loss': 0.00018406668095849454, 'train/grad_norm_pre_clip': 0.08339421451091766, 'timing/data': 0.00032406000536866486, 'timing/model': 0.5531593959894963, 'learning_rate/qwen_vl_interface': 2.824193786007294e-06, 'learning_rate/action_model': 2.8241937860072937e-05, 'epoch': 0.25}) +[2026-10-03 15:16:32,036][starVLA.training.train_starvla][INFO] - Step 3387, Loss: {'train/loss': 0.03578097000718117, 'train/grad_norm_pre_clip': 104.54412078857422, 'timing/data': 0.00036939000710844994, 'timing/model': 0.36425507100648247, 'learning_rate/qwen_vl_interface': 2.821575949160279e-06, 'learning_rate/action_model': 2.8215759491602787e-05, 'epoch': 0.25}) +[2026-10-03 15:16:32,683][starVLA.training.train_starvla][INFO] - Step 3388, Loss: {'train/loss': 1.6129768482642248e-05, 'train/grad_norm_pre_clip': 0.007196641527116299, 'timing/data': 0.00019756000256165862, 'timing/model': 0.3650726810010383, 'learning_rate/qwen_vl_interface': 2.8189591105466625e-06, 'learning_rate/action_model': 2.8189591105466624e-05, 'epoch': 0.25}) +[2026-10-03 15:16:33,330][starVLA.training.train_starvla][INFO] - Step 3389, Loss: {'train/loss': 1.1563006410142407e-05, 'train/grad_norm_pre_clip': 0.0023582973517477512, 'timing/data': 0.0003307699953438714, 'timing/model': 0.36438923099194653, 'learning_rate/qwen_vl_interface': 2.816343271242126e-06, 'learning_rate/action_model': 2.816343271242126e-05, 'epoch': 0.25}) +[2026-10-03 15:16:33,977][starVLA.training.train_starvla][INFO] - Step 3390, Loss: {'train/loss': 6.93664769642055e-05, 'train/grad_norm_pre_clip': 0.03442349284887314, 'timing/data': 0.000328909998643212, 'timing/model': 0.3643404709873721, 'learning_rate/qwen_vl_interface': 2.8137284323219467e-06, 'learning_rate/action_model': 2.8137284323219464e-05, 'epoch': 0.25}) +[2026-10-03 15:16:34,626][starVLA.training.train_starvla][INFO] - Step 3391, Loss: {'train/loss': 0.0002559996210038662, 'train/grad_norm_pre_clip': 0.13914351165294647, 'timing/data': 0.0003901399904862046, 'timing/model': 0.36475106098805554, 'learning_rate/qwen_vl_interface': 2.81111459486098e-06, 'learning_rate/action_model': 2.81111459486098e-05, 'epoch': 0.25}) +[2026-10-03 15:16:35,473][starVLA.training.train_starvla][INFO] - Step 3392, Loss: {'train/loss': 0.007135808002203703, 'train/grad_norm_pre_clip': 14.05235481262207, 'timing/data': 0.0003513300034683198, 'timing/model': 0.3646784409938846, 'learning_rate/qwen_vl_interface': 2.8085017599336806e-06, 'learning_rate/action_model': 2.8085017599336806e-05, 'epoch': 0.25}) +[2026-10-03 15:16:36,120][starVLA.training.train_starvla][INFO] - Step 3393, Loss: {'train/loss': 6.782275158911943e-05, 'train/grad_norm_pre_clip': 0.02020779438316822, 'timing/data': 0.000208609999390319, 'timing/model': 0.3641348810051568, 'learning_rate/qwen_vl_interface': 2.8058899286140828e-06, 'learning_rate/action_model': 2.805889928614083e-05, 'epoch': 0.25}) +[2026-10-03 15:16:36,768][starVLA.training.train_starvla][INFO] - Step 3394, Loss: {'train/loss': 0.2044217884540558, 'train/grad_norm_pre_clip': 125.75237274169922, 'timing/data': 0.00020342000061646104, 'timing/model': 0.3646603410015814, 'learning_rate/qwen_vl_interface': 2.8032791019758127e-06, 'learning_rate/action_model': 2.8032791019758127e-05, 'epoch': 0.25}) +[2026-10-03 15:16:37,418][starVLA.training.train_starvla][INFO] - Step 3395, Loss: {'train/loss': 2.3252352548297495e-05, 'train/grad_norm_pre_clip': 0.005344139877706766, 'timing/data': 0.00020543999562505633, 'timing/model': 0.3651518410042627, 'learning_rate/qwen_vl_interface': 2.800669281092084e-06, 'learning_rate/action_model': 2.8006692810920838e-05, 'epoch': 0.25}) +[2026-10-03 15:16:38,064][starVLA.training.train_starvla][INFO] - Step 3396, Loss: {'train/loss': 1.536993841000367e-05, 'train/grad_norm_pre_clip': 0.005421170964837074, 'timing/data': 0.0002393100003246218, 'timing/model': 0.36453837099543307, 'learning_rate/qwen_vl_interface': 2.7980604670356938e-06, 'learning_rate/action_model': 2.7980604670356937e-05, 'epoch': 0.25}) +[2026-10-03 15:16:38,888][starVLA.training.train_starvla][INFO] - Step 3397, Loss: {'train/loss': 0.0008304023649543524, 'train/grad_norm_pre_clip': 1.0959744453430176, 'timing/data': 0.000309199996991083, 'timing/model': 0.3418439439992653, 'learning_rate/qwen_vl_interface': 2.795452660879027e-06, 'learning_rate/action_model': 2.7954526608790273e-05, 'epoch': 0.25}) +[2026-10-03 15:16:39,535][starVLA.training.train_starvla][INFO] - Step 3398, Loss: {'train/loss': 2.4019025659072213e-05, 'train/grad_norm_pre_clip': 0.021616756916046143, 'timing/data': 0.00032699000439606607, 'timing/model': 0.36429667100310326, 'learning_rate/qwen_vl_interface': 2.7928458636940565e-06, 'learning_rate/action_model': 2.7928458636940564e-05, 'epoch': 0.25}) +[2026-10-03 15:16:40,184][starVLA.training.train_starvla][INFO] - Step 3399, Loss: {'train/loss': 1.1794048077717889e-05, 'train/grad_norm_pre_clip': 0.0037885089404881, 'timing/data': 0.000301949999993667, 'timing/model': 0.36468340099963825, 'learning_rate/qwen_vl_interface': 2.790240076552336e-06, 'learning_rate/action_model': 2.7902400765523355e-05, 'epoch': 0.25}) +[2026-10-03 15:16:40,831][starVLA.training.train_starvla][INFO] - Step 3400, Loss: {'train/loss': 0.006577508058398962, 'train/grad_norm_pre_clip': 16.280170440673828, 'timing/data': 0.00034936000884044915, 'timing/model': 0.3644206510070944, 'learning_rate/qwen_vl_interface': 2.787635300525006e-06, 'learning_rate/action_model': 2.787635300525006e-05, 'epoch': 0.25}) +[2026-10-03 15:16:41,478][starVLA.training.train_starvla][INFO] - Step 3401, Loss: {'train/loss': 0.00015977639122866094, 'train/grad_norm_pre_clip': 0.09744216501712799, 'timing/data': 0.0002482900017639622, 'timing/model': 0.36457534100918565, 'learning_rate/qwen_vl_interface': 2.7850315366827933e-06, 'learning_rate/action_model': 2.7850315366827934e-05, 'epoch': 0.25}) +[2026-10-03 15:16:42,317][starVLA.training.train_starvla][INFO] - Step 3402, Loss: {'train/loss': 4.723635356640443e-06, 'train/grad_norm_pre_clip': 0.001469733309932053, 'timing/data': 0.00022429999080486596, 'timing/model': 0.5563633650017437, 'learning_rate/qwen_vl_interface': 2.7824287860960076e-06, 'learning_rate/action_model': 2.7824287860960075e-05, 'epoch': 0.25}) +[2026-10-03 15:16:42,944][starVLA.training.train_starvla][INFO] - Step 3403, Loss: {'train/loss': 7.890056622272823e-06, 'train/grad_norm_pre_clip': 0.007109607569873333, 'timing/data': 0.0002864700072677806, 'timing/model': 0.3424729150137864, 'learning_rate/qwen_vl_interface': 2.7798270498345377e-06, 'learning_rate/action_model': 2.7798270498345376e-05, 'epoch': 0.25}) +[2026-10-03 15:16:43,591][starVLA.training.train_starvla][INFO] - Step 3404, Loss: {'train/loss': 6.094493983255234e-06, 'train/grad_norm_pre_clip': 0.00159072270616889, 'timing/data': 0.00032959999225568026, 'timing/model': 0.3645413810008904, 'learning_rate/qwen_vl_interface': 2.777226328967861e-06, 'learning_rate/action_model': 2.7772263289678607e-05, 'epoch': 0.25}) +[2026-10-03 15:16:44,239][starVLA.training.train_starvla][INFO] - Step 3405, Loss: {'train/loss': 2.919644975918345e-05, 'train/grad_norm_pre_clip': 0.011217176914215088, 'timing/data': 0.0002227400109404698, 'timing/model': 0.3649492010008544, 'learning_rate/qwen_vl_interface': 2.774626624565035e-06, 'learning_rate/action_model': 2.774626624565035e-05, 'epoch': 0.25}) +[2026-10-03 15:16:44,865][starVLA.training.train_starvla][INFO] - Step 3406, Loss: {'train/loss': 7.247945904964581e-05, 'train/grad_norm_pre_clip': 0.04294680431485176, 'timing/data': 0.0002557599946158007, 'timing/model': 0.3422414450033102, 'learning_rate/qwen_vl_interface': 2.772027937694701e-06, 'learning_rate/action_model': 2.772027937694701e-05, 'epoch': 0.25}) +[2026-10-03 15:16:45,721][starVLA.training.train_starvla][INFO] - Step 3407, Loss: {'train/loss': 8.589238859713078e-05, 'train/grad_norm_pre_clip': 0.09401968121528625, 'timing/data': 0.00021613998978864402, 'timing/model': 0.5720694939955138, 'learning_rate/qwen_vl_interface': 2.7694302694250806e-06, 'learning_rate/action_model': 2.7694302694250808e-05, 'epoch': 0.25}) +[2026-10-03 15:16:46,346][starVLA.training.train_starvla][INFO] - Step 3408, Loss: {'train/loss': 0.0003501975443214178, 'train/grad_norm_pre_clip': 0.292519748210907, 'timing/data': 0.0003025900077773258, 'timing/model': 0.3648770909931045, 'learning_rate/qwen_vl_interface': 2.7668336208239775e-06, 'learning_rate/action_model': 2.7668336208239776e-05, 'epoch': 0.25}) +[2026-10-03 15:16:46,994][starVLA.training.train_starvla][INFO] - Step 3409, Loss: {'train/loss': 2.0404491806402802e-05, 'train/grad_norm_pre_clip': 0.011486109346151352, 'timing/data': 0.00025132000155281276, 'timing/model': 0.3648917709942907, 'learning_rate/qwen_vl_interface': 2.7642379929587755e-06, 'learning_rate/action_model': 2.7642379929587754e-05, 'epoch': 0.25}) +[2026-10-03 15:16:47,641][starVLA.training.train_starvla][INFO] - Step 3410, Loss: {'train/loss': 4.194660959910834e-06, 'train/grad_norm_pre_clip': 0.0008444903651252389, 'timing/data': 0.00024095000117085874, 'timing/model': 0.36442826100392267, 'learning_rate/qwen_vl_interface': 2.7616433868964387e-06, 'learning_rate/action_model': 2.7616433868964388e-05, 'epoch': 0.25}) +[2026-10-03 15:16:48,291][starVLA.training.train_starvla][INFO] - Step 3411, Loss: {'train/loss': 3.29851136484649e-05, 'train/grad_norm_pre_clip': 0.009703362360596657, 'timing/data': 0.0003210499999113381, 'timing/model': 0.364472981003928, 'learning_rate/qwen_vl_interface': 2.7590498037035122e-06, 'learning_rate/action_model': 2.7590498037035124e-05, 'epoch': 0.25}) +[2026-10-03 15:16:49,145][starVLA.training.train_starvla][INFO] - Step 3412, Loss: {'train/loss': 4.1177572711603716e-05, 'train/grad_norm_pre_clip': 0.022978084161877632, 'timing/data': 0.00046847001067362726, 'timing/model': 0.5708771530044032, 'learning_rate/qwen_vl_interface': 2.7564572444461218e-06, 'learning_rate/action_model': 2.7564572444461217e-05, 'epoch': 0.25}) +[2026-10-03 15:16:49,791][starVLA.training.train_starvla][INFO] - Step 3413, Loss: {'train/loss': 2.4959333586593857e-06, 'train/grad_norm_pre_clip': 0.0004948278074152768, 'timing/data': 0.00019454999710433185, 'timing/model': 0.36437238099460956, 'learning_rate/qwen_vl_interface': 2.7538657101899674e-06, 'learning_rate/action_model': 2.7538657101899673e-05, 'epoch': 0.25}) +[2026-10-03 15:16:50,438][starVLA.training.train_starvla][INFO] - Step 3414, Loss: {'train/loss': 0.45635151863098145, 'train/grad_norm_pre_clip': 245.1783447265625, 'timing/data': 0.00021326998830772936, 'timing/model': 0.3639928710035747, 'learning_rate/qwen_vl_interface': 2.751275202000332e-06, 'learning_rate/action_model': 2.7512752020003317e-05, 'epoch': 0.25}) +[2026-10-03 15:16:51,088][starVLA.training.train_starvla][INFO] - Step 3415, Loss: {'train/loss': 2.908118949562777e-05, 'train/grad_norm_pre_clip': 0.020466359332203865, 'timing/data': 0.0003827299951808527, 'timing/model': 0.36531687100068666, 'learning_rate/qwen_vl_interface': 2.7486857209420763e-06, 'learning_rate/action_model': 2.7486857209420763e-05, 'epoch': 0.25}) +[2026-10-03 15:16:51,738][starVLA.training.train_starvla][INFO] - Step 3416, Loss: {'train/loss': 4.858677857555449e-05, 'train/grad_norm_pre_clip': 0.028254330158233643, 'timing/data': 0.0003242499951738864, 'timing/model': 0.3655177309992723, 'learning_rate/qwen_vl_interface': 2.7460972680796377e-06, 'learning_rate/action_model': 2.7460972680796375e-05, 'epoch': 0.25}) +[2026-10-03 15:16:52,386][starVLA.training.train_starvla][INFO] - Step 3417, Loss: {'train/loss': 0.0029387641698122025, 'train/grad_norm_pre_clip': 7.998147964477539, 'timing/data': 0.00019659000099636614, 'timing/model': 0.36513870100316126, 'learning_rate/qwen_vl_interface': 2.7435098444770314e-06, 'learning_rate/action_model': 2.7435098444770315e-05, 'epoch': 0.25}) +[2026-10-03 15:16:53,235][starVLA.training.train_starvla][INFO] - Step 3418, Loss: {'train/loss': 4.9434132961323485e-05, 'train/grad_norm_pre_clip': 0.015673095360398293, 'timing/data': 0.00025561999063938856, 'timing/model': 0.3646820809954079, 'learning_rate/qwen_vl_interface': 2.7409234511978495e-06, 'learning_rate/action_model': 2.7409234511978492e-05, 'epoch': 0.25}) +[2026-10-03 15:16:53,885][starVLA.training.train_starvla][INFO] - Step 3419, Loss: {'train/loss': 0.028597163036465645, 'train/grad_norm_pre_clip': 30.703248977661133, 'timing/data': 0.00020619999850168824, 'timing/model': 0.36542872099380475, 'learning_rate/qwen_vl_interface': 2.7383380893052606e-06, 'learning_rate/action_model': 2.7383380893052608e-05, 'epoch': 0.25}) +[2026-10-03 15:16:54,533][starVLA.training.train_starvla][INFO] - Step 3420, Loss: {'train/loss': 8.195632972274325e-07, 'train/grad_norm_pre_clip': 0.00023717987642157823, 'timing/data': 0.00035601000126916915, 'timing/model': 0.3652334609942045, 'learning_rate/qwen_vl_interface': 2.73575375986201e-06, 'learning_rate/action_model': 2.7357537598620096e-05, 'epoch': 0.25}) +[2026-10-03 15:16:55,182][starVLA.training.train_starvla][INFO] - Step 3421, Loss: {'train/loss': 1.6681318811606616e-05, 'train/grad_norm_pre_clip': 0.007591716013848782, 'timing/data': 0.00020030001178383827, 'timing/model': 0.3650262510054745, 'learning_rate/qwen_vl_interface': 2.733170463930417e-06, 'learning_rate/action_model': 2.7331704639304167e-05, 'epoch': 0.25}) +[2026-10-03 15:16:55,829][starVLA.training.train_starvla][INFO] - Step 3422, Loss: {'train/loss': 5.235680509940721e-05, 'train/grad_norm_pre_clip': 0.038469329476356506, 'timing/data': 0.0002484199940226972, 'timing/model': 0.36455198099429253, 'learning_rate/qwen_vl_interface': 2.7305882025723774e-06, 'learning_rate/action_model': 2.7305882025723774e-05, 'epoch': 0.25}) +[2026-10-03 15:16:56,682][starVLA.training.train_starvla][INFO] - Step 3423, Loss: {'train/loss': 3.3974149573623436e-06, 'train/grad_norm_pre_clip': 0.0015851258067414165, 'timing/data': 0.0002187999925808981, 'timing/model': 0.3654126309993444, 'learning_rate/qwen_vl_interface': 2.7280069768493626e-06, 'learning_rate/action_model': 2.7280069768493628e-05, 'epoch': 0.25}) +[2026-10-03 15:16:57,330][starVLA.training.train_starvla][INFO] - Step 3424, Loss: {'train/loss': 1.8498421923140995e-05, 'train/grad_norm_pre_clip': 0.0077287619933485985, 'timing/data': 0.00019428000086918473, 'timing/model': 0.364554630999919, 'learning_rate/qwen_vl_interface': 2.7254267878224144e-06, 'learning_rate/action_model': 2.7254267878224143e-05, 'epoch': 0.25}) +[2026-10-03 15:16:57,976][starVLA.training.train_starvla][INFO] - Step 3425, Loss: {'train/loss': 1.4677618764835643e-06, 'train/grad_norm_pre_clip': 0.00044458857155404985, 'timing/data': 0.00019926000095438212, 'timing/model': 0.3637480910110753, 'learning_rate/qwen_vl_interface': 2.7228476365521524e-06, 'learning_rate/action_model': 2.722847636552152e-05, 'epoch': 0.25}) +[2026-10-03 15:16:58,624][starVLA.training.train_starvla][INFO] - Step 3426, Loss: {'train/loss': 1.2173849427199457e-05, 'train/grad_norm_pre_clip': 0.0051154037937521935, 'timing/data': 0.00020018999930471182, 'timing/model': 0.36494471099285875, 'learning_rate/qwen_vl_interface': 2.7202695240987674e-06, 'learning_rate/action_model': 2.7202695240987676e-05, 'epoch': 0.25}) +[2026-10-03 15:16:59,274][starVLA.training.train_starvla][INFO] - Step 3427, Loss: {'train/loss': 0.0006715107592754066, 'train/grad_norm_pre_clip': 0.8630640506744385, 'timing/data': 0.00029556000663433224, 'timing/model': 0.36523285100702196, 'learning_rate/qwen_vl_interface': 2.717692451522025e-06, 'learning_rate/action_model': 2.7176924515220248e-05, 'epoch': 0.25}) +[2026-10-03 15:17:00,115][starVLA.training.train_starvla][INFO] - Step 3428, Loss: {'train/loss': 0.08400953561067581, 'train/grad_norm_pre_clip': 479.0429382324219, 'timing/data': 0.0002461300027789548, 'timing/model': 0.5575587850034935, 'learning_rate/qwen_vl_interface': 2.7151164198812624e-06, 'learning_rate/action_model': 2.7151164198812622e-05, 'epoch': 0.25}) +[2026-10-03 15:17:00,763][starVLA.training.train_starvla][INFO] - Step 3429, Loss: {'train/loss': 6.829050835222006e-05, 'train/grad_norm_pre_clip': 0.05439135804772377, 'timing/data': 0.0002835299965227023, 'timing/model': 0.365080781004508, 'learning_rate/qwen_vl_interface': 2.712541430235388e-06, 'learning_rate/action_model': 2.712541430235388e-05, 'epoch': 0.25}) +[2026-10-03 15:17:01,411][starVLA.training.train_starvla][INFO] - Step 3430, Loss: {'train/loss': 1.0803328223119024e-06, 'train/grad_norm_pre_clip': 0.00041313140536658466, 'timing/data': 0.00028203999681863934, 'timing/model': 0.36482181199244224, 'learning_rate/qwen_vl_interface': 2.7099674836428842e-06, 'learning_rate/action_model': 2.709967483642884e-05, 'epoch': 0.25}) +[2026-10-03 15:17:02,061][starVLA.training.train_starvla][INFO] - Step 3431, Loss: {'train/loss': 0.00016754865646362305, 'train/grad_norm_pre_clip': 0.1871079057455063, 'timing/data': 0.00023051000607665628, 'timing/model': 0.36603701100102626, 'learning_rate/qwen_vl_interface': 2.7073945811618018e-06, 'learning_rate/action_model': 2.7073945811618017e-05, 'epoch': 0.25}) +[2026-10-03 15:17:02,710][starVLA.training.train_starvla][INFO] - Step 3432, Loss: {'train/loss': 0.31004586815834045, 'train/grad_norm_pre_clip': 160.23672485351562, 'timing/data': 0.00029938000079710037, 'timing/model': 0.3651037709933007, 'learning_rate/qwen_vl_interface': 2.704822723849765e-06, 'learning_rate/action_model': 2.7048227238497648e-05, 'epoch': 0.25}) +[2026-10-03 15:17:03,358][starVLA.training.train_starvla][INFO] - Step 3433, Loss: {'train/loss': 3.7347213947214186e-05, 'train/grad_norm_pre_clip': 0.030700819566845894, 'timing/data': 0.0002784500102279708, 'timing/model': 0.3647932109888643, 'learning_rate/qwen_vl_interface': 2.7022519127639687e-06, 'learning_rate/action_model': 2.7022519127639682e-05, 'epoch': 0.25}) +[2026-10-03 15:17:04,202][starVLA.training.train_starvla][INFO] - Step 3434, Loss: {'train/loss': 0.11908165365457535, 'train/grad_norm_pre_clip': 205.0050811767578, 'timing/data': 0.0002544599992688745, 'timing/model': 0.36440394099918194, 'learning_rate/qwen_vl_interface': 2.6996821489611713e-06, 'learning_rate/action_model': 2.699682148961171e-05, 'epoch': 0.25}) +[2026-10-03 15:17:04,852][starVLA.training.train_starvla][INFO] - Step 3435, Loss: {'train/loss': 5.9574573242571205e-05, 'train/grad_norm_pre_clip': 0.042153868824243546, 'timing/data': 0.00028962000214960426, 'timing/model': 0.36531849099264946, 'learning_rate/qwen_vl_interface': 2.697113433497712e-06, 'learning_rate/action_model': 2.6971134334977115e-05, 'epoch': 0.25}) +[2026-10-03 15:17:05,500][starVLA.training.train_starvla][INFO] - Step 3436, Loss: {'train/loss': 3.851902874885127e-06, 'train/grad_norm_pre_clip': 0.0016269658226519823, 'timing/data': 0.000242110007093288, 'timing/model': 0.36536359100136906, 'learning_rate/qwen_vl_interface': 2.6945457674294863e-06, 'learning_rate/action_model': 2.694545767429486e-05, 'epoch': 0.25}) +[2026-10-03 15:17:06,128][starVLA.training.train_starvla][INFO] - Step 3437, Loss: {'train/loss': 0.000733687833417207, 'train/grad_norm_pre_clip': 2.334897994995117, 'timing/data': 0.000258400003076531, 'timing/model': 0.34288701400510035, 'learning_rate/qwen_vl_interface': 2.6919791518119676e-06, 'learning_rate/action_model': 2.6919791518119674e-05, 'epoch': 0.25}) +[2026-10-03 15:17:06,776][starVLA.training.train_starvla][INFO] - Step 3438, Loss: {'train/loss': 0.9681117534637451, 'train/grad_norm_pre_clip': 115.09686279296875, 'timing/data': 0.00032864899549167603, 'timing/model': 0.36513613199349493, 'learning_rate/qwen_vl_interface': 2.6894135877001946e-06, 'learning_rate/action_model': 2.689413587700195e-05, 'epoch': 0.25}) +[2026-10-03 15:17:07,619][starVLA.training.train_starvla][INFO] - Step 3439, Loss: {'train/loss': 0.0004866516392212361, 'train/grad_norm_pre_clip': 0.8923476338386536, 'timing/data': 0.0003145500086247921, 'timing/model': 0.3653008109977236, 'learning_rate/qwen_vl_interface': 2.686849076148775e-06, 'learning_rate/action_model': 2.6868490761487747e-05, 'epoch': 0.25}) +[2026-10-03 15:17:08,244][starVLA.training.train_starvla][INFO] - Step 3440, Loss: {'train/loss': 1.6761981896706857e-05, 'train/grad_norm_pre_clip': 0.016386423259973526, 'timing/data': 0.00034317000245209783, 'timing/model': 0.36454742099158466, 'learning_rate/qwen_vl_interface': 2.68428561821188e-06, 'learning_rate/action_model': 2.6842856182118802e-05, 'epoch': 0.25}) +[2026-10-03 15:17:08,892][starVLA.training.train_starvla][INFO] - Step 3441, Loss: {'train/loss': 0.025563830509781837, 'train/grad_norm_pre_clip': 111.24837493896484, 'timing/data': 0.0003003699966939166, 'timing/model': 0.36451229099475313, 'learning_rate/qwen_vl_interface': 2.681723214943253e-06, 'learning_rate/action_model': 2.681723214943253e-05, 'epoch': 0.25}) +[2026-10-03 15:17:09,538][starVLA.training.train_starvla][INFO] - Step 3442, Loss: {'train/loss': 7.510071100114146e-06, 'train/grad_norm_pre_clip': 0.003409615019336343, 'timing/data': 0.00031613999453838915, 'timing/model': 0.3641840409982251, 'learning_rate/qwen_vl_interface': 2.6791618673961996e-06, 'learning_rate/action_model': 2.6791618673961994e-05, 'epoch': 0.25}) +[2026-10-03 15:17:10,188][starVLA.training.train_starvla][INFO] - Step 3443, Loss: {'train/loss': 7.726093826931901e-06, 'train/grad_norm_pre_clip': 0.0036224471405148506, 'timing/data': 0.00034295998921152204, 'timing/model': 0.36488724101218395, 'learning_rate/qwen_vl_interface': 2.676601576623594e-06, 'learning_rate/action_model': 2.6766015766235937e-05, 'epoch': 0.25}) +[2026-10-03 15:17:11,040][starVLA.training.train_starvla][INFO] - Step 3444, Loss: {'train/loss': 0.6279040575027466, 'train/grad_norm_pre_clip': 41.02658462524414, 'timing/data': 0.00046813000517431647, 'timing/model': 0.36435360099130776, 'learning_rate/qwen_vl_interface': 2.674042343677874e-06, 'learning_rate/action_model': 2.6740423436778738e-05, 'epoch': 0.25}) +[2026-10-03 15:17:11,664][starVLA.training.train_starvla][INFO] - Step 3445, Loss: {'train/loss': 0.47039270401000977, 'train/grad_norm_pre_clip': 18.18934440612793, 'timing/data': 0.0002488300087861717, 'timing/model': 0.34214635399985127, 'learning_rate/qwen_vl_interface': 2.6714841696110456e-06, 'learning_rate/action_model': 2.6714841696110456e-05, 'epoch': 0.25}) +[2026-10-03 15:17:12,311][starVLA.training.train_starvla][INFO] - Step 3446, Loss: {'train/loss': 6.273305643844651e-06, 'train/grad_norm_pre_clip': 0.0020914191845804453, 'timing/data': 0.0002739199990173802, 'timing/model': 0.3647575009963475, 'learning_rate/qwen_vl_interface': 2.6689270554746743e-06, 'learning_rate/action_model': 2.6689270554746743e-05, 'epoch': 0.25}) +[2026-10-03 15:17:12,960][starVLA.training.train_starvla][INFO] - Step 3447, Loss: {'train/loss': 1.1473090125946328e-05, 'train/grad_norm_pre_clip': 0.01274034846574068, 'timing/data': 0.00026697000430431217, 'timing/model': 0.3642312409938313, 'learning_rate/qwen_vl_interface': 2.6663710023198956e-06, 'learning_rate/action_model': 2.6663710023198956e-05, 'epoch': 0.25}) +[2026-10-03 15:17:13,609][starVLA.training.train_starvla][INFO] - Step 3448, Loss: {'train/loss': 3.4719555515039247e-06, 'train/grad_norm_pre_clip': 0.001042813528329134, 'timing/data': 0.0002033199998550117, 'timing/model': 0.36508331100048963, 'learning_rate/qwen_vl_interface': 2.663816011197405e-06, 'learning_rate/action_model': 2.6638160111974048e-05, 'epoch': 0.25}) +[2026-10-03 15:17:14,457][starVLA.training.train_starvla][INFO] - Step 3449, Loss: {'train/loss': 0.0013206710573285818, 'train/grad_norm_pre_clip': 1.8890058994293213, 'timing/data': 0.0002081400016322732, 'timing/model': 0.56550730401068, 'learning_rate/qwen_vl_interface': 2.661262083157463e-06, 'learning_rate/action_model': 2.661262083157463e-05, 'epoch': 0.25}) +[2026-10-03 15:17:15,104][starVLA.training.train_starvla][INFO] - Step 3450, Loss: {'train/loss': 2.3543675524706487e-06, 'train/grad_norm_pre_clip': 0.0006256172782741487, 'timing/data': 0.00019774999236688018, 'timing/model': 0.3649068609956885, 'learning_rate/qwen_vl_interface': 2.6587092192498943e-06, 'learning_rate/action_model': 2.658709219249894e-05, 'epoch': 0.25}) +[2026-10-03 15:17:15,732][starVLA.training.train_starvla][INFO] - Step 3451, Loss: {'train/loss': 0.00015146407531574368, 'train/grad_norm_pre_clip': 0.1184937059879303, 'timing/data': 0.0002681799960555509, 'timing/model': 0.3650408709945623, 'learning_rate/qwen_vl_interface': 2.656157420524084e-06, 'learning_rate/action_model': 2.6561574205240842e-05, 'epoch': 0.25}) +[2026-10-03 15:17:16,379][starVLA.training.train_starvla][INFO] - Step 3452, Loss: {'train/loss': 5.945359589532018e-06, 'train/grad_norm_pre_clip': 0.00306807947345078, 'timing/data': 0.0003278299991507083, 'timing/model': 0.3646008710056776, 'learning_rate/qwen_vl_interface': 2.6536066880289795e-06, 'learning_rate/action_model': 2.6536066880289794e-05, 'epoch': 0.25}) +[2026-10-03 15:17:17,003][starVLA.training.train_starvla][INFO] - Step 3453, Loss: {'train/loss': 2.5230561732314527e-05, 'train/grad_norm_pre_clip': 0.017236050218343735, 'timing/data': 0.0002972000074805692, 'timing/model': 0.36430243200447876, 'learning_rate/qwen_vl_interface': 2.6510570228130916e-06, 'learning_rate/action_model': 2.6510570228130914e-05, 'epoch': 0.25}) +[2026-10-03 15:17:17,650][starVLA.training.train_starvla][INFO] - Step 3454, Loss: {'train/loss': 1.221893057845591e-06, 'train/grad_norm_pre_clip': 0.00017741569899953902, 'timing/data': 0.0003570700064301491, 'timing/model': 0.36384225099754985, 'learning_rate/qwen_vl_interface': 2.648508425924491e-06, 'learning_rate/action_model': 2.6485084259244907e-05, 'epoch': 0.25}) +[2026-10-03 15:17:18,500][starVLA.training.train_starvla][INFO] - Step 3455, Loss: {'train/loss': 0.0036546853370964527, 'train/grad_norm_pre_clip': 12.02172565460205, 'timing/data': 0.0003171600110363215, 'timing/model': 0.36476316100743134, 'learning_rate/qwen_vl_interface': 2.6459608984108093e-06, 'learning_rate/action_model': 2.645960898410809e-05, 'epoch': 0.25}) +[2026-10-03 15:17:19,148][starVLA.training.train_starvla][INFO] - Step 3456, Loss: {'train/loss': 2.4288756321766414e-06, 'train/grad_norm_pre_clip': 0.000584381865337491, 'timing/data': 0.0001994500053115189, 'timing/model': 0.3645333510066848, 'learning_rate/qwen_vl_interface': 2.6434144413192414e-06, 'learning_rate/action_model': 2.6434144413192414e-05, 'epoch': 0.25}) +[2026-10-03 15:17:19,795][starVLA.training.train_starvla][INFO] - Step 3457, Loss: {'train/loss': 4.835379058931721e-06, 'train/grad_norm_pre_clip': 0.0012509478256106377, 'timing/data': 0.0004102400125702843, 'timing/model': 0.36432120100653265, 'learning_rate/qwen_vl_interface': 2.6408690556965347e-06, 'learning_rate/action_model': 2.640869055696535e-05, 'epoch': 0.25}) +[2026-10-03 15:17:20,443][starVLA.training.train_starvla][INFO] - Step 3458, Loss: {'train/loss': 0.00018862279830500484, 'train/grad_norm_pre_clip': 0.36058565974235535, 'timing/data': 0.00033863999124150723, 'timing/model': 0.3643262320110807, 'learning_rate/qwen_vl_interface': 2.6383247425890047e-06, 'learning_rate/action_model': 2.6383247425890047e-05, 'epoch': 0.25}) +[2026-10-03 15:17:21,092][starVLA.training.train_starvla][INFO] - Step 3459, Loss: {'train/loss': 4.932231604470871e-06, 'train/grad_norm_pre_clip': 0.001136660110205412, 'timing/data': 0.0002655499993124977, 'timing/model': 0.3641558219969738, 'learning_rate/qwen_vl_interface': 2.6357815030425204e-06, 'learning_rate/action_model': 2.6357815030425205e-05, 'epoch': 0.25}) +[2026-10-03 15:17:21,934][starVLA.training.train_starvla][INFO] - Step 3460, Loss: {'train/loss': 2.4107548597385176e-05, 'train/grad_norm_pre_clip': 0.022610122337937355, 'timing/data': 0.00019715000235009938, 'timing/model': 0.5592245750012808, 'learning_rate/qwen_vl_interface': 2.6332393381025132e-06, 'learning_rate/action_model': 2.6332393381025133e-05, 'epoch': 0.25}) +[2026-10-03 15:17:22,582][starVLA.training.train_starvla][INFO] - Step 3461, Loss: {'train/loss': 5.826136111863889e-05, 'train/grad_norm_pre_clip': 0.0987447127699852, 'timing/data': 0.0002655099960975349, 'timing/model': 0.36509875100455247, 'learning_rate/qwen_vl_interface': 2.6306982488139704e-06, 'learning_rate/action_model': 2.6306982488139702e-05, 'epoch': 0.25}) +[2026-10-03 15:17:23,229][starVLA.training.train_starvla][INFO] - Step 3462, Loss: {'train/loss': 0.17963114380836487, 'train/grad_norm_pre_clip': 99.90259552001953, 'timing/data': 0.00025394999829586595, 'timing/model': 0.3642969409993384, 'learning_rate/qwen_vl_interface': 2.628158236221438e-06, 'learning_rate/action_model': 2.628158236221438e-05, 'epoch': 0.25}) +[2026-10-03 15:17:23,879][starVLA.training.train_starvla][INFO] - Step 3463, Loss: {'train/loss': 0.0003257312346249819, 'train/grad_norm_pre_clip': 1.1519484519958496, 'timing/data': 0.0002275099977850914, 'timing/model': 0.365170760996989, 'learning_rate/qwen_vl_interface': 2.6256193013690183e-06, 'learning_rate/action_model': 2.6256193013690182e-05, 'epoch': 0.25}) +[2026-10-03 15:17:24,525][starVLA.training.train_starvla][INFO] - Step 3464, Loss: {'train/loss': 4.03817102778703e-06, 'train/grad_norm_pre_clip': 0.0021179397590458393, 'timing/data': 0.00036544998874887824, 'timing/model': 0.3643246809951961, 'learning_rate/qwen_vl_interface': 2.623081445300373e-06, 'learning_rate/action_model': 2.6230814453003728e-05, 'epoch': 0.25}) +[2026-10-03 15:17:25,172][starVLA.training.train_starvla][INFO] - Step 3465, Loss: {'train/loss': 2.5331896722491365e-06, 'train/grad_norm_pre_clip': 0.0009963797638192773, 'timing/data': 0.0003251999878557399, 'timing/model': 0.3640202910028165, 'learning_rate/qwen_vl_interface': 2.6205446690587183e-06, 'learning_rate/action_model': 2.6205446690587182e-05, 'epoch': 0.25}) +[2026-10-03 15:17:26,011][starVLA.training.train_starvla][INFO] - Step 3466, Loss: {'train/loss': 1.3231856428319588e-05, 'train/grad_norm_pre_clip': 0.0061603328213095665, 'timing/data': 0.0003207700065104291, 'timing/model': 0.36443014099495485, 'learning_rate/qwen_vl_interface': 2.618008973686828e-06, 'learning_rate/action_model': 2.618008973686828e-05, 'epoch': 0.25}) +[2026-10-03 15:17:26,660][starVLA.training.train_starvla][INFO] - Step 3467, Loss: {'train/loss': 0.39184725284576416, 'train/grad_norm_pre_clip': 141.904541015625, 'timing/data': 0.0003457099955994636, 'timing/model': 0.36473473100340925, 'learning_rate/qwen_vl_interface': 2.615474360227028e-06, 'learning_rate/action_model': 2.615474360227028e-05, 'epoch': 0.25}) +[2026-10-03 15:17:27,309][starVLA.training.train_starvla][INFO] - Step 3468, Loss: {'train/loss': 0.0010361484019085765, 'train/grad_norm_pre_clip': 1.5395777225494385, 'timing/data': 0.00020370000856928527, 'timing/model': 0.3647979709930951, 'learning_rate/qwen_vl_interface': 2.612940829721204e-06, 'learning_rate/action_model': 2.612940829721204e-05, 'epoch': 0.25}) +[2026-10-03 15:17:27,959][starVLA.training.train_starvla][INFO] - Step 3469, Loss: {'train/loss': 4.172292847215431e-06, 'train/grad_norm_pre_clip': 0.001134792692027986, 'timing/data': 0.00036891001218464226, 'timing/model': 0.3644681019941345, 'learning_rate/qwen_vl_interface': 2.6104083832107948e-06, 'learning_rate/action_model': 2.6104083832107945e-05, 'epoch': 0.25}) +[2026-10-03 15:17:28,606][starVLA.training.train_starvla][INFO] - Step 3470, Loss: {'train/loss': 0.0006993178394623101, 'train/grad_norm_pre_clip': 0.9960882663726807, 'timing/data': 0.0002449400053592399, 'timing/model': 0.36473951100197155, 'learning_rate/qwen_vl_interface': 2.607877021736793e-06, 'learning_rate/action_model': 2.6078770217367926e-05, 'epoch': 0.25}) +[2026-10-03 15:17:29,448][starVLA.training.train_starvla][INFO] - Step 3471, Loss: {'train/loss': 0.24992157518863678, 'train/grad_norm_pre_clip': 92.31877136230469, 'timing/data': 0.00030402999254874885, 'timing/model': 0.3651700210029958, 'learning_rate/qwen_vl_interface': 2.605346746339745e-06, 'learning_rate/action_model': 2.605346746339745e-05, 'epoch': 0.25}) +[2026-10-03 15:17:30,096][starVLA.training.train_starvla][INFO] - Step 3472, Loss: {'train/loss': 0.00046947674127295613, 'train/grad_norm_pre_clip': 0.11158816516399384, 'timing/data': 0.00031956999737303704, 'timing/model': 0.36469185100577306, 'learning_rate/qwen_vl_interface': 2.602817558059753e-06, 'learning_rate/action_model': 2.6028175580597532e-05, 'epoch': 0.25}) +[2026-10-03 15:17:30,744][starVLA.training.train_starvla][INFO] - Step 3473, Loss: {'train/loss': 5.051430434832582e-06, 'train/grad_norm_pre_clip': 0.0011295367730781436, 'timing/data': 0.00027597999724093825, 'timing/model': 0.3649614099995233, 'learning_rate/qwen_vl_interface': 2.60028945793647e-06, 'learning_rate/action_model': 2.60028945793647e-05, 'epoch': 0.25}) +[2026-10-03 15:17:31,368][starVLA.training.train_starvla][INFO] - Step 3474, Loss: {'train/loss': 9.931374734151177e-06, 'train/grad_norm_pre_clip': 0.004402397200465202, 'timing/data': 0.00032952000037766993, 'timing/model': 0.34182658400095534, 'learning_rate/qwen_vl_interface': 2.5977624470091024e-06, 'learning_rate/action_model': 2.5977624470091023e-05, 'epoch': 0.25}) +[2026-10-03 15:17:32,018][starVLA.training.train_starvla][INFO] - Step 3475, Loss: {'train/loss': 0.1198945865035057, 'train/grad_norm_pre_clip': 197.4086456298828, 'timing/data': 0.00025823999021667987, 'timing/model': 0.36537190100352746, 'learning_rate/qwen_vl_interface': 2.59523652631641e-06, 'learning_rate/action_model': 2.5952365263164097e-05, 'epoch': 0.25}) +[2026-10-03 15:17:32,867][starVLA.training.train_starvla][INFO] - Step 3476, Loss: {'train/loss': 2.4810283321130555e-06, 'train/grad_norm_pre_clip': 0.0009169208933599293, 'timing/data': 0.00026201999571640044, 'timing/model': 0.36444946100527886, 'learning_rate/qwen_vl_interface': 2.5927116968967024e-06, 'learning_rate/action_model': 2.5927116968967025e-05, 'epoch': 0.25}) +[2026-10-03 15:17:33,513][starVLA.training.train_starvla][INFO] - Step 3477, Loss: {'train/loss': 0.00041629301267676055, 'train/grad_norm_pre_clip': 0.16032356023788452, 'timing/data': 0.0003378000110387802, 'timing/model': 0.36439933099609334, 'learning_rate/qwen_vl_interface': 2.5901879597878417e-06, 'learning_rate/action_model': 2.5901879597878415e-05, 'epoch': 0.25}) +[2026-10-03 15:17:34,160][starVLA.training.train_starvla][INFO] - Step 3478, Loss: {'train/loss': 7.897609748397372e-07, 'train/grad_norm_pre_clip': 0.00017193415260408074, 'timing/data': 0.0003182200016453862, 'timing/model': 0.36415813099301886, 'learning_rate/qwen_vl_interface': 2.5876653160272395e-06, 'learning_rate/action_model': 2.5876653160272397e-05, 'epoch': 0.25}) +[2026-10-03 15:17:34,810][starVLA.training.train_starvla][INFO] - Step 3479, Loss: {'train/loss': 7.495151294278912e-06, 'train/grad_norm_pre_clip': 0.002382011851295829, 'timing/data': 0.00020380999194458127, 'timing/model': 0.36524093100160826, 'learning_rate/qwen_vl_interface': 2.5851437666518607e-06, 'learning_rate/action_model': 2.5851437666518607e-05, 'epoch': 0.25}) +[2026-10-03 15:17:35,459][starVLA.training.train_starvla][INFO] - Step 3480, Loss: {'train/loss': 6.57190103083849e-05, 'train/grad_norm_pre_clip': 0.05391073599457741, 'timing/data': 0.00019552999583538622, 'timing/model': 0.3648979810095625, 'learning_rate/qwen_vl_interface': 2.5826233126982198e-06, 'learning_rate/action_model': 2.5826233126982196e-05, 'epoch': 0.25}) +[2026-10-03 15:17:36,301][starVLA.training.train_starvla][INFO] - Step 3481, Loss: {'train/loss': 0.31173568964004517, 'train/grad_norm_pre_clip': 109.5343017578125, 'timing/data': 0.00027908000629395247, 'timing/model': 0.5581940659903921, 'learning_rate/qwen_vl_interface': 2.580103955202381e-06, 'learning_rate/action_model': 2.580103955202381e-05, 'epoch': 0.25}) +[2026-10-03 15:17:36,949][starVLA.training.train_starvla][INFO] - Step 3482, Loss: {'train/loss': 0.0011791940778493881, 'train/grad_norm_pre_clip': 2.838122606277466, 'timing/data': 0.00020261001191101968, 'timing/model': 0.36517927100067027, 'learning_rate/qwen_vl_interface': 2.577585695199954e-06, 'learning_rate/action_model': 2.5775856951999538e-05, 'epoch': 0.25}) +[2026-10-03 15:17:37,602][starVLA.training.train_starvla][INFO] - Step 3483, Loss: {'train/loss': 0.2399430125951767, 'train/grad_norm_pre_clip': 377.2105407714844, 'timing/data': 0.0003264499973738566, 'timing/model': 0.3666921010008082, 'learning_rate/qwen_vl_interface': 2.575068533726101e-06, 'learning_rate/action_model': 2.575068533726101e-05, 'epoch': 0.25}) +[2026-10-03 15:17:38,250][starVLA.training.train_starvla][INFO] - Step 3484, Loss: {'train/loss': 0.3889724314212799, 'train/grad_norm_pre_clip': 46.135337829589844, 'timing/data': 0.0002445699938107282, 'timing/model': 0.3646263709961204, 'learning_rate/qwen_vl_interface': 2.5725524718155333e-06, 'learning_rate/action_model': 2.572552471815533e-05, 'epoch': 0.26}) +[2026-10-03 15:17:38,898][starVLA.training.train_starvla][INFO] - Step 3485, Loss: {'train/loss': 0.4883331060409546, 'train/grad_norm_pre_clip': 88.74036407470703, 'timing/data': 0.00022339999850373715, 'timing/model': 0.36466440199001227, 'learning_rate/qwen_vl_interface': 2.570037510502508e-06, 'learning_rate/action_model': 2.5700375105025083e-05, 'epoch': 0.26}) +[2026-10-03 15:17:39,546][starVLA.training.train_starvla][INFO] - Step 3486, Loss: {'train/loss': 7.830431968613993e-06, 'train/grad_norm_pre_clip': 0.002347827423363924, 'timing/data': 0.00023221998708322644, 'timing/model': 0.36464192099811044, 'learning_rate/qwen_vl_interface': 2.5675236508208313e-06, 'learning_rate/action_model': 2.5675236508208313e-05, 'epoch': 0.26}) +[2026-10-03 15:17:40,386][starVLA.training.train_starvla][INFO] - Step 3487, Loss: {'train/loss': 3.870627551805228e-05, 'train/grad_norm_pre_clip': 0.055973540991544724, 'timing/data': 0.00023176000104285777, 'timing/model': 0.3649632109882077, 'learning_rate/qwen_vl_interface': 2.565010893803856e-06, 'learning_rate/action_model': 2.565010893803856e-05, 'epoch': 0.26}) +[2026-10-03 15:17:41,034][starVLA.training.train_starvla][INFO] - Step 3488, Loss: {'train/loss': 4.790673301613424e-06, 'train/grad_norm_pre_clip': 0.0013576560886576772, 'timing/data': 0.0003815300005953759, 'timing/model': 0.3647536110074725, 'learning_rate/qwen_vl_interface': 2.562499240484482e-06, 'learning_rate/action_model': 2.5624992404844816e-05, 'epoch': 0.26}) +[2026-10-03 15:17:41,660][starVLA.training.train_starvla][INFO] - Step 3489, Loss: {'train/loss': 6.1465916587621905e-06, 'train/grad_norm_pre_clip': 0.003090518992394209, 'timing/data': 0.00030619998869951814, 'timing/model': 0.36459512100555, 'learning_rate/qwen_vl_interface': 2.559988691895155e-06, 'learning_rate/action_model': 2.559988691895155e-05, 'epoch': 0.26}) +[2026-10-03 15:17:42,307][starVLA.training.train_starvla][INFO] - Step 3490, Loss: {'train/loss': 3.3452970455982722e-06, 'train/grad_norm_pre_clip': 0.0009051519446074963, 'timing/data': 0.0002979800046887249, 'timing/model': 0.36442084099689964, 'learning_rate/qwen_vl_interface': 2.5574792490678648e-06, 'learning_rate/action_model': 2.5574792490678644e-05, 'epoch': 0.26}) +[2026-10-03 15:17:42,957][starVLA.training.train_starvla][INFO] - Step 3491, Loss: {'train/loss': 0.4598284959793091, 'train/grad_norm_pre_clip': 60.32077407836914, 'timing/data': 0.00030766001145821065, 'timing/model': 0.36500411199813243, 'learning_rate/qwen_vl_interface': 2.5549709130341493e-06, 'learning_rate/action_model': 2.5549709130341493e-05, 'epoch': 0.26}) +[2026-10-03 15:17:43,797][starVLA.training.train_starvla][INFO] - Step 3492, Loss: {'train/loss': 5.267489086691057e-06, 'train/grad_norm_pre_clip': 0.0016566957347095013, 'timing/data': 0.0003661399969132617, 'timing/model': 0.3651255110016791, 'learning_rate/qwen_vl_interface': 2.5524636848250924e-06, 'learning_rate/action_model': 2.552463684825092e-05, 'epoch': 0.26}) +[2026-10-03 15:17:44,444][starVLA.training.train_starvla][INFO] - Step 3493, Loss: {'train/loss': 4.239349891577149e-06, 'train/grad_norm_pre_clip': 0.0015674124006181955, 'timing/data': 0.00029846999677829444, 'timing/model': 0.36450790100207087, 'learning_rate/qwen_vl_interface': 2.5499575654713186e-06, 'learning_rate/action_model': 2.5499575654713185e-05, 'epoch': 0.26}) +[2026-10-03 15:17:45,091][starVLA.training.train_starvla][INFO] - Step 3494, Loss: {'train/loss': 9.620101627660915e-05, 'train/grad_norm_pre_clip': 0.06351902335882187, 'timing/data': 0.00019894000433851033, 'timing/model': 0.3647735809936421, 'learning_rate/qwen_vl_interface': 2.547452556003001e-06, 'learning_rate/action_model': 2.547452556003001e-05, 'epoch': 0.26}) +[2026-10-03 15:17:45,741][starVLA.training.train_starvla][INFO] - Step 3495, Loss: {'train/loss': 6.48193827146315e-06, 'train/grad_norm_pre_clip': 0.0019423783523961902, 'timing/data': 0.00030115900153759867, 'timing/model': 0.3651994520041626, 'learning_rate/qwen_vl_interface': 2.5449486574498535e-06, 'learning_rate/action_model': 2.5449486574498534e-05, 'epoch': 0.26}) +[2026-10-03 15:17:46,390][starVLA.training.train_starvla][INFO] - Step 3496, Loss: {'train/loss': 0.21424070000648499, 'train/grad_norm_pre_clip': 171.68006896972656, 'timing/data': 0.00022868999803904444, 'timing/model': 0.3651784110115841, 'learning_rate/qwen_vl_interface': 2.5424458708411335e-06, 'learning_rate/action_model': 2.5424458708411335e-05, 'epoch': 0.26}) +[2026-10-03 15:17:47,235][starVLA.training.train_starvla][INFO] - Step 3497, Loss: {'train/loss': 7.003430255281273e-06, 'train/grad_norm_pre_clip': 0.00323160900734365, 'timing/data': 0.00038608000613749027, 'timing/model': 0.36445834100595675, 'learning_rate/qwen_vl_interface': 2.5399441972056437e-06, 'learning_rate/action_model': 2.5399441972056438e-05, 'epoch': 0.26}) +[2026-10-03 15:17:47,860][starVLA.training.train_starvla][INFO] - Step 3498, Loss: {'train/loss': 0.00042602905887179077, 'train/grad_norm_pre_clip': 0.3782687783241272, 'timing/data': 0.0003067299985559657, 'timing/model': 0.3645027609891258, 'learning_rate/qwen_vl_interface': 2.537443637571726e-06, 'learning_rate/action_model': 2.537443637571726e-05, 'epoch': 0.26}) +[2026-10-03 15:17:48,488][starVLA.training.train_starvla][INFO] - Step 3499, Loss: {'train/loss': 0.0005565151222981513, 'train/grad_norm_pre_clip': 0.7399619817733765, 'timing/data': 0.0002972599904751405, 'timing/model': 0.3655214610043913, 'learning_rate/qwen_vl_interface': 2.534944192967269e-06, 'learning_rate/action_model': 2.534944192967269e-05, 'epoch': 0.26}) +[2026-10-03 15:17:49,137][starVLA.training.train_starvla][INFO] - Step 3500, Loss: {'train/loss': 1.768609945429489e-05, 'train/grad_norm_pre_clip': 0.009284387342631817, 'timing/data': 0.0002577099949121475, 'timing/model': 0.3649286209983984, 'learning_rate/qwen_vl_interface': 2.532445864419695e-06, 'learning_rate/action_model': 2.5324458644196952e-05, 'epoch': 0.26}) +[2026-10-03 15:17:49,785][starVLA.training.train_starvla][INFO] - Step 3501, Loss: {'train/loss': 0.5474269986152649, 'train/grad_norm_pre_clip': 197.0359649658203, 'timing/data': 0.00019360000442247838, 'timing/model': 0.3653425610100385, 'learning_rate/qwen_vl_interface': 2.5299486529559758e-06, 'learning_rate/action_model': 2.5299486529559756e-05, 'epoch': 0.26}) +[2026-10-03 15:17:50,633][starVLA.training.train_starvla][INFO] - Step 3502, Loss: {'train/loss': 0.4633677303791046, 'train/grad_norm_pre_clip': 45.59209442138672, 'timing/data': 0.000331880000885576, 'timing/model': 0.5651176839892287, 'learning_rate/qwen_vl_interface': 2.5274525596026202e-06, 'learning_rate/action_model': 2.52745255960262e-05, 'epoch': 0.26}) +[2026-10-03 15:17:51,282][starVLA.training.train_starvla][INFO] - Step 3503, Loss: {'train/loss': 0.6265575885772705, 'train/grad_norm_pre_clip': 62.872859954833984, 'timing/data': 0.00029132000054232776, 'timing/model': 0.36508514199522324, 'learning_rate/qwen_vl_interface': 2.524957585385677e-06, 'learning_rate/action_model': 2.524957585385677e-05, 'epoch': 0.26}) +[2026-10-03 15:17:51,930][starVLA.training.train_starvla][INFO] - Step 3504, Loss: {'train/loss': 0.24774381518363953, 'train/grad_norm_pre_clip': 159.6603546142578, 'timing/data': 0.00024172000121325254, 'timing/model': 0.36509153099905234, 'learning_rate/qwen_vl_interface': 2.522463731330738e-06, 'learning_rate/action_model': 2.5224637313307374e-05, 'epoch': 0.26}) +[2026-10-03 15:17:52,577][starVLA.training.train_starvla][INFO] - Step 3505, Loss: {'train/loss': 0.07682378590106964, 'train/grad_norm_pre_clip': 157.44337463378906, 'timing/data': 0.00023864000104367733, 'timing/model': 0.3645269010012271, 'learning_rate/qwen_vl_interface': 2.5199709984629294e-06, 'learning_rate/action_model': 2.5199709984629292e-05, 'epoch': 0.26}) +[2026-10-03 15:17:53,223][starVLA.training.train_starvla][INFO] - Step 3506, Loss: {'train/loss': 3.499200465739705e-05, 'train/grad_norm_pre_clip': 0.014843367040157318, 'timing/data': 0.00032140999974217266, 'timing/model': 0.3638701810123166, 'learning_rate/qwen_vl_interface': 2.517479387806922e-06, 'learning_rate/action_model': 2.517479387806922e-05, 'epoch': 0.26}) +[2026-10-03 15:17:53,872][starVLA.training.train_starvla][INFO] - Step 3507, Loss: {'train/loss': 0.2478940337896347, 'train/grad_norm_pre_clip': 357.2347412109375, 'timing/data': 0.0003304800047772005, 'timing/model': 0.36525573099788744, 'learning_rate/qwen_vl_interface': 2.5149889003869217e-06, 'learning_rate/action_model': 2.5149889003869214e-05, 'epoch': 0.26}) +[2026-10-03 15:17:54,721][starVLA.training.train_starvla][INFO] - Step 3508, Loss: {'train/loss': 6.991896952968091e-05, 'train/grad_norm_pre_clip': 0.03226514905691147, 'timing/data': 0.0003401499998290092, 'timing/model': 0.3648861109977588, 'learning_rate/qwen_vl_interface': 2.5124995372266736e-06, 'learning_rate/action_model': 2.5124995372266735e-05, 'epoch': 0.26}) +[2026-10-03 15:17:55,368][starVLA.training.train_starvla][INFO] - Step 3509, Loss: {'train/loss': 3.65295163646806e-05, 'train/grad_norm_pre_clip': 0.016086675226688385, 'timing/data': 0.00019722001161426306, 'timing/model': 0.3644266509945737, 'learning_rate/qwen_vl_interface': 2.5100112993494626e-06, 'learning_rate/action_model': 2.5100112993494624e-05, 'epoch': 0.26}) +[2026-10-03 15:17:55,993][starVLA.training.train_starvla][INFO] - Step 3510, Loss: {'train/loss': 0.00012277276255190372, 'train/grad_norm_pre_clip': 0.037603091448545456, 'timing/data': 0.0003039799921680242, 'timing/model': 0.34164321499702055, 'learning_rate/qwen_vl_interface': 2.5075241877781085e-06, 'learning_rate/action_model': 2.5075241877781086e-05, 'epoch': 0.26}) +[2026-10-03 15:17:56,642][starVLA.training.train_starvla][INFO] - Step 3511, Loss: {'train/loss': 0.00010479799675522372, 'train/grad_norm_pre_clip': 0.0807994157075882, 'timing/data': 0.00028268000460229814, 'timing/model': 0.3649303909915034, 'learning_rate/qwen_vl_interface': 2.5050382035349673e-06, 'learning_rate/action_model': 2.5050382035349674e-05, 'epoch': 0.26}) +[2026-10-03 15:17:57,267][starVLA.training.train_starvla][INFO] - Step 3512, Loss: {'train/loss': 0.22425955533981323, 'train/grad_norm_pre_clip': 96.23430633544922, 'timing/data': 0.000315849989419803, 'timing/model': 0.3645025810110383, 'learning_rate/qwen_vl_interface': 2.5025533476419343e-06, 'learning_rate/action_model': 2.5025533476419343e-05, 'epoch': 0.26}) +[2026-10-03 15:17:58,116][starVLA.training.train_starvla][INFO] - Step 3513, Loss: {'train/loss': 0.00035542427212931216, 'train/grad_norm_pre_clip': 0.28439009189605713, 'timing/data': 0.0002527699980419129, 'timing/model': 0.36447780100570526, 'learning_rate/qwen_vl_interface': 2.5000696211204394e-06, 'learning_rate/action_model': 2.5000696211204394e-05, 'epoch': 0.26}) +[2026-10-03 15:17:58,741][starVLA.training.train_starvla][INFO] - Step 3514, Loss: {'train/loss': 3.077629298786633e-05, 'train/grad_norm_pre_clip': 0.023136958479881287, 'timing/data': 0.00020122999558225274, 'timing/model': 0.3424471840116894, 'learning_rate/qwen_vl_interface': 2.4975870249914495e-06, 'learning_rate/action_model': 2.4975870249914497e-05, 'epoch': 0.26}) +[2026-10-03 15:17:59,390][starVLA.training.train_starvla][INFO] - Step 3515, Loss: {'train/loss': 0.019700875505805016, 'train/grad_norm_pre_clip': 52.58401870727539, 'timing/data': 0.00025749999622348696, 'timing/model': 0.36461950099328533, 'learning_rate/qwen_vl_interface': 2.4951055602754664e-06, 'learning_rate/action_model': 2.4951055602754665e-05, 'epoch': 0.26}) +[2026-10-03 15:18:00,037][starVLA.training.train_starvla][INFO] - Step 3516, Loss: {'train/loss': 0.0004143830738030374, 'train/grad_norm_pre_clip': 0.08374961465597153, 'timing/data': 0.00033522999729029834, 'timing/model': 0.36460373100999277, 'learning_rate/qwen_vl_interface': 2.4926252279925255e-06, 'learning_rate/action_model': 2.4926252279925254e-05, 'epoch': 0.26}) +[2026-10-03 15:18:00,684][starVLA.training.train_starvla][INFO] - Step 3517, Loss: {'train/loss': 0.05259028822183609, 'train/grad_norm_pre_clip': 104.93416595458984, 'timing/data': 0.0003089700039708987, 'timing/model': 0.3644113810005365, 'learning_rate/qwen_vl_interface': 2.490146029162199e-06, 'learning_rate/action_model': 2.4901460291621988e-05, 'epoch': 0.26}) +[2026-10-03 15:18:01,532][starVLA.training.train_starvla][INFO] - Step 3518, Loss: {'train/loss': 2.895105535571929e-05, 'train/grad_norm_pre_clip': 0.009887675754725933, 'timing/data': 0.00030810000316705555, 'timing/model': 0.5651539750106167, 'learning_rate/qwen_vl_interface': 2.4876679648035905e-06, 'learning_rate/action_model': 2.4876679648035903e-05, 'epoch': 0.26}) +[2026-10-03 15:18:02,182][starVLA.training.train_starvla][INFO] - Step 3519, Loss: {'train/loss': 2.5397579520358704e-05, 'train/grad_norm_pre_clip': 0.015066845342516899, 'timing/data': 0.00022694999643135816, 'timing/model': 0.36510492200613953, 'learning_rate/qwen_vl_interface': 2.4851910359353392e-06, 'learning_rate/action_model': 2.485191035935339e-05, 'epoch': 0.26}) +[2026-10-03 15:18:02,831][starVLA.training.train_starvla][INFO] - Step 3520, Loss: {'train/loss': 5.998832784825936e-05, 'train/grad_norm_pre_clip': 0.031239261850714684, 'timing/data': 0.0003335199871798977, 'timing/model': 0.3652103110071039, 'learning_rate/qwen_vl_interface': 2.4827152435756185e-06, 'learning_rate/action_model': 2.4827152435756185e-05, 'epoch': 0.26}) +[2026-10-03 15:18:03,477][starVLA.training.train_starvla][INFO] - Step 3521, Loss: {'train/loss': 0.549336850643158, 'train/grad_norm_pre_clip': 83.2092056274414, 'timing/data': 0.00019797000277321786, 'timing/model': 0.36436361201049294, 'learning_rate/qwen_vl_interface': 2.4802405887421307e-06, 'learning_rate/action_model': 2.4802405887421306e-05, 'epoch': 0.26}) +[2026-10-03 15:18:04,126][starVLA.training.train_starvla][INFO] - Step 3522, Loss: {'train/loss': 0.0005979629931971431, 'train/grad_norm_pre_clip': 0.3929690718650818, 'timing/data': 0.0002957599936053157, 'timing/model': 0.3645574510010192, 'learning_rate/qwen_vl_interface': 2.4777670724521134e-06, 'learning_rate/action_model': 2.4777670724521134e-05, 'epoch': 0.26}) +[2026-10-03 15:18:04,775][starVLA.training.train_starvla][INFO] - Step 3523, Loss: {'train/loss': 2.690837936825119e-05, 'train/grad_norm_pre_clip': 0.010503079742193222, 'timing/data': 0.0002829099976224825, 'timing/model': 0.36484476100304164, 'learning_rate/qwen_vl_interface': 2.475294695722337e-06, 'learning_rate/action_model': 2.475294695722337e-05, 'epoch': 0.26}) +[2026-10-03 15:18:05,619][starVLA.training.train_starvla][INFO] - Step 3524, Loss: {'train/loss': 0.2971777021884918, 'train/grad_norm_pre_clip': 138.0994110107422, 'timing/data': 0.0003175000019837171, 'timing/model': 0.36492092099797446, 'learning_rate/qwen_vl_interface': 2.4728234595691006e-06, 'learning_rate/action_model': 2.4728234595691005e-05, 'epoch': 0.26}) +[2026-10-03 15:18:06,267][starVLA.training.train_starvla][INFO] - Step 3525, Loss: {'train/loss': 0.04091091454029083, 'train/grad_norm_pre_clip': 97.76951599121094, 'timing/data': 0.0002803899988066405, 'timing/model': 0.3649879109871108, 'learning_rate/qwen_vl_interface': 2.4703533650082368e-06, 'learning_rate/action_model': 2.4703533650082366e-05, 'epoch': 0.26}) +[2026-10-03 15:18:06,915][starVLA.training.train_starvla][INFO] - Step 3526, Loss: {'train/loss': 2.2052550775697455e-05, 'train/grad_norm_pre_clip': 0.007468171417713165, 'timing/data': 0.0003599299961933866, 'timing/model': 0.3649203709937865, 'learning_rate/qwen_vl_interface': 2.4678844130551103e-06, 'learning_rate/action_model': 2.4678844130551103e-05, 'epoch': 0.26}) +[2026-10-03 15:18:07,543][starVLA.training.train_starvla][INFO] - Step 3527, Loss: {'train/loss': 3.8593907447648235e-06, 'train/grad_norm_pre_clip': 0.0009472384117543697, 'timing/data': 0.0003290699969511479, 'timing/model': 0.3652039610024076, 'learning_rate/qwen_vl_interface': 2.4654166047246102e-06, 'learning_rate/action_model': 2.46541660472461e-05, 'epoch': 0.26}) +[2026-10-03 15:18:08,192][starVLA.training.train_starvla][INFO] - Step 3528, Loss: {'train/loss': 3.0096067348495126e-05, 'train/grad_norm_pre_clip': 0.03563878685235977, 'timing/data': 0.0003497400030028075, 'timing/model': 0.36525226100638974, 'learning_rate/qwen_vl_interface': 2.4629499410311606e-06, 'learning_rate/action_model': 2.462949941031161e-05, 'epoch': 0.26}) +[2026-10-03 15:18:09,044][starVLA.training.train_starvla][INFO] - Step 3529, Loss: {'train/loss': 0.00018377247033640742, 'train/grad_norm_pre_clip': 0.17427486181259155, 'timing/data': 0.0003566499945009127, 'timing/model': 0.3654449400055455, 'learning_rate/qwen_vl_interface': 2.460484422988715e-06, 'learning_rate/action_model': 2.4604844229887146e-05, 'epoch': 0.26}) +[2026-10-03 15:18:09,694][starVLA.training.train_starvla][INFO] - Step 3530, Loss: {'train/loss': 0.00022552305017597973, 'train/grad_norm_pre_clip': 0.05855545774102211, 'timing/data': 0.0003413599915802479, 'timing/model': 0.3658864720055135, 'learning_rate/qwen_vl_interface': 2.458020051610753e-06, 'learning_rate/action_model': 2.4580200516107528e-05, 'epoch': 0.26}) +[2026-10-03 15:18:10,346][starVLA.training.train_starvla][INFO] - Step 3531, Loss: {'train/loss': 5.2973400670452975e-06, 'train/grad_norm_pre_clip': 0.001116955652832985, 'timing/data': 0.0003114989958703518, 'timing/model': 0.36618848200305365, 'learning_rate/qwen_vl_interface': 2.455556827910286e-06, 'learning_rate/action_model': 2.455556827910286e-05, 'epoch': 0.26}) +[2026-10-03 15:18:10,996][starVLA.training.train_starvla][INFO] - Step 3532, Loss: {'train/loss': 0.00016220525139942765, 'train/grad_norm_pre_clip': 0.09297455847263336, 'timing/data': 0.0003485100023681298, 'timing/model': 0.3656132299947785, 'learning_rate/qwen_vl_interface': 2.453094752899853e-06, 'learning_rate/action_model': 2.453094752899853e-05, 'epoch': 0.26}) +[2026-10-03 15:18:11,646][starVLA.training.train_starvla][INFO] - Step 3533, Loss: {'train/loss': 0.36717385053634644, 'train/grad_norm_pre_clip': 131.83860778808594, 'timing/data': 0.00020048000442329794, 'timing/model': 0.36617036200186703, 'learning_rate/qwen_vl_interface': 2.4506338275915173e-06, 'learning_rate/action_model': 2.4506338275915172e-05, 'epoch': 0.26}) +[2026-10-03 15:18:12,472][starVLA.training.train_starvla][INFO] - Step 3534, Loss: {'train/loss': 0.40689897537231445, 'train/grad_norm_pre_clip': 47.59886932373047, 'timing/data': 0.00030752000748179853, 'timing/model': 0.5640062949969433, 'learning_rate/qwen_vl_interface': 2.4481740529968735e-06, 'learning_rate/action_model': 2.4481740529968734e-05, 'epoch': 0.26}) +[2026-10-03 15:18:13,124][starVLA.training.train_starvla][INFO] - Step 3535, Loss: {'train/loss': 5.7114477385766804e-05, 'train/grad_norm_pre_clip': 0.028768783435225487, 'timing/data': 0.0003833700029645115, 'timing/model': 0.36681548999331426, 'learning_rate/qwen_vl_interface': 2.4457154301270427e-06, 'learning_rate/action_model': 2.4457154301270427e-05, 'epoch': 0.26}) +[2026-10-03 15:18:13,776][starVLA.training.train_starvla][INFO] - Step 3536, Loss: {'train/loss': 0.00010019988985732198, 'train/grad_norm_pre_clip': 0.1746443510055542, 'timing/data': 0.00023769999097567052, 'timing/model': 0.36674913100432605, 'learning_rate/qwen_vl_interface': 2.443257959992671e-06, 'learning_rate/action_model': 2.443257959992671e-05, 'epoch': 0.26}) +[2026-10-03 15:18:14,427][starVLA.training.train_starvla][INFO] - Step 3537, Loss: {'train/loss': 0.3574999272823334, 'train/grad_norm_pre_clip': 11.05220890045166, 'timing/data': 0.00031462000333704054, 'timing/model': 0.36648702100501396, 'learning_rate/qwen_vl_interface': 2.440801643603933e-06, 'learning_rate/action_model': 2.440801643603933e-05, 'epoch': 0.26}) +[2026-10-03 15:18:15,076][starVLA.training.train_starvla][INFO] - Step 3538, Loss: {'train/loss': 0.0001303579774685204, 'train/grad_norm_pre_clip': 0.13579051196575165, 'timing/data': 0.00023967000015545636, 'timing/model': 0.3654709310067119, 'learning_rate/qwen_vl_interface': 2.4383464819705252e-06, 'learning_rate/action_model': 2.438346481970525e-05, 'epoch': 0.26}) +[2026-10-03 15:18:15,728][starVLA.training.train_starvla][INFO] - Step 3539, Loss: {'train/loss': 0.00031422782922163606, 'train/grad_norm_pre_clip': 1.0237702131271362, 'timing/data': 0.0002582000015536323, 'timing/model': 0.366292032005731, 'learning_rate/qwen_vl_interface': 2.435892476101674e-06, 'learning_rate/action_model': 2.435892476101674e-05, 'epoch': 0.26}) +[2026-10-03 15:18:16,574][starVLA.training.train_starvla][INFO] - Step 3540, Loss: {'train/loss': 8.541828719899058e-05, 'train/grad_norm_pre_clip': 0.05128614231944084, 'timing/data': 0.0003385199961485341, 'timing/model': 0.36594648100435734, 'learning_rate/qwen_vl_interface': 2.433439627006128e-06, 'learning_rate/action_model': 2.4334396270061277e-05, 'epoch': 0.26}) +[2026-10-03 15:18:17,224][starVLA.training.train_starvla][INFO] - Step 3541, Loss: {'train/loss': 6.76257986924611e-05, 'train/grad_norm_pre_clip': 0.043642885982990265, 'timing/data': 0.00021488999482244253, 'timing/model': 0.36624821000441443, 'learning_rate/qwen_vl_interface': 2.4309879356921602e-06, 'learning_rate/action_model': 2.43098793569216e-05, 'epoch': 0.26}) +[2026-10-03 15:18:17,874][starVLA.training.train_starvla][INFO] - Step 3542, Loss: {'train/loss': 2.823761860781815e-06, 'train/grad_norm_pre_clip': 0.0005976484972052276, 'timing/data': 0.00027876999229192734, 'timing/model': 0.3660328020050656, 'learning_rate/qwen_vl_interface': 2.428537403167569e-06, 'learning_rate/action_model': 2.4285374031675687e-05, 'epoch': 0.26}) +[2026-10-03 15:18:18,527][starVLA.training.train_starvla][INFO] - Step 3543, Loss: {'train/loss': 0.00020385599054861814, 'train/grad_norm_pre_clip': 0.09189221262931824, 'timing/data': 0.0003100100002484396, 'timing/model': 0.36667483000201173, 'learning_rate/qwen_vl_interface': 2.4260880304396765e-06, 'learning_rate/action_model': 2.4260880304396765e-05, 'epoch': 0.26}) +[2026-10-03 15:18:19,178][starVLA.training.train_starvla][INFO] - Step 3544, Loss: {'train/loss': 0.011595151387155056, 'train/grad_norm_pre_clip': 16.62198829650879, 'timing/data': 0.00033881000126712024, 'timing/model': 0.36609769200731535, 'learning_rate/qwen_vl_interface': 2.4236398185153245e-06, 'learning_rate/action_model': 2.4236398185153245e-05, 'epoch': 0.26}) +[2026-10-03 15:18:20,020][starVLA.training.train_starvla][INFO] - Step 3545, Loss: {'train/loss': 4.211653504171409e-05, 'train/grad_norm_pre_clip': 0.019859839230775833, 'timing/data': 0.0002467699960106984, 'timing/model': 0.5575940849957988, 'learning_rate/qwen_vl_interface': 2.4211927684008823e-06, 'learning_rate/action_model': 2.4211927684008824e-05, 'epoch': 0.26}) +[2026-10-03 15:18:20,671][starVLA.training.train_starvla][INFO] - Step 3546, Loss: {'train/loss': 6.844249583082274e-05, 'train/grad_norm_pre_clip': 0.05177467316389084, 'timing/data': 0.00024322999524883926, 'timing/model': 0.3668521910003619, 'learning_rate/qwen_vl_interface': 2.4187468811022398e-06, 'learning_rate/action_model': 2.4187468811022396e-05, 'epoch': 0.26}) +[2026-10-03 15:18:21,324][starVLA.training.train_starvla][INFO] - Step 3547, Loss: {'train/loss': 0.23517146706581116, 'train/grad_norm_pre_clip': 45.389610290527344, 'timing/data': 0.00021984000341035426, 'timing/model': 0.36685483099427074, 'learning_rate/qwen_vl_interface': 2.416302157624809e-06, 'learning_rate/action_model': 2.4163021576248088e-05, 'epoch': 0.26}) +[2026-10-03 15:18:21,976][starVLA.training.train_starvla][INFO] - Step 3548, Loss: {'train/loss': 0.2826160490512848, 'train/grad_norm_pre_clip': 95.21613311767578, 'timing/data': 0.0002448599989293143, 'timing/model': 0.3668860309990123, 'learning_rate/qwen_vl_interface': 2.4138585989735227e-06, 'learning_rate/action_model': 2.4138585989735226e-05, 'epoch': 0.26}) +[2026-10-03 15:18:22,605][starVLA.training.train_starvla][INFO] - Step 3549, Loss: {'train/loss': 0.0003387320030014962, 'train/grad_norm_pre_clip': 0.1406940519809723, 'timing/data': 0.00019764000899158418, 'timing/model': 0.34436203399673104, 'learning_rate/qwen_vl_interface': 2.4114162061528357e-06, 'learning_rate/action_model': 2.4114162061528354e-05, 'epoch': 0.26}) +[2026-10-03 15:18:23,451][starVLA.training.train_starvla][INFO] - Step 3550, Loss: {'train/loss': 2.3182936274679378e-05, 'train/grad_norm_pre_clip': 0.013300087302923203, 'timing/data': 0.00033396999060641974, 'timing/model': 0.5621748650009977, 'learning_rate/qwen_vl_interface': 2.408974980166724e-06, 'learning_rate/action_model': 2.408974980166724e-05, 'epoch': 0.26}) +[2026-10-03 15:18:24,103][starVLA.training.train_starvla][INFO] - Step 3551, Loss: {'train/loss': 0.0001567685103509575, 'train/grad_norm_pre_clip': 0.06947865337133408, 'timing/data': 0.00037174001045059413, 'timing/model': 0.365932710992638, 'learning_rate/qwen_vl_interface': 2.406534922018683e-06, 'learning_rate/action_model': 2.4065349220186827e-05, 'epoch': 0.26}) +[2026-10-03 15:18:24,754][starVLA.training.train_starvla][INFO] - Step 3552, Loss: {'train/loss': 5.602788405667525e-06, 'train/grad_norm_pre_clip': 0.0016887157689779997, 'timing/data': 0.00020008999854326248, 'timing/model': 0.36633007100317627, 'learning_rate/qwen_vl_interface': 2.4040960327117287e-06, 'learning_rate/action_model': 2.4040960327117283e-05, 'epoch': 0.26}) +[2026-10-03 15:18:25,405][starVLA.training.train_starvla][INFO] - Step 3553, Loss: {'train/loss': 0.00025819672737270594, 'train/grad_norm_pre_clip': 0.2690742015838623, 'timing/data': 0.00019932999566663057, 'timing/model': 0.366090981013258, 'learning_rate/qwen_vl_interface': 2.4016583132483976e-06, 'learning_rate/action_model': 2.4016583132483976e-05, 'epoch': 0.26}) +[2026-10-03 15:18:26,054][starVLA.training.train_starvla][INFO] - Step 3554, Loss: {'train/loss': 0.0006476416019722819, 'train/grad_norm_pre_clip': 0.5195575952529907, 'timing/data': 0.0002611899981275201, 'timing/model': 0.3656821110052988, 'learning_rate/qwen_vl_interface': 2.399221764630741e-06, 'learning_rate/action_model': 2.3992217646307406e-05, 'epoch': 0.26}) +[2026-10-03 15:18:26,707][starVLA.training.train_starvla][INFO] - Step 3555, Loss: {'train/loss': 0.00025740277487784624, 'train/grad_norm_pre_clip': 0.09774746000766754, 'timing/data': 0.0003005700127687305, 'timing/model': 0.36653135099913925, 'learning_rate/qwen_vl_interface': 2.3967863878603334e-06, 'learning_rate/action_model': 2.3967863878603332e-05, 'epoch': 0.26}) +[2026-10-03 15:18:27,554][starVLA.training.train_starvla][INFO] - Step 3556, Loss: {'train/loss': 0.00028266411391086876, 'train/grad_norm_pre_clip': 0.24552589654922485, 'timing/data': 0.0003402599977562204, 'timing/model': 0.36637569099548273, 'learning_rate/qwen_vl_interface': 2.394352183938266e-06, 'learning_rate/action_model': 2.3943521839382664e-05, 'epoch': 0.26}) +[2026-10-03 15:18:28,205][starVLA.training.train_starvla][INFO] - Step 3557, Loss: {'train/loss': 0.0003497649449855089, 'train/grad_norm_pre_clip': 0.13189437985420227, 'timing/data': 0.00023979999241419137, 'timing/model': 0.36632501099666115, 'learning_rate/qwen_vl_interface': 2.39191915386515e-06, 'learning_rate/action_model': 2.39191915386515e-05, 'epoch': 0.26}) +[2026-10-03 15:18:28,832][starVLA.training.train_starvla][INFO] - Step 3558, Loss: {'train/loss': 0.00048050982877612114, 'train/grad_norm_pre_clip': 0.7903409600257874, 'timing/data': 0.00027801000396721065, 'timing/model': 0.36589113099034876, 'learning_rate/qwen_vl_interface': 2.3894872986411095e-06, 'learning_rate/action_model': 2.3894872986411094e-05, 'epoch': 0.26}) +[2026-10-03 15:18:29,485][starVLA.training.train_starvla][INFO] - Step 3559, Loss: {'train/loss': 7.093618478393182e-05, 'train/grad_norm_pre_clip': 0.03338133916258812, 'timing/data': 0.00037193999742157757, 'timing/model': 0.3669608110067202, 'learning_rate/qwen_vl_interface': 2.3870566192657903e-06, 'learning_rate/action_model': 2.38705661926579e-05, 'epoch': 0.26}) +[2026-10-03 15:18:30,136][starVLA.training.train_starvla][INFO] - Step 3560, Loss: {'train/loss': 0.6572555899620056, 'train/grad_norm_pre_clip': 29.30546760559082, 'timing/data': 0.000258529995335266, 'timing/model': 0.3660575409885496, 'learning_rate/qwen_vl_interface': 2.3846271167383515e-06, 'learning_rate/action_model': 2.3846271167383515e-05, 'epoch': 0.26}) +[2026-10-03 15:18:30,983][starVLA.training.train_starvla][INFO] - Step 3561, Loss: {'train/loss': 0.01965968869626522, 'train/grad_norm_pre_clip': 22.790332794189453, 'timing/data': 0.00019434999558143318, 'timing/model': 0.5628416649997234, 'learning_rate/qwen_vl_interface': 2.382198792057471e-06, 'learning_rate/action_model': 2.3821987920574707e-05, 'epoch': 0.26}) +[2026-10-03 15:18:31,632][starVLA.training.train_starvla][INFO] - Step 3562, Loss: {'train/loss': 0.0007691374630667269, 'train/grad_norm_pre_clip': 1.1049267053604126, 'timing/data': 0.00024374001077376306, 'timing/model': 0.36578973100404255, 'learning_rate/qwen_vl_interface': 2.3797716462213402e-06, 'learning_rate/action_model': 2.3797716462213402e-05, 'epoch': 0.26}) +[2026-10-03 15:18:32,285][starVLA.training.train_starvla][INFO] - Step 3563, Loss: {'train/loss': 0.0010472676949575543, 'train/grad_norm_pre_clip': 0.23334936797618866, 'timing/data': 0.0003004200116265565, 'timing/model': 0.3664566919906065, 'learning_rate/qwen_vl_interface': 2.37734568022767e-06, 'learning_rate/action_model': 2.3773456802276697e-05, 'epoch': 0.26}) +[2026-10-03 15:18:32,937][starVLA.training.train_starvla][INFO] - Step 3564, Loss: {'train/loss': 1.1995421118626837e-06, 'train/grad_norm_pre_clip': 0.00021951952658127993, 'timing/data': 0.0002853799960575998, 'timing/model': 0.3661029009963386, 'learning_rate/qwen_vl_interface': 2.3749208950736783e-06, 'learning_rate/action_model': 2.374920895073678e-05, 'epoch': 0.26}) +[2026-10-03 15:18:33,587][starVLA.training.train_starvla][INFO] - Step 3565, Loss: {'train/loss': 0.5142738819122314, 'train/grad_norm_pre_clip': 35.63885498046875, 'timing/data': 0.00031720999686513096, 'timing/model': 0.3658828220068244, 'learning_rate/qwen_vl_interface': 2.372497291756106e-06, 'learning_rate/action_model': 2.3724972917561058e-05, 'epoch': 0.26}) +[2026-10-03 15:18:34,434][starVLA.training.train_starvla][INFO] - Step 3566, Loss: {'train/loss': 0.0018320843810215592, 'train/grad_norm_pre_clip': 0.6005015969276428, 'timing/data': 0.00019731000065803528, 'timing/model': 0.5629925840039505, 'learning_rate/qwen_vl_interface': 2.3700748712712043e-06, 'learning_rate/action_model': 2.370074871271204e-05, 'epoch': 0.26}) +[2026-10-03 15:18:35,086][starVLA.training.train_starvla][INFO] - Step 3567, Loss: {'train/loss': 3.0398173294088338e-06, 'train/grad_norm_pre_clip': 0.001007327577099204, 'timing/data': 0.00028747000033035874, 'timing/model': 0.3663052920019254, 'learning_rate/qwen_vl_interface': 2.3676536346147388e-06, 'learning_rate/action_model': 2.3676536346147388e-05, 'epoch': 0.26}) +[2026-10-03 15:18:35,737][starVLA.training.train_starvla][INFO] - Step 3568, Loss: {'train/loss': 0.030303949490189552, 'train/grad_norm_pre_clip': 202.06082153320312, 'timing/data': 0.00033459000405855477, 'timing/model': 0.3666312699933769, 'learning_rate/qwen_vl_interface': 2.365233582781988e-06, 'learning_rate/action_model': 2.365233582781988e-05, 'epoch': 0.26}) +[2026-10-03 15:18:36,387][starVLA.training.train_starvla][INFO] - Step 3569, Loss: {'train/loss': 5.512291318154894e-05, 'train/grad_norm_pre_clip': 0.03289235010743141, 'timing/data': 0.00031011999817565084, 'timing/model': 0.36582161200931296, 'learning_rate/qwen_vl_interface': 2.3628147167677444e-06, 'learning_rate/action_model': 2.3628147167677443e-05, 'epoch': 0.26}) +[2026-10-03 15:18:37,037][starVLA.training.train_starvla][INFO] - Step 3570, Loss: {'train/loss': 0.004655126016587019, 'train/grad_norm_pre_clip': 4.981071472167969, 'timing/data': 0.0002534399973228574, 'timing/model': 0.3658208410051884, 'learning_rate/qwen_vl_interface': 2.3603970375663123e-06, 'learning_rate/action_model': 2.3603970375663125e-05, 'epoch': 0.26}) +[2026-10-03 15:18:37,689][starVLA.training.train_starvla][INFO] - Step 3571, Loss: {'train/loss': 2.0465698980842717e-05, 'train/grad_norm_pre_clip': 0.007495664991438389, 'timing/data': 0.0002841100067598745, 'timing/model': 0.36656399100320414, 'learning_rate/qwen_vl_interface': 2.35798054617151e-06, 'learning_rate/action_model': 2.3579805461715096e-05, 'epoch': 0.26}) +[2026-10-03 15:18:38,534][starVLA.training.train_starvla][INFO] - Step 3572, Loss: {'train/loss': 0.0012101527536287904, 'train/grad_norm_pre_clip': 0.35742858052253723, 'timing/data': 0.00038081000093370676, 'timing/model': 0.36612115100433584, 'learning_rate/qwen_vl_interface': 2.355565243576662e-06, 'learning_rate/action_model': 2.3555652435766617e-05, 'epoch': 0.26}) +[2026-10-03 15:18:39,186][starVLA.training.train_starvla][INFO] - Step 3573, Loss: {'train/loss': 1.7336918972432613e-05, 'train/grad_norm_pre_clip': 0.004564111120998859, 'timing/data': 0.0003042600001208484, 'timing/model': 0.3666912710032193, 'learning_rate/qwen_vl_interface': 2.35315113077461e-06, 'learning_rate/action_model': 2.3531511307746098e-05, 'epoch': 0.26}) +[2026-10-03 15:18:39,836][starVLA.training.train_starvla][INFO] - Step 3574, Loss: {'train/loss': 9.350349500891753e-06, 'train/grad_norm_pre_clip': 0.002619004575535655, 'timing/data': 0.00021287999697960913, 'timing/model': 0.36626759100181516, 'learning_rate/qwen_vl_interface': 2.3507382087577047e-06, 'learning_rate/action_model': 2.350738208757705e-05, 'epoch': 0.26}) +[2026-10-03 15:18:40,488][starVLA.training.train_starvla][INFO] - Step 3575, Loss: {'train/loss': 0.0005352041334845126, 'train/grad_norm_pre_clip': 0.16172026097774506, 'timing/data': 0.0003016400005435571, 'timing/model': 0.3667305800045142, 'learning_rate/qwen_vl_interface': 2.348326478517807e-06, 'learning_rate/action_model': 2.3483264785178072e-05, 'epoch': 0.26}) +[2026-10-03 15:18:41,140][starVLA.training.train_starvla][INFO] - Step 3576, Loss: {'train/loss': 0.44438767433166504, 'train/grad_norm_pre_clip': 178.07847595214844, 'timing/data': 0.00024404999567195773, 'timing/model': 0.3666220820014132, 'learning_rate/qwen_vl_interface': 2.3459159410462907e-06, 'learning_rate/action_model': 2.3459159410462905e-05, 'epoch': 0.26}) +[2026-10-03 15:18:41,977][starVLA.training.train_starvla][INFO] - Step 3577, Loss: {'train/loss': 2.3362686988548376e-05, 'train/grad_norm_pre_clip': 0.009333822876214981, 'timing/data': 0.00031689999741502106, 'timing/model': 0.5533260559896007, 'learning_rate/qwen_vl_interface': 2.343506597334031e-06, 'learning_rate/action_model': 2.343506597334031e-05, 'epoch': 0.26}) +[2026-10-03 15:18:42,627][starVLA.training.train_starvla][INFO] - Step 3578, Loss: {'train/loss': 0.0007981056696735322, 'train/grad_norm_pre_clip': 1.0798014402389526, 'timing/data': 0.00030292999872472137, 'timing/model': 0.3661501610040432, 'learning_rate/qwen_vl_interface': 2.341098448371423e-06, 'learning_rate/action_model': 2.3410984483714227e-05, 'epoch': 0.26}) +[2026-10-03 15:18:43,279][starVLA.training.train_starvla][INFO] - Step 3579, Loss: {'train/loss': 7.90492686064681e-06, 'train/grad_norm_pre_clip': 0.0038959409575909376, 'timing/data': 0.0003146499948343262, 'timing/model': 0.3666613710083766, 'learning_rate/qwen_vl_interface': 2.338691495148363e-06, 'learning_rate/action_model': 2.3386914951483626e-05, 'epoch': 0.26}) +[2026-10-03 15:18:43,931][starVLA.training.train_starvla][INFO] - Step 3580, Loss: {'train/loss': 0.0059839533641934395, 'train/grad_norm_pre_clip': 18.535249710083008, 'timing/data': 0.0003168800030834973, 'timing/model': 0.36711156099045184, 'learning_rate/qwen_vl_interface': 2.33628573865426e-06, 'learning_rate/action_model': 2.3362857386542596e-05, 'epoch': 0.26}) +[2026-10-03 15:18:44,581][starVLA.training.train_starvla][INFO] - Step 3581, Loss: {'train/loss': 0.0003991409612353891, 'train/grad_norm_pre_clip': 0.15372234582901, 'timing/data': 0.00026868999702855945, 'timing/model': 0.3660239219898358, 'learning_rate/qwen_vl_interface': 2.3338811798780282e-06, 'learning_rate/action_model': 2.333881179878028e-05, 'epoch': 0.26}) +[2026-10-03 15:18:45,424][starVLA.training.train_starvla][INFO] - Step 3582, Loss: {'train/loss': 0.30345165729522705, 'train/grad_norm_pre_clip': 82.36693572998047, 'timing/data': 0.0002891400072257966, 'timing/model': 0.5592656249937136, 'learning_rate/qwen_vl_interface': 2.3314778198080933e-06, 'learning_rate/action_model': 2.331477819808093e-05, 'epoch': 0.26}) +[2026-10-03 15:18:46,077][starVLA.training.train_starvla][INFO] - Step 3583, Loss: {'train/loss': 0.0001410122204106301, 'train/grad_norm_pre_clip': 0.10214384645223618, 'timing/data': 0.00026764000358525664, 'timing/model': 0.3666499509999994, 'learning_rate/qwen_vl_interface': 2.3290756594323834e-06, 'learning_rate/action_model': 2.3290756594323836e-05, 'epoch': 0.26}) +[2026-10-03 15:18:46,728][starVLA.training.train_starvla][INFO] - Step 3584, Loss: {'train/loss': 6.627138645853847e-05, 'train/grad_norm_pre_clip': 0.049829959869384766, 'timing/data': 0.0002564499882282689, 'timing/model': 0.36646337099955417, 'learning_rate/qwen_vl_interface': 2.3266746997383378e-06, 'learning_rate/action_model': 2.3266746997383377e-05, 'epoch': 0.26}) +[2026-10-03 15:18:47,379][starVLA.training.train_starvla][INFO] - Step 3585, Loss: {'train/loss': 0.08413981646299362, 'train/grad_norm_pre_clip': 179.61891174316406, 'timing/data': 0.0002612100070109591, 'timing/model': 0.3662939209898468, 'learning_rate/qwen_vl_interface': 2.324274941712899e-06, 'learning_rate/action_model': 2.324274941712899e-05, 'epoch': 0.26}) +[2026-10-03 15:18:48,029][starVLA.training.train_starvla][INFO] - Step 3586, Loss: {'train/loss': 0.00730357738211751, 'train/grad_norm_pre_clip': 28.886600494384766, 'timing/data': 0.00031856998975854367, 'timing/model': 0.36518603199510835, 'learning_rate/qwen_vl_interface': 2.321876386342519e-06, 'learning_rate/action_model': 2.3218763863425187e-05, 'epoch': 0.26}) +[2026-10-03 15:18:48,684][starVLA.training.train_starvla][INFO] - Step 3587, Loss: {'train/loss': 6.28823327133432e-05, 'train/grad_norm_pre_clip': 0.026705889031291008, 'timing/data': 0.0003703500115079805, 'timing/model': 0.3665838910092134, 'learning_rate/qwen_vl_interface': 2.31947903461315e-06, 'learning_rate/action_model': 2.3194790346131502e-05, 'epoch': 0.26}) +[2026-10-03 15:18:49,533][starVLA.training.train_starvla][INFO] - Step 3588, Loss: {'train/loss': 1.8528222426539287e-05, 'train/grad_norm_pre_clip': 0.011022049002349377, 'timing/data': 0.00020090999896638095, 'timing/model': 0.366644470996107, 'learning_rate/qwen_vl_interface': 2.3170828875102544e-06, 'learning_rate/action_model': 2.3170828875102543e-05, 'epoch': 0.26}) +[2026-10-03 15:18:50,183][starVLA.training.train_starvla][INFO] - Step 3589, Loss: {'train/loss': 3.174783705617301e-05, 'train/grad_norm_pre_clip': 0.022457655519247055, 'timing/data': 0.00020280000171624124, 'timing/model': 0.3665694410010474, 'learning_rate/qwen_vl_interface': 2.3146879460187984e-06, 'learning_rate/action_model': 2.3146879460187982e-05, 'epoch': 0.26}) +[2026-10-03 15:18:50,835][starVLA.training.train_starvla][INFO] - Step 3590, Loss: {'train/loss': 1.740406696626451e-05, 'train/grad_norm_pre_clip': 0.006636285223066807, 'timing/data': 0.00032423999800812453, 'timing/model': 0.3665381710015936, 'learning_rate/qwen_vl_interface': 2.3122942111232506e-06, 'learning_rate/action_model': 2.3122942111232506e-05, 'epoch': 0.26}) +[2026-10-03 15:18:51,487][starVLA.training.train_starvla][INFO] - Step 3591, Loss: {'train/loss': 0.0019468868849799037, 'train/grad_norm_pre_clip': 2.5356783866882324, 'timing/data': 0.0003740899992408231, 'timing/model': 0.36651599100150634, 'learning_rate/qwen_vl_interface': 2.3099016838075857e-06, 'learning_rate/action_model': 2.3099016838075856e-05, 'epoch': 0.26}) +[2026-10-03 15:18:52,138][starVLA.training.train_starvla][INFO] - Step 3592, Loss: {'train/loss': 8.128474291879684e-06, 'train/grad_norm_pre_clip': 0.0032161222770810127, 'timing/data': 0.0002086200111079961, 'timing/model': 0.36672046199964825, 'learning_rate/qwen_vl_interface': 2.307510365055282e-06, 'learning_rate/action_model': 2.307510365055282e-05, 'epoch': 0.26}) +[2026-10-03 15:18:52,990][starVLA.training.train_starvla][INFO] - Step 3593, Loss: {'train/loss': 0.014785942621529102, 'train/grad_norm_pre_clip': 37.6929931640625, 'timing/data': 0.00035249900247436017, 'timing/model': 0.36599517200374976, 'learning_rate/qwen_vl_interface': 2.3051202558493196e-06, 'learning_rate/action_model': 2.3051202558493192e-05, 'epoch': 0.26}) +[2026-10-03 15:18:53,639][starVLA.training.train_starvla][INFO] - Step 3594, Loss: {'train/loss': 5.7082692364929244e-05, 'train/grad_norm_pre_clip': 0.038350529968738556, 'timing/data': 0.00021908000053372234, 'timing/model': 0.36582531199383084, 'learning_rate/qwen_vl_interface': 2.302731357172182e-06, 'learning_rate/action_model': 2.302731357172182e-05, 'epoch': 0.26}) +[2026-10-03 15:18:54,292][starVLA.training.train_starvla][INFO] - Step 3595, Loss: {'train/loss': 0.0004844813665840775, 'train/grad_norm_pre_clip': 0.1195678561925888, 'timing/data': 0.00031458999728783965, 'timing/model': 0.3665579710068414, 'learning_rate/qwen_vl_interface': 2.3003436700058556e-06, 'learning_rate/action_model': 2.3003436700058554e-05, 'epoch': 0.26}) +[2026-10-03 15:18:54,943][starVLA.training.train_starvla][INFO] - Step 3596, Loss: {'train/loss': 0.0002828713913913816, 'train/grad_norm_pre_clip': 0.18511776626110077, 'timing/data': 0.0002621700114104897, 'timing/model': 0.3669063510023989, 'learning_rate/qwen_vl_interface': 2.297957195331828e-06, 'learning_rate/action_model': 2.2979571953318276e-05, 'epoch': 0.26}) +[2026-10-03 15:18:55,593][starVLA.training.train_starvla][INFO] - Step 3597, Loss: {'train/loss': 0.00016298075206577778, 'train/grad_norm_pre_clip': 0.05948469415307045, 'timing/data': 0.0003303099947515875, 'timing/model': 0.3658842419972643, 'learning_rate/qwen_vl_interface': 2.2955719341310903e-06, 'learning_rate/action_model': 2.2955719341310903e-05, 'epoch': 0.26}) +[2026-10-03 15:18:56,417][starVLA.training.train_starvla][INFO] - Step 3598, Loss: {'train/loss': 4.351109055278357e-06, 'train/grad_norm_pre_clip': 0.0016602366231381893, 'timing/data': 0.0002306600072188303, 'timing/model': 0.540155437993235, 'learning_rate/qwen_vl_interface': 2.2931878873841304e-06, 'learning_rate/action_model': 2.2931878873841302e-05, 'epoch': 0.26}) +[2026-10-03 15:18:57,069][starVLA.training.train_starvla][INFO] - Step 3599, Loss: {'train/loss': 1.5697851267759688e-05, 'train/grad_norm_pre_clip': 0.012596800923347473, 'timing/data': 0.000249080010689795, 'timing/model': 0.3662154509947868, 'learning_rate/qwen_vl_interface': 2.290805056070942e-06, 'learning_rate/action_model': 2.290805056070942e-05, 'epoch': 0.26}) +[2026-10-03 15:18:57,719][starVLA.training.train_starvla][INFO] - Step 3600, Loss: {'train/loss': 7.606880899402313e-06, 'train/grad_norm_pre_clip': 0.0018841889686882496, 'timing/data': 0.00033706000249367207, 'timing/model': 0.36577181100437883, 'learning_rate/qwen_vl_interface': 2.288423441171016e-06, 'learning_rate/action_model': 2.288423441171016e-05, 'epoch': 0.26}) +[2026-10-03 15:18:58,369][starVLA.training.train_starvla][INFO] - Step 3601, Loss: {'train/loss': 9.015190016725683e-07, 'train/grad_norm_pre_clip': 0.0003380406415089965, 'timing/data': 0.00024092000967357308, 'timing/model': 0.3658960109896725, 'learning_rate/qwen_vl_interface': 2.2860430436633447e-06, 'learning_rate/action_model': 2.2860430436633447e-05, 'epoch': 0.26}) +[2026-10-03 15:18:59,019][starVLA.training.train_starvla][INFO] - Step 3602, Loss: {'train/loss': 0.002280274871736765, 'train/grad_norm_pre_clip': 2.589224338531494, 'timing/data': 0.0002557400002842769, 'timing/model': 0.3654622309986735, 'learning_rate/qwen_vl_interface': 2.28366386452642e-06, 'learning_rate/action_model': 2.28366386452642e-05, 'epoch': 0.26}) +[2026-10-03 15:18:59,671][starVLA.training.train_starvla][INFO] - Step 3603, Loss: {'train/loss': 0.00041762980981729925, 'train/grad_norm_pre_clip': 0.22083796560764313, 'timing/data': 0.0003632300067692995, 'timing/model': 0.36474799099960364, 'learning_rate/qwen_vl_interface': 2.2812859047382326e-06, 'learning_rate/action_model': 2.2812859047382323e-05, 'epoch': 0.26}) +[2026-10-03 15:19:00,518][starVLA.training.train_starvla][INFO] - Step 3604, Loss: {'train/loss': 6.3478146330453455e-06, 'train/grad_norm_pre_clip': 0.001744598033837974, 'timing/data': 0.00020956998923793435, 'timing/model': 0.36484573100460693, 'learning_rate/qwen_vl_interface': 2.2789091652762712e-06, 'learning_rate/action_model': 2.2789091652762712e-05, 'epoch': 0.26}) +[2026-10-03 15:19:01,166][starVLA.training.train_starvla][INFO] - Step 3605, Loss: {'train/loss': 0.013448834419250488, 'train/grad_norm_pre_clip': 78.56402587890625, 'timing/data': 0.000221950002014637, 'timing/model': 0.36418629199033603, 'learning_rate/qwen_vl_interface': 2.276533647117524e-06, 'learning_rate/action_model': 2.2765336471175242e-05, 'epoch': 0.26}) +[2026-10-03 15:19:01,813][starVLA.training.train_starvla][INFO] - Step 3606, Loss: {'train/loss': 1.7448019207222387e-05, 'train/grad_norm_pre_clip': 0.006755023263394833, 'timing/data': 0.0003382300055818632, 'timing/model': 0.3643222320097266, 'learning_rate/qwen_vl_interface': 2.2741593512384773e-06, 'learning_rate/action_model': 2.2741593512384772e-05, 'epoch': 0.26}) +[2026-10-03 15:19:02,462][starVLA.training.train_starvla][INFO] - Step 3607, Loss: {'train/loss': 5.40161181561416e-06, 'train/grad_norm_pre_clip': 0.0011277565499767661, 'timing/data': 0.00026231000083498657, 'timing/model': 0.3646521710033994, 'learning_rate/qwen_vl_interface': 2.2717862786151153e-06, 'learning_rate/action_model': 2.271786278615115e-05, 'epoch': 0.26}) +[2026-10-03 15:19:03,110][starVLA.training.train_starvla][INFO] - Step 3608, Loss: {'train/loss': 0.5000107288360596, 'train/grad_norm_pre_clip': 50.723960876464844, 'timing/data': 0.0003356200031703338, 'timing/model': 0.3649686999997357, 'learning_rate/qwen_vl_interface': 2.2694144302229152e-06, 'learning_rate/action_model': 2.2694144302229152e-05, 'epoch': 0.26}) +[2026-10-03 15:19:03,952][starVLA.training.train_starvla][INFO] - Step 3609, Loss: {'train/loss': 0.00022794169490225613, 'train/grad_norm_pre_clip': 0.15529868006706238, 'timing/data': 0.000245189992710948, 'timing/model': 0.5596617850096663, 'learning_rate/qwen_vl_interface': 2.2670438070368568e-06, 'learning_rate/action_model': 2.2670438070368566e-05, 'epoch': 0.26}) +[2026-10-03 15:19:04,600][starVLA.training.train_starvla][INFO] - Step 3610, Loss: {'train/loss': 6.005089744576253e-06, 'train/grad_norm_pre_clip': 0.001279434422031045, 'timing/data': 0.00036383900442160666, 'timing/model': 0.3644714219990419, 'learning_rate/qwen_vl_interface': 2.264674410031413e-06, 'learning_rate/action_model': 2.2646744100314127e-05, 'epoch': 0.26}) +[2026-10-03 15:19:05,249][starVLA.training.train_starvla][INFO] - Step 3611, Loss: {'train/loss': 2.202184077759739e-05, 'train/grad_norm_pre_clip': 0.014790685847401619, 'timing/data': 0.00030834999051876366, 'timing/model': 0.3649209019931732, 'learning_rate/qwen_vl_interface': 2.2623062401805537e-06, 'learning_rate/action_model': 2.2623062401805535e-05, 'epoch': 0.26}) +[2026-10-03 15:19:05,897][starVLA.training.train_starvla][INFO] - Step 3612, Loss: {'train/loss': 0.002181250136345625, 'train/grad_norm_pre_clip': 3.9892537593841553, 'timing/data': 0.000245039991568774, 'timing/model': 0.3649002910096897, 'learning_rate/qwen_vl_interface': 2.2599392984577436e-06, 'learning_rate/action_model': 2.2599392984577438e-05, 'epoch': 0.26}) +[2026-10-03 15:19:06,546][starVLA.training.train_starvla][INFO] - Step 3613, Loss: {'train/loss': 1.681556568655651e-05, 'train/grad_norm_pre_clip': 0.005660432390868664, 'timing/data': 0.000338979996740818, 'timing/model': 0.3652404610038502, 'learning_rate/qwen_vl_interface': 2.2575735858359436e-06, 'learning_rate/action_model': 2.257573585835944e-05, 'epoch': 0.26}) +[2026-10-03 15:19:07,385][starVLA.training.train_starvla][INFO] - Step 3614, Loss: {'train/loss': 0.33872589468955994, 'train/grad_norm_pre_clip': 24.322114944458008, 'timing/data': 0.00022683999850414693, 'timing/model': 0.5559276059939293, 'learning_rate/qwen_vl_interface': 2.255209103287609e-06, 'learning_rate/action_model': 2.255209103287609e-05, 'epoch': 0.26}) +[2026-10-03 15:19:08,035][starVLA.training.train_starvla][INFO] - Step 3615, Loss: {'train/loss': 0.0005855630151927471, 'train/grad_norm_pre_clip': 0.8109052777290344, 'timing/data': 0.0002069100009975955, 'timing/model': 0.3651064809964737, 'learning_rate/qwen_vl_interface': 2.2528458517846886e-06, 'learning_rate/action_model': 2.252845851784689e-05, 'epoch': 0.26}) +[2026-10-03 15:19:08,684][starVLA.training.train_starvla][INFO] - Step 3616, Loss: {'train/loss': 2.3556345695396885e-05, 'train/grad_norm_pre_clip': 0.028968654572963715, 'timing/data': 0.0003348699974594638, 'timing/model': 0.36515461100498214, 'learning_rate/qwen_vl_interface': 2.250483832298627e-06, 'learning_rate/action_model': 2.250483832298627e-05, 'epoch': 0.26}) +[2026-10-03 15:19:09,331][starVLA.training.train_starvla][INFO] - Step 3617, Loss: {'train/loss': 2.1383011699072085e-06, 'train/grad_norm_pre_clip': 0.000496216001920402, 'timing/data': 0.00031805000617168844, 'timing/model': 0.36471699200046714, 'learning_rate/qwen_vl_interface': 2.2481230458003608e-06, 'learning_rate/action_model': 2.2481230458003606e-05, 'epoch': 0.26}) +[2026-10-03 15:19:09,978][starVLA.training.train_starvla][INFO] - Step 3618, Loss: {'train/loss': 2.7303998649585992e-05, 'train/grad_norm_pre_clip': 0.019480900838971138, 'timing/data': 0.0002607299975352362, 'timing/model': 0.36472732199763414, 'learning_rate/qwen_vl_interface': 2.2457634932603184e-06, 'learning_rate/action_model': 2.2457634932603183e-05, 'epoch': 0.26}) +[2026-10-03 15:19:10,628][starVLA.training.train_starvla][INFO] - Step 3619, Loss: {'train/loss': 1.2442415027180687e-06, 'train/grad_norm_pre_clip': 0.00023948555462993681, 'timing/data': 0.00030367900035344064, 'timing/model': 0.36531418199592736, 'learning_rate/qwen_vl_interface': 2.2434051756484255e-06, 'learning_rate/action_model': 2.2434051756484253e-05, 'epoch': 0.26}) +[2026-10-03 15:19:11,471][starVLA.training.train_starvla][INFO] - Step 3620, Loss: {'train/loss': 0.6039528846740723, 'train/grad_norm_pre_clip': 89.97259521484375, 'timing/data': 0.0003285300044808537, 'timing/model': 0.36522404199058656, 'learning_rate/qwen_vl_interface': 2.2410480939340945e-06, 'learning_rate/action_model': 2.2410480939340945e-05, 'epoch': 0.26}) +[2026-10-03 15:19:12,119][starVLA.training.train_starvla][INFO] - Step 3621, Loss: {'train/loss': 1.447618706151843e-05, 'train/grad_norm_pre_clip': 0.004343865439295769, 'timing/data': 0.00031505999504588544, 'timing/model': 0.36585480099893175, 'learning_rate/qwen_vl_interface': 2.238692249086234e-06, 'learning_rate/action_model': 2.2386922490862336e-05, 'epoch': 0.27}) +[2026-10-03 15:19:12,767][starVLA.training.train_starvla][INFO] - Step 3622, Loss: {'train/loss': 9.372544809593819e-06, 'train/grad_norm_pre_clip': 0.004318023566156626, 'timing/data': 0.0002442599943606183, 'timing/model': 0.3647162209963426, 'learning_rate/qwen_vl_interface': 2.2363376420732426e-06, 'learning_rate/action_model': 2.2363376420732423e-05, 'epoch': 0.27}) +[2026-10-03 15:19:13,416][starVLA.training.train_starvla][INFO] - Step 3623, Loss: {'train/loss': 7.822935003787279e-06, 'train/grad_norm_pre_clip': 0.0029759674798697233, 'timing/data': 0.00026748000527732074, 'timing/model': 0.36491902099805884, 'learning_rate/qwen_vl_interface': 2.233984273863011e-06, 'learning_rate/action_model': 2.233984273863011e-05, 'epoch': 0.27}) +[2026-10-03 15:19:14,065][starVLA.training.train_starvla][INFO] - Step 3624, Loss: {'train/loss': 0.3678602874279022, 'train/grad_norm_pre_clip': 23.272897720336914, 'timing/data': 0.00034186999255325645, 'timing/model': 0.3645805720007047, 'learning_rate/qwen_vl_interface': 2.231632145422919e-06, 'learning_rate/action_model': 2.231632145422919e-05, 'epoch': 0.27}) +[2026-10-03 15:19:14,910][starVLA.training.train_starvla][INFO] - Step 3625, Loss: {'train/loss': 8.620138032711111e-06, 'train/grad_norm_pre_clip': 0.0023861215449869633, 'timing/data': 0.00026082000113092363, 'timing/model': 0.36485745100071654, 'learning_rate/qwen_vl_interface': 2.2292812577198383e-06, 'learning_rate/action_model': 2.229281257719838e-05, 'epoch': 0.27}) +[2026-10-03 15:19:15,559][starVLA.training.train_starvla][INFO] - Step 3626, Loss: {'train/loss': 8.519069524481893e-05, 'train/grad_norm_pre_clip': 0.04913615435361862, 'timing/data': 0.0003503500047372654, 'timing/model': 0.3646462509932462, 'learning_rate/qwen_vl_interface': 2.2269316117201307e-06, 'learning_rate/action_model': 2.2269316117201307e-05, 'epoch': 0.27}) +[2026-10-03 15:19:16,209][starVLA.training.train_starvla][INFO] - Step 3627, Loss: {'train/loss': 3.3378332773281727e-06, 'train/grad_norm_pre_clip': 0.0007104373071342707, 'timing/data': 0.0002949600020656362, 'timing/model': 0.36511120200157166, 'learning_rate/qwen_vl_interface': 2.224583208389646e-06, 'learning_rate/action_model': 2.224583208389646e-05, 'epoch': 0.27}) +[2026-10-03 15:19:16,857][starVLA.training.train_starvla][INFO] - Step 3628, Loss: {'train/loss': 0.09052275121212006, 'train/grad_norm_pre_clip': 383.895263671875, 'timing/data': 0.00020080999820493162, 'timing/model': 0.36520841099263635, 'learning_rate/qwen_vl_interface': 2.2222360486937238e-06, 'learning_rate/action_model': 2.2222360486937235e-05, 'epoch': 0.27}) +[2026-10-03 15:19:17,505][starVLA.training.train_starvla][INFO] - Step 3629, Loss: {'train/loss': 7.964529686432797e-06, 'train/grad_norm_pre_clip': 0.0017010063165798783, 'timing/data': 0.0003313890047138557, 'timing/model': 0.3646795109962113, 'learning_rate/qwen_vl_interface': 2.219890133597194e-06, 'learning_rate/action_model': 2.219890133597194e-05, 'epoch': 0.27}) +[2026-10-03 15:19:18,352][starVLA.training.train_starvla][INFO] - Step 3630, Loss: {'train/loss': 0.0009245853871107101, 'train/grad_norm_pre_clip': 1.395219087600708, 'timing/data': 0.00031640999077353626, 'timing/model': 0.563674914999865, 'learning_rate/qwen_vl_interface': 2.217545464064374e-06, 'learning_rate/action_model': 2.217545464064374e-05, 'epoch': 0.27}) +[2026-10-03 15:19:19,002][starVLA.training.train_starvla][INFO] - Step 3631, Loss: {'train/loss': 0.24888499081134796, 'train/grad_norm_pre_clip': 40.82582092285156, 'timing/data': 0.00020837000920437276, 'timing/model': 0.36545217099774163, 'learning_rate/qwen_vl_interface': 2.215202041059067e-06, 'learning_rate/action_model': 2.215202041059067e-05, 'epoch': 0.27}) +[2026-10-03 15:19:19,628][starVLA.training.train_starvla][INFO] - Step 3632, Loss: {'train/loss': 3.792292318394175e-06, 'train/grad_norm_pre_clip': 0.0008912262273952365, 'timing/data': 0.0002492200001142919, 'timing/model': 0.3650588010059437, 'learning_rate/qwen_vl_interface': 2.212859865544566e-06, 'learning_rate/action_model': 2.212859865544566e-05, 'epoch': 0.27}) +[2026-10-03 15:19:20,276][starVLA.training.train_starvla][INFO] - Step 3633, Loss: {'train/loss': 6.185514939716086e-05, 'train/grad_norm_pre_clip': 0.09391659498214722, 'timing/data': 0.00031519999902229756, 'timing/model': 0.36458334200142417, 'learning_rate/qwen_vl_interface': 2.210518938483652e-06, 'learning_rate/action_model': 2.210518938483652e-05, 'epoch': 0.27}) +[2026-10-03 15:19:20,925][starVLA.training.train_starvla][INFO] - Step 3634, Loss: {'train/loss': 6.839441994088702e-06, 'train/grad_norm_pre_clip': 0.004258712287992239, 'timing/data': 0.00029267999343574047, 'timing/model': 0.3653739119909005, 'learning_rate/qwen_vl_interface': 2.20817926083859e-06, 'learning_rate/action_model': 2.20817926083859e-05, 'epoch': 0.27}) +[2026-10-03 15:19:21,773][starVLA.training.train_starvla][INFO] - Step 3635, Loss: {'train/loss': 1.3171471437090077e-05, 'train/grad_norm_pre_clip': 0.010290144942700863, 'timing/data': 0.0003430299984756857, 'timing/model': 0.5642834239988588, 'learning_rate/qwen_vl_interface': 2.205840833571135e-06, 'learning_rate/action_model': 2.205840833571135e-05, 'epoch': 0.27}) +[2026-10-03 15:19:22,422][starVLA.training.train_starvla][INFO] - Step 3636, Loss: {'train/loss': 0.00015623000217601657, 'train/grad_norm_pre_clip': 0.049891479313373566, 'timing/data': 0.0002039700048044324, 'timing/model': 0.36466312099946663, 'learning_rate/qwen_vl_interface': 2.2035036576425264e-06, 'learning_rate/action_model': 2.203503657642526e-05, 'epoch': 0.27}) +[2026-10-03 15:19:23,069][starVLA.training.train_starvla][INFO] - Step 3637, Loss: {'train/loss': 7.680933049414307e-05, 'train/grad_norm_pre_clip': 0.04347767308354378, 'timing/data': 0.0001972600002773106, 'timing/model': 0.36448871200263966, 'learning_rate/qwen_vl_interface': 2.2011677340134867e-06, 'learning_rate/action_model': 2.2011677340134867e-05, 'epoch': 0.27}) +[2026-10-03 15:19:23,724][starVLA.training.train_starvla][INFO] - Step 3638, Loss: {'train/loss': 1.4825674043095205e-05, 'train/grad_norm_pre_clip': 0.008811496198177338, 'timing/data': 0.0002608599897939712, 'timing/model': 0.3684532519982895, 'learning_rate/qwen_vl_interface': 2.198833063644227e-06, 'learning_rate/action_model': 2.198833063644227e-05, 'epoch': 0.27}) +[2026-10-03 15:19:24,382][starVLA.training.train_starvla][INFO] - Step 3639, Loss: {'train/loss': 0.0012094797566533089, 'train/grad_norm_pre_clip': 0.5672110915184021, 'timing/data': 0.0003371899947524071, 'timing/model': 0.3697738710034173, 'learning_rate/qwen_vl_interface': 2.196499647494443e-06, 'learning_rate/action_model': 2.196499647494443e-05, 'epoch': 0.27}) +[2026-10-03 15:19:25,016][starVLA.training.train_starvla][INFO] - Step 3640, Loss: {'train/loss': 0.00015698419883847237, 'train/grad_norm_pre_clip': 0.42434924840927124, 'timing/data': 0.00021803000709041953, 'timing/model': 0.3464684239879716, 'learning_rate/qwen_vl_interface': 2.1941674865233144e-06, 'learning_rate/action_model': 2.1941674865233143e-05, 'epoch': 0.27}) +[2026-10-03 15:19:25,847][starVLA.training.train_starvla][INFO] - Step 3641, Loss: {'train/loss': 0.2218545526266098, 'train/grad_norm_pre_clip': 25.430484771728516, 'timing/data': 0.00032681999437045306, 'timing/model': 0.3464012240001466, 'learning_rate/qwen_vl_interface': 2.1918365816895013e-06, 'learning_rate/action_model': 2.191836581689501e-05, 'epoch': 0.27}) +[2026-10-03 15:19:26,503][starVLA.training.train_starvla][INFO] - Step 3642, Loss: {'train/loss': 0.4614069163799286, 'train/grad_norm_pre_clip': 115.39484405517578, 'timing/data': 0.00025178999931085855, 'timing/model': 0.3682165520003764, 'learning_rate/qwen_vl_interface': 2.189506933951153e-06, 'learning_rate/action_model': 2.1895069339511525e-05, 'epoch': 0.27}) +[2026-10-03 15:19:27,161][starVLA.training.train_starvla][INFO] - Step 3643, Loss: {'train/loss': 2.569467505963985e-05, 'train/grad_norm_pre_clip': 0.008436094969511032, 'timing/data': 0.00038515999040100724, 'timing/model': 0.3695968910033116, 'learning_rate/qwen_vl_interface': 2.187178544265898e-06, 'learning_rate/action_model': 2.1871785442658983e-05, 'epoch': 0.27}) +[2026-10-03 15:19:27,819][starVLA.training.train_starvla][INFO] - Step 3644, Loss: {'train/loss': 0.7341535687446594, 'train/grad_norm_pre_clip': 34.06527328491211, 'timing/data': 0.0003449299983913079, 'timing/model': 0.37063290999503806, 'learning_rate/qwen_vl_interface': 2.184851413590852e-06, 'learning_rate/action_model': 2.1848514135908518e-05, 'epoch': 0.27}) +[2026-10-03 15:19:28,476][starVLA.training.train_starvla][INFO] - Step 3645, Loss: {'train/loss': 0.00013267189206089824, 'train/grad_norm_pre_clip': 0.07496403902769089, 'timing/data': 0.00033021999115590006, 'timing/model': 0.36946602100215387, 'learning_rate/qwen_vl_interface': 2.1825255428826087e-06, 'learning_rate/action_model': 2.182525542882609e-05, 'epoch': 0.27}) +[2026-10-03 15:19:29,335][starVLA.training.train_starvla][INFO] - Step 3646, Loss: {'train/loss': 0.004380055237561464, 'train/grad_norm_pre_clip': 16.38451385498047, 'timing/data': 0.0003168399998685345, 'timing/model': 0.36902353099139873, 'learning_rate/qwen_vl_interface': 2.180200933097246e-06, 'learning_rate/action_model': 2.180200933097246e-05, 'epoch': 0.27}) +[2026-10-03 15:19:29,992][starVLA.training.train_starvla][INFO] - Step 3647, Loss: {'train/loss': 4.343654381955275e-06, 'train/grad_norm_pre_clip': 0.0007746312185190618, 'timing/data': 0.0002194800035795197, 'timing/model': 0.36968411099223886, 'learning_rate/qwen_vl_interface': 2.1778775851903235e-06, 'learning_rate/action_model': 2.177877585190323e-05, 'epoch': 0.27}) +[2026-10-03 15:19:30,650][starVLA.training.train_starvla][INFO] - Step 3648, Loss: {'train/loss': 0.5908493995666504, 'train/grad_norm_pre_clip': 29.229230880737305, 'timing/data': 0.0002692499983822927, 'timing/model': 0.36985284100228455, 'learning_rate/qwen_vl_interface': 2.175555500116881e-06, 'learning_rate/action_model': 2.175555500116881e-05, 'epoch': 0.27}) +[2026-10-03 15:19:31,306][starVLA.training.train_starvla][INFO] - Step 3649, Loss: {'train/loss': 4.86520366393961e-06, 'train/grad_norm_pre_clip': 0.0014598388224840164, 'timing/data': 0.00031241000397130847, 'timing/model': 0.369168870995054, 'learning_rate/qwen_vl_interface': 2.1732346788314404e-06, 'learning_rate/action_model': 2.1732346788314402e-05, 'epoch': 0.27}) +[2026-10-03 15:19:31,940][starVLA.training.train_starvla][INFO] - Step 3650, Loss: {'train/loss': 2.6023581085610203e-05, 'train/grad_norm_pre_clip': 0.012952742166817188, 'timing/data': 0.0003236799966543913, 'timing/model': 0.36906937100866344, 'learning_rate/qwen_vl_interface': 2.170915122288006e-06, 'learning_rate/action_model': 2.170915122288006e-05, 'epoch': 0.27}) +[2026-10-03 15:19:32,798][starVLA.training.train_starvla][INFO] - Step 3651, Loss: {'train/loss': 4.497781628742814e-05, 'train/grad_norm_pre_clip': 0.03165685385465622, 'timing/data': 0.00033522999729029834, 'timing/model': 0.5693268450122559, 'learning_rate/qwen_vl_interface': 2.168596831440055e-06, 'learning_rate/action_model': 2.1685968314400552e-05, 'epoch': 0.27}) +[2026-10-03 15:19:33,455][starVLA.training.train_starvla][INFO] - Step 3652, Loss: {'train/loss': 3.1515824048256036e-06, 'train/grad_norm_pre_clip': 0.0006930209929123521, 'timing/data': 0.00024246000975836068, 'timing/model': 0.3703733110014582, 'learning_rate/qwen_vl_interface': 2.166279807240553e-06, 'learning_rate/action_model': 2.1662798072405528e-05, 'epoch': 0.27}) +[2026-10-03 15:19:34,111][starVLA.training.train_starvla][INFO] - Step 3653, Loss: {'train/loss': 1.1466012438177131e-05, 'train/grad_norm_pre_clip': 0.0038734336849302053, 'timing/data': 0.00034034000418614596, 'timing/model': 0.36912502101040445, 'learning_rate/qwen_vl_interface': 2.1639640506419394e-06, 'learning_rate/action_model': 2.163964050641939e-05, 'epoch': 0.27}) +[2026-10-03 15:19:34,768][starVLA.training.train_starvla][INFO] - Step 3654, Loss: {'train/loss': 0.6905534863471985, 'train/grad_norm_pre_clip': 98.73600769042969, 'timing/data': 0.00025335000827908516, 'timing/model': 0.3687927410064731, 'learning_rate/qwen_vl_interface': 2.161649562596135e-06, 'learning_rate/action_model': 2.161649562596135e-05, 'epoch': 0.27}) +[2026-10-03 15:19:35,403][starVLA.training.train_starvla][INFO] - Step 3655, Loss: {'train/loss': 4.0291222831001505e-05, 'train/grad_norm_pre_clip': 0.016493769362568855, 'timing/data': 0.00022000999888405204, 'timing/model': 0.3689441809983691, 'learning_rate/qwen_vl_interface': 2.1593363440545383e-06, 'learning_rate/action_model': 2.1593363440545384e-05, 'epoch': 0.27}) +[2026-10-03 15:19:36,061][starVLA.training.train_starvla][INFO] - Step 3656, Loss: {'train/loss': 0.2773607671260834, 'train/grad_norm_pre_clip': 35.08298110961914, 'timing/data': 0.00032217999978456646, 'timing/model': 0.3698472100077197, 'learning_rate/qwen_vl_interface': 2.157024395968025e-06, 'learning_rate/action_model': 2.1570243959680252e-05, 'epoch': 0.27}) +[2026-10-03 15:19:36,915][starVLA.training.train_starvla][INFO] - Step 3657, Loss: {'train/loss': 0.292817085981369, 'train/grad_norm_pre_clip': 108.74014282226562, 'timing/data': 0.00029561998962890357, 'timing/model': 0.3702566809952259, 'learning_rate/qwen_vl_interface': 2.154713719286952e-06, 'learning_rate/action_model': 2.154713719286952e-05, 'epoch': 0.27}) +[2026-10-03 15:19:37,572][starVLA.training.train_starvla][INFO] - Step 3658, Loss: {'train/loss': 0.09578953683376312, 'train/grad_norm_pre_clip': 175.43496704101562, 'timing/data': 0.0003258900105720386, 'timing/model': 0.36888769100187346, 'learning_rate/qwen_vl_interface': 2.152404314961149e-06, 'learning_rate/action_model': 2.152404314961149e-05, 'epoch': 0.27}) +[2026-10-03 15:19:38,229][starVLA.training.train_starvla][INFO] - Step 3659, Loss: {'train/loss': 3.7252652873576153e-06, 'train/grad_norm_pre_clip': 0.0006619427585974336, 'timing/data': 0.0002611099916975945, 'timing/model': 0.3689067410014104, 'learning_rate/qwen_vl_interface': 2.1500961839399264e-06, 'learning_rate/action_model': 2.1500961839399262e-05, 'epoch': 0.27}) +[2026-10-03 15:19:38,885][starVLA.training.train_starvla][INFO] - Step 3660, Loss: {'train/loss': 0.0004949428839609027, 'train/grad_norm_pre_clip': 0.946405291557312, 'timing/data': 0.00028747000033035874, 'timing/model': 0.3692972800054122, 'learning_rate/qwen_vl_interface': 2.1477893271720683e-06, 'learning_rate/action_model': 2.1477893271720682e-05, 'epoch': 0.27}) +[2026-10-03 15:19:39,541][starVLA.training.train_starvla][INFO] - Step 3661, Loss: {'train/loss': 0.000540130480658263, 'train/grad_norm_pre_clip': 1.0441306829452515, 'timing/data': 0.00031257000227924436, 'timing/model': 0.36915179099014495, 'learning_rate/qwen_vl_interface': 2.1454837456058385e-06, 'learning_rate/action_model': 2.1454837456058386e-05, 'epoch': 0.27}) +[2026-10-03 15:19:40,398][starVLA.training.train_starvla][INFO] - Step 3662, Loss: {'train/loss': 8.134684321703389e-05, 'train/grad_norm_pre_clip': 0.08653691411018372, 'timing/data': 0.00024161000328604132, 'timing/model': 0.5686688940040767, 'learning_rate/qwen_vl_interface': 2.1431794401889722e-06, 'learning_rate/action_model': 2.143179440188972e-05, 'epoch': 0.27}) +[2026-10-03 15:19:41,055][starVLA.training.train_starvla][INFO] - Step 3663, Loss: {'train/loss': 4.373713818495162e-05, 'train/grad_norm_pre_clip': 0.0241154246032238, 'timing/data': 0.00030177000735420734, 'timing/model': 0.3690356010047253, 'learning_rate/qwen_vl_interface': 2.1408764118686837e-06, 'learning_rate/action_model': 2.1408764118686838e-05, 'epoch': 0.27}) +[2026-10-03 15:19:41,711][starVLA.training.train_starvla][INFO] - Step 3664, Loss: {'train/loss': 7.083676609909162e-05, 'train/grad_norm_pre_clip': 0.03986414521932602, 'timing/data': 0.00022023000929038972, 'timing/model': 0.3693088500003796, 'learning_rate/qwen_vl_interface': 2.1385746615916598e-06, 'learning_rate/action_model': 2.1385746615916596e-05, 'epoch': 0.27}) +[2026-10-03 15:19:42,368][starVLA.training.train_starvla][INFO] - Step 3665, Loss: {'train/loss': 1.0050607670564204e-05, 'train/grad_norm_pre_clip': 0.003863279242068529, 'timing/data': 0.000245919989538379, 'timing/model': 0.3692983109940542, 'learning_rate/qwen_vl_interface': 2.1362741903040636e-06, 'learning_rate/action_model': 2.1362741903040633e-05, 'epoch': 0.27}) +[2026-10-03 15:19:43,025][starVLA.training.train_starvla][INFO] - Step 3666, Loss: {'train/loss': 3.477639256743714e-05, 'train/grad_norm_pre_clip': 0.010518468916416168, 'timing/data': 0.00033846999576780945, 'timing/model': 0.36947410099674016, 'learning_rate/qwen_vl_interface': 2.1339749989515318e-06, 'learning_rate/action_model': 2.1339749989515316e-05, 'epoch': 0.27}) +[2026-10-03 15:19:43,880][starVLA.training.train_starvla][INFO] - Step 3667, Loss: {'train/loss': 1.4014051885169465e-05, 'train/grad_norm_pre_clip': 0.005975346080958843, 'timing/data': 0.0003403000009711832, 'timing/model': 0.5661041650018888, 'learning_rate/qwen_vl_interface': 2.131677088479175e-06, 'learning_rate/action_model': 2.1316770884791752e-05, 'epoch': 0.27}) +[2026-10-03 15:19:44,537][starVLA.training.train_starvla][INFO] - Step 3668, Loss: {'train/loss': 2.1955647753202356e-05, 'train/grad_norm_pre_clip': 0.009607529267668724, 'timing/data': 0.00020759999461006373, 'timing/model': 0.37007218999497127, 'learning_rate/qwen_vl_interface': 2.1293804598315787e-06, 'learning_rate/action_model': 2.1293804598315787e-05, 'epoch': 0.27}) +[2026-10-03 15:19:45,194][starVLA.training.train_starvla][INFO] - Step 3669, Loss: {'train/loss': 1.5756988432258368e-05, 'train/grad_norm_pre_clip': 0.014671430923044682, 'timing/data': 0.00021189000108279288, 'timing/model': 0.3698114400031045, 'learning_rate/qwen_vl_interface': 2.1270851139527986e-06, 'learning_rate/action_model': 2.1270851139527983e-05, 'epoch': 0.27}) +[2026-10-03 15:19:45,852][starVLA.training.train_starvla][INFO] - Step 3670, Loss: {'train/loss': 0.00043603649828583, 'train/grad_norm_pre_clip': 0.3719465732574463, 'timing/data': 0.0002816190099110827, 'timing/model': 0.36937858200690243, 'learning_rate/qwen_vl_interface': 2.1247910517863653e-06, 'learning_rate/action_model': 2.124791051786365e-05, 'epoch': 0.27}) +[2026-10-03 15:19:46,511][starVLA.training.train_starvla][INFO] - Step 3671, Loss: {'train/loss': 0.0001666962925810367, 'train/grad_norm_pre_clip': 0.17659629881381989, 'timing/data': 0.0002949099871329963, 'timing/model': 0.3703242110059364, 'learning_rate/qwen_vl_interface': 2.1224982742752814e-06, 'learning_rate/action_model': 2.1224982742752812e-05, 'epoch': 0.27}) +[2026-10-03 15:19:47,146][starVLA.training.train_starvla][INFO] - Step 3672, Loss: {'train/loss': 1.2084388799848966e-05, 'train/grad_norm_pre_clip': 0.00507020577788353, 'timing/data': 0.00031042000045999885, 'timing/model': 0.34723774300073273, 'learning_rate/qwen_vl_interface': 2.1202067823620213e-06, 'learning_rate/action_model': 2.1202067823620212e-05, 'epoch': 0.27}) +[2026-10-03 15:19:47,999][starVLA.training.train_starvla][INFO] - Step 3673, Loss: {'train/loss': 1.601838448550552e-05, 'train/grad_norm_pre_clip': 0.005774488672614098, 'timing/data': 0.00032252000528387725, 'timing/model': 0.3693958410003688, 'learning_rate/qwen_vl_interface': 2.117916576988531e-06, 'learning_rate/action_model': 2.117916576988531e-05, 'epoch': 0.27}) +[2026-10-03 15:19:48,657][starVLA.training.train_starvla][INFO] - Step 3674, Loss: {'train/loss': 9.2087884695502e-06, 'train/grad_norm_pre_clip': 0.0030673001892864704, 'timing/data': 0.0002521000133128837, 'timing/model': 0.3692339809931582, 'learning_rate/qwen_vl_interface': 2.1156276590962257e-06, 'learning_rate/action_model': 2.1156276590962257e-05, 'epoch': 0.27}) +[2026-10-03 15:19:49,314][starVLA.training.train_starvla][INFO] - Step 3675, Loss: {'train/loss': 6.586253675777698e-06, 'train/grad_norm_pre_clip': 0.0019615169148892164, 'timing/data': 0.0003415500104892999, 'timing/model': 0.36900774099922273, 'learning_rate/qwen_vl_interface': 2.113340029625994e-06, 'learning_rate/action_model': 2.1133400296259938e-05, 'epoch': 0.27}) +[2026-10-03 15:19:49,971][starVLA.training.train_starvla][INFO] - Step 3676, Loss: {'train/loss': 7.852619091863744e-06, 'train/grad_norm_pre_clip': 0.0029090875759720802, 'timing/data': 0.0002459599927533418, 'timing/model': 0.36946725098823663, 'learning_rate/qwen_vl_interface': 2.111053689518194e-06, 'learning_rate/action_model': 2.111053689518194e-05, 'epoch': 0.27}) +[2026-10-03 15:19:50,630][starVLA.training.train_starvla][INFO] - Step 3677, Loss: {'train/loss': 2.0680890884250402e-05, 'train/grad_norm_pre_clip': 0.014211411587893963, 'timing/data': 0.0003757900121854618, 'timing/model': 0.37083873999654315, 'learning_rate/qwen_vl_interface': 2.1087686397126544e-06, 'learning_rate/action_model': 2.1087686397126545e-05, 'epoch': 0.27}) +[2026-10-03 15:19:51,484][starVLA.training.train_starvla][INFO] - Step 3678, Loss: {'train/loss': 3.126087904092856e-05, 'train/grad_norm_pre_clip': 0.008765296079218388, 'timing/data': 0.0003826000029221177, 'timing/model': 0.37062886999046896, 'learning_rate/qwen_vl_interface': 2.106484881148672e-06, 'learning_rate/action_model': 2.1064848811486718e-05, 'epoch': 0.27}) +[2026-10-03 15:19:52,143][starVLA.training.train_starvla][INFO] - Step 3679, Loss: {'train/loss': 0.3617597818374634, 'train/grad_norm_pre_clip': 77.19220733642578, 'timing/data': 0.00032234999525826424, 'timing/model': 0.370204871011083, 'learning_rate/qwen_vl_interface': 2.1042024147650137e-06, 'learning_rate/action_model': 2.1042024147650137e-05, 'epoch': 0.27}) +[2026-10-03 15:19:52,800][starVLA.training.train_starvla][INFO] - Step 3680, Loss: {'train/loss': 0.004018254578113556, 'train/grad_norm_pre_clip': 1.37043035030365, 'timing/data': 0.00034276899532414973, 'timing/model': 0.3694612820108887, 'learning_rate/qwen_vl_interface': 2.1019212414999163e-06, 'learning_rate/action_model': 2.101921241499916e-05, 'epoch': 0.27}) +[2026-10-03 15:19:53,457][starVLA.training.train_starvla][INFO] - Step 3681, Loss: {'train/loss': 0.00030005729058757424, 'train/grad_norm_pre_clip': 0.19377660751342773, 'timing/data': 0.00020308999228291214, 'timing/model': 0.3698563510115491, 'learning_rate/qwen_vl_interface': 2.099641362291082e-06, 'learning_rate/action_model': 2.099641362291082e-05, 'epoch': 0.27}) +[2026-10-03 15:19:54,092][starVLA.training.train_starvla][INFO] - Step 3682, Loss: {'train/loss': 1.82307976501761e-05, 'train/grad_norm_pre_clip': 0.007877173833549023, 'timing/data': 0.0002864499983843416, 'timing/model': 0.3690564509888645, 'learning_rate/qwen_vl_interface': 2.097362778075685e-06, 'learning_rate/action_model': 2.097362778075685e-05, 'epoch': 0.27}) +[2026-10-03 15:19:54,921][starVLA.training.train_starvla][INFO] - Step 3683, Loss: {'train/loss': 3.515139906085096e-05, 'train/grad_norm_pre_clip': 0.018867727369070053, 'timing/data': 0.0002158999996026978, 'timing/model': 0.5404021090071183, 'learning_rate/qwen_vl_interface': 2.0950854897903663e-06, 'learning_rate/action_model': 2.095085489790366e-05, 'epoch': 0.27}) +[2026-10-03 15:19:55,578][starVLA.training.train_starvla][INFO] - Step 3684, Loss: {'train/loss': 0.10741069167852402, 'train/grad_norm_pre_clip': 60.28004455566406, 'timing/data': 0.00020774001313839108, 'timing/model': 0.3697579610015964, 'learning_rate/qwen_vl_interface': 2.092809498371229e-06, 'learning_rate/action_model': 2.0928094983712292e-05, 'epoch': 0.27}) +[2026-10-03 15:19:56,236][starVLA.training.train_starvla][INFO] - Step 3685, Loss: {'train/loss': 1.3217003470344935e-05, 'train/grad_norm_pre_clip': 0.004481961484998465, 'timing/data': 0.0003085899952566251, 'timing/model': 0.36967456100683194, 'learning_rate/qwen_vl_interface': 2.0905348047538498e-06, 'learning_rate/action_model': 2.0905348047538497e-05, 'epoch': 0.27}) +[2026-10-03 15:19:56,893][starVLA.training.train_starvla][INFO] - Step 3686, Loss: {'train/loss': 0.7434421181678772, 'train/grad_norm_pre_clip': 346.6448059082031, 'timing/data': 0.00025033899873960763, 'timing/model': 0.3687225409958046, 'learning_rate/qwen_vl_interface': 2.0882614098732686e-06, 'learning_rate/action_model': 2.0882614098732685e-05, 'epoch': 0.27}) +[2026-10-03 15:19:57,551][starVLA.training.train_starvla][INFO] - Step 3687, Loss: {'train/loss': 0.0795506089925766, 'train/grad_norm_pre_clip': 180.97592163085938, 'timing/data': 0.0002668800007086247, 'timing/model': 0.36958775100356434, 'learning_rate/qwen_vl_interface': 2.085989314663993e-06, 'learning_rate/action_model': 2.0859893146639927e-05, 'epoch': 0.27}) +[2026-10-03 15:19:58,404][starVLA.training.train_starvla][INFO] - Step 3688, Loss: {'train/loss': 0.45693427324295044, 'train/grad_norm_pre_clip': 34.62347412109375, 'timing/data': 0.00030711998988408595, 'timing/model': 0.5655625850049546, 'learning_rate/qwen_vl_interface': 2.083718520059995e-06, 'learning_rate/action_model': 2.0837185200599946e-05, 'epoch': 0.27}) +[2026-10-03 15:19:59,060][starVLA.training.train_starvla][INFO] - Step 3689, Loss: {'train/loss': 1.3120113180775661e-05, 'train/grad_norm_pre_clip': 0.005023282021284103, 'timing/data': 0.00024096999550238252, 'timing/model': 0.3691973809909541, 'learning_rate/qwen_vl_interface': 2.0814490269947117e-06, 'learning_rate/action_model': 2.0814490269947118e-05, 'epoch': 0.27}) +[2026-10-03 15:19:59,716][starVLA.training.train_starvla][INFO] - Step 3690, Loss: {'train/loss': 0.0019346351036801934, 'train/grad_norm_pre_clip': 0.8744150400161743, 'timing/data': 0.0002042700070887804, 'timing/model': 0.3687311410030816, 'learning_rate/qwen_vl_interface': 2.079180836401047e-06, 'learning_rate/action_model': 2.079180836401047e-05, 'epoch': 0.27}) +[2026-10-03 15:20:00,375][starVLA.training.train_starvla][INFO] - Step 3691, Loss: {'train/loss': 0.20965202152729034, 'train/grad_norm_pre_clip': 40.88508605957031, 'timing/data': 0.00022405000345315784, 'timing/model': 0.37023888099065516, 'learning_rate/qwen_vl_interface': 2.0769139492113687e-06, 'learning_rate/action_model': 2.076913949211369e-05, 'epoch': 0.27}) +[2026-10-03 15:20:01,032][starVLA.training.train_starvla][INFO] - Step 3692, Loss: {'train/loss': 0.21905472874641418, 'train/grad_norm_pre_clip': 121.29882049560547, 'timing/data': 0.00024961000599432737, 'timing/model': 0.3687688009958947, 'learning_rate/qwen_vl_interface': 2.0746483663575075e-06, 'learning_rate/action_model': 2.0746483663575076e-05, 'epoch': 0.27}) +[2026-10-03 15:20:01,689][starVLA.training.train_starvla][INFO] - Step 3693, Loss: {'train/loss': 8.509751933161169e-05, 'train/grad_norm_pre_clip': 0.04084808751940727, 'timing/data': 0.0003170200070599094, 'timing/model': 0.36973764101276174, 'learning_rate/qwen_vl_interface': 2.0723840887707595e-06, 'learning_rate/action_model': 2.0723840887707594e-05, 'epoch': 0.27}) +[2026-10-03 15:20:02,539][starVLA.training.train_starvla][INFO] - Step 3694, Loss: {'train/loss': 0.0002465071447659284, 'train/grad_norm_pre_clip': 0.24064195156097412, 'timing/data': 0.00031988999398890883, 'timing/model': 0.36925117099599447, 'learning_rate/qwen_vl_interface': 2.0701211173818854e-06, 'learning_rate/action_model': 2.0701211173818853e-05, 'epoch': 0.27}) +[2026-10-03 15:20:03,198][starVLA.training.train_starvla][INFO] - Step 3695, Loss: {'train/loss': 7.651616215298418e-06, 'train/grad_norm_pre_clip': 0.001732917851768434, 'timing/data': 0.0002810000005410984, 'timing/model': 0.3696629309997661, 'learning_rate/qwen_vl_interface': 2.0678594531211036e-06, 'learning_rate/action_model': 2.0678594531211033e-05, 'epoch': 0.27}) +[2026-10-03 15:20:03,856][starVLA.training.train_starvla][INFO] - Step 3696, Loss: {'train/loss': 4.1984731069533154e-05, 'train/grad_norm_pre_clip': 0.024716971442103386, 'timing/data': 0.00020988000323995948, 'timing/model': 0.3704621210054029, 'learning_rate/qwen_vl_interface': 2.0655990969181e-06, 'learning_rate/action_model': 2.0655990969181003e-05, 'epoch': 0.27}) +[2026-10-03 15:20:04,513][starVLA.training.train_starvla][INFO] - Step 3697, Loss: {'train/loss': 2.6124975192942657e-05, 'train/grad_norm_pre_clip': 0.016030605882406235, 'timing/data': 0.00026929999876301736, 'timing/model': 0.3691680100018857, 'learning_rate/qwen_vl_interface': 2.0633400497020236e-06, 'learning_rate/action_model': 2.0633400497020232e-05, 'epoch': 0.27}) +[2026-10-03 15:20:05,170][starVLA.training.train_starvla][INFO] - Step 3698, Loss: {'train/loss': 0.1767861247062683, 'train/grad_norm_pre_clip': 89.37773895263672, 'timing/data': 0.00020711999968625605, 'timing/model': 0.36923098101397045, 'learning_rate/qwen_vl_interface': 2.061082312401481e-06, 'learning_rate/action_model': 2.061082312401481e-05, 'epoch': 0.27}) +[2026-10-03 15:20:06,017][starVLA.training.train_starvla][INFO] - Step 3699, Loss: {'train/loss': 9.04487296793377e-06, 'train/grad_norm_pre_clip': 0.0025483390782028437, 'timing/data': 0.00026470000739209354, 'timing/model': 0.5582780550030293, 'learning_rate/qwen_vl_interface': 2.058825885944544e-06, 'learning_rate/action_model': 2.0588258859445442e-05, 'epoch': 0.27}) +[2026-10-03 15:20:06,674][starVLA.training.train_starvla][INFO] - Step 3700, Loss: {'train/loss': 0.034870848059654236, 'train/grad_norm_pre_clip': 48.998573303222656, 'timing/data': 0.00028311999631114304, 'timing/model': 0.36976173099537846, 'learning_rate/qwen_vl_interface': 2.0565707712587447e-06, 'learning_rate/action_model': 2.0565707712587444e-05, 'epoch': 0.27}) +[2026-10-03 15:20:07,308][starVLA.training.train_starvla][INFO] - Step 3701, Loss: {'train/loss': 9.789747309696395e-06, 'train/grad_norm_pre_clip': 0.005347352009266615, 'timing/data': 0.00024855999799910933, 'timing/model': 0.36872992099961266, 'learning_rate/qwen_vl_interface': 2.0543169692710747e-06, 'learning_rate/action_model': 2.054316969271075e-05, 'epoch': 0.27}) +[2026-10-03 15:20:07,965][starVLA.training.train_starvla][INFO] - Step 3702, Loss: {'train/loss': 3.883819590555504e-05, 'train/grad_norm_pre_clip': 0.03434187173843384, 'timing/data': 0.00020388999837450683, 'timing/model': 0.36884531199757475, 'learning_rate/qwen_vl_interface': 2.0520644809079873e-06, 'learning_rate/action_model': 2.0520644809079873e-05, 'epoch': 0.27}) +[2026-10-03 15:20:08,623][starVLA.training.train_starvla][INFO] - Step 3703, Loss: {'train/loss': 3.245253174100071e-05, 'train/grad_norm_pre_clip': 0.015086849220097065, 'timing/data': 0.0003123299975413829, 'timing/model': 0.37045111099723727, 'learning_rate/qwen_vl_interface': 2.049813307095395e-06, 'learning_rate/action_model': 2.0498133070953947e-05, 'epoch': 0.27}) +[2026-10-03 15:20:09,471][starVLA.training.train_starvla][INFO] - Step 3704, Loss: {'train/loss': 2.053303069260437e-05, 'train/grad_norm_pre_clip': 0.006689047906547785, 'timing/data': 0.0003330899926368147, 'timing/model': 0.5598445850046119, 'learning_rate/qwen_vl_interface': 2.047563448758672e-06, 'learning_rate/action_model': 2.0475634487586717e-05, 'epoch': 0.27}) +[2026-10-03 15:20:10,128][starVLA.training.train_starvla][INFO] - Step 3705, Loss: {'train/loss': 1.894532579171937e-05, 'train/grad_norm_pre_clip': 0.007953978143632412, 'timing/data': 0.0003033199900528416, 'timing/model': 0.36969649099046364, 'learning_rate/qwen_vl_interface': 2.045314906822647e-06, 'learning_rate/action_model': 2.0453149068226467e-05, 'epoch': 0.27}) +[2026-10-03 15:20:10,784][starVLA.training.train_starvla][INFO] - Step 3706, Loss: {'train/loss': 0.017208827659487724, 'train/grad_norm_pre_clip': 33.52960968017578, 'timing/data': 0.0002062800049316138, 'timing/model': 0.3684470010048244, 'learning_rate/qwen_vl_interface': 2.0430676822116124e-06, 'learning_rate/action_model': 2.0430676822116123e-05, 'epoch': 0.27}) +[2026-10-03 15:20:11,442][starVLA.training.train_starvla][INFO] - Step 3707, Loss: {'train/loss': 5.766464164480567e-05, 'train/grad_norm_pre_clip': 0.06702844053506851, 'timing/data': 0.0003360400005476549, 'timing/model': 0.3693392420100281, 'learning_rate/qwen_vl_interface': 2.040821775849319e-06, 'learning_rate/action_model': 2.0408217758493188e-05, 'epoch': 0.27}) +[2026-10-03 15:20:12,100][starVLA.training.train_starvla][INFO] - Step 3708, Loss: {'train/loss': 6.0870606830576435e-06, 'train/grad_norm_pre_clip': 0.00119880517013371, 'timing/data': 0.0002109700144501403, 'timing/model': 0.37009583100734744, 'learning_rate/qwen_vl_interface': 2.03857718865897e-06, 'learning_rate/action_model': 2.0385771886589702e-05, 'epoch': 0.27}) +[2026-10-03 15:20:12,757][starVLA.training.train_starvla][INFO] - Step 3709, Loss: {'train/loss': 0.01665504463016987, 'train/grad_norm_pre_clip': 129.64535522460938, 'timing/data': 0.00024027000472415239, 'timing/model': 0.36955881099856924, 'learning_rate/qwen_vl_interface': 2.0363339215632337e-06, 'learning_rate/action_model': 2.0363339215632334e-05, 'epoch': 0.27}) +[2026-10-03 15:20:13,607][starVLA.training.train_starvla][INFO] - Step 3710, Loss: {'train/loss': 0.6690313816070557, 'train/grad_norm_pre_clip': 32.84348678588867, 'timing/data': 0.0002354099997319281, 'timing/model': 0.3691035509982612, 'learning_rate/qwen_vl_interface': 2.034091975484231e-06, 'learning_rate/action_model': 2.0340919754842307e-05, 'epoch': 0.27}) +[2026-10-03 15:20:14,265][starVLA.training.train_starvla][INFO] - Step 3711, Loss: {'train/loss': 4.6850571379764006e-05, 'train/grad_norm_pre_clip': 0.02006586641073227, 'timing/data': 0.0003394499944988638, 'timing/model': 0.36965523099934217, 'learning_rate/qwen_vl_interface': 2.0318513513435415e-06, 'learning_rate/action_model': 2.0318513513435415e-05, 'epoch': 0.27}) +[2026-10-03 15:20:14,923][starVLA.training.train_starvla][INFO] - Step 3712, Loss: {'train/loss': 9.067220162251033e-06, 'train/grad_norm_pre_clip': 0.002792746527120471, 'timing/data': 0.00020917999790981412, 'timing/model': 0.37028606100648176, 'learning_rate/qwen_vl_interface': 2.0296120500622018e-06, 'learning_rate/action_model': 2.0296120500622017e-05, 'epoch': 0.27}) +[2026-10-03 15:20:15,581][starVLA.training.train_starvla][INFO] - Step 3713, Loss: {'train/loss': 5.6421140470774844e-05, 'train/grad_norm_pre_clip': 0.018037492409348488, 'timing/data': 0.00032581000414211303, 'timing/model': 0.3696482310042484, 'learning_rate/qwen_vl_interface': 2.0273740725607034e-06, 'learning_rate/action_model': 2.0273740725607032e-05, 'epoch': 0.27}) +[2026-10-03 15:20:16,237][starVLA.training.train_starvla][INFO] - Step 3714, Loss: {'train/loss': 3.428978379815817e-05, 'train/grad_norm_pre_clip': 0.01873692311346531, 'timing/data': 0.00020537999807856977, 'timing/model': 0.3689456219872227, 'learning_rate/qwen_vl_interface': 2.0251374197589942e-06, 'learning_rate/action_model': 2.0251374197589943e-05, 'epoch': 0.27}) +[2026-10-03 15:20:17,089][starVLA.training.train_starvla][INFO] - Step 3715, Loss: {'train/loss': 0.0012351644691079855, 'train/grad_norm_pre_clip': 0.2599157691001892, 'timing/data': 0.0002766700054053217, 'timing/model': 0.3688307710108347, 'learning_rate/qwen_vl_interface': 2.0229020925764785e-06, 'learning_rate/action_model': 2.0229020925764785e-05, 'epoch': 0.27}) +[2026-10-03 15:20:17,745][starVLA.training.train_starvla][INFO] - Step 3716, Loss: {'train/loss': 0.0018446744652464986, 'train/grad_norm_pre_clip': 0.4564197063446045, 'timing/data': 0.00034470000537112355, 'timing/model': 0.36939689100836404, 'learning_rate/qwen_vl_interface': 2.0206680919320145e-06, 'learning_rate/action_model': 2.020668091932014e-05, 'epoch': 0.27}) +[2026-10-03 15:20:18,402][starVLA.training.train_starvla][INFO] - Step 3717, Loss: {'train/loss': 6.1094424381735735e-06, 'train/grad_norm_pre_clip': 0.0009751867619343102, 'timing/data': 0.00041708999196998775, 'timing/model': 0.3696238609991269, 'learning_rate/qwen_vl_interface': 2.0184354187439164e-06, 'learning_rate/action_model': 2.0184354187439162e-05, 'epoch': 0.27}) +[2026-10-03 15:20:19,059][starVLA.training.train_starvla][INFO] - Step 3718, Loss: {'train/loss': 9.303289698436856e-05, 'train/grad_norm_pre_clip': 0.066989004611969, 'timing/data': 0.00024321000091731548, 'timing/model': 0.36863724200520664, 'learning_rate/qwen_vl_interface': 2.016204073929949e-06, 'learning_rate/action_model': 2.016204073929949e-05, 'epoch': 0.27}) +[2026-10-03 15:20:19,671][starVLA.training.train_starvla][INFO] - Step 3719, Loss: {'train/loss': 0.6034287214279175, 'train/grad_norm_pre_clip': 99.41971588134766, 'timing/data': 0.0003278899966971949, 'timing/model': 0.346156374012935, 'learning_rate/qwen_vl_interface': 2.0139740584073363e-06, 'learning_rate/action_model': 2.0139740584073364e-05, 'epoch': 0.27}) +[2026-10-03 15:20:20,521][starVLA.training.train_starvla][INFO] - Step 3720, Loss: {'train/loss': 1.5027131667011417e-05, 'train/grad_norm_pre_clip': 0.004027752671390772, 'timing/data': 0.0002342100051464513, 'timing/model': 0.5635677649988793, 'learning_rate/qwen_vl_interface': 2.011745373092753e-06, 'learning_rate/action_model': 2.011745373092753e-05, 'epoch': 0.27}) +[2026-10-03 15:20:21,178][starVLA.training.train_starvla][INFO] - Step 3721, Loss: {'train/loss': 0.000302825472317636, 'train/grad_norm_pre_clip': 0.1697365641593933, 'timing/data': 0.0003012500092154369, 'timing/model': 0.36958524100191426, 'learning_rate/qwen_vl_interface': 2.0095180189023267e-06, 'learning_rate/action_model': 2.0095180189023265e-05, 'epoch': 0.27}) +[2026-10-03 15:20:21,835][starVLA.training.train_starvla][INFO] - Step 3722, Loss: {'train/loss': 8.523404176230542e-06, 'train/grad_norm_pre_clip': 0.0015303726540878415, 'timing/data': 0.00020985999435652047, 'timing/model': 0.36908510100329295, 'learning_rate/qwen_vl_interface': 2.0072919967516393e-06, 'learning_rate/action_model': 2.0072919967516392e-05, 'epoch': 0.27}) +[2026-10-03 15:20:22,492][starVLA.training.train_starvla][INFO] - Step 3723, Loss: {'train/loss': 2.1709554857807234e-05, 'train/grad_norm_pre_clip': 0.01314010750502348, 'timing/data': 0.00025627999275457114, 'timing/model': 0.3692222999961814, 'learning_rate/qwen_vl_interface': 2.0050673075557245e-06, 'learning_rate/action_model': 2.0050673075557243e-05, 'epoch': 0.27}) +[2026-10-03 15:20:23,126][starVLA.training.train_starvla][INFO] - Step 3724, Loss: {'train/loss': 2.7407779271015897e-05, 'train/grad_norm_pre_clip': 0.016654744744300842, 'timing/data': 0.0003030090010724962, 'timing/model': 0.3461762840015581, 'learning_rate/qwen_vl_interface': 2.0028439522290677e-06, 'learning_rate/action_model': 2.0028439522290676e-05, 'epoch': 0.27}) +[2026-10-03 15:20:23,782][starVLA.training.train_starvla][INFO] - Step 3725, Loss: {'train/loss': 0.587081789970398, 'train/grad_norm_pre_clip': 50.46061325073242, 'timing/data': 0.00024701000074855983, 'timing/model': 0.3696643699950073, 'learning_rate/qwen_vl_interface': 2.000621931685606e-06, 'learning_rate/action_model': 2.000621931685606e-05, 'epoch': 0.27}) +[2026-10-03 15:20:24,603][starVLA.training.train_starvla][INFO] - Step 3726, Loss: {'train/loss': 0.001609820406883955, 'train/grad_norm_pre_clip': 0.3861258029937744, 'timing/data': 0.0002742500073509291, 'timing/model': 0.36867857100151014, 'learning_rate/qwen_vl_interface': 1.9984012468387285e-06, 'learning_rate/action_model': 1.9984012468387285e-05, 'epoch': 0.27}) +[2026-10-03 15:20:25,260][starVLA.training.train_starvla][INFO] - Step 3727, Loss: {'train/loss': 0.5420093536376953, 'train/grad_norm_pre_clip': 50.37953567504883, 'timing/data': 0.00032761000329628587, 'timing/model': 0.3697579010040499, 'learning_rate/qwen_vl_interface': 1.9961818986012767e-06, 'learning_rate/action_model': 1.9961818986012763e-05, 'epoch': 0.27}) +[2026-10-03 15:20:25,917][starVLA.training.train_starvla][INFO] - Step 3728, Loss: {'train/loss': 4.3741820263676345e-05, 'train/grad_norm_pre_clip': 0.018997009843587875, 'timing/data': 0.00020326000230852515, 'timing/model': 0.3697429500025464, 'learning_rate/qwen_vl_interface': 1.9939638878855372e-06, 'learning_rate/action_model': 1.993963887885537e-05, 'epoch': 0.27}) +[2026-10-03 15:20:26,572][starVLA.training.train_starvla][INFO] - Step 3729, Loss: {'train/loss': 0.00872306153178215, 'train/grad_norm_pre_clip': 14.84766674041748, 'timing/data': 0.0002464300050633028, 'timing/model': 0.3690230009960942, 'learning_rate/qwen_vl_interface': 1.991747215603252e-06, 'learning_rate/action_model': 1.991747215603252e-05, 'epoch': 0.27}) +[2026-10-03 15:20:27,228][starVLA.training.train_starvla][INFO] - Step 3730, Loss: {'train/loss': 4.634236120182322e-06, 'train/grad_norm_pre_clip': 0.0007558813085779548, 'timing/data': 0.00024393000057898462, 'timing/model': 0.3682152809924446, 'learning_rate/qwen_vl_interface': 1.9895318826656116e-06, 'learning_rate/action_model': 1.9895318826656114e-05, 'epoch': 0.27}) +[2026-10-03 15:20:28,078][starVLA.training.train_starvla][INFO] - Step 3731, Loss: {'train/loss': 0.006887013092637062, 'train/grad_norm_pre_clip': 13.385564804077148, 'timing/data': 0.000404159989557229, 'timing/model': 0.36890559099265374, 'learning_rate/qwen_vl_interface': 1.9873178899832553e-06, 'learning_rate/action_model': 1.9873178899832554e-05, 'epoch': 0.27}) +[2026-10-03 15:20:28,735][starVLA.training.train_starvla][INFO] - Step 3732, Loss: {'train/loss': 0.0006669344729743898, 'train/grad_norm_pre_clip': 0.1726883351802826, 'timing/data': 0.00029716901190113276, 'timing/model': 0.370142350991955, 'learning_rate/qwen_vl_interface': 1.985105238466272e-06, 'learning_rate/action_model': 1.9851052384662722e-05, 'epoch': 0.27}) +[2026-10-03 15:20:29,391][starVLA.training.train_starvla][INFO] - Step 3733, Loss: {'train/loss': 0.00153935921844095, 'train/grad_norm_pre_clip': 1.9245494604110718, 'timing/data': 0.00025483999343123287, 'timing/model': 0.36883204100013245, 'learning_rate/qwen_vl_interface': 1.9828939290242e-06, 'learning_rate/action_model': 1.9828939290241998e-05, 'epoch': 0.27}) +[2026-10-03 15:20:30,047][starVLA.training.train_starvla][INFO] - Step 3734, Loss: {'train/loss': 1.0886778831481934, 'train/grad_norm_pre_clip': 137.2113494873047, 'timing/data': 0.00036145000194665045, 'timing/model': 0.36858658099663444, 'learning_rate/qwen_vl_interface': 1.980683962566024e-06, 'learning_rate/action_model': 1.980683962566024e-05, 'epoch': 0.27}) +[2026-10-03 15:20:30,682][starVLA.training.train_starvla][INFO] - Step 3735, Loss: {'train/loss': 5.363764284993522e-05, 'train/grad_norm_pre_clip': 0.01788940280675888, 'timing/data': 0.00028264000138733536, 'timing/model': 0.3696342309995089, 'learning_rate/qwen_vl_interface': 1.9784753400001784e-06, 'learning_rate/action_model': 1.978475340000178e-05, 'epoch': 0.27}) +[2026-10-03 15:20:31,535][starVLA.training.train_starvla][INFO] - Step 3736, Loss: {'train/loss': 0.3663804531097412, 'train/grad_norm_pre_clip': 53.42009735107422, 'timing/data': 0.00036777000059373677, 'timing/model': 0.37047960099880584, 'learning_rate/qwen_vl_interface': 1.9762680622345443e-06, 'learning_rate/action_model': 1.9762680622345442e-05, 'epoch': 0.27}) +[2026-10-03 15:20:32,193][starVLA.training.train_starvla][INFO] - Step 3737, Loss: {'train/loss': 7.202180859167129e-05, 'train/grad_norm_pre_clip': 0.02921183407306671, 'timing/data': 0.00023936000070534647, 'timing/model': 0.3697031109913951, 'learning_rate/qwen_vl_interface': 1.9740621301764513e-06, 'learning_rate/action_model': 1.9740621301764513e-05, 'epoch': 0.27}) +[2026-10-03 15:20:32,850][starVLA.training.train_starvla][INFO] - Step 3738, Loss: {'train/loss': 0.0002229449455626309, 'train/grad_norm_pre_clip': 0.09775487333536148, 'timing/data': 0.00028508999093901366, 'timing/model': 0.36930888101051096, 'learning_rate/qwen_vl_interface': 1.971857544732672e-06, 'learning_rate/action_model': 1.971857544732672e-05, 'epoch': 0.27}) +[2026-10-03 15:20:33,508][starVLA.training.train_starvla][INFO] - Step 3739, Loss: {'train/loss': 4.8163456085603684e-05, 'train/grad_norm_pre_clip': 0.016100477427244186, 'timing/data': 0.0002641199971549213, 'timing/model': 0.36905246099922806, 'learning_rate/qwen_vl_interface': 1.96965430680943e-06, 'learning_rate/action_model': 1.96965430680943e-05, 'epoch': 0.27}) +[2026-10-03 15:20:34,164][starVLA.training.train_starvla][INFO] - Step 3740, Loss: {'train/loss': 0.00017117042443715036, 'train/grad_norm_pre_clip': 0.0664968341588974, 'timing/data': 0.00029719999292865396, 'timing/model': 0.36930168999242596, 'learning_rate/qwen_vl_interface': 1.967452417312393e-06, 'learning_rate/action_model': 1.9674524173123927e-05, 'epoch': 0.27}) +[2026-10-03 15:20:35,013][starVLA.training.train_starvla][INFO] - Step 3741, Loss: {'train/loss': 3.359031688887626e-05, 'train/grad_norm_pre_clip': 0.013667909428477287, 'timing/data': 0.000245919989538379, 'timing/model': 0.561950795003213, 'learning_rate/qwen_vl_interface': 1.9652518771466746e-06, 'learning_rate/action_model': 1.9652518771466747e-05, 'epoch': 0.27}) +[2026-10-03 15:20:35,669][starVLA.training.train_starvla][INFO] - Step 3742, Loss: {'train/loss': 5.963336298009381e-05, 'train/grad_norm_pre_clip': 0.036333177238702774, 'timing/data': 0.00024259999918285757, 'timing/model': 0.3687546410073992, 'learning_rate/qwen_vl_interface': 1.963052687216834e-06, 'learning_rate/action_model': 1.963052687216834e-05, 'epoch': 0.27}) +[2026-10-03 15:20:36,326][starVLA.training.train_starvla][INFO] - Step 3743, Loss: {'train/loss': 0.0004452780995052308, 'train/grad_norm_pre_clip': 0.1886850893497467, 'timing/data': 0.00025753999943844974, 'timing/model': 0.36920101099531166, 'learning_rate/qwen_vl_interface': 1.960854848426874e-06, 'learning_rate/action_model': 1.960854848426874e-05, 'epoch': 0.27}) +[2026-10-03 15:20:36,983][starVLA.training.train_starvla][INFO] - Step 3744, Loss: {'train/loss': 0.0015264921821653843, 'train/grad_norm_pre_clip': 2.079622268676758, 'timing/data': 0.0002873899939004332, 'timing/model': 0.36890716099878773, 'learning_rate/qwen_vl_interface': 1.9586583616802446e-06, 'learning_rate/action_model': 1.9586583616802446e-05, 'epoch': 0.27}) +[2026-10-03 15:20:37,639][starVLA.training.train_starvla][INFO] - Step 3745, Loss: {'train/loss': 9.735903586260974e-05, 'train/grad_norm_pre_clip': 0.05563440918922424, 'timing/data': 0.00025332999939564615, 'timing/model': 0.36919607099844143, 'learning_rate/qwen_vl_interface': 1.956463227879838e-06, 'learning_rate/action_model': 1.9564632278798377e-05, 'epoch': 0.27}) +[2026-10-03 15:20:38,491][starVLA.training.train_starvla][INFO] - Step 3746, Loss: {'train/loss': 0.17269763350486755, 'train/grad_norm_pre_clip': 91.93831634521484, 'timing/data': 0.00024468000628985465, 'timing/model': 0.5642903650004882, 'learning_rate/qwen_vl_interface': 1.95426944792799e-06, 'learning_rate/action_model': 1.95426944792799e-05, 'epoch': 0.27}) +[2026-10-03 15:20:39,148][starVLA.training.train_starvla][INFO] - Step 3747, Loss: {'train/loss': 0.00088094687089324, 'train/grad_norm_pre_clip': 0.8400874733924866, 'timing/data': 0.00032099000236485153, 'timing/model': 0.3691569610091392, 'learning_rate/qwen_vl_interface': 1.9520770227264817e-06, 'learning_rate/action_model': 1.9520770227264817e-05, 'epoch': 0.27}) +[2026-10-03 15:20:39,805][starVLA.training.train_starvla][INFO] - Step 3748, Loss: {'train/loss': 7.262783765327185e-05, 'train/grad_norm_pre_clip': 0.04631682112812996, 'timing/data': 0.0003265699924668297, 'timing/model': 0.3702939310023794, 'learning_rate/qwen_vl_interface': 1.9498859531765373e-06, 'learning_rate/action_model': 1.9498859531765372e-05, 'epoch': 0.27}) +[2026-10-03 15:20:40,461][starVLA.training.train_starvla][INFO] - Step 3749, Loss: {'train/loss': 0.6198835372924805, 'train/grad_norm_pre_clip': 83.22837829589844, 'timing/data': 0.00024307999410666525, 'timing/model': 0.3691493009973783, 'learning_rate/qwen_vl_interface': 1.9476962401788204e-06, 'learning_rate/action_model': 1.9476962401788204e-05, 'epoch': 0.27}) +[2026-10-03 15:20:41,118][starVLA.training.train_starvla][INFO] - Step 3750, Loss: {'train/loss': 7.695767999393865e-05, 'train/grad_norm_pre_clip': 0.038744013756513596, 'timing/data': 0.00022215000353753567, 'timing/model': 0.3686708710010862, 'learning_rate/qwen_vl_interface': 1.9455078846334397e-06, 'learning_rate/action_model': 1.9455078846334395e-05, 'epoch': 0.27}) +[2026-10-03 15:20:41,752][starVLA.training.train_starvla][INFO] - Step 3751, Loss: {'train/loss': 0.00022339218412525952, 'train/grad_norm_pre_clip': 0.11772958189249039, 'timing/data': 0.0003258800134062767, 'timing/model': 0.36891589099832345, 'learning_rate/qwen_vl_interface': 1.9433208874399465e-06, 'learning_rate/action_model': 1.9433208874399464e-05, 'epoch': 0.27}) +[2026-10-03 15:20:42,605][starVLA.training.train_starvla][INFO] - Step 3752, Loss: {'train/loss': 1.592133594385814e-05, 'train/grad_norm_pre_clip': 0.004452876280993223, 'timing/data': 0.000258959989878349, 'timing/model': 0.3698605310055427, 'learning_rate/qwen_vl_interface': 1.9411352494973314e-06, 'learning_rate/action_model': 1.9411352494973315e-05, 'epoch': 0.27}) +[2026-10-03 15:20:43,262][starVLA.training.train_starvla][INFO] - Step 3753, Loss: {'train/loss': 6.478848808910698e-05, 'train/grad_norm_pre_clip': 0.02434026263654232, 'timing/data': 0.00024385999131482095, 'timing/model': 0.36886687100923155, 'learning_rate/qwen_vl_interface': 1.938950971704031e-06, 'learning_rate/action_model': 1.9389509717040308e-05, 'epoch': 0.27}) +[2026-10-03 15:20:43,918][starVLA.training.train_starvla][INFO] - Step 3754, Loss: {'train/loss': 1.2919131222588476e-05, 'train/grad_norm_pre_clip': 0.003134000115096569, 'timing/data': 0.00020357000175863504, 'timing/model': 0.3687720409943722, 'learning_rate/qwen_vl_interface': 1.936768054957915e-06, 'learning_rate/action_model': 1.9367680549579147e-05, 'epoch': 0.27}) +[2026-10-03 15:20:44,554][starVLA.training.train_starvla][INFO] - Step 3755, Loss: {'train/loss': 0.20673610270023346, 'train/grad_norm_pre_clip': 108.33261108398438, 'timing/data': 0.0003497800062177703, 'timing/model': 0.34676688400213607, 'learning_rate/qwen_vl_interface': 1.9345865001562994e-06, 'learning_rate/action_model': 1.9345865001562994e-05, 'epoch': 0.27}) +[2026-10-03 15:20:45,209][starVLA.training.train_starvla][INFO] - Step 3756, Loss: {'train/loss': 6.794567161705345e-05, 'train/grad_norm_pre_clip': 0.03544018417596817, 'timing/data': 0.0002469099999871105, 'timing/model': 0.368743701998028, 'learning_rate/qwen_vl_interface': 1.9324063081959394e-06, 'learning_rate/action_model': 1.9324063081959395e-05, 'epoch': 0.27}) +[2026-10-03 15:20:46,063][starVLA.training.train_starvla][INFO] - Step 3757, Loss: {'train/loss': 2.1293024474289268e-05, 'train/grad_norm_pre_clip': 0.006860360503196716, 'timing/data': 0.00023818000045139343, 'timing/model': 0.5658508140040794, 'learning_rate/qwen_vl_interface': 1.9302274799730304e-06, 'learning_rate/action_model': 1.9302274799730306e-05, 'epoch': 0.27}) +[2026-10-03 15:20:46,697][starVLA.training.train_starvla][INFO] - Step 3758, Loss: {'train/loss': 0.005374111235141754, 'train/grad_norm_pre_clip': 15.240033149719238, 'timing/data': 0.0002499299880582839, 'timing/model': 0.36954917100956663, 'learning_rate/qwen_vl_interface': 1.9280500163832055e-06, 'learning_rate/action_model': 1.9280500163832055e-05, 'epoch': 0.28}) +[2026-10-03 15:20:47,355][starVLA.training.train_starvla][INFO] - Step 3759, Loss: {'train/loss': 0.0005869233864359558, 'train/grad_norm_pre_clip': 0.739254355430603, 'timing/data': 0.00028308999026194215, 'timing/model': 0.3691903610015288, 'learning_rate/qwen_vl_interface': 1.925873918321537e-06, 'learning_rate/action_model': 1.925873918321537e-05, 'epoch': 0.28}) +[2026-10-03 15:20:48,011][starVLA.training.train_starvla][INFO] - Step 3760, Loss: {'train/loss': 6.35485484963283e-05, 'train/grad_norm_pre_clip': 0.032276254147291183, 'timing/data': 0.00030870999034959823, 'timing/model': 0.36877866099530365, 'learning_rate/qwen_vl_interface': 1.92369918668254e-06, 'learning_rate/action_model': 1.92369918668254e-05, 'epoch': 0.28}) +[2026-10-03 15:20:48,668][starVLA.training.train_starvla][INFO] - Step 3761, Loss: {'train/loss': 0.0004817806766368449, 'train/grad_norm_pre_clip': 0.5559629797935486, 'timing/data': 0.0003141000051982701, 'timing/model': 0.3691314310126472, 'learning_rate/qwen_vl_interface': 1.9215258223601605e-06, 'learning_rate/action_model': 1.9215258223601605e-05, 'epoch': 0.28}) +[2026-10-03 15:20:49,324][starVLA.training.train_starvla][INFO] - Step 3762, Loss: {'train/loss': 0.2690422534942627, 'train/grad_norm_pre_clip': 249.05860900878906, 'timing/data': 0.00028804999601561576, 'timing/model': 0.3683321610005805, 'learning_rate/qwen_vl_interface': 1.9193538262477874e-06, 'learning_rate/action_model': 1.9193538262477874e-05, 'epoch': 0.28}) +[2026-10-03 15:20:50,172][starVLA.training.train_starvla][INFO] - Step 3763, Loss: {'train/loss': 2.9778058888041414e-05, 'train/grad_norm_pre_clip': 0.01043740939348936, 'timing/data': 0.00032279999868478626, 'timing/model': 0.36901602099533193, 'learning_rate/qwen_vl_interface': 1.917183199238247e-06, 'learning_rate/action_model': 1.917183199238247e-05, 'epoch': 0.28}) +[2026-10-03 15:20:50,831][starVLA.training.train_starvla][INFO] - Step 3764, Loss: {'train/loss': 0.17337533831596375, 'train/grad_norm_pre_clip': 233.94956970214844, 'timing/data': 0.0003036499983863905, 'timing/model': 0.3698183309898013, 'learning_rate/qwen_vl_interface': 1.9150139422238016e-06, 'learning_rate/action_model': 1.9150139422238018e-05, 'epoch': 0.28}) +[2026-10-03 15:20:51,488][starVLA.training.train_starvla][INFO] - Step 3765, Loss: {'train/loss': 0.00014314551663119346, 'train/grad_norm_pre_clip': 0.1564285308122635, 'timing/data': 0.00020403000235091895, 'timing/model': 0.37009401099930983, 'learning_rate/qwen_vl_interface': 1.9128460560961513e-06, 'learning_rate/action_model': 1.9128460560961514e-05, 'epoch': 0.28}) +[2026-10-03 15:20:52,122][starVLA.training.train_starvla][INFO] - Step 3766, Loss: {'train/loss': 0.765373706817627, 'train/grad_norm_pre_clip': 30.952856063842773, 'timing/data': 0.00023348000831902027, 'timing/model': 0.3692622010130435, 'learning_rate/qwen_vl_interface': 1.910679541746432e-06, 'learning_rate/action_model': 1.9106795417464323e-05, 'epoch': 0.28}) +[2026-10-03 15:20:52,780][starVLA.training.train_starvla][INFO] - Step 3767, Loss: {'train/loss': 9.871934707916807e-06, 'train/grad_norm_pre_clip': 0.0020221734885126352, 'timing/data': 0.0003395200037630275, 'timing/model': 0.36926904099527746, 'learning_rate/qwen_vl_interface': 1.9085144000652166e-06, 'learning_rate/action_model': 1.9085144000652166e-05, 'epoch': 0.28}) +[2026-10-03 15:20:53,602][starVLA.training.train_starvla][INFO] - Step 3768, Loss: {'train/loss': 7.61411283747293e-05, 'train/grad_norm_pre_clip': 0.05863989144563675, 'timing/data': 0.00028664999990724027, 'timing/model': 0.5353844189958181, 'learning_rate/qwen_vl_interface': 1.9063506319425116e-06, 'learning_rate/action_model': 1.9063506319425114e-05, 'epoch': 0.28}) +[2026-10-03 15:20:54,259][starVLA.training.train_starvla][INFO] - Step 3769, Loss: {'train/loss': 0.0007230282644741237, 'train/grad_norm_pre_clip': 0.5849029421806335, 'timing/data': 0.00030005000007804483, 'timing/model': 0.3693446309916908, 'learning_rate/qwen_vl_interface': 1.9041882382677613e-06, 'learning_rate/action_model': 1.9041882382677612e-05, 'epoch': 0.28}) +[2026-10-03 15:20:54,917][starVLA.training.train_starvla][INFO] - Step 3770, Loss: {'train/loss': 4.255483509041369e-05, 'train/grad_norm_pre_clip': 0.02664550207555294, 'timing/data': 0.00024371000472456217, 'timing/model': 0.36912576100439765, 'learning_rate/qwen_vl_interface': 1.9020272199298452e-06, 'learning_rate/action_model': 1.902027219929845e-05, 'epoch': 0.28}) +[2026-10-03 15:20:55,574][starVLA.training.train_starvla][INFO] - Step 3771, Loss: {'train/loss': 0.00140154839027673, 'train/grad_norm_pre_clip': 2.9373414516448975, 'timing/data': 0.00027052999939769506, 'timing/model': 0.3690931309974985, 'learning_rate/qwen_vl_interface': 1.8998675778170736e-06, 'learning_rate/action_model': 1.8998675778170734e-05, 'epoch': 0.28}) +[2026-10-03 15:20:56,231][starVLA.training.train_starvla][INFO] - Step 3772, Loss: {'train/loss': 0.0004632288764696568, 'train/grad_norm_pre_clip': 0.4882568418979645, 'timing/data': 0.00025742000434547663, 'timing/model': 0.36974201099656057, 'learning_rate/qwen_vl_interface': 1.8977093128171955e-06, 'learning_rate/action_model': 1.8977093128171954e-05, 'epoch': 0.28}) +[2026-10-03 15:20:57,060][starVLA.training.train_starvla][INFO] - Step 3773, Loss: {'train/loss': 0.014289041049778461, 'train/grad_norm_pre_clip': 10.768841743469238, 'timing/data': 0.0002575899998191744, 'timing/model': 0.5636587859917199, 'learning_rate/qwen_vl_interface': 1.8955524258173919e-06, 'learning_rate/action_model': 1.895552425817392e-05, 'epoch': 0.28}) +[2026-10-03 15:20:57,717][starVLA.training.train_starvla][INFO] - Step 3774, Loss: {'train/loss': 2.2909563995199278e-05, 'train/grad_norm_pre_clip': 0.007712545804679394, 'timing/data': 0.0003652000013971701, 'timing/model': 0.3685439109976869, 'learning_rate/qwen_vl_interface': 1.8933969177042769e-06, 'learning_rate/action_model': 1.8933969177042768e-05, 'epoch': 0.28}) +[2026-10-03 15:20:58,375][starVLA.training.train_starvla][INFO] - Step 3775, Loss: {'train/loss': 0.000380926881916821, 'train/grad_norm_pre_clip': 0.17161805927753448, 'timing/data': 0.00025545999233145267, 'timing/model': 0.3694225910003297, 'learning_rate/qwen_vl_interface': 1.8912427893638996e-06, 'learning_rate/action_model': 1.8912427893638996e-05, 'epoch': 0.28}) +[2026-10-03 15:20:59,032][starVLA.training.train_starvla][INFO] - Step 3776, Loss: {'train/loss': 3.109670797130093e-05, 'train/grad_norm_pre_clip': 0.009586608968675137, 'timing/data': 0.00020156000391580164, 'timing/model': 0.36993257100402843, 'learning_rate/qwen_vl_interface': 1.8890900416817398e-06, 'learning_rate/action_model': 1.8890900416817397e-05, 'epoch': 0.28}) +[2026-10-03 15:20:59,689][starVLA.training.train_starvla][INFO] - Step 3777, Loss: {'train/loss': 0.006388901267200708, 'train/grad_norm_pre_clip': 19.607980728149414, 'timing/data': 0.00021674999152310193, 'timing/model': 0.3699540609959513, 'learning_rate/qwen_vl_interface': 1.8869386755427117e-06, 'learning_rate/action_model': 1.8869386755427114e-05, 'epoch': 0.28}) +[2026-10-03 15:21:00,346][starVLA.training.train_starvla][INFO] - Step 3778, Loss: {'train/loss': 0.00011006614658981562, 'train/grad_norm_pre_clip': 0.055085189640522, 'timing/data': 0.00028855999698862433, 'timing/model': 0.3689589209971018, 'learning_rate/qwen_vl_interface': 1.8847886918311597e-06, 'learning_rate/action_model': 1.8847886918311598e-05, 'epoch': 0.28}) +[2026-10-03 15:21:01,200][starVLA.training.train_starvla][INFO] - Step 3779, Loss: {'train/loss': 1.6971907825791277e-05, 'train/grad_norm_pre_clip': 0.005220317747443914, 'timing/data': 0.00026513999910093844, 'timing/model': 0.369364571000915, 'learning_rate/qwen_vl_interface': 1.8826400914308616e-06, 'learning_rate/action_model': 1.8826400914308613e-05, 'epoch': 0.28}) +[2026-10-03 15:21:01,857][starVLA.training.train_starvla][INFO] - Step 3780, Loss: {'train/loss': 1.1145935786771588e-05, 'train/grad_norm_pre_clip': 0.002186209661886096, 'timing/data': 0.000249469987466, 'timing/model': 0.369829250004841, 'learning_rate/qwen_vl_interface': 1.880492875225025e-06, 'learning_rate/action_model': 1.8804928752250252e-05, 'epoch': 0.28}) +[2026-10-03 15:21:02,515][starVLA.training.train_starvla][INFO] - Step 3781, Loss: {'train/loss': 0.000752605963498354, 'train/grad_norm_pre_clip': 2.145660638809204, 'timing/data': 0.00022828001237940043, 'timing/model': 0.3703767209954094, 'learning_rate/qwen_vl_interface': 1.8783470440962922e-06, 'learning_rate/action_model': 1.878347044096292e-05, 'epoch': 0.28}) +[2026-10-03 15:21:03,171][starVLA.training.train_starvla][INFO] - Step 3782, Loss: {'train/loss': 0.00016065177624113858, 'train/grad_norm_pre_clip': 0.1097385436296463, 'timing/data': 0.00030403998971451074, 'timing/model': 0.3688065009919228, 'learning_rate/qwen_vl_interface': 1.8762025989267293e-06, 'learning_rate/action_model': 1.876202598926729e-05, 'epoch': 0.28}) +[2026-10-03 15:21:03,807][starVLA.training.train_starvla][INFO] - Step 3783, Loss: {'train/loss': 1.0073092198581435e-05, 'train/grad_norm_pre_clip': 0.002094195457175374, 'timing/data': 0.00034822999441530555, 'timing/model': 0.3464611439994769, 'learning_rate/qwen_vl_interface': 1.874059540597839e-06, 'learning_rate/action_model': 1.874059540597839e-05, 'epoch': 0.28}) +[2026-10-03 15:21:04,652][starVLA.training.train_starvla][INFO] - Step 3784, Loss: {'train/loss': 0.03028699941933155, 'train/grad_norm_pre_clip': 143.29981994628906, 'timing/data': 0.00036292998993303627, 'timing/model': 0.556772826006636, 'learning_rate/qwen_vl_interface': 1.8719178699905508e-06, 'learning_rate/action_model': 1.8719178699905508e-05, 'epoch': 0.28}) +[2026-10-03 15:21:05,308][starVLA.training.train_starvla][INFO] - Step 3785, Loss: {'train/loss': 0.00018536680727265775, 'train/grad_norm_pre_clip': 0.1257912963628769, 'timing/data': 0.00021027000911999494, 'timing/model': 0.3693963010009611, 'learning_rate/qwen_vl_interface': 1.8697775879852242e-06, 'learning_rate/action_model': 1.8697775879852242e-05, 'epoch': 0.28}) +[2026-10-03 15:21:05,970][starVLA.training.train_starvla][INFO] - Step 3786, Loss: {'train/loss': 0.0004292741359677166, 'train/grad_norm_pre_clip': 0.1626904308795929, 'timing/data': 0.0003252299939049408, 'timing/model': 0.36878432100638747, 'learning_rate/qwen_vl_interface': 1.86763869546165e-06, 'learning_rate/action_model': 1.86763869546165e-05, 'epoch': 0.28}) +[2026-10-03 15:21:06,628][starVLA.training.train_starvla][INFO] - Step 3787, Loss: {'train/loss': 0.0002199788868892938, 'train/grad_norm_pre_clip': 0.14808940887451172, 'timing/data': 0.0002662699989741668, 'timing/model': 0.3700450010073837, 'learning_rate/qwen_vl_interface': 1.865501193299045e-06, 'learning_rate/action_model': 1.8655011932990447e-05, 'epoch': 0.28}) +[2026-10-03 15:21:07,286][starVLA.training.train_starvla][INFO] - Step 3788, Loss: {'train/loss': 0.00255562923848629, 'train/grad_norm_pre_clip': 5.517786502838135, 'timing/data': 0.00020392000442370772, 'timing/model': 0.3702873700094642, 'learning_rate/qwen_vl_interface': 1.8633650823760555e-06, 'learning_rate/action_model': 1.8633650823760555e-05, 'epoch': 0.28}) +[2026-10-03 15:21:08,146][starVLA.training.train_starvla][INFO] - Step 3789, Loss: {'train/loss': 0.35087382793426514, 'train/grad_norm_pre_clip': 86.04950714111328, 'timing/data': 0.00032405000820290297, 'timing/model': 0.36916212100186385, 'learning_rate/qwen_vl_interface': 1.8612303635707562e-06, 'learning_rate/action_model': 1.861230363570756e-05, 'epoch': 0.28}) +[2026-10-03 15:21:08,803][starVLA.training.train_starvla][INFO] - Step 3790, Loss: {'train/loss': 0.00017609127098694444, 'train/grad_norm_pre_clip': 0.1735919713973999, 'timing/data': 0.00023154000518843532, 'timing/model': 0.3686797919945093, 'learning_rate/qwen_vl_interface': 1.8590970377606494e-06, 'learning_rate/action_model': 1.859097037760649e-05, 'epoch': 0.28}) +[2026-10-03 15:21:09,460][starVLA.training.train_starvla][INFO] - Step 3791, Loss: {'train/loss': 0.4569488763809204, 'train/grad_norm_pre_clip': 63.86635208129883, 'timing/data': 0.0002687800006242469, 'timing/model': 0.3694675209990237, 'learning_rate/qwen_vl_interface': 1.8569651058226667e-06, 'learning_rate/action_model': 1.8569651058226666e-05, 'epoch': 0.28}) +[2026-10-03 15:21:10,117][starVLA.training.train_starvla][INFO] - Step 3792, Loss: {'train/loss': 2.07568627956789e-05, 'train/grad_norm_pre_clip': 0.005582207813858986, 'timing/data': 0.0003196199977537617, 'timing/model': 0.369385230995249, 'learning_rate/qwen_vl_interface': 1.854834568633161e-06, 'learning_rate/action_model': 1.854834568633161e-05, 'epoch': 0.28}) +[2026-10-03 15:21:10,774][starVLA.training.train_starvla][INFO] - Step 3793, Loss: {'train/loss': 0.0007785824127495289, 'train/grad_norm_pre_clip': 4.206152439117432, 'timing/data': 0.0002413500042166561, 'timing/model': 0.36988110101083294, 'learning_rate/qwen_vl_interface': 1.8527054270679174e-06, 'learning_rate/action_model': 1.8527054270679172e-05, 'epoch': 0.28}) +[2026-10-03 15:21:11,633][starVLA.training.train_starvla][INFO] - Step 3794, Loss: {'train/loss': 0.00011938602256122977, 'train/grad_norm_pre_clip': 0.03642984852194786, 'timing/data': 0.00021420999837573618, 'timing/model': 0.5707196639996255, 'learning_rate/qwen_vl_interface': 1.850577682002146e-06, 'learning_rate/action_model': 1.850577682002146e-05, 'epoch': 0.28}) +[2026-10-03 15:21:12,292][starVLA.training.train_starvla][INFO] - Step 3795, Loss: {'train/loss': 3.641488365246914e-05, 'train/grad_norm_pre_clip': 0.01222041342407465, 'timing/data': 0.0003261899983044714, 'timing/model': 0.3709052110061748, 'learning_rate/qwen_vl_interface': 1.848451334310482e-06, 'learning_rate/action_model': 1.848451334310482e-05, 'epoch': 0.28}) +[2026-10-03 15:21:12,949][starVLA.training.train_starvla][INFO] - Step 3796, Loss: {'train/loss': 0.3309181332588196, 'train/grad_norm_pre_clip': 55.90525817871094, 'timing/data': 0.00020545000734273344, 'timing/model': 0.3694695109879831, 'learning_rate/qwen_vl_interface': 1.8463263848669865e-06, 'learning_rate/action_model': 1.8463263848669864e-05, 'epoch': 0.28}) +[2026-10-03 15:21:13,606][starVLA.training.train_starvla][INFO] - Step 3797, Loss: {'train/loss': 3.6228535464033484e-05, 'train/grad_norm_pre_clip': 0.012024318799376488, 'timing/data': 0.000205080010346137, 'timing/model': 0.3698417910054559, 'learning_rate/qwen_vl_interface': 1.8442028345451468e-06, 'learning_rate/action_model': 1.8442028345451468e-05, 'epoch': 0.28}) +[2026-10-03 15:21:14,263][starVLA.training.train_starvla][INFO] - Step 3798, Loss: {'train/loss': 0.31241318583488464, 'train/grad_norm_pre_clip': 87.24230194091797, 'timing/data': 0.00032578001264482737, 'timing/model': 0.36902782099787146, 'learning_rate/qwen_vl_interface': 1.8420806842178746e-06, 'learning_rate/action_model': 1.8420806842178745e-05, 'epoch': 0.28}) +[2026-10-03 15:21:15,118][starVLA.training.train_starvla][INFO] - Step 3799, Loss: {'train/loss': 0.14966252446174622, 'train/grad_norm_pre_clip': 241.22117614746094, 'timing/data': 0.00028712999483104795, 'timing/model': 0.5662084150098963, 'learning_rate/qwen_vl_interface': 1.8399599347575034e-06, 'learning_rate/action_model': 1.8399599347575033e-05, 'epoch': 0.28}) +[2026-10-03 15:21:15,775][starVLA.training.train_starvla][INFO] - Step 3800, Loss: {'train/loss': 2.3207050617202185e-05, 'train/grad_norm_pre_clip': 0.008123076520860195, 'timing/data': 0.00020243000471964478, 'timing/model': 0.3697934320080094, 'learning_rate/qwen_vl_interface': 1.8378405870357943e-06, 'learning_rate/action_model': 1.837840587035794e-05, 'epoch': 0.28}) +[2026-10-03 15:21:16,432][starVLA.training.train_starvla][INFO] - Step 3801, Loss: {'train/loss': 3.996192754129879e-05, 'train/grad_norm_pre_clip': 0.014278728514909744, 'timing/data': 0.000281209999229759, 'timing/model': 0.369617181000649, 'learning_rate/qwen_vl_interface': 1.8357226419239316e-06, 'learning_rate/action_model': 1.8357226419239315e-05, 'epoch': 0.28}) +[2026-10-03 15:21:17,088][starVLA.training.train_starvla][INFO] - Step 3802, Loss: {'train/loss': 0.26259273290634155, 'train/grad_norm_pre_clip': 148.5780792236328, 'timing/data': 0.0002490799961378798, 'timing/model': 0.368409981005243, 'learning_rate/qwen_vl_interface': 1.8336061002925224e-06, 'learning_rate/action_model': 1.8336061002925224e-05, 'epoch': 0.28}) +[2026-10-03 15:21:17,746][starVLA.training.train_starvla][INFO] - Step 3803, Loss: {'train/loss': 0.00023676846467424184, 'train/grad_norm_pre_clip': 0.08409442007541656, 'timing/data': 0.00037526999949477613, 'timing/model': 0.36943157100176904, 'learning_rate/qwen_vl_interface': 1.8314909630115985e-06, 'learning_rate/action_model': 1.8314909630115983e-05, 'epoch': 0.28}) +[2026-10-03 15:21:18,403][starVLA.training.train_starvla][INFO] - Step 3804, Loss: {'train/loss': 0.00017339045007247478, 'train/grad_norm_pre_clip': 0.10206956416368484, 'timing/data': 0.0002064099971903488, 'timing/model': 0.3690697709971573, 'learning_rate/qwen_vl_interface': 1.8293772309506092e-06, 'learning_rate/action_model': 1.829377230950609e-05, 'epoch': 0.28}) +[2026-10-03 15:21:19,258][starVLA.training.train_starvla][INFO] - Step 3805, Loss: {'train/loss': 8.274650463135913e-05, 'train/grad_norm_pre_clip': 0.06165163964033127, 'timing/data': 0.00020701000175904483, 'timing/model': 0.3695753200008767, 'learning_rate/qwen_vl_interface': 1.8272649049784326e-06, 'learning_rate/action_model': 1.8272649049784323e-05, 'epoch': 0.28}) +[2026-10-03 15:21:19,915][starVLA.training.train_starvla][INFO] - Step 3806, Loss: {'train/loss': 3.614230081439018e-05, 'train/grad_norm_pre_clip': 0.019545355811715126, 'timing/data': 0.0002995500108227134, 'timing/model': 0.3686743909929646, 'learning_rate/qwen_vl_interface': 1.8251539859633653e-06, 'learning_rate/action_model': 1.825153985963365e-05, 'epoch': 0.28}) +[2026-10-03 15:21:20,573][starVLA.training.train_starvla][INFO] - Step 3807, Loss: {'train/loss': 0.001721344655379653, 'train/grad_norm_pre_clip': 3.7000715732574463, 'timing/data': 0.00034826999763026834, 'timing/model': 0.3696031309955288, 'learning_rate/qwen_vl_interface': 1.8230444747731265e-06, 'learning_rate/action_model': 1.8230444747731264e-05, 'epoch': 0.28}) +[2026-10-03 15:21:21,231][starVLA.training.train_starvla][INFO] - Step 3808, Loss: {'train/loss': 0.19053100049495697, 'train/grad_norm_pre_clip': 223.672607421875, 'timing/data': 0.00023752999550197273, 'timing/model': 0.36994357098592445, 'learning_rate/qwen_vl_interface': 1.820936372274857e-06, 'learning_rate/action_model': 1.820936372274857e-05, 'epoch': 0.28}) +[2026-10-03 15:21:21,887][starVLA.training.train_starvla][INFO] - Step 3809, Loss: {'train/loss': 1.0266739991493523e-05, 'train/grad_norm_pre_clip': 0.0019013544078916311, 'timing/data': 0.0003141899942420423, 'timing/model': 0.3693303409963846, 'learning_rate/qwen_vl_interface': 1.8188296793351178e-06, 'learning_rate/action_model': 1.818829679335118e-05, 'epoch': 0.28}) +[2026-10-03 15:21:22,738][starVLA.training.train_starvla][INFO] - Step 3810, Loss: {'train/loss': 1.1458821973064914e-05, 'train/grad_norm_pre_clip': 0.001912612933665514, 'timing/data': 0.000261950001004152, 'timing/model': 0.5625129250111058, 'learning_rate/qwen_vl_interface': 1.8167243968198905e-06, 'learning_rate/action_model': 1.8167243968198905e-05, 'epoch': 0.28}) +[2026-10-03 15:21:23,396][starVLA.training.train_starvla][INFO] - Step 3811, Loss: {'train/loss': 0.00015237020852509886, 'train/grad_norm_pre_clip': 0.36901921033859253, 'timing/data': 0.00026889999571722, 'timing/model': 0.3692031410027994, 'learning_rate/qwen_vl_interface': 1.8146205255945775e-06, 'learning_rate/action_model': 1.8146205255945777e-05, 'epoch': 0.28}) +[2026-10-03 15:21:24,054][starVLA.training.train_starvla][INFO] - Step 3812, Loss: {'train/loss': 1.0236975867883302e-05, 'train/grad_norm_pre_clip': 0.002184744691476226, 'timing/data': 0.00025861000176519156, 'timing/model': 0.3700348610000219, 'learning_rate/qwen_vl_interface': 1.8125180665240007e-06, 'learning_rate/action_model': 1.8125180665240008e-05, 'epoch': 0.28}) +[2026-10-03 15:21:24,711][starVLA.training.train_starvla][INFO] - Step 3813, Loss: {'train/loss': 1.1771588106057607e-05, 'train/grad_norm_pre_clip': 0.003115803236141801, 'timing/data': 0.0002458300004946068, 'timing/model': 0.37015410099411383, 'learning_rate/qwen_vl_interface': 1.8104170204724025e-06, 'learning_rate/action_model': 1.8104170204724024e-05, 'epoch': 0.28}) +[2026-10-03 15:21:25,368][starVLA.training.train_starvla][INFO] - Step 3814, Loss: {'train/loss': 0.2874450981616974, 'train/grad_norm_pre_clip': 85.22074890136719, 'timing/data': 0.0002607200003694743, 'timing/model': 0.36860635099583305, 'learning_rate/qwen_vl_interface': 1.8083173883034435e-06, 'learning_rate/action_model': 1.8083173883034435e-05, 'epoch': 0.28}) +[2026-10-03 15:21:26,026][starVLA.training.train_starvla][INFO] - Step 3815, Loss: {'train/loss': 0.0001630022597964853, 'train/grad_norm_pre_clip': 0.13351283967494965, 'timing/data': 0.0003086299984715879, 'timing/model': 0.36957787100982387, 'learning_rate/qwen_vl_interface': 1.8062191708802013e-06, 'learning_rate/action_model': 1.8062191708802012e-05, 'epoch': 0.28}) +[2026-10-03 15:21:26,874][starVLA.training.train_starvla][INFO] - Step 3816, Loss: {'train/loss': 4.5122553274268284e-05, 'train/grad_norm_pre_clip': 0.020132483914494514, 'timing/data': 0.00022093000006861985, 'timing/model': 0.36981651101086754, 'learning_rate/qwen_vl_interface': 1.804122369065175e-06, 'learning_rate/action_model': 1.8041223690651748e-05, 'epoch': 0.28}) +[2026-10-03 15:21:27,529][starVLA.training.train_starvla][INFO] - Step 3817, Loss: {'train/loss': 0.0001897009351523593, 'train/grad_norm_pre_clip': 0.10029754787683487, 'timing/data': 0.0002087299944832921, 'timing/model': 0.36920579099387396, 'learning_rate/qwen_vl_interface': 1.8020269837202806e-06, 'learning_rate/action_model': 1.8020269837202804e-05, 'epoch': 0.28}) +[2026-10-03 15:21:28,186][starVLA.training.train_starvla][INFO] - Step 3818, Loss: {'train/loss': 0.000399024662328884, 'train/grad_norm_pre_clip': 0.2802009880542755, 'timing/data': 0.00036331999581307173, 'timing/model': 0.3682066119945375, 'learning_rate/qwen_vl_interface': 1.7999330157068517e-06, 'learning_rate/action_model': 1.7999330157068515e-05, 'epoch': 0.28}) +[2026-10-03 15:21:28,843][starVLA.training.train_starvla][INFO] - Step 3819, Loss: {'train/loss': 0.0002221904433099553, 'train/grad_norm_pre_clip': 0.10386069118976593, 'timing/data': 0.0002211499959230423, 'timing/model': 0.36945012099749874, 'learning_rate/qwen_vl_interface': 1.7978404658856393e-06, 'learning_rate/action_model': 1.7978404658856394e-05, 'epoch': 0.28}) +[2026-10-03 15:21:29,500][starVLA.training.train_starvla][INFO] - Step 3820, Loss: {'train/loss': 0.421529620885849, 'train/grad_norm_pre_clip': 90.93310546875, 'timing/data': 0.0003048189973924309, 'timing/model': 0.36957105199689977, 'learning_rate/qwen_vl_interface': 1.7957493351168126e-06, 'learning_rate/action_model': 1.7957493351168126e-05, 'epoch': 0.28}) +[2026-10-03 15:21:30,350][starVLA.training.train_starvla][INFO] - Step 3821, Loss: {'train/loss': 7.939205534057692e-05, 'train/grad_norm_pre_clip': 0.033938489854335785, 'timing/data': 0.00022370999795384705, 'timing/model': 0.3691470009944169, 'learning_rate/qwen_vl_interface': 1.7936596242599552e-06, 'learning_rate/action_model': 1.7936596242599553e-05, 'epoch': 0.28}) +[2026-10-03 15:21:31,006][starVLA.training.train_starvla][INFO] - Step 3822, Loss: {'train/loss': 1.8819308024831116e-05, 'train/grad_norm_pre_clip': 0.0044179996475577354, 'timing/data': 0.0002605399931780994, 'timing/model': 0.3683718910033349, 'learning_rate/qwen_vl_interface': 1.7915713341740692e-06, 'learning_rate/action_model': 1.791571334174069e-05, 'epoch': 0.28}) +[2026-10-03 15:21:31,663][starVLA.training.train_starvla][INFO] - Step 3823, Loss: {'train/loss': 8.65750917000696e-05, 'train/grad_norm_pre_clip': 0.052289605140686035, 'timing/data': 0.0002313900040462613, 'timing/model': 0.36883830100123305, 'learning_rate/qwen_vl_interface': 1.7894844657175716e-06, 'learning_rate/action_model': 1.7894844657175717e-05, 'epoch': 0.28}) +[2026-10-03 15:21:32,319][starVLA.training.train_starvla][INFO] - Step 3824, Loss: {'train/loss': 0.00016377858992200345, 'train/grad_norm_pre_clip': 0.1551305651664734, 'timing/data': 0.00024750000739004463, 'timing/model': 0.36983705099555664, 'learning_rate/qwen_vl_interface': 1.7873990197482962e-06, 'learning_rate/action_model': 1.787399019748296e-05, 'epoch': 0.28}) +[2026-10-03 15:21:32,953][starVLA.training.train_starvla][INFO] - Step 3825, Loss: {'train/loss': 3.831430512946099e-05, 'train/grad_norm_pre_clip': 0.022797493264079094, 'timing/data': 0.00031466998916585, 'timing/model': 0.36891857100999914, 'learning_rate/qwen_vl_interface': 1.7853149971234883e-06, 'learning_rate/action_model': 1.785314997123488e-05, 'epoch': 0.28}) +[2026-10-03 15:21:33,586][starVLA.training.train_starvla][INFO] - Step 3826, Loss: {'train/loss': 0.14977307617664337, 'train/grad_norm_pre_clip': 128.6086883544922, 'timing/data': 0.00041360000614076853, 'timing/model': 0.36814443100593053, 'learning_rate/qwen_vl_interface': 1.7832323986998128e-06, 'learning_rate/action_model': 1.783232398699813e-05, 'epoch': 0.28}) +[2026-10-03 15:21:34,416][starVLA.training.train_starvla][INFO] - Step 3827, Loss: {'train/loss': 0.4293432831764221, 'train/grad_norm_pre_clip': 107.93186950683594, 'timing/data': 0.00032995999208651483, 'timing/model': 0.34648951300187036, 'learning_rate/qwen_vl_interface': 1.7811512253333468e-06, 'learning_rate/action_model': 1.7811512253333467e-05, 'epoch': 0.28}) +[2026-10-03 15:21:35,073][starVLA.training.train_starvla][INFO] - Step 3828, Loss: {'train/loss': 2.830033736245241e-05, 'train/grad_norm_pre_clip': 0.01459189411252737, 'timing/data': 0.0002148100029444322, 'timing/model': 0.36929549199703615, 'learning_rate/qwen_vl_interface': 1.7790714778795823e-06, 'learning_rate/action_model': 1.779071477879582e-05, 'epoch': 0.28}) +[2026-10-03 15:21:35,706][starVLA.training.train_starvla][INFO] - Step 3829, Loss: {'train/loss': 0.08715536445379257, 'train/grad_norm_pre_clip': 162.22055053710938, 'timing/data': 0.00023053999757394195, 'timing/model': 0.3466766950004967, 'learning_rate/qwen_vl_interface': 1.7769931571934246e-06, 'learning_rate/action_model': 1.7769931571934246e-05, 'epoch': 0.28}) +[2026-10-03 15:21:36,362][starVLA.training.train_starvla][INFO] - Step 3830, Loss: {'train/loss': 0.08446311950683594, 'train/grad_norm_pre_clip': 153.837646484375, 'timing/data': 0.0002578399871708825, 'timing/model': 0.3682971219968749, 'learning_rate/qwen_vl_interface': 1.7749162641291933e-06, 'learning_rate/action_model': 1.7749162641291933e-05, 'epoch': 0.28}) +[2026-10-03 15:21:37,019][starVLA.training.train_starvla][INFO] - Step 3831, Loss: {'train/loss': 4.0949285903479904e-05, 'train/grad_norm_pre_clip': 0.013195737265050411, 'timing/data': 0.0003202500083716586, 'timing/model': 0.3695115409936989, 'learning_rate/qwen_vl_interface': 1.7728407995406206e-06, 'learning_rate/action_model': 1.7728407995406205e-05, 'epoch': 0.28}) +[2026-10-03 15:21:37,876][starVLA.training.train_starvla][INFO] - Step 3832, Loss: {'train/loss': 6.157760071801022e-05, 'train/grad_norm_pre_clip': 0.027476675808429718, 'timing/data': 0.0003331399930175394, 'timing/model': 0.36928962101228535, 'learning_rate/qwen_vl_interface': 1.7707667642808513e-06, 'learning_rate/action_model': 1.7707667642808512e-05, 'epoch': 0.28}) +[2026-10-03 15:21:38,531][starVLA.training.train_starvla][INFO] - Step 3833, Loss: {'train/loss': 0.03123972751200199, 'train/grad_norm_pre_clip': 111.23577117919922, 'timing/data': 0.00025419901066925377, 'timing/model': 0.3689279819955118, 'learning_rate/qwen_vl_interface': 1.7686941592024423e-06, 'learning_rate/action_model': 1.7686941592024423e-05, 'epoch': 0.28}) +[2026-10-03 15:21:39,187][starVLA.training.train_starvla][INFO] - Step 3834, Loss: {'train/loss': 0.17499111592769623, 'train/grad_norm_pre_clip': 181.15997314453125, 'timing/data': 0.0003189200069755316, 'timing/model': 0.3682768309954554, 'learning_rate/qwen_vl_interface': 1.7666229851573662e-06, 'learning_rate/action_model': 1.7666229851573664e-05, 'epoch': 0.28}) +[2026-10-03 15:21:39,844][starVLA.training.train_starvla][INFO] - Step 3835, Loss: {'train/loss': 0.5465265512466431, 'train/grad_norm_pre_clip': 100.20699310302734, 'timing/data': 0.0003368099860381335, 'timing/model': 0.3690069910080638, 'learning_rate/qwen_vl_interface': 1.7645532429970006e-06, 'learning_rate/action_model': 1.7645532429970004e-05, 'epoch': 0.28}) +[2026-10-03 15:21:40,501][starVLA.training.train_starvla][INFO] - Step 3836, Loss: {'train/loss': 1.957974018296227e-05, 'train/grad_norm_pre_clip': 0.0038689582142978907, 'timing/data': 0.00028150998696219176, 'timing/model': 0.3699197409878252, 'learning_rate/qwen_vl_interface': 1.7624849335721402e-06, 'learning_rate/action_model': 1.7624849335721402e-05, 'epoch': 0.28}) +[2026-10-03 15:21:41,357][starVLA.training.train_starvla][INFO] - Step 3837, Loss: {'train/loss': 2.1396697775344364e-05, 'train/grad_norm_pre_clip': 0.006625671871006489, 'timing/data': 0.00021309999283403158, 'timing/model': 0.5686754060006933, 'learning_rate/qwen_vl_interface': 1.7604180577329883e-06, 'learning_rate/action_model': 1.7604180577329883e-05, 'epoch': 0.28}) +[2026-10-03 15:21:42,015][starVLA.training.train_starvla][INFO] - Step 3838, Loss: {'train/loss': 9.223780944012105e-05, 'train/grad_norm_pre_clip': 0.05411824584007263, 'timing/data': 0.00027087000489700586, 'timing/model': 0.3692991110001458, 'learning_rate/qwen_vl_interface': 1.7583526163291604e-06, 'learning_rate/action_model': 1.7583526163291604e-05, 'epoch': 0.28}) +[2026-10-03 15:21:42,674][starVLA.training.train_starvla][INFO] - Step 3839, Loss: {'train/loss': 0.00016187764413189143, 'train/grad_norm_pre_clip': 0.17402604222297668, 'timing/data': 0.0003100500034634024, 'timing/model': 0.3699262009904487, 'learning_rate/qwen_vl_interface': 1.7562886102096807e-06, 'learning_rate/action_model': 1.7562886102096807e-05, 'epoch': 0.28}) +[2026-10-03 15:21:43,331][starVLA.training.train_starvla][INFO] - Step 3840, Loss: {'train/loss': 0.0001735781697789207, 'train/grad_norm_pre_clip': 0.27877673506736755, 'timing/data': 0.0003690100129460916, 'timing/model': 0.36985932099923957, 'learning_rate/qwen_vl_interface': 1.7542260402229838e-06, 'learning_rate/action_model': 1.7542260402229838e-05, 'epoch': 0.28}) +[2026-10-03 15:21:43,987][starVLA.training.train_starvla][INFO] - Step 3841, Loss: {'train/loss': 1.4520974218612537e-05, 'train/grad_norm_pre_clip': 0.002744654892012477, 'timing/data': 0.0003084899944951758, 'timing/model': 0.3693651810026495, 'learning_rate/qwen_vl_interface': 1.7521649072169152e-06, 'learning_rate/action_model': 1.752164907216915e-05, 'epoch': 0.28}) +[2026-10-03 15:21:44,645][starVLA.training.train_starvla][INFO] - Step 3842, Loss: {'train/loss': 0.22495868802070618, 'train/grad_norm_pre_clip': 85.30792999267578, 'timing/data': 0.00024807000590953976, 'timing/model': 0.3692907210061094, 'learning_rate/qwen_vl_interface': 1.7501052120387285e-06, 'learning_rate/action_model': 1.7501052120387285e-05, 'epoch': 0.28}) +[2026-10-03 15:21:45,496][starVLA.training.train_starvla][INFO] - Step 3843, Loss: {'train/loss': 2.9070175514789298e-05, 'train/grad_norm_pre_clip': 0.009590438567101955, 'timing/data': 0.00034752000647131354, 'timing/model': 0.36929329099075403, 'learning_rate/qwen_vl_interface': 1.7480469555350866e-06, 'learning_rate/action_model': 1.7480469555350865e-05, 'epoch': 0.28}) +[2026-10-03 15:21:46,154][starVLA.training.train_starvla][INFO] - Step 3844, Loss: {'train/loss': 6.543644849443808e-05, 'train/grad_norm_pre_clip': 0.02774043194949627, 'timing/data': 0.00026402999355923384, 'timing/model': 0.3700395009946078, 'learning_rate/qwen_vl_interface': 1.7459901385520619e-06, 'learning_rate/action_model': 1.745990138552062e-05, 'epoch': 0.28}) +[2026-10-03 15:21:46,810][starVLA.training.train_starvla][INFO] - Step 3845, Loss: {'train/loss': 1.021459320327267e-05, 'train/grad_norm_pre_clip': 0.001841039047576487, 'timing/data': 0.00024600999313406646, 'timing/model': 0.3693740809976589, 'learning_rate/qwen_vl_interface': 1.7439347619351324e-06, 'learning_rate/action_model': 1.7439347619351324e-05, 'epoch': 0.28}) +[2026-10-03 15:21:47,467][starVLA.training.train_starvla][INFO] - Step 3846, Loss: {'train/loss': 3.0537798011209816e-05, 'train/grad_norm_pre_clip': 0.007778441533446312, 'timing/data': 0.000328180001815781, 'timing/model': 0.36903573099698406, 'learning_rate/qwen_vl_interface': 1.741880826529186e-06, 'learning_rate/action_model': 1.741880826529186e-05, 'epoch': 0.28}) +[2026-10-03 15:21:48,125][starVLA.training.train_starvla][INFO] - Step 3847, Loss: {'train/loss': 0.0007299090502783656, 'train/grad_norm_pre_clip': 0.7002894878387451, 'timing/data': 0.00027052999939769506, 'timing/model': 0.36943794100079685, 'learning_rate/qwen_vl_interface': 1.7398283331785205e-06, 'learning_rate/action_model': 1.7398283331785203e-05, 'epoch': 0.28}) +[2026-10-03 15:21:48,979][starVLA.training.train_starvla][INFO] - Step 3848, Loss: {'train/loss': 5.8396148233441636e-05, 'train/grad_norm_pre_clip': 0.01789327710866928, 'timing/data': 0.000361209997208789, 'timing/model': 0.37046756100608036, 'learning_rate/qwen_vl_interface': 1.737777282726835e-06, 'learning_rate/action_model': 1.737777282726835e-05, 'epoch': 0.28}) +[2026-10-03 15:21:49,636][starVLA.training.train_starvla][INFO] - Step 3849, Loss: {'train/loss': 1.194294418382924e-05, 'train/grad_norm_pre_clip': 0.002868610667064786, 'timing/data': 0.0002588799980003387, 'timing/model': 0.3691070209897589, 'learning_rate/qwen_vl_interface': 1.7357276760172403e-06, 'learning_rate/action_model': 1.7357276760172404e-05, 'epoch': 0.28}) +[2026-10-03 15:21:50,293][starVLA.training.train_starvla][INFO] - Step 3850, Loss: {'train/loss': 3.38680183631368e-05, 'train/grad_norm_pre_clip': 0.008819283917546272, 'timing/data': 0.0003544200008036569, 'timing/model': 0.36870821099728346, 'learning_rate/qwen_vl_interface': 1.733679513892252e-06, 'learning_rate/action_model': 1.7336795138922517e-05, 'epoch': 0.28}) +[2026-10-03 15:21:50,952][starVLA.training.train_starvla][INFO] - Step 3851, Loss: {'train/loss': 0.6697934865951538, 'train/grad_norm_pre_clip': 182.2688751220703, 'timing/data': 0.00035401999775785953, 'timing/model': 0.3695164310047403, 'learning_rate/qwen_vl_interface': 1.7316327971937926e-06, 'learning_rate/action_model': 1.7316327971937926e-05, 'epoch': 0.28}) +[2026-10-03 15:21:51,611][starVLA.training.train_starvla][INFO] - Step 3852, Loss: {'train/loss': 0.000424837606260553, 'train/grad_norm_pre_clip': 0.20181679725646973, 'timing/data': 0.0003402699949219823, 'timing/model': 0.370957281003939, 'learning_rate/qwen_vl_interface': 1.7295875267631904e-06, 'learning_rate/action_model': 1.7295875267631904e-05, 'epoch': 0.28}) +[2026-10-03 15:21:52,460][starVLA.training.train_starvla][INFO] - Step 3853, Loss: {'train/loss': 1.559330485179089e-05, 'train/grad_norm_pre_clip': 0.004507275298237801, 'timing/data': 0.00025815999833866954, 'timing/model': 0.5623786949872738, 'learning_rate/qwen_vl_interface': 1.7275437034411778e-06, 'learning_rate/action_model': 1.7275437034411776e-05, 'epoch': 0.28}) +[2026-10-03 15:21:53,094][starVLA.training.train_starvla][INFO] - Step 3854, Loss: {'train/loss': 0.0004318517167121172, 'train/grad_norm_pre_clip': 0.18708603084087372, 'timing/data': 0.0002872199984267354, 'timing/model': 0.3465060849994188, 'learning_rate/qwen_vl_interface': 1.7255013280678941e-06, 'learning_rate/action_model': 1.725501328067894e-05, 'epoch': 0.28}) +[2026-10-03 15:21:53,753][starVLA.training.train_starvla][INFO] - Step 3855, Loss: {'train/loss': 0.00014091262710280716, 'train/grad_norm_pre_clip': 0.05896966904401779, 'timing/data': 0.0003384999872650951, 'timing/model': 0.3697463809949113, 'learning_rate/qwen_vl_interface': 1.7234604014828824e-06, 'learning_rate/action_model': 1.7234604014828823e-05, 'epoch': 0.28}) +[2026-10-03 15:21:54,410][starVLA.training.train_starvla][INFO] - Step 3856, Loss: {'train/loss': 2.0569394109770656e-05, 'train/grad_norm_pre_clip': 0.006828388199210167, 'timing/data': 0.0003151900018565357, 'timing/model': 0.36944138100079726, 'learning_rate/qwen_vl_interface': 1.7214209245250906e-06, 'learning_rate/action_model': 1.7214209245250905e-05, 'epoch': 0.28}) +[2026-10-03 15:21:55,067][starVLA.training.train_starvla][INFO] - Step 3857, Loss: {'train/loss': 0.00033174428972415626, 'train/grad_norm_pre_clip': 0.5188962817192078, 'timing/data': 0.00021159999596420676, 'timing/model': 0.3690121110121254, 'learning_rate/qwen_vl_interface': 1.7193828980328723e-06, 'learning_rate/action_model': 1.719382898032872e-05, 'epoch': 0.28}) +[2026-10-03 15:21:55,915][starVLA.training.train_starvla][INFO] - Step 3858, Loss: {'train/loss': 0.00035140986437909305, 'train/grad_norm_pre_clip': 0.18284375965595245, 'timing/data': 0.0003076599969062954, 'timing/model': 0.5597293659957359, 'learning_rate/qwen_vl_interface': 1.71734632284398e-06, 'learning_rate/action_model': 1.7173463228439798e-05, 'epoch': 0.28}) +[2026-10-03 15:21:56,573][starVLA.training.train_starvla][INFO] - Step 3859, Loss: {'train/loss': 3.999491309514269e-05, 'train/grad_norm_pre_clip': 0.015682891011238098, 'timing/data': 0.0003614900051616132, 'timing/model': 0.36937316101102624, 'learning_rate/qwen_vl_interface': 1.7153111997955744e-06, 'learning_rate/action_model': 1.7153111997955743e-05, 'epoch': 0.28}) +[2026-10-03 15:21:57,226][starVLA.training.train_starvla][INFO] - Step 3860, Loss: {'train/loss': 0.09995795041322708, 'train/grad_norm_pre_clip': 503.7400817871094, 'timing/data': 0.00019855999562423676, 'timing/model': 0.368300221991376, 'learning_rate/qwen_vl_interface': 1.7132775297242175e-06, 'learning_rate/action_model': 1.7132775297242172e-05, 'epoch': 0.28}) +[2026-10-03 15:21:57,877][starVLA.training.train_starvla][INFO] - Step 3861, Loss: {'train/loss': 2.0063585907337256e-05, 'train/grad_norm_pre_clip': 0.004606928676366806, 'timing/data': 0.00020614000095520169, 'timing/model': 0.36696839198702946, 'learning_rate/qwen_vl_interface': 1.7112453134658747e-06, 'learning_rate/action_model': 1.7112453134658747e-05, 'epoch': 0.28}) +[2026-10-03 15:21:58,528][starVLA.training.train_starvla][INFO] - Step 3862, Loss: {'train/loss': 0.49727293848991394, 'train/grad_norm_pre_clip': 72.04711151123047, 'timing/data': 0.00037403899477794766, 'timing/model': 0.36523383200983517, 'learning_rate/qwen_vl_interface': 1.7092145518559129e-06, 'learning_rate/action_model': 1.709214551855913e-05, 'epoch': 0.28}) +[2026-10-03 15:21:59,380][starVLA.training.train_starvla][INFO] - Step 3863, Loss: {'train/loss': 2.9799321055179462e-05, 'train/grad_norm_pre_clip': 0.027454333379864693, 'timing/data': 0.00033382000401616096, 'timing/model': 0.5686947340000188, 'learning_rate/qwen_vl_interface': 1.7071852457291013e-06, 'learning_rate/action_model': 1.707185245729101e-05, 'epoch': 0.28}) +[2026-10-03 15:22:00,031][starVLA.training.train_starvla][INFO] - Step 3864, Loss: {'train/loss': 2.5308028853032738e-05, 'train/grad_norm_pre_clip': 0.009722444228827953, 'timing/data': 0.00020241999300196767, 'timing/model': 0.36777155099844094, 'learning_rate/qwen_vl_interface': 1.7051573959196122e-06, 'learning_rate/action_model': 1.7051573959196122e-05, 'epoch': 0.28}) +[2026-10-03 15:22:00,660][starVLA.training.train_starvla][INFO] - Step 3865, Loss: {'train/loss': 0.21743160486221313, 'train/grad_norm_pre_clip': 292.96649169921875, 'timing/data': 0.00029317000007722527, 'timing/model': 0.36697216099128127, 'learning_rate/qwen_vl_interface': 1.7031310032610166e-06, 'learning_rate/action_model': 1.7031310032610163e-05, 'epoch': 0.28}) +[2026-10-03 15:22:01,311][starVLA.training.train_starvla][INFO] - Step 3866, Loss: {'train/loss': 0.0001640487607801333, 'train/grad_norm_pre_clip': 0.2038617879152298, 'timing/data': 0.0002689100074348971, 'timing/model': 0.36572497099405155, 'learning_rate/qwen_vl_interface': 1.701106068586289e-06, 'learning_rate/action_model': 1.701106068586289e-05, 'epoch': 0.28}) +[2026-10-03 15:22:01,959][starVLA.training.train_starvla][INFO] - Step 3867, Loss: {'train/loss': 0.5575355887413025, 'train/grad_norm_pre_clip': 74.63371276855469, 'timing/data': 0.0003159299958497286, 'timing/model': 0.3644680610013893, 'learning_rate/qwen_vl_interface': 1.6990825927278032e-06, 'learning_rate/action_model': 1.699082592727803e-05, 'epoch': 0.28}) +[2026-10-03 15:22:02,611][starVLA.training.train_starvla][INFO] - Step 3868, Loss: {'train/loss': 0.0003449483774602413, 'train/grad_norm_pre_clip': 0.21806621551513672, 'timing/data': 0.00023548000899609178, 'timing/model': 0.3679199210018851, 'learning_rate/qwen_vl_interface': 1.6970605765173344e-06, 'learning_rate/action_model': 1.6970605765173346e-05, 'epoch': 0.28}) +[2026-10-03 15:22:03,457][starVLA.training.train_starvla][INFO] - Step 3869, Loss: {'train/loss': 2.3877724743215367e-05, 'train/grad_norm_pre_clip': 0.006800157018005848, 'timing/data': 0.00020542899437714368, 'timing/model': 0.367378151990124, 'learning_rate/qwen_vl_interface': 1.6950400207860557e-06, 'learning_rate/action_model': 1.6950400207860557e-05, 'epoch': 0.28}) +[2026-10-03 15:22:04,107][starVLA.training.train_starvla][INFO] - Step 3870, Loss: {'train/loss': 2.8280703190830536e-05, 'train/grad_norm_pre_clip': 0.013710813596844673, 'timing/data': 0.0003027400089194998, 'timing/model': 0.3650789520033868, 'learning_rate/qwen_vl_interface': 1.6930209263645412e-06, 'learning_rate/action_model': 1.6930209263645412e-05, 'epoch': 0.28}) +[2026-10-03 15:22:04,756][starVLA.training.train_starvla][INFO] - Step 3871, Loss: {'train/loss': 3.255895535403397e-06, 'train/grad_norm_pre_clip': 0.000607233087066561, 'timing/data': 0.00023720000172033906, 'timing/model': 0.36554643099952955, 'learning_rate/qwen_vl_interface': 1.6910032940827656e-06, 'learning_rate/action_model': 1.6910032940827657e-05, 'epoch': 0.28}) +[2026-10-03 15:22:05,407][starVLA.training.train_starvla][INFO] - Step 3872, Loss: {'train/loss': 2.1016858227085322e-05, 'train/grad_norm_pre_clip': 0.006288032978773117, 'timing/data': 0.00031200000375974923, 'timing/model': 0.36741637099476065, 'learning_rate/qwen_vl_interface': 1.6889871247701003e-06, 'learning_rate/action_model': 1.6889871247701002e-05, 'epoch': 0.28}) +[2026-10-03 15:22:06,035][starVLA.training.train_starvla][INFO] - Step 3873, Loss: {'train/loss': 4.485220415517688e-06, 'train/grad_norm_pre_clip': 0.0010368924122303724, 'timing/data': 0.0002540299901738763, 'timing/model': 0.34416489499562886, 'learning_rate/qwen_vl_interface': 1.6869724192553168e-06, 'learning_rate/action_model': 1.686972419255317e-05, 'epoch': 0.28}) +[2026-10-03 15:22:06,881][starVLA.training.train_starvla][INFO] - Step 3874, Loss: {'train/loss': 2.265602961415425e-05, 'train/grad_norm_pre_clip': 0.007770031690597534, 'timing/data': 0.0002556999970693141, 'timing/model': 0.5605991350021213, 'learning_rate/qwen_vl_interface': 1.6849591783665847e-06, 'learning_rate/action_model': 1.6849591783665847e-05, 'epoch': 0.28}) +[2026-10-03 15:22:07,530][starVLA.training.train_starvla][INFO] - Step 3875, Loss: {'train/loss': 1.640587652218528e-05, 'train/grad_norm_pre_clip': 0.002932374831289053, 'timing/data': 0.0001997700019273907, 'timing/model': 0.3657489110046299, 'learning_rate/qwen_vl_interface': 1.6829474029314713e-06, 'learning_rate/action_model': 1.6829474029314713e-05, 'epoch': 0.28}) +[2026-10-03 15:22:08,181][starVLA.training.train_starvla][INFO] - Step 3876, Loss: {'train/loss': 1.4088601346884388e-05, 'train/grad_norm_pre_clip': 0.0036476212553679943, 'timing/data': 0.00031396000122185796, 'timing/model': 0.36737134199938737, 'learning_rate/qwen_vl_interface': 1.6809370937769408e-06, 'learning_rate/action_model': 1.680937093776941e-05, 'epoch': 0.28}) +[2026-10-03 15:22:08,835][starVLA.training.train_starvla][INFO] - Step 3877, Loss: {'train/loss': 3.686950367409736e-05, 'train/grad_norm_pre_clip': 0.01903008297085762, 'timing/data': 0.0003030400112038478, 'timing/model': 0.36967691099562217, 'learning_rate/qwen_vl_interface': 1.678928251729356e-06, 'learning_rate/action_model': 1.6789282517293557e-05, 'epoch': 0.28}) +[2026-10-03 15:22:09,477][starVLA.training.train_starvla][INFO] - Step 3878, Loss: {'train/loss': 0.011056589893996716, 'train/grad_norm_pre_clip': 11.259466171264648, 'timing/data': 0.00029495000489987433, 'timing/model': 0.342831504996866, 'learning_rate/qwen_vl_interface': 1.6769208776144767e-06, 'learning_rate/action_model': 1.6769208776144764e-05, 'epoch': 0.28}) +[2026-10-03 15:22:10,320][starVLA.training.train_starvla][INFO] - Step 3879, Loss: {'train/loss': 2.3385484382743016e-05, 'train/grad_norm_pre_clip': 0.014661035500466824, 'timing/data': 0.00028221000684425235, 'timing/model': 0.5598731859936379, 'learning_rate/qwen_vl_interface': 1.6749149722574564e-06, 'learning_rate/action_model': 1.6749149722574564e-05, 'epoch': 0.28}) +[2026-10-03 15:22:10,973][starVLA.training.train_starvla][INFO] - Step 3880, Loss: {'train/loss': 0.00010080277570523322, 'train/grad_norm_pre_clip': 0.2539021670818329, 'timing/data': 0.00030636000155936927, 'timing/model': 0.3686262510018423, 'learning_rate/qwen_vl_interface': 1.6729105364828486e-06, 'learning_rate/action_model': 1.6729105364828484e-05, 'epoch': 0.28}) +[2026-10-03 15:22:11,628][starVLA.training.train_starvla][INFO] - Step 3881, Loss: {'train/loss': 7.487708899134304e-06, 'train/grad_norm_pre_clip': 0.0020902245305478573, 'timing/data': 0.0002508100005798042, 'timing/model': 0.3697627810033737, 'learning_rate/qwen_vl_interface': 1.6709075711146e-06, 'learning_rate/action_model': 1.6709075711146e-05, 'epoch': 0.28}) +[2026-10-03 15:22:12,292][starVLA.training.train_starvla][INFO] - Step 3882, Loss: {'train/loss': 0.00010474280134076253, 'train/grad_norm_pre_clip': 0.046222176402807236, 'timing/data': 0.00031444999331142753, 'timing/model': 0.36555220099398866, 'learning_rate/qwen_vl_interface': 1.6689060769760556e-06, 'learning_rate/action_model': 1.6689060769760557e-05, 'epoch': 0.28}) +[2026-10-03 15:22:12,941][starVLA.training.train_starvla][INFO] - Step 3883, Loss: {'train/loss': 2.293882607773412e-05, 'train/grad_norm_pre_clip': 0.007396536413580179, 'timing/data': 0.0003101600013906136, 'timing/model': 0.365117512003053, 'learning_rate/qwen_vl_interface': 1.6669060548899528e-06, 'learning_rate/action_model': 1.6669060548899525e-05, 'epoch': 0.28}) +[2026-10-03 15:22:13,767][starVLA.training.train_starvla][INFO] - Step 3884, Loss: {'train/loss': 0.00028428039513528347, 'train/grad_norm_pre_clip': 0.0684586763381958, 'timing/data': 0.0002999300049850717, 'timing/model': 0.5642208649951499, 'learning_rate/qwen_vl_interface': 1.6649075056784253e-06, 'learning_rate/action_model': 1.6649075056784253e-05, 'epoch': 0.28}) +[2026-10-03 15:22:14,422][starVLA.training.train_starvla][INFO] - Step 3885, Loss: {'train/loss': 0.48310092091560364, 'train/grad_norm_pre_clip': 57.71651077270508, 'timing/data': 0.00032292999094352126, 'timing/model': 0.36869957199087366, 'learning_rate/qwen_vl_interface': 1.6629104301630019e-06, 'learning_rate/action_model': 1.662910430163002e-05, 'epoch': 0.28}) +[2026-10-03 15:22:15,082][starVLA.training.train_starvla][INFO] - Step 3886, Loss: {'train/loss': 0.032144252210855484, 'train/grad_norm_pre_clip': 60.77527618408203, 'timing/data': 0.00020680000307038426, 'timing/model': 0.3652199319913052, 'learning_rate/qwen_vl_interface': 1.6609148291646043e-06, 'learning_rate/action_model': 1.6609148291646042e-05, 'epoch': 0.28}) +[2026-10-03 15:22:15,731][starVLA.training.train_starvla][INFO] - Step 3887, Loss: {'train/loss': 0.0001078971617971547, 'train/grad_norm_pre_clip': 0.047231484204530716, 'timing/data': 0.0002830500015988946, 'timing/model': 0.3658949009986827, 'learning_rate/qwen_vl_interface': 1.658920703503549e-06, 'learning_rate/action_model': 1.6589207035035488e-05, 'epoch': 0.28}) +[2026-10-03 15:22:16,337][starVLA.training.train_starvla][INFO] - Step 3888, Loss: {'train/loss': 5.736901130148908e-06, 'train/grad_norm_pre_clip': 0.0009824143489822745, 'timing/data': 0.00019783001334872097, 'timing/model': 0.34425832401029766, 'learning_rate/qwen_vl_interface': 1.6569280539995467e-06, 'learning_rate/action_model': 1.6569280539995467e-05, 'epoch': 0.28}) +[2026-10-03 15:22:16,987][starVLA.training.train_starvla][INFO] - Step 3889, Loss: {'train/loss': 0.0007139510125853121, 'train/grad_norm_pre_clip': 0.8909307718276978, 'timing/data': 0.0002070299960905686, 'timing/model': 0.3668115710024722, 'learning_rate/qwen_vl_interface': 1.6549368814716977e-06, 'learning_rate/action_model': 1.6549368814716977e-05, 'epoch': 0.28}) +[2026-10-03 15:22:17,830][starVLA.training.train_starvla][INFO] - Step 3890, Loss: {'train/loss': 2.1001951608923264e-05, 'train/grad_norm_pre_clip': 0.00916758831590414, 'timing/data': 0.00020845999824814498, 'timing/model': 0.36620575201231986, 'learning_rate/qwen_vl_interface': 1.6529471867385009e-06, 'learning_rate/action_model': 1.6529471867385007e-05, 'epoch': 0.28}) +[2026-10-03 15:22:18,479][starVLA.training.train_starvla][INFO] - Step 3891, Loss: {'train/loss': 0.5049987435340881, 'train/grad_norm_pre_clip': 84.6548843383789, 'timing/data': 0.0003368299949215725, 'timing/model': 0.36529469200468156, 'learning_rate/qwen_vl_interface': 1.6509589706178417e-06, 'learning_rate/action_model': 1.6509589706178417e-05, 'epoch': 0.28}) +[2026-10-03 15:22:19,129][starVLA.training.train_starvla][INFO] - Step 3892, Loss: {'train/loss': 0.0002099279226968065, 'train/grad_norm_pre_clip': 0.09750087559223175, 'timing/data': 0.00019597999926190823, 'timing/model': 0.3669959410035517, 'learning_rate/qwen_vl_interface': 1.648972233927002e-06, 'learning_rate/action_model': 1.648972233927002e-05, 'epoch': 0.28}) +[2026-10-03 15:22:19,781][starVLA.training.train_starvla][INFO] - Step 3893, Loss: {'train/loss': 9.87649618764408e-05, 'train/grad_norm_pre_clip': 0.06839326024055481, 'timing/data': 0.0003224499960197136, 'timing/model': 0.36761589100933634, 'learning_rate/qwen_vl_interface': 1.646986977482654e-06, 'learning_rate/action_model': 1.6469869774826538e-05, 'epoch': 0.28}) +[2026-10-03 15:22:20,432][starVLA.training.train_starvla][INFO] - Step 3894, Loss: {'train/loss': 1.683006303210277e-05, 'train/grad_norm_pre_clip': 0.006898137275129557, 'timing/data': 0.00031991000287234783, 'timing/model': 0.36537515198870096, 'learning_rate/qwen_vl_interface': 1.6450032021008618e-06, 'learning_rate/action_model': 1.6450032021008617e-05, 'epoch': 0.29}) +[2026-10-03 15:22:21,254][starVLA.training.train_starvla][INFO] - Step 3895, Loss: {'train/loss': 0.00025666248984634876, 'train/grad_norm_pre_clip': 0.07100682705640793, 'timing/data': 0.00032715999986976385, 'timing/model': 0.34244307498738635, 'learning_rate/qwen_vl_interface': 1.6430209085970812e-06, 'learning_rate/action_model': 1.643020908597081e-05, 'epoch': 0.29}) +[2026-10-03 15:22:21,905][starVLA.training.train_starvla][INFO] - Step 3896, Loss: {'train/loss': 0.11668738722801208, 'train/grad_norm_pre_clip': 180.85147094726562, 'timing/data': 0.00034423898614477366, 'timing/model': 0.36711903200193774, 'learning_rate/qwen_vl_interface': 1.641040097786157e-06, 'learning_rate/action_model': 1.641040097786157e-05, 'epoch': 0.29}) +[2026-10-03 15:22:22,557][starVLA.training.train_starvla][INFO] - Step 3897, Loss: {'train/loss': 3.0455914384219795e-05, 'train/grad_norm_pre_clip': 0.008457843214273453, 'timing/data': 0.00028851999377366155, 'timing/model': 0.36809250099759083, 'learning_rate/qwen_vl_interface': 1.6390607704823267e-06, 'learning_rate/action_model': 1.6390607704823265e-05, 'epoch': 0.29}) +[2026-10-03 15:22:23,208][starVLA.training.train_starvla][INFO] - Step 3898, Loss: {'train/loss': 0.00012777198571711779, 'train/grad_norm_pre_clip': 0.07656539231538773, 'timing/data': 0.00027196999872103333, 'timing/model': 0.36598664200573694, 'learning_rate/qwen_vl_interface': 1.6370829274992162e-06, 'learning_rate/action_model': 1.6370829274992162e-05, 'epoch': 0.29}) +[2026-10-03 15:22:23,860][starVLA.training.train_starvla][INFO] - Step 3899, Loss: {'train/loss': 0.001158210332505405, 'train/grad_norm_pre_clip': 2.3173470497131348, 'timing/data': 0.00020817999029532075, 'timing/model': 0.36763495099148713, 'learning_rate/qwen_vl_interface': 1.6351065696498422e-06, 'learning_rate/action_model': 1.6351065696498422e-05, 'epoch': 0.29}) +[2026-10-03 15:22:24,682][starVLA.training.train_starvla][INFO] - Step 3900, Loss: {'train/loss': 1.0311439837096259e-05, 'train/grad_norm_pre_clip': 0.0023541906848549843, 'timing/data': 0.0002552700025262311, 'timing/model': 0.5604878449958051, 'learning_rate/qwen_vl_interface': 1.6331316977466117e-06, 'learning_rate/action_model': 1.6331316977466115e-05, 'epoch': 0.29}) +[2026-10-03 15:22:25,335][starVLA.training.train_starvla][INFO] - Step 3901, Loss: {'train/loss': 4.360891034593806e-05, 'train/grad_norm_pre_clip': 0.01634197309613228, 'timing/data': 0.00020476999634411186, 'timing/model': 0.3676291909941938, 'learning_rate/qwen_vl_interface': 1.6311583126013198e-06, 'learning_rate/action_model': 1.6311583126013198e-05, 'epoch': 0.29}) +[2026-10-03 15:22:25,987][starVLA.training.train_starvla][INFO] - Step 3902, Loss: {'train/loss': 3.482198735582642e-05, 'train/grad_norm_pre_clip': 0.01293128915131092, 'timing/data': 0.00041094000334851444, 'timing/model': 0.36618468099914026, 'learning_rate/qwen_vl_interface': 1.6291864150251488e-06, 'learning_rate/action_model': 1.629186415025149e-05, 'epoch': 0.29}) +[2026-10-03 15:22:26,641][starVLA.training.train_starvla][INFO] - Step 3903, Loss: {'train/loss': 0.00014835753245279193, 'train/grad_norm_pre_clip': 0.09520574659109116, 'timing/data': 0.0003231499867979437, 'timing/model': 0.3690959019877482, 'learning_rate/qwen_vl_interface': 1.6272160058286727e-06, 'learning_rate/action_model': 1.6272160058286726e-05, 'epoch': 0.29}) +[2026-10-03 15:22:27,300][starVLA.training.train_starvla][INFO] - Step 3904, Loss: {'train/loss': 0.00017521943664178252, 'train/grad_norm_pre_clip': 0.02918357029557228, 'timing/data': 0.0003427199990255758, 'timing/model': 0.3715241110039642, 'learning_rate/qwen_vl_interface': 1.6252470858218519e-06, 'learning_rate/action_model': 1.6252470858218517e-05, 'epoch': 0.29}) +[2026-10-03 15:22:28,158][starVLA.training.train_starvla][INFO] - Step 3905, Loss: {'train/loss': 7.919892595964484e-06, 'train/grad_norm_pre_clip': 0.0018568361410871148, 'timing/data': 0.0003593400033423677, 'timing/model': 0.5697324040083913, 'learning_rate/qwen_vl_interface': 1.6232796558140357e-06, 'learning_rate/action_model': 1.6232796558140357e-05, 'epoch': 0.29}) +[2026-10-03 15:22:28,817][starVLA.training.train_starvla][INFO] - Step 3906, Loss: {'train/loss': 1.222618084284477e-05, 'train/grad_norm_pre_clip': 0.0027293572202324867, 'timing/data': 0.0002796100015984848, 'timing/model': 0.37024233100237325, 'learning_rate/qwen_vl_interface': 1.6213137166139592e-06, 'learning_rate/action_model': 1.621313716613959e-05, 'epoch': 0.29}) +[2026-10-03 15:22:29,475][starVLA.training.train_starvla][INFO] - Step 3907, Loss: {'train/loss': 4.1922383388737217e-05, 'train/grad_norm_pre_clip': 0.03959505632519722, 'timing/data': 0.00034406999475322664, 'timing/model': 0.36985748099687044, 'learning_rate/qwen_vl_interface': 1.6193492690297463e-06, 'learning_rate/action_model': 1.6193492690297464e-05, 'epoch': 0.29}) +[2026-10-03 15:22:30,135][starVLA.training.train_starvla][INFO] - Step 3908, Loss: {'train/loss': 0.2884998917579651, 'train/grad_norm_pre_clip': 88.35102081298828, 'timing/data': 0.00042658000893425196, 'timing/model': 0.3719580809993204, 'learning_rate/qwen_vl_interface': 1.6173863138689078e-06, 'learning_rate/action_model': 1.617386313868908e-05, 'epoch': 0.29}) +[2026-10-03 15:22:30,794][starVLA.training.train_starvla][INFO] - Step 3909, Loss: {'train/loss': 0.20113302767276764, 'train/grad_norm_pre_clip': 164.84127807617188, 'timing/data': 0.00026296000578440726, 'timing/model': 0.3714335600088816, 'learning_rate/qwen_vl_interface': 1.6154248519383396e-06, 'learning_rate/action_model': 1.6154248519383395e-05, 'epoch': 0.29}) +[2026-10-03 15:22:31,454][starVLA.training.train_starvla][INFO] - Step 3910, Loss: {'train/loss': 0.26444268226623535, 'train/grad_norm_pre_clip': 97.42758178710938, 'timing/data': 0.00038681901060044765, 'timing/model': 0.370042320995708, 'learning_rate/qwen_vl_interface': 1.6134648840443248e-06, 'learning_rate/action_model': 1.613464884044325e-05, 'epoch': 0.29}) +[2026-10-03 15:22:32,314][starVLA.training.train_starvla][INFO] - Step 3911, Loss: {'train/loss': 9.90155604085885e-06, 'train/grad_norm_pre_clip': 0.003620034083724022, 'timing/data': 0.0003703900001710281, 'timing/model': 0.37039787200046703, 'learning_rate/qwen_vl_interface': 1.6115064109925334e-06, 'learning_rate/action_model': 1.6115064109925336e-05, 'epoch': 0.29}) +[2026-10-03 15:22:32,974][starVLA.training.train_starvla][INFO] - Step 3912, Loss: {'train/loss': 1.4885343261994421e-05, 'train/grad_norm_pre_clip': 0.007282341364771128, 'timing/data': 0.00035150999610777944, 'timing/model': 0.3715629599901149, 'learning_rate/qwen_vl_interface': 1.6095494335880173e-06, 'learning_rate/action_model': 1.6095494335880174e-05, 'epoch': 0.29}) +[2026-10-03 15:22:33,633][starVLA.training.train_starvla][INFO] - Step 3913, Loss: {'train/loss': 0.0001621297706151381, 'train/grad_norm_pre_clip': 0.07466666400432587, 'timing/data': 0.0002562199952080846, 'timing/model': 0.37043614099093247, 'learning_rate/qwen_vl_interface': 1.6075939526352175e-06, 'learning_rate/action_model': 1.6075939526352175e-05, 'epoch': 0.29}) +[2026-10-03 15:22:34,291][starVLA.training.train_starvla][INFO] - Step 3914, Loss: {'train/loss': 0.0001266085891984403, 'train/grad_norm_pre_clip': 0.03803400322794914, 'timing/data': 0.000341090009897016, 'timing/model': 0.36957563100440893, 'learning_rate/qwen_vl_interface': 1.6056399689379576e-06, 'learning_rate/action_model': 1.6056399689379575e-05, 'epoch': 0.29}) +[2026-10-03 15:22:34,949][starVLA.training.train_starvla][INFO] - Step 3915, Loss: {'train/loss': 0.00024810642935335636, 'train/grad_norm_pre_clip': 0.22809463739395142, 'timing/data': 0.00022764000459574163, 'timing/model': 0.3694571609958075, 'learning_rate/qwen_vl_interface': 1.6036874832994473e-06, 'learning_rate/action_model': 1.603687483299447e-05, 'epoch': 0.29}) +[2026-10-03 15:22:35,804][starVLA.training.train_starvla][INFO] - Step 3916, Loss: {'train/loss': 3.426038892939687e-05, 'train/grad_norm_pre_clip': 0.018789609894156456, 'timing/data': 0.0002637599973240867, 'timing/model': 0.5672879640042083, 'learning_rate/qwen_vl_interface': 1.6017364965222796e-06, 'learning_rate/action_model': 1.6017364965222793e-05, 'epoch': 0.29}) +[2026-10-03 15:22:36,463][starVLA.training.train_starvla][INFO] - Step 3917, Loss: {'train/loss': 0.0004515453474596143, 'train/grad_norm_pre_clip': 0.27282437682151794, 'timing/data': 0.00021371999173425138, 'timing/model': 0.37083608099783305, 'learning_rate/qwen_vl_interface': 1.5997870094084305e-06, 'learning_rate/action_model': 1.5997870094084306e-05, 'epoch': 0.29}) +[2026-10-03 15:22:37,122][starVLA.training.train_starvla][INFO] - Step 3918, Loss: {'train/loss': 1.3045373634668067e-05, 'train/grad_norm_pre_clip': 0.003318328410387039, 'timing/data': 0.0003130900004180148, 'timing/model': 0.36982427099428605, 'learning_rate/qwen_vl_interface': 1.5978390227592616e-06, 'learning_rate/action_model': 1.5978390227592615e-05, 'epoch': 0.29}) +[2026-10-03 15:22:37,756][starVLA.training.train_starvla][INFO] - Step 3919, Loss: {'train/loss': 0.6017715334892273, 'train/grad_norm_pre_clip': 49.858280181884766, 'timing/data': 0.0003328099992359057, 'timing/model': 0.3690500910015544, 'learning_rate/qwen_vl_interface': 1.5958925373755164e-06, 'learning_rate/action_model': 1.5958925373755164e-05, 'epoch': 0.29}) +[2026-10-03 15:22:38,415][starVLA.training.train_starvla][INFO] - Step 3920, Loss: {'train/loss': 0.3633354902267456, 'train/grad_norm_pre_clip': 57.12934875488281, 'timing/data': 0.0002734599984250963, 'timing/model': 0.37100176099920645, 'learning_rate/qwen_vl_interface': 1.5939475540573213e-06, 'learning_rate/action_model': 1.5939475540573212e-05, 'epoch': 0.29}) +[2026-10-03 15:22:39,273][starVLA.training.train_starvla][INFO] - Step 3921, Loss: {'train/loss': 0.00010695039964048192, 'train/grad_norm_pre_clip': 0.07342381775379181, 'timing/data': 0.0002586499904282391, 'timing/model': 0.5696815150004113, 'learning_rate/qwen_vl_interface': 1.5920040736041867e-06, 'learning_rate/action_model': 1.5920040736041866e-05, 'epoch': 0.29}) +[2026-10-03 15:22:39,932][starVLA.training.train_starvla][INFO] - Step 3922, Loss: {'train/loss': 0.4366482198238373, 'train/grad_norm_pre_clip': 120.86299133300781, 'timing/data': 0.0003081299946643412, 'timing/model': 0.3700299819902284, 'learning_rate/qwen_vl_interface': 1.5900620968150015e-06, 'learning_rate/action_model': 1.5900620968150015e-05, 'epoch': 0.29}) +[2026-10-03 15:22:40,588][starVLA.training.train_starvla][INFO] - Step 3923, Loss: {'train/loss': 7.867761269153561e-06, 'train/grad_norm_pre_clip': 0.0014061040710657835, 'timing/data': 0.00033666001399978995, 'timing/model': 0.3691854310018243, 'learning_rate/qwen_vl_interface': 1.5881216244880407e-06, 'learning_rate/action_model': 1.5881216244880407e-05, 'epoch': 0.29}) +[2026-10-03 15:22:41,249][starVLA.training.train_starvla][INFO] - Step 3924, Loss: {'train/loss': 0.000564013491384685, 'train/grad_norm_pre_clip': 0.8159377574920654, 'timing/data': 0.0003357800014782697, 'timing/model': 0.3715318010072224, 'learning_rate/qwen_vl_interface': 1.5861826574209587e-06, 'learning_rate/action_model': 1.5861826574209587e-05, 'epoch': 0.29}) +[2026-10-03 15:22:41,908][starVLA.training.train_starvla][INFO] - Step 3925, Loss: {'train/loss': 1.9855106074828655e-05, 'train/grad_norm_pre_clip': 0.006697594188153744, 'timing/data': 0.0003222390078008175, 'timing/model': 0.3712674520065775, 'learning_rate/qwen_vl_interface': 1.5842451964107916e-06, 'learning_rate/action_model': 1.5842451964107916e-05, 'epoch': 0.29}) +[2026-10-03 15:22:42,567][starVLA.training.train_starvla][INFO] - Step 3926, Loss: {'train/loss': 0.0017615888500586152, 'train/grad_norm_pre_clip': 9.809138298034668, 'timing/data': 0.0002729699917836115, 'timing/model': 0.36996904099942185, 'learning_rate/qwen_vl_interface': 1.5823092422539567e-06, 'learning_rate/action_model': 1.582309242253957e-05, 'epoch': 0.29}) +[2026-10-03 15:22:43,425][starVLA.training.train_starvla][INFO] - Step 3927, Loss: {'train/loss': 9.028131898958236e-05, 'train/grad_norm_pre_clip': 0.05058998987078667, 'timing/data': 0.0002610199881019071, 'timing/model': 0.3699225909949746, 'learning_rate/qwen_vl_interface': 1.5803747957462514e-06, 'learning_rate/action_model': 1.5803747957462514e-05, 'epoch': 0.29}) +[2026-10-03 15:22:44,085][starVLA.training.train_starvla][INFO] - Step 3928, Loss: {'train/loss': 0.01556171290576458, 'train/grad_norm_pre_clip': 92.1069564819336, 'timing/data': 0.0003403799928491935, 'timing/model': 0.37171846099954564, 'learning_rate/qwen_vl_interface': 1.5784418576828541e-06, 'learning_rate/action_model': 1.5784418576828543e-05, 'epoch': 0.29}) +[2026-10-03 15:22:44,745][starVLA.training.train_starvla][INFO] - Step 3929, Loss: {'train/loss': 0.00019171778694726527, 'train/grad_norm_pre_clip': 0.07023600488901138, 'timing/data': 0.0002791799925034866, 'timing/model': 0.37160172099538613, 'learning_rate/qwen_vl_interface': 1.576510428858322e-06, 'learning_rate/action_model': 1.5765104288583217e-05, 'epoch': 0.29}) +[2026-10-03 15:22:45,405][starVLA.training.train_starvla][INFO] - Step 3930, Loss: {'train/loss': 1.711324330244679e-05, 'train/grad_norm_pre_clip': 0.005224902648478746, 'timing/data': 0.0003011700027855113, 'timing/model': 0.36974457099859137, 'learning_rate/qwen_vl_interface': 1.5745805100665922e-06, 'learning_rate/action_model': 1.5745805100665923e-05, 'epoch': 0.29}) +[2026-10-03 15:22:46,064][starVLA.training.train_starvla][INFO] - Step 3931, Loss: {'train/loss': 5.558100383495912e-06, 'train/grad_norm_pre_clip': 0.001291234279051423, 'timing/data': 0.000404050006181933, 'timing/model': 0.369988641992677, 'learning_rate/qwen_vl_interface': 1.572652102100983e-06, 'learning_rate/action_model': 1.572652102100983e-05, 'epoch': 0.29}) +[2026-10-03 15:22:46,919][starVLA.training.train_starvla][INFO] - Step 3932, Loss: {'train/loss': 0.34597504138946533, 'train/grad_norm_pre_clip': 142.2493133544922, 'timing/data': 0.00025779999850783497, 'timing/model': 0.5657263559987769, 'learning_rate/qwen_vl_interface': 1.5707252057541898e-06, 'learning_rate/action_model': 1.5707252057541896e-05, 'epoch': 0.29}) +[2026-10-03 15:22:47,579][starVLA.training.train_starvla][INFO] - Step 3933, Loss: {'train/loss': 1.844718644861132e-05, 'train/grad_norm_pre_clip': 0.00359743507578969, 'timing/data': 0.00027072000375483185, 'timing/model': 0.37164152099285275, 'learning_rate/qwen_vl_interface': 1.5687998218182849e-06, 'learning_rate/action_model': 1.5687998218182847e-05, 'epoch': 0.29}) +[2026-10-03 15:22:48,239][starVLA.training.train_starvla][INFO] - Step 3934, Loss: {'train/loss': 2.347526060475502e-05, 'train/grad_norm_pre_clip': 0.009915267117321491, 'timing/data': 0.0003382400027476251, 'timing/model': 0.37080171100387815, 'learning_rate/qwen_vl_interface': 1.5668759510847235e-06, 'learning_rate/action_model': 1.5668759510847233e-05, 'epoch': 0.29}) +[2026-10-03 15:22:48,898][starVLA.training.train_starvla][INFO] - Step 3935, Loss: {'train/loss': 0.7084111571311951, 'train/grad_norm_pre_clip': 55.61699676513672, 'timing/data': 0.00032577000092715025, 'timing/model': 0.370371431010426, 'learning_rate/qwen_vl_interface': 1.5649535943443331e-06, 'learning_rate/action_model': 1.564953594344333e-05, 'epoch': 0.29}) +[2026-10-03 15:22:49,559][starVLA.training.train_starvla][INFO] - Step 3936, Loss: {'train/loss': 2.948027031379752e-05, 'train/grad_norm_pre_clip': 0.007788811810314655, 'timing/data': 0.00025425000058021396, 'timing/model': 0.37214314099401236, 'learning_rate/qwen_vl_interface': 1.5630327523873244e-06, 'learning_rate/action_model': 1.563032752387324e-05, 'epoch': 0.29}) +[2026-10-03 15:22:50,414][starVLA.training.train_starvla][INFO] - Step 3937, Loss: {'train/loss': 7.336541602853686e-05, 'train/grad_norm_pre_clip': 0.05486401915550232, 'timing/data': 0.00030497000261675566, 'timing/model': 0.5654414849996101, 'learning_rate/qwen_vl_interface': 1.561113426003282e-06, 'learning_rate/action_model': 1.561113426003282e-05, 'epoch': 0.29}) +[2026-10-03 15:22:51,073][starVLA.training.train_starvla][INFO] - Step 3938, Loss: {'train/loss': 0.005723233800381422, 'train/grad_norm_pre_clip': 11.84821605682373, 'timing/data': 0.0002739999908953905, 'timing/model': 0.3702858210017439, 'learning_rate/qwen_vl_interface': 1.5591956159811686e-06, 'learning_rate/action_model': 1.5591956159811687e-05, 'epoch': 0.29}) +[2026-10-03 15:22:51,732][starVLA.training.train_starvla][INFO] - Step 3939, Loss: {'train/loss': 0.0009748543379828334, 'train/grad_norm_pre_clip': 1.3858498334884644, 'timing/data': 0.00023982000129763037, 'timing/model': 0.37010141200153157, 'learning_rate/qwen_vl_interface': 1.5572793231093235e-06, 'learning_rate/action_model': 1.5572793231093236e-05, 'epoch': 0.29}) +[2026-10-03 15:22:52,391][starVLA.training.train_starvla][INFO] - Step 3940, Loss: {'train/loss': 1.6606631106697023e-05, 'train/grad_norm_pre_clip': 0.005282486788928509, 'timing/data': 0.00022510999406222254, 'timing/model': 0.3710134510038188, 'learning_rate/qwen_vl_interface': 1.5553645481754626e-06, 'learning_rate/action_model': 1.5553645481754626e-05, 'epoch': 0.29}) +[2026-10-03 15:22:53,051][starVLA.training.train_starvla][INFO] - Step 3941, Loss: {'train/loss': 0.00013034173753112555, 'train/grad_norm_pre_clip': 0.11736250668764114, 'timing/data': 0.0002912299969466403, 'timing/model': 0.3713085510098608, 'learning_rate/qwen_vl_interface': 1.5534512919666775e-06, 'learning_rate/action_model': 1.5534512919666776e-05, 'epoch': 0.29}) +[2026-10-03 15:22:53,711][starVLA.training.train_starvla][INFO] - Step 3942, Loss: {'train/loss': 8.523325050191488e-06, 'train/grad_norm_pre_clip': 0.0016009531682357192, 'timing/data': 0.0002613399992696941, 'timing/model': 0.37029385099594947, 'learning_rate/qwen_vl_interface': 1.551539555269435e-06, 'learning_rate/action_model': 1.5515395552694348e-05, 'epoch': 0.29}) +[2026-10-03 15:22:54,562][starVLA.training.train_starvla][INFO] - Step 3943, Loss: {'train/loss': 5.143102316651493e-05, 'train/grad_norm_pre_clip': 0.01755085028707981, 'timing/data': 0.00034204999974463135, 'timing/model': 0.37026041198987514, 'learning_rate/qwen_vl_interface': 1.5496293388695787e-06, 'learning_rate/action_model': 1.5496293388695786e-05, 'epoch': 0.29}) +[2026-10-03 15:22:55,222][starVLA.training.train_starvla][INFO] - Step 3944, Loss: {'train/loss': 0.22619348764419556, 'train/grad_norm_pre_clip': 118.7316665649414, 'timing/data': 0.00032993999775499105, 'timing/model': 0.3715842500096187, 'learning_rate/qwen_vl_interface': 1.547720643552327e-06, 'learning_rate/action_model': 1.547720643552327e-05, 'epoch': 0.29}) +[2026-10-03 15:22:55,882][starVLA.training.train_starvla][INFO] - Step 3945, Loss: {'train/loss': 0.4547286033630371, 'train/grad_norm_pre_clip': 60.95198059082031, 'timing/data': 0.0002669399982551113, 'timing/model': 0.3702233309886651, 'learning_rate/qwen_vl_interface': 1.5458134701022706e-06, 'learning_rate/action_model': 1.5458134701022705e-05, 'epoch': 0.29}) +[2026-10-03 15:22:56,540][starVLA.training.train_starvla][INFO] - Step 3946, Loss: {'train/loss': 0.00039972516242414713, 'train/grad_norm_pre_clip': 0.38689982891082764, 'timing/data': 0.0003557799936970696, 'timing/model': 0.3700400109955808, 'learning_rate/qwen_vl_interface': 1.5439078193033765e-06, 'learning_rate/action_model': 1.5439078193033766e-05, 'epoch': 0.29}) +[2026-10-03 15:22:57,197][starVLA.training.train_starvla][INFO] - Step 3947, Loss: {'train/loss': 7.934642781037837e-06, 'train/grad_norm_pre_clip': 0.0019309442723169923, 'timing/data': 0.00026992999482899904, 'timing/model': 0.3695282410044456, 'learning_rate/qwen_vl_interface': 1.5420036919389872e-06, 'learning_rate/action_model': 1.5420036919389873e-05, 'epoch': 0.29}) +[2026-10-03 15:22:58,056][starVLA.training.train_starvla][INFO] - Step 3948, Loss: {'train/loss': 0.0005906834849156439, 'train/grad_norm_pre_clip': 0.28355932235717773, 'timing/data': 0.00027438999677542597, 'timing/model': 0.3701560709887417, 'learning_rate/qwen_vl_interface': 1.540101088791817e-06, 'learning_rate/action_model': 1.5401010887918168e-05, 'epoch': 0.29}) +[2026-10-03 15:22:58,715][starVLA.training.train_starvla][INFO] - Step 3949, Loss: {'train/loss': 0.46619731187820435, 'train/grad_norm_pre_clip': 52.049320220947266, 'timing/data': 0.0003206100082024932, 'timing/model': 0.3713229010027135, 'learning_rate/qwen_vl_interface': 1.5382000106439532e-06, 'learning_rate/action_model': 1.5382000106439532e-05, 'epoch': 0.29}) +[2026-10-03 15:22:59,352][starVLA.training.train_starvla][INFO] - Step 3950, Loss: {'train/loss': 6.0363126976881176e-05, 'train/grad_norm_pre_clip': 0.02452935092151165, 'timing/data': 0.0003890799998771399, 'timing/model': 0.3699550010060193, 'learning_rate/qwen_vl_interface': 1.5363004582768584e-06, 'learning_rate/action_model': 1.5363004582768584e-05, 'epoch': 0.29}) +[2026-10-03 15:23:00,010][starVLA.training.train_starvla][INFO] - Step 3951, Loss: {'train/loss': 2.6515437639318407e-05, 'train/grad_norm_pre_clip': 0.007049488835036755, 'timing/data': 0.00021533000108320266, 'timing/model': 0.36973655100155156, 'learning_rate/qwen_vl_interface': 1.534402432471367e-06, 'learning_rate/action_model': 1.534402432471367e-05, 'epoch': 0.29}) +[2026-10-03 15:23:00,670][starVLA.training.train_starvla][INFO] - Step 3952, Loss: {'train/loss': 0.0001466695684939623, 'train/grad_norm_pre_clip': 0.11480636149644852, 'timing/data': 0.0002555699902586639, 'timing/model': 0.3711799510056153, 'learning_rate/qwen_vl_interface': 1.532505934007685e-06, 'learning_rate/action_model': 1.532505934007685e-05, 'epoch': 0.29}) +[2026-10-03 15:23:01,528][starVLA.training.train_starvla][INFO] - Step 3953, Loss: {'train/loss': 4.6091536205494776e-05, 'train/grad_norm_pre_clip': 0.01632816344499588, 'timing/data': 0.0002897300000768155, 'timing/model': 0.5696710549964337, 'learning_rate/qwen_vl_interface': 1.5306109636653915e-06, 'learning_rate/action_model': 1.5306109636653914e-05, 'epoch': 0.29}) +[2026-10-03 15:23:02,187][starVLA.training.train_starvla][INFO] - Step 3954, Loss: {'train/loss': 0.3838781416416168, 'train/grad_norm_pre_clip': 69.1220932006836, 'timing/data': 0.0002550799981690943, 'timing/model': 0.37014065199764445, 'learning_rate/qwen_vl_interface': 1.5287175222234383e-06, 'learning_rate/action_model': 1.528717522223438e-05, 'epoch': 0.29}) +[2026-10-03 15:23:02,845][starVLA.training.train_starvla][INFO] - Step 3955, Loss: {'train/loss': 0.0002821950474753976, 'train/grad_norm_pre_clip': 0.08250848948955536, 'timing/data': 0.0003608900005929172, 'timing/model': 0.36999849100538995, 'learning_rate/qwen_vl_interface': 1.5268256104601457e-06, 'learning_rate/action_model': 1.5268256104601454e-05, 'epoch': 0.29}) +[2026-10-03 15:23:03,505][starVLA.training.train_starvla][INFO] - Step 3956, Loss: {'train/loss': 9.164054972643498e-06, 'train/grad_norm_pre_clip': 0.0027605402283370495, 'timing/data': 0.0002641099999891594, 'timing/model': 0.3720260109985247, 'learning_rate/qwen_vl_interface': 1.5249352291532075e-06, 'learning_rate/action_model': 1.5249352291532073e-05, 'epoch': 0.29}) +[2026-10-03 15:23:04,164][starVLA.training.train_starvla][INFO] - Step 3957, Loss: {'train/loss': 5.263980710878968e-05, 'train/grad_norm_pre_clip': 0.03841405361890793, 'timing/data': 0.00036207999801263213, 'timing/model': 0.37037085100018885, 'learning_rate/qwen_vl_interface': 1.523046379079689e-06, 'learning_rate/action_model': 1.523046379079689e-05, 'epoch': 0.29}) +[2026-10-03 15:23:04,823][starVLA.training.train_starvla][INFO] - Step 3958, Loss: {'train/loss': 1.1078891475335695e-05, 'train/grad_norm_pre_clip': 0.002409952925518155, 'timing/data': 0.00038918000063858926, 'timing/model': 0.36954443200374953, 'learning_rate/qwen_vl_interface': 1.5211590610160244e-06, 'learning_rate/action_model': 1.5211590610160242e-05, 'epoch': 0.29}) +[2026-10-03 15:23:05,676][starVLA.training.train_starvla][INFO] - Step 3959, Loss: {'train/loss': 0.0021005463786423206, 'train/grad_norm_pre_clip': 4.299173355102539, 'timing/data': 0.00034526998933870345, 'timing/model': 0.37020463100634515, 'learning_rate/qwen_vl_interface': 1.5192732757380193e-06, 'learning_rate/action_model': 1.5192732757380194e-05, 'epoch': 0.29}) +[2026-10-03 15:23:06,311][starVLA.training.train_starvla][INFO] - Step 3960, Loss: {'train/loss': 4.716202965937555e-06, 'train/grad_norm_pre_clip': 0.000815990730188787, 'timing/data': 0.00026713000261224806, 'timing/model': 0.3475265839952044, 'learning_rate/qwen_vl_interface': 1.517389024020849e-06, 'learning_rate/action_model': 1.5173890240208488e-05, 'epoch': 0.29}) +[2026-10-03 15:23:06,969][starVLA.training.train_starvla][INFO] - Step 3961, Loss: {'train/loss': 7.113029278116301e-05, 'train/grad_norm_pre_clip': 0.033125344663858414, 'timing/data': 0.00029350999102462083, 'timing/model': 0.3706316110037733, 'learning_rate/qwen_vl_interface': 1.515506306639057e-06, 'learning_rate/action_model': 1.515506306639057e-05, 'epoch': 0.29}) +[2026-10-03 15:23:07,628][starVLA.training.train_starvla][INFO] - Step 3962, Loss: {'train/loss': 0.1617373526096344, 'train/grad_norm_pre_clip': 136.45358276367188, 'timing/data': 0.0003762499982258305, 'timing/model': 0.36977642100828234, 'learning_rate/qwen_vl_interface': 1.5136251243665589e-06, 'learning_rate/action_model': 1.5136251243665589e-05, 'epoch': 0.29}) +[2026-10-03 15:23:08,285][starVLA.training.train_starvla][INFO] - Step 3963, Loss: {'train/loss': 4.1268802306149155e-05, 'train/grad_norm_pre_clip': 0.018678385764360428, 'timing/data': 0.00026976999652106315, 'timing/model': 0.3691606009961106, 'learning_rate/qwen_vl_interface': 1.5117454779766372e-06, 'learning_rate/action_model': 1.5117454779766371e-05, 'epoch': 0.29}) +[2026-10-03 15:23:09,136][starVLA.training.train_starvla][INFO] - Step 3964, Loss: {'train/loss': 7.703827577643096e-06, 'train/grad_norm_pre_clip': 0.0012194346636533737, 'timing/data': 0.0002899899991462007, 'timing/model': 0.5635672360076569, 'learning_rate/qwen_vl_interface': 1.5098673682419428e-06, 'learning_rate/action_model': 1.5098673682419429e-05, 'epoch': 0.29}) +[2026-10-03 15:23:09,795][starVLA.training.train_starvla][INFO] - Step 3965, Loss: {'train/loss': 0.07267861068248749, 'train/grad_norm_pre_clip': 162.09259033203125, 'timing/data': 0.0003095800057053566, 'timing/model': 0.3704145410010824, 'learning_rate/qwen_vl_interface': 1.5079907959344977e-06, 'learning_rate/action_model': 1.5079907959344978e-05, 'epoch': 0.29}) +[2026-10-03 15:23:10,454][starVLA.training.train_starvla][INFO] - Step 3966, Loss: {'train/loss': 2.0391129510244355e-05, 'train/grad_norm_pre_clip': 0.006735390052199364, 'timing/data': 0.00022277000243775547, 'timing/model': 0.3697698809992289, 'learning_rate/qwen_vl_interface': 1.5061157618256873e-06, 'learning_rate/action_model': 1.5061157618256871e-05, 'epoch': 0.29}) +[2026-10-03 15:23:11,112][starVLA.training.train_starvla][INFO] - Step 3967, Loss: {'train/loss': 4.1283536120317876e-05, 'train/grad_norm_pre_clip': 0.017333926633000374, 'timing/data': 0.00033272999280598015, 'timing/model': 0.3690960909880232, 'learning_rate/qwen_vl_interface': 1.5042422666862675e-06, 'learning_rate/action_model': 1.5042422666862673e-05, 'epoch': 0.29}) +[2026-10-03 15:23:11,771][starVLA.training.train_starvla][INFO] - Step 3968, Loss: {'train/loss': 9.335400136478711e-06, 'train/grad_norm_pre_clip': 0.0031462903134524822, 'timing/data': 0.00021461000142153352, 'timing/model': 0.371374330992694, 'learning_rate/qwen_vl_interface': 1.5023703112863628e-06, 'learning_rate/action_model': 1.5023703112863626e-05, 'epoch': 0.29}) +[2026-10-03 15:23:12,628][starVLA.training.train_starvla][INFO] - Step 3969, Loss: {'train/loss': 1.4923220078344457e-05, 'train/grad_norm_pre_clip': 0.003837299533188343, 'timing/data': 0.0003236499906051904, 'timing/model': 0.5678863450011704, 'learning_rate/qwen_vl_interface': 1.5004998963954629e-06, 'learning_rate/action_model': 1.5004998963954628e-05, 'epoch': 0.29}) +[2026-10-03 15:23:13,287][starVLA.training.train_starvla][INFO] - Step 3970, Loss: {'train/loss': 4.861014167545363e-05, 'train/grad_norm_pre_clip': 0.013398855924606323, 'timing/data': 0.00035973999183624983, 'timing/model': 0.369917980991886, 'learning_rate/qwen_vl_interface': 1.498631022782424e-06, 'learning_rate/action_model': 1.4986310227824241e-05, 'epoch': 0.29}) +[2026-10-03 15:23:13,944][starVLA.training.train_starvla][INFO] - Step 3971, Loss: {'train/loss': 0.003980154637247324, 'train/grad_norm_pre_clip': 9.641525268554688, 'timing/data': 0.0002585199981695041, 'timing/model': 0.3694888909958536, 'learning_rate/qwen_vl_interface': 1.4967636912154707e-06, 'learning_rate/action_model': 1.4967636912154706e-05, 'epoch': 0.29}) +[2026-10-03 15:23:14,581][starVLA.training.train_starvla][INFO] - Step 3972, Loss: {'train/loss': 0.00048677483573555946, 'train/grad_norm_pre_clip': 0.11138829588890076, 'timing/data': 0.000311869996949099, 'timing/model': 0.37095419100660365, 'learning_rate/qwen_vl_interface': 1.4948979024621915e-06, 'learning_rate/action_model': 1.4948979024621914e-05, 'epoch': 0.29}) +[2026-10-03 15:23:15,241][starVLA.training.train_starvla][INFO] - Step 3973, Loss: {'train/loss': 0.2382904440164566, 'train/grad_norm_pre_clip': 55.03557205200195, 'timing/data': 0.0002659899910213426, 'timing/model': 0.37148407100175973, 'learning_rate/qwen_vl_interface': 1.493033657289542e-06, 'learning_rate/action_model': 1.493033657289542e-05, 'epoch': 0.29}) +[2026-10-03 15:23:15,901][starVLA.training.train_starvla][INFO] - Step 3974, Loss: {'train/loss': 0.006690431386232376, 'train/grad_norm_pre_clip': 12.757150650024414, 'timing/data': 0.0002937800018116832, 'timing/model': 0.36992716099484824, 'learning_rate/qwen_vl_interface': 1.4911709564638428e-06, 'learning_rate/action_model': 1.4911709564638427e-05, 'epoch': 0.29}) +[2026-10-03 15:23:16,761][starVLA.training.train_starvla][INFO] - Step 3975, Loss: {'train/loss': 0.3161940574645996, 'train/grad_norm_pre_clip': 250.8439178466797, 'timing/data': 0.0002236700092907995, 'timing/model': 0.3702496210025856, 'learning_rate/qwen_vl_interface': 1.4893098007507817e-06, 'learning_rate/action_model': 1.4893098007507816e-05, 'epoch': 0.29}) +[2026-10-03 15:23:17,397][starVLA.training.train_starvla][INFO] - Step 3976, Loss: {'train/loss': 0.020375847816467285, 'train/grad_norm_pre_clip': 26.26204490661621, 'timing/data': 0.0003233199968235567, 'timing/model': 0.3710949520027498, 'learning_rate/qwen_vl_interface': 1.487450190915406e-06, 'learning_rate/action_model': 1.487450190915406e-05, 'epoch': 0.29}) +[2026-10-03 15:23:18,058][starVLA.training.train_starvla][INFO] - Step 3977, Loss: {'train/loss': 0.0004972019232809544, 'train/grad_norm_pre_clip': 0.5131981372833252, 'timing/data': 0.00025819000438787043, 'timing/model': 0.3713114810088882, 'learning_rate/qwen_vl_interface': 1.4855921277221336e-06, 'learning_rate/action_model': 1.4855921277221336e-05, 'epoch': 0.29}) +[2026-10-03 15:23:18,717][starVLA.training.train_starvla][INFO] - Step 3978, Loss: {'train/loss': 7.518989878008142e-05, 'train/grad_norm_pre_clip': 0.031336382031440735, 'timing/data': 0.00036964999162591994, 'timing/model': 0.3698172110016458, 'learning_rate/qwen_vl_interface': 1.4837356119347423e-06, 'learning_rate/action_model': 1.4837356119347423e-05, 'epoch': 0.29}) +[2026-10-03 15:23:19,376][starVLA.training.train_starvla][INFO] - Step 3979, Loss: {'train/loss': 0.14870890974998474, 'train/grad_norm_pre_clip': 226.26113891601562, 'timing/data': 0.0002662899933056906, 'timing/model': 0.3707188709959155, 'learning_rate/qwen_vl_interface': 1.4818806443163774e-06, 'learning_rate/action_model': 1.4818806443163775e-05, 'epoch': 0.29}) +[2026-10-03 15:23:20,235][starVLA.training.train_starvla][INFO] - Step 3980, Loss: {'train/loss': 2.7169704480911605e-05, 'train/grad_norm_pre_clip': 0.010306615382432938, 'timing/data': 0.0003106499934801832, 'timing/model': 0.37174355100432876, 'learning_rate/qwen_vl_interface': 1.480027225629547e-06, 'learning_rate/action_model': 1.480027225629547e-05, 'epoch': 0.29}) +[2026-10-03 15:23:20,896][starVLA.training.train_starvla][INFO] - Step 3981, Loss: {'train/loss': 0.000931062619201839, 'train/grad_norm_pre_clip': 1.2894281148910522, 'timing/data': 0.0003709399898070842, 'timing/model': 0.37274401099421084, 'learning_rate/qwen_vl_interface': 1.4781753566361182e-06, 'learning_rate/action_model': 1.4781753566361182e-05, 'epoch': 0.29}) +[2026-10-03 15:23:21,556][starVLA.training.train_starvla][INFO] - Step 3982, Loss: {'train/loss': 0.221339151263237, 'train/grad_norm_pre_clip': 152.05174255371094, 'timing/data': 0.00031644999398849905, 'timing/model': 0.3702220010018209, 'learning_rate/qwen_vl_interface': 1.4763250380973255e-06, 'learning_rate/action_model': 1.4763250380973254e-05, 'epoch': 0.29}) +[2026-10-03 15:23:22,191][starVLA.training.train_starvla][INFO] - Step 3983, Loss: {'train/loss': 0.0001039537091855891, 'train/grad_norm_pre_clip': 0.05235302075743675, 'timing/data': 0.00022450000687967986, 'timing/model': 0.34688068499963265, 'learning_rate/qwen_vl_interface': 1.4744762707737662e-06, 'learning_rate/action_model': 1.4744762707737662e-05, 'epoch': 0.29}) +[2026-10-03 15:23:22,849][starVLA.training.train_starvla][INFO] - Step 3984, Loss: {'train/loss': 1.46252459671814e-05, 'train/grad_norm_pre_clip': 0.003011325141415, 'timing/data': 0.00026744999922811985, 'timing/model': 0.3699391510017449, 'learning_rate/qwen_vl_interface': 1.4726290554253977e-06, 'learning_rate/action_model': 1.4726290554253977e-05, 'epoch': 0.29}) +[2026-10-03 15:23:23,704][starVLA.training.train_starvla][INFO] - Step 3985, Loss: {'train/loss': 1.5586259905830957e-05, 'train/grad_norm_pre_clip': 0.004304481204599142, 'timing/data': 0.00024984001356642693, 'timing/model': 0.5660735550045501, 'learning_rate/qwen_vl_interface': 1.4707833928115406e-06, 'learning_rate/action_model': 1.4707833928115405e-05, 'epoch': 0.29}) +[2026-10-03 15:23:24,365][starVLA.training.train_starvla][INFO] - Step 3986, Loss: {'train/loss': 0.014213410206139088, 'train/grad_norm_pre_clip': 38.3292236328125, 'timing/data': 0.0003933300031349063, 'timing/model': 0.36856753200117964, 'learning_rate/qwen_vl_interface': 1.4689392836908775e-06, 'learning_rate/action_model': 1.4689392836908774e-05, 'epoch': 0.29}) +[2026-10-03 15:23:25,023][starVLA.training.train_starvla][INFO] - Step 3987, Loss: {'train/loss': 8.955486009654123e-06, 'train/grad_norm_pre_clip': 0.0023641891311854124, 'timing/data': 0.0002993300004163757, 'timing/model': 0.36940845199569594, 'learning_rate/qwen_vl_interface': 1.4670967288214514e-06, 'learning_rate/action_model': 1.4670967288214513e-05, 'epoch': 0.29}) +[2026-10-03 15:23:25,657][starVLA.training.train_starvla][INFO] - Step 3988, Loss: {'train/loss': 0.0002299538318766281, 'train/grad_norm_pre_clip': 0.12125866860151291, 'timing/data': 0.00030702000367455184, 'timing/model': 0.3473166440089699, 'learning_rate/qwen_vl_interface': 1.4652557289606687e-06, 'learning_rate/action_model': 1.4652557289606685e-05, 'epoch': 0.29}) +[2026-10-03 15:23:26,316][starVLA.training.train_starvla][INFO] - Step 3989, Loss: {'train/loss': 9.871901056612842e-06, 'train/grad_norm_pre_clip': 0.002297972794622183, 'timing/data': 0.00030957999115344137, 'timing/model': 0.3707060110027669, 'learning_rate/qwen_vl_interface': 1.4634162848652918e-06, 'learning_rate/action_model': 1.4634162848652919e-05, 'epoch': 0.29}) +[2026-10-03 15:23:27,171][starVLA.training.train_starvla][INFO] - Step 3990, Loss: {'train/loss': 2.7654741643345915e-05, 'train/grad_norm_pre_clip': 0.006548888515681028, 'timing/data': 0.0003293390036560595, 'timing/model': 0.5669178349926369, 'learning_rate/qwen_vl_interface': 1.4615783972914494e-06, 'learning_rate/action_model': 1.4615783972914493e-05, 'epoch': 0.29}) +[2026-10-03 15:23:27,827][starVLA.training.train_starvla][INFO] - Step 3991, Loss: {'train/loss': 3.605583333410323e-05, 'train/grad_norm_pre_clip': 0.015299820341169834, 'timing/data': 0.00020926000433973968, 'timing/model': 0.36945092100359034, 'learning_rate/qwen_vl_interface': 1.459742066994626e-06, 'learning_rate/action_model': 1.4597420669946261e-05, 'epoch': 0.29}) +[2026-10-03 15:23:28,486][starVLA.training.train_starvla][INFO] - Step 3992, Loss: {'train/loss': 2.8012214897898957e-05, 'train/grad_norm_pre_clip': 0.009645543992519379, 'timing/data': 0.00028967000253032893, 'timing/model': 0.3702793710108381, 'learning_rate/qwen_vl_interface': 1.4579072947296687e-06, 'learning_rate/action_model': 1.4579072947296687e-05, 'epoch': 0.29}) +[2026-10-03 15:23:29,145][starVLA.training.train_starvla][INFO] - Step 3993, Loss: {'train/loss': 0.0003151570854242891, 'train/grad_norm_pre_clip': 0.20049940049648285, 'timing/data': 0.00037400001019705087, 'timing/model': 0.37125701100740116, 'learning_rate/qwen_vl_interface': 1.4560740812507823e-06, 'learning_rate/action_model': 1.4560740812507822e-05, 'epoch': 0.29}) +[2026-10-03 15:23:29,804][starVLA.training.train_starvla][INFO] - Step 3994, Loss: {'train/loss': 1.3552406926464755e-05, 'train/grad_norm_pre_clip': 0.0034330885391682386, 'timing/data': 0.00036843000270891935, 'timing/model': 0.36932440199598204, 'learning_rate/qwen_vl_interface': 1.4542424273115322e-06, 'learning_rate/action_model': 1.454242427311532e-05, 'epoch': 0.29}) +[2026-10-03 15:23:30,462][starVLA.training.train_starvla][INFO] - Step 3995, Loss: {'train/loss': 2.931500057457015e-05, 'train/grad_norm_pre_clip': 0.012231359258294106, 'timing/data': 0.0003580000047804788, 'timing/model': 0.369559801009018, 'learning_rate/qwen_vl_interface': 1.4524123336648419e-06, 'learning_rate/action_model': 1.4524123336648418e-05, 'epoch': 0.29}) +[2026-10-03 15:23:31,318][starVLA.training.train_starvla][INFO] - Step 3996, Loss: {'train/loss': 8.719897596165538e-05, 'train/grad_norm_pre_clip': 0.013704109005630016, 'timing/data': 0.0003205600078217685, 'timing/model': 0.3712196909909835, 'learning_rate/qwen_vl_interface': 1.4505838010629942e-06, 'learning_rate/action_model': 1.4505838010629941e-05, 'epoch': 0.29}) +[2026-10-03 15:23:31,976][starVLA.training.train_starvla][INFO] - Step 3997, Loss: {'train/loss': 5.3653056966140866e-05, 'train/grad_norm_pre_clip': 0.028743209317326546, 'timing/data': 0.0002594499965198338, 'timing/model': 0.3700491509953281, 'learning_rate/qwen_vl_interface': 1.4487568302576284e-06, 'learning_rate/action_model': 1.4487568302576282e-05, 'epoch': 0.29}) +[2026-10-03 15:23:32,634][starVLA.training.train_starvla][INFO] - Step 3998, Loss: {'train/loss': 7.897511750343256e-06, 'train/grad_norm_pre_clip': 0.002043466316536069, 'timing/data': 0.0003312399931019172, 'timing/model': 0.3696022709918907, 'learning_rate/qwen_vl_interface': 1.446931421999746e-06, 'learning_rate/action_model': 1.4469314219997459e-05, 'epoch': 0.29}) +[2026-10-03 15:23:33,290][starVLA.training.train_starvla][INFO] - Step 3999, Loss: {'train/loss': 1.6926887838053517e-05, 'train/grad_norm_pre_clip': 0.0064582545310258865, 'timing/data': 0.00031266998848877847, 'timing/model': 0.36838894100219477, 'learning_rate/qwen_vl_interface': 1.4451075770397004e-06, 'learning_rate/action_model': 1.4451075770397004e-05, 'epoch': 0.29}) +[2026-10-03 15:23:33,948][starVLA.training.train_starvla][INFO] - Step 4000, Loss: {'train/loss': 0.0008819039212539792, 'train/grad_norm_pre_clip': 0.8741279244422913, 'timing/data': 0.00031335999665316194, 'timing/model': 0.37072508099663537, 'learning_rate/qwen_vl_interface': 1.4432852961272068e-06, 'learning_rate/action_model': 1.4432852961272066e-05, 'epoch': 0.29}) +[2026-10-03 15:23:34,799][starVLA.training.train_starvla][INFO] - Step 4001, Loss: {'train/loss': 1.2166459782747552e-05, 'train/grad_norm_pre_clip': 0.0036317764315754175, 'timing/data': 0.00030871000490151346, 'timing/model': 0.5625250960001722, 'learning_rate/qwen_vl_interface': 1.4414645800113353e-06, 'learning_rate/action_model': 1.4414645800113352e-05, 'epoch': 0.29}) +[2026-10-03 15:23:35,457][starVLA.training.train_starvla][INFO] - Step 4002, Loss: {'train/loss': 0.41610628366470337, 'train/grad_norm_pre_clip': 17.17605209350586, 'timing/data': 0.0002621099993120879, 'timing/model': 0.36948006101010833, 'learning_rate/qwen_vl_interface': 1.4396454294405138e-06, 'learning_rate/action_model': 1.4396454294405136e-05, 'epoch': 0.29}) +[2026-10-03 15:23:36,114][starVLA.training.train_starvla][INFO] - Step 4003, Loss: {'train/loss': 1.7522816051496193e-05, 'train/grad_norm_pre_clip': 0.01105947233736515, 'timing/data': 0.0003157600003760308, 'timing/model': 0.3693226009927457, 'learning_rate/qwen_vl_interface': 1.4378278451625271e-06, 'learning_rate/action_model': 1.4378278451625271e-05, 'epoch': 0.29}) +[2026-10-03 15:23:36,773][starVLA.training.train_starvla][INFO] - Step 4004, Loss: {'train/loss': 0.0003065364435315132, 'train/grad_norm_pre_clip': 0.48815783858299255, 'timing/data': 0.00031540000054519624, 'timing/model': 0.37142966999090277, 'learning_rate/qwen_vl_interface': 1.4360118279245143e-06, 'learning_rate/action_model': 1.4360118279245144e-05, 'epoch': 0.29}) +[2026-10-03 15:23:37,431][starVLA.training.train_starvla][INFO] - Step 4005, Loss: {'train/loss': 0.00010872568236663938, 'train/grad_norm_pre_clip': 0.1975077986717224, 'timing/data': 0.00020725000649690628, 'timing/model': 0.3707766310108127, 'learning_rate/qwen_vl_interface': 1.434197378472973e-06, 'learning_rate/action_model': 1.4341973784729729e-05, 'epoch': 0.29}) +[2026-10-03 15:23:38,288][starVLA.training.train_starvla][INFO] - Step 4006, Loss: {'train/loss': 1.536991294415202e-05, 'train/grad_norm_pre_clip': 0.005342170130461454, 'timing/data': 0.0002689199900487438, 'timing/model': 0.5682172839879058, 'learning_rate/qwen_vl_interface': 1.4323844975537535e-06, 'learning_rate/action_model': 1.4323844975537534e-05, 'epoch': 0.29}) +[2026-10-03 15:23:38,945][starVLA.training.train_starvla][INFO] - Step 4007, Loss: {'train/loss': 9.931335625879001e-06, 'train/grad_norm_pre_clip': 0.003140798769891262, 'timing/data': 0.00021430999913718551, 'timing/model': 0.3698122210043948, 'learning_rate/qwen_vl_interface': 1.4305731859120628e-06, 'learning_rate/action_model': 1.4305731859120628e-05, 'epoch': 0.29}) +[2026-10-03 15:23:39,603][starVLA.training.train_starvla][INFO] - Step 4008, Loss: {'train/loss': 0.0001350574311800301, 'train/grad_norm_pre_clip': 0.06721365451812744, 'timing/data': 0.00026431999867781997, 'timing/model': 0.3707908910000697, 'learning_rate/qwen_vl_interface': 1.4287634442924647e-06, 'learning_rate/action_model': 1.4287634442924646e-05, 'epoch': 0.29}) +[2026-10-03 15:23:40,262][starVLA.training.train_starvla][INFO] - Step 4009, Loss: {'train/loss': 4.970124427927658e-05, 'train/grad_norm_pre_clip': 0.03210805729031563, 'timing/data': 0.00025445000210311264, 'timing/model': 0.37047988099220674, 'learning_rate/qwen_vl_interface': 1.4269552734388725e-06, 'learning_rate/action_model': 1.4269552734388724e-05, 'epoch': 0.29}) +[2026-10-03 15:23:40,920][starVLA.training.train_starvla][INFO] - Step 4010, Loss: {'train/loss': 5.2272567700129e-05, 'train/grad_norm_pre_clip': 0.026331746950745583, 'timing/data': 0.00022383999021258205, 'timing/model': 0.3694356810010504, 'learning_rate/qwen_vl_interface': 1.4251486740945584e-06, 'learning_rate/action_model': 1.4251486740945583e-05, 'epoch': 0.29}) +[2026-10-03 15:23:41,578][starVLA.training.train_starvla][INFO] - Step 4011, Loss: {'train/loss': 0.19145655632019043, 'train/grad_norm_pre_clip': 423.9978332519531, 'timing/data': 0.0002561799919931218, 'timing/model': 0.3695937720040092, 'learning_rate/qwen_vl_interface': 1.423343647002147e-06, 'learning_rate/action_model': 1.423343647002147e-05, 'epoch': 0.29}) +[2026-10-03 15:23:42,426][starVLA.training.train_starvla][INFO] - Step 4012, Loss: {'train/loss': 9.379956281918567e-06, 'train/grad_norm_pre_clip': 0.00334891676902771, 'timing/data': 0.0004206999874440953, 'timing/model': 0.3721663210017141, 'learning_rate/qwen_vl_interface': 1.421540192903617e-06, 'learning_rate/action_model': 1.421540192903617e-05, 'epoch': 0.29}) +[2026-10-03 15:23:43,085][starVLA.training.train_starvla][INFO] - Step 4013, Loss: {'train/loss': 3.182328146067448e-05, 'train/grad_norm_pre_clip': 0.014215550385415554, 'timing/data': 0.0003458499995758757, 'timing/model': 0.37062249099835753, 'learning_rate/qwen_vl_interface': 1.4197383125402997e-06, 'learning_rate/action_model': 1.4197383125402997e-05, 'epoch': 0.29}) +[2026-10-03 15:23:43,743][starVLA.training.train_starvla][INFO] - Step 4014, Loss: {'train/loss': 9.521637366560753e-06, 'train/grad_norm_pre_clip': 0.003146663075312972, 'timing/data': 0.0003167799877701327, 'timing/model': 0.3691873510106234, 'learning_rate/qwen_vl_interface': 1.4179380066528807e-06, 'learning_rate/action_model': 1.4179380066528807e-05, 'epoch': 0.29}) +[2026-10-03 15:23:44,400][starVLA.training.train_starvla][INFO] - Step 4015, Loss: {'train/loss': 0.0032423699740320444, 'train/grad_norm_pre_clip': 2.71956467628479, 'timing/data': 0.0002938800025731325, 'timing/model': 0.3692995910096215, 'learning_rate/qwen_vl_interface': 1.4161392759813984e-06, 'learning_rate/action_model': 1.4161392759813983e-05, 'epoch': 0.29}) +[2026-10-03 15:23:45,058][starVLA.training.train_starvla][INFO] - Step 4016, Loss: {'train/loss': 0.004188219550997019, 'train/grad_norm_pre_clip': 31.81638526916504, 'timing/data': 0.00033008999889716506, 'timing/model': 0.3706382909876993, 'learning_rate/qwen_vl_interface': 1.4143421212652411e-06, 'learning_rate/action_model': 1.414342121265241e-05, 'epoch': 0.29}) +[2026-10-03 15:23:45,917][starVLA.training.train_starvla][INFO] - Step 4017, Loss: {'train/loss': 0.0052839601412415504, 'train/grad_norm_pre_clip': 9.848244667053223, 'timing/data': 0.000335429998813197, 'timing/model': 0.37043691099097487, 'learning_rate/qwen_vl_interface': 1.4125465432431529e-06, 'learning_rate/action_model': 1.4125465432431529e-05, 'epoch': 0.29}) +[2026-10-03 15:23:46,576][starVLA.training.train_starvla][INFO] - Step 4018, Loss: {'train/loss': 1.876677924883552e-05, 'train/grad_norm_pre_clip': 0.0067346771247684956, 'timing/data': 0.00040084999636746943, 'timing/model': 0.3695217810018221, 'learning_rate/qwen_vl_interface': 1.4107525426532273e-06, 'learning_rate/action_model': 1.4107525426532272e-05, 'epoch': 0.29}) +[2026-10-03 15:23:47,209][starVLA.training.train_starvla][INFO] - Step 4019, Loss: {'train/loss': 0.0013092750450596213, 'train/grad_norm_pre_clip': 2.1401920318603516, 'timing/data': 0.00034312999923713505, 'timing/model': 0.3687342020130018, 'learning_rate/qwen_vl_interface': 1.4089601202329116e-06, 'learning_rate/action_model': 1.4089601202329114e-05, 'epoch': 0.29}) +[2026-10-03 15:23:47,868][starVLA.training.train_starvla][INFO] - Step 4020, Loss: {'train/loss': 0.00018399742839392275, 'train/grad_norm_pre_clip': 0.030957533046603203, 'timing/data': 0.00033614000130910426, 'timing/model': 0.37062721099937335, 'learning_rate/qwen_vl_interface': 1.407169276719e-06, 'learning_rate/action_model': 1.4071692767190001e-05, 'epoch': 0.29}) +[2026-10-03 15:23:48,526][starVLA.training.train_starvla][INFO] - Step 4021, Loss: {'train/loss': 2.3198528651846573e-05, 'train/grad_norm_pre_clip': 0.016816124320030212, 'timing/data': 0.00025519999326206744, 'timing/model': 0.3708114009932615, 'learning_rate/qwen_vl_interface': 1.4053800128476426e-06, 'learning_rate/action_model': 1.4053800128476427e-05, 'epoch': 0.29}) +[2026-10-03 15:23:49,374][starVLA.training.train_starvla][INFO] - Step 4022, Loss: {'train/loss': 0.005041885655373335, 'train/grad_norm_pre_clip': 16.045255661010742, 'timing/data': 0.00025872999685816467, 'timing/model': 0.5587508259923197, 'learning_rate/qwen_vl_interface': 1.403592329354338e-06, 'learning_rate/action_model': 1.403592329354338e-05, 'epoch': 0.29}) +[2026-10-03 15:23:50,008][starVLA.training.train_starvla][INFO] - Step 4023, Loss: {'train/loss': 2.2759373678127304e-05, 'train/grad_norm_pre_clip': 0.009876178577542305, 'timing/data': 0.0002861200046027079, 'timing/model': 0.3463641039998038, 'learning_rate/qwen_vl_interface': 1.4018062269739354e-06, 'learning_rate/action_model': 1.4018062269739355e-05, 'epoch': 0.29}) +[2026-10-03 15:23:50,666][starVLA.training.train_starvla][INFO] - Step 4024, Loss: {'train/loss': 0.22409673035144806, 'train/grad_norm_pre_clip': 96.06226348876953, 'timing/data': 0.0003331100015202537, 'timing/model': 0.3706294609874021, 'learning_rate/qwen_vl_interface': 1.4000217064406339e-06, 'learning_rate/action_model': 1.4000217064406337e-05, 'epoch': 0.29}) +[2026-10-03 15:23:51,325][starVLA.training.train_starvla][INFO] - Step 4025, Loss: {'train/loss': 1.0922059118456673e-05, 'train/grad_norm_pre_clip': 0.005958071444183588, 'timing/data': 0.0003219700010959059, 'timing/model': 0.3704330309992656, 'learning_rate/qwen_vl_interface': 1.3982387684879836e-06, 'learning_rate/action_model': 1.3982387684879836e-05, 'epoch': 0.29}) +[2026-10-03 15:23:51,961][starVLA.training.train_starvla][INFO] - Step 4026, Loss: {'train/loss': 0.02362980507314205, 'train/grad_norm_pre_clip': 69.66397857666016, 'timing/data': 0.0003795600059675053, 'timing/model': 0.37013953099085484, 'learning_rate/qwen_vl_interface': 1.3964574138488804e-06, 'learning_rate/action_model': 1.3964574138488803e-05, 'epoch': 0.29}) +[2026-10-03 15:23:52,618][starVLA.training.train_starvla][INFO] - Step 4027, Loss: {'train/loss': 4.3220745283178985e-05, 'train/grad_norm_pre_clip': 0.014525284990668297, 'timing/data': 0.0003134599974146113, 'timing/model': 0.36959494100301526, 'learning_rate/qwen_vl_interface': 1.3946776432555743e-06, 'learning_rate/action_model': 1.3946776432555741e-05, 'epoch': 0.29}) +[2026-10-03 15:23:53,476][starVLA.training.train_starvla][INFO] - Step 4028, Loss: {'train/loss': 7.323825229832437e-06, 'train/grad_norm_pre_clip': 0.001724684378132224, 'timing/data': 0.00025039000320248306, 'timing/model': 0.370761971003958, 'learning_rate/qwen_vl_interface': 1.3928994574396595e-06, 'learning_rate/action_model': 1.3928994574396594e-05, 'epoch': 0.29}) +[2026-10-03 15:23:54,112][starVLA.training.train_starvla][INFO] - Step 4029, Loss: {'train/loss': 1.2486359992180951e-05, 'train/grad_norm_pre_clip': 0.005416693165898323, 'timing/data': 0.00038055000186432153, 'timing/model': 0.34759524499531835, 'learning_rate/qwen_vl_interface': 1.391122857132083e-06, 'learning_rate/action_model': 1.3911228571320831e-05, 'epoch': 0.29}) +[2026-10-03 15:23:54,770][starVLA.training.train_starvla][INFO] - Step 4030, Loss: {'train/loss': 0.3718430697917938, 'train/grad_norm_pre_clip': 98.28218078613281, 'timing/data': 0.00036102000740356743, 'timing/model': 0.36929154199606273, 'learning_rate/qwen_vl_interface': 1.3893478430631376e-06, 'learning_rate/action_model': 1.3893478430631376e-05, 'epoch': 0.29}) +[2026-10-03 15:23:55,427][starVLA.training.train_starvla][INFO] - Step 4031, Loss: {'train/loss': 3.114328819719958e-06, 'train/grad_norm_pre_clip': 0.0009853991214185953, 'timing/data': 0.00021180999465286732, 'timing/model': 0.369737211993197, 'learning_rate/qwen_vl_interface': 1.3875744159624652e-06, 'learning_rate/action_model': 1.3875744159624652e-05, 'epoch': 0.3}) +[2026-10-03 15:23:56,087][starVLA.training.train_starvla][INFO] - Step 4032, Loss: {'train/loss': 0.3228902518749237, 'train/grad_norm_pre_clip': 72.93083190917969, 'timing/data': 0.0003267400024924427, 'timing/model': 0.37122861099487636, 'learning_rate/qwen_vl_interface': 1.3858025765590525e-06, 'learning_rate/action_model': 1.3858025765590525e-05, 'epoch': 0.3}) +[2026-10-03 15:23:56,942][starVLA.training.train_starvla][INFO] - Step 4033, Loss: {'train/loss': 7.606965482409578e-06, 'train/grad_norm_pre_clip': 0.0019547862466424704, 'timing/data': 0.00021916998957749456, 'timing/model': 0.3709886110009393, 'learning_rate/qwen_vl_interface': 1.384032325581237e-06, 'learning_rate/action_model': 1.3840323255812371e-05, 'epoch': 0.3}) +[2026-10-03 15:23:57,577][starVLA.training.train_starvla][INFO] - Step 4034, Loss: {'train/loss': 3.933881998818833e-06, 'train/grad_norm_pre_clip': 0.0010043852962553501, 'timing/data': 0.0002828500000759959, 'timing/model': 0.3461560050054686, 'learning_rate/qwen_vl_interface': 1.3822636637567025e-06, 'learning_rate/action_model': 1.3822636637567024e-05, 'epoch': 0.3}) +[2026-10-03 15:23:58,234][starVLA.training.train_starvla][INFO] - Step 4035, Loss: {'train/loss': 0.0017170265782624483, 'train/grad_norm_pre_clip': 3.0860211849212646, 'timing/data': 0.00026617999537847936, 'timing/model': 0.3692888520017732, 'learning_rate/qwen_vl_interface': 1.3804965918124788e-06, 'learning_rate/action_model': 1.3804965918124787e-05, 'epoch': 0.3}) +[2026-10-03 15:23:58,892][starVLA.training.train_starvla][INFO] - Step 4036, Loss: {'train/loss': 3.189057679264806e-05, 'train/grad_norm_pre_clip': 0.010843353345990181, 'timing/data': 0.00030251000134740025, 'timing/model': 0.3708196210063761, 'learning_rate/qwen_vl_interface': 1.3787311104749418e-06, 'learning_rate/action_model': 1.3787311104749418e-05, 'epoch': 0.3}) +[2026-10-03 15:23:59,551][starVLA.training.train_starvla][INFO] - Step 4037, Loss: {'train/loss': 1.8729635485215113e-05, 'train/grad_norm_pre_clip': 0.005989359691739082, 'timing/data': 0.00031377999403048307, 'timing/model': 0.3705762020108523, 'learning_rate/qwen_vl_interface': 1.3769672204698149e-06, 'learning_rate/action_model': 1.3769672204698147e-05, 'epoch': 0.3}) +[2026-10-03 15:24:00,400][starVLA.training.train_starvla][INFO] - Step 4038, Loss: {'train/loss': 1.1317313692416064e-05, 'train/grad_norm_pre_clip': 0.004008097108453512, 'timing/data': 0.00031041000329423696, 'timing/model': 0.560101205992396, 'learning_rate/qwen_vl_interface': 1.3752049225221664e-06, 'learning_rate/action_model': 1.3752049225221663e-05, 'epoch': 0.3}) +[2026-10-03 15:24:01,034][starVLA.training.train_starvla][INFO] - Step 4039, Loss: {'train/loss': 8.426390195381828e-06, 'train/grad_norm_pre_clip': 0.004821707960218191, 'timing/data': 0.0002086000022245571, 'timing/model': 0.3692279620008776, 'learning_rate/qwen_vl_interface': 1.3734442173564103e-06, 'learning_rate/action_model': 1.3734442173564102e-05, 'epoch': 0.3}) +[2026-10-03 15:24:01,693][starVLA.training.train_starvla][INFO] - Step 4040, Loss: {'train/loss': 0.000318941252771765, 'train/grad_norm_pre_clip': 0.39376288652420044, 'timing/data': 0.000334739001118578, 'timing/model': 0.37082301201007795, 'learning_rate/qwen_vl_interface': 1.3716851056963062e-06, 'learning_rate/action_model': 1.3716851056963062e-05, 'epoch': 0.3}) +[2026-10-03 15:24:02,351][starVLA.training.train_starvla][INFO] - Step 4041, Loss: {'train/loss': 2.5629938136262354e-06, 'train/grad_norm_pre_clip': 0.0004640237893909216, 'timing/data': 0.0002611899981275201, 'timing/model': 0.3710698609938845, 'learning_rate/qwen_vl_interface': 1.3699275882649597e-06, 'learning_rate/action_model': 1.3699275882649596e-05, 'epoch': 0.3}) +[2026-10-03 15:24:02,987][starVLA.training.train_starvla][INFO] - Step 4042, Loss: {'train/loss': 0.0872107520699501, 'train/grad_norm_pre_clip': 141.62091064453125, 'timing/data': 0.0003208799898857251, 'timing/model': 0.3468947939982172, 'learning_rate/qwen_vl_interface': 1.368171665784817e-06, 'learning_rate/action_model': 1.3681716657848171e-05, 'epoch': 0.3}) +[2026-10-03 15:24:03,845][starVLA.training.train_starvla][INFO] - Step 4043, Loss: {'train/loss': 0.7872939109802246, 'train/grad_norm_pre_clip': 42.00762176513672, 'timing/data': 0.00033853000786621124, 'timing/model': 0.570883264997974, 'learning_rate/qwen_vl_interface': 1.3664173389776735e-06, 'learning_rate/action_model': 1.3664173389776735e-05, 'epoch': 0.3}) +[2026-10-03 15:24:04,504][starVLA.training.train_starvla][INFO] - Step 4044, Loss: {'train/loss': 0.004971260204911232, 'train/grad_norm_pre_clip': 8.887693405151367, 'timing/data': 0.0002504699950804934, 'timing/model': 0.3712665810016915, 'learning_rate/qwen_vl_interface': 1.3646646085646658e-06, 'learning_rate/action_model': 1.3646646085646659e-05, 'epoch': 0.3}) +[2026-10-03 15:24:05,141][starVLA.training.train_starvla][INFO] - Step 4045, Loss: {'train/loss': 0.0005383717361837626, 'train/grad_norm_pre_clip': 0.461470365524292, 'timing/data': 0.0003054900007555261, 'timing/model': 0.3703575209947303, 'learning_rate/qwen_vl_interface': 1.3629134752662756e-06, 'learning_rate/action_model': 1.3629134752662756e-05, 'epoch': 0.3}) +[2026-10-03 15:24:05,800][starVLA.training.train_starvla][INFO] - Step 4046, Loss: {'train/loss': 4.2542628762021195e-06, 'train/grad_norm_pre_clip': 0.001042005605995655, 'timing/data': 0.00028377000126056373, 'timing/model': 0.36986909099505283, 'learning_rate/qwen_vl_interface': 1.3611639398023282e-06, 'learning_rate/action_model': 1.3611639398023283e-05, 'epoch': 0.3}) +[2026-10-03 15:24:06,456][starVLA.training.train_starvla][INFO] - Step 4047, Loss: {'train/loss': 1.050511946232291e-05, 'train/grad_norm_pre_clip': 0.0037988866679370403, 'timing/data': 0.00021186000958550721, 'timing/model': 0.36916960099188145, 'learning_rate/qwen_vl_interface': 1.35941600289199e-06, 'learning_rate/action_model': 1.35941600289199e-05, 'epoch': 0.3}) +[2026-10-03 15:24:07,115][starVLA.training.train_starvla][INFO] - Step 4048, Loss: {'train/loss': 5.118504759593634e-06, 'train/grad_norm_pre_clip': 0.0017015805933624506, 'timing/data': 0.0002610200026538223, 'timing/model': 0.37116955099918414, 'learning_rate/qwen_vl_interface': 1.3576696652537735e-06, 'learning_rate/action_model': 1.3576696652537736e-05, 'epoch': 0.3}) +[2026-10-03 15:24:07,976][starVLA.training.train_starvla][INFO] - Step 4049, Loss: {'train/loss': 0.0007275138050317764, 'train/grad_norm_pre_clip': 0.9208941459655762, 'timing/data': 0.00027057000261265785, 'timing/model': 0.3714287410111865, 'learning_rate/qwen_vl_interface': 1.355924927605531e-06, 'learning_rate/action_model': 1.3559249276055311e-05, 'epoch': 0.3}) +[2026-10-03 15:24:08,612][starVLA.training.train_starvla][INFO] - Step 4050, Loss: {'train/loss': 1.992081342905294e-05, 'train/grad_norm_pre_clip': 0.009449699893593788, 'timing/data': 0.0003701299865497276, 'timing/model': 0.36967964099312667, 'learning_rate/qwen_vl_interface': 1.3541817906644591e-06, 'learning_rate/action_model': 1.3541817906644591e-05, 'epoch': 0.3}) +[2026-10-03 15:24:09,269][starVLA.training.train_starvla][INFO] - Step 4051, Loss: {'train/loss': 3.439457941567525e-05, 'train/grad_norm_pre_clip': 0.01957240141928196, 'timing/data': 0.0003753400087589398, 'timing/model': 0.36920725100208074, 'learning_rate/qwen_vl_interface': 1.3524402551470955e-06, 'learning_rate/action_model': 1.3524402551470956e-05, 'epoch': 0.3}) +[2026-10-03 15:24:09,927][starVLA.training.train_starvla][INFO] - Step 4052, Loss: {'train/loss': 0.000511612743139267, 'train/grad_norm_pre_clip': 0.7766914963722229, 'timing/data': 0.0002604100009193644, 'timing/model': 0.37045556100201793, 'learning_rate/qwen_vl_interface': 1.3507003217693196e-06, 'learning_rate/action_model': 1.3507003217693196e-05, 'epoch': 0.3}) +[2026-10-03 15:24:10,586][starVLA.training.train_starvla][INFO] - Step 4053, Loss: {'train/loss': 0.0001555665221530944, 'train/grad_norm_pre_clip': 0.06432389467954636, 'timing/data': 0.0003097400040132925, 'timing/model': 0.37052357000357006, 'learning_rate/qwen_vl_interface': 1.3489619912463515e-06, 'learning_rate/action_model': 1.3489619912463514e-05, 'epoch': 0.3}) +[2026-10-03 15:24:11,443][starVLA.training.train_starvla][INFO] - Step 4054, Loss: {'train/loss': 3.3222731872228906e-05, 'train/grad_norm_pre_clip': 0.025032589212059975, 'timing/data': 0.00036558999272529036, 'timing/model': 0.5677412340010051, 'learning_rate/qwen_vl_interface': 1.3472252642927534e-06, 'learning_rate/action_model': 1.3472252642927533e-05, 'epoch': 0.3}) +[2026-10-03 15:24:12,099][starVLA.training.train_starvla][INFO] - Step 4055, Loss: {'train/loss': 3.278245458204765e-06, 'train/grad_norm_pre_clip': 0.0006986365769989789, 'timing/data': 0.00024819000100251287, 'timing/model': 0.3689840209990507, 'learning_rate/qwen_vl_interface': 1.3454901416224284e-06, 'learning_rate/action_model': 1.3454901416224284e-05, 'epoch': 0.3}) +[2026-10-03 15:24:12,759][starVLA.training.train_starvla][INFO] - Step 4056, Loss: {'train/loss': 5.457720180856995e-05, 'train/grad_norm_pre_clip': 0.017047805711627007, 'timing/data': 0.00021658001060131937, 'timing/model': 0.3712677310104482, 'learning_rate/qwen_vl_interface': 1.34375662394862e-06, 'learning_rate/action_model': 1.34375662394862e-05, 'epoch': 0.3}) +[2026-10-03 15:24:13,417][starVLA.training.train_starvla][INFO] - Step 4057, Loss: {'train/loss': 1.5846786482143216e-05, 'train/grad_norm_pre_clip': 0.005137182772159576, 'timing/data': 0.0003245900006731972, 'timing/model': 0.37035722100699786, 'learning_rate/qwen_vl_interface': 1.3420247119839116e-06, 'learning_rate/action_model': 1.3420247119839117e-05, 'epoch': 0.3}) +[2026-10-03 15:24:14,076][starVLA.training.train_starvla][INFO] - Step 4058, Loss: {'train/loss': 2.784000571409706e-05, 'train/grad_norm_pre_clip': 0.02143797092139721, 'timing/data': 0.0003738899977179244, 'timing/model': 0.3697338009951636, 'learning_rate/qwen_vl_interface': 1.340294406440227e-06, 'learning_rate/action_model': 1.340294406440227e-05, 'epoch': 0.3}) +[2026-10-03 15:24:14,942][starVLA.training.train_starvla][INFO] - Step 4059, Loss: {'train/loss': 0.7864346504211426, 'train/grad_norm_pre_clip': 168.4678497314453, 'timing/data': 0.00034073900314979255, 'timing/model': 0.5786235550040146, 'learning_rate/qwen_vl_interface': 1.3385657080288297e-06, 'learning_rate/action_model': 1.3385657080288297e-05, 'epoch': 0.3}) +[2026-10-03 15:24:15,600][starVLA.training.train_starvla][INFO] - Step 4060, Loss: {'train/loss': 3.3676406019367278e-06, 'train/grad_norm_pre_clip': 0.0015152120031416416, 'timing/data': 0.00030186999356374145, 'timing/model': 0.36880466200818773, 'learning_rate/qwen_vl_interface': 1.336838617460322e-06, 'learning_rate/action_model': 1.336838617460322e-05, 'epoch': 0.3}) +[2026-10-03 15:24:16,257][starVLA.training.train_starvla][INFO] - Step 4061, Loss: {'train/loss': 5.653765401802957e-05, 'train/grad_norm_pre_clip': 0.03017374314367771, 'timing/data': 0.00025720999110490084, 'timing/model': 0.3689370619977126, 'learning_rate/qwen_vl_interface': 1.335113135444646e-06, 'learning_rate/action_model': 1.335113135444646e-05, 'epoch': 0.3}) +[2026-10-03 15:24:16,914][starVLA.training.train_starvla][INFO] - Step 4062, Loss: {'train/loss': 6.385058441082947e-06, 'train/grad_norm_pre_clip': 0.002448927378281951, 'timing/data': 0.00033605001226533204, 'timing/model': 0.3692409819923341, 'learning_rate/qwen_vl_interface': 1.3333892626910823e-06, 'learning_rate/action_model': 1.3333892626910823e-05, 'epoch': 0.3}) +[2026-10-03 15:24:17,570][starVLA.training.train_starvla][INFO] - Step 4063, Loss: {'train/loss': 5.5562719353474677e-05, 'train/grad_norm_pre_clip': 0.02896098792552948, 'timing/data': 0.0002330290008103475, 'timing/model': 0.36924960199394263, 'learning_rate/qwen_vl_interface': 1.3316669999082486e-06, 'learning_rate/action_model': 1.3316669999082484e-05, 'epoch': 0.3}) +[2026-10-03 15:24:18,205][starVLA.training.train_starvla][INFO] - Step 4064, Loss: {'train/loss': 8.448658263660036e-06, 'train/grad_norm_pre_clip': 0.0043400186114013195, 'timing/data': 0.00025213000481016934, 'timing/model': 0.3463782850012649, 'learning_rate/qwen_vl_interface': 1.3299463478041024e-06, 'learning_rate/action_model': 1.3299463478041024e-05, 'epoch': 0.3}) +[2026-10-03 15:24:19,060][starVLA.training.train_starvla][INFO] - Step 4065, Loss: {'train/loss': 0.009765983559191227, 'train/grad_norm_pre_clip': 16.614389419555664, 'timing/data': 0.0002681500045582652, 'timing/model': 0.3690692609961843, 'learning_rate/qwen_vl_interface': 1.3282273070859396e-06, 'learning_rate/action_model': 1.3282273070859396e-05, 'epoch': 0.3}) +[2026-10-03 15:24:19,717][starVLA.training.train_starvla][INFO] - Step 4066, Loss: {'train/loss': 5.468662038765615e-06, 'train/grad_norm_pre_clip': 0.001536053721792996, 'timing/data': 0.0003264099941588938, 'timing/model': 0.3690044510003645, 'learning_rate/qwen_vl_interface': 1.3265098784603924e-06, 'learning_rate/action_model': 1.3265098784603924e-05, 'epoch': 0.3}) +[2026-10-03 15:24:20,373][starVLA.training.train_starvla][INFO] - Step 4067, Loss: {'train/loss': 0.00016849071835167706, 'train/grad_norm_pre_clip': 0.027781015262007713, 'timing/data': 0.0002100200072163716, 'timing/model': 0.3692366619943641, 'learning_rate/qwen_vl_interface': 1.324794062633431e-06, 'learning_rate/action_model': 1.324794062633431e-05, 'epoch': 0.3}) +[2026-10-03 15:24:21,031][starVLA.training.train_starvla][INFO] - Step 4068, Loss: {'train/loss': 0.00042079162085428834, 'train/grad_norm_pre_clip': 0.2354343682527542, 'timing/data': 0.0003298800002085045, 'timing/model': 0.36966813101025764, 'learning_rate/qwen_vl_interface': 1.323079860310363e-06, 'learning_rate/action_model': 1.3230798603103628e-05, 'epoch': 0.3}) +[2026-10-03 15:24:21,691][starVLA.training.train_starvla][INFO] - Step 4069, Loss: {'train/loss': 0.003401687368750572, 'train/grad_norm_pre_clip': 5.415198802947998, 'timing/data': 0.00028315000236034393, 'timing/model': 0.37045928099541925, 'learning_rate/qwen_vl_interface': 1.3213672721958309e-06, 'learning_rate/action_model': 1.321367272195831e-05, 'epoch': 0.3}) +[2026-10-03 15:24:22,545][starVLA.training.train_starvla][INFO] - Step 4070, Loss: {'train/loss': 0.00010238824324915186, 'train/grad_norm_pre_clip': 0.07886108756065369, 'timing/data': 0.0002490500046405941, 'timing/model': 0.567030906007858, 'learning_rate/qwen_vl_interface': 1.3196562989938175e-06, 'learning_rate/action_model': 1.3196562989938174e-05, 'epoch': 0.3}) +[2026-10-03 15:24:23,201][starVLA.training.train_starvla][INFO] - Step 4071, Loss: {'train/loss': 1.4311910490505397e-05, 'train/grad_norm_pre_clip': 0.005081893410533667, 'timing/data': 0.00020556000526994467, 'timing/model': 0.3695609720016364, 'learning_rate/qwen_vl_interface': 1.3179469414076357e-06, 'learning_rate/action_model': 1.3179469414076356e-05, 'epoch': 0.3}) +[2026-10-03 15:24:23,860][starVLA.training.train_starvla][INFO] - Step 4072, Loss: {'train/loss': 1.2546199286589399e-05, 'train/grad_norm_pre_clip': 0.005931973457336426, 'timing/data': 0.00033214999712072313, 'timing/model': 0.37013626100088004, 'learning_rate/qwen_vl_interface': 1.3162392001399401e-06, 'learning_rate/action_model': 1.3162392001399401e-05, 'epoch': 0.3}) +[2026-10-03 15:24:24,518][starVLA.training.train_starvla][INFO] - Step 4073, Loss: {'train/loss': 0.0001139691230491735, 'train/grad_norm_pre_clip': 0.05725596100091934, 'timing/data': 0.00033101000008173287, 'timing/model': 0.3694525019964203, 'learning_rate/qwen_vl_interface': 1.3145330758927182e-06, 'learning_rate/action_model': 1.3145330758927182e-05, 'epoch': 0.3}) +[2026-10-03 15:24:25,174][starVLA.training.train_starvla][INFO] - Step 4074, Loss: {'train/loss': 0.00013227970339357853, 'train/grad_norm_pre_clip': 0.07437954843044281, 'timing/data': 0.00035234000824857503, 'timing/model': 0.36932632100069895, 'learning_rate/qwen_vl_interface': 1.3128285693672935e-06, 'learning_rate/action_model': 1.3128285693672934e-05, 'epoch': 0.3}) +[2026-10-03 15:24:25,831][starVLA.training.train_starvla][INFO] - Step 4075, Loss: {'train/loss': 0.544410765171051, 'train/grad_norm_pre_clip': 45.29621887207031, 'timing/data': 0.0003195099998265505, 'timing/model': 0.36917825099953916, 'learning_rate/qwen_vl_interface': 1.311125681264325e-06, 'learning_rate/action_model': 1.311125681264325e-05, 'epoch': 0.3}) +[2026-10-03 15:24:26,684][starVLA.training.train_starvla][INFO] - Step 4076, Loss: {'train/loss': 0.00011974639346590266, 'train/grad_norm_pre_clip': 0.0525357723236084, 'timing/data': 0.0003728099982254207, 'timing/model': 0.36877639099839143, 'learning_rate/qwen_vl_interface': 1.309424412283803e-06, 'learning_rate/action_model': 1.3094244122838029e-05, 'epoch': 0.3}) +[2026-10-03 15:24:27,341][starVLA.training.train_starvla][INFO] - Step 4077, Loss: {'train/loss': 1.0318557542632334e-05, 'train/grad_norm_pre_clip': 0.008561156690120697, 'timing/data': 0.00039636000292375684, 'timing/model': 0.3691872719937237, 'learning_rate/qwen_vl_interface': 1.307724763125057e-06, 'learning_rate/action_model': 1.307724763125057e-05, 'epoch': 0.3}) +[2026-10-03 15:24:27,975][starVLA.training.train_starvla][INFO] - Step 4078, Loss: {'train/loss': 0.00020631107327062637, 'train/grad_norm_pre_clip': 0.22294270992279053, 'timing/data': 0.00026944999990519136, 'timing/model': 0.3464299440092873, 'learning_rate/qwen_vl_interface': 1.3060267344867478e-06, 'learning_rate/action_model': 1.3060267344867478e-05, 'epoch': 0.3}) +[2026-10-03 15:24:28,630][starVLA.training.train_starvla][INFO] - Step 4079, Loss: {'train/loss': 4.649134098144714e-06, 'train/grad_norm_pre_clip': 0.0011419372167438269, 'timing/data': 0.000338979996740818, 'timing/model': 0.368661812011851, 'learning_rate/qwen_vl_interface': 1.3043303270668715e-06, 'learning_rate/action_model': 1.3043303270668713e-05, 'epoch': 0.3}) +[2026-10-03 15:24:29,288][starVLA.training.train_starvla][INFO] - Step 4080, Loss: {'train/loss': 4.380313112051226e-05, 'train/grad_norm_pre_clip': 0.03177047148346901, 'timing/data': 0.0002135700051439926, 'timing/model': 0.36932631200761534, 'learning_rate/qwen_vl_interface': 1.3026355415627565e-06, 'learning_rate/action_model': 1.3026355415627564e-05, 'epoch': 0.3}) +[2026-10-03 15:24:30,148][starVLA.training.train_starvla][INFO] - Step 4081, Loss: {'train/loss': 3.062181349378079e-06, 'train/grad_norm_pre_clip': 0.000540041015483439, 'timing/data': 0.0003555799921741709, 'timing/model': 0.3691917609976372, 'learning_rate/qwen_vl_interface': 1.3009423786710653e-06, 'learning_rate/action_model': 1.3009423786710653e-05, 'epoch': 0.3}) +[2026-10-03 15:24:30,804][starVLA.training.train_starvla][INFO] - Step 4082, Loss: {'train/loss': 0.0004009562253486365, 'train/grad_norm_pre_clip': 1.1675792932510376, 'timing/data': 0.0002591199881862849, 'timing/model': 0.36888371100940276, 'learning_rate/qwen_vl_interface': 1.2992508390877933e-06, 'learning_rate/action_model': 1.2992508390877933e-05, 'epoch': 0.3}) +[2026-10-03 15:24:31,462][starVLA.training.train_starvla][INFO] - Step 4083, Loss: {'train/loss': 0.00011474356870166957, 'train/grad_norm_pre_clip': 0.05489364638924599, 'timing/data': 0.00025877000007312745, 'timing/model': 0.3695796510000946, 'learning_rate/qwen_vl_interface': 1.297560923508269e-06, 'learning_rate/action_model': 1.297560923508269e-05, 'epoch': 0.3}) +[2026-10-03 15:24:32,120][starVLA.training.train_starvla][INFO] - Step 4084, Loss: {'train/loss': 8.12106009107083e-06, 'train/grad_norm_pre_clip': 0.002052947413176298, 'timing/data': 0.00021405999723356217, 'timing/model': 0.3690137509984197, 'learning_rate/qwen_vl_interface': 1.295872632627152e-06, 'learning_rate/action_model': 1.2958726326271519e-05, 'epoch': 0.3}) +[2026-10-03 15:24:32,778][starVLA.training.train_starvla][INFO] - Step 4085, Loss: {'train/loss': 3.9189881135826e-06, 'train/grad_norm_pre_clip': 0.0009960498427972198, 'timing/data': 0.0003480299928924069, 'timing/model': 0.36967697099316865, 'learning_rate/qwen_vl_interface': 1.2941859671384364e-06, 'learning_rate/action_model': 1.2941859671384362e-05, 'epoch': 0.3}) +[2026-10-03 15:24:33,625][starVLA.training.train_starvla][INFO] - Step 4086, Loss: {'train/loss': 7.076897600200027e-05, 'train/grad_norm_pre_clip': 0.14086920022964478, 'timing/data': 0.00024033000227063894, 'timing/model': 0.5595704369916348, 'learning_rate/qwen_vl_interface': 1.2925009277354444e-06, 'learning_rate/action_model': 1.2925009277354444e-05, 'epoch': 0.3}) +[2026-10-03 15:24:34,283][starVLA.training.train_starvla][INFO] - Step 4087, Loss: {'train/loss': 3.5986773582408205e-05, 'train/grad_norm_pre_clip': 0.024067232385277748, 'timing/data': 0.0002911300107371062, 'timing/model': 0.36957827099831775, 'learning_rate/qwen_vl_interface': 1.2908175151108335e-06, 'learning_rate/action_model': 1.2908175151108334e-05, 'epoch': 0.3}) +[2026-10-03 15:24:34,940][starVLA.training.train_starvla][INFO] - Step 4088, Loss: {'train/loss': 4.835223808186129e-05, 'train/grad_norm_pre_clip': 0.018840551376342773, 'timing/data': 0.0003478599974187091, 'timing/model': 0.36931594100315124, 'learning_rate/qwen_vl_interface': 1.2891357299565899e-06, 'learning_rate/action_model': 1.2891357299565898e-05, 'epoch': 0.3}) +[2026-10-03 15:24:35,598][starVLA.training.train_starvla][INFO] - Step 4089, Loss: {'train/loss': 0.15641844272613525, 'train/grad_norm_pre_clip': 119.50736236572266, 'timing/data': 0.0003476999991107732, 'timing/model': 0.3694568619976053, 'learning_rate/qwen_vl_interface': 1.2874555729640333e-06, 'learning_rate/action_model': 1.2874555729640334e-05, 'epoch': 0.3}) +[2026-10-03 15:24:36,254][starVLA.training.train_starvla][INFO] - Step 4090, Loss: {'train/loss': 0.1793004274368286, 'train/grad_norm_pre_clip': 89.04035186767578, 'timing/data': 0.0003060000017285347, 'timing/model': 0.3690507319988683, 'learning_rate/qwen_vl_interface': 1.2857770448238124e-06, 'learning_rate/action_model': 1.2857770448238123e-05, 'epoch': 0.3}) +[2026-10-03 15:24:37,113][starVLA.training.train_starvla][INFO] - Step 4091, Loss: {'train/loss': 0.38505175709724426, 'train/grad_norm_pre_clip': 44.91268539428711, 'timing/data': 0.00021148999803699553, 'timing/model': 0.571910504004336, 'learning_rate/qwen_vl_interface': 1.2841001462259067e-06, 'learning_rate/action_model': 1.2841001462259067e-05, 'epoch': 0.3}) +[2026-10-03 15:24:37,771][starVLA.training.train_starvla][INFO] - Step 4092, Loss: {'train/loss': 0.00011824579996755347, 'train/grad_norm_pre_clip': 0.09784381836652756, 'timing/data': 0.0003903399920091033, 'timing/model': 0.36974034100421704, 'learning_rate/qwen_vl_interface': 1.282424877859626e-06, 'learning_rate/action_model': 1.282424877859626e-05, 'epoch': 0.3}) +[2026-10-03 15:24:38,428][starVLA.training.train_starvla][INFO] - Step 4093, Loss: {'train/loss': 0.4107200801372528, 'train/grad_norm_pre_clip': 21.443214416503906, 'timing/data': 0.0002605500048957765, 'timing/model': 0.3691748810088029, 'learning_rate/qwen_vl_interface': 1.2807512404136088e-06, 'learning_rate/action_model': 1.2807512404136087e-05, 'epoch': 0.3}) +[2026-10-03 15:24:39,085][starVLA.training.train_starvla][INFO] - Step 4094, Loss: {'train/loss': 0.00018114859994966537, 'train/grad_norm_pre_clip': 0.159342959523201, 'timing/data': 0.00021149001258891076, 'timing/model': 0.3695027819921961, 'learning_rate/qwen_vl_interface': 1.2790792345758255e-06, 'learning_rate/action_model': 1.2790792345758254e-05, 'epoch': 0.3}) +[2026-10-03 15:24:39,742][starVLA.training.train_starvla][INFO] - Step 4095, Loss: {'train/loss': 0.0006147095700725913, 'train/grad_norm_pre_clip': 2.764566659927368, 'timing/data': 0.00032079999800771475, 'timing/model': 0.36932337199687026, 'learning_rate/qwen_vl_interface': 1.2774088610335752e-06, 'learning_rate/action_model': 1.2774088610335752e-05, 'epoch': 0.3}) +[2026-10-03 15:24:40,401][starVLA.training.train_starvla][INFO] - Step 4096, Loss: {'train/loss': 2.4279435820062645e-05, 'train/grad_norm_pre_clip': 0.012051604688167572, 'timing/data': 0.00028603999817278236, 'timing/model': 0.36917572100355756, 'learning_rate/qwen_vl_interface': 1.2757401204734827e-06, 'learning_rate/action_model': 1.2757401204734826e-05, 'epoch': 0.3}) +[2026-10-03 15:24:41,251][starVLA.training.train_starvla][INFO] - Step 4097, Loss: {'train/loss': 0.3569745421409607, 'train/grad_norm_pre_clip': 13.346431732177734, 'timing/data': 0.00027226000383961946, 'timing/model': 0.3707752409973182, 'learning_rate/qwen_vl_interface': 1.2740730135815064e-06, 'learning_rate/action_model': 1.2740730135815062e-05, 'epoch': 0.3}) +[2026-10-03 15:24:41,911][starVLA.training.train_starvla][INFO] - Step 4098, Loss: {'train/loss': 1.0594279046927113e-05, 'train/grad_norm_pre_clip': 0.004516596905887127, 'timing/data': 0.0003492800024105236, 'timing/model': 0.37043478099803906, 'learning_rate/qwen_vl_interface': 1.2724075410429298e-06, 'learning_rate/action_model': 1.2724075410429299e-05, 'epoch': 0.3}) +[2026-10-03 15:24:42,570][starVLA.training.train_starvla][INFO] - Step 4099, Loss: {'train/loss': 0.0012127798981964588, 'train/grad_norm_pre_clip': 4.769604206085205, 'timing/data': 0.00028994999593123794, 'timing/model': 0.37086179200559855, 'learning_rate/qwen_vl_interface': 1.2707437035423664e-06, 'learning_rate/action_model': 1.2707437035423664e-05, 'epoch': 0.3}) +[2026-10-03 15:24:43,228][starVLA.training.train_starvla][INFO] - Step 4100, Loss: {'train/loss': 9.807480091694742e-05, 'train/grad_norm_pre_clip': 0.07147390395402908, 'timing/data': 0.00035123000270687044, 'timing/model': 0.3685154520062497, 'learning_rate/qwen_vl_interface': 1.2690815017637573e-06, 'learning_rate/action_model': 1.2690815017637572e-05, 'epoch': 0.3}) +[2026-10-03 15:24:43,886][starVLA.training.train_starvla][INFO] - Step 4101, Loss: {'train/loss': 0.00020288271480239928, 'train/grad_norm_pre_clip': 0.04620501399040222, 'timing/data': 0.00035327000659890473, 'timing/model': 0.36905329099681694, 'learning_rate/qwen_vl_interface': 1.2674209363903694e-06, 'learning_rate/action_model': 1.2674209363903693e-05, 'epoch': 0.3}) +[2026-10-03 15:24:44,735][starVLA.training.train_starvla][INFO] - Step 4102, Loss: {'train/loss': 9.886772204481531e-06, 'train/grad_norm_pre_clip': 0.0026461943052709103, 'timing/data': 0.0003788899921346456, 'timing/model': 0.3682577909930842, 'learning_rate/qwen_vl_interface': 1.2657620081048005e-06, 'learning_rate/action_model': 1.2657620081048006e-05, 'epoch': 0.3}) +[2026-10-03 15:24:45,392][starVLA.training.train_starvla][INFO] - Step 4103, Loss: {'train/loss': 0.0002937513345386833, 'train/grad_norm_pre_clip': 0.12205320596694946, 'timing/data': 0.00028295999800320715, 'timing/model': 0.36899065200123005, 'learning_rate/qwen_vl_interface': 1.2641047175889714e-06, 'learning_rate/action_model': 1.2641047175889714e-05, 'epoch': 0.3}) +[2026-10-03 15:24:46,049][starVLA.training.train_starvla][INFO] - Step 4104, Loss: {'train/loss': 3.864927566610277e-05, 'train/grad_norm_pre_clip': 0.020370762795209885, 'timing/data': 0.00024850999761838466, 'timing/model': 0.3688335309998365, 'learning_rate/qwen_vl_interface': 1.2624490655241332e-06, 'learning_rate/action_model': 1.262449065524133e-05, 'epoch': 0.3}) +[2026-10-03 15:24:46,706][starVLA.training.train_starvla][INFO] - Step 4105, Loss: {'train/loss': 3.9515925891464576e-05, 'train/grad_norm_pre_clip': 0.016247762367129326, 'timing/data': 0.00034666999999899417, 'timing/model': 0.36890086099447217, 'learning_rate/qwen_vl_interface': 1.2607950525908602e-06, 'learning_rate/action_model': 1.2607950525908601e-05, 'epoch': 0.3}) +[2026-10-03 15:24:47,362][starVLA.training.train_starvla][INFO] - Step 4106, Loss: {'train/loss': 0.48176249861717224, 'train/grad_norm_pre_clip': 134.7560577392578, 'timing/data': 0.0003876900009345263, 'timing/model': 0.36912467099318746, 'learning_rate/qwen_vl_interface': 1.2591426794690575e-06, 'learning_rate/action_model': 1.2591426794690576e-05, 'epoch': 0.3}) +[2026-10-03 15:24:48,184][starVLA.training.train_starvla][INFO] - Step 4107, Loss: {'train/loss': 0.5368845462799072, 'train/grad_norm_pre_clip': 60.063289642333984, 'timing/data': 0.0003457200073171407, 'timing/model': 0.5353093090088805, 'learning_rate/qwen_vl_interface': 1.25749194683795e-06, 'learning_rate/action_model': 1.25749194683795e-05, 'epoch': 0.3}) +[2026-10-03 15:24:48,845][starVLA.training.train_starvla][INFO] - Step 4108, Loss: {'train/loss': 1.3090399079374038e-05, 'train/grad_norm_pre_clip': 0.004351080860942602, 'timing/data': 0.00035064000985585153, 'timing/model': 0.3703024610003922, 'learning_rate/qwen_vl_interface': 1.2558428553760933e-06, 'learning_rate/action_model': 1.2558428553760932e-05, 'epoch': 0.3}) +[2026-10-03 15:24:49,506][starVLA.training.train_starvla][INFO] - Step 4109, Loss: {'train/loss': 1.1391493899282068e-05, 'train/grad_norm_pre_clip': 0.004443584941327572, 'timing/data': 0.0002307700051460415, 'timing/model': 0.3705463710066397, 'learning_rate/qwen_vl_interface': 1.2541954057613659e-06, 'learning_rate/action_model': 1.2541954057613658e-05, 'epoch': 0.3}) +[2026-10-03 15:24:50,164][starVLA.training.train_starvla][INFO] - Step 4110, Loss: {'train/loss': 0.00027847697492688894, 'train/grad_norm_pre_clip': 0.36370226740837097, 'timing/data': 0.00030762000824324787, 'timing/model': 0.37034658099582884, 'learning_rate/qwen_vl_interface': 1.2525495986709727e-06, 'learning_rate/action_model': 1.2525495986709726e-05, 'epoch': 0.3}) +[2026-10-03 15:24:50,823][starVLA.training.train_starvla][INFO] - Step 4111, Loss: {'train/loss': 0.006304721813648939, 'train/grad_norm_pre_clip': 36.478763580322266, 'timing/data': 0.00033860999974422157, 'timing/model': 0.3703281100024469, 'learning_rate/qwen_vl_interface': 1.2509054347814423e-06, 'learning_rate/action_model': 1.2509054347814423e-05, 'epoch': 0.3}) +[2026-10-03 15:24:51,682][starVLA.training.train_starvla][INFO] - Step 4112, Loss: {'train/loss': 1.6315929315169342e-05, 'train/grad_norm_pre_clip': 0.00974914152175188, 'timing/data': 0.00021485000615939498, 'timing/model': 0.5694161249994067, 'learning_rate/qwen_vl_interface': 1.2492629147686288e-06, 'learning_rate/action_model': 1.2492629147686289e-05, 'epoch': 0.3}) +[2026-10-03 15:24:52,343][starVLA.training.train_starvla][INFO] - Step 4113, Loss: {'train/loss': 0.00012145027721999213, 'train/grad_norm_pre_clip': 0.09565451741218567, 'timing/data': 0.00034100000630132854, 'timing/model': 0.3714425020007184, 'learning_rate/qwen_vl_interface': 1.2476220393077098e-06, 'learning_rate/action_model': 1.2476220393077098e-05, 'epoch': 0.3}) +[2026-10-03 15:24:53,002][starVLA.training.train_starvla][INFO] - Step 4114, Loss: {'train/loss': 9.320443496108055e-06, 'train/grad_norm_pre_clip': 0.003092141356319189, 'timing/data': 0.0003338100068503991, 'timing/model': 0.3710599710029783, 'learning_rate/qwen_vl_interface': 1.2459828090731877e-06, 'learning_rate/action_model': 1.2459828090731876e-05, 'epoch': 0.3}) +[2026-10-03 15:24:53,660][starVLA.training.train_starvla][INFO] - Step 4115, Loss: {'train/loss': 0.20439749956130981, 'train/grad_norm_pre_clip': 182.34942626953125, 'timing/data': 0.000245939998421818, 'timing/model': 0.37027505099831615, 'learning_rate/qwen_vl_interface': 1.2443452247388876e-06, 'learning_rate/action_model': 1.2443452247388877e-05, 'epoch': 0.3}) +[2026-10-03 15:24:54,320][starVLA.training.train_starvla][INFO] - Step 4116, Loss: {'train/loss': 6.21367507847026e-06, 'train/grad_norm_pre_clip': 0.002665191888809204, 'timing/data': 0.00031226999999489635, 'timing/model': 0.3700862010009587, 'learning_rate/qwen_vl_interface': 1.2427092869779603e-06, 'learning_rate/action_model': 1.2427092869779602e-05, 'epoch': 0.3}) +[2026-10-03 15:24:54,980][starVLA.training.train_starvla][INFO] - Step 4117, Loss: {'train/loss': 5.019035597797483e-05, 'train/grad_norm_pre_clip': 0.04754599183797836, 'timing/data': 0.00021688999549951404, 'timing/model': 0.37085788100375794, 'learning_rate/qwen_vl_interface': 1.2410749964628751e-06, 'learning_rate/action_model': 1.2410749964628751e-05, 'epoch': 0.3}) +[2026-10-03 15:24:55,831][starVLA.training.train_starvla][INFO] - Step 4118, Loss: {'train/loss': 5.112284634378739e-05, 'train/grad_norm_pre_clip': 0.05059845745563507, 'timing/data': 0.00021328000002540648, 'timing/model': 0.37051612099458, 'learning_rate/qwen_vl_interface': 1.2394423538654296e-06, 'learning_rate/action_model': 1.2394423538654296e-05, 'epoch': 0.3}) +[2026-10-03 15:24:56,490][starVLA.training.train_starvla][INFO] - Step 4119, Loss: {'train/loss': 0.000851157121360302, 'train/grad_norm_pre_clip': 1.4114338159561157, 'timing/data': 0.000268509000306949, 'timing/model': 0.37056629099242855, 'learning_rate/qwen_vl_interface': 1.2378113598567394e-06, 'learning_rate/action_model': 1.2378113598567392e-05, 'epoch': 0.3}) +[2026-10-03 15:24:57,151][starVLA.training.train_starvla][INFO] - Step 4120, Loss: {'train/loss': 7.703835763095412e-06, 'train/grad_norm_pre_clip': 0.001880113733932376, 'timing/data': 0.00027021999994758517, 'timing/model': 0.3702821509941714, 'learning_rate/qwen_vl_interface': 1.2361820151072454e-06, 'learning_rate/action_model': 1.2361820151072453e-05, 'epoch': 0.3}) +[2026-10-03 15:24:57,812][starVLA.training.train_starvla][INFO] - Step 4121, Loss: {'train/loss': 0.46395179629325867, 'train/grad_norm_pre_clip': 39.48175811767578, 'timing/data': 0.0003399800043553114, 'timing/model': 0.37069937100750394, 'learning_rate/qwen_vl_interface': 1.23455432028671e-06, 'learning_rate/action_model': 1.23455432028671e-05, 'epoch': 0.3}) +[2026-10-03 15:24:58,471][starVLA.training.train_starvla][INFO] - Step 4122, Loss: {'train/loss': 4.4350232201395556e-05, 'train/grad_norm_pre_clip': 0.022234655916690826, 'timing/data': 0.0003758990060305223, 'timing/model': 0.37066387198865414, 'learning_rate/qwen_vl_interface': 1.2329282760642167e-06, 'learning_rate/action_model': 1.2329282760642166e-05, 'epoch': 0.3}) +[2026-10-03 15:24:59,335][starVLA.training.train_starvla][INFO] - Step 4123, Loss: {'train/loss': 0.0016074574086815119, 'train/grad_norm_pre_clip': 3.3002469539642334, 'timing/data': 0.0003148800024064258, 'timing/model': 0.3690487720014062, 'learning_rate/qwen_vl_interface': 1.2313038831081716e-06, 'learning_rate/action_model': 1.2313038831081717e-05, 'epoch': 0.3}) +[2026-10-03 15:24:59,992][starVLA.training.train_starvla][INFO] - Step 4124, Loss: {'train/loss': 8.612785677541979e-06, 'train/grad_norm_pre_clip': 0.0026727821677923203, 'timing/data': 0.00021348000154830515, 'timing/model': 0.3691481410060078, 'learning_rate/qwen_vl_interface': 1.2296811420863002e-06, 'learning_rate/action_model': 1.2296811420863003e-05, 'epoch': 0.3}) +[2026-10-03 15:25:00,649][starVLA.training.train_starvla][INFO] - Step 4125, Loss: {'train/loss': 0.0003096162108704448, 'train/grad_norm_pre_clip': 0.44177019596099854, 'timing/data': 0.00031227999716065824, 'timing/model': 0.36871291100396775, 'learning_rate/qwen_vl_interface': 1.2280600536656504e-06, 'learning_rate/action_model': 1.2280600536656503e-05, 'epoch': 0.3}) +[2026-10-03 15:25:01,304][starVLA.training.train_starvla][INFO] - Step 4126, Loss: {'train/loss': 0.0001144041380030103, 'train/grad_norm_pre_clip': 0.05103970319032669, 'timing/data': 0.0003711000026669353, 'timing/model': 0.3683421319874469, 'learning_rate/qwen_vl_interface': 1.2264406185125907e-06, 'learning_rate/action_model': 1.2264406185125907e-05, 'epoch': 0.3}) +[2026-10-03 15:25:01,961][starVLA.training.train_starvla][INFO] - Step 4127, Loss: {'train/loss': 0.4978999197483063, 'train/grad_norm_pre_clip': 33.23383712768555, 'timing/data': 0.00027100999432150275, 'timing/model': 0.3699238710105419, 'learning_rate/qwen_vl_interface': 1.2248228372928098e-06, 'learning_rate/action_model': 1.2248228372928097e-05, 'epoch': 0.3}) +[2026-10-03 15:25:02,792][starVLA.training.train_starvla][INFO] - Step 4128, Loss: {'train/loss': 0.3040061593055725, 'train/grad_norm_pre_clip': 23.101797103881836, 'timing/data': 0.0003601500065997243, 'timing/model': 0.5652035859966418, 'learning_rate/qwen_vl_interface': 1.2232067106713174e-06, 'learning_rate/action_model': 1.2232067106713174e-05, 'epoch': 0.3}) +[2026-10-03 15:25:03,450][starVLA.training.train_starvla][INFO] - Step 4129, Loss: {'train/loss': 4.449556581676006e-05, 'train/grad_norm_pre_clip': 0.022724222391843796, 'timing/data': 0.00036580000596586615, 'timing/model': 0.36876962101086974, 'learning_rate/qwen_vl_interface': 1.22159223931244e-06, 'learning_rate/action_model': 1.2215922393124399e-05, 'epoch': 0.3}) +[2026-10-03 15:25:04,106][starVLA.training.train_starvla][INFO] - Step 4130, Loss: {'train/loss': 0.0003665073309093714, 'train/grad_norm_pre_clip': 0.08205150812864304, 'timing/data': 0.0002764390083029866, 'timing/model': 0.36873420199844986, 'learning_rate/qwen_vl_interface': 1.2199794238798266e-06, 'learning_rate/action_model': 1.2199794238798264e-05, 'epoch': 0.3}) +[2026-10-03 15:25:04,762][starVLA.training.train_starvla][INFO] - Step 4131, Loss: {'train/loss': 9.13055264391005e-05, 'train/grad_norm_pre_clip': 0.11581224948167801, 'timing/data': 0.00024316000053659081, 'timing/model': 0.36934446200029925, 'learning_rate/qwen_vl_interface': 1.2183682650364447e-06, 'learning_rate/action_model': 1.2183682650364447e-05, 'epoch': 0.3}) +[2026-10-03 15:25:05,396][starVLA.training.train_starvla][INFO] - Step 4132, Loss: {'train/loss': 0.00040190338040702045, 'train/grad_norm_pre_clip': 0.09810135513544083, 'timing/data': 0.00032267000642605126, 'timing/model': 0.36863205100235064, 'learning_rate/qwen_vl_interface': 1.216758763444581e-06, 'learning_rate/action_model': 1.2167587634445809e-05, 'epoch': 0.3}) +[2026-10-03 15:25:06,054][starVLA.training.train_starvla][INFO] - Step 4133, Loss: {'train/loss': 0.0005415435298345983, 'train/grad_norm_pre_clip': 0.11733084917068481, 'timing/data': 0.0003435200051171705, 'timing/model': 0.36894120200304314, 'learning_rate/qwen_vl_interface': 1.215150919765841e-06, 'learning_rate/action_model': 1.215150919765841e-05, 'epoch': 0.3}) +[2026-10-03 15:25:06,903][starVLA.training.train_starvla][INFO] - Step 4134, Loss: {'train/loss': 1.6062904251157306e-05, 'train/grad_norm_pre_clip': 0.004844327457249165, 'timing/data': 0.00021826999727636576, 'timing/model': 0.3692101420019753, 'learning_rate/qwen_vl_interface': 1.213544734661148e-06, 'learning_rate/action_model': 1.2135447346611478e-05, 'epoch': 0.3}) +[2026-10-03 15:25:07,559][starVLA.training.train_starvla][INFO] - Step 4135, Loss: {'train/loss': 0.0006845347816124558, 'train/grad_norm_pre_clip': 0.28562623262405396, 'timing/data': 0.0002811700105667114, 'timing/model': 0.36899225099477917, 'learning_rate/qwen_vl_interface': 1.2119402087907448e-06, 'learning_rate/action_model': 1.2119402087907448e-05, 'epoch': 0.3}) +[2026-10-03 15:25:08,217][starVLA.training.train_starvla][INFO] - Step 4136, Loss: {'train/loss': 2.4115890482789837e-05, 'train/grad_norm_pre_clip': 0.008764862082898617, 'timing/data': 0.00030443001014646143, 'timing/model': 0.3688273610023316, 'learning_rate/qwen_vl_interface': 1.2103373428141903e-06, 'learning_rate/action_model': 1.2103373428141903e-05, 'epoch': 0.3}) +[2026-10-03 15:25:08,874][starVLA.training.train_starvla][INFO] - Step 4137, Loss: {'train/loss': 9.924018741003238e-06, 'train/grad_norm_pre_clip': 0.002431252971291542, 'timing/data': 0.0003134099970338866, 'timing/model': 0.3689745409938041, 'learning_rate/qwen_vl_interface': 1.2087361373903633e-06, 'learning_rate/action_model': 1.2087361373903633e-05, 'epoch': 0.3}) +[2026-10-03 15:25:09,529][starVLA.training.train_starvla][INFO] - Step 4138, Loss: {'train/loss': 0.00011830026051029563, 'train/grad_norm_pre_clip': 0.06890422105789185, 'timing/data': 0.00024833901261445135, 'timing/model': 0.36856155199347995, 'learning_rate/qwen_vl_interface': 1.2071365931774584e-06, 'learning_rate/action_model': 1.2071365931774584e-05, 'epoch': 0.3}) +[2026-10-03 15:25:10,376][starVLA.training.train_starvla][INFO] - Step 4139, Loss: {'train/loss': 2.524647607060615e-05, 'train/grad_norm_pre_clip': 0.011675672605633736, 'timing/data': 0.0002508000034140423, 'timing/model': 0.3686207819991978, 'learning_rate/qwen_vl_interface': 1.2055387108329885e-06, 'learning_rate/action_model': 1.2055387108329886e-05, 'epoch': 0.3}) +[2026-10-03 15:25:11,034][starVLA.training.train_starvla][INFO] - Step 4140, Loss: {'train/loss': 0.032046087086200714, 'train/grad_norm_pre_clip': 43.611656188964844, 'timing/data': 0.0002678190066944808, 'timing/model': 0.3690755920106312, 'learning_rate/qwen_vl_interface': 1.2039424910137807e-06, 'learning_rate/action_model': 1.2039424910137806e-05, 'epoch': 0.3}) +[2026-10-03 15:25:11,669][starVLA.training.train_starvla][INFO] - Step 4141, Loss: {'train/loss': 0.5510807037353516, 'train/grad_norm_pre_clip': 59.03992462158203, 'timing/data': 0.0002522799914004281, 'timing/model': 0.3686540119961137, 'learning_rate/qwen_vl_interface': 1.2023479343759822e-06, 'learning_rate/action_model': 1.2023479343759823e-05, 'epoch': 0.3}) +[2026-10-03 15:25:12,325][starVLA.training.train_starvla][INFO] - Step 4142, Loss: {'train/loss': 1.778345540515147e-05, 'train/grad_norm_pre_clip': 0.008782429620623589, 'timing/data': 0.00030886000604368746, 'timing/model': 0.36929849100124557, 'learning_rate/qwen_vl_interface': 1.2007550415750546e-06, 'learning_rate/action_model': 1.2007550415750545e-05, 'epoch': 0.3}) +[2026-10-03 15:25:12,982][starVLA.training.train_starvla][INFO] - Step 4143, Loss: {'train/loss': 3.8146815768413944e-06, 'train/grad_norm_pre_clip': 0.0008270554826594889, 'timing/data': 0.0002623899927129969, 'timing/model': 0.36966475201188587, 'learning_rate/qwen_vl_interface': 1.1991638132657759e-06, 'learning_rate/action_model': 1.199163813265776e-05, 'epoch': 0.3}) +[2026-10-03 15:25:13,843][starVLA.training.train_starvla][INFO] - Step 4144, Loss: {'train/loss': 0.00025396901764906943, 'train/grad_norm_pre_clip': 0.11568474024534225, 'timing/data': 0.00031149999995250255, 'timing/model': 0.5725339349883143, 'learning_rate/qwen_vl_interface': 1.1975742501022406e-06, 'learning_rate/action_model': 1.1975742501022405e-05, 'epoch': 0.3}) +[2026-10-03 15:25:14,500][starVLA.training.train_starvla][INFO] - Step 4145, Loss: {'train/loss': 0.6554788947105408, 'train/grad_norm_pre_clip': 56.38129806518555, 'timing/data': 0.00038374999712686986, 'timing/model': 0.36896274199534673, 'learning_rate/qwen_vl_interface': 1.1959863527378569e-06, 'learning_rate/action_model': 1.1959863527378567e-05, 'epoch': 0.3}) +[2026-10-03 15:25:15,134][starVLA.training.train_starvla][INFO] - Step 4146, Loss: {'train/loss': 6.847019903943874e-06, 'train/grad_norm_pre_clip': 0.0015081767924129963, 'timing/data': 0.0003012699889950454, 'timing/model': 0.3685326120030368, 'learning_rate/qwen_vl_interface': 1.1944001218253505e-06, 'learning_rate/action_model': 1.1944001218253503e-05, 'epoch': 0.3}) +[2026-10-03 15:25:15,790][starVLA.training.train_starvla][INFO] - Step 4147, Loss: {'train/loss': 0.0020162267610430717, 'train/grad_norm_pre_clip': 4.940938949584961, 'timing/data': 0.00020603000302799046, 'timing/model': 0.36925431099371053, 'learning_rate/qwen_vl_interface': 1.1928155580167608e-06, 'learning_rate/action_model': 1.1928155580167608e-05, 'epoch': 0.3}) +[2026-10-03 15:25:16,447][starVLA.training.train_starvla][INFO] - Step 4148, Loss: {'train/loss': 0.3644983470439911, 'train/grad_norm_pre_clip': 39.86479568481445, 'timing/data': 0.00038717000279575586, 'timing/model': 0.3692163410014473, 'learning_rate/qwen_vl_interface': 1.1912326619634412e-06, 'learning_rate/action_model': 1.1912326619634412e-05, 'epoch': 0.3}) +[2026-10-03 15:25:17,105][starVLA.training.train_starvla][INFO] - Step 4149, Loss: {'train/loss': 0.0008227885700762272, 'train/grad_norm_pre_clip': 0.4584779739379883, 'timing/data': 0.0002105799940181896, 'timing/model': 0.3690773709968198, 'learning_rate/qwen_vl_interface': 1.1896514343160632e-06, 'learning_rate/action_model': 1.189651434316063e-05, 'epoch': 0.3}) +[2026-10-03 15:25:17,958][starVLA.training.train_starvla][INFO] - Step 4150, Loss: {'train/loss': 8.850984158925712e-06, 'train/grad_norm_pre_clip': 0.003749089315533638, 'timing/data': 0.00037352999788708985, 'timing/model': 0.36870157199155074, 'learning_rate/qwen_vl_interface': 1.1880718757246062e-06, 'learning_rate/action_model': 1.1880718757246063e-05, 'epoch': 0.3}) +[2026-10-03 15:25:18,615][starVLA.training.train_starvla][INFO] - Step 4151, Loss: {'train/loss': 6.854463208583184e-06, 'train/grad_norm_pre_clip': 0.002386941574513912, 'timing/data': 0.00030035000236239284, 'timing/model': 0.3700466619920917, 'learning_rate/qwen_vl_interface': 1.186493986838369e-06, 'learning_rate/action_model': 1.186493986838369e-05, 'epoch': 0.3}) +[2026-10-03 15:25:19,274][starVLA.training.train_starvla][INFO] - Step 4152, Loss: {'train/loss': 0.00036873595672659576, 'train/grad_norm_pre_clip': 0.08648495376110077, 'timing/data': 0.00037167000118643045, 'timing/model': 0.36935071100015193, 'learning_rate/qwen_vl_interface': 1.1849177683059612e-06, 'learning_rate/action_model': 1.1849177683059612e-05, 'epoch': 0.3}) +[2026-10-03 15:25:19,908][starVLA.training.train_starvla][INFO] - Step 4153, Loss: {'train/loss': 2.4718292479519732e-05, 'train/grad_norm_pre_clip': 0.009917082265019417, 'timing/data': 0.00036700999771710485, 'timing/model': 0.345867714000633, 'learning_rate/qwen_vl_interface': 1.1833432207753075e-06, 'learning_rate/action_model': 1.1833432207753076e-05, 'epoch': 0.3}) +[2026-10-03 15:25:20,565][starVLA.training.train_starvla][INFO] - Step 4154, Loss: {'train/loss': 5.464235437102616e-05, 'train/grad_norm_pre_clip': 0.020509567111730576, 'timing/data': 0.0003646600089268759, 'timing/model': 0.36953919001098257, 'learning_rate/qwen_vl_interface': 1.1817703448936448e-06, 'learning_rate/action_model': 1.181770344893645e-05, 'epoch': 0.3}) +[2026-10-03 15:25:21,412][starVLA.training.train_starvla][INFO] - Step 4155, Loss: {'train/loss': 2.263336682517547e-05, 'train/grad_norm_pre_clip': 0.013076397590339184, 'timing/data': 0.0002893799974117428, 'timing/model': 0.5592691750061931, 'learning_rate/qwen_vl_interface': 1.1801991413075228e-06, 'learning_rate/action_model': 1.1801991413075227e-05, 'epoch': 0.3}) +[2026-10-03 15:25:22,071][starVLA.training.train_starvla][INFO] - Step 4156, Loss: {'train/loss': 0.08408138155937195, 'train/grad_norm_pre_clip': 123.23319244384766, 'timing/data': 0.0004639799881260842, 'timing/model': 0.36900626099668443, 'learning_rate/qwen_vl_interface': 1.1786296106628038e-06, 'learning_rate/action_model': 1.1786296106628036e-05, 'epoch': 0.3}) +[2026-10-03 15:25:22,727][starVLA.training.train_starvla][INFO] - Step 4157, Loss: {'train/loss': 0.017346123233437538, 'train/grad_norm_pre_clip': 20.461885452270508, 'timing/data': 0.0003529399982653558, 'timing/model': 0.36866716999793425, 'learning_rate/qwen_vl_interface': 1.1770617536046633e-06, 'learning_rate/action_model': 1.1770617536046632e-05, 'epoch': 0.3}) +[2026-10-03 15:25:23,384][starVLA.training.train_starvla][INFO] - Step 4158, Loss: {'train/loss': 0.0003747891169041395, 'train/grad_norm_pre_clip': 0.29390519857406616, 'timing/data': 0.00020574001246131957, 'timing/model': 0.36930721999669913, 'learning_rate/qwen_vl_interface': 1.1754955707775873e-06, 'learning_rate/action_model': 1.1754955707775872e-05, 'epoch': 0.3}) +[2026-10-03 15:25:24,041][starVLA.training.train_starvla][INFO] - Step 4159, Loss: {'train/loss': 0.0002671901893336326, 'train/grad_norm_pre_clip': 0.27968984842300415, 'timing/data': 0.00037553999572992325, 'timing/model': 0.3692878500005463, 'learning_rate/qwen_vl_interface': 1.1739310628253753e-06, 'learning_rate/action_model': 1.1739310628253753e-05, 'epoch': 0.3}) +[2026-10-03 15:25:24,893][starVLA.training.train_starvla][INFO] - Step 4160, Loss: {'train/loss': 3.526663931552321e-05, 'train/grad_norm_pre_clip': 0.03287702053785324, 'timing/data': 0.00026470900047570467, 'timing/model': 0.5643898849957623, 'learning_rate/qwen_vl_interface': 1.172368230391136e-06, 'learning_rate/action_model': 1.1723682303911359e-05, 'epoch': 0.3}) +[2026-10-03 15:25:25,549][starVLA.training.train_starvla][INFO] - Step 4161, Loss: {'train/loss': 4.455347152543254e-05, 'train/grad_norm_pre_clip': 0.021436387673020363, 'timing/data': 0.00023689999943599105, 'timing/model': 0.3688757700001588, 'learning_rate/qwen_vl_interface': 1.1708070741172928e-06, 'learning_rate/action_model': 1.1708070741172928e-05, 'epoch': 0.3}) +[2026-10-03 15:25:26,159][starVLA.training.train_starvla][INFO] - Step 4162, Loss: {'train/loss': 1.2613490980584174e-05, 'train/grad_norm_pre_clip': 0.0038232996594160795, 'timing/data': 0.000352019997080788, 'timing/model': 0.3459949240059359, 'learning_rate/qwen_vl_interface': 1.1692475946455755e-06, 'learning_rate/action_model': 1.1692475946455755e-05, 'epoch': 0.3}) +[2026-10-03 15:25:26,816][starVLA.training.train_starvla][INFO] - Step 4163, Loss: {'train/loss': 3.9306876715272665e-05, 'train/grad_norm_pre_clip': 0.014720444567501545, 'timing/data': 0.00027312999009154737, 'timing/model': 0.36922511100419797, 'learning_rate/qwen_vl_interface': 1.1676897926170289e-06, 'learning_rate/action_model': 1.1676897926170289e-05, 'epoch': 0.3}) +[2026-10-03 15:25:27,474][starVLA.training.train_starvla][INFO] - Step 4164, Loss: {'train/loss': 1.0817882866831496e-05, 'train/grad_norm_pre_clip': 0.0032836757600307465, 'timing/data': 0.0003622499934863299, 'timing/model': 0.3691960209980607, 'learning_rate/qwen_vl_interface': 1.166133668672006e-06, 'learning_rate/action_model': 1.166133668672006e-05, 'epoch': 0.3}) +[2026-10-03 15:25:28,107][starVLA.training.train_starvla][INFO] - Step 4165, Loss: {'train/loss': 0.08150909096002579, 'train/grad_norm_pre_clip': 326.5951843261719, 'timing/data': 0.00020802000653930008, 'timing/model': 0.3688053310033865, 'learning_rate/qwen_vl_interface': 1.1645792234501706e-06, 'learning_rate/action_model': 1.1645792234501706e-05, 'epoch': 0.3}) +[2026-10-03 15:25:28,960][starVLA.training.train_starvla][INFO] - Step 4166, Loss: {'train/loss': 1.5884081221884117e-05, 'train/grad_norm_pre_clip': 0.004255932290107012, 'timing/data': 0.0004032099968753755, 'timing/model': 0.369275529999868, 'learning_rate/qwen_vl_interface': 1.1630264575904972e-06, 'learning_rate/action_model': 1.1630264575904973e-05, 'epoch': 0.3}) +[2026-10-03 15:25:29,616][starVLA.training.train_starvla][INFO] - Step 4167, Loss: {'train/loss': 2.010104071814567e-05, 'train/grad_norm_pre_clip': 0.006713423877954483, 'timing/data': 0.0002625600027386099, 'timing/model': 0.36878771099145524, 'learning_rate/qwen_vl_interface': 1.1614753717312683e-06, 'learning_rate/action_model': 1.1614753717312681e-05, 'epoch': 0.3}) +[2026-10-03 15:25:30,274][starVLA.training.train_starvla][INFO] - Step 4168, Loss: {'train/loss': 0.0012439304264262319, 'train/grad_norm_pre_clip': 1.3190817832946777, 'timing/data': 0.0002959400007966906, 'timing/model': 0.36972773999150377, 'learning_rate/qwen_vl_interface': 1.1599259665100767e-06, 'learning_rate/action_model': 1.1599259665100767e-05, 'epoch': 0.31}) +[2026-10-03 15:25:30,929][starVLA.training.train_starvla][INFO] - Step 4169, Loss: {'train/loss': 2.0897772628813982e-05, 'train/grad_norm_pre_clip': 0.007929305545985699, 'timing/data': 0.00021279999054968357, 'timing/model': 0.36911700999189634, 'learning_rate/qwen_vl_interface': 1.1583782425638252e-06, 'learning_rate/action_model': 1.158378242563825e-05, 'epoch': 0.31}) +[2026-10-03 15:25:31,584][starVLA.training.train_starvla][INFO] - Step 4170, Loss: {'train/loss': 2.8144939278718084e-05, 'train/grad_norm_pre_clip': 0.012844347395002842, 'timing/data': 0.0002076800010399893, 'timing/model': 0.3686070709954947, 'learning_rate/qwen_vl_interface': 1.1568322005287237e-06, 'learning_rate/action_model': 1.1568322005287235e-05, 'epoch': 0.31}) +[2026-10-03 15:25:32,431][starVLA.training.train_starvla][INFO] - Step 4171, Loss: {'train/loss': 3.879165160469711e-05, 'train/grad_norm_pre_clip': 0.02120346948504448, 'timing/data': 0.00023941999825183302, 'timing/model': 0.5596899249940179, 'learning_rate/qwen_vl_interface': 1.1552878410402923e-06, 'learning_rate/action_model': 1.1552878410402924e-05, 'epoch': 0.31}) +[2026-10-03 15:25:33,088][starVLA.training.train_starvla][INFO] - Step 4172, Loss: {'train/loss': 0.3451746106147766, 'train/grad_norm_pre_clip': 73.06990814208984, 'timing/data': 0.0002822599926730618, 'timing/model': 0.36840374099847395, 'learning_rate/qwen_vl_interface': 1.1537451647333594e-06, 'learning_rate/action_model': 1.1537451647333593e-05, 'epoch': 0.31}) +[2026-10-03 15:25:33,743][starVLA.training.train_starvla][INFO] - Step 4173, Loss: {'train/loss': 3.8471625884994864e-05, 'train/grad_norm_pre_clip': 0.05600159615278244, 'timing/data': 0.0003649199934443459, 'timing/model': 0.3687328210071428, 'learning_rate/qwen_vl_interface': 1.1522041722420594e-06, 'learning_rate/action_model': 1.1522041722420594e-05, 'epoch': 0.31}) +[2026-10-03 15:25:34,399][starVLA.training.train_starvla][INFO] - Step 4174, Loss: {'train/loss': 0.00012453633826225996, 'train/grad_norm_pre_clip': 0.11769431084394455, 'timing/data': 0.0003391799982637167, 'timing/model': 0.3688574310071999, 'learning_rate/qwen_vl_interface': 1.1506648641998365e-06, 'learning_rate/action_model': 1.1506648641998365e-05, 'epoch': 0.31}) +[2026-10-03 15:25:35,033][starVLA.training.train_starvla][INFO] - Step 4175, Loss: {'train/loss': 4.415025614434853e-05, 'train/grad_norm_pre_clip': 0.01828361116349697, 'timing/data': 0.00025910000840667635, 'timing/model': 0.3458340739889536, 'learning_rate/qwen_vl_interface': 1.1491272412394433e-06, 'learning_rate/action_model': 1.1491272412394433e-05, 'epoch': 0.31}) +[2026-10-03 15:25:35,886][starVLA.training.train_starvla][INFO] - Step 4176, Loss: {'train/loss': 0.0001637394307181239, 'train/grad_norm_pre_clip': 0.12352051585912704, 'timing/data': 0.00021302999812178314, 'timing/model': 0.560716774998582, 'learning_rate/qwen_vl_interface': 1.1475913039929373e-06, 'learning_rate/action_model': 1.1475913039929372e-05, 'epoch': 0.31}) +[2026-10-03 15:25:36,540][starVLA.training.train_starvla][INFO] - Step 4177, Loss: {'train/loss': 0.00022470849216915667, 'train/grad_norm_pre_clip': 0.07043613493442535, 'timing/data': 0.0002534399973228574, 'timing/model': 0.36830140100209974, 'learning_rate/qwen_vl_interface': 1.1460570530916844e-06, 'learning_rate/action_model': 1.1460570530916844e-05, 'epoch': 0.31}) +[2026-10-03 15:25:37,196][starVLA.training.train_starvla][INFO] - Step 4178, Loss: {'train/loss': 0.00013518176274374127, 'train/grad_norm_pre_clip': 0.08183691650629044, 'timing/data': 0.0002121900033671409, 'timing/model': 0.3687158009997802, 'learning_rate/qwen_vl_interface': 1.1445244891663581e-06, 'learning_rate/action_model': 1.1445244891663581e-05, 'epoch': 0.31}) +[2026-10-03 15:25:37,852][starVLA.training.train_starvla][INFO] - Step 4179, Loss: {'train/loss': 7.733604434179142e-06, 'train/grad_norm_pre_clip': 0.0018839772092178464, 'timing/data': 0.0003465590125415474, 'timing/model': 0.368727041000966, 'learning_rate/qwen_vl_interface': 1.1429936128469366e-06, 'learning_rate/action_model': 1.1429936128469366e-05, 'epoch': 0.31}) +[2026-10-03 15:25:38,512][starVLA.training.train_starvla][INFO] - Step 4180, Loss: {'train/loss': 0.00013227533781901002, 'train/grad_norm_pre_clip': 0.08919006586074829, 'timing/data': 0.00033780999365262687, 'timing/model': 0.3708342600002652, 'learning_rate/qwen_vl_interface': 1.1414644247627056e-06, 'learning_rate/action_model': 1.1414644247627058e-05, 'epoch': 0.31}) +[2026-10-03 15:25:39,365][starVLA.training.train_starvla][INFO] - Step 4181, Loss: {'train/loss': 1.0326344636268914e-05, 'train/grad_norm_pre_clip': 0.003101898357272148, 'timing/data': 0.0002976500109070912, 'timing/model': 0.5652685639943229, 'learning_rate/qwen_vl_interface': 1.1399369255422568e-06, 'learning_rate/action_model': 1.1399369255422568e-05, 'epoch': 0.31}) +[2026-10-03 15:25:40,023][starVLA.training.train_starvla][INFO] - Step 4182, Loss: {'train/loss': 0.005319423042237759, 'train/grad_norm_pre_clip': 6.138298511505127, 'timing/data': 0.00036859999818261713, 'timing/model': 0.36903887998778373, 'learning_rate/qwen_vl_interface': 1.138411115813488e-06, 'learning_rate/action_model': 1.138411115813488e-05, 'epoch': 0.31}) +[2026-10-03 15:25:40,659][starVLA.training.train_starvla][INFO] - Step 4183, Loss: {'train/loss': 0.5044830441474915, 'train/grad_norm_pre_clip': 197.9166259765625, 'timing/data': 0.0003517299919622019, 'timing/model': 0.3467060740076704, 'learning_rate/qwen_vl_interface': 1.1368869962036006e-06, 'learning_rate/action_model': 1.1368869962036005e-05, 'epoch': 0.31}) +[2026-10-03 15:25:41,319][starVLA.training.train_starvla][INFO] - Step 4184, Loss: {'train/loss': 2.0859592041233554e-05, 'train/grad_norm_pre_clip': 0.01342606358230114, 'timing/data': 0.00026201999571640044, 'timing/model': 0.3706309299886925, 'learning_rate/qwen_vl_interface': 1.1353645673391025e-06, 'learning_rate/action_model': 1.1353645673391026e-05, 'epoch': 0.31}) +[2026-10-03 15:25:41,979][starVLA.training.train_starvla][INFO] - Step 4185, Loss: {'train/loss': 0.0004154477210249752, 'train/grad_norm_pre_clip': 0.42266979813575745, 'timing/data': 0.0003300099924672395, 'timing/model': 0.3709392300079344, 'learning_rate/qwen_vl_interface': 1.1338438298458078e-06, 'learning_rate/action_model': 1.1338438298458078e-05, 'epoch': 0.31}) +[2026-10-03 15:25:42,841][starVLA.training.train_starvla][INFO] - Step 4186, Loss: {'train/loss': 0.0020723629277199507, 'train/grad_norm_pre_clip': 4.021008014678955, 'timing/data': 0.00029567000456154346, 'timing/model': 0.5706568629975663, 'learning_rate/qwen_vl_interface': 1.1323247843488345e-06, 'learning_rate/action_model': 1.1323247843488344e-05, 'epoch': 0.31}) +[2026-10-03 15:25:43,499][starVLA.training.train_starvla][INFO] - Step 4187, Loss: {'train/loss': 0.21282042562961578, 'train/grad_norm_pre_clip': 160.4385223388672, 'timing/data': 0.00020686000061687082, 'timing/model': 0.37007212100434117, 'learning_rate/qwen_vl_interface': 1.130807431472603e-06, 'learning_rate/action_model': 1.130807431472603e-05, 'epoch': 0.31}) +[2026-10-03 15:25:44,159][starVLA.training.train_starvla][INFO] - Step 4188, Loss: {'train/loss': 0.0010685960296541452, 'train/grad_norm_pre_clip': 1.1725198030471802, 'timing/data': 0.00027066998882219195, 'timing/model': 0.3704306299914606, 'learning_rate/qwen_vl_interface': 1.1292917718408416e-06, 'learning_rate/action_model': 1.1292917718408416e-05, 'epoch': 0.31}) +[2026-10-03 15:25:44,819][starVLA.training.train_starvla][INFO] - Step 4189, Loss: {'train/loss': 8.384415559703484e-05, 'train/grad_norm_pre_clip': 0.03811163455247879, 'timing/data': 0.00029355000879149884, 'timing/model': 0.37138327000138815, 'learning_rate/qwen_vl_interface': 1.12777780607658e-06, 'learning_rate/action_model': 1.12777780607658e-05, 'epoch': 0.31}) +[2026-10-03 15:25:45,478][starVLA.training.train_starvla][INFO] - Step 4190, Loss: {'train/loss': 0.00015792649355717003, 'train/grad_norm_pre_clip': 0.19527050852775574, 'timing/data': 0.00037902001349721104, 'timing/model': 0.3701223800017033, 'learning_rate/qwen_vl_interface': 1.1262655348021523e-06, 'learning_rate/action_model': 1.1262655348021523e-05, 'epoch': 0.31}) +[2026-10-03 15:25:46,140][starVLA.training.train_starvla][INFO] - Step 4191, Loss: {'train/loss': 5.2973318815929815e-06, 'train/grad_norm_pre_clip': 0.0009408933110535145, 'timing/data': 0.00034314999356865883, 'timing/model': 0.36998894000134896, 'learning_rate/qwen_vl_interface': 1.1247549586391966e-06, 'learning_rate/action_model': 1.1247549586391964e-05, 'epoch': 0.31}) +[2026-10-03 15:25:47,015][starVLA.training.train_starvla][INFO] - Step 4192, Loss: {'train/loss': 0.6906273365020752, 'train/grad_norm_pre_clip': 161.96029663085938, 'timing/data': 0.00036530000215861946, 'timing/model': 0.36985222100338433, 'learning_rate/qwen_vl_interface': 1.1232460782086534e-06, 'learning_rate/action_model': 1.1232460782086534e-05, 'epoch': 0.31}) +[2026-10-03 15:25:47,680][starVLA.training.train_starvla][INFO] - Step 4193, Loss: {'train/loss': 0.31500542163848877, 'train/grad_norm_pre_clip': 74.79855346679688, 'timing/data': 0.0003373100043972954, 'timing/model': 0.3744450300000608, 'learning_rate/qwen_vl_interface': 1.1217388941307658e-06, 'learning_rate/action_model': 1.1217388941307658e-05, 'epoch': 0.31}) +[2026-10-03 15:25:48,345][starVLA.training.train_starvla][INFO] - Step 4194, Loss: {'train/loss': 1.166729907708941e-05, 'train/grad_norm_pre_clip': 0.002930351532995701, 'timing/data': 0.00027119000151287764, 'timing/model': 0.3718829499994172, 'learning_rate/qwen_vl_interface': 1.1202334070250807e-06, 'learning_rate/action_model': 1.1202334070250806e-05, 'epoch': 0.31}) +[2026-10-03 15:25:48,980][starVLA.training.train_starvla][INFO] - Step 4195, Loss: {'train/loss': 1.309028266405221e-05, 'train/grad_norm_pre_clip': 0.004648442845791578, 'timing/data': 0.0003377299872227013, 'timing/model': 0.3463195939984871, 'learning_rate/qwen_vl_interface': 1.1187296175104472e-06, 'learning_rate/action_model': 1.1187296175104471e-05, 'epoch': 0.31}) +[2026-10-03 15:25:49,642][starVLA.training.train_starvla][INFO] - Step 4196, Loss: {'train/loss': 6.183907316881232e-06, 'train/grad_norm_pre_clip': 0.0012530413223430514, 'timing/data': 0.0003382699942449108, 'timing/model': 0.3716071910021128, 'learning_rate/qwen_vl_interface': 1.1172275262050158e-06, 'learning_rate/action_model': 1.1172275262050159e-05, 'epoch': 0.31}) +[2026-10-03 15:25:50,501][starVLA.training.train_starvla][INFO] - Step 4197, Loss: {'train/loss': 9.073927503777668e-05, 'train/grad_norm_pre_clip': 0.07929662615060806, 'timing/data': 0.00023581999994348735, 'timing/model': 0.5701827729935758, 'learning_rate/qwen_vl_interface': 1.1157271337262401e-06, 'learning_rate/action_model': 1.11572713372624e-05, 'epoch': 0.31}) +[2026-10-03 15:25:51,160][starVLA.training.train_starvla][INFO] - Step 4198, Loss: {'train/loss': 2.0562638383125886e-05, 'train/grad_norm_pre_clip': 0.00559456879273057, 'timing/data': 0.0004655700031435117, 'timing/model': 0.36988494999241084, 'learning_rate/qwen_vl_interface': 1.1142284406908751e-06, 'learning_rate/action_model': 1.114228440690875e-05, 'epoch': 0.31}) +[2026-10-03 15:25:51,819][starVLA.training.train_starvla][INFO] - Step 4199, Loss: {'train/loss': 0.0008864727569743991, 'train/grad_norm_pre_clip': 1.001346468925476, 'timing/data': 0.00032605999149382114, 'timing/model': 0.3698711900069611, 'learning_rate/qwen_vl_interface': 1.1127314477149758e-06, 'learning_rate/action_model': 1.1127314477149759e-05, 'epoch': 0.31}) +[2026-10-03 15:25:52,479][starVLA.training.train_starvla][INFO] - Step 4200, Loss: {'train/loss': 0.00020449122530408204, 'train/grad_norm_pre_clip': 0.18248842656612396, 'timing/data': 0.00022038001043256372, 'timing/model': 0.37082016999193, 'learning_rate/qwen_vl_interface': 1.1112361554139003e-06, 'learning_rate/action_model': 1.1112361554139004e-05, 'epoch': 0.31}) +[2026-10-03 15:25:53,116][starVLA.training.train_starvla][INFO] - Step 4201, Loss: {'train/loss': 7.137520697142463e-06, 'train/grad_norm_pre_clip': 0.0018379685934633017, 'timing/data': 0.00030019000405445695, 'timing/model': 0.3478257030074019, 'learning_rate/qwen_vl_interface': 1.1097425644023065e-06, 'learning_rate/action_model': 1.1097425644023065e-05, 'epoch': 0.31}) +[2026-10-03 15:25:53,775][starVLA.training.train_starvla][INFO] - Step 4202, Loss: {'train/loss': 6.645866051258054e-06, 'train/grad_norm_pre_clip': 0.0014647203497588634, 'timing/data': 0.0002562099980423227, 'timing/model': 0.370160651000333, 'learning_rate/qwen_vl_interface': 1.1082506752941529e-06, 'learning_rate/action_model': 1.1082506752941528e-05, 'epoch': 0.31}) +[2026-10-03 15:25:54,635][starVLA.training.train_starvla][INFO] - Step 4203, Loss: {'train/loss': 0.001801488222554326, 'train/grad_norm_pre_clip': 1.1915080547332764, 'timing/data': 0.0003425699978834018, 'timing/model': 0.36961009999504313, 'learning_rate/qwen_vl_interface': 1.1067604887027004e-06, 'learning_rate/action_model': 1.1067604887027004e-05, 'epoch': 0.31}) +[2026-10-03 15:25:55,295][starVLA.training.train_starvla][INFO] - Step 4204, Loss: {'train/loss': 0.00012644643720705062, 'train/grad_norm_pre_clip': 0.08085421472787857, 'timing/data': 0.0003162700013490394, 'timing/model': 0.37009217000741046, 'learning_rate/qwen_vl_interface': 1.1052720052405066e-06, 'learning_rate/action_model': 1.1052720052405064e-05, 'epoch': 0.31}) +[2026-10-03 15:25:55,955][starVLA.training.train_starvla][INFO] - Step 4205, Loss: {'train/loss': 0.000284121633740142, 'train/grad_norm_pre_clip': 0.112451933324337, 'timing/data': 0.0003731500037247315, 'timing/model': 0.3717205699940678, 'learning_rate/qwen_vl_interface': 1.1037852255194309e-06, 'learning_rate/action_model': 1.1037852255194308e-05, 'epoch': 0.31}) +[2026-10-03 15:25:56,614][starVLA.training.train_starvla][INFO] - Step 4206, Loss: {'train/loss': 1.6963589587248862e-05, 'train/grad_norm_pre_clip': 0.015898777171969414, 'timing/data': 0.00033325000549666584, 'timing/model': 0.37008171099296305, 'learning_rate/qwen_vl_interface': 1.1023001501506329e-06, 'learning_rate/action_model': 1.102300150150633e-05, 'epoch': 0.31}) +[2026-10-03 15:25:57,273][starVLA.training.train_starvla][INFO] - Step 4207, Loss: {'train/loss': 0.374789834022522, 'train/grad_norm_pre_clip': 71.55729675292969, 'timing/data': 0.0002563900052336976, 'timing/model': 0.36977059999480844, 'learning_rate/qwen_vl_interface': 1.1008167797445717e-06, 'learning_rate/action_model': 1.1008167797445718e-05, 'epoch': 0.31}) +[2026-10-03 15:25:58,123][starVLA.training.train_starvla][INFO] - Step 4208, Loss: {'train/loss': 0.4152187407016754, 'train/grad_norm_pre_clip': 145.4026641845703, 'timing/data': 0.0003591799904825166, 'timing/model': 0.5610463049961254, 'learning_rate/qwen_vl_interface': 1.099335114911003e-06, 'learning_rate/action_model': 1.0993351149110031e-05, 'epoch': 0.31}) +[2026-10-03 15:25:58,760][starVLA.training.train_starvla][INFO] - Step 4209, Loss: {'train/loss': 0.0001225117448484525, 'train/grad_norm_pre_clip': 0.10644064098596573, 'timing/data': 0.0003112099948339164, 'timing/model': 0.37092853100330103, 'learning_rate/qwen_vl_interface': 1.0978551562589846e-06, 'learning_rate/action_model': 1.0978551562589845e-05, 'epoch': 0.31}) +[2026-10-03 15:25:59,420][starVLA.training.train_starvla][INFO] - Step 4210, Loss: {'train/loss': 0.0006828975747339427, 'train/grad_norm_pre_clip': 0.6054621934890747, 'timing/data': 0.00021800999820698053, 'timing/model': 0.369901240002946, 'learning_rate/qwen_vl_interface': 1.0963769043968708e-06, 'learning_rate/action_model': 1.0963769043968708e-05, 'epoch': 0.31}) +[2026-10-03 15:26:00,078][starVLA.training.train_starvla][INFO] - Step 4211, Loss: {'train/loss': 0.0006994458963163197, 'train/grad_norm_pre_clip': 0.8344730138778687, 'timing/data': 0.00033318999339826405, 'timing/model': 0.3699373699928401, 'learning_rate/qwen_vl_interface': 1.0949003599323158e-06, 'learning_rate/action_model': 1.0949003599323157e-05, 'epoch': 0.31}) +[2026-10-03 15:26:00,738][starVLA.training.train_starvla][INFO] - Step 4212, Loss: {'train/loss': 5.446338491310598e-06, 'train/grad_norm_pre_clip': 0.0010910599958151579, 'timing/data': 0.00032979999377857894, 'timing/model': 0.3712059599929489, 'learning_rate/qwen_vl_interface': 1.0934255234722708e-06, 'learning_rate/action_model': 1.0934255234722706e-05, 'epoch': 0.31}) +[2026-10-03 15:26:01,593][starVLA.training.train_starvla][INFO] - Step 4213, Loss: {'train/loss': 1.7135596863226965e-05, 'train/grad_norm_pre_clip': 0.006104776635766029, 'timing/data': 0.00026359999901615083, 'timing/model': 0.5655504729948007, 'learning_rate/qwen_vl_interface': 1.0919523956229858e-06, 'learning_rate/action_model': 1.0919523956229857e-05, 'epoch': 0.31}) +[2026-10-03 15:26:02,252][starVLA.training.train_starvla][INFO] - Step 4214, Loss: {'train/loss': 2.3490334569942206e-05, 'train/grad_norm_pre_clip': 0.013274455443024635, 'timing/data': 0.0003799700061790645, 'timing/model': 0.36963469999318477, 'learning_rate/qwen_vl_interface': 1.0904809769900075e-06, 'learning_rate/action_model': 1.0904809769900073e-05, 'epoch': 0.31}) +[2026-10-03 15:26:02,910][starVLA.training.train_starvla][INFO] - Step 4215, Loss: {'train/loss': 0.000301061401842162, 'train/grad_norm_pre_clip': 0.10527359694242477, 'timing/data': 0.0003034800029126927, 'timing/model': 0.370292840001639, 'learning_rate/qwen_vl_interface': 1.0890112681781821e-06, 'learning_rate/action_model': 1.089011268178182e-05, 'epoch': 0.31}) +[2026-10-03 15:26:03,571][starVLA.training.train_starvla][INFO] - Step 4216, Loss: {'train/loss': 8.344544767169282e-06, 'train/grad_norm_pre_clip': 0.0023552279453724623, 'timing/data': 0.0002939099940704182, 'timing/model': 0.37114872099482454, 'learning_rate/qwen_vl_interface': 1.0875432697916492e-06, 'learning_rate/action_model': 1.087543269791649e-05, 'epoch': 0.31}) +[2026-10-03 15:26:04,231][starVLA.training.train_starvla][INFO] - Step 4217, Loss: {'train/loss': 0.00011794574675150216, 'train/grad_norm_pre_clip': 0.06987196952104568, 'timing/data': 0.00034022000909317285, 'timing/model': 0.3706942499993602, 'learning_rate/qwen_vl_interface': 1.0860769824338477e-06, 'learning_rate/action_model': 1.0860769824338478e-05, 'epoch': 0.31}) +[2026-10-03 15:26:05,086][starVLA.training.train_starvla][INFO] - Step 4218, Loss: {'train/loss': 0.06548353284597397, 'train/grad_norm_pre_clip': 131.3203125, 'timing/data': 0.00035149999894201756, 'timing/model': 0.5664609639934497, 'learning_rate/qwen_vl_interface': 1.084612406707514e-06, 'learning_rate/action_model': 1.084612406707514e-05, 'epoch': 0.31}) +[2026-10-03 15:26:05,745][starVLA.training.train_starvla][INFO] - Step 4219, Loss: {'train/loss': 4.398441524244845e-05, 'train/grad_norm_pre_clip': 0.021615775302052498, 'timing/data': 0.00033025999437086284, 'timing/model': 0.3708391899999697, 'learning_rate/qwen_vl_interface': 1.0831495432146795e-06, 'learning_rate/action_model': 1.0831495432146795e-05, 'epoch': 0.31}) +[2026-10-03 15:26:06,405][starVLA.training.train_starvla][INFO] - Step 4220, Loss: {'train/loss': 0.4288630187511444, 'train/grad_norm_pre_clip': 41.47052001953125, 'timing/data': 0.0002698600001167506, 'timing/model': 0.37014020000060555, 'learning_rate/qwen_vl_interface': 1.0816883925566717e-06, 'learning_rate/action_model': 1.0816883925566716e-05, 'epoch': 0.31}) +[2026-10-03 15:26:07,065][starVLA.training.train_starvla][INFO] - Step 4221, Loss: {'train/loss': 0.6303219199180603, 'train/grad_norm_pre_clip': 33.97076416015625, 'timing/data': 0.0003390199999557808, 'timing/model': 0.3708960899966769, 'learning_rate/qwen_vl_interface': 1.0802289553341153e-06, 'learning_rate/action_model': 1.0802289553341152e-05, 'epoch': 0.31}) +[2026-10-03 15:26:07,701][starVLA.training.train_starvla][INFO] - Step 4222, Loss: {'train/loss': 0.00014226097846403718, 'train/grad_norm_pre_clip': 0.11708910018205643, 'timing/data': 0.00033806001010816544, 'timing/model': 0.34653226399677806, 'learning_rate/qwen_vl_interface': 1.0787712321469289e-06, 'learning_rate/action_model': 1.0787712321469288e-05, 'epoch': 0.31}) +[2026-10-03 15:26:08,563][starVLA.training.train_starvla][INFO] - Step 4223, Loss: {'train/loss': 2.7790597414423246e-06, 'train/grad_norm_pre_clip': 0.0004178271337877959, 'timing/data': 0.00029039999935775995, 'timing/model': 0.5722667530062608, 'learning_rate/qwen_vl_interface': 1.077315223594328e-06, 'learning_rate/action_model': 1.077315223594328e-05, 'epoch': 0.31}) +[2026-10-03 15:26:09,223][starVLA.training.train_starvla][INFO] - Step 4224, Loss: {'train/loss': 0.0002073246578220278, 'train/grad_norm_pre_clip': 0.5231696963310242, 'timing/data': 0.000257949999650009, 'timing/model': 0.3700755699974252, 'learning_rate/qwen_vl_interface': 1.0758609302748227e-06, 'learning_rate/action_model': 1.0758609302748227e-05, 'epoch': 0.31}) +[2026-10-03 15:26:09,883][starVLA.training.train_starvla][INFO] - Step 4225, Loss: {'train/loss': 0.12282396107912064, 'train/grad_norm_pre_clip': 111.00572967529297, 'timing/data': 0.00033912000071723014, 'timing/model': 0.3713273499888601, 'learning_rate/qwen_vl_interface': 1.0744083527862182e-06, 'learning_rate/action_model': 1.074408352786218e-05, 'epoch': 0.31}) +[2026-10-03 15:26:10,519][starVLA.training.train_starvla][INFO] - Step 4226, Loss: {'train/loss': 1.5011732102720998e-05, 'train/grad_norm_pre_clip': 0.007568763103336096, 'timing/data': 0.00034764000156428665, 'timing/model': 0.3701111900008982, 'learning_rate/qwen_vl_interface': 1.0729574917256144e-06, 'learning_rate/action_model': 1.0729574917256145e-05, 'epoch': 0.31}) +[2026-10-03 15:26:11,178][starVLA.training.train_starvla][INFO] - Step 4227, Loss: {'train/loss': 4.81304778077174e-06, 'train/grad_norm_pre_clip': 0.000819354725535959, 'timing/data': 0.00031625998963136226, 'timing/model': 0.36938690999522805, 'learning_rate/qwen_vl_interface': 1.0715083476894047e-06, 'learning_rate/action_model': 1.0715083476894047e-05, 'epoch': 0.31}) +[2026-10-03 15:26:11,838][starVLA.training.train_starvla][INFO] - Step 4228, Loss: {'train/loss': 1.2598626199178398e-05, 'train/grad_norm_pre_clip': 0.0036297703627496958, 'timing/data': 0.00035566999576985836, 'timing/model': 0.37027891000616364, 'learning_rate/qwen_vl_interface': 1.0700609212732773e-06, 'learning_rate/action_model': 1.0700609212732774e-05, 'epoch': 0.31}) +[2026-10-03 15:26:12,698][starVLA.training.train_starvla][INFO] - Step 4229, Loss: {'train/loss': 4.151857865508646e-05, 'train/grad_norm_pre_clip': 0.022262291982769966, 'timing/data': 0.00036211899714544415, 'timing/model': 0.3710860510036582, 'learning_rate/qwen_vl_interface': 1.0686152130722155e-06, 'learning_rate/action_model': 1.0686152130722154e-05, 'epoch': 0.31}) +[2026-10-03 15:26:13,357][starVLA.training.train_starvla][INFO] - Step 4230, Loss: {'train/loss': 4.962061666446971e-06, 'train/grad_norm_pre_clip': 0.00141006032936275, 'timing/data': 0.0002957599936053157, 'timing/model': 0.3699608410097426, 'learning_rate/qwen_vl_interface': 1.0671712236804947e-06, 'learning_rate/action_model': 1.0671712236804946e-05, 'epoch': 0.31}) +[2026-10-03 15:26:14,016][starVLA.training.train_starvla][INFO] - Step 4231, Loss: {'train/loss': 3.024927309525083e-06, 'train/grad_norm_pre_clip': 0.0007023739162832499, 'timing/data': 0.0003684700059238821, 'timing/model': 0.36940486000094097, 'learning_rate/qwen_vl_interface': 1.0657289536916845e-06, 'learning_rate/action_model': 1.0657289536916845e-05, 'epoch': 0.31}) +[2026-10-03 15:26:14,653][starVLA.training.train_starvla][INFO] - Step 4232, Loss: {'train/loss': 0.0006311660981737077, 'train/grad_norm_pre_clip': 0.7421110272407532, 'timing/data': 0.0002631299867061898, 'timing/model': 0.3707052600075258, 'learning_rate/qwen_vl_interface': 1.0642884036986481e-06, 'learning_rate/action_model': 1.064288403698648e-05, 'epoch': 0.31}) +[2026-10-03 15:26:15,315][starVLA.training.train_starvla][INFO] - Step 4233, Loss: {'train/loss': 9.573710485710762e-06, 'train/grad_norm_pre_clip': 0.0037031949032098055, 'timing/data': 0.0004266900068614632, 'timing/model': 0.3724233999964781, 'learning_rate/qwen_vl_interface': 1.0628495742935404e-06, 'learning_rate/action_model': 1.0628495742935405e-05, 'epoch': 0.31}) +[2026-10-03 15:26:16,198][starVLA.training.train_starvla][INFO] - Step 4234, Loss: {'train/loss': 1.1175475265190471e-05, 'train/grad_norm_pre_clip': 0.003154695499688387, 'timing/data': 0.00042301999928895384, 'timing/model': 0.37328969000373036, 'learning_rate/qwen_vl_interface': 1.0614124660678103e-06, 'learning_rate/action_model': 1.0614124660678102e-05, 'epoch': 0.31}) +[2026-10-03 15:26:16,860][starVLA.training.train_starvla][INFO] - Step 4235, Loss: {'train/loss': 0.00017155034583993256, 'train/grad_norm_pre_clip': 0.14409559965133667, 'timing/data': 0.00034864000917878, 'timing/model': 0.3709994709934108, 'learning_rate/qwen_vl_interface': 1.0599770796121988e-06, 'learning_rate/action_model': 1.0599770796121987e-05, 'epoch': 0.31}) +[2026-10-03 15:26:17,525][starVLA.training.train_starvla][INFO] - Step 4236, Loss: {'train/loss': 6.102847328293137e-05, 'train/grad_norm_pre_clip': 0.08466888219118118, 'timing/data': 0.00034548000257927924, 'timing/model': 0.37401507000322454, 'learning_rate/qwen_vl_interface': 1.0585434155167394e-06, 'learning_rate/action_model': 1.0585434155167393e-05, 'epoch': 0.31}) +[2026-10-03 15:26:18,190][starVLA.training.train_starvla][INFO] - Step 4237, Loss: {'train/loss': 0.3243016302585602, 'train/grad_norm_pre_clip': 71.4781723022461, 'timing/data': 0.0003510999958962202, 'timing/model': 0.37300737999612466, 'learning_rate/qwen_vl_interface': 1.0571114743707556e-06, 'learning_rate/action_model': 1.0571114743707555e-05, 'epoch': 0.31}) +[2026-10-03 15:26:18,852][starVLA.training.train_starvla][INFO] - Step 4238, Loss: {'train/loss': 0.3750048279762268, 'train/grad_norm_pre_clip': 56.18779373168945, 'timing/data': 0.0003287699946667999, 'timing/model': 0.371468930010451, 'learning_rate/qwen_vl_interface': 1.0556812567628661e-06, 'learning_rate/action_model': 1.055681256762866e-05, 'epoch': 0.31}) +[2026-10-03 15:26:19,489][starVLA.training.train_starvla][INFO] - Step 4239, Loss: {'train/loss': 0.5080935955047607, 'train/grad_norm_pre_clip': 42.95823287963867, 'timing/data': 0.00021494000975508243, 'timing/model': 0.37044590098957997, 'learning_rate/qwen_vl_interface': 1.0542527632809783e-06, 'learning_rate/action_model': 1.0542527632809784e-05, 'epoch': 0.31}) +[2026-10-03 15:26:20,345][starVLA.training.train_starvla][INFO] - Step 4240, Loss: {'train/loss': 0.0007658322574570775, 'train/grad_norm_pre_clip': 1.7116197347640991, 'timing/data': 0.00031493999995291233, 'timing/model': 0.3713058400026057, 'learning_rate/qwen_vl_interface': 1.0528259945122925e-06, 'learning_rate/action_model': 1.0528259945122925e-05, 'epoch': 0.31}) +[2026-10-03 15:26:20,984][starVLA.training.train_starvla][INFO] - Step 4241, Loss: {'train/loss': 3.143011417705566e-05, 'train/grad_norm_pre_clip': 0.023373732343316078, 'timing/data': 0.0003586999955587089, 'timing/model': 0.3483304230030626, 'learning_rate/qwen_vl_interface': 1.0514009510432988e-06, 'learning_rate/action_model': 1.0514009510432988e-05, 'epoch': 0.31}) +[2026-10-03 15:26:21,643][starVLA.training.train_starvla][INFO] - Step 4242, Loss: {'train/loss': 1.414215087890625, 'train/grad_norm_pre_clip': 31.020952224731445, 'timing/data': 0.0004584199923556298, 'timing/model': 0.3702092699968489, 'learning_rate/qwen_vl_interface': 1.0499776334597796e-06, 'learning_rate/action_model': 1.0499776334597797e-05, 'epoch': 0.31}) +[2026-10-03 15:26:22,301][starVLA.training.train_starvla][INFO] - Step 4243, Loss: {'train/loss': 6.874662358313799e-05, 'train/grad_norm_pre_clip': 0.04210184887051582, 'timing/data': 0.0003167400136590004, 'timing/model': 0.3700468600000022, 'learning_rate/qwen_vl_interface': 1.0485560423468064e-06, 'learning_rate/action_model': 1.0485560423468063e-05, 'epoch': 0.31}) +[2026-10-03 15:26:22,962][starVLA.training.train_starvla][INFO] - Step 4244, Loss: {'train/loss': 0.0032961517572402954, 'train/grad_norm_pre_clip': 3.506985902786255, 'timing/data': 0.00031709000177215785, 'timing/model': 0.3710543009947287, 'learning_rate/qwen_vl_interface': 1.0471361782887421e-06, 'learning_rate/action_model': 1.0471361782887421e-05, 'epoch': 0.31}) +[2026-10-03 15:26:23,817][starVLA.training.train_starvla][INFO] - Step 4245, Loss: {'train/loss': 2.5480926524323877e-06, 'train/grad_norm_pre_clip': 0.0007316021947190166, 'timing/data': 0.0003405600000405684, 'timing/model': 0.371601089995238, 'learning_rate/qwen_vl_interface': 1.0457180418692385e-06, 'learning_rate/action_model': 1.0457180418692383e-05, 'epoch': 0.31}) +[2026-10-03 15:26:24,476][starVLA.training.train_starvla][INFO] - Step 4246, Loss: {'train/loss': 1.7828322597779334e-05, 'train/grad_norm_pre_clip': 0.005202229600399733, 'timing/data': 0.00026471000455785543, 'timing/model': 0.3697663009952521, 'learning_rate/qwen_vl_interface': 1.0443016336712395e-06, 'learning_rate/action_model': 1.0443016336712393e-05, 'epoch': 0.31}) +[2026-10-03 15:26:25,135][starVLA.training.train_starvla][INFO] - Step 4247, Loss: {'train/loss': 0.00018077477579936385, 'train/grad_norm_pre_clip': 0.20054258406162262, 'timing/data': 0.0003386000025784597, 'timing/model': 0.3700216309953248, 'learning_rate/qwen_vl_interface': 1.0428869542769744e-06, 'learning_rate/action_model': 1.0428869542769744e-05, 'epoch': 0.31}) +[2026-10-03 15:26:25,795][starVLA.training.train_starvla][INFO] - Step 4248, Loss: {'train/loss': 0.33218783140182495, 'train/grad_norm_pre_clip': 80.08126831054688, 'timing/data': 0.0003260500088799745, 'timing/model': 0.37075800099410117, 'learning_rate/qwen_vl_interface': 1.0414740042679663e-06, 'learning_rate/action_model': 1.0414740042679662e-05, 'epoch': 0.31}) +[2026-10-03 15:26:26,455][starVLA.training.train_starvla][INFO] - Step 4249, Loss: {'train/loss': 0.40241512656211853, 'train/grad_norm_pre_clip': 44.73230743408203, 'timing/data': 0.00033894000807777047, 'timing/model': 0.37075652000203263, 'learning_rate/qwen_vl_interface': 1.040062784225025e-06, 'learning_rate/action_model': 1.0400627842250249e-05, 'epoch': 0.31}) +[2026-10-03 15:26:27,294][starVLA.training.train_starvla][INFO] - Step 4250, Loss: {'train/loss': 2.0248191503924318e-05, 'train/grad_norm_pre_clip': 0.007507889997214079, 'timing/data': 0.0003661499940790236, 'timing/model': 0.34770310299063567, 'learning_rate/qwen_vl_interface': 1.0386532947282496e-06, 'learning_rate/action_model': 1.0386532947282495e-05, 'epoch': 0.31}) +[2026-10-03 15:26:27,959][starVLA.training.train_starvla][INFO] - Step 4251, Loss: {'train/loss': 0.0006322374101728201, 'train/grad_norm_pre_clip': 0.6570133566856384, 'timing/data': 0.0003331299958517775, 'timing/model': 0.37319302999821957, 'learning_rate/qwen_vl_interface': 1.0372455363570282e-06, 'learning_rate/action_model': 1.0372455363570283e-05, 'epoch': 0.31}) +[2026-10-03 15:26:28,627][starVLA.training.train_starvla][INFO] - Step 4252, Loss: {'train/loss': 0.0002363290113862604, 'train/grad_norm_pre_clip': 0.44525623321533203, 'timing/data': 0.00035026999830733985, 'timing/model': 0.3743183900078293, 'learning_rate/qwen_vl_interface': 1.035839509690039e-06, 'learning_rate/action_model': 1.0358395096900388e-05, 'epoch': 0.31}) +[2026-10-03 15:26:29,293][starVLA.training.train_starvla][INFO] - Step 4253, Loss: {'train/loss': 0.08206532150506973, 'train/grad_norm_pre_clip': 157.11349487304688, 'timing/data': 0.00034108999534510076, 'timing/model': 0.3732968399999663, 'learning_rate/qwen_vl_interface': 1.0344352153052429e-06, 'learning_rate/action_model': 1.0344352153052427e-05, 'epoch': 0.31}) +[2026-10-03 15:26:29,956][starVLA.training.train_starvla][INFO] - Step 4254, Loss: {'train/loss': 0.3363646864891052, 'train/grad_norm_pre_clip': 174.922119140625, 'timing/data': 0.0002067199966404587, 'timing/model': 0.37261192000005394, 'learning_rate/qwen_vl_interface': 1.033032653779894e-06, 'learning_rate/action_model': 1.033032653779894e-05, 'epoch': 0.31}) +[2026-10-03 15:26:30,818][starVLA.training.train_starvla][INFO] - Step 4255, Loss: {'train/loss': 0.1876145899295807, 'train/grad_norm_pre_clip': 172.85586547851562, 'timing/data': 0.00020719000895041972, 'timing/model': 0.5707890939956997, 'learning_rate/qwen_vl_interface': 1.0316318256905325e-06, 'learning_rate/action_model': 1.0316318256905324e-05, 'epoch': 0.31}) +[2026-10-03 15:26:31,483][starVLA.training.train_starvla][INFO] - Step 4256, Loss: {'train/loss': 7.219527105917223e-06, 'train/grad_norm_pre_clip': 0.002173860091716051, 'timing/data': 0.0003946700016967952, 'timing/model': 0.37367550999624655, 'learning_rate/qwen_vl_interface': 1.0302327316129856e-06, 'learning_rate/action_model': 1.0302327316129855e-05, 'epoch': 0.31}) +[2026-10-03 15:26:32,142][starVLA.training.train_starvla][INFO] - Step 4257, Loss: {'train/loss': 9.864459570962936e-06, 'train/grad_norm_pre_clip': 0.0022205410059541464, 'timing/data': 0.00028685999859590083, 'timing/model': 0.3707407600013539, 'learning_rate/qwen_vl_interface': 1.028835372122368e-06, 'learning_rate/action_model': 1.0288353721223682e-05, 'epoch': 0.31}) +[2026-10-03 15:26:32,800][starVLA.training.train_starvla][INFO] - Step 4258, Loss: {'train/loss': 3.104399729636498e-05, 'train/grad_norm_pre_clip': 0.023084571585059166, 'timing/data': 0.0002492200001142919, 'timing/model': 0.37006097999983467, 'learning_rate/qwen_vl_interface': 1.0274397477930822e-06, 'learning_rate/action_model': 1.0274397477930822e-05, 'epoch': 0.31}) +[2026-10-03 15:26:33,459][starVLA.training.train_starvla][INFO] - Step 4259, Loss: {'train/loss': 0.0009782480774447322, 'train/grad_norm_pre_clip': 0.7734466195106506, 'timing/data': 0.00034571999276522547, 'timing/model': 0.3699476010078797, 'learning_rate/qwen_vl_interface': 1.0260458591988164e-06, 'learning_rate/action_model': 1.0260458591988165e-05, 'epoch': 0.31}) +[2026-10-03 15:26:34,327][starVLA.training.train_starvla][INFO] - Step 4260, Loss: {'train/loss': 0.00022772655938751996, 'train/grad_norm_pre_clip': 0.0943078026175499, 'timing/data': 0.00028001000464428216, 'timing/model': 0.5783637919957982, 'learning_rate/qwen_vl_interface': 1.0246537069125445e-06, 'learning_rate/action_model': 1.0246537069125444e-05, 'epoch': 0.31}) +[2026-10-03 15:26:34,987][starVLA.training.train_starvla][INFO] - Step 4261, Loss: {'train/loss': 1.4952788660593797e-05, 'train/grad_norm_pre_clip': 0.0041229682974517345, 'timing/data': 0.00033822000841610134, 'timing/model': 0.3707060500018997, 'learning_rate/qwen_vl_interface': 1.0232632915065286e-06, 'learning_rate/action_model': 1.0232632915065285e-05, 'epoch': 0.31}) +[2026-10-03 15:26:35,646][starVLA.training.train_starvla][INFO] - Step 4262, Loss: {'train/loss': 4.537917993729934e-05, 'train/grad_norm_pre_clip': 0.0329035259783268, 'timing/data': 0.0002543599985074252, 'timing/model': 0.36978973999794107, 'learning_rate/qwen_vl_interface': 1.021874613552316e-06, 'learning_rate/action_model': 1.0218746135523157e-05, 'epoch': 0.31}) +[2026-10-03 15:26:36,305][starVLA.training.train_starvla][INFO] - Step 4263, Loss: {'train/loss': 0.000501524715218693, 'train/grad_norm_pre_clip': 0.9232535362243652, 'timing/data': 0.00029950999305583537, 'timing/model': 0.36991937999846414, 'learning_rate/qwen_vl_interface': 1.0204876736207384e-06, 'learning_rate/action_model': 1.0204876736207384e-05, 'epoch': 0.31}) +[2026-10-03 15:26:36,966][starVLA.training.train_starvla][INFO] - Step 4264, Loss: {'train/loss': 0.00010083267989102751, 'train/grad_norm_pre_clip': 0.07034455239772797, 'timing/data': 0.0003909999941242859, 'timing/model': 0.37140456099587027, 'learning_rate/qwen_vl_interface': 1.0191024722819163e-06, 'learning_rate/action_model': 1.0191024722819162e-05, 'epoch': 0.31}) +[2026-10-03 15:26:37,626][starVLA.training.train_starvla][INFO] - Step 4265, Loss: {'train/loss': 8.947872629505582e-06, 'train/grad_norm_pre_clip': 0.00255390047095716, 'timing/data': 0.00032670999644324183, 'timing/model': 0.37142841999593657, 'learning_rate/qwen_vl_interface': 1.017719010105253e-06, 'learning_rate/action_model': 1.017719010105253e-05, 'epoch': 0.31}) +[2026-10-03 15:26:38,462][starVLA.training.train_starvla][INFO] - Step 4266, Loss: {'train/loss': 3.6609657399822026e-05, 'train/grad_norm_pre_clip': 0.011013247072696686, 'timing/data': 0.00035519999801181257, 'timing/model': 0.34742901400022674, 'learning_rate/qwen_vl_interface': 1.016337287659437e-06, 'learning_rate/action_model': 1.0163372876594368e-05, 'epoch': 0.31}) +[2026-10-03 15:26:39,121][starVLA.training.train_starvla][INFO] - Step 4267, Loss: {'train/loss': 0.031234852969646454, 'train/grad_norm_pre_clip': 93.77438354492188, 'timing/data': 0.00035369000397622585, 'timing/model': 0.37025221101066563, 'learning_rate/qwen_vl_interface': 1.0149573055124427e-06, 'learning_rate/action_model': 1.0149573055124426e-05, 'epoch': 0.31}) +[2026-10-03 15:26:39,758][starVLA.training.train_starvla][INFO] - Step 4268, Loss: {'train/loss': 1.4714412827743217e-05, 'train/grad_norm_pre_clip': 0.004682475700974464, 'timing/data': 0.00027393999334890395, 'timing/model': 0.34750579298997764, 'learning_rate/qwen_vl_interface': 1.013579064231529e-06, 'learning_rate/action_model': 1.013579064231529e-05, 'epoch': 0.31}) +[2026-10-03 15:26:40,419][starVLA.training.train_starvla][INFO] - Step 4269, Loss: {'train/loss': 6.721440877299756e-05, 'train/grad_norm_pre_clip': 0.024090003222227097, 'timing/data': 0.0002980499994009733, 'timing/model': 0.3705113309988519, 'learning_rate/qwen_vl_interface': 1.012202564383239e-06, 'learning_rate/action_model': 1.0122025643832388e-05, 'epoch': 0.31}) +[2026-10-03 15:26:41,054][starVLA.training.train_starvla][INFO] - Step 4270, Loss: {'train/loss': 1.4281903531809803e-05, 'train/grad_norm_pre_clip': 0.008480529300868511, 'timing/data': 0.00036539998836815357, 'timing/model': 0.3466142130055232, 'learning_rate/qwen_vl_interface': 1.0108278065333978e-06, 'learning_rate/action_model': 1.0108278065333978e-05, 'epoch': 0.31}) +[2026-10-03 15:26:41,888][starVLA.training.train_starvla][INFO] - Step 4271, Loss: {'train/loss': 6.414892595785204e-06, 'train/grad_norm_pre_clip': 0.001541712787002325, 'timing/data': 0.00034876000427175313, 'timing/model': 0.5689204439986497, 'learning_rate/qwen_vl_interface': 1.0094547912471183e-06, 'learning_rate/action_model': 1.0094547912471184e-05, 'epoch': 0.31}) +[2026-10-03 15:26:42,548][starVLA.training.train_starvla][INFO] - Step 4272, Loss: {'train/loss': 0.04956507682800293, 'train/grad_norm_pre_clip': 126.25796508789062, 'timing/data': 0.0003493900003377348, 'timing/model': 0.37128805999236647, 'learning_rate/qwen_vl_interface': 1.0080835190887949e-06, 'learning_rate/action_model': 1.0080835190887947e-05, 'epoch': 0.31}) +[2026-10-03 15:26:43,209][starVLA.training.train_starvla][INFO] - Step 4273, Loss: {'train/loss': 4.030744548799703e-06, 'train/grad_norm_pre_clip': 0.0010001902701333165, 'timing/data': 0.0003166499955113977, 'timing/model': 0.37085202999878675, 'learning_rate/qwen_vl_interface': 1.0067139906221052e-06, 'learning_rate/action_model': 1.0067139906221052e-05, 'epoch': 0.31}) +[2026-10-03 15:26:43,868][starVLA.training.train_starvla][INFO] - Step 4274, Loss: {'train/loss': 2.6574383809929714e-05, 'train/grad_norm_pre_clip': 0.007802979089319706, 'timing/data': 0.0002756300091277808, 'timing/model': 0.3697257199964952, 'learning_rate/qwen_vl_interface': 1.0053462064100112e-06, 'learning_rate/action_model': 1.0053462064100112e-05, 'epoch': 0.31}) +[2026-10-03 15:26:44,527][starVLA.training.train_starvla][INFO] - Step 4275, Loss: {'train/loss': 4.596972303261282e-06, 'train/grad_norm_pre_clip': 0.0012117886217311025, 'timing/data': 0.0003512599942041561, 'timing/model': 0.37012761100777425, 'learning_rate/qwen_vl_interface': 1.003980167014757e-06, 'learning_rate/action_model': 1.003980167014757e-05, 'epoch': 0.31}) +[2026-10-03 15:26:45,379][starVLA.training.train_starvla][INFO] - Step 4276, Loss: {'train/loss': 0.00016396073624491692, 'train/grad_norm_pre_clip': 0.10787201672792435, 'timing/data': 0.0002983900049002841, 'timing/model': 0.5627778150083032, 'learning_rate/qwen_vl_interface': 1.0026158729978702e-06, 'learning_rate/action_model': 1.0026158729978702e-05, 'epoch': 0.31}) +[2026-10-03 15:26:46,042][starVLA.training.train_starvla][INFO] - Step 4277, Loss: {'train/loss': 0.2459687441587448, 'train/grad_norm_pre_clip': 105.0854721069336, 'timing/data': 0.0003475399862509221, 'timing/model': 0.3706793099991046, 'learning_rate/qwen_vl_interface': 1.0012533249201604e-06, 'learning_rate/action_model': 1.0012533249201602e-05, 'epoch': 0.31}) +[2026-10-03 15:26:46,707][starVLA.training.train_starvla][INFO] - Step 4278, Loss: {'train/loss': 9.669906285125762e-05, 'train/grad_norm_pre_clip': 0.036016467958688736, 'timing/data': 0.00036673000431619585, 'timing/model': 0.3729382000019541, 'learning_rate/qwen_vl_interface': 9.99892523341719e-07, 'learning_rate/action_model': 9.99892523341719e-06, 'epoch': 0.31}) +[2026-10-03 15:26:47,369][starVLA.training.train_starvla][INFO] - Step 4279, Loss: {'train/loss': 0.0007693850202485919, 'train/grad_norm_pre_clip': 0.6795182824134827, 'timing/data': 0.00034250000317115337, 'timing/model': 0.3719176199956564, 'learning_rate/qwen_vl_interface': 9.98533468821922e-07, 'learning_rate/action_model': 9.98533468821922e-06, 'epoch': 0.31}) +[2026-10-03 15:26:48,034][starVLA.training.train_starvla][INFO] - Step 4280, Loss: {'train/loss': 9.684944234322757e-05, 'train/grad_norm_pre_clip': 0.26593053340911865, 'timing/data': 0.0003468100039754063, 'timing/model': 0.37356631000875495, 'learning_rate/qwen_vl_interface': 9.971761619194231e-07, 'learning_rate/action_model': 9.97176161919423e-06, 'epoch': 0.31}) +[2026-10-03 15:26:48,700][starVLA.training.train_starvla][INFO] - Step 4281, Loss: {'train/loss': 7.83041468821466e-06, 'train/grad_norm_pre_clip': 0.008248821832239628, 'timing/data': 0.00036407999868970364, 'timing/model': 0.37405388000479434, 'learning_rate/qwen_vl_interface': 9.958206031921606e-07, 'learning_rate/action_model': 9.958206031921604e-06, 'epoch': 0.31}) +[2026-10-03 15:26:49,543][starVLA.training.train_starvla][INFO] - Step 4282, Loss: {'train/loss': 0.0008745311642996967, 'train/grad_norm_pre_clip': 0.8180174231529236, 'timing/data': 0.00022432999685406685, 'timing/model': 0.34956335299648345, 'learning_rate/qwen_vl_interface': 9.944667931973537e-07, 'learning_rate/action_model': 9.944667931973536e-06, 'epoch': 0.31}) +[2026-10-03 15:26:50,202][starVLA.training.train_starvla][INFO] - Step 4283, Loss: {'train/loss': 0.00026061030803248286, 'train/grad_norm_pre_clip': 0.086018405854702, 'timing/data': 0.0003760800027521327, 'timing/model': 0.3698606510006357, 'learning_rate/qwen_vl_interface': 9.93114732491503e-07, 'learning_rate/action_model': 9.93114732491503e-06, 'epoch': 0.31}) +[2026-10-03 15:26:50,864][starVLA.training.train_starvla][INFO] - Step 4284, Loss: {'train/loss': 0.0003990507684648037, 'train/grad_norm_pre_clip': 0.2952691614627838, 'timing/data': 0.0002650300011737272, 'timing/model': 0.3716453700035345, 'learning_rate/qwen_vl_interface': 9.91764421630389e-07, 'learning_rate/action_model': 9.91764421630389e-06, 'epoch': 0.31}) +[2026-10-03 15:26:51,522][starVLA.training.train_starvla][INFO] - Step 4285, Loss: {'train/loss': 0.00013566216512117535, 'train/grad_norm_pre_clip': 0.04408600181341171, 'timing/data': 0.0003367400058778003, 'timing/model': 0.3702935100009199, 'learning_rate/qwen_vl_interface': 9.904158611690732e-07, 'learning_rate/action_model': 9.90415861169073e-06, 'epoch': 0.31}) +[2026-10-03 15:26:52,181][starVLA.training.train_starvla][INFO] - Step 4286, Loss: {'train/loss': 0.0002025314752245322, 'train/grad_norm_pre_clip': 0.18147039413452148, 'timing/data': 0.00033660999906715006, 'timing/model': 0.36938038999505807, 'learning_rate/qwen_vl_interface': 9.890690516618987e-07, 'learning_rate/action_model': 9.890690516618985e-06, 'epoch': 0.31}) +[2026-10-03 15:26:53,031][starVLA.training.train_starvla][INFO] - Step 4287, Loss: {'train/loss': 0.6314719319343567, 'train/grad_norm_pre_clip': 132.0747833251953, 'timing/data': 0.000344469997799024, 'timing/model': 0.5608597149985144, 'learning_rate/qwen_vl_interface': 9.877239936624875e-07, 'learning_rate/action_model': 9.877239936624875e-06, 'epoch': 0.31}) +[2026-10-03 15:26:53,690][starVLA.training.train_starvla][INFO] - Step 4288, Loss: {'train/loss': 1.4848516912024934e-05, 'train/grad_norm_pre_clip': 0.003450771328061819, 'timing/data': 0.00032929998997133225, 'timing/model': 0.37016335100634024, 'learning_rate/qwen_vl_interface': 9.863806877237421e-07, 'learning_rate/action_model': 9.863806877237422e-06, 'epoch': 0.31}) +[2026-10-03 15:26:54,349][starVLA.training.train_starvla][INFO] - Step 4289, Loss: {'train/loss': 2.0145436792518012e-05, 'train/grad_norm_pre_clip': 0.0066294423304498196, 'timing/data': 0.0002785300021059811, 'timing/model': 0.3702099310030462, 'learning_rate/qwen_vl_interface': 9.850391343978454e-07, 'learning_rate/action_model': 9.850391343978454e-06, 'epoch': 0.31}) +[2026-10-03 15:26:55,008][starVLA.training.train_starvla][INFO] - Step 4290, Loss: {'train/loss': 6.981098977121292e-06, 'train/grad_norm_pre_clip': 0.002738717244938016, 'timing/data': 0.0002552899968577549, 'timing/model': 0.37044169098953716, 'learning_rate/qwen_vl_interface': 9.836993342362596e-07, 'learning_rate/action_model': 9.836993342362596e-06, 'epoch': 0.31}) +[2026-10-03 15:26:55,665][starVLA.training.train_starvla][INFO] - Step 4291, Loss: {'train/loss': 1.2442080333130434e-05, 'train/grad_norm_pre_clip': 0.004033278673887253, 'timing/data': 0.00030852999771013856, 'timing/model': 0.36962652098736726, 'learning_rate/qwen_vl_interface': 9.823612877897246e-07, 'learning_rate/action_model': 9.823612877897245e-06, 'epoch': 0.31}) +[2026-10-03 15:26:56,523][starVLA.training.train_starvla][INFO] - Step 4292, Loss: {'train/loss': 2.8272341296542436e-05, 'train/grad_norm_pre_clip': 0.033557962626218796, 'timing/data': 0.00030153999978210777, 'timing/model': 0.5681345439952565, 'learning_rate/qwen_vl_interface': 9.81024995608261e-07, 'learning_rate/action_model': 9.810249956082609e-06, 'epoch': 0.31}) +[2026-10-03 15:26:57,180][starVLA.training.train_starvla][INFO] - Step 4293, Loss: {'train/loss': 2.5209570594597608e-05, 'train/grad_norm_pre_clip': 0.01193266175687313, 'timing/data': 0.0003274899936513975, 'timing/model': 0.36976659001084045, 'learning_rate/qwen_vl_interface': 9.796904582411685e-07, 'learning_rate/action_model': 9.796904582411684e-06, 'epoch': 0.31}) +[2026-10-03 15:26:57,838][starVLA.training.train_starvla][INFO] - Step 4294, Loss: {'train/loss': 0.0001965286210179329, 'train/grad_norm_pre_clip': 0.07224135845899582, 'timing/data': 0.00021531000675167888, 'timing/model': 0.3698224910040153, 'learning_rate/qwen_vl_interface': 9.78357676237025e-07, 'learning_rate/action_model': 9.78357676237025e-06, 'epoch': 0.31}) +[2026-10-03 15:26:58,496][starVLA.training.train_starvla][INFO] - Step 4295, Loss: {'train/loss': 1.5444238670170307e-05, 'train/grad_norm_pre_clip': 0.005739905871450901, 'timing/data': 0.00026812001306097955, 'timing/model': 0.36924688098952174, 'learning_rate/qwen_vl_interface': 9.77026650143687e-07, 'learning_rate/action_model': 9.77026650143687e-06, 'epoch': 0.31}) +[2026-10-03 15:26:59,155][starVLA.training.train_starvla][INFO] - Step 4296, Loss: {'train/loss': 0.00012880699068773538, 'train/grad_norm_pre_clip': 0.03950202092528343, 'timing/data': 0.0002599599974928424, 'timing/model': 0.37039620100404136, 'learning_rate/qwen_vl_interface': 9.756973805082882e-07, 'learning_rate/action_model': 9.756973805082883e-06, 'epoch': 0.31}) +[2026-10-03 15:26:59,816][starVLA.training.train_starvla][INFO] - Step 4297, Loss: {'train/loss': 0.0009238672209903598, 'train/grad_norm_pre_clip': 0.7508137822151184, 'timing/data': 0.0002121700090356171, 'timing/model': 0.3716668099950766, 'learning_rate/qwen_vl_interface': 9.743698678772434e-07, 'learning_rate/action_model': 9.743698678772433e-06, 'epoch': 0.31}) +[2026-10-03 15:27:00,668][starVLA.training.train_starvla][INFO] - Step 4298, Loss: {'train/loss': 0.014838878065347672, 'train/grad_norm_pre_clip': 31.166179656982422, 'timing/data': 0.00028856001154053956, 'timing/model': 0.369950811000308, 'learning_rate/qwen_vl_interface': 9.730441127962395e-07, 'learning_rate/action_model': 9.730441127962395e-06, 'epoch': 0.31}) +[2026-10-03 15:27:01,325][starVLA.training.train_starvla][INFO] - Step 4299, Loss: {'train/loss': 0.0004911960568279028, 'train/grad_norm_pre_clip': 0.3860850930213928, 'timing/data': 0.00030546000925824046, 'timing/model': 0.3692888310033595, 'learning_rate/qwen_vl_interface': 9.717201158102465e-07, 'learning_rate/action_model': 9.717201158102466e-06, 'epoch': 0.31}) +[2026-10-03 15:27:01,987][starVLA.training.train_starvla][INFO] - Step 4300, Loss: {'train/loss': 4.796242865268141e-05, 'train/grad_norm_pre_clip': 0.017709694802761078, 'timing/data': 0.00031662000401411206, 'timing/model': 0.3707259009970585, 'learning_rate/qwen_vl_interface': 9.703978774635094e-07, 'learning_rate/action_model': 9.703978774635095e-06, 'epoch': 0.31}) +[2026-10-03 15:27:02,647][starVLA.training.train_starvla][INFO] - Step 4301, Loss: {'train/loss': 4.678931418311549e-06, 'train/grad_norm_pre_clip': 0.0015067511703819036, 'timing/data': 0.00032928999280557036, 'timing/model': 0.37017090000153985, 'learning_rate/qwen_vl_interface': 9.6907739829955e-07, 'learning_rate/action_model': 9.6907739829955e-06, 'epoch': 0.31}) +[2026-10-03 15:27:03,306][starVLA.training.train_starvla][INFO] - Step 4302, Loss: {'train/loss': 0.8653683662414551, 'train/grad_norm_pre_clip': 28.835020065307617, 'timing/data': 0.0002517200045986101, 'timing/model': 0.37034423000295646, 'learning_rate/qwen_vl_interface': 9.677586788611681e-07, 'learning_rate/action_model': 9.677586788611681e-06, 'epoch': 0.31}) +[2026-10-03 15:27:04,174][starVLA.training.train_starvla][INFO] - Step 4303, Loss: {'train/loss': 3.386656680959277e-05, 'train/grad_norm_pre_clip': 0.015983350574970245, 'timing/data': 0.0002945200103567913, 'timing/model': 0.3704324709979119, 'learning_rate/qwen_vl_interface': 9.664417196904382e-07, 'learning_rate/action_model': 9.664417196904382e-06, 'epoch': 0.31}) +[2026-10-03 15:27:04,834][starVLA.training.train_starvla][INFO] - Step 4304, Loss: {'train/loss': 2.9197426556493156e-05, 'train/grad_norm_pre_clip': 0.010276468470692635, 'timing/data': 0.0002742299984674901, 'timing/model': 0.3709382899978664, 'learning_rate/qwen_vl_interface': 9.651265213287134e-07, 'learning_rate/action_model': 9.651265213287134e-06, 'epoch': 0.32}) +[2026-10-03 15:27:05,494][starVLA.training.train_starvla][INFO] - Step 4305, Loss: {'train/loss': 0.2490057349205017, 'train/grad_norm_pre_clip': 177.6767120361328, 'timing/data': 0.000344770000083372, 'timing/model': 0.3709149299975252, 'learning_rate/qwen_vl_interface': 9.638130843166218e-07, 'learning_rate/action_model': 9.638130843166219e-06, 'epoch': 0.32}) +[2026-10-03 15:27:06,153][starVLA.training.train_starvla][INFO] - Step 4306, Loss: {'train/loss': 5.789056558569428e-06, 'train/grad_norm_pre_clip': 0.0015258328057825565, 'timing/data': 0.00026421000075060874, 'timing/model': 0.3695317009987775, 'learning_rate/qwen_vl_interface': 9.62501409194068e-07, 'learning_rate/action_model': 9.625014091940678e-06, 'epoch': 0.32}) +[2026-10-03 15:27:06,810][starVLA.training.train_starvla][INFO] - Step 4307, Loss: {'train/loss': 0.0002938111138064414, 'train/grad_norm_pre_clip': 0.2052411437034607, 'timing/data': 0.0002115399984177202, 'timing/model': 0.3698181609943276, 'learning_rate/qwen_vl_interface': 9.611914965002317e-07, 'learning_rate/action_model': 9.611914965002316e-06, 'epoch': 0.32}) +[2026-10-03 15:27:07,676][starVLA.training.train_starvla][INFO] - Step 4308, Loss: {'train/loss': 4.4463777157943696e-05, 'train/grad_norm_pre_clip': 0.020815333351492882, 'timing/data': 0.0003309599997010082, 'timing/model': 0.5751143829984358, 'learning_rate/qwen_vl_interface': 9.598833467735686e-07, 'learning_rate/action_model': 9.598833467735685e-06, 'epoch': 0.32}) +[2026-10-03 15:27:08,335][starVLA.training.train_starvla][INFO] - Step 4309, Loss: {'train/loss': 0.23519274592399597, 'train/grad_norm_pre_clip': 291.8124084472656, 'timing/data': 0.0002711200068006292, 'timing/model': 0.37089968999498524, 'learning_rate/qwen_vl_interface': 9.585769605518097e-07, 'learning_rate/action_model': 9.585769605518097e-06, 'epoch': 0.32}) +[2026-10-03 15:27:08,994][starVLA.training.train_starvla][INFO] - Step 4310, Loss: {'train/loss': 1.7605094399186783e-05, 'train/grad_norm_pre_clip': 0.005132781807333231, 'timing/data': 0.0003022799937753007, 'timing/model': 0.3702865800005384, 'learning_rate/qwen_vl_interface': 9.572723383719611e-07, 'learning_rate/action_model': 9.572723383719612e-06, 'epoch': 0.32}) +[2026-10-03 15:27:09,652][starVLA.training.train_starvla][INFO] - Step 4311, Loss: {'train/loss': 0.4327264726161957, 'train/grad_norm_pre_clip': 69.99559020996094, 'timing/data': 0.00025522999931126833, 'timing/model': 0.369840571001987, 'learning_rate/qwen_vl_interface': 9.559694807703041e-07, 'learning_rate/action_model': 9.55969480770304e-06, 'epoch': 0.32}) +[2026-10-03 15:27:10,289][starVLA.training.train_starvla][INFO] - Step 4312, Loss: {'train/loss': 0.12659002840518951, 'train/grad_norm_pre_clip': 138.1648712158203, 'timing/data': 0.0002951399947050959, 'timing/model': 0.3475534739991417, 'learning_rate/qwen_vl_interface': 9.546683882823945e-07, 'learning_rate/action_model': 9.546683882823945e-06, 'epoch': 0.32}) +[2026-10-03 15:27:10,948][starVLA.training.train_starvla][INFO] - Step 4313, Loss: {'train/loss': 5.738062827731483e-05, 'train/grad_norm_pre_clip': 0.01984211429953575, 'timing/data': 0.0002275300066685304, 'timing/model': 0.370093451012508, 'learning_rate/qwen_vl_interface': 9.533690614430608e-07, 'learning_rate/action_model': 9.533690614430608e-06, 'epoch': 0.32}) +[2026-10-03 15:27:11,803][starVLA.training.train_starvla][INFO] - Step 4314, Loss: {'train/loss': 0.0002202905307058245, 'train/grad_norm_pre_clip': 0.03988068178296089, 'timing/data': 0.00039710999408271164, 'timing/model': 0.3695771600032458, 'learning_rate/qwen_vl_interface': 9.520715007864082e-07, 'learning_rate/action_model': 9.52071500786408e-06, 'epoch': 0.32}) +[2026-10-03 15:27:12,460][starVLA.training.train_starvla][INFO] - Step 4315, Loss: {'train/loss': 1.3187267541070469e-05, 'train/grad_norm_pre_clip': 0.0042401510290801525, 'timing/data': 0.00032348000968340784, 'timing/model': 0.3693677999981446, 'learning_rate/qwen_vl_interface': 9.50775706845814e-07, 'learning_rate/action_model': 9.50775706845814e-06, 'epoch': 0.32}) +[2026-10-03 15:27:13,120][starVLA.training.train_starvla][INFO] - Step 4316, Loss: {'train/loss': 0.0002536729152780026, 'train/grad_norm_pre_clip': 0.31181302666664124, 'timing/data': 0.0003100399917457253, 'timing/model': 0.370032060003723, 'learning_rate/qwen_vl_interface': 9.49481680153931e-07, 'learning_rate/action_model': 9.494816801539309e-06, 'epoch': 0.32}) +[2026-10-03 15:27:13,778][starVLA.training.train_starvla][INFO] - Step 4317, Loss: {'train/loss': 0.0028292930219322443, 'train/grad_norm_pre_clip': 12.883684158325195, 'timing/data': 0.0002542300062486902, 'timing/model': 0.3700590010121232, 'learning_rate/qwen_vl_interface': 9.481894212426848e-07, 'learning_rate/action_model': 9.481894212426847e-06, 'epoch': 0.32}) +[2026-10-03 15:27:14,436][starVLA.training.train_starvla][INFO] - Step 4318, Loss: {'train/loss': 2.9054619517410174e-05, 'train/grad_norm_pre_clip': 0.014476819895207882, 'timing/data': 0.00025132000155281276, 'timing/model': 0.3697290510026505, 'learning_rate/qwen_vl_interface': 9.468989306432738e-07, 'learning_rate/action_model': 9.468989306432738e-06, 'epoch': 0.32}) +[2026-10-03 15:27:15,293][starVLA.training.train_starvla][INFO] - Step 4319, Loss: {'train/loss': 4.9537680752109736e-05, 'train/grad_norm_pre_clip': 0.022871237248182297, 'timing/data': 0.0003327999875182286, 'timing/model': 0.5695229440025287, 'learning_rate/qwen_vl_interface': 9.456102088861702e-07, 'learning_rate/action_model': 9.456102088861701e-06, 'epoch': 0.32}) +[2026-10-03 15:27:15,955][starVLA.training.train_starvla][INFO] - Step 4320, Loss: {'train/loss': 1.087020427803509e-05, 'train/grad_norm_pre_clip': 0.002709544962272048, 'timing/data': 0.0003503889893181622, 'timing/model': 0.37163230100122746, 'learning_rate/qwen_vl_interface': 9.443232565011188e-07, 'learning_rate/action_model': 9.443232565011187e-06, 'epoch': 0.32}) +[2026-10-03 15:27:16,616][starVLA.training.train_starvla][INFO] - Step 4321, Loss: {'train/loss': 0.3332946300506592, 'train/grad_norm_pre_clip': 51.4173583984375, 'timing/data': 0.00032963999547064304, 'timing/model': 0.3715994910016889, 'learning_rate/qwen_vl_interface': 9.430380740171369e-07, 'learning_rate/action_model': 9.430380740171368e-06, 'epoch': 0.32}) +[2026-10-03 15:27:17,275][starVLA.training.train_starvla][INFO] - Step 4322, Loss: {'train/loss': 4.000931767222937e-06, 'train/grad_norm_pre_clip': 0.0008589840144850314, 'timing/data': 0.00035103000118397176, 'timing/model': 0.36975544098822866, 'learning_rate/qwen_vl_interface': 9.417546619625143e-07, 'learning_rate/action_model': 9.417546619625143e-06, 'epoch': 0.32}) +[2026-10-03 15:27:17,934][starVLA.training.train_starvla][INFO] - Step 4323, Loss: {'train/loss': 1.652466926316265e-05, 'train/grad_norm_pre_clip': 0.005539341829717159, 'timing/data': 0.0002029799943557009, 'timing/model': 0.3709594910033047, 'learning_rate/qwen_vl_interface': 9.404730208648144e-07, 'learning_rate/action_model': 9.404730208648144e-06, 'epoch': 0.32}) +[2026-10-03 15:27:18,788][starVLA.training.train_starvla][INFO] - Step 4324, Loss: {'train/loss': 2.129195308953058e-05, 'train/grad_norm_pre_clip': 0.019482571631669998, 'timing/data': 0.0002695099974516779, 'timing/model': 0.5643548049993115, 'learning_rate/qwen_vl_interface': 9.3919315125087e-07, 'learning_rate/action_model': 9.3919315125087e-06, 'epoch': 0.32}) +[2026-10-03 15:27:19,449][starVLA.training.train_starvla][INFO] - Step 4325, Loss: {'train/loss': 0.00012570107355713844, 'train/grad_norm_pre_clip': 0.03492187336087227, 'timing/data': 0.00024311000015586615, 'timing/model': 0.3701875109982211, 'learning_rate/qwen_vl_interface': 9.379150536467879e-07, 'learning_rate/action_model': 9.379150536467878e-06, 'epoch': 0.32}) +[2026-10-03 15:27:20,107][starVLA.training.train_starvla][INFO] - Step 4326, Loss: {'train/loss': 7.159951110224938e-06, 'train/grad_norm_pre_clip': 0.0014036702923476696, 'timing/data': 0.00026590999914333224, 'timing/model': 0.36914289100968745, 'learning_rate/qwen_vl_interface': 9.366387285779453e-07, 'learning_rate/action_model': 9.366387285779453e-06, 'epoch': 0.32}) +[2026-10-03 15:27:20,765][starVLA.training.train_starvla][INFO] - Step 4327, Loss: {'train/loss': 0.9899958968162537, 'train/grad_norm_pre_clip': 52.63334655761719, 'timing/data': 0.0003184600063832477, 'timing/model': 0.3701047509966884, 'learning_rate/qwen_vl_interface': 9.353641765689915e-07, 'learning_rate/action_model': 9.353641765689915e-06, 'epoch': 0.32}) +[2026-10-03 15:27:21,427][starVLA.training.train_starvla][INFO] - Step 4328, Loss: {'train/loss': 5.461240107251797e-06, 'train/grad_norm_pre_clip': 0.0013141080271452665, 'timing/data': 0.00033926000469364226, 'timing/model': 0.37169488100335, 'learning_rate/qwen_vl_interface': 9.340913981438462e-07, 'learning_rate/action_model': 9.340913981438463e-06, 'epoch': 0.32}) +[2026-10-03 15:27:22,088][starVLA.training.train_starvla][INFO] - Step 4329, Loss: {'train/loss': 6.295590992522193e-06, 'train/grad_norm_pre_clip': 0.005230538081377745, 'timing/data': 0.00037106999661773443, 'timing/model': 0.3709997400001157, 'learning_rate/qwen_vl_interface': 9.328203938257019e-07, 'learning_rate/action_model': 9.328203938257018e-06, 'epoch': 0.32}) +[2026-10-03 15:27:22,950][starVLA.training.train_starvla][INFO] - Step 4330, Loss: {'train/loss': 0.3940737247467041, 'train/grad_norm_pre_clip': 56.11157989501953, 'timing/data': 0.00034178000350948423, 'timing/model': 0.3704487500071991, 'learning_rate/qwen_vl_interface': 9.315511641370191e-07, 'learning_rate/action_model': 9.315511641370191e-06, 'epoch': 0.32}) +[2026-10-03 15:27:23,609][starVLA.training.train_starvla][INFO] - Step 4331, Loss: {'train/loss': 3.6851059121545404e-05, 'train/grad_norm_pre_clip': 0.015074940398335457, 'timing/data': 0.0002994800015585497, 'timing/model': 0.3699453310109675, 'learning_rate/qwen_vl_interface': 9.302837095995306e-07, 'learning_rate/action_model': 9.302837095995306e-06, 'epoch': 0.32}) +[2026-10-03 15:27:24,271][starVLA.training.train_starvla][INFO] - Step 4332, Loss: {'train/loss': 0.446412593126297, 'train/grad_norm_pre_clip': 23.470508575439453, 'timing/data': 0.0004210500046610832, 'timing/model': 0.3703282110072905, 'learning_rate/qwen_vl_interface': 9.290180307342392e-07, 'learning_rate/action_model': 9.290180307342391e-06, 'epoch': 0.32}) +[2026-10-03 15:27:24,932][starVLA.training.train_starvla][INFO] - Step 4333, Loss: {'train/loss': 6.653305263171205e-06, 'train/grad_norm_pre_clip': 0.001290192361921072, 'timing/data': 0.00021593000565189868, 'timing/model': 0.3713093400001526, 'learning_rate/qwen_vl_interface': 9.277541280614183e-07, 'learning_rate/action_model': 9.277541280614183e-06, 'epoch': 0.32}) +[2026-10-03 15:27:25,591][starVLA.training.train_starvla][INFO] - Step 4334, Loss: {'train/loss': 0.0002522958384361118, 'train/grad_norm_pre_clip': 0.1317068487405777, 'timing/data': 0.00021000999549869448, 'timing/model': 0.37038011000549886, 'learning_rate/qwen_vl_interface': 9.264920021006093e-07, 'learning_rate/action_model': 9.264920021006093e-06, 'epoch': 0.32}) +[2026-10-03 15:27:26,452][starVLA.training.train_starvla][INFO] - Step 4335, Loss: {'train/loss': 0.3805612325668335, 'train/grad_norm_pre_clip': 56.75014877319336, 'timing/data': 0.00030735001200810075, 'timing/model': 0.3702995909989113, 'learning_rate/qwen_vl_interface': 9.252316533706246e-07, 'learning_rate/action_model': 9.252316533706246e-06, 'epoch': 0.32}) +[2026-10-03 15:27:27,113][starVLA.training.train_starvla][INFO] - Step 4336, Loss: {'train/loss': 1.126514507632237e-05, 'train/grad_norm_pre_clip': 0.0023364881053566933, 'timing/data': 0.000259129999903962, 'timing/model': 0.3711896299937507, 'learning_rate/qwen_vl_interface': 9.239730823895469e-07, 'learning_rate/action_model': 9.23973082389547e-06, 'epoch': 0.32}) +[2026-10-03 15:27:27,780][starVLA.training.train_starvla][INFO] - Step 4337, Loss: {'train/loss': 0.0005047601880505681, 'train/grad_norm_pre_clip': 1.3928875923156738, 'timing/data': 0.0002850699966074899, 'timing/model': 0.3741583899973193, 'learning_rate/qwen_vl_interface': 9.227162896747264e-07, 'learning_rate/action_model': 9.227162896747264e-06, 'epoch': 0.32}) +[2026-10-03 15:27:28,443][starVLA.training.train_starvla][INFO] - Step 4338, Loss: {'train/loss': 0.4037315845489502, 'train/grad_norm_pre_clip': 58.49974822998047, 'timing/data': 0.00036576000275090337, 'timing/model': 0.37270459000137635, 'learning_rate/qwen_vl_interface': 9.214612757427833e-07, 'learning_rate/action_model': 9.214612757427833e-06, 'epoch': 0.32}) +[2026-10-03 15:27:29,109][starVLA.training.train_starvla][INFO] - Step 4339, Loss: {'train/loss': 0.22843556106090546, 'train/grad_norm_pre_clip': 168.4198760986328, 'timing/data': 0.00033819999953266233, 'timing/model': 0.3729882300103782, 'learning_rate/qwen_vl_interface': 9.202080411096071e-07, 'learning_rate/action_model': 9.20208041109607e-06, 'epoch': 0.32}) +[2026-10-03 15:27:29,983][starVLA.training.train_starvla][INFO] - Step 4340, Loss: {'train/loss': 0.31202664971351624, 'train/grad_norm_pre_clip': 38.95158767700195, 'timing/data': 0.00022560000070370734, 'timing/model': 0.5825230320042465, 'learning_rate/qwen_vl_interface': 9.189565862903543e-07, 'learning_rate/action_model': 9.189565862903543e-06, 'epoch': 0.32}) +[2026-10-03 15:27:30,647][starVLA.training.train_starvla][INFO] - Step 4341, Loss: {'train/loss': 7.532443305535708e-06, 'train/grad_norm_pre_clip': 0.002195602748543024, 'timing/data': 0.00036217999877408147, 'timing/model': 0.3744184000097448, 'learning_rate/qwen_vl_interface': 9.17706911799451e-07, 'learning_rate/action_model': 9.17706911799451e-06, 'epoch': 0.32}) +[2026-10-03 15:27:31,313][starVLA.training.train_starvla][INFO] - Step 4342, Loss: {'train/loss': 0.007824108004570007, 'train/grad_norm_pre_clip': 33.209835052490234, 'timing/data': 0.0002651000104378909, 'timing/model': 0.37310989999969024, 'learning_rate/qwen_vl_interface': 9.164590181505912e-07, 'learning_rate/action_model': 9.164590181505911e-06, 'epoch': 0.32}) +[2026-10-03 15:27:31,971][starVLA.training.train_starvla][INFO] - Step 4343, Loss: {'train/loss': 0.0019634997006505728, 'train/grad_norm_pre_clip': 4.443159103393555, 'timing/data': 0.0002595699916128069, 'timing/model': 0.369915810006205, 'learning_rate/qwen_vl_interface': 9.15212905856738e-07, 'learning_rate/action_model': 9.15212905856738e-06, 'epoch': 0.32}) +[2026-10-03 15:27:32,608][starVLA.training.train_starvla][INFO] - Step 4344, Loss: {'train/loss': 7.657753303647041e-05, 'train/grad_norm_pre_clip': 0.029902562499046326, 'timing/data': 0.00038610000046901405, 'timing/model': 0.37087728999904357, 'learning_rate/qwen_vl_interface': 9.139685754301189e-07, 'learning_rate/action_model': 9.139685754301189e-06, 'epoch': 0.32}) +[2026-10-03 15:27:33,459][starVLA.training.train_starvla][INFO] - Step 4345, Loss: {'train/loss': 3.6368837754707783e-05, 'train/grad_norm_pre_clip': 0.02059686928987503, 'timing/data': 0.00020726000366266817, 'timing/model': 0.5621809849981219, 'learning_rate/qwen_vl_interface': 9.127260273822324e-07, 'learning_rate/action_model': 9.127260273822325e-06, 'epoch': 0.32}) +[2026-10-03 15:27:34,096][starVLA.training.train_starvla][INFO] - Step 4346, Loss: {'train/loss': 4.2901840060949326e-05, 'train/grad_norm_pre_clip': 0.019928641617298126, 'timing/data': 0.0003513699921313673, 'timing/model': 0.34731887400266714, 'learning_rate/qwen_vl_interface': 9.114852622238435e-07, 'learning_rate/action_model': 9.114852622238435e-06, 'epoch': 0.32}) +[2026-10-03 15:27:34,754][starVLA.training.train_starvla][INFO] - Step 4347, Loss: {'train/loss': 2.3222106392495334e-05, 'train/grad_norm_pre_clip': 0.011567597277462482, 'timing/data': 0.00021029000345151871, 'timing/model': 0.3703270300029544, 'learning_rate/qwen_vl_interface': 9.102462804649827e-07, 'learning_rate/action_model': 9.102462804649827e-06, 'epoch': 0.32}) +[2026-10-03 15:27:35,391][starVLA.training.train_starvla][INFO] - Step 4348, Loss: {'train/loss': 0.37667348980903625, 'train/grad_norm_pre_clip': 21.810535430908203, 'timing/data': 0.0003713490004884079, 'timing/model': 0.34746073400310706, 'learning_rate/qwen_vl_interface': 9.090090826149489e-07, 'learning_rate/action_model': 9.09009082614949e-06, 'epoch': 0.32}) +[2026-10-03 15:27:36,052][starVLA.training.train_starvla][INFO] - Step 4349, Loss: {'train/loss': 0.0007329117506742477, 'train/grad_norm_pre_clip': 0.41834890842437744, 'timing/data': 0.000277810002444312, 'timing/model': 0.371314350006287, 'learning_rate/qwen_vl_interface': 9.077736691823081e-07, 'learning_rate/action_model': 9.07773669182308e-06, 'epoch': 0.32}) +[2026-10-03 15:27:36,711][starVLA.training.train_starvla][INFO] - Step 4350, Loss: {'train/loss': 2.12557970371563e-05, 'train/grad_norm_pre_clip': 0.0075782183557748795, 'timing/data': 0.00021081999875605106, 'timing/model': 0.3710925799969118, 'learning_rate/qwen_vl_interface': 9.065400406748912e-07, 'learning_rate/action_model': 9.06540040674891e-06, 'epoch': 0.32}) +[2026-10-03 15:27:37,567][starVLA.training.train_starvla][INFO] - Step 4351, Loss: {'train/loss': 0.25290584564208984, 'train/grad_norm_pre_clip': 225.82862854003906, 'timing/data': 0.00033085999893955886, 'timing/model': 0.3695295100042131, 'learning_rate/qwen_vl_interface': 9.053081975997969e-07, 'learning_rate/action_model': 9.053081975997968e-06, 'epoch': 0.32}) +[2026-10-03 15:27:38,204][starVLA.training.train_starvla][INFO] - Step 4352, Loss: {'train/loss': 1.7969605323742144e-05, 'train/grad_norm_pre_clip': 0.00617254339158535, 'timing/data': 0.0003247400018153712, 'timing/model': 0.3703906599985203, 'learning_rate/qwen_vl_interface': 9.040781404633894e-07, 'learning_rate/action_model': 9.040781404633895e-06, 'epoch': 0.32}) +[2026-10-03 15:27:38,864][starVLA.training.train_starvla][INFO] - Step 4353, Loss: {'train/loss': 8.266177610494196e-05, 'train/grad_norm_pre_clip': 0.04662131518125534, 'timing/data': 0.0002644190099090338, 'timing/model': 0.37103010099963285, 'learning_rate/qwen_vl_interface': 9.028498697712982e-07, 'learning_rate/action_model': 9.028498697712981e-06, 'epoch': 0.32}) +[2026-10-03 15:27:39,524][starVLA.training.train_starvla][INFO] - Step 4354, Loss: {'train/loss': 0.0017332707066088915, 'train/grad_norm_pre_clip': 5.626520156860352, 'timing/data': 0.0003329399914946407, 'timing/model': 0.3701244999974733, 'learning_rate/qwen_vl_interface': 9.016233860284195e-07, 'learning_rate/action_model': 9.016233860284195e-06, 'epoch': 0.32}) +[2026-10-03 15:27:40,183][starVLA.training.train_starvla][INFO] - Step 4355, Loss: {'train/loss': 1.6099991626106203e-05, 'train/grad_norm_pre_clip': 0.0038882277440279722, 'timing/data': 0.0003393099905224517, 'timing/model': 0.37040570100361947, 'learning_rate/qwen_vl_interface': 9.003986897389142e-07, 'learning_rate/action_model': 9.003986897389141e-06, 'epoch': 0.32}) +[2026-10-03 15:27:41,042][starVLA.training.train_starvla][INFO] - Step 4356, Loss: {'train/loss': 0.4958157539367676, 'train/grad_norm_pre_clip': 38.80998611450195, 'timing/data': 0.00023037999926600605, 'timing/model': 0.5694262140023056, 'learning_rate/qwen_vl_interface': 8.991757814062094e-07, 'learning_rate/action_model': 8.991757814062093e-06, 'epoch': 0.32}) +[2026-10-03 15:27:41,703][starVLA.training.train_starvla][INFO] - Step 4357, Loss: {'train/loss': 4.2785068217199296e-05, 'train/grad_norm_pre_clip': 0.015281966887414455, 'timing/data': 0.0003760899999178946, 'timing/model': 0.3714787399949273, 'learning_rate/qwen_vl_interface': 8.979546615329954e-07, 'learning_rate/action_model': 8.979546615329953e-06, 'epoch': 0.32}) +[2026-10-03 15:27:42,361][starVLA.training.train_starvla][INFO] - Step 4358, Loss: {'train/loss': 0.001426728325895965, 'train/grad_norm_pre_clip': 1.9560973644256592, 'timing/data': 0.00027631000557448715, 'timing/model': 0.36977627100714017, 'learning_rate/qwen_vl_interface': 8.967353306212286e-07, 'learning_rate/action_model': 8.967353306212286e-06, 'epoch': 0.32}) +[2026-10-03 15:27:43,020][starVLA.training.train_starvla][INFO] - Step 4359, Loss: {'train/loss': 3.311891487101093e-05, 'train/grad_norm_pre_clip': 0.023376844823360443, 'timing/data': 0.00032995999208651483, 'timing/model': 0.37008624999725726, 'learning_rate/qwen_vl_interface': 8.955177891721303e-07, 'learning_rate/action_model': 8.955177891721303e-06, 'epoch': 0.32}) +[2026-10-03 15:27:43,682][starVLA.training.train_starvla][INFO] - Step 4360, Loss: {'train/loss': 0.45990920066833496, 'train/grad_norm_pre_clip': 38.7797737121582, 'timing/data': 0.0003662899980554357, 'timing/model': 0.37132601099438034, 'learning_rate/qwen_vl_interface': 8.943020376861858e-07, 'learning_rate/action_model': 8.943020376861857e-06, 'epoch': 0.32}) +[2026-10-03 15:27:44,320][starVLA.training.train_starvla][INFO] - Step 4361, Loss: {'train/loss': 9.566421795170754e-06, 'train/grad_norm_pre_clip': 0.0019126124680042267, 'timing/data': 0.0003500400052871555, 'timing/model': 0.37149735099228565, 'learning_rate/qwen_vl_interface': 8.930880766631444e-07, 'learning_rate/action_model': 8.930880766631445e-06, 'epoch': 0.32}) +[2026-10-03 15:27:45,188][starVLA.training.train_starvla][INFO] - Step 4362, Loss: {'train/loss': 8.191910455934703e-05, 'train/grad_norm_pre_clip': 0.05509190261363983, 'timing/data': 0.00022401999740395695, 'timing/model': 0.37052446098823566, 'learning_rate/qwen_vl_interface': 8.918759066020201e-07, 'learning_rate/action_model': 8.9187590660202e-06, 'epoch': 0.32}) +[2026-10-03 15:27:45,851][starVLA.training.train_starvla][INFO] - Step 4363, Loss: {'train/loss': 5.3382264013634995e-05, 'train/grad_norm_pre_clip': 0.015511085279285908, 'timing/data': 0.0004176599904894829, 'timing/model': 0.373066239990294, 'learning_rate/qwen_vl_interface': 8.906655280010895e-07, 'learning_rate/action_model': 8.906655280010896e-06, 'epoch': 0.32}) +[2026-10-03 15:27:46,520][starVLA.training.train_starvla][INFO] - Step 4364, Loss: {'train/loss': 2.3609612981090322e-05, 'train/grad_norm_pre_clip': 0.007750784978270531, 'timing/data': 0.00027551999664865434, 'timing/model': 0.37328193000575993, 'learning_rate/qwen_vl_interface': 8.894569413578938e-07, 'learning_rate/action_model': 8.894569413578937e-06, 'epoch': 0.32}) +[2026-10-03 15:27:47,187][starVLA.training.train_starvla][INFO] - Step 4365, Loss: {'train/loss': 1.5451338185812347e-05, 'train/grad_norm_pre_clip': 0.006988024804741144, 'timing/data': 0.00038285000482574105, 'timing/model': 0.373805141003686, 'learning_rate/qwen_vl_interface': 8.882501471692379e-07, 'learning_rate/action_model': 8.882501471692378e-06, 'epoch': 0.32}) +[2026-10-03 15:27:47,826][starVLA.training.train_starvla][INFO] - Step 4366, Loss: {'train/loss': 0.0005810058210045099, 'train/grad_norm_pre_clip': 0.46980157494544983, 'timing/data': 0.00039722998917568475, 'timing/model': 0.3704555110016372, 'learning_rate/qwen_vl_interface': 8.870451459311894e-07, 'learning_rate/action_model': 8.870451459311893e-06, 'epoch': 0.32}) +[2026-10-03 15:27:48,663][starVLA.training.train_starvla][INFO] - Step 4367, Loss: {'train/loss': 0.32970473170280457, 'train/grad_norm_pre_clip': 17.51236915588379, 'timing/data': 0.00031756999669596553, 'timing/model': 0.5479321370075922, 'learning_rate/qwen_vl_interface': 8.85841938139077e-07, 'learning_rate/action_model': 8.85841938139077e-06, 'epoch': 0.32}) +[2026-10-03 15:27:49,327][starVLA.training.train_starvla][INFO] - Step 4368, Loss: {'train/loss': 5.401645012170775e-06, 'train/grad_norm_pre_clip': 0.001281394506804645, 'timing/data': 0.0002859999949578196, 'timing/model': 0.3736989699973492, 'learning_rate/qwen_vl_interface': 8.846405242874955e-07, 'learning_rate/action_model': 8.846405242874954e-06, 'epoch': 0.32}) +[2026-10-03 15:27:49,967][starVLA.training.train_starvla][INFO] - Step 4369, Loss: {'train/loss': 0.353074312210083, 'train/grad_norm_pre_clip': 25.06879425048828, 'timing/data': 0.00028545000532176346, 'timing/model': 0.347889992990531, 'learning_rate/qwen_vl_interface': 8.834409048703004e-07, 'learning_rate/action_model': 8.834409048703005e-06, 'epoch': 0.32}) +[2026-10-03 15:27:50,625][starVLA.training.train_starvla][INFO] - Step 4370, Loss: {'train/loss': 7.717196422163397e-05, 'train/grad_norm_pre_clip': 0.1112859696149826, 'timing/data': 0.0002917700039688498, 'timing/model': 0.3696541709941812, 'learning_rate/qwen_vl_interface': 8.822430803806098e-07, 'learning_rate/action_model': 8.822430803806097e-06, 'epoch': 0.32}) +[2026-10-03 15:27:51,284][starVLA.training.train_starvla][INFO] - Step 4371, Loss: {'train/loss': 0.0003358558169566095, 'train/grad_norm_pre_clip': 0.1163397878408432, 'timing/data': 0.00028029999521095306, 'timing/model': 0.37043957100831904, 'learning_rate/qwen_vl_interface': 8.810470513108041e-07, 'learning_rate/action_model': 8.81047051310804e-06, 'epoch': 0.32}) +[2026-10-03 15:27:52,136][starVLA.training.train_starvla][INFO] - Step 4372, Loss: {'train/loss': 2.270784170832485e-05, 'train/grad_norm_pre_clip': 0.018899276852607727, 'timing/data': 0.0003143299982184544, 'timing/model': 0.5627320750063518, 'learning_rate/qwen_vl_interface': 8.798528181525262e-07, 'learning_rate/action_model': 8.798528181525261e-06, 'epoch': 0.32}) +[2026-10-03 15:27:52,795][starVLA.training.train_starvla][INFO] - Step 4373, Loss: {'train/loss': 0.22214195132255554, 'train/grad_norm_pre_clip': 161.68215942382812, 'timing/data': 0.0002081300044665113, 'timing/model': 0.3704295700008515, 'learning_rate/qwen_vl_interface': 8.786603813966798e-07, 'learning_rate/action_model': 8.786603813966798e-06, 'epoch': 0.32}) +[2026-10-03 15:27:53,454][starVLA.training.train_starvla][INFO] - Step 4374, Loss: {'train/loss': 1.7172635125461966e-05, 'train/grad_norm_pre_clip': 0.00592883350327611, 'timing/data': 0.00020855000184383243, 'timing/model': 0.36986212100600824, 'learning_rate/qwen_vl_interface': 8.774697415334304e-07, 'learning_rate/action_model': 8.774697415334304e-06, 'epoch': 0.32}) +[2026-10-03 15:27:54,111][starVLA.training.train_starvla][INFO] - Step 4375, Loss: {'train/loss': 2.9621660360135138e-05, 'train/grad_norm_pre_clip': 0.009412677958607674, 'timing/data': 0.000354110001353547, 'timing/model': 0.36967678100336343, 'learning_rate/qwen_vl_interface': 8.762808990522056e-07, 'learning_rate/action_model': 8.762808990522057e-06, 'epoch': 0.32}) +[2026-10-03 15:27:54,749][starVLA.training.train_starvla][INFO] - Step 4376, Loss: {'train/loss': 7.659151742700487e-05, 'train/grad_norm_pre_clip': 0.042923636734485626, 'timing/data': 0.00032715000270400196, 'timing/model': 0.34755440399749205, 'learning_rate/qwen_vl_interface': 8.750938544416934e-07, 'learning_rate/action_model': 8.750938544416934e-06, 'epoch': 0.32}) +[2026-10-03 15:27:55,609][starVLA.training.train_starvla][INFO] - Step 4377, Loss: {'train/loss': 0.2500537037849426, 'train/grad_norm_pre_clip': 151.72317504882812, 'timing/data': 0.00030435000371653587, 'timing/model': 0.5715128440060653, 'learning_rate/qwen_vl_interface': 8.739086081898437e-07, 'learning_rate/action_model': 8.739086081898437e-06, 'epoch': 0.32}) +[2026-10-03 15:27:56,245][starVLA.training.train_starvla][INFO] - Step 4378, Loss: {'train/loss': 1.2754982890328392e-05, 'train/grad_norm_pre_clip': 0.00328082381747663, 'timing/data': 0.0002619599981699139, 'timing/model': 0.3696485109976493, 'learning_rate/qwen_vl_interface': 8.727251607838647e-07, 'learning_rate/action_model': 8.727251607838647e-06, 'epoch': 0.32}) +[2026-10-03 15:27:56,904][starVLA.training.train_starvla][INFO] - Step 4379, Loss: {'train/loss': 1.3701081115868874e-05, 'train/grad_norm_pre_clip': 0.004421865567564964, 'timing/data': 0.00030237999453675, 'timing/model': 0.3697094009985449, 'learning_rate/qwen_vl_interface': 8.715435127102281e-07, 'learning_rate/action_model': 8.715435127102281e-06, 'epoch': 0.32}) +[2026-10-03 15:27:57,565][starVLA.training.train_starvla][INFO] - Step 4380, Loss: {'train/loss': 0.0001090947407647036, 'train/grad_norm_pre_clip': 0.055826298892498016, 'timing/data': 0.0003162500070175156, 'timing/model': 0.37132199000916444, 'learning_rate/qwen_vl_interface': 8.703636644546651e-07, 'learning_rate/action_model': 8.703636644546651e-06, 'epoch': 0.32}) +[2026-10-03 15:27:58,202][starVLA.training.train_starvla][INFO] - Step 4381, Loss: {'train/loss': 0.00040582427754998207, 'train/grad_norm_pre_clip': 0.11137599498033524, 'timing/data': 0.00021149999520275742, 'timing/model': 0.34771431400440633, 'learning_rate/qwen_vl_interface': 8.69185616502166e-07, 'learning_rate/action_model': 8.69185616502166e-06, 'epoch': 0.32}) +[2026-10-03 15:27:58,863][starVLA.training.train_starvla][INFO] - Step 4382, Loss: {'train/loss': 1.4758767974853981e-05, 'train/grad_norm_pre_clip': 0.0079267593100667, 'timing/data': 0.0002112299989676103, 'timing/model': 0.37086181101039983, 'learning_rate/qwen_vl_interface': 8.680093693369821e-07, 'learning_rate/action_model': 8.68009369336982e-06, 'epoch': 0.32}) +[2026-10-03 15:27:59,713][starVLA.training.train_starvla][INFO] - Step 4383, Loss: {'train/loss': 0.0016869710525497794, 'train/grad_norm_pre_clip': 3.093181610107422, 'timing/data': 0.00029013000312261283, 'timing/model': 0.3700418110092869, 'learning_rate/qwen_vl_interface': 8.66834923442624e-07, 'learning_rate/action_model': 8.66834923442624e-06, 'epoch': 0.32}) +[2026-10-03 15:28:00,372][starVLA.training.train_starvla][INFO] - Step 4384, Loss: {'train/loss': 6.005137038300745e-06, 'train/grad_norm_pre_clip': 0.0010694446973502636, 'timing/data': 0.0003590300038922578, 'timing/model': 0.3693344409985002, 'learning_rate/qwen_vl_interface': 8.65662279301863e-07, 'learning_rate/action_model': 8.656622793018628e-06, 'epoch': 0.32}) +[2026-10-03 15:28:01,029][starVLA.training.train_starvla][INFO] - Step 4385, Loss: {'train/loss': 0.027597511187195778, 'train/grad_norm_pre_clip': 116.5729751586914, 'timing/data': 0.00035972999467048794, 'timing/model': 0.3693948810105212, 'learning_rate/qwen_vl_interface': 8.644914373967274e-07, 'learning_rate/action_model': 8.644914373967273e-06, 'epoch': 0.32}) +[2026-10-03 15:28:01,684][starVLA.training.train_starvla][INFO] - Step 4386, Loss: {'train/loss': 1.3894957191951107e-05, 'train/grad_norm_pre_clip': 0.0042455168440938, 'timing/data': 0.0002622699976200238, 'timing/model': 0.3687568010063842, 'learning_rate/qwen_vl_interface': 8.633223982085068e-07, 'learning_rate/action_model': 8.633223982085068e-06, 'epoch': 0.32}) +[2026-10-03 15:28:02,342][starVLA.training.train_starvla][INFO] - Step 4387, Loss: {'train/loss': 2.6202615117654204e-05, 'train/grad_norm_pre_clip': 0.008636233396828175, 'timing/data': 0.00030283001251518726, 'timing/model': 0.36957758099015336, 'learning_rate/qwen_vl_interface': 8.621551622177499e-07, 'learning_rate/action_model': 8.621551622177499e-06, 'epoch': 0.32}) +[2026-10-03 15:28:03,179][starVLA.training.train_starvla][INFO] - Step 4388, Loss: {'train/loss': 6.668215519312071e-06, 'train/grad_norm_pre_clip': 0.0011095729423686862, 'timing/data': 0.0002955899981316179, 'timing/model': 0.34685793399694376, 'learning_rate/qwen_vl_interface': 8.609897299042615e-07, 'learning_rate/action_model': 8.609897299042614e-06, 'epoch': 0.32}) +[2026-10-03 15:28:03,837][starVLA.training.train_starvla][INFO] - Step 4389, Loss: {'train/loss': 1.4476177966571413e-05, 'train/grad_norm_pre_clip': 0.0033258465118706226, 'timing/data': 0.00020831999427173287, 'timing/model': 0.36994065099861473, 'learning_rate/qwen_vl_interface': 8.598261017471088e-07, 'learning_rate/action_model': 8.598261017471088e-06, 'epoch': 0.32}) +[2026-10-03 15:28:04,494][starVLA.training.train_starvla][INFO] - Step 4390, Loss: {'train/loss': 0.000975935545284301, 'train/grad_norm_pre_clip': 1.8472133874893188, 'timing/data': 0.00020701999892480671, 'timing/model': 0.36950381100177765, 'learning_rate/qwen_vl_interface': 8.586642782246133e-07, 'learning_rate/action_model': 8.586642782246131e-06, 'epoch': 0.32}) +[2026-10-03 15:28:05,151][starVLA.training.train_starvla][INFO] - Step 4391, Loss: {'train/loss': 2.288677205797285e-05, 'train/grad_norm_pre_clip': 0.006557552609592676, 'timing/data': 0.00020488999143708497, 'timing/model': 0.3695692800101824, 'learning_rate/qwen_vl_interface': 8.57504259814358e-07, 'learning_rate/action_model': 8.57504259814358e-06, 'epoch': 0.32}) +[2026-10-03 15:28:05,809][starVLA.training.train_starvla][INFO] - Step 4392, Loss: {'train/loss': 0.31054040789604187, 'train/grad_norm_pre_clip': 20.295337677001953, 'timing/data': 0.00035604000731837004, 'timing/model': 0.36881468100182246, 'learning_rate/qwen_vl_interface': 8.563460469931823e-07, 'learning_rate/action_model': 8.563460469931822e-06, 'epoch': 0.32}) +[2026-10-03 15:28:06,659][starVLA.training.train_starvla][INFO] - Step 4393, Loss: {'train/loss': 0.000162072159582749, 'train/grad_norm_pre_clip': 0.1940307766199112, 'timing/data': 0.0003630999999586493, 'timing/model': 0.5615111840015743, 'learning_rate/qwen_vl_interface': 8.551896402371834e-07, 'learning_rate/action_model': 8.551896402371834e-06, 'epoch': 0.32}) +[2026-10-03 15:28:07,315][starVLA.training.train_starvla][INFO] - Step 4394, Loss: {'train/loss': 9.083010809263214e-05, 'train/grad_norm_pre_clip': 0.022869758307933807, 'timing/data': 0.0002106100000673905, 'timing/model': 0.3694058499968378, 'learning_rate/qwen_vl_interface': 8.540350400217177e-07, 'learning_rate/action_model': 8.540350400217176e-06, 'epoch': 0.32}) +[2026-10-03 15:28:07,974][starVLA.training.train_starvla][INFO] - Step 4395, Loss: {'train/loss': 7.990957237780094e-05, 'train/grad_norm_pre_clip': 0.038029495626688004, 'timing/data': 0.0002112299989676103, 'timing/model': 0.36981048100278713, 'learning_rate/qwen_vl_interface': 8.528822468213964e-07, 'learning_rate/action_model': 8.528822468213964e-06, 'epoch': 0.32}) +[2026-10-03 15:28:08,632][starVLA.training.train_starvla][INFO] - Step 4396, Loss: {'train/loss': 5.908285857003648e-06, 'train/grad_norm_pre_clip': 0.0009698314242996275, 'timing/data': 0.0003586200036806986, 'timing/model': 0.36906669099698775, 'learning_rate/qwen_vl_interface': 8.517312611100898e-07, 'learning_rate/action_model': 8.517312611100897e-06, 'epoch': 0.32}) +[2026-10-03 15:28:09,289][starVLA.training.train_starvla][INFO] - Step 4397, Loss: {'train/loss': 0.003219609148800373, 'train/grad_norm_pre_clip': 1.9250742197036743, 'timing/data': 0.0003693590115290135, 'timing/model': 0.3695336210075766, 'learning_rate/qwen_vl_interface': 8.505820833609249e-07, 'learning_rate/action_model': 8.50582083360925e-06, 'epoch': 0.32}) +[2026-10-03 15:28:10,150][starVLA.training.train_starvla][INFO] - Step 4398, Loss: {'train/loss': 0.00014787414693273604, 'train/grad_norm_pre_clip': 0.2656087279319763, 'timing/data': 0.000314389995764941, 'timing/model': 0.571959594002692, 'learning_rate/qwen_vl_interface': 8.494347140462847e-07, 'learning_rate/action_model': 8.494347140462846e-06, 'epoch': 0.32}) +[2026-10-03 15:28:10,808][starVLA.training.train_starvla][INFO] - Step 4399, Loss: {'train/loss': 6.637104524997994e-05, 'train/grad_norm_pre_clip': 0.028351036831736565, 'timing/data': 0.00037663000694010407, 'timing/model': 0.3701322200067807, 'learning_rate/qwen_vl_interface': 8.482891536378105e-07, 'learning_rate/action_model': 8.482891536378105e-06, 'epoch': 0.32}) +[2026-10-03 15:28:11,468][starVLA.training.train_starvla][INFO] - Step 4400, Loss: {'train/loss': 5.1056977099506184e-05, 'train/grad_norm_pre_clip': 0.05446420609951019, 'timing/data': 0.0003147200040984899, 'timing/model': 0.36959326099895407, 'learning_rate/qwen_vl_interface': 8.471454026063972e-07, 'learning_rate/action_model': 8.47145402606397e-06, 'epoch': 0.32}) +[2026-10-03 15:28:12,125][starVLA.training.train_starvla][INFO] - Step 4401, Loss: {'train/loss': 0.4126647710800171, 'train/grad_norm_pre_clip': 56.35521697998047, 'timing/data': 0.00020799999765586108, 'timing/model': 0.3695593410084257, 'learning_rate/qwen_vl_interface': 8.460034614221988e-07, 'learning_rate/action_model': 8.460034614221988e-06, 'epoch': 0.32}) +[2026-10-03 15:28:12,782][starVLA.training.train_starvla][INFO] - Step 4402, Loss: {'train/loss': 0.0016815520357340574, 'train/grad_norm_pre_clip': 0.8384358286857605, 'timing/data': 0.00021553000260610133, 'timing/model': 0.36941088100138586, 'learning_rate/qwen_vl_interface': 8.448633305546236e-07, 'learning_rate/action_model': 8.448633305546235e-06, 'epoch': 0.32}) +[2026-10-03 15:28:13,440][starVLA.training.train_starvla][INFO] - Step 4403, Loss: {'train/loss': 0.0004590938042383641, 'train/grad_norm_pre_clip': 0.16543470323085785, 'timing/data': 0.00022528000408783555, 'timing/model': 0.37042893099715, 'learning_rate/qwen_vl_interface': 8.437250104723369e-07, 'learning_rate/action_model': 8.437250104723369e-06, 'epoch': 0.32}) +[2026-10-03 15:28:14,292][starVLA.training.train_starvla][INFO] - Step 4404, Loss: {'train/loss': 7.47708254493773e-05, 'train/grad_norm_pre_clip': 0.053674809634685516, 'timing/data': 0.00025752000510692596, 'timing/model': 0.3695652099995641, 'learning_rate/qwen_vl_interface': 8.42588501643259e-07, 'learning_rate/action_model': 8.42588501643259e-06, 'epoch': 0.32}) +[2026-10-03 15:28:14,949][starVLA.training.train_starvla][INFO] - Step 4405, Loss: {'train/loss': 4.429127511684783e-05, 'train/grad_norm_pre_clip': 0.026228593662381172, 'timing/data': 0.0003618000046117231, 'timing/model': 0.36921148099645507, 'learning_rate/qwen_vl_interface': 8.414538045345655e-07, 'learning_rate/action_model': 8.414538045345654e-06, 'epoch': 0.32}) +[2026-10-03 15:28:15,606][starVLA.training.train_starvla][INFO] - Step 4406, Loss: {'train/loss': 0.45339223742485046, 'train/grad_norm_pre_clip': 213.38528442382812, 'timing/data': 0.0002598799910629168, 'timing/model': 0.36913313099648803, 'learning_rate/qwen_vl_interface': 8.403209196126871e-07, 'learning_rate/action_model': 8.403209196126871e-06, 'epoch': 0.32}) +[2026-10-03 15:28:16,264][starVLA.training.train_starvla][INFO] - Step 4407, Loss: {'train/loss': 0.00046179801574908197, 'train/grad_norm_pre_clip': 0.08406279981136322, 'timing/data': 0.00030051000067032874, 'timing/model': 0.3697250109980814, 'learning_rate/qwen_vl_interface': 8.391898473433111e-07, 'learning_rate/action_model': 8.39189847343311e-06, 'epoch': 0.32}) +[2026-10-03 15:28:16,923][starVLA.training.train_starvla][INFO] - Step 4408, Loss: {'train/loss': 0.00010471332643646747, 'train/grad_norm_pre_clip': 0.03073422983288765, 'timing/data': 0.0003755000070668757, 'timing/model': 0.3697960610006703, 'learning_rate/qwen_vl_interface': 8.380605881913778e-07, 'learning_rate/action_model': 8.380605881913779e-06, 'epoch': 0.32}) +[2026-10-03 15:28:17,776][starVLA.training.train_starvla][INFO] - Step 4409, Loss: {'train/loss': 3.7291476473910734e-05, 'train/grad_norm_pre_clip': 0.01636497676372528, 'timing/data': 0.00034276899532414973, 'timing/model': 0.36980641199625097, 'learning_rate/qwen_vl_interface': 8.369331426210832e-07, 'learning_rate/action_model': 8.369331426210832e-06, 'epoch': 0.32}) +[2026-10-03 15:28:18,434][starVLA.training.train_starvla][INFO] - Step 4410, Loss: {'train/loss': 0.6055529713630676, 'train/grad_norm_pre_clip': 32.42667007446289, 'timing/data': 0.00021302999812178314, 'timing/model': 0.3697188410005765, 'learning_rate/qwen_vl_interface': 8.358075110958785e-07, 'learning_rate/action_model': 8.358075110958784e-06, 'epoch': 0.32}) +[2026-10-03 15:28:19,092][starVLA.training.train_starvla][INFO] - Step 4411, Loss: {'train/loss': 0.20389631390571594, 'train/grad_norm_pre_clip': 228.7845916748047, 'timing/data': 0.00025293001090176404, 'timing/model': 0.36985006000031717, 'learning_rate/qwen_vl_interface': 8.346836940784661e-07, 'learning_rate/action_model': 8.346836940784661e-06, 'epoch': 0.32}) +[2026-10-03 15:28:19,751][starVLA.training.train_starvla][INFO] - Step 4412, Loss: {'train/loss': 0.00018111293320544064, 'train/grad_norm_pre_clip': 0.06772913783788681, 'timing/data': 0.0002695599978324026, 'timing/model': 0.3696324910124531, 'learning_rate/qwen_vl_interface': 8.335616920308061e-07, 'learning_rate/action_model': 8.33561692030806e-06, 'epoch': 0.32}) +[2026-10-03 15:28:20,409][starVLA.training.train_starvla][INFO] - Step 4413, Loss: {'train/loss': 8.98532016435638e-06, 'train/grad_norm_pre_clip': 0.0023268130607903004, 'timing/data': 0.0002511200000299141, 'timing/model': 0.36949659101082943, 'learning_rate/qwen_vl_interface': 8.324415054141111e-07, 'learning_rate/action_model': 8.32441505414111e-06, 'epoch': 0.32}) +[2026-10-03 15:28:21,240][starVLA.training.train_starvla][INFO] - Step 4414, Loss: {'train/loss': 2.7498233976075426e-05, 'train/grad_norm_pre_clip': 0.010649395175278187, 'timing/data': 0.00021598998864647, 'timing/model': 0.5660133050114382, 'learning_rate/qwen_vl_interface': 8.313231346888476e-07, 'learning_rate/action_model': 8.313231346888476e-06, 'epoch': 0.32}) +[2026-10-03 15:28:21,899][starVLA.training.train_starvla][INFO] - Step 4415, Loss: {'train/loss': 0.3131154775619507, 'train/grad_norm_pre_clip': 11.927438735961914, 'timing/data': 0.00029999999969732016, 'timing/model': 0.36963168100919574, 'learning_rate/qwen_vl_interface': 8.302065803147353e-07, 'learning_rate/action_model': 8.302065803147353e-06, 'epoch': 0.32}) +[2026-10-03 15:28:22,558][starVLA.training.train_starvla][INFO] - Step 4416, Loss: {'train/loss': 0.003411124926060438, 'train/grad_norm_pre_clip': 4.414663314819336, 'timing/data': 0.00021268999262247235, 'timing/model': 0.37052837100054603, 'learning_rate/qwen_vl_interface': 8.29091842750748e-07, 'learning_rate/action_model': 8.29091842750748e-06, 'epoch': 0.32}) +[2026-10-03 15:28:23,215][starVLA.training.train_starvla][INFO] - Step 4417, Loss: {'train/loss': 5.088725174573483e-06, 'train/grad_norm_pre_clip': 0.0009070283849723637, 'timing/data': 0.0002977600088343024, 'timing/model': 0.3695321699924534, 'learning_rate/qwen_vl_interface': 8.279789224551116e-07, 'learning_rate/action_model': 8.279789224551116e-06, 'epoch': 0.32}) +[2026-10-03 15:28:23,873][starVLA.training.train_starvla][INFO] - Step 4418, Loss: {'train/loss': 0.00011153604282299057, 'train/grad_norm_pre_clip': 0.05201788619160652, 'timing/data': 0.0002703299978747964, 'timing/model': 0.37010661100794096, 'learning_rate/qwen_vl_interface': 8.268678198853058e-07, 'learning_rate/action_model': 8.268678198853059e-06, 'epoch': 0.32}) +[2026-10-03 15:28:24,732][starVLA.training.train_starvla][INFO] - Step 4419, Loss: {'train/loss': 1.3872734598407988e-05, 'train/grad_norm_pre_clip': 0.0032161111012101173, 'timing/data': 0.00031150999711826444, 'timing/model': 0.569861152995145, 'learning_rate/qwen_vl_interface': 8.257585354980628e-07, 'learning_rate/action_model': 8.257585354980628e-06, 'epoch': 0.32}) +[2026-10-03 15:28:25,391][starVLA.training.train_starvla][INFO] - Step 4420, Loss: {'train/loss': 0.00013708478945773095, 'train/grad_norm_pre_clip': 0.07340586185455322, 'timing/data': 0.00031972999568097293, 'timing/model': 0.3697103410086129, 'learning_rate/qwen_vl_interface': 8.246510697493681e-07, 'learning_rate/action_model': 8.246510697493682e-06, 'epoch': 0.32}) +[2026-10-03 15:28:26,048][starVLA.training.train_starvla][INFO] - Step 4421, Loss: {'train/loss': 0.2283879965543747, 'train/grad_norm_pre_clip': 58.84563446044922, 'timing/data': 0.00028510999982245266, 'timing/model': 0.3692382920125965, 'learning_rate/qwen_vl_interface': 8.235454230944584e-07, 'learning_rate/action_model': 8.235454230944584e-06, 'epoch': 0.32}) +[2026-10-03 15:28:26,705][starVLA.training.train_starvla][INFO] - Step 4422, Loss: {'train/loss': 2.632817268022336e-05, 'train/grad_norm_pre_clip': 0.01377101056277752, 'timing/data': 0.0002916899975389242, 'timing/model': 0.36908019099792, 'learning_rate/qwen_vl_interface': 8.224415959878234e-07, 'learning_rate/action_model': 8.224415959878234e-06, 'epoch': 0.32}) +[2026-10-03 15:28:27,363][starVLA.training.train_starvla][INFO] - Step 4423, Loss: {'train/loss': 0.354583203792572, 'train/grad_norm_pre_clip': 58.51859664916992, 'timing/data': 0.00029812999127898365, 'timing/model': 0.3693849900009809, 'learning_rate/qwen_vl_interface': 8.213395888832052e-07, 'learning_rate/action_model': 8.213395888832051e-06, 'epoch': 0.32}) +[2026-10-03 15:28:28,022][starVLA.training.train_starvla][INFO] - Step 4424, Loss: {'train/loss': 7.38676535547711e-05, 'train/grad_norm_pre_clip': 0.02601434849202633, 'timing/data': 0.000275849990430288, 'timing/model': 0.3703749109990895, 'learning_rate/qwen_vl_interface': 8.202394022335958e-07, 'learning_rate/action_model': 8.202394022335957e-06, 'epoch': 0.32}) +[2026-10-03 15:28:28,874][starVLA.training.train_starvla][INFO] - Step 4425, Loss: {'train/loss': 2.353341733396519e-05, 'train/grad_norm_pre_clip': 0.010756428353488445, 'timing/data': 0.00033402000553905964, 'timing/model': 0.3698624110111268, 'learning_rate/qwen_vl_interface': 8.191410364912424e-07, 'learning_rate/action_model': 8.191410364912424e-06, 'epoch': 0.32}) +[2026-10-03 15:28:29,531][starVLA.training.train_starvla][INFO] - Step 4426, Loss: {'train/loss': 3.662962626549415e-05, 'train/grad_norm_pre_clip': 0.024670500308275223, 'timing/data': 0.00022743998852092773, 'timing/model': 0.3697085510066245, 'learning_rate/qwen_vl_interface': 8.1804449210764e-07, 'learning_rate/action_model': 8.1804449210764e-06, 'epoch': 0.32}) +[2026-10-03 15:28:30,193][starVLA.training.train_starvla][INFO] - Step 4427, Loss: {'train/loss': 9.462049092689995e-06, 'train/grad_norm_pre_clip': 0.0026386852841824293, 'timing/data': 0.0002568700001575053, 'timing/model': 0.3700304509984562, 'learning_rate/qwen_vl_interface': 8.169497695335377e-07, 'learning_rate/action_model': 8.169497695335376e-06, 'epoch': 0.32}) +[2026-10-03 15:28:30,860][starVLA.training.train_starvla][INFO] - Step 4428, Loss: {'train/loss': 1.1369362255209126e-05, 'train/grad_norm_pre_clip': 0.0027479343116283417, 'timing/data': 0.0003873899986501783, 'timing/model': 0.37349754000024404, 'learning_rate/qwen_vl_interface': 8.158568692189342e-07, 'learning_rate/action_model': 8.158568692189342e-06, 'epoch': 0.32}) +[2026-10-03 15:28:31,525][starVLA.training.train_starvla][INFO] - Step 4429, Loss: {'train/loss': 1.4610306607210077e-05, 'train/grad_norm_pre_clip': 0.004530074540525675, 'timing/data': 0.00021183000353630632, 'timing/model': 0.3733478100039065, 'learning_rate/qwen_vl_interface': 8.147657916130797e-07, 'learning_rate/action_model': 8.147657916130796e-06, 'epoch': 0.32}) +[2026-10-03 15:28:32,363][starVLA.training.train_starvla][INFO] - Step 4430, Loss: {'train/loss': 9.787079034140334e-05, 'train/grad_norm_pre_clip': 0.037649717181921005, 'timing/data': 0.00022813999385107309, 'timing/model': 0.34741222400043625, 'learning_rate/qwen_vl_interface': 8.136765371644753e-07, 'learning_rate/action_model': 8.136765371644753e-06, 'epoch': 0.32}) +[2026-10-03 15:28:33,023][starVLA.training.train_starvla][INFO] - Step 4431, Loss: {'train/loss': 0.2842966616153717, 'train/grad_norm_pre_clip': 211.87362670898438, 'timing/data': 0.0002590200019767508, 'timing/model': 0.3706860300007975, 'learning_rate/qwen_vl_interface': 8.125891063208713e-07, 'learning_rate/action_model': 8.125891063208712e-06, 'epoch': 0.32}) +[2026-10-03 15:28:33,683][starVLA.training.train_starvla][INFO] - Step 4432, Loss: {'train/loss': 5.409641016740352e-05, 'train/grad_norm_pre_clip': 0.023407844826579094, 'timing/data': 0.0002781999937724322, 'timing/model': 0.37024588100030087, 'learning_rate/qwen_vl_interface': 8.115034995292698e-07, 'learning_rate/action_model': 8.115034995292698e-06, 'epoch': 0.32}) +[2026-10-03 15:28:34,342][starVLA.training.train_starvla][INFO] - Step 4433, Loss: {'train/loss': 8.754319424042478e-06, 'train/grad_norm_pre_clip': 0.0019836407154798508, 'timing/data': 0.00035937999200541526, 'timing/model': 0.3701413609960582, 'learning_rate/qwen_vl_interface': 8.104197172359235e-07, 'learning_rate/action_model': 8.104197172359234e-06, 'epoch': 0.32}) +[2026-10-03 15:28:35,000][starVLA.training.train_starvla][INFO] - Step 4434, Loss: {'train/loss': 0.0003701020614244044, 'train/grad_norm_pre_clip': 0.22231315076351166, 'timing/data': 0.00021703999664168805, 'timing/model': 0.37033761000202503, 'learning_rate/qwen_vl_interface': 8.093377598863334e-07, 'learning_rate/action_model': 8.093377598863334e-06, 'epoch': 0.32}) +[2026-10-03 15:28:35,862][starVLA.training.train_starvla][INFO] - Step 4435, Loss: {'train/loss': 0.00010447404201840982, 'train/grad_norm_pre_clip': 0.17882148921489716, 'timing/data': 0.0002370899892412126, 'timing/model': 0.5727619139943272, 'learning_rate/qwen_vl_interface': 8.082576279252519e-07, 'learning_rate/action_model': 8.082576279252518e-06, 'epoch': 0.32}) +[2026-10-03 15:28:36,522][starVLA.training.train_starvla][INFO] - Step 4436, Loss: {'train/loss': 0.0001591819163877517, 'train/grad_norm_pre_clip': 0.06355765461921692, 'timing/data': 0.0002782200026558712, 'timing/model': 0.369960001000436, 'learning_rate/qwen_vl_interface': 8.071793217966803e-07, 'learning_rate/action_model': 8.071793217966802e-06, 'epoch': 0.32}) +[2026-10-03 15:28:37,181][starVLA.training.train_starvla][INFO] - Step 4437, Loss: {'train/loss': 0.01921798475086689, 'train/grad_norm_pre_clip': 79.52137756347656, 'timing/data': 0.00021610000112559646, 'timing/model': 0.3702449999982491, 'learning_rate/qwen_vl_interface': 8.061028419438691e-07, 'learning_rate/action_model': 8.06102841943869e-06, 'epoch': 0.32}) +[2026-10-03 15:28:37,838][starVLA.training.train_starvla][INFO] - Step 4438, Loss: {'train/loss': 5.975356907583773e-05, 'train/grad_norm_pre_clip': 0.016336917877197266, 'timing/data': 0.00030375999631360173, 'timing/model': 0.3695376209943788, 'learning_rate/qwen_vl_interface': 8.050281888093193e-07, 'learning_rate/action_model': 8.050281888093192e-06, 'epoch': 0.32}) +[2026-10-03 15:28:38,474][starVLA.training.train_starvla][INFO] - Step 4439, Loss: {'train/loss': 0.001033352455124259, 'train/grad_norm_pre_clip': 0.3959185481071472, 'timing/data': 0.00026001999503932893, 'timing/model': 0.34714107400213834, 'learning_rate/qwen_vl_interface': 8.039553628347794e-07, 'learning_rate/action_model': 8.039553628347793e-06, 'epoch': 0.32}) +[2026-10-03 15:28:39,135][starVLA.training.train_starvla][INFO] - Step 4440, Loss: {'train/loss': 0.22940784692764282, 'train/grad_norm_pre_clip': 38.87700653076172, 'timing/data': 0.00038607999158557504, 'timing/model': 0.3703691510017961, 'learning_rate/qwen_vl_interface': 8.028843644612483e-07, 'learning_rate/action_model': 8.028843644612482e-06, 'epoch': 0.32}) +[2026-10-03 15:28:39,992][starVLA.training.train_starvla][INFO] - Step 4441, Loss: {'train/loss': 4.5936376409372315e-05, 'train/grad_norm_pre_clip': 0.014298063702881336, 'timing/data': 0.00031088000105228275, 'timing/model': 0.37019164000230376, 'learning_rate/qwen_vl_interface': 8.01815194128973e-07, 'learning_rate/action_model': 8.01815194128973e-06, 'epoch': 0.33}) +[2026-10-03 15:28:40,648][starVLA.training.train_starvla][INFO] - Step 4442, Loss: {'train/loss': 5.944489748799242e-05, 'train/grad_norm_pre_clip': 0.02650197595357895, 'timing/data': 0.0002126299950759858, 'timing/model': 0.3692393609962892, 'learning_rate/qwen_vl_interface': 8.007478522774481e-07, 'learning_rate/action_model': 8.00747852277448e-06, 'epoch': 0.33}) +[2026-10-03 15:28:41,308][starVLA.training.train_starvla][INFO] - Step 4443, Loss: {'train/loss': 0.42115354537963867, 'train/grad_norm_pre_clip': 43.01185607910156, 'timing/data': 0.00027231899730395526, 'timing/model': 0.36948160199972335, 'learning_rate/qwen_vl_interface': 7.996823393454197e-07, 'learning_rate/action_model': 7.996823393454196e-06, 'epoch': 0.33}) +[2026-10-03 15:28:41,967][starVLA.training.train_starvla][INFO] - Step 4444, Loss: {'train/loss': 0.0009364680154249072, 'train/grad_norm_pre_clip': 0.1814843863248825, 'timing/data': 0.0003313799970783293, 'timing/model': 0.3698697310028365, 'learning_rate/qwen_vl_interface': 7.986186557708773e-07, 'learning_rate/action_model': 7.986186557708771e-06, 'epoch': 0.33}) +[2026-10-03 15:28:42,626][starVLA.training.train_starvla][INFO] - Step 4445, Loss: {'train/loss': 2.4071339794318192e-05, 'train/grad_norm_pre_clip': 0.007400491274893284, 'timing/data': 0.00029183000151533633, 'timing/model': 0.370130520008388, 'learning_rate/qwen_vl_interface': 7.975568019910617e-07, 'learning_rate/action_model': 7.975568019910617e-06, 'epoch': 0.33}) +[2026-10-03 15:28:43,486][starVLA.training.train_starvla][INFO] - Step 4446, Loss: {'train/loss': 1.6554713511141017e-05, 'train/grad_norm_pre_clip': 0.003369402140378952, 'timing/data': 0.0002872500044759363, 'timing/model': 0.36972660099854693, 'learning_rate/qwen_vl_interface': 7.964967784424618e-07, 'learning_rate/action_model': 7.964967784424618e-06, 'epoch': 0.33}) +[2026-10-03 15:28:44,145][starVLA.training.train_starvla][INFO] - Step 4447, Loss: {'train/loss': 0.0005679401219822466, 'train/grad_norm_pre_clip': 0.10415783524513245, 'timing/data': 0.00020881999807897955, 'timing/model': 0.3699703610036522, 'learning_rate/qwen_vl_interface': 7.954385855608129e-07, 'learning_rate/action_model': 7.95438585560813e-06, 'epoch': 0.33}) +[2026-10-03 15:28:44,804][starVLA.training.train_starvla][INFO] - Step 4448, Loss: {'train/loss': 0.0006376850651577115, 'train/grad_norm_pre_clip': 1.0577428340911865, 'timing/data': 0.000344770000083372, 'timing/model': 0.37010410099173896, 'learning_rate/qwen_vl_interface': 7.943822237810974e-07, 'learning_rate/action_model': 7.943822237810975e-06, 'epoch': 0.33}) +[2026-10-03 15:28:45,462][starVLA.training.train_starvla][INFO] - Step 4449, Loss: {'train/loss': 1.560091186547652e-05, 'train/grad_norm_pre_clip': 0.006754371337592602, 'timing/data': 0.00020895000488962978, 'timing/model': 0.3699419609911274, 'learning_rate/qwen_vl_interface': 7.93327693537547e-07, 'learning_rate/action_model': 7.93327693537547e-06, 'epoch': 0.33}) +[2026-10-03 15:28:46,123][starVLA.training.train_starvla][INFO] - Step 4450, Loss: {'train/loss': 9.772974590305239e-05, 'train/grad_norm_pre_clip': 0.04779145121574402, 'timing/data': 0.0003297399962320924, 'timing/model': 0.369853839991265, 'learning_rate/qwen_vl_interface': 7.92274995263638e-07, 'learning_rate/action_model': 7.92274995263638e-06, 'epoch': 0.33}) +[2026-10-03 15:28:46,997][starVLA.training.train_starvla][INFO] - Step 4451, Loss: {'train/loss': 0.0003849928907584399, 'train/grad_norm_pre_clip': 0.7561346292495728, 'timing/data': 0.0002997499977936968, 'timing/model': 0.5815305029973388, 'learning_rate/qwen_vl_interface': 7.912241293920958e-07, 'learning_rate/action_model': 7.912241293920959e-06, 'epoch': 0.33}) +[2026-10-03 15:28:47,662][starVLA.training.train_starvla][INFO] - Step 4452, Loss: {'train/loss': 0.00023705113562755287, 'train/grad_norm_pre_clip': 0.09072458744049072, 'timing/data': 0.0004530300066107884, 'timing/model': 0.37363894999725744, 'learning_rate/qwen_vl_interface': 7.90175096354891e-07, 'learning_rate/action_model': 7.90175096354891e-06, 'epoch': 0.33}) +[2026-10-03 15:28:48,326][starVLA.training.train_starvla][INFO] - Step 4453, Loss: {'train/loss': 2.2074764274293557e-05, 'train/grad_norm_pre_clip': 0.011119984090328217, 'timing/data': 0.00036078999983146787, 'timing/model': 0.3724976899975445, 'learning_rate/qwen_vl_interface': 7.891278965832431e-07, 'learning_rate/action_model': 7.891278965832432e-06, 'epoch': 0.33}) +[2026-10-03 15:28:48,988][starVLA.training.train_starvla][INFO] - Step 4454, Loss: {'train/loss': 0.34378504753112793, 'train/grad_norm_pre_clip': 45.54080581665039, 'timing/data': 0.00041962999966926873, 'timing/model': 0.3706687600060832, 'learning_rate/qwen_vl_interface': 7.880825305076138e-07, 'learning_rate/action_model': 7.880825305076139e-06, 'epoch': 0.33}) +[2026-10-03 15:28:49,653][starVLA.training.train_starvla][INFO] - Step 4455, Loss: {'train/loss': 1.4922326954547316e-05, 'train/grad_norm_pre_clip': 0.009319769218564034, 'timing/data': 0.00021254000603221357, 'timing/model': 0.3737134709954262, 'learning_rate/qwen_vl_interface': 7.870389985577148e-07, 'learning_rate/action_model': 7.870389985577148e-06, 'epoch': 0.33}) +[2026-10-03 15:28:50,314][starVLA.training.train_starvla][INFO] - Step 4456, Loss: {'train/loss': 0.0009926651837304235, 'train/grad_norm_pre_clip': 2.2060675621032715, 'timing/data': 0.0002799299982143566, 'timing/model': 0.37033796099422034, 'learning_rate/qwen_vl_interface': 7.859973011625028e-07, 'learning_rate/action_model': 7.859973011625026e-06, 'epoch': 0.33}) +[2026-10-03 15:28:51,168][starVLA.training.train_starvla][INFO] - Step 4457, Loss: {'train/loss': 2.125506944139488e-05, 'train/grad_norm_pre_clip': 0.01012432761490345, 'timing/data': 0.00032083899714052677, 'timing/model': 0.3696566620055819, 'learning_rate/qwen_vl_interface': 7.849574387501801e-07, 'learning_rate/action_model': 7.8495743875018e-06, 'epoch': 0.33}) +[2026-10-03 15:28:51,826][starVLA.training.train_starvla][INFO] - Step 4458, Loss: {'train/loss': 0.0001431787823094055, 'train/grad_norm_pre_clip': 0.0865616425871849, 'timing/data': 0.00040952999552246183, 'timing/model': 0.3704644209938124, 'learning_rate/qwen_vl_interface': 7.839194117481952e-07, 'learning_rate/action_model': 7.839194117481952e-06, 'epoch': 0.33}) +[2026-10-03 15:28:52,484][starVLA.training.train_starvla][INFO] - Step 4459, Loss: {'train/loss': 6.548999863298377e-06, 'train/grad_norm_pre_clip': 0.0015977544244378805, 'timing/data': 0.0003333400090923533, 'timing/model': 0.36946255998918787, 'learning_rate/qwen_vl_interface': 7.828832205832418e-07, 'learning_rate/action_model': 7.828832205832418e-06, 'epoch': 0.33}) +[2026-10-03 15:28:53,143][starVLA.training.train_starvla][INFO] - Step 4460, Loss: {'train/loss': 7.5534000643529e-05, 'train/grad_norm_pre_clip': 0.04188938066363335, 'timing/data': 0.0003411000070627779, 'timing/model': 0.36978716000157874, 'learning_rate/qwen_vl_interface': 7.818488656812581e-07, 'learning_rate/action_model': 7.818488656812582e-06, 'epoch': 0.33}) +[2026-10-03 15:28:53,800][starVLA.training.train_starvla][INFO] - Step 4461, Loss: {'train/loss': 4.84981392219197e-05, 'train/grad_norm_pre_clip': 0.014306837692856789, 'timing/data': 0.0003295899950899184, 'timing/model': 0.3689129900012631, 'learning_rate/qwen_vl_interface': 7.808163474674291e-07, 'learning_rate/action_model': 7.80816347467429e-06, 'epoch': 0.33}) +[2026-10-03 15:28:54,649][starVLA.training.train_starvla][INFO] - Step 4462, Loss: {'train/loss': 0.005848797503858805, 'train/grad_norm_pre_clip': 14.664636611938477, 'timing/data': 0.00035266899794805795, 'timing/model': 0.368679441002314, 'learning_rate/qwen_vl_interface': 7.797856663661836e-07, 'learning_rate/action_model': 7.797856663661835e-06, 'epoch': 0.33}) +[2026-10-03 15:28:55,306][starVLA.training.train_starvla][INFO] - Step 4463, Loss: {'train/loss': 2.615815174067393e-05, 'train/grad_norm_pre_clip': 0.008093212731182575, 'timing/data': 0.0002503100113244727, 'timing/model': 0.36854020100145135, 'learning_rate/qwen_vl_interface': 7.787568228011959e-07, 'learning_rate/action_model': 7.78756822801196e-06, 'epoch': 0.33}) +[2026-10-03 15:28:55,964][starVLA.training.train_starvla][INFO] - Step 4464, Loss: {'train/loss': 2.52338304562727e-05, 'train/grad_norm_pre_clip': 0.008996090851724148, 'timing/data': 0.00032495999766979367, 'timing/model': 0.3696616010129219, 'learning_rate/qwen_vl_interface': 7.777298171953858e-07, 'learning_rate/action_model': 7.777298171953857e-06, 'epoch': 0.33}) +[2026-10-03 15:28:56,620][starVLA.training.train_starvla][INFO] - Step 4465, Loss: {'train/loss': 0.0008791768341325223, 'train/grad_norm_pre_clip': 0.2157420814037323, 'timing/data': 0.0003278600051999092, 'timing/model': 0.3684569519973593, 'learning_rate/qwen_vl_interface': 7.767046499709144e-07, 'learning_rate/action_model': 7.767046499709143e-06, 'epoch': 0.33}) +[2026-10-03 15:28:57,276][starVLA.training.train_starvla][INFO] - Step 4466, Loss: {'train/loss': 0.30966776609420776, 'train/grad_norm_pre_clip': 12.557255744934082, 'timing/data': 0.0002480999974068254, 'timing/model': 0.3691503810114227, 'learning_rate/qwen_vl_interface': 7.756813215491904e-07, 'learning_rate/action_model': 7.756813215491904e-06, 'epoch': 0.33}) +[2026-10-03 15:28:58,125][starVLA.training.train_starvla][INFO] - Step 4467, Loss: {'train/loss': 0.03148831054568291, 'train/grad_norm_pre_clip': 65.82605743408203, 'timing/data': 0.0003228799905627966, 'timing/model': 0.36900620099913795, 'learning_rate/qwen_vl_interface': 7.746598323508655e-07, 'learning_rate/action_model': 7.746598323508655e-06, 'epoch': 0.33}) +[2026-10-03 15:28:58,759][starVLA.training.train_starvla][INFO] - Step 4468, Loss: {'train/loss': 4.664684092858806e-05, 'train/grad_norm_pre_clip': 0.026146531105041504, 'timing/data': 0.0002408500004094094, 'timing/model': 0.34598272400035057, 'learning_rate/qwen_vl_interface': 7.736401827958351e-07, 'learning_rate/action_model': 7.73640182795835e-06, 'epoch': 0.33}) +[2026-10-03 15:28:59,415][starVLA.training.train_starvla][INFO] - Step 4469, Loss: {'train/loss': 6.785805453546345e-05, 'train/grad_norm_pre_clip': 0.054619934409856796, 'timing/data': 0.00033194999559782445, 'timing/model': 0.3684892010060139, 'learning_rate/qwen_vl_interface': 7.726223733032382e-07, 'learning_rate/action_model': 7.72622373303238e-06, 'epoch': 0.33}) +[2026-10-03 15:29:00,070][starVLA.training.train_starvla][INFO] - Step 4470, Loss: {'train/loss': 0.00016066926764324307, 'train/grad_norm_pre_clip': 0.06651229411363602, 'timing/data': 0.00025950999406632036, 'timing/model': 0.36862599100277293, 'learning_rate/qwen_vl_interface': 7.716064042914584e-07, 'learning_rate/action_model': 7.716064042914584e-06, 'epoch': 0.33}) +[2026-10-03 15:29:00,726][starVLA.training.train_starvla][INFO] - Step 4471, Loss: {'train/loss': 0.29154932498931885, 'train/grad_norm_pre_clip': 14.632964134216309, 'timing/data': 0.00022540999634657055, 'timing/model': 0.36828870099270716, 'learning_rate/qwen_vl_interface': 7.705922761781226e-07, 'learning_rate/action_model': 7.705922761781226e-06, 'epoch': 0.33}) +[2026-10-03 15:29:01,573][starVLA.training.train_starvla][INFO] - Step 4472, Loss: {'train/loss': 0.0005389094003476202, 'train/grad_norm_pre_clip': 0.6046339869499207, 'timing/data': 0.00023920999956317246, 'timing/model': 0.5587675450078677, 'learning_rate/qwen_vl_interface': 7.695799893800996e-07, 'learning_rate/action_model': 7.695799893800996e-06, 'epoch': 0.33}) +[2026-10-03 15:29:02,229][starVLA.training.train_starvla][INFO] - Step 4473, Loss: {'train/loss': 4.999316843168344e-06, 'train/grad_norm_pre_clip': 0.0007943561649881303, 'timing/data': 0.00032616998942103237, 'timing/model': 0.36885876100859605, 'learning_rate/qwen_vl_interface': 7.685695443135028e-07, 'learning_rate/action_model': 7.685695443135027e-06, 'epoch': 0.33}) +[2026-10-03 15:29:02,884][starVLA.training.train_starvla][INFO] - Step 4474, Loss: {'train/loss': 5.481143307406455e-05, 'train/grad_norm_pre_clip': 0.029654884710907936, 'timing/data': 0.00029565999284386635, 'timing/model': 0.3689642210083548, 'learning_rate/qwen_vl_interface': 7.675609413936888e-07, 'learning_rate/action_model': 7.675609413936888e-06, 'epoch': 0.33}) +[2026-10-03 15:29:03,541][starVLA.training.train_starvla][INFO] - Step 4475, Loss: {'train/loss': 0.27396199107170105, 'train/grad_norm_pre_clip': 22.29960823059082, 'timing/data': 0.0002048800088232383, 'timing/model': 0.3687120410031639, 'learning_rate/qwen_vl_interface': 7.665541810352541e-07, 'learning_rate/action_model': 7.66554181035254e-06, 'epoch': 0.33}) +[2026-10-03 15:29:04,176][starVLA.training.train_starvla][INFO] - Step 4476, Loss: {'train/loss': 6.479965668404475e-05, 'train/grad_norm_pre_clip': 0.02421736903488636, 'timing/data': 0.0002579400024842471, 'timing/model': 0.368925060989568, 'learning_rate/qwen_vl_interface': 7.655492636520415e-07, 'learning_rate/action_model': 7.655492636520414e-06, 'epoch': 0.33}) +[2026-10-03 15:29:05,038][starVLA.training.train_starvla][INFO] - Step 4477, Loss: {'train/loss': 0.3934255540370941, 'train/grad_norm_pre_clip': 73.99713897705078, 'timing/data': 0.0003247499989811331, 'timing/model': 0.5742151140002534, 'learning_rate/qwen_vl_interface': 7.64546189657134e-07, 'learning_rate/action_model': 7.645461896571339e-06, 'epoch': 0.33}) +[2026-10-03 15:29:05,693][starVLA.training.train_starvla][INFO] - Step 4478, Loss: {'train/loss': 0.00012562001938931644, 'train/grad_norm_pre_clip': 0.04001501575112343, 'timing/data': 0.0003289299929747358, 'timing/model': 0.3687462309899274, 'learning_rate/qwen_vl_interface': 7.635449594628571e-07, 'learning_rate/action_model': 7.635449594628571e-06, 'epoch': 0.33}) +[2026-10-03 15:29:06,351][starVLA.training.train_starvla][INFO] - Step 4479, Loss: {'train/loss': 3.475999255897477e-05, 'train/grad_norm_pre_clip': 0.015343818813562393, 'timing/data': 0.00025427000946365297, 'timing/model': 0.3694307810073951, 'learning_rate/qwen_vl_interface': 7.625455734807803e-07, 'learning_rate/action_model': 7.625455734807803e-06, 'epoch': 0.33}) +[2026-10-03 15:29:07,010][starVLA.training.train_starvla][INFO] - Step 4480, Loss: {'train/loss': 4.538859138847329e-05, 'train/grad_norm_pre_clip': 0.020364493131637573, 'timing/data': 0.0003486299974611029, 'timing/model': 0.3701771810010541, 'learning_rate/qwen_vl_interface': 7.615480321217116e-07, 'learning_rate/action_model': 7.615480321217117e-06, 'epoch': 0.33}) +[2026-10-03 15:29:07,667][starVLA.training.train_starvla][INFO] - Step 4481, Loss: {'train/loss': 0.00029491237364709377, 'train/grad_norm_pre_clip': 0.22634926438331604, 'timing/data': 0.0003672099992400035, 'timing/model': 0.37007787100446876, 'learning_rate/qwen_vl_interface': 7.605523357957028e-07, 'learning_rate/action_model': 7.605523357957028e-06, 'epoch': 0.33}) +[2026-10-03 15:29:08,324][starVLA.training.train_starvla][INFO] - Step 4482, Loss: {'train/loss': 0.0008468183223158121, 'train/grad_norm_pre_clip': 0.15979605913162231, 'timing/data': 0.00031146999390330166, 'timing/model': 0.369090911000967, 'learning_rate/qwen_vl_interface': 7.59558484912048e-07, 'learning_rate/action_model': 7.59558484912048e-06, 'epoch': 0.33}) +[2026-10-03 15:29:09,184][starVLA.training.train_starvla][INFO] - Step 4483, Loss: {'train/loss': 2.430243512208108e-05, 'train/grad_norm_pre_clip': 0.007055861409753561, 'timing/data': 0.000218679997487925, 'timing/model': 0.369407771009719, 'learning_rate/qwen_vl_interface': 7.585664798792807e-07, 'learning_rate/action_model': 7.585664798792807e-06, 'epoch': 0.33}) +[2026-10-03 15:29:09,846][starVLA.training.train_starvla][INFO] - Step 4484, Loss: {'train/loss': 4.037181497551501e-05, 'train/grad_norm_pre_clip': 0.01808321662247181, 'timing/data': 0.0004020100022898987, 'timing/model': 0.3702053509914549, 'learning_rate/qwen_vl_interface': 7.575763211051779e-07, 'learning_rate/action_model': 7.575763211051778e-06, 'epoch': 0.33}) +[2026-10-03 15:29:10,503][starVLA.training.train_starvla][INFO] - Step 4485, Loss: {'train/loss': 0.1890677958726883, 'train/grad_norm_pre_clip': 205.0030059814453, 'timing/data': 0.00033363999682478607, 'timing/model': 0.3694245999940904, 'learning_rate/qwen_vl_interface': 7.565880089967553e-07, 'learning_rate/action_model': 7.565880089967553e-06, 'epoch': 0.33}) +[2026-10-03 15:29:11,137][starVLA.training.train_starvla][INFO] - Step 4486, Loss: {'train/loss': 0.0002088324399664998, 'train/grad_norm_pre_clip': 0.1726197451353073, 'timing/data': 0.0002475700021022931, 'timing/model': 0.34620328299934044, 'learning_rate/qwen_vl_interface': 7.556015439602725e-07, 'learning_rate/action_model': 7.556015439602725e-06, 'epoch': 0.33}) +[2026-10-03 15:29:11,794][starVLA.training.train_starvla][INFO] - Step 4487, Loss: {'train/loss': 0.00012448630877770483, 'train/grad_norm_pre_clip': 0.040851760655641556, 'timing/data': 0.00029381000786088407, 'timing/model': 0.3691287910041865, 'learning_rate/qwen_vl_interface': 7.546169264012252e-07, 'learning_rate/action_model': 7.5461692640122525e-06, 'epoch': 0.33}) +[2026-10-03 15:29:12,650][starVLA.training.train_starvla][INFO] - Step 4488, Loss: {'train/loss': 4.90031270601321e-05, 'train/grad_norm_pre_clip': 0.02223091758787632, 'timing/data': 0.0003779600083362311, 'timing/model': 0.5668061540054623, 'learning_rate/qwen_vl_interface': 7.536341567243544e-07, 'learning_rate/action_model': 7.536341567243543e-06, 'epoch': 0.33}) +[2026-10-03 15:29:13,307][starVLA.training.train_starvla][INFO] - Step 4489, Loss: {'train/loss': 5.4362855735234916e-05, 'train/grad_norm_pre_clip': 0.04125164449214935, 'timing/data': 0.00022016999719198793, 'timing/model': 0.36923698098689783, 'learning_rate/qwen_vl_interface': 7.526532353336395e-07, 'learning_rate/action_model': 7.526532353336395e-06, 'epoch': 0.33}) +[2026-10-03 15:29:13,963][starVLA.training.train_starvla][INFO] - Step 4490, Loss: {'train/loss': 2.928553476522211e-05, 'train/grad_norm_pre_clip': 0.013956691138446331, 'timing/data': 0.0003398099943296984, 'timing/model': 0.36919618099636864, 'learning_rate/qwen_vl_interface': 7.516741626322994e-07, 'learning_rate/action_model': 7.516741626322994e-06, 'epoch': 0.33}) +[2026-10-03 15:29:14,598][starVLA.training.train_starvla][INFO] - Step 4491, Loss: {'train/loss': 4.321319465816487e-06, 'train/grad_norm_pre_clip': 0.000843350775539875, 'timing/data': 0.00031316999229602516, 'timing/model': 0.36888422099582385, 'learning_rate/qwen_vl_interface': 7.506969390227945e-07, 'learning_rate/action_model': 7.506969390227945e-06, 'epoch': 0.33}) +[2026-10-03 15:29:15,234][starVLA.training.train_starvla][INFO] - Step 4492, Loss: {'train/loss': 0.37167397141456604, 'train/grad_norm_pre_clip': 85.87628173828125, 'timing/data': 0.0002640200109453872, 'timing/model': 0.3463322940078797, 'learning_rate/qwen_vl_interface': 7.497215649068244e-07, 'learning_rate/action_model': 7.497215649068244e-06, 'epoch': 0.33}) +[2026-10-03 15:29:16,095][starVLA.training.train_starvla][INFO] - Step 4493, Loss: {'train/loss': 0.3086782991886139, 'train/grad_norm_pre_clip': 143.80409240722656, 'timing/data': 0.0003602700016926974, 'timing/model': 0.5733355740085244, 'learning_rate/qwen_vl_interface': 7.48748040685329e-07, 'learning_rate/action_model': 7.487480406853289e-06, 'epoch': 0.33}) +[2026-10-03 15:29:16,753][starVLA.training.train_starvla][INFO] - Step 4494, Loss: {'train/loss': 0.3015657365322113, 'train/grad_norm_pre_clip': 48.40920639038086, 'timing/data': 0.00034270000469405204, 'timing/model': 0.3695846310001798, 'learning_rate/qwen_vl_interface': 7.477763667584867e-07, 'learning_rate/action_model': 7.477763667584866e-06, 'epoch': 0.33}) +[2026-10-03 15:29:17,411][starVLA.training.train_starvla][INFO] - Step 4495, Loss: {'train/loss': 0.2044689804315567, 'train/grad_norm_pre_clip': 42.34245300292969, 'timing/data': 0.00028099000337533653, 'timing/model': 0.3694292010040954, 'learning_rate/qwen_vl_interface': 7.468065435257159e-07, 'learning_rate/action_model': 7.468065435257159e-06, 'epoch': 0.33}) +[2026-10-03 15:29:18,070][starVLA.training.train_starvla][INFO] - Step 4496, Loss: {'train/loss': 1.7120943084592e-05, 'train/grad_norm_pre_clip': 0.004764102399349213, 'timing/data': 0.00021440000273287296, 'timing/model': 0.3691996709967498, 'learning_rate/qwen_vl_interface': 7.458385713856745e-07, 'learning_rate/action_model': 7.458385713856744e-06, 'epoch': 0.33}) +[2026-10-03 15:29:18,727][starVLA.training.train_starvla][INFO] - Step 4497, Loss: {'train/loss': 0.37300705909729004, 'train/grad_norm_pre_clip': 119.59785461425781, 'timing/data': 0.00033943999733310193, 'timing/model': 0.36965797100856435, 'learning_rate/qwen_vl_interface': 7.448724507362605e-07, 'learning_rate/action_model': 7.448724507362605e-06, 'epoch': 0.33}) +[2026-10-03 15:29:19,383][starVLA.training.train_starvla][INFO] - Step 4498, Loss: {'train/loss': 0.001027841353788972, 'train/grad_norm_pre_clip': 0.26424577832221985, 'timing/data': 0.00033325000549666584, 'timing/model': 0.36874319100752473, 'learning_rate/qwen_vl_interface': 7.439081819746078e-07, 'learning_rate/action_model': 7.4390818197460775e-06, 'epoch': 0.33}) +[2026-10-03 15:29:20,230][starVLA.training.train_starvla][INFO] - Step 4499, Loss: {'train/loss': 0.00020435513579286635, 'train/grad_norm_pre_clip': 0.10794240981340408, 'timing/data': 0.00025594999897293746, 'timing/model': 0.3696017100010067, 'learning_rate/qwen_vl_interface': 7.42945765497091e-07, 'learning_rate/action_model': 7.42945765497091e-06, 'epoch': 0.33}) +[2026-10-03 15:29:20,888][starVLA.training.train_starvla][INFO] - Step 4500, Loss: {'train/loss': 6.007643241900951e-05, 'train/grad_norm_pre_clip': 0.021139120683073997, 'timing/data': 0.0002647600049385801, 'timing/model': 0.36928432100103237, 'learning_rate/qwen_vl_interface': 7.41985201699324e-07, 'learning_rate/action_model': 7.41985201699324e-06, 'epoch': 0.33}) +[2026-10-03 15:29:21,544][starVLA.training.train_starvla][INFO] - Step 4501, Loss: {'train/loss': 0.0013519813073799014, 'train/grad_norm_pre_clip': 2.107450008392334, 'timing/data': 0.00032114000350702554, 'timing/model': 0.3685769409930799, 'learning_rate/qwen_vl_interface': 7.410264909761581e-07, 'learning_rate/action_model': 7.41026490976158e-06, 'epoch': 0.33}) +[2026-10-03 15:29:22,201][starVLA.training.train_starvla][INFO] - Step 4502, Loss: {'train/loss': 9.083615441340953e-05, 'train/grad_norm_pre_clip': 0.0309870857745409, 'timing/data': 0.00033756000630091876, 'timing/model': 0.36875564100046176, 'learning_rate/qwen_vl_interface': 7.40069633721683e-07, 'learning_rate/action_model': 7.400696337216829e-06, 'epoch': 0.33}) +[2026-10-03 15:29:22,858][starVLA.training.train_starvla][INFO] - Step 4503, Loss: {'train/loss': 0.8195968866348267, 'train/grad_norm_pre_clip': 24.18793487548828, 'timing/data': 0.0002537399996072054, 'timing/model': 0.3692049209930701, 'learning_rate/qwen_vl_interface': 7.391146303292262e-07, 'learning_rate/action_model': 7.391146303292262e-06, 'epoch': 0.33}) +[2026-10-03 15:29:23,714][starVLA.training.train_starvla][INFO] - Step 4504, Loss: {'train/loss': 0.0010514192981645465, 'train/grad_norm_pre_clip': 0.3855975568294525, 'timing/data': 0.00021849000768270344, 'timing/model': 0.3701767909951741, 'learning_rate/qwen_vl_interface': 7.381614811913538e-07, 'learning_rate/action_model': 7.381614811913537e-06, 'epoch': 0.33}) +[2026-10-03 15:29:24,372][starVLA.training.train_starvla][INFO] - Step 4505, Loss: {'train/loss': 2.5881159672280774e-05, 'train/grad_norm_pre_clip': 0.009292485192418098, 'timing/data': 0.0003079899906879291, 'timing/model': 0.3697273709985893, 'learning_rate/qwen_vl_interface': 7.372101866998693e-07, 'learning_rate/action_model': 7.372101866998693e-06, 'epoch': 0.33}) +[2026-10-03 15:29:25,031][starVLA.training.train_starvla][INFO] - Step 4506, Loss: {'train/loss': 1.6159439837792888e-05, 'train/grad_norm_pre_clip': 0.009606363251805305, 'timing/data': 0.0002129200001945719, 'timing/model': 0.37087919100304134, 'learning_rate/qwen_vl_interface': 7.362607472458147e-07, 'learning_rate/action_model': 7.362607472458147e-06, 'epoch': 0.33}) +[2026-10-03 15:29:25,690][starVLA.training.train_starvla][INFO] - Step 4507, Loss: {'train/loss': 0.003189223352819681, 'train/grad_norm_pre_clip': 6.400158405303955, 'timing/data': 0.000212610000744462, 'timing/model': 0.37008296100248117, 'learning_rate/qwen_vl_interface': 7.353131632194682e-07, 'learning_rate/action_model': 7.3531316321946816e-06, 'epoch': 0.33}) +[2026-10-03 15:29:26,351][starVLA.training.train_starvla][INFO] - Step 4508, Loss: {'train/loss': 5.818759382236749e-05, 'train/grad_norm_pre_clip': 0.016815070062875748, 'timing/data': 0.00021688999549951404, 'timing/model': 0.37080157099990174, 'learning_rate/qwen_vl_interface': 7.343674350103455e-07, 'learning_rate/action_model': 7.343674350103455e-06, 'epoch': 0.33}) +[2026-10-03 15:29:27,203][starVLA.training.train_starvla][INFO] - Step 4509, Loss: {'train/loss': 6.463308818638325e-05, 'train/grad_norm_pre_clip': 0.02963060326874256, 'timing/data': 0.0002505799930077046, 'timing/model': 0.5637379249965306, 'learning_rate/qwen_vl_interface': 7.334235630071995e-07, 'learning_rate/action_model': 7.334235630071995e-06, 'epoch': 0.33}) +[2026-10-03 15:29:27,860][starVLA.training.train_starvla][INFO] - Step 4510, Loss: {'train/loss': 8.960023114923388e-05, 'train/grad_norm_pre_clip': 0.07113111764192581, 'timing/data': 0.0003260200028307736, 'timing/model': 0.3693400510092033, 'learning_rate/qwen_vl_interface': 7.324815475980207e-07, 'learning_rate/action_model': 7.324815475980207e-06, 'epoch': 0.33}) +[2026-10-03 15:29:28,518][starVLA.training.train_starvla][INFO] - Step 4511, Loss: {'train/loss': 2.7610736651695333e-05, 'train/grad_norm_pre_clip': 0.007641888689249754, 'timing/data': 0.00030780000088270754, 'timing/model': 0.3694525610044366, 'learning_rate/qwen_vl_interface': 7.315413891700363e-07, 'learning_rate/action_model': 7.315413891700362e-06, 'epoch': 0.33}) +[2026-10-03 15:29:29,178][starVLA.training.train_starvla][INFO] - Step 4512, Loss: {'train/loss': 1.0303972885594703e-05, 'train/grad_norm_pre_clip': 0.0025169013533741236, 'timing/data': 0.0003786499873967841, 'timing/model': 0.3707833100052085, 'learning_rate/qwen_vl_interface': 7.306030881097092e-07, 'learning_rate/action_model': 7.3060308810970925e-06, 'epoch': 0.33}) +[2026-10-03 15:29:29,835][starVLA.training.train_starvla][INFO] - Step 4513, Loss: {'train/loss': 0.38268882036209106, 'train/grad_norm_pre_clip': 59.355567932128906, 'timing/data': 0.00025113999436143786, 'timing/model': 0.36925425101071596, 'learning_rate/qwen_vl_interface': 7.296666448027406e-07, 'learning_rate/action_model': 7.296666448027406e-06, 'epoch': 0.33}) +[2026-10-03 15:29:30,491][starVLA.training.train_starvla][INFO] - Step 4514, Loss: {'train/loss': 0.0014429150614887476, 'train/grad_norm_pre_clip': 0.523934543132782, 'timing/data': 0.00021909001225139946, 'timing/model': 0.3694671210105298, 'learning_rate/qwen_vl_interface': 7.287320596340664e-07, 'learning_rate/action_model': 7.287320596340663e-06, 'epoch': 0.33}) +[2026-10-03 15:29:31,340][starVLA.training.train_starvla][INFO] - Step 4515, Loss: {'train/loss': 1.447595059289597e-05, 'train/grad_norm_pre_clip': 0.005463260691612959, 'timing/data': 0.0003250699955970049, 'timing/model': 0.36962481099180877, 'learning_rate/qwen_vl_interface': 7.277993329878593e-07, 'learning_rate/action_model': 7.277993329878594e-06, 'epoch': 0.33}) +[2026-10-03 15:29:31,999][starVLA.training.train_starvla][INFO] - Step 4516, Loss: {'train/loss': 0.020397936925292015, 'train/grad_norm_pre_clip': 102.07659149169922, 'timing/data': 0.00032563001150265336, 'timing/model': 0.3695321310078725, 'learning_rate/qwen_vl_interface': 7.268684652475277e-07, 'learning_rate/action_model': 7.268684652475277e-06, 'epoch': 0.33}) +[2026-10-03 15:29:32,659][starVLA.training.train_starvla][INFO] - Step 4517, Loss: {'train/loss': 0.0002344878885196522, 'train/grad_norm_pre_clip': 0.08185577392578125, 'timing/data': 0.00032369900145567954, 'timing/model': 0.3710092609981075, 'learning_rate/qwen_vl_interface': 7.259394567957169e-07, 'learning_rate/action_model': 7.259394567957169e-06, 'epoch': 0.33}) +[2026-10-03 15:29:33,293][starVLA.training.train_starvla][INFO] - Step 4518, Loss: {'train/loss': 0.04480763152241707, 'train/grad_norm_pre_clip': 112.68754577636719, 'timing/data': 0.00032454000029247254, 'timing/model': 0.3465184539963957, 'learning_rate/qwen_vl_interface': 7.250123080143066e-07, 'learning_rate/action_model': 7.2501230801430665e-06, 'epoch': 0.33}) +[2026-10-03 15:29:33,951][starVLA.training.train_starvla][INFO] - Step 4519, Loss: {'train/loss': 8.70094881975092e-05, 'train/grad_norm_pre_clip': 0.042458854615688324, 'timing/data': 0.0003035100089618936, 'timing/model': 0.3702235410019057, 'learning_rate/qwen_vl_interface': 7.240870192844127e-07, 'learning_rate/action_model': 7.240870192844127e-06, 'epoch': 0.33}) +[2026-10-03 15:29:34,802][starVLA.training.train_starvla][INFO] - Step 4520, Loss: {'train/loss': 0.012728920206427574, 'train/grad_norm_pre_clip': 19.082727432250977, 'timing/data': 0.0003424599999561906, 'timing/model': 0.3700238400051603, 'learning_rate/qwen_vl_interface': 7.231635909863858e-07, 'learning_rate/action_model': 7.2316359098638585e-06, 'epoch': 0.33}) +[2026-10-03 15:29:35,459][starVLA.training.train_starvla][INFO] - Step 4521, Loss: {'train/loss': 0.0013009171234443784, 'train/grad_norm_pre_clip': 1.4308993816375732, 'timing/data': 0.0003113999991910532, 'timing/model': 0.36926158001006115, 'learning_rate/qwen_vl_interface': 7.222420234998138e-07, 'learning_rate/action_model': 7.2224202349981385e-06, 'epoch': 0.33}) +[2026-10-03 15:29:36,116][starVLA.training.train_starvla][INFO] - Step 4522, Loss: {'train/loss': 0.00011991884821327403, 'train/grad_norm_pre_clip': 0.028334731236100197, 'timing/data': 0.0003515290009090677, 'timing/model': 0.369605910993414, 'learning_rate/qwen_vl_interface': 7.213223172035175e-07, 'learning_rate/action_model': 7.213223172035174e-06, 'epoch': 0.33}) +[2026-10-03 15:29:36,774][starVLA.training.train_starvla][INFO] - Step 4523, Loss: {'train/loss': 0.00011020589590771124, 'train/grad_norm_pre_clip': 0.08912185579538345, 'timing/data': 0.0003115599974989891, 'timing/model': 0.3697033710050164, 'learning_rate/qwen_vl_interface': 7.204044724755528e-07, 'learning_rate/action_model': 7.204044724755528e-06, 'epoch': 0.33}) +[2026-10-03 15:29:37,433][starVLA.training.train_starvla][INFO] - Step 4524, Loss: {'train/loss': 9.118761954596266e-05, 'train/grad_norm_pre_clip': 0.02719137631356716, 'timing/data': 0.00034264998976141214, 'timing/model': 0.36934009101241827, 'learning_rate/qwen_vl_interface': 7.194884896932132e-07, 'learning_rate/action_model': 7.194884896932132e-06, 'epoch': 0.33}) +[2026-10-03 15:29:38,251][starVLA.training.train_starvla][INFO] - Step 4525, Loss: {'train/loss': 0.1649179458618164, 'train/grad_norm_pre_clip': 191.94854736328125, 'timing/data': 0.0003513300034683198, 'timing/model': 0.3691747710108757, 'learning_rate/qwen_vl_interface': 7.185743692330221e-07, 'learning_rate/action_model': 7.185743692330221e-06, 'epoch': 0.33}) +[2026-10-03 15:29:38,886][starVLA.training.train_starvla][INFO] - Step 4526, Loss: {'train/loss': 1.6218902601394802e-05, 'train/grad_norm_pre_clip': 0.007732835598289967, 'timing/data': 0.00021529999503400177, 'timing/model': 0.3694085309980437, 'learning_rate/qwen_vl_interface': 7.176621114707408e-07, 'learning_rate/action_model': 7.176621114707407e-06, 'epoch': 0.33}) +[2026-10-03 15:29:39,545][starVLA.training.train_starvla][INFO] - Step 4527, Loss: {'train/loss': 0.0006912598619237542, 'train/grad_norm_pre_clip': 0.14325383305549622, 'timing/data': 0.00028761000430677086, 'timing/model': 0.36984513099014293, 'learning_rate/qwen_vl_interface': 7.167517167813642e-07, 'learning_rate/action_model': 7.167517167813642e-06, 'epoch': 0.33}) +[2026-10-03 15:29:40,205][starVLA.training.train_starvla][INFO] - Step 4528, Loss: {'train/loss': 0.00010046902752947062, 'train/grad_norm_pre_clip': 0.05549941211938858, 'timing/data': 0.0002994299866259098, 'timing/model': 0.3705121410021093, 'learning_rate/qwen_vl_interface': 7.15843185539121e-07, 'learning_rate/action_model': 7.15843185539121e-06, 'epoch': 0.33}) +[2026-10-03 15:29:40,862][starVLA.training.train_starvla][INFO] - Step 4529, Loss: {'train/loss': 6.692083843518049e-05, 'train/grad_norm_pre_clip': 0.02697017416357994, 'timing/data': 0.0003653899912023917, 'timing/model': 0.3695426909980597, 'learning_rate/qwen_vl_interface': 7.149365181174739e-07, 'learning_rate/action_model': 7.149365181174739e-06, 'epoch': 0.33}) +[2026-10-03 15:29:41,714][starVLA.training.train_starvla][INFO] - Step 4530, Loss: {'train/loss': 0.000985652208328247, 'train/grad_norm_pre_clip': 1.0216739177703857, 'timing/data': 0.00025615000049583614, 'timing/model': 0.5638731449871557, 'learning_rate/qwen_vl_interface': 7.14031714889121e-07, 'learning_rate/action_model': 7.14031714889121e-06, 'epoch': 0.33}) +[2026-10-03 15:29:42,371][starVLA.training.train_starvla][INFO] - Step 4531, Loss: {'train/loss': 0.42780354619026184, 'train/grad_norm_pre_clip': 64.94599151611328, 'timing/data': 0.0002928800095105544, 'timing/model': 0.3690572610066738, 'learning_rate/qwen_vl_interface': 7.131287762259907e-07, 'learning_rate/action_model': 7.131287762259907e-06, 'epoch': 0.33}) +[2026-10-03 15:29:43,030][starVLA.training.train_starvla][INFO] - Step 4532, Loss: {'train/loss': 1.4491132787952665e-05, 'train/grad_norm_pre_clip': 0.003211883595213294, 'timing/data': 0.00036174998967908323, 'timing/model': 0.36897581099765375, 'learning_rate/qwen_vl_interface': 7.12227702499248e-07, 'learning_rate/action_model': 7.12227702499248e-06, 'epoch': 0.33}) +[2026-10-03 15:29:43,686][starVLA.training.train_starvla][INFO] - Step 4533, Loss: {'train/loss': 0.25952234864234924, 'train/grad_norm_pre_clip': 10.979771614074707, 'timing/data': 0.00033783999970182776, 'timing/model': 0.3690411310089985, 'learning_rate/qwen_vl_interface': 7.113284940792903e-07, 'learning_rate/action_model': 7.113284940792903e-06, 'epoch': 0.33}) +[2026-10-03 15:29:44,342][starVLA.training.train_starvla][INFO] - Step 4534, Loss: {'train/loss': 0.2804027795791626, 'train/grad_norm_pre_clip': 20.686288833618164, 'timing/data': 0.00027593999402597547, 'timing/model': 0.3690567409939831, 'learning_rate/qwen_vl_interface': 7.104311513357476e-07, 'learning_rate/action_model': 7.1043115133574755e-06, 'epoch': 0.33}) +[2026-10-03 15:29:44,998][starVLA.training.train_starvla][INFO] - Step 4535, Loss: {'train/loss': 5.014182534068823e-06, 'train/grad_norm_pre_clip': 0.0009562353952787817, 'timing/data': 0.00025240999821107835, 'timing/model': 0.36893589100509416, 'learning_rate/qwen_vl_interface': 7.095356746374845e-07, 'learning_rate/action_model': 7.095356746374844e-06, 'epoch': 0.33}) +[2026-10-03 15:29:45,843][starVLA.training.train_starvla][INFO] - Step 4536, Loss: {'train/loss': 0.3131152391433716, 'train/grad_norm_pre_clip': 47.052833557128906, 'timing/data': 0.00034314999356865883, 'timing/model': 0.36906553100561723, 'learning_rate/qwen_vl_interface': 7.086420643525973e-07, 'learning_rate/action_model': 7.086420643525972e-06, 'epoch': 0.33}) +[2026-10-03 15:29:46,501][starVLA.training.train_starvla][INFO] - Step 4537, Loss: {'train/loss': 4.5469314500223845e-05, 'train/grad_norm_pre_clip': 0.025149229913949966, 'timing/data': 0.0003410299977986142, 'timing/model': 0.36918291098845657, 'learning_rate/qwen_vl_interface': 7.077503208484151e-07, 'learning_rate/action_model': 7.0775032084841515e-06, 'epoch': 0.33}) +[2026-10-03 15:29:47,157][starVLA.training.train_starvla][INFO] - Step 4538, Loss: {'train/loss': 2.1113364709890448e-05, 'train/grad_norm_pre_clip': 0.007977195084095001, 'timing/data': 0.00034505000803619623, 'timing/model': 0.3687637410039315, 'learning_rate/qwen_vl_interface': 7.068604444915003e-07, 'learning_rate/action_model': 7.068604444915003e-06, 'epoch': 0.33}) +[2026-10-03 15:29:47,815][starVLA.training.train_starvla][INFO] - Step 4539, Loss: {'train/loss': 4.797644942300394e-05, 'train/grad_norm_pre_clip': 0.01499080378562212, 'timing/data': 0.0002065699954982847, 'timing/model': 0.3692932709964225, 'learning_rate/qwen_vl_interface': 7.059724356476475e-07, 'learning_rate/action_model': 7.059724356476474e-06, 'epoch': 0.33}) +[2026-10-03 15:29:48,474][starVLA.training.train_starvla][INFO] - Step 4540, Loss: {'train/loss': 2.6796287784236483e-05, 'train/grad_norm_pre_clip': 0.015460668131709099, 'timing/data': 0.00033428000460844487, 'timing/model': 0.36945145099889487, 'learning_rate/qwen_vl_interface': 7.050862946818836e-07, 'learning_rate/action_model': 7.050862946818836e-06, 'epoch': 0.33}) +[2026-10-03 15:29:49,327][starVLA.training.train_starvla][INFO] - Step 4541, Loss: {'train/loss': 0.3146309554576874, 'train/grad_norm_pre_clip': 63.60992431640625, 'timing/data': 0.00032063000253401697, 'timing/model': 0.36858760099858046, 'learning_rate/qwen_vl_interface': 7.042020219584673e-07, 'learning_rate/action_model': 7.042020219584673e-06, 'epoch': 0.33}) +[2026-10-03 15:29:49,984][starVLA.training.train_starvla][INFO] - Step 4542, Loss: {'train/loss': 4.055727185914293e-05, 'train/grad_norm_pre_clip': 0.022003494203090668, 'timing/data': 0.00034456000139471143, 'timing/model': 0.3690593009960139, 'learning_rate/qwen_vl_interface': 7.0331961784089e-07, 'learning_rate/action_model': 7.033196178408899e-06, 'epoch': 0.33}) +[2026-10-03 15:29:50,642][starVLA.training.train_starvla][INFO] - Step 4543, Loss: {'train/loss': 0.27912846207618713, 'train/grad_norm_pre_clip': 49.336402893066406, 'timing/data': 0.0003033600078197196, 'timing/model': 0.3692041499889456, 'learning_rate/qwen_vl_interface': 7.024390826918748e-07, 'learning_rate/action_model': 7.024390826918747e-06, 'epoch': 0.33}) +[2026-10-03 15:29:51,299][starVLA.training.train_starvla][INFO] - Step 4544, Loss: {'train/loss': 5.237710411165608e-06, 'train/grad_norm_pre_clip': 0.0021191982086747885, 'timing/data': 0.0003475700068520382, 'timing/model': 0.36857373200473376, 'learning_rate/qwen_vl_interface': 7.01560416873376e-07, 'learning_rate/action_model': 7.015604168733759e-06, 'epoch': 0.33}) +[2026-10-03 15:29:51,956][starVLA.training.train_starvla][INFO] - Step 4545, Loss: {'train/loss': 0.14391373097896576, 'train/grad_norm_pre_clip': 75.10819244384766, 'timing/data': 0.00030913999944459647, 'timing/model': 0.3696573510096641, 'learning_rate/qwen_vl_interface': 7.006836207465803e-07, 'learning_rate/action_model': 7.006836207465803e-06, 'epoch': 0.33}) +[2026-10-03 15:29:52,807][starVLA.training.train_starvla][INFO] - Step 4546, Loss: {'train/loss': 4.916034231428057e-05, 'train/grad_norm_pre_clip': 0.027424221858382225, 'timing/data': 0.00041364999196957797, 'timing/model': 0.5633710849942872, 'learning_rate/qwen_vl_interface': 6.998086946719058e-07, 'learning_rate/action_model': 6.998086946719057e-06, 'epoch': 0.33}) +[2026-10-03 15:29:53,464][starVLA.training.train_starvla][INFO] - Step 4547, Loss: {'train/loss': 5.787634654552676e-05, 'train/grad_norm_pre_clip': 0.028093550354242325, 'timing/data': 0.0002462999982526526, 'timing/model': 0.36928734100365546, 'learning_rate/qwen_vl_interface': 6.989356390090008e-07, 'learning_rate/action_model': 6.989356390090007e-06, 'epoch': 0.33}) +[2026-10-03 15:29:54,122][starVLA.training.train_starvla][INFO] - Step 4548, Loss: {'train/loss': 0.0003418528358452022, 'train/grad_norm_pre_clip': 0.3205106258392334, 'timing/data': 0.0003690000012284145, 'timing/model': 0.36908844098798, 'learning_rate/qwen_vl_interface': 6.980644541167459e-07, 'learning_rate/action_model': 6.980644541167458e-06, 'epoch': 0.33}) +[2026-10-03 15:29:54,778][starVLA.training.train_starvla][INFO] - Step 4549, Loss: {'train/loss': 0.3814164996147156, 'train/grad_norm_pre_clip': 43.089683532714844, 'timing/data': 0.00020784999651368707, 'timing/model': 0.36922241101274267, 'learning_rate/qwen_vl_interface': 6.971951403532524e-07, 'learning_rate/action_model': 6.971951403532523e-06, 'epoch': 0.33}) +[2026-10-03 15:29:55,435][starVLA.training.train_starvla][INFO] - Step 4550, Loss: {'train/loss': 0.0012672431766986847, 'train/grad_norm_pre_clip': 0.35820460319519043, 'timing/data': 0.00020614999812096357, 'timing/model': 0.36949764100427274, 'learning_rate/qwen_vl_interface': 6.963276980758622e-07, 'learning_rate/action_model': 6.963276980758622e-06, 'epoch': 0.33}) +[2026-10-03 15:29:56,070][starVLA.training.train_starvla][INFO] - Step 4551, Loss: {'train/loss': 0.4643588364124298, 'train/grad_norm_pre_clip': 33.068363189697266, 'timing/data': 0.0002597699931357056, 'timing/model': 0.36949892100528814, 'learning_rate/qwen_vl_interface': 6.954621276411485e-07, 'learning_rate/action_model': 6.954621276411484e-06, 'epoch': 0.33}) +[2026-10-03 15:29:56,927][starVLA.training.train_starvla][INFO] - Step 4552, Loss: {'train/loss': 3.826237661996856e-05, 'train/grad_norm_pre_clip': 0.010171301662921906, 'timing/data': 0.00028111001302022487, 'timing/model': 0.37476538099872414, 'learning_rate/qwen_vl_interface': 6.945984294049136e-07, 'learning_rate/action_model': 6.9459842940491364e-06, 'epoch': 0.33}) +[2026-10-03 15:29:57,584][starVLA.training.train_starvla][INFO] - Step 4553, Loss: {'train/loss': 0.0005061601405031979, 'train/grad_norm_pre_clip': 0.43403056263923645, 'timing/data': 0.0002062299899989739, 'timing/model': 0.36994118099391926, 'learning_rate/qwen_vl_interface': 6.937366037221919e-07, 'learning_rate/action_model': 6.937366037221918e-06, 'epoch': 0.33}) +[2026-10-03 15:29:58,240][starVLA.training.train_starvla][INFO] - Step 4554, Loss: {'train/loss': 6.3478228184976615e-06, 'train/grad_norm_pre_clip': 0.0020327065140008926, 'timing/data': 0.0003520900063449517, 'timing/model': 0.36916619099793024, 'learning_rate/qwen_vl_interface': 6.928766509472471e-07, 'learning_rate/action_model': 6.92876650947247e-06, 'epoch': 0.33}) +[2026-10-03 15:29:58,898][starVLA.training.train_starvla][INFO] - Step 4555, Loss: {'train/loss': 7.500396895920858e-05, 'train/grad_norm_pre_clip': 0.04916253313422203, 'timing/data': 0.0002158699935534969, 'timing/model': 0.3698610599967651, 'learning_rate/qwen_vl_interface': 6.920185714335737e-07, 'learning_rate/action_model': 6.9201857143357375e-06, 'epoch': 0.33}) +[2026-10-03 15:29:59,557][starVLA.training.train_starvla][INFO] - Step 4556, Loss: {'train/loss': 2.7036163373850286e-05, 'train/grad_norm_pre_clip': 0.009510626085102558, 'timing/data': 0.0003377499961061403, 'timing/model': 0.3700092610088177, 'learning_rate/qwen_vl_interface': 6.911623655338959e-07, 'learning_rate/action_model': 6.911623655338959e-06, 'epoch': 0.33}) +[2026-10-03 15:30:00,408][starVLA.training.train_starvla][INFO] - Step 4557, Loss: {'train/loss': 0.0007129076402634382, 'train/grad_norm_pre_clip': 1.1499258279800415, 'timing/data': 0.0002508000034140423, 'timing/model': 0.562735444997088, 'learning_rate/qwen_vl_interface': 6.903080336001666e-07, 'learning_rate/action_model': 6.903080336001666e-06, 'epoch': 0.33}) +[2026-10-03 15:30:01,066][starVLA.training.train_starvla][INFO] - Step 4558, Loss: {'train/loss': 0.0011845022672787309, 'train/grad_norm_pre_clip': 0.37351661920547485, 'timing/data': 0.00020966000738553703, 'timing/model': 0.3698254810005892, 'learning_rate/qwen_vl_interface': 6.894555759835707e-07, 'learning_rate/action_model': 6.894555759835707e-06, 'epoch': 0.33}) +[2026-10-03 15:30:01,725][starVLA.training.train_starvla][INFO] - Step 4559, Loss: {'train/loss': 2.8347843908704817e-05, 'train/grad_norm_pre_clip': 0.013032298535108566, 'timing/data': 0.00029463000828400254, 'timing/model': 0.3699226109893061, 'learning_rate/qwen_vl_interface': 6.886049930345204e-07, 'learning_rate/action_model': 6.8860499303452035e-06, 'epoch': 0.33}) +[2026-10-03 15:30:02,384][starVLA.training.train_starvla][INFO] - Step 4560, Loss: {'train/loss': 7.835475844331086e-05, 'train/grad_norm_pre_clip': 0.04476028308272362, 'timing/data': 0.00029175999225117266, 'timing/model': 0.36991130100796, 'learning_rate/qwen_vl_interface': 6.877562851026584e-07, 'learning_rate/action_model': 6.877562851026584e-06, 'epoch': 0.33}) +[2026-10-03 15:30:03,019][starVLA.training.train_starvla][INFO] - Step 4561, Loss: {'train/loss': 0.004433807916939259, 'train/grad_norm_pre_clip': 3.3262550830841064, 'timing/data': 0.00034174999746028334, 'timing/model': 0.3696789310051827, 'learning_rate/qwen_vl_interface': 6.869094525368572e-07, 'learning_rate/action_model': 6.8690945253685715e-06, 'epoch': 0.33}) +[2026-10-03 15:30:03,876][starVLA.training.train_starvla][INFO] - Step 4562, Loss: {'train/loss': 0.00014006537094246596, 'train/grad_norm_pre_clip': 0.050636060535907745, 'timing/data': 0.00020707999647129327, 'timing/model': 0.5700448139978107, 'learning_rate/qwen_vl_interface': 6.860644956852168e-07, 'learning_rate/action_model': 6.860644956852168e-06, 'epoch': 0.33}) +[2026-10-03 15:30:04,533][starVLA.training.train_starvla][INFO] - Step 4563, Loss: {'train/loss': 0.10591869801282883, 'train/grad_norm_pre_clip': 180.44602966308594, 'timing/data': 0.00025270000332966447, 'timing/model': 0.36945148099039216, 'learning_rate/qwen_vl_interface': 6.852214148950667e-07, 'learning_rate/action_model': 6.852214148950667e-06, 'epoch': 0.33}) +[2026-10-03 15:30:05,192][starVLA.training.train_starvla][INFO] - Step 4564, Loss: {'train/loss': 0.2931610643863678, 'train/grad_norm_pre_clip': 80.56047821044922, 'timing/data': 0.0002652300026966259, 'timing/model': 0.3695530620025238, 'learning_rate/qwen_vl_interface': 6.843802105129667e-07, 'learning_rate/action_model': 6.843802105129666e-06, 'epoch': 0.33}) +[2026-10-03 15:30:05,851][starVLA.training.train_starvla][INFO] - Step 4565, Loss: {'train/loss': 0.00020458691869862378, 'train/grad_norm_pre_clip': 0.07253588736057281, 'timing/data': 0.0002116999967256561, 'timing/model': 0.3703383510001004, 'learning_rate/qwen_vl_interface': 6.835408828847035e-07, 'learning_rate/action_model': 6.8354088288470356e-06, 'epoch': 0.33}) +[2026-10-03 15:30:06,508][starVLA.training.train_starvla][INFO] - Step 4566, Loss: {'train/loss': 0.0009084743214771152, 'train/grad_norm_pre_clip': 0.19707085192203522, 'timing/data': 0.00032959999225568026, 'timing/model': 0.3695475209970027, 'learning_rate/qwen_vl_interface': 6.827034323552929e-07, 'learning_rate/action_model': 6.827034323552929e-06, 'epoch': 0.33}) +[2026-10-03 15:30:07,167][starVLA.training.train_starvla][INFO] - Step 4567, Loss: {'train/loss': 0.0007923523662611842, 'train/grad_norm_pre_clip': 0.12420392781496048, 'timing/data': 0.00034455000422894955, 'timing/model': 0.3690691010124283, 'learning_rate/qwen_vl_interface': 6.818678592689794e-07, 'learning_rate/action_model': 6.818678592689794e-06, 'epoch': 0.33}) +[2026-10-03 15:30:08,020][starVLA.training.train_starvla][INFO] - Step 4568, Loss: {'train/loss': 0.4582176208496094, 'train/grad_norm_pre_clip': 99.69532012939453, 'timing/data': 0.0002733389992499724, 'timing/model': 0.3699825320072705, 'learning_rate/qwen_vl_interface': 6.810341639692357e-07, 'learning_rate/action_model': 6.810341639692357e-06, 'epoch': 0.33}) +[2026-10-03 15:30:08,677][starVLA.training.train_starvla][INFO] - Step 4569, Loss: {'train/loss': 9.408945334143937e-05, 'train/grad_norm_pre_clip': 0.03735321760177612, 'timing/data': 0.00021104000916238874, 'timing/model': 0.3701469410007121, 'learning_rate/qwen_vl_interface': 6.802023467987631e-07, 'learning_rate/action_model': 6.802023467987631e-06, 'epoch': 0.33}) +[2026-10-03 15:30:09,335][starVLA.training.train_starvla][INFO] - Step 4570, Loss: {'train/loss': 0.15233363211154938, 'train/grad_norm_pre_clip': 99.53341674804688, 'timing/data': 0.0003190199931850657, 'timing/model': 0.37016625099931844, 'learning_rate/qwen_vl_interface': 6.793724080994884e-07, 'learning_rate/action_model': 6.793724080994884e-06, 'epoch': 0.33}) +[2026-10-03 15:30:09,994][starVLA.training.train_starvla][INFO] - Step 4571, Loss: {'train/loss': 0.0001498553465353325, 'train/grad_norm_pre_clip': 0.150558203458786, 'timing/data': 0.00033925999014172703, 'timing/model': 0.3699702209996758, 'learning_rate/qwen_vl_interface': 6.785443482125693e-07, 'learning_rate/action_model': 6.7854434821256915e-06, 'epoch': 0.33}) +[2026-10-03 15:30:10,654][starVLA.training.train_starvla][INFO] - Step 4572, Loss: {'train/loss': 1.9475235603749752e-05, 'train/grad_norm_pre_clip': 0.005723345093429089, 'timing/data': 0.0003355000080773607, 'timing/model': 0.36983070099086035, 'learning_rate/qwen_vl_interface': 6.777181674783895e-07, 'learning_rate/action_model': 6.777181674783895e-06, 'epoch': 0.33}) +[2026-10-03 15:30:11,502][starVLA.training.train_starvla][INFO] - Step 4573, Loss: {'train/loss': 5.632603915728396e-06, 'train/grad_norm_pre_clip': 0.001309265848249197, 'timing/data': 0.0002583299938123673, 'timing/model': 0.5599458559881896, 'learning_rate/qwen_vl_interface': 6.76893866236561e-07, 'learning_rate/action_model': 6.7689386623656094e-06, 'epoch': 0.33}) +[2026-10-03 15:30:12,160][starVLA.training.train_starvla][INFO] - Step 4574, Loss: {'train/loss': 0.00017263107292819768, 'train/grad_norm_pre_clip': 0.30999094247817993, 'timing/data': 0.00035524000122677535, 'timing/model': 0.36978099000407383, 'learning_rate/qwen_vl_interface': 6.76071444825923e-07, 'learning_rate/action_model': 6.76071444825923e-06, 'epoch': 0.33}) +[2026-10-03 15:30:12,820][starVLA.training.train_starvla][INFO] - Step 4575, Loss: {'train/loss': 3.9417780499206856e-05, 'train/grad_norm_pre_clip': 0.014066868461668491, 'timing/data': 0.00029819000337738544, 'timing/model': 0.37113281099300366, 'learning_rate/qwen_vl_interface': 6.752509035845413e-07, 'learning_rate/action_model': 6.7525090358454126e-06, 'epoch': 0.33}) +[2026-10-03 15:30:13,480][starVLA.training.train_starvla][INFO] - Step 4576, Loss: {'train/loss': 0.003362291492521763, 'train/grad_norm_pre_clip': 4.956920623779297, 'timing/data': 0.00030582000908907503, 'timing/model': 0.37039588099287357, 'learning_rate/qwen_vl_interface': 6.744322428497095e-07, 'learning_rate/action_model': 6.744322428497095e-06, 'epoch': 0.33}) +[2026-10-03 15:30:14,138][starVLA.training.train_starvla][INFO] - Step 4577, Loss: {'train/loss': 7.957157322380226e-06, 'train/grad_norm_pre_clip': 0.0016563785029575229, 'timing/data': 0.00021263999224174768, 'timing/model': 0.3700878109957557, 'learning_rate/qwen_vl_interface': 6.736154629579487e-07, 'learning_rate/action_model': 6.736154629579487e-06, 'epoch': 0.34}) +[2026-10-03 15:30:14,994][starVLA.training.train_starvla][INFO] - Step 4578, Loss: {'train/loss': 0.0011309677502140403, 'train/grad_norm_pre_clip': 3.753680944442749, 'timing/data': 0.00025193000328727067, 'timing/model': 0.5680712449975545, 'learning_rate/qwen_vl_interface': 6.728005642450055e-07, 'learning_rate/action_model': 6.7280056424500554e-06, 'epoch': 0.34}) +[2026-10-03 15:30:15,652][starVLA.training.train_starvla][INFO] - Step 4579, Loss: {'train/loss': 1.8878454284276813e-05, 'train/grad_norm_pre_clip': 0.012258682399988174, 'timing/data': 0.0002896000078180805, 'timing/model': 0.3699142810073681, 'learning_rate/qwen_vl_interface': 6.719875470458553e-07, 'learning_rate/action_model': 6.719875470458553e-06, 'epoch': 0.34}) +[2026-10-03 15:30:16,313][starVLA.training.train_starvla][INFO] - Step 4580, Loss: {'train/loss': 0.00013844840577803552, 'train/grad_norm_pre_clip': 0.05834265053272247, 'timing/data': 0.00021950999507680535, 'timing/model': 0.3710951209941413, 'learning_rate/qwen_vl_interface': 6.711764116946972e-07, 'learning_rate/action_model': 6.7117641169469715e-06, 'epoch': 0.34}) +[2026-10-03 15:30:16,951][starVLA.training.train_starvla][INFO] - Step 4581, Loss: {'train/loss': 0.4588879942893982, 'train/grad_norm_pre_clip': 41.343177795410156, 'timing/data': 0.00029987000743858516, 'timing/model': 0.3716579509928124, 'learning_rate/qwen_vl_interface': 6.703671585249587e-07, 'learning_rate/action_model': 6.703671585249586e-06, 'epoch': 0.34}) +[2026-10-03 15:30:17,586][starVLA.training.train_starvla][INFO] - Step 4582, Loss: {'train/loss': 0.0025862541515380144, 'train/grad_norm_pre_clip': 6.98518705368042, 'timing/data': 0.0003405500028748065, 'timing/model': 0.3465682139940327, 'learning_rate/qwen_vl_interface': 6.69559787869294e-07, 'learning_rate/action_model': 6.69559787869294e-06, 'epoch': 0.34}) +[2026-10-03 15:30:18,455][starVLA.training.train_starvla][INFO] - Step 4583, Loss: {'train/loss': 4.313868430472212e-06, 'train/grad_norm_pre_clip': 0.0006489211227744818, 'timing/data': 0.0003392400103621185, 'timing/model': 0.5798631030047545, 'learning_rate/qwen_vl_interface': 6.68754300059582e-07, 'learning_rate/action_model': 6.687543000595819e-06, 'epoch': 0.34}) +[2026-10-03 15:30:19,119][starVLA.training.train_starvla][INFO] - Step 4584, Loss: {'train/loss': 0.3122658431529999, 'train/grad_norm_pre_clip': 119.76490020751953, 'timing/data': 0.00035813900467474014, 'timing/model': 0.3728394310019212, 'learning_rate/qwen_vl_interface': 6.67950695426929e-07, 'learning_rate/action_model': 6.679506954269291e-06, 'epoch': 0.34}) +[2026-10-03 15:30:19,777][starVLA.training.train_starvla][INFO] - Step 4585, Loss: {'train/loss': 0.23138277232646942, 'train/grad_norm_pre_clip': 84.51200103759766, 'timing/data': 0.00034223000693600625, 'timing/model': 0.36994918099662755, 'learning_rate/qwen_vl_interface': 6.671489743016651e-07, 'learning_rate/action_model': 6.671489743016651e-06, 'epoch': 0.34}) +[2026-10-03 15:30:20,435][starVLA.training.train_starvla][INFO] - Step 4586, Loss: {'train/loss': 2.3355798475677148e-05, 'train/grad_norm_pre_clip': 0.010077490471303463, 'timing/data': 0.0003236600023228675, 'timing/model': 0.36986552098824177, 'learning_rate/qwen_vl_interface': 6.663491370133487e-07, 'learning_rate/action_model': 6.663491370133487e-06, 'epoch': 0.34}) +[2026-10-03 15:30:21,094][starVLA.training.train_starvla][INFO] - Step 4587, Loss: {'train/loss': 2.588208553788718e-05, 'train/grad_norm_pre_clip': 0.00670276116579771, 'timing/data': 0.00023877000785432756, 'timing/model': 0.3702952710009413, 'learning_rate/qwen_vl_interface': 6.655511838907629e-07, 'learning_rate/action_model': 6.655511838907629e-06, 'epoch': 0.34}) +[2026-10-03 15:30:21,754][starVLA.training.train_starvla][INFO] - Step 4588, Loss: {'train/loss': 3.874659523717128e-05, 'train/grad_norm_pre_clip': 0.020248547196388245, 'timing/data': 0.00021641999774146825, 'timing/model': 0.37037981099274475, 'learning_rate/qwen_vl_interface': 6.647551152619149e-07, 'learning_rate/action_model': 6.647551152619148e-06, 'epoch': 0.34}) +[2026-10-03 15:30:22,603][starVLA.training.train_starvla][INFO] - Step 4589, Loss: {'train/loss': 1.6785341358627193e-05, 'train/grad_norm_pre_clip': 0.004847417119890451, 'timing/data': 0.00022005000209901482, 'timing/model': 0.3716482510062633, 'learning_rate/qwen_vl_interface': 6.639609314540396e-07, 'learning_rate/action_model': 6.639609314540396e-06, 'epoch': 0.34}) +[2026-10-03 15:30:23,262][starVLA.training.train_starvla][INFO] - Step 4590, Loss: {'train/loss': 2.6426114345667884e-05, 'train/grad_norm_pre_clip': 0.00766599690541625, 'timing/data': 0.00021363000269047916, 'timing/model': 0.3706105509918416, 'learning_rate/qwen_vl_interface': 6.631686327935954e-07, 'learning_rate/action_model': 6.631686327935954e-06, 'epoch': 0.34}) +[2026-10-03 15:30:23,924][starVLA.training.train_starvla][INFO] - Step 4591, Loss: {'train/loss': 2.1523985196836293e-05, 'train/grad_norm_pre_clip': 0.005694322753697634, 'timing/data': 0.0003309900057502091, 'timing/model': 0.3716847699979553, 'learning_rate/qwen_vl_interface': 6.623782196062665e-07, 'learning_rate/action_model': 6.623782196062664e-06, 'epoch': 0.34}) +[2026-10-03 15:30:24,560][starVLA.training.train_starvla][INFO] - Step 4592, Loss: {'train/loss': 0.0012634770246222615, 'train/grad_norm_pre_clip': 0.4651222825050354, 'timing/data': 0.00020902999676764011, 'timing/model': 0.3700063209980726, 'learning_rate/qwen_vl_interface': 6.615896922169622e-07, 'learning_rate/action_model': 6.615896922169622e-06, 'epoch': 0.34}) +[2026-10-03 15:30:25,219][starVLA.training.train_starvla][INFO] - Step 4593, Loss: {'train/loss': 0.00022704557341057807, 'train/grad_norm_pre_clip': 0.14928393065929413, 'timing/data': 0.0003381099959369749, 'timing/model': 0.3706165910116397, 'learning_rate/qwen_vl_interface': 6.608030509498156e-07, 'learning_rate/action_model': 6.608030509498156e-06, 'epoch': 0.34}) +[2026-10-03 15:30:26,072][starVLA.training.train_starvla][INFO] - Step 4594, Loss: {'train/loss': 0.4514903724193573, 'train/grad_norm_pre_clip': 34.345523834228516, 'timing/data': 0.0002170000079786405, 'timing/model': 0.37086519099830184, 'learning_rate/qwen_vl_interface': 6.600182961281865e-07, 'learning_rate/action_model': 6.600182961281864e-06, 'epoch': 0.34}) +[2026-10-03 15:30:26,709][starVLA.training.train_starvla][INFO] - Step 4595, Loss: {'train/loss': 0.32215723395347595, 'train/grad_norm_pre_clip': 37.4195442199707, 'timing/data': 0.00026927998987957835, 'timing/model': 0.37042010101140477, 'learning_rate/qwen_vl_interface': 6.592354280746561e-07, 'learning_rate/action_model': 6.5923542807465595e-06, 'epoch': 0.34}) +[2026-10-03 15:30:27,369][starVLA.training.train_starvla][INFO] - Step 4596, Loss: {'train/loss': 9.116373985307291e-05, 'train/grad_norm_pre_clip': 0.035035621374845505, 'timing/data': 0.0003216299955965951, 'timing/model': 0.37057156099763233, 'learning_rate/qwen_vl_interface': 6.584544471110335e-07, 'learning_rate/action_model': 6.584544471110335e-06, 'epoch': 0.34}) +[2026-10-03 15:30:28,006][starVLA.training.train_starvla][INFO] - Step 4597, Loss: {'train/loss': 6.607826799154282e-05, 'train/grad_norm_pre_clip': 0.03835957124829292, 'timing/data': 0.00030000999686308205, 'timing/model': 0.3474371639895253, 'learning_rate/qwen_vl_interface': 6.576753535583495e-07, 'learning_rate/action_model': 6.576753535583495e-06, 'epoch': 0.34}) +[2026-10-03 15:30:28,665][starVLA.training.train_starvla][INFO] - Step 4598, Loss: {'train/loss': 2.661860708030872e-05, 'train/grad_norm_pre_clip': 0.0163727980107069, 'timing/data': 0.0002577599952928722, 'timing/model': 0.37031719999504276, 'learning_rate/qwen_vl_interface': 6.5689814773686e-07, 'learning_rate/action_model': 6.568981477368599e-06, 'epoch': 0.34}) +[2026-10-03 15:30:29,526][starVLA.training.train_starvla][INFO] - Step 4599, Loss: {'train/loss': 3.678656503325328e-05, 'train/grad_norm_pre_clip': 0.013794795610010624, 'timing/data': 0.0003335499932290986, 'timing/model': 0.37156350098666735, 'learning_rate/qwen_vl_interface': 6.56122829966046e-07, 'learning_rate/action_model': 6.56122829966046e-06, 'epoch': 0.34}) +[2026-10-03 15:30:30,186][starVLA.training.train_starvla][INFO] - Step 4600, Loss: {'train/loss': 1.4237335562938824e-05, 'train/grad_norm_pre_clip': 0.011275692842900753, 'timing/data': 0.0003313100023660809, 'timing/model': 0.3707778509997297, 'learning_rate/qwen_vl_interface': 6.553494005646101e-07, 'learning_rate/action_model': 6.553494005646101e-06, 'epoch': 0.34}) +[2026-10-03 15:30:30,844][starVLA.training.train_starvla][INFO] - Step 4601, Loss: {'train/loss': 0.00023155994131229818, 'train/grad_norm_pre_clip': 0.12855318188667297, 'timing/data': 0.00021269998978823423, 'timing/model': 0.3698376809916226, 'learning_rate/qwen_vl_interface': 6.545778598504807e-07, 'learning_rate/action_model': 6.545778598504807e-06, 'epoch': 0.34}) +[2026-10-03 15:30:31,503][starVLA.training.train_starvla][INFO] - Step 4602, Loss: {'train/loss': 5.246317232376896e-05, 'train/grad_norm_pre_clip': 0.02360222302377224, 'timing/data': 0.0002081300044665113, 'timing/model': 0.370901590998983, 'learning_rate/qwen_vl_interface': 6.538082081408093e-07, 'learning_rate/action_model': 6.538082081408093e-06, 'epoch': 0.34}) +[2026-10-03 15:30:32,163][starVLA.training.train_starvla][INFO] - Step 4603, Loss: {'train/loss': 3.456663398537785e-05, 'train/grad_norm_pre_clip': 0.013218157924711704, 'timing/data': 0.0003106700023636222, 'timing/model': 0.37059884099289775, 'learning_rate/qwen_vl_interface': 6.530404457519697e-07, 'learning_rate/action_model': 6.530404457519697e-06, 'epoch': 0.34}) +[2026-10-03 15:30:33,020][starVLA.training.train_starvla][INFO] - Step 4604, Loss: {'train/loss': 0.3736690878868103, 'train/grad_norm_pre_clip': 177.37646484375, 'timing/data': 0.00027644999499898404, 'timing/model': 0.568543483997928, 'learning_rate/qwen_vl_interface': 6.522745729995614e-07, 'learning_rate/action_model': 6.522745729995613e-06, 'epoch': 0.34}) +[2026-10-03 15:30:33,679][starVLA.training.train_starvla][INFO] - Step 4605, Loss: {'train/loss': 7.359468145295978e-05, 'train/grad_norm_pre_clip': 0.03788645192980766, 'timing/data': 0.00021405000006780028, 'timing/model': 0.3705285400064895, 'learning_rate/qwen_vl_interface': 6.515105901984037e-07, 'learning_rate/action_model': 6.515105901984037e-06, 'epoch': 0.34}) +[2026-10-03 15:30:34,316][starVLA.training.train_starvla][INFO] - Step 4606, Loss: {'train/loss': 0.05264831334352493, 'train/grad_norm_pre_clip': 82.46575927734375, 'timing/data': 0.00022325999452732503, 'timing/model': 0.370227850988158, 'learning_rate/qwen_vl_interface': 6.50748497662543e-07, 'learning_rate/action_model': 6.507484976625429e-06, 'epoch': 0.34}) +[2026-10-03 15:30:34,975][starVLA.training.train_starvla][INFO] - Step 4607, Loss: {'train/loss': 0.39606422185897827, 'train/grad_norm_pre_clip': 209.0536651611328, 'timing/data': 0.00031953999132383615, 'timing/model': 0.3701551309932256, 'learning_rate/qwen_vl_interface': 6.499882957052459e-07, 'learning_rate/action_model': 6.4998829570524586e-06, 'epoch': 0.34}) +[2026-10-03 15:30:35,634][starVLA.training.train_starvla][INFO] - Step 4608, Loss: {'train/loss': 6.262605893425643e-05, 'train/grad_norm_pre_clip': 0.050605423748493195, 'timing/data': 0.0002475700021022931, 'timing/model': 0.3707245309924474, 'learning_rate/qwen_vl_interface': 6.492299846390024e-07, 'learning_rate/action_model': 6.492299846390023e-06, 'epoch': 0.34}) +[2026-10-03 15:30:36,481][starVLA.training.train_starvla][INFO] - Step 4609, Loss: {'train/loss': 9.672784653957933e-05, 'train/grad_norm_pre_clip': 0.07101470232009888, 'timing/data': 0.0003132999991066754, 'timing/model': 0.5581767859985121, 'learning_rate/qwen_vl_interface': 6.484735647755259e-07, 'learning_rate/action_model': 6.48473564775526e-06, 'epoch': 0.34}) +[2026-10-03 15:30:37,140][starVLA.training.train_starvla][INFO] - Step 4610, Loss: {'train/loss': 0.3126353621482849, 'train/grad_norm_pre_clip': 146.22000122070312, 'timing/data': 0.00021071999799460173, 'timing/model': 0.37089509099314455, 'learning_rate/qwen_vl_interface': 6.477190364257524e-07, 'learning_rate/action_model': 6.477190364257524e-06, 'epoch': 0.34}) +[2026-10-03 15:30:37,776][starVLA.training.train_starvla][INFO] - Step 4611, Loss: {'train/loss': 0.0015546104405075312, 'train/grad_norm_pre_clip': 0.7897660732269287, 'timing/data': 0.00020986999152228236, 'timing/model': 0.34744753398990724, 'learning_rate/qwen_vl_interface': 6.469663998998393e-07, 'learning_rate/action_model': 6.469663998998393e-06, 'epoch': 0.34}) +[2026-10-03 15:30:38,436][starVLA.training.train_starvla][INFO] - Step 4612, Loss: {'train/loss': 0.0009293875191360712, 'train/grad_norm_pre_clip': 0.3182528614997864, 'timing/data': 0.00021148000087123364, 'timing/model': 0.3709346010000445, 'learning_rate/qwen_vl_interface': 6.462156555071681e-07, 'learning_rate/action_model': 6.46215655507168e-06, 'epoch': 0.34}) +[2026-10-03 15:30:39,072][starVLA.training.train_starvla][INFO] - Step 4613, Loss: {'train/loss': 2.930777736764867e-05, 'train/grad_norm_pre_clip': 0.010571892373263836, 'timing/data': 0.00029892000020481646, 'timing/model': 0.3475802040047711, 'learning_rate/qwen_vl_interface': 6.454668035563401e-07, 'learning_rate/action_model': 6.454668035563401e-06, 'epoch': 0.34}) +[2026-10-03 15:30:39,731][starVLA.training.train_starvla][INFO] - Step 4614, Loss: {'train/loss': 0.0019996820483356714, 'train/grad_norm_pre_clip': 3.278841018676758, 'timing/data': 0.00021495000692084432, 'timing/model': 0.3701343610009644, 'learning_rate/qwen_vl_interface': 6.447198443551809e-07, 'learning_rate/action_model': 6.4471984435518075e-06, 'epoch': 0.34}) +[2026-10-03 15:30:40,585][starVLA.training.train_starvla][INFO] - Step 4615, Loss: {'train/loss': 3.344227297930047e-05, 'train/grad_norm_pre_clip': 0.011304753832519054, 'timing/data': 0.0003499000013107434, 'timing/model': 0.36974500100768637, 'learning_rate/qwen_vl_interface': 6.439747782107373e-07, 'learning_rate/action_model': 6.4397477821073735e-06, 'epoch': 0.34}) +[2026-10-03 15:30:41,222][starVLA.training.train_starvla][INFO] - Step 4616, Loss: {'train/loss': 2.290951852046419e-05, 'train/grad_norm_pre_clip': 0.008503585122525692, 'timing/data': 0.00020936000510118902, 'timing/model': 0.37100925100094173, 'learning_rate/qwen_vl_interface': 6.432316054292779e-07, 'learning_rate/action_model': 6.43231605429278e-06, 'epoch': 0.34}) +[2026-10-03 15:30:41,881][starVLA.training.train_starvla][INFO] - Step 4617, Loss: {'train/loss': 1.4342182112159207e-05, 'train/grad_norm_pre_clip': 0.002454390749335289, 'timing/data': 0.0003879799915011972, 'timing/model': 0.370335931002046, 'learning_rate/qwen_vl_interface': 6.424903263162924e-07, 'learning_rate/action_model': 6.424903263162923e-06, 'epoch': 0.34}) +[2026-10-03 15:30:42,540][starVLA.training.train_starvla][INFO] - Step 4618, Loss: {'train/loss': 1.0318928616470657e-05, 'train/grad_norm_pre_clip': 0.0027984841726720333, 'timing/data': 0.0002146199985872954, 'timing/model': 0.3702600410033483, 'learning_rate/qwen_vl_interface': 6.417509411764924e-07, 'learning_rate/action_model': 6.4175094117649244e-06, 'epoch': 0.34}) +[2026-10-03 15:30:43,200][starVLA.training.train_starvla][INFO] - Step 4619, Loss: {'train/loss': 4.670152338803746e-05, 'train/grad_norm_pre_clip': 0.02041773684322834, 'timing/data': 0.0003120600013062358, 'timing/model': 0.37052159100130666, 'learning_rate/qwen_vl_interface': 6.410134503138122e-07, 'learning_rate/action_model': 6.4101345031381224e-06, 'epoch': 0.34}) +[2026-10-03 15:30:44,058][starVLA.training.train_starvla][INFO] - Step 4620, Loss: {'train/loss': 0.0012977353762835264, 'train/grad_norm_pre_clip': 4.7650465965271, 'timing/data': 0.0003818800032604486, 'timing/model': 0.3714854899881175, 'learning_rate/qwen_vl_interface': 6.402778540314054e-07, 'learning_rate/action_model': 6.402778540314054e-06, 'epoch': 0.34}) +[2026-10-03 15:30:44,719][starVLA.training.train_starvla][INFO] - Step 4621, Loss: {'train/loss': 6.220058276085183e-05, 'train/grad_norm_pre_clip': 0.05226103216409683, 'timing/data': 0.0002644199994392693, 'timing/model': 0.37139048099925276, 'learning_rate/qwen_vl_interface': 6.39544152631649e-07, 'learning_rate/action_model': 6.3954415263164895e-06, 'epoch': 0.34}) +[2026-10-03 15:30:45,378][starVLA.training.train_starvla][INFO] - Step 4622, Loss: {'train/loss': 1.5675446775276214e-05, 'train/grad_norm_pre_clip': 0.004524996969848871, 'timing/data': 0.00036880900734104216, 'timing/model': 0.37008879099448677, 'learning_rate/qwen_vl_interface': 6.388123464161388e-07, 'learning_rate/action_model': 6.388123464161389e-06, 'epoch': 0.34}) +[2026-10-03 15:30:46,036][starVLA.training.train_starvla][INFO] - Step 4623, Loss: {'train/loss': 0.49533653259277344, 'train/grad_norm_pre_clip': 45.08782196044922, 'timing/data': 0.00021765999554190785, 'timing/model': 0.36996839099447243, 'learning_rate/qwen_vl_interface': 6.380824356856936e-07, 'learning_rate/action_model': 6.380824356856936e-06, 'epoch': 0.34}) +[2026-10-03 15:30:46,696][starVLA.training.train_starvla][INFO] - Step 4624, Loss: {'train/loss': 0.21135669946670532, 'train/grad_norm_pre_clip': 112.80416107177734, 'timing/data': 0.0003232200106140226, 'timing/model': 0.37144349099253304, 'learning_rate/qwen_vl_interface': 6.373544207403513e-07, 'learning_rate/action_model': 6.373544207403513e-06, 'epoch': 0.34}) +[2026-10-03 15:30:47,550][starVLA.training.train_starvla][INFO] - Step 4625, Loss: {'train/loss': 2.087538450723514e-05, 'train/grad_norm_pre_clip': 0.006946716457605362, 'timing/data': 0.0003274400078225881, 'timing/model': 0.5648714450071566, 'learning_rate/qwen_vl_interface': 6.366283018793723e-07, 'learning_rate/action_model': 6.366283018793723e-06, 'epoch': 0.34}) +[2026-10-03 15:30:48,210][starVLA.training.train_starvla][INFO] - Step 4626, Loss: {'train/loss': 0.28972920775413513, 'train/grad_norm_pre_clip': 100.78018188476562, 'timing/data': 0.00035432000004220754, 'timing/model': 0.37041504100488964, 'learning_rate/qwen_vl_interface': 6.359040794012358e-07, 'learning_rate/action_model': 6.359040794012358e-06, 'epoch': 0.34}) +[2026-10-03 15:30:48,844][starVLA.training.train_starvla][INFO] - Step 4627, Loss: {'train/loss': 1.683791560935788e-05, 'train/grad_norm_pre_clip': 0.004192896652966738, 'timing/data': 0.00020990999473724514, 'timing/model': 0.34625836499617435, 'learning_rate/qwen_vl_interface': 6.351817536036431e-07, 'learning_rate/action_model': 6.351817536036431e-06, 'epoch': 0.34}) +[2026-10-03 15:30:49,502][starVLA.training.train_starvla][INFO] - Step 4628, Loss: {'train/loss': 1.917660301842261e-05, 'train/grad_norm_pre_clip': 0.007207788992673159, 'timing/data': 0.00026190000062342733, 'timing/model': 0.3705336409911979, 'learning_rate/qwen_vl_interface': 6.344613247835135e-07, 'learning_rate/action_model': 6.344613247835135e-06, 'epoch': 0.34}) +[2026-10-03 15:30:50,139][starVLA.training.train_starvla][INFO] - Step 4629, Loss: {'train/loss': 0.27153581380844116, 'train/grad_norm_pre_clip': 19.050487518310547, 'timing/data': 0.0002652600087458268, 'timing/model': 0.3706860710080946, 'learning_rate/qwen_vl_interface': 6.337427932369898e-07, 'learning_rate/action_model': 6.337427932369898e-06, 'epoch': 0.34}) +[2026-10-03 15:30:50,798][starVLA.training.train_starvla][INFO] - Step 4630, Loss: {'train/loss': 2.7669611881719902e-05, 'train/grad_norm_pre_clip': 0.00875016301870346, 'timing/data': 0.00027196000155527145, 'timing/model': 0.36982959099987056, 'learning_rate/qwen_vl_interface': 6.330261592594328e-07, 'learning_rate/action_model': 6.330261592594327e-06, 'epoch': 0.34}) +[2026-10-03 15:30:51,653][starVLA.training.train_starvla][INFO] - Step 4631, Loss: {'train/loss': 0.00023244068142957985, 'train/grad_norm_pre_clip': 0.12440665811300278, 'timing/data': 0.00036673000431619585, 'timing/model': 0.3697917109966511, 'learning_rate/qwen_vl_interface': 6.323114231454229e-07, 'learning_rate/action_model': 6.323114231454229e-06, 'epoch': 0.34}) +[2026-10-03 15:30:52,312][starVLA.training.train_starvla][INFO] - Step 4632, Loss: {'train/loss': 8.129946218105033e-05, 'train/grad_norm_pre_clip': 0.050699979066848755, 'timing/data': 0.0003066299977945164, 'timing/model': 0.37144519099092577, 'learning_rate/qwen_vl_interface': 6.315985851887618e-07, 'learning_rate/action_model': 6.315985851887618e-06, 'epoch': 0.34}) +[2026-10-03 15:30:52,972][starVLA.training.train_starvla][INFO] - Step 4633, Loss: {'train/loss': 0.37504148483276367, 'train/grad_norm_pre_clip': 60.85066223144531, 'timing/data': 0.0002178799913963303, 'timing/model': 0.37115100100345444, 'learning_rate/qwen_vl_interface': 6.308876456824697e-07, 'learning_rate/action_model': 6.3088764568246975e-06, 'epoch': 0.34}) +[2026-10-03 15:30:53,631][starVLA.training.train_starvla][INFO] - Step 4634, Loss: {'train/loss': 0.4270143508911133, 'train/grad_norm_pre_clip': 32.58298873901367, 'timing/data': 0.00024742999812588096, 'timing/model': 0.3694379309890792, 'learning_rate/qwen_vl_interface': 6.301786049187875e-07, 'learning_rate/action_model': 6.3017860491878745e-06, 'epoch': 0.34}) +[2026-10-03 15:30:54,289][starVLA.training.train_starvla][INFO] - Step 4635, Loss: {'train/loss': 0.0003368437464814633, 'train/grad_norm_pre_clip': 0.6489114165306091, 'timing/data': 0.0003228999994462356, 'timing/model': 0.36964837099367287, 'learning_rate/qwen_vl_interface': 6.294714631891755e-07, 'learning_rate/action_model': 6.294714631891754e-06, 'epoch': 0.34}) +[2026-10-03 15:30:55,149][starVLA.training.train_starvla][INFO] - Step 4636, Loss: {'train/loss': 0.4938051998615265, 'train/grad_norm_pre_clip': 32.66547775268555, 'timing/data': 0.000377349992049858, 'timing/model': 0.3709554910019506, 'learning_rate/qwen_vl_interface': 6.287662207843115e-07, 'learning_rate/action_model': 6.287662207843114e-06, 'epoch': 0.34}) +[2026-10-03 15:30:55,786][starVLA.training.train_starvla][INFO] - Step 4637, Loss: {'train/loss': 1.4885884411341976e-05, 'train/grad_norm_pre_clip': 0.005036463961005211, 'timing/data': 0.0003237000055378303, 'timing/model': 0.37055724099627696, 'learning_rate/qwen_vl_interface': 6.280628779940956e-07, 'learning_rate/action_model': 6.280628779940956e-06, 'epoch': 0.34}) +[2026-10-03 15:30:56,436][starVLA.training.train_starvla][INFO] - Step 4638, Loss: {'train/loss': 0.0002170673105865717, 'train/grad_norm_pre_clip': 0.07454419881105423, 'timing/data': 0.00020149999181739986, 'timing/model': 0.3655437319976045, 'learning_rate/qwen_vl_interface': 6.273614351076445e-07, 'learning_rate/action_model': 6.273614351076445e-06, 'epoch': 0.34}) +[2026-10-03 15:30:57,099][starVLA.training.train_starvla][INFO] - Step 4639, Loss: {'train/loss': 1.725390029605478e-05, 'train/grad_norm_pre_clip': 0.005770637188106775, 'timing/data': 0.0002479199902154505, 'timing/model': 0.3652897709980607, 'learning_rate/qwen_vl_interface': 6.266618924132951e-07, 'learning_rate/action_model': 6.266618924132951e-06, 'epoch': 0.34}) +[2026-10-03 15:30:57,748][starVLA.training.train_starvla][INFO] - Step 4640, Loss: {'train/loss': 0.00026703401817940176, 'train/grad_norm_pre_clip': 0.2242312729358673, 'timing/data': 0.0003004300087923184, 'timing/model': 0.36481620199629106, 'learning_rate/qwen_vl_interface': 6.25964250198603e-07, 'learning_rate/action_model': 6.2596425019860294e-06, 'epoch': 0.34}) +[2026-10-03 15:30:58,585][starVLA.training.train_starvla][INFO] - Step 4641, Loss: {'train/loss': 0.17039242386817932, 'train/grad_norm_pre_clip': 129.67640686035156, 'timing/data': 0.00019626998982857913, 'timing/model': 0.5533674559992505, 'learning_rate/qwen_vl_interface': 6.25268508750342e-07, 'learning_rate/action_model': 6.25268508750342e-06, 'epoch': 0.34}) +[2026-10-03 15:30:59,236][starVLA.training.train_starvla][INFO] - Step 4642, Loss: {'train/loss': 2.4339537048945203e-05, 'train/grad_norm_pre_clip': 0.007790548726916313, 'timing/data': 0.0002158999996026978, 'timing/model': 0.36627693100308534, 'learning_rate/qwen_vl_interface': 6.245746683545062e-07, 'learning_rate/action_model': 6.245746683545062e-06, 'epoch': 0.34}) +[2026-10-03 15:30:59,899][starVLA.training.train_starvla][INFO] - Step 4643, Loss: {'train/loss': 0.00020011265587527305, 'train/grad_norm_pre_clip': 0.08714179694652557, 'timing/data': 0.00033059000270441175, 'timing/model': 0.3648222710035043, 'learning_rate/qwen_vl_interface': 6.238827292963063e-07, 'learning_rate/action_model': 6.238827292963063e-06, 'epoch': 0.34}) +[2026-10-03 15:31:00,549][starVLA.training.train_starvla][INFO] - Step 4644, Loss: {'train/loss': 4.611349868355319e-05, 'train/grad_norm_pre_clip': 0.016626210883259773, 'timing/data': 0.00025599000218790025, 'timing/model': 0.3660339210036909, 'learning_rate/qwen_vl_interface': 6.231926918601731e-07, 'learning_rate/action_model': 6.2319269186017305e-06, 'epoch': 0.34}) +[2026-10-03 15:31:01,198][starVLA.training.train_starvla][INFO] - Step 4645, Loss: {'train/loss': 8.629882358945906e-05, 'train/grad_norm_pre_clip': 0.05016234144568443, 'timing/data': 0.00019958001212216914, 'timing/model': 0.36569775200041477, 'learning_rate/qwen_vl_interface': 6.22504556329754e-07, 'learning_rate/action_model': 6.22504556329754e-06, 'epoch': 0.34}) +[2026-10-03 15:31:02,006][starVLA.training.train_starvla][INFO] - Step 4646, Loss: {'train/loss': 0.0011665462516248226, 'train/grad_norm_pre_clip': 0.30389031767845154, 'timing/data': 0.00021914999524597079, 'timing/model': 0.5249411809927551, 'learning_rate/qwen_vl_interface': 6.218183229879164e-07, 'learning_rate/action_model': 6.218183229879164e-06, 'epoch': 0.34}) +[2026-10-03 15:31:02,668][starVLA.training.train_starvla][INFO] - Step 4647, Loss: {'train/loss': 0.1946679651737213, 'train/grad_norm_pre_clip': 205.35536193847656, 'timing/data': 0.00020515000505838543, 'timing/model': 0.36632913201174233, 'learning_rate/qwen_vl_interface': 6.211339921167447e-07, 'learning_rate/action_model': 6.211339921167446e-06, 'epoch': 0.34}) +[2026-10-03 15:31:03,297][starVLA.training.train_starvla][INFO] - Step 4648, Loss: {'train/loss': 5.3643689170712605e-06, 'train/grad_norm_pre_clip': 0.001158289727754891, 'timing/data': 0.00030562000756617635, 'timing/model': 0.366213950997917, 'learning_rate/qwen_vl_interface': 6.204515639975415e-07, 'learning_rate/action_model': 6.204515639975414e-06, 'epoch': 0.34}) +[2026-10-03 15:31:03,947][starVLA.training.train_starvla][INFO] - Step 4649, Loss: {'train/loss': 0.00011788661504397169, 'train/grad_norm_pre_clip': 0.04166577383875847, 'timing/data': 0.0002095099916914478, 'timing/model': 0.36625616099627223, 'learning_rate/qwen_vl_interface': 6.197710389108262e-07, 'learning_rate/action_model': 6.197710389108262e-06, 'epoch': 0.34}) +[2026-10-03 15:31:04,598][starVLA.training.train_starvla][INFO] - Step 4650, Loss: {'train/loss': 0.258792519569397, 'train/grad_norm_pre_clip': 155.4010009765625, 'timing/data': 0.00020244999905116856, 'timing/model': 0.36617547199421097, 'learning_rate/qwen_vl_interface': 6.190924171363378e-07, 'learning_rate/action_model': 6.190924171363378e-06, 'epoch': 0.34}) +[2026-10-03 15:31:05,438][starVLA.training.train_starvla][INFO] - Step 4651, Loss: {'train/loss': 0.00011755140440072864, 'train/grad_norm_pre_clip': 0.06200810894370079, 'timing/data': 0.0003299999953014776, 'timing/model': 0.5537027770042187, 'learning_rate/qwen_vl_interface': 6.184156989530326e-07, 'learning_rate/action_model': 6.184156989530325e-06, 'epoch': 0.34}) +[2026-10-03 15:31:06,090][starVLA.training.train_starvla][INFO] - Step 4652, Loss: {'train/loss': 4.6846707846270874e-05, 'train/grad_norm_pre_clip': 0.016195330768823624, 'timing/data': 0.00019308000628370792, 'timing/model': 0.36677742199390195, 'learning_rate/qwen_vl_interface': 6.17740884639083e-07, 'learning_rate/action_model': 6.177408846390829e-06, 'epoch': 0.34}) +[2026-10-03 15:31:06,718][starVLA.training.train_starvla][INFO] - Step 4653, Loss: {'train/loss': 0.12776561081409454, 'train/grad_norm_pre_clip': 181.13929748535156, 'timing/data': 0.0003092500119237229, 'timing/model': 0.3660312319989316, 'learning_rate/qwen_vl_interface': 6.170679744718798e-07, 'learning_rate/action_model': 6.1706797447187975e-06, 'epoch': 0.34}) +[2026-10-03 15:31:07,370][starVLA.training.train_starvla][INFO] - Step 4654, Loss: {'train/loss': 3.292730252724141e-05, 'train/grad_norm_pre_clip': 0.018255123868584633, 'timing/data': 0.00023782000062055886, 'timing/model': 0.3665870410040952, 'learning_rate/qwen_vl_interface': 6.163969687280311e-07, 'learning_rate/action_model': 6.163969687280311e-06, 'epoch': 0.34}) +[2026-10-03 15:31:08,022][starVLA.training.train_starvla][INFO] - Step 4655, Loss: {'train/loss': 0.0002571535296738148, 'train/grad_norm_pre_clip': 0.08239144831895828, 'timing/data': 0.00026278999575879425, 'timing/model': 0.36680782098846976, 'learning_rate/qwen_vl_interface': 6.157278676833627e-07, 'learning_rate/action_model': 6.157278676833627e-06, 'epoch': 0.34}) +[2026-10-03 15:31:08,673][starVLA.training.train_starvla][INFO] - Step 4656, Loss: {'train/loss': 0.00022502875071950257, 'train/grad_norm_pre_clip': 0.19051863253116608, 'timing/data': 0.0003236800112063065, 'timing/model': 0.36645522199978586, 'learning_rate/qwen_vl_interface': 6.150606716129155e-07, 'learning_rate/action_model': 6.150606716129154e-06, 'epoch': 0.34}) +[2026-10-03 15:31:09,522][starVLA.training.train_starvla][INFO] - Step 4657, Loss: {'train/loss': 0.11862339824438095, 'train/grad_norm_pre_clip': 92.45892333984375, 'timing/data': 0.0003377800021553412, 'timing/model': 0.36611386100412346, 'learning_rate/qwen_vl_interface': 6.143953807909497e-07, 'learning_rate/action_model': 6.143953807909496e-06, 'epoch': 0.34}) +[2026-10-03 15:31:10,150][starVLA.training.train_starvla][INFO] - Step 4658, Loss: {'train/loss': 6.644744280492887e-05, 'train/grad_norm_pre_clip': 0.03243261203169823, 'timing/data': 0.00024879998818505555, 'timing/model': 0.36627939199388493, 'learning_rate/qwen_vl_interface': 6.137319954909407e-07, 'learning_rate/action_model': 6.137319954909406e-06, 'epoch': 0.34}) +[2026-10-03 15:31:10,802][starVLA.training.train_starvla][INFO] - Step 4659, Loss: {'train/loss': 2.8594340619747527e-05, 'train/grad_norm_pre_clip': 0.008628475479781628, 'timing/data': 0.0003509300004225224, 'timing/model': 0.366368531002081, 'learning_rate/qwen_vl_interface': 6.130705159855815e-07, 'learning_rate/action_model': 6.130705159855814e-06, 'epoch': 0.34}) +[2026-10-03 15:31:11,452][starVLA.training.train_starvla][INFO] - Step 4660, Loss: {'train/loss': 0.04203985258936882, 'train/grad_norm_pre_clip': 100.24503326416016, 'timing/data': 0.0002734200097620487, 'timing/model': 0.3662359709996963, 'learning_rate/qwen_vl_interface': 6.124109425467805e-07, 'learning_rate/action_model': 6.124109425467805e-06, 'epoch': 0.34}) +[2026-10-03 15:31:12,081][starVLA.training.train_starvla][INFO] - Step 4661, Loss: {'train/loss': 0.5427637696266174, 'train/grad_norm_pre_clip': 26.72878646850586, 'timing/data': 0.0002959600096801296, 'timing/model': 0.36627364199375734, 'learning_rate/qwen_vl_interface': 6.117532754456642e-07, 'learning_rate/action_model': 6.117532754456642e-06, 'epoch': 0.34}) +[2026-10-03 15:31:12,928][starVLA.training.train_starvla][INFO] - Step 4662, Loss: {'train/loss': 0.0016546723200008273, 'train/grad_norm_pre_clip': 0.6041074991226196, 'timing/data': 0.0003557399904821068, 'timing/model': 0.3664250519941561, 'learning_rate/qwen_vl_interface': 6.110975149525746e-07, 'learning_rate/action_model': 6.110975149525746e-06, 'epoch': 0.34}) +[2026-10-03 15:31:13,580][starVLA.training.train_starvla][INFO] - Step 4663, Loss: {'train/loss': 2.2664095013169572e-05, 'train/grad_norm_pre_clip': 0.005410643294453621, 'timing/data': 0.00023180000425782055, 'timing/model': 0.36663387100270484, 'learning_rate/qwen_vl_interface': 6.104436613370692e-07, 'learning_rate/action_model': 6.104436613370692e-06, 'epoch': 0.34}) +[2026-10-03 15:31:14,232][starVLA.training.train_starvla][INFO] - Step 4664, Loss: {'train/loss': 2.0413917809491977e-05, 'train/grad_norm_pre_clip': 0.005878133233636618, 'timing/data': 0.00024024999584071338, 'timing/model': 0.36668760199972894, 'learning_rate/qwen_vl_interface': 6.09791714867924e-07, 'learning_rate/action_model': 6.0979171486792396e-06, 'epoch': 0.34}) +[2026-10-03 15:31:14,883][starVLA.training.train_starvla][INFO] - Step 4665, Loss: {'train/loss': 0.001500129234045744, 'train/grad_norm_pre_clip': 0.48266392946243286, 'timing/data': 0.00023548000899609178, 'timing/model': 0.36638628201035317, 'learning_rate/qwen_vl_interface': 6.091416758131287e-07, 'learning_rate/action_model': 6.091416758131286e-06, 'epoch': 0.34}) +[2026-10-03 15:31:15,533][starVLA.training.train_starvla][INFO] - Step 4666, Loss: {'train/loss': 0.0016236434457823634, 'train/grad_norm_pre_clip': 0.819728434085846, 'timing/data': 0.000308040005620569, 'timing/model': 0.36648420100391377, 'learning_rate/qwen_vl_interface': 6.084935444398895e-07, 'learning_rate/action_model': 6.084935444398894e-06, 'epoch': 0.34}) +[2026-10-03 15:31:16,381][starVLA.training.train_starvla][INFO] - Step 4667, Loss: {'train/loss': 4.808061203220859e-05, 'train/grad_norm_pre_clip': 0.015666795894503593, 'timing/data': 0.00028854999982286245, 'timing/model': 0.562436735999654, 'learning_rate/qwen_vl_interface': 6.078473210146294e-07, 'learning_rate/action_model': 6.078473210146294e-06, 'epoch': 0.34}) +[2026-10-03 15:31:17,031][starVLA.training.train_starvla][INFO] - Step 4668, Loss: {'train/loss': 2.372174640186131e-05, 'train/grad_norm_pre_clip': 0.008055026642978191, 'timing/data': 0.00023643000167794526, 'timing/model': 0.3658093919948442, 'learning_rate/qwen_vl_interface': 6.072030058029862e-07, 'learning_rate/action_model': 6.072030058029861e-06, 'epoch': 0.34}) +[2026-10-03 15:31:17,680][starVLA.training.train_starvla][INFO] - Step 4669, Loss: {'train/loss': 0.0001344281481578946, 'train/grad_norm_pre_clip': 0.08707232773303986, 'timing/data': 0.0003018199931830168, 'timing/model': 0.36528180200548377, 'learning_rate/qwen_vl_interface': 6.065605990698136e-07, 'learning_rate/action_model': 6.065605990698136e-06, 'epoch': 0.34}) +[2026-10-03 15:31:18,330][starVLA.training.train_starvla][INFO] - Step 4670, Loss: {'train/loss': 7.438674219883978e-05, 'train/grad_norm_pre_clip': 0.03264223784208298, 'timing/data': 0.00024761000531725585, 'timing/model': 0.36571685200033244, 'learning_rate/qwen_vl_interface': 6.059201010791811e-07, 'learning_rate/action_model': 6.059201010791811e-06, 'epoch': 0.34}) +[2026-10-03 15:31:18,982][starVLA.training.train_starvla][INFO] - Step 4671, Loss: {'train/loss': 0.0002026526053668931, 'train/grad_norm_pre_clip': 0.07126239687204361, 'timing/data': 0.0002973399969050661, 'timing/model': 0.3660667719959747, 'learning_rate/qwen_vl_interface': 6.052815120943721e-07, 'learning_rate/action_model': 6.052815120943721e-06, 'epoch': 0.34}) +[2026-10-03 15:31:19,807][starVLA.training.train_starvla][INFO] - Step 4672, Loss: {'train/loss': 4.891342541668564e-05, 'train/grad_norm_pre_clip': 0.018672065809369087, 'timing/data': 0.00031812999804969877, 'timing/model': 0.5630878750089323, 'learning_rate/qwen_vl_interface': 6.046448323778874e-07, 'learning_rate/action_model': 6.046448323778874e-06, 'epoch': 0.34}) +[2026-10-03 15:31:20,458][starVLA.training.train_starvla][INFO] - Step 4673, Loss: {'train/loss': 0.00012485632032621652, 'train/grad_norm_pre_clip': 0.058284591883420944, 'timing/data': 0.0002424300037091598, 'timing/model': 0.3659306109911995, 'learning_rate/qwen_vl_interface': 6.040100621914419e-07, 'learning_rate/action_model': 6.040100621914418e-06, 'epoch': 0.34}) +[2026-10-03 15:31:21,108][starVLA.training.train_starvla][INFO] - Step 4674, Loss: {'train/loss': 5.455628706840798e-05, 'train/grad_norm_pre_clip': 0.01672469452023506, 'timing/data': 0.0004207199963275343, 'timing/model': 0.3660280519980006, 'learning_rate/qwen_vl_interface': 6.033772017959652e-07, 'learning_rate/action_model': 6.033772017959652e-06, 'epoch': 0.34}) +[2026-10-03 15:31:21,760][starVLA.training.train_starvla][INFO] - Step 4675, Loss: {'train/loss': 9.372281783726066e-05, 'train/grad_norm_pre_clip': 0.041617561131715775, 'timing/data': 0.0002731499989749864, 'timing/model': 0.3658194619929418, 'learning_rate/qwen_vl_interface': 6.027462514516022e-07, 'learning_rate/action_model': 6.0274625145160214e-06, 'epoch': 0.34}) +[2026-10-03 15:31:22,410][starVLA.training.train_starvla][INFO] - Step 4676, Loss: {'train/loss': 0.3053344190120697, 'train/grad_norm_pre_clip': 117.4300537109375, 'timing/data': 0.0002493499923730269, 'timing/model': 0.3656417820020579, 'learning_rate/qwen_vl_interface': 6.021172114177139e-07, 'learning_rate/action_model': 6.021172114177139e-06, 'epoch': 0.34}) +[2026-10-03 15:31:23,060][starVLA.training.train_starvla][INFO] - Step 4677, Loss: {'train/loss': 0.000102214929938782, 'train/grad_norm_pre_clip': 0.0887974202632904, 'timing/data': 0.00029804000223521143, 'timing/model': 0.36592918100359384, 'learning_rate/qwen_vl_interface': 6.014900819528734e-07, 'learning_rate/action_model': 6.0149008195287335e-06, 'epoch': 0.34}) +[2026-10-03 15:31:23,912][starVLA.training.train_starvla][INFO] - Step 4678, Loss: {'train/loss': 0.0003522516926750541, 'train/grad_norm_pre_clip': 0.19111882150173187, 'timing/data': 0.000251730001764372, 'timing/model': 0.36693606099288445, 'learning_rate/qwen_vl_interface': 6.008648633148707e-07, 'learning_rate/action_model': 6.008648633148707e-06, 'epoch': 0.34}) +[2026-10-03 15:31:24,564][starVLA.training.train_starvla][INFO] - Step 4679, Loss: {'train/loss': 0.31479334831237793, 'train/grad_norm_pre_clip': 53.5316276550293, 'timing/data': 0.00029910000739619136, 'timing/model': 0.3664412220095983, 'learning_rate/qwen_vl_interface': 6.002415557607097e-07, 'learning_rate/action_model': 6.002415557607096e-06, 'epoch': 0.34}) +[2026-10-03 15:31:25,215][starVLA.training.train_starvla][INFO] - Step 4680, Loss: {'train/loss': 5.480008985614404e-05, 'train/grad_norm_pre_clip': 0.021442875266075134, 'timing/data': 0.00020369999401737005, 'timing/model': 0.36663123199832626, 'learning_rate/qwen_vl_interface': 5.996201595466088e-07, 'learning_rate/action_model': 5.996201595466088e-06, 'epoch': 0.34}) +[2026-10-03 15:31:25,867][starVLA.training.train_starvla][INFO] - Step 4681, Loss: {'train/loss': 1.5221223293337971e-05, 'train/grad_norm_pre_clip': 0.003745722584426403, 'timing/data': 0.0003095400024903938, 'timing/model': 0.36679025199555326, 'learning_rate/qwen_vl_interface': 5.990006749279998e-07, 'learning_rate/action_model': 5.990006749279998e-06, 'epoch': 0.34}) +[2026-10-03 15:31:26,517][starVLA.training.train_starvla][INFO] - Step 4682, Loss: {'train/loss': 0.00015717912174295634, 'train/grad_norm_pre_clip': 0.08337608724832535, 'timing/data': 0.000242110007093288, 'timing/model': 0.3657715410081437, 'learning_rate/qwen_vl_interface': 5.983831021595296e-07, 'learning_rate/action_model': 5.9838310215952956e-06, 'epoch': 0.34}) +[2026-10-03 15:31:27,369][starVLA.training.train_starvla][INFO] - Step 4683, Loss: {'train/loss': 0.0005732052377425134, 'train/grad_norm_pre_clip': 1.382753849029541, 'timing/data': 0.0002797899942379445, 'timing/model': 0.3663932119961828, 'learning_rate/qwen_vl_interface': 5.977674414950595e-07, 'learning_rate/action_model': 5.9776744149505955e-06, 'epoch': 0.34}) +[2026-10-03 15:31:28,019][starVLA.training.train_starvla][INFO] - Step 4684, Loss: {'train/loss': 0.4531542956829071, 'train/grad_norm_pre_clip': 29.790647506713867, 'timing/data': 0.00024090999795589596, 'timing/model': 0.36642552200646605, 'learning_rate/qwen_vl_interface': 5.971536931876635e-07, 'learning_rate/action_model': 5.971536931876634e-06, 'epoch': 0.34}) +[2026-10-03 15:31:28,670][starVLA.training.train_starvla][INFO] - Step 4685, Loss: {'train/loss': 0.00014254607958719134, 'train/grad_norm_pre_clip': 0.06423249840736389, 'timing/data': 0.00033274000452365726, 'timing/model': 0.36634639199473895, 'learning_rate/qwen_vl_interface': 5.965418574896315e-07, 'learning_rate/action_model': 5.9654185748963145e-06, 'epoch': 0.34}) +[2026-10-03 15:31:29,320][starVLA.training.train_starvla][INFO] - Step 4686, Loss: {'train/loss': 0.00496907252818346, 'train/grad_norm_pre_clip': 19.383304595947266, 'timing/data': 0.0002422700054012239, 'timing/model': 0.36610565199225675, 'learning_rate/qwen_vl_interface': 5.959319346524646e-07, 'learning_rate/action_model': 5.959319346524646e-06, 'epoch': 0.34}) +[2026-10-03 15:31:29,973][starVLA.training.train_starvla][INFO] - Step 4687, Loss: {'train/loss': 0.0001963303511729464, 'train/grad_norm_pre_clip': 0.11609463393688202, 'timing/data': 0.00029305000498425215, 'timing/model': 0.366804000994307, 'learning_rate/qwen_vl_interface': 5.953239249268804e-07, 'learning_rate/action_model': 5.953239249268804e-06, 'epoch': 0.34}) +[2026-10-03 15:31:30,791][starVLA.training.train_starvla][INFO] - Step 4688, Loss: {'train/loss': 7.525523687945679e-05, 'train/grad_norm_pre_clip': 0.030807631090283394, 'timing/data': 0.00021482999727595598, 'timing/model': 0.5332402790081687, 'learning_rate/qwen_vl_interface': 5.947178285628081e-07, 'learning_rate/action_model': 5.94717828562808e-06, 'epoch': 0.34}) +[2026-10-03 15:31:31,442][starVLA.training.train_starvla][INFO] - Step 4689, Loss: {'train/loss': 0.0006162795471027493, 'train/grad_norm_pre_clip': 0.23886461555957794, 'timing/data': 0.0002622900065034628, 'timing/model': 0.36717794099240564, 'learning_rate/qwen_vl_interface': 5.941136458093907e-07, 'learning_rate/action_model': 5.941136458093907e-06, 'epoch': 0.34}) +[2026-10-03 15:31:32,071][starVLA.training.train_starvla][INFO] - Step 4690, Loss: {'train/loss': 0.35813236236572266, 'train/grad_norm_pre_clip': 55.05973434448242, 'timing/data': 0.00032524998823646456, 'timing/model': 0.36639024199394044, 'learning_rate/qwen_vl_interface': 5.935113769149855e-07, 'learning_rate/action_model': 5.935113769149856e-06, 'epoch': 0.34}) +[2026-10-03 15:31:32,721][starVLA.training.train_starvla][INFO] - Step 4691, Loss: {'train/loss': 0.09810002148151398, 'train/grad_norm_pre_clip': 273.5706787109375, 'timing/data': 0.0002600499865366146, 'timing/model': 0.36437601200304925, 'learning_rate/qwen_vl_interface': 5.92911022127163e-07, 'learning_rate/action_model': 5.929110221271629e-06, 'epoch': 0.34}) +[2026-10-03 15:31:33,346][starVLA.training.train_starvla][INFO] - Step 4692, Loss: {'train/loss': 0.00013645077706314623, 'train/grad_norm_pre_clip': 0.08899033814668655, 'timing/data': 0.0003100399917457253, 'timing/model': 0.34208169499470387, 'learning_rate/qwen_vl_interface': 5.92312581692705e-07, 'learning_rate/action_model': 5.923125816927049e-06, 'epoch': 0.34}) +[2026-10-03 15:31:33,993][starVLA.training.train_starvla][INFO] - Step 4693, Loss: {'train/loss': 0.00010994571493938565, 'train/grad_norm_pre_clip': 0.0334710069000721, 'timing/data': 0.00023895999765954912, 'timing/model': 0.3648704009974608, 'learning_rate/qwen_vl_interface': 5.917160558576088e-07, 'learning_rate/action_model': 5.917160558576088e-06, 'epoch': 0.34}) +[2026-10-03 15:31:34,832][starVLA.training.train_starvla][INFO] - Step 4694, Loss: {'train/loss': 0.00947003997862339, 'train/grad_norm_pre_clip': 20.3241024017334, 'timing/data': 0.0003218199999537319, 'timing/model': 0.3648227720113937, 'learning_rate/qwen_vl_interface': 5.911214448670829e-07, 'learning_rate/action_model': 5.911214448670828e-06, 'epoch': 0.34}) +[2026-10-03 15:31:35,481][starVLA.training.train_starvla][INFO] - Step 4695, Loss: {'train/loss': 0.0009018363198265433, 'train/grad_norm_pre_clip': 0.5377433896064758, 'timing/data': 0.0002648700028657913, 'timing/model': 0.36473576100252103, 'learning_rate/qwen_vl_interface': 5.905287489655499e-07, 'learning_rate/action_model': 5.905287489655499e-06, 'epoch': 0.34}) +[2026-10-03 15:31:36,128][starVLA.training.train_starvla][INFO] - Step 4696, Loss: {'train/loss': 6.470047810580581e-05, 'train/grad_norm_pre_clip': 0.02539166994392872, 'timing/data': 0.0002408599975751713, 'timing/model': 0.36441459199704695, 'learning_rate/qwen_vl_interface': 5.899379683966445e-07, 'learning_rate/action_model': 5.899379683966445e-06, 'epoch': 0.34}) +[2026-10-03 15:31:36,776][starVLA.training.train_starvla][INFO] - Step 4697, Loss: {'train/loss': 0.009798996150493622, 'train/grad_norm_pre_clip': 47.49075698852539, 'timing/data': 0.00031386000046040863, 'timing/model': 0.36472383099317085, 'learning_rate/qwen_vl_interface': 5.893491034032142e-07, 'learning_rate/action_model': 5.893491034032142e-06, 'epoch': 0.34}) +[2026-10-03 15:31:37,422][starVLA.training.train_starvla][INFO] - Step 4698, Loss: {'train/loss': 5.290783155942336e-05, 'train/grad_norm_pre_clip': 0.03143564984202385, 'timing/data': 0.0002179300063289702, 'timing/model': 0.36436961199797224, 'learning_rate/qwen_vl_interface': 5.887621542273196e-07, 'learning_rate/action_model': 5.887621542273196e-06, 'epoch': 0.34}) +[2026-10-03 15:31:38,265][starVLA.training.train_starvla][INFO] - Step 4699, Loss: {'train/loss': 9.688533464213833e-05, 'train/grad_norm_pre_clip': 0.03800007328391075, 'timing/data': 0.00024046999169513583, 'timing/model': 0.5582327260053717, 'learning_rate/qwen_vl_interface': 5.881771211102327e-07, 'learning_rate/action_model': 5.881771211102327e-06, 'epoch': 0.34}) +[2026-10-03 15:31:38,912][starVLA.training.train_starvla][INFO] - Step 4700, Loss: {'train/loss': 7.226999514386989e-06, 'train/grad_norm_pre_clip': 0.0015508988872170448, 'timing/data': 0.00023135000083129853, 'timing/model': 0.3648340219951933, 'learning_rate/qwen_vl_interface': 5.875940042924391e-07, 'learning_rate/action_model': 5.875940042924391e-06, 'epoch': 0.34}) +[2026-10-03 15:31:39,560][starVLA.training.train_starvla][INFO] - Step 4701, Loss: {'train/loss': 6.462149758590385e-05, 'train/grad_norm_pre_clip': 0.025822853669524193, 'timing/data': 0.00024006998864933848, 'timing/model': 0.3646714409987908, 'learning_rate/qwen_vl_interface': 5.870128040136368e-07, 'learning_rate/action_model': 5.870128040136367e-06, 'epoch': 0.34}) +[2026-10-03 15:31:40,207][starVLA.training.train_starvla][INFO] - Step 4702, Loss: {'train/loss': 3.192354051861912e-05, 'train/grad_norm_pre_clip': 0.014851952902972698, 'timing/data': 0.00031805899925529957, 'timing/model': 0.3645849430031376, 'learning_rate/qwen_vl_interface': 5.864335205127332e-07, 'learning_rate/action_model': 5.864335205127332e-06, 'epoch': 0.34}) +[2026-10-03 15:31:40,857][starVLA.training.train_starvla][INFO] - Step 4703, Loss: {'train/loss': 0.00026798874023370445, 'train/grad_norm_pre_clip': 0.14014142751693726, 'timing/data': 0.0002894500066759065, 'timing/model': 0.3647289910004474, 'learning_rate/qwen_vl_interface': 5.858561540278513e-07, 'learning_rate/action_model': 5.858561540278513e-06, 'epoch': 0.34}) +[2026-10-03 15:31:41,670][starVLA.training.train_starvla][INFO] - Step 4704, Loss: {'train/loss': 5.7430130254942924e-05, 'train/grad_norm_pre_clip': 0.025229237973690033, 'timing/data': 0.00034004999906755984, 'timing/model': 0.5522082969982876, 'learning_rate/qwen_vl_interface': 5.852807047963244e-07, 'learning_rate/action_model': 5.852807047963243e-06, 'epoch': 0.34}) +[2026-10-03 15:31:42,317][starVLA.training.train_starvla][INFO] - Step 4705, Loss: {'train/loss': 0.45548269152641296, 'train/grad_norm_pre_clip': 51.3658332824707, 'timing/data': 0.00022922999050933868, 'timing/model': 0.3649196309997933, 'learning_rate/qwen_vl_interface': 5.847071730546974e-07, 'learning_rate/action_model': 5.847071730546974e-06, 'epoch': 0.34}) +[2026-10-03 15:31:42,965][starVLA.training.train_starvla][INFO] - Step 4706, Loss: {'train/loss': 3.596088936319575e-05, 'train/grad_norm_pre_clip': 0.010387187823653221, 'timing/data': 0.0002293900033691898, 'timing/model': 0.36474458199518267, 'learning_rate/qwen_vl_interface': 5.841355590387291e-07, 'learning_rate/action_model': 5.841355590387291e-06, 'epoch': 0.34}) +[2026-10-03 15:31:43,592][starVLA.training.train_starvla][INFO] - Step 4707, Loss: {'train/loss': 7.629347237525508e-06, 'train/grad_norm_pre_clip': 0.0013481093337759376, 'timing/data': 0.0002489600010449067, 'timing/model': 0.3422785550064873, 'learning_rate/qwen_vl_interface': 5.835658629833866e-07, 'learning_rate/action_model': 5.8356586298338654e-06, 'epoch': 0.34}) +[2026-10-03 15:31:44,239][starVLA.training.train_starvla][INFO] - Step 4708, Loss: {'train/loss': 0.0010560917435213923, 'train/grad_norm_pre_clip': 0.2221958488225937, 'timing/data': 0.0003140900080325082, 'timing/model': 0.3643324529984966, 'learning_rate/qwen_vl_interface': 5.829980851228512e-07, 'learning_rate/action_model': 5.829980851228512e-06, 'epoch': 0.34}) +[2026-10-03 15:31:44,887][starVLA.training.train_starvla][INFO] - Step 4709, Loss: {'train/loss': 0.0003166158276144415, 'train/grad_norm_pre_clip': 0.389609158039093, 'timing/data': 0.000309049995848909, 'timing/model': 0.3642696220049402, 'learning_rate/qwen_vl_interface': 5.824322256905155e-07, 'learning_rate/action_model': 5.824322256905154e-06, 'epoch': 0.34}) +[2026-10-03 15:31:45,727][starVLA.training.train_starvla][INFO] - Step 4710, Loss: {'train/loss': 0.0001069630598067306, 'train/grad_norm_pre_clip': 0.04007275775074959, 'timing/data': 0.00022969998826738447, 'timing/model': 0.36453213199274614, 'learning_rate/qwen_vl_interface': 5.818682849189818e-07, 'learning_rate/action_model': 5.818682849189817e-06, 'epoch': 0.34}) +[2026-10-03 15:31:46,376][starVLA.training.train_starvla][INFO] - Step 4711, Loss: {'train/loss': 0.2315794974565506, 'train/grad_norm_pre_clip': 107.33943939208984, 'timing/data': 0.00028685000143013895, 'timing/model': 0.3643448719958542, 'learning_rate/qwen_vl_interface': 5.813062630400653e-07, 'learning_rate/action_model': 5.813062630400653e-06, 'epoch': 0.34}) +[2026-10-03 15:31:47,000][starVLA.training.train_starvla][INFO] - Step 4712, Loss: {'train/loss': 2.3155213057179935e-05, 'train/grad_norm_pre_clip': 0.009081672877073288, 'timing/data': 0.0003034900000784546, 'timing/model': 0.3416419749992201, 'learning_rate/qwen_vl_interface': 5.807461602847922e-07, 'learning_rate/action_model': 5.807461602847922e-06, 'epoch': 0.34}) +[2026-10-03 15:31:47,625][starVLA.training.train_starvla][INFO] - Step 4713, Loss: {'train/loss': 3.963892231695354e-05, 'train/grad_norm_pre_clip': 0.01528154220432043, 'timing/data': 0.0003279800002928823, 'timing/model': 0.34227483499853406, 'learning_rate/qwen_vl_interface': 5.801879768833998e-07, 'learning_rate/action_model': 5.801879768833998e-06, 'epoch': 0.34}) +[2026-10-03 15:31:48,273][starVLA.training.train_starvla][INFO] - Step 4714, Loss: {'train/loss': 0.00013921756180934608, 'train/grad_norm_pre_clip': 0.04474113881587982, 'timing/data': 0.0003236600023228675, 'timing/model': 0.36465095200401265, 'learning_rate/qwen_vl_interface': 5.796317130653363e-07, 'learning_rate/action_model': 5.796317130653363e-06, 'epoch': 0.35}) +[2026-10-03 15:31:49,115][starVLA.training.train_starvla][INFO] - Step 4715, Loss: {'train/loss': 3.154326986987144e-05, 'train/grad_norm_pre_clip': 0.009722946211695671, 'timing/data': 0.0003439399879425764, 'timing/model': 0.3644842620124109, 'learning_rate/qwen_vl_interface': 5.790773690592592e-07, 'learning_rate/action_model': 5.7907736905925925e-06, 'epoch': 0.35}) +[2026-10-03 15:31:49,762][starVLA.training.train_starvla][INFO] - Step 4716, Loss: {'train/loss': 1.0654147445166018e-05, 'train/grad_norm_pre_clip': 0.0025876942090690136, 'timing/data': 0.00023925999994389713, 'timing/model': 0.36429957101063337, 'learning_rate/qwen_vl_interface': 5.7852494509304e-07, 'learning_rate/action_model': 5.7852494509304e-06, 'epoch': 0.35}) +[2026-10-03 15:31:50,411][starVLA.training.train_starvla][INFO] - Step 4717, Loss: {'train/loss': 0.0006370717310346663, 'train/grad_norm_pre_clip': 0.8840482234954834, 'timing/data': 0.0003724799898918718, 'timing/model': 0.3652213020104682, 'learning_rate/qwen_vl_interface': 5.779744413937592e-07, 'learning_rate/action_model': 5.779744413937592e-06, 'epoch': 0.35}) +[2026-10-03 15:31:51,061][starVLA.training.train_starvla][INFO] - Step 4718, Loss: {'train/loss': 0.42002949118614197, 'train/grad_norm_pre_clip': 37.06251907348633, 'timing/data': 0.00032471001031808555, 'timing/model': 0.3653367530059768, 'learning_rate/qwen_vl_interface': 5.774258581877073e-07, 'learning_rate/action_model': 5.774258581877073e-06, 'epoch': 0.35}) +[2026-10-03 15:31:51,710][starVLA.training.train_starvla][INFO] - Step 4719, Loss: {'train/loss': 3.2718558941269293e-05, 'train/grad_norm_pre_clip': 0.019220741465687752, 'timing/data': 0.0002684599894564599, 'timing/model': 0.3646433009998873, 'learning_rate/qwen_vl_interface': 5.768791957003866e-07, 'learning_rate/action_model': 5.768791957003865e-06, 'epoch': 0.35}) +[2026-10-03 15:31:52,551][starVLA.training.train_starvla][INFO] - Step 4720, Loss: {'train/loss': 2.605375811981503e-05, 'train/grad_norm_pre_clip': 0.008710517548024654, 'timing/data': 0.0002798999921651557, 'timing/model': 0.557628456997918, 'learning_rate/qwen_vl_interface': 5.763344541565091e-07, 'learning_rate/action_model': 5.7633445415650905e-06, 'epoch': 0.35}) +[2026-10-03 15:31:53,198][starVLA.training.train_starvla][INFO] - Step 4721, Loss: {'train/loss': 0.0005734537844546139, 'train/grad_norm_pre_clip': 0.3180779814720154, 'timing/data': 0.00021158000163268298, 'timing/model': 0.3647625220037298, 'learning_rate/qwen_vl_interface': 5.75791633779998e-07, 'learning_rate/action_model': 5.75791633779998e-06, 'epoch': 0.35}) +[2026-10-03 15:31:53,824][starVLA.training.train_starvla][INFO] - Step 4722, Loss: {'train/loss': 2.9471802918124013e-05, 'train/grad_norm_pre_clip': 0.011340390890836716, 'timing/data': 0.00023577999672852457, 'timing/model': 0.342707044997951, 'learning_rate/qwen_vl_interface': 5.752507347939857e-07, 'learning_rate/action_model': 5.752507347939858e-06, 'epoch': 0.35}) +[2026-10-03 15:31:54,474][starVLA.training.train_starvla][INFO] - Step 4723, Loss: {'train/loss': 1.5057346899993718e-05, 'train/grad_norm_pre_clip': 0.003291709814220667, 'timing/data': 0.0003163600049447268, 'timing/model': 0.36519116099225357, 'learning_rate/qwen_vl_interface': 5.74711757420816e-07, 'learning_rate/action_model': 5.747117574208159e-06, 'epoch': 0.35}) +[2026-10-03 15:31:55,102][starVLA.training.train_starvla][INFO] - Step 4724, Loss: {'train/loss': 0.002933658193796873, 'train/grad_norm_pre_clip': 7.315537929534912, 'timing/data': 0.00025580999499652535, 'timing/model': 0.36603378200379666, 'learning_rate/qwen_vl_interface': 5.74174701882042e-07, 'learning_rate/action_model': 5.74174701882042e-06, 'epoch': 0.35}) +[2026-10-03 15:31:55,942][starVLA.training.train_starvla][INFO] - Step 4725, Loss: {'train/loss': 0.3339766263961792, 'train/grad_norm_pre_clip': 63.13429260253906, 'timing/data': 0.00032140000257641077, 'timing/model': 0.5553492170001846, 'learning_rate/qwen_vl_interface': 5.736395683984262e-07, 'learning_rate/action_model': 5.736395683984261e-06, 'epoch': 0.35}) +[2026-10-03 15:31:56,594][starVLA.training.train_starvla][INFO] - Step 4726, Loss: {'train/loss': 0.46992090344429016, 'train/grad_norm_pre_clip': 26.267202377319336, 'timing/data': 0.00022533998708240688, 'timing/model': 0.3668313820089679, 'learning_rate/qwen_vl_interface': 5.731063571899421e-07, 'learning_rate/action_model': 5.731063571899421e-06, 'epoch': 0.35}) +[2026-10-03 15:31:57,246][starVLA.training.train_starvla][INFO] - Step 4727, Loss: {'train/loss': 1.4900815585860983e-05, 'train/grad_norm_pre_clip': 0.003836605465039611, 'timing/data': 0.0003340599942021072, 'timing/model': 0.36647413100581616, 'learning_rate/qwen_vl_interface': 5.72575068475773e-07, 'learning_rate/action_model': 5.72575068475773e-06, 'epoch': 0.35}) +[2026-10-03 15:31:57,897][starVLA.training.train_starvla][INFO] - Step 4728, Loss: {'train/loss': 3.321895928820595e-05, 'train/grad_norm_pre_clip': 0.013038079254329205, 'timing/data': 0.000405359998694621, 'timing/model': 0.3664350509934593, 'learning_rate/qwen_vl_interface': 5.720457024743115e-07, 'learning_rate/action_model': 5.720457024743115e-06, 'epoch': 0.35}) +[2026-10-03 15:31:58,548][starVLA.training.train_starvla][INFO] - Step 4729, Loss: {'train/loss': 4.552431346382946e-05, 'train/grad_norm_pre_clip': 0.020574375987052917, 'timing/data': 0.00036391000321600586, 'timing/model': 0.3664931509993039, 'learning_rate/qwen_vl_interface': 5.715182594031601e-07, 'learning_rate/action_model': 5.7151825940316e-06, 'epoch': 0.35}) +[2026-10-03 15:31:59,199][starVLA.training.train_starvla][INFO] - Step 4730, Loss: {'train/loss': 0.00017993432993534952, 'train/grad_norm_pre_clip': 0.04805229231715202, 'timing/data': 0.00034775999665725976, 'timing/model': 0.3670854810043238, 'learning_rate/qwen_vl_interface': 5.709927394791305e-07, 'learning_rate/action_model': 5.709927394791305e-06, 'epoch': 0.35}) +[2026-10-03 15:32:00,048][starVLA.training.train_starvla][INFO] - Step 4731, Loss: {'train/loss': 6.776979716960341e-05, 'train/grad_norm_pre_clip': 0.08714012056589127, 'timing/data': 0.00027970000519417226, 'timing/model': 0.3665087109984597, 'learning_rate/qwen_vl_interface': 5.70469142918244e-07, 'learning_rate/action_model': 5.704691429182441e-06, 'epoch': 0.35}) +[2026-10-03 15:32:00,699][starVLA.training.train_starvla][INFO] - Step 4732, Loss: {'train/loss': 0.017912618815898895, 'train/grad_norm_pre_clip': 49.5693359375, 'timing/data': 0.000198840003577061, 'timing/model': 0.36661710099724587, 'learning_rate/qwen_vl_interface': 5.699474699357317e-07, 'learning_rate/action_model': 5.699474699357317e-06, 'epoch': 0.35}) +[2026-10-03 15:32:01,349][starVLA.training.train_starvla][INFO] - Step 4733, Loss: {'train/loss': 0.003918426111340523, 'train/grad_norm_pre_clip': 13.353711128234863, 'timing/data': 0.00027471999055705965, 'timing/model': 0.36580332199810073, 'learning_rate/qwen_vl_interface': 5.694277207460331e-07, 'learning_rate/action_model': 5.694277207460332e-06, 'epoch': 0.35}) +[2026-10-03 15:32:02,001][starVLA.training.train_starvla][INFO] - Step 4734, Loss: {'train/loss': 0.035416558384895325, 'train/grad_norm_pre_clip': 182.65252685546875, 'timing/data': 0.000291750009637326, 'timing/model': 0.36659780199988745, 'learning_rate/qwen_vl_interface': 5.689098955627982e-07, 'learning_rate/action_model': 5.689098955627982e-06, 'epoch': 0.35}) +[2026-10-03 15:32:02,631][starVLA.training.train_starvla][INFO] - Step 4735, Loss: {'train/loss': 0.0002625097695272416, 'train/grad_norm_pre_clip': 0.08058395236730576, 'timing/data': 0.0002734199952101335, 'timing/model': 0.34457612500409596, 'learning_rate/qwen_vl_interface': 5.683939945988853e-07, 'learning_rate/action_model': 5.683939945988853e-06, 'epoch': 0.35}) +[2026-10-03 15:32:03,482][starVLA.training.train_starvla][INFO] - Step 4736, Loss: {'train/loss': 2.071157723548822e-05, 'train/grad_norm_pre_clip': 0.010121943429112434, 'timing/data': 0.00026436999905854464, 'timing/model': 0.3661420709977392, 'learning_rate/qwen_vl_interface': 5.678800180663608e-07, 'learning_rate/action_model': 5.678800180663609e-06, 'epoch': 0.35}) +[2026-10-03 15:32:04,110][starVLA.training.train_starvla][INFO] - Step 4737, Loss: {'train/loss': 0.0007934995810501277, 'train/grad_norm_pre_clip': 0.2032393515110016, 'timing/data': 0.00029363999783527106, 'timing/model': 0.3439059840020491, 'learning_rate/qwen_vl_interface': 5.673679661765018e-07, 'learning_rate/action_model': 5.6736796617650175e-06, 'epoch': 0.35}) +[2026-10-03 15:32:04,761][starVLA.training.train_starvla][INFO] - Step 4738, Loss: {'train/loss': 0.0005200223531574011, 'train/grad_norm_pre_clip': 0.366318017244339, 'timing/data': 0.0002788299898384139, 'timing/model': 0.3662748710048618, 'learning_rate/qwen_vl_interface': 5.668578391397938e-07, 'learning_rate/action_model': 5.668578391397937e-06, 'epoch': 0.35}) +[2026-10-03 15:32:05,414][starVLA.training.train_starvla][INFO] - Step 4739, Loss: {'train/loss': 0.0006606116658076644, 'train/grad_norm_pre_clip': 0.8205951452255249, 'timing/data': 0.0002805099938996136, 'timing/model': 0.3665540820074966, 'learning_rate/qwen_vl_interface': 5.663496371659303e-07, 'learning_rate/action_model': 5.663496371659303e-06, 'epoch': 0.35}) +[2026-10-03 15:32:06,066][starVLA.training.train_starvla][INFO] - Step 4740, Loss: {'train/loss': 0.0013091780710965395, 'train/grad_norm_pre_clip': 1.2099477052688599, 'timing/data': 0.0003324200079077855, 'timing/model': 0.36633919199812226, 'learning_rate/qwen_vl_interface': 5.658433604638141e-07, 'learning_rate/action_model': 5.65843360463814e-06, 'epoch': 0.35}) +[2026-10-03 15:32:06,920][starVLA.training.train_starvla][INFO] - Step 4741, Loss: {'train/loss': 0.018121754750609398, 'train/grad_norm_pre_clip': 66.00762939453125, 'timing/data': 0.00022718998661730438, 'timing/model': 0.5657423149968963, 'learning_rate/qwen_vl_interface': 5.653390092415559e-07, 'learning_rate/action_model': 5.653390092415559e-06, 'epoch': 0.35}) +[2026-10-03 15:32:07,571][starVLA.training.train_starvla][INFO] - Step 4742, Loss: {'train/loss': 6.84443220961839e-05, 'train/grad_norm_pre_clip': 0.02311471290886402, 'timing/data': 0.00019910999981220812, 'timing/model': 0.3661168309918139, 'learning_rate/qwen_vl_interface': 5.648365837064759e-07, 'learning_rate/action_model': 5.648365837064759e-06, 'epoch': 0.35}) +[2026-10-03 15:32:08,224][starVLA.training.train_starvla][INFO] - Step 4743, Loss: {'train/loss': 5.9245230659144e-05, 'train/grad_norm_pre_clip': 0.02975700981914997, 'timing/data': 0.00031740999838802963, 'timing/model': 0.36650573200313374, 'learning_rate/qwen_vl_interface': 5.643360840651025e-07, 'learning_rate/action_model': 5.643360840651025e-06, 'epoch': 0.35}) +[2026-10-03 15:32:08,875][starVLA.training.train_starvla][INFO] - Step 4744, Loss: {'train/loss': 0.001014917274005711, 'train/grad_norm_pre_clip': 1.5697731971740723, 'timing/data': 0.00024182999914046377, 'timing/model': 0.3663533520011697, 'learning_rate/qwen_vl_interface': 5.638375105231719e-07, 'learning_rate/action_model': 5.638375105231719e-06, 'epoch': 0.35}) +[2026-10-03 15:32:09,525][starVLA.training.train_starvla][INFO] - Step 4745, Loss: {'train/loss': 3.1439398298971355e-05, 'train/grad_norm_pre_clip': 0.00820454303175211, 'timing/data': 0.0002333100128453225, 'timing/model': 0.36601376099861227, 'learning_rate/qwen_vl_interface': 5.633408632856294e-07, 'learning_rate/action_model': 5.6334086328562945e-06, 'epoch': 0.35}) +[2026-10-03 15:32:10,371][starVLA.training.train_starvla][INFO] - Step 4746, Loss: {'train/loss': 0.5508898496627808, 'train/grad_norm_pre_clip': 46.2258415222168, 'timing/data': 0.0003110899997409433, 'timing/model': 0.5619031060050474, 'learning_rate/qwen_vl_interface': 5.628461425566268e-07, 'learning_rate/action_model': 5.6284614255662675e-06, 'epoch': 0.35}) +[2026-10-03 15:32:11,024][starVLA.training.train_starvla][INFO] - Step 4747, Loss: {'train/loss': 0.0002988862106576562, 'train/grad_norm_pre_clip': 0.09325376898050308, 'timing/data': 0.0002938599936896935, 'timing/model': 0.36661984099191613, 'learning_rate/qwen_vl_interface': 5.62353348539526e-07, 'learning_rate/action_model': 5.623533485395259e-06, 'epoch': 0.35}) +[2026-10-03 15:32:11,676][starVLA.training.train_starvla][INFO] - Step 4748, Loss: {'train/loss': 2.6955296561936848e-05, 'train/grad_norm_pre_clip': 0.010129624977707863, 'timing/data': 0.0002712299901759252, 'timing/model': 0.3669492210028693, 'learning_rate/qwen_vl_interface': 5.618624814368956e-07, 'learning_rate/action_model': 5.618624814368956e-06, 'epoch': 0.35}) +[2026-10-03 15:32:12,304][starVLA.training.train_starvla][INFO] - Step 4749, Loss: {'train/loss': 0.03623849153518677, 'train/grad_norm_pre_clip': 116.21119689941406, 'timing/data': 0.0003124899958493188, 'timing/model': 0.3656831609987421, 'learning_rate/qwen_vl_interface': 5.613735414505126e-07, 'learning_rate/action_model': 5.613735414505125e-06, 'epoch': 0.35}) +[2026-10-03 15:32:12,932][starVLA.training.train_starvla][INFO] - Step 4750, Loss: {'train/loss': 8.27006260806229e-06, 'train/grad_norm_pre_clip': 0.001586625468917191, 'timing/data': 0.00025378000282216817, 'timing/model': 0.34379374400305096, 'learning_rate/qwen_vl_interface': 5.608865287813617e-07, 'learning_rate/action_model': 5.608865287813616e-06, 'epoch': 0.35}) +[2026-10-03 15:32:13,563][starVLA.training.train_starvla][INFO] - Step 4751, Loss: {'train/loss': 0.21127444505691528, 'train/grad_norm_pre_clip': 101.13531494140625, 'timing/data': 0.0002780799986794591, 'timing/model': 0.3663797909975983, 'learning_rate/qwen_vl_interface': 5.604014436296362e-07, 'learning_rate/action_model': 5.604014436296362e-06, 'epoch': 0.35}) +[2026-10-03 15:32:14,414][starVLA.training.train_starvla][INFO] - Step 4752, Loss: {'train/loss': 0.0001129880765802227, 'train/grad_norm_pre_clip': 0.05063708499073982, 'timing/data': 0.0003026100021088496, 'timing/model': 0.36649469200347085, 'learning_rate/qwen_vl_interface': 5.599182861947348e-07, 'learning_rate/action_model': 5.599182861947347e-06, 'epoch': 0.35}) +[2026-10-03 15:32:15,065][starVLA.training.train_starvla][INFO] - Step 4753, Loss: {'train/loss': 2.7275504180579446e-05, 'train/grad_norm_pre_clip': 0.007260921411216259, 'timing/data': 0.00033115998667199165, 'timing/model': 0.36604766199889127, 'learning_rate/qwen_vl_interface': 5.594370566752665e-07, 'learning_rate/action_model': 5.5943705667526645e-06, 'epoch': 0.35}) +[2026-10-03 15:32:15,716][starVLA.training.train_starvla][INFO] - Step 4754, Loss: {'train/loss': 4.012159115518443e-05, 'train/grad_norm_pre_clip': 0.019197093322873116, 'timing/data': 0.00019608000002335757, 'timing/model': 0.36676701199030504, 'learning_rate/qwen_vl_interface': 5.589577552690456e-07, 'learning_rate/action_model': 5.589577552690456e-06, 'epoch': 0.35}) +[2026-10-03 15:32:16,369][starVLA.training.train_starvla][INFO] - Step 4755, Loss: {'train/loss': 0.00039829168235883117, 'train/grad_norm_pre_clip': 0.3615027964115143, 'timing/data': 0.00027651900018099695, 'timing/model': 0.36642005199973937, 'learning_rate/qwen_vl_interface': 5.584803821730961e-07, 'learning_rate/action_model': 5.58480382173096e-06, 'epoch': 0.35}) +[2026-10-03 15:32:17,019][starVLA.training.train_starvla][INFO] - Step 4756, Loss: {'train/loss': 0.003540937090292573, 'train/grad_norm_pre_clip': 12.165560722351074, 'timing/data': 0.0002920299884863198, 'timing/model': 0.366020370987826, 'learning_rate/qwen_vl_interface': 5.580049375836466e-07, 'learning_rate/action_model': 5.580049375836466e-06, 'epoch': 0.35}) +[2026-10-03 15:32:17,859][starVLA.training.train_starvla][INFO] - Step 4757, Loss: {'train/loss': 4.1666575270937756e-05, 'train/grad_norm_pre_clip': 0.01246584951877594, 'timing/data': 0.0002518700057407841, 'timing/model': 0.5561607760027982, 'learning_rate/qwen_vl_interface': 5.575314216961356e-07, 'learning_rate/action_model': 5.575314216961356e-06, 'epoch': 0.35}) +[2026-10-03 15:32:18,509][starVLA.training.train_starvla][INFO] - Step 4758, Loss: {'train/loss': 5.061499905423261e-05, 'train/grad_norm_pre_clip': 0.026354201138019562, 'timing/data': 0.00024352000036742538, 'timing/model': 0.36614400199323427, 'learning_rate/qwen_vl_interface': 5.570598347052073e-07, 'learning_rate/action_model': 5.570598347052072e-06, 'epoch': 0.35}) +[2026-10-03 15:32:19,162][starVLA.training.train_starvla][INFO] - Step 4759, Loss: {'train/loss': 4.817137960344553e-05, 'train/grad_norm_pre_clip': 0.13229429721832275, 'timing/data': 0.0002053299976978451, 'timing/model': 0.36746739099908154, 'learning_rate/qwen_vl_interface': 5.565901768047126e-07, 'learning_rate/action_model': 5.5659017680471255e-06, 'epoch': 0.35}) +[2026-10-03 15:32:19,811][starVLA.training.train_starvla][INFO] - Step 4760, Loss: {'train/loss': 0.0008568721241317689, 'train/grad_norm_pre_clip': 0.7888699173927307, 'timing/data': 0.00019690999761223793, 'timing/model': 0.3659089610009687, 'learning_rate/qwen_vl_interface': 5.561224481877105e-07, 'learning_rate/action_model': 5.561224481877105e-06, 'epoch': 0.35}) +[2026-10-03 15:32:20,462][starVLA.training.train_starvla][INFO] - Step 4761, Loss: {'train/loss': 0.004612869583070278, 'train/grad_norm_pre_clip': 5.569653511047363, 'timing/data': 0.0002869199961423874, 'timing/model': 0.366595690997201, 'learning_rate/qwen_vl_interface': 5.55656649046467e-07, 'learning_rate/action_model': 5.556566490464669e-06, 'epoch': 0.35}) +[2026-10-03 15:32:21,308][starVLA.training.train_starvla][INFO] - Step 4762, Loss: {'train/loss': 3.6963876482332125e-05, 'train/grad_norm_pre_clip': 0.020896155387163162, 'timing/data': 0.00020054999913554639, 'timing/model': 0.5617404659860767, 'learning_rate/qwen_vl_interface': 5.551927795724544e-07, 'learning_rate/action_model': 5.551927795724543e-06, 'epoch': 0.35}) +[2026-10-03 15:32:21,962][starVLA.training.train_starvla][INFO] - Step 4763, Loss: {'train/loss': 3.0417115340242162e-05, 'train/grad_norm_pre_clip': 0.015392833389341831, 'timing/data': 0.00020609000057447702, 'timing/model': 0.36842523100494873, 'learning_rate/qwen_vl_interface': 5.547308399563515e-07, 'learning_rate/action_model': 5.547308399563514e-06, 'epoch': 0.35}) +[2026-10-03 15:32:22,612][starVLA.training.train_starvla][INFO] - Step 4764, Loss: {'train/loss': 0.0007259934791363776, 'train/grad_norm_pre_clip': 1.3362383842468262, 'timing/data': 0.00019184000848326832, 'timing/model': 0.3664902210002765, 'learning_rate/qwen_vl_interface': 5.542708303880442e-07, 'learning_rate/action_model': 5.542708303880442e-06, 'epoch': 0.35}) +[2026-10-03 15:32:23,262][starVLA.training.train_starvla][INFO] - Step 4765, Loss: {'train/loss': 0.07434042543172836, 'train/grad_norm_pre_clip': 160.35617065429688, 'timing/data': 0.0001964999974006787, 'timing/model': 0.3662497909972444, 'learning_rate/qwen_vl_interface': 5.538127510566258e-07, 'learning_rate/action_model': 5.538127510566258e-06, 'epoch': 0.35}) +[2026-10-03 15:32:23,912][starVLA.training.train_starvla][INFO] - Step 4766, Loss: {'train/loss': 0.0026655616238713264, 'train/grad_norm_pre_clip': 1.526792049407959, 'timing/data': 0.0001935099862748757, 'timing/model': 0.36657494099927135, 'learning_rate/qwen_vl_interface': 5.533566021503944e-07, 'learning_rate/action_model': 5.533566021503943e-06, 'epoch': 0.35}) +[2026-10-03 15:32:24,566][starVLA.training.train_starvla][INFO] - Step 4767, Loss: {'train/loss': 5.473914643516764e-05, 'train/grad_norm_pre_clip': 0.02468971721827984, 'timing/data': 0.00020167000184301287, 'timing/model': 0.3674973710003542, 'learning_rate/qwen_vl_interface': 5.529023838568566e-07, 'learning_rate/action_model': 5.529023838568566e-06, 'epoch': 0.35}) +[2026-10-03 15:32:25,407][starVLA.training.train_starvla][INFO] - Step 4768, Loss: {'train/loss': 0.007023894693702459, 'train/grad_norm_pre_clip': 55.56501770019531, 'timing/data': 0.00020036000933032483, 'timing/model': 0.3663034219935071, 'learning_rate/qwen_vl_interface': 5.524500963627239e-07, 'learning_rate/action_model': 5.524500963627239e-06, 'epoch': 0.35}) +[2026-10-03 15:32:26,056][starVLA.training.train_starvla][INFO] - Step 4769, Loss: {'train/loss': 2.2083011572249234e-05, 'train/grad_norm_pre_clip': 0.005188894458115101, 'timing/data': 0.00024002999998629093, 'timing/model': 0.36598534100630786, 'learning_rate/qwen_vl_interface': 5.519997398539145e-07, 'learning_rate/action_model': 5.519997398539145e-06, 'epoch': 0.35}) +[2026-10-03 15:32:26,710][starVLA.training.train_starvla][INFO] - Step 4770, Loss: {'train/loss': 7.394235581159592e-05, 'train/grad_norm_pre_clip': 0.06130843609571457, 'timing/data': 0.00024418000248260796, 'timing/model': 0.3702029309934005, 'learning_rate/qwen_vl_interface': 5.515513145155531e-07, 'learning_rate/action_model': 5.515513145155531e-06, 'epoch': 0.35}) +[2026-10-03 15:32:27,364][starVLA.training.train_starvla][INFO] - Step 4771, Loss: {'train/loss': 0.3777236342430115, 'train/grad_norm_pre_clip': 64.68045806884766, 'timing/data': 0.0002835500054061413, 'timing/model': 0.3678850010037422, 'learning_rate/qwen_vl_interface': 5.511048205319702e-07, 'learning_rate/action_model': 5.511048205319701e-06, 'epoch': 0.35}) +[2026-10-03 15:32:28,015][starVLA.training.train_starvla][INFO] - Step 4772, Loss: {'train/loss': 3.7998033803887665e-05, 'train/grad_norm_pre_clip': 0.02705441229045391, 'timing/data': 0.000341070001013577, 'timing/model': 0.36580040099215694, 'learning_rate/qwen_vl_interface': 5.506602580867034e-07, 'learning_rate/action_model': 5.506602580867034e-06, 'epoch': 0.35}) +[2026-10-03 15:32:28,841][starVLA.training.train_starvla][INFO] - Step 4773, Loss: {'train/loss': 0.3461659550666809, 'train/grad_norm_pre_clip': 43.685394287109375, 'timing/data': 0.0002498199901310727, 'timing/model': 0.36577049100014847, 'learning_rate/qwen_vl_interface': 5.502176273624942e-07, 'learning_rate/action_model': 5.502176273624942e-06, 'epoch': 0.35}) +[2026-10-03 15:32:29,494][starVLA.training.train_starvla][INFO] - Step 4774, Loss: {'train/loss': 0.0008432099130004644, 'train/grad_norm_pre_clip': 0.21534185111522675, 'timing/data': 0.0001938399946084246, 'timing/model': 0.36792163099744357, 'learning_rate/qwen_vl_interface': 5.497769285412921e-07, 'learning_rate/action_model': 5.497769285412921e-06, 'epoch': 0.35}) +[2026-10-03 15:32:30,125][starVLA.training.train_starvla][INFO] - Step 4775, Loss: {'train/loss': 0.00024872488575056195, 'train/grad_norm_pre_clip': 0.12357202917337418, 'timing/data': 0.00022108999837655574, 'timing/model': 0.3679065320029622, 'learning_rate/qwen_vl_interface': 5.493381618042519e-07, 'learning_rate/action_model': 5.493381618042518e-06, 'epoch': 0.35}) +[2026-10-03 15:32:30,753][starVLA.training.train_starvla][INFO] - Step 4776, Loss: {'train/loss': 0.00010406888759462163, 'train/grad_norm_pre_clip': 0.04671170562505722, 'timing/data': 0.00019885000074282289, 'timing/model': 0.34315096501086373, 'learning_rate/qwen_vl_interface': 5.489013273317343e-07, 'learning_rate/action_model': 5.489013273317343e-06, 'epoch': 0.35}) +[2026-10-03 15:32:31,404][starVLA.training.train_starvla][INFO] - Step 4777, Loss: {'train/loss': 0.0005049339961260557, 'train/grad_norm_pre_clip': 0.14968881011009216, 'timing/data': 0.00024723999376874417, 'timing/model': 0.36616362098720856, 'learning_rate/qwen_vl_interface': 5.484664253033042e-07, 'learning_rate/action_model': 5.484664253033042e-06, 'epoch': 0.35}) +[2026-10-03 15:32:32,247][starVLA.training.train_starvla][INFO] - Step 4778, Loss: {'train/loss': 0.00010882740025408566, 'train/grad_norm_pre_clip': 0.08032497763633728, 'timing/data': 0.0002529600023990497, 'timing/model': 0.5591146249935264, 'learning_rate/qwen_vl_interface': 5.480334558977351e-07, 'learning_rate/action_model': 5.480334558977351e-06, 'epoch': 0.35}) +[2026-10-03 15:32:32,900][starVLA.training.train_starvla][INFO] - Step 4779, Loss: {'train/loss': 0.001463688793592155, 'train/grad_norm_pre_clip': 0.7128722071647644, 'timing/data': 0.0002751599968178198, 'timing/model': 0.36714892101008445, 'learning_rate/qwen_vl_interface': 5.476024192930029e-07, 'learning_rate/action_model': 5.47602419293003e-06, 'epoch': 0.35}) +[2026-10-03 15:32:33,548][starVLA.training.train_starvla][INFO] - Step 4780, Loss: {'train/loss': 7.472883226000704e-06, 'train/grad_norm_pre_clip': 0.0014606972690671682, 'timing/data': 0.0001983599941013381, 'timing/model': 0.3645541720034089, 'learning_rate/qwen_vl_interface': 5.471733156662914e-07, 'learning_rate/action_model': 5.4717331566629135e-06, 'epoch': 0.35}) +[2026-10-03 15:32:34,196][starVLA.training.train_starvla][INFO] - Step 4781, Loss: {'train/loss': 1.534042712592054e-05, 'train/grad_norm_pre_clip': 0.003980790730565786, 'timing/data': 0.0001961599919013679, 'timing/model': 0.3657130820065504, 'learning_rate/qwen_vl_interface': 5.467461451939882e-07, 'learning_rate/action_model': 5.467461451939882e-06, 'epoch': 0.35}) +[2026-10-03 15:32:34,845][starVLA.training.train_starvla][INFO] - Step 4782, Loss: {'train/loss': 1.4215491319191642e-05, 'train/grad_norm_pre_clip': 0.004274400416761637, 'timing/data': 0.00019162001262884587, 'timing/model': 0.3666475909994915, 'learning_rate/qwen_vl_interface': 5.463209080516873e-07, 'learning_rate/action_model': 5.463209080516873e-06, 'epoch': 0.35}) +[2026-10-03 15:32:35,497][starVLA.training.train_starvla][INFO] - Step 4783, Loss: {'train/loss': 7.857641321606934e-05, 'train/grad_norm_pre_clip': 0.03661763295531273, 'timing/data': 0.00023891001183073968, 'timing/model': 0.36681002100522164, 'learning_rate/qwen_vl_interface': 5.458976044141875e-07, 'learning_rate/action_model': 5.458976044141875e-06, 'epoch': 0.35}) +[2026-10-03 15:32:36,333][starVLA.training.train_starvla][INFO] - Step 4784, Loss: {'train/loss': 0.3586907684803009, 'train/grad_norm_pre_clip': 125.0657958984375, 'timing/data': 0.0002769799903035164, 'timing/model': 0.36443294200580567, 'learning_rate/qwen_vl_interface': 5.454762344554931e-07, 'learning_rate/action_model': 5.4547623445549316e-06, 'epoch': 0.35}) +[2026-10-03 15:32:36,958][starVLA.training.train_starvla][INFO] - Step 4785, Loss: {'train/loss': 0.12662114202976227, 'train/grad_norm_pre_clip': 169.75473022460938, 'timing/data': 0.00019258000247646123, 'timing/model': 0.3640256620128639, 'learning_rate/qwen_vl_interface': 5.450567983488137e-07, 'learning_rate/action_model': 5.450567983488137e-06, 'epoch': 0.35}) +[2026-10-03 15:32:37,607][starVLA.training.train_starvla][INFO] - Step 4786, Loss: {'train/loss': 0.0004128179862163961, 'train/grad_norm_pre_clip': 0.18038150668144226, 'timing/data': 0.00029830000130459666, 'timing/model': 0.36646897200262174, 'learning_rate/qwen_vl_interface': 5.446392962665629e-07, 'learning_rate/action_model': 5.4463929626656296e-06, 'epoch': 0.35}) +[2026-10-03 15:32:38,259][starVLA.training.train_starvla][INFO] - Step 4787, Loss: {'train/loss': 2.976317955472041e-05, 'train/grad_norm_pre_clip': 0.006937466096132994, 'timing/data': 0.0002533999941078946, 'timing/model': 0.3671146209962899, 'learning_rate/qwen_vl_interface': 5.442237283803605e-07, 'learning_rate/action_model': 5.442237283803604e-06, 'epoch': 0.35}) +[2026-10-03 15:32:39,095][starVLA.training.train_starvla][INFO] - Step 4788, Loss: {'train/loss': 0.00010893375292653218, 'train/grad_norm_pre_clip': 0.10033673048019409, 'timing/data': 0.0002513499930500984, 'timing/model': 0.552274257002864, 'learning_rate/qwen_vl_interface': 5.43810094861031e-07, 'learning_rate/action_model': 5.43810094861031e-06, 'epoch': 0.35}) +[2026-10-03 15:32:39,743][starVLA.training.train_starvla][INFO] - Step 4789, Loss: {'train/loss': 0.06738585233688354, 'train/grad_norm_pre_clip': 197.21498107910156, 'timing/data': 0.00022976000036578625, 'timing/model': 0.36484740099695046, 'learning_rate/qwen_vl_interface': 5.433983958786027e-07, 'learning_rate/action_model': 5.433983958786027e-06, 'epoch': 0.35}) +[2026-10-03 15:32:40,393][starVLA.training.train_starvla][INFO] - Step 4790, Loss: {'train/loss': 0.0002468091552145779, 'train/grad_norm_pre_clip': 0.15720631182193756, 'timing/data': 0.00023195000539999455, 'timing/model': 0.36657310200098436, 'learning_rate/qwen_vl_interface': 5.429886316023102e-07, 'learning_rate/action_model': 5.429886316023101e-06, 'epoch': 0.35}) +[2026-10-03 15:32:41,044][starVLA.training.train_starvla][INFO] - Step 4791, Loss: {'train/loss': 0.00016537701594643295, 'train/grad_norm_pre_clip': 0.09207221865653992, 'timing/data': 0.00026793000870384276, 'timing/model': 0.3668939019989921, 'learning_rate/qwen_vl_interface': 5.425808022005916e-07, 'learning_rate/action_model': 5.425808022005916e-06, 'epoch': 0.35}) +[2026-10-03 15:32:41,692][starVLA.training.train_starvla][INFO] - Step 4792, Loss: {'train/loss': 0.3742538392543793, 'train/grad_norm_pre_clip': 50.34843063354492, 'timing/data': 0.00030989899823907763, 'timing/model': 0.3645192620024318, 'learning_rate/qwen_vl_interface': 5.421749078410903e-07, 'learning_rate/action_model': 5.421749078410903e-06, 'epoch': 0.35}) +[2026-10-03 15:32:42,340][starVLA.training.train_starvla][INFO] - Step 4793, Loss: {'train/loss': 0.8689383268356323, 'train/grad_norm_pre_clip': 18.427228927612305, 'timing/data': 0.00019762999727390707, 'timing/model': 0.3648899310064735, 'learning_rate/qwen_vl_interface': 5.417709486906547e-07, 'learning_rate/action_model': 5.417709486906547e-06, 'epoch': 0.35}) +[2026-10-03 15:32:43,184][starVLA.training.train_starvla][INFO] - Step 4794, Loss: {'train/loss': 0.3528929650783539, 'train/grad_norm_pre_clip': 129.02691650390625, 'timing/data': 0.0002250199904665351, 'timing/model': 0.3664075020060409, 'learning_rate/qwen_vl_interface': 5.413689249153367e-07, 'learning_rate/action_model': 5.413689249153366e-06, 'epoch': 0.35}) +[2026-10-03 15:32:43,835][starVLA.training.train_starvla][INFO] - Step 4795, Loss: {'train/loss': 0.00016397649596910924, 'train/grad_norm_pre_clip': 0.0784730464220047, 'timing/data': 0.0002384489926043898, 'timing/model': 0.3668306519975886, 'learning_rate/qwen_vl_interface': 5.409688366803931e-07, 'learning_rate/action_model': 5.409688366803931e-06, 'epoch': 0.35}) +[2026-10-03 15:32:44,482][starVLA.training.train_starvla][INFO] - Step 4796, Loss: {'train/loss': 1.3544852663471829e-05, 'train/grad_norm_pre_clip': 0.0030666766688227654, 'timing/data': 0.0002085699961753562, 'timing/model': 0.36436881199188065, 'learning_rate/qwen_vl_interface': 5.405706841502853e-07, 'learning_rate/action_model': 5.4057068415028525e-06, 'epoch': 0.35}) +[2026-10-03 15:32:45,106][starVLA.training.train_starvla][INFO] - Step 4797, Loss: {'train/loss': 6.018016938469373e-05, 'train/grad_norm_pre_clip': 0.03264068067073822, 'timing/data': 0.00019003999477718025, 'timing/model': 0.3640869520022534, 'learning_rate/qwen_vl_interface': 5.401744674886781e-07, 'learning_rate/action_model': 5.40174467488678e-06, 'epoch': 0.35}) +[2026-10-03 15:32:45,755][starVLA.training.train_starvla][INFO] - Step 4798, Loss: {'train/loss': 0.00019895643345080316, 'train/grad_norm_pre_clip': 0.11910871416330338, 'timing/data': 0.0002040699910139665, 'timing/model': 0.3662975409970386, 'learning_rate/qwen_vl_interface': 5.397801868584418e-07, 'learning_rate/action_model': 5.397801868584417e-06, 'epoch': 0.35}) +[2026-10-03 15:32:46,594][starVLA.training.train_starvla][INFO] - Step 4799, Loss: {'train/loss': 0.0002667791268322617, 'train/grad_norm_pre_clip': 0.11086776852607727, 'timing/data': 0.00020871999731753021, 'timing/model': 0.554087607000838, 'learning_rate/qwen_vl_interface': 5.393878424216502e-07, 'learning_rate/action_model': 5.393878424216501e-06, 'epoch': 0.35}) +[2026-10-03 15:32:47,241][starVLA.training.train_starvla][INFO] - Step 4800, Loss: {'train/loss': 0.0008839954971335828, 'train/grad_norm_pre_clip': 1.0963646173477173, 'timing/data': 0.0002346800029044971, 'timing/model': 0.3642774809995899, 'learning_rate/qwen_vl_interface': 5.389974343395813e-07, 'learning_rate/action_model': 5.389974343395812e-06, 'epoch': 0.35}) +[2026-10-03 15:32:47,887][starVLA.training.train_starvla][INFO] - Step 4801, Loss: {'train/loss': 5.540720303542912e-05, 'train/grad_norm_pre_clip': 0.015262902714312077, 'timing/data': 0.00018959000590257347, 'timing/model': 0.36468096199678257, 'learning_rate/qwen_vl_interface': 5.38608962772717e-07, 'learning_rate/action_model': 5.38608962772717e-06, 'epoch': 0.35}) +[2026-10-03 15:32:48,537][starVLA.training.train_starvla][INFO] - Step 4802, Loss: {'train/loss': 3.7531321140704677e-05, 'train/grad_norm_pre_clip': 0.021765340119600296, 'timing/data': 0.00024027000472415239, 'timing/model': 0.36636118100432213, 'learning_rate/qwen_vl_interface': 5.382224278807429e-07, 'learning_rate/action_model': 5.382224278807429e-06, 'epoch': 0.35}) +[2026-10-03 15:32:49,188][starVLA.training.train_starvla][INFO] - Step 4803, Loss: {'train/loss': 5.646309000439942e-05, 'train/grad_norm_pre_clip': 0.03266431763768196, 'timing/data': 0.00025432000984437764, 'timing/model': 0.36683796100260224, 'learning_rate/qwen_vl_interface': 5.378378298225496e-07, 'learning_rate/action_model': 5.378378298225496e-06, 'epoch': 0.35}) +[2026-10-03 15:32:50,051][starVLA.training.train_starvla][INFO] - Step 4804, Loss: {'train/loss': 0.0002144236641470343, 'train/grad_norm_pre_clip': 0.10295092314481735, 'timing/data': 0.0003956299915444106, 'timing/model': 0.5721584139973857, 'learning_rate/qwen_vl_interface': 5.374551687562306e-07, 'learning_rate/action_model': 5.374551687562306e-06, 'epoch': 0.35}) +[2026-10-03 15:32:50,711][starVLA.training.train_starvla][INFO] - Step 4805, Loss: {'train/loss': 0.013933368027210236, 'train/grad_norm_pre_clip': 24.78022003173828, 'timing/data': 0.0004027399991173297, 'timing/model': 0.37137144000735134, 'learning_rate/qwen_vl_interface': 5.37074444839083e-07, 'learning_rate/action_model': 5.370744448390829e-06, 'epoch': 0.35}) +[2026-10-03 15:32:51,348][starVLA.training.train_starvla][INFO] - Step 4806, Loss: {'train/loss': 2.7454234441393055e-05, 'train/grad_norm_pre_clip': 0.010283200070261955, 'timing/data': 0.0003040000010514632, 'timing/model': 0.3478017940069549, 'learning_rate/qwen_vl_interface': 5.366956582276085e-07, 'learning_rate/action_model': 5.366956582276084e-06, 'epoch': 0.35}) +[2026-10-03 15:32:52,007][starVLA.training.train_starvla][INFO] - Step 4807, Loss: {'train/loss': 2.8756679967045784e-05, 'train/grad_norm_pre_clip': 0.014457624405622482, 'timing/data': 0.00029476999770849943, 'timing/model': 0.37010883100447245, 'learning_rate/qwen_vl_interface': 5.363188090775122e-07, 'learning_rate/action_model': 5.363188090775122e-06, 'epoch': 0.35}) +[2026-10-03 15:32:52,668][starVLA.training.train_starvla][INFO] - Step 4808, Loss: {'train/loss': 0.00043629080755636096, 'train/grad_norm_pre_clip': 0.0821104496717453, 'timing/data': 0.000361530008376576, 'timing/model': 0.372289801001898, 'learning_rate/qwen_vl_interface': 5.359438975437018e-07, 'learning_rate/action_model': 5.359438975437017e-06, 'epoch': 0.35}) +[2026-10-03 15:32:53,328][starVLA.training.train_starvla][INFO] - Step 4809, Loss: {'train/loss': 1.9102395526715554e-05, 'train/grad_norm_pre_clip': 0.0070462096482515335, 'timing/data': 0.00027148000663146377, 'timing/model': 0.37087472100392915, 'learning_rate/qwen_vl_interface': 5.355709237802895e-07, 'learning_rate/action_model': 5.3557092378028955e-06, 'epoch': 0.35}) +[2026-10-03 15:32:54,179][starVLA.training.train_starvla][INFO] - Step 4810, Loss: {'train/loss': 6.648516864515841e-05, 'train/grad_norm_pre_clip': 0.03034024126827717, 'timing/data': 0.00021934999676886946, 'timing/model': 0.3708031319984002, 'learning_rate/qwen_vl_interface': 5.351998879405911e-07, 'learning_rate/action_model': 5.351998879405911e-06, 'epoch': 0.35}) +[2026-10-03 15:32:54,839][starVLA.training.train_starvla][INFO] - Step 4811, Loss: {'train/loss': 5.05269636050798e-05, 'train/grad_norm_pre_clip': 0.0219217911362648, 'timing/data': 0.00030282000079751015, 'timing/model': 0.3703845209965948, 'learning_rate/qwen_vl_interface': 5.348307901771256e-07, 'learning_rate/action_model': 5.348307901771256e-06, 'epoch': 0.35}) +[2026-10-03 15:32:55,498][starVLA.training.train_starvla][INFO] - Step 4812, Loss: {'train/loss': 3.1268300517695025e-05, 'train/grad_norm_pre_clip': 0.010117631405591965, 'timing/data': 0.00021050999930594116, 'timing/model': 0.37131747099920176, 'learning_rate/qwen_vl_interface': 5.344636306416147e-07, 'learning_rate/action_model': 5.344636306416146e-06, 'epoch': 0.35}) +[2026-10-03 15:32:56,159][starVLA.training.train_starvla][INFO] - Step 4813, Loss: {'train/loss': 0.0006661778315901756, 'train/grad_norm_pre_clip': 0.1273280829191208, 'timing/data': 0.00037758999678771943, 'timing/model': 0.37089185100921895, 'learning_rate/qwen_vl_interface': 5.340984094849836e-07, 'learning_rate/action_model': 5.340984094849836e-06, 'epoch': 0.35}) +[2026-10-03 15:32:56,817][starVLA.training.train_starvla][INFO] - Step 4814, Loss: {'train/loss': 0.023732544854283333, 'train/grad_norm_pre_clip': 50.189605712890625, 'timing/data': 0.0003175000019837171, 'timing/model': 0.3694341810041806, 'learning_rate/qwen_vl_interface': 5.337351268573606e-07, 'learning_rate/action_model': 5.337351268573606e-06, 'epoch': 0.35}) +[2026-10-03 15:32:57,674][starVLA.training.train_starvla][INFO] - Step 4815, Loss: {'train/loss': 0.009144306182861328, 'train/grad_norm_pre_clip': 30.95081901550293, 'timing/data': 0.00022007999359630048, 'timing/model': 0.3703273610008182, 'learning_rate/qwen_vl_interface': 5.33373782908079e-07, 'learning_rate/action_model': 5.333737829080789e-06, 'epoch': 0.35}) +[2026-10-03 15:32:58,308][starVLA.training.train_starvla][INFO] - Step 4816, Loss: {'train/loss': 1.969072582141962e-05, 'train/grad_norm_pre_clip': 0.008736268617212772, 'timing/data': 0.0003226100088795647, 'timing/model': 0.347215513989795, 'learning_rate/qwen_vl_interface': 5.330143777856718e-07, 'learning_rate/action_model': 5.330143777856718e-06, 'epoch': 0.35}) +[2026-10-03 15:32:58,969][starVLA.training.train_starvla][INFO] - Step 4817, Loss: {'train/loss': 6.315043719951063e-05, 'train/grad_norm_pre_clip': 0.02942630462348461, 'timing/data': 0.00022106000687927008, 'timing/model': 0.37099997099721804, 'learning_rate/qwen_vl_interface': 5.326569116378786e-07, 'learning_rate/action_model': 5.326569116378786e-06, 'epoch': 0.35}) +[2026-10-03 15:32:59,628][starVLA.training.train_starvla][INFO] - Step 4818, Loss: {'train/loss': 1.0875487327575684, 'train/grad_norm_pre_clip': 56.92979431152344, 'timing/data': 0.00032094000198412687, 'timing/model': 0.37023024099471513, 'learning_rate/qwen_vl_interface': 5.323013846116387e-07, 'learning_rate/action_model': 5.323013846116386e-06, 'epoch': 0.35}) +[2026-10-03 15:33:00,288][starVLA.training.train_starvla][INFO] - Step 4819, Loss: {'train/loss': 0.00016192957991734147, 'train/grad_norm_pre_clip': 0.0747784674167633, 'timing/data': 0.00021207000827416778, 'timing/model': 0.3710659409989603, 'learning_rate/qwen_vl_interface': 5.31947796853097e-07, 'learning_rate/action_model': 5.3194779685309696e-06, 'epoch': 0.35}) +[2026-10-03 15:33:01,142][starVLA.training.train_starvla][INFO] - Step 4820, Loss: {'train/loss': 9.311967733083293e-05, 'train/grad_norm_pre_clip': 0.03325246647000313, 'timing/data': 0.00025143999664578587, 'timing/model': 0.5658235949958907, 'learning_rate/qwen_vl_interface': 5.315961485075992e-07, 'learning_rate/action_model': 5.315961485075993e-06, 'epoch': 0.35}) +[2026-10-03 15:33:01,803][starVLA.training.train_starvla][INFO] - Step 4821, Loss: {'train/loss': 6.21205399511382e-05, 'train/grad_norm_pre_clip': 0.027237046509981155, 'timing/data': 0.000302100001135841, 'timing/model': 0.371466270997189, 'learning_rate/qwen_vl_interface': 5.312464397196958e-07, 'learning_rate/action_model': 5.312464397196958e-06, 'epoch': 0.35}) +[2026-10-03 15:33:02,461][starVLA.training.train_starvla][INFO] - Step 4822, Loss: {'train/loss': 9.730221790960059e-06, 'train/grad_norm_pre_clip': 0.006073443219065666, 'timing/data': 0.0003034800029126927, 'timing/model': 0.36964875100238714, 'learning_rate/qwen_vl_interface': 5.308986706331377e-07, 'learning_rate/action_model': 5.308986706331376e-06, 'epoch': 0.35}) +[2026-10-03 15:33:03,120][starVLA.training.train_starvla][INFO] - Step 4823, Loss: {'train/loss': 0.0003585680387914181, 'train/grad_norm_pre_clip': 0.3259008824825287, 'timing/data': 0.00021253999148029834, 'timing/model': 0.37032903100771364, 'learning_rate/qwen_vl_interface': 5.305528413908797e-07, 'learning_rate/action_model': 5.305528413908798e-06, 'epoch': 0.35}) +[2026-10-03 15:33:03,780][starVLA.training.train_starvla][INFO] - Step 4824, Loss: {'train/loss': 0.00038753298576921225, 'train/grad_norm_pre_clip': 0.1391151249408722, 'timing/data': 0.0002585600013844669, 'timing/model': 0.37162479100516066, 'learning_rate/qwen_vl_interface': 5.302089521350799e-07, 'learning_rate/action_model': 5.3020895213507996e-06, 'epoch': 0.35}) +[2026-10-03 15:33:04,636][starVLA.training.train_starvla][INFO] - Step 4825, Loss: {'train/loss': 4.363018524600193e-05, 'train/grad_norm_pre_clip': 0.01424342766404152, 'timing/data': 0.0002540400018915534, 'timing/model': 0.56669330499426, 'learning_rate/qwen_vl_interface': 5.298670030070977e-07, 'learning_rate/action_model': 5.298670030070976e-06, 'epoch': 0.35}) +[2026-10-03 15:33:05,295][starVLA.training.train_starvla][INFO] - Step 4826, Loss: {'train/loss': 3.733907578862272e-05, 'train/grad_norm_pre_clip': 0.011083549819886684, 'timing/data': 0.00021652999566867948, 'timing/model': 0.37015356100164354, 'learning_rate/qwen_vl_interface': 5.295269941474957e-07, 'learning_rate/action_model': 5.295269941474957e-06, 'epoch': 0.35}) +[2026-10-03 15:33:05,954][starVLA.training.train_starvla][INFO] - Step 4827, Loss: {'train/loss': 0.13192139565944672, 'train/grad_norm_pre_clip': 173.7788543701172, 'timing/data': 0.0002650099922902882, 'timing/model': 0.37033173099916894, 'learning_rate/qwen_vl_interface': 5.291889256960384e-07, 'learning_rate/action_model': 5.291889256960384e-06, 'epoch': 0.35}) +[2026-10-03 15:33:06,613][starVLA.training.train_starvla][INFO] - Step 4828, Loss: {'train/loss': 0.048057861626148224, 'train/grad_norm_pre_clip': 143.63795471191406, 'timing/data': 0.0002580199943622574, 'timing/model': 0.37113347100967076, 'learning_rate/qwen_vl_interface': 5.288527977916929e-07, 'learning_rate/action_model': 5.288527977916929e-06, 'epoch': 0.35}) +[2026-10-03 15:33:07,250][starVLA.training.train_starvla][INFO] - Step 4829, Loss: {'train/loss': 1.6532161680515856e-05, 'train/grad_norm_pre_clip': 0.007985190488398075, 'timing/data': 0.00026889999571722, 'timing/model': 0.37129764100245666, 'learning_rate/qwen_vl_interface': 5.28518610572629e-07, 'learning_rate/action_model': 5.28518610572629e-06, 'epoch': 0.35}) +[2026-10-03 15:33:07,910][starVLA.training.train_starvla][INFO] - Step 4830, Loss: {'train/loss': 0.3767006993293762, 'train/grad_norm_pre_clip': 88.70420837402344, 'timing/data': 0.0003267299907747656, 'timing/model': 0.36982097099826206, 'learning_rate/qwen_vl_interface': 5.281863641762184e-07, 'learning_rate/action_model': 5.281863641762184e-06, 'epoch': 0.35}) +[2026-10-03 15:33:08,759][starVLA.training.train_starvla][INFO] - Step 4831, Loss: {'train/loss': 2.5300967536168173e-05, 'train/grad_norm_pre_clip': 0.010383222252130508, 'timing/data': 0.0003369200130691752, 'timing/model': 0.3705947909911629, 'learning_rate/qwen_vl_interface': 5.278560587390355e-07, 'learning_rate/action_model': 5.278560587390355e-06, 'epoch': 0.35}) +[2026-10-03 15:33:09,418][starVLA.training.train_starvla][INFO] - Step 4832, Loss: {'train/loss': 0.01626310683786869, 'train/grad_norm_pre_clip': 26.261207580566406, 'timing/data': 0.0003552799898898229, 'timing/model': 0.3709603109891759, 'learning_rate/qwen_vl_interface': 5.275276943968557e-07, 'learning_rate/action_model': 5.275276943968557e-06, 'epoch': 0.35}) +[2026-10-03 15:33:10,078][starVLA.training.train_starvla][INFO] - Step 4833, Loss: {'train/loss': 5.85613133807783e-06, 'train/grad_norm_pre_clip': 0.0010292035294696689, 'timing/data': 0.0003360999980941415, 'timing/model': 0.3710875710094115, 'learning_rate/qwen_vl_interface': 5.272012712846566e-07, 'learning_rate/action_model': 5.2720127128465654e-06, 'epoch': 0.35}) +[2026-10-03 15:33:10,738][starVLA.training.train_starvla][INFO] - Step 4834, Loss: {'train/loss': 0.050297465175390244, 'train/grad_norm_pre_clip': 176.03221130371094, 'timing/data': 0.00025040999753400683, 'timing/model': 0.3698784820007859, 'learning_rate/qwen_vl_interface': 5.268767895366198e-07, 'learning_rate/action_model': 5.268767895366198e-06, 'epoch': 0.35}) +[2026-10-03 15:33:11,396][starVLA.training.train_starvla][INFO] - Step 4835, Loss: {'train/loss': 0.012924234382808208, 'train/grad_norm_pre_clip': 47.10880661010742, 'timing/data': 0.0002081099955830723, 'timing/model': 0.36978344099770766, 'learning_rate/qwen_vl_interface': 5.265542492861268e-07, 'learning_rate/action_model': 5.2655424928612676e-06, 'epoch': 0.35}) +[2026-10-03 15:33:12,247][starVLA.training.train_starvla][INFO] - Step 4836, Loss: {'train/loss': 3.255515912314877e-05, 'train/grad_norm_pre_clip': 0.015644241124391556, 'timing/data': 0.0003123789938399568, 'timing/model': 0.37049489200580865, 'learning_rate/qwen_vl_interface': 5.262336506657617e-07, 'learning_rate/action_model': 5.2623365066576165e-06, 'epoch': 0.35}) +[2026-10-03 15:33:12,907][starVLA.training.train_starvla][INFO] - Step 4837, Loss: {'train/loss': 0.000207630917429924, 'train/grad_norm_pre_clip': 0.10422128438949585, 'timing/data': 0.00021853999351151288, 'timing/model': 0.3711915619933279, 'learning_rate/qwen_vl_interface': 5.259149938073103e-07, 'learning_rate/action_model': 5.259149938073102e-06, 'epoch': 0.35}) +[2026-10-03 15:33:13,566][starVLA.training.train_starvla][INFO] - Step 4838, Loss: {'train/loss': 7.284953608177602e-05, 'train/grad_norm_pre_clip': 0.0254326481372118, 'timing/data': 0.0003087600052822381, 'timing/model': 0.3701728410087526, 'learning_rate/qwen_vl_interface': 5.255982788417609e-07, 'learning_rate/action_model': 5.255982788417608e-06, 'epoch': 0.35}) +[2026-10-03 15:33:14,223][starVLA.training.train_starvla][INFO] - Step 4839, Loss: {'train/loss': 0.0002334623277420178, 'train/grad_norm_pre_clip': 0.20678764581680298, 'timing/data': 0.00026805000379681587, 'timing/model': 0.36982476100092754, 'learning_rate/qwen_vl_interface': 5.252835058993029e-07, 'learning_rate/action_model': 5.252835058993029e-06, 'epoch': 0.35}) +[2026-10-03 15:33:14,881][starVLA.training.train_starvla][INFO] - Step 4840, Loss: {'train/loss': 0.00011206719500478357, 'train/grad_norm_pre_clip': 0.05229543149471283, 'timing/data': 0.00027737999334931374, 'timing/model': 0.3698630709986901, 'learning_rate/qwen_vl_interface': 5.249706751093274e-07, 'learning_rate/action_model': 5.249706751093273e-06, 'epoch': 0.35}) +[2026-10-03 15:33:15,731][starVLA.training.train_starvla][INFO] - Step 4841, Loss: {'train/loss': 3.627988917287439e-05, 'train/grad_norm_pre_clip': 0.011527634225785732, 'timing/data': 0.0002970200002891943, 'timing/model': 0.5607397159910761, 'learning_rate/qwen_vl_interface': 5.246597866004273e-07, 'learning_rate/action_model': 5.246597866004273e-06, 'epoch': 0.35}) +[2026-10-03 15:33:16,366][starVLA.training.train_starvla][INFO] - Step 4842, Loss: {'train/loss': 4.0687169530428946e-05, 'train/grad_norm_pre_clip': 0.016635490581393242, 'timing/data': 0.0002634799893712625, 'timing/model': 0.36928418099705596, 'learning_rate/qwen_vl_interface': 5.243508405003975e-07, 'learning_rate/action_model': 5.243508405003975e-06, 'epoch': 0.35}) +[2026-10-03 15:33:17,023][starVLA.training.train_starvla][INFO] - Step 4843, Loss: {'train/loss': 0.01086094044148922, 'train/grad_norm_pre_clip': 42.6307373046875, 'timing/data': 0.0003216799959773198, 'timing/model': 0.3692772310023429, 'learning_rate/qwen_vl_interface': 5.240438369362336e-07, 'learning_rate/action_model': 5.2404383693623366e-06, 'epoch': 0.35}) +[2026-10-03 15:33:17,682][starVLA.training.train_starvla][INFO] - Step 4844, Loss: {'train/loss': 0.3202460706233978, 'train/grad_norm_pre_clip': 17.871809005737305, 'timing/data': 0.0003114889987045899, 'timing/model': 0.3706551320065046, 'learning_rate/qwen_vl_interface': 5.237387760341335e-07, 'learning_rate/action_model': 5.237387760341335e-06, 'epoch': 0.35}) +[2026-10-03 15:33:18,319][starVLA.training.train_starvla][INFO] - Step 4845, Loss: {'train/loss': 0.0002406923595117405, 'train/grad_norm_pre_clip': 0.10359805822372437, 'timing/data': 0.0002701500052353367, 'timing/model': 0.3705878110049525, 'learning_rate/qwen_vl_interface': 5.234356579194963e-07, 'learning_rate/action_model': 5.234356579194963e-06, 'epoch': 0.35}) +[2026-10-03 15:33:19,144][starVLA.training.train_starvla][INFO] - Step 4846, Loss: {'train/loss': 3.0508566851494834e-05, 'train/grad_norm_pre_clip': 0.01057747844606638, 'timing/data': 0.00036985000770073384, 'timing/model': 0.5358462790027261, 'learning_rate/qwen_vl_interface': 5.231344827169218e-07, 'learning_rate/action_model': 5.231344827169217e-06, 'epoch': 0.35}) +[2026-10-03 15:33:19,801][starVLA.training.train_starvla][INFO] - Step 4847, Loss: {'train/loss': 3.7302106647985056e-05, 'train/grad_norm_pre_clip': 0.012656092643737793, 'timing/data': 0.0002570000069681555, 'timing/model': 0.36967461100721266, 'learning_rate/qwen_vl_interface': 5.228352505502119e-07, 'learning_rate/action_model': 5.228352505502119e-06, 'epoch': 0.35}) +[2026-10-03 15:33:20,459][starVLA.training.train_starvla][INFO] - Step 4848, Loss: {'train/loss': 1.8998525774804875e-05, 'train/grad_norm_pre_clip': 0.0039912620559334755, 'timing/data': 0.0003069200029131025, 'timing/model': 0.370278080998105, 'learning_rate/qwen_vl_interface': 5.225379615423701e-07, 'learning_rate/action_model': 5.2253796154237e-06, 'epoch': 0.35}) +[2026-10-03 15:33:21,119][starVLA.training.train_starvla][INFO] - Step 4849, Loss: {'train/loss': 0.0003055021516047418, 'train/grad_norm_pre_clip': 0.11400988698005676, 'timing/data': 0.0002900499966926873, 'timing/model': 0.37123730100574903, 'learning_rate/qwen_vl_interface': 5.222426158156008e-07, 'learning_rate/action_model': 5.222426158156008e-06, 'epoch': 0.35}) +[2026-10-03 15:33:21,768][starVLA.training.train_starvla][INFO] - Step 4850, Loss: {'train/loss': 5.861086174263619e-05, 'train/grad_norm_pre_clip': 0.01986389234662056, 'timing/data': 0.0002171200030716136, 'timing/model': 0.3698832109948853, 'learning_rate/qwen_vl_interface': 5.219492134913084e-07, 'learning_rate/action_model': 5.219492134913084e-06, 'epoch': 0.35}) +[2026-10-03 15:33:22,425][starVLA.training.train_starvla][INFO] - Step 4851, Loss: {'train/loss': 6.583587673958391e-05, 'train/grad_norm_pre_clip': 0.03033910132944584, 'timing/data': 0.00025943000218831003, 'timing/model': 0.3696955909981625, 'learning_rate/qwen_vl_interface': 5.216577546901008e-07, 'learning_rate/action_model': 5.216577546901007e-06, 'epoch': 0.36}) +[2026-10-03 15:33:23,277][starVLA.training.train_starvla][INFO] - Step 4852, Loss: {'train/loss': 4.8662139306543395e-05, 'train/grad_norm_pre_clip': 0.046667326241731644, 'timing/data': 0.0003295299975434318, 'timing/model': 0.36999567100428976, 'learning_rate/qwen_vl_interface': 5.213682395317852e-07, 'learning_rate/action_model': 5.213682395317851e-06, 'epoch': 0.36}) +[2026-10-03 15:33:23,937][starVLA.training.train_starvla][INFO] - Step 4853, Loss: {'train/loss': 4.478185292100534e-05, 'train/grad_norm_pre_clip': 0.015919333323836327, 'timing/data': 0.0003229199937777594, 'timing/model': 0.37101010099286214, 'learning_rate/qwen_vl_interface': 5.2108066813537e-07, 'learning_rate/action_model': 5.2108066813537e-06, 'epoch': 0.36}) +[2026-10-03 15:33:24,607][starVLA.training.train_starvla][INFO] - Step 4854, Loss: {'train/loss': 4.227858153171837e-05, 'train/grad_norm_pre_clip': 0.027015235275030136, 'timing/data': 0.00021598000603262335, 'timing/model': 0.36979464199976064, 'learning_rate/qwen_vl_interface': 5.207950406190654e-07, 'learning_rate/action_model': 5.207950406190653e-06, 'epoch': 0.36}) +[2026-10-03 15:33:25,264][starVLA.training.train_starvla][INFO] - Step 4855, Loss: {'train/loss': 5.485282963491045e-05, 'train/grad_norm_pre_clip': 0.022252436727285385, 'timing/data': 0.0002436600043438375, 'timing/model': 0.3693795709987171, 'learning_rate/qwen_vl_interface': 5.205113571002818e-07, 'learning_rate/action_model': 5.205113571002818e-06, 'epoch': 0.36}) +[2026-10-03 15:33:25,922][starVLA.training.train_starvla][INFO] - Step 4856, Loss: {'train/loss': 1.6726136891520582e-05, 'train/grad_norm_pre_clip': 0.005264026112854481, 'timing/data': 0.00034759999834932387, 'timing/model': 0.37014238099800423, 'learning_rate/qwen_vl_interface': 5.202296176956308e-07, 'learning_rate/action_model': 5.202296176956308e-06, 'epoch': 0.36}) +[2026-10-03 15:33:26,776][starVLA.training.train_starvla][INFO] - Step 4857, Loss: {'train/loss': 7.145034032873809e-06, 'train/grad_norm_pre_clip': 0.0016482558567076921, 'timing/data': 0.00032589999318588525, 'timing/model': 0.5659642859973246, 'learning_rate/qwen_vl_interface': 5.19949822520925e-07, 'learning_rate/action_model': 5.19949822520925e-06, 'epoch': 0.36}) +[2026-10-03 15:33:27,447][starVLA.training.train_starvla][INFO] - Step 4858, Loss: {'train/loss': 0.49535346031188965, 'train/grad_norm_pre_clip': 33.043701171875, 'timing/data': 0.0002643899933900684, 'timing/model': 0.3698739309911616, 'learning_rate/qwen_vl_interface': 5.196719716911777e-07, 'learning_rate/action_model': 5.1967197169117764e-06, 'epoch': 0.36}) +[2026-10-03 15:33:28,105][starVLA.training.train_starvla][INFO] - Step 4859, Loss: {'train/loss': 0.0005747562972828746, 'train/grad_norm_pre_clip': 0.20904125273227692, 'timing/data': 0.00041448000411037356, 'timing/model': 0.369586510991212, 'learning_rate/qwen_vl_interface': 5.193960653206029e-07, 'learning_rate/action_model': 5.193960653206029e-06, 'epoch': 0.36}) +[2026-10-03 15:33:28,763][starVLA.training.train_starvla][INFO] - Step 4860, Loss: {'train/loss': 0.00015008720220066607, 'train/grad_norm_pre_clip': 0.06827592104673386, 'timing/data': 0.00025999000354204327, 'timing/model': 0.3705738600110635, 'learning_rate/qwen_vl_interface': 5.191221035226144e-07, 'learning_rate/action_model': 5.191221035226144e-06, 'epoch': 0.36}) +[2026-10-03 15:33:29,422][starVLA.training.train_starvla][INFO] - Step 4861, Loss: {'train/loss': 0.00041143549606204033, 'train/grad_norm_pre_clip': 0.4148881733417511, 'timing/data': 0.0003095800057053566, 'timing/model': 0.3708754210092593, 'learning_rate/qwen_vl_interface': 5.188500864098292e-07, 'learning_rate/action_model': 5.1885008640982915e-06, 'epoch': 0.36}) +[2026-10-03 15:33:30,095][starVLA.training.train_starvla][INFO] - Step 4862, Loss: {'train/loss': 7.360434392467141e-05, 'train/grad_norm_pre_clip': 0.02325555868446827, 'timing/data': 0.0002177199930883944, 'timing/model': 0.36995211099565495, 'learning_rate/qwen_vl_interface': 5.185800140940613e-07, 'learning_rate/action_model': 5.185800140940613e-06, 'epoch': 0.36}) +[2026-10-03 15:33:30,945][starVLA.training.train_starvla][INFO] - Step 4863, Loss: {'train/loss': 8.466441067866981e-05, 'train/grad_norm_pre_clip': 0.05137073993682861, 'timing/data': 0.00033815999631769955, 'timing/model': 0.3697223210037919, 'learning_rate/qwen_vl_interface': 5.183118866863288e-07, 'learning_rate/action_model': 5.1831188668632875e-06, 'epoch': 0.36}) +[2026-10-03 15:33:31,602][starVLA.training.train_starvla][INFO] - Step 4864, Loss: {'train/loss': 0.23115812242031097, 'train/grad_norm_pre_clip': 62.963233947753906, 'timing/data': 0.00035867000406142324, 'timing/model': 0.370411101001082, 'learning_rate/qwen_vl_interface': 5.180457042968484e-07, 'learning_rate/action_model': 5.1804570429684835e-06, 'epoch': 0.36}) +[2026-10-03 15:33:32,262][starVLA.training.train_starvla][INFO] - Step 4865, Loss: {'train/loss': 0.0001207123787025921, 'train/grad_norm_pre_clip': 0.04208197444677353, 'timing/data': 0.00029992000781930983, 'timing/model': 0.37093248100427445, 'learning_rate/qwen_vl_interface': 5.17781467035037e-07, 'learning_rate/action_model': 5.177814670350371e-06, 'epoch': 0.36}) +[2026-10-03 15:33:32,921][starVLA.training.train_starvla][INFO] - Step 4866, Loss: {'train/loss': 0.0001874744484666735, 'train/grad_norm_pre_clip': 0.11380137503147125, 'timing/data': 0.00027180000324733555, 'timing/model': 0.3699391310074134, 'learning_rate/qwen_vl_interface': 5.175191750095133e-07, 'learning_rate/action_model': 5.175191750095132e-06, 'epoch': 0.36}) +[2026-10-03 15:33:33,579][starVLA.training.train_starvla][INFO] - Step 4867, Loss: {'train/loss': 0.31844422221183777, 'train/grad_norm_pre_clip': 132.47415161132812, 'timing/data': 0.0002090699999826029, 'timing/model': 0.3699259810091462, 'learning_rate/qwen_vl_interface': 5.172588283280953e-07, 'learning_rate/action_model': 5.172588283280953e-06, 'epoch': 0.36}) +[2026-10-03 15:33:34,436][starVLA.training.train_starvla][INFO] - Step 4868, Loss: {'train/loss': 0.00015615252777934074, 'train/grad_norm_pre_clip': 0.10805200040340424, 'timing/data': 0.00028759999258909374, 'timing/model': 0.3701224410033319, 'learning_rate/qwen_vl_interface': 5.170004270978019e-07, 'learning_rate/action_model': 5.170004270978018e-06, 'epoch': 0.36}) +[2026-10-03 15:33:35,073][starVLA.training.train_starvla][INFO] - Step 4869, Loss: {'train/loss': 8.421098027611151e-05, 'train/grad_norm_pre_clip': 0.02694530040025711, 'timing/data': 0.00026745999639388174, 'timing/model': 0.3481339739955729, 'learning_rate/qwen_vl_interface': 5.167439714248512e-07, 'learning_rate/action_model': 5.1674397142485116e-06, 'epoch': 0.36}) +[2026-10-03 15:33:35,732][starVLA.training.train_starvla][INFO] - Step 4870, Loss: {'train/loss': 4.556958447210491e-05, 'train/grad_norm_pre_clip': 0.028836054727435112, 'timing/data': 0.0002955000090878457, 'timing/model': 0.3701159610063769, 'learning_rate/qwen_vl_interface': 5.164894614146633e-07, 'learning_rate/action_model': 5.164894614146634e-06, 'epoch': 0.36}) +[2026-10-03 15:33:36,367][starVLA.training.train_starvla][INFO] - Step 4871, Loss: {'train/loss': 5.441610119305551e-05, 'train/grad_norm_pre_clip': 0.013977576978504658, 'timing/data': 0.00020908999431412667, 'timing/model': 0.3693712419917574, 'learning_rate/qwen_vl_interface': 5.16236897171858e-07, 'learning_rate/action_model': 5.16236897171858e-06, 'epoch': 0.36}) +[2026-10-03 15:33:36,979][starVLA.training.train_starvla][INFO] - Step 4872, Loss: {'train/loss': 0.00011944276047870517, 'train/grad_norm_pre_clip': 0.03105791285634041, 'timing/data': 0.00021405999723356217, 'timing/model': 0.3475919439952122, 'learning_rate/qwen_vl_interface': 5.15986278800254e-07, 'learning_rate/action_model': 5.15986278800254e-06, 'epoch': 0.36}) +[2026-10-03 15:33:37,639][starVLA.training.train_starvla][INFO] - Step 4873, Loss: {'train/loss': 0.00016495914314873517, 'train/grad_norm_pre_clip': 0.09145338088274002, 'timing/data': 0.00027129999944008887, 'timing/model': 0.3715412109886529, 'learning_rate/qwen_vl_interface': 5.157376064028719e-07, 'learning_rate/action_model': 5.157376064028718e-06, 'epoch': 0.36}) +[2026-10-03 15:33:38,500][starVLA.training.train_starvla][INFO] - Step 4874, Loss: {'train/loss': 0.0001349405647488311, 'train/grad_norm_pre_clip': 0.050902802497148514, 'timing/data': 0.00046098999155219644, 'timing/model': 0.37026483099907637, 'learning_rate/qwen_vl_interface': 5.154908800819307e-07, 'learning_rate/action_model': 5.154908800819306e-06, 'epoch': 0.36}) +[2026-10-03 15:33:39,159][starVLA.training.train_starvla][INFO] - Step 4875, Loss: {'train/loss': 0.00013919519551564008, 'train/grad_norm_pre_clip': 0.08270172029733658, 'timing/data': 0.00027003999275621027, 'timing/model': 0.37031107099028304, 'learning_rate/qwen_vl_interface': 5.152460999388504e-07, 'learning_rate/action_model': 5.152460999388503e-06, 'epoch': 0.36}) +[2026-10-03 15:33:39,818][starVLA.training.train_starvla][INFO] - Step 4876, Loss: {'train/loss': 1.4535715308738872e-05, 'train/grad_norm_pre_clip': 0.004568586125969887, 'timing/data': 0.0002961999998660758, 'timing/model': 0.37023561100068036, 'learning_rate/qwen_vl_interface': 5.150032660742517e-07, 'learning_rate/action_model': 5.150032660742517e-06, 'epoch': 0.36}) +[2026-10-03 15:33:40,478][starVLA.training.train_starvla][INFO] - Step 4877, Loss: {'train/loss': 3.821792779490352e-05, 'train/grad_norm_pre_clip': 0.012997139245271683, 'timing/data': 0.00025955001183319837, 'timing/model': 0.37094846098625567, 'learning_rate/qwen_vl_interface': 5.147623785879535e-07, 'learning_rate/action_model': 5.147623785879534e-06, 'epoch': 0.36}) +[2026-10-03 15:33:41,137][starVLA.training.train_starvla][INFO] - Step 4878, Loss: {'train/loss': 8.921434346120805e-05, 'train/grad_norm_pre_clip': 0.0475887805223465, 'timing/data': 0.0003624300006777048, 'timing/model': 0.3700112119986443, 'learning_rate/qwen_vl_interface': 5.145234375789758e-07, 'learning_rate/action_model': 5.145234375789758e-06, 'epoch': 0.36}) +[2026-10-03 15:33:41,991][starVLA.training.train_starvla][INFO] - Step 4879, Loss: {'train/loss': 0.4313446283340454, 'train/grad_norm_pre_clip': 47.14116287231445, 'timing/data': 0.0003373500076122582, 'timing/model': 0.3697049809998134, 'learning_rate/qwen_vl_interface': 5.142864431455387e-07, 'learning_rate/action_model': 5.1428644314553875e-06, 'epoch': 0.36}) +[2026-10-03 15:33:42,648][starVLA.training.train_starvla][INFO] - Step 4880, Loss: {'train/loss': 4.562049434753135e-05, 'train/grad_norm_pre_clip': 0.01698417030274868, 'timing/data': 0.0002521199930924922, 'timing/model': 0.3699425609956961, 'learning_rate/qwen_vl_interface': 5.140513953850612e-07, 'learning_rate/action_model': 5.140513953850612e-06, 'epoch': 0.36}) +[2026-10-03 15:33:43,310][starVLA.training.train_starvla][INFO] - Step 4881, Loss: {'train/loss': 1.0981961167999543e-05, 'train/grad_norm_pre_clip': 0.0030708007980138063, 'timing/data': 0.0003666700067697093, 'timing/model': 0.3725520809966838, 'learning_rate/qwen_vl_interface': 5.138182943941623e-07, 'learning_rate/action_model': 5.138182943941623e-06, 'epoch': 0.36}) +[2026-10-03 15:33:43,972][starVLA.training.train_starvla][INFO] - Step 4882, Loss: {'train/loss': 5.0634273065952584e-05, 'train/grad_norm_pre_clip': 0.024646859616041183, 'timing/data': 0.0003591599961509928, 'timing/model': 0.37208535098761786, 'learning_rate/qwen_vl_interface': 5.135871402686615e-07, 'learning_rate/action_model': 5.135871402686615e-06, 'epoch': 0.36}) +[2026-10-03 15:33:44,634][starVLA.training.train_starvla][INFO] - Step 4883, Loss: {'train/loss': 0.0003463165776338428, 'train/grad_norm_pre_clip': 0.20409363508224487, 'timing/data': 0.00024194001161959022, 'timing/model': 0.3717431409895653, 'learning_rate/qwen_vl_interface': 5.133579331035777e-07, 'learning_rate/action_model': 5.133579331035777e-06, 'epoch': 0.36}) +[2026-10-03 15:33:45,495][starVLA.training.train_starvla][INFO] - Step 4884, Loss: {'train/loss': 0.01347668468952179, 'train/grad_norm_pre_clip': 44.57902145385742, 'timing/data': 0.0002469099999871105, 'timing/model': 0.5713400339882355, 'learning_rate/qwen_vl_interface': 5.131306729931284e-07, 'learning_rate/action_model': 5.1313067299312836e-06, 'epoch': 0.36}) +[2026-10-03 15:33:46,158][starVLA.training.train_starvla][INFO] - Step 4885, Loss: {'train/loss': 0.40686678886413574, 'train/grad_norm_pre_clip': 35.245243072509766, 'timing/data': 0.00027396999939810485, 'timing/model': 0.3726604010007577, 'learning_rate/qwen_vl_interface': 5.129053600307326e-07, 'learning_rate/action_model': 5.129053600307326e-06, 'epoch': 0.36}) +[2026-10-03 15:33:46,822][starVLA.training.train_starvla][INFO] - Step 4886, Loss: {'train/loss': 0.00013559425133280456, 'train/grad_norm_pre_clip': 0.053650908172130585, 'timing/data': 0.00033518999407533556, 'timing/model': 0.37281344100483693, 'learning_rate/qwen_vl_interface': 5.126819943090078e-07, 'learning_rate/action_model': 5.1268199430900775e-06, 'epoch': 0.36}) +[2026-10-03 15:33:47,483][starVLA.training.train_starvla][INFO] - Step 4887, Loss: {'train/loss': 0.0029263165779411793, 'train/grad_norm_pre_clip': 7.6454362869262695, 'timing/data': 0.00032374999136663973, 'timing/model': 0.37188432000402827, 'learning_rate/qwen_vl_interface': 5.124605759197706e-07, 'learning_rate/action_model': 5.124605759197706e-06, 'epoch': 0.36}) +[2026-10-03 15:33:48,146][starVLA.training.train_starvla][INFO] - Step 4888, Loss: {'train/loss': 5.307681567501277e-05, 'train/grad_norm_pre_clip': 0.013726083561778069, 'timing/data': 0.0002529699995648116, 'timing/model': 0.3726077399915084, 'learning_rate/qwen_vl_interface': 5.122411049540383e-07, 'learning_rate/action_model': 5.1224110495403825e-06, 'epoch': 0.36}) +[2026-10-03 15:33:48,810][starVLA.training.train_starvla][INFO] - Step 4889, Loss: {'train/loss': 4.703294689534232e-05, 'train/grad_norm_pre_clip': 0.021320682018995285, 'timing/data': 0.00028610001027118415, 'timing/model': 0.3728163019986823, 'learning_rate/qwen_vl_interface': 5.120235815020269e-07, 'learning_rate/action_model': 5.120235815020269e-06, 'epoch': 0.36}) +[2026-10-03 15:33:49,664][starVLA.training.train_starvla][INFO] - Step 4890, Loss: {'train/loss': 2.6231751689920202e-05, 'train/grad_norm_pre_clip': 0.00813246425241232, 'timing/data': 0.00028931000269949436, 'timing/model': 0.3719872709916672, 'learning_rate/qwen_vl_interface': 5.118080056531524e-07, 'learning_rate/action_model': 5.118080056531524e-06, 'epoch': 0.36}) +[2026-10-03 15:33:50,326][starVLA.training.train_starvla][INFO] - Step 4891, Loss: {'train/loss': 0.058353111147880554, 'train/grad_norm_pre_clip': 122.79721069335938, 'timing/data': 0.0003454300021985546, 'timing/model': 0.37197749099868815, 'learning_rate/qwen_vl_interface': 5.115943774960295e-07, 'learning_rate/action_model': 5.115943774960295e-06, 'epoch': 0.36}) +[2026-10-03 15:33:50,988][starVLA.training.train_starvla][INFO] - Step 4892, Loss: {'train/loss': 2.1881633074372075e-05, 'train/grad_norm_pre_clip': 0.005415116902440786, 'timing/data': 0.0003141799970762804, 'timing/model': 0.3723356019909261, 'learning_rate/qwen_vl_interface': 5.113826971184727e-07, 'learning_rate/action_model': 5.113826971184727e-06, 'epoch': 0.36}) +[2026-10-03 15:33:51,653][starVLA.training.train_starvla][INFO] - Step 4893, Loss: {'train/loss': 0.0003418555133976042, 'train/grad_norm_pre_clip': 0.06444957107305527, 'timing/data': 0.0002622799947857857, 'timing/model': 0.3727971299958881, 'learning_rate/qwen_vl_interface': 5.111729646074954e-07, 'learning_rate/action_model': 5.111729646074954e-06, 'epoch': 0.36}) +[2026-10-03 15:33:52,316][starVLA.training.train_starvla][INFO] - Step 4894, Loss: {'train/loss': 0.00012525224883574992, 'train/grad_norm_pre_clip': 0.16955062747001648, 'timing/data': 0.0003377699904376641, 'timing/model': 0.372066140989773, 'learning_rate/qwen_vl_interface': 5.10965180049312e-07, 'learning_rate/action_model': 5.10965180049312e-06, 'epoch': 0.36}) +[2026-10-03 15:33:53,179][starVLA.training.train_starvla][INFO] - Step 4895, Loss: {'train/loss': 0.00018294909386895597, 'train/grad_norm_pre_clip': 0.10426747053861618, 'timing/data': 0.0003444500034675002, 'timing/model': 0.3715979210101068, 'learning_rate/qwen_vl_interface': 5.107593435293336e-07, 'learning_rate/action_model': 5.107593435293336e-06, 'epoch': 0.36}) +[2026-10-03 15:33:53,842][starVLA.training.train_starvla][INFO] - Step 4896, Loss: {'train/loss': 0.0236051082611084, 'train/grad_norm_pre_clip': 48.841705322265625, 'timing/data': 0.00028199999360367656, 'timing/model': 0.3722302009991836, 'learning_rate/qwen_vl_interface': 5.105554551321724e-07, 'learning_rate/action_model': 5.105554551321723e-06, 'epoch': 0.36}) +[2026-10-03 15:33:54,505][starVLA.training.train_starvla][INFO] - Step 4897, Loss: {'train/loss': 1.0676569218048826e-05, 'train/grad_norm_pre_clip': 0.002285365480929613, 'timing/data': 0.00026824000815395266, 'timing/model': 0.37260174099355936, 'learning_rate/qwen_vl_interface': 5.103535149416389e-07, 'learning_rate/action_model': 5.1035351494163885e-06, 'epoch': 0.36}) +[2026-10-03 15:33:55,167][starVLA.training.train_starvla][INFO] - Step 4898, Loss: {'train/loss': 0.0005646042991429567, 'train/grad_norm_pre_clip': 0.23738263547420502, 'timing/data': 0.0002152399974875152, 'timing/model': 0.371654579997994, 'learning_rate/qwen_vl_interface': 5.101535230407436e-07, 'learning_rate/action_model': 5.101535230407436e-06, 'epoch': 0.36}) +[2026-10-03 15:33:55,830][starVLA.training.train_starvla][INFO] - Step 4899, Loss: {'train/loss': 0.0004441087949089706, 'train/grad_norm_pre_clip': 0.1191682443022728, 'timing/data': 0.00026902000536210835, 'timing/model': 0.372257401002571, 'learning_rate/qwen_vl_interface': 5.099554795116956e-07, 'learning_rate/action_model': 5.0995547951169565e-06, 'epoch': 0.36}) +[2026-10-03 15:33:56,689][starVLA.training.train_starvla][INFO] - Step 4900, Loss: {'train/loss': 0.00601683696731925, 'train/grad_norm_pre_clip': 16.67626190185547, 'timing/data': 0.00028759999258909374, 'timing/model': 0.37346839100064244, 'learning_rate/qwen_vl_interface': 5.097593844359024e-07, 'learning_rate/action_model': 5.0975938443590245e-06, 'epoch': 0.36}) +[2026-10-03 15:33:57,353][starVLA.training.train_starvla][INFO] - Step 4901, Loss: {'train/loss': 0.9136807918548584, 'train/grad_norm_pre_clip': 49.11577606201172, 'timing/data': 0.0002740199997788295, 'timing/model': 0.3728721509978641, 'learning_rate/qwen_vl_interface': 5.095652378939718e-07, 'learning_rate/action_model': 5.095652378939717e-06, 'epoch': 0.36}) +[2026-10-03 15:33:58,015][starVLA.training.train_starvla][INFO] - Step 4902, Loss: {'train/loss': 0.00020934719941578805, 'train/grad_norm_pre_clip': 0.10638724267482758, 'timing/data': 0.00024743999529164284, 'timing/model': 0.3717747209884692, 'learning_rate/qwen_vl_interface': 5.093730399657101e-07, 'learning_rate/action_model': 5.0937303996571e-06, 'epoch': 0.36}) +[2026-10-03 15:33:58,655][starVLA.training.train_starvla][INFO] - Step 4903, Loss: {'train/loss': 5.12732622155454e-05, 'train/grad_norm_pre_clip': 0.0203593447804451, 'timing/data': 0.00029163999715819955, 'timing/model': 0.3490492639975855, 'learning_rate/qwen_vl_interface': 5.091827907301227e-07, 'learning_rate/action_model': 5.091827907301227e-06, 'epoch': 0.36}) +[2026-10-03 15:33:59,317][starVLA.training.train_starvla][INFO] - Step 4904, Loss: {'train/loss': 2.8605141778825782e-05, 'train/grad_norm_pre_clip': 0.014755360782146454, 'timing/data': 0.00030762000824324787, 'timing/model': 0.37256189101026393, 'learning_rate/qwen_vl_interface': 5.089944902654134e-07, 'learning_rate/action_model': 5.089944902654134e-06, 'epoch': 0.36}) +[2026-10-03 15:34:00,171][starVLA.training.train_starvla][INFO] - Step 4905, Loss: {'train/loss': 0.0006822457071393728, 'train/grad_norm_pre_clip': 0.8173492550849915, 'timing/data': 0.00022098000044934452, 'timing/model': 0.5648690050002187, 'learning_rate/qwen_vl_interface': 5.088081386489856e-07, 'learning_rate/action_model': 5.088081386489857e-06, 'epoch': 0.36}) +[2026-10-03 15:34:00,829][starVLA.training.train_starvla][INFO] - Step 4906, Loss: {'train/loss': 0.2607070803642273, 'train/grad_norm_pre_clip': 45.570743560791016, 'timing/data': 0.0002945099986391142, 'timing/model': 0.3698821710131597, 'learning_rate/qwen_vl_interface': 5.086237359574419e-07, 'learning_rate/action_model': 5.086237359574419e-06, 'epoch': 0.36}) +[2026-10-03 15:34:01,466][starVLA.training.train_starvla][INFO] - Step 4907, Loss: {'train/loss': 6.47272972855717e-05, 'train/grad_norm_pre_clip': 0.03806683421134949, 'timing/data': 0.0002946199965663254, 'timing/model': 0.3475251640047645, 'learning_rate/qwen_vl_interface': 5.084412822665825e-07, 'learning_rate/action_model': 5.084412822665825e-06, 'epoch': 0.36}) +[2026-10-03 15:34:02,124][starVLA.training.train_starvla][INFO] - Step 4908, Loss: {'train/loss': 0.00020793237490579486, 'train/grad_norm_pre_clip': 0.1579296886920929, 'timing/data': 0.00027385000430513173, 'timing/model': 0.37067720100458246, 'learning_rate/qwen_vl_interface': 5.082607776514081e-07, 'learning_rate/action_model': 5.082607776514081e-06, 'epoch': 0.36}) +[2026-10-03 15:34:02,785][starVLA.training.train_starvla][INFO] - Step 4909, Loss: {'train/loss': 0.0005218198639340699, 'train/grad_norm_pre_clip': 0.8246444463729858, 'timing/data': 0.0003638299967860803, 'timing/model': 0.37169473100220785, 'learning_rate/qwen_vl_interface': 5.080822221861167e-07, 'learning_rate/action_model': 5.080822221861166e-06, 'epoch': 0.36}) +[2026-10-03 15:34:03,642][starVLA.training.train_starvla][INFO] - Step 4910, Loss: {'train/loss': 0.027739936485886574, 'train/grad_norm_pre_clip': 112.96422576904297, 'timing/data': 0.00033552999957464635, 'timing/model': 0.5684169849992031, 'learning_rate/qwen_vl_interface': 5.07905615944106e-07, 'learning_rate/action_model': 5.0790561594410595e-06, 'epoch': 0.36}) +[2026-10-03 15:34:04,302][starVLA.training.train_starvla][INFO] - Step 4911, Loss: {'train/loss': 2.8646070859394968e-05, 'train/grad_norm_pre_clip': 0.010575487278401852, 'timing/data': 0.00023207000049296767, 'timing/model': 0.3705757809948409, 'learning_rate/qwen_vl_interface': 5.077309589979716e-07, 'learning_rate/action_model': 5.077309589979715e-06, 'epoch': 0.36}) +[2026-10-03 15:34:04,960][starVLA.training.train_starvla][INFO] - Step 4912, Loss: {'train/loss': 0.32520395517349243, 'train/grad_norm_pre_clip': 17.219486236572266, 'timing/data': 0.000304490007692948, 'timing/model': 0.3703795109904604, 'learning_rate/qwen_vl_interface': 5.075582514195089e-07, 'learning_rate/action_model': 5.0755825141950895e-06, 'epoch': 0.36}) +[2026-10-03 15:34:05,620][starVLA.training.train_starvla][INFO] - Step 4913, Loss: {'train/loss': 2.492799103492871e-05, 'train/grad_norm_pre_clip': 0.008484559133648872, 'timing/data': 0.0002599800063762814, 'timing/model': 0.37139930100238416, 'learning_rate/qwen_vl_interface': 5.073874932797115e-07, 'learning_rate/action_model': 5.0738749327971155e-06, 'epoch': 0.36}) +[2026-10-03 15:34:06,279][starVLA.training.train_starvla][INFO] - Step 4914, Loss: {'train/loss': 0.000333597999997437, 'train/grad_norm_pre_clip': 0.12815351784229279, 'timing/data': 0.0003468200011411682, 'timing/model': 0.3696551309985807, 'learning_rate/qwen_vl_interface': 5.072186846487713e-07, 'learning_rate/action_model': 5.072186846487713e-06, 'epoch': 0.36}) +[2026-10-03 15:34:06,937][starVLA.training.train_starvla][INFO] - Step 4915, Loss: {'train/loss': 9.201356988342013e-06, 'train/grad_norm_pre_clip': 0.0022456329315900803, 'timing/data': 0.00025630000163801014, 'timing/model': 0.36967415099206846, 'learning_rate/qwen_vl_interface': 5.070518255960796e-07, 'learning_rate/action_model': 5.070518255960795e-06, 'epoch': 0.36}) +[2026-10-03 15:34:07,772][starVLA.training.train_starvla][INFO] - Step 4916, Loss: {'train/loss': 3.1618637876817957e-05, 'train/grad_norm_pre_clip': 0.0117898378521204, 'timing/data': 0.0003133200079901144, 'timing/model': 0.3477563549968181, 'learning_rate/qwen_vl_interface': 5.068869161902256e-07, 'learning_rate/action_model': 5.068869161902256e-06, 'epoch': 0.36}) +[2026-10-03 15:34:08,433][starVLA.training.train_starvla][INFO] - Step 4917, Loss: {'train/loss': 9.030020009959117e-06, 'train/grad_norm_pre_clip': 0.002038327045738697, 'timing/data': 0.0002653599949553609, 'timing/model': 0.37098958098795265, 'learning_rate/qwen_vl_interface': 5.067239564989968e-07, 'learning_rate/action_model': 5.067239564989968e-06, 'epoch': 0.36}) +[2026-10-03 15:34:09,068][starVLA.training.train_starvla][INFO] - Step 4918, Loss: {'train/loss': 1.578740557306446e-05, 'train/grad_norm_pre_clip': 0.005742689594626427, 'timing/data': 0.0003067700017709285, 'timing/model': 0.36960804200498387, 'learning_rate/qwen_vl_interface': 5.065629465893808e-07, 'learning_rate/action_model': 5.065629465893807e-06, 'epoch': 0.36}) +[2026-10-03 15:34:09,728][starVLA.training.train_starvla][INFO] - Step 4919, Loss: {'train/loss': 0.3620471656322479, 'train/grad_norm_pre_clip': 139.2984619140625, 'timing/data': 0.00023061000683810562, 'timing/model': 0.3700118620035937, 'learning_rate/qwen_vl_interface': 5.064038865275625e-07, 'learning_rate/action_model': 5.0640388652756245e-06, 'epoch': 0.36}) +[2026-10-03 15:34:10,386][starVLA.training.train_starvla][INFO] - Step 4920, Loss: {'train/loss': 0.7030540108680725, 'train/grad_norm_pre_clip': 31.40906524658203, 'timing/data': 0.0002922300045611337, 'timing/model': 0.37011682099546306, 'learning_rate/qwen_vl_interface': 5.062467763789242e-07, 'learning_rate/action_model': 5.062467763789242e-06, 'epoch': 0.36}) +[2026-10-03 15:34:11,242][starVLA.training.train_starvla][INFO] - Step 4921, Loss: {'train/loss': 6.367008609231561e-05, 'train/grad_norm_pre_clip': 0.021156392991542816, 'timing/data': 0.00022471000556834042, 'timing/model': 0.5667590049997671, 'learning_rate/qwen_vl_interface': 5.060916162080496e-07, 'learning_rate/action_model': 5.060916162080496e-06, 'epoch': 0.36}) +[2026-10-03 15:34:11,902][starVLA.training.train_starvla][INFO] - Step 4922, Loss: {'train/loss': 1.6115176549647003e-05, 'train/grad_norm_pre_clip': 0.005733451806008816, 'timing/data': 0.0003178999904775992, 'timing/model': 0.3706879210076295, 'learning_rate/qwen_vl_interface': 5.059384060787182e-07, 'learning_rate/action_model': 5.059384060787183e-06, 'epoch': 0.36}) +[2026-10-03 15:34:12,561][starVLA.training.train_starvla][INFO] - Step 4923, Loss: {'train/loss': 8.379280916415155e-05, 'train/grad_norm_pre_clip': 0.029033396393060684, 'timing/data': 0.0003227700071875006, 'timing/model': 0.3699069409922231, 'learning_rate/qwen_vl_interface': 5.05787146053909e-07, 'learning_rate/action_model': 5.0578714605390904e-06, 'epoch': 0.36}) +[2026-10-03 15:34:13,220][starVLA.training.train_starvla][INFO] - Step 4924, Loss: {'train/loss': 0.009548739530146122, 'train/grad_norm_pre_clip': 38.74428176879883, 'timing/data': 0.0002537200052756816, 'timing/model': 0.37125129099877086, 'learning_rate/qwen_vl_interface': 5.056378361957998e-07, 'learning_rate/action_model': 5.056378361957998e-06, 'epoch': 0.36}) +[2026-10-03 15:34:13,881][starVLA.training.train_starvla][INFO] - Step 4925, Loss: {'train/loss': 0.002655548043549061, 'train/grad_norm_pre_clip': 10.046346664428711, 'timing/data': 0.00022329999774228781, 'timing/model': 0.37196398999367375, 'learning_rate/qwen_vl_interface': 5.05490476565766e-07, 'learning_rate/action_model': 5.054904765657659e-06, 'epoch': 0.36}) +[2026-10-03 15:34:14,540][starVLA.training.train_starvla][INFO] - Step 4926, Loss: {'train/loss': 0.00017123376892413944, 'train/grad_norm_pre_clip': 0.07518694549798965, 'timing/data': 0.00026393000734969974, 'timing/model': 0.3704974510037573, 'learning_rate/qwen_vl_interface': 5.053450672243811e-07, 'learning_rate/action_model': 5.0534506722438114e-06, 'epoch': 0.36}) +[2026-10-03 15:34:15,388][starVLA.training.train_starvla][INFO] - Step 4927, Loss: {'train/loss': 9.05983142729383e-06, 'train/grad_norm_pre_clip': 0.0018575048306956887, 'timing/data': 0.00033169001108035445, 'timing/model': 0.3697936219978146, 'learning_rate/qwen_vl_interface': 5.052016082314184e-07, 'learning_rate/action_model': 5.052016082314184e-06, 'epoch': 0.36}) +[2026-10-03 15:34:16,026][starVLA.training.train_starvla][INFO] - Step 4928, Loss: {'train/loss': 5.434115155367181e-05, 'train/grad_norm_pre_clip': 0.019354717805981636, 'timing/data': 0.0002567100018495694, 'timing/model': 0.37057303100300487, 'learning_rate/qwen_vl_interface': 5.050600996458473e-07, 'learning_rate/action_model': 5.050600996458473e-06, 'epoch': 0.36}) +[2026-10-03 15:34:16,691][starVLA.training.train_starvla][INFO] - Step 4929, Loss: {'train/loss': 5.354708628146909e-05, 'train/grad_norm_pre_clip': 0.020911602303385735, 'timing/data': 0.00030156999127939343, 'timing/model': 0.37056490100803785, 'learning_rate/qwen_vl_interface': 5.049205415258373e-07, 'learning_rate/action_model': 5.049205415258373e-06, 'epoch': 0.36}) +[2026-10-03 15:34:17,353][starVLA.training.train_starvla][INFO] - Step 4930, Loss: {'train/loss': 0.0003174099256284535, 'train/grad_norm_pre_clip': 0.17703087627887726, 'timing/data': 0.0003997000021627173, 'timing/model': 0.3716729509906145, 'learning_rate/qwen_vl_interface': 5.047829339287557e-07, 'learning_rate/action_model': 5.047829339287557e-06, 'epoch': 0.36}) +[2026-10-03 15:34:18,205][starVLA.training.train_starvla][INFO] - Step 4931, Loss: {'train/loss': 3.700876550283283e-05, 'train/grad_norm_pre_clip': 0.021412352100014687, 'timing/data': 0.000288159993942827, 'timing/model': 0.5614717650023522, 'learning_rate/qwen_vl_interface': 5.046472769111668e-07, 'learning_rate/action_model': 5.046472769111667e-06, 'epoch': 0.36}) +[2026-10-03 15:34:18,867][starVLA.training.train_starvla][INFO] - Step 4932, Loss: {'train/loss': 0.0006685812841169536, 'train/grad_norm_pre_clip': 0.10980580002069473, 'timing/data': 0.00027644999499898404, 'timing/model': 0.37142776101245545, 'learning_rate/qwen_vl_interface': 5.045135705288353e-07, 'learning_rate/action_model': 5.0451357052883525e-06, 'epoch': 0.36}) +[2026-10-03 15:34:19,523][starVLA.training.train_starvla][INFO] - Step 4933, Loss: {'train/loss': 6.171526911202818e-05, 'train/grad_norm_pre_clip': 0.05455989018082619, 'timing/data': 0.00028516000020317733, 'timing/model': 0.36693644199112896, 'learning_rate/qwen_vl_interface': 5.043818148367221e-07, 'learning_rate/action_model': 5.043818148367221e-06, 'epoch': 0.36}) +[2026-10-03 15:34:20,173][starVLA.training.train_starvla][INFO] - Step 4934, Loss: {'train/loss': 4.870044722338207e-05, 'train/grad_norm_pre_clip': 0.01373603567481041, 'timing/data': 0.0002929299953393638, 'timing/model': 0.367018481993, 'learning_rate/qwen_vl_interface': 5.042520098889869e-07, 'learning_rate/action_model': 5.0425200988898685e-06, 'epoch': 0.36}) +[2026-10-03 15:34:20,826][starVLA.training.train_starvla][INFO] - Step 4935, Loss: {'train/loss': 1.683757182036061e-05, 'train/grad_norm_pre_clip': 0.005043248180299997, 'timing/data': 0.0002538800035836175, 'timing/model': 0.36691890099609736, 'learning_rate/qwen_vl_interface': 5.041241557389879e-07, 'learning_rate/action_model': 5.041241557389879e-06, 'epoch': 0.36}) +[2026-10-03 15:34:21,474][starVLA.training.train_starvla][INFO] - Step 4936, Loss: {'train/loss': 0.001054505817592144, 'train/grad_norm_pre_clip': 1.7967591285705566, 'timing/data': 0.0002844300033757463, 'timing/model': 0.36535504199855495, 'learning_rate/qwen_vl_interface': 5.039982524392814e-07, 'learning_rate/action_model': 5.039982524392814e-06, 'epoch': 0.36}) +[2026-10-03 15:34:22,328][starVLA.training.train_starvla][INFO] - Step 4937, Loss: {'train/loss': 0.3447626233100891, 'train/grad_norm_pre_clip': 53.204063415527344, 'timing/data': 0.00025332999939564615, 'timing/model': 0.36496077199990395, 'learning_rate/qwen_vl_interface': 5.038743000416207e-07, 'learning_rate/action_model': 5.0387430004162064e-06, 'epoch': 0.36}) +[2026-10-03 15:34:22,979][starVLA.training.train_starvla][INFO] - Step 4938, Loss: {'train/loss': 0.381612092256546, 'train/grad_norm_pre_clip': 51.898841857910156, 'timing/data': 0.00023963999410625547, 'timing/model': 0.3676753910112893, 'learning_rate/qwen_vl_interface': 5.037522985969588e-07, 'learning_rate/action_model': 5.037522985969587e-06, 'epoch': 0.36}) +[2026-10-03 15:34:23,631][starVLA.training.train_starvla][INFO] - Step 4939, Loss: {'train/loss': 4.894566154689528e-05, 'train/grad_norm_pre_clip': 0.015708571299910545, 'timing/data': 0.00029738999728579074, 'timing/model': 0.3670645710080862, 'learning_rate/qwen_vl_interface': 5.036322481554451e-07, 'learning_rate/action_model': 5.036322481554451e-06, 'epoch': 0.36}) +[2026-10-03 15:34:24,279][starVLA.training.train_starvla][INFO] - Step 4940, Loss: {'train/loss': 3.1752897484693676e-05, 'train/grad_norm_pre_clip': 0.00940179917961359, 'timing/data': 0.00025582000671420246, 'timing/model': 0.3644067920104135, 'learning_rate/qwen_vl_interface': 5.035141487664285e-07, 'learning_rate/action_model': 5.035141487664285e-06, 'epoch': 0.36}) +[2026-10-03 15:34:24,927][starVLA.training.train_starvla][INFO] - Step 4941, Loss: {'train/loss': 7.942197953525465e-06, 'train/grad_norm_pre_clip': 0.0027419908437877893, 'timing/data': 0.00025209999876096845, 'timing/model': 0.36476234199653845, 'learning_rate/qwen_vl_interface': 5.03398000478455e-07, 'learning_rate/action_model': 5.03398000478455e-06, 'epoch': 0.36}) +[2026-10-03 15:34:25,770][starVLA.training.train_starvla][INFO] - Step 4942, Loss: {'train/loss': 0.03581788018345833, 'train/grad_norm_pre_clip': 70.97606658935547, 'timing/data': 0.000267869996605441, 'timing/model': 0.36687189200893044, 'learning_rate/qwen_vl_interface': 5.032838033392686e-07, 'learning_rate/action_model': 5.032838033392685e-06, 'epoch': 0.36}) +[2026-10-03 15:34:26,422][starVLA.training.train_starvla][INFO] - Step 4943, Loss: {'train/loss': 0.046896666288375854, 'train/grad_norm_pre_clip': 193.2054901123047, 'timing/data': 0.00026446999981999397, 'timing/model': 0.3669102620042395, 'learning_rate/qwen_vl_interface': 5.031715573958111e-07, 'learning_rate/action_model': 5.031715573958111e-06, 'epoch': 0.36}) +[2026-10-03 15:34:27,069][starVLA.training.train_starvla][INFO] - Step 4944, Loss: {'train/loss': 3.519983511068858e-05, 'train/grad_norm_pre_clip': 0.013683435507118702, 'timing/data': 0.0002495399967301637, 'timing/model': 0.3646852609963389, 'learning_rate/qwen_vl_interface': 5.030612626942235e-07, 'learning_rate/action_model': 5.030612626942235e-06, 'epoch': 0.36}) +[2026-10-03 15:34:27,717][starVLA.training.train_starvla][INFO] - Step 4945, Loss: {'train/loss': 0.13004879653453827, 'train/grad_norm_pre_clip': 144.4720458984375, 'timing/data': 0.00019526999676600099, 'timing/model': 0.3651168419892201, 'learning_rate/qwen_vl_interface': 5.029529192798433e-07, 'learning_rate/action_model': 5.029529192798433e-06, 'epoch': 0.36}) +[2026-10-03 15:34:28,367][starVLA.training.train_starvla][INFO] - Step 4946, Loss: {'train/loss': 0.3185150921344757, 'train/grad_norm_pre_clip': 99.01580810546875, 'timing/data': 0.00021314999321475625, 'timing/model': 0.3668746710027335, 'learning_rate/qwen_vl_interface': 5.028465271972058e-07, 'learning_rate/action_model': 5.028465271972058e-06, 'epoch': 0.36}) +[2026-10-03 15:34:29,213][starVLA.training.train_starvla][INFO] - Step 4947, Loss: {'train/loss': 0.00015559131861664355, 'train/grad_norm_pre_clip': 0.15599505603313446, 'timing/data': 0.00031629999284632504, 'timing/model': 0.5614507659920491, 'learning_rate/qwen_vl_interface': 5.027420864900457e-07, 'learning_rate/action_model': 5.027420864900456e-06, 'epoch': 0.36}) +[2026-10-03 15:34:29,860][starVLA.training.train_starvla][INFO] - Step 4948, Loss: {'train/loss': 7.458156323991716e-05, 'train/grad_norm_pre_clip': 0.03550317510962486, 'timing/data': 0.00024527999630663544, 'timing/model': 0.36455848200421315, 'learning_rate/qwen_vl_interface': 5.026395972012943e-07, 'learning_rate/action_model': 5.0263959720129436e-06, 'epoch': 0.36}) +[2026-10-03 15:34:30,508][starVLA.training.train_starvla][INFO] - Step 4949, Loss: {'train/loss': 6.119765021139756e-05, 'train/grad_norm_pre_clip': 0.019751064479351044, 'timing/data': 0.00026903899561148137, 'timing/model': 0.36503509200701956, 'learning_rate/qwen_vl_interface': 5.025390593730809e-07, 'learning_rate/action_model': 5.025390593730809e-06, 'epoch': 0.36}) +[2026-10-03 15:34:31,158][starVLA.training.train_starvla][INFO] - Step 4950, Loss: {'train/loss': 0.1342829018831253, 'train/grad_norm_pre_clip': 98.03123474121094, 'timing/data': 0.0001934300089487806, 'timing/model': 0.36687058200186584, 'learning_rate/qwen_vl_interface': 5.024404730467325e-07, 'learning_rate/action_model': 5.024404730467325e-06, 'epoch': 0.36}) +[2026-10-03 15:34:31,809][starVLA.training.train_starvla][INFO] - Step 4951, Loss: {'train/loss': 5.8984987845178694e-05, 'train/grad_norm_pre_clip': 0.020540187135338783, 'timing/data': 0.0001965300034498796, 'timing/model': 0.36658842099132016, 'learning_rate/qwen_vl_interface': 5.023438382627749e-07, 'learning_rate/action_model': 5.0234383826277486e-06, 'epoch': 0.36}) +[2026-10-03 15:34:32,629][starVLA.training.train_starvla][INFO] - Step 4952, Loss: {'train/loss': 0.00043394131353124976, 'train/grad_norm_pre_clip': 0.36574071645736694, 'timing/data': 0.00031692000629846007, 'timing/model': 0.5592424259957625, 'learning_rate/qwen_vl_interface': 5.022491550609311e-07, 'learning_rate/action_model': 5.022491550609311e-06, 'epoch': 0.36}) +[2026-10-03 15:34:33,277][starVLA.training.train_starvla][INFO] - Step 4953, Loss: {'train/loss': 0.4163540303707123, 'train/grad_norm_pre_clip': 31.509963989257812, 'timing/data': 0.00028036000730935484, 'timing/model': 0.3648574329999974, 'learning_rate/qwen_vl_interface': 5.021564234801213e-07, 'learning_rate/action_model': 5.021564234801213e-06, 'epoch': 0.36}) +[2026-10-03 15:34:33,927][starVLA.training.train_starvla][INFO] - Step 4954, Loss: {'train/loss': 3.9133112295530736e-05, 'train/grad_norm_pre_clip': 0.012659412808716297, 'timing/data': 0.00024326000129804015, 'timing/model': 0.3665817919973051, 'learning_rate/qwen_vl_interface': 5.020656435584636e-07, 'learning_rate/action_model': 5.020656435584635e-06, 'epoch': 0.36}) +[2026-10-03 15:34:34,578][starVLA.training.train_starvla][INFO] - Step 4955, Loss: {'train/loss': 5.660829265252687e-05, 'train/grad_norm_pre_clip': 0.016351662576198578, 'timing/data': 0.000272149991360493, 'timing/model': 0.3668451909907162, 'learning_rate/qwen_vl_interface': 5.019768153332752e-07, 'learning_rate/action_model': 5.019768153332752e-06, 'epoch': 0.36}) +[2026-10-03 15:34:35,203][starVLA.training.train_starvla][INFO] - Step 4956, Loss: {'train/loss': 0.0003998295869678259, 'train/grad_norm_pre_clip': 0.07219836115837097, 'timing/data': 0.00023854999744798988, 'timing/model': 0.3423879750043852, 'learning_rate/qwen_vl_interface': 5.018899388410693e-07, 'learning_rate/action_model': 5.0188993884106935e-06, 'epoch': 0.36}) +[2026-10-03 15:34:35,852][starVLA.training.train_starvla][INFO] - Step 4957, Loss: {'train/loss': 9.055624104803428e-05, 'train/grad_norm_pre_clip': 0.0708392783999443, 'timing/data': 0.000202410010388121, 'timing/model': 0.3651042820129078, 'learning_rate/qwen_vl_interface': 5.018050141175576e-07, 'learning_rate/action_model': 5.018050141175575e-06, 'epoch': 0.36}) +[2026-10-03 15:34:36,694][starVLA.training.train_starvla][INFO] - Step 4958, Loss: {'train/loss': 0.01774304173886776, 'train/grad_norm_pre_clip': 33.06472396850586, 'timing/data': 0.00025363000167999417, 'timing/model': 0.36684301200148184, 'learning_rate/qwen_vl_interface': 5.017220411976496e-07, 'learning_rate/action_model': 5.0172204119764965e-06, 'epoch': 0.36}) +[2026-10-03 15:34:37,346][starVLA.training.train_starvla][INFO] - Step 4959, Loss: {'train/loss': 0.0010576944332569838, 'train/grad_norm_pre_clip': 0.42626333236694336, 'timing/data': 0.0002340200007893145, 'timing/model': 0.366828671991243, 'learning_rate/qwen_vl_interface': 5.016410201154523e-07, 'learning_rate/action_model': 5.016410201154523e-06, 'epoch': 0.36}) +[2026-10-03 15:34:37,994][starVLA.training.train_starvla][INFO] - Step 4960, Loss: {'train/loss': 0.0006312895566225052, 'train/grad_norm_pre_clip': 0.4666810631752014, 'timing/data': 0.0002399200020590797, 'timing/model': 0.3646095119911479, 'learning_rate/qwen_vl_interface': 5.015619509042709e-07, 'learning_rate/action_model': 5.015619509042709e-06, 'epoch': 0.36}) +[2026-10-03 15:34:38,641][starVLA.training.train_starvla][INFO] - Step 4961, Loss: {'train/loss': 0.0004063145024701953, 'train/grad_norm_pre_clip': 0.5112414956092834, 'timing/data': 0.00034830000367946923, 'timing/model': 0.36441624200961087, 'learning_rate/qwen_vl_interface': 5.014848335966067e-07, 'learning_rate/action_model': 5.014848335966067e-06, 'epoch': 0.36}) +[2026-10-03 15:34:39,290][starVLA.training.train_starvla][INFO] - Step 4962, Loss: {'train/loss': 4.227034150972031e-05, 'train/grad_norm_pre_clip': 0.012718034908175468, 'timing/data': 0.00023677000717725605, 'timing/model': 0.36664135099272244, 'learning_rate/qwen_vl_interface': 5.014096682241608e-07, 'learning_rate/action_model': 5.014096682241608e-06, 'epoch': 0.36}) +[2026-10-03 15:34:40,129][starVLA.training.train_starvla][INFO] - Step 4963, Loss: {'train/loss': 5.851626701769419e-05, 'train/grad_norm_pre_clip': 0.03490746021270752, 'timing/data': 0.0002757299953373149, 'timing/model': 0.5540684269944904, 'learning_rate/qwen_vl_interface': 5.013364548178299e-07, 'learning_rate/action_model': 5.013364548178299e-06, 'epoch': 0.36}) +[2026-10-03 15:34:40,777][starVLA.training.train_starvla][INFO] - Step 4964, Loss: {'train/loss': 4.324572364566848e-05, 'train/grad_norm_pre_clip': 0.01355217769742012, 'timing/data': 0.0002407300053164363, 'timing/model': 0.3648284910013899, 'learning_rate/qwen_vl_interface': 5.012651934077101e-07, 'learning_rate/action_model': 5.0126519340771e-06, 'epoch': 0.36}) +[2026-10-03 15:34:41,425][starVLA.training.train_starvla][INFO] - Step 4965, Loss: {'train/loss': 1.686769064690452e-05, 'train/grad_norm_pre_clip': 0.00538315623998642, 'timing/data': 0.00023506999423261732, 'timing/model': 0.3645750220021, 'learning_rate/qwen_vl_interface': 5.01195884023094e-07, 'learning_rate/action_model': 5.0119588402309395e-06, 'epoch': 0.36}) +[2026-10-03 15:34:42,073][starVLA.training.train_starvla][INFO] - Step 4966, Loss: {'train/loss': 9.332509944215417e-05, 'train/grad_norm_pre_clip': 0.06235663220286369, 'timing/data': 0.00024152999685611576, 'timing/model': 0.3662874520086916, 'learning_rate/qwen_vl_interface': 5.011285266924714e-07, 'learning_rate/action_model': 5.011285266924714e-06, 'epoch': 0.36}) +[2026-10-03 15:34:42,726][starVLA.training.train_starvla][INFO] - Step 4967, Loss: {'train/loss': 4.594915662892163e-05, 'train/grad_norm_pre_clip': 0.01642206311225891, 'timing/data': 0.0001996100036194548, 'timing/model': 0.3672290719987359, 'learning_rate/qwen_vl_interface': 5.010631214435312e-07, 'learning_rate/action_model': 5.010631214435312e-06, 'epoch': 0.36}) +[2026-10-03 15:34:43,540][starVLA.training.train_starvla][INFO] - Step 4968, Loss: {'train/loss': 5.829330257256515e-05, 'train/grad_norm_pre_clip': 0.017133697867393494, 'timing/data': 0.00024311000015586615, 'timing/model': 0.5294363010034431, 'learning_rate/qwen_vl_interface': 5.009996683031586e-07, 'learning_rate/action_model': 5.009996683031586e-06, 'epoch': 0.36}) +[2026-10-03 15:34:44,188][starVLA.training.train_starvla][INFO] - Step 4969, Loss: {'train/loss': 0.00041927554411813617, 'train/grad_norm_pre_clip': 0.15264883637428284, 'timing/data': 0.000248910000664182, 'timing/model': 0.3647789420065237, 'learning_rate/qwen_vl_interface': 5.009381672974375e-07, 'learning_rate/action_model': 5.009381672974375e-06, 'epoch': 0.36}) +[2026-10-03 15:34:44,838][starVLA.training.train_starvla][INFO] - Step 4970, Loss: {'train/loss': 2.285645496158395e-05, 'train/grad_norm_pre_clip': 0.01522009540349245, 'timing/data': 0.0002996800030814484, 'timing/model': 0.36678761200164445, 'learning_rate/qwen_vl_interface': 5.008786184516475e-07, 'learning_rate/action_model': 5.008786184516474e-06, 'epoch': 0.36}) +[2026-10-03 15:34:45,499][starVLA.training.train_starvla][INFO] - Step 4971, Loss: {'train/loss': 1.945282201631926e-05, 'train/grad_norm_pre_clip': 0.00582668324932456, 'timing/data': 0.0002171300002373755, 'timing/model': 0.37077440100256354, 'learning_rate/qwen_vl_interface': 5.008210217902676e-07, 'learning_rate/action_model': 5.008210217902676e-06, 'epoch': 0.36}) +[2026-10-03 15:34:46,159][starVLA.training.train_starvla][INFO] - Step 4972, Loss: {'train/loss': 0.31212547421455383, 'train/grad_norm_pre_clip': 211.48829650878906, 'timing/data': 0.00021072999516036361, 'timing/model': 0.37165439101227093, 'learning_rate/qwen_vl_interface': 5.007653773369737e-07, 'learning_rate/action_model': 5.007653773369736e-06, 'epoch': 0.36}) +[2026-10-03 15:34:46,796][starVLA.training.train_starvla][INFO] - Step 4973, Loss: {'train/loss': 0.0007247234461829066, 'train/grad_norm_pre_clip': 0.23145432770252228, 'timing/data': 0.0003320499963592738, 'timing/model': 0.3712672920082696, 'learning_rate/qwen_vl_interface': 5.007116851146386e-07, 'learning_rate/action_model': 5.007116851146386e-06, 'epoch': 0.36}) +[2026-10-03 15:34:47,636][starVLA.training.train_starvla][INFO] - Step 4974, Loss: {'train/loss': 0.00040621592779643834, 'train/grad_norm_pre_clip': 0.07205326855182648, 'timing/data': 0.0002723499928833917, 'timing/model': 0.3475712440122152, 'learning_rate/qwen_vl_interface': 5.006599451453334e-07, 'learning_rate/action_model': 5.006599451453334e-06, 'epoch': 0.36}) +[2026-10-03 15:34:48,296][starVLA.training.train_starvla][INFO] - Step 4975, Loss: {'train/loss': 0.000578150968067348, 'train/grad_norm_pre_clip': 0.13178886473178864, 'timing/data': 0.0002830099983839318, 'timing/model': 0.3708199910033727, 'learning_rate/qwen_vl_interface': 5.006101574503268e-07, 'learning_rate/action_model': 5.006101574503267e-06, 'epoch': 0.36}) +[2026-10-03 15:34:48,957][starVLA.training.train_starvla][INFO] - Step 4976, Loss: {'train/loss': 4.952702875016257e-05, 'train/grad_norm_pre_clip': 0.012520112097263336, 'timing/data': 0.000212589991861023, 'timing/model': 0.37180518099921755, 'learning_rate/qwen_vl_interface': 5.005623220500839e-07, 'learning_rate/action_model': 5.005623220500839e-06, 'epoch': 0.36}) +[2026-10-03 15:34:49,616][starVLA.training.train_starvla][INFO] - Step 4977, Loss: {'train/loss': 0.03274308890104294, 'train/grad_norm_pre_clip': 196.1638946533203, 'timing/data': 0.00022681998962070793, 'timing/model': 0.37139251100597903, 'learning_rate/qwen_vl_interface': 5.005164389642687e-07, 'learning_rate/action_model': 5.005164389642687e-06, 'epoch': 0.36}) +[2026-10-03 15:34:50,276][starVLA.training.train_starvla][INFO] - Step 4978, Loss: {'train/loss': 0.00021228566765785217, 'train/grad_norm_pre_clip': 0.14121009409427643, 'timing/data': 0.00038123999547678977, 'timing/model': 0.37007204099791124, 'learning_rate/qwen_vl_interface': 5.004725082117423e-07, 'learning_rate/action_model': 5.004725082117422e-06, 'epoch': 0.36}) +[2026-10-03 15:34:51,129][starVLA.training.train_starvla][INFO] - Step 4979, Loss: {'train/loss': 0.0003403391456231475, 'train/grad_norm_pre_clip': 0.06483390182256699, 'timing/data': 0.0002642600011313334, 'timing/model': 0.5644039750040974, 'learning_rate/qwen_vl_interface': 5.004305298105616e-07, 'learning_rate/action_model': 5.004305298105616e-06, 'epoch': 0.36}) +[2026-10-03 15:34:51,790][starVLA.training.train_starvla][INFO] - Step 4980, Loss: {'train/loss': 3.507473593344912e-05, 'train/grad_norm_pre_clip': 0.00971846841275692, 'timing/data': 0.0002270199911436066, 'timing/model': 0.37211185100022703, 'learning_rate/qwen_vl_interface': 5.003905037779834e-07, 'learning_rate/action_model': 5.0039050377798345e-06, 'epoch': 0.36}) +[2026-10-03 15:34:52,405][starVLA.training.train_starvla][INFO] - Step 4981, Loss: {'train/loss': 7.10733438609168e-05, 'train/grad_norm_pre_clip': 0.05210602283477783, 'timing/data': 0.00026417999470140785, 'timing/model': 0.3486254040035419, 'learning_rate/qwen_vl_interface': 5.003524301304608e-07, 'learning_rate/action_model': 5.003524301304608e-06, 'epoch': 0.36}) +[2026-10-03 15:34:53,067][starVLA.training.train_starvla][INFO] - Step 4982, Loss: {'train/loss': 4.374270429252647e-05, 'train/grad_norm_pre_clip': 0.01229187473654747, 'timing/data': 0.0002850400051102042, 'timing/model': 0.3719483909953851, 'learning_rate/qwen_vl_interface': 5.003163088836445e-07, 'learning_rate/action_model': 5.003163088836446e-06, 'epoch': 0.36}) +[2026-10-03 15:34:53,704][starVLA.training.train_starvla][INFO] - Step 4983, Loss: {'train/loss': 9.544035492581315e-06, 'train/grad_norm_pre_clip': 0.004734020680189133, 'timing/data': 0.00021531000675167888, 'timing/model': 0.37084692099597305, 'learning_rate/qwen_vl_interface': 5.002821400523818e-07, 'learning_rate/action_model': 5.002821400523818e-06, 'epoch': 0.36}) +[2026-10-03 15:34:54,551][starVLA.training.train_starvla][INFO] - Step 4984, Loss: {'train/loss': 1.7172962543554604e-05, 'train/grad_norm_pre_clip': 0.004578634165227413, 'timing/data': 0.0004190990002825856, 'timing/model': 0.5574834470025962, 'learning_rate/qwen_vl_interface': 5.002499236507192e-07, 'learning_rate/action_model': 5.0024992365071925e-06, 'epoch': 0.36}) +[2026-10-03 15:34:55,209][starVLA.training.train_starvla][INFO] - Step 4985, Loss: {'train/loss': 0.000536261242814362, 'train/grad_norm_pre_clip': 0.1754063069820404, 'timing/data': 0.00033676000020932406, 'timing/model': 0.36950524199346546, 'learning_rate/qwen_vl_interface': 5.002196596918993e-07, 'learning_rate/action_model': 5.002196596918993e-06, 'epoch': 0.36}) +[2026-10-03 15:34:55,866][starVLA.training.train_starvla][INFO] - Step 4986, Loss: {'train/loss': 0.31828340888023376, 'train/grad_norm_pre_clip': 69.74363708496094, 'timing/data': 0.00021315000776667148, 'timing/model': 0.36997186100052204, 'learning_rate/qwen_vl_interface': 5.001913481883623e-07, 'learning_rate/action_model': 5.001913481883623e-06, 'epoch': 0.36}) +[2026-10-03 15:34:56,524][starVLA.training.train_starvla][INFO] - Step 4987, Loss: {'train/loss': 0.0005950559861958027, 'train/grad_norm_pre_clip': 0.23277242481708527, 'timing/data': 0.00030347000574693084, 'timing/model': 0.37000926199834794, 'learning_rate/qwen_vl_interface': 5.001649891517459e-07, 'learning_rate/action_model': 5.001649891517459e-06, 'epoch': 0.37}) +[2026-10-03 15:34:57,182][starVLA.training.train_starvla][INFO] - Step 4988, Loss: {'train/loss': 0.000448265258455649, 'train/grad_norm_pre_clip': 0.28743889927864075, 'timing/data': 0.00021192998974584043, 'timing/model': 0.3699986210122006, 'learning_rate/qwen_vl_interface': 5.001405825928858e-07, 'learning_rate/action_model': 5.0014058259288575e-06, 'epoch': 0.37}) +[2026-10-03 15:34:58,034][starVLA.training.train_starvla][INFO] - Step 4989, Loss: {'train/loss': 4.186477235634811e-05, 'train/grad_norm_pre_clip': 0.03703544661402702, 'timing/data': 0.00034093999420292675, 'timing/model': 0.5636876360076712, 'learning_rate/qwen_vl_interface': 5.001181285218143e-07, 'learning_rate/action_model': 5.0011812852181434e-06, 'epoch': 0.37}) +[2026-10-03 15:34:58,691][starVLA.training.train_starvla][INFO] - Step 4990, Loss: {'train/loss': 3.397871842025779e-05, 'train/grad_norm_pre_clip': 0.009995454922318459, 'timing/data': 0.00034902000334113836, 'timing/model': 0.369344230988645, 'learning_rate/qwen_vl_interface': 5.000976269477613e-07, 'learning_rate/action_model': 5.0009762694776125e-06, 'epoch': 0.37}) +[2026-10-03 15:34:59,326][starVLA.training.train_starvla][INFO] - Step 4991, Loss: {'train/loss': 0.0033921042922884226, 'train/grad_norm_pre_clip': 7.4845123291015625, 'timing/data': 0.0003098900051554665, 'timing/model': 0.34686739499738906, 'learning_rate/qwen_vl_interface': 5.000790778791542e-07, 'learning_rate/action_model': 5.000790778791543e-06, 'epoch': 0.37}) +[2026-10-03 15:34:59,981][starVLA.training.train_starvla][INFO] - Step 4992, Loss: {'train/loss': 0.3435322642326355, 'train/grad_norm_pre_clip': 10.802048683166504, 'timing/data': 0.00033255999733228236, 'timing/model': 0.36659655200492125, 'learning_rate/qwen_vl_interface': 5.000624813236185e-07, 'learning_rate/action_model': 5.000624813236185e-06, 'epoch': 0.37}) +[2026-10-03 15:35:00,629][starVLA.training.train_starvla][INFO] - Step 4993, Loss: {'train/loss': 8.98534563020803e-06, 'train/grad_norm_pre_clip': 0.0019791547674685717, 'timing/data': 0.0001977399952011183, 'timing/model': 0.36460068200540263, 'learning_rate/qwen_vl_interface': 5.000478372879752e-07, 'learning_rate/action_model': 5.000478372879752e-06, 'epoch': 0.37}) +[2026-10-03 15:35:01,255][starVLA.training.train_starvla][INFO] - Step 4994, Loss: {'train/loss': 0.00030670559499412775, 'train/grad_norm_pre_clip': 0.7855512499809265, 'timing/data': 0.00034093999420292675, 'timing/model': 0.3651355809997767, 'learning_rate/qwen_vl_interface': 5.000351457782452e-07, 'learning_rate/action_model': 5.000351457782451e-06, 'epoch': 0.37}) +[2026-10-03 15:35:02,100][starVLA.training.train_starvla][INFO] - Step 4995, Loss: {'train/loss': 4.237634493620135e-05, 'train/grad_norm_pre_clip': 0.02740430273115635, 'timing/data': 0.0002865189890144393, 'timing/model': 0.3648981919977814, 'learning_rate/qwen_vl_interface': 5.000244067996445e-07, 'learning_rate/action_model': 5.000244067996445e-06, 'epoch': 0.37}) +[2026-10-03 15:35:02,726][starVLA.training.train_starvla][INFO] - Step 4996, Loss: {'train/loss': 0.29935503005981445, 'train/grad_norm_pre_clip': 89.99073791503906, 'timing/data': 0.00023149000480771065, 'timing/model': 0.3648355619952781, 'learning_rate/qwen_vl_interface': 5.000156203565884e-07, 'learning_rate/action_model': 5.000156203565883e-06, 'epoch': 0.37}) +[2026-10-03 15:35:03,373][starVLA.training.train_starvla][INFO] - Step 4997, Loss: {'train/loss': 0.14336971938610077, 'train/grad_norm_pre_clip': 234.33547973632812, 'timing/data': 0.0002461399999447167, 'timing/model': 0.36461041298753116, 'learning_rate/qwen_vl_interface': 5.000087864526878e-07, 'learning_rate/action_model': 5.000087864526878e-06, 'epoch': 0.37}) +[2026-10-03 15:35:04,022][starVLA.training.train_starvla][INFO] - Step 4998, Loss: {'train/loss': 0.00016512579168193042, 'train/grad_norm_pre_clip': 0.05917973443865776, 'timing/data': 0.0002453299966873601, 'timing/model': 0.36550497100688517, 'learning_rate/qwen_vl_interface': 5.000039050907523e-07, 'learning_rate/action_model': 5.000039050907523e-06, 'epoch': 0.37}) +[2026-10-03 15:35:04,670][starVLA.training.train_starvla][INFO] - Step 4999, Loss: {'train/loss': 0.7362297177314758, 'train/grad_norm_pre_clip': 46.582176208496094, 'timing/data': 0.0002420999953756109, 'timing/model': 0.36494296201271936, 'learning_rate/qwen_vl_interface': 5.000009762727885e-07, 'learning_rate/action_model': 5.000009762727884e-06, 'epoch': 0.37}) +[2026-10-03 15:35:05,518][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 5000 +[2026-10-03 15:35:30,830][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 5000 +[2026-10-03 15:35:30,830][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 5000 +[2026-10-03 15:35:30,830][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 5000 +[2026-10-03 15:35:30,831][starVLA.training.train_starvla][INFO] - Step 5000, Loss: {'train/loss': 0.25075680017471313, 'train/grad_norm_pre_clip': 89.25469207763672, 'eval/action_loss/seconds': 25.31004265799129, 'eval/action_loss/samples': 120.0, 'eval/loss': 1.006046175956726, 'eval/latency_0/loss': 1.461010456085205, 'eval/latency_2/loss': 1.4413483142852783, 'eval/latency_4/loss': 0.11578015238046646, 'eval/inverted_pendulum_rgb_state/loss': 1.006046175956726, 'eval/inverted_pendulum_rgb_state/latency_0/loss': 1.461010456085205, 'eval/inverted_pendulum_rgb_state/latency_2/loss': 1.4413483142852783, 'eval/inverted_pendulum_rgb_state/latency_4/loss': 0.11578015238046646, 'timing/data': 0.00039149999793153256, 'timing/model': 0.3650884020025842, 'learning_rate/qwen_vl_interface': 5e-07, 'learning_rate/action_model': 4.9999999999999996e-06, 'epoch': 0.37}) +[2026-10-03 15:35:30,898][root][INFO] - /data/ubuntu/conda_envs/starvla_rl_games_openvla/bin/x86_64-conda-linux-gnu-cc -march=nocona -mtune=haswell -ftree-vectorize -fPIC -fstack-protector-strong -fno-plt -O2 -ffunction-sections -pipe -isystem /data/ubuntu/conda_envs/starvla_rl_games_openvla/include -I/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/include -L/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/lib -L/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/lib/stubs -DNDEBUG -D_FORTIFY_SOURCE=2 -O2 -isystem /data/ubuntu/conda_envs/starvla_rl_games_openvla/include -I/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/include -L/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/lib -L/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/lib/stubs -fPIC -march=nocona -mtune=haswell -ftree-vectorize -fPIC -fstack-protector-strong -fno-plt -O2 -ffunction-sections -pipe -isystem /data/ubuntu/conda_envs/starvla_rl_games_openvla/include -I/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/include -L/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/lib -L/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/lib/stubs -c /tmp/tmper40lzh9/test.c -o /tmp/tmper40lzh9/test.o +[2026-10-03 15:35:30,907][root][INFO] - /data/ubuntu/conda_envs/starvla_rl_games_openvla/bin/x86_64-conda-linux-gnu-cc -Wl,-O2 -Wl,--sort-common -Wl,--as-needed -Wl,-z,relro -Wl,-z,now -Wl,--disable-new-dtags -Wl,--gc-sections -Wl,--allow-shlib-undefined -Wl,-rpath,/data/ubuntu/conda_envs/starvla_rl_games_openvla/lib -Wl,-rpath-link,/data/ubuntu/conda_envs/starvla_rl_games_openvla/lib -L/data/ubuntu/conda_envs/starvla_rl_games_openvla/lib -L/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/lib -L/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/lib/stubs /tmp/tmper40lzh9/test.o -laio -o /tmp/tmper40lzh9/a.out +[2026-10-03 15:35:30,915][root][INFO] - /data/ubuntu/conda_envs/starvla_rl_games_openvla/bin/x86_64-conda-linux-gnu-cc -march=nocona -mtune=haswell -ftree-vectorize -fPIC -fstack-protector-strong -fno-plt -O2 -ffunction-sections -pipe -isystem /data/ubuntu/conda_envs/starvla_rl_games_openvla/include -I/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/include -L/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/lib -L/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/lib/stubs -DNDEBUG -D_FORTIFY_SOURCE=2 -O2 -isystem /data/ubuntu/conda_envs/starvla_rl_games_openvla/include -I/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/include -L/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/lib -L/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/lib/stubs -fPIC -march=nocona -mtune=haswell -ftree-vectorize -fPIC -fstack-protector-strong -fno-plt -O2 -ffunction-sections -pipe -isystem /data/ubuntu/conda_envs/starvla_rl_games_openvla/include -I/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/include -L/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/lib -L/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/lib/stubs -c /tmp/tmp9_dzba6z/test.c -o /tmp/tmp9_dzba6z/test.o +[2026-10-03 15:35:30,923][root][INFO] - /data/ubuntu/conda_envs/starvla_rl_games_openvla/bin/x86_64-conda-linux-gnu-cc -Wl,-O2 -Wl,--sort-common -Wl,--as-needed -Wl,-z,relro -Wl,-z,now -Wl,--disable-new-dtags -Wl,--gc-sections -Wl,--allow-shlib-undefined -Wl,-rpath,/data/ubuntu/conda_envs/starvla_rl_games_openvla/lib -Wl,-rpath-link,/data/ubuntu/conda_envs/starvla_rl_games_openvla/lib -L/data/ubuntu/conda_envs/starvla_rl_games_openvla/lib -L/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/lib -L/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/lib/stubs /tmp/tmp9_dzba6z/test.o -L/data/ubuntu/conda_envs/starvla_rl_games_openvla -L/data/ubuntu/conda_envs/starvla_rl_games_openvla/lib64 -lcufile -o /tmp/tmp9_dzba6z/a.out +[2026-10-03 15:35:30,936][root][INFO] - /data/ubuntu/conda_envs/starvla_rl_games_openvla/bin/x86_64-conda-linux-gnu-cc -march=nocona -mtune=haswell -ftree-vectorize -fPIC -fstack-protector-strong -fno-plt -O2 -ffunction-sections -pipe -isystem /data/ubuntu/conda_envs/starvla_rl_games_openvla/include -I/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/include -L/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/lib -L/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/lib/stubs -DNDEBUG -D_FORTIFY_SOURCE=2 -O2 -isystem /data/ubuntu/conda_envs/starvla_rl_games_openvla/include -I/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/include -L/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/lib -L/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/lib/stubs -fPIC -march=nocona -mtune=haswell -ftree-vectorize -fPIC -fstack-protector-strong -fno-plt -O2 -ffunction-sections -pipe -isystem /data/ubuntu/conda_envs/starvla_rl_games_openvla/include -I/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/include -L/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/lib -L/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/lib/stubs -c /tmp/tmpkfdvu4ni/test.c -o /tmp/tmpkfdvu4ni/test.o +[2026-10-03 15:35:30,943][root][INFO] - /data/ubuntu/conda_envs/starvla_rl_games_openvla/bin/x86_64-conda-linux-gnu-cc -Wl,-O2 -Wl,--sort-common -Wl,--as-needed -Wl,-z,relro -Wl,-z,now -Wl,--disable-new-dtags -Wl,--gc-sections -Wl,--allow-shlib-undefined -Wl,-rpath,/data/ubuntu/conda_envs/starvla_rl_games_openvla/lib -Wl,-rpath-link,/data/ubuntu/conda_envs/starvla_rl_games_openvla/lib -L/data/ubuntu/conda_envs/starvla_rl_games_openvla/lib -L/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/lib -L/data/ubuntu/conda_envs/starvla_rl_games_openvla/targets/x86_64-linux/lib/stubs /tmp/tmpkfdvu4ni/test.o -laio -o /tmp/tmpkfdvu4ni/a.out +[2026-10-03 15:35:35,390][starVLA.training.rl_games.checkpoint_sync][WARNING] - HF checkpoint sync failed for repo talha15032/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2: (Request ID: Root=1-6ac120c7-008f11f93e9f95e5123df384;44ca0c47-6cfd-4210-9ebc-4bb7ffa60ee7) + +403 Forbidden: You don't have the rights to create a model under the namespace "talha15032". +Cannot access content at: https://huggingface.co/api/repos/create. +Make sure your token has the correct permissions. +[2026-10-03 15:35:35,390][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-10-03 15:35:35,412][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-10-03 15:35:35,413][starVLA.training.train_starvla][INFO] - Training complete. Final training state is saved under /home/ubuntu/talha/results/Checkpoints/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2/checkpoints diff --git a/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2/summary.jsonl b/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2/summary.jsonl new file mode 100644 index 0000000000000000000000000000000000000000..04f41c132dfb80a0ccb65212db19ce96e77fd03c --- /dev/null +++ b/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/openvla_inverted_pendulum_mixed_latency_024_og_l0_native_no_latency_prompt_exp2/summary.jsonl @@ -0,0 +1 @@ +{"steps": 5000} diff --git a/zero-latency/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_0_200ep/checkpoints/steps_7000_model.safetensors b/zero-latency/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_0_200ep/checkpoints/steps_7000_model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..a8c995c30649e35b4ef34987495a299ed5e37d18 --- /dev/null +++ b/zero-latency/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_0_200ep/checkpoints/steps_7000_model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:25a15d0460a63a7ca036e9300d39a4f57b8f9e656c6ff42941348b6a570bf767 +size 9138230516 diff --git a/zero-latency/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_0_200ep/config.full.yaml b/zero-latency/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_0_200ep/config.full.yaml new file mode 100644 index 0000000000000000000000000000000000000000..4213697c0079691dba722c336de50535e7d2cc11 --- /dev/null +++ b/zero-latency/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_0_200ep/config.full.yaml @@ -0,0 +1,209 @@ +framework: + name: QwenOFT + qwenvl: + base_vlm: /workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + attn_implementation: flash_attention_2 + enable_gradient_checkpointing: true + action_model: + action_model_type: MLP + action_dim: 7 + action_hidden_dim: 2560 + future_action_window_size: 0 + past_action_window_size: 0 + loss_type: discrete_ce + state_dim: 7 + action_horizon: 1 + action_env_dim: 6 +datasets: + vla_data: + dataset_py: lerobot_datasets + include_state: true + data_root_dir: /workspace/starVLA/data/demon_attack_fix_latency_0_200ep + data_mix: demon_attack_train__bridge + eval_data_mix: demon_attack_train__bridge__val + custom_mixtures_path: null + action_type: discrete + sequential_step_sampling: false + shuffle: true + action_balance: + enabled: false + strategy: balanced_epoch + action_key: action_id + target_flap_fraction: 0.3 + noop_id: 0 + flap_id: 1 + latency_curriculum: + enabled: false + strategy: exclusive + latencies: null + phase_steps: null + per_device_batch_size: 32 + load_all_data_for_training: true + obs_image_size: null + video_backend: torchvision_av +dataset: + source_hf: '' + config_name: null + source_subdir: null + converted_name: demon_attack_train + single_source_hf: '' + mixed_source_hf: '' + single_converted_name: demon_attack_train + mixed_converted_name: demon_attack_mixed_latency_train + single_latency_filter: null + mixed_latency_filter: null + force_download: false + setup_force: false + skip_verification: false + verify_rows: 200 + max_episodes: null + episodes_per_latency: null + latency_filter: null + debug_subset: + enabled: false + max_episodes: 5 + suffix: debug +base_model: + repo_id: Qwen/Qwen3-VL-4B-Instruct +initialization: + checkpoint_local_dir: playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_hf_repo_id: StarVLA/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_filename: checkpoints/steps_5000_pytorch_model.pt +trainer: + max_train_steps: 7000 + num_warmup_steps: 100 + save_interval: 100 + eval_interval: 100 + eval_num_batches: 100 + per_latency_eval_num_batches: null + eval_action_classification: true + cc_f1_tolerance: 1 + learning_rate: + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + action_model: 0.0001 + lr_scheduler_type: cosine_with_min_lr + scheduler_specific_kwargs: + min_lr: 1.0e-06 + freeze_modules: '' + loss_scale: + vla: 1.0 + vlm: 0.1 + max_grad_norm: 1.0 + weight_decay: 0.0 + logging_frequency: 1 + gradient_clipping: 1.0 + gradient_accumulation_steps: 4 + distributed_backend: none + is_resume: false + pretrained_checkpoint: null + resume_step: 0 + reload_modules: null + optimizer: + name: AdamW + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + weight_decay: 1.0e-08 + fused: false + save_format: pt +workspace_dir: WORKSPACE_DIR +run_root_dir: /workspace/starVLA/results/Checkpoints +seed: 42 +wandb_entity: dongqianyu99-zhejiang-university +wandb_project: starVLA_rl_games +auth: + env_file: null + hf_token_env: HF_TOKEN + wandb_api_key_env: WANDB_API_KEY +paths: + run_root_dir: results/Checkpoints + dataset_local_dir: playground/Datasets/rl_games + dataset_cache_dir: null + base_model_dir: playground/Pretrained_models/Qwen3-VL-4B-Instruct + accelerate_config: starVLA/config/deepseeds/deepspeed_zero2.yaml +launch: + use_accelerate: true + gpus: null + num_processes: 1 + dry_run: false +conda: + enabled: true + env_name: null +rl_games: + model_alias: openvla + env_eval: + image_size: 224 + frameskip: 4 + seed: 42 + fixed_episode_seeds: true + latency_seed_stride: 0 + task_seed_stride: 0 + task_description: 'You are playing Demon Attack from a single game image. Choose + exactly one action from: NOOP, FIRE, RIGHT, LEFT, RIGHTFIRE, LEFTFIRE.' + demon_attack: + noop_max: 30 + enabled: true + distributed_mode: none + vectorized: + enabled: false + batch_size: 1 + latency: + prompt_map_path: /workspace/starVLA/data/demon_attack_fix_latency_0_200ep/demon_attack_train__bridge/latency_prompt_map.json + mode: single + values: + - 0 + mid_train: + enabled: true + interval_steps: 500 + latencies: + - 0 + num_episodes: 5 + max_steps_per_episode: 3600 + post_train: + enabled: false + latencies: + - 0 + - 1 + - 2 + - 3 + - 4 + num_episodes: 5 + max_steps_per_episode: 3600 + task: demon_attack + initialization_mode: bridge + action_carrier: bridge +model: openvla +env: demon_attack +init: bridge +bridge_base_model: + repo_id: + openvla: Qwen/Qwen3-VL-4B-Instruct + pi0: StarVLA/Qwen2.5-VL-3B-Instruct-Action + pi05: Qwen/Qwen3-VL-4B-Instruct + gr00t: Qwen/Qwen3-VL-4B-Instruct + local_dir: + openvla: playground/Pretrained_models/Qwen3-VL-4B-Instruct + pi0: playground/Pretrained_models/Qwen2.5-VL-3B-Instruct-Action + pi05: playground/Pretrained_models/Qwen3-VL-4B-Instruct + gr00t: playground/Pretrained_models/Qwen3-VL-4B-Instruct +mode: single +checkpoint: + load: auto + hf_repo_id: null + save_best_model: false + save_pt_file: false + local: + keep_last_n: 1 + sync: + enabled: false + repo_id: null + keep_last_n: 0 + sync_every_n_checkpoints: 1 + resume_policy: local_latest +run_id: demon_attack_fix_latency_0_200ep +output_dir: /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep +config_yaml: null +is_debug: false +version_id: '0.21' diff --git a/zero-latency/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_0_200ep/config.yaml b/zero-latency/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_0_200ep/config.yaml new file mode 100644 index 0000000000000000000000000000000000000000..12e5a90c95efdeb5e98ee90c2b29736c95b1edca --- /dev/null +++ b/zero-latency/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_0_200ep/config.yaml @@ -0,0 +1,97 @@ +checkpoint: + local: + keep_last_n: 1 + save_best_model: false + save_pt_file: false + sync: + enabled: false + keep_last_n: 0 + repo_id: null +datasets: + vla_data: + data_mix: demon_attack_train__bridge + dataset_py: lerobot_datasets + eval_data_mix: demon_attack_train__bridge__val + include_state: true + latency_curriculum: + enabled: false + obs_image_size: null + per_device_batch_size: 32 +framework: + action_model: + action_dim: 7 + action_env_dim: 6 + action_hidden_dim: 2560 + action_horizon: 1 + action_model_type: MLP + loss_type: discrete_ce + state_dim: 7 + name: QwenOFT + qwenvl: + attn_implementation: flash_attention_2 + base_vlm: /workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + enable_gradient_checkpointing: true +output_dir: /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep +rl_games: + env_eval: + demon_attack: + noop_max: 30 + distributed_mode: none + enabled: true + fixed_episode_seeds: true + frameskip: 4 + image_size: 224 + latency: + prompt_map_path: /workspace/starVLA/data/demon_attack_fix_latency_0_200ep/demon_attack_train__bridge/latency_prompt_map.json + latency_seed_stride: 0 + mid_train: + enabled: true + interval_steps: 500 + latencies: + - 0 + max_steps_per_episode: 3600 + num_episodes: 5 + seed: 42 + task_description: 'You are playing Demon Attack from a single game image. Choose + exactly one action from: NOOP, FIRE, RIGHT, LEFT, RIGHTFIRE, LEFTFIRE.' + task_seed_stride: 0 + vectorized: + enabled: false + model_alias: openvla + task: demon_attack +run_id: demon_attack_fix_latency_0_200ep +run_root_dir: /workspace/starVLA/results/Checkpoints +seed: 42 +trainer: + cc_f1_tolerance: 1 + distributed_backend: none + eval_action_classification: true + eval_interval: 100 + eval_num_batches: 100 + freeze_modules: '' + gradient_accumulation_steps: 4 + gradient_clipping: 1.0 + is_resume: false + learning_rate: + action_model: 0.0001 + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + logging_frequency: 1 + lr_scheduler_type: cosine_with_min_lr + max_train_steps: 7000 + num_warmup_steps: 100 + optimizer: + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + fused: false + weight_decay: 1.0e-08 + per_latency_eval_num_batches: null + pretrained_checkpoint: null + reload_modules: null + save_interval: 100 + scheduler_specific_kwargs: + min_lr: 1.0e-06 +wandb_entity: dongqianyu99-zhejiang-university +wandb_project: starVLA_rl_games diff --git a/zero-latency/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_0_200ep/dataset_statistics.json b/zero-latency/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_0_200ep/dataset_statistics.json new file mode 100644 index 0000000000000000000000000000000000000000..7febd6939621463780cd1080aa0c1db18b578c41 --- /dev/null +++ b/zero-latency/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_0_200ep/dataset_statistics.json @@ -0,0 +1,127 @@ +{ + "new_embodiment": { + "action": { + "mean": [ + 0.009211494587361813, + 0.00885747093707323, + 0.17263218760490417, + 0.1680275797843933, + 0.3163609206676483, + 0.32491034269332886, + 0.0 + ], + "std": [ + 0.09535734355449677, + 0.09356050193309784, + 0.3784366250038147, + 0.3730704188346863, + 0.46596139669418335, + 0.4689920246601105, + 0.0 + ], + "max": [ + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 0.0, + 0.0, + 1.0, + 1.0, + 1.0, + 1.0, + 0.0 + ], + "mask": [ + true, + true, + true, + true, + true, + true, + true + ] + }, + "state": { + "mean": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ] + }, + "num_transitions": 435000, + "num_trajectories": 180 + } +} \ No newline at end of file diff --git a/zero-latency/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_0_200ep/dataset_statistics_eval.json b/zero-latency/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_0_200ep/dataset_statistics_eval.json new file mode 100644 index 0000000000000000000000000000000000000000..6e4c2d4796156a19030e26dbaa2c79b94d94f38b --- /dev/null +++ b/zero-latency/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_0_200ep/dataset_statistics_eval.json @@ -0,0 +1,127 @@ +{ + "new_embodiment": { + "action": { + "mean": [ + 0.009733290411531925, + 0.009130473248660564, + 0.17506690323352814, + 0.17256741225719452, + 0.3143470585346222, + 0.31915488839149475, + 0.0 + ], + "std": [ + 0.09819413721561432, + 0.09510758519172668, + 0.3800537884235382, + 0.37778180837631226, + 0.46411585807800293, + 0.466117799282074, + 0.0 + ], + "max": [ + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + 1.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 0.0, + 0.0, + 1.0, + 1.0, + 1.0, + 1.0, + 0.0 + ], + "mask": [ + true, + true, + true, + true, + true, + true, + true + ] + }, + "state": { + "mean": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ] + }, + "num_transitions": 68014, + "num_trajectories": 20 + } +} \ No newline at end of file diff --git a/zero-latency/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_0_200ep/hydra/.hydra/config.yaml b/zero-latency/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_0_200ep/hydra/.hydra/config.yaml new file mode 100644 index 0000000000000000000000000000000000000000..c4b70220a7c5d6c67de0638a907852a0cb81e8bd --- /dev/null +++ b/zero-latency/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_0_200ep/hydra/.hydra/config.yaml @@ -0,0 +1,207 @@ +framework: + qwenvl: + base_vlm: /workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + attn_implementation: flash_attention_2 + enable_gradient_checkpointing: true + action_model: + state_dim: 7 + loss_type: discrete_ce + action_horizon: 1 + future_action_window_size: 0 + past_action_window_size: 0 + action_dim: 7 + action_env_dim: 6 + name: QwenOFT +datasets: + vla_data: + dataset_py: lerobot_datasets + include_state: true + data_root_dir: /workspace/starVLA/data/demon_attack_fix_latency_0_200ep + data_mix: demon_attack_train__bridge + eval_data_mix: demon_attack_train__bridge__val + custom_mixtures_path: null + action_type: discrete + sequential_step_sampling: false + shuffle: true + action_balance: + enabled: false + strategy: balanced_epoch + action_key: action_id + target_flap_fraction: 0.3 + noop_id: 0 + flap_id: 1 + latency_curriculum: + enabled: false + strategy: exclusive + latencies: null + phase_steps: null + per_device_batch_size: 32 + load_all_data_for_training: true + obs_image_size: null + video_backend: torchvision_av +dataset: + source_hf: ${dataset.single_source_hf} + config_name: null + source_subdir: null + converted_name: ${dataset.single_converted_name} + single_source_hf: '' + mixed_source_hf: '' + single_converted_name: demon_attack_train + mixed_converted_name: demon_attack_mixed_latency_train + single_latency_filter: null + mixed_latency_filter: null + force_download: false + setup_force: false + skip_verification: false + verify_rows: 200 + max_episodes: null + episodes_per_latency: null + latency_filter: ${dataset.single_latency_filter} + debug_subset: + enabled: false + max_episodes: 5 + suffix: debug +base_model: + repo_id: ${bridge_base_model.repo_id.${model}} +initialization: + checkpoint_local_dir: playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_hf_repo_id: StarVLA/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_filename: checkpoints/steps_5000_pytorch_model.pt +trainer: + max_train_steps: 7000 + num_warmup_steps: 100 + save_interval: 100 + eval_interval: 100 + eval_num_batches: 100 + per_latency_eval_num_batches: null + eval_action_classification: true + cc_f1_tolerance: 1 + learning_rate: + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + action_model: 0.0001 + lr_scheduler_type: cosine_with_min_lr + scheduler_specific_kwargs: + min_lr: 1.0e-06 + freeze_modules: '' + loss_scale: + vla: 1.0 + vlm: 0.1 + max_grad_norm: 1.0 + weight_decay: 0.0 + logging_frequency: 1 + gradient_clipping: 1.0 + gradient_accumulation_steps: 4 + distributed_backend: none + is_resume: false + pretrained_checkpoint: /workspace/starVLA/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt + resume_step: 0 + reload_modules: null + optimizer: + name: AdamW + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + weight_decay: 1.0e-08 + fused: false + save_format: pt +workspace_dir: WORKSPACE_DIR +run_root_dir: /workspace/starVLA/results/Checkpoints +seed: 42 +wandb_entity: dongqianyu99-zhejiang-university +wandb_project: starVLA_rl_games +auth: + env_file: null + hf_token_env: HF_TOKEN + wandb_api_key_env: WANDB_API_KEY +paths: + run_root_dir: results/Checkpoints + dataset_local_dir: playground/Datasets/rl_games + dataset_cache_dir: null + base_model_dir: ${bridge_base_model.local_dir.${model}} + accelerate_config: starVLA/config/deepseeds/deepspeed_zero2.yaml +launch: + use_accelerate: true + gpus: null + num_processes: 1 + dry_run: false +conda: + enabled: true + env_name: null +rl_games: + model_alias: openvla + env_eval: + image_size: 224 + frameskip: 4 + seed: 42 + fixed_episode_seeds: true + latency_seed_stride: 0 + task_seed_stride: 0 + task_description: 'You are playing Demon Attack from a single game image. Choose + exactly one action from: NOOP, FIRE, RIGHT, LEFT, RIGHTFIRE, LEFTFIRE.' + demon_attack: + noop_max: 30 + enabled: true + distributed_mode: none + vectorized: + enabled: false + batch_size: 1 + latency: + prompt_map_path: /workspace/starVLA/data/demon_attack_fix_latency_0_200ep/demon_attack_train__bridge/latency_prompt_map.json + mode: single + values: + - 0 + mid_train: + enabled: true + interval_steps: 500 + latencies: + - 0 + num_episodes: 5 + max_steps_per_episode: 3600 + post_train: + enabled: false + latencies: + - 0 + - 1 + - 2 + - 3 + - 4 + num_episodes: 5 + max_steps_per_episode: 3600 + task: demon_attack + initialization_mode: bridge + action_carrier: bridge +model: openvla +env: demon_attack +init: bridge +bridge_base_model: + repo_id: + openvla: Qwen/Qwen3-VL-4B-Instruct + pi0: StarVLA/Qwen2.5-VL-3B-Instruct-Action + pi05: Qwen/Qwen3-VL-4B-Instruct + gr00t: Qwen/Qwen3-VL-4B-Instruct + local_dir: + openvla: playground/Pretrained_models/Qwen3-VL-4B-Instruct + pi0: playground/Pretrained_models/Qwen2.5-VL-3B-Instruct-Action + pi05: playground/Pretrained_models/Qwen3-VL-4B-Instruct + gr00t: playground/Pretrained_models/Qwen3-VL-4B-Instruct +mode: single +checkpoint: + load: auto + hf_repo_id: null + save_best_model: false + save_pt_file: false + local: + keep_last_n: 1 + sync: + enabled: false + repo_id: null + keep_last_n: 0 + sync_every_n_checkpoints: 1 + resume_policy: local_latest +run_id: demon_attack_fix_latency_0_200ep +output_dir: null +config_yaml: null +is_debug: false +version_id: '0.21' diff --git a/zero-latency/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_0_200ep/hydra/.hydra/hydra.yaml b/zero-latency/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_0_200ep/hydra/.hydra/hydra.yaml new file mode 100644 index 0000000000000000000000000000000000000000..e4881485112c39b9309fb436c9c156e1b58ee6fa --- /dev/null +++ b/zero-latency/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_0_200ep/hydra/.hydra/hydra.yaml @@ -0,0 +1,259 @@ +hydra: + run: + dir: ${run_root_dir}/${run_id}/hydra + sweep: + dir: multirun/${now:%Y-%m-%d}/${now:%H-%M-%S} + subdir: ${hydra.job.num} + launcher: + _target_: hydra._internal.core_plugins.basic_launcher.BasicLauncher + sweeper: + _target_: hydra._internal.core_plugins.basic_sweeper.BasicSweeper + max_batch_size: null + params: null + help: + app_name: ${hydra.job.name} + header: '${hydra.help.app_name} is powered by Hydra. + + ' + footer: 'Powered by Hydra (https://hydra.cc) + + Use --hydra-help to view Hydra specific help + + ' + template: '${hydra.help.header} + + == Configuration groups == + + Compose your configuration from those groups (group=option) + + + $APP_CONFIG_GROUPS + + + == Config == + + Override anything in the config (foo.bar=value) + + + $CONFIG + + + ${hydra.help.footer} + + ' + hydra_help: + template: 'Hydra (${hydra.runtime.version}) + + See https://hydra.cc for more info. + + + == Flags == + + $FLAGS_HELP + + + == Configuration groups == + + Compose your configuration from those groups (For example, append hydra/job_logging=disabled + to command line) + + + $HYDRA_CONFIG_GROUPS + + + Use ''--cfg hydra'' to Show the Hydra config. + + ' + hydra_help: ??? + hydra_logging: + version: 1 + formatters: + simple: + format: '[%(asctime)s][HYDRA] %(message)s' + handlers: + console: + class: logging.StreamHandler + formatter: simple + stream: ext://sys.stdout + root: + level: INFO + handlers: + - console + loggers: + logging_example: + level: DEBUG + disable_existing_loggers: false + job_logging: + version: 1 + formatters: + simple: + format: '[%(asctime)s][%(name)s][%(levelname)s] - %(message)s' + handlers: + console: + class: logging.StreamHandler + formatter: simple + stream: ext://sys.stdout + file: + class: logging.FileHandler + formatter: simple + filename: ${hydra.runtime.output_dir}/${hydra.job.name}.log + root: + level: INFO + handlers: + - console + - file + disable_existing_loggers: false + env: {} + mode: RUN + searchpath: [] + callbacks: {} + output_subdir: .hydra + overrides: + hydra: + - hydra.mode=RUN + task: + - model=openvla + - env=demon_attack + - init=bridge + - mode=single + - run_id=demon_attack_fix_latency_0_200ep + - run_root_dir=/workspace/starVLA/results/Checkpoints + - seed=42 + - wandb_entity=dongqianyu99-zhejiang-university + - wandb_project=starVLA_rl_games + - rl_games.env_eval.enabled=true + - checkpoint.sync.enabled=false + - checkpoint.sync.keep_last_n=0 + - checkpoint.local.keep_last_n=1 + - checkpoint.save_best_model=false + - checkpoint.save_pt_file=false + - trainer.is_resume=false + - trainer.pretrained_checkpoint=/workspace/starVLA/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt + - trainer.resume_step=0 + - trainer.max_train_steps=7000 + - trainer.num_warmup_steps=100 + - trainer.save_interval=100 + - trainer.eval_interval=100 + - trainer.eval_num_batches=100 + - trainer.logging_frequency=1 + - trainer.gradient_accumulation_steps=4 + - trainer.distributed_backend=none + - trainer.learning_rate.base=2e-05 + - trainer.learning_rate.qwen_vl_interface=1e-05 + - trainer.learning_rate.action_model=0.0001 + - trainer.lr_scheduler_type=cosine_with_min_lr + - trainer.scheduler_specific_kwargs.min_lr=1e-06 + - trainer.loss_scale.vla=1.0 + - trainer.loss_scale.vlm=0.1 + - trainer.max_grad_norm=1.0 + - trainer.weight_decay=0.0 + - trainer.gradient_clipping=1.0 + - trainer.optimizer.name=AdamW + - trainer.optimizer.betas=[0.9,0.95] + - trainer.optimizer.eps=1e-08 + - trainer.optimizer.weight_decay=1e-08 + - trainer.optimizer.fused=false + - trainer.save_format=pt + - framework.name=QwenOFT + - framework.qwenvl.attn_implementation=flash_attention_2 + - framework.qwenvl.enable_gradient_checkpointing=true + - framework.action_model.action_dim=7 + - framework.action_model.action_env_dim=6 + - framework.action_model.state_dim=7 + - framework.action_model.loss_type=discrete_ce + - framework.action_model.action_horizon=1 + - framework.action_model.future_action_window_size=0 + - framework.action_model.past_action_window_size=0 + - datasets.vla_data.include_state=true + - datasets.vla_data.action_type=discrete + - datasets.vla_data.sequential_step_sampling=false + - datasets.vla_data.shuffle=true + - datasets.vla_data.action_balance.enabled=false + - datasets.vla_data.action_balance.strategy=balanced_epoch + - datasets.vla_data.action_balance.action_key=action_id + - datasets.vla_data.action_balance.target_flap_fraction=0.3 + - datasets.vla_data.action_balance.noop_id=0 + - datasets.vla_data.action_balance.flap_id=1 + - datasets.vla_data.latency_curriculum.enabled=false + - datasets.vla_data.latency_curriculum.strategy=exclusive + - datasets.vla_data.per_device_batch_size=32 + - datasets.vla_data.load_all_data_for_training=true + - datasets.vla_data.video_backend=torchvision_av + - datasets.vla_data.data_root_dir=/workspace/starVLA/data/demon_attack_fix_latency_0_200ep + - datasets.vla_data.data_mix=demon_attack_train__bridge + - datasets.vla_data.eval_data_mix=demon_attack_train__bridge__val + - framework.qwenvl.base_vlm=/workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + - rl_games.task=demon_attack + - rl_games.model_alias=openvla + - rl_games.initialization_mode=bridge + - rl_games.action_carrier=bridge + - rl_games.env_eval.distributed_mode=none + - rl_games.env_eval.latency.mode=single + - rl_games.env_eval.frameskip=4 + - rl_games.env_eval.image_size=224 + - rl_games.env_eval.seed=42 + - rl_games.env_eval.fixed_episode_seeds=true + - rl_games.env_eval.latency_seed_stride=0 + - rl_games.env_eval.task_seed_stride=0 + - 'rl_games.env_eval.task_description=''You are playing Demon Attack from a single + game image. Choose exactly one action from: NOOP, FIRE, RIGHT, LEFT, RIGHTFIRE, + LEFTFIRE.''' + - rl_games.env_eval.latency.values=[0] + - rl_games.env_eval.latency.prompt_map_path=/workspace/starVLA/data/demon_attack_fix_latency_0_200ep/demon_attack_train__bridge/latency_prompt_map.json + - rl_games.env_eval.mid_train.enabled=true + - rl_games.env_eval.mid_train.interval_steps=500 + - rl_games.env_eval.mid_train.num_episodes=5 + - rl_games.env_eval.mid_train.max_steps_per_episode=3600 + - rl_games.env_eval.mid_train.latencies=[0] + - rl_games.env_eval.post_train.enabled=false + - rl_games.env_eval.post_train.num_episodes=5 + - rl_games.env_eval.post_train.max_steps_per_episode=3600 + - rl_games.env_eval.post_train.latencies=[0,1,2,3,4] + job: + name: train_starvla_hydra + chdir: false + override_dirname: 'checkpoint.local.keep_last_n=1,checkpoint.save_best_model=false,checkpoint.save_pt_file=false,checkpoint.sync.enabled=false,checkpoint.sync.keep_last_n=0,datasets.vla_data.action_balance.action_key=action_id,datasets.vla_data.action_balance.enabled=false,datasets.vla_data.action_balance.flap_id=1,datasets.vla_data.action_balance.noop_id=0,datasets.vla_data.action_balance.strategy=balanced_epoch,datasets.vla_data.action_balance.target_flap_fraction=0.3,datasets.vla_data.action_type=discrete,datasets.vla_data.data_mix=demon_attack_train__bridge,datasets.vla_data.data_root_dir=/workspace/starVLA/data/demon_attack_fix_latency_0_200ep,datasets.vla_data.eval_data_mix=demon_attack_train__bridge__val,datasets.vla_data.include_state=true,datasets.vla_data.latency_curriculum.enabled=false,datasets.vla_data.latency_curriculum.strategy=exclusive,datasets.vla_data.load_all_data_for_training=true,datasets.vla_data.per_device_batch_size=32,datasets.vla_data.sequential_step_sampling=false,datasets.vla_data.shuffle=true,datasets.vla_data.video_backend=torchvision_av,env=demon_attack,framework.action_model.action_dim=7,framework.action_model.action_env_dim=6,framework.action_model.action_horizon=1,framework.action_model.future_action_window_size=0,framework.action_model.loss_type=discrete_ce,framework.action_model.past_action_window_size=0,framework.action_model.state_dim=7,framework.name=QwenOFT,framework.qwenvl.attn_implementation=flash_attention_2,framework.qwenvl.base_vlm=/workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct,framework.qwenvl.enable_gradient_checkpointing=true,init=bridge,mode=single,model=openvla,rl_games.action_carrier=bridge,rl_games.env_eval.distributed_mode=none,rl_games.env_eval.enabled=true,rl_games.env_eval.fixed_episode_seeds=true,rl_games.env_eval.frameskip=4,rl_games.env_eval.image_size=224,rl_games.env_eval.latency.mode=single,rl_games.env_eval.latency.prompt_map_path=/workspace/starVLA/data/demon_attack_fix_latency_0_200ep/demon_attack_train__bridge/latency_prompt_map.json,rl_games.env_eval.latency.values=[0],rl_games.env_eval.latency_seed_stride=0,rl_games.env_eval.mid_train.enabled=true,rl_games.env_eval.mid_train.interval_steps=500,rl_games.env_eval.mid_train.latencies=[0],rl_games.env_eval.mid_train.max_steps_per_episode=3600,rl_games.env_eval.mid_train.num_episodes=5,rl_games.env_eval.post_train.enabled=false,rl_games.env_eval.post_train.latencies=[0,1,2,3,4],rl_games.env_eval.post_train.max_steps_per_episode=3600,rl_games.env_eval.post_train.num_episodes=5,rl_games.env_eval.seed=42,rl_games.env_eval.task_description=''You + are playing Demon Attack from a single game image. Choose exactly one action + from: NOOP, FIRE, RIGHT, LEFT, RIGHTFIRE, LEFTFIRE.'',rl_games.env_eval.task_seed_stride=0,rl_games.initialization_mode=bridge,rl_games.model_alias=openvla,rl_games.task=demon_attack,run_id=demon_attack_fix_latency_0_200ep,run_root_dir=/workspace/starVLA/results/Checkpoints,seed=42,trainer.distributed_backend=none,trainer.eval_interval=100,trainer.eval_num_batches=100,trainer.gradient_accumulation_steps=4,trainer.gradient_clipping=1.0,trainer.is_resume=false,trainer.learning_rate.action_model=0.0001,trainer.learning_rate.base=2e-05,trainer.learning_rate.qwen_vl_interface=1e-05,trainer.logging_frequency=1,trainer.loss_scale.vla=1.0,trainer.loss_scale.vlm=0.1,trainer.lr_scheduler_type=cosine_with_min_lr,trainer.max_grad_norm=1.0,trainer.max_train_steps=7000,trainer.num_warmup_steps=100,trainer.optimizer.betas=[0.9,0.95],trainer.optimizer.eps=1e-08,trainer.optimizer.fused=false,trainer.optimizer.name=AdamW,trainer.optimizer.weight_decay=1e-08,trainer.pretrained_checkpoint=/workspace/starVLA/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt,trainer.resume_step=0,trainer.save_format=pt,trainer.save_interval=100,trainer.scheduler_specific_kwargs.min_lr=1e-06,trainer.weight_decay=0.0,wandb_entity=dongqianyu99-zhejiang-university,wandb_project=starVLA_rl_games' + id: ??? + num: ??? + config_name: train + env_set: {} + env_copy: [] + config: + override_dirname: + kv_sep: '=' + item_sep: ',' + exclude_keys: [] + runtime: + version: 1.3.3 + version_base: '1.1' + cwd: /workspace/starVLA + config_sources: + - path: hydra.conf + schema: pkg + provider: hydra + - path: /workspace/starVLA/examples/rl_games/config + schema: file + provider: main + - path: '' + schema: structured + provider: schema + output_dir: /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/hydra + choices: + cross_task_setup: null + checkpoint: default + mode: single + init: bridge + env: demon_attack + model: openvla + hydra/env: default + hydra/callbacks: null + hydra/job_logging: default + hydra/hydra_logging: default + hydra/hydra_help: default + hydra/help: default + hydra/sweeper: basic + hydra/launcher: basic + hydra/output: default + verbose: false diff --git a/zero-latency/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_0_200ep/hydra/.hydra/overrides.yaml b/zero-latency/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_0_200ep/hydra/.hydra/overrides.yaml new file mode 100644 index 0000000000000000000000000000000000000000..9904de61773098b1da33e67af68f1c52753fa1fb --- /dev/null +++ b/zero-latency/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_0_200ep/hydra/.hydra/overrides.yaml @@ -0,0 +1,97 @@ +- model=openvla +- env=demon_attack +- init=bridge +- mode=single +- run_id=demon_attack_fix_latency_0_200ep +- run_root_dir=/workspace/starVLA/results/Checkpoints +- seed=42 +- wandb_entity=dongqianyu99-zhejiang-university +- wandb_project=starVLA_rl_games +- rl_games.env_eval.enabled=true +- checkpoint.sync.enabled=false +- checkpoint.sync.keep_last_n=0 +- checkpoint.local.keep_last_n=1 +- checkpoint.save_best_model=false +- checkpoint.save_pt_file=false +- trainer.is_resume=false +- trainer.pretrained_checkpoint=/workspace/starVLA/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt +- trainer.resume_step=0 +- trainer.max_train_steps=7000 +- trainer.num_warmup_steps=100 +- trainer.save_interval=100 +- trainer.eval_interval=100 +- trainer.eval_num_batches=100 +- trainer.logging_frequency=1 +- trainer.gradient_accumulation_steps=4 +- trainer.distributed_backend=none +- trainer.learning_rate.base=2e-05 +- trainer.learning_rate.qwen_vl_interface=1e-05 +- trainer.learning_rate.action_model=0.0001 +- trainer.lr_scheduler_type=cosine_with_min_lr +- trainer.scheduler_specific_kwargs.min_lr=1e-06 +- trainer.loss_scale.vla=1.0 +- trainer.loss_scale.vlm=0.1 +- trainer.max_grad_norm=1.0 +- trainer.weight_decay=0.0 +- trainer.gradient_clipping=1.0 +- trainer.optimizer.name=AdamW +- trainer.optimizer.betas=[0.9,0.95] +- trainer.optimizer.eps=1e-08 +- trainer.optimizer.weight_decay=1e-08 +- trainer.optimizer.fused=false +- trainer.save_format=pt +- framework.name=QwenOFT +- framework.qwenvl.attn_implementation=flash_attention_2 +- framework.qwenvl.enable_gradient_checkpointing=true +- framework.action_model.action_dim=7 +- framework.action_model.action_env_dim=6 +- framework.action_model.state_dim=7 +- framework.action_model.loss_type=discrete_ce +- framework.action_model.action_horizon=1 +- framework.action_model.future_action_window_size=0 +- framework.action_model.past_action_window_size=0 +- datasets.vla_data.include_state=true +- datasets.vla_data.action_type=discrete +- datasets.vla_data.sequential_step_sampling=false +- datasets.vla_data.shuffle=true +- datasets.vla_data.action_balance.enabled=false +- datasets.vla_data.action_balance.strategy=balanced_epoch +- datasets.vla_data.action_balance.action_key=action_id +- datasets.vla_data.action_balance.target_flap_fraction=0.3 +- datasets.vla_data.action_balance.noop_id=0 +- datasets.vla_data.action_balance.flap_id=1 +- datasets.vla_data.latency_curriculum.enabled=false +- datasets.vla_data.latency_curriculum.strategy=exclusive +- datasets.vla_data.per_device_batch_size=32 +- datasets.vla_data.load_all_data_for_training=true +- datasets.vla_data.video_backend=torchvision_av +- datasets.vla_data.data_root_dir=/workspace/starVLA/data/demon_attack_fix_latency_0_200ep +- datasets.vla_data.data_mix=demon_attack_train__bridge +- datasets.vla_data.eval_data_mix=demon_attack_train__bridge__val +- framework.qwenvl.base_vlm=/workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct +- rl_games.task=demon_attack +- rl_games.model_alias=openvla +- rl_games.initialization_mode=bridge +- rl_games.action_carrier=bridge +- rl_games.env_eval.distributed_mode=none +- rl_games.env_eval.latency.mode=single +- rl_games.env_eval.frameskip=4 +- rl_games.env_eval.image_size=224 +- rl_games.env_eval.seed=42 +- rl_games.env_eval.fixed_episode_seeds=true +- rl_games.env_eval.latency_seed_stride=0 +- rl_games.env_eval.task_seed_stride=0 +- 'rl_games.env_eval.task_description=''You are playing Demon Attack from a single + game image. Choose exactly one action from: NOOP, FIRE, RIGHT, LEFT, RIGHTFIRE, + LEFTFIRE.''' +- rl_games.env_eval.latency.values=[0] +- rl_games.env_eval.latency.prompt_map_path=/workspace/starVLA/data/demon_attack_fix_latency_0_200ep/demon_attack_train__bridge/latency_prompt_map.json +- rl_games.env_eval.mid_train.enabled=true +- rl_games.env_eval.mid_train.interval_steps=500 +- rl_games.env_eval.mid_train.num_episodes=5 +- rl_games.env_eval.mid_train.max_steps_per_episode=3600 +- rl_games.env_eval.mid_train.latencies=[0] +- rl_games.env_eval.post_train.enabled=false +- rl_games.env_eval.post_train.num_episodes=5 +- rl_games.env_eval.post_train.max_steps_per_episode=3600 +- rl_games.env_eval.post_train.latencies=[0,1,2,3,4] diff --git a/zero-latency/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_0_200ep/hydra/train_starvla_hydra.log b/zero-latency/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_0_200ep/hydra/train_starvla_hydra.log new file mode 100644 index 0000000000000000000000000000000000000000..032572f0e0ef18ecac3190703d2a137effd85867 --- /dev/null +++ b/zero-latency/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_0_200ep/hydra/train_starvla_hydra.log @@ -0,0 +1,7766 @@ +[2026-06-17 06:12:34,226][starVLA.training.train_starvla][INFO] - VLA Training :: Warming Up +[2026-06-17 06:12:34,235][starVLA.training.train_starvla][INFO] - ✅ Configuration wrapped for access tracking +[2026-06-17 06:12:39,509][starVLA.training.train_starvla][INFO] - Preloaded model checkpoint before Accelerator/DeepSpeed initialization: /workspace/starVLA/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt +[2026-06-17 06:12:39,541][starVLA.training.train_starvla][INFO] - Creating VLA Dataset with Mixture `demon_attack_train__bridge` +[2026-06-17 06:12:40,067][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Behavior (data_config): ['BEHAVIOR_challenge'] +[2026-06-17 06:12:40,068][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from DOMINO: ['robotwin'] +[2026-06-17 06:12:40,068][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from DOMINO (data_config): [] +[2026-06-17 06:12:40,068][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from DOMINO (data_config): ['domino', 'domino_clean', 'domino_random', 'domino_cotrain'] +[2026-06-17 06:12:40,068][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Franka: ['custom_robot_config', 'demo_sim_franka_delta_joints', 'SO101'] +[2026-06-17 06:12:40,068][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Franka (data_config): [] +[2026-06-17 06:12:40,068][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Franka (data_config): ['custom_dataset', 'custom_dataset_2', 'demo_sim_pick_place', 'SO101_pick'] +[2026-06-17 06:12:40,068][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from LIBERO: ['libero_franka'] +[2026-06-17 06:12:40,068][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from LIBERO (data_config): [] +[2026-06-17 06:12:40,068][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from LIBERO (data_config): ['libero_all', 'libero_goal', 'multi_robot'] +[2026-06-17 06:12:40,068][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from RoboChallenge_table30v2: ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-17 06:12:40,068][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from RoboChallenge_table30v2 (data_config): ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-17 06:12:40,068][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from RoboChallenge_table30v2 (data_config): ['robochallenge_table30v2_shred_paper', 'robochallenge_table30v2_ur5_all', 'robochallenge_table30v2_arx5_all', 'robochallenge_table30v2_dosw1_all'] +[2026-06-17 06:12:40,068][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_365: ['panda_omron_robocasa365'] +[2026-06-17 06:12:40,068][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_365 (data_config): [] +[2026-06-17 06:12:40,068][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_365 (data_config): ['robocasa365_open_drawer_target_human', 'robocasa365_atomic_target_human_all', 'robocasa365_composite_target_human_all', 'robocasa365_target_human_all'] +[2026-06-17 06:12:40,068][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_tabletop: ['fourier_gr1_arms_waist'] +[2026-06-17 06:12:40,069][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_tabletop (data_config): [] +[2026-06-17 06:12:40,069][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_tabletop (data_config): ['fourier_gr1_unified_1000'] +[2026-06-17 06:12:40,069][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robotwin: ['robotwin', 'robotwin50', 'arx_x5'] +[2026-06-17 06:12:40,069][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robotwin (data_config): [] +[2026-06-17 06:12:40,069][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robotwin (data_config): ['robotwin_all', 'robotwin_all_50', 'robotwin', 'robotwin_task1', 'robotwin_task2', 'arx_x5'] +[2026-06-17 06:12:40,069][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from SimplerEnv: ['oxe_droid', 'oxe_bridge', 'oxe_rt1'] +[2026-06-17 06:12:40,069][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from SimplerEnv (data_config): [] +[2026-06-17 06:12:40,069][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from SimplerEnv (data_config): ['bridge', 'bridge_rt_1'] +[2026-06-17 06:12:40,069][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from VLA-Arena: ['vla_arena_franka'] +[2026-06-17 06:12:40,069][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from VLA-Arena (data_config): [] +[2026-06-17 06:12:40,069][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from VLA-Arena (data_config): ['vla_arena_L0_S', 'vla_arena_L0_M', 'vla_arena_L0_L'] +[2026-06-17 06:12:40,069][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from rl_games: ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-17 06:12:40,069][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from rl_games (data_config): ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-17 06:12:40,069][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from rl_games (data_config): ['flappy_train', 'flappy_train__bridge', 'flappy_mixed_latency_train', 'flappy_mixed_latency_train__bridge', 'demon_attack_train', 'demon_attack_train__bridge', 'demon_attack_mixed_latency_train', 'demon_attack_mixed_latency_train__bridge', 'deadly_corridor_train', 'deadly_corridor_train__bridge', 'deadly_corridor_mixed_latency_train', 'deadly_corridor_mixed_latency_train__bridge'] +[2026-06-17 06:12:40,102][starVLA.training.train_starvla][INFO] - Creating VLA Eval Dataset with Mixture `demon_attack_train__bridge__val` +[2026-06-17 06:12:40,130][starVLA.training.train_starvla][INFO] - 📝 Full config saved at /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/config.full.yaml +[2026-06-17 06:12:40,136][starVLA.training.train_starvla][INFO] - 📊 Accessed config snapshot saved at /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/config.yaml +[2026-06-17 06:12:40,136][starVLA.training.train_starvla][INFO] - No pretrained checkpoint provided. Starting training from scratch. +[2026-06-17 06:12:42,943][starVLA.training.train_starvla][INFO] - ***** Training Configuration ***** +[2026-06-17 06:12:42,943][starVLA.training.train_starvla][INFO] - Total optimization steps = 7000 +[2026-06-17 06:12:42,943][starVLA.training.train_starvla][INFO] - Per device batch size = 32 +[2026-06-17 06:12:42,943][starVLA.training.train_starvla][INFO] - Gradient accumulation steps = 4 +[2026-06-17 06:12:42,943][starVLA.training.train_starvla][INFO] - Total batch size = 128 +[2026-06-17 06:12:58,109][starVLA.training.train_starvla][INFO] - Step 1, Loss: {'train/loss': 1.765942096710205, 'train/grad_norm_pre_clip': 5.909287929534912, 'timing/data': 0.005427437950856984, 'timing/model': 0.9704256789991632, 'learning_rate/qwen_vl_interface': 1.0000000000000001e-07, 'learning_rate/action_model': 1.0000000000000002e-06, 'epoch': 0.0}) +[2026-06-17 06:13:01,764][starVLA.training.train_starvla][INFO] - Step 2, Loss: {'train/loss': 1.799099326133728, 'train/grad_norm_pre_clip': 5.838852882385254, 'timing/data': 0.0050287049962207675, 'timing/model': 0.9554097630316392, 'learning_rate/qwen_vl_interface': 2.0000000000000002e-07, 'learning_rate/action_model': 2.0000000000000003e-06, 'epoch': 0.0}) +[2026-06-17 06:13:05,405][starVLA.training.train_starvla][INFO] - Step 3, Loss: {'train/loss': 1.8063819408416748, 'train/grad_norm_pre_clip': 5.96204948425293, 'timing/data': 0.004633551929146051, 'timing/model': 0.949869039002806, 'learning_rate/qwen_vl_interface': 3.0000000000000004e-07, 'learning_rate/action_model': 3e-06, 'epoch': 0.0}) +[2026-06-17 06:13:09,080][starVLA.training.train_starvla][INFO] - Step 4, Loss: {'train/loss': 1.7493840456008911, 'train/grad_norm_pre_clip': 5.84586238861084, 'timing/data': 0.005218824953772128, 'timing/model': 0.9574469709768891, 'learning_rate/qwen_vl_interface': 4.0000000000000003e-07, 'learning_rate/action_model': 4.000000000000001e-06, 'epoch': 0.0}) +[2026-06-17 06:13:12,765][starVLA.training.train_starvla][INFO] - Step 5, Loss: {'train/loss': 1.7331013679504395, 'train/grad_norm_pre_clip': 5.58683967590332, 'timing/data': 0.005152080091647804, 'timing/model': 0.964058147976175, 'learning_rate/qwen_vl_interface': 5.000000000000001e-07, 'learning_rate/action_model': 5e-06, 'epoch': 0.0}) +[2026-06-17 06:13:16,433][starVLA.training.train_starvla][INFO] - Step 6, Loss: {'train/loss': 1.742085337638855, 'train/grad_norm_pre_clip': 8.12669563293457, 'timing/data': 0.0055234580067917705, 'timing/model': 0.9592478648992255, 'learning_rate/qwen_vl_interface': 6.000000000000001e-07, 'learning_rate/action_model': 6e-06, 'epoch': 0.0}) +[2026-06-17 06:13:20,102][starVLA.training.train_starvla][INFO] - Step 7, Loss: {'train/loss': 1.7063170671463013, 'train/grad_norm_pre_clip': 5.519586563110352, 'timing/data': 0.005250134970992804, 'timing/model': 0.9607808029977605, 'learning_rate/qwen_vl_interface': 7.000000000000001e-07, 'learning_rate/action_model': 7.000000000000001e-06, 'epoch': 0.0}) +[2026-06-17 06:13:23,805][starVLA.training.train_starvla][INFO] - Step 8, Loss: {'train/loss': 1.6758681535720825, 'train/grad_norm_pre_clip': 6.769966125488281, 'timing/data': 0.004932626034133136, 'timing/model': 0.9612020390341058, 'learning_rate/qwen_vl_interface': 8.000000000000001e-07, 'learning_rate/action_model': 8.000000000000001e-06, 'epoch': 0.0}) +[2026-06-17 06:13:27,482][starVLA.training.train_starvla][INFO] - Step 9, Loss: {'train/loss': 1.6399633884429932, 'train/grad_norm_pre_clip': 4.1098761558532715, 'timing/data': 0.0048758049961179495, 'timing/model': 0.9608742849668488, 'learning_rate/qwen_vl_interface': 9.000000000000001e-07, 'learning_rate/action_model': 9e-06, 'epoch': 0.0}) +[2026-06-17 06:13:31,265][starVLA.training.train_starvla][INFO] - Step 10, Loss: {'train/loss': 1.594245433807373, 'train/grad_norm_pre_clip': 4.79411506652832, 'timing/data': 0.004927441012114286, 'timing/model': 0.9622151500079781, 'learning_rate/qwen_vl_interface': 1.0000000000000002e-06, 'learning_rate/action_model': 1e-05, 'epoch': 0.0}) +[2026-06-17 06:13:34,958][starVLA.training.train_starvla][INFO] - Step 11, Loss: {'train/loss': 1.5468186140060425, 'train/grad_norm_pre_clip': 6.175330638885498, 'timing/data': 0.005391562939621508, 'timing/model': 0.9664135180646554, 'learning_rate/qwen_vl_interface': 1.1e-06, 'learning_rate/action_model': 1.1000000000000001e-05, 'epoch': 0.0}) +[2026-06-17 06:13:38,660][starVLA.training.train_starvla][INFO] - Step 12, Loss: {'train/loss': 1.5806705951690674, 'train/grad_norm_pre_clip': 3.7164247035980225, 'timing/data': 0.004993617068976164, 'timing/model': 0.9666743500856683, 'learning_rate/qwen_vl_interface': 1.2000000000000002e-06, 'learning_rate/action_model': 1.2e-05, 'epoch': 0.0}) +[2026-06-17 06:13:42,375][starVLA.training.train_starvla][INFO] - Step 13, Loss: {'train/loss': 1.5354396104812622, 'train/grad_norm_pre_clip': 3.0252583026885986, 'timing/data': 0.004719290998764336, 'timing/model': 0.9700163590023294, 'learning_rate/qwen_vl_interface': 1.3e-06, 'learning_rate/action_model': 1.3000000000000001e-05, 'epoch': 0.0}) +[2026-06-17 06:13:46,101][starVLA.training.train_starvla][INFO] - Step 14, Loss: {'train/loss': 1.5755318403244019, 'train/grad_norm_pre_clip': 7.66812801361084, 'timing/data': 0.004936696030199528, 'timing/model': 0.9645182169042528, 'learning_rate/qwen_vl_interface': 1.4000000000000001e-06, 'learning_rate/action_model': 1.4000000000000001e-05, 'epoch': 0.0}) +[2026-06-17 06:13:49,827][starVLA.training.train_starvla][INFO] - Step 15, Loss: {'train/loss': 1.4813178777694702, 'train/grad_norm_pre_clip': 2.4736461639404297, 'timing/data': 0.004374775104224682, 'timing/model': 0.9673580389935523, 'learning_rate/qwen_vl_interface': 1.5e-06, 'learning_rate/action_model': 1.5e-05, 'epoch': 0.0}) +[2026-06-17 06:13:53,541][starVLA.training.train_starvla][INFO] - Step 16, Loss: {'train/loss': 1.597292184829712, 'train/grad_norm_pre_clip': 3.6410536766052246, 'timing/data': 0.005283495062030852, 'timing/model': 0.976070664008148, 'learning_rate/qwen_vl_interface': 1.6000000000000001e-06, 'learning_rate/action_model': 1.6000000000000003e-05, 'epoch': 0.0}) +[2026-06-17 06:13:57,270][starVLA.training.train_starvla][INFO] - Step 17, Loss: {'train/loss': 1.432063102722168, 'train/grad_norm_pre_clip': 2.0716333389282227, 'timing/data': 0.005544680985622108, 'timing/model': 0.9785256360191852, 'learning_rate/qwen_vl_interface': 1.7000000000000002e-06, 'learning_rate/action_model': 1.7000000000000003e-05, 'epoch': 0.01}) +[2026-06-17 06:14:00,997][starVLA.training.train_starvla][INFO] - Step 18, Loss: {'train/loss': 1.4971965551376343, 'train/grad_norm_pre_clip': 2.603409767150879, 'timing/data': 0.004493449931032956, 'timing/model': 0.9683816799661145, 'learning_rate/qwen_vl_interface': 1.8000000000000001e-06, 'learning_rate/action_model': 1.8e-05, 'epoch': 0.01}) +[2026-06-17 06:14:04,746][starVLA.training.train_starvla][INFO] - Step 19, Loss: {'train/loss': 1.4862035512924194, 'train/grad_norm_pre_clip': 3.361445665359497, 'timing/data': 0.005957913002930582, 'timing/model': 0.9730242749210447, 'learning_rate/qwen_vl_interface': 1.9000000000000002e-06, 'learning_rate/action_model': 1.9e-05, 'epoch': 0.01}) +[2026-06-17 06:14:08,488][starVLA.training.train_starvla][INFO] - Step 20, Loss: {'train/loss': 1.2672958374023438, 'train/grad_norm_pre_clip': 4.192598342895508, 'timing/data': 0.005406938027590513, 'timing/model': 0.9717239460442215, 'learning_rate/qwen_vl_interface': 2.0000000000000003e-06, 'learning_rate/action_model': 2e-05, 'epoch': 0.01}) +[2026-06-17 06:14:12,211][starVLA.training.train_starvla][INFO] - Step 21, Loss: {'train/loss': 1.5116440057754517, 'train/grad_norm_pre_clip': 3.2047510147094727, 'timing/data': 0.0050416659796610475, 'timing/model': 0.971865119994618, 'learning_rate/qwen_vl_interface': 2.1000000000000002e-06, 'learning_rate/action_model': 2.1e-05, 'epoch': 0.01}) +[2026-06-17 06:14:15,937][starVLA.training.train_starvla][INFO] - Step 22, Loss: {'train/loss': 1.3815007209777832, 'train/grad_norm_pre_clip': 1.80289888381958, 'timing/data': 0.005495758028700948, 'timing/model': 0.9707425000378862, 'learning_rate/qwen_vl_interface': 2.2e-06, 'learning_rate/action_model': 2.2000000000000003e-05, 'epoch': 0.01}) +[2026-06-17 06:14:19,658][starVLA.training.train_starvla][INFO] - Step 23, Loss: {'train/loss': 1.3217928409576416, 'train/grad_norm_pre_clip': 3.670703411102295, 'timing/data': 0.004712054040282965, 'timing/model': 0.9710917809279636, 'learning_rate/qwen_vl_interface': 2.3000000000000004e-06, 'learning_rate/action_model': 2.3000000000000003e-05, 'epoch': 0.01}) +[2026-06-17 06:14:23,384][starVLA.training.train_starvla][INFO] - Step 24, Loss: {'train/loss': 1.4716907739639282, 'train/grad_norm_pre_clip': 3.3291680812835693, 'timing/data': 0.005296675022691488, 'timing/model': 0.9702605940401554, 'learning_rate/qwen_vl_interface': 2.4000000000000003e-06, 'learning_rate/action_model': 2.4e-05, 'epoch': 0.01}) +[2026-06-17 06:14:27,113][starVLA.training.train_starvla][INFO] - Step 25, Loss: {'train/loss': 1.5406726598739624, 'train/grad_norm_pre_clip': 2.044748306274414, 'timing/data': 0.005065995035693049, 'timing/model': 0.9743113300064579, 'learning_rate/qwen_vl_interface': 2.5e-06, 'learning_rate/action_model': 2.5e-05, 'epoch': 0.01}) +[2026-06-17 06:14:30,852][starVLA.training.train_starvla][INFO] - Step 26, Loss: {'train/loss': 1.3784496784210205, 'train/grad_norm_pre_clip': 3.1370537281036377, 'timing/data': 0.005292571964673698, 'timing/model': 0.977201457018964, 'learning_rate/qwen_vl_interface': 2.6e-06, 'learning_rate/action_model': 2.6000000000000002e-05, 'epoch': 0.01}) +[2026-06-17 06:14:34,584][starVLA.training.train_starvla][INFO] - Step 27, Loss: {'train/loss': 1.3235620260238647, 'train/grad_norm_pre_clip': 3.0117030143737793, 'timing/data': 0.005146358977071941, 'timing/model': 0.9727213449077681, 'learning_rate/qwen_vl_interface': 2.7000000000000004e-06, 'learning_rate/action_model': 2.7000000000000002e-05, 'epoch': 0.01}) +[2026-06-17 06:14:38,316][starVLA.training.train_starvla][INFO] - Step 28, Loss: {'train/loss': 1.3475474119186401, 'train/grad_norm_pre_clip': 2.545253038406372, 'timing/data': 0.004823201918043196, 'timing/model': 0.9745212440611795, 'learning_rate/qwen_vl_interface': 2.8000000000000003e-06, 'learning_rate/action_model': 2.8000000000000003e-05, 'epoch': 0.01}) +[2026-06-17 06:14:42,050][starVLA.training.train_starvla][INFO] - Step 29, Loss: {'train/loss': 1.5505503416061401, 'train/grad_norm_pre_clip': 2.5257437229156494, 'timing/data': 0.004916623001918197, 'timing/model': 0.9731136430054903, 'learning_rate/qwen_vl_interface': 2.9e-06, 'learning_rate/action_model': 2.9e-05, 'epoch': 0.01}) +[2026-06-17 06:14:45,784][starVLA.training.train_starvla][INFO] - Step 30, Loss: {'train/loss': 1.4613418579101562, 'train/grad_norm_pre_clip': 4.402811527252197, 'timing/data': 0.004868492949754, 'timing/model': 0.9738066459540278, 'learning_rate/qwen_vl_interface': 3e-06, 'learning_rate/action_model': 3e-05, 'epoch': 0.01}) +[2026-06-17 06:14:49,526][starVLA.training.train_starvla][INFO] - Step 31, Loss: {'train/loss': 1.413316249847412, 'train/grad_norm_pre_clip': 3.399016857147217, 'timing/data': 0.00521034502889961, 'timing/model': 0.9818997229449451, 'learning_rate/qwen_vl_interface': 3.1000000000000004e-06, 'learning_rate/action_model': 3.1e-05, 'epoch': 0.01}) +[2026-06-17 06:14:53,277][starVLA.training.train_starvla][INFO] - Step 32, Loss: {'train/loss': 1.5316030979156494, 'train/grad_norm_pre_clip': 2.5253825187683105, 'timing/data': 0.0053794150007888675, 'timing/model': 0.9740492749260738, 'learning_rate/qwen_vl_interface': 3.2000000000000003e-06, 'learning_rate/action_model': 3.2000000000000005e-05, 'epoch': 0.01}) +[2026-06-17 06:14:57,002][starVLA.training.train_starvla][INFO] - Step 33, Loss: {'train/loss': 1.4831663370132446, 'train/grad_norm_pre_clip': 1.8588489294052124, 'timing/data': 0.004601793014444411, 'timing/model': 0.9716676069656387, 'learning_rate/qwen_vl_interface': 3.3000000000000006e-06, 'learning_rate/action_model': 3.3e-05, 'epoch': 0.01}) +[2026-06-17 06:15:00,750][starVLA.training.train_starvla][INFO] - Step 34, Loss: {'train/loss': 1.3533263206481934, 'train/grad_norm_pre_clip': 1.2674179077148438, 'timing/data': 0.004981471924111247, 'timing/model': 0.9835997980553657, 'learning_rate/qwen_vl_interface': 3.4000000000000005e-06, 'learning_rate/action_model': 3.4000000000000007e-05, 'epoch': 0.01}) +[2026-06-17 06:15:04,494][starVLA.training.train_starvla][INFO] - Step 35, Loss: {'train/loss': 1.4290567636489868, 'train/grad_norm_pre_clip': 3.7061426639556885, 'timing/data': 0.004836934036575258, 'timing/model': 0.976337039959617, 'learning_rate/qwen_vl_interface': 3.5e-06, 'learning_rate/action_model': 3.5e-05, 'epoch': 0.01}) +[2026-06-17 06:15:08,231][starVLA.training.train_starvla][INFO] - Step 36, Loss: {'train/loss': 1.433957815170288, 'train/grad_norm_pre_clip': 2.8418526649475098, 'timing/data': 0.0051003380212932825, 'timing/model': 0.975951335974969, 'learning_rate/qwen_vl_interface': 3.6000000000000003e-06, 'learning_rate/action_model': 3.6e-05, 'epoch': 0.01}) +[2026-06-17 06:15:12,125][starVLA.training.train_starvla][INFO] - Step 37, Loss: {'train/loss': 1.5783942937850952, 'train/grad_norm_pre_clip': 4.8490471839904785, 'timing/data': 0.005109040066599846, 'timing/model': 1.0492632569512352, 'learning_rate/qwen_vl_interface': 3.7e-06, 'learning_rate/action_model': 3.7e-05, 'epoch': 0.01}) +[2026-06-17 06:15:15,868][starVLA.training.train_starvla][INFO] - Step 38, Loss: {'train/loss': 1.674981951713562, 'train/grad_norm_pre_clip': 7.010296821594238, 'timing/data': 0.005140664987266064, 'timing/model': 0.9761018530698493, 'learning_rate/qwen_vl_interface': 3.8000000000000005e-06, 'learning_rate/action_model': 3.8e-05, 'epoch': 0.01}) +[2026-06-17 06:15:19,616][starVLA.training.train_starvla][INFO] - Step 39, Loss: {'train/loss': 1.3917827606201172, 'train/grad_norm_pre_clip': 3.7277655601501465, 'timing/data': 0.005488545983098447, 'timing/model': 0.9757889989996329, 'learning_rate/qwen_vl_interface': 3.900000000000001e-06, 'learning_rate/action_model': 3.9000000000000006e-05, 'epoch': 0.01}) +[2026-06-17 06:15:23,364][starVLA.training.train_starvla][INFO] - Step 40, Loss: {'train/loss': 1.4765102863311768, 'train/grad_norm_pre_clip': 6.304587364196777, 'timing/data': 0.00563486001919955, 'timing/model': 0.979969767970033, 'learning_rate/qwen_vl_interface': 4.000000000000001e-06, 'learning_rate/action_model': 4e-05, 'epoch': 0.01}) +[2026-06-17 06:15:27,103][starVLA.training.train_starvla][INFO] - Step 41, Loss: {'train/loss': 1.4177196025848389, 'train/grad_norm_pre_clip': 7.486145973205566, 'timing/data': 0.0055685300612822175, 'timing/model': 0.9733758820220828, 'learning_rate/qwen_vl_interface': 4.1e-06, 'learning_rate/action_model': 4.1e-05, 'epoch': 0.01}) +[2026-06-17 06:15:30,843][starVLA.training.train_starvla][INFO] - Step 42, Loss: {'train/loss': 1.333661675453186, 'train/grad_norm_pre_clip': 4.541993141174316, 'timing/data': 0.0054521020501852036, 'timing/model': 0.9734543008962646, 'learning_rate/qwen_vl_interface': 4.2000000000000004e-06, 'learning_rate/action_model': 4.2e-05, 'epoch': 0.01}) +[2026-06-17 06:15:34,578][starVLA.training.train_starvla][INFO] - Step 43, Loss: {'train/loss': 1.4190038442611694, 'train/grad_norm_pre_clip': 5.536563396453857, 'timing/data': 0.004863496054895222, 'timing/model': 0.9753398379543796, 'learning_rate/qwen_vl_interface': 4.3e-06, 'learning_rate/action_model': 4.3e-05, 'epoch': 0.01}) +[2026-06-17 06:15:38,323][starVLA.training.train_starvla][INFO] - Step 44, Loss: {'train/loss': 1.5914994478225708, 'train/grad_norm_pre_clip': 2.5332846641540527, 'timing/data': 0.0054462680127471685, 'timing/model': 0.9769008780131117, 'learning_rate/qwen_vl_interface': 4.4e-06, 'learning_rate/action_model': 4.4000000000000006e-05, 'epoch': 0.01}) +[2026-06-17 06:15:42,066][starVLA.training.train_starvla][INFO] - Step 45, Loss: {'train/loss': 1.3105796575546265, 'train/grad_norm_pre_clip': 3.2185730934143066, 'timing/data': 0.005032506072893739, 'timing/model': 0.9745797209907323, 'learning_rate/qwen_vl_interface': 4.5e-06, 'learning_rate/action_model': 4.5e-05, 'epoch': 0.01}) +[2026-06-17 06:15:45,816][starVLA.training.train_starvla][INFO] - Step 46, Loss: {'train/loss': 1.4652459621429443, 'train/grad_norm_pre_clip': 3.6609408855438232, 'timing/data': 0.005426481016911566, 'timing/model': 0.9758149170083925, 'learning_rate/qwen_vl_interface': 4.600000000000001e-06, 'learning_rate/action_model': 4.600000000000001e-05, 'epoch': 0.01}) +[2026-06-17 06:15:49,560][starVLA.training.train_starvla][INFO] - Step 47, Loss: {'train/loss': 1.3712540864944458, 'train/grad_norm_pre_clip': 3.901329517364502, 'timing/data': 0.005388673976995051, 'timing/model': 0.9787669650977477, 'learning_rate/qwen_vl_interface': 4.7e-06, 'learning_rate/action_model': 4.7e-05, 'epoch': 0.01}) +[2026-06-17 06:15:53,308][starVLA.training.train_starvla][INFO] - Step 48, Loss: {'train/loss': 1.4338268041610718, 'train/grad_norm_pre_clip': 3.320006847381592, 'timing/data': 0.005137900006957352, 'timing/model': 0.9777292250655591, 'learning_rate/qwen_vl_interface': 4.800000000000001e-06, 'learning_rate/action_model': 4.8e-05, 'epoch': 0.01}) +[2026-06-17 06:15:57,077][starVLA.training.train_starvla][INFO] - Step 49, Loss: {'train/loss': 1.2467654943466187, 'train/grad_norm_pre_clip': 2.08606219291687, 'timing/data': 0.0050630830228328705, 'timing/model': 0.989256575005129, 'learning_rate/qwen_vl_interface': 4.9000000000000005e-06, 'learning_rate/action_model': 4.9e-05, 'epoch': 0.01}) +[2026-06-17 06:16:00,840][starVLA.training.train_starvla][INFO] - Step 50, Loss: {'train/loss': 1.26407790184021, 'train/grad_norm_pre_clip': 10.923394203186035, 'timing/data': 0.005707664997316897, 'timing/model': 0.9757998479763046, 'learning_rate/qwen_vl_interface': 5e-06, 'learning_rate/action_model': 5e-05, 'epoch': 0.01}) +[2026-06-17 06:16:04,602][starVLA.training.train_starvla][INFO] - Step 51, Loss: {'train/loss': 1.5007885694503784, 'train/grad_norm_pre_clip': 5.2761921882629395, 'timing/data': 0.005735327024012804, 'timing/model': 0.9776431070640683, 'learning_rate/qwen_vl_interface': 5.1e-06, 'learning_rate/action_model': 5.1000000000000006e-05, 'epoch': 0.02}) +[2026-06-17 06:16:08,351][starVLA.training.train_starvla][INFO] - Step 52, Loss: {'train/loss': 1.4008266925811768, 'train/grad_norm_pre_clip': 3.164130687713623, 'timing/data': 0.005075962981209159, 'timing/model': 0.9800480309640989, 'learning_rate/qwen_vl_interface': 5.2e-06, 'learning_rate/action_model': 5.2000000000000004e-05, 'epoch': 0.02}) +[2026-06-17 06:16:12,100][starVLA.training.train_starvla][INFO] - Step 53, Loss: {'train/loss': 1.343871831893921, 'train/grad_norm_pre_clip': 4.954222202301025, 'timing/data': 0.0051981579745188355, 'timing/model': 0.9770092910621315, 'learning_rate/qwen_vl_interface': 5.300000000000001e-06, 'learning_rate/action_model': 5.300000000000001e-05, 'epoch': 0.02}) +[2026-06-17 06:16:15,849][starVLA.training.train_starvla][INFO] - Step 54, Loss: {'train/loss': 1.2764333486557007, 'train/grad_norm_pre_clip': 6.4715423583984375, 'timing/data': 0.004259287961758673, 'timing/model': 0.9790722199250013, 'learning_rate/qwen_vl_interface': 5.400000000000001e-06, 'learning_rate/action_model': 5.4000000000000005e-05, 'epoch': 0.02}) +[2026-06-17 06:16:19,606][starVLA.training.train_starvla][INFO] - Step 55, Loss: {'train/loss': 1.4994372129440308, 'train/grad_norm_pre_clip': 5.335907936096191, 'timing/data': 0.005606260034255683, 'timing/model': 0.9805269809439778, 'learning_rate/qwen_vl_interface': 5.500000000000001e-06, 'learning_rate/action_model': 5.500000000000001e-05, 'epoch': 0.02}) +[2026-06-17 06:16:23,362][starVLA.training.train_starvla][INFO] - Step 56, Loss: {'train/loss': 1.4625808000564575, 'train/grad_norm_pre_clip': 8.623259544372559, 'timing/data': 0.005239955964498222, 'timing/model': 0.978401112020947, 'learning_rate/qwen_vl_interface': 5.600000000000001e-06, 'learning_rate/action_model': 5.6000000000000006e-05, 'epoch': 0.02}) +[2026-06-17 06:16:27,115][starVLA.training.train_starvla][INFO] - Step 57, Loss: {'train/loss': 1.2165786027908325, 'train/grad_norm_pre_clip': 11.109613418579102, 'timing/data': 0.004756103968247771, 'timing/model': 0.9796363620553166, 'learning_rate/qwen_vl_interface': 5.7e-06, 'learning_rate/action_model': 5.6999999999999996e-05, 'epoch': 0.02}) +[2026-06-17 06:16:30,867][starVLA.training.train_starvla][INFO] - Step 58, Loss: {'train/loss': 1.3226248025894165, 'train/grad_norm_pre_clip': 5.876341819763184, 'timing/data': 0.005662882002070546, 'timing/model': 0.9808402820490301, 'learning_rate/qwen_vl_interface': 5.8e-06, 'learning_rate/action_model': 5.8e-05, 'epoch': 0.02}) +[2026-06-17 06:16:34,620][starVLA.training.train_starvla][INFO] - Step 59, Loss: {'train/loss': 1.481703281402588, 'train/grad_norm_pre_clip': 10.593324661254883, 'timing/data': 0.006402142927981913, 'timing/model': 0.9741965159773827, 'learning_rate/qwen_vl_interface': 5.9e-06, 'learning_rate/action_model': 5.9e-05, 'epoch': 0.02}) +[2026-06-17 06:16:38,398][starVLA.training.train_starvla][INFO] - Step 60, Loss: {'train/loss': 1.4029676914215088, 'train/grad_norm_pre_clip': 7.874946117401123, 'timing/data': 0.004765805904753506, 'timing/model': 0.9741008570417762, 'learning_rate/qwen_vl_interface': 6e-06, 'learning_rate/action_model': 6e-05, 'epoch': 0.02}) +[2026-06-17 06:16:42,148][starVLA.training.train_starvla][INFO] - Step 61, Loss: {'train/loss': 1.4046286344528198, 'train/grad_norm_pre_clip': 13.892961502075195, 'timing/data': 0.005339456954970956, 'timing/model': 0.9777719110716134, 'learning_rate/qwen_vl_interface': 6.1e-06, 'learning_rate/action_model': 6.1e-05, 'epoch': 0.02}) +[2026-06-17 06:16:45,893][starVLA.training.train_starvla][INFO] - Step 62, Loss: {'train/loss': 1.4149152040481567, 'train/grad_norm_pre_clip': 8.609235763549805, 'timing/data': 0.004258010070770979, 'timing/model': 0.977638843934983, 'learning_rate/qwen_vl_interface': 6.200000000000001e-06, 'learning_rate/action_model': 6.2e-05, 'epoch': 0.02}) +[2026-06-17 06:16:49,642][starVLA.training.train_starvla][INFO] - Step 63, Loss: {'train/loss': 1.2767469882965088, 'train/grad_norm_pre_clip': 11.053550720214844, 'timing/data': 0.004503255011513829, 'timing/model': 0.9775820530485362, 'learning_rate/qwen_vl_interface': 6.300000000000001e-06, 'learning_rate/action_model': 6.3e-05, 'epoch': 0.02}) +[2026-06-17 06:16:53,489][starVLA.training.train_starvla][INFO] - Step 64, Loss: {'train/loss': 1.2747294902801514, 'train/grad_norm_pre_clip': 7.201638698577881, 'timing/data': 0.004360195016488433, 'timing/model': 0.9875416320282966, 'learning_rate/qwen_vl_interface': 6.4000000000000006e-06, 'learning_rate/action_model': 6.400000000000001e-05, 'epoch': 0.02}) +[2026-06-17 06:16:57,248][starVLA.training.train_starvla][INFO] - Step 65, Loss: {'train/loss': 1.339299201965332, 'train/grad_norm_pre_clip': 6.034546375274658, 'timing/data': 0.00560917507391423, 'timing/model': 0.9810048670042306, 'learning_rate/qwen_vl_interface': 6.5000000000000004e-06, 'learning_rate/action_model': 6.500000000000001e-05, 'epoch': 0.02}) +[2026-06-17 06:17:01,077][starVLA.training.train_starvla][INFO] - Step 66, Loss: {'train/loss': 1.304283857345581, 'train/grad_norm_pre_clip': 6.43575382232666, 'timing/data': 0.005783434025943279, 'timing/model': 0.9785978969885036, 'learning_rate/qwen_vl_interface': 6.600000000000001e-06, 'learning_rate/action_model': 6.6e-05, 'epoch': 0.02}) +[2026-06-17 06:17:04,831][starVLA.training.train_starvla][INFO] - Step 67, Loss: {'train/loss': 1.368280053138733, 'train/grad_norm_pre_clip': 7.018569469451904, 'timing/data': 0.005617625894956291, 'timing/model': 0.9784243749454618, 'learning_rate/qwen_vl_interface': 6.700000000000001e-06, 'learning_rate/action_model': 6.7e-05, 'epoch': 0.02}) +[2026-06-17 06:17:08,578][starVLA.training.train_starvla][INFO] - Step 68, Loss: {'train/loss': 1.2818366289138794, 'train/grad_norm_pre_clip': 9.371101379394531, 'timing/data': 0.005820440943352878, 'timing/model': 0.9753886730177328, 'learning_rate/qwen_vl_interface': 6.800000000000001e-06, 'learning_rate/action_model': 6.800000000000001e-05, 'epoch': 0.02}) +[2026-06-17 06:17:12,324][starVLA.training.train_starvla][INFO] - Step 69, Loss: {'train/loss': 1.3793678283691406, 'train/grad_norm_pre_clip': 9.502737045288086, 'timing/data': 0.004988610977306962, 'timing/model': 0.9765085619874299, 'learning_rate/qwen_vl_interface': 6.9e-06, 'learning_rate/action_model': 6.9e-05, 'epoch': 0.02}) +[2026-06-17 06:17:16,073][starVLA.training.train_starvla][INFO] - Step 70, Loss: {'train/loss': 1.287865400314331, 'train/grad_norm_pre_clip': 12.41684341430664, 'timing/data': 0.004591312957927585, 'timing/model': 0.9765162379480898, 'learning_rate/qwen_vl_interface': 7e-06, 'learning_rate/action_model': 7e-05, 'epoch': 0.02}) +[2026-06-17 06:17:19,822][starVLA.training.train_starvla][INFO] - Step 71, Loss: {'train/loss': 1.5090115070343018, 'train/grad_norm_pre_clip': 12.699021339416504, 'timing/data': 0.004362927051261067, 'timing/model': 0.9765993830515072, 'learning_rate/qwen_vl_interface': 7.100000000000001e-06, 'learning_rate/action_model': 7.1e-05, 'epoch': 0.02}) +[2026-06-17 06:17:23,568][starVLA.training.train_starvla][INFO] - Step 72, Loss: {'train/loss': 1.4082601070404053, 'train/grad_norm_pre_clip': 16.95681381225586, 'timing/data': 0.006199681083671749, 'timing/model': 0.9770736600039527, 'learning_rate/qwen_vl_interface': 7.2000000000000005e-06, 'learning_rate/action_model': 7.2e-05, 'epoch': 0.02}) +[2026-06-17 06:17:27,318][starVLA.training.train_starvla][INFO] - Step 73, Loss: {'train/loss': 1.473010778427124, 'train/grad_norm_pre_clip': 6.997933864593506, 'timing/data': 0.0049341010162606835, 'timing/model': 0.9776476949919015, 'learning_rate/qwen_vl_interface': 7.3e-06, 'learning_rate/action_model': 7.3e-05, 'epoch': 0.02}) +[2026-06-17 06:17:31,057][starVLA.training.train_starvla][INFO] - Step 74, Loss: {'train/loss': 1.3547903299331665, 'train/grad_norm_pre_clip': 9.715606689453125, 'timing/data': 0.005582031095400453, 'timing/model': 0.9745155489072204, 'learning_rate/qwen_vl_interface': 7.4e-06, 'learning_rate/action_model': 7.4e-05, 'epoch': 0.02}) +[2026-06-17 06:17:34,810][starVLA.training.train_starvla][INFO] - Step 75, Loss: {'train/loss': 1.5350944995880127, 'train/grad_norm_pre_clip': 8.071575164794922, 'timing/data': 0.005106341908685863, 'timing/model': 0.9766096579842269, 'learning_rate/qwen_vl_interface': 7.500000000000001e-06, 'learning_rate/action_model': 7.500000000000001e-05, 'epoch': 0.02}) +[2026-06-17 06:17:38,561][starVLA.training.train_starvla][INFO] - Step 76, Loss: {'train/loss': 1.496644377708435, 'train/grad_norm_pre_clip': 8.68978500366211, 'timing/data': 0.005744264926761389, 'timing/model': 0.9763097469694912, 'learning_rate/qwen_vl_interface': 7.600000000000001e-06, 'learning_rate/action_model': 7.6e-05, 'epoch': 0.02}) +[2026-06-17 06:17:42,327][starVLA.training.train_starvla][INFO] - Step 77, Loss: {'train/loss': 1.2714087963104248, 'train/grad_norm_pre_clip': 9.13891315460205, 'timing/data': 0.004472868982702494, 'timing/model': 0.9752813170198351, 'learning_rate/qwen_vl_interface': 7.7e-06, 'learning_rate/action_model': 7.7e-05, 'epoch': 0.02}) +[2026-06-17 06:17:46,077][starVLA.training.train_starvla][INFO] - Step 78, Loss: {'train/loss': 1.223048210144043, 'train/grad_norm_pre_clip': 6.13691520690918, 'timing/data': 0.005569202941842377, 'timing/model': 0.9800896419910714, 'learning_rate/qwen_vl_interface': 7.800000000000002e-06, 'learning_rate/action_model': 7.800000000000001e-05, 'epoch': 0.02}) +[2026-06-17 06:17:49,824][starVLA.training.train_starvla][INFO] - Step 79, Loss: {'train/loss': 1.6319559812545776, 'train/grad_norm_pre_clip': 11.398066520690918, 'timing/data': 0.005584014928899705, 'timing/model': 0.9771679129917175, 'learning_rate/qwen_vl_interface': 7.9e-06, 'learning_rate/action_model': 7.900000000000001e-05, 'epoch': 0.02}) +[2026-06-17 06:17:53,568][starVLA.training.train_starvla][INFO] - Step 80, Loss: {'train/loss': 1.4027694463729858, 'train/grad_norm_pre_clip': 9.501684188842773, 'timing/data': 0.005226838984526694, 'timing/model': 0.9747335819993168, 'learning_rate/qwen_vl_interface': 8.000000000000001e-06, 'learning_rate/action_model': 8e-05, 'epoch': 0.02}) +[2026-06-17 06:17:57,317][starVLA.training.train_starvla][INFO] - Step 81, Loss: {'train/loss': 1.5607389211654663, 'train/grad_norm_pre_clip': 7.518157958984375, 'timing/data': 0.004661332000978291, 'timing/model': 0.9779993259580806, 'learning_rate/qwen_vl_interface': 8.1e-06, 'learning_rate/action_model': 8.1e-05, 'epoch': 0.02}) +[2026-06-17 06:18:01,063][starVLA.training.train_starvla][INFO] - Step 82, Loss: {'train/loss': 1.4496064186096191, 'train/grad_norm_pre_clip': 6.91839075088501, 'timing/data': 0.00491757201962173, 'timing/model': 0.9759562920080498, 'learning_rate/qwen_vl_interface': 8.2e-06, 'learning_rate/action_model': 8.2e-05, 'epoch': 0.02}) +[2026-06-17 06:18:04,818][starVLA.training.train_starvla][INFO] - Step 83, Loss: {'train/loss': 1.3952723741531372, 'train/grad_norm_pre_clip': 5.358388900756836, 'timing/data': 0.005251231021247804, 'timing/model': 0.9781974289799109, 'learning_rate/qwen_vl_interface': 8.3e-06, 'learning_rate/action_model': 8.3e-05, 'epoch': 0.02}) +[2026-06-17 06:18:08,576][starVLA.training.train_starvla][INFO] - Step 84, Loss: {'train/loss': 1.19498872756958, 'train/grad_norm_pre_clip': 6.7746171951293945, 'timing/data': 0.005964349955320358, 'timing/model': 0.9769796840846539, 'learning_rate/qwen_vl_interface': 8.400000000000001e-06, 'learning_rate/action_model': 8.4e-05, 'epoch': 0.02}) +[2026-06-17 06:18:12,334][starVLA.training.train_starvla][INFO] - Step 85, Loss: {'train/loss': 1.4879968166351318, 'train/grad_norm_pre_clip': 6.152397155761719, 'timing/data': 0.004258051980286837, 'timing/model': 0.9805968549335375, 'learning_rate/qwen_vl_interface': 8.5e-06, 'learning_rate/action_model': 8.5e-05, 'epoch': 0.03}) +[2026-06-17 06:18:16,085][starVLA.training.train_starvla][INFO] - Step 86, Loss: {'train/loss': 1.3703163862228394, 'train/grad_norm_pre_clip': 4.716137409210205, 'timing/data': 0.0057163049932569265, 'timing/model': 0.976534710964188, 'learning_rate/qwen_vl_interface': 8.6e-06, 'learning_rate/action_model': 8.6e-05, 'epoch': 0.03}) +[2026-06-17 06:18:19,838][starVLA.training.train_starvla][INFO] - Step 87, Loss: {'train/loss': 1.3544669151306152, 'train/grad_norm_pre_clip': 5.6734185218811035, 'timing/data': 0.0057399459183216095, 'timing/model': 0.9787479980150238, 'learning_rate/qwen_vl_interface': 8.700000000000001e-06, 'learning_rate/action_model': 8.7e-05, 'epoch': 0.03}) +[2026-06-17 06:18:23,585][starVLA.training.train_starvla][INFO] - Step 88, Loss: {'train/loss': 1.6013495922088623, 'train/grad_norm_pre_clip': 5.76531982421875, 'timing/data': 0.0046194769674912095, 'timing/model': 0.9771535630570725, 'learning_rate/qwen_vl_interface': 8.8e-06, 'learning_rate/action_model': 8.800000000000001e-05, 'epoch': 0.03}) +[2026-06-17 06:18:27,343][starVLA.training.train_starvla][INFO] - Step 89, Loss: {'train/loss': 1.3080775737762451, 'train/grad_norm_pre_clip': 5.472917556762695, 'timing/data': 0.0055856689577922225, 'timing/model': 0.9772552220383659, 'learning_rate/qwen_vl_interface': 8.900000000000001e-06, 'learning_rate/action_model': 8.900000000000001e-05, 'epoch': 0.03}) +[2026-06-17 06:18:31,191][starVLA.training.train_starvla][INFO] - Step 90, Loss: {'train/loss': 1.2990316152572632, 'train/grad_norm_pre_clip': 10.854802131652832, 'timing/data': 0.005562194972299039, 'timing/model': 1.065348420990631, 'learning_rate/qwen_vl_interface': 9e-06, 'learning_rate/action_model': 9e-05, 'epoch': 0.03}) +[2026-06-17 06:18:34,945][starVLA.training.train_starvla][INFO] - Step 91, Loss: {'train/loss': 1.3402059078216553, 'train/grad_norm_pre_clip': 6.76000452041626, 'timing/data': 0.005617016926407814, 'timing/model': 0.9762589229503646, 'learning_rate/qwen_vl_interface': 9.100000000000001e-06, 'learning_rate/action_model': 9.1e-05, 'epoch': 0.03}) +[2026-06-17 06:18:38,770][starVLA.training.train_starvla][INFO] - Step 92, Loss: {'train/loss': 1.3521465063095093, 'train/grad_norm_pre_clip': 6.282928466796875, 'timing/data': 0.004525894997641444, 'timing/model': 0.9764780380064622, 'learning_rate/qwen_vl_interface': 9.200000000000002e-06, 'learning_rate/action_model': 9.200000000000001e-05, 'epoch': 0.03}) +[2026-06-17 06:18:42,520][starVLA.training.train_starvla][INFO] - Step 93, Loss: {'train/loss': 1.4760782718658447, 'train/grad_norm_pre_clip': 8.646024703979492, 'timing/data': 0.004463350051082671, 'timing/model': 0.9782755989581347, 'learning_rate/qwen_vl_interface': 9.3e-06, 'learning_rate/action_model': 9.300000000000001e-05, 'epoch': 0.03}) +[2026-06-17 06:18:46,278][starVLA.training.train_starvla][INFO] - Step 94, Loss: {'train/loss': 1.3151177167892456, 'train/grad_norm_pre_clip': 3.7607696056365967, 'timing/data': 0.006350503070279956, 'timing/model': 0.9779969080118462, 'learning_rate/qwen_vl_interface': 9.4e-06, 'learning_rate/action_model': 9.4e-05, 'epoch': 0.03}) +[2026-06-17 06:18:50,027][starVLA.training.train_starvla][INFO] - Step 95, Loss: {'train/loss': 1.32670259475708, 'train/grad_norm_pre_clip': 7.092619895935059, 'timing/data': 0.005665709963068366, 'timing/model': 0.9775603289017454, 'learning_rate/qwen_vl_interface': 9.5e-06, 'learning_rate/action_model': 9.5e-05, 'epoch': 0.03}) +[2026-06-17 06:18:53,861][starVLA.training.train_starvla][INFO] - Step 96, Loss: {'train/loss': 1.406962275505066, 'train/grad_norm_pre_clip': 7.156660556793213, 'timing/data': 0.006237584981136024, 'timing/model': 0.9783395129488781, 'learning_rate/qwen_vl_interface': 9.600000000000001e-06, 'learning_rate/action_model': 9.6e-05, 'epoch': 0.03}) +[2026-06-17 06:18:57,615][starVLA.training.train_starvla][INFO] - Step 97, Loss: {'train/loss': 1.365391492843628, 'train/grad_norm_pre_clip': 5.680599212646484, 'timing/data': 0.004367970046587288, 'timing/model': 0.9821377170737833, 'learning_rate/qwen_vl_interface': 9.7e-06, 'learning_rate/action_model': 9.7e-05, 'epoch': 0.03}) +[2026-06-17 06:19:01,372][starVLA.training.train_starvla][INFO] - Step 98, Loss: {'train/loss': 1.4983429908752441, 'train/grad_norm_pre_clip': 17.121522903442383, 'timing/data': 0.005561805912293494, 'timing/model': 0.9782089999644086, 'learning_rate/qwen_vl_interface': 9.800000000000001e-06, 'learning_rate/action_model': 9.8e-05, 'epoch': 0.03}) +[2026-06-17 06:19:05,125][starVLA.training.train_starvla][INFO] - Step 99, Loss: {'train/loss': 1.2950125932693481, 'train/grad_norm_pre_clip': 7.64937162399292, 'timing/data': 0.005390056991018355, 'timing/model': 0.9802327509969473, 'learning_rate/qwen_vl_interface': 9.9e-06, 'learning_rate/action_model': 9.900000000000001e-05, 'epoch': 0.03}) +[2026-06-17 06:19:51,398][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_openvla/include -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_openvla/include -fPIC -c /tmp/tmpccqhukb8/test.c -o /tmp/tmpccqhukb8/test.o +[2026-06-17 06:19:51,408][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat /tmp/tmpccqhukb8/test.o -laio -o /tmp/tmpccqhukb8/a.out +[2026-06-17 06:19:51,635][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_openvla/include -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_openvla/include -fPIC -c /tmp/tmpwen85gyq/test.c -o /tmp/tmpwen85gyq/test.o +[2026-06-17 06:19:51,646][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat /tmp/tmpwen85gyq/test.o -L/usr/local/cuda -L/usr/local/cuda/lib64 -lcufile -o /tmp/tmpwen85gyq/a.out +[2026-06-17 06:19:51,669][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_openvla/include -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_openvla/include -fPIC -c /tmp/tmphh2pk4nt/test.c -o /tmp/tmphh2pk4nt/test.o +[2026-06-17 06:19:51,678][root][INFO] - gcc -pthread -B /venv/starvla_openvla/compiler_compat /tmp/tmphh2pk4nt/test.o -laio -o /tmp/tmphh2pk4nt/a.out +[2026-06-17 06:20:56,578][starVLA.training.train_starvla][INFO] - Step 100, Loss: {'train/loss': 1.332024097442627, 'train/grad_norm_pre_clip': 4.935667991638184, 'eval/loss': 1.3992472887039185, 'eval/latency_0/loss': 1.3992472887039185, 'eval/demon_fire_precision': 0.8646895273401297, 'eval/demon_fire_recall': 0.7749169435215947, 'eval/demon_fire_f1': 0.8173455978975032, 'eval/demon_fire_rate_model': 0.600111234705228, 'eval/demon_fire_rate_teacher': 0.6696329254727474, 'eval/demon_left_precision': 0.5959183673469388, 'eval/demon_left_recall': 0.7444759206798867, 'eval/demon_left_f1': 0.6619647355163729, 'eval/demon_left_rate_model': 0.6131813125695217, 'eval/demon_left_rate_teacher': 0.49082313681868744, 'eval/demon_right_precision': 0.6290438533429188, 'eval/demon_right_recall': 0.4943502824858757, 'eval/demon_right_f1': 0.5536222714330908, 'eval/demon_right_rate_model': 0.3868186874304783, 'eval/demon_right_rate_teacher': 0.4922135706340378, 'eval/demon_move_f1': 0.6077935034747319, 'eval/demon_cc_f1': 0.7407300086428421, 'eval/latency_0/demon_fire_precision': 0.8646895273401297, 'eval/latency_0/demon_fire_recall': 0.7749169435215947, 'eval/latency_0/demon_fire_f1': 0.8173455978975032, 'eval/latency_0/demon_fire_rate_model': 0.600111234705228, 'eval/latency_0/demon_fire_rate_teacher': 0.6696329254727474, 'eval/latency_0/demon_left_precision': 0.5959183673469388, 'eval/latency_0/demon_left_recall': 0.7444759206798867, 'eval/latency_0/demon_left_f1': 0.6619647355163729, 'eval/latency_0/demon_left_rate_model': 0.6131813125695217, 'eval/latency_0/demon_left_rate_teacher': 0.49082313681868744, 'eval/latency_0/demon_right_precision': 0.6290438533429188, 'eval/latency_0/demon_right_recall': 0.4943502824858757, 'eval/latency_0/demon_right_f1': 0.5536222714330908, 'eval/latency_0/demon_right_rate_model': 0.3868186874304783, 'eval/latency_0/demon_right_rate_teacher': 0.4922135706340378, 'eval/latency_0/demon_move_f1': 0.6077935034747319, 'eval/latency_0/demon_cc_f1': 0.7407300086428421, 'timing/data': 0.005674590938724577, 'timing/model': 0.9754386510467157, 'learning_rate/qwen_vl_interface': 1e-05, 'learning_rate/action_model': 0.0001, 'epoch': 0.03}) +[2026-06-17 06:20:56,579][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_100_state +[2026-06-17 06:20:56,582][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 06:21:00,011][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_100_state/model.safetensors +[2026-06-17 06:21:07,578][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_100_state/optimizer.bin +[2026-06-17 06:21:07,578][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_100_state/sampler.bin +[2026-06-17 06:21:07,579][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_100_state/sampler_1.bin +[2026-06-17 06:21:07,579][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_100_state/random_states_0.pkl +[2026-06-17 06:21:07,581][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 100 +[2026-06-17 06:21:07,581][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 06:21:07,597][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 06:21:12,314][starVLA.training.train_starvla][INFO] - Step 101, Loss: {'train/loss': 1.3541146516799927, 'train/grad_norm_pre_clip': 8.463521957397461, 'timing/data': 0.004845109069719911, 'timing/model': 1.1915535220177844, 'learning_rate/qwen_vl_interface': 9.999999507659945e-06, 'learning_rate/action_model': 9.999999507659945e-05, 'epoch': 0.03}) +[2026-06-17 06:21:17,030][starVLA.training.train_starvla][INFO] - Step 102, Loss: {'train/loss': 1.3368749618530273, 'train/grad_norm_pre_clip': 8.105624198913574, 'timing/data': 0.0048247650265693665, 'timing/model': 1.1888889280380681, 'learning_rate/qwen_vl_interface': 9.99999803063988e-06, 'learning_rate/action_model': 9.99999803063988e-05, 'epoch': 0.03}) +[2026-06-17 06:21:21,765][starVLA.training.train_starvla][INFO] - Step 103, Loss: {'train/loss': 1.2432210445404053, 'train/grad_norm_pre_clip': 6.248234748840332, 'timing/data': 0.004861710011027753, 'timing/model': 1.318585149012506, 'learning_rate/qwen_vl_interface': 9.999995568940113e-06, 'learning_rate/action_model': 9.999995568940113e-05, 'epoch': 0.03}) +[2026-06-17 06:21:26,388][starVLA.training.train_starvla][INFO] - Step 104, Loss: {'train/loss': 1.382083535194397, 'train/grad_norm_pre_clip': 9.63544750213623, 'timing/data': 0.0054071389604359865, 'timing/model': 1.206409239093773, 'learning_rate/qwen_vl_interface': 9.999992122561152e-06, 'learning_rate/action_model': 9.999992122561152e-05, 'epoch': 0.03}) +[2026-06-17 06:21:31,044][starVLA.training.train_starvla][INFO] - Step 105, Loss: {'train/loss': 1.3417365550994873, 'train/grad_norm_pre_clip': 5.093509197235107, 'timing/data': 0.004420747049152851, 'timing/model': 1.1948955019470304, 'learning_rate/qwen_vl_interface': 9.999987691503714e-06, 'learning_rate/action_model': 9.999987691503714e-05, 'epoch': 0.03}) +[2026-06-17 06:21:35,806][starVLA.training.train_starvla][INFO] - Step 106, Loss: {'train/loss': 1.4651026725769043, 'train/grad_norm_pre_clip': 6.490249156951904, 'timing/data': 0.004693259950727224, 'timing/model': 1.1958086399827152, 'learning_rate/qwen_vl_interface': 9.999982275768716e-06, 'learning_rate/action_model': 9.999982275768715e-05, 'epoch': 0.03}) +[2026-06-17 06:21:40,559][starVLA.training.train_starvla][INFO] - Step 107, Loss: {'train/loss': 1.5037109851837158, 'train/grad_norm_pre_clip': 8.757758140563965, 'timing/data': 0.004533033934421837, 'timing/model': 1.1975163009483367, 'learning_rate/qwen_vl_interface': 9.99997587535728e-06, 'learning_rate/action_model': 9.99997587535728e-05, 'epoch': 0.03}) +[2026-06-17 06:21:45,335][starVLA.training.train_starvla][INFO] - Step 108, Loss: {'train/loss': 1.4330861568450928, 'train/grad_norm_pre_clip': 3.426468849182129, 'timing/data': 0.004536170046776533, 'timing/model': 1.2013934259302914, 'learning_rate/qwen_vl_interface': 9.999968490270735e-06, 'learning_rate/action_model': 9.999968490270735e-05, 'epoch': 0.03}) +[2026-06-17 06:21:50,093][starVLA.training.train_starvla][INFO] - Step 109, Loss: {'train/loss': 1.3532884120941162, 'train/grad_norm_pre_clip': 6.446109771728516, 'timing/data': 0.00461374397855252, 'timing/model': 1.3226093919947743, 'learning_rate/qwen_vl_interface': 9.99996012051061e-06, 'learning_rate/action_model': 9.99996012051061e-05, 'epoch': 0.03}) +[2026-06-17 06:21:54,756][starVLA.training.train_starvla][INFO] - Step 110, Loss: {'train/loss': 1.3382599353790283, 'train/grad_norm_pre_clip': 4.212738037109375, 'timing/data': 0.0047901199432089925, 'timing/model': 1.2207008550176397, 'learning_rate/qwen_vl_interface': 9.999950766078641e-06, 'learning_rate/action_model': 9.99995076607864e-05, 'epoch': 0.03}) +[2026-06-17 06:21:59,394][starVLA.training.train_starvla][INFO] - Step 111, Loss: {'train/loss': 1.2391225099563599, 'train/grad_norm_pre_clip': 6.418701171875, 'timing/data': 0.0048386609414592385, 'timing/model': 1.2006263489602134, 'learning_rate/qwen_vl_interface': 9.999940426976768e-06, 'learning_rate/action_model': 9.999940426976768e-05, 'epoch': 0.03}) +[2026-06-17 06:22:04,091][starVLA.training.train_starvla][INFO] - Step 112, Loss: {'train/loss': 1.3597811460494995, 'train/grad_norm_pre_clip': 5.3917107582092285, 'timing/data': 0.004667542991228402, 'timing/model': 1.2018400070955977, 'learning_rate/qwen_vl_interface': 9.999929103207133e-06, 'learning_rate/action_model': 9.999929103207132e-05, 'epoch': 0.03}) +[2026-06-17 06:22:08,816][starVLA.training.train_starvla][INFO] - Step 113, Loss: {'train/loss': 1.6541755199432373, 'train/grad_norm_pre_clip': 7.667626857757568, 'timing/data': 0.004976431024260819, 'timing/model': 1.1990647669881582, 'learning_rate/qwen_vl_interface': 9.999916794772083e-06, 'learning_rate/action_model': 9.999916794772082e-05, 'epoch': 0.03}) +[2026-06-17 06:22:13,586][starVLA.training.train_starvla][INFO] - Step 114, Loss: {'train/loss': 1.3477152585983276, 'train/grad_norm_pre_clip': 6.429100036621094, 'timing/data': 0.004340491956099868, 'timing/model': 1.202216780046001, 'learning_rate/qwen_vl_interface': 9.99990350167417e-06, 'learning_rate/action_model': 9.999903501674169e-05, 'epoch': 0.03}) +[2026-06-17 06:22:18,377][starVLA.training.train_starvla][INFO] - Step 115, Loss: {'train/loss': 1.4346637725830078, 'train/grad_norm_pre_clip': 6.709927558898926, 'timing/data': 0.004527514916844666, 'timing/model': 1.3397592970868573, 'learning_rate/qwen_vl_interface': 9.99988922391615e-06, 'learning_rate/action_model': 9.99988922391615e-05, 'epoch': 0.03}) +[2026-06-17 06:22:23,027][starVLA.training.train_starvla][INFO] - Step 116, Loss: {'train/loss': 1.4291538000106812, 'train/grad_norm_pre_clip': 7.311657428741455, 'timing/data': 0.0046626350376755, 'timing/model': 1.2025293650804088, 'learning_rate/qwen_vl_interface': 9.999873961500985e-06, 'learning_rate/action_model': 9.999873961500984e-05, 'epoch': 0.03}) +[2026-06-17 06:22:27,796][starVLA.training.train_starvla][INFO] - Step 117, Loss: {'train/loss': 1.3962490558624268, 'train/grad_norm_pre_clip': 5.2152628898620605, 'timing/data': 0.004898392013274133, 'timing/model': 1.2008165960432962, 'learning_rate/qwen_vl_interface': 9.999857714431835e-06, 'learning_rate/action_model': 9.999857714431834e-05, 'epoch': 0.03}) +[2026-06-17 06:22:32,728][starVLA.training.train_starvla][INFO] - Step 118, Loss: {'train/loss': 1.2035906314849854, 'train/grad_norm_pre_clip': 5.126040458679199, 'timing/data': 0.004521789029240608, 'timing/model': 1.20423460600432, 'learning_rate/qwen_vl_interface': 9.999840482712069e-06, 'learning_rate/action_model': 9.999840482712068e-05, 'epoch': 0.03}) +[2026-06-17 06:22:37,479][starVLA.training.train_starvla][INFO] - Step 119, Loss: {'train/loss': 1.5158121585845947, 'train/grad_norm_pre_clip': 5.641120433807373, 'timing/data': 0.004930260009132326, 'timing/model': 1.2041032520355657, 'learning_rate/qwen_vl_interface': 9.99982226634526e-06, 'learning_rate/action_model': 9.99982226634526e-05, 'epoch': 0.04}) +[2026-06-17 06:22:42,337][starVLA.training.train_starvla][INFO] - Step 120, Loss: {'train/loss': 1.5541775226593018, 'train/grad_norm_pre_clip': 9.410624504089355, 'timing/data': 0.004814587067812681, 'timing/model': 1.2025510149542242, 'learning_rate/qwen_vl_interface': 9.999803065335185e-06, 'learning_rate/action_model': 9.999803065335185e-05, 'epoch': 0.04}) +[2026-06-17 06:22:47,105][starVLA.training.train_starvla][INFO] - Step 121, Loss: {'train/loss': 1.402390956878662, 'train/grad_norm_pre_clip': 9.306436538696289, 'timing/data': 0.004931229981593788, 'timing/model': 1.327959498972632, 'learning_rate/qwen_vl_interface': 9.999782879685825e-06, 'learning_rate/action_model': 9.999782879685824e-05, 'epoch': 0.04}) +[2026-06-17 06:22:51,786][starVLA.training.train_starvla][INFO] - Step 122, Loss: {'train/loss': 1.4260289669036865, 'train/grad_norm_pre_clip': 7.7514519691467285, 'timing/data': 0.0047837699530646205, 'timing/model': 1.2027447649743408, 'learning_rate/qwen_vl_interface': 9.99976170940136e-06, 'learning_rate/action_model': 9.99976170940136e-05, 'epoch': 0.04}) +[2026-06-17 06:22:56,559][starVLA.training.train_starvla][INFO] - Step 123, Loss: {'train/loss': 1.5123651027679443, 'train/grad_norm_pre_clip': 7.223578929901123, 'timing/data': 0.004945003893226385, 'timing/model': 1.2034415830858052, 'learning_rate/qwen_vl_interface': 9.999739554486184e-06, 'learning_rate/action_model': 9.999739554486182e-05, 'epoch': 0.04}) +[2026-06-17 06:23:01,300][starVLA.training.train_starvla][INFO] - Step 124, Loss: {'train/loss': 1.411625623703003, 'train/grad_norm_pre_clip': 6.772844314575195, 'timing/data': 0.004828664939850569, 'timing/model': 1.2056361919967458, 'learning_rate/qwen_vl_interface': 9.999716414944887e-06, 'learning_rate/action_model': 9.999716414944886e-05, 'epoch': 0.04}) +[2026-06-17 06:23:06,088][starVLA.training.train_starvla][INFO] - Step 125, Loss: {'train/loss': 1.264580488204956, 'train/grad_norm_pre_clip': 5.618868350982666, 'timing/data': 0.005000287084840238, 'timing/model': 1.205279992078431, 'learning_rate/qwen_vl_interface': 9.999692290782264e-06, 'learning_rate/action_model': 9.999692290782263e-05, 'epoch': 0.04}) +[2026-06-17 06:23:10,872][starVLA.training.train_starvla][INFO] - Step 126, Loss: {'train/loss': 1.4146243333816528, 'train/grad_norm_pre_clip': 6.540126323699951, 'timing/data': 0.005001188023015857, 'timing/model': 1.2073892150074244, 'learning_rate/qwen_vl_interface': 9.999667182003321e-06, 'learning_rate/action_model': 9.999667182003321e-05, 'epoch': 0.04}) +[2026-06-17 06:23:15,664][starVLA.training.train_starvla][INFO] - Step 127, Loss: {'train/loss': 1.2650779485702515, 'train/grad_norm_pre_clip': 4.895943641662598, 'timing/data': 0.004900743952021003, 'timing/model': 1.3275241090450436, 'learning_rate/qwen_vl_interface': 9.99964108861326e-06, 'learning_rate/action_model': 9.99964108861326e-05, 'epoch': 0.04}) +[2026-06-17 06:23:20,361][starVLA.training.train_starvla][INFO] - Step 128, Loss: {'train/loss': 1.1410828828811646, 'train/grad_norm_pre_clip': 9.318401336669922, 'timing/data': 0.005287728970870376, 'timing/model': 1.250130004948005, 'learning_rate/qwen_vl_interface': 9.99961401061749e-06, 'learning_rate/action_model': 9.99961401061749e-05, 'epoch': 0.04}) +[2026-06-17 06:23:25,010][starVLA.training.train_starvla][INFO] - Step 129, Loss: {'train/loss': 1.2677786350250244, 'train/grad_norm_pre_clip': 8.13568115234375, 'timing/data': 0.004599613952450454, 'timing/model': 1.2036188310012221, 'learning_rate/qwen_vl_interface': 9.999585948021623e-06, 'learning_rate/action_model': 9.999585948021624e-05, 'epoch': 0.04}) +[2026-06-17 06:23:29,749][starVLA.training.train_starvla][INFO] - Step 130, Loss: {'train/loss': 1.4095969200134277, 'train/grad_norm_pre_clip': 8.767226219177246, 'timing/data': 0.005022738012485206, 'timing/model': 1.2053964440710843, 'learning_rate/qwen_vl_interface': 9.99955690083148e-06, 'learning_rate/action_model': 9.99955690083148e-05, 'epoch': 0.04}) +[2026-06-17 06:23:34,533][starVLA.training.train_starvla][INFO] - Step 131, Loss: {'train/loss': 1.2812286615371704, 'train/grad_norm_pre_clip': 5.584260940551758, 'timing/data': 0.004812377039343119, 'timing/model': 1.2091965459985659, 'learning_rate/qwen_vl_interface': 9.99952686905308e-06, 'learning_rate/action_model': 9.999526869053081e-05, 'epoch': 0.04}) +[2026-06-17 06:23:39,324][starVLA.training.train_starvla][INFO] - Step 132, Loss: {'train/loss': 1.4475153684616089, 'train/grad_norm_pre_clip': 12.904044151306152, 'timing/data': 0.0048376600025221705, 'timing/model': 1.20660927798599, 'learning_rate/qwen_vl_interface': 9.99949585269265e-06, 'learning_rate/action_model': 9.99949585269265e-05, 'epoch': 0.04}) +[2026-06-17 06:23:44,126][starVLA.training.train_starvla][INFO] - Step 133, Loss: {'train/loss': 1.2617777585983276, 'train/grad_norm_pre_clip': 10.085713386535645, 'timing/data': 0.004627960966899991, 'timing/model': 1.3412855480564758, 'learning_rate/qwen_vl_interface': 9.99946385175662e-06, 'learning_rate/action_model': 9.999463851756619e-05, 'epoch': 0.04}) +[2026-06-17 06:23:48,905][starVLA.training.train_starvla][INFO] - Step 134, Loss: {'train/loss': 1.565779447555542, 'train/grad_norm_pre_clip': 16.95638656616211, 'timing/data': 0.005293909925967455, 'timing/model': 1.323252068948932, 'learning_rate/qwen_vl_interface': 9.999430866251623e-06, 'learning_rate/action_model': 9.999430866251622e-05, 'epoch': 0.04}) +[2026-06-17 06:23:53,562][starVLA.training.train_starvla][INFO] - Step 135, Loss: {'train/loss': 1.3132495880126953, 'train/grad_norm_pre_clip': 15.186938285827637, 'timing/data': 0.004870400996878743, 'timing/model': 1.2060774000128731, 'learning_rate/qwen_vl_interface': 9.999396896184494e-06, 'learning_rate/action_model': 9.999396896184494e-05, 'epoch': 0.04}) +[2026-06-17 06:23:58,345][starVLA.training.train_starvla][INFO] - Step 136, Loss: {'train/loss': 1.3959710597991943, 'train/grad_norm_pre_clip': 7.276411056518555, 'timing/data': 0.004786380915902555, 'timing/model': 1.2056487560039386, 'learning_rate/qwen_vl_interface': 9.99936194156228e-06, 'learning_rate/action_model': 9.99936194156228e-05, 'epoch': 0.04}) +[2026-06-17 06:24:03,149][starVLA.training.train_starvla][INFO] - Step 137, Loss: {'train/loss': 1.3131097555160522, 'train/grad_norm_pre_clip': 6.283360481262207, 'timing/data': 0.005132612073794007, 'timing/model': 1.206063924008049, 'learning_rate/qwen_vl_interface': 9.999326002392225e-06, 'learning_rate/action_model': 9.999326002392224e-05, 'epoch': 0.04}) +[2026-06-17 06:24:07,938][starVLA.training.train_starvla][INFO] - Step 138, Loss: {'train/loss': 1.3057043552398682, 'train/grad_norm_pre_clip': 8.519429206848145, 'timing/data': 0.004857464926317334, 'timing/model': 1.2075759510044008, 'learning_rate/qwen_vl_interface': 9.999289078681779e-06, 'learning_rate/action_model': 9.999289078681778e-05, 'epoch': 0.04}) +[2026-06-17 06:24:12,721][starVLA.training.train_starvla][INFO] - Step 139, Loss: {'train/loss': 1.2457448244094849, 'train/grad_norm_pre_clip': 5.1638970375061035, 'timing/data': 0.004498107940889895, 'timing/model': 1.332798933959566, 'learning_rate/qwen_vl_interface': 9.999251170438594e-06, 'learning_rate/action_model': 9.999251170438594e-05, 'epoch': 0.04}) +[2026-06-17 06:24:17,408][starVLA.training.train_starvla][INFO] - Step 140, Loss: {'train/loss': 1.4821140766143799, 'train/grad_norm_pre_clip': 9.257843017578125, 'timing/data': 0.004716749885119498, 'timing/model': 1.20608537690714, 'learning_rate/qwen_vl_interface': 9.999212277670533e-06, 'learning_rate/action_model': 9.999212277670533e-05, 'epoch': 0.04}) +[2026-06-17 06:24:22,195][starVLA.training.train_starvla][INFO] - Step 141, Loss: {'train/loss': 1.2005984783172607, 'train/grad_norm_pre_clip': 6.9589924812316895, 'timing/data': 0.0050523229874670506, 'timing/model': 1.2071739719249308, 'learning_rate/qwen_vl_interface': 9.999172400385658e-06, 'learning_rate/action_model': 9.999172400385658e-05, 'epoch': 0.04}) +[2026-06-17 06:24:26,973][starVLA.training.train_starvla][INFO] - Step 142, Loss: {'train/loss': 1.1798709630966187, 'train/grad_norm_pre_clip': 5.163106918334961, 'timing/data': 0.004854260943830013, 'timing/model': 1.2036708660889417, 'learning_rate/qwen_vl_interface': 9.999131538592232e-06, 'learning_rate/action_model': 9.999131538592232e-05, 'epoch': 0.04}) +[2026-06-17 06:24:31,756][starVLA.training.train_starvla][INFO] - Step 143, Loss: {'train/loss': 1.4923443794250488, 'train/grad_norm_pre_clip': 4.554978370666504, 'timing/data': 0.004924916080199182, 'timing/model': 1.2049183879280463, 'learning_rate/qwen_vl_interface': 9.999089692298729e-06, 'learning_rate/action_model': 9.99908969229873e-05, 'epoch': 0.04}) +[2026-06-17 06:24:36,503][starVLA.training.train_starvla][INFO] - Step 144, Loss: {'train/loss': 1.3321055173873901, 'train/grad_norm_pre_clip': 5.536668300628662, 'timing/data': 0.00500159605871886, 'timing/model': 1.2471573809161782, 'learning_rate/qwen_vl_interface': 9.999046861513822e-06, 'learning_rate/action_model': 9.999046861513821e-05, 'epoch': 0.04}) +[2026-06-17 06:24:41,235][starVLA.training.train_starvla][INFO] - Step 145, Loss: {'train/loss': 1.1965306997299194, 'train/grad_norm_pre_clip': 4.547173023223877, 'timing/data': 0.004725624923594296, 'timing/model': 1.2786182319978252, 'learning_rate/qwen_vl_interface': 9.99900304624639e-06, 'learning_rate/action_model': 9.99900304624639e-05, 'epoch': 0.04}) +[2026-06-17 06:24:45,970][starVLA.training.train_starvla][INFO] - Step 146, Loss: {'train/loss': 1.1217989921569824, 'train/grad_norm_pre_clip': 6.065626621246338, 'timing/data': 0.004969215951859951, 'timing/model': 1.2047653070185333, 'learning_rate/qwen_vl_interface': 9.998958246505518e-06, 'learning_rate/action_model': 9.998958246505516e-05, 'epoch': 0.04}) +[2026-06-17 06:24:50,771][starVLA.training.train_starvla][INFO] - Step 147, Loss: {'train/loss': 1.086106538772583, 'train/grad_norm_pre_clip': 4.403839588165283, 'timing/data': 0.00659985700622201, 'timing/model': 1.2182183359982446, 'learning_rate/qwen_vl_interface': 9.998912462300489e-06, 'learning_rate/action_model': 9.998912462300488e-05, 'epoch': 0.04}) +[2026-06-17 06:24:55,552][starVLA.training.train_starvla][INFO] - Step 148, Loss: {'train/loss': 1.5825810432434082, 'train/grad_norm_pre_clip': 17.302799224853516, 'timing/data': 0.005307564977556467, 'timing/model': 1.2052606609649956, 'learning_rate/qwen_vl_interface': 9.998865693640797e-06, 'learning_rate/action_model': 9.998865693640796e-05, 'epoch': 0.04}) +[2026-06-17 06:25:00,346][starVLA.training.train_starvla][INFO] - Step 149, Loss: {'train/loss': 1.2166320085525513, 'train/grad_norm_pre_clip': 8.894855499267578, 'timing/data': 0.00454961892683059, 'timing/model': 1.205089358962141, 'learning_rate/qwen_vl_interface': 9.998817940536138e-06, 'learning_rate/action_model': 9.998817940536137e-05, 'epoch': 0.04}) +[2026-06-17 06:25:05,161][starVLA.training.train_starvla][INFO] - Step 150, Loss: {'train/loss': 1.039193034172058, 'train/grad_norm_pre_clip': 5.057856559753418, 'timing/data': 0.004743003053590655, 'timing/model': 1.2423242130316794, 'learning_rate/qwen_vl_interface': 9.99876920299641e-06, 'learning_rate/action_model': 9.998769202996408e-05, 'epoch': 0.04}) +[2026-06-17 06:25:09,986][starVLA.training.train_starvla][INFO] - Step 151, Loss: {'train/loss': 1.436366081237793, 'train/grad_norm_pre_clip': 11.84085750579834, 'timing/data': 0.004926525987684727, 'timing/model': 1.288608951959759, 'learning_rate/qwen_vl_interface': 9.998719481031715e-06, 'learning_rate/action_model': 9.998719481031715e-05, 'epoch': 0.04}) +[2026-06-17 06:25:14,668][starVLA.training.train_starvla][INFO] - Step 152, Loss: {'train/loss': 1.428068995475769, 'train/grad_norm_pre_clip': 13.998936653137207, 'timing/data': 0.0051632310496643186, 'timing/model': 1.2152969009475783, 'learning_rate/qwen_vl_interface': 9.998668774652363e-06, 'learning_rate/action_model': 9.998668774652362e-05, 'epoch': 0.04}) +[2026-06-17 06:25:19,451][starVLA.training.train_starvla][INFO] - Step 153, Loss: {'train/loss': 1.318762183189392, 'train/grad_norm_pre_clip': 9.272568702697754, 'timing/data': 0.004758154973387718, 'timing/model': 1.2050931539852172, 'learning_rate/qwen_vl_interface': 9.998617083868863e-06, 'learning_rate/action_model': 9.998617083868863e-05, 'epoch': 0.05}) +[2026-06-17 06:25:24,244][starVLA.training.train_starvla][INFO] - Step 154, Loss: {'train/loss': 1.4070872068405151, 'train/grad_norm_pre_clip': 8.128662109375, 'timing/data': 0.004721597069874406, 'timing/model': 1.2100074369227514, 'learning_rate/qwen_vl_interface': 9.998564408691933e-06, 'learning_rate/action_model': 9.998564408691932e-05, 'epoch': 0.05}) +[2026-06-17 06:25:29,041][starVLA.training.train_starvla][INFO] - Step 155, Loss: {'train/loss': 1.2430704832077026, 'train/grad_norm_pre_clip': 9.964194297790527, 'timing/data': 0.004666605964303017, 'timing/model': 1.2066301100421697, 'learning_rate/qwen_vl_interface': 9.998510749132491e-06, 'learning_rate/action_model': 9.998510749132491e-05, 'epoch': 0.05}) +[2026-06-17 06:25:33,825][starVLA.training.train_starvla][INFO] - Step 156, Loss: {'train/loss': 1.3685239553451538, 'train/grad_norm_pre_clip': 7.407119274139404, 'timing/data': 0.005045472062192857, 'timing/model': 1.3318883809261024, 'learning_rate/qwen_vl_interface': 9.998456105201659e-06, 'learning_rate/action_model': 9.99845610520166e-05, 'epoch': 0.05}) +[2026-06-17 06:25:38,488][starVLA.training.train_starvla][INFO] - Step 157, Loss: {'train/loss': 1.447432041168213, 'train/grad_norm_pre_clip': 9.627692222595215, 'timing/data': 0.004754319088533521, 'timing/model': 1.2069044020026922, 'learning_rate/qwen_vl_interface': 9.99840047691077e-06, 'learning_rate/action_model': 9.99840047691077e-05, 'epoch': 0.05}) +[2026-06-17 06:25:43,276][starVLA.training.train_starvla][INFO] - Step 158, Loss: {'train/loss': 1.195473313331604, 'train/grad_norm_pre_clip': 6.409840106964111, 'timing/data': 0.004566134070046246, 'timing/model': 1.207622884074226, 'learning_rate/qwen_vl_interface': 9.998343864271352e-06, 'learning_rate/action_model': 9.998343864271351e-05, 'epoch': 0.05}) +[2026-06-17 06:25:47,960][starVLA.training.train_starvla][INFO] - Step 159, Loss: {'train/loss': 1.3406428098678589, 'train/grad_norm_pre_clip': 6.616327285766602, 'timing/data': 0.004727128078229725, 'timing/model': 1.2072514690225944, 'learning_rate/qwen_vl_interface': 9.99828626729514e-06, 'learning_rate/action_model': 9.998286267295139e-05, 'epoch': 0.05}) +[2026-06-17 06:25:52,673][starVLA.training.train_starvla][INFO] - Step 160, Loss: {'train/loss': 1.2094320058822632, 'train/grad_norm_pre_clip': 7.254509449005127, 'timing/data': 0.0048154700780287385, 'timing/model': 1.207759415032342, 'learning_rate/qwen_vl_interface': 9.998227685994077e-06, 'learning_rate/action_model': 9.998227685994075e-05, 'epoch': 0.05}) +[2026-06-17 06:25:57,408][starVLA.training.train_starvla][INFO] - Step 161, Loss: {'train/loss': 1.2262362241744995, 'train/grad_norm_pre_clip': 3.9451277256011963, 'timing/data': 0.004910445073619485, 'timing/model': 1.204686701996252, 'learning_rate/qwen_vl_interface': 9.998168120380303e-06, 'learning_rate/action_model': 9.998168120380303e-05, 'epoch': 0.05}) +[2026-06-17 06:26:02,188][starVLA.training.train_starvla][INFO] - Step 162, Loss: {'train/loss': 1.2765825986862183, 'train/grad_norm_pre_clip': 5.388712406158447, 'timing/data': 0.004859243985265493, 'timing/model': 1.3312931780237705, 'learning_rate/qwen_vl_interface': 9.99810757046617e-06, 'learning_rate/action_model': 9.99810757046617e-05, 'epoch': 0.05}) +[2026-06-17 06:26:06,843][starVLA.training.train_starvla][INFO] - Step 163, Loss: {'train/loss': 1.3510311841964722, 'train/grad_norm_pre_clip': 6.5897393226623535, 'timing/data': 0.004991858033463359, 'timing/model': 1.2048336799489334, 'learning_rate/qwen_vl_interface': 9.998046036264229e-06, 'learning_rate/action_model': 9.99804603626423e-05, 'epoch': 0.05}) +[2026-06-17 06:26:11,634][starVLA.training.train_starvla][INFO] - Step 164, Loss: {'train/loss': 1.3018029928207397, 'train/grad_norm_pre_clip': 7.13700008392334, 'timing/data': 0.005130799021571875, 'timing/model': 1.2054481910308823, 'learning_rate/qwen_vl_interface': 9.997983517787236e-06, 'learning_rate/action_model': 9.997983517787236e-05, 'epoch': 0.05}) +[2026-06-17 06:26:16,421][starVLA.training.train_starvla][INFO] - Step 165, Loss: {'train/loss': 1.2430607080459595, 'train/grad_norm_pre_clip': 3.45620059967041, 'timing/data': 0.0047166700242087245, 'timing/model': 1.2066317000426352, 'learning_rate/qwen_vl_interface': 9.99792001504815e-06, 'learning_rate/action_model': 9.997920015048149e-05, 'epoch': 0.05}) +[2026-06-17 06:26:21,203][starVLA.training.train_starvla][INFO] - Step 166, Loss: {'train/loss': 1.3150131702423096, 'train/grad_norm_pre_clip': 4.488545894622803, 'timing/data': 0.004896452999673784, 'timing/model': 1.2073175410041586, 'learning_rate/qwen_vl_interface': 9.997855528060136e-06, 'learning_rate/action_model': 9.997855528060135e-05, 'epoch': 0.05}) +[2026-06-17 06:26:25,951][starVLA.training.train_starvla][INFO] - Step 167, Loss: {'train/loss': 1.322770118713379, 'train/grad_norm_pre_clip': 4.717031955718994, 'timing/data': 0.004862754023633897, 'timing/model': 1.208502196939662, 'learning_rate/qwen_vl_interface': 9.997790056836562e-06, 'learning_rate/action_model': 9.997790056836562e-05, 'epoch': 0.05}) +[2026-06-17 06:26:30,762][starVLA.training.train_starvla][INFO] - Step 168, Loss: {'train/loss': 1.2817199230194092, 'train/grad_norm_pre_clip': 7.383274078369141, 'timing/data': 0.004918650025501847, 'timing/model': 1.3392884019995108, 'learning_rate/qwen_vl_interface': 9.997723601390998e-06, 'learning_rate/action_model': 9.997723601390998e-05, 'epoch': 0.05}) +[2026-06-17 06:26:35,430][starVLA.training.train_starvla][INFO] - Step 169, Loss: {'train/loss': 1.3770411014556885, 'train/grad_norm_pre_clip': 12.111967086791992, 'timing/data': 0.005086787045001984, 'timing/model': 1.2118564159609377, 'learning_rate/qwen_vl_interface': 9.997656161737227e-06, 'learning_rate/action_model': 9.997656161737226e-05, 'epoch': 0.05}) +[2026-06-17 06:26:40,232][starVLA.training.train_starvla][INFO] - Step 170, Loss: {'train/loss': 1.3041640520095825, 'train/grad_norm_pre_clip': 4.985361099243164, 'timing/data': 0.004926012014038861, 'timing/model': 1.2071896009147167, 'learning_rate/qwen_vl_interface': 9.997587737889222e-06, 'learning_rate/action_model': 9.997587737889221e-05, 'epoch': 0.05}) +[2026-06-17 06:26:45,022][starVLA.training.train_starvla][INFO] - Step 171, Loss: {'train/loss': 1.3039910793304443, 'train/grad_norm_pre_clip': 6.822546482086182, 'timing/data': 0.004512001993134618, 'timing/model': 1.2069614740321413, 'learning_rate/qwen_vl_interface': 9.997518329861172e-06, 'learning_rate/action_model': 9.997518329861171e-05, 'epoch': 0.05}) +[2026-06-17 06:26:49,811][starVLA.training.train_starvla][INFO] - Step 172, Loss: {'train/loss': 1.3085947036743164, 'train/grad_norm_pre_clip': 7.86067533493042, 'timing/data': 0.00491077802143991, 'timing/model': 1.206646803067997, 'learning_rate/qwen_vl_interface': 9.997447937667462e-06, 'learning_rate/action_model': 9.997447937667462e-05, 'epoch': 0.05}) +[2026-06-17 06:26:54,573][starVLA.training.train_starvla][INFO] - Step 173, Loss: {'train/loss': 1.413532018661499, 'train/grad_norm_pre_clip': 15.145554542541504, 'timing/data': 0.004823518916964531, 'timing/model': 1.2063485539983958, 'learning_rate/qwen_vl_interface': 9.997376561322686e-06, 'learning_rate/action_model': 9.997376561322686e-05, 'epoch': 0.05}) +[2026-06-17 06:26:59,369][starVLA.training.train_starvla][INFO] - Step 174, Loss: {'train/loss': 1.1040669679641724, 'train/grad_norm_pre_clip': 7.487828731536865, 'timing/data': 0.0047609490575268865, 'timing/model': 1.3363647309597582, 'learning_rate/qwen_vl_interface': 9.997304200841642e-06, 'learning_rate/action_model': 9.997304200841641e-05, 'epoch': 0.05}) +[2026-06-17 06:27:04,037][starVLA.training.train_starvla][INFO] - Step 175, Loss: {'train/loss': 0.9916265606880188, 'train/grad_norm_pre_clip': 5.265015125274658, 'timing/data': 0.0048656120197847486, 'timing/model': 1.2101401260588318, 'learning_rate/qwen_vl_interface': 9.997230856239327e-06, 'learning_rate/action_model': 9.997230856239327e-05, 'epoch': 0.05}) +[2026-06-17 06:27:08,828][starVLA.training.train_starvla][INFO] - Step 176, Loss: {'train/loss': 1.1441714763641357, 'train/grad_norm_pre_clip': 6.160147190093994, 'timing/data': 0.004670445923693478, 'timing/model': 1.2079693480627611, 'learning_rate/qwen_vl_interface': 9.997156527530948e-06, 'learning_rate/action_model': 9.997156527530948e-05, 'epoch': 0.05}) +[2026-06-17 06:27:13,641][starVLA.training.train_starvla][INFO] - Step 177, Loss: {'train/loss': 1.201850414276123, 'train/grad_norm_pre_clip': 9.287919998168945, 'timing/data': 0.0048977750120684505, 'timing/model': 1.2077043569879606, 'learning_rate/qwen_vl_interface': 9.997081214731915e-06, 'learning_rate/action_model': 9.997081214731914e-05, 'epoch': 0.05}) +[2026-06-17 06:27:18,383][starVLA.training.train_starvla][INFO] - Step 178, Loss: {'train/loss': 1.2589328289031982, 'train/grad_norm_pre_clip': 16.304641723632812, 'timing/data': 0.004692138056270778, 'timing/model': 1.207681588944979, 'learning_rate/qwen_vl_interface': 9.997004917857834e-06, 'learning_rate/action_model': 9.997004917857834e-05, 'epoch': 0.05}) +[2026-06-17 06:27:23,173][starVLA.training.train_starvla][INFO] - Step 179, Loss: {'train/loss': 1.283174991607666, 'train/grad_norm_pre_clip': 12.242267608642578, 'timing/data': 0.004930904018692672, 'timing/model': 1.205884635914117, 'learning_rate/qwen_vl_interface': 9.996927636924529e-06, 'learning_rate/action_model': 9.996927636924528e-05, 'epoch': 0.05}) +[2026-06-17 06:27:28,003][starVLA.training.train_starvla][INFO] - Step 180, Loss: {'train/loss': 1.4269872903823853, 'train/grad_norm_pre_clip': 8.966265678405762, 'timing/data': 0.005193716962821782, 'timing/model': 1.372854745015502, 'learning_rate/qwen_vl_interface': 9.996849371948016e-06, 'learning_rate/action_model': 9.996849371948016e-05, 'epoch': 0.05}) +[2026-06-17 06:27:32,677][starVLA.training.train_starvla][INFO] - Step 181, Loss: {'train/loss': 1.2835923433303833, 'train/grad_norm_pre_clip': 7.122099876403809, 'timing/data': 0.005016003968194127, 'timing/model': 1.2089279450010508, 'learning_rate/qwen_vl_interface': 9.996770122944518e-06, 'learning_rate/action_model': 9.996770122944518e-05, 'epoch': 0.05}) +[2026-06-17 06:27:37,529][starVLA.training.train_starvla][INFO] - Step 182, Loss: {'train/loss': 1.376161813735962, 'train/grad_norm_pre_clip': 10.48200798034668, 'timing/data': 0.004586230963468552, 'timing/model': 1.2076894430210814, 'learning_rate/qwen_vl_interface': 9.99668988993047e-06, 'learning_rate/action_model': 9.99668988993047e-05, 'epoch': 0.05}) +[2026-06-17 06:27:42,278][starVLA.training.train_starvla][INFO] - Step 183, Loss: {'train/loss': 1.3048065900802612, 'train/grad_norm_pre_clip': 7.423925876617432, 'timing/data': 0.004655212978832424, 'timing/model': 1.2072252549696714, 'learning_rate/qwen_vl_interface': 9.996608672922496e-06, 'learning_rate/action_model': 9.996608672922496e-05, 'epoch': 0.05}) +[2026-06-17 06:27:47,059][starVLA.training.train_starvla][INFO] - Step 184, Loss: {'train/loss': 1.120481252670288, 'train/grad_norm_pre_clip': 5.316213607788086, 'timing/data': 0.004741062060929835, 'timing/model': 1.2122591740917414, 'learning_rate/qwen_vl_interface': 9.996526471937438e-06, 'learning_rate/action_model': 9.996526471937439e-05, 'epoch': 0.05}) +[2026-06-17 06:27:51,919][starVLA.training.train_starvla][INFO] - Step 185, Loss: {'train/loss': 1.2333343029022217, 'train/grad_norm_pre_clip': 12.2181978225708, 'timing/data': 0.004886000999249518, 'timing/model': 1.2109900790965185, 'learning_rate/qwen_vl_interface': 9.996443286992337e-06, 'learning_rate/action_model': 9.996443286992336e-05, 'epoch': 0.05}) +[2026-06-17 06:27:56,732][starVLA.training.train_starvla][INFO] - Step 186, Loss: {'train/loss': 1.408318281173706, 'train/grad_norm_pre_clip': 9.0071439743042, 'timing/data': 0.0042395469499751925, 'timing/model': 1.3429933190345764, 'learning_rate/qwen_vl_interface': 9.996359118104432e-06, 'learning_rate/action_model': 9.996359118104432e-05, 'epoch': 0.05}) +[2026-06-17 06:28:01,427][starVLA.training.train_starvla][INFO] - Step 187, Loss: {'train/loss': 1.1745140552520752, 'train/grad_norm_pre_clip': 7.889152526855469, 'timing/data': 0.004885138012468815, 'timing/model': 1.2216167809674516, 'learning_rate/qwen_vl_interface': 9.996273965291177e-06, 'learning_rate/action_model': 9.996273965291176e-05, 'epoch': 0.06}) +[2026-06-17 06:28:06,235][starVLA.training.train_starvla][INFO] - Step 188, Loss: {'train/loss': 1.5789470672607422, 'train/grad_norm_pre_clip': 13.661553382873535, 'timing/data': 0.00472140999045223, 'timing/model': 1.2211302869254723, 'learning_rate/qwen_vl_interface': 9.996187828570221e-06, 'learning_rate/action_model': 9.996187828570221e-05, 'epoch': 0.06}) +[2026-06-17 06:28:11,023][starVLA.training.train_starvla][INFO] - Step 189, Loss: {'train/loss': 1.2944403886795044, 'train/grad_norm_pre_clip': 7.516049861907959, 'timing/data': 0.005105169024318457, 'timing/model': 1.2057479869108647, 'learning_rate/qwen_vl_interface': 9.996100707959422e-06, 'learning_rate/action_model': 9.996100707959421e-05, 'epoch': 0.06}) +[2026-06-17 06:28:15,794][starVLA.training.train_starvla][INFO] - Step 190, Loss: {'train/loss': 1.1353390216827393, 'train/grad_norm_pre_clip': 5.939338207244873, 'timing/data': 0.004737496026791632, 'timing/model': 1.2133423999184743, 'learning_rate/qwen_vl_interface': 9.996012603476836e-06, 'learning_rate/action_model': 9.996012603476837e-05, 'epoch': 0.06}) +[2026-06-17 06:28:20,614][starVLA.training.train_starvla][INFO] - Step 191, Loss: {'train/loss': 1.3485100269317627, 'train/grad_norm_pre_clip': 11.244085311889648, 'timing/data': 0.00474231003317982, 'timing/model': 1.21470205206424, 'learning_rate/qwen_vl_interface': 9.995923515140734e-06, 'learning_rate/action_model': 9.995923515140732e-05, 'epoch': 0.06}) +[2026-06-17 06:28:25,441][starVLA.training.train_starvla][INFO] - Step 192, Loss: {'train/loss': 1.1242401599884033, 'train/grad_norm_pre_clip': 6.684394359588623, 'timing/data': 0.0049400549614802, 'timing/model': 1.346525424043648, 'learning_rate/qwen_vl_interface': 9.995833442969578e-06, 'learning_rate/action_model': 9.995833442969578e-05, 'epoch': 0.06}) +[2026-06-17 06:28:30,113][starVLA.training.train_starvla][INFO] - Step 193, Loss: {'train/loss': 1.1176035404205322, 'train/grad_norm_pre_clip': 10.984086036682129, 'timing/data': 0.004766872967593372, 'timing/model': 1.210147061967291, 'learning_rate/qwen_vl_interface': 9.995742386982045e-06, 'learning_rate/action_model': 9.995742386982045e-05, 'epoch': 0.06}) +[2026-06-17 06:28:34,912][starVLA.training.train_starvla][INFO] - Step 194, Loss: {'train/loss': 1.3671284914016724, 'train/grad_norm_pre_clip': 7.616445064544678, 'timing/data': 0.004992370959371328, 'timing/model': 1.214824800961651, 'learning_rate/qwen_vl_interface': 9.995650347197006e-06, 'learning_rate/action_model': 9.995650347197005e-05, 'epoch': 0.06}) +[2026-06-17 06:28:39,684][starVLA.training.train_starvla][INFO] - Step 195, Loss: {'train/loss': 1.5157073736190796, 'train/grad_norm_pre_clip': 9.861824035644531, 'timing/data': 0.004888135939836502, 'timing/model': 1.206513129058294, 'learning_rate/qwen_vl_interface': 9.995557323633544e-06, 'learning_rate/action_model': 9.995557323633543e-05, 'epoch': 0.06}) +[2026-06-17 06:28:44,501][starVLA.training.train_starvla][INFO] - Step 196, Loss: {'train/loss': 1.2067763805389404, 'train/grad_norm_pre_clip': 8.414593696594238, 'timing/data': 0.005111249978654087, 'timing/model': 1.2081721770809963, 'learning_rate/qwen_vl_interface': 9.995463316310942e-06, 'learning_rate/action_model': 9.995463316310943e-05, 'epoch': 0.06}) +[2026-06-17 06:28:49,304][starVLA.training.train_starvla][INFO] - Step 197, Loss: {'train/loss': 1.3502140045166016, 'train/grad_norm_pre_clip': 10.313666343688965, 'timing/data': 0.004811485996469855, 'timing/model': 1.2092429810436442, 'learning_rate/qwen_vl_interface': 9.995368325248688e-06, 'learning_rate/action_model': 9.995368325248688e-05, 'epoch': 0.06}) +[2026-06-17 06:28:54,235][starVLA.training.train_starvla][INFO] - Step 198, Loss: {'train/loss': 1.5663281679153442, 'train/grad_norm_pre_clip': 14.949321746826172, 'timing/data': 0.004793213098309934, 'timing/model': 1.4575084829702973, 'learning_rate/qwen_vl_interface': 9.995272350466476e-06, 'learning_rate/action_model': 9.995272350466476e-05, 'epoch': 0.06}) +[2026-06-17 06:28:58,952][starVLA.training.train_starvla][INFO] - Step 199, Loss: {'train/loss': 1.4855027198791504, 'train/grad_norm_pre_clip': 9.0687894821167, 'timing/data': 0.004358754958957434, 'timing/model': 1.2528476109728217, 'learning_rate/qwen_vl_interface': 9.995175391984199e-06, 'learning_rate/action_model': 9.995175391984198e-05, 'epoch': 0.06}) +[2026-06-17 06:30:40,866][starVLA.training.train_starvla][INFO] - Step 200, Loss: {'train/loss': 1.065890908241272, 'train/grad_norm_pre_clip': 6.304017543792725, 'eval/loss': 1.3122532367706299, 'eval/latency_0/loss': 1.3122532367706299, 'eval/demon_fire_precision': 0.9196661450182577, 'eval/demon_fire_recall': 0.7321428571428571, 'eval/demon_fire_f1': 0.8152601156069363, 'eval/demon_fire_rate_model': 0.5330923248053393, 'eval/demon_fire_rate_teacher': 0.6696329254727474, 'eval/demon_left_precision': 0.5671417854463616, 'eval/demon_left_recall': 0.856657223796034, 'eval/demon_left_f1': 0.6824644549763034, 'eval/demon_left_rate_model': 0.7413793103448276, 'eval/demon_left_rate_teacher': 0.49082313681868744, 'eval/demon_right_precision': 0.6709677419354839, 'eval/demon_right_recall': 0.3525423728813559, 'eval/demon_right_f1': 0.46222222222222215, 'eval/demon_right_rate_model': 0.25862068965517243, 'eval/demon_right_rate_teacher': 0.4922135706340378, 'eval/demon_move_f1': 0.5723433385992628, 'eval/demon_cc_f1': 0.7231794952811992, 'eval/latency_0/demon_fire_precision': 0.9196661450182577, 'eval/latency_0/demon_fire_recall': 0.7321428571428571, 'eval/latency_0/demon_fire_f1': 0.8152601156069363, 'eval/latency_0/demon_fire_rate_model': 0.5330923248053393, 'eval/latency_0/demon_fire_rate_teacher': 0.6696329254727474, 'eval/latency_0/demon_left_precision': 0.5671417854463616, 'eval/latency_0/demon_left_recall': 0.856657223796034, 'eval/latency_0/demon_left_f1': 0.6824644549763034, 'eval/latency_0/demon_left_rate_model': 0.7413793103448276, 'eval/latency_0/demon_left_rate_teacher': 0.49082313681868744, 'eval/latency_0/demon_right_precision': 0.6709677419354839, 'eval/latency_0/demon_right_recall': 0.3525423728813559, 'eval/latency_0/demon_right_f1': 0.46222222222222215, 'eval/latency_0/demon_right_rate_model': 0.25862068965517243, 'eval/latency_0/demon_right_rate_teacher': 0.4922135706340378, 'eval/latency_0/demon_move_f1': 0.5723433385992628, 'eval/latency_0/demon_cc_f1': 0.7231794952811992, 'timing/data': 0.005421150010079145, 'timing/model': 1.2065144020598382, 'learning_rate/qwen_vl_interface': 9.995077449821957e-06, 'learning_rate/action_model': 9.995077449821956e-05, 'epoch': 0.06}) +[2026-06-17 06:30:40,866][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_200_state +[2026-06-17 06:30:40,874][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 06:30:44,329][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_200_state/model.safetensors +[2026-06-17 06:30:52,350][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_200_state/optimizer.bin +[2026-06-17 06:30:52,351][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_200_state/sampler.bin +[2026-06-17 06:30:52,351][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_200_state/sampler_1.bin +[2026-06-17 06:30:52,352][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_200_state/random_states_0.pkl +[2026-06-17 06:30:52,373][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 200 +[2026-06-17 06:30:52,373][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 06:30:52,381][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 06:30:57,095][starVLA.training.train_starvla][INFO] - Step 201, Loss: {'train/loss': 1.469387173652649, 'train/grad_norm_pre_clip': 10.435904502868652, 'timing/data': 0.004672327893786132, 'timing/model': 1.316538453917019, 'learning_rate/qwen_vl_interface': 9.994978524000053e-06, 'learning_rate/action_model': 9.994978524000052e-05, 'epoch': 0.06}) +[2026-06-17 06:31:01,704][starVLA.training.train_starvla][INFO] - Step 202, Loss: {'train/loss': 1.441853642463684, 'train/grad_norm_pre_clip': 12.843385696411133, 'timing/data': 0.004908233997412026, 'timing/model': 1.198188778012991, 'learning_rate/qwen_vl_interface': 9.994878614538996e-06, 'learning_rate/action_model': 9.994878614538996e-05, 'epoch': 0.06}) +[2026-06-17 06:31:06,452][starVLA.training.train_starvla][INFO] - Step 203, Loss: {'train/loss': 1.523168683052063, 'train/grad_norm_pre_clip': 6.557844638824463, 'timing/data': 0.0049765100702643394, 'timing/model': 1.194384987000376, 'learning_rate/qwen_vl_interface': 9.994777721459496e-06, 'learning_rate/action_model': 9.994777721459497e-05, 'epoch': 0.06}) +[2026-06-17 06:31:11,266][starVLA.training.train_starvla][INFO] - Step 204, Loss: {'train/loss': 1.39655339717865, 'train/grad_norm_pre_clip': 8.1596040725708, 'timing/data': 0.00467620010022074, 'timing/model': 1.1957191380206496, 'learning_rate/qwen_vl_interface': 9.994675844782468e-06, 'learning_rate/action_model': 9.994675844782468e-05, 'epoch': 0.06}) +[2026-06-17 06:31:16,078][starVLA.training.train_starvla][INFO] - Step 205, Loss: {'train/loss': 1.1961050033569336, 'train/grad_norm_pre_clip': 5.77410364151001, 'timing/data': 0.004427765030413866, 'timing/model': 1.199888748000376, 'learning_rate/qwen_vl_interface': 9.994572984529033e-06, 'learning_rate/action_model': 9.994572984529034e-05, 'epoch': 0.06}) +[2026-06-17 06:31:20,824][starVLA.training.train_starvla][INFO] - Step 206, Loss: {'train/loss': 1.3091312646865845, 'train/grad_norm_pre_clip': 5.084348201751709, 'timing/data': 0.004599904990755022, 'timing/model': 1.2000697969924659, 'learning_rate/qwen_vl_interface': 9.994469140720513e-06, 'learning_rate/action_model': 9.994469140720512e-05, 'epoch': 0.06}) +[2026-06-17 06:31:25,507][starVLA.training.train_starvla][INFO] - Step 207, Loss: {'train/loss': 1.1648437976837158, 'train/grad_norm_pre_clip': 6.284719944000244, 'timing/data': 0.004265410010702908, 'timing/model': 1.2515595250297338, 'learning_rate/qwen_vl_interface': 9.994364313378435e-06, 'learning_rate/action_model': 9.994364313378433e-05, 'epoch': 0.06}) +[2026-06-17 06:31:30,143][starVLA.training.train_starvla][INFO] - Step 208, Loss: {'train/loss': 1.2910789251327515, 'train/grad_norm_pre_clip': 8.083893775939941, 'timing/data': 0.004861409077420831, 'timing/model': 1.2031172690913081, 'learning_rate/qwen_vl_interface': 9.99425850252453e-06, 'learning_rate/action_model': 9.994258502524529e-05, 'epoch': 0.06}) +[2026-06-17 06:31:34,916][starVLA.training.train_starvla][INFO] - Step 209, Loss: {'train/loss': 1.437989592552185, 'train/grad_norm_pre_clip': 12.573095321655273, 'timing/data': 0.004885133006609976, 'timing/model': 1.241652413038537, 'learning_rate/qwen_vl_interface': 9.994151708180731e-06, 'learning_rate/action_model': 9.99415170818073e-05, 'epoch': 0.06}) +[2026-06-17 06:31:39,683][starVLA.training.train_starvla][INFO] - Step 210, Loss: {'train/loss': 1.2814841270446777, 'train/grad_norm_pre_clip': 7.288584232330322, 'timing/data': 0.005204341025091708, 'timing/model': 1.2022482759784907, 'learning_rate/qwen_vl_interface': 9.994043930369179e-06, 'learning_rate/action_model': 9.994043930369178e-05, 'epoch': 0.06}) +[2026-06-17 06:31:44,461][starVLA.training.train_starvla][INFO] - Step 211, Loss: {'train/loss': 1.4001274108886719, 'train/grad_norm_pre_clip': 7.628810405731201, 'timing/data': 0.004769849008880556, 'timing/model': 1.2016981249907985, 'learning_rate/qwen_vl_interface': 9.993935169112213e-06, 'learning_rate/action_model': 9.993935169112213e-05, 'epoch': 0.06}) +[2026-06-17 06:31:49,253][starVLA.training.train_starvla][INFO] - Step 212, Loss: {'train/loss': 1.1382015943527222, 'train/grad_norm_pre_clip': 4.6823506355285645, 'timing/data': 0.006556125939823687, 'timing/model': 1.3436411370057613, 'learning_rate/qwen_vl_interface': 9.993825424432385e-06, 'learning_rate/action_model': 9.993825424432385e-05, 'epoch': 0.06}) +[2026-06-17 06:31:53,912][starVLA.training.train_starvla][INFO] - Step 213, Loss: {'train/loss': 1.3652762174606323, 'train/grad_norm_pre_clip': 4.9811692237854, 'timing/data': 0.004666651017032564, 'timing/model': 1.2034661669749767, 'learning_rate/qwen_vl_interface': 9.99371469635244e-06, 'learning_rate/action_model': 9.99371469635244e-05, 'epoch': 0.06}) +[2026-06-17 06:31:58,703][starVLA.training.train_starvla][INFO] - Step 214, Loss: {'train/loss': 1.3154550790786743, 'train/grad_norm_pre_clip': 9.815840721130371, 'timing/data': 0.004893676959909499, 'timing/model': 1.2053381599253044, 'learning_rate/qwen_vl_interface': 9.993602984895336e-06, 'learning_rate/action_model': 9.993602984895335e-05, 'epoch': 0.06}) +[2026-06-17 06:32:03,488][starVLA.training.train_starvla][INFO] - Step 215, Loss: {'train/loss': 1.3234344720840454, 'train/grad_norm_pre_clip': 4.40635347366333, 'timing/data': 0.0046625250251963735, 'timing/model': 1.202138904016465, 'learning_rate/qwen_vl_interface': 9.993490290084225e-06, 'learning_rate/action_model': 9.993490290084225e-05, 'epoch': 0.06}) +[2026-06-17 06:32:08,263][starVLA.training.train_starvla][INFO] - Step 216, Loss: {'train/loss': 1.4041696786880493, 'train/grad_norm_pre_clip': 6.376291751861572, 'timing/data': 0.004512069979682565, 'timing/model': 1.2027469599852338, 'learning_rate/qwen_vl_interface': 9.993376611942477e-06, 'learning_rate/action_model': 9.993376611942477e-05, 'epoch': 0.06}) +[2026-06-17 06:32:12,970][starVLA.training.train_starvla][INFO] - Step 217, Loss: {'train/loss': 1.3856972455978394, 'train/grad_norm_pre_clip': 7.955872058868408, 'timing/data': 0.004774082917720079, 'timing/model': 1.2043436140520498, 'learning_rate/qwen_vl_interface': 9.99326195049365e-06, 'learning_rate/action_model': 9.99326195049365e-05, 'epoch': 0.06}) +[2026-06-17 06:32:17,702][starVLA.training.train_starvla][INFO] - Step 218, Loss: {'train/loss': 1.332877516746521, 'train/grad_norm_pre_clip': 5.5044965744018555, 'timing/data': 0.0046517939772456884, 'timing/model': 1.2804229279281572, 'learning_rate/qwen_vl_interface': 9.99314630576152e-06, 'learning_rate/action_model': 9.993146305761519e-05, 'epoch': 0.06}) +[2026-06-17 06:32:22,363][starVLA.training.train_starvla][INFO] - Step 219, Loss: {'train/loss': 1.0792020559310913, 'train/grad_norm_pre_clip': 4.0215959548950195, 'timing/data': 0.00484793889336288, 'timing/model': 1.2060990070458502, 'learning_rate/qwen_vl_interface': 9.993029677770055e-06, 'learning_rate/action_model': 9.993029677770055e-05, 'epoch': 0.06}) +[2026-06-17 06:32:27,157][starVLA.training.train_starvla][INFO] - Step 220, Loss: {'train/loss': 1.1797120571136475, 'train/grad_norm_pre_clip': 5.236654758453369, 'timing/data': 0.004648678936064243, 'timing/model': 1.2054269940126687, 'learning_rate/qwen_vl_interface': 9.992912066543433e-06, 'learning_rate/action_model': 9.992912066543433e-05, 'epoch': 0.06}) +[2026-06-17 06:32:31,946][starVLA.training.train_starvla][INFO] - Step 221, Loss: {'train/loss': 1.4236990213394165, 'train/grad_norm_pre_clip': 8.752727508544922, 'timing/data': 0.004753408022224903, 'timing/model': 1.209703573025763, 'learning_rate/qwen_vl_interface': 9.992793472106039e-06, 'learning_rate/action_model': 9.992793472106038e-05, 'epoch': 0.07}) +[2026-06-17 06:32:36,786][starVLA.training.train_starvla][INFO] - Step 222, Loss: {'train/loss': 1.288926124572754, 'train/grad_norm_pre_clip': 7.729488372802734, 'timing/data': 0.004539987072348595, 'timing/model': 1.2027601739391685, 'learning_rate/qwen_vl_interface': 9.992673894482452e-06, 'learning_rate/action_model': 9.992673894482451e-05, 'epoch': 0.07}) +[2026-06-17 06:32:41,552][starVLA.training.train_starvla][INFO] - Step 223, Loss: {'train/loss': 1.0871295928955078, 'train/grad_norm_pre_clip': 5.16439151763916, 'timing/data': 0.004725285922177136, 'timing/model': 1.2035525589017197, 'learning_rate/qwen_vl_interface': 9.992553333697464e-06, 'learning_rate/action_model': 9.992553333697463e-05, 'epoch': 0.07}) +[2026-06-17 06:32:46,324][starVLA.training.train_starvla][INFO] - Step 224, Loss: {'train/loss': 1.3458176851272583, 'train/grad_norm_pre_clip': 7.650910377502441, 'timing/data': 0.00492215808480978, 'timing/model': 1.3005100950831547, 'learning_rate/qwen_vl_interface': 9.992431789776067e-06, 'learning_rate/action_model': 9.992431789776066e-05, 'epoch': 0.07}) +[2026-06-17 06:32:50,988][starVLA.training.train_starvla][INFO] - Step 225, Loss: {'train/loss': 1.138575792312622, 'train/grad_norm_pre_clip': 6.6029157638549805, 'timing/data': 0.004823187948204577, 'timing/model': 1.2048164550215006, 'learning_rate/qwen_vl_interface': 9.992309262743458e-06, 'learning_rate/action_model': 9.992309262743457e-05, 'epoch': 0.07}) +[2026-06-17 06:32:55,847][starVLA.training.train_starvla][INFO] - Step 226, Loss: {'train/loss': 1.2255421876907349, 'train/grad_norm_pre_clip': 10.027942657470703, 'timing/data': 0.004635169985704124, 'timing/model': 1.2055898118996993, 'learning_rate/qwen_vl_interface': 9.992185752625035e-06, 'learning_rate/action_model': 9.992185752625033e-05, 'epoch': 0.07}) +[2026-06-17 06:33:00,632][starVLA.training.train_starvla][INFO] - Step 227, Loss: {'train/loss': 1.18545663356781, 'train/grad_norm_pre_clip': 7.205291271209717, 'timing/data': 0.0046265110140666366, 'timing/model': 1.204540934995748, 'learning_rate/qwen_vl_interface': 9.992061259446402e-06, 'learning_rate/action_model': 9.992061259446401e-05, 'epoch': 0.07}) +[2026-06-17 06:33:05,416][starVLA.training.train_starvla][INFO] - Step 228, Loss: {'train/loss': 1.130127191543579, 'train/grad_norm_pre_clip': 5.380744934082031, 'timing/data': 0.004637446021661162, 'timing/model': 1.2072602029656991, 'learning_rate/qwen_vl_interface': 9.991935783233368e-06, 'learning_rate/action_model': 9.991935783233367e-05, 'epoch': 0.07}) +[2026-06-17 06:33:10,202][starVLA.training.train_starvla][INFO] - Step 229, Loss: {'train/loss': 1.0320578813552856, 'train/grad_norm_pre_clip': 7.58750581741333, 'timing/data': 0.0047396449372172356, 'timing/model': 1.3312171520665288, 'learning_rate/qwen_vl_interface': 9.991809324011943e-06, 'learning_rate/action_model': 9.991809324011943e-05, 'epoch': 0.07}) +[2026-06-17 06:33:14,854][starVLA.training.train_starvla][INFO] - Step 230, Loss: {'train/loss': 1.440077781677246, 'train/grad_norm_pre_clip': 10.14986801147461, 'timing/data': 0.004935011966153979, 'timing/model': 1.2048442979576066, 'learning_rate/qwen_vl_interface': 9.991681881808343e-06, 'learning_rate/action_model': 9.991681881808343e-05, 'epoch': 0.07}) +[2026-06-17 06:33:19,634][starVLA.training.train_starvla][INFO] - Step 231, Loss: {'train/loss': 1.157051682472229, 'train/grad_norm_pre_clip': 8.375277519226074, 'timing/data': 0.005166057031601667, 'timing/model': 1.2048247429775074, 'learning_rate/qwen_vl_interface': 9.991553456648986e-06, 'learning_rate/action_model': 9.991553456648985e-05, 'epoch': 0.07}) +[2026-06-17 06:33:24,309][starVLA.training.train_starvla][INFO] - Step 232, Loss: {'train/loss': 1.1039913892745972, 'train/grad_norm_pre_clip': 7.915460586547852, 'timing/data': 0.004993456997908652, 'timing/model': 1.2051380060147494, 'learning_rate/qwen_vl_interface': 9.991424048560497e-06, 'learning_rate/action_model': 9.991424048560496e-05, 'epoch': 0.07}) +[2026-06-17 06:33:29,070][starVLA.training.train_starvla][INFO] - Step 233, Loss: {'train/loss': 1.0979053974151611, 'train/grad_norm_pre_clip': 5.812190055847168, 'timing/data': 0.0051103029400110245, 'timing/model': 1.2471385140670463, 'learning_rate/qwen_vl_interface': 9.991293657569698e-06, 'learning_rate/action_model': 9.991293657569698e-05, 'epoch': 0.07}) +[2026-06-17 06:33:33,807][starVLA.training.train_starvla][INFO] - Step 234, Loss: {'train/loss': 1.1424288749694824, 'train/grad_norm_pre_clip': 5.300462245941162, 'timing/data': 0.004944297950714827, 'timing/model': 1.2066935839829966, 'learning_rate/qwen_vl_interface': 9.991162283703624e-06, 'learning_rate/action_model': 9.991162283703624e-05, 'epoch': 0.07}) +[2026-06-17 06:33:38,611][starVLA.training.train_starvla][INFO] - Step 235, Loss: {'train/loss': 1.3985230922698975, 'train/grad_norm_pre_clip': 5.538122653961182, 'timing/data': 0.004869002033956349, 'timing/model': 1.3327526029897854, 'learning_rate/qwen_vl_interface': 9.991029926989506e-06, 'learning_rate/action_model': 9.991029926989506e-05, 'epoch': 0.07}) +[2026-06-17 06:33:43,277][starVLA.training.train_starvla][INFO] - Step 236, Loss: {'train/loss': 1.219693660736084, 'train/grad_norm_pre_clip': 8.073339462280273, 'timing/data': 0.004512366955168545, 'timing/model': 1.2069194880314171, 'learning_rate/qwen_vl_interface': 9.990896587454783e-06, 'learning_rate/action_model': 9.990896587454783e-05, 'epoch': 0.07}) +[2026-06-17 06:33:48,065][starVLA.training.train_starvla][INFO] - Step 237, Loss: {'train/loss': 1.3434526920318604, 'train/grad_norm_pre_clip': 7.527748107910156, 'timing/data': 0.004873925005085766, 'timing/model': 1.2085088830208406, 'learning_rate/qwen_vl_interface': 9.990762265127096e-06, 'learning_rate/action_model': 9.990762265127096e-05, 'epoch': 0.07}) +[2026-06-17 06:33:52,856][starVLA.training.train_starvla][INFO] - Step 238, Loss: {'train/loss': 1.3781250715255737, 'train/grad_norm_pre_clip': 10.324706077575684, 'timing/data': 0.004363557090982795, 'timing/model': 1.2063622989226133, 'learning_rate/qwen_vl_interface': 9.990626960034291e-06, 'learning_rate/action_model': 9.990626960034291e-05, 'epoch': 0.07}) +[2026-06-17 06:33:57,639][starVLA.training.train_starvla][INFO] - Step 239, Loss: {'train/loss': 1.1505545377731323, 'train/grad_norm_pre_clip': 8.290020942687988, 'timing/data': 0.0046280729584395885, 'timing/model': 1.207283868920058, 'learning_rate/qwen_vl_interface': 9.990490672204417e-06, 'learning_rate/action_model': 9.990490672204415e-05, 'epoch': 0.07}) +[2026-06-17 06:34:02,388][starVLA.training.train_starvla][INFO] - Step 240, Loss: {'train/loss': 1.3672617673873901, 'train/grad_norm_pre_clip': 5.550635814666748, 'timing/data': 0.005089842015877366, 'timing/model': 1.2075813100673258, 'learning_rate/qwen_vl_interface': 9.99035340166572e-06, 'learning_rate/action_model': 9.990353401665721e-05, 'epoch': 0.07}) +[2026-06-17 06:34:07,190][starVLA.training.train_starvla][INFO] - Step 241, Loss: {'train/loss': 1.3231251239776611, 'train/grad_norm_pre_clip': 8.860194206237793, 'timing/data': 0.004506147000938654, 'timing/model': 1.3462419700808823, 'learning_rate/qwen_vl_interface': 9.990215148446667e-06, 'learning_rate/action_model': 9.990215148446666e-05, 'epoch': 0.07}) +[2026-06-17 06:34:11,848][starVLA.training.train_starvla][INFO] - Step 242, Loss: {'train/loss': 1.0690016746520996, 'train/grad_norm_pre_clip': 6.467916488647461, 'timing/data': 0.00454481306951493, 'timing/model': 1.205961302970536, 'learning_rate/qwen_vl_interface': 9.990075912575912e-06, 'learning_rate/action_model': 9.990075912575911e-05, 'epoch': 0.07}) +[2026-06-17 06:34:16,639][starVLA.training.train_starvla][INFO] - Step 243, Loss: {'train/loss': 1.193244457244873, 'train/grad_norm_pre_clip': 13.943153381347656, 'timing/data': 0.004680244950577617, 'timing/model': 1.2073868579464033, 'learning_rate/qwen_vl_interface': 9.98993569408232e-06, 'learning_rate/action_model': 9.989935694082321e-05, 'epoch': 0.07}) +[2026-06-17 06:34:21,429][starVLA.training.train_starvla][INFO] - Step 244, Loss: {'train/loss': 1.331485629081726, 'train/grad_norm_pre_clip': 9.597563743591309, 'timing/data': 0.00501284294296056, 'timing/model': 1.2063312419923022, 'learning_rate/qwen_vl_interface': 9.989794492994958e-06, 'learning_rate/action_model': 9.989794492994958e-05, 'epoch': 0.07}) +[2026-06-17 06:34:26,220][starVLA.training.train_starvla][INFO] - Step 245, Loss: {'train/loss': 1.1053400039672852, 'train/grad_norm_pre_clip': 6.378699779510498, 'timing/data': 0.004689669003710151, 'timing/model': 1.2062025839695707, 'learning_rate/qwen_vl_interface': 9.989652309343097e-06, 'learning_rate/action_model': 9.989652309343098e-05, 'epoch': 0.07}) +[2026-06-17 06:34:30,982][starVLA.training.train_starvla][INFO] - Step 246, Loss: {'train/loss': 1.1047532558441162, 'train/grad_norm_pre_clip': 8.167986869812012, 'timing/data': 0.00426511699333787, 'timing/model': 1.2078652760246769, 'learning_rate/qwen_vl_interface': 9.989509143156213e-06, 'learning_rate/action_model': 9.989509143156213e-05, 'epoch': 0.07}) +[2026-06-17 06:34:35,780][starVLA.training.train_starvla][INFO] - Step 247, Loss: {'train/loss': 1.3174630403518677, 'train/grad_norm_pre_clip': 10.570189476013184, 'timing/data': 0.004542432026937604, 'timing/model': 1.3366532160434872, 'learning_rate/qwen_vl_interface': 9.989364994463983e-06, 'learning_rate/action_model': 9.989364994463983e-05, 'epoch': 0.07}) +[2026-06-17 06:34:40,457][starVLA.training.train_starvla][INFO] - Step 248, Loss: {'train/loss': 1.294090747833252, 'train/grad_norm_pre_clip': 6.421431064605713, 'timing/data': 0.004935912904329598, 'timing/model': 1.2089474269887432, 'learning_rate/qwen_vl_interface': 9.989219863296291e-06, 'learning_rate/action_model': 9.989219863296291e-05, 'epoch': 0.07}) +[2026-06-17 06:34:45,260][starVLA.training.train_starvla][INFO] - Step 249, Loss: {'train/loss': 1.1223586797714233, 'train/grad_norm_pre_clip': 5.236485481262207, 'timing/data': 0.004684392013587058, 'timing/model': 1.2078034640289843, 'learning_rate/qwen_vl_interface': 9.98907374968322e-06, 'learning_rate/action_model': 9.989073749683219e-05, 'epoch': 0.07}) +[2026-06-17 06:34:50,048][starVLA.training.train_starvla][INFO] - Step 250, Loss: {'train/loss': 1.7587451934814453, 'train/grad_norm_pre_clip': 12.44000244140625, 'timing/data': 0.004653798998333514, 'timing/model': 1.2075262899743393, 'learning_rate/qwen_vl_interface': 9.988926653655062e-06, 'learning_rate/action_model': 9.988926653655062e-05, 'epoch': 0.07}) +[2026-06-17 06:34:54,788][starVLA.training.train_starvla][INFO] - Step 251, Loss: {'train/loss': 1.1160415410995483, 'train/grad_norm_pre_clip': 6.003049850463867, 'timing/data': 0.005213066004216671, 'timing/model': 1.206740533001721, 'learning_rate/qwen_vl_interface': 9.988778575242309e-06, 'learning_rate/action_model': 9.988778575242309e-05, 'epoch': 0.07}) +[2026-06-17 06:34:59,580][starVLA.training.train_starvla][INFO] - Step 252, Loss: {'train/loss': 1.3299285173416138, 'train/grad_norm_pre_clip': 7.55756139755249, 'timing/data': 0.0045746860560029745, 'timing/model': 1.2069082059897482, 'learning_rate/qwen_vl_interface': 9.98862951447566e-06, 'learning_rate/action_model': 9.988629514475659e-05, 'epoch': 0.07}) +[2026-06-17 06:35:04,376][starVLA.training.train_starvla][INFO] - Step 253, Loss: {'train/loss': 1.2873890399932861, 'train/grad_norm_pre_clip': 7.250579357147217, 'timing/data': 0.00464651407673955, 'timing/model': 1.3353246320039034, 'learning_rate/qwen_vl_interface': 9.98847947138601e-06, 'learning_rate/action_model': 9.98847947138601e-05, 'epoch': 0.07}) +[2026-06-17 06:35:09,063][starVLA.training.train_starvla][INFO] - Step 254, Loss: {'train/loss': 1.3387707471847534, 'train/grad_norm_pre_clip': 10.937705039978027, 'timing/data': 0.004669466055929661, 'timing/model': 1.2156309050042182, 'learning_rate/qwen_vl_interface': 9.98832844600447e-06, 'learning_rate/action_model': 9.98832844600447e-05, 'epoch': 0.07}) +[2026-06-17 06:35:13,880][starVLA.training.train_starvla][INFO] - Step 255, Loss: {'train/loss': 1.237831711769104, 'train/grad_norm_pre_clip': 7.473071098327637, 'timing/data': 0.006912106997333467, 'timing/model': 1.225126325036399, 'learning_rate/qwen_vl_interface': 9.988176438362344e-06, 'learning_rate/action_model': 9.988176438362343e-05, 'epoch': 0.08}) +[2026-06-17 06:35:18,659][starVLA.training.train_starvla][INFO] - Step 256, Loss: {'train/loss': 1.1514050960540771, 'train/grad_norm_pre_clip': 6.026932239532471, 'timing/data': 0.005152355995960534, 'timing/model': 1.21526334004011, 'learning_rate/qwen_vl_interface': 9.988023448491143e-06, 'learning_rate/action_model': 9.988023448491143e-05, 'epoch': 0.08}) +[2026-06-17 06:35:23,443][starVLA.training.train_starvla][INFO] - Step 257, Loss: {'train/loss': 1.2061796188354492, 'train/grad_norm_pre_clip': 5.513364791870117, 'timing/data': 0.005002524005249143, 'timing/model': 1.2093679259996861, 'learning_rate/qwen_vl_interface': 9.987869476422582e-06, 'learning_rate/action_model': 9.987869476422583e-05, 'epoch': 0.08}) +[2026-06-17 06:35:28,268][starVLA.training.train_starvla][INFO] - Step 258, Loss: {'train/loss': 1.4639283418655396, 'train/grad_norm_pre_clip': 4.868852138519287, 'timing/data': 0.0049471009988337755, 'timing/model': 1.2151658879593015, 'learning_rate/qwen_vl_interface': 9.987714522188583e-06, 'learning_rate/action_model': 9.987714522188582e-05, 'epoch': 0.08}) +[2026-06-17 06:35:33,099][starVLA.training.train_starvla][INFO] - Step 259, Loss: {'train/loss': 1.219844937324524, 'train/grad_norm_pre_clip': 7.959705352783203, 'timing/data': 0.0050758559955284, 'timing/model': 1.3562301259953529, 'learning_rate/qwen_vl_interface': 9.987558585821264e-06, 'learning_rate/action_model': 9.987558585821263e-05, 'epoch': 0.08}) +[2026-06-17 06:35:37,796][starVLA.training.train_starvla][INFO] - Step 260, Loss: {'train/loss': 1.161290168762207, 'train/grad_norm_pre_clip': 8.326601028442383, 'timing/data': 0.004935830016620457, 'timing/model': 1.206249888986349, 'learning_rate/qwen_vl_interface': 9.987401667352952e-06, 'learning_rate/action_model': 9.987401667352951e-05, 'epoch': 0.08}) +[2026-06-17 06:35:42,595][starVLA.training.train_starvla][INFO] - Step 261, Loss: {'train/loss': 1.2585433721542358, 'train/grad_norm_pre_clip': 7.279851913452148, 'timing/data': 0.005205987021327019, 'timing/model': 1.2089365379652008, 'learning_rate/qwen_vl_interface': 9.987243766816175e-06, 'learning_rate/action_model': 9.987243766816176e-05, 'epoch': 0.08}) +[2026-06-17 06:35:47,397][starVLA.training.train_starvla][INFO] - Step 262, Loss: {'train/loss': 1.2449228763580322, 'train/grad_norm_pre_clip': 7.784147262573242, 'timing/data': 0.004778378992341459, 'timing/model': 1.2065405689645559, 'learning_rate/qwen_vl_interface': 9.987084884243671e-06, 'learning_rate/action_model': 9.987084884243672e-05, 'epoch': 0.08}) +[2026-06-17 06:35:52,151][starVLA.training.train_starvla][INFO] - Step 263, Loss: {'train/loss': 1.3360048532485962, 'train/grad_norm_pre_clip': 10.500532150268555, 'timing/data': 0.0046543789794668555, 'timing/model': 1.2068284369306639, 'learning_rate/qwen_vl_interface': 9.986925019668374e-06, 'learning_rate/action_model': 9.986925019668373e-05, 'epoch': 0.08}) +[2026-06-17 06:35:56,945][starVLA.training.train_starvla][INFO] - Step 264, Loss: {'train/loss': 1.1926085948944092, 'train/grad_norm_pre_clip': 6.582650184631348, 'timing/data': 0.00470135803334415, 'timing/model': 1.2073271579574794, 'learning_rate/qwen_vl_interface': 9.98676417312342e-06, 'learning_rate/action_model': 9.986764173123419e-05, 'epoch': 0.08}) +[2026-06-17 06:36:01,728][starVLA.training.train_starvla][INFO] - Step 265, Loss: {'train/loss': 1.2524372339248657, 'train/grad_norm_pre_clip': 6.915437698364258, 'timing/data': 0.0048027579905465245, 'timing/model': 1.3312376589747146, 'learning_rate/qwen_vl_interface': 9.986602344642157e-06, 'learning_rate/action_model': 9.986602344642156e-05, 'epoch': 0.08}) +[2026-06-17 06:36:06,406][starVLA.training.train_starvla][INFO] - Step 266, Loss: {'train/loss': 1.1652181148529053, 'train/grad_norm_pre_clip': 6.331781387329102, 'timing/data': 0.004825967946089804, 'timing/model': 1.2043738539796323, 'learning_rate/qwen_vl_interface': 9.986439534258132e-06, 'learning_rate/action_model': 9.986439534258131e-05, 'epoch': 0.08}) +[2026-06-17 06:36:11,193][starVLA.training.train_starvla][INFO] - Step 267, Loss: {'train/loss': 1.2175253629684448, 'train/grad_norm_pre_clip': 7.417968273162842, 'timing/data': 0.004275304032489657, 'timing/model': 1.210582587053068, 'learning_rate/qwen_vl_interface': 9.986275742005093e-06, 'learning_rate/action_model': 9.986275742005093e-05, 'epoch': 0.08}) +[2026-06-17 06:36:16,066][starVLA.training.train_starvla][INFO] - Step 268, Loss: {'train/loss': 1.6155322790145874, 'train/grad_norm_pre_clip': 12.548137664794922, 'timing/data': 0.004794526030309498, 'timing/model': 1.2093254129868, 'learning_rate/qwen_vl_interface': 9.986110967916997e-06, 'learning_rate/action_model': 9.986110967916996e-05, 'epoch': 0.08}) +[2026-06-17 06:36:20,873][starVLA.training.train_starvla][INFO] - Step 269, Loss: {'train/loss': 1.2656521797180176, 'train/grad_norm_pre_clip': 7.080617904663086, 'timing/data': 0.004521608934737742, 'timing/model': 1.2081240970874205, 'learning_rate/qwen_vl_interface': 9.985945212028002e-06, 'learning_rate/action_model': 9.985945212028001e-05, 'epoch': 0.08}) +[2026-06-17 06:36:25,669][starVLA.training.train_starvla][INFO] - Step 270, Loss: {'train/loss': 1.5174651145935059, 'train/grad_norm_pre_clip': 11.680587768554688, 'timing/data': 0.004719997989013791, 'timing/model': 1.2070637670112774, 'learning_rate/qwen_vl_interface': 9.985778474372467e-06, 'learning_rate/action_model': 9.985778474372466e-05, 'epoch': 0.08}) +[2026-06-17 06:36:30,462][starVLA.training.train_starvla][INFO] - Step 271, Loss: {'train/loss': 1.1277931928634644, 'train/grad_norm_pre_clip': 6.399317741394043, 'timing/data': 0.00457230000756681, 'timing/model': 1.3317670159740373, 'learning_rate/qwen_vl_interface': 9.98561075498496e-06, 'learning_rate/action_model': 9.98561075498496e-05, 'epoch': 0.08}) +[2026-06-17 06:36:35,172][starVLA.training.train_starvla][INFO] - Step 272, Loss: {'train/loss': 1.2364599704742432, 'train/grad_norm_pre_clip': 8.304000854492188, 'timing/data': 0.004572337958961725, 'timing/model': 1.2528055059956387, 'learning_rate/qwen_vl_interface': 9.985442053900246e-06, 'learning_rate/action_model': 9.985442053900246e-05, 'epoch': 0.08}) +[2026-06-17 06:36:39,857][starVLA.training.train_starvla][INFO] - Step 273, Loss: {'train/loss': 1.1405619382858276, 'train/grad_norm_pre_clip': 8.22479248046875, 'timing/data': 0.004733690991997719, 'timing/model': 1.2130542800296098, 'learning_rate/qwen_vl_interface': 9.9852723711533e-06, 'learning_rate/action_model': 9.9852723711533e-05, 'epoch': 0.08}) +[2026-06-17 06:36:44,613][starVLA.training.train_starvla][INFO] - Step 274, Loss: {'train/loss': 1.5249427556991577, 'train/grad_norm_pre_clip': 15.344060897827148, 'timing/data': 0.004866115981712937, 'timing/model': 1.2060799519531429, 'learning_rate/qwen_vl_interface': 9.985101706779297e-06, 'learning_rate/action_model': 9.985101706779296e-05, 'epoch': 0.08}) +[2026-06-17 06:36:49,426][starVLA.training.train_starvla][INFO] - Step 275, Loss: {'train/loss': 1.3624553680419922, 'train/grad_norm_pre_clip': 10.157423973083496, 'timing/data': 0.004956981050781906, 'timing/model': 1.207515540998429, 'learning_rate/qwen_vl_interface': 9.984930060813613e-06, 'learning_rate/action_model': 9.984930060813613e-05, 'epoch': 0.08}) +[2026-06-17 06:36:54,222][starVLA.training.train_starvla][INFO] - Step 276, Loss: {'train/loss': 1.4878833293914795, 'train/grad_norm_pre_clip': 16.368558883666992, 'timing/data': 0.004568494041450322, 'timing/model': 1.207965765031986, 'learning_rate/qwen_vl_interface': 9.984757433291832e-06, 'learning_rate/action_model': 9.984757433291832e-05, 'epoch': 0.08}) +[2026-06-17 06:36:59,022][starVLA.training.train_starvla][INFO] - Step 277, Loss: {'train/loss': 1.186783790588379, 'train/grad_norm_pre_clip': 5.481582164764404, 'timing/data': 0.00500045099761337, 'timing/model': 1.3391661220230162, 'learning_rate/qwen_vl_interface': 9.984583824249742e-06, 'learning_rate/action_model': 9.984583824249742e-05, 'epoch': 0.08}) +[2026-06-17 06:37:03,805][starVLA.training.train_starvla][INFO] - Step 278, Loss: {'train/loss': 1.17607581615448, 'train/grad_norm_pre_clip': 6.818796634674072, 'timing/data': 0.0048048620810732245, 'timing/model': 1.3291101539507508, 'learning_rate/qwen_vl_interface': 9.98440923372333e-06, 'learning_rate/action_model': 9.984409233723328e-05, 'epoch': 0.08}) +[2026-06-17 06:37:08,464][starVLA.training.train_starvla][INFO] - Step 279, Loss: {'train/loss': 1.175194263458252, 'train/grad_norm_pre_clip': 6.022982597351074, 'timing/data': 0.004739596042782068, 'timing/model': 1.2052867430029437, 'learning_rate/qwen_vl_interface': 9.984233661748788e-06, 'learning_rate/action_model': 9.984233661748788e-05, 'epoch': 0.08}) +[2026-06-17 06:37:13,261][starVLA.training.train_starvla][INFO] - Step 280, Loss: {'train/loss': 1.1011292934417725, 'train/grad_norm_pre_clip': 6.185116291046143, 'timing/data': 0.005351946922019124, 'timing/model': 1.2061902659479529, 'learning_rate/qwen_vl_interface': 9.984057108362516e-06, 'learning_rate/action_model': 9.984057108362515e-05, 'epoch': 0.08}) +[2026-06-17 06:37:18,046][starVLA.training.train_starvla][INFO] - Step 281, Loss: {'train/loss': 1.2775261402130127, 'train/grad_norm_pre_clip': 7.366758346557617, 'timing/data': 0.004902685061097145, 'timing/model': 1.2035943319788203, 'learning_rate/qwen_vl_interface': 9.98387957360111e-06, 'learning_rate/action_model': 9.98387957360111e-05, 'epoch': 0.08}) +[2026-06-17 06:37:22,835][starVLA.training.train_starvla][INFO] - Step 282, Loss: {'train/loss': 1.2501592636108398, 'train/grad_norm_pre_clip': 8.4271879196167, 'timing/data': 0.00492614193353802, 'timing/model': 1.2060250060167164, 'learning_rate/qwen_vl_interface': 9.983701057501375e-06, 'learning_rate/action_model': 9.983701057501375e-05, 'epoch': 0.08}) +[2026-06-17 06:37:27,621][starVLA.training.train_starvla][INFO] - Step 283, Loss: {'train/loss': 1.100351095199585, 'train/grad_norm_pre_clip': 4.909266948699951, 'timing/data': 0.004771266016177833, 'timing/model': 1.3333918650168926, 'learning_rate/qwen_vl_interface': 9.983521560100317e-06, 'learning_rate/action_model': 9.983521560100316e-05, 'epoch': 0.08}) +[2026-06-17 06:37:32,288][starVLA.training.train_starvla][INFO] - Step 284, Loss: {'train/loss': 1.0944675207138062, 'train/grad_norm_pre_clip': 5.356212139129639, 'timing/data': 0.005277886986732483, 'timing/model': 1.2069874029839411, 'learning_rate/qwen_vl_interface': 9.983341081435145e-06, 'learning_rate/action_model': 9.983341081435145e-05, 'epoch': 0.08}) +[2026-06-17 06:37:37,088][starVLA.training.train_starvla][INFO] - Step 285, Loss: {'train/loss': 1.141471028327942, 'train/grad_norm_pre_clip': 10.215046882629395, 'timing/data': 0.004977395990863442, 'timing/model': 1.2073761080391705, 'learning_rate/qwen_vl_interface': 9.983159621543275e-06, 'learning_rate/action_model': 9.983159621543275e-05, 'epoch': 0.08}) +[2026-06-17 06:37:41,878][starVLA.training.train_starvla][INFO] - Step 286, Loss: {'train/loss': 1.2051286697387695, 'train/grad_norm_pre_clip': 4.741508483886719, 'timing/data': 0.00538485893048346, 'timing/model': 1.2078382200561464, 'learning_rate/qwen_vl_interface': 9.982977180462324e-06, 'learning_rate/action_model': 9.982977180462323e-05, 'epoch': 0.08}) +[2026-06-17 06:37:46,722][starVLA.training.train_starvla][INFO] - Step 287, Loss: {'train/loss': 0.8711421489715576, 'train/grad_norm_pre_clip': 5.841929912567139, 'timing/data': 0.004694992094300687, 'timing/model': 1.2547763569746166, 'learning_rate/qwen_vl_interface': 9.982793758230107e-06, 'learning_rate/action_model': 9.982793758230108e-05, 'epoch': 0.08}) +[2026-06-17 06:37:51,443][starVLA.training.train_starvla][INFO] - Step 288, Loss: {'train/loss': 1.1810452938079834, 'train/grad_norm_pre_clip': 6.639911651611328, 'timing/data': 0.004391482099890709, 'timing/model': 1.2116185890045017, 'learning_rate/qwen_vl_interface': 9.982609354884653e-06, 'learning_rate/action_model': 9.982609354884653e-05, 'epoch': 0.08}) +[2026-06-17 06:37:56,253][starVLA.training.train_starvla][INFO] - Step 289, Loss: {'train/loss': 1.227410078048706, 'train/grad_norm_pre_clip': 9.383049011230469, 'timing/data': 0.004620071034878492, 'timing/model': 1.3487539030611515, 'learning_rate/qwen_vl_interface': 9.982423970464187e-06, 'learning_rate/action_model': 9.982423970464186e-05, 'epoch': 0.09}) +[2026-06-17 06:38:00,916][starVLA.training.train_starvla][INFO] - Step 290, Loss: {'train/loss': 1.1113439798355103, 'train/grad_norm_pre_clip': 6.8739495277404785, 'timing/data': 0.004720509052276611, 'timing/model': 1.20756527595222, 'learning_rate/qwen_vl_interface': 9.98223760500714e-06, 'learning_rate/action_model': 9.982237605007139e-05, 'epoch': 0.09}) +[2026-06-17 06:38:05,725][starVLA.training.train_starvla][INFO] - Step 291, Loss: {'train/loss': 1.1911321878433228, 'train/grad_norm_pre_clip': 8.231680870056152, 'timing/data': 0.004728545085527003, 'timing/model': 1.2110137370182201, 'learning_rate/qwen_vl_interface': 9.982050258552146e-06, 'learning_rate/action_model': 9.982050258552145e-05, 'epoch': 0.09}) +[2026-06-17 06:38:10,526][starVLA.training.train_starvla][INFO] - Step 292, Loss: {'train/loss': 1.0703107118606567, 'train/grad_norm_pre_clip': 6.536579132080078, 'timing/data': 0.004518078989349306, 'timing/model': 1.2080980989849195, 'learning_rate/qwen_vl_interface': 9.981861931138039e-06, 'learning_rate/action_model': 9.981861931138039e-05, 'epoch': 0.09}) +[2026-06-17 06:38:15,318][starVLA.training.train_starvla][INFO] - Step 293, Loss: {'train/loss': 0.9176262021064758, 'train/grad_norm_pre_clip': 9.26313591003418, 'timing/data': 0.004684176994487643, 'timing/model': 1.207281193928793, 'learning_rate/qwen_vl_interface': 9.981672622803863e-06, 'learning_rate/action_model': 9.981672622803862e-05, 'epoch': 0.09}) +[2026-06-17 06:38:20,104][starVLA.training.train_starvla][INFO] - Step 294, Loss: {'train/loss': 1.3694766759872437, 'train/grad_norm_pre_clip': 10.065855979919434, 'timing/data': 0.004660074017010629, 'timing/model': 1.2080151119735092, 'learning_rate/qwen_vl_interface': 9.981482333588863e-06, 'learning_rate/action_model': 9.981482333588862e-05, 'epoch': 0.09}) +[2026-06-17 06:38:24,854][starVLA.training.train_starvla][INFO] - Step 295, Loss: {'train/loss': 1.1513229608535767, 'train/grad_norm_pre_clip': 6.087423324584961, 'timing/data': 0.0051364380633458495, 'timing/model': 1.2921140739927068, 'learning_rate/qwen_vl_interface': 9.981291063532481e-06, 'learning_rate/action_model': 9.981291063532481e-05, 'epoch': 0.09}) +[2026-06-17 06:38:29,529][starVLA.training.train_starvla][INFO] - Step 296, Loss: {'train/loss': 1.6573894023895264, 'train/grad_norm_pre_clip': 9.776368141174316, 'timing/data': 0.004693615948781371, 'timing/model': 1.2138274629833177, 'learning_rate/qwen_vl_interface': 9.981098812674372e-06, 'learning_rate/action_model': 9.981098812674371e-05, 'epoch': 0.09}) +[2026-06-17 06:38:34,317][starVLA.training.train_starvla][INFO] - Step 297, Loss: {'train/loss': 1.3454300165176392, 'train/grad_norm_pre_clip': 8.864301681518555, 'timing/data': 0.005042406963184476, 'timing/model': 1.2092225609812886, 'learning_rate/qwen_vl_interface': 9.980905581054388e-06, 'learning_rate/action_model': 9.980905581054387e-05, 'epoch': 0.09}) +[2026-06-17 06:38:39,136][starVLA.training.train_starvla][INFO] - Step 298, Loss: {'train/loss': 1.0168370008468628, 'train/grad_norm_pre_clip': 17.366931915283203, 'timing/data': 0.004618076025508344, 'timing/model': 1.2168654520064592, 'learning_rate/qwen_vl_interface': 9.980711368712586e-06, 'learning_rate/action_model': 9.980711368712585e-05, 'epoch': 0.09}) +[2026-06-17 06:38:43,940][starVLA.training.train_starvla][INFO] - Step 299, Loss: {'train/loss': 1.3932288885116577, 'train/grad_norm_pre_clip': 8.30005168914795, 'timing/data': 0.0044949790462851524, 'timing/model': 1.2147101019509137, 'learning_rate/qwen_vl_interface': 9.980516175689228e-06, 'learning_rate/action_model': 9.980516175689227e-05, 'epoch': 0.09}) +[2026-06-17 06:40:26,040][starVLA.training.train_starvla][INFO] - Step 300, Loss: {'train/loss': 1.1309478282928467, 'train/grad_norm_pre_clip': 6.908997058868408, 'eval/loss': 1.1710530519485474, 'eval/latency_0/loss': 1.1710530519485474, 'eval/demon_fire_precision': 0.8764091858037578, 'eval/demon_fire_recall': 0.8716777408637874, 'eval/demon_fire_f1': 0.8740370601707267, 'eval/demon_fire_rate_model': 0.6660177975528365, 'eval/demon_fire_rate_teacher': 0.6696329254727474, 'eval/demon_left_precision': 0.611671785870996, 'eval/demon_left_recall': 0.7898016997167139, 'eval/demon_left_f1': 0.6894164193867458, 'eval/demon_left_rate_model': 0.6337597330367074, 'eval/demon_left_rate_teacher': 0.49082313681868744, 'eval/demon_right_precision': 0.6795747911921033, 'eval/demon_right_recall': 0.5056497175141242, 'eval/demon_right_f1': 0.5798509880142532, 'eval/demon_right_rate_model': 0.36624026696329254, 'eval/demon_right_rate_teacher': 0.4922135706340378, 'eval/demon_move_f1': 0.6346337037004994, 'eval/demon_cc_f1': 0.7851789623766529, 'eval/latency_0/demon_fire_precision': 0.8764091858037578, 'eval/latency_0/demon_fire_recall': 0.8716777408637874, 'eval/latency_0/demon_fire_f1': 0.8740370601707267, 'eval/latency_0/demon_fire_rate_model': 0.6660177975528365, 'eval/latency_0/demon_fire_rate_teacher': 0.6696329254727474, 'eval/latency_0/demon_left_precision': 0.611671785870996, 'eval/latency_0/demon_left_recall': 0.7898016997167139, 'eval/latency_0/demon_left_f1': 0.6894164193867458, 'eval/latency_0/demon_left_rate_model': 0.6337597330367074, 'eval/latency_0/demon_left_rate_teacher': 0.49082313681868744, 'eval/latency_0/demon_right_precision': 0.6795747911921033, 'eval/latency_0/demon_right_recall': 0.5056497175141242, 'eval/latency_0/demon_right_f1': 0.5798509880142532, 'eval/latency_0/demon_right_rate_model': 0.36624026696329254, 'eval/latency_0/demon_right_rate_teacher': 0.4922135706340378, 'eval/latency_0/demon_move_f1': 0.6346337037004994, 'eval/latency_0/demon_cc_f1': 0.7851789623766529, 'timing/data': 0.004637095029465854, 'timing/model': 1.350646825041622, 'learning_rate/qwen_vl_interface': 9.980320002024774e-06, 'learning_rate/action_model': 9.980320002024774e-05, 'epoch': 0.09}) +[2026-06-17 06:40:26,041][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_300_state +[2026-06-17 06:40:26,044][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 06:40:29,570][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_300_state/model.safetensors +[2026-06-17 06:40:37,448][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_300_state/optimizer.bin +[2026-06-17 06:40:37,449][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_300_state/sampler.bin +[2026-06-17 06:40:37,449][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_300_state/sampler_1.bin +[2026-06-17 06:40:37,450][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_300_state/random_states_0.pkl +[2026-06-17 06:40:37,482][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 300 +[2026-06-17 06:40:37,482][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 06:40:37,491][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 06:40:42,161][starVLA.training.train_starvla][INFO] - Step 301, Loss: {'train/loss': 1.2410383224487305, 'train/grad_norm_pre_clip': 9.553288459777832, 'timing/data': 0.005064644035883248, 'timing/model': 1.1964441880118102, 'learning_rate/qwen_vl_interface': 9.980122847759896e-06, 'learning_rate/action_model': 9.980122847759896e-05, 'epoch': 0.09}) +[2026-06-17 06:40:46,873][starVLA.training.train_starvla][INFO] - Step 302, Loss: {'train/loss': 1.278980016708374, 'train/grad_norm_pre_clip': 8.437713623046875, 'timing/data': 0.004861297085881233, 'timing/model': 1.2991242839489132, 'learning_rate/qwen_vl_interface': 9.97992471293546e-06, 'learning_rate/action_model': 9.97992471293546e-05, 'epoch': 0.09}) +[2026-06-17 06:40:51,496][starVLA.training.train_starvla][INFO] - Step 303, Loss: {'train/loss': 1.280626654624939, 'train/grad_norm_pre_clip': 7.466181755065918, 'timing/data': 0.004833047976717353, 'timing/model': 1.2000189939280972, 'learning_rate/qwen_vl_interface': 9.979725597592543e-06, 'learning_rate/action_model': 9.979725597592541e-05, 'epoch': 0.09}) +[2026-06-17 06:40:56,242][starVLA.training.train_starvla][INFO] - Step 304, Loss: {'train/loss': 1.2996103763580322, 'train/grad_norm_pre_clip': 13.284299850463867, 'timing/data': 0.005084407050162554, 'timing/model': 1.1991394279757515, 'learning_rate/qwen_vl_interface': 9.979525501772418e-06, 'learning_rate/action_model': 9.979525501772417e-05, 'epoch': 0.09}) +[2026-06-17 06:41:00,995][starVLA.training.train_starvla][INFO] - Step 305, Loss: {'train/loss': 1.2942670583724976, 'train/grad_norm_pre_clip': 12.249183654785156, 'timing/data': 0.005164511036127806, 'timing/model': 1.2036444830009714, 'learning_rate/qwen_vl_interface': 9.97932442551657e-06, 'learning_rate/action_model': 9.979324425516569e-05, 'epoch': 0.09}) +[2026-06-17 06:41:05,763][starVLA.training.train_starvla][INFO] - Step 306, Loss: {'train/loss': 1.1294599771499634, 'train/grad_norm_pre_clip': 6.629249095916748, 'timing/data': 0.0050900839269161224, 'timing/model': 1.1970901420572773, 'learning_rate/qwen_vl_interface': 9.979122368866678e-06, 'learning_rate/action_model': 9.979122368866679e-05, 'epoch': 0.09}) +[2026-06-17 06:41:10,469][starVLA.training.train_starvla][INFO] - Step 307, Loss: {'train/loss': 1.389185905456543, 'train/grad_norm_pre_clip': 12.671422004699707, 'timing/data': 0.004570464952848852, 'timing/model': 1.200026259990409, 'learning_rate/qwen_vl_interface': 9.978919331864631e-06, 'learning_rate/action_model': 9.97891933186463e-05, 'epoch': 0.09}) +[2026-06-17 06:41:15,237][starVLA.training.train_starvla][INFO] - Step 308, Loss: {'train/loss': 1.3752751350402832, 'train/grad_norm_pre_clip': 15.099003791809082, 'timing/data': 0.0049724510172382, 'timing/model': 1.3308957499684766, 'learning_rate/qwen_vl_interface': 9.978715314552518e-06, 'learning_rate/action_model': 9.978715314552518e-05, 'epoch': 0.09}) +[2026-06-17 06:41:19,873][starVLA.training.train_starvla][INFO] - Step 309, Loss: {'train/loss': 1.3563092947006226, 'train/grad_norm_pre_clip': 11.13310432434082, 'timing/data': 0.0049156820168718696, 'timing/model': 1.198224324034527, 'learning_rate/qwen_vl_interface': 9.978510316972631e-06, 'learning_rate/action_model': 9.978510316972631e-05, 'epoch': 0.09}) +[2026-06-17 06:41:24,648][starVLA.training.train_starvla][INFO] - Step 310, Loss: {'train/loss': 1.126779317855835, 'train/grad_norm_pre_clip': 5.096144199371338, 'timing/data': 0.005267639993689954, 'timing/model': 1.205534512992017, 'learning_rate/qwen_vl_interface': 9.978304339167467e-06, 'learning_rate/action_model': 9.978304339167467e-05, 'epoch': 0.09}) +[2026-06-17 06:41:29,418][starVLA.training.train_starvla][INFO] - Step 311, Loss: {'train/loss': 1.2419869899749756, 'train/grad_norm_pre_clip': 6.101656436920166, 'timing/data': 0.004462752025574446, 'timing/model': 1.199400977930054, 'learning_rate/qwen_vl_interface': 9.978097381179726e-06, 'learning_rate/action_model': 9.978097381179726e-05, 'epoch': 0.09}) +[2026-06-17 06:41:34,250][starVLA.training.train_starvla][INFO] - Step 312, Loss: {'train/loss': 1.0781055688858032, 'train/grad_norm_pre_clip': 6.2691755294799805, 'timing/data': 0.0049026289489120245, 'timing/model': 1.2703580759698525, 'learning_rate/qwen_vl_interface': 9.977889443052312e-06, 'learning_rate/action_model': 9.977889443052311e-05, 'epoch': 0.09}) +[2026-06-17 06:41:39,053][starVLA.training.train_starvla][INFO] - Step 313, Loss: {'train/loss': 1.2427637577056885, 'train/grad_norm_pre_clip': 9.200340270996094, 'timing/data': 0.0050171640468761325, 'timing/model': 1.203071920084767, 'learning_rate/qwen_vl_interface': 9.977680524828327e-06, 'learning_rate/action_model': 9.977680524828327e-05, 'epoch': 0.09}) +[2026-06-17 06:41:43,834][starVLA.training.train_starvla][INFO] - Step 314, Loss: {'train/loss': 1.0832018852233887, 'train/grad_norm_pre_clip': 12.788369178771973, 'timing/data': 0.004927536007016897, 'timing/model': 1.3291733029764146, 'learning_rate/qwen_vl_interface': 9.977470626551083e-06, 'learning_rate/action_model': 9.977470626551083e-05, 'epoch': 0.09}) +[2026-06-17 06:41:48,485][starVLA.training.train_starvla][INFO] - Step 315, Loss: {'train/loss': 1.1918731927871704, 'train/grad_norm_pre_clip': 15.171224594116211, 'timing/data': 0.005037199007347226, 'timing/model': 1.2021230790996924, 'learning_rate/qwen_vl_interface': 9.977259748264089e-06, 'learning_rate/action_model': 9.97725974826409e-05, 'epoch': 0.09}) +[2026-06-17 06:41:53,264][starVLA.training.train_starvla][INFO] - Step 316, Loss: {'train/loss': 1.2803977727890015, 'train/grad_norm_pre_clip': 12.38052749633789, 'timing/data': 0.004918167018331587, 'timing/model': 1.2073068199679255, 'learning_rate/qwen_vl_interface': 9.977047890011065e-06, 'learning_rate/action_model': 9.977047890011065e-05, 'epoch': 0.09}) +[2026-06-17 06:41:58,059][starVLA.training.train_starvla][INFO] - Step 317, Loss: {'train/loss': 1.132411003112793, 'train/grad_norm_pre_clip': 10.330419540405273, 'timing/data': 0.004838526016101241, 'timing/model': 1.2034827070310712, 'learning_rate/qwen_vl_interface': 9.976835051835928e-06, 'learning_rate/action_model': 9.976835051835926e-05, 'epoch': 0.09}) +[2026-06-17 06:42:02,864][starVLA.training.train_starvla][INFO] - Step 318, Loss: {'train/loss': 1.2370784282684326, 'train/grad_norm_pre_clip': 25.8045711517334, 'timing/data': 0.004992441041395068, 'timing/model': 1.2042340269545093, 'learning_rate/qwen_vl_interface': 9.976621233782795e-06, 'learning_rate/action_model': 9.976621233782795e-05, 'epoch': 0.09}) +[2026-06-17 06:42:07,615][starVLA.training.train_starvla][INFO] - Step 319, Loss: {'train/loss': 1.2305337190628052, 'train/grad_norm_pre_clip': 8.573630332946777, 'timing/data': 0.0049640260403975844, 'timing/model': 1.2045852270675823, 'learning_rate/qwen_vl_interface': 9.976406435895995e-06, 'learning_rate/action_model': 9.976406435895995e-05, 'epoch': 0.09}) +[2026-06-17 06:42:12,395][starVLA.training.train_starvla][INFO] - Step 320, Loss: {'train/loss': 0.9571241736412048, 'train/grad_norm_pre_clip': 8.48023796081543, 'timing/data': 0.004904004978016019, 'timing/model': 1.331815261975862, 'learning_rate/qwen_vl_interface': 9.976190658220056e-06, 'learning_rate/action_model': 9.976190658220056e-05, 'epoch': 0.09}) +[2026-06-17 06:42:17,050][starVLA.training.train_starvla][INFO] - Step 321, Loss: {'train/loss': 1.18491792678833, 'train/grad_norm_pre_clip': 9.964686393737793, 'timing/data': 0.005068718106485903, 'timing/model': 1.2053008859511465, 'learning_rate/qwen_vl_interface': 9.975973900799709e-06, 'learning_rate/action_model': 9.975973900799708e-05, 'epoch': 0.09}) +[2026-06-17 06:42:21,848][starVLA.training.train_starvla][INFO] - Step 322, Loss: {'train/loss': 1.2517939805984497, 'train/grad_norm_pre_clip': 11.753874778747559, 'timing/data': 0.004796429071575403, 'timing/model': 1.2074851029319689, 'learning_rate/qwen_vl_interface': 9.975756163679884e-06, 'learning_rate/action_model': 9.975756163679885e-05, 'epoch': 0.09}) +[2026-06-17 06:42:26,633][starVLA.training.train_starvla][INFO] - Step 323, Loss: {'train/loss': 1.1156880855560303, 'train/grad_norm_pre_clip': 7.745456695556641, 'timing/data': 0.004842171096242964, 'timing/model': 1.2069089899305254, 'learning_rate/qwen_vl_interface': 9.975537446905722e-06, 'learning_rate/action_model': 9.975537446905723e-05, 'epoch': 0.1}) +[2026-06-17 06:42:31,371][starVLA.training.train_starvla][INFO] - Step 324, Loss: {'train/loss': 1.1242759227752686, 'train/grad_norm_pre_clip': 11.088654518127441, 'timing/data': 0.005076125031337142, 'timing/model': 1.20666710997466, 'learning_rate/qwen_vl_interface': 9.975317750522565e-06, 'learning_rate/action_model': 9.975317750522564e-05, 'epoch': 0.1}) +[2026-06-17 06:42:36,154][starVLA.training.train_starvla][INFO] - Step 325, Loss: {'train/loss': 1.249685287475586, 'train/grad_norm_pre_clip': 10.927783012390137, 'timing/data': 0.005128760007210076, 'timing/model': 1.2060496259946376, 'learning_rate/qwen_vl_interface': 9.97509707457595e-06, 'learning_rate/action_model': 9.97509707457595e-05, 'epoch': 0.1}) +[2026-06-17 06:42:40,965][starVLA.training.train_starvla][INFO] - Step 326, Loss: {'train/loss': 1.0784708261489868, 'train/grad_norm_pre_clip': 13.696396827697754, 'timing/data': 0.004980447003617883, 'timing/model': 1.3346732879290357, 'learning_rate/qwen_vl_interface': 9.974875419111631e-06, 'learning_rate/action_model': 9.97487541911163e-05, 'epoch': 0.1}) +[2026-06-17 06:42:45,626][starVLA.training.train_starvla][INFO] - Step 327, Loss: {'train/loss': 1.2751673460006714, 'train/grad_norm_pre_clip': 14.320773124694824, 'timing/data': 0.004891004064120352, 'timing/model': 1.2080711129819974, 'learning_rate/qwen_vl_interface': 9.974652784175552e-06, 'learning_rate/action_model': 9.974652784175551e-05, 'epoch': 0.1}) +[2026-06-17 06:42:50,414][starVLA.training.train_starvla][INFO] - Step 328, Loss: {'train/loss': 1.011894702911377, 'train/grad_norm_pre_clip': 15.568739891052246, 'timing/data': 0.005009663989767432, 'timing/model': 1.207141168997623, 'learning_rate/qwen_vl_interface': 9.974429169813865e-06, 'learning_rate/action_model': 9.974429169813864e-05, 'epoch': 0.1}) +[2026-06-17 06:42:55,143][starVLA.training.train_starvla][INFO] - Step 329, Loss: {'train/loss': 1.3079442977905273, 'train/grad_norm_pre_clip': 12.148482322692871, 'timing/data': 0.004964415915310383, 'timing/model': 1.2073803240200505, 'learning_rate/qwen_vl_interface': 9.974204576072928e-06, 'learning_rate/action_model': 9.974204576072928e-05, 'epoch': 0.1}) +[2026-06-17 06:42:59,890][starVLA.training.train_starvla][INFO] - Step 330, Loss: {'train/loss': 1.203843355178833, 'train/grad_norm_pre_clip': 10.104950904846191, 'timing/data': 0.004623422049917281, 'timing/model': 1.2069557189242914, 'learning_rate/qwen_vl_interface': 9.973979002999299e-06, 'learning_rate/action_model': 9.973979002999299e-05, 'epoch': 0.1}) +[2026-06-17 06:43:04,683][starVLA.training.train_starvla][INFO] - Step 331, Loss: {'train/loss': 1.1097451448440552, 'train/grad_norm_pre_clip': 8.917977333068848, 'timing/data': 0.004868759890086949, 'timing/model': 1.2067053379723802, 'learning_rate/qwen_vl_interface': 9.97375245063974e-06, 'learning_rate/action_model': 9.97375245063974e-05, 'epoch': 0.1}) +[2026-06-17 06:43:09,480][starVLA.training.train_starvla][INFO] - Step 332, Loss: {'train/loss': 1.0719410181045532, 'train/grad_norm_pre_clip': 8.277466773986816, 'timing/data': 0.004924794076941907, 'timing/model': 1.3335163909941912, 'learning_rate/qwen_vl_interface': 9.973524919041215e-06, 'learning_rate/action_model': 9.973524919041215e-05, 'epoch': 0.1}) +[2026-06-17 06:43:14,156][starVLA.training.train_starvla][INFO] - Step 333, Loss: {'train/loss': 1.2670824527740479, 'train/grad_norm_pre_clip': 10.316252708435059, 'timing/data': 0.005161736975423992, 'timing/model': 1.2071931150276214, 'learning_rate/qwen_vl_interface': 9.97329640825089e-06, 'learning_rate/action_model': 9.973296408250891e-05, 'epoch': 0.1}) +[2026-06-17 06:43:18,936][starVLA.training.train_starvla][INFO] - Step 334, Loss: {'train/loss': 1.3017888069152832, 'train/grad_norm_pre_clip': 9.79356575012207, 'timing/data': 0.0048708789981901646, 'timing/model': 1.2050680419197306, 'learning_rate/qwen_vl_interface': 9.97306691831614e-06, 'learning_rate/action_model': 9.97306691831614e-05, 'epoch': 0.1}) +[2026-06-17 06:43:23,715][starVLA.training.train_starvla][INFO] - Step 335, Loss: {'train/loss': 1.4181073904037476, 'train/grad_norm_pre_clip': 13.949965476989746, 'timing/data': 0.004978029988706112, 'timing/model': 1.205602355999872, 'learning_rate/qwen_vl_interface': 9.972836449284533e-06, 'learning_rate/action_model': 9.972836449284532e-05, 'epoch': 0.1}) +[2026-06-17 06:43:28,468][starVLA.training.train_starvla][INFO] - Step 336, Loss: {'train/loss': 1.2170323133468628, 'train/grad_norm_pre_clip': 9.847538948059082, 'timing/data': 0.004804819007404149, 'timing/model': 1.2091130840126425, 'learning_rate/qwen_vl_interface': 9.97260500120385e-06, 'learning_rate/action_model': 9.97260500120385e-05, 'epoch': 0.1}) +[2026-06-17 06:43:33,259][starVLA.training.train_starvla][INFO] - Step 337, Loss: {'train/loss': 1.2281571626663208, 'train/grad_norm_pre_clip': 14.377132415771484, 'timing/data': 0.005442641908302903, 'timing/model': 1.2077925169141963, 'learning_rate/qwen_vl_interface': 9.972372574122067e-06, 'learning_rate/action_model': 9.972372574122068e-05, 'epoch': 0.1}) +[2026-06-17 06:43:38,050][starVLA.training.train_starvla][INFO] - Step 338, Loss: {'train/loss': 1.2029685974121094, 'train/grad_norm_pre_clip': 5.184378623962402, 'timing/data': 0.0047348939115181565, 'timing/model': 1.3336643780348822, 'learning_rate/qwen_vl_interface': 9.972139168087368e-06, 'learning_rate/action_model': 9.972139168087369e-05, 'epoch': 0.1}) +[2026-06-17 06:43:42,718][starVLA.training.train_starvla][INFO] - Step 339, Loss: {'train/loss': 1.0546497106552124, 'train/grad_norm_pre_clip': 10.835695266723633, 'timing/data': 0.0049808359472081065, 'timing/model': 1.20767057093326, 'learning_rate/qwen_vl_interface': 9.97190478314814e-06, 'learning_rate/action_model': 9.97190478314814e-05, 'epoch': 0.1}) +[2026-06-17 06:43:47,585][starVLA.training.train_starvla][INFO] - Step 340, Loss: {'train/loss': 1.1553925275802612, 'train/grad_norm_pre_clip': 7.228399276733398, 'timing/data': 0.0049062310717999935, 'timing/model': 1.2110392400063574, 'learning_rate/qwen_vl_interface': 9.971669419352968e-06, 'learning_rate/action_model': 9.971669419352966e-05, 'epoch': 0.1}) +[2026-06-17 06:43:52,332][starVLA.training.train_starvla][INFO] - Step 341, Loss: {'train/loss': 1.0353634357452393, 'train/grad_norm_pre_clip': 17.16690444946289, 'timing/data': 0.004758487921208143, 'timing/model': 1.207777373958379, 'learning_rate/qwen_vl_interface': 9.971433076750644e-06, 'learning_rate/action_model': 9.971433076750643e-05, 'epoch': 0.1}) +[2026-06-17 06:43:57,145][starVLA.training.train_starvla][INFO] - Step 342, Loss: {'train/loss': 1.057621717453003, 'train/grad_norm_pre_clip': 9.462837219238281, 'timing/data': 0.004871188895776868, 'timing/model': 1.2142923680366948, 'learning_rate/qwen_vl_interface': 9.971195755390166e-06, 'learning_rate/action_model': 9.971195755390164e-05, 'epoch': 0.1}) +[2026-06-17 06:44:01,935][starVLA.training.train_starvla][INFO] - Step 343, Loss: {'train/loss': 1.0146852731704712, 'train/grad_norm_pre_clip': 7.427692413330078, 'timing/data': 0.004876860999502242, 'timing/model': 1.206772059085779, 'learning_rate/qwen_vl_interface': 9.970957455320723e-06, 'learning_rate/action_model': 9.970957455320723e-05, 'epoch': 0.1}) +[2026-06-17 06:44:06,736][starVLA.training.train_starvla][INFO] - Step 344, Loss: {'train/loss': 1.2955282926559448, 'train/grad_norm_pre_clip': 9.723734855651855, 'timing/data': 0.004561426001600921, 'timing/model': 1.3346338520059362, 'learning_rate/qwen_vl_interface': 9.970718176591723e-06, 'learning_rate/action_model': 9.970718176591722e-05, 'epoch': 0.1}) +[2026-06-17 06:44:11,466][starVLA.training.train_starvla][INFO] - Step 345, Loss: {'train/loss': 1.089978575706482, 'train/grad_norm_pre_clip': 7.306041717529297, 'timing/data': 0.0047659180127084255, 'timing/model': 1.2689413710031658, 'learning_rate/qwen_vl_interface': 9.970477919252763e-06, 'learning_rate/action_model': 9.970477919252763e-05, 'epoch': 0.1}) +[2026-06-17 06:44:16,142][starVLA.training.train_starvla][INFO] - Step 346, Loss: {'train/loss': 1.1279568672180176, 'train/grad_norm_pre_clip': 12.335379600524902, 'timing/data': 0.004836865933611989, 'timing/model': 1.2125848809955642, 'learning_rate/qwen_vl_interface': 9.970236683353652e-06, 'learning_rate/action_model': 9.970236683353652e-05, 'epoch': 0.1}) +[2026-06-17 06:44:20,901][starVLA.training.train_starvla][INFO] - Step 347, Loss: {'train/loss': 1.5098662376403809, 'train/grad_norm_pre_clip': 15.202117919921875, 'timing/data': 0.004944875021465123, 'timing/model': 1.2066738120047376, 'learning_rate/qwen_vl_interface': 9.969994468944399e-06, 'learning_rate/action_model': 9.969994468944398e-05, 'epoch': 0.1}) +[2026-06-17 06:44:25,795][starVLA.training.train_starvla][INFO] - Step 348, Loss: {'train/loss': 1.3060325384140015, 'train/grad_norm_pre_clip': 7.5060224533081055, 'timing/data': 0.005089164013043046, 'timing/model': 1.2218861349392682, 'learning_rate/qwen_vl_interface': 9.969751276075212e-06, 'learning_rate/action_model': 9.969751276075212e-05, 'epoch': 0.1}) +[2026-06-17 06:44:30,594][starVLA.training.train_starvla][INFO] - Step 349, Loss: {'train/loss': 1.3332102298736572, 'train/grad_norm_pre_clip': 9.227615356445312, 'timing/data': 0.0047560089733451605, 'timing/model': 1.2135180899640545, 'learning_rate/qwen_vl_interface': 9.969507104796506e-06, 'learning_rate/action_model': 9.969507104796505e-05, 'epoch': 0.1}) +[2026-06-17 06:44:35,381][starVLA.training.train_starvla][INFO] - Step 350, Loss: {'train/loss': 0.8960390686988831, 'train/grad_norm_pre_clip': 7.625740051269531, 'timing/data': 0.005067048943601549, 'timing/model': 1.33160414500162, 'learning_rate/qwen_vl_interface': 9.969261955158898e-06, 'learning_rate/action_model': 9.969261955158898e-05, 'epoch': 0.1}) +[2026-06-17 06:44:40,164][starVLA.training.train_starvla][INFO] - Step 351, Loss: {'train/loss': 1.151258111000061, 'train/grad_norm_pre_clip': 12.36746597290039, 'timing/data': 0.005099572008475661, 'timing/model': 1.3261313480325043, 'learning_rate/qwen_vl_interface': 9.96901582721321e-06, 'learning_rate/action_model': 9.96901582721321e-05, 'epoch': 0.1}) +[2026-06-17 06:44:44,831][starVLA.training.train_starvla][INFO] - Step 352, Loss: {'train/loss': 1.1794357299804688, 'train/grad_norm_pre_clip': 129.6964874267578, 'timing/data': 0.005089327925816178, 'timing/model': 1.2059750719927251, 'learning_rate/qwen_vl_interface': 9.968768721010463e-06, 'learning_rate/action_model': 9.968768721010463e-05, 'epoch': 0.1}) +[2026-06-17 06:44:49,599][starVLA.training.train_starvla][INFO] - Step 353, Loss: {'train/loss': 1.3663734197616577, 'train/grad_norm_pre_clip': 6.976163387298584, 'timing/data': 0.004861138993874192, 'timing/model': 1.2082234889967367, 'learning_rate/qwen_vl_interface': 9.968520636601883e-06, 'learning_rate/action_model': 9.968520636601883e-05, 'epoch': 0.1}) +[2026-06-17 06:44:54,382][starVLA.training.train_starvla][INFO] - Step 354, Loss: {'train/loss': 1.152440071105957, 'train/grad_norm_pre_clip': 6.7789154052734375, 'timing/data': 0.004782879026606679, 'timing/model': 1.2070004650158808, 'learning_rate/qwen_vl_interface': 9.968271574038899e-06, 'learning_rate/action_model': 9.968271574038898e-05, 'epoch': 0.1}) +[2026-06-17 06:44:59,177][starVLA.training.train_starvla][INFO] - Step 355, Loss: {'train/loss': 0.9601109623908997, 'train/grad_norm_pre_clip': 12.36873722076416, 'timing/data': 0.0048593629617244005, 'timing/model': 1.207996545941569, 'learning_rate/qwen_vl_interface': 9.968021533373138e-06, 'learning_rate/action_model': 9.968021533373138e-05, 'epoch': 0.1}) +[2026-06-17 06:45:03,970][starVLA.training.train_starvla][INFO] - Step 356, Loss: {'train/loss': 1.362486720085144, 'train/grad_norm_pre_clip': 7.110135078430176, 'timing/data': 0.00498815905302763, 'timing/model': 1.3351963569875807, 'learning_rate/qwen_vl_interface': 9.967770514656439e-06, 'learning_rate/action_model': 9.967770514656438e-05, 'epoch': 0.1}) +[2026-06-17 06:45:08,634][starVLA.training.train_starvla][INFO] - Step 357, Loss: {'train/loss': 1.1850444078445435, 'train/grad_norm_pre_clip': 95.28673553466797, 'timing/data': 0.004947773995809257, 'timing/model': 1.2077897159615532, 'learning_rate/qwen_vl_interface': 9.967518517940832e-06, 'learning_rate/action_model': 9.967518517940832e-05, 'epoch': 0.11}) +[2026-06-17 06:45:13,421][starVLA.training.train_starvla][INFO] - Step 358, Loss: {'train/loss': 1.3898531198501587, 'train/grad_norm_pre_clip': 10.713064193725586, 'timing/data': 0.004866143921390176, 'timing/model': 1.2053109710104764, 'learning_rate/qwen_vl_interface': 9.967265543278563e-06, 'learning_rate/action_model': 9.967265543278562e-05, 'epoch': 0.11}) +[2026-06-17 06:45:18,204][starVLA.training.train_starvla][INFO] - Step 359, Loss: {'train/loss': 0.8363810181617737, 'train/grad_norm_pre_clip': 9.503345489501953, 'timing/data': 0.00511633011046797, 'timing/model': 1.2044052209239453, 'learning_rate/qwen_vl_interface': 9.96701159072207e-06, 'learning_rate/action_model': 9.96701159072207e-05, 'epoch': 0.11}) +[2026-06-17 06:45:22,992][starVLA.training.train_starvla][INFO] - Step 360, Loss: {'train/loss': 0.8897997140884399, 'train/grad_norm_pre_clip': 9.10400104522705, 'timing/data': 0.005099882953800261, 'timing/model': 1.206086502992548, 'learning_rate/qwen_vl_interface': 9.966756660323998e-06, 'learning_rate/action_model': 9.966756660323999e-05, 'epoch': 0.11}) +[2026-06-17 06:45:27,721][starVLA.training.train_starvla][INFO] - Step 361, Loss: {'train/loss': 0.9486809968948364, 'train/grad_norm_pre_clip': 7.549408912658691, 'timing/data': 0.004424892948009074, 'timing/model': 1.2089879380073398, 'learning_rate/qwen_vl_interface': 9.966500752137195e-06, 'learning_rate/action_model': 9.966500752137196e-05, 'epoch': 0.11}) +[2026-06-17 06:45:32,470][starVLA.training.train_starvla][INFO] - Step 362, Loss: {'train/loss': 1.251782774925232, 'train/grad_norm_pre_clip': 9.25328540802002, 'timing/data': 0.004912497941404581, 'timing/model': 1.2815947689814493, 'learning_rate/qwen_vl_interface': 9.966243866214712e-06, 'learning_rate/action_model': 9.966243866214712e-05, 'epoch': 0.11}) +[2026-06-17 06:45:37,136][starVLA.training.train_starvla][INFO] - Step 363, Loss: {'train/loss': 1.2237094640731812, 'train/grad_norm_pre_clip': 11.057405471801758, 'timing/data': 0.004732878995127976, 'timing/model': 1.2067549979547039, 'learning_rate/qwen_vl_interface': 9.965986002609799e-06, 'learning_rate/action_model': 9.965986002609799e-05, 'epoch': 0.11}) +[2026-06-17 06:45:42,001][starVLA.training.train_starvla][INFO] - Step 364, Loss: {'train/loss': 1.1802773475646973, 'train/grad_norm_pre_clip': 8.78126049041748, 'timing/data': 0.005015985923819244, 'timing/model': 1.2069287110352889, 'learning_rate/qwen_vl_interface': 9.965727161375915e-06, 'learning_rate/action_model': 9.965727161375914e-05, 'epoch': 0.11}) +[2026-06-17 06:45:46,791][starVLA.training.train_starvla][INFO] - Step 365, Loss: {'train/loss': 1.2042630910873413, 'train/grad_norm_pre_clip': 51.21685791015625, 'timing/data': 0.005369229009374976, 'timing/model': 1.20540103293024, 'learning_rate/qwen_vl_interface': 9.965467342566715e-06, 'learning_rate/action_model': 9.965467342566715e-05, 'epoch': 0.11}) +[2026-06-17 06:45:51,580][starVLA.training.train_starvla][INFO] - Step 366, Loss: {'train/loss': 1.229568362236023, 'train/grad_norm_pre_clip': 8.09192180633545, 'timing/data': 0.004685597028583288, 'timing/model': 1.20423152402509, 'learning_rate/qwen_vl_interface': 9.965206546236061e-06, 'learning_rate/action_model': 9.96520654623606e-05, 'epoch': 0.11}) +[2026-06-17 06:45:56,369][starVLA.training.train_starvla][INFO] - Step 367, Loss: {'train/loss': 1.0850943326950073, 'train/grad_norm_pre_clip': 8.644665718078613, 'timing/data': 0.004855016013607383, 'timing/model': 1.2092403449350968, 'learning_rate/qwen_vl_interface': 9.964944772438014e-06, 'learning_rate/action_model': 9.964944772438013e-05, 'epoch': 0.11}) +[2026-06-17 06:46:01,119][starVLA.training.train_starvla][INFO] - Step 368, Loss: {'train/loss': 1.0188963413238525, 'train/grad_norm_pre_clip': 9.388039588928223, 'timing/data': 0.00505615200381726, 'timing/model': 1.2929443269968033, 'learning_rate/qwen_vl_interface': 9.964682021226844e-06, 'learning_rate/action_model': 9.964682021226844e-05, 'epoch': 0.11}) +[2026-06-17 06:46:05,783][starVLA.training.train_starvla][INFO] - Step 369, Loss: {'train/loss': 1.0843976736068726, 'train/grad_norm_pre_clip': 8.021102905273438, 'timing/data': 0.005029845982789993, 'timing/model': 1.2066532570170239, 'learning_rate/qwen_vl_interface': 9.964418292657019e-06, 'learning_rate/action_model': 9.964418292657018e-05, 'epoch': 0.11}) +[2026-06-17 06:46:10,583][starVLA.training.train_starvla][INFO] - Step 370, Loss: {'train/loss': 1.2318288087844849, 'train/grad_norm_pre_clip': 14.328551292419434, 'timing/data': 0.004833110026083887, 'timing/model': 1.209499757969752, 'learning_rate/qwen_vl_interface': 9.96415358678321e-06, 'learning_rate/action_model': 9.964153586783209e-05, 'epoch': 0.11}) +[2026-06-17 06:46:15,388][starVLA.training.train_starvla][INFO] - Step 371, Loss: {'train/loss': 1.2893562316894531, 'train/grad_norm_pre_clip': 13.478060722351074, 'timing/data': 0.004689185065217316, 'timing/model': 1.2130048479884863, 'learning_rate/qwen_vl_interface': 9.963887903660285e-06, 'learning_rate/action_model': 9.963887903660285e-05, 'epoch': 0.11}) +[2026-06-17 06:46:20,176][starVLA.training.train_starvla][INFO] - Step 372, Loss: {'train/loss': 1.4752135276794434, 'train/grad_norm_pre_clip': 11.120468139648438, 'timing/data': 0.005378596018999815, 'timing/model': 1.2068469730438665, 'learning_rate/qwen_vl_interface': 9.96362124334333e-06, 'learning_rate/action_model': 9.963621243343329e-05, 'epoch': 0.11}) +[2026-06-17 06:46:24,996][starVLA.training.train_starvla][INFO] - Step 373, Loss: {'train/loss': 0.8747707009315491, 'train/grad_norm_pre_clip': 62.062644958496094, 'timing/data': 0.004914039047434926, 'timing/model': 1.3371965870028362, 'learning_rate/qwen_vl_interface': 9.963353605887617e-06, 'learning_rate/action_model': 9.963353605887616e-05, 'epoch': 0.11}) +[2026-06-17 06:46:29,652][starVLA.training.train_starvla][INFO] - Step 374, Loss: {'train/loss': 1.145472764968872, 'train/grad_norm_pre_clip': 8.597118377685547, 'timing/data': 0.0047607149463146925, 'timing/model': 1.2061172659741715, 'learning_rate/qwen_vl_interface': 9.963084991348628e-06, 'learning_rate/action_model': 9.963084991348628e-05, 'epoch': 0.11}) +[2026-06-17 06:46:34,442][starVLA.training.train_starvla][INFO] - Step 375, Loss: {'train/loss': 1.0083611011505127, 'train/grad_norm_pre_clip': 31.975854873657227, 'timing/data': 0.005086079938337207, 'timing/model': 1.209690295974724, 'learning_rate/qwen_vl_interface': 9.962815399782051e-06, 'learning_rate/action_model': 9.96281539978205e-05, 'epoch': 0.11}) +[2026-06-17 06:46:39,118][starVLA.training.train_starvla][INFO] - Step 376, Loss: {'train/loss': 0.951239287853241, 'train/grad_norm_pre_clip': 6.61107063293457, 'timing/data': 0.004604812012985349, 'timing/model': 1.205717921955511, 'learning_rate/qwen_vl_interface': 9.96254483124377e-06, 'learning_rate/action_model': 9.96254483124377e-05, 'epoch': 0.11}) +[2026-06-17 06:46:43,824][starVLA.training.train_starvla][INFO] - Step 377, Loss: {'train/loss': 1.2691882848739624, 'train/grad_norm_pre_clip': 13.486201286315918, 'timing/data': 0.0049121229676529765, 'timing/model': 1.2052734610624611, 'learning_rate/qwen_vl_interface': 9.962273285789876e-06, 'learning_rate/action_model': 9.962273285789875e-05, 'epoch': 0.11}) +[2026-06-17 06:46:48,572][starVLA.training.train_starvla][INFO] - Step 378, Loss: {'train/loss': 1.1224218606948853, 'train/grad_norm_pre_clip': 7.928979873657227, 'timing/data': 0.005068205064162612, 'timing/model': 1.2079450190067291, 'learning_rate/qwen_vl_interface': 9.962000763476658e-06, 'learning_rate/action_model': 9.962000763476658e-05, 'epoch': 0.11}) +[2026-06-17 06:46:53,365][starVLA.training.train_starvla][INFO] - Step 379, Loss: {'train/loss': 1.146233320236206, 'train/grad_norm_pre_clip': 10.802294731140137, 'timing/data': 0.00443155609536916, 'timing/model': 1.335793740930967, 'learning_rate/qwen_vl_interface': 9.96172726436061e-06, 'learning_rate/action_model': 9.96172726436061e-05, 'epoch': 0.11}) +[2026-06-17 06:46:58,031][starVLA.training.train_starvla][INFO] - Step 380, Loss: {'train/loss': 1.1464614868164062, 'train/grad_norm_pre_clip': 6.4022440910339355, 'timing/data': 0.004774576053023338, 'timing/model': 1.2089185109362006, 'learning_rate/qwen_vl_interface': 9.961452788498432e-06, 'learning_rate/action_model': 9.961452788498433e-05, 'epoch': 0.11}) +[2026-06-17 06:47:02,818][starVLA.training.train_starvla][INFO] - Step 381, Loss: {'train/loss': 1.2726452350616455, 'train/grad_norm_pre_clip': 8.007349014282227, 'timing/data': 0.005057204980403185, 'timing/model': 1.2062953989952803, 'learning_rate/qwen_vl_interface': 9.96117733594702e-06, 'learning_rate/action_model': 9.96117733594702e-05, 'epoch': 0.11}) +[2026-06-17 06:47:07,609][starVLA.training.train_starvla][INFO] - Step 382, Loss: {'train/loss': 1.252253770828247, 'train/grad_norm_pre_clip': 11.515222549438477, 'timing/data': 0.004918533959425986, 'timing/model': 1.2062515859724954, 'learning_rate/qwen_vl_interface': 9.960900906763478e-06, 'learning_rate/action_model': 9.960900906763477e-05, 'epoch': 0.11}) +[2026-06-17 06:47:12,419][starVLA.training.train_starvla][INFO] - Step 383, Loss: {'train/loss': 1.0824064016342163, 'train/grad_norm_pre_clip': 7.845432281494141, 'timing/data': 0.0049579390324652195, 'timing/model': 1.2179900540504605, 'learning_rate/qwen_vl_interface': 9.960623501005108e-06, 'learning_rate/action_model': 9.960623501005107e-05, 'epoch': 0.11}) +[2026-06-17 06:47:17,188][starVLA.training.train_starvla][INFO] - Step 384, Loss: {'train/loss': 0.8516802787780762, 'train/grad_norm_pre_clip': 7.3198561668396, 'timing/data': 0.004728464991785586, 'timing/model': 1.2112581268884242, 'learning_rate/qwen_vl_interface': 9.960345118729416e-06, 'learning_rate/action_model': 9.960345118729417e-05, 'epoch': 0.11}) +[2026-06-17 06:47:21,988][starVLA.training.train_starvla][INFO] - Step 385, Loss: {'train/loss': 1.0832427740097046, 'train/grad_norm_pre_clip': 7.6336750984191895, 'timing/data': 0.004582786932587624, 'timing/model': 1.3446775369811803, 'learning_rate/qwen_vl_interface': 9.960065759994116e-06, 'learning_rate/action_model': 9.960065759994114e-05, 'epoch': 0.11}) +[2026-06-17 06:47:26,672][starVLA.training.train_starvla][INFO] - Step 386, Loss: {'train/loss': 1.1941337585449219, 'train/grad_norm_pre_clip': 6.684764385223389, 'timing/data': 0.004966722917743027, 'timing/model': 1.2087346919579431, 'learning_rate/qwen_vl_interface': 9.959785424857113e-06, 'learning_rate/action_model': 9.959785424857112e-05, 'epoch': 0.11}) +[2026-06-17 06:47:31,477][starVLA.training.train_starvla][INFO] - Step 387, Loss: {'train/loss': 1.2007213830947876, 'train/grad_norm_pre_clip': 18.883676528930664, 'timing/data': 0.005059906048700213, 'timing/model': 1.2082626819610596, 'learning_rate/qwen_vl_interface': 9.959504113376523e-06, 'learning_rate/action_model': 9.959504113376523e-05, 'epoch': 0.11}) +[2026-06-17 06:47:36,272][starVLA.training.train_starvla][INFO] - Step 388, Loss: {'train/loss': 1.0411059856414795, 'train/grad_norm_pre_clip': 10.316153526306152, 'timing/data': 0.004703338956460357, 'timing/model': 1.207298162044026, 'learning_rate/qwen_vl_interface': 9.959221825610664e-06, 'learning_rate/action_model': 9.959221825610663e-05, 'epoch': 0.11}) +[2026-06-17 06:47:41,080][starVLA.training.train_starvla][INFO] - Step 389, Loss: {'train/loss': 1.1479928493499756, 'train/grad_norm_pre_clip': 5.8548665046691895, 'timing/data': 0.005222879000939429, 'timing/model': 1.206482767011039, 'learning_rate/qwen_vl_interface': 9.958938561618053e-06, 'learning_rate/action_model': 9.958938561618053e-05, 'epoch': 0.11}) +[2026-06-17 06:47:45,881][starVLA.training.train_starvla][INFO] - Step 390, Loss: {'train/loss': 1.0165519714355469, 'train/grad_norm_pre_clip': 8.240989685058594, 'timing/data': 0.00498981901910156, 'timing/model': 1.2055493230000138, 'learning_rate/qwen_vl_interface': 9.958654321457412e-06, 'learning_rate/action_model': 9.958654321457412e-05, 'epoch': 0.11}) +[2026-06-17 06:47:50,697][starVLA.training.train_starvla][INFO] - Step 391, Loss: {'train/loss': 1.220051646232605, 'train/grad_norm_pre_clip': 5.366488456726074, 'timing/data': 0.005166689050383866, 'timing/model': 1.3314648669911548, 'learning_rate/qwen_vl_interface': 9.958369105187661e-06, 'learning_rate/action_model': 9.958369105187661e-05, 'epoch': 0.12}) +[2026-06-17 06:47:55,363][starVLA.training.train_starvla][INFO] - Step 392, Loss: {'train/loss': 1.116590142250061, 'train/grad_norm_pre_clip': 32.368080139160156, 'timing/data': 0.005252895993180573, 'timing/model': 1.2076529350597411, 'learning_rate/qwen_vl_interface': 9.958082912867932e-06, 'learning_rate/action_model': 9.958082912867931e-05, 'epoch': 0.12}) +[2026-06-17 06:48:00,154][starVLA.training.train_starvla][INFO] - Step 393, Loss: {'train/loss': 1.0325158834457397, 'train/grad_norm_pre_clip': 6.363896369934082, 'timing/data': 0.00473615899682045, 'timing/model': 1.2058183549670503, 'learning_rate/qwen_vl_interface': 9.957795744557545e-06, 'learning_rate/action_model': 9.957795744557546e-05, 'epoch': 0.12}) +[2026-06-17 06:48:04,944][starVLA.training.train_starvla][INFO] - Step 394, Loss: {'train/loss': 1.3014706373214722, 'train/grad_norm_pre_clip': 6.1119842529296875, 'timing/data': 0.004965594038367271, 'timing/model': 1.2077268389984965, 'learning_rate/qwen_vl_interface': 9.957507600316036e-06, 'learning_rate/action_model': 9.957507600316036e-05, 'epoch': 0.12}) +[2026-06-17 06:48:09,708][starVLA.training.train_starvla][INFO] - Step 395, Loss: {'train/loss': 1.4130957126617432, 'train/grad_norm_pre_clip': 7.406765460968018, 'timing/data': 0.004849124001339078, 'timing/model': 1.2133755079703405, 'learning_rate/qwen_vl_interface': 9.957218480203137e-06, 'learning_rate/action_model': 9.957218480203136e-05, 'epoch': 0.12}) +[2026-06-17 06:48:14,508][starVLA.training.train_starvla][INFO] - Step 396, Loss: {'train/loss': 1.084688663482666, 'train/grad_norm_pre_clip': 7.7801899909973145, 'timing/data': 0.005377180990763009, 'timing/model': 1.2070005829446018, 'learning_rate/qwen_vl_interface': 9.95692838427878e-06, 'learning_rate/action_model': 9.956928384278778e-05, 'epoch': 0.12}) +[2026-06-17 06:48:19,307][starVLA.training.train_starvla][INFO] - Step 397, Loss: {'train/loss': 0.9892758727073669, 'train/grad_norm_pre_clip': 6.138365268707275, 'timing/data': 0.005342805990949273, 'timing/model': 1.3385724700056016, 'learning_rate/qwen_vl_interface': 9.956637312603105e-06, 'learning_rate/action_model': 9.956637312603104e-05, 'epoch': 0.12}) +[2026-06-17 06:48:23,976][starVLA.training.train_starvla][INFO] - Step 398, Loss: {'train/loss': 0.9755669832229614, 'train/grad_norm_pre_clip': 5.064846515655518, 'timing/data': 0.005357854999601841, 'timing/model': 1.2071174290031195, 'learning_rate/qwen_vl_interface': 9.95634526523645e-06, 'learning_rate/action_model': 9.956345265236449e-05, 'epoch': 0.12}) +[2026-06-17 06:48:28,776][starVLA.training.train_starvla][INFO] - Step 399, Loss: {'train/loss': 1.1079615354537964, 'train/grad_norm_pre_clip': 7.408055782318115, 'timing/data': 0.0048373970203101635, 'timing/model': 1.2097399280173704, 'learning_rate/qwen_vl_interface': 9.956052242239358e-06, 'learning_rate/action_model': 9.956052242239358e-05, 'epoch': 0.12}) +[2026-06-17 06:50:11,213][starVLA.training.train_starvla][INFO] - Step 400, Loss: {'train/loss': 1.1485507488250732, 'train/grad_norm_pre_clip': 7.890916347503662, 'eval/loss': 1.1359812021255493, 'eval/latency_0/loss': 1.1359812021255493, 'eval/demon_fire_precision': 0.8304221251819505, 'eval/demon_fire_recall': 0.9476744186046512, 'eval/demon_fire_f1': 0.8851823118696663, 'eval/demon_fire_rate_model': 0.764182424916574, 'eval/demon_fire_rate_teacher': 0.6696329254727474, 'eval/demon_left_precision': 0.6576942610462163, 'eval/demon_left_recall': 0.7337110481586402, 'eval/demon_left_f1': 0.6936261381896089, 'eval/demon_left_rate_model': 0.5475528364849833, 'eval/demon_left_rate_teacher': 0.49082313681868744, 'eval/demon_right_precision': 0.6939151813153043, 'eval/demon_right_recall': 0.6378531073446327, 'eval/demon_right_f1': 0.6647041507212246, 'eval/demon_right_rate_model': 0.4524471635150167, 'eval/demon_right_rate_teacher': 0.4922135706340378, 'eval/demon_move_f1': 0.6791651444554168, 'eval/demon_cc_f1': 0.8113482780982784, 'eval/latency_0/demon_fire_precision': 0.8304221251819505, 'eval/latency_0/demon_fire_recall': 0.9476744186046512, 'eval/latency_0/demon_fire_f1': 0.8851823118696663, 'eval/latency_0/demon_fire_rate_model': 0.764182424916574, 'eval/latency_0/demon_fire_rate_teacher': 0.6696329254727474, 'eval/latency_0/demon_left_precision': 0.6576942610462163, 'eval/latency_0/demon_left_recall': 0.7337110481586402, 'eval/latency_0/demon_left_f1': 0.6936261381896089, 'eval/latency_0/demon_left_rate_model': 0.5475528364849833, 'eval/latency_0/demon_left_rate_teacher': 0.49082313681868744, 'eval/latency_0/demon_right_precision': 0.6939151813153043, 'eval/latency_0/demon_right_recall': 0.6378531073446327, 'eval/latency_0/demon_right_f1': 0.6647041507212246, 'eval/latency_0/demon_right_rate_model': 0.4524471635150167, 'eval/latency_0/demon_right_rate_teacher': 0.4922135706340378, 'eval/latency_0/demon_move_f1': 0.6791651444554168, 'eval/latency_0/demon_cc_f1': 0.8113482780982784, 'timing/data': 0.004618934937752783, 'timing/model': 1.2069414359284565, 'learning_rate/qwen_vl_interface': 9.955758243672572e-06, 'learning_rate/action_model': 9.955758243672572e-05, 'epoch': 0.12}) +[2026-06-17 06:50:11,214][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_400_state +[2026-06-17 06:50:11,217][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 06:50:14,764][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_400_state/model.safetensors +[2026-06-17 06:50:22,541][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_400_state/optimizer.bin +[2026-06-17 06:50:22,541][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_400_state/sampler.bin +[2026-06-17 06:50:22,541][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_400_state/sampler_1.bin +[2026-06-17 06:50:22,542][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_400_state/random_states_0.pkl +[2026-06-17 06:50:22,575][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 400 +[2026-06-17 06:50:22,576][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 06:50:22,607][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 06:50:27,343][starVLA.training.train_starvla][INFO] - Step 401, Loss: {'train/loss': 1.316318392753601, 'train/grad_norm_pre_clip': 21.403905868530273, 'timing/data': 0.0050291180377826095, 'timing/model': 1.1883657639846206, 'learning_rate/qwen_vl_interface': 9.955463269597038e-06, 'learning_rate/action_model': 9.955463269597038e-05, 'epoch': 0.12}) +[2026-06-17 06:50:32,071][starVLA.training.train_starvla][INFO] - Step 402, Loss: {'train/loss': 1.193626046180725, 'train/grad_norm_pre_clip': 6.270876884460449, 'timing/data': 0.004712998983450234, 'timing/model': 1.191094606067054, 'learning_rate/qwen_vl_interface': 9.955167320073905e-06, 'learning_rate/action_model': 9.955167320073905e-05, 'epoch': 0.12}) +[2026-06-17 06:50:36,884][starVLA.training.train_starvla][INFO] - Step 403, Loss: {'train/loss': 1.0766910314559937, 'train/grad_norm_pre_clip': 8.731127738952637, 'timing/data': 0.005028453073464334, 'timing/model': 1.1930078529985622, 'learning_rate/qwen_vl_interface': 9.954870395164524e-06, 'learning_rate/action_model': 9.954870395164524e-05, 'epoch': 0.12}) +[2026-06-17 06:50:41,591][starVLA.training.train_starvla][INFO] - Step 404, Loss: {'train/loss': 1.025285005569458, 'train/grad_norm_pre_clip': 12.69742488861084, 'timing/data': 0.004905537003651261, 'timing/model': 1.1946133259916678, 'learning_rate/qwen_vl_interface': 9.954572494930448e-06, 'learning_rate/action_model': 9.954572494930448e-05, 'epoch': 0.12}) +[2026-06-17 06:50:46,415][starVLA.training.train_starvla][INFO] - Step 405, Loss: {'train/loss': 1.2879931926727295, 'train/grad_norm_pre_clip': 13.38841438293457, 'timing/data': 0.004356308025307953, 'timing/model': 1.3374860379844904, 'learning_rate/qwen_vl_interface': 9.95427361943343e-06, 'learning_rate/action_model': 9.954273619433431e-05, 'epoch': 0.12}) +[2026-06-17 06:50:51,047][starVLA.training.train_starvla][INFO] - Step 406, Loss: {'train/loss': 1.0076606273651123, 'train/grad_norm_pre_clip': 7.605193138122559, 'timing/data': 0.0050343479961156845, 'timing/model': 1.1999405520036817, 'learning_rate/qwen_vl_interface': 9.95397376873543e-06, 'learning_rate/action_model': 9.95397376873543e-05, 'epoch': 0.12}) +[2026-06-17 06:50:55,808][starVLA.training.train_starvla][INFO] - Step 407, Loss: {'train/loss': 1.014979362487793, 'train/grad_norm_pre_clip': 10.479633331298828, 'timing/data': 0.004685522988438606, 'timing/model': 1.198666170006618, 'learning_rate/qwen_vl_interface': 9.953672942898606e-06, 'learning_rate/action_model': 9.953672942898605e-05, 'epoch': 0.12}) +[2026-06-17 06:51:00,583][starVLA.training.train_starvla][INFO] - Step 408, Loss: {'train/loss': 0.9696931838989258, 'train/grad_norm_pre_clip': 9.82138729095459, 'timing/data': 0.004571718047372997, 'timing/model': 1.2020338410511613, 'learning_rate/qwen_vl_interface': 9.953371141985318e-06, 'learning_rate/action_model': 9.953371141985319e-05, 'epoch': 0.12}) +[2026-06-17 06:51:05,298][starVLA.training.train_starvla][INFO] - Step 409, Loss: {'train/loss': 1.0710275173187256, 'train/grad_norm_pre_clip': 11.713286399841309, 'timing/data': 0.005270116962492466, 'timing/model': 1.198696564999409, 'learning_rate/qwen_vl_interface': 9.953068366058133e-06, 'learning_rate/action_model': 9.953068366058133e-05, 'epoch': 0.12}) +[2026-06-17 06:51:10,065][starVLA.training.train_starvla][INFO] - Step 410, Loss: {'train/loss': 0.9422893524169922, 'train/grad_norm_pre_clip': 10.189826011657715, 'timing/data': 0.0048043240094557405, 'timing/model': 1.1998813740210608, 'learning_rate/qwen_vl_interface': 9.952764615179813e-06, 'learning_rate/action_model': 9.952764615179812e-05, 'epoch': 0.12}) +[2026-06-17 06:51:14,838][starVLA.training.train_starvla][INFO] - Step 411, Loss: {'train/loss': 1.397695779800415, 'train/grad_norm_pre_clip': 18.345260620117188, 'timing/data': 0.004796541994437575, 'timing/model': 1.3305395729839802, 'learning_rate/qwen_vl_interface': 9.95245988941333e-06, 'learning_rate/action_model': 9.95245988941333e-05, 'epoch': 0.12}) +[2026-06-17 06:51:19,492][starVLA.training.train_starvla][INFO] - Step 412, Loss: {'train/loss': 1.4280457496643066, 'train/grad_norm_pre_clip': 10.708327293395996, 'timing/data': 0.0047694939421489835, 'timing/model': 1.211835250025615, 'learning_rate/qwen_vl_interface': 9.952154188821852e-06, 'learning_rate/action_model': 9.952154188821851e-05, 'epoch': 0.12}) +[2026-06-17 06:51:24,261][starVLA.training.train_starvla][INFO] - Step 413, Loss: {'train/loss': 1.1283502578735352, 'train/grad_norm_pre_clip': 20.407434463500977, 'timing/data': 0.004625112051144242, 'timing/model': 1.2017748829675838, 'learning_rate/qwen_vl_interface': 9.95184751346875e-06, 'learning_rate/action_model': 9.95184751346875e-05, 'epoch': 0.12}) +[2026-06-17 06:51:28,970][starVLA.training.train_starvla][INFO] - Step 414, Loss: {'train/loss': 0.9076273441314697, 'train/grad_norm_pre_clip': 10.882816314697266, 'timing/data': 0.004412440932355821, 'timing/model': 1.2066861030180007, 'learning_rate/qwen_vl_interface': 9.951539863417598e-06, 'learning_rate/action_model': 9.951539863417599e-05, 'epoch': 0.12}) +[2026-06-17 06:51:33,702][starVLA.training.train_starvla][INFO] - Step 415, Loss: {'train/loss': 1.0633944272994995, 'train/grad_norm_pre_clip': 9.846006393432617, 'timing/data': 0.004553477978333831, 'timing/model': 1.2029531099833548, 'learning_rate/qwen_vl_interface': 9.951231238732174e-06, 'learning_rate/action_model': 9.951231238732173e-05, 'epoch': 0.12}) +[2026-06-17 06:51:38,490][starVLA.training.train_starvla][INFO] - Step 416, Loss: {'train/loss': 1.0762336254119873, 'train/grad_norm_pre_clip': 6.480626106262207, 'timing/data': 0.0048035141080617905, 'timing/model': 1.2049237209139392, 'learning_rate/qwen_vl_interface': 9.950921639476455e-06, 'learning_rate/action_model': 9.950921639476455e-05, 'epoch': 0.12}) +[2026-06-17 06:51:43,281][starVLA.training.train_starvla][INFO] - Step 417, Loss: {'train/loss': 1.100443720817566, 'train/grad_norm_pre_clip': 8.539575576782227, 'timing/data': 0.004824056988582015, 'timing/model': 1.332967013004236, 'learning_rate/qwen_vl_interface': 9.950611065714623e-06, 'learning_rate/action_model': 9.950611065714623e-05, 'epoch': 0.12}) +[2026-06-17 06:51:47,931][starVLA.training.train_starvla][INFO] - Step 418, Loss: {'train/loss': 0.9933093786239624, 'train/grad_norm_pre_clip': 10.942248344421387, 'timing/data': 0.004781924071721733, 'timing/model': 1.2036739269969985, 'learning_rate/qwen_vl_interface': 9.950299517511058e-06, 'learning_rate/action_model': 9.950299517511058e-05, 'epoch': 0.12}) +[2026-06-17 06:51:52,711][starVLA.training.train_starvla][INFO] - Step 419, Loss: {'train/loss': 1.0186010599136353, 'train/grad_norm_pre_clip': 10.834427833557129, 'timing/data': 0.0046664089895784855, 'timing/model': 1.2028721949318424, 'learning_rate/qwen_vl_interface': 9.949986994930346e-06, 'learning_rate/action_model': 9.949986994930346e-05, 'epoch': 0.12}) +[2026-06-17 06:51:57,536][starVLA.training.train_starvla][INFO] - Step 420, Loss: {'train/loss': 1.0147050619125366, 'train/grad_norm_pre_clip': 11.14522933959961, 'timing/data': 0.0058654909953475, 'timing/model': 1.2042075649369508, 'learning_rate/qwen_vl_interface': 9.949673498037272e-06, 'learning_rate/action_model': 9.949673498037272e-05, 'epoch': 0.12}) +[2026-06-17 06:52:02,290][starVLA.training.train_starvla][INFO] - Step 421, Loss: {'train/loss': 0.929977297782898, 'train/grad_norm_pre_clip': 6.841572284698486, 'timing/data': 0.0046040830202400684, 'timing/model': 1.2048086860449985, 'learning_rate/qwen_vl_interface': 9.949359026896826e-06, 'learning_rate/action_model': 9.949359026896826e-05, 'epoch': 0.12}) +[2026-06-17 06:52:07,087][starVLA.training.train_starvla][INFO] - Step 422, Loss: {'train/loss': 0.8988968729972839, 'train/grad_norm_pre_clip': 7.116512775421143, 'timing/data': 0.0050588109297677875, 'timing/model': 1.2044208879815415, 'learning_rate/qwen_vl_interface': 9.949043581574196e-06, 'learning_rate/action_model': 9.949043581574196e-05, 'epoch': 0.12}) +[2026-06-17 06:52:11,878][starVLA.training.train_starvla][INFO] - Step 423, Loss: {'train/loss': 0.9067239761352539, 'train/grad_norm_pre_clip': 8.833588600158691, 'timing/data': 0.004927207017317414, 'timing/model': 1.3404070839751512, 'learning_rate/qwen_vl_interface': 9.948727162134775e-06, 'learning_rate/action_model': 9.948727162134775e-05, 'epoch': 0.12}) +[2026-06-17 06:52:16,534][starVLA.training.train_starvla][INFO] - Step 424, Loss: {'train/loss': 1.1455692052841187, 'train/grad_norm_pre_clip': 9.122514724731445, 'timing/data': 0.004970257985405624, 'timing/model': 1.205969992093742, 'learning_rate/qwen_vl_interface': 9.948409768644157e-06, 'learning_rate/action_model': 9.948409768644156e-05, 'epoch': 0.12}) +[2026-06-17 06:52:21,364][starVLA.training.train_starvla][INFO] - Step 425, Loss: {'train/loss': 0.9764204621315002, 'train/grad_norm_pre_clip': 12.10328197479248, 'timing/data': 0.0051310049602761865, 'timing/model': 1.2104557109996676, 'learning_rate/qwen_vl_interface': 9.948091401168139e-06, 'learning_rate/action_model': 9.948091401168138e-05, 'epoch': 0.13}) +[2026-06-17 06:52:26,166][starVLA.training.train_starvla][INFO] - Step 426, Loss: {'train/loss': 1.1290937662124634, 'train/grad_norm_pre_clip': 7.259683132171631, 'timing/data': 0.004720691009424627, 'timing/model': 1.2477061609970406, 'learning_rate/qwen_vl_interface': 9.947772059772718e-06, 'learning_rate/action_model': 9.947772059772717e-05, 'epoch': 0.13}) +[2026-06-17 06:52:31,010][starVLA.training.train_starvla][INFO] - Step 427, Loss: {'train/loss': 1.0632445812225342, 'train/grad_norm_pre_clip': 5.719996452331543, 'timing/data': 0.004992906004190445, 'timing/model': 1.2040103799663484, 'learning_rate/qwen_vl_interface': 9.947451744524094e-06, 'learning_rate/action_model': 9.947451744524093e-05, 'epoch': 0.13}) +[2026-06-17 06:52:35,896][starVLA.training.train_starvla][INFO] - Step 428, Loss: {'train/loss': 1.0068938732147217, 'train/grad_norm_pre_clip': 6.362029075622559, 'timing/data': 0.005274091032333672, 'timing/model': 1.2067292799474671, 'learning_rate/qwen_vl_interface': 9.947130455488668e-06, 'learning_rate/action_model': 9.947130455488666e-05, 'epoch': 0.13}) +[2026-06-17 06:52:40,692][starVLA.training.train_starvla][INFO] - Step 429, Loss: {'train/loss': 1.1141966581344604, 'train/grad_norm_pre_clip': 10.937841415405273, 'timing/data': 0.005142510985024273, 'timing/model': 1.3321258840151131, 'learning_rate/qwen_vl_interface': 9.946808192733045e-06, 'learning_rate/action_model': 9.946808192733044e-05, 'epoch': 0.13}) +[2026-06-17 06:52:45,393][starVLA.training.train_starvla][INFO] - Step 430, Loss: {'train/loss': 1.2895647287368774, 'train/grad_norm_pre_clip': 14.870613098144531, 'timing/data': 0.004936118028126657, 'timing/model': 1.245517487055622, 'learning_rate/qwen_vl_interface': 9.946484956324028e-06, 'learning_rate/action_model': 9.946484956324028e-05, 'epoch': 0.13}) +[2026-06-17 06:52:50,048][starVLA.training.train_starvla][INFO] - Step 431, Loss: {'train/loss': 1.0030114650726318, 'train/grad_norm_pre_clip': 10.469006538391113, 'timing/data': 0.004537526983767748, 'timing/model': 1.2051638249540702, 'learning_rate/qwen_vl_interface': 9.946160746328628e-06, 'learning_rate/action_model': 9.946160746328627e-05, 'epoch': 0.13}) +[2026-06-17 06:52:54,810][starVLA.training.train_starvla][INFO] - Step 432, Loss: {'train/loss': 1.01921808719635, 'train/grad_norm_pre_clip': 17.273618698120117, 'timing/data': 0.004663485917262733, 'timing/model': 1.2070290540577844, 'learning_rate/qwen_vl_interface': 9.945835562814052e-06, 'learning_rate/action_model': 9.945835562814051e-05, 'epoch': 0.13}) +[2026-06-17 06:52:59,606][starVLA.training.train_starvla][INFO] - Step 433, Loss: {'train/loss': 1.142257571220398, 'train/grad_norm_pre_clip': 8.887928009033203, 'timing/data': 0.0048507609171792865, 'timing/model': 1.2069795089773834, 'learning_rate/qwen_vl_interface': 9.94550940584771e-06, 'learning_rate/action_model': 9.945509405847709e-05, 'epoch': 0.13}) +[2026-06-17 06:53:04,395][starVLA.training.train_starvla][INFO] - Step 434, Loss: {'train/loss': 0.8897915482521057, 'train/grad_norm_pre_clip': 8.014548301696777, 'timing/data': 0.004896753001958132, 'timing/model': 1.205771206994541, 'learning_rate/qwen_vl_interface': 9.945182275497215e-06, 'learning_rate/action_model': 9.945182275497215e-05, 'epoch': 0.13}) +[2026-06-17 06:53:09,191][starVLA.training.train_starvla][INFO] - Step 435, Loss: {'train/loss': 1.0306313037872314, 'train/grad_norm_pre_clip': 9.020444869995117, 'timing/data': 0.00531803397461772, 'timing/model': 1.3391961120069027, 'learning_rate/qwen_vl_interface': 9.944854171830382e-06, 'learning_rate/action_model': 9.944854171830383e-05, 'epoch': 0.13}) +[2026-06-17 06:53:13,979][starVLA.training.train_starvla][INFO] - Step 436, Loss: {'train/loss': 0.9387744069099426, 'train/grad_norm_pre_clip': 11.557827949523926, 'timing/data': 0.004789161030203104, 'timing/model': 1.3303015349665657, 'learning_rate/qwen_vl_interface': 9.94452509491523e-06, 'learning_rate/action_model': 9.94452509491523e-05, 'epoch': 0.13}) +[2026-06-17 06:53:18,652][starVLA.training.train_starvla][INFO] - Step 437, Loss: {'train/loss': 1.0466241836547852, 'train/grad_norm_pre_clip': 8.628229141235352, 'timing/data': 0.004687740001827478, 'timing/model': 1.2104348649736494, 'learning_rate/qwen_vl_interface': 9.944195044819973e-06, 'learning_rate/action_model': 9.944195044819972e-05, 'epoch': 0.13}) +[2026-06-17 06:53:23,434][starVLA.training.train_starvla][INFO] - Step 438, Loss: {'train/loss': 0.9385237097740173, 'train/grad_norm_pre_clip': 6.535266399383545, 'timing/data': 0.005213164957240224, 'timing/model': 1.2026198230450973, 'learning_rate/qwen_vl_interface': 9.94386402161303e-06, 'learning_rate/action_model': 9.943864021613029e-05, 'epoch': 0.13}) +[2026-06-17 06:53:28,222][starVLA.training.train_starvla][INFO] - Step 439, Loss: {'train/loss': 1.0487782955169678, 'train/grad_norm_pre_clip': 10.250701904296875, 'timing/data': 0.005386341945268214, 'timing/model': 1.2072368799708784, 'learning_rate/qwen_vl_interface': 9.943532025363026e-06, 'learning_rate/action_model': 9.943532025363025e-05, 'epoch': 0.13}) +[2026-06-17 06:53:33,013][starVLA.training.train_starvla][INFO] - Step 440, Loss: {'train/loss': 1.3669259548187256, 'train/grad_norm_pre_clip': 8.284232139587402, 'timing/data': 0.004857565043494105, 'timing/model': 1.2060100300004706, 'learning_rate/qwen_vl_interface': 9.943199056138781e-06, 'learning_rate/action_model': 9.943199056138782e-05, 'epoch': 0.13}) +[2026-06-17 06:53:37,810][starVLA.training.train_starvla][INFO] - Step 441, Loss: {'train/loss': 0.9976691603660583, 'train/grad_norm_pre_clip': 6.831756591796875, 'timing/data': 0.004895466030575335, 'timing/model': 1.3351226639933884, 'learning_rate/qwen_vl_interface': 9.942865114009325e-06, 'learning_rate/action_model': 9.942865114009324e-05, 'epoch': 0.13}) +[2026-06-17 06:53:42,468][starVLA.training.train_starvla][INFO] - Step 442, Loss: {'train/loss': 1.2248830795288086, 'train/grad_norm_pre_clip': 9.870429039001465, 'timing/data': 0.0048818939831107855, 'timing/model': 1.205963682034053, 'learning_rate/qwen_vl_interface': 9.94253019904388e-06, 'learning_rate/action_model': 9.942530199043879e-05, 'epoch': 0.13}) +[2026-06-17 06:53:47,256][starVLA.training.train_starvla][INFO] - Step 443, Loss: {'train/loss': 1.1274421215057373, 'train/grad_norm_pre_clip': 9.08699893951416, 'timing/data': 0.004899178980849683, 'timing/model': 1.2069964020047337, 'learning_rate/qwen_vl_interface': 9.942194311311874e-06, 'learning_rate/action_model': 9.942194311311874e-05, 'epoch': 0.13}) +[2026-06-17 06:53:52,073][starVLA.training.train_starvla][INFO] - Step 444, Loss: {'train/loss': 1.2700412273406982, 'train/grad_norm_pre_clip': 8.456321716308594, 'timing/data': 0.005181452026590705, 'timing/model': 1.215135536971502, 'learning_rate/qwen_vl_interface': 9.941857450882939e-06, 'learning_rate/action_model': 9.941857450882937e-05, 'epoch': 0.13}) +[2026-06-17 06:53:56,890][starVLA.training.train_starvla][INFO] - Step 445, Loss: {'train/loss': 1.04812753200531, 'train/grad_norm_pre_clip': 6.136264801025391, 'timing/data': 0.004750719061121345, 'timing/model': 1.2119221009779721, 'learning_rate/qwen_vl_interface': 9.941519617826906e-06, 'learning_rate/action_model': 9.941519617826906e-05, 'epoch': 0.13}) +[2026-06-17 06:54:01,614][starVLA.training.train_starvla][INFO] - Step 446, Loss: {'train/loss': 1.030005693435669, 'train/grad_norm_pre_clip': 8.45301342010498, 'timing/data': 0.004944983986206353, 'timing/model': 1.2062839289428666, 'learning_rate/qwen_vl_interface': 9.941180812213807e-06, 'learning_rate/action_model': 9.941180812213806e-05, 'epoch': 0.13}) +[2026-06-17 06:54:06,376][starVLA.training.train_starvla][INFO] - Step 447, Loss: {'train/loss': 1.2521320581436157, 'train/grad_norm_pre_clip': 8.664322853088379, 'timing/data': 0.004743951954878867, 'timing/model': 1.2813459628960118, 'learning_rate/qwen_vl_interface': 9.940841034113878e-06, 'learning_rate/action_model': 9.940841034113877e-05, 'epoch': 0.13}) +[2026-06-17 06:54:11,039][starVLA.training.train_starvla][INFO] - Step 448, Loss: {'train/loss': 1.0308408737182617, 'train/grad_norm_pre_clip': 8.601608276367188, 'timing/data': 0.004890308016911149, 'timing/model': 1.206822804058902, 'learning_rate/qwen_vl_interface': 9.940500283597557e-06, 'learning_rate/action_model': 9.940500283597556e-05, 'epoch': 0.13}) +[2026-06-17 06:54:15,845][starVLA.training.train_starvla][INFO] - Step 449, Loss: {'train/loss': 1.1196335554122925, 'train/grad_norm_pre_clip': 11.098649978637695, 'timing/data': 0.004808638943359256, 'timing/model': 1.21335293201264, 'learning_rate/qwen_vl_interface': 9.940158560735478e-06, 'learning_rate/action_model': 9.940158560735477e-05, 'epoch': 0.13}) +[2026-06-17 06:54:20,664][starVLA.training.train_starvla][INFO] - Step 450, Loss: {'train/loss': 1.0396345853805542, 'train/grad_norm_pre_clip': 8.637434959411621, 'timing/data': 0.004906526068225503, 'timing/model': 1.2100048150168732, 'learning_rate/qwen_vl_interface': 9.939815865598482e-06, 'learning_rate/action_model': 9.93981586559848e-05, 'epoch': 0.13}) +[2026-06-17 06:54:25,475][starVLA.training.train_starvla][INFO] - Step 451, Loss: {'train/loss': 1.1365231275558472, 'train/grad_norm_pre_clip': 7.4624924659729, 'timing/data': 0.005296163028106093, 'timing/model': 1.2106852430151775, 'learning_rate/qwen_vl_interface': 9.939472198257611e-06, 'learning_rate/action_model': 9.93947219825761e-05, 'epoch': 0.13}) +[2026-06-17 06:54:30,279][starVLA.training.train_starvla][INFO] - Step 452, Loss: {'train/loss': 0.9077826738357544, 'train/grad_norm_pre_clip': 5.452702045440674, 'timing/data': 0.004445002996362746, 'timing/model': 1.212851713062264, 'learning_rate/qwen_vl_interface': 9.939127558784108e-06, 'learning_rate/action_model': 9.939127558784107e-05, 'epoch': 0.13}) +[2026-06-17 06:54:35,034][starVLA.training.train_starvla][INFO] - Step 453, Loss: {'train/loss': 1.2250643968582153, 'train/grad_norm_pre_clip': 7.754754543304443, 'timing/data': 0.005042767035774887, 'timing/model': 1.3026639560703188, 'learning_rate/qwen_vl_interface': 9.938781947249417e-06, 'learning_rate/action_model': 9.938781947249416e-05, 'epoch': 0.13}) +[2026-06-17 06:54:39,730][starVLA.training.train_starvla][INFO] - Step 454, Loss: {'train/loss': 1.264979362487793, 'train/grad_norm_pre_clip': 8.219212532043457, 'timing/data': 0.004692172980867326, 'timing/model': 1.2247486389242113, 'learning_rate/qwen_vl_interface': 9.93843536372518e-06, 'learning_rate/action_model': 9.93843536372518e-05, 'epoch': 0.13}) +[2026-06-17 06:54:44,538][starVLA.training.train_starvla][INFO] - Step 455, Loss: {'train/loss': 1.1256206035614014, 'train/grad_norm_pre_clip': 7.915899753570557, 'timing/data': 0.004984450992196798, 'timing/model': 1.2127052010037005, 'learning_rate/qwen_vl_interface': 9.938087808283248e-06, 'learning_rate/action_model': 9.938087808283248e-05, 'epoch': 0.13}) +[2026-06-17 06:54:49,443][starVLA.training.train_starvla][INFO] - Step 456, Loss: {'train/loss': 1.151829719543457, 'train/grad_norm_pre_clip': 8.505342483520508, 'timing/data': 0.004922557040117681, 'timing/model': 1.2108533189166337, 'learning_rate/qwen_vl_interface': 9.937739280995672e-06, 'learning_rate/action_model': 9.937739280995672e-05, 'epoch': 0.13}) +[2026-06-17 06:54:54,257][starVLA.training.train_starvla][INFO] - Step 457, Loss: {'train/loss': 1.1806479692459106, 'train/grad_norm_pre_clip': 7.589693069458008, 'timing/data': 0.004887292045168579, 'timing/model': 1.2180139709962532, 'learning_rate/qwen_vl_interface': 9.937389781934697e-06, 'learning_rate/action_model': 9.937389781934696e-05, 'epoch': 0.13}) +[2026-06-17 06:54:59,050][starVLA.training.train_starvla][INFO] - Step 458, Loss: {'train/loss': 1.049741268157959, 'train/grad_norm_pre_clip': 5.209029197692871, 'timing/data': 0.004846152965910733, 'timing/model': 1.3335744290379807, 'learning_rate/qwen_vl_interface': 9.937039311172776e-06, 'learning_rate/action_model': 9.937039311172776e-05, 'epoch': 0.13}) +[2026-06-17 06:55:03,717][starVLA.training.train_starvla][INFO] - Step 459, Loss: {'train/loss': 1.3148527145385742, 'train/grad_norm_pre_clip': 6.230165481567383, 'timing/data': 0.005025555961765349, 'timing/model': 1.2093371800146997, 'learning_rate/qwen_vl_interface': 9.936687868782562e-06, 'learning_rate/action_model': 9.936687868782561e-05, 'epoch': 0.14}) +[2026-06-17 06:55:08,507][starVLA.training.train_starvla][INFO] - Step 460, Loss: {'train/loss': 0.8850060701370239, 'train/grad_norm_pre_clip': 7.129647731781006, 'timing/data': 0.004988287924788892, 'timing/model': 1.2077085840282962, 'learning_rate/qwen_vl_interface': 9.93633545483691e-06, 'learning_rate/action_model': 9.93633545483691e-05, 'epoch': 0.14}) +[2026-06-17 06:55:13,192][starVLA.training.train_starvla][INFO] - Step 461, Loss: {'train/loss': 1.2516887187957764, 'train/grad_norm_pre_clip': 6.428090572357178, 'timing/data': 0.005108725046738982, 'timing/model': 1.2073771400609985, 'learning_rate/qwen_vl_interface': 9.935982069408877e-06, 'learning_rate/action_model': 9.935982069408877e-05, 'epoch': 0.14}) +[2026-06-17 06:55:17,899][starVLA.training.train_starvla][INFO] - Step 462, Loss: {'train/loss': 1.044256567955017, 'train/grad_norm_pre_clip': 10.167860984802246, 'timing/data': 0.004760370007716119, 'timing/model': 1.2064252260606736, 'learning_rate/qwen_vl_interface': 9.935627712571716e-06, 'learning_rate/action_model': 9.935627712571715e-05, 'epoch': 0.14}) +[2026-06-17 06:55:22,641][starVLA.training.train_starvla][INFO] - Step 463, Loss: {'train/loss': 0.9742461442947388, 'train/grad_norm_pre_clip': 4.861021518707275, 'timing/data': 0.0049528180388733745, 'timing/model': 1.2065609350102022, 'learning_rate/qwen_vl_interface': 9.93527238439889e-06, 'learning_rate/action_model': 9.935272384398889e-05, 'epoch': 0.14}) +[2026-06-17 06:55:27,436][starVLA.training.train_starvla][INFO] - Step 464, Loss: {'train/loss': 1.27690589427948, 'train/grad_norm_pre_clip': 9.88356876373291, 'timing/data': 0.005038585979491472, 'timing/model': 1.3376620300114155, 'learning_rate/qwen_vl_interface': 9.934916084964056e-06, 'learning_rate/action_model': 9.934916084964056e-05, 'epoch': 0.14}) +[2026-06-17 06:55:32,106][starVLA.training.train_starvla][INFO] - Step 465, Loss: {'train/loss': 1.1644258499145508, 'train/grad_norm_pre_clip': 8.986367225646973, 'timing/data': 0.005121253081597388, 'timing/model': 1.207489310996607, 'learning_rate/qwen_vl_interface': 9.934558814341077e-06, 'learning_rate/action_model': 9.934558814341077e-05, 'epoch': 0.14}) +[2026-06-17 06:55:36,975][starVLA.training.train_starvla][INFO] - Step 466, Loss: {'train/loss': 1.196763038635254, 'train/grad_norm_pre_clip': 21.50958251953125, 'timing/data': 0.0049376440001651645, 'timing/model': 1.2676450150320306, 'learning_rate/qwen_vl_interface': 9.934200572604015e-06, 'learning_rate/action_model': 9.934200572604015e-05, 'epoch': 0.14}) +[2026-06-17 06:55:41,795][starVLA.training.train_starvla][INFO] - Step 467, Loss: {'train/loss': 0.9230258464813232, 'train/grad_norm_pre_clip': 6.658090114593506, 'timing/data': 0.00504379509948194, 'timing/model': 1.2068202670197934, 'learning_rate/qwen_vl_interface': 9.933841359827133e-06, 'learning_rate/action_model': 9.933841359827133e-05, 'epoch': 0.14}) +[2026-06-17 06:55:46,580][starVLA.training.train_starvla][INFO] - Step 468, Loss: {'train/loss': 0.981959879398346, 'train/grad_norm_pre_clip': 6.2242112159729, 'timing/data': 0.0047857260797172785, 'timing/model': 1.211697281920351, 'learning_rate/qwen_vl_interface': 9.933481176084899e-06, 'learning_rate/action_model': 9.933481176084899e-05, 'epoch': 0.14}) +[2026-06-17 06:55:51,358][starVLA.training.train_starvla][INFO] - Step 469, Loss: {'train/loss': 1.0723094940185547, 'train/grad_norm_pre_clip': 7.671039581298828, 'timing/data': 0.004765399964526296, 'timing/model': 1.2146214758977294, 'learning_rate/qwen_vl_interface': 9.933120021451976e-06, 'learning_rate/action_model': 9.933120021451976e-05, 'epoch': 0.14}) +[2026-06-17 06:55:56,188][starVLA.training.train_starvla][INFO] - Step 470, Loss: {'train/loss': 0.936403214931488, 'train/grad_norm_pre_clip': 5.9666619300842285, 'timing/data': 0.006752568995580077, 'timing/model': 1.3528878280194476, 'learning_rate/qwen_vl_interface': 9.932757896003235e-06, 'learning_rate/action_model': 9.932757896003235e-05, 'epoch': 0.14}) +[2026-06-17 06:56:00,860][starVLA.training.train_starvla][INFO] - Step 471, Loss: {'train/loss': 0.9362052083015442, 'train/grad_norm_pre_clip': 5.91493558883667, 'timing/data': 0.005187037982977927, 'timing/model': 1.2123490739613771, 'learning_rate/qwen_vl_interface': 9.932394799813742e-06, 'learning_rate/action_model': 9.932394799813741e-05, 'epoch': 0.14}) +[2026-06-17 06:56:05,658][starVLA.training.train_starvla][INFO] - Step 472, Loss: {'train/loss': 1.2402076721191406, 'train/grad_norm_pre_clip': 13.42149829864502, 'timing/data': 0.0050689768977463245, 'timing/model': 1.2075869879918173, 'learning_rate/qwen_vl_interface': 9.932030732958768e-06, 'learning_rate/action_model': 9.932030732958769e-05, 'epoch': 0.14}) +[2026-06-17 06:56:10,464][starVLA.training.train_starvla][INFO] - Step 473, Loss: {'train/loss': 1.3049691915512085, 'train/grad_norm_pre_clip': 11.570467948913574, 'timing/data': 0.00523499003611505, 'timing/model': 1.2068439859431237, 'learning_rate/qwen_vl_interface': 9.931665695513786e-06, 'learning_rate/action_model': 9.931665695513786e-05, 'epoch': 0.14}) +[2026-06-17 06:56:15,253][starVLA.training.train_starvla][INFO] - Step 474, Loss: {'train/loss': 1.2465944290161133, 'train/grad_norm_pre_clip': 7.7897515296936035, 'timing/data': 0.004865135997533798, 'timing/model': 1.204628037987277, 'learning_rate/qwen_vl_interface': 9.931299687554466e-06, 'learning_rate/action_model': 9.931299687554466e-05, 'epoch': 0.14}) +[2026-06-17 06:56:20,030][starVLA.training.train_starvla][INFO] - Step 475, Loss: {'train/loss': 1.3933825492858887, 'train/grad_norm_pre_clip': 14.513564109802246, 'timing/data': 0.004734973888844252, 'timing/model': 1.206359778996557, 'learning_rate/qwen_vl_interface': 9.930932709156686e-06, 'learning_rate/action_model': 9.930932709156685e-05, 'epoch': 0.14}) +[2026-06-17 06:56:24,855][starVLA.training.train_starvla][INFO] - Step 476, Loss: {'train/loss': 1.0624078512191772, 'train/grad_norm_pre_clip': 11.724912643432617, 'timing/data': 0.005238986108452082, 'timing/model': 1.3403554499382153, 'learning_rate/qwen_vl_interface': 9.930564760396518e-06, 'learning_rate/action_model': 9.930564760396518e-05, 'epoch': 0.14}) +[2026-06-17 06:56:29,518][starVLA.training.train_starvla][INFO] - Step 477, Loss: {'train/loss': 0.956301748752594, 'train/grad_norm_pre_clip': 10.542461395263672, 'timing/data': 0.005052714026533067, 'timing/model': 1.2073872310575098, 'learning_rate/qwen_vl_interface': 9.930195841350237e-06, 'learning_rate/action_model': 9.930195841350236e-05, 'epoch': 0.14}) +[2026-06-17 06:56:34,327][starVLA.training.train_starvla][INFO] - Step 478, Loss: {'train/loss': 1.1976757049560547, 'train/grad_norm_pre_clip': 13.035061836242676, 'timing/data': 0.004847711999900639, 'timing/model': 1.206412723986432, 'learning_rate/qwen_vl_interface': 9.929825952094323e-06, 'learning_rate/action_model': 9.929825952094323e-05, 'epoch': 0.14}) +[2026-06-17 06:56:39,131][starVLA.training.train_starvla][INFO] - Step 479, Loss: {'train/loss': 1.2341653108596802, 'train/grad_norm_pre_clip': 8.857062339782715, 'timing/data': 0.00494762696325779, 'timing/model': 1.2100715240230784, 'learning_rate/qwen_vl_interface': 9.929455092705453e-06, 'learning_rate/action_model': 9.929455092705453e-05, 'epoch': 0.14}) +[2026-06-17 06:56:43,871][starVLA.training.train_starvla][INFO] - Step 480, Loss: {'train/loss': 1.4980764389038086, 'train/grad_norm_pre_clip': 17.657527923583984, 'timing/data': 0.00497500307392329, 'timing/model': 1.207775860093534, 'learning_rate/qwen_vl_interface': 9.929083263260508e-06, 'learning_rate/action_model': 9.929083263260507e-05, 'epoch': 0.14}) +[2026-06-17 06:56:48,684][starVLA.training.train_starvla][INFO] - Step 481, Loss: {'train/loss': 0.9764271974563599, 'train/grad_norm_pre_clip': 8.99547290802002, 'timing/data': 0.005040068994276226, 'timing/model': 1.21948423597496, 'learning_rate/qwen_vl_interface': 9.928710463836566e-06, 'learning_rate/action_model': 9.928710463836565e-05, 'epoch': 0.14}) +[2026-06-17 06:56:53,479][starVLA.training.train_starvla][INFO] - Step 482, Loss: {'train/loss': 1.4138270616531372, 'train/grad_norm_pre_clip': 14.590377807617188, 'timing/data': 0.004982828977517784, 'timing/model': 1.3385900009889156, 'learning_rate/qwen_vl_interface': 9.928336694510912e-06, 'learning_rate/action_model': 9.928336694510911e-05, 'epoch': 0.14}) +[2026-06-17 06:56:58,149][starVLA.training.train_starvla][INFO] - Step 483, Loss: {'train/loss': 1.1716469526290894, 'train/grad_norm_pre_clip': 12.753214836120605, 'timing/data': 0.0047577370423823595, 'timing/model': 1.2058288829866797, 'learning_rate/qwen_vl_interface': 9.927961955361025e-06, 'learning_rate/action_model': 9.927961955361025e-05, 'epoch': 0.14}) +[2026-06-17 06:57:02,934][starVLA.training.train_starvla][INFO] - Step 484, Loss: {'train/loss': 1.060624122619629, 'train/grad_norm_pre_clip': 8.035353660583496, 'timing/data': 0.005031762993894517, 'timing/model': 1.207906209048815, 'learning_rate/qwen_vl_interface': 9.927586246464593e-06, 'learning_rate/action_model': 9.927586246464593e-05, 'epoch': 0.14}) +[2026-06-17 06:57:07,656][starVLA.training.train_starvla][INFO] - Step 485, Loss: {'train/loss': 0.9923862814903259, 'train/grad_norm_pre_clip': 9.505773544311523, 'timing/data': 0.004874772042967379, 'timing/model': 1.2060398160247132, 'learning_rate/qwen_vl_interface': 9.927209567899499e-06, 'learning_rate/action_model': 9.927209567899498e-05, 'epoch': 0.14}) +[2026-06-17 06:57:12,434][starVLA.training.train_starvla][INFO] - Step 486, Loss: {'train/loss': 1.1415268182754517, 'train/grad_norm_pre_clip': 11.986656188964844, 'timing/data': 0.005703179049305618, 'timing/model': 1.2082965220324695, 'learning_rate/qwen_vl_interface': 9.926831919743827e-06, 'learning_rate/action_model': 9.926831919743827e-05, 'epoch': 0.14}) +[2026-06-17 06:57:17,226][starVLA.training.train_starvla][INFO] - Step 487, Loss: {'train/loss': 1.0357857942581177, 'train/grad_norm_pre_clip': 11.002969741821289, 'timing/data': 0.005052778054960072, 'timing/model': 1.2071604779921472, 'learning_rate/qwen_vl_interface': 9.926453302075866e-06, 'learning_rate/action_model': 9.926453302075867e-05, 'epoch': 0.14}) +[2026-06-17 06:57:22,031][starVLA.training.train_starvla][INFO] - Step 488, Loss: {'train/loss': 1.1086831092834473, 'train/grad_norm_pre_clip': 9.184335708618164, 'timing/data': 0.0050539609510451555, 'timing/model': 1.3322144590783864, 'learning_rate/qwen_vl_interface': 9.926073714974103e-06, 'learning_rate/action_model': 9.926073714974104e-05, 'epoch': 0.14}) +[2026-06-17 06:57:26,697][starVLA.training.train_starvla][INFO] - Step 489, Loss: {'train/loss': 0.8922778964042664, 'train/grad_norm_pre_clip': 8.245501518249512, 'timing/data': 0.004906240967102349, 'timing/model': 1.208883014973253, 'learning_rate/qwen_vl_interface': 9.925693158517228e-06, 'learning_rate/action_model': 9.925693158517228e-05, 'epoch': 0.14}) +[2026-06-17 06:57:31,498][starVLA.training.train_starvla][INFO] - Step 490, Loss: {'train/loss': 1.2140570878982544, 'train/grad_norm_pre_clip': 7.558679580688477, 'timing/data': 0.004809091915376484, 'timing/model': 1.2153932910878211, 'learning_rate/qwen_vl_interface': 9.92531163278413e-06, 'learning_rate/action_model': 9.92531163278413e-05, 'epoch': 0.14}) +[2026-06-17 06:57:36,298][starVLA.training.train_starvla][INFO] - Step 491, Loss: {'train/loss': 0.9558364748954773, 'train/grad_norm_pre_clip': 7.73134708404541, 'timing/data': 0.005118403001688421, 'timing/model': 1.206728248973377, 'learning_rate/qwen_vl_interface': 9.924929137853899e-06, 'learning_rate/action_model': 9.924929137853898e-05, 'epoch': 0.14}) +[2026-06-17 06:57:41,060][starVLA.training.train_starvla][INFO] - Step 492, Loss: {'train/loss': 1.0847421884536743, 'train/grad_norm_pre_clip': 9.176589012145996, 'timing/data': 0.004677767981775105, 'timing/model': 1.2073385330149904, 'learning_rate/qwen_vl_interface': 9.92454567380583e-06, 'learning_rate/action_model': 9.924545673805829e-05, 'epoch': 0.14}) +[2026-06-17 06:57:45,849][starVLA.training.train_starvla][INFO] - Step 493, Loss: {'train/loss': 1.26780104637146, 'train/grad_norm_pre_clip': 7.584985256195068, 'timing/data': 0.004813338979147375, 'timing/model': 1.2071440849686041, 'learning_rate/qwen_vl_interface': 9.924161240719409e-06, 'learning_rate/action_model': 9.924161240719408e-05, 'epoch': 0.15}) +[2026-06-17 06:57:50,689][starVLA.training.train_starvla][INFO] - Step 494, Loss: {'train/loss': 1.2396252155303955, 'train/grad_norm_pre_clip': 8.871820449829102, 'timing/data': 0.005000830045901239, 'timing/model': 1.3439783670473844, 'learning_rate/qwen_vl_interface': 9.923775838674336e-06, 'learning_rate/action_model': 9.923775838674335e-05, 'epoch': 0.15}) +[2026-06-17 06:57:55,359][starVLA.training.train_starvla][INFO] - Step 495, Loss: {'train/loss': 0.9175165295600891, 'train/grad_norm_pre_clip': 5.8695173263549805, 'timing/data': 0.004884385038167238, 'timing/model': 1.210770815028809, 'learning_rate/qwen_vl_interface': 9.9233894677505e-06, 'learning_rate/action_model': 9.9233894677505e-05, 'epoch': 0.15}) +[2026-06-17 06:58:00,151][starVLA.training.train_starvla][INFO] - Step 496, Loss: {'train/loss': 1.2613739967346191, 'train/grad_norm_pre_clip': 9.284224510192871, 'timing/data': 0.004779067006893456, 'timing/model': 1.2085586580215022, 'learning_rate/qwen_vl_interface': 9.923002128028e-06, 'learning_rate/action_model': 9.923002128028e-05, 'epoch': 0.15}) +[2026-06-17 06:58:04,931][starVLA.training.train_starvla][INFO] - Step 497, Loss: {'train/loss': 1.2806774377822876, 'train/grad_norm_pre_clip': 7.9183454513549805, 'timing/data': 0.00493138306774199, 'timing/model': 1.2071239309152588, 'learning_rate/qwen_vl_interface': 9.922613819587131e-06, 'learning_rate/action_model': 9.92261381958713e-05, 'epoch': 0.15}) +[2026-06-17 06:58:09,743][starVLA.training.train_starvla][INFO] - Step 498, Loss: {'train/loss': 1.1436654329299927, 'train/grad_norm_pre_clip': 6.545547008514404, 'timing/data': 0.004753548069857061, 'timing/model': 1.2075543910032138, 'learning_rate/qwen_vl_interface': 9.922224542508389e-06, 'learning_rate/action_model': 9.922224542508388e-05, 'epoch': 0.15}) +[2026-06-17 06:58:14,543][starVLA.training.train_starvla][INFO] - Step 499, Loss: {'train/loss': 0.9864628314971924, 'train/grad_norm_pre_clip': 6.179842948913574, 'timing/data': 0.004701850935816765, 'timing/model': 1.2068041709717363, 'learning_rate/qwen_vl_interface': 9.92183429687247e-06, 'learning_rate/action_model': 9.92183429687247e-05, 'epoch': 0.15}) +[2026-06-17 07:07:44,734][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 500 +[2026-06-17 07:07:44,735][starVLA.training.train_starvla][INFO] - Step 500, Loss: {'train/loss': 1.230555534362793, 'train/grad_norm_pre_clip': 8.965926170349121, 'eval/loss': 1.0706918239593506, 'eval/latency_0/loss': 1.0706918239593506, 'eval/demon_fire_precision': 0.8669596484218938, 'eval/demon_fire_recall': 0.9011627906976745, 'eval/demon_fire_f1': 0.8837304011402973, 'eval/demon_fire_rate_model': 0.6960511679644049, 'eval/demon_fire_rate_teacher': 0.6696329254727474, 'eval/demon_left_precision': 0.7018528916339135, 'eval/demon_left_recall': 0.7082152974504249, 'eval/demon_left_f1': 0.7050197405527355, 'eval/demon_left_rate_model': 0.49527252502780866, 'eval/demon_left_rate_teacher': 0.49082313681868744, 'eval/demon_right_precision': 0.7002754820936639, 'eval/demon_right_recall': 0.7180790960451977, 'eval/demon_right_f1': 0.7090655509065551, 'eval/demon_right_rate_model': 0.5047274749721913, 'eval/demon_right_rate_teacher': 0.4922135706340378, 'eval/demon_move_f1': 0.7070426457296453, 'eval/demon_cc_f1': 0.8220984484719316, 'eval/latency_0/demon_fire_precision': 0.8669596484218938, 'eval/latency_0/demon_fire_recall': 0.9011627906976745, 'eval/latency_0/demon_fire_f1': 0.8837304011402973, 'eval/latency_0/demon_fire_rate_model': 0.6960511679644049, 'eval/latency_0/demon_fire_rate_teacher': 0.6696329254727474, 'eval/latency_0/demon_left_precision': 0.7018528916339135, 'eval/latency_0/demon_left_recall': 0.7082152974504249, 'eval/latency_0/demon_left_f1': 0.7050197405527355, 'eval/latency_0/demon_left_rate_model': 0.49527252502780866, 'eval/latency_0/demon_left_rate_teacher': 0.49082313681868744, 'eval/latency_0/demon_right_precision': 0.7002754820936639, 'eval/latency_0/demon_right_recall': 0.7180790960451977, 'eval/latency_0/demon_right_f1': 0.7090655509065551, 'eval/latency_0/demon_right_rate_model': 0.5047274749721913, 'eval/latency_0/demon_right_rate_teacher': 0.4922135706340378, 'eval/latency_0/demon_move_f1': 0.7070426457296453, 'eval/latency_0/demon_cc_f1': 0.8220984484719316, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 4306.0, 'rl_games_eval/mid_train/mean_length': 3311.8, 'rl_games_eval/mid_train/std_reward': 1051.1060840847606, 'rl_games_eval/mid_train/std_length': 101.09678530991972, 'rl_games_eval/mid_train/macro_mean_reward': 4306.0, 'rl_games_eval/mid_train/macro_mean_length': 3311.8, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/demon_attack__latency_0/mean_reward': 4306.0, 'rl_games_eval/mid_train/demon_attack__latency_0/mean_length': 3311.8, 'rl_games_eval/mid_train/demon_attack__latency_0/std_reward': 1051.1060840847606, 'rl_games_eval/mid_train/demon_attack__latency_0/std_length': 101.09678530991972, 'rl_games_eval/mid_train/demon_attack__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 4306.0, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.004475686000660062, 'timing/model': 1.340637827059254, 'learning_rate/qwen_vl_interface': 9.921443082760276e-06, 'learning_rate/action_model': 9.921443082760276e-05, 'epoch': 0.15}) +[2026-06-17 07:07:44,735][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_500_state +[2026-06-17 07:07:44,738][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 07:07:48,282][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_500_state/model.safetensors +[2026-06-17 07:07:56,065][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_500_state/optimizer.bin +[2026-06-17 07:07:56,065][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_500_state/sampler.bin +[2026-06-17 07:07:56,065][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_500_state/sampler_1.bin +[2026-06-17 07:07:56,066][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_500_state/random_states_0.pkl +[2026-06-17 07:07:56,102][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 500 +[2026-06-17 07:07:56,102][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 07:07:56,128][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 07:08:00,696][starVLA.training.train_starvla][INFO] - Step 501, Loss: {'train/loss': 1.3056628704071045, 'train/grad_norm_pre_clip': 6.87960147857666, 'timing/data': 0.004241509013809264, 'timing/model': 1.1842883150093257, 'learning_rate/qwen_vl_interface': 9.921050900252904e-06, 'learning_rate/action_model': 9.921050900252903e-05, 'epoch': 0.15}) +[2026-06-17 07:08:05,411][starVLA.training.train_starvla][INFO] - Step 502, Loss: {'train/loss': 0.9103166460990906, 'train/grad_norm_pre_clip': 5.23355770111084, 'timing/data': 0.005245610955171287, 'timing/model': 1.1884390539489686, 'learning_rate/qwen_vl_interface': 9.920657749431654e-06, 'learning_rate/action_model': 9.920657749431654e-05, 'epoch': 0.15}) +[2026-06-17 07:08:10,123][starVLA.training.train_starvla][INFO] - Step 503, Loss: {'train/loss': 1.0474817752838135, 'train/grad_norm_pre_clip': 6.197608947753906, 'timing/data': 0.004379974096082151, 'timing/model': 1.1915593679295853, 'learning_rate/qwen_vl_interface': 9.920263630378026e-06, 'learning_rate/action_model': 9.920263630378024e-05, 'epoch': 0.15}) +[2026-06-17 07:08:14,854][starVLA.training.train_starvla][INFO] - Step 504, Loss: {'train/loss': 1.007947325706482, 'train/grad_norm_pre_clip': 6.565374851226807, 'timing/data': 0.004704902064986527, 'timing/model': 1.1901926090940833, 'learning_rate/qwen_vl_interface': 9.919868543173722e-06, 'learning_rate/action_model': 9.919868543173721e-05, 'epoch': 0.15}) +[2026-06-17 07:08:19,668][starVLA.training.train_starvla][INFO] - Step 505, Loss: {'train/loss': 0.9207340478897095, 'train/grad_norm_pre_clip': 7.200809478759766, 'timing/data': 0.004963960964232683, 'timing/model': 1.2008494000183418, 'learning_rate/qwen_vl_interface': 9.919472487900645e-06, 'learning_rate/action_model': 9.919472487900644e-05, 'epoch': 0.15}) +[2026-06-17 07:08:24,430][starVLA.training.train_starvla][INFO] - Step 506, Loss: {'train/loss': 1.091787576675415, 'train/grad_norm_pre_clip': 6.751642227172852, 'timing/data': 0.005044505000114441, 'timing/model': 1.3261158090317622, 'learning_rate/qwen_vl_interface': 9.919075464640895e-06, 'learning_rate/action_model': 9.919075464640895e-05, 'epoch': 0.15}) +[2026-06-17 07:08:29,080][starVLA.training.train_starvla][INFO] - Step 507, Loss: {'train/loss': 0.9468942880630493, 'train/grad_norm_pre_clip': 8.291748046875, 'timing/data': 0.005552773014642298, 'timing/model': 1.2011899170465767, 'learning_rate/qwen_vl_interface': 9.91867747347678e-06, 'learning_rate/action_model': 9.918677473476779e-05, 'epoch': 0.15}) +[2026-06-17 07:08:33,847][starVLA.training.train_starvla][INFO] - Step 508, Loss: {'train/loss': 0.9971381425857544, 'train/grad_norm_pre_clip': 5.971292972564697, 'timing/data': 0.0047426020028069615, 'timing/model': 1.2021653089905158, 'learning_rate/qwen_vl_interface': 9.918278514490798e-06, 'learning_rate/action_model': 9.918278514490799e-05, 'epoch': 0.15}) +[2026-06-17 07:08:38,605][starVLA.training.train_starvla][INFO] - Step 509, Loss: {'train/loss': 1.135199785232544, 'train/grad_norm_pre_clip': 6.95456075668335, 'timing/data': 0.004654553020372987, 'timing/model': 1.198532111942768, 'learning_rate/qwen_vl_interface': 9.917878587765658e-06, 'learning_rate/action_model': 9.917878587765658e-05, 'epoch': 0.15}) +[2026-06-17 07:08:43,390][starVLA.training.train_starvla][INFO] - Step 510, Loss: {'train/loss': 0.9729202389717102, 'train/grad_norm_pre_clip': 5.354851245880127, 'timing/data': 0.004430242930538952, 'timing/model': 1.204656824003905, 'learning_rate/qwen_vl_interface': 9.917477693384264e-06, 'learning_rate/action_model': 9.917477693384264e-05, 'epoch': 0.15}) +[2026-06-17 07:08:48,168][starVLA.training.train_starvla][INFO] - Step 511, Loss: {'train/loss': 1.0409177541732788, 'train/grad_norm_pre_clip': 6.443646430969238, 'timing/data': 0.00491799192968756, 'timing/model': 1.3414532319875434, 'learning_rate/qwen_vl_interface': 9.917075831429721e-06, 'learning_rate/action_model': 9.917075831429722e-05, 'epoch': 0.15}) +[2026-06-17 07:08:52,918][starVLA.training.train_starvla][INFO] - Step 512, Loss: {'train/loss': 1.0065617561340332, 'train/grad_norm_pre_clip': 5.78425931930542, 'timing/data': 0.005042843986302614, 'timing/model': 1.3063524859026074, 'learning_rate/qwen_vl_interface': 9.916673001985339e-06, 'learning_rate/action_model': 9.916673001985338e-05, 'epoch': 0.15}) +[2026-06-17 07:08:57,555][starVLA.training.train_starvla][INFO] - Step 513, Loss: {'train/loss': 0.9412546753883362, 'train/grad_norm_pre_clip': 8.368058204650879, 'timing/data': 0.004742309916764498, 'timing/model': 1.201900337007828, 'learning_rate/qwen_vl_interface': 9.91626920513462e-06, 'learning_rate/action_model': 9.916269205134619e-05, 'epoch': 0.15}) +[2026-06-17 07:09:02,319][starVLA.training.train_starvla][INFO] - Step 514, Loss: {'train/loss': 1.101610541343689, 'train/grad_norm_pre_clip': 7.934214115142822, 'timing/data': 0.0045556360855698586, 'timing/model': 1.1988033819943666, 'learning_rate/qwen_vl_interface': 9.915864440961272e-06, 'learning_rate/action_model': 9.915864440961272e-05, 'epoch': 0.15}) +[2026-06-17 07:09:07,083][starVLA.training.train_starvla][INFO] - Step 515, Loss: {'train/loss': 1.0151782035827637, 'train/grad_norm_pre_clip': 8.156209945678711, 'timing/data': 0.004461635020561516, 'timing/model': 1.2003559119766578, 'learning_rate/qwen_vl_interface': 9.915458709549205e-06, 'learning_rate/action_model': 9.915458709549206e-05, 'epoch': 0.15}) +[2026-06-17 07:09:11,857][starVLA.training.train_starvla][INFO] - Step 516, Loss: {'train/loss': 0.9850828051567078, 'train/grad_norm_pre_clip': 8.180912971496582, 'timing/data': 0.004610336967743933, 'timing/model': 1.202214602031745, 'learning_rate/qwen_vl_interface': 9.91505201098253e-06, 'learning_rate/action_model': 9.91505201098253e-05, 'epoch': 0.15}) +[2026-06-17 07:09:16,583][starVLA.training.train_starvla][INFO] - Step 517, Loss: {'train/loss': 1.016266942024231, 'train/grad_norm_pre_clip': 5.969072341918945, 'timing/data': 0.004630881012417376, 'timing/model': 1.2018548130290583, 'learning_rate/qwen_vl_interface': 9.914644345345551e-06, 'learning_rate/action_model': 9.914644345345552e-05, 'epoch': 0.15}) +[2026-06-17 07:09:21,356][starVLA.training.train_starvla][INFO] - Step 518, Loss: {'train/loss': 1.196736454963684, 'train/grad_norm_pre_clip': 9.453055381774902, 'timing/data': 0.004765277029946446, 'timing/model': 1.3313583550043404, 'learning_rate/qwen_vl_interface': 9.914235712722782e-06, 'learning_rate/action_model': 9.914235712722781e-05, 'epoch': 0.15}) +[2026-06-17 07:09:25,999][starVLA.training.train_starvla][INFO] - Step 519, Loss: {'train/loss': 0.8210621476173401, 'train/grad_norm_pre_clip': 5.993118762969971, 'timing/data': 0.004448694991879165, 'timing/model': 1.2010038529988378, 'learning_rate/qwen_vl_interface': 9.913826113198927e-06, 'learning_rate/action_model': 9.913826113198927e-05, 'epoch': 0.15}) +[2026-06-17 07:09:30,771][starVLA.training.train_starvla][INFO] - Step 520, Loss: {'train/loss': 1.231560230255127, 'train/grad_norm_pre_clip': 8.263551712036133, 'timing/data': 0.0042930919444188476, 'timing/model': 1.2048300419701263, 'learning_rate/qwen_vl_interface': 9.913415546858904e-06, 'learning_rate/action_model': 9.913415546858904e-05, 'epoch': 0.15}) +[2026-06-17 07:09:35,558][starVLA.training.train_starvla][INFO] - Step 521, Loss: {'train/loss': 1.1342419385910034, 'train/grad_norm_pre_clip': 19.369726181030273, 'timing/data': 0.004787852987647057, 'timing/model': 1.2021972360089421, 'learning_rate/qwen_vl_interface': 9.913004013787818e-06, 'learning_rate/action_model': 9.913004013787819e-05, 'epoch': 0.15}) +[2026-06-17 07:09:40,321][starVLA.training.train_starvla][INFO] - Step 522, Loss: {'train/loss': 1.0255961418151855, 'train/grad_norm_pre_clip': 11.005826950073242, 'timing/data': 0.004498940077610314, 'timing/model': 1.2011422789655626, 'learning_rate/qwen_vl_interface': 9.912591514070983e-06, 'learning_rate/action_model': 9.912591514070983e-05, 'epoch': 0.15}) +[2026-06-17 07:09:45,059][starVLA.training.train_starvla][INFO] - Step 523, Loss: {'train/loss': 1.1499214172363281, 'train/grad_norm_pre_clip': 7.717512607574463, 'timing/data': 0.004743669065646827, 'timing/model': 1.2063109679147601, 'learning_rate/qwen_vl_interface': 9.912178047793912e-06, 'learning_rate/action_model': 9.91217804779391e-05, 'epoch': 0.15}) +[2026-06-17 07:09:49,828][starVLA.training.train_starvla][INFO] - Step 524, Loss: {'train/loss': 1.2364321947097778, 'train/grad_norm_pre_clip': 11.439546585083008, 'timing/data': 0.004922041087411344, 'timing/model': 1.323291547014378, 'learning_rate/qwen_vl_interface': 9.911763615042311e-06, 'learning_rate/action_model': 9.91176361504231e-05, 'epoch': 0.15}) +[2026-06-17 07:09:54,480][starVLA.training.train_starvla][INFO] - Step 525, Loss: {'train/loss': 1.31051504611969, 'train/grad_norm_pre_clip': 9.732528686523438, 'timing/data': 0.004729248001240194, 'timing/model': 1.2035930970450863, 'learning_rate/qwen_vl_interface': 9.911348215902099e-06, 'learning_rate/action_model': 9.911348215902099e-05, 'epoch': 0.15}) +[2026-06-17 07:09:59,257][starVLA.training.train_starvla][INFO] - Step 526, Loss: {'train/loss': 0.9074405431747437, 'train/grad_norm_pre_clip': 4.829401969909668, 'timing/data': 0.004447802086360753, 'timing/model': 1.2033600480062887, 'learning_rate/qwen_vl_interface': 9.910931850459385e-06, 'learning_rate/action_model': 9.910931850459385e-05, 'epoch': 0.15}) +[2026-06-17 07:10:04,046][starVLA.training.train_starvla][INFO] - Step 527, Loss: {'train/loss': 0.9743651151657104, 'train/grad_norm_pre_clip': 5.9124531745910645, 'timing/data': 0.004572256933897734, 'timing/model': 1.206239715917036, 'learning_rate/qwen_vl_interface': 9.910514518800485e-06, 'learning_rate/action_model': 9.910514518800484e-05, 'epoch': 0.16}) +[2026-06-17 07:10:08,836][starVLA.training.train_starvla][INFO] - Step 528, Loss: {'train/loss': 1.0629180669784546, 'train/grad_norm_pre_clip': 11.173080444335938, 'timing/data': 0.004800050985068083, 'timing/model': 1.2081326059997082, 'learning_rate/qwen_vl_interface': 9.910096221011908e-06, 'learning_rate/action_model': 9.910096221011907e-05, 'epoch': 0.16}) +[2026-06-17 07:10:13,626][starVLA.training.train_starvla][INFO] - Step 529, Loss: {'train/loss': 1.3925036191940308, 'train/grad_norm_pre_clip': 10.136930465698242, 'timing/data': 0.004381499951705337, 'timing/model': 1.3356573090422899, 'learning_rate/qwen_vl_interface': 9.90967695718037e-06, 'learning_rate/action_model': 9.90967695718037e-05, 'epoch': 0.16}) +[2026-06-17 07:10:18,279][starVLA.training.train_starvla][INFO] - Step 530, Loss: {'train/loss': 0.8387643694877625, 'train/grad_norm_pre_clip': 8.052203178405762, 'timing/data': 0.0050361460307613015, 'timing/model': 1.20495895401109, 'learning_rate/qwen_vl_interface': 9.909256727392786e-06, 'learning_rate/action_model': 9.909256727392784e-05, 'epoch': 0.16}) +[2026-06-17 07:10:23,058][starVLA.training.train_starvla][INFO] - Step 531, Loss: {'train/loss': 1.0324020385742188, 'train/grad_norm_pre_clip': 9.509175300598145, 'timing/data': 0.0047972750617191195, 'timing/model': 1.2044296069070697, 'learning_rate/qwen_vl_interface': 9.908835531736269e-06, 'learning_rate/action_model': 9.908835531736267e-05, 'epoch': 0.16}) +[2026-06-17 07:10:27,754][starVLA.training.train_starvla][INFO] - Step 532, Loss: {'train/loss': 1.478687047958374, 'train/grad_norm_pre_clip': 27.726394653320312, 'timing/data': 0.004678595927543938, 'timing/model': 1.2037610489642248, 'learning_rate/qwen_vl_interface': 9.908413370298131e-06, 'learning_rate/action_model': 9.90841337029813e-05, 'epoch': 0.16}) +[2026-06-17 07:10:32,484][starVLA.training.train_starvla][INFO] - Step 533, Loss: {'train/loss': 1.1007050275802612, 'train/grad_norm_pre_clip': 6.675577640533447, 'timing/data': 0.004421988967806101, 'timing/model': 1.204344401950948, 'learning_rate/qwen_vl_interface': 9.90799024316589e-06, 'learning_rate/action_model': 9.90799024316589e-05, 'epoch': 0.16}) +[2026-06-17 07:10:37,273][starVLA.training.train_starvla][INFO] - Step 534, Loss: {'train/loss': 1.0887678861618042, 'train/grad_norm_pre_clip': 7.919898509979248, 'timing/data': 0.00398187804967165, 'timing/model': 1.203333823941648, 'learning_rate/qwen_vl_interface': 9.90756615042726e-06, 'learning_rate/action_model': 9.907566150427259e-05, 'epoch': 0.16}) +[2026-06-17 07:10:42,069][starVLA.training.train_starvla][INFO] - Step 535, Loss: {'train/loss': 1.0920768976211548, 'train/grad_norm_pre_clip': 7.292707443237305, 'timing/data': 0.004647976020351052, 'timing/model': 1.3388563679764047, 'learning_rate/qwen_vl_interface': 9.907141092170154e-06, 'learning_rate/action_model': 9.907141092170153e-05, 'epoch': 0.16}) +[2026-06-17 07:10:46,730][starVLA.training.train_starvla][INFO] - Step 536, Loss: {'train/loss': 1.3338903188705444, 'train/grad_norm_pre_clip': 8.107795715332031, 'timing/data': 0.004145438899286091, 'timing/model': 1.2068954019341618, 'learning_rate/qwen_vl_interface': 9.906715068482689e-06, 'learning_rate/action_model': 9.90671506848269e-05, 'epoch': 0.16}) +[2026-06-17 07:10:51,592][starVLA.training.train_starvla][INFO] - Step 537, Loss: {'train/loss': 1.236013650894165, 'train/grad_norm_pre_clip': 7.021608829498291, 'timing/data': 0.004153237910941243, 'timing/model': 1.2071825969032943, 'learning_rate/qwen_vl_interface': 9.90628807945318e-06, 'learning_rate/action_model': 9.90628807945318e-05, 'epoch': 0.16}) +[2026-06-17 07:10:56,380][starVLA.training.train_starvla][INFO] - Step 538, Loss: {'train/loss': 0.9185874462127686, 'train/grad_norm_pre_clip': 10.316178321838379, 'timing/data': 0.005234400974586606, 'timing/model': 1.206855823053047, 'learning_rate/qwen_vl_interface': 9.90586012517014e-06, 'learning_rate/action_model': 9.90586012517014e-05, 'epoch': 0.16}) +[2026-06-17 07:11:01,168][starVLA.training.train_starvla][INFO] - Step 539, Loss: {'train/loss': 1.0980939865112305, 'train/grad_norm_pre_clip': 9.407266616821289, 'timing/data': 0.004647864028811455, 'timing/model': 1.243232178967446, 'learning_rate/qwen_vl_interface': 9.905431205722288e-06, 'learning_rate/action_model': 9.905431205722288e-05, 'epoch': 0.16}) +[2026-06-17 07:11:05,959][starVLA.training.train_starvla][INFO] - Step 540, Loss: {'train/loss': 1.2582297325134277, 'train/grad_norm_pre_clip': 10.980615615844727, 'timing/data': 0.0046795200323686, 'timing/model': 1.2045956139918417, 'learning_rate/qwen_vl_interface': 9.905001321198536e-06, 'learning_rate/action_model': 9.905001321198536e-05, 'epoch': 0.16}) +[2026-06-17 07:11:10,765][starVLA.training.train_starvla][INFO] - Step 541, Loss: {'train/loss': 1.3350114822387695, 'train/grad_norm_pre_clip': 8.796216011047363, 'timing/data': 0.005069436971098185, 'timing/model': 1.3454294760013, 'learning_rate/qwen_vl_interface': 9.904570471688003e-06, 'learning_rate/action_model': 9.904570471688002e-05, 'epoch': 0.16}) +[2026-06-17 07:11:15,500][starVLA.training.train_starvla][INFO] - Step 542, Loss: {'train/loss': 0.9876981377601624, 'train/grad_norm_pre_clip': 6.377254486083984, 'timing/data': 0.004267846932634711, 'timing/model': 1.2094643220771104, 'learning_rate/qwen_vl_interface': 9.90413865728e-06, 'learning_rate/action_model': 9.90413865728e-05, 'epoch': 0.16}) +[2026-06-17 07:11:20,286][starVLA.training.train_starvla][INFO] - Step 543, Loss: {'train/loss': 1.1282598972320557, 'train/grad_norm_pre_clip': 7.659557342529297, 'timing/data': 0.0047738110879436135, 'timing/model': 1.2041281079873443, 'learning_rate/qwen_vl_interface': 9.903705878064047e-06, 'learning_rate/action_model': 9.903705878064046e-05, 'epoch': 0.16}) +[2026-06-17 07:11:25,074][starVLA.training.train_starvla][INFO] - Step 544, Loss: {'train/loss': 1.0028225183486938, 'train/grad_norm_pre_clip': 6.63109827041626, 'timing/data': 0.004304928006604314, 'timing/model': 1.2065110229887068, 'learning_rate/qwen_vl_interface': 9.903272134129858e-06, 'learning_rate/action_model': 9.903272134129858e-05, 'epoch': 0.16}) +[2026-06-17 07:11:29,910][starVLA.training.train_starvla][INFO] - Step 545, Loss: {'train/loss': 1.0094295740127563, 'train/grad_norm_pre_clip': 7.194832801818848, 'timing/data': 0.004177565919235349, 'timing/model': 1.2056380308931693, 'learning_rate/qwen_vl_interface': 9.90283742556735e-06, 'learning_rate/action_model': 9.902837425567349e-05, 'epoch': 0.16}) +[2026-06-17 07:11:34,582][starVLA.training.train_starvla][INFO] - Step 546, Loss: {'train/loss': 1.257512092590332, 'train/grad_norm_pre_clip': 7.220686435699463, 'timing/data': 0.004744856967590749, 'timing/model': 1.2032832010881975, 'learning_rate/qwen_vl_interface': 9.902401752466635e-06, 'learning_rate/action_model': 9.902401752466634e-05, 'epoch': 0.16}) +[2026-06-17 07:11:39,289][starVLA.training.train_starvla][INFO] - Step 547, Loss: {'train/loss': 1.0595613718032837, 'train/grad_norm_pre_clip': 5.521566867828369, 'timing/data': 0.004782119067385793, 'timing/model': 1.2454652320593596, 'learning_rate/qwen_vl_interface': 9.90196511491803e-06, 'learning_rate/action_model': 9.901965114918029e-05, 'epoch': 0.16}) +[2026-06-17 07:11:43,947][starVLA.training.train_starvla][INFO] - Step 548, Loss: {'train/loss': 0.9564540386199951, 'train/grad_norm_pre_clip': 6.082054615020752, 'timing/data': 0.004520965972915292, 'timing/model': 1.206672157975845, 'learning_rate/qwen_vl_interface': 9.901527513012053e-06, 'learning_rate/action_model': 9.901527513012052e-05, 'epoch': 0.16}) +[2026-06-17 07:11:48,689][starVLA.training.train_starvla][INFO] - Step 549, Loss: {'train/loss': 0.9711413383483887, 'train/grad_norm_pre_clip': 6.2222723960876465, 'timing/data': 0.004842560971155763, 'timing/model': 1.206687586964108, 'learning_rate/qwen_vl_interface': 9.901088946839415e-06, 'learning_rate/action_model': 9.901088946839414e-05, 'epoch': 0.16}) +[2026-06-17 07:11:53,477][starVLA.training.train_starvla][INFO] - Step 550, Loss: {'train/loss': 0.876756489276886, 'train/grad_norm_pre_clip': 7.551896572113037, 'timing/data': 0.0043625799007713795, 'timing/model': 1.20697061996907, 'learning_rate/qwen_vl_interface': 9.900649416491035e-06, 'learning_rate/action_model': 9.900649416491035e-05, 'epoch': 0.16}) +[2026-06-17 07:11:58,260][starVLA.training.train_starvla][INFO] - Step 551, Loss: {'train/loss': 1.0528069734573364, 'train/grad_norm_pre_clip': 6.935674667358398, 'timing/data': 0.004522838979028165, 'timing/model': 1.2054398889886215, 'learning_rate/qwen_vl_interface': 9.900208922058026e-06, 'learning_rate/action_model': 9.900208922058025e-05, 'epoch': 0.16}) +[2026-06-17 07:12:03,046][starVLA.training.train_starvla][INFO] - Step 552, Loss: {'train/loss': 0.8872196674346924, 'train/grad_norm_pre_clip': 6.495443820953369, 'timing/data': 0.004786155070178211, 'timing/model': 1.331168586970307, 'learning_rate/qwen_vl_interface': 9.899767463631704e-06, 'learning_rate/action_model': 9.899767463631704e-05, 'epoch': 0.16}) +[2026-06-17 07:12:07,848][starVLA.training.train_starvla][INFO] - Step 553, Loss: {'train/loss': 1.0742344856262207, 'train/grad_norm_pre_clip': 8.131888389587402, 'timing/data': 0.004349233000539243, 'timing/model': 1.3262720630737022, 'learning_rate/qwen_vl_interface': 9.899325041303582e-06, 'learning_rate/action_model': 9.899325041303583e-05, 'epoch': 0.16}) +[2026-06-17 07:12:12,561][starVLA.training.train_starvla][INFO] - Step 554, Loss: {'train/loss': 1.0913231372833252, 'train/grad_norm_pre_clip': 6.713127613067627, 'timing/data': 0.004883883986622095, 'timing/model': 1.2097651610383764, 'learning_rate/qwen_vl_interface': 9.89888165516538e-06, 'learning_rate/action_model': 9.898881655165379e-05, 'epoch': 0.16}) +[2026-06-17 07:12:17,317][starVLA.training.train_starvla][INFO] - Step 555, Loss: {'train/loss': 0.8853164315223694, 'train/grad_norm_pre_clip': 5.071136951446533, 'timing/data': 0.004872210090979934, 'timing/model': 1.213047747965902, 'learning_rate/qwen_vl_interface': 9.898437305309004e-06, 'learning_rate/action_model': 9.898437305309003e-05, 'epoch': 0.16}) +[2026-06-17 07:12:22,108][starVLA.training.train_starvla][INFO] - Step 556, Loss: {'train/loss': 0.9130797386169434, 'train/grad_norm_pre_clip': 8.47476577758789, 'timing/data': 0.004485912970267236, 'timing/model': 1.2067972850054502, 'learning_rate/qwen_vl_interface': 9.897991991826575e-06, 'learning_rate/action_model': 9.897991991826574e-05, 'epoch': 0.16}) +[2026-06-17 07:12:26,904][starVLA.training.train_starvla][INFO] - Step 557, Loss: {'train/loss': 0.8169929385185242, 'train/grad_norm_pre_clip': 10.394559860229492, 'timing/data': 0.004930316004902124, 'timing/model': 1.2052059270208701, 'learning_rate/qwen_vl_interface': 9.897545714810404e-06, 'learning_rate/action_model': 9.897545714810403e-05, 'epoch': 0.16}) +[2026-06-17 07:12:31,689][starVLA.training.train_starvla][INFO] - Step 558, Loss: {'train/loss': 1.2244203090667725, 'train/grad_norm_pre_clip': 11.743326187133789, 'timing/data': 0.0046980270417407155, 'timing/model': 1.332890360034071, 'learning_rate/qwen_vl_interface': 9.897098474353004e-06, 'learning_rate/action_model': 9.897098474353004e-05, 'epoch': 0.16}) +[2026-06-17 07:12:36,471][starVLA.training.train_starvla][INFO] - Step 559, Loss: {'train/loss': 1.1906918287277222, 'train/grad_norm_pre_clip': 27.664318084716797, 'timing/data': 0.004878810956142843, 'timing/model': 1.325776614015922, 'learning_rate/qwen_vl_interface': 9.896650270547092e-06, 'learning_rate/action_model': 9.89665027054709e-05, 'epoch': 0.16}) +[2026-06-17 07:12:41,133][starVLA.training.train_starvla][INFO] - Step 560, Loss: {'train/loss': 0.9618725776672363, 'train/grad_norm_pre_clip': 9.174591064453125, 'timing/data': 0.004806036013178527, 'timing/model': 1.2065747949527577, 'learning_rate/qwen_vl_interface': 9.896201103485577e-06, 'learning_rate/action_model': 9.896201103485577e-05, 'epoch': 0.16}) +[2026-06-17 07:12:45,891][starVLA.training.train_starvla][INFO] - Step 561, Loss: {'train/loss': 1.2466397285461426, 'train/grad_norm_pre_clip': 11.436456680297852, 'timing/data': 0.004445821978151798, 'timing/model': 1.2057244970928878, 'learning_rate/qwen_vl_interface': 9.895750973261576e-06, 'learning_rate/action_model': 9.895750973261574e-05, 'epoch': 0.17}) +[2026-06-17 07:12:50,677][starVLA.training.train_starvla][INFO] - Step 562, Loss: {'train/loss': 1.325380802154541, 'train/grad_norm_pre_clip': 9.875274658203125, 'timing/data': 0.004633761942386627, 'timing/model': 1.2048305859789252, 'learning_rate/qwen_vl_interface': 9.895299879968398e-06, 'learning_rate/action_model': 9.895299879968398e-05, 'epoch': 0.17}) +[2026-06-17 07:12:55,468][starVLA.training.train_starvla][INFO] - Step 563, Loss: {'train/loss': 1.1619421243667603, 'train/grad_norm_pre_clip': 6.772523880004883, 'timing/data': 0.004734253045171499, 'timing/model': 1.2065393179655075, 'learning_rate/qwen_vl_interface': 9.894847823699555e-06, 'learning_rate/action_model': 9.894847823699555e-05, 'epoch': 0.17}) +[2026-06-17 07:13:00,202][starVLA.training.train_starvla][INFO] - Step 564, Loss: {'train/loss': 1.0396758317947388, 'train/grad_norm_pre_clip': 6.917368412017822, 'timing/data': 0.004588843090459704, 'timing/model': 1.2737983409315348, 'learning_rate/qwen_vl_interface': 9.894394804548763e-06, 'learning_rate/action_model': 9.894394804548762e-05, 'epoch': 0.17}) +[2026-06-17 07:13:04,874][starVLA.training.train_starvla][INFO] - Step 565, Loss: {'train/loss': 0.7233288288116455, 'train/grad_norm_pre_clip': 13.47473430633545, 'timing/data': 0.004460214986465871, 'timing/model': 1.2085558840772137, 'learning_rate/qwen_vl_interface': 9.893940822609928e-06, 'learning_rate/action_model': 9.893940822609927e-05, 'epoch': 0.17}) +[2026-06-17 07:13:09,639][starVLA.training.train_starvla][INFO] - Step 566, Loss: {'train/loss': 1.0583045482635498, 'train/grad_norm_pre_clip': 10.43657398223877, 'timing/data': 0.00497532996814698, 'timing/model': 1.2062003399478272, 'learning_rate/qwen_vl_interface': 9.893485877977163e-06, 'learning_rate/action_model': 9.893485877977163e-05, 'epoch': 0.17}) +[2026-06-17 07:13:14,433][starVLA.training.train_starvla][INFO] - Step 567, Loss: {'train/loss': 0.9121124148368835, 'train/grad_norm_pre_clip': 9.09938907623291, 'timing/data': 0.004660837003029883, 'timing/model': 1.2083977890433744, 'learning_rate/qwen_vl_interface': 9.89302997074478e-06, 'learning_rate/action_model': 9.89302997074478e-05, 'epoch': 0.17}) +[2026-06-17 07:13:19,245][starVLA.training.train_starvla][INFO] - Step 568, Loss: {'train/loss': 0.9647120237350464, 'train/grad_norm_pre_clip': 8.830949783325195, 'timing/data': 0.004310964955948293, 'timing/model': 1.2079396459739655, 'learning_rate/qwen_vl_interface': 9.892573101007288e-06, 'learning_rate/action_model': 9.892573101007289e-05, 'epoch': 0.17}) +[2026-06-17 07:13:24,029][starVLA.training.train_starvla][INFO] - Step 569, Loss: {'train/loss': 0.9267542958259583, 'train/grad_norm_pre_clip': 6.642467498779297, 'timing/data': 0.004271093988791108, 'timing/model': 1.3311556059634313, 'learning_rate/qwen_vl_interface': 9.892115268859398e-06, 'learning_rate/action_model': 9.892115268859397e-05, 'epoch': 0.17}) +[2026-06-17 07:13:28,764][starVLA.training.train_starvla][INFO] - Step 570, Loss: {'train/loss': 0.8907955288887024, 'train/grad_norm_pre_clip': 10.848466873168945, 'timing/data': 0.004379479098133743, 'timing/model': 1.2834074579877779, 'learning_rate/qwen_vl_interface': 9.891656474396015e-06, 'learning_rate/action_model': 9.891656474396015e-05, 'epoch': 0.17}) +[2026-06-17 07:13:33,424][starVLA.training.train_starvla][INFO] - Step 571, Loss: {'train/loss': 1.1663819551467896, 'train/grad_norm_pre_clip': 7.693465709686279, 'timing/data': 0.004779801005497575, 'timing/model': 1.2074458179995418, 'learning_rate/qwen_vl_interface': 9.891196717712253e-06, 'learning_rate/action_model': 9.891196717712252e-05, 'epoch': 0.17}) +[2026-06-17 07:13:38,204][starVLA.training.train_starvla][INFO] - Step 572, Loss: {'train/loss': 1.0312188863754272, 'train/grad_norm_pre_clip': 9.587735176086426, 'timing/data': 0.004297171952202916, 'timing/model': 1.2048855039756745, 'learning_rate/qwen_vl_interface': 9.890735998903414e-06, 'learning_rate/action_model': 9.890735998903414e-05, 'epoch': 0.17}) +[2026-06-17 07:13:42,993][starVLA.training.train_starvla][INFO] - Step 573, Loss: {'train/loss': 1.0594311952590942, 'train/grad_norm_pre_clip': 8.293608665466309, 'timing/data': 0.004609772004187107, 'timing/model': 1.2066518070641905, 'learning_rate/qwen_vl_interface': 9.890274318065012e-06, 'learning_rate/action_model': 9.890274318065011e-05, 'epoch': 0.17}) +[2026-06-17 07:13:47,782][starVLA.training.train_starvla][INFO] - Step 574, Loss: {'train/loss': 1.0573785305023193, 'train/grad_norm_pre_clip': 11.208386421203613, 'timing/data': 0.00449186097830534, 'timing/model': 1.208175502019003, 'learning_rate/qwen_vl_interface': 9.889811675292748e-06, 'learning_rate/action_model': 9.889811675292749e-05, 'epoch': 0.17}) +[2026-06-17 07:13:52,574][starVLA.training.train_starvla][INFO] - Step 575, Loss: {'train/loss': 0.6892600655555725, 'train/grad_norm_pre_clip': 8.608230590820312, 'timing/data': 0.0046359190018847585, 'timing/model': 1.3355489489622414, 'learning_rate/qwen_vl_interface': 9.889348070682533e-06, 'learning_rate/action_model': 9.889348070682533e-05, 'epoch': 0.17}) +[2026-06-17 07:13:57,236][starVLA.training.train_starvla][INFO] - Step 576, Loss: {'train/loss': 1.0074608325958252, 'train/grad_norm_pre_clip': 16.68812370300293, 'timing/data': 0.00446924299467355, 'timing/model': 1.2074139699107036, 'learning_rate/qwen_vl_interface': 9.888883504330469e-06, 'learning_rate/action_model': 9.888883504330468e-05, 'epoch': 0.17}) +[2026-06-17 07:14:02,031][starVLA.training.train_starvla][INFO] - Step 577, Loss: {'train/loss': 1.0546220541000366, 'train/grad_norm_pre_clip': 11.73118782043457, 'timing/data': 0.004745661048218608, 'timing/model': 1.2088408849667758, 'learning_rate/qwen_vl_interface': 9.888417976332865e-06, 'learning_rate/action_model': 9.888417976332865e-05, 'epoch': 0.17}) +[2026-06-17 07:14:06,828][starVLA.training.train_starvla][INFO] - Step 578, Loss: {'train/loss': 0.9237208366394043, 'train/grad_norm_pre_clip': 7.3478312492370605, 'timing/data': 0.004474326968193054, 'timing/model': 1.207049548975192, 'learning_rate/qwen_vl_interface': 9.887951486786222e-06, 'learning_rate/action_model': 9.887951486786221e-05, 'epoch': 0.17}) +[2026-06-17 07:14:11,570][starVLA.training.train_starvla][INFO] - Step 579, Loss: {'train/loss': 1.273012399673462, 'train/grad_norm_pre_clip': 9.576668739318848, 'timing/data': 0.004777658963575959, 'timing/model': 1.2064502050634474, 'learning_rate/qwen_vl_interface': 9.887484035787246e-06, 'learning_rate/action_model': 9.887484035787246e-05, 'epoch': 0.17}) +[2026-06-17 07:14:16,354][starVLA.training.train_starvla][INFO] - Step 580, Loss: {'train/loss': 0.8613487482070923, 'train/grad_norm_pre_clip': 7.189638137817383, 'timing/data': 0.004794177017174661, 'timing/model': 1.205526705016382, 'learning_rate/qwen_vl_interface': 9.887015623432837e-06, 'learning_rate/action_model': 9.887015623432837e-05, 'epoch': 0.17}) +[2026-06-17 07:14:21,116][starVLA.training.train_starvla][INFO] - Step 581, Loss: {'train/loss': 0.9132448434829712, 'train/grad_norm_pre_clip': 7.633723735809326, 'timing/data': 0.005007177009247243, 'timing/model': 1.30372709792573, 'learning_rate/qwen_vl_interface': 9.886546249820101e-06, 'learning_rate/action_model': 9.886546249820102e-05, 'epoch': 0.17}) +[2026-06-17 07:14:25,779][starVLA.training.train_starvla][INFO] - Step 582, Loss: {'train/loss': 0.8999451398849487, 'train/grad_norm_pre_clip': 7.4064788818359375, 'timing/data': 0.0043687159195542336, 'timing/model': 1.205996097996831, 'learning_rate/qwen_vl_interface': 9.886075915046339e-06, 'learning_rate/action_model': 9.886075915046338e-05, 'epoch': 0.17}) +[2026-06-17 07:14:30,586][starVLA.training.train_starvla][INFO] - Step 583, Loss: {'train/loss': 1.354406714439392, 'train/grad_norm_pre_clip': 10.385233879089355, 'timing/data': 0.005386137985624373, 'timing/model': 1.2090416109422222, 'learning_rate/qwen_vl_interface': 9.885604619209052e-06, 'learning_rate/action_model': 9.885604619209051e-05, 'epoch': 0.17}) +[2026-06-17 07:14:35,443][starVLA.training.train_starvla][INFO] - Step 584, Loss: {'train/loss': 0.9457313418388367, 'train/grad_norm_pre_clip': 10.80337142944336, 'timing/data': 0.004598764004185796, 'timing/model': 1.2108126500388607, 'learning_rate/qwen_vl_interface': 9.885132362405935e-06, 'learning_rate/action_model': 9.885132362405936e-05, 'epoch': 0.17}) +[2026-06-17 07:14:40,235][starVLA.training.train_starvla][INFO] - Step 585, Loss: {'train/loss': 1.11764395236969, 'train/grad_norm_pre_clip': 10.992107391357422, 'timing/data': 0.004713251022621989, 'timing/model': 1.207144447020255, 'learning_rate/qwen_vl_interface': 9.884659144734896e-06, 'learning_rate/action_model': 9.884659144734896e-05, 'epoch': 0.17}) +[2026-06-17 07:14:45,003][starVLA.training.train_starvla][INFO] - Step 586, Loss: {'train/loss': 0.971139669418335, 'train/grad_norm_pre_clip': 8.657437324523926, 'timing/data': 0.004450779990293086, 'timing/model': 1.2131815609754995, 'learning_rate/qwen_vl_interface': 9.884184966294028e-06, 'learning_rate/action_model': 9.884184966294028e-05, 'epoch': 0.17}) +[2026-06-17 07:14:49,811][starVLA.training.train_starvla][INFO] - Step 587, Loss: {'train/loss': 0.8258695602416992, 'train/grad_norm_pre_clip': 7.197385787963867, 'timing/data': 0.004498352995142341, 'timing/model': 1.3411166130099446, 'learning_rate/qwen_vl_interface': 9.883709827181632e-06, 'learning_rate/action_model': 9.88370982718163e-05, 'epoch': 0.17}) +[2026-06-17 07:14:54,489][starVLA.training.train_starvla][INFO] - Step 588, Loss: {'train/loss': 1.161035180091858, 'train/grad_norm_pre_clip': 8.445181846618652, 'timing/data': 0.004185560974292457, 'timing/model': 1.208455552929081, 'learning_rate/qwen_vl_interface': 9.8832337274962e-06, 'learning_rate/action_model': 9.883233727496198e-05, 'epoch': 0.17}) +[2026-06-17 07:14:59,278][starVLA.training.train_starvla][INFO] - Step 589, Loss: {'train/loss': 1.2850514650344849, 'train/grad_norm_pre_clip': 9.395966529846191, 'timing/data': 0.004241983988322318, 'timing/model': 1.208688111975789, 'learning_rate/qwen_vl_interface': 9.88275666733643e-06, 'learning_rate/action_model': 9.882756667336429e-05, 'epoch': 0.17}) +[2026-06-17 07:15:04,084][starVLA.training.train_starvla][INFO] - Step 590, Loss: {'train/loss': 1.1646779775619507, 'train/grad_norm_pre_clip': 7.921474456787109, 'timing/data': 0.004452842054888606, 'timing/model': 1.2095758629729971, 'learning_rate/qwen_vl_interface': 9.88227864680122e-06, 'learning_rate/action_model': 9.88227864680122e-05, 'epoch': 0.17}) +[2026-06-17 07:15:08,862][starVLA.training.train_starvla][INFO] - Step 591, Loss: {'train/loss': 0.8972280025482178, 'train/grad_norm_pre_clip': 14.185601234436035, 'timing/data': 0.00475683796685189, 'timing/model': 1.2053926180815324, 'learning_rate/qwen_vl_interface': 9.88179966598966e-06, 'learning_rate/action_model': 9.88179966598966e-05, 'epoch': 0.17}) +[2026-06-17 07:15:13,646][starVLA.training.train_starvla][INFO] - Step 592, Loss: {'train/loss': 1.064279317855835, 'train/grad_norm_pre_clip': 11.20753288269043, 'timing/data': 0.0042310310527682304, 'timing/model': 1.2080175250303, 'learning_rate/qwen_vl_interface': 9.881319725001046e-06, 'learning_rate/action_model': 9.881319725001046e-05, 'epoch': 0.17}) +[2026-06-17 07:15:18,455][starVLA.training.train_starvla][INFO] - Step 593, Loss: {'train/loss': 1.0141855478286743, 'train/grad_norm_pre_clip': 10.234054565429688, 'timing/data': 0.004496967070735991, 'timing/model': 1.3378304820507765, 'learning_rate/qwen_vl_interface': 9.88083882393487e-06, 'learning_rate/action_model': 9.880838823934869e-05, 'epoch': 0.17}) +[2026-06-17 07:15:23,137][starVLA.training.train_starvla][INFO] - Step 594, Loss: {'train/loss': 1.1086513996124268, 'train/grad_norm_pre_clip': 9.105399131774902, 'timing/data': 0.0048127330373972654, 'timing/model': 1.2079062530538067, 'learning_rate/qwen_vl_interface': 9.880356962890823e-06, 'learning_rate/action_model': 9.880356962890822e-05, 'epoch': 0.17}) +[2026-06-17 07:15:27,953][starVLA.training.train_starvla][INFO] - Step 595, Loss: {'train/loss': 0.8756678700447083, 'train/grad_norm_pre_clip': 7.250060558319092, 'timing/data': 0.004601227003149688, 'timing/model': 1.2131236280547455, 'learning_rate/qwen_vl_interface': 9.879874141968792e-06, 'learning_rate/action_model': 9.879874141968792e-05, 'epoch': 0.18}) +[2026-06-17 07:15:32,792][starVLA.training.train_starvla][INFO] - Step 596, Loss: {'train/loss': 0.9454721808433533, 'train/grad_norm_pre_clip': 7.05964469909668, 'timing/data': 0.004705972969532013, 'timing/model': 1.2085727440426126, 'learning_rate/qwen_vl_interface': 9.879390361268873e-06, 'learning_rate/action_model': 9.879390361268873e-05, 'epoch': 0.18}) +[2026-06-17 07:15:37,595][starVLA.training.train_starvla][INFO] - Step 597, Loss: {'train/loss': 1.1668074131011963, 'train/grad_norm_pre_clip': 7.777029037475586, 'timing/data': 0.004145840066485107, 'timing/model': 1.2077217060141265, 'learning_rate/qwen_vl_interface': 9.878905620891349e-06, 'learning_rate/action_model': 9.878905620891349e-05, 'epoch': 0.18}) +[2026-06-17 07:15:42,393][starVLA.training.train_starvla][INFO] - Step 598, Loss: {'train/loss': 1.0672097206115723, 'train/grad_norm_pre_clip': 8.794377326965332, 'timing/data': 0.004290882032364607, 'timing/model': 1.3435504760127515, 'learning_rate/qwen_vl_interface': 9.87841992093671e-06, 'learning_rate/action_model': 9.87841992093671e-05, 'epoch': 0.18}) +[2026-06-17 07:15:47,057][starVLA.training.train_starvla][INFO] - Step 599, Loss: {'train/loss': 1.0089071989059448, 'train/grad_norm_pre_clip': 10.951715469360352, 'timing/data': 0.004760734969750047, 'timing/model': 1.2096645899582654, 'learning_rate/qwen_vl_interface': 9.877933261505642e-06, 'learning_rate/action_model': 9.877933261505641e-05, 'epoch': 0.18}) +[2026-06-17 07:17:29,070][starVLA.training.train_starvla][INFO] - Step 600, Loss: {'train/loss': 1.0333871841430664, 'train/grad_norm_pre_clip': 7.9298415184021, 'eval/loss': 1.032431721687317, 'eval/latency_0/loss': 1.032431721687317, 'eval/demon_fire_precision': 0.9084730403262347, 'eval/demon_fire_recall': 0.832641196013289, 'eval/demon_fire_f1': 0.8689057421451788, 'eval/demon_fire_rate_model': 0.6137374860956618, 'eval/demon_fire_rate_teacher': 0.6696329254727474, 'eval/demon_left_precision': 0.6743436204514048, 'eval/demon_left_recall': 0.8294617563739377, 'eval/demon_left_f1': 0.7439024390243902, 'eval/demon_left_rate_model': 0.603726362625139, 'eval/demon_left_rate_teacher': 0.49082313681868744, 'eval/demon_right_precision': 0.7698245614035087, 'eval/demon_right_recall': 0.619774011299435, 'eval/demon_right_f1': 0.6866979655712051, 'eval/demon_right_rate_model': 0.39627363737486093, 'eval/demon_right_rate_teacher': 0.4922135706340378, 'eval/demon_move_f1': 0.7153002022977977, 'eval/demon_cc_f1': 0.8163163352265544, 'eval/latency_0/demon_fire_precision': 0.9084730403262347, 'eval/latency_0/demon_fire_recall': 0.832641196013289, 'eval/latency_0/demon_fire_f1': 0.8689057421451788, 'eval/latency_0/demon_fire_rate_model': 0.6137374860956618, 'eval/latency_0/demon_fire_rate_teacher': 0.6696329254727474, 'eval/latency_0/demon_left_precision': 0.6743436204514048, 'eval/latency_0/demon_left_recall': 0.8294617563739377, 'eval/latency_0/demon_left_f1': 0.7439024390243902, 'eval/latency_0/demon_left_rate_model': 0.603726362625139, 'eval/latency_0/demon_left_rate_teacher': 0.49082313681868744, 'eval/latency_0/demon_right_precision': 0.7698245614035087, 'eval/latency_0/demon_right_recall': 0.619774011299435, 'eval/latency_0/demon_right_f1': 0.6866979655712051, 'eval/latency_0/demon_right_rate_model': 0.39627363737486093, 'eval/latency_0/demon_right_rate_teacher': 0.4922135706340378, 'eval/latency_0/demon_move_f1': 0.7153002022977977, 'eval/latency_0/demon_cc_f1': 0.8163163352265544, 'timing/data': 0.004503816948272288, 'timing/model': 1.213308023987338, 'learning_rate/qwen_vl_interface': 9.877445642699026e-06, 'learning_rate/action_model': 9.877445642699026e-05, 'epoch': 0.18}) +[2026-06-17 07:17:29,070][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_600_state +[2026-06-17 07:17:29,073][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 07:17:32,609][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_600_state/model.safetensors +[2026-06-17 07:17:40,439][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_600_state/optimizer.bin +[2026-06-17 07:17:40,439][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_600_state/sampler.bin +[2026-06-17 07:17:40,440][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_600_state/sampler_1.bin +[2026-06-17 07:17:40,440][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_600_state/random_states_0.pkl +[2026-06-17 07:17:40,478][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 600 +[2026-06-17 07:17:40,478][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 07:17:40,497][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 07:17:45,091][starVLA.training.train_starvla][INFO] - Step 601, Loss: {'train/loss': 0.8400707840919495, 'train/grad_norm_pre_clip': 8.13883113861084, 'timing/data': 0.004554755054414272, 'timing/model': 1.1898044189438224, 'learning_rate/qwen_vl_interface': 9.876957064617951e-06, 'learning_rate/action_model': 9.876957064617951e-05, 'epoch': 0.18}) +[2026-06-17 07:17:49,863][starVLA.training.train_starvla][INFO] - Step 602, Loss: {'train/loss': 1.1517552137374878, 'train/grad_norm_pre_clip': 6.398939609527588, 'timing/data': 0.0045672080013900995, 'timing/model': 1.2011159079847857, 'learning_rate/qwen_vl_interface': 9.876467527363696e-06, 'learning_rate/action_model': 9.876467527363697e-05, 'epoch': 0.18}) +[2026-06-17 07:17:54,602][starVLA.training.train_starvla][INFO] - Step 603, Loss: {'train/loss': 1.072817087173462, 'train/grad_norm_pre_clip': 6.550455093383789, 'timing/data': 0.004537516972050071, 'timing/model': 1.1918870299123228, 'learning_rate/qwen_vl_interface': 9.875977031037746e-06, 'learning_rate/action_model': 9.875977031037745e-05, 'epoch': 0.18}) +[2026-06-17 07:17:59,377][starVLA.training.train_starvla][INFO] - Step 604, Loss: {'train/loss': 1.0507795810699463, 'train/grad_norm_pre_clip': 5.941954612731934, 'timing/data': 0.006224035983905196, 'timing/model': 1.2000300170620903, 'learning_rate/qwen_vl_interface': 9.875485575741779e-06, 'learning_rate/action_model': 9.875485575741779e-05, 'epoch': 0.18}) +[2026-06-17 07:18:04,141][starVLA.training.train_starvla][INFO] - Step 605, Loss: {'train/loss': 1.0672080516815186, 'train/grad_norm_pre_clip': 9.919397354125977, 'timing/data': 0.004548675962723792, 'timing/model': 1.196365326992236, 'learning_rate/qwen_vl_interface': 9.874993161577674e-06, 'learning_rate/action_model': 9.874993161577675e-05, 'epoch': 0.18}) +[2026-06-17 07:18:08,978][starVLA.training.train_starvla][INFO] - Step 606, Loss: {'train/loss': 1.3853962421417236, 'train/grad_norm_pre_clip': 6.73069429397583, 'timing/data': 0.004714380018413067, 'timing/model': 1.3241929908981547, 'learning_rate/qwen_vl_interface': 9.874499788647512e-06, 'learning_rate/action_model': 9.874499788647513e-05, 'epoch': 0.18}) +[2026-06-17 07:18:13,613][starVLA.training.train_starvla][INFO] - Step 607, Loss: {'train/loss': 0.9145570397377014, 'train/grad_norm_pre_clip': 52.104679107666016, 'timing/data': 0.005134753999300301, 'timing/model': 1.1991440310375765, 'learning_rate/qwen_vl_interface': 9.874005457053565e-06, 'learning_rate/action_model': 9.874005457053565e-05, 'epoch': 0.18}) +[2026-06-17 07:18:18,384][starVLA.training.train_starvla][INFO] - Step 608, Loss: {'train/loss': 1.113641619682312, 'train/grad_norm_pre_clip': 5.964258670806885, 'timing/data': 0.0047732070088386536, 'timing/model': 1.2004664990818128, 'learning_rate/qwen_vl_interface': 9.873510166898315e-06, 'learning_rate/action_model': 9.873510166898314e-05, 'epoch': 0.18}) +[2026-06-17 07:18:23,116][starVLA.training.train_starvla][INFO] - Step 609, Loss: {'train/loss': 1.0938907861709595, 'train/grad_norm_pre_clip': 7.260024070739746, 'timing/data': 0.004842316964641213, 'timing/model': 1.1986625510035083, 'learning_rate/qwen_vl_interface': 9.87301391828443e-06, 'learning_rate/action_model': 9.873013918284428e-05, 'epoch': 0.18}) +[2026-06-17 07:18:27,935][starVLA.training.train_starvla][INFO] - Step 610, Loss: {'train/loss': 1.1854174137115479, 'train/grad_norm_pre_clip': 22.6144962310791, 'timing/data': 0.004698742995969951, 'timing/model': 1.2101128799840808, 'learning_rate/qwen_vl_interface': 9.872516711314784e-06, 'learning_rate/action_model': 9.872516711314784e-05, 'epoch': 0.18}) +[2026-06-17 07:18:32,741][starVLA.training.train_starvla][INFO] - Step 611, Loss: {'train/loss': 0.9991589188575745, 'train/grad_norm_pre_clip': 8.275267601013184, 'timing/data': 0.00480677792802453, 'timing/model': 1.2095286808907986, 'learning_rate/qwen_vl_interface': 9.872018546092452e-06, 'learning_rate/action_model': 9.872018546092451e-05, 'epoch': 0.18}) +[2026-06-17 07:18:37,699][starVLA.training.train_starvla][INFO] - Step 612, Loss: {'train/loss': 1.0491660833358765, 'train/grad_norm_pre_clip': 7.5695366859436035, 'timing/data': 0.0045825360575690866, 'timing/model': 1.4973890860565007, 'learning_rate/qwen_vl_interface': 9.871519422720699e-06, 'learning_rate/action_model': 9.871519422720698e-05, 'epoch': 0.18}) +[2026-06-17 07:18:42,499][starVLA.training.train_starvla][INFO] - Step 613, Loss: {'train/loss': 0.925554096698761, 'train/grad_norm_pre_clip': 6.6689653396606445, 'timing/data': 0.00496732397004962, 'timing/model': 1.332870040088892, 'learning_rate/qwen_vl_interface': 9.871019341303e-06, 'learning_rate/action_model': 9.871019341303e-05, 'epoch': 0.18}) +[2026-06-17 07:18:47,169][starVLA.training.train_starvla][INFO] - Step 614, Loss: {'train/loss': 0.9288319945335388, 'train/grad_norm_pre_clip': 7.6909074783325195, 'timing/data': 0.004994728951714933, 'timing/model': 1.2074671380687505, 'learning_rate/qwen_vl_interface': 9.870518301943018e-06, 'learning_rate/action_model': 9.870518301943018e-05, 'epoch': 0.18}) +[2026-06-17 07:18:51,929][starVLA.training.train_starvla][INFO] - Step 615, Loss: {'train/loss': 1.1616356372833252, 'train/grad_norm_pre_clip': 9.434414863586426, 'timing/data': 0.004794872947968543, 'timing/model': 1.210627743974328, 'learning_rate/qwen_vl_interface': 9.870016304744617e-06, 'learning_rate/action_model': 9.870016304744616e-05, 'epoch': 0.18}) +[2026-06-17 07:18:56,740][starVLA.training.train_starvla][INFO] - Step 616, Loss: {'train/loss': 0.9585555791854858, 'train/grad_norm_pre_clip': 6.005756378173828, 'timing/data': 0.004921993939206004, 'timing/model': 1.2113005860010162, 'learning_rate/qwen_vl_interface': 9.869513349811868e-06, 'learning_rate/action_model': 9.869513349811868e-05, 'epoch': 0.18}) +[2026-06-17 07:19:01,580][starVLA.training.train_starvla][INFO] - Step 617, Loss: {'train/loss': 0.9891989231109619, 'train/grad_norm_pre_clip': 8.922435760498047, 'timing/data': 0.0063252439722418785, 'timing/model': 1.2200483810156584, 'learning_rate/qwen_vl_interface': 9.86900943724903e-06, 'learning_rate/action_model': 9.869009437249031e-05, 'epoch': 0.18}) +[2026-06-17 07:19:06,370][starVLA.training.train_starvla][INFO] - Step 618, Loss: {'train/loss': 1.035862922668457, 'train/grad_norm_pre_clip': 6.2481560707092285, 'timing/data': 0.004336719983257353, 'timing/model': 1.3309000240406021, 'learning_rate/qwen_vl_interface': 9.868504567160566e-06, 'learning_rate/action_model': 9.868504567160565e-05, 'epoch': 0.18}) +[2026-06-17 07:19:11,182][starVLA.training.train_starvla][INFO] - Step 619, Loss: {'train/loss': 1.1905659437179565, 'train/grad_norm_pre_clip': 8.977563858032227, 'timing/data': 0.0062021529302001, 'timing/model': 1.324795680004172, 'learning_rate/qwen_vl_interface': 9.867998739651134e-06, 'learning_rate/action_model': 9.867998739651134e-05, 'epoch': 0.18}) +[2026-06-17 07:19:15,886][starVLA.training.train_starvla][INFO] - Step 620, Loss: {'train/loss': 0.9938451051712036, 'train/grad_norm_pre_clip': 7.8383612632751465, 'timing/data': 0.00426590908318758, 'timing/model': 1.224940118030645, 'learning_rate/qwen_vl_interface': 9.867491954825595e-06, 'learning_rate/action_model': 9.867491954825595e-05, 'epoch': 0.18}) +[2026-06-17 07:19:20,682][starVLA.training.train_starvla][INFO] - Step 621, Loss: {'train/loss': 1.2987501621246338, 'train/grad_norm_pre_clip': 9.01424503326416, 'timing/data': 0.004526142030954361, 'timing/model': 1.2143962520640343, 'learning_rate/qwen_vl_interface': 9.866984212789007e-06, 'learning_rate/action_model': 9.866984212789006e-05, 'epoch': 0.18}) +[2026-06-17 07:19:25,499][starVLA.training.train_starvla][INFO] - Step 622, Loss: {'train/loss': 1.070490837097168, 'train/grad_norm_pre_clip': 8.921146392822266, 'timing/data': 0.005000227014534175, 'timing/model': 1.2100295050768182, 'learning_rate/qwen_vl_interface': 9.86647551364662e-06, 'learning_rate/action_model': 9.86647551364662e-05, 'epoch': 0.18}) +[2026-06-17 07:19:30,309][starVLA.training.train_starvla][INFO] - Step 623, Loss: {'train/loss': 1.0070923566818237, 'train/grad_norm_pre_clip': 8.644222259521484, 'timing/data': 0.005055474932305515, 'timing/model': 1.2183882230892777, 'learning_rate/qwen_vl_interface': 9.865965857503893e-06, 'learning_rate/action_model': 9.865965857503894e-05, 'epoch': 0.18}) +[2026-06-17 07:19:35,058][starVLA.training.train_starvla][INFO] - Step 624, Loss: {'train/loss': 1.0310760736465454, 'train/grad_norm_pre_clip': 13.10063362121582, 'timing/data': 0.004863306065090001, 'timing/model': 1.2762250680243596, 'learning_rate/qwen_vl_interface': 9.865455244466478e-06, 'learning_rate/action_model': 9.865455244466477e-05, 'epoch': 0.18}) +[2026-06-17 07:19:39,724][starVLA.training.train_starvla][INFO] - Step 625, Loss: {'train/loss': 0.9401010870933533, 'train/grad_norm_pre_clip': 8.879707336425781, 'timing/data': 0.004554545972496271, 'timing/model': 1.2123085119528696, 'learning_rate/qwen_vl_interface': 9.864943674640223e-06, 'learning_rate/action_model': 9.864943674640222e-05, 'epoch': 0.18}) +[2026-06-17 07:19:44,507][starVLA.training.train_starvla][INFO] - Step 626, Loss: {'train/loss': 0.9394683837890625, 'train/grad_norm_pre_clip': 6.10302734375, 'timing/data': 0.004878536914475262, 'timing/model': 1.2067177420249209, 'learning_rate/qwen_vl_interface': 9.864431148131178e-06, 'learning_rate/action_model': 9.864431148131178e-05, 'epoch': 0.18}) +[2026-06-17 07:19:49,329][starVLA.training.train_starvla][INFO] - Step 627, Loss: {'train/loss': 0.9700384140014648, 'train/grad_norm_pre_clip': 7.504593849182129, 'timing/data': 0.004707904066890478, 'timing/model': 1.2063984140986577, 'learning_rate/qwen_vl_interface': 9.863917665045591e-06, 'learning_rate/action_model': 9.863917665045592e-05, 'epoch': 0.18}) +[2026-06-17 07:19:54,128][starVLA.training.train_starvla][INFO] - Step 628, Loss: {'train/loss': 0.849330723285675, 'train/grad_norm_pre_clip': 5.426644802093506, 'timing/data': 0.005046093952842057, 'timing/model': 1.2071989109972492, 'learning_rate/qwen_vl_interface': 9.863403225489909e-06, 'learning_rate/action_model': 9.863403225489909e-05, 'epoch': 0.18}) +[2026-06-17 07:19:58,955][starVLA.training.train_starvla][INFO] - Step 629, Loss: {'train/loss': 0.6992565989494324, 'train/grad_norm_pre_clip': 6.266414642333984, 'timing/data': 0.004983684979379177, 'timing/model': 1.348173859063536, 'learning_rate/qwen_vl_interface': 9.862887829570771e-06, 'learning_rate/action_model': 9.862887829570772e-05, 'epoch': 0.19}) +[2026-06-17 07:20:03,713][starVLA.training.train_starvla][INFO] - Step 630, Loss: {'train/loss': 0.919585108757019, 'train/grad_norm_pre_clip': 7.223989963531494, 'timing/data': 0.00483025296125561, 'timing/model': 1.2889084839262068, 'learning_rate/qwen_vl_interface': 9.862371477395024e-06, 'learning_rate/action_model': 9.862371477395025e-05, 'epoch': 0.19}) +[2026-06-17 07:20:08,390][starVLA.training.train_starvla][INFO] - Step 631, Loss: {'train/loss': 1.2935032844543457, 'train/grad_norm_pre_clip': 8.867372512817383, 'timing/data': 0.0045616429997608066, 'timing/model': 1.2108212920138612, 'learning_rate/qwen_vl_interface': 9.861854169069708e-06, 'learning_rate/action_model': 9.861854169069707e-05, 'epoch': 0.19}) +[2026-06-17 07:20:13,178][starVLA.training.train_starvla][INFO] - Step 632, Loss: {'train/loss': 1.054316759109497, 'train/grad_norm_pre_clip': 14.51819896697998, 'timing/data': 0.004641411011107266, 'timing/model': 1.2050098299514502, 'learning_rate/qwen_vl_interface': 9.861335904702058e-06, 'learning_rate/action_model': 9.861335904702058e-05, 'epoch': 0.19}) +[2026-06-17 07:20:17,972][starVLA.training.train_starvla][INFO] - Step 633, Loss: {'train/loss': 1.2078404426574707, 'train/grad_norm_pre_clip': 9.213605880737305, 'timing/data': 0.0048836920177564025, 'timing/model': 1.2108664000406861, 'learning_rate/qwen_vl_interface': 9.860816684399514e-06, 'learning_rate/action_model': 9.860816684399514e-05, 'epoch': 0.19}) +[2026-06-17 07:20:22,762][starVLA.training.train_starvla][INFO] - Step 634, Loss: {'train/loss': 0.9999677538871765, 'train/grad_norm_pre_clip': 8.01135540008545, 'timing/data': 0.0048742389772087336, 'timing/model': 1.2082904019625857, 'learning_rate/qwen_vl_interface': 9.86029650826971e-06, 'learning_rate/action_model': 9.86029650826971e-05, 'epoch': 0.19}) +[2026-06-17 07:20:27,562][starVLA.training.train_starvla][INFO] - Step 635, Loss: {'train/loss': 1.1338996887207031, 'train/grad_norm_pre_clip': 8.243108749389648, 'timing/data': 0.004752070060931146, 'timing/model': 1.336664048023522, 'learning_rate/qwen_vl_interface': 9.859775376420478e-06, 'learning_rate/action_model': 9.859775376420478e-05, 'epoch': 0.19}) +[2026-06-17 07:20:32,228][starVLA.training.train_starvla][INFO] - Step 636, Loss: {'train/loss': 1.0441035032272339, 'train/grad_norm_pre_clip': 8.431600570678711, 'timing/data': 0.004993380047380924, 'timing/model': 1.2073119359556586, 'learning_rate/qwen_vl_interface': 9.85925328895985e-06, 'learning_rate/action_model': 9.85925328895985e-05, 'epoch': 0.19}) +[2026-06-17 07:20:37,080][starVLA.training.train_starvla][INFO] - Step 637, Loss: {'train/loss': 1.0728886127471924, 'train/grad_norm_pre_clip': 11.092301368713379, 'timing/data': 0.00501175201497972, 'timing/model': 1.2655442390823737, 'learning_rate/qwen_vl_interface': 9.858730245996057e-06, 'learning_rate/action_model': 9.858730245996057e-05, 'epoch': 0.19}) +[2026-06-17 07:20:41,877][starVLA.training.train_starvla][INFO] - Step 638, Loss: {'train/loss': 1.1171241998672485, 'train/grad_norm_pre_clip': 7.53553581237793, 'timing/data': 0.005139486049301922, 'timing/model': 1.20536116999574, 'learning_rate/qwen_vl_interface': 9.858206247637522e-06, 'learning_rate/action_model': 9.858206247637523e-05, 'epoch': 0.19}) +[2026-06-17 07:20:46,572][starVLA.training.train_starvla][INFO] - Step 639, Loss: {'train/loss': 1.2215074300765991, 'train/grad_norm_pre_clip': 14.10367202758789, 'timing/data': 0.004390627960674465, 'timing/model': 1.2066189360339195, 'learning_rate/qwen_vl_interface': 9.857681293992876e-06, 'learning_rate/action_model': 9.857681293992877e-05, 'epoch': 0.19}) +[2026-06-17 07:20:51,400][starVLA.training.train_starvla][INFO] - Step 640, Loss: {'train/loss': 0.9932142496109009, 'train/grad_norm_pre_clip': 11.228353500366211, 'timing/data': 0.004447906045243144, 'timing/model': 1.2114422160666436, 'learning_rate/qwen_vl_interface': 9.85715538517094e-06, 'learning_rate/action_model': 9.857155385170938e-05, 'epoch': 0.19}) +[2026-06-17 07:20:56,182][starVLA.training.train_starvla][INFO] - Step 641, Loss: {'train/loss': 1.0620911121368408, 'train/grad_norm_pre_clip': 14.13864517211914, 'timing/data': 0.0050265740137547255, 'timing/model': 1.311148272943683, 'learning_rate/qwen_vl_interface': 9.856628521280733e-06, 'learning_rate/action_model': 9.856628521280733e-05, 'epoch': 0.19}) +[2026-06-17 07:21:00,866][starVLA.training.train_starvla][INFO] - Step 642, Loss: {'train/loss': 0.8491032123565674, 'train/grad_norm_pre_clip': 8.836451530456543, 'timing/data': 0.005134318955242634, 'timing/model': 1.2150139290606603, 'learning_rate/qwen_vl_interface': 9.856100702431477e-06, 'learning_rate/action_model': 9.856100702431477e-05, 'epoch': 0.19}) +[2026-06-17 07:21:05,665][starVLA.training.train_starvla][INFO] - Step 643, Loss: {'train/loss': 1.1737874746322632, 'train/grad_norm_pre_clip': 8.639196395874023, 'timing/data': 0.005021522054448724, 'timing/model': 1.2069647359894589, 'learning_rate/qwen_vl_interface': 9.85557192873259e-06, 'learning_rate/action_model': 9.85557192873259e-05, 'epoch': 0.19}) +[2026-06-17 07:21:10,455][starVLA.training.train_starvla][INFO] - Step 644, Loss: {'train/loss': 0.8803818225860596, 'train/grad_norm_pre_clip': 7.994479656219482, 'timing/data': 0.004818634013645351, 'timing/model': 1.20634613011498, 'learning_rate/qwen_vl_interface': 9.855042200293684e-06, 'learning_rate/action_model': 9.855042200293685e-05, 'epoch': 0.19}) +[2026-06-17 07:21:15,244][starVLA.training.train_starvla][INFO] - Step 645, Loss: {'train/loss': 0.9073423743247986, 'train/grad_norm_pre_clip': 25.304853439331055, 'timing/data': 0.004440788994543254, 'timing/model': 1.2060011400608346, 'learning_rate/qwen_vl_interface': 9.854511517224578e-06, 'learning_rate/action_model': 9.854511517224578e-05, 'epoch': 0.19}) +[2026-06-17 07:21:20,003][starVLA.training.train_starvla][INFO] - Step 646, Loss: {'train/loss': 1.0898228883743286, 'train/grad_norm_pre_clip': 7.483989238739014, 'timing/data': 0.004613291006535292, 'timing/model': 1.206116148037836, 'learning_rate/qwen_vl_interface': 9.853979879635278e-06, 'learning_rate/action_model': 9.853979879635278e-05, 'epoch': 0.19}) +[2026-06-17 07:21:24,799][starVLA.training.train_starvla][INFO] - Step 647, Loss: {'train/loss': 1.0893388986587524, 'train/grad_norm_pre_clip': 7.641624450683594, 'timing/data': 0.004642796935513616, 'timing/model': 1.3380441009067, 'learning_rate/qwen_vl_interface': 9.853447287635995e-06, 'learning_rate/action_model': 9.853447287635995e-05, 'epoch': 0.19}) +[2026-06-17 07:21:29,467][starVLA.training.train_starvla][INFO] - Step 648, Loss: {'train/loss': 0.9963766932487488, 'train/grad_norm_pre_clip': 7.734907627105713, 'timing/data': 0.004107662010937929, 'timing/model': 1.2064090110361576, 'learning_rate/qwen_vl_interface': 9.852913741337136e-06, 'learning_rate/action_model': 9.852913741337136e-05, 'epoch': 0.19}) +[2026-06-17 07:21:34,262][starVLA.training.train_starvla][INFO] - Step 649, Loss: {'train/loss': 0.8199461102485657, 'train/grad_norm_pre_clip': 8.315958023071289, 'timing/data': 0.004471341031603515, 'timing/model': 1.2077702670358121, 'learning_rate/qwen_vl_interface': 9.852379240849306e-06, 'learning_rate/action_model': 9.852379240849305e-05, 'epoch': 0.19}) +[2026-06-17 07:21:39,055][starVLA.training.train_starvla][INFO] - Step 650, Loss: {'train/loss': 0.9298624396324158, 'train/grad_norm_pre_clip': 9.696181297302246, 'timing/data': 0.004590085009112954, 'timing/model': 1.2094371010316536, 'learning_rate/qwen_vl_interface': 9.851843786283306e-06, 'learning_rate/action_model': 9.851843786283306e-05, 'epoch': 0.19}) +[2026-06-17 07:21:43,854][starVLA.training.train_starvla][INFO] - Step 651, Loss: {'train/loss': 1.075096607208252, 'train/grad_norm_pre_clip': 8.60389518737793, 'timing/data': 0.004396351054310799, 'timing/model': 1.2098270510323346, 'learning_rate/qwen_vl_interface': 9.851307377750137e-06, 'learning_rate/action_model': 9.851307377750137e-05, 'epoch': 0.19}) +[2026-06-17 07:21:48,613][starVLA.training.train_starvla][INFO] - Step 652, Loss: {'train/loss': 1.157779574394226, 'train/grad_norm_pre_clip': 56.655853271484375, 'timing/data': 0.004848220036365092, 'timing/model': 1.2061161580495536, 'learning_rate/qwen_vl_interface': 9.850770015360999e-06, 'learning_rate/action_model': 9.850770015360999e-05, 'epoch': 0.19}) +[2026-06-17 07:21:53,396][starVLA.training.train_starvla][INFO] - Step 653, Loss: {'train/loss': 1.3214478492736816, 'train/grad_norm_pre_clip': 7.5541839599609375, 'timing/data': 0.004696017014794052, 'timing/model': 1.3229575620498508, 'learning_rate/qwen_vl_interface': 9.850231699227284e-06, 'learning_rate/action_model': 9.850231699227284e-05, 'epoch': 0.19}) +[2026-06-17 07:21:58,069][starVLA.training.train_starvla][INFO] - Step 654, Loss: {'train/loss': 0.881618082523346, 'train/grad_norm_pre_clip': 8.685364723205566, 'timing/data': 0.004381809965707362, 'timing/model': 1.2046923349844292, 'learning_rate/qwen_vl_interface': 9.84969242946059e-06, 'learning_rate/action_model': 9.84969242946059e-05, 'epoch': 0.19}) +[2026-06-17 07:22:02,873][starVLA.training.train_starvla][INFO] - Step 655, Loss: {'train/loss': 0.9916753172874451, 'train/grad_norm_pre_clip': 6.971075534820557, 'timing/data': 0.004483798053115606, 'timing/model': 1.2092126059578732, 'learning_rate/qwen_vl_interface': 9.849152206172704e-06, 'learning_rate/action_model': 9.849152206172704e-05, 'epoch': 0.19}) +[2026-06-17 07:22:07,665][starVLA.training.train_starvla][INFO] - Step 656, Loss: {'train/loss': 0.9911653399467468, 'train/grad_norm_pre_clip': 10.95157527923584, 'timing/data': 0.004497844958677888, 'timing/model': 1.2080861729336902, 'learning_rate/qwen_vl_interface': 9.84861102947562e-06, 'learning_rate/action_model': 9.848611029475619e-05, 'epoch': 0.19}) +[2026-06-17 07:22:12,465][starVLA.training.train_starvla][INFO] - Step 657, Loss: {'train/loss': 1.1320551633834839, 'train/grad_norm_pre_clip': 8.665780067443848, 'timing/data': 0.004698751028627157, 'timing/model': 1.2077450409997255, 'learning_rate/qwen_vl_interface': 9.848068899481516e-06, 'learning_rate/action_model': 9.848068899481516e-05, 'epoch': 0.19}) +[2026-06-17 07:22:17,440][starVLA.training.train_starvla][INFO] - Step 658, Loss: {'train/loss': 0.9680280089378357, 'train/grad_norm_pre_clip': 5.9681715965271, 'timing/data': 0.004853316931985319, 'timing/model': 1.3431251300498843, 'learning_rate/qwen_vl_interface': 9.847525816302784e-06, 'learning_rate/action_model': 9.847525816302784e-05, 'epoch': 0.19}) +[2026-06-17 07:22:22,163][starVLA.training.train_starvla][INFO] - Step 659, Loss: {'train/loss': 0.8854621052742004, 'train/grad_norm_pre_clip': 8.335378646850586, 'timing/data': 0.004520419053733349, 'timing/model': 1.2073561690049246, 'learning_rate/qwen_vl_interface': 9.846981780052005e-06, 'learning_rate/action_model': 9.846981780052003e-05, 'epoch': 0.19}) +[2026-06-17 07:22:26,956][starVLA.training.train_starvla][INFO] - Step 660, Loss: {'train/loss': 1.018665075302124, 'train/grad_norm_pre_clip': 9.426287651062012, 'timing/data': 0.004538742010481656, 'timing/model': 1.2089640410849825, 'learning_rate/qwen_vl_interface': 9.846436790841955e-06, 'learning_rate/action_model': 9.846436790841955e-05, 'epoch': 0.19}) +[2026-06-17 07:22:31,675][starVLA.training.train_starvla][INFO] - Step 661, Loss: {'train/loss': 0.9174926280975342, 'train/grad_norm_pre_clip': 7.811099529266357, 'timing/data': 0.004527886980213225, 'timing/model': 1.216945421998389, 'learning_rate/qwen_vl_interface': 9.84589084878561e-06, 'learning_rate/action_model': 9.845890848785611e-05, 'epoch': 0.19}) +[2026-06-17 07:22:36,424][starVLA.training.train_starvla][INFO] - Step 662, Loss: {'train/loss': 1.0097733736038208, 'train/grad_norm_pre_clip': 23.88768768310547, 'timing/data': 0.004353818017989397, 'timing/model': 1.2068938090233132, 'learning_rate/qwen_vl_interface': 9.84534395399615e-06, 'learning_rate/action_model': 9.84534395399615e-05, 'epoch': 0.19}) +[2026-06-17 07:22:41,255][starVLA.training.train_starvla][INFO] - Step 663, Loss: {'train/loss': 1.2039602994918823, 'train/grad_norm_pre_clip': 12.444695472717285, 'timing/data': 0.004578827996738255, 'timing/model': 1.2088845380349085, 'learning_rate/qwen_vl_interface': 9.844796106586942e-06, 'learning_rate/action_model': 9.844796106586941e-05, 'epoch': 0.2}) +[2026-06-17 07:22:46,052][starVLA.training.train_starvla][INFO] - Step 664, Loss: {'train/loss': 1.0010470151901245, 'train/grad_norm_pre_clip': 8.332337379455566, 'timing/data': 0.004613413009792566, 'timing/model': 1.3390265440102667, 'learning_rate/qwen_vl_interface': 9.844247306671557e-06, 'learning_rate/action_model': 9.844247306671555e-05, 'epoch': 0.2}) +[2026-06-17 07:22:50,720][starVLA.training.train_starvla][INFO] - Step 665, Loss: {'train/loss': 1.1149628162384033, 'train/grad_norm_pre_clip': 8.270331382751465, 'timing/data': 0.00436656700912863, 'timing/model': 1.2085733050480485, 'learning_rate/qwen_vl_interface': 9.843697554363762e-06, 'learning_rate/action_model': 9.843697554363762e-05, 'epoch': 0.2}) +[2026-06-17 07:22:55,546][starVLA.training.train_starvla][INFO] - Step 666, Loss: {'train/loss': 0.9619805812835693, 'train/grad_norm_pre_clip': 8.108734130859375, 'timing/data': 0.0059534909669309855, 'timing/model': 1.2317038570763543, 'learning_rate/qwen_vl_interface': 9.84314684977752e-06, 'learning_rate/action_model': 9.84314684977752e-05, 'epoch': 0.2}) +[2026-06-17 07:23:00,329][starVLA.training.train_starvla][INFO] - Step 667, Loss: {'train/loss': 0.946578860282898, 'train/grad_norm_pre_clip': 6.801882743835449, 'timing/data': 0.004555194056592882, 'timing/model': 1.2077412470243871, 'learning_rate/qwen_vl_interface': 9.842595193026993e-06, 'learning_rate/action_model': 9.842595193026993e-05, 'epoch': 0.2}) +[2026-06-17 07:23:05,098][starVLA.training.train_starvla][INFO] - Step 668, Loss: {'train/loss': 1.2618701457977295, 'train/grad_norm_pre_clip': 7.49548864364624, 'timing/data': 0.004730539978481829, 'timing/model': 1.21510536596179, 'learning_rate/qwen_vl_interface': 9.842042584226545e-06, 'learning_rate/action_model': 9.842042584226544e-05, 'epoch': 0.2}) +[2026-06-17 07:23:09,927][starVLA.training.train_starvla][INFO] - Step 669, Loss: {'train/loss': 1.0397446155548096, 'train/grad_norm_pre_clip': 7.716268539428711, 'timing/data': 0.004911290016025305, 'timing/model': 1.2115713570965454, 'learning_rate/qwen_vl_interface': 9.841489023490723e-06, 'learning_rate/action_model': 9.841489023490723e-05, 'epoch': 0.2}) +[2026-06-17 07:23:14,747][starVLA.training.train_starvla][INFO] - Step 670, Loss: {'train/loss': 1.2153512239456177, 'train/grad_norm_pre_clip': 11.644824028015137, 'timing/data': 0.004875794984400272, 'timing/model': 1.3531732000410557, 'learning_rate/qwen_vl_interface': 9.840934510934288e-06, 'learning_rate/action_model': 9.840934510934288e-05, 'epoch': 0.2}) +[2026-06-17 07:23:19,438][starVLA.training.train_starvla][INFO] - Step 671, Loss: {'train/loss': 0.9324035048484802, 'train/grad_norm_pre_clip': 11.65656566619873, 'timing/data': 0.004715231945738196, 'timing/model': 1.2132421459537, 'learning_rate/qwen_vl_interface': 9.840379046672189e-06, 'learning_rate/action_model': 9.840379046672189e-05, 'epoch': 0.2}) +[2026-06-17 07:23:24,320][starVLA.training.train_starvla][INFO] - Step 672, Loss: {'train/loss': 1.0167688131332397, 'train/grad_norm_pre_clip': 16.901342391967773, 'timing/data': 0.004774947999976575, 'timing/model': 1.2090229550376534, 'learning_rate/qwen_vl_interface': 9.839822630819573e-06, 'learning_rate/action_model': 9.839822630819572e-05, 'epoch': 0.2}) +[2026-06-17 07:23:29,145][starVLA.training.train_starvla][INFO] - Step 673, Loss: {'train/loss': 1.105919361114502, 'train/grad_norm_pre_clip': 7.840555667877197, 'timing/data': 0.006552002974785864, 'timing/model': 1.211900015012361, 'learning_rate/qwen_vl_interface': 9.839265263491789e-06, 'learning_rate/action_model': 9.839265263491789e-05, 'epoch': 0.2}) +[2026-06-17 07:23:34,000][starVLA.training.train_starvla][INFO] - Step 674, Loss: {'train/loss': 0.9660019874572754, 'train/grad_norm_pre_clip': 7.86010217666626, 'timing/data': 0.004311020020395517, 'timing/model': 1.2174227540381253, 'learning_rate/qwen_vl_interface': 9.838706944804376e-06, 'learning_rate/action_model': 9.838706944804374e-05, 'epoch': 0.2}) +[2026-06-17 07:23:38,707][starVLA.training.train_starvla][INFO] - Step 675, Loss: {'train/loss': 1.0202914476394653, 'train/grad_norm_pre_clip': 8.0020170211792, 'timing/data': 0.004095389042049646, 'timing/model': 1.2110533830709755, 'learning_rate/qwen_vl_interface': 9.838147674873076e-06, 'learning_rate/action_model': 9.838147674873075e-05, 'epoch': 0.2}) +[2026-06-17 07:23:43,439][starVLA.training.train_starvla][INFO] - Step 676, Loss: {'train/loss': 1.1754409074783325, 'train/grad_norm_pre_clip': 7.624622821807861, 'timing/data': 0.00462591799441725, 'timing/model': 1.2612102390266955, 'learning_rate/qwen_vl_interface': 9.837587453813826e-06, 'learning_rate/action_model': 9.837587453813824e-05, 'epoch': 0.2}) +[2026-06-17 07:23:48,123][starVLA.training.train_starvla][INFO] - Step 677, Loss: {'train/loss': 0.9580828547477722, 'train/grad_norm_pre_clip': 6.118900775909424, 'timing/data': 0.004701816942542791, 'timing/model': 1.2117775590159, 'learning_rate/qwen_vl_interface': 9.83702628174276e-06, 'learning_rate/action_model': 9.837026281742759e-05, 'epoch': 0.2}) +[2026-06-17 07:23:52,880][starVLA.training.train_starvla][INFO] - Step 678, Loss: {'train/loss': 1.0938267707824707, 'train/grad_norm_pre_clip': 10.063455581665039, 'timing/data': 0.004894291050732136, 'timing/model': 1.2111160421045497, 'learning_rate/qwen_vl_interface': 9.83646415877621e-06, 'learning_rate/action_model': 9.83646415877621e-05, 'epoch': 0.2}) +[2026-06-17 07:23:57,672][starVLA.training.train_starvla][INFO] - Step 679, Loss: {'train/loss': 1.0835847854614258, 'train/grad_norm_pre_clip': 9.180896759033203, 'timing/data': 0.004435208044014871, 'timing/model': 1.206680846051313, 'learning_rate/qwen_vl_interface': 9.835901085030703e-06, 'learning_rate/action_model': 9.835901085030703e-05, 'epoch': 0.2}) +[2026-06-17 07:24:02,488][starVLA.training.train_starvla][INFO] - Step 680, Loss: {'train/loss': 0.9341109395027161, 'train/grad_norm_pre_clip': 6.84384822845459, 'timing/data': 0.004491528030484915, 'timing/model': 1.2054832190042362, 'learning_rate/qwen_vl_interface': 9.83533706062297e-06, 'learning_rate/action_model': 9.835337060622969e-05, 'epoch': 0.2}) +[2026-06-17 07:24:07,284][starVLA.training.train_starvla][INFO] - Step 681, Loss: {'train/loss': 1.0884313583374023, 'train/grad_norm_pre_clip': 6.718865871429443, 'timing/data': 0.004458392970263958, 'timing/model': 1.3367449509678409, 'learning_rate/qwen_vl_interface': 9.834772085669927e-06, 'learning_rate/action_model': 9.834772085669927e-05, 'epoch': 0.2}) +[2026-06-17 07:24:12,067][starVLA.training.train_starvla][INFO] - Step 682, Loss: {'train/loss': 1.0448755025863647, 'train/grad_norm_pre_clip': 7.2860493659973145, 'timing/data': 0.0045219139428809285, 'timing/model': 1.3273279769346118, 'learning_rate/qwen_vl_interface': 9.834206160288698e-06, 'learning_rate/action_model': 9.834206160288698e-05, 'epoch': 0.2}) +[2026-06-17 07:24:16,731][starVLA.training.train_starvla][INFO] - Step 683, Loss: {'train/loss': 0.5642669200897217, 'train/grad_norm_pre_clip': 4.505457878112793, 'timing/data': 0.0050295289838686585, 'timing/model': 1.2051735289860517, 'learning_rate/qwen_vl_interface': 9.8336392845966e-06, 'learning_rate/action_model': 9.833639284596599e-05, 'epoch': 0.2}) +[2026-06-17 07:24:21,481][starVLA.training.train_starvla][INFO] - Step 684, Loss: {'train/loss': 0.8749884366989136, 'train/grad_norm_pre_clip': 6.245537757873535, 'timing/data': 0.004512330051511526, 'timing/model': 1.206713935942389, 'learning_rate/qwen_vl_interface': 9.833071458711146e-06, 'learning_rate/action_model': 9.833071458711146e-05, 'epoch': 0.2}) +[2026-06-17 07:24:26,271][starVLA.training.train_starvla][INFO] - Step 685, Loss: {'train/loss': 1.0190056562423706, 'train/grad_norm_pre_clip': 8.927634239196777, 'timing/data': 0.004665390006266534, 'timing/model': 1.2064199830638245, 'learning_rate/qwen_vl_interface': 9.832502682750048e-06, 'learning_rate/action_model': 9.832502682750047e-05, 'epoch': 0.2}) +[2026-06-17 07:24:31,067][starVLA.training.train_starvla][INFO] - Step 686, Loss: {'train/loss': 0.8705577254295349, 'train/grad_norm_pre_clip': 7.879497528076172, 'timing/data': 0.00464477704372257, 'timing/model': 1.2061674149008468, 'learning_rate/qwen_vl_interface': 9.831932956831212e-06, 'learning_rate/action_model': 9.831932956831211e-05, 'epoch': 0.2}) +[2026-06-17 07:24:35,859][starVLA.training.train_starvla][INFO] - Step 687, Loss: {'train/loss': 1.1388331651687622, 'train/grad_norm_pre_clip': 8.861711502075195, 'timing/data': 0.004617631901055574, 'timing/model': 1.3362603129353374, 'learning_rate/qwen_vl_interface': 9.831362281072743e-06, 'learning_rate/action_model': 9.831362281072742e-05, 'epoch': 0.2}) +[2026-06-17 07:24:40,642][starVLA.training.train_starvla][INFO] - Step 688, Loss: {'train/loss': 0.8626202940940857, 'train/grad_norm_pre_clip': 7.941851615905762, 'timing/data': 0.004921377985738218, 'timing/model': 1.3236771889496595, 'learning_rate/qwen_vl_interface': 9.830790655592944e-06, 'learning_rate/action_model': 9.830790655592943e-05, 'epoch': 0.2}) +[2026-06-17 07:24:45,315][starVLA.training.train_starvla][INFO] - Step 689, Loss: {'train/loss': 0.8529242277145386, 'train/grad_norm_pre_clip': 7.7656569480896, 'timing/data': 0.004680734011344612, 'timing/model': 1.2085006309207529, 'learning_rate/qwen_vl_interface': 9.830218080510313e-06, 'learning_rate/action_model': 9.830218080510312e-05, 'epoch': 0.2}) +[2026-06-17 07:24:50,082][starVLA.training.train_starvla][INFO] - Step 690, Loss: {'train/loss': 0.7932501435279846, 'train/grad_norm_pre_clip': 9.565594673156738, 'timing/data': 0.005438823951408267, 'timing/model': 1.2080765889259055, 'learning_rate/qwen_vl_interface': 9.829644555943545e-06, 'learning_rate/action_model': 9.829644555943545e-05, 'epoch': 0.2}) +[2026-06-17 07:24:54,869][starVLA.training.train_starvla][INFO] - Step 691, Loss: {'train/loss': 1.043086051940918, 'train/grad_norm_pre_clip': 7.746206760406494, 'timing/data': 0.004539672052487731, 'timing/model': 1.208903313963674, 'learning_rate/qwen_vl_interface': 9.82907008201153e-06, 'learning_rate/action_model': 9.829070082011531e-05, 'epoch': 0.2}) +[2026-06-17 07:24:59,685][starVLA.training.train_starvla][INFO] - Step 692, Loss: {'train/loss': 0.9049538969993591, 'train/grad_norm_pre_clip': 10.829312324523926, 'timing/data': 0.005168672068975866, 'timing/model': 1.2067254749126732, 'learning_rate/qwen_vl_interface': 9.828494658833366e-06, 'learning_rate/action_model': 9.828494658833365e-05, 'epoch': 0.2}) +[2026-06-17 07:25:04,445][starVLA.training.train_starvla][INFO] - Step 693, Loss: {'train/loss': 1.248189926147461, 'train/grad_norm_pre_clip': 10.974311828613281, 'timing/data': 0.004489851067773998, 'timing/model': 1.2774979539681226, 'learning_rate/qwen_vl_interface': 9.82791828652833e-06, 'learning_rate/action_model': 9.827918286528329e-05, 'epoch': 0.2}) +[2026-06-17 07:25:09,107][starVLA.training.train_starvla][INFO] - Step 694, Loss: {'train/loss': 1.2978901863098145, 'train/grad_norm_pre_clip': 11.834282875061035, 'timing/data': 0.004434079979546368, 'timing/model': 1.2085487499134615, 'learning_rate/qwen_vl_interface': 9.827340965215903e-06, 'learning_rate/action_model': 9.827340965215904e-05, 'epoch': 0.2}) +[2026-06-17 07:25:13,877][starVLA.training.train_starvla][INFO] - Step 695, Loss: {'train/loss': 0.9572732448577881, 'train/grad_norm_pre_clip': 11.231049537658691, 'timing/data': 0.004318624036386609, 'timing/model': 1.2072460990166292, 'learning_rate/qwen_vl_interface': 9.826762695015773e-06, 'learning_rate/action_model': 9.826762695015772e-05, 'epoch': 0.2}) +[2026-06-17 07:25:18,711][starVLA.training.train_starvla][INFO] - Step 696, Loss: {'train/loss': 1.0122483968734741, 'train/grad_norm_pre_clip': 20.336917877197266, 'timing/data': 0.004619495011866093, 'timing/model': 1.2080042040906847, 'learning_rate/qwen_vl_interface': 9.82618347604781e-06, 'learning_rate/action_model': 9.82618347604781e-05, 'epoch': 0.2}) +[2026-06-17 07:25:23,516][starVLA.training.train_starvla][INFO] - Step 697, Loss: {'train/loss': 0.9174851179122925, 'train/grad_norm_pre_clip': 8.652259826660156, 'timing/data': 0.0047672990476712584, 'timing/model': 1.2055058319820091, 'learning_rate/qwen_vl_interface': 9.825603308432089e-06, 'learning_rate/action_model': 9.825603308432088e-05, 'epoch': 0.21}) +[2026-06-17 07:25:28,307][starVLA.training.train_starvla][INFO] - Step 698, Loss: {'train/loss': 1.2935450077056885, 'train/grad_norm_pre_clip': 64.96460723876953, 'timing/data': 0.004660062957555056, 'timing/model': 1.333891346002929, 'learning_rate/qwen_vl_interface': 9.825022192288878e-06, 'learning_rate/action_model': 9.825022192288877e-05, 'epoch': 0.21}) +[2026-06-17 07:25:33,063][starVLA.training.train_starvla][INFO] - Step 699, Loss: {'train/loss': 1.0984951257705688, 'train/grad_norm_pre_clip': 13.046208381652832, 'timing/data': 0.004589854972437024, 'timing/model': 1.288133380934596, 'learning_rate/qwen_vl_interface': 9.824440127738643e-06, 'learning_rate/action_model': 9.824440127738642e-05, 'epoch': 0.21}) +[2026-06-17 07:27:15,213][starVLA.training.train_starvla][INFO] - Step 700, Loss: {'train/loss': 1.1737459897994995, 'train/grad_norm_pre_clip': 12.58040714263916, 'eval/loss': 1.0181373357772827, 'eval/latency_0/loss': 1.0181373357772827, 'eval/demon_fire_precision': 0.8640316205533597, 'eval/demon_fire_recall': 0.9078073089700996, 'eval/demon_fire_f1': 0.8853786958282706, 'eval/demon_fire_rate_model': 0.703559510567297, 'eval/demon_fire_rate_teacher': 0.6696329254727474, 'eval/demon_left_precision': 0.7914614121510674, 'eval/demon_left_recall': 0.5461756373937677, 'eval/demon_left_f1': 0.6463291987931612, 'eval/demon_left_rate_model': 0.3387096774193548, 'eval/demon_left_rate_teacher': 0.49082313681868744, 'eval/demon_right_precision': 0.6505466778805719, 'eval/demon_right_recall': 0.8740112994350282, 'eval/demon_right_f1': 0.7459016393442621, 'eval/demon_right_rate_model': 0.6612903225806451, 'eval/demon_right_rate_teacher': 0.4922135706340378, 'eval/demon_move_f1': 0.6961154190687117, 'eval/demon_cc_f1': 0.8186085062722319, 'eval/latency_0/demon_fire_precision': 0.8640316205533597, 'eval/latency_0/demon_fire_recall': 0.9078073089700996, 'eval/latency_0/demon_fire_f1': 0.8853786958282706, 'eval/latency_0/demon_fire_rate_model': 0.703559510567297, 'eval/latency_0/demon_fire_rate_teacher': 0.6696329254727474, 'eval/latency_0/demon_left_precision': 0.7914614121510674, 'eval/latency_0/demon_left_recall': 0.5461756373937677, 'eval/latency_0/demon_left_f1': 0.6463291987931612, 'eval/latency_0/demon_left_rate_model': 0.3387096774193548, 'eval/latency_0/demon_left_rate_teacher': 0.49082313681868744, 'eval/latency_0/demon_right_precision': 0.6505466778805719, 'eval/latency_0/demon_right_recall': 0.8740112994350282, 'eval/latency_0/demon_right_f1': 0.7459016393442621, 'eval/latency_0/demon_right_rate_model': 0.6612903225806451, 'eval/latency_0/demon_right_rate_teacher': 0.4922135706340378, 'eval/latency_0/demon_move_f1': 0.6961154190687117, 'eval/latency_0/demon_cc_f1': 0.8186085062722319, 'timing/data': 0.004282254027202725, 'timing/model': 1.2079900050302967, 'learning_rate/qwen_vl_interface': 9.823857114902047e-06, 'learning_rate/action_model': 9.823857114902047e-05, 'epoch': 0.21}) +[2026-06-17 07:27:15,213][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_700_state +[2026-06-17 07:27:15,216][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 07:27:18,734][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_700_state/model.safetensors +[2026-06-17 07:27:26,466][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_700_state/optimizer.bin +[2026-06-17 07:27:26,466][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_700_state/sampler.bin +[2026-06-17 07:27:26,467][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_700_state/sampler_1.bin +[2026-06-17 07:27:26,467][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_700_state/random_states_0.pkl +[2026-06-17 07:27:26,507][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 700 +[2026-06-17 07:27:26,507][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 07:27:26,544][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 07:27:31,173][starVLA.training.train_starvla][INFO] - Step 701, Loss: {'train/loss': 0.9922186136245728, 'train/grad_norm_pre_clip': 12.363601684570312, 'timing/data': 0.005000483011826873, 'timing/model': 1.2020854250295088, 'learning_rate/qwen_vl_interface': 9.823273153899951e-06, 'learning_rate/action_model': 9.82327315389995e-05, 'epoch': 0.21}) +[2026-06-17 07:27:35,779][starVLA.training.train_starvla][INFO] - Step 702, Loss: {'train/loss': 1.1309154033660889, 'train/grad_norm_pre_clip': 8.89065933227539, 'timing/data': 0.004618839011527598, 'timing/model': 1.1908846509177238, 'learning_rate/qwen_vl_interface': 9.822688244853407e-06, 'learning_rate/action_model': 9.822688244853408e-05, 'epoch': 0.21}) +[2026-06-17 07:27:40,499][starVLA.training.train_starvla][INFO] - Step 703, Loss: {'train/loss': 1.2253031730651855, 'train/grad_norm_pre_clip': 8.31715202331543, 'timing/data': 0.004861744935624301, 'timing/model': 1.2024416829226539, 'learning_rate/qwen_vl_interface': 9.822102387883672e-06, 'learning_rate/action_model': 9.822102387883671e-05, 'epoch': 0.21}) +[2026-06-17 07:27:45,163][starVLA.training.train_starvla][INFO] - Step 704, Loss: {'train/loss': 0.8517147898674011, 'train/grad_norm_pre_clip': 7.350673198699951, 'timing/data': 0.0048500909470021725, 'timing/model': 1.1978057739324868, 'learning_rate/qwen_vl_interface': 9.82151558311219e-06, 'learning_rate/action_model': 9.821515583112191e-05, 'epoch': 0.21}) +[2026-06-17 07:27:49,914][starVLA.training.train_starvla][INFO] - Step 705, Loss: {'train/loss': 0.8658341765403748, 'train/grad_norm_pre_clip': 7.8898115158081055, 'timing/data': 0.004675325937569141, 'timing/model': 1.1967602749355137, 'learning_rate/qwen_vl_interface': 9.82092783066061e-06, 'learning_rate/action_model': 9.82092783066061e-05, 'epoch': 0.21}) +[2026-06-17 07:27:54,730][starVLA.training.train_starvla][INFO] - Step 706, Loss: {'train/loss': 1.1667027473449707, 'train/grad_norm_pre_clip': 10.092893600463867, 'timing/data': 0.0055628829868510365, 'timing/model': 1.3058053300483152, 'learning_rate/qwen_vl_interface': 9.820339130650771e-06, 'learning_rate/action_model': 9.820339130650771e-05, 'epoch': 0.21}) +[2026-06-17 07:27:59,476][starVLA.training.train_starvla][INFO] - Step 707, Loss: {'train/loss': 1.1690670251846313, 'train/grad_norm_pre_clip': 9.596442222595215, 'timing/data': 0.004624382010661066, 'timing/model': 1.1995887940283865, 'learning_rate/qwen_vl_interface': 9.819749483204715e-06, 'learning_rate/action_model': 9.819749483204715e-05, 'epoch': 0.21}) +[2026-06-17 07:28:04,243][starVLA.training.train_starvla][INFO] - Step 708, Loss: {'train/loss': 1.0087798833847046, 'train/grad_norm_pre_clip': 10.85287857055664, 'timing/data': 0.004515143926255405, 'timing/model': 1.1984505440341309, 'learning_rate/qwen_vl_interface': 9.81915888844467e-06, 'learning_rate/action_model': 9.81915888844467e-05, 'epoch': 0.21}) +[2026-06-17 07:28:09,014][starVLA.training.train_starvla][INFO] - Step 709, Loss: {'train/loss': 0.871373176574707, 'train/grad_norm_pre_clip': 7.032249927520752, 'timing/data': 0.004639070015400648, 'timing/model': 1.2005908370483667, 'learning_rate/qwen_vl_interface': 9.818567346493072e-06, 'learning_rate/action_model': 9.818567346493071e-05, 'epoch': 0.21}) +[2026-06-17 07:28:13,790][starVLA.training.train_starvla][INFO] - Step 710, Loss: {'train/loss': 1.0172513723373413, 'train/grad_norm_pre_clip': 14.166349411010742, 'timing/data': 0.004791759070940316, 'timing/model': 1.202781518921256, 'learning_rate/qwen_vl_interface': 9.817974857472549e-06, 'learning_rate/action_model': 9.817974857472548e-05, 'epoch': 0.21}) +[2026-06-17 07:28:18,516][starVLA.training.train_starvla][INFO] - Step 711, Loss: {'train/loss': 0.673382580280304, 'train/grad_norm_pre_clip': 6.327634334564209, 'timing/data': 0.004652556963264942, 'timing/model': 1.2013848730130121, 'learning_rate/qwen_vl_interface': 9.817381421505922e-06, 'learning_rate/action_model': 9.817381421505921e-05, 'epoch': 0.21}) +[2026-06-17 07:28:23,302][starVLA.training.train_starvla][INFO] - Step 712, Loss: {'train/loss': 1.3056859970092773, 'train/grad_norm_pre_clip': 7.997232913970947, 'timing/data': 0.004616993013769388, 'timing/model': 1.3309538839384913, 'learning_rate/qwen_vl_interface': 9.81678703871621e-06, 'learning_rate/action_model': 9.81678703871621e-05, 'epoch': 0.21}) +[2026-06-17 07:28:27,949][starVLA.training.train_starvla][INFO] - Step 713, Loss: {'train/loss': 1.1544876098632812, 'train/grad_norm_pre_clip': 9.820037841796875, 'timing/data': 0.004933554911985993, 'timing/model': 1.2043308930005878, 'learning_rate/qwen_vl_interface': 9.816191709226631e-06, 'learning_rate/action_model': 9.816191709226632e-05, 'epoch': 0.21}) +[2026-06-17 07:28:32,758][starVLA.training.train_starvla][INFO] - Step 714, Loss: {'train/loss': 0.7813801765441895, 'train/grad_norm_pre_clip': 8.619942665100098, 'timing/data': 0.004607593989931047, 'timing/model': 1.2032507340190932, 'learning_rate/qwen_vl_interface': 9.815595433160599e-06, 'learning_rate/action_model': 9.815595433160598e-05, 'epoch': 0.21}) +[2026-06-17 07:28:37,532][starVLA.training.train_starvla][INFO] - Step 715, Loss: {'train/loss': 1.2654922008514404, 'train/grad_norm_pre_clip': 8.278552055358887, 'timing/data': 0.004809412988834083, 'timing/model': 1.202691156999208, 'learning_rate/qwen_vl_interface': 9.81499821064172e-06, 'learning_rate/action_model': 9.81499821064172e-05, 'epoch': 0.21}) +[2026-06-17 07:28:42,315][starVLA.training.train_starvla][INFO] - Step 716, Loss: {'train/loss': 1.0533661842346191, 'train/grad_norm_pre_clip': 17.964296340942383, 'timing/data': 0.004675064934417605, 'timing/model': 1.2048071069875732, 'learning_rate/qwen_vl_interface': 9.814400041793799e-06, 'learning_rate/action_model': 9.814400041793798e-05, 'epoch': 0.21}) +[2026-06-17 07:28:47,070][starVLA.training.train_starvla][INFO] - Step 717, Loss: {'train/loss': 0.9008931517601013, 'train/grad_norm_pre_clip': 8.877382278442383, 'timing/data': 0.005084962002001703, 'timing/model': 1.2028298799414188, 'learning_rate/qwen_vl_interface': 9.813800926740839e-06, 'learning_rate/action_model': 9.813800926740838e-05, 'epoch': 0.21}) +[2026-06-17 07:28:51,849][starVLA.training.train_starvla][INFO] - Step 718, Loss: {'train/loss': 1.2253226041793823, 'train/grad_norm_pre_clip': 10.417803764343262, 'timing/data': 0.00470615504309535, 'timing/model': 1.3279870530823246, 'learning_rate/qwen_vl_interface': 9.813200865607034e-06, 'learning_rate/action_model': 9.813200865607034e-05, 'epoch': 0.21}) +[2026-06-17 07:28:56,543][starVLA.training.train_starvla][INFO] - Step 719, Loss: {'train/loss': 0.731217086315155, 'train/grad_norm_pre_clip': 5.8343610763549805, 'timing/data': 0.0044365080539137125, 'timing/model': 1.2037885149475187, 'learning_rate/qwen_vl_interface': 9.812599858516782e-06, 'learning_rate/action_model': 9.812599858516782e-05, 'epoch': 0.21}) +[2026-06-17 07:29:01,328][starVLA.training.train_starvla][INFO] - Step 720, Loss: {'train/loss': 1.0585806369781494, 'train/grad_norm_pre_clip': 9.94904613494873, 'timing/data': 0.004705336992628872, 'timing/model': 1.20508005307056, 'learning_rate/qwen_vl_interface': 9.81199790559467e-06, 'learning_rate/action_model': 9.811997905594668e-05, 'epoch': 0.21}) +[2026-06-17 07:29:06,136][starVLA.training.train_starvla][INFO] - Step 721, Loss: {'train/loss': 1.068676471710205, 'train/grad_norm_pre_clip': 8.550434112548828, 'timing/data': 0.004732309957034886, 'timing/model': 1.2125600239960477, 'learning_rate/qwen_vl_interface': 9.811395006965481e-06, 'learning_rate/action_model': 9.81139500696548e-05, 'epoch': 0.21}) +[2026-06-17 07:29:10,958][starVLA.training.train_starvla][INFO] - Step 722, Loss: {'train/loss': 1.0114753246307373, 'train/grad_norm_pre_clip': 8.68110466003418, 'timing/data': 0.004950056900270283, 'timing/model': 1.2101896740496159, 'learning_rate/qwen_vl_interface': 9.810791162754199e-06, 'learning_rate/action_model': 9.810791162754198e-05, 'epoch': 0.21}) +[2026-06-17 07:29:15,766][starVLA.training.train_starvla][INFO] - Step 723, Loss: {'train/loss': 1.1672089099884033, 'train/grad_norm_pre_clip': 9.439878463745117, 'timing/data': 0.004637312958948314, 'timing/model': 1.350662634940818, 'learning_rate/qwen_vl_interface': 9.810186373086002e-06, 'learning_rate/action_model': 9.810186373086002e-05, 'epoch': 0.21}) +[2026-06-17 07:29:20,437][starVLA.training.train_starvla][INFO] - Step 724, Loss: {'train/loss': 1.0916460752487183, 'train/grad_norm_pre_clip': 9.07809066772461, 'timing/data': 0.004829148994758725, 'timing/model': 1.2105124109657481, 'learning_rate/qwen_vl_interface': 9.809580638086264e-06, 'learning_rate/action_model': 9.809580638086263e-05, 'epoch': 0.21}) +[2026-06-17 07:29:25,233][starVLA.training.train_starvla][INFO] - Step 725, Loss: {'train/loss': 1.0763683319091797, 'train/grad_norm_pre_clip': 8.572784423828125, 'timing/data': 0.004681983031332493, 'timing/model': 1.2049312599701807, 'learning_rate/qwen_vl_interface': 9.80897395788055e-06, 'learning_rate/action_model': 9.808973957880551e-05, 'epoch': 0.21}) +[2026-06-17 07:29:29,942][starVLA.training.train_starvla][INFO] - Step 726, Loss: {'train/loss': 1.0286225080490112, 'train/grad_norm_pre_clip': 7.778501987457275, 'timing/data': 0.005132730002515018, 'timing/model': 1.2062236610800028, 'learning_rate/qwen_vl_interface': 9.808366332594632e-06, 'learning_rate/action_model': 9.808366332594632e-05, 'epoch': 0.21}) +[2026-06-17 07:29:34,684][starVLA.training.train_starvla][INFO] - Step 727, Loss: {'train/loss': 1.2580697536468506, 'train/grad_norm_pre_clip': 8.45603084564209, 'timing/data': 0.004780089017003775, 'timing/model': 1.2079371500294656, 'learning_rate/qwen_vl_interface': 9.807757762354467e-06, 'learning_rate/action_model': 9.807757762354467e-05, 'epoch': 0.21}) +[2026-06-17 07:29:39,493][starVLA.training.train_starvla][INFO] - Step 728, Loss: {'train/loss': 1.1609842777252197, 'train/grad_norm_pre_clip': 16.716047286987305, 'timing/data': 0.004742547054775059, 'timing/model': 1.2070045330328867, 'learning_rate/qwen_vl_interface': 9.807148247286215e-06, 'learning_rate/action_model': 9.807148247286215e-05, 'epoch': 0.21}) +[2026-06-17 07:29:44,366][starVLA.training.train_starvla][INFO] - Step 729, Loss: {'train/loss': 1.0102014541625977, 'train/grad_norm_pre_clip': 19.262500762939453, 'timing/data': 0.0047923329984769225, 'timing/model': 1.3476933760102838, 'learning_rate/qwen_vl_interface': 9.806537787516226e-06, 'learning_rate/action_model': 9.806537787516226e-05, 'epoch': 0.21}) +[2026-06-17 07:29:49,048][starVLA.training.train_starvla][INFO] - Step 730, Loss: {'train/loss': 0.968076765537262, 'train/grad_norm_pre_clip': 7.944334506988525, 'timing/data': 0.004876798018813133, 'timing/model': 1.2070814239559695, 'learning_rate/qwen_vl_interface': 9.80592638317105e-06, 'learning_rate/action_model': 9.80592638317105e-05, 'epoch': 0.21}) +[2026-06-17 07:29:53,856][starVLA.training.train_starvla][INFO] - Step 731, Loss: {'train/loss': 0.9218912124633789, 'train/grad_norm_pre_clip': 8.143880844116211, 'timing/data': 0.0043815739918500185, 'timing/model': 1.210879899095744, 'learning_rate/qwen_vl_interface': 9.805314034377434e-06, 'learning_rate/action_model': 9.805314034377434e-05, 'epoch': 0.22}) +[2026-06-17 07:29:58,682][starVLA.training.train_starvla][INFO] - Step 732, Loss: {'train/loss': 0.9034087657928467, 'train/grad_norm_pre_clip': 7.53599214553833, 'timing/data': 0.004634003969840705, 'timing/model': 1.220474564936012, 'learning_rate/qwen_vl_interface': 9.804700741262318e-06, 'learning_rate/action_model': 9.804700741262317e-05, 'epoch': 0.22}) +[2026-06-17 07:30:03,465][starVLA.training.train_starvla][INFO] - Step 733, Loss: {'train/loss': 0.7930718660354614, 'train/grad_norm_pre_clip': 8.012310981750488, 'timing/data': 0.0046754940412938595, 'timing/model': 1.2159446630394086, 'learning_rate/qwen_vl_interface': 9.804086503952835e-06, 'learning_rate/action_model': 9.804086503952835e-05, 'epoch': 0.22}) +[2026-06-17 07:30:08,271][starVLA.training.train_starvla][INFO] - Step 734, Loss: {'train/loss': 0.8482650518417358, 'train/grad_norm_pre_clip': 7.976062774658203, 'timing/data': 0.0045184900518506765, 'timing/model': 1.2051628299523145, 'learning_rate/qwen_vl_interface': 9.803471322576319e-06, 'learning_rate/action_model': 9.803471322576319e-05, 'epoch': 0.22}) +[2026-06-17 07:30:13,069][starVLA.training.train_starvla][INFO] - Step 735, Loss: {'train/loss': 0.8525248169898987, 'train/grad_norm_pre_clip': 7.371574878692627, 'timing/data': 0.004682119004428387, 'timing/model': 1.340027056983672, 'learning_rate/qwen_vl_interface': 9.8028551972603e-06, 'learning_rate/action_model': 9.802855197260299e-05, 'epoch': 0.22}) +[2026-06-17 07:30:17,734][starVLA.training.train_starvla][INFO] - Step 736, Loss: {'train/loss': 1.194260835647583, 'train/grad_norm_pre_clip': 7.684018611907959, 'timing/data': 0.004732219967991114, 'timing/model': 1.2071701879613101, 'learning_rate/qwen_vl_interface': 9.802238128132498e-06, 'learning_rate/action_model': 9.802238128132498e-05, 'epoch': 0.22}) +[2026-06-17 07:30:22,531][starVLA.training.train_starvla][INFO] - Step 737, Loss: {'train/loss': 1.0474156141281128, 'train/grad_norm_pre_clip': 13.018638610839844, 'timing/data': 0.0047255579847842455, 'timing/model': 1.2053000800078735, 'learning_rate/qwen_vl_interface': 9.801620115320836e-06, 'learning_rate/action_model': 9.801620115320835e-05, 'epoch': 0.22}) +[2026-06-17 07:30:27,332][starVLA.training.train_starvla][INFO] - Step 738, Loss: {'train/loss': 1.1384235620498657, 'train/grad_norm_pre_clip': 10.618363380432129, 'timing/data': 0.004509652964770794, 'timing/model': 1.2070695899892598, 'learning_rate/qwen_vl_interface': 9.801001158953425e-06, 'learning_rate/action_model': 9.801001158953423e-05, 'epoch': 0.22}) +[2026-06-17 07:30:32,136][starVLA.training.train_starvla][INFO] - Step 739, Loss: {'train/loss': 0.9921662211418152, 'train/grad_norm_pre_clip': 9.383599281311035, 'timing/data': 0.005316239083185792, 'timing/model': 1.216138733085245, 'learning_rate/qwen_vl_interface': 9.800381259158575e-06, 'learning_rate/action_model': 9.800381259158574e-05, 'epoch': 0.22}) +[2026-06-17 07:30:36,824][starVLA.training.train_starvla][INFO] - Step 740, Loss: {'train/loss': 1.077879548072815, 'train/grad_norm_pre_clip': 7.9979424476623535, 'timing/data': 0.00445468001998961, 'timing/model': 1.206705637043342, 'learning_rate/qwen_vl_interface': 9.799760416064794e-06, 'learning_rate/action_model': 9.799760416064793e-05, 'epoch': 0.22}) +[2026-06-17 07:30:41,545][starVLA.training.train_starvla][INFO] - Step 741, Loss: {'train/loss': 1.039915680885315, 'train/grad_norm_pre_clip': 6.837924957275391, 'timing/data': 0.004535465035587549, 'timing/model': 1.2574842410394922, 'learning_rate/qwen_vl_interface': 9.799138629800785e-06, 'learning_rate/action_model': 9.799138629800785e-05, 'epoch': 0.22}) +[2026-06-17 07:30:46,207][starVLA.training.train_starvla][INFO] - Step 742, Loss: {'train/loss': 1.0449837446212769, 'train/grad_norm_pre_clip': 10.252657890319824, 'timing/data': 0.004619132028892636, 'timing/model': 1.2073011320317164, 'learning_rate/qwen_vl_interface': 9.798515900495441e-06, 'learning_rate/action_model': 9.79851590049544e-05, 'epoch': 0.22}) +[2026-06-17 07:30:50,946][starVLA.training.train_starvla][INFO] - Step 743, Loss: {'train/loss': 1.1705912351608276, 'train/grad_norm_pre_clip': 7.696895599365234, 'timing/data': 0.004312740988098085, 'timing/model': 1.208964160992764, 'learning_rate/qwen_vl_interface': 9.797892228277857e-06, 'learning_rate/action_model': 9.797892228277857e-05, 'epoch': 0.22}) +[2026-06-17 07:30:55,756][starVLA.training.train_starvla][INFO] - Step 744, Loss: {'train/loss': 1.023128628730774, 'train/grad_norm_pre_clip': 7.925278186798096, 'timing/data': 0.004295454011298716, 'timing/model': 1.2097186649916694, 'learning_rate/qwen_vl_interface': 9.797267613277319e-06, 'learning_rate/action_model': 9.797267613277319e-05, 'epoch': 0.22}) +[2026-06-17 07:31:00,565][starVLA.training.train_starvla][INFO] - Step 745, Loss: {'train/loss': 0.9637953639030457, 'train/grad_norm_pre_clip': 10.098705291748047, 'timing/data': 0.0046333810314536095, 'timing/model': 1.2073857309296727, 'learning_rate/qwen_vl_interface': 9.796642055623314e-06, 'learning_rate/action_model': 9.796642055623314e-05, 'epoch': 0.22}) +[2026-06-17 07:31:05,374][starVLA.training.train_starvla][INFO] - Step 746, Loss: {'train/loss': 0.8411317467689514, 'train/grad_norm_pre_clip': 7.701117992401123, 'timing/data': 0.004502388997934759, 'timing/model': 1.3354888960020617, 'learning_rate/qwen_vl_interface': 9.796015555445518e-06, 'learning_rate/action_model': 9.796015555445517e-05, 'epoch': 0.22}) +[2026-06-17 07:31:10,172][starVLA.training.train_starvla][INFO] - Step 747, Loss: {'train/loss': 1.1005165576934814, 'train/grad_norm_pre_clip': 7.197655200958252, 'timing/data': 0.0048650449607521296, 'timing/model': 1.3321837580297142, 'learning_rate/qwen_vl_interface': 9.795388112873806e-06, 'learning_rate/action_model': 9.795388112873806e-05, 'epoch': 0.22}) +[2026-06-17 07:31:14,848][starVLA.training.train_starvla][INFO] - Step 748, Loss: {'train/loss': 0.9189667105674744, 'train/grad_norm_pre_clip': 5.74833345413208, 'timing/data': 0.004602228989824653, 'timing/model': 1.2117560299811885, 'learning_rate/qwen_vl_interface': 9.794759728038246e-06, 'learning_rate/action_model': 9.794759728038246e-05, 'epoch': 0.22}) +[2026-06-17 07:31:19,618][starVLA.training.train_starvla][INFO] - Step 749, Loss: {'train/loss': 0.9162732362747192, 'train/grad_norm_pre_clip': 8.578031539916992, 'timing/data': 0.004525658907368779, 'timing/model': 1.2143962680129334, 'learning_rate/qwen_vl_interface': 9.794130401069105e-06, 'learning_rate/action_model': 9.794130401069104e-05, 'epoch': 0.22}) +[2026-06-17 07:31:24,449][starVLA.training.train_starvla][INFO] - Step 750, Loss: {'train/loss': 1.05158531665802, 'train/grad_norm_pre_clip': 9.457328796386719, 'timing/data': 0.004416463081724942, 'timing/model': 1.216600484913215, 'learning_rate/qwen_vl_interface': 9.79350013209684e-06, 'learning_rate/action_model': 9.79350013209684e-05, 'epoch': 0.22}) +[2026-06-17 07:31:29,248][starVLA.training.train_starvla][INFO] - Step 751, Loss: {'train/loss': 0.9675568342208862, 'train/grad_norm_pre_clip': 8.885971069335938, 'timing/data': 0.004384389030747116, 'timing/model': 1.2119746500393376, 'learning_rate/qwen_vl_interface': 9.792868921252112e-06, 'learning_rate/action_model': 9.792868921252111e-05, 'epoch': 0.22}) +[2026-06-17 07:31:34,068][starVLA.training.train_starvla][INFO] - Step 752, Loss: {'train/loss': 0.9573696255683899, 'train/grad_norm_pre_clip': 8.727095603942871, 'timing/data': 0.004684615065343678, 'timing/model': 1.345220401068218, 'learning_rate/qwen_vl_interface': 9.792236768665765e-06, 'learning_rate/action_model': 9.792236768665764e-05, 'epoch': 0.22}) +[2026-06-17 07:31:38,864][starVLA.training.train_starvla][INFO] - Step 753, Loss: {'train/loss': 0.8080765604972839, 'train/grad_norm_pre_clip': 7.333009719848633, 'timing/data': 0.00435241695959121, 'timing/model': 1.3363941869465634, 'learning_rate/qwen_vl_interface': 9.79160367446885e-06, 'learning_rate/action_model': 9.79160367446885e-05, 'epoch': 0.22}) +[2026-06-17 07:31:43,528][starVLA.training.train_starvla][INFO] - Step 754, Loss: {'train/loss': 1.0288872718811035, 'train/grad_norm_pre_clip': 6.489475727081299, 'timing/data': 0.004422605969011784, 'timing/model': 1.2125689020613208, 'learning_rate/qwen_vl_interface': 9.790969638792606e-06, 'learning_rate/action_model': 9.790969638792606e-05, 'epoch': 0.22}) +[2026-06-17 07:31:48,330][starVLA.training.train_starvla][INFO] - Step 755, Loss: {'train/loss': 0.8035586476325989, 'train/grad_norm_pre_clip': 6.465859889984131, 'timing/data': 0.0046125269727781415, 'timing/model': 1.2083301859674975, 'learning_rate/qwen_vl_interface': 9.790334661768471e-06, 'learning_rate/action_model': 9.79033466176847e-05, 'epoch': 0.22}) +[2026-06-17 07:31:53,135][starVLA.training.train_starvla][INFO] - Step 756, Loss: {'train/loss': 0.7695139646530151, 'train/grad_norm_pre_clip': 7.209883213043213, 'timing/data': 0.004719362943433225, 'timing/model': 1.209684618981555, 'learning_rate/qwen_vl_interface': 9.78969874352807e-06, 'learning_rate/action_model': 9.78969874352807e-05, 'epoch': 0.22}) +[2026-06-17 07:31:57,933][starVLA.training.train_starvla][INFO] - Step 757, Loss: {'train/loss': 0.8691049814224243, 'train/grad_norm_pre_clip': 7.652405738830566, 'timing/data': 0.004528549965471029, 'timing/model': 1.20904026995413, 'learning_rate/qwen_vl_interface': 9.78906188420324e-06, 'learning_rate/action_model': 9.78906188420324e-05, 'epoch': 0.22}) +[2026-06-17 07:32:02,685][starVLA.training.train_starvla][INFO] - Step 758, Loss: {'train/loss': 0.7483629584312439, 'train/grad_norm_pre_clip': 19.275625228881836, 'timing/data': 0.004863762995228171, 'timing/model': 1.283524307073094, 'learning_rate/qwen_vl_interface': 9.788424083925996e-06, 'learning_rate/action_model': 9.788424083925995e-05, 'epoch': 0.22}) +[2026-06-17 07:32:07,359][starVLA.training.train_starvla][INFO] - Step 759, Loss: {'train/loss': 0.9577150940895081, 'train/grad_norm_pre_clip': 9.268375396728516, 'timing/data': 0.004635702003724873, 'timing/model': 1.2099081409396604, 'learning_rate/qwen_vl_interface': 9.787785342828554e-06, 'learning_rate/action_model': 9.787785342828553e-05, 'epoch': 0.22}) +[2026-06-17 07:32:12,133][starVLA.training.train_starvla][INFO] - Step 760, Loss: {'train/loss': 1.024100661277771, 'train/grad_norm_pre_clip': 7.194891929626465, 'timing/data': 0.004650954972021282, 'timing/model': 1.2083525949856266, 'learning_rate/qwen_vl_interface': 9.78714566104333e-06, 'learning_rate/action_model': 9.787145661043329e-05, 'epoch': 0.22}) +[2026-06-17 07:32:16,924][starVLA.training.train_starvla][INFO] - Step 761, Loss: {'train/loss': 1.14792001247406, 'train/grad_norm_pre_clip': 10.4568452835083, 'timing/data': 0.004599591018632054, 'timing/model': 1.2068364539882168, 'learning_rate/qwen_vl_interface': 9.786505038702926e-06, 'learning_rate/action_model': 9.786505038702925e-05, 'epoch': 0.22}) +[2026-06-17 07:32:21,715][starVLA.training.train_starvla][INFO] - Step 762, Loss: {'train/loss': 0.8894978165626526, 'train/grad_norm_pre_clip': 7.665806770324707, 'timing/data': 0.0045140840811654925, 'timing/model': 1.2074664259562269, 'learning_rate/qwen_vl_interface': 9.785863475940147e-06, 'learning_rate/action_model': 9.785863475940145e-05, 'epoch': 0.22}) +[2026-06-17 07:32:26,518][starVLA.training.train_starvla][INFO] - Step 763, Loss: {'train/loss': 0.6831657886505127, 'train/grad_norm_pre_clip': 7.656139373779297, 'timing/data': 0.004830282996408641, 'timing/model': 1.3471373840002343, 'learning_rate/qwen_vl_interface': 9.785220972887988e-06, 'learning_rate/action_model': 9.785220972887989e-05, 'epoch': 0.22}) +[2026-06-17 07:32:31,279][starVLA.training.train_starvla][INFO] - Step 764, Loss: {'train/loss': 1.0836634635925293, 'train/grad_norm_pre_clip': 13.30543041229248, 'timing/data': 0.004579097963869572, 'timing/model': 1.2909586309688166, 'learning_rate/qwen_vl_interface': 9.784577529679643e-06, 'learning_rate/action_model': 9.784577529679643e-05, 'epoch': 0.22}) +[2026-06-17 07:32:35,947][starVLA.training.train_starvla][INFO] - Step 765, Loss: {'train/loss': 1.0775268077850342, 'train/grad_norm_pre_clip': 8.141190528869629, 'timing/data': 0.0044856289168819785, 'timing/model': 1.2082226870115846, 'learning_rate/qwen_vl_interface': 9.783933146448495e-06, 'learning_rate/action_model': 9.783933146448495e-05, 'epoch': 0.23}) +[2026-06-17 07:32:40,762][starVLA.training.train_starvla][INFO] - Step 766, Loss: {'train/loss': 0.6918458342552185, 'train/grad_norm_pre_clip': 6.601411819458008, 'timing/data': 0.005110090947709978, 'timing/model': 1.2068316499935463, 'learning_rate/qwen_vl_interface': 9.783287823328127e-06, 'learning_rate/action_model': 9.783287823328127e-05, 'epoch': 0.23}) +[2026-06-17 07:32:45,571][starVLA.training.train_starvla][INFO] - Step 767, Loss: {'train/loss': 1.264961838722229, 'train/grad_norm_pre_clip': 8.38205337524414, 'timing/data': 0.004867142997682095, 'timing/model': 1.207259224029258, 'learning_rate/qwen_vl_interface': 9.782641560452317e-06, 'learning_rate/action_model': 9.782641560452316e-05, 'epoch': 0.23}) +[2026-06-17 07:32:50,375][starVLA.training.train_starvla][INFO] - Step 768, Loss: {'train/loss': 1.0267245769500732, 'train/grad_norm_pre_clip': 10.8897123336792, 'timing/data': 0.004598458996042609, 'timing/model': 1.2086953389225528, 'learning_rate/qwen_vl_interface': 9.781994357955032e-06, 'learning_rate/action_model': 9.781994357955032e-05, 'epoch': 0.23}) +[2026-06-17 07:32:55,185][starVLA.training.train_starvla][INFO] - Step 769, Loss: {'train/loss': 1.2437443733215332, 'train/grad_norm_pre_clip': 14.463810920715332, 'timing/data': 0.004511932958848774, 'timing/model': 1.3472312829690054, 'learning_rate/qwen_vl_interface': 9.78134621597044e-06, 'learning_rate/action_model': 9.781346215970441e-05, 'epoch': 0.23}) +[2026-06-17 07:32:59,861][starVLA.training.train_starvla][INFO] - Step 770, Loss: {'train/loss': 1.2396358251571655, 'train/grad_norm_pre_clip': 11.286087036132812, 'timing/data': 0.005242869956418872, 'timing/model': 1.213831530069001, 'learning_rate/qwen_vl_interface': 9.780697134632903e-06, 'learning_rate/action_model': 9.780697134632902e-05, 'epoch': 0.23}) +[2026-06-17 07:33:04,683][starVLA.training.train_starvla][INFO] - Step 771, Loss: {'train/loss': 1.1195447444915771, 'train/grad_norm_pre_clip': 11.924596786499023, 'timing/data': 0.004978996003046632, 'timing/model': 1.2204412700375542, 'learning_rate/qwen_vl_interface': 9.780047114076974e-06, 'learning_rate/action_model': 9.780047114076973e-05, 'epoch': 0.23}) +[2026-06-17 07:33:09,482][starVLA.training.train_starvla][INFO] - Step 772, Loss: {'train/loss': 1.344894528388977, 'train/grad_norm_pre_clip': 11.050817489624023, 'timing/data': 0.004840475972741842, 'timing/model': 1.2087980610085651, 'learning_rate/qwen_vl_interface': 9.779396154437404e-06, 'learning_rate/action_model': 9.779396154437403e-05, 'epoch': 0.23}) +[2026-06-17 07:33:14,201][starVLA.training.train_starvla][INFO] - Step 773, Loss: {'train/loss': 0.7620619535446167, 'train/grad_norm_pre_clip': 7.647411823272705, 'timing/data': 0.004642363055609167, 'timing/model': 1.2143689319491386, 'learning_rate/qwen_vl_interface': 9.778744255849138e-06, 'learning_rate/action_model': 9.778744255849138e-05, 'epoch': 0.23}) +[2026-06-17 07:33:18,996][starVLA.training.train_starvla][INFO] - Step 774, Loss: {'train/loss': 1.3133686780929565, 'train/grad_norm_pre_clip': 15.697487831115723, 'timing/data': 0.0045811719028279185, 'timing/model': 1.2067683869972825, 'learning_rate/qwen_vl_interface': 9.778091418447313e-06, 'learning_rate/action_model': 9.778091418447312e-05, 'epoch': 0.23}) +[2026-06-17 07:33:23,762][starVLA.training.train_starvla][INFO] - Step 775, Loss: {'train/loss': 0.8355026841163635, 'train/grad_norm_pre_clip': 7.527587890625, 'timing/data': 0.004442312056198716, 'timing/model': 1.304477600962855, 'learning_rate/qwen_vl_interface': 9.777437642367265e-06, 'learning_rate/action_model': 9.777437642367264e-05, 'epoch': 0.23}) +[2026-06-17 07:33:28,430][starVLA.training.train_starvla][INFO] - Step 776, Loss: {'train/loss': 1.1961272954940796, 'train/grad_norm_pre_clip': 57.11073684692383, 'timing/data': 0.0048374050529673696, 'timing/model': 1.211743028019555, 'learning_rate/qwen_vl_interface': 9.776782927744523e-06, 'learning_rate/action_model': 9.776782927744522e-05, 'epoch': 0.23}) +[2026-06-17 07:33:33,240][starVLA.training.train_starvla][INFO] - Step 777, Loss: {'train/loss': 1.1994529962539673, 'train/grad_norm_pre_clip': 9.964193344116211, 'timing/data': 0.006220531999133527, 'timing/model': 1.2182563410606235, 'learning_rate/qwen_vl_interface': 9.776127274714808e-06, 'learning_rate/action_model': 9.776127274714807e-05, 'epoch': 0.23}) +[2026-06-17 07:33:38,030][starVLA.training.train_starvla][INFO] - Step 778, Loss: {'train/loss': 1.1246771812438965, 'train/grad_norm_pre_clip': 11.534682273864746, 'timing/data': 0.004538628971204162, 'timing/model': 1.2062758640386164, 'learning_rate/qwen_vl_interface': 9.775470683414038e-06, 'learning_rate/action_model': 9.775470683414038e-05, 'epoch': 0.23}) +[2026-06-17 07:33:42,835][starVLA.training.train_starvla][INFO] - Step 779, Loss: {'train/loss': 1.0349557399749756, 'train/grad_norm_pre_clip': 9.5381498336792, 'timing/data': 0.004601225955411792, 'timing/model': 1.215643442934379, 'learning_rate/qwen_vl_interface': 9.774813153978327e-06, 'learning_rate/action_model': 9.774813153978327e-05, 'epoch': 0.23}) +[2026-06-17 07:33:47,595][starVLA.training.train_starvla][INFO] - Step 780, Loss: {'train/loss': 0.8203921318054199, 'train/grad_norm_pre_clip': 12.785853385925293, 'timing/data': 0.004745402024127543, 'timing/model': 1.2101335739716887, 'learning_rate/qwen_vl_interface': 9.77415468654398e-06, 'learning_rate/action_model': 9.77415468654398e-05, 'epoch': 0.23}) +[2026-06-17 07:33:52,490][starVLA.training.train_starvla][INFO] - Step 781, Loss: {'train/loss': 0.8496288657188416, 'train/grad_norm_pre_clip': 22.39423179626465, 'timing/data': 0.0046289319870993495, 'timing/model': 1.353043518960476, 'learning_rate/qwen_vl_interface': 9.773495281247497e-06, 'learning_rate/action_model': 9.773495281247497e-05, 'epoch': 0.23}) +[2026-06-17 07:33:57,156][starVLA.training.train_starvla][INFO] - Step 782, Loss: {'train/loss': 0.8963217735290527, 'train/grad_norm_pre_clip': 8.193318367004395, 'timing/data': 0.004569369019009173, 'timing/model': 1.2059825700707734, 'learning_rate/qwen_vl_interface': 9.772834938225577e-06, 'learning_rate/action_model': 9.772834938225576e-05, 'epoch': 0.23}) +[2026-06-17 07:34:01,948][starVLA.training.train_starvla][INFO] - Step 783, Loss: {'train/loss': 1.0561696290969849, 'train/grad_norm_pre_clip': 10.51742172241211, 'timing/data': 0.004817552049644291, 'timing/model': 1.2078944010427222, 'learning_rate/qwen_vl_interface': 9.772173657615107e-06, 'learning_rate/action_model': 9.772173657615106e-05, 'epoch': 0.23}) +[2026-06-17 07:34:06,749][starVLA.training.train_starvla][INFO] - Step 784, Loss: {'train/loss': 1.0461703538894653, 'train/grad_norm_pre_clip': 11.318099021911621, 'timing/data': 0.004561107023619115, 'timing/model': 1.2155232740333304, 'learning_rate/qwen_vl_interface': 9.77151143955317e-06, 'learning_rate/action_model': 9.77151143955317e-05, 'epoch': 0.23}) +[2026-06-17 07:34:11,538][starVLA.training.train_starvla][INFO] - Step 785, Loss: {'train/loss': 0.9331804513931274, 'train/grad_norm_pre_clip': 10.145590782165527, 'timing/data': 0.0048738730838522315, 'timing/model': 1.211490556015633, 'learning_rate/qwen_vl_interface': 9.770848284177048e-06, 'learning_rate/action_model': 9.770848284177047e-05, 'epoch': 0.23}) +[2026-06-17 07:34:16,294][starVLA.training.train_starvla][INFO] - Step 786, Loss: {'train/loss': 0.7363969087600708, 'train/grad_norm_pre_clip': 137.11277770996094, 'timing/data': 0.0047668240731582046, 'timing/model': 1.21234327100683, 'learning_rate/qwen_vl_interface': 9.77018419162421e-06, 'learning_rate/action_model': 9.77018419162421e-05, 'epoch': 0.23}) +[2026-06-17 07:34:21,093][starVLA.training.train_starvla][INFO] - Step 787, Loss: {'train/loss': 0.8571935892105103, 'train/grad_norm_pre_clip': 8.219276428222656, 'timing/data': 0.004829765995964408, 'timing/model': 1.3231310719856992, 'learning_rate/qwen_vl_interface': 9.769519162032328e-06, 'learning_rate/action_model': 9.769519162032327e-05, 'epoch': 0.23}) +[2026-06-17 07:34:25,774][starVLA.training.train_starvla][INFO] - Step 788, Loss: {'train/loss': 0.8361333608627319, 'train/grad_norm_pre_clip': 6.576010704040527, 'timing/data': 0.004593395977281034, 'timing/model': 1.2126999089960009, 'learning_rate/qwen_vl_interface': 9.768853195539259e-06, 'learning_rate/action_model': 9.768853195539258e-05, 'epoch': 0.23}) +[2026-06-17 07:34:30,595][starVLA.training.train_starvla][INFO] - Step 789, Loss: {'train/loss': 0.8741225600242615, 'train/grad_norm_pre_clip': 8.512618064880371, 'timing/data': 0.004494966007769108, 'timing/model': 1.21567504608538, 'learning_rate/qwen_vl_interface': 9.76818629228306e-06, 'learning_rate/action_model': 9.768186292283059e-05, 'epoch': 0.23}) +[2026-06-17 07:34:35,393][starVLA.training.train_starvla][INFO] - Step 790, Loss: {'train/loss': 1.0352261066436768, 'train/grad_norm_pre_clip': 12.856513977050781, 'timing/data': 0.00458223803434521, 'timing/model': 1.2085846369154751, 'learning_rate/qwen_vl_interface': 9.767518452401981e-06, 'learning_rate/action_model': 9.76751845240198e-05, 'epoch': 0.23}) +[2026-06-17 07:34:40,209][starVLA.training.train_starvla][INFO] - Step 791, Loss: {'train/loss': 1.1110392808914185, 'train/grad_norm_pre_clip': 10.180336952209473, 'timing/data': 0.005015936912968755, 'timing/model': 1.2073706289520487, 'learning_rate/qwen_vl_interface': 9.766849676034465e-06, 'learning_rate/action_model': 9.766849676034464e-05, 'epoch': 0.23}) +[2026-06-17 07:34:45,034][starVLA.training.train_starvla][INFO] - Step 792, Loss: {'train/loss': 1.24488365650177, 'train/grad_norm_pre_clip': 14.125944137573242, 'timing/data': 0.00447478995192796, 'timing/model': 1.342248525004834, 'learning_rate/qwen_vl_interface': 9.766179963319152e-06, 'learning_rate/action_model': 9.76617996331915e-05, 'epoch': 0.23}) +[2026-06-17 07:34:49,697][starVLA.training.train_starvla][INFO] - Step 793, Loss: {'train/loss': 1.1687097549438477, 'train/grad_norm_pre_clip': 13.121160507202148, 'timing/data': 0.004384739906527102, 'timing/model': 1.2057933789910749, 'learning_rate/qwen_vl_interface': 9.765509314394872e-06, 'learning_rate/action_model': 9.765509314394872e-05, 'epoch': 0.23}) +[2026-06-17 07:34:54,486][starVLA.training.train_starvla][INFO] - Step 794, Loss: {'train/loss': 1.2867298126220703, 'train/grad_norm_pre_clip': 16.514751434326172, 'timing/data': 0.004731304012238979, 'timing/model': 1.2062661609379575, 'learning_rate/qwen_vl_interface': 9.764837729400652e-06, 'learning_rate/action_model': 9.764837729400653e-05, 'epoch': 0.23}) +[2026-06-17 07:34:59,196][starVLA.training.train_starvla][INFO] - Step 795, Loss: {'train/loss': 0.9911518692970276, 'train/grad_norm_pre_clip': 10.36092758178711, 'timing/data': 0.004389702924527228, 'timing/model': 1.2069142439868301, 'learning_rate/qwen_vl_interface': 9.764165208475714e-06, 'learning_rate/action_model': 9.764165208475714e-05, 'epoch': 0.23}) +[2026-06-17 07:35:03,956][starVLA.training.train_starvla][INFO] - Step 796, Loss: {'train/loss': 0.9082820415496826, 'train/grad_norm_pre_clip': 6.629059791564941, 'timing/data': 0.00474423507694155, 'timing/model': 1.2060986020369455, 'learning_rate/qwen_vl_interface': 9.76349175175947e-06, 'learning_rate/action_model': 9.763491751759469e-05, 'epoch': 0.23}) +[2026-06-17 07:35:08,753][starVLA.training.train_starvla][INFO] - Step 797, Loss: {'train/loss': 0.7693963646888733, 'train/grad_norm_pre_clip': 7.246853828430176, 'timing/data': 0.0043246649438515306, 'timing/model': 1.205184113001451, 'learning_rate/qwen_vl_interface': 9.76281735939153e-06, 'learning_rate/action_model': 9.762817359391529e-05, 'epoch': 0.23}) +[2026-06-17 07:35:13,549][starVLA.training.train_starvla][INFO] - Step 798, Loss: {'train/loss': 1.2226605415344238, 'train/grad_norm_pre_clip': 12.84561538696289, 'timing/data': 0.00444397795945406, 'timing/model': 1.3354784059338272, 'learning_rate/qwen_vl_interface': 9.762142031511695e-06, 'learning_rate/action_model': 9.762142031511695e-05, 'epoch': 0.23}) +[2026-06-17 07:35:18,228][starVLA.training.train_starvla][INFO] - Step 799, Loss: {'train/loss': 1.1938587427139282, 'train/grad_norm_pre_clip': 10.660517692565918, 'timing/data': 0.00513172906357795, 'timing/model': 1.2138119629817083, 'learning_rate/qwen_vl_interface': 9.76146576825996e-06, 'learning_rate/action_model': 9.761465768259959e-05, 'epoch': 0.24}) +[2026-06-17 07:37:00,063][starVLA.training.train_starvla][INFO] - Step 800, Loss: {'train/loss': 0.9877334237098694, 'train/grad_norm_pre_clip': 10.4191312789917, 'eval/loss': 1.0160605907440186, 'eval/latency_0/loss': 1.0160605907440186, 'eval/demon_fire_precision': 0.8975340136054422, 'eval/demon_fire_recall': 0.8766611295681063, 'eval/demon_fire_f1': 0.8869747899159665, 'eval/demon_fire_rate_model': 0.6540600667408232, 'eval/demon_fire_rate_teacher': 0.6696329254727474, 'eval/demon_left_precision': 0.803411131059246, 'eval/demon_left_recall': 0.5070821529745042, 'eval/demon_left_f1': 0.6217436609934005, 'eval/demon_left_rate_model': 0.3097886540600667, 'eval/demon_left_rate_teacher': 0.49082313681868744, 'eval/demon_right_precision': 0.6353746978243352, 'eval/demon_right_recall': 0.8909604519774011, 'eval/demon_right_f1': 0.7417685794920038, 'eval/demon_right_rate_model': 0.6902113459399333, 'eval/demon_right_rate_teacher': 0.4922135706340378, 'eval/demon_move_f1': 0.6817561202427022, 'eval/demon_cc_f1': 0.8135111275841421, 'eval/latency_0/demon_fire_precision': 0.8975340136054422, 'eval/latency_0/demon_fire_recall': 0.8766611295681063, 'eval/latency_0/demon_fire_f1': 0.8869747899159665, 'eval/latency_0/demon_fire_rate_model': 0.6540600667408232, 'eval/latency_0/demon_fire_rate_teacher': 0.6696329254727474, 'eval/latency_0/demon_left_precision': 0.803411131059246, 'eval/latency_0/demon_left_recall': 0.5070821529745042, 'eval/latency_0/demon_left_f1': 0.6217436609934005, 'eval/latency_0/demon_left_rate_model': 0.3097886540600667, 'eval/latency_0/demon_left_rate_teacher': 0.49082313681868744, 'eval/latency_0/demon_right_precision': 0.6353746978243352, 'eval/latency_0/demon_right_recall': 0.8909604519774011, 'eval/latency_0/demon_right_f1': 0.7417685794920038, 'eval/latency_0/demon_right_rate_model': 0.6902113459399333, 'eval/latency_0/demon_right_rate_teacher': 0.4922135706340378, 'eval/latency_0/demon_move_f1': 0.6817561202427022, 'eval/latency_0/demon_cc_f1': 0.8135111275841421, 'timing/data': 0.004610177013091743, 'timing/model': 1.2081884209765121, 'learning_rate/qwen_vl_interface': 9.760788569776518e-06, 'learning_rate/action_model': 9.760788569776517e-05, 'epoch': 0.24}) +[2026-06-17 07:37:00,064][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_800_state +[2026-06-17 07:37:00,066][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 07:37:03,634][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_800_state/model.safetensors +[2026-06-17 07:37:11,391][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_800_state/optimizer.bin +[2026-06-17 07:37:11,392][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_800_state/sampler.bin +[2026-06-17 07:37:11,392][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_800_state/sampler_1.bin +[2026-06-17 07:37:11,393][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_800_state/random_states_0.pkl +[2026-06-17 07:37:11,434][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 800 +[2026-06-17 07:37:11,434][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 07:37:11,472][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 07:37:16,103][starVLA.training.train_starvla][INFO] - Step 801, Loss: {'train/loss': 0.9701759219169617, 'train/grad_norm_pre_clip': 7.009166240692139, 'timing/data': 0.004994231043383479, 'timing/model': 1.1909381839213893, 'learning_rate/qwen_vl_interface': 9.760110436201752e-06, 'learning_rate/action_model': 9.76011043620175e-05, 'epoch': 0.24}) +[2026-06-17 07:37:20,837][starVLA.training.train_starvla][INFO] - Step 802, Loss: {'train/loss': 1.360884189605713, 'train/grad_norm_pre_clip': 12.39353084564209, 'timing/data': 0.005007639992982149, 'timing/model': 1.1928009310504422, 'learning_rate/qwen_vl_interface': 9.759431367676238e-06, 'learning_rate/action_model': 9.759431367676237e-05, 'epoch': 0.24}) +[2026-06-17 07:37:25,525][starVLA.training.train_starvla][INFO] - Step 803, Loss: {'train/loss': 1.0301679372787476, 'train/grad_norm_pre_clip': 7.1900811195373535, 'timing/data': 0.004950513946823776, 'timing/model': 1.19397670798935, 'learning_rate/qwen_vl_interface': 9.758751364340751e-06, 'learning_rate/action_model': 9.75875136434075e-05, 'epoch': 0.24}) +[2026-06-17 07:37:30,267][starVLA.training.train_starvla][INFO] - Step 804, Loss: {'train/loss': 0.9515025019645691, 'train/grad_norm_pre_clip': 10.03825855255127, 'timing/data': 0.005370176979340613, 'timing/model': 1.1933208329137415, 'learning_rate/qwen_vl_interface': 9.758070426336252e-06, 'learning_rate/action_model': 9.758070426336252e-05, 'epoch': 0.24}) +[2026-06-17 07:37:35,008][starVLA.training.train_starvla][INFO] - Step 805, Loss: {'train/loss': 1.3028056621551514, 'train/grad_norm_pre_clip': 8.219051361083984, 'timing/data': 0.004918302991427481, 'timing/model': 1.194304139004089, 'learning_rate/qwen_vl_interface': 9.757388553803905e-06, 'learning_rate/action_model': 9.757388553803905e-05, 'epoch': 0.24}) +[2026-06-17 07:37:39,752][starVLA.training.train_starvla][INFO] - Step 806, Loss: {'train/loss': 1.0988184213638306, 'train/grad_norm_pre_clip': 7.7133636474609375, 'timing/data': 0.004933136980980635, 'timing/model': 1.3209825049852952, 'learning_rate/qwen_vl_interface': 9.75670574688506e-06, 'learning_rate/action_model': 9.75670574688506e-05, 'epoch': 0.24}) +[2026-06-17 07:37:44,586][starVLA.training.train_starvla][INFO] - Step 807, Loss: {'train/loss': 1.1411316394805908, 'train/grad_norm_pre_clip': 7.469799518585205, 'timing/data': 0.004613445955328643, 'timing/model': 1.3226784550352022, 'learning_rate/qwen_vl_interface': 9.756022005721265e-06, 'learning_rate/action_model': 9.756022005721265e-05, 'epoch': 0.24}) +[2026-06-17 07:37:49,219][starVLA.training.train_starvla][INFO] - Step 808, Loss: {'train/loss': 1.0462291240692139, 'train/grad_norm_pre_clip': 11.026143074035645, 'timing/data': 0.004926541936583817, 'timing/model': 1.1990912999026477, 'learning_rate/qwen_vl_interface': 9.755337330454257e-06, 'learning_rate/action_model': 9.755337330454257e-05, 'epoch': 0.24}) +[2026-06-17 07:37:53,951][starVLA.training.train_starvla][INFO] - Step 809, Loss: {'train/loss': 1.1278738975524902, 'train/grad_norm_pre_clip': 14.102046966552734, 'timing/data': 0.005297004012390971, 'timing/model': 1.1999046999262646, 'learning_rate/qwen_vl_interface': 9.754651721225975e-06, 'learning_rate/action_model': 9.754651721225975e-05, 'epoch': 0.24}) +[2026-06-17 07:37:58,736][starVLA.training.train_starvla][INFO] - Step 810, Loss: {'train/loss': 0.9916684627532959, 'train/grad_norm_pre_clip': 6.647624492645264, 'timing/data': 0.0046907729702070355, 'timing/model': 1.200844544917345, 'learning_rate/qwen_vl_interface': 9.753965178178543e-06, 'learning_rate/action_model': 9.753965178178543e-05, 'epoch': 0.24}) +[2026-06-17 07:38:03,519][starVLA.training.train_starvla][INFO] - Step 811, Loss: {'train/loss': 1.028476595878601, 'train/grad_norm_pre_clip': 8.578556060791016, 'timing/data': 0.004943992011249065, 'timing/model': 1.2005589390173554, 'learning_rate/qwen_vl_interface': 9.753277701454285e-06, 'learning_rate/action_model': 9.753277701454284e-05, 'epoch': 0.24}) +[2026-06-17 07:38:08,286][starVLA.training.train_starvla][INFO] - Step 812, Loss: {'train/loss': 0.8685882091522217, 'train/grad_norm_pre_clip': 10.835689544677734, 'timing/data': 0.004775758017785847, 'timing/model': 1.3291542119113728, 'learning_rate/qwen_vl_interface': 9.752589291195711e-06, 'learning_rate/action_model': 9.75258929119571e-05, 'epoch': 0.24}) +[2026-06-17 07:38:13,097][starVLA.training.train_starvla][INFO] - Step 813, Loss: {'train/loss': 0.898649275302887, 'train/grad_norm_pre_clip': 8.896363258361816, 'timing/data': 0.005021924967877567, 'timing/model': 1.3266596199246123, 'learning_rate/qwen_vl_interface': 9.751899947545534e-06, 'learning_rate/action_model': 9.751899947545533e-05, 'epoch': 0.24}) +[2026-06-17 07:38:17,737][starVLA.training.train_starvla][INFO] - Step 814, Loss: {'train/loss': 0.9802244305610657, 'train/grad_norm_pre_clip': 6.976622104644775, 'timing/data': 0.004356874036602676, 'timing/model': 1.2014133649645373, 'learning_rate/qwen_vl_interface': 9.751209670646653e-06, 'learning_rate/action_model': 9.751209670646652e-05, 'epoch': 0.24}) +[2026-06-17 07:38:22,482][starVLA.training.train_starvla][INFO] - Step 815, Loss: {'train/loss': 1.037920594215393, 'train/grad_norm_pre_clip': 10.011919021606445, 'timing/data': 0.004963038023561239, 'timing/model': 1.2012899019755423, 'learning_rate/qwen_vl_interface': 9.750518460642162e-06, 'learning_rate/action_model': 9.75051846064216e-05, 'epoch': 0.24}) +[2026-06-17 07:38:27,271][starVLA.training.train_starvla][INFO] - Step 816, Loss: {'train/loss': 1.080459475517273, 'train/grad_norm_pre_clip': 7.2181525230407715, 'timing/data': 0.004586644005030394, 'timing/model': 1.2010668630246073, 'learning_rate/qwen_vl_interface': 9.749826317675352e-06, 'learning_rate/action_model': 9.749826317675352e-05, 'epoch': 0.24}) +[2026-06-17 07:38:32,047][starVLA.training.train_starvla][INFO] - Step 817, Loss: {'train/loss': 1.1120719909667969, 'train/grad_norm_pre_clip': 8.096360206604004, 'timing/data': 0.004851225996389985, 'timing/model': 1.2043421230046079, 'learning_rate/qwen_vl_interface': 9.749133241889706e-06, 'learning_rate/action_model': 9.749133241889705e-05, 'epoch': 0.24}) +[2026-06-17 07:38:36,780][starVLA.training.train_starvla][INFO] - Step 818, Loss: {'train/loss': 1.0268229246139526, 'train/grad_norm_pre_clip': 10.8541898727417, 'timing/data': 0.005636535934172571, 'timing/model': 1.2766823109705001, 'learning_rate/qwen_vl_interface': 9.748439233428897e-06, 'learning_rate/action_model': 9.748439233428896e-05, 'epoch': 0.24}) +[2026-06-17 07:38:41,440][starVLA.training.train_starvla][INFO] - Step 819, Loss: {'train/loss': 0.9090412855148315, 'train/grad_norm_pre_clip': 6.558040618896484, 'timing/data': 0.005048509920015931, 'timing/model': 1.2036589650670066, 'learning_rate/qwen_vl_interface': 9.747744292436794e-06, 'learning_rate/action_model': 9.747744292436793e-05, 'epoch': 0.24}) +[2026-06-17 07:38:46,242][starVLA.training.train_starvla][INFO] - Step 820, Loss: {'train/loss': 0.8585875630378723, 'train/grad_norm_pre_clip': 6.502280235290527, 'timing/data': 0.004367340006865561, 'timing/model': 1.209209561930038, 'learning_rate/qwen_vl_interface': 9.747048419057458e-06, 'learning_rate/action_model': 9.747048419057458e-05, 'epoch': 0.24}) +[2026-06-17 07:38:51,040][starVLA.training.train_starvla][INFO] - Step 821, Loss: {'train/loss': 1.167435646057129, 'train/grad_norm_pre_clip': 10.647126197814941, 'timing/data': 0.004898695042356849, 'timing/model': 1.2001286080339924, 'learning_rate/qwen_vl_interface': 9.746351613435148e-06, 'learning_rate/action_model': 9.746351613435147e-05, 'epoch': 0.24}) +[2026-06-17 07:38:55,816][starVLA.training.train_starvla][INFO] - Step 822, Loss: {'train/loss': 0.9210227727890015, 'train/grad_norm_pre_clip': 8.993193626403809, 'timing/data': 0.004500139970332384, 'timing/model': 1.2034353909548372, 'learning_rate/qwen_vl_interface': 9.74565387571431e-06, 'learning_rate/action_model': 9.745653875714309e-05, 'epoch': 0.24}) +[2026-06-17 07:39:00,594][starVLA.training.train_starvla][INFO] - Step 823, Loss: {'train/loss': 0.8546223044395447, 'train/grad_norm_pre_clip': 6.175327301025391, 'timing/data': 0.004939299076795578, 'timing/model': 1.3295565149746835, 'learning_rate/qwen_vl_interface': 9.744955206039585e-06, 'learning_rate/action_model': 9.744955206039585e-05, 'epoch': 0.24}) +[2026-06-17 07:39:05,363][starVLA.training.train_starvla][INFO] - Step 824, Loss: {'train/loss': 1.2244998216629028, 'train/grad_norm_pre_clip': 9.450063705444336, 'timing/data': 0.004993316950276494, 'timing/model': 1.2847446219529957, 'learning_rate/qwen_vl_interface': 9.74425560455581e-06, 'learning_rate/action_model': 9.74425560455581e-05, 'epoch': 0.24}) +[2026-06-17 07:39:10,025][starVLA.training.train_starvla][INFO] - Step 825, Loss: {'train/loss': 0.7918543219566345, 'train/grad_norm_pre_clip': 5.477985858917236, 'timing/data': 0.004940347978845239, 'timing/model': 1.2094174739904702, 'learning_rate/qwen_vl_interface': 9.743555071408011e-06, 'learning_rate/action_model': 9.743555071408011e-05, 'epoch': 0.24}) +[2026-06-17 07:39:14,819][starVLA.training.train_starvla][INFO] - Step 826, Loss: {'train/loss': 0.8113136887550354, 'train/grad_norm_pre_clip': 7.635879039764404, 'timing/data': 0.004949461901560426, 'timing/model': 1.2056176260812208, 'learning_rate/qwen_vl_interface': 9.742853606741412e-06, 'learning_rate/action_model': 9.742853606741411e-05, 'epoch': 0.24}) +[2026-06-17 07:39:19,599][starVLA.training.train_starvla][INFO] - Step 827, Loss: {'train/loss': 0.946358323097229, 'train/grad_norm_pre_clip': 5.962440490722656, 'timing/data': 0.004602069966495037, 'timing/model': 1.205936742015183, 'learning_rate/qwen_vl_interface': 9.742151210701423e-06, 'learning_rate/action_model': 9.742151210701424e-05, 'epoch': 0.24}) +[2026-06-17 07:39:24,388][starVLA.training.train_starvla][INFO] - Step 828, Loss: {'train/loss': 1.2537107467651367, 'train/grad_norm_pre_clip': 10.15704345703125, 'timing/data': 0.005159797961823642, 'timing/model': 1.2048394089797512, 'learning_rate/qwen_vl_interface': 9.741447883433656e-06, 'learning_rate/action_model': 9.741447883433656e-05, 'epoch': 0.24}) +[2026-06-17 07:39:29,209][starVLA.training.train_starvla][INFO] - Step 829, Loss: {'train/loss': 1.2221474647521973, 'train/grad_norm_pre_clip': 9.41797924041748, 'timing/data': 0.004869036958552897, 'timing/model': 1.204455264029093, 'learning_rate/qwen_vl_interface': 9.74074362508391e-06, 'learning_rate/action_model': 9.740743625083911e-05, 'epoch': 0.24}) +[2026-06-17 07:39:33,962][starVLA.training.train_starvla][INFO] - Step 830, Loss: {'train/loss': 0.9218409657478333, 'train/grad_norm_pre_clip': 8.898457527160645, 'timing/data': 0.004850854980759323, 'timing/model': 1.2931786690605804, 'learning_rate/qwen_vl_interface': 9.74003843579818e-06, 'learning_rate/action_model': 9.740038435798178e-05, 'epoch': 0.24}) +[2026-06-17 07:39:38,628][starVLA.training.train_starvla][INFO] - Step 831, Loss: {'train/loss': 0.973933219909668, 'train/grad_norm_pre_clip': 8.264130592346191, 'timing/data': 0.004892130964435637, 'timing/model': 1.2088669739896432, 'learning_rate/qwen_vl_interface': 9.739332315722647e-06, 'learning_rate/action_model': 9.739332315722646e-05, 'epoch': 0.24}) +[2026-06-17 07:39:43,414][starVLA.training.train_starvla][INFO] - Step 832, Loss: {'train/loss': 1.058884620666504, 'train/grad_norm_pre_clip': 7.183849811553955, 'timing/data': 0.00474127009510994, 'timing/model': 1.2067964150337502, 'learning_rate/qwen_vl_interface': 9.738625265003695e-06, 'learning_rate/action_model': 9.738625265003695e-05, 'epoch': 0.24}) +[2026-06-17 07:39:48,218][starVLA.training.train_starvla][INFO] - Step 833, Loss: {'train/loss': 1.107483148574829, 'train/grad_norm_pre_clip': 12.11763858795166, 'timing/data': 0.005030314088799059, 'timing/model': 1.2107312419684604, 'learning_rate/qwen_vl_interface': 9.7379172837879e-06, 'learning_rate/action_model': 9.737917283787898e-05, 'epoch': 0.25}) +[2026-06-17 07:39:53,014][starVLA.training.train_starvla][INFO] - Step 834, Loss: {'train/loss': 1.1554847955703735, 'train/grad_norm_pre_clip': 10.452390670776367, 'timing/data': 0.005324389087036252, 'timing/model': 1.2100420789793134, 'learning_rate/qwen_vl_interface': 9.737208372222018e-06, 'learning_rate/action_model': 9.737208372222017e-05, 'epoch': 0.25}) +[2026-06-17 07:39:57,831][starVLA.training.train_starvla][INFO] - Step 835, Loss: {'train/loss': 0.983226478099823, 'train/grad_norm_pre_clip': 8.418509483337402, 'timing/data': 0.004703060956671834, 'timing/model': 1.3548773410730064, 'learning_rate/qwen_vl_interface': 9.736498530453015e-06, 'learning_rate/action_model': 9.736498530453014e-05, 'epoch': 0.25}) +[2026-06-17 07:40:02,525][starVLA.training.train_starvla][INFO] - Step 836, Loss: {'train/loss': 0.872602641582489, 'train/grad_norm_pre_clip': 7.172839641571045, 'timing/data': 0.0047551930183544755, 'timing/model': 1.2104437680682167, 'learning_rate/qwen_vl_interface': 9.735787758628036e-06, 'learning_rate/action_model': 9.735787758628035e-05, 'epoch': 0.25}) +[2026-06-17 07:40:07,379][starVLA.training.train_starvla][INFO] - Step 837, Loss: {'train/loss': 1.1776235103607178, 'train/grad_norm_pre_clip': 11.67212200164795, 'timing/data': 0.005003275931812823, 'timing/model': 1.2041687400778756, 'learning_rate/qwen_vl_interface': 9.73507605689443e-06, 'learning_rate/action_model': 9.735076056894429e-05, 'epoch': 0.25}) +[2026-06-17 07:40:12,197][starVLA.training.train_starvla][INFO] - Step 838, Loss: {'train/loss': 0.674248993396759, 'train/grad_norm_pre_clip': 6.033360004425049, 'timing/data': 0.004566280986182392, 'timing/model': 1.2420878659468144, 'learning_rate/qwen_vl_interface': 9.73436342539973e-06, 'learning_rate/action_model': 9.734363425399728e-05, 'epoch': 0.25}) +[2026-06-17 07:40:16,925][starVLA.training.train_starvla][INFO] - Step 839, Loss: {'train/loss': 1.0287971496582031, 'train/grad_norm_pre_clip': 9.389942169189453, 'timing/data': 0.004745912970975041, 'timing/model': 1.2119830109877512, 'learning_rate/qwen_vl_interface': 9.733649864291667e-06, 'learning_rate/action_model': 9.733649864291666e-05, 'epoch': 0.25}) +[2026-06-17 07:40:21,680][starVLA.training.train_starvla][INFO] - Step 840, Loss: {'train/loss': 0.8448386788368225, 'train/grad_norm_pre_clip': 6.912036418914795, 'timing/data': 0.004875548998825252, 'timing/model': 1.2040020899148658, 'learning_rate/qwen_vl_interface': 9.732935373718162e-06, 'learning_rate/action_model': 9.732935373718161e-05, 'epoch': 0.25}) +[2026-06-17 07:40:26,481][starVLA.training.train_starvla][INFO] - Step 841, Loss: {'train/loss': 1.0840458869934082, 'train/grad_norm_pre_clip': 13.551342964172363, 'timing/data': 0.005064178025349975, 'timing/model': 1.3467332309810445, 'learning_rate/qwen_vl_interface': 9.73221995382733e-06, 'learning_rate/action_model': 9.73221995382733e-05, 'epoch': 0.25}) +[2026-06-17 07:40:31,180][starVLA.training.train_starvla][INFO] - Step 842, Loss: {'train/loss': 1.2518317699432373, 'train/grad_norm_pre_clip': 11.51904296875, 'timing/data': 0.005058867973275483, 'timing/model': 1.244568238966167, 'learning_rate/qwen_vl_interface': 9.73150360476748e-06, 'learning_rate/action_model': 9.73150360476748e-05, 'epoch': 0.25}) +[2026-06-17 07:40:35,986][starVLA.training.train_starvla][INFO] - Step 843, Loss: {'train/loss': 1.058000087738037, 'train/grad_norm_pre_clip': 10.07055950164795, 'timing/data': 0.0051289210096001625, 'timing/model': 1.2072250429773703, 'learning_rate/qwen_vl_interface': 9.73078632668711e-06, 'learning_rate/action_model': 9.73078632668711e-05, 'epoch': 0.25}) +[2026-06-17 07:40:40,769][starVLA.training.train_starvla][INFO] - Step 844, Loss: {'train/loss': 1.0671931505203247, 'train/grad_norm_pre_clip': 10.626260757446289, 'timing/data': 0.004788556951098144, 'timing/model': 1.2071486359927803, 'learning_rate/qwen_vl_interface': 9.730068119734912e-06, 'learning_rate/action_model': 9.730068119734912e-05, 'epoch': 0.25}) +[2026-06-17 07:40:45,567][starVLA.training.train_starvla][INFO] - Step 845, Loss: {'train/loss': 1.1678744554519653, 'train/grad_norm_pre_clip': 8.986555099487305, 'timing/data': 0.004854677943512797, 'timing/model': 1.2060139320092276, 'learning_rate/qwen_vl_interface': 9.729348984059773e-06, 'learning_rate/action_model': 9.729348984059774e-05, 'epoch': 0.25}) +[2026-06-17 07:40:50,356][starVLA.training.train_starvla][INFO] - Step 846, Loss: {'train/loss': 0.8543844223022461, 'train/grad_norm_pre_clip': 11.25279712677002, 'timing/data': 0.00516364595387131, 'timing/model': 1.204615069902502, 'learning_rate/qwen_vl_interface': 9.728628919810769e-06, 'learning_rate/action_model': 9.728628919810768e-05, 'epoch': 0.25}) +[2026-06-17 07:40:55,150][starVLA.training.train_starvla][INFO] - Step 847, Loss: {'train/loss': 0.7668900489807129, 'train/grad_norm_pre_clip': 6.274844646453857, 'timing/data': 0.005012255976907909, 'timing/model': 1.3335181259317324, 'learning_rate/qwen_vl_interface': 9.727907927137171e-06, 'learning_rate/action_model': 9.727907927137171e-05, 'epoch': 0.25}) +[2026-06-17 07:40:59,824][starVLA.training.train_starvla][INFO] - Step 848, Loss: {'train/loss': 1.041791558265686, 'train/grad_norm_pre_clip': 9.824231147766113, 'timing/data': 0.005037060938775539, 'timing/model': 1.2070851341122761, 'learning_rate/qwen_vl_interface': 9.72718600618844e-06, 'learning_rate/action_model': 9.727186006188441e-05, 'epoch': 0.25}) +[2026-06-17 07:41:04,659][starVLA.training.train_starvla][INFO] - Step 849, Loss: {'train/loss': 1.0749568939208984, 'train/grad_norm_pre_clip': 27.278249740600586, 'timing/data': 0.005394994979724288, 'timing/model': 1.214380782097578, 'learning_rate/qwen_vl_interface': 9.726463157114233e-06, 'learning_rate/action_model': 9.726463157114234e-05, 'epoch': 0.25}) +[2026-06-17 07:41:09,455][starVLA.training.train_starvla][INFO] - Step 850, Loss: {'train/loss': 0.70652174949646, 'train/grad_norm_pre_clip': 6.57431173324585, 'timing/data': 0.004884941037744284, 'timing/model': 1.2059811890358105, 'learning_rate/qwen_vl_interface': 9.725739380064399e-06, 'learning_rate/action_model': 9.725739380064398e-05, 'epoch': 0.25}) +[2026-06-17 07:41:14,242][starVLA.training.train_starvla][INFO] - Step 851, Loss: {'train/loss': 0.835125744342804, 'train/grad_norm_pre_clip': 7.333611965179443, 'timing/data': 0.005111483973450959, 'timing/model': 1.2063236430985853, 'learning_rate/qwen_vl_interface': 9.725014675188972e-06, 'learning_rate/action_model': 9.725014675188973e-05, 'epoch': 0.25}) +[2026-06-17 07:41:19,042][starVLA.training.train_starvla][INFO] - Step 852, Loss: {'train/loss': 0.8504632711410522, 'train/grad_norm_pre_clip': 8.665876388549805, 'timing/data': 0.004953630035743117, 'timing/model': 1.347267166012898, 'learning_rate/qwen_vl_interface': 9.72428904263819e-06, 'learning_rate/action_model': 9.72428904263819e-05, 'epoch': 0.25}) +[2026-06-17 07:41:23,703][starVLA.training.train_starvla][INFO] - Step 853, Loss: {'train/loss': 0.7883790135383606, 'train/grad_norm_pre_clip': 9.281975746154785, 'timing/data': 0.005032213986851275, 'timing/model': 1.2066610309993848, 'learning_rate/qwen_vl_interface': 9.723562482562474e-06, 'learning_rate/action_model': 9.723562482562473e-05, 'epoch': 0.25}) +[2026-06-17 07:41:28,490][starVLA.training.train_starvla][INFO] - Step 854, Loss: {'train/loss': 0.7631248831748962, 'train/grad_norm_pre_clip': 7.324528694152832, 'timing/data': 0.005036158952862024, 'timing/model': 1.2056727630551904, 'learning_rate/qwen_vl_interface': 9.72283499511244e-06, 'learning_rate/action_model': 9.722834995112439e-05, 'epoch': 0.25}) +[2026-06-17 07:41:33,213][starVLA.training.train_starvla][INFO] - Step 855, Loss: {'train/loss': 1.034974455833435, 'train/grad_norm_pre_clip': 7.497371196746826, 'timing/data': 0.0053962309611961246, 'timing/model': 1.2054003810044378, 'learning_rate/qwen_vl_interface': 9.722106580438901e-06, 'learning_rate/action_model': 9.7221065804389e-05, 'epoch': 0.25}) +[2026-06-17 07:41:37,970][starVLA.training.train_starvla][INFO] - Step 856, Loss: {'train/loss': 1.2535163164138794, 'train/grad_norm_pre_clip': 10.916348457336426, 'timing/data': 0.005130582954734564, 'timing/model': 1.2075206979643553, 'learning_rate/qwen_vl_interface': 9.721377238692855e-06, 'learning_rate/action_model': 9.721377238692854e-05, 'epoch': 0.25}) +[2026-06-17 07:41:42,764][starVLA.training.train_starvla][INFO] - Step 857, Loss: {'train/loss': 1.0332599878311157, 'train/grad_norm_pre_clip': 10.155630111694336, 'timing/data': 0.005126887932419777, 'timing/model': 1.2054578539682552, 'learning_rate/qwen_vl_interface': 9.720646970025494e-06, 'learning_rate/action_model': 9.720646970025494e-05, 'epoch': 0.25}) +[2026-06-17 07:41:47,547][starVLA.training.train_starvla][INFO] - Step 858, Loss: {'train/loss': 0.6895159482955933, 'train/grad_norm_pre_clip': 5.714079856872559, 'timing/data': 0.004797755973413587, 'timing/model': 1.3334923080401495, 'learning_rate/qwen_vl_interface': 9.719915774588209e-06, 'learning_rate/action_model': 9.719915774588209e-05, 'epoch': 0.25}) +[2026-06-17 07:41:52,217][starVLA.training.train_starvla][INFO] - Step 859, Loss: {'train/loss': 0.8729547262191772, 'train/grad_norm_pre_clip': 8.504430770874023, 'timing/data': 0.004767999984323978, 'timing/model': 1.205852692015469, 'learning_rate/qwen_vl_interface': 9.719183652532572e-06, 'learning_rate/action_model': 9.719183652532572e-05, 'epoch': 0.25}) +[2026-06-17 07:41:57,007][starVLA.training.train_starvla][INFO] - Step 860, Loss: {'train/loss': 0.9301385879516602, 'train/grad_norm_pre_clip': 7.341432094573975, 'timing/data': 0.0050231629284098744, 'timing/model': 1.2049904599552974, 'learning_rate/qwen_vl_interface': 9.718450604010355e-06, 'learning_rate/action_model': 9.718450604010355e-05, 'epoch': 0.25}) +[2026-06-17 07:42:01,712][starVLA.training.train_starvla][INFO] - Step 861, Loss: {'train/loss': 0.8693529963493347, 'train/grad_norm_pre_clip': 7.6752119064331055, 'timing/data': 0.004736154922284186, 'timing/model': 1.2053563579684123, 'learning_rate/qwen_vl_interface': 9.717716629173521e-06, 'learning_rate/action_model': 9.71771662917352e-05, 'epoch': 0.25}) +[2026-06-17 07:42:06,450][starVLA.training.train_starvla][INFO] - Step 862, Loss: {'train/loss': 0.8917874693870544, 'train/grad_norm_pre_clip': 6.269628524780273, 'timing/data': 0.005073075066320598, 'timing/model': 1.2073323829099536, 'learning_rate/qwen_vl_interface': 9.71698172817422e-06, 'learning_rate/action_model': 9.716981728174219e-05, 'epoch': 0.25}) +[2026-06-17 07:42:11,247][starVLA.training.train_starvla][INFO] - Step 863, Loss: {'train/loss': 0.6955184936523438, 'train/grad_norm_pre_clip': 7.569412708282471, 'timing/data': 0.004751775995828211, 'timing/model': 1.2112096040509641, 'learning_rate/qwen_vl_interface': 9.7162459011648e-06, 'learning_rate/action_model': 9.7162459011648e-05, 'epoch': 0.25}) +[2026-06-17 07:42:16,030][starVLA.training.train_starvla][INFO] - Step 864, Loss: {'train/loss': 0.9808002710342407, 'train/grad_norm_pre_clip': 13.40461540222168, 'timing/data': 0.004794472013600171, 'timing/model': 1.3323050279868767, 'learning_rate/qwen_vl_interface': 9.7155091482978e-06, 'learning_rate/action_model': 9.7155091482978e-05, 'epoch': 0.25}) +[2026-06-17 07:42:20,708][starVLA.training.train_starvla][INFO] - Step 865, Loss: {'train/loss': 1.0273278951644897, 'train/grad_norm_pre_clip': 8.514395713806152, 'timing/data': 0.005028352024964988, 'timing/model': 1.2043501329608262, 'learning_rate/qwen_vl_interface': 9.714771469725947e-06, 'learning_rate/action_model': 9.714771469725947e-05, 'epoch': 0.25}) +[2026-06-17 07:42:25,505][starVLA.training.train_starvla][INFO] - Step 866, Loss: {'train/loss': 0.8107415437698364, 'train/grad_norm_pre_clip': 8.593217849731445, 'timing/data': 0.005004924954846501, 'timing/model': 1.2073475889628753, 'learning_rate/qwen_vl_interface': 9.714032865602167e-06, 'learning_rate/action_model': 9.714032865602166e-05, 'epoch': 0.25}) +[2026-06-17 07:42:30,280][starVLA.training.train_starvla][INFO] - Step 867, Loss: {'train/loss': 1.0951995849609375, 'train/grad_norm_pre_clip': 9.855939865112305, 'timing/data': 0.0046187820844352245, 'timing/model': 1.205887092044577, 'learning_rate/qwen_vl_interface': 9.713293336079568e-06, 'learning_rate/action_model': 9.713293336079568e-05, 'epoch': 0.26}) +[2026-06-17 07:42:35,029][starVLA.training.train_starvla][INFO] - Step 868, Loss: {'train/loss': 0.7669082880020142, 'train/grad_norm_pre_clip': 10.100210189819336, 'timing/data': 0.004433152033016086, 'timing/model': 1.2052250250708312, 'learning_rate/qwen_vl_interface': 9.712552881311459e-06, 'learning_rate/action_model': 9.712552881311458e-05, 'epoch': 0.26}) +[2026-06-17 07:42:39,809][starVLA.training.train_starvla][INFO] - Step 869, Loss: {'train/loss': 1.0795308351516724, 'train/grad_norm_pre_clip': 10.614622116088867, 'timing/data': 0.004796126973815262, 'timing/model': 1.206330757937394, 'learning_rate/qwen_vl_interface': 9.711811501451336e-06, 'learning_rate/action_model': 9.711811501451335e-05, 'epoch': 0.26}) +[2026-06-17 07:42:44,611][starVLA.training.train_starvla][INFO] - Step 870, Loss: {'train/loss': 0.7172603607177734, 'train/grad_norm_pre_clip': 9.017394065856934, 'timing/data': 0.0049565129447728395, 'timing/model': 1.3400813539046794, 'learning_rate/qwen_vl_interface': 9.711069196652886e-06, 'learning_rate/action_model': 9.711069196652886e-05, 'epoch': 0.26}) +[2026-06-17 07:42:49,280][starVLA.training.train_starvla][INFO] - Step 871, Loss: {'train/loss': 1.1458109617233276, 'train/grad_norm_pre_clip': 11.693366050720215, 'timing/data': 0.007852121023461223, 'timing/model': 1.2128319359617308, 'learning_rate/qwen_vl_interface': 9.710325967069992e-06, 'learning_rate/action_model': 9.710325967069991e-05, 'epoch': 0.26}) +[2026-06-17 07:42:54,069][starVLA.training.train_starvla][INFO] - Step 872, Loss: {'train/loss': 1.2174382209777832, 'train/grad_norm_pre_clip': 10.076223373413086, 'timing/data': 0.005037583992816508, 'timing/model': 1.2060627419268712, 'learning_rate/qwen_vl_interface': 9.709581812856725e-06, 'learning_rate/action_model': 9.709581812856726e-05, 'epoch': 0.26}) +[2026-06-17 07:42:58,954][starVLA.training.train_starvla][INFO] - Step 873, Loss: {'train/loss': 0.9202675223350525, 'train/grad_norm_pre_clip': 10.023394584655762, 'timing/data': 0.004835717030800879, 'timing/model': 1.205534866079688, 'learning_rate/qwen_vl_interface': 9.708836734167351e-06, 'learning_rate/action_model': 9.708836734167351e-05, 'epoch': 0.26}) +[2026-06-17 07:43:03,736][starVLA.training.train_starvla][INFO] - Step 874, Loss: {'train/loss': 0.8001081347465515, 'train/grad_norm_pre_clip': 10.101001739501953, 'timing/data': 0.004665368003770709, 'timing/model': 1.2039811669383198, 'learning_rate/qwen_vl_interface': 9.708090731156321e-06, 'learning_rate/action_model': 9.708090731156321e-05, 'epoch': 0.26}) +[2026-06-17 07:43:08,411][starVLA.training.train_starvla][INFO] - Step 875, Loss: {'train/loss': 1.1033730506896973, 'train/grad_norm_pre_clip': 10.506006240844727, 'timing/data': 0.004830110934562981, 'timing/model': 1.2075182530097663, 'learning_rate/qwen_vl_interface': 9.707343803978288e-06, 'learning_rate/action_model': 9.707343803978288e-05, 'epoch': 0.26}) +[2026-06-17 07:43:13,107][starVLA.training.train_starvla][INFO] - Step 876, Loss: {'train/loss': 0.9608099460601807, 'train/grad_norm_pre_clip': 7.4224467277526855, 'timing/data': 0.0049304739804938436, 'timing/model': 1.2429505889303982, 'learning_rate/qwen_vl_interface': 9.706595952788088e-06, 'learning_rate/action_model': 9.706595952788088e-05, 'epoch': 0.26}) +[2026-06-17 07:43:17,766][starVLA.training.train_starvla][INFO] - Step 877, Loss: {'train/loss': 1.0698291063308716, 'train/grad_norm_pre_clip': 8.765910148620605, 'timing/data': 0.005251973983831704, 'timing/model': 1.2051055369665846, 'learning_rate/qwen_vl_interface': 9.705847177740751e-06, 'learning_rate/action_model': 9.705847177740751e-05, 'epoch': 0.26}) +[2026-06-17 07:43:22,509][starVLA.training.train_starvla][INFO] - Step 878, Loss: {'train/loss': 0.8532307147979736, 'train/grad_norm_pre_clip': 6.674243450164795, 'timing/data': 0.004741276963613927, 'timing/model': 1.2088148049078882, 'learning_rate/qwen_vl_interface': 9.7050974789915e-06, 'learning_rate/action_model': 9.7050974789915e-05, 'epoch': 0.26}) +[2026-06-17 07:43:27,297][starVLA.training.train_starvla][INFO] - Step 879, Loss: {'train/loss': 0.9156507849693298, 'train/grad_norm_pre_clip': 12.163783073425293, 'timing/data': 0.005118286935612559, 'timing/model': 1.2046109220245853, 'learning_rate/qwen_vl_interface': 9.704346856695747e-06, 'learning_rate/action_model': 9.704346856695746e-05, 'epoch': 0.26}) +[2026-06-17 07:43:32,089][starVLA.training.train_starvla][INFO] - Step 880, Loss: {'train/loss': 1.0609503984451294, 'train/grad_norm_pre_clip': 12.47839641571045, 'timing/data': 0.0051242869812995195, 'timing/model': 1.2093449499225244, 'learning_rate/qwen_vl_interface': 9.703595311009098e-06, 'learning_rate/action_model': 9.703595311009097e-05, 'epoch': 0.26}) +[2026-06-17 07:43:36,886][starVLA.training.train_starvla][INFO] - Step 881, Loss: {'train/loss': 1.0636842250823975, 'train/grad_norm_pre_clip': 8.812801361083984, 'timing/data': 0.005002267076633871, 'timing/model': 1.34182035503909, 'learning_rate/qwen_vl_interface': 9.702842842087348e-06, 'learning_rate/action_model': 9.702842842087348e-05, 'epoch': 0.26}) +[2026-06-17 07:43:41,664][starVLA.training.train_starvla][INFO] - Step 882, Loss: {'train/loss': 0.8825325965881348, 'train/grad_norm_pre_clip': 7.257394313812256, 'timing/data': 0.004953817930072546, 'timing/model': 1.3244417230598629, 'learning_rate/qwen_vl_interface': 9.702089450086486e-06, 'learning_rate/action_model': 9.702089450086486e-05, 'epoch': 0.26}) +[2026-06-17 07:43:46,339][starVLA.training.train_starvla][INFO] - Step 883, Loss: {'train/loss': 0.9359846115112305, 'train/grad_norm_pre_clip': 6.4498724937438965, 'timing/data': 0.004971272079274058, 'timing/model': 1.2071981980698183, 'learning_rate/qwen_vl_interface': 9.70133513516269e-06, 'learning_rate/action_model': 9.701335135162688e-05, 'epoch': 0.26}) +[2026-06-17 07:43:51,092][starVLA.training.train_starvla][INFO] - Step 884, Loss: {'train/loss': 1.3439936637878418, 'train/grad_norm_pre_clip': 10.931800842285156, 'timing/data': 0.0049743239069357514, 'timing/model': 1.2038871980039403, 'learning_rate/qwen_vl_interface': 9.70057989747233e-06, 'learning_rate/action_model': 9.700579897472329e-05, 'epoch': 0.26}) +[2026-06-17 07:43:55,879][starVLA.training.train_starvla][INFO] - Step 885, Loss: {'train/loss': 0.9519528150558472, 'train/grad_norm_pre_clip': 9.114585876464844, 'timing/data': 0.004941141000017524, 'timing/model': 1.2081268039764836, 'learning_rate/qwen_vl_interface': 9.699823737171969e-06, 'learning_rate/action_model': 9.699823737171969e-05, 'epoch': 0.26}) +[2026-06-17 07:44:00,679][starVLA.training.train_starvla][INFO] - Step 886, Loss: {'train/loss': 0.9981256127357483, 'train/grad_norm_pre_clip': 5.950284004211426, 'timing/data': 0.0051468960009515285, 'timing/model': 1.2101946669863537, 'learning_rate/qwen_vl_interface': 9.699066654418358e-06, 'learning_rate/action_model': 9.699066654418358e-05, 'epoch': 0.26}) +[2026-06-17 07:44:05,471][starVLA.training.train_starvla][INFO] - Step 887, Loss: {'train/loss': 0.9453280568122864, 'train/grad_norm_pre_clip': 7.091585636138916, 'timing/data': 0.005130838020704687, 'timing/model': 1.3307527060387656, 'learning_rate/qwen_vl_interface': 9.698308649368444e-06, 'learning_rate/action_model': 9.698308649368444e-05, 'epoch': 0.26}) +[2026-06-17 07:44:10,288][starVLA.training.train_starvla][INFO] - Step 888, Loss: {'train/loss': 0.9494481682777405, 'train/grad_norm_pre_clip': 6.596848011016846, 'timing/data': 0.0049543899949640036, 'timing/model': 1.3278902570018545, 'learning_rate/qwen_vl_interface': 9.69754972217936e-06, 'learning_rate/action_model': 9.69754972217936e-05, 'epoch': 0.26}) +[2026-06-17 07:44:14,957][starVLA.training.train_starvla][INFO] - Step 889, Loss: {'train/loss': 0.8918300867080688, 'train/grad_norm_pre_clip': 9.681488037109375, 'timing/data': 0.005091949016787112, 'timing/model': 1.2061899560503662, 'learning_rate/qwen_vl_interface': 9.696789873008432e-06, 'learning_rate/action_model': 9.696789873008433e-05, 'epoch': 0.26}) +[2026-06-17 07:44:19,734][starVLA.training.train_starvla][INFO] - Step 890, Loss: {'train/loss': 0.8047776222229004, 'train/grad_norm_pre_clip': 15.045892715454102, 'timing/data': 0.004898815997876227, 'timing/model': 1.2063274920219555, 'learning_rate/qwen_vl_interface': 9.69602910201318e-06, 'learning_rate/action_model': 9.696029102013179e-05, 'epoch': 0.26}) +[2026-06-17 07:44:24,534][starVLA.training.train_starvla][INFO] - Step 891, Loss: {'train/loss': 1.0690456628799438, 'train/grad_norm_pre_clip': 9.048664093017578, 'timing/data': 0.005025950958952308, 'timing/model': 1.2161741449963301, 'learning_rate/qwen_vl_interface': 9.695267409351308e-06, 'learning_rate/action_model': 9.695267409351307e-05, 'epoch': 0.26}) +[2026-06-17 07:44:29,336][starVLA.training.train_starvla][INFO] - Step 892, Loss: {'train/loss': 0.9392785429954529, 'train/grad_norm_pre_clip': 9.52857780456543, 'timing/data': 0.004468985018320382, 'timing/model': 1.2079493609489873, 'learning_rate/qwen_vl_interface': 9.694504795180722e-06, 'learning_rate/action_model': 9.694504795180721e-05, 'epoch': 0.26}) +[2026-06-17 07:44:34,115][starVLA.training.train_starvla][INFO] - Step 893, Loss: {'train/loss': 1.1368999481201172, 'train/grad_norm_pre_clip': 15.180068969726562, 'timing/data': 0.00477025902364403, 'timing/model': 1.2841615159995854, 'learning_rate/qwen_vl_interface': 9.693741259659507e-06, 'learning_rate/action_model': 9.693741259659507e-05, 'epoch': 0.26}) +[2026-06-17 07:44:38,784][starVLA.training.train_starvla][INFO] - Step 894, Loss: {'train/loss': 1.0572201013565063, 'train/grad_norm_pre_clip': 13.484810829162598, 'timing/data': 0.004907689057290554, 'timing/model': 1.2081856839358807, 'learning_rate/qwen_vl_interface': 9.692976802945947e-06, 'learning_rate/action_model': 9.692976802945948e-05, 'epoch': 0.26}) +[2026-06-17 07:44:43,585][starVLA.training.train_starvla][INFO] - Step 895, Loss: {'train/loss': 1.0049169063568115, 'train/grad_norm_pre_clip': 9.588767051696777, 'timing/data': 0.004912785021588206, 'timing/model': 1.2197268760064617, 'learning_rate/qwen_vl_interface': 9.692211425198517e-06, 'learning_rate/action_model': 9.692211425198517e-05, 'epoch': 0.26}) +[2026-06-17 07:44:48,391][starVLA.training.train_starvla][INFO] - Step 896, Loss: {'train/loss': 0.9887135624885559, 'train/grad_norm_pre_clip': 7.915616035461426, 'timing/data': 0.004885388072580099, 'timing/model': 1.204266811022535, 'learning_rate/qwen_vl_interface': 9.691445126575879e-06, 'learning_rate/action_model': 9.691445126575877e-05, 'epoch': 0.26}) +[2026-06-17 07:44:53,193][starVLA.training.train_starvla][INFO] - Step 897, Loss: {'train/loss': 0.8349800705909729, 'train/grad_norm_pre_clip': 9.888270378112793, 'timing/data': 0.004910053918138146, 'timing/model': 1.2079087300226092, 'learning_rate/qwen_vl_interface': 9.690677907236884e-06, 'learning_rate/action_model': 9.690677907236884e-05, 'epoch': 0.26}) +[2026-06-17 07:44:57,978][starVLA.training.train_starvla][INFO] - Step 898, Loss: {'train/loss': 0.8946576118469238, 'train/grad_norm_pre_clip': 6.3983354568481445, 'timing/data': 0.004792187944985926, 'timing/model': 1.3316363641060889, 'learning_rate/qwen_vl_interface': 9.689909767340582e-06, 'learning_rate/action_model': 9.689909767340583e-05, 'epoch': 0.26}) +[2026-06-17 07:45:02,745][starVLA.training.train_starvla][INFO] - Step 899, Loss: {'train/loss': 1.1409995555877686, 'train/grad_norm_pre_clip': 10.324097633361816, 'timing/data': 0.004759325995109975, 'timing/model': 1.3051868370966986, 'learning_rate/qwen_vl_interface': 9.689140707046209e-06, 'learning_rate/action_model': 9.689140707046209e-05, 'epoch': 0.26}) +[2026-06-17 07:46:45,486][starVLA.training.train_starvla][INFO] - Step 900, Loss: {'train/loss': 0.8582795858383179, 'train/grad_norm_pre_clip': 11.84123706817627, 'eval/loss': 0.9745705127716064, 'eval/latency_0/loss': 0.9745705127716064, 'eval/demon_fire_precision': 0.8717235901509134, 'eval/demon_fire_recall': 0.9115448504983389, 'eval/demon_fire_f1': 0.8911896061713358, 'eval/demon_fire_rate_model': 0.7002224694104561, 'eval/demon_fire_rate_teacher': 0.6696329254727474, 'eval/demon_left_precision': 0.8102246320681642, 'eval/demon_left_recall': 0.5926345609065156, 'eval/demon_left_f1': 0.6845549738219896, 'eval/demon_left_rate_model': 0.35901001112347053, 'eval/demon_left_rate_teacher': 0.49082313681868744, 'eval/demon_right_precision': 0.6815618221258134, 'eval/demon_right_recall': 0.8875706214689265, 'eval/demon_right_f1': 0.7710429447852761, 'eval/demon_right_rate_model': 0.6409899888765295, 'eval/demon_right_rate_teacher': 0.4922135706340378, 'eval/demon_move_f1': 0.7277989593036329, 'eval/demon_cc_f1': 0.8349554332836279, 'eval/latency_0/demon_fire_precision': 0.8717235901509134, 'eval/latency_0/demon_fire_recall': 0.9115448504983389, 'eval/latency_0/demon_fire_f1': 0.8911896061713358, 'eval/latency_0/demon_fire_rate_model': 0.7002224694104561, 'eval/latency_0/demon_fire_rate_teacher': 0.6696329254727474, 'eval/latency_0/demon_left_precision': 0.8102246320681642, 'eval/latency_0/demon_left_recall': 0.5926345609065156, 'eval/latency_0/demon_left_f1': 0.6845549738219896, 'eval/latency_0/demon_left_rate_model': 0.35901001112347053, 'eval/latency_0/demon_left_rate_teacher': 0.49082313681868744, 'eval/latency_0/demon_right_precision': 0.6815618221258134, 'eval/latency_0/demon_right_recall': 0.8875706214689265, 'eval/latency_0/demon_right_f1': 0.7710429447852761, 'eval/latency_0/demon_right_rate_model': 0.6409899888765295, 'eval/latency_0/demon_right_rate_teacher': 0.4922135706340378, 'eval/latency_0/demon_move_f1': 0.7277989593036329, 'eval/latency_0/demon_cc_f1': 0.8349554332836279, 'timing/data': 0.00515386404003948, 'timing/model': 1.2059872299432755, 'learning_rate/qwen_vl_interface': 9.68837072651319e-06, 'learning_rate/action_model': 9.688370726513188e-05, 'epoch': 0.26}) +[2026-06-17 07:46:45,487][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_900_state +[2026-06-17 07:46:45,490][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 07:46:49,040][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_900_state/model.safetensors +[2026-06-17 07:46:56,756][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_900_state/optimizer.bin +[2026-06-17 07:46:56,757][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_900_state/sampler.bin +[2026-06-17 07:46:56,757][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_900_state/sampler_1.bin +[2026-06-17 07:46:56,758][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_900_state/random_states_0.pkl +[2026-06-17 07:46:56,800][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 900 +[2026-06-17 07:46:56,800][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 07:46:56,812][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 07:47:01,418][starVLA.training.train_starvla][INFO] - Step 901, Loss: {'train/loss': 0.7519100308418274, 'train/grad_norm_pre_clip': 9.127577781677246, 'timing/data': 0.004859709995798767, 'timing/model': 1.2019106190418825, 'learning_rate/qwen_vl_interface': 9.687599825901141e-06, 'learning_rate/action_model': 9.687599825901142e-05, 'epoch': 0.27}) +[2026-06-17 07:47:06,021][starVLA.training.train_starvla][INFO] - Step 902, Loss: {'train/loss': 1.2516535520553589, 'train/grad_norm_pre_clip': 10.37881851196289, 'timing/data': 0.004689283086918294, 'timing/model': 1.1889601669972762, 'learning_rate/qwen_vl_interface': 9.686828005369876e-06, 'learning_rate/action_model': 9.686828005369875e-05, 'epoch': 0.27}) +[2026-06-17 07:47:10,640][starVLA.training.train_starvla][INFO] - Step 903, Loss: {'train/loss': 0.8548821806907654, 'train/grad_norm_pre_clip': 10.846076011657715, 'timing/data': 0.004907730035483837, 'timing/model': 1.1924337700475007, 'learning_rate/qwen_vl_interface': 9.68605526507939e-06, 'learning_rate/action_model': 9.686055265079389e-05, 'epoch': 0.27}) +[2026-06-17 07:47:15,329][starVLA.training.train_starvla][INFO] - Step 904, Loss: {'train/loss': 1.1745426654815674, 'train/grad_norm_pre_clip': 12.314491271972656, 'timing/data': 0.00463590701110661, 'timing/model': 1.1927198480116203, 'learning_rate/qwen_vl_interface': 9.685281605189874e-06, 'learning_rate/action_model': 9.685281605189874e-05, 'epoch': 0.27}) +[2026-06-17 07:47:20,069][starVLA.training.train_starvla][INFO] - Step 905, Loss: {'train/loss': 0.7967792749404907, 'train/grad_norm_pre_clip': 7.893943786621094, 'timing/data': 0.004668441019020975, 'timing/model': 1.198903830954805, 'learning_rate/qwen_vl_interface': 9.68450702586171e-06, 'learning_rate/action_model': 9.684507025861709e-05, 'epoch': 0.27}) +[2026-06-17 07:47:24,801][starVLA.training.train_starvla][INFO] - Step 906, Loss: {'train/loss': 0.7885813117027283, 'train/grad_norm_pre_clip': 10.746667861938477, 'timing/data': 0.004453467088751495, 'timing/model': 1.298344439943321, 'learning_rate/qwen_vl_interface': 9.683731527255467e-06, 'learning_rate/action_model': 9.683731527255467e-05, 'epoch': 0.27}) +[2026-06-17 07:47:29,427][starVLA.training.train_starvla][INFO] - Step 907, Loss: {'train/loss': 1.1434093713760376, 'train/grad_norm_pre_clip': 12.501367568969727, 'timing/data': 0.004609128111042082, 'timing/model': 1.198489562026225, 'learning_rate/qwen_vl_interface': 9.682955109531906e-06, 'learning_rate/action_model': 9.682955109531905e-05, 'epoch': 0.27}) +[2026-06-17 07:47:34,193][starVLA.training.train_starvla][INFO] - Step 908, Loss: {'train/loss': 1.0581886768341064, 'train/grad_norm_pre_clip': 10.404486656188965, 'timing/data': 0.004497361951507628, 'timing/model': 1.2007011369569227, 'learning_rate/qwen_vl_interface': 9.682177772851983e-06, 'learning_rate/action_model': 9.682177772851982e-05, 'epoch': 0.27}) +[2026-06-17 07:47:38,966][starVLA.training.train_starvla][INFO] - Step 909, Loss: {'train/loss': 0.8759210109710693, 'train/grad_norm_pre_clip': 7.858521938323975, 'timing/data': 0.004947321955114603, 'timing/model': 1.1989929230185226, 'learning_rate/qwen_vl_interface': 9.681399517376837e-06, 'learning_rate/action_model': 9.681399517376837e-05, 'epoch': 0.27}) +[2026-06-17 07:47:43,739][starVLA.training.train_starvla][INFO] - Step 910, Loss: {'train/loss': 1.1929931640625, 'train/grad_norm_pre_clip': 11.416265487670898, 'timing/data': 0.00489442003890872, 'timing/model': 1.2031934059923515, 'learning_rate/qwen_vl_interface': 9.680620343267803e-06, 'learning_rate/action_model': 9.680620343267802e-05, 'epoch': 0.27}) +[2026-06-17 07:47:48,459][starVLA.training.train_starvla][INFO] - Step 911, Loss: {'train/loss': 1.217394471168518, 'train/grad_norm_pre_clip': 9.337447166442871, 'timing/data': 0.004971889895386994, 'timing/model': 1.2018561830045655, 'learning_rate/qwen_vl_interface': 9.679840250686403e-06, 'learning_rate/action_model': 9.679840250686404e-05, 'epoch': 0.27}) +[2026-06-17 07:47:53,247][starVLA.training.train_starvla][INFO] - Step 912, Loss: {'train/loss': 0.9505494832992554, 'train/grad_norm_pre_clip': 9.931774139404297, 'timing/data': 0.004939523991197348, 'timing/model': 1.3433395409956574, 'learning_rate/qwen_vl_interface': 9.679059239794355e-06, 'learning_rate/action_model': 9.679059239794355e-05, 'epoch': 0.27}) +[2026-06-17 07:47:57,899][starVLA.training.train_starvla][INFO] - Step 913, Loss: {'train/loss': 0.8338261842727661, 'train/grad_norm_pre_clip': 9.151403427124023, 'timing/data': 0.00475931097753346, 'timing/model': 1.2108278940431774, 'learning_rate/qwen_vl_interface': 9.67827731075356e-06, 'learning_rate/action_model': 9.67827731075356e-05, 'epoch': 0.27}) +[2026-06-17 07:48:02,664][starVLA.training.train_starvla][INFO] - Step 914, Loss: {'train/loss': 0.7710613012313843, 'train/grad_norm_pre_clip': 10.68864631652832, 'timing/data': 0.004942892934195697, 'timing/model': 1.2000378479715437, 'learning_rate/qwen_vl_interface': 9.677494463726112e-06, 'learning_rate/action_model': 9.677494463726112e-05, 'epoch': 0.27}) +[2026-06-17 07:48:07,445][starVLA.training.train_starvla][INFO] - Step 915, Loss: {'train/loss': 1.012969732284546, 'train/grad_norm_pre_clip': 9.695956230163574, 'timing/data': 0.0049656040500849485, 'timing/model': 1.2026017479365692, 'learning_rate/qwen_vl_interface': 9.676710698874299e-06, 'learning_rate/action_model': 9.676710698874298e-05, 'epoch': 0.27}) +[2026-06-17 07:48:12,215][starVLA.training.train_starvla][INFO] - Step 916, Loss: {'train/loss': 1.2456024885177612, 'train/grad_norm_pre_clip': 8.073407173156738, 'timing/data': 0.004989763023331761, 'timing/model': 1.2051935429917648, 'learning_rate/qwen_vl_interface': 9.675926016360595e-06, 'learning_rate/action_model': 9.675926016360595e-05, 'epoch': 0.27}) +[2026-06-17 07:48:16,962][starVLA.training.train_starvla][INFO] - Step 917, Loss: {'train/loss': 1.0019742250442505, 'train/grad_norm_pre_clip': 10.08398723602295, 'timing/data': 0.004825151991099119, 'timing/model': 1.2020188540918753, 'learning_rate/qwen_vl_interface': 9.675140416347664e-06, 'learning_rate/action_model': 9.675140416347663e-05, 'epoch': 0.27}) +[2026-06-17 07:48:21,727][starVLA.training.train_starvla][INFO] - Step 918, Loss: {'train/loss': 0.845309853553772, 'train/grad_norm_pre_clip': 6.830197334289551, 'timing/data': 0.004948433022946119, 'timing/model': 1.3202245619613677, 'learning_rate/qwen_vl_interface': 9.674353898998362e-06, 'learning_rate/action_model': 9.674353898998363e-05, 'epoch': 0.27}) +[2026-06-17 07:48:26,384][starVLA.training.train_starvla][INFO] - Step 919, Loss: {'train/loss': 0.8665958046913147, 'train/grad_norm_pre_clip': 14.858590126037598, 'timing/data': 0.004874553997069597, 'timing/model': 1.2037960489979014, 'learning_rate/qwen_vl_interface': 9.673566464475738e-06, 'learning_rate/action_model': 9.673566464475739e-05, 'epoch': 0.27}) +[2026-06-17 07:48:31,183][starVLA.training.train_starvla][INFO] - Step 920, Loss: {'train/loss': 1.01383376121521, 'train/grad_norm_pre_clip': 10.30693244934082, 'timing/data': 0.004619410960003734, 'timing/model': 1.2086126829963177, 'learning_rate/qwen_vl_interface': 9.672778112943026e-06, 'learning_rate/action_model': 9.672778112943026e-05, 'epoch': 0.27}) +[2026-06-17 07:48:35,974][starVLA.training.train_starvla][INFO] - Step 921, Loss: {'train/loss': 1.1082204580307007, 'train/grad_norm_pre_clip': 8.691512107849121, 'timing/data': 0.005000412929803133, 'timing/model': 1.2025798039976507, 'learning_rate/qwen_vl_interface': 9.671988844563652e-06, 'learning_rate/action_model': 9.671988844563652e-05, 'epoch': 0.27}) +[2026-06-17 07:48:40,764][starVLA.training.train_starvla][INFO] - Step 922, Loss: {'train/loss': 1.0826361179351807, 'train/grad_norm_pre_clip': 9.527748107910156, 'timing/data': 0.004952277056872845, 'timing/model': 1.2026449459372088, 'learning_rate/qwen_vl_interface': 9.671198659501233e-06, 'learning_rate/action_model': 9.671198659501231e-05, 'epoch': 0.27}) +[2026-06-17 07:48:45,554][starVLA.training.train_starvla][INFO] - Step 923, Loss: {'train/loss': 1.1581380367279053, 'train/grad_norm_pre_clip': 11.8081693649292, 'timing/data': 0.004680681973695755, 'timing/model': 1.3318308889865875, 'learning_rate/qwen_vl_interface': 9.670407557919573e-06, 'learning_rate/action_model': 9.670407557919572e-05, 'epoch': 0.27}) +[2026-06-17 07:48:50,214][starVLA.training.train_starvla][INFO] - Step 924, Loss: {'train/loss': 0.8628352880477905, 'train/grad_norm_pre_clip': 7.287886619567871, 'timing/data': 0.004951372975483537, 'timing/model': 1.204718486056663, 'learning_rate/qwen_vl_interface': 9.669615539982668e-06, 'learning_rate/action_model': 9.669615539982668e-05, 'epoch': 0.27}) +[2026-06-17 07:48:54,999][starVLA.training.train_starvla][INFO] - Step 925, Loss: {'train/loss': 1.3597891330718994, 'train/grad_norm_pre_clip': 10.711333274841309, 'timing/data': 0.005134848994202912, 'timing/model': 1.202405244926922, 'learning_rate/qwen_vl_interface': 9.668822605854708e-06, 'learning_rate/action_model': 9.668822605854706e-05, 'epoch': 0.27}) +[2026-06-17 07:48:59,712][starVLA.training.train_starvla][INFO] - Step 926, Loss: {'train/loss': 0.9173598289489746, 'train/grad_norm_pre_clip': 7.278032302856445, 'timing/data': 0.004640814964659512, 'timing/model': 1.2117667599814013, 'learning_rate/qwen_vl_interface': 9.668028755700067e-06, 'learning_rate/action_model': 9.668028755700067e-05, 'epoch': 0.27}) +[2026-06-17 07:49:04,449][starVLA.training.train_starvla][INFO] - Step 927, Loss: {'train/loss': 0.9659620523452759, 'train/grad_norm_pre_clip': 8.660672187805176, 'timing/data': 0.004675316973589361, 'timing/model': 1.2055514520034194, 'learning_rate/qwen_vl_interface': 9.66723398968331e-06, 'learning_rate/action_model': 9.66723398968331e-05, 'epoch': 0.27}) +[2026-06-17 07:49:09,232][starVLA.training.train_starvla][INFO] - Step 928, Loss: {'train/loss': 1.186919093132019, 'train/grad_norm_pre_clip': 8.5702486038208, 'timing/data': 0.004798860056325793, 'timing/model': 1.2050992279546335, 'learning_rate/qwen_vl_interface': 9.666438307969194e-06, 'learning_rate/action_model': 9.666438307969195e-05, 'epoch': 0.27}) +[2026-06-17 07:49:14,016][starVLA.training.train_starvla][INFO] - Step 929, Loss: {'train/loss': 0.943979799747467, 'train/grad_norm_pre_clip': 7.606630802154541, 'timing/data': 0.004999979049898684, 'timing/model': 1.3341437180060893, 'learning_rate/qwen_vl_interface': 9.665641710722666e-06, 'learning_rate/action_model': 9.665641710722665e-05, 'epoch': 0.27}) +[2026-06-17 07:49:18,689][starVLA.training.train_starvla][INFO] - Step 930, Loss: {'train/loss': 0.7892399430274963, 'train/grad_norm_pre_clip': 6.88925838470459, 'timing/data': 0.004494098015129566, 'timing/model': 1.206376873073168, 'learning_rate/qwen_vl_interface': 9.664844198108858e-06, 'learning_rate/action_model': 9.664844198108858e-05, 'epoch': 0.27}) +[2026-06-17 07:49:23,502][starVLA.training.train_starvla][INFO] - Step 931, Loss: {'train/loss': 0.8498684763908386, 'train/grad_norm_pre_clip': 8.335271835327148, 'timing/data': 0.004677477991208434, 'timing/model': 1.213438122998923, 'learning_rate/qwen_vl_interface': 9.6640457702931e-06, 'learning_rate/action_model': 9.664045770293099e-05, 'epoch': 0.27}) +[2026-06-17 07:49:28,306][starVLA.training.train_starvla][INFO] - Step 932, Loss: {'train/loss': 0.8627428412437439, 'train/grad_norm_pre_clip': 6.405423641204834, 'timing/data': 0.004939508973620832, 'timing/model': 1.2100828799884766, 'learning_rate/qwen_vl_interface': 9.663246427440901e-06, 'learning_rate/action_model': 9.6632464274409e-05, 'epoch': 0.27}) +[2026-06-17 07:49:33,099][starVLA.training.train_starvla][INFO] - Step 933, Loss: {'train/loss': 0.8294199705123901, 'train/grad_norm_pre_clip': 9.10384750366211, 'timing/data': 0.005090342951007187, 'timing/model': 1.2133927550166845, 'learning_rate/qwen_vl_interface': 9.662446169717969e-06, 'learning_rate/action_model': 9.662446169717969e-05, 'epoch': 0.27}) +[2026-06-17 07:49:37,913][starVLA.training.train_starvla][INFO] - Step 934, Loss: {'train/loss': 1.17323899269104, 'train/grad_norm_pre_clip': 10.983784675598145, 'timing/data': 0.004774762084707618, 'timing/model': 1.2206719469977543, 'learning_rate/qwen_vl_interface': 9.6616449972902e-06, 'learning_rate/action_model': 9.6616449972902e-05, 'epoch': 0.27}) +[2026-06-17 07:49:42,734][starVLA.training.train_starvla][INFO] - Step 935, Loss: {'train/loss': 0.9706887006759644, 'train/grad_norm_pre_clip': 9.489508628845215, 'timing/data': 0.004837944055907428, 'timing/model': 1.3418066879967228, 'learning_rate/qwen_vl_interface': 9.660842910323675e-06, 'learning_rate/action_model': 9.660842910323674e-05, 'epoch': 0.28}) +[2026-06-17 07:49:47,423][starVLA.training.train_starvla][INFO] - Step 936, Loss: {'train/loss': 1.076306939125061, 'train/grad_norm_pre_clip': 12.299203872680664, 'timing/data': 0.00478307600133121, 'timing/model': 1.2120291590690613, 'learning_rate/qwen_vl_interface': 9.660039908984667e-06, 'learning_rate/action_model': 9.660039908984666e-05, 'epoch': 0.28}) +[2026-06-17 07:49:52,233][starVLA.training.train_starvla][INFO] - Step 937, Loss: {'train/loss': 0.9439483880996704, 'train/grad_norm_pre_clip': 7.852482318878174, 'timing/data': 0.0047111299354583025, 'timing/model': 1.2123147300444543, 'learning_rate/qwen_vl_interface': 9.659235993439643e-06, 'learning_rate/action_model': 9.659235993439643e-05, 'epoch': 0.28}) +[2026-06-17 07:49:57,064][starVLA.training.train_starvla][INFO] - Step 938, Loss: {'train/loss': 0.9013692140579224, 'train/grad_norm_pre_clip': 6.938438892364502, 'timing/data': 0.004318254999816418, 'timing/model': 1.2086131040705368, 'learning_rate/qwen_vl_interface': 9.658431163855251e-06, 'learning_rate/action_model': 9.658431163855251e-05, 'epoch': 0.28}) +[2026-06-17 07:50:01,902][starVLA.training.train_starvla][INFO] - Step 939, Loss: {'train/loss': 0.7225313782691956, 'train/grad_norm_pre_clip': 6.990149974822998, 'timing/data': 0.005168291972950101, 'timing/model': 1.2105536540038884, 'learning_rate/qwen_vl_interface': 9.657625420398334e-06, 'learning_rate/action_model': 9.657625420398333e-05, 'epoch': 0.28}) +[2026-06-17 07:50:06,640][starVLA.training.train_starvla][INFO] - Step 940, Loss: {'train/loss': 0.7581099271774292, 'train/grad_norm_pre_clip': 6.921260833740234, 'timing/data': 0.004742257064208388, 'timing/model': 1.20615587499924, 'learning_rate/qwen_vl_interface': 9.656818763235927e-06, 'learning_rate/action_model': 9.656818763235926e-05, 'epoch': 0.28}) +[2026-06-17 07:50:11,336][starVLA.training.train_starvla][INFO] - Step 941, Loss: {'train/loss': 0.7882016897201538, 'train/grad_norm_pre_clip': 7.172599792480469, 'timing/data': 0.004705331986770034, 'timing/model': 1.2459991569630802, 'learning_rate/qwen_vl_interface': 9.656011192535248e-06, 'learning_rate/action_model': 9.656011192535247e-05, 'epoch': 0.28}) +[2026-06-17 07:50:16,003][starVLA.training.train_starvla][INFO] - Step 942, Loss: {'train/loss': 0.7994620203971863, 'train/grad_norm_pre_clip': 10.433930397033691, 'timing/data': 0.004772714921273291, 'timing/model': 1.213840555981733, 'learning_rate/qwen_vl_interface': 9.655202708463706e-06, 'learning_rate/action_model': 9.655202708463707e-05, 'epoch': 0.28}) +[2026-06-17 07:50:20,756][starVLA.training.train_starvla][INFO] - Step 943, Loss: {'train/loss': 0.9946398138999939, 'train/grad_norm_pre_clip': 10.192429542541504, 'timing/data': 0.004844330018386245, 'timing/model': 1.2036493390332907, 'learning_rate/qwen_vl_interface': 9.654393311188905e-06, 'learning_rate/action_model': 9.654393311188905e-05, 'epoch': 0.28}) +[2026-06-17 07:50:25,550][starVLA.training.train_starvla][INFO] - Step 944, Loss: {'train/loss': 1.1254298686981201, 'train/grad_norm_pre_clip': 7.850892543792725, 'timing/data': 0.0047403560020029545, 'timing/model': 1.2054658500710502, 'learning_rate/qwen_vl_interface': 9.653583000878631e-06, 'learning_rate/action_model': 9.65358300087863e-05, 'epoch': 0.28}) +[2026-06-17 07:50:30,352][starVLA.training.train_starvla][INFO] - Step 945, Loss: {'train/loss': 1.1457793712615967, 'train/grad_norm_pre_clip': 16.226045608520508, 'timing/data': 0.004699331941083074, 'timing/model': 1.2076579510467127, 'learning_rate/qwen_vl_interface': 9.65277177770086e-06, 'learning_rate/action_model': 9.65277177770086e-05, 'epoch': 0.28}) +[2026-06-17 07:50:35,154][starVLA.training.train_starvla][INFO] - Step 946, Loss: {'train/loss': 1.101453423500061, 'train/grad_norm_pre_clip': 10.95927906036377, 'timing/data': 0.004920060047879815, 'timing/model': 1.3350301260361448, 'learning_rate/qwen_vl_interface': 9.651959641823764e-06, 'learning_rate/action_model': 9.651959641823763e-05, 'epoch': 0.28}) +[2026-06-17 07:50:39,937][starVLA.training.train_starvla][INFO] - Step 947, Loss: {'train/loss': 1.0738564729690552, 'train/grad_norm_pre_clip': 8.554916381835938, 'timing/data': 0.004911958007141948, 'timing/model': 1.3294265079312027, 'learning_rate/qwen_vl_interface': 9.651146593415696e-06, 'learning_rate/action_model': 9.651146593415695e-05, 'epoch': 0.28}) +[2026-06-17 07:50:44,613][starVLA.training.train_starvla][INFO] - Step 948, Loss: {'train/loss': 0.8619143962860107, 'train/grad_norm_pre_clip': 8.5215425491333, 'timing/data': 0.0047688649501651525, 'timing/model': 1.2055694089503959, 'learning_rate/qwen_vl_interface': 9.650332632645206e-06, 'learning_rate/action_model': 9.650332632645205e-05, 'epoch': 0.28}) +[2026-06-17 07:50:49,374][starVLA.training.train_starvla][INFO] - Step 949, Loss: {'train/loss': 0.8636782765388489, 'train/grad_norm_pre_clip': 7.906495571136475, 'timing/data': 0.004634637967683375, 'timing/model': 1.2091120450058952, 'learning_rate/qwen_vl_interface': 9.649517759681022e-06, 'learning_rate/action_model': 9.649517759681023e-05, 'epoch': 0.28}) +[2026-06-17 07:50:54,172][starVLA.training.train_starvla][INFO] - Step 950, Loss: {'train/loss': 0.9659254550933838, 'train/grad_norm_pre_clip': 8.712700843811035, 'timing/data': 0.004583170986734331, 'timing/model': 1.2084995310287923, 'learning_rate/qwen_vl_interface': 9.648701974692075e-06, 'learning_rate/action_model': 9.648701974692075e-05, 'epoch': 0.28}) +[2026-06-17 07:50:58,964][starVLA.training.train_starvla][INFO] - Step 951, Loss: {'train/loss': 1.046936273574829, 'train/grad_norm_pre_clip': 7.774526596069336, 'timing/data': 0.004689236055128276, 'timing/model': 1.2055755089968443, 'learning_rate/qwen_vl_interface': 9.647885277847476e-06, 'learning_rate/action_model': 9.647885277847476e-05, 'epoch': 0.28}) +[2026-06-17 07:51:03,754][starVLA.training.train_starvla][INFO] - Step 952, Loss: {'train/loss': 1.0073009729385376, 'train/grad_norm_pre_clip': 9.270345687866211, 'timing/data': 0.004747926956042647, 'timing/model': 1.3329374400200322, 'learning_rate/qwen_vl_interface': 9.647067669316525e-06, 'learning_rate/action_model': 9.647067669316525e-05, 'epoch': 0.28}) +[2026-06-17 07:51:08,532][starVLA.training.train_starvla][INFO] - Step 953, Loss: {'train/loss': 1.054619550704956, 'train/grad_norm_pre_clip': 10.676243782043457, 'timing/data': 0.004909777897410095, 'timing/model': 1.3235501409508288, 'learning_rate/qwen_vl_interface': 9.646249149268716e-06, 'learning_rate/action_model': 9.646249149268715e-05, 'epoch': 0.28}) +[2026-06-17 07:51:13,201][starVLA.training.train_starvla][INFO] - Step 954, Loss: {'train/loss': 1.198073387145996, 'train/grad_norm_pre_clip': 8.309778213500977, 'timing/data': 0.004979651072062552, 'timing/model': 1.210187511984259, 'learning_rate/qwen_vl_interface': 9.645429717873728e-06, 'learning_rate/action_model': 9.645429717873727e-05, 'epoch': 0.28}) +[2026-06-17 07:51:17,966][starVLA.training.train_starvla][INFO] - Step 955, Loss: {'train/loss': 1.0686100721359253, 'train/grad_norm_pre_clip': 8.17003345489502, 'timing/data': 0.005386764998547733, 'timing/model': 1.2056980159832165, 'learning_rate/qwen_vl_interface': 9.644609375301429e-06, 'learning_rate/action_model': 9.64460937530143e-05, 'epoch': 0.28}) +[2026-06-17 07:51:22,766][starVLA.training.train_starvla][INFO] - Step 956, Loss: {'train/loss': 1.0392818450927734, 'train/grad_norm_pre_clip': 10.983925819396973, 'timing/data': 0.005109959980472922, 'timing/model': 1.209115622099489, 'learning_rate/qwen_vl_interface': 9.643788121721877e-06, 'learning_rate/action_model': 9.643788121721876e-05, 'epoch': 0.28}) +[2026-06-17 07:51:27,568][starVLA.training.train_starvla][INFO] - Step 957, Loss: {'train/loss': 1.294847846031189, 'train/grad_norm_pre_clip': 10.258402824401855, 'timing/data': 0.00486767606344074, 'timing/model': 1.2049513849196956, 'learning_rate/qwen_vl_interface': 9.64296595730532e-06, 'learning_rate/action_model': 9.64296595730532e-05, 'epoch': 0.28}) +[2026-06-17 07:51:32,304][starVLA.training.train_starvla][INFO] - Step 958, Loss: {'train/loss': 0.7972836494445801, 'train/grad_norm_pre_clip': 7.669461727142334, 'timing/data': 0.004634286044165492, 'timing/model': 1.2779480559984222, 'learning_rate/qwen_vl_interface': 9.642142882222193e-06, 'learning_rate/action_model': 9.642142882222193e-05, 'epoch': 0.28}) +[2026-06-17 07:51:36,966][starVLA.training.train_starvla][INFO] - Step 959, Loss: {'train/loss': 0.8946126103401184, 'train/grad_norm_pre_clip': 9.937472343444824, 'timing/data': 0.004843604052439332, 'timing/model': 1.206733277067542, 'learning_rate/qwen_vl_interface': 9.641318896643123e-06, 'learning_rate/action_model': 9.641318896643122e-05, 'epoch': 0.28}) +[2026-06-17 07:51:41,770][starVLA.training.train_starvla][INFO] - Step 960, Loss: {'train/loss': 1.067124605178833, 'train/grad_norm_pre_clip': 7.626865863800049, 'timing/data': 0.00459567504003644, 'timing/model': 1.207283959025517, 'learning_rate/qwen_vl_interface': 9.640494000738917e-06, 'learning_rate/action_model': 9.640494000738917e-05, 'epoch': 0.28}) +[2026-06-17 07:51:46,565][starVLA.training.train_starvla][INFO] - Step 961, Loss: {'train/loss': 0.8441185355186462, 'train/grad_norm_pre_clip': 7.53316068649292, 'timing/data': 0.00478671898599714, 'timing/model': 1.2064518129918724, 'learning_rate/qwen_vl_interface': 9.63966819468058e-06, 'learning_rate/action_model': 9.63966819468058e-05, 'epoch': 0.28}) +[2026-06-17 07:51:51,356][starVLA.training.train_starvla][INFO] - Step 962, Loss: {'train/loss': 0.9752436280250549, 'train/grad_norm_pre_clip': 7.168154239654541, 'timing/data': 0.004706355044618249, 'timing/model': 1.204050803091377, 'learning_rate/qwen_vl_interface': 9.638841478639306e-06, 'learning_rate/action_model': 9.638841478639306e-05, 'epoch': 0.28}) +[2026-06-17 07:51:56,151][starVLA.training.train_starvla][INFO] - Step 963, Loss: {'train/loss': 1.0106264352798462, 'train/grad_norm_pre_clip': 6.7666144371032715, 'timing/data': 0.004839403904043138, 'timing/model': 1.343854292994365, 'learning_rate/qwen_vl_interface': 9.638013852786468e-06, 'learning_rate/action_model': 9.638013852786467e-05, 'epoch': 0.28}) +[2026-06-17 07:52:00,906][starVLA.training.train_starvla][INFO] - Step 964, Loss: {'train/loss': 0.7232262492179871, 'train/grad_norm_pre_clip': 6.753535747528076, 'timing/data': 0.0047360609751194715, 'timing/model': 1.2835952669847757, 'learning_rate/qwen_vl_interface': 9.637185317293638e-06, 'learning_rate/action_model': 9.637185317293638e-05, 'epoch': 0.28}) +[2026-06-17 07:52:05,569][starVLA.training.train_starvla][INFO] - Step 965, Loss: {'train/loss': 1.1140773296356201, 'train/grad_norm_pre_clip': 10.927139282226562, 'timing/data': 0.004848740063607693, 'timing/model': 1.205866583972238, 'learning_rate/qwen_vl_interface': 9.63635587233257e-06, 'learning_rate/action_model': 9.63635587233257e-05, 'epoch': 0.28}) +[2026-06-17 07:52:10,379][starVLA.training.train_starvla][INFO] - Step 966, Loss: {'train/loss': 1.0303136110305786, 'train/grad_norm_pre_clip': 8.871883392333984, 'timing/data': 0.004944375948980451, 'timing/model': 1.2105167130939662, 'learning_rate/qwen_vl_interface': 9.63552551807521e-06, 'learning_rate/action_model': 9.63552551807521e-05, 'epoch': 0.28}) +[2026-06-17 07:52:15,184][starVLA.training.train_starvla][INFO] - Step 967, Loss: {'train/loss': 0.7964726686477661, 'train/grad_norm_pre_clip': 6.053084850311279, 'timing/data': 0.004973856033757329, 'timing/model': 1.2116656400030479, 'learning_rate/qwen_vl_interface': 9.634694254693691e-06, 'learning_rate/action_model': 9.634694254693692e-05, 'epoch': 0.28}) +[2026-06-17 07:52:20,061][starVLA.training.train_starvla][INFO] - Step 968, Loss: {'train/loss': 0.8203756213188171, 'train/grad_norm_pre_clip': 6.651911735534668, 'timing/data': 0.006017161067575216, 'timing/model': 1.2126996270380914, 'learning_rate/qwen_vl_interface': 9.633862082360336e-06, 'learning_rate/action_model': 9.633862082360335e-05, 'epoch': 0.28}) +[2026-06-17 07:52:24,851][starVLA.training.train_starvla][INFO] - Step 969, Loss: {'train/loss': 1.2598804235458374, 'train/grad_norm_pre_clip': 6.609481334686279, 'timing/data': 0.004795423010364175, 'timing/model': 1.335629104054533, 'learning_rate/qwen_vl_interface': 9.633029001247654e-06, 'learning_rate/action_model': 9.633029001247653e-05, 'epoch': 0.29}) +[2026-06-17 07:52:29,504][starVLA.training.train_starvla][INFO] - Step 970, Loss: {'train/loss': 1.0764755010604858, 'train/grad_norm_pre_clip': 6.535632133483887, 'timing/data': 0.005007126019336283, 'timing/model': 1.205998124089092, 'learning_rate/qwen_vl_interface': 9.63219501152834e-06, 'learning_rate/action_model': 9.632195011528341e-05, 'epoch': 0.29}) +[2026-06-17 07:52:34,289][starVLA.training.train_starvla][INFO] - Step 971, Loss: {'train/loss': 0.7989193797111511, 'train/grad_norm_pre_clip': 11.442047119140625, 'timing/data': 0.004844935960136354, 'timing/model': 1.2036538200918585, 'learning_rate/qwen_vl_interface': 9.63136011337529e-06, 'learning_rate/action_model': 9.63136011337529e-05, 'epoch': 0.29}) +[2026-06-17 07:52:39,068][starVLA.training.train_starvla][INFO] - Step 972, Loss: {'train/loss': 0.9027624130249023, 'train/grad_norm_pre_clip': 9.43518352508545, 'timing/data': 0.005019554984755814, 'timing/model': 1.2041739349951968, 'learning_rate/qwen_vl_interface': 9.63052430696157e-06, 'learning_rate/action_model': 9.63052430696157e-05, 'epoch': 0.29}) +[2026-06-17 07:52:43,940][starVLA.training.train_starvla][INFO] - Step 973, Loss: {'train/loss': 1.1894475221633911, 'train/grad_norm_pre_clip': 10.918354988098145, 'timing/data': 0.004644389031454921, 'timing/model': 1.20810386701487, 'learning_rate/qwen_vl_interface': 9.629687592460449e-06, 'learning_rate/action_model': 9.62968759246045e-05, 'epoch': 0.29}) +[2026-06-17 07:52:48,717][starVLA.training.train_starvla][INFO] - Step 974, Loss: {'train/loss': 0.9309527277946472, 'train/grad_norm_pre_clip': 18.26508331298828, 'timing/data': 0.004823917988687754, 'timing/model': 1.2052196930162609, 'learning_rate/qwen_vl_interface': 9.628849970045378e-06, 'learning_rate/action_model': 9.628849970045378e-05, 'epoch': 0.29}) +[2026-06-17 07:52:53,475][starVLA.training.train_starvla][INFO] - Step 975, Loss: {'train/loss': 1.050523281097412, 'train/grad_norm_pre_clip': 8.491194725036621, 'timing/data': 0.004831189988180995, 'timing/model': 1.303304648026824, 'learning_rate/qwen_vl_interface': 9.628011439889997e-06, 'learning_rate/action_model': 9.628011439889996e-05, 'epoch': 0.29}) +[2026-06-17 07:52:58,141][starVLA.training.train_starvla][INFO] - Step 976, Loss: {'train/loss': 0.8125418424606323, 'train/grad_norm_pre_clip': 7.942999362945557, 'timing/data': 0.004575292929075658, 'timing/model': 1.2134066889993846, 'learning_rate/qwen_vl_interface': 9.62717200216813e-06, 'learning_rate/action_model': 9.62717200216813e-05, 'epoch': 0.29}) +[2026-06-17 07:53:02,938][starVLA.training.train_starvla][INFO] - Step 977, Loss: {'train/loss': 0.8407948613166809, 'train/grad_norm_pre_clip': 8.645635604858398, 'timing/data': 0.004753350047394633, 'timing/model': 1.205102534033358, 'learning_rate/qwen_vl_interface': 9.6263316570538e-06, 'learning_rate/action_model': 9.6263316570538e-05, 'epoch': 0.29}) +[2026-06-17 07:53:07,723][starVLA.training.train_starvla][INFO] - Step 978, Loss: {'train/loss': 1.1278046369552612, 'train/grad_norm_pre_clip': 6.159210681915283, 'timing/data': 0.00427896308247, 'timing/model': 1.2046514230314642, 'learning_rate/qwen_vl_interface': 9.625490404721208e-06, 'learning_rate/action_model': 9.625490404721208e-05, 'epoch': 0.29}) +[2026-06-17 07:53:12,517][starVLA.training.train_starvla][INFO] - Step 979, Loss: {'train/loss': 1.012518286705017, 'train/grad_norm_pre_clip': 6.985346794128418, 'timing/data': 0.005131371086463332, 'timing/model': 1.2078953679883853, 'learning_rate/qwen_vl_interface': 9.624648245344747e-06, 'learning_rate/action_model': 9.624648245344746e-05, 'epoch': 0.29}) +[2026-06-17 07:53:17,305][starVLA.training.train_starvla][INFO] - Step 980, Loss: {'train/loss': 1.0243552923202515, 'train/grad_norm_pre_clip': 7.201263904571533, 'timing/data': 0.004834833089262247, 'timing/model': 1.2068925969069824, 'learning_rate/qwen_vl_interface': 9.623805179098996e-06, 'learning_rate/action_model': 9.623805179098996e-05, 'epoch': 0.29}) +[2026-06-17 07:53:22,092][starVLA.training.train_starvla][INFO] - Step 981, Loss: {'train/loss': 1.0266551971435547, 'train/grad_norm_pre_clip': 7.612673759460449, 'timing/data': 0.00473111099563539, 'timing/model': 1.3342422000132501, 'learning_rate/qwen_vl_interface': 9.622961206158726e-06, 'learning_rate/action_model': 9.622961206158726e-05, 'epoch': 0.29}) +[2026-06-17 07:53:26,749][starVLA.training.train_starvla][INFO] - Step 982, Loss: {'train/loss': 0.8186625838279724, 'train/grad_norm_pre_clip': 7.842198848724365, 'timing/data': 0.004841370973736048, 'timing/model': 1.2061652459669858, 'learning_rate/qwen_vl_interface': 9.622116326698894e-06, 'learning_rate/action_model': 9.622116326698893e-05, 'epoch': 0.29}) +[2026-06-17 07:53:31,621][starVLA.training.train_starvla][INFO] - Step 983, Loss: {'train/loss': 0.7878480553627014, 'train/grad_norm_pre_clip': 8.146931648254395, 'timing/data': 0.005016357987187803, 'timing/model': 1.2060342079494148, 'learning_rate/qwen_vl_interface': 9.621270540894639e-06, 'learning_rate/action_model': 9.621270540894639e-05, 'epoch': 0.29}) +[2026-06-17 07:53:36,418][starVLA.training.train_starvla][INFO] - Step 984, Loss: {'train/loss': 0.9323185086250305, 'train/grad_norm_pre_clip': 7.203476428985596, 'timing/data': 0.004893834935501218, 'timing/model': 1.2067574959946796, 'learning_rate/qwen_vl_interface': 9.6204238489213e-06, 'learning_rate/action_model': 9.6204238489213e-05, 'epoch': 0.29}) +[2026-06-17 07:53:41,205][starVLA.training.train_starvla][INFO] - Step 985, Loss: {'train/loss': 1.138209581375122, 'train/grad_norm_pre_clip': 10.751530647277832, 'timing/data': 0.0049242760287597775, 'timing/model': 1.2099956199526787, 'learning_rate/qwen_vl_interface': 9.619576250954393e-06, 'learning_rate/action_model': 9.619576250954393e-05, 'epoch': 0.29}) +[2026-06-17 07:53:46,047][starVLA.training.train_starvla][INFO] - Step 986, Loss: {'train/loss': 0.8611827492713928, 'train/grad_norm_pre_clip': 7.229556083679199, 'timing/data': 0.005253324052318931, 'timing/model': 1.2096027249936014, 'learning_rate/qwen_vl_interface': 9.618727747169628e-06, 'learning_rate/action_model': 9.618727747169627e-05, 'epoch': 0.29}) +[2026-06-17 07:53:50,915][starVLA.training.train_starvla][INFO] - Step 987, Loss: {'train/loss': 1.0972226858139038, 'train/grad_norm_pre_clip': 9.45593547821045, 'timing/data': 0.005167947965674102, 'timing/model': 1.3264039510395378, 'learning_rate/qwen_vl_interface': 9.617878337742902e-06, 'learning_rate/action_model': 9.617878337742902e-05, 'epoch': 0.29}) +[2026-06-17 07:53:55,591][starVLA.training.train_starvla][INFO] - Step 988, Loss: {'train/loss': 0.7142619490623474, 'train/grad_norm_pre_clip': 7.543315410614014, 'timing/data': 0.004595909966155887, 'timing/model': 1.2099721520207822, 'learning_rate/qwen_vl_interface': 9.617028022850294e-06, 'learning_rate/action_model': 9.617028022850293e-05, 'epoch': 0.29}) +[2026-06-17 07:54:00,437][starVLA.training.train_starvla][INFO] - Step 989, Loss: {'train/loss': 1.0177578926086426, 'train/grad_norm_pre_clip': 8.976430892944336, 'timing/data': 0.004884163034148514, 'timing/model': 1.2064896289957687, 'learning_rate/qwen_vl_interface': 9.616176802668078e-06, 'learning_rate/action_model': 9.616176802668078e-05, 'epoch': 0.29}) +[2026-06-17 07:54:05,227][starVLA.training.train_starvla][INFO] - Step 990, Loss: {'train/loss': 0.9637030959129333, 'train/grad_norm_pre_clip': 10.490777969360352, 'timing/data': 0.004788120044395328, 'timing/model': 1.205987337976694, 'learning_rate/qwen_vl_interface': 9.615324677372712e-06, 'learning_rate/action_model': 9.615324677372713e-05, 'epoch': 0.29}) +[2026-06-17 07:54:10,044][starVLA.training.train_starvla][INFO] - Step 991, Loss: {'train/loss': 0.5964638590812683, 'train/grad_norm_pre_clip': 5.610760688781738, 'timing/data': 0.006076798075810075, 'timing/model': 1.2054734689882025, 'learning_rate/qwen_vl_interface': 9.614471647140844e-06, 'learning_rate/action_model': 9.614471647140844e-05, 'epoch': 0.29}) +[2026-06-17 07:54:14,902][starVLA.training.train_starvla][INFO] - Step 992, Loss: {'train/loss': 0.9243485331535339, 'train/grad_norm_pre_clip': 8.867772102355957, 'timing/data': 0.004745495971292257, 'timing/model': 1.3344636900583282, 'learning_rate/qwen_vl_interface': 9.613617712149309e-06, 'learning_rate/action_model': 9.613617712149308e-05, 'epoch': 0.29}) +[2026-06-17 07:54:19,562][starVLA.training.train_starvla][INFO] - Step 993, Loss: {'train/loss': 0.9360330700874329, 'train/grad_norm_pre_clip': 11.045432090759277, 'timing/data': 0.004729819018393755, 'timing/model': 1.2065539399627596, 'learning_rate/qwen_vl_interface': 9.612762872575124e-06, 'learning_rate/action_model': 9.612762872575124e-05, 'epoch': 0.29}) +[2026-06-17 07:54:24,353][starVLA.training.train_starvla][INFO] - Step 994, Loss: {'train/loss': 0.8241182565689087, 'train/grad_norm_pre_clip': 7.652454853057861, 'timing/data': 0.004503667936660349, 'timing/model': 1.2066197340609506, 'learning_rate/qwen_vl_interface': 9.611907128595503e-06, 'learning_rate/action_model': 9.611907128595502e-05, 'epoch': 0.29}) +[2026-06-17 07:54:29,087][starVLA.training.train_starvla][INFO] - Step 995, Loss: {'train/loss': 1.3455955982208252, 'train/grad_norm_pre_clip': 13.62139892578125, 'timing/data': 0.004709589993581176, 'timing/model': 1.2063076900085434, 'learning_rate/qwen_vl_interface': 9.61105048038784e-06, 'learning_rate/action_model': 9.611050480387839e-05, 'epoch': 0.29}) +[2026-06-17 07:54:33,842][starVLA.training.train_starvla][INFO] - Step 996, Loss: {'train/loss': 0.8817501664161682, 'train/grad_norm_pre_clip': 8.785999298095703, 'timing/data': 0.00488687795586884, 'timing/model': 1.209877207991667, 'learning_rate/qwen_vl_interface': 9.610192928129716e-06, 'learning_rate/action_model': 9.610192928129717e-05, 'epoch': 0.29}) +[2026-06-17 07:54:38,605][starVLA.training.train_starvla][INFO] - Step 997, Loss: {'train/loss': 0.6976178884506226, 'train/grad_norm_pre_clip': 9.065990447998047, 'timing/data': 0.004708630032837391, 'timing/model': 1.2072367270011455, 'learning_rate/qwen_vl_interface': 9.609334471998911e-06, 'learning_rate/action_model': 9.609334471998912e-05, 'epoch': 0.29}) +[2026-06-17 07:54:43,398][starVLA.training.train_starvla][INFO] - Step 998, Loss: {'train/loss': 0.8079546689987183, 'train/grad_norm_pre_clip': 7.627168655395508, 'timing/data': 0.005775270983576775, 'timing/model': 1.3385706170229241, 'learning_rate/qwen_vl_interface': 9.608475112173379e-06, 'learning_rate/action_model': 9.608475112173379e-05, 'epoch': 0.29}) +[2026-06-17 07:54:48,097][starVLA.training.train_starvla][INFO] - Step 999, Loss: {'train/loss': 0.7203737497329712, 'train/grad_norm_pre_clip': 9.111276626586914, 'timing/data': 0.00507372198626399, 'timing/model': 1.2075888360850513, 'learning_rate/qwen_vl_interface': 9.607614848831264e-06, 'learning_rate/action_model': 9.607614848831264e-05, 'epoch': 0.29}) +[2026-06-17 08:03:09,286][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1000 +[2026-06-17 08:03:09,287][starVLA.training.train_starvla][INFO] - Step 1000, Loss: {'train/loss': 0.876621663570404, 'train/grad_norm_pre_clip': 15.471301078796387, 'eval/loss': 0.9607166051864624, 'eval/latency_0/loss': 0.9607166051864624, 'eval/demon_fire_precision': 0.8860445912469034, 'eval/demon_fire_recall': 0.8911960132890365, 'eval/demon_fire_f1': 0.8886128364389234, 'eval/demon_fire_rate_model': 0.6735261401557285, 'eval/demon_fire_rate_teacher': 0.6696329254727474, 'eval/demon_left_precision': 0.6779586449626045, 'eval/demon_left_recall': 0.8730878186968839, 'eval/demon_left_f1': 0.7632491332342745, 'eval/demon_left_rate_model': 0.632091212458287, 'eval/demon_left_rate_teacher': 0.49082313681868744, 'eval/demon_right_precision': 0.8201058201058201, 'eval/demon_right_recall': 0.6129943502824858, 'eval/demon_right_f1': 0.7015842224377626, 'eval/demon_right_rate_model': 0.367908787541713, 'eval/demon_right_rate_teacher': 0.4922135706340378, 'eval/demon_move_f1': 0.7324166778360186, 'eval/demon_cc_f1': 0.8351793946035451, 'eval/latency_0/demon_fire_precision': 0.8860445912469034, 'eval/latency_0/demon_fire_recall': 0.8911960132890365, 'eval/latency_0/demon_fire_f1': 0.8886128364389234, 'eval/latency_0/demon_fire_rate_model': 0.6735261401557285, 'eval/latency_0/demon_fire_rate_teacher': 0.6696329254727474, 'eval/latency_0/demon_left_precision': 0.6779586449626045, 'eval/latency_0/demon_left_recall': 0.8730878186968839, 'eval/latency_0/demon_left_f1': 0.7632491332342745, 'eval/latency_0/demon_left_rate_model': 0.632091212458287, 'eval/latency_0/demon_left_rate_teacher': 0.49082313681868744, 'eval/latency_0/demon_right_precision': 0.8201058201058201, 'eval/latency_0/demon_right_recall': 0.6129943502824858, 'eval/latency_0/demon_right_f1': 0.7015842224377626, 'eval/latency_0/demon_right_rate_model': 0.367908787541713, 'eval/latency_0/demon_right_rate_teacher': 0.4922135706340378, 'eval/latency_0/demon_move_f1': 0.7324166778360186, 'eval/latency_0/demon_cc_f1': 0.8351793946035451, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 2497.0, 'rl_games_eval/mid_train/mean_length': 2830.6, 'rl_games_eval/mid_train/std_reward': 1304.0191716382087, 'rl_games_eval/mid_train/std_length': 990.1094080958932, 'rl_games_eval/mid_train/macro_mean_reward': 2497.0, 'rl_games_eval/mid_train/macro_mean_length': 2830.6, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/demon_attack__latency_0/mean_reward': 2497.0, 'rl_games_eval/mid_train/demon_attack__latency_0/mean_length': 2830.6, 'rl_games_eval/mid_train/demon_attack__latency_0/std_reward': 1304.0191716382087, 'rl_games_eval/mid_train/demon_attack__latency_0/std_length': 990.1094080958932, 'rl_games_eval/mid_train/demon_attack__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 2497.0, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.004487915080972016, 'timing/model': 1.2064865470165387, 'learning_rate/qwen_vl_interface': 9.606753682150904e-06, 'learning_rate/action_model': 9.606753682150903e-05, 'epoch': 0.29}) +[2026-06-17 08:03:09,287][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_1000_state +[2026-06-17 08:03:09,290][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 08:03:12,842][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_1000_state/model.safetensors +[2026-06-17 08:03:20,562][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_1000_state/optimizer.bin +[2026-06-17 08:03:20,562][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_1000_state/sampler.bin +[2026-06-17 08:03:20,562][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_1000_state/sampler_1.bin +[2026-06-17 08:03:20,563][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_1000_state/random_states_0.pkl +[2026-06-17 08:03:20,608][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1000 +[2026-06-17 08:03:20,608][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 08:03:20,636][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 08:03:25,379][starVLA.training.train_starvla][INFO] - Step 1001, Loss: {'train/loss': 0.6283727884292603, 'train/grad_norm_pre_clip': 6.544304847717285, 'timing/data': 0.0051224230555817485, 'timing/model': 1.1830919500207528, 'learning_rate/qwen_vl_interface': 9.605891612310815e-06, 'learning_rate/action_model': 9.605891612310814e-05, 'epoch': 0.29}) +[2026-06-17 08:03:30,123][starVLA.training.train_starvla][INFO] - Step 1002, Loss: {'train/loss': 0.7312019467353821, 'train/grad_norm_pre_clip': 9.104350090026855, 'timing/data': 0.004849742981605232, 'timing/model': 1.3145627890480682, 'learning_rate/qwen_vl_interface': 9.60502863948971e-06, 'learning_rate/action_model': 9.60502863948971e-05, 'epoch': 0.29}) +[2026-06-17 08:03:34,732][starVLA.training.train_starvla][INFO] - Step 1003, Loss: {'train/loss': 1.3125309944152832, 'train/grad_norm_pre_clip': 11.66145133972168, 'timing/data': 0.005866575986146927, 'timing/model': 1.1902458670083433, 'learning_rate/qwen_vl_interface': 9.604164763866479e-06, 'learning_rate/action_model': 9.60416476386648e-05, 'epoch': 0.3}) +[2026-06-17 08:03:39,472][starVLA.training.train_starvla][INFO] - Step 1004, Loss: {'train/loss': 0.8181111216545105, 'train/grad_norm_pre_clip': 12.5498628616333, 'timing/data': 0.00498570897616446, 'timing/model': 1.1905730670550838, 'learning_rate/qwen_vl_interface': 9.60329998562021e-06, 'learning_rate/action_model': 9.60329998562021e-05, 'epoch': 0.3}) +[2026-06-17 08:03:44,230][starVLA.training.train_starvla][INFO] - Step 1005, Loss: {'train/loss': 1.177069902420044, 'train/grad_norm_pre_clip': 10.263777732849121, 'timing/data': 0.005584720987826586, 'timing/model': 1.1926199229201302, 'learning_rate/qwen_vl_interface': 9.602434304930168e-06, 'learning_rate/action_model': 9.602434304930168e-05, 'epoch': 0.3}) +[2026-06-17 08:03:48,893][starVLA.training.train_starvla][INFO] - Step 1006, Loss: {'train/loss': 1.1462739706039429, 'train/grad_norm_pre_clip': 9.557759284973145, 'timing/data': 0.0051400590455159545, 'timing/model': 1.1932333779986948, 'learning_rate/qwen_vl_interface': 9.601567721975811e-06, 'learning_rate/action_model': 9.601567721975812e-05, 'epoch': 0.3}) +[2026-06-17 08:03:53,619][starVLA.training.train_starvla][INFO] - Step 1007, Loss: {'train/loss': 1.2913479804992676, 'train/grad_norm_pre_clip': 11.053426742553711, 'timing/data': 0.005661169067025185, 'timing/model': 1.193486970034428, 'learning_rate/qwen_vl_interface': 9.600700236936785e-06, 'learning_rate/action_model': 9.600700236936785e-05, 'epoch': 0.3}) +[2026-06-17 08:03:58,339][starVLA.training.train_starvla][INFO] - Step 1008, Loss: {'train/loss': 1.0303981304168701, 'train/grad_norm_pre_clip': 7.9672040939331055, 'timing/data': 0.004702393081970513, 'timing/model': 1.3005671940045431, 'learning_rate/qwen_vl_interface': 9.599831849992916e-06, 'learning_rate/action_model': 9.599831849992916e-05, 'epoch': 0.3}) +[2026-06-17 08:04:02,963][starVLA.training.train_starvla][INFO] - Step 1009, Loss: {'train/loss': 0.82330322265625, 'train/grad_norm_pre_clip': 7.8158063888549805, 'timing/data': 0.005238195997662842, 'timing/model': 1.1964481200557202, 'learning_rate/qwen_vl_interface': 9.598962561324225e-06, 'learning_rate/action_model': 9.598962561324225e-05, 'epoch': 0.3}) +[2026-06-17 08:04:07,717][starVLA.training.train_starvla][INFO] - Step 1010, Loss: {'train/loss': 0.8146910071372986, 'train/grad_norm_pre_clip': 9.918421745300293, 'timing/data': 0.005375323002226651, 'timing/model': 1.1960294810123742, 'learning_rate/qwen_vl_interface': 9.598092371110916e-06, 'learning_rate/action_model': 9.598092371110916e-05, 'epoch': 0.3}) +[2026-06-17 08:04:12,481][starVLA.training.train_starvla][INFO] - Step 1011, Loss: {'train/loss': 0.8266905546188354, 'train/grad_norm_pre_clip': 9.118518829345703, 'timing/data': 0.005288371001370251, 'timing/model': 1.1999194669770077, 'learning_rate/qwen_vl_interface': 9.59722127953338e-06, 'learning_rate/action_model': 9.597221279533379e-05, 'epoch': 0.3}) +[2026-06-17 08:04:17,234][starVLA.training.train_starvla][INFO] - Step 1012, Loss: {'train/loss': 0.9454041123390198, 'train/grad_norm_pre_clip': 14.806502342224121, 'timing/data': 0.0051658329321071506, 'timing/model': 1.1986330959480256, 'learning_rate/qwen_vl_interface': 9.596349286772194e-06, 'learning_rate/action_model': 9.596349286772193e-05, 'epoch': 0.3}) +[2026-06-17 08:04:21,954][starVLA.training.train_starvla][INFO] - Step 1013, Loss: {'train/loss': 1.0083421468734741, 'train/grad_norm_pre_clip': 7.5449910163879395, 'timing/data': 0.005295965005643666, 'timing/model': 1.1975385900586843, 'learning_rate/qwen_vl_interface': 9.595476393008123e-06, 'learning_rate/action_model': 9.595476393008123e-05, 'epoch': 0.3}) +[2026-06-17 08:04:26,721][starVLA.training.train_starvla][INFO] - Step 1014, Loss: {'train/loss': 0.722009539604187, 'train/grad_norm_pre_clip': 7.209970474243164, 'timing/data': 0.0052196470787748694, 'timing/model': 1.332114455057308, 'learning_rate/qwen_vl_interface': 9.59460259842212e-06, 'learning_rate/action_model': 9.59460259842212e-05, 'epoch': 0.3}) +[2026-06-17 08:04:31,351][starVLA.training.train_starvla][INFO] - Step 1015, Loss: {'train/loss': 0.8990228772163391, 'train/grad_norm_pre_clip': 7.614672660827637, 'timing/data': 0.005242432001978159, 'timing/model': 1.1993573741056025, 'learning_rate/qwen_vl_interface': 9.593727903195323e-06, 'learning_rate/action_model': 9.593727903195324e-05, 'epoch': 0.3}) +[2026-06-17 08:04:36,111][starVLA.training.train_starvla][INFO] - Step 1016, Loss: {'train/loss': 0.8143967986106873, 'train/grad_norm_pre_clip': 15.227699279785156, 'timing/data': 0.0049470660742372274, 'timing/model': 1.1992301670834422, 'learning_rate/qwen_vl_interface': 9.59285230750906e-06, 'learning_rate/action_model': 9.592852307509059e-05, 'epoch': 0.3}) +[2026-06-17 08:04:40,875][starVLA.training.train_starvla][INFO] - Step 1017, Loss: {'train/loss': 0.9742817878723145, 'train/grad_norm_pre_clip': 9.987401962280273, 'timing/data': 0.005499718943610787, 'timing/model': 1.2007200010120869, 'learning_rate/qwen_vl_interface': 9.591975811544839e-06, 'learning_rate/action_model': 9.591975811544839e-05, 'epoch': 0.3}) +[2026-06-17 08:04:45,627][starVLA.training.train_starvla][INFO] - Step 1018, Loss: {'train/loss': 0.8350905776023865, 'train/grad_norm_pre_clip': 10.538991928100586, 'timing/data': 0.0053043930092826486, 'timing/model': 1.1983659529360011, 'learning_rate/qwen_vl_interface': 9.59109841548436e-06, 'learning_rate/action_model': 9.591098415484359e-05, 'epoch': 0.3}) +[2026-06-17 08:04:50,429][starVLA.training.train_starvla][INFO] - Step 1019, Loss: {'train/loss': 0.8866956830024719, 'train/grad_norm_pre_clip': 13.635832786560059, 'timing/data': 0.00535113993100822, 'timing/model': 1.2385747389635071, 'learning_rate/qwen_vl_interface': 9.590220119509508e-06, 'learning_rate/action_model': 9.590220119509508e-05, 'epoch': 0.3}) +[2026-06-17 08:04:55,183][starVLA.training.train_starvla][INFO] - Step 1020, Loss: {'train/loss': 0.8285511136054993, 'train/grad_norm_pre_clip': 19.955078125, 'timing/data': 0.005134435021318495, 'timing/model': 1.3157430100254714, 'learning_rate/qwen_vl_interface': 9.589340923802355e-06, 'learning_rate/action_model': 9.589340923802354e-05, 'epoch': 0.3}) +[2026-06-17 08:04:59,818][starVLA.training.train_starvla][INFO] - Step 1021, Loss: {'train/loss': 0.8318800330162048, 'train/grad_norm_pre_clip': 8.797341346740723, 'timing/data': 0.005234442069195211, 'timing/model': 1.198700625100173, 'learning_rate/qwen_vl_interface': 9.58846082854516e-06, 'learning_rate/action_model': 9.58846082854516e-05, 'epoch': 0.3}) +[2026-06-17 08:05:04,653][starVLA.training.train_starvla][INFO] - Step 1022, Loss: {'train/loss': 0.7660917639732361, 'train/grad_norm_pre_clip': 13.302986145019531, 'timing/data': 0.005363867967389524, 'timing/model': 1.2016056169522926, 'learning_rate/qwen_vl_interface': 9.587579833920366e-06, 'learning_rate/action_model': 9.587579833920365e-05, 'epoch': 0.3}) +[2026-06-17 08:05:09,425][starVLA.training.train_starvla][INFO] - Step 1023, Loss: {'train/loss': 1.1161048412322998, 'train/grad_norm_pre_clip': 12.571228981018066, 'timing/data': 0.005189660005271435, 'timing/model': 1.2003337650094181, 'learning_rate/qwen_vl_interface': 9.586697940110605e-06, 'learning_rate/action_model': 9.586697940110605e-05, 'epoch': 0.3}) +[2026-06-17 08:05:14,197][starVLA.training.train_starvla][INFO] - Step 1024, Loss: {'train/loss': 1.0208503007888794, 'train/grad_norm_pre_clip': 10.0559663772583, 'timing/data': 0.005364519078284502, 'timing/model': 1.2064402060350403, 'learning_rate/qwen_vl_interface': 9.585815147298696e-06, 'learning_rate/action_model': 9.585815147298696e-05, 'epoch': 0.3}) +[2026-06-17 08:05:18,978][starVLA.training.train_starvla][INFO] - Step 1025, Loss: {'train/loss': 0.9470886588096619, 'train/grad_norm_pre_clip': 9.451176643371582, 'timing/data': 0.005206870031543076, 'timing/model': 1.3346201439853758, 'learning_rate/qwen_vl_interface': 9.58493145566764e-06, 'learning_rate/action_model': 9.58493145566764e-05, 'epoch': 0.3}) +[2026-06-17 08:05:23,623][starVLA.training.train_starvla][INFO] - Step 1026, Loss: {'train/loss': 0.8910353183746338, 'train/grad_norm_pre_clip': 16.239086151123047, 'timing/data': 0.005270440015010536, 'timing/model': 1.2036855049664155, 'learning_rate/qwen_vl_interface': 9.584046865400627e-06, 'learning_rate/action_model': 9.584046865400627e-05, 'epoch': 0.3}) +[2026-06-17 08:05:28,407][starVLA.training.train_starvla][INFO] - Step 1027, Loss: {'train/loss': 0.688235878944397, 'train/grad_norm_pre_clip': 6.181099891662598, 'timing/data': 0.0056603000266477466, 'timing/model': 1.2057130939792842, 'learning_rate/qwen_vl_interface': 9.583161376681039e-06, 'learning_rate/action_model': 9.583161376681038e-05, 'epoch': 0.3}) +[2026-06-17 08:05:33,107][starVLA.training.train_starvla][INFO] - Step 1028, Loss: {'train/loss': 0.9199538826942444, 'train/grad_norm_pre_clip': 11.272176742553711, 'timing/data': 0.005612531094811857, 'timing/model': 1.2041069699916989, 'learning_rate/qwen_vl_interface': 9.582274989692432e-06, 'learning_rate/action_model': 9.582274989692432e-05, 'epoch': 0.3}) +[2026-06-17 08:05:37,832][starVLA.training.train_starvla][INFO] - Step 1029, Loss: {'train/loss': 1.2494477033615112, 'train/grad_norm_pre_clip': 10.041191101074219, 'timing/data': 0.005517506971955299, 'timing/model': 1.2031231549335644, 'learning_rate/qwen_vl_interface': 9.58138770461856e-06, 'learning_rate/action_model': 9.58138770461856e-05, 'epoch': 0.3}) +[2026-06-17 08:05:42,609][starVLA.training.train_starvla][INFO] - Step 1030, Loss: {'train/loss': 1.0861482620239258, 'train/grad_norm_pre_clip': 10.808342933654785, 'timing/data': 0.00536058796569705, 'timing/model': 1.2012114969547838, 'learning_rate/qwen_vl_interface': 9.580499521643355e-06, 'learning_rate/action_model': 9.580499521643355e-05, 'epoch': 0.3}) +[2026-06-17 08:05:47,381][starVLA.training.train_starvla][INFO] - Step 1031, Loss: {'train/loss': 0.8993744254112244, 'train/grad_norm_pre_clip': 10.116037368774414, 'timing/data': 0.005390388076193631, 'timing/model': 1.331095114001073, 'learning_rate/qwen_vl_interface': 9.579610440950939e-06, 'learning_rate/action_model': 9.579610440950937e-05, 'epoch': 0.3}) +[2026-06-17 08:05:52,027][starVLA.training.train_starvla][INFO] - Step 1032, Loss: {'train/loss': 0.9103485345840454, 'train/grad_norm_pre_clip': 9.375146865844727, 'timing/data': 0.005204856977798045, 'timing/model': 1.2023387750377879, 'learning_rate/qwen_vl_interface': 9.57872046272562e-06, 'learning_rate/action_model': 9.57872046272562e-05, 'epoch': 0.3}) +[2026-06-17 08:05:56,811][starVLA.training.train_starvla][INFO] - Step 1033, Loss: {'train/loss': 0.816185712814331, 'train/grad_norm_pre_clip': 9.136113166809082, 'timing/data': 0.005381901981309056, 'timing/model': 1.2017437980975956, 'learning_rate/qwen_vl_interface': 9.577829587151892e-06, 'learning_rate/action_model': 9.577829587151891e-05, 'epoch': 0.3}) +[2026-06-17 08:06:01,577][starVLA.training.train_starvla][INFO] - Step 1034, Loss: {'train/loss': 0.7702887058258057, 'train/grad_norm_pre_clip': 8.825185775756836, 'timing/data': 0.005455670063383877, 'timing/model': 1.2013910259120166, 'learning_rate/qwen_vl_interface': 9.576937814414433e-06, 'learning_rate/action_model': 9.576937814414432e-05, 'epoch': 0.3}) +[2026-06-17 08:06:06,310][starVLA.training.train_starvla][INFO] - Step 1035, Loss: {'train/loss': 0.918535590171814, 'train/grad_norm_pre_clip': 9.238016128540039, 'timing/data': 0.0049612660659477115, 'timing/model': 1.201768858009018, 'learning_rate/qwen_vl_interface': 9.576045144698109e-06, 'learning_rate/action_model': 9.576045144698108e-05, 'epoch': 0.3}) +[2026-06-17 08:06:11,070][starVLA.training.train_starvla][INFO] - Step 1036, Loss: {'train/loss': 0.7715073823928833, 'train/grad_norm_pre_clip': 10.106022834777832, 'timing/data': 0.005257213953882456, 'timing/model': 1.2014444030355662, 'learning_rate/qwen_vl_interface': 9.57515157818797e-06, 'learning_rate/action_model': 9.57515157818797e-05, 'epoch': 0.3}) +[2026-06-17 08:06:15,859][starVLA.training.train_starvla][INFO] - Step 1037, Loss: {'train/loss': 1.0668087005615234, 'train/grad_norm_pre_clip': 9.965723991394043, 'timing/data': 0.0054500020341947675, 'timing/model': 1.3372896820073947, 'learning_rate/qwen_vl_interface': 9.574257115069256e-06, 'learning_rate/action_model': 9.574257115069256e-05, 'epoch': 0.31}) +[2026-06-17 08:06:20,514][starVLA.training.train_starvla][INFO] - Step 1038, Loss: {'train/loss': 1.3115674257278442, 'train/grad_norm_pre_clip': 17.57616424560547, 'timing/data': 0.005530536058358848, 'timing/model': 1.2079520450206473, 'learning_rate/qwen_vl_interface': 9.573361755527389e-06, 'learning_rate/action_model': 9.573361755527389e-05, 'epoch': 0.31}) +[2026-06-17 08:06:25,309][starVLA.training.train_starvla][INFO] - Step 1039, Loss: {'train/loss': 1.087573766708374, 'train/grad_norm_pre_clip': 13.026385307312012, 'timing/data': 0.0055062370374798775, 'timing/model': 1.2046345750568435, 'learning_rate/qwen_vl_interface': 9.572465499747978e-06, 'learning_rate/action_model': 9.572465499747978e-05, 'epoch': 0.31}) +[2026-06-17 08:06:30,091][starVLA.training.train_starvla][INFO] - Step 1040, Loss: {'train/loss': 0.9794859886169434, 'train/grad_norm_pre_clip': 14.151219367980957, 'timing/data': 0.005381003022193909, 'timing/model': 1.2045272500254214, 'learning_rate/qwen_vl_interface': 9.571568347916819e-06, 'learning_rate/action_model': 9.571568347916819e-05, 'epoch': 0.31}) +[2026-06-17 08:06:34,875][starVLA.training.train_starvla][INFO] - Step 1041, Loss: {'train/loss': 1.2787127494812012, 'train/grad_norm_pre_clip': 9.463025093078613, 'timing/data': 0.005192140000872314, 'timing/model': 1.2035852510016412, 'learning_rate/qwen_vl_interface': 9.570670300219889e-06, 'learning_rate/action_model': 9.570670300219889e-05, 'epoch': 0.31}) +[2026-06-17 08:06:39,542][starVLA.training.train_starvla][INFO] - Step 1042, Loss: {'train/loss': 0.8593867421150208, 'train/grad_norm_pre_clip': 9.7388277053833, 'timing/data': 0.005576194962486625, 'timing/model': 1.2035535520408303, 'learning_rate/qwen_vl_interface': 9.56977135684336e-06, 'learning_rate/action_model': 9.569771356843359e-05, 'epoch': 0.31}) +[2026-06-17 08:06:44,299][starVLA.training.train_starvla][INFO] - Step 1043, Loss: {'train/loss': 0.686089038848877, 'train/grad_norm_pre_clip': 6.612357139587402, 'timing/data': 0.0055758130038157105, 'timing/model': 1.2418353769462556, 'learning_rate/qwen_vl_interface': 9.568871517973576e-06, 'learning_rate/action_model': 9.568871517973577e-05, 'epoch': 0.31}) +[2026-06-17 08:06:48,951][starVLA.training.train_starvla][INFO] - Step 1044, Loss: {'train/loss': 1.1065865755081177, 'train/grad_norm_pre_clip': 11.583194732666016, 'timing/data': 0.005361710907891393, 'timing/model': 1.204424130031839, 'learning_rate/qwen_vl_interface': 9.567970783797081e-06, 'learning_rate/action_model': 9.567970783797082e-05, 'epoch': 0.31}) +[2026-06-17 08:06:53,693][starVLA.training.train_starvla][INFO] - Step 1045, Loss: {'train/loss': 1.1736642122268677, 'train/grad_norm_pre_clip': 11.780489921569824, 'timing/data': 0.005390119971707463, 'timing/model': 1.2035151700256392, 'learning_rate/qwen_vl_interface': 9.567069154500597e-06, 'learning_rate/action_model': 9.567069154500596e-05, 'epoch': 0.31}) +[2026-06-17 08:06:58,496][starVLA.training.train_starvla][INFO] - Step 1046, Loss: {'train/loss': 1.0065175294876099, 'train/grad_norm_pre_clip': 13.31721019744873, 'timing/data': 0.005320545984432101, 'timing/model': 1.2042519500246271, 'learning_rate/qwen_vl_interface': 9.56616663027103e-06, 'learning_rate/action_model': 9.566166630271029e-05, 'epoch': 0.31}) +[2026-06-17 08:07:03,279][starVLA.training.train_starvla][INFO] - Step 1047, Loss: {'train/loss': 1.0790503025054932, 'train/grad_norm_pre_clip': 14.584304809570312, 'timing/data': 0.005428893957287073, 'timing/model': 1.2032099280040711, 'learning_rate/qwen_vl_interface': 9.56526321129548e-06, 'learning_rate/action_model': 9.565263211295479e-05, 'epoch': 0.31}) +[2026-06-17 08:07:08,068][starVLA.training.train_starvla][INFO] - Step 1048, Loss: {'train/loss': 1.2101987600326538, 'train/grad_norm_pre_clip': 12.82569694519043, 'timing/data': 0.005202282918617129, 'timing/model': 1.327470207004808, 'learning_rate/qwen_vl_interface': 9.56435889776122e-06, 'learning_rate/action_model': 9.564358897761218e-05, 'epoch': 0.31}) +[2026-06-17 08:07:12,858][starVLA.training.train_starvla][INFO] - Step 1049, Loss: {'train/loss': 0.9200834631919861, 'train/grad_norm_pre_clip': 8.213547706604004, 'timing/data': 0.005465434049256146, 'timing/model': 1.337304302956909, 'learning_rate/qwen_vl_interface': 9.563453689855717e-06, 'learning_rate/action_model': 9.563453689855716e-05, 'epoch': 0.31}) +[2026-06-17 08:07:17,534][starVLA.training.train_starvla][INFO] - Step 1050, Loss: {'train/loss': 0.9963881373405457, 'train/grad_norm_pre_clip': 7.510779857635498, 'timing/data': 0.0054754429729655385, 'timing/model': 1.2051601619459689, 'learning_rate/qwen_vl_interface': 9.562547587766624e-06, 'learning_rate/action_model': 9.562547587766623e-05, 'epoch': 0.31}) +[2026-06-17 08:07:22,295][starVLA.training.train_starvla][INFO] - Step 1051, Loss: {'train/loss': 0.7970194220542908, 'train/grad_norm_pre_clip': 7.321937084197998, 'timing/data': 0.005542101920582354, 'timing/model': 1.2089733920292929, 'learning_rate/qwen_vl_interface': 9.561640591681775e-06, 'learning_rate/action_model': 9.561640591681774e-05, 'epoch': 0.31}) +[2026-06-17 08:07:27,082][starVLA.training.train_starvla][INFO] - Step 1052, Loss: {'train/loss': 0.6901207566261292, 'train/grad_norm_pre_clip': 8.798420906066895, 'timing/data': 0.005412152037024498, 'timing/model': 1.2066487319534644, 'learning_rate/qwen_vl_interface': 9.560732701789193e-06, 'learning_rate/action_model': 9.560732701789193e-05, 'epoch': 0.31}) +[2026-06-17 08:07:31,859][starVLA.training.train_starvla][INFO] - Step 1053, Loss: {'train/loss': 0.8604012131690979, 'train/grad_norm_pre_clip': 10.965678215026855, 'timing/data': 0.005020114942453802, 'timing/model': 1.2032310279319063, 'learning_rate/qwen_vl_interface': 9.55982391827708e-06, 'learning_rate/action_model': 9.559823918277081e-05, 'epoch': 0.31}) +[2026-06-17 08:07:36,641][starVLA.training.train_starvla][INFO] - Step 1054, Loss: {'train/loss': 0.9203014969825745, 'train/grad_norm_pre_clip': 9.619589805603027, 'timing/data': 0.005332699045538902, 'timing/model': 1.3337519650813192, 'learning_rate/qwen_vl_interface': 9.558914241333835e-06, 'learning_rate/action_model': 9.558914241333835e-05, 'epoch': 0.31}) +[2026-06-17 08:07:41,422][starVLA.training.train_starvla][INFO] - Step 1055, Loss: {'train/loss': 0.7318788766860962, 'train/grad_norm_pre_clip': 8.151421546936035, 'timing/data': 0.005136153078638017, 'timing/model': 1.3343146319966763, 'learning_rate/qwen_vl_interface': 9.55800367114803e-06, 'learning_rate/action_model': 9.55800367114803e-05, 'epoch': 0.31}) +[2026-06-17 08:07:46,075][starVLA.training.train_starvla][INFO] - Step 1056, Loss: {'train/loss': 0.8272324204444885, 'train/grad_norm_pre_clip': 7.430156230926514, 'timing/data': 0.005339207942597568, 'timing/model': 1.2040041020372882, 'learning_rate/qwen_vl_interface': 9.557092207908428e-06, 'learning_rate/action_model': 9.557092207908428e-05, 'epoch': 0.31}) +[2026-06-17 08:07:50,831][starVLA.training.train_starvla][INFO] - Step 1057, Loss: {'train/loss': 0.847812294960022, 'train/grad_norm_pre_clip': 9.775291442871094, 'timing/data': 0.0054254160495474935, 'timing/model': 1.2047694500070065, 'learning_rate/qwen_vl_interface': 9.556179851803978e-06, 'learning_rate/action_model': 9.556179851803978e-05, 'epoch': 0.31}) +[2026-06-17 08:07:55,663][starVLA.training.train_starvla][INFO] - Step 1058, Loss: {'train/loss': 0.935884416103363, 'train/grad_norm_pre_clip': 10.569212913513184, 'timing/data': 0.005782291060313582, 'timing/model': 1.2093940309714526, 'learning_rate/qwen_vl_interface': 9.555266603023809e-06, 'learning_rate/action_model': 9.555266603023808e-05, 'epoch': 0.31}) +[2026-06-17 08:08:00,473][starVLA.training.train_starvla][INFO] - Step 1059, Loss: {'train/loss': 0.8188561201095581, 'train/grad_norm_pre_clip': 7.308971405029297, 'timing/data': 0.005267305066809058, 'timing/model': 1.2119154080282897, 'learning_rate/qwen_vl_interface': 9.554352461757243e-06, 'learning_rate/action_model': 9.554352461757242e-05, 'epoch': 0.31}) +[2026-06-17 08:08:05,207][starVLA.training.train_starvla][INFO] - Step 1060, Loss: {'train/loss': 0.8817918300628662, 'train/grad_norm_pre_clip': 5.830512046813965, 'timing/data': 0.0054540380369871855, 'timing/model': 1.284209129982628, 'learning_rate/qwen_vl_interface': 9.553437428193778e-06, 'learning_rate/action_model': 9.553437428193778e-05, 'epoch': 0.31}) +[2026-06-17 08:08:09,857][starVLA.training.train_starvla][INFO] - Step 1061, Loss: {'train/loss': 1.0623074769973755, 'train/grad_norm_pre_clip': 10.902482986450195, 'timing/data': 0.005053477943874896, 'timing/model': 1.2036864960100502, 'learning_rate/qwen_vl_interface': 9.552521502523104e-06, 'learning_rate/action_model': 9.552521502523102e-05, 'epoch': 0.31}) +[2026-06-17 08:08:14,618][starVLA.training.train_starvla][INFO] - Step 1062, Loss: {'train/loss': 0.9608346223831177, 'train/grad_norm_pre_clip': 10.214940071105957, 'timing/data': 0.005280190031044185, 'timing/model': 1.2046380930114537, 'learning_rate/qwen_vl_interface': 9.551604684935093e-06, 'learning_rate/action_model': 9.551604684935092e-05, 'epoch': 0.31}) +[2026-06-17 08:08:19,496][starVLA.training.train_starvla][INFO] - Step 1063, Loss: {'train/loss': 0.9702079892158508, 'train/grad_norm_pre_clip': 7.143630027770996, 'timing/data': 0.005133177968673408, 'timing/model': 1.2032115630572662, 'learning_rate/qwen_vl_interface': 9.550686975619803e-06, 'learning_rate/action_model': 9.550686975619802e-05, 'epoch': 0.31}) +[2026-06-17 08:08:24,278][starVLA.training.train_starvla][INFO] - Step 1064, Loss: {'train/loss': 0.9281398057937622, 'train/grad_norm_pre_clip': 9.237105369567871, 'timing/data': 0.005302962032146752, 'timing/model': 1.2031967380316928, 'learning_rate/qwen_vl_interface': 9.549768374767473e-06, 'learning_rate/action_model': 9.549768374767473e-05, 'epoch': 0.31}) +[2026-06-17 08:08:29,063][starVLA.training.train_starvla][INFO] - Step 1065, Loss: {'train/loss': 0.7378977537155151, 'train/grad_norm_pre_clip': 10.2088623046875, 'timing/data': 0.005406087031587958, 'timing/model': 1.3337028790265322, 'learning_rate/qwen_vl_interface': 9.548848882568533e-06, 'learning_rate/action_model': 9.548848882568532e-05, 'epoch': 0.31}) +[2026-06-17 08:08:33,812][starVLA.training.train_starvla][INFO] - Step 1066, Loss: {'train/loss': 1.1096858978271484, 'train/grad_norm_pre_clip': 8.684842109680176, 'timing/data': 0.0055033169919624925, 'timing/model': 1.2973297610878944, 'learning_rate/qwen_vl_interface': 9.547928499213593e-06, 'learning_rate/action_model': 9.547928499213593e-05, 'epoch': 0.31}) +[2026-06-17 08:08:38,463][starVLA.training.train_starvla][INFO] - Step 1067, Loss: {'train/loss': 0.9843787550926208, 'train/grad_norm_pre_clip': 11.239518165588379, 'timing/data': 0.0051025800639763474, 'timing/model': 1.2037950799567625, 'learning_rate/qwen_vl_interface': 9.547007224893452e-06, 'learning_rate/action_model': 9.547007224893452e-05, 'epoch': 0.31}) +[2026-06-17 08:08:43,254][starVLA.training.train_starvla][INFO] - Step 1068, Loss: {'train/loss': 0.8297829627990723, 'train/grad_norm_pre_clip': 11.92837905883789, 'timing/data': 0.0054962459253147244, 'timing/model': 1.2065290410537273, 'learning_rate/qwen_vl_interface': 9.546085059799088e-06, 'learning_rate/action_model': 9.546085059799088e-05, 'epoch': 0.31}) +[2026-06-17 08:08:48,056][starVLA.training.train_starvla][INFO] - Step 1069, Loss: {'train/loss': 0.8971291184425354, 'train/grad_norm_pre_clip': 9.531349182128906, 'timing/data': 0.0062148780561983585, 'timing/model': 1.2053126619430259, 'learning_rate/qwen_vl_interface': 9.545162004121672e-06, 'learning_rate/action_model': 9.54516200412167e-05, 'epoch': 0.31}) +[2026-06-17 08:08:52,932][starVLA.training.train_starvla][INFO] - Step 1070, Loss: {'train/loss': 1.0880721807479858, 'train/grad_norm_pre_clip': 9.681096076965332, 'timing/data': 0.0054330750135704875, 'timing/model': 1.203377939062193, 'learning_rate/qwen_vl_interface': 9.544238058052546e-06, 'learning_rate/action_model': 9.544238058052545e-05, 'epoch': 0.31}) +[2026-06-17 08:08:57,720][starVLA.training.train_starvla][INFO] - Step 1071, Loss: {'train/loss': 1.0033323764801025, 'train/grad_norm_pre_clip': 10.108999252319336, 'timing/data': 0.005300899967551231, 'timing/model': 1.3336744499392807, 'learning_rate/qwen_vl_interface': 9.54331322178325e-06, 'learning_rate/action_model': 9.543313221783251e-05, 'epoch': 0.32}) +[2026-06-17 08:09:02,400][starVLA.training.train_starvla][INFO] - Step 1072, Loss: {'train/loss': 0.9539014101028442, 'train/grad_norm_pre_clip': 8.79256534576416, 'timing/data': 0.005370225990191102, 'timing/model': 1.2045513360062614, 'learning_rate/qwen_vl_interface': 9.542387495505505e-06, 'learning_rate/action_model': 9.542387495505506e-05, 'epoch': 0.32}) +[2026-06-17 08:09:07,188][starVLA.training.train_starvla][INFO] - Step 1073, Loss: {'train/loss': 0.796100378036499, 'train/grad_norm_pre_clip': 9.081249237060547, 'timing/data': 0.00505386502481997, 'timing/model': 1.2065269019221887, 'learning_rate/qwen_vl_interface': 9.541460879411214e-06, 'learning_rate/action_model': 9.541460879411215e-05, 'epoch': 0.32}) +[2026-06-17 08:09:11,982][starVLA.training.train_starvla][INFO] - Step 1074, Loss: {'train/loss': 1.0324314832687378, 'train/grad_norm_pre_clip': 11.40288257598877, 'timing/data': 0.005369399907067418, 'timing/model': 1.2076367810368538, 'learning_rate/qwen_vl_interface': 9.540533373692466e-06, 'learning_rate/action_model': 9.540533373692466e-05, 'epoch': 0.32}) +[2026-06-17 08:09:16,683][starVLA.training.train_starvla][INFO] - Step 1075, Loss: {'train/loss': 0.6434779167175293, 'train/grad_norm_pre_clip': 5.717074871063232, 'timing/data': 0.006418947014026344, 'timing/model': 1.2164162049302831, 'learning_rate/qwen_vl_interface': 9.539604978541531e-06, 'learning_rate/action_model': 9.539604978541531e-05, 'epoch': 0.32}) +[2026-06-17 08:09:21,454][starVLA.training.train_starvla][INFO] - Step 1076, Loss: {'train/loss': 0.8939999341964722, 'train/grad_norm_pre_clip': 9.140301704406738, 'timing/data': 0.005240485072135925, 'timing/model': 1.2031222860096022, 'learning_rate/qwen_vl_interface': 9.538675694150868e-06, 'learning_rate/action_model': 9.538675694150868e-05, 'epoch': 0.32}) +[2026-06-17 08:09:26,210][starVLA.training.train_starvla][INFO] - Step 1077, Loss: {'train/loss': 1.0754314661026, 'train/grad_norm_pre_clip': 9.060812950134277, 'timing/data': 0.005870572989806533, 'timing/model': 1.30831571505405, 'learning_rate/qwen_vl_interface': 9.53774552071312e-06, 'learning_rate/action_model': 9.537745520713119e-05, 'epoch': 0.32}) +[2026-06-17 08:09:30,868][starVLA.training.train_starvla][INFO] - Step 1078, Loss: {'train/loss': 1.2926576137542725, 'train/grad_norm_pre_clip': 13.352761268615723, 'timing/data': 0.005142882000654936, 'timing/model': 1.2036179599817842, 'learning_rate/qwen_vl_interface': 9.53681445842111e-06, 'learning_rate/action_model': 9.53681445842111e-05, 'epoch': 0.32}) +[2026-06-17 08:09:35,648][starVLA.training.train_starvla][INFO] - Step 1079, Loss: {'train/loss': 1.1520689725875854, 'train/grad_norm_pre_clip': 9.367873191833496, 'timing/data': 0.004685521009378135, 'timing/model': 1.2048369409749284, 'learning_rate/qwen_vl_interface': 9.53588250746785e-06, 'learning_rate/action_model': 9.53588250746785e-05, 'epoch': 0.32}) +[2026-06-17 08:09:40,501][starVLA.training.train_starvla][INFO] - Step 1080, Loss: {'train/loss': 0.7063279151916504, 'train/grad_norm_pre_clip': 8.519521713256836, 'timing/data': 0.004757323069497943, 'timing/model': 1.2082444900879636, 'learning_rate/qwen_vl_interface': 9.534949668046537e-06, 'learning_rate/action_model': 9.534949668046536e-05, 'epoch': 0.32}) +[2026-06-17 08:09:45,274][starVLA.training.train_starvla][INFO] - Step 1081, Loss: {'train/loss': 0.919906497001648, 'train/grad_norm_pre_clip': 9.218465805053711, 'timing/data': 0.005188666051253676, 'timing/model': 1.2013156920438632, 'learning_rate/qwen_vl_interface': 9.534015940350545e-06, 'learning_rate/action_model': 9.534015940350544e-05, 'epoch': 0.32}) +[2026-06-17 08:09:50,024][starVLA.training.train_starvla][INFO] - Step 1082, Loss: {'train/loss': 0.9866538643836975, 'train/grad_norm_pre_clip': 9.741193771362305, 'timing/data': 0.004636052995920181, 'timing/model': 1.2040345839923248, 'learning_rate/qwen_vl_interface': 9.533081324573437e-06, 'learning_rate/action_model': 9.533081324573437e-05, 'epoch': 0.32}) +[2026-06-17 08:09:54,816][starVLA.training.train_starvla][INFO] - Step 1083, Loss: {'train/loss': 0.9753152132034302, 'train/grad_norm_pre_clip': 10.069628715515137, 'timing/data': 0.004737229901365936, 'timing/model': 1.3376952939433977, 'learning_rate/qwen_vl_interface': 9.532145820908965e-06, 'learning_rate/action_model': 9.532145820908964e-05, 'epoch': 0.32}) +[2026-06-17 08:09:59,525][starVLA.training.train_starvla][INFO] - Step 1084, Loss: {'train/loss': 0.8498684763908386, 'train/grad_norm_pre_clip': 7.325879096984863, 'timing/data': 0.005048849969170988, 'timing/model': 1.2025863829767331, 'learning_rate/qwen_vl_interface': 9.531209429551056e-06, 'learning_rate/action_model': 9.531209429551054e-05, 'epoch': 0.32}) +[2026-06-17 08:10:04,317][starVLA.training.train_starvla][INFO] - Step 1085, Loss: {'train/loss': 0.6547591090202332, 'train/grad_norm_pre_clip': 18.681047439575195, 'timing/data': 0.005344095057807863, 'timing/model': 1.2070978910196573, 'learning_rate/qwen_vl_interface': 9.530272150693823e-06, 'learning_rate/action_model': 9.530272150693824e-05, 'epoch': 0.32}) +[2026-06-17 08:10:09,101][starVLA.training.train_starvla][INFO] - Step 1086, Loss: {'train/loss': 1.0475510358810425, 'train/grad_norm_pre_clip': 14.720580101013184, 'timing/data': 0.005443983944132924, 'timing/model': 1.2040071840165183, 'learning_rate/qwen_vl_interface': 9.52933398453157e-06, 'learning_rate/action_model': 9.52933398453157e-05, 'epoch': 0.32}) +[2026-06-17 08:10:13,882][starVLA.training.train_starvla][INFO] - Step 1087, Loss: {'train/loss': 1.1229149103164673, 'train/grad_norm_pre_clip': 9.703792572021484, 'timing/data': 0.005363308009691536, 'timing/model': 1.2167569530429319, 'learning_rate/qwen_vl_interface': 9.528394931258777e-06, 'learning_rate/action_model': 9.528394931258776e-05, 'epoch': 0.32}) +[2026-06-17 08:10:18,619][starVLA.training.train_starvla][INFO] - Step 1088, Loss: {'train/loss': 0.9783709645271301, 'train/grad_norm_pre_clip': 9.097798347473145, 'timing/data': 0.005456320941448212, 'timing/model': 1.2061871089972556, 'learning_rate/qwen_vl_interface': 9.527454991070108e-06, 'learning_rate/action_model': 9.527454991070108e-05, 'epoch': 0.32}) +[2026-06-17 08:10:23,388][starVLA.training.train_starvla][INFO] - Step 1089, Loss: {'train/loss': 0.9775267839431763, 'train/grad_norm_pre_clip': 10.426159858703613, 'timing/data': 0.0052581579657271504, 'timing/model': 1.3236471429700032, 'learning_rate/qwen_vl_interface': 9.52651416416042e-06, 'learning_rate/action_model': 9.526514164160421e-05, 'epoch': 0.32}) +[2026-06-17 08:10:28,032][starVLA.training.train_starvla][INFO] - Step 1090, Loss: {'train/loss': 0.6797711849212646, 'train/grad_norm_pre_clip': 6.691551208496094, 'timing/data': 0.0051902070408686996, 'timing/model': 1.2010305509902537, 'learning_rate/qwen_vl_interface': 9.525572450724743e-06, 'learning_rate/action_model': 9.525572450724743e-05, 'epoch': 0.32}) +[2026-06-17 08:10:32,813][starVLA.training.train_starvla][INFO] - Step 1091, Loss: {'train/loss': 0.900136411190033, 'train/grad_norm_pre_clip': 9.911319732666016, 'timing/data': 0.005239479010924697, 'timing/model': 1.2096346379257739, 'learning_rate/qwen_vl_interface': 9.524629850958297e-06, 'learning_rate/action_model': 9.524629850958297e-05, 'epoch': 0.32}) +[2026-06-17 08:10:37,605][starVLA.training.train_starvla][INFO] - Step 1092, Loss: {'train/loss': 0.9721840620040894, 'train/grad_norm_pre_clip': 9.189862251281738, 'timing/data': 0.005260123987682164, 'timing/model': 1.211343204951845, 'learning_rate/qwen_vl_interface': 9.523686365056482e-06, 'learning_rate/action_model': 9.523686365056481e-05, 'epoch': 0.32}) +[2026-06-17 08:10:42,405][starVLA.training.train_starvla][INFO] - Step 1093, Loss: {'train/loss': 0.9208115339279175, 'train/grad_norm_pre_clip': 11.50805377960205, 'timing/data': 0.0075068880105391145, 'timing/model': 1.2103189140325412, 'learning_rate/qwen_vl_interface': 9.522741993214886e-06, 'learning_rate/action_model': 9.522741993214886e-05, 'epoch': 0.32}) +[2026-06-17 08:10:47,202][starVLA.training.train_starvla][INFO] - Step 1094, Loss: {'train/loss': 0.7771639823913574, 'train/grad_norm_pre_clip': 6.9247565269470215, 'timing/data': 0.005211587995290756, 'timing/model': 1.3389110529096797, 'learning_rate/qwen_vl_interface': 9.521796735629275e-06, 'learning_rate/action_model': 9.521796735629275e-05, 'epoch': 0.32}) +[2026-06-17 08:10:51,859][starVLA.training.train_starvla][INFO] - Step 1095, Loss: {'train/loss': 0.7500707507133484, 'train/grad_norm_pre_clip': 12.483724594116211, 'timing/data': 0.00484857801347971, 'timing/model': 1.2044293370563537, 'learning_rate/qwen_vl_interface': 9.520850592495603e-06, 'learning_rate/action_model': 9.520850592495604e-05, 'epoch': 0.32}) +[2026-06-17 08:10:56,635][starVLA.training.train_starvla][INFO] - Step 1096, Loss: {'train/loss': 0.9022226929664612, 'train/grad_norm_pre_clip': 9.736770629882812, 'timing/data': 0.005035114008933306, 'timing/model': 1.2030924139544368, 'learning_rate/qwen_vl_interface': 9.519903564010012e-06, 'learning_rate/action_model': 9.51990356401001e-05, 'epoch': 0.32}) +[2026-06-17 08:11:01,335][starVLA.training.train_starvla][INFO] - Step 1097, Loss: {'train/loss': 1.1226491928100586, 'train/grad_norm_pre_clip': 10.220027923583984, 'timing/data': 0.00537652091588825, 'timing/model': 1.2039413179736584, 'learning_rate/qwen_vl_interface': 9.518955650368816e-06, 'learning_rate/action_model': 9.518955650368814e-05, 'epoch': 0.32}) +[2026-06-17 08:11:06,072][starVLA.training.train_starvla][INFO] - Step 1098, Loss: {'train/loss': 1.108967900276184, 'train/grad_norm_pre_clip': 31.201778411865234, 'timing/data': 0.005396721069701016, 'timing/model': 1.2032204640563577, 'learning_rate/qwen_vl_interface': 9.51800685176852e-06, 'learning_rate/action_model': 9.518006851768518e-05, 'epoch': 0.32}) +[2026-06-17 08:11:10,855][starVLA.training.train_starvla][INFO] - Step 1099, Loss: {'train/loss': 1.1220769882202148, 'train/grad_norm_pre_clip': 15.568624496459961, 'timing/data': 0.005106181022711098, 'timing/model': 1.204610064974986, 'learning_rate/qwen_vl_interface': 9.51705716840581e-06, 'learning_rate/action_model': 9.517057168405808e-05, 'epoch': 0.32}) +[2026-06-17 08:12:51,525][starVLA.training.train_starvla][INFO] - Step 1100, Loss: {'train/loss': 0.9690477848052979, 'train/grad_norm_pre_clip': 12.935277938842773, 'eval/loss': 0.9829543828964233, 'eval/latency_0/loss': 0.9829543828964233, 'eval/demon_fire_precision': 0.8365560014593214, 'eval/demon_fire_recall': 0.9522425249169435, 'eval/demon_fire_f1': 0.8906583802680131, 'eval/demon_fire_rate_model': 0.7622358175750834, 'eval/demon_fire_rate_teacher': 0.6696329254727474, 'eval/demon_left_precision': 0.670407315045719, 'eval/demon_left_recall': 0.9138810198300283, 'eval/demon_left_f1': 0.7734356269479741, 'eval/demon_left_rate_model': 0.6690767519466073, 'eval/demon_left_rate_teacher': 0.49082313681868744, 'eval/demon_right_precision': 0.8697478991596639, 'eval/demon_right_recall': 0.5847457627118644, 'eval/demon_right_f1': 0.6993243243243243, 'eval/demon_right_rate_model': 0.3309232480533927, 'eval/demon_right_rate_teacher': 0.4922135706340378, 'eval/demon_move_f1': 0.7363799756361492, 'eval/demon_cc_f1': 0.8379884827314852, 'eval/latency_0/demon_fire_precision': 0.8365560014593214, 'eval/latency_0/demon_fire_recall': 0.9522425249169435, 'eval/latency_0/demon_fire_f1': 0.8906583802680131, 'eval/latency_0/demon_fire_rate_model': 0.7622358175750834, 'eval/latency_0/demon_fire_rate_teacher': 0.6696329254727474, 'eval/latency_0/demon_left_precision': 0.670407315045719, 'eval/latency_0/demon_left_recall': 0.9138810198300283, 'eval/latency_0/demon_left_f1': 0.7734356269479741, 'eval/latency_0/demon_left_rate_model': 0.6690767519466073, 'eval/latency_0/demon_left_rate_teacher': 0.49082313681868744, 'eval/latency_0/demon_right_precision': 0.8697478991596639, 'eval/latency_0/demon_right_recall': 0.5847457627118644, 'eval/latency_0/demon_right_f1': 0.6993243243243243, 'eval/latency_0/demon_right_rate_model': 0.3309232480533927, 'eval/latency_0/demon_right_rate_teacher': 0.4922135706340378, 'eval/latency_0/demon_move_f1': 0.7363799756361492, 'eval/latency_0/demon_cc_f1': 0.8379884827314852, 'timing/data': 0.005638259928673506, 'timing/model': 1.3328785449266434, 'learning_rate/qwen_vl_interface': 9.516106600477557e-06, 'learning_rate/action_model': 9.516106600477557e-05, 'epoch': 0.32}) +[2026-06-17 08:12:51,526][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_1100_state +[2026-06-17 08:12:51,528][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 08:12:55,056][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_1100_state/model.safetensors +[2026-06-17 08:13:02,766][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_1100_state/optimizer.bin +[2026-06-17 08:13:02,766][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_1100_state/sampler.bin +[2026-06-17 08:13:02,766][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_1100_state/sampler_1.bin +[2026-06-17 08:13:02,767][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_1100_state/random_states_0.pkl +[2026-06-17 08:13:02,812][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1100 +[2026-06-17 08:13:02,812][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 08:13:02,822][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 08:13:07,424][starVLA.training.train_starvla][INFO] - Step 1101, Loss: {'train/loss': 1.165541172027588, 'train/grad_norm_pre_clip': 11.86465835571289, 'timing/data': 0.004447632934898138, 'timing/model': 1.1855520160170272, 'learning_rate/qwen_vl_interface': 9.515155148180816e-06, 'learning_rate/action_model': 9.515155148180816e-05, 'epoch': 0.32}) +[2026-06-17 08:13:12,029][starVLA.training.train_starvla][INFO] - Step 1102, Loss: {'train/loss': 1.1595009565353394, 'train/grad_norm_pre_clip': 14.819429397583008, 'timing/data': 0.004662569030188024, 'timing/model': 1.2064457720844075, 'learning_rate/qwen_vl_interface': 9.514202811712825e-06, 'learning_rate/action_model': 9.514202811712825e-05, 'epoch': 0.32}) +[2026-06-17 08:13:16,629][starVLA.training.train_starvla][INFO] - Step 1103, Loss: {'train/loss': 0.7894253730773926, 'train/grad_norm_pre_clip': 7.9840922355651855, 'timing/data': 0.004582724999636412, 'timing/model': 1.1907127499580383, 'learning_rate/qwen_vl_interface': 9.513249591271e-06, 'learning_rate/action_model': 9.513249591271e-05, 'epoch': 0.32}) +[2026-06-17 08:13:21,304][starVLA.training.train_starvla][INFO] - Step 1104, Loss: {'train/loss': 0.8608949184417725, 'train/grad_norm_pre_clip': 7.823958396911621, 'timing/data': 0.004890629090368748, 'timing/model': 1.1892814009916037, 'learning_rate/qwen_vl_interface': 9.512295487052949e-06, 'learning_rate/action_model': 9.512295487052949e-05, 'epoch': 0.32}) +[2026-06-17 08:13:26,019][starVLA.training.train_starvla][INFO] - Step 1105, Loss: {'train/loss': 0.9389159679412842, 'train/grad_norm_pre_clip': 8.648383140563965, 'timing/data': 0.005137074971571565, 'timing/model': 1.1948002299759537, 'learning_rate/qwen_vl_interface': 9.511340499256455e-06, 'learning_rate/action_model': 9.511340499256455e-05, 'epoch': 0.33}) +[2026-06-17 08:13:30,762][starVLA.training.train_starvla][INFO] - Step 1106, Loss: {'train/loss': 0.9744211435317993, 'train/grad_norm_pre_clip': 12.037409782409668, 'timing/data': 0.004775455920025706, 'timing/model': 1.2037378930253908, 'learning_rate/qwen_vl_interface': 9.510384628079493e-06, 'learning_rate/action_model': 9.510384628079493e-05, 'epoch': 0.33}) +[2026-06-17 08:13:35,582][starVLA.training.train_starvla][INFO] - Step 1107, Loss: {'train/loss': 0.9237269163131714, 'train/grad_norm_pre_clip': 13.080737113952637, 'timing/data': 0.004668619949370623, 'timing/model': 1.3268930630292743, 'learning_rate/qwen_vl_interface': 9.50942787372021e-06, 'learning_rate/action_model': 9.50942787372021e-05, 'epoch': 0.33}) +[2026-06-17 08:13:40,201][starVLA.training.train_starvla][INFO] - Step 1108, Loss: {'train/loss': 1.0799434185028076, 'train/grad_norm_pre_clip': 10.202742576599121, 'timing/data': 0.004751123953610659, 'timing/model': 1.197069383924827, 'learning_rate/qwen_vl_interface': 9.508470236376948e-06, 'learning_rate/action_model': 9.508470236376948e-05, 'epoch': 0.33}) +[2026-06-17 08:13:44,967][starVLA.training.train_starvla][INFO] - Step 1109, Loss: {'train/loss': 0.8228595852851868, 'train/grad_norm_pre_clip': 5.883849620819092, 'timing/data': 0.0048572010127827525, 'timing/model': 1.2019854380050674, 'learning_rate/qwen_vl_interface': 9.507511716248224e-06, 'learning_rate/action_model': 9.507511716248224e-05, 'epoch': 0.33}) +[2026-06-17 08:13:49,738][starVLA.training.train_starvla][INFO] - Step 1110, Loss: {'train/loss': 0.656416118144989, 'train/grad_norm_pre_clip': 20.061166763305664, 'timing/data': 0.00471300003118813, 'timing/model': 1.2011093040928245, 'learning_rate/qwen_vl_interface': 9.506552313532739e-06, 'learning_rate/action_model': 9.506552313532739e-05, 'epoch': 0.33}) +[2026-06-17 08:13:54,605][starVLA.training.train_starvla][INFO] - Step 1111, Loss: {'train/loss': 0.9860435724258423, 'train/grad_norm_pre_clip': 7.362353324890137, 'timing/data': 0.03969346103258431, 'timing/model': 1.2045441450318322, 'learning_rate/qwen_vl_interface': 9.505592028429377e-06, 'learning_rate/action_model': 9.505592028429377e-05, 'epoch': 0.33}) +[2026-06-17 08:13:59,258][starVLA.training.train_starvla][INFO] - Step 1112, Loss: {'train/loss': 0.8024318814277649, 'train/grad_norm_pre_clip': 6.334773063659668, 'timing/data': 0.004266818985342979, 'timing/model': 1.1999159309780225, 'learning_rate/qwen_vl_interface': 9.504630861137212e-06, 'learning_rate/action_model': 9.504630861137212e-05, 'epoch': 0.33}) +[2026-06-17 08:14:03,931][starVLA.training.train_starvla][INFO] - Step 1113, Loss: {'train/loss': 0.8203598856925964, 'train/grad_norm_pre_clip': 7.419096946716309, 'timing/data': 0.004666493972763419, 'timing/model': 1.2393835070542991, 'learning_rate/qwen_vl_interface': 9.503668811855489e-06, 'learning_rate/action_model': 9.503668811855488e-05, 'epoch': 0.33}) +[2026-06-17 08:14:08,564][starVLA.training.train_starvla][INFO] - Step 1114, Loss: {'train/loss': 0.9007577300071716, 'train/grad_norm_pre_clip': 9.266923904418945, 'timing/data': 0.005035897018387914, 'timing/model': 1.197483362047933, 'learning_rate/qwen_vl_interface': 9.502705880783644e-06, 'learning_rate/action_model': 9.502705880783645e-05, 'epoch': 0.33}) +[2026-06-17 08:14:13,273][starVLA.training.train_starvla][INFO] - Step 1115, Loss: {'train/loss': 0.793956458568573, 'train/grad_norm_pre_clip': 9.181248664855957, 'timing/data': 0.004671468981541693, 'timing/model': 1.1999457029160112, 'learning_rate/qwen_vl_interface': 9.501742068121296e-06, 'learning_rate/action_model': 9.501742068121296e-05, 'epoch': 0.33}) +[2026-06-17 08:14:18,035][starVLA.training.train_starvla][INFO] - Step 1116, Loss: {'train/loss': 0.8575252890586853, 'train/grad_norm_pre_clip': 8.792451858520508, 'timing/data': 0.004837785963900387, 'timing/model': 1.2014685600297526, 'learning_rate/qwen_vl_interface': 9.500777374068241e-06, 'learning_rate/action_model': 9.50077737406824e-05, 'epoch': 0.33}) +[2026-06-17 08:14:22,796][starVLA.training.train_starvla][INFO] - Step 1117, Loss: {'train/loss': 1.0117344856262207, 'train/grad_norm_pre_clip': 9.05624008178711, 'timing/data': 0.004500415991060436, 'timing/model': 1.1970128940884024, 'learning_rate/qwen_vl_interface': 9.49981179882446e-06, 'learning_rate/action_model': 9.499811798824461e-05, 'epoch': 0.33}) +[2026-06-17 08:14:27,564][starVLA.training.train_starvla][INFO] - Step 1118, Loss: {'train/loss': 0.9224879145622253, 'train/grad_norm_pre_clip': 8.944371223449707, 'timing/data': 0.004618211998604238, 'timing/model': 1.326809423044324, 'learning_rate/qwen_vl_interface': 9.498845342590124e-06, 'learning_rate/action_model': 9.498845342590124e-05, 'epoch': 0.33}) +[2026-06-17 08:14:32,322][starVLA.training.train_starvla][INFO] - Step 1119, Loss: {'train/loss': 0.8428018689155579, 'train/grad_norm_pre_clip': 7.995926380157471, 'timing/data': 0.004583650967106223, 'timing/model': 1.3192425060551614, 'learning_rate/qwen_vl_interface': 9.497878005565575e-06, 'learning_rate/action_model': 9.497878005565574e-05, 'epoch': 0.33}) +[2026-06-17 08:14:37,045][starVLA.training.train_starvla][INFO] - Step 1120, Loss: {'train/loss': 1.1330358982086182, 'train/grad_norm_pre_clip': 13.351624488830566, 'timing/data': 0.004825521027669311, 'timing/model': 1.1996710499515757, 'learning_rate/qwen_vl_interface': 9.496909787951344e-06, 'learning_rate/action_model': 9.496909787951343e-05, 'epoch': 0.33}) +[2026-06-17 08:14:41,764][starVLA.training.train_starvla][INFO] - Step 1121, Loss: {'train/loss': 0.7930750846862793, 'train/grad_norm_pre_clip': 9.312785148620605, 'timing/data': 0.004612250020727515, 'timing/model': 1.2015293720178306, 'learning_rate/qwen_vl_interface': 9.495940689948144e-06, 'learning_rate/action_model': 9.495940689948144e-05, 'epoch': 0.33}) +[2026-06-17 08:14:46,530][starVLA.training.train_starvla][INFO] - Step 1122, Loss: {'train/loss': 0.8890763521194458, 'train/grad_norm_pre_clip': 11.566658973693848, 'timing/data': 0.004862976027652621, 'timing/model': 1.2023071290459484, 'learning_rate/qwen_vl_interface': 9.49497071175687e-06, 'learning_rate/action_model': 9.49497071175687e-05, 'epoch': 0.33}) +[2026-06-17 08:14:51,298][starVLA.training.train_starvla][INFO] - Step 1123, Loss: {'train/loss': 0.8468039035797119, 'train/grad_norm_pre_clip': 60.76325607299805, 'timing/data': 0.005046568927355111, 'timing/model': 1.2006052159704268, 'learning_rate/qwen_vl_interface': 9.493999853578601e-06, 'learning_rate/action_model': 9.493999853578602e-05, 'epoch': 0.33}) +[2026-06-17 08:14:56,060][starVLA.training.train_starvla][INFO] - Step 1124, Loss: {'train/loss': 0.9767662882804871, 'train/grad_norm_pre_clip': 10.670652389526367, 'timing/data': 0.004723651916719973, 'timing/model': 1.3247723110252991, 'learning_rate/qwen_vl_interface': 9.493028115614595e-06, 'learning_rate/action_model': 9.493028115614595e-05, 'epoch': 0.33}) +[2026-06-17 08:15:00,832][starVLA.training.train_starvla][INFO] - Step 1125, Loss: {'train/loss': 0.9769139289855957, 'train/grad_norm_pre_clip': 14.121217727661133, 'timing/data': 0.004755898960866034, 'timing/model': 1.3288042260101065, 'learning_rate/qwen_vl_interface': 9.492055498066295e-06, 'learning_rate/action_model': 9.492055498066294e-05, 'epoch': 0.33}) +[2026-06-17 08:15:05,493][starVLA.training.train_starvla][INFO] - Step 1126, Loss: {'train/loss': 1.048080563545227, 'train/grad_norm_pre_clip': 10.408963203430176, 'timing/data': 0.004994258051738143, 'timing/model': 1.2039980289991945, 'learning_rate/qwen_vl_interface': 9.491082001135326e-06, 'learning_rate/action_model': 9.491082001135325e-05, 'epoch': 0.33}) +[2026-06-17 08:15:10,267][starVLA.training.train_starvla][INFO] - Step 1127, Loss: {'train/loss': 1.1154406070709229, 'train/grad_norm_pre_clip': 11.4984769821167, 'timing/data': 0.0044729020446538925, 'timing/model': 1.20843641506508, 'learning_rate/qwen_vl_interface': 9.490107625023492e-06, 'learning_rate/action_model': 9.490107625023492e-05, 'epoch': 0.33}) +[2026-06-17 08:15:15,067][starVLA.training.train_starvla][INFO] - Step 1128, Loss: {'train/loss': 0.7494543790817261, 'train/grad_norm_pre_clip': 9.850359916687012, 'timing/data': 0.004563419031910598, 'timing/model': 1.212045483989641, 'learning_rate/qwen_vl_interface': 9.489132369932788e-06, 'learning_rate/action_model': 9.489132369932788e-05, 'epoch': 0.33}) +[2026-06-17 08:15:19,880][starVLA.training.train_starvla][INFO] - Step 1129, Loss: {'train/loss': 0.9067574143409729, 'train/grad_norm_pre_clip': 10.11312484741211, 'timing/data': 0.004352809977717698, 'timing/model': 1.2057039020583034, 'learning_rate/qwen_vl_interface': 9.488156236065378e-06, 'learning_rate/action_model': 9.488156236065379e-05, 'epoch': 0.33}) +[2026-06-17 08:15:24,622][starVLA.training.train_starvla][INFO] - Step 1130, Loss: {'train/loss': 0.9936909079551697, 'train/grad_norm_pre_clip': 9.387397766113281, 'timing/data': 0.004455916001461446, 'timing/model': 1.2780269719660282, 'learning_rate/qwen_vl_interface': 9.487179223623626e-06, 'learning_rate/action_model': 9.487179223623624e-05, 'epoch': 0.33}) +[2026-06-17 08:15:29,358][starVLA.training.train_starvla][INFO] - Step 1131, Loss: {'train/loss': 1.09443998336792, 'train/grad_norm_pre_clip': 13.743480682373047, 'timing/data': 0.004649515962228179, 'timing/model': 1.2086146130459383, 'learning_rate/qwen_vl_interface': 9.486201332810058e-06, 'learning_rate/action_model': 9.486201332810057e-05, 'epoch': 0.33}) +[2026-06-17 08:15:34,145][starVLA.training.train_starvla][INFO] - Step 1132, Loss: {'train/loss': 1.0363436937332153, 'train/grad_norm_pre_clip': 11.782730102539062, 'timing/data': 0.004880373016931117, 'timing/model': 1.2064045300940052, 'learning_rate/qwen_vl_interface': 9.485222563827395e-06, 'learning_rate/action_model': 9.485222563827395e-05, 'epoch': 0.33}) +[2026-06-17 08:15:38,941][starVLA.training.train_starvla][INFO] - Step 1133, Loss: {'train/loss': 0.8433454632759094, 'train/grad_norm_pre_clip': 11.851438522338867, 'timing/data': 0.004415074945427477, 'timing/model': 1.2079379520146176, 'learning_rate/qwen_vl_interface': 9.484242916878537e-06, 'learning_rate/action_model': 9.484242916878536e-05, 'epoch': 0.33}) +[2026-06-17 08:15:43,736][starVLA.training.train_starvla][INFO] - Step 1134, Loss: {'train/loss': 0.9023475646972656, 'train/grad_norm_pre_clip': 9.658379554748535, 'timing/data': 0.0046733219642192125, 'timing/model': 1.2090775119140744, 'learning_rate/qwen_vl_interface': 9.483262392166567e-06, 'learning_rate/action_model': 9.483262392166566e-05, 'epoch': 0.33}) +[2026-06-17 08:15:48,537][starVLA.training.train_starvla][INFO] - Step 1135, Loss: {'train/loss': 0.807239830493927, 'train/grad_norm_pre_clip': 7.007627010345459, 'timing/data': 0.0047290639486163855, 'timing/model': 1.3398569759447128, 'learning_rate/qwen_vl_interface': 9.482280989894748e-06, 'learning_rate/action_model': 9.482280989894748e-05, 'epoch': 0.33}) +[2026-06-17 08:15:53,280][starVLA.training.train_starvla][INFO] - Step 1136, Loss: {'train/loss': 0.7558145523071289, 'train/grad_norm_pre_clip': 6.732998847961426, 'timing/data': 0.00483734009321779, 'timing/model': 1.2980554649839178, 'learning_rate/qwen_vl_interface': 9.481298710266525e-06, 'learning_rate/action_model': 9.481298710266526e-05, 'epoch': 0.33}) +[2026-06-17 08:15:58,064][starVLA.training.train_starvla][INFO] - Step 1137, Loss: {'train/loss': 0.7429853677749634, 'train/grad_norm_pre_clip': 7.901322364807129, 'timing/data': 0.005609381012618542, 'timing/model': 1.2105961930938065, 'learning_rate/qwen_vl_interface': 9.480315553485527e-06, 'learning_rate/action_model': 9.480315553485527e-05, 'epoch': 0.33}) +[2026-06-17 08:16:02,859][starVLA.training.train_starvla][INFO] - Step 1138, Loss: {'train/loss': 0.9177132248878479, 'train/grad_norm_pre_clip': 9.859017372131348, 'timing/data': 0.00450477097183466, 'timing/model': 1.2075893009314314, 'learning_rate/qwen_vl_interface': 9.479331519755563e-06, 'learning_rate/action_model': 9.479331519755562e-05, 'epoch': 0.33}) +[2026-06-17 08:16:07,639][starVLA.training.train_starvla][INFO] - Step 1139, Loss: {'train/loss': 0.8654071688652039, 'train/grad_norm_pre_clip': 10.489795684814453, 'timing/data': 0.004656538018025458, 'timing/model': 1.2034674290334806, 'learning_rate/qwen_vl_interface': 9.478346609280622e-06, 'learning_rate/action_model': 9.478346609280623e-05, 'epoch': 0.34}) +[2026-06-17 08:16:12,433][starVLA.training.train_starvla][INFO] - Step 1140, Loss: {'train/loss': 1.4348310232162476, 'train/grad_norm_pre_clip': 14.256525993347168, 'timing/data': 0.0043490249663591385, 'timing/model': 1.2085969629697502, 'learning_rate/qwen_vl_interface': 9.477360822264883e-06, 'learning_rate/action_model': 9.477360822264882e-05, 'epoch': 0.34}) +[2026-06-17 08:16:17,204][starVLA.training.train_starvla][INFO] - Step 1141, Loss: {'train/loss': 0.9216736555099487, 'train/grad_norm_pre_clip': 9.342148780822754, 'timing/data': 0.0057379319332540035, 'timing/model': 1.206657845992595, 'learning_rate/qwen_vl_interface': 9.476374158912692e-06, 'learning_rate/action_model': 9.476374158912691e-05, 'epoch': 0.34}) +[2026-06-17 08:16:21,953][starVLA.training.train_starvla][INFO] - Step 1142, Loss: {'train/loss': 0.72296541929245, 'train/grad_norm_pre_clip': 8.187848091125488, 'timing/data': 0.004994172020815313, 'timing/model': 1.3007902189856395, 'learning_rate/qwen_vl_interface': 9.475386619428593e-06, 'learning_rate/action_model': 9.475386619428594e-05, 'epoch': 0.34}) +[2026-06-17 08:16:26,609][starVLA.training.train_starvla][INFO] - Step 1143, Loss: {'train/loss': 0.6835915446281433, 'train/grad_norm_pre_clip': 6.430528163909912, 'timing/data': 0.004846740979701281, 'timing/model': 1.20246458100155, 'learning_rate/qwen_vl_interface': 9.4743982040173e-06, 'learning_rate/action_model': 9.4743982040173e-05, 'epoch': 0.34}) +[2026-06-17 08:16:31,399][starVLA.training.train_starvla][INFO] - Step 1144, Loss: {'train/loss': 0.87921541929245, 'train/grad_norm_pre_clip': 8.865228652954102, 'timing/data': 0.004703460028395057, 'timing/model': 1.2112114679766819, 'learning_rate/qwen_vl_interface': 9.473408912883714e-06, 'learning_rate/action_model': 9.473408912883714e-05, 'epoch': 0.34}) +[2026-06-17 08:16:36,194][starVLA.training.train_starvla][INFO] - Step 1145, Loss: {'train/loss': 0.7468594908714294, 'train/grad_norm_pre_clip': 8.674447059631348, 'timing/data': 0.0046210079453885555, 'timing/model': 1.2015117860864848, 'learning_rate/qwen_vl_interface': 9.472418746232916e-06, 'learning_rate/action_model': 9.472418746232916e-05, 'epoch': 0.34}) +[2026-06-17 08:16:40,966][starVLA.training.train_starvla][INFO] - Step 1146, Loss: {'train/loss': 1.1327201128005981, 'train/grad_norm_pre_clip': 7.508848667144775, 'timing/data': 0.004689588909968734, 'timing/model': 1.2027969140326604, 'learning_rate/qwen_vl_interface': 9.471427704270168e-06, 'learning_rate/action_model': 9.471427704270169e-05, 'epoch': 0.34}) +[2026-06-17 08:16:45,769][starVLA.training.train_starvla][INFO] - Step 1147, Loss: {'train/loss': 0.860697865486145, 'train/grad_norm_pre_clip': 6.7408223152160645, 'timing/data': 0.0045950779458507895, 'timing/model': 1.3317310730926692, 'learning_rate/qwen_vl_interface': 9.470435787200914e-06, 'learning_rate/action_model': 9.470435787200915e-05, 'epoch': 0.34}) +[2026-06-17 08:16:50,446][starVLA.training.train_starvla][INFO] - Step 1148, Loss: {'train/loss': 0.9392421841621399, 'train/grad_norm_pre_clip': 15.934206008911133, 'timing/data': 0.004588739946484566, 'timing/model': 1.2089103780454025, 'learning_rate/qwen_vl_interface': 9.469442995230781e-06, 'learning_rate/action_model': 9.469442995230782e-05, 'epoch': 0.34}) +[2026-06-17 08:16:55,254][starVLA.training.train_starvla][INFO] - Step 1149, Loss: {'train/loss': 0.9114898443222046, 'train/grad_norm_pre_clip': 10.606484413146973, 'timing/data': 0.0051708180690184236, 'timing/model': 1.2141318080248311, 'learning_rate/qwen_vl_interface': 9.468449328565574e-06, 'learning_rate/action_model': 9.468449328565573e-05, 'epoch': 0.34}) +[2026-06-17 08:17:00,059][starVLA.training.train_starvla][INFO] - Step 1150, Loss: {'train/loss': 0.7927941083908081, 'train/grad_norm_pre_clip': 8.034146308898926, 'timing/data': 0.00459427700843662, 'timing/model': 1.2121018529869616, 'learning_rate/qwen_vl_interface': 9.467454787411282e-06, 'learning_rate/action_model': 9.467454787411281e-05, 'epoch': 0.34}) +[2026-06-17 08:17:04,802][starVLA.training.train_starvla][INFO] - Step 1151, Loss: {'train/loss': 1.0949907302856445, 'train/grad_norm_pre_clip': 13.418330192565918, 'timing/data': 0.004801418981514871, 'timing/model': 1.2027263020863757, 'learning_rate/qwen_vl_interface': 9.466459371974076e-06, 'learning_rate/action_model': 9.466459371974076e-05, 'epoch': 0.34}) +[2026-06-17 08:17:09,568][starVLA.training.train_starvla][INFO] - Step 1152, Loss: {'train/loss': 0.6439962387084961, 'train/grad_norm_pre_clip': 6.6300177574157715, 'timing/data': 0.004872668068856001, 'timing/model': 1.2022641089279205, 'learning_rate/qwen_vl_interface': 9.465463082460304e-06, 'learning_rate/action_model': 9.465463082460304e-05, 'epoch': 0.34}) +[2026-06-17 08:17:14,345][starVLA.training.train_starvla][INFO] - Step 1153, Loss: {'train/loss': 0.9197781085968018, 'train/grad_norm_pre_clip': 10.839383125305176, 'timing/data': 0.0044232739601284266, 'timing/model': 1.3339295639889315, 'learning_rate/qwen_vl_interface': 9.4644659190765e-06, 'learning_rate/action_model': 9.464465919076499e-05, 'epoch': 0.34}) +[2026-06-17 08:17:18,988][starVLA.training.train_starvla][INFO] - Step 1154, Loss: {'train/loss': 1.1950565576553345, 'train/grad_norm_pre_clip': 8.983317375183105, 'timing/data': 0.004804380005225539, 'timing/model': 1.2014941809466109, 'learning_rate/qwen_vl_interface': 9.463467882029374e-06, 'learning_rate/action_model': 9.463467882029373e-05, 'epoch': 0.34}) +[2026-06-17 08:17:23,763][starVLA.training.train_starvla][INFO] - Step 1155, Loss: {'train/loss': 0.8494239449501038, 'train/grad_norm_pre_clip': 9.905147552490234, 'timing/data': 0.004795879009179771, 'timing/model': 1.2022886520717293, 'learning_rate/qwen_vl_interface': 9.462468971525822e-06, 'learning_rate/action_model': 9.462468971525821e-05, 'epoch': 0.34}) +[2026-06-17 08:17:28,541][starVLA.training.train_starvla][INFO] - Step 1156, Loss: {'train/loss': 1.125042200088501, 'train/grad_norm_pre_clip': 9.781058311462402, 'timing/data': 0.0046406949404627085, 'timing/model': 1.2020134970080107, 'learning_rate/qwen_vl_interface': 9.461469187772921e-06, 'learning_rate/action_model': 9.46146918777292e-05, 'epoch': 0.34}) +[2026-06-17 08:17:33,364][starVLA.training.train_starvla][INFO] - Step 1157, Loss: {'train/loss': 0.7418146133422852, 'train/grad_norm_pre_clip': 6.084537506103516, 'timing/data': 0.00469205598346889, 'timing/model': 1.2013958279276267, 'learning_rate/qwen_vl_interface': 9.460468530977924e-06, 'learning_rate/action_model': 9.460468530977924e-05, 'epoch': 0.34}) +[2026-06-17 08:17:38,116][starVLA.training.train_starvla][INFO] - Step 1158, Loss: {'train/loss': 0.8229897618293762, 'train/grad_norm_pre_clip': 6.7655930519104, 'timing/data': 0.004917077021673322, 'timing/model': 1.2026670459890738, 'learning_rate/qwen_vl_interface': 9.459467001348271e-06, 'learning_rate/action_model': 9.45946700134827e-05, 'epoch': 0.34}) +[2026-06-17 08:17:42,886][starVLA.training.train_starvla][INFO] - Step 1159, Loss: {'train/loss': 1.010398268699646, 'train/grad_norm_pre_clip': 9.065973281860352, 'timing/data': 0.004318471997976303, 'timing/model': 1.3191447160206735, 'learning_rate/qwen_vl_interface': 9.458464599091578e-06, 'learning_rate/action_model': 9.458464599091578e-05, 'epoch': 0.34}) +[2026-06-17 08:17:47,534][starVLA.training.train_starvla][INFO] - Step 1160, Loss: {'train/loss': 1.0934760570526123, 'train/grad_norm_pre_clip': 10.289576530456543, 'timing/data': 0.004700802965089679, 'timing/model': 1.2013181479414925, 'learning_rate/qwen_vl_interface': 9.457461324415647e-06, 'learning_rate/action_model': 9.457461324415646e-05, 'epoch': 0.34}) +[2026-06-17 08:17:52,312][starVLA.training.train_starvla][INFO] - Step 1161, Loss: {'train/loss': 1.1370964050292969, 'train/grad_norm_pre_clip': 9.891109466552734, 'timing/data': 0.004896068014204502, 'timing/model': 1.2021614930126816, 'learning_rate/qwen_vl_interface': 9.456457177528454e-06, 'learning_rate/action_model': 9.456457177528453e-05, 'epoch': 0.34}) +[2026-06-17 08:17:57,110][starVLA.training.train_starvla][INFO] - Step 1162, Loss: {'train/loss': 1.1707412004470825, 'train/grad_norm_pre_clip': 10.252957344055176, 'timing/data': 0.004772774060256779, 'timing/model': 1.205533048021607, 'learning_rate/qwen_vl_interface': 9.455452158638163e-06, 'learning_rate/action_model': 9.455452158638163e-05, 'epoch': 0.34}) +[2026-06-17 08:18:01,943][starVLA.training.train_starvla][INFO] - Step 1163, Loss: {'train/loss': 1.0586493015289307, 'train/grad_norm_pre_clip': 7.723910331726074, 'timing/data': 0.004980952944606543, 'timing/model': 1.2023845609510317, 'learning_rate/qwen_vl_interface': 9.454446267953114e-06, 'learning_rate/action_model': 9.454446267953114e-05, 'epoch': 0.34}) +[2026-06-17 08:18:06,739][starVLA.training.train_starvla][INFO] - Step 1164, Loss: {'train/loss': 0.7925726771354675, 'train/grad_norm_pre_clip': 6.141153812408447, 'timing/data': 0.004784028045833111, 'timing/model': 1.3439737521111965, 'learning_rate/qwen_vl_interface': 9.453439505681829e-06, 'learning_rate/action_model': 9.453439505681829e-05, 'epoch': 0.34}) +[2026-06-17 08:18:11,391][starVLA.training.train_starvla][INFO] - Step 1165, Loss: {'train/loss': 0.8549715280532837, 'train/grad_norm_pre_clip': 8.60356616973877, 'timing/data': 0.004547547898255289, 'timing/model': 1.2034205310046673, 'learning_rate/qwen_vl_interface': 9.452431872033012e-06, 'learning_rate/action_model': 9.452431872033012e-05, 'epoch': 0.34}) +[2026-06-17 08:18:16,190][starVLA.training.train_starvla][INFO] - Step 1166, Loss: {'train/loss': 0.9543432593345642, 'train/grad_norm_pre_clip': 7.783798694610596, 'timing/data': 0.004764414974488318, 'timing/model': 1.2043280539801344, 'learning_rate/qwen_vl_interface': 9.451423367215546e-06, 'learning_rate/action_model': 9.451423367215545e-05, 'epoch': 0.34}) +[2026-06-17 08:18:20,968][starVLA.training.train_starvla][INFO] - Step 1167, Loss: {'train/loss': 0.810336709022522, 'train/grad_norm_pre_clip': 7.848388671875, 'timing/data': 0.004697385942563415, 'timing/model': 1.2072817930020392, 'learning_rate/qwen_vl_interface': 9.450413991438494e-06, 'learning_rate/action_model': 9.450413991438494e-05, 'epoch': 0.34}) +[2026-06-17 08:18:25,697][starVLA.training.train_starvla][INFO] - Step 1168, Loss: {'train/loss': 0.8026206493377686, 'train/grad_norm_pre_clip': 20.893207550048828, 'timing/data': 0.004654709016904235, 'timing/model': 1.2036704620113596, 'learning_rate/qwen_vl_interface': 9.449403744911104e-06, 'learning_rate/action_model': 9.449403744911103e-05, 'epoch': 0.34}) +[2026-06-17 08:18:30,471][starVLA.training.train_starvla][INFO] - Step 1169, Loss: {'train/loss': 1.0942015647888184, 'train/grad_norm_pre_clip': 10.030155181884766, 'timing/data': 0.004598165978677571, 'timing/model': 1.202422096976079, 'learning_rate/qwen_vl_interface': 9.448392627842797e-06, 'learning_rate/action_model': 9.448392627842795e-05, 'epoch': 0.34}) +[2026-06-17 08:18:35,245][starVLA.training.train_starvla][INFO] - Step 1170, Loss: {'train/loss': 0.6647500991821289, 'train/grad_norm_pre_clip': 7.612040996551514, 'timing/data': 0.004744385019876063, 'timing/model': 1.334421989042312, 'learning_rate/qwen_vl_interface': 9.44738064044318e-06, 'learning_rate/action_model': 9.44738064044318e-05, 'epoch': 0.34}) +[2026-06-17 08:18:39,909][starVLA.training.train_starvla][INFO] - Step 1171, Loss: {'train/loss': 0.8535705208778381, 'train/grad_norm_pre_clip': 9.239901542663574, 'timing/data': 0.004757633083499968, 'timing/model': 1.218930192058906, 'learning_rate/qwen_vl_interface': 9.44636778292204e-06, 'learning_rate/action_model': 9.44636778292204e-05, 'epoch': 0.34}) +[2026-06-17 08:18:44,705][starVLA.training.train_starvla][INFO] - Step 1172, Loss: {'train/loss': 1.0067628622055054, 'train/grad_norm_pre_clip': 8.060955047607422, 'timing/data': 0.004765111021697521, 'timing/model': 1.2027061131084338, 'learning_rate/qwen_vl_interface': 9.445354055489345e-06, 'learning_rate/action_model': 9.445354055489345e-05, 'epoch': 0.34}) +[2026-06-17 08:18:49,485][starVLA.training.train_starvla][INFO] - Step 1173, Loss: {'train/loss': 0.7818604707717896, 'train/grad_norm_pre_clip': 7.383519649505615, 'timing/data': 0.004783221986144781, 'timing/model': 1.2099689550232142, 'learning_rate/qwen_vl_interface': 9.444339458355239e-06, 'learning_rate/action_model': 9.444339458355239e-05, 'epoch': 0.35}) +[2026-06-17 08:18:54,229][starVLA.training.train_starvla][INFO] - Step 1174, Loss: {'train/loss': 0.7850095629692078, 'train/grad_norm_pre_clip': 8.847525596618652, 'timing/data': 0.004244492040015757, 'timing/model': 1.2023531530285254, 'learning_rate/qwen_vl_interface': 9.443323991730049e-06, 'learning_rate/action_model': 9.443323991730049e-05, 'epoch': 0.35}) +[2026-06-17 08:18:59,042][starVLA.training.train_starvla][INFO] - Step 1175, Loss: {'train/loss': 1.1555848121643066, 'train/grad_norm_pre_clip': 9.731870651245117, 'timing/data': 0.004474111017771065, 'timing/model': 1.2028322330443189, 'learning_rate/qwen_vl_interface': 9.442307655824284e-06, 'learning_rate/action_model': 9.442307655824285e-05, 'epoch': 0.35}) +[2026-06-17 08:19:03,825][starVLA.training.train_starvla][INFO] - Step 1176, Loss: {'train/loss': 1.0442063808441162, 'train/grad_norm_pre_clip': 8.836078643798828, 'timing/data': 0.004808090976439416, 'timing/model': 1.3410998330218717, 'learning_rate/qwen_vl_interface': 9.441290450848631e-06, 'learning_rate/action_model': 9.44129045084863e-05, 'epoch': 0.35}) +[2026-06-17 08:19:08,479][starVLA.training.train_starvla][INFO] - Step 1177, Loss: {'train/loss': 0.879715621471405, 'train/grad_norm_pre_clip': 8.057705879211426, 'timing/data': 0.004890625015832484, 'timing/model': 1.2100192720536143, 'learning_rate/qwen_vl_interface': 9.440272377013957e-06, 'learning_rate/action_model': 9.440272377013956e-05, 'epoch': 0.35}) +[2026-06-17 08:19:13,264][starVLA.training.train_starvla][INFO] - Step 1178, Loss: {'train/loss': 0.8743340373039246, 'train/grad_norm_pre_clip': 7.592819690704346, 'timing/data': 0.004419111064635217, 'timing/model': 1.2032836190192029, 'learning_rate/qwen_vl_interface': 9.43925343453131e-06, 'learning_rate/action_model': 9.43925343453131e-05, 'epoch': 0.35}) +[2026-06-17 08:19:18,048][starVLA.training.train_starvla][INFO] - Step 1179, Loss: {'train/loss': 0.9028398990631104, 'train/grad_norm_pre_clip': 8.824085235595703, 'timing/data': 0.004646041081286967, 'timing/model': 1.2053642029641196, 'learning_rate/qwen_vl_interface': 9.438233623611918e-06, 'learning_rate/action_model': 9.438233623611918e-05, 'epoch': 0.35}) +[2026-06-17 08:19:22,819][starVLA.training.train_starvla][INFO] - Step 1180, Loss: {'train/loss': 0.860432505607605, 'train/grad_norm_pre_clip': 9.6591796875, 'timing/data': 0.004677935969084501, 'timing/model': 1.2011588480090722, 'learning_rate/qwen_vl_interface': 9.43721294446719e-06, 'learning_rate/action_model': 9.437212944467191e-05, 'epoch': 0.35}) +[2026-06-17 08:19:27,483][starVLA.training.train_starvla][INFO] - Step 1181, Loss: {'train/loss': 0.9891892075538635, 'train/grad_norm_pre_clip': 8.137123107910156, 'timing/data': 0.004770816070958972, 'timing/model': 1.2041453510755673, 'learning_rate/qwen_vl_interface': 9.436191397308713e-06, 'learning_rate/action_model': 9.436191397308713e-05, 'epoch': 0.35}) +[2026-06-17 08:19:32,167][starVLA.training.train_starvla][INFO] - Step 1182, Loss: {'train/loss': 0.6936003565788269, 'train/grad_norm_pre_clip': 7.448586940765381, 'timing/data': 0.004970551002770662, 'timing/model': 1.242166301002726, 'learning_rate/qwen_vl_interface': 9.435168982348254e-06, 'learning_rate/action_model': 9.435168982348255e-05, 'epoch': 0.35}) +[2026-06-17 08:19:36,841][starVLA.training.train_starvla][INFO] - Step 1183, Loss: {'train/loss': 0.7100011706352234, 'train/grad_norm_pre_clip': 6.911325931549072, 'timing/data': 0.0044785820646211505, 'timing/model': 1.2039881579112262, 'learning_rate/qwen_vl_interface': 9.434145699797763e-06, 'learning_rate/action_model': 9.434145699797764e-05, 'epoch': 0.35}) +[2026-06-17 08:19:41,567][starVLA.training.train_starvla][INFO] - Step 1184, Loss: {'train/loss': 0.9137929677963257, 'train/grad_norm_pre_clip': 7.754240036010742, 'timing/data': 0.0046545249642804265, 'timing/model': 1.2060169809265062, 'learning_rate/qwen_vl_interface': 9.433121549869365e-06, 'learning_rate/action_model': 9.433121549869365e-05, 'epoch': 0.35}) +[2026-06-17 08:19:46,353][starVLA.training.train_starvla][INFO] - Step 1185, Loss: {'train/loss': 0.9176836609840393, 'train/grad_norm_pre_clip': 18.3068904876709, 'timing/data': 0.004401911981403828, 'timing/model': 1.2058556930860505, 'learning_rate/qwen_vl_interface': 9.43209653277537e-06, 'learning_rate/action_model': 9.432096532775369e-05, 'epoch': 0.35}) +[2026-06-17 08:19:51,135][starVLA.training.train_starvla][INFO] - Step 1186, Loss: {'train/loss': 0.7863615155220032, 'train/grad_norm_pre_clip': 8.947171211242676, 'timing/data': 0.004658584948629141, 'timing/model': 1.2038334810640663, 'learning_rate/qwen_vl_interface': 9.431070648728262e-06, 'learning_rate/action_model': 9.431070648728261e-05, 'epoch': 0.35}) +[2026-06-17 08:19:55,914][starVLA.training.train_starvla][INFO] - Step 1187, Loss: {'train/loss': 0.8985128998756409, 'train/grad_norm_pre_clip': 8.455713272094727, 'timing/data': 0.0047254739329218864, 'timing/model': 1.3313069480936974, 'learning_rate/qwen_vl_interface': 9.430043897940709e-06, 'learning_rate/action_model': 9.43004389794071e-05, 'epoch': 0.35}) +[2026-06-17 08:20:00,695][starVLA.training.train_starvla][INFO] - Step 1188, Loss: {'train/loss': 0.7263996601104736, 'train/grad_norm_pre_clip': 6.687170505523682, 'timing/data': 0.004875660059042275, 'timing/model': 1.3355002340395004, 'learning_rate/qwen_vl_interface': 9.429016280625559e-06, 'learning_rate/action_model': 9.429016280625559e-05, 'epoch': 0.35}) +[2026-06-17 08:20:05,357][starVLA.training.train_starvla][INFO] - Step 1189, Loss: {'train/loss': 0.8476030826568604, 'train/grad_norm_pre_clip': 9.148029327392578, 'timing/data': 0.004591056960634887, 'timing/model': 1.2030008400324732, 'learning_rate/qwen_vl_interface': 9.427987796995838e-06, 'learning_rate/action_model': 9.427987796995837e-05, 'epoch': 0.35}) +[2026-06-17 08:20:10,103][starVLA.training.train_starvla][INFO] - Step 1190, Loss: {'train/loss': 0.9573552012443542, 'train/grad_norm_pre_clip': 8.66324520111084, 'timing/data': 0.0048584300093352795, 'timing/model': 1.2045919540105388, 'learning_rate/qwen_vl_interface': 9.42695844726475e-06, 'learning_rate/action_model': 9.426958447264749e-05, 'epoch': 0.35}) +[2026-06-17 08:20:14,874][starVLA.training.train_starvla][INFO] - Step 1191, Loss: {'train/loss': 0.8329781293869019, 'train/grad_norm_pre_clip': 8.061311721801758, 'timing/data': 0.00476980593521148, 'timing/model': 1.2036011710297316, 'learning_rate/qwen_vl_interface': 9.42592823164568e-06, 'learning_rate/action_model': 9.42592823164568e-05, 'epoch': 0.35}) +[2026-06-17 08:20:19,673][starVLA.training.train_starvla][INFO] - Step 1192, Loss: {'train/loss': 0.8887380957603455, 'train/grad_norm_pre_clip': 11.028116226196289, 'timing/data': 0.004822343005798757, 'timing/model': 1.2020742549793795, 'learning_rate/qwen_vl_interface': 9.424897150352197e-06, 'learning_rate/action_model': 9.424897150352197e-05, 'epoch': 0.35}) +[2026-06-17 08:20:24,388][starVLA.training.train_starvla][INFO] - Step 1193, Loss: {'train/loss': 1.1314246654510498, 'train/grad_norm_pre_clip': 10.033300399780273, 'timing/data': 0.004687701002694666, 'timing/model': 1.2709456098964438, 'learning_rate/qwen_vl_interface': 9.42386520359804e-06, 'learning_rate/action_model': 9.42386520359804e-05, 'epoch': 0.35}) +[2026-06-17 08:20:29,032][starVLA.training.train_starvla][INFO] - Step 1194, Loss: {'train/loss': 0.6646782159805298, 'train/grad_norm_pre_clip': 6.391173362731934, 'timing/data': 0.004480900941416621, 'timing/model': 1.2018666219664738, 'learning_rate/qwen_vl_interface': 9.422832391597136e-06, 'learning_rate/action_model': 9.422832391597136e-05, 'epoch': 0.35}) +[2026-06-17 08:20:33,787][starVLA.training.train_starvla][INFO] - Step 1195, Loss: {'train/loss': 0.8032928705215454, 'train/grad_norm_pre_clip': 8.594389915466309, 'timing/data': 0.00505019398406148, 'timing/model': 1.2048037649365142, 'learning_rate/qwen_vl_interface': 9.421798714563586e-06, 'learning_rate/action_model': 9.421798714563586e-05, 'epoch': 0.35}) +[2026-06-17 08:20:38,609][starVLA.training.train_starvla][INFO] - Step 1196, Loss: {'train/loss': 0.9520636200904846, 'train/grad_norm_pre_clip': 9.49036979675293, 'timing/data': 0.004948551068082452, 'timing/model': 1.2034626479726285, 'learning_rate/qwen_vl_interface': 9.420764172711676e-06, 'learning_rate/action_model': 9.420764172711676e-05, 'epoch': 0.35}) +[2026-06-17 08:20:43,377][starVLA.training.train_starvla][INFO] - Step 1197, Loss: {'train/loss': 0.8222878575325012, 'train/grad_norm_pre_clip': 7.479653358459473, 'timing/data': 0.0047213450307026505, 'timing/model': 1.202332000946626, 'learning_rate/qwen_vl_interface': 9.419728766255863e-06, 'learning_rate/action_model': 9.419728766255862e-05, 'epoch': 0.35}) +[2026-06-17 08:20:48,165][starVLA.training.train_starvla][INFO] - Step 1198, Loss: {'train/loss': 0.7225059270858765, 'train/grad_norm_pre_clip': 31.878780364990234, 'timing/data': 0.004759407020173967, 'timing/model': 1.3311491420026869, 'learning_rate/qwen_vl_interface': 9.418692495410789e-06, 'learning_rate/action_model': 9.418692495410789e-05, 'epoch': 0.35}) +[2026-06-17 08:20:52,895][starVLA.training.train_starvla][INFO] - Step 1199, Loss: {'train/loss': 0.9188195466995239, 'train/grad_norm_pre_clip': 8.292407989501953, 'timing/data': 0.004986146930605173, 'timing/model': 1.28505010495428, 'learning_rate/qwen_vl_interface': 9.417655360391277e-06, 'learning_rate/action_model': 9.417655360391277e-05, 'epoch': 0.35}) +[2026-06-17 08:22:34,636][starVLA.training.train_starvla][INFO] - Step 1200, Loss: {'train/loss': 0.9482487440109253, 'train/grad_norm_pre_clip': 9.332783699035645, 'eval/loss': 0.911517322063446, 'eval/latency_0/loss': 0.911517322063446, 'eval/demon_fire_precision': 0.8936708860759494, 'eval/demon_fire_recall': 0.8795681063122923, 'eval/demon_fire_f1': 0.8865634156550858, 'eval/demon_fire_rate_model': 0.6590656284760845, 'eval/demon_fire_rate_teacher': 0.6696329254727474, 'eval/demon_left_precision': 0.754758020663404, 'eval/demon_left_recall': 0.7864022662889518, 'eval/demon_left_f1': 0.7702552719200888, 'eval/demon_left_rate_model': 0.5114015572858732, 'eval/demon_left_rate_teacher': 0.49082313681868744, 'eval/demon_right_precision': 0.7706317586795675, 'eval/demon_right_recall': 0.7649717514124293, 'eval/demon_right_f1': 0.7677913240714489, 'eval/demon_right_rate_model': 0.48859844271412683, 'eval/demon_right_rate_teacher': 0.4922135706340378, 'eval/demon_move_f1': 0.7690232979957689, 'eval/demon_cc_f1': 0.8476940876163104, 'eval/latency_0/demon_fire_precision': 0.8936708860759494, 'eval/latency_0/demon_fire_recall': 0.8795681063122923, 'eval/latency_0/demon_fire_f1': 0.8865634156550858, 'eval/latency_0/demon_fire_rate_model': 0.6590656284760845, 'eval/latency_0/demon_fire_rate_teacher': 0.6696329254727474, 'eval/latency_0/demon_left_precision': 0.754758020663404, 'eval/latency_0/demon_left_recall': 0.7864022662889518, 'eval/latency_0/demon_left_f1': 0.7702552719200888, 'eval/latency_0/demon_left_rate_model': 0.5114015572858732, 'eval/latency_0/demon_left_rate_teacher': 0.49082313681868744, 'eval/latency_0/demon_right_precision': 0.7706317586795675, 'eval/latency_0/demon_right_recall': 0.7649717514124293, 'eval/latency_0/demon_right_f1': 0.7677913240714489, 'eval/latency_0/demon_right_rate_model': 0.48859844271412683, 'eval/latency_0/demon_right_rate_teacher': 0.4922135706340378, 'eval/latency_0/demon_move_f1': 0.7690232979957689, 'eval/latency_0/demon_cc_f1': 0.8476940876163104, 'timing/data': 0.004570438992232084, 'timing/model': 1.2047697020461783, 'learning_rate/qwen_vl_interface': 9.416617361412325e-06, 'learning_rate/action_model': 9.416617361412325e-05, 'epoch': 0.35}) +[2026-06-17 08:22:34,637][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_1200_state +[2026-06-17 08:22:34,640][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 08:22:38,227][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_1200_state/model.safetensors +[2026-06-17 08:22:45,948][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_1200_state/optimizer.bin +[2026-06-17 08:22:45,949][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_1200_state/sampler.bin +[2026-06-17 08:22:45,949][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_1200_state/sampler_1.bin +[2026-06-17 08:22:45,950][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_1200_state/random_states_0.pkl +[2026-06-17 08:22:45,994][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1200 +[2026-06-17 08:22:45,994][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 08:22:46,010][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 08:22:50,611][starVLA.training.train_starvla][INFO] - Step 1201, Loss: {'train/loss': 0.8996754288673401, 'train/grad_norm_pre_clip': 7.561862468719482, 'timing/data': 0.004504291107878089, 'timing/model': 1.2031146750086918, 'learning_rate/qwen_vl_interface': 9.415578498689106e-06, 'learning_rate/action_model': 9.415578498689106e-05, 'epoch': 0.35}) +[2026-06-17 08:22:55,210][starVLA.training.train_starvla][INFO] - Step 1202, Loss: {'train/loss': 1.1653969287872314, 'train/grad_norm_pre_clip': 9.369002342224121, 'timing/data': 0.00436472799628973, 'timing/model': 1.192438785918057, 'learning_rate/qwen_vl_interface': 9.414538772436987e-06, 'learning_rate/action_model': 9.414538772436986e-05, 'epoch': 0.35}) +[2026-06-17 08:22:59,831][starVLA.training.train_starvla][INFO] - Step 1203, Loss: {'train/loss': 0.78847736120224, 'train/grad_norm_pre_clip': 11.449546813964844, 'timing/data': 0.004298073938116431, 'timing/model': 1.194150059018284, 'learning_rate/qwen_vl_interface': 9.413498182871496e-06, 'learning_rate/action_model': 9.413498182871496e-05, 'epoch': 0.35}) +[2026-06-17 08:23:04,499][starVLA.training.train_starvla][INFO] - Step 1204, Loss: {'train/loss': 0.8727244734764099, 'train/grad_norm_pre_clip': 8.75002670288086, 'timing/data': 0.004865022958256304, 'timing/model': 1.1950302150798962, 'learning_rate/qwen_vl_interface': 9.412456730208352e-06, 'learning_rate/action_model': 9.412456730208352e-05, 'epoch': 0.35}) +[2026-06-17 08:23:09,247][starVLA.training.train_starvla][INFO] - Step 1205, Loss: {'train/loss': 0.7173594236373901, 'train/grad_norm_pre_clip': 5.181254863739014, 'timing/data': 0.004525435972027481, 'timing/model': 1.1960665209917352, 'learning_rate/qwen_vl_interface': 9.41141441466345e-06, 'learning_rate/action_model': 9.41141441466345e-05, 'epoch': 0.35}) +[2026-06-17 08:23:13,970][starVLA.training.train_starvla][INFO] - Step 1206, Loss: {'train/loss': 1.0408194065093994, 'train/grad_norm_pre_clip': 9.983460426330566, 'timing/data': 0.0048542749136686325, 'timing/model': 1.2956617729505524, 'learning_rate/qwen_vl_interface': 9.41037123645286e-06, 'learning_rate/action_model': 9.41037123645286e-05, 'epoch': 0.35}) +[2026-06-17 08:23:18,603][starVLA.training.train_starvla][INFO] - Step 1207, Loss: {'train/loss': 0.7424327731132507, 'train/grad_norm_pre_clip': 6.279033660888672, 'timing/data': 0.004655890981666744, 'timing/model': 1.1979037020355463, 'learning_rate/qwen_vl_interface': 9.409327195792838e-06, 'learning_rate/action_model': 9.409327195792838e-05, 'epoch': 0.36}) +[2026-06-17 08:23:23,357][starVLA.training.train_starvla][INFO] - Step 1208, Loss: {'train/loss': 0.7834571599960327, 'train/grad_norm_pre_clip': 8.736838340759277, 'timing/data': 0.004699136014096439, 'timing/model': 1.1983764740871266, 'learning_rate/qwen_vl_interface': 9.408282292899812e-06, 'learning_rate/action_model': 9.408282292899812e-05, 'epoch': 0.36}) +[2026-06-17 08:23:28,129][starVLA.training.train_starvla][INFO] - Step 1209, Loss: {'train/loss': 0.8260480761528015, 'train/grad_norm_pre_clip': 8.406845092773438, 'timing/data': 0.006180458003655076, 'timing/model': 1.2100374990841374, 'learning_rate/qwen_vl_interface': 9.407236527990394e-06, 'learning_rate/action_model': 9.407236527990393e-05, 'epoch': 0.36}) +[2026-06-17 08:23:32,900][starVLA.training.train_starvla][INFO] - Step 1210, Loss: {'train/loss': 0.5041655898094177, 'train/grad_norm_pre_clip': 7.316661834716797, 'timing/data': 0.004712780937552452, 'timing/model': 1.2001276890514418, 'learning_rate/qwen_vl_interface': 9.406189901281369e-06, 'learning_rate/action_model': 9.406189901281368e-05, 'epoch': 0.36}) +[2026-06-17 08:23:37,612][starVLA.training.train_starvla][INFO] - Step 1211, Loss: {'train/loss': 0.976995587348938, 'train/grad_norm_pre_clip': 17.572004318237305, 'timing/data': 0.004333821008913219, 'timing/model': 1.2004232879262418, 'learning_rate/qwen_vl_interface': 9.405142412989705e-06, 'learning_rate/action_model': 9.405142412989704e-05, 'epoch': 0.36}) +[2026-06-17 08:23:42,388][starVLA.training.train_starvla][INFO] - Step 1212, Loss: {'train/loss': 0.9388346672058105, 'train/grad_norm_pre_clip': 10.154831886291504, 'timing/data': 0.004653775016777217, 'timing/model': 1.3366531849605963, 'learning_rate/qwen_vl_interface': 9.404094063332546e-06, 'learning_rate/action_model': 9.404094063332546e-05, 'epoch': 0.36}) +[2026-06-17 08:23:47,035][starVLA.training.train_starvla][INFO] - Step 1213, Loss: {'train/loss': 0.9240666627883911, 'train/grad_norm_pre_clip': 9.555978775024414, 'timing/data': 0.004912156029604375, 'timing/model': 1.2055738109629601, 'learning_rate/qwen_vl_interface': 9.40304485252722e-06, 'learning_rate/action_model': 9.403044852527219e-05, 'epoch': 0.36}) +[2026-06-17 08:23:51,808][starVLA.training.train_starvla][INFO] - Step 1214, Loss: {'train/loss': 0.9806661009788513, 'train/grad_norm_pre_clip': 9.678674697875977, 'timing/data': 0.006161539000459015, 'timing/model': 1.2018910550978035, 'learning_rate/qwen_vl_interface': 9.401994780791226e-06, 'learning_rate/action_model': 9.401994780791226e-05, 'epoch': 0.36}) +[2026-06-17 08:23:56,582][starVLA.training.train_starvla][INFO] - Step 1215, Loss: {'train/loss': 0.6921575665473938, 'train/grad_norm_pre_clip': 7.538130760192871, 'timing/data': 0.004780529998242855, 'timing/model': 1.2058204320492223, 'learning_rate/qwen_vl_interface': 9.400943848342246e-06, 'learning_rate/action_model': 9.400943848342246e-05, 'epoch': 0.36}) +[2026-06-17 08:24:01,538][starVLA.training.train_starvla][INFO] - Step 1216, Loss: {'train/loss': 0.948630690574646, 'train/grad_norm_pre_clip': 9.104022026062012, 'timing/data': 0.004547141958028078, 'timing/model': 1.2021271040430292, 'learning_rate/qwen_vl_interface': 9.39989205539814e-06, 'learning_rate/action_model': 9.39989205539814e-05, 'epoch': 0.36}) +[2026-06-17 08:24:06,277][starVLA.training.train_starvla][INFO] - Step 1217, Loss: {'train/loss': 0.8713560700416565, 'train/grad_norm_pre_clip': 7.822469711303711, 'timing/data': 0.004512743907980621, 'timing/model': 1.2020871239947155, 'learning_rate/qwen_vl_interface': 9.398839402176946e-06, 'learning_rate/action_model': 9.398839402176946e-05, 'epoch': 0.36}) +[2026-06-17 08:24:11,036][starVLA.training.train_starvla][INFO] - Step 1218, Loss: {'train/loss': 0.9975835084915161, 'train/grad_norm_pre_clip': 8.757081985473633, 'timing/data': 0.004866115981712937, 'timing/model': 1.3172684960300103, 'learning_rate/qwen_vl_interface': 9.39778588889688e-06, 'learning_rate/action_model': 9.39778588889688e-05, 'epoch': 0.36}) +[2026-06-17 08:24:15,700][starVLA.training.train_starvla][INFO] - Step 1219, Loss: {'train/loss': 0.7955357432365417, 'train/grad_norm_pre_clip': 9.525827407836914, 'timing/data': 0.004611060954630375, 'timing/model': 1.2033349899575114, 'learning_rate/qwen_vl_interface': 9.396731515776335e-06, 'learning_rate/action_model': 9.396731515776333e-05, 'epoch': 0.36}) +[2026-06-17 08:24:20,527][starVLA.training.train_starvla][INFO] - Step 1220, Loss: {'train/loss': 0.8934628367424011, 'train/grad_norm_pre_clip': 9.58547306060791, 'timing/data': 0.004816691041924059, 'timing/model': 1.24916333402507, 'learning_rate/qwen_vl_interface': 9.395676283033884e-06, 'learning_rate/action_model': 9.395676283033885e-05, 'epoch': 0.36}) +[2026-06-17 08:24:25,319][starVLA.training.train_starvla][INFO] - Step 1221, Loss: {'train/loss': 0.8413947224617004, 'train/grad_norm_pre_clip': 10.527735710144043, 'timing/data': 0.004376586060971022, 'timing/model': 1.2022479879669845, 'learning_rate/qwen_vl_interface': 9.39462019088828e-06, 'learning_rate/action_model': 9.394620190888279e-05, 'epoch': 0.36}) +[2026-06-17 08:24:30,089][starVLA.training.train_starvla][INFO] - Step 1222, Loss: {'train/loss': 0.8436533212661743, 'train/grad_norm_pre_clip': 9.995904922485352, 'timing/data': 0.004619157058186829, 'timing/model': 1.201399418991059, 'learning_rate/qwen_vl_interface': 9.393563239558449e-06, 'learning_rate/action_model': 9.39356323955845e-05, 'epoch': 0.36}) +[2026-06-17 08:24:34,867][starVLA.training.train_starvla][INFO] - Step 1223, Loss: {'train/loss': 0.8577730655670166, 'train/grad_norm_pre_clip': 9.925904273986816, 'timing/data': 0.004639706923626363, 'timing/model': 1.3314438989618793, 'learning_rate/qwen_vl_interface': 9.392505429263501e-06, 'learning_rate/action_model': 9.392505429263501e-05, 'epoch': 0.36}) +[2026-06-17 08:24:39,517][starVLA.training.train_starvla][INFO] - Step 1224, Loss: {'train/loss': 0.934949517250061, 'train/grad_norm_pre_clip': 10.380717277526855, 'timing/data': 0.004591827979311347, 'timing/model': 1.2033603410236537, 'learning_rate/qwen_vl_interface': 9.39144676022272e-06, 'learning_rate/action_model': 9.39144676022272e-05, 'epoch': 0.36}) +[2026-06-17 08:24:44,301][starVLA.training.train_starvla][INFO] - Step 1225, Loss: {'train/loss': 0.568960428237915, 'train/grad_norm_pre_clip': 6.9441633224487305, 'timing/data': 0.004596379934810102, 'timing/model': 1.2054766661021858, 'learning_rate/qwen_vl_interface': 9.390387232655568e-06, 'learning_rate/action_model': 9.390387232655569e-05, 'epoch': 0.36}) +[2026-06-17 08:24:49,017][starVLA.training.train_starvla][INFO] - Step 1226, Loss: {'train/loss': 0.7992091178894043, 'train/grad_norm_pre_clip': 11.83968448638916, 'timing/data': 0.004897261969745159, 'timing/model': 1.2070410810410976, 'learning_rate/qwen_vl_interface': 9.38932684678169e-06, 'learning_rate/action_model': 9.38932684678169e-05, 'epoch': 0.36}) +[2026-06-17 08:24:53,766][starVLA.training.train_starvla][INFO] - Step 1227, Loss: {'train/loss': 0.8875751495361328, 'train/grad_norm_pre_clip': 10.769323348999023, 'timing/data': 0.00472726603038609, 'timing/model': 1.2108935349388048, 'learning_rate/qwen_vl_interface': 9.388265602820901e-06, 'learning_rate/action_model': 9.3882656028209e-05, 'epoch': 0.36}) +[2026-06-17 08:24:58,564][starVLA.training.train_starvla][INFO] - Step 1228, Loss: {'train/loss': 1.177256464958191, 'train/grad_norm_pre_clip': 12.639580726623535, 'timing/data': 0.004807260003872216, 'timing/model': 1.2057686829939485, 'learning_rate/qwen_vl_interface': 9.387203500993199e-06, 'learning_rate/action_model': 9.387203500993199e-05, 'epoch': 0.36}) +[2026-06-17 08:25:03,375][starVLA.training.train_starvla][INFO] - Step 1229, Loss: {'train/loss': 0.7765541672706604, 'train/grad_norm_pre_clip': 8.484502792358398, 'timing/data': 0.004760817973874509, 'timing/model': 1.355405590031296, 'learning_rate/qwen_vl_interface': 9.38614054151876e-06, 'learning_rate/action_model': 9.38614054151876e-05, 'epoch': 0.36}) +[2026-06-17 08:25:08,024][starVLA.training.train_starvla][INFO] - Step 1230, Loss: {'train/loss': 1.000882863998413, 'train/grad_norm_pre_clip': 10.57760238647461, 'timing/data': 0.004761252086609602, 'timing/model': 1.2036207140190527, 'learning_rate/qwen_vl_interface': 9.385076724617937e-06, 'learning_rate/action_model': 9.385076724617935e-05, 'epoch': 0.36}) +[2026-06-17 08:25:12,818][starVLA.training.train_starvla][INFO] - Step 1231, Loss: {'train/loss': 0.5328170657157898, 'train/grad_norm_pre_clip': 7.684252738952637, 'timing/data': 0.004851393052376807, 'timing/model': 1.2128379719797522, 'learning_rate/qwen_vl_interface': 9.384012050511258e-06, 'learning_rate/action_model': 9.384012050511258e-05, 'epoch': 0.36}) +[2026-06-17 08:25:17,591][starVLA.training.train_starvla][INFO] - Step 1232, Loss: {'train/loss': 0.8069592714309692, 'train/grad_norm_pre_clip': 14.473949432373047, 'timing/data': 0.0046222429955378175, 'timing/model': 1.2078488538973033, 'learning_rate/qwen_vl_interface': 9.382946519419433e-06, 'learning_rate/action_model': 9.382946519419434e-05, 'epoch': 0.36}) +[2026-06-17 08:25:22,344][starVLA.training.train_starvla][INFO] - Step 1233, Loss: {'train/loss': 0.9164468050003052, 'train/grad_norm_pre_clip': 12.418066024780273, 'timing/data': 0.004946023924276233, 'timing/model': 1.2114133180584759, 'learning_rate/qwen_vl_interface': 9.381880131563347e-06, 'learning_rate/action_model': 9.381880131563346e-05, 'epoch': 0.36}) +[2026-06-17 08:25:27,114][starVLA.training.train_starvla][INFO] - Step 1234, Loss: {'train/loss': 0.7989845871925354, 'train/grad_norm_pre_clip': 9.989327430725098, 'timing/data': 0.004736855975352228, 'timing/model': 1.2046653589932248, 'learning_rate/qwen_vl_interface': 9.380812887164065e-06, 'learning_rate/action_model': 9.380812887164065e-05, 'epoch': 0.36}) +[2026-06-17 08:25:31,920][starVLA.training.train_starvla][INFO] - Step 1235, Loss: {'train/loss': 0.9574431777000427, 'train/grad_norm_pre_clip': 12.517566680908203, 'timing/data': 0.0047359910095110536, 'timing/model': 1.352439956041053, 'learning_rate/qwen_vl_interface': 9.379744786442824e-06, 'learning_rate/action_model': 9.379744786442824e-05, 'epoch': 0.36}) +[2026-06-17 08:25:36,575][starVLA.training.train_starvla][INFO] - Step 1236, Loss: {'train/loss': 0.8756799697875977, 'train/grad_norm_pre_clip': 9.658105850219727, 'timing/data': 0.005208197049796581, 'timing/model': 1.2061870790785179, 'learning_rate/qwen_vl_interface': 9.378675829621047e-06, 'learning_rate/action_model': 9.378675829621046e-05, 'epoch': 0.36}) +[2026-06-17 08:25:41,364][starVLA.training.train_starvla][INFO] - Step 1237, Loss: {'train/loss': 0.9140231013298035, 'train/grad_norm_pre_clip': 11.021543502807617, 'timing/data': 0.004234532010741532, 'timing/model': 1.2046007349854335, 'learning_rate/qwen_vl_interface': 9.377606016920328e-06, 'learning_rate/action_model': 9.377606016920327e-05, 'epoch': 0.36}) +[2026-06-17 08:25:46,157][starVLA.training.train_starvla][INFO] - Step 1238, Loss: {'train/loss': 0.780919075012207, 'train/grad_norm_pre_clip': 10.079191207885742, 'timing/data': 0.004817088949494064, 'timing/model': 1.2117784039583057, 'learning_rate/qwen_vl_interface': 9.376535348562438e-06, 'learning_rate/action_model': 9.376535348562439e-05, 'epoch': 0.36}) +[2026-06-17 08:25:50,962][starVLA.training.train_starvla][INFO] - Step 1239, Loss: {'train/loss': 0.9229990243911743, 'train/grad_norm_pre_clip': 9.379274368286133, 'timing/data': 0.004786501987837255, 'timing/model': 1.2061276669846848, 'learning_rate/qwen_vl_interface': 9.375463824769332e-06, 'learning_rate/action_model': 9.375463824769333e-05, 'epoch': 0.36}) +[2026-06-17 08:25:55,691][starVLA.training.train_starvla][INFO] - Step 1240, Loss: {'train/loss': 1.3750464916229248, 'train/grad_norm_pre_clip': 14.944921493530273, 'timing/data': 0.004915772005915642, 'timing/model': 1.20998573792167, 'learning_rate/qwen_vl_interface': 9.374391445763135e-06, 'learning_rate/action_model': 9.374391445763135e-05, 'epoch': 0.36}) +[2026-06-17 08:26:00,405][starVLA.training.train_starvla][INFO] - Step 1241, Loss: {'train/loss': 0.7893213033676147, 'train/grad_norm_pre_clip': 7.904428482055664, 'timing/data': 0.004853322985582054, 'timing/model': 1.2472852519713342, 'learning_rate/qwen_vl_interface': 9.373318211766155e-06, 'learning_rate/action_model': 9.373318211766155e-05, 'epoch': 0.37}) +[2026-06-17 08:26:05,061][starVLA.training.train_starvla][INFO] - Step 1242, Loss: {'train/loss': 0.8407156467437744, 'train/grad_norm_pre_clip': 9.124781608581543, 'timing/data': 0.004812908009625971, 'timing/model': 1.205625390051864, 'learning_rate/qwen_vl_interface': 9.37224412300087e-06, 'learning_rate/action_model': 9.37224412300087e-05, 'epoch': 0.37}) +[2026-06-17 08:26:09,829][starVLA.training.train_starvla][INFO] - Step 1243, Loss: {'train/loss': 0.9166271686553955, 'train/grad_norm_pre_clip': 12.53681468963623, 'timing/data': 0.004621997009962797, 'timing/model': 1.2040976979769766, 'learning_rate/qwen_vl_interface': 9.371169179689945e-06, 'learning_rate/action_model': 9.371169179689945e-05, 'epoch': 0.37}) +[2026-06-17 08:26:14,626][starVLA.training.train_starvla][INFO] - Step 1244, Loss: {'train/loss': 1.4083117246627808, 'train/grad_norm_pre_clip': 14.36304759979248, 'timing/data': 0.00455760897602886, 'timing/model': 1.2056066319346428, 'learning_rate/qwen_vl_interface': 9.370093382056213e-06, 'learning_rate/action_model': 9.370093382056214e-05, 'epoch': 0.37}) +[2026-06-17 08:26:19,540][starVLA.training.train_starvla][INFO] - Step 1245, Loss: {'train/loss': 0.7897324562072754, 'train/grad_norm_pre_clip': 10.526262283325195, 'timing/data': 0.004633471951819956, 'timing/model': 1.2034587670350447, 'learning_rate/qwen_vl_interface': 9.369016730322692e-06, 'learning_rate/action_model': 9.36901673032269e-05, 'epoch': 0.37}) +[2026-06-17 08:26:24,336][starVLA.training.train_starvla][INFO] - Step 1246, Loss: {'train/loss': 0.9127963781356812, 'train/grad_norm_pre_clip': 9.818412780761719, 'timing/data': 0.004683054983615875, 'timing/model': 1.3370136950397864, 'learning_rate/qwen_vl_interface': 9.367939224712566e-06, 'learning_rate/action_model': 9.367939224712566e-05, 'epoch': 0.37}) +[2026-06-17 08:26:29,123][starVLA.training.train_starvla][INFO] - Step 1247, Loss: {'train/loss': 1.0232868194580078, 'train/grad_norm_pre_clip': 11.179644584655762, 'timing/data': 0.004630317911505699, 'timing/model': 1.3277236289577559, 'learning_rate/qwen_vl_interface': 9.366860865449213e-06, 'learning_rate/action_model': 9.366860865449212e-05, 'epoch': 0.37}) +[2026-06-17 08:26:33,790][starVLA.training.train_starvla][INFO] - Step 1248, Loss: {'train/loss': 0.9785984754562378, 'train/grad_norm_pre_clip': 13.030591011047363, 'timing/data': 0.004779929993674159, 'timing/model': 1.2135678400518373, 'learning_rate/qwen_vl_interface': 9.365781652756168e-06, 'learning_rate/action_model': 9.365781652756168e-05, 'epoch': 0.37}) +[2026-06-17 08:26:38,544][starVLA.training.train_starvla][INFO] - Step 1249, Loss: {'train/loss': 0.8832350969314575, 'train/grad_norm_pre_clip': 9.81020450592041, 'timing/data': 0.004907623981125653, 'timing/model': 1.210915329051204, 'learning_rate/qwen_vl_interface': 9.36470158685716e-06, 'learning_rate/action_model': 9.36470158685716e-05, 'epoch': 0.37}) +[2026-06-17 08:26:43,360][starVLA.training.train_starvla][INFO] - Step 1250, Loss: {'train/loss': 0.7492677569389343, 'train/grad_norm_pre_clip': 8.57206916809082, 'timing/data': 0.0049211010336875916, 'timing/model': 1.2114129449473694, 'learning_rate/qwen_vl_interface': 9.363620667976085e-06, 'learning_rate/action_model': 9.363620667976085e-05, 'epoch': 0.37}) +[2026-06-17 08:26:48,152][starVLA.training.train_starvla][INFO] - Step 1251, Loss: {'train/loss': 0.8382765054702759, 'train/grad_norm_pre_clip': 6.760122776031494, 'timing/data': 0.0044867569813504815, 'timing/model': 1.2124699030537158, 'learning_rate/qwen_vl_interface': 9.362538896337019e-06, 'learning_rate/action_model': 9.362538896337019e-05, 'epoch': 0.37}) +[2026-06-17 08:26:52,976][starVLA.training.train_starvla][INFO] - Step 1252, Loss: {'train/loss': 0.889690101146698, 'train/grad_norm_pre_clip': 10.679581642150879, 'timing/data': 0.004616005928255618, 'timing/model': 1.3409450160106644, 'learning_rate/qwen_vl_interface': 9.361456272164215e-06, 'learning_rate/action_model': 9.361456272164214e-05, 'epoch': 0.37}) +[2026-06-17 08:26:57,777][starVLA.training.train_starvla][INFO] - Step 1253, Loss: {'train/loss': 0.7698169946670532, 'train/grad_norm_pre_clip': 7.837069034576416, 'timing/data': 0.004892785917036235, 'timing/model': 1.3397006059531122, 'learning_rate/qwen_vl_interface': 9.3603727956821e-06, 'learning_rate/action_model': 9.3603727956821e-05, 'epoch': 0.37}) +[2026-06-17 08:27:02,477][starVLA.training.train_starvla][INFO] - Step 1254, Loss: {'train/loss': 0.6682186126708984, 'train/grad_norm_pre_clip': 7.385738849639893, 'timing/data': 0.008468273910693824, 'timing/model': 1.2165942239807919, 'learning_rate/qwen_vl_interface': 9.359288467115283e-06, 'learning_rate/action_model': 9.359288467115283e-05, 'epoch': 0.37}) +[2026-06-17 08:27:07,256][starVLA.training.train_starvla][INFO] - Step 1255, Loss: {'train/loss': 0.9648562669754028, 'train/grad_norm_pre_clip': 12.102046012878418, 'timing/data': 0.004550634068436921, 'timing/model': 1.2043672429863364, 'learning_rate/qwen_vl_interface': 9.358203286688544e-06, 'learning_rate/action_model': 9.358203286688544e-05, 'epoch': 0.37}) +[2026-06-17 08:27:12,044][starVLA.training.train_starvla][INFO] - Step 1256, Loss: {'train/loss': 0.7694900631904602, 'train/grad_norm_pre_clip': 9.078274726867676, 'timing/data': 0.004587356001138687, 'timing/model': 1.2046638029860333, 'learning_rate/qwen_vl_interface': 9.357117254626844e-06, 'learning_rate/action_model': 9.357117254626844e-05, 'epoch': 0.37}) +[2026-06-17 08:27:16,839][starVLA.training.train_starvla][INFO] - Step 1257, Loss: {'train/loss': 1.223514199256897, 'train/grad_norm_pre_clip': 21.380762100219727, 'timing/data': 0.004736406030133367, 'timing/model': 1.2128736090380698, 'learning_rate/qwen_vl_interface': 9.356030371155316e-06, 'learning_rate/action_model': 9.356030371155316e-05, 'epoch': 0.37}) +[2026-06-17 08:27:21,577][starVLA.training.train_starvla][INFO] - Step 1258, Loss: {'train/loss': 1.124456763267517, 'train/grad_norm_pre_clip': 11.813483238220215, 'timing/data': 0.004970213049091399, 'timing/model': 1.2841528380522504, 'learning_rate/qwen_vl_interface': 9.354942636499275e-06, 'learning_rate/action_model': 9.354942636499276e-05, 'epoch': 0.37}) +[2026-06-17 08:27:26,261][starVLA.training.train_starvla][INFO] - Step 1259, Loss: {'train/loss': 0.7653518915176392, 'train/grad_norm_pre_clip': 9.103557586669922, 'timing/data': 0.004674560972489417, 'timing/model': 1.2208475150400773, 'learning_rate/qwen_vl_interface': 9.35385405088421e-06, 'learning_rate/action_model': 9.353854050884208e-05, 'epoch': 0.37}) +[2026-06-17 08:27:31,050][starVLA.training.train_starvla][INFO] - Step 1260, Loss: {'train/loss': 0.8486657738685608, 'train/grad_norm_pre_clip': 12.740606307983398, 'timing/data': 0.004693957976996899, 'timing/model': 1.2134177510160953, 'learning_rate/qwen_vl_interface': 9.352764614535782e-06, 'learning_rate/action_model': 9.352764614535781e-05, 'epoch': 0.37}) +[2026-06-17 08:27:35,846][starVLA.training.train_starvla][INFO] - Step 1261, Loss: {'train/loss': 0.7999085783958435, 'train/grad_norm_pre_clip': 9.29708194732666, 'timing/data': 0.004401154932565987, 'timing/model': 1.2129771040054038, 'learning_rate/qwen_vl_interface': 9.351674327679837e-06, 'learning_rate/action_model': 9.351674327679836e-05, 'epoch': 0.37}) +[2026-06-17 08:27:40,632][starVLA.training.train_starvla][INFO] - Step 1262, Loss: {'train/loss': 0.8209154009819031, 'train/grad_norm_pre_clip': 10.720166206359863, 'timing/data': 0.004694006056524813, 'timing/model': 1.206308642984368, 'learning_rate/qwen_vl_interface': 9.350583190542387e-06, 'learning_rate/action_model': 9.350583190542387e-05, 'epoch': 0.37}) +[2026-06-17 08:27:45,441][starVLA.training.train_starvla][INFO] - Step 1263, Loss: {'train/loss': 0.9039195775985718, 'train/grad_norm_pre_clip': 9.1694974899292, 'timing/data': 0.00485120399389416, 'timing/model': 1.3513709240360186, 'learning_rate/qwen_vl_interface': 9.34949120334963e-06, 'learning_rate/action_model': 9.34949120334963e-05, 'epoch': 0.37}) +[2026-06-17 08:27:50,189][starVLA.training.train_starvla][INFO] - Step 1264, Loss: {'train/loss': 0.7832817435264587, 'train/grad_norm_pre_clip': 10.572479248046875, 'timing/data': 0.005305301980115473, 'timing/model': 1.2864085820037872, 'learning_rate/qwen_vl_interface': 9.348398366327938e-06, 'learning_rate/action_model': 9.348398366327937e-05, 'epoch': 0.37}) +[2026-06-17 08:27:54,870][starVLA.training.train_starvla][INFO] - Step 1265, Loss: {'train/loss': 0.7185201048851013, 'train/grad_norm_pre_clip': 12.680058479309082, 'timing/data': 0.004413502058014274, 'timing/model': 1.2102900869213045, 'learning_rate/qwen_vl_interface': 9.347304679703852e-06, 'learning_rate/action_model': 9.347304679703851e-05, 'epoch': 0.37}) +[2026-06-17 08:27:59,739][starVLA.training.train_starvla][INFO] - Step 1266, Loss: {'train/loss': 0.996636152267456, 'train/grad_norm_pre_clip': 16.271251678466797, 'timing/data': 0.004730546963401139, 'timing/model': 1.2097772329580039, 'learning_rate/qwen_vl_interface': 9.346210143704099e-06, 'learning_rate/action_model': 9.346210143704098e-05, 'epoch': 0.37}) +[2026-06-17 08:28:04,523][starVLA.training.train_starvla][INFO] - Step 1267, Loss: {'train/loss': 0.775081217288971, 'train/grad_norm_pre_clip': 9.750371932983398, 'timing/data': 0.004275875980965793, 'timing/model': 1.2043671219144017, 'learning_rate/qwen_vl_interface': 9.345114758555573e-06, 'learning_rate/action_model': 9.345114758555573e-05, 'epoch': 0.37}) +[2026-06-17 08:28:09,307][starVLA.training.train_starvla][INFO] - Step 1268, Loss: {'train/loss': 1.0014032125473022, 'train/grad_norm_pre_clip': 8.373562812805176, 'timing/data': 0.004493726999498904, 'timing/model': 1.2085578750120476, 'learning_rate/qwen_vl_interface': 9.344018524485352e-06, 'learning_rate/action_model': 9.344018524485351e-05, 'epoch': 0.37}) +[2026-06-17 08:28:14,102][starVLA.training.train_starvla][INFO] - Step 1269, Loss: {'train/loss': 0.8364620804786682, 'train/grad_norm_pre_clip': 48.395118713378906, 'timing/data': 0.004868333926424384, 'timing/model': 1.3425201830687001, 'learning_rate/qwen_vl_interface': 9.342921441720686e-06, 'learning_rate/action_model': 9.342921441720685e-05, 'epoch': 0.37}) +[2026-06-17 08:28:18,777][starVLA.training.train_starvla][INFO] - Step 1270, Loss: {'train/loss': 0.76485276222229, 'train/grad_norm_pre_clip': 10.990727424621582, 'timing/data': 0.004788386053405702, 'timing/model': 1.2048001089133322, 'learning_rate/qwen_vl_interface': 9.341823510489e-06, 'learning_rate/action_model': 9.341823510489e-05, 'epoch': 0.37}) +[2026-06-17 08:28:23,567][starVLA.training.train_starvla][INFO] - Step 1271, Loss: {'train/loss': 0.7812209725379944, 'train/grad_norm_pre_clip': 8.277430534362793, 'timing/data': 0.004725178005173802, 'timing/model': 1.206053560017608, 'learning_rate/qwen_vl_interface': 9.340724731017898e-06, 'learning_rate/action_model': 9.340724731017898e-05, 'epoch': 0.37}) +[2026-06-17 08:28:28,358][starVLA.training.train_starvla][INFO] - Step 1272, Loss: {'train/loss': 0.7289570569992065, 'train/grad_norm_pre_clip': 11.275383949279785, 'timing/data': 0.00470267899800092, 'timing/model': 1.2078218889655545, 'learning_rate/qwen_vl_interface': 9.339625103535158e-06, 'learning_rate/action_model': 9.339625103535156e-05, 'epoch': 0.37}) +[2026-06-17 08:28:33,075][starVLA.training.train_starvla][INFO] - Step 1273, Loss: {'train/loss': 0.5816858410835266, 'train/grad_norm_pre_clip': 6.720918655395508, 'timing/data': 0.0045902159763500094, 'timing/model': 1.212529449025169, 'learning_rate/qwen_vl_interface': 9.338524628268734e-06, 'learning_rate/action_model': 9.338524628268733e-05, 'epoch': 0.37}) +[2026-06-17 08:28:37,883][starVLA.training.train_starvla][INFO] - Step 1274, Loss: {'train/loss': 0.7945799231529236, 'train/grad_norm_pre_clip': 8.03006649017334, 'timing/data': 0.0044555290369316936, 'timing/model': 1.2112245380412787, 'learning_rate/qwen_vl_interface': 9.337423305446753e-06, 'learning_rate/action_model': 9.337423305446753e-05, 'epoch': 0.37}) +[2026-06-17 08:28:42,672][starVLA.training.train_starvla][INFO] - Step 1275, Loss: {'train/loss': 0.9658074975013733, 'train/grad_norm_pre_clip': 9.930291175842285, 'timing/data': 0.006155588082037866, 'timing/model': 1.3279777260031551, 'learning_rate/qwen_vl_interface': 9.336321135297525e-06, 'learning_rate/action_model': 9.336321135297525e-05, 'epoch': 0.38}) +[2026-06-17 08:28:47,345][starVLA.training.train_starvla][INFO] - Step 1276, Loss: {'train/loss': 0.7898690104484558, 'train/grad_norm_pre_clip': 8.418980598449707, 'timing/data': 0.004762010998092592, 'timing/model': 1.2044725259765983, 'learning_rate/qwen_vl_interface': 9.335218118049526e-06, 'learning_rate/action_model': 9.335218118049526e-05, 'epoch': 0.38}) +[2026-06-17 08:28:52,127][starVLA.training.train_starvla][INFO] - Step 1277, Loss: {'train/loss': 0.9247899055480957, 'train/grad_norm_pre_clip': 18.47431755065918, 'timing/data': 0.004511367063969374, 'timing/model': 1.2046805509598926, 'learning_rate/qwen_vl_interface': 9.334114253931417e-06, 'learning_rate/action_model': 9.334114253931416e-05, 'epoch': 0.38}) +[2026-06-17 08:28:56,948][starVLA.training.train_starvla][INFO] - Step 1278, Loss: {'train/loss': 1.2944025993347168, 'train/grad_norm_pre_clip': 10.410025596618652, 'timing/data': 0.004820537054911256, 'timing/model': 1.2104922629659995, 'learning_rate/qwen_vl_interface': 9.333009543172027e-06, 'learning_rate/action_model': 9.333009543172027e-05, 'epoch': 0.38}) +[2026-06-17 08:29:01,748][starVLA.training.train_starvla][INFO] - Step 1279, Loss: {'train/loss': 0.8626329898834229, 'train/grad_norm_pre_clip': 7.839193344116211, 'timing/data': 0.004706113948486745, 'timing/model': 1.2027334780432284, 'learning_rate/qwen_vl_interface': 9.331903986000365e-06, 'learning_rate/action_model': 9.331903986000365e-05, 'epoch': 0.38}) +[2026-06-17 08:29:06,500][starVLA.training.train_starvla][INFO] - Step 1280, Loss: {'train/loss': 0.7572228908538818, 'train/grad_norm_pre_clip': 10.593632698059082, 'timing/data': 0.004690115922130644, 'timing/model': 1.2158964509144425, 'learning_rate/qwen_vl_interface': 9.330797582645615e-06, 'learning_rate/action_model': 9.330797582645615e-05, 'epoch': 0.38}) +[2026-06-17 08:29:11,294][starVLA.training.train_starvla][INFO] - Step 1281, Loss: {'train/loss': 1.1495248079299927, 'train/grad_norm_pre_clip': 10.466364860534668, 'timing/data': 0.004808939062058926, 'timing/model': 1.338145897956565, 'learning_rate/qwen_vl_interface': 9.329690333337137e-06, 'learning_rate/action_model': 9.329690333337137e-05, 'epoch': 0.38}) +[2026-06-17 08:29:15,973][starVLA.training.train_starvla][INFO] - Step 1282, Loss: {'train/loss': 0.8976510167121887, 'train/grad_norm_pre_clip': 9.080101013183594, 'timing/data': 0.004659586003981531, 'timing/model': 1.2133260068949312, 'learning_rate/qwen_vl_interface': 9.328582238304461e-06, 'learning_rate/action_model': 9.328582238304461e-05, 'epoch': 0.38}) +[2026-06-17 08:29:20,781][starVLA.training.train_starvla][INFO] - Step 1283, Loss: {'train/loss': 1.1005586385726929, 'train/grad_norm_pre_clip': 10.461629867553711, 'timing/data': 0.004594363970682025, 'timing/model': 1.211622679955326, 'learning_rate/qwen_vl_interface': 9.327473297777297e-06, 'learning_rate/action_model': 9.327473297777297e-05, 'epoch': 0.38}) +[2026-06-17 08:29:25,565][starVLA.training.train_starvla][INFO] - Step 1284, Loss: {'train/loss': 1.0264946222305298, 'train/grad_norm_pre_clip': 9.41364860534668, 'timing/data': 0.004555898020043969, 'timing/model': 1.2070425170240924, 'learning_rate/qwen_vl_interface': 9.326363511985533e-06, 'learning_rate/action_model': 9.326363511985533e-05, 'epoch': 0.38}) +[2026-06-17 08:29:30,337][starVLA.training.train_starvla][INFO] - Step 1285, Loss: {'train/loss': 1.1273268461227417, 'train/grad_norm_pre_clip': 8.20495319366455, 'timing/data': 0.004757734015583992, 'timing/model': 1.2033765870146453, 'learning_rate/qwen_vl_interface': 9.325252881159226e-06, 'learning_rate/action_model': 9.325252881159226e-05, 'epoch': 0.38}) +[2026-06-17 08:29:35,084][starVLA.training.train_starvla][INFO] - Step 1286, Loss: {'train/loss': 0.9859142899513245, 'train/grad_norm_pre_clip': 10.661738395690918, 'timing/data': 0.004598714993335307, 'timing/model': 1.2044173079775646, 'learning_rate/qwen_vl_interface': 9.324141405528612e-06, 'learning_rate/action_model': 9.324141405528612e-05, 'epoch': 0.38}) +[2026-06-17 08:29:39,854][starVLA.training.train_starvla][INFO] - Step 1287, Loss: {'train/loss': 1.0734214782714844, 'train/grad_norm_pre_clip': 12.101140975952148, 'timing/data': 0.004728520987555385, 'timing/model': 1.3218905769754201, 'learning_rate/qwen_vl_interface': 9.3230290853241e-06, 'learning_rate/action_model': 9.3230290853241e-05, 'epoch': 0.38}) +[2026-06-17 08:29:44,503][starVLA.training.train_starvla][INFO] - Step 1288, Loss: {'train/loss': 0.845668613910675, 'train/grad_norm_pre_clip': 6.036715984344482, 'timing/data': 0.004551023943349719, 'timing/model': 1.2039778709877282, 'learning_rate/qwen_vl_interface': 9.321915920776278e-06, 'learning_rate/action_model': 9.321915920776278e-05, 'epoch': 0.38}) +[2026-06-17 08:29:49,288][starVLA.training.train_starvla][INFO] - Step 1289, Loss: {'train/loss': 0.9712274670600891, 'train/grad_norm_pre_clip': 11.595321655273438, 'timing/data': 0.004645603941753507, 'timing/model': 1.204359823022969, 'learning_rate/qwen_vl_interface': 9.320801912115901e-06, 'learning_rate/action_model': 9.320801912115901e-05, 'epoch': 0.38}) +[2026-06-17 08:29:54,072][starVLA.training.train_starvla][INFO] - Step 1290, Loss: {'train/loss': 1.2090200185775757, 'train/grad_norm_pre_clip': 8.698238372802734, 'timing/data': 0.005277757067233324, 'timing/model': 1.2031420600833371, 'learning_rate/qwen_vl_interface': 9.319687059573909e-06, 'learning_rate/action_model': 9.319687059573909e-05, 'epoch': 0.38}) +[2026-06-17 08:29:58,864][starVLA.training.train_starvla][INFO] - Step 1291, Loss: {'train/loss': 0.9396181106567383, 'train/grad_norm_pre_clip': 7.272407054901123, 'timing/data': 0.004859647015109658, 'timing/model': 1.2116832960164174, 'learning_rate/qwen_vl_interface': 9.318571363381408e-06, 'learning_rate/action_model': 9.318571363381408e-05, 'epoch': 0.38}) +[2026-06-17 08:30:03,687][starVLA.training.train_starvla][INFO] - Step 1292, Loss: {'train/loss': 0.7302477359771729, 'train/grad_norm_pre_clip': 7.110558032989502, 'timing/data': 0.004832052974961698, 'timing/model': 1.3461750880815089, 'learning_rate/qwen_vl_interface': 9.317454823769686e-06, 'learning_rate/action_model': 9.317454823769687e-05, 'epoch': 0.38}) +[2026-06-17 08:30:08,358][starVLA.training.train_starvla][INFO] - Step 1293, Loss: {'train/loss': 1.0278810262680054, 'train/grad_norm_pre_clip': 9.089920997619629, 'timing/data': 0.004741844022646546, 'timing/model': 1.2100169560872018, 'learning_rate/qwen_vl_interface': 9.316337440970206e-06, 'learning_rate/action_model': 9.316337440970206e-05, 'epoch': 0.38}) +[2026-06-17 08:30:13,158][starVLA.training.train_starvla][INFO] - Step 1294, Loss: {'train/loss': 0.7747297286987305, 'train/grad_norm_pre_clip': 6.940401077270508, 'timing/data': 0.004786013043485582, 'timing/model': 1.2183534710202366, 'learning_rate/qwen_vl_interface': 9.315219215214596e-06, 'learning_rate/action_model': 9.315219215214595e-05, 'epoch': 0.38}) +[2026-06-17 08:30:17,877][starVLA.training.train_starvla][INFO] - Step 1295, Loss: {'train/loss': 0.7269406318664551, 'train/grad_norm_pre_clip': 10.895781517028809, 'timing/data': 0.004745911923237145, 'timing/model': 1.2160465660272166, 'learning_rate/qwen_vl_interface': 9.314100146734668e-06, 'learning_rate/action_model': 9.314100146734668e-05, 'epoch': 0.38}) +[2026-06-17 08:30:22,632][starVLA.training.train_starvla][INFO] - Step 1296, Loss: {'train/loss': 0.7835217118263245, 'train/grad_norm_pre_clip': 6.983101844787598, 'timing/data': 0.006431299960240722, 'timing/model': 1.2198650669306517, 'learning_rate/qwen_vl_interface': 9.312980235762408e-06, 'learning_rate/action_model': 9.312980235762408e-05, 'epoch': 0.38}) +[2026-06-17 08:30:27,415][starVLA.training.train_starvla][INFO] - Step 1297, Loss: {'train/loss': 0.8346816897392273, 'train/grad_norm_pre_clip': 38.56009292602539, 'timing/data': 0.004782820004038513, 'timing/model': 1.2056098569883034, 'learning_rate/qwen_vl_interface': 9.311859482529971e-06, 'learning_rate/action_model': 9.31185948252997e-05, 'epoch': 0.38}) +[2026-06-17 08:30:32,199][starVLA.training.train_starvla][INFO] - Step 1298, Loss: {'train/loss': 0.8305119276046753, 'train/grad_norm_pre_clip': 7.7393717765808105, 'timing/data': 0.004757447983138263, 'timing/model': 1.3335154550150037, 'learning_rate/qwen_vl_interface': 9.310737887269694e-06, 'learning_rate/action_model': 9.310737887269694e-05, 'epoch': 0.38}) +[2026-06-17 08:30:36,849][starVLA.training.train_starvla][INFO] - Step 1299, Loss: {'train/loss': 0.9191803336143494, 'train/grad_norm_pre_clip': 7.911438941955566, 'timing/data': 0.004854871076531708, 'timing/model': 1.2044742880389094, 'learning_rate/qwen_vl_interface': 9.309615450214084e-06, 'learning_rate/action_model': 9.309615450214084e-05, 'epoch': 0.38}) +[2026-06-17 08:32:18,784][starVLA.training.train_starvla][INFO] - Step 1300, Loss: {'train/loss': 1.013613224029541, 'train/grad_norm_pre_clip': 7.373806953430176, 'eval/loss': 0.9470853209495544, 'eval/latency_0/loss': 0.9470853209495544, 'eval/demon_fire_precision': 0.896, 'eval/demon_fire_recall': 0.8837209302325582, 'eval/demon_fire_f1': 0.8898181057913443, 'eval/demon_fire_rate_model': 0.6604560622914349, 'eval/demon_fire_rate_teacher': 0.6696329254727474, 'eval/demon_left_precision': 0.8751091703056768, 'eval/demon_left_recall': 0.5677053824362607, 'eval/demon_left_f1': 0.688659793814433, 'eval/demon_left_rate_model': 0.31840934371523916, 'eval/demon_left_rate_teacher': 0.49082313681868744, 'eval/demon_right_precision': 0.6879554222031719, 'eval/demon_right_recall': 0.9067796610169492, 'eval/demon_right_f1': 0.7823543748476723, 'eval/demon_right_rate_model': 0.6487764182424917, 'eval/demon_right_rate_teacher': 0.4922135706340378, 'eval/demon_move_f1': 0.7355070843310527, 'eval/demon_cc_f1': 0.8371286690349471, 'eval/latency_0/demon_fire_precision': 0.896, 'eval/latency_0/demon_fire_recall': 0.8837209302325582, 'eval/latency_0/demon_fire_f1': 0.8898181057913443, 'eval/latency_0/demon_fire_rate_model': 0.6604560622914349, 'eval/latency_0/demon_fire_rate_teacher': 0.6696329254727474, 'eval/latency_0/demon_left_precision': 0.8751091703056768, 'eval/latency_0/demon_left_recall': 0.5677053824362607, 'eval/latency_0/demon_left_f1': 0.688659793814433, 'eval/latency_0/demon_left_rate_model': 0.31840934371523916, 'eval/latency_0/demon_left_rate_teacher': 0.49082313681868744, 'eval/latency_0/demon_right_precision': 0.6879554222031719, 'eval/latency_0/demon_right_recall': 0.9067796610169492, 'eval/latency_0/demon_right_f1': 0.7823543748476723, 'eval/latency_0/demon_right_rate_model': 0.6487764182424917, 'eval/latency_0/demon_right_rate_teacher': 0.4922135706340378, 'eval/latency_0/demon_move_f1': 0.7355070843310527, 'eval/latency_0/demon_cc_f1': 0.8371286690349471, 'timing/data': 0.004835334024392068, 'timing/model': 1.2129398620454594, 'learning_rate/qwen_vl_interface': 9.308492171595821e-06, 'learning_rate/action_model': 9.30849217159582e-05, 'epoch': 0.38}) +[2026-06-17 08:32:18,784][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_1300_state +[2026-06-17 08:32:18,787][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 08:32:22,330][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_1300_state/model.safetensors +[2026-06-17 08:32:29,982][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_1300_state/optimizer.bin +[2026-06-17 08:32:29,982][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_1300_state/sampler.bin +[2026-06-17 08:32:29,983][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_1300_state/sampler_1.bin +[2026-06-17 08:32:29,983][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_1300_state/random_states_0.pkl +[2026-06-17 08:32:30,047][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1300 +[2026-06-17 08:32:30,047][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 08:32:30,076][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 08:32:34,662][starVLA.training.train_starvla][INFO] - Step 1301, Loss: {'train/loss': 0.8094009160995483, 'train/grad_norm_pre_clip': 7.707120418548584, 'timing/data': 0.004887899965979159, 'timing/model': 1.188955778023228, 'learning_rate/qwen_vl_interface': 9.307368051647765e-06, 'learning_rate/action_model': 9.307368051647765e-05, 'epoch': 0.38}) +[2026-06-17 08:32:39,387][starVLA.training.train_starvla][INFO] - Step 1302, Loss: {'train/loss': 0.9656105041503906, 'train/grad_norm_pre_clip': 7.922793388366699, 'timing/data': 0.004684333922341466, 'timing/model': 1.1905231289565563, 'learning_rate/qwen_vl_interface': 9.306243090602945e-06, 'learning_rate/action_model': 9.306243090602946e-05, 'epoch': 0.38}) +[2026-06-17 08:32:44,115][starVLA.training.train_starvla][INFO] - Step 1303, Loss: {'train/loss': 1.1225600242614746, 'train/grad_norm_pre_clip': 20.830141067504883, 'timing/data': 0.004535353044047952, 'timing/model': 1.1917170869419351, 'learning_rate/qwen_vl_interface': 9.30511728869457e-06, 'learning_rate/action_model': 9.305117288694569e-05, 'epoch': 0.38}) +[2026-06-17 08:32:48,883][starVLA.training.train_starvla][INFO] - Step 1304, Loss: {'train/loss': 0.7277281880378723, 'train/grad_norm_pre_clip': 9.911203384399414, 'timing/data': 0.004484819946810603, 'timing/model': 1.1926884959684685, 'learning_rate/qwen_vl_interface': 9.303990646156015e-06, 'learning_rate/action_model': 9.303990646156014e-05, 'epoch': 0.38}) +[2026-06-17 08:32:53,628][starVLA.training.train_starvla][INFO] - Step 1305, Loss: {'train/loss': 1.0109208822250366, 'train/grad_norm_pre_clip': 11.142549514770508, 'timing/data': 0.004367506015114486, 'timing/model': 1.1937081150244921, 'learning_rate/qwen_vl_interface': 9.302863163220837e-06, 'learning_rate/action_model': 9.302863163220837e-05, 'epoch': 0.38}) +[2026-06-17 08:32:58,382][starVLA.training.train_starvla][INFO] - Step 1306, Loss: {'train/loss': 0.8567301034927368, 'train/grad_norm_pre_clip': 8.709176063537598, 'timing/data': 0.0044645129237324, 'timing/model': 1.3237828880082816, 'learning_rate/qwen_vl_interface': 9.301734840122765e-06, 'learning_rate/action_model': 9.301734840122766e-05, 'epoch': 0.38}) +[2026-06-17 08:33:03,003][starVLA.training.train_starvla][INFO] - Step 1307, Loss: {'train/loss': 0.9842165112495422, 'train/grad_norm_pre_clip': 8.916997909545898, 'timing/data': 0.004490876919589937, 'timing/model': 1.1957396429497749, 'learning_rate/qwen_vl_interface': 9.300605677095701e-06, 'learning_rate/action_model': 9.3006056770957e-05, 'epoch': 0.38}) +[2026-06-17 08:33:07,768][starVLA.training.train_starvla][INFO] - Step 1308, Loss: {'train/loss': 0.889190673828125, 'train/grad_norm_pre_clip': 10.093985557556152, 'timing/data': 0.004645220004022121, 'timing/model': 1.2017187440069392, 'learning_rate/qwen_vl_interface': 9.299475674373722e-06, 'learning_rate/action_model': 9.299475674373721e-05, 'epoch': 0.38}) +[2026-06-17 08:33:12,485][starVLA.training.train_starvla][INFO] - Step 1309, Loss: {'train/loss': 1.0015063285827637, 'train/grad_norm_pre_clip': 11.405166625976562, 'timing/data': 0.0046048329677432775, 'timing/model': 1.1998053379356861, 'learning_rate/qwen_vl_interface': 9.298344832191076e-06, 'learning_rate/action_model': 9.298344832191076e-05, 'epoch': 0.39}) +[2026-06-17 08:33:17,270][starVLA.training.train_starvla][INFO] - Step 1310, Loss: {'train/loss': 0.9506688117980957, 'train/grad_norm_pre_clip': 9.716390609741211, 'timing/data': 0.00464785797521472, 'timing/model': 1.1984048440353945, 'learning_rate/qwen_vl_interface': 9.297213150782192e-06, 'learning_rate/action_model': 9.297213150782192e-05, 'epoch': 0.39}) +[2026-06-17 08:33:22,040][starVLA.training.train_starvla][INFO] - Step 1311, Loss: {'train/loss': 0.8284195065498352, 'train/grad_norm_pre_clip': 11.81874942779541, 'timing/data': 0.004625392029993236, 'timing/model': 1.2008067219285294, 'learning_rate/qwen_vl_interface': 9.296080630381666e-06, 'learning_rate/action_model': 9.296080630381666e-05, 'epoch': 0.39}) +[2026-06-17 08:33:26,815][starVLA.training.train_starvla][INFO] - Step 1312, Loss: {'train/loss': 1.0524015426635742, 'train/grad_norm_pre_clip': 11.304667472839355, 'timing/data': 0.004388123983517289, 'timing/model': 1.3360621770843863, 'learning_rate/qwen_vl_interface': 9.294947271224271e-06, 'learning_rate/action_model': 9.294947271224272e-05, 'epoch': 0.39}) +[2026-06-17 08:33:31,584][starVLA.training.train_starvla][INFO] - Step 1313, Loss: {'train/loss': 1.4236721992492676, 'train/grad_norm_pre_clip': 9.345104217529297, 'timing/data': 0.004581974004395306, 'timing/model': 1.3227313651004806, 'learning_rate/qwen_vl_interface': 9.293813073544957e-06, 'learning_rate/action_model': 9.293813073544956e-05, 'epoch': 0.39}) +[2026-06-17 08:33:36,231][starVLA.training.train_starvla][INFO] - Step 1314, Loss: {'train/loss': 1.0527414083480835, 'train/grad_norm_pre_clip': 7.62871789932251, 'timing/data': 0.004504487034864724, 'timing/model': 1.2069450890412554, 'learning_rate/qwen_vl_interface': 9.292678037578838e-06, 'learning_rate/action_model': 9.292678037578837e-05, 'epoch': 0.39}) +[2026-06-17 08:33:40,964][starVLA.training.train_starvla][INFO] - Step 1315, Loss: {'train/loss': 0.8125836253166199, 'train/grad_norm_pre_clip': 6.928922176361084, 'timing/data': 0.00464520591776818, 'timing/model': 1.2025204329984263, 'learning_rate/qwen_vl_interface': 9.291542163561214e-06, 'learning_rate/action_model': 9.291542163561213e-05, 'epoch': 0.39}) +[2026-06-17 08:33:45,746][starVLA.training.train_starvla][INFO] - Step 1316, Loss: {'train/loss': 1.2413512468338013, 'train/grad_norm_pre_clip': 8.778290748596191, 'timing/data': 0.0046216140035539865, 'timing/model': 1.2050835160771385, 'learning_rate/qwen_vl_interface': 9.29040545172755e-06, 'learning_rate/action_model': 9.290405451727549e-05, 'epoch': 0.39}) +[2026-06-17 08:33:50,515][starVLA.training.train_starvla][INFO] - Step 1317, Loss: {'train/loss': 0.779296338558197, 'train/grad_norm_pre_clip': 13.072251319885254, 'timing/data': 0.004769803956151009, 'timing/model': 1.2023829859681427, 'learning_rate/qwen_vl_interface': 9.289267902313486e-06, 'learning_rate/action_model': 9.289267902313487e-05, 'epoch': 0.39}) +[2026-06-17 08:33:55,290][starVLA.training.train_starvla][INFO] - Step 1318, Loss: {'train/loss': 0.8929142951965332, 'train/grad_norm_pre_clip': 29.2781925201416, 'timing/data': 0.004511323990300298, 'timing/model': 1.3316977960057557, 'learning_rate/qwen_vl_interface': 9.288129515554843e-06, 'learning_rate/action_model': 9.288129515554843e-05, 'epoch': 0.39}) +[2026-06-17 08:34:00,056][starVLA.training.train_starvla][INFO] - Step 1319, Loss: {'train/loss': 1.0564383268356323, 'train/grad_norm_pre_clip': 13.904901504516602, 'timing/data': 0.00435600196942687, 'timing/model': 1.3215371889527887, 'learning_rate/qwen_vl_interface': 9.286990291687605e-06, 'learning_rate/action_model': 9.286990291687604e-05, 'epoch': 0.39}) +[2026-06-17 08:34:04,704][starVLA.training.train_starvla][INFO] - Step 1320, Loss: {'train/loss': 0.8823283910751343, 'train/grad_norm_pre_clip': 9.250027656555176, 'timing/data': 0.0043102429481223226, 'timing/model': 1.2078684449661523, 'learning_rate/qwen_vl_interface': 9.285850230947934e-06, 'learning_rate/action_model': 9.285850230947934e-05, 'epoch': 0.39}) +[2026-06-17 08:34:09,458][starVLA.training.train_starvla][INFO] - Step 1321, Loss: {'train/loss': 0.8629536032676697, 'train/grad_norm_pre_clip': 7.855433464050293, 'timing/data': 0.004447938990779221, 'timing/model': 1.2029417980229482, 'learning_rate/qwen_vl_interface': 9.28470933357217e-06, 'learning_rate/action_model': 9.28470933357217e-05, 'epoch': 0.39}) +[2026-06-17 08:34:14,239][starVLA.training.train_starvla][INFO] - Step 1322, Loss: {'train/loss': 0.7935806512832642, 'train/grad_norm_pre_clip': 8.45160961151123, 'timing/data': 0.004557815962471068, 'timing/model': 1.2040238319896162, 'learning_rate/qwen_vl_interface': 9.283567599796819e-06, 'learning_rate/action_model': 9.283567599796818e-05, 'epoch': 0.39}) +[2026-06-17 08:34:19,016][starVLA.training.train_starvla][INFO] - Step 1323, Loss: {'train/loss': 0.8684937357902527, 'train/grad_norm_pre_clip': 6.454057693481445, 'timing/data': 0.004564617993310094, 'timing/model': 1.201387829030864, 'learning_rate/qwen_vl_interface': 9.282425029858561e-06, 'learning_rate/action_model': 9.282425029858561e-05, 'epoch': 0.39}) +[2026-06-17 08:34:23,738][starVLA.training.train_starvla][INFO] - Step 1324, Loss: {'train/loss': 0.7736661434173584, 'train/grad_norm_pre_clip': 6.683426380157471, 'timing/data': 0.004575924016535282, 'timing/model': 1.2744549720082432, 'learning_rate/qwen_vl_interface': 9.281281623994258e-06, 'learning_rate/action_model': 9.281281623994258e-05, 'epoch': 0.39}) +[2026-06-17 08:34:28,391][starVLA.training.train_starvla][INFO] - Step 1325, Loss: {'train/loss': 0.9870333075523376, 'train/grad_norm_pre_clip': 8.65235424041748, 'timing/data': 0.004778918926604092, 'timing/model': 1.2052854609210044, 'learning_rate/qwen_vl_interface': 9.280137382440936e-06, 'learning_rate/action_model': 9.280137382440937e-05, 'epoch': 0.39}) +[2026-06-17 08:34:33,154][starVLA.training.train_starvla][INFO] - Step 1326, Loss: {'train/loss': 0.8567508459091187, 'train/grad_norm_pre_clip': 11.768573760986328, 'timing/data': 0.004570679040625691, 'timing/model': 1.2052631539991125, 'learning_rate/qwen_vl_interface': 9.278992305435798e-06, 'learning_rate/action_model': 9.278992305435797e-05, 'epoch': 0.39}) +[2026-06-17 08:34:37,943][starVLA.training.train_starvla][INFO] - Step 1327, Loss: {'train/loss': 0.8226253986358643, 'train/grad_norm_pre_clip': 8.182829856872559, 'timing/data': 0.004393802955746651, 'timing/model': 1.204990238067694, 'learning_rate/qwen_vl_interface': 9.27784639321622e-06, 'learning_rate/action_model': 9.27784639321622e-05, 'epoch': 0.39}) +[2026-06-17 08:34:42,733][starVLA.training.train_starvla][INFO] - Step 1328, Loss: {'train/loss': 0.6225618124008179, 'train/grad_norm_pre_clip': 5.155083656311035, 'timing/data': 0.004439031006768346, 'timing/model': 1.2039970179321244, 'learning_rate/qwen_vl_interface': 9.27669964601975e-06, 'learning_rate/action_model': 9.27669964601975e-05, 'epoch': 0.39}) +[2026-06-17 08:34:47,515][starVLA.training.train_starvla][INFO] - Step 1329, Loss: {'train/loss': 0.715616762638092, 'train/grad_norm_pre_clip': 15.25643253326416, 'timing/data': 0.00476412195712328, 'timing/model': 1.3339249819982797, 'learning_rate/qwen_vl_interface': 9.27555206408411e-06, 'learning_rate/action_model': 9.275552064084111e-05, 'epoch': 0.39}) +[2026-06-17 08:34:52,248][starVLA.training.train_starvla][INFO] - Step 1330, Loss: {'train/loss': 0.8867442607879639, 'train/grad_norm_pre_clip': 9.34764575958252, 'timing/data': 0.004144730046391487, 'timing/model': 1.280608730041422, 'learning_rate/qwen_vl_interface': 9.274403647647197e-06, 'learning_rate/action_model': 9.274403647647197e-05, 'epoch': 0.39}) +[2026-06-17 08:34:56,908][starVLA.training.train_starvla][INFO] - Step 1331, Loss: {'train/loss': 0.8186548948287964, 'train/grad_norm_pre_clip': 6.584742069244385, 'timing/data': 0.0046095100697129965, 'timing/model': 1.2034351479960606, 'learning_rate/qwen_vl_interface': 9.273254396947079e-06, 'learning_rate/action_model': 9.273254396947078e-05, 'epoch': 0.39}) +[2026-06-17 08:35:01,717][starVLA.training.train_starvla][INFO] - Step 1332, Loss: {'train/loss': 1.0534632205963135, 'train/grad_norm_pre_clip': 8.099634170532227, 'timing/data': 0.004569642944261432, 'timing/model': 1.2128538439283147, 'learning_rate/qwen_vl_interface': 9.27210431222199e-06, 'learning_rate/action_model': 9.27210431222199e-05, 'epoch': 0.39}) +[2026-06-17 08:35:06,513][starVLA.training.train_starvla][INFO] - Step 1333, Loss: {'train/loss': 0.8054149746894836, 'train/grad_norm_pre_clip': 9.23855209350586, 'timing/data': 0.0044666200410574675, 'timing/model': 1.204456120962277, 'learning_rate/qwen_vl_interface': 9.270953393710353e-06, 'learning_rate/action_model': 9.270953393710352e-05, 'epoch': 0.39}) +[2026-06-17 08:35:11,296][starVLA.training.train_starvla][INFO] - Step 1334, Loss: {'train/loss': 0.7482994794845581, 'train/grad_norm_pre_clip': 10.528852462768555, 'timing/data': 0.004705194034613669, 'timing/model': 1.2053003049222752, 'learning_rate/qwen_vl_interface': 9.269801641650751e-06, 'learning_rate/action_model': 9.26980164165075e-05, 'epoch': 0.39}) +[2026-06-17 08:35:16,071][starVLA.training.train_starvla][INFO] - Step 1335, Loss: {'train/loss': 0.6836286783218384, 'train/grad_norm_pre_clip': 7.200263500213623, 'timing/data': 0.004729775013402104, 'timing/model': 1.2040209060069174, 'learning_rate/qwen_vl_interface': 9.268649056281941e-06, 'learning_rate/action_model': 9.268649056281941e-05, 'epoch': 0.39}) +[2026-06-17 08:35:20,858][starVLA.training.train_starvla][INFO] - Step 1336, Loss: {'train/loss': 0.8199909925460815, 'train/grad_norm_pre_clip': 8.725775718688965, 'timing/data': 0.004381962935440242, 'timing/model': 1.2896994529291987, 'learning_rate/qwen_vl_interface': 9.267495637842859e-06, 'learning_rate/action_model': 9.267495637842857e-05, 'epoch': 0.39}) +[2026-06-17 08:35:25,513][starVLA.training.train_starvla][INFO] - Step 1337, Loss: {'train/loss': 0.9077467918395996, 'train/grad_norm_pre_clip': 7.791637897491455, 'timing/data': 0.004665457061491907, 'timing/model': 1.2042500600218773, 'learning_rate/qwen_vl_interface': 9.266341386572607e-06, 'learning_rate/action_model': 9.266341386572605e-05, 'epoch': 0.39}) +[2026-06-17 08:35:30,285][starVLA.training.train_starvla][INFO] - Step 1338, Loss: {'train/loss': 0.664588212966919, 'train/grad_norm_pre_clip': 7.3611884117126465, 'timing/data': 0.004363271989859641, 'timing/model': 1.2034524150658399, 'learning_rate/qwen_vl_interface': 9.265186302710465e-06, 'learning_rate/action_model': 9.265186302710465e-05, 'epoch': 0.39}) +[2026-06-17 08:35:35,068][starVLA.training.train_starvla][INFO] - Step 1339, Loss: {'train/loss': 0.9258065223693848, 'train/grad_norm_pre_clip': 8.331171989440918, 'timing/data': 0.004632280906662345, 'timing/model': 1.2044811839004979, 'learning_rate/qwen_vl_interface': 9.26403038649588e-06, 'learning_rate/action_model': 9.26403038649588e-05, 'epoch': 0.39}) +[2026-06-17 08:35:39,850][starVLA.training.train_starvla][INFO] - Step 1340, Loss: {'train/loss': 0.6793895363807678, 'train/grad_norm_pre_clip': 9.367729187011719, 'timing/data': 0.004451598972082138, 'timing/model': 1.2028239809442312, 'learning_rate/qwen_vl_interface': 9.262873638168478e-06, 'learning_rate/action_model': 9.262873638168478e-05, 'epoch': 0.39}) +[2026-06-17 08:35:44,645][starVLA.training.train_starvla][INFO] - Step 1341, Loss: {'train/loss': 1.017880916595459, 'train/grad_norm_pre_clip': 11.366352081298828, 'timing/data': 0.004234830033965409, 'timing/model': 1.3496324960142374, 'learning_rate/qwen_vl_interface': 9.261716057968054e-06, 'learning_rate/action_model': 9.261716057968053e-05, 'epoch': 0.39}) +[2026-06-17 08:35:49,303][starVLA.training.train_starvla][INFO] - Step 1342, Loss: {'train/loss': 0.9463111162185669, 'train/grad_norm_pre_clip': 9.100054740905762, 'timing/data': 0.00450092600658536, 'timing/model': 1.2055013590725139, 'learning_rate/qwen_vl_interface': 9.260557646134572e-06, 'learning_rate/action_model': 9.260557646134572e-05, 'epoch': 0.39}) +[2026-06-17 08:35:54,102][starVLA.training.train_starvla][INFO] - Step 1343, Loss: {'train/loss': 0.6048139929771423, 'train/grad_norm_pre_clip': 6.364327430725098, 'timing/data': 0.0047178949462249875, 'timing/model': 1.2043673919979483, 'learning_rate/qwen_vl_interface': 9.259398402908175e-06, 'learning_rate/action_model': 9.259398402908174e-05, 'epoch': 0.4}) +[2026-06-17 08:35:58,880][starVLA.training.train_starvla][INFO] - Step 1344, Loss: {'train/loss': 0.90216064453125, 'train/grad_norm_pre_clip': 8.095531463623047, 'timing/data': 0.0045040230033919215, 'timing/model': 1.204931756015867, 'learning_rate/qwen_vl_interface': 9.258238328529177e-06, 'learning_rate/action_model': 9.258238328529176e-05, 'epoch': 0.4}) +[2026-06-17 08:36:03,586][starVLA.training.train_starvla][INFO] - Step 1345, Loss: {'train/loss': 0.7219231128692627, 'train/grad_norm_pre_clip': 7.41546630859375, 'timing/data': 0.005352435982786119, 'timing/model': 1.2066981060197577, 'learning_rate/qwen_vl_interface': 9.257077423238058e-06, 'learning_rate/action_model': 9.257077423238058e-05, 'epoch': 0.4}) +[2026-06-17 08:36:08,322][starVLA.training.train_starvla][INFO] - Step 1346, Loss: {'train/loss': 1.2289577722549438, 'train/grad_norm_pre_clip': 12.0871000289917, 'timing/data': 0.0046049830270931125, 'timing/model': 1.2048595449887216, 'learning_rate/qwen_vl_interface': 9.25591568727548e-06, 'learning_rate/action_model': 9.255915687275478e-05, 'epoch': 0.4}) +[2026-06-17 08:36:13,110][starVLA.training.train_starvla][INFO] - Step 1347, Loss: {'train/loss': 1.183557152748108, 'train/grad_norm_pre_clip': 14.468587875366211, 'timing/data': 0.004567247931845486, 'timing/model': 1.3410601950017735, 'learning_rate/qwen_vl_interface': 9.254753120882265e-06, 'learning_rate/action_model': 9.254753120882266e-05, 'epoch': 0.4}) +[2026-06-17 08:36:17,770][starVLA.training.train_starvla][INFO] - Step 1348, Loss: {'train/loss': 0.6739754676818848, 'train/grad_norm_pre_clip': 7.2852396965026855, 'timing/data': 0.004510760889388621, 'timing/model': 1.2061285570962355, 'learning_rate/qwen_vl_interface': 9.25358972429942e-06, 'learning_rate/action_model': 9.253589724299421e-05, 'epoch': 0.4}) +[2026-06-17 08:36:22,587][starVLA.training.train_starvla][INFO] - Step 1349, Loss: {'train/loss': 0.8298186659812927, 'train/grad_norm_pre_clip': 9.627120971679688, 'timing/data': 0.0046549380058422685, 'timing/model': 1.2025880119763315, 'learning_rate/qwen_vl_interface': 9.25242549776812e-06, 'learning_rate/action_model': 9.252425497768119e-05, 'epoch': 0.4}) +[2026-06-17 08:36:27,372][starVLA.training.train_starvla][INFO] - Step 1350, Loss: {'train/loss': 0.9632823467254639, 'train/grad_norm_pre_clip': 8.915114402770996, 'timing/data': 0.004374590003862977, 'timing/model': 1.2046993310796097, 'learning_rate/qwen_vl_interface': 9.251260441529706e-06, 'learning_rate/action_model': 9.251260441529705e-05, 'epoch': 0.4}) +[2026-06-17 08:36:32,155][starVLA.training.train_starvla][INFO] - Step 1351, Loss: {'train/loss': 0.8629000186920166, 'train/grad_norm_pre_clip': 8.56259536743164, 'timing/data': 0.00463384494651109, 'timing/model': 1.2096214910270646, 'learning_rate/qwen_vl_interface': 9.250094555825697e-06, 'learning_rate/action_model': 9.250094555825696e-05, 'epoch': 0.4}) +[2026-06-17 08:36:36,909][starVLA.training.train_starvla][INFO] - Step 1352, Loss: {'train/loss': 0.829501211643219, 'train/grad_norm_pre_clip': 10.142345428466797, 'timing/data': 0.0045537169789895415, 'timing/model': 1.205755305942148, 'learning_rate/qwen_vl_interface': 9.248927840897782e-06, 'learning_rate/action_model': 9.248927840897781e-05, 'epoch': 0.4}) +[2026-06-17 08:36:41,689][starVLA.training.train_starvla][INFO] - Step 1353, Loss: {'train/loss': 0.7718232274055481, 'train/grad_norm_pre_clip': 8.461167335510254, 'timing/data': 0.004531748942099512, 'timing/model': 1.3236597399227321, 'learning_rate/qwen_vl_interface': 9.247760296987822e-06, 'learning_rate/action_model': 9.247760296987821e-05, 'epoch': 0.4}) +[2026-06-17 08:36:46,349][starVLA.training.train_starvla][INFO] - Step 1354, Loss: {'train/loss': 0.7797051668167114, 'train/grad_norm_pre_clip': 10.4595947265625, 'timing/data': 0.004213299020193517, 'timing/model': 1.2057574120117351, 'learning_rate/qwen_vl_interface': 9.246591924337852e-06, 'learning_rate/action_model': 9.246591924337852e-05, 'epoch': 0.4}) +[2026-06-17 08:36:51,135][starVLA.training.train_starvla][INFO] - Step 1355, Loss: {'train/loss': 0.480776846408844, 'train/grad_norm_pre_clip': 5.801819324493408, 'timing/data': 0.004497878020629287, 'timing/model': 1.2047321810387075, 'learning_rate/qwen_vl_interface': 9.245422723190076e-06, 'learning_rate/action_model': 9.245422723190074e-05, 'epoch': 0.4}) +[2026-06-17 08:36:55,934][starVLA.training.train_starvla][INFO] - Step 1356, Loss: {'train/loss': 0.8187692761421204, 'train/grad_norm_pre_clip': 9.21265983581543, 'timing/data': 0.0043658510548993945, 'timing/model': 1.2041824270272627, 'learning_rate/qwen_vl_interface': 9.244252693786868e-06, 'learning_rate/action_model': 9.244252693786869e-05, 'epoch': 0.4}) +[2026-06-17 08:37:00,735][starVLA.training.train_starvla][INFO] - Step 1357, Loss: {'train/loss': 0.90236496925354, 'train/grad_norm_pre_clip': 8.876221656799316, 'timing/data': 0.004226127057336271, 'timing/model': 1.2049759889487177, 'learning_rate/qwen_vl_interface': 9.24308183637078e-06, 'learning_rate/action_model': 9.24308183637078e-05, 'epoch': 0.4}) +[2026-06-17 08:37:05,528][starVLA.training.train_starvla][INFO] - Step 1358, Loss: {'train/loss': 0.7628923654556274, 'train/grad_norm_pre_clip': 7.9691996574401855, 'timing/data': 0.0041247650515288115, 'timing/model': 1.3367885909974575, 'learning_rate/qwen_vl_interface': 9.24191015118453e-06, 'learning_rate/action_model': 9.24191015118453e-05, 'epoch': 0.4}) +[2026-06-17 08:37:10,198][starVLA.training.train_starvla][INFO] - Step 1359, Loss: {'train/loss': 1.161394715309143, 'train/grad_norm_pre_clip': 11.963014602661133, 'timing/data': 0.004849968012422323, 'timing/model': 1.2112039549974725, 'learning_rate/qwen_vl_interface': 9.24073763847101e-06, 'learning_rate/action_model': 9.24073763847101e-05, 'epoch': 0.4}) +[2026-06-17 08:37:15,001][starVLA.training.train_starvla][INFO] - Step 1360, Loss: {'train/loss': 0.9847855567932129, 'train/grad_norm_pre_clip': 11.790484428405762, 'timing/data': 0.004224392003379762, 'timing/model': 1.2106932869646698, 'learning_rate/qwen_vl_interface': 9.239564298473284e-06, 'learning_rate/action_model': 9.239564298473284e-05, 'epoch': 0.4}) +[2026-06-17 08:37:19,705][starVLA.training.train_starvla][INFO] - Step 1361, Loss: {'train/loss': 0.676338255405426, 'train/grad_norm_pre_clip': 10.386629104614258, 'timing/data': 0.004258695058524609, 'timing/model': 1.205194387002848, 'learning_rate/qwen_vl_interface': 9.238390131434586e-06, 'learning_rate/action_model': 9.238390131434586e-05, 'epoch': 0.4}) +[2026-06-17 08:37:24,453][starVLA.training.train_starvla][INFO] - Step 1362, Loss: {'train/loss': 0.9618977308273315, 'train/grad_norm_pre_clip': 10.226669311523438, 'timing/data': 0.004532520077191293, 'timing/model': 1.2044147478882223, 'learning_rate/qwen_vl_interface': 9.237215137598322e-06, 'learning_rate/action_model': 9.237215137598322e-05, 'epoch': 0.4}) +[2026-06-17 08:37:29,243][starVLA.training.train_starvla][INFO] - Step 1363, Loss: {'train/loss': 0.6962181329727173, 'train/grad_norm_pre_clip': 19.735015869140625, 'timing/data': 0.004564476083032787, 'timing/model': 1.2056983900256455, 'learning_rate/qwen_vl_interface': 9.236039317208068e-06, 'learning_rate/action_model': 9.236039317208068e-05, 'epoch': 0.4}) +[2026-06-17 08:37:34,030][starVLA.training.train_starvla][INFO] - Step 1364, Loss: {'train/loss': 0.8549943566322327, 'train/grad_norm_pre_clip': 8.850006103515625, 'timing/data': 0.004640729981474578, 'timing/model': 1.333755286061205, 'learning_rate/qwen_vl_interface': 9.234862670507577e-06, 'learning_rate/action_model': 9.234862670507576e-05, 'epoch': 0.4}) +[2026-06-17 08:37:38,693][starVLA.training.train_starvla][INFO] - Step 1365, Loss: {'train/loss': 0.6372213959693909, 'train/grad_norm_pre_clip': 7.210105895996094, 'timing/data': 0.0048024519346654415, 'timing/model': 1.2043721979716793, 'learning_rate/qwen_vl_interface': 9.233685197740765e-06, 'learning_rate/action_model': 9.233685197740764e-05, 'epoch': 0.4}) +[2026-06-17 08:37:43,483][starVLA.training.train_starvla][INFO] - Step 1366, Loss: {'train/loss': 0.6399343013763428, 'train/grad_norm_pre_clip': 6.732353210449219, 'timing/data': 0.004389288020320237, 'timing/model': 1.204205538961105, 'learning_rate/qwen_vl_interface': 9.232506899151725e-06, 'learning_rate/action_model': 9.232506899151725e-05, 'epoch': 0.4}) +[2026-06-17 08:37:48,262][starVLA.training.train_starvla][INFO] - Step 1367, Loss: {'train/loss': 1.4573631286621094, 'train/grad_norm_pre_clip': 16.6043701171875, 'timing/data': 0.004436251940205693, 'timing/model': 1.2041873269481584, 'learning_rate/qwen_vl_interface': 9.23132777498472e-06, 'learning_rate/action_model': 9.231327774984719e-05, 'epoch': 0.4}) +[2026-06-17 08:37:53,016][starVLA.training.train_starvla][INFO] - Step 1368, Loss: {'train/loss': 0.8235408067703247, 'train/grad_norm_pre_clip': 11.04619312286377, 'timing/data': 0.004633630975149572, 'timing/model': 1.204715765081346, 'learning_rate/qwen_vl_interface': 9.230147825484182e-06, 'learning_rate/action_model': 9.230147825484183e-05, 'epoch': 0.4}) +[2026-06-17 08:37:57,784][starVLA.training.train_starvla][INFO] - Step 1369, Loss: {'train/loss': 0.8977580666542053, 'train/grad_norm_pre_clip': 9.667160987854004, 'timing/data': 0.004335117992013693, 'timing/model': 1.2041579990182072, 'learning_rate/qwen_vl_interface': 9.228967050894717e-06, 'learning_rate/action_model': 9.228967050894716e-05, 'epoch': 0.4}) +[2026-06-17 08:38:02,574][starVLA.training.train_starvla][INFO] - Step 1370, Loss: {'train/loss': 1.0138821601867676, 'train/grad_norm_pre_clip': 12.32621955871582, 'timing/data': 0.004080022103153169, 'timing/model': 1.3368364489870146, 'learning_rate/qwen_vl_interface': 9.227785451461103e-06, 'learning_rate/action_model': 9.227785451461102e-05, 'epoch': 0.4}) +[2026-06-17 08:38:07,239][starVLA.training.train_starvla][INFO] - Step 1371, Loss: {'train/loss': 0.8281669020652771, 'train/grad_norm_pre_clip': 22.237699508666992, 'timing/data': 0.00430572999175638, 'timing/model': 1.2092584180645645, 'learning_rate/qwen_vl_interface': 9.226603027428281e-06, 'learning_rate/action_model': 9.226603027428281e-05, 'epoch': 0.4}) +[2026-06-17 08:38:12,094][starVLA.training.train_starvla][INFO] - Step 1372, Loss: {'train/loss': 0.88334721326828, 'train/grad_norm_pre_clip': 16.687009811401367, 'timing/data': 0.004462494049221277, 'timing/model': 1.2066273060627282, 'learning_rate/qwen_vl_interface': 9.225419779041376e-06, 'learning_rate/action_model': 9.225419779041376e-05, 'epoch': 0.4}) +[2026-06-17 08:38:16,882][starVLA.training.train_starvla][INFO] - Step 1373, Loss: {'train/loss': 0.8166179060935974, 'train/grad_norm_pre_clip': 7.759292125701904, 'timing/data': 0.004602259024977684, 'timing/model': 1.2066420100163668, 'learning_rate/qwen_vl_interface': 9.224235706545672e-06, 'learning_rate/action_model': 9.224235706545671e-05, 'epoch': 0.4}) +[2026-06-17 08:38:21,669][starVLA.training.train_starvla][INFO] - Step 1374, Loss: {'train/loss': 0.7583180069923401, 'train/grad_norm_pre_clip': 11.446759223937988, 'timing/data': 0.004385653999634087, 'timing/model': 1.2055139469448477, 'learning_rate/qwen_vl_interface': 9.22305081018663e-06, 'learning_rate/action_model': 9.223050810186629e-05, 'epoch': 0.4}) +[2026-06-17 08:38:26,349][starVLA.training.train_starvla][INFO] - Step 1375, Loss: {'train/loss': 0.8891414403915405, 'train/grad_norm_pre_clip': 10.022130966186523, 'timing/data': 0.004410170018672943, 'timing/model': 1.2061734619783238, 'learning_rate/qwen_vl_interface': 9.221865090209878e-06, 'learning_rate/action_model': 9.221865090209878e-05, 'epoch': 0.4}) +[2026-06-17 08:38:31,040][starVLA.training.train_starvla][INFO] - Step 1376, Loss: {'train/loss': 0.8996617197990417, 'train/grad_norm_pre_clip': 9.598246574401855, 'timing/data': 0.004557327949441969, 'timing/model': 1.2405941830947995, 'learning_rate/qwen_vl_interface': 9.22067854686122e-06, 'learning_rate/action_model': 9.22067854686122e-05, 'epoch': 0.4}) +[2026-06-17 08:38:35,688][starVLA.training.train_starvla][INFO] - Step 1377, Loss: {'train/loss': 0.8586314916610718, 'train/grad_norm_pre_clip': 10.806814193725586, 'timing/data': 0.004505497985519469, 'timing/model': 1.2038116479525343, 'learning_rate/qwen_vl_interface': 9.219491180386626e-06, 'learning_rate/action_model': 9.219491180386626e-05, 'epoch': 0.41}) +[2026-06-17 08:38:40,455][starVLA.training.train_starvla][INFO] - Step 1378, Loss: {'train/loss': 0.9080596566200256, 'train/grad_norm_pre_clip': 8.498194694519043, 'timing/data': 0.0042885810835286975, 'timing/model': 1.2050936970626935, 'learning_rate/qwen_vl_interface': 9.21830299103224e-06, 'learning_rate/action_model': 9.21830299103224e-05, 'epoch': 0.41}) +[2026-06-17 08:38:45,245][starVLA.training.train_starvla][INFO] - Step 1379, Loss: {'train/loss': 1.0408169031143188, 'train/grad_norm_pre_clip': 10.050682067871094, 'timing/data': 0.004466568003408611, 'timing/model': 1.2098961840383708, 'learning_rate/qwen_vl_interface': 9.217113979044373e-06, 'learning_rate/action_model': 9.217113979044373e-05, 'epoch': 0.41}) +[2026-06-17 08:38:50,089][starVLA.training.train_starvla][INFO] - Step 1380, Loss: {'train/loss': 0.8710231781005859, 'train/grad_norm_pre_clip': 20.09339714050293, 'timing/data': 0.004518390982411802, 'timing/model': 1.2050036260625347, 'learning_rate/qwen_vl_interface': 9.215924144669509e-06, 'learning_rate/action_model': 9.215924144669509e-05, 'epoch': 0.41}) +[2026-06-17 08:38:54,865][starVLA.training.train_starvla][INFO] - Step 1381, Loss: {'train/loss': 0.9039709568023682, 'train/grad_norm_pre_clip': 12.039552688598633, 'timing/data': 0.004614058067090809, 'timing/model': 1.3291419669985771, 'learning_rate/qwen_vl_interface': 9.214733488154304e-06, 'learning_rate/action_model': 9.214733488154304e-05, 'epoch': 0.41}) +[2026-06-17 08:38:59,669][starVLA.training.train_starvla][INFO] - Step 1382, Loss: {'train/loss': 0.8359825015068054, 'train/grad_norm_pre_clip': 9.895352363586426, 'timing/data': 0.004308147937990725, 'timing/model': 1.3247784350533038, 'learning_rate/qwen_vl_interface': 9.213542009745579e-06, 'learning_rate/action_model': 9.213542009745578e-05, 'epoch': 0.41}) +[2026-06-17 08:39:04,330][starVLA.training.train_starvla][INFO] - Step 1383, Loss: {'train/loss': 0.6677770018577576, 'train/grad_norm_pre_clip': 7.19570255279541, 'timing/data': 0.004396845935843885, 'timing/model': 1.2051100389799103, 'learning_rate/qwen_vl_interface': 9.21234970969033e-06, 'learning_rate/action_model': 9.21234970969033e-05, 'epoch': 0.41}) +[2026-06-17 08:39:09,081][starVLA.training.train_starvla][INFO] - Step 1384, Loss: {'train/loss': 1.0311872959136963, 'train/grad_norm_pre_clip': 9.84560489654541, 'timing/data': 0.004500743001699448, 'timing/model': 1.204518657992594, 'learning_rate/qwen_vl_interface': 9.211156588235724e-06, 'learning_rate/action_model': 9.211156588235725e-05, 'epoch': 0.41}) +[2026-06-17 08:39:13,866][starVLA.training.train_starvla][INFO] - Step 1385, Loss: {'train/loss': 0.6876575350761414, 'train/grad_norm_pre_clip': 8.708868980407715, 'timing/data': 0.0045055509544909, 'timing/model': 1.202893320005387, 'learning_rate/qwen_vl_interface': 9.209962645629095e-06, 'learning_rate/action_model': 9.209962645629095e-05, 'epoch': 0.41}) +[2026-06-17 08:39:18,642][starVLA.training.train_starvla][INFO] - Step 1386, Loss: {'train/loss': 1.1295101642608643, 'train/grad_norm_pre_clip': 10.01480770111084, 'timing/data': 0.004646757966838777, 'timing/model': 1.2055244089569896, 'learning_rate/qwen_vl_interface': 9.208767882117947e-06, 'learning_rate/action_model': 9.208767882117947e-05, 'epoch': 0.41}) +[2026-06-17 08:39:23,426][starVLA.training.train_starvla][INFO] - Step 1387, Loss: {'train/loss': 0.7326560020446777, 'train/grad_norm_pre_clip': 6.653053283691406, 'timing/data': 0.004473809967748821, 'timing/model': 1.3307767799124122, 'learning_rate/qwen_vl_interface': 9.207572297949958e-06, 'learning_rate/action_model': 9.207572297949958e-05, 'epoch': 0.41}) +[2026-06-17 08:39:28,221][starVLA.training.train_starvla][INFO] - Step 1388, Loss: {'train/loss': 0.6561980247497559, 'train/grad_norm_pre_clip': 9.73131275177002, 'timing/data': 0.004612184013240039, 'timing/model': 1.3308001919649541, 'learning_rate/qwen_vl_interface': 9.206375893372973e-06, 'learning_rate/action_model': 9.206375893372973e-05, 'epoch': 0.41}) +[2026-06-17 08:39:32,905][starVLA.training.train_starvla][INFO] - Step 1389, Loss: {'train/loss': 0.8786114454269409, 'train/grad_norm_pre_clip': 13.180923461914062, 'timing/data': 0.004366172943264246, 'timing/model': 1.21174247097224, 'learning_rate/qwen_vl_interface': 9.20517866863501e-06, 'learning_rate/action_model': 9.205178668635009e-05, 'epoch': 0.41}) +[2026-06-17 08:39:37,696][starVLA.training.train_starvla][INFO] - Step 1390, Loss: {'train/loss': 0.9468203186988831, 'train/grad_norm_pre_clip': 9.088037490844727, 'timing/data': 0.004647545982152224, 'timing/model': 1.2130570859881118, 'learning_rate/qwen_vl_interface': 9.20398062398425e-06, 'learning_rate/action_model': 9.20398062398425e-05, 'epoch': 0.41}) +[2026-06-17 08:39:42,672][starVLA.training.train_starvla][INFO] - Step 1391, Loss: {'train/loss': 0.9512295722961426, 'train/grad_norm_pre_clip': 8.797496795654297, 'timing/data': 0.004373915959149599, 'timing/model': 1.401505807065405, 'learning_rate/qwen_vl_interface': 9.202781759669053e-06, 'learning_rate/action_model': 9.202781759669052e-05, 'epoch': 0.41}) +[2026-06-17 08:39:47,500][starVLA.training.train_starvla][INFO] - Step 1392, Loss: {'train/loss': 0.8331899046897888, 'train/grad_norm_pre_clip': 7.646701812744141, 'timing/data': 0.004738855990581214, 'timing/model': 1.2126689680153504, 'learning_rate/qwen_vl_interface': 9.201582075937943e-06, 'learning_rate/action_model': 9.201582075937944e-05, 'epoch': 0.41}) +[2026-06-17 08:39:52,239][starVLA.training.train_starvla][INFO] - Step 1393, Loss: {'train/loss': 1.1558294296264648, 'train/grad_norm_pre_clip': 10.489361763000488, 'timing/data': 0.004364102962426841, 'timing/model': 1.2737838750472292, 'learning_rate/qwen_vl_interface': 9.200381573039619e-06, 'learning_rate/action_model': 9.200381573039619e-05, 'epoch': 0.41}) +[2026-06-17 08:39:56,928][starVLA.training.train_starvla][INFO] - Step 1394, Loss: {'train/loss': 0.9183610677719116, 'train/grad_norm_pre_clip': 8.3220853805542, 'timing/data': 0.004503639996983111, 'timing/model': 1.2233357289806008, 'learning_rate/qwen_vl_interface': 9.199180251222942e-06, 'learning_rate/action_model': 9.199180251222942e-05, 'epoch': 0.41}) +[2026-06-17 08:40:01,749][starVLA.training.train_starvla][INFO] - Step 1395, Loss: {'train/loss': 0.7025425434112549, 'train/grad_norm_pre_clip': 9.85183048248291, 'timing/data': 0.004023577901534736, 'timing/model': 1.2245173529954627, 'learning_rate/qwen_vl_interface': 9.197978110736951e-06, 'learning_rate/action_model': 9.19797811073695e-05, 'epoch': 0.41}) +[2026-06-17 08:40:06,558][starVLA.training.train_starvla][INFO] - Step 1396, Loss: {'train/loss': 0.7354231476783752, 'train/grad_norm_pre_clip': 7.94968843460083, 'timing/data': 0.004523950046859682, 'timing/model': 1.2118478670017794, 'learning_rate/qwen_vl_interface': 9.196775151830849e-06, 'learning_rate/action_model': 9.196775151830849e-05, 'epoch': 0.41}) +[2026-06-17 08:40:11,382][starVLA.training.train_starvla][INFO] - Step 1397, Loss: {'train/loss': 0.9517797231674194, 'train/grad_norm_pre_clip': 10.028219223022461, 'timing/data': 0.004836712963879108, 'timing/model': 1.2102203470421955, 'learning_rate/qwen_vl_interface': 9.19557137475401e-06, 'learning_rate/action_model': 9.19557137475401e-05, 'epoch': 0.41}) +[2026-06-17 08:40:16,192][starVLA.training.train_starvla][INFO] - Step 1398, Loss: {'train/loss': 0.5207890272140503, 'train/grad_norm_pre_clip': 7.378370761871338, 'timing/data': 0.0045504969311878085, 'timing/model': 1.3460706860059872, 'learning_rate/qwen_vl_interface': 9.19436677975598e-06, 'learning_rate/action_model': 9.19436677975598e-05, 'epoch': 0.41}) +[2026-06-17 08:40:20,961][starVLA.training.train_starvla][INFO] - Step 1399, Loss: {'train/loss': 0.8162636756896973, 'train/grad_norm_pre_clip': 9.350601196289062, 'timing/data': 0.0047512518940493464, 'timing/model': 1.2858602409251034, 'learning_rate/qwen_vl_interface': 9.193161367086472e-06, 'learning_rate/action_model': 9.193161367086472e-05, 'epoch': 0.41}) +[2026-06-17 08:42:02,788][starVLA.training.train_starvla][INFO] - Step 1400, Loss: {'train/loss': 0.7067258358001709, 'train/grad_norm_pre_clip': 9.101737976074219, 'eval/loss': 0.8910677433013916, 'eval/latency_0/loss': 0.8910677433013916, 'eval/demon_fire_precision': 0.9019118869492935, 'eval/demon_fire_recall': 0.9011627906976745, 'eval/demon_fire_f1': 0.901537183215621, 'eval/demon_fire_rate_model': 0.6690767519466073, 'eval/demon_fire_rate_teacher': 0.6696329254727474, 'eval/demon_left_precision': 0.7447462839569452, 'eval/demon_left_recall': 0.8232294617563739, 'eval/demon_left_f1': 0.7820236813778256, 'eval/demon_left_rate_model': 0.5425472747497219, 'eval/demon_left_rate_teacher': 0.49082313681868744, 'eval/demon_right_precision': 0.8124602670057216, 'eval/demon_right_recall': 0.7220338983050848, 'eval/demon_right_f1': 0.7645827101405923, 'eval/demon_right_rate_model': 0.4374304783092325, 'eval/demon_right_rate_teacher': 0.4922135706340378, 'eval/demon_move_f1': 0.7733031957592089, 'eval/demon_cc_f1': 0.858813052694534, 'eval/latency_0/demon_fire_precision': 0.9019118869492935, 'eval/latency_0/demon_fire_recall': 0.9011627906976745, 'eval/latency_0/demon_fire_f1': 0.901537183215621, 'eval/latency_0/demon_fire_rate_model': 0.6690767519466073, 'eval/latency_0/demon_fire_rate_teacher': 0.6696329254727474, 'eval/latency_0/demon_left_precision': 0.7447462839569452, 'eval/latency_0/demon_left_recall': 0.8232294617563739, 'eval/latency_0/demon_left_f1': 0.7820236813778256, 'eval/latency_0/demon_left_rate_model': 0.5425472747497219, 'eval/latency_0/demon_left_rate_teacher': 0.49082313681868744, 'eval/latency_0/demon_right_precision': 0.8124602670057216, 'eval/latency_0/demon_right_recall': 0.7220338983050848, 'eval/latency_0/demon_right_f1': 0.7645827101405923, 'eval/latency_0/demon_right_rate_model': 0.4374304783092325, 'eval/latency_0/demon_right_rate_teacher': 0.4922135706340378, 'eval/latency_0/demon_move_f1': 0.7733031957592089, 'eval/latency_0/demon_cc_f1': 0.858813052694534, 'timing/data': 0.004670119960792363, 'timing/model': 1.2063484321115538, 'learning_rate/qwen_vl_interface': 9.191955136995371e-06, 'learning_rate/action_model': 9.191955136995371e-05, 'epoch': 0.41}) +[2026-06-17 08:42:02,788][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_1400_state +[2026-06-17 08:42:02,791][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 08:42:06,322][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_1400_state/model.safetensors +[2026-06-17 08:42:13,643][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_1400_state/optimizer.bin +[2026-06-17 08:42:13,644][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_1400_state/sampler.bin +[2026-06-17 08:42:13,644][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_1400_state/sampler_1.bin +[2026-06-17 08:42:13,645][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_1400_state/random_states_0.pkl +[2026-06-17 08:42:13,713][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1400 +[2026-06-17 08:42:13,713][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 08:42:13,723][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 08:42:18,330][starVLA.training.train_starvla][INFO] - Step 1401, Loss: {'train/loss': 1.0031300783157349, 'train/grad_norm_pre_clip': 11.287391662597656, 'timing/data': 0.004883550922386348, 'timing/model': 1.2023598860250786, 'learning_rate/qwen_vl_interface': 9.190748089732727e-06, 'learning_rate/action_model': 9.190748089732728e-05, 'epoch': 0.41}) +[2026-06-17 08:42:22,989][starVLA.training.train_starvla][INFO] - Step 1402, Loss: {'train/loss': 0.7131104469299316, 'train/grad_norm_pre_clip': 8.671241760253906, 'timing/data': 0.005571946036070585, 'timing/model': 1.1914743189699948, 'learning_rate/qwen_vl_interface': 9.189540225548763e-06, 'learning_rate/action_model': 9.189540225548764e-05, 'epoch': 0.41}) +[2026-06-17 08:42:27,616][starVLA.training.train_starvla][INFO] - Step 1403, Loss: {'train/loss': 0.8176935911178589, 'train/grad_norm_pre_clip': 11.178726196289062, 'timing/data': 0.004989854991436005, 'timing/model': 1.1943999930517748, 'learning_rate/qwen_vl_interface': 9.188331544693872e-06, 'learning_rate/action_model': 9.188331544693873e-05, 'epoch': 0.41}) +[2026-06-17 08:42:32,266][starVLA.training.train_starvla][INFO] - Step 1404, Loss: {'train/loss': 0.86638343334198, 'train/grad_norm_pre_clip': 13.035456657409668, 'timing/data': 0.004592784913256764, 'timing/model': 1.1948082989547402, 'learning_rate/qwen_vl_interface': 9.187122047418614e-06, 'learning_rate/action_model': 9.187122047418614e-05, 'epoch': 0.41}) +[2026-06-17 08:42:36,949][starVLA.training.train_starvla][INFO] - Step 1405, Loss: {'train/loss': 0.6584067344665527, 'train/grad_norm_pre_clip': 9.492781639099121, 'timing/data': 0.005087940953671932, 'timing/model': 1.1977308180648834, 'learning_rate/qwen_vl_interface': 9.18591173397372e-06, 'learning_rate/action_model': 9.185911733973719e-05, 'epoch': 0.41}) +[2026-06-17 08:42:41,667][starVLA.training.train_starvla][INFO] - Step 1406, Loss: {'train/loss': 0.8293733596801758, 'train/grad_norm_pre_clip': 9.48757553100586, 'timing/data': 0.004999471013434231, 'timing/model': 1.2843582381028682, 'learning_rate/qwen_vl_interface': 9.184700604610087e-06, 'learning_rate/action_model': 9.184700604610085e-05, 'epoch': 0.41}) +[2026-06-17 08:42:46,363][starVLA.training.train_starvla][INFO] - Step 1407, Loss: {'train/loss': 0.9567839503288269, 'train/grad_norm_pre_clip': 9.970831871032715, 'timing/data': 0.004946765024214983, 'timing/model': 1.1989108819980174, 'learning_rate/qwen_vl_interface': 9.183488659578786e-06, 'learning_rate/action_model': 9.183488659578785e-05, 'epoch': 0.41}) +[2026-06-17 08:42:51,212][starVLA.training.train_starvla][INFO] - Step 1408, Loss: {'train/loss': 0.7874308824539185, 'train/grad_norm_pre_clip': 10.9143648147583, 'timing/data': 0.0050785839557647705, 'timing/model': 1.199639527942054, 'learning_rate/qwen_vl_interface': 9.18227589913105e-06, 'learning_rate/action_model': 9.18227589913105e-05, 'epoch': 0.41}) +[2026-06-17 08:42:55,979][starVLA.training.train_starvla][INFO] - Step 1409, Loss: {'train/loss': 0.8861265778541565, 'train/grad_norm_pre_clip': 10.483660697937012, 'timing/data': 0.005158681073226035, 'timing/model': 1.2005841829814017, 'learning_rate/qwen_vl_interface': 9.181062323518293e-06, 'learning_rate/action_model': 9.181062323518292e-05, 'epoch': 0.41}) +[2026-06-17 08:43:00,757][starVLA.training.train_starvla][INFO] - Step 1410, Loss: {'train/loss': 1.1165752410888672, 'train/grad_norm_pre_clip': 12.118739128112793, 'timing/data': 0.005178924999199808, 'timing/model': 1.2011125789722428, 'learning_rate/qwen_vl_interface': 9.179847932992084e-06, 'learning_rate/action_model': 9.179847932992083e-05, 'epoch': 0.41}) +[2026-06-17 08:43:05,535][starVLA.training.train_starvla][INFO] - Step 1411, Loss: {'train/loss': 1.0149781703948975, 'train/grad_norm_pre_clip': 14.579780578613281, 'timing/data': 0.005053504952229559, 'timing/model': 1.3321109740063548, 'learning_rate/qwen_vl_interface': 9.17863272780417e-06, 'learning_rate/action_model': 9.178632727804168e-05, 'epoch': 0.42}) +[2026-06-17 08:43:10,174][starVLA.training.train_starvla][INFO] - Step 1412, Loss: {'train/loss': 0.938003420829773, 'train/grad_norm_pre_clip': 10.230303764343262, 'timing/data': 0.0049917560536414385, 'timing/model': 1.2016581740463153, 'learning_rate/qwen_vl_interface': 9.177416708206463e-06, 'learning_rate/action_model': 9.177416708206463e-05, 'epoch': 0.42}) +[2026-06-17 08:43:14,952][starVLA.training.train_starvla][INFO] - Step 1413, Loss: {'train/loss': 0.9387075901031494, 'train/grad_norm_pre_clip': 11.730384826660156, 'timing/data': 0.00500813196413219, 'timing/model': 1.2020151680335402, 'learning_rate/qwen_vl_interface': 9.176199874451049e-06, 'learning_rate/action_model': 9.176199874451048e-05, 'epoch': 0.42}) +[2026-06-17 08:43:19,737][starVLA.training.train_starvla][INFO] - Step 1414, Loss: {'train/loss': 0.7098801136016846, 'train/grad_norm_pre_clip': 28.82895278930664, 'timing/data': 0.005072698928415775, 'timing/model': 1.2050167960114777, 'learning_rate/qwen_vl_interface': 9.174982226790175e-06, 'learning_rate/action_model': 9.174982226790175e-05, 'epoch': 0.42}) +[2026-06-17 08:43:24,423][starVLA.training.train_starvla][INFO] - Step 1415, Loss: {'train/loss': 0.9803892970085144, 'train/grad_norm_pre_clip': 11.697668075561523, 'timing/data': 0.005218982929363847, 'timing/model': 1.2013580850325525, 'learning_rate/qwen_vl_interface': 9.173763765476262e-06, 'learning_rate/action_model': 9.17376376547626e-05, 'epoch': 0.42}) +[2026-06-17 08:43:29,147][starVLA.training.train_starvla][INFO] - Step 1416, Loss: {'train/loss': 1.100605845451355, 'train/grad_norm_pre_clip': 13.210845947265625, 'timing/data': 0.004971324931830168, 'timing/model': 1.2015659149037674, 'learning_rate/qwen_vl_interface': 9.172544490761897e-06, 'learning_rate/action_model': 9.172544490761897e-05, 'epoch': 0.42}) +[2026-06-17 08:43:34,005][starVLA.training.train_starvla][INFO] - Step 1417, Loss: {'train/loss': 0.7785680294036865, 'train/grad_norm_pre_clip': 8.257028579711914, 'timing/data': 0.005190425086766481, 'timing/model': 1.3307949299924076, 'learning_rate/qwen_vl_interface': 9.171324402899841e-06, 'learning_rate/action_model': 9.17132440289984e-05, 'epoch': 0.42}) +[2026-06-17 08:43:38,645][starVLA.training.train_starvla][INFO] - Step 1418, Loss: {'train/loss': 0.7562843561172485, 'train/grad_norm_pre_clip': 8.432291030883789, 'timing/data': 0.0048473740462213755, 'timing/model': 1.2017416529124603, 'learning_rate/qwen_vl_interface': 9.170103502143014e-06, 'learning_rate/action_model': 9.170103502143013e-05, 'epoch': 0.42}) +[2026-06-17 08:43:43,418][starVLA.training.train_starvla][INFO] - Step 1419, Loss: {'train/loss': 1.138353943824768, 'train/grad_norm_pre_clip': 33.08879470825195, 'timing/data': 0.0049704930279403925, 'timing/model': 1.2016823299927637, 'learning_rate/qwen_vl_interface': 9.168881788744513e-06, 'learning_rate/action_model': 9.168881788744513e-05, 'epoch': 0.42}) +[2026-06-17 08:43:48,262][starVLA.training.train_starvla][INFO] - Step 1420, Loss: {'train/loss': 0.8767724633216858, 'train/grad_norm_pre_clip': 12.24997329711914, 'timing/data': 0.005264953011646867, 'timing/model': 1.2049334530020133, 'learning_rate/qwen_vl_interface': 9.1676592629576e-06, 'learning_rate/action_model': 9.167659262957601e-05, 'epoch': 0.42}) +[2026-06-17 08:43:53,044][starVLA.training.train_starvla][INFO] - Step 1421, Loss: {'train/loss': 1.0716060400009155, 'train/grad_norm_pre_clip': 9.274991989135742, 'timing/data': 0.004992707050405443, 'timing/model': 1.2038578999927267, 'learning_rate/qwen_vl_interface': 9.16643592503571e-06, 'learning_rate/action_model': 9.166435925035709e-05, 'epoch': 0.42}) +[2026-06-17 08:43:57,777][starVLA.training.train_starvla][INFO] - Step 1422, Loss: {'train/loss': 0.7989556193351746, 'train/grad_norm_pre_clip': 7.742855548858643, 'timing/data': 0.00465044192969799, 'timing/model': 1.2037081320304424, 'learning_rate/qwen_vl_interface': 9.165211775232436e-06, 'learning_rate/action_model': 9.165211775232435e-05, 'epoch': 0.42}) +[2026-06-17 08:44:02,554][starVLA.training.train_starvla][INFO] - Step 1423, Loss: {'train/loss': 0.7256718277931213, 'train/grad_norm_pre_clip': 8.348787307739258, 'timing/data': 0.004728929954580963, 'timing/model': 1.3315861739683896, 'learning_rate/qwen_vl_interface': 9.163986813801548e-06, 'learning_rate/action_model': 9.163986813801548e-05, 'epoch': 0.42}) +[2026-06-17 08:44:07,216][starVLA.training.train_starvla][INFO] - Step 1424, Loss: {'train/loss': 0.8405324220657349, 'train/grad_norm_pre_clip': 12.980409622192383, 'timing/data': 0.004840275039896369, 'timing/model': 1.2076919060200453, 'learning_rate/qwen_vl_interface': 9.162761040996979e-06, 'learning_rate/action_model': 9.162761040996978e-05, 'epoch': 0.42}) +[2026-06-17 08:44:11,995][starVLA.training.train_starvla][INFO] - Step 1425, Loss: {'train/loss': 0.9297492504119873, 'train/grad_norm_pre_clip': 9.194184303283691, 'timing/data': 0.005281250923871994, 'timing/model': 1.2040157039882615, 'learning_rate/qwen_vl_interface': 9.161534457072839e-06, 'learning_rate/action_model': 9.161534457072839e-05, 'epoch': 0.42}) +[2026-06-17 08:44:16,779][starVLA.training.train_starvla][INFO] - Step 1426, Loss: {'train/loss': 0.939878523349762, 'train/grad_norm_pre_clip': 8.60578727722168, 'timing/data': 0.005123102921061218, 'timing/model': 1.2032386109931394, 'learning_rate/qwen_vl_interface': 9.160307062283397e-06, 'learning_rate/action_model': 9.160307062283397e-05, 'epoch': 0.42}) +[2026-06-17 08:44:21,561][starVLA.training.train_starvla][INFO] - Step 1427, Loss: {'train/loss': 0.8423522114753723, 'train/grad_norm_pre_clip': 10.2499418258667, 'timing/data': 0.00520266208332032, 'timing/model': 1.2042135000228882, 'learning_rate/qwen_vl_interface': 9.159078856883093e-06, 'learning_rate/action_model': 9.159078856883093e-05, 'epoch': 0.42}) +[2026-06-17 08:44:26,363][starVLA.training.train_starvla][INFO] - Step 1428, Loss: {'train/loss': 0.8092182874679565, 'train/grad_norm_pre_clip': 22.282222747802734, 'timing/data': 0.005196493002586067, 'timing/model': 1.332029968034476, 'learning_rate/qwen_vl_interface': 9.157849841126535e-06, 'learning_rate/action_model': 9.157849841126535e-05, 'epoch': 0.42}) +[2026-06-17 08:44:31,020][starVLA.training.train_starvla][INFO] - Step 1429, Loss: {'train/loss': 0.9963973760604858, 'train/grad_norm_pre_clip': 10.557798385620117, 'timing/data': 0.005088911973871291, 'timing/model': 1.2034197160974145, 'learning_rate/qwen_vl_interface': 9.156620015268502e-06, 'learning_rate/action_model': 9.1566200152685e-05, 'epoch': 0.42}) +[2026-06-17 08:44:35,806][starVLA.training.train_starvla][INFO] - Step 1430, Loss: {'train/loss': 0.9920223355293274, 'train/grad_norm_pre_clip': 9.71081829071045, 'timing/data': 0.005308390944264829, 'timing/model': 1.2060770029202104, 'learning_rate/qwen_vl_interface': 9.155389379563934e-06, 'learning_rate/action_model': 9.155389379563933e-05, 'epoch': 0.42}) +[2026-06-17 08:44:40,511][starVLA.training.train_starvla][INFO] - Step 1431, Loss: {'train/loss': 0.6227917075157166, 'train/grad_norm_pre_clip': 6.522750377655029, 'timing/data': 0.005229834001511335, 'timing/model': 1.2055945300962776, 'learning_rate/qwen_vl_interface': 9.154157934267946e-06, 'learning_rate/action_model': 9.154157934267944e-05, 'epoch': 0.42}) +[2026-06-17 08:44:45,298][starVLA.training.train_starvla][INFO] - Step 1432, Loss: {'train/loss': 1.0597800016403198, 'train/grad_norm_pre_clip': 9.775525093078613, 'timing/data': 0.004965594038367271, 'timing/model': 1.2039301679469645, 'learning_rate/qwen_vl_interface': 9.152925679635817e-06, 'learning_rate/action_model': 9.152925679635817e-05, 'epoch': 0.42}) +[2026-06-17 08:44:50,162][starVLA.training.train_starvla][INFO] - Step 1433, Loss: {'train/loss': 0.8525682091712952, 'train/grad_norm_pre_clip': 8.513636589050293, 'timing/data': 0.00466293899808079, 'timing/model': 1.2016850460786372, 'learning_rate/qwen_vl_interface': 9.151692615922995e-06, 'learning_rate/action_model': 9.151692615922994e-05, 'epoch': 0.42}) +[2026-06-17 08:44:54,958][starVLA.training.train_starvla][INFO] - Step 1434, Loss: {'train/loss': 1.140202283859253, 'train/grad_norm_pre_clip': 35.139774322509766, 'timing/data': 0.005151742021553218, 'timing/model': 1.3364527779631317, 'learning_rate/qwen_vl_interface': 9.150458743385095e-06, 'learning_rate/action_model': 9.150458743385094e-05, 'epoch': 0.42}) +[2026-06-17 08:44:59,613][starVLA.training.train_starvla][INFO] - Step 1435, Loss: {'train/loss': 0.7112483382225037, 'train/grad_norm_pre_clip': 10.937684059143066, 'timing/data': 0.005248895031400025, 'timing/model': 1.2040063290623948, 'learning_rate/qwen_vl_interface': 9.1492240622779e-06, 'learning_rate/action_model': 9.1492240622779e-05, 'epoch': 0.42}) +[2026-06-17 08:45:04,411][starVLA.training.train_starvla][INFO] - Step 1436, Loss: {'train/loss': 0.9633365869522095, 'train/grad_norm_pre_clip': 12.015439987182617, 'timing/data': 0.004830557038076222, 'timing/model': 1.2047517839819193, 'learning_rate/qwen_vl_interface': 9.147988572857362e-06, 'learning_rate/action_model': 9.147988572857362e-05, 'epoch': 0.42}) +[2026-06-17 08:45:09,191][starVLA.training.train_starvla][INFO] - Step 1437, Loss: {'train/loss': 1.2592843770980835, 'train/grad_norm_pre_clip': 34.69891357421875, 'timing/data': 0.004825387033633888, 'timing/model': 1.205145713989623, 'learning_rate/qwen_vl_interface': 9.146752275379599e-06, 'learning_rate/action_model': 9.146752275379598e-05, 'epoch': 0.42}) +[2026-06-17 08:45:13,944][starVLA.training.train_starvla][INFO] - Step 1438, Loss: {'train/loss': 0.663419246673584, 'train/grad_norm_pre_clip': 12.534634590148926, 'timing/data': 0.005067578982561827, 'timing/model': 1.2042375369928777, 'learning_rate/qwen_vl_interface': 9.145515170100894e-06, 'learning_rate/action_model': 9.145515170100894e-05, 'epoch': 0.42}) +[2026-06-17 08:45:18,724][starVLA.training.train_starvla][INFO] - Step 1439, Loss: {'train/loss': 1.0208841562271118, 'train/grad_norm_pre_clip': 15.128449440002441, 'timing/data': 0.0052818539552390575, 'timing/model': 1.2047971209976822, 'learning_rate/qwen_vl_interface': 9.144277257277705e-06, 'learning_rate/action_model': 9.144277257277703e-05, 'epoch': 0.42}) +[2026-06-17 08:45:23,517][starVLA.training.train_starvla][INFO] - Step 1440, Loss: {'train/loss': 1.254971981048584, 'train/grad_norm_pre_clip': 16.684114456176758, 'timing/data': 0.005209163064137101, 'timing/model': 1.3391905039316043, 'learning_rate/qwen_vl_interface': 9.14303853716665e-06, 'learning_rate/action_model': 9.143038537166649e-05, 'epoch': 0.42}) +[2026-06-17 08:45:28,174][starVLA.training.train_starvla][INFO] - Step 1441, Loss: {'train/loss': 0.911144495010376, 'train/grad_norm_pre_clip': 11.869119644165039, 'timing/data': 0.004677421995438635, 'timing/model': 1.2066783129703254, 'learning_rate/qwen_vl_interface': 9.141799010024516e-06, 'learning_rate/action_model': 9.141799010024515e-05, 'epoch': 0.42}) +[2026-06-17 08:45:32,972][starVLA.training.train_starvla][INFO] - Step 1442, Loss: {'train/loss': 0.9917735457420349, 'train/grad_norm_pre_clip': 10.087007522583008, 'timing/data': 0.005141856032423675, 'timing/model': 1.2063504460966215, 'learning_rate/qwen_vl_interface': 9.140558676108258e-06, 'learning_rate/action_model': 9.140558676108259e-05, 'epoch': 0.42}) +[2026-06-17 08:45:37,758][starVLA.training.train_starvla][INFO] - Step 1443, Loss: {'train/loss': 0.862731397151947, 'train/grad_norm_pre_clip': 11.014524459838867, 'timing/data': 0.005066815996542573, 'timing/model': 1.2048611650243402, 'learning_rate/qwen_vl_interface': 9.139317535675002e-06, 'learning_rate/action_model': 9.139317535675002e-05, 'epoch': 0.42}) +[2026-06-17 08:45:42,558][starVLA.training.train_starvla][INFO] - Step 1444, Loss: {'train/loss': 1.1934322118759155, 'train/grad_norm_pre_clip': 9.29537582397461, 'timing/data': 0.004924089997075498, 'timing/model': 1.2105731859337538, 'learning_rate/qwen_vl_interface': 9.138075588982035e-06, 'learning_rate/action_model': 9.138075588982036e-05, 'epoch': 0.42}) +[2026-06-17 08:45:47,232][starVLA.training.train_starvla][INFO] - Step 1445, Loss: {'train/loss': 0.7315915822982788, 'train/grad_norm_pre_clip': 9.198155403137207, 'timing/data': 0.004852295038290322, 'timing/model': 1.204059874988161, 'learning_rate/qwen_vl_interface': 9.136832836286816e-06, 'learning_rate/action_model': 9.136832836286816e-05, 'epoch': 0.43}) +[2026-06-17 08:45:51,934][starVLA.training.train_starvla][INFO] - Step 1446, Loss: {'train/loss': 0.7574658393859863, 'train/grad_norm_pre_clip': 6.841588020324707, 'timing/data': 0.004944437998346984, 'timing/model': 1.2474586659809574, 'learning_rate/qwen_vl_interface': 9.135589277846967e-06, 'learning_rate/action_model': 9.135589277846966e-05, 'epoch': 0.43}) +[2026-06-17 08:45:56,603][starVLA.training.train_starvla][INFO] - Step 1447, Loss: {'train/loss': 0.7824515104293823, 'train/grad_norm_pre_clip': 8.555276870727539, 'timing/data': 0.005150074954144657, 'timing/model': 1.2083007509354502, 'learning_rate/qwen_vl_interface': 9.134344913920278e-06, 'learning_rate/action_model': 9.134344913920278e-05, 'epoch': 0.43}) +[2026-06-17 08:46:01,382][starVLA.training.train_starvla][INFO] - Step 1448, Loss: {'train/loss': 0.8209143877029419, 'train/grad_norm_pre_clip': 11.326741218566895, 'timing/data': 0.005138676962815225, 'timing/model': 1.2064475360093638, 'learning_rate/qwen_vl_interface': 9.133099744764711e-06, 'learning_rate/action_model': 9.133099744764711e-05, 'epoch': 0.43}) +[2026-06-17 08:46:06,172][starVLA.training.train_starvla][INFO] - Step 1449, Loss: {'train/loss': 0.9501250982284546, 'train/grad_norm_pre_clip': 9.320528984069824, 'timing/data': 0.004908481030724943, 'timing/model': 1.203833639039658, 'learning_rate/qwen_vl_interface': 9.131853770638386e-06, 'learning_rate/action_model': 9.131853770638386e-05, 'epoch': 0.43}) +[2026-06-17 08:46:10,956][starVLA.training.train_starvla][INFO] - Step 1450, Loss: {'train/loss': 0.7888403534889221, 'train/grad_norm_pre_clip': 28.05371856689453, 'timing/data': 0.005018001073040068, 'timing/model': 1.2055297020124272, 'learning_rate/qwen_vl_interface': 9.130606991799599e-06, 'learning_rate/action_model': 9.130606991799599e-05, 'epoch': 0.43}) +[2026-06-17 08:46:15,741][starVLA.training.train_starvla][INFO] - Step 1451, Loss: {'train/loss': 1.0313390493392944, 'train/grad_norm_pre_clip': 7.5677595138549805, 'timing/data': 0.00499840104021132, 'timing/model': 1.331115919048898, 'learning_rate/qwen_vl_interface': 9.129359408506808e-06, 'learning_rate/action_model': 9.129359408506808e-05, 'epoch': 0.43}) +[2026-06-17 08:46:20,532][starVLA.training.train_starvla][INFO] - Step 1452, Loss: {'train/loss': 1.0888707637786865, 'train/grad_norm_pre_clip': 9.433987617492676, 'timing/data': 0.0049534529680386186, 'timing/model': 1.3388089209329337, 'learning_rate/qwen_vl_interface': 9.128111021018636e-06, 'learning_rate/action_model': 9.128111021018636e-05, 'epoch': 0.43}) +[2026-06-17 08:46:25,204][starVLA.training.train_starvla][INFO] - Step 1453, Loss: {'train/loss': 0.8202002644538879, 'train/grad_norm_pre_clip': 7.647855758666992, 'timing/data': 0.005232007009908557, 'timing/model': 1.2078541449736804, 'learning_rate/qwen_vl_interface': 9.126861829593876e-06, 'learning_rate/action_model': 9.126861829593876e-05, 'epoch': 0.43}) +[2026-06-17 08:46:29,952][starVLA.training.train_starvla][INFO] - Step 1454, Loss: {'train/loss': 0.8723172545433044, 'train/grad_norm_pre_clip': 8.182107925415039, 'timing/data': 0.005094066029414535, 'timing/model': 1.2077712250174955, 'learning_rate/qwen_vl_interface': 9.125611834491487e-06, 'learning_rate/action_model': 9.125611834491486e-05, 'epoch': 0.43}) +[2026-06-17 08:46:34,764][starVLA.training.train_starvla][INFO] - Step 1455, Loss: {'train/loss': 0.8381664752960205, 'train/grad_norm_pre_clip': 8.568127632141113, 'timing/data': 0.005354503984563053, 'timing/model': 1.2056737079983577, 'learning_rate/qwen_vl_interface': 9.124361035970594e-06, 'learning_rate/action_model': 9.124361035970594e-05, 'epoch': 0.43}) +[2026-06-17 08:46:39,569][starVLA.training.train_starvla][INFO] - Step 1456, Loss: {'train/loss': 0.8569368124008179, 'train/grad_norm_pre_clip': 8.773971557617188, 'timing/data': 0.005055728019215167, 'timing/model': 1.2127691260538995, 'learning_rate/qwen_vl_interface': 9.123109434290491e-06, 'learning_rate/action_model': 9.123109434290492e-05, 'epoch': 0.43}) +[2026-06-17 08:46:44,378][starVLA.training.train_starvla][INFO] - Step 1457, Loss: {'train/loss': 0.7776751518249512, 'train/grad_norm_pre_clip': 14.945822715759277, 'timing/data': 0.004867697018198669, 'timing/model': 1.3454048660350963, 'learning_rate/qwen_vl_interface': 9.121857029710632e-06, 'learning_rate/action_model': 9.121857029710633e-05, 'epoch': 0.43}) +[2026-06-17 08:46:49,169][starVLA.training.train_starvla][INFO] - Step 1458, Loss: {'train/loss': 0.7531656622886658, 'train/grad_norm_pre_clip': 12.491186141967773, 'timing/data': 0.004994781920686364, 'timing/model': 1.3296673749573529, 'learning_rate/qwen_vl_interface': 9.120603822490647e-06, 'learning_rate/action_model': 9.120603822490647e-05, 'epoch': 0.43}) +[2026-06-17 08:46:53,849][starVLA.training.train_starvla][INFO] - Step 1459, Loss: {'train/loss': 0.8237811923027039, 'train/grad_norm_pre_clip': 7.509021759033203, 'timing/data': 0.005254367948509753, 'timing/model': 1.2139938819454983, 'learning_rate/qwen_vl_interface': 9.119349812890323e-06, 'learning_rate/action_model': 9.119349812890323e-05, 'epoch': 0.43}) +[2026-06-17 08:46:58,619][starVLA.training.train_starvla][INFO] - Step 1460, Loss: {'train/loss': 0.6544157862663269, 'train/grad_norm_pre_clip': 7.532724380493164, 'timing/data': 0.005214060074649751, 'timing/model': 1.2122103930450976, 'learning_rate/qwen_vl_interface': 9.11809500116962e-06, 'learning_rate/action_model': 9.11809500116962e-05, 'epoch': 0.43}) +[2026-06-17 08:47:03,441][starVLA.training.train_starvla][INFO] - Step 1461, Loss: {'train/loss': 1.030906319618225, 'train/grad_norm_pre_clip': 11.092085838317871, 'timing/data': 0.004767804988659918, 'timing/model': 1.212463929085061, 'learning_rate/qwen_vl_interface': 9.116839387588662e-06, 'learning_rate/action_model': 9.116839387588661e-05, 'epoch': 0.43}) +[2026-06-17 08:47:08,264][starVLA.training.train_starvla][INFO] - Step 1462, Loss: {'train/loss': 0.7997467517852783, 'train/grad_norm_pre_clip': 8.747854232788086, 'timing/data': 0.005919266026467085, 'timing/model': 1.2169789189938456, 'learning_rate/qwen_vl_interface': 9.115582972407735e-06, 'learning_rate/action_model': 9.115582972407736e-05, 'epoch': 0.43}) +[2026-06-17 08:47:13,028][starVLA.training.train_starvla][INFO] - Step 1463, Loss: {'train/loss': 0.8827083110809326, 'train/grad_norm_pre_clip': 7.758939743041992, 'timing/data': 0.005128098069690168, 'timing/model': 1.2928528300253674, 'learning_rate/qwen_vl_interface': 9.114325755887301e-06, 'learning_rate/action_model': 9.114325755887301e-05, 'epoch': 0.43}) +[2026-06-17 08:47:17,715][starVLA.training.train_starvla][INFO] - Step 1464, Loss: {'train/loss': 0.9690366387367249, 'train/grad_norm_pre_clip': 7.273102760314941, 'timing/data': 0.005131045007146895, 'timing/model': 1.2122430709423497, 'learning_rate/qwen_vl_interface': 9.113067738287978e-06, 'learning_rate/action_model': 9.113067738287979e-05, 'epoch': 0.43}) +[2026-06-17 08:47:22,591][starVLA.training.train_starvla][INFO] - Step 1465, Loss: {'train/loss': 0.9812485575675964, 'train/grad_norm_pre_clip': 14.996278762817383, 'timing/data': 0.00508245499804616, 'timing/model': 1.2098137440625578, 'learning_rate/qwen_vl_interface': 9.111808919870557e-06, 'learning_rate/action_model': 9.111808919870557e-05, 'epoch': 0.43}) +[2026-06-17 08:47:27,404][starVLA.training.train_starvla][INFO] - Step 1466, Loss: {'train/loss': 1.1497658491134644, 'train/grad_norm_pre_clip': 10.066954612731934, 'timing/data': 0.004858447937294841, 'timing/model': 1.2091555279912427, 'learning_rate/qwen_vl_interface': 9.11054930089599e-06, 'learning_rate/action_model': 9.110549300895991e-05, 'epoch': 0.43}) +[2026-06-17 08:47:32,203][starVLA.training.train_starvla][INFO] - Step 1467, Loss: {'train/loss': 0.8741794228553772, 'train/grad_norm_pre_clip': 7.906120777130127, 'timing/data': 0.00506983301602304, 'timing/model': 1.2084720629500225, 'learning_rate/qwen_vl_interface': 9.109288881625402e-06, 'learning_rate/action_model': 9.109288881625402e-05, 'epoch': 0.43}) +[2026-06-17 08:47:36,993][starVLA.training.train_starvla][INFO] - Step 1468, Loss: {'train/loss': 0.5827561020851135, 'train/grad_norm_pre_clip': 6.284041881561279, 'timing/data': 0.004825055017136037, 'timing/model': 1.3327599019976333, 'learning_rate/qwen_vl_interface': 9.108027662320073e-06, 'learning_rate/action_model': 9.108027662320073e-05, 'epoch': 0.43}) +[2026-06-17 08:47:41,730][starVLA.training.train_starvla][INFO] - Step 1469, Loss: {'train/loss': 0.9626305103302002, 'train/grad_norm_pre_clip': 8.959468841552734, 'timing/data': 0.005270116031169891, 'timing/model': 1.278729246929288, 'learning_rate/qwen_vl_interface': 9.10676564324146e-06, 'learning_rate/action_model': 9.10676564324146e-05, 'epoch': 0.43}) +[2026-06-17 08:47:46,392][starVLA.training.train_starvla][INFO] - Step 1470, Loss: {'train/loss': 1.113041639328003, 'train/grad_norm_pre_clip': 11.81753158569336, 'timing/data': 0.004946849076077342, 'timing/model': 1.2115261600119993, 'learning_rate/qwen_vl_interface': 9.105502824651179e-06, 'learning_rate/action_model': 9.105502824651178e-05, 'epoch': 0.43}) +[2026-06-17 08:47:51,196][starVLA.training.train_starvla][INFO] - Step 1471, Loss: {'train/loss': 0.8678041100502014, 'train/grad_norm_pre_clip': 103.13716888427734, 'timing/data': 0.004726625978946686, 'timing/model': 1.2066008689580485, 'learning_rate/qwen_vl_interface': 9.104239206811012e-06, 'learning_rate/action_model': 9.104239206811011e-05, 'epoch': 0.43}) +[2026-06-17 08:47:56,023][starVLA.training.train_starvla][INFO] - Step 1472, Loss: {'train/loss': 0.643186628818512, 'train/grad_norm_pre_clip': 6.5636467933654785, 'timing/data': 0.00519217096734792, 'timing/model': 1.2166965879732743, 'learning_rate/qwen_vl_interface': 9.102974789982911e-06, 'learning_rate/action_model': 9.102974789982912e-05, 'epoch': 0.43}) +[2026-06-17 08:48:00,812][starVLA.training.train_starvla][INFO] - Step 1473, Loss: {'train/loss': 0.8303898572921753, 'train/grad_norm_pre_clip': 11.127671241760254, 'timing/data': 0.005048788036219776, 'timing/model': 1.2062460830202326, 'learning_rate/qwen_vl_interface': 9.10170957442899e-06, 'learning_rate/action_model': 9.101709574428991e-05, 'epoch': 0.43}) +[2026-06-17 08:48:05,602][starVLA.training.train_starvla][INFO] - Step 1474, Loss: {'train/loss': 0.7670249342918396, 'train/grad_norm_pre_clip': 6.7416090965271, 'timing/data': 0.005205065943300724, 'timing/model': 1.204885974060744, 'learning_rate/qwen_vl_interface': 9.10044356041153e-06, 'learning_rate/action_model': 9.10044356041153e-05, 'epoch': 0.43}) +[2026-06-17 08:48:10,358][starVLA.training.train_starvla][INFO] - Step 1475, Loss: {'train/loss': 0.6586663722991943, 'train/grad_norm_pre_clip': 9.042728424072266, 'timing/data': 0.004724515951238573, 'timing/model': 1.2996236679609865, 'learning_rate/qwen_vl_interface': 9.099176748192976e-06, 'learning_rate/action_model': 9.099176748192976e-05, 'epoch': 0.43}) +[2026-06-17 08:48:15,031][starVLA.training.train_starvla][INFO] - Step 1476, Loss: {'train/loss': 0.9707996249198914, 'train/grad_norm_pre_clip': 9.570014953613281, 'timing/data': 0.004982099053449929, 'timing/model': 1.214191954000853, 'learning_rate/qwen_vl_interface': 9.097909138035943e-06, 'learning_rate/action_model': 9.097909138035941e-05, 'epoch': 0.43}) +[2026-06-17 08:48:19,823][starVLA.training.train_starvla][INFO] - Step 1477, Loss: {'train/loss': 0.9498374462127686, 'train/grad_norm_pre_clip': 10.531777381896973, 'timing/data': 0.005410877987742424, 'timing/model': 1.2046611639671028, 'learning_rate/qwen_vl_interface': 9.096640730203202e-06, 'learning_rate/action_model': 9.096640730203202e-05, 'epoch': 0.43}) +[2026-06-17 08:48:24,612][starVLA.training.train_starvla][INFO] - Step 1478, Loss: {'train/loss': 0.5821656584739685, 'train/grad_norm_pre_clip': 20.721940994262695, 'timing/data': 0.004879178013652563, 'timing/model': 1.2057904379907995, 'learning_rate/qwen_vl_interface': 9.095371524957698e-06, 'learning_rate/action_model': 9.095371524957698e-05, 'epoch': 0.43}) +[2026-06-17 08:48:29,400][starVLA.training.train_starvla][INFO] - Step 1479, Loss: {'train/loss': 0.8179858326911926, 'train/grad_norm_pre_clip': 8.065515518188477, 'timing/data': 0.0053231349447742105, 'timing/model': 1.206117296940647, 'learning_rate/qwen_vl_interface': 9.094101522562541e-06, 'learning_rate/action_model': 9.09410152256254e-05, 'epoch': 0.44}) +[2026-06-17 08:48:34,196][starVLA.training.train_starvla][INFO] - Step 1480, Loss: {'train/loss': 0.9489162564277649, 'train/grad_norm_pre_clip': 9.562505722045898, 'timing/data': 0.004905026056803763, 'timing/model': 1.3424639769364148, 'learning_rate/qwen_vl_interface': 9.092830723281e-06, 'learning_rate/action_model': 9.092830723281e-05, 'epoch': 0.44}) +[2026-06-17 08:48:38,853][starVLA.training.train_starvla][INFO] - Step 1481, Loss: {'train/loss': 0.9356681704521179, 'train/grad_norm_pre_clip': 12.179215431213379, 'timing/data': 0.004335941979661584, 'timing/model': 1.2037146799266338, 'learning_rate/qwen_vl_interface': 9.091559127376519e-06, 'learning_rate/action_model': 9.091559127376517e-05, 'epoch': 0.44}) +[2026-06-17 08:48:43,650][starVLA.training.train_starvla][INFO] - Step 1482, Loss: {'train/loss': 0.731780469417572, 'train/grad_norm_pre_clip': 8.095592498779297, 'timing/data': 0.004970782087184489, 'timing/model': 1.2044823130127043, 'learning_rate/qwen_vl_interface': 9.090286735112693e-06, 'learning_rate/action_model': 9.090286735112693e-05, 'epoch': 0.44}) +[2026-06-17 08:48:48,442][starVLA.training.train_starvla][INFO] - Step 1483, Loss: {'train/loss': 1.0768592357635498, 'train/grad_norm_pre_clip': 9.325448989868164, 'timing/data': 0.005116803920827806, 'timing/model': 1.2054062690585852, 'learning_rate/qwen_vl_interface': 9.089013546753296e-06, 'learning_rate/action_model': 9.089013546753295e-05, 'epoch': 0.44}) +[2026-06-17 08:48:53,151][starVLA.training.train_starvla][INFO] - Step 1484, Loss: {'train/loss': 0.9402099251747131, 'train/grad_norm_pre_clip': 11.106223106384277, 'timing/data': 0.005072594038210809, 'timing/model': 1.2081659189425409, 'learning_rate/qwen_vl_interface': 9.08773956256226e-06, 'learning_rate/action_model': 9.087739562562259e-05, 'epoch': 0.44}) +[2026-06-17 08:48:57,905][starVLA.training.train_starvla][INFO] - Step 1485, Loss: {'train/loss': 0.9715659618377686, 'train/grad_norm_pre_clip': 9.196051597595215, 'timing/data': 0.00513257400598377, 'timing/model': 1.2047249019378796, 'learning_rate/qwen_vl_interface': 9.086464782803682e-06, 'learning_rate/action_model': 9.086464782803682e-05, 'epoch': 0.44}) +[2026-06-17 08:49:02,705][starVLA.training.train_starvla][INFO] - Step 1486, Loss: {'train/loss': 0.7863428592681885, 'train/grad_norm_pre_clip': 9.927620887756348, 'timing/data': 0.004922681953758001, 'timing/model': 1.3422562340274453, 'learning_rate/qwen_vl_interface': 9.085189207741827e-06, 'learning_rate/action_model': 9.085189207741827e-05, 'epoch': 0.44}) +[2026-06-17 08:49:07,377][starVLA.training.train_starvla][INFO] - Step 1487, Loss: {'train/loss': 0.7410614490509033, 'train/grad_norm_pre_clip': 7.549377918243408, 'timing/data': 0.0050817669834941626, 'timing/model': 1.2063656620448455, 'learning_rate/qwen_vl_interface': 9.083912837641123e-06, 'learning_rate/action_model': 9.083912837641123e-05, 'epoch': 0.44}) +[2026-06-17 08:49:12,178][starVLA.training.train_starvla][INFO] - Step 1488, Loss: {'train/loss': 0.741369903087616, 'train/grad_norm_pre_clip': 11.427281379699707, 'timing/data': 0.005049764993600547, 'timing/model': 1.2050906689837575, 'learning_rate/qwen_vl_interface': 9.082635672766162e-06, 'learning_rate/action_model': 9.082635672766161e-05, 'epoch': 0.44}) +[2026-06-17 08:49:16,967][starVLA.training.train_starvla][INFO] - Step 1489, Loss: {'train/loss': 1.041098713874817, 'train/grad_norm_pre_clip': 10.796895027160645, 'timing/data': 0.005424674949608743, 'timing/model': 1.207204052945599, 'learning_rate/qwen_vl_interface': 9.081357713381702e-06, 'learning_rate/action_model': 9.081357713381701e-05, 'epoch': 0.44}) +[2026-06-17 08:49:21,749][starVLA.training.train_starvla][INFO] - Step 1490, Loss: {'train/loss': 1.0584630966186523, 'train/grad_norm_pre_clip': 9.337634086608887, 'timing/data': 0.004963965038768947, 'timing/model': 1.2079329050611705, 'learning_rate/qwen_vl_interface': 9.080078959752664e-06, 'learning_rate/action_model': 9.080078959752664e-05, 'epoch': 0.44}) +[2026-06-17 08:49:26,496][starVLA.training.train_starvla][INFO] - Step 1491, Loss: {'train/loss': 0.9985615611076355, 'train/grad_norm_pre_clip': 9.557108879089355, 'timing/data': 0.005364684038795531, 'timing/model': 1.2052374189952388, 'learning_rate/qwen_vl_interface': 9.078799412144137e-06, 'learning_rate/action_model': 9.078799412144137e-05, 'epoch': 0.44}) +[2026-06-17 08:49:31,276][starVLA.training.train_starvla][INFO] - Step 1492, Loss: {'train/loss': 0.9698383808135986, 'train/grad_norm_pre_clip': 25.027849197387695, 'timing/data': 0.005451370030641556, 'timing/model': 1.3228290400002152, 'learning_rate/qwen_vl_interface': 9.077519070821373e-06, 'learning_rate/action_model': 9.077519070821373e-05, 'epoch': 0.44}) +[2026-06-17 08:49:35,942][starVLA.training.train_starvla][INFO] - Step 1493, Loss: {'train/loss': 0.9640282392501831, 'train/grad_norm_pre_clip': 9.59299373626709, 'timing/data': 0.005032640998251736, 'timing/model': 1.2063235739478841, 'learning_rate/qwen_vl_interface': 9.076237936049787e-06, 'learning_rate/action_model': 9.076237936049786e-05, 'epoch': 0.44}) +[2026-06-17 08:49:40,728][starVLA.training.train_starvla][INFO] - Step 1494, Loss: {'train/loss': 0.7231737375259399, 'train/grad_norm_pre_clip': 9.788902282714844, 'timing/data': 0.005082478979602456, 'timing/model': 1.2058457729872316, 'learning_rate/qwen_vl_interface': 9.074956008094958e-06, 'learning_rate/action_model': 9.074956008094958e-05, 'epoch': 0.44}) +[2026-06-17 08:49:45,512][starVLA.training.train_starvla][INFO] - Step 1495, Loss: {'train/loss': 0.8426201343536377, 'train/grad_norm_pre_clip': 7.909496784210205, 'timing/data': 0.004926334018819034, 'timing/model': 1.2051516380161047, 'learning_rate/qwen_vl_interface': 9.073673287222634e-06, 'learning_rate/action_model': 9.073673287222633e-05, 'epoch': 0.44}) +[2026-06-17 08:49:50,294][starVLA.training.train_starvla][INFO] - Step 1496, Loss: {'train/loss': 0.8636517524719238, 'train/grad_norm_pre_clip': 10.29857349395752, 'timing/data': 0.00477219105232507, 'timing/model': 1.2041236819932237, 'learning_rate/qwen_vl_interface': 9.072389773698723e-06, 'learning_rate/action_model': 9.072389773698723e-05, 'epoch': 0.44}) +[2026-06-17 08:49:55,075][starVLA.training.train_starvla][INFO] - Step 1497, Loss: {'train/loss': 0.8001429438591003, 'train/grad_norm_pre_clip': 22.333858489990234, 'timing/data': 0.005238238023594022, 'timing/model': 1.332514496985823, 'learning_rate/qwen_vl_interface': 9.071105467789299e-06, 'learning_rate/action_model': 9.071105467789298e-05, 'epoch': 0.44}) +[2026-06-17 08:49:59,730][starVLA.training.train_starvla][INFO] - Step 1498, Loss: {'train/loss': 1.2000218629837036, 'train/grad_norm_pre_clip': 11.520096778869629, 'timing/data': 0.004801834002137184, 'timing/model': 1.204369576065801, 'learning_rate/qwen_vl_interface': 9.0698203697606e-06, 'learning_rate/action_model': 9.0698203697606e-05, 'epoch': 0.44}) +[2026-06-17 08:50:04,515][starVLA.training.train_starvla][INFO] - Step 1499, Loss: {'train/loss': 0.6524394154548645, 'train/grad_norm_pre_clip': 8.026850700378418, 'timing/data': 0.004832786042243242, 'timing/model': 1.204535315046087, 'learning_rate/qwen_vl_interface': 9.068534479879026e-06, 'learning_rate/action_model': 9.068534479879024e-05, 'epoch': 0.44}) +[2026-06-17 09:00:13,001][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1500 +[2026-06-17 09:00:13,002][starVLA.training.train_starvla][INFO] - Step 1500, Loss: {'train/loss': 0.8214970827102661, 'train/grad_norm_pre_clip': 22.18044090270996, 'eval/loss': 0.9012959003448486, 'eval/latency_0/loss': 0.9012959003448486, 'eval/demon_fire_precision': 0.844208211143695, 'eval/demon_fire_recall': 0.9563953488372093, 'eval/demon_fire_f1': 0.8968068535825545, 'eval/demon_fire_rate_model': 0.7586206896551724, 'eval/demon_fire_rate_teacher': 0.6696329254727474, 'eval/demon_left_precision': 0.7349042709867453, 'eval/demon_left_recall': 0.8481586402266289, 'eval/demon_left_f1': 0.7874802735402421, 'eval/demon_left_rate_model': 0.5664627363737486, 'eval/demon_left_rate_teacher': 0.49082313681868744, 'eval/demon_right_precision': 0.822963438101347, 'eval/demon_right_recall': 0.7248587570621469, 'eval/demon_right_f1': 0.770802042655452, 'eval/demon_right_rate_model': 0.4335372636262514, 'eval/demon_right_rate_teacher': 0.4922135706340378, 'eval/demon_move_f1': 0.7791411580978471, 'eval/demon_cc_f1': 0.8579372168350012, 'eval/latency_0/demon_fire_precision': 0.844208211143695, 'eval/latency_0/demon_fire_recall': 0.9563953488372093, 'eval/latency_0/demon_fire_f1': 0.8968068535825545, 'eval/latency_0/demon_fire_rate_model': 0.7586206896551724, 'eval/latency_0/demon_fire_rate_teacher': 0.6696329254727474, 'eval/latency_0/demon_left_precision': 0.7349042709867453, 'eval/latency_0/demon_left_recall': 0.8481586402266289, 'eval/latency_0/demon_left_f1': 0.7874802735402421, 'eval/latency_0/demon_left_rate_model': 0.5664627363737486, 'eval/latency_0/demon_left_rate_teacher': 0.49082313681868744, 'eval/latency_0/demon_right_precision': 0.822963438101347, 'eval/latency_0/demon_right_recall': 0.7248587570621469, 'eval/latency_0/demon_right_f1': 0.770802042655452, 'eval/latency_0/demon_right_rate_model': 0.4335372636262514, 'eval/latency_0/demon_right_rate_teacher': 0.4922135706340378, 'eval/latency_0/demon_move_f1': 0.7791411580978471, 'eval/latency_0/demon_cc_f1': 0.8579372168350012, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 6878.0, 'rl_games_eval/mid_train/mean_length': 3600.0, 'rl_games_eval/mid_train/std_reward': 524.1908049556001, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/macro_mean_reward': 6878.0, 'rl_games_eval/mid_train/macro_mean_length': 3600.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/demon_attack__latency_0/mean_reward': 6878.0, 'rl_games_eval/mid_train/demon_attack__latency_0/mean_length': 3600.0, 'rl_games_eval/mid_train/demon_attack__latency_0/std_reward': 524.1908049556001, 'rl_games_eval/mid_train/demon_attack__latency_0/std_length': 0.0, 'rl_games_eval/mid_train/demon_attack__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 6878.0, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.005146245006471872, 'timing/model': 1.2050261630211025, 'learning_rate/qwen_vl_interface': 9.067247798411147e-06, 'learning_rate/action_model': 9.067247798411147e-05, 'epoch': 0.44}) +[2026-06-17 09:00:13,002][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_1500_state +[2026-06-17 09:00:13,005][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 09:00:16,577][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_1500_state/model.safetensors +[2026-06-17 09:00:24,311][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_1500_state/optimizer.bin +[2026-06-17 09:00:24,311][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_1500_state/sampler.bin +[2026-06-17 09:00:24,312][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_1500_state/sampler_1.bin +[2026-06-17 09:00:24,312][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_1500_state/random_states_0.pkl +[2026-06-17 09:00:24,353][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1500 +[2026-06-17 09:00:24,354][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 09:00:24,364][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 09:00:29,084][starVLA.training.train_starvla][INFO] - Step 1501, Loss: {'train/loss': 0.6334917545318604, 'train/grad_norm_pre_clip': 6.950628757476807, 'timing/data': 0.005054090986959636, 'timing/model': 1.188769314903766, 'learning_rate/qwen_vl_interface': 9.065960325623692e-06, 'learning_rate/action_model': 9.065960325623692e-05, 'epoch': 0.44}) +[2026-06-17 09:00:33,858][starVLA.training.train_starvla][INFO] - Step 1502, Loss: {'train/loss': 0.9454114437103271, 'train/grad_norm_pre_clip': 17.643627166748047, 'timing/data': 0.004487881087698042, 'timing/model': 1.1959221179131418, 'learning_rate/qwen_vl_interface': 9.064672061783555e-06, 'learning_rate/action_model': 9.064672061783554e-05, 'epoch': 0.44}) +[2026-06-17 09:00:38,555][starVLA.training.train_starvla][INFO] - Step 1503, Loss: {'train/loss': 1.1048669815063477, 'train/grad_norm_pre_clip': 9.941200256347656, 'timing/data': 0.006405748892575502, 'timing/model': 1.206801739986986, 'learning_rate/qwen_vl_interface': 9.063383007157795e-06, 'learning_rate/action_model': 9.063383007157796e-05, 'epoch': 0.44}) +[2026-06-17 09:00:43,273][starVLA.training.train_starvla][INFO] - Step 1504, Loss: {'train/loss': 0.9518879055976868, 'train/grad_norm_pre_clip': 9.4464750289917, 'timing/data': 0.004489882965572178, 'timing/model': 1.286836149985902, 'learning_rate/qwen_vl_interface': 9.062093162013635e-06, 'learning_rate/action_model': 9.062093162013634e-05, 'epoch': 0.44}) +[2026-06-17 09:00:47,898][starVLA.training.train_starvla][INFO] - Step 1505, Loss: {'train/loss': 0.6329221725463867, 'train/grad_norm_pre_clip': 6.9795966148376465, 'timing/data': 0.00498662400059402, 'timing/model': 1.19802508095745, 'learning_rate/qwen_vl_interface': 9.06080252661846e-06, 'learning_rate/action_model': 9.060802526618458e-05, 'epoch': 0.44}) +[2026-06-17 09:00:52,631][starVLA.training.train_starvla][INFO] - Step 1506, Loss: {'train/loss': 0.8090838193893433, 'train/grad_norm_pre_clip': 10.992593765258789, 'timing/data': 0.004816107917577028, 'timing/model': 1.193024224950932, 'learning_rate/qwen_vl_interface': 9.059511101239818e-06, 'learning_rate/action_model': 9.059511101239818e-05, 'epoch': 0.44}) +[2026-06-17 09:00:57,394][starVLA.training.train_starvla][INFO] - Step 1507, Loss: {'train/loss': 0.7778315544128418, 'train/grad_norm_pre_clip': 21.80083656311035, 'timing/data': 0.004743375000543892, 'timing/model': 1.203693110961467, 'learning_rate/qwen_vl_interface': 9.058218886145427e-06, 'learning_rate/action_model': 9.058218886145427e-05, 'epoch': 0.44}) +[2026-06-17 09:01:02,158][starVLA.training.train_starvla][INFO] - Step 1508, Loss: {'train/loss': 1.0180855989456177, 'train/grad_norm_pre_clip': 7.742976188659668, 'timing/data': 0.004843765986151993, 'timing/model': 1.1958831360097975, 'learning_rate/qwen_vl_interface': 9.056925881603165e-06, 'learning_rate/action_model': 9.056925881603166e-05, 'epoch': 0.44}) +[2026-06-17 09:01:06,909][starVLA.training.train_starvla][INFO] - Step 1509, Loss: {'train/loss': 0.7486599087715149, 'train/grad_norm_pre_clip': 10.73369312286377, 'timing/data': 0.004683541017584503, 'timing/model': 1.329907919978723, 'learning_rate/qwen_vl_interface': 9.05563208788107e-06, 'learning_rate/action_model': 9.05563208788107e-05, 'epoch': 0.44}) +[2026-06-17 09:01:11,551][starVLA.training.train_starvla][INFO] - Step 1510, Loss: {'train/loss': 0.8492558002471924, 'train/grad_norm_pre_clip': 7.222179889678955, 'timing/data': 0.004194737994112074, 'timing/model': 1.1990500580286607, 'learning_rate/qwen_vl_interface': 9.054337505247349e-06, 'learning_rate/action_model': 9.054337505247348e-05, 'epoch': 0.44}) +[2026-06-17 09:01:16,326][starVLA.training.train_starvla][INFO] - Step 1511, Loss: {'train/loss': 0.8366256952285767, 'train/grad_norm_pre_clip': 11.73737621307373, 'timing/data': 0.004861166933551431, 'timing/model': 1.2029412729898468, 'learning_rate/qwen_vl_interface': 9.05304213397037e-06, 'learning_rate/action_model': 9.053042133970369e-05, 'epoch': 0.44}) +[2026-06-17 09:01:21,089][starVLA.training.train_starvla][INFO] - Step 1512, Loss: {'train/loss': 0.842846691608429, 'train/grad_norm_pre_clip': 15.364177703857422, 'timing/data': 0.00460219697561115, 'timing/model': 1.2002793930005282, 'learning_rate/qwen_vl_interface': 9.051745974318663e-06, 'learning_rate/action_model': 9.051745974318662e-05, 'epoch': 0.44}) +[2026-06-17 09:01:25,772][starVLA.training.train_starvla][INFO] - Step 1513, Loss: {'train/loss': 0.8363703489303589, 'train/grad_norm_pre_clip': 9.076611518859863, 'timing/data': 0.0046779250260442495, 'timing/model': 1.200990843004547, 'learning_rate/qwen_vl_interface': 9.050449026560923e-06, 'learning_rate/action_model': 9.050449026560924e-05, 'epoch': 0.45}) +[2026-06-17 09:01:30,507][starVLA.training.train_starvla][INFO] - Step 1514, Loss: {'train/loss': 0.8720493912696838, 'train/grad_norm_pre_clip': 13.273229598999023, 'timing/data': 0.004722187994047999, 'timing/model': 1.2115633260691538, 'learning_rate/qwen_vl_interface': 9.049151290966013e-06, 'learning_rate/action_model': 9.049151290966012e-05, 'epoch': 0.45}) +[2026-06-17 09:01:35,277][starVLA.training.train_starvla][INFO] - Step 1515, Loss: {'train/loss': 0.9683247208595276, 'train/grad_norm_pre_clip': 13.477293968200684, 'timing/data': 0.004337889025919139, 'timing/model': 1.3326138750417158, 'learning_rate/qwen_vl_interface': 9.04785276780295e-06, 'learning_rate/action_model': 9.04785276780295e-05, 'epoch': 0.45}) +[2026-06-17 09:01:39,921][starVLA.training.train_starvla][INFO] - Step 1516, Loss: {'train/loss': 0.953783392906189, 'train/grad_norm_pre_clip': 12.240616798400879, 'timing/data': 0.004766519065015018, 'timing/model': 1.20052246900741, 'learning_rate/qwen_vl_interface': 9.046553457340924e-06, 'learning_rate/action_model': 9.046553457340924e-05, 'epoch': 0.45}) +[2026-06-17 09:01:44,742][starVLA.training.train_starvla][INFO] - Step 1517, Loss: {'train/loss': 0.9784106016159058, 'train/grad_norm_pre_clip': 13.078835487365723, 'timing/data': 0.004732009954750538, 'timing/model': 1.202127347001806, 'learning_rate/qwen_vl_interface': 9.04525335984928e-06, 'learning_rate/action_model': 9.04525335984928e-05, 'epoch': 0.45}) +[2026-06-17 09:01:49,566][starVLA.training.train_starvla][INFO] - Step 1518, Loss: {'train/loss': 0.8021878600120544, 'train/grad_norm_pre_clip': 13.29804515838623, 'timing/data': 0.004872728022746742, 'timing/model': 1.2046145600033924, 'learning_rate/qwen_vl_interface': 9.043952475597532e-06, 'learning_rate/action_model': 9.043952475597532e-05, 'epoch': 0.45}) +[2026-06-17 09:01:54,344][starVLA.training.train_starvla][INFO] - Step 1519, Loss: {'train/loss': 0.9212135076522827, 'train/grad_norm_pre_clip': 8.117542266845703, 'timing/data': 0.0048312179278582335, 'timing/model': 1.20222836593166, 'learning_rate/qwen_vl_interface': 9.04265080485535e-06, 'learning_rate/action_model': 9.042650804855351e-05, 'epoch': 0.45}) +[2026-06-17 09:01:59,096][starVLA.training.train_starvla][INFO] - Step 1520, Loss: {'train/loss': 0.9535501003265381, 'train/grad_norm_pre_clip': 10.266493797302246, 'timing/data': 0.0046903390903025866, 'timing/model': 1.2068045559572056, 'learning_rate/qwen_vl_interface': 9.041348347892578e-06, 'learning_rate/action_model': 9.041348347892579e-05, 'epoch': 0.45}) +[2026-06-17 09:02:03,870][starVLA.training.train_starvla][INFO] - Step 1521, Loss: {'train/loss': 0.869605541229248, 'train/grad_norm_pre_clip': 8.133959770202637, 'timing/data': 0.004879532963968813, 'timing/model': 1.3248830779921263, 'learning_rate/qwen_vl_interface': 9.040045104979213e-06, 'learning_rate/action_model': 9.040045104979214e-05, 'epoch': 0.45}) +[2026-06-17 09:02:08,513][starVLA.training.train_starvla][INFO] - Step 1522, Loss: {'train/loss': 0.8064830899238586, 'train/grad_norm_pre_clip': 8.901581764221191, 'timing/data': 0.004959134967066348, 'timing/model': 1.200730766984634, 'learning_rate/qwen_vl_interface': 9.03874107638542e-06, 'learning_rate/action_model': 9.038741076385419e-05, 'epoch': 0.45}) +[2026-06-17 09:02:13,294][starVLA.training.train_starvla][INFO] - Step 1523, Loss: {'train/loss': 1.1257638931274414, 'train/grad_norm_pre_clip': 10.63211727142334, 'timing/data': 0.004533723113127053, 'timing/model': 1.2047444679774344, 'learning_rate/qwen_vl_interface': 9.037436262381525e-06, 'learning_rate/action_model': 9.037436262381525e-05, 'epoch': 0.45}) +[2026-06-17 09:02:18,076][starVLA.training.train_starvla][INFO] - Step 1524, Loss: {'train/loss': 0.6967341303825378, 'train/grad_norm_pre_clip': 9.598382949829102, 'timing/data': 0.004606745089404285, 'timing/model': 1.2037090209778398, 'learning_rate/qwen_vl_interface': 9.036130663238018e-06, 'learning_rate/action_model': 9.036130663238017e-05, 'epoch': 0.45}) +[2026-06-17 09:02:22,855][starVLA.training.train_starvla][INFO] - Step 1525, Loss: {'train/loss': 0.7337351441383362, 'train/grad_norm_pre_clip': 16.280935287475586, 'timing/data': 0.004601776949129999, 'timing/model': 1.2036972659407184, 'learning_rate/qwen_vl_interface': 9.034824279225547e-06, 'learning_rate/action_model': 9.034824279225547e-05, 'epoch': 0.45}) +[2026-06-17 09:02:27,648][starVLA.training.train_starvla][INFO] - Step 1526, Loss: {'train/loss': 0.7717148661613464, 'train/grad_norm_pre_clip': 7.818522930145264, 'timing/data': 0.004304397967644036, 'timing/model': 1.339018450002186, 'learning_rate/qwen_vl_interface': 9.033517110614933e-06, 'learning_rate/action_model': 9.033517110614933e-05, 'epoch': 0.45}) +[2026-06-17 09:02:32,343][starVLA.training.train_starvla][INFO] - Step 1527, Loss: {'train/loss': 0.8388509750366211, 'train/grad_norm_pre_clip': 10.846044540405273, 'timing/data': 0.004975695046596229, 'timing/model': 1.205990748014301, 'learning_rate/qwen_vl_interface': 9.032209157677151e-06, 'learning_rate/action_model': 9.032209157677151e-05, 'epoch': 0.45}) +[2026-06-17 09:02:37,130][starVLA.training.train_starvla][INFO] - Step 1528, Loss: {'train/loss': 1.006890892982483, 'train/grad_norm_pre_clip': 10.418680191040039, 'timing/data': 0.004525933996774256, 'timing/model': 1.203547776909545, 'learning_rate/qwen_vl_interface': 9.03090042068334e-06, 'learning_rate/action_model': 9.030900420683339e-05, 'epoch': 0.45}) +[2026-06-17 09:02:41,845][starVLA.training.train_starvla][INFO] - Step 1529, Loss: {'train/loss': 0.7269681096076965, 'train/grad_norm_pre_clip': 8.218451499938965, 'timing/data': 0.004922124091535807, 'timing/model': 1.2081836119759828, 'learning_rate/qwen_vl_interface': 9.029590899904803e-06, 'learning_rate/action_model': 9.029590899904803e-05, 'epoch': 0.45}) +[2026-06-17 09:02:46,587][starVLA.training.train_starvla][INFO] - Step 1530, Loss: {'train/loss': 0.8993501663208008, 'train/grad_norm_pre_clip': 10.44383430480957, 'timing/data': 0.004892839002422988, 'timing/model': 1.2078573019243777, 'learning_rate/qwen_vl_interface': 9.028280595613006e-06, 'learning_rate/action_model': 9.028280595613005e-05, 'epoch': 0.45}) +[2026-06-17 09:02:51,373][starVLA.training.train_starvla][INFO] - Step 1531, Loss: {'train/loss': 0.721465528011322, 'train/grad_norm_pre_clip': 11.9534330368042, 'timing/data': 0.004692935035564005, 'timing/model': 1.2049836229998618, 'learning_rate/qwen_vl_interface': 9.026969508079575e-06, 'learning_rate/action_model': 9.026969508079575e-05, 'epoch': 0.45}) +[2026-06-17 09:02:56,162][starVLA.training.train_starvla][INFO] - Step 1532, Loss: {'train/loss': 0.9161233305931091, 'train/grad_norm_pre_clip': 11.448376655578613, 'timing/data': 0.004498819005675614, 'timing/model': 1.3420446349773556, 'learning_rate/qwen_vl_interface': 9.025657637576302e-06, 'learning_rate/action_model': 9.025657637576302e-05, 'epoch': 0.45}) +[2026-06-17 09:03:00,837][starVLA.training.train_starvla][INFO] - Step 1533, Loss: {'train/loss': 0.8714982271194458, 'train/grad_norm_pre_clip': 9.823514938354492, 'timing/data': 0.004828384029679, 'timing/model': 1.206579390913248, 'learning_rate/qwen_vl_interface': 9.024344984375138e-06, 'learning_rate/action_model': 9.024344984375137e-05, 'epoch': 0.45}) +[2026-06-17 09:03:05,631][starVLA.training.train_starvla][INFO] - Step 1534, Loss: {'train/loss': 1.0221974849700928, 'train/grad_norm_pre_clip': 9.394476890563965, 'timing/data': 0.004428932094015181, 'timing/model': 1.2045223489403725, 'learning_rate/qwen_vl_interface': 9.023031548748197e-06, 'learning_rate/action_model': 9.023031548748196e-05, 'epoch': 0.45}) +[2026-06-17 09:03:10,421][starVLA.training.train_starvla][INFO] - Step 1535, Loss: {'train/loss': 0.9763925075531006, 'train/grad_norm_pre_clip': 8.311280250549316, 'timing/data': 0.004598136059939861, 'timing/model': 1.203736622002907, 'learning_rate/qwen_vl_interface': 9.021717330967755e-06, 'learning_rate/action_model': 9.021717330967754e-05, 'epoch': 0.45}) +[2026-06-17 09:03:15,219][starVLA.training.train_starvla][INFO] - Step 1536, Loss: {'train/loss': 1.0723506212234497, 'train/grad_norm_pre_clip': 9.331072807312012, 'timing/data': 0.0050270239589735866, 'timing/model': 1.2129315530182794, 'learning_rate/qwen_vl_interface': 9.020402331306254e-06, 'learning_rate/action_model': 9.020402331306253e-05, 'epoch': 0.45}) +[2026-06-17 09:03:20,079][starVLA.training.train_starvla][INFO] - Step 1537, Loss: {'train/loss': 0.6713594198226929, 'train/grad_norm_pre_clip': 7.100977420806885, 'timing/data': 0.004803612013347447, 'timing/model': 1.2880248259752989, 'learning_rate/qwen_vl_interface': 9.01908655003629e-06, 'learning_rate/action_model': 9.01908655003629e-05, 'epoch': 0.45}) +[2026-06-17 09:03:24,888][starVLA.training.train_starvla][INFO] - Step 1538, Loss: {'train/loss': 0.8860368132591248, 'train/grad_norm_pre_clip': 10.561019897460938, 'timing/data': 0.004699609940871596, 'timing/model': 1.3470822280505672, 'learning_rate/qwen_vl_interface': 9.017769987430629e-06, 'learning_rate/action_model': 9.017769987430628e-05, 'epoch': 0.45}) +[2026-06-17 09:03:29,543][starVLA.training.train_starvla][INFO] - Step 1539, Loss: {'train/loss': 0.6796109080314636, 'train/grad_norm_pre_clip': 11.04546070098877, 'timing/data': 0.004930294002406299, 'timing/model': 1.2103884810348973, 'learning_rate/qwen_vl_interface': 9.016452643762197e-06, 'learning_rate/action_model': 9.016452643762197e-05, 'epoch': 0.45}) +[2026-06-17 09:03:34,340][starVLA.training.train_starvla][INFO] - Step 1540, Loss: {'train/loss': 0.72026526927948, 'train/grad_norm_pre_clip': 9.28565788269043, 'timing/data': 0.00469079299364239, 'timing/model': 1.2096249209716916, 'learning_rate/qwen_vl_interface': 9.015134519304078e-06, 'learning_rate/action_model': 9.015134519304078e-05, 'epoch': 0.45}) +[2026-06-17 09:03:39,174][starVLA.training.train_starvla][INFO] - Step 1541, Loss: {'train/loss': 0.9612730145454407, 'train/grad_norm_pre_clip': 10.99209213256836, 'timing/data': 0.004545517032966018, 'timing/model': 1.2055219459580258, 'learning_rate/qwen_vl_interface': 9.013815614329521e-06, 'learning_rate/action_model': 9.013815614329522e-05, 'epoch': 0.45}) +[2026-06-17 09:03:43,967][starVLA.training.train_starvla][INFO] - Step 1542, Loss: {'train/loss': 0.8816461563110352, 'train/grad_norm_pre_clip': 11.072494506835938, 'timing/data': 0.0046947289956733584, 'timing/model': 1.2066768109798431, 'learning_rate/qwen_vl_interface': 9.01249592911194e-06, 'learning_rate/action_model': 9.012495929111938e-05, 'epoch': 0.45}) +[2026-06-17 09:03:48,636][starVLA.training.train_starvla][INFO] - Step 1543, Loss: {'train/loss': 0.9019504189491272, 'train/grad_norm_pre_clip': 8.488445281982422, 'timing/data': 0.004477933049201965, 'timing/model': 1.2070705429650843, 'learning_rate/qwen_vl_interface': 9.0111754639249e-06, 'learning_rate/action_model': 9.0111754639249e-05, 'epoch': 0.45}) +[2026-06-17 09:03:53,337][starVLA.training.train_starvla][INFO] - Step 1544, Loss: {'train/loss': 0.7278772592544556, 'train/grad_norm_pre_clip': 6.856381893157959, 'timing/data': 0.0046371889766305685, 'timing/model': 1.2526655809488147, 'learning_rate/qwen_vl_interface': 9.009854219042142e-06, 'learning_rate/action_model': 9.009854219042142e-05, 'epoch': 0.45}) +[2026-06-17 09:03:58,088][starVLA.training.train_starvla][INFO] - Step 1545, Loss: {'train/loss': 0.7828255295753479, 'train/grad_norm_pre_clip': 7.23869514465332, 'timing/data': 0.0045767000410705805, 'timing/model': 1.2111642489908263, 'learning_rate/qwen_vl_interface': 9.008532194737556e-06, 'learning_rate/action_model': 9.008532194737557e-05, 'epoch': 0.45}) +[2026-06-17 09:04:02,840][starVLA.training.train_starvla][INFO] - Step 1546, Loss: {'train/loss': 0.8251431584358215, 'train/grad_norm_pre_clip': 7.958887577056885, 'timing/data': 0.004506118013523519, 'timing/model': 1.2058234490687028, 'learning_rate/qwen_vl_interface': 9.007209391285205e-06, 'learning_rate/action_model': 9.007209391285205e-05, 'epoch': 0.45}) +[2026-06-17 09:04:07,657][starVLA.training.train_starvla][INFO] - Step 1547, Loss: {'train/loss': 0.8046391606330872, 'train/grad_norm_pre_clip': 9.12890338897705, 'timing/data': 0.004406814929097891, 'timing/model': 1.204962974996306, 'learning_rate/qwen_vl_interface': 9.005885808959301e-06, 'learning_rate/action_model': 9.005885808959302e-05, 'epoch': 0.46}) +[2026-06-17 09:04:12,454][starVLA.training.train_starvla][INFO] - Step 1548, Loss: {'train/loss': 0.983582615852356, 'train/grad_norm_pre_clip': 9.48457145690918, 'timing/data': 0.005197949940338731, 'timing/model': 1.2104075269307941, 'learning_rate/qwen_vl_interface': 9.00456144803423e-06, 'learning_rate/action_model': 9.00456144803423e-05, 'epoch': 0.46}) +[2026-06-17 09:04:17,249][starVLA.training.train_starvla][INFO] - Step 1549, Loss: {'train/loss': 0.7943453788757324, 'train/grad_norm_pre_clip': 10.726457595825195, 'timing/data': 0.004669000976718962, 'timing/model': 1.3336730729788542, 'learning_rate/qwen_vl_interface': 9.00323630878453e-06, 'learning_rate/action_model': 9.00323630878453e-05, 'epoch': 0.46}) +[2026-06-17 09:04:22,040][starVLA.training.train_starvla][INFO] - Step 1550, Loss: {'train/loss': 0.9633491039276123, 'train/grad_norm_pre_clip': 13.593718528747559, 'timing/data': 0.005066314013674855, 'timing/model': 1.3321889219805598, 'learning_rate/qwen_vl_interface': 9.001910391484903e-06, 'learning_rate/action_model': 9.001910391484904e-05, 'epoch': 0.46}) +[2026-06-17 09:04:26,716][starVLA.training.train_starvla][INFO] - Step 1551, Loss: {'train/loss': 0.6760079264640808, 'train/grad_norm_pre_clip': 13.141918182373047, 'timing/data': 0.004489470040425658, 'timing/model': 1.2103385649388656, 'learning_rate/qwen_vl_interface': 9.000583696410216e-06, 'learning_rate/action_model': 9.000583696410217e-05, 'epoch': 0.46}) +[2026-06-17 09:04:31,479][starVLA.training.train_starvla][INFO] - Step 1552, Loss: {'train/loss': 0.7137130498886108, 'train/grad_norm_pre_clip': 7.337149620056152, 'timing/data': 0.0045396770583465695, 'timing/model': 1.206220610998571, 'learning_rate/qwen_vl_interface': 8.999256223835493e-06, 'learning_rate/action_model': 8.999256223835494e-05, 'epoch': 0.46}) +[2026-06-17 09:04:36,286][starVLA.training.train_starvla][INFO] - Step 1553, Loss: {'train/loss': 0.827976644039154, 'train/grad_norm_pre_clip': 8.538500785827637, 'timing/data': 0.004695543902926147, 'timing/model': 1.2122377949999645, 'learning_rate/qwen_vl_interface': 8.997927974035922e-06, 'learning_rate/action_model': 8.997927974035921e-05, 'epoch': 0.46}) +[2026-06-17 09:04:41,072][starVLA.training.train_starvla][INFO] - Step 1554, Loss: {'train/loss': 0.8440328240394592, 'train/grad_norm_pre_clip': 9.251611709594727, 'timing/data': 0.004641273990273476, 'timing/model': 1.2083073219982907, 'learning_rate/qwen_vl_interface': 8.996598947286846e-06, 'learning_rate/action_model': 8.996598947286846e-05, 'epoch': 0.46}) +[2026-06-17 09:04:45,911][starVLA.training.train_starvla][INFO] - Step 1555, Loss: {'train/loss': 1.0435450077056885, 'train/grad_norm_pre_clip': 14.143767356872559, 'timing/data': 0.0046207859413698316, 'timing/model': 1.374669631011784, 'learning_rate/qwen_vl_interface': 8.99526914386378e-06, 'learning_rate/action_model': 8.995269143863779e-05, 'epoch': 0.46}) +[2026-06-17 09:04:50,686][starVLA.training.train_starvla][INFO] - Step 1556, Loss: {'train/loss': 0.9739053249359131, 'train/grad_norm_pre_clip': 9.567037582397461, 'timing/data': 0.004707115003839135, 'timing/model': 1.3233638850506395, 'learning_rate/qwen_vl_interface': 8.993938564042386e-06, 'learning_rate/action_model': 8.993938564042387e-05, 'epoch': 0.46}) +[2026-06-17 09:04:55,342][starVLA.training.train_starvla][INFO] - Step 1557, Loss: {'train/loss': 1.1674681901931763, 'train/grad_norm_pre_clip': 23.74375343322754, 'timing/data': 0.004667733912356198, 'timing/model': 1.203950286959298, 'learning_rate/qwen_vl_interface': 8.992607208098502e-06, 'learning_rate/action_model': 8.992607208098503e-05, 'epoch': 0.46}) +[2026-06-17 09:05:00,115][starVLA.training.train_starvla][INFO] - Step 1558, Loss: {'train/loss': 0.8474869132041931, 'train/grad_norm_pre_clip': 10.800680160522461, 'timing/data': 0.004718352924101055, 'timing/model': 1.208422036957927, 'learning_rate/qwen_vl_interface': 8.991275076308117e-06, 'learning_rate/action_model': 8.991275076308117e-05, 'epoch': 0.46}) +[2026-06-17 09:05:04,928][starVLA.training.train_starvla][INFO] - Step 1559, Loss: {'train/loss': 1.2353441715240479, 'train/grad_norm_pre_clip': 13.506300926208496, 'timing/data': 0.004535235930234194, 'timing/model': 1.2047182449605316, 'learning_rate/qwen_vl_interface': 8.989942168947382e-06, 'learning_rate/action_model': 8.989942168947381e-05, 'epoch': 0.46}) +[2026-06-17 09:05:09,716][starVLA.training.train_starvla][INFO] - Step 1560, Loss: {'train/loss': 0.760057806968689, 'train/grad_norm_pre_clip': 9.10909652709961, 'timing/data': 0.004535202053375542, 'timing/model': 1.2065819429699332, 'learning_rate/qwen_vl_interface': 8.98860848629261e-06, 'learning_rate/action_model': 8.988608486292611e-05, 'epoch': 0.46}) +[2026-06-17 09:05:14,441][starVLA.training.train_starvla][INFO] - Step 1561, Loss: {'train/loss': 0.9153788685798645, 'train/grad_norm_pre_clip': 9.75716781616211, 'timing/data': 0.004666018998250365, 'timing/model': 1.2768619620474055, 'learning_rate/qwen_vl_interface': 8.987274028620276e-06, 'learning_rate/action_model': 8.987274028620276e-05, 'epoch': 0.46}) +[2026-06-17 09:05:19,106][starVLA.training.train_starvla][INFO] - Step 1562, Loss: {'train/loss': 0.6603779196739197, 'train/grad_norm_pre_clip': 53.62776184082031, 'timing/data': 0.004461124073714018, 'timing/model': 1.2120759269455448, 'learning_rate/qwen_vl_interface': 8.985938796207017e-06, 'learning_rate/action_model': 8.985938796207017e-05, 'epoch': 0.46}) +[2026-06-17 09:05:23,867][starVLA.training.train_starvla][INFO] - Step 1563, Loss: {'train/loss': 0.8471795916557312, 'train/grad_norm_pre_clip': 8.314562797546387, 'timing/data': 0.004648440983146429, 'timing/model': 1.202707637916319, 'learning_rate/qwen_vl_interface': 8.984602789329623e-06, 'learning_rate/action_model': 8.984602789329623e-05, 'epoch': 0.46}) +[2026-06-17 09:05:28,650][starVLA.training.train_starvla][INFO] - Step 1564, Loss: {'train/loss': 0.7302610278129578, 'train/grad_norm_pre_clip': 7.177783012390137, 'timing/data': 0.0046134559670463204, 'timing/model': 1.205567495082505, 'learning_rate/qwen_vl_interface': 8.983266008265052e-06, 'learning_rate/action_model': 8.983266008265052e-05, 'epoch': 0.46}) +[2026-06-17 09:05:33,475][starVLA.training.train_starvla][INFO] - Step 1565, Loss: {'train/loss': 1.0163638591766357, 'train/grad_norm_pre_clip': 18.287933349609375, 'timing/data': 0.004595678998157382, 'timing/model': 1.204265805077739, 'learning_rate/qwen_vl_interface': 8.981928453290421e-06, 'learning_rate/action_model': 8.981928453290421e-05, 'epoch': 0.46}) +[2026-06-17 09:05:38,253][starVLA.training.train_starvla][INFO] - Step 1566, Loss: {'train/loss': 0.8213961720466614, 'train/grad_norm_pre_clip': 10.643227577209473, 'timing/data': 0.0045297740725800395, 'timing/model': 1.3293086329940706, 'learning_rate/qwen_vl_interface': 8.980590124683007e-06, 'learning_rate/action_model': 8.980590124683007e-05, 'epoch': 0.46}) +[2026-06-17 09:05:42,989][starVLA.training.train_starvla][INFO] - Step 1567, Loss: {'train/loss': 0.8220248222351074, 'train/grad_norm_pre_clip': 11.238448143005371, 'timing/data': 0.004807936027646065, 'timing/model': 1.287214769050479, 'learning_rate/qwen_vl_interface': 8.979251022720244e-06, 'learning_rate/action_model': 8.979251022720244e-05, 'epoch': 0.46}) +[2026-06-17 09:05:47,648][starVLA.training.train_starvla][INFO] - Step 1568, Loss: {'train/loss': 0.8958181738853455, 'train/grad_norm_pre_clip': 8.894919395446777, 'timing/data': 0.004831418977119029, 'timing/model': 1.205320538021624, 'learning_rate/qwen_vl_interface': 8.977911147679733e-06, 'learning_rate/action_model': 8.977911147679733e-05, 'epoch': 0.46}) +[2026-06-17 09:05:52,441][starVLA.training.train_starvla][INFO] - Step 1569, Loss: {'train/loss': 1.1123206615447998, 'train/grad_norm_pre_clip': 10.589040756225586, 'timing/data': 0.004402923979796469, 'timing/model': 1.2051992559572682, 'learning_rate/qwen_vl_interface': 8.976570499839229e-06, 'learning_rate/action_model': 8.976570499839228e-05, 'epoch': 0.46}) +[2026-06-17 09:05:57,238][starVLA.training.train_starvla][INFO] - Step 1570, Loss: {'train/loss': 0.7896857857704163, 'train/grad_norm_pre_clip': 9.128193855285645, 'timing/data': 0.004619850078597665, 'timing/model': 1.2113042889395729, 'learning_rate/qwen_vl_interface': 8.97522907947665e-06, 'learning_rate/action_model': 8.975229079476649e-05, 'epoch': 0.46}) +[2026-06-17 09:06:02,027][starVLA.training.train_starvla][INFO] - Step 1571, Loss: {'train/loss': 0.8327140212059021, 'train/grad_norm_pre_clip': 7.428704261779785, 'timing/data': 0.004578487016260624, 'timing/model': 1.203468597959727, 'learning_rate/qwen_vl_interface': 8.973886886870075e-06, 'learning_rate/action_model': 8.973886886870074e-05, 'epoch': 0.46}) +[2026-06-17 09:06:06,810][starVLA.training.train_starvla][INFO] - Step 1572, Loss: {'train/loss': 0.8290948271751404, 'train/grad_norm_pre_clip': 8.87095832824707, 'timing/data': 0.0047492970479652286, 'timing/model': 1.3336039950372651, 'learning_rate/qwen_vl_interface': 8.972543922297739e-06, 'learning_rate/action_model': 8.972543922297739e-05, 'epoch': 0.46}) +[2026-06-17 09:06:11,460][starVLA.training.train_starvla][INFO] - Step 1573, Loss: {'train/loss': 1.1891571283340454, 'train/grad_norm_pre_clip': 9.962841033935547, 'timing/data': 0.004660197999328375, 'timing/model': 1.2037772990297526, 'learning_rate/qwen_vl_interface': 8.971200186038046e-06, 'learning_rate/action_model': 8.971200186038045e-05, 'epoch': 0.46}) +[2026-06-17 09:06:16,248][starVLA.training.train_starvla][INFO] - Step 1574, Loss: {'train/loss': 0.7325760722160339, 'train/grad_norm_pre_clip': 9.458789825439453, 'timing/data': 0.0046715299831703305, 'timing/model': 1.205600816057995, 'learning_rate/qwen_vl_interface': 8.969855678369547e-06, 'learning_rate/action_model': 8.969855678369546e-05, 'epoch': 0.46}) +[2026-06-17 09:06:21,078][starVLA.training.train_starvla][INFO] - Step 1575, Loss: {'train/loss': 0.8333964943885803, 'train/grad_norm_pre_clip': 11.64753532409668, 'timing/data': 0.004611416952684522, 'timing/model': 1.2427751949289814, 'learning_rate/qwen_vl_interface': 8.968510399570964e-06, 'learning_rate/action_model': 8.968510399570965e-05, 'epoch': 0.46}) +[2026-06-17 09:06:25,775][starVLA.training.train_starvla][INFO] - Step 1576, Loss: {'train/loss': 1.0987462997436523, 'train/grad_norm_pre_clip': 11.267889976501465, 'timing/data': 0.004991302965208888, 'timing/model': 1.207554040942341, 'learning_rate/qwen_vl_interface': 8.967164349921173e-06, 'learning_rate/action_model': 8.967164349921174e-05, 'epoch': 0.46}) +[2026-06-17 09:06:30,555][starVLA.training.train_starvla][INFO] - Step 1577, Loss: {'train/loss': 0.6454233527183533, 'train/grad_norm_pre_clip': 5.789244174957275, 'timing/data': 0.004952132934704423, 'timing/model': 1.2047024900093675, 'learning_rate/qwen_vl_interface': 8.965817529699216e-06, 'learning_rate/action_model': 8.965817529699215e-05, 'epoch': 0.46}) +[2026-06-17 09:06:35,317][starVLA.training.train_starvla][INFO] - Step 1578, Loss: {'train/loss': 0.7633110284805298, 'train/grad_norm_pre_clip': 7.206693649291992, 'timing/data': 0.005536110955290496, 'timing/model': 1.3047286929795519, 'learning_rate/qwen_vl_interface': 8.964469939184283e-06, 'learning_rate/action_model': 8.964469939184282e-05, 'epoch': 0.46}) +[2026-06-17 09:06:39,991][starVLA.training.train_starvla][INFO] - Step 1579, Loss: {'train/loss': 0.6255499124526978, 'train/grad_norm_pre_clip': 7.826174736022949, 'timing/data': 0.004950456088408828, 'timing/model': 1.2073157229460776, 'learning_rate/qwen_vl_interface': 8.963121578655736e-06, 'learning_rate/action_model': 8.963121578655736e-05, 'epoch': 0.46}) +[2026-06-17 09:06:44,857][starVLA.training.train_starvla][INFO] - Step 1580, Loss: {'train/loss': 0.9617652893066406, 'train/grad_norm_pre_clip': 14.623274803161621, 'timing/data': 0.004736269940622151, 'timing/model': 1.2064409760059789, 'learning_rate/qwen_vl_interface': 8.96177244839309e-06, 'learning_rate/action_model': 8.961772448393092e-05, 'epoch': 0.46}) +[2026-06-17 09:06:49,654][starVLA.training.train_starvla][INFO] - Step 1581, Loss: {'train/loss': 0.5905600786209106, 'train/grad_norm_pre_clip': 6.562248229980469, 'timing/data': 0.004642655956558883, 'timing/model': 1.2092315329937264, 'learning_rate/qwen_vl_interface': 8.960422548676024e-06, 'learning_rate/action_model': 8.960422548676022e-05, 'epoch': 0.47}) +[2026-06-17 09:06:54,574][starVLA.training.train_starvla][INFO] - Step 1582, Loss: {'train/loss': 0.743973433971405, 'train/grad_norm_pre_clip': 8.305782318115234, 'timing/data': 0.004761034972034395, 'timing/model': 1.2087127979611978, 'learning_rate/qwen_vl_interface': 8.959071879784368e-06, 'learning_rate/action_model': 8.959071879784368e-05, 'epoch': 0.47}) +[2026-06-17 09:06:59,328][starVLA.training.train_starvla][INFO] - Step 1583, Loss: {'train/loss': 0.7517935037612915, 'train/grad_norm_pre_clip': 6.8342061042785645, 'timing/data': 0.004528735997155309, 'timing/model': 1.2104945339960977, 'learning_rate/qwen_vl_interface': 8.95772044199812e-06, 'learning_rate/action_model': 8.95772044199812e-05, 'epoch': 0.47}) +[2026-06-17 09:07:04,159][starVLA.training.train_starvla][INFO] - Step 1584, Loss: {'train/loss': 1.3718403577804565, 'train/grad_norm_pre_clip': 12.442619323730469, 'timing/data': 0.004616767982952297, 'timing/model': 1.3471140899928287, 'learning_rate/qwen_vl_interface': 8.956368235597437e-06, 'learning_rate/action_model': 8.956368235597436e-05, 'epoch': 0.47}) +[2026-06-17 09:07:08,842][starVLA.training.train_starvla][INFO] - Step 1585, Loss: {'train/loss': 0.5824060440063477, 'train/grad_norm_pre_clip': 7.523100852966309, 'timing/data': 0.004960059886798263, 'timing/model': 1.2121712530497462, 'learning_rate/qwen_vl_interface': 8.955015260862627e-06, 'learning_rate/action_model': 8.955015260862627e-05, 'epoch': 0.47}) +[2026-06-17 09:07:13,642][starVLA.training.train_starvla][INFO] - Step 1586, Loss: {'train/loss': 0.6764335632324219, 'train/grad_norm_pre_clip': 27.23607063293457, 'timing/data': 0.004918721970170736, 'timing/model': 1.2096929210238159, 'learning_rate/qwen_vl_interface': 8.95366151807417e-06, 'learning_rate/action_model': 8.95366151807417e-05, 'epoch': 0.47}) +[2026-06-17 09:07:18,451][starVLA.training.train_starvla][INFO] - Step 1587, Loss: {'train/loss': 0.6713176369667053, 'train/grad_norm_pre_clip': 11.646553993225098, 'timing/data': 0.004495461005717516, 'timing/model': 1.2122952400241047, 'learning_rate/qwen_vl_interface': 8.952307007512694e-06, 'learning_rate/action_model': 8.952307007512694e-05, 'epoch': 0.47}) +[2026-06-17 09:07:23,329][starVLA.training.train_starvla][INFO] - Step 1588, Loss: {'train/loss': 0.7285057306289673, 'train/grad_norm_pre_clip': 26.073192596435547, 'timing/data': 0.004823624971322715, 'timing/model': 1.2086264899699017, 'learning_rate/qwen_vl_interface': 8.950951729458991e-06, 'learning_rate/action_model': 8.950951729458991e-05, 'epoch': 0.47}) +[2026-06-17 09:07:28,171][starVLA.training.train_starvla][INFO] - Step 1589, Loss: {'train/loss': 0.8917406797409058, 'train/grad_norm_pre_clip': 7.631448268890381, 'timing/data': 0.0047138649970293045, 'timing/model': 1.208046465064399, 'learning_rate/qwen_vl_interface': 8.949595684194012e-06, 'learning_rate/action_model': 8.949595684194012e-05, 'epoch': 0.47}) +[2026-06-17 09:07:32,989][starVLA.training.train_starvla][INFO] - Step 1590, Loss: {'train/loss': 0.7357701063156128, 'train/grad_norm_pre_clip': 10.016226768493652, 'timing/data': 0.004634047974832356, 'timing/model': 1.342985825962387, 'learning_rate/qwen_vl_interface': 8.948238871998866e-06, 'learning_rate/action_model': 8.948238871998866e-05, 'epoch': 0.47}) +[2026-06-17 09:07:37,648][starVLA.training.train_starvla][INFO] - Step 1591, Loss: {'train/loss': 0.9571859836578369, 'train/grad_norm_pre_clip': 8.77707576751709, 'timing/data': 0.004650567891076207, 'timing/model': 1.2088328569661826, 'learning_rate/qwen_vl_interface': 8.946881293154825e-06, 'learning_rate/action_model': 8.946881293154825e-05, 'epoch': 0.47}) +[2026-06-17 09:07:42,460][starVLA.training.train_starvla][INFO] - Step 1592, Loss: {'train/loss': 0.7311668992042542, 'train/grad_norm_pre_clip': 9.851465225219727, 'timing/data': 0.004608059069141746, 'timing/model': 1.2096181840170175, 'learning_rate/qwen_vl_interface': 8.945522947943312e-06, 'learning_rate/action_model': 8.945522947943313e-05, 'epoch': 0.47}) +[2026-06-17 09:07:47,308][starVLA.training.train_starvla][INFO] - Step 1593, Loss: {'train/loss': 0.6274962425231934, 'train/grad_norm_pre_clip': 8.66731071472168, 'timing/data': 0.005170895019546151, 'timing/model': 1.203501024050638, 'learning_rate/qwen_vl_interface': 8.944163836645916e-06, 'learning_rate/action_model': 8.944163836645915e-05, 'epoch': 0.47}) +[2026-06-17 09:07:52,177][starVLA.training.train_starvla][INFO] - Step 1594, Loss: {'train/loss': 0.7898083925247192, 'train/grad_norm_pre_clip': 9.1724214553833, 'timing/data': 0.004568257019855082, 'timing/model': 1.2048956360667944, 'learning_rate/qwen_vl_interface': 8.942803959544381e-06, 'learning_rate/action_model': 8.942803959544381e-05, 'epoch': 0.47}) +[2026-06-17 09:07:56,978][starVLA.training.train_starvla][INFO] - Step 1595, Loss: {'train/loss': 0.7001277208328247, 'train/grad_norm_pre_clip': 10.137957572937012, 'timing/data': 0.004745939979329705, 'timing/model': 1.3318646480329335, 'learning_rate/qwen_vl_interface': 8.941443316920613e-06, 'learning_rate/action_model': 8.941443316920612e-05, 'epoch': 0.47}) +[2026-06-17 09:08:01,640][starVLA.training.train_starvla][INFO] - Step 1596, Loss: {'train/loss': 0.8818283081054688, 'train/grad_norm_pre_clip': 9.558720588684082, 'timing/data': 0.004629615927115083, 'timing/model': 1.2060091990279034, 'learning_rate/qwen_vl_interface': 8.940081909056671e-06, 'learning_rate/action_model': 8.940081909056671e-05, 'epoch': 0.47}) +[2026-06-17 09:08:06,426][starVLA.training.train_starvla][INFO] - Step 1597, Loss: {'train/loss': 0.8504653573036194, 'train/grad_norm_pre_clip': 10.178237915039062, 'timing/data': 0.004546032985672355, 'timing/model': 1.2072909630369395, 'learning_rate/qwen_vl_interface': 8.93871973623478e-06, 'learning_rate/action_model': 8.93871973623478e-05, 'epoch': 0.47}) +[2026-06-17 09:08:11,132][starVLA.training.train_starvla][INFO] - Step 1598, Loss: {'train/loss': 0.6218152642250061, 'train/grad_norm_pre_clip': 8.038729667663574, 'timing/data': 0.004593855002894998, 'timing/model': 1.2065755759831518, 'learning_rate/qwen_vl_interface': 8.937356798737317e-06, 'learning_rate/action_model': 8.937356798737317e-05, 'epoch': 0.47}) +[2026-06-17 09:08:15,870][starVLA.training.train_starvla][INFO] - Step 1599, Loss: {'train/loss': 0.9355098009109497, 'train/grad_norm_pre_clip': 13.987751960754395, 'timing/data': 0.004820286063477397, 'timing/model': 1.207532376050949, 'learning_rate/qwen_vl_interface': 8.935993096846825e-06, 'learning_rate/action_model': 8.935993096846825e-05, 'epoch': 0.47}) +[2026-06-17 09:09:57,260][starVLA.training.train_starvla][INFO] - Step 1600, Loss: {'train/loss': 0.9424468278884888, 'train/grad_norm_pre_clip': 12.257044792175293, 'eval/loss': 0.8811562061309814, 'eval/latency_0/loss': 0.8811562061309814, 'eval/demon_fire_precision': 0.8676358071920428, 'eval/demon_fire_recall': 0.9418604651162791, 'eval/demon_fire_f1': 0.9032258064516129, 'eval/demon_fire_rate_model': 0.7269187986651835, 'eval/demon_fire_rate_teacher': 0.6696329254727474, 'eval/demon_left_precision': 0.8222222222222222, 'eval/demon_left_recall': 0.7127478753541077, 'eval/demon_left_f1': 0.7635811836115327, 'eval/demon_left_rate_model': 0.42547274749721914, 'eval/demon_left_rate_teacher': 0.49082313681868744, 'eval/demon_right_precision': 0.747923790913532, 'eval/demon_right_recall': 0.8649717514124293, 'eval/demon_right_f1': 0.8022006811632171, 'eval/demon_right_rate_model': 0.5692436040044494, 'eval/demon_right_rate_teacher': 0.4922135706340378, 'eval/demon_move_f1': 0.782890932387375, 'eval/demon_cc_f1': 0.8634123611382611, 'eval/latency_0/demon_fire_precision': 0.8676358071920428, 'eval/latency_0/demon_fire_recall': 0.9418604651162791, 'eval/latency_0/demon_fire_f1': 0.9032258064516129, 'eval/latency_0/demon_fire_rate_model': 0.7269187986651835, 'eval/latency_0/demon_fire_rate_teacher': 0.6696329254727474, 'eval/latency_0/demon_left_precision': 0.8222222222222222, 'eval/latency_0/demon_left_recall': 0.7127478753541077, 'eval/latency_0/demon_left_f1': 0.7635811836115327, 'eval/latency_0/demon_left_rate_model': 0.42547274749721914, 'eval/latency_0/demon_left_rate_teacher': 0.49082313681868744, 'eval/latency_0/demon_right_precision': 0.747923790913532, 'eval/latency_0/demon_right_recall': 0.8649717514124293, 'eval/latency_0/demon_right_f1': 0.8022006811632171, 'eval/latency_0/demon_right_rate_model': 0.5692436040044494, 'eval/latency_0/demon_right_rate_teacher': 0.4922135706340378, 'eval/latency_0/demon_move_f1': 0.782890932387375, 'eval/latency_0/demon_cc_f1': 0.8634123611382611, 'timing/data': 0.004838338936679065, 'timing/model': 1.2059094790602103, 'learning_rate/qwen_vl_interface': 8.934628630845994e-06, 'learning_rate/action_model': 8.934628630845995e-05, 'epoch': 0.47}) +[2026-06-17 09:09:57,260][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_1600_state +[2026-06-17 09:09:57,263][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 09:10:00,833][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_1600_state/model.safetensors +[2026-06-17 09:10:08,517][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_1600_state/optimizer.bin +[2026-06-17 09:10:08,518][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_1600_state/sampler.bin +[2026-06-17 09:10:08,518][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_1600_state/sampler_1.bin +[2026-06-17 09:10:08,519][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_1600_state/random_states_0.pkl +[2026-06-17 09:10:08,562][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1600 +[2026-06-17 09:10:08,562][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 09:10:08,600][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 09:10:13,312][starVLA.training.train_starvla][INFO] - Step 1601, Loss: {'train/loss': 1.00566565990448, 'train/grad_norm_pre_clip': 13.336651802062988, 'timing/data': 0.004968230030499399, 'timing/model': 1.1869992569554597, 'learning_rate/qwen_vl_interface': 8.933263401017685e-06, 'learning_rate/action_model': 8.933263401017684e-05, 'epoch': 0.47}) +[2026-06-17 09:10:18,007][starVLA.training.train_starvla][INFO] - Step 1602, Loss: {'train/loss': 0.8014239072799683, 'train/grad_norm_pre_clip': 8.354164123535156, 'timing/data': 0.005163194960914552, 'timing/model': 1.1908935699611902, 'learning_rate/qwen_vl_interface': 8.93189740764491e-06, 'learning_rate/action_model': 8.931897407644909e-05, 'epoch': 0.47}) +[2026-06-17 09:10:22,752][starVLA.training.train_starvla][INFO] - Step 1603, Loss: {'train/loss': 1.14969801902771, 'train/grad_norm_pre_clip': 13.747236251831055, 'timing/data': 0.005093078012578189, 'timing/model': 1.3210412689950317, 'learning_rate/qwen_vl_interface': 8.93053065101084e-06, 'learning_rate/action_model': 8.93053065101084e-05, 'epoch': 0.47}) +[2026-06-17 09:10:27,365][starVLA.training.train_starvla][INFO] - Step 1604, Loss: {'train/loss': 0.803875744342804, 'train/grad_norm_pre_clip': 10.337379455566406, 'timing/data': 0.00492033502086997, 'timing/model': 1.1957281000213698, 'learning_rate/qwen_vl_interface': 8.929163131398804e-06, 'learning_rate/action_model': 8.929163131398805e-05, 'epoch': 0.47}) +[2026-06-17 09:10:32,230][starVLA.training.train_starvla][INFO] - Step 1605, Loss: {'train/loss': 1.0564545392990112, 'train/grad_norm_pre_clip': 11.722127914428711, 'timing/data': 0.005085956072434783, 'timing/model': 1.2028725689742714, 'learning_rate/qwen_vl_interface': 8.927794849092294e-06, 'learning_rate/action_model': 8.927794849092293e-05, 'epoch': 0.47}) +[2026-06-17 09:10:36,991][starVLA.training.train_starvla][INFO] - Step 1606, Loss: {'train/loss': 1.2524781227111816, 'train/grad_norm_pre_clip': 13.174062728881836, 'timing/data': 0.004997854004614055, 'timing/model': 1.2012321130605415, 'learning_rate/qwen_vl_interface': 8.926425804374953e-06, 'learning_rate/action_model': 8.926425804374954e-05, 'epoch': 0.47}) +[2026-06-17 09:10:41,699][starVLA.training.train_starvla][INFO] - Step 1607, Loss: {'train/loss': 0.6580049991607666, 'train/grad_norm_pre_clip': 7.231608867645264, 'timing/data': 0.004953079973347485, 'timing/model': 1.1979393319925293, 'learning_rate/qwen_vl_interface': 8.925055997530587e-06, 'learning_rate/action_model': 8.925055997530587e-05, 'epoch': 0.47}) +[2026-06-17 09:10:46,467][starVLA.training.train_starvla][INFO] - Step 1608, Loss: {'train/loss': 0.698545515537262, 'train/grad_norm_pre_clip': 8.653729438781738, 'timing/data': 0.00491333391983062, 'timing/model': 1.2049421169795096, 'learning_rate/qwen_vl_interface': 8.92368542884316e-06, 'learning_rate/action_model': 8.923685428843159e-05, 'epoch': 0.47}) +[2026-06-17 09:10:51,226][starVLA.training.train_starvla][INFO] - Step 1609, Loss: {'train/loss': 1.014823079109192, 'train/grad_norm_pre_clip': 8.419159889221191, 'timing/data': 0.004950529895722866, 'timing/model': 1.3257414500694722, 'learning_rate/qwen_vl_interface': 8.922314098596789e-06, 'learning_rate/action_model': 8.922314098596789e-05, 'epoch': 0.47}) +[2026-06-17 09:10:55,854][starVLA.training.train_starvla][INFO] - Step 1610, Loss: {'train/loss': 1.576078176498413, 'train/grad_norm_pre_clip': 15.818427085876465, 'timing/data': 0.005042781005613506, 'timing/model': 1.1996796460589394, 'learning_rate/qwen_vl_interface': 8.920942007075752e-06, 'learning_rate/action_model': 8.920942007075752e-05, 'epoch': 0.47}) +[2026-06-17 09:11:00,619][starVLA.training.train_starvla][INFO] - Step 1611, Loss: {'train/loss': 0.6424216032028198, 'train/grad_norm_pre_clip': 8.340243339538574, 'timing/data': 0.004992133937776089, 'timing/model': 1.2017813959391788, 'learning_rate/qwen_vl_interface': 8.91956915456449e-06, 'learning_rate/action_model': 8.91956915456449e-05, 'epoch': 0.47}) +[2026-06-17 09:11:05,402][starVLA.training.train_starvla][INFO] - Step 1612, Loss: {'train/loss': 0.7167649269104004, 'train/grad_norm_pre_clip': 8.583724021911621, 'timing/data': 0.0048283860087394714, 'timing/model': 1.2069791039684787, 'learning_rate/qwen_vl_interface': 8.918195541347593e-06, 'learning_rate/action_model': 8.918195541347593e-05, 'epoch': 0.47}) +[2026-06-17 09:11:10,145][starVLA.training.train_starvla][INFO] - Step 1613, Loss: {'train/loss': 0.9613211154937744, 'train/grad_norm_pre_clip': 19.868154525756836, 'timing/data': 0.005020605982281268, 'timing/model': 1.2001495440490544, 'learning_rate/qwen_vl_interface': 8.916821167709811e-06, 'learning_rate/action_model': 8.916821167709812e-05, 'epoch': 0.47}) +[2026-06-17 09:11:14,909][starVLA.training.train_starvla][INFO] - Step 1614, Loss: {'train/loss': 1.128637433052063, 'train/grad_norm_pre_clip': 8.582215309143066, 'timing/data': 0.0050441770581528544, 'timing/model': 1.2018139589345083, 'learning_rate/qwen_vl_interface': 8.915446033936057e-06, 'learning_rate/action_model': 8.915446033936056e-05, 'epoch': 0.47}) +[2026-06-17 09:11:19,708][starVLA.training.train_starvla][INFO] - Step 1615, Loss: {'train/loss': 0.818216860294342, 'train/grad_norm_pre_clip': 6.391651153564453, 'timing/data': 0.005009443964809179, 'timing/model': 1.33580019895453, 'learning_rate/qwen_vl_interface': 8.914070140311395e-06, 'learning_rate/action_model': 8.914070140311395e-05, 'epoch': 0.48}) +[2026-06-17 09:11:24,361][starVLA.training.train_starvla][INFO] - Step 1616, Loss: {'train/loss': 0.6978124380111694, 'train/grad_norm_pre_clip': 8.061566352844238, 'timing/data': 0.004906282061710954, 'timing/model': 1.2088478039950132, 'learning_rate/qwen_vl_interface': 8.912693487121053e-06, 'learning_rate/action_model': 8.912693487121052e-05, 'epoch': 0.48}) +[2026-06-17 09:11:29,151][starVLA.training.train_starvla][INFO] - Step 1617, Loss: {'train/loss': 0.8557443022727966, 'train/grad_norm_pre_clip': 7.7990617752075195, 'timing/data': 0.005094593041576445, 'timing/model': 1.208088328014128, 'learning_rate/qwen_vl_interface': 8.911316074650407e-06, 'learning_rate/action_model': 8.911316074650406e-05, 'epoch': 0.48}) +[2026-06-17 09:11:33,927][starVLA.training.train_starvla][INFO] - Step 1618, Loss: {'train/loss': 0.6154638528823853, 'train/grad_norm_pre_clip': 8.695747375488281, 'timing/data': 0.004926668014377356, 'timing/model': 1.2043398750247434, 'learning_rate/qwen_vl_interface': 8.909937903185e-06, 'learning_rate/action_model': 8.909937903185e-05, 'epoch': 0.48}) +[2026-06-17 09:11:38,716][starVLA.training.train_starvla][INFO] - Step 1619, Loss: {'train/loss': 0.8937563300132751, 'train/grad_norm_pre_clip': 6.295786380767822, 'timing/data': 0.00493576203007251, 'timing/model': 1.2077330179745331, 'learning_rate/qwen_vl_interface': 8.908558973010528e-06, 'learning_rate/action_model': 8.908558973010526e-05, 'epoch': 0.48}) +[2026-06-17 09:11:43,389][starVLA.training.train_starvla][INFO] - Step 1620, Loss: {'train/loss': 0.812603235244751, 'train/grad_norm_pre_clip': 11.980286598205566, 'timing/data': 0.004540332010947168, 'timing/model': 1.210344334016554, 'learning_rate/qwen_vl_interface': 8.907179284412841e-06, 'learning_rate/action_model': 8.90717928441284e-05, 'epoch': 0.48}) +[2026-06-17 09:11:48,084][starVLA.training.train_starvla][INFO] - Step 1621, Loss: {'train/loss': 1.0172055959701538, 'train/grad_norm_pre_clip': 9.225645065307617, 'timing/data': 0.005020100041292608, 'timing/model': 1.244192700018175, 'learning_rate/qwen_vl_interface': 8.905798837677956e-06, 'learning_rate/action_model': 8.905798837677956e-05, 'epoch': 0.48}) +[2026-06-17 09:11:52,747][starVLA.training.train_starvla][INFO] - Step 1622, Loss: {'train/loss': 1.0207990407943726, 'train/grad_norm_pre_clip': 12.894034385681152, 'timing/data': 0.004993650014512241, 'timing/model': 1.210117512033321, 'learning_rate/qwen_vl_interface': 8.904417633092037e-06, 'learning_rate/action_model': 8.904417633092037e-05, 'epoch': 0.48}) +[2026-06-17 09:11:57,491][starVLA.training.train_starvla][INFO] - Step 1623, Loss: {'train/loss': 0.8912989497184753, 'train/grad_norm_pre_clip': 11.583828926086426, 'timing/data': 0.005263638915494084, 'timing/model': 1.203370617935434, 'learning_rate/qwen_vl_interface': 8.903035670941411e-06, 'learning_rate/action_model': 8.903035670941412e-05, 'epoch': 0.48}) +[2026-06-17 09:12:02,274][starVLA.training.train_starvla][INFO] - Step 1624, Loss: {'train/loss': 0.7255811095237732, 'train/grad_norm_pre_clip': 14.746132850646973, 'timing/data': 0.0051457349909469485, 'timing/model': 1.205282306065783, 'learning_rate/qwen_vl_interface': 8.901652951512559e-06, 'learning_rate/action_model': 8.90165295151256e-05, 'epoch': 0.48}) +[2026-06-17 09:12:07,065][starVLA.training.train_starvla][INFO] - Step 1625, Loss: {'train/loss': 1.0895695686340332, 'train/grad_norm_pre_clip': 14.077880859375, 'timing/data': 0.00523335998877883, 'timing/model': 1.210753696039319, 'learning_rate/qwen_vl_interface': 8.900269475092122e-06, 'learning_rate/action_model': 8.900269475092122e-05, 'epoch': 0.48}) +[2026-06-17 09:12:11,843][starVLA.training.train_starvla][INFO] - Step 1626, Loss: {'train/loss': 1.108458399772644, 'train/grad_norm_pre_clip': 18.186845779418945, 'timing/data': 0.004671541042625904, 'timing/model': 1.3287013790104538, 'learning_rate/qwen_vl_interface': 8.898885241966894e-06, 'learning_rate/action_model': 8.898885241966893e-05, 'epoch': 0.48}) +[2026-06-17 09:12:16,632][starVLA.training.train_starvla][INFO] - Step 1627, Loss: {'train/loss': 1.2528539896011353, 'train/grad_norm_pre_clip': 10.141037940979004, 'timing/data': 0.005086756078526378, 'timing/model': 1.339821891975589, 'learning_rate/qwen_vl_interface': 8.89750025242383e-06, 'learning_rate/action_model': 8.89750025242383e-05, 'epoch': 0.48}) +[2026-06-17 09:12:21,293][starVLA.training.train_starvla][INFO] - Step 1628, Loss: {'train/loss': 1.0461288690567017, 'train/grad_norm_pre_clip': 8.234657287597656, 'timing/data': 0.0053179089445620775, 'timing/model': 1.2090103180380538, 'learning_rate/qwen_vl_interface': 8.896114506750037e-06, 'learning_rate/action_model': 8.896114506750036e-05, 'epoch': 0.48}) +[2026-06-17 09:12:26,033][starVLA.training.train_starvla][INFO] - Step 1629, Loss: {'train/loss': 1.0262163877487183, 'train/grad_norm_pre_clip': 11.300220489501953, 'timing/data': 0.0050760000012815, 'timing/model': 1.2044796450063586, 'learning_rate/qwen_vl_interface': 8.894728005232785e-06, 'learning_rate/action_model': 8.894728005232785e-05, 'epoch': 0.48}) +[2026-06-17 09:12:30,812][starVLA.training.train_starvla][INFO] - Step 1630, Loss: {'train/loss': 0.658550500869751, 'train/grad_norm_pre_clip': 7.8564324378967285, 'timing/data': 0.004915619036182761, 'timing/model': 1.2028658950002864, 'learning_rate/qwen_vl_interface': 8.893340748159494e-06, 'learning_rate/action_model': 8.893340748159493e-05, 'epoch': 0.48}) +[2026-06-17 09:12:35,587][starVLA.training.train_starvla][INFO] - Step 1631, Loss: {'train/loss': 1.029093623161316, 'train/grad_norm_pre_clip': 10.026058197021484, 'timing/data': 0.00495945499278605, 'timing/model': 1.2034103069454432, 'learning_rate/qwen_vl_interface': 8.891952735817746e-06, 'learning_rate/action_model': 8.891952735817745e-05, 'epoch': 0.48}) +[2026-06-17 09:12:40,381][starVLA.training.train_starvla][INFO] - Step 1632, Loss: {'train/loss': 0.9083933234214783, 'train/grad_norm_pre_clip': 8.82103443145752, 'timing/data': 0.00514887401368469, 'timing/model': 1.3345782089745626, 'learning_rate/qwen_vl_interface': 8.890563968495278e-06, 'learning_rate/action_model': 8.890563968495278e-05, 'epoch': 0.48}) +[2026-06-17 09:12:45,155][starVLA.training.train_starvla][INFO] - Step 1633, Loss: {'train/loss': 1.018827199935913, 'train/grad_norm_pre_clip': 8.775845527648926, 'timing/data': 0.005217650905251503, 'timing/model': 1.3221136080101132, 'learning_rate/qwen_vl_interface': 8.88917444647998e-06, 'learning_rate/action_model': 8.889174446479981e-05, 'epoch': 0.48}) +[2026-06-17 09:12:49,815][starVLA.training.train_starvla][INFO] - Step 1634, Loss: {'train/loss': 1.1074836254119873, 'train/grad_norm_pre_clip': 9.697993278503418, 'timing/data': 0.005237748031504452, 'timing/model': 1.209183917962946, 'learning_rate/qwen_vl_interface': 8.887784170059905e-06, 'learning_rate/action_model': 8.887784170059905e-05, 'epoch': 0.48}) +[2026-06-17 09:12:54,586][starVLA.training.train_starvla][INFO] - Step 1635, Loss: {'train/loss': 1.095559000968933, 'train/grad_norm_pre_clip': 8.055017471313477, 'timing/data': 0.005049768020398915, 'timing/model': 1.206003470113501, 'learning_rate/qwen_vl_interface': 8.886393139523258e-06, 'learning_rate/action_model': 8.886393139523257e-05, 'epoch': 0.48}) +[2026-06-17 09:12:59,370][starVLA.training.train_starvla][INFO] - Step 1636, Loss: {'train/loss': 1.032800555229187, 'train/grad_norm_pre_clip': 10.443809509277344, 'timing/data': 0.004841485992074013, 'timing/model': 1.2029499349882826, 'learning_rate/qwen_vl_interface': 8.885001355158397e-06, 'learning_rate/action_model': 8.885001355158397e-05, 'epoch': 0.48}) +[2026-06-17 09:13:04,189][starVLA.training.train_starvla][INFO] - Step 1637, Loss: {'train/loss': 0.9751445651054382, 'train/grad_norm_pre_clip': 8.022814750671387, 'timing/data': 0.0051288100657984614, 'timing/model': 1.2093024159548804, 'learning_rate/qwen_vl_interface': 8.883608817253847e-06, 'learning_rate/action_model': 8.883608817253847e-05, 'epoch': 0.48}) +[2026-06-17 09:13:08,911][starVLA.training.train_starvla][INFO] - Step 1638, Loss: {'train/loss': 0.959055483341217, 'train/grad_norm_pre_clip': 10.627941131591797, 'timing/data': 0.004945217981003225, 'timing/model': 1.2732831480680034, 'learning_rate/qwen_vl_interface': 8.882215526098276e-06, 'learning_rate/action_model': 8.882215526098277e-05, 'epoch': 0.48}) +[2026-06-17 09:13:13,564][starVLA.training.train_starvla][INFO] - Step 1639, Loss: {'train/loss': 0.8502833843231201, 'train/grad_norm_pre_clip': 23.10735321044922, 'timing/data': 0.0055551460245624185, 'timing/model': 1.204467777046375, 'learning_rate/qwen_vl_interface': 8.880821481980521e-06, 'learning_rate/action_model': 8.880821481980521e-05, 'epoch': 0.48}) +[2026-06-17 09:13:18,331][starVLA.training.train_starvla][INFO] - Step 1640, Loss: {'train/loss': 0.7717908024787903, 'train/grad_norm_pre_clip': 8.166335105895996, 'timing/data': 0.005124422023072839, 'timing/model': 1.2060960850212723, 'learning_rate/qwen_vl_interface': 8.879426685189565e-06, 'learning_rate/action_model': 8.879426685189565e-05, 'epoch': 0.48}) +[2026-06-17 09:13:23,120][starVLA.training.train_starvla][INFO] - Step 1641, Loss: {'train/loss': 0.9215894937515259, 'train/grad_norm_pre_clip': 8.353564262390137, 'timing/data': 0.005077334935776889, 'timing/model': 1.208134707994759, 'learning_rate/qwen_vl_interface': 8.878031136014552e-06, 'learning_rate/action_model': 8.878031136014551e-05, 'epoch': 0.48}) +[2026-06-17 09:13:27,905][starVLA.training.train_starvla][INFO] - Step 1642, Loss: {'train/loss': 0.7824950814247131, 'train/grad_norm_pre_clip': 8.400002479553223, 'timing/data': 0.0059006549417972565, 'timing/model': 1.2038130359724164, 'learning_rate/qwen_vl_interface': 8.87663483474478e-06, 'learning_rate/action_model': 8.87663483474478e-05, 'epoch': 0.48}) +[2026-06-17 09:13:32,730][starVLA.training.train_starvla][INFO] - Step 1643, Loss: {'train/loss': 0.8689073324203491, 'train/grad_norm_pre_clip': 8.420180320739746, 'timing/data': 0.0050809739623218775, 'timing/model': 1.3321331369224936, 'learning_rate/qwen_vl_interface': 8.875237781669704e-06, 'learning_rate/action_model': 8.875237781669703e-05, 'epoch': 0.48}) +[2026-06-17 09:13:37,465][starVLA.training.train_starvla][INFO] - Step 1644, Loss: {'train/loss': 0.9676831960678101, 'train/grad_norm_pre_clip': 8.292236328125, 'timing/data': 0.005188075010664761, 'timing/model': 1.2853372331010178, 'learning_rate/qwen_vl_interface': 8.873839977078933e-06, 'learning_rate/action_model': 8.873839977078934e-05, 'epoch': 0.48}) +[2026-06-17 09:13:42,123][starVLA.training.train_starvla][INFO] - Step 1645, Loss: {'train/loss': 0.9936206340789795, 'train/grad_norm_pre_clip': 9.267621040344238, 'timing/data': 0.005225700093433261, 'timing/model': 1.2056012409739196, 'learning_rate/qwen_vl_interface': 8.87244142126224e-06, 'learning_rate/action_model': 8.87244142126224e-05, 'epoch': 0.48}) +[2026-06-17 09:13:46,989][starVLA.training.train_starvla][INFO] - Step 1646, Loss: {'train/loss': 0.8443089723587036, 'train/grad_norm_pre_clip': 8.888544082641602, 'timing/data': 0.0047964469995349646, 'timing/model': 1.2061555630061775, 'learning_rate/qwen_vl_interface': 8.871042114509539e-06, 'learning_rate/action_model': 8.87104211450954e-05, 'epoch': 0.48}) +[2026-06-17 09:13:51,783][starVLA.training.train_starvla][INFO] - Step 1647, Loss: {'train/loss': 0.826079249382019, 'train/grad_norm_pre_clip': 7.990134239196777, 'timing/data': 0.005218998994678259, 'timing/model': 1.2057056420017034, 'learning_rate/qwen_vl_interface': 8.869642057110912e-06, 'learning_rate/action_model': 8.869642057110911e-05, 'epoch': 0.48}) +[2026-06-17 09:13:56,585][starVLA.training.train_starvla][INFO] - Step 1648, Loss: {'train/loss': 0.67005854845047, 'train/grad_norm_pre_clip': 6.798141956329346, 'timing/data': 0.005042436998337507, 'timing/model': 1.215759575017728, 'learning_rate/qwen_vl_interface': 8.868241249356592e-06, 'learning_rate/action_model': 8.868241249356592e-05, 'epoch': 0.48}) +[2026-06-17 09:14:01,364][starVLA.training.train_starvla][INFO] - Step 1649, Loss: {'train/loss': 0.7501839995384216, 'train/grad_norm_pre_clip': 8.230875015258789, 'timing/data': 0.0049618169432505965, 'timing/model': 1.2045887629501522, 'learning_rate/qwen_vl_interface': 8.866839691536968e-06, 'learning_rate/action_model': 8.866839691536968e-05, 'epoch': 0.49}) +[2026-06-17 09:14:06,126][starVLA.training.train_starvla][INFO] - Step 1650, Loss: {'train/loss': 0.9685842990875244, 'train/grad_norm_pre_clip': 9.028839111328125, 'timing/data': 0.004854165017604828, 'timing/model': 1.2901281940285116, 'learning_rate/qwen_vl_interface': 8.865437383942584e-06, 'learning_rate/action_model': 8.865437383942583e-05, 'epoch': 0.49}) +[2026-06-17 09:14:10,790][starVLA.training.train_starvla][INFO] - Step 1651, Loss: {'train/loss': 0.7721478939056396, 'train/grad_norm_pre_clip': 9.75981330871582, 'timing/data': 0.005057540023699403, 'timing/model': 1.2065417639678344, 'learning_rate/qwen_vl_interface': 8.86403432686414e-06, 'learning_rate/action_model': 8.86403432686414e-05, 'epoch': 0.49}) +[2026-06-17 09:14:15,590][starVLA.training.train_starvla][INFO] - Step 1652, Loss: {'train/loss': 1.22589910030365, 'train/grad_norm_pre_clip': 10.21253776550293, 'timing/data': 0.005042078904807568, 'timing/model': 1.2051311959512532, 'learning_rate/qwen_vl_interface': 8.86263052059249e-06, 'learning_rate/action_model': 8.862630520592489e-05, 'epoch': 0.49}) +[2026-06-17 09:14:20,380][starVLA.training.train_starvla][INFO] - Step 1653, Loss: {'train/loss': 1.043867826461792, 'train/grad_norm_pre_clip': 8.258482933044434, 'timing/data': 0.005230862996540964, 'timing/model': 1.2068768640747294, 'learning_rate/qwen_vl_interface': 8.861225965418647e-06, 'learning_rate/action_model': 8.861225965418646e-05, 'epoch': 0.49}) +[2026-06-17 09:14:25,166][starVLA.training.train_starvla][INFO] - Step 1654, Loss: {'train/loss': 0.8499292731285095, 'train/grad_norm_pre_clip': 11.542518615722656, 'timing/data': 0.005016868002712727, 'timing/model': 1.2046438300749287, 'learning_rate/qwen_vl_interface': 8.859820661633775e-06, 'learning_rate/action_model': 8.859820661633775e-05, 'epoch': 0.49}) +[2026-06-17 09:14:29,953][starVLA.training.train_starvla][INFO] - Step 1655, Loss: {'train/loss': 0.6122094392776489, 'train/grad_norm_pre_clip': 6.141118049621582, 'timing/data': 0.004881707951426506, 'timing/model': 1.337991345091723, 'learning_rate/qwen_vl_interface': 8.858414609529193e-06, 'learning_rate/action_model': 8.858414609529193e-05, 'epoch': 0.49}) +[2026-06-17 09:14:34,607][starVLA.training.train_starvla][INFO] - Step 1656, Loss: {'train/loss': 0.9769768714904785, 'train/grad_norm_pre_clip': 9.882987022399902, 'timing/data': 0.004852197947911918, 'timing/model': 1.2036463260883465, 'learning_rate/qwen_vl_interface': 8.857007809396384e-06, 'learning_rate/action_model': 8.857007809396383e-05, 'epoch': 0.49}) +[2026-06-17 09:14:39,396][starVLA.training.train_starvla][INFO] - Step 1657, Loss: {'train/loss': 0.9193554520606995, 'train/grad_norm_pre_clip': 8.665519714355469, 'timing/data': 0.004553221049718559, 'timing/model': 1.2041875580325723, 'learning_rate/qwen_vl_interface': 8.855600261526973e-06, 'learning_rate/action_model': 8.855600261526973e-05, 'epoch': 0.49}) +[2026-06-17 09:14:44,183][starVLA.training.train_starvla][INFO] - Step 1658, Loss: {'train/loss': 1.0675867795944214, 'train/grad_norm_pre_clip': 9.019216537475586, 'timing/data': 0.005172770004719496, 'timing/model': 1.2075407849624753, 'learning_rate/qwen_vl_interface': 8.854191966212747e-06, 'learning_rate/action_model': 8.854191966212748e-05, 'epoch': 0.49}) +[2026-06-17 09:14:48,886][starVLA.training.train_starvla][INFO] - Step 1659, Loss: {'train/loss': 1.1000710725784302, 'train/grad_norm_pre_clip': 283.40509033203125, 'timing/data': 0.004457233007997274, 'timing/model': 1.2070191089296713, 'learning_rate/qwen_vl_interface': 8.852782923745648e-06, 'learning_rate/action_model': 8.852782923745648e-05, 'epoch': 0.49}) +[2026-06-17 09:14:53,659][starVLA.training.train_starvla][INFO] - Step 1660, Loss: {'train/loss': 0.9355123043060303, 'train/grad_norm_pre_clip': 8.984538078308105, 'timing/data': 0.005257067969068885, 'timing/model': 1.2051330539397895, 'learning_rate/qwen_vl_interface': 8.851373134417773e-06, 'learning_rate/action_model': 8.851373134417773e-05, 'epoch': 0.49}) +[2026-06-17 09:14:58,480][starVLA.training.train_starvla][INFO] - Step 1661, Loss: {'train/loss': 0.838221549987793, 'train/grad_norm_pre_clip': 9.910831451416016, 'timing/data': 0.004891418037004769, 'timing/model': 1.3353642449947074, 'learning_rate/qwen_vl_interface': 8.849962598521373e-06, 'learning_rate/action_model': 8.849962598521373e-05, 'epoch': 0.49}) +[2026-06-17 09:15:03,153][starVLA.training.train_starvla][INFO] - Step 1662, Loss: {'train/loss': 0.6665990948677063, 'train/grad_norm_pre_clip': 8.418071746826172, 'timing/data': 0.00562325504142791, 'timing/model': 1.2173296180553734, 'learning_rate/qwen_vl_interface': 8.848551316348852e-06, 'learning_rate/action_model': 8.848551316348853e-05, 'epoch': 0.49}) +[2026-06-17 09:15:07,972][starVLA.training.train_starvla][INFO] - Step 1663, Loss: {'train/loss': 1.125207781791687, 'train/grad_norm_pre_clip': 11.409071922302246, 'timing/data': 0.00514040095731616, 'timing/model': 1.2108110949629918, 'learning_rate/qwen_vl_interface': 8.847139288192771e-06, 'learning_rate/action_model': 8.84713928819277e-05, 'epoch': 0.49}) +[2026-06-17 09:15:12,767][starVLA.training.train_starvla][INFO] - Step 1664, Loss: {'train/loss': 1.0080037117004395, 'train/grad_norm_pre_clip': 11.387588500976562, 'timing/data': 0.0050604690331965685, 'timing/model': 1.2076887630391866, 'learning_rate/qwen_vl_interface': 8.845726514345845e-06, 'learning_rate/action_model': 8.845726514345843e-05, 'epoch': 0.49}) +[2026-06-17 09:15:17,555][starVLA.training.train_starvla][INFO] - Step 1665, Loss: {'train/loss': 0.8239423632621765, 'train/grad_norm_pre_clip': 8.23258113861084, 'timing/data': 0.00503364996984601, 'timing/model': 1.2057655550306663, 'learning_rate/qwen_vl_interface': 8.844312995100943e-06, 'learning_rate/action_model': 8.844312995100942e-05, 'epoch': 0.49}) +[2026-06-17 09:15:22,295][starVLA.training.train_starvla][INFO] - Step 1666, Loss: {'train/loss': 0.6721755266189575, 'train/grad_norm_pre_clip': 9.304363250732422, 'timing/data': 0.005421918001957238, 'timing/model': 1.2035053740255535, 'learning_rate/qwen_vl_interface': 8.842898730751088e-06, 'learning_rate/action_model': 8.842898730751087e-05, 'epoch': 0.49}) +[2026-06-17 09:15:27,069][starVLA.training.train_starvla][INFO] - Step 1667, Loss: {'train/loss': 0.7210641503334045, 'train/grad_norm_pre_clip': 7.41753625869751, 'timing/data': 0.005220622988417745, 'timing/model': 1.3232318330556154, 'learning_rate/qwen_vl_interface': 8.841483721589462e-06, 'learning_rate/action_model': 8.84148372158946e-05, 'epoch': 0.49}) +[2026-06-17 09:15:31,729][starVLA.training.train_starvla][INFO] - Step 1668, Loss: {'train/loss': 1.057485580444336, 'train/grad_norm_pre_clip': 10.602727890014648, 'timing/data': 0.004738331073895097, 'timing/model': 1.205767225008458, 'learning_rate/qwen_vl_interface': 8.840067967909396e-06, 'learning_rate/action_model': 8.840067967909394e-05, 'epoch': 0.49}) +[2026-06-17 09:15:36,514][starVLA.training.train_starvla][INFO] - Step 1669, Loss: {'train/loss': 0.7149643301963806, 'train/grad_norm_pre_clip': 9.002974510192871, 'timing/data': 0.005131221027113497, 'timing/model': 1.2052946819458157, 'learning_rate/qwen_vl_interface': 8.838651470004376e-06, 'learning_rate/action_model': 8.838651470004375e-05, 'epoch': 0.49}) +[2026-06-17 09:15:41,323][starVLA.training.train_starvla][INFO] - Step 1670, Loss: {'train/loss': 0.6437743902206421, 'train/grad_norm_pre_clip': 8.54789924621582, 'timing/data': 0.004944198997691274, 'timing/model': 1.2028282950632274, 'learning_rate/qwen_vl_interface': 8.837234228168045e-06, 'learning_rate/action_model': 8.837234228168044e-05, 'epoch': 0.49}) +[2026-06-17 09:15:46,129][starVLA.training.train_starvla][INFO] - Step 1671, Loss: {'train/loss': 0.7252024412155151, 'train/grad_norm_pre_clip': 10.12955379486084, 'timing/data': 0.005010791937820613, 'timing/model': 1.2162421100074425, 'learning_rate/qwen_vl_interface': 8.835816242694197e-06, 'learning_rate/action_model': 8.835816242694197e-05, 'epoch': 0.49}) +[2026-06-17 09:15:50,939][starVLA.training.train_starvla][INFO] - Step 1672, Loss: {'train/loss': 0.8390321135520935, 'train/grad_norm_pre_clip': 9.88937759399414, 'timing/data': 0.004998382064513862, 'timing/model': 1.343111829017289, 'learning_rate/qwen_vl_interface': 8.834397513876786e-06, 'learning_rate/action_model': 8.834397513876785e-05, 'epoch': 0.49}) +[2026-06-17 09:15:55,594][starVLA.training.train_starvla][INFO] - Step 1673, Loss: {'train/loss': 0.8139686584472656, 'train/grad_norm_pre_clip': 10.877752304077148, 'timing/data': 0.004833586979657412, 'timing/model': 1.2048398540355265, 'learning_rate/qwen_vl_interface': 8.832978042009913e-06, 'learning_rate/action_model': 8.832978042009912e-05, 'epoch': 0.49}) +[2026-06-17 09:16:00,379][starVLA.training.train_starvla][INFO] - Step 1674, Loss: {'train/loss': 0.9678146243095398, 'train/grad_norm_pre_clip': 10.612125396728516, 'timing/data': 0.005179341998882592, 'timing/model': 1.2053639170480892, 'learning_rate/qwen_vl_interface': 8.831557827387835e-06, 'learning_rate/action_model': 8.831557827387834e-05, 'epoch': 0.49}) +[2026-06-17 09:16:05,089][starVLA.training.train_starvla][INFO] - Step 1675, Loss: {'train/loss': 0.9647538065910339, 'train/grad_norm_pre_clip': 9.233763694763184, 'timing/data': 0.004999059019610286, 'timing/model': 1.2050144139211625, 'learning_rate/qwen_vl_interface': 8.830136870304966e-06, 'learning_rate/action_model': 8.830136870304965e-05, 'epoch': 0.49}) +[2026-06-17 09:16:09,823][starVLA.training.train_starvla][INFO] - Step 1676, Loss: {'train/loss': 0.8061195611953735, 'train/grad_norm_pre_clip': 8.600997924804688, 'timing/data': 0.005218209000304341, 'timing/model': 1.2050062339985743, 'learning_rate/qwen_vl_interface': 8.828715171055871e-06, 'learning_rate/action_model': 8.828715171055871e-05, 'epoch': 0.49}) +[2026-06-17 09:16:14,695][starVLA.training.train_starvla][INFO] - Step 1677, Loss: {'train/loss': 0.8345863223075867, 'train/grad_norm_pre_clip': 9.945358276367188, 'timing/data': 0.004928378970362246, 'timing/model': 1.2081900651101023, 'learning_rate/qwen_vl_interface': 8.827292729935271e-06, 'learning_rate/action_model': 8.827292729935271e-05, 'epoch': 0.49}) +[2026-06-17 09:16:19,487][starVLA.training.train_starvla][INFO] - Step 1678, Loss: {'train/loss': 0.8540011048316956, 'train/grad_norm_pre_clip': 9.932235717773438, 'timing/data': 0.004840777022764087, 'timing/model': 1.330401633051224, 'learning_rate/qwen_vl_interface': 8.825869547238038e-06, 'learning_rate/action_model': 8.825869547238038e-05, 'epoch': 0.49}) +[2026-06-17 09:16:24,144][starVLA.training.train_starvla][INFO] - Step 1679, Loss: {'train/loss': 0.8052914142608643, 'train/grad_norm_pre_clip': 9.97829818725586, 'timing/data': 0.0048057029489427805, 'timing/model': 1.2054818359902129, 'learning_rate/qwen_vl_interface': 8.824445623259202e-06, 'learning_rate/action_model': 8.8244456232592e-05, 'epoch': 0.49}) +[2026-06-17 09:16:28,923][starVLA.training.train_starvla][INFO] - Step 1680, Loss: {'train/loss': 0.7369204759597778, 'train/grad_norm_pre_clip': 8.507181167602539, 'timing/data': 0.00515629502478987, 'timing/model': 1.2043221580097452, 'learning_rate/qwen_vl_interface': 8.82302095829394e-06, 'learning_rate/action_model': 8.82302095829394e-05, 'epoch': 0.49}) +[2026-06-17 09:16:33,712][starVLA.training.train_starvla][INFO] - Step 1681, Loss: {'train/loss': 0.6015194058418274, 'train/grad_norm_pre_clip': 25.812210083007812, 'timing/data': 0.005282709025777876, 'timing/model': 1.2112229840131477, 'learning_rate/qwen_vl_interface': 8.82159555263759e-06, 'learning_rate/action_model': 8.821595552637589e-05, 'epoch': 0.49}) +[2026-06-17 09:16:38,461][starVLA.training.train_starvla][INFO] - Step 1682, Loss: {'train/loss': 1.025256633758545, 'train/grad_norm_pre_clip': 9.831433296203613, 'timing/data': 0.005066605983301997, 'timing/model': 1.206201752065681, 'learning_rate/qwen_vl_interface': 8.820169406585638e-06, 'learning_rate/action_model': 8.820169406585637e-05, 'epoch': 0.49}) +[2026-06-17 09:16:43,246][starVLA.training.train_starvla][INFO] - Step 1683, Loss: {'train/loss': 0.8060047626495361, 'train/grad_norm_pre_clip': 11.347248077392578, 'timing/data': 0.004999010008759797, 'timing/model': 1.2099971259012818, 'learning_rate/qwen_vl_interface': 8.818742520433727e-06, 'learning_rate/action_model': 8.818742520433726e-05, 'epoch': 0.5}) +[2026-06-17 09:16:48,043][starVLA.training.train_starvla][INFO] - Step 1684, Loss: {'train/loss': 0.9150307178497314, 'train/grad_norm_pre_clip': 10.102899551391602, 'timing/data': 0.004829723970033228, 'timing/model': 1.3458260199986398, 'learning_rate/qwen_vl_interface': 8.81731489447765e-06, 'learning_rate/action_model': 8.81731489447765e-05, 'epoch': 0.5}) +[2026-06-17 09:16:52,701][starVLA.training.train_starvla][INFO] - Step 1685, Loss: {'train/loss': 0.8815312385559082, 'train/grad_norm_pre_clip': 9.916468620300293, 'timing/data': 0.005131700891070068, 'timing/model': 1.2041722980793566, 'learning_rate/qwen_vl_interface': 8.815886529013356e-06, 'learning_rate/action_model': 8.815886529013356e-05, 'epoch': 0.5}) +[2026-06-17 09:16:57,505][starVLA.training.train_starvla][INFO] - Step 1686, Loss: {'train/loss': 0.8918505907058716, 'train/grad_norm_pre_clip': 10.21970272064209, 'timing/data': 0.005151767982169986, 'timing/model': 1.2122791289584711, 'learning_rate/qwen_vl_interface': 8.814457424336951e-06, 'learning_rate/action_model': 8.81445742433695e-05, 'epoch': 0.5}) +[2026-06-17 09:17:02,314][starVLA.training.train_starvla][INFO] - Step 1687, Loss: {'train/loss': 0.7864176630973816, 'train/grad_norm_pre_clip': 10.99561882019043, 'timing/data': 0.004973615054041147, 'timing/model': 1.205361844971776, 'learning_rate/qwen_vl_interface': 8.813027580744685e-06, 'learning_rate/action_model': 8.813027580744684e-05, 'epoch': 0.5}) +[2026-06-17 09:17:07,121][starVLA.training.train_starvla][INFO] - Step 1688, Loss: {'train/loss': 0.7503288984298706, 'train/grad_norm_pre_clip': 10.49806022644043, 'timing/data': 0.005079997004941106, 'timing/model': 1.2047954149311408, 'learning_rate/qwen_vl_interface': 8.811596998532965e-06, 'learning_rate/action_model': 8.811596998532965e-05, 'epoch': 0.5}) +[2026-06-17 09:17:11,794][starVLA.training.train_starvla][INFO] - Step 1689, Loss: {'train/loss': 0.8789485692977905, 'train/grad_norm_pre_clip': 12.727404594421387, 'timing/data': 0.0049568620743229985, 'timing/model': 1.2043503109598532, 'learning_rate/qwen_vl_interface': 8.810165677998356e-06, 'learning_rate/action_model': 8.810165677998356e-05, 'epoch': 0.5}) +[2026-06-17 09:17:16,496][starVLA.training.train_starvla][INFO] - Step 1690, Loss: {'train/loss': 0.8735703229904175, 'train/grad_norm_pre_clip': 10.683126449584961, 'timing/data': 0.005063057062216103, 'timing/model': 1.2461592429317534, 'learning_rate/qwen_vl_interface': 8.808733619437572e-06, 'learning_rate/action_model': 8.808733619437572e-05, 'epoch': 0.5}) +[2026-06-17 09:17:21,164][starVLA.training.train_starvla][INFO] - Step 1691, Loss: {'train/loss': 0.824664831161499, 'train/grad_norm_pre_clip': 8.45760726928711, 'timing/data': 0.005377894034609199, 'timing/model': 1.211107345065102, 'learning_rate/qwen_vl_interface': 8.80730082314748e-06, 'learning_rate/action_model': 8.80730082314748e-05, 'epoch': 0.5}) +[2026-06-17 09:17:25,933][starVLA.training.train_starvla][INFO] - Step 1692, Loss: {'train/loss': 0.9499525427818298, 'train/grad_norm_pre_clip': 8.376611709594727, 'timing/data': 0.005092934938147664, 'timing/model': 1.2061474439688027, 'learning_rate/qwen_vl_interface': 8.805867289425098e-06, 'learning_rate/action_model': 8.805867289425098e-05, 'epoch': 0.5}) +[2026-06-17 09:17:30,729][starVLA.training.train_starvla][INFO] - Step 1693, Loss: {'train/loss': 0.8923177123069763, 'train/grad_norm_pre_clip': 10.412710189819336, 'timing/data': 0.005049000959843397, 'timing/model': 1.2068523720372468, 'learning_rate/qwen_vl_interface': 8.804433018567601e-06, 'learning_rate/action_model': 8.8044330185676e-05, 'epoch': 0.5}) +[2026-06-17 09:17:35,557][starVLA.training.train_starvla][INFO] - Step 1694, Loss: {'train/loss': 0.7944825887680054, 'train/grad_norm_pre_clip': 9.865955352783203, 'timing/data': 0.0046561131021007895, 'timing/model': 1.2074729949235916, 'learning_rate/qwen_vl_interface': 8.802998010872316e-06, 'learning_rate/action_model': 8.802998010872316e-05, 'epoch': 0.5}) +[2026-06-17 09:17:40,341][starVLA.training.train_starvla][INFO] - Step 1695, Loss: {'train/loss': 0.671704113483429, 'train/grad_norm_pre_clip': 7.681407928466797, 'timing/data': 0.005367659032344818, 'timing/model': 1.333447384997271, 'learning_rate/qwen_vl_interface': 8.801562266636719e-06, 'learning_rate/action_model': 8.801562266636719e-05, 'epoch': 0.5}) +[2026-06-17 09:17:45,135][starVLA.training.train_starvla][INFO] - Step 1696, Loss: {'train/loss': 0.914524257183075, 'train/grad_norm_pre_clip': 9.214401245117188, 'timing/data': 0.004655443015508354, 'timing/model': 1.3387428929563612, 'learning_rate/qwen_vl_interface': 8.800125786158442e-06, 'learning_rate/action_model': 8.800125786158442e-05, 'epoch': 0.5}) +[2026-06-17 09:17:49,791][starVLA.training.train_starvla][INFO] - Step 1697, Loss: {'train/loss': 0.7749671339988708, 'train/grad_norm_pre_clip': 11.084868431091309, 'timing/data': 0.005251219961792231, 'timing/model': 1.2052955250255764, 'learning_rate/qwen_vl_interface': 8.798688569735268e-06, 'learning_rate/action_model': 8.798688569735268e-05, 'epoch': 0.5}) +[2026-06-17 09:17:54,650][starVLA.training.train_starvla][INFO] - Step 1698, Loss: {'train/loss': 0.9586830139160156, 'train/grad_norm_pre_clip': 12.328779220581055, 'timing/data': 0.005265668034553528, 'timing/model': 1.2046690799761564, 'learning_rate/qwen_vl_interface': 8.797250617665139e-06, 'learning_rate/action_model': 8.797250617665138e-05, 'epoch': 0.5}) +[2026-06-17 09:17:59,437][starVLA.training.train_starvla][INFO] - Step 1699, Loss: {'train/loss': 0.8563353419303894, 'train/grad_norm_pre_clip': 17.851572036743164, 'timing/data': 0.004962355946190655, 'timing/model': 1.2047346020117402, 'learning_rate/qwen_vl_interface': 8.795811930246136e-06, 'learning_rate/action_model': 8.795811930246136e-05, 'epoch': 0.5}) +[2026-06-17 09:19:41,256][starVLA.training.train_starvla][INFO] - Step 1700, Loss: {'train/loss': 0.7906737327575684, 'train/grad_norm_pre_clip': 9.210209846496582, 'eval/loss': 0.8923904299736023, 'eval/latency_0/loss': 0.8923904299736023, 'eval/demon_fire_precision': 0.8630293780999618, 'eval/demon_fire_recall': 0.9393687707641196, 'eval/demon_fire_f1': 0.8995824219526745, 'eval/demon_fire_rate_model': 0.728865406006674, 'eval/demon_fire_rate_teacher': 0.6696329254727474, 'eval/demon_left_precision': 0.7896701929060361, 'eval/demon_left_recall': 0.7189801699716714, 'eval/demon_left_f1': 0.7526690391459074, 'eval/demon_left_rate_model': 0.4468854282536151, 'eval/demon_left_rate_teacher': 0.49082313681868744, 'eval/demon_right_precision': 0.7405731523378583, 'eval/demon_right_recall': 0.8322033898305085, 'eval/demon_right_f1': 0.7837190742218674, 'eval/demon_right_rate_model': 0.5531145717463849, 'eval/demon_right_rate_teacher': 0.4922135706340378, 'eval/demon_move_f1': 0.7681940566838874, 'eval/demon_cc_f1': 0.8556771034308074, 'eval/latency_0/demon_fire_precision': 0.8630293780999618, 'eval/latency_0/demon_fire_recall': 0.9393687707641196, 'eval/latency_0/demon_fire_f1': 0.8995824219526745, 'eval/latency_0/demon_fire_rate_model': 0.728865406006674, 'eval/latency_0/demon_fire_rate_teacher': 0.6696329254727474, 'eval/latency_0/demon_left_precision': 0.7896701929060361, 'eval/latency_0/demon_left_recall': 0.7189801699716714, 'eval/latency_0/demon_left_f1': 0.7526690391459074, 'eval/latency_0/demon_left_rate_model': 0.4468854282536151, 'eval/latency_0/demon_left_rate_teacher': 0.49082313681868744, 'eval/latency_0/demon_right_precision': 0.7405731523378583, 'eval/latency_0/demon_right_recall': 0.8322033898305085, 'eval/latency_0/demon_right_f1': 0.7837190742218674, 'eval/latency_0/demon_right_rate_model': 0.5531145717463849, 'eval/latency_0/demon_right_rate_teacher': 0.4922135706340378, 'eval/latency_0/demon_move_f1': 0.7681940566838874, 'eval/latency_0/demon_cc_f1': 0.8556771034308074, 'timing/data': 0.0045170929515734315, 'timing/model': 1.206096229958348, 'learning_rate/qwen_vl_interface': 8.794372507776508e-06, 'learning_rate/action_model': 8.794372507776508e-05, 'epoch': 0.5}) +[2026-06-17 09:19:41,256][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_1700_state +[2026-06-17 09:19:41,259][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 09:19:44,778][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_1700_state/model.safetensors +[2026-06-17 09:19:52,120][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_1700_state/optimizer.bin +[2026-06-17 09:19:52,120][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_1700_state/sampler.bin +[2026-06-17 09:19:52,121][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_1700_state/sampler_1.bin +[2026-06-17 09:19:52,121][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_1700_state/random_states_0.pkl +[2026-06-17 09:19:52,166][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1700 +[2026-06-17 09:19:52,166][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 09:19:52,247][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 09:19:56,968][starVLA.training.train_starvla][INFO] - Step 1701, Loss: {'train/loss': 0.8610181212425232, 'train/grad_norm_pre_clip': 8.72404670715332, 'timing/data': 0.004490294959396124, 'timing/model': 1.1945545949274674, 'learning_rate/qwen_vl_interface': 8.792932350554642e-06, 'learning_rate/action_model': 8.792932350554643e-05, 'epoch': 0.5}) +[2026-06-17 09:20:01,727][starVLA.training.train_starvla][INFO] - Step 1702, Loss: {'train/loss': 0.8450290560722351, 'train/grad_norm_pre_clip': 8.044072151184082, 'timing/data': 0.004763975040987134, 'timing/model': 1.2002479819348082, 'learning_rate/qwen_vl_interface': 8.791491458879089e-06, 'learning_rate/action_model': 8.791491458879089e-05, 'epoch': 0.5}) +[2026-06-17 09:20:06,480][starVLA.training.train_starvla][INFO] - Step 1703, Loss: {'train/loss': 1.0154157876968384, 'train/grad_norm_pre_clip': 15.286351203918457, 'timing/data': 0.004539627931080759, 'timing/model': 1.3370306049473584, 'learning_rate/qwen_vl_interface': 8.790049833048545e-06, 'learning_rate/action_model': 8.790049833048545e-05, 'epoch': 0.5}) +[2026-06-17 09:20:11,101][starVLA.training.train_starvla][INFO] - Step 1704, Loss: {'train/loss': 0.6926590204238892, 'train/grad_norm_pre_clip': 8.248218536376953, 'timing/data': 0.004606469999998808, 'timing/model': 1.1960339969955385, 'learning_rate/qwen_vl_interface': 8.788607473361862e-06, 'learning_rate/action_model': 8.788607473361862e-05, 'epoch': 0.5}) +[2026-06-17 09:20:15,856][starVLA.training.train_starvla][INFO] - Step 1705, Loss: {'train/loss': 0.7203111052513123, 'train/grad_norm_pre_clip': 7.631176948547363, 'timing/data': 0.004689441062510014, 'timing/model': 1.201090595102869, 'learning_rate/qwen_vl_interface': 8.787164380118044e-06, 'learning_rate/action_model': 8.787164380118043e-05, 'epoch': 0.5}) +[2026-06-17 09:20:20,606][starVLA.training.train_starvla][INFO] - Step 1706, Loss: {'train/loss': 1.0325992107391357, 'train/grad_norm_pre_clip': 9.58721923828125, 'timing/data': 0.004758826922625303, 'timing/model': 1.1967076520668343, 'learning_rate/qwen_vl_interface': 8.785720553616241e-06, 'learning_rate/action_model': 8.78572055361624e-05, 'epoch': 0.5}) +[2026-06-17 09:20:25,361][starVLA.training.train_starvla][INFO] - Step 1707, Loss: {'train/loss': 0.5493519306182861, 'train/grad_norm_pre_clip': 8.250628471374512, 'timing/data': 0.004774386994540691, 'timing/model': 1.2015354810282588, 'learning_rate/qwen_vl_interface': 8.784275994155762e-06, 'learning_rate/action_model': 8.784275994155762e-05, 'epoch': 0.5}) +[2026-06-17 09:20:30,119][starVLA.training.train_starvla][INFO] - Step 1708, Loss: {'train/loss': 0.8541398644447327, 'train/grad_norm_pre_clip': 7.319522380828857, 'timing/data': 0.0050080150831490755, 'timing/model': 1.328355427016504, 'learning_rate/qwen_vl_interface': 8.782830702036067e-06, 'learning_rate/action_model': 8.782830702036067e-05, 'epoch': 0.5}) +[2026-06-17 09:20:34,747][starVLA.training.train_starvla][INFO] - Step 1709, Loss: {'train/loss': 0.9518914818763733, 'train/grad_norm_pre_clip': 9.592928886413574, 'timing/data': 0.004245031042955816, 'timing/model': 1.2006200939649716, 'learning_rate/qwen_vl_interface': 8.781384677556767e-06, 'learning_rate/action_model': 8.781384677556767e-05, 'epoch': 0.5}) +[2026-06-17 09:20:39,503][starVLA.training.train_starvla][INFO] - Step 1710, Loss: {'train/loss': 0.7038213014602661, 'train/grad_norm_pre_clip': 9.38990592956543, 'timing/data': 0.004622846026904881, 'timing/model': 1.198607653961517, 'learning_rate/qwen_vl_interface': 8.779937921017624e-06, 'learning_rate/action_model': 8.779937921017623e-05, 'epoch': 0.5}) +[2026-06-17 09:20:44,149][starVLA.training.train_starvla][INFO] - Step 1711, Loss: {'train/loss': 0.7191975712776184, 'train/grad_norm_pre_clip': 8.11847972869873, 'timing/data': 0.004706224892288446, 'timing/model': 1.1995921359630302, 'learning_rate/qwen_vl_interface': 8.77849043271855e-06, 'learning_rate/action_model': 8.778490432718551e-05, 'epoch': 0.5}) +[2026-06-17 09:20:48,804][starVLA.training.train_starvla][INFO] - Step 1712, Loss: {'train/loss': 0.9313579797744751, 'train/grad_norm_pre_clip': 8.591416358947754, 'timing/data': 0.004524442949332297, 'timing/model': 1.202121181995608, 'learning_rate/qwen_vl_interface': 8.777042212959615e-06, 'learning_rate/action_model': 8.777042212959614e-05, 'epoch': 0.5}) +[2026-06-17 09:20:53,492][starVLA.training.train_starvla][INFO] - Step 1713, Loss: {'train/loss': 0.6686241626739502, 'train/grad_norm_pre_clip': 10.018715858459473, 'timing/data': 0.004593376070261002, 'timing/model': 1.201051232055761, 'learning_rate/qwen_vl_interface': 8.775593262041032e-06, 'learning_rate/action_model': 8.77559326204103e-05, 'epoch': 0.5}) +[2026-06-17 09:20:58,221][starVLA.training.train_starvla][INFO] - Step 1714, Loss: {'train/loss': 0.6734127998352051, 'train/grad_norm_pre_clip': 7.369417190551758, 'timing/data': 0.004680734942667186, 'timing/model': 1.2854308460373431, 'learning_rate/qwen_vl_interface': 8.774143580263172e-06, 'learning_rate/action_model': 8.77414358026317e-05, 'epoch': 0.5}) +[2026-06-17 09:21:02,871][starVLA.training.train_starvla][INFO] - Step 1715, Loss: {'train/loss': 0.8874594569206238, 'train/grad_norm_pre_clip': 8.881651878356934, 'timing/data': 0.0047347089275717735, 'timing/model': 1.2027764610247687, 'learning_rate/qwen_vl_interface': 8.772693167926554e-06, 'learning_rate/action_model': 8.772693167926555e-05, 'epoch': 0.5}) +[2026-06-17 09:21:07,645][starVLA.training.train_starvla][INFO] - Step 1716, Loss: {'train/loss': 0.9254369139671326, 'train/grad_norm_pre_clip': 9.667593002319336, 'timing/data': 0.004481792915612459, 'timing/model': 1.2027980600250885, 'learning_rate/qwen_vl_interface': 8.771242025331855e-06, 'learning_rate/action_model': 8.771242025331854e-05, 'epoch': 0.5}) +[2026-06-17 09:21:12,421][starVLA.training.train_starvla][INFO] - Step 1717, Loss: {'train/loss': 0.8618890643119812, 'train/grad_norm_pre_clip': 10.522037506103516, 'timing/data': 0.004523562034592032, 'timing/model': 1.2010694060008973, 'learning_rate/qwen_vl_interface': 8.769790152779893e-06, 'learning_rate/action_model': 8.769790152779892e-05, 'epoch': 0.51}) +[2026-06-17 09:21:17,198][starVLA.training.train_starvla][INFO] - Step 1718, Loss: {'train/loss': 0.5587041974067688, 'train/grad_norm_pre_clip': 7.43039083480835, 'timing/data': 0.00453230005223304, 'timing/model': 1.2061248449608684, 'learning_rate/qwen_vl_interface': 8.768337550571644e-06, 'learning_rate/action_model': 8.768337550571645e-05, 'epoch': 0.51}) +[2026-06-17 09:21:21,974][starVLA.training.train_starvla][INFO] - Step 1719, Loss: {'train/loss': 0.8657875657081604, 'train/grad_norm_pre_clip': 10.890761375427246, 'timing/data': 0.00473445700481534, 'timing/model': 1.3338835509493947, 'learning_rate/qwen_vl_interface': 8.766884219008238e-06, 'learning_rate/action_model': 8.766884219008238e-05, 'epoch': 0.51}) +[2026-06-17 09:21:26,633][starVLA.training.train_starvla][INFO] - Step 1720, Loss: {'train/loss': 0.6945140361785889, 'train/grad_norm_pre_clip': 8.924196243286133, 'timing/data': 0.00438737403601408, 'timing/model': 1.2025717090582475, 'learning_rate/qwen_vl_interface': 8.765430158390948e-06, 'learning_rate/action_model': 8.765430158390947e-05, 'epoch': 0.51}) +[2026-06-17 09:21:31,417][starVLA.training.train_starvla][INFO] - Step 1721, Loss: {'train/loss': 0.5561974048614502, 'train/grad_norm_pre_clip': 11.037795066833496, 'timing/data': 0.004793255007825792, 'timing/model': 1.2060286329360679, 'learning_rate/qwen_vl_interface': 8.763975369021202e-06, 'learning_rate/action_model': 8.763975369021203e-05, 'epoch': 0.51}) +[2026-06-17 09:21:36,203][starVLA.training.train_starvla][INFO] - Step 1722, Loss: {'train/loss': 0.6697278022766113, 'train/grad_norm_pre_clip': 8.590201377868652, 'timing/data': 0.004587537026964128, 'timing/model': 1.2065444730687886, 'learning_rate/qwen_vl_interface': 8.762519851200581e-06, 'learning_rate/action_model': 8.762519851200581e-05, 'epoch': 0.51}) +[2026-06-17 09:21:40,890][starVLA.training.train_starvla][INFO] - Step 1723, Loss: {'train/loss': 0.96013343334198, 'train/grad_norm_pre_clip': 9.577099800109863, 'timing/data': 0.004312259028665721, 'timing/model': 1.2019602480577305, 'learning_rate/qwen_vl_interface': 8.761063605230818e-06, 'learning_rate/action_model': 8.761063605230818e-05, 'epoch': 0.51}) +[2026-06-17 09:21:45,655][starVLA.training.train_starvla][INFO] - Step 1724, Loss: {'train/loss': 0.8686541318893433, 'train/grad_norm_pre_clip': 10.358596801757812, 'timing/data': 0.004659979022108018, 'timing/model': 1.2011242000153288, 'learning_rate/qwen_vl_interface': 8.759606631413788e-06, 'learning_rate/action_model': 8.759606631413788e-05, 'epoch': 0.51}) +[2026-06-17 09:21:50,399][starVLA.training.train_starvla][INFO] - Step 1725, Loss: {'train/loss': 0.6879015564918518, 'train/grad_norm_pre_clip': 7.818778038024902, 'timing/data': 0.004383281106129289, 'timing/model': 1.299330156063661, 'learning_rate/qwen_vl_interface': 8.75814893005153e-06, 'learning_rate/action_model': 8.75814893005153e-05, 'epoch': 0.51}) +[2026-06-17 09:21:55,064][starVLA.training.train_starvla][INFO] - Step 1726, Loss: {'train/loss': 0.818446934223175, 'train/grad_norm_pre_clip': 9.91538143157959, 'timing/data': 0.00462444603908807, 'timing/model': 1.2031317630317062, 'learning_rate/qwen_vl_interface': 8.756690501446222e-06, 'learning_rate/action_model': 8.756690501446221e-05, 'epoch': 0.51}) +[2026-06-17 09:21:59,854][starVLA.training.train_starvla][INFO] - Step 1727, Loss: {'train/loss': 0.5910475850105286, 'train/grad_norm_pre_clip': 6.867257595062256, 'timing/data': 0.004669965011999011, 'timing/model': 1.2093888079980388, 'learning_rate/qwen_vl_interface': 8.755231345900199e-06, 'learning_rate/action_model': 8.755231345900198e-05, 'epoch': 0.51}) +[2026-06-17 09:22:04,636][starVLA.training.train_starvla][INFO] - Step 1728, Loss: {'train/loss': 1.2732548713684082, 'train/grad_norm_pre_clip': 15.251587867736816, 'timing/data': 0.004480586969293654, 'timing/model': 1.202848048065789, 'learning_rate/qwen_vl_interface': 8.753771463715947e-06, 'learning_rate/action_model': 8.753771463715947e-05, 'epoch': 0.51}) +[2026-06-17 09:22:09,423][starVLA.training.train_starvla][INFO] - Step 1729, Loss: {'train/loss': 0.9244835376739502, 'train/grad_norm_pre_clip': 14.497637748718262, 'timing/data': 0.0044261489529162645, 'timing/model': 1.2041271139169112, 'learning_rate/qwen_vl_interface': 8.7523108551961e-06, 'learning_rate/action_model': 8.7523108551961e-05, 'epoch': 0.51}) +[2026-06-17 09:22:14,152][starVLA.training.train_starvla][INFO] - Step 1730, Loss: {'train/loss': 0.9105460047721863, 'train/grad_norm_pre_clip': 9.644450187683105, 'timing/data': 0.004583830013871193, 'timing/model': 1.2053086230298504, 'learning_rate/qwen_vl_interface': 8.750849520643445e-06, 'learning_rate/action_model': 8.750849520643444e-05, 'epoch': 0.51}) +[2026-06-17 09:22:18,934][starVLA.training.train_starvla][INFO] - Step 1731, Loss: {'train/loss': 0.7008923888206482, 'train/grad_norm_pre_clip': 10.6429443359375, 'timing/data': 0.004561495967209339, 'timing/model': 1.3358502669725567, 'learning_rate/qwen_vl_interface': 8.749387460360917e-06, 'learning_rate/action_model': 8.749387460360916e-05, 'epoch': 0.51}) +[2026-06-17 09:22:23,602][starVLA.training.train_starvla][INFO] - Step 1732, Loss: {'train/loss': 1.0829863548278809, 'train/grad_norm_pre_clip': 11.406307220458984, 'timing/data': 0.004530632053501904, 'timing/model': 1.2048700849991292, 'learning_rate/qwen_vl_interface': 8.747924674651602e-06, 'learning_rate/action_model': 8.747924674651602e-05, 'epoch': 0.51}) +[2026-06-17 09:22:28,385][starVLA.training.train_starvla][INFO] - Step 1733, Loss: {'train/loss': 0.7044511437416077, 'train/grad_norm_pre_clip': 13.772114753723145, 'timing/data': 0.004815311054699123, 'timing/model': 1.2082259330200031, 'learning_rate/qwen_vl_interface': 8.74646116381874e-06, 'learning_rate/action_model': 8.74646116381874e-05, 'epoch': 0.51}) +[2026-06-17 09:22:33,166][starVLA.training.train_starvla][INFO] - Step 1734, Loss: {'train/loss': 1.208717703819275, 'train/grad_norm_pre_clip': 15.576082229614258, 'timing/data': 0.004482069052755833, 'timing/model': 1.2031736939679831, 'learning_rate/qwen_vl_interface': 8.744996928165714e-06, 'learning_rate/action_model': 8.744996928165714e-05, 'epoch': 0.51}) +[2026-06-17 09:22:37,939][starVLA.training.train_starvla][INFO] - Step 1735, Loss: {'train/loss': 0.6365243196487427, 'train/grad_norm_pre_clip': 11.163734436035156, 'timing/data': 0.004544924013316631, 'timing/model': 1.2051768180681393, 'learning_rate/qwen_vl_interface': 8.743531967996066e-06, 'learning_rate/action_model': 8.743531967996065e-05, 'epoch': 0.51}) +[2026-06-17 09:22:42,684][starVLA.training.train_starvla][INFO] - Step 1736, Loss: {'train/loss': 0.7940150499343872, 'train/grad_norm_pre_clip': 8.74301528930664, 'timing/data': 0.004384095082059503, 'timing/model': 1.202217147918418, 'learning_rate/qwen_vl_interface': 8.742066283613481e-06, 'learning_rate/action_model': 8.74206628361348e-05, 'epoch': 0.51}) +[2026-06-17 09:22:47,454][starVLA.training.train_starvla][INFO] - Step 1737, Loss: {'train/loss': 0.8856518864631653, 'train/grad_norm_pre_clip': 9.069145202636719, 'timing/data': 0.004612735938280821, 'timing/model': 1.319031324936077, 'learning_rate/qwen_vl_interface': 8.740599875321797e-06, 'learning_rate/action_model': 8.740599875321797e-05, 'epoch': 0.51}) +[2026-06-17 09:22:52,153][starVLA.training.train_starvla][INFO] - Step 1738, Loss: {'train/loss': 0.6965926885604858, 'train/grad_norm_pre_clip': 8.914961814880371, 'timing/data': 0.0044297060230746865, 'timing/model': 1.2036636089906096, 'learning_rate/qwen_vl_interface': 8.739132743425004e-06, 'learning_rate/action_model': 8.739132743425002e-05, 'epoch': 0.51}) +[2026-06-17 09:22:56,943][starVLA.training.train_starvla][INFO] - Step 1739, Loss: {'train/loss': 1.0897822380065918, 'train/grad_norm_pre_clip': 11.504997253417969, 'timing/data': 0.0047721590381115675, 'timing/model': 1.205481062992476, 'learning_rate/qwen_vl_interface': 8.737664888227239e-06, 'learning_rate/action_model': 8.737664888227239e-05, 'epoch': 0.51}) +[2026-06-17 09:23:01,720][starVLA.training.train_starvla][INFO] - Step 1740, Loss: {'train/loss': 0.8146436214447021, 'train/grad_norm_pre_clip': 11.707403182983398, 'timing/data': 0.004393331939354539, 'timing/model': 1.2033264780184254, 'learning_rate/qwen_vl_interface': 8.736196310032787e-06, 'learning_rate/action_model': 8.736196310032787e-05, 'epoch': 0.51}) +[2026-06-17 09:23:06,498][starVLA.training.train_starvla][INFO] - Step 1741, Loss: {'train/loss': 0.8799643516540527, 'train/grad_norm_pre_clip': 9.08974552154541, 'timing/data': 0.004550341982394457, 'timing/model': 1.2046542110620067, 'learning_rate/qwen_vl_interface': 8.734727009146092e-06, 'learning_rate/action_model': 8.734727009146092e-05, 'epoch': 0.51}) +[2026-06-17 09:23:11,279][starVLA.training.train_starvla][INFO] - Step 1742, Loss: {'train/loss': 0.745881974697113, 'train/grad_norm_pre_clip': 9.846057891845703, 'timing/data': 0.004628013004548848, 'timing/model': 1.3324645350221545, 'learning_rate/qwen_vl_interface': 8.733256985871737e-06, 'learning_rate/action_model': 8.733256985871737e-05, 'epoch': 0.51}) +[2026-06-17 09:23:15,943][starVLA.training.train_starvla][INFO] - Step 1743, Loss: {'train/loss': 0.9857363700866699, 'train/grad_norm_pre_clip': 11.119658470153809, 'timing/data': 0.004558278014883399, 'timing/model': 1.2052192880073562, 'learning_rate/qwen_vl_interface': 8.73178624051446e-06, 'learning_rate/action_model': 8.73178624051446e-05, 'epoch': 0.51}) +[2026-06-17 09:23:20,733][starVLA.training.train_starvla][INFO] - Step 1744, Loss: {'train/loss': 0.7670040130615234, 'train/grad_norm_pre_clip': 11.913829803466797, 'timing/data': 0.004454036010429263, 'timing/model': 1.2077504330081865, 'learning_rate/qwen_vl_interface': 8.73031477337915e-06, 'learning_rate/action_model': 8.73031477337915e-05, 'epoch': 0.51}) +[2026-06-17 09:23:25,442][starVLA.training.train_starvla][INFO] - Step 1745, Loss: {'train/loss': 0.6556771397590637, 'train/grad_norm_pre_clip': 21.885730743408203, 'timing/data': 0.0045860339887440205, 'timing/model': 1.205499469069764, 'learning_rate/qwen_vl_interface': 8.728842584770842e-06, 'learning_rate/action_model': 8.728842584770842e-05, 'epoch': 0.51}) +[2026-06-17 09:23:30,190][starVLA.training.train_starvla][INFO] - Step 1746, Loss: {'train/loss': 0.7914004921913147, 'train/grad_norm_pre_clip': 8.66703987121582, 'timing/data': 0.0043130089761689305, 'timing/model': 1.208554562064819, 'learning_rate/qwen_vl_interface': 8.727369674994722e-06, 'learning_rate/action_model': 8.727369674994722e-05, 'epoch': 0.51}) +[2026-06-17 09:23:34,979][starVLA.training.train_starvla][INFO] - Step 1747, Loss: {'train/loss': 0.9592269659042358, 'train/grad_norm_pre_clip': 9.908482551574707, 'timing/data': 0.004755259957164526, 'timing/model': 1.2089779119705781, 'learning_rate/qwen_vl_interface': 8.725896044356128e-06, 'learning_rate/action_model': 8.725896044356128e-05, 'epoch': 0.51}) +[2026-06-17 09:23:39,773][starVLA.training.train_starvla][INFO] - Step 1748, Loss: {'train/loss': 1.2466288805007935, 'train/grad_norm_pre_clip': 11.523303985595703, 'timing/data': 0.0046012530801817775, 'timing/model': 1.3348849180620164, 'learning_rate/qwen_vl_interface': 8.724421693160544e-06, 'learning_rate/action_model': 8.724421693160543e-05, 'epoch': 0.51}) +[2026-06-17 09:23:44,439][starVLA.training.train_starvla][INFO] - Step 1749, Loss: {'train/loss': 0.9005088806152344, 'train/grad_norm_pre_clip': 8.673504829406738, 'timing/data': 0.004281681962311268, 'timing/model': 1.204969662008807, 'learning_rate/qwen_vl_interface': 8.722946621713603e-06, 'learning_rate/action_model': 8.722946621713602e-05, 'epoch': 0.51}) +[2026-06-17 09:23:49,241][starVLA.training.train_starvla][INFO] - Step 1750, Loss: {'train/loss': 0.7006203532218933, 'train/grad_norm_pre_clip': 6.101257801055908, 'timing/data': 0.004145369050092995, 'timing/model': 1.2118858270114288, 'learning_rate/qwen_vl_interface': 8.721470830321091e-06, 'learning_rate/action_model': 8.72147083032109e-05, 'epoch': 0.51}) +[2026-06-17 09:23:54,036][starVLA.training.train_starvla][INFO] - Step 1751, Loss: {'train/loss': 0.6863971948623657, 'train/grad_norm_pre_clip': 5.227353096008301, 'timing/data': 0.004456716938875616, 'timing/model': 1.2066555449273437, 'learning_rate/qwen_vl_interface': 8.719994319288939e-06, 'learning_rate/action_model': 8.71999431928894e-05, 'epoch': 0.52}) +[2026-06-17 09:23:58,823][starVLA.training.train_starvla][INFO] - Step 1752, Loss: {'train/loss': 0.8985122442245483, 'train/grad_norm_pre_clip': 8.70480728149414, 'timing/data': 0.004638097947463393, 'timing/model': 1.2052327650599182, 'learning_rate/qwen_vl_interface': 8.718517088923233e-06, 'learning_rate/action_model': 8.718517088923233e-05, 'epoch': 0.52}) +[2026-06-17 09:24:03,609][starVLA.training.train_starvla][INFO] - Step 1753, Loss: {'train/loss': 0.8401866555213928, 'train/grad_norm_pre_clip': 7.0363240242004395, 'timing/data': 0.004584506037645042, 'timing/model': 1.209963927976787, 'learning_rate/qwen_vl_interface': 8.7170391395302e-06, 'learning_rate/action_model': 8.7170391395302e-05, 'epoch': 0.52}) +[2026-06-17 09:24:08,409][starVLA.training.train_starvla][INFO] - Step 1754, Loss: {'train/loss': 0.7456109523773193, 'train/grad_norm_pre_clip': 7.443566799163818, 'timing/data': 0.0050686049507930875, 'timing/model': 1.3418137260014191, 'learning_rate/qwen_vl_interface': 8.715560471416223e-06, 'learning_rate/action_model': 8.715560471416224e-05, 'epoch': 0.52}) +[2026-06-17 09:24:13,081][starVLA.training.train_starvla][INFO] - Step 1755, Loss: {'train/loss': 0.8306211829185486, 'train/grad_norm_pre_clip': 7.9608235359191895, 'timing/data': 0.005070518935099244, 'timing/model': 1.2079272120026872, 'learning_rate/qwen_vl_interface': 8.714081084887833e-06, 'learning_rate/action_model': 8.714081084887833e-05, 'epoch': 0.52}) +[2026-06-17 09:24:17,892][starVLA.training.train_starvla][INFO] - Step 1756, Loss: {'train/loss': 0.7201592326164246, 'train/grad_norm_pre_clip': 8.046710968017578, 'timing/data': 0.004188834107480943, 'timing/model': 1.207233076915145, 'learning_rate/qwen_vl_interface': 8.712600980251706e-06, 'learning_rate/action_model': 8.712600980251705e-05, 'epoch': 0.52}) +[2026-06-17 09:24:22,686][starVLA.training.train_starvla][INFO] - Step 1757, Loss: {'train/loss': 0.8508940935134888, 'train/grad_norm_pre_clip': 8.816263198852539, 'timing/data': 0.004538952023722231, 'timing/model': 1.2061899550026283, 'learning_rate/qwen_vl_interface': 8.71112015781467e-06, 'learning_rate/action_model': 8.71112015781467e-05, 'epoch': 0.52}) +[2026-06-17 09:24:27,485][starVLA.training.train_starvla][INFO] - Step 1758, Loss: {'train/loss': 0.972554087638855, 'train/grad_norm_pre_clip': 8.792214393615723, 'timing/data': 0.004640903091058135, 'timing/model': 1.2115367389051244, 'learning_rate/qwen_vl_interface': 8.709638617883699e-06, 'learning_rate/action_model': 8.709638617883699e-05, 'epoch': 0.52}) +[2026-06-17 09:24:32,166][starVLA.training.train_starvla][INFO] - Step 1759, Loss: {'train/loss': 0.8870832920074463, 'train/grad_norm_pre_clip': 9.136686325073242, 'timing/data': 0.004940303973853588, 'timing/model': 1.2101412559859455, 'learning_rate/qwen_vl_interface': 8.708156360765922e-06, 'learning_rate/action_model': 8.708156360765922e-05, 'epoch': 0.52}) +[2026-06-17 09:24:36,870][starVLA.training.train_starvla][INFO] - Step 1760, Loss: {'train/loss': 0.6984926462173462, 'train/grad_norm_pre_clip': 6.945311546325684, 'timing/data': 0.004666005959734321, 'timing/model': 1.2467287379549816, 'learning_rate/qwen_vl_interface': 8.706673386768609e-06, 'learning_rate/action_model': 8.706673386768609e-05, 'epoch': 0.52}) +[2026-06-17 09:24:41,533][starVLA.training.train_starvla][INFO] - Step 1761, Loss: {'train/loss': 0.9846652746200562, 'train/grad_norm_pre_clip': 11.109528541564941, 'timing/data': 0.00470194302033633, 'timing/model': 1.2073963599978015, 'learning_rate/qwen_vl_interface': 8.705189696199184e-06, 'learning_rate/action_model': 8.705189696199183e-05, 'epoch': 0.52}) +[2026-06-17 09:24:46,278][starVLA.training.train_starvla][INFO] - Step 1762, Loss: {'train/loss': 1.0018386840820312, 'train/grad_norm_pre_clip': 12.491466522216797, 'timing/data': 0.004761217045597732, 'timing/model': 1.2071021660231054, 'learning_rate/qwen_vl_interface': 8.703705289365216e-06, 'learning_rate/action_model': 8.703705289365216e-05, 'epoch': 0.52}) +[2026-06-17 09:24:51,069][starVLA.training.train_starvla][INFO] - Step 1763, Loss: {'train/loss': 0.8615511059761047, 'train/grad_norm_pre_clip': 16.231538772583008, 'timing/data': 0.004543257993645966, 'timing/model': 1.2069926139665768, 'learning_rate/qwen_vl_interface': 8.702220166574426e-06, 'learning_rate/action_model': 8.702220166574425e-05, 'epoch': 0.52}) +[2026-06-17 09:24:55,868][starVLA.training.train_starvla][INFO] - Step 1764, Loss: {'train/loss': 0.5470036864280701, 'train/grad_norm_pre_clip': 6.130809307098389, 'timing/data': 0.00484280101954937, 'timing/model': 1.2066872519208118, 'learning_rate/qwen_vl_interface': 8.700734328134681e-06, 'learning_rate/action_model': 8.70073432813468e-05, 'epoch': 0.52}) +[2026-06-17 09:25:00,689][starVLA.training.train_starvla][INFO] - Step 1765, Loss: {'train/loss': 0.9495387673377991, 'train/grad_norm_pre_clip': 9.678366661071777, 'timing/data': 0.0047923719976097345, 'timing/model': 1.3363027969608083, 'learning_rate/qwen_vl_interface': 8.699247774353997e-06, 'learning_rate/action_model': 8.699247774353996e-05, 'epoch': 0.52}) +[2026-06-17 09:25:05,513][starVLA.training.train_starvla][INFO] - Step 1766, Loss: {'train/loss': 0.7707387208938599, 'train/grad_norm_pre_clip': 9.042713165283203, 'timing/data': 0.004452932975254953, 'timing/model': 1.3392616129713133, 'learning_rate/qwen_vl_interface': 8.697760505540536e-06, 'learning_rate/action_model': 8.697760505540536e-05, 'epoch': 0.52}) +[2026-06-17 09:25:10,206][starVLA.training.train_starvla][INFO] - Step 1767, Loss: {'train/loss': 0.6852409243583679, 'train/grad_norm_pre_clip': 8.410051345825195, 'timing/data': 0.004466304904781282, 'timing/model': 1.2396625800756738, 'learning_rate/qwen_vl_interface': 8.696272522002614e-06, 'learning_rate/action_model': 8.696272522002615e-05, 'epoch': 0.52}) +[2026-06-17 09:25:14,976][starVLA.training.train_starvla][INFO] - Step 1768, Loss: {'train/loss': 0.9318762421607971, 'train/grad_norm_pre_clip': 10.079350471496582, 'timing/data': 0.004684485960751772, 'timing/model': 1.209748319000937, 'learning_rate/qwen_vl_interface': 8.69478382404869e-06, 'learning_rate/action_model': 8.69478382404869e-05, 'epoch': 0.52}) +[2026-06-17 09:25:19,783][starVLA.training.train_starvla][INFO] - Step 1769, Loss: {'train/loss': 0.9025087356567383, 'train/grad_norm_pre_clip': 8.328878402709961, 'timing/data': 0.004696262069046497, 'timing/model': 1.216709602973424, 'learning_rate/qwen_vl_interface': 8.693294411987371e-06, 'learning_rate/action_model': 8.693294411987371e-05, 'epoch': 0.52}) +[2026-06-17 09:25:24,619][starVLA.training.train_starvla][INFO] - Step 1770, Loss: {'train/loss': 0.8107081651687622, 'train/grad_norm_pre_clip': 10.94810962677002, 'timing/data': 0.004418703960254788, 'timing/model': 1.2041115739848465, 'learning_rate/qwen_vl_interface': 8.691804286127416e-06, 'learning_rate/action_model': 8.691804286127415e-05, 'epoch': 0.52}) +[2026-06-17 09:25:29,407][starVLA.training.train_starvla][INFO] - Step 1771, Loss: {'train/loss': 0.8765286207199097, 'train/grad_norm_pre_clip': 11.7743558883667, 'timing/data': 0.004308506031520665, 'timing/model': 1.3368995629716665, 'learning_rate/qwen_vl_interface': 8.690313446777729e-06, 'learning_rate/action_model': 8.690313446777728e-05, 'epoch': 0.52}) +[2026-06-17 09:25:34,197][starVLA.training.train_starvla][INFO] - Step 1772, Loss: {'train/loss': 1.056401014328003, 'train/grad_norm_pre_clip': 12.904994010925293, 'timing/data': 0.0047343639889732, 'timing/model': 1.3299245330272242, 'learning_rate/qwen_vl_interface': 8.688821894247363e-06, 'learning_rate/action_model': 8.688821894247362e-05, 'epoch': 0.52}) +[2026-06-17 09:25:38,860][starVLA.training.train_starvla][INFO] - Step 1773, Loss: {'train/loss': 0.88112473487854, 'train/grad_norm_pre_clip': 39.541053771972656, 'timing/data': 0.004439848009496927, 'timing/model': 1.2055396290961653, 'learning_rate/qwen_vl_interface': 8.687329628845517e-06, 'learning_rate/action_model': 8.687329628845516e-05, 'epoch': 0.52}) +[2026-06-17 09:25:43,621][starVLA.training.train_starvla][INFO] - Step 1774, Loss: {'train/loss': 1.051072120666504, 'train/grad_norm_pre_clip': 26.242189407348633, 'timing/data': 0.004626825102604926, 'timing/model': 1.2146511049941182, 'learning_rate/qwen_vl_interface': 8.68583665088154e-06, 'learning_rate/action_model': 8.68583665088154e-05, 'epoch': 0.52}) +[2026-06-17 09:25:48,501][starVLA.training.train_starvla][INFO] - Step 1775, Loss: {'train/loss': 0.9577003121376038, 'train/grad_norm_pre_clip': 10.679366111755371, 'timing/data': 0.004673053044825792, 'timing/model': 1.2059479260351509, 'learning_rate/qwen_vl_interface': 8.684342960664928e-06, 'learning_rate/action_model': 8.684342960664928e-05, 'epoch': 0.52}) +[2026-06-17 09:25:53,303][starVLA.training.train_starvla][INFO] - Step 1776, Loss: {'train/loss': 0.7755447030067444, 'train/grad_norm_pre_clip': 8.30943775177002, 'timing/data': 0.004287849995307624, 'timing/model': 1.2248605110216886, 'learning_rate/qwen_vl_interface': 8.682848558505327e-06, 'learning_rate/action_model': 8.682848558505326e-05, 'epoch': 0.52}) +[2026-06-17 09:25:58,026][starVLA.training.train_starvla][INFO] - Step 1777, Loss: {'train/loss': 0.7083488702774048, 'train/grad_norm_pre_clip': 8.234519958496094, 'timing/data': 0.004542144946753979, 'timing/model': 1.2733957800082862, 'learning_rate/qwen_vl_interface': 8.681353444712525e-06, 'learning_rate/action_model': 8.681353444712524e-05, 'epoch': 0.52}) +[2026-06-17 09:26:02,689][starVLA.training.train_starvla][INFO] - Step 1778, Loss: {'train/loss': 1.0295556783676147, 'train/grad_norm_pre_clip': 12.626153945922852, 'timing/data': 0.005012511974200606, 'timing/model': 1.2142125759273767, 'learning_rate/qwen_vl_interface': 8.67985761959646e-06, 'learning_rate/action_model': 8.67985761959646e-05, 'epoch': 0.52}) +[2026-06-17 09:26:07,477][starVLA.training.train_starvla][INFO] - Step 1779, Loss: {'train/loss': 0.7955685257911682, 'train/grad_norm_pre_clip': 9.063197135925293, 'timing/data': 0.004703402053564787, 'timing/model': 1.2137137399986386, 'learning_rate/qwen_vl_interface': 8.678361083467221e-06, 'learning_rate/action_model': 8.67836108346722e-05, 'epoch': 0.52}) +[2026-06-17 09:26:12,273][starVLA.training.train_starvla][INFO] - Step 1780, Loss: {'train/loss': 0.9457976818084717, 'train/grad_norm_pre_clip': 10.646377563476562, 'timing/data': 0.004611069918610156, 'timing/model': 1.2068934480194002, 'learning_rate/qwen_vl_interface': 8.67686383663504e-06, 'learning_rate/action_model': 8.67686383663504e-05, 'epoch': 0.52}) +[2026-06-17 09:26:17,098][starVLA.training.train_starvla][INFO] - Step 1781, Loss: {'train/loss': 0.6920928359031677, 'train/grad_norm_pre_clip': 7.616393089294434, 'timing/data': 0.004748692037537694, 'timing/model': 1.2099455060670152, 'learning_rate/qwen_vl_interface': 8.6753658794103e-06, 'learning_rate/action_model': 8.675365879410298e-05, 'epoch': 0.52}) +[2026-06-17 09:26:21,911][starVLA.training.train_starvla][INFO] - Step 1782, Loss: {'train/loss': 0.9358670115470886, 'train/grad_norm_pre_clip': 11.948599815368652, 'timing/data': 0.004611648968420923, 'timing/model': 1.354762127972208, 'learning_rate/qwen_vl_interface': 8.673867212103525e-06, 'learning_rate/action_model': 8.673867212103524e-05, 'epoch': 0.52}) +[2026-06-17 09:26:26,678][starVLA.training.train_starvla][INFO] - Step 1783, Loss: {'train/loss': 1.1817851066589355, 'train/grad_norm_pre_clip': 12.425050735473633, 'timing/data': 0.004172910936176777, 'timing/model': 1.3054912280058488, 'learning_rate/qwen_vl_interface': 8.672367835025395e-06, 'learning_rate/action_model': 8.672367835025393e-05, 'epoch': 0.52}) +[2026-06-17 09:26:31,354][starVLA.training.train_starvla][INFO] - Step 1784, Loss: {'train/loss': 0.9181716442108154, 'train/grad_norm_pre_clip': 10.025408744812012, 'timing/data': 0.004527229000814259, 'timing/model': 1.210645320941694, 'learning_rate/qwen_vl_interface': 8.670867748486727e-06, 'learning_rate/action_model': 8.670867748486726e-05, 'epoch': 0.52}) +[2026-06-17 09:26:36,151][starVLA.training.train_starvla][INFO] - Step 1785, Loss: {'train/loss': 0.8583905100822449, 'train/grad_norm_pre_clip': 8.45756721496582, 'timing/data': 0.005448875017464161, 'timing/model': 1.2065704609267414, 'learning_rate/qwen_vl_interface': 8.669366952798494e-06, 'learning_rate/action_model': 8.669366952798494e-05, 'epoch': 0.53}) +[2026-06-17 09:26:40,952][starVLA.training.train_starvla][INFO] - Step 1786, Loss: {'train/loss': 1.0779926776885986, 'train/grad_norm_pre_clip': 10.982972145080566, 'timing/data': 0.004534333944320679, 'timing/model': 1.2061794479377568, 'learning_rate/qwen_vl_interface': 8.667865448271814e-06, 'learning_rate/action_model': 8.667865448271814e-05, 'epoch': 0.53}) +[2026-06-17 09:26:45,765][starVLA.training.train_starvla][INFO] - Step 1787, Loss: {'train/loss': 0.8158683180809021, 'train/grad_norm_pre_clip': 6.860051155090332, 'timing/data': 0.004765753983519971, 'timing/model': 1.2087004199856892, 'learning_rate/qwen_vl_interface': 8.666363235217949e-06, 'learning_rate/action_model': 8.66636323521795e-05, 'epoch': 0.53}) +[2026-06-17 09:26:50,556][starVLA.training.train_starvla][INFO] - Step 1788, Loss: {'train/loss': 0.9258078932762146, 'train/grad_norm_pre_clip': 11.53014087677002, 'timing/data': 0.004526835051365197, 'timing/model': 1.2081883299397305, 'learning_rate/qwen_vl_interface': 8.66486031394831e-06, 'learning_rate/action_model': 8.664860313948309e-05, 'epoch': 0.53}) +[2026-06-17 09:26:55,299][starVLA.training.train_starvla][INFO] - Step 1789, Loss: {'train/loss': 0.873013436794281, 'train/grad_norm_pre_clip': 8.57662582397461, 'timing/data': 0.004655112978070974, 'timing/model': 1.2913384600542486, 'learning_rate/qwen_vl_interface': 8.66335668477445e-06, 'learning_rate/action_model': 8.66335668477445e-05, 'epoch': 0.53}) +[2026-06-17 09:26:59,965][starVLA.training.train_starvla][INFO] - Step 1790, Loss: {'train/loss': 0.6868126392364502, 'train/grad_norm_pre_clip': 7.33594274520874, 'timing/data': 0.004411812056787312, 'timing/model': 1.2098724619718269, 'learning_rate/qwen_vl_interface': 8.661852348008079e-06, 'learning_rate/action_model': 8.661852348008078e-05, 'epoch': 0.53}) +[2026-06-17 09:27:04,752][starVLA.training.train_starvla][INFO] - Step 1791, Loss: {'train/loss': 0.7239304780960083, 'train/grad_norm_pre_clip': 6.632890224456787, 'timing/data': 0.0048768880078569055, 'timing/model': 1.2042106409789994, 'learning_rate/qwen_vl_interface': 8.660347303961044e-06, 'learning_rate/action_model': 8.660347303961044e-05, 'epoch': 0.53}) +[2026-06-17 09:27:09,570][starVLA.training.train_starvla][INFO] - Step 1792, Loss: {'train/loss': 0.6775418519973755, 'train/grad_norm_pre_clip': 12.292686462402344, 'timing/data': 0.004458620096556842, 'timing/model': 1.2124809139641002, 'learning_rate/qwen_vl_interface': 8.658841552945342e-06, 'learning_rate/action_model': 8.658841552945342e-05, 'epoch': 0.53}) +[2026-06-17 09:27:14,386][starVLA.training.train_starvla][INFO] - Step 1793, Loss: {'train/loss': 1.2411959171295166, 'train/grad_norm_pre_clip': 12.461644172668457, 'timing/data': 0.004692141083069146, 'timing/model': 1.2061679098987952, 'learning_rate/qwen_vl_interface': 8.65733509527312e-06, 'learning_rate/action_model': 8.65733509527312e-05, 'epoch': 0.53}) +[2026-06-17 09:27:19,183][starVLA.training.train_starvla][INFO] - Step 1794, Loss: {'train/loss': 0.9696078896522522, 'train/grad_norm_pre_clip': 11.694235801696777, 'timing/data': 0.004326788010075688, 'timing/model': 1.337267330964096, 'learning_rate/qwen_vl_interface': 8.655827931256667e-06, 'learning_rate/action_model': 8.655827931256667e-05, 'epoch': 0.53}) +[2026-06-17 09:27:23,860][starVLA.training.train_starvla][INFO] - Step 1795, Loss: {'train/loss': 1.1759445667266846, 'train/grad_norm_pre_clip': 10.908239364624023, 'timing/data': 0.004497401998378336, 'timing/model': 1.206616441020742, 'learning_rate/qwen_vl_interface': 8.654320061208418e-06, 'learning_rate/action_model': 8.654320061208417e-05, 'epoch': 0.53}) +[2026-06-17 09:27:28,658][starVLA.training.train_starvla][INFO] - Step 1796, Loss: {'train/loss': 0.787477433681488, 'train/grad_norm_pre_clip': 29.82322120666504, 'timing/data': 0.004602151922881603, 'timing/model': 1.2063330119708553, 'learning_rate/qwen_vl_interface': 8.652811485440956e-06, 'learning_rate/action_model': 8.652811485440956e-05, 'epoch': 0.53}) +[2026-06-17 09:27:33,470][starVLA.training.train_starvla][INFO] - Step 1797, Loss: {'train/loss': 0.944472074508667, 'train/grad_norm_pre_clip': 82.28306579589844, 'timing/data': 0.004552925005555153, 'timing/model': 1.2121421120828018, 'learning_rate/qwen_vl_interface': 8.651302204267013e-06, 'learning_rate/action_model': 8.651302204267013e-05, 'epoch': 0.53}) +[2026-06-17 09:27:38,202][starVLA.training.train_starvla][INFO] - Step 1798, Loss: {'train/loss': 0.7215096950531006, 'train/grad_norm_pre_clip': 10.373796463012695, 'timing/data': 0.004486034973524511, 'timing/model': 1.207909179967828, 'learning_rate/qwen_vl_interface': 8.649792217999461e-06, 'learning_rate/action_model': 8.649792217999462e-05, 'epoch': 0.53}) +[2026-06-17 09:27:43,004][starVLA.training.train_starvla][INFO] - Step 1799, Loss: {'train/loss': 0.9696313738822937, 'train/grad_norm_pre_clip': 13.20543384552002, 'timing/data': 0.0045391389867290854, 'timing/model': 1.2430628139991313, 'learning_rate/qwen_vl_interface': 8.648281526951327e-06, 'learning_rate/action_model': 8.648281526951327e-05, 'epoch': 0.53}) +[2026-06-17 09:29:24,213][starVLA.training.train_starvla][INFO] - Step 1800, Loss: {'train/loss': 0.9033865928649902, 'train/grad_norm_pre_clip': 10.719136238098145, 'eval/loss': 0.8462547063827515, 'eval/latency_0/loss': 0.8462547063827515, 'eval/demon_fire_precision': 0.904228855721393, 'eval/demon_fire_recall': 0.9057308970099668, 'eval/demon_fire_f1': 0.9049792531120332, 'eval/demon_fire_rate_model': 0.6707452725250278, 'eval/demon_fire_rate_teacher': 0.6696329254727474, 'eval/demon_left_precision': 0.7898921067575241, 'eval/demon_left_recall': 0.7881019830028329, 'eval/demon_left_f1': 0.7889960294951787, 'eval/demon_left_rate_model': 0.48971078976640714, 'eval/demon_left_rate_teacher': 0.49082313681868744, 'eval/demon_right_precision': 0.784506273867976, 'eval/demon_right_recall': 0.8124293785310734, 'eval/demon_right_f1': 0.7982237024701637, 'eval/demon_right_rate_model': 0.5097330367074527, 'eval/demon_right_rate_teacher': 0.4922135706340378, 'eval/demon_move_f1': 0.7936098659826711, 'eval/demon_cc_f1': 0.8684189879472002, 'eval/latency_0/demon_fire_precision': 0.904228855721393, 'eval/latency_0/demon_fire_recall': 0.9057308970099668, 'eval/latency_0/demon_fire_f1': 0.9049792531120332, 'eval/latency_0/demon_fire_rate_model': 0.6707452725250278, 'eval/latency_0/demon_fire_rate_teacher': 0.6696329254727474, 'eval/latency_0/demon_left_precision': 0.7898921067575241, 'eval/latency_0/demon_left_recall': 0.7881019830028329, 'eval/latency_0/demon_left_f1': 0.7889960294951787, 'eval/latency_0/demon_left_rate_model': 0.48971078976640714, 'eval/latency_0/demon_left_rate_teacher': 0.49082313681868744, 'eval/latency_0/demon_right_precision': 0.784506273867976, 'eval/latency_0/demon_right_recall': 0.8124293785310734, 'eval/latency_0/demon_right_f1': 0.7982237024701637, 'eval/latency_0/demon_right_rate_model': 0.5097330367074527, 'eval/latency_0/demon_right_rate_teacher': 0.4922135706340378, 'eval/latency_0/demon_move_f1': 0.7936098659826711, 'eval/latency_0/demon_cc_f1': 0.8684189879472002, 'timing/data': 0.004691347014158964, 'timing/model': 1.3353765830397606, 'learning_rate/qwen_vl_interface': 8.646770131435774e-06, 'learning_rate/action_model': 8.646770131435773e-05, 'epoch': 0.53}) +[2026-06-17 09:29:24,213][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_1800_state +[2026-06-17 09:29:24,216][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 09:29:27,756][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_1800_state/model.safetensors +[2026-06-17 09:29:35,402][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_1800_state/optimizer.bin +[2026-06-17 09:29:35,402][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_1800_state/sampler.bin +[2026-06-17 09:29:35,402][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_1800_state/sampler_1.bin +[2026-06-17 09:29:35,403][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_1800_state/random_states_0.pkl +[2026-06-17 09:29:35,456][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1800 +[2026-06-17 09:29:35,457][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 09:29:35,483][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 09:29:40,207][starVLA.training.train_starvla][INFO] - Step 1801, Loss: {'train/loss': 1.085147500038147, 'train/grad_norm_pre_clip': 8.999738693237305, 'timing/data': 0.004639057908207178, 'timing/model': 1.3295810830313712, 'learning_rate/qwen_vl_interface': 8.645258031766116e-06, 'learning_rate/action_model': 8.645258031766116e-05, 'epoch': 0.53}) +[2026-06-17 09:29:44,821][starVLA.training.train_starvla][INFO] - Step 1802, Loss: {'train/loss': 0.5973457098007202, 'train/grad_norm_pre_clip': 7.182732582092285, 'timing/data': 0.005030378000810742, 'timing/model': 1.203911755932495, 'learning_rate/qwen_vl_interface': 8.643745228255816e-06, 'learning_rate/action_model': 8.643745228255817e-05, 'epoch': 0.53}) +[2026-06-17 09:29:49,416][starVLA.training.train_starvla][INFO] - Step 1803, Loss: {'train/loss': 0.9596961140632629, 'train/grad_norm_pre_clip': 13.048187255859375, 'timing/data': 0.004974158015102148, 'timing/model': 1.1911878159735352, 'learning_rate/qwen_vl_interface': 8.64223172121848e-06, 'learning_rate/action_model': 8.642231721218479e-05, 'epoch': 0.53}) +[2026-06-17 09:29:54,052][starVLA.training.train_starvla][INFO] - Step 1804, Loss: {'train/loss': 1.10760498046875, 'train/grad_norm_pre_clip': 11.154596328735352, 'timing/data': 0.005069041973911226, 'timing/model': 1.1910915819462389, 'learning_rate/qwen_vl_interface': 8.640717510967857e-06, 'learning_rate/action_model': 8.640717510967856e-05, 'epoch': 0.53}) +[2026-06-17 09:29:58,727][starVLA.training.train_starvla][INFO] - Step 1805, Loss: {'train/loss': 0.7037864923477173, 'train/grad_norm_pre_clip': 8.366230964660645, 'timing/data': 0.0051154280081391335, 'timing/model': 1.1910383040085435, 'learning_rate/qwen_vl_interface': 8.639202597817844e-06, 'learning_rate/action_model': 8.639202597817844e-05, 'epoch': 0.53}) +[2026-06-17 09:30:03,521][starVLA.training.train_starvla][INFO] - Step 1806, Loss: {'train/loss': 1.0027681589126587, 'train/grad_norm_pre_clip': 8.329239845275879, 'timing/data': 0.005276384996250272, 'timing/model': 1.1970850119832903, 'learning_rate/qwen_vl_interface': 8.637686982082487e-06, 'learning_rate/action_model': 8.637686982082487e-05, 'epoch': 0.53}) +[2026-06-17 09:30:08,268][starVLA.training.train_starvla][INFO] - Step 1807, Loss: {'train/loss': 0.8206239342689514, 'train/grad_norm_pre_clip': 10.60187816619873, 'timing/data': 0.005068702972494066, 'timing/model': 1.3283211530651897, 'learning_rate/qwen_vl_interface': 8.636170664075972e-06, 'learning_rate/action_model': 8.636170664075972e-05, 'epoch': 0.53}) +[2026-06-17 09:30:12,893][starVLA.training.train_starvla][INFO] - Step 1808, Loss: {'train/loss': 0.7024301886558533, 'train/grad_norm_pre_clip': 10.2428560256958, 'timing/data': 0.005139503977261484, 'timing/model': 1.1977925180690363, 'learning_rate/qwen_vl_interface': 8.634653644112636e-06, 'learning_rate/action_model': 8.634653644112635e-05, 'epoch': 0.53}) +[2026-06-17 09:30:17,644][starVLA.training.train_starvla][INFO] - Step 1809, Loss: {'train/loss': 0.6585260033607483, 'train/grad_norm_pre_clip': 9.049598693847656, 'timing/data': 0.004860940971411765, 'timing/model': 1.1965329590020701, 'learning_rate/qwen_vl_interface': 8.633135922506957e-06, 'learning_rate/action_model': 8.633135922506957e-05, 'epoch': 0.53}) +[2026-06-17 09:30:22,426][starVLA.training.train_starvla][INFO] - Step 1810, Loss: {'train/loss': 0.8599261045455933, 'train/grad_norm_pre_clip': 8.024063110351562, 'timing/data': 0.004956212011165917, 'timing/model': 1.196707791998051, 'learning_rate/qwen_vl_interface': 8.63161749957356e-06, 'learning_rate/action_model': 8.63161749957356e-05, 'epoch': 0.53}) +[2026-06-17 09:30:27,204][starVLA.training.train_starvla][INFO] - Step 1811, Loss: {'train/loss': 0.9180209636688232, 'train/grad_norm_pre_clip': 8.000925064086914, 'timing/data': 0.004915000987239182, 'timing/model': 1.1957246060483158, 'learning_rate/qwen_vl_interface': 8.630098375627216e-06, 'learning_rate/action_model': 8.630098375627217e-05, 'epoch': 0.53}) +[2026-06-17 09:30:31,928][starVLA.training.train_starvla][INFO] - Step 1812, Loss: {'train/loss': 0.9480717182159424, 'train/grad_norm_pre_clip': 10.172606468200684, 'timing/data': 0.005067300982773304, 'timing/model': 1.1949592899763957, 'learning_rate/qwen_vl_interface': 8.628578550982846e-06, 'learning_rate/action_model': 8.628578550982845e-05, 'epoch': 0.53}) +[2026-06-17 09:30:36,694][starVLA.training.train_starvla][INFO] - Step 1813, Loss: {'train/loss': 0.9005630016326904, 'train/grad_norm_pre_clip': 9.755560874938965, 'timing/data': 0.004943603067658842, 'timing/model': 1.328301458968781, 'learning_rate/qwen_vl_interface': 8.627058025955504e-06, 'learning_rate/action_model': 8.627058025955503e-05, 'epoch': 0.53}) +[2026-06-17 09:30:41,342][starVLA.training.train_starvla][INFO] - Step 1814, Loss: {'train/loss': 0.5285512804985046, 'train/grad_norm_pre_clip': 7.090073585510254, 'timing/data': 0.00488459097687155, 'timing/model': 1.197632108000107, 'learning_rate/qwen_vl_interface': 8.6255368008604e-06, 'learning_rate/action_model': 8.625536800860398e-05, 'epoch': 0.53}) +[2026-06-17 09:30:46,113][starVLA.training.train_starvla][INFO] - Step 1815, Loss: {'train/loss': 0.9279404878616333, 'train/grad_norm_pre_clip': 13.088489532470703, 'timing/data': 0.004948180983774364, 'timing/model': 1.1994030959904194, 'learning_rate/qwen_vl_interface': 8.624014876012883e-06, 'learning_rate/action_model': 8.624014876012882e-05, 'epoch': 0.53}) +[2026-06-17 09:30:50,893][starVLA.training.train_starvla][INFO] - Step 1816, Loss: {'train/loss': 0.790844738483429, 'train/grad_norm_pre_clip': 9.453046798706055, 'timing/data': 0.0050898370100185275, 'timing/model': 1.2019604990491644, 'learning_rate/qwen_vl_interface': 8.622492251728453e-06, 'learning_rate/action_model': 8.622492251728452e-05, 'epoch': 0.53}) +[2026-06-17 09:30:55,617][starVLA.training.train_starvla][INFO] - Step 1817, Loss: {'train/loss': 0.5580281019210815, 'train/grad_norm_pre_clip': 6.779331684112549, 'timing/data': 0.0048786160768941045, 'timing/model': 1.2024538309779018, 'learning_rate/qwen_vl_interface': 8.62096892832275e-06, 'learning_rate/action_model': 8.620968928322749e-05, 'epoch': 0.53}) +[2026-06-17 09:31:00,429][starVLA.training.train_starvla][INFO] - Step 1818, Loss: {'train/loss': 0.8103203177452087, 'train/grad_norm_pre_clip': 9.881467819213867, 'timing/data': 0.005100628011859953, 'timing/model': 1.2000300110084936, 'learning_rate/qwen_vl_interface': 8.619444906111562e-06, 'learning_rate/action_model': 8.619444906111561e-05, 'epoch': 0.53}) +[2026-06-17 09:31:05,205][starVLA.training.train_starvla][INFO] - Step 1819, Loss: {'train/loss': 0.8281816840171814, 'train/grad_norm_pre_clip': 10.399621963500977, 'timing/data': 0.005210899049416184, 'timing/model': 1.33773441798985, 'learning_rate/qwen_vl_interface': 8.617920185410817e-06, 'learning_rate/action_model': 8.617920185410817e-05, 'epoch': 0.54}) +[2026-06-17 09:31:09,848][starVLA.training.train_starvla][INFO] - Step 1820, Loss: {'train/loss': 0.9626835584640503, 'train/grad_norm_pre_clip': 9.888201713562012, 'timing/data': 0.0050834419671446085, 'timing/model': 1.2013548179529607, 'learning_rate/qwen_vl_interface': 8.616394766536594e-06, 'learning_rate/action_model': 8.616394766536593e-05, 'epoch': 0.54}) +[2026-06-17 09:31:14,617][starVLA.training.train_starvla][INFO] - Step 1821, Loss: {'train/loss': 0.9989281296730042, 'train/grad_norm_pre_clip': 11.226452827453613, 'timing/data': 0.005025623948313296, 'timing/model': 1.2043467069743201, 'learning_rate/qwen_vl_interface': 8.614868649805115e-06, 'learning_rate/action_model': 8.614868649805114e-05, 'epoch': 0.54}) +[2026-06-17 09:31:19,398][starVLA.training.train_starvla][INFO] - Step 1822, Loss: {'train/loss': 0.6693123579025269, 'train/grad_norm_pre_clip': 10.544469833374023, 'timing/data': 0.004870592965744436, 'timing/model': 1.2017640370177105, 'learning_rate/qwen_vl_interface': 8.613341835532743e-06, 'learning_rate/action_model': 8.613341835532743e-05, 'epoch': 0.54}) +[2026-06-17 09:31:24,147][starVLA.training.train_starvla][INFO] - Step 1823, Loss: {'train/loss': 0.9392816424369812, 'train/grad_norm_pre_clip': 9.62213134765625, 'timing/data': 0.00034616701304912567, 'timing/model': 1.201550920959562, 'learning_rate/qwen_vl_interface': 8.611814324035989e-06, 'learning_rate/action_model': 8.61181432403599e-05, 'epoch': 0.54}) +[2026-06-17 09:31:28,932][starVLA.training.train_starvla][INFO] - Step 1824, Loss: {'train/loss': 0.9368793964385986, 'train/grad_norm_pre_clip': 10.843888282775879, 'timing/data': 0.0003777510719373822, 'timing/model': 1.2032125459518284, 'learning_rate/qwen_vl_interface': 8.610286115631508e-06, 'learning_rate/action_model': 8.610286115631509e-05, 'epoch': 0.54}) +[2026-06-17 09:31:33,721][starVLA.training.train_starvla][INFO] - Step 1825, Loss: {'train/loss': 1.0660943984985352, 'train/grad_norm_pre_clip': 10.983765602111816, 'timing/data': 0.0003574569709599018, 'timing/model': 1.3486005730228499, 'learning_rate/qwen_vl_interface': 8.608757210636101e-06, 'learning_rate/action_model': 8.608757210636101e-05, 'epoch': 0.54}) +[2026-06-17 09:31:38,364][starVLA.training.train_starvla][INFO] - Step 1826, Loss: {'train/loss': 0.9223124384880066, 'train/grad_norm_pre_clip': 14.393054962158203, 'timing/data': 0.00037620996590703726, 'timing/model': 1.2008744339691475, 'learning_rate/qwen_vl_interface': 8.60722760936671e-06, 'learning_rate/action_model': 8.60722760936671e-05, 'epoch': 0.54}) +[2026-06-17 09:31:43,140][starVLA.training.train_starvla][INFO] - Step 1827, Loss: {'train/loss': 0.667462170124054, 'train/grad_norm_pre_clip': 8.478116989135742, 'timing/data': 0.0003449409268796444, 'timing/model': 1.2032822360051796, 'learning_rate/qwen_vl_interface': 8.605697312140421e-06, 'learning_rate/action_model': 8.605697312140422e-05, 'epoch': 0.54}) +[2026-06-17 09:31:47,874][starVLA.training.train_starvla][INFO] - Step 1828, Loss: {'train/loss': 0.5364868640899658, 'train/grad_norm_pre_clip': 7.684990882873535, 'timing/data': 0.00036123208701610565, 'timing/model': 1.2071017919806764, 'learning_rate/qwen_vl_interface': 8.604166319274472e-06, 'learning_rate/action_model': 8.604166319274471e-05, 'epoch': 0.54}) +[2026-06-17 09:31:52,648][starVLA.training.train_starvla][INFO] - Step 1829, Loss: {'train/loss': 0.9607529640197754, 'train/grad_norm_pre_clip': 7.837617874145508, 'timing/data': 0.00034048594534397125, 'timing/model': 1.2040519720176235, 'learning_rate/qwen_vl_interface': 8.602634631086233e-06, 'learning_rate/action_model': 8.602634631086233e-05, 'epoch': 0.54}) +[2026-06-17 09:31:57,509][starVLA.training.train_starvla][INFO] - Step 1830, Loss: {'train/loss': 0.560634195804596, 'train/grad_norm_pre_clip': 5.74641227722168, 'timing/data': 0.000347682973369956, 'timing/model': 1.2062945150537416, 'learning_rate/qwen_vl_interface': 8.601102247893227e-06, 'learning_rate/action_model': 8.601102247893226e-05, 'epoch': 0.54}) +[2026-06-17 09:32:02,309][starVLA.training.train_starvla][INFO] - Step 1831, Loss: {'train/loss': 0.5616065859794617, 'train/grad_norm_pre_clip': 8.833409309387207, 'timing/data': 0.0003474500263109803, 'timing/model': 1.3390725569333881, 'learning_rate/qwen_vl_interface': 8.599569170013121e-06, 'learning_rate/action_model': 8.59956917001312e-05, 'epoch': 0.54}) +[2026-06-17 09:32:06,956][starVLA.training.train_starvla][INFO] - Step 1832, Loss: {'train/loss': 0.590707004070282, 'train/grad_norm_pre_clip': 9.303507804870605, 'timing/data': 0.0004195489455014467, 'timing/model': 1.2031625530216843, 'learning_rate/qwen_vl_interface': 8.59803539776372e-06, 'learning_rate/action_model': 8.59803539776372e-05, 'epoch': 0.54}) +[2026-06-17 09:32:11,740][starVLA.training.train_starvla][INFO] - Step 1833, Loss: {'train/loss': 1.2275456190109253, 'train/grad_norm_pre_clip': 14.79857349395752, 'timing/data': 0.0003553200513124466, 'timing/model': 1.2056318409740925, 'learning_rate/qwen_vl_interface': 8.596500931462979e-06, 'learning_rate/action_model': 8.596500931462979e-05, 'epoch': 0.54}) +[2026-06-17 09:32:16,527][starVLA.training.train_starvla][INFO] - Step 1834, Loss: {'train/loss': 0.534675657749176, 'train/grad_norm_pre_clip': 9.069244384765625, 'timing/data': 0.00037081295158714056, 'timing/model': 1.205616605002433, 'learning_rate/qwen_vl_interface': 8.594965771428993e-06, 'learning_rate/action_model': 8.594965771428993e-05, 'epoch': 0.54}) +[2026-06-17 09:32:21,304][starVLA.training.train_starvla][INFO] - Step 1835, Loss: {'train/loss': 0.8830180764198303, 'train/grad_norm_pre_clip': 8.45922565460205, 'timing/data': 0.0004545100964605808, 'timing/model': 1.2031593370484188, 'learning_rate/qwen_vl_interface': 8.593429917980006e-06, 'learning_rate/action_model': 8.593429917980006e-05, 'epoch': 0.54}) +[2026-06-17 09:32:26,086][starVLA.training.train_starvla][INFO] - Step 1836, Loss: {'train/loss': 0.7722312808036804, 'train/grad_norm_pre_clip': 10.620735168457031, 'timing/data': 0.00035364495124667883, 'timing/model': 1.202683061012067, 'learning_rate/qwen_vl_interface': 8.591893371434397e-06, 'learning_rate/action_model': 8.591893371434396e-05, 'epoch': 0.54}) +[2026-06-17 09:32:30,945][starVLA.training.train_starvla][INFO] - Step 1837, Loss: {'train/loss': 0.8892660737037659, 'train/grad_norm_pre_clip': 11.306886672973633, 'timing/data': 0.0007950380677357316, 'timing/model': 1.3360725139500573, 'learning_rate/qwen_vl_interface': 8.590356132110695e-06, 'learning_rate/action_model': 8.590356132110695e-05, 'epoch': 0.54}) +[2026-06-17 09:32:35,600][starVLA.training.train_starvla][INFO] - Step 1838, Loss: {'train/loss': 1.1557916402816772, 'train/grad_norm_pre_clip': 12.216550827026367, 'timing/data': 0.0003552120178937912, 'timing/model': 1.2041035509901121, 'learning_rate/qwen_vl_interface': 8.588818200327575e-06, 'learning_rate/action_model': 8.588818200327575e-05, 'epoch': 0.54}) +[2026-06-17 09:32:40,379][starVLA.training.train_starvla][INFO] - Step 1839, Loss: {'train/loss': 0.8435591459274292, 'train/grad_norm_pre_clip': 9.843704223632812, 'timing/data': 0.0003489389782771468, 'timing/model': 1.2035247299354523, 'learning_rate/qwen_vl_interface': 8.587279576403848e-06, 'learning_rate/action_model': 8.587279576403847e-05, 'epoch': 0.54}) +[2026-06-17 09:32:45,155][starVLA.training.train_starvla][INFO] - Step 1840, Loss: {'train/loss': 0.6844821572303772, 'train/grad_norm_pre_clip': 9.65604019165039, 'timing/data': 0.0003667169949039817, 'timing/model': 1.2026632779743522, 'learning_rate/qwen_vl_interface': 8.585740260658475e-06, 'learning_rate/action_model': 8.585740260658475e-05, 'epoch': 0.54}) +[2026-06-17 09:32:49,950][starVLA.training.train_starvla][INFO] - Step 1841, Loss: {'train/loss': 1.1347532272338867, 'train/grad_norm_pre_clip': 10.531593322753906, 'timing/data': 0.0003514341078698635, 'timing/model': 1.2093021880136803, 'learning_rate/qwen_vl_interface': 8.584200253410557e-06, 'learning_rate/action_model': 8.584200253410556e-05, 'epoch': 0.54}) +[2026-06-17 09:32:54,698][starVLA.training.train_starvla][INFO] - Step 1842, Loss: {'train/loss': 0.8054512143135071, 'train/grad_norm_pre_clip': 10.670501708984375, 'timing/data': 0.00036357901990413666, 'timing/model': 1.2106861100764945, 'learning_rate/qwen_vl_interface': 8.582659554979336e-06, 'learning_rate/action_model': 8.582659554979336e-05, 'epoch': 0.54}) +[2026-06-17 09:32:59,492][starVLA.training.train_starvla][INFO] - Step 1843, Loss: {'train/loss': 0.6485397815704346, 'train/grad_norm_pre_clip': 11.978778839111328, 'timing/data': 0.00036031100898981094, 'timing/model': 1.3491923590190709, 'learning_rate/qwen_vl_interface': 8.581118165684204e-06, 'learning_rate/action_model': 8.581118165684204e-05, 'epoch': 0.54}) +[2026-06-17 09:33:04,140][starVLA.training.train_starvla][INFO] - Step 1844, Loss: {'train/loss': 0.8586326241493225, 'train/grad_norm_pre_clip': 10.303156852722168, 'timing/data': 0.00038817396853119135, 'timing/model': 1.2017899649217725, 'learning_rate/qwen_vl_interface': 8.579576085844695e-06, 'learning_rate/action_model': 8.579576085844694e-05, 'epoch': 0.54}) +[2026-06-17 09:33:08,929][starVLA.training.train_starvla][INFO] - Step 1845, Loss: {'train/loss': 0.7678167819976807, 'train/grad_norm_pre_clip': 8.310154914855957, 'timing/data': 0.00033075292594730854, 'timing/model': 1.2060701689915732, 'learning_rate/qwen_vl_interface': 8.578033315780474e-06, 'learning_rate/action_model': 8.578033315780475e-05, 'epoch': 0.54}) +[2026-06-17 09:33:13,711][starVLA.training.train_starvla][INFO] - Step 1846, Loss: {'train/loss': 0.7188906073570251, 'train/grad_norm_pre_clip': 9.28347110748291, 'timing/data': 0.0003254340263083577, 'timing/model': 1.2040587249211967, 'learning_rate/qwen_vl_interface': 8.576489855811371e-06, 'learning_rate/action_model': 8.57648985581137e-05, 'epoch': 0.54}) +[2026-06-17 09:33:18,482][starVLA.training.train_starvla][INFO] - Step 1847, Loss: {'train/loss': 0.8530707359313965, 'train/grad_norm_pre_clip': 15.316886901855469, 'timing/data': 0.00034980010241270065, 'timing/model': 1.2024026630679145, 'learning_rate/qwen_vl_interface': 8.57494570625734e-06, 'learning_rate/action_model': 8.57494570625734e-05, 'epoch': 0.54}) +[2026-06-17 09:33:23,235][starVLA.training.train_starvla][INFO] - Step 1848, Loss: {'train/loss': 0.8212748765945435, 'train/grad_norm_pre_clip': 10.299025535583496, 'timing/data': 0.0003621740033850074, 'timing/model': 1.20375210698694, 'learning_rate/qwen_vl_interface': 8.573400867438486e-06, 'learning_rate/action_model': 8.573400867438486e-05, 'epoch': 0.54}) +[2026-06-17 09:33:28,024][starVLA.training.train_starvla][INFO] - Step 1849, Loss: {'train/loss': 0.7489072680473328, 'train/grad_norm_pre_clip': 11.169238090515137, 'timing/data': 0.00040948507376015186, 'timing/model': 1.338466916931793, 'learning_rate/qwen_vl_interface': 8.571855339675057e-06, 'learning_rate/action_model': 8.571855339675057e-05, 'epoch': 0.54}) +[2026-06-17 09:33:32,675][starVLA.training.train_starvla][INFO] - Step 1850, Loss: {'train/loss': 0.8008151650428772, 'train/grad_norm_pre_clip': 20.211463928222656, 'timing/data': 0.00035980099346488714, 'timing/model': 1.2035311429062858, 'learning_rate/qwen_vl_interface': 8.57030912328744e-06, 'learning_rate/action_model': 8.57030912328744e-05, 'epoch': 0.54}) +[2026-06-17 09:33:37,466][starVLA.training.train_starvla][INFO] - Step 1851, Loss: {'train/loss': 0.8720055818557739, 'train/grad_norm_pre_clip': 11.164363861083984, 'timing/data': 0.0003807470202445984, 'timing/model': 1.204857102013193, 'learning_rate/qwen_vl_interface': 8.568762218596172e-06, 'learning_rate/action_model': 8.568762218596171e-05, 'epoch': 0.54}) +[2026-06-17 09:33:42,243][starVLA.training.train_starvla][INFO] - Step 1852, Loss: {'train/loss': 0.7198857069015503, 'train/grad_norm_pre_clip': 8.7783784866333, 'timing/data': 0.0003420140128582716, 'timing/model': 1.202169856056571, 'learning_rate/qwen_vl_interface': 8.567214625921922e-06, 'learning_rate/action_model': 8.567214625921922e-05, 'epoch': 0.54}) +[2026-06-17 09:33:47,027][starVLA.training.train_starvla][INFO] - Step 1853, Loss: {'train/loss': 0.8708585500717163, 'train/grad_norm_pre_clip': 10.244063377380371, 'timing/data': 0.0004247379256412387, 'timing/model': 1.204590886947699, 'learning_rate/qwen_vl_interface': 8.565666345585511e-06, 'learning_rate/action_model': 8.56566634558551e-05, 'epoch': 0.55}) +[2026-06-17 09:33:51,817][starVLA.training.train_starvla][INFO] - Step 1854, Loss: {'train/loss': 0.7573534846305847, 'train/grad_norm_pre_clip': 11.166682243347168, 'timing/data': 0.0003646200057119131, 'timing/model': 1.3376873990055174, 'learning_rate/qwen_vl_interface': 8.564117377907898e-06, 'learning_rate/action_model': 8.564117377907898e-05, 'epoch': 0.55}) +[2026-06-17 09:33:56,513][starVLA.training.train_starvla][INFO] - Step 1855, Loss: {'train/loss': 0.6799956560134888, 'train/grad_norm_pre_clip': 8.064840316772461, 'timing/data': 0.0005310389678925276, 'timing/model': 1.2537561369827017, 'learning_rate/qwen_vl_interface': 8.56256772321019e-06, 'learning_rate/action_model': 8.562567723210188e-05, 'epoch': 0.55}) +[2026-06-17 09:34:01,160][starVLA.training.train_starvla][INFO] - Step 1856, Loss: {'train/loss': 1.0328783988952637, 'train/grad_norm_pre_clip': 15.1056489944458, 'timing/data': 0.00035323004703968763, 'timing/model': 1.2044840659946203, 'learning_rate/qwen_vl_interface': 8.561017381813624e-06, 'learning_rate/action_model': 8.561017381813624e-05, 'epoch': 0.55}) +[2026-06-17 09:34:05,882][starVLA.training.train_starvla][INFO] - Step 1857, Loss: {'train/loss': 0.8331807851791382, 'train/grad_norm_pre_clip': 33.64201736450195, 'timing/data': 0.00036408903542906046, 'timing/model': 1.2024077749811113, 'learning_rate/qwen_vl_interface': 8.559466354039594e-06, 'learning_rate/action_model': 8.559466354039594e-05, 'epoch': 0.55}) +[2026-06-17 09:34:10,665][starVLA.training.train_starvla][INFO] - Step 1858, Loss: {'train/loss': 0.7642657160758972, 'train/grad_norm_pre_clip': 8.84244441986084, 'timing/data': 0.00035694800317287445, 'timing/model': 1.2061934310477227, 'learning_rate/qwen_vl_interface': 8.557914640209628e-06, 'learning_rate/action_model': 8.557914640209627e-05, 'epoch': 0.55}) +[2026-06-17 09:34:15,439][starVLA.training.train_starvla][INFO] - Step 1859, Loss: {'train/loss': 1.0355137586593628, 'train/grad_norm_pre_clip': 9.997200012207031, 'timing/data': 0.00034679495729506016, 'timing/model': 1.2026382840704173, 'learning_rate/qwen_vl_interface': 8.556362240645397e-06, 'learning_rate/action_model': 8.556362240645397e-05, 'epoch': 0.55}) +[2026-06-17 09:34:20,206][starVLA.training.train_starvla][INFO] - Step 1860, Loss: {'train/loss': 0.9466252326965332, 'train/grad_norm_pre_clip': 9.985330581665039, 'timing/data': 0.0003752961056306958, 'timing/model': 1.3288360360311344, 'learning_rate/qwen_vl_interface': 8.554809155668716e-06, 'learning_rate/action_model': 8.554809155668716e-05, 'epoch': 0.55}) +[2026-06-17 09:34:24,847][starVLA.training.train_starvla][INFO] - Step 1861, Loss: {'train/loss': 0.9066625833511353, 'train/grad_norm_pre_clip': 8.649088859558105, 'timing/data': 0.0003589709522202611, 'timing/model': 1.2012568950885907, 'learning_rate/qwen_vl_interface': 8.553255385601541e-06, 'learning_rate/action_model': 8.55325538560154e-05, 'epoch': 0.55}) +[2026-06-17 09:34:29,624][starVLA.training.train_starvla][INFO] - Step 1862, Loss: {'train/loss': 1.029694676399231, 'train/grad_norm_pre_clip': 9.462203025817871, 'timing/data': 0.00036826496943831444, 'timing/model': 1.2028901509474963, 'learning_rate/qwen_vl_interface': 8.551700930765969e-06, 'learning_rate/action_model': 8.551700930765969e-05, 'epoch': 0.55}) +[2026-06-17 09:34:34,406][starVLA.training.train_starvla][INFO] - Step 1863, Loss: {'train/loss': 0.8015344738960266, 'train/grad_norm_pre_clip': 8.967659950256348, 'timing/data': 0.0003330379258841276, 'timing/model': 1.2022118339082226, 'learning_rate/qwen_vl_interface': 8.550145791484243e-06, 'learning_rate/action_model': 8.550145791484242e-05, 'epoch': 0.55}) +[2026-06-17 09:34:39,177][starVLA.training.train_starvla][INFO] - Step 1864, Loss: {'train/loss': 1.070988655090332, 'train/grad_norm_pre_clip': 9.16386890411377, 'timing/data': 0.00043238396756350994, 'timing/model': 1.2009569950168952, 'learning_rate/qwen_vl_interface': 8.548589968078743e-06, 'learning_rate/action_model': 8.548589968078743e-05, 'epoch': 0.55}) +[2026-06-17 09:34:44,007][starVLA.training.train_starvla][INFO] - Step 1865, Loss: {'train/loss': 0.8595362901687622, 'train/grad_norm_pre_clip': 8.070680618286133, 'timing/data': 0.00041771004907786846, 'timing/model': 1.2056455879937857, 'learning_rate/qwen_vl_interface': 8.547033460871992e-06, 'learning_rate/action_model': 8.547033460871992e-05, 'epoch': 0.55}) +[2026-06-17 09:34:48,782][starVLA.training.train_starvla][INFO] - Step 1866, Loss: {'train/loss': 0.675690233707428, 'train/grad_norm_pre_clip': 8.756274223327637, 'timing/data': 0.00037603103555738926, 'timing/model': 1.334103819099255, 'learning_rate/qwen_vl_interface': 8.545476270186658e-06, 'learning_rate/action_model': 8.545476270186659e-05, 'epoch': 0.55}) +[2026-06-17 09:34:53,440][starVLA.training.train_starvla][INFO] - Step 1867, Loss: {'train/loss': 0.9398666024208069, 'train/grad_norm_pre_clip': 8.435761451721191, 'timing/data': 0.0003405259922146797, 'timing/model': 1.2020354530541226, 'learning_rate/qwen_vl_interface': 8.543918396345546e-06, 'learning_rate/action_model': 8.543918396345546e-05, 'epoch': 0.55}) +[2026-06-17 09:34:58,222][starVLA.training.train_starvla][INFO] - Step 1868, Loss: {'train/loss': 0.8574239611625671, 'train/grad_norm_pre_clip': 9.45699691772461, 'timing/data': 0.0005582839949056506, 'timing/model': 1.2101791040040553, 'learning_rate/qwen_vl_interface': 8.542359839671607e-06, 'learning_rate/action_model': 8.542359839671606e-05, 'epoch': 0.55}) +[2026-06-17 09:35:03,034][starVLA.training.train_starvla][INFO] - Step 1869, Loss: {'train/loss': 0.7882793545722961, 'train/grad_norm_pre_clip': 7.990987777709961, 'timing/data': 0.0003480709856376052, 'timing/model': 1.2038864680798724, 'learning_rate/qwen_vl_interface': 8.540800600487931e-06, 'learning_rate/action_model': 8.54080060048793e-05, 'epoch': 0.55}) +[2026-06-17 09:35:07,809][starVLA.training.train_starvla][INFO] - Step 1870, Loss: {'train/loss': 0.7194591760635376, 'train/grad_norm_pre_clip': 8.63489818572998, 'timing/data': 0.00037429400254040956, 'timing/model': 1.201474630041048, 'learning_rate/qwen_vl_interface': 8.539240679117748e-06, 'learning_rate/action_model': 8.539240679117748e-05, 'epoch': 0.55}) +[2026-06-17 09:35:12,530][starVLA.training.train_starvla][INFO] - Step 1871, Loss: {'train/loss': 0.7391648292541504, 'train/grad_norm_pre_clip': 7.516799449920654, 'timing/data': 0.00036391301546245813, 'timing/model': 1.202154119964689, 'learning_rate/qwen_vl_interface': 8.537680075884433e-06, 'learning_rate/action_model': 8.537680075884433e-05, 'epoch': 0.55}) +[2026-06-17 09:35:17,307][starVLA.training.train_starvla][INFO] - Step 1872, Loss: {'train/loss': 0.6959608793258667, 'train/grad_norm_pre_clip': 7.051557540893555, 'timing/data': 0.0003825300373136997, 'timing/model': 1.3314694659784436, 'learning_rate/qwen_vl_interface': 8.5361187911115e-06, 'learning_rate/action_model': 8.5361187911115e-05, 'epoch': 0.55}) +[2026-06-17 09:35:21,949][starVLA.training.train_starvla][INFO] - Step 1873, Loss: {'train/loss': 0.8166048526763916, 'train/grad_norm_pre_clip': 8.87084674835205, 'timing/data': 0.00039064500015228987, 'timing/model': 1.2005681609734893, 'learning_rate/qwen_vl_interface': 8.534556825122608e-06, 'learning_rate/action_model': 8.534556825122607e-05, 'epoch': 0.55}) +[2026-06-17 09:35:26,741][starVLA.training.train_starvla][INFO] - Step 1874, Loss: {'train/loss': 1.0080397129058838, 'train/grad_norm_pre_clip': 8.794382095336914, 'timing/data': 0.0003607729449868202, 'timing/model': 1.2033047599252313, 'learning_rate/qwen_vl_interface': 8.532994178241551e-06, 'learning_rate/action_model': 8.53299417824155e-05, 'epoch': 0.55}) +[2026-06-17 09:35:31,528][starVLA.training.train_starvla][INFO] - Step 1875, Loss: {'train/loss': 1.0382226705551147, 'train/grad_norm_pre_clip': 9.13188648223877, 'timing/data': 0.0005652880063280463, 'timing/model': 1.2020860259654, 'learning_rate/qwen_vl_interface': 8.531430850792267e-06, 'learning_rate/action_model': 8.531430850792266e-05, 'epoch': 0.55}) +[2026-06-17 09:35:36,300][starVLA.training.train_starvla][INFO] - Step 1876, Loss: {'train/loss': 0.7459393739700317, 'train/grad_norm_pre_clip': 7.209883689880371, 'timing/data': 0.00042574398685246706, 'timing/model': 1.2043780089588836, 'learning_rate/qwen_vl_interface': 8.529866843098838e-06, 'learning_rate/action_model': 8.529866843098837e-05, 'epoch': 0.55}) +[2026-06-17 09:35:41,066][starVLA.training.train_starvla][INFO] - Step 1877, Loss: {'train/loss': 0.6575482487678528, 'train/grad_norm_pre_clip': 8.64501953125, 'timing/data': 0.0003312809858471155, 'timing/model': 1.202224217937328, 'learning_rate/qwen_vl_interface': 8.528302155485482e-06, 'learning_rate/action_model': 8.528302155485481e-05, 'epoch': 0.55}) +[2026-06-17 09:35:45,865][starVLA.training.train_starvla][INFO] - Step 1878, Loss: {'train/loss': 0.5703635811805725, 'train/grad_norm_pre_clip': 10.366540908813477, 'timing/data': 0.00036677601747214794, 'timing/model': 1.345243044081144, 'learning_rate/qwen_vl_interface': 8.526736788276563e-06, 'learning_rate/action_model': 8.526736788276563e-05, 'epoch': 0.55}) +[2026-06-17 09:35:50,519][starVLA.training.train_starvla][INFO] - Step 1879, Loss: {'train/loss': 0.920171320438385, 'train/grad_norm_pre_clip': 8.283320426940918, 'timing/data': 0.00035881204530596733, 'timing/model': 1.2041076649911702, 'learning_rate/qwen_vl_interface': 8.525170741796578e-06, 'learning_rate/action_model': 8.525170741796579e-05, 'epoch': 0.55}) +[2026-06-17 09:35:55,306][starVLA.training.train_starvla][INFO] - Step 1880, Loss: {'train/loss': 0.9975025653839111, 'train/grad_norm_pre_clip': 10.845429420471191, 'timing/data': 0.00037388503551483154, 'timing/model': 1.2108308060560375, 'learning_rate/qwen_vl_interface': 8.523604016370179e-06, 'learning_rate/action_model': 8.523604016370179e-05, 'epoch': 0.55}) +[2026-06-17 09:36:00,092][starVLA.training.train_starvla][INFO] - Step 1881, Loss: {'train/loss': 0.8622934222221375, 'train/grad_norm_pre_clip': 8.060064315795898, 'timing/data': 0.0003738559316843748, 'timing/model': 1.2026749330107123, 'learning_rate/qwen_vl_interface': 8.522036612322143e-06, 'learning_rate/action_model': 8.522036612322142e-05, 'epoch': 0.55}) +[2026-06-17 09:36:04,876][starVLA.training.train_starvla][INFO] - Step 1882, Loss: {'train/loss': 1.1432280540466309, 'train/grad_norm_pre_clip': 11.14249324798584, 'timing/data': 0.0003990200348198414, 'timing/model': 1.2039613369852304, 'learning_rate/qwen_vl_interface': 8.520468529977395e-06, 'learning_rate/action_model': 8.520468529977395e-05, 'epoch': 0.55}) +[2026-06-17 09:36:09,672][starVLA.training.train_starvla][INFO] - Step 1883, Loss: {'train/loss': 0.8889210224151611, 'train/grad_norm_pre_clip': 7.150068283081055, 'timing/data': 0.00035168102476745844, 'timing/model': 1.3449822940165177, 'learning_rate/qwen_vl_interface': 8.518899769661002e-06, 'learning_rate/action_model': 8.518899769661002e-05, 'epoch': 0.55}) +[2026-06-17 09:36:14,371][starVLA.training.train_starvla][INFO] - Step 1884, Loss: {'train/loss': 0.8155732750892639, 'train/grad_norm_pre_clip': 10.680453300476074, 'timing/data': 0.0003702569520100951, 'timing/model': 1.2564809280447662, 'learning_rate/qwen_vl_interface': 8.517330331698171e-06, 'learning_rate/action_model': 8.51733033169817e-05, 'epoch': 0.55}) +[2026-06-17 09:36:19,033][starVLA.training.train_starvla][INFO] - Step 1885, Loss: {'train/loss': 0.9936010837554932, 'train/grad_norm_pre_clip': 11.104718208312988, 'timing/data': 0.00038430001586675644, 'timing/model': 1.2060404679505154, 'learning_rate/qwen_vl_interface': 8.515760216414244e-06, 'learning_rate/action_model': 8.515760216414244e-05, 'epoch': 0.55}) +[2026-06-17 09:36:23,828][starVLA.training.train_starvla][INFO] - Step 1886, Loss: {'train/loss': 0.6843218803405762, 'train/grad_norm_pre_clip': 9.759984970092773, 'timing/data': 0.00040471600368618965, 'timing/model': 1.2617044229991734, 'learning_rate/qwen_vl_interface': 8.514189424134712e-06, 'learning_rate/action_model': 8.514189424134713e-05, 'epoch': 0.55}) +[2026-06-17 09:36:28,604][starVLA.training.train_starvla][INFO] - Step 1887, Loss: {'train/loss': 0.8894938826560974, 'train/grad_norm_pre_clip': 8.347872734069824, 'timing/data': 0.00042416504584252834, 'timing/model': 1.2026436359155923, 'learning_rate/qwen_vl_interface': 8.512617955185199e-06, 'learning_rate/action_model': 8.5126179551852e-05, 'epoch': 0.56}) +[2026-06-17 09:36:33,379][starVLA.training.train_starvla][INFO] - Step 1888, Loss: {'train/loss': 0.7810585498809814, 'train/grad_norm_pre_clip': 18.890926361083984, 'timing/data': 0.0003782169660553336, 'timing/model': 1.2029425440123305, 'learning_rate/qwen_vl_interface': 8.511045809891472e-06, 'learning_rate/action_model': 8.511045809891472e-05, 'epoch': 0.56}) +[2026-06-17 09:36:38,173][starVLA.training.train_starvla][INFO] - Step 1889, Loss: {'train/loss': 0.6588740348815918, 'train/grad_norm_pre_clip': 9.808913230895996, 'timing/data': 0.0003645180258899927, 'timing/model': 1.3324243749957532, 'learning_rate/qwen_vl_interface': 8.509472988579443e-06, 'learning_rate/action_model': 8.509472988579443e-05, 'epoch': 0.56}) +[2026-06-17 09:36:42,858][starVLA.training.train_starvla][INFO] - Step 1890, Loss: {'train/loss': 0.9358654022216797, 'train/grad_norm_pre_clip': 8.406707763671875, 'timing/data': 0.00035547895822674036, 'timing/model': 1.2072864799993113, 'learning_rate/qwen_vl_interface': 8.507899491575153e-06, 'learning_rate/action_model': 8.507899491575154e-05, 'epoch': 0.56}) +[2026-06-17 09:36:47,626][starVLA.training.train_starvla][INFO] - Step 1891, Loss: {'train/loss': 0.8257548809051514, 'train/grad_norm_pre_clip': 8.818792343139648, 'timing/data': 0.0003429000498726964, 'timing/model': 1.2026387229561806, 'learning_rate/qwen_vl_interface': 8.506325319204795e-06, 'learning_rate/action_model': 8.506325319204796e-05, 'epoch': 0.56}) +[2026-06-17 09:36:52,426][starVLA.training.train_starvla][INFO] - Step 1892, Loss: {'train/loss': 0.9600870609283447, 'train/grad_norm_pre_clip': 10.362933158874512, 'timing/data': 0.0003595480229705572, 'timing/model': 1.2133162280078977, 'learning_rate/qwen_vl_interface': 8.504750471794694e-06, 'learning_rate/action_model': 8.504750471794693e-05, 'epoch': 0.56}) +[2026-06-17 09:36:57,263][starVLA.training.train_starvla][INFO] - Step 1893, Loss: {'train/loss': 0.6763784885406494, 'train/grad_norm_pre_clip': 11.242069244384766, 'timing/data': 0.00036024709697812796, 'timing/model': 1.201903617940843, 'learning_rate/qwen_vl_interface': 8.503174949671315e-06, 'learning_rate/action_model': 8.503174949671315e-05, 'epoch': 0.56}) +[2026-06-17 09:37:02,055][starVLA.training.train_starvla][INFO] - Step 1894, Loss: {'train/loss': 0.6928840279579163, 'train/grad_norm_pre_clip': 6.5757904052734375, 'timing/data': 0.0003613940207287669, 'timing/model': 1.2120325369760394, 'learning_rate/qwen_vl_interface': 8.501598753161273e-06, 'learning_rate/action_model': 8.501598753161272e-05, 'epoch': 0.56}) +[2026-06-17 09:37:06,845][starVLA.training.train_starvla][INFO] - Step 1895, Loss: {'train/loss': 0.8661529421806335, 'train/grad_norm_pre_clip': 9.48697566986084, 'timing/data': 0.0003779809921979904, 'timing/model': 1.3405470619909465, 'learning_rate/qwen_vl_interface': 8.500021882591308e-06, 'learning_rate/action_model': 8.500021882591308e-05, 'epoch': 0.56}) +[2026-06-17 09:37:11,497][starVLA.training.train_starvla][INFO] - Step 1896, Loss: {'train/loss': 0.8606101870536804, 'train/grad_norm_pre_clip': 8.682512283325195, 'timing/data': 0.0005084590520709753, 'timing/model': 1.2022162630455568, 'learning_rate/qwen_vl_interface': 8.498444338288308e-06, 'learning_rate/action_model': 8.498444338288309e-05, 'epoch': 0.56}) +[2026-06-17 09:37:16,298][starVLA.training.train_starvla][INFO] - Step 1897, Loss: {'train/loss': 0.7293289303779602, 'train/grad_norm_pre_clip': 9.620118141174316, 'timing/data': 0.0008618399733677506, 'timing/model': 1.2059150299755856, 'learning_rate/qwen_vl_interface': 8.496866120579304e-06, 'learning_rate/action_model': 8.496866120579303e-05, 'epoch': 0.56}) +[2026-06-17 09:37:21,113][starVLA.training.train_starvla][INFO] - Step 1898, Loss: {'train/loss': 0.9800145626068115, 'train/grad_norm_pre_clip': 8.929633140563965, 'timing/data': 0.000526105985045433, 'timing/model': 1.2095247759716585, 'learning_rate/qwen_vl_interface': 8.495287229791459e-06, 'learning_rate/action_model': 8.495287229791459e-05, 'epoch': 0.56}) +[2026-06-17 09:37:25,912][starVLA.training.train_starvla][INFO] - Step 1899, Loss: {'train/loss': 0.8325816988945007, 'train/grad_norm_pre_clip': 6.501212120056152, 'timing/data': 0.0003591750282794237, 'timing/model': 1.2029032100690529, 'learning_rate/qwen_vl_interface': 8.493707666252078e-06, 'learning_rate/action_model': 8.493707666252078e-05, 'epoch': 0.56}) +[2026-06-17 09:39:06,217][starVLA.training.train_starvla][INFO] - Step 1900, Loss: {'train/loss': 0.7464642524719238, 'train/grad_norm_pre_clip': 8.463943481445312, 'eval/loss': 0.8563010692596436, 'eval/latency_0/loss': 0.8563010692596436, 'eval/demon_fire_precision': 0.8792569659442725, 'eval/demon_fire_recall': 0.9435215946843853, 'eval/demon_fire_f1': 0.9102564102564102, 'eval/demon_fire_rate_model': 0.7185761957730812, 'eval/demon_fire_rate_teacher': 0.6696329254727474, 'eval/demon_left_precision': 0.8522809558291093, 'eval/demon_left_recall': 0.6668555240793201, 'eval/demon_left_f1': 0.7482517482517483, 'eval/demon_left_rate_model': 0.3840378197997775, 'eval/demon_left_rate_teacher': 0.49082313681868744, 'eval/demon_right_precision': 0.7256596906278435, 'eval/demon_right_recall': 0.9011299435028248, 'eval/demon_right_f1': 0.8039314516129031, 'eval/demon_right_rate_model': 0.6112347052280311, 'eval/demon_right_rate_teacher': 0.4922135706340378, 'eval/demon_move_f1': 0.7760915999323257, 'eval/demon_cc_f1': 0.8653764989725993, 'eval/latency_0/demon_fire_precision': 0.8792569659442725, 'eval/latency_0/demon_fire_recall': 0.9435215946843853, 'eval/latency_0/demon_fire_f1': 0.9102564102564102, 'eval/latency_0/demon_fire_rate_model': 0.7185761957730812, 'eval/latency_0/demon_fire_rate_teacher': 0.6696329254727474, 'eval/latency_0/demon_left_precision': 0.8522809558291093, 'eval/latency_0/demon_left_recall': 0.6668555240793201, 'eval/latency_0/demon_left_f1': 0.7482517482517483, 'eval/latency_0/demon_left_rate_model': 0.3840378197997775, 'eval/latency_0/demon_left_rate_teacher': 0.49082313681868744, 'eval/latency_0/demon_right_precision': 0.7256596906278435, 'eval/latency_0/demon_right_recall': 0.9011299435028248, 'eval/latency_0/demon_right_f1': 0.8039314516129031, 'eval/latency_0/demon_right_rate_model': 0.6112347052280311, 'eval/latency_0/demon_right_rate_teacher': 0.4922135706340378, 'eval/latency_0/demon_move_f1': 0.7760915999323257, 'eval/latency_0/demon_cc_f1': 0.8653764989725993, 'timing/data': 0.0003423239104449749, 'timing/model': 1.2060999779496342, 'learning_rate/qwen_vl_interface': 8.492127430288609e-06, 'learning_rate/action_model': 8.492127430288608e-05, 'epoch': 0.56}) +[2026-06-17 09:39:06,217][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_1900_state +[2026-06-17 09:39:06,220][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 09:39:09,709][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_1900_state/model.safetensors +[2026-06-17 09:39:17,093][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_1900_state/optimizer.bin +[2026-06-17 09:39:17,093][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_1900_state/sampler.bin +[2026-06-17 09:39:17,093][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_1900_state/sampler_1.bin +[2026-06-17 09:39:17,094][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_1900_state/random_states_0.pkl +[2026-06-17 09:39:17,144][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1900 +[2026-06-17 09:39:17,144][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 09:39:17,154][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 09:39:21,902][starVLA.training.train_starvla][INFO] - Step 1901, Loss: {'train/loss': 0.8582944869995117, 'train/grad_norm_pre_clip': 11.279176712036133, 'timing/data': 0.0003412381047382951, 'timing/model': 1.1875191070139408, 'learning_rate/qwen_vl_interface': 8.490546522228633e-06, 'learning_rate/action_model': 8.490546522228633e-05, 'epoch': 0.56}) +[2026-06-17 09:39:26,627][starVLA.training.train_starvla][INFO] - Step 1902, Loss: {'train/loss': 0.9475270509719849, 'train/grad_norm_pre_clip': 11.095346450805664, 'timing/data': 0.00037826993502676487, 'timing/model': 1.3158994580153376, 'learning_rate/qwen_vl_interface': 8.488964942399876e-06, 'learning_rate/action_model': 8.488964942399877e-05, 'epoch': 0.56}) +[2026-06-17 09:39:31,229][starVLA.training.train_starvla][INFO] - Step 1903, Loss: {'train/loss': 0.9338648915290833, 'train/grad_norm_pre_clip': 7.423999786376953, 'timing/data': 0.0003704819828271866, 'timing/model': 1.191064883954823, 'learning_rate/qwen_vl_interface': 8.487382691130203e-06, 'learning_rate/action_model': 8.487382691130202e-05, 'epoch': 0.56}) +[2026-06-17 09:39:35,977][starVLA.training.train_starvla][INFO] - Step 1904, Loss: {'train/loss': 0.7616439461708069, 'train/grad_norm_pre_clip': 9.142118453979492, 'timing/data': 0.00037338200490921736, 'timing/model': 1.2010538979666308, 'learning_rate/qwen_vl_interface': 8.485799768747614e-06, 'learning_rate/action_model': 8.485799768747614e-05, 'epoch': 0.56}) +[2026-06-17 09:39:40,616][starVLA.training.train_starvla][INFO] - Step 1905, Loss: {'train/loss': 0.6187494397163391, 'train/grad_norm_pre_clip': 6.712761878967285, 'timing/data': 0.0004146450664848089, 'timing/model': 1.1957630710676312, 'learning_rate/qwen_vl_interface': 8.484216175580248e-06, 'learning_rate/action_model': 8.484216175580249e-05, 'epoch': 0.56}) +[2026-06-17 09:39:45,284][starVLA.training.train_starvla][INFO] - Step 1906, Loss: {'train/loss': 0.7056599259376526, 'train/grad_norm_pre_clip': 8.367161750793457, 'timing/data': 0.00036320206709206104, 'timing/model': 1.1967530370457098, 'learning_rate/qwen_vl_interface': 8.482631911956395e-06, 'learning_rate/action_model': 8.482631911956395e-05, 'epoch': 0.56}) +[2026-06-17 09:39:49,991][starVLA.training.train_starvla][INFO] - Step 1907, Loss: {'train/loss': 0.8664283156394958, 'train/grad_norm_pre_clip': 12.002598762512207, 'timing/data': 0.0003776939120143652, 'timing/model': 1.1989341779844835, 'learning_rate/qwen_vl_interface': 8.481046978204462e-06, 'learning_rate/action_model': 8.481046978204463e-05, 'epoch': 0.56}) +[2026-06-17 09:39:54,752][starVLA.training.train_starvla][INFO] - Step 1908, Loss: {'train/loss': 1.1760280132293701, 'train/grad_norm_pre_clip': 9.779375076293945, 'timing/data': 0.0003733650082722306, 'timing/model': 1.3302904249867424, 'learning_rate/qwen_vl_interface': 8.479461374653017e-06, 'learning_rate/action_model': 8.479461374653018e-05, 'epoch': 0.56}) +[2026-06-17 09:39:59,380][starVLA.training.train_starvla][INFO] - Step 1909, Loss: {'train/loss': 0.9763660430908203, 'train/grad_norm_pre_clip': 8.45787525177002, 'timing/data': 0.0004218049580231309, 'timing/model': 1.1994866610039026, 'learning_rate/qwen_vl_interface': 8.477875101630752e-06, 'learning_rate/action_model': 8.477875101630753e-05, 'epoch': 0.56}) +[2026-06-17 09:40:04,149][starVLA.training.train_starvla][INFO] - Step 1910, Loss: {'train/loss': 0.7141633629798889, 'train/grad_norm_pre_clip': 12.07220458984375, 'timing/data': 0.0003249200526624918, 'timing/model': 1.2052972059464082, 'learning_rate/qwen_vl_interface': 8.476288159466508e-06, 'learning_rate/action_model': 8.476288159466508e-05, 'epoch': 0.56}) +[2026-06-17 09:40:08,934][starVLA.training.train_starvla][INFO] - Step 1911, Loss: {'train/loss': 1.0283417701721191, 'train/grad_norm_pre_clip': 8.51713752746582, 'timing/data': 0.0004159800009801984, 'timing/model': 1.2033380910288543, 'learning_rate/qwen_vl_interface': 8.474700548489255e-06, 'learning_rate/action_model': 8.474700548489255e-05, 'epoch': 0.56}) +[2026-06-17 09:40:13,689][starVLA.training.train_starvla][INFO] - Step 1912, Loss: {'train/loss': 0.980230450630188, 'train/grad_norm_pre_clip': 9.674345970153809, 'timing/data': 0.00041488592978566885, 'timing/model': 1.200526550062932, 'learning_rate/qwen_vl_interface': 8.473112269028108e-06, 'learning_rate/action_model': 8.473112269028107e-05, 'epoch': 0.56}) +[2026-06-17 09:40:18,453][starVLA.training.train_starvla][INFO] - Step 1913, Loss: {'train/loss': 0.5009739398956299, 'train/grad_norm_pre_clip': 10.897517204284668, 'timing/data': 0.0003719649976119399, 'timing/model': 1.203860781970434, 'learning_rate/qwen_vl_interface': 8.47152332141232e-06, 'learning_rate/action_model': 8.47152332141232e-05, 'epoch': 0.56}) +[2026-06-17 09:40:23,235][starVLA.training.train_starvla][INFO] - Step 1914, Loss: {'train/loss': 0.8762106895446777, 'train/grad_norm_pre_clip': 11.084203720092773, 'timing/data': 0.00033046596217900515, 'timing/model': 1.3359313909895718, 'learning_rate/qwen_vl_interface': 8.469933705971282e-06, 'learning_rate/action_model': 8.469933705971282e-05, 'epoch': 0.56}) +[2026-06-17 09:40:27,931][starVLA.training.train_starvla][INFO] - Step 1915, Loss: {'train/loss': 0.6848196983337402, 'train/grad_norm_pre_clip': 9.125340461730957, 'timing/data': 0.00031590904109179974, 'timing/model': 1.2002232549712062, 'learning_rate/qwen_vl_interface': 8.468343423034521e-06, 'learning_rate/action_model': 8.46834342303452e-05, 'epoch': 0.56}) +[2026-06-17 09:40:32,760][starVLA.training.train_starvla][INFO] - Step 1916, Loss: {'train/loss': 0.7576056718826294, 'train/grad_norm_pre_clip': 13.050655364990234, 'timing/data': 0.00038141594268381596, 'timing/model': 1.2050246660364792, 'learning_rate/qwen_vl_interface': 8.466752472931704e-06, 'learning_rate/action_model': 8.466752472931704e-05, 'epoch': 0.56}) +[2026-06-17 09:40:37,538][starVLA.training.train_starvla][INFO] - Step 1917, Loss: {'train/loss': 0.9542320370674133, 'train/grad_norm_pre_clip': 9.672483444213867, 'timing/data': 0.00037504907231777906, 'timing/model': 1.204503673943691, 'learning_rate/qwen_vl_interface': 8.465160855992639e-06, 'learning_rate/action_model': 8.465160855992638e-05, 'epoch': 0.56}) +[2026-06-17 09:40:42,313][starVLA.training.train_starvla][INFO] - Step 1918, Loss: {'train/loss': 1.1835908889770508, 'train/grad_norm_pre_clip': 16.821678161621094, 'timing/data': 0.0003753629280254245, 'timing/model': 1.2023167780134827, 'learning_rate/qwen_vl_interface': 8.46356857254727e-06, 'learning_rate/action_model': 8.46356857254727e-05, 'epoch': 0.56}) +[2026-06-17 09:40:47,098][starVLA.training.train_starvla][INFO] - Step 1919, Loss: {'train/loss': 0.6448772549629211, 'train/grad_norm_pre_clip': 8.991174697875977, 'timing/data': 0.00038678001146763563, 'timing/model': 1.329271161928773, 'learning_rate/qwen_vl_interface': 8.461975622925676e-06, 'learning_rate/action_model': 8.461975622925676e-05, 'epoch': 0.56}) +[2026-06-17 09:40:51,790][starVLA.training.train_starvla][INFO] - Step 1920, Loss: {'train/loss': 0.8076894283294678, 'train/grad_norm_pre_clip': 11.299680709838867, 'timing/data': 0.00038310710806399584, 'timing/model': 1.241567136021331, 'learning_rate/qwen_vl_interface': 8.460382007458081e-06, 'learning_rate/action_model': 8.460382007458081e-05, 'epoch': 0.56}) +[2026-06-17 09:40:56,442][starVLA.training.train_starvla][INFO] - Step 1921, Loss: {'train/loss': 1.1804144382476807, 'train/grad_norm_pre_clip': 9.54843807220459, 'timing/data': 0.00037685607094317675, 'timing/model': 1.202948548947461, 'learning_rate/qwen_vl_interface': 8.45878772647484e-06, 'learning_rate/action_model': 8.458787726474839e-05, 'epoch': 0.57}) +[2026-06-17 09:41:01,177][starVLA.training.train_starvla][INFO] - Step 1922, Loss: {'train/loss': 0.9820032119750977, 'train/grad_norm_pre_clip': 10.970246315002441, 'timing/data': 0.00035321200266480446, 'timing/model': 1.204508455004543, 'learning_rate/qwen_vl_interface': 8.457192780306452e-06, 'learning_rate/action_model': 8.457192780306451e-05, 'epoch': 0.57}) +[2026-06-17 09:41:05,959][starVLA.training.train_starvla][INFO] - Step 1923, Loss: {'train/loss': 1.059893012046814, 'train/grad_norm_pre_clip': 12.650415420532227, 'timing/data': 0.0005707310047000647, 'timing/model': 1.2039043619297445, 'learning_rate/qwen_vl_interface': 8.455597169283549e-06, 'learning_rate/action_model': 8.455597169283548e-05, 'epoch': 0.57}) +[2026-06-17 09:41:10,909][starVLA.training.train_starvla][INFO] - Step 1924, Loss: {'train/loss': 0.8236439228057861, 'train/grad_norm_pre_clip': 21.798376083374023, 'timing/data': 0.0003657740307971835, 'timing/model': 1.20738147594966, 'learning_rate/qwen_vl_interface': 8.454000893736901e-06, 'learning_rate/action_model': 8.4540008937369e-05, 'epoch': 0.57}) +[2026-06-17 09:41:15,695][starVLA.training.train_starvla][INFO] - Step 1925, Loss: {'train/loss': 0.816078245639801, 'train/grad_norm_pre_clip': 7.420344829559326, 'timing/data': 0.00035702704917639494, 'timing/model': 1.3352755659725517, 'learning_rate/qwen_vl_interface': 8.452403953997422e-06, 'learning_rate/action_model': 8.452403953997422e-05, 'epoch': 0.57}) +[2026-06-17 09:41:20,428][starVLA.training.train_starvla][INFO] - Step 1926, Loss: {'train/loss': 0.6552659869194031, 'train/grad_norm_pre_clip': 9.006898880004883, 'timing/data': 0.0004934880416840315, 'timing/model': 1.2076960889389738, 'learning_rate/qwen_vl_interface': 8.450806350396158e-06, 'learning_rate/action_model': 8.450806350396158e-05, 'epoch': 0.57}) +[2026-06-17 09:41:25,320][starVLA.training.train_starvla][INFO] - Step 1927, Loss: {'train/loss': 0.808786928653717, 'train/grad_norm_pre_clip': 7.707031726837158, 'timing/data': 0.0004128720611333847, 'timing/model': 1.2054755890276283, 'learning_rate/qwen_vl_interface': 8.449208083264291e-06, 'learning_rate/action_model': 8.449208083264292e-05, 'epoch': 0.57}) +[2026-06-17 09:41:30,208][starVLA.training.train_starvla][INFO] - Step 1928, Loss: {'train/loss': 0.7093229293823242, 'train/grad_norm_pre_clip': 8.197500228881836, 'timing/data': 0.0003468180075287819, 'timing/model': 1.2056433509569615, 'learning_rate/qwen_vl_interface': 8.447609152933149e-06, 'learning_rate/action_model': 8.447609152933149e-05, 'epoch': 0.57}) +[2026-06-17 09:41:35,016][starVLA.training.train_starvla][INFO] - Step 1929, Loss: {'train/loss': 0.9439107775688171, 'train/grad_norm_pre_clip': 9.625375747680664, 'timing/data': 0.00035822298377752304, 'timing/model': 1.2039493980119005, 'learning_rate/qwen_vl_interface': 8.44600955973419e-06, 'learning_rate/action_model': 8.446009559734188e-05, 'epoch': 0.57}) +[2026-06-17 09:41:39,792][starVLA.training.train_starvla][INFO] - Step 1930, Loss: {'train/loss': 0.6670225858688354, 'train/grad_norm_pre_clip': 9.131377220153809, 'timing/data': 0.00037783291190862656, 'timing/model': 1.2027141849976033, 'learning_rate/qwen_vl_interface': 8.444409303999005e-06, 'learning_rate/action_model': 8.444409303999005e-05, 'epoch': 0.57}) +[2026-06-17 09:41:44,600][starVLA.training.train_starvla][INFO] - Step 1931, Loss: {'train/loss': 0.9276866316795349, 'train/grad_norm_pre_clip': 8.277832984924316, 'timing/data': 0.0003394669620320201, 'timing/model': 1.3437842020066455, 'learning_rate/qwen_vl_interface': 8.442808386059337e-06, 'learning_rate/action_model': 8.442808386059337e-05, 'epoch': 0.57}) +[2026-06-17 09:41:49,268][starVLA.training.train_starvla][INFO] - Step 1932, Loss: {'train/loss': 0.9020907878875732, 'train/grad_norm_pre_clip': 8.450948715209961, 'timing/data': 0.0005887229926884174, 'timing/model': 1.205160730984062, 'learning_rate/qwen_vl_interface': 8.441206806247055e-06, 'learning_rate/action_model': 8.441206806247054e-05, 'epoch': 0.57}) +[2026-06-17 09:41:54,080][starVLA.training.train_starvla][INFO] - Step 1933, Loss: {'train/loss': 0.5836420059204102, 'train/grad_norm_pre_clip': 7.0599565505981445, 'timing/data': 0.00037649902515113354, 'timing/model': 1.2051274429541081, 'learning_rate/qwen_vl_interface': 8.439604564894167e-06, 'learning_rate/action_model': 8.439604564894167e-05, 'epoch': 0.57}) +[2026-06-17 09:41:58,982][starVLA.training.train_starvla][INFO] - Step 1934, Loss: {'train/loss': 0.8662040829658508, 'train/grad_norm_pre_clip': 13.01852035522461, 'timing/data': 0.0005127480253577232, 'timing/model': 1.2116187950596213, 'learning_rate/qwen_vl_interface': 8.438001662332821e-06, 'learning_rate/action_model': 8.43800166233282e-05, 'epoch': 0.57}) +[2026-06-17 09:42:03,824][starVLA.training.train_starvla][INFO] - Step 1935, Loss: {'train/loss': 1.0686691999435425, 'train/grad_norm_pre_clip': 13.533215522766113, 'timing/data': 0.0003697420470416546, 'timing/model': 1.2056685420684516, 'learning_rate/qwen_vl_interface': 8.4363980988953e-06, 'learning_rate/action_model': 8.436398098895301e-05, 'epoch': 0.57}) +[2026-06-17 09:42:08,606][starVLA.training.train_starvla][INFO] - Step 1936, Loss: {'train/loss': 0.7227416038513184, 'train/grad_norm_pre_clip': 7.629142761230469, 'timing/data': 0.0003922659670934081, 'timing/model': 1.2054166999878362, 'learning_rate/qwen_vl_interface': 8.434793874914026e-06, 'learning_rate/action_model': 8.434793874914026e-05, 'epoch': 0.57}) +[2026-06-17 09:42:13,391][starVLA.training.train_starvla][INFO] - Step 1937, Loss: {'train/loss': 1.0009468793869019, 'train/grad_norm_pre_clip': 9.71339225769043, 'timing/data': 0.00036544003523886204, 'timing/model': 1.3336508050560951, 'learning_rate/qwen_vl_interface': 8.433188990721553e-06, 'learning_rate/action_model': 8.433188990721552e-05, 'epoch': 0.57}) +[2026-06-17 09:42:18,124][starVLA.training.train_starvla][INFO] - Step 1938, Loss: {'train/loss': 1.0766557455062866, 'train/grad_norm_pre_clip': 9.78322696685791, 'timing/data': 0.0003961570328101516, 'timing/model': 1.2688525749836117, 'learning_rate/qwen_vl_interface': 8.431583446650577e-06, 'learning_rate/action_model': 8.431583446650578e-05, 'epoch': 0.57}) +[2026-06-17 09:42:23,017][starVLA.training.train_starvla][INFO] - Step 1939, Loss: {'train/loss': 0.8297116160392761, 'train/grad_norm_pre_clip': 8.509909629821777, 'timing/data': 0.00039864901918917894, 'timing/model': 1.2126150040421635, 'learning_rate/qwen_vl_interface': 8.429977243033928e-06, 'learning_rate/action_model': 8.429977243033928e-05, 'epoch': 0.57}) +[2026-06-17 09:42:27,821][starVLA.training.train_starvla][INFO] - Step 1940, Loss: {'train/loss': 0.8660181760787964, 'train/grad_norm_pre_clip': 7.324880123138428, 'timing/data': 0.00038499897345900536, 'timing/model': 1.2101366019342095, 'learning_rate/qwen_vl_interface': 8.428370380204577e-06, 'learning_rate/action_model': 8.428370380204577e-05, 'epoch': 0.57}) +[2026-06-17 09:42:32,623][starVLA.training.train_starvla][INFO] - Step 1941, Loss: {'train/loss': 0.7064470052719116, 'train/grad_norm_pre_clip': 8.608992576599121, 'timing/data': 0.000369424931704998, 'timing/model': 1.2116288180695847, 'learning_rate/qwen_vl_interface': 8.426762858495625e-06, 'learning_rate/action_model': 8.426762858495625e-05, 'epoch': 0.57}) +[2026-06-17 09:42:37,394][starVLA.training.train_starvla][INFO] - Step 1942, Loss: {'train/loss': 0.4747544825077057, 'train/grad_norm_pre_clip': 9.24759292602539, 'timing/data': 0.0006464279722422361, 'timing/model': 1.208952417015098, 'learning_rate/qwen_vl_interface': 8.425154678240315e-06, 'learning_rate/action_model': 8.425154678240314e-05, 'epoch': 0.57}) +[2026-06-17 09:42:42,218][starVLA.training.train_starvla][INFO] - Step 1943, Loss: {'train/loss': 0.961091160774231, 'train/grad_norm_pre_clip': 11.358314514160156, 'timing/data': 0.00032585998997092247, 'timing/model': 1.34797735104803, 'learning_rate/qwen_vl_interface': 8.423545839772021e-06, 'learning_rate/action_model': 8.423545839772021e-05, 'epoch': 0.57}) +[2026-06-17 09:42:46,889][starVLA.training.train_starvla][INFO] - Step 1944, Loss: {'train/loss': 0.8586134314537048, 'train/grad_norm_pre_clip': 7.837164402008057, 'timing/data': 0.00036530999932438135, 'timing/model': 1.2134202040033415, 'learning_rate/qwen_vl_interface': 8.421936343424264e-06, 'learning_rate/action_model': 8.421936343424262e-05, 'epoch': 0.57}) +[2026-06-17 09:42:51,671][starVLA.training.train_starvla][INFO] - Step 1945, Loss: {'train/loss': 0.8910219073295593, 'train/grad_norm_pre_clip': 10.020118713378906, 'timing/data': 0.00039489101618528366, 'timing/model': 1.2055088609922677, 'learning_rate/qwen_vl_interface': 8.420326189530685e-06, 'learning_rate/action_model': 8.420326189530685e-05, 'epoch': 0.57}) +[2026-06-17 09:42:56,477][starVLA.training.train_starvla][INFO] - Step 1946, Loss: {'train/loss': 0.9714986085891724, 'train/grad_norm_pre_clip': 11.188864707946777, 'timing/data': 0.00036798801738768816, 'timing/model': 1.2047937280731276, 'learning_rate/qwen_vl_interface': 8.418715378425078e-06, 'learning_rate/action_model': 8.418715378425078e-05, 'epoch': 0.57}) +[2026-06-17 09:43:01,331][starVLA.training.train_starvla][INFO] - Step 1947, Loss: {'train/loss': 1.047903299331665, 'train/grad_norm_pre_clip': 11.301513671875, 'timing/data': 0.0003597959876060486, 'timing/model': 1.2125322580104694, 'learning_rate/qwen_vl_interface': 8.417103910441364e-06, 'learning_rate/action_model': 8.417103910441363e-05, 'epoch': 0.57}) +[2026-06-17 09:43:06,111][starVLA.training.train_starvla][INFO] - Step 1948, Loss: {'train/loss': 0.7836328148841858, 'train/grad_norm_pre_clip': 8.113450050354004, 'timing/data': 0.0003323199925944209, 'timing/model': 1.3310632939683273, 'learning_rate/qwen_vl_interface': 8.415491785913599e-06, 'learning_rate/action_model': 8.4154917859136e-05, 'epoch': 0.57}) +[2026-06-17 09:43:10,814][starVLA.training.train_starvla][INFO] - Step 1949, Loss: {'train/loss': 0.6927922368049622, 'train/grad_norm_pre_clip': 7.46888542175293, 'timing/data': 0.00034725305158644915, 'timing/model': 1.2490538960555568, 'learning_rate/qwen_vl_interface': 8.413879005175983e-06, 'learning_rate/action_model': 8.413879005175983e-05, 'epoch': 0.57}) +[2026-06-17 09:43:15,480][starVLA.training.train_starvla][INFO] - Step 1950, Loss: {'train/loss': 0.7088141441345215, 'train/grad_norm_pre_clip': 9.270008087158203, 'timing/data': 0.0003697649808600545, 'timing/model': 1.2042930700117722, 'learning_rate/qwen_vl_interface': 8.412265568562843e-06, 'learning_rate/action_model': 8.412265568562842e-05, 'epoch': 0.57}) +[2026-06-17 09:43:20,226][starVLA.training.train_starvla][INFO] - Step 1951, Loss: {'train/loss': 0.8119879961013794, 'train/grad_norm_pre_clip': 10.108526229858398, 'timing/data': 0.0004622719716280699, 'timing/model': 1.208390787942335, 'learning_rate/qwen_vl_interface': 8.41065147640865e-06, 'learning_rate/action_model': 8.41065147640865e-05, 'epoch': 0.57}) +[2026-06-17 09:43:25,019][starVLA.training.train_starvla][INFO] - Step 1952, Loss: {'train/loss': 0.7727844715118408, 'train/grad_norm_pre_clip': 9.409969329833984, 'timing/data': 0.0003640359500423074, 'timing/model': 1.2065885870251805, 'learning_rate/qwen_vl_interface': 8.409036729048002e-06, 'learning_rate/action_model': 8.409036729048002e-05, 'epoch': 0.57}) +[2026-06-17 09:43:29,809][starVLA.training.train_starvla][INFO] - Step 1953, Loss: {'train/loss': 0.8406853079795837, 'train/grad_norm_pre_clip': 9.102957725524902, 'timing/data': 0.0003628319827839732, 'timing/model': 1.2053171310108155, 'learning_rate/qwen_vl_interface': 8.407421326815644e-06, 'learning_rate/action_model': 8.407421326815644e-05, 'epoch': 0.57}) +[2026-06-17 09:43:34,640][starVLA.training.train_starvla][INFO] - Step 1954, Loss: {'train/loss': 0.8794131278991699, 'train/grad_norm_pre_clip': 8.778956413269043, 'timing/data': 0.00034743105061352253, 'timing/model': 1.3368865529773757, 'learning_rate/qwen_vl_interface': 8.405805270046445e-06, 'learning_rate/action_model': 8.405805270046445e-05, 'epoch': 0.57}) +[2026-06-17 09:43:39,308][starVLA.training.train_starvla][INFO] - Step 1955, Loss: {'train/loss': 0.9023601412773132, 'train/grad_norm_pre_clip': 8.755532264709473, 'timing/data': 0.0003726179711520672, 'timing/model': 1.204757662024349, 'learning_rate/qwen_vl_interface': 8.404188559075419e-06, 'learning_rate/action_model': 8.404188559075419e-05, 'epoch': 0.58}) +[2026-06-17 09:43:44,087][starVLA.training.train_starvla][INFO] - Step 1956, Loss: {'train/loss': 0.7845655083656311, 'train/grad_norm_pre_clip': 11.31270694732666, 'timing/data': 0.00036808499135077, 'timing/model': 1.204042315017432, 'learning_rate/qwen_vl_interface': 8.40257119423771e-06, 'learning_rate/action_model': 8.402571194237709e-05, 'epoch': 0.58}) +[2026-06-17 09:43:48,865][starVLA.training.train_starvla][INFO] - Step 1957, Loss: {'train/loss': 0.7520301342010498, 'train/grad_norm_pre_clip': 8.187402725219727, 'timing/data': 0.00040323298890143633, 'timing/model': 1.2055515419924632, 'learning_rate/qwen_vl_interface': 8.4009531758686e-06, 'learning_rate/action_model': 8.4009531758686e-05, 'epoch': 0.58}) +[2026-06-17 09:43:53,648][starVLA.training.train_starvla][INFO] - Step 1958, Loss: {'train/loss': 0.8979963660240173, 'train/grad_norm_pre_clip': 15.613303184509277, 'timing/data': 0.00033917895052582026, 'timing/model': 1.2052751570008695, 'learning_rate/qwen_vl_interface': 8.399334504303507e-06, 'learning_rate/action_model': 8.399334504303507e-05, 'epoch': 0.58}) +[2026-06-17 09:43:58,435][starVLA.training.train_starvla][INFO] - Step 1959, Loss: {'train/loss': 1.2296323776245117, 'train/grad_norm_pre_clip': 8.508185386657715, 'timing/data': 0.00033751793671399355, 'timing/model': 1.2040609329706058, 'learning_rate/qwen_vl_interface': 8.397715179877982e-06, 'learning_rate/action_model': 8.397715179877982e-05, 'epoch': 0.58}) +[2026-06-17 09:44:03,219][starVLA.training.train_starvla][INFO] - Step 1960, Loss: {'train/loss': 0.8534672856330872, 'train/grad_norm_pre_clip': 9.715425491333008, 'timing/data': 0.00034760194830596447, 'timing/model': 1.3335144879529253, 'learning_rate/qwen_vl_interface': 8.396095202927712e-06, 'learning_rate/action_model': 8.396095202927712e-05, 'epoch': 0.58}) +[2026-06-17 09:44:07,871][starVLA.training.train_starvla][INFO] - Step 1961, Loss: {'train/loss': 0.7559118270874023, 'train/grad_norm_pre_clip': 9.468147277832031, 'timing/data': 0.0004474489251151681, 'timing/model': 1.2044363330351189, 'learning_rate/qwen_vl_interface': 8.394474573788522e-06, 'learning_rate/action_model': 8.394474573788522e-05, 'epoch': 0.58}) +[2026-06-17 09:44:12,664][starVLA.training.train_starvla][INFO] - Step 1962, Loss: {'train/loss': 0.7943921685218811, 'train/grad_norm_pre_clip': 8.219560623168945, 'timing/data': 0.0004735219990834594, 'timing/model': 1.2037274980684742, 'learning_rate/qwen_vl_interface': 8.39285329279637e-06, 'learning_rate/action_model': 8.39285329279637e-05, 'epoch': 0.58}) +[2026-06-17 09:44:17,492][starVLA.training.train_starvla][INFO] - Step 1963, Loss: {'train/loss': 0.9564481973648071, 'train/grad_norm_pre_clip': 10.08476448059082, 'timing/data': 0.0003794049844145775, 'timing/model': 1.212946059065871, 'learning_rate/qwen_vl_interface': 8.391231360287348e-06, 'learning_rate/action_model': 8.391231360287347e-05, 'epoch': 0.58}) +[2026-06-17 09:44:22,323][starVLA.training.train_starvla][INFO] - Step 1964, Loss: {'train/loss': 0.8161714673042297, 'train/grad_norm_pre_clip': 7.102302074432373, 'timing/data': 0.0003471469972282648, 'timing/model': 1.210786952986382, 'learning_rate/qwen_vl_interface': 8.389608776597684e-06, 'learning_rate/action_model': 8.389608776597684e-05, 'epoch': 0.58}) +[2026-06-17 09:44:27,067][starVLA.training.train_starvla][INFO] - Step 1965, Loss: {'train/loss': 0.688288152217865, 'train/grad_norm_pre_clip': 9.131308555603027, 'timing/data': 0.00034886994399130344, 'timing/model': 1.207232236978598, 'learning_rate/qwen_vl_interface': 8.387985542063743e-06, 'learning_rate/action_model': 8.387985542063743e-05, 'epoch': 0.58}) +[2026-06-17 09:44:31,903][starVLA.training.train_starvla][INFO] - Step 1966, Loss: {'train/loss': 0.6141367554664612, 'train/grad_norm_pre_clip': 6.174773693084717, 'timing/data': 0.00036923098377883434, 'timing/model': 1.352144390111789, 'learning_rate/qwen_vl_interface': 8.386361657022024e-06, 'learning_rate/action_model': 8.386361657022023e-05, 'epoch': 0.58}) +[2026-06-17 09:44:36,564][starVLA.training.train_starvla][INFO] - Step 1967, Loss: {'train/loss': 0.8579780459403992, 'train/grad_norm_pre_clip': 9.06307601928711, 'timing/data': 0.0005285079823806882, 'timing/model': 1.2048613290535286, 'learning_rate/qwen_vl_interface': 8.384737121809156e-06, 'learning_rate/action_model': 8.384737121809156e-05, 'epoch': 0.58}) +[2026-06-17 09:44:41,352][starVLA.training.train_starvla][INFO] - Step 1968, Loss: {'train/loss': 0.8047233819961548, 'train/grad_norm_pre_clip': 8.22786808013916, 'timing/data': 0.0003602360375225544, 'timing/model': 1.2071468599606305, 'learning_rate/qwen_vl_interface': 8.38311193676191e-06, 'learning_rate/action_model': 8.383111936761909e-05, 'epoch': 0.58}) +[2026-06-17 09:44:46,152][starVLA.training.train_starvla][INFO] - Step 1969, Loss: {'train/loss': 0.849198579788208, 'train/grad_norm_pre_clip': 10.838040351867676, 'timing/data': 0.000372081995010376, 'timing/model': 1.211803791928105, 'learning_rate/qwen_vl_interface': 8.381486102217188e-06, 'learning_rate/action_model': 8.381486102217188e-05, 'epoch': 0.58}) +[2026-06-17 09:44:50,943][starVLA.training.train_starvla][INFO] - Step 1970, Loss: {'train/loss': 0.8097274303436279, 'train/grad_norm_pre_clip': 8.000764846801758, 'timing/data': 0.00034341891296207905, 'timing/model': 1.207840693066828, 'learning_rate/qwen_vl_interface': 8.379859618512029e-06, 'learning_rate/action_model': 8.379859618512029e-05, 'epoch': 0.58}) +[2026-06-17 09:44:55,734][starVLA.training.train_starvla][INFO] - Step 1971, Loss: {'train/loss': 0.9102339744567871, 'train/grad_norm_pre_clip': 9.256917953491211, 'timing/data': 0.0003869169158861041, 'timing/model': 1.2047059920150787, 'learning_rate/qwen_vl_interface': 8.378232485983603e-06, 'learning_rate/action_model': 8.378232485983603e-05, 'epoch': 0.58}) +[2026-06-17 09:45:00,517][starVLA.training.train_starvla][INFO] - Step 1972, Loss: {'train/loss': 0.7600301504135132, 'train/grad_norm_pre_clip': 8.429884910583496, 'timing/data': 0.0003593950532376766, 'timing/model': 1.3330769910244271, 'learning_rate/qwen_vl_interface': 8.376604704969216e-06, 'learning_rate/action_model': 8.376604704969216e-05, 'epoch': 0.58}) +[2026-06-17 09:45:05,197][starVLA.training.train_starvla][INFO] - Step 1973, Loss: {'train/loss': 0.629483163356781, 'train/grad_norm_pre_clip': 7.0220513343811035, 'timing/data': 0.00041532807517796755, 'timing/model': 1.2077406879980117, 'learning_rate/qwen_vl_interface': 8.374976275806307e-06, 'learning_rate/action_model': 8.374976275806308e-05, 'epoch': 0.58}) +[2026-06-17 09:45:10,011][starVLA.training.train_starvla][INFO] - Step 1974, Loss: {'train/loss': 0.6395109295845032, 'train/grad_norm_pre_clip': 7.972295761108398, 'timing/data': 0.0003890719963237643, 'timing/model': 1.2135801480617374, 'learning_rate/qwen_vl_interface': 8.373347198832457e-06, 'learning_rate/action_model': 8.373347198832457e-05, 'epoch': 0.58}) +[2026-06-17 09:45:14,840][starVLA.training.train_starvla][INFO] - Step 1975, Loss: {'train/loss': 0.7485814690589905, 'train/grad_norm_pre_clip': 11.172611236572266, 'timing/data': 0.00036076700780540705, 'timing/model': 1.2136392359388992, 'learning_rate/qwen_vl_interface': 8.37171747438537e-06, 'learning_rate/action_model': 8.37171747438537e-05, 'epoch': 0.58}) +[2026-06-17 09:45:19,650][starVLA.training.train_starvla][INFO] - Step 1976, Loss: {'train/loss': 0.8851938843727112, 'train/grad_norm_pre_clip': 8.614094734191895, 'timing/data': 0.0003187709953635931, 'timing/model': 1.204463304951787, 'learning_rate/qwen_vl_interface': 8.370087102802892e-06, 'learning_rate/action_model': 8.370087102802892e-05, 'epoch': 0.58}) +[2026-06-17 09:45:24,429][starVLA.training.train_starvla][INFO] - Step 1977, Loss: {'train/loss': 0.9426851272583008, 'train/grad_norm_pre_clip': 11.58347225189209, 'timing/data': 0.0003450610674917698, 'timing/model': 1.332389852963388, 'learning_rate/qwen_vl_interface': 8.368456084423e-06, 'learning_rate/action_model': 8.368456084422999e-05, 'epoch': 0.58}) +[2026-06-17 09:45:29,096][starVLA.training.train_starvla][INFO] - Step 1978, Loss: {'train/loss': 0.8682012557983398, 'train/grad_norm_pre_clip': 9.544160842895508, 'timing/data': 0.0003351030172780156, 'timing/model': 1.2056834950344637, 'learning_rate/qwen_vl_interface': 8.366824419583806e-06, 'learning_rate/action_model': 8.366824419583806e-05, 'epoch': 0.58}) +[2026-06-17 09:45:33,892][starVLA.training.train_starvla][INFO] - Step 1979, Loss: {'train/loss': 0.953184962272644, 'train/grad_norm_pre_clip': 13.020251274108887, 'timing/data': 0.0003820930141955614, 'timing/model': 1.2135011670179665, 'learning_rate/qwen_vl_interface': 8.365192108623555e-06, 'learning_rate/action_model': 8.365192108623555e-05, 'epoch': 0.58}) +[2026-06-17 09:45:38,618][starVLA.training.train_starvla][INFO] - Step 1980, Loss: {'train/loss': 1.1118751764297485, 'train/grad_norm_pre_clip': 8.334331512451172, 'timing/data': 0.0005237880395725369, 'timing/model': 1.2039602410513908, 'learning_rate/qwen_vl_interface': 8.363559151880631e-06, 'learning_rate/action_model': 8.36355915188063e-05, 'epoch': 0.58}) +[2026-06-17 09:45:43,378][starVLA.training.train_starvla][INFO] - Step 1981, Loss: {'train/loss': 0.5553529262542725, 'train/grad_norm_pre_clip': 6.374904632568359, 'timing/data': 0.0003311140462756157, 'timing/model': 1.204178993939422, 'learning_rate/qwen_vl_interface': 8.361925549693541e-06, 'learning_rate/action_model': 8.36192554969354e-05, 'epoch': 0.58}) +[2026-06-17 09:45:48,164][starVLA.training.train_starvla][INFO] - Step 1982, Loss: {'train/loss': 1.0684094429016113, 'train/grad_norm_pre_clip': 9.609318733215332, 'timing/data': 0.00035353703424334526, 'timing/model': 1.2044638510560617, 'learning_rate/qwen_vl_interface': 8.360291302400937e-06, 'learning_rate/action_model': 8.360291302400937e-05, 'epoch': 0.58}) +[2026-06-17 09:45:52,985][starVLA.training.train_starvla][INFO] - Step 1983, Loss: {'train/loss': 0.7974937558174133, 'train/grad_norm_pre_clip': 7.540820121765137, 'timing/data': 0.0003755579236894846, 'timing/model': 1.3471404489828274, 'learning_rate/qwen_vl_interface': 8.3586564103416e-06, 'learning_rate/action_model': 8.358656410341599e-05, 'epoch': 0.58}) +[2026-06-17 09:45:57,645][starVLA.training.train_starvla][INFO] - Step 1984, Loss: {'train/loss': 0.925632655620575, 'train/grad_norm_pre_clip': 17.158483505249023, 'timing/data': 0.0003444820176810026, 'timing/model': 1.207125277025625, 'learning_rate/qwen_vl_interface': 8.357020873854445e-06, 'learning_rate/action_model': 8.357020873854444e-05, 'epoch': 0.58}) +[2026-06-17 09:46:02,427][starVLA.training.train_starvla][INFO] - Step 1985, Loss: {'train/loss': 0.8429579734802246, 'train/grad_norm_pre_clip': 8.536715507507324, 'timing/data': 0.0005165099864825606, 'timing/model': 1.2069720959989354, 'learning_rate/qwen_vl_interface': 8.355384693278516e-06, 'learning_rate/action_model': 8.355384693278517e-05, 'epoch': 0.58}) +[2026-06-17 09:46:07,168][starVLA.training.train_starvla][INFO] - Step 1986, Loss: {'train/loss': 0.6892091035842896, 'train/grad_norm_pre_clip': 9.440642356872559, 'timing/data': 0.0003569909604266286, 'timing/model': 1.203890208969824, 'learning_rate/qwen_vl_interface': 8.353747868953002e-06, 'learning_rate/action_model': 8.353747868953003e-05, 'epoch': 0.58}) +[2026-06-17 09:46:11,974][starVLA.training.train_starvla][INFO] - Step 1987, Loss: {'train/loss': 0.6291892528533936, 'train/grad_norm_pre_clip': 8.27297306060791, 'timing/data': 0.0004172429908066988, 'timing/model': 1.2032694269437343, 'learning_rate/qwen_vl_interface': 8.352110401217213e-06, 'learning_rate/action_model': 8.352110401217213e-05, 'epoch': 0.58}) +[2026-06-17 09:46:16,765][starVLA.training.train_starvla][INFO] - Step 1988, Loss: {'train/loss': 0.9266618490219116, 'train/grad_norm_pre_clip': 8.476763725280762, 'timing/data': 0.00043564208317548037, 'timing/model': 1.2052516200346872, 'learning_rate/qwen_vl_interface': 8.350472290410602e-06, 'learning_rate/action_model': 8.350472290410602e-05, 'epoch': 0.58}) +[2026-06-17 09:46:21,548][starVLA.training.train_starvla][INFO] - Step 1989, Loss: {'train/loss': 0.6895356774330139, 'train/grad_norm_pre_clip': 6.605350017547607, 'timing/data': 0.00040276499930769205, 'timing/model': 1.3360657970188186, 'learning_rate/qwen_vl_interface': 8.348833536872747e-06, 'learning_rate/action_model': 8.348833536872747e-05, 'epoch': 0.59}) +[2026-06-17 09:46:26,208][starVLA.training.train_starvla][INFO] - Step 1990, Loss: {'train/loss': 0.7169021964073181, 'train/grad_norm_pre_clip': 7.75750207901001, 'timing/data': 0.00036427099257707596, 'timing/model': 1.2078088220441714, 'learning_rate/qwen_vl_interface': 8.347194140943366e-06, 'learning_rate/action_model': 8.347194140943366e-05, 'epoch': 0.59}) +[2026-06-17 09:46:31,026][starVLA.training.train_starvla][INFO] - Step 1991, Loss: {'train/loss': 0.7194666266441345, 'train/grad_norm_pre_clip': 9.697949409484863, 'timing/data': 0.00035582203418016434, 'timing/model': 1.2120161760831252, 'learning_rate/qwen_vl_interface': 8.345554102962306e-06, 'learning_rate/action_model': 8.345554102962305e-05, 'epoch': 0.59}) +[2026-06-17 09:46:35,723][starVLA.training.train_starvla][INFO] - Step 1992, Loss: {'train/loss': 0.6414294242858887, 'train/grad_norm_pre_clip': 7.001974582672119, 'timing/data': 0.00042537099216133356, 'timing/model': 1.2091260770102963, 'learning_rate/qwen_vl_interface': 8.34391342326955e-06, 'learning_rate/action_model': 8.343913423269549e-05, 'epoch': 0.59}) +[2026-06-17 09:46:40,381][starVLA.training.train_starvla][INFO] - Step 1993, Loss: {'train/loss': 0.7436210513114929, 'train/grad_norm_pre_clip': 7.994506359100342, 'timing/data': 0.00041193398647010326, 'timing/model': 1.201260442030616, 'learning_rate/qwen_vl_interface': 8.342272102205213e-06, 'learning_rate/action_model': 8.342272102205212e-05, 'epoch': 0.59}) +[2026-06-17 09:46:45,086][starVLA.training.train_starvla][INFO] - Step 1994, Loss: {'train/loss': 0.9424470067024231, 'train/grad_norm_pre_clip': 9.82848834991455, 'timing/data': 0.000351092079654336, 'timing/model': 1.203369744005613, 'learning_rate/qwen_vl_interface': 8.340630140109539e-06, 'learning_rate/action_model': 8.340630140109539e-05, 'epoch': 0.59}) +[2026-06-17 09:46:49,848][starVLA.training.train_starvla][INFO] - Step 1995, Loss: {'train/loss': 0.8150194883346558, 'train/grad_norm_pre_clip': 8.298713684082031, 'timing/data': 0.00035246892366558313, 'timing/model': 1.281869397033006, 'learning_rate/qwen_vl_interface': 8.338987537322916e-06, 'learning_rate/action_model': 8.338987537322916e-05, 'epoch': 0.59}) +[2026-06-17 09:46:54,551][starVLA.training.train_starvla][INFO] - Step 1996, Loss: {'train/loss': 0.9479153752326965, 'train/grad_norm_pre_clip': 9.46689224243164, 'timing/data': 0.00035762100014835596, 'timing/model': 1.22813446004875, 'learning_rate/qwen_vl_interface': 8.337344294185848e-06, 'learning_rate/action_model': 8.337344294185849e-05, 'epoch': 0.59}) +[2026-06-17 09:46:59,343][starVLA.training.train_starvla][INFO] - Step 1997, Loss: {'train/loss': 0.6171426773071289, 'train/grad_norm_pre_clip': 8.095756530761719, 'timing/data': 0.0004026070237159729, 'timing/model': 1.2058525360189378, 'learning_rate/qwen_vl_interface': 8.335700411038988e-06, 'learning_rate/action_model': 8.335700411038988e-05, 'epoch': 0.59}) +[2026-06-17 09:47:04,140][starVLA.training.train_starvla][INFO] - Step 1998, Loss: {'train/loss': 0.8157868385314941, 'train/grad_norm_pre_clip': 7.313575267791748, 'timing/data': 0.00034738401882350445, 'timing/model': 1.2101643930654973, 'learning_rate/qwen_vl_interface': 8.334055888223111e-06, 'learning_rate/action_model': 8.334055888223111e-05, 'epoch': 0.59}) +[2026-06-17 09:47:08,993][starVLA.training.train_starvla][INFO] - Step 1999, Loss: {'train/loss': 0.818706750869751, 'train/grad_norm_pre_clip': 13.34058952331543, 'timing/data': 0.0003685730043798685, 'timing/model': 1.2059089570539072, 'learning_rate/qwen_vl_interface': 8.332410726079132e-06, 'learning_rate/action_model': 8.332410726079132e-05, 'epoch': 0.59}) +[2026-06-17 09:57:17,462][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 2000 +[2026-06-17 09:57:17,463][starVLA.training.train_starvla][INFO] - Step 2000, Loss: {'train/loss': 0.7307977676391602, 'train/grad_norm_pre_clip': 8.626206398010254, 'eval/loss': 0.8650821447372437, 'eval/latency_0/loss': 0.8650821447372437, 'eval/demon_fire_precision': 0.9041737649063032, 'eval/demon_fire_recall': 0.8816445182724253, 'eval/demon_fire_f1': 0.8927670311185871, 'eval/demon_fire_rate_model': 0.6529477196885428, 'eval/demon_fire_rate_teacher': 0.6696329254727474, 'eval/demon_left_precision': 0.8452803406671399, 'eval/demon_left_recall': 0.6747875354107649, 'eval/demon_left_f1': 0.7504725897920606, 'eval/demon_left_rate_model': 0.3918242491657397, 'eval/demon_left_rate_teacher': 0.49082313681868744, 'eval/demon_right_precision': 0.7248168498168498, 'eval/demon_right_recall': 0.8943502824858757, 'eval/demon_right_f1': 0.8007081436519979, 'eval/demon_right_rate_model': 0.60734149054505, 'eval/demon_right_rate_teacher': 0.4922135706340378, 'eval/demon_move_f1': 0.7755903667220292, 'eval/demon_cc_f1': 0.8540575699605402, 'eval/latency_0/demon_fire_precision': 0.9041737649063032, 'eval/latency_0/demon_fire_recall': 0.8816445182724253, 'eval/latency_0/demon_fire_f1': 0.8927670311185871, 'eval/latency_0/demon_fire_rate_model': 0.6529477196885428, 'eval/latency_0/demon_fire_rate_teacher': 0.6696329254727474, 'eval/latency_0/demon_left_precision': 0.8452803406671399, 'eval/latency_0/demon_left_recall': 0.6747875354107649, 'eval/latency_0/demon_left_f1': 0.7504725897920606, 'eval/latency_0/demon_left_rate_model': 0.3918242491657397, 'eval/latency_0/demon_left_rate_teacher': 0.49082313681868744, 'eval/latency_0/demon_right_precision': 0.7248168498168498, 'eval/latency_0/demon_right_recall': 0.8943502824858757, 'eval/latency_0/demon_right_f1': 0.8007081436519979, 'eval/latency_0/demon_right_rate_model': 0.60734149054505, 'eval/latency_0/demon_right_rate_teacher': 0.4922135706340378, 'eval/latency_0/demon_move_f1': 0.7755903667220292, 'eval/latency_0/demon_cc_f1': 0.8540575699605402, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 3078.0, 'rl_games_eval/mid_train/mean_length': 3600.0, 'rl_games_eval/mid_train/std_reward': 2452.6426564014578, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/macro_mean_reward': 3078.0, 'rl_games_eval/mid_train/macro_mean_length': 3600.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/demon_attack__latency_0/mean_reward': 3078.0, 'rl_games_eval/mid_train/demon_attack__latency_0/mean_length': 3600.0, 'rl_games_eval/mid_train/demon_attack__latency_0/std_reward': 2452.6426564014578, 'rl_games_eval/mid_train/demon_attack__latency_0/std_length': 0.0, 'rl_games_eval/mid_train/demon_attack__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 3078.0, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.000540334964171052, 'timing/model': 1.3378858540672809, 'learning_rate/qwen_vl_interface': 8.330764924948093e-06, 'learning_rate/action_model': 8.330764924948093e-05, 'epoch': 0.59}) +[2026-06-17 09:57:17,463][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_2000_state +[2026-06-17 09:57:17,465][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 09:57:21,000][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_2000_state/model.safetensors +[2026-06-17 09:57:28,321][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_2000_state/optimizer.bin +[2026-06-17 09:57:28,321][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_2000_state/sampler.bin +[2026-06-17 09:57:28,321][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_2000_state/sampler_1.bin +[2026-06-17 09:57:28,322][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_2000_state/random_states_0.pkl +[2026-06-17 09:57:28,367][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2000 +[2026-06-17 09:57:28,367][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 09:57:28,393][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 09:57:33,140][starVLA.training.train_starvla][INFO] - Step 2001, Loss: {'train/loss': 0.8241666555404663, 'train/grad_norm_pre_clip': 8.618900299072266, 'timing/data': 0.0004095210460945964, 'timing/model': 1.32164537103381, 'learning_rate/qwen_vl_interface': 8.329118485171167e-06, 'learning_rate/action_model': 8.329118485171167e-05, 'epoch': 0.59}) +[2026-06-17 09:57:37,807][starVLA.training.train_starvla][INFO] - Step 2002, Loss: {'train/loss': 0.894390344619751, 'train/grad_norm_pre_clip': 7.837673664093018, 'timing/data': 0.00037797598633915186, 'timing/model': 1.1886490639299154, 'learning_rate/qwen_vl_interface': 8.327471407089667e-06, 'learning_rate/action_model': 8.327471407089668e-05, 'epoch': 0.59}) +[2026-06-17 09:57:42,528][starVLA.training.train_starvla][INFO] - Step 2003, Loss: {'train/loss': 0.7433844208717346, 'train/grad_norm_pre_clip': 9.040830612182617, 'timing/data': 0.00034826307091861963, 'timing/model': 1.1913880540523678, 'learning_rate/qwen_vl_interface': 8.325823691045032e-06, 'learning_rate/action_model': 8.325823691045033e-05, 'epoch': 0.59}) +[2026-06-17 09:57:47,281][starVLA.training.train_starvla][INFO] - Step 2004, Loss: {'train/loss': 0.9115620255470276, 'train/grad_norm_pre_clip': 11.218026161193848, 'timing/data': 0.00036665902007371187, 'timing/model': 1.1952875059796497, 'learning_rate/qwen_vl_interface': 8.32417533737884e-06, 'learning_rate/action_model': 8.324175337378839e-05, 'epoch': 0.59}) +[2026-06-17 09:57:52,030][starVLA.training.train_starvla][INFO] - Step 2005, Loss: {'train/loss': 0.8631922006607056, 'train/grad_norm_pre_clip': 8.377124786376953, 'timing/data': 0.0003482530592009425, 'timing/model': 1.1950687469216064, 'learning_rate/qwen_vl_interface': 8.322526346432789e-06, 'learning_rate/action_model': 8.322526346432789e-05, 'epoch': 0.59}) +[2026-06-17 09:57:56,775][starVLA.training.train_starvla][INFO] - Step 2006, Loss: {'train/loss': 1.072462797164917, 'train/grad_norm_pre_clip': 9.24420166015625, 'timing/data': 0.0003852330846711993, 'timing/model': 1.3243219029391184, 'learning_rate/qwen_vl_interface': 8.32087671854872e-06, 'learning_rate/action_model': 8.32087671854872e-05, 'epoch': 0.59}) +[2026-06-17 09:58:01,391][starVLA.training.train_starvla][INFO] - Step 2007, Loss: {'train/loss': 0.553625762462616, 'train/grad_norm_pre_clip': 6.4018235206604, 'timing/data': 0.0004208579193800688, 'timing/model': 1.1952058549504727, 'learning_rate/qwen_vl_interface': 8.319226454068604e-06, 'learning_rate/action_model': 8.319226454068604e-05, 'epoch': 0.59}) +[2026-06-17 09:58:06,139][starVLA.training.train_starvla][INFO] - Step 2008, Loss: {'train/loss': 1.0240203142166138, 'train/grad_norm_pre_clip': 10.610831260681152, 'timing/data': 0.00038416997995227575, 'timing/model': 1.1982930810190737, 'learning_rate/qwen_vl_interface': 8.317575553334542e-06, 'learning_rate/action_model': 8.317575553334542e-05, 'epoch': 0.59}) +[2026-06-17 09:58:10,870][starVLA.training.train_starvla][INFO] - Step 2009, Loss: {'train/loss': 1.0672049522399902, 'train/grad_norm_pre_clip': 10.768159866333008, 'timing/data': 0.0003736549988389015, 'timing/model': 1.1962578700622544, 'learning_rate/qwen_vl_interface': 8.315924016688767e-06, 'learning_rate/action_model': 8.315924016688766e-05, 'epoch': 0.59}) +[2026-06-17 09:58:15,625][starVLA.training.train_starvla][INFO] - Step 2010, Loss: {'train/loss': 0.734550416469574, 'train/grad_norm_pre_clip': 8.395222663879395, 'timing/data': 0.0003530830144882202, 'timing/model': 1.1971609400352463, 'learning_rate/qwen_vl_interface': 8.314271844473644e-06, 'learning_rate/action_model': 8.314271844473644e-05, 'epoch': 0.59}) +[2026-06-17 09:58:20,378][starVLA.training.train_starvla][INFO] - Step 2011, Loss: {'train/loss': 1.0625616312026978, 'train/grad_norm_pre_clip': 8.550683975219727, 'timing/data': 0.00035253597889095545, 'timing/model': 1.1985633849399164, 'learning_rate/qwen_vl_interface': 8.312619037031668e-06, 'learning_rate/action_model': 8.312619037031668e-05, 'epoch': 0.59}) +[2026-06-17 09:58:25,153][starVLA.training.train_starvla][INFO] - Step 2012, Loss: {'train/loss': 0.7629637718200684, 'train/grad_norm_pre_clip': 9.028075218200684, 'timing/data': 0.00035792996641248465, 'timing/model': 1.3401264540152624, 'learning_rate/qwen_vl_interface': 8.310965594705473e-06, 'learning_rate/action_model': 8.310965594705472e-05, 'epoch': 0.59}) +[2026-06-17 09:58:29,798][starVLA.training.train_starvla][INFO] - Step 2013, Loss: {'train/loss': 0.7119665741920471, 'train/grad_norm_pre_clip': 8.160521507263184, 'timing/data': 0.0003327239537611604, 'timing/model': 1.2001588119892403, 'learning_rate/qwen_vl_interface': 8.309311517837814e-06, 'learning_rate/action_model': 8.309311517837813e-05, 'epoch': 0.59}) +[2026-06-17 09:58:34,587][starVLA.training.train_starvla][INFO] - Step 2014, Loss: {'train/loss': 0.8222454786300659, 'train/grad_norm_pre_clip': 9.383002281188965, 'timing/data': 0.0005878430092707276, 'timing/model': 1.2007670809980482, 'learning_rate/qwen_vl_interface': 8.307656806771587e-06, 'learning_rate/action_model': 8.307656806771588e-05, 'epoch': 0.59}) +[2026-06-17 09:58:39,357][starVLA.training.train_starvla][INFO] - Step 2015, Loss: {'train/loss': 0.7783476710319519, 'train/grad_norm_pre_clip': 11.616342544555664, 'timing/data': 0.0003706669667735696, 'timing/model': 1.2007736850064248, 'learning_rate/qwen_vl_interface': 8.306001461849814e-06, 'learning_rate/action_model': 8.306001461849814e-05, 'epoch': 0.59}) +[2026-06-17 09:58:44,171][starVLA.training.train_starvla][INFO] - Step 2016, Loss: {'train/loss': 0.6751022338867188, 'train/grad_norm_pre_clip': 6.562548637390137, 'timing/data': 0.000377992051653564, 'timing/model': 1.2015247220406309, 'learning_rate/qwen_vl_interface': 8.30434548341565e-06, 'learning_rate/action_model': 8.304345483415649e-05, 'epoch': 0.59}) +[2026-06-17 09:58:48,953][starVLA.training.train_starvla][INFO] - Step 2017, Loss: {'train/loss': 0.7607367634773254, 'train/grad_norm_pre_clip': 7.996793746948242, 'timing/data': 0.0003996489103883505, 'timing/model': 1.202170431963168, 'learning_rate/qwen_vl_interface': 8.30268887181238e-06, 'learning_rate/action_model': 8.302688871812378e-05, 'epoch': 0.59}) +[2026-06-17 09:58:53,747][starVLA.training.train_starvla][INFO] - Step 2018, Loss: {'train/loss': 0.9413581490516663, 'train/grad_norm_pre_clip': 8.337053298950195, 'timing/data': 0.0006250819424167275, 'timing/model': 1.3421740779886022, 'learning_rate/qwen_vl_interface': 8.301031627383422e-06, 'learning_rate/action_model': 8.301031627383421e-05, 'epoch': 0.59}) +[2026-06-17 09:58:58,432][starVLA.training.train_starvla][INFO] - Step 2019, Loss: {'train/loss': 0.969106137752533, 'train/grad_norm_pre_clip': 8.71243953704834, 'timing/data': 0.0006629109848290682, 'timing/model': 1.2081333099631593, 'learning_rate/qwen_vl_interface': 8.299373750472323e-06, 'learning_rate/action_model': 8.299373750472324e-05, 'epoch': 0.59}) +[2026-06-17 09:59:03,225][starVLA.training.train_starvla][INFO] - Step 2020, Loss: {'train/loss': 0.6861099600791931, 'train/grad_norm_pre_clip': 6.911151885986328, 'timing/data': 0.0003744550049304962, 'timing/model': 1.2017426380189136, 'learning_rate/qwen_vl_interface': 8.297715241422765e-06, 'learning_rate/action_model': 8.297715241422766e-05, 'epoch': 0.59}) +[2026-06-17 09:59:08,002][starVLA.training.train_starvla][INFO] - Step 2021, Loss: {'train/loss': 0.9394280314445496, 'train/grad_norm_pre_clip': 11.2792329788208, 'timing/data': 0.0006873880047351122, 'timing/model': 1.203121867030859, 'learning_rate/qwen_vl_interface': 8.29605610057856e-06, 'learning_rate/action_model': 8.29605610057856e-05, 'epoch': 0.59}) +[2026-06-17 09:59:12,740][starVLA.training.train_starvla][INFO] - Step 2022, Loss: {'train/loss': 0.6034677028656006, 'train/grad_norm_pre_clip': 9.101998329162598, 'timing/data': 0.00035097100771963596, 'timing/model': 1.2080231700092554, 'learning_rate/qwen_vl_interface': 8.294396328283647e-06, 'learning_rate/action_model': 8.294396328283645e-05, 'epoch': 0.59}) +[2026-06-17 09:59:17,535][starVLA.training.train_starvla][INFO] - Step 2023, Loss: {'train/loss': 0.7462306022644043, 'train/grad_norm_pre_clip': 7.137380599975586, 'timing/data': 0.00037363299634307623, 'timing/model': 1.2112046760739759, 'learning_rate/qwen_vl_interface': 8.292735924882097e-06, 'learning_rate/action_model': 8.292735924882097e-05, 'epoch': 0.6}) +[2026-06-17 09:59:22,342][starVLA.training.train_starvla][INFO] - Step 2024, Loss: {'train/loss': 1.041197657585144, 'train/grad_norm_pre_clip': 9.092692375183105, 'timing/data': 0.0004090330330654979, 'timing/model': 1.3472850989783183, 'learning_rate/qwen_vl_interface': 8.291074890718117e-06, 'learning_rate/action_model': 8.291074890718117e-05, 'epoch': 0.6}) +[2026-06-17 09:59:26,993][starVLA.training.train_starvla][INFO] - Step 2025, Loss: {'train/loss': 0.8692563772201538, 'train/grad_norm_pre_clip': 10.525300979614258, 'timing/data': 0.0004023179644718766, 'timing/model': 1.2046157639706507, 'learning_rate/qwen_vl_interface': 8.28941322613604e-06, 'learning_rate/action_model': 8.289413226136039e-05, 'epoch': 0.6}) +[2026-06-17 09:59:31,790][starVLA.training.train_starvla][INFO] - Step 2026, Loss: {'train/loss': 0.9178404211997986, 'train/grad_norm_pre_clip': 11.174295425415039, 'timing/data': 0.0003678640350699425, 'timing/model': 1.2072494289604947, 'learning_rate/qwen_vl_interface': 8.287750931480332e-06, 'learning_rate/action_model': 8.287750931480332e-05, 'epoch': 0.6}) +[2026-06-17 09:59:36,584][starVLA.training.train_starvla][INFO] - Step 2027, Loss: {'train/loss': 0.7953577637672424, 'train/grad_norm_pre_clip': 12.74122142791748, 'timing/data': 0.0003952690167352557, 'timing/model': 1.2034303259570152, 'learning_rate/qwen_vl_interface': 8.286088007095584e-06, 'learning_rate/action_model': 8.286088007095584e-05, 'epoch': 0.6}) +[2026-06-17 09:59:41,373][starVLA.training.train_starvla][INFO] - Step 2028, Loss: {'train/loss': 0.8404489755630493, 'train/grad_norm_pre_clip': 9.71845531463623, 'timing/data': 0.00035632296930998564, 'timing/model': 1.2039976780069992, 'learning_rate/qwen_vl_interface': 8.284424453326528e-06, 'learning_rate/action_model': 8.284424453326528e-05, 'epoch': 0.6}) +[2026-06-17 09:59:46,157][starVLA.training.train_starvla][INFO] - Step 2029, Loss: {'train/loss': 0.8579344749450684, 'train/grad_norm_pre_clip': 10.28789234161377, 'timing/data': 0.00036149495281279087, 'timing/model': 1.2049474839586765, 'learning_rate/qwen_vl_interface': 8.282760270518017e-06, 'learning_rate/action_model': 8.282760270518017e-05, 'epoch': 0.6}) +[2026-06-17 09:59:50,863][starVLA.training.train_starvla][INFO] - Step 2030, Loss: {'train/loss': 0.9168851971626282, 'train/grad_norm_pre_clip': 8.723408699035645, 'timing/data': 0.0003784459549933672, 'timing/model': 1.2572139030089602, 'learning_rate/qwen_vl_interface': 8.281095459015037e-06, 'learning_rate/action_model': 8.281095459015036e-05, 'epoch': 0.6}) +[2026-06-17 09:59:55,564][starVLA.training.train_starvla][INFO] - Step 2031, Loss: {'train/loss': 0.8736118078231812, 'train/grad_norm_pre_clip': 12.429919242858887, 'timing/data': 0.0003902319585904479, 'timing/model': 1.2040212029824033, 'learning_rate/qwen_vl_interface': 8.279430019162709e-06, 'learning_rate/action_model': 8.279430019162708e-05, 'epoch': 0.6}) +[2026-06-17 10:00:00,310][starVLA.training.train_starvla][INFO] - Step 2032, Loss: {'train/loss': 0.8421192169189453, 'train/grad_norm_pre_clip': 8.13962173461914, 'timing/data': 0.0006272540194913745, 'timing/model': 1.2050758400000632, 'learning_rate/qwen_vl_interface': 8.277763951306277e-06, 'learning_rate/action_model': 8.277763951306277e-05, 'epoch': 0.6}) +[2026-06-17 10:00:05,114][starVLA.training.train_starvla][INFO] - Step 2033, Loss: {'train/loss': 0.8008452653884888, 'train/grad_norm_pre_clip': 9.701838493347168, 'timing/data': 0.000685616978444159, 'timing/model': 1.2087207660079002, 'learning_rate/qwen_vl_interface': 8.27609725579112e-06, 'learning_rate/action_model': 8.27609725579112e-05, 'epoch': 0.6}) +[2026-06-17 10:00:09,915][starVLA.training.train_starvla][INFO] - Step 2034, Loss: {'train/loss': 0.7506653666496277, 'train/grad_norm_pre_clip': 7.775769233703613, 'timing/data': 0.00043091189581900835, 'timing/model': 1.2044697109377012, 'learning_rate/qwen_vl_interface': 8.274429932962745e-06, 'learning_rate/action_model': 8.274429932962746e-05, 'epoch': 0.6}) +[2026-06-17 10:00:14,702][starVLA.training.train_starvla][INFO] - Step 2035, Loss: {'train/loss': 0.8435612320899963, 'train/grad_norm_pre_clip': 8.74626350402832, 'timing/data': 0.00037011306267231703, 'timing/model': 1.3345396450022236, 'learning_rate/qwen_vl_interface': 8.272761983166793e-06, 'learning_rate/action_model': 8.272761983166792e-05, 'epoch': 0.6}) +[2026-06-17 10:00:19,351][starVLA.training.train_starvla][INFO] - Step 2036, Loss: {'train/loss': 0.7857890129089355, 'train/grad_norm_pre_clip': 9.697138786315918, 'timing/data': 0.0005526799941435456, 'timing/model': 1.201238619047217, 'learning_rate/qwen_vl_interface': 8.271093406749025e-06, 'learning_rate/action_model': 8.271093406749025e-05, 'epoch': 0.6}) +[2026-06-17 10:00:24,144][starVLA.training.train_starvla][INFO] - Step 2037, Loss: {'train/loss': 0.5941405296325684, 'train/grad_norm_pre_clip': 8.976435661315918, 'timing/data': 0.0003777090460062027, 'timing/model': 1.2026540930382907, 'learning_rate/qwen_vl_interface': 8.269424204055346e-06, 'learning_rate/action_model': 8.269424204055346e-05, 'epoch': 0.6}) +[2026-06-17 10:00:28,819][starVLA.training.train_starvla][INFO] - Step 2038, Loss: {'train/loss': 0.6855347752571106, 'train/grad_norm_pre_clip': 8.161293983459473, 'timing/data': 0.0003575399750843644, 'timing/model': 1.2073287250241265, 'learning_rate/qwen_vl_interface': 8.26775437543178e-06, 'learning_rate/action_model': 8.26775437543178e-05, 'epoch': 0.6}) +[2026-06-17 10:00:33,495][starVLA.training.train_starvla][INFO] - Step 2039, Loss: {'train/loss': 1.0347458124160767, 'train/grad_norm_pre_clip': 10.915897369384766, 'timing/data': 0.0003532009432092309, 'timing/model': 1.2091476910281926, 'learning_rate/qwen_vl_interface': 8.266083921224484e-06, 'learning_rate/action_model': 8.266083921224485e-05, 'epoch': 0.6}) +[2026-06-17 10:00:38,197][starVLA.training.train_starvla][INFO] - Step 2040, Loss: {'train/loss': 0.7796981930732727, 'train/grad_norm_pre_clip': 11.372546195983887, 'timing/data': 0.0003561079502105713, 'timing/model': 1.2081602569669485, 'learning_rate/qwen_vl_interface': 8.264412841779747e-06, 'learning_rate/action_model': 8.264412841779746e-05, 'epoch': 0.6}) +[2026-06-17 10:00:42,934][starVLA.training.train_starvla][INFO] - Step 2041, Loss: {'train/loss': 1.1676441431045532, 'train/grad_norm_pre_clip': 10.43127727508545, 'timing/data': 0.0003474390832707286, 'timing/model': 1.2836605629418045, 'learning_rate/qwen_vl_interface': 8.262741137443984e-06, 'learning_rate/action_model': 8.262741137443985e-05, 'epoch': 0.6}) +[2026-06-17 10:00:47,587][starVLA.training.train_starvla][INFO] - Step 2042, Loss: {'train/loss': 0.8536794781684875, 'train/grad_norm_pre_clip': 10.165144920349121, 'timing/data': 0.0003590319538488984, 'timing/model': 1.2038082838989794, 'learning_rate/qwen_vl_interface': 8.26106880856374e-06, 'learning_rate/action_model': 8.26106880856374e-05, 'epoch': 0.6}) +[2026-06-17 10:00:52,380][starVLA.training.train_starvla][INFO] - Step 2043, Loss: {'train/loss': 0.9653003811836243, 'train/grad_norm_pre_clip': 11.23718547821045, 'timing/data': 0.0004119439981877804, 'timing/model': 1.2046228490071371, 'learning_rate/qwen_vl_interface': 8.259395855485693e-06, 'learning_rate/action_model': 8.259395855485694e-05, 'epoch': 0.6}) +[2026-06-17 10:00:57,172][starVLA.training.train_starvla][INFO] - Step 2044, Loss: {'train/loss': 0.47558897733688354, 'train/grad_norm_pre_clip': 6.217447757720947, 'timing/data': 0.0003882749006152153, 'timing/model': 1.2058018390089273, 'learning_rate/qwen_vl_interface': 8.257722278556647e-06, 'learning_rate/action_model': 8.257722278556646e-05, 'epoch': 0.6}) +[2026-06-17 10:01:01,953][starVLA.training.train_starvla][INFO] - Step 2045, Loss: {'train/loss': 0.6713402271270752, 'train/grad_norm_pre_clip': 8.60858154296875, 'timing/data': 0.00038828200194984674, 'timing/model': 1.204630627995357, 'learning_rate/qwen_vl_interface': 8.256048078123536e-06, 'learning_rate/action_model': 8.256048078123535e-05, 'epoch': 0.6}) +[2026-06-17 10:01:06,734][starVLA.training.train_starvla][INFO] - Step 2046, Loss: {'train/loss': 1.1138845682144165, 'train/grad_norm_pre_clip': 9.48697566986084, 'timing/data': 0.0006894400576129556, 'timing/model': 1.3314211399992928, 'learning_rate/qwen_vl_interface': 8.254373254533424e-06, 'learning_rate/action_model': 8.254373254533423e-05, 'epoch': 0.6}) +[2026-06-17 10:01:11,405][starVLA.training.train_starvla][INFO] - Step 2047, Loss: {'train/loss': 0.9698687791824341, 'train/grad_norm_pre_clip': 7.978503704071045, 'timing/data': 0.00034859206061810255, 'timing/model': 1.217167590977624, 'learning_rate/qwen_vl_interface': 8.252697808133501e-06, 'learning_rate/action_model': 8.252697808133501e-05, 'epoch': 0.6}) +[2026-06-17 10:01:16,086][starVLA.training.train_starvla][INFO] - Step 2048, Loss: {'train/loss': 1.1824969053268433, 'train/grad_norm_pre_clip': 9.855608940124512, 'timing/data': 0.0003854889655485749, 'timing/model': 1.209857823094353, 'learning_rate/qwen_vl_interface': 8.251021739271095e-06, 'learning_rate/action_model': 8.251021739271094e-05, 'epoch': 0.6}) +[2026-06-17 10:01:20,874][starVLA.training.train_starvla][INFO] - Step 2049, Loss: {'train/loss': 0.8785335421562195, 'train/grad_norm_pre_clip': 8.781769752502441, 'timing/data': 0.0003510230453684926, 'timing/model': 1.205777778988704, 'learning_rate/qwen_vl_interface': 8.249345048293652e-06, 'learning_rate/action_model': 8.24934504829365e-05, 'epoch': 0.6}) +[2026-06-17 10:01:25,681][starVLA.training.train_starvla][INFO] - Step 2050, Loss: {'train/loss': 0.7406068444252014, 'train/grad_norm_pre_clip': 9.200072288513184, 'timing/data': 0.0003535960568115115, 'timing/model': 1.2067428180016577, 'learning_rate/qwen_vl_interface': 8.247667735548753e-06, 'learning_rate/action_model': 8.247667735548752e-05, 'epoch': 0.6}) +[2026-06-17 10:01:30,476][starVLA.training.train_starvla][INFO] - Step 2051, Loss: {'train/loss': 1.00784170627594, 'train/grad_norm_pre_clip': 14.86099910736084, 'timing/data': 0.00035469699651002884, 'timing/model': 1.2044711740454659, 'learning_rate/qwen_vl_interface': 8.245989801384104e-06, 'learning_rate/action_model': 8.245989801384104e-05, 'epoch': 0.6}) +[2026-06-17 10:01:35,262][starVLA.training.train_starvla][INFO] - Step 2052, Loss: {'train/loss': 0.8239848613739014, 'train/grad_norm_pre_clip': 11.61740779876709, 'timing/data': 0.0003684009425342083, 'timing/model': 1.3381461879471317, 'learning_rate/qwen_vl_interface': 8.24431124614755e-06, 'learning_rate/action_model': 8.244311246147549e-05, 'epoch': 0.6}) +[2026-06-17 10:01:39,927][starVLA.training.train_starvla][INFO] - Step 2053, Loss: {'train/loss': 0.7734763622283936, 'train/grad_norm_pre_clip': 8.785314559936523, 'timing/data': 0.0003475330304354429, 'timing/model': 1.2047222330002114, 'learning_rate/qwen_vl_interface': 8.24263207018705e-06, 'learning_rate/action_model': 8.24263207018705e-05, 'epoch': 0.6}) +[2026-06-17 10:01:44,716][starVLA.training.train_starvla][INFO] - Step 2054, Loss: {'train/loss': 0.7613082528114319, 'train/grad_norm_pre_clip': 9.163857460021973, 'timing/data': 0.0003622609656304121, 'timing/model': 1.2049056090181693, 'learning_rate/qwen_vl_interface': 8.2409522738507e-06, 'learning_rate/action_model': 8.240952273850701e-05, 'epoch': 0.6}) +[2026-06-17 10:01:49,614][starVLA.training.train_starvla][INFO] - Step 2055, Loss: {'train/loss': 0.9351800084114075, 'train/grad_norm_pre_clip': 9.002131462097168, 'timing/data': 0.000616310047917068, 'timing/model': 1.2054291750537232, 'learning_rate/qwen_vl_interface': 8.23927185748673e-06, 'learning_rate/action_model': 8.239271857486729e-05, 'epoch': 0.6}) +[2026-06-17 10:01:54,350][starVLA.training.train_starvla][INFO] - Step 2056, Loss: {'train/loss': 0.6462491154670715, 'train/grad_norm_pre_clip': 7.01178503036499, 'timing/data': 0.00038580503314733505, 'timing/model': 1.2028597589815035, 'learning_rate/qwen_vl_interface': 8.237590821443484e-06, 'learning_rate/action_model': 8.237590821443483e-05, 'epoch': 0.6}) +[2026-06-17 10:01:59,141][starVLA.training.train_starvla][INFO] - Step 2057, Loss: {'train/loss': 0.7306526303291321, 'train/grad_norm_pre_clip': 8.39742374420166, 'timing/data': 0.00037506804801523685, 'timing/model': 1.2064877919619903, 'learning_rate/qwen_vl_interface': 8.235909166069447e-06, 'learning_rate/action_model': 8.235909166069447e-05, 'epoch': 0.61}) +[2026-06-17 10:02:03,939][starVLA.training.train_starvla][INFO] - Step 2058, Loss: {'train/loss': 0.9604462385177612, 'train/grad_norm_pre_clip': 11.4548921585083, 'timing/data': 0.0003511259565129876, 'timing/model': 1.3394937310367823, 'learning_rate/qwen_vl_interface': 8.234226891713225e-06, 'learning_rate/action_model': 8.234226891713225e-05, 'epoch': 0.61}) +[2026-06-17 10:02:08,597][starVLA.training.train_starvla][INFO] - Step 2059, Loss: {'train/loss': 0.7432758212089539, 'train/grad_norm_pre_clip': 7.933923721313477, 'timing/data': 0.00035989098250865936, 'timing/model': 1.20472434500698, 'learning_rate/qwen_vl_interface': 8.232543998723559e-06, 'learning_rate/action_model': 8.232543998723559e-05, 'epoch': 0.61}) +[2026-06-17 10:02:13,384][starVLA.training.train_starvla][INFO] - Step 2060, Loss: {'train/loss': 0.8558110594749451, 'train/grad_norm_pre_clip': 11.046880722045898, 'timing/data': 0.0003535889554768801, 'timing/model': 1.2027488179737702, 'learning_rate/qwen_vl_interface': 8.230860487449314e-06, 'learning_rate/action_model': 8.230860487449313e-05, 'epoch': 0.61}) +[2026-06-17 10:02:18,164][starVLA.training.train_starvla][INFO] - Step 2061, Loss: {'train/loss': 1.007568597793579, 'train/grad_norm_pre_clip': 10.06280517578125, 'timing/data': 0.00036426400765776634, 'timing/model': 1.2033416799968109, 'learning_rate/qwen_vl_interface': 8.22917635823948e-06, 'learning_rate/action_model': 8.229176358239478e-05, 'epoch': 0.61}) +[2026-06-17 10:02:22,940][starVLA.training.train_starvla][INFO] - Step 2062, Loss: {'train/loss': 0.7555230855941772, 'train/grad_norm_pre_clip': 7.6226043701171875, 'timing/data': 0.0003537109587341547, 'timing/model': 1.205086195957847, 'learning_rate/qwen_vl_interface': 8.227491611443182e-06, 'learning_rate/action_model': 8.227491611443182e-05, 'epoch': 0.61}) +[2026-06-17 10:02:27,716][starVLA.training.train_starvla][INFO] - Step 2063, Loss: {'train/loss': 0.7252599000930786, 'train/grad_norm_pre_clip': 7.917375564575195, 'timing/data': 0.00036257703322917223, 'timing/model': 1.2039150360506028, 'learning_rate/qwen_vl_interface': 8.225806247409667e-06, 'learning_rate/action_model': 8.225806247409668e-05, 'epoch': 0.61}) +[2026-06-17 10:02:32,419][starVLA.training.train_starvla][INFO] - Step 2064, Loss: {'train/loss': 1.120954990386963, 'train/grad_norm_pre_clip': 13.249510765075684, 'timing/data': 0.0003941520117223263, 'timing/model': 1.249780718004331, 'learning_rate/qwen_vl_interface': 8.224120266488318e-06, 'learning_rate/action_model': 8.224120266488318e-05, 'epoch': 0.61}) +[2026-06-17 10:02:37,073][starVLA.training.train_starvla][INFO] - Step 2065, Loss: {'train/loss': 0.6548368334770203, 'train/grad_norm_pre_clip': 7.1243486404418945, 'timing/data': 0.00034380401484668255, 'timing/model': 1.20617937494535, 'learning_rate/qwen_vl_interface': 8.222433669028636e-06, 'learning_rate/action_model': 8.222433669028635e-05, 'epoch': 0.61}) +[2026-06-17 10:02:41,817][starVLA.training.train_starvla][INFO] - Step 2066, Loss: {'train/loss': 1.2021464109420776, 'train/grad_norm_pre_clip': 9.898159980773926, 'timing/data': 0.00034708704333752394, 'timing/model': 1.2023933960590512, 'learning_rate/qwen_vl_interface': 8.220746455380257e-06, 'learning_rate/action_model': 8.220746455380257e-05, 'epoch': 0.61}) +[2026-06-17 10:02:46,669][starVLA.training.train_starvla][INFO] - Step 2067, Loss: {'train/loss': 0.8924668431282043, 'train/grad_norm_pre_clip': 13.842519760131836, 'timing/data': 0.00038492202293127775, 'timing/model': 1.2045069909654558, 'learning_rate/qwen_vl_interface': 8.219058625892939e-06, 'learning_rate/action_model': 8.21905862589294e-05, 'epoch': 0.61}) +[2026-06-17 10:02:51,478][starVLA.training.train_starvla][INFO] - Step 2068, Loss: {'train/loss': 0.7906286120414734, 'train/grad_norm_pre_clip': 7.572958946228027, 'timing/data': 0.0005058670649304986, 'timing/model': 1.2108347261091694, 'learning_rate/qwen_vl_interface': 8.217370180916574e-06, 'learning_rate/action_model': 8.217370180916575e-05, 'epoch': 0.61}) +[2026-06-17 10:02:56,270][starVLA.training.train_starvla][INFO] - Step 2069, Loss: {'train/loss': 0.6141802668571472, 'train/grad_norm_pre_clip': 6.65616512298584, 'timing/data': 0.00034123705700039864, 'timing/model': 1.3370510169770569, 'learning_rate/qwen_vl_interface': 8.215681120801179e-06, 'learning_rate/action_model': 8.215681120801179e-05, 'epoch': 0.61}) +[2026-06-17 10:03:00,987][starVLA.training.train_starvla][INFO] - Step 2070, Loss: {'train/loss': 0.9901083707809448, 'train/grad_norm_pre_clip': 7.237191200256348, 'timing/data': 0.0003524579806253314, 'timing/model': 1.203803269076161, 'learning_rate/qwen_vl_interface': 8.213991445896895e-06, 'learning_rate/action_model': 8.213991445896895e-05, 'epoch': 0.61}) +[2026-06-17 10:03:05,869][starVLA.training.train_starvla][INFO] - Step 2071, Loss: {'train/loss': 0.9145001769065857, 'train/grad_norm_pre_clip': 7.24901819229126, 'timing/data': 0.00034992198925465345, 'timing/model': 1.2059182070661336, 'learning_rate/qwen_vl_interface': 8.212301156553995e-06, 'learning_rate/action_model': 8.212301156553995e-05, 'epoch': 0.61}) +[2026-06-17 10:03:10,534][starVLA.training.train_starvla][INFO] - Step 2072, Loss: {'train/loss': 0.8378930687904358, 'train/grad_norm_pre_clip': 8.323491096496582, 'timing/data': 0.00034697900991886854, 'timing/model': 1.2045197800034657, 'learning_rate/qwen_vl_interface': 8.210610253122879e-06, 'learning_rate/action_model': 8.210610253122879e-05, 'epoch': 0.61}) +[2026-06-17 10:03:15,203][starVLA.training.train_starvla][INFO] - Step 2073, Loss: {'train/loss': 0.7359856367111206, 'train/grad_norm_pre_clip': 7.6609206199646, 'timing/data': 0.000373880029655993, 'timing/model': 1.2069171919720247, 'learning_rate/qwen_vl_interface': 8.20891873595407e-06, 'learning_rate/action_model': 8.20891873595407e-05, 'epoch': 0.61}) +[2026-06-17 10:03:19,936][starVLA.training.train_starvla][INFO] - Step 2074, Loss: {'train/loss': 0.9081268310546875, 'train/grad_norm_pre_clip': 10.465076446533203, 'timing/data': 0.000365586020052433, 'timing/model': 1.2120921979658306, 'learning_rate/qwen_vl_interface': 8.207226605398225e-06, 'learning_rate/action_model': 8.207226605398224e-05, 'epoch': 0.61}) +[2026-06-17 10:03:24,693][starVLA.training.train_starvla][INFO] - Step 2075, Loss: {'train/loss': 0.8617973327636719, 'train/grad_norm_pre_clip': 10.094444274902344, 'timing/data': 0.0003539980389177799, 'timing/model': 1.2943436269415542, 'learning_rate/qwen_vl_interface': 8.20553386180612e-06, 'learning_rate/action_model': 8.205533861806119e-05, 'epoch': 0.61}) +[2026-06-17 10:03:29,356][starVLA.training.train_starvla][INFO] - Step 2076, Loss: {'train/loss': 1.221224069595337, 'train/grad_norm_pre_clip': 9.084417343139648, 'timing/data': 0.0003662148956209421, 'timing/model': 1.2094150129705667, 'learning_rate/qwen_vl_interface': 8.203840505528665e-06, 'learning_rate/action_model': 8.203840505528665e-05, 'epoch': 0.61}) +[2026-06-17 10:03:34,175][starVLA.training.train_starvla][INFO] - Step 2077, Loss: {'train/loss': 0.7653154134750366, 'train/grad_norm_pre_clip': 12.696712493896484, 'timing/data': 0.00042605295311659575, 'timing/model': 1.2142548970878124, 'learning_rate/qwen_vl_interface': 8.202146536916899e-06, 'learning_rate/action_model': 8.202146536916899e-05, 'epoch': 0.61}) +[2026-06-17 10:03:39,097][starVLA.training.train_starvla][INFO] - Step 2078, Loss: {'train/loss': 1.141477108001709, 'train/grad_norm_pre_clip': 9.434761047363281, 'timing/data': 0.00036759499926120043, 'timing/model': 1.2117268659640104, 'learning_rate/qwen_vl_interface': 8.200451956321974e-06, 'learning_rate/action_model': 8.200451956321974e-05, 'epoch': 0.61}) +[2026-06-17 10:03:43,910][starVLA.training.train_starvla][INFO] - Step 2079, Loss: {'train/loss': 0.6783809065818787, 'train/grad_norm_pre_clip': 7.942905426025391, 'timing/data': 0.0004172079497948289, 'timing/model': 1.2115786779904738, 'learning_rate/qwen_vl_interface': 8.198756764095187e-06, 'learning_rate/action_model': 8.198756764095187e-05, 'epoch': 0.61}) +[2026-06-17 10:03:48,736][starVLA.training.train_starvla][INFO] - Step 2080, Loss: {'train/loss': 0.7175562381744385, 'train/grad_norm_pre_clip': 8.172647476196289, 'timing/data': 0.00035546591971069574, 'timing/model': 1.349484171019867, 'learning_rate/qwen_vl_interface': 8.19706096058795e-06, 'learning_rate/action_model': 8.197060960587949e-05, 'epoch': 0.61}) +[2026-06-17 10:03:53,441][starVLA.training.train_starvla][INFO] - Step 2081, Loss: {'train/loss': 0.9100324511528015, 'train/grad_norm_pre_clip': 9.187933921813965, 'timing/data': 0.0003489709924906492, 'timing/model': 1.235945847001858, 'learning_rate/qwen_vl_interface': 8.195364546151802e-06, 'learning_rate/action_model': 8.195364546151801e-05, 'epoch': 0.61}) +[2026-06-17 10:03:58,121][starVLA.training.train_starvla][INFO] - Step 2082, Loss: {'train/loss': 1.0841691493988037, 'train/grad_norm_pre_clip': 11.654669761657715, 'timing/data': 0.000337225035764277, 'timing/model': 1.2114913079421967, 'learning_rate/qwen_vl_interface': 8.193667521138416e-06, 'learning_rate/action_model': 8.193667521138416e-05, 'epoch': 0.61}) +[2026-06-17 10:04:02,918][starVLA.training.train_starvla][INFO] - Step 2083, Loss: {'train/loss': 0.6977338194847107, 'train/grad_norm_pre_clip': 9.040616989135742, 'timing/data': 0.0003497130237519741, 'timing/model': 1.2139192169997841, 'learning_rate/qwen_vl_interface': 8.191969885899585e-06, 'learning_rate/action_model': 8.191969885899585e-05, 'epoch': 0.61}) +[2026-06-17 10:04:07,722][starVLA.training.train_starvla][INFO] - Step 2084, Loss: {'train/loss': 1.0303075313568115, 'train/grad_norm_pre_clip': 9.024949073791504, 'timing/data': 0.0005985349416732788, 'timing/model': 1.2049484410090372, 'learning_rate/qwen_vl_interface': 8.190271640787232e-06, 'learning_rate/action_model': 8.190271640787232e-05, 'epoch': 0.61}) +[2026-06-17 10:04:12,532][starVLA.training.train_starvla][INFO] - Step 2085, Loss: {'train/loss': 0.7786258459091187, 'train/grad_norm_pre_clip': 7.646689414978027, 'timing/data': 0.0004051401047036052, 'timing/model': 1.2132199179613963, 'learning_rate/qwen_vl_interface': 8.188572786153403e-06, 'learning_rate/action_model': 8.188572786153402e-05, 'epoch': 0.61}) +[2026-06-17 10:04:17,340][starVLA.training.train_starvla][INFO] - Step 2086, Loss: {'train/loss': 0.6765264272689819, 'train/grad_norm_pre_clip': 18.977746963500977, 'timing/data': 0.00038470898289233446, 'timing/model': 1.3321827440522611, 'learning_rate/qwen_vl_interface': 8.186873322350274e-06, 'learning_rate/action_model': 8.186873322350273e-05, 'epoch': 0.61}) +[2026-06-17 10:04:22,056][starVLA.training.train_starvla][INFO] - Step 2087, Loss: {'train/loss': 1.0556309223175049, 'train/grad_norm_pre_clip': 9.087332725524902, 'timing/data': 0.0003325290745124221, 'timing/model': 1.203846491058357, 'learning_rate/qwen_vl_interface': 8.185173249730143e-06, 'learning_rate/action_model': 8.185173249730142e-05, 'epoch': 0.61}) +[2026-06-17 10:04:26,881][starVLA.training.train_starvla][INFO] - Step 2088, Loss: {'train/loss': 0.8061186075210571, 'train/grad_norm_pre_clip': 12.084477424621582, 'timing/data': 0.00036225689109414816, 'timing/model': 1.2062792669748887, 'learning_rate/qwen_vl_interface': 8.183472568645441e-06, 'learning_rate/action_model': 8.18347256864544e-05, 'epoch': 0.61}) +[2026-06-17 10:04:31,665][starVLA.training.train_starvla][INFO] - Step 2089, Loss: {'train/loss': 0.5085617899894714, 'train/grad_norm_pre_clip': 6.709988117218018, 'timing/data': 0.0003848309861496091, 'timing/model': 1.2035225469153374, 'learning_rate/qwen_vl_interface': 8.181771279448718e-06, 'learning_rate/action_model': 8.181771279448718e-05, 'epoch': 0.61}) +[2026-06-17 10:04:36,327][starVLA.training.train_starvla][INFO] - Step 2090, Loss: {'train/loss': 0.8596416711807251, 'train/grad_norm_pre_clip': 7.549249649047852, 'timing/data': 0.00037758902180939913, 'timing/model': 1.2034577790182084, 'learning_rate/qwen_vl_interface': 8.180069382492655e-06, 'learning_rate/action_model': 8.180069382492654e-05, 'epoch': 0.61}) +[2026-06-17 10:04:41,030][starVLA.training.train_starvla][INFO] - Step 2091, Loss: {'train/loss': 0.930030882358551, 'train/grad_norm_pre_clip': 8.490028381347656, 'timing/data': 0.0005875109927728772, 'timing/model': 1.210353055037558, 'learning_rate/qwen_vl_interface': 8.178366878130056e-06, 'learning_rate/action_model': 8.178366878130056e-05, 'epoch': 0.62}) +[2026-06-17 10:04:45,732][starVLA.training.train_starvla][INFO] - Step 2092, Loss: {'train/loss': 0.661747395992279, 'train/grad_norm_pre_clip': 6.462093353271484, 'timing/data': 0.00037174695171415806, 'timing/model': 1.2462369600543752, 'learning_rate/qwen_vl_interface': 8.176663766713852e-06, 'learning_rate/action_model': 8.176663766713852e-05, 'epoch': 0.62}) +[2026-06-17 10:04:50,417][starVLA.training.train_starvla][INFO] - Step 2093, Loss: {'train/loss': 0.8474998474121094, 'train/grad_norm_pre_clip': 10.40101146697998, 'timing/data': 0.00036365597043186426, 'timing/model': 1.2107379430672154, 'learning_rate/qwen_vl_interface': 8.174960048597099e-06, 'learning_rate/action_model': 8.174960048597099e-05, 'epoch': 0.62}) +[2026-06-17 10:04:55,156][starVLA.training.train_starvla][INFO] - Step 2094, Loss: {'train/loss': 1.080220341682434, 'train/grad_norm_pre_clip': 9.43710708618164, 'timing/data': 0.0003479729639366269, 'timing/model': 1.208837136044167, 'learning_rate/qwen_vl_interface': 8.173255724132982e-06, 'learning_rate/action_model': 8.17325572413298e-05, 'epoch': 0.62}) +[2026-06-17 10:04:59,918][starVLA.training.train_starvla][INFO] - Step 2095, Loss: {'train/loss': 0.5064656138420105, 'train/grad_norm_pre_clip': 6.417356014251709, 'timing/data': 0.0006736170034855604, 'timing/model': 1.2105342029826716, 'learning_rate/qwen_vl_interface': 8.171550793674806e-06, 'learning_rate/action_model': 8.171550793674805e-05, 'epoch': 0.62}) +[2026-06-17 10:05:04,735][starVLA.training.train_starvla][INFO] - Step 2096, Loss: {'train/loss': 0.6403694748878479, 'train/grad_norm_pre_clip': 6.169316291809082, 'timing/data': 0.0003661880036816001, 'timing/model': 1.2054513440234587, 'learning_rate/qwen_vl_interface': 8.169845257576008e-06, 'learning_rate/action_model': 8.169845257576007e-05, 'epoch': 0.62}) +[2026-06-17 10:05:09,562][starVLA.training.train_starvla][INFO] - Step 2097, Loss: {'train/loss': 0.8156145215034485, 'train/grad_norm_pre_clip': 7.629873275756836, 'timing/data': 0.000394799979403615, 'timing/model': 1.35361656290479, 'learning_rate/qwen_vl_interface': 8.168139116190146e-06, 'learning_rate/action_model': 8.168139116190146e-05, 'epoch': 0.62}) +[2026-06-17 10:05:14,228][starVLA.training.train_starvla][INFO] - Step 2098, Loss: {'train/loss': 0.8069478273391724, 'train/grad_norm_pre_clip': 10.387077331542969, 'timing/data': 0.0004394400166347623, 'timing/model': 1.2050904240459204, 'learning_rate/qwen_vl_interface': 8.166432369870905e-06, 'learning_rate/action_model': 8.166432369870904e-05, 'epoch': 0.62}) +[2026-06-17 10:05:19,029][starVLA.training.train_starvla][INFO] - Step 2099, Loss: {'train/loss': 1.2440271377563477, 'train/grad_norm_pre_clip': 12.699560165405273, 'timing/data': 0.0005516899982467294, 'timing/model': 1.2083846819587052, 'learning_rate/qwen_vl_interface': 8.164725018972096e-06, 'learning_rate/action_model': 8.164725018972096e-05, 'epoch': 0.62}) +[2026-06-17 10:07:01,306][starVLA.training.train_starvla][INFO] - Step 2100, Loss: {'train/loss': 0.7021304965019226, 'train/grad_norm_pre_clip': 9.152483940124512, 'eval/loss': 0.8601734638214111, 'eval/latency_0/loss': 0.8601734638214111, 'eval/demon_fire_precision': 0.8909531502423264, 'eval/demon_fire_recall': 0.9161129568106312, 'eval/demon_fire_f1': 0.9033579033579034, 'eval/demon_fire_rate_model': 0.6885428253615128, 'eval/demon_fire_rate_teacher': 0.6696329254727474, 'eval/demon_left_precision': 0.7335926105979582, 'eval/demon_left_recall': 0.854957507082153, 'eval/demon_left_f1': 0.7896389324960754, 'eval/demon_left_rate_model': 0.5720244716351501, 'eval/demon_left_rate_teacher': 0.49082313681868744, 'eval/demon_right_precision': 0.8305312710154674, 'eval/demon_right_recall': 0.6977401129943502, 'eval/demon_right_f1': 0.7583665950260976, 'eval/demon_right_rate_model': 0.4135150166852058, 'eval/demon_right_rate_teacher': 0.4922135706340378, 'eval/demon_move_f1': 0.7740027637610865, 'eval/demon_cc_f1': 0.860228234362768, 'eval/latency_0/demon_fire_precision': 0.8909531502423264, 'eval/latency_0/demon_fire_recall': 0.9161129568106312, 'eval/latency_0/demon_fire_f1': 0.9033579033579034, 'eval/latency_0/demon_fire_rate_model': 0.6885428253615128, 'eval/latency_0/demon_fire_rate_teacher': 0.6696329254727474, 'eval/latency_0/demon_left_precision': 0.7335926105979582, 'eval/latency_0/demon_left_recall': 0.854957507082153, 'eval/latency_0/demon_left_f1': 0.7896389324960754, 'eval/latency_0/demon_left_rate_model': 0.5720244716351501, 'eval/latency_0/demon_left_rate_teacher': 0.49082313681868744, 'eval/latency_0/demon_right_precision': 0.8305312710154674, 'eval/latency_0/demon_right_recall': 0.6977401129943502, 'eval/latency_0/demon_right_f1': 0.7583665950260976, 'eval/latency_0/demon_right_rate_model': 0.4135150166852058, 'eval/latency_0/demon_right_rate_teacher': 0.4922135706340378, 'eval/latency_0/demon_move_f1': 0.7740027637610865, 'eval/latency_0/demon_cc_f1': 0.860228234362768, 'timing/data': 0.0003499520244076848, 'timing/model': 1.2114368960028514, 'learning_rate/qwen_vl_interface': 8.163017063847654e-06, 'learning_rate/action_model': 8.163017063847654e-05, 'epoch': 0.62}) +[2026-06-17 10:07:01,307][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_2100_state +[2026-06-17 10:07:01,309][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 10:07:04,869][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_2100_state/model.safetensors +[2026-06-17 10:07:12,532][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_2100_state/optimizer.bin +[2026-06-17 10:07:12,533][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_2100_state/sampler.bin +[2026-06-17 10:07:12,533][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_2100_state/sampler_1.bin +[2026-06-17 10:07:12,533][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_2100_state/random_states_0.pkl +[2026-06-17 10:07:12,587][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2100 +[2026-06-17 10:07:12,588][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 10:07:12,627][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 10:07:17,352][starVLA.training.train_starvla][INFO] - Step 2101, Loss: {'train/loss': 0.92091965675354, 'train/grad_norm_pre_clip': 8.873771667480469, 'timing/data': 0.00035890506114810705, 'timing/model': 1.187453336082399, 'learning_rate/qwen_vl_interface': 8.16130850485164e-06, 'learning_rate/action_model': 8.161308504851639e-05, 'epoch': 0.62}) +[2026-06-17 10:07:22,054][starVLA.training.train_starvla][INFO] - Step 2102, Loss: {'train/loss': 0.8604891896247864, 'train/grad_norm_pre_clip': 9.905537605285645, 'timing/data': 0.00040992104914039373, 'timing/model': 1.1894164379918948, 'learning_rate/qwen_vl_interface': 8.15959934233824e-06, 'learning_rate/action_model': 8.15959934233824e-05, 'epoch': 0.62}) +[2026-06-17 10:07:26,856][starVLA.training.train_starvla][INFO] - Step 2103, Loss: {'train/loss': 1.08053719997406, 'train/grad_norm_pre_clip': 12.35064697265625, 'timing/data': 0.00037479191087186337, 'timing/model': 1.1930680010700598, 'learning_rate/qwen_vl_interface': 8.157889576661765e-06, 'learning_rate/action_model': 8.157889576661764e-05, 'epoch': 0.62}) +[2026-06-17 10:07:31,659][starVLA.training.train_starvla][INFO] - Step 2104, Loss: {'train/loss': 0.9582471251487732, 'train/grad_norm_pre_clip': 10.93014144897461, 'timing/data': 0.00039695308078080416, 'timing/model': 1.1931640279944986, 'learning_rate/qwen_vl_interface': 8.156179208176652e-06, 'learning_rate/action_model': 8.156179208176652e-05, 'epoch': 0.62}) +[2026-06-17 10:07:36,439][starVLA.training.train_starvla][INFO] - Step 2105, Loss: {'train/loss': 0.7506586313247681, 'train/grad_norm_pre_clip': 11.73814582824707, 'timing/data': 0.0003663769457489252, 'timing/model': 1.2745436319382861, 'learning_rate/qwen_vl_interface': 8.154468237237463e-06, 'learning_rate/action_model': 8.154468237237462e-05, 'epoch': 0.62}) +[2026-06-17 10:07:41,058][starVLA.training.train_starvla][INFO] - Step 2106, Loss: {'train/loss': 0.6629002690315247, 'train/grad_norm_pre_clip': 7.090382099151611, 'timing/data': 0.000375150004401803, 'timing/model': 1.1962607820751145, 'learning_rate/qwen_vl_interface': 8.152756664198882e-06, 'learning_rate/action_model': 8.152756664198882e-05, 'epoch': 0.62}) +[2026-06-17 10:07:45,795][starVLA.training.train_starvla][INFO] - Step 2107, Loss: {'train/loss': 0.6157114505767822, 'train/grad_norm_pre_clip': 7.926961421966553, 'timing/data': 0.0003859719727188349, 'timing/model': 1.1983657609671354, 'learning_rate/qwen_vl_interface': 8.151044489415722e-06, 'learning_rate/action_model': 8.151044489415722e-05, 'epoch': 0.62}) +[2026-06-17 10:07:50,559][starVLA.training.train_starvla][INFO] - Step 2108, Loss: {'train/loss': 1.143930435180664, 'train/grad_norm_pre_clip': 13.18615436553955, 'timing/data': 0.00033781793899834156, 'timing/model': 1.1962690020445734, 'learning_rate/qwen_vl_interface': 8.149331713242916e-06, 'learning_rate/action_model': 8.149331713242916e-05, 'epoch': 0.62}) +[2026-06-17 10:07:55,314][starVLA.training.train_starvla][INFO] - Step 2109, Loss: {'train/loss': 0.782651960849762, 'train/grad_norm_pre_clip': 10.124267578125, 'timing/data': 0.0003494770498946309, 'timing/model': 1.197919863043353, 'learning_rate/qwen_vl_interface': 8.147618336035526e-06, 'learning_rate/action_model': 8.147618336035526e-05, 'epoch': 0.62}) +[2026-06-17 10:08:00,068][starVLA.training.train_starvla][INFO] - Step 2110, Loss: {'train/loss': 1.0363435745239258, 'train/grad_norm_pre_clip': 9.046350479125977, 'timing/data': 0.0003497350262477994, 'timing/model': 1.3268509870395064, 'learning_rate/qwen_vl_interface': 8.14590435814874e-06, 'learning_rate/action_model': 8.14590435814874e-05, 'epoch': 0.62}) +[2026-06-17 10:08:04,712][starVLA.training.train_starvla][INFO] - Step 2111, Loss: {'train/loss': 0.5849485993385315, 'train/grad_norm_pre_clip': 6.831969261169434, 'timing/data': 0.0005723469657823443, 'timing/model': 1.2026796819409356, 'learning_rate/qwen_vl_interface': 8.144189779937862e-06, 'learning_rate/action_model': 8.144189779937862e-05, 'epoch': 0.62}) +[2026-06-17 10:08:09,503][starVLA.training.train_starvla][INFO] - Step 2112, Loss: {'train/loss': 0.7813185453414917, 'train/grad_norm_pre_clip': 7.803801536560059, 'timing/data': 0.00034631602466106415, 'timing/model': 1.2031391309574246, 'learning_rate/qwen_vl_interface': 8.142474601758331e-06, 'learning_rate/action_model': 8.142474601758329e-05, 'epoch': 0.62}) +[2026-06-17 10:08:14,278][starVLA.training.train_starvla][INFO] - Step 2113, Loss: {'train/loss': 0.9404700994491577, 'train/grad_norm_pre_clip': 9.527288436889648, 'timing/data': 0.00034152204170823097, 'timing/model': 1.2021647079382092, 'learning_rate/qwen_vl_interface': 8.1407588239657e-06, 'learning_rate/action_model': 8.140758823965698e-05, 'epoch': 0.62}) +[2026-06-17 10:08:19,061][starVLA.training.train_starvla][INFO] - Step 2114, Loss: {'train/loss': 1.0031951665878296, 'train/grad_norm_pre_clip': 9.262688636779785, 'timing/data': 0.0003854050301015377, 'timing/model': 1.1990719619207084, 'learning_rate/qwen_vl_interface': 8.139042446915653e-06, 'learning_rate/action_model': 8.139042446915654e-05, 'epoch': 0.62}) +[2026-06-17 10:08:23,853][starVLA.training.train_starvla][INFO] - Step 2115, Loss: {'train/loss': 0.8842873573303223, 'train/grad_norm_pre_clip': 10.267708778381348, 'timing/data': 0.0003502259496599436, 'timing/model': 1.2021863129921257, 'learning_rate/qwen_vl_interface': 8.137325470964001e-06, 'learning_rate/action_model': 8.137325470964e-05, 'epoch': 0.62}) +[2026-06-17 10:08:28,624][starVLA.training.train_starvla][INFO] - Step 2116, Loss: {'train/loss': 1.1254324913024902, 'train/grad_norm_pre_clip': 14.326814651489258, 'timing/data': 0.0003582249628379941, 'timing/model': 1.3285687669413164, 'learning_rate/qwen_vl_interface': 8.135607896466669e-06, 'learning_rate/action_model': 8.13560789646667e-05, 'epoch': 0.62}) +[2026-06-17 10:08:33,263][starVLA.training.train_starvla][INFO] - Step 2117, Loss: {'train/loss': 0.8490455746650696, 'train/grad_norm_pre_clip': 8.581069946289062, 'timing/data': 0.00035890203434973955, 'timing/model': 1.2004811979131773, 'learning_rate/qwen_vl_interface': 8.133889723779718e-06, 'learning_rate/action_model': 8.133889723779717e-05, 'epoch': 0.62}) +[2026-06-17 10:08:38,034][starVLA.training.train_starvla][INFO] - Step 2118, Loss: {'train/loss': 0.8840577006340027, 'train/grad_norm_pre_clip': 7.845103740692139, 'timing/data': 0.000396291958168149, 'timing/model': 1.2010773789370432, 'learning_rate/qwen_vl_interface': 8.132170953259321e-06, 'learning_rate/action_model': 8.132170953259322e-05, 'epoch': 0.62}) +[2026-06-17 10:08:42,819][starVLA.training.train_starvla][INFO] - Step 2119, Loss: {'train/loss': 0.9349063634872437, 'train/grad_norm_pre_clip': 9.672403335571289, 'timing/data': 0.0003753430210053921, 'timing/model': 1.2017811709083617, 'learning_rate/qwen_vl_interface': 8.130451585261786e-06, 'learning_rate/action_model': 8.130451585261786e-05, 'epoch': 0.62}) +[2026-06-17 10:08:47,548][starVLA.training.train_starvla][INFO] - Step 2120, Loss: {'train/loss': 0.8467856049537659, 'train/grad_norm_pre_clip': 9.510225296020508, 'timing/data': 0.0003782260464504361, 'timing/model': 1.2045786939561367, 'learning_rate/qwen_vl_interface': 8.128731620143538e-06, 'learning_rate/action_model': 8.128731620143538e-05, 'epoch': 0.62}) +[2026-06-17 10:08:52,328][starVLA.training.train_starvla][INFO] - Step 2121, Loss: {'train/loss': 0.7672266960144043, 'train/grad_norm_pre_clip': 7.708453178405762, 'timing/data': 0.0003332309424877167, 'timing/model': 1.2035848679952323, 'learning_rate/qwen_vl_interface': 8.127011058261129e-06, 'learning_rate/action_model': 8.127011058261128e-05, 'epoch': 0.62}) +[2026-06-17 10:08:57,180][starVLA.training.train_starvla][INFO] - Step 2122, Loss: {'train/loss': 0.6571903824806213, 'train/grad_norm_pre_clip': 6.622488975524902, 'timing/data': 0.00037232006434351206, 'timing/model': 1.3320277930470183, 'learning_rate/qwen_vl_interface': 8.125289899971229e-06, 'learning_rate/action_model': 8.125289899971228e-05, 'epoch': 0.62}) +[2026-06-17 10:09:01,831][starVLA.training.train_starvla][INFO] - Step 2123, Loss: {'train/loss': 0.9067370891571045, 'train/grad_norm_pre_clip': 8.800432205200195, 'timing/data': 0.000412505934946239, 'timing/model': 1.202260214020498, 'learning_rate/qwen_vl_interface': 8.123568145630641e-06, 'learning_rate/action_model': 8.12356814563064e-05, 'epoch': 0.62}) +[2026-06-17 10:09:06,628][starVLA.training.train_starvla][INFO] - Step 2124, Loss: {'train/loss': 1.0190739631652832, 'train/grad_norm_pre_clip': 14.98851490020752, 'timing/data': 0.0006129060639068484, 'timing/model': 1.2082790629938245, 'learning_rate/qwen_vl_interface': 8.121845795596282e-06, 'learning_rate/action_model': 8.121845795596282e-05, 'epoch': 0.62}) +[2026-06-17 10:09:11,427][starVLA.training.train_starvla][INFO] - Step 2125, Loss: {'train/loss': 0.773664116859436, 'train/grad_norm_pre_clip': 8.343173027038574, 'timing/data': 0.00037930801045149565, 'timing/model': 1.2087773280218244, 'learning_rate/qwen_vl_interface': 8.1201228502252e-06, 'learning_rate/action_model': 8.120122850225199e-05, 'epoch': 0.63}) +[2026-06-17 10:09:16,235][starVLA.training.train_starvla][INFO] - Step 2126, Loss: {'train/loss': 0.8876814246177673, 'train/grad_norm_pre_clip': 8.335491180419922, 'timing/data': 0.0003390469355508685, 'timing/model': 1.224084137007594, 'learning_rate/qwen_vl_interface': 8.118399309874563e-06, 'learning_rate/action_model': 8.118399309874563e-05, 'epoch': 0.63}) +[2026-06-17 10:09:21,020][starVLA.training.train_starvla][INFO] - Step 2127, Loss: {'train/loss': 0.8002246618270874, 'train/grad_norm_pre_clip': 11.465367317199707, 'timing/data': 0.00034799904096871614, 'timing/model': 1.2019541279878467, 'learning_rate/qwen_vl_interface': 8.116675174901663e-06, 'learning_rate/action_model': 8.116675174901662e-05, 'epoch': 0.63}) +[2026-06-17 10:09:25,809][starVLA.training.train_starvla][INFO] - Step 2128, Loss: {'train/loss': 0.8485163450241089, 'train/grad_norm_pre_clip': 7.451671123504639, 'timing/data': 0.0004783520707860589, 'timing/model': 1.3447785100433975, 'learning_rate/qwen_vl_interface': 8.114950445663915e-06, 'learning_rate/action_model': 8.114950445663913e-05, 'epoch': 0.63}) +[2026-06-17 10:09:30,459][starVLA.training.train_starvla][INFO] - Step 2129, Loss: {'train/loss': 0.5813248157501221, 'train/grad_norm_pre_clip': 8.107322692871094, 'timing/data': 0.0007137670181691647, 'timing/model': 1.2014565210556611, 'learning_rate/qwen_vl_interface': 8.113225122518854e-06, 'learning_rate/action_model': 8.113225122518854e-05, 'epoch': 0.63}) +[2026-06-17 10:09:35,249][starVLA.training.train_starvla][INFO] - Step 2130, Loss: {'train/loss': 1.0141364336013794, 'train/grad_norm_pre_clip': 10.486371040344238, 'timing/data': 0.0004429830005392432, 'timing/model': 1.2035837540170178, 'learning_rate/qwen_vl_interface': 8.111499205824148e-06, 'learning_rate/action_model': 8.111499205824147e-05, 'epoch': 0.63}) +[2026-06-17 10:09:40,030][starVLA.training.train_starvla][INFO] - Step 2131, Loss: {'train/loss': 0.8808408975601196, 'train/grad_norm_pre_clip': 10.520721435546875, 'timing/data': 0.0004120949888601899, 'timing/model': 1.2040120570454746, 'learning_rate/qwen_vl_interface': 8.109772695937575e-06, 'learning_rate/action_model': 8.109772695937574e-05, 'epoch': 0.63}) +[2026-06-17 10:09:44,815][starVLA.training.train_starvla][INFO] - Step 2132, Loss: {'train/loss': 0.6566514372825623, 'train/grad_norm_pre_clip': 8.75950813293457, 'timing/data': 0.00041468499694019556, 'timing/model': 1.2025949760572985, 'learning_rate/qwen_vl_interface': 8.108045593217045e-06, 'learning_rate/action_model': 8.108045593217045e-05, 'epoch': 0.63}) +[2026-06-17 10:09:49,588][starVLA.training.train_starvla][INFO] - Step 2133, Loss: {'train/loss': 0.7303927540779114, 'train/grad_norm_pre_clip': 8.06708812713623, 'timing/data': 0.0003578849136829376, 'timing/model': 1.3310069900471717, 'learning_rate/qwen_vl_interface': 8.106317898020589e-06, 'learning_rate/action_model': 8.106317898020589e-05, 'epoch': 0.63}) +[2026-06-17 10:09:54,330][starVLA.training.train_starvla][INFO] - Step 2134, Loss: {'train/loss': 0.8076069951057434, 'train/grad_norm_pre_clip': 9.684835433959961, 'timing/data': 0.00037658808287233114, 'timing/model': 1.2893729659263045, 'learning_rate/qwen_vl_interface': 8.104589610706359e-06, 'learning_rate/action_model': 8.104589610706359e-05, 'epoch': 0.63}) +[2026-06-17 10:09:58,979][starVLA.training.train_starvla][INFO] - Step 2135, Loss: {'train/loss': 0.6616266369819641, 'train/grad_norm_pre_clip': 7.0392231941223145, 'timing/data': 0.0003605419769883156, 'timing/model': 1.2040636690799147, 'learning_rate/qwen_vl_interface': 8.102860731632632e-06, 'learning_rate/action_model': 8.102860731632631e-05, 'epoch': 0.63}) +[2026-06-17 10:10:03,764][starVLA.training.train_starvla][INFO] - Step 2136, Loss: {'train/loss': 0.7673442363739014, 'train/grad_norm_pre_clip': 7.8419108390808105, 'timing/data': 0.0003724100533872843, 'timing/model': 1.2037897560512647, 'learning_rate/qwen_vl_interface': 8.101131261157802e-06, 'learning_rate/action_model': 8.101131261157802e-05, 'epoch': 0.63}) +[2026-06-17 10:10:08,541][starVLA.training.train_starvla][INFO] - Step 2137, Loss: {'train/loss': 1.0208004713058472, 'train/grad_norm_pre_clip': 10.503203392028809, 'timing/data': 0.0005040720570832491, 'timing/model': 1.2028611010173336, 'learning_rate/qwen_vl_interface': 8.099401199640397e-06, 'learning_rate/action_model': 8.099401199640397e-05, 'epoch': 0.63}) +[2026-06-17 10:10:13,321][starVLA.training.train_starvla][INFO] - Step 2138, Loss: {'train/loss': 0.9058319330215454, 'train/grad_norm_pre_clip': 11.059263229370117, 'timing/data': 0.0005742840003222227, 'timing/model': 1.2052564900368452, 'learning_rate/qwen_vl_interface': 8.097670547439057e-06, 'learning_rate/action_model': 8.097670547439058e-05, 'epoch': 0.63}) +[2026-06-17 10:10:18,098][starVLA.training.train_starvla][INFO] - Step 2139, Loss: {'train/loss': 1.0350182056427002, 'train/grad_norm_pre_clip': 8.707290649414062, 'timing/data': 0.00041382608469575644, 'timing/model': 1.3300981119973585, 'learning_rate/qwen_vl_interface': 8.095939304912549e-06, 'learning_rate/action_model': 8.095939304912548e-05, 'epoch': 0.63}) +[2026-06-17 10:10:22,761][starVLA.training.train_starvla][INFO] - Step 2140, Loss: {'train/loss': 1.153460144996643, 'train/grad_norm_pre_clip': 12.073945999145508, 'timing/data': 0.00039637996815145016, 'timing/model': 1.2138349590823054, 'learning_rate/qwen_vl_interface': 8.094207472419759e-06, 'learning_rate/action_model': 8.094207472419759e-05, 'epoch': 0.63}) +[2026-06-17 10:10:27,421][starVLA.training.train_starvla][INFO] - Step 2141, Loss: {'train/loss': 0.6593098044395447, 'train/grad_norm_pre_clip': 9.418545722961426, 'timing/data': 0.00042295095045119524, 'timing/model': 1.2050385850016028, 'learning_rate/qwen_vl_interface': 8.092475050319699e-06, 'learning_rate/action_model': 8.092475050319698e-05, 'epoch': 0.63}) +[2026-06-17 10:10:32,234][starVLA.training.train_starvla][INFO] - Step 2142, Loss: {'train/loss': 0.9002348780632019, 'train/grad_norm_pre_clip': 9.972118377685547, 'timing/data': 0.0003430629149079323, 'timing/model': 1.2042645229957998, 'learning_rate/qwen_vl_interface': 8.090742038971503e-06, 'learning_rate/action_model': 8.090742038971503e-05, 'epoch': 0.63}) +[2026-06-17 10:10:37,026][starVLA.training.train_starvla][INFO] - Step 2143, Loss: {'train/loss': 0.8038345575332642, 'train/grad_norm_pre_clip': 8.713166236877441, 'timing/data': 0.000401264987885952, 'timing/model': 1.205751314992085, 'learning_rate/qwen_vl_interface': 8.089008438734427e-06, 'learning_rate/action_model': 8.089008438734427e-05, 'epoch': 0.63}) +[2026-06-17 10:10:41,820][starVLA.training.train_starvla][INFO] - Step 2144, Loss: {'train/loss': 0.7850713729858398, 'train/grad_norm_pre_clip': 7.867225170135498, 'timing/data': 0.0004274190869182348, 'timing/model': 1.204521854990162, 'learning_rate/qwen_vl_interface': 8.087274249967844e-06, 'learning_rate/action_model': 8.087274249967845e-05, 'epoch': 0.63}) +[2026-06-17 10:10:46,606][starVLA.training.train_starvla][INFO] - Step 2145, Loss: {'train/loss': 0.6280835270881653, 'train/grad_norm_pre_clip': 7.248603343963623, 'timing/data': 0.00037947308737784624, 'timing/model': 1.3357511010253802, 'learning_rate/qwen_vl_interface': 8.08553947303126e-06, 'learning_rate/action_model': 8.085539473031259e-05, 'epoch': 0.63}) +[2026-06-17 10:10:51,264][starVLA.training.train_starvla][INFO] - Step 2146, Loss: {'train/loss': 0.7973201274871826, 'train/grad_norm_pre_clip': 8.965242385864258, 'timing/data': 0.00041246891487389803, 'timing/model': 1.2062737250234932, 'learning_rate/qwen_vl_interface': 8.083804108284288e-06, 'learning_rate/action_model': 8.083804108284287e-05, 'epoch': 0.63}) +[2026-06-17 10:10:56,049][starVLA.training.train_starvla][INFO] - Step 2147, Loss: {'train/loss': 0.7909018397331238, 'train/grad_norm_pre_clip': 7.493372917175293, 'timing/data': 0.00043431599624454975, 'timing/model': 1.20763162500225, 'learning_rate/qwen_vl_interface': 8.082068156086675e-06, 'learning_rate/action_model': 8.082068156086675e-05, 'epoch': 0.63}) +[2026-06-17 10:11:00,830][starVLA.training.train_starvla][INFO] - Step 2148, Loss: {'train/loss': 0.6206516027450562, 'train/grad_norm_pre_clip': 9.049577713012695, 'timing/data': 0.00034630997106432915, 'timing/model': 1.2050974640296772, 'learning_rate/qwen_vl_interface': 8.080331616798289e-06, 'learning_rate/action_model': 8.080331616798287e-05, 'epoch': 0.63}) +[2026-06-17 10:11:05,562][starVLA.training.train_starvla][INFO] - Step 2149, Loss: {'train/loss': 1.137671947479248, 'train/grad_norm_pre_clip': 14.45192813873291, 'timing/data': 0.0003867499763146043, 'timing/model': 1.2091095610521734, 'learning_rate/qwen_vl_interface': 8.078594490779112e-06, 'learning_rate/action_model': 8.078594490779111e-05, 'epoch': 0.63}) +[2026-06-17 10:11:10,348][starVLA.training.train_starvla][INFO] - Step 2150, Loss: {'train/loss': 0.719353199005127, 'train/grad_norm_pre_clip': 8.239571571350098, 'timing/data': 0.0003548399545252323, 'timing/model': 1.2059235500637442, 'learning_rate/qwen_vl_interface': 8.076856778389253e-06, 'learning_rate/action_model': 8.076856778389252e-05, 'epoch': 0.63}) +[2026-06-17 10:11:15,129][starVLA.training.train_starvla][INFO] - Step 2151, Loss: {'train/loss': 1.107806921005249, 'train/grad_norm_pre_clip': 10.497332572937012, 'timing/data': 0.0003511989489197731, 'timing/model': 1.33292211801745, 'learning_rate/qwen_vl_interface': 8.075118479988942e-06, 'learning_rate/action_model': 8.075118479988941e-05, 'epoch': 0.63}) +[2026-06-17 10:11:19,850][starVLA.training.train_starvla][INFO] - Step 2152, Loss: {'train/loss': 0.7896811366081238, 'train/grad_norm_pre_clip': 8.163718223571777, 'timing/data': 0.0003624639939516783, 'timing/model': 1.203395493910648, 'learning_rate/qwen_vl_interface': 8.07337959593853e-06, 'learning_rate/action_model': 8.073379595938529e-05, 'epoch': 0.63}) +[2026-06-17 10:11:24,636][starVLA.training.train_starvla][INFO] - Step 2153, Loss: {'train/loss': 0.896294116973877, 'train/grad_norm_pre_clip': 8.868425369262695, 'timing/data': 0.0005105379968881607, 'timing/model': 1.2033800199860707, 'learning_rate/qwen_vl_interface': 8.071640126598488e-06, 'learning_rate/action_model': 8.071640126598488e-05, 'epoch': 0.63}) +[2026-06-17 10:11:29,429][starVLA.training.train_starvla][INFO] - Step 2154, Loss: {'train/loss': 0.8543679714202881, 'train/grad_norm_pre_clip': 6.884687900543213, 'timing/data': 0.0003640599315986037, 'timing/model': 1.2039563331054524, 'learning_rate/qwen_vl_interface': 8.069900072329414e-06, 'learning_rate/action_model': 8.069900072329415e-05, 'epoch': 0.63}) +[2026-06-17 10:11:34,204][starVLA.training.train_starvla][INFO] - Step 2155, Loss: {'train/loss': 0.61445552110672, 'train/grad_norm_pre_clip': 7.180268287658691, 'timing/data': 0.00036896904930472374, 'timing/model': 1.204352986998856, 'learning_rate/qwen_vl_interface': 8.068159433492021e-06, 'learning_rate/action_model': 8.068159433492021e-05, 'epoch': 0.63}) +[2026-06-17 10:11:38,989][starVLA.training.train_starvla][INFO] - Step 2156, Loss: {'train/loss': 0.7966589331626892, 'train/grad_norm_pre_clip': 7.429232120513916, 'timing/data': 0.0005440609529614449, 'timing/model': 1.204733697930351, 'learning_rate/qwen_vl_interface': 8.066418210447144e-06, 'learning_rate/action_model': 8.066418210447143e-05, 'epoch': 0.63}) +[2026-06-17 10:11:43,785][starVLA.training.train_starvla][INFO] - Step 2157, Loss: {'train/loss': 0.7110028266906738, 'train/grad_norm_pre_clip': 8.331738471984863, 'timing/data': 0.00033838406670838594, 'timing/model': 1.3342491509392858, 'learning_rate/qwen_vl_interface': 8.064676403555741e-06, 'learning_rate/action_model': 8.064676403555741e-05, 'epoch': 0.63}) +[2026-06-17 10:11:48,476][starVLA.training.train_starvla][INFO] - Step 2158, Loss: {'train/loss': 1.0150567293167114, 'train/grad_norm_pre_clip': 15.289554595947266, 'timing/data': 0.0003736510407179594, 'timing/model': 1.2126993630081415, 'learning_rate/qwen_vl_interface': 8.06293401317889e-06, 'learning_rate/action_model': 8.062934013178891e-05, 'epoch': 0.63}) +[2026-06-17 10:11:53,269][starVLA.training.train_starvla][INFO] - Step 2159, Loss: {'train/loss': 0.8929020762443542, 'train/grad_norm_pre_clip': 11.493989944458008, 'timing/data': 0.0003359799738973379, 'timing/model': 1.2106898730853572, 'learning_rate/qwen_vl_interface': 8.061191039677793e-06, 'learning_rate/action_model': 8.061191039677793e-05, 'epoch': 0.64}) +[2026-06-17 10:11:58,048][starVLA.training.train_starvla][INFO] - Step 2160, Loss: {'train/loss': 0.9592207074165344, 'train/grad_norm_pre_clip': 9.428488731384277, 'timing/data': 0.0005733640864491463, 'timing/model': 1.2042800170602277, 'learning_rate/qwen_vl_interface': 8.059447483413768e-06, 'learning_rate/action_model': 8.059447483413767e-05, 'epoch': 0.64}) +[2026-06-17 10:12:02,841][starVLA.training.train_starvla][INFO] - Step 2161, Loss: {'train/loss': 0.8111619353294373, 'train/grad_norm_pre_clip': 9.497093200683594, 'timing/data': 0.00037876598071306944, 'timing/model': 1.2107292159926146, 'learning_rate/qwen_vl_interface': 8.057703344748258e-06, 'learning_rate/action_model': 8.057703344748257e-05, 'epoch': 0.64}) +[2026-06-17 10:12:07,631][starVLA.training.train_starvla][INFO] - Step 2162, Loss: {'train/loss': 0.9156438112258911, 'train/grad_norm_pre_clip': 8.40119457244873, 'timing/data': 0.00035884103272110224, 'timing/model': 1.3344962319824845, 'learning_rate/qwen_vl_interface': 8.05595862404282e-06, 'learning_rate/action_model': 8.05595862404282e-05, 'epoch': 0.64}) +[2026-06-17 10:12:12,363][starVLA.training.train_starvla][INFO] - Step 2163, Loss: {'train/loss': 0.850617527961731, 'train/grad_norm_pre_clip': 8.043534278869629, 'timing/data': 0.00041699595749378204, 'timing/model': 1.2808383509982377, 'learning_rate/qwen_vl_interface': 8.054213321659143e-06, 'learning_rate/action_model': 8.054213321659142e-05, 'epoch': 0.64}) +[2026-06-17 10:12:17,018][starVLA.training.train_starvla][INFO] - Step 2164, Loss: {'train/loss': 0.7905012369155884, 'train/grad_norm_pre_clip': 8.685640335083008, 'timing/data': 0.0003709070151671767, 'timing/model': 1.2034153329441324, 'learning_rate/qwen_vl_interface': 8.052467437959025e-06, 'learning_rate/action_model': 8.052467437959025e-05, 'epoch': 0.64}) +[2026-06-17 10:12:21,802][starVLA.training.train_starvla][INFO] - Step 2165, Loss: {'train/loss': 0.640200674533844, 'train/grad_norm_pre_clip': 8.7542085647583, 'timing/data': 0.00038239394780248404, 'timing/model': 1.2042876650812104, 'learning_rate/qwen_vl_interface': 8.050720973304393e-06, 'learning_rate/action_model': 8.050720973304391e-05, 'epoch': 0.64}) +[2026-06-17 10:12:26,589][starVLA.training.train_starvla][INFO] - Step 2166, Loss: {'train/loss': 0.6923935413360596, 'train/grad_norm_pre_clip': 8.513514518737793, 'timing/data': 0.00034644303377717733, 'timing/model': 1.2053825079929084, 'learning_rate/qwen_vl_interface': 8.048973928057288e-06, 'learning_rate/action_model': 8.048973928057288e-05, 'epoch': 0.64}) +[2026-06-17 10:12:31,378][starVLA.training.train_starvla][INFO] - Step 2167, Loss: {'train/loss': 0.8718826770782471, 'train/grad_norm_pre_clip': 8.899645805358887, 'timing/data': 0.0005665040807798505, 'timing/model': 1.206111463950947, 'learning_rate/qwen_vl_interface': 8.047226302579877e-06, 'learning_rate/action_model': 8.047226302579877e-05, 'epoch': 0.64}) +[2026-06-17 10:12:36,165][starVLA.training.train_starvla][INFO] - Step 2168, Loss: {'train/loss': 0.7936477065086365, 'train/grad_norm_pre_clip': 8.469681739807129, 'timing/data': 0.0003800679696723819, 'timing/model': 1.3307666439795867, 'learning_rate/qwen_vl_interface': 8.045478097234443e-06, 'learning_rate/action_model': 8.045478097234444e-05, 'epoch': 0.64}) +[2026-06-17 10:12:40,831][starVLA.training.train_starvla][INFO] - Step 2169, Loss: {'train/loss': 0.7894257307052612, 'train/grad_norm_pre_clip': 10.672043800354004, 'timing/data': 0.0003479029983282089, 'timing/model': 1.2161519399378449, 'learning_rate/qwen_vl_interface': 8.043729312383393e-06, 'learning_rate/action_model': 8.043729312383392e-05, 'epoch': 0.64}) +[2026-06-17 10:12:45,491][starVLA.training.train_starvla][INFO] - Step 2170, Loss: {'train/loss': 0.7339268922805786, 'train/grad_norm_pre_clip': 8.849571228027344, 'timing/data': 0.00044845195952802896, 'timing/model': 1.205544349970296, 'learning_rate/qwen_vl_interface': 8.041979948389248e-06, 'learning_rate/action_model': 8.041979948389248e-05, 'epoch': 0.64}) +[2026-06-17 10:12:50,292][starVLA.training.train_starvla][INFO] - Step 2171, Loss: {'train/loss': 0.9318597912788391, 'train/grad_norm_pre_clip': 9.696826934814453, 'timing/data': 0.00040477991569787264, 'timing/model': 1.2080670889699832, 'learning_rate/qwen_vl_interface': 8.040230005614656e-06, 'learning_rate/action_model': 8.040230005614656e-05, 'epoch': 0.64}) +[2026-06-17 10:12:55,077][starVLA.training.train_starvla][INFO] - Step 2172, Loss: {'train/loss': 0.6898372769355774, 'train/grad_norm_pre_clip': 10.550564765930176, 'timing/data': 0.00040869496297091246, 'timing/model': 1.2044863399351016, 'learning_rate/qwen_vl_interface': 8.038479484422383e-06, 'learning_rate/action_model': 8.038479484422383e-05, 'epoch': 0.64}) +[2026-06-17 10:12:59,880][starVLA.training.train_starvla][INFO] - Step 2173, Loss: {'train/loss': 0.684673011302948, 'train/grad_norm_pre_clip': 8.602041244506836, 'timing/data': 0.000354332965798676, 'timing/model': 1.2057637580437586, 'learning_rate/qwen_vl_interface': 8.03672838517531e-06, 'learning_rate/action_model': 8.036728385175311e-05, 'epoch': 0.64}) +[2026-06-17 10:13:04,661][starVLA.training.train_starvla][INFO] - Step 2174, Loss: {'train/loss': 0.9633951187133789, 'train/grad_norm_pre_clip': 11.141318321228027, 'timing/data': 0.00039812608156353235, 'timing/model': 1.3322517749620602, 'learning_rate/qwen_vl_interface': 8.034976708236446e-06, 'learning_rate/action_model': 8.034976708236445e-05, 'epoch': 0.64}) +[2026-06-17 10:13:09,321][starVLA.training.train_starvla][INFO] - Step 2175, Loss: {'train/loss': 0.7590557932853699, 'train/grad_norm_pre_clip': 8.00140380859375, 'timing/data': 0.00037783896550536156, 'timing/model': 1.2063497680937871, 'learning_rate/qwen_vl_interface': 8.033224453968913e-06, 'learning_rate/action_model': 8.033224453968913e-05, 'epoch': 0.64}) +[2026-06-17 10:13:14,101][starVLA.training.train_starvla][INFO] - Step 2176, Loss: {'train/loss': 0.6411822438240051, 'train/grad_norm_pre_clip': 9.27746295928955, 'timing/data': 0.0003629320999607444, 'timing/model': 1.2027013550978154, 'learning_rate/qwen_vl_interface': 8.031471622735956e-06, 'learning_rate/action_model': 8.031471622735956e-05, 'epoch': 0.64}) +[2026-06-17 10:13:18,882][starVLA.training.train_starvla][INFO] - Step 2177, Loss: {'train/loss': 0.5480669140815735, 'train/grad_norm_pre_clip': 6.79184103012085, 'timing/data': 0.0003405190072953701, 'timing/model': 1.204286961001344, 'learning_rate/qwen_vl_interface': 8.029718214900937e-06, 'learning_rate/action_model': 8.029718214900937e-05, 'epoch': 0.64}) +[2026-06-17 10:13:23,615][starVLA.training.train_starvla][INFO] - Step 2178, Loss: {'train/loss': 0.8711524605751038, 'train/grad_norm_pre_clip': 8.147970199584961, 'timing/data': 0.0004245250020176172, 'timing/model': 1.2042199859861284, 'learning_rate/qwen_vl_interface': 8.027964230827342e-06, 'learning_rate/action_model': 8.027964230827342e-05, 'epoch': 0.64}) +[2026-06-17 10:13:28,389][starVLA.training.train_starvla][INFO] - Step 2179, Loss: {'train/loss': 0.6716415286064148, 'train/grad_norm_pre_clip': 7.834355354309082, 'timing/data': 0.0003907890059053898, 'timing/model': 1.2064267770620063, 'learning_rate/qwen_vl_interface': 8.026209670878776e-06, 'learning_rate/action_model': 8.026209670878775e-05, 'epoch': 0.64}) +[2026-06-17 10:13:33,244][starVLA.training.train_starvla][INFO] - Step 2180, Loss: {'train/loss': 0.5970444083213806, 'train/grad_norm_pre_clip': 8.066003799438477, 'timing/data': 0.0003547370433807373, 'timing/model': 1.3532351540634409, 'learning_rate/qwen_vl_interface': 8.024454535418952e-06, 'learning_rate/action_model': 8.024454535418951e-05, 'epoch': 0.64}) +[2026-06-17 10:13:37,904][starVLA.training.train_starvla][INFO] - Step 2181, Loss: {'train/loss': 0.973887026309967, 'train/grad_norm_pre_clip': 10.70189094543457, 'timing/data': 0.00033608300145715475, 'timing/model': 1.20747099397704, 'learning_rate/qwen_vl_interface': 8.02269882481172e-06, 'learning_rate/action_model': 8.022698824811721e-05, 'epoch': 0.64}) +[2026-06-17 10:13:42,687][starVLA.training.train_starvla][INFO] - Step 2182, Loss: {'train/loss': 0.6790812611579895, 'train/grad_norm_pre_clip': 9.16413402557373, 'timing/data': 0.0003655370092019439, 'timing/model': 1.2037154149729759, 'learning_rate/qwen_vl_interface': 8.02094253942104e-06, 'learning_rate/action_model': 8.02094253942104e-05, 'epoch': 0.64}) +[2026-06-17 10:13:47,479][starVLA.training.train_starvla][INFO] - Step 2183, Loss: {'train/loss': 1.3238019943237305, 'train/grad_norm_pre_clip': 13.14898681640625, 'timing/data': 0.0004193219356238842, 'timing/model': 1.2048670849762857, 'learning_rate/qwen_vl_interface': 8.019185679610988e-06, 'learning_rate/action_model': 8.019185679610988e-05, 'epoch': 0.64}) +[2026-06-17 10:13:52,347][starVLA.training.train_starvla][INFO] - Step 2184, Loss: {'train/loss': 0.6354545950889587, 'train/grad_norm_pre_clip': 8.4208402633667, 'timing/data': 0.00037473998963832855, 'timing/model': 1.2100100580137223, 'learning_rate/qwen_vl_interface': 8.017428245745765e-06, 'learning_rate/action_model': 8.017428245745763e-05, 'epoch': 0.64}) +[2026-06-17 10:13:57,209][starVLA.training.train_starvla][INFO] - Step 2185, Loss: {'train/loss': 0.5973459482192993, 'train/grad_norm_pre_clip': 5.797667980194092, 'timing/data': 0.00038341106846928596, 'timing/model': 1.2130888659739867, 'learning_rate/qwen_vl_interface': 8.015670238189688e-06, 'learning_rate/action_model': 8.015670238189688e-05, 'epoch': 0.64}) +[2026-06-17 10:14:01,910][starVLA.training.train_starvla][INFO] - Step 2186, Loss: {'train/loss': 0.7835983037948608, 'train/grad_norm_pre_clip': 7.508190631866455, 'timing/data': 0.00035257195122539997, 'timing/model': 1.2517717810114846, 'learning_rate/qwen_vl_interface': 8.013911657307196e-06, 'learning_rate/action_model': 8.013911657307196e-05, 'epoch': 0.64}) +[2026-06-17 10:14:06,573][starVLA.training.train_starvla][INFO] - Step 2187, Loss: {'train/loss': 0.8556650876998901, 'train/grad_norm_pre_clip': 7.260438919067383, 'timing/data': 0.0003841270226985216, 'timing/model': 1.2078963030362502, 'learning_rate/qwen_vl_interface': 8.012152503462843e-06, 'learning_rate/action_model': 8.012152503462842e-05, 'epoch': 0.64}) +[2026-06-17 10:14:11,317][starVLA.training.train_starvla][INFO] - Step 2188, Loss: {'train/loss': 0.7512331604957581, 'train/grad_norm_pre_clip': 9.874403953552246, 'timing/data': 0.0003334319917485118, 'timing/model': 1.2062067080987617, 'learning_rate/qwen_vl_interface': 8.010392777021304e-06, 'learning_rate/action_model': 8.010392777021303e-05, 'epoch': 0.64}) +[2026-06-17 10:14:16,205][starVLA.training.train_starvla][INFO] - Step 2189, Loss: {'train/loss': 1.175949215888977, 'train/grad_norm_pre_clip': 14.42194938659668, 'timing/data': 0.00036988803185522556, 'timing/model': 1.2098147959914058, 'learning_rate/qwen_vl_interface': 8.008632478347372e-06, 'learning_rate/action_model': 8.008632478347371e-05, 'epoch': 0.64}) +[2026-06-17 10:14:21,048][starVLA.training.train_starvla][INFO] - Step 2190, Loss: {'train/loss': 0.7356377243995667, 'train/grad_norm_pre_clip': 9.681612014770508, 'timing/data': 0.00045142206363379955, 'timing/model': 1.2047000039601699, 'learning_rate/qwen_vl_interface': 8.006871607805957e-06, 'learning_rate/action_model': 8.006871607805957e-05, 'epoch': 0.64}) +[2026-06-17 10:14:25,832][starVLA.training.train_starvla][INFO] - Step 2191, Loss: {'train/loss': 0.814592719078064, 'train/grad_norm_pre_clip': 8.980332374572754, 'timing/data': 0.0004279640270397067, 'timing/model': 1.3325376559514552, 'learning_rate/qwen_vl_interface': 8.005110165762093e-06, 'learning_rate/action_model': 8.005110165762093e-05, 'epoch': 0.64}) +[2026-06-17 10:14:30,513][starVLA.training.train_starvla][INFO] - Step 2192, Loss: {'train/loss': 0.7689720392227173, 'train/grad_norm_pre_clip': 10.440796852111816, 'timing/data': 0.000440087984316051, 'timing/model': 1.206936675007455, 'learning_rate/qwen_vl_interface': 8.003348152580927e-06, 'learning_rate/action_model': 8.003348152580927e-05, 'epoch': 0.65}) +[2026-06-17 10:14:35,326][starVLA.training.train_starvla][INFO] - Step 2193, Loss: {'train/loss': 0.8525648713111877, 'train/grad_norm_pre_clip': 9.854742050170898, 'timing/data': 0.00036460801493376493, 'timing/model': 1.2102173450402915, 'learning_rate/qwen_vl_interface': 8.001585568627724e-06, 'learning_rate/action_model': 8.001585568627725e-05, 'epoch': 0.65}) +[2026-06-17 10:14:40,014][starVLA.training.train_starvla][INFO] - Step 2194, Loss: {'train/loss': 0.7695782780647278, 'train/grad_norm_pre_clip': 7.65356969833374, 'timing/data': 0.00037263904232531786, 'timing/model': 1.2091986919986084, 'learning_rate/qwen_vl_interface': 7.999822414267874e-06, 'learning_rate/action_model': 7.999822414267874e-05, 'epoch': 0.65}) +[2026-06-17 10:14:44,706][starVLA.training.train_starvla][INFO] - Step 2195, Loss: {'train/loss': 0.668484091758728, 'train/grad_norm_pre_clip': 7.067237377166748, 'timing/data': 0.00034661591053009033, 'timing/model': 1.203441774006933, 'learning_rate/qwen_vl_interface': 7.998058689866879e-06, 'learning_rate/action_model': 7.998058689866879e-05, 'epoch': 0.65}) +[2026-06-17 10:14:49,423][starVLA.training.train_starvla][INFO] - Step 2196, Loss: {'train/loss': 0.6196275949478149, 'train/grad_norm_pre_clip': 9.178557395935059, 'timing/data': 0.0005939600523561239, 'timing/model': 1.2098579009762034, 'learning_rate/qwen_vl_interface': 7.99629439579036e-06, 'learning_rate/action_model': 7.99629439579036e-05, 'epoch': 0.65}) +[2026-06-17 10:14:54,161][starVLA.training.train_starvla][INFO] - Step 2197, Loss: {'train/loss': 0.9304410815238953, 'train/grad_norm_pre_clip': 11.310990333557129, 'timing/data': 0.00034273904748260975, 'timing/model': 1.286645506042987, 'learning_rate/qwen_vl_interface': 7.994529532404059e-06, 'learning_rate/action_model': 7.994529532404058e-05, 'epoch': 0.65}) +[2026-06-17 10:14:58,844][starVLA.training.train_starvla][INFO] - Step 2198, Loss: {'train/loss': 0.6536409258842468, 'train/grad_norm_pre_clip': 9.94382095336914, 'timing/data': 0.0003711640601977706, 'timing/model': 1.2098210629774258, 'learning_rate/qwen_vl_interface': 7.99276410007383e-06, 'learning_rate/action_model': 7.99276410007383e-05, 'epoch': 0.65}) +[2026-06-17 10:15:03,647][starVLA.training.train_starvla][INFO] - Step 2199, Loss: {'train/loss': 1.0904605388641357, 'train/grad_norm_pre_clip': 18.027982711791992, 'timing/data': 0.00036197504960000515, 'timing/model': 1.2067524099256843, 'learning_rate/qwen_vl_interface': 7.990998099165652e-06, 'learning_rate/action_model': 7.990998099165652e-05, 'epoch': 0.65}) +[2026-06-17 10:16:45,979][starVLA.training.train_starvla][INFO] - Step 2200, Loss: {'train/loss': 0.8267753720283508, 'train/grad_norm_pre_clip': 8.347670555114746, 'eval/loss': 0.8329077959060669, 'eval/latency_0/loss': 0.8329077959060669, 'eval/demon_fire_precision': 0.9057635675220866, 'eval/demon_fire_recall': 0.8941029900332226, 'eval/demon_fire_f1': 0.8998955067920584, 'eval/demon_fire_rate_model': 0.6610122358175751, 'eval/demon_fire_rate_teacher': 0.6696329254727474, 'eval/demon_left_precision': 0.7759197324414716, 'eval/demon_left_recall': 0.7886685552407932, 'eval/demon_left_f1': 0.7822422028659736, 'eval/demon_left_rate_model': 0.4988876529477197, 'eval/demon_left_rate_teacher': 0.49082313681868744, 'eval/demon_right_precision': 0.7841037204058624, 'eval/demon_right_recall': 0.7858757062146893, 'eval/demon_right_f1': 0.7849887133182843, 'eval/demon_right_rate_model': 0.4933259176863181, 'eval/demon_right_rate_teacher': 0.4922135706340378, 'eval/demon_move_f1': 0.783615458092129, 'eval/demon_cc_f1': 0.861542436260743, 'eval/latency_0/demon_fire_precision': 0.9057635675220866, 'eval/latency_0/demon_fire_recall': 0.8941029900332226, 'eval/latency_0/demon_fire_f1': 0.8998955067920584, 'eval/latency_0/demon_fire_rate_model': 0.6610122358175751, 'eval/latency_0/demon_fire_rate_teacher': 0.6696329254727474, 'eval/latency_0/demon_left_precision': 0.7759197324414716, 'eval/latency_0/demon_left_recall': 0.7886685552407932, 'eval/latency_0/demon_left_f1': 0.7822422028659736, 'eval/latency_0/demon_left_rate_model': 0.4988876529477197, 'eval/latency_0/demon_left_rate_teacher': 0.49082313681868744, 'eval/latency_0/demon_right_precision': 0.7841037204058624, 'eval/latency_0/demon_right_recall': 0.7858757062146893, 'eval/latency_0/demon_right_f1': 0.7849887133182843, 'eval/latency_0/demon_right_rate_model': 0.4933259176863181, 'eval/latency_0/demon_right_rate_teacher': 0.4922135706340378, 'eval/latency_0/demon_move_f1': 0.783615458092129, 'eval/latency_0/demon_cc_f1': 0.861542436260743, 'timing/data': 0.0003704150440171361, 'timing/model': 1.2127665800508112, 'learning_rate/qwen_vl_interface': 7.98923153004562e-06, 'learning_rate/action_model': 7.98923153004562e-05, 'epoch': 0.65}) +[2026-06-17 10:16:45,980][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_2200_state +[2026-06-17 10:16:45,983][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 10:16:49,548][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_2200_state/model.safetensors +[2026-06-17 10:16:57,221][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_2200_state/optimizer.bin +[2026-06-17 10:16:57,221][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_2200_state/sampler.bin +[2026-06-17 10:16:57,221][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_2200_state/sampler_1.bin +[2026-06-17 10:16:57,222][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_2200_state/random_states_0.pkl +[2026-06-17 10:16:57,276][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2200 +[2026-06-17 10:16:57,276][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 10:16:57,317][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 10:17:02,040][starVLA.training.train_starvla][INFO] - Step 2201, Loss: {'train/loss': 0.9589052200317383, 'train/grad_norm_pre_clip': 10.488105773925781, 'timing/data': 0.0003601530333980918, 'timing/model': 1.1875741360709071, 'learning_rate/qwen_vl_interface': 7.987464393079943e-06, 'learning_rate/action_model': 7.987464393079943e-05, 'epoch': 0.65}) +[2026-06-17 10:17:06,774][starVLA.training.train_starvla][INFO] - Step 2202, Loss: {'train/loss': 0.7737165689468384, 'train/grad_norm_pre_clip': 8.227948188781738, 'timing/data': 0.0004809510428458452, 'timing/model': 1.1883412069873884, 'learning_rate/qwen_vl_interface': 7.985696688634953e-06, 'learning_rate/action_model': 7.985696688634952e-05, 'epoch': 0.65}) +[2026-06-17 10:17:11,510][starVLA.training.train_starvla][INFO] - Step 2203, Loss: {'train/loss': 0.9430286288261414, 'train/grad_norm_pre_clip': 13.030861854553223, 'timing/data': 0.000364272971637547, 'timing/model': 1.191677900031209, 'learning_rate/qwen_vl_interface': 7.983928417077095e-06, 'learning_rate/action_model': 7.983928417077095e-05, 'epoch': 0.65}) +[2026-06-17 10:17:16,249][starVLA.training.train_starvla][INFO] - Step 2204, Loss: {'train/loss': 0.7841342687606812, 'train/grad_norm_pre_clip': 9.015166282653809, 'timing/data': 0.0003727229777723551, 'timing/model': 1.3227597379591316, 'learning_rate/qwen_vl_interface': 7.982159578772935e-06, 'learning_rate/action_model': 7.982159578772934e-05, 'epoch': 0.65}) +[2026-06-17 10:17:20,869][starVLA.training.train_starvla][INFO] - Step 2205, Loss: {'train/loss': 0.7972093820571899, 'train/grad_norm_pre_clip': 10.105992317199707, 'timing/data': 0.00036683399230241776, 'timing/model': 1.1940969310235232, 'learning_rate/qwen_vl_interface': 7.980390174089154e-06, 'learning_rate/action_model': 7.980390174089153e-05, 'epoch': 0.65}) +[2026-06-17 10:17:25,614][starVLA.training.train_starvla][INFO] - Step 2206, Loss: {'train/loss': 0.788750946521759, 'train/grad_norm_pre_clip': 62.203617095947266, 'timing/data': 0.00039276108145713806, 'timing/model': 1.1952612299937755, 'learning_rate/qwen_vl_interface': 7.978620203392551e-06, 'learning_rate/action_model': 7.97862020339255e-05, 'epoch': 0.65}) +[2026-06-17 10:17:30,413][starVLA.training.train_starvla][INFO] - Step 2207, Loss: {'train/loss': 0.6606331467628479, 'train/grad_norm_pre_clip': 8.062406539916992, 'timing/data': 0.000488925026729703, 'timing/model': 1.1964803920127451, 'learning_rate/qwen_vl_interface': 7.976849667050044e-06, 'learning_rate/action_model': 7.976849667050044e-05, 'epoch': 0.65}) +[2026-06-17 10:17:35,151][starVLA.training.train_starvla][INFO] - Step 2208, Loss: {'train/loss': 0.8072746396064758, 'train/grad_norm_pre_clip': 13.250936508178711, 'timing/data': 0.0003549790708348155, 'timing/model': 1.1984014209592715, 'learning_rate/qwen_vl_interface': 7.975078565428667e-06, 'learning_rate/action_model': 7.975078565428668e-05, 'epoch': 0.65}) +[2026-06-17 10:17:39,896][starVLA.training.train_starvla][INFO] - Step 2209, Loss: {'train/loss': 0.9282350540161133, 'train/grad_norm_pre_clip': 8.326329231262207, 'timing/data': 0.0003465120680630207, 'timing/model': 1.1942365799332038, 'learning_rate/qwen_vl_interface': 7.97330689889557e-06, 'learning_rate/action_model': 7.97330689889557e-05, 'epoch': 0.65}) +[2026-06-17 10:17:44,664][starVLA.training.train_starvla][INFO] - Step 2210, Loss: {'train/loss': 0.5689451098442078, 'train/grad_norm_pre_clip': 8.499361991882324, 'timing/data': 0.00034348201006650925, 'timing/model': 1.335088333929889, 'learning_rate/qwen_vl_interface': 7.971534667818026e-06, 'learning_rate/action_model': 7.971534667818025e-05, 'epoch': 0.65}) +[2026-06-17 10:17:49,299][starVLA.training.train_starvla][INFO] - Step 2211, Loss: {'train/loss': 1.1636734008789062, 'train/grad_norm_pre_clip': 10.916502952575684, 'timing/data': 0.0006285669514909387, 'timing/model': 1.198998254025355, 'learning_rate/qwen_vl_interface': 7.969761872563413e-06, 'learning_rate/action_model': 7.969761872563413e-05, 'epoch': 0.65}) +[2026-06-17 10:17:54,088][starVLA.training.train_starvla][INFO] - Step 2212, Loss: {'train/loss': 0.6235938668251038, 'train/grad_norm_pre_clip': 7.749293804168701, 'timing/data': 0.0003698060754686594, 'timing/model': 1.1999333560233936, 'learning_rate/qwen_vl_interface': 7.96798851349924e-06, 'learning_rate/action_model': 7.96798851349924e-05, 'epoch': 0.65}) +[2026-06-17 10:17:58,794][starVLA.training.train_starvla][INFO] - Step 2213, Loss: {'train/loss': 0.7385499477386475, 'train/grad_norm_pre_clip': 10.653961181640625, 'timing/data': 0.0003818940604105592, 'timing/model': 1.210545901907608, 'learning_rate/qwen_vl_interface': 7.966214590993122e-06, 'learning_rate/action_model': 7.966214590993122e-05, 'epoch': 0.65}) +[2026-06-17 10:18:03,512][starVLA.training.train_starvla][INFO] - Step 2214, Loss: {'train/loss': 0.9326186180114746, 'train/grad_norm_pre_clip': 10.349700927734375, 'timing/data': 0.0003608010010793805, 'timing/model': 1.199380764970556, 'learning_rate/qwen_vl_interface': 7.964440105412797e-06, 'learning_rate/action_model': 7.964440105412796e-05, 'epoch': 0.65}) +[2026-06-17 10:18:08,288][starVLA.training.train_starvla][INFO] - Step 2215, Loss: {'train/loss': 0.8469106554985046, 'train/grad_norm_pre_clip': 9.656702995300293, 'timing/data': 0.0003650449216365814, 'timing/model': 1.199951704009436, 'learning_rate/qwen_vl_interface': 7.962665057126116e-06, 'learning_rate/action_model': 7.962665057126117e-05, 'epoch': 0.65}) +[2026-06-17 10:18:13,066][starVLA.training.train_starvla][INFO] - Step 2216, Loss: {'train/loss': 0.8428338766098022, 'train/grad_norm_pre_clip': 11.103921890258789, 'timing/data': 0.0003768099704757333, 'timing/model': 1.3354365730192512, 'learning_rate/qwen_vl_interface': 7.960889446501052e-06, 'learning_rate/action_model': 7.960889446501052e-05, 'epoch': 0.65}) +[2026-06-17 10:18:17,724][starVLA.training.train_starvla][INFO] - Step 2217, Loss: {'train/loss': 0.8048932552337646, 'train/grad_norm_pre_clip': 11.366887092590332, 'timing/data': 0.00038328301161527634, 'timing/model': 1.2048977320082486, 'learning_rate/qwen_vl_interface': 7.959113273905688e-06, 'learning_rate/action_model': 7.959113273905688e-05, 'epoch': 0.65}) +[2026-06-17 10:18:22,504][starVLA.training.train_starvla][INFO] - Step 2218, Loss: {'train/loss': 1.0952860116958618, 'train/grad_norm_pre_clip': 9.537683486938477, 'timing/data': 0.00035417801700532436, 'timing/model': 1.2030265049543232, 'learning_rate/qwen_vl_interface': 7.957336539708226e-06, 'learning_rate/action_model': 7.957336539708225e-05, 'epoch': 0.65}) +[2026-06-17 10:18:27,274][starVLA.training.train_starvla][INFO] - Step 2219, Loss: {'train/loss': 0.7616920471191406, 'train/grad_norm_pre_clip': 11.888594627380371, 'timing/data': 0.0003582289209589362, 'timing/model': 1.2027396970661357, 'learning_rate/qwen_vl_interface': 7.955559244276987e-06, 'learning_rate/action_model': 7.955559244276987e-05, 'epoch': 0.65}) +[2026-06-17 10:18:32,072][starVLA.training.train_starvla][INFO] - Step 2220, Loss: {'train/loss': 0.8890801668167114, 'train/grad_norm_pre_clip': 8.490310668945312, 'timing/data': 0.00035271700471639633, 'timing/model': 1.2099236520007253, 'learning_rate/qwen_vl_interface': 7.953781387980404e-06, 'learning_rate/action_model': 7.953781387980404e-05, 'epoch': 0.65}) +[2026-06-17 10:18:36,846][starVLA.training.train_starvla][INFO] - Step 2221, Loss: {'train/loss': 0.5144329071044922, 'train/grad_norm_pre_clip': 13.314101219177246, 'timing/data': 0.0003294609487056732, 'timing/model': 1.202785397064872, 'learning_rate/qwen_vl_interface': 7.952002971187032e-06, 'learning_rate/action_model': 7.952002971187032e-05, 'epoch': 0.65}) +[2026-06-17 10:18:41,623][starVLA.training.train_starvla][INFO] - Step 2222, Loss: {'train/loss': 0.8344408273696899, 'train/grad_norm_pre_clip': 10.281668663024902, 'timing/data': 0.0003694529877975583, 'timing/model': 1.3324635479366407, 'learning_rate/qwen_vl_interface': 7.950223994265534e-06, 'learning_rate/action_model': 7.950223994265533e-05, 'epoch': 0.65}) +[2026-06-17 10:18:46,270][starVLA.training.train_starvla][INFO] - Step 2223, Loss: {'train/loss': 0.8258189558982849, 'train/grad_norm_pre_clip': 9.606438636779785, 'timing/data': 0.0003767759772017598, 'timing/model': 1.2051378620089963, 'learning_rate/qwen_vl_interface': 7.9484444575847e-06, 'learning_rate/action_model': 7.948444457584699e-05, 'epoch': 0.65}) +[2026-06-17 10:18:51,064][starVLA.training.train_starvla][INFO] - Step 2224, Loss: {'train/loss': 0.7610435485839844, 'train/grad_norm_pre_clip': 8.972055435180664, 'timing/data': 0.00036930793430656195, 'timing/model': 1.2061258039902896, 'learning_rate/qwen_vl_interface': 7.946664361513423e-06, 'learning_rate/action_model': 7.946664361513421e-05, 'epoch': 0.65}) +[2026-06-17 10:18:55,850][starVLA.training.train_starvla][INFO] - Step 2225, Loss: {'train/loss': 0.718090832233429, 'train/grad_norm_pre_clip': 28.250661849975586, 'timing/data': 0.00035024306271225214, 'timing/model': 1.206553939031437, 'learning_rate/qwen_vl_interface': 7.944883706420724e-06, 'learning_rate/action_model': 7.944883706420724e-05, 'epoch': 0.65}) +[2026-06-17 10:19:00,629][starVLA.training.train_starvla][INFO] - Step 2226, Loss: {'train/loss': 0.7843595743179321, 'train/grad_norm_pre_clip': 7.949249267578125, 'timing/data': 0.0003447659546509385, 'timing/model': 1.207440200028941, 'learning_rate/qwen_vl_interface': 7.943102492675732e-06, 'learning_rate/action_model': 7.943102492675732e-05, 'epoch': 0.66}) +[2026-06-17 10:19:05,434][starVLA.training.train_starvla][INFO] - Step 2227, Loss: {'train/loss': 0.9128661751747131, 'train/grad_norm_pre_clip': 11.013382911682129, 'timing/data': 0.00035267695784568787, 'timing/model': 1.341697697993368, 'learning_rate/qwen_vl_interface': 7.941320720647694e-06, 'learning_rate/action_model': 7.941320720647695e-05, 'epoch': 0.66}) +[2026-06-17 10:19:10,177][starVLA.training.train_starvla][INFO] - Step 2228, Loss: {'train/loss': 0.6478790640830994, 'train/grad_norm_pre_clip': 8.207971572875977, 'timing/data': 0.0003503910265862942, 'timing/model': 1.2911390570225194, 'learning_rate/qwen_vl_interface': 7.939538390705976e-06, 'learning_rate/action_model': 7.939538390705976e-05, 'epoch': 0.66}) +[2026-06-17 10:19:14,837][starVLA.training.train_starvla][INFO] - Step 2229, Loss: {'train/loss': 0.7109139561653137, 'train/grad_norm_pre_clip': 11.006096839904785, 'timing/data': 0.00047400000039488077, 'timing/model': 1.2031423170119524, 'learning_rate/qwen_vl_interface': 7.937755503220057e-06, 'learning_rate/action_model': 7.937755503220056e-05, 'epoch': 0.66}) +[2026-06-17 10:19:19,619][starVLA.training.train_starvla][INFO] - Step 2230, Loss: {'train/loss': 0.6869795322418213, 'train/grad_norm_pre_clip': 9.01516056060791, 'timing/data': 0.00032979995012283325, 'timing/model': 1.207172255963087, 'learning_rate/qwen_vl_interface': 7.93597205855953e-06, 'learning_rate/action_model': 7.935972058559529e-05, 'epoch': 0.66}) +[2026-06-17 10:19:24,406][starVLA.training.train_starvla][INFO] - Step 2231, Loss: {'train/loss': 0.9314661622047424, 'train/grad_norm_pre_clip': 8.63909912109375, 'timing/data': 0.0003421000437811017, 'timing/model': 1.2041790030198172, 'learning_rate/qwen_vl_interface': 7.934188057094103e-06, 'learning_rate/action_model': 7.934188057094103e-05, 'epoch': 0.66}) +[2026-06-17 10:19:29,201][starVLA.training.train_starvla][INFO] - Step 2232, Loss: {'train/loss': 0.6761325001716614, 'train/grad_norm_pre_clip': 7.814753532409668, 'timing/data': 0.0007310799555853009, 'timing/model': 1.2052443309221417, 'learning_rate/qwen_vl_interface': 7.932403499193604e-06, 'learning_rate/action_model': 7.932403499193605e-05, 'epoch': 0.66}) +[2026-06-17 10:19:34,020][starVLA.training.train_starvla][INFO] - Step 2233, Loss: {'train/loss': 1.0448405742645264, 'train/grad_norm_pre_clip': 10.503643989562988, 'timing/data': 0.0003928069490939379, 'timing/model': 1.32902158703655, 'learning_rate/qwen_vl_interface': 7.930618385227975e-06, 'learning_rate/action_model': 7.930618385227973e-05, 'epoch': 0.66}) +[2026-06-17 10:19:38,678][starVLA.training.train_starvla][INFO] - Step 2234, Loss: {'train/loss': 0.7900015115737915, 'train/grad_norm_pre_clip': 9.934686660766602, 'timing/data': 0.0004187899176031351, 'timing/model': 1.2122492250055075, 'learning_rate/qwen_vl_interface': 7.928832715567268e-06, 'learning_rate/action_model': 7.928832715567267e-05, 'epoch': 0.66}) +[2026-06-17 10:19:43,331][starVLA.training.train_starvla][INFO] - Step 2235, Loss: {'train/loss': 0.5970677733421326, 'train/grad_norm_pre_clip': 8.399968147277832, 'timing/data': 0.0003633539890870452, 'timing/model': 1.203690274967812, 'learning_rate/qwen_vl_interface': 7.927046490581655e-06, 'learning_rate/action_model': 7.927046490581655e-05, 'epoch': 0.66}) +[2026-06-17 10:19:48,104][starVLA.training.train_starvla][INFO] - Step 2236, Loss: {'train/loss': 0.9484743475914001, 'train/grad_norm_pre_clip': 10.056534767150879, 'timing/data': 0.0003452280070632696, 'timing/model': 1.2022928580408916, 'learning_rate/qwen_vl_interface': 7.925259710641428e-06, 'learning_rate/action_model': 7.925259710641427e-05, 'epoch': 0.66}) +[2026-06-17 10:19:52,881][starVLA.training.train_starvla][INFO] - Step 2237, Loss: {'train/loss': 0.786723792552948, 'train/grad_norm_pre_clip': 13.273900985717773, 'timing/data': 0.0003535500727593899, 'timing/model': 1.2026219780091196, 'learning_rate/qwen_vl_interface': 7.923472376116982e-06, 'learning_rate/action_model': 7.923472376116982e-05, 'epoch': 0.66}) +[2026-06-17 10:19:57,666][starVLA.training.train_starvla][INFO] - Step 2238, Loss: {'train/loss': 0.9181665182113647, 'train/grad_norm_pre_clip': 11.531681060791016, 'timing/data': 0.0006527360528707504, 'timing/model': 1.205889274016954, 'learning_rate/qwen_vl_interface': 7.921684487378836e-06, 'learning_rate/action_model': 7.921684487378837e-05, 'epoch': 0.66}) +[2026-06-17 10:20:02,469][starVLA.training.train_starvla][INFO] - Step 2239, Loss: {'train/loss': 0.5950451493263245, 'train/grad_norm_pre_clip': 8.877344131469727, 'timing/data': 0.0003689200384542346, 'timing/model': 1.3494558059610426, 'learning_rate/qwen_vl_interface': 7.919896044797622e-06, 'learning_rate/action_model': 7.91989604479762e-05, 'epoch': 0.66}) +[2026-06-17 10:20:07,117][starVLA.training.train_starvla][INFO] - Step 2240, Loss: {'train/loss': 0.850919246673584, 'train/grad_norm_pre_clip': 9.479161262512207, 'timing/data': 0.00041827093809843063, 'timing/model': 1.2027906429721043, 'learning_rate/qwen_vl_interface': 7.918107048744083e-06, 'learning_rate/action_model': 7.918107048744083e-05, 'epoch': 0.66}) +[2026-06-17 10:20:11,896][starVLA.training.train_starvla][INFO] - Step 2241, Loss: {'train/loss': 0.975178062915802, 'train/grad_norm_pre_clip': 9.206753730773926, 'timing/data': 0.00035090011078864336, 'timing/model': 1.2051837489707395, 'learning_rate/qwen_vl_interface': 7.916317499589083e-06, 'learning_rate/action_model': 7.916317499589083e-05, 'epoch': 0.66}) +[2026-06-17 10:20:16,667][starVLA.training.train_starvla][INFO] - Step 2242, Loss: {'train/loss': 0.9841415882110596, 'train/grad_norm_pre_clip': 8.05402946472168, 'timing/data': 0.00043428002391010523, 'timing/model': 1.2049514130922034, 'learning_rate/qwen_vl_interface': 7.914527397703596e-06, 'learning_rate/action_model': 7.914527397703595e-05, 'epoch': 0.66}) +[2026-06-17 10:20:21,449][starVLA.training.train_starvla][INFO] - Step 2243, Loss: {'train/loss': 1.0000687837600708, 'train/grad_norm_pre_clip': 9.764863967895508, 'timing/data': 0.00042107398621737957, 'timing/model': 1.2024768000701442, 'learning_rate/qwen_vl_interface': 7.912736743458712e-06, 'learning_rate/action_model': 7.912736743458711e-05, 'epoch': 0.66}) +[2026-06-17 10:20:26,241][starVLA.training.train_starvla][INFO] - Step 2244, Loss: {'train/loss': 0.7146568298339844, 'train/grad_norm_pre_clip': 9.633234977722168, 'timing/data': 0.000351571012288332, 'timing/model': 1.205460648983717, 'learning_rate/qwen_vl_interface': 7.910945537225638e-06, 'learning_rate/action_model': 7.910945537225637e-05, 'epoch': 0.66}) +[2026-06-17 10:20:31,044][starVLA.training.train_starvla][INFO] - Step 2245, Loss: {'train/loss': 0.9616594910621643, 'train/grad_norm_pre_clip': 9.071160316467285, 'timing/data': 0.0006693219766020775, 'timing/model': 1.3332937409868464, 'learning_rate/qwen_vl_interface': 7.909153779375688e-06, 'learning_rate/action_model': 7.909153779375688e-05, 'epoch': 0.66}) +[2026-06-17 10:20:35,702][starVLA.training.train_starvla][INFO] - Step 2246, Loss: {'train/loss': 0.6879386901855469, 'train/grad_norm_pre_clip': 9.88797378540039, 'timing/data': 0.00034245895221829414, 'timing/model': 1.2068412119988352, 'learning_rate/qwen_vl_interface': 7.907361470280299e-06, 'learning_rate/action_model': 7.907361470280299e-05, 'epoch': 0.66}) +[2026-06-17 10:20:40,511][starVLA.training.train_starvla][INFO] - Step 2247, Loss: {'train/loss': 0.6669531464576721, 'train/grad_norm_pre_clip': 7.662152290344238, 'timing/data': 0.0003489799564704299, 'timing/model': 1.204575237003155, 'learning_rate/qwen_vl_interface': 7.905568610311018e-06, 'learning_rate/action_model': 7.905568610311017e-05, 'epoch': 0.66}) +[2026-06-17 10:20:45,222][starVLA.training.train_starvla][INFO] - Step 2248, Loss: {'train/loss': 0.5579113960266113, 'train/grad_norm_pre_clip': 6.298892974853516, 'timing/data': 0.00032646802719682455, 'timing/model': 1.2057086650747806, 'learning_rate/qwen_vl_interface': 7.903775199839506e-06, 'learning_rate/action_model': 7.903775199839505e-05, 'epoch': 0.66}) +[2026-06-17 10:20:49,963][starVLA.training.train_starvla][INFO] - Step 2249, Loss: {'train/loss': 0.7237384915351868, 'train/grad_norm_pre_clip': 7.5013861656188965, 'timing/data': 0.0004044140223413706, 'timing/model': 1.2070223490009084, 'learning_rate/qwen_vl_interface': 7.901981239237539e-06, 'learning_rate/action_model': 7.901981239237539e-05, 'epoch': 0.66}) +[2026-06-17 10:20:54,744][starVLA.training.train_starvla][INFO] - Step 2250, Loss: {'train/loss': 0.8104068040847778, 'train/grad_norm_pre_clip': 8.003625869750977, 'timing/data': 0.0003672089660540223, 'timing/model': 1.200940506067127, 'learning_rate/qwen_vl_interface': 7.900186728877009e-06, 'learning_rate/action_model': 7.90018672887701e-05, 'epoch': 0.66}) +[2026-06-17 10:20:59,525][starVLA.training.train_starvla][INFO] - Step 2251, Loss: {'train/loss': 0.7101343274116516, 'train/grad_norm_pre_clip': 7.732643127441406, 'timing/data': 0.00035434402525424957, 'timing/model': 1.3316736159613356, 'learning_rate/qwen_vl_interface': 7.898391669129919e-06, 'learning_rate/action_model': 7.898391669129918e-05, 'epoch': 0.66}) +[2026-06-17 10:21:04,183][starVLA.training.train_starvla][INFO] - Step 2252, Loss: {'train/loss': 1.0128511190414429, 'train/grad_norm_pre_clip': 9.743181228637695, 'timing/data': 0.0003501379396766424, 'timing/model': 1.2080453881062567, 'learning_rate/qwen_vl_interface': 7.896596060368384e-06, 'learning_rate/action_model': 7.896596060368384e-05, 'epoch': 0.66}) +[2026-06-17 10:21:08,972][starVLA.training.train_starvla][INFO] - Step 2253, Loss: {'train/loss': 0.5546651482582092, 'train/grad_norm_pre_clip': 8.224976539611816, 'timing/data': 0.0004226970486342907, 'timing/model': 1.2069770520320162, 'learning_rate/qwen_vl_interface': 7.89479990296464e-06, 'learning_rate/action_model': 7.894799902964639e-05, 'epoch': 0.66}) +[2026-06-17 10:21:13,753][starVLA.training.train_starvla][INFO] - Step 2254, Loss: {'train/loss': 0.7672391533851624, 'train/grad_norm_pre_clip': 8.78127384185791, 'timing/data': 0.00035248405765742064, 'timing/model': 1.2038435169961303, 'learning_rate/qwen_vl_interface': 7.893003197291028e-06, 'learning_rate/action_model': 7.893003197291027e-05, 'epoch': 0.66}) +[2026-06-17 10:21:18,522][starVLA.training.train_starvla][INFO] - Step 2255, Loss: {'train/loss': 0.7507067918777466, 'train/grad_norm_pre_clip': 12.988814353942871, 'timing/data': 0.0003653720486909151, 'timing/model': 1.2031969190575182, 'learning_rate/qwen_vl_interface': 7.89120594372001e-06, 'learning_rate/action_model': 7.89120594372001e-05, 'epoch': 0.66}) +[2026-06-17 10:21:23,311][starVLA.training.train_starvla][INFO] - Step 2256, Loss: {'train/loss': 0.980469822883606, 'train/grad_norm_pre_clip': 13.860543251037598, 'timing/data': 0.0003579959738999605, 'timing/model': 1.205272878985852, 'learning_rate/qwen_vl_interface': 7.88940814262416e-06, 'learning_rate/action_model': 7.889408142624158e-05, 'epoch': 0.66}) +[2026-06-17 10:21:28,110][starVLA.training.train_starvla][INFO] - Step 2257, Loss: {'train/loss': 0.9556125402450562, 'train/grad_norm_pre_clip': 10.098454475402832, 'timing/data': 0.0003662139642983675, 'timing/model': 1.3422962060431018, 'learning_rate/qwen_vl_interface': 7.887609794376158e-06, 'learning_rate/action_model': 7.887609794376158e-05, 'epoch': 0.66}) +[2026-06-17 10:21:32,772][starVLA.training.train_starvla][INFO] - Step 2258, Loss: {'train/loss': 1.135270357131958, 'train/grad_norm_pre_clip': 15.219876289367676, 'timing/data': 0.0003438349813222885, 'timing/model': 1.2040808820165694, 'learning_rate/qwen_vl_interface': 7.88581089934881e-06, 'learning_rate/action_model': 7.88581089934881e-05, 'epoch': 0.66}) +[2026-06-17 10:21:37,562][starVLA.training.train_starvla][INFO] - Step 2259, Loss: {'train/loss': 0.7413668036460876, 'train/grad_norm_pre_clip': 9.917052268981934, 'timing/data': 0.0003500219900161028, 'timing/model': 1.2075544099789113, 'learning_rate/qwen_vl_interface': 7.884011457915025e-06, 'learning_rate/action_model': 7.884011457915024e-05, 'epoch': 0.66}) +[2026-06-17 10:21:42,349][starVLA.training.train_starvla][INFO] - Step 2260, Loss: {'train/loss': 0.8708024621009827, 'train/grad_norm_pre_clip': 11.432892799377441, 'timing/data': 0.00036187702789902687, 'timing/model': 1.2094828829867765, 'learning_rate/qwen_vl_interface': 7.882211470447832e-06, 'learning_rate/action_model': 7.882211470447831e-05, 'epoch': 0.67}) +[2026-06-17 10:21:47,125][starVLA.training.train_starvla][INFO] - Step 2261, Loss: {'train/loss': 0.8250477313995361, 'train/grad_norm_pre_clip': 10.465668678283691, 'timing/data': 0.0003470679512247443, 'timing/model': 1.2063241809373721, 'learning_rate/qwen_vl_interface': 7.880410937320367e-06, 'learning_rate/action_model': 7.880410937320368e-05, 'epoch': 0.67}) +[2026-06-17 10:21:51,903][starVLA.training.train_starvla][INFO] - Step 2262, Loss: {'train/loss': 0.7922075986862183, 'train/grad_norm_pre_clip': 13.257819175720215, 'timing/data': 0.0003418140113353729, 'timing/model': 1.3309062630869448, 'learning_rate/qwen_vl_interface': 7.878609858905885e-06, 'learning_rate/action_model': 7.878609858905886e-05, 'epoch': 0.67}) +[2026-06-17 10:21:56,640][starVLA.training.train_starvla][INFO] - Step 2263, Loss: {'train/loss': 0.5328677892684937, 'train/grad_norm_pre_clip': 6.586428165435791, 'timing/data': 0.0003583680372685194, 'timing/model': 1.2906836499460042, 'learning_rate/qwen_vl_interface': 7.876808235577751e-06, 'learning_rate/action_model': 7.87680823557775e-05, 'epoch': 0.67}) +[2026-06-17 10:22:01,293][starVLA.training.train_starvla][INFO] - Step 2264, Loss: {'train/loss': 0.6728081107139587, 'train/grad_norm_pre_clip': 16.4830379486084, 'timing/data': 0.0003734281053766608, 'timing/model': 1.2040104300249368, 'learning_rate/qwen_vl_interface': 7.875006067709444e-06, 'learning_rate/action_model': 7.875006067709443e-05, 'epoch': 0.67}) +[2026-06-17 10:22:06,081][starVLA.training.train_starvla][INFO] - Step 2265, Loss: {'train/loss': 1.0989834070205688, 'train/grad_norm_pre_clip': 11.863086700439453, 'timing/data': 0.0003890119260177016, 'timing/model': 1.2031160539481789, 'learning_rate/qwen_vl_interface': 7.873203355674553e-06, 'learning_rate/action_model': 7.873203355674552e-05, 'epoch': 0.67}) +[2026-06-17 10:22:10,861][starVLA.training.train_starvla][INFO] - Step 2266, Loss: {'train/loss': 0.9152323007583618, 'train/grad_norm_pre_clip': 10.277807235717773, 'timing/data': 0.00035882589872926474, 'timing/model': 1.2048077170038596, 'learning_rate/qwen_vl_interface': 7.871400099846784e-06, 'learning_rate/action_model': 7.871400099846783e-05, 'epoch': 0.67}) +[2026-06-17 10:22:15,650][starVLA.training.train_starvla][INFO] - Step 2267, Loss: {'train/loss': 0.9969674944877625, 'train/grad_norm_pre_clip': 12.120402336120605, 'timing/data': 0.00034864305052906275, 'timing/model': 1.2055272969882935, 'learning_rate/qwen_vl_interface': 7.869596300599954e-06, 'learning_rate/action_model': 7.869596300599953e-05, 'epoch': 0.67}) +[2026-06-17 10:22:20,434][starVLA.training.train_starvla][INFO] - Step 2268, Loss: {'train/loss': 1.003667950630188, 'train/grad_norm_pre_clip': 13.281390190124512, 'timing/data': 0.00038146390579640865, 'timing/model': 1.3344929019222036, 'learning_rate/qwen_vl_interface': 7.867791958307992e-06, 'learning_rate/action_model': 7.867791958307992e-05, 'epoch': 0.67}) +[2026-06-17 10:22:25,109][starVLA.training.train_starvla][INFO] - Step 2269, Loss: {'train/loss': 1.0326361656188965, 'train/grad_norm_pre_clip': 11.862129211425781, 'timing/data': 0.000332736992277205, 'timing/model': 1.2156731659779325, 'learning_rate/qwen_vl_interface': 7.865987073344942e-06, 'learning_rate/action_model': 7.865987073344942e-05, 'epoch': 0.67}) +[2026-06-17 10:22:29,760][starVLA.training.train_starvla][INFO] - Step 2270, Loss: {'train/loss': 0.5023161172866821, 'train/grad_norm_pre_clip': 6.890783786773682, 'timing/data': 0.000371305039152503, 'timing/model': 1.204253304982558, 'learning_rate/qwen_vl_interface': 7.864181646084953e-06, 'learning_rate/action_model': 7.864181646084953e-05, 'epoch': 0.67}) +[2026-06-17 10:22:34,537][starVLA.training.train_starvla][INFO] - Step 2271, Loss: {'train/loss': 0.6880204677581787, 'train/grad_norm_pre_clip': 7.4194769859313965, 'timing/data': 0.0003542529884725809, 'timing/model': 1.2063359919702634, 'learning_rate/qwen_vl_interface': 7.862375676902298e-06, 'learning_rate/action_model': 7.862375676902297e-05, 'epoch': 0.67}) +[2026-06-17 10:22:39,324][starVLA.training.train_starvla][INFO] - Step 2272, Loss: {'train/loss': 0.628891110420227, 'train/grad_norm_pre_clip': 7.427760601043701, 'timing/data': 0.000348650966770947, 'timing/model': 1.2058312279405072, 'learning_rate/qwen_vl_interface': 7.860569166171353e-06, 'learning_rate/action_model': 7.860569166171354e-05, 'epoch': 0.67}) +[2026-06-17 10:22:44,107][starVLA.training.train_starvla][INFO] - Step 2273, Loss: {'train/loss': 0.7528725862503052, 'train/grad_norm_pre_clip': 8.937149047851562, 'timing/data': 0.00033878302201628685, 'timing/model': 1.203730596927926, 'learning_rate/qwen_vl_interface': 7.858762114266613e-06, 'learning_rate/action_model': 7.858762114266612e-05, 'epoch': 0.67}) +[2026-06-17 10:22:48,888][starVLA.training.train_starvla][INFO] - Step 2274, Loss: {'train/loss': 0.8915714025497437, 'train/grad_norm_pre_clip': 13.383142471313477, 'timing/data': 0.00033879000693559647, 'timing/model': 1.3317730070557445, 'learning_rate/qwen_vl_interface': 7.856954521562677e-06, 'learning_rate/action_model': 7.856954521562677e-05, 'epoch': 0.67}) +[2026-06-17 10:22:53,553][starVLA.training.train_starvla][INFO] - Step 2275, Loss: {'train/loss': 0.823487401008606, 'train/grad_norm_pre_clip': 9.105825424194336, 'timing/data': 0.0003347969613969326, 'timing/model': 1.2049422659911215, 'learning_rate/qwen_vl_interface': 7.855146388434264e-06, 'learning_rate/action_model': 7.855146388434264e-05, 'epoch': 0.67}) +[2026-06-17 10:22:58,331][starVLA.training.train_starvla][INFO] - Step 2276, Loss: {'train/loss': 0.7267741560935974, 'train/grad_norm_pre_clip': 8.27756118774414, 'timing/data': 0.0008894269121810794, 'timing/model': 1.2053128470433876, 'learning_rate/qwen_vl_interface': 7.853337715256202e-06, 'learning_rate/action_model': 7.853337715256203e-05, 'epoch': 0.67}) +[2026-06-17 10:23:03,124][starVLA.training.train_starvla][INFO] - Step 2277, Loss: {'train/loss': 1.000885248184204, 'train/grad_norm_pre_clip': 13.042871475219727, 'timing/data': 0.0006847389740869403, 'timing/model': 1.2067928479518741, 'learning_rate/qwen_vl_interface': 7.851528502403431e-06, 'learning_rate/action_model': 7.851528502403431e-05, 'epoch': 0.67}) +[2026-06-17 10:23:07,802][starVLA.training.train_starvla][INFO] - Step 2278, Loss: {'train/loss': 1.0611956119537354, 'train/grad_norm_pre_clip': 11.849173545837402, 'timing/data': 0.0005646368954330683, 'timing/model': 1.2134369640843943, 'learning_rate/qwen_vl_interface': 7.849718750251001e-06, 'learning_rate/action_model': 7.849718750251002e-05, 'epoch': 0.67}) +[2026-06-17 10:23:12,470][starVLA.training.train_starvla][INFO] - Step 2279, Loss: {'train/loss': 0.6974061727523804, 'train/grad_norm_pre_clip': 9.3080415725708, 'timing/data': 0.00040895689744502306, 'timing/model': 1.2036062850384042, 'learning_rate/qwen_vl_interface': 7.847908459174077e-06, 'learning_rate/action_model': 7.847908459174078e-05, 'epoch': 0.67}) +[2026-06-17 10:23:17,152][starVLA.training.train_starvla][INFO] - Step 2280, Loss: {'train/loss': 0.7480376362800598, 'train/grad_norm_pre_clip': 12.505732536315918, 'timing/data': 0.0003778430400416255, 'timing/model': 1.2308922749944031, 'learning_rate/qwen_vl_interface': 7.846097629547934e-06, 'learning_rate/action_model': 7.846097629547933e-05, 'epoch': 0.67}) +[2026-06-17 10:23:21,808][starVLA.training.train_starvla][INFO] - Step 2281, Loss: {'train/loss': 0.9505165815353394, 'train/grad_norm_pre_clip': 10.507270812988281, 'timing/data': 0.0003675539046525955, 'timing/model': 1.202778666978702, 'learning_rate/qwen_vl_interface': 7.844286261747959e-06, 'learning_rate/action_model': 7.844286261747958e-05, 'epoch': 0.67}) +[2026-06-17 10:23:26,603][starVLA.training.train_starvla][INFO] - Step 2282, Loss: {'train/loss': 0.6677471399307251, 'train/grad_norm_pre_clip': 9.234139442443848, 'timing/data': 0.00034435291308909655, 'timing/model': 1.2054800590267405, 'learning_rate/qwen_vl_interface': 7.84247435614965e-06, 'learning_rate/action_model': 7.84247435614965e-05, 'epoch': 0.67}) +[2026-06-17 10:23:31,360][starVLA.training.train_starvla][INFO] - Step 2283, Loss: {'train/loss': 0.8986150026321411, 'train/grad_norm_pre_clip': 8.955934524536133, 'timing/data': 0.000350181944668293, 'timing/model': 1.2114227860001847, 'learning_rate/qwen_vl_interface': 7.840661913128618e-06, 'learning_rate/action_model': 7.840661913128619e-05, 'epoch': 0.67}) +[2026-06-17 10:23:36,162][starVLA.training.train_starvla][INFO] - Step 2284, Loss: {'train/loss': 0.6063153147697449, 'train/grad_norm_pre_clip': 8.46094036102295, 'timing/data': 0.0003492090618237853, 'timing/model': 1.2053975110175088, 'learning_rate/qwen_vl_interface': 7.838848933060583e-06, 'learning_rate/action_model': 7.838848933060583e-05, 'epoch': 0.67}) +[2026-06-17 10:23:40,955][starVLA.training.train_starvla][INFO] - Step 2285, Loss: {'train/loss': 0.7241251468658447, 'train/grad_norm_pre_clip': 9.2855806350708, 'timing/data': 0.000421482021920383, 'timing/model': 1.3325912770815194, 'learning_rate/qwen_vl_interface': 7.83703541632138e-06, 'learning_rate/action_model': 7.83703541632138e-05, 'epoch': 0.67}) +[2026-06-17 10:23:45,619][starVLA.training.train_starvla][INFO] - Step 2286, Loss: {'train/loss': 0.7273889780044556, 'train/grad_norm_pre_clip': 11.08360481262207, 'timing/data': 0.00037035997956991196, 'timing/model': 1.2045868040295318, 'learning_rate/qwen_vl_interface': 7.835221363286949e-06, 'learning_rate/action_model': 7.835221363286949e-05, 'epoch': 0.67}) +[2026-06-17 10:23:50,421][starVLA.training.train_starvla][INFO] - Step 2287, Loss: {'train/loss': 0.8765421509742737, 'train/grad_norm_pre_clip': 11.811468124389648, 'timing/data': 0.0003872539382427931, 'timing/model': 1.204676699009724, 'learning_rate/qwen_vl_interface': 7.83340677433335e-06, 'learning_rate/action_model': 7.83340677433335e-05, 'epoch': 0.67}) +[2026-06-17 10:23:55,176][starVLA.training.train_starvla][INFO] - Step 2288, Loss: {'train/loss': 0.9018357992172241, 'train/grad_norm_pre_clip': 14.010519027709961, 'timing/data': 0.00040683301631361246, 'timing/model': 1.206947116064839, 'learning_rate/qwen_vl_interface': 7.831591649836745e-06, 'learning_rate/action_model': 7.831591649836744e-05, 'epoch': 0.67}) +[2026-06-17 10:23:59,959][starVLA.training.train_starvla][INFO] - Step 2289, Loss: {'train/loss': 0.8895105719566345, 'train/grad_norm_pre_clip': 19.056581497192383, 'timing/data': 0.00036032998468726873, 'timing/model': 1.2038820780580863, 'learning_rate/qwen_vl_interface': 7.829775990173413e-06, 'learning_rate/action_model': 7.829775990173413e-05, 'epoch': 0.67}) +[2026-06-17 10:24:04,761][starVLA.training.train_starvla][INFO] - Step 2290, Loss: {'train/loss': 1.0418481826782227, 'train/grad_norm_pre_clip': 11.954726219177246, 'timing/data': 0.0003436970291659236, 'timing/model': 1.2073816299671307, 'learning_rate/qwen_vl_interface': 7.827959795719744e-06, 'learning_rate/action_model': 7.827959795719744e-05, 'epoch': 0.67}) +[2026-06-17 10:24:09,584][starVLA.training.train_starvla][INFO] - Step 2291, Loss: {'train/loss': 0.5940684676170349, 'train/grad_norm_pre_clip': 6.46882438659668, 'timing/data': 0.00036601792089641094, 'timing/model': 1.33434750395827, 'learning_rate/qwen_vl_interface': 7.826143066852233e-06, 'learning_rate/action_model': 7.826143066852233e-05, 'epoch': 0.67}) +[2026-06-17 10:24:14,245][starVLA.training.train_starvla][INFO] - Step 2292, Loss: {'train/loss': 1.0101455450057983, 'train/grad_norm_pre_clip': 11.716583251953125, 'timing/data': 0.00038031896110624075, 'timing/model': 1.2073945930460468, 'learning_rate/qwen_vl_interface': 7.824325803947494e-06, 'learning_rate/action_model': 7.824325803947495e-05, 'epoch': 0.67}) +[2026-06-17 10:24:19,036][starVLA.training.train_starvla][INFO] - Step 2293, Loss: {'train/loss': 0.7644768953323364, 'train/grad_norm_pre_clip': 9.045872688293457, 'timing/data': 0.0003465119516476989, 'timing/model': 1.2052466260502115, 'learning_rate/qwen_vl_interface': 7.822508007382246e-06, 'learning_rate/action_model': 7.822508007382245e-05, 'epoch': 0.67}) +[2026-06-17 10:24:23,712][starVLA.training.train_starvla][INFO] - Step 2294, Loss: {'train/loss': 0.7427349090576172, 'train/grad_norm_pre_clip': 8.851567268371582, 'timing/data': 0.00033990503288805485, 'timing/model': 1.2054059359943494, 'learning_rate/qwen_vl_interface': 7.820689677533319e-06, 'learning_rate/action_model': 7.82068967753332e-05, 'epoch': 0.68}) +[2026-06-17 10:24:28,489][starVLA.training.train_starvla][INFO] - Step 2295, Loss: {'train/loss': 0.9271414875984192, 'train/grad_norm_pre_clip': 9.300508499145508, 'timing/data': 0.0003443020395934582, 'timing/model': 1.2093608770519495, 'learning_rate/qwen_vl_interface': 7.818870814777655e-06, 'learning_rate/action_model': 7.818870814777655e-05, 'epoch': 0.68}) +[2026-06-17 10:24:33,199][starVLA.training.train_starvla][INFO] - Step 2296, Loss: {'train/loss': 1.1076579093933105, 'train/grad_norm_pre_clip': 12.924487113952637, 'timing/data': 0.00033337902277708054, 'timing/model': 1.2099469299428165, 'learning_rate/qwen_vl_interface': 7.81705141949231e-06, 'learning_rate/action_model': 7.817051419492308e-05, 'epoch': 0.68}) +[2026-06-17 10:24:37,942][starVLA.training.train_starvla][INFO] - Step 2297, Loss: {'train/loss': 0.6619301438331604, 'train/grad_norm_pre_clip': 8.00635051727295, 'timing/data': 0.0005887870211154222, 'timing/model': 1.2797909800428897, 'learning_rate/qwen_vl_interface': 7.81523149205444e-06, 'learning_rate/action_model': 7.81523149205444e-05, 'epoch': 0.68}) +[2026-06-17 10:24:42,595][starVLA.training.train_starvla][INFO] - Step 2298, Loss: {'train/loss': 1.1212172508239746, 'train/grad_norm_pre_clip': 20.114337921142578, 'timing/data': 0.000347261899150908, 'timing/model': 1.2050421120366082, 'learning_rate/qwen_vl_interface': 7.813411032841323e-06, 'learning_rate/action_model': 7.813411032841323e-05, 'epoch': 0.68}) +[2026-06-17 10:24:47,440][starVLA.training.train_starvla][INFO] - Step 2299, Loss: {'train/loss': 0.8276929259300232, 'train/grad_norm_pre_clip': 11.005621910095215, 'timing/data': 0.0003681069938465953, 'timing/model': 1.204520751023665, 'learning_rate/qwen_vl_interface': 7.81159004223034e-06, 'learning_rate/action_model': 7.81159004223034e-05, 'epoch': 0.68}) +[2026-06-17 10:26:29,298][starVLA.training.train_starvla][INFO] - Step 2300, Loss: {'train/loss': 0.6406108140945435, 'train/grad_norm_pre_clip': 10.041158676147461, 'eval/loss': 0.8654117584228516, 'eval/latency_0/loss': 0.8654117584228516, 'eval/demon_fire_precision': 0.9050371593724195, 'eval/demon_fire_recall': 0.9102990033222591, 'eval/demon_fire_f1': 0.9076604554865425, 'eval/demon_fire_rate_model': 0.6735261401557285, 'eval/demon_fire_rate_teacher': 0.6696329254727474, 'eval/demon_left_precision': 0.8584905660377359, 'eval/demon_left_recall': 0.6186968838526912, 'eval/demon_left_f1': 0.719130721106355, 'eval/demon_left_rate_model': 0.35372636262513907, 'eval/demon_left_rate_teacher': 0.49082313681868744, 'eval/demon_right_precision': 0.6990500863557858, 'eval/demon_right_recall': 0.9146892655367231, 'eval/demon_right_f1': 0.7924620655898188, 'eval/demon_right_rate_model': 0.6440489432703004, 'eval/demon_right_rate_teacher': 0.4922135706340378, 'eval/demon_move_f1': 0.7557963933480869, 'eval/demon_cc_f1': 0.856057521002665, 'eval/latency_0/demon_fire_precision': 0.9050371593724195, 'eval/latency_0/demon_fire_recall': 0.9102990033222591, 'eval/latency_0/demon_fire_f1': 0.9076604554865425, 'eval/latency_0/demon_fire_rate_model': 0.6735261401557285, 'eval/latency_0/demon_fire_rate_teacher': 0.6696329254727474, 'eval/latency_0/demon_left_precision': 0.8584905660377359, 'eval/latency_0/demon_left_recall': 0.6186968838526912, 'eval/latency_0/demon_left_f1': 0.719130721106355, 'eval/latency_0/demon_left_rate_model': 0.35372636262513907, 'eval/latency_0/demon_left_rate_teacher': 0.49082313681868744, 'eval/latency_0/demon_right_precision': 0.6990500863557858, 'eval/latency_0/demon_right_recall': 0.9146892655367231, 'eval/latency_0/demon_right_f1': 0.7924620655898188, 'eval/latency_0/demon_right_rate_model': 0.6440489432703004, 'eval/latency_0/demon_right_rate_teacher': 0.4922135706340378, 'eval/latency_0/demon_move_f1': 0.7557963933480869, 'eval/latency_0/demon_cc_f1': 0.856057521002665, 'timing/data': 0.0003927110228687525, 'timing/model': 1.209377457969822, 'learning_rate/qwen_vl_interface': 7.809768520598986e-06, 'learning_rate/action_model': 7.809768520598984e-05, 'epoch': 0.68}) +[2026-06-17 10:26:29,298][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_2300_state +[2026-06-17 10:26:29,301][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 10:26:32,865][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_2300_state/model.safetensors +[2026-06-17 10:26:40,504][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_2300_state/optimizer.bin +[2026-06-17 10:26:40,504][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_2300_state/sampler.bin +[2026-06-17 10:26:40,504][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_2300_state/sampler_1.bin +[2026-06-17 10:26:40,505][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_2300_state/random_states_0.pkl +[2026-06-17 10:26:40,546][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2300 +[2026-06-17 10:26:40,546][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 10:26:40,562][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 10:26:45,181][starVLA.training.train_starvla][INFO] - Step 2301, Loss: {'train/loss': 0.9523672461509705, 'train/grad_norm_pre_clip': 9.722792625427246, 'timing/data': 0.0003336709924042225, 'timing/model': 1.1899701190413907, 'learning_rate/qwen_vl_interface': 7.80794646832486e-06, 'learning_rate/action_model': 7.80794646832486e-05, 'epoch': 0.68}) +[2026-06-17 10:26:49,823][starVLA.training.train_starvla][INFO] - Step 2302, Loss: {'train/loss': 1.1409708261489868, 'train/grad_norm_pre_clip': 14.846693992614746, 'timing/data': 0.0003429669886827469, 'timing/model': 1.190151880029589, 'learning_rate/qwen_vl_interface': 7.80612388578568e-06, 'learning_rate/action_model': 7.80612388578568e-05, 'epoch': 0.68}) +[2026-06-17 10:26:54,595][starVLA.training.train_starvla][INFO] - Step 2303, Loss: {'train/loss': 0.9493408203125, 'train/grad_norm_pre_clip': 10.305001258850098, 'timing/data': 0.00035344797652214766, 'timing/model': 1.192702918080613, 'learning_rate/qwen_vl_interface': 7.80430077335927e-06, 'learning_rate/action_model': 7.80430077335927e-05, 'epoch': 0.68}) +[2026-06-17 10:26:59,297][starVLA.training.train_starvla][INFO] - Step 2304, Loss: {'train/loss': 0.4865078926086426, 'train/grad_norm_pre_clip': 11.963685989379883, 'timing/data': 0.0004098119679838419, 'timing/model': 1.2957849899539724, 'learning_rate/qwen_vl_interface': 7.802477131423561e-06, 'learning_rate/action_model': 7.80247713142356e-05, 'epoch': 0.68}) +[2026-06-17 10:27:03,911][starVLA.training.train_starvla][INFO] - Step 2305, Loss: {'train/loss': 0.7113754153251648, 'train/grad_norm_pre_clip': 10.495001792907715, 'timing/data': 0.0003494719276204705, 'timing/model': 1.1954643499339, 'learning_rate/qwen_vl_interface': 7.800652960356594e-06, 'learning_rate/action_model': 7.800652960356594e-05, 'epoch': 0.68}) +[2026-06-17 10:27:08,662][starVLA.training.train_starvla][INFO] - Step 2306, Loss: {'train/loss': 0.7804191708564758, 'train/grad_norm_pre_clip': 9.556785583496094, 'timing/data': 0.000352138071320951, 'timing/model': 1.197967577027157, 'learning_rate/qwen_vl_interface': 7.798828260536524e-06, 'learning_rate/action_model': 7.798828260536524e-05, 'epoch': 0.68}) +[2026-06-17 10:27:13,421][starVLA.training.train_starvla][INFO] - Step 2307, Loss: {'train/loss': 0.6182055473327637, 'train/grad_norm_pre_clip': 7.872123718261719, 'timing/data': 0.00035999296233057976, 'timing/model': 1.198443248984404, 'learning_rate/qwen_vl_interface': 7.797003032341615e-06, 'learning_rate/action_model': 7.797003032341614e-05, 'epoch': 0.68}) +[2026-06-17 10:27:18,171][starVLA.training.train_starvla][INFO] - Step 2308, Loss: {'train/loss': 0.5540406703948975, 'train/grad_norm_pre_clip': 8.746525764465332, 'timing/data': 0.0003758709644898772, 'timing/model': 1.1985933351097628, 'learning_rate/qwen_vl_interface': 7.795177276150234e-06, 'learning_rate/action_model': 7.795177276150234e-05, 'epoch': 0.68}) +[2026-06-17 10:27:22,941][starVLA.training.train_starvla][INFO] - Step 2309, Loss: {'train/loss': 0.8209336400032043, 'train/grad_norm_pre_clip': 24.5798397064209, 'timing/data': 0.00036328507121652365, 'timing/model': 1.3381192570086569, 'learning_rate/qwen_vl_interface': 7.793350992340867e-06, 'learning_rate/action_model': 7.793350992340867e-05, 'epoch': 0.68}) +[2026-06-17 10:27:27,615][starVLA.training.train_starvla][INFO] - Step 2310, Loss: {'train/loss': 0.735865592956543, 'train/grad_norm_pre_clip': 8.866177558898926, 'timing/data': 0.00037555303424596786, 'timing/model': 1.2236137869767845, 'learning_rate/qwen_vl_interface': 7.791524181292102e-06, 'learning_rate/action_model': 7.791524181292101e-05, 'epoch': 0.68}) +[2026-06-17 10:27:32,277][starVLA.training.train_starvla][INFO] - Step 2311, Loss: {'train/loss': 0.8448610305786133, 'train/grad_norm_pre_clip': 9.505786895751953, 'timing/data': 0.0003789050970226526, 'timing/model': 1.207120870007202, 'learning_rate/qwen_vl_interface': 7.789696843382638e-06, 'learning_rate/action_model': 7.789696843382639e-05, 'epoch': 0.68}) +[2026-06-17 10:27:37,132][starVLA.training.train_starvla][INFO] - Step 2312, Loss: {'train/loss': 0.647711992263794, 'train/grad_norm_pre_clip': 8.548927307128906, 'timing/data': 0.0003997760359197855, 'timing/model': 1.2160661799134687, 'learning_rate/qwen_vl_interface': 7.787868978991288e-06, 'learning_rate/action_model': 7.787868978991287e-05, 'epoch': 0.68}) +[2026-06-17 10:27:41,870][starVLA.training.train_starvla][INFO] - Step 2313, Loss: {'train/loss': 0.8825403451919556, 'train/grad_norm_pre_clip': 12.64151382446289, 'timing/data': 0.00036399299278855324, 'timing/model': 1.2068970700493082, 'learning_rate/qwen_vl_interface': 7.786040588496966e-06, 'learning_rate/action_model': 7.786040588496965e-05, 'epoch': 0.68}) +[2026-06-17 10:27:46,630][starVLA.training.train_starvla][INFO] - Step 2314, Loss: {'train/loss': 0.8561556935310364, 'train/grad_norm_pre_clip': 10.861346244812012, 'timing/data': 0.0003432669909670949, 'timing/model': 1.2026465170783922, 'learning_rate/qwen_vl_interface': 7.784211672278703e-06, 'learning_rate/action_model': 7.784211672278703e-05, 'epoch': 0.68}) +[2026-06-17 10:27:51,409][starVLA.training.train_starvla][INFO] - Step 2315, Loss: {'train/loss': 0.9779662489891052, 'train/grad_norm_pre_clip': 9.812753677368164, 'timing/data': 0.00038091104943305254, 'timing/model': 1.3400232209824026, 'learning_rate/qwen_vl_interface': 7.78238223071563e-06, 'learning_rate/action_model': 7.78238223071563e-05, 'epoch': 0.68}) +[2026-06-17 10:27:56,073][starVLA.training.train_starvla][INFO] - Step 2316, Loss: {'train/loss': 0.8731325268745422, 'train/grad_norm_pre_clip': 9.63846492767334, 'timing/data': 0.0003646360710263252, 'timing/model': 1.2026795099955052, 'learning_rate/qwen_vl_interface': 7.780552264187e-06, 'learning_rate/action_model': 7.780552264186999e-05, 'epoch': 0.68}) +[2026-06-17 10:28:00,871][starVLA.training.train_starvla][INFO] - Step 2317, Loss: {'train/loss': 0.850110650062561, 'train/grad_norm_pre_clip': 13.620159149169922, 'timing/data': 0.00034171296283602715, 'timing/model': 1.2080087140202522, 'learning_rate/qwen_vl_interface': 7.77872177307216e-06, 'learning_rate/action_model': 7.778721773072159e-05, 'epoch': 0.68}) +[2026-06-17 10:28:05,642][starVLA.training.train_starvla][INFO] - Step 2318, Loss: {'train/loss': 0.7183846235275269, 'train/grad_norm_pre_clip': 12.027640342712402, 'timing/data': 0.0003566020168364048, 'timing/model': 1.200756017002277, 'learning_rate/qwen_vl_interface': 7.776890757750575e-06, 'learning_rate/action_model': 7.776890757750575e-05, 'epoch': 0.68}) +[2026-06-17 10:28:10,415][starVLA.training.train_starvla][INFO] - Step 2319, Loss: {'train/loss': 1.1643601655960083, 'train/grad_norm_pre_clip': 11.962287902832031, 'timing/data': 0.00034785293973982334, 'timing/model': 1.2011498459614813, 'learning_rate/qwen_vl_interface': 7.775059218601816e-06, 'learning_rate/action_model': 7.775059218601816e-05, 'epoch': 0.68}) +[2026-06-17 10:28:15,190][starVLA.training.train_starvla][INFO] - Step 2320, Loss: {'train/loss': 0.6130595207214355, 'train/grad_norm_pre_clip': 7.757535934448242, 'timing/data': 0.0003526640357449651, 'timing/model': 1.3294521800708026, 'learning_rate/qwen_vl_interface': 7.773227156005567e-06, 'learning_rate/action_model': 7.773227156005567e-05, 'epoch': 0.68}) +[2026-06-17 10:28:19,873][starVLA.training.train_starvla][INFO] - Step 2321, Loss: {'train/loss': 0.7774484157562256, 'train/grad_norm_pre_clip': 8.105762481689453, 'timing/data': 0.0003490080125629902, 'timing/model': 1.2360268189804628, 'learning_rate/qwen_vl_interface': 7.771394570341612e-06, 'learning_rate/action_model': 7.771394570341612e-05, 'epoch': 0.68}) +[2026-06-17 10:28:24,530][starVLA.training.train_starvla][INFO] - Step 2322, Loss: {'train/loss': 0.4617270231246948, 'train/grad_norm_pre_clip': 6.852250576019287, 'timing/data': 0.00036003394052386284, 'timing/model': 1.2134399589849636, 'learning_rate/qwen_vl_interface': 7.769561461989848e-06, 'learning_rate/action_model': 7.769561461989848e-05, 'epoch': 0.68}) +[2026-06-17 10:28:29,266][starVLA.training.train_starvla][INFO] - Step 2323, Loss: {'train/loss': 0.9710835814476013, 'train/grad_norm_pre_clip': 10.514381408691406, 'timing/data': 0.0004255051026120782, 'timing/model': 1.2051269670482725, 'learning_rate/qwen_vl_interface': 7.767727831330285e-06, 'learning_rate/action_model': 7.767727831330285e-05, 'epoch': 0.68}) +[2026-06-17 10:28:34,049][starVLA.training.train_starvla][INFO] - Step 2324, Loss: {'train/loss': 0.5904393196105957, 'train/grad_norm_pre_clip': 8.504678726196289, 'timing/data': 0.00037208106368780136, 'timing/model': 1.2016183719970286, 'learning_rate/qwen_vl_interface': 7.765893678743029e-06, 'learning_rate/action_model': 7.76589367874303e-05, 'epoch': 0.68}) +[2026-06-17 10:28:38,850][starVLA.training.train_starvla][INFO] - Step 2325, Loss: {'train/loss': 0.6091851592063904, 'train/grad_norm_pre_clip': 7.103052616119385, 'timing/data': 0.00040805793832987547, 'timing/model': 1.2040371880866587, 'learning_rate/qwen_vl_interface': 7.76405900460831e-06, 'learning_rate/action_model': 7.76405900460831e-05, 'epoch': 0.68}) +[2026-06-17 10:28:43,637][starVLA.training.train_starvla][INFO] - Step 2326, Loss: {'train/loss': 0.7081632018089294, 'train/grad_norm_pre_clip': 13.684537887573242, 'timing/data': 0.00033482094295322895, 'timing/model': 1.3361781650455669, 'learning_rate/qwen_vl_interface': 7.762223809306453e-06, 'learning_rate/action_model': 7.762223809306452e-05, 'epoch': 0.68}) +[2026-06-17 10:28:48,292][starVLA.training.train_starvla][INFO] - Step 2327, Loss: {'train/loss': 0.8154391646385193, 'train/grad_norm_pre_clip': 8.561654090881348, 'timing/data': 0.0003782670246437192, 'timing/model': 1.2057228479534388, 'learning_rate/qwen_vl_interface': 7.760388093217897e-06, 'learning_rate/action_model': 7.760388093217898e-05, 'epoch': 0.68}) +[2026-06-17 10:28:53,078][starVLA.training.train_starvla][INFO] - Step 2328, Loss: {'train/loss': 0.8192497491836548, 'train/grad_norm_pre_clip': 10.1256742477417, 'timing/data': 0.00034075905568897724, 'timing/model': 1.203869380056858, 'learning_rate/qwen_vl_interface': 7.758551856723187e-06, 'learning_rate/action_model': 7.758551856723188e-05, 'epoch': 0.69}) +[2026-06-17 10:28:57,868][starVLA.training.train_starvla][INFO] - Step 2329, Loss: {'train/loss': 0.9689756631851196, 'train/grad_norm_pre_clip': 11.505919456481934, 'timing/data': 0.0004196629161015153, 'timing/model': 1.2092433900106698, 'learning_rate/qwen_vl_interface': 7.756715100202979e-06, 'learning_rate/action_model': 7.75671510020298e-05, 'epoch': 0.69}) +[2026-06-17 10:29:02,677][starVLA.training.train_starvla][INFO] - Step 2330, Loss: {'train/loss': 0.6823896765708923, 'train/grad_norm_pre_clip': 13.103119850158691, 'timing/data': 0.00035807001404464245, 'timing/model': 1.2108087559463456, 'learning_rate/qwen_vl_interface': 7.754877824038035e-06, 'learning_rate/action_model': 7.754877824038035e-05, 'epoch': 0.69}) +[2026-06-17 10:29:07,488][starVLA.training.train_starvla][INFO] - Step 2331, Loss: {'train/loss': 1.0856136083602905, 'train/grad_norm_pre_clip': 12.635562896728516, 'timing/data': 0.0004058149643242359, 'timing/model': 1.2090291529893875, 'learning_rate/qwen_vl_interface': 7.75304002860922e-06, 'learning_rate/action_model': 7.75304002860922e-05, 'epoch': 0.69}) +[2026-06-17 10:29:12,291][starVLA.training.train_starvla][INFO] - Step 2332, Loss: {'train/loss': 0.9206007122993469, 'train/grad_norm_pre_clip': 10.589554786682129, 'timing/data': 0.000392664922401309, 'timing/model': 1.3422598610632122, 'learning_rate/qwen_vl_interface': 7.751201714297513e-06, 'learning_rate/action_model': 7.751201714297512e-05, 'epoch': 0.69}) +[2026-06-17 10:29:16,961][starVLA.training.train_starvla][INFO] - Step 2333, Loss: {'train/loss': 0.7900664806365967, 'train/grad_norm_pre_clip': 10.734641075134277, 'timing/data': 0.0003616489702835679, 'timing/model': 1.207503359997645, 'learning_rate/qwen_vl_interface': 7.749362881484002e-06, 'learning_rate/action_model': 7.749362881484001e-05, 'epoch': 0.69}) +[2026-06-17 10:29:21,742][starVLA.training.train_starvla][INFO] - Step 2334, Loss: {'train/loss': 0.7120086550712585, 'train/grad_norm_pre_clip': 31.089784622192383, 'timing/data': 0.00035787501838058233, 'timing/model': 1.2060387070523575, 'learning_rate/qwen_vl_interface': 7.747523530549874e-06, 'learning_rate/action_model': 7.747523530549875e-05, 'epoch': 0.69}) +[2026-06-17 10:29:26,520][starVLA.training.train_starvla][INFO] - Step 2335, Loss: {'train/loss': 1.074162483215332, 'train/grad_norm_pre_clip': 11.407490730285645, 'timing/data': 0.0003359660040587187, 'timing/model': 1.2032108870334923, 'learning_rate/qwen_vl_interface': 7.745683661876432e-06, 'learning_rate/action_model': 7.745683661876432e-05, 'epoch': 0.69}) +[2026-06-17 10:29:31,296][starVLA.training.train_starvla][INFO] - Step 2336, Loss: {'train/loss': 0.8128191828727722, 'train/grad_norm_pre_clip': 8.260025978088379, 'timing/data': 0.0003656459739431739, 'timing/model': 1.203164318925701, 'learning_rate/qwen_vl_interface': 7.743843275845081e-06, 'learning_rate/action_model': 7.74384327584508e-05, 'epoch': 0.69}) +[2026-06-17 10:29:36,024][starVLA.training.train_starvla][INFO] - Step 2337, Loss: {'train/loss': 0.8120927810668945, 'train/grad_norm_pre_clip': 8.362366676330566, 'timing/data': 0.00034415104892104864, 'timing/model': 1.2024758990155533, 'learning_rate/qwen_vl_interface': 7.742002372837336e-06, 'learning_rate/action_model': 7.742002372837335e-05, 'epoch': 0.69}) +[2026-06-17 10:29:40,862][starVLA.training.train_starvla][INFO] - Step 2338, Loss: {'train/loss': 0.6819133162498474, 'train/grad_norm_pre_clip': 9.575437545776367, 'timing/data': 0.0005575959803536534, 'timing/model': 1.350625657942146, 'learning_rate/qwen_vl_interface': 7.740160953234815e-06, 'learning_rate/action_model': 7.740160953234815e-05, 'epoch': 0.69}) +[2026-06-17 10:29:45,526][starVLA.training.train_starvla][INFO] - Step 2339, Loss: {'train/loss': 0.8783963322639465, 'train/grad_norm_pre_clip': 9.33806324005127, 'timing/data': 0.0006509320810437202, 'timing/model': 1.2084480030462146, 'learning_rate/qwen_vl_interface': 7.738319017419252e-06, 'learning_rate/action_model': 7.738319017419252e-05, 'epoch': 0.69}) +[2026-06-17 10:29:50,341][starVLA.training.train_starvla][INFO] - Step 2340, Loss: {'train/loss': 0.8046959042549133, 'train/grad_norm_pre_clip': 10.916419982910156, 'timing/data': 0.0003670579753816128, 'timing/model': 1.2070053030038252, 'learning_rate/qwen_vl_interface': 7.736476565772478e-06, 'learning_rate/action_model': 7.736476565772478e-05, 'epoch': 0.69}) +[2026-06-17 10:29:55,166][starVLA.training.train_starvla][INFO] - Step 2341, Loss: {'train/loss': 0.903343915939331, 'train/grad_norm_pre_clip': 9.452247619628906, 'timing/data': 0.00033784995321184397, 'timing/model': 1.2033390899887308, 'learning_rate/qwen_vl_interface': 7.734633598676435e-06, 'learning_rate/action_model': 7.734633598676434e-05, 'epoch': 0.69}) +[2026-06-17 10:29:59,932][starVLA.training.train_starvla][INFO] - Step 2342, Loss: {'train/loss': 0.7792141437530518, 'train/grad_norm_pre_clip': 8.82003402709961, 'timing/data': 0.00038483203388750553, 'timing/model': 1.202835363917984, 'learning_rate/qwen_vl_interface': 7.732790116513176e-06, 'learning_rate/action_model': 7.732790116513176e-05, 'epoch': 0.69}) +[2026-06-17 10:30:04,719][starVLA.training.train_starvla][INFO] - Step 2343, Loss: {'train/loss': 0.6859004497528076, 'train/grad_norm_pre_clip': 7.136237621307373, 'timing/data': 0.00040486291982233524, 'timing/model': 1.2062821970321238, 'learning_rate/qwen_vl_interface': 7.730946119664852e-06, 'learning_rate/action_model': 7.730946119664853e-05, 'epoch': 0.69}) +[2026-06-17 10:30:09,523][starVLA.training.train_starvla][INFO] - Step 2344, Loss: {'train/loss': 0.6983392238616943, 'train/grad_norm_pre_clip': 8.21456241607666, 'timing/data': 0.0005024860147386789, 'timing/model': 1.335872262949124, 'learning_rate/qwen_vl_interface': 7.729101608513729e-06, 'learning_rate/action_model': 7.72910160851373e-05, 'epoch': 0.69}) +[2026-06-17 10:30:14,188][starVLA.training.train_starvla][INFO] - Step 2345, Loss: {'train/loss': 1.0932866334915161, 'train/grad_norm_pre_clip': 11.458840370178223, 'timing/data': 0.0003566339146345854, 'timing/model': 1.2079159950371832, 'learning_rate/qwen_vl_interface': 7.727256583442175e-06, 'learning_rate/action_model': 7.727256583442175e-05, 'epoch': 0.69}) +[2026-06-17 10:30:18,987][starVLA.training.train_starvla][INFO] - Step 2346, Loss: {'train/loss': 0.5160951614379883, 'train/grad_norm_pre_clip': 6.38591194152832, 'timing/data': 0.0003462120657786727, 'timing/model': 1.2045453299069777, 'learning_rate/qwen_vl_interface': 7.725411044832665e-06, 'learning_rate/action_model': 7.725411044832665e-05, 'epoch': 0.69}) +[2026-06-17 10:30:23,780][starVLA.training.train_starvla][INFO] - Step 2347, Loss: {'train/loss': 0.5510313510894775, 'train/grad_norm_pre_clip': 10.526773452758789, 'timing/data': 0.000359266996383667, 'timing/model': 1.2045791780110449, 'learning_rate/qwen_vl_interface': 7.723564993067783e-06, 'learning_rate/action_model': 7.723564993067783e-05, 'epoch': 0.69}) +[2026-06-17 10:30:28,603][starVLA.training.train_starvla][INFO] - Step 2348, Loss: {'train/loss': 0.8536263108253479, 'train/grad_norm_pre_clip': 8.755084037780762, 'timing/data': 0.0003749489551410079, 'timing/model': 1.2113481969572604, 'learning_rate/qwen_vl_interface': 7.721718428530214e-06, 'learning_rate/action_model': 7.721718428530214e-05, 'epoch': 0.69}) +[2026-06-17 10:30:33,390][starVLA.training.train_starvla][INFO] - Step 2349, Loss: {'train/loss': 0.772020161151886, 'train/grad_norm_pre_clip': 7.935525894165039, 'timing/data': 0.00038537103682756424, 'timing/model': 1.3315594400046393, 'learning_rate/qwen_vl_interface': 7.719871351602757e-06, 'learning_rate/action_model': 7.719871351602757e-05, 'epoch': 0.69}) +[2026-06-17 10:30:38,095][starVLA.training.train_starvla][INFO] - Step 2350, Loss: {'train/loss': 0.8374469876289368, 'train/grad_norm_pre_clip': 11.738473892211914, 'timing/data': 0.00040727993473410606, 'timing/model': 1.2468769699335098, 'learning_rate/qwen_vl_interface': 7.718023762668309e-06, 'learning_rate/action_model': 7.718023762668309e-05, 'epoch': 0.69}) +[2026-06-17 10:30:42,796][starVLA.training.train_starvla][INFO] - Step 2351, Loss: {'train/loss': 0.7087457776069641, 'train/grad_norm_pre_clip': 8.845320701599121, 'timing/data': 0.000371222966350615, 'timing/model': 1.2087173459585756, 'learning_rate/qwen_vl_interface': 7.716175662109882e-06, 'learning_rate/action_model': 7.716175662109881e-05, 'epoch': 0.69}) +[2026-06-17 10:30:47,573][starVLA.training.train_starvla][INFO] - Step 2352, Loss: {'train/loss': 0.8018172979354858, 'train/grad_norm_pre_clip': 7.504945755004883, 'timing/data': 0.00035652797669172287, 'timing/model': 1.2137869400903583, 'learning_rate/qwen_vl_interface': 7.714327050310586e-06, 'learning_rate/action_model': 7.714327050310585e-05, 'epoch': 0.69}) +[2026-06-17 10:30:52,403][starVLA.training.train_starvla][INFO] - Step 2353, Loss: {'train/loss': 1.015899419784546, 'train/grad_norm_pre_clip': 11.772109031677246, 'timing/data': 0.0003625020617619157, 'timing/model': 1.211315362015739, 'learning_rate/qwen_vl_interface': 7.71247792765364e-06, 'learning_rate/action_model': 7.712477927653638e-05, 'epoch': 0.69}) +[2026-06-17 10:30:57,200][starVLA.training.train_starvla][INFO] - Step 2354, Loss: {'train/loss': 0.7922556400299072, 'train/grad_norm_pre_clip': 9.32275676727295, 'timing/data': 0.0003507420187816024, 'timing/model': 1.2066000940976664, 'learning_rate/qwen_vl_interface': 7.71062829452237e-06, 'learning_rate/action_model': 7.710628294522369e-05, 'epoch': 0.69}) +[2026-06-17 10:31:02,014][starVLA.training.train_starvla][INFO] - Step 2355, Loss: {'train/loss': 0.7081893682479858, 'train/grad_norm_pre_clip': 12.958463668823242, 'timing/data': 0.0003970089601352811, 'timing/model': 1.350669376901351, 'learning_rate/qwen_vl_interface': 7.708778151300206e-06, 'learning_rate/action_model': 7.708778151300206e-05, 'epoch': 0.69}) +[2026-06-17 10:31:06,699][starVLA.training.train_starvla][INFO] - Step 2356, Loss: {'train/loss': 0.7432347536087036, 'train/grad_norm_pre_clip': 10.397196769714355, 'timing/data': 0.00034615001641213894, 'timing/model': 1.2099027970107272, 'learning_rate/qwen_vl_interface': 7.706927498370687e-06, 'learning_rate/action_model': 7.706927498370686e-05, 'epoch': 0.69}) +[2026-06-17 10:31:11,505][starVLA.training.train_starvla][INFO] - Step 2357, Loss: {'train/loss': 0.8440070152282715, 'train/grad_norm_pre_clip': 9.849210739135742, 'timing/data': 0.000689313979819417, 'timing/model': 1.211441042018123, 'learning_rate/qwen_vl_interface': 7.705076336117452e-06, 'learning_rate/action_model': 7.705076336117451e-05, 'epoch': 0.69}) +[2026-06-17 10:31:16,320][starVLA.training.train_starvla][INFO] - Step 2358, Loss: {'train/loss': 0.7870534062385559, 'train/grad_norm_pre_clip': 10.440204620361328, 'timing/data': 0.0005818079225718975, 'timing/model': 1.2071843380108476, 'learning_rate/qwen_vl_interface': 7.703224664924252e-06, 'learning_rate/action_model': 7.703224664924251e-05, 'epoch': 0.69}) +[2026-06-17 10:31:21,119][starVLA.training.train_starvla][INFO] - Step 2359, Loss: {'train/loss': 1.25443696975708, 'train/grad_norm_pre_clip': 14.190240859985352, 'timing/data': 0.0004140749806538224, 'timing/model': 1.2098445020383224, 'learning_rate/qwen_vl_interface': 7.70137248517494e-06, 'learning_rate/action_model': 7.701372485174939e-05, 'epoch': 0.69}) +[2026-06-17 10:31:25,912][starVLA.training.train_starvla][INFO] - Step 2360, Loss: {'train/loss': 0.9117757081985474, 'train/grad_norm_pre_clip': 10.433684349060059, 'timing/data': 0.00037471007090061903, 'timing/model': 1.2106649848865345, 'learning_rate/qwen_vl_interface': 7.699519797253471e-06, 'learning_rate/action_model': 7.699519797253472e-05, 'epoch': 0.69}) +[2026-06-17 10:31:30,741][starVLA.training.train_starvla][INFO] - Step 2361, Loss: {'train/loss': 0.8335483074188232, 'train/grad_norm_pre_clip': 9.506421089172363, 'timing/data': 0.00040547503158450127, 'timing/model': 1.3451350119430572, 'learning_rate/qwen_vl_interface': 7.697666601543915e-06, 'learning_rate/action_model': 7.697666601543915e-05, 'epoch': 0.69}) +[2026-06-17 10:31:35,431][starVLA.training.train_starvla][INFO] - Step 2362, Loss: {'train/loss': 0.9469690322875977, 'train/grad_norm_pre_clip': 8.90514850616455, 'timing/data': 0.00035758502781391144, 'timing/model': 1.2278303808998317, 'learning_rate/qwen_vl_interface': 7.695812898430438e-06, 'learning_rate/action_model': 7.695812898430437e-05, 'epoch': 0.7}) +[2026-06-17 10:31:40,237][starVLA.training.train_starvla][INFO] - Step 2363, Loss: {'train/loss': 0.8523452281951904, 'train/grad_norm_pre_clip': 9.271471977233887, 'timing/data': 0.0005431370809674263, 'timing/model': 1.2063492520246655, 'learning_rate/qwen_vl_interface': 7.693958688297316e-06, 'learning_rate/action_model': 7.693958688297314e-05, 'epoch': 0.7}) +[2026-06-17 10:31:45,035][starVLA.training.train_starvla][INFO] - Step 2364, Loss: {'train/loss': 0.656302273273468, 'train/grad_norm_pre_clip': 8.684602737426758, 'timing/data': 0.00038381910417228937, 'timing/model': 1.2114099250175059, 'learning_rate/qwen_vl_interface': 7.692103971528925e-06, 'learning_rate/action_model': 7.692103971528926e-05, 'epoch': 0.7}) +[2026-06-17 10:31:49,827][starVLA.training.train_starvla][INFO] - Step 2365, Loss: {'train/loss': 0.8235297203063965, 'train/grad_norm_pre_clip': 10.406444549560547, 'timing/data': 0.00037449400406330824, 'timing/model': 1.2095056099351496, 'learning_rate/qwen_vl_interface': 7.690248748509757e-06, 'learning_rate/action_model': 7.690248748509757e-05, 'epoch': 0.7}) +[2026-06-17 10:31:54,575][starVLA.training.train_starvla][INFO] - Step 2366, Loss: {'train/loss': 0.6489296555519104, 'train/grad_norm_pre_clip': 8.133014678955078, 'timing/data': 0.00032888201531022787, 'timing/model': 1.2059077490121126, 'learning_rate/qwen_vl_interface': 7.688393019624396e-06, 'learning_rate/action_model': 7.688393019624395e-05, 'epoch': 0.7}) +[2026-06-17 10:31:59,363][starVLA.training.train_starvla][INFO] - Step 2367, Loss: {'train/loss': 0.6077809929847717, 'train/grad_norm_pre_clip': 7.246346950531006, 'timing/data': 0.000602290965616703, 'timing/model': 1.3363473580684513, 'learning_rate/qwen_vl_interface': 7.686536785257539e-06, 'learning_rate/action_model': 7.686536785257538e-05, 'epoch': 0.7}) +[2026-06-17 10:32:04,052][starVLA.training.train_starvla][INFO] - Step 2368, Loss: {'train/loss': 0.9561675190925598, 'train/grad_norm_pre_clip': 7.953044414520264, 'timing/data': 0.0005302600329741836, 'timing/model': 1.2195632939692587, 'learning_rate/qwen_vl_interface': 7.684680045793982e-06, 'learning_rate/action_model': 7.684680045793983e-05, 'epoch': 0.7}) +[2026-06-17 10:32:08,849][starVLA.training.train_starvla][INFO] - Step 2369, Loss: {'train/loss': 0.656087338924408, 'train/grad_norm_pre_clip': 10.705005645751953, 'timing/data': 0.0003653020830824971, 'timing/model': 1.2047685390571132, 'learning_rate/qwen_vl_interface': 7.682822801618634e-06, 'learning_rate/action_model': 7.682822801618633e-05, 'epoch': 0.7}) +[2026-06-17 10:32:13,690][starVLA.training.train_starvla][INFO] - Step 2370, Loss: {'train/loss': 0.7616079449653625, 'train/grad_norm_pre_clip': 9.266800880432129, 'timing/data': 0.00041253201197832823, 'timing/model': 1.218384578009136, 'learning_rate/qwen_vl_interface': 7.680965053116498e-06, 'learning_rate/action_model': 7.680965053116499e-05, 'epoch': 0.7}) +[2026-06-17 10:32:18,478][starVLA.training.train_starvla][INFO] - Step 2371, Loss: {'train/loss': 0.7715666890144348, 'train/grad_norm_pre_clip': 9.126121520996094, 'timing/data': 0.00035744206979870796, 'timing/model': 1.2112283930182457, 'learning_rate/qwen_vl_interface': 7.679106800672695e-06, 'learning_rate/action_model': 7.679106800672695e-05, 'epoch': 0.7}) +[2026-06-17 10:32:23,235][starVLA.training.train_starvla][INFO] - Step 2372, Loss: {'train/loss': 0.6937838792800903, 'train/grad_norm_pre_clip': 8.369319915771484, 'timing/data': 0.0005934069631621242, 'timing/model': 1.202539889025502, 'learning_rate/qwen_vl_interface': 7.677248044672438e-06, 'learning_rate/action_model': 7.677248044672437e-05, 'epoch': 0.7}) +[2026-06-17 10:32:28,043][starVLA.training.train_starvla][INFO] - Step 2373, Loss: {'train/loss': 0.6933718919754028, 'train/grad_norm_pre_clip': 7.862736701965332, 'timing/data': 0.0003506829962134361, 'timing/model': 1.340676540043205, 'learning_rate/qwen_vl_interface': 7.675388785501048e-06, 'learning_rate/action_model': 7.675388785501047e-05, 'epoch': 0.7}) +[2026-06-17 10:32:32,702][starVLA.training.train_starvla][INFO] - Step 2374, Loss: {'train/loss': 0.7114582657814026, 'train/grad_norm_pre_clip': 8.079758644104004, 'timing/data': 0.00041643402073532343, 'timing/model': 1.2074277009814978, 'learning_rate/qwen_vl_interface': 7.673529023543951e-06, 'learning_rate/action_model': 7.673529023543951e-05, 'epoch': 0.7}) +[2026-06-17 10:32:37,488][starVLA.training.train_starvla][INFO] - Step 2375, Loss: {'train/loss': 0.777788519859314, 'train/grad_norm_pre_clip': 12.011693954467773, 'timing/data': 0.0003415449755266309, 'timing/model': 1.2032185440184548, 'learning_rate/qwen_vl_interface': 7.671668759186683e-06, 'learning_rate/action_model': 7.671668759186683e-05, 'epoch': 0.7}) +[2026-06-17 10:32:42,272][starVLA.training.train_starvla][INFO] - Step 2376, Loss: {'train/loss': 0.6711933612823486, 'train/grad_norm_pre_clip': 8.202439308166504, 'timing/data': 0.00042109296191483736, 'timing/model': 1.2051324569620192, 'learning_rate/qwen_vl_interface': 7.669807992814875e-06, 'learning_rate/action_model': 7.669807992814875e-05, 'epoch': 0.7}) +[2026-06-17 10:32:47,059][starVLA.training.train_starvla][INFO] - Step 2377, Loss: {'train/loss': 0.7388006448745728, 'train/grad_norm_pre_clip': 8.27390193939209, 'timing/data': 0.00035968399606645107, 'timing/model': 1.2042085640132427, 'learning_rate/qwen_vl_interface': 7.667946724814264e-06, 'learning_rate/action_model': 7.667946724814265e-05, 'epoch': 0.7}) +[2026-06-17 10:32:51,922][starVLA.training.train_starvla][INFO] - Step 2378, Loss: {'train/loss': 0.42261213064193726, 'train/grad_norm_pre_clip': 6.149807929992676, 'timing/data': 0.0003611700376495719, 'timing/model': 1.3379411839414388, 'learning_rate/qwen_vl_interface': 7.666084955570698e-06, 'learning_rate/action_model': 7.666084955570698e-05, 'epoch': 0.7}) +[2026-06-17 10:32:56,628][starVLA.training.train_starvla][INFO] - Step 2379, Loss: {'train/loss': 0.9787164330482483, 'train/grad_norm_pre_clip': 19.39433479309082, 'timing/data': 0.0003323199925944209, 'timing/model': 1.2423514650436118, 'learning_rate/qwen_vl_interface': 7.664222685470119e-06, 'learning_rate/action_model': 7.664222685470118e-05, 'epoch': 0.7}) +[2026-06-17 10:33:01,285][starVLA.training.train_starvla][INFO] - Step 2380, Loss: {'train/loss': 0.9165060520172119, 'train/grad_norm_pre_clip': 11.16001033782959, 'timing/data': 0.000391694949939847, 'timing/model': 1.205709294998087, 'learning_rate/qwen_vl_interface': 7.66235991489858e-06, 'learning_rate/action_model': 7.66235991489858e-05, 'epoch': 0.7}) +[2026-06-17 10:33:06,035][starVLA.training.train_starvla][INFO] - Step 2381, Loss: {'train/loss': 0.8236847519874573, 'train/grad_norm_pre_clip': 11.447330474853516, 'timing/data': 0.0003831640351563692, 'timing/model': 1.2052246188977733, 'learning_rate/qwen_vl_interface': 7.660496644242233e-06, 'learning_rate/action_model': 7.660496644242233e-05, 'epoch': 0.7}) +[2026-06-17 10:33:10,861][starVLA.training.train_starvla][INFO] - Step 2382, Loss: {'train/loss': 0.6752314567565918, 'train/grad_norm_pre_clip': 8.038362503051758, 'timing/data': 0.0003939169691875577, 'timing/model': 1.2210417430615053, 'learning_rate/qwen_vl_interface': 7.658632873887342e-06, 'learning_rate/action_model': 7.658632873887342e-05, 'epoch': 0.7}) +[2026-06-17 10:33:15,683][starVLA.training.train_starvla][INFO] - Step 2383, Loss: {'train/loss': 0.7991492748260498, 'train/grad_norm_pre_clip': 8.619221687316895, 'timing/data': 0.000372720998711884, 'timing/model': 1.203946843976155, 'learning_rate/qwen_vl_interface': 7.656768604220262e-06, 'learning_rate/action_model': 7.656768604220261e-05, 'epoch': 0.7}) +[2026-06-17 10:33:20,466][starVLA.training.train_starvla][INFO] - Step 2384, Loss: {'train/loss': 0.7746067643165588, 'train/grad_norm_pre_clip': 9.249911308288574, 'timing/data': 0.000402023084461689, 'timing/model': 1.33297876897268, 'learning_rate/qwen_vl_interface': 7.654903835627459e-06, 'learning_rate/action_model': 7.654903835627458e-05, 'epoch': 0.7}) +[2026-06-17 10:33:25,128][starVLA.training.train_starvla][INFO] - Step 2385, Loss: {'train/loss': 0.6487160921096802, 'train/grad_norm_pre_clip': 7.8397297859191895, 'timing/data': 0.0003716719802469015, 'timing/model': 1.2059082679916173, 'learning_rate/qwen_vl_interface': 7.653038568495505e-06, 'learning_rate/action_model': 7.653038568495506e-05, 'epoch': 0.7}) +[2026-06-17 10:33:29,989][starVLA.training.train_starvla][INFO] - Step 2386, Loss: {'train/loss': 0.7711982727050781, 'train/grad_norm_pre_clip': 9.994614601135254, 'timing/data': 0.0003402730217203498, 'timing/model': 1.2031901450827718, 'learning_rate/qwen_vl_interface': 7.651172803211068e-06, 'learning_rate/action_model': 7.651172803211068e-05, 'epoch': 0.7}) +[2026-06-17 10:33:34,778][starVLA.training.train_starvla][INFO] - Step 2387, Loss: {'train/loss': 0.7283377051353455, 'train/grad_norm_pre_clip': 12.161765098571777, 'timing/data': 0.0003649418940767646, 'timing/model': 1.2056103060021996, 'learning_rate/qwen_vl_interface': 7.649306540160928e-06, 'learning_rate/action_model': 7.649306540160927e-05, 'epoch': 0.7}) +[2026-06-17 10:33:39,555][starVLA.training.train_starvla][INFO] - Step 2388, Loss: {'train/loss': 0.6888020038604736, 'train/grad_norm_pre_clip': 11.32221794128418, 'timing/data': 0.0003904229961335659, 'timing/model': 1.2025609519332647, 'learning_rate/qwen_vl_interface': 7.647439779731959e-06, 'learning_rate/action_model': 7.647439779731958e-05, 'epoch': 0.7}) +[2026-06-17 10:33:44,343][starVLA.training.train_starvla][INFO] - Step 2389, Loss: {'train/loss': 0.8371812105178833, 'train/grad_norm_pre_clip': 9.313901901245117, 'timing/data': 0.0003768230089917779, 'timing/model': 1.202846649917774, 'learning_rate/qwen_vl_interface': 7.645572522311144e-06, 'learning_rate/action_model': 7.645572522311144e-05, 'epoch': 0.7}) +[2026-06-17 10:33:49,126][starVLA.training.train_starvla][INFO] - Step 2390, Loss: {'train/loss': 1.0194743871688843, 'train/grad_norm_pre_clip': 12.358183860778809, 'timing/data': 0.0005163030000403523, 'timing/model': 1.3348755639744923, 'learning_rate/qwen_vl_interface': 7.643704768285568e-06, 'learning_rate/action_model': 7.643704768285568e-05, 'epoch': 0.7}) +[2026-06-17 10:33:53,781][starVLA.training.train_starvla][INFO] - Step 2391, Loss: {'train/loss': 0.7106689810752869, 'train/grad_norm_pre_clip': 11.009908676147461, 'timing/data': 0.0003410410135984421, 'timing/model': 1.2054436870384961, 'learning_rate/qwen_vl_interface': 7.641836518042417e-06, 'learning_rate/action_model': 7.641836518042416e-05, 'epoch': 0.7}) +[2026-06-17 10:33:58,559][starVLA.training.train_starvla][INFO] - Step 2392, Loss: {'train/loss': 0.7802167534828186, 'train/grad_norm_pre_clip': 9.88891315460205, 'timing/data': 0.00042320601642131805, 'timing/model': 1.2026694330852479, 'learning_rate/qwen_vl_interface': 7.639967771968983e-06, 'learning_rate/action_model': 7.639967771968983e-05, 'epoch': 0.7}) +[2026-06-17 10:34:03,333][starVLA.training.train_starvla][INFO] - Step 2393, Loss: {'train/loss': 0.8543564677238464, 'train/grad_norm_pre_clip': 10.618596076965332, 'timing/data': 0.0003695129416882992, 'timing/model': 1.2043654429726303, 'learning_rate/qwen_vl_interface': 7.638098530452657e-06, 'learning_rate/action_model': 7.638098530452658e-05, 'epoch': 0.7}) +[2026-06-17 10:34:08,109][starVLA.training.train_starvla][INFO] - Step 2394, Loss: {'train/loss': 0.6430135369300842, 'train/grad_norm_pre_clip': 8.381234169006348, 'timing/data': 0.00037259701639413834, 'timing/model': 1.205602523055859, 'learning_rate/qwen_vl_interface': 7.636228793880935e-06, 'learning_rate/action_model': 7.636228793880935e-05, 'epoch': 0.7}) +[2026-06-17 10:34:12,851][starVLA.training.train_starvla][INFO] - Step 2395, Loss: {'train/loss': 0.5455282330513, 'train/grad_norm_pre_clip': 8.131120681762695, 'timing/data': 0.0003763770218938589, 'timing/model': 1.2059512819396332, 'learning_rate/qwen_vl_interface': 7.63435856264142e-06, 'learning_rate/action_model': 7.63435856264142e-05, 'epoch': 0.7}) +[2026-06-17 10:34:17,668][starVLA.training.train_starvla][INFO] - Step 2396, Loss: {'train/loss': 0.8713265061378479, 'train/grad_norm_pre_clip': 10.855679512023926, 'timing/data': 0.00040495104622095823, 'timing/model': 1.3457292839884758, 'learning_rate/qwen_vl_interface': 7.632487837121806e-06, 'learning_rate/action_model': 7.632487837121806e-05, 'epoch': 0.71}) +[2026-06-17 10:34:22,362][starVLA.training.train_starvla][INFO] - Step 2397, Loss: {'train/loss': 0.5863310098648071, 'train/grad_norm_pre_clip': 6.967587471008301, 'timing/data': 0.000731050968170166, 'timing/model': 1.2197588720591739, 'learning_rate/qwen_vl_interface': 7.630616617709902e-06, 'learning_rate/action_model': 7.630616617709902e-05, 'epoch': 0.71}) +[2026-06-17 10:34:27,165][starVLA.training.train_starvla][INFO] - Step 2398, Loss: {'train/loss': 1.3306210041046143, 'train/grad_norm_pre_clip': 10.934465408325195, 'timing/data': 0.0003864630125463009, 'timing/model': 1.2194042350165546, 'learning_rate/qwen_vl_interface': 7.6287449047936095e-06, 'learning_rate/action_model': 7.628744904793609e-05, 'epoch': 0.71}) +[2026-06-17 10:34:31,973][starVLA.training.train_starvla][INFO] - Step 2399, Loss: {'train/loss': 1.0088317394256592, 'train/grad_norm_pre_clip': 10.751895904541016, 'timing/data': 0.0003224499523639679, 'timing/model': 1.2041255199583247, 'learning_rate/qwen_vl_interface': 7.6268726987609384e-06, 'learning_rate/action_model': 7.626872698760939e-05, 'epoch': 0.71}) +[2026-06-17 10:36:13,570][starVLA.training.train_starvla][INFO] - Step 2400, Loss: {'train/loss': 0.8692829012870789, 'train/grad_norm_pre_clip': 9.165996551513672, 'eval/loss': 0.8212173581123352, 'eval/latency_0/loss': 0.8212173581123352, 'eval/demon_fire_precision': 0.888976069046685, 'eval/demon_fire_recall': 0.9410299003322259, 'eval/demon_fire_f1': 0.9142626588662499, 'eval/demon_fire_rate_model': 0.7088431590656284, 'eval/demon_fire_rate_teacher': 0.6696329254727474, 'eval/demon_left_precision': 0.849082256968049, 'eval/demon_left_recall': 0.7076487252124646, 'eval/demon_left_f1': 0.7719406674907292, 'eval/demon_left_rate_model': 0.4090656284760845, 'eval/demon_left_rate_teacher': 0.49082313681868744, 'eval/demon_right_precision': 0.7435294117647059, 'eval/demon_right_recall': 0.8926553672316384, 'eval/demon_right_f1': 0.8112965340179716, 'eval/demon_right_rate_model': 0.5909343715239155, 'eval/demon_right_rate_teacher': 0.4922135706340378, 'eval/demon_move_f1': 0.7916186007543504, 'eval/demon_cc_f1': 0.8736564482104902, 'eval/latency_0/demon_fire_precision': 0.888976069046685, 'eval/latency_0/demon_fire_recall': 0.9410299003322259, 'eval/latency_0/demon_fire_f1': 0.9142626588662499, 'eval/latency_0/demon_fire_rate_model': 0.7088431590656284, 'eval/latency_0/demon_fire_rate_teacher': 0.6696329254727474, 'eval/latency_0/demon_left_precision': 0.849082256968049, 'eval/latency_0/demon_left_recall': 0.7076487252124646, 'eval/latency_0/demon_left_f1': 0.7719406674907292, 'eval/latency_0/demon_left_rate_model': 0.4090656284760845, 'eval/latency_0/demon_left_rate_teacher': 0.49082313681868744, 'eval/latency_0/demon_right_precision': 0.7435294117647059, 'eval/latency_0/demon_right_recall': 0.8926553672316384, 'eval/latency_0/demon_right_f1': 0.8112965340179716, 'eval/latency_0/demon_right_rate_model': 0.5909343715239155, 'eval/latency_0/demon_right_rate_teacher': 0.4922135706340378, 'eval/latency_0/demon_move_f1': 0.7916186007543504, 'eval/latency_0/demon_cc_f1': 0.8736564482104902, 'timing/data': 0.0005857080686837435, 'timing/model': 1.2058949060738087, 'learning_rate/qwen_vl_interface': 7.625e-06, 'learning_rate/action_model': 7.625e-05, 'epoch': 0.71}) +[2026-06-17 10:36:13,570][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_2400_state +[2026-06-17 10:36:13,574][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 10:36:17,115][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_2400_state/model.safetensors +[2026-06-17 10:36:24,322][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_2400_state/optimizer.bin +[2026-06-17 10:36:24,323][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_2400_state/sampler.bin +[2026-06-17 10:36:24,323][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_2400_state/sampler_1.bin +[2026-06-17 10:36:24,324][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_2400_state/random_states_0.pkl +[2026-06-17 10:36:24,372][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2400 +[2026-06-17 10:36:24,373][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 10:36:24,382][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 10:36:29,085][starVLA.training.train_starvla][INFO] - Step 2401, Loss: {'train/loss': 1.0339701175689697, 'train/grad_norm_pre_clip': 10.785401344299316, 'timing/data': 0.0003802110441029072, 'timing/model': 1.1879085350083187, 'learning_rate/qwen_vl_interface': 7.6231268088990075e-06, 'learning_rate/action_model': 7.623126808899008e-05, 'epoch': 0.71}) +[2026-06-17 10:36:33,813][starVLA.training.train_starvla][INFO] - Step 2402, Loss: {'train/loss': 0.8231126070022583, 'train/grad_norm_pre_clip': 10.034340858459473, 'timing/data': 0.00036781199742108583, 'timing/model': 1.1871957529801875, 'learning_rate/qwen_vl_interface': 7.621253125846273e-06, 'learning_rate/action_model': 7.621253125846273e-05, 'epoch': 0.71}) +[2026-06-17 10:36:38,556][starVLA.training.train_starvla][INFO] - Step 2403, Loss: {'train/loss': 0.6584799885749817, 'train/grad_norm_pre_clip': 9.780338287353516, 'timing/data': 0.000349409063346684, 'timing/model': 1.3252372610149905, 'learning_rate/qwen_vl_interface': 7.619378951230213e-06, 'learning_rate/action_model': 7.619378951230213e-05, 'epoch': 0.71}) +[2026-06-17 10:36:43,276][starVLA.training.train_starvla][INFO] - Step 2404, Loss: {'train/loss': 0.6734803915023804, 'train/grad_norm_pre_clip': 7.7027153968811035, 'timing/data': 0.0003448600182309747, 'timing/model': 1.3098002220503986, 'learning_rate/qwen_vl_interface': 7.617504285439346e-06, 'learning_rate/action_model': 7.617504285439346e-05, 'epoch': 0.71}) +[2026-06-17 10:36:47,871][starVLA.training.train_starvla][INFO] - Step 2405, Loss: {'train/loss': 0.8572071194648743, 'train/grad_norm_pre_clip': 10.843469619750977, 'timing/data': 0.0003502060426399112, 'timing/model': 1.1930013770470396, 'learning_rate/qwen_vl_interface': 7.615629128862295e-06, 'learning_rate/action_model': 7.615629128862295e-05, 'epoch': 0.71}) +[2026-06-17 10:36:52,538][starVLA.training.train_starvla][INFO] - Step 2406, Loss: {'train/loss': 0.84897780418396, 'train/grad_norm_pre_clip': 10.202668190002441, 'timing/data': 0.0003647319972515106, 'timing/model': 1.1937160709640011, 'learning_rate/qwen_vl_interface': 7.61375348188778e-06, 'learning_rate/action_model': 7.61375348188778e-05, 'epoch': 0.71}) +[2026-06-17 10:36:57,291][starVLA.training.train_starvla][INFO] - Step 2407, Loss: {'train/loss': 0.9707046747207642, 'train/grad_norm_pre_clip': 9.616458892822266, 'timing/data': 0.0003188889240846038, 'timing/model': 1.194551756954752, 'learning_rate/qwen_vl_interface': 7.611877344904624e-06, 'learning_rate/action_model': 7.611877344904624e-05, 'epoch': 0.71}) +[2026-06-17 10:37:02,130][starVLA.training.train_starvla][INFO] - Step 2408, Loss: {'train/loss': 0.7317211627960205, 'train/grad_norm_pre_clip': 9.170568466186523, 'timing/data': 0.00036154710687696934, 'timing/model': 1.1942135840654373, 'learning_rate/qwen_vl_interface': 7.610000718301753e-06, 'learning_rate/action_model': 7.610000718301753e-05, 'epoch': 0.71}) +[2026-06-17 10:37:06,871][starVLA.training.train_starvla][INFO] - Step 2409, Loss: {'train/loss': 0.7674411535263062, 'train/grad_norm_pre_clip': 7.559568405151367, 'timing/data': 0.0005599999567493796, 'timing/model': 1.3214611809235066, 'learning_rate/qwen_vl_interface': 7.608123602468195e-06, 'learning_rate/action_model': 7.608123602468194e-05, 'epoch': 0.71}) +[2026-06-17 10:37:11,489][starVLA.training.train_starvla][INFO] - Step 2410, Loss: {'train/loss': 0.8189029097557068, 'train/grad_norm_pre_clip': 7.037827491760254, 'timing/data': 0.00036527100019156933, 'timing/model': 1.195937521988526, 'learning_rate/qwen_vl_interface': 7.606245997793074e-06, 'learning_rate/action_model': 7.606245997793075e-05, 'epoch': 0.71}) +[2026-06-17 10:37:16,234][starVLA.training.train_starvla][INFO] - Step 2411, Loss: {'train/loss': 0.8797075748443604, 'train/grad_norm_pre_clip': 13.160442352294922, 'timing/data': 0.0003506210632622242, 'timing/model': 1.193840917898342, 'learning_rate/qwen_vl_interface': 7.604367904665625e-06, 'learning_rate/action_model': 7.604367904665624e-05, 'epoch': 0.71}) +[2026-06-17 10:37:20,952][starVLA.training.train_starvla][INFO] - Step 2412, Loss: {'train/loss': 1.0062153339385986, 'train/grad_norm_pre_clip': 15.039870262145996, 'timing/data': 0.0003538919845595956, 'timing/model': 1.1964697130024433, 'learning_rate/qwen_vl_interface': 7.602489323475174e-06, 'learning_rate/action_model': 7.602489323475174e-05, 'epoch': 0.71}) +[2026-06-17 10:37:25,723][starVLA.training.train_starvla][INFO] - Step 2413, Loss: {'train/loss': 0.6695267558097839, 'train/grad_norm_pre_clip': 9.935869216918945, 'timing/data': 0.00035080104134976864, 'timing/model': 1.1974786469945684, 'learning_rate/qwen_vl_interface': 7.600610254611155e-06, 'learning_rate/action_model': 7.600610254611155e-05, 'epoch': 0.71}) +[2026-06-17 10:37:30,519][starVLA.training.train_starvla][INFO] - Step 2414, Loss: {'train/loss': 0.8345515131950378, 'train/grad_norm_pre_clip': 10.038311958312988, 'timing/data': 0.0003425150644034147, 'timing/model': 1.1976544309873134, 'learning_rate/qwen_vl_interface': 7.5987306984631025e-06, 'learning_rate/action_model': 7.598730698463102e-05, 'epoch': 0.71}) +[2026-06-17 10:37:35,290][starVLA.training.train_starvla][INFO] - Step 2415, Loss: {'train/loss': 0.8460801243782043, 'train/grad_norm_pre_clip': 9.465572357177734, 'timing/data': 0.0003662899835035205, 'timing/model': 1.3260057809529826, 'learning_rate/qwen_vl_interface': 7.596850655420647e-06, 'learning_rate/action_model': 7.596850655420646e-05, 'epoch': 0.71}) +[2026-06-17 10:37:39,981][starVLA.training.train_starvla][INFO] - Step 2416, Loss: {'train/loss': 0.7572184801101685, 'train/grad_norm_pre_clip': 9.342288970947266, 'timing/data': 0.00037176895420998335, 'timing/model': 1.1985836999956518, 'learning_rate/qwen_vl_interface': 7.5949701258735265e-06, 'learning_rate/action_model': 7.594970125873526e-05, 'epoch': 0.71}) +[2026-06-17 10:37:44,822][starVLA.training.train_starvla][INFO] - Step 2417, Loss: {'train/loss': 0.8416313529014587, 'train/grad_norm_pre_clip': 9.165809631347656, 'timing/data': 0.00035582203418016434, 'timing/model': 1.2093667200533673, 'learning_rate/qwen_vl_interface': 7.5930891102115735e-06, 'learning_rate/action_model': 7.593089110211573e-05, 'epoch': 0.71}) +[2026-06-17 10:37:49,612][starVLA.training.train_starvla][INFO] - Step 2418, Loss: {'train/loss': 1.0616356134414673, 'train/grad_norm_pre_clip': 11.291029930114746, 'timing/data': 0.00038790295366197824, 'timing/model': 1.2047575089382008, 'learning_rate/qwen_vl_interface': 7.59120760882473e-06, 'learning_rate/action_model': 7.59120760882473e-05, 'epoch': 0.71}) +[2026-06-17 10:37:54,392][starVLA.training.train_starvla][INFO] - Step 2419, Loss: {'train/loss': 0.7933726906776428, 'train/grad_norm_pre_clip': 9.81687068939209, 'timing/data': 0.0003538440214470029, 'timing/model': 1.2067072649952024, 'learning_rate/qwen_vl_interface': 7.5893256221030275e-06, 'learning_rate/action_model': 7.589325622103027e-05, 'epoch': 0.71}) +[2026-06-17 10:37:59,202][starVLA.training.train_starvla][INFO] - Step 2420, Loss: {'train/loss': 1.1372952461242676, 'train/grad_norm_pre_clip': 11.244190216064453, 'timing/data': 0.0003633020678535104, 'timing/model': 1.205663874046877, 'learning_rate/qwen_vl_interface': 7.587443150436608e-06, 'learning_rate/action_model': 7.587443150436608e-05, 'epoch': 0.71}) +[2026-06-17 10:38:03,975][starVLA.training.train_starvla][INFO] - Step 2421, Loss: {'train/loss': 0.7677009701728821, 'train/grad_norm_pre_clip': 9.093791007995605, 'timing/data': 0.0007105240365490317, 'timing/model': 1.3412967700278386, 'learning_rate/qwen_vl_interface': 7.585560194215706e-06, 'learning_rate/action_model': 7.585560194215706e-05, 'epoch': 0.71}) +[2026-06-17 10:38:08,618][starVLA.training.train_starvla][INFO] - Step 2422, Loss: {'train/loss': 0.8014867305755615, 'train/grad_norm_pre_clip': 10.388761520385742, 'timing/data': 0.00037721404805779457, 'timing/model': 1.1983724849997088, 'learning_rate/qwen_vl_interface': 7.583676753830665e-06, 'learning_rate/action_model': 7.583676753830664e-05, 'epoch': 0.71}) +[2026-06-17 10:38:13,406][starVLA.training.train_starvla][INFO] - Step 2423, Loss: {'train/loss': 0.8010555505752563, 'train/grad_norm_pre_clip': 10.202616691589355, 'timing/data': 0.0005974739324301481, 'timing/model': 1.202919906936586, 'learning_rate/qwen_vl_interface': 7.58179282967192e-06, 'learning_rate/action_model': 7.58179282967192e-05, 'epoch': 0.71}) +[2026-06-17 10:38:18,198][starVLA.training.train_starvla][INFO] - Step 2424, Loss: {'train/loss': 0.5751978754997253, 'train/grad_norm_pre_clip': 6.952056407928467, 'timing/data': 0.00038745906203985214, 'timing/model': 1.2047227000584826, 'learning_rate/qwen_vl_interface': 7.579908422130013e-06, 'learning_rate/action_model': 7.579908422130014e-05, 'epoch': 0.71}) +[2026-06-17 10:38:22,986][starVLA.training.train_starvla][INFO] - Step 2425, Loss: {'train/loss': 0.9455938339233398, 'train/grad_norm_pre_clip': 17.905187606811523, 'timing/data': 0.0003388490295037627, 'timing/model': 1.21041124407202, 'learning_rate/qwen_vl_interface': 7.578023531595583e-06, 'learning_rate/action_model': 7.578023531595583e-05, 'epoch': 0.71}) +[2026-06-17 10:38:27,732][starVLA.training.train_starvla][INFO] - Step 2426, Loss: {'train/loss': 0.7391119003295898, 'train/grad_norm_pre_clip': 10.408528327941895, 'timing/data': 0.0003901938907802105, 'timing/model': 1.2120036970591173, 'learning_rate/qwen_vl_interface': 7.576138158459369e-06, 'learning_rate/action_model': 7.576138158459368e-05, 'epoch': 0.71}) +[2026-06-17 10:38:32,511][starVLA.training.train_starvla][INFO] - Step 2427, Loss: {'train/loss': 0.8918693661689758, 'train/grad_norm_pre_clip': 9.306694030761719, 'timing/data': 0.00042209308594465256, 'timing/model': 1.339256535982713, 'learning_rate/qwen_vl_interface': 7.574252303112213e-06, 'learning_rate/action_model': 7.574252303112212e-05, 'epoch': 0.71}) +[2026-06-17 10:38:37,153][starVLA.training.train_starvla][INFO] - Step 2428, Loss: {'train/loss': 1.1673067808151245, 'train/grad_norm_pre_clip': 9.578181266784668, 'timing/data': 0.0005832799943163991, 'timing/model': 1.2022912569809705, 'learning_rate/qwen_vl_interface': 7.572365965945053e-06, 'learning_rate/action_model': 7.572365965945052e-05, 'epoch': 0.71}) +[2026-06-17 10:38:41,929][starVLA.training.train_starvla][INFO] - Step 2429, Loss: {'train/loss': 0.6747785210609436, 'train/grad_norm_pre_clip': 8.745737075805664, 'timing/data': 0.0003634369932115078, 'timing/model': 1.2013953910209239, 'learning_rate/qwen_vl_interface': 7.570479147348929e-06, 'learning_rate/action_model': 7.570479147348929e-05, 'epoch': 0.71}) +[2026-06-17 10:38:46,712][starVLA.training.train_starvla][INFO] - Step 2430, Loss: {'train/loss': 0.8747697472572327, 'train/grad_norm_pre_clip': 13.806196212768555, 'timing/data': 0.0003420670982450247, 'timing/model': 1.2069369530072436, 'learning_rate/qwen_vl_interface': 7.568591847714983e-06, 'learning_rate/action_model': 7.568591847714982e-05, 'epoch': 0.72}) +[2026-06-17 10:38:51,485][starVLA.training.train_starvla][INFO] - Step 2431, Loss: {'train/loss': 0.8039141297340393, 'train/grad_norm_pre_clip': 8.36432933807373, 'timing/data': 0.00037998007610440254, 'timing/model': 1.204874500981532, 'learning_rate/qwen_vl_interface': 7.5667040674344495e-06, 'learning_rate/action_model': 7.566704067434449e-05, 'epoch': 0.72}) +[2026-06-17 10:38:56,242][starVLA.training.train_starvla][INFO] - Step 2432, Loss: {'train/loss': 0.6130271553993225, 'train/grad_norm_pre_clip': 8.667800903320312, 'timing/data': 0.0003370599588379264, 'timing/model': 1.2024012770270929, 'learning_rate/qwen_vl_interface': 7.564815806898669e-06, 'learning_rate/action_model': 7.564815806898668e-05, 'epoch': 0.72}) +[2026-06-17 10:39:01,011][starVLA.training.train_starvla][INFO] - Step 2433, Loss: {'train/loss': 0.607092022895813, 'train/grad_norm_pre_clip': 7.1862006187438965, 'timing/data': 0.00037353497464209795, 'timing/model': 1.3270837529562414, 'learning_rate/qwen_vl_interface': 7.562927066499084e-06, 'learning_rate/action_model': 7.562927066499083e-05, 'epoch': 0.72}) +[2026-06-17 10:39:05,654][starVLA.training.train_starvla][INFO] - Step 2434, Loss: {'train/loss': 0.8644780516624451, 'train/grad_norm_pre_clip': 8.069672584533691, 'timing/data': 0.00035888899583369493, 'timing/model': 1.2000981390010566, 'learning_rate/qwen_vl_interface': 7.5610378466272274e-06, 'learning_rate/action_model': 7.561037846627227e-05, 'epoch': 0.72}) +[2026-06-17 10:39:10,502][starVLA.training.train_starvla][INFO] - Step 2435, Loss: {'train/loss': 1.0415536165237427, 'train/grad_norm_pre_clip': 10.29443645477295, 'timing/data': 0.0003978860331699252, 'timing/model': 1.286565855029039, 'learning_rate/qwen_vl_interface': 7.5591481476747384e-06, 'learning_rate/action_model': 7.559148147674739e-05, 'epoch': 0.72}) +[2026-06-17 10:39:15,285][starVLA.training.train_starvla][INFO] - Step 2436, Loss: {'train/loss': 0.8936458826065063, 'train/grad_norm_pre_clip': 15.303367614746094, 'timing/data': 0.0003507809014990926, 'timing/model': 1.2033302090130746, 'learning_rate/qwen_vl_interface': 7.557257970033355e-06, 'learning_rate/action_model': 7.557257970033354e-05, 'epoch': 0.72}) +[2026-06-17 10:39:20,072][starVLA.training.train_starvla][INFO] - Step 2437, Loss: {'train/loss': 0.8877591490745544, 'train/grad_norm_pre_clip': 8.256543159484863, 'timing/data': 0.00035390397533774376, 'timing/model': 1.2063543100375682, 'learning_rate/qwen_vl_interface': 7.55536731409491e-06, 'learning_rate/action_model': 7.55536731409491e-05, 'epoch': 0.72}) +[2026-06-17 10:39:24,852][starVLA.training.train_starvla][INFO] - Step 2438, Loss: {'train/loss': 0.9698487520217896, 'train/grad_norm_pre_clip': 10.910136222839355, 'timing/data': 0.0003432310186326504, 'timing/model': 1.3318964469945058, 'learning_rate/qwen_vl_interface': 7.553476180251341e-06, 'learning_rate/action_model': 7.55347618025134e-05, 'epoch': 0.72}) +[2026-06-17 10:39:29,551][starVLA.training.train_starvla][INFO] - Step 2439, Loss: {'train/loss': 0.962123453617096, 'train/grad_norm_pre_clip': 9.435248374938965, 'timing/data': 0.0003432030789554119, 'timing/model': 1.2441380210220814, 'learning_rate/qwen_vl_interface': 7.551584568894681e-06, 'learning_rate/action_model': 7.551584568894681e-05, 'epoch': 0.72}) +[2026-06-17 10:39:34,204][starVLA.training.train_starvla][INFO] - Step 2440, Loss: {'train/loss': 0.8026541471481323, 'train/grad_norm_pre_clip': 8.885128021240234, 'timing/data': 0.0005221180617809296, 'timing/model': 1.2025236849440262, 'learning_rate/qwen_vl_interface': 7.5496924804170635e-06, 'learning_rate/action_model': 7.549692480417063e-05, 'epoch': 0.72}) +[2026-06-17 10:39:38,942][starVLA.training.train_starvla][INFO] - Step 2441, Loss: {'train/loss': 0.8493154048919678, 'train/grad_norm_pre_clip': 8.945504188537598, 'timing/data': 0.00039956707041710615, 'timing/model': 1.2120798289543018, 'learning_rate/qwen_vl_interface': 7.5477999152107195e-06, 'learning_rate/action_model': 7.54779991521072e-05, 'epoch': 0.72}) +[2026-06-17 10:39:43,713][starVLA.training.train_starvla][INFO] - Step 2442, Loss: {'train/loss': 0.6812223792076111, 'train/grad_norm_pre_clip': 8.791739463806152, 'timing/data': 0.0003988490207120776, 'timing/model': 1.2022253789473325, 'learning_rate/qwen_vl_interface': 7.545906873667982e-06, 'learning_rate/action_model': 7.545906873667981e-05, 'epoch': 0.72}) +[2026-06-17 10:39:48,552][starVLA.training.train_starvla][INFO] - Step 2443, Loss: {'train/loss': 0.9060872793197632, 'train/grad_norm_pre_clip': 11.520103454589844, 'timing/data': 0.00035552994813770056, 'timing/model': 1.214796622050926, 'learning_rate/qwen_vl_interface': 7.544013356181277e-06, 'learning_rate/action_model': 7.544013356181277e-05, 'epoch': 0.72}) +[2026-06-17 10:39:53,318][starVLA.training.train_starvla][INFO] - Step 2444, Loss: {'train/loss': 1.0681239366531372, 'train/grad_norm_pre_clip': 10.052578926086426, 'timing/data': 0.00034102704375982285, 'timing/model': 1.3291892589768395, 'learning_rate/qwen_vl_interface': 7.542119363143135e-06, 'learning_rate/action_model': 7.542119363143136e-05, 'epoch': 0.72}) +[2026-06-17 10:39:57,959][starVLA.training.train_starvla][INFO] - Step 2445, Loss: {'train/loss': 0.7699369788169861, 'train/grad_norm_pre_clip': 9.114020347595215, 'timing/data': 0.0004817289300262928, 'timing/model': 1.2005153339123353, 'learning_rate/qwen_vl_interface': 7.540224894946184e-06, 'learning_rate/action_model': 7.540224894946184e-05, 'epoch': 0.72}) +[2026-06-17 10:40:02,738][starVLA.training.train_starvla][INFO] - Step 2446, Loss: {'train/loss': 0.7552560567855835, 'train/grad_norm_pre_clip': 11.87267017364502, 'timing/data': 0.0006919869920238853, 'timing/model': 1.2098732739686966, 'learning_rate/qwen_vl_interface': 7.538329951983148e-06, 'learning_rate/action_model': 7.538329951983147e-05, 'epoch': 0.72}) +[2026-06-17 10:40:07,526][starVLA.training.train_starvla][INFO] - Step 2447, Loss: {'train/loss': 0.832683265209198, 'train/grad_norm_pre_clip': 6.955721855163574, 'timing/data': 0.0006264899857342243, 'timing/model': 1.2034513379912823, 'learning_rate/qwen_vl_interface': 7.536434534646851e-06, 'learning_rate/action_model': 7.53643453464685e-05, 'epoch': 0.72}) +[2026-06-17 10:40:12,312][starVLA.training.train_starvla][INFO] - Step 2448, Loss: {'train/loss': 0.9452101588249207, 'train/grad_norm_pre_clip': 9.309089660644531, 'timing/data': 0.00036227796226739883, 'timing/model': 1.2005692719249055, 'learning_rate/qwen_vl_interface': 7.534538643330213e-06, 'learning_rate/action_model': 7.534538643330214e-05, 'epoch': 0.72}) +[2026-06-17 10:40:17,100][starVLA.training.train_starvla][INFO] - Step 2449, Loss: {'train/loss': 0.6956347823143005, 'train/grad_norm_pre_clip': 7.413999557495117, 'timing/data': 0.0003280990058556199, 'timing/model': 1.2008079789811745, 'learning_rate/qwen_vl_interface': 7.5326422784262566e-06, 'learning_rate/action_model': 7.532642278426256e-05, 'epoch': 0.72}) +[2026-06-17 10:40:21,873][starVLA.training.train_starvla][INFO] - Step 2450, Loss: {'train/loss': 1.1417516469955444, 'train/grad_norm_pre_clip': 11.545821189880371, 'timing/data': 0.00038494099862873554, 'timing/model': 1.3338587230537087, 'learning_rate/qwen_vl_interface': 7.5307454403281e-06, 'learning_rate/action_model': 7.5307454403281e-05, 'epoch': 0.72}) +[2026-06-17 10:40:26,531][starVLA.training.train_starvla][INFO] - Step 2451, Loss: {'train/loss': 0.6215904355049133, 'train/grad_norm_pre_clip': 6.771060943603516, 'timing/data': 0.0007856989977881312, 'timing/model': 1.204699064954184, 'learning_rate/qwen_vl_interface': 7.528848129428958e-06, 'learning_rate/action_model': 7.528848129428958e-05, 'epoch': 0.72}) +[2026-06-17 10:40:31,308][starVLA.training.train_starvla][INFO] - Step 2452, Loss: {'train/loss': 0.7584801912307739, 'train/grad_norm_pre_clip': 8.749500274658203, 'timing/data': 0.00038509699515998363, 'timing/model': 1.204458498978056, 'learning_rate/qwen_vl_interface': 7.526950346122149e-06, 'learning_rate/action_model': 7.526950346122148e-05, 'epoch': 0.72}) +[2026-06-17 10:40:36,072][starVLA.training.train_starvla][INFO] - Step 2453, Loss: {'train/loss': 0.9066503047943115, 'train/grad_norm_pre_clip': 8.178812980651855, 'timing/data': 0.0003780919359996915, 'timing/model': 1.1995769659988582, 'learning_rate/qwen_vl_interface': 7.525052090801083e-06, 'learning_rate/action_model': 7.525052090801083e-05, 'epoch': 0.72}) +[2026-06-17 10:40:40,863][starVLA.training.train_starvla][INFO] - Step 2454, Loss: {'train/loss': 0.735929548740387, 'train/grad_norm_pre_clip': 9.354269981384277, 'timing/data': 0.0005929120816290379, 'timing/model': 1.2069201930426061, 'learning_rate/qwen_vl_interface': 7.523153363859269e-06, 'learning_rate/action_model': 7.523153363859269e-05, 'epoch': 0.72}) +[2026-06-17 10:40:45,598][starVLA.training.train_starvla][INFO] - Step 2455, Loss: {'train/loss': 0.9464646577835083, 'train/grad_norm_pre_clip': 27.040788650512695, 'timing/data': 0.0003704419359564781, 'timing/model': 1.2015998660353944, 'learning_rate/qwen_vl_interface': 7.521254165690317e-06, 'learning_rate/action_model': 7.521254165690317e-05, 'epoch': 0.72}) +[2026-06-17 10:40:50,384][starVLA.training.train_starvla][INFO] - Step 2456, Loss: {'train/loss': 0.8325629830360413, 'train/grad_norm_pre_clip': 12.274249076843262, 'timing/data': 0.0004105689004063606, 'timing/model': 1.3301208979682997, 'learning_rate/qwen_vl_interface': 7.519354496687933e-06, 'learning_rate/action_model': 7.519354496687933e-05, 'epoch': 0.72}) +[2026-06-17 10:40:55,030][starVLA.training.train_starvla][INFO] - Step 2457, Loss: {'train/loss': 0.6492886543273926, 'train/grad_norm_pre_clip': 8.898665428161621, 'timing/data': 0.0003664370160549879, 'timing/model': 1.2012687620008364, 'learning_rate/qwen_vl_interface': 7.517454357245918e-06, 'learning_rate/action_model': 7.517454357245917e-05, 'epoch': 0.72}) +[2026-06-17 10:40:59,813][starVLA.training.train_starvla][INFO] - Step 2458, Loss: {'train/loss': 0.7603484392166138, 'train/grad_norm_pre_clip': 8.581559181213379, 'timing/data': 0.0003701519453898072, 'timing/model': 1.201805499033071, 'learning_rate/qwen_vl_interface': 7.515553747758176e-06, 'learning_rate/action_model': 7.515553747758176e-05, 'epoch': 0.72}) +[2026-06-17 10:41:04,642][starVLA.training.train_starvla][INFO] - Step 2459, Loss: {'train/loss': 0.8926980495452881, 'train/grad_norm_pre_clip': 11.64412784576416, 'timing/data': 0.0003470160299912095, 'timing/model': 1.2054396929452196, 'learning_rate/qwen_vl_interface': 7.513652668618705e-06, 'learning_rate/action_model': 7.513652668618705e-05, 'epoch': 0.72}) +[2026-06-17 10:41:09,421][starVLA.training.train_starvla][INFO] - Step 2460, Loss: {'train/loss': 0.7570425868034363, 'train/grad_norm_pre_clip': 7.661818027496338, 'timing/data': 0.0005888149607926607, 'timing/model': 1.200171446078457, 'learning_rate/qwen_vl_interface': 7.5117511202215994e-06, 'learning_rate/action_model': 7.5117511202216e-05, 'epoch': 0.72}) +[2026-06-17 10:41:14,177][starVLA.training.train_starvla][INFO] - Step 2461, Loss: {'train/loss': 0.8055052757263184, 'train/grad_norm_pre_clip': 7.505646705627441, 'timing/data': 0.00034090899862349033, 'timing/model': 1.201399683021009, 'learning_rate/qwen_vl_interface': 7.509849102961053e-06, 'learning_rate/action_model': 7.509849102961053e-05, 'epoch': 0.72}) +[2026-06-17 10:41:18,953][starVLA.training.train_starvla][INFO] - Step 2462, Loss: {'train/loss': 0.9821714162826538, 'train/grad_norm_pre_clip': 10.528642654418945, 'timing/data': 0.0003667839337140322, 'timing/model': 1.3336051609367132, 'learning_rate/qwen_vl_interface': 7.507946617231357e-06, 'learning_rate/action_model': 7.507946617231356e-05, 'epoch': 0.72}) +[2026-06-17 10:41:23,602][starVLA.training.train_starvla][INFO] - Step 2463, Loss: {'train/loss': 0.6407706141471863, 'train/grad_norm_pre_clip': 19.678525924682617, 'timing/data': 0.0003538009477779269, 'timing/model': 1.2018812929745764, 'learning_rate/qwen_vl_interface': 7.5060436634268965e-06, 'learning_rate/action_model': 7.506043663426896e-05, 'epoch': 0.72}) +[2026-06-17 10:41:28,395][starVLA.training.train_starvla][INFO] - Step 2464, Loss: {'train/loss': 0.9985042214393616, 'train/grad_norm_pre_clip': 9.369683265686035, 'timing/data': 0.00036360800731927156, 'timing/model': 1.2026840359903872, 'learning_rate/qwen_vl_interface': 7.504140241942157e-06, 'learning_rate/action_model': 7.504140241942157e-05, 'epoch': 0.73}) +[2026-06-17 10:41:33,187][starVLA.training.train_starvla][INFO] - Step 2465, Loss: {'train/loss': 0.7846351861953735, 'train/grad_norm_pre_clip': 8.996893882751465, 'timing/data': 0.000373765011318028, 'timing/model': 1.2012609029188752, 'learning_rate/qwen_vl_interface': 7.502236353171718e-06, 'learning_rate/action_model': 7.502236353171718e-05, 'epoch': 0.73}) +[2026-06-17 10:41:37,981][starVLA.training.train_starvla][INFO] - Step 2466, Loss: {'train/loss': 0.933148205280304, 'train/grad_norm_pre_clip': 10.32628059387207, 'timing/data': 0.0003592419670894742, 'timing/model': 1.202099372050725, 'learning_rate/qwen_vl_interface': 7.500331997510262e-06, 'learning_rate/action_model': 7.500331997510262e-05, 'epoch': 0.73}) +[2026-06-17 10:41:42,771][starVLA.training.train_starvla][INFO] - Step 2467, Loss: {'train/loss': 0.8202611207962036, 'train/grad_norm_pre_clip': 7.383228778839111, 'timing/data': 0.00035805790685117245, 'timing/model': 1.3359999649692327, 'learning_rate/qwen_vl_interface': 7.498427175352561e-06, 'learning_rate/action_model': 7.49842717535256e-05, 'epoch': 0.73}) +[2026-06-17 10:41:47,446][starVLA.training.train_starvla][INFO] - Step 2468, Loss: {'train/loss': 1.1614383459091187, 'train/grad_norm_pre_clip': 13.303194999694824, 'timing/data': 0.00034237292129546404, 'timing/model': 1.2336997019592673, 'learning_rate/qwen_vl_interface': 7.496521887093487e-06, 'learning_rate/action_model': 7.496521887093486e-05, 'epoch': 0.73}) +[2026-06-17 10:41:52,102][starVLA.training.train_starvla][INFO] - Step 2469, Loss: {'train/loss': 0.734727680683136, 'train/grad_norm_pre_clip': 7.843549728393555, 'timing/data': 0.0007135800551623106, 'timing/model': 1.2023297080304474, 'learning_rate/qwen_vl_interface': 7.494616133128008e-06, 'learning_rate/action_model': 7.494616133128008e-05, 'epoch': 0.73}) +[2026-06-17 10:41:56,835][starVLA.training.train_starvla][INFO] - Step 2470, Loss: {'train/loss': 0.8039518594741821, 'train/grad_norm_pre_clip': 8.848122596740723, 'timing/data': 0.00035481201484799385, 'timing/model': 1.2031162929488346, 'learning_rate/qwen_vl_interface': 7.4927099138511895e-06, 'learning_rate/action_model': 7.492709913851189e-05, 'epoch': 0.73}) +[2026-06-17 10:42:01,618][starVLA.training.train_starvla][INFO] - Step 2471, Loss: {'train/loss': 0.6223120093345642, 'train/grad_norm_pre_clip': 8.489964485168457, 'timing/data': 0.0003736380022019148, 'timing/model': 1.2020578749943525, 'learning_rate/qwen_vl_interface': 7.490803229658192e-06, 'learning_rate/action_model': 7.490803229658192e-05, 'epoch': 0.73}) +[2026-06-17 10:42:06,400][starVLA.training.train_starvla][INFO] - Step 2472, Loss: {'train/loss': 0.7417994737625122, 'train/grad_norm_pre_clip': 10.821538925170898, 'timing/data': 0.0003837699769064784, 'timing/model': 1.2017816860461608, 'learning_rate/qwen_vl_interface': 7.488896080944274e-06, 'learning_rate/action_model': 7.488896080944273e-05, 'epoch': 0.73}) +[2026-06-17 10:42:11,173][starVLA.training.train_starvla][INFO] - Step 2473, Loss: {'train/loss': 0.6731295585632324, 'train/grad_norm_pre_clip': 8.057293891906738, 'timing/data': 0.00034650403540581465, 'timing/model': 1.3307533179176971, 'learning_rate/qwen_vl_interface': 7.4869884681047896e-06, 'learning_rate/action_model': 7.486988468104789e-05, 'epoch': 0.73}) +[2026-06-17 10:42:15,821][starVLA.training.train_starvla][INFO] - Step 2474, Loss: {'train/loss': 0.6737654805183411, 'train/grad_norm_pre_clip': 8.400654792785645, 'timing/data': 0.0003471310483291745, 'timing/model': 1.2026506820693612, 'learning_rate/qwen_vl_interface': 7.485080391535189e-06, 'learning_rate/action_model': 7.485080391535188e-05, 'epoch': 0.73}) +[2026-06-17 10:42:20,625][starVLA.training.train_starvla][INFO] - Step 2475, Loss: {'train/loss': 0.8168728351593018, 'train/grad_norm_pre_clip': 6.99249792098999, 'timing/data': 0.0003479180159047246, 'timing/model': 1.208503810921684, 'learning_rate/qwen_vl_interface': 7.483171851631018e-06, 'learning_rate/action_model': 7.483171851631017e-05, 'epoch': 0.73}) +[2026-06-17 10:42:25,391][starVLA.training.train_starvla][INFO] - Step 2476, Loss: {'train/loss': 0.727184534072876, 'train/grad_norm_pre_clip': 7.686174392700195, 'timing/data': 0.00036643107887357473, 'timing/model': 1.2014730049995705, 'learning_rate/qwen_vl_interface': 7.481262848787919e-06, 'learning_rate/action_model': 7.481262848787919e-05, 'epoch': 0.73}) +[2026-06-17 10:42:30,162][starVLA.training.train_starvla][INFO] - Step 2477, Loss: {'train/loss': 0.9268488883972168, 'train/grad_norm_pre_clip': 10.12736988067627, 'timing/data': 0.0004367120563983917, 'timing/model': 1.2017798479646444, 'learning_rate/qwen_vl_interface': 7.479353383401631e-06, 'learning_rate/action_model': 7.47935338340163e-05, 'epoch': 0.73}) +[2026-06-17 10:42:34,978][starVLA.training.train_starvla][INFO] - Step 2478, Loss: {'train/loss': 0.8762876391410828, 'train/grad_norm_pre_clip': 9.032230377197266, 'timing/data': 0.0003757580416277051, 'timing/model': 1.1998378990683705, 'learning_rate/qwen_vl_interface': 7.477443455867987e-06, 'learning_rate/action_model': 7.477443455867988e-05, 'epoch': 0.73}) +[2026-06-17 10:42:39,763][starVLA.training.train_starvla][INFO] - Step 2479, Loss: {'train/loss': 0.8158600926399231, 'train/grad_norm_pre_clip': 9.240628242492676, 'timing/data': 0.00035560305695980787, 'timing/model': 1.3306892189430073, 'learning_rate/qwen_vl_interface': 7.475533066582917e-06, 'learning_rate/action_model': 7.475533066582918e-05, 'epoch': 0.73}) +[2026-06-17 10:42:44,412][starVLA.training.train_starvla][INFO] - Step 2480, Loss: {'train/loss': 0.8457528352737427, 'train/grad_norm_pre_clip': 15.772331237792969, 'timing/data': 0.0003298550145700574, 'timing/model': 1.20244806795381, 'learning_rate/qwen_vl_interface': 7.47362221594245e-06, 'learning_rate/action_model': 7.47362221594245e-05, 'epoch': 0.73}) +[2026-06-17 10:42:49,188][starVLA.training.train_starvla][INFO] - Step 2481, Loss: {'train/loss': 0.6361956000328064, 'train/grad_norm_pre_clip': 7.412302494049072, 'timing/data': 0.0003434339305385947, 'timing/model': 1.206295636948198, 'learning_rate/qwen_vl_interface': 7.4717109043427025e-06, 'learning_rate/action_model': 7.471710904342702e-05, 'epoch': 0.73}) +[2026-06-17 10:42:53,963][starVLA.training.train_starvla][INFO] - Step 2482, Loss: {'train/loss': 0.7164145112037659, 'train/grad_norm_pre_clip': 7.749622821807861, 'timing/data': 0.0003560309996828437, 'timing/model': 1.2013671699678525, 'learning_rate/qwen_vl_interface': 7.469799132179894e-06, 'learning_rate/action_model': 7.469799132179895e-05, 'epoch': 0.73}) +[2026-06-17 10:42:58,731][starVLA.training.train_starvla][INFO] - Step 2483, Loss: {'train/loss': 0.8533210158348083, 'train/grad_norm_pre_clip': 8.178924560546875, 'timing/data': 0.0004027090035378933, 'timing/model': 1.2034045689506456, 'learning_rate/qwen_vl_interface': 7.467886899850338e-06, 'learning_rate/action_model': 7.467886899850338e-05, 'epoch': 0.73}) +[2026-06-17 10:43:03,454][starVLA.training.train_starvla][INFO] - Step 2484, Loss: {'train/loss': 0.8471338152885437, 'train/grad_norm_pre_clip': 8.996017456054688, 'timing/data': 0.0003592000575736165, 'timing/model': 1.200857050018385, 'learning_rate/qwen_vl_interface': 7.465974207750442e-06, 'learning_rate/action_model': 7.465974207750442e-05, 'epoch': 0.73}) +[2026-06-17 10:43:08,235][starVLA.training.train_starvla][INFO] - Step 2485, Loss: {'train/loss': 0.8386775851249695, 'train/grad_norm_pre_clip': 9.61447525024414, 'timing/data': 0.000372120994143188, 'timing/model': 1.3371717310510576, 'learning_rate/qwen_vl_interface': 7.464061056276707e-06, 'learning_rate/action_model': 7.464061056276707e-05, 'epoch': 0.73}) +[2026-06-17 10:43:12,874][starVLA.training.train_starvla][INFO] - Step 2486, Loss: {'train/loss': 1.1473748683929443, 'train/grad_norm_pre_clip': 14.29140853881836, 'timing/data': 0.0007087590638548136, 'timing/model': 1.2000099970027804, 'learning_rate/qwen_vl_interface': 7.462147445825733e-06, 'learning_rate/action_model': 7.462147445825733e-05, 'epoch': 0.73}) +[2026-06-17 10:43:17,643][starVLA.training.train_starvla][INFO] - Step 2487, Loss: {'train/loss': 0.7972599864006042, 'train/grad_norm_pre_clip': 10.27736759185791, 'timing/data': 0.0003893610555678606, 'timing/model': 1.2027373800519854, 'learning_rate/qwen_vl_interface': 7.460233376794211e-06, 'learning_rate/action_model': 7.46023337679421e-05, 'epoch': 0.73}) +[2026-06-17 10:43:22,417][starVLA.training.train_starvla][INFO] - Step 2488, Loss: {'train/loss': 1.2351360321044922, 'train/grad_norm_pre_clip': 11.840107917785645, 'timing/data': 0.00037168199196457863, 'timing/model': 1.2115999199450016, 'learning_rate/qwen_vl_interface': 7.458318849578934e-06, 'learning_rate/action_model': 7.458318849578933e-05, 'epoch': 0.73}) +[2026-06-17 10:43:27,185][starVLA.training.train_starvla][INFO] - Step 2489, Loss: {'train/loss': 0.8063877820968628, 'train/grad_norm_pre_clip': 10.64573860168457, 'timing/data': 0.00035708595532923937, 'timing/model': 1.20280329301022, 'learning_rate/qwen_vl_interface': 7.456403864576782e-06, 'learning_rate/action_model': 7.456403864576782e-05, 'epoch': 0.73}) +[2026-06-17 10:43:31,932][starVLA.training.train_starvla][INFO] - Step 2490, Loss: {'train/loss': 0.8639019727706909, 'train/grad_norm_pre_clip': 11.119599342346191, 'timing/data': 0.0003705560229718685, 'timing/model': 1.2001902649644762, 'learning_rate/qwen_vl_interface': 7.4544884221847354e-06, 'learning_rate/action_model': 7.454488422184735e-05, 'epoch': 0.73}) +[2026-06-17 10:43:36,714][starVLA.training.train_starvla][INFO] - Step 2491, Loss: {'train/loss': 0.6002762913703918, 'train/grad_norm_pre_clip': 7.633735179901123, 'timing/data': 0.00035132095217704773, 'timing/model': 1.335987125057727, 'learning_rate/qwen_vl_interface': 7.452572522799867e-06, 'learning_rate/action_model': 7.452572522799866e-05, 'epoch': 0.73}) +[2026-06-17 10:43:41,366][starVLA.training.train_starvla][INFO] - Step 2492, Loss: {'train/loss': 0.6169676184654236, 'train/grad_norm_pre_clip': 6.906113624572754, 'timing/data': 0.000518969027325511, 'timing/model': 1.2028056940762326, 'learning_rate/qwen_vl_interface': 7.450656166819344e-06, 'learning_rate/action_model': 7.450656166819344e-05, 'epoch': 0.73}) +[2026-06-17 10:43:46,145][starVLA.training.train_starvla][INFO] - Step 2493, Loss: {'train/loss': 0.8062171339988708, 'train/grad_norm_pre_clip': 7.997616767883301, 'timing/data': 0.00037752394564449787, 'timing/model': 1.2085461650276557, 'learning_rate/qwen_vl_interface': 7.4487393546404295e-06, 'learning_rate/action_model': 7.44873935464043e-05, 'epoch': 0.73}) +[2026-06-17 10:43:50,915][starVLA.training.train_starvla][INFO] - Step 2494, Loss: {'train/loss': 0.6694338321685791, 'train/grad_norm_pre_clip': 7.9339776039123535, 'timing/data': 0.0003469110233709216, 'timing/model': 1.2007206269772723, 'learning_rate/qwen_vl_interface': 7.446822086660481e-06, 'learning_rate/action_model': 7.446822086660481e-05, 'epoch': 0.73}) +[2026-06-17 10:43:55,725][starVLA.training.train_starvla][INFO] - Step 2495, Loss: {'train/loss': 0.8529514670372009, 'train/grad_norm_pre_clip': 9.15495777130127, 'timing/data': 0.0004432318964973092, 'timing/model': 1.2022081959294155, 'learning_rate/qwen_vl_interface': 7.444904363276951e-06, 'learning_rate/action_model': 7.444904363276951e-05, 'epoch': 0.73}) +[2026-06-17 10:44:00,506][starVLA.training.train_starvla][INFO] - Step 2496, Loss: {'train/loss': 0.9244638681411743, 'train/grad_norm_pre_clip': 9.836297035217285, 'timing/data': 0.0003597360337153077, 'timing/model': 1.3394336850615218, 'learning_rate/qwen_vl_interface': 7.442986184887383e-06, 'learning_rate/action_model': 7.442986184887383e-05, 'epoch': 0.73}) +[2026-06-17 10:44:05,202][starVLA.training.train_starvla][INFO] - Step 2497, Loss: {'train/loss': 0.7741041779518127, 'train/grad_norm_pre_clip': 8.336591720581055, 'timing/data': 0.00040570099372416735, 'timing/model': 1.2547116270288825, 'learning_rate/qwen_vl_interface': 7.4410675518894205e-06, 'learning_rate/action_model': 7.44106755188942e-05, 'epoch': 0.73}) +[2026-06-17 10:44:09,857][starVLA.training.train_starvla][INFO] - Step 2498, Loss: {'train/loss': 0.8646706342697144, 'train/grad_norm_pre_clip': 9.88497257232666, 'timing/data': 0.00041205494198948145, 'timing/model': 1.2175399051047862, 'learning_rate/qwen_vl_interface': 7.439148464680795e-06, 'learning_rate/action_model': 7.439148464680796e-05, 'epoch': 0.74}) +[2026-06-17 10:44:14,586][starVLA.training.train_starvla][INFO] - Step 2499, Loss: {'train/loss': 0.6884714365005493, 'train/grad_norm_pre_clip': 9.290081024169922, 'timing/data': 0.00038051395677030087, 'timing/model': 1.2051612919894978, 'learning_rate/qwen_vl_interface': 7.4372289236593395e-06, 'learning_rate/action_model': 7.43722892365934e-05, 'epoch': 0.74}) +[2026-06-17 10:53:45,892][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 2500 +[2026-06-17 10:53:45,893][starVLA.training.train_starvla][INFO] - Step 2500, Loss: {'train/loss': 0.7485191822052002, 'train/grad_norm_pre_clip': 8.55802059173584, 'eval/loss': 0.8231983780860901, 'eval/latency_0/loss': 0.8231983780860901, 'eval/demon_fire_precision': 0.9217926186291739, 'eval/demon_fire_recall': 0.8712624584717608, 'eval/demon_fire_f1': 0.8958155422715627, 'eval/demon_fire_rate_model': 0.6329254727474972, 'eval/demon_fire_rate_teacher': 0.6696329254727474, 'eval/demon_left_precision': 0.7683377308707124, 'eval/demon_left_recall': 0.8249291784702549, 'eval/demon_left_f1': 0.7956284153005466, 'eval/demon_left_rate_model': 0.5269744160177976, 'eval/demon_left_rate_teacher': 0.49082313681868744, 'eval/demon_right_precision': 0.8068517424689899, 'eval/demon_right_recall': 0.7717514124293785, 'eval/demon_right_f1': 0.7889113485417267, 'eval/demon_right_rate_model': 0.47080088987764185, 'eval/demon_right_rate_teacher': 0.4922135706340378, 'eval/demon_move_f1': 0.7922698819211367, 'eval/demon_cc_f1': 0.8620171594069679, 'eval/latency_0/demon_fire_precision': 0.9217926186291739, 'eval/latency_0/demon_fire_recall': 0.8712624584717608, 'eval/latency_0/demon_fire_f1': 0.8958155422715627, 'eval/latency_0/demon_fire_rate_model': 0.6329254727474972, 'eval/latency_0/demon_fire_rate_teacher': 0.6696329254727474, 'eval/latency_0/demon_left_precision': 0.7683377308707124, 'eval/latency_0/demon_left_recall': 0.8249291784702549, 'eval/latency_0/demon_left_f1': 0.7956284153005466, 'eval/latency_0/demon_left_rate_model': 0.5269744160177976, 'eval/latency_0/demon_left_rate_teacher': 0.49082313681868744, 'eval/latency_0/demon_right_precision': 0.8068517424689899, 'eval/latency_0/demon_right_recall': 0.7717514124293785, 'eval/latency_0/demon_right_f1': 0.7889113485417267, 'eval/latency_0/demon_right_rate_model': 0.47080088987764185, 'eval/latency_0/demon_right_rate_teacher': 0.4922135706340378, 'eval/latency_0/demon_move_f1': 0.7922698819211367, 'eval/latency_0/demon_cc_f1': 0.8620171594069679, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 7588.0, 'rl_games_eval/mid_train/mean_length': 3336.2, 'rl_games_eval/mid_train/std_reward': 2504.1956792551177, 'rl_games_eval/mid_train/std_length': 527.5999999999999, 'rl_games_eval/mid_train/macro_mean_reward': 7588.0, 'rl_games_eval/mid_train/macro_mean_length': 3336.2, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/demon_attack__latency_0/mean_reward': 7588.0, 'rl_games_eval/mid_train/demon_attack__latency_0/mean_length': 3336.2, 'rl_games_eval/mid_train/demon_attack__latency_0/std_reward': 2504.1956792551177, 'rl_games_eval/mid_train/demon_attack__latency_0/std_length': 527.5999999999999, 'rl_games_eval/mid_train/demon_attack__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 7588.0, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.00035716104321181774, 'timing/model': 1.2020348989171907, 'learning_rate/qwen_vl_interface': 7.435308929222973e-06, 'learning_rate/action_model': 7.435308929222972e-05, 'epoch': 0.74}) +[2026-06-17 10:53:45,893][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_2500_state +[2026-06-17 10:53:45,896][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 10:53:49,450][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_2500_state/model.safetensors +[2026-06-17 10:53:57,154][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_2500_state/optimizer.bin +[2026-06-17 10:53:57,154][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_2500_state/sampler.bin +[2026-06-17 10:53:57,154][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_2500_state/sampler_1.bin +[2026-06-17 10:53:57,155][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_2500_state/random_states_0.pkl +[2026-06-17 10:53:57,203][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2500 +[2026-06-17 10:53:57,203][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 10:53:57,213][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 10:54:01,891][starVLA.training.train_starvla][INFO] - Step 2501, Loss: {'train/loss': 0.7652539610862732, 'train/grad_norm_pre_clip': 7.5049147605896, 'timing/data': 0.0003616129979491234, 'timing/model': 1.1830539470538497, 'learning_rate/qwen_vl_interface': 7.433388481769716e-06, 'learning_rate/action_model': 7.433388481769715e-05, 'epoch': 0.74}) +[2026-06-17 10:54:06,596][starVLA.training.train_starvla][INFO] - Step 2502, Loss: {'train/loss': 0.4959520399570465, 'train/grad_norm_pre_clip': 6.271543502807617, 'timing/data': 0.000712668988853693, 'timing/model': 1.2931595139671117, 'learning_rate/qwen_vl_interface': 7.431467581697675e-06, 'learning_rate/action_model': 7.431467581697674e-05, 'epoch': 0.74}) +[2026-06-17 10:54:11,205][starVLA.training.train_starvla][INFO] - Step 2503, Loss: {'train/loss': 0.5848332643508911, 'train/grad_norm_pre_clip': 7.353785037994385, 'timing/data': 0.00034187897108495235, 'timing/model': 1.1932942919665948, 'learning_rate/qwen_vl_interface': 7.429546229405057e-06, 'learning_rate/action_model': 7.429546229405057e-05, 'epoch': 0.74}) +[2026-06-17 10:54:15,938][starVLA.training.train_starvla][INFO] - Step 2504, Loss: {'train/loss': 0.9596981406211853, 'train/grad_norm_pre_clip': 8.999168395996094, 'timing/data': 0.000350340036675334, 'timing/model': 1.192295414977707, 'learning_rate/qwen_vl_interface': 7.427624425290162e-06, 'learning_rate/action_model': 7.427624425290161e-05, 'epoch': 0.74}) +[2026-06-17 10:54:20,685][starVLA.training.train_starvla][INFO] - Step 2505, Loss: {'train/loss': 0.6118381023406982, 'train/grad_norm_pre_clip': 7.09285831451416, 'timing/data': 0.00042482593562453985, 'timing/model': 1.1931828130036592, 'learning_rate/qwen_vl_interface': 7.425702169751379e-06, 'learning_rate/action_model': 7.425702169751378e-05, 'epoch': 0.74}) +[2026-06-17 10:54:25,428][starVLA.training.train_starvla][INFO] - Step 2506, Loss: {'train/loss': 0.6515058875083923, 'train/grad_norm_pre_clip': 8.345867156982422, 'timing/data': 0.0003600700292736292, 'timing/model': 1.1939570349641144, 'learning_rate/qwen_vl_interface': 7.4237794631871965e-06, 'learning_rate/action_model': 7.423779463187196e-05, 'epoch': 0.74}) +[2026-06-17 10:54:30,165][starVLA.training.train_starvla][INFO] - Step 2507, Loss: {'train/loss': 0.5957547426223755, 'train/grad_norm_pre_clip': 7.23959493637085, 'timing/data': 0.0003670079167932272, 'timing/model': 1.3218589689349756, 'learning_rate/qwen_vl_interface': 7.42185630599619e-06, 'learning_rate/action_model': 7.42185630599619e-05, 'epoch': 0.74}) +[2026-06-17 10:54:34,787][starVLA.training.train_starvla][INFO] - Step 2508, Loss: {'train/loss': 0.7480701208114624, 'train/grad_norm_pre_clip': 8.88209056854248, 'timing/data': 0.0003656089538708329, 'timing/model': 1.2032664419384673, 'learning_rate/qwen_vl_interface': 7.419932698577036e-06, 'learning_rate/action_model': 7.419932698577035e-05, 'epoch': 0.74}) +[2026-06-17 10:54:39,415][starVLA.training.train_starvla][INFO] - Step 2509, Loss: {'train/loss': 0.6829858422279358, 'train/grad_norm_pre_clip': 8.88999080657959, 'timing/data': 0.00038278300780802965, 'timing/model': 1.196161902975291, 'learning_rate/qwen_vl_interface': 7.418008641328498e-06, 'learning_rate/action_model': 7.418008641328497e-05, 'epoch': 0.74}) +[2026-06-17 10:54:44,088][starVLA.training.train_starvla][INFO] - Step 2510, Loss: {'train/loss': 1.1856722831726074, 'train/grad_norm_pre_clip': 14.562735557556152, 'timing/data': 0.0003760980907827616, 'timing/model': 1.1995307200122625, 'learning_rate/qwen_vl_interface': 7.416084134649435e-06, 'learning_rate/action_model': 7.416084134649435e-05, 'epoch': 0.74}) +[2026-06-17 10:54:48,794][starVLA.training.train_starvla][INFO] - Step 2511, Loss: {'train/loss': 0.8648372888565063, 'train/grad_norm_pre_clip': 16.147701263427734, 'timing/data': 0.00036569300573319197, 'timing/model': 1.1986259700497612, 'learning_rate/qwen_vl_interface': 7.414159178938801e-06, 'learning_rate/action_model': 7.414159178938801e-05, 'epoch': 0.74}) +[2026-06-17 10:54:53,538][starVLA.training.train_starvla][INFO] - Step 2512, Loss: {'train/loss': 0.7527748942375183, 'train/grad_norm_pre_clip': 9.250353813171387, 'timing/data': 0.0003642779774963856, 'timing/model': 1.1983875370351598, 'learning_rate/qwen_vl_interface': 7.412233774595639e-06, 'learning_rate/action_model': 7.41223377459564e-05, 'epoch': 0.74}) +[2026-06-17 10:54:58,315][starVLA.training.train_starvla][INFO] - Step 2513, Loss: {'train/loss': 0.6979780793190002, 'train/grad_norm_pre_clip': 8.09566593170166, 'timing/data': 0.0004036780446767807, 'timing/model': 1.3338021209929138, 'learning_rate/qwen_vl_interface': 7.410307922019091e-06, 'learning_rate/action_model': 7.41030792201909e-05, 'epoch': 0.74}) +[2026-06-17 10:55:02,963][starVLA.training.train_starvla][INFO] - Step 2514, Loss: {'train/loss': 0.887650728225708, 'train/grad_norm_pre_clip': 8.927414894104004, 'timing/data': 0.0007382379844784737, 'timing/model': 1.2072820171015337, 'learning_rate/qwen_vl_interface': 7.408381621608387e-06, 'learning_rate/action_model': 7.408381621608387e-05, 'epoch': 0.74}) +[2026-06-17 10:55:07,733][starVLA.training.train_starvla][INFO] - Step 2515, Loss: {'train/loss': 0.66059410572052, 'train/grad_norm_pre_clip': 9.178532600402832, 'timing/data': 0.0003555819857865572, 'timing/model': 1.202552616945468, 'learning_rate/qwen_vl_interface': 7.406454873762848e-06, 'learning_rate/action_model': 7.406454873762848e-05, 'epoch': 0.74}) +[2026-06-17 10:55:12,518][starVLA.training.train_starvla][INFO] - Step 2516, Loss: {'train/loss': 0.7706744074821472, 'train/grad_norm_pre_clip': 9.517175674438477, 'timing/data': 0.0003519899910315871, 'timing/model': 1.2020831499248743, 'learning_rate/qwen_vl_interface': 7.404527678881894e-06, 'learning_rate/action_model': 7.404527678881894e-05, 'epoch': 0.74}) +[2026-06-17 10:55:17,291][starVLA.training.train_starvla][INFO] - Step 2517, Loss: {'train/loss': 0.8342087268829346, 'train/grad_norm_pre_clip': 12.234426498413086, 'timing/data': 0.0003624309320002794, 'timing/model': 1.2026160339592025, 'learning_rate/qwen_vl_interface': 7.402600037365034e-06, 'learning_rate/action_model': 7.402600037365034e-05, 'epoch': 0.74}) +[2026-06-17 10:55:22,063][starVLA.training.train_starvla][INFO] - Step 2518, Loss: {'train/loss': 0.9079023003578186, 'train/grad_norm_pre_clip': 10.676612854003906, 'timing/data': 0.00036117900162935257, 'timing/model': 1.331266307970509, 'learning_rate/qwen_vl_interface': 7.400671949611869e-06, 'learning_rate/action_model': 7.40067194961187e-05, 'epoch': 0.74}) +[2026-06-17 10:55:26,743][starVLA.training.train_starvla][INFO] - Step 2519, Loss: {'train/loss': 0.8392946720123291, 'train/grad_norm_pre_clip': 11.79815673828125, 'timing/data': 0.0003734750207513571, 'timing/model': 1.2350253749173135, 'learning_rate/qwen_vl_interface': 7.398743416022095e-06, 'learning_rate/action_model': 7.398743416022095e-05, 'epoch': 0.74}) +[2026-06-17 10:55:31,385][starVLA.training.train_starvla][INFO] - Step 2520, Loss: {'train/loss': 0.9410606622695923, 'train/grad_norm_pre_clip': 10.821014404296875, 'timing/data': 0.00035309302620589733, 'timing/model': 1.2034934120019898, 'learning_rate/qwen_vl_interface': 7.3968144369954995e-06, 'learning_rate/action_model': 7.396814436995499e-05, 'epoch': 0.74}) +[2026-06-17 10:55:36,102][starVLA.training.train_starvla][INFO] - Step 2521, Loss: {'train/loss': 0.8119121789932251, 'train/grad_norm_pre_clip': 9.599541664123535, 'timing/data': 0.00040898495353758335, 'timing/model': 1.2025818909751251, 'learning_rate/qwen_vl_interface': 7.39488501293196e-06, 'learning_rate/action_model': 7.39488501293196e-05, 'epoch': 0.74}) +[2026-06-17 10:55:40,872][starVLA.training.train_starvla][INFO] - Step 2522, Loss: {'train/loss': 0.8777742981910706, 'train/grad_norm_pre_clip': 11.608222007751465, 'timing/data': 0.00037878297735005617, 'timing/model': 1.2013630260480568, 'learning_rate/qwen_vl_interface': 7.392955144231452e-06, 'learning_rate/action_model': 7.392955144231452e-05, 'epoch': 0.74}) +[2026-06-17 10:55:45,693][starVLA.training.train_starvla][INFO] - Step 2523, Loss: {'train/loss': 0.661706805229187, 'train/grad_norm_pre_clip': 8.887422561645508, 'timing/data': 0.0003757709637284279, 'timing/model': 1.202389682061039, 'learning_rate/qwen_vl_interface': 7.391024831294034e-06, 'learning_rate/action_model': 7.391024831294033e-05, 'epoch': 0.74}) +[2026-06-17 10:55:50,486][starVLA.training.train_starvla][INFO] - Step 2524, Loss: {'train/loss': 0.7072941660881042, 'train/grad_norm_pre_clip': 7.486490249633789, 'timing/data': 0.00043395301327109337, 'timing/model': 1.341697779018432, 'learning_rate/qwen_vl_interface': 7.389094074519866e-06, 'learning_rate/action_model': 7.389094074519866e-05, 'epoch': 0.74}) +[2026-06-17 10:55:55,135][starVLA.training.train_starvla][INFO] - Step 2525, Loss: {'train/loss': 0.9027199745178223, 'train/grad_norm_pre_clip': 9.371304512023926, 'timing/data': 0.0003381789429113269, 'timing/model': 1.2030651960521936, 'learning_rate/qwen_vl_interface': 7.387162874309194e-06, 'learning_rate/action_model': 7.387162874309193e-05, 'epoch': 0.74}) +[2026-06-17 10:55:59,910][starVLA.training.train_starvla][INFO] - Step 2526, Loss: {'train/loss': 0.6627850532531738, 'train/grad_norm_pre_clip': 8.95341968536377, 'timing/data': 0.0003563410136848688, 'timing/model': 1.2020610240288079, 'learning_rate/qwen_vl_interface': 7.385231231062359e-06, 'learning_rate/action_model': 7.385231231062358e-05, 'epoch': 0.74}) +[2026-06-17 10:56:04,709][starVLA.training.train_starvla][INFO] - Step 2527, Loss: {'train/loss': 0.674035906791687, 'train/grad_norm_pre_clip': 10.402308464050293, 'timing/data': 0.00037952593993395567, 'timing/model': 1.2025193379959092, 'learning_rate/qwen_vl_interface': 7.383299145179789e-06, 'learning_rate/action_model': 7.383299145179789e-05, 'epoch': 0.74}) +[2026-06-17 10:56:09,528][starVLA.training.train_starvla][INFO] - Step 2528, Loss: {'train/loss': 0.8019250631332397, 'train/grad_norm_pre_clip': 12.272592544555664, 'timing/data': 0.0007222209824249148, 'timing/model': 1.2110388239379972, 'learning_rate/qwen_vl_interface': 7.381366617062014e-06, 'learning_rate/action_model': 7.381366617062014e-05, 'epoch': 0.74}) +[2026-06-17 10:56:14,416][starVLA.training.train_starvla][INFO] - Step 2529, Loss: {'train/loss': 0.8185694217681885, 'train/grad_norm_pre_clip': 13.731417655944824, 'timing/data': 0.00035241409204900265, 'timing/model': 1.2034585409564897, 'learning_rate/qwen_vl_interface': 7.379433647109645e-06, 'learning_rate/action_model': 7.379433647109645e-05, 'epoch': 0.74}) +[2026-06-17 10:56:19,195][starVLA.training.train_starvla][INFO] - Step 2530, Loss: {'train/loss': 0.5583329200744629, 'train/grad_norm_pre_clip': 8.69546890258789, 'timing/data': 0.0003724670968949795, 'timing/model': 1.338056717067957, 'learning_rate/qwen_vl_interface': 7.37750023572339e-06, 'learning_rate/action_model': 7.37750023572339e-05, 'epoch': 0.74}) +[2026-06-17 10:56:23,841][starVLA.training.train_starvla][INFO] - Step 2531, Loss: {'train/loss': 0.9397692680358887, 'train/grad_norm_pre_clip': 18.18871307373047, 'timing/data': 0.00035277800634503365, 'timing/model': 1.2034180390182883, 'learning_rate/qwen_vl_interface': 7.375566383304047e-06, 'learning_rate/action_model': 7.375566383304047e-05, 'epoch': 0.74}) +[2026-06-17 10:56:28,618][starVLA.training.train_starvla][INFO] - Step 2532, Loss: {'train/loss': 0.7241974472999573, 'train/grad_norm_pre_clip': 9.46261978149414, 'timing/data': 0.00039414200000464916, 'timing/model': 1.2014359759632498, 'learning_rate/qwen_vl_interface': 7.373632090252505e-06, 'learning_rate/action_model': 7.373632090252505e-05, 'epoch': 0.75}) +[2026-06-17 10:56:33,391][starVLA.training.train_starvla][INFO] - Step 2533, Loss: {'train/loss': 0.8919414281845093, 'train/grad_norm_pre_clip': 10.510165214538574, 'timing/data': 0.00040096999146044254, 'timing/model': 1.203249415033497, 'learning_rate/qwen_vl_interface': 7.371697356969746e-06, 'learning_rate/action_model': 7.371697356969746e-05, 'epoch': 0.75}) +[2026-06-17 10:56:38,161][starVLA.training.train_starvla][INFO] - Step 2534, Loss: {'train/loss': 0.753491222858429, 'train/grad_norm_pre_clip': 11.013887405395508, 'timing/data': 0.0003752330085262656, 'timing/model': 1.2035715100355446, 'learning_rate/qwen_vl_interface': 7.36976218385684e-06, 'learning_rate/action_model': 7.36976218385684e-05, 'epoch': 0.75}) +[2026-06-17 10:56:42,886][starVLA.training.train_starvla][INFO] - Step 2535, Loss: {'train/loss': 0.8381809592247009, 'train/grad_norm_pre_clip': 10.493680953979492, 'timing/data': 0.00040410703513771296, 'timing/model': 1.2035294950474054, 'learning_rate/qwen_vl_interface': 7.367826571314955e-06, 'learning_rate/action_model': 7.367826571314954e-05, 'epoch': 0.75}) +[2026-06-17 10:56:47,715][starVLA.training.train_starvla][INFO] - Step 2536, Loss: {'train/loss': 0.6666109561920166, 'train/grad_norm_pre_clip': 8.326952934265137, 'timing/data': 0.0003415109822526574, 'timing/model': 1.3506399929756299, 'learning_rate/qwen_vl_interface': 7.36589051974534e-06, 'learning_rate/action_model': 7.36589051974534e-05, 'epoch': 0.75}) +[2026-06-17 10:56:52,556][starVLA.training.train_starvla][INFO] - Step 2537, Loss: {'train/loss': 0.6965652108192444, 'train/grad_norm_pre_clip': 10.93044376373291, 'timing/data': 0.00037754199001938105, 'timing/model': 1.2125367280095816, 'learning_rate/qwen_vl_interface': 7.363954029549348e-06, 'learning_rate/action_model': 7.363954029549348e-05, 'epoch': 0.75}) +[2026-06-17 10:56:57,379][starVLA.training.train_starvla][INFO] - Step 2538, Loss: {'train/loss': 0.9988088607788086, 'train/grad_norm_pre_clip': 12.083819389343262, 'timing/data': 0.00037541904021054506, 'timing/model': 1.2034382009878755, 'learning_rate/qwen_vl_interface': 7.362017101128406e-06, 'learning_rate/action_model': 7.362017101128406e-05, 'epoch': 0.75}) +[2026-06-17 10:57:02,195][starVLA.training.train_starvla][INFO] - Step 2539, Loss: {'train/loss': 0.9859741926193237, 'train/grad_norm_pre_clip': 12.858302116394043, 'timing/data': 0.0003546320367604494, 'timing/model': 1.2105217929929495, 'learning_rate/qwen_vl_interface': 7.360079734884049e-06, 'learning_rate/action_model': 7.360079734884049e-05, 'epoch': 0.75}) +[2026-06-17 10:57:06,977][starVLA.training.train_starvla][INFO] - Step 2540, Loss: {'train/loss': 0.9040470123291016, 'train/grad_norm_pre_clip': 11.599873542785645, 'timing/data': 0.0003938430454581976, 'timing/model': 1.2026562920073047, 'learning_rate/qwen_vl_interface': 7.358141931217893e-06, 'learning_rate/action_model': 7.358141931217892e-05, 'epoch': 0.75}) +[2026-06-17 10:57:11,759][starVLA.training.train_starvla][INFO] - Step 2541, Loss: {'train/loss': 0.6723631620407104, 'train/grad_norm_pre_clip': 9.821569442749023, 'timing/data': 0.00035751296672970057, 'timing/model': 1.205214931978844, 'learning_rate/qwen_vl_interface': 7.356203690531643e-06, 'learning_rate/action_model': 7.356203690531643e-05, 'epoch': 0.75}) +[2026-06-17 10:57:16,536][starVLA.training.train_starvla][INFO] - Step 2542, Loss: {'train/loss': 0.8457410335540771, 'train/grad_norm_pre_clip': 11.429121017456055, 'timing/data': 0.00036979198921471834, 'timing/model': 1.3285400009481236, 'learning_rate/qwen_vl_interface': 7.354265013227104e-06, 'learning_rate/action_model': 7.354265013227104e-05, 'epoch': 0.75}) +[2026-06-17 10:57:21,191][starVLA.training.train_starvla][INFO] - Step 2543, Loss: {'train/loss': 0.7671849131584167, 'train/grad_norm_pre_clip': 10.184822082519531, 'timing/data': 0.0003543220227584243, 'timing/model': 1.2052398150553927, 'learning_rate/qwen_vl_interface': 7.352325899706163e-06, 'learning_rate/action_model': 7.352325899706162e-05, 'epoch': 0.75}) +[2026-06-17 10:57:25,968][starVLA.training.train_starvla][INFO] - Step 2544, Loss: {'train/loss': 0.6582257151603699, 'train/grad_norm_pre_clip': 8.693344116210938, 'timing/data': 0.0003690409939736128, 'timing/model': 1.2038040150655434, 'learning_rate/qwen_vl_interface': 7.350386350370802e-06, 'learning_rate/action_model': 7.350386350370801e-05, 'epoch': 0.75}) +[2026-06-17 10:57:30,746][starVLA.training.train_starvla][INFO] - Step 2545, Loss: {'train/loss': 0.5241091847419739, 'train/grad_norm_pre_clip': 7.118736267089844, 'timing/data': 0.00035943498369306326, 'timing/model': 1.2037302799290046, 'learning_rate/qwen_vl_interface': 7.348446365623089e-06, 'learning_rate/action_model': 7.348446365623088e-05, 'epoch': 0.75}) +[2026-06-17 10:57:35,556][starVLA.training.train_starvla][INFO] - Step 2546, Loss: {'train/loss': 0.8150208592414856, 'train/grad_norm_pre_clip': 8.389565467834473, 'timing/data': 0.00034836505074054003, 'timing/model': 1.2032102019293234, 'learning_rate/qwen_vl_interface': 7.346505945865186e-06, 'learning_rate/action_model': 7.346505945865186e-05, 'epoch': 0.75}) +[2026-06-17 10:57:40,347][starVLA.training.train_starvla][INFO] - Step 2547, Loss: {'train/loss': 0.7645987868309021, 'train/grad_norm_pre_clip': 8.892935752868652, 'timing/data': 0.0004204340511932969, 'timing/model': 1.3316699999850243, 'learning_rate/qwen_vl_interface': 7.344565091499345e-06, 'learning_rate/action_model': 7.344565091499344e-05, 'epoch': 0.75}) +[2026-06-17 10:57:45,059][starVLA.training.train_starvla][INFO] - Step 2548, Loss: {'train/loss': 0.7717859148979187, 'train/grad_norm_pre_clip': 12.470071792602539, 'timing/data': 0.00039349799044430256, 'timing/model': 1.2397463360102847, 'learning_rate/qwen_vl_interface': 7.342623802927907e-06, 'learning_rate/action_model': 7.342623802927907e-05, 'epoch': 0.75}) +[2026-06-17 10:57:49,711][starVLA.training.train_starvla][INFO] - Step 2549, Loss: {'train/loss': 0.6786596179008484, 'train/grad_norm_pre_clip': 8.71017837524414, 'timing/data': 0.0003476849524304271, 'timing/model': 1.2049994450062513, 'learning_rate/qwen_vl_interface': 7.340682080553303e-06, 'learning_rate/action_model': 7.340682080553303e-05, 'epoch': 0.75}) +[2026-06-17 10:57:54,441][starVLA.training.train_starvla][INFO] - Step 2550, Loss: {'train/loss': 1.0100517272949219, 'train/grad_norm_pre_clip': 9.451130867004395, 'timing/data': 0.0003724429989233613, 'timing/model': 1.2054522139951587, 'learning_rate/qwen_vl_interface': 7.338739924778053e-06, 'learning_rate/action_model': 7.338739924778053e-05, 'epoch': 0.75}) +[2026-06-17 10:57:59,220][starVLA.training.train_starvla][INFO] - Step 2551, Loss: {'train/loss': 0.8819502592086792, 'train/grad_norm_pre_clip': 10.618349075317383, 'timing/data': 0.000334674958139658, 'timing/model': 1.2043199829058722, 'learning_rate/qwen_vl_interface': 7.336797336004771e-06, 'learning_rate/action_model': 7.33679733600477e-05, 'epoch': 0.75}) +[2026-06-17 10:58:04,074][starVLA.training.train_starvla][INFO] - Step 2552, Loss: {'train/loss': 0.8486126661300659, 'train/grad_norm_pre_clip': 8.75393009185791, 'timing/data': 0.000354824005626142, 'timing/model': 1.2059359840350226, 'learning_rate/qwen_vl_interface': 7.334854314636153e-06, 'learning_rate/action_model': 7.334854314636152e-05, 'epoch': 0.75}) +[2026-06-17 10:58:08,918][starVLA.training.train_starvla][INFO] - Step 2553, Loss: {'train/loss': 0.8946669101715088, 'train/grad_norm_pre_clip': 11.750943183898926, 'timing/data': 0.0003649740247055888, 'timing/model': 1.3491708090296015, 'learning_rate/qwen_vl_interface': 7.3329108610749955e-06, 'learning_rate/action_model': 7.332910861074995e-05, 'epoch': 0.75}) +[2026-06-17 10:58:13,571][starVLA.training.train_starvla][INFO] - Step 2554, Loss: {'train/loss': 0.9372974634170532, 'train/grad_norm_pre_clip': 12.310182571411133, 'timing/data': 0.0003582980716601014, 'timing/model': 1.2039154910016805, 'learning_rate/qwen_vl_interface': 7.330966975724173e-06, 'learning_rate/action_model': 7.330966975724173e-05, 'epoch': 0.75}) +[2026-06-17 10:58:18,361][starVLA.training.train_starvla][INFO] - Step 2555, Loss: {'train/loss': 0.933571994304657, 'train/grad_norm_pre_clip': 12.487665176391602, 'timing/data': 0.0003460460575297475, 'timing/model': 1.2088023320538923, 'learning_rate/qwen_vl_interface': 7.329022658986657e-06, 'learning_rate/action_model': 7.329022658986658e-05, 'epoch': 0.75}) +[2026-06-17 10:58:23,159][starVLA.training.train_starvla][INFO] - Step 2556, Loss: {'train/loss': 0.7942708730697632, 'train/grad_norm_pre_clip': 11.218280792236328, 'timing/data': 0.00034483999479562044, 'timing/model': 1.204206258058548, 'learning_rate/qwen_vl_interface': 7.327077911265508e-06, 'learning_rate/action_model': 7.327077911265507e-05, 'epoch': 0.75}) +[2026-06-17 10:58:27,940][starVLA.training.train_starvla][INFO] - Step 2557, Loss: {'train/loss': 0.8317194581031799, 'train/grad_norm_pre_clip': 11.36844253540039, 'timing/data': 0.0003521270118653774, 'timing/model': 1.203977518947795, 'learning_rate/qwen_vl_interface': 7.325132732963872e-06, 'learning_rate/action_model': 7.325132732963872e-05, 'epoch': 0.75}) +[2026-06-17 10:58:32,745][starVLA.training.train_starvla][INFO] - Step 2558, Loss: {'train/loss': 0.5012658834457397, 'train/grad_norm_pre_clip': 7.86247444152832, 'timing/data': 0.0003715469501912594, 'timing/model': 1.2034052959643304, 'learning_rate/qwen_vl_interface': 7.323187124484988e-06, 'learning_rate/action_model': 7.323187124484988e-05, 'epoch': 0.75}) +[2026-06-17 10:58:37,541][starVLA.training.train_starvla][INFO] - Step 2559, Loss: {'train/loss': 0.8644582033157349, 'train/grad_norm_pre_clip': 10.093002319335938, 'timing/data': 0.0003454890102148056, 'timing/model': 1.3362512600142509, 'learning_rate/qwen_vl_interface': 7.3212410862321815e-06, 'learning_rate/action_model': 7.321241086232182e-05, 'epoch': 0.75}) +[2026-06-17 10:58:42,207][starVLA.training.train_starvla][INFO] - Step 2560, Loss: {'train/loss': 0.8212225437164307, 'train/grad_norm_pre_clip': 11.961081504821777, 'timing/data': 0.0004020839696750045, 'timing/model': 1.2060517970239744, 'learning_rate/qwen_vl_interface': 7.319294618608869e-06, 'learning_rate/action_model': 7.319294618608868e-05, 'epoch': 0.75}) +[2026-06-17 10:58:46,990][starVLA.training.train_starvla][INFO] - Step 2561, Loss: {'train/loss': 0.6442556381225586, 'train/grad_norm_pre_clip': 24.731689453125, 'timing/data': 0.0004120110534131527, 'timing/model': 1.2047361991135404, 'learning_rate/qwen_vl_interface': 7.317347722018557e-06, 'learning_rate/action_model': 7.317347722018557e-05, 'epoch': 0.75}) +[2026-06-17 10:58:51,806][starVLA.training.train_starvla][INFO] - Step 2562, Loss: {'train/loss': 0.8292564749717712, 'train/grad_norm_pre_clip': 10.096113204956055, 'timing/data': 0.0003718440420925617, 'timing/model': 1.2097418130142614, 'learning_rate/qwen_vl_interface': 7.3154003968648355e-06, 'learning_rate/action_model': 7.315400396864835e-05, 'epoch': 0.75}) +[2026-06-17 10:58:56,584][starVLA.training.train_starvla][INFO] - Step 2563, Loss: {'train/loss': 0.6150654554367065, 'train/grad_norm_pre_clip': 12.976563453674316, 'timing/data': 0.0004373131087049842, 'timing/model': 1.211440845974721, 'learning_rate/qwen_vl_interface': 7.313452643551392e-06, 'learning_rate/action_model': 7.313452643551392e-05, 'epoch': 0.75}) +[2026-06-17 10:59:01,310][starVLA.training.train_starvla][INFO] - Step 2564, Loss: {'train/loss': 0.8173869848251343, 'train/grad_norm_pre_clip': 11.03154468536377, 'timing/data': 0.0003648520214483142, 'timing/model': 1.2056723450077698, 'learning_rate/qwen_vl_interface': 7.311504462481993e-06, 'learning_rate/action_model': 7.311504462481993e-05, 'epoch': 0.75}) +[2026-06-17 10:59:06,087][starVLA.training.train_starvla][INFO] - Step 2565, Loss: {'train/loss': 0.5758053064346313, 'train/grad_norm_pre_clip': 9.76330280303955, 'timing/data': 0.00032597198151052, 'timing/model': 1.330681607942097, 'learning_rate/qwen_vl_interface': 7.3095558540605e-06, 'learning_rate/action_model': 7.309555854060499e-05, 'epoch': 0.75}) +[2026-06-17 10:59:10,768][starVLA.training.train_starvla][INFO] - Step 2566, Loss: {'train/loss': 0.7835999727249146, 'train/grad_norm_pre_clip': 9.579719543457031, 'timing/data': 0.00037033401895314455, 'timing/model': 1.2147930910577998, 'learning_rate/qwen_vl_interface': 7.307606818690865e-06, 'learning_rate/action_model': 7.307606818690864e-05, 'epoch': 0.76}) +[2026-06-17 10:59:15,578][starVLA.training.train_starvla][INFO] - Step 2567, Loss: {'train/loss': 0.7935071587562561, 'train/grad_norm_pre_clip': 10.218988418579102, 'timing/data': 0.00040446792263537645, 'timing/model': 1.2118009170517325, 'learning_rate/qwen_vl_interface': 7.30565735677712e-06, 'learning_rate/action_model': 7.305657356777119e-05, 'epoch': 0.76}) +[2026-06-17 10:59:20,357][starVLA.training.train_starvla][INFO] - Step 2568, Loss: {'train/loss': 0.5390494465827942, 'train/grad_norm_pre_clip': 7.237085819244385, 'timing/data': 0.0003589480184018612, 'timing/model': 1.2057259370340034, 'learning_rate/qwen_vl_interface': 7.303707468723395e-06, 'learning_rate/action_model': 7.303707468723395e-05, 'epoch': 0.76}) +[2026-06-17 10:59:25,131][starVLA.training.train_starvla][INFO] - Step 2569, Loss: {'train/loss': 0.5825645327568054, 'train/grad_norm_pre_clip': 8.14054012298584, 'timing/data': 0.0003546599764376879, 'timing/model': 1.2047010749811307, 'learning_rate/qwen_vl_interface': 7.301757154933899e-06, 'learning_rate/action_model': 7.3017571549339e-05, 'epoch': 0.76}) +[2026-06-17 10:59:29,896][starVLA.training.train_starvla][INFO] - Step 2570, Loss: {'train/loss': 0.609157383441925, 'train/grad_norm_pre_clip': 7.955887794494629, 'timing/data': 0.0005779649363830686, 'timing/model': 1.2042981310514733, 'learning_rate/qwen_vl_interface': 7.29980641581294e-06, 'learning_rate/action_model': 7.299806415812939e-05, 'epoch': 0.76}) +[2026-06-17 10:59:34,686][starVLA.training.train_starvla][INFO] - Step 2571, Loss: {'train/loss': 0.7458909153938293, 'train/grad_norm_pre_clip': 10.630584716796875, 'timing/data': 0.00037729297764599323, 'timing/model': 1.3323122629662976, 'learning_rate/qwen_vl_interface': 7.297855251764903e-06, 'learning_rate/action_model': 7.297855251764903e-05, 'epoch': 0.76}) +[2026-06-17 10:59:39,376][starVLA.training.train_starvla][INFO] - Step 2572, Loss: {'train/loss': 0.8915798664093018, 'train/grad_norm_pre_clip': 12.31567668914795, 'timing/data': 0.0003586970269680023, 'timing/model': 1.208491149940528, 'learning_rate/qwen_vl_interface': 7.295903663194269e-06, 'learning_rate/action_model': 7.29590366319427e-05, 'epoch': 0.76}) +[2026-06-17 10:59:44,190][starVLA.training.train_starvla][INFO] - Step 2573, Loss: {'train/loss': 0.6673708558082581, 'train/grad_norm_pre_clip': 8.9115571975708, 'timing/data': 0.00035669305361807346, 'timing/model': 1.2066115379566327, 'learning_rate/qwen_vl_interface': 7.293951650505606e-06, 'learning_rate/action_model': 7.293951650505607e-05, 'epoch': 0.76}) +[2026-06-17 10:59:49,011][starVLA.training.train_starvla][INFO] - Step 2574, Loss: {'train/loss': 0.6411508321762085, 'train/grad_norm_pre_clip': 9.281014442443848, 'timing/data': 0.000377572956494987, 'timing/model': 1.2126522139878944, 'learning_rate/qwen_vl_interface': 7.291999214103565e-06, 'learning_rate/action_model': 7.291999214103564e-05, 'epoch': 0.76}) +[2026-06-17 10:59:53,802][starVLA.training.train_starvla][INFO] - Step 2575, Loss: {'train/loss': 0.7787544131278992, 'train/grad_norm_pre_clip': 8.10901927947998, 'timing/data': 0.0003479910083115101, 'timing/model': 1.2162256808951497, 'learning_rate/qwen_vl_interface': 7.29004635439289e-06, 'learning_rate/action_model': 7.29004635439289e-05, 'epoch': 0.76}) +[2026-06-17 10:59:58,597][starVLA.training.train_starvla][INFO] - Step 2576, Loss: {'train/loss': 0.4934333264827728, 'train/grad_norm_pre_clip': 6.564239501953125, 'timing/data': 0.0003542399499565363, 'timing/model': 1.3328636209480464, 'learning_rate/qwen_vl_interface': 7.28809307177841e-06, 'learning_rate/action_model': 7.28809307177841e-05, 'epoch': 0.76}) +[2026-06-17 11:00:03,296][starVLA.training.train_starvla][INFO] - Step 2577, Loss: {'train/loss': 0.6995311975479126, 'train/grad_norm_pre_clip': 8.212508201599121, 'timing/data': 0.00042024196591228247, 'timing/model': 1.2427540440112352, 'learning_rate/qwen_vl_interface': 7.286139366665044e-06, 'learning_rate/action_model': 7.286139366665043e-05, 'epoch': 0.76}) +[2026-06-17 11:00:07,962][starVLA.training.train_starvla][INFO] - Step 2578, Loss: {'train/loss': 0.4976012408733368, 'train/grad_norm_pre_clip': 8.671246528625488, 'timing/data': 0.00034262496046721935, 'timing/model': 1.2062712900806218, 'learning_rate/qwen_vl_interface': 7.284185239457796e-06, 'learning_rate/action_model': 7.284185239457795e-05, 'epoch': 0.76}) +[2026-06-17 11:00:12,716][starVLA.training.train_starvla][INFO] - Step 2579, Loss: {'train/loss': 0.673625648021698, 'train/grad_norm_pre_clip': 19.412975311279297, 'timing/data': 0.0003776010125875473, 'timing/model': 1.210878803045489, 'learning_rate/qwen_vl_interface': 7.282230690561759e-06, 'learning_rate/action_model': 7.28223069056176e-05, 'epoch': 0.76}) +[2026-06-17 11:00:17,521][starVLA.training.train_starvla][INFO] - Step 2580, Loss: {'train/loss': 0.6595750451087952, 'train/grad_norm_pre_clip': 9.287893295288086, 'timing/data': 0.0003577120369300246, 'timing/model': 1.2107074919622391, 'learning_rate/qwen_vl_interface': 7.280275720382113e-06, 'learning_rate/action_model': 7.280275720382113e-05, 'epoch': 0.76}) +[2026-06-17 11:00:22,333][starVLA.training.train_starvla][INFO] - Step 2581, Loss: {'train/loss': 0.8461813926696777, 'train/grad_norm_pre_clip': 11.282954216003418, 'timing/data': 0.00033859896939247847, 'timing/model': 1.2038773340173066, 'learning_rate/qwen_vl_interface': 7.2783203293241246e-06, 'learning_rate/action_model': 7.278320329324125e-05, 'epoch': 0.76}) +[2026-06-17 11:00:27,165][starVLA.training.train_starvla][INFO] - Step 2582, Loss: {'train/loss': 0.8906623125076294, 'train/grad_norm_pre_clip': 10.501775741577148, 'timing/data': 0.00037850497756153345, 'timing/model': 1.3489165770588443, 'learning_rate/qwen_vl_interface': 7.276364517793152e-06, 'learning_rate/action_model': 7.276364517793151e-05, 'epoch': 0.76}) +[2026-06-17 11:00:31,828][starVLA.training.train_starvla][INFO] - Step 2583, Loss: {'train/loss': 0.7753956913948059, 'train/grad_norm_pre_clip': 9.119524955749512, 'timing/data': 0.00035590596962720156, 'timing/model': 1.20749760302715, 'learning_rate/qwen_vl_interface': 7.274408286194631e-06, 'learning_rate/action_model': 7.27440828619463e-05, 'epoch': 0.76}) +[2026-06-17 11:00:36,679][starVLA.training.train_starvla][INFO] - Step 2584, Loss: {'train/loss': 0.9659702777862549, 'train/grad_norm_pre_clip': 12.203136444091797, 'timing/data': 0.000560445012524724, 'timing/model': 1.2137445589760318, 'learning_rate/qwen_vl_interface': 7.272451634934096e-06, 'learning_rate/action_model': 7.272451634934096e-05, 'epoch': 0.76}) +[2026-06-17 11:00:41,621][starVLA.training.train_starvla][INFO] - Step 2585, Loss: {'train/loss': 0.7918938398361206, 'train/grad_norm_pre_clip': 10.191123008728027, 'timing/data': 0.0003595789894461632, 'timing/model': 1.2134699139278382, 'learning_rate/qwen_vl_interface': 7.270494564417161e-06, 'learning_rate/action_model': 7.270494564417161e-05, 'epoch': 0.76}) +[2026-06-17 11:00:46,446][starVLA.training.train_starvla][INFO] - Step 2586, Loss: {'train/loss': 0.9328180551528931, 'train/grad_norm_pre_clip': 10.167149543762207, 'timing/data': 0.0003608160186558962, 'timing/model': 1.2154621959198266, 'learning_rate/qwen_vl_interface': 7.268537075049527e-06, 'learning_rate/action_model': 7.268537075049526e-05, 'epoch': 0.76}) +[2026-06-17 11:00:51,263][starVLA.training.train_starvla][INFO] - Step 2587, Loss: {'train/loss': 0.43010103702545166, 'train/grad_norm_pre_clip': 11.760963439941406, 'timing/data': 0.00035574391949921846, 'timing/model': 1.2135060250293463, 'learning_rate/qwen_vl_interface': 7.2665791672369875e-06, 'learning_rate/action_model': 7.266579167236988e-05, 'epoch': 0.76}) +[2026-06-17 11:00:56,105][starVLA.training.train_starvla][INFO] - Step 2588, Loss: {'train/loss': 0.8076762557029724, 'train/grad_norm_pre_clip': 12.535675048828125, 'timing/data': 0.00036633294075727463, 'timing/model': 1.3565925339935347, 'learning_rate/qwen_vl_interface': 7.264620841385415e-06, 'learning_rate/action_model': 7.264620841385415e-05, 'epoch': 0.76}) +[2026-06-17 11:01:00,784][starVLA.training.train_starvla][INFO] - Step 2589, Loss: {'train/loss': 0.9624670743942261, 'train/grad_norm_pre_clip': 9.9417724609375, 'timing/data': 0.0003606959944590926, 'timing/model': 1.2123374450020492, 'learning_rate/qwen_vl_interface': 7.262662097900775e-06, 'learning_rate/action_model': 7.262662097900774e-05, 'epoch': 0.76}) +[2026-06-17 11:01:05,610][starVLA.training.train_starvla][INFO] - Step 2590, Loss: {'train/loss': 0.6900034546852112, 'train/grad_norm_pre_clip': 13.249338150024414, 'timing/data': 0.0004024709342047572, 'timing/model': 1.2205019820248708, 'learning_rate/qwen_vl_interface': 7.260702937189116e-06, 'learning_rate/action_model': 7.260702937189115e-05, 'epoch': 0.76}) +[2026-06-17 11:01:10,419][starVLA.training.train_starvla][INFO] - Step 2591, Loss: {'train/loss': 1.0602047443389893, 'train/grad_norm_pre_clip': 18.23349380493164, 'timing/data': 0.000344799947924912, 'timing/model': 1.2090511029819027, 'learning_rate/qwen_vl_interface': 7.258743359656574e-06, 'learning_rate/action_model': 7.258743359656573e-05, 'epoch': 0.76}) +[2026-06-17 11:01:15,260][starVLA.training.train_starvla][INFO] - Step 2592, Loss: {'train/loss': 0.7999688386917114, 'train/grad_norm_pre_clip': 11.724288940429688, 'timing/data': 0.00033306702971458435, 'timing/model': 1.2062667969148606, 'learning_rate/qwen_vl_interface': 7.2567833657093725e-06, 'learning_rate/action_model': 7.256783365709371e-05, 'epoch': 0.76}) +[2026-06-17 11:01:19,987][starVLA.training.train_starvla][INFO] - Step 2593, Loss: {'train/loss': 0.7683160305023193, 'train/grad_norm_pre_clip': 8.810173988342285, 'timing/data': 0.0003640430513769388, 'timing/model': 1.20288676104974, 'learning_rate/qwen_vl_interface': 7.254822955753821e-06, 'learning_rate/action_model': 7.25482295575382e-05, 'epoch': 0.76}) +[2026-06-17 11:01:24,766][starVLA.training.train_starvla][INFO] - Step 2594, Loss: {'train/loss': 1.000983715057373, 'train/grad_norm_pre_clip': 17.59637451171875, 'timing/data': 0.00037498900201171637, 'timing/model': 1.3320301460335031, 'learning_rate/qwen_vl_interface': 7.2528621301963135e-06, 'learning_rate/action_model': 7.252862130196313e-05, 'epoch': 0.76}) +[2026-06-17 11:01:29,416][starVLA.training.train_starvla][INFO] - Step 2595, Loss: {'train/loss': 0.8655589818954468, 'train/grad_norm_pre_clip': 16.73438835144043, 'timing/data': 0.0003929109079763293, 'timing/model': 1.204708441044204, 'learning_rate/qwen_vl_interface': 7.25090088944333e-06, 'learning_rate/action_model': 7.250900889443329e-05, 'epoch': 0.76}) +[2026-06-17 11:01:34,216][starVLA.training.train_starvla][INFO] - Step 2596, Loss: {'train/loss': 0.782960057258606, 'train/grad_norm_pre_clip': 10.591723442077637, 'timing/data': 0.00034848705399781466, 'timing/model': 1.2189022289821878, 'learning_rate/qwen_vl_interface': 7.248939233901438e-06, 'learning_rate/action_model': 7.248939233901439e-05, 'epoch': 0.76}) +[2026-06-17 11:01:39,003][starVLA.training.train_starvla][INFO] - Step 2597, Loss: {'train/loss': 1.2594265937805176, 'train/grad_norm_pre_clip': 12.070259094238281, 'timing/data': 0.0006456129485741258, 'timing/model': 1.20457002799958, 'learning_rate/qwen_vl_interface': 7.2469771639772944e-06, 'learning_rate/action_model': 7.246977163977294e-05, 'epoch': 0.76}) +[2026-06-17 11:01:43,779][starVLA.training.train_starvla][INFO] - Step 2598, Loss: {'train/loss': 0.6997880935668945, 'train/grad_norm_pre_clip': 8.998888969421387, 'timing/data': 0.00035030790604650974, 'timing/model': 1.2036845620023087, 'learning_rate/qwen_vl_interface': 7.245014680077633e-06, 'learning_rate/action_model': 7.245014680077633e-05, 'epoch': 0.76}) +[2026-06-17 11:01:48,539][starVLA.training.train_starvla][INFO] - Step 2599, Loss: {'train/loss': 0.4911036491394043, 'train/grad_norm_pre_clip': 10.302483558654785, 'timing/data': 0.00036083697341382504, 'timing/model': 1.2024665270000696, 'learning_rate/qwen_vl_interface': 7.243051782609285e-06, 'learning_rate/action_model': 7.243051782609285e-05, 'epoch': 0.76}) +[2026-06-17 11:03:30,345][starVLA.training.train_starvla][INFO] - Step 2600, Loss: {'train/loss': 0.6662542819976807, 'train/grad_norm_pre_clip': 9.097738265991211, 'eval/loss': 0.8022074103355408, 'eval/latency_0/loss': 0.8022074103355408, 'eval/demon_fire_precision': 0.9006109979633401, 'eval/demon_fire_recall': 0.9181893687707641, 'eval/demon_fire_f1': 0.9093152375077113, 'eval/demon_fire_rate_model': 0.6827030033370411, 'eval/demon_fire_rate_teacher': 0.6696329254727474, 'eval/demon_left_precision': 0.827652733118971, 'eval/demon_left_recall': 0.7291784702549575, 'eval/demon_left_f1': 0.775301204819277, 'eval/demon_left_rate_model': 0.4324249165739711, 'eval/demon_left_rate_teacher': 0.49082313681868744, 'eval/demon_right_precision': 0.7556427870461236, 'eval/demon_right_recall': 0.8700564971751412, 'eval/demon_right_f1': 0.8088235294117647, 'eval/demon_right_rate_model': 0.5667408231368187, 'eval/demon_right_rate_teacher': 0.4922135706340378, 'eval/demon_move_f1': 0.7920623671155209, 'eval/demon_cc_f1': 0.8706492941130782, 'eval/latency_0/demon_fire_precision': 0.9006109979633401, 'eval/latency_0/demon_fire_recall': 0.9181893687707641, 'eval/latency_0/demon_fire_f1': 0.9093152375077113, 'eval/latency_0/demon_fire_rate_model': 0.6827030033370411, 'eval/latency_0/demon_fire_rate_teacher': 0.6696329254727474, 'eval/latency_0/demon_left_precision': 0.827652733118971, 'eval/latency_0/demon_left_recall': 0.7291784702549575, 'eval/latency_0/demon_left_f1': 0.775301204819277, 'eval/latency_0/demon_left_rate_model': 0.4324249165739711, 'eval/latency_0/demon_left_rate_teacher': 0.49082313681868744, 'eval/latency_0/demon_right_precision': 0.7556427870461236, 'eval/latency_0/demon_right_recall': 0.8700564971751412, 'eval/latency_0/demon_right_f1': 0.8088235294117647, 'eval/latency_0/demon_right_rate_model': 0.5667408231368187, 'eval/latency_0/demon_right_rate_teacher': 0.4922135706340378, 'eval/latency_0/demon_move_f1': 0.7920623671155209, 'eval/latency_0/demon_cc_f1': 0.8706492941130782, 'timing/data': 0.0006957580335438251, 'timing/model': 1.3355296429945156, 'learning_rate/qwen_vl_interface': 7.241088471979155e-06, 'learning_rate/action_model': 7.241088471979155e-05, 'epoch': 0.77}) +[2026-06-17 11:03:30,346][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_2600_state +[2026-06-17 11:03:30,349][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 11:03:33,908][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_2600_state/model.safetensors +[2026-06-17 11:03:41,535][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_2600_state/optimizer.bin +[2026-06-17 11:03:41,536][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_2600_state/sampler.bin +[2026-06-17 11:03:41,536][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_2600_state/sampler_1.bin +[2026-06-17 11:03:41,536][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_2600_state/random_states_0.pkl +[2026-06-17 11:03:41,585][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2600 +[2026-06-17 11:03:41,586][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 11:03:41,623][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 11:03:46,406][starVLA.training.train_starvla][INFO] - Step 2601, Loss: {'train/loss': 0.6190279722213745, 'train/grad_norm_pre_clip': 7.580379009246826, 'timing/data': 0.00035104400012642145, 'timing/model': 1.3181912989821285, 'learning_rate/qwen_vl_interface': 7.239124748594244e-06, 'learning_rate/action_model': 7.239124748594244e-05, 'epoch': 0.77}) +[2026-06-17 11:03:51,022][starVLA.training.train_starvla][INFO] - Step 2602, Loss: {'train/loss': 0.6310542225837708, 'train/grad_norm_pre_clip': 8.388922691345215, 'timing/data': 0.00036116898991167545, 'timing/model': 1.2088999400148168, 'learning_rate/qwen_vl_interface': 7.237160612861632e-06, 'learning_rate/action_model': 7.237160612861632e-05, 'epoch': 0.77}) +[2026-06-17 11:03:55,631][starVLA.training.train_starvla][INFO] - Step 2603, Loss: {'train/loss': 0.7941111922264099, 'train/grad_norm_pre_clip': 9.794425010681152, 'timing/data': 0.00034923304338008165, 'timing/model': 1.1933323779376224, 'learning_rate/qwen_vl_interface': 7.235196065188485e-06, 'learning_rate/action_model': 7.235196065188485e-05, 'epoch': 0.77}) +[2026-06-17 11:04:00,334][starVLA.training.train_starvla][INFO] - Step 2604, Loss: {'train/loss': 0.6824268102645874, 'train/grad_norm_pre_clip': 8.956500053405762, 'timing/data': 0.0003737409133464098, 'timing/model': 1.1958844469627365, 'learning_rate/qwen_vl_interface': 7.23323110598206e-06, 'learning_rate/action_model': 7.23323110598206e-05, 'epoch': 0.77}) +[2026-06-17 11:04:05,024][starVLA.training.train_starvla][INFO] - Step 2605, Loss: {'train/loss': 0.7607904076576233, 'train/grad_norm_pre_clip': 10.712382316589355, 'timing/data': 0.00036555598489940166, 'timing/model': 1.1970166949322447, 'learning_rate/qwen_vl_interface': 7.231265735649691e-06, 'learning_rate/action_model': 7.231265735649691e-05, 'epoch': 0.77}) +[2026-06-17 11:04:09,772][starVLA.training.train_starvla][INFO] - Step 2606, Loss: {'train/loss': 0.9028909206390381, 'train/grad_norm_pre_clip': 10.059821128845215, 'timing/data': 0.0003636640030890703, 'timing/model': 1.1965432880679145, 'learning_rate/qwen_vl_interface': 7.229299954598805e-06, 'learning_rate/action_model': 7.229299954598804e-05, 'epoch': 0.77}) +[2026-06-17 11:04:14,526][starVLA.training.train_starvla][INFO] - Step 2607, Loss: {'train/loss': 0.8488546013832092, 'train/grad_norm_pre_clip': 10.243888854980469, 'timing/data': 0.00036104803439229727, 'timing/model': 1.3319403459317982, 'learning_rate/qwen_vl_interface': 7.227333763236907e-06, 'learning_rate/action_model': 7.227333763236906e-05, 'epoch': 0.77}) +[2026-06-17 11:04:19,159][starVLA.training.train_starvla][INFO] - Step 2608, Loss: {'train/loss': 0.9726066589355469, 'train/grad_norm_pre_clip': 10.394493103027344, 'timing/data': 0.00038125202991068363, 'timing/model': 1.1986900459742174, 'learning_rate/qwen_vl_interface': 7.225367161971592e-06, 'learning_rate/action_model': 7.225367161971591e-05, 'epoch': 0.77}) +[2026-06-17 11:04:23,917][starVLA.training.train_starvla][INFO] - Step 2609, Loss: {'train/loss': 0.8035175204277039, 'train/grad_norm_pre_clip': 9.058398246765137, 'timing/data': 0.0006304780254140496, 'timing/model': 1.1996377520263195, 'learning_rate/qwen_vl_interface': 7.223400151210539e-06, 'learning_rate/action_model': 7.223400151210538e-05, 'epoch': 0.77}) +[2026-06-17 11:04:28,680][starVLA.training.train_starvla][INFO] - Step 2610, Loss: {'train/loss': 0.7368966341018677, 'train/grad_norm_pre_clip': 19.06389045715332, 'timing/data': 0.0003881789743900299, 'timing/model': 1.1998004489578307, 'learning_rate/qwen_vl_interface': 7.2214327313615105e-06, 'learning_rate/action_model': 7.22143273136151e-05, 'epoch': 0.77}) +[2026-06-17 11:04:33,470][starVLA.training.train_starvla][INFO] - Step 2611, Loss: {'train/loss': 0.4791596829891205, 'train/grad_norm_pre_clip': 7.738241195678711, 'timing/data': 0.0003568889806047082, 'timing/model': 1.1994824419962242, 'learning_rate/qwen_vl_interface': 7.219464902832357e-06, 'learning_rate/action_model': 7.219464902832356e-05, 'epoch': 0.77}) +[2026-06-17 11:04:38,231][starVLA.training.train_starvla][INFO] - Step 2612, Loss: {'train/loss': 0.5414963364601135, 'train/grad_norm_pre_clip': 7.6667609214782715, 'timing/data': 0.0005752199795097113, 'timing/model': 1.197800722089596, 'learning_rate/qwen_vl_interface': 7.2174966660310085e-06, 'learning_rate/action_model': 7.217496666031008e-05, 'epoch': 0.77}) +[2026-06-17 11:04:42,995][starVLA.training.train_starvla][INFO] - Step 2613, Loss: {'train/loss': 0.5932992100715637, 'train/grad_norm_pre_clip': 12.2307710647583, 'timing/data': 0.00035828095860779285, 'timing/model': 1.330717617063783, 'learning_rate/qwen_vl_interface': 7.215528021365487e-06, 'learning_rate/action_model': 7.215528021365487e-05, 'epoch': 0.77}) +[2026-06-17 11:04:47,626][starVLA.training.train_starvla][INFO] - Step 2614, Loss: {'train/loss': 0.6279611587524414, 'train/grad_norm_pre_clip': 8.85887336730957, 'timing/data': 0.0003500889288261533, 'timing/model': 1.198853690060787, 'learning_rate/qwen_vl_interface': 7.213558969243888e-06, 'learning_rate/action_model': 7.213558969243888e-05, 'epoch': 0.77}) +[2026-06-17 11:04:52,384][starVLA.training.train_starvla][INFO] - Step 2615, Loss: {'train/loss': 0.7113553285598755, 'train/grad_norm_pre_clip': 8.50986385345459, 'timing/data': 0.00036136701237410307, 'timing/model': 1.1999023699900135, 'learning_rate/qwen_vl_interface': 7.211589510074404e-06, 'learning_rate/action_model': 7.211589510074404e-05, 'epoch': 0.77}) +[2026-06-17 11:04:57,164][starVLA.training.train_starvla][INFO] - Step 2616, Loss: {'train/loss': 0.6628310084342957, 'train/grad_norm_pre_clip': 8.715902328491211, 'timing/data': 0.00036091997753828764, 'timing/model': 1.201777926995419, 'learning_rate/qwen_vl_interface': 7.209619644265302e-06, 'learning_rate/action_model': 7.209619644265302e-05, 'epoch': 0.77}) +[2026-06-17 11:05:01,931][starVLA.training.train_starvla][INFO] - Step 2617, Loss: {'train/loss': 0.7018985748291016, 'train/grad_norm_pre_clip': 11.718948364257812, 'timing/data': 0.0003431179793551564, 'timing/model': 1.200444440008141, 'learning_rate/qwen_vl_interface': 7.20764937222494e-06, 'learning_rate/action_model': 7.20764937222494e-05, 'epoch': 0.77}) +[2026-06-17 11:05:06,725][starVLA.training.train_starvla][INFO] - Step 2618, Loss: {'train/loss': 0.5944612622261047, 'train/grad_norm_pre_clip': 12.054961204528809, 'timing/data': 0.0006041330052539706, 'timing/model': 1.3419681859668344, 'learning_rate/qwen_vl_interface': 7.205678694361757e-06, 'learning_rate/action_model': 7.205678694361757e-05, 'epoch': 0.77}) +[2026-06-17 11:05:11,470][starVLA.training.train_starvla][INFO] - Step 2619, Loss: {'train/loss': 0.7749220728874207, 'train/grad_norm_pre_clip': 10.625079154968262, 'timing/data': 0.0006039830623194575, 'timing/model': 1.2835231850622222, 'learning_rate/qwen_vl_interface': 7.203707611084276e-06, 'learning_rate/action_model': 7.203707611084275e-05, 'epoch': 0.77}) +[2026-06-17 11:05:16,116][starVLA.training.train_starvla][INFO] - Step 2620, Loss: {'train/loss': 0.9815252423286438, 'train/grad_norm_pre_clip': 20.972379684448242, 'timing/data': 0.00038314005360007286, 'timing/model': 1.2022947539808229, 'learning_rate/qwen_vl_interface': 7.2017361228011055e-06, 'learning_rate/action_model': 7.201736122801104e-05, 'epoch': 0.77}) +[2026-06-17 11:05:20,890][starVLA.training.train_starvla][INFO] - Step 2621, Loss: {'train/loss': 0.9107785224914551, 'train/grad_norm_pre_clip': 11.782910346984863, 'timing/data': 0.00037770799826830626, 'timing/model': 1.2010731920599937, 'learning_rate/qwen_vl_interface': 7.199764229920935e-06, 'learning_rate/action_model': 7.199764229920934e-05, 'epoch': 0.77}) +[2026-06-17 11:05:25,676][starVLA.training.train_starvla][INFO] - Step 2622, Loss: {'train/loss': 0.7783263921737671, 'train/grad_norm_pre_clip': 9.53049087524414, 'timing/data': 0.00036228110548108816, 'timing/model': 1.2066162059782073, 'learning_rate/qwen_vl_interface': 7.197791932852543e-06, 'learning_rate/action_model': 7.197791932852542e-05, 'epoch': 0.77}) +[2026-06-17 11:05:30,456][starVLA.training.train_starvla][INFO] - Step 2623, Loss: {'train/loss': 0.91799396276474, 'train/grad_norm_pre_clip': 10.571313858032227, 'timing/data': 0.0003561869962140918, 'timing/model': 1.203047889051959, 'learning_rate/qwen_vl_interface': 7.195819232004787e-06, 'learning_rate/action_model': 7.195819232004787e-05, 'epoch': 0.77}) +[2026-06-17 11:05:35,232][starVLA.training.train_starvla][INFO] - Step 2624, Loss: {'train/loss': 0.7582389116287231, 'train/grad_norm_pre_clip': 10.502288818359375, 'timing/data': 0.0003650879953056574, 'timing/model': 1.327495548990555, 'learning_rate/qwen_vl_interface': 7.1938461277866105e-06, 'learning_rate/action_model': 7.19384612778661e-05, 'epoch': 0.77}) +[2026-06-17 11:05:39,879][starVLA.training.train_starvla][INFO] - Step 2625, Loss: {'train/loss': 0.8940065503120422, 'train/grad_norm_pre_clip': 10.993474006652832, 'timing/data': 0.00033327704295516014, 'timing/model': 1.2086575540015474, 'learning_rate/qwen_vl_interface': 7.19187262060704e-06, 'learning_rate/action_model': 7.19187262060704e-05, 'epoch': 0.77}) +[2026-06-17 11:05:44,524][starVLA.training.train_starvla][INFO] - Step 2626, Loss: {'train/loss': 0.452950656414032, 'train/grad_norm_pre_clip': 7.823479175567627, 'timing/data': 0.0003556689480319619, 'timing/model': 1.201589728007093, 'learning_rate/qwen_vl_interface': 7.189898710875185e-06, 'learning_rate/action_model': 7.189898710875184e-05, 'epoch': 0.77}) +[2026-06-17 11:05:49,313][starVLA.training.train_starvla][INFO] - Step 2627, Loss: {'train/loss': 0.734122097492218, 'train/grad_norm_pre_clip': 11.120919227600098, 'timing/data': 0.000332775991410017, 'timing/model': 1.2031827860046178, 'learning_rate/qwen_vl_interface': 7.187924399000241e-06, 'learning_rate/action_model': 7.18792439900024e-05, 'epoch': 0.77}) +[2026-06-17 11:05:54,093][starVLA.training.train_starvla][INFO] - Step 2628, Loss: {'train/loss': 0.6126735210418701, 'train/grad_norm_pre_clip': 11.007136344909668, 'timing/data': 0.00036826194263994694, 'timing/model': 1.203389746020548, 'learning_rate/qwen_vl_interface': 7.185949685391483e-06, 'learning_rate/action_model': 7.185949685391483e-05, 'epoch': 0.77}) +[2026-06-17 11:05:58,870][starVLA.training.train_starvla][INFO] - Step 2629, Loss: {'train/loss': 1.056517243385315, 'train/grad_norm_pre_clip': 13.383506774902344, 'timing/data': 0.00036201393231749535, 'timing/model': 1.202618362964131, 'learning_rate/qwen_vl_interface': 7.183974570458272e-06, 'learning_rate/action_model': 7.183974570458272e-05, 'epoch': 0.77}) +[2026-06-17 11:06:03,644][starVLA.training.train_starvla][INFO] - Step 2630, Loss: {'train/loss': 0.7684025764465332, 'train/grad_norm_pre_clip': 16.083229064941406, 'timing/data': 0.0004621220286935568, 'timing/model': 1.32810312602669, 'learning_rate/qwen_vl_interface': 7.181999054610053e-06, 'learning_rate/action_model': 7.181999054610052e-05, 'epoch': 0.77}) +[2026-06-17 11:06:08,291][starVLA.training.train_starvla][INFO] - Step 2631, Loss: {'train/loss': 0.7861385345458984, 'train/grad_norm_pre_clip': 11.248577117919922, 'timing/data': 0.000392670975998044, 'timing/model': 1.2030110399937257, 'learning_rate/qwen_vl_interface': 7.180023138256349e-06, 'learning_rate/action_model': 7.180023138256348e-05, 'epoch': 0.77}) +[2026-06-17 11:06:13,056][starVLA.training.train_starvla][INFO] - Step 2632, Loss: {'train/loss': 0.9746600389480591, 'train/grad_norm_pre_clip': 17.566585540771484, 'timing/data': 0.0003568950342014432, 'timing/model': 1.2002805100055411, 'learning_rate/qwen_vl_interface': 7.178046821806771e-06, 'learning_rate/action_model': 7.17804682180677e-05, 'epoch': 0.77}) +[2026-06-17 11:06:17,737][starVLA.training.train_starvla][INFO] - Step 2633, Loss: {'train/loss': 0.7443955540657043, 'train/grad_norm_pre_clip': 9.69019603729248, 'timing/data': 0.0005654580891132355, 'timing/model': 1.2009201960172504, 'learning_rate/qwen_vl_interface': 7.176070105671016e-06, 'learning_rate/action_model': 7.176070105671015e-05, 'epoch': 0.77}) +[2026-06-17 11:06:22,450][starVLA.training.train_starvla][INFO] - Step 2634, Loss: {'train/loss': 0.6807059049606323, 'train/grad_norm_pre_clip': 10.271015167236328, 'timing/data': 0.00042203301563858986, 'timing/model': 1.2031336029758677, 'learning_rate/qwen_vl_interface': 7.174092990258852e-06, 'learning_rate/action_model': 7.174092990258851e-05, 'epoch': 0.78}) +[2026-06-17 11:06:27,208][starVLA.training.train_starvla][INFO] - Step 2635, Loss: {'train/loss': 1.0250869989395142, 'train/grad_norm_pre_clip': 14.389252662658691, 'timing/data': 0.0003888449864462018, 'timing/model': 1.204143854090944, 'learning_rate/qwen_vl_interface': 7.172115475980142e-06, 'learning_rate/action_model': 7.172115475980142e-05, 'epoch': 0.78}) +[2026-06-17 11:06:32,136][starVLA.training.train_starvla][INFO] - Step 2636, Loss: {'train/loss': 0.6537175178527832, 'train/grad_norm_pre_clip': 9.899065971374512, 'timing/data': 0.00036015594378113747, 'timing/model': 1.3432318690465763, 'learning_rate/qwen_vl_interface': 7.170137563244825e-06, 'learning_rate/action_model': 7.170137563244824e-05, 'epoch': 0.78}) +[2026-06-17 11:06:36,809][starVLA.training.train_starvla][INFO] - Step 2637, Loss: {'train/loss': 0.814658522605896, 'train/grad_norm_pre_clip': 11.773332595825195, 'timing/data': 0.0003723530098795891, 'timing/model': 1.202961663948372, 'learning_rate/qwen_vl_interface': 7.168159252462927e-06, 'learning_rate/action_model': 7.168159252462926e-05, 'epoch': 0.78}) +[2026-06-17 11:06:41,617][starVLA.training.train_starvla][INFO] - Step 2638, Loss: {'train/loss': 0.6007717251777649, 'train/grad_norm_pre_clip': 9.715299606323242, 'timing/data': 0.00035740400198847055, 'timing/model': 1.2033415499608964, 'learning_rate/qwen_vl_interface': 7.166180544044549e-06, 'learning_rate/action_model': 7.166180544044549e-05, 'epoch': 0.78}) +[2026-06-17 11:06:46,415][starVLA.training.train_starvla][INFO] - Step 2639, Loss: {'train/loss': 0.9071886539459229, 'train/grad_norm_pre_clip': 9.6923189163208, 'timing/data': 0.0003493879921734333, 'timing/model': 1.20375684206374, 'learning_rate/qwen_vl_interface': 7.164201438399884e-06, 'learning_rate/action_model': 7.164201438399884e-05, 'epoch': 0.78}) +[2026-06-17 11:06:51,200][starVLA.training.train_starvla][INFO] - Step 2640, Loss: {'train/loss': 0.4602604806423187, 'train/grad_norm_pre_clip': 6.812514781951904, 'timing/data': 0.00037319306284189224, 'timing/model': 1.2106226339237764, 'learning_rate/qwen_vl_interface': 7.1622219359392e-06, 'learning_rate/action_model': 7.162221935939199e-05, 'epoch': 0.78}) +[2026-06-17 11:06:55,996][starVLA.training.train_starvla][INFO] - Step 2641, Loss: {'train/loss': 0.533611536026001, 'train/grad_norm_pre_clip': 7.227332592010498, 'timing/data': 0.0003692921018227935, 'timing/model': 1.337764163967222, 'learning_rate/qwen_vl_interface': 7.160242037072852e-06, 'learning_rate/action_model': 7.160242037072851e-05, 'epoch': 0.78}) +[2026-06-17 11:07:00,732][starVLA.training.train_starvla][INFO] - Step 2642, Loss: {'train/loss': 0.8414257764816284, 'train/grad_norm_pre_clip': 11.446352005004883, 'timing/data': 0.00037528807297348976, 'timing/model': 1.2710859179496765, 'learning_rate/qwen_vl_interface': 7.158261742211275e-06, 'learning_rate/action_model': 7.158261742211274e-05, 'epoch': 0.78}) +[2026-06-17 11:07:05,400][starVLA.training.train_starvla][INFO] - Step 2643, Loss: {'train/loss': 0.8154294490814209, 'train/grad_norm_pre_clip': 10.815034866333008, 'timing/data': 0.000412218039855361, 'timing/model': 1.205815249006264, 'learning_rate/qwen_vl_interface': 7.156281051764983e-06, 'learning_rate/action_model': 7.156281051764983e-05, 'epoch': 0.78}) +[2026-06-17 11:07:10,142][starVLA.training.train_starvla][INFO] - Step 2644, Loss: {'train/loss': 1.0369764566421509, 'train/grad_norm_pre_clip': 11.670918464660645, 'timing/data': 0.0003576700109988451, 'timing/model': 1.2109239629935473, 'learning_rate/qwen_vl_interface': 7.15429996614458e-06, 'learning_rate/action_model': 7.15429996614458e-05, 'epoch': 0.78}) +[2026-06-17 11:07:14,946][starVLA.training.train_starvla][INFO] - Step 2645, Loss: {'train/loss': 1.0172264575958252, 'train/grad_norm_pre_clip': 10.457361221313477, 'timing/data': 0.0003683569375425577, 'timing/model': 1.2076741540804505, 'learning_rate/qwen_vl_interface': 7.152318485760743e-06, 'learning_rate/action_model': 7.152318485760742e-05, 'epoch': 0.78}) +[2026-06-17 11:07:19,735][starVLA.training.train_starvla][INFO] - Step 2646, Loss: {'train/loss': 0.722565233707428, 'train/grad_norm_pre_clip': 8.589972496032715, 'timing/data': 0.00034458492882549763, 'timing/model': 1.208813171950169, 'learning_rate/qwen_vl_interface': 7.150336611024239e-06, 'learning_rate/action_model': 7.150336611024239e-05, 'epoch': 0.78}) +[2026-06-17 11:07:24,519][starVLA.training.train_starvla][INFO] - Step 2647, Loss: {'train/loss': 0.5836114287376404, 'train/grad_norm_pre_clip': 9.7210693359375, 'timing/data': 0.0004666280001401901, 'timing/model': 1.3311090230708942, 'learning_rate/qwen_vl_interface': 7.1483543423459086e-06, 'learning_rate/action_model': 7.148354342345908e-05, 'epoch': 0.78}) +[2026-06-17 11:07:29,290][starVLA.training.train_starvla][INFO] - Step 2648, Loss: {'train/loss': 0.9049376249313354, 'train/grad_norm_pre_clip': 9.020450592041016, 'timing/data': 0.0003634900785982609, 'timing/model': 1.3275060229934752, 'learning_rate/qwen_vl_interface': 7.146371680136682e-06, 'learning_rate/action_model': 7.146371680136681e-05, 'epoch': 0.78}) +[2026-06-17 11:07:33,959][starVLA.training.train_starvla][INFO] - Step 2649, Loss: {'train/loss': 0.6909055709838867, 'train/grad_norm_pre_clip': 9.008660316467285, 'timing/data': 0.0003962418995797634, 'timing/model': 1.2055398699594662, 'learning_rate/qwen_vl_interface': 7.144388624807565e-06, 'learning_rate/action_model': 7.144388624807566e-05, 'epoch': 0.78}) +[2026-06-17 11:07:38,718][starVLA.training.train_starvla][INFO] - Step 2650, Loss: {'train/loss': 0.636182427406311, 'train/grad_norm_pre_clip': 8.317758560180664, 'timing/data': 0.00033480708952993155, 'timing/model': 1.2046200929908082, 'learning_rate/qwen_vl_interface': 7.142405176769648e-06, 'learning_rate/action_model': 7.142405176769648e-05, 'epoch': 0.78}) +[2026-06-17 11:07:43,539][starVLA.training.train_starvla][INFO] - Step 2651, Loss: {'train/loss': 0.7268656492233276, 'train/grad_norm_pre_clip': 7.927145481109619, 'timing/data': 0.0003864290192723274, 'timing/model': 1.2049737389897928, 'learning_rate/qwen_vl_interface': 7.140421336434102e-06, 'learning_rate/action_model': 7.140421336434102e-05, 'epoch': 0.78}) +[2026-06-17 11:07:48,332][starVLA.training.train_starvla][INFO] - Step 2652, Loss: {'train/loss': 0.6880218386650085, 'train/grad_norm_pre_clip': 9.776972770690918, 'timing/data': 0.000527304015122354, 'timing/model': 1.202500156010501, 'learning_rate/qwen_vl_interface': 7.138437104212176e-06, 'learning_rate/action_model': 7.138437104212176e-05, 'epoch': 0.78}) +[2026-06-17 11:07:53,117][starVLA.training.train_starvla][INFO] - Step 2653, Loss: {'train/loss': 0.8421387076377869, 'train/grad_norm_pre_clip': 13.201582908630371, 'timing/data': 0.0004106920678168535, 'timing/model': 1.3329320900375023, 'learning_rate/qwen_vl_interface': 7.136452480515208e-06, 'learning_rate/action_model': 7.136452480515208e-05, 'epoch': 0.78}) +[2026-06-17 11:07:57,767][starVLA.training.train_starvla][INFO] - Step 2654, Loss: {'train/loss': 0.6804634928703308, 'train/grad_norm_pre_clip': 9.718006134033203, 'timing/data': 0.00038321304600685835, 'timing/model': 1.2021575779654086, 'learning_rate/qwen_vl_interface': 7.134467465754612e-06, 'learning_rate/action_model': 7.134467465754612e-05, 'epoch': 0.78}) +[2026-06-17 11:08:02,565][starVLA.training.train_starvla][INFO] - Step 2655, Loss: {'train/loss': 0.7970303297042847, 'train/grad_norm_pre_clip': 9.698990821838379, 'timing/data': 0.00035932601895183325, 'timing/model': 1.20634188700933, 'learning_rate/qwen_vl_interface': 7.132482060341881e-06, 'learning_rate/action_model': 7.13248206034188e-05, 'epoch': 0.78}) +[2026-06-17 11:08:07,337][starVLA.training.train_starvla][INFO] - Step 2656, Loss: {'train/loss': 1.1102137565612793, 'train/grad_norm_pre_clip': 12.653120040893555, 'timing/data': 0.00033493898808956146, 'timing/model': 1.201611615018919, 'learning_rate/qwen_vl_interface': 7.1304962646885935e-06, 'learning_rate/action_model': 7.130496264688593e-05, 'epoch': 0.78}) +[2026-06-17 11:08:12,112][starVLA.training.train_starvla][INFO] - Step 2657, Loss: {'train/loss': 0.717533528804779, 'train/grad_norm_pre_clip': 8.4119291305542, 'timing/data': 0.00037320307455956936, 'timing/model': 1.204693753970787, 'learning_rate/qwen_vl_interface': 7.128510079206408e-06, 'learning_rate/action_model': 7.128510079206407e-05, 'epoch': 0.78}) +[2026-06-17 11:08:16,896][starVLA.training.train_starvla][INFO] - Step 2658, Loss: {'train/loss': 0.800922691822052, 'train/grad_norm_pre_clip': 9.701922416687012, 'timing/data': 0.000593362026847899, 'timing/model': 1.203546907985583, 'learning_rate/qwen_vl_interface': 7.126523504307061e-06, 'learning_rate/action_model': 7.126523504307061e-05, 'epoch': 0.78}) +[2026-06-17 11:08:21,654][starVLA.training.train_starvla][INFO] - Step 2659, Loss: {'train/loss': 0.9732348322868347, 'train/grad_norm_pre_clip': 10.417135238647461, 'timing/data': 0.0003614160232245922, 'timing/model': 1.3000345980981365, 'learning_rate/qwen_vl_interface': 7.124536540402371e-06, 'learning_rate/action_model': 7.124536540402372e-05, 'epoch': 0.78}) +[2026-06-17 11:08:26,301][starVLA.training.train_starvla][INFO] - Step 2660, Loss: {'train/loss': 0.8602144718170166, 'train/grad_norm_pre_clip': 10.129067420959473, 'timing/data': 0.0003416020190343261, 'timing/model': 1.2023522269446403, 'learning_rate/qwen_vl_interface': 7.1225491879042405e-06, 'learning_rate/action_model': 7.12254918790424e-05, 'epoch': 0.78}) +[2026-06-17 11:08:31,093][starVLA.training.train_starvla][INFO] - Step 2661, Loss: {'train/loss': 1.1159043312072754, 'train/grad_norm_pre_clip': 9.83924674987793, 'timing/data': 0.0005438880762085319, 'timing/model': 1.212654757895507, 'learning_rate/qwen_vl_interface': 7.120561447224649e-06, 'learning_rate/action_model': 7.120561447224649e-05, 'epoch': 0.78}) +[2026-06-17 11:08:35,880][starVLA.training.train_starvla][INFO] - Step 2662, Loss: {'train/loss': 0.7221555113792419, 'train/grad_norm_pre_clip': 7.553206920623779, 'timing/data': 0.0003969270037487149, 'timing/model': 1.2079476040089503, 'learning_rate/qwen_vl_interface': 7.118573318775656e-06, 'learning_rate/action_model': 7.118573318775655e-05, 'epoch': 0.78}) +[2026-06-17 11:08:40,671][starVLA.training.train_starvla][INFO] - Step 2663, Loss: {'train/loss': 0.5843708515167236, 'train/grad_norm_pre_clip': 9.082897186279297, 'timing/data': 0.0005088869947940111, 'timing/model': 1.2045979410177097, 'learning_rate/qwen_vl_interface': 7.116584802969403e-06, 'learning_rate/action_model': 7.116584802969402e-05, 'epoch': 0.78}) +[2026-06-17 11:08:45,450][starVLA.training.train_starvla][INFO] - Step 2664, Loss: {'train/loss': 0.8652197122573853, 'train/grad_norm_pre_clip': 10.501663208007812, 'timing/data': 0.00035714590921998024, 'timing/model': 1.3309979010373354, 'learning_rate/qwen_vl_interface': 7.114595900218112e-06, 'learning_rate/action_model': 7.114595900218112e-05, 'epoch': 0.78}) +[2026-06-17 11:08:50,114][starVLA.training.train_starvla][INFO] - Step 2665, Loss: {'train/loss': 0.8891870379447937, 'train/grad_norm_pre_clip': 8.999800682067871, 'timing/data': 0.0003703049151226878, 'timing/model': 1.2155057239579037, 'learning_rate/qwen_vl_interface': 7.112606610934084e-06, 'learning_rate/action_model': 7.112606610934084e-05, 'epoch': 0.78}) +[2026-06-17 11:08:54,767][starVLA.training.train_starvla][INFO] - Step 2666, Loss: {'train/loss': 0.7300992608070374, 'train/grad_norm_pre_clip': 8.613686561584473, 'timing/data': 0.0004063269589096308, 'timing/model': 1.2069535260088742, 'learning_rate/qwen_vl_interface': 7.110616935529702e-06, 'learning_rate/action_model': 7.110616935529702e-05, 'epoch': 0.78}) +[2026-06-17 11:08:59,464][starVLA.training.train_starvla][INFO] - Step 2667, Loss: {'train/loss': 0.9038851261138916, 'train/grad_norm_pre_clip': 7.929039478302002, 'timing/data': 0.0008637879509478807, 'timing/model': 1.2047332259826362, 'learning_rate/qwen_vl_interface': 7.108626874417426e-06, 'learning_rate/action_model': 7.108626874417426e-05, 'epoch': 0.78}) +[2026-06-17 11:09:04,182][starVLA.training.train_starvla][INFO] - Step 2668, Loss: {'train/loss': 0.7936671376228333, 'train/grad_norm_pre_clip': 43.343345642089844, 'timing/data': 0.0003545369254425168, 'timing/model': 1.2045741379261017, 'learning_rate/qwen_vl_interface': 7.106636428009801e-06, 'learning_rate/action_model': 7.1066364280098e-05, 'epoch': 0.79}) +[2026-06-17 11:09:08,943][starVLA.training.train_starvla][INFO] - Step 2669, Loss: {'train/loss': 0.5868752598762512, 'train/grad_norm_pre_clip': 7.4660444259643555, 'timing/data': 0.00035076693166047335, 'timing/model': 1.2069633449427783, 'learning_rate/qwen_vl_interface': 7.104645596719445e-06, 'learning_rate/action_model': 7.104645596719445e-05, 'epoch': 0.79}) +[2026-06-17 11:09:13,757][starVLA.training.train_starvla][INFO] - Step 2670, Loss: {'train/loss': 0.9641186594963074, 'train/grad_norm_pre_clip': 10.789361953735352, 'timing/data': 0.000448797014541924, 'timing/model': 1.354371375986375, 'learning_rate/qwen_vl_interface': 7.102654380959063e-06, 'learning_rate/action_model': 7.102654380959063e-05, 'epoch': 0.79}) +[2026-06-17 11:09:18,426][starVLA.training.train_starvla][INFO] - Step 2671, Loss: {'train/loss': 0.8570607304573059, 'train/grad_norm_pre_clip': 13.307920455932617, 'timing/data': 0.000668765977025032, 'timing/model': 1.2080573469866067, 'learning_rate/qwen_vl_interface': 7.100662781141432e-06, 'learning_rate/action_model': 7.100662781141432e-05, 'epoch': 0.79}) +[2026-06-17 11:09:23,228][starVLA.training.train_starvla][INFO] - Step 2672, Loss: {'train/loss': 1.0355379581451416, 'train/grad_norm_pre_clip': 9.979318618774414, 'timing/data': 0.0003591949352994561, 'timing/model': 1.2041467699455097, 'learning_rate/qwen_vl_interface': 7.0986707976794155e-06, 'learning_rate/action_model': 7.098670797679415e-05, 'epoch': 0.79}) +[2026-06-17 11:09:28,014][starVLA.training.train_starvla][INFO] - Step 2673, Loss: {'train/loss': 1.0236525535583496, 'train/grad_norm_pre_clip': 11.654465675354004, 'timing/data': 0.0003331269836053252, 'timing/model': 1.205689423950389, 'learning_rate/qwen_vl_interface': 7.096678430985955e-06, 'learning_rate/action_model': 7.096678430985954e-05, 'epoch': 0.79}) +[2026-06-17 11:09:32,794][starVLA.training.train_starvla][INFO] - Step 2674, Loss: {'train/loss': 0.780631959438324, 'train/grad_norm_pre_clip': 10.2528657913208, 'timing/data': 0.0003476149868220091, 'timing/model': 1.2045946680009365, 'learning_rate/qwen_vl_interface': 7.094685681474066e-06, 'learning_rate/action_model': 7.094685681474065e-05, 'epoch': 0.79}) +[2026-06-17 11:09:37,570][starVLA.training.train_starvla][INFO] - Step 2675, Loss: {'train/loss': 0.7157363891601562, 'train/grad_norm_pre_clip': 9.49444580078125, 'timing/data': 0.0003466119524091482, 'timing/model': 1.3290583429625258, 'learning_rate/qwen_vl_interface': 7.092692549556853e-06, 'learning_rate/action_model': 7.092692549556853e-05, 'epoch': 0.79}) +[2026-06-17 11:09:42,276][starVLA.training.train_starvla][INFO] - Step 2676, Loss: {'train/loss': 0.5877667665481567, 'train/grad_norm_pre_clip': 7.79183292388916, 'timing/data': 0.00035427301190793514, 'timing/model': 1.2416717479936779, 'learning_rate/qwen_vl_interface': 7.0906990356474895e-06, 'learning_rate/action_model': 7.090699035647489e-05, 'epoch': 0.79}) +[2026-06-17 11:09:46,938][starVLA.training.train_starvla][INFO] - Step 2677, Loss: {'train/loss': 0.7128956317901611, 'train/grad_norm_pre_clip': 8.73131275177002, 'timing/data': 0.0004206090234220028, 'timing/model': 1.20306549500674, 'learning_rate/qwen_vl_interface': 7.088705140159238e-06, 'learning_rate/action_model': 7.088705140159237e-05, 'epoch': 0.79}) +[2026-06-17 11:09:51,691][starVLA.training.train_starvla][INFO] - Step 2678, Loss: {'train/loss': 0.7480033040046692, 'train/grad_norm_pre_clip': 12.088797569274902, 'timing/data': 0.0003476820420473814, 'timing/model': 1.2101660950575024, 'learning_rate/qwen_vl_interface': 7.086710863505428e-06, 'learning_rate/action_model': 7.086710863505428e-05, 'epoch': 0.79}) +[2026-06-17 11:09:56,485][starVLA.training.train_starvla][INFO] - Step 2679, Loss: {'train/loss': 0.8338847160339355, 'train/grad_norm_pre_clip': 9.673345565795898, 'timing/data': 0.00031924108043313026, 'timing/model': 1.201806353055872, 'learning_rate/qwen_vl_interface': 7.084716206099485e-06, 'learning_rate/action_model': 7.084716206099484e-05, 'epoch': 0.79}) +[2026-06-17 11:10:01,280][starVLA.training.train_starvla][INFO] - Step 2680, Loss: {'train/loss': 0.6660963892936707, 'train/grad_norm_pre_clip': 9.379096031188965, 'timing/data': 0.000355313066393137, 'timing/model': 1.204660678980872, 'learning_rate/qwen_vl_interface': 7.082721168354894e-06, 'learning_rate/action_model': 7.082721168354894e-05, 'epoch': 0.79}) +[2026-06-17 11:10:06,087][starVLA.training.train_starvla][INFO] - Step 2681, Loss: {'train/loss': 0.7316932082176208, 'train/grad_norm_pre_clip': 9.439613342285156, 'timing/data': 0.0003372390056028962, 'timing/model': 1.3515660359989852, 'learning_rate/qwen_vl_interface': 7.080725750685235e-06, 'learning_rate/action_model': 7.080725750685234e-05, 'epoch': 0.79}) +[2026-06-17 11:10:10,747][starVLA.training.train_starvla][INFO] - Step 2682, Loss: {'train/loss': 0.7664126753807068, 'train/grad_norm_pre_clip': 9.078014373779297, 'timing/data': 0.00037586293183267117, 'timing/model': 1.2049395209178329, 'learning_rate/qwen_vl_interface': 7.078729953504157e-06, 'learning_rate/action_model': 7.078729953504156e-05, 'epoch': 0.79}) +[2026-06-17 11:10:15,625][starVLA.training.train_starvla][INFO] - Step 2683, Loss: {'train/loss': 0.72379469871521, 'train/grad_norm_pre_clip': 8.998147010803223, 'timing/data': 0.0007895899470895529, 'timing/model': 1.2039686880307272, 'learning_rate/qwen_vl_interface': 7.076733777225391e-06, 'learning_rate/action_model': 7.07673377722539e-05, 'epoch': 0.79}) +[2026-06-17 11:10:20,427][starVLA.training.train_starvla][INFO] - Step 2684, Loss: {'train/loss': 0.9542552828788757, 'train/grad_norm_pre_clip': 10.079816818237305, 'timing/data': 0.00033488101325929165, 'timing/model': 1.2058390760794282, 'learning_rate/qwen_vl_interface': 7.074737222262749e-06, 'learning_rate/action_model': 7.074737222262748e-05, 'epoch': 0.79}) +[2026-06-17 11:10:25,244][starVLA.training.train_starvla][INFO] - Step 2685, Loss: {'train/loss': 0.6080119609832764, 'train/grad_norm_pre_clip': 9.158818244934082, 'timing/data': 0.00036765600088983774, 'timing/model': 1.2034470430808142, 'learning_rate/qwen_vl_interface': 7.072740289030115e-06, 'learning_rate/action_model': 7.072740289030115e-05, 'epoch': 0.79}) +[2026-06-17 11:10:30,057][starVLA.training.train_starvla][INFO] - Step 2686, Loss: {'train/loss': 0.6419883370399475, 'train/grad_norm_pre_clip': 8.343253135681152, 'timing/data': 0.00047287391498684883, 'timing/model': 1.2075842970516533, 'learning_rate/qwen_vl_interface': 7.07074297794146e-06, 'learning_rate/action_model': 7.07074297794146e-05, 'epoch': 0.79}) +[2026-06-17 11:10:34,856][starVLA.training.train_starvla][INFO] - Step 2687, Loss: {'train/loss': 0.7053906321525574, 'train/grad_norm_pre_clip': 9.415916442871094, 'timing/data': 0.00033564900513738394, 'timing/model': 1.3354833170305938, 'learning_rate/qwen_vl_interface': 7.068745289410826e-06, 'learning_rate/action_model': 7.068745289410827e-05, 'epoch': 0.79}) +[2026-06-17 11:10:39,519][starVLA.training.train_starvla][INFO] - Step 2688, Loss: {'train/loss': 0.9462388157844543, 'train/grad_norm_pre_clip': 11.662690162658691, 'timing/data': 0.00033892004285007715, 'timing/model': 1.2059596700128168, 'learning_rate/qwen_vl_interface': 7.066747223852337e-06, 'learning_rate/action_model': 7.066747223852337e-05, 'epoch': 0.79}) +[2026-06-17 11:10:44,326][starVLA.training.train_starvla][INFO] - Step 2689, Loss: {'train/loss': 0.5832421779632568, 'train/grad_norm_pre_clip': 7.447483062744141, 'timing/data': 0.0004062911029905081, 'timing/model': 1.204216836951673, 'learning_rate/qwen_vl_interface': 7.064748781680192e-06, 'learning_rate/action_model': 7.064748781680192e-05, 'epoch': 0.79}) +[2026-06-17 11:10:49,138][starVLA.training.train_starvla][INFO] - Step 2690, Loss: {'train/loss': 0.9478460550308228, 'train/grad_norm_pre_clip': 10.868681907653809, 'timing/data': 0.00037929194513708353, 'timing/model': 1.2047782860463485, 'learning_rate/qwen_vl_interface': 7.062749963308673e-06, 'learning_rate/action_model': 7.062749963308672e-05, 'epoch': 0.79}) +[2026-06-17 11:10:53,928][starVLA.training.train_starvla][INFO] - Step 2691, Loss: {'train/loss': 0.5927437543869019, 'train/grad_norm_pre_clip': 7.901645660400391, 'timing/data': 0.0003175749443471432, 'timing/model': 1.2093624100089073, 'learning_rate/qwen_vl_interface': 7.060750769152135e-06, 'learning_rate/action_model': 7.060750769152135e-05, 'epoch': 0.79}) +[2026-06-17 11:10:58,656][starVLA.training.train_starvla][INFO] - Step 2692, Loss: {'train/loss': 0.7392506003379822, 'train/grad_norm_pre_clip': 7.670388698577881, 'timing/data': 0.00040015706326812506, 'timing/model': 1.2051951499888673, 'learning_rate/qwen_vl_interface': 7.058751199625014e-06, 'learning_rate/action_model': 7.058751199625015e-05, 'epoch': 0.79}) +[2026-06-17 11:11:03,467][starVLA.training.train_starvla][INFO] - Step 2693, Loss: {'train/loss': 0.9004895091056824, 'train/grad_norm_pre_clip': 10.260562896728516, 'timing/data': 0.00033264898229390383, 'timing/model': 1.3426746539771557, 'learning_rate/qwen_vl_interface': 7.056751255141824e-06, 'learning_rate/action_model': 7.056751255141824e-05, 'epoch': 0.79}) +[2026-06-17 11:11:08,129][starVLA.training.train_starvla][INFO] - Step 2694, Loss: {'train/loss': 0.6940887570381165, 'train/grad_norm_pre_clip': 11.58871078491211, 'timing/data': 0.00034430890809744596, 'timing/model': 1.2051582179265097, 'learning_rate/qwen_vl_interface': 7.054750936117155e-06, 'learning_rate/action_model': 7.054750936117156e-05, 'epoch': 0.79}) +[2026-06-17 11:11:12,906][starVLA.training.train_starvla][INFO] - Step 2695, Loss: {'train/loss': 0.703998327255249, 'train/grad_norm_pre_clip': 7.922219753265381, 'timing/data': 0.0003497410798445344, 'timing/model': 1.2041543039958924, 'learning_rate/qwen_vl_interface': 7.0527502429656745e-06, 'learning_rate/action_model': 7.052750242965674e-05, 'epoch': 0.79}) +[2026-06-17 11:11:17,695][starVLA.training.train_starvla][INFO] - Step 2696, Loss: {'train/loss': 0.7830315232276917, 'train/grad_norm_pre_clip': 8.204898834228516, 'timing/data': 0.0004048000555485487, 'timing/model': 1.2083710409933701, 'learning_rate/qwen_vl_interface': 7.050749176102128e-06, 'learning_rate/action_model': 7.050749176102129e-05, 'epoch': 0.79}) +[2026-06-17 11:11:22,470][starVLA.training.train_starvla][INFO] - Step 2697, Loss: {'train/loss': 1.227508544921875, 'train/grad_norm_pre_clip': 13.910942077636719, 'timing/data': 0.00034738797694444656, 'timing/model': 1.2049017560202628, 'learning_rate/qwen_vl_interface': 7.048747735941342e-06, 'learning_rate/action_model': 7.048747735941341e-05, 'epoch': 0.79}) +[2026-06-17 11:11:27,230][starVLA.training.train_starvla][INFO] - Step 2698, Loss: {'train/loss': 0.8637809753417969, 'train/grad_norm_pre_clip': 11.993873596191406, 'timing/data': 0.0004129890585318208, 'timing/model': 1.2041563430102542, 'learning_rate/qwen_vl_interface': 7.046745922898212e-06, 'learning_rate/action_model': 7.046745922898211e-05, 'epoch': 0.79}) +[2026-06-17 11:11:32,017][starVLA.training.train_starvla][INFO] - Step 2699, Loss: {'train/loss': 0.7841053605079651, 'train/grad_norm_pre_clip': 10.370643615722656, 'timing/data': 0.0003809769405052066, 'timing/model': 1.3383509000996128, 'learning_rate/qwen_vl_interface': 7.04474373738772e-06, 'learning_rate/action_model': 7.04474373738772e-05, 'epoch': 0.79}) +[2026-06-17 11:13:14,026][starVLA.training.train_starvla][INFO] - Step 2700, Loss: {'train/loss': 0.5649778246879578, 'train/grad_norm_pre_clip': 8.88620662689209, 'eval/loss': 0.8322875499725342, 'eval/latency_0/loss': 0.8322875499725342, 'eval/demon_fire_precision': 0.9125589369909987, 'eval/demon_fire_recall': 0.8841362126245847, 'eval/demon_fire_f1': 0.8981227589116221, 'eval/demon_fire_rate_model': 0.6487764182424917, 'eval/demon_fire_rate_teacher': 0.6696329254727474, 'eval/demon_left_precision': 0.7430488974113135, 'eval/demon_left_recall': 0.8781869688385269, 'eval/demon_left_f1': 0.8049857179953259, 'eval/demon_left_rate_model': 0.5800889877641824, 'eval/demon_left_rate_teacher': 0.49082313681868744, 'eval/demon_right_precision': 0.8455392809587217, 'eval/demon_right_recall': 0.7175141242937854, 'eval/demon_right_f1': 0.7762836185819071, 'eval/demon_right_rate_model': 0.41768631813125695, 'eval/demon_right_rate_teacher': 0.4922135706340378, 'eval/demon_move_f1': 0.7906346682886165, 'eval/demon_cc_f1': 0.8629278196799618, 'eval/latency_0/demon_fire_precision': 0.9125589369909987, 'eval/latency_0/demon_fire_recall': 0.8841362126245847, 'eval/latency_0/demon_fire_f1': 0.8981227589116221, 'eval/latency_0/demon_fire_rate_model': 0.6487764182424917, 'eval/latency_0/demon_fire_rate_teacher': 0.6696329254727474, 'eval/latency_0/demon_left_precision': 0.7430488974113135, 'eval/latency_0/demon_left_recall': 0.8781869688385269, 'eval/latency_0/demon_left_f1': 0.8049857179953259, 'eval/latency_0/demon_left_rate_model': 0.5800889877641824, 'eval/latency_0/demon_left_rate_teacher': 0.49082313681868744, 'eval/latency_0/demon_right_precision': 0.8455392809587217, 'eval/latency_0/demon_right_recall': 0.7175141242937854, 'eval/latency_0/demon_right_f1': 0.7762836185819071, 'eval/latency_0/demon_right_rate_model': 0.41768631813125695, 'eval/latency_0/demon_right_rate_teacher': 0.4922135706340378, 'eval/latency_0/demon_move_f1': 0.7906346682886165, 'eval/latency_0/demon_cc_f1': 0.8629278196799618, 'timing/data': 0.00034926796797662973, 'timing/model': 1.206519708968699, 'learning_rate/qwen_vl_interface': 7.042741179824919e-06, 'learning_rate/action_model': 7.042741179824918e-05, 'epoch': 0.79}) +[2026-06-17 11:13:14,027][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_2700_state +[2026-06-17 11:13:14,030][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 11:13:17,555][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_2700_state/model.safetensors +[2026-06-17 11:13:24,848][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_2700_state/optimizer.bin +[2026-06-17 11:13:24,849][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_2700_state/sampler.bin +[2026-06-17 11:13:24,849][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_2700_state/sampler_1.bin +[2026-06-17 11:13:24,849][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_2700_state/random_states_0.pkl +[2026-06-17 11:13:24,898][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2700 +[2026-06-17 11:13:24,898][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 11:13:24,926][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 11:13:29,564][starVLA.training.train_starvla][INFO] - Step 2701, Loss: {'train/loss': 0.7326185703277588, 'train/grad_norm_pre_clip': 8.727115631103516, 'timing/data': 0.00048583897296339273, 'timing/model': 1.2134635320398957, 'learning_rate/qwen_vl_interface': 7.040738250624944e-06, 'learning_rate/action_model': 7.040738250624944e-05, 'epoch': 0.79}) +[2026-06-17 11:13:34,158][starVLA.training.train_starvla][INFO] - Step 2702, Loss: {'train/loss': 0.976182222366333, 'train/grad_norm_pre_clip': 11.453866958618164, 'timing/data': 0.00036369904410094023, 'timing/model': 1.1887712569441646, 'learning_rate/qwen_vl_interface': 7.038734950203001e-06, 'learning_rate/action_model': 7.038734950203001e-05, 'epoch': 0.8}) +[2026-06-17 11:13:38,825][starVLA.training.train_starvla][INFO] - Step 2703, Loss: {'train/loss': 0.9309343695640564, 'train/grad_norm_pre_clip': 9.873833656311035, 'timing/data': 0.00035798700992017984, 'timing/model': 1.192489210050553, 'learning_rate/qwen_vl_interface': 7.036731278974379e-06, 'learning_rate/action_model': 7.036731278974378e-05, 'epoch': 0.8}) +[2026-06-17 11:13:43,540][starVLA.training.train_starvla][INFO] - Step 2704, Loss: {'train/loss': 0.5977820158004761, 'train/grad_norm_pre_clip': 9.374774932861328, 'timing/data': 0.00037694897036999464, 'timing/model': 1.1996714060660452, 'learning_rate/qwen_vl_interface': 7.034727237354442e-06, 'learning_rate/action_model': 7.034727237354441e-05, 'epoch': 0.8}) +[2026-06-17 11:13:48,292][starVLA.training.train_starvla][INFO] - Step 2705, Loss: {'train/loss': 0.589968740940094, 'train/grad_norm_pre_clip': 9.664857864379883, 'timing/data': 0.0003809920744970441, 'timing/model': 1.1946348539786413, 'learning_rate/qwen_vl_interface': 7.0327228257586244e-06, 'learning_rate/action_model': 7.032722825758625e-05, 'epoch': 0.8}) +[2026-06-17 11:13:53,055][starVLA.training.train_starvla][INFO] - Step 2706, Loss: {'train/loss': 1.0092905759811401, 'train/grad_norm_pre_clip': 10.786087036132812, 'timing/data': 0.00038056704215705395, 'timing/model': 1.3244063799502328, 'learning_rate/qwen_vl_interface': 7.030718044602451e-06, 'learning_rate/action_model': 7.03071804460245e-05, 'epoch': 0.8}) +[2026-06-17 11:13:57,674][starVLA.training.train_starvla][INFO] - Step 2707, Loss: {'train/loss': 0.7116712927818298, 'train/grad_norm_pre_clip': 8.026697158813477, 'timing/data': 0.000441401032730937, 'timing/model': 1.1959433179581538, 'learning_rate/qwen_vl_interface': 7.028712894301507e-06, 'learning_rate/action_model': 7.028712894301506e-05, 'epoch': 0.8}) +[2026-06-17 11:14:02,438][starVLA.training.train_starvla][INFO] - Step 2708, Loss: {'train/loss': 0.8234403133392334, 'train/grad_norm_pre_clip': 9.243828773498535, 'timing/data': 0.0005612981040030718, 'timing/model': 1.198488385998644, 'learning_rate/qwen_vl_interface': 7.026707375271469e-06, 'learning_rate/action_model': 7.026707375271468e-05, 'epoch': 0.8}) +[2026-06-17 11:14:07,162][starVLA.training.train_starvla][INFO] - Step 2709, Loss: {'train/loss': 0.5851486921310425, 'train/grad_norm_pre_clip': 9.238792419433594, 'timing/data': 0.00036419194657355547, 'timing/model': 1.2006803350523114, 'learning_rate/qwen_vl_interface': 7.024701487928079e-06, 'learning_rate/action_model': 7.024701487928078e-05, 'epoch': 0.8}) +[2026-06-17 11:14:11,924][starVLA.training.train_starvla][INFO] - Step 2710, Loss: {'train/loss': 0.8330401182174683, 'train/grad_norm_pre_clip': 8.741598129272461, 'timing/data': 0.0003603539662435651, 'timing/model': 1.197960525052622, 'learning_rate/qwen_vl_interface': 7.02269523268716e-06, 'learning_rate/action_model': 7.022695232687159e-05, 'epoch': 0.8}) +[2026-06-17 11:14:16,685][starVLA.training.train_starvla][INFO] - Step 2711, Loss: {'train/loss': 0.8427659273147583, 'train/grad_norm_pre_clip': 41.368839263916016, 'timing/data': 0.00036455399822443724, 'timing/model': 1.2014663009904325, 'learning_rate/qwen_vl_interface': 7.020688609964612e-06, 'learning_rate/action_model': 7.020688609964612e-05, 'epoch': 0.8}) +[2026-06-17 11:14:21,449][starVLA.training.train_starvla][INFO] - Step 2712, Loss: {'train/loss': 0.7393983006477356, 'train/grad_norm_pre_clip': 8.900212287902832, 'timing/data': 0.00038776593282818794, 'timing/model': 1.3278688149293885, 'learning_rate/qwen_vl_interface': 7.018681620176412e-06, 'learning_rate/action_model': 7.018681620176412e-05, 'epoch': 0.8}) +[2026-06-17 11:14:26,088][starVLA.training.train_starvla][INFO] - Step 2713, Loss: {'train/loss': 0.574449896812439, 'train/grad_norm_pre_clip': 7.545895576477051, 'timing/data': 0.00036704400554299355, 'timing/model': 1.2003805020358413, 'learning_rate/qwen_vl_interface': 7.016674263738606e-06, 'learning_rate/action_model': 7.016674263738607e-05, 'epoch': 0.8}) +[2026-06-17 11:14:30,855][starVLA.training.train_starvla][INFO] - Step 2714, Loss: {'train/loss': 0.726523756980896, 'train/grad_norm_pre_clip': 8.88991641998291, 'timing/data': 0.0005267889937385917, 'timing/model': 1.1984806079417467, 'learning_rate/qwen_vl_interface': 7.014666541067325e-06, 'learning_rate/action_model': 7.014666541067325e-05, 'epoch': 0.8}) +[2026-06-17 11:14:35,656][starVLA.training.train_starvla][INFO] - Step 2715, Loss: {'train/loss': 0.5604061484336853, 'train/grad_norm_pre_clip': 9.481770515441895, 'timing/data': 0.0006511380197480321, 'timing/model': 1.1987472829641774, 'learning_rate/qwen_vl_interface': 7.012658452578774e-06, 'learning_rate/action_model': 7.012658452578773e-05, 'epoch': 0.8}) +[2026-06-17 11:14:40,423][starVLA.training.train_starvla][INFO] - Step 2716, Loss: {'train/loss': 0.84234619140625, 'train/grad_norm_pre_clip': 9.980731010437012, 'timing/data': 0.0003390240017324686, 'timing/model': 1.201139886979945, 'learning_rate/qwen_vl_interface': 7.010649998689225e-06, 'learning_rate/action_model': 7.010649998689224e-05, 'epoch': 0.8}) +[2026-06-17 11:14:45,259][starVLA.training.train_starvla][INFO] - Step 2717, Loss: {'train/loss': 0.6646246314048767, 'train/grad_norm_pre_clip': 9.016985893249512, 'timing/data': 0.000356990029104054, 'timing/model': 1.1982016609981656, 'learning_rate/qwen_vl_interface': 7.0086411798150385e-06, 'learning_rate/action_model': 7.008641179815038e-05, 'epoch': 0.8}) +[2026-06-17 11:14:50,035][starVLA.training.train_starvla][INFO] - Step 2718, Loss: {'train/loss': 0.7732052803039551, 'train/grad_norm_pre_clip': 10.710661888122559, 'timing/data': 0.00034479901660233736, 'timing/model': 1.3416388989426196, 'learning_rate/qwen_vl_interface': 7.006631996372642e-06, 'learning_rate/action_model': 7.006631996372643e-05, 'epoch': 0.8}) +[2026-06-17 11:14:54,677][starVLA.training.train_starvla][INFO] - Step 2719, Loss: {'train/loss': 0.7690994739532471, 'train/grad_norm_pre_clip': 9.679608345031738, 'timing/data': 0.00036565307527780533, 'timing/model': 1.2015698390314355, 'learning_rate/qwen_vl_interface': 7.004622448778545e-06, 'learning_rate/action_model': 7.004622448778544e-05, 'epoch': 0.8}) +[2026-06-17 11:14:59,450][starVLA.training.train_starvla][INFO] - Step 2720, Loss: {'train/loss': 0.9389349222183228, 'train/grad_norm_pre_clip': 12.413472175598145, 'timing/data': 0.00041974999476224184, 'timing/model': 1.2034913229290396, 'learning_rate/qwen_vl_interface': 7.0026125374493244e-06, 'learning_rate/action_model': 7.002612537449324e-05, 'epoch': 0.8}) +[2026-06-17 11:15:04,245][starVLA.training.train_starvla][INFO] - Step 2721, Loss: {'train/loss': 0.9164181351661682, 'train/grad_norm_pre_clip': 10.395105361938477, 'timing/data': 0.0003783089341595769, 'timing/model': 1.203779507894069, 'learning_rate/qwen_vl_interface': 7.000602262801639e-06, 'learning_rate/action_model': 7.000602262801638e-05, 'epoch': 0.8}) +[2026-06-17 11:15:09,034][starVLA.training.train_starvla][INFO] - Step 2722, Loss: {'train/loss': 0.7544811367988586, 'train/grad_norm_pre_clip': 9.802042961120605, 'timing/data': 0.0003765400033444166, 'timing/model': 1.2030124430311844, 'learning_rate/qwen_vl_interface': 6.9985916252522225e-06, 'learning_rate/action_model': 6.998591625252222e-05, 'epoch': 0.8}) +[2026-06-17 11:15:13,781][starVLA.training.train_starvla][INFO] - Step 2723, Loss: {'train/loss': 0.615341067314148, 'train/grad_norm_pre_clip': 8.946279525756836, 'timing/data': 0.0003588399849832058, 'timing/model': 1.201187878032215, 'learning_rate/qwen_vl_interface': 6.9965806252178785e-06, 'learning_rate/action_model': 6.996580625217877e-05, 'epoch': 0.8}) +[2026-06-17 11:15:18,566][starVLA.training.train_starvla][INFO] - Step 2724, Loss: {'train/loss': 0.8563004732131958, 'train/grad_norm_pre_clip': 9.724783897399902, 'timing/data': 0.00034937600139528513, 'timing/model': 1.3388796979561448, 'learning_rate/qwen_vl_interface': 6.994569263115491e-06, 'learning_rate/action_model': 6.994569263115491e-05, 'epoch': 0.8}) +[2026-06-17 11:15:23,216][starVLA.training.train_starvla][INFO] - Step 2725, Loss: {'train/loss': 0.7863263487815857, 'train/grad_norm_pre_clip': 10.18547248840332, 'timing/data': 0.0003821280552074313, 'timing/model': 1.208033152972348, 'learning_rate/qwen_vl_interface': 6.992557539362019e-06, 'learning_rate/action_model': 6.99255753936202e-05, 'epoch': 0.8}) +[2026-06-17 11:15:28,001][starVLA.training.train_starvla][INFO] - Step 2726, Loss: {'train/loss': 0.7634358406066895, 'train/grad_norm_pre_clip': 9.108233451843262, 'timing/data': 0.00036737509071826935, 'timing/model': 1.2067487710155547, 'learning_rate/qwen_vl_interface': 6.990545454374495e-06, 'learning_rate/action_model': 6.990545454374494e-05, 'epoch': 0.8}) +[2026-06-17 11:15:32,783][starVLA.training.train_starvla][INFO] - Step 2727, Loss: {'train/loss': 0.6672523617744446, 'train/grad_norm_pre_clip': 11.932783126831055, 'timing/data': 0.00035845302045345306, 'timing/model': 1.2033122059656307, 'learning_rate/qwen_vl_interface': 6.988533008570024e-06, 'learning_rate/action_model': 6.988533008570024e-05, 'epoch': 0.8}) +[2026-06-17 11:15:37,561][starVLA.training.train_starvla][INFO] - Step 2728, Loss: {'train/loss': 0.5888685584068298, 'train/grad_norm_pre_clip': 9.153818130493164, 'timing/data': 0.00037148408591747284, 'timing/model': 1.2028303010156378, 'learning_rate/qwen_vl_interface': 6.98652020236579e-06, 'learning_rate/action_model': 6.986520202365789e-05, 'epoch': 0.8}) +[2026-06-17 11:15:42,337][starVLA.training.train_starvla][INFO] - Step 2729, Loss: {'train/loss': 0.5131572484970093, 'train/grad_norm_pre_clip': 8.143881797790527, 'timing/data': 0.0004631810588762164, 'timing/model': 1.205949730006978, 'learning_rate/qwen_vl_interface': 6.98450703617905e-06, 'learning_rate/action_model': 6.98450703617905e-05, 'epoch': 0.8}) +[2026-06-17 11:15:47,128][starVLA.training.train_starvla][INFO] - Step 2730, Loss: {'train/loss': 0.8033102750778198, 'train/grad_norm_pre_clip': 42.27028274536133, 'timing/data': 0.0003960080211982131, 'timing/model': 1.3389568889979273, 'learning_rate/qwen_vl_interface': 6.982493510427134e-06, 'learning_rate/action_model': 6.982493510427133e-05, 'epoch': 0.8}) +[2026-06-17 11:15:51,784][starVLA.training.train_starvla][INFO] - Step 2731, Loss: {'train/loss': 1.1186485290527344, 'train/grad_norm_pre_clip': 20.90229034423828, 'timing/data': 0.0003619319759309292, 'timing/model': 1.2051006159745157, 'learning_rate/qwen_vl_interface': 6.980479625527449e-06, 'learning_rate/action_model': 6.980479625527448e-05, 'epoch': 0.8}) +[2026-06-17 11:15:56,594][starVLA.training.train_starvla][INFO] - Step 2732, Loss: {'train/loss': 0.7877259850502014, 'train/grad_norm_pre_clip': 10.699131965637207, 'timing/data': 0.0003506189677864313, 'timing/model': 1.2030326779931784, 'learning_rate/qwen_vl_interface': 6.978465381897477e-06, 'learning_rate/action_model': 6.978465381897477e-05, 'epoch': 0.8}) +[2026-06-17 11:16:01,380][starVLA.training.train_starvla][INFO] - Step 2733, Loss: {'train/loss': 0.9752838015556335, 'train/grad_norm_pre_clip': 9.593302726745605, 'timing/data': 0.0008601970039308071, 'timing/model': 1.2030779380584136, 'learning_rate/qwen_vl_interface': 6.97645077995477e-06, 'learning_rate/action_model': 6.97645077995477e-05, 'epoch': 0.8}) +[2026-06-17 11:16:06,154][starVLA.training.train_starvla][INFO] - Step 2734, Loss: {'train/loss': 0.8921917676925659, 'train/grad_norm_pre_clip': 11.85025405883789, 'timing/data': 0.0005601169541478157, 'timing/model': 1.2003495550015941, 'learning_rate/qwen_vl_interface': 6.97443582011696e-06, 'learning_rate/action_model': 6.974435820116959e-05, 'epoch': 0.8}) +[2026-06-17 11:16:10,933][starVLA.training.train_starvla][INFO] - Step 2735, Loss: {'train/loss': 0.5927290916442871, 'train/grad_norm_pre_clip': 9.26794719696045, 'timing/data': 0.00042251800186932087, 'timing/model': 1.3335490019526333, 'learning_rate/qwen_vl_interface': 6.972420502801748e-06, 'learning_rate/action_model': 6.972420502801747e-05, 'epoch': 0.8}) +[2026-06-17 11:16:15,619][starVLA.training.train_starvla][INFO] - Step 2736, Loss: {'train/loss': 0.6741281151771545, 'train/grad_norm_pre_clip': 8.594047546386719, 'timing/data': 0.00040869403164833784, 'timing/model': 1.2436394409742206, 'learning_rate/qwen_vl_interface': 6.970404828426911e-06, 'learning_rate/action_model': 6.97040482842691e-05, 'epoch': 0.81}) +[2026-06-17 11:16:20,282][starVLA.training.train_starvla][INFO] - Step 2737, Loss: {'train/loss': 0.6716142296791077, 'train/grad_norm_pre_clip': 7.077814102172852, 'timing/data': 0.0003577970201149583, 'timing/model': 1.2022938739974052, 'learning_rate/qwen_vl_interface': 6.9683887974103044e-06, 'learning_rate/action_model': 6.968388797410305e-05, 'epoch': 0.81}) +[2026-06-17 11:16:25,035][starVLA.training.train_starvla][INFO] - Step 2738, Loss: {'train/loss': 0.7903392314910889, 'train/grad_norm_pre_clip': 12.424070358276367, 'timing/data': 0.00035832496359944344, 'timing/model': 1.2156484320294112, 'learning_rate/qwen_vl_interface': 6.966372410169848e-06, 'learning_rate/action_model': 6.966372410169848e-05, 'epoch': 0.81}) +[2026-06-17 11:16:29,832][starVLA.training.train_starvla][INFO] - Step 2739, Loss: {'train/loss': 0.7232450842857361, 'train/grad_norm_pre_clip': 9.144225120544434, 'timing/data': 0.0003607249818742275, 'timing/model': 1.2035388660151511, 'learning_rate/qwen_vl_interface': 6.964355667123548e-06, 'learning_rate/action_model': 6.964355667123547e-05, 'epoch': 0.81}) +[2026-06-17 11:16:34,623][starVLA.training.train_starvla][INFO] - Step 2740, Loss: {'train/loss': 0.7392425537109375, 'train/grad_norm_pre_clip': 8.806713104248047, 'timing/data': 0.00037107092794030905, 'timing/model': 1.2082475369097665, 'learning_rate/qwen_vl_interface': 6.96233856868947e-06, 'learning_rate/action_model': 6.962338568689469e-05, 'epoch': 0.81}) +[2026-06-17 11:16:39,422][starVLA.training.train_starvla][INFO] - Step 2741, Loss: {'train/loss': 0.8769277930259705, 'train/grad_norm_pre_clip': 10.193684577941895, 'timing/data': 0.0003688719589263201, 'timing/model': 1.349867420969531, 'learning_rate/qwen_vl_interface': 6.960321115285767e-06, 'learning_rate/action_model': 6.960321115285767e-05, 'epoch': 0.81}) +[2026-06-17 11:16:44,118][starVLA.training.train_starvla][INFO] - Step 2742, Loss: {'train/loss': 1.0735666751861572, 'train/grad_norm_pre_clip': 10.1896390914917, 'timing/data': 0.0003378320252522826, 'timing/model': 1.2075665100710467, 'learning_rate/qwen_vl_interface': 6.958303307330655e-06, 'learning_rate/action_model': 6.958303307330655e-05, 'epoch': 0.81}) +[2026-06-17 11:16:48,913][starVLA.training.train_starvla][INFO] - Step 2743, Loss: {'train/loss': 0.8656628727912903, 'train/grad_norm_pre_clip': 15.039834022521973, 'timing/data': 0.0003332099877297878, 'timing/model': 1.2085418740753084, 'learning_rate/qwen_vl_interface': 6.956285145242429e-06, 'learning_rate/action_model': 6.956285145242428e-05, 'epoch': 0.81}) +[2026-06-17 11:16:53,699][starVLA.training.train_starvla][INFO] - Step 2744, Loss: {'train/loss': 0.8491833806037903, 'train/grad_norm_pre_clip': 10.258430480957031, 'timing/data': 0.000372197013348341, 'timing/model': 1.2076285150833428, 'learning_rate/qwen_vl_interface': 6.954266629439456e-06, 'learning_rate/action_model': 6.954266629439457e-05, 'epoch': 0.81}) +[2026-06-17 11:16:58,478][starVLA.training.train_starvla][INFO] - Step 2745, Loss: {'train/loss': 0.8221315145492554, 'train/grad_norm_pre_clip': 25.10844612121582, 'timing/data': 0.0004344930639490485, 'timing/model': 1.2043672089930624, 'learning_rate/qwen_vl_interface': 6.952247760340176e-06, 'learning_rate/action_model': 6.952247760340176e-05, 'epoch': 0.81}) +[2026-06-17 11:17:03,281][starVLA.training.train_starvla][INFO] - Step 2746, Loss: {'train/loss': 0.776374101638794, 'train/grad_norm_pre_clip': 9.577457427978516, 'timing/data': 0.00039145490154623985, 'timing/model': 1.2062846720218658, 'learning_rate/qwen_vl_interface': 6.950228538363108e-06, 'learning_rate/action_model': 6.950228538363107e-05, 'epoch': 0.81}) +[2026-06-17 11:17:08,085][starVLA.training.train_starvla][INFO] - Step 2747, Loss: {'train/loss': 0.8182197213172913, 'train/grad_norm_pre_clip': 10.628338813781738, 'timing/data': 0.00040984107181429863, 'timing/model': 1.348363709053956, 'learning_rate/qwen_vl_interface': 6.948208963926829e-06, 'learning_rate/action_model': 6.94820896392683e-05, 'epoch': 0.81}) +[2026-06-17 11:17:12,756][starVLA.training.train_starvla][INFO] - Step 2748, Loss: {'train/loss': 0.783721923828125, 'train/grad_norm_pre_clip': 8.994636535644531, 'timing/data': 0.0003736920189112425, 'timing/model': 1.2020811290713027, 'learning_rate/qwen_vl_interface': 6.946189037450007e-06, 'learning_rate/action_model': 6.946189037450007e-05, 'epoch': 0.81}) +[2026-06-17 11:17:17,567][starVLA.training.train_starvla][INFO] - Step 2749, Loss: {'train/loss': 0.8968669772148132, 'train/grad_norm_pre_clip': 11.401941299438477, 'timing/data': 0.00037926307413727045, 'timing/model': 1.2091666050255299, 'learning_rate/qwen_vl_interface': 6.9441687593513715e-06, 'learning_rate/action_model': 6.944168759351372e-05, 'epoch': 0.81}) +[2026-06-17 11:17:22,367][starVLA.training.train_starvla][INFO] - Step 2750, Loss: {'train/loss': 0.5913366079330444, 'train/grad_norm_pre_clip': 8.175322532653809, 'timing/data': 0.00038654403761029243, 'timing/model': 1.204332065070048, 'learning_rate/qwen_vl_interface': 6.942148130049732e-06, 'learning_rate/action_model': 6.942148130049731e-05, 'epoch': 0.81}) +[2026-06-17 11:17:27,342][starVLA.training.train_starvla][INFO] - Step 2751, Loss: {'train/loss': 0.7118270993232727, 'train/grad_norm_pre_clip': 10.455998420715332, 'timing/data': 0.0004038719926029444, 'timing/model': 1.2019672190072015, 'learning_rate/qwen_vl_interface': 6.940127149963961e-06, 'learning_rate/action_model': 6.94012714996396e-05, 'epoch': 0.81}) +[2026-06-17 11:17:32,065][starVLA.training.train_starvla][INFO] - Step 2752, Loss: {'train/loss': 0.5322652459144592, 'train/grad_norm_pre_clip': 7.648962497711182, 'timing/data': 0.00042999594006687403, 'timing/model': 1.2023212490603328, 'learning_rate/qwen_vl_interface': 6.9381058195130146e-06, 'learning_rate/action_model': 6.938105819513014e-05, 'epoch': 0.81}) +[2026-06-17 11:17:36,849][starVLA.training.train_starvla][INFO] - Step 2753, Loss: {'train/loss': 0.6868781447410583, 'train/grad_norm_pre_clip': 8.323653221130371, 'timing/data': 0.0004591429606080055, 'timing/model': 1.3349783848971128, 'learning_rate/qwen_vl_interface': 6.936084139115915e-06, 'learning_rate/action_model': 6.936084139115915e-05, 'epoch': 0.81}) +[2026-06-17 11:17:41,498][starVLA.training.train_starvla][INFO] - Step 2754, Loss: {'train/loss': 0.657248854637146, 'train/grad_norm_pre_clip': 7.958954811096191, 'timing/data': 0.00037450704257935286, 'timing/model': 1.202915853005834, 'learning_rate/qwen_vl_interface': 6.934062109191759e-06, 'learning_rate/action_model': 6.934062109191759e-05, 'epoch': 0.81}) +[2026-06-17 11:17:46,291][starVLA.training.train_starvla][INFO] - Step 2755, Loss: {'train/loss': 0.7782572507858276, 'train/grad_norm_pre_clip': 9.735698699951172, 'timing/data': 0.00035228789784014225, 'timing/model': 1.2051088080042973, 'learning_rate/qwen_vl_interface': 6.9320397301597155e-06, 'learning_rate/action_model': 6.932039730159715e-05, 'epoch': 0.81}) +[2026-06-17 11:17:51,065][starVLA.training.train_starvla][INFO] - Step 2756, Loss: {'train/loss': 0.8835391402244568, 'train/grad_norm_pre_clip': 10.59650993347168, 'timing/data': 0.00042596401181071997, 'timing/model': 1.204810626921244, 'learning_rate/qwen_vl_interface': 6.930017002439027e-06, 'learning_rate/action_model': 6.930017002439027e-05, 'epoch': 0.81}) +[2026-06-17 11:17:55,861][starVLA.training.train_starvla][INFO] - Step 2757, Loss: {'train/loss': 1.1535756587982178, 'train/grad_norm_pre_clip': 12.250568389892578, 'timing/data': 0.000425187055952847, 'timing/model': 1.2102751720231026, 'learning_rate/qwen_vl_interface': 6.927993926449005e-06, 'learning_rate/action_model': 6.927993926449005e-05, 'epoch': 0.81}) +[2026-06-17 11:18:00,653][starVLA.training.train_starvla][INFO] - Step 2758, Loss: {'train/loss': 0.573384165763855, 'train/grad_norm_pre_clip': 6.954035758972168, 'timing/data': 0.00045904808212071657, 'timing/model': 1.210118280025199, 'learning_rate/qwen_vl_interface': 6.925970502609036e-06, 'learning_rate/action_model': 6.925970502609036e-05, 'epoch': 0.81}) +[2026-06-17 11:18:05,435][starVLA.training.train_starvla][INFO] - Step 2759, Loss: {'train/loss': 0.4915202856063843, 'train/grad_norm_pre_clip': 9.829222679138184, 'timing/data': 0.00040983897633850574, 'timing/model': 1.331336366944015, 'learning_rate/qwen_vl_interface': 6.92394673133858e-06, 'learning_rate/action_model': 6.92394673133858e-05, 'epoch': 0.81}) +[2026-06-17 11:18:10,100][starVLA.training.train_starvla][INFO] - Step 2760, Loss: {'train/loss': 0.47121214866638184, 'train/grad_norm_pre_clip': 7.641408443450928, 'timing/data': 0.0005768119590356946, 'timing/model': 1.2082762859063223, 'learning_rate/qwen_vl_interface': 6.921922613057165e-06, 'learning_rate/action_model': 6.921922613057164e-05, 'epoch': 0.81}) +[2026-06-17 11:18:14,919][starVLA.training.train_starvla][INFO] - Step 2761, Loss: {'train/loss': 0.9246700406074524, 'train/grad_norm_pre_clip': 11.620949745178223, 'timing/data': 0.00041548802983015776, 'timing/model': 1.206388246966526, 'learning_rate/qwen_vl_interface': 6.9198981481843915e-06, 'learning_rate/action_model': 6.919898148184392e-05, 'epoch': 0.81}) +[2026-06-17 11:18:19,727][starVLA.training.train_starvla][INFO] - Step 2762, Loss: {'train/loss': 0.7449331879615784, 'train/grad_norm_pre_clip': 8.95472240447998, 'timing/data': 0.0004277639091014862, 'timing/model': 1.2084424099884927, 'learning_rate/qwen_vl_interface': 6.917873337139936e-06, 'learning_rate/action_model': 6.917873337139937e-05, 'epoch': 0.81}) +[2026-06-17 11:18:24,529][starVLA.training.train_starvla][INFO] - Step 2763, Loss: {'train/loss': 0.7564488053321838, 'train/grad_norm_pre_clip': 9.768677711486816, 'timing/data': 0.00037903606425970793, 'timing/model': 1.2063428539549932, 'learning_rate/qwen_vl_interface': 6.915848180343545e-06, 'learning_rate/action_model': 6.915848180343545e-05, 'epoch': 0.81}) +[2026-06-17 11:18:29,375][starVLA.training.train_starvla][INFO] - Step 2764, Loss: {'train/loss': 0.6662890911102295, 'train/grad_norm_pre_clip': 11.081037521362305, 'timing/data': 0.00036229705438017845, 'timing/model': 1.340373249957338, 'learning_rate/qwen_vl_interface': 6.913822678215031e-06, 'learning_rate/action_model': 6.913822678215031e-05, 'epoch': 0.81}) +[2026-06-17 11:18:34,099][starVLA.training.train_starvla][INFO] - Step 2765, Loss: {'train/loss': 0.6433106064796448, 'train/grad_norm_pre_clip': 8.399309158325195, 'timing/data': 0.00035094004124403, 'timing/model': 1.2633943000109866, 'learning_rate/qwen_vl_interface': 6.911796831174287e-06, 'learning_rate/action_model': 6.911796831174286e-05, 'epoch': 0.81}) +[2026-06-17 11:18:38,774][starVLA.training.train_starvla][INFO] - Step 2766, Loss: {'train/loss': 1.0241608619689941, 'train/grad_norm_pre_clip': 11.456029891967773, 'timing/data': 0.0005089259939268231, 'timing/model': 1.210548089002259, 'learning_rate/qwen_vl_interface': 6.909770639641271e-06, 'learning_rate/action_model': 6.909770639641271e-05, 'epoch': 0.81}) +[2026-06-17 11:18:43,539][starVLA.training.train_starvla][INFO] - Step 2767, Loss: {'train/loss': 0.9891017079353333, 'train/grad_norm_pre_clip': 8.419661521911621, 'timing/data': 0.0004356378922238946, 'timing/model': 1.2109870710410178, 'learning_rate/qwen_vl_interface': 6.907744104036015e-06, 'learning_rate/action_model': 6.907744104036015e-05, 'epoch': 0.81}) +[2026-06-17 11:18:48,337][starVLA.training.train_starvla][INFO] - Step 2768, Loss: {'train/loss': 0.5660881400108337, 'train/grad_norm_pre_clip': 8.00680160522461, 'timing/data': 0.0005953770596534014, 'timing/model': 1.205458219978027, 'learning_rate/qwen_vl_interface': 6.905717224778623e-06, 'learning_rate/action_model': 6.905717224778622e-05, 'epoch': 0.81}) +[2026-06-17 11:18:53,129][starVLA.training.train_starvla][INFO] - Step 2769, Loss: {'train/loss': 0.7525205612182617, 'train/grad_norm_pre_clip': 9.308555603027344, 'timing/data': 0.0003522229380905628, 'timing/model': 1.202827065018937, 'learning_rate/qwen_vl_interface': 6.903690002289268e-06, 'learning_rate/action_model': 6.903690002289267e-05, 'epoch': 0.81}) +[2026-06-17 11:18:57,926][starVLA.training.train_starvla][INFO] - Step 2770, Loss: {'train/loss': 0.6243789792060852, 'train/grad_norm_pre_clip': 10.704895973205566, 'timing/data': 0.00039116910193115473, 'timing/model': 1.3314804410329089, 'learning_rate/qwen_vl_interface': 6.901662436988199e-06, 'learning_rate/action_model': 6.901662436988199e-05, 'epoch': 0.82}) +[2026-06-17 11:19:02,613][starVLA.training.train_starvla][INFO] - Step 2771, Loss: {'train/loss': 0.8372020721435547, 'train/grad_norm_pre_clip': 10.52220344543457, 'timing/data': 0.00035964895505458117, 'timing/model': 1.2083811990451068, 'learning_rate/qwen_vl_interface': 6.899634529295726e-06, 'learning_rate/action_model': 6.899634529295726e-05, 'epoch': 0.82}) +[2026-06-17 11:19:07,390][starVLA.training.train_starvla][INFO] - Step 2772, Loss: {'train/loss': 0.8631236553192139, 'train/grad_norm_pre_clip': 9.914227485656738, 'timing/data': 0.0003353230422362685, 'timing/model': 1.203678956022486, 'learning_rate/qwen_vl_interface': 6.897606279632241e-06, 'learning_rate/action_model': 6.897606279632241e-05, 'epoch': 0.82}) +[2026-06-17 11:19:12,179][starVLA.training.train_starvla][INFO] - Step 2773, Loss: {'train/loss': 0.9148502945899963, 'train/grad_norm_pre_clip': 10.600205421447754, 'timing/data': 0.0003451609518378973, 'timing/model': 1.2064251260599121, 'learning_rate/qwen_vl_interface': 6.895577688418202e-06, 'learning_rate/action_model': 6.895577688418202e-05, 'epoch': 0.82}) +[2026-06-17 11:19:16,979][starVLA.training.train_starvla][INFO] - Step 2774, Loss: {'train/loss': 1.0897458791732788, 'train/grad_norm_pre_clip': 14.049160957336426, 'timing/data': 0.00033008598256856203, 'timing/model': 1.203102245926857, 'learning_rate/qwen_vl_interface': 6.893548756074136e-06, 'learning_rate/action_model': 6.893548756074136e-05, 'epoch': 0.82}) +[2026-06-17 11:19:21,786][starVLA.training.train_starvla][INFO] - Step 2775, Loss: {'train/loss': 0.583290696144104, 'train/grad_norm_pre_clip': 9.87094783782959, 'timing/data': 0.0003653729800134897, 'timing/model': 1.2029103199020028, 'learning_rate/qwen_vl_interface': 6.891519483020647e-06, 'learning_rate/action_model': 6.891519483020647e-05, 'epoch': 0.82}) +[2026-06-17 11:19:26,627][starVLA.training.train_starvla][INFO] - Step 2776, Loss: {'train/loss': 0.7685900330543518, 'train/grad_norm_pre_clip': 9.10415267944336, 'timing/data': 0.0003504140768200159, 'timing/model': 1.3321710070595145, 'learning_rate/qwen_vl_interface': 6.8894898696784e-06, 'learning_rate/action_model': 6.889489869678399e-05, 'epoch': 0.82}) +[2026-06-17 11:19:31,286][starVLA.training.train_starvla][INFO] - Step 2777, Loss: {'train/loss': 0.782236635684967, 'train/grad_norm_pre_clip': 9.47547435760498, 'timing/data': 0.00036372803151607513, 'timing/model': 1.2029144029365852, 'learning_rate/qwen_vl_interface': 6.887459916468141e-06, 'learning_rate/action_model': 6.88745991646814e-05, 'epoch': 0.82}) +[2026-06-17 11:19:36,073][starVLA.training.train_starvla][INFO] - Step 2778, Loss: {'train/loss': 0.6829383373260498, 'train/grad_norm_pre_clip': 8.510674476623535, 'timing/data': 0.00041059497743844986, 'timing/model': 1.203743199002929, 'learning_rate/qwen_vl_interface': 6.885429623810678e-06, 'learning_rate/action_model': 6.885429623810677e-05, 'epoch': 0.82}) +[2026-06-17 11:19:40,958][starVLA.training.train_starvla][INFO] - Step 2779, Loss: {'train/loss': 0.5587487816810608, 'train/grad_norm_pre_clip': 8.92038631439209, 'timing/data': 0.0006883089663460851, 'timing/model': 1.222398811019957, 'learning_rate/qwen_vl_interface': 6.883398992126894e-06, 'learning_rate/action_model': 6.883398992126894e-05, 'epoch': 0.82}) +[2026-06-17 11:19:45,761][starVLA.training.train_starvla][INFO] - Step 2780, Loss: {'train/loss': 0.694566011428833, 'train/grad_norm_pre_clip': 8.95865535736084, 'timing/data': 0.00039223802741616964, 'timing/model': 1.2022129209944978, 'learning_rate/qwen_vl_interface': 6.8813680218377425e-06, 'learning_rate/action_model': 6.881368021837741e-05, 'epoch': 0.82}) +[2026-06-17 11:19:50,505][starVLA.training.train_starvla][INFO] - Step 2781, Loss: {'train/loss': 0.8404418230056763, 'train/grad_norm_pre_clip': 12.856310844421387, 'timing/data': 0.00039771408773958683, 'timing/model': 1.2070972690125927, 'learning_rate/qwen_vl_interface': 6.8793367133642445e-06, 'learning_rate/action_model': 6.879336713364243e-05, 'epoch': 0.82}) +[2026-06-17 11:19:55,324][starVLA.training.train_starvla][INFO] - Step 2782, Loss: {'train/loss': 0.8047746419906616, 'train/grad_norm_pre_clip': 10.27699089050293, 'timing/data': 0.00034234696067869663, 'timing/model': 1.3355313889915124, 'learning_rate/qwen_vl_interface': 6.877305067127492e-06, 'learning_rate/action_model': 6.877305067127492e-05, 'epoch': 0.82}) +[2026-06-17 11:19:59,987][starVLA.training.train_starvla][INFO] - Step 2783, Loss: {'train/loss': 0.6380575895309448, 'train/grad_norm_pre_clip': 8.004866600036621, 'timing/data': 0.00035308091901242733, 'timing/model': 1.2032799780135974, 'learning_rate/qwen_vl_interface': 6.875273083548648e-06, 'learning_rate/action_model': 6.875273083548648e-05, 'epoch': 0.82}) +[2026-06-17 11:20:04,778][starVLA.training.train_starvla][INFO] - Step 2784, Loss: {'train/loss': 0.7859275341033936, 'train/grad_norm_pre_clip': 7.91666841506958, 'timing/data': 0.0003612750442698598, 'timing/model': 1.2062735689105466, 'learning_rate/qwen_vl_interface': 6.873240763048945e-06, 'learning_rate/action_model': 6.873240763048945e-05, 'epoch': 0.82}) +[2026-06-17 11:20:09,570][starVLA.training.train_starvla][INFO] - Step 2785, Loss: {'train/loss': 0.7656028866767883, 'train/grad_norm_pre_clip': 11.174603462219238, 'timing/data': 0.0003963580820709467, 'timing/model': 1.2040407680906355, 'learning_rate/qwen_vl_interface': 6.871208106049685e-06, 'learning_rate/action_model': 6.871208106049685e-05, 'epoch': 0.82}) +[2026-06-17 11:20:14,339][starVLA.training.train_starvla][INFO] - Step 2786, Loss: {'train/loss': 0.6953640580177307, 'train/grad_norm_pre_clip': 11.11467456817627, 'timing/data': 0.00037577003240585327, 'timing/model': 1.203065415029414, 'learning_rate/qwen_vl_interface': 6.869175112972239e-06, 'learning_rate/action_model': 6.86917511297224e-05, 'epoch': 0.82}) +[2026-06-17 11:20:19,159][starVLA.training.train_starvla][INFO] - Step 2787, Loss: {'train/loss': 0.8951271772384644, 'train/grad_norm_pre_clip': 9.026225090026855, 'timing/data': 0.0003505449276417494, 'timing/model': 1.2026473610894755, 'learning_rate/qwen_vl_interface': 6.867141784238052e-06, 'learning_rate/action_model': 6.867141784238051e-05, 'epoch': 0.82}) +[2026-06-17 11:20:23,959][starVLA.training.train_starvla][INFO] - Step 2788, Loss: {'train/loss': 0.6887294054031372, 'train/grad_norm_pre_clip': 50.56752395629883, 'timing/data': 0.00037664303090423346, 'timing/model': 1.3474090269301087, 'learning_rate/qwen_vl_interface': 6.865108120268632e-06, 'learning_rate/action_model': 6.865108120268632e-05, 'epoch': 0.82}) +[2026-06-17 11:20:28,609][starVLA.training.train_starvla][INFO] - Step 2789, Loss: {'train/loss': 0.6428731083869934, 'train/grad_norm_pre_clip': 8.437355041503906, 'timing/data': 0.00034226698335260153, 'timing/model': 1.2046485679456964, 'learning_rate/qwen_vl_interface': 6.863074121485561e-06, 'learning_rate/action_model': 6.86307412148556e-05, 'epoch': 0.82}) +[2026-06-17 11:20:33,400][starVLA.training.train_starvla][INFO] - Step 2790, Loss: {'train/loss': 0.5869262218475342, 'train/grad_norm_pre_clip': 7.145466327667236, 'timing/data': 0.0003777950769290328, 'timing/model': 1.205318577005528, 'learning_rate/qwen_vl_interface': 6.861039788310489e-06, 'learning_rate/action_model': 6.861039788310489e-05, 'epoch': 0.82}) +[2026-06-17 11:20:38,193][starVLA.training.train_starvla][INFO] - Step 2791, Loss: {'train/loss': 0.5529961585998535, 'train/grad_norm_pre_clip': 8.281522750854492, 'timing/data': 0.0003556320443749428, 'timing/model': 1.2151395760010928, 'learning_rate/qwen_vl_interface': 6.859005121165135e-06, 'learning_rate/action_model': 6.859005121165135e-05, 'epoch': 0.82}) +[2026-06-17 11:20:43,006][starVLA.training.train_starvla][INFO] - Step 2792, Loss: {'train/loss': 0.8383256196975708, 'train/grad_norm_pre_clip': 12.295804977416992, 'timing/data': 0.0003250220324844122, 'timing/model': 1.2055653959978372, 'learning_rate/qwen_vl_interface': 6.856970120471289e-06, 'learning_rate/action_model': 6.856970120471289e-05, 'epoch': 0.82}) +[2026-06-17 11:20:47,815][starVLA.training.train_starvla][INFO] - Step 2793, Loss: {'train/loss': 0.7909326553344727, 'train/grad_norm_pre_clip': 9.060907363891602, 'timing/data': 0.00039317901246249676, 'timing/model': 1.3392894050339237, 'learning_rate/qwen_vl_interface': 6.854934786650805e-06, 'learning_rate/action_model': 6.854934786650806e-05, 'epoch': 0.82}) +[2026-06-17 11:20:52,534][starVLA.training.train_starvla][INFO] - Step 2794, Loss: {'train/loss': 0.7519250512123108, 'train/grad_norm_pre_clip': 16.581470489501953, 'timing/data': 0.00037828797940164804, 'timing/model': 1.253004149068147, 'learning_rate/qwen_vl_interface': 6.852899120125615e-06, 'learning_rate/action_model': 6.852899120125616e-05, 'epoch': 0.82}) +[2026-06-17 11:20:57,200][starVLA.training.train_starvla][INFO] - Step 2795, Loss: {'train/loss': 0.6170102953910828, 'train/grad_norm_pre_clip': 9.873466491699219, 'timing/data': 0.00033954589162021875, 'timing/model': 1.2113430378958583, 'learning_rate/qwen_vl_interface': 6.850863121317712e-06, 'learning_rate/action_model': 6.850863121317712e-05, 'epoch': 0.82}) +[2026-06-17 11:21:01,949][starVLA.training.train_starvla][INFO] - Step 2796, Loss: {'train/loss': 0.8044647574424744, 'train/grad_norm_pre_clip': 17.989290237426758, 'timing/data': 0.0004084990359842777, 'timing/model': 1.2096722449641675, 'learning_rate/qwen_vl_interface': 6.848826790649162e-06, 'learning_rate/action_model': 6.848826790649163e-05, 'epoch': 0.82}) +[2026-06-17 11:21:06,761][starVLA.training.train_starvla][INFO] - Step 2797, Loss: {'train/loss': 0.9973505735397339, 'train/grad_norm_pre_clip': 13.134117126464844, 'timing/data': 0.00043824396561831236, 'timing/model': 1.2175699120853096, 'learning_rate/qwen_vl_interface': 6.846790128542098e-06, 'learning_rate/action_model': 6.846790128542098e-05, 'epoch': 0.82}) +[2026-06-17 11:21:11,593][starVLA.training.train_starvla][INFO] - Step 2798, Loss: {'train/loss': 0.658602237701416, 'train/grad_norm_pre_clip': 11.136614799499512, 'timing/data': 0.0003858229611068964, 'timing/model': 1.2032281990395859, 'learning_rate/qwen_vl_interface': 6.8447531354187214e-06, 'learning_rate/action_model': 6.844753135418722e-05, 'epoch': 0.82}) +[2026-06-17 11:21:16,390][starVLA.training.train_starvla][INFO] - Step 2799, Loss: {'train/loss': 0.6530184745788574, 'train/grad_norm_pre_clip': 7.98649787902832, 'timing/data': 0.00036385690327733755, 'timing/model': 1.3301494790939614, 'learning_rate/qwen_vl_interface': 6.842715811701305e-06, 'learning_rate/action_model': 6.842715811701304e-05, 'epoch': 0.82}) +[2026-06-17 11:22:58,463][starVLA.training.train_starvla][INFO] - Step 2800, Loss: {'train/loss': 1.0587995052337646, 'train/grad_norm_pre_clip': 12.485339164733887, 'eval/loss': 0.805877149105072, 'eval/latency_0/loss': 0.805877149105072, 'eval/demon_fire_precision': 0.9202266782911944, 'eval/demon_fire_recall': 0.8766611295681063, 'eval/demon_fire_f1': 0.8979157805189281, 'eval/demon_fire_rate_model': 0.6379310344827587, 'eval/demon_fire_rate_teacher': 0.6696329254727474, 'eval/demon_left_precision': 0.8157582938388626, 'eval/demon_left_recall': 0.7801699716713881, 'eval/demon_left_f1': 0.7975673327541268, 'eval/demon_left_rate_model': 0.4694104560622914, 'eval/demon_left_rate_teacher': 0.49082313681868744, 'eval/demon_right_precision': 0.7845911949685535, 'eval/demon_right_recall': 0.8457627118644068, 'eval/demon_right_f1': 0.8140293637846656, 'eval/demon_right_rate_model': 0.5305895439377085, 'eval/demon_right_rate_teacher': 0.4922135706340378, 'eval/demon_move_f1': 0.8057983482693962, 'eval/demon_cc_f1': 0.8681424314722918, 'eval/latency_0/demon_fire_precision': 0.9202266782911944, 'eval/latency_0/demon_fire_recall': 0.8766611295681063, 'eval/latency_0/demon_fire_f1': 0.8979157805189281, 'eval/latency_0/demon_fire_rate_model': 0.6379310344827587, 'eval/latency_0/demon_fire_rate_teacher': 0.6696329254727474, 'eval/latency_0/demon_left_precision': 0.8157582938388626, 'eval/latency_0/demon_left_recall': 0.7801699716713881, 'eval/latency_0/demon_left_f1': 0.7975673327541268, 'eval/latency_0/demon_left_rate_model': 0.4694104560622914, 'eval/latency_0/demon_left_rate_teacher': 0.49082313681868744, 'eval/latency_0/demon_right_precision': 0.7845911949685535, 'eval/latency_0/demon_right_recall': 0.8457627118644068, 'eval/latency_0/demon_right_f1': 0.8140293637846656, 'eval/latency_0/demon_right_rate_model': 0.5305895439377085, 'eval/latency_0/demon_right_rate_teacher': 0.4922135706340378, 'eval/latency_0/demon_move_f1': 0.8057983482693962, 'eval/latency_0/demon_cc_f1': 0.8681424314722918, 'timing/data': 0.0006665610708296299, 'timing/model': 1.2139405950438231, 'learning_rate/qwen_vl_interface': 6.840678157812184e-06, 'learning_rate/action_model': 6.840678157812184e-05, 'epoch': 0.82}) +[2026-06-17 11:22:58,464][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_2800_state +[2026-06-17 11:22:58,467][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 11:23:02,009][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_2800_state/model.safetensors +[2026-06-17 11:23:09,300][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_2800_state/optimizer.bin +[2026-06-17 11:23:09,300][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_2800_state/sampler.bin +[2026-06-17 11:23:09,300][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_2800_state/sampler_1.bin +[2026-06-17 11:23:09,301][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_2800_state/random_states_0.pkl +[2026-06-17 11:23:09,350][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2800 +[2026-06-17 11:23:09,350][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 11:23:09,360][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 11:23:14,163][starVLA.training.train_starvla][INFO] - Step 2801, Loss: {'train/loss': 0.7841283679008484, 'train/grad_norm_pre_clip': 88.3719482421875, 'timing/data': 0.0003756439546123147, 'timing/model': 1.3185830330476165, 'learning_rate/qwen_vl_interface': 6.8386401741737725e-06, 'learning_rate/action_model': 6.838640174173773e-05, 'epoch': 0.82}) +[2026-06-17 11:23:18,761][starVLA.training.train_starvla][INFO] - Step 2802, Loss: {'train/loss': 0.6411582827568054, 'train/grad_norm_pre_clip': 8.725035667419434, 'timing/data': 0.0004979260265827179, 'timing/model': 1.1952337650582194, 'learning_rate/qwen_vl_interface': 6.836601861208542e-06, 'learning_rate/action_model': 6.836601861208541e-05, 'epoch': 0.82}) +[2026-06-17 11:23:23,513][starVLA.training.train_starvla][INFO] - Step 2803, Loss: {'train/loss': 0.6015013456344604, 'train/grad_norm_pre_clip': 7.309932231903076, 'timing/data': 0.0003565489314496517, 'timing/model': 1.2011351439869031, 'learning_rate/qwen_vl_interface': 6.834563219339039e-06, 'learning_rate/action_model': 6.834563219339038e-05, 'epoch': 0.82}) +[2026-06-17 11:23:28,223][starVLA.training.train_starvla][INFO] - Step 2804, Loss: {'train/loss': 0.6213236451148987, 'train/grad_norm_pre_clip': 9.256911277770996, 'timing/data': 0.000367421074770391, 'timing/model': 1.1932448260486126, 'learning_rate/qwen_vl_interface': 6.832524248987874e-06, 'learning_rate/action_model': 6.832524248987874e-05, 'epoch': 0.83}) +[2026-06-17 11:23:32,996][starVLA.training.train_starvla][INFO] - Step 2805, Loss: {'train/loss': 0.9518266320228577, 'train/grad_norm_pre_clip': 10.958781242370605, 'timing/data': 0.0004224400036036968, 'timing/model': 1.1943160409573466, 'learning_rate/qwen_vl_interface': 6.8304849505777314e-06, 'learning_rate/action_model': 6.830484950577732e-05, 'epoch': 0.83}) +[2026-06-17 11:23:37,754][starVLA.training.train_starvla][INFO] - Step 2806, Loss: {'train/loss': 0.6269444823265076, 'train/grad_norm_pre_clip': 8.24709701538086, 'timing/data': 0.00038582703564316034, 'timing/model': 1.2031860069837421, 'learning_rate/qwen_vl_interface': 6.828445324531357e-06, 'learning_rate/action_model': 6.828445324531357e-05, 'epoch': 0.83}) +[2026-06-17 11:23:42,508][starVLA.training.train_starvla][INFO] - Step 2807, Loss: {'train/loss': 0.6960590481758118, 'train/grad_norm_pre_clip': 7.675304412841797, 'timing/data': 0.00035975000355392694, 'timing/model': 1.3300718539394438, 'learning_rate/qwen_vl_interface': 6.826405371271567e-06, 'learning_rate/action_model': 6.826405371271567e-05, 'epoch': 0.83}) +[2026-06-17 11:23:47,147][starVLA.training.train_starvla][INFO] - Step 2808, Loss: {'train/loss': 0.5876981019973755, 'train/grad_norm_pre_clip': 9.3844633102417, 'timing/data': 0.0003648590063676238, 'timing/model': 1.2093859410379082, 'learning_rate/qwen_vl_interface': 6.82436509122125e-06, 'learning_rate/action_model': 6.82436509122125e-05, 'epoch': 0.83}) +[2026-06-17 11:23:51,906][starVLA.training.train_starvla][INFO] - Step 2809, Loss: {'train/loss': 0.9187048673629761, 'train/grad_norm_pre_clip': 10.160823822021484, 'timing/data': 0.0003593079745769501, 'timing/model': 1.1982443180168048, 'learning_rate/qwen_vl_interface': 6.822324484803353e-06, 'learning_rate/action_model': 6.822324484803353e-05, 'epoch': 0.83}) +[2026-06-17 11:23:56,639][starVLA.training.train_starvla][INFO] - Step 2810, Loss: {'train/loss': 0.5526617169380188, 'train/grad_norm_pre_clip': 9.604334831237793, 'timing/data': 0.0003766319714486599, 'timing/model': 1.1984215640695766, 'learning_rate/qwen_vl_interface': 6.8202835524409e-06, 'learning_rate/action_model': 6.8202835524409e-05, 'epoch': 0.83}) +[2026-06-17 11:24:01,288][starVLA.training.train_starvla][INFO] - Step 2811, Loss: {'train/loss': 1.2258217334747314, 'train/grad_norm_pre_clip': 14.246749877929688, 'timing/data': 0.0003751490730792284, 'timing/model': 1.1962112110340968, 'learning_rate/qwen_vl_interface': 6.818242294556977e-06, 'learning_rate/action_model': 6.818242294556977e-05, 'epoch': 0.83}) +[2026-06-17 11:24:05,970][starVLA.training.train_starvla][INFO] - Step 2812, Loss: {'train/loss': 0.6757906675338745, 'train/grad_norm_pre_clip': 8.604851722717285, 'timing/data': 0.0003587729297578335, 'timing/model': 1.1986046209931374, 'learning_rate/qwen_vl_interface': 6.816200711574738e-06, 'learning_rate/action_model': 6.816200711574738e-05, 'epoch': 0.83}) +[2026-06-17 11:24:10,685][starVLA.training.train_starvla][INFO] - Step 2813, Loss: {'train/loss': 0.5946381092071533, 'train/grad_norm_pre_clip': 11.618622779846191, 'timing/data': 0.00035490700975060463, 'timing/model': 1.2802737839519978, 'learning_rate/qwen_vl_interface': 6.814158803917406e-06, 'learning_rate/action_model': 6.814158803917406e-05, 'epoch': 0.83}) +[2026-06-17 11:24:15,329][starVLA.training.train_starvla][INFO] - Step 2814, Loss: {'train/loss': 0.9627828598022461, 'train/grad_norm_pre_clip': 11.02599048614502, 'timing/data': 0.0003621010109782219, 'timing/model': 1.2001776789547876, 'learning_rate/qwen_vl_interface': 6.812116572008271e-06, 'learning_rate/action_model': 6.812116572008271e-05, 'epoch': 0.83}) +[2026-06-17 11:24:20,095][starVLA.training.train_starvla][INFO] - Step 2815, Loss: {'train/loss': 0.5979455709457397, 'train/grad_norm_pre_clip': 8.578641891479492, 'timing/data': 0.00035801995545625687, 'timing/model': 1.2023254610830918, 'learning_rate/qwen_vl_interface': 6.8100740162706906e-06, 'learning_rate/action_model': 6.81007401627069e-05, 'epoch': 0.83}) +[2026-06-17 11:24:24,856][starVLA.training.train_starvla][INFO] - Step 2816, Loss: {'train/loss': 0.6452413201332092, 'train/grad_norm_pre_clip': 10.232824325561523, 'timing/data': 0.000340471975505352, 'timing/model': 1.1998228540178388, 'learning_rate/qwen_vl_interface': 6.808031137128087e-06, 'learning_rate/action_model': 6.808031137128086e-05, 'epoch': 0.83}) +[2026-06-17 11:24:29,646][starVLA.training.train_starvla][INFO] - Step 2817, Loss: {'train/loss': 0.8295533657073975, 'train/grad_norm_pre_clip': 11.181808471679688, 'timing/data': 0.0003574700094759464, 'timing/model': 1.2039698640583083, 'learning_rate/qwen_vl_interface': 6.805987935003951e-06, 'learning_rate/action_model': 6.805987935003951e-05, 'epoch': 0.83}) +[2026-06-17 11:24:34,437][starVLA.training.train_starvla][INFO] - Step 2818, Loss: {'train/loss': 1.2636585235595703, 'train/grad_norm_pre_clip': 13.071434020996094, 'timing/data': 0.0004130540182814002, 'timing/model': 1.3405624930746853, 'learning_rate/qwen_vl_interface': 6.803944410321843e-06, 'learning_rate/action_model': 6.803944410321844e-05, 'epoch': 0.83}) +[2026-06-17 11:24:39,094][starVLA.training.train_starvla][INFO] - Step 2819, Loss: {'train/loss': 0.8469281792640686, 'train/grad_norm_pre_clip': 10.444994926452637, 'timing/data': 0.00035515998024493456, 'timing/model': 1.2108205509139225, 'learning_rate/qwen_vl_interface': 6.801900563505388e-06, 'learning_rate/action_model': 6.801900563505388e-05, 'epoch': 0.83}) +[2026-06-17 11:24:43,750][starVLA.training.train_starvla][INFO] - Step 2820, Loss: {'train/loss': 1.1342144012451172, 'train/grad_norm_pre_clip': 14.848834037780762, 'timing/data': 0.00035113701596856117, 'timing/model': 1.2058884500293061, 'learning_rate/qwen_vl_interface': 6.7998563949782755e-06, 'learning_rate/action_model': 6.799856394978276e-05, 'epoch': 0.83}) +[2026-06-17 11:24:48,541][starVLA.training.train_starvla][INFO] - Step 2821, Loss: {'train/loss': 0.8295636773109436, 'train/grad_norm_pre_clip': 13.507408142089844, 'timing/data': 0.00038278091233223677, 'timing/model': 1.201915026991628, 'learning_rate/qwen_vl_interface': 6.797811905164263e-06, 'learning_rate/action_model': 6.797811905164263e-05, 'epoch': 0.83}) +[2026-06-17 11:24:53,315][starVLA.training.train_starvla][INFO] - Step 2822, Loss: {'train/loss': 0.9940086007118225, 'train/grad_norm_pre_clip': 13.367339134216309, 'timing/data': 0.0004298670683056116, 'timing/model': 1.201440827921033, 'learning_rate/qwen_vl_interface': 6.795767094487179e-06, 'learning_rate/action_model': 6.795767094487179e-05, 'epoch': 0.83}) +[2026-06-17 11:24:58,087][starVLA.training.train_starvla][INFO] - Step 2823, Loss: {'train/loss': 0.8144598007202148, 'train/grad_norm_pre_clip': 12.69696044921875, 'timing/data': 0.00040341797284781933, 'timing/model': 1.2019383369479328, 'learning_rate/qwen_vl_interface': 6.793721963370913e-06, 'learning_rate/action_model': 6.793721963370913e-05, 'epoch': 0.83}) +[2026-06-17 11:25:02,877][starVLA.training.train_starvla][INFO] - Step 2824, Loss: {'train/loss': 0.488254189491272, 'train/grad_norm_pre_clip': 7.434986591339111, 'timing/data': 0.00033841002732515335, 'timing/model': 1.3396611419739202, 'learning_rate/qwen_vl_interface': 6.7916765122394215e-06, 'learning_rate/action_model': 6.791676512239421e-05, 'epoch': 0.83}) +[2026-06-17 11:25:07,535][starVLA.training.train_starvla][INFO] - Step 2825, Loss: {'train/loss': 0.7926005125045776, 'train/grad_norm_pre_clip': 12.591243743896484, 'timing/data': 0.0003630670253187418, 'timing/model': 1.2053577560000122, 'learning_rate/qwen_vl_interface': 6.7896307415167305e-06, 'learning_rate/action_model': 6.78963074151673e-05, 'epoch': 0.83}) +[2026-06-17 11:25:12,313][starVLA.training.train_starvla][INFO] - Step 2826, Loss: {'train/loss': 0.8034182190895081, 'train/grad_norm_pre_clip': 10.67810344696045, 'timing/data': 0.00038950296584516764, 'timing/model': 1.2056846850318834, 'learning_rate/qwen_vl_interface': 6.787584651626929e-06, 'learning_rate/action_model': 6.78758465162693e-05, 'epoch': 0.83}) +[2026-06-17 11:25:17,255][starVLA.training.train_starvla][INFO] - Step 2827, Loss: {'train/loss': 0.9158594608306885, 'train/grad_norm_pre_clip': 13.301217079162598, 'timing/data': 0.0004191900370642543, 'timing/model': 1.2035192189505324, 'learning_rate/qwen_vl_interface': 6.785538242994177e-06, 'learning_rate/action_model': 6.785538242994177e-05, 'epoch': 0.83}) +[2026-06-17 11:25:22,043][starVLA.training.train_starvla][INFO] - Step 2828, Loss: {'train/loss': 0.5608394145965576, 'train/grad_norm_pre_clip': 10.585287094116211, 'timing/data': 0.0005619790172204375, 'timing/model': 1.2063868059776723, 'learning_rate/qwen_vl_interface': 6.783491516042694e-06, 'learning_rate/action_model': 6.783491516042693e-05, 'epoch': 0.83}) +[2026-06-17 11:25:26,920][starVLA.training.train_starvla][INFO] - Step 2829, Loss: {'train/loss': 0.6711836457252502, 'train/grad_norm_pre_clip': 10.196145057678223, 'timing/data': 0.00037363998126238585, 'timing/model': 1.2043972180690616, 'learning_rate/qwen_vl_interface': 6.781444471196768e-06, 'learning_rate/action_model': 6.781444471196769e-05, 'epoch': 0.83}) +[2026-06-17 11:25:31,717][starVLA.training.train_starvla][INFO] - Step 2830, Loss: {'train/loss': 0.6685429811477661, 'train/grad_norm_pre_clip': 9.136144638061523, 'timing/data': 0.0003654040629044175, 'timing/model': 1.332900898065418, 'learning_rate/qwen_vl_interface': 6.7793971088807575e-06, 'learning_rate/action_model': 6.779397108880757e-05, 'epoch': 0.83}) +[2026-06-17 11:25:36,381][starVLA.training.train_starvla][INFO] - Step 2831, Loss: {'train/loss': 0.8422145247459412, 'train/grad_norm_pre_clip': 10.088260650634766, 'timing/data': 0.0003679769579321146, 'timing/model': 1.206649666884914, 'learning_rate/qwen_vl_interface': 6.77734942951908e-06, 'learning_rate/action_model': 6.77734942951908e-05, 'epoch': 0.83}) +[2026-06-17 11:25:41,181][starVLA.training.train_starvla][INFO] - Step 2832, Loss: {'train/loss': 0.5954283475875854, 'train/grad_norm_pre_clip': 9.511570930480957, 'timing/data': 0.00035930704325437546, 'timing/model': 1.2042942639673129, 'learning_rate/qwen_vl_interface': 6.7753014335362224e-06, 'learning_rate/action_model': 6.775301433536222e-05, 'epoch': 0.83}) +[2026-06-17 11:25:45,884][starVLA.training.train_starvla][INFO] - Step 2833, Loss: {'train/loss': 0.6290700435638428, 'train/grad_norm_pre_clip': 8.999218940734863, 'timing/data': 0.00036175805144011974, 'timing/model': 1.2036507120355964, 'learning_rate/qwen_vl_interface': 6.773253121356735e-06, 'learning_rate/action_model': 6.773253121356735e-05, 'epoch': 0.83}) +[2026-06-17 11:25:50,618][starVLA.training.train_starvla][INFO] - Step 2834, Loss: {'train/loss': 0.75139981508255, 'train/grad_norm_pre_clip': 10.588835716247559, 'timing/data': 0.0004049270646646619, 'timing/model': 1.2050942240748554, 'learning_rate/qwen_vl_interface': 6.771204493405239e-06, 'learning_rate/action_model': 6.771204493405239e-05, 'epoch': 0.83}) +[2026-06-17 11:25:55,391][starVLA.training.train_starvla][INFO] - Step 2835, Loss: {'train/loss': 0.8835980296134949, 'train/grad_norm_pre_clip': 13.994604110717773, 'timing/data': 0.00035047694109380245, 'timing/model': 1.2030339949997142, 'learning_rate/qwen_vl_interface': 6.769155550106413e-06, 'learning_rate/action_model': 6.769155550106413e-05, 'epoch': 0.83}) +[2026-06-17 11:26:00,171][starVLA.training.train_starvla][INFO] - Step 2836, Loss: {'train/loss': 0.6763046979904175, 'train/grad_norm_pre_clip': 10.662342071533203, 'timing/data': 0.00035688700154423714, 'timing/model': 1.330664103035815, 'learning_rate/qwen_vl_interface': 6.767106291885006e-06, 'learning_rate/action_model': 6.767106291885006e-05, 'epoch': 0.83}) +[2026-06-17 11:26:04,818][starVLA.training.train_starvla][INFO] - Step 2837, Loss: {'train/loss': 0.9104536175727844, 'train/grad_norm_pre_clip': 11.182175636291504, 'timing/data': 0.0003902530297636986, 'timing/model': 1.2033757749013603, 'learning_rate/qwen_vl_interface': 6.765056719165833e-06, 'learning_rate/action_model': 6.765056719165834e-05, 'epoch': 0.83}) +[2026-06-17 11:26:09,598][starVLA.training.train_starvla][INFO] - Step 2838, Loss: {'train/loss': 1.0137819051742554, 'train/grad_norm_pre_clip': 12.560776710510254, 'timing/data': 0.000352125964127481, 'timing/model': 1.2037863950245082, 'learning_rate/qwen_vl_interface': 6.763006832373772e-06, 'learning_rate/action_model': 6.763006832373771e-05, 'epoch': 0.84}) +[2026-06-17 11:26:14,397][starVLA.training.train_starvla][INFO] - Step 2839, Loss: {'train/loss': 0.5684200525283813, 'train/grad_norm_pre_clip': 17.269872665405273, 'timing/data': 0.0003441219450905919, 'timing/model': 1.2041581500088796, 'learning_rate/qwen_vl_interface': 6.760956631933766e-06, 'learning_rate/action_model': 6.760956631933765e-05, 'epoch': 0.84}) +[2026-06-17 11:26:19,175][starVLA.training.train_starvla][INFO] - Step 2840, Loss: {'train/loss': 0.9464128017425537, 'train/grad_norm_pre_clip': 11.952771186828613, 'timing/data': 0.00037369399797171354, 'timing/model': 1.20391501695849, 'learning_rate/qwen_vl_interface': 6.758906118270826e-06, 'learning_rate/action_model': 6.758906118270825e-05, 'epoch': 0.84}) +[2026-06-17 11:26:23,953][starVLA.training.train_starvla][INFO] - Step 2841, Loss: {'train/loss': 0.4800686538219452, 'train/grad_norm_pre_clip': 6.78068208694458, 'timing/data': 0.0003542209742590785, 'timing/model': 1.2027287799865007, 'learning_rate/qwen_vl_interface': 6.756855291810024e-06, 'learning_rate/action_model': 6.756855291810024e-05, 'epoch': 0.84}) +[2026-06-17 11:26:28,736][starVLA.training.train_starvla][INFO] - Step 2842, Loss: {'train/loss': 0.8043044805526733, 'train/grad_norm_pre_clip': 13.116602897644043, 'timing/data': 0.0003567109815776348, 'timing/model': 1.3340180979575962, 'learning_rate/qwen_vl_interface': 6.754804152976495e-06, 'learning_rate/action_model': 6.754804152976495e-05, 'epoch': 0.84}) +[2026-06-17 11:26:33,457][starVLA.training.train_starvla][INFO] - Step 2843, Loss: {'train/loss': 0.675011157989502, 'train/grad_norm_pre_clip': 12.321357727050781, 'timing/data': 0.0003753150813281536, 'timing/model': 1.2041786620393395, 'learning_rate/qwen_vl_interface': 6.752752702195448e-06, 'learning_rate/action_model': 6.752752702195448e-05, 'epoch': 0.84}) +[2026-06-17 11:26:38,236][starVLA.training.train_starvla][INFO] - Step 2844, Loss: {'train/loss': 0.8664771914482117, 'train/grad_norm_pre_clip': 10.506340980529785, 'timing/data': 0.000366356922313571, 'timing/model': 1.2051007420523092, 'learning_rate/qwen_vl_interface': 6.750700939892151e-06, 'learning_rate/action_model': 6.75070093989215e-05, 'epoch': 0.84}) +[2026-06-17 11:26:43,006][starVLA.training.train_starvla][INFO] - Step 2845, Loss: {'train/loss': 0.8451034426689148, 'train/grad_norm_pre_clip': 10.574623107910156, 'timing/data': 0.0003717989893630147, 'timing/model': 1.20152895397041, 'learning_rate/qwen_vl_interface': 6.748648866491931e-06, 'learning_rate/action_model': 6.748648866491932e-05, 'epoch': 0.84}) +[2026-06-17 11:26:47,890][starVLA.training.train_starvla][INFO] - Step 2846, Loss: {'train/loss': 1.1040631532669067, 'train/grad_norm_pre_clip': 11.85175895690918, 'timing/data': 0.0003853129455819726, 'timing/model': 1.2035694268997759, 'learning_rate/qwen_vl_interface': 6.7465964824201915e-06, 'learning_rate/action_model': 6.746596482420192e-05, 'epoch': 0.84}) +[2026-06-17 11:26:52,689][starVLA.training.train_starvla][INFO] - Step 2847, Loss: {'train/loss': 0.9058387279510498, 'train/grad_norm_pre_clip': 13.790731430053711, 'timing/data': 0.0003522509941831231, 'timing/model': 1.3294680850813165, 'learning_rate/qwen_vl_interface': 6.744543788102388e-06, 'learning_rate/action_model': 6.744543788102388e-05, 'epoch': 0.84}) +[2026-06-17 11:26:57,429][starVLA.training.train_starvla][INFO] - Step 2848, Loss: {'train/loss': 0.6354590058326721, 'train/grad_norm_pre_clip': 9.918724060058594, 'timing/data': 0.0003697850042954087, 'timing/model': 1.2839566480834037, 'learning_rate/qwen_vl_interface': 6.742490783964051e-06, 'learning_rate/action_model': 6.742490783964051e-05, 'epoch': 0.84}) +[2026-06-17 11:27:02,110][starVLA.training.train_starvla][INFO] - Step 2849, Loss: {'train/loss': 0.6581328511238098, 'train/grad_norm_pre_clip': 8.709858894348145, 'timing/data': 0.00038654706440865993, 'timing/model': 1.2084909429540858, 'learning_rate/qwen_vl_interface': 6.740437470430766e-06, 'learning_rate/action_model': 6.740437470430766e-05, 'epoch': 0.84}) +[2026-06-17 11:27:06,956][starVLA.training.train_starvla][INFO] - Step 2850, Loss: {'train/loss': 0.6959435343742371, 'train/grad_norm_pre_clip': 9.732394218444824, 'timing/data': 0.000354600022546947, 'timing/model': 1.2027858850779012, 'learning_rate/qwen_vl_interface': 6.7383838479281915e-06, 'learning_rate/action_model': 6.738383847928191e-05, 'epoch': 0.84}) +[2026-06-17 11:27:11,734][starVLA.training.train_starvla][INFO] - Step 2851, Loss: {'train/loss': 1.16946280002594, 'train/grad_norm_pre_clip': 11.899645805358887, 'timing/data': 0.00043695210479199886, 'timing/model': 1.2023396349977702, 'learning_rate/qwen_vl_interface': 6.736329916882044e-06, 'learning_rate/action_model': 6.736329916882044e-05, 'epoch': 0.84}) +[2026-06-17 11:27:16,507][starVLA.training.train_starvla][INFO] - Step 2852, Loss: {'train/loss': 0.5676426887512207, 'train/grad_norm_pre_clip': 10.389758110046387, 'timing/data': 0.0003594739828258753, 'timing/model': 1.2023576729698107, 'learning_rate/qwen_vl_interface': 6.734275677718105e-06, 'learning_rate/action_model': 6.734275677718105e-05, 'epoch': 0.84}) +[2026-06-17 11:27:21,280][starVLA.training.train_starvla][INFO] - Step 2853, Loss: {'train/loss': 0.7041580080986023, 'train/grad_norm_pre_clip': 10.490559577941895, 'timing/data': 0.00036977999843657017, 'timing/model': 1.3304858680348843, 'learning_rate/qwen_vl_interface': 6.73222113086222e-06, 'learning_rate/action_model': 6.73222113086222e-05, 'epoch': 0.84}) +[2026-06-17 11:27:25,966][starVLA.training.train_starvla][INFO] - Step 2854, Loss: {'train/loss': 0.7802930474281311, 'train/grad_norm_pre_clip': 12.036336898803711, 'timing/data': 0.0003636549226939678, 'timing/model': 1.2230180570622906, 'learning_rate/qwen_vl_interface': 6.7301662767403e-06, 'learning_rate/action_model': 6.7301662767403e-05, 'epoch': 0.84}) +[2026-06-17 11:27:30,613][starVLA.training.train_starvla][INFO] - Step 2855, Loss: {'train/loss': 0.7174844741821289, 'train/grad_norm_pre_clip': 9.803014755249023, 'timing/data': 0.00037226208951324224, 'timing/model': 1.2023784150369465, 'learning_rate/qwen_vl_interface': 6.728111115778319e-06, 'learning_rate/action_model': 6.728111115778318e-05, 'epoch': 0.84}) +[2026-06-17 11:27:35,389][starVLA.training.train_starvla][INFO] - Step 2856, Loss: {'train/loss': 0.6686989068984985, 'train/grad_norm_pre_clip': 10.436232566833496, 'timing/data': 0.00036462198477238417, 'timing/model': 1.2012094949604943, 'learning_rate/qwen_vl_interface': 6.7260556484023135e-06, 'learning_rate/action_model': 6.726055648402313e-05, 'epoch': 0.84}) +[2026-06-17 11:27:40,224][starVLA.training.train_starvla][INFO] - Step 2857, Loss: {'train/loss': 0.6423138380050659, 'train/grad_norm_pre_clip': 10.985620498657227, 'timing/data': 0.00033288903068751097, 'timing/model': 1.202087876969017, 'learning_rate/qwen_vl_interface': 6.723999875038381e-06, 'learning_rate/action_model': 6.72399987503838e-05, 'epoch': 0.84}) +[2026-06-17 11:27:45,008][starVLA.training.train_starvla][INFO] - Step 2858, Loss: {'train/loss': 0.8691249489784241, 'train/grad_norm_pre_clip': 11.653953552246094, 'timing/data': 0.00038163899444043636, 'timing/model': 1.200932636973448, 'learning_rate/qwen_vl_interface': 6.721943796112691e-06, 'learning_rate/action_model': 6.721943796112691e-05, 'epoch': 0.84}) +[2026-06-17 11:27:49,784][starVLA.training.train_starvla][INFO] - Step 2859, Loss: {'train/loss': 0.8183979392051697, 'train/grad_norm_pre_clip': 16.717041015625, 'timing/data': 0.0003657939378172159, 'timing/model': 1.3327363890130073, 'learning_rate/qwen_vl_interface': 6.719887412051468e-06, 'learning_rate/action_model': 6.719887412051468e-05, 'epoch': 0.84}) +[2026-06-17 11:27:54,499][starVLA.training.train_starvla][INFO] - Step 2860, Loss: {'train/loss': 0.6538894772529602, 'train/grad_norm_pre_clip': 9.529656410217285, 'timing/data': 0.0005195240955799818, 'timing/model': 1.2023822519695386, 'learning_rate/qwen_vl_interface': 6.7178307232810015e-06, 'learning_rate/action_model': 6.717830723281e-05, 'epoch': 0.84}) +[2026-06-17 11:27:59,353][starVLA.training.train_starvla][INFO] - Step 2861, Loss: {'train/loss': 0.8961535692214966, 'train/grad_norm_pre_clip': 11.521523475646973, 'timing/data': 0.0005958270048722625, 'timing/model': 1.2023886210517958, 'learning_rate/qwen_vl_interface': 6.715773730227646e-06, 'learning_rate/action_model': 6.715773730227646e-05, 'epoch': 0.84}) +[2026-06-17 11:28:04,137][starVLA.training.train_starvla][INFO] - Step 2862, Loss: {'train/loss': 0.8331705331802368, 'train/grad_norm_pre_clip': 9.374744415283203, 'timing/data': 0.0003599849296733737, 'timing/model': 1.205131967086345, 'learning_rate/qwen_vl_interface': 6.7137164333178216e-06, 'learning_rate/action_model': 6.713716433317821e-05, 'epoch': 0.84}) +[2026-06-17 11:28:08,866][starVLA.training.train_starvla][INFO] - Step 2863, Loss: {'train/loss': 0.9176009893417358, 'train/grad_norm_pre_clip': 12.217948913574219, 'timing/data': 0.0003723220434039831, 'timing/model': 1.2027992790099233, 'learning_rate/qwen_vl_interface': 6.711658832978004e-06, 'learning_rate/action_model': 6.711658832978003e-05, 'epoch': 0.84}) +[2026-06-17 11:28:13,644][starVLA.training.train_starvla][INFO] - Step 2864, Loss: {'train/loss': 0.6567981243133545, 'train/grad_norm_pre_clip': 12.074287414550781, 'timing/data': 0.00035699200816452503, 'timing/model': 1.2044859839370474, 'learning_rate/qwen_vl_interface': 6.709600929634737e-06, 'learning_rate/action_model': 6.709600929634738e-05, 'epoch': 0.84}) +[2026-06-17 11:28:18,439][starVLA.training.train_starvla][INFO] - Step 2865, Loss: {'train/loss': 0.6758076548576355, 'train/grad_norm_pre_clip': 11.43775749206543, 'timing/data': 0.00036376307252794504, 'timing/model': 1.338690541102551, 'learning_rate/qwen_vl_interface': 6.707542723714629e-06, 'learning_rate/action_model': 6.707542723714629e-05, 'epoch': 0.84}) +[2026-06-17 11:28:23,107][starVLA.training.train_starvla][INFO] - Step 2866, Loss: {'train/loss': 0.7519116401672363, 'train/grad_norm_pre_clip': 11.044893264770508, 'timing/data': 0.0003621779615059495, 'timing/model': 1.2064294470474124, 'learning_rate/qwen_vl_interface': 6.7054842156443435e-06, 'learning_rate/action_model': 6.705484215644343e-05, 'epoch': 0.84}) +[2026-06-17 11:28:27,890][starVLA.training.train_starvla][INFO] - Step 2867, Loss: {'train/loss': 0.7464897632598877, 'train/grad_norm_pre_clip': 12.432597160339355, 'timing/data': 0.00042202207259833813, 'timing/model': 1.2040556661086157, 'learning_rate/qwen_vl_interface': 6.703425405850614e-06, 'learning_rate/action_model': 6.703425405850613e-05, 'epoch': 0.84}) +[2026-06-17 11:28:32,668][starVLA.training.train_starvla][INFO] - Step 2868, Loss: {'train/loss': 0.789344847202301, 'train/grad_norm_pre_clip': 9.73874282836914, 'timing/data': 0.000377909978851676, 'timing/model': 1.2036452899919823, 'learning_rate/qwen_vl_interface': 6.701366294760236e-06, 'learning_rate/action_model': 6.701366294760235e-05, 'epoch': 0.84}) +[2026-06-17 11:28:37,440][starVLA.training.train_starvla][INFO] - Step 2869, Loss: {'train/loss': 0.7711306810379028, 'train/grad_norm_pre_clip': 12.733034133911133, 'timing/data': 0.0004152499604970217, 'timing/model': 1.2039795860182494, 'learning_rate/qwen_vl_interface': 6.6993068828000605e-06, 'learning_rate/action_model': 6.69930688280006e-05, 'epoch': 0.84}) +[2026-06-17 11:28:42,236][starVLA.training.train_starvla][INFO] - Step 2870, Loss: {'train/loss': 0.4261777102947235, 'train/grad_norm_pre_clip': 7.764477729797363, 'timing/data': 0.00037714000791311264, 'timing/model': 1.2055392829934135, 'learning_rate/qwen_vl_interface': 6.697247170397011e-06, 'learning_rate/action_model': 6.697247170397011e-05, 'epoch': 0.84}) +[2026-06-17 11:28:47,021][starVLA.training.train_starvla][INFO] - Step 2871, Loss: {'train/loss': 0.9567888379096985, 'train/grad_norm_pre_clip': 13.372115135192871, 'timing/data': 0.00037540600169450045, 'timing/model': 1.3354136239504442, 'learning_rate/qwen_vl_interface': 6.6951871579780645e-06, 'learning_rate/action_model': 6.695187157978064e-05, 'epoch': 0.84}) +[2026-06-17 11:28:51,685][starVLA.training.train_starvla][INFO] - Step 2872, Loss: {'train/loss': 0.47885778546333313, 'train/grad_norm_pre_clip': 7.210153579711914, 'timing/data': 0.0003647439880296588, 'timing/model': 1.2059845919720829, 'learning_rate/qwen_vl_interface': 6.693126845970267e-06, 'learning_rate/action_model': 6.693126845970267e-05, 'epoch': 0.85}) +[2026-06-17 11:28:56,514][starVLA.training.train_starvla][INFO] - Step 2873, Loss: {'train/loss': 0.8612426519393921, 'train/grad_norm_pre_clip': 12.100967407226562, 'timing/data': 0.0005550209898501635, 'timing/model': 1.2479357830015942, 'learning_rate/qwen_vl_interface': 6.691066234800719e-06, 'learning_rate/action_model': 6.691066234800719e-05, 'epoch': 0.85}) +[2026-06-17 11:29:01,307][starVLA.training.train_starvla][INFO] - Step 2874, Loss: {'train/loss': 0.6841248273849487, 'train/grad_norm_pre_clip': 11.101012229919434, 'timing/data': 0.00037643301766365767, 'timing/model': 1.2083395048975945, 'learning_rate/qwen_vl_interface': 6.689005324896591e-06, 'learning_rate/action_model': 6.689005324896591e-05, 'epoch': 0.85}) +[2026-06-17 11:29:06,097][starVLA.training.train_starvla][INFO] - Step 2875, Loss: {'train/loss': 0.8001974821090698, 'train/grad_norm_pre_clip': 13.421098709106445, 'timing/data': 0.0003361939452588558, 'timing/model': 1.2086311839520931, 'learning_rate/qwen_vl_interface': 6.686944116685111e-06, 'learning_rate/action_model': 6.686944116685111e-05, 'epoch': 0.85}) +[2026-06-17 11:29:10,878][starVLA.training.train_starvla][INFO] - Step 2876, Loss: {'train/loss': 0.8509007692337036, 'train/grad_norm_pre_clip': 9.896810531616211, 'timing/data': 0.00034496502485126257, 'timing/model': 1.329285038053058, 'learning_rate/qwen_vl_interface': 6.684882610593568e-06, 'learning_rate/action_model': 6.684882610593567e-05, 'epoch': 0.85}) +[2026-06-17 11:29:15,616][starVLA.training.train_starvla][INFO] - Step 2877, Loss: {'train/loss': 0.6549726724624634, 'train/grad_norm_pre_clip': 9.292458534240723, 'timing/data': 0.00033301999792456627, 'timing/model': 1.2840040479786694, 'learning_rate/qwen_vl_interface': 6.682820807049317e-06, 'learning_rate/action_model': 6.682820807049317e-05, 'epoch': 0.85}) +[2026-06-17 11:29:20,276][starVLA.training.train_starvla][INFO] - Step 2878, Loss: {'train/loss': 0.7754740715026855, 'train/grad_norm_pre_clip': 11.732551574707031, 'timing/data': 0.00043107406236231327, 'timing/model': 1.2043714079773054, 'learning_rate/qwen_vl_interface': 6.680758706479772e-06, 'learning_rate/action_model': 6.680758706479772e-05, 'epoch': 0.85}) +[2026-06-17 11:29:25,059][starVLA.training.train_starvla][INFO] - Step 2879, Loss: {'train/loss': 0.6114078164100647, 'train/grad_norm_pre_clip': 9.349181175231934, 'timing/data': 0.00039488309994339943, 'timing/model': 1.2050109460251406, 'learning_rate/qwen_vl_interface': 6.678696309312405e-06, 'learning_rate/action_model': 6.678696309312406e-05, 'epoch': 0.85}) +[2026-06-17 11:29:29,846][starVLA.training.train_starvla][INFO] - Step 2880, Loss: {'train/loss': 0.8039217591285706, 'train/grad_norm_pre_clip': 8.1063814163208, 'timing/data': 0.00034436595160514116, 'timing/model': 1.204444168950431, 'learning_rate/qwen_vl_interface': 6.676633615974758e-06, 'learning_rate/action_model': 6.676633615974757e-05, 'epoch': 0.85}) +[2026-06-17 11:29:34,637][starVLA.training.train_starvla][INFO] - Step 2881, Loss: {'train/loss': 0.6186593174934387, 'train/grad_norm_pre_clip': 14.244893074035645, 'timing/data': 0.0003802829887717962, 'timing/model': 1.2063007430406287, 'learning_rate/qwen_vl_interface': 6.6745706268944265e-06, 'learning_rate/action_model': 6.674570626894426e-05, 'epoch': 0.85}) +[2026-06-17 11:29:39,429][starVLA.training.train_starvla][INFO] - Step 2882, Loss: {'train/loss': 0.6906930208206177, 'train/grad_norm_pre_clip': 9.63748836517334, 'timing/data': 0.0003946570213884115, 'timing/model': 1.3331439339090139, 'learning_rate/qwen_vl_interface': 6.672507342499071e-06, 'learning_rate/action_model': 6.67250734249907e-05, 'epoch': 0.85}) +[2026-06-17 11:29:44,103][starVLA.training.train_starvla][INFO] - Step 2883, Loss: {'train/loss': 0.901648759841919, 'train/grad_norm_pre_clip': 14.784501075744629, 'timing/data': 0.00044096901547163725, 'timing/model': 1.2135753389447927, 'learning_rate/qwen_vl_interface': 6.670443763216412e-06, 'learning_rate/action_model': 6.670443763216411e-05, 'epoch': 0.85}) +[2026-06-17 11:29:48,771][starVLA.training.train_starvla][INFO] - Step 2884, Loss: {'train/loss': 0.779035747051239, 'train/grad_norm_pre_clip': 9.614910125732422, 'timing/data': 0.0003797949757426977, 'timing/model': 1.2149742050096393, 'learning_rate/qwen_vl_interface': 6.668379889474233e-06, 'learning_rate/action_model': 6.668379889474234e-05, 'epoch': 0.85}) +[2026-06-17 11:29:53,565][starVLA.training.train_starvla][INFO] - Step 2885, Loss: {'train/loss': 0.8307610750198364, 'train/grad_norm_pre_clip': 11.310284614562988, 'timing/data': 0.0003557389136403799, 'timing/model': 1.205018967972137, 'learning_rate/qwen_vl_interface': 6.666315721700378e-06, 'learning_rate/action_model': 6.666315721700377e-05, 'epoch': 0.85}) +[2026-06-17 11:29:58,358][starVLA.training.train_starvla][INFO] - Step 2886, Loss: {'train/loss': 0.7308468222618103, 'train/grad_norm_pre_clip': 10.67850399017334, 'timing/data': 0.0005456770304590464, 'timing/model': 1.205989649053663, 'learning_rate/qwen_vl_interface': 6.664251260322748e-06, 'learning_rate/action_model': 6.664251260322749e-05, 'epoch': 0.85}) +[2026-06-17 11:30:03,158][starVLA.training.train_starvla][INFO] - Step 2887, Loss: {'train/loss': 0.5926793217658997, 'train/grad_norm_pre_clip': 10.827049255371094, 'timing/data': 0.00044904393143951893, 'timing/model': 1.2112682959996164, 'learning_rate/qwen_vl_interface': 6.66218650576931e-06, 'learning_rate/action_model': 6.66218650576931e-05, 'epoch': 0.85}) +[2026-06-17 11:30:07,952][starVLA.training.train_starvla][INFO] - Step 2888, Loss: {'train/loss': 0.7258402109146118, 'train/grad_norm_pre_clip': 14.959200859069824, 'timing/data': 0.00035971507895737886, 'timing/model': 1.3374168199952692, 'learning_rate/qwen_vl_interface': 6.66012145846809e-06, 'learning_rate/action_model': 6.660121458468091e-05, 'epoch': 0.85}) +[2026-06-17 11:30:12,620][starVLA.training.train_starvla][INFO] - Step 2889, Loss: {'train/loss': 0.9820374846458435, 'train/grad_norm_pre_clip': 13.012800216674805, 'timing/data': 0.000369130983017385, 'timing/model': 1.206052466062829, 'learning_rate/qwen_vl_interface': 6.658056118847176e-06, 'learning_rate/action_model': 6.658056118847175e-05, 'epoch': 0.85}) +[2026-06-17 11:30:17,441][starVLA.training.train_starvla][INFO] - Step 2890, Loss: {'train/loss': 0.9324192404747009, 'train/grad_norm_pre_clip': 11.368894577026367, 'timing/data': 0.0004294429672881961, 'timing/model': 1.2033375869505107, 'learning_rate/qwen_vl_interface': 6.655990487334713e-06, 'learning_rate/action_model': 6.655990487334712e-05, 'epoch': 0.85}) +[2026-06-17 11:30:22,215][starVLA.training.train_starvla][INFO] - Step 2891, Loss: {'train/loss': 1.0439282655715942, 'train/grad_norm_pre_clip': 13.13895034790039, 'timing/data': 0.0004154130583629012, 'timing/model': 1.203168676001951, 'learning_rate/qwen_vl_interface': 6.653924564358907e-06, 'learning_rate/action_model': 6.653924564358907e-05, 'epoch': 0.85}) +[2026-06-17 11:30:26,941][starVLA.training.train_starvla][INFO] - Step 2892, Loss: {'train/loss': 0.824994683265686, 'train/grad_norm_pre_clip': 14.540170669555664, 'timing/data': 0.00035638001281768084, 'timing/model': 1.205474764923565, 'learning_rate/qwen_vl_interface': 6.6518583503480294e-06, 'learning_rate/action_model': 6.651858350348029e-05, 'epoch': 0.85}) +[2026-06-17 11:30:31,750][starVLA.training.train_starvla][INFO] - Step 2893, Loss: {'train/loss': 0.5327032804489136, 'train/grad_norm_pre_clip': 9.410639762878418, 'timing/data': 0.00034628191497176886, 'timing/model': 1.2068923310143873, 'learning_rate/qwen_vl_interface': 6.649791845730408e-06, 'learning_rate/action_model': 6.649791845730408e-05, 'epoch': 0.85}) +[2026-06-17 11:30:36,538][starVLA.training.train_starvla][INFO] - Step 2894, Loss: {'train/loss': 0.5294962525367737, 'train/grad_norm_pre_clip': 10.945262908935547, 'timing/data': 0.00036758207716047764, 'timing/model': 1.3338988629402593, 'learning_rate/qwen_vl_interface': 6.64772505093443e-06, 'learning_rate/action_model': 6.647725050934429e-05, 'epoch': 0.85}) +[2026-06-17 11:30:41,193][starVLA.training.train_starvla][INFO] - Step 2895, Loss: {'train/loss': 1.010979175567627, 'train/grad_norm_pre_clip': 12.378458023071289, 'timing/data': 0.0003912620013579726, 'timing/model': 1.2057955950731412, 'learning_rate/qwen_vl_interface': 6.645657966388543e-06, 'learning_rate/action_model': 6.645657966388543e-05, 'epoch': 0.85}) +[2026-06-17 11:30:45,970][starVLA.training.train_starvla][INFO] - Step 2896, Loss: {'train/loss': 0.7052395343780518, 'train/grad_norm_pre_clip': 10.519969940185547, 'timing/data': 0.0003938910085707903, 'timing/model': 1.203705226071179, 'learning_rate/qwen_vl_interface': 6.643590592521261e-06, 'learning_rate/action_model': 6.643590592521261e-05, 'epoch': 0.85}) +[2026-06-17 11:30:50,753][starVLA.training.train_starvla][INFO] - Step 2897, Loss: {'train/loss': 0.7456824779510498, 'train/grad_norm_pre_clip': 12.249927520751953, 'timing/data': 0.0003370320191606879, 'timing/model': 1.2050070409895852, 'learning_rate/qwen_vl_interface': 6.641522929761148e-06, 'learning_rate/action_model': 6.641522929761148e-05, 'epoch': 0.85}) +[2026-06-17 11:30:55,579][starVLA.training.train_starvla][INFO] - Step 2898, Loss: {'train/loss': 0.47411659359931946, 'train/grad_norm_pre_clip': 9.660640716552734, 'timing/data': 0.00037001504097133875, 'timing/model': 1.20526694401633, 'learning_rate/qwen_vl_interface': 6.639454978536835e-06, 'learning_rate/action_model': 6.639454978536834e-05, 'epoch': 0.85}) +[2026-06-17 11:31:00,357][starVLA.training.train_starvla][INFO] - Step 2899, Loss: {'train/loss': 0.7630992531776428, 'train/grad_norm_pre_clip': 10.902976989746094, 'timing/data': 0.00035355298314243555, 'timing/model': 1.2068585969973356, 'learning_rate/qwen_vl_interface': 6.63738673927701e-06, 'learning_rate/action_model': 6.63738673927701e-05, 'epoch': 0.85}) +[2026-06-17 11:32:41,760][starVLA.training.train_starvla][INFO] - Step 2900, Loss: {'train/loss': 0.5181257128715515, 'train/grad_norm_pre_clip': 11.064690589904785, 'eval/loss': 0.7849210500717163, 'eval/latency_0/loss': 0.7849210500717163, 'eval/demon_fire_precision': 0.8915566226490597, 'eval/demon_fire_recall': 0.925249169435216, 'eval/demon_fire_f1': 0.908090482983493, 'eval/demon_fire_rate_model': 0.6949388209121246, 'eval/demon_fire_rate_teacher': 0.6696329254727474, 'eval/demon_left_precision': 0.7815482502651113, 'eval/demon_left_recall': 0.8351274787535411, 'eval/demon_left_f1': 0.8074500136948781, 'eval/demon_left_rate_model': 0.5244716351501668, 'eval/demon_left_rate_teacher': 0.49082313681868744, 'eval/demon_right_precision': 0.8220140515222483, 'eval/demon_right_recall': 0.7932203389830509, 'eval/demon_right_f1': 0.8073605520414031, 'eval/demon_right_rate_model': 0.474972191323693, 'eval/demon_right_rate_teacher': 0.4922135706340378, 'eval/demon_move_f1': 0.8074052828681406, 'eval/demon_cc_f1': 0.8753433166436265, 'eval/latency_0/demon_fire_precision': 0.8915566226490597, 'eval/latency_0/demon_fire_recall': 0.925249169435216, 'eval/latency_0/demon_fire_f1': 0.908090482983493, 'eval/latency_0/demon_fire_rate_model': 0.6949388209121246, 'eval/latency_0/demon_fire_rate_teacher': 0.6696329254727474, 'eval/latency_0/demon_left_precision': 0.7815482502651113, 'eval/latency_0/demon_left_recall': 0.8351274787535411, 'eval/latency_0/demon_left_f1': 0.8074500136948781, 'eval/latency_0/demon_left_rate_model': 0.5244716351501668, 'eval/latency_0/demon_left_rate_teacher': 0.49082313681868744, 'eval/latency_0/demon_right_precision': 0.8220140515222483, 'eval/latency_0/demon_right_recall': 0.7932203389830509, 'eval/latency_0/demon_right_f1': 0.8073605520414031, 'eval/latency_0/demon_right_rate_model': 0.474972191323693, 'eval/latency_0/demon_right_rate_teacher': 0.4922135706340378, 'eval/latency_0/demon_move_f1': 0.8074052828681406, 'eval/latency_0/demon_cc_f1': 0.8753433166436265, 'timing/data': 0.00042899500112980604, 'timing/model': 1.2665404650615528, 'learning_rate/qwen_vl_interface': 6.63531821241042e-06, 'learning_rate/action_model': 6.63531821241042e-05, 'epoch': 0.85}) +[2026-06-17 11:32:41,760][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_2900_state +[2026-06-17 11:32:41,764][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 11:32:45,344][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_2900_state/model.safetensors +[2026-06-17 11:32:52,966][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_2900_state/optimizer.bin +[2026-06-17 11:32:52,966][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_2900_state/sampler.bin +[2026-06-17 11:32:52,966][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_2900_state/sampler_1.bin +[2026-06-17 11:32:52,967][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_2900_state/random_states_0.pkl +[2026-06-17 11:32:53,013][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2900 +[2026-06-17 11:32:53,013][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 11:32:53,056][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 11:32:57,816][starVLA.training.train_starvla][INFO] - Step 2901, Loss: {'train/loss': 0.6463243961334229, 'train/grad_norm_pre_clip': 9.495705604553223, 'timing/data': 0.0003785770386457443, 'timing/model': 1.3193420759635046, 'learning_rate/qwen_vl_interface': 6.633249398365874e-06, 'learning_rate/action_model': 6.633249398365873e-05, 'epoch': 0.85}) +[2026-06-17 11:33:02,416][starVLA.training.train_starvla][INFO] - Step 2902, Loss: {'train/loss': 0.8953779935836792, 'train/grad_norm_pre_clip': 10.176480293273926, 'timing/data': 0.00035241199657320976, 'timing/model': 1.1900558409979567, 'learning_rate/qwen_vl_interface': 6.6311802975722395e-06, 'learning_rate/action_model': 6.63118029757224e-05, 'epoch': 0.85}) +[2026-06-17 11:33:07,148][starVLA.training.train_starvla][INFO] - Step 2903, Loss: {'train/loss': 0.5823959112167358, 'train/grad_norm_pre_clip': 9.650174140930176, 'timing/data': 0.0003923861077055335, 'timing/model': 1.1927215430187061, 'learning_rate/qwen_vl_interface': 6.629110910458444e-06, 'learning_rate/action_model': 6.629110910458444e-05, 'epoch': 0.85}) +[2026-06-17 11:33:11,880][starVLA.training.train_starvla][INFO] - Step 2904, Loss: {'train/loss': 0.8841736912727356, 'train/grad_norm_pre_clip': 12.156375885009766, 'timing/data': 0.00036594702396541834, 'timing/model': 1.1932122089201584, 'learning_rate/qwen_vl_interface': 6.627041237453471e-06, 'learning_rate/action_model': 6.627041237453471e-05, 'epoch': 0.85}) +[2026-06-17 11:33:16,629][starVLA.training.train_starvla][INFO] - Step 2905, Loss: {'train/loss': 0.7585353851318359, 'train/grad_norm_pre_clip': 8.9843111038208, 'timing/data': 0.00037312100175768137, 'timing/model': 1.1952893709531054, 'learning_rate/qwen_vl_interface': 6.624971278986369e-06, 'learning_rate/action_model': 6.62497127898637e-05, 'epoch': 0.85}) +[2026-06-17 11:33:21,376][starVLA.training.train_starvla][INFO] - Step 2906, Loss: {'train/loss': 0.627903938293457, 'train/grad_norm_pre_clip': 11.500459671020508, 'timing/data': 0.00036370393354445696, 'timing/model': 1.196261832024902, 'learning_rate/qwen_vl_interface': 6.622901035486241e-06, 'learning_rate/action_model': 6.622901035486241e-05, 'epoch': 0.86}) +[2026-06-17 11:33:26,222][starVLA.training.train_starvla][INFO] - Step 2907, Loss: {'train/loss': 0.799214243888855, 'train/grad_norm_pre_clip': 11.603318214416504, 'timing/data': 0.0003760620020329952, 'timing/model': 1.32721230504103, 'learning_rate/qwen_vl_interface': 6.620830507382251e-06, 'learning_rate/action_model': 6.620830507382251e-05, 'epoch': 0.86}) +[2026-06-17 11:33:30,904][starVLA.training.train_starvla][INFO] - Step 2908, Loss: {'train/loss': 0.7019507884979248, 'train/grad_norm_pre_clip': 8.519465446472168, 'timing/data': 0.0003861959557980299, 'timing/model': 1.198843116988428, 'learning_rate/qwen_vl_interface': 6.618759695103623e-06, 'learning_rate/action_model': 6.618759695103622e-05, 'epoch': 0.86}) +[2026-06-17 11:33:35,669][starVLA.training.train_starvla][INFO] - Step 2909, Loss: {'train/loss': 0.703161358833313, 'train/grad_norm_pre_clip': 9.830613136291504, 'timing/data': 0.0006911590462550521, 'timing/model': 1.1982511340174824, 'learning_rate/qwen_vl_interface': 6.616688599079635e-06, 'learning_rate/action_model': 6.616688599079634e-05, 'epoch': 0.86}) +[2026-06-17 11:33:40,333][starVLA.training.train_starvla][INFO] - Step 2910, Loss: {'train/loss': 0.6585705876350403, 'train/grad_norm_pre_clip': 8.881608009338379, 'timing/data': 0.0003780609695240855, 'timing/model': 1.2039676420390606, 'learning_rate/qwen_vl_interface': 6.614617219739632e-06, 'learning_rate/action_model': 6.614617219739632e-05, 'epoch': 0.86}) +[2026-06-17 11:33:44,994][starVLA.training.train_starvla][INFO] - Step 2911, Loss: {'train/loss': 0.49122902750968933, 'train/grad_norm_pre_clip': 9.841163635253906, 'timing/data': 0.0003529649693518877, 'timing/model': 1.200258429045789, 'learning_rate/qwen_vl_interface': 6.612545557513008e-06, 'learning_rate/action_model': 6.612545557513008e-05, 'epoch': 0.86}) +[2026-06-17 11:33:49,702][starVLA.training.train_starvla][INFO] - Step 2912, Loss: {'train/loss': 0.9293336868286133, 'train/grad_norm_pre_clip': 11.998724937438965, 'timing/data': 0.0003531260881572962, 'timing/model': 1.1999350240221247, 'learning_rate/qwen_vl_interface': 6.610473612829227e-06, 'learning_rate/action_model': 6.610473612829226e-05, 'epoch': 0.86}) +[2026-06-17 11:33:54,443][starVLA.training.train_starvla][INFO] - Step 2913, Loss: {'train/loss': 0.7311521172523499, 'train/grad_norm_pre_clip': 10.795944213867188, 'timing/data': 0.00034852290991693735, 'timing/model': 1.2879965349566191, 'learning_rate/qwen_vl_interface': 6.608401386117801e-06, 'learning_rate/action_model': 6.6084013861178e-05, 'epoch': 0.86}) +[2026-06-17 11:33:59,117][starVLA.training.train_starvla][INFO] - Step 2914, Loss: {'train/loss': 0.7424284815788269, 'train/grad_norm_pre_clip': 9.947610855102539, 'timing/data': 0.00037478108424693346, 'timing/model': 1.214741406030953, 'learning_rate/qwen_vl_interface': 6.606328877808306e-06, 'learning_rate/action_model': 6.606328877808305e-05, 'epoch': 0.86}) +[2026-06-17 11:34:03,891][starVLA.training.train_starvla][INFO] - Step 2915, Loss: {'train/loss': 0.8887910842895508, 'train/grad_norm_pre_clip': 11.832551002502441, 'timing/data': 0.000521537964232266, 'timing/model': 1.1994647199753672, 'learning_rate/qwen_vl_interface': 6.604256088330374e-06, 'learning_rate/action_model': 6.604256088330374e-05, 'epoch': 0.86}) +[2026-06-17 11:34:08,745][starVLA.training.train_starvla][INFO] - Step 2916, Loss: {'train/loss': 0.5458858013153076, 'train/grad_norm_pre_clip': 8.949122428894043, 'timing/data': 0.00038987898733466864, 'timing/model': 1.2080007119802758, 'learning_rate/qwen_vl_interface': 6.602183018113698e-06, 'learning_rate/action_model': 6.602183018113698e-05, 'epoch': 0.86}) +[2026-06-17 11:34:13,563][starVLA.training.train_starvla][INFO] - Step 2917, Loss: {'train/loss': 0.8250879049301147, 'train/grad_norm_pre_clip': 9.973724365234375, 'timing/data': 0.0003669620491564274, 'timing/model': 1.203654665965587, 'learning_rate/qwen_vl_interface': 6.600109667588029e-06, 'learning_rate/action_model': 6.60010966758803e-05, 'epoch': 0.86}) +[2026-06-17 11:34:18,342][starVLA.training.train_starvla][INFO] - Step 2918, Loss: {'train/loss': 0.7984550595283508, 'train/grad_norm_pre_clip': 12.004472732543945, 'timing/data': 0.00034308491740375757, 'timing/model': 1.327247839071788, 'learning_rate/qwen_vl_interface': 6.598036037183172e-06, 'learning_rate/action_model': 6.598036037183172e-05, 'epoch': 0.86}) +[2026-06-17 11:34:22,986][starVLA.training.train_starvla][INFO] - Step 2919, Loss: {'train/loss': 0.5263379812240601, 'train/grad_norm_pre_clip': 8.333683013916016, 'timing/data': 0.000346521963365376, 'timing/model': 1.201513702981174, 'learning_rate/qwen_vl_interface': 6.595962127328993e-06, 'learning_rate/action_model': 6.595962127328993e-05, 'epoch': 0.86}) +[2026-06-17 11:34:27,762][starVLA.training.train_starvla][INFO] - Step 2920, Loss: {'train/loss': 0.8446624875068665, 'train/grad_norm_pre_clip': 12.093001365661621, 'timing/data': 0.0003432669909670949, 'timing/model': 1.203505843062885, 'learning_rate/qwen_vl_interface': 6.593887938455418e-06, 'learning_rate/action_model': 6.593887938455418e-05, 'epoch': 0.86}) +[2026-06-17 11:34:32,524][starVLA.training.train_starvla][INFO] - Step 2921, Loss: {'train/loss': 0.6111459732055664, 'train/grad_norm_pre_clip': 9.187396049499512, 'timing/data': 0.00032430700957775116, 'timing/model': 1.204011669033207, 'learning_rate/qwen_vl_interface': 6.591813470992428e-06, 'learning_rate/action_model': 6.591813470992427e-05, 'epoch': 0.86}) +[2026-06-17 11:34:37,296][starVLA.training.train_starvla][INFO] - Step 2922, Loss: {'train/loss': 0.7415550351142883, 'train/grad_norm_pre_clip': 12.185903549194336, 'timing/data': 0.0003457790007814765, 'timing/model': 1.2012969549978152, 'learning_rate/qwen_vl_interface': 6.58973872537006e-06, 'learning_rate/action_model': 6.589738725370059e-05, 'epoch': 0.86}) +[2026-06-17 11:34:42,073][starVLA.training.train_starvla][INFO] - Step 2923, Loss: {'train/loss': 0.6018646955490112, 'train/grad_norm_pre_clip': 7.862196922302246, 'timing/data': 0.00038416089955717325, 'timing/model': 1.2026750888908282, 'learning_rate/qwen_vl_interface': 6.587663702018413e-06, 'learning_rate/action_model': 6.587663702018412e-05, 'epoch': 0.86}) +[2026-06-17 11:34:46,885][starVLA.training.train_starvla][INFO] - Step 2924, Loss: {'train/loss': 0.8676964640617371, 'train/grad_norm_pre_clip': 16.928800582885742, 'timing/data': 0.00038120802491903305, 'timing/model': 1.3650130049791187, 'learning_rate/qwen_vl_interface': 6.585588401367641e-06, 'learning_rate/action_model': 6.58558840136764e-05, 'epoch': 0.86}) +[2026-06-17 11:34:51,549][starVLA.training.train_starvla][INFO] - Step 2925, Loss: {'train/loss': 0.7035406827926636, 'train/grad_norm_pre_clip': 9.759204864501953, 'timing/data': 0.000620263977907598, 'timing/model': 1.20502477302216, 'learning_rate/qwen_vl_interface': 6.583512823847957e-06, 'learning_rate/action_model': 6.583512823847958e-05, 'epoch': 0.86}) +[2026-06-17 11:34:56,364][starVLA.training.train_starvla][INFO] - Step 2926, Loss: {'train/loss': 0.8176153302192688, 'train/grad_norm_pre_clip': 9.357512474060059, 'timing/data': 0.0003412860678508878, 'timing/model': 1.2038447939557955, 'learning_rate/qwen_vl_interface': 6.5814369698896285e-06, 'learning_rate/action_model': 6.581436969889628e-05, 'epoch': 0.86}) +[2026-06-17 11:35:01,063][starVLA.training.train_starvla][INFO] - Step 2927, Loss: {'train/loss': 0.5826071500778198, 'train/grad_norm_pre_clip': 8.433632850646973, 'timing/data': 0.000426111975684762, 'timing/model': 1.204114094027318, 'learning_rate/qwen_vl_interface': 6.5793608399229855e-06, 'learning_rate/action_model': 6.579360839922985e-05, 'epoch': 0.86}) +[2026-06-17 11:35:05,851][starVLA.training.train_starvla][INFO] - Step 2928, Loss: {'train/loss': 0.6357608437538147, 'train/grad_norm_pre_clip': 9.522028923034668, 'timing/data': 0.0003596720052883029, 'timing/model': 1.203314250917174, 'learning_rate/qwen_vl_interface': 6.577284434378408e-06, 'learning_rate/action_model': 6.577284434378408e-05, 'epoch': 0.86}) +[2026-06-17 11:35:10,607][starVLA.training.train_starvla][INFO] - Step 2929, Loss: {'train/loss': 0.9139506816864014, 'train/grad_norm_pre_clip': 12.78007984161377, 'timing/data': 0.0003509910311549902, 'timing/model': 1.2031933470861986, 'learning_rate/qwen_vl_interface': 6.5752077536863404e-06, 'learning_rate/action_model': 6.57520775368634e-05, 'epoch': 0.86}) +[2026-06-17 11:35:15,413][starVLA.training.train_starvla][INFO] - Step 2930, Loss: {'train/loss': 1.0208169221878052, 'train/grad_norm_pre_clip': 14.535628318786621, 'timing/data': 0.00035297300200909376, 'timing/model': 1.3363585559418425, 'learning_rate/qwen_vl_interface': 6.573130798277277e-06, 'learning_rate/action_model': 6.573130798277277e-05, 'epoch': 0.86}) +[2026-06-17 11:35:20,058][starVLA.training.train_starvla][INFO] - Step 2931, Loss: {'train/loss': 0.561935305595398, 'train/grad_norm_pre_clip': 9.561911582946777, 'timing/data': 0.0003574470756575465, 'timing/model': 1.202386960037984, 'learning_rate/qwen_vl_interface': 6.571053568581778e-06, 'learning_rate/action_model': 6.571053568581777e-05, 'epoch': 0.86}) +[2026-06-17 11:35:24,833][starVLA.training.train_starvla][INFO] - Step 2932, Loss: {'train/loss': 0.48756739497184753, 'train/grad_norm_pre_clip': 9.136576652526855, 'timing/data': 0.0003927210345864296, 'timing/model': 1.207023447030224, 'learning_rate/qwen_vl_interface': 6.568976065030453e-06, 'learning_rate/action_model': 6.568976065030453e-05, 'epoch': 0.86}) +[2026-06-17 11:35:29,637][starVLA.training.train_starvla][INFO] - Step 2933, Loss: {'train/loss': 0.6735983490943909, 'train/grad_norm_pre_clip': 10.58508586883545, 'timing/data': 0.00039833702612668276, 'timing/model': 1.214469442027621, 'learning_rate/qwen_vl_interface': 6.566898288053968e-06, 'learning_rate/action_model': 6.566898288053968e-05, 'epoch': 0.86}) +[2026-06-17 11:35:34,452][starVLA.training.train_starvla][INFO] - Step 2934, Loss: {'train/loss': 0.8077094554901123, 'train/grad_norm_pre_clip': 11.184220314025879, 'timing/data': 0.0004020240157842636, 'timing/model': 1.201700177974999, 'learning_rate/qwen_vl_interface': 6.564820238083053e-06, 'learning_rate/action_model': 6.564820238083052e-05, 'epoch': 0.86}) +[2026-06-17 11:35:39,265][starVLA.training.train_starvla][INFO] - Step 2935, Loss: {'train/loss': 0.5975741147994995, 'train/grad_norm_pre_clip': 10.137615203857422, 'timing/data': 0.000371636007912457, 'timing/model': 1.3551753349602222, 'learning_rate/qwen_vl_interface': 6.562741915548486e-06, 'learning_rate/action_model': 6.562741915548486e-05, 'epoch': 0.86}) +[2026-06-17 11:35:43,980][starVLA.training.train_starvla][INFO] - Step 2936, Loss: {'train/loss': 0.4777041971683502, 'train/grad_norm_pre_clip': 7.382531642913818, 'timing/data': 0.0003749449970200658, 'timing/model': 1.2523050099844113, 'learning_rate/qwen_vl_interface': 6.5606633208811096e-06, 'learning_rate/action_model': 6.560663320881109e-05, 'epoch': 0.86}) +[2026-06-17 11:35:48,639][starVLA.training.train_starvla][INFO] - Step 2937, Loss: {'train/loss': 0.93012535572052, 'train/grad_norm_pre_clip': 11.338550567626953, 'timing/data': 0.00037153693847358227, 'timing/model': 1.2080742219695821, 'learning_rate/qwen_vl_interface': 6.558584454511814e-06, 'learning_rate/action_model': 6.558584454511814e-05, 'epoch': 0.86}) +[2026-06-17 11:35:53,394][starVLA.training.train_starvla][INFO] - Step 2938, Loss: {'train/loss': 0.8950485587120056, 'train/grad_norm_pre_clip': 11.775550842285156, 'timing/data': 0.0004065249813720584, 'timing/model': 1.2076539180707186, 'learning_rate/qwen_vl_interface': 6.556505316871555e-06, 'learning_rate/action_model': 6.556505316871554e-05, 'epoch': 0.86}) +[2026-06-17 11:35:58,179][starVLA.training.train_starvla][INFO] - Step 2939, Loss: {'train/loss': 0.6862180233001709, 'train/grad_norm_pre_clip': 9.481568336486816, 'timing/data': 0.0003524860367178917, 'timing/model': 1.2072301929583773, 'learning_rate/qwen_vl_interface': 6.554425908391337e-06, 'learning_rate/action_model': 6.554425908391336e-05, 'epoch': 0.86}) +[2026-06-17 11:36:03,020][starVLA.training.train_starvla][INFO] - Step 2940, Loss: {'train/loss': 0.623500406742096, 'train/grad_norm_pre_clip': 11.131525993347168, 'timing/data': 0.00044139300007373095, 'timing/model': 1.2032343299360946, 'learning_rate/qwen_vl_interface': 6.552346229502224e-06, 'learning_rate/action_model': 6.552346229502224e-05, 'epoch': 0.87}) +[2026-06-17 11:36:07,851][starVLA.training.train_starvla][INFO] - Step 2941, Loss: {'train/loss': 0.6084828972816467, 'train/grad_norm_pre_clip': 14.013772010803223, 'timing/data': 0.0003488330403342843, 'timing/model': 1.3535916310502216, 'learning_rate/qwen_vl_interface': 6.550266280635337e-06, 'learning_rate/action_model': 6.550266280635336e-05, 'epoch': 0.87}) +[2026-06-17 11:36:12,513][starVLA.training.train_starvla][INFO] - Step 2942, Loss: {'train/loss': 0.805029034614563, 'train/grad_norm_pre_clip': 9.27117919921875, 'timing/data': 0.00035298510920256376, 'timing/model': 1.203815654036589, 'learning_rate/qwen_vl_interface': 6.54818606222185e-06, 'learning_rate/action_model': 6.548186062221849e-05, 'epoch': 0.87}) +[2026-06-17 11:36:17,321][starVLA.training.train_starvla][INFO] - Step 2943, Loss: {'train/loss': 0.4927094280719757, 'train/grad_norm_pre_clip': 11.790432929992676, 'timing/data': 0.00039864203426986933, 'timing/model': 1.2012936550891027, 'learning_rate/qwen_vl_interface': 6.546105574692997e-06, 'learning_rate/action_model': 6.546105574692997e-05, 'epoch': 0.87}) +[2026-06-17 11:36:22,098][starVLA.training.train_starvla][INFO] - Step 2944, Loss: {'train/loss': 0.8400177955627441, 'train/grad_norm_pre_clip': 10.496711730957031, 'timing/data': 0.00034368596971035004, 'timing/model': 1.2035935800522566, 'learning_rate/qwen_vl_interface': 6.544024818480061e-06, 'learning_rate/action_model': 6.544024818480062e-05, 'epoch': 0.87}) +[2026-06-17 11:36:26,882][starVLA.training.train_starvla][INFO] - Step 2945, Loss: {'train/loss': 0.7274605631828308, 'train/grad_norm_pre_clip': 10.570067405700684, 'timing/data': 0.0003617530455812812, 'timing/model': 1.2040414650691673, 'learning_rate/qwen_vl_interface': 6.5419437940143905e-06, 'learning_rate/action_model': 6.54194379401439e-05, 'epoch': 0.87}) +[2026-06-17 11:36:31,663][starVLA.training.train_starvla][INFO] - Step 2946, Loss: {'train/loss': 0.715522050857544, 'train/grad_norm_pre_clip': 10.646535873413086, 'timing/data': 0.0003504749620333314, 'timing/model': 1.2039698279695585, 'learning_rate/qwen_vl_interface': 6.5398625017273795e-06, 'learning_rate/action_model': 6.539862501727379e-05, 'epoch': 0.87}) +[2026-06-17 11:36:36,465][starVLA.training.train_starvla][INFO] - Step 2947, Loss: {'train/loss': 0.9152028560638428, 'train/grad_norm_pre_clip': 13.213078498840332, 'timing/data': 0.00035597605165094137, 'timing/model': 1.3355564440134913, 'learning_rate/qwen_vl_interface': 6.537780942050486e-06, 'learning_rate/action_model': 6.537780942050486e-05, 'epoch': 0.87}) +[2026-06-17 11:36:41,127][starVLA.training.train_starvla][INFO] - Step 2948, Loss: {'train/loss': 0.8821956515312195, 'train/grad_norm_pre_clip': 10.182218551635742, 'timing/data': 0.0003735870122909546, 'timing/model': 1.2064795690821484, 'learning_rate/qwen_vl_interface': 6.535699115415216e-06, 'learning_rate/action_model': 6.535699115415216e-05, 'epoch': 0.87}) +[2026-06-17 11:36:45,933][starVLA.training.train_starvla][INFO] - Step 2949, Loss: {'train/loss': 0.7767015099525452, 'train/grad_norm_pre_clip': 9.093087196350098, 'timing/data': 0.00038027099799364805, 'timing/model': 1.2064306599786505, 'learning_rate/qwen_vl_interface': 6.533617022253137e-06, 'learning_rate/action_model': 6.533617022253137e-05, 'epoch': 0.87}) +[2026-06-17 11:36:50,722][starVLA.training.train_starvla][INFO] - Step 2950, Loss: {'train/loss': 0.6521145701408386, 'train/grad_norm_pre_clip': 10.530618667602539, 'timing/data': 0.00037986994720995426, 'timing/model': 1.2072631190530956, 'learning_rate/qwen_vl_interface': 6.531534662995866e-06, 'learning_rate/action_model': 6.531534662995866e-05, 'epoch': 0.87}) +[2026-06-17 11:36:55,526][starVLA.training.train_starvla][INFO] - Step 2951, Loss: {'train/loss': 1.0729845762252808, 'train/grad_norm_pre_clip': 12.057609558105469, 'timing/data': 0.0003570360131561756, 'timing/model': 1.203026186907664, 'learning_rate/qwen_vl_interface': 6.529452038075082e-06, 'learning_rate/action_model': 6.529452038075081e-05, 'epoch': 0.87}) +[2026-06-17 11:37:00,296][starVLA.training.train_starvla][INFO] - Step 2952, Loss: {'train/loss': 0.5004767775535583, 'train/grad_norm_pre_clip': 8.210158348083496, 'timing/data': 0.0003607620019465685, 'timing/model': 1.2173757249256596, 'learning_rate/qwen_vl_interface': 6.5273691479225124e-06, 'learning_rate/action_model': 6.527369147922511e-05, 'epoch': 0.87}) +[2026-06-17 11:37:05,123][starVLA.training.train_starvla][INFO] - Step 2953, Loss: {'train/loss': 0.8664103150367737, 'train/grad_norm_pre_clip': 12.39345645904541, 'timing/data': 0.0007275630487129092, 'timing/model': 1.3548463810002431, 'learning_rate/qwen_vl_interface': 6.525285992969945e-06, 'learning_rate/action_model': 6.525285992969945e-05, 'epoch': 0.87}) +[2026-06-17 11:37:09,777][starVLA.training.train_starvla][INFO] - Step 2954, Loss: {'train/loss': 0.7468303442001343, 'train/grad_norm_pre_clip': 10.034196853637695, 'timing/data': 0.000351731083355844, 'timing/model': 1.2046338900690898, 'learning_rate/qwen_vl_interface': 6.523202573649219e-06, 'learning_rate/action_model': 6.523202573649219e-05, 'epoch': 0.87}) +[2026-06-17 11:37:14,557][starVLA.training.train_starvla][INFO] - Step 2955, Loss: {'train/loss': 0.6459509134292603, 'train/grad_norm_pre_clip': 10.114041328430176, 'timing/data': 0.00036205595824867487, 'timing/model': 1.2036789619596675, 'learning_rate/qwen_vl_interface': 6.5211188903922294e-06, 'learning_rate/action_model': 6.52111889039223e-05, 'epoch': 0.87}) +[2026-06-17 11:37:19,339][starVLA.training.train_starvla][INFO] - Step 2956, Loss: {'train/loss': 0.8126453161239624, 'train/grad_norm_pre_clip': 10.726076126098633, 'timing/data': 0.00034546595998108387, 'timing/model': 1.205924735055305, 'learning_rate/qwen_vl_interface': 6.519034943630926e-06, 'learning_rate/action_model': 6.519034943630927e-05, 'epoch': 0.87}) +[2026-06-17 11:37:24,103][starVLA.training.train_starvla][INFO] - Step 2957, Loss: {'train/loss': 0.6999021172523499, 'train/grad_norm_pre_clip': 10.644084930419922, 'timing/data': 0.0003395539242774248, 'timing/model': 1.2014353449922055, 'learning_rate/qwen_vl_interface': 6.5169507337973125e-06, 'learning_rate/action_model': 6.516950733797312e-05, 'epoch': 0.87}) +[2026-06-17 11:37:28,862][starVLA.training.train_starvla][INFO] - Step 2958, Loss: {'train/loss': 0.7904369235038757, 'train/grad_norm_pre_clip': 10.719136238098145, 'timing/data': 0.0003175900783389807, 'timing/model': 1.2034483469324186, 'learning_rate/qwen_vl_interface': 6.51486626132345e-06, 'learning_rate/action_model': 6.51486626132345e-05, 'epoch': 0.87}) +[2026-06-17 11:37:33,649][starVLA.training.train_starvla][INFO] - Step 2959, Loss: {'train/loss': 0.5598861575126648, 'train/grad_norm_pre_clip': 9.38640308380127, 'timing/data': 0.00035020499490201473, 'timing/model': 1.3349801870062947, 'learning_rate/qwen_vl_interface': 6.512781526641449e-06, 'learning_rate/action_model': 6.512781526641449e-05, 'epoch': 0.87}) +[2026-06-17 11:37:38,319][starVLA.training.train_starvla][INFO] - Step 2960, Loss: {'train/loss': 0.8559693098068237, 'train/grad_norm_pre_clip': 12.186166763305664, 'timing/data': 0.0003445519832894206, 'timing/model': 1.206288953893818, 'learning_rate/qwen_vl_interface': 6.51069653018348e-06, 'learning_rate/action_model': 6.51069653018348e-05, 'epoch': 0.87}) +[2026-06-17 11:37:43,118][starVLA.training.train_starvla][INFO] - Step 2961, Loss: {'train/loss': 0.6304249167442322, 'train/grad_norm_pre_clip': 8.291065216064453, 'timing/data': 0.00034944096114486456, 'timing/model': 1.2042597809340805, 'learning_rate/qwen_vl_interface': 6.508611272381763e-06, 'learning_rate/action_model': 6.508611272381763e-05, 'epoch': 0.87}) +[2026-06-17 11:37:47,910][starVLA.training.train_starvla][INFO] - Step 2962, Loss: {'train/loss': 1.0070679187774658, 'train/grad_norm_pre_clip': 12.079870223999023, 'timing/data': 0.0004175259964540601, 'timing/model': 1.2025687689892948, 'learning_rate/qwen_vl_interface': 6.506525753668576e-06, 'learning_rate/action_model': 6.506525753668576e-05, 'epoch': 0.87}) +[2026-06-17 11:37:52,856][starVLA.training.train_starvla][INFO] - Step 2963, Loss: {'train/loss': 0.4270382523536682, 'train/grad_norm_pre_clip': 6.996088981628418, 'timing/data': 0.0003610550193116069, 'timing/model': 1.2083915389375761, 'learning_rate/qwen_vl_interface': 6.504439974476246e-06, 'learning_rate/action_model': 6.504439974476246e-05, 'epoch': 0.87}) +[2026-06-17 11:37:57,670][starVLA.training.train_starvla][INFO] - Step 2964, Loss: {'train/loss': 0.7633124589920044, 'train/grad_norm_pre_clip': 12.26180362701416, 'timing/data': 0.0004306139890104532, 'timing/model': 1.339372129063122, 'learning_rate/qwen_vl_interface': 6.50235393523716e-06, 'learning_rate/action_model': 6.502353935237159e-05, 'epoch': 0.87}) +[2026-06-17 11:38:02,392][starVLA.training.train_starvla][INFO] - Step 2965, Loss: {'train/loss': 0.6708287000656128, 'train/grad_norm_pre_clip': 11.882549285888672, 'timing/data': 0.000420736032538116, 'timing/model': 1.2610095579875633, 'learning_rate/qwen_vl_interface': 6.500267636383758e-06, 'learning_rate/action_model': 6.500267636383758e-05, 'epoch': 0.87}) +[2026-06-17 11:38:07,075][starVLA.training.train_starvla][INFO] - Step 2966, Loss: {'train/loss': 0.8506009578704834, 'train/grad_norm_pre_clip': 10.461435317993164, 'timing/data': 0.0003800109261646867, 'timing/model': 1.2131580259883776, 'learning_rate/qwen_vl_interface': 6.498181078348527e-06, 'learning_rate/action_model': 6.498181078348527e-05, 'epoch': 0.87}) +[2026-06-17 11:38:11,846][starVLA.training.train_starvla][INFO] - Step 2967, Loss: {'train/loss': 0.8902674913406372, 'train/grad_norm_pre_clip': 11.927390098571777, 'timing/data': 0.00036655692383646965, 'timing/model': 1.2107166909845546, 'learning_rate/qwen_vl_interface': 6.496094261564018e-06, 'learning_rate/action_model': 6.496094261564018e-05, 'epoch': 0.87}) +[2026-06-17 11:38:16,671][starVLA.training.train_starvla][INFO] - Step 2968, Loss: {'train/loss': 0.5764198303222656, 'train/grad_norm_pre_clip': 8.985764503479004, 'timing/data': 0.000365325016900897, 'timing/model': 1.2025606529787183, 'learning_rate/qwen_vl_interface': 6.494007186462824e-06, 'learning_rate/action_model': 6.494007186462823e-05, 'epoch': 0.87}) +[2026-06-17 11:38:21,450][starVLA.training.train_starvla][INFO] - Step 2969, Loss: {'train/loss': 0.481685072183609, 'train/grad_norm_pre_clip': 6.466231346130371, 'timing/data': 0.0003498500445857644, 'timing/model': 1.2037979230517522, 'learning_rate/qwen_vl_interface': 6.491919853477604e-06, 'learning_rate/action_model': 6.491919853477604e-05, 'epoch': 0.87}) +[2026-06-17 11:38:26,240][starVLA.training.train_starvla][INFO] - Step 2970, Loss: {'train/loss': 0.8697243928909302, 'train/grad_norm_pre_clip': 12.53173828125, 'timing/data': 0.0003578040050342679, 'timing/model': 1.3410207570996135, 'learning_rate/qwen_vl_interface': 6.4898322630410615e-06, 'learning_rate/action_model': 6.489832263041062e-05, 'epoch': 0.87}) +[2026-06-17 11:38:30,908][starVLA.training.train_starvla][INFO] - Step 2971, Loss: {'train/loss': 0.8268574476242065, 'train/grad_norm_pre_clip': 9.61580753326416, 'timing/data': 0.00035818200558423996, 'timing/model': 1.2049597969744354, 'learning_rate/qwen_vl_interface': 6.487744415585958e-06, 'learning_rate/action_model': 6.487744415585958e-05, 'epoch': 0.87}) +[2026-06-17 11:38:35,736][starVLA.training.train_starvla][INFO] - Step 2972, Loss: {'train/loss': 0.8458840250968933, 'train/grad_norm_pre_clip': 14.84853744506836, 'timing/data': 0.0003615369787439704, 'timing/model': 1.2120711919851601, 'learning_rate/qwen_vl_interface': 6.485656311545103e-06, 'learning_rate/action_model': 6.485656311545103e-05, 'epoch': 0.87}) +[2026-06-17 11:38:40,556][starVLA.training.train_starvla][INFO] - Step 2973, Loss: {'train/loss': 1.0320994853973389, 'train/grad_norm_pre_clip': 15.945023536682129, 'timing/data': 0.0003632140578702092, 'timing/model': 1.2077363450080156, 'learning_rate/qwen_vl_interface': 6.483567951351366e-06, 'learning_rate/action_model': 6.483567951351365e-05, 'epoch': 0.87}) +[2026-06-17 11:38:45,357][starVLA.training.train_starvla][INFO] - Step 2974, Loss: {'train/loss': 0.9398403167724609, 'train/grad_norm_pre_clip': 12.359639167785645, 'timing/data': 0.00035659701097756624, 'timing/model': 1.209212590008974, 'learning_rate/qwen_vl_interface': 6.481479335437667e-06, 'learning_rate/action_model': 6.481479335437667e-05, 'epoch': 0.88}) +[2026-06-17 11:38:50,160][starVLA.training.train_starvla][INFO] - Step 2975, Loss: {'train/loss': 0.8398817181587219, 'train/grad_norm_pre_clip': 12.14739990234375, 'timing/data': 0.00036183407064527273, 'timing/model': 1.2095045780297369, 'learning_rate/qwen_vl_interface': 6.479390464236974e-06, 'learning_rate/action_model': 6.479390464236974e-05, 'epoch': 0.88}) +[2026-06-17 11:38:54,985][starVLA.training.train_starvla][INFO] - Step 2976, Loss: {'train/loss': 0.5098295211791992, 'train/grad_norm_pre_clip': 8.96523666381836, 'timing/data': 0.0003593210130929947, 'timing/model': 1.3487594299949706, 'learning_rate/qwen_vl_interface': 6.477301338182317e-06, 'learning_rate/action_model': 6.477301338182317e-05, 'epoch': 0.88}) +[2026-06-17 11:38:59,659][starVLA.training.train_starvla][INFO] - Step 2977, Loss: {'train/loss': 0.5518351793289185, 'train/grad_norm_pre_clip': 8.899189949035645, 'timing/data': 0.00035630492493510246, 'timing/model': 1.2106137230293825, 'learning_rate/qwen_vl_interface': 6.47521195770677e-06, 'learning_rate/action_model': 6.47521195770677e-05, 'epoch': 0.88}) +[2026-06-17 11:39:04,458][starVLA.training.train_starvla][INFO] - Step 2978, Loss: {'train/loss': 0.5239484906196594, 'train/grad_norm_pre_clip': 10.691747665405273, 'timing/data': 0.0003594179870560765, 'timing/model': 1.2050369940698147, 'learning_rate/qwen_vl_interface': 6.473122323243467e-06, 'learning_rate/action_model': 6.473122323243467e-05, 'epoch': 0.88}) +[2026-06-17 11:39:09,305][starVLA.training.train_starvla][INFO] - Step 2979, Loss: {'train/loss': 0.6129668354988098, 'train/grad_norm_pre_clip': 8.837867736816406, 'timing/data': 0.0006355489604175091, 'timing/model': 1.2464851690456271, 'learning_rate/qwen_vl_interface': 6.471032435225591e-06, 'learning_rate/action_model': 6.47103243522559e-05, 'epoch': 0.88}) +[2026-06-17 11:39:14,084][starVLA.training.train_starvla][INFO] - Step 2980, Loss: {'train/loss': 0.6888960599899292, 'train/grad_norm_pre_clip': 14.171566009521484, 'timing/data': 0.000396367977373302, 'timing/model': 1.2038383909966797, 'learning_rate/qwen_vl_interface': 6.468942294086376e-06, 'learning_rate/action_model': 6.468942294086376e-05, 'epoch': 0.88}) +[2026-06-17 11:39:18,816][starVLA.training.train_starvla][INFO] - Step 2981, Loss: {'train/loss': 0.793773889541626, 'train/grad_norm_pre_clip': 20.753557205200195, 'timing/data': 0.0003653500461950898, 'timing/model': 1.2043923240853474, 'learning_rate/qwen_vl_interface': 6.466851900259113e-06, 'learning_rate/action_model': 6.466851900259112e-05, 'epoch': 0.88}) +[2026-06-17 11:39:23,601][starVLA.training.train_starvla][INFO] - Step 2982, Loss: {'train/loss': 0.6048321723937988, 'train/grad_norm_pre_clip': 9.907127380371094, 'timing/data': 0.00037639192305505276, 'timing/model': 1.334582638926804, 'learning_rate/qwen_vl_interface': 6.464761254177139e-06, 'learning_rate/action_model': 6.464761254177138e-05, 'epoch': 0.88}) +[2026-06-17 11:39:28,256][starVLA.training.train_starvla][INFO] - Step 2983, Loss: {'train/loss': 0.5993401408195496, 'train/grad_norm_pre_clip': 7.464720726013184, 'timing/data': 0.00036651792470365763, 'timing/model': 1.2024220729945228, 'learning_rate/qwen_vl_interface': 6.462670356273852e-06, 'learning_rate/action_model': 6.462670356273852e-05, 'epoch': 0.88}) +[2026-06-17 11:39:33,035][starVLA.training.train_starvla][INFO] - Step 2984, Loss: {'train/loss': 0.9384425282478333, 'train/grad_norm_pre_clip': 12.327797889709473, 'timing/data': 0.0003769799368456006, 'timing/model': 1.2043120400048792, 'learning_rate/qwen_vl_interface': 6.460579206982695e-06, 'learning_rate/action_model': 6.460579206982696e-05, 'epoch': 0.88}) +[2026-06-17 11:39:37,808][starVLA.training.train_starvla][INFO] - Step 2985, Loss: {'train/loss': 0.7378298044204712, 'train/grad_norm_pre_clip': 11.42896556854248, 'timing/data': 0.0006387109169736505, 'timing/model': 1.2018733429722488, 'learning_rate/qwen_vl_interface': 6.458487806737166e-06, 'learning_rate/action_model': 6.458487806737165e-05, 'epoch': 0.88}) +[2026-06-17 11:39:42,595][starVLA.training.train_starvla][INFO] - Step 2986, Loss: {'train/loss': 1.0272067785263062, 'train/grad_norm_pre_clip': 13.11742877960205, 'timing/data': 0.00036954996176064014, 'timing/model': 1.2054681959562004, 'learning_rate/qwen_vl_interface': 6.456396155970814e-06, 'learning_rate/action_model': 6.456396155970814e-05, 'epoch': 0.88}) +[2026-06-17 11:39:47,369][starVLA.training.train_starvla][INFO] - Step 2987, Loss: {'train/loss': 0.8684333562850952, 'train/grad_norm_pre_clip': 13.265395164489746, 'timing/data': 0.00036076095420867205, 'timing/model': 1.2094204579479992, 'learning_rate/qwen_vl_interface': 6.4543042551172404e-06, 'learning_rate/action_model': 6.45430425511724e-05, 'epoch': 0.88}) +[2026-06-17 11:39:52,162][starVLA.training.train_starvla][INFO] - Step 2988, Loss: {'train/loss': 0.5276796817779541, 'train/grad_norm_pre_clip': 10.00649642944336, 'timing/data': 0.00035742297768592834, 'timing/model': 1.3434147059451789, 'learning_rate/qwen_vl_interface': 6.4522121046101e-06, 'learning_rate/action_model': 6.452212104610099e-05, 'epoch': 0.88}) +[2026-06-17 11:39:56,812][starVLA.training.train_starvla][INFO] - Step 2989, Loss: {'train/loss': 0.39829355478286743, 'train/grad_norm_pre_clip': 6.805962562561035, 'timing/data': 0.00033566902857273817, 'timing/model': 1.2039203910389915, 'learning_rate/qwen_vl_interface': 6.450119704883095e-06, 'learning_rate/action_model': 6.450119704883096e-05, 'epoch': 0.88}) +[2026-06-17 11:40:01,616][starVLA.training.train_starvla][INFO] - Step 2990, Loss: {'train/loss': 0.49229440093040466, 'train/grad_norm_pre_clip': 7.899308204650879, 'timing/data': 0.0003906979691237211, 'timing/model': 1.213039007037878, 'learning_rate/qwen_vl_interface': 6.448027056369983e-06, 'learning_rate/action_model': 6.448027056369983e-05, 'epoch': 0.88}) +[2026-06-17 11:40:06,401][starVLA.training.train_starvla][INFO] - Step 2991, Loss: {'train/loss': 0.5076131224632263, 'train/grad_norm_pre_clip': 13.212928771972656, 'timing/data': 0.0003614260349422693, 'timing/model': 1.2089154890272766, 'learning_rate/qwen_vl_interface': 6.445934159504576e-06, 'learning_rate/action_model': 6.445934159504576e-05, 'epoch': 0.88}) +[2026-06-17 11:40:11,187][starVLA.training.train_starvla][INFO] - Step 2992, Loss: {'train/loss': 1.0634559392929077, 'train/grad_norm_pre_clip': 13.089982986450195, 'timing/data': 0.0003849300555884838, 'timing/model': 1.2076641500461847, 'learning_rate/qwen_vl_interface': 6.44384101472073e-06, 'learning_rate/action_model': 6.443841014720729e-05, 'epoch': 0.88}) +[2026-06-17 11:40:15,993][starVLA.training.train_starvla][INFO] - Step 2993, Loss: {'train/loss': 0.6032909154891968, 'train/grad_norm_pre_clip': 7.2143635749816895, 'timing/data': 0.00043621601071208715, 'timing/model': 1.3536804880714044, 'learning_rate/qwen_vl_interface': 6.441747622452357e-06, 'learning_rate/action_model': 6.441747622452356e-05, 'epoch': 0.88}) +[2026-06-17 11:40:20,693][starVLA.training.train_starvla][INFO] - Step 2994, Loss: {'train/loss': 0.9546576738357544, 'train/grad_norm_pre_clip': 13.207916259765625, 'timing/data': 0.00038194400258362293, 'timing/model': 1.2489818220492452, 'learning_rate/qwen_vl_interface': 6.4396539831334195e-06, 'learning_rate/action_model': 6.439653983133419e-05, 'epoch': 0.88}) +[2026-06-17 11:40:25,353][starVLA.training.train_starvla][INFO] - Step 2995, Loss: {'train/loss': 0.6228981018066406, 'train/grad_norm_pre_clip': 9.426321983337402, 'timing/data': 0.0003440129803493619, 'timing/model': 1.2057673400267959, 'learning_rate/qwen_vl_interface': 6.437560097197928e-06, 'learning_rate/action_model': 6.437560097197927e-05, 'epoch': 0.88}) +[2026-06-17 11:40:30,087][starVLA.training.train_starvla][INFO] - Step 2996, Loss: {'train/loss': 0.6790916919708252, 'train/grad_norm_pre_clip': 14.60290241241455, 'timing/data': 0.0003434049431234598, 'timing/model': 1.2026645140722394, 'learning_rate/qwen_vl_interface': 6.4354659650799546e-06, 'learning_rate/action_model': 6.435465965079955e-05, 'epoch': 0.88}) +[2026-06-17 11:40:34,884][starVLA.training.train_starvla][INFO] - Step 2997, Loss: {'train/loss': 0.740513026714325, 'train/grad_norm_pre_clip': 9.44806957244873, 'timing/data': 0.0003426650073379278, 'timing/model': 1.2099193419562653, 'learning_rate/qwen_vl_interface': 6.433371587213609e-06, 'learning_rate/action_model': 6.433371587213607e-05, 'epoch': 0.88}) +[2026-06-17 11:40:39,689][starVLA.training.train_starvla][INFO] - Step 2998, Loss: {'train/loss': 0.5233307480812073, 'train/grad_norm_pre_clip': 8.292439460754395, 'timing/data': 0.000381608959287405, 'timing/model': 1.2064166279742494, 'learning_rate/qwen_vl_interface': 6.431276964033062e-06, 'learning_rate/action_model': 6.431276964033062e-05, 'epoch': 0.88}) +[2026-06-17 11:40:44,483][starVLA.training.train_starvla][INFO] - Step 2999, Loss: {'train/loss': 0.7092621326446533, 'train/grad_norm_pre_clip': 19.22372055053711, 'timing/data': 0.0003808080218732357, 'timing/model': 1.3383216399233788, 'learning_rate/qwen_vl_interface': 6.429182095972526e-06, 'learning_rate/action_model': 6.429182095972526e-05, 'epoch': 0.88}) +[2026-06-17 11:50:52,833][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 3000 +[2026-06-17 11:50:52,834][starVLA.training.train_starvla][INFO] - Step 3000, Loss: {'train/loss': 0.6811566352844238, 'train/grad_norm_pre_clip': 9.465124130249023, 'eval/loss': 0.7952085733413696, 'eval/latency_0/loss': 0.7952085733413696, 'eval/demon_fire_precision': 0.8935496636327661, 'eval/demon_fire_recall': 0.9377076411960132, 'eval/demon_fire_f1': 0.9150962512664639, 'eval/demon_fire_rate_model': 0.7027252502780867, 'eval/demon_fire_rate_teacher': 0.6696329254727474, 'eval/demon_left_precision': 0.8433503836317136, 'eval/demon_left_recall': 0.7473087818696884, 'eval/demon_left_f1': 0.792430159206969, 'eval/demon_left_rate_model': 0.43492769744160176, 'eval/demon_left_rate_teacher': 0.49082313681868744, 'eval/demon_right_precision': 0.7714708785784797, 'eval/demon_right_recall': 0.8830508474576271, 'eval/demon_right_f1': 0.8234984193888304, 'eval/demon_right_rate_model': 0.5634037819799778, 'eval/demon_right_rate_teacher': 0.4922135706340378, 'eval/demon_move_f1': 0.8079642892978998, 'eval/demon_cc_f1': 0.8800877031445365, 'eval/latency_0/demon_fire_precision': 0.8935496636327661, 'eval/latency_0/demon_fire_recall': 0.9377076411960132, 'eval/latency_0/demon_fire_f1': 0.9150962512664639, 'eval/latency_0/demon_fire_rate_model': 0.7027252502780867, 'eval/latency_0/demon_fire_rate_teacher': 0.6696329254727474, 'eval/latency_0/demon_left_precision': 0.8433503836317136, 'eval/latency_0/demon_left_recall': 0.7473087818696884, 'eval/latency_0/demon_left_f1': 0.792430159206969, 'eval/latency_0/demon_left_rate_model': 0.43492769744160176, 'eval/latency_0/demon_left_rate_teacher': 0.49082313681868744, 'eval/latency_0/demon_right_precision': 0.7714708785784797, 'eval/latency_0/demon_right_recall': 0.8830508474576271, 'eval/latency_0/demon_right_f1': 0.8234984193888304, 'eval/latency_0/demon_right_rate_model': 0.5634037819799778, 'eval/latency_0/demon_right_rate_teacher': 0.4922135706340378, 'eval/latency_0/demon_move_f1': 0.8079642892978998, 'eval/latency_0/demon_cc_f1': 0.8800877031445365, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 4170.0, 'rl_games_eval/mid_train/mean_length': 3600.0, 'rl_games_eval/mid_train/std_reward': 3118.414661330337, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/macro_mean_reward': 4170.0, 'rl_games_eval/mid_train/macro_mean_length': 3600.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/demon_attack__latency_0/mean_reward': 4170.0, 'rl_games_eval/mid_train/demon_attack__latency_0/mean_length': 3600.0, 'rl_games_eval/mid_train/demon_attack__latency_0/std_reward': 3118.414661330337, 'rl_games_eval/mid_train/demon_attack__latency_0/std_length': 0.0, 'rl_games_eval/mid_train/demon_attack__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 4170.0, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.00034926494117826223, 'timing/model': 1.2039320819312707, 'learning_rate/qwen_vl_interface': 6.427086983466275e-06, 'learning_rate/action_model': 6.427086983466275e-05, 'epoch': 0.88}) +[2026-06-17 11:50:52,834][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_3000_state +[2026-06-17 11:50:52,837][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 11:50:56,374][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_3000_state/model.safetensors +[2026-06-17 11:51:03,910][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_3000_state/optimizer.bin +[2026-06-17 11:51:03,910][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_3000_state/sampler.bin +[2026-06-17 11:51:03,911][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_3000_state/sampler_1.bin +[2026-06-17 11:51:03,911][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_3000_state/random_states_0.pkl +[2026-06-17 11:51:03,954][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3000 +[2026-06-17 11:51:03,954][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 11:51:03,964][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 11:51:08,562][starVLA.training.train_starvla][INFO] - Step 3001, Loss: {'train/loss': 0.6708146333694458, 'train/grad_norm_pre_clip': 11.13609504699707, 'timing/data': 0.0004396130098029971, 'timing/model': 1.189083067001775, 'learning_rate/qwen_vl_interface': 6.4249916269486255e-06, 'learning_rate/action_model': 6.424991626948626e-05, 'epoch': 0.88}) +[2026-06-17 11:51:13,280][starVLA.training.train_starvla][INFO] - Step 3002, Loss: {'train/loss': 0.5891076922416687, 'train/grad_norm_pre_clip': 9.766380310058594, 'timing/data': 0.0004659510450437665, 'timing/model': 1.1893997839652002, 'learning_rate/qwen_vl_interface': 6.422896026853948e-06, 'learning_rate/action_model': 6.422896026853947e-05, 'epoch': 0.88}) +[2026-06-17 11:51:18,007][starVLA.training.train_starvla][INFO] - Step 3003, Loss: {'train/loss': 0.563669741153717, 'train/grad_norm_pre_clip': 16.694616317749023, 'timing/data': 0.0003179550403729081, 'timing/model': 1.1919411040144041, 'learning_rate/qwen_vl_interface': 6.420800183616661e-06, 'learning_rate/action_model': 6.420800183616661e-05, 'epoch': 0.88}) +[2026-06-17 11:51:22,743][starVLA.training.train_starvla][INFO] - Step 3004, Loss: {'train/loss': 0.7185144424438477, 'train/grad_norm_pre_clip': 11.660331726074219, 'timing/data': 0.00036085699684917927, 'timing/model': 1.192982433945872, 'learning_rate/qwen_vl_interface': 6.4187040976712355e-06, 'learning_rate/action_model': 6.418704097671235e-05, 'epoch': 0.88}) +[2026-06-17 11:51:27,490][starVLA.training.train_starvla][INFO] - Step 3005, Loss: {'train/loss': 0.869516134262085, 'train/grad_norm_pre_clip': 13.028095245361328, 'timing/data': 0.0003890040097758174, 'timing/model': 1.321960919070989, 'learning_rate/qwen_vl_interface': 6.416607769452195e-06, 'learning_rate/action_model': 6.416607769452194e-05, 'epoch': 0.88}) +[2026-06-17 11:51:32,106][starVLA.training.train_starvla][INFO] - Step 3006, Loss: {'train/loss': 0.6787261962890625, 'train/grad_norm_pre_clip': 10.208550453186035, 'timing/data': 0.0003749110037460923, 'timing/model': 1.1970760490512475, 'learning_rate/qwen_vl_interface': 6.414511199394106e-06, 'learning_rate/action_model': 6.414511199394105e-05, 'epoch': 0.88}) +[2026-06-17 11:51:36,847][starVLA.training.train_starvla][INFO] - Step 3007, Loss: {'train/loss': 0.5728099346160889, 'train/grad_norm_pre_clip': 13.370390892028809, 'timing/data': 0.0003589800326153636, 'timing/model': 1.1971519149374217, 'learning_rate/qwen_vl_interface': 6.4124143879315935e-06, 'learning_rate/action_model': 6.412414387931593e-05, 'epoch': 0.88}) +[2026-06-17 11:51:41,508][starVLA.training.train_starvla][INFO] - Step 3008, Loss: {'train/loss': 0.6792609691619873, 'train/grad_norm_pre_clip': 26.842374801635742, 'timing/data': 0.00045101193245500326, 'timing/model': 1.1947627280605957, 'learning_rate/qwen_vl_interface': 6.410317335499328e-06, 'learning_rate/action_model': 6.410317335499328e-05, 'epoch': 0.89}) +[2026-06-17 11:51:46,212][starVLA.training.train_starvla][INFO] - Step 3009, Loss: {'train/loss': 0.7061680555343628, 'train/grad_norm_pre_clip': 10.740274429321289, 'timing/data': 0.0004745590267702937, 'timing/model': 1.1995500609045848, 'learning_rate/qwen_vl_interface': 6.40822004253203e-06, 'learning_rate/action_model': 6.40822004253203e-05, 'epoch': 0.89}) +[2026-06-17 11:51:50,948][starVLA.training.train_starvla][INFO] - Step 3010, Loss: {'train/loss': 1.080405831336975, 'train/grad_norm_pre_clip': 17.234411239624023, 'timing/data': 0.00033642095513641834, 'timing/model': 1.1955925449728966, 'learning_rate/qwen_vl_interface': 6.406122509464471e-06, 'learning_rate/action_model': 6.40612250946447e-05, 'epoch': 0.89}) +[2026-06-17 11:51:55,709][starVLA.training.train_starvla][INFO] - Step 3011, Loss: {'train/loss': 0.4505876898765564, 'train/grad_norm_pre_clip': 9.109318733215332, 'timing/data': 0.00034945295192301273, 'timing/model': 1.326549672987312, 'learning_rate/qwen_vl_interface': 6.404024736731472e-06, 'learning_rate/action_model': 6.404024736731472e-05, 'epoch': 0.89}) +[2026-06-17 11:52:00,349][starVLA.training.train_starvla][INFO] - Step 3012, Loss: {'train/loss': 0.8329527974128723, 'train/grad_norm_pre_clip': 21.456079483032227, 'timing/data': 0.0003812799695879221, 'timing/model': 1.2059001229936257, 'learning_rate/qwen_vl_interface': 6.401926724767903e-06, 'learning_rate/action_model': 6.401926724767902e-05, 'epoch': 0.89}) +[2026-06-17 11:52:05,120][starVLA.training.train_starvla][INFO] - Step 3013, Loss: {'train/loss': 0.8275848031044006, 'train/grad_norm_pre_clip': 11.48990249633789, 'timing/data': 0.0003930269740521908, 'timing/model': 1.2002946769353002, 'learning_rate/qwen_vl_interface': 6.399828474008683e-06, 'learning_rate/action_model': 6.399828474008682e-05, 'epoch': 0.89}) +[2026-06-17 11:52:09,887][starVLA.training.train_starvla][INFO] - Step 3014, Loss: {'train/loss': 0.8253219127655029, 'train/grad_norm_pre_clip': 13.65949535369873, 'timing/data': 0.00042724201921373606, 'timing/model': 1.1996081649558619, 'learning_rate/qwen_vl_interface': 6.397729984888785e-06, 'learning_rate/action_model': 6.397729984888785e-05, 'epoch': 0.89}) +[2026-06-17 11:52:14,672][starVLA.training.train_starvla][INFO] - Step 3015, Loss: {'train/loss': 0.759254515171051, 'train/grad_norm_pre_clip': 18.84113311767578, 'timing/data': 0.0003657930064946413, 'timing/model': 1.209200558019802, 'learning_rate/qwen_vl_interface': 6.395631257843224e-06, 'learning_rate/action_model': 6.395631257843224e-05, 'epoch': 0.89}) +[2026-06-17 11:52:19,438][starVLA.training.train_starvla][INFO] - Step 3016, Loss: {'train/loss': 0.5178864002227783, 'train/grad_norm_pre_clip': 10.17190170288086, 'timing/data': 0.0003648150013759732, 'timing/model': 1.3244234479498118, 'learning_rate/qwen_vl_interface': 6.393532293307073e-06, 'learning_rate/action_model': 6.393532293307072e-05, 'epoch': 0.89}) +[2026-06-17 11:52:24,128][starVLA.training.train_starvla][INFO] - Step 3017, Loss: {'train/loss': 0.6966987252235413, 'train/grad_norm_pre_clip': 11.68230152130127, 'timing/data': 0.0003951039398089051, 'timing/model': 1.2439376510446891, 'learning_rate/qwen_vl_interface': 6.391433091715447e-06, 'learning_rate/action_model': 6.391433091715446e-05, 'epoch': 0.89}) +[2026-06-17 11:52:28,785][starVLA.training.train_starvla][INFO] - Step 3018, Loss: {'train/loss': 0.7521438598632812, 'train/grad_norm_pre_clip': 13.71457576751709, 'timing/data': 0.0006047420902177691, 'timing/model': 1.208312762901187, 'learning_rate/qwen_vl_interface': 6.389333653503511e-06, 'learning_rate/action_model': 6.38933365350351e-05, 'epoch': 0.89}) +[2026-06-17 11:52:33,505][starVLA.training.train_starvla][INFO] - Step 3019, Loss: {'train/loss': 0.6714163422584534, 'train/grad_norm_pre_clip': 13.704078674316406, 'timing/data': 0.0004244609735906124, 'timing/model': 1.202756336890161, 'learning_rate/qwen_vl_interface': 6.387233979106484e-06, 'learning_rate/action_model': 6.387233979106483e-05, 'epoch': 0.89}) +[2026-06-17 11:52:38,273][starVLA.training.train_starvla][INFO] - Step 3020, Loss: {'train/loss': 0.8581680655479431, 'train/grad_norm_pre_clip': 16.899471282958984, 'timing/data': 0.00036575901322066784, 'timing/model': 1.2000980759039521, 'learning_rate/qwen_vl_interface': 6.385134068959629e-06, 'learning_rate/action_model': 6.385134068959628e-05, 'epoch': 0.89}) +[2026-06-17 11:52:43,052][starVLA.training.train_starvla][INFO] - Step 3021, Loss: {'train/loss': 0.5728188157081604, 'train/grad_norm_pre_clip': 14.031837463378906, 'timing/data': 0.00036906893365085125, 'timing/model': 1.2022817919496447, 'learning_rate/qwen_vl_interface': 6.3830339234982596e-06, 'learning_rate/action_model': 6.38303392349826e-05, 'epoch': 0.89}) +[2026-06-17 11:52:47,824][starVLA.training.train_starvla][INFO] - Step 3022, Loss: {'train/loss': 0.56949782371521, 'train/grad_norm_pre_clip': 8.898042678833008, 'timing/data': 0.00033826997969299555, 'timing/model': 1.3317046420415863, 'learning_rate/qwen_vl_interface': 6.38093354315774e-06, 'learning_rate/action_model': 6.38093354315774e-05, 'epoch': 0.89}) +[2026-06-17 11:52:52,468][starVLA.training.train_starvla][INFO] - Step 3023, Loss: {'train/loss': 0.8336756825447083, 'train/grad_norm_pre_clip': 9.885708808898926, 'timing/data': 0.00035174807999283075, 'timing/model': 1.2007364450255409, 'learning_rate/qwen_vl_interface': 6.378832928373479e-06, 'learning_rate/action_model': 6.378832928373478e-05, 'epoch': 0.89}) +[2026-06-17 11:52:57,238][starVLA.training.train_starvla][INFO] - Step 3024, Loss: {'train/loss': 0.5989095568656921, 'train/grad_norm_pre_clip': 9.634607315063477, 'timing/data': 0.00038678792770951986, 'timing/model': 1.2026682250434533, 'learning_rate/qwen_vl_interface': 6.376732079580936e-06, 'learning_rate/action_model': 6.376732079580936e-05, 'epoch': 0.89}) +[2026-06-17 11:53:02,037][starVLA.training.train_starvla][INFO] - Step 3025, Loss: {'train/loss': 0.7086573243141174, 'train/grad_norm_pre_clip': 10.920842170715332, 'timing/data': 0.00042356597259640694, 'timing/model': 1.2048826720565557, 'learning_rate/qwen_vl_interface': 6.374630997215623e-06, 'learning_rate/action_model': 6.374630997215623e-05, 'epoch': 0.89}) +[2026-06-17 11:53:06,841][starVLA.training.train_starvla][INFO] - Step 3026, Loss: {'train/loss': 0.47183510661125183, 'train/grad_norm_pre_clip': 8.222750663757324, 'timing/data': 0.0004409520188346505, 'timing/model': 1.2043780620442703, 'learning_rate/qwen_vl_interface': 6.372529681713093e-06, 'learning_rate/action_model': 6.372529681713094e-05, 'epoch': 0.89}) +[2026-06-17 11:53:11,626][starVLA.training.train_starvla][INFO] - Step 3027, Loss: {'train/loss': 1.1280467510223389, 'train/grad_norm_pre_clip': 13.51344108581543, 'timing/data': 0.0005825590342283249, 'timing/model': 1.2034582999767736, 'learning_rate/qwen_vl_interface': 6.370428133508953e-06, 'learning_rate/action_model': 6.370428133508953e-05, 'epoch': 0.89}) +[2026-06-17 11:53:16,421][starVLA.training.train_starvla][INFO] - Step 3028, Loss: {'train/loss': 0.8568809032440186, 'train/grad_norm_pre_clip': 17.31995964050293, 'timing/data': 0.00042508007027208805, 'timing/model': 1.351655664970167, 'learning_rate/qwen_vl_interface': 6.368326353038852e-06, 'learning_rate/action_model': 6.368326353038853e-05, 'epoch': 0.89}) +[2026-06-17 11:53:21,081][starVLA.training.train_starvla][INFO] - Step 3029, Loss: {'train/loss': 0.5992715358734131, 'train/grad_norm_pre_clip': 8.138543128967285, 'timing/data': 0.0003458170685917139, 'timing/model': 1.204481861088425, 'learning_rate/qwen_vl_interface': 6.366224340738499e-06, 'learning_rate/action_model': 6.366224340738499e-05, 'epoch': 0.89}) +[2026-06-17 11:53:25,868][starVLA.training.train_starvla][INFO] - Step 3030, Loss: {'train/loss': 0.8586939573287964, 'train/grad_norm_pre_clip': 13.622160911560059, 'timing/data': 0.0007028479594737291, 'timing/model': 1.204444871051237, 'learning_rate/qwen_vl_interface': 6.3641220970436355e-06, 'learning_rate/action_model': 6.364122097043635e-05, 'epoch': 0.89}) +[2026-06-17 11:53:30,644][starVLA.training.train_starvla][INFO] - Step 3031, Loss: {'train/loss': 0.44494128227233887, 'train/grad_norm_pre_clip': 7.1590752601623535, 'timing/data': 0.0005880380049347878, 'timing/model': 1.207465612096712, 'learning_rate/qwen_vl_interface': 6.362019622390065e-06, 'learning_rate/action_model': 6.362019622390066e-05, 'epoch': 0.89}) +[2026-06-17 11:53:35,438][starVLA.training.train_starvla][INFO] - Step 3032, Loss: {'train/loss': 0.618208110332489, 'train/grad_norm_pre_clip': 8.966824531555176, 'timing/data': 0.00037540006451308727, 'timing/model': 1.2078276929678395, 'learning_rate/qwen_vl_interface': 6.359916917213629e-06, 'learning_rate/action_model': 6.359916917213629e-05, 'epoch': 0.89}) +[2026-06-17 11:53:40,175][starVLA.training.train_starvla][INFO] - Step 3033, Loss: {'train/loss': 0.5928434729576111, 'train/grad_norm_pre_clip': 11.861410140991211, 'timing/data': 0.0004127329448238015, 'timing/model': 1.2024100850103423, 'learning_rate/qwen_vl_interface': 6.357813981950223e-06, 'learning_rate/action_model': 6.357813981950223e-05, 'epoch': 0.89}) +[2026-06-17 11:53:44,955][starVLA.training.train_starvla][INFO] - Step 3034, Loss: {'train/loss': 0.5120709538459778, 'train/grad_norm_pre_clip': 10.186507225036621, 'timing/data': 0.00041127600707113743, 'timing/model': 1.334658648003824, 'learning_rate/qwen_vl_interface': 6.355710817035784e-06, 'learning_rate/action_model': 6.355710817035783e-05, 'epoch': 0.89}) +[2026-06-17 11:53:49,622][starVLA.training.train_starvla][INFO] - Step 3035, Loss: {'train/loss': 0.6536892652511597, 'train/grad_norm_pre_clip': 9.98274040222168, 'timing/data': 0.00035622005816549063, 'timing/model': 1.2045404720120132, 'learning_rate/qwen_vl_interface': 6.353607422906304e-06, 'learning_rate/action_model': 6.353607422906304e-05, 'epoch': 0.89}) +[2026-06-17 11:53:54,410][starVLA.training.train_starvla][INFO] - Step 3036, Loss: {'train/loss': 0.6163139343261719, 'train/grad_norm_pre_clip': 10.24746036529541, 'timing/data': 0.0005602609599009156, 'timing/model': 1.203787176986225, 'learning_rate/qwen_vl_interface': 6.351503799997819e-06, 'learning_rate/action_model': 6.351503799997818e-05, 'epoch': 0.89}) +[2026-06-17 11:53:59,188][starVLA.training.train_starvla][INFO] - Step 3037, Loss: {'train/loss': 0.8254467248916626, 'train/grad_norm_pre_clip': 11.45549488067627, 'timing/data': 0.0007984600961208344, 'timing/model': 1.2051226700423285, 'learning_rate/qwen_vl_interface': 6.349399948746409e-06, 'learning_rate/action_model': 6.349399948746409e-05, 'epoch': 0.89}) +[2026-06-17 11:54:03,971][starVLA.training.train_starvla][INFO] - Step 3038, Loss: {'train/loss': 0.6530497074127197, 'train/grad_norm_pre_clip': 10.521312713623047, 'timing/data': 0.0003249020082876086, 'timing/model': 1.2039593999506906, 'learning_rate/qwen_vl_interface': 6.347295869588208e-06, 'learning_rate/action_model': 6.347295869588208e-05, 'epoch': 0.89}) +[2026-06-17 11:54:08,741][starVLA.training.train_starvla][INFO] - Step 3039, Loss: {'train/loss': 0.4722347855567932, 'train/grad_norm_pre_clip': 9.03124713897705, 'timing/data': 0.00037526502273976803, 'timing/model': 1.2026509359711781, 'learning_rate/qwen_vl_interface': 6.345191562959392e-06, 'learning_rate/action_model': 6.345191562959391e-05, 'epoch': 0.89}) +[2026-06-17 11:54:13,524][starVLA.training.train_starvla][INFO] - Step 3040, Loss: {'train/loss': 0.7411580681800842, 'train/grad_norm_pre_clip': 12.088637351989746, 'timing/data': 0.0003851429792121053, 'timing/model': 1.3334000429604203, 'learning_rate/qwen_vl_interface': 6.343087029296187e-06, 'learning_rate/action_model': 6.343087029296186e-05, 'epoch': 0.89}) +[2026-06-17 11:54:18,178][starVLA.training.train_starvla][INFO] - Step 3041, Loss: {'train/loss': 0.49190518260002136, 'train/grad_norm_pre_clip': 12.53629207611084, 'timing/data': 0.0003740119282156229, 'timing/model': 1.2048798589967191, 'learning_rate/qwen_vl_interface': 6.340982269034863e-06, 'learning_rate/action_model': 6.340982269034863e-05, 'epoch': 0.89}) +[2026-06-17 11:54:22,966][starVLA.training.train_starvla][INFO] - Step 3042, Loss: {'train/loss': 0.5790376663208008, 'train/grad_norm_pre_clip': 8.32812213897705, 'timing/data': 0.00036438496317714453, 'timing/model': 1.2032378689618781, 'learning_rate/qwen_vl_interface': 6.3388772826117415e-06, 'learning_rate/action_model': 6.338877282611742e-05, 'epoch': 0.9}) +[2026-06-17 11:54:27,757][starVLA.training.train_starvla][INFO] - Step 3043, Loss: {'train/loss': 0.8377667665481567, 'train/grad_norm_pre_clip': 13.078892707824707, 'timing/data': 0.0003655980108305812, 'timing/model': 1.2050112759461626, 'learning_rate/qwen_vl_interface': 6.3367720704631895e-06, 'learning_rate/action_model': 6.33677207046319e-05, 'epoch': 0.9}) +[2026-06-17 11:54:32,610][starVLA.training.train_starvla][INFO] - Step 3044, Loss: {'train/loss': 0.6114934086799622, 'train/grad_norm_pre_clip': 11.481730461120605, 'timing/data': 0.0003783910069614649, 'timing/model': 1.205554286018014, 'learning_rate/qwen_vl_interface': 6.334666633025616e-06, 'learning_rate/action_model': 6.334666633025615e-05, 'epoch': 0.9}) +[2026-06-17 11:54:37,400][starVLA.training.train_starvla][INFO] - Step 3045, Loss: {'train/loss': 0.7980977892875671, 'train/grad_norm_pre_clip': 11.242563247680664, 'timing/data': 0.0004146859282627702, 'timing/model': 1.33746973995585, 'learning_rate/qwen_vl_interface': 6.332560970735482e-06, 'learning_rate/action_model': 6.332560970735482e-05, 'epoch': 0.9}) +[2026-06-17 11:54:42,107][starVLA.training.train_starvla][INFO] - Step 3046, Loss: {'train/loss': 0.7673851251602173, 'train/grad_norm_pre_clip': 10.736603736877441, 'timing/data': 0.0003971339901909232, 'timing/model': 1.2567359749227762, 'learning_rate/qwen_vl_interface': 6.3304550840292945e-06, 'learning_rate/action_model': 6.330455084029294e-05, 'epoch': 0.9}) +[2026-06-17 11:54:46,780][starVLA.training.train_starvla][INFO] - Step 3047, Loss: {'train/loss': 0.9096404910087585, 'train/grad_norm_pre_clip': 50.863243103027344, 'timing/data': 0.0003900530282407999, 'timing/model': 1.2052334840409458, 'learning_rate/qwen_vl_interface': 6.328348973343605e-06, 'learning_rate/action_model': 6.328348973343605e-05, 'epoch': 0.9}) +[2026-06-17 11:54:51,516][starVLA.training.train_starvla][INFO] - Step 3048, Loss: {'train/loss': 0.8441747426986694, 'train/grad_norm_pre_clip': 9.544992446899414, 'timing/data': 0.0003858960699290037, 'timing/model': 1.204598851967603, 'learning_rate/qwen_vl_interface': 6.326242639115014e-06, 'learning_rate/action_model': 6.326242639115015e-05, 'epoch': 0.9}) +[2026-06-17 11:54:56,302][starVLA.training.train_starvla][INFO] - Step 3049, Loss: {'train/loss': 0.7549949288368225, 'train/grad_norm_pre_clip': 10.445748329162598, 'timing/data': 0.0003985939547419548, 'timing/model': 1.204263168037869, 'learning_rate/qwen_vl_interface': 6.324136081780166e-06, 'learning_rate/action_model': 6.324136081780165e-05, 'epoch': 0.9}) +[2026-06-17 11:55:01,085][starVLA.training.train_starvla][INFO] - Step 3050, Loss: {'train/loss': 0.602601170539856, 'train/grad_norm_pre_clip': 9.096538543701172, 'timing/data': 0.0005688669625669718, 'timing/model': 1.2046305909752846, 'learning_rate/qwen_vl_interface': 6.322029301775751e-06, 'learning_rate/action_model': 6.32202930177575e-05, 'epoch': 0.9}) +[2026-06-17 11:55:05,873][starVLA.training.train_starvla][INFO] - Step 3051, Loss: {'train/loss': 0.7084932327270508, 'train/grad_norm_pre_clip': 10.765803337097168, 'timing/data': 0.00036051205825060606, 'timing/model': 1.330282784998417, 'learning_rate/qwen_vl_interface': 6.3199222995385105e-06, 'learning_rate/action_model': 6.31992229953851e-05, 'epoch': 0.9}) +[2026-06-17 11:55:10,553][starVLA.training.train_starvla][INFO] - Step 3052, Loss: {'train/loss': 0.4706338047981262, 'train/grad_norm_pre_clip': 8.226461410522461, 'timing/data': 0.0003546319203451276, 'timing/model': 1.2066486929543316, 'learning_rate/qwen_vl_interface': 6.317815075505223e-06, 'learning_rate/action_model': 6.317815075505222e-05, 'epoch': 0.9}) +[2026-06-17 11:55:15,354][starVLA.training.train_starvla][INFO] - Step 3053, Loss: {'train/loss': 0.8774784207344055, 'train/grad_norm_pre_clip': 12.87668514251709, 'timing/data': 0.0005162579473108053, 'timing/model': 1.2097521339310333, 'learning_rate/qwen_vl_interface': 6.315707630112724e-06, 'learning_rate/action_model': 6.315707630112723e-05, 'epoch': 0.9}) +[2026-06-17 11:55:20,153][starVLA.training.train_starvla][INFO] - Step 3054, Loss: {'train/loss': 0.5615027546882629, 'train/grad_norm_pre_clip': 9.247629165649414, 'timing/data': 0.00035934196785092354, 'timing/model': 1.2096212380565703, 'learning_rate/qwen_vl_interface': 6.313599963797884e-06, 'learning_rate/action_model': 6.313599963797884e-05, 'epoch': 0.9}) +[2026-06-17 11:55:25,009][starVLA.training.train_starvla][INFO] - Step 3055, Loss: {'train/loss': 0.9233260154724121, 'train/grad_norm_pre_clip': 14.119293212890625, 'timing/data': 0.00035383901558816433, 'timing/model': 1.2070844440022483, 'learning_rate/qwen_vl_interface': 6.311492076997629e-06, 'learning_rate/action_model': 6.311492076997629e-05, 'epoch': 0.9}) +[2026-06-17 11:55:29,801][starVLA.training.train_starvla][INFO] - Step 3056, Loss: {'train/loss': 0.7691720128059387, 'train/grad_norm_pre_clip': 11.225637435913086, 'timing/data': 0.00036726100370287895, 'timing/model': 1.2046912389341742, 'learning_rate/qwen_vl_interface': 6.309383970148922e-06, 'learning_rate/action_model': 6.309383970148922e-05, 'epoch': 0.9}) +[2026-06-17 11:55:34,592][starVLA.training.train_starvla][INFO] - Step 3057, Loss: {'train/loss': 0.843917965888977, 'train/grad_norm_pre_clip': 13.630043029785156, 'timing/data': 0.0003715190105140209, 'timing/model': 1.3377266860334203, 'learning_rate/qwen_vl_interface': 6.3072756436887805e-06, 'learning_rate/action_model': 6.30727564368878e-05, 'epoch': 0.9}) +[2026-06-17 11:55:39,257][starVLA.training.train_starvla][INFO] - Step 3058, Loss: {'train/loss': 1.015787124633789, 'train/grad_norm_pre_clip': 12.233081817626953, 'timing/data': 0.00038309500087052584, 'timing/model': 1.2056013320107013, 'learning_rate/qwen_vl_interface': 6.30516709805426e-06, 'learning_rate/action_model': 6.30516709805426e-05, 'epoch': 0.9}) +[2026-06-17 11:55:44,050][starVLA.training.train_starvla][INFO] - Step 3059, Loss: {'train/loss': 0.6603624820709229, 'train/grad_norm_pre_clip': 9.724824905395508, 'timing/data': 0.0003469840157777071, 'timing/model': 1.2053897820878774, 'learning_rate/qwen_vl_interface': 6.303058333682464e-06, 'learning_rate/action_model': 6.303058333682463e-05, 'epoch': 0.9}) +[2026-06-17 11:55:48,842][starVLA.training.train_starvla][INFO] - Step 3060, Loss: {'train/loss': 0.5524576306343079, 'train/grad_norm_pre_clip': 9.1708984375, 'timing/data': 0.00041104701813310385, 'timing/model': 1.2054027230478823, 'learning_rate/qwen_vl_interface': 6.300949351010542e-06, 'learning_rate/action_model': 6.300949351010543e-05, 'epoch': 0.9}) +[2026-06-17 11:55:53,666][starVLA.training.train_starvla][INFO] - Step 3061, Loss: {'train/loss': 0.8617762923240662, 'train/grad_norm_pre_clip': 10.785895347595215, 'timing/data': 0.0006019639549776912, 'timing/model': 1.2212290239986032, 'learning_rate/qwen_vl_interface': 6.298840150475689e-06, 'learning_rate/action_model': 6.298840150475689e-05, 'epoch': 0.9}) +[2026-06-17 11:55:58,424][starVLA.training.train_starvla][INFO] - Step 3062, Loss: {'train/loss': 0.6884461045265198, 'train/grad_norm_pre_clip': 9.382081031799316, 'timing/data': 0.00038139906246215105, 'timing/model': 1.2116297029424459, 'learning_rate/qwen_vl_interface': 6.296730732515145e-06, 'learning_rate/action_model': 6.296730732515145e-05, 'epoch': 0.9}) +[2026-06-17 11:56:03,211][starVLA.training.train_starvla][INFO] - Step 3063, Loss: {'train/loss': 1.1889352798461914, 'train/grad_norm_pre_clip': 13.76986026763916, 'timing/data': 0.00035105005372315645, 'timing/model': 1.3350197579711676, 'learning_rate/qwen_vl_interface': 6.294621097566193e-06, 'learning_rate/action_model': 6.294621097566193e-05, 'epoch': 0.9}) +[2026-06-17 11:56:07,870][starVLA.training.train_starvla][INFO] - Step 3064, Loss: {'train/loss': 0.4892829954624176, 'train/grad_norm_pre_clip': 9.519017219543457, 'timing/data': 0.00033088400959968567, 'timing/model': 1.2051720099989325, 'learning_rate/qwen_vl_interface': 6.292511246066163e-06, 'learning_rate/action_model': 6.292511246066163e-05, 'epoch': 0.9}) +[2026-06-17 11:56:12,658][starVLA.training.train_starvla][INFO] - Step 3065, Loss: {'train/loss': 0.659722089767456, 'train/grad_norm_pre_clip': 9.135876655578613, 'timing/data': 0.0004321970045566559, 'timing/model': 1.2068545459769666, 'learning_rate/qwen_vl_interface': 6.290401178452431e-06, 'learning_rate/action_model': 6.29040117845243e-05, 'epoch': 0.9}) +[2026-06-17 11:56:17,443][starVLA.training.train_starvla][INFO] - Step 3066, Loss: {'train/loss': 0.9314537644386292, 'train/grad_norm_pre_clip': 13.019824028015137, 'timing/data': 0.0003733220510184765, 'timing/model': 1.2044473030837253, 'learning_rate/qwen_vl_interface': 6.288290895162414e-06, 'learning_rate/action_model': 6.288290895162413e-05, 'epoch': 0.9}) +[2026-06-17 11:56:22,227][starVLA.training.train_starvla][INFO] - Step 3067, Loss: {'train/loss': 1.1145925521850586, 'train/grad_norm_pre_clip': 14.02945327758789, 'timing/data': 0.0005777280312031507, 'timing/model': 1.2086209920234978, 'learning_rate/qwen_vl_interface': 6.286180396633579e-06, 'learning_rate/action_model': 6.286180396633579e-05, 'epoch': 0.9}) +[2026-06-17 11:56:26,994][starVLA.training.train_starvla][INFO] - Step 3068, Loss: {'train/loss': 0.8003717660903931, 'train/grad_norm_pre_clip': 12.703006744384766, 'timing/data': 0.00036215700674802065, 'timing/model': 1.205883684917353, 'learning_rate/qwen_vl_interface': 6.28406968330343e-06, 'learning_rate/action_model': 6.28406968330343e-05, 'epoch': 0.9}) +[2026-06-17 11:56:31,826][starVLA.training.train_starvla][INFO] - Step 3069, Loss: {'train/loss': 0.7396670579910278, 'train/grad_norm_pre_clip': 12.748567581176758, 'timing/data': 0.0003551898989826441, 'timing/model': 1.3350588399916887, 'learning_rate/qwen_vl_interface': 6.281958755609526e-06, 'learning_rate/action_model': 6.281958755609526e-05, 'epoch': 0.9}) +[2026-06-17 11:56:36,476][starVLA.training.train_starvla][INFO] - Step 3070, Loss: {'train/loss': 0.9627415537834167, 'train/grad_norm_pre_clip': 11.368242263793945, 'timing/data': 0.000384898972697556, 'timing/model': 1.2045230040093884, 'learning_rate/qwen_vl_interface': 6.279847613989459e-06, 'learning_rate/action_model': 6.279847613989459e-05, 'epoch': 0.9}) +[2026-06-17 11:56:41,280][starVLA.training.train_starvla][INFO] - Step 3071, Loss: {'train/loss': 0.5245850682258606, 'train/grad_norm_pre_clip': 8.679390907287598, 'timing/data': 0.0003529679961502552, 'timing/model': 1.2082520769909024, 'learning_rate/qwen_vl_interface': 6.2777362588808745e-06, 'learning_rate/action_model': 6.277736258880874e-05, 'epoch': 0.9}) +[2026-06-17 11:56:46,088][starVLA.training.train_starvla][INFO] - Step 3072, Loss: {'train/loss': 0.7292951345443726, 'train/grad_norm_pre_clip': 9.863812446594238, 'timing/data': 0.00035188498441129923, 'timing/model': 1.204440897097811, 'learning_rate/qwen_vl_interface': 6.275624690721458e-06, 'learning_rate/action_model': 6.275624690721458e-05, 'epoch': 0.9}) +[2026-06-17 11:56:50,922][starVLA.training.train_starvla][INFO] - Step 3073, Loss: {'train/loss': 0.5615824460983276, 'train/grad_norm_pre_clip': 9.753045082092285, 'timing/data': 0.0007044919766485691, 'timing/model': 1.2032604690175503, 'learning_rate/qwen_vl_interface': 6.27351290994894e-06, 'learning_rate/action_model': 6.27351290994894e-05, 'epoch': 0.9}) +[2026-06-17 11:56:55,706][starVLA.training.train_starvla][INFO] - Step 3074, Loss: {'train/loss': 0.8010514378547668, 'train/grad_norm_pre_clip': 15.83329963684082, 'timing/data': 0.00036199402529746294, 'timing/model': 1.3307484660763294, 'learning_rate/qwen_vl_interface': 6.271400917001092e-06, 'learning_rate/action_model': 6.271400917001092e-05, 'epoch': 0.9}) +[2026-06-17 11:57:00,401][starVLA.training.train_starvla][INFO] - Step 3075, Loss: {'train/loss': 0.9569206237792969, 'train/grad_norm_pre_clip': 12.70350456237793, 'timing/data': 0.00039854703936725855, 'timing/model': 1.2443634499795735, 'learning_rate/qwen_vl_interface': 6.269288712315737e-06, 'learning_rate/action_model': 6.269288712315737e-05, 'epoch': 0.9}) +[2026-06-17 11:57:05,052][starVLA.training.train_starvla][INFO] - Step 3076, Loss: {'train/loss': 0.5740832686424255, 'train/grad_norm_pre_clip': 7.589285850524902, 'timing/data': 0.0003641920629888773, 'timing/model': 1.2055275429738685, 'learning_rate/qwen_vl_interface': 6.267176296330734e-06, 'learning_rate/action_model': 6.267176296330733e-05, 'epoch': 0.91}) +[2026-06-17 11:57:09,799][starVLA.training.train_starvla][INFO] - Step 3077, Loss: {'train/loss': 0.43447455763816833, 'train/grad_norm_pre_clip': 7.030217170715332, 'timing/data': 0.0003590020351111889, 'timing/model': 1.2071563230128959, 'learning_rate/qwen_vl_interface': 6.265063669483992e-06, 'learning_rate/action_model': 6.265063669483991e-05, 'epoch': 0.91}) +[2026-06-17 11:57:14,590][starVLA.training.train_starvla][INFO] - Step 3078, Loss: {'train/loss': 0.8777758479118347, 'train/grad_norm_pre_clip': 15.688898086547852, 'timing/data': 0.0003526960499584675, 'timing/model': 1.2036711659748107, 'learning_rate/qwen_vl_interface': 6.262950832213457e-06, 'learning_rate/action_model': 6.262950832213456e-05, 'epoch': 0.91}) +[2026-06-17 11:57:19,367][starVLA.training.train_starvla][INFO] - Step 3079, Loss: {'train/loss': 0.86246657371521, 'train/grad_norm_pre_clip': 9.42684555053711, 'timing/data': 0.0004035590682178736, 'timing/model': 1.2030690639512613, 'learning_rate/qwen_vl_interface': 6.260837784957125e-06, 'learning_rate/action_model': 6.260837784957126e-05, 'epoch': 0.91}) +[2026-06-17 11:57:24,148][starVLA.training.train_starvla][INFO] - Step 3080, Loss: {'train/loss': 0.5822796821594238, 'train/grad_norm_pre_clip': 9.158268928527832, 'timing/data': 0.00035002699587494135, 'timing/model': 1.3342150480020791, 'learning_rate/qwen_vl_interface': 6.258724528153034e-06, 'learning_rate/action_model': 6.258724528153033e-05, 'epoch': 0.91}) +[2026-06-17 11:57:28,805][starVLA.training.train_starvla][INFO] - Step 3081, Loss: {'train/loss': 0.5994176268577576, 'train/grad_norm_pre_clip': 8.974831581115723, 'timing/data': 0.00040079199243336916, 'timing/model': 1.2058310360880569, 'learning_rate/qwen_vl_interface': 6.256611062239261e-06, 'learning_rate/action_model': 6.25661106223926e-05, 'epoch': 0.91}) +[2026-06-17 11:57:33,593][starVLA.training.train_starvla][INFO] - Step 3082, Loss: {'train/loss': 0.8441003561019897, 'train/grad_norm_pre_clip': 10.006375312805176, 'timing/data': 0.0003750050673261285, 'timing/model': 1.206921446020715, 'learning_rate/qwen_vl_interface': 6.254497387653933e-06, 'learning_rate/action_model': 6.254497387653932e-05, 'epoch': 0.91}) +[2026-06-17 11:57:38,377][starVLA.training.train_starvla][INFO] - Step 3083, Loss: {'train/loss': 0.8135212659835815, 'train/grad_norm_pre_clip': 10.084071159362793, 'timing/data': 0.0003730760654434562, 'timing/model': 1.2099189769942313, 'learning_rate/qwen_vl_interface': 6.252383504835213e-06, 'learning_rate/action_model': 6.252383504835213e-05, 'epoch': 0.91}) +[2026-06-17 11:57:43,164][starVLA.training.train_starvla][INFO] - Step 3084, Loss: {'train/loss': 0.8760932087898254, 'train/grad_norm_pre_clip': 9.487791061401367, 'timing/data': 0.0003682180540636182, 'timing/model': 1.2051111599430442, 'learning_rate/qwen_vl_interface': 6.250269414221317e-06, 'learning_rate/action_model': 6.250269414221317e-05, 'epoch': 0.91}) +[2026-06-17 11:57:47,974][starVLA.training.train_starvla][INFO] - Step 3085, Loss: {'train/loss': 0.8470481634140015, 'train/grad_norm_pre_clip': 11.843205451965332, 'timing/data': 0.0004028070252388716, 'timing/model': 1.2067839240189642, 'learning_rate/qwen_vl_interface': 6.248155116250494e-06, 'learning_rate/action_model': 6.248155116250493e-05, 'epoch': 0.91}) +[2026-06-17 11:57:52,788][starVLA.training.train_starvla][INFO] - Step 3086, Loss: {'train/loss': 0.8615174293518066, 'train/grad_norm_pre_clip': 12.540860176086426, 'timing/data': 0.0006028279894962907, 'timing/model': 1.3473530309274793, 'learning_rate/qwen_vl_interface': 6.246040611361042e-06, 'learning_rate/action_model': 6.246040611361041e-05, 'epoch': 0.91}) +[2026-06-17 11:57:57,449][starVLA.training.train_starvla][INFO] - Step 3087, Loss: {'train/loss': 0.9688998460769653, 'train/grad_norm_pre_clip': 13.87814712524414, 'timing/data': 0.0003614820307120681, 'timing/model': 1.2059794070664793, 'learning_rate/qwen_vl_interface': 6.2439258999912976e-06, 'learning_rate/action_model': 6.243925899991298e-05, 'epoch': 0.91}) +[2026-06-17 11:58:02,231][starVLA.training.train_starvla][INFO] - Step 3088, Loss: {'train/loss': 0.5703029036521912, 'train/grad_norm_pre_clip': 8.441774368286133, 'timing/data': 0.0003571279812604189, 'timing/model': 1.2048969049938023, 'learning_rate/qwen_vl_interface': 6.241810982579646e-06, 'learning_rate/action_model': 6.241810982579646e-05, 'epoch': 0.91}) +[2026-06-17 11:58:07,028][starVLA.training.train_starvla][INFO] - Step 3089, Loss: {'train/loss': 0.5272689461708069, 'train/grad_norm_pre_clip': 9.06291389465332, 'timing/data': 0.00037702592089772224, 'timing/model': 1.204830392030999, 'learning_rate/qwen_vl_interface': 6.239695859564511e-06, 'learning_rate/action_model': 6.23969585956451e-05, 'epoch': 0.91}) +[2026-06-17 11:58:11,811][starVLA.training.train_starvla][INFO] - Step 3090, Loss: {'train/loss': 0.5127602219581604, 'train/grad_norm_pre_clip': 11.434392929077148, 'timing/data': 0.0003469109069555998, 'timing/model': 1.20597736898344, 'learning_rate/qwen_vl_interface': 6.2375805313843564e-06, 'learning_rate/action_model': 6.237580531384356e-05, 'epoch': 0.91}) +[2026-06-17 11:58:16,544][starVLA.training.train_starvla][INFO] - Step 3091, Loss: {'train/loss': 0.9195387363433838, 'train/grad_norm_pre_clip': 12.92846393585205, 'timing/data': 0.0003550619585439563, 'timing/model': 1.2052531789522618, 'learning_rate/qwen_vl_interface': 6.2354649984777e-06, 'learning_rate/action_model': 6.235464998477699e-05, 'epoch': 0.91}) +[2026-06-17 11:58:21,339][starVLA.training.train_starvla][INFO] - Step 3092, Loss: {'train/loss': 0.7644298672676086, 'train/grad_norm_pre_clip': 11.170796394348145, 'timing/data': 0.0003541669575497508, 'timing/model': 1.3356346869841218, 'learning_rate/qwen_vl_interface': 6.2333492612830835e-06, 'learning_rate/action_model': 6.233349261283083e-05, 'epoch': 0.91}) +[2026-06-17 11:58:25,999][starVLA.training.train_starvla][INFO] - Step 3093, Loss: {'train/loss': 0.6635094285011292, 'train/grad_norm_pre_clip': 10.638521194458008, 'timing/data': 0.0003445440670475364, 'timing/model': 1.2053255030186847, 'learning_rate/qwen_vl_interface': 6.231233320239112e-06, 'learning_rate/action_model': 6.231233320239112e-05, 'epoch': 0.91}) +[2026-06-17 11:58:30,817][starVLA.training.train_starvla][INFO] - Step 3094, Loss: {'train/loss': 0.6232781410217285, 'train/grad_norm_pre_clip': 9.350231170654297, 'timing/data': 0.00037280202377587557, 'timing/model': 1.2038934140000492, 'learning_rate/qwen_vl_interface': 6.229117175784414e-06, 'learning_rate/action_model': 6.229117175784413e-05, 'epoch': 0.91}) +[2026-06-17 11:58:35,617][starVLA.training.train_starvla][INFO] - Step 3095, Loss: {'train/loss': 0.9360255002975464, 'train/grad_norm_pre_clip': 12.485391616821289, 'timing/data': 0.0003313430352136493, 'timing/model': 1.2210926200496033, 'learning_rate/qwen_vl_interface': 6.227000828357672e-06, 'learning_rate/action_model': 6.227000828357672e-05, 'epoch': 0.91}) +[2026-06-17 11:58:40,398][starVLA.training.train_starvla][INFO] - Step 3096, Loss: {'train/loss': 0.736566424369812, 'train/grad_norm_pre_clip': 13.727036476135254, 'timing/data': 0.0003644530661404133, 'timing/model': 1.2058109870413318, 'learning_rate/qwen_vl_interface': 6.2248842783976085e-06, 'learning_rate/action_model': 6.224884278397608e-05, 'epoch': 0.91}) +[2026-06-17 11:58:45,177][starVLA.training.train_starvla][INFO] - Step 3097, Loss: {'train/loss': 0.7115879058837891, 'train/grad_norm_pre_clip': 12.217885971069336, 'timing/data': 0.0004061800427734852, 'timing/model': 1.2078381449682638, 'learning_rate/qwen_vl_interface': 6.222767526342983e-06, 'learning_rate/action_model': 6.222767526342983e-05, 'epoch': 0.91}) +[2026-06-17 11:58:49,957][starVLA.training.train_starvla][INFO] - Step 3098, Loss: {'train/loss': 0.9131996035575867, 'train/grad_norm_pre_clip': 13.304926872253418, 'timing/data': 0.0003547470550984144, 'timing/model': 1.3318169560516253, 'learning_rate/qwen_vl_interface': 6.220650572632605e-06, 'learning_rate/action_model': 6.220650572632604e-05, 'epoch': 0.91}) +[2026-06-17 11:58:54,619][starVLA.training.train_starvla][INFO] - Step 3099, Loss: {'train/loss': 0.5993614196777344, 'train/grad_norm_pre_clip': 11.213349342346191, 'timing/data': 0.00035050802398473024, 'timing/model': 1.2046052439836785, 'learning_rate/qwen_vl_interface': 6.218533417705318e-06, 'learning_rate/action_model': 6.218533417705318e-05, 'epoch': 0.91}) +[2026-06-17 12:00:36,607][starVLA.training.train_starvla][INFO] - Step 3100, Loss: {'train/loss': 0.5916202068328857, 'train/grad_norm_pre_clip': 10.990775108337402, 'eval/loss': 0.7582953572273254, 'eval/latency_0/loss': 0.7582953572273254, 'eval/demon_fire_precision': 0.9024390243902439, 'eval/demon_fire_recall': 0.9219269102990033, 'eval/demon_fire_f1': 0.9120788824979458, 'eval/demon_fire_rate_model': 0.6840934371523916, 'eval/demon_fire_rate_teacher': 0.6696329254727474, 'eval/demon_left_precision': 0.7952069716775599, 'eval/demon_left_recall': 0.8271954674220963, 'eval/demon_left_f1': 0.8108858650374895, 'eval/demon_left_rate_model': 0.5105672969966629, 'eval/demon_left_rate_teacher': 0.49082313681868744, 'eval/demon_right_precision': 0.8222857142857143, 'eval/demon_right_recall': 0.8129943502824859, 'eval/demon_right_f1': 0.8176136363636364, 'eval/demon_right_rate_model': 0.4866518353726363, 'eval/demon_right_rate_teacher': 0.4922135706340378, 'eval/demon_move_f1': 0.8142497507005629, 'eval/demon_cc_f1': 0.8803477143280716, 'eval/latency_0/demon_fire_precision': 0.9024390243902439, 'eval/latency_0/demon_fire_recall': 0.9219269102990033, 'eval/latency_0/demon_fire_f1': 0.9120788824979458, 'eval/latency_0/demon_fire_rate_model': 0.6840934371523916, 'eval/latency_0/demon_fire_rate_teacher': 0.6696329254727474, 'eval/latency_0/demon_left_precision': 0.7952069716775599, 'eval/latency_0/demon_left_recall': 0.8271954674220963, 'eval/latency_0/demon_left_f1': 0.8108858650374895, 'eval/latency_0/demon_left_rate_model': 0.5105672969966629, 'eval/latency_0/demon_left_rate_teacher': 0.49082313681868744, 'eval/latency_0/demon_right_precision': 0.8222857142857143, 'eval/latency_0/demon_right_recall': 0.8129943502824859, 'eval/latency_0/demon_right_f1': 0.8176136363636364, 'eval/latency_0/demon_right_rate_model': 0.4866518353726363, 'eval/latency_0/demon_right_rate_teacher': 0.4922135706340378, 'eval/latency_0/demon_move_f1': 0.8142497507005629, 'eval/latency_0/demon_cc_f1': 0.8803477143280716, 'timing/data': 0.0003480209270492196, 'timing/model': 1.2090161269297823, 'learning_rate/qwen_vl_interface': 6.216416062000012e-06, 'learning_rate/action_model': 6.216416062000013e-05, 'epoch': 0.91}) +[2026-06-17 12:00:36,607][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_3100_state +[2026-06-17 12:00:36,610][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 12:00:40,174][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_3100_state/model.safetensors +[2026-06-17 12:00:47,931][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_3100_state/optimizer.bin +[2026-06-17 12:00:47,932][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_3100_state/sampler.bin +[2026-06-17 12:00:47,932][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_3100_state/sampler_1.bin +[2026-06-17 12:00:47,933][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_3100_state/random_states_0.pkl +[2026-06-17 12:00:47,981][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3100 +[2026-06-17 12:00:47,981][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 12:00:47,991][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 12:00:52,698][starVLA.training.train_starvla][INFO] - Step 3101, Loss: {'train/loss': 0.7809706926345825, 'train/grad_norm_pre_clip': 11.628762245178223, 'timing/data': 0.0003798330435529351, 'timing/model': 1.1859072669176385, 'learning_rate/qwen_vl_interface': 6.214298505955615e-06, 'learning_rate/action_model': 6.214298505955614e-05, 'epoch': 0.91}) +[2026-06-17 12:00:57,347][starVLA.training.train_starvla][INFO] - Step 3102, Loss: {'train/loss': 0.8340325355529785, 'train/grad_norm_pre_clip': 10.235621452331543, 'timing/data': 0.000427871011197567, 'timing/model': 1.1869930579559878, 'learning_rate/qwen_vl_interface': 6.2121807500111e-06, 'learning_rate/action_model': 6.212180750011099e-05, 'epoch': 0.91}) +[2026-06-17 12:01:02,044][starVLA.training.train_starvla][INFO] - Step 3103, Loss: {'train/loss': 0.945860743522644, 'train/grad_norm_pre_clip': 14.38088321685791, 'timing/data': 0.0004150370368734002, 'timing/model': 1.197712210007012, 'learning_rate/qwen_vl_interface': 6.210062794605482e-06, 'learning_rate/action_model': 6.21006279460548e-05, 'epoch': 0.91}) +[2026-06-17 12:01:06,773][starVLA.training.train_starvla][INFO] - Step 3104, Loss: {'train/loss': 0.47752824425697327, 'train/grad_norm_pre_clip': 9.886127471923828, 'timing/data': 0.00037143309600651264, 'timing/model': 1.1947328890673816, 'learning_rate/qwen_vl_interface': 6.207944640177812e-06, 'learning_rate/action_model': 6.207944640177812e-05, 'epoch': 0.91}) +[2026-06-17 12:01:11,728][starVLA.training.train_starvla][INFO] - Step 3105, Loss: {'train/loss': 0.5660233497619629, 'train/grad_norm_pre_clip': 10.29547119140625, 'timing/data': 0.0003701000241562724, 'timing/model': 1.3344797119498253, 'learning_rate/qwen_vl_interface': 6.205826287167188e-06, 'learning_rate/action_model': 6.205826287167188e-05, 'epoch': 0.91}) +[2026-06-17 12:01:16,358][starVLA.training.train_starvla][INFO] - Step 3106, Loss: {'train/loss': 0.7425097227096558, 'train/grad_norm_pre_clip': 40.14935302734375, 'timing/data': 0.0003726840950548649, 'timing/model': 1.195181813905947, 'learning_rate/qwen_vl_interface': 6.203707736012745e-06, 'learning_rate/action_model': 6.203707736012745e-05, 'epoch': 0.91}) +[2026-06-17 12:01:21,117][starVLA.training.train_starvla][INFO] - Step 3107, Loss: {'train/loss': 0.6539528965950012, 'train/grad_norm_pre_clip': 12.025517463684082, 'timing/data': 0.00032973301131278276, 'timing/model': 1.1957160729216412, 'learning_rate/qwen_vl_interface': 6.201588987153665e-06, 'learning_rate/action_model': 6.201588987153664e-05, 'epoch': 0.91}) +[2026-06-17 12:01:25,880][starVLA.training.train_starvla][INFO] - Step 3108, Loss: {'train/loss': 0.521851658821106, 'train/grad_norm_pre_clip': 9.314773559570312, 'timing/data': 0.0004157830262556672, 'timing/model': 1.197400143952109, 'learning_rate/qwen_vl_interface': 6.19947004102916e-06, 'learning_rate/action_model': 6.199470041029159e-05, 'epoch': 0.91}) +[2026-06-17 12:01:30,653][starVLA.training.train_starvla][INFO] - Step 3109, Loss: {'train/loss': 1.0501958131790161, 'train/grad_norm_pre_clip': 14.512040138244629, 'timing/data': 0.0004001080524176359, 'timing/model': 1.2035675530787557, 'learning_rate/qwen_vl_interface': 6.197350898078496e-06, 'learning_rate/action_model': 6.197350898078495e-05, 'epoch': 0.91}) +[2026-06-17 12:01:35,413][starVLA.training.train_starvla][INFO] - Step 3110, Loss: {'train/loss': 0.7471771836280823, 'train/grad_norm_pre_clip': 10.943982124328613, 'timing/data': 0.0003576249582692981, 'timing/model': 1.1983873770805076, 'learning_rate/qwen_vl_interface': 6.195231558740972e-06, 'learning_rate/action_model': 6.195231558740971e-05, 'epoch': 0.92}) +[2026-06-17 12:01:40,115][starVLA.training.train_starvla][INFO] - Step 3111, Loss: {'train/loss': 0.7534542083740234, 'train/grad_norm_pre_clip': 20.9750919342041, 'timing/data': 0.00040163390804082155, 'timing/model': 1.2578571980120614, 'learning_rate/qwen_vl_interface': 6.1931120234559275e-06, 'learning_rate/action_model': 6.193112023455927e-05, 'epoch': 0.92}) +[2026-06-17 12:01:44,758][starVLA.training.train_starvla][INFO] - Step 3112, Loss: {'train/loss': 0.7534592151641846, 'train/grad_norm_pre_clip': 9.46297550201416, 'timing/data': 0.00037697702646255493, 'timing/model': 1.2034226319519803, 'learning_rate/qwen_vl_interface': 6.190992292662747e-06, 'learning_rate/action_model': 6.190992292662746e-05, 'epoch': 0.92}) +[2026-06-17 12:01:49,515][starVLA.training.train_starvla][INFO] - Step 3113, Loss: {'train/loss': 0.620333194732666, 'train/grad_norm_pre_clip': 9.651676177978516, 'timing/data': 0.0003438210114836693, 'timing/model': 1.1998843280598521, 'learning_rate/qwen_vl_interface': 6.18887236680085e-06, 'learning_rate/action_model': 6.18887236680085e-05, 'epoch': 0.92}) +[2026-06-17 12:01:54,291][starVLA.training.train_starvla][INFO] - Step 3114, Loss: {'train/loss': 0.668982982635498, 'train/grad_norm_pre_clip': 11.527153968811035, 'timing/data': 0.0003685340052470565, 'timing/model': 1.2017389609245583, 'learning_rate/qwen_vl_interface': 6.186752246309702e-06, 'learning_rate/action_model': 6.186752246309702e-05, 'epoch': 0.92}) +[2026-06-17 12:01:59,097][starVLA.training.train_starvla][INFO] - Step 3115, Loss: {'train/loss': 0.9424149990081787, 'train/grad_norm_pre_clip': 15.370041847229004, 'timing/data': 0.00035774707794189453, 'timing/model': 1.2029905800009146, 'learning_rate/qwen_vl_interface': 6.184631931628806e-06, 'learning_rate/action_model': 6.184631931628806e-05, 'epoch': 0.92}) +[2026-06-17 12:02:03,875][starVLA.training.train_starvla][INFO] - Step 3116, Loss: {'train/loss': 1.2069429159164429, 'train/grad_norm_pre_clip': 13.293846130371094, 'timing/data': 0.00037380808498710394, 'timing/model': 1.3324208909180015, 'learning_rate/qwen_vl_interface': 6.1825114231977036e-06, 'learning_rate/action_model': 6.182511423197704e-05, 'epoch': 0.92}) +[2026-06-17 12:02:08,528][starVLA.training.train_starvla][INFO] - Step 3117, Loss: {'train/loss': 0.5965265035629272, 'train/grad_norm_pre_clip': 9.864537239074707, 'timing/data': 0.00037413404788821936, 'timing/model': 1.2046677940525115, 'learning_rate/qwen_vl_interface': 6.18039072145598e-06, 'learning_rate/action_model': 6.18039072145598e-05, 'epoch': 0.92}) +[2026-06-17 12:02:13,306][starVLA.training.train_starvla][INFO] - Step 3118, Loss: {'train/loss': 0.40998780727386475, 'train/grad_norm_pre_clip': 9.61684799194336, 'timing/data': 0.00035833101719617844, 'timing/model': 1.202136019943282, 'learning_rate/qwen_vl_interface': 6.17826982684326e-06, 'learning_rate/action_model': 6.178269826843259e-05, 'epoch': 0.92}) +[2026-06-17 12:02:17,966][starVLA.training.train_starvla][INFO] - Step 3119, Loss: {'train/loss': 0.7894642353057861, 'train/grad_norm_pre_clip': 11.94444751739502, 'timing/data': 0.0003469879738986492, 'timing/model': 1.2035361139569432, 'learning_rate/qwen_vl_interface': 6.176148739799205e-06, 'learning_rate/action_model': 6.176148739799205e-05, 'epoch': 0.92}) +[2026-06-17 12:02:22,633][starVLA.training.train_starvla][INFO] - Step 3120, Loss: {'train/loss': 0.71941077709198, 'train/grad_norm_pre_clip': 10.746976852416992, 'timing/data': 0.00043179094791412354, 'timing/model': 1.203539984067902, 'learning_rate/qwen_vl_interface': 6.1740274607635195e-06, 'learning_rate/action_model': 6.174027460763519e-05, 'epoch': 0.92}) +[2026-06-17 12:02:27,380][starVLA.training.train_starvla][INFO] - Step 3121, Loss: {'train/loss': 0.8929076790809631, 'train/grad_norm_pre_clip': 11.503432273864746, 'timing/data': 0.0003387050237506628, 'timing/model': 1.2028966710204259, 'learning_rate/qwen_vl_interface': 6.1719059901759474e-06, 'learning_rate/action_model': 6.171905990175947e-05, 'epoch': 0.92}) +[2026-06-17 12:02:32,120][starVLA.training.train_starvla][INFO] - Step 3122, Loss: {'train/loss': 0.845000147819519, 'train/grad_norm_pre_clip': 14.021409034729004, 'timing/data': 0.00034761393908411264, 'timing/model': 1.2969480950850993, 'learning_rate/qwen_vl_interface': 6.169784328476272e-06, 'learning_rate/action_model': 6.169784328476271e-05, 'epoch': 0.92}) +[2026-06-17 12:02:36,782][starVLA.training.train_starvla][INFO] - Step 3123, Loss: {'train/loss': 0.4880572259426117, 'train/grad_norm_pre_clip': 10.581221580505371, 'timing/data': 0.0006256790366023779, 'timing/model': 1.2021195159759372, 'learning_rate/qwen_vl_interface': 6.167662476104314e-06, 'learning_rate/action_model': 6.167662476104314e-05, 'epoch': 0.92}) +[2026-06-17 12:02:41,559][starVLA.training.train_starvla][INFO] - Step 3124, Loss: {'train/loss': 0.7704668641090393, 'train/grad_norm_pre_clip': 13.79797077178955, 'timing/data': 0.00036085094325244427, 'timing/model': 1.2041105990065262, 'learning_rate/qwen_vl_interface': 6.165540433499939e-06, 'learning_rate/action_model': 6.165540433499939e-05, 'epoch': 0.92}) +[2026-06-17 12:02:46,343][starVLA.training.train_starvla][INFO] - Step 3125, Loss: {'train/loss': 0.7670824527740479, 'train/grad_norm_pre_clip': 10.857970237731934, 'timing/data': 0.00035719003062695265, 'timing/model': 1.2041164129041135, 'learning_rate/qwen_vl_interface': 6.163418201103045e-06, 'learning_rate/action_model': 6.163418201103045e-05, 'epoch': 0.92}) +[2026-06-17 12:02:51,121][starVLA.training.train_starvla][INFO] - Step 3126, Loss: {'train/loss': 0.5480656623840332, 'train/grad_norm_pre_clip': 11.818965911865234, 'timing/data': 0.0003340189578011632, 'timing/model': 1.203592812991701, 'learning_rate/qwen_vl_interface': 6.161295779353578e-06, 'learning_rate/action_model': 6.161295779353577e-05, 'epoch': 0.92}) +[2026-06-17 12:02:55,902][starVLA.training.train_starvla][INFO] - Step 3127, Loss: {'train/loss': 0.6986919045448303, 'train/grad_norm_pre_clip': 11.894253730773926, 'timing/data': 0.000334752956405282, 'timing/model': 1.3332398679340258, 'learning_rate/qwen_vl_interface': 6.159173168691512e-06, 'learning_rate/action_model': 6.159173168691512e-05, 'epoch': 0.92}) +[2026-06-17 12:03:00,549][starVLA.training.train_starvla][INFO] - Step 3128, Loss: {'train/loss': 0.8953772187232971, 'train/grad_norm_pre_clip': 11.302083015441895, 'timing/data': 0.000350742950104177, 'timing/model': 1.2060475220205262, 'learning_rate/qwen_vl_interface': 6.157050369556871e-06, 'learning_rate/action_model': 6.157050369556872e-05, 'epoch': 0.92}) +[2026-06-17 12:03:05,326][starVLA.training.train_starvla][INFO] - Step 3129, Loss: {'train/loss': 0.8346879482269287, 'train/grad_norm_pre_clip': 14.17728328704834, 'timing/data': 0.0003649079008027911, 'timing/model': 1.2020846740342677, 'learning_rate/qwen_vl_interface': 6.154927382389714e-06, 'learning_rate/action_model': 6.154927382389714e-05, 'epoch': 0.92}) +[2026-06-17 12:03:10,089][starVLA.training.train_starvla][INFO] - Step 3130, Loss: {'train/loss': 0.571431577205658, 'train/grad_norm_pre_clip': 8.913413047790527, 'timing/data': 0.0003531889524310827, 'timing/model': 1.2042444329708815, 'learning_rate/qwen_vl_interface': 6.152804207630135e-06, 'learning_rate/action_model': 6.152804207630136e-05, 'epoch': 0.92}) +[2026-06-17 12:03:14,874][starVLA.training.train_starvla][INFO] - Step 3131, Loss: {'train/loss': 0.8962971568107605, 'train/grad_norm_pre_clip': 13.242035865783691, 'timing/data': 0.00035113596823066473, 'timing/model': 1.2034959560260177, 'learning_rate/qwen_vl_interface': 6.150680845718274e-06, 'learning_rate/action_model': 6.150680845718274e-05, 'epoch': 0.92}) +[2026-06-17 12:03:19,657][starVLA.training.train_starvla][INFO] - Step 3132, Loss: {'train/loss': 0.8126387000083923, 'train/grad_norm_pre_clip': 10.56933307647705, 'timing/data': 0.0003686649724841118, 'timing/model': 1.2014757879078388, 'learning_rate/qwen_vl_interface': 6.148557297094303e-06, 'learning_rate/action_model': 6.148557297094303e-05, 'epoch': 0.92}) +[2026-06-17 12:03:24,473][starVLA.training.train_starvla][INFO] - Step 3133, Loss: {'train/loss': 0.814339816570282, 'train/grad_norm_pre_clip': 19.174808502197266, 'timing/data': 0.0005369869759306312, 'timing/model': 1.3317451119655743, 'learning_rate/qwen_vl_interface': 6.1464335621984375e-06, 'learning_rate/action_model': 6.146433562198437e-05, 'epoch': 0.92}) +[2026-06-17 12:03:29,164][starVLA.training.train_starvla][INFO] - Step 3134, Loss: {'train/loss': 0.5923494696617126, 'train/grad_norm_pre_clip': 8.976207733154297, 'timing/data': 0.0004036050522699952, 'timing/model': 1.2033393330639228, 'learning_rate/qwen_vl_interface': 6.14430964147093e-06, 'learning_rate/action_model': 6.14430964147093e-05, 'epoch': 0.92}) +[2026-06-17 12:03:33,943][starVLA.training.train_starvla][INFO] - Step 3135, Loss: {'train/loss': 0.8299385905265808, 'train/grad_norm_pre_clip': 12.437858581542969, 'timing/data': 0.0005361509975045919, 'timing/model': 1.203575863968581, 'learning_rate/qwen_vl_interface': 6.1421855353520735e-06, 'learning_rate/action_model': 6.142185535352073e-05, 'epoch': 0.92}) +[2026-06-17 12:03:38,654][starVLA.training.train_starvla][INFO] - Step 3136, Loss: {'train/loss': 0.8175735473632812, 'train/grad_norm_pre_clip': 10.68605899810791, 'timing/data': 0.0003728739684447646, 'timing/model': 1.2059588830452412, 'learning_rate/qwen_vl_interface': 6.140061244282194e-06, 'learning_rate/action_model': 6.140061244282194e-05, 'epoch': 0.92}) +[2026-06-17 12:03:43,394][starVLA.training.train_starvla][INFO] - Step 3137, Loss: {'train/loss': 1.00423002243042, 'train/grad_norm_pre_clip': 14.691548347473145, 'timing/data': 0.00036837300285696983, 'timing/model': 1.2045559959951788, 'learning_rate/qwen_vl_interface': 6.137936768701659e-06, 'learning_rate/action_model': 6.137936768701659e-05, 'epoch': 0.92}) +[2026-06-17 12:03:48,177][starVLA.training.train_starvla][INFO] - Step 3138, Loss: {'train/loss': 0.8012961149215698, 'train/grad_norm_pre_clip': 18.417465209960938, 'timing/data': 0.0003581700148060918, 'timing/model': 1.2076364499516785, 'learning_rate/qwen_vl_interface': 6.135812109050879e-06, 'learning_rate/action_model': 6.135812109050879e-05, 'epoch': 0.92}) +[2026-06-17 12:03:53,008][starVLA.training.train_starvla][INFO] - Step 3139, Loss: {'train/loss': 0.8211671113967896, 'train/grad_norm_pre_clip': 10.76509952545166, 'timing/data': 0.00045105803292244673, 'timing/model': 1.337309738039039, 'learning_rate/qwen_vl_interface': 6.1336872657702936e-06, 'learning_rate/action_model': 6.133687265770293e-05, 'epoch': 0.92}) +[2026-06-17 12:03:57,659][starVLA.training.train_starvla][INFO] - Step 3140, Loss: {'train/loss': 0.5593082308769226, 'train/grad_norm_pre_clip': 10.717625617980957, 'timing/data': 0.0004173470661044121, 'timing/model': 1.2032222039997578, 'learning_rate/qwen_vl_interface': 6.131562239300389e-06, 'learning_rate/action_model': 6.13156223930039e-05, 'epoch': 0.92}) +[2026-06-17 12:04:02,438][starVLA.training.train_starvla][INFO] - Step 3141, Loss: {'train/loss': 0.6080349683761597, 'train/grad_norm_pre_clip': 10.862815856933594, 'timing/data': 0.00035037193447351456, 'timing/model': 1.2035865960642695, 'learning_rate/qwen_vl_interface': 6.129437030081683e-06, 'learning_rate/action_model': 6.129437030081682e-05, 'epoch': 0.92}) +[2026-06-17 12:04:07,247][starVLA.training.train_starvla][INFO] - Step 3142, Loss: {'train/loss': 0.8098147511482239, 'train/grad_norm_pre_clip': 13.559771537780762, 'timing/data': 0.0005374769680202007, 'timing/model': 1.213180165970698, 'learning_rate/qwen_vl_interface': 6.1273116385547336e-06, 'learning_rate/action_model': 6.127311638554734e-05, 'epoch': 0.92}) +[2026-06-17 12:04:12,064][starVLA.training.train_starvla][INFO] - Step 3143, Loss: {'train/loss': 0.7652331590652466, 'train/grad_norm_pre_clip': 14.80460262298584, 'timing/data': 0.0003441799199208617, 'timing/model': 1.2026942650554702, 'learning_rate/qwen_vl_interface': 6.125186065160138e-06, 'learning_rate/action_model': 6.125186065160138e-05, 'epoch': 0.92}) +[2026-06-17 12:04:16,852][starVLA.training.train_starvla][INFO] - Step 3144, Loss: {'train/loss': 0.5154118537902832, 'train/grad_norm_pre_clip': 13.933228492736816, 'timing/data': 0.00034193298779428005, 'timing/model': 1.2210447189863771, 'learning_rate/qwen_vl_interface': 6.123060310338529e-06, 'learning_rate/action_model': 6.123060310338529e-05, 'epoch': 0.93}) +[2026-06-17 12:04:21,593][starVLA.training.train_starvla][INFO] - Step 3145, Loss: {'train/loss': 0.7675975561141968, 'train/grad_norm_pre_clip': 10.489214897155762, 'timing/data': 0.00033817102666944265, 'timing/model': 1.258368998998776, 'learning_rate/qwen_vl_interface': 6.120934374530579e-06, 'learning_rate/action_model': 6.12093437453058e-05, 'epoch': 0.93}) +[2026-06-17 12:04:26,263][starVLA.training.train_starvla][INFO] - Step 3146, Loss: {'train/loss': 0.6655228734016418, 'train/grad_norm_pre_clip': 10.152100563049316, 'timing/data': 0.0006444680038839579, 'timing/model': 1.209650991950184, 'learning_rate/qwen_vl_interface': 6.118808258176996e-06, 'learning_rate/action_model': 6.118808258176995e-05, 'epoch': 0.93}) +[2026-06-17 12:04:31,010][starVLA.training.train_starvla][INFO] - Step 3147, Loss: {'train/loss': 0.5305953025817871, 'train/grad_norm_pre_clip': 9.991768836975098, 'timing/data': 0.0005393499741330743, 'timing/model': 1.202558116056025, 'learning_rate/qwen_vl_interface': 6.116681961718525e-06, 'learning_rate/action_model': 6.116681961718524e-05, 'epoch': 0.93}) +[2026-06-17 12:04:35,808][starVLA.training.train_starvla][INFO] - Step 3148, Loss: {'train/loss': 0.7451058030128479, 'train/grad_norm_pre_clip': 12.96127700805664, 'timing/data': 0.00036139495205134153, 'timing/model': 1.2052258029580116, 'learning_rate/qwen_vl_interface': 6.114555485595952e-06, 'learning_rate/action_model': 6.114555485595952e-05, 'epoch': 0.93}) +[2026-06-17 12:04:40,646][starVLA.training.train_starvla][INFO] - Step 3149, Loss: {'train/loss': 1.0050392150878906, 'train/grad_norm_pre_clip': 15.464646339416504, 'timing/data': 0.0003793130163103342, 'timing/model': 1.2019997090101242, 'learning_rate/qwen_vl_interface': 6.112428830250096e-06, 'learning_rate/action_model': 6.112428830250095e-05, 'epoch': 0.93}) +[2026-06-17 12:04:45,437][starVLA.training.train_starvla][INFO] - Step 3150, Loss: {'train/loss': 0.8024230003356934, 'train/grad_norm_pre_clip': 12.646347045898438, 'timing/data': 0.00036877195816487074, 'timing/model': 1.3380364469485357, 'learning_rate/qwen_vl_interface': 6.110301996121816e-06, 'learning_rate/action_model': 6.110301996121816e-05, 'epoch': 0.93}) +[2026-06-17 12:04:50,106][starVLA.training.train_starvla][INFO] - Step 3151, Loss: {'train/loss': 0.45128485560417175, 'train/grad_norm_pre_clip': 7.7037272453308105, 'timing/data': 0.00035478000063449144, 'timing/model': 1.2130519050406292, 'learning_rate/qwen_vl_interface': 6.108174983652007e-06, 'learning_rate/action_model': 6.108174983652006e-05, 'epoch': 0.93}) +[2026-06-17 12:04:54,894][starVLA.training.train_starvla][INFO] - Step 3152, Loss: {'train/loss': 0.7910464406013489, 'train/grad_norm_pre_clip': 10.775385856628418, 'timing/data': 0.00035572994966059923, 'timing/model': 1.2126204860396683, 'learning_rate/qwen_vl_interface': 6.106047793281598e-06, 'learning_rate/action_model': 6.106047793281598e-05, 'epoch': 0.93}) +[2026-06-17 12:04:59,561][starVLA.training.train_starvla][INFO] - Step 3153, Loss: {'train/loss': 0.5439075827598572, 'train/grad_norm_pre_clip': 16.32944679260254, 'timing/data': 0.000667833024635911, 'timing/model': 1.2028088070219383, 'learning_rate/qwen_vl_interface': 6.1039204254515626e-06, 'learning_rate/action_model': 6.103920425451563e-05, 'epoch': 0.93}) +[2026-06-17 12:05:04,240][starVLA.training.train_starvla][INFO] - Step 3154, Loss: {'train/loss': 0.9149056673049927, 'train/grad_norm_pre_clip': 12.538135528564453, 'timing/data': 0.00033765099942684174, 'timing/model': 1.2083711620653048, 'learning_rate/qwen_vl_interface': 6.101792880602901e-06, 'learning_rate/action_model': 6.101792880602901e-05, 'epoch': 0.93}) +[2026-06-17 12:05:08,965][starVLA.training.train_starvla][INFO] - Step 3155, Loss: {'train/loss': 0.7281639575958252, 'train/grad_norm_pre_clip': 11.421602249145508, 'timing/data': 0.0004325109766796231, 'timing/model': 1.2164120450615883, 'learning_rate/qwen_vl_interface': 6.099665159176663e-06, 'learning_rate/action_model': 6.099665159176663e-05, 'epoch': 0.93}) +[2026-06-17 12:05:13,706][starVLA.training.train_starvla][INFO] - Step 3156, Loss: {'train/loss': 0.5688897371292114, 'train/grad_norm_pre_clip': 8.495198249816895, 'timing/data': 0.0006098219892010093, 'timing/model': 1.2839729180559516, 'learning_rate/qwen_vl_interface': 6.09753726161392e-06, 'learning_rate/action_model': 6.09753726161392e-05, 'epoch': 0.93}) +[2026-06-17 12:05:18,373][starVLA.training.train_starvla][INFO] - Step 3157, Loss: {'train/loss': 0.8571683168411255, 'train/grad_norm_pre_clip': 12.859606742858887, 'timing/data': 0.0003671830054372549, 'timing/model': 1.2080464359605685, 'learning_rate/qwen_vl_interface': 6.095409188355793e-06, 'learning_rate/action_model': 6.095409188355793e-05, 'epoch': 0.93}) +[2026-06-17 12:05:23,187][starVLA.training.train_starvla][INFO] - Step 3158, Loss: {'train/loss': 0.6662174463272095, 'train/grad_norm_pre_clip': 9.838541984558105, 'timing/data': 0.0003725260030478239, 'timing/model': 1.2078020960325375, 'learning_rate/qwen_vl_interface': 6.0932809398434315e-06, 'learning_rate/action_model': 6.093280939843431e-05, 'epoch': 0.93}) +[2026-06-17 12:05:28,029][starVLA.training.train_starvla][INFO] - Step 3159, Loss: {'train/loss': 0.7189417481422424, 'train/grad_norm_pre_clip': 11.21027660369873, 'timing/data': 0.0003592029679566622, 'timing/model': 1.205037118983455, 'learning_rate/qwen_vl_interface': 6.091152516518022e-06, 'learning_rate/action_model': 6.0911525165180216e-05, 'epoch': 0.93}) +[2026-06-17 12:05:32,842][starVLA.training.train_starvla][INFO] - Step 3160, Loss: {'train/loss': 0.6491231322288513, 'train/grad_norm_pre_clip': 11.371646881103516, 'timing/data': 0.00035250792279839516, 'timing/model': 1.2053443419281393, 'learning_rate/qwen_vl_interface': 6.089023918820792e-06, 'learning_rate/action_model': 6.089023918820792e-05, 'epoch': 0.93}) +[2026-06-17 12:05:37,654][starVLA.training.train_starvla][INFO] - Step 3161, Loss: {'train/loss': 0.9095641374588013, 'train/grad_norm_pre_clip': 13.581666946411133, 'timing/data': 0.000365910935215652, 'timing/model': 1.3387912890175357, 'learning_rate/qwen_vl_interface': 6.0868951471929986e-06, 'learning_rate/action_model': 6.086895147192998e-05, 'epoch': 0.93}) +[2026-06-17 12:05:42,333][starVLA.training.train_starvla][INFO] - Step 3162, Loss: {'train/loss': 0.6830470561981201, 'train/grad_norm_pre_clip': 9.526907920837402, 'timing/data': 0.00036524294409900904, 'timing/model': 1.2143399589695036, 'learning_rate/qwen_vl_interface': 6.084766202075944e-06, 'learning_rate/action_model': 6.0847662020759435e-05, 'epoch': 0.93}) +[2026-06-17 12:05:46,994][starVLA.training.train_starvla][INFO] - Step 3163, Loss: {'train/loss': 0.971163809299469, 'train/grad_norm_pre_clip': 12.859269142150879, 'timing/data': 0.000373549060896039, 'timing/model': 1.2070547989569604, 'learning_rate/qwen_vl_interface': 6.082637083910955e-06, 'learning_rate/action_model': 6.0826370839109546e-05, 'epoch': 0.93}) +[2026-06-17 12:05:51,790][starVLA.training.train_starvla][INFO] - Step 3164, Loss: {'train/loss': 0.6748696565628052, 'train/grad_norm_pre_clip': 9.557626724243164, 'timing/data': 0.00034771300852298737, 'timing/model': 1.2117595439776778, 'learning_rate/qwen_vl_interface': 6.080507793139403e-06, 'learning_rate/action_model': 6.080507793139403e-05, 'epoch': 0.93}) +[2026-06-17 12:05:56,591][starVLA.training.train_starvla][INFO] - Step 3165, Loss: {'train/loss': 0.7211659550666809, 'train/grad_norm_pre_clip': 10.06454849243164, 'timing/data': 0.00035526405554264784, 'timing/model': 1.2047151940641925, 'learning_rate/qwen_vl_interface': 6.0783783302026925e-06, 'learning_rate/action_model': 6.078378330202692e-05, 'epoch': 0.93}) +[2026-06-17 12:06:01,410][starVLA.training.train_starvla][INFO] - Step 3166, Loss: {'train/loss': 0.7417492866516113, 'train/grad_norm_pre_clip': 10.749191284179688, 'timing/data': 0.00041571096517145634, 'timing/model': 1.2083211479475722, 'learning_rate/qwen_vl_interface': 6.076248695542262e-06, 'learning_rate/action_model': 6.076248695542261e-05, 'epoch': 0.93}) +[2026-06-17 12:06:06,206][starVLA.training.train_starvla][INFO] - Step 3167, Loss: {'train/loss': 0.7423217296600342, 'train/grad_norm_pre_clip': 10.703369140625, 'timing/data': 0.00035561900585889816, 'timing/model': 1.331377569003962, 'learning_rate/qwen_vl_interface': 6.074118889599589e-06, 'learning_rate/action_model': 6.0741188895995884e-05, 'epoch': 0.93}) +[2026-06-17 12:06:10,856][starVLA.training.train_starvla][INFO] - Step 3168, Loss: {'train/loss': 1.0012693405151367, 'train/grad_norm_pre_clip': 13.711371421813965, 'timing/data': 0.0003867560299113393, 'timing/model': 1.2042812269646674, 'learning_rate/qwen_vl_interface': 6.07198891281618e-06, 'learning_rate/action_model': 6.07198891281618e-05, 'epoch': 0.93}) +[2026-06-17 12:06:15,649][starVLA.training.train_starvla][INFO] - Step 3169, Loss: {'train/loss': 0.7259423136711121, 'train/grad_norm_pre_clip': 12.960421562194824, 'timing/data': 0.0005111249629408121, 'timing/model': 1.2052837869850919, 'learning_rate/qwen_vl_interface': 6.0698587656335895e-06, 'learning_rate/action_model': 6.069858765633589e-05, 'epoch': 0.93}) +[2026-06-17 12:06:20,446][starVLA.training.train_starvla][INFO] - Step 3170, Loss: {'train/loss': 0.8467302918434143, 'train/grad_norm_pre_clip': 13.751995086669922, 'timing/data': 0.00036504899617284536, 'timing/model': 1.2054456680780277, 'learning_rate/qwen_vl_interface': 6.067728448493392e-06, 'learning_rate/action_model': 6.067728448493392e-05, 'epoch': 0.93}) +[2026-06-17 12:06:25,202][starVLA.training.train_starvla][INFO] - Step 3171, Loss: {'train/loss': 0.765465259552002, 'train/grad_norm_pre_clip': 11.63575553894043, 'timing/data': 0.0003930979873985052, 'timing/model': 1.205005232943222, 'learning_rate/qwen_vl_interface': 6.065597961837209e-06, 'learning_rate/action_model': 6.0655979618372086e-05, 'epoch': 0.93}) +[2026-06-17 12:06:30,000][starVLA.training.train_starvla][INFO] - Step 3172, Loss: {'train/loss': 0.8264084458351135, 'train/grad_norm_pre_clip': 10.774444580078125, 'timing/data': 0.00037030200473964214, 'timing/model': 1.2051275060512125, 'learning_rate/qwen_vl_interface': 6.063467306106688e-06, 'learning_rate/action_model': 6.0634673061066885e-05, 'epoch': 0.93}) +[2026-06-17 12:06:34,786][starVLA.training.train_starvla][INFO] - Step 3173, Loss: {'train/loss': 0.5487759113311768, 'train/grad_norm_pre_clip': 27.30474090576172, 'timing/data': 0.0003915420966222882, 'timing/model': 1.3332642709137872, 'learning_rate/qwen_vl_interface': 6.061336481743522e-06, 'learning_rate/action_model': 6.061336481743522e-05, 'epoch': 0.93}) +[2026-06-17 12:06:39,437][starVLA.training.train_starvla][INFO] - Step 3174, Loss: {'train/loss': 0.3959914743900299, 'train/grad_norm_pre_clip': 10.026792526245117, 'timing/data': 0.0003885290352627635, 'timing/model': 1.203698800993152, 'learning_rate/qwen_vl_interface': 6.059205489189429e-06, 'learning_rate/action_model': 6.059205489189429e-05, 'epoch': 0.93}) +[2026-06-17 12:06:44,255][starVLA.training.train_starvla][INFO] - Step 3175, Loss: {'train/loss': 0.5636993646621704, 'train/grad_norm_pre_clip': 9.562438011169434, 'timing/data': 0.00041434098966419697, 'timing/model': 1.2207729129586369, 'learning_rate/qwen_vl_interface': 6.057074328886168e-06, 'learning_rate/action_model': 6.057074328886168e-05, 'epoch': 0.93}) +[2026-06-17 12:06:49,091][starVLA.training.train_starvla][INFO] - Step 3176, Loss: {'train/loss': 0.4617388844490051, 'train/grad_norm_pre_clip': 8.54570198059082, 'timing/data': 0.00036275095771998167, 'timing/model': 1.2034772359766066, 'learning_rate/qwen_vl_interface': 6.0549430012755295e-06, 'learning_rate/action_model': 6.0549430012755295e-05, 'epoch': 0.93}) +[2026-06-17 12:06:53,868][starVLA.training.train_starvla][INFO] - Step 3177, Loss: {'train/loss': 0.8296858072280884, 'train/grad_norm_pre_clip': 9.435898780822754, 'timing/data': 0.0007724869064986706, 'timing/model': 1.2034135139547288, 'learning_rate/qwen_vl_interface': 6.052811506799341e-06, 'learning_rate/action_model': 6.052811506799341e-05, 'epoch': 0.93}) +[2026-06-17 12:06:58,661][starVLA.training.train_starvla][INFO] - Step 3178, Loss: {'train/loss': 0.7555669546127319, 'train/grad_norm_pre_clip': 11.39292049407959, 'timing/data': 0.00041194004006683826, 'timing/model': 1.2051744149066508, 'learning_rate/qwen_vl_interface': 6.050679845899462e-06, 'learning_rate/action_model': 6.050679845899462e-05, 'epoch': 0.94}) +[2026-06-17 12:07:03,440][starVLA.training.train_starvla][INFO] - Step 3179, Loss: {'train/loss': 0.7089876532554626, 'train/grad_norm_pre_clip': 10.334086418151855, 'timing/data': 0.0003569949185475707, 'timing/model': 1.3327939660521224, 'learning_rate/qwen_vl_interface': 6.048548019017791e-06, 'learning_rate/action_model': 6.04854801901779e-05, 'epoch': 0.94}) +[2026-06-17 12:07:08,102][starVLA.training.train_starvla][INFO] - Step 3180, Loss: {'train/loss': 0.7812303304672241, 'train/grad_norm_pre_clip': 11.479254722595215, 'timing/data': 0.00037376489490270615, 'timing/model': 1.211311973980628, 'learning_rate/qwen_vl_interface': 6.046416026596256e-06, 'learning_rate/action_model': 6.046416026596255e-05, 'epoch': 0.94}) +[2026-06-17 12:07:12,894][starVLA.training.train_starvla][INFO] - Step 3181, Loss: {'train/loss': 0.7475734353065491, 'train/grad_norm_pre_clip': 10.512235641479492, 'timing/data': 0.00037454499397426844, 'timing/model': 1.210028923000209, 'learning_rate/qwen_vl_interface': 6.044283869076822e-06, 'learning_rate/action_model': 6.0442838690768224e-05, 'epoch': 0.94}) +[2026-06-17 12:07:17,683][starVLA.training.train_starvla][INFO] - Step 3182, Loss: {'train/loss': 0.6509993076324463, 'train/grad_norm_pre_clip': 10.406451225280762, 'timing/data': 0.000570676987990737, 'timing/model': 1.2071089929668233, 'learning_rate/qwen_vl_interface': 6.042151546901487e-06, 'learning_rate/action_model': 6.042151546901487e-05, 'epoch': 0.94}) +[2026-06-17 12:07:22,464][starVLA.training.train_starvla][INFO] - Step 3183, Loss: {'train/loss': 0.8824065923690796, 'train/grad_norm_pre_clip': 10.95130443572998, 'timing/data': 0.00039709999691694975, 'timing/model': 1.2045596030075103, 'learning_rate/qwen_vl_interface': 6.0400190605122846e-06, 'learning_rate/action_model': 6.040019060512284e-05, 'epoch': 0.94}) +[2026-06-17 12:07:27,238][starVLA.training.train_starvla][INFO] - Step 3184, Loss: {'train/loss': 0.901973307132721, 'train/grad_norm_pre_clip': 9.713406562805176, 'timing/data': 0.00033494201488792896, 'timing/model': 1.3269191179424524, 'learning_rate/qwen_vl_interface': 6.0378864103512805e-06, 'learning_rate/action_model': 6.03788641035128e-05, 'epoch': 0.94}) +[2026-06-17 12:07:31,980][starVLA.training.train_starvla][INFO] - Step 3185, Loss: {'train/loss': 0.6017749905586243, 'train/grad_norm_pre_clip': 8.578407287597656, 'timing/data': 0.00036645098589360714, 'timing/model': 1.2940315030282363, 'learning_rate/qwen_vl_interface': 6.035753596860573e-06, 'learning_rate/action_model': 6.0357535968605725e-05, 'epoch': 0.94}) +[2026-06-17 12:07:36,643][starVLA.training.train_starvla][INFO] - Step 3186, Loss: {'train/loss': 1.0061534643173218, 'train/grad_norm_pre_clip': 13.409486770629883, 'timing/data': 0.00037424592301249504, 'timing/model': 1.205716872937046, 'learning_rate/qwen_vl_interface': 6.033620620482303e-06, 'learning_rate/action_model': 6.033620620482303e-05, 'epoch': 0.94}) +[2026-06-17 12:07:41,429][starVLA.training.train_starvla][INFO] - Step 3187, Loss: {'train/loss': 0.6435256600379944, 'train/grad_norm_pre_clip': 12.1847562789917, 'timing/data': 0.0006249030120670795, 'timing/model': 1.203723531914875, 'learning_rate/qwen_vl_interface': 6.031487481658633e-06, 'learning_rate/action_model': 6.031487481658633e-05, 'epoch': 0.94}) +[2026-06-17 12:07:46,293][starVLA.training.train_starvla][INFO] - Step 3188, Loss: {'train/loss': 0.8759573698043823, 'train/grad_norm_pre_clip': 12.433605194091797, 'timing/data': 0.0005297289462760091, 'timing/model': 1.208785132970661, 'learning_rate/qwen_vl_interface': 6.029354180831769e-06, 'learning_rate/action_model': 6.029354180831769e-05, 'epoch': 0.94}) +[2026-06-17 12:07:51,067][starVLA.training.train_starvla][INFO] - Step 3189, Loss: {'train/loss': 0.5210832953453064, 'train/grad_norm_pre_clip': 13.732537269592285, 'timing/data': 0.0003565700026229024, 'timing/model': 1.2023398730671033, 'learning_rate/qwen_vl_interface': 6.027220718443943e-06, 'learning_rate/action_model': 6.027220718443943e-05, 'epoch': 0.94}) +[2026-06-17 12:07:55,843][starVLA.training.train_starvla][INFO] - Step 3190, Loss: {'train/loss': 0.877056896686554, 'train/grad_norm_pre_clip': 12.312585830688477, 'timing/data': 0.0003512949915602803, 'timing/model': 1.3331823209300637, 'learning_rate/qwen_vl_interface': 6.025087094937426e-06, 'learning_rate/action_model': 6.025087094937426e-05, 'epoch': 0.94}) +[2026-06-17 12:08:00,503][starVLA.training.train_starvla][INFO] - Step 3191, Loss: {'train/loss': 0.6850926876068115, 'train/grad_norm_pre_clip': 12.471344947814941, 'timing/data': 0.00041121593676507473, 'timing/model': 1.2104456080123782, 'learning_rate/qwen_vl_interface': 6.022953310754521e-06, 'learning_rate/action_model': 6.022953310754521e-05, 'epoch': 0.94}) +[2026-06-17 12:08:05,152][starVLA.training.train_starvla][INFO] - Step 3192, Loss: {'train/loss': 0.9550613760948181, 'train/grad_norm_pre_clip': 14.372037887573242, 'timing/data': 0.0004017370520159602, 'timing/model': 1.2024532300420105, 'learning_rate/qwen_vl_interface': 6.020819366337562e-06, 'learning_rate/action_model': 6.020819366337562e-05, 'epoch': 0.94}) +[2026-06-17 12:08:09,926][starVLA.training.train_starvla][INFO] - Step 3193, Loss: {'train/loss': 0.6249880194664001, 'train/grad_norm_pre_clip': 14.422965049743652, 'timing/data': 0.0003404750023037195, 'timing/model': 1.2031861180439591, 'learning_rate/qwen_vl_interface': 6.018685262128921e-06, 'learning_rate/action_model': 6.01868526212892e-05, 'epoch': 0.94}) +[2026-06-17 12:08:14,714][starVLA.training.train_starvla][INFO] - Step 3194, Loss: {'train/loss': 0.6833234429359436, 'train/grad_norm_pre_clip': 10.787936210632324, 'timing/data': 0.00037437991704791784, 'timing/model': 1.203076792997308, 'learning_rate/qwen_vl_interface': 6.016550998570995e-06, 'learning_rate/action_model': 6.0165509985709944e-05, 'epoch': 0.94}) +[2026-06-17 12:08:19,496][starVLA.training.train_starvla][INFO] - Step 3195, Loss: {'train/loss': 1.2272557020187378, 'train/grad_norm_pre_clip': 15.29822826385498, 'timing/data': 0.0003442550078034401, 'timing/model': 1.2047757080290467, 'learning_rate/qwen_vl_interface': 6.014416576106224e-06, 'learning_rate/action_model': 6.0144165761062244e-05, 'epoch': 0.94}) +[2026-06-17 12:08:24,277][starVLA.training.train_starvla][INFO] - Step 3196, Loss: {'train/loss': 0.7279362082481384, 'train/grad_norm_pre_clip': 9.75841999053955, 'timing/data': 0.0003442430170252919, 'timing/model': 1.335925890947692, 'learning_rate/qwen_vl_interface': 6.012281995177074e-06, 'learning_rate/action_model': 6.012281995177074e-05, 'epoch': 0.94}) +[2026-06-17 12:08:28,932][starVLA.training.train_starvla][INFO] - Step 3197, Loss: {'train/loss': 0.5909889936447144, 'train/grad_norm_pre_clip': 9.806114196777344, 'timing/data': 0.0003682660171762109, 'timing/model': 1.2037721689557657, 'learning_rate/qwen_vl_interface': 6.010147256226046e-06, 'learning_rate/action_model': 6.0101472562260454e-05, 'epoch': 0.94}) +[2026-06-17 12:08:33,732][starVLA.training.train_starvla][INFO] - Step 3198, Loss: {'train/loss': 1.0067458152770996, 'train/grad_norm_pre_clip': 20.45515251159668, 'timing/data': 0.0003650059225037694, 'timing/model': 1.2066743150353432, 'learning_rate/qwen_vl_interface': 6.008012359695673e-06, 'learning_rate/action_model': 6.0080123596956726e-05, 'epoch': 0.94}) +[2026-06-17 12:08:38,518][starVLA.training.train_starvla][INFO] - Step 3199, Loss: {'train/loss': 0.7109032273292542, 'train/grad_norm_pre_clip': 10.51917839050293, 'timing/data': 0.0003415970131754875, 'timing/model': 1.204652048065327, 'learning_rate/qwen_vl_interface': 6.005877306028522e-06, 'learning_rate/action_model': 6.005877306028522e-05, 'epoch': 0.94}) +[2026-06-17 12:10:20,618][starVLA.training.train_starvla][INFO] - Step 3200, Loss: {'train/loss': 0.4651646912097931, 'train/grad_norm_pre_clip': 8.227983474731445, 'eval/loss': 0.7764977812767029, 'eval/latency_0/loss': 0.7764977812767029, 'eval/demon_fire_precision': 0.8675696012039127, 'eval/demon_fire_recall': 0.957641196013289, 'eval/demon_fire_f1': 0.9103829451243584, 'eval/demon_fire_rate_model': 0.739154616240267, 'eval/demon_fire_rate_teacher': 0.6696329254727474, 'eval/demon_left_precision': 0.8424050632911393, 'eval/demon_left_recall': 0.7541076487252124, 'eval/demon_left_f1': 0.7958146487294468, 'eval/demon_left_rate_model': 0.43937708565072303, 'eval/demon_left_rate_teacher': 0.49082313681868744, 'eval/demon_right_precision': 0.7736318407960199, 'eval/demon_right_recall': 0.8785310734463276, 'eval/demon_right_f1': 0.8227513227513227, 'eval/demon_right_rate_model': 0.5589543937708565, 'eval/demon_right_rate_teacher': 0.4922135706340378, 'eval/demon_move_f1': 0.8092829857403847, 'eval/demon_cc_f1': 0.8774964789757351, 'eval/latency_0/demon_fire_precision': 0.8675696012039127, 'eval/latency_0/demon_fire_recall': 0.957641196013289, 'eval/latency_0/demon_fire_f1': 0.9103829451243584, 'eval/latency_0/demon_fire_rate_model': 0.739154616240267, 'eval/latency_0/demon_fire_rate_teacher': 0.6696329254727474, 'eval/latency_0/demon_left_precision': 0.8424050632911393, 'eval/latency_0/demon_left_recall': 0.7541076487252124, 'eval/latency_0/demon_left_f1': 0.7958146487294468, 'eval/latency_0/demon_left_rate_model': 0.43937708565072303, 'eval/latency_0/demon_left_rate_teacher': 0.49082313681868744, 'eval/latency_0/demon_right_precision': 0.7736318407960199, 'eval/latency_0/demon_right_recall': 0.8785310734463276, 'eval/latency_0/demon_right_f1': 0.8227513227513227, 'eval/latency_0/demon_right_rate_model': 0.5589543937708565, 'eval/latency_0/demon_right_rate_teacher': 0.4922135706340378, 'eval/latency_0/demon_move_f1': 0.8092829857403847, 'eval/latency_0/demon_cc_f1': 0.8774964789757351, 'timing/data': 0.0005788670387119055, 'timing/model': 1.2048045949777588, 'learning_rate/qwen_vl_interface': 6.003742095667194e-06, 'learning_rate/action_model': 6.003742095667194e-05, 'epoch': 0.94}) +[2026-06-17 12:10:20,618][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_3200_state +[2026-06-17 12:10:20,621][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 12:10:24,164][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_3200_state/model.safetensors +[2026-06-17 12:10:31,795][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_3200_state/optimizer.bin +[2026-06-17 12:10:31,795][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_3200_state/sampler.bin +[2026-06-17 12:10:31,795][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_3200_state/sampler_1.bin +[2026-06-17 12:10:31,796][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_3200_state/random_states_0.pkl +[2026-06-17 12:10:31,848][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3200 +[2026-06-17 12:10:31,848][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 12:10:31,921][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 12:10:36,651][starVLA.training.train_starvla][INFO] - Step 3201, Loss: {'train/loss': 0.5523355007171631, 'train/grad_norm_pre_clip': 9.376959800720215, 'timing/data': 0.0003367930185049772, 'timing/model': 1.190213070018217, 'learning_rate/qwen_vl_interface': 6.001606729054318e-06, 'learning_rate/action_model': 6.0016067290543176e-05, 'epoch': 0.94}) +[2026-06-17 12:10:41,399][starVLA.training.train_starvla][INFO] - Step 3202, Loss: {'train/loss': 0.6764367818832397, 'train/grad_norm_pre_clip': 10.52531909942627, 'timing/data': 0.00033345306292176247, 'timing/model': 1.1898258249275386, 'learning_rate/qwen_vl_interface': 5.999471206632559e-06, 'learning_rate/action_model': 5.999471206632559e-05, 'epoch': 0.94}) +[2026-06-17 12:10:46,132][starVLA.training.train_starvla][INFO] - Step 3203, Loss: {'train/loss': 0.9284226894378662, 'train/grad_norm_pre_clip': 11.755367279052734, 'timing/data': 0.0004023300716653466, 'timing/model': 1.1909509151009843, 'learning_rate/qwen_vl_interface': 5.99733552884461e-06, 'learning_rate/action_model': 5.99733552884461e-05, 'epoch': 0.94}) +[2026-06-17 12:10:50,879][starVLA.training.train_starvla][INFO] - Step 3204, Loss: {'train/loss': 1.0443332195281982, 'train/grad_norm_pre_clip': 11.62108325958252, 'timing/data': 0.00047885498497635126, 'timing/model': 1.3356589450268075, 'learning_rate/qwen_vl_interface': 5.995199696133204e-06, 'learning_rate/action_model': 5.9951996961332034e-05, 'epoch': 0.94}) +[2026-06-17 12:10:55,499][starVLA.training.train_starvla][INFO] - Step 3205, Loss: {'train/loss': 0.5288564562797546, 'train/grad_norm_pre_clip': 7.959262847900391, 'timing/data': 0.0003666799748316407, 'timing/model': 1.1959894360043108, 'learning_rate/qwen_vl_interface': 5.993063708941098e-06, 'learning_rate/action_model': 5.993063708941098e-05, 'epoch': 0.94}) +[2026-06-17 12:11:00,275][starVLA.training.train_starvla][INFO] - Step 3206, Loss: {'train/loss': 0.6420038938522339, 'train/grad_norm_pre_clip': 12.723865509033203, 'timing/data': 0.0003684191033244133, 'timing/model': 1.203596057021059, 'learning_rate/qwen_vl_interface': 5.990927567711085e-06, 'learning_rate/action_model': 5.990927567711085e-05, 'epoch': 0.94}) +[2026-06-17 12:11:05,033][starVLA.training.train_starvla][INFO] - Step 3207, Loss: {'train/loss': 0.9391172528266907, 'train/grad_norm_pre_clip': 11.998698234558105, 'timing/data': 0.00038401701021939516, 'timing/model': 1.1970941389445215, 'learning_rate/qwen_vl_interface': 5.988791272885991e-06, 'learning_rate/action_model': 5.988791272885991e-05, 'epoch': 0.94}) +[2026-06-17 12:11:09,879][starVLA.training.train_starvla][INFO] - Step 3208, Loss: {'train/loss': 0.5895518064498901, 'train/grad_norm_pre_clip': 8.998735427856445, 'timing/data': 0.00037186103872954845, 'timing/model': 1.1984012559987605, 'learning_rate/qwen_vl_interface': 5.986654824908671e-06, 'learning_rate/action_model': 5.9866548249086706e-05, 'epoch': 0.94}) +[2026-06-17 12:11:14,665][starVLA.training.train_starvla][INFO] - Step 3209, Loss: {'train/loss': 0.5905860066413879, 'train/grad_norm_pre_clip': 15.999248504638672, 'timing/data': 0.00037496990989893675, 'timing/model': 1.3416467939969152, 'learning_rate/qwen_vl_interface': 5.9845182242220125e-06, 'learning_rate/action_model': 5.984518224222012e-05, 'epoch': 0.94}) +[2026-06-17 12:11:19,381][starVLA.training.train_starvla][INFO] - Step 3210, Loss: {'train/loss': 0.8959685564041138, 'train/grad_norm_pre_clip': 11.096946716308594, 'timing/data': 0.0003872489323839545, 'timing/model': 1.2776493660639971, 'learning_rate/qwen_vl_interface': 5.982381471268935e-06, 'learning_rate/action_model': 5.9823814712689354e-05, 'epoch': 0.94}) +[2026-06-17 12:11:24,021][starVLA.training.train_starvla][INFO] - Step 3211, Loss: {'train/loss': 0.3675116002559662, 'train/grad_norm_pre_clip': 6.934577465057373, 'timing/data': 0.00040884793270379305, 'timing/model': 1.1991687450790778, 'learning_rate/qwen_vl_interface': 5.980244566492392e-06, 'learning_rate/action_model': 5.9802445664923925e-05, 'epoch': 0.94}) +[2026-06-17 12:11:28,794][starVLA.training.train_starvla][INFO] - Step 3212, Loss: {'train/loss': 0.8951819539070129, 'train/grad_norm_pre_clip': 12.754293441772461, 'timing/data': 0.0003980121109634638, 'timing/model': 1.1992533369921148, 'learning_rate/qwen_vl_interface': 5.978107510335365e-06, 'learning_rate/action_model': 5.9781075103353647e-05, 'epoch': 0.95}) +[2026-06-17 12:11:33,561][starVLA.training.train_starvla][INFO] - Step 3213, Loss: {'train/loss': 0.6554601192474365, 'train/grad_norm_pre_clip': 10.466065406799316, 'timing/data': 0.00036296702455729246, 'timing/model': 1.2003723869565874, 'learning_rate/qwen_vl_interface': 5.975970303240867e-06, 'learning_rate/action_model': 5.9759703032408664e-05, 'epoch': 0.95}) +[2026-06-17 12:11:38,340][starVLA.training.train_starvla][INFO] - Step 3214, Loss: {'train/loss': 0.8296739459037781, 'train/grad_norm_pre_clip': 41.45439529418945, 'timing/data': 0.00034400797449052334, 'timing/model': 1.2064155070111156, 'learning_rate/qwen_vl_interface': 5.973832945651944e-06, 'learning_rate/action_model': 5.9738329456519435e-05, 'epoch': 0.95}) +[2026-06-17 12:11:43,201][starVLA.training.train_starvla][INFO] - Step 3215, Loss: {'train/loss': 0.687603771686554, 'train/grad_norm_pre_clip': 10.147279739379883, 'timing/data': 0.00034858391154557467, 'timing/model': 1.3288363290484995, 'learning_rate/qwen_vl_interface': 5.971695438011672e-06, 'learning_rate/action_model': 5.971695438011672e-05, 'epoch': 0.95}) +[2026-06-17 12:11:47,870][starVLA.training.train_starvla][INFO] - Step 3216, Loss: {'train/loss': 0.8804216384887695, 'train/grad_norm_pre_clip': 8.427905082702637, 'timing/data': 0.0003735930658876896, 'timing/model': 1.228025613934733, 'learning_rate/qwen_vl_interface': 5.969557780763159e-06, 'learning_rate/action_model': 5.9695577807631585e-05, 'epoch': 0.95}) +[2026-06-17 12:11:52,510][starVLA.training.train_starvla][INFO] - Step 3217, Loss: {'train/loss': 0.8007320165634155, 'train/grad_norm_pre_clip': 12.326680183410645, 'timing/data': 0.00038414797745645046, 'timing/model': 1.2001766118919477, 'learning_rate/qwen_vl_interface': 5.9674199743495456e-06, 'learning_rate/action_model': 5.967419974349545e-05, 'epoch': 0.95}) +[2026-06-17 12:11:57,281][starVLA.training.train_starvla][INFO] - Step 3218, Loss: {'train/loss': 0.8737528324127197, 'train/grad_norm_pre_clip': 16.07478904724121, 'timing/data': 0.00045551301445811987, 'timing/model': 1.2002480759983882, 'learning_rate/qwen_vl_interface': 5.965282019213996e-06, 'learning_rate/action_model': 5.965282019213996e-05, 'epoch': 0.95}) +[2026-06-17 12:12:02,088][starVLA.training.train_starvla][INFO] - Step 3219, Loss: {'train/loss': 0.6900787353515625, 'train/grad_norm_pre_clip': 12.23654842376709, 'timing/data': 0.0003649670397862792, 'timing/model': 1.2006088629132137, 'learning_rate/qwen_vl_interface': 5.963143915799718e-06, 'learning_rate/action_model': 5.9631439157997184e-05, 'epoch': 0.95}) +[2026-06-17 12:12:06,885][starVLA.training.train_starvla][INFO] - Step 3220, Loss: {'train/loss': 0.7980185747146606, 'train/grad_norm_pre_clip': 10.588699340820312, 'timing/data': 0.0003576120361685753, 'timing/model': 1.2035164620028809, 'learning_rate/qwen_vl_interface': 5.961005664549936e-06, 'learning_rate/action_model': 5.9610056645499354e-05, 'epoch': 0.95}) +[2026-06-17 12:12:11,668][starVLA.training.train_starvla][INFO] - Step 3221, Loss: {'train/loss': 0.7999886274337769, 'train/grad_norm_pre_clip': 11.430281639099121, 'timing/data': 0.0003747990122064948, 'timing/model': 1.3327262599486858, 'learning_rate/qwen_vl_interface': 5.958867265907916e-06, 'learning_rate/action_model': 5.958867265907916e-05, 'epoch': 0.95}) +[2026-06-17 12:12:16,334][starVLA.training.train_starvla][INFO] - Step 3222, Loss: {'train/loss': 0.7351208329200745, 'train/grad_norm_pre_clip': 10.782769203186035, 'timing/data': 0.0003839649725705385, 'timing/model': 1.201044088928029, 'learning_rate/qwen_vl_interface': 5.956728720316949e-06, 'learning_rate/action_model': 5.9567287203169485e-05, 'epoch': 0.95}) +[2026-06-17 12:12:21,155][starVLA.training.train_starvla][INFO] - Step 3223, Loss: {'train/loss': 0.7151932120323181, 'train/grad_norm_pre_clip': 9.664813995361328, 'timing/data': 0.0003749459283426404, 'timing/model': 1.206247398047708, 'learning_rate/qwen_vl_interface': 5.954590028220357e-06, 'learning_rate/action_model': 5.954590028220357e-05, 'epoch': 0.95}) +[2026-06-17 12:12:25,946][starVLA.training.train_starvla][INFO] - Step 3224, Loss: {'train/loss': 0.737920343875885, 'train/grad_norm_pre_clip': 10.894325256347656, 'timing/data': 0.0003672180464491248, 'timing/model': 1.2153569760266691, 'learning_rate/qwen_vl_interface': 5.952451190061496e-06, 'learning_rate/action_model': 5.9524511900614956e-05, 'epoch': 0.95}) +[2026-06-17 12:12:30,633][starVLA.training.train_starvla][INFO] - Step 3225, Loss: {'train/loss': 0.572247326374054, 'train/grad_norm_pre_clip': 8.056386947631836, 'timing/data': 0.00035844394005835056, 'timing/model': 1.2060590490000322, 'learning_rate/qwen_vl_interface': 5.950312206283743e-06, 'learning_rate/action_model': 5.9503122062837434e-05, 'epoch': 0.95}) +[2026-06-17 12:12:35,316][starVLA.training.train_starvla][INFO] - Step 3226, Loss: {'train/loss': 0.4076426029205322, 'train/grad_norm_pre_clip': 6.277346134185791, 'timing/data': 0.00035785301588475704, 'timing/model': 1.2154216449707747, 'learning_rate/qwen_vl_interface': 5.948173077330522e-06, 'learning_rate/action_model': 5.948173077330522e-05, 'epoch': 0.95}) +[2026-06-17 12:12:40,038][starVLA.training.train_starvla][INFO] - Step 3227, Loss: {'train/loss': 0.296142578125, 'train/grad_norm_pre_clip': 7.666316986083984, 'timing/data': 0.00035953696351498365, 'timing/model': 1.2498253900557756, 'learning_rate/qwen_vl_interface': 5.9460338036452665e-06, 'learning_rate/action_model': 5.946033803645267e-05, 'epoch': 0.95}) +[2026-06-17 12:12:44,772][starVLA.training.train_starvla][INFO] - Step 3228, Loss: {'train/loss': 0.7290371656417847, 'train/grad_norm_pre_clip': 10.295286178588867, 'timing/data': 0.000420068041421473, 'timing/model': 1.2014582350384444, 'learning_rate/qwen_vl_interface': 5.943894385671456e-06, 'learning_rate/action_model': 5.9438943856714556e-05, 'epoch': 0.95}) +[2026-06-17 12:12:49,526][starVLA.training.train_starvla][INFO] - Step 3229, Loss: {'train/loss': 0.6820767521858215, 'train/grad_norm_pre_clip': 12.670323371887207, 'timing/data': 0.00040446699131280184, 'timing/model': 1.2129151640692726, 'learning_rate/qwen_vl_interface': 5.941754823852593e-06, 'learning_rate/action_model': 5.9417548238525924e-05, 'epoch': 0.95}) +[2026-06-17 12:12:54,290][starVLA.training.train_starvla][INFO] - Step 3230, Loss: {'train/loss': 0.786188006401062, 'train/grad_norm_pre_clip': 10.687525749206543, 'timing/data': 0.00034447305370122194, 'timing/model': 1.208104739896953, 'learning_rate/qwen_vl_interface': 5.939615118632208e-06, 'learning_rate/action_model': 5.939615118632208e-05, 'epoch': 0.95}) +[2026-06-17 12:12:59,097][starVLA.training.train_starvla][INFO] - Step 3231, Loss: {'train/loss': 0.7702062129974365, 'train/grad_norm_pre_clip': 14.27541732788086, 'timing/data': 0.00040447700303047895, 'timing/model': 1.2144881850108504, 'learning_rate/qwen_vl_interface': 5.937475270453872e-06, 'learning_rate/action_model': 5.937475270453872e-05, 'epoch': 0.95}) +[2026-06-17 12:13:03,912][starVLA.training.train_starvla][INFO] - Step 3232, Loss: {'train/loss': 0.7939709424972534, 'train/grad_norm_pre_clip': 14.86052131652832, 'timing/data': 0.0003738650120794773, 'timing/model': 1.3555337670259178, 'learning_rate/qwen_vl_interface': 5.935335279761169e-06, 'learning_rate/action_model': 5.935335279761168e-05, 'epoch': 0.95}) +[2026-06-17 12:13:08,605][starVLA.training.train_starvla][INFO] - Step 3233, Loss: {'train/loss': 0.5910846590995789, 'train/grad_norm_pre_clip': 15.804473876953125, 'timing/data': 0.00035202596336603165, 'timing/model': 1.2159217110602185, 'learning_rate/qwen_vl_interface': 5.933195146997725e-06, 'learning_rate/action_model': 5.9331951469977254e-05, 'epoch': 0.95}) +[2026-06-17 12:13:13,397][starVLA.training.train_starvla][INFO] - Step 3234, Loss: {'train/loss': 0.8298214077949524, 'train/grad_norm_pre_clip': 11.330273628234863, 'timing/data': 0.000327872927300632, 'timing/model': 1.2116056549130008, 'learning_rate/qwen_vl_interface': 5.931054872607193e-06, 'learning_rate/action_model': 5.9310548726071925e-05, 'epoch': 0.95}) +[2026-06-17 12:13:18,164][starVLA.training.train_starvla][INFO] - Step 3235, Loss: {'train/loss': 0.6251033544540405, 'train/grad_norm_pre_clip': 9.475728034973145, 'timing/data': 0.0004166560247540474, 'timing/model': 1.2035888449754566, 'learning_rate/qwen_vl_interface': 5.928914457033252e-06, 'learning_rate/action_model': 5.928914457033252e-05, 'epoch': 0.95}) +[2026-06-17 12:13:22,951][starVLA.training.train_starvla][INFO] - Step 3236, Loss: {'train/loss': 0.7976133823394775, 'train/grad_norm_pre_clip': 12.211371421813965, 'timing/data': 0.0003313759807497263, 'timing/model': 1.2049405060242862, 'learning_rate/qwen_vl_interface': 5.926773900719616e-06, 'learning_rate/action_model': 5.9267739007196157e-05, 'epoch': 0.95}) +[2026-06-17 12:13:27,734][starVLA.training.train_starvla][INFO] - Step 3237, Loss: {'train/loss': 0.8230661749839783, 'train/grad_norm_pre_clip': 11.490870475769043, 'timing/data': 0.0003384280717000365, 'timing/model': 1.2059341740095988, 'learning_rate/qwen_vl_interface': 5.92463320411002e-06, 'learning_rate/action_model': 5.92463320411002e-05, 'epoch': 0.95}) +[2026-06-17 12:13:32,536][starVLA.training.train_starvla][INFO] - Step 3238, Loss: {'train/loss': 0.8381226658821106, 'train/grad_norm_pre_clip': 16.59906768798828, 'timing/data': 0.000533160986378789, 'timing/model': 1.3465840090066195, 'learning_rate/qwen_vl_interface': 5.922492367648236e-06, 'learning_rate/action_model': 5.922492367648236e-05, 'epoch': 0.95}) +[2026-06-17 12:13:37,214][starVLA.training.train_starvla][INFO] - Step 3239, Loss: {'train/loss': 0.6320934295654297, 'train/grad_norm_pre_clip': 31.449234008789062, 'timing/data': 0.0005240739556029439, 'timing/model': 1.207185449078679, 'learning_rate/qwen_vl_interface': 5.920351391778059e-06, 'learning_rate/action_model': 5.920351391778059e-05, 'epoch': 0.95}) +[2026-06-17 12:13:42,016][starVLA.training.train_starvla][INFO] - Step 3240, Loss: {'train/loss': 0.707489550113678, 'train/grad_norm_pre_clip': 9.477639198303223, 'timing/data': 0.0003679930232465267, 'timing/model': 1.2188572050072253, 'learning_rate/qwen_vl_interface': 5.918210276943318e-06, 'learning_rate/action_model': 5.918210276943318e-05, 'epoch': 0.95}) +[2026-06-17 12:13:46,690][starVLA.training.train_starvla][INFO] - Step 3241, Loss: {'train/loss': 0.7603927254676819, 'train/grad_norm_pre_clip': 10.99902057647705, 'timing/data': 0.0003649899736046791, 'timing/model': 1.2060937619535252, 'learning_rate/qwen_vl_interface': 5.916069023587867e-06, 'learning_rate/action_model': 5.916069023587867e-05, 'epoch': 0.95}) +[2026-06-17 12:13:51,386][starVLA.training.train_starvla][INFO] - Step 3242, Loss: {'train/loss': 0.988143265247345, 'train/grad_norm_pre_clip': 14.122126579284668, 'timing/data': 0.0003527830122038722, 'timing/model': 1.2107865640427917, 'learning_rate/qwen_vl_interface': 5.913927632155592e-06, 'learning_rate/action_model': 5.913927632155591e-05, 'epoch': 0.95}) +[2026-06-17 12:13:56,106][starVLA.training.train_starvla][INFO] - Step 3243, Loss: {'train/loss': 0.7739038467407227, 'train/grad_norm_pre_clip': 11.931418418884277, 'timing/data': 0.000740660005249083, 'timing/model': 1.204981924034655, 'learning_rate/qwen_vl_interface': 5.911786103090402e-06, 'learning_rate/action_model': 5.911786103090402e-05, 'epoch': 0.95}) +[2026-06-17 12:14:00,849][starVLA.training.train_starvla][INFO] - Step 3244, Loss: {'train/loss': 0.9148104786872864, 'train/grad_norm_pre_clip': 12.1256103515625, 'timing/data': 0.0005942810093984008, 'timing/model': 1.2839222269831225, 'learning_rate/qwen_vl_interface': 5.909644436836243e-06, 'learning_rate/action_model': 5.909644436836244e-05, 'epoch': 0.95}) +[2026-06-17 12:14:05,508][starVLA.training.train_starvla][INFO] - Step 3245, Loss: {'train/loss': 1.173107385635376, 'train/grad_norm_pre_clip': 17.17082405090332, 'timing/data': 0.0003332509659230709, 'timing/model': 1.2047865419881418, 'learning_rate/qwen_vl_interface': 5.907502633837082e-06, 'learning_rate/action_model': 5.907502633837082e-05, 'epoch': 0.95}) +[2026-06-17 12:14:10,296][starVLA.training.train_starvla][INFO] - Step 3246, Loss: {'train/loss': 0.8306053876876831, 'train/grad_norm_pre_clip': 13.346446990966797, 'timing/data': 0.00035420898348093033, 'timing/model': 1.2056971499696374, 'learning_rate/qwen_vl_interface': 5.905360694536918e-06, 'learning_rate/action_model': 5.905360694536918e-05, 'epoch': 0.96}) +[2026-06-17 12:14:15,086][starVLA.training.train_starvla][INFO] - Step 3247, Loss: {'train/loss': 0.46602484583854675, 'train/grad_norm_pre_clip': 9.167570114135742, 'timing/data': 0.00036093301605433226, 'timing/model': 1.205813659937121, 'learning_rate/qwen_vl_interface': 5.903218619379776e-06, 'learning_rate/action_model': 5.903218619379776e-05, 'epoch': 0.96}) +[2026-06-17 12:14:19,872][starVLA.training.train_starvla][INFO] - Step 3248, Loss: {'train/loss': 0.5927696824073792, 'train/grad_norm_pre_clip': 10.189574241638184, 'timing/data': 0.0003495789133012295, 'timing/model': 1.2047755410894752, 'learning_rate/qwen_vl_interface': 5.901076408809712e-06, 'learning_rate/action_model': 5.9010764088097114e-05, 'epoch': 0.96}) +[2026-06-17 12:14:24,675][starVLA.training.train_starvla][INFO] - Step 3249, Loss: {'train/loss': 0.8188174962997437, 'train/grad_norm_pre_clip': 30.17349624633789, 'timing/data': 0.00038455903995782137, 'timing/model': 1.3307153730420396, 'learning_rate/qwen_vl_interface': 5.898934063270806e-06, 'learning_rate/action_model': 5.8989340632708056e-05, 'epoch': 0.96}) +[2026-06-17 12:14:29,366][starVLA.training.train_starvla][INFO] - Step 3250, Loss: {'train/loss': 0.6752297282218933, 'train/grad_norm_pre_clip': 11.043875694274902, 'timing/data': 0.0003493789117783308, 'timing/model': 1.2219399240566418, 'learning_rate/qwen_vl_interface': 5.896791583207173e-06, 'learning_rate/action_model': 5.896791583207173e-05, 'epoch': 0.96}) +[2026-06-17 12:14:34,028][starVLA.training.train_starvla][INFO] - Step 3251, Loss: {'train/loss': 0.8494430184364319, 'train/grad_norm_pre_clip': 14.031723022460938, 'timing/data': 0.0003672459861263633, 'timing/model': 1.2054879439529032, 'learning_rate/qwen_vl_interface': 5.894648969062947e-06, 'learning_rate/action_model': 5.894648969062947e-05, 'epoch': 0.96}) +[2026-06-17 12:14:38,833][starVLA.training.train_starvla][INFO] - Step 3252, Loss: {'train/loss': 0.7579703330993652, 'train/grad_norm_pre_clip': 10.257471084594727, 'timing/data': 0.000778402085416019, 'timing/model': 1.2047070150729269, 'learning_rate/qwen_vl_interface': 5.892506221282297e-06, 'learning_rate/action_model': 5.892506221282297e-05, 'epoch': 0.96}) +[2026-06-17 12:14:43,634][starVLA.training.train_starvla][INFO] - Step 3253, Loss: {'train/loss': 0.7263698577880859, 'train/grad_norm_pre_clip': 13.354063987731934, 'timing/data': 0.00040140002965927124, 'timing/model': 1.2043289070716128, 'learning_rate/qwen_vl_interface': 5.890363340309415e-06, 'learning_rate/action_model': 5.890363340309415e-05, 'epoch': 0.96}) +[2026-06-17 12:14:48,429][starVLA.training.train_starvla][INFO] - Step 3254, Loss: {'train/loss': 0.5643478035926819, 'train/grad_norm_pre_clip': 15.254463195800781, 'timing/data': 0.0003784419968724251, 'timing/model': 1.2059667209396139, 'learning_rate/qwen_vl_interface': 5.888220326588523e-06, 'learning_rate/action_model': 5.8882203265885226e-05, 'epoch': 0.96}) +[2026-06-17 12:14:53,227][starVLA.training.train_starvla][INFO] - Step 3255, Loss: {'train/loss': 0.7083624601364136, 'train/grad_norm_pre_clip': 13.194077491760254, 'timing/data': 0.0003410229692235589, 'timing/model': 1.3323900639079511, 'learning_rate/qwen_vl_interface': 5.8860771805638725e-06, 'learning_rate/action_model': 5.886077180563872e-05, 'epoch': 0.96}) +[2026-06-17 12:14:57,894][starVLA.training.train_starvla][INFO] - Step 3256, Loss: {'train/loss': 0.6170501708984375, 'train/grad_norm_pre_clip': 9.644777297973633, 'timing/data': 0.0003564979415386915, 'timing/model': 1.2049079809803516, 'learning_rate/qwen_vl_interface': 5.883933902679736e-06, 'learning_rate/action_model': 5.8839339026797355e-05, 'epoch': 0.96}) +[2026-06-17 12:15:02,683][starVLA.training.train_starvla][INFO] - Step 3257, Loss: {'train/loss': 0.5411805510520935, 'train/grad_norm_pre_clip': 9.444973945617676, 'timing/data': 0.0003458049613982439, 'timing/model': 1.2049627860542387, 'learning_rate/qwen_vl_interface': 5.8817904933804214e-06, 'learning_rate/action_model': 5.881790493380421e-05, 'epoch': 0.96}) +[2026-06-17 12:15:07,466][starVLA.training.train_starvla][INFO] - Step 3258, Loss: {'train/loss': 0.7186095714569092, 'train/grad_norm_pre_clip': 10.487292289733887, 'timing/data': 0.0003614220768213272, 'timing/model': 1.2035537830088288, 'learning_rate/qwen_vl_interface': 5.8796469531102565e-06, 'learning_rate/action_model': 5.879646953110256e-05, 'epoch': 0.96}) +[2026-06-17 12:15:12,281][starVLA.training.train_starvla][INFO] - Step 3259, Loss: {'train/loss': 0.7114371657371521, 'train/grad_norm_pre_clip': 8.842251777648926, 'timing/data': 0.0003421870060265064, 'timing/model': 1.202773350989446, 'learning_rate/qwen_vl_interface': 5.877503282313601e-06, 'learning_rate/action_model': 5.877503282313601e-05, 'epoch': 0.96}) +[2026-06-17 12:15:17,063][starVLA.training.train_starvla][INFO] - Step 3260, Loss: {'train/loss': 0.7746548652648926, 'train/grad_norm_pre_clip': 13.413556098937988, 'timing/data': 0.0003621089272201061, 'timing/model': 1.2034995779395103, 'learning_rate/qwen_vl_interface': 5.875359481434839e-06, 'learning_rate/action_model': 5.875359481434839e-05, 'epoch': 0.96}) +[2026-06-17 12:15:21,845][starVLA.training.train_starvla][INFO] - Step 3261, Loss: {'train/loss': 0.565462052822113, 'train/grad_norm_pre_clip': 8.55056381225586, 'timing/data': 0.000364540028385818, 'timing/model': 1.3309228740399703, 'learning_rate/qwen_vl_interface': 5.873215550918384e-06, 'learning_rate/action_model': 5.873215550918384e-05, 'epoch': 0.96}) +[2026-06-17 12:15:26,504][starVLA.training.train_starvla][INFO] - Step 3262, Loss: {'train/loss': 0.44559067487716675, 'train/grad_norm_pre_clip': 8.457784652709961, 'timing/data': 0.0003480679588392377, 'timing/model': 1.202400748967193, 'learning_rate/qwen_vl_interface': 5.871071491208676e-06, 'learning_rate/action_model': 5.871071491208676e-05, 'epoch': 0.96}) +[2026-06-17 12:15:31,288][starVLA.training.train_starvla][INFO] - Step 3263, Loss: {'train/loss': 0.8927838206291199, 'train/grad_norm_pre_clip': 32.87765121459961, 'timing/data': 0.0003656800836324692, 'timing/model': 1.209003823925741, 'learning_rate/qwen_vl_interface': 5.868927302750178e-06, 'learning_rate/action_model': 5.868927302750178e-05, 'epoch': 0.96}) +[2026-06-17 12:15:36,073][starVLA.training.train_starvla][INFO] - Step 3264, Loss: {'train/loss': 0.5146842002868652, 'train/grad_norm_pre_clip': 9.720440864562988, 'timing/data': 0.0004166950238868594, 'timing/model': 1.20348492404446, 'learning_rate/qwen_vl_interface': 5.866782985987386e-06, 'learning_rate/action_model': 5.866782985987386e-05, 'epoch': 0.96}) +[2026-06-17 12:15:40,734][starVLA.training.train_starvla][INFO] - Step 3265, Loss: {'train/loss': 0.7088406085968018, 'train/grad_norm_pre_clip': 11.565096855163574, 'timing/data': 0.00037863897159695625, 'timing/model': 1.2016454690601677, 'learning_rate/qwen_vl_interface': 5.864638541364815e-06, 'learning_rate/action_model': 5.8646385413648154e-05, 'epoch': 0.96}) +[2026-06-17 12:15:45,395][starVLA.training.train_starvla][INFO] - Step 3266, Loss: {'train/loss': 0.8564361333847046, 'train/grad_norm_pre_clip': 14.359159469604492, 'timing/data': 0.0003487049834802747, 'timing/model': 1.2037482649320737, 'learning_rate/qwen_vl_interface': 5.862493969327015e-06, 'learning_rate/action_model': 5.862493969327014e-05, 'epoch': 0.96}) +[2026-06-17 12:15:50,075][starVLA.training.train_starvla][INFO] - Step 3267, Loss: {'train/loss': 0.6421438455581665, 'train/grad_norm_pre_clip': 11.812172889709473, 'timing/data': 0.00031977100297808647, 'timing/model': 1.2311224390286952, 'learning_rate/qwen_vl_interface': 5.860349270318555e-06, 'learning_rate/action_model': 5.860349270318555e-05, 'epoch': 0.96}) +[2026-06-17 12:15:54,840][starVLA.training.train_starvla][INFO] - Step 3268, Loss: {'train/loss': 0.8081759214401245, 'train/grad_norm_pre_clip': 12.45782470703125, 'timing/data': 0.0003775750519707799, 'timing/model': 1.2072774049593136, 'learning_rate/qwen_vl_interface': 5.8582044447840346e-06, 'learning_rate/action_model': 5.858204444784034e-05, 'epoch': 0.96}) +[2026-06-17 12:15:59,572][starVLA.training.train_starvla][INFO] - Step 3269, Loss: {'train/loss': 0.7156617641448975, 'train/grad_norm_pre_clip': 11.526386260986328, 'timing/data': 0.00041750201489776373, 'timing/model': 1.2054599979892373, 'learning_rate/qwen_vl_interface': 5.8560594931680785e-06, 'learning_rate/action_model': 5.856059493168078e-05, 'epoch': 0.96}) +[2026-06-17 12:16:04,340][starVLA.training.train_starvla][INFO] - Step 3270, Loss: {'train/loss': 0.59371018409729, 'train/grad_norm_pre_clip': 10.246577262878418, 'timing/data': 0.00035733391996473074, 'timing/model': 1.2055275950115174, 'learning_rate/qwen_vl_interface': 5.853914415915337e-06, 'learning_rate/action_model': 5.853914415915337e-05, 'epoch': 0.96}) +[2026-06-17 12:16:09,142][starVLA.training.train_starvla][INFO] - Step 3271, Loss: {'train/loss': 0.5906616449356079, 'train/grad_norm_pre_clip': 9.467053413391113, 'timing/data': 0.00034178898204118013, 'timing/model': 1.2045256930869073, 'learning_rate/qwen_vl_interface': 5.8517692134704875e-06, 'learning_rate/action_model': 5.851769213470487e-05, 'epoch': 0.96}) +[2026-06-17 12:16:13,927][starVLA.training.train_starvla][INFO] - Step 3272, Loss: {'train/loss': 0.6140049695968628, 'train/grad_norm_pre_clip': 9.853988647460938, 'timing/data': 0.00035673310048878193, 'timing/model': 1.3316503970418125, 'learning_rate/qwen_vl_interface': 5.849623886278232e-06, 'learning_rate/action_model': 5.849623886278231e-05, 'epoch': 0.96}) +[2026-06-17 12:16:18,587][starVLA.training.train_starvla][INFO] - Step 3273, Loss: {'train/loss': 0.6700868606567383, 'train/grad_norm_pre_clip': 10.564329147338867, 'timing/data': 0.0003630860010161996, 'timing/model': 1.2067567310296, 'learning_rate/qwen_vl_interface': 5.8474784347833e-06, 'learning_rate/action_model': 5.8474784347833e-05, 'epoch': 0.96}) +[2026-06-17 12:16:23,377][starVLA.training.train_starvla][INFO] - Step 3274, Loss: {'train/loss': 0.8936964869499207, 'train/grad_norm_pre_clip': 11.363235473632812, 'timing/data': 0.00033772503957152367, 'timing/model': 1.2050162239465863, 'learning_rate/qwen_vl_interface': 5.845332859430446e-06, 'learning_rate/action_model': 5.845332859430446e-05, 'epoch': 0.96}) +[2026-06-17 12:16:28,130][starVLA.training.train_starvla][INFO] - Step 3275, Loss: {'train/loss': 0.7487804889678955, 'train/grad_norm_pre_clip': 12.435516357421875, 'timing/data': 0.00042303698137402534, 'timing/model': 1.2088045380078256, 'learning_rate/qwen_vl_interface': 5.843187160664449e-06, 'learning_rate/action_model': 5.8431871606644495e-05, 'epoch': 0.96}) +[2026-06-17 12:16:32,935][starVLA.training.train_starvla][INFO] - Step 3276, Loss: {'train/loss': 0.5784319639205933, 'train/grad_norm_pre_clip': 8.762179374694824, 'timing/data': 0.0003621239447966218, 'timing/model': 1.2031353689962998, 'learning_rate/qwen_vl_interface': 5.841041338930115e-06, 'learning_rate/action_model': 5.8410413389301155e-05, 'epoch': 0.96}) +[2026-06-17 12:16:37,735][starVLA.training.train_starvla][INFO] - Step 3277, Loss: {'train/loss': 0.43070662021636963, 'train/grad_norm_pre_clip': 11.180253982543945, 'timing/data': 0.0004207650199532509, 'timing/model': 1.2047242750413716, 'learning_rate/qwen_vl_interface': 5.8388953946722775e-06, 'learning_rate/action_model': 5.8388953946722766e-05, 'epoch': 0.96}) +[2026-06-17 12:16:42,529][starVLA.training.train_starvla][INFO] - Step 3278, Loss: {'train/loss': 0.818488359451294, 'train/grad_norm_pre_clip': 14.36535358428955, 'timing/data': 0.0003677400527521968, 'timing/model': 1.3364518089219928, 'learning_rate/qwen_vl_interface': 5.836749328335789e-06, 'learning_rate/action_model': 5.8367493283357885e-05, 'epoch': 0.96}) +[2026-06-17 12:16:47,191][starVLA.training.train_starvla][INFO] - Step 3279, Loss: {'train/loss': 0.676267147064209, 'train/grad_norm_pre_clip': 9.67085075378418, 'timing/data': 0.00033844891004264355, 'timing/model': 1.2076353030279279, 'learning_rate/qwen_vl_interface': 5.8346031403655345e-06, 'learning_rate/action_model': 5.8346031403655345e-05, 'epoch': 0.96}) +[2026-06-17 12:16:51,987][starVLA.training.train_starvla][INFO] - Step 3280, Loss: {'train/loss': 0.5116000771522522, 'train/grad_norm_pre_clip': 10.474946975708008, 'timing/data': 0.0003680940717458725, 'timing/model': 1.2024694429710507, 'learning_rate/qwen_vl_interface': 5.832456831206418e-06, 'learning_rate/action_model': 5.8324568312064176e-05, 'epoch': 0.97}) +[2026-06-17 12:16:56,672][starVLA.training.train_starvla][INFO] - Step 3281, Loss: {'train/loss': 0.7563061118125916, 'train/grad_norm_pre_clip': 12.011512756347656, 'timing/data': 0.00034385197795927525, 'timing/model': 1.204335756949149, 'learning_rate/qwen_vl_interface': 5.830310401303377e-06, 'learning_rate/action_model': 5.830310401303377e-05, 'epoch': 0.97}) +[2026-06-17 12:17:01,352][starVLA.training.train_starvla][INFO] - Step 3282, Loss: {'train/loss': 0.7671166062355042, 'train/grad_norm_pre_clip': 14.414068222045898, 'timing/data': 0.0003430439392104745, 'timing/model': 1.2092328020371497, 'learning_rate/qwen_vl_interface': 5.8281638511013625e-06, 'learning_rate/action_model': 5.828163851101363e-05, 'epoch': 0.97}) +[2026-06-17 12:17:06,061][starVLA.training.train_starvla][INFO] - Step 3283, Loss: {'train/loss': 1.0004043579101562, 'train/grad_norm_pre_clip': 15.371867179870605, 'timing/data': 0.0003275810740888119, 'timing/model': 1.2060924120014533, 'learning_rate/qwen_vl_interface': 5.8260171810453605e-06, 'learning_rate/action_model': 5.82601718104536e-05, 'epoch': 0.97}) +[2026-06-17 12:17:10,802][starVLA.training.train_starvla][INFO] - Step 3284, Loss: {'train/loss': 0.5458648800849915, 'train/grad_norm_pre_clip': 9.977389335632324, 'timing/data': 0.0003486310597509146, 'timing/model': 1.281069527962245, 'learning_rate/qwen_vl_interface': 5.823870391580376e-06, 'learning_rate/action_model': 5.823870391580376e-05, 'epoch': 0.97}) +[2026-06-17 12:17:15,468][starVLA.training.train_starvla][INFO] - Step 3285, Loss: {'train/loss': 0.7874426245689392, 'train/grad_norm_pre_clip': 11.36022663116455, 'timing/data': 0.000492547987960279, 'timing/model': 1.2038661170518026, 'learning_rate/qwen_vl_interface': 5.821723483151444e-06, 'learning_rate/action_model': 5.821723483151443e-05, 'epoch': 0.97}) +[2026-06-17 12:17:20,253][starVLA.training.train_starvla][INFO] - Step 3286, Loss: {'train/loss': 0.8852702379226685, 'train/grad_norm_pre_clip': 14.231852531433105, 'timing/data': 0.00037920998875051737, 'timing/model': 1.2019716179929674, 'learning_rate/qwen_vl_interface': 5.819576456203617e-06, 'learning_rate/action_model': 5.8195764562036174e-05, 'epoch': 0.97}) +[2026-06-17 12:17:25,037][starVLA.training.train_starvla][INFO] - Step 3287, Loss: {'train/loss': 0.4526733458042145, 'train/grad_norm_pre_clip': 9.262840270996094, 'timing/data': 0.0004372060066089034, 'timing/model': 1.2054154590005055, 'learning_rate/qwen_vl_interface': 5.817429311181977e-06, 'learning_rate/action_model': 5.8174293111819764e-05, 'epoch': 0.97}) +[2026-06-17 12:17:29,826][starVLA.training.train_starvla][INFO] - Step 3288, Loss: {'train/loss': 0.5071980953216553, 'train/grad_norm_pre_clip': 10.262299537658691, 'timing/data': 0.0004612869815900922, 'timing/model': 1.2059426190098748, 'learning_rate/qwen_vl_interface': 5.815282048531633e-06, 'learning_rate/action_model': 5.815282048531633e-05, 'epoch': 0.97}) +[2026-06-17 12:17:34,613][starVLA.training.train_starvla][INFO] - Step 3289, Loss: {'train/loss': 0.8763959407806396, 'train/grad_norm_pre_clip': 13.465070724487305, 'timing/data': 0.0004165800055488944, 'timing/model': 1.329253970994614, 'learning_rate/qwen_vl_interface': 5.81313466869771e-06, 'learning_rate/action_model': 5.8131346686977096e-05, 'epoch': 0.97}) +[2026-06-17 12:17:39,280][starVLA.training.train_starvla][INFO] - Step 3290, Loss: {'train/loss': 0.6222661137580872, 'train/grad_norm_pre_clip': 13.894164085388184, 'timing/data': 0.00039487704634666443, 'timing/model': 1.2126620039343834, 'learning_rate/qwen_vl_interface': 5.810987172125364e-06, 'learning_rate/action_model': 5.810987172125364e-05, 'epoch': 0.97}) +[2026-06-17 12:17:43,934][starVLA.training.train_starvla][INFO] - Step 3291, Loss: {'train/loss': 0.8640670776367188, 'train/grad_norm_pre_clip': 13.55976390838623, 'timing/data': 0.0003739760722965002, 'timing/model': 1.2019547279924154, 'learning_rate/qwen_vl_interface': 5.808839559259774e-06, 'learning_rate/action_model': 5.8088395592597746e-05, 'epoch': 0.97}) +[2026-06-17 12:17:48,771][starVLA.training.train_starvla][INFO] - Step 3292, Loss: {'train/loss': 0.6678194403648376, 'train/grad_norm_pre_clip': 10.98300552368164, 'timing/data': 0.0003935829736292362, 'timing/model': 1.2058885809965432, 'learning_rate/qwen_vl_interface': 5.806691830546143e-06, 'learning_rate/action_model': 5.8066918305461427e-05, 'epoch': 0.97}) +[2026-06-17 12:17:53,565][starVLA.training.train_starvla][INFO] - Step 3293, Loss: {'train/loss': 0.5210138559341431, 'train/grad_norm_pre_clip': 9.02403736114502, 'timing/data': 0.00036999001167714596, 'timing/model': 1.2030442350078374, 'learning_rate/qwen_vl_interface': 5.804543986429696e-06, 'learning_rate/action_model': 5.804543986429696e-05, 'epoch': 0.97}) +[2026-06-17 12:17:58,367][starVLA.training.train_starvla][INFO] - Step 3294, Loss: {'train/loss': 0.8521696329116821, 'train/grad_norm_pre_clip': 9.925177574157715, 'timing/data': 0.00040917505975812674, 'timing/model': 1.2051195439416915, 'learning_rate/qwen_vl_interface': 5.802396027355683e-06, 'learning_rate/action_model': 5.8023960273556835e-05, 'epoch': 0.97}) +[2026-06-17 12:18:03,154][starVLA.training.train_starvla][INFO] - Step 3295, Loss: {'train/loss': 0.5643004775047302, 'train/grad_norm_pre_clip': 9.212379455566406, 'timing/data': 0.0003645599354058504, 'timing/model': 1.3339088129578158, 'learning_rate/qwen_vl_interface': 5.800247953769381e-06, 'learning_rate/action_model': 5.800247953769381e-05, 'epoch': 0.97}) +[2026-06-17 12:18:07,806][starVLA.training.train_starvla][INFO] - Step 3296, Loss: {'train/loss': 0.9633946418762207, 'train/grad_norm_pre_clip': 12.038834571838379, 'timing/data': 0.00033416307996958494, 'timing/model': 1.2027857720386237, 'learning_rate/qwen_vl_interface': 5.798099766116084e-06, 'learning_rate/action_model': 5.798099766116084e-05, 'epoch': 0.97}) +[2026-06-17 12:18:12,607][starVLA.training.train_starvla][INFO] - Step 3297, Loss: {'train/loss': 0.8125283718109131, 'train/grad_norm_pre_clip': 12.07326889038086, 'timing/data': 0.00042365200351923704, 'timing/model': 1.2023538999492303, 'learning_rate/qwen_vl_interface': 5.7959514648411175e-06, 'learning_rate/action_model': 5.795951464841117e-05, 'epoch': 0.97}) +[2026-06-17 12:18:17,377][starVLA.training.train_starvla][INFO] - Step 3298, Loss: {'train/loss': 0.48052436113357544, 'train/grad_norm_pre_clip': 8.787627220153809, 'timing/data': 0.0003942749463021755, 'timing/model': 1.2024564309976995, 'learning_rate/qwen_vl_interface': 5.793803050389824e-06, 'learning_rate/action_model': 5.793803050389824e-05, 'epoch': 0.97}) +[2026-06-17 12:18:22,166][starVLA.training.train_starvla][INFO] - Step 3299, Loss: {'train/loss': 0.708110511302948, 'train/grad_norm_pre_clip': 10.770068168640137, 'timing/data': 0.000346221961081028, 'timing/model': 1.2054280210286379, 'learning_rate/qwen_vl_interface': 5.7916545232075725e-06, 'learning_rate/action_model': 5.791654523207573e-05, 'epoch': 0.97}) +[2026-06-17 12:20:04,013][starVLA.training.train_starvla][INFO] - Step 3300, Loss: {'train/loss': 0.7192428112030029, 'train/grad_norm_pre_clip': 12.263264656066895, 'eval/loss': 0.7669571042060852, 'eval/latency_0/loss': 0.7669571042060852, 'eval/demon_fire_precision': 0.9105657237936772, 'eval/demon_fire_recall': 0.9090531561461794, 'eval/demon_fire_f1': 0.9098088113050705, 'eval/demon_fire_rate_model': 0.6685205784204672, 'eval/demon_fire_rate_teacher': 0.6696329254727474, 'eval/demon_left_precision': 0.8170731707317073, 'eval/demon_left_recall': 0.7971671388101983, 'eval/demon_left_f1': 0.8069974189848007, 'eval/demon_left_rate_model': 0.4788654060066741, 'eval/demon_left_rate_teacher': 0.49082313681868744, 'eval/demon_right_precision': 0.7991430101767542, 'eval/demon_right_recall': 0.8429378531073446, 'eval/demon_right_f1': 0.8204564201264778, 'eval/demon_right_rate_model': 0.5191879866518354, 'eval/demon_right_rate_teacher': 0.4922135706340378, 'eval/demon_move_f1': 0.8137269195556392, 'eval/demon_cc_f1': 0.8786833003798931, 'eval/latency_0/demon_fire_precision': 0.9105657237936772, 'eval/latency_0/demon_fire_recall': 0.9090531561461794, 'eval/latency_0/demon_fire_f1': 0.9098088113050705, 'eval/latency_0/demon_fire_rate_model': 0.6685205784204672, 'eval/latency_0/demon_fire_rate_teacher': 0.6696329254727474, 'eval/latency_0/demon_left_precision': 0.8170731707317073, 'eval/latency_0/demon_left_recall': 0.7971671388101983, 'eval/latency_0/demon_left_f1': 0.8069974189848007, 'eval/latency_0/demon_left_rate_model': 0.4788654060066741, 'eval/latency_0/demon_left_rate_teacher': 0.49082313681868744, 'eval/latency_0/demon_right_precision': 0.7991430101767542, 'eval/latency_0/demon_right_recall': 0.8429378531073446, 'eval/latency_0/demon_right_f1': 0.8204564201264778, 'eval/latency_0/demon_right_rate_model': 0.5191879866518354, 'eval/latency_0/demon_right_rate_teacher': 0.4922135706340378, 'eval/latency_0/demon_move_f1': 0.8137269195556392, 'eval/latency_0/demon_cc_f1': 0.8786833003798931, 'timing/data': 0.0003550660330802202, 'timing/model': 1.201890830998309, 'learning_rate/qwen_vl_interface': 5.789505883739756e-06, 'learning_rate/action_model': 5.7895058837397554e-05, 'epoch': 0.97}) +[2026-06-17 12:20:04,013][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_3300_state +[2026-06-17 12:20:04,016][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 12:20:07,568][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_3300_state/model.safetensors +[2026-06-17 12:20:15,246][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_3300_state/optimizer.bin +[2026-06-17 12:20:15,247][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_3300_state/sampler.bin +[2026-06-17 12:20:15,247][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_3300_state/sampler_1.bin +[2026-06-17 12:20:15,248][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_3300_state/random_states_0.pkl +[2026-06-17 12:20:15,301][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3300 +[2026-06-17 12:20:15,301][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 12:20:15,338][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 12:20:20,070][starVLA.training.train_starvla][INFO] - Step 3301, Loss: {'train/loss': 0.8554363250732422, 'train/grad_norm_pre_clip': 13.989198684692383, 'timing/data': 0.0003444879548624158, 'timing/model': 1.1869563169311732, 'learning_rate/qwen_vl_interface': 5.78735713243179e-06, 'learning_rate/action_model': 5.78735713243179e-05, 'epoch': 0.97}) +[2026-06-17 12:20:24,773][starVLA.training.train_starvla][INFO] - Step 3302, Loss: {'train/loss': 0.5496039986610413, 'train/grad_norm_pre_clip': 11.416404724121094, 'timing/data': 0.0005925910081714392, 'timing/model': 1.1959974199999124, 'learning_rate/qwen_vl_interface': 5.785208269729111e-06, 'learning_rate/action_model': 5.785208269729111e-05, 'epoch': 0.97}) +[2026-06-17 12:20:29,553][starVLA.training.train_starvla][INFO] - Step 3303, Loss: {'train/loss': 0.7151992917060852, 'train/grad_norm_pre_clip': 12.425143241882324, 'timing/data': 0.0003589079715311527, 'timing/model': 1.320791956037283, 'learning_rate/qwen_vl_interface': 5.783059296077183e-06, 'learning_rate/action_model': 5.7830592960771826e-05, 'epoch': 0.97}) +[2026-06-17 12:20:34,166][starVLA.training.train_starvla][INFO] - Step 3304, Loss: {'train/loss': 0.8222670555114746, 'train/grad_norm_pre_clip': 13.750822067260742, 'timing/data': 0.00041108299046754837, 'timing/model': 1.1931299719726667, 'learning_rate/qwen_vl_interface': 5.780910211921488e-06, 'learning_rate/action_model': 5.7809102119214874e-05, 'epoch': 0.97}) +[2026-06-17 12:20:38,908][starVLA.training.train_starvla][INFO] - Step 3305, Loss: {'train/loss': 0.5320443511009216, 'train/grad_norm_pre_clip': 9.364157676696777, 'timing/data': 0.0003744519781321287, 'timing/model': 1.1934302910231054, 'learning_rate/qwen_vl_interface': 5.778761017707536e-06, 'learning_rate/action_model': 5.778761017707536e-05, 'epoch': 0.97}) +[2026-06-17 12:20:43,655][starVLA.training.train_starvla][INFO] - Step 3306, Loss: {'train/loss': 0.4471409320831299, 'train/grad_norm_pre_clip': 7.547456741333008, 'timing/data': 0.0004083069507032633, 'timing/model': 1.1960490370402113, 'learning_rate/qwen_vl_interface': 5.7766117138808545e-06, 'learning_rate/action_model': 5.776611713880854e-05, 'epoch': 0.97}) +[2026-06-17 12:20:48,285][starVLA.training.train_starvla][INFO] - Step 3307, Loss: {'train/loss': 0.5882283449172974, 'train/grad_norm_pre_clip': 11.051989555358887, 'timing/data': 0.0003563300706446171, 'timing/model': 1.1963148150825873, 'learning_rate/qwen_vl_interface': 5.774462300886999e-06, 'learning_rate/action_model': 5.774462300886999e-05, 'epoch': 0.97}) +[2026-06-17 12:20:53,039][starVLA.training.train_starvla][INFO] - Step 3308, Loss: {'train/loss': 0.6713151931762695, 'train/grad_norm_pre_clip': 11.549903869628906, 'timing/data': 0.00037838402204215527, 'timing/model': 1.2012539779534563, 'learning_rate/qwen_vl_interface': 5.772312779171545e-06, 'learning_rate/action_model': 5.7723127791715445e-05, 'epoch': 0.97}) +[2026-06-17 12:20:57,812][starVLA.training.train_starvla][INFO] - Step 3309, Loss: {'train/loss': 0.9584289789199829, 'train/grad_norm_pre_clip': 17.346141815185547, 'timing/data': 0.0003614579327404499, 'timing/model': 1.3379774659406394, 'learning_rate/qwen_vl_interface': 5.770163149180089e-06, 'learning_rate/action_model': 5.770163149180089e-05, 'epoch': 0.97}) +[2026-06-17 12:21:02,447][starVLA.training.train_starvla][INFO] - Step 3310, Loss: {'train/loss': 0.4901508092880249, 'train/grad_norm_pre_clip': 9.589853286743164, 'timing/data': 0.0004928340204060078, 'timing/model': 1.2008843419607729, 'learning_rate/qwen_vl_interface': 5.768013411358252e-06, 'learning_rate/action_model': 5.768013411358252e-05, 'epoch': 0.97}) +[2026-06-17 12:21:07,232][starVLA.training.train_starvla][INFO] - Step 3311, Loss: {'train/loss': 0.5401151180267334, 'train/grad_norm_pre_clip': 9.612783432006836, 'timing/data': 0.0003744229907169938, 'timing/model': 1.2000128449872136, 'learning_rate/qwen_vl_interface': 5.765863566151677e-06, 'learning_rate/action_model': 5.7658635661516767e-05, 'epoch': 0.97}) +[2026-06-17 12:21:11,986][starVLA.training.train_starvla][INFO] - Step 3312, Loss: {'train/loss': 0.714330792427063, 'train/grad_norm_pre_clip': 13.094097137451172, 'timing/data': 0.00037097500171512365, 'timing/model': 1.2006576829589903, 'learning_rate/qwen_vl_interface': 5.763713614006033e-06, 'learning_rate/action_model': 5.763713614006032e-05, 'epoch': 0.97}) +[2026-06-17 12:21:16,745][starVLA.training.train_starvla][INFO] - Step 3313, Loss: {'train/loss': 0.6525357365608215, 'train/grad_norm_pre_clip': 11.017051696777344, 'timing/data': 0.00041512493044137955, 'timing/model': 1.2064110819483176, 'learning_rate/qwen_vl_interface': 5.761563555367e-06, 'learning_rate/action_model': 5.761563555367e-05, 'epoch': 0.97}) +[2026-06-17 12:21:21,513][starVLA.training.train_starvla][INFO] - Step 3314, Loss: {'train/loss': 0.8110658526420593, 'train/grad_norm_pre_clip': 18.549449920654297, 'timing/data': 0.0005992850055918097, 'timing/model': 1.2000392259797081, 'learning_rate/qwen_vl_interface': 5.759413390680293e-06, 'learning_rate/action_model': 5.759413390680293e-05, 'epoch': 0.98}) +[2026-06-17 12:21:26,277][starVLA.training.train_starvla][INFO] - Step 3315, Loss: {'train/loss': 0.7456207871437073, 'train/grad_norm_pre_clip': 11.94756031036377, 'timing/data': 0.00035242398735135794, 'timing/model': 1.3292987139429897, 'learning_rate/qwen_vl_interface': 5.757263120391643e-06, 'learning_rate/action_model': 5.757263120391643e-05, 'epoch': 0.98}) +[2026-06-17 12:21:30,915][starVLA.training.train_starvla][INFO] - Step 3316, Loss: {'train/loss': 0.7380659580230713, 'train/grad_norm_pre_clip': 13.461525917053223, 'timing/data': 0.0003315489739179611, 'timing/model': 1.1982534839771688, 'learning_rate/qwen_vl_interface': 5.7551127449468005e-06, 'learning_rate/action_model': 5.7551127449468004e-05, 'epoch': 0.98}) +[2026-06-17 12:21:35,685][starVLA.training.train_starvla][INFO] - Step 3317, Loss: {'train/loss': 0.6989388465881348, 'train/grad_norm_pre_clip': 12.54755973815918, 'timing/data': 0.00033485901076346636, 'timing/model': 1.1998241849942133, 'learning_rate/qwen_vl_interface': 5.752962264791543e-06, 'learning_rate/action_model': 5.752962264791543e-05, 'epoch': 0.98}) +[2026-06-17 12:21:40,457][starVLA.training.train_starvla][INFO] - Step 3318, Loss: {'train/loss': 0.738604724407196, 'train/grad_norm_pre_clip': 13.052128791809082, 'timing/data': 0.0003334489883854985, 'timing/model': 1.2003136819694191, 'learning_rate/qwen_vl_interface': 5.750811680371666e-06, 'learning_rate/action_model': 5.750811680371666e-05, 'epoch': 0.98}) +[2026-06-17 12:21:45,226][starVLA.training.train_starvla][INFO] - Step 3319, Loss: {'train/loss': 0.5042658448219299, 'train/grad_norm_pre_clip': 10.04748821258545, 'timing/data': 0.00055661890655756, 'timing/model': 1.2016833430388942, 'learning_rate/qwen_vl_interface': 5.7486609921329914e-06, 'learning_rate/action_model': 5.748660992132992e-05, 'epoch': 0.98}) +[2026-06-17 12:21:50,011][starVLA.training.train_starvla][INFO] - Step 3320, Loss: {'train/loss': 0.497644305229187, 'train/grad_norm_pre_clip': 9.673954963684082, 'timing/data': 0.00033135106787085533, 'timing/model': 1.3351376490900293, 'learning_rate/qwen_vl_interface': 5.746510200521354e-06, 'learning_rate/action_model': 5.746510200521355e-05, 'epoch': 0.98}) +[2026-06-17 12:21:54,750][starVLA.training.train_starvla][INFO] - Step 3321, Loss: {'train/loss': 0.8214133381843567, 'train/grad_norm_pre_clip': 13.470165252685547, 'timing/data': 0.0003541379701346159, 'timing/model': 1.2849964270135388, 'learning_rate/qwen_vl_interface': 5.7443593059826206e-06, 'learning_rate/action_model': 5.744359305982621e-05, 'epoch': 0.98}) +[2026-06-17 12:21:59,400][starVLA.training.train_starvla][INFO] - Step 3322, Loss: {'train/loss': 0.9148396253585815, 'train/grad_norm_pre_clip': 14.168270111083984, 'timing/data': 0.00036066805478185415, 'timing/model': 1.2040142989717424, 'learning_rate/qwen_vl_interface': 5.742208308962669e-06, 'learning_rate/action_model': 5.742208308962669e-05, 'epoch': 0.98}) +[2026-06-17 12:22:04,177][starVLA.training.train_starvla][INFO] - Step 3323, Loss: {'train/loss': 0.678943395614624, 'train/grad_norm_pre_clip': 15.345145225524902, 'timing/data': 0.0004050110001116991, 'timing/model': 1.2023696179967374, 'learning_rate/qwen_vl_interface': 5.740057209907405e-06, 'learning_rate/action_model': 5.7400572099074046e-05, 'epoch': 0.98}) +[2026-06-17 12:22:08,961][starVLA.training.train_starvla][INFO] - Step 3324, Loss: {'train/loss': 0.4760480225086212, 'train/grad_norm_pre_clip': 32.91904830932617, 'timing/data': 0.00039744796231389046, 'timing/model': 1.2054594659712166, 'learning_rate/qwen_vl_interface': 5.7379060092627545e-06, 'learning_rate/action_model': 5.7379060092627544e-05, 'epoch': 0.98}) +[2026-06-17 12:22:13,744][starVLA.training.train_starvla][INFO] - Step 3325, Loss: {'train/loss': 0.9755544066429138, 'train/grad_norm_pre_clip': 16.9791316986084, 'timing/data': 0.00035174994263798, 'timing/model': 1.2017939350334927, 'learning_rate/qwen_vl_interface': 5.735754707474663e-06, 'learning_rate/action_model': 5.735754707474663e-05, 'epoch': 0.98}) +[2026-06-17 12:22:18,525][starVLA.training.train_starvla][INFO] - Step 3326, Loss: {'train/loss': 0.38196080923080444, 'train/grad_norm_pre_clip': 8.870768547058105, 'timing/data': 0.0006226650439202785, 'timing/model': 1.3388529539806768, 'learning_rate/qwen_vl_interface': 5.7336033049891e-06, 'learning_rate/action_model': 5.733603304989099e-05, 'epoch': 0.98}) +[2026-06-17 12:22:23,177][starVLA.training.train_starvla][INFO] - Step 3327, Loss: {'train/loss': 0.9419385194778442, 'train/grad_norm_pre_clip': 11.83112621307373, 'timing/data': 0.0003850160865113139, 'timing/model': 1.2111255240160972, 'learning_rate/qwen_vl_interface': 5.731451802252049e-06, 'learning_rate/action_model': 5.731451802252048e-05, 'epoch': 0.98}) +[2026-06-17 12:22:27,829][starVLA.training.train_starvla][INFO] - Step 3328, Loss: {'train/loss': 0.7426618337631226, 'train/grad_norm_pre_clip': 14.811423301696777, 'timing/data': 0.00042304699309170246, 'timing/model': 1.2076297020539641, 'learning_rate/qwen_vl_interface': 5.729300199709522e-06, 'learning_rate/action_model': 5.729300199709522e-05, 'epoch': 0.98}) +[2026-06-17 12:22:32,611][starVLA.training.train_starvla][INFO] - Step 3329, Loss: {'train/loss': 0.6706404685974121, 'train/grad_norm_pre_clip': 14.285553932189941, 'timing/data': 0.0006072618998587132, 'timing/model': 1.2031251389998943, 'learning_rate/qwen_vl_interface': 5.7271484978075475e-06, 'learning_rate/action_model': 5.7271484978075475e-05, 'epoch': 0.98}) +[2026-06-17 12:22:37,389][starVLA.training.train_starvla][INFO] - Step 3330, Loss: {'train/loss': 0.7050690054893494, 'train/grad_norm_pre_clip': 11.888222694396973, 'timing/data': 0.00037430610973387957, 'timing/model': 1.2018451699987054, 'learning_rate/qwen_vl_interface': 5.724996696992176e-06, 'learning_rate/action_model': 5.7249966969921764e-05, 'epoch': 0.98}) +[2026-06-17 12:22:42,176][starVLA.training.train_starvla][INFO] - Step 3331, Loss: {'train/loss': 0.7653113603591919, 'train/grad_norm_pre_clip': 16.718873977661133, 'timing/data': 0.0005312019493430853, 'timing/model': 1.2019269629381597, 'learning_rate/qwen_vl_interface': 5.72284479770948e-06, 'learning_rate/action_model': 5.72284479770948e-05, 'epoch': 0.98}) +[2026-06-17 12:22:46,955][starVLA.training.train_starvla][INFO] - Step 3332, Loss: {'train/loss': 0.709756076335907, 'train/grad_norm_pre_clip': 13.669498443603516, 'timing/data': 0.00037165696267038584, 'timing/model': 1.3320625299820676, 'learning_rate/qwen_vl_interface': 5.720692800405547e-06, 'learning_rate/action_model': 5.720692800405547e-05, 'epoch': 0.98}) +[2026-06-17 12:22:51,614][starVLA.training.train_starvla][INFO] - Step 3333, Loss: {'train/loss': 0.5132469534873962, 'train/grad_norm_pre_clip': 12.71688175201416, 'timing/data': 0.0003637300105765462, 'timing/model': 1.2050243730191141, 'learning_rate/qwen_vl_interface': 5.71854070552649e-06, 'learning_rate/action_model': 5.7185407055264904e-05, 'epoch': 0.98}) +[2026-06-17 12:22:56,400][starVLA.training.train_starvla][INFO] - Step 3334, Loss: {'train/loss': 0.954881489276886, 'train/grad_norm_pre_clip': 15.201308250427246, 'timing/data': 0.00038417603354901075, 'timing/model': 1.204006621032022, 'learning_rate/qwen_vl_interface': 5.71638851351844e-06, 'learning_rate/action_model': 5.71638851351844e-05, 'epoch': 0.98}) +[2026-06-17 12:23:01,177][starVLA.training.train_starvla][INFO] - Step 3335, Loss: {'train/loss': 0.7900591492652893, 'train/grad_norm_pre_clip': 14.271113395690918, 'timing/data': 0.0003536790609359741, 'timing/model': 1.2036884679691866, 'learning_rate/qwen_vl_interface': 5.714236224827551e-06, 'learning_rate/action_model': 5.714236224827551e-05, 'epoch': 0.98}) +[2026-06-17 12:23:05,909][starVLA.training.train_starvla][INFO] - Step 3336, Loss: {'train/loss': 0.6680815815925598, 'train/grad_norm_pre_clip': 11.016739845275879, 'timing/data': 0.00036313198506832123, 'timing/model': 1.2041690120240673, 'learning_rate/qwen_vl_interface': 5.712083839899991e-06, 'learning_rate/action_model': 5.712083839899991e-05, 'epoch': 0.98}) +[2026-06-17 12:23:10,707][starVLA.training.train_starvla][INFO] - Step 3337, Loss: {'train/loss': 0.47257429361343384, 'train/grad_norm_pre_clip': 11.853652000427246, 'timing/data': 0.0003728159936144948, 'timing/model': 1.2048355450388044, 'learning_rate/qwen_vl_interface': 5.709931359181956e-06, 'learning_rate/action_model': 5.709931359181956e-05, 'epoch': 0.98}) +[2026-06-17 12:23:15,521][starVLA.training.train_starvla][INFO] - Step 3338, Loss: {'train/loss': 0.6406657695770264, 'train/grad_norm_pre_clip': 11.74427318572998, 'timing/data': 0.0003617479233071208, 'timing/model': 1.3343434589914978, 'learning_rate/qwen_vl_interface': 5.707778783119654e-06, 'learning_rate/action_model': 5.707778783119654e-05, 'epoch': 0.98}) +[2026-06-17 12:23:20,171][starVLA.training.train_starvla][INFO] - Step 3339, Loss: {'train/loss': 0.5784783959388733, 'train/grad_norm_pre_clip': 11.259682655334473, 'timing/data': 0.00038735195994377136, 'timing/model': 1.2081855869619176, 'learning_rate/qwen_vl_interface': 5.705626112159318e-06, 'learning_rate/action_model': 5.7056261121593177e-05, 'epoch': 0.98}) +[2026-06-17 12:23:24,957][starVLA.training.train_starvla][INFO] - Step 3340, Loss: {'train/loss': 0.7161381840705872, 'train/grad_norm_pre_clip': 11.940274238586426, 'timing/data': 0.00041487300768494606, 'timing/model': 1.2049067120533437, 'learning_rate/qwen_vl_interface': 5.703473346747199e-06, 'learning_rate/action_model': 5.703473346747199e-05, 'epoch': 0.98}) +[2026-06-17 12:23:29,749][starVLA.training.train_starvla][INFO] - Step 3341, Loss: {'train/loss': 0.9578917026519775, 'train/grad_norm_pre_clip': 14.69446086883545, 'timing/data': 0.0003628380363807082, 'timing/model': 1.2073805580148473, 'learning_rate/qwen_vl_interface': 5.701320487329567e-06, 'learning_rate/action_model': 5.701320487329567e-05, 'epoch': 0.98}) +[2026-06-17 12:23:34,543][starVLA.training.train_starvla][INFO] - Step 3342, Loss: {'train/loss': 0.7503418326377869, 'train/grad_norm_pre_clip': 11.669909477233887, 'timing/data': 0.00035165599547326565, 'timing/model': 1.219809511094354, 'learning_rate/qwen_vl_interface': 5.699167534352712e-06, 'learning_rate/action_model': 5.699167534352712e-05, 'epoch': 0.98}) +[2026-06-17 12:23:39,335][starVLA.training.train_starvla][INFO] - Step 3343, Loss: {'train/loss': 0.726100742816925, 'train/grad_norm_pre_clip': 10.983298301696777, 'timing/data': 0.0005297050811350346, 'timing/model': 1.2057822369970381, 'learning_rate/qwen_vl_interface': 5.697014488262945e-06, 'learning_rate/action_model': 5.697014488262945e-05, 'epoch': 0.98}) +[2026-06-17 12:23:44,129][starVLA.training.train_starvla][INFO] - Step 3344, Loss: {'train/loss': 0.6420305371284485, 'train/grad_norm_pre_clip': 9.410589218139648, 'timing/data': 0.00035401596687734127, 'timing/model': 1.344721846980974, 'learning_rate/qwen_vl_interface': 5.694861349506591e-06, 'learning_rate/action_model': 5.6948613495065903e-05, 'epoch': 0.98}) +[2026-06-17 12:23:48,797][starVLA.training.train_starvla][INFO] - Step 3345, Loss: {'train/loss': 0.9520112872123718, 'train/grad_norm_pre_clip': 15.175243377685547, 'timing/data': 0.0008025349816307425, 'timing/model': 1.2074069739319384, 'learning_rate/qwen_vl_interface': 5.692708118530003e-06, 'learning_rate/action_model': 5.6927081185300025e-05, 'epoch': 0.98}) +[2026-06-17 12:23:53,595][starVLA.training.train_starvla][INFO] - Step 3346, Loss: {'train/loss': 0.9063496589660645, 'train/grad_norm_pre_clip': 15.309955596923828, 'timing/data': 0.0003629629500210285, 'timing/model': 1.2034963140031323, 'learning_rate/qwen_vl_interface': 5.6905547957795446e-06, 'learning_rate/action_model': 5.690554795779545e-05, 'epoch': 0.98}) +[2026-06-17 12:23:58,463][starVLA.training.train_starvla][INFO] - Step 3347, Loss: {'train/loss': 0.7006727457046509, 'train/grad_norm_pre_clip': 13.035147666931152, 'timing/data': 0.0004151280736550689, 'timing/model': 1.2031265730038285, 'learning_rate/qwen_vl_interface': 5.688401381701604e-06, 'learning_rate/action_model': 5.688401381701603e-05, 'epoch': 0.98}) +[2026-06-17 12:24:03,239][starVLA.training.train_starvla][INFO] - Step 3348, Loss: {'train/loss': 0.6353776454925537, 'train/grad_norm_pre_clip': 13.364053726196289, 'timing/data': 0.0004393829731270671, 'timing/model': 1.2040304690599442, 'learning_rate/qwen_vl_interface': 5.686247876742583e-06, 'learning_rate/action_model': 5.686247876742583e-05, 'epoch': 0.99}) +[2026-06-17 12:24:08,024][starVLA.training.train_starvla][INFO] - Step 3349, Loss: {'train/loss': 0.6946184635162354, 'train/grad_norm_pre_clip': 14.311898231506348, 'timing/data': 0.00038266193587332964, 'timing/model': 1.3307925240369514, 'learning_rate/qwen_vl_interface': 5.684094281348908e-06, 'learning_rate/action_model': 5.684094281348908e-05, 'epoch': 0.99}) +[2026-06-17 12:24:12,760][starVLA.training.train_starvla][INFO] - Step 3350, Loss: {'train/loss': 1.0321760177612305, 'train/grad_norm_pre_clip': 15.303980827331543, 'timing/data': 0.0003593600122258067, 'timing/model': 1.282698614988476, 'learning_rate/qwen_vl_interface': 5.681940595967023e-06, 'learning_rate/action_model': 5.681940595967024e-05, 'epoch': 0.99}) +[2026-06-17 12:24:17,414][starVLA.training.train_starvla][INFO] - Step 3351, Loss: {'train/loss': 0.7808831334114075, 'train/grad_norm_pre_clip': 12.84298038482666, 'timing/data': 0.00041357206646353006, 'timing/model': 1.2125992719084024, 'learning_rate/qwen_vl_interface': 5.679786821043388e-06, 'learning_rate/action_model': 5.679786821043388e-05, 'epoch': 0.99}) +[2026-06-17 12:24:22,198][starVLA.training.train_starvla][INFO] - Step 3352, Loss: {'train/loss': 0.6877648830413818, 'train/grad_norm_pre_clip': 27.778461456298828, 'timing/data': 0.0003721880493685603, 'timing/model': 1.20523632701952, 'learning_rate/qwen_vl_interface': 5.677632957024482e-06, 'learning_rate/action_model': 5.6776329570244826e-05, 'epoch': 0.99}) +[2026-06-17 12:24:27,006][starVLA.training.train_starvla][INFO] - Step 3353, Loss: {'train/loss': 0.8566398024559021, 'train/grad_norm_pre_clip': 13.730195045471191, 'timing/data': 0.00036669394467025995, 'timing/model': 1.2037191530689597, 'learning_rate/qwen_vl_interface': 5.675479004356803e-06, 'learning_rate/action_model': 5.675479004356803e-05, 'epoch': 0.99}) +[2026-06-17 12:24:31,799][starVLA.training.train_starvla][INFO] - Step 3354, Loss: {'train/loss': 0.6981362700462341, 'train/grad_norm_pre_clip': 16.027891159057617, 'timing/data': 0.00038241303991526365, 'timing/model': 1.2032130520092323, 'learning_rate/qwen_vl_interface': 5.673324963486871e-06, 'learning_rate/action_model': 5.67332496348687e-05, 'epoch': 0.99}) +[2026-06-17 12:24:36,605][starVLA.training.train_starvla][INFO] - Step 3355, Loss: {'train/loss': 0.8623216152191162, 'train/grad_norm_pre_clip': 10.525714874267578, 'timing/data': 0.00034073705319315195, 'timing/model': 1.3317969270283356, 'learning_rate/qwen_vl_interface': 5.671170834861215e-06, 'learning_rate/action_model': 5.671170834861215e-05, 'epoch': 0.99}) +[2026-06-17 12:24:41,269][starVLA.training.train_starvla][INFO] - Step 3356, Loss: {'train/loss': 0.8634998798370361, 'train/grad_norm_pre_clip': 10.406078338623047, 'timing/data': 0.0003600510535761714, 'timing/model': 1.2130618609953672, 'learning_rate/qwen_vl_interface': 5.6690166189263925e-06, 'learning_rate/action_model': 5.669016618926393e-05, 'epoch': 0.99}) +[2026-06-17 12:24:45,921][starVLA.training.train_starvla][INFO] - Step 3357, Loss: {'train/loss': 0.5924990177154541, 'train/grad_norm_pre_clip': 10.799995422363281, 'timing/data': 0.00036755006294697523, 'timing/model': 1.2047919110627845, 'learning_rate/qwen_vl_interface': 5.666862316128977e-06, 'learning_rate/action_model': 5.666862316128977e-05, 'epoch': 0.99}) +[2026-06-17 12:24:50,701][starVLA.training.train_starvla][INFO] - Step 3358, Loss: {'train/loss': 0.6453645825386047, 'train/grad_norm_pre_clip': 12.278437614440918, 'timing/data': 0.0003488529473543167, 'timing/model': 1.2038100080098957, 'learning_rate/qwen_vl_interface': 5.664707926915551e-06, 'learning_rate/action_model': 5.66470792691555e-05, 'epoch': 0.99}) +[2026-06-17 12:24:55,488][starVLA.training.train_starvla][INFO] - Step 3359, Loss: {'train/loss': 0.7352993488311768, 'train/grad_norm_pre_clip': 14.038228034973145, 'timing/data': 0.00033460499253124, 'timing/model': 1.2048986469162628, 'learning_rate/qwen_vl_interface': 5.662553451732729e-06, 'learning_rate/action_model': 5.6625534517327285e-05, 'epoch': 0.99}) +[2026-06-17 12:25:00,272][starVLA.training.train_starvla][INFO] - Step 3360, Loss: {'train/loss': 0.7423374652862549, 'train/grad_norm_pre_clip': 12.418298721313477, 'timing/data': 0.00034064799547195435, 'timing/model': 1.2053841640008613, 'learning_rate/qwen_vl_interface': 5.660398891027129e-06, 'learning_rate/action_model': 5.660398891027129e-05, 'epoch': 0.99}) +[2026-06-17 12:25:05,094][starVLA.training.train_starvla][INFO] - Step 3361, Loss: {'train/loss': 0.9156611561775208, 'train/grad_norm_pre_clip': 13.31618881225586, 'timing/data': 0.0003710889723151922, 'timing/model': 1.330998943070881, 'learning_rate/qwen_vl_interface': 5.658244245245399e-06, 'learning_rate/action_model': 5.658244245245399e-05, 'epoch': 0.99}) +[2026-06-17 12:25:09,752][starVLA.training.train_starvla][INFO] - Step 3362, Loss: {'train/loss': 0.8330147862434387, 'train/grad_norm_pre_clip': 16.12164306640625, 'timing/data': 0.00035718700382858515, 'timing/model': 1.2123942220350727, 'learning_rate/qwen_vl_interface': 5.6560895148341985e-06, 'learning_rate/action_model': 5.656089514834199e-05, 'epoch': 0.99}) +[2026-06-17 12:25:14,530][starVLA.training.train_starvla][INFO] - Step 3363, Loss: {'train/loss': 0.6958189606666565, 'train/grad_norm_pre_clip': 11.095303535461426, 'timing/data': 0.0003530799876898527, 'timing/model': 1.2046901729190722, 'learning_rate/qwen_vl_interface': 5.653934700240204e-06, 'learning_rate/action_model': 5.6539347002402034e-05, 'epoch': 0.99}) +[2026-06-17 12:25:19,308][starVLA.training.train_starvla][INFO] - Step 3364, Loss: {'train/loss': 0.6149501204490662, 'train/grad_norm_pre_clip': 10.647151947021484, 'timing/data': 0.00034369598142802715, 'timing/model': 1.2045419289497659, 'learning_rate/qwen_vl_interface': 5.651779801910111e-06, 'learning_rate/action_model': 5.651779801910111e-05, 'epoch': 0.99}) +[2026-06-17 12:25:24,046][starVLA.training.train_starvla][INFO] - Step 3365, Loss: {'train/loss': 0.6521214246749878, 'train/grad_norm_pre_clip': 13.689624786376953, 'timing/data': 0.0003849719651043415, 'timing/model': 1.2060439620399848, 'learning_rate/qwen_vl_interface': 5.649624820290631e-06, 'learning_rate/action_model': 5.649624820290631e-05, 'epoch': 0.99}) +[2026-06-17 12:25:28,900][starVLA.training.train_starvla][INFO] - Step 3366, Loss: {'train/loss': 0.8295671939849854, 'train/grad_norm_pre_clip': 12.859728813171387, 'timing/data': 0.00035830901470035315, 'timing/model': 1.206554069998674, 'learning_rate/qwen_vl_interface': 5.647469755828501e-06, 'learning_rate/action_model': 5.647469755828501e-05, 'epoch': 0.99}) +[2026-06-17 12:25:33,693][starVLA.training.train_starvla][INFO] - Step 3367, Loss: {'train/loss': 0.9330736398696899, 'train/grad_norm_pre_clip': 13.466018676757812, 'timing/data': 0.00036974402610212564, 'timing/model': 1.3357475069351494, 'learning_rate/qwen_vl_interface': 5.645314608970457e-06, 'learning_rate/action_model': 5.6453146089704566e-05, 'epoch': 0.99}) +[2026-06-17 12:25:38,391][starVLA.training.train_starvla][INFO] - Step 3368, Loss: {'train/loss': 0.8337675333023071, 'train/grad_norm_pre_clip': 13.045804977416992, 'timing/data': 0.0003835629904642701, 'timing/model': 1.2048474790062755, 'learning_rate/qwen_vl_interface': 5.643159380163273e-06, 'learning_rate/action_model': 5.643159380163273e-05, 'epoch': 0.99}) +[2026-06-17 12:25:43,177][starVLA.training.train_starvla][INFO] - Step 3369, Loss: {'train/loss': 0.7021172046661377, 'train/grad_norm_pre_clip': 24.6919002532959, 'timing/data': 0.00036350300069898367, 'timing/model': 1.2059550659032539, 'learning_rate/qwen_vl_interface': 5.641004069853725e-06, 'learning_rate/action_model': 5.641004069853725e-05, 'epoch': 0.99}) +[2026-06-17 12:25:47,963][starVLA.training.train_starvla][INFO] - Step 3370, Loss: {'train/loss': 0.552568256855011, 'train/grad_norm_pre_clip': 13.63740062713623, 'timing/data': 0.0005975989624857903, 'timing/model': 1.2042685049818829, 'learning_rate/qwen_vl_interface': 5.638848678488613e-06, 'learning_rate/action_model': 5.638848678488613e-05, 'epoch': 0.99}) +[2026-06-17 12:25:52,744][starVLA.training.train_starvla][INFO] - Step 3371, Loss: {'train/loss': 0.7656356692314148, 'train/grad_norm_pre_clip': 11.179506301879883, 'timing/data': 0.000709276064299047, 'timing/model': 1.203423427999951, 'learning_rate/qwen_vl_interface': 5.636693206514752e-06, 'learning_rate/action_model': 5.636693206514752e-05, 'epoch': 0.99}) +[2026-06-17 12:25:57,549][starVLA.training.train_starvla][INFO] - Step 3372, Loss: {'train/loss': 0.5914623141288757, 'train/grad_norm_pre_clip': 10.982681274414062, 'timing/data': 0.000601797946728766, 'timing/model': 1.2034010840579867, 'learning_rate/qwen_vl_interface': 5.634537654378973e-06, 'learning_rate/action_model': 5.634537654378973e-05, 'epoch': 0.99}) +[2026-06-17 12:26:02,353][starVLA.training.train_starvla][INFO] - Step 3373, Loss: {'train/loss': 0.7728757262229919, 'train/grad_norm_pre_clip': 12.190940856933594, 'timing/data': 0.00037397805135697126, 'timing/model': 1.3400223690550774, 'learning_rate/qwen_vl_interface': 5.632382022528123e-06, 'learning_rate/action_model': 5.632382022528123e-05, 'epoch': 0.99}) +[2026-06-17 12:26:07,103][starVLA.training.train_starvla][INFO] - Step 3374, Loss: {'train/loss': 0.7210355401039124, 'train/grad_norm_pre_clip': 11.101133346557617, 'timing/data': 0.0003550139954313636, 'timing/model': 1.2069143690168858, 'learning_rate/qwen_vl_interface': 5.6302263114090705e-06, 'learning_rate/action_model': 5.63022631140907e-05, 'epoch': 0.99}) +[2026-06-17 12:26:11,883][starVLA.training.train_starvla][INFO] - Step 3375, Loss: {'train/loss': 0.9074554443359375, 'train/grad_norm_pre_clip': 11.883475303649902, 'timing/data': 0.00033725297544151545, 'timing/model': 1.2027549559716135, 'learning_rate/qwen_vl_interface': 5.628070521468693e-06, 'learning_rate/action_model': 5.628070521468692e-05, 'epoch': 0.99}) +[2026-06-17 12:26:16,666][starVLA.training.train_starvla][INFO] - Step 3376, Loss: {'train/loss': 0.6108081936836243, 'train/grad_norm_pre_clip': 8.924750328063965, 'timing/data': 0.00032289198134094477, 'timing/model': 1.2036632610252127, 'learning_rate/qwen_vl_interface': 5.625914653153889e-06, 'learning_rate/action_model': 5.625914653153889e-05, 'epoch': 0.99}) +[2026-06-17 12:26:21,444][starVLA.training.train_starvla][INFO] - Step 3377, Loss: {'train/loss': 0.7511540651321411, 'train/grad_norm_pre_clip': 10.704987525939941, 'timing/data': 0.00038361502811312675, 'timing/model': 1.2026675490196794, 'learning_rate/qwen_vl_interface': 5.623758706911574e-06, 'learning_rate/action_model': 5.6237587069115737e-05, 'epoch': 0.99}) +[2026-06-17 12:26:26,224][starVLA.training.train_starvla][INFO] - Step 3378, Loss: {'train/loss': 0.7926110625267029, 'train/grad_norm_pre_clip': 12.780499458312988, 'timing/data': 0.0003367280587553978, 'timing/model': 1.331123725976795, 'learning_rate/qwen_vl_interface': 5.621602683188677e-06, 'learning_rate/action_model': 5.621602683188677e-05, 'epoch': 0.99}) +[2026-06-17 12:26:30,948][starVLA.training.train_starvla][INFO] - Step 3379, Loss: {'train/loss': 0.8580238819122314, 'train/grad_norm_pre_clip': 14.444297790527344, 'timing/data': 0.00036851305048912764, 'timing/model': 1.2797439090209082, 'learning_rate/qwen_vl_interface': 5.619446582432143e-06, 'learning_rate/action_model': 5.619446582432143e-05, 'epoch': 0.99}) +[2026-06-17 12:26:35,599][starVLA.training.train_starvla][INFO] - Step 3380, Loss: {'train/loss': 1.059852957725525, 'train/grad_norm_pre_clip': 15.25652027130127, 'timing/data': 0.0003668790450319648, 'timing/model': 1.2045743290800601, 'learning_rate/qwen_vl_interface': 5.617290405088936e-06, 'learning_rate/action_model': 5.617290405088936e-05, 'epoch': 0.99}) +[2026-06-17 12:26:40,390][starVLA.training.train_starvla][INFO] - Step 3381, Loss: {'train/loss': 0.742339015007019, 'train/grad_norm_pre_clip': 10.917167663574219, 'timing/data': 0.000356359058059752, 'timing/model': 1.203475989983417, 'learning_rate/qwen_vl_interface': 5.615134151606033e-06, 'learning_rate/action_model': 5.615134151606033e-05, 'epoch': 0.99}) +[2026-06-17 12:26:45,191][starVLA.training.train_starvla][INFO] - Step 3382, Loss: {'train/loss': 0.9117545485496521, 'train/grad_norm_pre_clip': 16.37599754333496, 'timing/data': 0.0003759560640901327, 'timing/model': 1.2121132359607145, 'learning_rate/qwen_vl_interface': 5.612977822430425e-06, 'learning_rate/action_model': 5.612977822430425e-05, 'epoch': 1.0}) +[2026-06-17 12:26:50,000][starVLA.training.train_starvla][INFO] - Step 3383, Loss: {'train/loss': 0.8702130913734436, 'train/grad_norm_pre_clip': 16.383729934692383, 'timing/data': 0.00035041000228375196, 'timing/model': 1.2048154689837247, 'learning_rate/qwen_vl_interface': 5.610821418009127e-06, 'learning_rate/action_model': 5.610821418009127e-05, 'epoch': 1.0}) +[2026-06-17 12:26:54,789][starVLA.training.train_starvla][INFO] - Step 3384, Loss: {'train/loss': 1.1056334972381592, 'train/grad_norm_pre_clip': 15.591286659240723, 'timing/data': 0.0003633670276030898, 'timing/model': 1.3381225380580872, 'learning_rate/qwen_vl_interface': 5.608664938789159e-06, 'learning_rate/action_model': 5.608664938789159e-05, 'epoch': 1.0}) +[2026-06-17 12:26:59,459][starVLA.training.train_starvla][INFO] - Step 3385, Loss: {'train/loss': 0.8241905570030212, 'train/grad_norm_pre_clip': 13.119998931884766, 'timing/data': 0.0005854930495843291, 'timing/model': 1.212241108994931, 'learning_rate/qwen_vl_interface': 5.606508385217564e-06, 'learning_rate/action_model': 5.6065083852175636e-05, 'epoch': 1.0}) +[2026-06-17 12:27:04,126][starVLA.training.train_starvla][INFO] - Step 3386, Loss: {'train/loss': 0.5592207908630371, 'train/grad_norm_pre_clip': 11.654251098632812, 'timing/data': 0.00037135498132556677, 'timing/model': 1.20525740692392, 'learning_rate/qwen_vl_interface': 5.604351757741396e-06, 'learning_rate/action_model': 5.6043517577413964e-05, 'epoch': 1.0}) +[2026-06-17 12:27:08,917][starVLA.training.train_starvla][INFO] - Step 3387, Loss: {'train/loss': 0.8541948199272156, 'train/grad_norm_pre_clip': 14.476853370666504, 'timing/data': 0.0004088419955223799, 'timing/model': 1.2040437230607495, 'learning_rate/qwen_vl_interface': 5.602195056807727e-06, 'learning_rate/action_model': 5.602195056807727e-05, 'epoch': 1.0}) +[2026-06-17 12:27:13,723][starVLA.training.train_starvla][INFO] - Step 3388, Loss: {'train/loss': 0.8679063320159912, 'train/grad_norm_pre_clip': 12.999480247497559, 'timing/data': 0.0003666189732030034, 'timing/model': 1.2112815310247242, 'learning_rate/qwen_vl_interface': 5.600038282863647e-06, 'learning_rate/action_model': 5.6000382828636466e-05, 'epoch': 1.0}) +[2026-06-17 12:27:18,514][starVLA.training.train_starvla][INFO] - Step 3389, Loss: {'train/loss': 0.6640847325325012, 'train/grad_norm_pre_clip': 15.162670135498047, 'timing/data': 0.0003396200481802225, 'timing/model': 1.2042563050054014, 'learning_rate/qwen_vl_interface': 5.597881436356252e-06, 'learning_rate/action_model': 5.597881436356252e-05, 'epoch': 1.0}) +[2026-06-17 12:27:23,292][starVLA.training.train_starvla][INFO] - Step 3390, Loss: {'train/loss': 0.7543565034866333, 'train/grad_norm_pre_clip': 10.305476188659668, 'timing/data': 0.00037530006375163794, 'timing/model': 1.3312458820873871, 'learning_rate/qwen_vl_interface': 5.595724517732661e-06, 'learning_rate/action_model': 5.595724517732661e-05, 'epoch': 1.0}) +[2026-06-17 12:27:27,985][starVLA.training.train_starvla][INFO] - Step 3391, Loss: {'train/loss': 0.5452267527580261, 'train/grad_norm_pre_clip': 8.643051147460938, 'timing/data': 0.00043882301542907953, 'timing/model': 1.2052404939895496, 'learning_rate/qwen_vl_interface': 5.5935675274400046e-06, 'learning_rate/action_model': 5.593567527440005e-05, 'epoch': 1.0}) +[2026-06-17 12:27:32,770][starVLA.training.train_starvla][INFO] - Step 3392, Loss: {'train/loss': 0.6733636856079102, 'train/grad_norm_pre_clip': 9.771889686584473, 'timing/data': 0.0006293309852480888, 'timing/model': 1.2045127380406484, 'learning_rate/qwen_vl_interface': 5.591410465925431e-06, 'learning_rate/action_model': 5.5914104659254306e-05, 'epoch': 1.0}) +[2026-06-17 12:27:37,545][starVLA.training.train_starvla][INFO] - Step 3393, Loss: {'train/loss': 0.6807960867881775, 'train/grad_norm_pre_clip': 10.914663314819336, 'timing/data': 0.00038669805508106947, 'timing/model': 1.2034596360754222, 'learning_rate/qwen_vl_interface': 5.589253333636101e-06, 'learning_rate/action_model': 5.589253333636101e-05, 'epoch': 1.0}) +[2026-06-17 12:27:42,335][starVLA.training.train_starvla][INFO] - Step 3394, Loss: {'train/loss': 0.9963968396186829, 'train/grad_norm_pre_clip': 12.071040153503418, 'timing/data': 0.0005983508890494704, 'timing/model': 1.2054841839708388, 'learning_rate/qwen_vl_interface': 5.587096131019189e-06, 'learning_rate/action_model': 5.5870961310191896e-05, 'epoch': 1.0}) +[2026-06-17 12:27:47,120][starVLA.training.train_starvla][INFO] - Step 3395, Loss: {'train/loss': 0.8445783853530884, 'train/grad_norm_pre_clip': 12.167681694030762, 'timing/data': 0.00041215401142835617, 'timing/model': 1.2040150630054995, 'learning_rate/qwen_vl_interface': 5.5849388585218865e-06, 'learning_rate/action_model': 5.5849388585218866e-05, 'epoch': 1.0}) +[2026-06-17 12:27:52,011][starVLA.training.train_starvla][INFO] - Step 3396, Loss: {'train/loss': 0.6214386224746704, 'train/grad_norm_pre_clip': 9.994601249694824, 'timing/data': 0.00035172002390027046, 'timing/model': 1.3504874870413914, 'learning_rate/qwen_vl_interface': 5.582781516591398e-06, 'learning_rate/action_model': 5.5827815165913977e-05, 'epoch': 1.0}) +[2026-06-17 12:27:56,699][starVLA.training.train_starvla][INFO] - Step 3397, Loss: {'train/loss': 0.9238577485084534, 'train/grad_norm_pre_clip': 12.26815128326416, 'timing/data': 0.0005002120742574334, 'timing/model': 1.210889718029648, 'learning_rate/qwen_vl_interface': 5.580624105674945e-06, 'learning_rate/action_model': 5.580624105674944e-05, 'epoch': 1.0}) +[2026-06-17 12:28:01,483][starVLA.training.train_starvla][INFO] - Step 3398, Loss: {'train/loss': 0.730891227722168, 'train/grad_norm_pre_clip': 13.390693664550781, 'timing/data': 0.0003266769926995039, 'timing/model': 1.2041898920433596, 'learning_rate/qwen_vl_interface': 5.578466626219756e-06, 'learning_rate/action_model': 5.578466626219756e-05, 'epoch': 1.0}) +[2026-06-17 12:28:06,271][starVLA.training.train_starvla][INFO] - Step 3399, Loss: {'train/loss': 0.9940956830978394, 'train/grad_norm_pre_clip': 12.448944091796875, 'timing/data': 0.0003356561064720154, 'timing/model': 1.211264505982399, 'learning_rate/qwen_vl_interface': 5.576309078673083e-06, 'learning_rate/action_model': 5.576309078673083e-05, 'epoch': 1.0}) +[2026-06-17 12:29:48,325][starVLA.training.train_starvla][INFO] - Step 3400, Loss: {'train/loss': 0.6333645582199097, 'train/grad_norm_pre_clip': 8.844937324523926, 'eval/loss': 0.759041428565979, 'eval/latency_0/loss': 0.759041428565979, 'eval/demon_fire_precision': 0.9024887800897593, 'eval/demon_fire_recall': 0.9186046511627907, 'eval/demon_fire_f1': 0.910475406462235, 'eval/demon_fire_rate_model': 0.6815906562847609, 'eval/demon_fire_rate_teacher': 0.6696329254727474, 'eval/demon_left_precision': 0.8376741368867353, 'eval/demon_left_recall': 0.7835694050991502, 'eval/demon_left_f1': 0.8097189695550352, 'eval/demon_left_rate_model': 0.45912124582869857, 'eval/demon_left_rate_teacher': 0.49082313681868744, 'eval/demon_right_precision': 0.7958441558441558, 'eval/demon_right_recall': 0.8655367231638418, 'eval/demon_right_f1': 0.8292286874154262, 'eval/demon_right_rate_model': 0.5353170189098999, 'eval/demon_right_rate_teacher': 0.4922135706340378, 'eval/demon_move_f1': 0.8194738284852308, 'eval/demon_cc_f1': 0.8811211741076888, 'eval/latency_0/demon_fire_precision': 0.9024887800897593, 'eval/latency_0/demon_fire_recall': 0.9186046511627907, 'eval/latency_0/demon_fire_f1': 0.910475406462235, 'eval/latency_0/demon_fire_rate_model': 0.6815906562847609, 'eval/latency_0/demon_fire_rate_teacher': 0.6696329254727474, 'eval/latency_0/demon_left_precision': 0.8376741368867353, 'eval/latency_0/demon_left_recall': 0.7835694050991502, 'eval/latency_0/demon_left_f1': 0.8097189695550352, 'eval/latency_0/demon_left_rate_model': 0.45912124582869857, 'eval/latency_0/demon_left_rate_teacher': 0.49082313681868744, 'eval/latency_0/demon_right_precision': 0.7958441558441558, 'eval/latency_0/demon_right_recall': 0.8655367231638418, 'eval/latency_0/demon_right_f1': 0.8292286874154262, 'eval/latency_0/demon_right_rate_model': 0.5353170189098999, 'eval/latency_0/demon_right_rate_teacher': 0.4922135706340378, 'eval/latency_0/demon_move_f1': 0.8194738284852308, 'eval/latency_0/demon_cc_f1': 0.8811211741076888, 'timing/data': 0.004635013057850301, 'timing/model': 1.2032593119656667, 'learning_rate/qwen_vl_interface': 5.5741514634821885e-06, 'learning_rate/action_model': 5.574151463482188e-05, 'epoch': 1.0}) +[2026-06-17 12:29:48,326][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_3400_state +[2026-06-17 12:29:48,329][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 12:29:51,883][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_3400_state/model.safetensors +[2026-06-17 12:29:59,504][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_3400_state/optimizer.bin +[2026-06-17 12:29:59,504][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_3400_state/sampler.bin +[2026-06-17 12:29:59,505][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_3400_state/sampler_1.bin +[2026-06-17 12:29:59,505][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_3400_state/random_states_0.pkl +[2026-06-17 12:29:59,568][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3400 +[2026-06-17 12:29:59,568][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 12:29:59,578][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 12:30:04,291][starVLA.training.train_starvla][INFO] - Step 3401, Loss: {'train/loss': 0.7707832455635071, 'train/grad_norm_pre_clip': 9.588128089904785, 'timing/data': 0.005099131027236581, 'timing/model': 1.1886218920117244, 'learning_rate/qwen_vl_interface': 5.571993781094345e-06, 'learning_rate/action_model': 5.5719937810943444e-05, 'epoch': 1.0}) +[2026-06-17 12:30:09,014][starVLA.training.train_starvla][INFO] - Step 3402, Loss: {'train/loss': 0.9457432627677917, 'train/grad_norm_pre_clip': 22.61182975769043, 'timing/data': 0.0048183390172198415, 'timing/model': 1.19186278199777, 'learning_rate/qwen_vl_interface': 5.569836031956845e-06, 'learning_rate/action_model': 5.5698360319568454e-05, 'epoch': 1.0}) +[2026-06-17 12:30:13,745][starVLA.training.train_starvla][INFO] - Step 3403, Loss: {'train/loss': 1.0209946632385254, 'train/grad_norm_pre_clip': 14.342766761779785, 'timing/data': 0.0050063899252563715, 'timing/model': 1.1934680870035663, 'learning_rate/qwen_vl_interface': 5.5676782165169915e-06, 'learning_rate/action_model': 5.567678216516991e-05, 'epoch': 1.0}) +[2026-06-17 12:30:18,508][starVLA.training.train_starvla][INFO] - Step 3404, Loss: {'train/loss': 0.8208447694778442, 'train/grad_norm_pre_clip': 16.2681941986084, 'timing/data': 0.004744074074551463, 'timing/model': 1.3295089229941368, 'learning_rate/qwen_vl_interface': 5.565520335222101e-06, 'learning_rate/action_model': 5.565520335222101e-05, 'epoch': 1.0}) +[2026-06-17 12:30:23,126][starVLA.training.train_starvla][INFO] - Step 3405, Loss: {'train/loss': 0.6785215139389038, 'train/grad_norm_pre_clip': 10.636866569519043, 'timing/data': 0.00491318095009774, 'timing/model': 1.1950959300156683, 'learning_rate/qwen_vl_interface': 5.563362388519507e-06, 'learning_rate/action_model': 5.5633623885195066e-05, 'epoch': 1.0}) +[2026-06-17 12:30:27,880][starVLA.training.train_starvla][INFO] - Step 3406, Loss: {'train/loss': 0.9374685883522034, 'train/grad_norm_pre_clip': 11.893848419189453, 'timing/data': 0.004385584965348244, 'timing/model': 1.1946927100652829, 'learning_rate/qwen_vl_interface': 5.56120437685655e-06, 'learning_rate/action_model': 5.56120437685655e-05, 'epoch': 1.0}) +[2026-06-17 12:30:32,643][starVLA.training.train_starvla][INFO] - Step 3407, Loss: {'train/loss': 1.0450457334518433, 'train/grad_norm_pre_clip': 13.474976539611816, 'timing/data': 0.004869754076935351, 'timing/model': 1.1974481249926612, 'learning_rate/qwen_vl_interface': 5.559046300680593e-06, 'learning_rate/action_model': 5.5590463006805925e-05, 'epoch': 1.0}) +[2026-06-17 12:30:37,398][starVLA.training.train_starvla][INFO] - Step 3408, Loss: {'train/loss': 0.6315476298332214, 'train/grad_norm_pre_clip': 76.9030532836914, 'timing/data': 0.004548260010778904, 'timing/model': 1.1963420079555362, 'learning_rate/qwen_vl_interface': 5.556888160439006e-06, 'learning_rate/action_model': 5.556888160439006e-05, 'epoch': 1.0}) +[2026-06-17 12:30:42,167][starVLA.training.train_starvla][INFO] - Step 3409, Loss: {'train/loss': 0.5509851574897766, 'train/grad_norm_pre_clip': 7.730167388916016, 'timing/data': 0.004914976074360311, 'timing/model': 1.3351646839873865, 'learning_rate/qwen_vl_interface': 5.5547299565791725e-06, 'learning_rate/action_model': 5.5547299565791723e-05, 'epoch': 1.0}) +[2026-06-17 12:30:46,925][starVLA.training.train_starvla][INFO] - Step 3410, Loss: {'train/loss': 0.654579222202301, 'train/grad_norm_pre_clip': 10.474197387695312, 'timing/data': 0.004965147934854031, 'timing/model': 1.319546933984384, 'learning_rate/qwen_vl_interface': 5.55257168954849e-06, 'learning_rate/action_model': 5.5525716895484905e-05, 'epoch': 1.0}) +[2026-06-17 12:30:51,566][starVLA.training.train_starvla][INFO] - Step 3411, Loss: {'train/loss': 0.9186210036277771, 'train/grad_norm_pre_clip': 10.754436492919922, 'timing/data': 0.004621888976544142, 'timing/model': 1.2035947409458458, 'learning_rate/qwen_vl_interface': 5.550413359794371e-06, 'learning_rate/action_model': 5.550413359794371e-05, 'epoch': 1.0}) +[2026-06-17 12:30:56,274][starVLA.training.train_starvla][INFO] - Step 3412, Loss: {'train/loss': 0.8630966544151306, 'train/grad_norm_pre_clip': 16.289295196533203, 'timing/data': 0.005016318988054991, 'timing/model': 1.2014532539760694, 'learning_rate/qwen_vl_interface': 5.548254967764239e-06, 'learning_rate/action_model': 5.5482549677642394e-05, 'epoch': 1.0}) +[2026-06-17 12:31:01,014][starVLA.training.train_starvla][INFO] - Step 3413, Loss: {'train/loss': 0.6761801242828369, 'train/grad_norm_pre_clip': 10.611221313476562, 'timing/data': 0.004686763044446707, 'timing/model': 1.2054764050990343, 'learning_rate/qwen_vl_interface': 5.546096513905532e-06, 'learning_rate/action_model': 5.546096513905532e-05, 'epoch': 1.0}) +[2026-06-17 12:31:05,823][starVLA.training.train_starvla][INFO] - Step 3414, Loss: {'train/loss': 0.7540209889411926, 'train/grad_norm_pre_clip': 11.440316200256348, 'timing/data': 0.004870545933954418, 'timing/model': 1.2001157669583336, 'learning_rate/qwen_vl_interface': 5.5439379986657e-06, 'learning_rate/action_model': 5.543937998665699e-05, 'epoch': 1.0}) +[2026-06-17 12:31:10,589][starVLA.training.train_starvla][INFO] - Step 3415, Loss: {'train/loss': 0.7786045074462891, 'train/grad_norm_pre_clip': 11.835968971252441, 'timing/data': 0.004745816928334534, 'timing/model': 1.326100369100459, 'learning_rate/qwen_vl_interface': 5.5417794224922025e-06, 'learning_rate/action_model': 5.5417794224922023e-05, 'epoch': 1.0}) +[2026-06-17 12:31:15,235][starVLA.training.train_starvla][INFO] - Step 3416, Loss: {'train/loss': 0.8557186722755432, 'train/grad_norm_pre_clip': 11.976476669311523, 'timing/data': 0.004685512976720929, 'timing/model': 1.2008439099881798, 'learning_rate/qwen_vl_interface': 5.539620785832517e-06, 'learning_rate/action_model': 5.5396207858325166e-05, 'epoch': 1.01}) +[2026-06-17 12:31:20,002][starVLA.training.train_starvla][INFO] - Step 3417, Loss: {'train/loss': 0.7846137881278992, 'train/grad_norm_pre_clip': 13.39505386352539, 'timing/data': 0.0045757381012663245, 'timing/model': 1.2010786309838295, 'learning_rate/qwen_vl_interface': 5.537462089134133e-06, 'learning_rate/action_model': 5.537462089134132e-05, 'epoch': 1.01}) +[2026-06-17 12:31:24,730][starVLA.training.train_starvla][INFO] - Step 3418, Loss: {'train/loss': 0.9353307485580444, 'train/grad_norm_pre_clip': 14.431270599365234, 'timing/data': 0.004936792887747288, 'timing/model': 1.2016399829881266, 'learning_rate/qwen_vl_interface': 5.535303332844546e-06, 'learning_rate/action_model': 5.5353033328445455e-05, 'epoch': 1.01}) +[2026-06-17 12:31:29,513][starVLA.training.train_starvla][INFO] - Step 3419, Loss: {'train/loss': 0.7464798092842102, 'train/grad_norm_pre_clip': 8.674492835998535, 'timing/data': 0.005214608972892165, 'timing/model': 1.203012457001023, 'learning_rate/qwen_vl_interface': 5.533144517411273e-06, 'learning_rate/action_model': 5.533144517411273e-05, 'epoch': 1.01}) +[2026-06-17 12:31:34,295][starVLA.training.train_starvla][INFO] - Step 3420, Loss: {'train/loss': 0.774673581123352, 'train/grad_norm_pre_clip': 9.824203491210938, 'timing/data': 0.004670377937145531, 'timing/model': 1.2064608989749104, 'learning_rate/qwen_vl_interface': 5.530985643281834e-06, 'learning_rate/action_model': 5.530985643281834e-05, 'epoch': 1.01}) +[2026-06-17 12:31:39,080][starVLA.training.train_starvla][INFO] - Step 3421, Loss: {'train/loss': 0.6293160319328308, 'train/grad_norm_pre_clip': 9.198439598083496, 'timing/data': 0.0051650130189955235, 'timing/model': 1.3317971200449392, 'learning_rate/qwen_vl_interface': 5.528826710903774e-06, 'learning_rate/action_model': 5.528826710903774e-05, 'epoch': 1.01}) +[2026-06-17 12:31:43,854][starVLA.training.train_starvla][INFO] - Step 3422, Loss: {'train/loss': 1.0222264528274536, 'train/grad_norm_pre_clip': 16.893451690673828, 'timing/data': 0.004830749938264489, 'timing/model': 1.325776448007673, 'learning_rate/qwen_vl_interface': 5.526667720724632e-06, 'learning_rate/action_model': 5.526667720724632e-05, 'epoch': 1.01}) +[2026-06-17 12:31:48,537][starVLA.training.train_starvla][INFO] - Step 3423, Loss: {'train/loss': 0.5605515241622925, 'train/grad_norm_pre_clip': 7.871151924133301, 'timing/data': 0.004956866032443941, 'timing/model': 1.204672042047605, 'learning_rate/qwen_vl_interface': 5.5245086731919764e-06, 'learning_rate/action_model': 5.524508673191976e-05, 'epoch': 1.01}) +[2026-06-17 12:31:53,282][starVLA.training.train_starvla][INFO] - Step 3424, Loss: {'train/loss': 0.5987155437469482, 'train/grad_norm_pre_clip': 7.691425323486328, 'timing/data': 0.004886295064352453, 'timing/model': 1.2058358110953122, 'learning_rate/qwen_vl_interface': 5.522349568753376e-06, 'learning_rate/action_model': 5.522349568753377e-05, 'epoch': 1.01}) +[2026-06-17 12:31:58,064][starVLA.training.train_starvla][INFO] - Step 3425, Loss: {'train/loss': 0.784824550151825, 'train/grad_norm_pre_clip': 10.421456336975098, 'timing/data': 0.00451033900026232, 'timing/model': 1.2043950069928542, 'learning_rate/qwen_vl_interface': 5.520190407856419e-06, 'learning_rate/action_model': 5.5201904078564185e-05, 'epoch': 1.01}) +[2026-06-17 12:32:02,841][starVLA.training.train_starvla][INFO] - Step 3426, Loss: {'train/loss': 0.6351065039634705, 'train/grad_norm_pre_clip': 9.198090553283691, 'timing/data': 0.00463027402292937, 'timing/model': 1.2022526499349624, 'learning_rate/qwen_vl_interface': 5.5180311909486974e-06, 'learning_rate/action_model': 5.518031190948697e-05, 'epoch': 1.01}) +[2026-06-17 12:32:07,625][starVLA.training.train_starvla][INFO] - Step 3427, Loss: {'train/loss': 0.7030512094497681, 'train/grad_norm_pre_clip': 20.179704666137695, 'timing/data': 0.004987729014828801, 'timing/model': 1.3299468230688944, 'learning_rate/qwen_vl_interface': 5.515871918477821e-06, 'learning_rate/action_model': 5.515871918477821e-05, 'epoch': 1.01}) +[2026-06-17 12:32:12,401][starVLA.training.train_starvla][INFO] - Step 3428, Loss: {'train/loss': 1.1235085725784302, 'train/grad_norm_pre_clip': 12.327101707458496, 'timing/data': 0.00470391009002924, 'timing/model': 1.325307153048925, 'learning_rate/qwen_vl_interface': 5.513712590891414e-06, 'learning_rate/action_model': 5.513712590891413e-05, 'epoch': 1.01}) +[2026-06-17 12:32:17,048][starVLA.training.train_starvla][INFO] - Step 3429, Loss: {'train/loss': 0.7737538814544678, 'train/grad_norm_pre_clip': 11.99739933013916, 'timing/data': 0.004884394002147019, 'timing/model': 1.2025127039523795, 'learning_rate/qwen_vl_interface': 5.511553208637098e-06, 'learning_rate/action_model': 5.511553208637098e-05, 'epoch': 1.01}) +[2026-06-17 12:32:21,798][starVLA.training.train_starvla][INFO] - Step 3430, Loss: {'train/loss': 0.631482720375061, 'train/grad_norm_pre_clip': 9.381094932556152, 'timing/data': 0.004974841955117881, 'timing/model': 1.2040634710574523, 'learning_rate/qwen_vl_interface': 5.509393772162524e-06, 'learning_rate/action_model': 5.509393772162523e-05, 'epoch': 1.01}) +[2026-06-17 12:32:26,581][starVLA.training.train_starvla][INFO] - Step 3431, Loss: {'train/loss': 0.9871418476104736, 'train/grad_norm_pre_clip': 10.719788551330566, 'timing/data': 0.004625246045179665, 'timing/model': 1.2054173829965293, 'learning_rate/qwen_vl_interface': 5.50723428191534e-06, 'learning_rate/action_model': 5.50723428191534e-05, 'epoch': 1.01}) +[2026-06-17 12:32:31,365][starVLA.training.train_starvla][INFO] - Step 3432, Loss: {'train/loss': 0.7530624270439148, 'train/grad_norm_pre_clip': 12.957709312438965, 'timing/data': 0.0055749251041561365, 'timing/model': 1.2021867090370506, 'learning_rate/qwen_vl_interface': 5.505074738343212e-06, 'learning_rate/action_model': 5.505074738343212e-05, 'epoch': 1.01}) +[2026-06-17 12:32:36,097][starVLA.training.train_starvla][INFO] - Step 3433, Loss: {'train/loss': 0.6509722471237183, 'train/grad_norm_pre_clip': 10.691543579101562, 'timing/data': 0.004761457908898592, 'timing/model': 1.2724651700118557, 'learning_rate/qwen_vl_interface': 5.502915141893819e-06, 'learning_rate/action_model': 5.5029151418938184e-05, 'epoch': 1.01}) +[2026-06-17 12:32:40,752][starVLA.training.train_starvla][INFO] - Step 3434, Loss: {'train/loss': 0.6038061380386353, 'train/grad_norm_pre_clip': 8.535940170288086, 'timing/data': 0.004951864015311003, 'timing/model': 1.2039680109592155, 'learning_rate/qwen_vl_interface': 5.5007554930148436e-06, 'learning_rate/action_model': 5.5007554930148434e-05, 'epoch': 1.01}) +[2026-06-17 12:32:45,525][starVLA.training.train_starvla][INFO] - Step 3435, Loss: {'train/loss': 0.6327398419380188, 'train/grad_norm_pre_clip': 13.451302528381348, 'timing/data': 0.004834454040974379, 'timing/model': 1.2043966320343316, 'learning_rate/qwen_vl_interface': 5.498595792153985e-06, 'learning_rate/action_model': 5.498595792153984e-05, 'epoch': 1.01}) +[2026-06-17 12:32:50,334][starVLA.training.train_starvla][INFO] - Step 3436, Loss: {'train/loss': 0.4756956696510315, 'train/grad_norm_pre_clip': 8.241092681884766, 'timing/data': 0.004911091993562877, 'timing/model': 1.2160231759771705, 'learning_rate/qwen_vl_interface': 5.49643603975895e-06, 'learning_rate/action_model': 5.4964360397589495e-05, 'epoch': 1.01}) +[2026-06-17 12:32:55,157][starVLA.training.train_starvla][INFO] - Step 3437, Loss: {'train/loss': 0.5351946353912354, 'train/grad_norm_pre_clip': 8.981953620910645, 'timing/data': 0.004526581964455545, 'timing/model': 1.2047426390927285, 'learning_rate/qwen_vl_interface': 5.494276236277458e-06, 'learning_rate/action_model': 5.4942762362774584e-05, 'epoch': 1.01}) +[2026-06-17 12:32:59,951][starVLA.training.train_starvla][INFO] - Step 3438, Loss: {'train/loss': 0.5690956711769104, 'train/grad_norm_pre_clip': 12.988273620605469, 'timing/data': 0.00495427695568651, 'timing/model': 1.3321175309829414, 'learning_rate/qwen_vl_interface': 5.492116382157242e-06, 'learning_rate/action_model': 5.492116382157242e-05, 'epoch': 1.01}) +[2026-06-17 12:33:04,689][starVLA.training.train_starvla][INFO] - Step 3439, Loss: {'train/loss': 0.597404956817627, 'train/grad_norm_pre_clip': 9.872309684753418, 'timing/data': 0.004856430925428867, 'timing/model': 1.2862918239552528, 'learning_rate/qwen_vl_interface': 5.489956477846038e-06, 'learning_rate/action_model': 5.489956477846038e-05, 'epoch': 1.01}) +[2026-06-17 12:33:09,364][starVLA.training.train_starvla][INFO] - Step 3440, Loss: {'train/loss': 0.9594249725341797, 'train/grad_norm_pre_clip': 36.32573699951172, 'timing/data': 0.0048288400284945965, 'timing/model': 1.2061478330288082, 'learning_rate/qwen_vl_interface': 5.4877965237915984e-06, 'learning_rate/action_model': 5.4877965237915986e-05, 'epoch': 1.01}) +[2026-06-17 12:33:14,148][starVLA.training.train_starvla][INFO] - Step 3441, Loss: {'train/loss': 0.8391304612159729, 'train/grad_norm_pre_clip': 18.089576721191406, 'timing/data': 0.004603222012519836, 'timing/model': 1.2045153920771554, 'learning_rate/qwen_vl_interface': 5.485636520441684e-06, 'learning_rate/action_model': 5.4856365204416836e-05, 'epoch': 1.01}) +[2026-06-17 12:33:18,931][starVLA.training.train_starvla][INFO] - Step 3442, Loss: {'train/loss': 0.8951131105422974, 'train/grad_norm_pre_clip': 18.9597110748291, 'timing/data': 0.004877174040302634, 'timing/model': 1.2048645730828866, 'learning_rate/qwen_vl_interface': 5.483476468244063e-06, 'learning_rate/action_model': 5.483476468244063e-05, 'epoch': 1.01}) +[2026-06-17 12:33:23,717][starVLA.training.train_starvla][INFO] - Step 3443, Loss: {'train/loss': 0.6323391199111938, 'train/grad_norm_pre_clip': 9.325078964233398, 'timing/data': 0.004788742051459849, 'timing/model': 1.2044853209517896, 'learning_rate/qwen_vl_interface': 5.481316367646522e-06, 'learning_rate/action_model': 5.4813163676465216e-05, 'epoch': 1.01}) +[2026-06-17 12:33:28,494][starVLA.training.train_starvla][INFO] - Step 3444, Loss: {'train/loss': 0.6359342336654663, 'train/grad_norm_pre_clip': 10.51471996307373, 'timing/data': 0.004869782947935164, 'timing/model': 1.2049800771055743, 'learning_rate/qwen_vl_interface': 5.479156219096847e-06, 'learning_rate/action_model': 5.479156219096847e-05, 'epoch': 1.01}) +[2026-06-17 12:33:33,238][starVLA.training.train_starvla][INFO] - Step 3445, Loss: {'train/loss': 0.6640322208404541, 'train/grad_norm_pre_clip': 9.457093238830566, 'timing/data': 0.005119687062688172, 'timing/model': 1.2906849229475483, 'learning_rate/qwen_vl_interface': 5.476996023042843e-06, 'learning_rate/action_model': 5.476996023042843e-05, 'epoch': 1.01}) +[2026-06-17 12:33:37,912][starVLA.training.train_starvla][INFO] - Step 3446, Loss: {'train/loss': 0.5304979085922241, 'train/grad_norm_pre_clip': 9.190397262573242, 'timing/data': 0.0049672359600663185, 'timing/model': 1.2042949730530381, 'learning_rate/qwen_vl_interface': 5.474835779932317e-06, 'learning_rate/action_model': 5.474835779932317e-05, 'epoch': 1.01}) +[2026-06-17 12:33:42,700][starVLA.training.train_starvla][INFO] - Step 3447, Loss: {'train/loss': 0.7641623616218567, 'train/grad_norm_pre_clip': 11.421797752380371, 'timing/data': 0.004812039085663855, 'timing/model': 1.2033705699723214, 'learning_rate/qwen_vl_interface': 5.472675490213093e-06, 'learning_rate/action_model': 5.472675490213093e-05, 'epoch': 1.01}) +[2026-06-17 12:33:47,491][starVLA.training.train_starvla][INFO] - Step 3448, Loss: {'train/loss': 0.8340062499046326, 'train/grad_norm_pre_clip': 13.014935493469238, 'timing/data': 0.00508471392095089, 'timing/model': 1.2056376669788733, 'learning_rate/qwen_vl_interface': 5.470515154333e-06, 'learning_rate/action_model': 5.470515154333e-05, 'epoch': 1.01}) +[2026-06-17 12:33:52,303][starVLA.training.train_starvla][INFO] - Step 3449, Loss: {'train/loss': 0.5741254091262817, 'train/grad_norm_pre_clip': 10.531458854675293, 'timing/data': 0.004901345004327595, 'timing/model': 1.2139017520239577, 'learning_rate/qwen_vl_interface': 5.468354772739879e-06, 'learning_rate/action_model': 5.468354772739879e-05, 'epoch': 1.01}) +[2026-06-17 12:33:57,109][starVLA.training.train_starvla][INFO] - Step 3450, Loss: {'train/loss': 0.5747449398040771, 'train/grad_norm_pre_clip': 11.251463890075684, 'timing/data': 0.004949180991388857, 'timing/model': 1.33398291899357, 'learning_rate/qwen_vl_interface': 5.466194345881578e-06, 'learning_rate/action_model': 5.4661943458815776e-05, 'epoch': 1.02}) +[2026-06-17 12:34:01,764][starVLA.training.train_starvla][INFO] - Step 3451, Loss: {'train/loss': 0.7687218189239502, 'train/grad_norm_pre_clip': 13.603004455566406, 'timing/data': 0.0048580270959064364, 'timing/model': 1.203729308093898, 'learning_rate/qwen_vl_interface': 5.4640338742059564e-06, 'learning_rate/action_model': 5.464033874205956e-05, 'epoch': 1.02}) +[2026-06-17 12:34:06,603][starVLA.training.train_starvla][INFO] - Step 3452, Loss: {'train/loss': 0.8776068687438965, 'train/grad_norm_pre_clip': 22.322982788085938, 'timing/data': 0.004783190903253853, 'timing/model': 1.2056108390679583, 'learning_rate/qwen_vl_interface': 5.461873358160883e-06, 'learning_rate/action_model': 5.461873358160883e-05, 'epoch': 1.02}) +[2026-06-17 12:34:11,381][starVLA.training.train_starvla][INFO] - Step 3453, Loss: {'train/loss': 0.5261943340301514, 'train/grad_norm_pre_clip': 12.883832931518555, 'timing/data': 0.004674199968576431, 'timing/model': 1.2024178570136428, 'learning_rate/qwen_vl_interface': 5.459712798194234e-06, 'learning_rate/action_model': 5.459712798194233e-05, 'epoch': 1.02}) +[2026-06-17 12:34:16,102][starVLA.training.train_starvla][INFO] - Step 3454, Loss: {'train/loss': 0.5778031349182129, 'train/grad_norm_pre_clip': 11.515809059143066, 'timing/data': 0.004982130019925535, 'timing/model': 1.210326120024547, 'learning_rate/qwen_vl_interface': 5.457552194753896e-06, 'learning_rate/action_model': 5.457552194753895e-05, 'epoch': 1.02}) +[2026-06-17 12:34:20,849][starVLA.training.train_starvla][INFO] - Step 3455, Loss: {'train/loss': 0.5845058560371399, 'train/grad_norm_pre_clip': 11.012398719787598, 'timing/data': 0.004768033977597952, 'timing/model': 1.2049641410121694, 'learning_rate/qwen_vl_interface': 5.455391548287764e-06, 'learning_rate/action_model': 5.455391548287764e-05, 'epoch': 1.02}) +[2026-06-17 12:34:25,643][starVLA.training.train_starvla][INFO] - Step 3456, Loss: {'train/loss': 0.5995840430259705, 'train/grad_norm_pre_clip': 9.858981132507324, 'timing/data': 0.004549968987703323, 'timing/model': 1.3381331330165267, 'learning_rate/qwen_vl_interface': 5.453230859243745e-06, 'learning_rate/action_model': 5.4532308592437444e-05, 'epoch': 1.02}) +[2026-06-17 12:34:30,306][starVLA.training.train_starvla][INFO] - Step 3457, Loss: {'train/loss': 0.7719433307647705, 'train/grad_norm_pre_clip': 13.18594741821289, 'timing/data': 0.004946510074660182, 'timing/model': 1.2032608520239592, 'learning_rate/qwen_vl_interface': 5.451070128069748e-06, 'learning_rate/action_model': 5.4510701280697485e-05, 'epoch': 1.02}) +[2026-06-17 12:34:35,092][starVLA.training.train_starvla][INFO] - Step 3458, Loss: {'train/loss': 0.5301574468612671, 'train/grad_norm_pre_clip': 9.697278022766113, 'timing/data': 0.004861159948632121, 'timing/model': 1.2063669859198853, 'learning_rate/qwen_vl_interface': 5.4489093552136986e-06, 'learning_rate/action_model': 5.4489093552136984e-05, 'epoch': 1.02}) +[2026-06-17 12:34:39,889][starVLA.training.train_starvla][INFO] - Step 3459, Loss: {'train/loss': 0.9300987124443054, 'train/grad_norm_pre_clip': 13.365429878234863, 'timing/data': 0.0050419908948242664, 'timing/model': 1.2047963319346309, 'learning_rate/qwen_vl_interface': 5.446748541123526e-06, 'learning_rate/action_model': 5.4467485411235254e-05, 'epoch': 1.02}) +[2026-06-17 12:34:44,674][starVLA.training.train_starvla][INFO] - Step 3460, Loss: {'train/loss': 0.8740155100822449, 'train/grad_norm_pre_clip': 14.510610580444336, 'timing/data': 0.0048798739444464445, 'timing/model': 1.2034573699347675, 'learning_rate/qwen_vl_interface': 5.444587686247168e-06, 'learning_rate/action_model': 5.4445876862471675e-05, 'epoch': 1.02}) +[2026-06-17 12:34:49,426][starVLA.training.train_starvla][INFO] - Step 3461, Loss: {'train/loss': 0.7221319675445557, 'train/grad_norm_pre_clip': 13.375321388244629, 'timing/data': 0.00486698595341295, 'timing/model': 1.2046670040581375, 'learning_rate/qwen_vl_interface': 5.442426791032574e-06, 'learning_rate/action_model': 5.4424267910325746e-05, 'epoch': 1.02}) +[2026-06-17 12:34:54,213][starVLA.training.train_starvla][INFO] - Step 3462, Loss: {'train/loss': 0.9692263603210449, 'train/grad_norm_pre_clip': 16.253129959106445, 'timing/data': 0.004733195994049311, 'timing/model': 1.3210840040119365, 'learning_rate/qwen_vl_interface': 5.4402658559277e-06, 'learning_rate/action_model': 5.4402658559276995e-05, 'epoch': 1.02}) +[2026-06-17 12:34:58,871][starVLA.training.train_starvla][INFO] - Step 3463, Loss: {'train/loss': 0.9210615754127502, 'train/grad_norm_pre_clip': 18.45722770690918, 'timing/data': 0.004781480063684285, 'timing/model': 1.2047943889629096, 'learning_rate/qwen_vl_interface': 5.438104881380507e-06, 'learning_rate/action_model': 5.438104881380507e-05, 'epoch': 1.02}) +[2026-06-17 12:35:03,661][starVLA.training.train_starvla][INFO] - Step 3464, Loss: {'train/loss': 0.6201335787773132, 'train/grad_norm_pre_clip': 13.604504585266113, 'timing/data': 0.004816271015442908, 'timing/model': 1.2044431739486754, 'learning_rate/qwen_vl_interface': 5.435943867838971e-06, 'learning_rate/action_model': 5.435943867838971e-05, 'epoch': 1.02}) +[2026-06-17 12:35:08,451][starVLA.training.train_starvla][INFO] - Step 3465, Loss: {'train/loss': 0.7244508862495422, 'train/grad_norm_pre_clip': 11.549751281738281, 'timing/data': 0.005024648038670421, 'timing/model': 1.2077529419912025, 'learning_rate/qwen_vl_interface': 5.433782815751072e-06, 'learning_rate/action_model': 5.433782815751072e-05, 'epoch': 1.02}) +[2026-06-17 12:35:13,241][starVLA.training.train_starvla][INFO] - Step 3466, Loss: {'train/loss': 0.5101245045661926, 'train/grad_norm_pre_clip': 11.512558937072754, 'timing/data': 0.004683576058596373, 'timing/model': 1.2057467580307275, 'learning_rate/qwen_vl_interface': 5.431621725564797e-06, 'learning_rate/action_model': 5.4316217255647974e-05, 'epoch': 1.02}) +[2026-06-17 12:35:18,060][starVLA.training.train_starvla][INFO] - Step 3467, Loss: {'train/loss': 0.6119808554649353, 'train/grad_norm_pre_clip': 12.15109634399414, 'timing/data': 0.004872435005381703, 'timing/model': 1.3481665230356157, 'learning_rate/qwen_vl_interface': 5.429460597728142e-06, 'learning_rate/action_model': 5.429460597728142e-05, 'epoch': 1.02}) +[2026-06-17 12:35:22,717][starVLA.training.train_starvla][INFO] - Step 3468, Loss: {'train/loss': 0.9192987680435181, 'train/grad_norm_pre_clip': 12.82711124420166, 'timing/data': 0.004925618995912373, 'timing/model': 1.2054146779701114, 'learning_rate/qwen_vl_interface': 5.4272994326891114e-06, 'learning_rate/action_model': 5.427299432689111e-05, 'epoch': 1.02}) +[2026-06-17 12:35:27,548][starVLA.training.train_starvla][INFO] - Step 3469, Loss: {'train/loss': 0.6257233619689941, 'train/grad_norm_pre_clip': 10.565735816955566, 'timing/data': 0.005008824053220451, 'timing/model': 1.2114184929523617, 'learning_rate/qwen_vl_interface': 5.425138230895719e-06, 'learning_rate/action_model': 5.425138230895719e-05, 'epoch': 1.02}) +[2026-06-17 12:35:32,271][starVLA.training.train_starvla][INFO] - Step 3470, Loss: {'train/loss': 0.3651035726070404, 'train/grad_norm_pre_clip': 8.650979995727539, 'timing/data': 0.004646940971724689, 'timing/model': 1.203966944012791, 'learning_rate/qwen_vl_interface': 5.422976992795984e-06, 'learning_rate/action_model': 5.422976992795984e-05, 'epoch': 1.02}) +[2026-06-17 12:35:37,042][starVLA.training.train_starvla][INFO] - Step 3471, Loss: {'train/loss': 0.7795110940933228, 'train/grad_norm_pre_clip': 11.82317066192627, 'timing/data': 0.004702843958511949, 'timing/model': 1.2103140400722623, 'learning_rate/qwen_vl_interface': 5.420815718837929e-06, 'learning_rate/action_model': 5.42081571883793e-05, 'epoch': 1.02}) +[2026-06-17 12:35:41,854][starVLA.training.train_starvla][INFO] - Step 3472, Loss: {'train/loss': 0.8159250020980835, 'train/grad_norm_pre_clip': 12.274068832397461, 'timing/data': 0.00454009894747287, 'timing/model': 1.211697445018217, 'learning_rate/qwen_vl_interface': 5.418654409469595e-06, 'learning_rate/action_model': 5.4186544094695946e-05, 'epoch': 1.02}) +[2026-06-17 12:35:46,647][starVLA.training.train_starvla][INFO] - Step 3473, Loss: {'train/loss': 0.6491334438323975, 'train/grad_norm_pre_clip': 9.081524848937988, 'timing/data': 0.00513061392121017, 'timing/model': 1.3308342109667137, 'learning_rate/qwen_vl_interface': 5.416493065139019e-06, 'learning_rate/action_model': 5.4164930651390186e-05, 'epoch': 1.02}) +[2026-06-17 12:35:51,309][starVLA.training.train_starvla][INFO] - Step 3474, Loss: {'train/loss': 0.4739181399345398, 'train/grad_norm_pre_clip': 10.378242492675781, 'timing/data': 0.00470858602784574, 'timing/model': 1.2041584650287405, 'learning_rate/qwen_vl_interface': 5.4143316862942516e-06, 'learning_rate/action_model': 5.4143316862942514e-05, 'epoch': 1.02}) +[2026-06-17 12:35:56,143][starVLA.training.train_starvla][INFO] - Step 3475, Loss: {'train/loss': 0.8060040473937988, 'train/grad_norm_pre_clip': 15.057268142700195, 'timing/data': 0.004902992979623377, 'timing/model': 1.21460320497863, 'learning_rate/qwen_vl_interface': 5.412170273383348e-06, 'learning_rate/action_model': 5.412170273383348e-05, 'epoch': 1.02}) +[2026-06-17 12:36:00,916][starVLA.training.train_starvla][INFO] - Step 3476, Loss: {'train/loss': 0.7738674879074097, 'train/grad_norm_pre_clip': 11.60933780670166, 'timing/data': 0.004874062025919557, 'timing/model': 1.2044256059452891, 'learning_rate/qwen_vl_interface': 5.410008826854374e-06, 'learning_rate/action_model': 5.410008826854373e-05, 'epoch': 1.02}) +[2026-06-17 12:36:05,673][starVLA.training.train_starvla][INFO] - Step 3477, Loss: {'train/loss': 0.7839200496673584, 'train/grad_norm_pre_clip': 18.939844131469727, 'timing/data': 0.004340377985499799, 'timing/model': 1.204361020005308, 'learning_rate/qwen_vl_interface': 5.407847347155395e-06, 'learning_rate/action_model': 5.407847347155395e-05, 'epoch': 1.02}) +[2026-06-17 12:36:10,470][starVLA.training.train_starvla][INFO] - Step 3478, Loss: {'train/loss': 0.9585998058319092, 'train/grad_norm_pre_clip': 13.631491661071777, 'timing/data': 0.0050729489885270596, 'timing/model': 1.2214652620023116, 'learning_rate/qwen_vl_interface': 5.4056858347344934e-06, 'learning_rate/action_model': 5.405685834734493e-05, 'epoch': 1.02}) +[2026-06-17 12:36:15,280][starVLA.training.train_starvla][INFO] - Step 3479, Loss: {'train/loss': 0.8200017213821411, 'train/grad_norm_pre_clip': 11.612886428833008, 'timing/data': 0.004863703972660005, 'timing/model': 1.3501645890064538, 'learning_rate/qwen_vl_interface': 5.403524290039751e-06, 'learning_rate/action_model': 5.403524290039751e-05, 'epoch': 1.02}) +[2026-06-17 12:36:19,950][starVLA.training.train_starvla][INFO] - Step 3480, Loss: {'train/loss': 0.539160966873169, 'train/grad_norm_pre_clip': 20.98834991455078, 'timing/data': 0.004972669994458556, 'timing/model': 1.2124827760271728, 'learning_rate/qwen_vl_interface': 5.401362713519257e-06, 'learning_rate/action_model': 5.4013627135192565e-05, 'epoch': 1.02}) +[2026-06-17 12:36:24,822][starVLA.training.train_starvla][INFO] - Step 3481, Loss: {'train/loss': 0.6483212113380432, 'train/grad_norm_pre_clip': 11.682719230651855, 'timing/data': 0.004621216910891235, 'timing/model': 1.2060677399858832, 'learning_rate/qwen_vl_interface': 5.399201105621111e-06, 'learning_rate/action_model': 5.39920110562111e-05, 'epoch': 1.02}) +[2026-06-17 12:36:29,718][starVLA.training.train_starvla][INFO] - Step 3482, Loss: {'train/loss': 0.6030592322349548, 'train/grad_norm_pre_clip': 10.869038581848145, 'timing/data': 0.004990500048734248, 'timing/model': 1.2133993279421702, 'learning_rate/qwen_vl_interface': 5.397039466793412e-06, 'learning_rate/action_model': 5.3970394667934123e-05, 'epoch': 1.02}) +[2026-06-17 12:36:34,517][starVLA.training.train_starvla][INFO] - Step 3483, Loss: {'train/loss': 0.5194905400276184, 'train/grad_norm_pre_clip': 9.63905143737793, 'timing/data': 0.00501553004141897, 'timing/model': 1.2089319679653272, 'learning_rate/qwen_vl_interface': 5.394877797484278e-06, 'learning_rate/action_model': 5.3948777974842777e-05, 'epoch': 1.02}) +[2026-06-17 12:36:39,210][starVLA.training.train_starvla][INFO] - Step 3484, Loss: {'train/loss': 0.5889286994934082, 'train/grad_norm_pre_clip': 9.044515609741211, 'timing/data': 0.005171741941012442, 'timing/model': 1.2065624439856037, 'learning_rate/qwen_vl_interface': 5.392716098141817e-06, 'learning_rate/action_model': 5.3927160981418165e-05, 'epoch': 1.03}) +[2026-06-17 12:36:43,929][starVLA.training.train_starvla][INFO] - Step 3485, Loss: {'train/loss': 0.8432698845863342, 'train/grad_norm_pre_clip': 12.32309341430664, 'timing/data': 0.004652686999179423, 'timing/model': 1.2427714789519086, 'learning_rate/qwen_vl_interface': 5.390554369214157e-06, 'learning_rate/action_model': 5.390554369214157e-05, 'epoch': 1.03}) +[2026-06-17 12:36:48,588][starVLA.training.train_starvla][INFO] - Step 3486, Loss: {'train/loss': 0.5943493247032166, 'train/grad_norm_pre_clip': 20.084026336669922, 'timing/data': 0.004888463066890836, 'timing/model': 1.2041869760723785, 'learning_rate/qwen_vl_interface': 5.388392611149425e-06, 'learning_rate/action_model': 5.388392611149424e-05, 'epoch': 1.03}) +[2026-06-17 12:36:53,322][starVLA.training.train_starvla][INFO] - Step 3487, Loss: {'train/loss': 0.6405870318412781, 'train/grad_norm_pre_clip': 10.618593215942383, 'timing/data': 0.004868685966357589, 'timing/model': 1.2066142960684374, 'learning_rate/qwen_vl_interface': 5.386230824395754e-06, 'learning_rate/action_model': 5.386230824395754e-05, 'epoch': 1.03}) +[2026-06-17 12:36:58,137][starVLA.training.train_starvla][INFO] - Step 3488, Loss: {'train/loss': 0.6792059540748596, 'train/grad_norm_pre_clip': 14.154169082641602, 'timing/data': 0.004761705058626831, 'timing/model': 1.2099682970438153, 'learning_rate/qwen_vl_interface': 5.384069009401288e-06, 'learning_rate/action_model': 5.384069009401288e-05, 'epoch': 1.03}) +[2026-06-17 12:37:02,936][starVLA.training.train_starvla][INFO] - Step 3489, Loss: {'train/loss': 0.6575343608856201, 'train/grad_norm_pre_clip': 10.37859058380127, 'timing/data': 0.004835717962123454, 'timing/model': 1.2049649968976155, 'learning_rate/qwen_vl_interface': 5.381907166614171e-06, 'learning_rate/action_model': 5.3819071666141704e-05, 'epoch': 1.03}) +[2026-06-17 12:37:07,739][starVLA.training.train_starvla][INFO] - Step 3490, Loss: {'train/loss': 0.5316373109817505, 'train/grad_norm_pre_clip': 15.610517501831055, 'timing/data': 0.0046377640683203936, 'timing/model': 1.336005971999839, 'learning_rate/qwen_vl_interface': 5.379745296482559e-06, 'learning_rate/action_model': 5.379745296482559e-05, 'epoch': 1.03}) +[2026-06-17 12:37:12,457][starVLA.training.train_starvla][INFO] - Step 3491, Loss: {'train/loss': 0.6320852041244507, 'train/grad_norm_pre_clip': 17.268468856811523, 'timing/data': 0.0052328689489513636, 'timing/model': 1.2598998689791188, 'learning_rate/qwen_vl_interface': 5.377583399454605e-06, 'learning_rate/action_model': 5.3775833994546044e-05, 'epoch': 1.03}) +[2026-06-17 12:37:17,118][starVLA.training.train_starvla][INFO] - Step 3492, Loss: {'train/loss': 0.6425533890724182, 'train/grad_norm_pre_clip': 10.57401180267334, 'timing/data': 0.0047102998942136765, 'timing/model': 1.2054889269638807, 'learning_rate/qwen_vl_interface': 5.375421475978476e-06, 'learning_rate/action_model': 5.375421475978476e-05, 'epoch': 1.03}) +[2026-06-17 12:37:21,889][starVLA.training.train_starvla][INFO] - Step 3493, Loss: {'train/loss': 0.7801529169082642, 'train/grad_norm_pre_clip': 11.412335395812988, 'timing/data': 0.004768813960254192, 'timing/model': 1.2094022730598226, 'learning_rate/qwen_vl_interface': 5.37325952650234e-06, 'learning_rate/action_model': 5.37325952650234e-05, 'epoch': 1.03}) +[2026-06-17 12:37:26,782][starVLA.training.train_starvla][INFO] - Step 3494, Loss: {'train/loss': 0.5471712350845337, 'train/grad_norm_pre_clip': 11.572406768798828, 'timing/data': 0.00516639300622046, 'timing/model': 1.207455005031079, 'learning_rate/qwen_vl_interface': 5.371097551474371e-06, 'learning_rate/action_model': 5.371097551474371e-05, 'epoch': 1.03}) +[2026-06-17 12:37:31,589][starVLA.training.train_starvla][INFO] - Step 3495, Loss: {'train/loss': 0.889916181564331, 'train/grad_norm_pre_clip': 13.261024475097656, 'timing/data': 0.005123637965880334, 'timing/model': 1.2075184569694102, 'learning_rate/qwen_vl_interface': 5.36893555134275e-06, 'learning_rate/action_model': 5.3689355513427494e-05, 'epoch': 1.03}) +[2026-06-17 12:37:36,434][starVLA.training.train_starvla][INFO] - Step 3496, Loss: {'train/loss': 0.90154629945755, 'train/grad_norm_pre_clip': 12.43930435180664, 'timing/data': 0.005117814056575298, 'timing/model': 1.330454976996407, 'learning_rate/qwen_vl_interface': 5.366773526555661e-06, 'learning_rate/action_model': 5.366773526555661e-05, 'epoch': 1.03}) +[2026-06-17 12:37:41,229][starVLA.training.train_starvla][INFO] - Step 3497, Loss: {'train/loss': 0.6603355407714844, 'train/grad_norm_pre_clip': 15.179153442382812, 'timing/data': 0.0047664709854871035, 'timing/model': 1.343979757046327, 'learning_rate/qwen_vl_interface': 5.364611477561294e-06, 'learning_rate/action_model': 5.3646114775612945e-05, 'epoch': 1.03}) +[2026-06-17 12:37:45,887][starVLA.training.train_starvla][INFO] - Step 3498, Loss: {'train/loss': 0.7453862428665161, 'train/grad_norm_pre_clip': 10.495979309082031, 'timing/data': 0.005123454029671848, 'timing/model': 1.2047583360690624, 'learning_rate/qwen_vl_interface': 5.362449404807845e-06, 'learning_rate/action_model': 5.362449404807844e-05, 'epoch': 1.03}) +[2026-06-17 12:37:50,663][starVLA.training.train_starvla][INFO] - Step 3499, Loss: {'train/loss': 0.6602067351341248, 'train/grad_norm_pre_clip': 11.841936111450195, 'timing/data': 0.004735947004519403, 'timing/model': 1.2097423149971291, 'learning_rate/qwen_vl_interface': 5.360287308743512e-06, 'learning_rate/action_model': 5.3602873087435126e-05, 'epoch': 1.03}) +[2026-06-17 12:47:58,523][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 3500 +[2026-06-17 12:47:58,524][starVLA.training.train_starvla][INFO] - Step 3500, Loss: {'train/loss': 0.9273319244384766, 'train/grad_norm_pre_clip': 12.88499641418457, 'eval/loss': 0.7814648151397705, 'eval/latency_0/loss': 0.7814648151397705, 'eval/demon_fire_precision': 0.8976537216828478, 'eval/demon_fire_recall': 0.9215116279069767, 'eval/demon_fire_f1': 0.9094262295081967, 'eval/demon_fire_rate_model': 0.6874304783092324, 'eval/demon_fire_rate_teacher': 0.6696329254727474, 'eval/demon_left_precision': 0.9068627450980392, 'eval/demon_left_recall': 0.6288951841359773, 'eval/demon_left_f1': 0.742723318835731, 'eval/demon_left_rate_model': 0.3403781979977753, 'eval/demon_left_rate_teacher': 0.49082313681868744, 'eval/demon_right_precision': 0.7153222364489971, 'eval/demon_right_recall': 0.9468926553672317, 'eval/demon_right_f1': 0.8149769025042548, 'eval/demon_right_rate_model': 0.6515572858731924, 'eval/demon_right_rate_teacher': 0.4922135706340378, 'eval/demon_move_f1': 0.7788501106699929, 'eval/demon_cc_f1': 0.8658763457477804, 'eval/latency_0/demon_fire_precision': 0.8976537216828478, 'eval/latency_0/demon_fire_recall': 0.9215116279069767, 'eval/latency_0/demon_fire_f1': 0.9094262295081967, 'eval/latency_0/demon_fire_rate_model': 0.6874304783092324, 'eval/latency_0/demon_fire_rate_teacher': 0.6696329254727474, 'eval/latency_0/demon_left_precision': 0.9068627450980392, 'eval/latency_0/demon_left_recall': 0.6288951841359773, 'eval/latency_0/demon_left_f1': 0.742723318835731, 'eval/latency_0/demon_left_rate_model': 0.3403781979977753, 'eval/latency_0/demon_left_rate_teacher': 0.49082313681868744, 'eval/latency_0/demon_right_precision': 0.7153222364489971, 'eval/latency_0/demon_right_recall': 0.9468926553672317, 'eval/latency_0/demon_right_f1': 0.8149769025042548, 'eval/latency_0/demon_right_rate_model': 0.6515572858731924, 'eval/latency_0/demon_right_rate_teacher': 0.4922135706340378, 'eval/latency_0/demon_move_f1': 0.7788501106699929, 'eval/latency_0/demon_cc_f1': 0.8658763457477804, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 3993.0, 'rl_games_eval/mid_train/mean_length': 3600.0, 'rl_games_eval/mid_train/std_reward': 1864.852809205059, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/macro_mean_reward': 3993.0, 'rl_games_eval/mid_train/macro_mean_length': 3600.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/demon_attack__latency_0/mean_reward': 3993.0, 'rl_games_eval/mid_train/demon_attack__latency_0/mean_length': 3600.0, 'rl_games_eval/mid_train/demon_attack__latency_0/std_reward': 1864.852809205059, 'rl_games_eval/mid_train/demon_attack__latency_0/std_length': 0.0, 'rl_games_eval/mid_train/demon_attack__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 3993.0, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.004765212070196867, 'timing/model': 1.2059404520550743, 'learning_rate/qwen_vl_interface': 5.358125189816502e-06, 'learning_rate/action_model': 5.358125189816502e-05, 'epoch': 1.03}) +[2026-06-17 12:47:58,524][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_3500_state +[2026-06-17 12:47:58,527][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 12:48:02,079][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_3500_state/model.safetensors +[2026-06-17 12:48:09,693][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_3500_state/optimizer.bin +[2026-06-17 12:48:09,693][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_3500_state/sampler.bin +[2026-06-17 12:48:09,693][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_3500_state/sampler_1.bin +[2026-06-17 12:48:09,694][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_3500_state/random_states_0.pkl +[2026-06-17 12:48:09,743][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3500 +[2026-06-17 12:48:09,743][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 12:48:09,753][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 12:48:14,454][starVLA.training.train_starvla][INFO] - Step 3501, Loss: {'train/loss': 0.4616486132144928, 'train/grad_norm_pre_clip': 8.00900650024414, 'timing/data': 0.005284592974931002, 'timing/model': 1.188124097068794, 'learning_rate/qwen_vl_interface': 5.355963048475024e-06, 'learning_rate/action_model': 5.3559630484750244e-05, 'epoch': 1.03}) +[2026-06-17 12:48:19,185][starVLA.training.train_starvla][INFO] - Step 3502, Loss: {'train/loss': 0.8983851075172424, 'train/grad_norm_pre_clip': 55.709415435791016, 'timing/data': 0.005294497008435428, 'timing/model': 1.32333026395645, 'learning_rate/qwen_vl_interface': 5.3538008851672916e-06, 'learning_rate/action_model': 5.3538008851672916e-05, 'epoch': 1.03}) +[2026-06-17 12:48:23,784][starVLA.training.train_starvla][INFO] - Step 3503, Loss: {'train/loss': 0.6325538754463196, 'train/grad_norm_pre_clip': 12.43288516998291, 'timing/data': 0.005449990043416619, 'timing/model': 1.189710760023445, 'learning_rate/qwen_vl_interface': 5.351638700341525e-06, 'learning_rate/action_model': 5.351638700341525e-05, 'epoch': 1.03}) +[2026-06-17 12:48:28,511][starVLA.training.train_starvla][INFO] - Step 3504, Loss: {'train/loss': 0.5953181385993958, 'train/grad_norm_pre_clip': 10.437654495239258, 'timing/data': 0.005291732959449291, 'timing/model': 1.1926276609301567, 'learning_rate/qwen_vl_interface': 5.349476494445947e-06, 'learning_rate/action_model': 5.3494764944459466e-05, 'epoch': 1.03}) +[2026-06-17 12:48:33,155][starVLA.training.train_starvla][INFO] - Step 3505, Loss: {'train/loss': 0.935200035572052, 'train/grad_norm_pre_clip': 14.968513488769531, 'timing/data': 0.0049420990981161594, 'timing/model': 1.1909921889891848, 'learning_rate/qwen_vl_interface': 5.3473142679287824e-06, 'learning_rate/action_model': 5.3473142679287826e-05, 'epoch': 1.03}) +[2026-06-17 12:48:37,845][starVLA.training.train_starvla][INFO] - Step 3506, Loss: {'train/loss': 0.9003910422325134, 'train/grad_norm_pre_clip': 14.33655071258545, 'timing/data': 0.005146467941813171, 'timing/model': 1.1963860400719568, 'learning_rate/qwen_vl_interface': 5.345152021238268e-06, 'learning_rate/action_model': 5.345152021238268e-05, 'epoch': 1.03}) +[2026-06-17 12:48:42,564][starVLA.training.train_starvla][INFO] - Step 3507, Loss: {'train/loss': 0.5948234796524048, 'train/grad_norm_pre_clip': 19.80388641357422, 'timing/data': 0.00482368899974972, 'timing/model': 1.1984025300480425, 'learning_rate/qwen_vl_interface': 5.342989754822635e-06, 'learning_rate/action_model': 5.342989754822635e-05, 'epoch': 1.03}) +[2026-06-17 12:48:47,324][starVLA.training.train_starvla][INFO] - Step 3508, Loss: {'train/loss': 0.720622181892395, 'train/grad_norm_pre_clip': 10.137956619262695, 'timing/data': 0.00519974494818598, 'timing/model': 1.3261930039152503, 'learning_rate/qwen_vl_interface': 5.340827469130126e-06, 'learning_rate/action_model': 5.3408274691301255e-05, 'epoch': 1.03}) +[2026-06-17 12:48:51,954][starVLA.training.train_starvla][INFO] - Step 3509, Loss: {'train/loss': 0.6039906740188599, 'train/grad_norm_pre_clip': 11.013912200927734, 'timing/data': 0.005509507027454674, 'timing/model': 1.1981595950201154, 'learning_rate/qwen_vl_interface': 5.338665164608985e-06, 'learning_rate/action_model': 5.3386651646089845e-05, 'epoch': 1.03}) +[2026-06-17 12:48:56,715][starVLA.training.train_starvla][INFO] - Step 3510, Loss: {'train/loss': 0.6451274752616882, 'train/grad_norm_pre_clip': 8.360957145690918, 'timing/data': 0.005303442943841219, 'timing/model': 1.1968580560060218, 'learning_rate/qwen_vl_interface': 5.336502841707461e-06, 'learning_rate/action_model': 5.33650284170746e-05, 'epoch': 1.03}) +[2026-06-17 12:49:01,468][starVLA.training.train_starvla][INFO] - Step 3511, Loss: {'train/loss': 0.6104363799095154, 'train/grad_norm_pre_clip': 12.381315231323242, 'timing/data': 0.0052385469898581505, 'timing/model': 1.1969942839350551, 'learning_rate/qwen_vl_interface': 5.334340500873803e-06, 'learning_rate/action_model': 5.3343405008738034e-05, 'epoch': 1.03}) +[2026-06-17 12:49:06,184][starVLA.training.train_starvla][INFO] - Step 3512, Loss: {'train/loss': 0.7141537666320801, 'train/grad_norm_pre_clip': 15.145219802856445, 'timing/data': 0.005257967044599354, 'timing/model': 1.2012751959264278, 'learning_rate/qwen_vl_interface': 5.33217814255627e-06, 'learning_rate/action_model': 5.33217814255627e-05, 'epoch': 1.03}) +[2026-06-17 12:49:10,956][starVLA.training.train_starvla][INFO] - Step 3513, Loss: {'train/loss': 0.8477702140808105, 'train/grad_norm_pre_clip': 14.792577743530273, 'timing/data': 0.0055248679127544165, 'timing/model': 1.200650362065062, 'learning_rate/qwen_vl_interface': 5.330015767203118e-06, 'learning_rate/action_model': 5.330015767203118e-05, 'epoch': 1.03}) +[2026-06-17 12:49:15,723][starVLA.training.train_starvla][INFO] - Step 3514, Loss: {'train/loss': 0.5874621272087097, 'train/grad_norm_pre_clip': 12.258838653564453, 'timing/data': 0.005338053917512298, 'timing/model': 1.3271034599747509, 'learning_rate/qwen_vl_interface': 5.3278533752626135e-06, 'learning_rate/action_model': 5.3278533752626136e-05, 'epoch': 1.03}) +[2026-06-17 12:49:20,444][starVLA.training.train_starvla][INFO] - Step 3515, Loss: {'train/loss': 0.7638480067253113, 'train/grad_norm_pre_clip': 10.422721862792969, 'timing/data': 0.005792380077764392, 'timing/model': 1.1994520369917154, 'learning_rate/qwen_vl_interface': 5.325690967183019e-06, 'learning_rate/action_model': 5.325690967183019e-05, 'epoch': 1.03}) +[2026-06-17 12:49:25,215][starVLA.training.train_starvla][INFO] - Step 3516, Loss: {'train/loss': 0.5622000098228455, 'train/grad_norm_pre_clip': 9.638398170471191, 'timing/data': 0.00523325998801738, 'timing/model': 1.2016862280433998, 'learning_rate/qwen_vl_interface': 5.323528543412607e-06, 'learning_rate/action_model': 5.323528543412607e-05, 'epoch': 1.03}) +[2026-06-17 12:49:30,019][starVLA.training.train_starvla][INFO] - Step 3517, Loss: {'train/loss': 0.8179219365119934, 'train/grad_norm_pre_clip': 12.247276306152344, 'timing/data': 0.005288230022415519, 'timing/model': 1.2008954400662333, 'learning_rate/qwen_vl_interface': 5.321366104399648e-06, 'learning_rate/action_model': 5.321366104399648e-05, 'epoch': 1.03}) +[2026-06-17 12:49:34,752][starVLA.training.train_starvla][INFO] - Step 3518, Loss: {'train/loss': 0.8398678302764893, 'train/grad_norm_pre_clip': 16.452253341674805, 'timing/data': 0.0057666609063744545, 'timing/model': 1.2023818080779165, 'learning_rate/qwen_vl_interface': 5.319203650592419e-06, 'learning_rate/action_model': 5.3192036505924194e-05, 'epoch': 1.04}) +[2026-06-17 12:49:39,513][starVLA.training.train_starvla][INFO] - Step 3519, Loss: {'train/loss': 0.7484496831893921, 'train/grad_norm_pre_clip': 10.927273750305176, 'timing/data': 0.005192331038415432, 'timing/model': 1.2007115710293874, 'learning_rate/qwen_vl_interface': 5.317041182439199e-06, 'learning_rate/action_model': 5.3170411824391985e-05, 'epoch': 1.04}) +[2026-06-17 12:49:44,305][starVLA.training.train_starvla][INFO] - Step 3520, Loss: {'train/loss': 0.6795530915260315, 'train/grad_norm_pre_clip': 12.645269393920898, 'timing/data': 0.005177902989089489, 'timing/model': 1.3493301010457799, 'learning_rate/qwen_vl_interface': 5.314878700388267e-06, 'learning_rate/action_model': 5.314878700388267e-05, 'epoch': 1.04}) +[2026-06-17 12:49:48,956][starVLA.training.train_starvla][INFO] - Step 3521, Loss: {'train/loss': 0.602462649345398, 'train/grad_norm_pre_clip': 11.360697746276855, 'timing/data': 0.0052290200255811214, 'timing/model': 1.205282193957828, 'learning_rate/qwen_vl_interface': 5.312716204887914e-06, 'learning_rate/action_model': 5.3127162048879145e-05, 'epoch': 1.04}) +[2026-06-17 12:49:53,740][starVLA.training.train_starvla][INFO] - Step 3522, Loss: {'train/loss': 0.8305248618125916, 'train/grad_norm_pre_clip': 10.915580749511719, 'timing/data': 0.005138505948707461, 'timing/model': 1.2062154948944226, 'learning_rate/qwen_vl_interface': 5.3105536963864226e-06, 'learning_rate/action_model': 5.310553696386422e-05, 'epoch': 1.04}) +[2026-06-17 12:49:58,530][starVLA.training.train_starvla][INFO] - Step 3523, Loss: {'train/loss': 0.6824952363967896, 'train/grad_norm_pre_clip': 11.121522903442383, 'timing/data': 0.004961258033290505, 'timing/model': 1.2028181429486722, 'learning_rate/qwen_vl_interface': 5.3083911753320875e-06, 'learning_rate/action_model': 5.308391175332087e-05, 'epoch': 1.04}) +[2026-06-17 12:50:03,326][starVLA.training.train_starvla][INFO] - Step 3524, Loss: {'train/loss': 0.882986843585968, 'train/grad_norm_pre_clip': 12.275325775146484, 'timing/data': 0.005291357985697687, 'timing/model': 1.2051076060160995, 'learning_rate/qwen_vl_interface': 5.306228642173194e-06, 'learning_rate/action_model': 5.306228642173194e-05, 'epoch': 1.04}) +[2026-06-17 12:50:08,003][starVLA.training.train_starvla][INFO] - Step 3525, Loss: {'train/loss': 0.7108320593833923, 'train/grad_norm_pre_clip': 10.294270515441895, 'timing/data': 0.005169407930225134, 'timing/model': 1.2028268609428778, 'learning_rate/qwen_vl_interface': 5.304066097358046e-06, 'learning_rate/action_model': 5.304066097358046e-05, 'epoch': 1.04}) +[2026-06-17 12:50:12,703][starVLA.training.train_starvla][INFO] - Step 3526, Loss: {'train/loss': 0.6767033934593201, 'train/grad_norm_pre_clip': 11.933061599731445, 'timing/data': 0.005265717045404017, 'timing/model': 1.2444313080050051, 'learning_rate/qwen_vl_interface': 5.301903541334937e-06, 'learning_rate/action_model': 5.301903541334937e-05, 'epoch': 1.04}) +[2026-06-17 12:50:17,352][starVLA.training.train_starvla][INFO] - Step 3527, Loss: {'train/loss': 0.6888293623924255, 'train/grad_norm_pre_clip': 11.581846237182617, 'timing/data': 0.0054357999470084906, 'timing/model': 1.20239118905738, 'learning_rate/qwen_vl_interface': 5.299740974552168e-06, 'learning_rate/action_model': 5.299740974552167e-05, 'epoch': 1.04}) +[2026-06-17 12:50:22,087][starVLA.training.train_starvla][INFO] - Step 3528, Loss: {'train/loss': 0.5984323024749756, 'train/grad_norm_pre_clip': 11.001169204711914, 'timing/data': 0.005286094034090638, 'timing/model': 1.2027069699252024, 'learning_rate/qwen_vl_interface': 5.297578397458042e-06, 'learning_rate/action_model': 5.297578397458042e-05, 'epoch': 1.04}) +[2026-06-17 12:50:26,875][starVLA.training.train_starvla][INFO] - Step 3529, Loss: {'train/loss': 0.5778952836990356, 'train/grad_norm_pre_clip': 23.973247528076172, 'timing/data': 0.005056375986896455, 'timing/model': 1.2033354769228026, 'learning_rate/qwen_vl_interface': 5.295415810500862e-06, 'learning_rate/action_model': 5.295415810500862e-05, 'epoch': 1.04}) +[2026-06-17 12:50:31,666][starVLA.training.train_starvla][INFO] - Step 3530, Loss: {'train/loss': 0.7541632652282715, 'train/grad_norm_pre_clip': 15.033491134643555, 'timing/data': 0.005298340925946832, 'timing/model': 1.2021927109453827, 'learning_rate/qwen_vl_interface': 5.293253214128938e-06, 'learning_rate/action_model': 5.293253214128937e-05, 'epoch': 1.04}) +[2026-06-17 12:50:36,447][starVLA.training.train_starvla][INFO] - Step 3531, Loss: {'train/loss': 0.7191427946090698, 'train/grad_norm_pre_clip': 13.083333015441895, 'timing/data': 0.005303118028678, 'timing/model': 1.3341319869505242, 'learning_rate/qwen_vl_interface': 5.291090608790572e-06, 'learning_rate/action_model': 5.291090608790572e-05, 'epoch': 1.04}) +[2026-06-17 12:50:41,165][starVLA.training.train_starvla][INFO] - Step 3532, Loss: {'train/loss': 0.7066810727119446, 'train/grad_norm_pre_clip': 11.877411842346191, 'timing/data': 0.004893600009381771, 'timing/model': 1.2519863389898092, 'learning_rate/qwen_vl_interface': 5.28892799493408e-06, 'learning_rate/action_model': 5.2889279949340796e-05, 'epoch': 1.04}) +[2026-06-17 12:50:45,834][starVLA.training.train_starvla][INFO] - Step 3533, Loss: {'train/loss': 0.7377381324768066, 'train/grad_norm_pre_clip': 12.241321563720703, 'timing/data': 0.005281090037897229, 'timing/model': 1.2063109209993854, 'learning_rate/qwen_vl_interface': 5.286765373007774e-06, 'learning_rate/action_model': 5.2867653730077735e-05, 'epoch': 1.04}) +[2026-06-17 12:50:50,592][starVLA.training.train_starvla][INFO] - Step 3534, Loss: {'train/loss': 0.7633751630783081, 'train/grad_norm_pre_clip': 12.92065715789795, 'timing/data': 0.005019817966967821, 'timing/model': 1.206594847026281, 'learning_rate/qwen_vl_interface': 5.284602743459968e-06, 'learning_rate/action_model': 5.284602743459968e-05, 'epoch': 1.04}) +[2026-06-17 12:50:55,377][starVLA.training.train_starvla][INFO] - Step 3535, Loss: {'train/loss': 0.7352068424224854, 'train/grad_norm_pre_clip': 11.052096366882324, 'timing/data': 0.005268480977974832, 'timing/model': 1.2032487889518961, 'learning_rate/qwen_vl_interface': 5.282440106738973e-06, 'learning_rate/action_model': 5.282440106738973e-05, 'epoch': 1.04}) +[2026-06-17 12:51:00,166][starVLA.training.train_starvla][INFO] - Step 3536, Loss: {'train/loss': 0.7425020933151245, 'train/grad_norm_pre_clip': 12.718120574951172, 'timing/data': 0.00522189203184098, 'timing/model': 1.2039170879870653, 'learning_rate/qwen_vl_interface': 5.280277463293111e-06, 'learning_rate/action_model': 5.280277463293111e-05, 'epoch': 1.04}) +[2026-06-17 12:51:04,948][starVLA.training.train_starvla][INFO] - Step 3537, Loss: {'train/loss': 0.8467358350753784, 'train/grad_norm_pre_clip': 18.89669418334961, 'timing/data': 0.004903292050585151, 'timing/model': 1.3289931210456416, 'learning_rate/qwen_vl_interface': 5.2781148135706975e-06, 'learning_rate/action_model': 5.278114813570697e-05, 'epoch': 1.04}) +[2026-06-17 12:51:09,722][starVLA.training.train_starvla][INFO] - Step 3538, Loss: {'train/loss': 0.41786226630210876, 'train/grad_norm_pre_clip': 9.178287506103516, 'timing/data': 0.005159736960195005, 'timing/model': 1.3219374669715762, 'learning_rate/qwen_vl_interface': 5.275952158020054e-06, 'learning_rate/action_model': 5.275952158020053e-05, 'epoch': 1.04}) +[2026-06-17 12:51:14,376][starVLA.training.train_starvla][INFO] - Step 3539, Loss: {'train/loss': 0.6401182413101196, 'train/grad_norm_pre_clip': 12.424506187438965, 'timing/data': 0.0051067579770460725, 'timing/model': 1.2050923249917105, 'learning_rate/qwen_vl_interface': 5.273789497089497e-06, 'learning_rate/action_model': 5.2737894970894974e-05, 'epoch': 1.04}) +[2026-06-17 12:51:19,121][starVLA.training.train_starvla][INFO] - Step 3540, Loss: {'train/loss': 0.5051192045211792, 'train/grad_norm_pre_clip': 9.235520362854004, 'timing/data': 0.005093181040138006, 'timing/model': 1.2041496010497212, 'learning_rate/qwen_vl_interface': 5.271626831227355e-06, 'learning_rate/action_model': 5.2716268312273544e-05, 'epoch': 1.04}) +[2026-06-17 12:51:23,905][starVLA.training.train_starvla][INFO] - Step 3541, Loss: {'train/loss': 0.6762160658836365, 'train/grad_norm_pre_clip': 19.91036605834961, 'timing/data': 0.005120018031448126, 'timing/model': 1.2074185140663758, 'learning_rate/qwen_vl_interface': 5.2694641608819465e-06, 'learning_rate/action_model': 5.269464160881946e-05, 'epoch': 1.04}) +[2026-06-17 12:51:28,701][starVLA.training.train_starvla][INFO] - Step 3542, Loss: {'train/loss': 0.7354306578636169, 'train/grad_norm_pre_clip': 13.162755966186523, 'timing/data': 0.005251506110653281, 'timing/model': 1.2092010650085285, 'learning_rate/qwen_vl_interface': 5.267301486501595e-06, 'learning_rate/action_model': 5.2673014865015945e-05, 'epoch': 1.04}) +[2026-06-17 12:51:33,426][starVLA.training.train_starvla][INFO] - Step 3543, Loss: {'train/loss': 0.9394727945327759, 'train/grad_norm_pre_clip': 12.22025203704834, 'timing/data': 0.005273254006169736, 'timing/model': 1.2720170429674909, 'learning_rate/qwen_vl_interface': 5.265138808534628e-06, 'learning_rate/action_model': 5.265138808534628e-05, 'epoch': 1.04}) +[2026-06-17 12:51:38,089][starVLA.training.train_starvla][INFO] - Step 3544, Loss: {'train/loss': 0.6100441217422485, 'train/grad_norm_pre_clip': 10.50869083404541, 'timing/data': 0.005175926955416799, 'timing/model': 1.203386734938249, 'learning_rate/qwen_vl_interface': 5.26297612742937e-06, 'learning_rate/action_model': 5.26297612742937e-05, 'epoch': 1.04}) +[2026-06-17 12:51:42,866][starVLA.training.train_starvla][INFO] - Step 3545, Loss: {'train/loss': 1.051028847694397, 'train/grad_norm_pre_clip': 12.99448013305664, 'timing/data': 0.005123254028148949, 'timing/model': 1.2043093390529975, 'learning_rate/qwen_vl_interface': 5.2608134436341476e-06, 'learning_rate/action_model': 5.260813443634147e-05, 'epoch': 1.04}) +[2026-06-17 12:51:47,655][starVLA.training.train_starvla][INFO] - Step 3546, Loss: {'train/loss': 0.6766698956489563, 'train/grad_norm_pre_clip': 11.485487937927246, 'timing/data': 0.005049749044701457, 'timing/model': 1.2055936359101906, 'learning_rate/qwen_vl_interface': 5.258650757597285e-06, 'learning_rate/action_model': 5.258650757597286e-05, 'epoch': 1.04}) +[2026-06-17 12:51:52,483][starVLA.training.train_starvla][INFO] - Step 3547, Loss: {'train/loss': 0.9050931930541992, 'train/grad_norm_pre_clip': 16.940671920776367, 'timing/data': 0.0051917629316449165, 'timing/model': 1.2043918269919232, 'learning_rate/qwen_vl_interface': 5.256488069767113e-06, 'learning_rate/action_model': 5.256488069767113e-05, 'epoch': 1.04}) +[2026-06-17 12:51:57,268][starVLA.training.train_starvla][INFO] - Step 3548, Loss: {'train/loss': 0.6317941546440125, 'train/grad_norm_pre_clip': 10.384477615356445, 'timing/data': 0.00522664503660053, 'timing/model': 1.3321932209655643, 'learning_rate/qwen_vl_interface': 5.254325380591955e-06, 'learning_rate/action_model': 5.254325380591955e-05, 'epoch': 1.04}) +[2026-06-17 12:52:02,001][starVLA.training.train_starvla][INFO] - Step 3549, Loss: {'train/loss': 0.6509838700294495, 'train/grad_norm_pre_clip': 10.672764778137207, 'timing/data': 0.005080924951471388, 'timing/model': 1.2825221220264211, 'learning_rate/qwen_vl_interface': 5.252162690520142e-06, 'learning_rate/action_model': 5.252162690520142e-05, 'epoch': 1.04}) +[2026-06-17 12:52:06,671][starVLA.training.train_starvla][INFO] - Step 3550, Loss: {'train/loss': 0.752258837223053, 'train/grad_norm_pre_clip': 12.714250564575195, 'timing/data': 0.005212439922615886, 'timing/model': 1.204014049959369, 'learning_rate/qwen_vl_interface': 5.2500000000000006e-06, 'learning_rate/action_model': 5.25e-05, 'epoch': 1.04}) +[2026-06-17 12:52:11,466][starVLA.training.train_starvla][INFO] - Step 3551, Loss: {'train/loss': 0.75015789270401, 'train/grad_norm_pre_clip': 10.59184455871582, 'timing/data': 0.005321788950823247, 'timing/model': 1.206719890004024, 'learning_rate/qwen_vl_interface': 5.24783730947986e-06, 'learning_rate/action_model': 5.2478373094798594e-05, 'epoch': 1.04}) +[2026-06-17 12:52:16,253][starVLA.training.train_starvla][INFO] - Step 3552, Loss: {'train/loss': 0.6056373119354248, 'train/grad_norm_pre_clip': 13.765715599060059, 'timing/data': 0.005112139042466879, 'timing/model': 1.20601787394844, 'learning_rate/qwen_vl_interface': 5.245674619408045e-06, 'learning_rate/action_model': 5.245674619408045e-05, 'epoch': 1.05}) +[2026-06-17 12:52:21,047][starVLA.training.train_starvla][INFO] - Step 3553, Loss: {'train/loss': 0.8804587125778198, 'train/grad_norm_pre_clip': 12.059710502624512, 'timing/data': 0.005860353005118668, 'timing/model': 1.2050973649602383, 'learning_rate/qwen_vl_interface': 5.243511930232889e-06, 'learning_rate/action_model': 5.243511930232888e-05, 'epoch': 1.05}) +[2026-06-17 12:52:25,821][starVLA.training.train_starvla][INFO] - Step 3554, Loss: {'train/loss': 0.5552340149879456, 'train/grad_norm_pre_clip': 12.27584457397461, 'timing/data': 0.005150696029886603, 'timing/model': 1.2041653619380668, 'learning_rate/qwen_vl_interface': 5.2413492424027165e-06, 'learning_rate/action_model': 5.241349242402717e-05, 'epoch': 1.05}) +[2026-06-17 12:52:30,567][starVLA.training.train_starvla][INFO] - Step 3555, Loss: {'train/loss': 0.6629733443260193, 'train/grad_norm_pre_clip': 9.590630531311035, 'timing/data': 0.005402427050285041, 'timing/model': 1.2914821229642257, 'learning_rate/qwen_vl_interface': 5.239186556365855e-06, 'learning_rate/action_model': 5.239186556365855e-05, 'epoch': 1.05}) +[2026-06-17 12:52:35,231][starVLA.training.train_starvla][INFO] - Step 3556, Loss: {'train/loss': 1.2555339336395264, 'train/grad_norm_pre_clip': 18.52075958251953, 'timing/data': 0.005474753095768392, 'timing/model': 1.207208473002538, 'learning_rate/qwen_vl_interface': 5.23702387257063e-06, 'learning_rate/action_model': 5.23702387257063e-05, 'epoch': 1.05}) +[2026-06-17 12:52:40,020][starVLA.training.train_starvla][INFO] - Step 3557, Loss: {'train/loss': 0.939336895942688, 'train/grad_norm_pre_clip': 12.154484748840332, 'timing/data': 0.005257242009975016, 'timing/model': 1.2060171130578965, 'learning_rate/qwen_vl_interface': 5.234861191465373e-06, 'learning_rate/action_model': 5.234861191465372e-05, 'epoch': 1.05}) +[2026-06-17 12:52:44,810][starVLA.training.train_starvla][INFO] - Step 3558, Loss: {'train/loss': 1.0896472930908203, 'train/grad_norm_pre_clip': 17.866775512695312, 'timing/data': 0.005295785027556121, 'timing/model': 1.204434369923547, 'learning_rate/qwen_vl_interface': 5.232698513498407e-06, 'learning_rate/action_model': 5.2326985134984066e-05, 'epoch': 1.05}) +[2026-06-17 12:52:49,707][starVLA.training.train_starvla][INFO] - Step 3559, Loss: {'train/loss': 0.6699821352958679, 'train/grad_norm_pre_clip': 11.430304527282715, 'timing/data': 0.005083562922663987, 'timing/model': 1.2045505020068958, 'learning_rate/qwen_vl_interface': 5.230535839118057e-06, 'learning_rate/action_model': 5.230535839118057e-05, 'epoch': 1.05}) +[2026-06-17 12:52:54,499][starVLA.training.train_starvla][INFO] - Step 3560, Loss: {'train/loss': 0.5144277215003967, 'train/grad_norm_pre_clip': 36.94020462036133, 'timing/data': 0.0048434900818392634, 'timing/model': 1.3355605800170451, 'learning_rate/qwen_vl_interface': 5.2283731687726455e-06, 'learning_rate/action_model': 5.2283731687726454e-05, 'epoch': 1.05}) +[2026-06-17 12:52:59,159][starVLA.training.train_starvla][INFO] - Step 3561, Loss: {'train/loss': 0.7393821477890015, 'train/grad_norm_pre_clip': 11.203742980957031, 'timing/data': 0.005295997951179743, 'timing/model': 1.2069952508900315, 'learning_rate/qwen_vl_interface': 5.226210502910504e-06, 'learning_rate/action_model': 5.226210502910504e-05, 'epoch': 1.05}) +[2026-06-17 12:53:03,959][starVLA.training.train_starvla][INFO] - Step 3562, Loss: {'train/loss': 0.6791109442710876, 'train/grad_norm_pre_clip': 10.75583553314209, 'timing/data': 0.004969937959685922, 'timing/model': 1.2067983619635925, 'learning_rate/qwen_vl_interface': 5.224047841979949e-06, 'learning_rate/action_model': 5.2240478419799485e-05, 'epoch': 1.05}) +[2026-06-17 12:53:08,751][starVLA.training.train_starvla][INFO] - Step 3563, Loss: {'train/loss': 0.8669320940971375, 'train/grad_norm_pre_clip': 12.82521915435791, 'timing/data': 0.005129455006681383, 'timing/model': 1.204907428007573, 'learning_rate/qwen_vl_interface': 5.2218851864293044e-06, 'learning_rate/action_model': 5.221885186429304e-05, 'epoch': 1.05}) +[2026-06-17 12:53:13,470][starVLA.training.train_starvla][INFO] - Step 3564, Loss: {'train/loss': 0.5099295377731323, 'train/grad_norm_pre_clip': 10.84409236907959, 'timing/data': 0.005339798983186483, 'timing/model': 1.209593644947745, 'learning_rate/qwen_vl_interface': 5.21972253670689e-06, 'learning_rate/action_model': 5.21972253670689e-05, 'epoch': 1.05}) +[2026-06-17 12:53:18,217][starVLA.training.train_starvla][INFO] - Step 3565, Loss: {'train/loss': 0.675771176815033, 'train/grad_norm_pre_clip': 9.890910148620605, 'timing/data': 0.005655050044879317, 'timing/model': 1.2052262499928474, 'learning_rate/qwen_vl_interface': 5.217559893261028e-06, 'learning_rate/action_model': 5.217559893261028e-05, 'epoch': 1.05}) +[2026-06-17 12:53:23,009][starVLA.training.train_starvla][INFO] - Step 3566, Loss: {'train/loss': 0.7045201659202576, 'train/grad_norm_pre_clip': 50.53160095214844, 'timing/data': 0.005410021054558456, 'timing/model': 1.3357622709590942, 'learning_rate/qwen_vl_interface': 5.2153972565400355e-06, 'learning_rate/action_model': 5.215397256540035e-05, 'epoch': 1.05}) +[2026-06-17 12:53:27,678][starVLA.training.train_starvla][INFO] - Step 3567, Loss: {'train/loss': 0.6955738663673401, 'train/grad_norm_pre_clip': 8.673507690429688, 'timing/data': 0.0051470930920913815, 'timing/model': 1.2096725010778755, 'learning_rate/qwen_vl_interface': 5.213234626992227e-06, 'learning_rate/action_model': 5.213234626992226e-05, 'epoch': 1.05}) +[2026-06-17 12:53:32,480][starVLA.training.train_starvla][INFO] - Step 3568, Loss: {'train/loss': 0.5907867550849915, 'train/grad_norm_pre_clip': 9.550901412963867, 'timing/data': 0.005112007958814502, 'timing/model': 1.2082048760494217, 'learning_rate/qwen_vl_interface': 5.21107200506592e-06, 'learning_rate/action_model': 5.21107200506592e-05, 'epoch': 1.05}) +[2026-06-17 12:53:37,274][starVLA.training.train_starvla][INFO] - Step 3569, Loss: {'train/loss': 0.7727969884872437, 'train/grad_norm_pre_clip': 18.129684448242188, 'timing/data': 0.005168681964278221, 'timing/model': 1.2110378900542855, 'learning_rate/qwen_vl_interface': 5.208909391209429e-06, 'learning_rate/action_model': 5.208909391209429e-05, 'epoch': 1.05}) +[2026-06-17 12:53:42,069][starVLA.training.train_starvla][INFO] - Step 3570, Loss: {'train/loss': 0.8718468546867371, 'train/grad_norm_pre_clip': 13.072283744812012, 'timing/data': 0.005337066948413849, 'timing/model': 1.204921381897293, 'learning_rate/qwen_vl_interface': 5.206746785871066e-06, 'learning_rate/action_model': 5.2067467858710664e-05, 'epoch': 1.05}) +[2026-06-17 12:53:46,825][starVLA.training.train_starvla][INFO] - Step 3571, Loss: {'train/loss': 0.5690425038337708, 'train/grad_norm_pre_clip': 8.116158485412598, 'timing/data': 0.005425384966656566, 'timing/model': 1.207369062001817, 'learning_rate/qwen_vl_interface': 5.204584189499139e-06, 'learning_rate/action_model': 5.2045841894991385e-05, 'epoch': 1.05}) +[2026-06-17 12:53:51,608][starVLA.training.train_starvla][INFO] - Step 3572, Loss: {'train/loss': 0.5961468815803528, 'train/grad_norm_pre_clip': 11.839993476867676, 'timing/data': 0.004935450037010014, 'timing/model': 1.3245562579249963, 'learning_rate/qwen_vl_interface': 5.202421602541959e-06, 'learning_rate/action_model': 5.2024216025419594e-05, 'epoch': 1.05}) +[2026-06-17 12:53:56,265][starVLA.training.train_starvla][INFO] - Step 3573, Loss: {'train/loss': 0.6731691956520081, 'train/grad_norm_pre_clip': 9.938423156738281, 'timing/data': 0.005156330997124314, 'timing/model': 1.2050217350479215, 'learning_rate/qwen_vl_interface': 5.200259025447835e-06, 'learning_rate/action_model': 5.2002590254478345e-05, 'epoch': 1.05}) +[2026-06-17 12:54:01,054][starVLA.training.train_starvla][INFO] - Step 3574, Loss: {'train/loss': 0.5607011318206787, 'train/grad_norm_pre_clip': 10.023119926452637, 'timing/data': 0.00495801295619458, 'timing/model': 1.206596453092061, 'learning_rate/qwen_vl_interface': 5.198096458665064e-06, 'learning_rate/action_model': 5.198096458665064e-05, 'epoch': 1.05}) +[2026-06-17 12:54:05,852][starVLA.training.train_starvla][INFO] - Step 3575, Loss: {'train/loss': 0.5877423286437988, 'train/grad_norm_pre_clip': 8.880193710327148, 'timing/data': 0.005456520011648536, 'timing/model': 1.205678096972406, 'learning_rate/qwen_vl_interface': 5.195933902641955e-06, 'learning_rate/action_model': 5.195933902641955e-05, 'epoch': 1.05}) +[2026-06-17 12:54:10,664][starVLA.training.train_starvla][INFO] - Step 3576, Loss: {'train/loss': 0.7759044766426086, 'train/grad_norm_pre_clip': 12.870719909667969, 'timing/data': 0.005475368001498282, 'timing/model': 1.208102182019502, 'learning_rate/qwen_vl_interface': 5.193771357826807e-06, 'learning_rate/action_model': 5.193771357826806e-05, 'epoch': 1.05}) +[2026-06-17 12:54:15,454][starVLA.training.train_starvla][INFO] - Step 3577, Loss: {'train/loss': 0.6083559989929199, 'train/grad_norm_pre_clip': 10.167421340942383, 'timing/data': 0.005932137952186167, 'timing/model': 1.3316751540405676, 'learning_rate/qwen_vl_interface': 5.191608824667916e-06, 'learning_rate/action_model': 5.1916088246679165e-05, 'epoch': 1.05}) +[2026-06-17 12:54:20,113][starVLA.training.train_starvla][INFO] - Step 3578, Loss: {'train/loss': 1.0479013919830322, 'train/grad_norm_pre_clip': 16.424423217773438, 'timing/data': 0.004987680003978312, 'timing/model': 1.204231601092033, 'learning_rate/qwen_vl_interface': 5.189446303613578e-06, 'learning_rate/action_model': 5.1894463036135775e-05, 'epoch': 1.05}) +[2026-06-17 12:54:24,901][starVLA.training.train_starvla][INFO] - Step 3579, Loss: {'train/loss': 0.6364701390266418, 'train/grad_norm_pre_clip': 11.906622886657715, 'timing/data': 0.00504460500087589, 'timing/model': 1.205985078937374, 'learning_rate/qwen_vl_interface': 5.187283795112087e-06, 'learning_rate/action_model': 5.1872837951120866e-05, 'epoch': 1.05}) +[2026-06-17 12:54:29,620][starVLA.training.train_starvla][INFO] - Step 3580, Loss: {'train/loss': 0.5858278870582581, 'train/grad_norm_pre_clip': 9.302027702331543, 'timing/data': 0.005181677057407796, 'timing/model': 1.2125324269291013, 'learning_rate/qwen_vl_interface': 5.185121299611734e-06, 'learning_rate/action_model': 5.1851212996117335e-05, 'epoch': 1.05}) +[2026-06-17 12:54:34,358][starVLA.training.train_starvla][INFO] - Step 3581, Loss: {'train/loss': 0.46442896127700806, 'train/grad_norm_pre_clip': 8.161018371582031, 'timing/data': 0.005264565930701792, 'timing/model': 1.2065836979309097, 'learning_rate/qwen_vl_interface': 5.182958817560803e-06, 'learning_rate/action_model': 5.1829588175608026e-05, 'epoch': 1.05}) +[2026-06-17 12:54:39,156][starVLA.training.train_starvla][INFO] - Step 3582, Loss: {'train/loss': 0.8897804617881775, 'train/grad_norm_pre_clip': 13.343481063842773, 'timing/data': 0.005356842069886625, 'timing/model': 1.2067686599912122, 'learning_rate/qwen_vl_interface': 5.180796349407583e-06, 'learning_rate/action_model': 5.180796349407583e-05, 'epoch': 1.05}) +[2026-06-17 12:54:43,952][starVLA.training.train_starvla][INFO] - Step 3583, Loss: {'train/loss': 0.6597574949264526, 'train/grad_norm_pre_clip': 12.362923622131348, 'timing/data': 0.005443981965072453, 'timing/model': 1.3341448560822755, 'learning_rate/qwen_vl_interface': 5.178633895600354e-06, 'learning_rate/action_model': 5.178633895600353e-05, 'epoch': 1.05}) +[2026-06-17 12:54:48,622][starVLA.training.train_starvla][INFO] - Step 3584, Loss: {'train/loss': 0.5197975039482117, 'train/grad_norm_pre_clip': 8.850966453552246, 'timing/data': 0.005180376931093633, 'timing/model': 1.2094468580326065, 'learning_rate/qwen_vl_interface': 5.176471456587394e-06, 'learning_rate/action_model': 5.176471456587394e-05, 'epoch': 1.05}) +[2026-06-17 12:54:53,412][starVLA.training.train_starvla][INFO] - Step 3585, Loss: {'train/loss': 0.6303305625915527, 'train/grad_norm_pre_clip': 9.830416679382324, 'timing/data': 0.004959885962307453, 'timing/model': 1.2062096209265292, 'learning_rate/qwen_vl_interface': 5.1743090328169805e-06, 'learning_rate/action_model': 5.1743090328169805e-05, 'epoch': 1.05}) +[2026-06-17 12:54:58,212][starVLA.training.train_starvla][INFO] - Step 3586, Loss: {'train/loss': 0.716585099697113, 'train/grad_norm_pre_clip': 12.5231351852417, 'timing/data': 0.0054721609922125936, 'timing/model': 1.2108589640120044, 'learning_rate/qwen_vl_interface': 5.172146624737388e-06, 'learning_rate/action_model': 5.1721466247373875e-05, 'epoch': 1.06}) +[2026-06-17 12:55:02,964][starVLA.training.train_starvla][INFO] - Step 3587, Loss: {'train/loss': 0.8662613034248352, 'train/grad_norm_pre_clip': 14.858101844787598, 'timing/data': 0.0054228350054472685, 'timing/model': 1.2082469330634922, 'learning_rate/qwen_vl_interface': 5.169984232796883e-06, 'learning_rate/action_model': 5.169984232796883e-05, 'epoch': 1.06}) +[2026-06-17 12:55:07,747][starVLA.training.train_starvla][INFO] - Step 3588, Loss: {'train/loss': 0.619098424911499, 'train/grad_norm_pre_clip': 12.083606719970703, 'timing/data': 0.005581473931670189, 'timing/model': 1.2071224859682843, 'learning_rate/qwen_vl_interface': 5.1678218574437325e-06, 'learning_rate/action_model': 5.167821857443732e-05, 'epoch': 1.06}) +[2026-06-17 12:55:12,541][starVLA.training.train_starvla][INFO] - Step 3589, Loss: {'train/loss': 0.8931556344032288, 'train/grad_norm_pre_clip': 10.906303405761719, 'timing/data': 0.005415435996837914, 'timing/model': 1.3369561929721385, 'learning_rate/qwen_vl_interface': 5.165659499126198e-06, 'learning_rate/action_model': 5.165659499126198e-05, 'epoch': 1.06}) +[2026-06-17 12:55:17,205][starVLA.training.train_starvla][INFO] - Step 3590, Loss: {'train/loss': 0.6647008657455444, 'train/grad_norm_pre_clip': 15.411001205444336, 'timing/data': 0.00536846905015409, 'timing/model': 1.2097229739883915, 'learning_rate/qwen_vl_interface': 5.1634971582925405e-06, 'learning_rate/action_model': 5.163497158292541e-05, 'epoch': 1.06}) +[2026-06-17 12:55:22,015][starVLA.training.train_starvla][INFO] - Step 3591, Loss: {'train/loss': 0.7719166278839111, 'train/grad_norm_pre_clip': 12.30844783782959, 'timing/data': 0.0059292789082974195, 'timing/model': 1.207257094909437, 'learning_rate/qwen_vl_interface': 5.161334835391016e-06, 'learning_rate/action_model': 5.161334835391016e-05, 'epoch': 1.06}) +[2026-06-17 12:55:26,821][starVLA.training.train_starvla][INFO] - Step 3592, Loss: {'train/loss': 0.7139001488685608, 'train/grad_norm_pre_clip': 13.380733489990234, 'timing/data': 0.005075125023722649, 'timing/model': 1.204559849924408, 'learning_rate/qwen_vl_interface': 5.159172530869875e-06, 'learning_rate/action_model': 5.159172530869874e-05, 'epoch': 1.06}) +[2026-06-17 12:55:31,612][starVLA.training.train_starvla][INFO] - Step 3593, Loss: {'train/loss': 0.7045516967773438, 'train/grad_norm_pre_clip': 14.919215202331543, 'timing/data': 0.005117050022818148, 'timing/model': 1.2064633809495717, 'learning_rate/qwen_vl_interface': 5.157010245177366e-06, 'learning_rate/action_model': 5.1570102451773664e-05, 'epoch': 1.06}) +[2026-06-17 12:55:36,297][starVLA.training.train_starvla][INFO] - Step 3594, Loss: {'train/loss': 0.6914952993392944, 'train/grad_norm_pre_clip': 9.18942928314209, 'timing/data': 0.005155847989954054, 'timing/model': 1.2120658999774605, 'learning_rate/qwen_vl_interface': 5.1548479787617346e-06, 'learning_rate/action_model': 5.1548479787617346e-05, 'epoch': 1.06}) +[2026-06-17 12:55:41,006][starVLA.training.train_starvla][INFO] - Step 3595, Loss: {'train/loss': 0.5708452463150024, 'train/grad_norm_pre_clip': 9.880112648010254, 'timing/data': 0.004930976079776883, 'timing/model': 1.2503299650270492, 'learning_rate/qwen_vl_interface': 5.152685732071219e-06, 'learning_rate/action_model': 5.1526857320712185e-05, 'epoch': 1.06}) +[2026-06-17 12:55:45,668][starVLA.training.train_starvla][INFO] - Step 3596, Loss: {'train/loss': 0.5912892818450928, 'train/grad_norm_pre_clip': 12.316730499267578, 'timing/data': 0.005277844960801303, 'timing/model': 1.2059592489385977, 'learning_rate/qwen_vl_interface': 5.1505235055540545e-06, 'learning_rate/action_model': 5.1505235055540545e-05, 'epoch': 1.06}) +[2026-06-17 12:55:50,402][starVLA.training.train_starvla][INFO] - Step 3597, Loss: {'train/loss': 0.6847997903823853, 'train/grad_norm_pre_clip': 13.653456687927246, 'timing/data': 0.005276267067529261, 'timing/model': 1.2042983659775928, 'learning_rate/qwen_vl_interface': 5.148361299658477e-06, 'learning_rate/action_model': 5.148361299658476e-05, 'epoch': 1.06}) +[2026-06-17 12:55:55,191][starVLA.training.train_starvla][INFO] - Step 3598, Loss: {'train/loss': 0.7399548292160034, 'train/grad_norm_pre_clip': 11.393622398376465, 'timing/data': 0.0051396689377725124, 'timing/model': 1.205036285915412, 'learning_rate/qwen_vl_interface': 5.1461991148327095e-06, 'learning_rate/action_model': 5.1461991148327095e-05, 'epoch': 1.06}) +[2026-06-17 12:55:59,985][starVLA.training.train_starvla][INFO] - Step 3599, Loss: {'train/loss': 0.6066887974739075, 'train/grad_norm_pre_clip': 14.117696762084961, 'timing/data': 0.005285336985252798, 'timing/model': 1.209284820011817, 'learning_rate/qwen_vl_interface': 5.144036951524977e-06, 'learning_rate/action_model': 5.144036951524977e-05, 'epoch': 1.06}) +[2026-06-17 12:57:41,779][starVLA.training.train_starvla][INFO] - Step 3600, Loss: {'train/loss': 0.6435831189155579, 'train/grad_norm_pre_clip': 10.160106658935547, 'eval/loss': 0.7788996696472168, 'eval/latency_0/loss': 0.7788996696472168, 'eval/demon_fire_precision': 0.9057461761058289, 'eval/demon_fire_recall': 0.9098837209302325, 'eval/demon_fire_f1': 0.907810234099855, 'eval/demon_fire_rate_model': 0.6726918798665183, 'eval/demon_fire_rate_teacher': 0.6696329254727474, 'eval/demon_left_precision': 0.848692403486924, 'eval/demon_left_recall': 0.7722379603399433, 'eval/demon_left_f1': 0.8086621180658558, 'eval/demon_left_rate_model': 0.44660734149054504, 'eval/demon_left_rate_teacher': 0.49082313681868744, 'eval/demon_right_precision': 0.7841654059505799, 'eval/demon_right_recall': 0.8785310734463276, 'eval/demon_right_f1': 0.8286703970157208, 'eval/demon_right_rate_model': 0.5514460511679644, 'eval/demon_right_rate_teacher': 0.4922135706340378, 'eval/demon_move_f1': 0.8186662575407883, 'eval/demon_cc_f1': 0.8790930908812636, 'eval/latency_0/demon_fire_precision': 0.9057461761058289, 'eval/latency_0/demon_fire_recall': 0.9098837209302325, 'eval/latency_0/demon_fire_f1': 0.907810234099855, 'eval/latency_0/demon_fire_rate_model': 0.6726918798665183, 'eval/latency_0/demon_fire_rate_teacher': 0.6696329254727474, 'eval/latency_0/demon_left_precision': 0.848692403486924, 'eval/latency_0/demon_left_recall': 0.7722379603399433, 'eval/latency_0/demon_left_f1': 0.8086621180658558, 'eval/latency_0/demon_left_rate_model': 0.44660734149054504, 'eval/latency_0/demon_left_rate_teacher': 0.49082313681868744, 'eval/latency_0/demon_right_precision': 0.7841654059505799, 'eval/latency_0/demon_right_recall': 0.8785310734463276, 'eval/latency_0/demon_right_f1': 0.8286703970157208, 'eval/latency_0/demon_right_rate_model': 0.5514460511679644, 'eval/latency_0/demon_right_rate_teacher': 0.4922135706340378, 'eval/latency_0/demon_move_f1': 0.8186662575407883, 'eval/latency_0/demon_cc_f1': 0.8790930908812636, 'timing/data': 0.00541213306132704, 'timing/model': 1.3377575379563496, 'learning_rate/qwen_vl_interface': 5.141874810183498e-06, 'learning_rate/action_model': 5.141874810183498e-05, 'epoch': 1.06}) +[2026-06-17 12:57:41,780][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_3600_state +[2026-06-17 12:57:41,782][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 12:57:45,332][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_3600_state/model.safetensors +[2026-06-17 12:57:52,980][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_3600_state/optimizer.bin +[2026-06-17 12:57:52,981][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_3600_state/sampler.bin +[2026-06-17 12:57:52,981][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_3600_state/sampler_1.bin +[2026-06-17 12:57:52,982][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_3600_state/random_states_0.pkl +[2026-06-17 12:57:53,027][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3600 +[2026-06-17 12:57:53,027][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 12:57:53,038][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 12:57:57,764][starVLA.training.train_starvla][INFO] - Step 3601, Loss: {'train/loss': 1.0267144441604614, 'train/grad_norm_pre_clip': 14.401704788208008, 'timing/data': 0.004960143007338047, 'timing/model': 1.1887605199590325, 'learning_rate/qwen_vl_interface': 5.139712691256489e-06, 'learning_rate/action_model': 5.1397126912564885e-05, 'epoch': 1.06}) +[2026-06-17 12:58:02,486][starVLA.training.train_starvla][INFO] - Step 3602, Loss: {'train/loss': 0.8347026705741882, 'train/grad_norm_pre_clip': 12.235401153564453, 'timing/data': 0.005383925046771765, 'timing/model': 1.3158816789509729, 'learning_rate/qwen_vl_interface': 5.137550595192156e-06, 'learning_rate/action_model': 5.137550595192156e-05, 'epoch': 1.06}) +[2026-06-17 12:58:07,088][starVLA.training.train_starvla][INFO] - Step 3603, Loss: {'train/loss': 0.4718896448612213, 'train/grad_norm_pre_clip': 9.986135482788086, 'timing/data': 0.0050769009394571185, 'timing/model': 1.1913386079249904, 'learning_rate/qwen_vl_interface': 5.135388522438707e-06, 'learning_rate/action_model': 5.1353885224387066e-05, 'epoch': 1.06}) +[2026-06-17 12:58:11,831][starVLA.training.train_starvla][INFO] - Step 3604, Loss: {'train/loss': 0.865930438041687, 'train/grad_norm_pre_clip': 13.274697303771973, 'timing/data': 0.0052027839701622725, 'timing/model': 1.1926848539151251, 'learning_rate/qwen_vl_interface': 5.133226473444341e-06, 'learning_rate/action_model': 5.13322647344434e-05, 'epoch': 1.06}) +[2026-06-17 12:58:16,536][starVLA.training.train_starvla][INFO] - Step 3605, Loss: {'train/loss': 0.5262832045555115, 'train/grad_norm_pre_clip': 20.13201141357422, 'timing/data': 0.005151132005266845, 'timing/model': 1.1961679949890822, 'learning_rate/qwen_vl_interface': 5.1310644486572515e-06, 'learning_rate/action_model': 5.131064448657251e-05, 'epoch': 1.06}) +[2026-06-17 12:58:21,302][starVLA.training.train_starvla][INFO] - Step 3606, Loss: {'train/loss': 0.7504165768623352, 'train/grad_norm_pre_clip': 12.34638786315918, 'timing/data': 0.005036914953961968, 'timing/model': 1.1963177449069917, 'learning_rate/qwen_vl_interface': 5.12890244852563e-06, 'learning_rate/action_model': 5.12890244852563e-05, 'epoch': 1.06}) +[2026-06-17 12:58:26,060][starVLA.training.train_starvla][INFO] - Step 3607, Loss: {'train/loss': 0.8211400508880615, 'train/grad_norm_pre_clip': 13.969232559204102, 'timing/data': 0.005317319999448955, 'timing/model': 1.199521475005895, 'learning_rate/qwen_vl_interface': 5.126740473497661e-06, 'learning_rate/action_model': 5.126740473497661e-05, 'epoch': 1.06}) +[2026-06-17 12:58:30,817][starVLA.training.train_starvla][INFO] - Step 3608, Loss: {'train/loss': 0.7023835182189941, 'train/grad_norm_pre_clip': 11.857126235961914, 'timing/data': 0.005346294026821852, 'timing/model': 1.3247416659723967, 'learning_rate/qwen_vl_interface': 5.124578524021525e-06, 'learning_rate/action_model': 5.124578524021525e-05, 'epoch': 1.06}) +[2026-06-17 12:58:35,569][starVLA.training.train_starvla][INFO] - Step 3609, Loss: {'train/loss': 0.6652567386627197, 'train/grad_norm_pre_clip': 12.707687377929688, 'timing/data': 0.004896119004115462, 'timing/model': 1.3200262320460752, 'learning_rate/qwen_vl_interface': 5.122416600545396e-06, 'learning_rate/action_model': 5.122416600545397e-05, 'epoch': 1.06}) +[2026-06-17 12:58:40,201][starVLA.training.train_starvla][INFO] - Step 3610, Loss: {'train/loss': 0.5654659867286682, 'train/grad_norm_pre_clip': 11.88433837890625, 'timing/data': 0.004983405000530183, 'timing/model': 1.1991650779964402, 'learning_rate/qwen_vl_interface': 5.1202547035174436e-06, 'learning_rate/action_model': 5.120254703517443e-05, 'epoch': 1.06}) +[2026-06-17 12:58:44,962][starVLA.training.train_starvla][INFO] - Step 3611, Loss: {'train/loss': 0.47262364625930786, 'train/grad_norm_pre_clip': 11.357342720031738, 'timing/data': 0.007330929976888001, 'timing/model': 1.214030436007306, 'learning_rate/qwen_vl_interface': 5.118092833385829e-06, 'learning_rate/action_model': 5.1180928333858293e-05, 'epoch': 1.06}) +[2026-06-17 12:58:49,728][starVLA.training.train_starvla][INFO] - Step 3612, Loss: {'train/loss': 0.8256199955940247, 'train/grad_norm_pre_clip': 10.890411376953125, 'timing/data': 0.005324393045157194, 'timing/model': 1.1977735219988972, 'learning_rate/qwen_vl_interface': 5.115930990598713e-06, 'learning_rate/action_model': 5.115930990598713e-05, 'epoch': 1.06}) +[2026-06-17 12:58:54,502][starVLA.training.train_starvla][INFO] - Step 3613, Loss: {'train/loss': 0.9418861269950867, 'train/grad_norm_pre_clip': 16.986425399780273, 'timing/data': 0.005178963067010045, 'timing/model': 1.1985411760397255, 'learning_rate/qwen_vl_interface': 5.113769175604248e-06, 'learning_rate/action_model': 5.113769175604247e-05, 'epoch': 1.06}) +[2026-06-17 12:58:59,219][starVLA.training.train_starvla][INFO] - Step 3614, Loss: {'train/loss': 0.7468515038490295, 'train/grad_norm_pre_clip': 15.56479263305664, 'timing/data': 0.0048274260479956865, 'timing/model': 1.2708289959700778, 'learning_rate/qwen_vl_interface': 5.111607388850576e-06, 'learning_rate/action_model': 5.111607388850576e-05, 'epoch': 1.06}) +[2026-06-17 12:59:03,868][starVLA.training.train_starvla][INFO] - Step 3615, Loss: {'train/loss': 0.9560278654098511, 'train/grad_norm_pre_clip': 12.545212745666504, 'timing/data': 0.00485245103482157, 'timing/model': 1.203527444973588, 'learning_rate/qwen_vl_interface': 5.1094456307858444e-06, 'learning_rate/action_model': 5.109445630785844e-05, 'epoch': 1.06}) +[2026-06-17 12:59:08,630][starVLA.training.train_starvla][INFO] - Step 3616, Loss: {'train/loss': 0.7832049131393433, 'train/grad_norm_pre_clip': 18.242523193359375, 'timing/data': 0.005365688935853541, 'timing/model': 1.2011832379503176, 'learning_rate/qwen_vl_interface': 5.107283901858184e-06, 'learning_rate/action_model': 5.107283901858184e-05, 'epoch': 1.06}) +[2026-06-17 12:59:13,407][starVLA.training.train_starvla][INFO] - Step 3617, Loss: {'train/loss': 0.6723875403404236, 'train/grad_norm_pre_clip': 12.216931343078613, 'timing/data': 0.0050617080414667726, 'timing/model': 1.2037510019727051, 'learning_rate/qwen_vl_interface': 5.105122202515726e-06, 'learning_rate/action_model': 5.1051222025157255e-05, 'epoch': 1.06}) +[2026-06-17 12:59:18,181][starVLA.training.train_starvla][INFO] - Step 3618, Loss: {'train/loss': 0.7883821725845337, 'train/grad_norm_pre_clip': 14.791375160217285, 'timing/data': 0.005435052909888327, 'timing/model': 1.2034222070360556, 'learning_rate/qwen_vl_interface': 5.102960533206588e-06, 'learning_rate/action_model': 5.1029605332065874e-05, 'epoch': 1.06}) +[2026-06-17 12:59:22,964][starVLA.training.train_starvla][INFO] - Step 3619, Loss: {'train/loss': 0.6034796237945557, 'train/grad_norm_pre_clip': 10.651464462280273, 'timing/data': 0.00497498398181051, 'timing/model': 1.3298606859752908, 'learning_rate/qwen_vl_interface': 5.100798894378891e-06, 'learning_rate/action_model': 5.100798894378891e-05, 'epoch': 1.06}) +[2026-06-17 12:59:27,683][starVLA.training.train_starvla][INFO] - Step 3620, Loss: {'train/loss': 0.5790867209434509, 'train/grad_norm_pre_clip': 11.3223876953125, 'timing/data': 0.005244137020781636, 'timing/model': 1.2746888539986685, 'learning_rate/qwen_vl_interface': 5.098637286480744e-06, 'learning_rate/action_model': 5.098637286480744e-05, 'epoch': 1.07}) +[2026-06-17 12:59:32,414][starVLA.training.train_starvla][INFO] - Step 3621, Loss: {'train/loss': 0.7923214435577393, 'train/grad_norm_pre_clip': 14.014313697814941, 'timing/data': 0.005128268036060035, 'timing/model': 1.2069143049884588, 'learning_rate/qwen_vl_interface': 5.096475709960252e-06, 'learning_rate/action_model': 5.096475709960251e-05, 'epoch': 1.07}) +[2026-06-17 12:59:37,215][starVLA.training.train_starvla][INFO] - Step 3622, Loss: {'train/loss': 0.6179569363594055, 'train/grad_norm_pre_clip': 9.626842498779297, 'timing/data': 0.005291073000989854, 'timing/model': 1.2045915080234408, 'learning_rate/qwen_vl_interface': 5.094314165265506e-06, 'learning_rate/action_model': 5.0943141652655056e-05, 'epoch': 1.07}) +[2026-06-17 12:59:42,154][starVLA.training.train_starvla][INFO] - Step 3623, Loss: {'train/loss': 0.695120632648468, 'train/grad_norm_pre_clip': 11.84242057800293, 'timing/data': 0.004969291971065104, 'timing/model': 1.2097253309329972, 'learning_rate/qwen_vl_interface': 5.092152652844606e-06, 'learning_rate/action_model': 5.092152652844606e-05, 'epoch': 1.07}) +[2026-06-17 12:59:46,943][starVLA.training.train_starvla][INFO] - Step 3624, Loss: {'train/loss': 0.7236989140510559, 'train/grad_norm_pre_clip': 11.347297668457031, 'timing/data': 0.005138557055033743, 'timing/model': 1.2042631519725546, 'learning_rate/qwen_vl_interface': 5.089991173145628e-06, 'learning_rate/action_model': 5.0899911731456285e-05, 'epoch': 1.07}) +[2026-06-17 12:59:51,715][starVLA.training.train_starvla][INFO] - Step 3625, Loss: {'train/loss': 0.5986047983169556, 'train/grad_norm_pre_clip': 10.88379192352295, 'timing/data': 0.00528629298787564, 'timing/model': 1.2036017220234498, 'learning_rate/qwen_vl_interface': 5.087829726616653e-06, 'learning_rate/action_model': 5.087829726616653e-05, 'epoch': 1.07}) +[2026-06-17 12:59:56,455][starVLA.training.train_starvla][INFO] - Step 3626, Loss: {'train/loss': 0.835480272769928, 'train/grad_norm_pre_clip': 12.69990348815918, 'timing/data': 0.005042334902100265, 'timing/model': 1.2872487669810653, 'learning_rate/qwen_vl_interface': 5.08566831370575e-06, 'learning_rate/action_model': 5.0856683137057504e-05, 'epoch': 1.07}) +[2026-06-17 13:00:01,121][starVLA.training.train_starvla][INFO] - Step 3627, Loss: {'train/loss': 0.7274039387702942, 'train/grad_norm_pre_clip': 23.31095314025879, 'timing/data': 0.005672302097082138, 'timing/model': 1.2096710440237075, 'learning_rate/qwen_vl_interface': 5.083506934860982e-06, 'learning_rate/action_model': 5.083506934860982e-05, 'epoch': 1.07}) +[2026-06-17 13:00:05,907][starVLA.training.train_starvla][INFO] - Step 3628, Loss: {'train/loss': 0.6513476967811584, 'train/grad_norm_pre_clip': 12.025286674499512, 'timing/data': 0.005129135912284255, 'timing/model': 1.2060346259968355, 'learning_rate/qwen_vl_interface': 5.081345590530407e-06, 'learning_rate/action_model': 5.081345590530407e-05, 'epoch': 1.07}) +[2026-06-17 13:00:10,703][starVLA.training.train_starvla][INFO] - Step 3629, Loss: {'train/loss': 0.9484066963195801, 'train/grad_norm_pre_clip': 12.649531364440918, 'timing/data': 0.005348465987481177, 'timing/model': 1.204557367018424, 'learning_rate/qwen_vl_interface': 5.079184281162069e-06, 'learning_rate/action_model': 5.0791842811620694e-05, 'epoch': 1.07}) +[2026-06-17 13:00:15,489][starVLA.training.train_starvla][INFO] - Step 3630, Loss: {'train/loss': 0.707421600818634, 'train/grad_norm_pre_clip': 9.580953598022461, 'timing/data': 0.005278981989249587, 'timing/model': 1.2025826929602772, 'learning_rate/qwen_vl_interface': 5.077023007204018e-06, 'learning_rate/action_model': 5.077023007204018e-05, 'epoch': 1.07}) +[2026-06-17 13:00:20,270][starVLA.training.train_starvla][INFO] - Step 3631, Loss: {'train/loss': 0.6601549983024597, 'train/grad_norm_pre_clip': 10.349870681762695, 'timing/data': 0.005309115047566593, 'timing/model': 1.3324903988977894, 'learning_rate/qwen_vl_interface': 5.074861769104281e-06, 'learning_rate/action_model': 5.074861769104281e-05, 'epoch': 1.07}) +[2026-06-17 13:00:24,931][starVLA.training.train_starvla][INFO] - Step 3632, Loss: {'train/loss': 0.6641876697540283, 'train/grad_norm_pre_clip': 10.099027633666992, 'timing/data': 0.005099494941532612, 'timing/model': 1.2019706059945747, 'learning_rate/qwen_vl_interface': 5.07270056731089e-06, 'learning_rate/action_model': 5.07270056731089e-05, 'epoch': 1.07}) +[2026-06-17 13:00:29,725][starVLA.training.train_starvla][INFO] - Step 3633, Loss: {'train/loss': 0.6447126269340515, 'train/grad_norm_pre_clip': 11.393169403076172, 'timing/data': 0.005164760979823768, 'timing/model': 1.2046572300605476, 'learning_rate/qwen_vl_interface': 5.070539402271859e-06, 'learning_rate/action_model': 5.070539402271859e-05, 'epoch': 1.07}) +[2026-06-17 13:00:34,518][starVLA.training.train_starvla][INFO] - Step 3634, Loss: {'train/loss': 0.9917057156562805, 'train/grad_norm_pre_clip': 12.170013427734375, 'timing/data': 0.005013293935917318, 'timing/model': 1.2050817760173231, 'learning_rate/qwen_vl_interface': 5.068378274435205e-06, 'learning_rate/action_model': 5.068378274435205e-05, 'epoch': 1.07}) +[2026-06-17 13:00:39,249][starVLA.training.train_starvla][INFO] - Step 3635, Loss: {'train/loss': 0.5925471186637878, 'train/grad_norm_pre_clip': 10.115970611572266, 'timing/data': 0.005732888006605208, 'timing/model': 1.2329022589838132, 'learning_rate/qwen_vl_interface': 5.066217184248931e-06, 'learning_rate/action_model': 5.0662171842489305e-05, 'epoch': 1.07}) +[2026-06-17 13:00:43,992][starVLA.training.train_starvla][INFO] - Step 3636, Loss: {'train/loss': 0.7033487558364868, 'train/grad_norm_pre_clip': 11.56815242767334, 'timing/data': 0.005240102065727115, 'timing/model': 1.2038784199394286, 'learning_rate/qwen_vl_interface': 5.0640561321610295e-06, 'learning_rate/action_model': 5.064056132161029e-05, 'epoch': 1.07}) +[2026-06-17 13:00:48,790][starVLA.training.train_starvla][INFO] - Step 3637, Loss: {'train/loss': 1.1809948682785034, 'train/grad_norm_pre_clip': 19.46076202392578, 'timing/data': 0.005082524032332003, 'timing/model': 1.3438766859471798, 'learning_rate/qwen_vl_interface': 5.061895118619493e-06, 'learning_rate/action_model': 5.061895118619493e-05, 'epoch': 1.07}) +[2026-06-17 13:00:53,453][starVLA.training.train_starvla][INFO] - Step 3638, Loss: {'train/loss': 0.6869943737983704, 'train/grad_norm_pre_clip': 10.430829048156738, 'timing/data': 0.004673762945458293, 'timing/model': 1.2077608230756596, 'learning_rate/qwen_vl_interface': 5.059734144072302e-06, 'learning_rate/action_model': 5.059734144072302e-05, 'epoch': 1.07}) +[2026-06-17 13:00:58,240][starVLA.training.train_starvla][INFO] - Step 3639, Loss: {'train/loss': 0.5029418468475342, 'train/grad_norm_pre_clip': 8.454627990722656, 'timing/data': 0.00532689003739506, 'timing/model': 1.2046044819289818, 'learning_rate/qwen_vl_interface': 5.0575732089674275e-06, 'learning_rate/action_model': 5.057573208967428e-05, 'epoch': 1.07}) +[2026-06-17 13:01:03,052][starVLA.training.train_starvla][INFO] - Step 3640, Loss: {'train/loss': 0.7595343589782715, 'train/grad_norm_pre_clip': 11.960997581481934, 'timing/data': 0.00526098592672497, 'timing/model': 1.209814204950817, 'learning_rate/qwen_vl_interface': 5.055412313752832e-06, 'learning_rate/action_model': 5.055412313752832e-05, 'epoch': 1.07}) +[2026-06-17 13:01:07,850][starVLA.training.train_starvla][INFO] - Step 3641, Loss: {'train/loss': 0.7153323888778687, 'train/grad_norm_pre_clip': 13.696198463439941, 'timing/data': 0.005428706994280219, 'timing/model': 1.2098969989456236, 'learning_rate/qwen_vl_interface': 5.053251458876474e-06, 'learning_rate/action_model': 5.0532514588764744e-05, 'epoch': 1.07}) +[2026-06-17 13:01:12,596][starVLA.training.train_starvla][INFO] - Step 3642, Loss: {'train/loss': 0.7633899450302124, 'train/grad_norm_pre_clip': 11.784280776977539, 'timing/data': 0.005196335958316922, 'timing/model': 1.2050031720427796, 'learning_rate/qwen_vl_interface': 5.0510906447863025e-06, 'learning_rate/action_model': 5.051090644786303e-05, 'epoch': 1.07}) +[2026-06-17 13:01:17,379][starVLA.training.train_starvla][INFO] - Step 3643, Loss: {'train/loss': 0.7606432437896729, 'train/grad_norm_pre_clip': 10.592434883117676, 'timing/data': 0.005323361954651773, 'timing/model': 1.3309498389717191, 'learning_rate/qwen_vl_interface': 5.048929871930252e-06, 'learning_rate/action_model': 5.048929871930251e-05, 'epoch': 1.07}) +[2026-06-17 13:01:22,057][starVLA.training.train_starvla][INFO] - Step 3644, Loss: {'train/loss': 0.63850337266922, 'train/grad_norm_pre_clip': 11.913719177246094, 'timing/data': 0.005120750050991774, 'timing/model': 1.2101238729665056, 'learning_rate/qwen_vl_interface': 5.046769140756257e-06, 'learning_rate/action_model': 5.046769140756257e-05, 'epoch': 1.07}) +[2026-06-17 13:01:26,871][starVLA.training.train_starvla][INFO] - Step 3645, Loss: {'train/loss': 0.5121761560440063, 'train/grad_norm_pre_clip': 10.73490047454834, 'timing/data': 0.00489147799089551, 'timing/model': 1.2086570169776678, 'learning_rate/qwen_vl_interface': 5.044608451712237e-06, 'learning_rate/action_model': 5.044608451712237e-05, 'epoch': 1.07}) +[2026-06-17 13:01:31,656][starVLA.training.train_starvla][INFO] - Step 3646, Loss: {'train/loss': 0.6505645513534546, 'train/grad_norm_pre_clip': 10.555540084838867, 'timing/data': 0.005429282085970044, 'timing/model': 1.2046621900517493, 'learning_rate/qwen_vl_interface': 5.042447805246105e-06, 'learning_rate/action_model': 5.042447805246106e-05, 'epoch': 1.07}) +[2026-06-17 13:01:36,461][starVLA.training.train_starvla][INFO] - Step 3647, Loss: {'train/loss': 0.6615451574325562, 'train/grad_norm_pre_clip': 9.659382820129395, 'timing/data': 0.005321233067661524, 'timing/model': 1.2092213539872319, 'learning_rate/qwen_vl_interface': 5.040287201805767e-06, 'learning_rate/action_model': 5.040287201805767e-05, 'epoch': 1.07}) +[2026-06-17 13:01:41,265][starVLA.training.train_starvla][INFO] - Step 3648, Loss: {'train/loss': 0.8184143900871277, 'train/grad_norm_pre_clip': 14.480111122131348, 'timing/data': 0.005356451962143183, 'timing/model': 1.3434049860807136, 'learning_rate/qwen_vl_interface': 5.038126641839119e-06, 'learning_rate/action_model': 5.0381266418391184e-05, 'epoch': 1.07}) +[2026-06-17 13:01:45,929][starVLA.training.train_starvla][INFO] - Step 3649, Loss: {'train/loss': 0.9974610209465027, 'train/grad_norm_pre_clip': 15.670571327209473, 'timing/data': 0.005717021063901484, 'timing/model': 1.207780226948671, 'learning_rate/qwen_vl_interface': 5.0359661257940455e-06, 'learning_rate/action_model': 5.035966125794045e-05, 'epoch': 1.07}) +[2026-06-17 13:01:50,745][starVLA.training.train_starvla][INFO] - Step 3650, Loss: {'train/loss': 0.6928830742835999, 'train/grad_norm_pre_clip': 10.554472923278809, 'timing/data': 0.004937135032378137, 'timing/model': 1.205210625892505, 'learning_rate/qwen_vl_interface': 5.033805654118424e-06, 'learning_rate/action_model': 5.033805654118424e-05, 'epoch': 1.07}) +[2026-06-17 13:01:55,453][starVLA.training.train_starvla][INFO] - Step 3651, Loss: {'train/loss': 0.6030058264732361, 'train/grad_norm_pre_clip': 11.78011417388916, 'timing/data': 0.005070392042398453, 'timing/model': 1.2049756849883124, 'learning_rate/qwen_vl_interface': 5.031645227260122e-06, 'learning_rate/action_model': 5.0316452272601224e-05, 'epoch': 1.07}) +[2026-06-17 13:02:00,201][starVLA.training.train_starvla][INFO] - Step 3652, Loss: {'train/loss': 0.5308070778846741, 'train/grad_norm_pre_clip': 8.811692237854004, 'timing/data': 0.005062205018475652, 'timing/model': 1.2070296499878168, 'learning_rate/qwen_vl_interface': 5.029484845667001e-06, 'learning_rate/action_model': 5.029484845667001e-05, 'epoch': 1.07}) +[2026-06-17 13:02:04,983][starVLA.training.train_starvla][INFO] - Step 3653, Loss: {'train/loss': 0.5225999355316162, 'train/grad_norm_pre_clip': 9.688431739807129, 'timing/data': 0.005092613049782813, 'timing/model': 1.2032638469245285, 'learning_rate/qwen_vl_interface': 5.027324509786908e-06, 'learning_rate/action_model': 5.027324509786908e-05, 'epoch': 1.07}) +[2026-06-17 13:02:09,780][starVLA.training.train_starvla][INFO] - Step 3654, Loss: {'train/loss': 0.8280502557754517, 'train/grad_norm_pre_clip': 12.92803955078125, 'timing/data': 0.004926418885588646, 'timing/model': 1.3377267840551212, 'learning_rate/qwen_vl_interface': 5.025164220067683e-06, 'learning_rate/action_model': 5.025164220067683e-05, 'epoch': 1.08}) +[2026-06-17 13:02:14,442][starVLA.training.train_starvla][INFO] - Step 3655, Loss: {'train/loss': 0.5968623757362366, 'train/grad_norm_pre_clip': 8.796765327453613, 'timing/data': 0.005162341985851526, 'timing/model': 1.2064345399849117, 'learning_rate/qwen_vl_interface': 5.023003976957159e-06, 'learning_rate/action_model': 5.023003976957159e-05, 'epoch': 1.08}) +[2026-06-17 13:02:19,238][starVLA.training.train_starvla][INFO] - Step 3656, Loss: {'train/loss': 0.42919155955314636, 'train/grad_norm_pre_clip': 13.661603927612305, 'timing/data': 0.005015595001168549, 'timing/model': 1.2047919710166752, 'learning_rate/qwen_vl_interface': 5.020843780903155e-06, 'learning_rate/action_model': 5.020843780903155e-05, 'epoch': 1.08}) +[2026-06-17 13:02:24,047][starVLA.training.train_starvla][INFO] - Step 3657, Loss: {'train/loss': 0.673046886920929, 'train/grad_norm_pre_clip': 11.290369033813477, 'timing/data': 0.0053438369650393724, 'timing/model': 1.214085001964122, 'learning_rate/qwen_vl_interface': 5.01868363235348e-06, 'learning_rate/action_model': 5.01868363235348e-05, 'epoch': 1.08}) +[2026-06-17 13:02:28,810][starVLA.training.train_starvla][INFO] - Step 3658, Loss: {'train/loss': 0.7572832107543945, 'train/grad_norm_pre_clip': 16.26923370361328, 'timing/data': 0.00484182802028954, 'timing/model': 1.2061522220028564, 'learning_rate/qwen_vl_interface': 5.016523531755937e-06, 'learning_rate/action_model': 5.016523531755937e-05, 'epoch': 1.08}) +[2026-06-17 13:02:33,594][starVLA.training.train_starvla][INFO] - Step 3659, Loss: {'train/loss': 1.1616637706756592, 'train/grad_norm_pre_clip': 17.332218170166016, 'timing/data': 0.0050899890484288335, 'timing/model': 1.2054046699777246, 'learning_rate/qwen_vl_interface': 5.014363479558318e-06, 'learning_rate/action_model': 5.014363479558318e-05, 'epoch': 1.08}) +[2026-06-17 13:02:38,383][starVLA.training.train_starvla][INFO] - Step 3660, Loss: {'train/loss': 0.41303062438964844, 'train/grad_norm_pre_clip': 7.306293487548828, 'timing/data': 0.0048790909349918365, 'timing/model': 1.3365398790920153, 'learning_rate/qwen_vl_interface': 5.012203476208403e-06, 'learning_rate/action_model': 5.0122034762084025e-05, 'epoch': 1.08}) +[2026-06-17 13:02:43,043][starVLA.training.train_starvla][INFO] - Step 3661, Loss: {'train/loss': 0.8619442582130432, 'train/grad_norm_pre_clip': 12.860773086547852, 'timing/data': 0.005165200913324952, 'timing/model': 1.204991637961939, 'learning_rate/qwen_vl_interface': 5.010043522153965e-06, 'learning_rate/action_model': 5.0100435221539646e-05, 'epoch': 1.08}) +[2026-06-17 13:02:47,840][starVLA.training.train_starvla][INFO] - Step 3662, Loss: {'train/loss': 0.8249742388725281, 'train/grad_norm_pre_clip': 11.093938827514648, 'timing/data': 0.005203249049372971, 'timing/model': 1.2108853460522369, 'learning_rate/qwen_vl_interface': 5.007883617842759e-06, 'learning_rate/action_model': 5.007883617842759e-05, 'epoch': 1.08}) +[2026-06-17 13:02:52,626][starVLA.training.train_starvla][INFO] - Step 3663, Loss: {'train/loss': 0.7848892211914062, 'train/grad_norm_pre_clip': 11.59113597869873, 'timing/data': 0.004964495892636478, 'timing/model': 1.2036712779663503, 'learning_rate/qwen_vl_interface': 5.005723763722543e-06, 'learning_rate/action_model': 5.005723763722543e-05, 'epoch': 1.08}) +[2026-06-17 13:02:57,409][starVLA.training.train_starvla][INFO] - Step 3664, Loss: {'train/loss': 0.5955411791801453, 'train/grad_norm_pre_clip': 14.356499671936035, 'timing/data': 0.005091064958833158, 'timing/model': 1.205921774962917, 'learning_rate/qwen_vl_interface': 5.003563960241052e-06, 'learning_rate/action_model': 5.003563960241052e-05, 'epoch': 1.08}) +[2026-06-17 13:03:02,088][starVLA.training.train_starvla][INFO] - Step 3665, Loss: {'train/loss': 0.8393203616142273, 'train/grad_norm_pre_clip': 11.082015037536621, 'timing/data': 0.005142162088304758, 'timing/model': 1.2051406770478934, 'learning_rate/qwen_vl_interface': 5.001404207846017e-06, 'learning_rate/action_model': 5.001404207846017e-05, 'epoch': 1.08}) +[2026-06-17 13:03:06,784][starVLA.training.train_starvla][INFO] - Step 3666, Loss: {'train/loss': 0.7239251136779785, 'train/grad_norm_pre_clip': 15.250035285949707, 'timing/data': 0.005299704964272678, 'timing/model': 1.2407353210728616, 'learning_rate/qwen_vl_interface': 4.999244506985158e-06, 'learning_rate/action_model': 4.999244506985158e-05, 'epoch': 1.08}) +[2026-06-17 13:03:11,436][starVLA.training.train_starvla][INFO] - Step 3667, Loss: {'train/loss': 0.7908449769020081, 'train/grad_norm_pre_clip': 13.72732925415039, 'timing/data': 0.005013093003071845, 'timing/model': 1.2045141910202801, 'learning_rate/qwen_vl_interface': 4.997084858106182e-06, 'learning_rate/action_model': 4.997084858106183e-05, 'epoch': 1.08}) +[2026-06-17 13:03:16,161][starVLA.training.train_starvla][INFO] - Step 3668, Loss: {'train/loss': 1.0839471817016602, 'train/grad_norm_pre_clip': 14.308505058288574, 'timing/data': 0.004925591056235135, 'timing/model': 1.203460237942636, 'learning_rate/qwen_vl_interface': 4.994925261656789e-06, 'learning_rate/action_model': 4.994925261656789e-05, 'epoch': 1.08}) +[2026-06-17 13:03:20,968][starVLA.training.train_starvla][INFO] - Step 3669, Loss: {'train/loss': 0.8226475715637207, 'train/grad_norm_pre_clip': 30.888065338134766, 'timing/data': 0.005010441993363202, 'timing/model': 1.210883617051877, 'learning_rate/qwen_vl_interface': 4.99276571808466e-06, 'learning_rate/action_model': 4.9927657180846605e-05, 'epoch': 1.08}) +[2026-06-17 13:03:25,771][starVLA.training.train_starvla][INFO] - Step 3670, Loss: {'train/loss': 0.7698413133621216, 'train/grad_norm_pre_clip': 12.265382766723633, 'timing/data': 0.00527169497217983, 'timing/model': 1.2044596300693229, 'learning_rate/qwen_vl_interface': 4.9906062278374775e-06, 'learning_rate/action_model': 4.990606227837478e-05, 'epoch': 1.08}) +[2026-06-17 13:03:30,558][starVLA.training.train_starvla][INFO] - Step 3671, Loss: {'train/loss': 0.8031935691833496, 'train/grad_norm_pre_clip': 79.22782897949219, 'timing/data': 0.005526615073904395, 'timing/model': 1.3350477480562404, 'learning_rate/qwen_vl_interface': 4.988446791362903e-06, 'learning_rate/action_model': 4.988446791362903e-05, 'epoch': 1.08}) +[2026-06-17 13:03:35,291][starVLA.training.train_starvla][INFO] - Step 3672, Loss: {'train/loss': 0.9371100068092346, 'train/grad_norm_pre_clip': 14.232100486755371, 'timing/data': 0.005175587022677064, 'timing/model': 1.2604601050261408, 'learning_rate/qwen_vl_interface': 4.986287409108589e-06, 'learning_rate/action_model': 4.9862874091085893e-05, 'epoch': 1.08}) +[2026-06-17 13:03:39,995][starVLA.training.train_starvla][INFO] - Step 3673, Loss: {'train/loss': 0.8273462057113647, 'train/grad_norm_pre_clip': 13.479387283325195, 'timing/data': 0.00521641899831593, 'timing/model': 1.2053164259996265, 'learning_rate/qwen_vl_interface': 4.984128081522179e-06, 'learning_rate/action_model': 4.984128081522178e-05, 'epoch': 1.08}) +[2026-06-17 13:03:44,727][starVLA.training.train_starvla][INFO] - Step 3674, Loss: {'train/loss': 0.6617650985717773, 'train/grad_norm_pre_clip': 10.832698822021484, 'timing/data': 0.004874303936958313, 'timing/model': 1.2066226389724761, 'learning_rate/qwen_vl_interface': 4.981968809051304e-06, 'learning_rate/action_model': 4.981968809051304e-05, 'epoch': 1.08}) +[2026-06-17 13:03:49,510][starVLA.training.train_starvla][INFO] - Step 3675, Loss: {'train/loss': 0.6720903515815735, 'train/grad_norm_pre_clip': 10.435623168945312, 'timing/data': 0.004791033919900656, 'timing/model': 1.2044898289022967, 'learning_rate/qwen_vl_interface': 4.979809592143583e-06, 'learning_rate/action_model': 4.979809592143583e-05, 'epoch': 1.08}) +[2026-06-17 13:03:54,300][starVLA.training.train_starvla][INFO] - Step 3676, Loss: {'train/loss': 0.523661196231842, 'train/grad_norm_pre_clip': 9.58867359161377, 'timing/data': 0.0050965879345312715, 'timing/model': 1.2041051529813558, 'learning_rate/qwen_vl_interface': 4.977650431246624e-06, 'learning_rate/action_model': 4.977650431246624e-05, 'epoch': 1.08}) +[2026-06-17 13:03:59,087][starVLA.training.train_starvla][INFO] - Step 3677, Loss: {'train/loss': 0.6755757331848145, 'train/grad_norm_pre_clip': 12.968640327453613, 'timing/data': 0.005471420008689165, 'timing/model': 1.332171816029586, 'learning_rate/qwen_vl_interface': 4.9754913268080255e-06, 'learning_rate/action_model': 4.975491326808025e-05, 'epoch': 1.08}) +[2026-06-17 13:04:03,820][starVLA.training.train_starvla][INFO] - Step 3678, Loss: {'train/loss': 0.7087615132331848, 'train/grad_norm_pre_clip': 11.42958927154541, 'timing/data': 0.005255738040432334, 'timing/model': 1.2716148879844695, 'learning_rate/qwen_vl_interface': 4.973332279275369e-06, 'learning_rate/action_model': 4.973332279275369e-05, 'epoch': 1.08}) +[2026-06-17 13:04:08,478][starVLA.training.train_starvla][INFO] - Step 3679, Loss: {'train/loss': 0.7862769365310669, 'train/grad_norm_pre_clip': 13.103976249694824, 'timing/data': 0.00513525097630918, 'timing/model': 1.2042809179984033, 'learning_rate/qwen_vl_interface': 4.971173289096229e-06, 'learning_rate/action_model': 4.9711732890962286e-05, 'epoch': 1.08}) +[2026-06-17 13:04:13,299][starVLA.training.train_starvla][INFO] - Step 3680, Loss: {'train/loss': 0.8432988524436951, 'train/grad_norm_pre_clip': 13.467277526855469, 'timing/data': 0.033960478962399065, 'timing/model': 1.2265320969745517, 'learning_rate/qwen_vl_interface': 4.969014356718165e-06, 'learning_rate/action_model': 4.969014356718165e-05, 'epoch': 1.08}) +[2026-06-17 13:04:18,102][starVLA.training.train_starvla][INFO] - Step 3681, Loss: {'train/loss': 0.9890064597129822, 'train/grad_norm_pre_clip': 12.431609153747559, 'timing/data': 0.004814149928279221, 'timing/model': 1.2154656309867278, 'learning_rate/qwen_vl_interface': 4.9668554825887285e-06, 'learning_rate/action_model': 4.966855482588728e-05, 'epoch': 1.08}) +[2026-06-17 13:04:22,897][starVLA.training.train_starvla][INFO] - Step 3682, Loss: {'train/loss': 0.6586484313011169, 'train/grad_norm_pre_clip': 10.4487886428833, 'timing/data': 0.005205961060710251, 'timing/model': 1.2077376990346238, 'learning_rate/qwen_vl_interface': 4.964696667155455e-06, 'learning_rate/action_model': 4.964696667155454e-05, 'epoch': 1.08}) +[2026-06-17 13:04:27,694][starVLA.training.train_starvla][INFO] - Step 3683, Loss: {'train/loss': 0.7343887090682983, 'train/grad_norm_pre_clip': 10.842756271362305, 'timing/data': 0.005269794026389718, 'timing/model': 1.3321715609636158, 'learning_rate/qwen_vl_interface': 4.96253791086587e-06, 'learning_rate/action_model': 4.96253791086587e-05, 'epoch': 1.08}) +[2026-06-17 13:04:32,486][starVLA.training.train_starvla][INFO] - Step 3684, Loss: {'train/loss': 0.7498451471328735, 'train/grad_norm_pre_clip': 9.820399284362793, 'timing/data': 0.005227071000263095, 'timing/model': 1.3257449120283127, 'learning_rate/qwen_vl_interface': 4.960379214167483e-06, 'learning_rate/action_model': 4.960379214167483e-05, 'epoch': 1.08}) +[2026-06-17 13:04:37,164][starVLA.training.train_starvla][INFO] - Step 3685, Loss: {'train/loss': 0.7050260901451111, 'train/grad_norm_pre_clip': 14.999576568603516, 'timing/data': 0.005075917928479612, 'timing/model': 1.216742065967992, 'learning_rate/qwen_vl_interface': 4.9582205775077994e-06, 'learning_rate/action_model': 4.958220577507799e-05, 'epoch': 1.08}) +[2026-06-17 13:04:41,925][starVLA.training.train_starvla][INFO] - Step 3686, Loss: {'train/loss': 0.423383891582489, 'train/grad_norm_pre_clip': 7.969003200531006, 'timing/data': 0.005098021007142961, 'timing/model': 1.2067987580085173, 'learning_rate/qwen_vl_interface': 4.956062001334302e-06, 'learning_rate/action_model': 4.956062001334302e-05, 'epoch': 1.08}) +[2026-06-17 13:04:46,715][starVLA.training.train_starvla][INFO] - Step 3687, Loss: {'train/loss': 0.5473672747612, 'train/grad_norm_pre_clip': 18.943954467773438, 'timing/data': 0.005229142028838396, 'timing/model': 1.2043163490016013, 'learning_rate/qwen_vl_interface': 4.953903486094469e-06, 'learning_rate/action_model': 4.953903486094469e-05, 'epoch': 1.08}) +[2026-06-17 13:04:51,513][starVLA.training.train_starvla][INFO] - Step 3688, Loss: {'train/loss': 0.613218367099762, 'train/grad_norm_pre_clip': 9.917424201965332, 'timing/data': 0.005636727903038263, 'timing/model': 1.2054703530156985, 'learning_rate/qwen_vl_interface': 4.951745032235762e-06, 'learning_rate/action_model': 4.951745032235762e-05, 'epoch': 1.09}) +[2026-06-17 13:04:56,249][starVLA.training.train_starvla][INFO] - Step 3689, Loss: {'train/loss': 0.6205666661262512, 'train/grad_norm_pre_clip': 9.852448463439941, 'timing/data': 0.005208570975810289, 'timing/model': 1.2720890030032024, 'learning_rate/qwen_vl_interface': 4.949586640205631e-06, 'learning_rate/action_model': 4.94958664020563e-05, 'epoch': 1.09}) +[2026-06-17 13:05:00,946][starVLA.training.train_starvla][INFO] - Step 3690, Loss: {'train/loss': 0.5801209807395935, 'train/grad_norm_pre_clip': 11.016647338867188, 'timing/data': 0.005082170944660902, 'timing/model': 1.2182559410575777, 'learning_rate/qwen_vl_interface': 4.9474283104515124e-06, 'learning_rate/action_model': 4.9474283104515126e-05, 'epoch': 1.09}) +[2026-06-17 13:05:05,737][starVLA.training.train_starvla][INFO] - Step 3691, Loss: {'train/loss': 0.5987386107444763, 'train/grad_norm_pre_clip': 11.258079528808594, 'timing/data': 0.005294772097840905, 'timing/model': 1.2053244810085744, 'learning_rate/qwen_vl_interface': 4.945270043420829e-06, 'learning_rate/action_model': 4.945270043420828e-05, 'epoch': 1.09}) +[2026-06-17 13:05:10,527][starVLA.training.train_starvla][INFO] - Step 3692, Loss: {'train/loss': 0.6872276663780212, 'train/grad_norm_pre_clip': 12.525311470031738, 'timing/data': 0.005877632996998727, 'timing/model': 1.2038755459943786, 'learning_rate/qwen_vl_interface': 4.9431118395609954e-06, 'learning_rate/action_model': 4.9431118395609954e-05, 'epoch': 1.09}) +[2026-06-17 13:05:15,341][starVLA.training.train_starvla][INFO] - Step 3693, Loss: {'train/loss': 0.800422728061676, 'train/grad_norm_pre_clip': 12.791730880737305, 'timing/data': 0.005109314923174679, 'timing/model': 1.2182705170707777, 'learning_rate/qwen_vl_interface': 4.940953699319407e-06, 'learning_rate/action_model': 4.940953699319407e-05, 'epoch': 1.09}) +[2026-06-17 13:05:20,138][starVLA.training.train_starvla][INFO] - Step 3694, Loss: {'train/loss': 0.6970016360282898, 'train/grad_norm_pre_clip': 12.568774223327637, 'timing/data': 0.004689697059802711, 'timing/model': 1.3451150020118803, 'learning_rate/qwen_vl_interface': 4.93879562314345e-06, 'learning_rate/action_model': 4.93879562314345e-05, 'epoch': 1.09}) +[2026-06-17 13:05:24,890][starVLA.training.train_starvla][INFO] - Step 3695, Loss: {'train/loss': 0.5242172479629517, 'train/grad_norm_pre_clip': 8.598389625549316, 'timing/data': 0.004964821040630341, 'timing/model': 1.2925182139733806, 'learning_rate/qwen_vl_interface': 4.936637611480495e-06, 'learning_rate/action_model': 4.9366376114804945e-05, 'epoch': 1.09}) +[2026-06-17 13:05:29,571][starVLA.training.train_starvla][INFO] - Step 3696, Loss: {'train/loss': 0.7055486440658569, 'train/grad_norm_pre_clip': 12.792961120605469, 'timing/data': 0.005088848993182182, 'timing/model': 1.2114114760188386, 'learning_rate/qwen_vl_interface': 4.9344796647779e-06, 'learning_rate/action_model': 4.9344796647778995e-05, 'epoch': 1.09}) +[2026-06-17 13:05:34,364][starVLA.training.train_starvla][INFO] - Step 3697, Loss: {'train/loss': 0.7952895164489746, 'train/grad_norm_pre_clip': 13.316558837890625, 'timing/data': 0.005426143063232303, 'timing/model': 1.2053457840811461, 'learning_rate/qwen_vl_interface': 4.9323217834830105e-06, 'learning_rate/action_model': 4.9323217834830105e-05, 'epoch': 1.09}) +[2026-06-17 13:05:39,159][starVLA.training.train_starvla][INFO] - Step 3698, Loss: {'train/loss': 0.6771534085273743, 'train/grad_norm_pre_clip': 11.09029769897461, 'timing/data': 0.004918375983834267, 'timing/model': 1.2048838959308341, 'learning_rate/qwen_vl_interface': 4.930163968043155e-06, 'learning_rate/action_model': 4.930163968043155e-05, 'epoch': 1.09}) +[2026-06-17 13:05:43,948][starVLA.training.train_starvla][INFO] - Step 3699, Loss: {'train/loss': 0.6039659976959229, 'train/grad_norm_pre_clip': 12.26852035522461, 'timing/data': 0.005217429017648101, 'timing/model': 1.2047485279617831, 'learning_rate/qwen_vl_interface': 4.928006218905656e-06, 'learning_rate/action_model': 4.928006218905656e-05, 'epoch': 1.09}) +[2026-06-17 13:07:26,089][starVLA.training.train_starvla][INFO] - Step 3700, Loss: {'train/loss': 0.7916014790534973, 'train/grad_norm_pre_clip': 11.50838565826416, 'eval/loss': 0.7580769062042236, 'eval/latency_0/loss': 0.7580769062042236, 'eval/demon_fire_precision': 0.889589905362776, 'eval/demon_fire_recall': 0.9368770764119602, 'eval/demon_fire_f1': 0.912621359223301, 'eval/demon_fire_rate_model': 0.7052280311457174, 'eval/demon_fire_rate_teacher': 0.6696329254727474, 'eval/demon_left_precision': 0.8158947970268725, 'eval/demon_left_recall': 0.8084985835694051, 'eval/demon_left_f1': 0.8121798520204894, 'eval/demon_left_rate_model': 0.48637374860956617, 'eval/demon_left_rate_teacher': 0.49082313681868744, 'eval/demon_right_precision': 0.806027397260274, 'eval/demon_right_recall': 0.8310734463276837, 'eval/demon_right_f1': 0.8183588317107094, 'eval/demon_right_rate_model': 0.5075083426028921, 'eval/demon_right_rate_teacher': 0.4922135706340378, 'eval/demon_move_f1': 0.8152693418655994, 'eval/demon_cc_f1': 0.8810589556095281, 'eval/latency_0/demon_fire_precision': 0.889589905362776, 'eval/latency_0/demon_fire_recall': 0.9368770764119602, 'eval/latency_0/demon_fire_f1': 0.912621359223301, 'eval/latency_0/demon_fire_rate_model': 0.7052280311457174, 'eval/latency_0/demon_fire_rate_teacher': 0.6696329254727474, 'eval/latency_0/demon_left_precision': 0.8158947970268725, 'eval/latency_0/demon_left_recall': 0.8084985835694051, 'eval/latency_0/demon_left_f1': 0.8121798520204894, 'eval/latency_0/demon_left_rate_model': 0.48637374860956617, 'eval/latency_0/demon_left_rate_teacher': 0.49082313681868744, 'eval/latency_0/demon_right_precision': 0.806027397260274, 'eval/latency_0/demon_right_recall': 0.8310734463276837, 'eval/latency_0/demon_right_f1': 0.8183588317107094, 'eval/latency_0/demon_right_rate_model': 0.5075083426028921, 'eval/latency_0/demon_right_rate_teacher': 0.4922135706340378, 'eval/latency_0/demon_move_f1': 0.8152693418655994, 'eval/latency_0/demon_cc_f1': 0.8810589556095281, 'timing/data': 0.005223373998887837, 'timing/model': 1.2056076300796121, 'learning_rate/qwen_vl_interface': 4.9258485365178135e-06, 'learning_rate/action_model': 4.925848536517814e-05, 'epoch': 1.09}) +[2026-06-17 13:07:26,090][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_3700_state +[2026-06-17 13:07:26,093][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 13:07:29,664][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_3700_state/model.safetensors +[2026-06-17 13:07:37,380][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_3700_state/optimizer.bin +[2026-06-17 13:07:37,380][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_3700_state/sampler.bin +[2026-06-17 13:07:37,380][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_3700_state/sampler_1.bin +[2026-06-17 13:07:37,381][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_3700_state/random_states_0.pkl +[2026-06-17 13:07:37,445][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3700 +[2026-06-17 13:07:37,445][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 13:07:37,475][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 13:07:42,195][starVLA.training.train_starvla][INFO] - Step 3701, Loss: {'train/loss': 0.7253775596618652, 'train/grad_norm_pre_clip': 11.758248329162598, 'timing/data': 0.005421959911473095, 'timing/model': 1.1869951270055026, 'learning_rate/qwen_vl_interface': 4.923690921326918e-06, 'learning_rate/action_model': 4.9236909213269175e-05, 'epoch': 1.09}) +[2026-06-17 13:07:46,849][starVLA.training.train_starvla][INFO] - Step 3702, Loss: {'train/loss': 0.8055634498596191, 'train/grad_norm_pre_clip': 11.657851219177246, 'timing/data': 0.005321298958733678, 'timing/model': 1.1913642539875582, 'learning_rate/qwen_vl_interface': 4.921533373780244e-06, 'learning_rate/action_model': 4.921533373780244e-05, 'epoch': 1.09}) +[2026-06-17 13:07:51,595][starVLA.training.train_starvla][INFO] - Step 3703, Loss: {'train/loss': 0.7464116811752319, 'train/grad_norm_pre_clip': 26.31603240966797, 'timing/data': 0.005197683000005782, 'timing/model': 1.2718270849436522, 'learning_rate/qwen_vl_interface': 4.919375894325057e-06, 'learning_rate/action_model': 4.9193758943250566e-05, 'epoch': 1.09}) +[2026-06-17 13:07:56,236][starVLA.training.train_starvla][INFO] - Step 3704, Loss: {'train/loss': 0.7089866995811462, 'train/grad_norm_pre_clip': 13.181716918945312, 'timing/data': 0.0050548589788377285, 'timing/model': 1.2117427890188992, 'learning_rate/qwen_vl_interface': 4.917218483408603e-06, 'learning_rate/action_model': 4.9172184834086035e-05, 'epoch': 1.09}) +[2026-06-17 13:08:00,987][starVLA.training.train_starvla][INFO] - Step 3705, Loss: {'train/loss': 0.577616810798645, 'train/grad_norm_pre_clip': 8.660755157470703, 'timing/data': 0.00534466200042516, 'timing/model': 1.1953373620053753, 'learning_rate/qwen_vl_interface': 4.915061141478114e-06, 'learning_rate/action_model': 4.915061141478113e-05, 'epoch': 1.09}) +[2026-06-17 13:08:05,754][starVLA.training.train_starvla][INFO] - Step 3706, Loss: {'train/loss': 0.5304632782936096, 'train/grad_norm_pre_clip': 10.622020721435547, 'timing/data': 0.005079641006886959, 'timing/model': 1.1985780639806762, 'learning_rate/qwen_vl_interface': 4.912903868980812e-06, 'learning_rate/action_model': 4.9129038689808115e-05, 'epoch': 1.09}) +[2026-06-17 13:08:10,591][starVLA.training.train_starvla][INFO] - Step 3707, Loss: {'train/loss': 0.4348074197769165, 'train/grad_norm_pre_clip': 10.106337547302246, 'timing/data': 0.005268711014650762, 'timing/model': 1.2813240579562262, 'learning_rate/qwen_vl_interface': 4.9107466663639005e-06, 'learning_rate/action_model': 4.910746666363901e-05, 'epoch': 1.09}) +[2026-06-17 13:08:15,342][starVLA.training.train_starvla][INFO] - Step 3708, Loss: {'train/loss': 0.5670982003211975, 'train/grad_norm_pre_clip': 12.201249122619629, 'timing/data': 0.004990663030184805, 'timing/model': 1.197560001979582, 'learning_rate/qwen_vl_interface': 4.9085895340745705e-06, 'learning_rate/action_model': 4.9085895340745705e-05, 'epoch': 1.09}) +[2026-06-17 13:08:20,070][starVLA.training.train_starvla][INFO] - Step 3709, Loss: {'train/loss': 0.8915634751319885, 'train/grad_norm_pre_clip': 18.806304931640625, 'timing/data': 0.005326195969246328, 'timing/model': 1.295539234066382, 'learning_rate/qwen_vl_interface': 4.906432472559996e-06, 'learning_rate/action_model': 4.9064324725599955e-05, 'epoch': 1.09}) +[2026-06-17 13:08:24,713][starVLA.training.train_starvla][INFO] - Step 3710, Loss: {'train/loss': 0.513409435749054, 'train/grad_norm_pre_clip': 10.573060035705566, 'timing/data': 0.00530048506334424, 'timing/model': 1.2003311649896204, 'learning_rate/qwen_vl_interface': 4.90427548226734e-06, 'learning_rate/action_model': 4.9042754822673396e-05, 'epoch': 1.09}) +[2026-06-17 13:08:29,484][starVLA.training.train_starvla][INFO] - Step 3711, Loss: {'train/loss': 0.7289042472839355, 'train/grad_norm_pre_clip': 14.755812644958496, 'timing/data': 0.0050764090847224, 'timing/model': 1.2015385290142149, 'learning_rate/qwen_vl_interface': 4.902118563643751e-06, 'learning_rate/action_model': 4.902118563643751e-05, 'epoch': 1.09}) +[2026-06-17 13:08:34,260][starVLA.training.train_starvla][INFO] - Step 3712, Loss: {'train/loss': 0.9119201898574829, 'train/grad_norm_pre_clip': 13.452875137329102, 'timing/data': 0.005536339012905955, 'timing/model': 1.2011502119712532, 'learning_rate/qwen_vl_interface': 4.899961717136356e-06, 'learning_rate/action_model': 4.899961717136355e-05, 'epoch': 1.09}) +[2026-06-17 13:08:39,034][starVLA.training.train_starvla][INFO] - Step 3713, Loss: {'train/loss': 0.5701514482498169, 'train/grad_norm_pre_clip': 13.482563972473145, 'timing/data': 0.0052501700120046735, 'timing/model': 1.2039244249463081, 'learning_rate/qwen_vl_interface': 4.897804943192273e-06, 'learning_rate/action_model': 4.897804943192273e-05, 'epoch': 1.09}) +[2026-06-17 13:08:43,820][starVLA.training.train_starvla][INFO] - Step 3714, Loss: {'train/loss': 0.7968578338623047, 'train/grad_norm_pre_clip': 11.965182304382324, 'timing/data': 0.0050519200740382075, 'timing/model': 1.333354125963524, 'learning_rate/qwen_vl_interface': 4.895648242258605e-06, 'learning_rate/action_model': 4.895648242258605e-05, 'epoch': 1.09}) +[2026-06-17 13:08:48,474][starVLA.training.train_starvla][INFO] - Step 3715, Loss: {'train/loss': 0.5662553310394287, 'train/grad_norm_pre_clip': 10.880128860473633, 'timing/data': 0.005325636942870915, 'timing/model': 1.2010942879132926, 'learning_rate/qwen_vl_interface': 4.893491614782438e-06, 'learning_rate/action_model': 4.8934916147824375e-05, 'epoch': 1.09}) +[2026-06-17 13:08:53,284][starVLA.training.train_starvla][INFO] - Step 3716, Loss: {'train/loss': 0.4666095972061157, 'train/grad_norm_pre_clip': 11.085966110229492, 'timing/data': 0.005168998031876981, 'timing/model': 1.2033763349754736, 'learning_rate/qwen_vl_interface': 4.891335061210841e-06, 'learning_rate/action_model': 4.891335061210841e-05, 'epoch': 1.09}) +[2026-06-17 13:08:58,064][starVLA.training.train_starvla][INFO] - Step 3717, Loss: {'train/loss': 0.7063812017440796, 'train/grad_norm_pre_clip': 15.221399307250977, 'timing/data': 0.005196545040234923, 'timing/model': 1.2079778299666941, 'learning_rate/qwen_vl_interface': 4.889178581990874e-06, 'learning_rate/action_model': 4.8891785819908734e-05, 'epoch': 1.09}) +[2026-06-17 13:09:02,759][starVLA.training.train_starvla][INFO] - Step 3718, Loss: {'train/loss': 0.7696877717971802, 'train/grad_norm_pre_clip': 13.690939903259277, 'timing/data': 0.0051919210236519575, 'timing/model': 1.20139308902435, 'learning_rate/qwen_vl_interface': 4.887022177569576e-06, 'learning_rate/action_model': 4.8870221775695754e-05, 'epoch': 1.09}) +[2026-06-17 13:09:07,566][starVLA.training.train_starvla][INFO] - Step 3719, Loss: {'train/loss': 0.8791744709014893, 'train/grad_norm_pre_clip': 15.34144401550293, 'timing/data': 0.005087440018542111, 'timing/model': 1.2024826079141349, 'learning_rate/qwen_vl_interface': 4.88486584839397e-06, 'learning_rate/action_model': 4.8848658483939695e-05, 'epoch': 1.09}) +[2026-06-17 13:09:12,348][starVLA.training.train_starvla][INFO] - Step 3720, Loss: {'train/loss': 0.6546064019203186, 'train/grad_norm_pre_clip': 11.905187606811523, 'timing/data': 0.005279686884023249, 'timing/model': 1.3328578890068457, 'learning_rate/qwen_vl_interface': 4.882709594911065e-06, 'learning_rate/action_model': 4.8827095949110655e-05, 'epoch': 1.09}) +[2026-06-17 13:09:16,999][starVLA.training.train_starvla][INFO] - Step 3721, Loss: {'train/loss': 0.8329042196273804, 'train/grad_norm_pre_clip': 14.25047779083252, 'timing/data': 0.005397030035965145, 'timing/model': 1.2073835000628605, 'learning_rate/qwen_vl_interface': 4.880553417567858e-06, 'learning_rate/action_model': 4.880553417567858e-05, 'epoch': 1.09}) +[2026-06-17 13:09:21,783][starVLA.training.train_starvla][INFO] - Step 3722, Loss: {'train/loss': 0.9402250051498413, 'train/grad_norm_pre_clip': 13.327710151672363, 'timing/data': 0.005092182080261409, 'timing/model': 1.205588785931468, 'learning_rate/qwen_vl_interface': 4.878397316811324e-06, 'learning_rate/action_model': 4.878397316811324e-05, 'epoch': 1.1}) +[2026-06-17 13:09:26,559][starVLA.training.train_starvla][INFO] - Step 3723, Loss: {'train/loss': 0.7051154375076294, 'train/grad_norm_pre_clip': 11.128007888793945, 'timing/data': 0.00516674795653671, 'timing/model': 1.2057567560113966, 'learning_rate/qwen_vl_interface': 4.876241293088428e-06, 'learning_rate/action_model': 4.876241293088428e-05, 'epoch': 1.1}) +[2026-06-17 13:09:31,340][starVLA.training.train_starvla][INFO] - Step 3724, Loss: {'train/loss': 0.6376292705535889, 'train/grad_norm_pre_clip': 10.385351181030273, 'timing/data': 0.0052068670047447085, 'timing/model': 1.20810862199869, 'learning_rate/qwen_vl_interface': 4.874085346846111e-06, 'learning_rate/action_model': 4.87408534684611e-05, 'epoch': 1.1}) +[2026-06-17 13:09:36,122][starVLA.training.train_starvla][INFO] - Step 3725, Loss: {'train/loss': 0.5319780707359314, 'train/grad_norm_pre_clip': 11.281421661376953, 'timing/data': 0.005136738996952772, 'timing/model': 1.203816590947099, 'learning_rate/qwen_vl_interface': 4.871929478531309e-06, 'learning_rate/action_model': 4.871929478531309e-05, 'epoch': 1.1}) +[2026-06-17 13:09:40,901][starVLA.training.train_starvla][INFO] - Step 3726, Loss: {'train/loss': 0.615462601184845, 'train/grad_norm_pre_clip': 10.570878982543945, 'timing/data': 0.00509739900007844, 'timing/model': 1.3271260620094836, 'learning_rate/qwen_vl_interface': 4.869773688590932e-06, 'learning_rate/action_model': 4.8697736885909325e-05, 'epoch': 1.1}) +[2026-06-17 13:09:45,555][starVLA.training.train_starvla][INFO] - Step 3727, Loss: {'train/loss': 0.5724198222160339, 'train/grad_norm_pre_clip': 13.876715660095215, 'timing/data': 0.005240864935331047, 'timing/model': 1.2047483150381595, 'learning_rate/qwen_vl_interface': 4.867617977471878e-06, 'learning_rate/action_model': 4.867617977471878e-05, 'epoch': 1.1}) +[2026-06-17 13:09:50,367][starVLA.training.train_starvla][INFO] - Step 3728, Loss: {'train/loss': 0.6840898990631104, 'train/grad_norm_pre_clip': 15.523956298828125, 'timing/data': 0.004872558987699449, 'timing/model': 1.2217581060249358, 'learning_rate/qwen_vl_interface': 4.865462345621029e-06, 'learning_rate/action_model': 4.865462345621029e-05, 'epoch': 1.1}) +[2026-06-17 13:09:55,157][starVLA.training.train_starvla][INFO] - Step 3729, Loss: {'train/loss': 0.6022353172302246, 'train/grad_norm_pre_clip': 11.889772415161133, 'timing/data': 0.005064851022325456, 'timing/model': 1.2064237460726872, 'learning_rate/qwen_vl_interface': 4.86330679348525e-06, 'learning_rate/action_model': 4.863306793485249e-05, 'epoch': 1.1}) +[2026-06-17 13:09:59,952][starVLA.training.train_starvla][INFO] - Step 3730, Loss: {'train/loss': 0.8387665748596191, 'train/grad_norm_pre_clip': 15.306708335876465, 'timing/data': 0.005229196976870298, 'timing/model': 1.2036483390256763, 'learning_rate/qwen_vl_interface': 4.8611513215113894e-06, 'learning_rate/action_model': 4.861151321511389e-05, 'epoch': 1.1}) +[2026-06-17 13:10:04,743][starVLA.training.train_starvla][INFO] - Step 3731, Loss: {'train/loss': 0.616993248462677, 'train/grad_norm_pre_clip': 12.456802368164062, 'timing/data': 0.005018354044295847, 'timing/model': 1.3351553829852492, 'learning_rate/qwen_vl_interface': 4.858995930146276e-06, 'learning_rate/action_model': 4.858995930146275e-05, 'epoch': 1.1}) +[2026-06-17 13:10:09,401][starVLA.training.train_starvla][INFO] - Step 3732, Loss: {'train/loss': 0.800088107585907, 'train/grad_norm_pre_clip': 17.236513137817383, 'timing/data': 0.00499778997618705, 'timing/model': 1.2061134719988331, 'learning_rate/qwen_vl_interface': 4.856840619836729e-06, 'learning_rate/action_model': 4.856840619836728e-05, 'epoch': 1.1}) +[2026-06-17 13:10:14,203][starVLA.training.train_starvla][INFO] - Step 3733, Loss: {'train/loss': 0.41885316371917725, 'train/grad_norm_pre_clip': 10.012054443359375, 'timing/data': 0.005118455970659852, 'timing/model': 1.2108009048970416, 'learning_rate/qwen_vl_interface': 4.854685391029543e-06, 'learning_rate/action_model': 4.854685391029543e-05, 'epoch': 1.1}) +[2026-06-17 13:10:18,913][starVLA.training.train_starvla][INFO] - Step 3734, Loss: {'train/loss': 0.701815128326416, 'train/grad_norm_pre_clip': 10.349257469177246, 'timing/data': 0.00549279700499028, 'timing/model': 1.2070102790603414, 'learning_rate/qwen_vl_interface': 4.852530244171502e-06, 'learning_rate/action_model': 4.852530244171502e-05, 'epoch': 1.1}) +[2026-06-17 13:10:23,647][starVLA.training.train_starvla][INFO] - Step 3735, Loss: {'train/loss': 0.7755095362663269, 'train/grad_norm_pre_clip': 14.282125473022461, 'timing/data': 0.005004231003113091, 'timing/model': 1.2080773790366948, 'learning_rate/qwen_vl_interface': 4.850375179709368e-06, 'learning_rate/action_model': 4.8503751797093674e-05, 'epoch': 1.1}) +[2026-06-17 13:10:28,433][starVLA.training.train_starvla][INFO] - Step 3736, Loss: {'train/loss': 0.6672287583351135, 'train/grad_norm_pre_clip': 16.588546752929688, 'timing/data': 0.005114685976877809, 'timing/model': 1.203920591971837, 'learning_rate/qwen_vl_interface': 4.84822019808989e-06, 'learning_rate/action_model': 4.84822019808989e-05, 'epoch': 1.1}) +[2026-06-17 13:10:33,216][starVLA.training.train_starvla][INFO] - Step 3737, Loss: {'train/loss': 0.5244879722595215, 'train/grad_norm_pre_clip': 11.677510261535645, 'timing/data': 0.005182155058719218, 'timing/model': 1.3360661669867113, 'learning_rate/qwen_vl_interface': 4.846065299759799e-06, 'learning_rate/action_model': 4.846065299759799e-05, 'epoch': 1.1}) +[2026-06-17 13:10:37,871][starVLA.training.train_starvla][INFO] - Step 3738, Loss: {'train/loss': 0.7443843483924866, 'train/grad_norm_pre_clip': 13.082938194274902, 'timing/data': 0.005303652957081795, 'timing/model': 1.2052755519980565, 'learning_rate/qwen_vl_interface': 4.843910485165802e-06, 'learning_rate/action_model': 4.843910485165802e-05, 'epoch': 1.1}) +[2026-06-17 13:10:42,685][starVLA.training.train_starvla][INFO] - Step 3739, Loss: {'train/loss': 0.8115355372428894, 'train/grad_norm_pre_clip': 13.424880981445312, 'timing/data': 0.005198369035497308, 'timing/model': 1.204120310023427, 'learning_rate/qwen_vl_interface': 4.8417557547546015e-06, 'learning_rate/action_model': 4.8417557547546015e-05, 'epoch': 1.1}) +[2026-06-17 13:10:47,461][starVLA.training.train_starvla][INFO] - Step 3740, Loss: {'train/loss': 0.4842163324356079, 'train/grad_norm_pre_clip': 9.933927536010742, 'timing/data': 0.005181577056646347, 'timing/model': 1.2050715269288048, 'learning_rate/qwen_vl_interface': 4.839601108972872e-06, 'learning_rate/action_model': 4.839601108972871e-05, 'epoch': 1.1}) +[2026-06-17 13:10:52,211][starVLA.training.train_starvla][INFO] - Step 3741, Loss: {'train/loss': 0.9460443258285522, 'train/grad_norm_pre_clip': 15.573002815246582, 'timing/data': 0.005200693034566939, 'timing/model': 1.2044863810297102, 'learning_rate/qwen_vl_interface': 4.837446548267275e-06, 'learning_rate/action_model': 4.837446548267274e-05, 'epoch': 1.1}) +[2026-06-17 13:10:56,987][starVLA.training.train_starvla][INFO] - Step 3742, Loss: {'train/loss': 0.9273116588592529, 'train/grad_norm_pre_clip': 12.148743629455566, 'timing/data': 0.005350727937184274, 'timing/model': 1.2043839449761435, 'learning_rate/qwen_vl_interface': 4.835292073084449e-06, 'learning_rate/action_model': 4.8352920730844486e-05, 'epoch': 1.1}) +[2026-06-17 13:11:01,779][starVLA.training.train_starvla][INFO] - Step 3743, Loss: {'train/loss': 0.8285005688667297, 'train/grad_norm_pre_clip': 13.09416675567627, 'timing/data': 0.005242619896307588, 'timing/model': 1.337657144991681, 'learning_rate/qwen_vl_interface': 4.833137683871025e-06, 'learning_rate/action_model': 4.833137683871025e-05, 'epoch': 1.1}) +[2026-06-17 13:11:06,436][starVLA.training.train_starvla][INFO] - Step 3744, Loss: {'train/loss': 0.4962601065635681, 'train/grad_norm_pre_clip': 8.96217155456543, 'timing/data': 0.004774293047375977, 'timing/model': 1.2078050440177321, 'learning_rate/qwen_vl_interface': 4.830983381073609e-06, 'learning_rate/action_model': 4.830983381073608e-05, 'epoch': 1.1}) +[2026-06-17 13:11:11,221][starVLA.training.train_starvla][INFO] - Step 3745, Loss: {'train/loss': 0.6698975563049316, 'train/grad_norm_pre_clip': 11.415247917175293, 'timing/data': 0.005062047042883933, 'timing/model': 1.204984307056293, 'learning_rate/qwen_vl_interface': 4.828829165138787e-06, 'learning_rate/action_model': 4.828829165138787e-05, 'epoch': 1.1}) +[2026-06-17 13:11:16,004][starVLA.training.train_starvla][INFO] - Step 3746, Loss: {'train/loss': 0.9202126264572144, 'train/grad_norm_pre_clip': 13.306896209716797, 'timing/data': 0.005337354028597474, 'timing/model': 1.2058592040557414, 'learning_rate/qwen_vl_interface': 4.826675036513132e-06, 'learning_rate/action_model': 4.826675036513132e-05, 'epoch': 1.1}) +[2026-06-17 13:11:20,789][starVLA.training.train_starvla][INFO] - Step 3747, Loss: {'train/loss': 1.0364538431167603, 'train/grad_norm_pre_clip': 15.727676391601562, 'timing/data': 0.005315121030434966, 'timing/model': 1.2039501150138676, 'learning_rate/qwen_vl_interface': 4.824520995643198e-06, 'learning_rate/action_model': 4.8245209956431984e-05, 'epoch': 1.1}) +[2026-06-17 13:11:25,462][starVLA.training.train_starvla][INFO] - Step 3748, Loss: {'train/loss': 0.6130005121231079, 'train/grad_norm_pre_clip': 12.161065101623535, 'timing/data': 0.005062931915745139, 'timing/model': 1.205112818046473, 'learning_rate/qwen_vl_interface': 4.82236704297552e-06, 'learning_rate/action_model': 4.82236704297552e-05, 'epoch': 1.1}) +[2026-06-17 13:11:30,165][starVLA.training.train_starvla][INFO] - Step 3749, Loss: {'train/loss': 1.0105297565460205, 'train/grad_norm_pre_clip': 17.892316818237305, 'timing/data': 0.005001438898034394, 'timing/model': 1.2446123149711639, 'learning_rate/qwen_vl_interface': 4.820213178956613e-06, 'learning_rate/action_model': 4.820213178956612e-05, 'epoch': 1.1}) +[2026-06-17 13:11:34,832][starVLA.training.train_starvla][INFO] - Step 3750, Loss: {'train/loss': 0.3611176908016205, 'train/grad_norm_pre_clip': 10.711369514465332, 'timing/data': 0.005198589991778135, 'timing/model': 1.2090663610724732, 'learning_rate/qwen_vl_interface': 4.818059404032977e-06, 'learning_rate/action_model': 4.818059404032977e-05, 'epoch': 1.1}) +[2026-06-17 13:11:39,586][starVLA.training.train_starvla][INFO] - Step 3751, Loss: {'train/loss': 0.6087236404418945, 'train/grad_norm_pre_clip': 10.96889591217041, 'timing/data': 0.006965604028664529, 'timing/model': 1.217280367971398, 'learning_rate/qwen_vl_interface': 4.815905718651092e-06, 'learning_rate/action_model': 4.815905718651092e-05, 'epoch': 1.1}) +[2026-06-17 13:11:44,378][starVLA.training.train_starvla][INFO] - Step 3752, Loss: {'train/loss': 0.760535717010498, 'train/grad_norm_pre_clip': 16.65150260925293, 'timing/data': 0.0053251440403983, 'timing/model': 1.2050419190200046, 'learning_rate/qwen_vl_interface': 4.813752123257419e-06, 'learning_rate/action_model': 4.813752123257418e-05, 'epoch': 1.1}) +[2026-06-17 13:11:49,160][starVLA.training.train_starvla][INFO] - Step 3753, Loss: {'train/loss': 0.8314878344535828, 'train/grad_norm_pre_clip': 11.416451454162598, 'timing/data': 0.00543669203761965, 'timing/model': 1.2023073920281604, 'learning_rate/qwen_vl_interface': 4.811598618298398e-06, 'learning_rate/action_model': 4.811598618298398e-05, 'epoch': 1.1}) +[2026-06-17 13:11:53,942][starVLA.training.train_starvla][INFO] - Step 3754, Loss: {'train/loss': 0.7328976988792419, 'train/grad_norm_pre_clip': 53.07053756713867, 'timing/data': 0.004910072078928351, 'timing/model': 1.3310863389633596, 'learning_rate/qwen_vl_interface': 4.809445204220457e-06, 'learning_rate/action_model': 4.809445204220457e-05, 'epoch': 1.1}) +[2026-06-17 13:11:58,645][starVLA.training.train_starvla][INFO] - Step 3755, Loss: {'train/loss': 0.7086418271064758, 'train/grad_norm_pre_clip': 11.624120712280273, 'timing/data': 0.005178157007321715, 'timing/model': 1.2501589059829712, 'learning_rate/qwen_vl_interface': 4.8072918814699985e-06, 'learning_rate/action_model': 4.8072918814699986e-05, 'epoch': 1.1}) +[2026-06-17 13:12:03,377][starVLA.training.train_starvla][INFO] - Step 3756, Loss: {'train/loss': 0.6349053382873535, 'train/grad_norm_pre_clip': 38.232486724853516, 'timing/data': 0.0048504100413993, 'timing/model': 1.2030612010275945, 'learning_rate/qwen_vl_interface': 4.80513865049341e-06, 'learning_rate/action_model': 4.80513865049341e-05, 'epoch': 1.11}) +[2026-06-17 13:12:08,128][starVLA.training.train_starvla][INFO] - Step 3757, Loss: {'train/loss': 0.7586900591850281, 'train/grad_norm_pre_clip': 32.751243591308594, 'timing/data': 0.00554415094666183, 'timing/model': 1.2144075749674812, 'learning_rate/qwen_vl_interface': 4.802985511737057e-06, 'learning_rate/action_model': 4.802985511737057e-05, 'epoch': 1.11}) +[2026-06-17 13:12:12,916][starVLA.training.train_starvla][INFO] - Step 3758, Loss: {'train/loss': 0.6549393534660339, 'train/grad_norm_pre_clip': 9.713706970214844, 'timing/data': 0.00611869094427675, 'timing/model': 1.203519818955101, 'learning_rate/qwen_vl_interface': 4.80083246564729e-06, 'learning_rate/action_model': 4.8008324656472894e-05, 'epoch': 1.11}) +[2026-06-17 13:12:17,715][starVLA.training.train_starvla][INFO] - Step 3759, Loss: {'train/loss': 0.6597688794136047, 'train/grad_norm_pre_clip': 13.049570083618164, 'timing/data': 0.005257987999357283, 'timing/model': 1.2072520660003647, 'learning_rate/qwen_vl_interface': 4.798679512670436e-06, 'learning_rate/action_model': 4.798679512670435e-05, 'epoch': 1.11}) +[2026-06-17 13:12:22,503][starVLA.training.train_starvla][INFO] - Step 3760, Loss: {'train/loss': 0.7000753283500671, 'train/grad_norm_pre_clip': 10.987308502197266, 'timing/data': 0.005203747074119747, 'timing/model': 1.337192971026525, 'learning_rate/qwen_vl_interface': 4.796526653252802e-06, 'learning_rate/action_model': 4.796526653252802e-05, 'epoch': 1.11}) +[2026-06-17 13:12:27,165][starVLA.training.train_starvla][INFO] - Step 3761, Loss: {'train/loss': 0.578008770942688, 'train/grad_norm_pre_clip': 10.733999252319336, 'timing/data': 0.0051553669618442655, 'timing/model': 1.2088629700010642, 'learning_rate/qwen_vl_interface': 4.794373887840683e-06, 'learning_rate/action_model': 4.794373887840682e-05, 'epoch': 1.11}) +[2026-06-17 13:12:31,990][starVLA.training.train_starvla][INFO] - Step 3762, Loss: {'train/loss': 0.40325525403022766, 'train/grad_norm_pre_clip': 7.792588233947754, 'timing/data': 0.005168617935851216, 'timing/model': 1.2027304359944537, 'learning_rate/qwen_vl_interface': 4.792221216880347e-06, 'learning_rate/action_model': 4.792221216880347e-05, 'epoch': 1.11}) +[2026-06-17 13:12:36,775][starVLA.training.train_starvla][INFO] - Step 3763, Loss: {'train/loss': 1.1908512115478516, 'train/grad_norm_pre_clip': 31.774993896484375, 'timing/data': 0.005199986044317484, 'timing/model': 1.2033233150141314, 'learning_rate/qwen_vl_interface': 4.790068640818046e-06, 'learning_rate/action_model': 4.790068640818045e-05, 'epoch': 1.11}) +[2026-06-17 13:12:41,568][starVLA.training.train_starvla][INFO] - Step 3764, Loss: {'train/loss': 0.5592638850212097, 'train/grad_norm_pre_clip': 10.402438163757324, 'timing/data': 0.005148484953679144, 'timing/model': 1.209747274988331, 'learning_rate/qwen_vl_interface': 4.787916160100008e-06, 'learning_rate/action_model': 4.787916160100008e-05, 'epoch': 1.11}) +[2026-06-17 13:12:46,386][starVLA.training.train_starvla][INFO] - Step 3765, Loss: {'train/loss': 0.5717481374740601, 'train/grad_norm_pre_clip': 11.717123985290527, 'timing/data': 0.005300516961142421, 'timing/model': 1.346462595043704, 'learning_rate/qwen_vl_interface': 4.78576377517245e-06, 'learning_rate/action_model': 4.7857637751724504e-05, 'epoch': 1.11}) +[2026-06-17 13:12:51,078][starVLA.training.train_starvla][INFO] - Step 3766, Loss: {'train/loss': 0.7910266518592834, 'train/grad_norm_pre_clip': 10.826830863952637, 'timing/data': 0.005682956078089774, 'timing/model': 1.203869970049709, 'learning_rate/qwen_vl_interface': 4.783611486481561e-06, 'learning_rate/action_model': 4.7836114864815604e-05, 'epoch': 1.11}) +[2026-06-17 13:12:55,866][starVLA.training.train_starvla][INFO] - Step 3767, Loss: {'train/loss': 0.7090231776237488, 'train/grad_norm_pre_clip': 11.118151664733887, 'timing/data': 0.005081722978502512, 'timing/model': 1.2093273510690778, 'learning_rate/qwen_vl_interface': 4.781459294473511e-06, 'learning_rate/action_model': 4.781459294473511e-05, 'epoch': 1.11}) +[2026-06-17 13:13:00,623][starVLA.training.train_starvla][INFO] - Step 3768, Loss: {'train/loss': 0.7993621826171875, 'train/grad_norm_pre_clip': 13.663466453552246, 'timing/data': 0.005716255982406437, 'timing/model': 1.2030756179010496, 'learning_rate/qwen_vl_interface': 4.779307199594454e-06, 'learning_rate/action_model': 4.779307199594454e-05, 'epoch': 1.11}) +[2026-06-17 13:13:05,420][starVLA.training.train_starvla][INFO] - Step 3769, Loss: {'train/loss': 0.6172459125518799, 'train/grad_norm_pre_clip': 13.426838874816895, 'timing/data': 0.005106822005473077, 'timing/model': 1.2116440739482641, 'learning_rate/qwen_vl_interface': 4.777155202290522e-06, 'learning_rate/action_model': 4.777155202290521e-05, 'epoch': 1.11}) +[2026-06-17 13:13:10,235][starVLA.training.train_starvla][INFO] - Step 3770, Loss: {'train/loss': 0.6739338040351868, 'train/grad_norm_pre_clip': 13.240523338317871, 'timing/data': 0.004917145008221269, 'timing/model': 1.2149005529936403, 'learning_rate/qwen_vl_interface': 4.775003303007824e-06, 'learning_rate/action_model': 4.775003303007824e-05, 'epoch': 1.11}) +[2026-06-17 13:13:15,043][starVLA.training.train_starvla][INFO] - Step 3771, Loss: {'train/loss': 0.6260496377944946, 'train/grad_norm_pre_clip': 10.00081729888916, 'timing/data': 0.0051347470143809915, 'timing/model': 1.3439002470113337, 'learning_rate/qwen_vl_interface': 4.772851502192453e-06, 'learning_rate/action_model': 4.772851502192452e-05, 'epoch': 1.11}) +[2026-06-17 13:13:19,789][starVLA.training.train_starvla][INFO] - Step 3772, Loss: {'train/loss': 0.591171145439148, 'train/grad_norm_pre_clip': 10.670939445495605, 'timing/data': 0.005396784981712699, 'timing/model': 1.2810122050577775, 'learning_rate/qwen_vl_interface': 4.770699800290478e-06, 'learning_rate/action_model': 4.770699800290478e-05, 'epoch': 1.11}) +[2026-06-17 13:13:24,463][starVLA.training.train_starvla][INFO] - Step 3773, Loss: {'train/loss': 0.5097149014472961, 'train/grad_norm_pre_clip': 9.923744201660156, 'timing/data': 0.005369017017073929, 'timing/model': 1.2105376739054918, 'learning_rate/qwen_vl_interface': 4.768548197747953e-06, 'learning_rate/action_model': 4.768548197747953e-05, 'epoch': 1.11}) +[2026-06-17 13:13:29,271][starVLA.training.train_starvla][INFO] - Step 3774, Loss: {'train/loss': 0.730013370513916, 'train/grad_norm_pre_clip': 13.296323776245117, 'timing/data': 0.004725032020360231, 'timing/model': 1.2056135709863156, 'learning_rate/qwen_vl_interface': 4.766396695010903e-06, 'learning_rate/action_model': 4.766396695010903e-05, 'epoch': 1.11}) +[2026-06-17 13:13:34,053][starVLA.training.train_starvla][INFO] - Step 3775, Loss: {'train/loss': 0.7234790325164795, 'train/grad_norm_pre_clip': 12.661201477050781, 'timing/data': 0.005228574969805777, 'timing/model': 1.2044216610956937, 'learning_rate/qwen_vl_interface': 4.764245292525337e-06, 'learning_rate/action_model': 4.7642452925253376e-05, 'epoch': 1.11}) +[2026-06-17 13:13:38,847][starVLA.training.train_starvla][INFO] - Step 3776, Loss: {'train/loss': 1.2307536602020264, 'train/grad_norm_pre_clip': 15.330191612243652, 'timing/data': 0.005473183002322912, 'timing/model': 1.2055561350425705, 'learning_rate/qwen_vl_interface': 4.762093990737247e-06, 'learning_rate/action_model': 4.762093990737246e-05, 'epoch': 1.11}) +[2026-06-17 13:13:43,633][starVLA.training.train_starvla][INFO] - Step 3777, Loss: {'train/loss': 0.4786909818649292, 'train/grad_norm_pre_clip': 9.47179126739502, 'timing/data': 0.005115878069773316, 'timing/model': 1.2155586211010814, 'learning_rate/qwen_vl_interface': 4.759942790092596e-06, 'learning_rate/action_model': 4.759942790092596e-05, 'epoch': 1.11}) +[2026-06-17 13:13:48,376][starVLA.training.train_starvla][INFO] - Step 3778, Loss: {'train/loss': 0.7573566436767578, 'train/grad_norm_pre_clip': 11.170929908752441, 'timing/data': 0.0055413569789379835, 'timing/model': 1.2909302150364965, 'learning_rate/qwen_vl_interface': 4.757791691037332e-06, 'learning_rate/action_model': 4.757791691037332e-05, 'epoch': 1.11}) +[2026-06-17 13:13:53,032][starVLA.training.train_starvla][INFO] - Step 3779, Loss: {'train/loss': 0.8506913781166077, 'train/grad_norm_pre_clip': 14.64479923248291, 'timing/data': 0.0052324129501357675, 'timing/model': 1.204035855946131, 'learning_rate/qwen_vl_interface': 4.755640694017381e-06, 'learning_rate/action_model': 4.755640694017381e-05, 'epoch': 1.11}) +[2026-06-17 13:13:57,808][starVLA.training.train_starvla][INFO] - Step 3780, Loss: {'train/loss': 0.5868886709213257, 'train/grad_norm_pre_clip': 11.062704086303711, 'timing/data': 0.004867983050644398, 'timing/model': 1.2029466760577634, 'learning_rate/qwen_vl_interface': 4.753489799478647e-06, 'learning_rate/action_model': 4.7534897994786464e-05, 'epoch': 1.11}) +[2026-06-17 13:14:02,595][starVLA.training.train_starvla][INFO] - Step 3781, Loss: {'train/loss': 0.7230075597763062, 'train/grad_norm_pre_clip': 11.917129516601562, 'timing/data': 0.005043182987719774, 'timing/model': 1.2031086439965293, 'learning_rate/qwen_vl_interface': 4.7513390078670105e-06, 'learning_rate/action_model': 4.75133900786701e-05, 'epoch': 1.11}) +[2026-06-17 13:14:07,374][starVLA.training.train_starvla][INFO] - Step 3782, Loss: {'train/loss': 0.46988803148269653, 'train/grad_norm_pre_clip': 9.586668014526367, 'timing/data': 0.005324174999259412, 'timing/model': 1.2023209600010887, 'learning_rate/qwen_vl_interface': 4.749188319628333e-06, 'learning_rate/action_model': 4.749188319628333e-05, 'epoch': 1.11}) +[2026-06-17 13:14:12,161][starVLA.training.train_starvla][INFO] - Step 3783, Loss: {'train/loss': 0.7571873664855957, 'train/grad_norm_pre_clip': 10.53109359741211, 'timing/data': 0.005262225982733071, 'timing/model': 1.3361666740383953, 'learning_rate/qwen_vl_interface': 4.747037735208457e-06, 'learning_rate/action_model': 4.7470377352084575e-05, 'epoch': 1.11}) +[2026-06-17 13:14:16,845][starVLA.training.train_starvla][INFO] - Step 3784, Loss: {'train/loss': 0.7910606861114502, 'train/grad_norm_pre_clip': 12.387295722961426, 'timing/data': 0.005431639030575752, 'timing/model': 1.2042163310106844, 'learning_rate/qwen_vl_interface': 4.7448872550532006e-06, 'learning_rate/action_model': 4.7448872550532e-05, 'epoch': 1.11}) +[2026-06-17 13:14:21,639][starVLA.training.train_starvla][INFO] - Step 3785, Loss: {'train/loss': 0.7427191734313965, 'train/grad_norm_pre_clip': 15.83341121673584, 'timing/data': 0.007433581980876625, 'timing/model': 1.2106144540011883, 'learning_rate/qwen_vl_interface': 4.74273687960836e-06, 'learning_rate/action_model': 4.74273687960836e-05, 'epoch': 1.11}) +[2026-06-17 13:14:26,431][starVLA.training.train_starvla][INFO] - Step 3786, Loss: {'train/loss': 0.5980557799339294, 'train/grad_norm_pre_clip': 12.131499290466309, 'timing/data': 0.005454918020404875, 'timing/model': 1.2050592900486663, 'learning_rate/qwen_vl_interface': 4.7405866093197075e-06, 'learning_rate/action_model': 4.740586609319707e-05, 'epoch': 1.11}) +[2026-06-17 13:14:31,137][starVLA.training.train_starvla][INFO] - Step 3787, Loss: {'train/loss': 0.7166012525558472, 'train/grad_norm_pre_clip': 13.978809356689453, 'timing/data': 0.005031365086324513, 'timing/model': 1.2037525660125539, 'learning_rate/qwen_vl_interface': 4.738436444633002e-06, 'learning_rate/action_model': 4.738436444633001e-05, 'epoch': 1.11}) +[2026-06-17 13:14:35,879][starVLA.training.train_starvla][INFO] - Step 3788, Loss: {'train/loss': 0.43430617451667786, 'train/grad_norm_pre_clip': 9.964046478271484, 'timing/data': 0.004955711076036096, 'timing/model': 1.2046174999559298, 'learning_rate/qwen_vl_interface': 4.73628638599397e-06, 'learning_rate/action_model': 4.73628638599397e-05, 'epoch': 1.11}) +[2026-06-17 13:14:40,677][starVLA.training.train_starvla][INFO] - Step 3789, Loss: {'train/loss': 0.6593680381774902, 'train/grad_norm_pre_clip': 11.914643287658691, 'timing/data': 0.005139451008290052, 'timing/model': 1.3389806130435318, 'learning_rate/qwen_vl_interface': 4.734136433848323e-06, 'learning_rate/action_model': 4.7341364338483224e-05, 'epoch': 1.11}) +[2026-06-17 13:14:45,334][starVLA.training.train_starvla][INFO] - Step 3790, Loss: {'train/loss': 0.7171244025230408, 'train/grad_norm_pre_clip': 14.516901969909668, 'timing/data': 0.005168581963516772, 'timing/model': 1.204603441990912, 'learning_rate/qwen_vl_interface': 4.731986588641748e-06, 'learning_rate/action_model': 4.7319865886417484e-05, 'epoch': 1.12}) +[2026-06-17 13:14:50,129][starVLA.training.train_starvla][INFO] - Step 3791, Loss: {'train/loss': 0.7044807076454163, 'train/grad_norm_pre_clip': 14.537501335144043, 'timing/data': 0.004909179988317192, 'timing/model': 1.2065612780861557, 'learning_rate/qwen_vl_interface': 4.729836850819912e-06, 'learning_rate/action_model': 4.729836850819912e-05, 'epoch': 1.12}) +[2026-06-17 13:14:54,945][starVLA.training.train_starvla][INFO] - Step 3792, Loss: {'train/loss': 0.7342355251312256, 'train/grad_norm_pre_clip': 13.987232208251953, 'timing/data': 0.004995297989808023, 'timing/model': 1.2022057429421693, 'learning_rate/qwen_vl_interface': 4.727687220828458e-06, 'learning_rate/action_model': 4.727687220828458e-05, 'epoch': 1.12}) +[2026-06-17 13:14:59,723][starVLA.training.train_starvla][INFO] - Step 3793, Loss: {'train/loss': 0.746056854724884, 'train/grad_norm_pre_clip': 19.05487823486328, 'timing/data': 0.0055959150195121765, 'timing/model': 1.2065326180309057, 'learning_rate/qwen_vl_interface': 4.7255376991130005e-06, 'learning_rate/action_model': 4.7255376991130005e-05, 'epoch': 1.12}) +[2026-06-17 13:15:04,473][starVLA.training.train_starvla][INFO] - Step 3794, Loss: {'train/loss': 0.7949806451797485, 'train/grad_norm_pre_clip': 12.004806518554688, 'timing/data': 0.005329858977347612, 'timing/model': 1.203547234996222, 'learning_rate/qwen_vl_interface': 4.723388286119147e-06, 'learning_rate/action_model': 4.723388286119146e-05, 'epoch': 1.12}) +[2026-06-17 13:15:09,254][starVLA.training.train_starvla][INFO] - Step 3795, Loss: {'train/loss': 0.7020630240440369, 'train/grad_norm_pre_clip': 26.27774429321289, 'timing/data': 0.0051472639897838235, 'timing/model': 1.3288508410332724, 'learning_rate/qwen_vl_interface': 4.721238982292467e-06, 'learning_rate/action_model': 4.721238982292466e-05, 'epoch': 1.12}) +[2026-06-17 13:15:13,913][starVLA.training.train_starvla][INFO] - Step 3796, Loss: {'train/loss': 0.6815916895866394, 'train/grad_norm_pre_clip': 11.916034698486328, 'timing/data': 0.005368258920498192, 'timing/model': 1.204232249991037, 'learning_rate/qwen_vl_interface': 4.719089788078514e-06, 'learning_rate/action_model': 4.719089788078514e-05, 'epoch': 1.12}) +[2026-06-17 13:15:18,701][starVLA.training.train_starvla][INFO] - Step 3797, Loss: {'train/loss': 0.6895865797996521, 'train/grad_norm_pre_clip': 12.396495819091797, 'timing/data': 0.005886705010198057, 'timing/model': 1.2037002770230174, 'learning_rate/qwen_vl_interface': 4.716940703922818e-06, 'learning_rate/action_model': 4.716940703922818e-05, 'epoch': 1.12}) +[2026-06-17 13:15:23,665][starVLA.training.train_starvla][INFO] - Step 3798, Loss: {'train/loss': 0.4841585159301758, 'train/grad_norm_pre_clip': 9.15889835357666, 'timing/data': 0.004563549067825079, 'timing/model': 1.2116741680074483, 'learning_rate/qwen_vl_interface': 4.714791730270889e-06, 'learning_rate/action_model': 4.71479173027089e-05, 'epoch': 1.12}) +[2026-06-17 13:15:28,454][starVLA.training.train_starvla][INFO] - Step 3799, Loss: {'train/loss': 0.7476494312286377, 'train/grad_norm_pre_clip': 12.475286483764648, 'timing/data': 0.004849179065786302, 'timing/model': 1.202050901018083, 'learning_rate/qwen_vl_interface': 4.712642867568213e-06, 'learning_rate/action_model': 4.712642867568212e-05, 'epoch': 1.12}) +[2026-06-17 13:17:10,209][starVLA.training.train_starvla][INFO] - Step 3800, Loss: {'train/loss': 0.48536407947540283, 'train/grad_norm_pre_clip': 10.176009178161621, 'eval/loss': 0.7457560896873474, 'eval/latency_0/loss': 0.7457560896873474, 'eval/demon_fire_precision': 0.9150272080368355, 'eval/demon_fire_recall': 0.9078073089700996, 'eval/demon_fire_f1': 0.911402960183448, 'eval/demon_fire_rate_model': 0.664349276974416, 'eval/demon_fire_rate_teacher': 0.6696329254727474, 'eval/demon_left_precision': 0.792778649921507, 'eval/demon_left_recall': 0.8583569405099151, 'eval/demon_left_f1': 0.8242655059847661, 'eval/demon_left_rate_model': 0.5314238042269188, 'eval/demon_left_rate_teacher': 0.49082313681868744, 'eval/demon_right_precision': 0.8433373349339736, 'eval/demon_right_recall': 0.7937853107344632, 'eval/demon_right_f1': 0.8178114086146683, 'eval/demon_right_rate_model': 0.4632925472747497, 'eval/demon_right_rate_teacher': 0.4922135706340378, 'eval/demon_move_f1': 0.8210384572997171, 'eval/demon_cc_f1': 0.8822717847099397, 'eval/latency_0/demon_fire_precision': 0.9150272080368355, 'eval/latency_0/demon_fire_recall': 0.9078073089700996, 'eval/latency_0/demon_fire_f1': 0.911402960183448, 'eval/latency_0/demon_fire_rate_model': 0.664349276974416, 'eval/latency_0/demon_fire_rate_teacher': 0.6696329254727474, 'eval/latency_0/demon_left_precision': 0.792778649921507, 'eval/latency_0/demon_left_recall': 0.8583569405099151, 'eval/latency_0/demon_left_f1': 0.8242655059847661, 'eval/latency_0/demon_left_rate_model': 0.5314238042269188, 'eval/latency_0/demon_left_rate_teacher': 0.49082313681868744, 'eval/latency_0/demon_right_precision': 0.8433373349339736, 'eval/latency_0/demon_right_recall': 0.7937853107344632, 'eval/latency_0/demon_right_f1': 0.8178114086146683, 'eval/latency_0/demon_right_rate_model': 0.4632925472747497, 'eval/latency_0/demon_right_rate_teacher': 0.4922135706340378, 'eval/latency_0/demon_move_f1': 0.8210384572997171, 'eval/latency_0/demon_cc_f1': 0.8822717847099397, 'timing/data': 0.005144583992660046, 'timing/model': 1.3335849250433967, 'learning_rate/qwen_vl_interface': 4.7104941162602436e-06, 'learning_rate/action_model': 4.710494116260244e-05, 'epoch': 1.12}) +[2026-06-17 13:17:10,210][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_3800_state +[2026-06-17 13:17:10,213][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 13:17:13,775][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_3800_state/model.safetensors +[2026-06-17 13:17:21,522][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_3800_state/optimizer.bin +[2026-06-17 13:17:21,522][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_3800_state/sampler.bin +[2026-06-17 13:17:21,522][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_3800_state/sampler_1.bin +[2026-06-17 13:17:21,523][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_3800_state/random_states_0.pkl +[2026-06-17 13:17:21,569][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3800 +[2026-06-17 13:17:21,569][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 13:17:21,607][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 13:17:26,211][starVLA.training.train_starvla][INFO] - Step 3801, Loss: {'train/loss': 0.869962751865387, 'train/grad_norm_pre_clip': 14.571163177490234, 'timing/data': 0.004957693978212774, 'timing/model': 1.1860483690397814, 'learning_rate/qwen_vl_interface': 4.708345476792428e-06, 'learning_rate/action_model': 4.7083454767924276e-05, 'epoch': 1.12}) +[2026-06-17 13:17:30,830][starVLA.training.train_starvla][INFO] - Step 3802, Loss: {'train/loss': 0.6955544948577881, 'train/grad_norm_pre_clip': 12.10511302947998, 'timing/data': 0.005285399034619331, 'timing/model': 1.2119425940327346, 'learning_rate/qwen_vl_interface': 4.706196949610178e-06, 'learning_rate/action_model': 4.7061969496101776e-05, 'epoch': 1.12}) +[2026-06-17 13:17:35,422][starVLA.training.train_starvla][INFO] - Step 3803, Loss: {'train/loss': 1.015119194984436, 'train/grad_norm_pre_clip': 14.134923934936523, 'timing/data': 0.0050438460893929005, 'timing/model': 1.189085362944752, 'learning_rate/qwen_vl_interface': 4.7040485351588845e-06, 'learning_rate/action_model': 4.7040485351588845e-05, 'epoch': 1.12}) +[2026-06-17 13:17:40,064][starVLA.training.train_starvla][INFO] - Step 3804, Loss: {'train/loss': 0.6919665932655334, 'train/grad_norm_pre_clip': 12.84992504119873, 'timing/data': 0.005015685921534896, 'timing/model': 1.1900553279556334, 'learning_rate/qwen_vl_interface': 4.701900233883916e-06, 'learning_rate/action_model': 4.7019002338839156e-05, 'epoch': 1.12}) +[2026-06-17 13:17:44,734][starVLA.training.train_starvla][INFO] - Step 3805, Loss: {'train/loss': 0.6955718994140625, 'train/grad_norm_pre_clip': 15.675603866577148, 'timing/data': 0.004481711075641215, 'timing/model': 1.1909263640409335, 'learning_rate/qwen_vl_interface': 4.69975204623062e-06, 'learning_rate/action_model': 4.69975204623062e-05, 'epoch': 1.12}) +[2026-06-17 13:17:49,440][starVLA.training.train_starvla][INFO] - Step 3806, Loss: {'train/loss': 0.7911056876182556, 'train/grad_norm_pre_clip': 13.576570510864258, 'timing/data': 0.004722060984931886, 'timing/model': 1.1940547139383852, 'learning_rate/qwen_vl_interface': 4.697603972644319e-06, 'learning_rate/action_model': 4.697603972644318e-05, 'epoch': 1.12}) +[2026-06-17 13:17:54,216][starVLA.training.train_starvla][INFO] - Step 3807, Loss: {'train/loss': 0.8785266876220703, 'train/grad_norm_pre_clip': 14.307612419128418, 'timing/data': 0.004708759021013975, 'timing/model': 1.3250060359714553, 'learning_rate/qwen_vl_interface': 4.695456013570306e-06, 'learning_rate/action_model': 4.695456013570306e-05, 'epoch': 1.12}) +[2026-06-17 13:17:58,828][starVLA.training.train_starvla][INFO] - Step 3808, Loss: {'train/loss': 0.8826708197593689, 'train/grad_norm_pre_clip': 15.902690887451172, 'timing/data': 0.004717824049293995, 'timing/model': 1.1952147639822215, 'learning_rate/qwen_vl_interface': 4.693308169453858e-06, 'learning_rate/action_model': 4.693308169453858e-05, 'epoch': 1.12}) +[2026-06-17 13:18:03,609][starVLA.training.train_starvla][INFO] - Step 3809, Loss: {'train/loss': 0.6509819030761719, 'train/grad_norm_pre_clip': 14.282637596130371, 'timing/data': 0.005004922044463456, 'timing/model': 1.1955194100737572, 'learning_rate/qwen_vl_interface': 4.691160440740226e-06, 'learning_rate/action_model': 4.691160440740226e-05, 'epoch': 1.12}) +[2026-06-17 13:18:08,377][starVLA.training.train_starvla][INFO] - Step 3810, Loss: {'train/loss': 0.9519850015640259, 'train/grad_norm_pre_clip': 14.027290344238281, 'timing/data': 0.005025232909247279, 'timing/model': 1.2000126170460135, 'learning_rate/qwen_vl_interface': 4.689012827874637e-06, 'learning_rate/action_model': 4.689012827874637e-05, 'epoch': 1.12}) +[2026-06-17 13:18:13,125][starVLA.training.train_starvla][INFO] - Step 3811, Loss: {'train/loss': 0.722560465335846, 'train/grad_norm_pre_clip': 11.481202125549316, 'timing/data': 0.005014974041841924, 'timing/model': 1.198300332063809, 'learning_rate/qwen_vl_interface': 4.6868653313022905e-06, 'learning_rate/action_model': 4.68686533130229e-05, 'epoch': 1.12}) +[2026-06-17 13:18:17,896][starVLA.training.train_starvla][INFO] - Step 3812, Loss: {'train/loss': 0.923270046710968, 'train/grad_norm_pre_clip': 13.662193298339844, 'timing/data': 0.004773745080456138, 'timing/model': 1.1984153960365802, 'learning_rate/qwen_vl_interface': 4.684717951468368e-06, 'learning_rate/action_model': 4.6847179514683684e-05, 'epoch': 1.12}) +[2026-06-17 13:18:22,679][starVLA.training.train_starvla][INFO] - Step 3813, Loss: {'train/loss': 0.7102460861206055, 'train/grad_norm_pre_clip': 13.014080047607422, 'timing/data': 0.004890416981652379, 'timing/model': 1.3429806519998237, 'learning_rate/qwen_vl_interface': 4.682570688818024e-06, 'learning_rate/action_model': 4.682570688818024e-05, 'epoch': 1.12}) +[2026-06-17 13:18:27,318][starVLA.training.train_starvla][INFO] - Step 3814, Loss: {'train/loss': 0.8843472599983215, 'train/grad_norm_pre_clip': 12.117147445678711, 'timing/data': 0.0052745750872418284, 'timing/model': 1.199323571054265, 'learning_rate/qwen_vl_interface': 4.680423543796385e-06, 'learning_rate/action_model': 4.680423543796385e-05, 'epoch': 1.12}) +[2026-06-17 13:18:32,092][starVLA.training.train_starvla][INFO] - Step 3815, Loss: {'train/loss': 0.6400517225265503, 'train/grad_norm_pre_clip': 15.06847095489502, 'timing/data': 0.005065592005848885, 'timing/model': 1.201854121987708, 'learning_rate/qwen_vl_interface': 4.678276516848558e-06, 'learning_rate/action_model': 4.678276516848557e-05, 'epoch': 1.12}) +[2026-06-17 13:18:36,883][starVLA.training.train_starvla][INFO] - Step 3816, Loss: {'train/loss': 0.5826115012168884, 'train/grad_norm_pre_clip': 17.591981887817383, 'timing/data': 0.0047249950002878904, 'timing/model': 1.2075770660303533, 'learning_rate/qwen_vl_interface': 4.676129608419625e-06, 'learning_rate/action_model': 4.676129608419625e-05, 'epoch': 1.12}) +[2026-06-17 13:18:41,616][starVLA.training.train_starvla][INFO] - Step 3817, Loss: {'train/loss': 0.5818800926208496, 'train/grad_norm_pre_clip': 10.343245506286621, 'timing/data': 0.005282780970446765, 'timing/model': 1.2015249520773068, 'learning_rate/qwen_vl_interface': 4.6739828189546415e-06, 'learning_rate/action_model': 4.6739828189546417e-05, 'epoch': 1.12}) +[2026-06-17 13:18:46,417][starVLA.training.train_starvla][INFO] - Step 3818, Loss: {'train/loss': 0.6119703650474548, 'train/grad_norm_pre_clip': 13.198654174804688, 'timing/data': 0.004904998932033777, 'timing/model': 1.2040483230957761, 'learning_rate/qwen_vl_interface': 4.671836148898641e-06, 'learning_rate/action_model': 4.6718361488986404e-05, 'epoch': 1.12}) +[2026-06-17 13:18:51,205][starVLA.training.train_starvla][INFO] - Step 3819, Loss: {'train/loss': 0.6937891244888306, 'train/grad_norm_pre_clip': 10.542097091674805, 'timing/data': 0.004591227974742651, 'timing/model': 1.3407374490052462, 'learning_rate/qwen_vl_interface': 4.6696895986966254e-06, 'learning_rate/action_model': 4.669689598696625e-05, 'epoch': 1.12}) +[2026-06-17 13:18:55,868][starVLA.training.train_starvla][INFO] - Step 3820, Loss: {'train/loss': 0.5508503317832947, 'train/grad_norm_pre_clip': 9.634980201721191, 'timing/data': 0.0046736199874430895, 'timing/model': 1.2034505989868194, 'learning_rate/qwen_vl_interface': 4.667543168793583e-06, 'learning_rate/action_model': 4.667543168793583e-05, 'epoch': 1.12}) +[2026-06-17 13:19:00,650][starVLA.training.train_starvla][INFO] - Step 3821, Loss: {'train/loss': 0.67087322473526, 'train/grad_norm_pre_clip': 11.538012504577637, 'timing/data': 0.004746431950479746, 'timing/model': 1.2043723710812628, 'learning_rate/qwen_vl_interface': 4.665396859634468e-06, 'learning_rate/action_model': 4.665396859634468e-05, 'epoch': 1.12}) +[2026-06-17 13:19:05,426][starVLA.training.train_starvla][INFO] - Step 3822, Loss: {'train/loss': 0.8018196821212769, 'train/grad_norm_pre_clip': 34.10982894897461, 'timing/data': 0.005051292944699526, 'timing/model': 1.202186553971842, 'learning_rate/qwen_vl_interface': 4.663250671664212e-06, 'learning_rate/action_model': 4.663250671664211e-05, 'epoch': 1.12}) +[2026-06-17 13:19:10,221][starVLA.training.train_starvla][INFO] - Step 3823, Loss: {'train/loss': 0.5753822922706604, 'train/grad_norm_pre_clip': 10.1560640335083, 'timing/data': 0.005085821961984038, 'timing/model': 1.2073079330148175, 'learning_rate/qwen_vl_interface': 4.6611046053277245e-06, 'learning_rate/action_model': 4.661104605327724e-05, 'epoch': 1.12}) +[2026-06-17 13:19:14,890][starVLA.training.train_starvla][INFO] - Step 3824, Loss: {'train/loss': 0.6872761249542236, 'train/grad_norm_pre_clip': 12.567642211914062, 'timing/data': 0.004758161958307028, 'timing/model': 1.2044339700369164, 'learning_rate/qwen_vl_interface': 4.658958661069886e-06, 'learning_rate/action_model': 4.6589586610698856e-05, 'epoch': 1.13}) +[2026-06-17 13:19:19,599][starVLA.training.train_starvla][INFO] - Step 3825, Loss: {'train/loss': 0.5736896991729736, 'train/grad_norm_pre_clip': 10.59375286102295, 'timing/data': 0.005093613988719881, 'timing/model': 1.248662436963059, 'learning_rate/qwen_vl_interface': 4.6568128393355535e-06, 'learning_rate/action_model': 4.656812839335553e-05, 'epoch': 1.13}) +[2026-06-17 13:19:24,267][starVLA.training.train_starvla][INFO] - Step 3826, Loss: {'train/loss': 0.8517073392868042, 'train/grad_norm_pre_clip': 14.344578742980957, 'timing/data': 0.0049161280039697886, 'timing/model': 1.2046398209640756, 'learning_rate/qwen_vl_interface': 4.654667140569555e-06, 'learning_rate/action_model': 4.6546671405695544e-05, 'epoch': 1.13}) +[2026-06-17 13:19:29,027][starVLA.training.train_starvla][INFO] - Step 3827, Loss: {'train/loss': 1.1102570295333862, 'train/grad_norm_pre_clip': 14.832294464111328, 'timing/data': 0.00485810695681721, 'timing/model': 1.2117414749227464, 'learning_rate/qwen_vl_interface': 4.652521565216701e-06, 'learning_rate/action_model': 4.652521565216701e-05, 'epoch': 1.13}) +[2026-06-17 13:19:33,826][starVLA.training.train_starvla][INFO] - Step 3828, Loss: {'train/loss': 0.7004314064979553, 'train/grad_norm_pre_clip': 11.1983060836792, 'timing/data': 0.00478765310253948, 'timing/model': 1.206201542983763, 'learning_rate/qwen_vl_interface': 4.65037611372177e-06, 'learning_rate/action_model': 4.650376113721769e-05, 'epoch': 1.13}) +[2026-06-17 13:19:38,605][starVLA.training.train_starvla][INFO] - Step 3829, Loss: {'train/loss': 0.7274613976478577, 'train/grad_norm_pre_clip': 12.988760948181152, 'timing/data': 0.004822136950679123, 'timing/model': 1.199958362034522, 'learning_rate/qwen_vl_interface': 4.648230786529514e-06, 'learning_rate/action_model': 4.648230786529513e-05, 'epoch': 1.13}) +[2026-06-17 13:19:43,392][starVLA.training.train_starvla][INFO] - Step 3830, Loss: {'train/loss': 0.9383824467658997, 'train/grad_norm_pre_clip': 15.714377403259277, 'timing/data': 0.00497920298948884, 'timing/model': 1.3398135609459132, 'learning_rate/qwen_vl_interface': 4.646085584084664e-06, 'learning_rate/action_model': 4.6460855840846637e-05, 'epoch': 1.13}) +[2026-06-17 13:19:48,108][starVLA.training.train_starvla][INFO] - Step 3831, Loss: {'train/loss': 0.9443861246109009, 'train/grad_norm_pre_clip': 14.4844388961792, 'timing/data': 0.0048013749765232205, 'timing/model': 1.2604663849342614, 'learning_rate/qwen_vl_interface': 4.643940506831923e-06, 'learning_rate/action_model': 4.6439405068319226e-05, 'epoch': 1.13}) +[2026-06-17 13:19:52,783][starVLA.training.train_starvla][INFO] - Step 3832, Loss: {'train/loss': 0.5755872130393982, 'train/grad_norm_pre_clip': 12.257755279541016, 'timing/data': 0.004805299919098616, 'timing/model': 1.2091598600381985, 'learning_rate/qwen_vl_interface': 4.641795555215967e-06, 'learning_rate/action_model': 4.6417955552159676e-05, 'epoch': 1.13}) +[2026-06-17 13:19:57,534][starVLA.training.train_starvla][INFO] - Step 3833, Loss: {'train/loss': 0.5532234907150269, 'train/grad_norm_pre_clip': 9.384196281433105, 'timing/data': 0.0048873970517888665, 'timing/model': 1.2025342200649902, 'learning_rate/qwen_vl_interface': 4.639650729681445e-06, 'learning_rate/action_model': 4.6396507296814454e-05, 'epoch': 1.13}) +[2026-06-17 13:20:02,331][starVLA.training.train_starvla][INFO] - Step 3834, Loss: {'train/loss': 0.4882151186466217, 'train/grad_norm_pre_clip': 11.134380340576172, 'timing/data': 0.004798915004357696, 'timing/model': 1.2166915360139683, 'learning_rate/qwen_vl_interface': 4.637506030672987e-06, 'learning_rate/action_model': 4.637506030672987e-05, 'epoch': 1.13}) +[2026-06-17 13:20:07,148][starVLA.training.train_starvla][INFO] - Step 3835, Loss: {'train/loss': 0.8568626642227173, 'train/grad_norm_pre_clip': 11.792510986328125, 'timing/data': 0.005087874946184456, 'timing/model': 1.2343490769853815, 'learning_rate/qwen_vl_interface': 4.635361458635186e-06, 'learning_rate/action_model': 4.6353614586351864e-05, 'epoch': 1.13}) +[2026-06-17 13:20:11,972][starVLA.training.train_starvla][INFO] - Step 3836, Loss: {'train/loss': 0.7106266021728516, 'train/grad_norm_pre_clip': 11.974047660827637, 'timing/data': 0.004889530944637954, 'timing/model': 1.3463110519805923, 'learning_rate/qwen_vl_interface': 4.633217014012616e-06, 'learning_rate/action_model': 4.633217014012616e-05, 'epoch': 1.13}) +[2026-06-17 13:20:16,753][starVLA.training.train_starvla][INFO] - Step 3837, Loss: {'train/loss': 0.632515549659729, 'train/grad_norm_pre_clip': 10.138711929321289, 'timing/data': 0.005139601998962462, 'timing/model': 1.3224330219672993, 'learning_rate/qwen_vl_interface': 4.631072697249822e-06, 'learning_rate/action_model': 4.631072697249822e-05, 'epoch': 1.13}) +[2026-06-17 13:20:21,402][starVLA.training.train_starvla][INFO] - Step 3838, Loss: {'train/loss': 0.6016939878463745, 'train/grad_norm_pre_clip': 9.630754470825195, 'timing/data': 0.004846888012252748, 'timing/model': 1.2051785580115393, 'learning_rate/qwen_vl_interface': 4.628928508791325e-06, 'learning_rate/action_model': 4.628928508791324e-05, 'epoch': 1.13}) +[2026-06-17 13:20:26,159][starVLA.training.train_starvla][INFO] - Step 3839, Loss: {'train/loss': 0.666336178779602, 'train/grad_norm_pre_clip': 12.100282669067383, 'timing/data': 0.004771498031914234, 'timing/model': 1.2042133690556511, 'learning_rate/qwen_vl_interface': 4.626784449081617e-06, 'learning_rate/action_model': 4.626784449081617e-05, 'epoch': 1.13}) +[2026-06-17 13:20:30,957][starVLA.training.train_starvla][INFO] - Step 3840, Loss: {'train/loss': 0.7797108292579651, 'train/grad_norm_pre_clip': 13.442048072814941, 'timing/data': 0.004964674008078873, 'timing/model': 1.205362157896161, 'learning_rate/qwen_vl_interface': 4.62464051856516e-06, 'learning_rate/action_model': 4.6246405185651596e-05, 'epoch': 1.13}) +[2026-06-17 13:20:35,753][starVLA.training.train_starvla][INFO] - Step 3841, Loss: {'train/loss': 0.5112490057945251, 'train/grad_norm_pre_clip': 10.894776344299316, 'timing/data': 0.005003473022952676, 'timing/model': 1.2033986140741035, 'learning_rate/qwen_vl_interface': 4.622496717686401e-06, 'learning_rate/action_model': 4.6224967176864005e-05, 'epoch': 1.13}) +[2026-06-17 13:20:40,488][starVLA.training.train_starvla][INFO] - Step 3842, Loss: {'train/loss': 0.6955907940864563, 'train/grad_norm_pre_clip': 10.747499465942383, 'timing/data': 0.005193967022933066, 'timing/model': 1.2816422879695892, 'learning_rate/qwen_vl_interface': 4.620353046889745e-06, 'learning_rate/action_model': 4.620353046889745e-05, 'epoch': 1.13}) +[2026-06-17 13:20:45,154][starVLA.training.train_starvla][INFO] - Step 3843, Loss: {'train/loss': 0.6643280386924744, 'train/grad_norm_pre_clip': 12.297651290893555, 'timing/data': 0.004805561969988048, 'timing/model': 1.206122009898536, 'learning_rate/qwen_vl_interface': 4.6182095066195805e-06, 'learning_rate/action_model': 4.61820950661958e-05, 'epoch': 1.13}) +[2026-06-17 13:20:49,918][starVLA.training.train_starvla][INFO] - Step 3844, Loss: {'train/loss': 0.7261727452278137, 'train/grad_norm_pre_clip': 12.5897855758667, 'timing/data': 0.004990371991880238, 'timing/model': 1.2040922290179878, 'learning_rate/qwen_vl_interface': 4.616066097320264e-06, 'learning_rate/action_model': 4.6160660973202636e-05, 'epoch': 1.13}) +[2026-06-17 13:20:54,707][starVLA.training.train_starvla][INFO] - Step 3845, Loss: {'train/loss': 0.730678379535675, 'train/grad_norm_pre_clip': 12.232158660888672, 'timing/data': 0.005090903025120497, 'timing/model': 1.2051700330339372, 'learning_rate/qwen_vl_interface': 4.613922819436129e-06, 'learning_rate/action_model': 4.613922819436128e-05, 'epoch': 1.13}) +[2026-06-17 13:20:59,489][starVLA.training.train_starvla][INFO] - Step 3846, Loss: {'train/loss': 0.6975432634353638, 'train/grad_norm_pre_clip': 13.152480125427246, 'timing/data': 0.005209725932218134, 'timing/model': 1.2046024329029024, 'learning_rate/qwen_vl_interface': 4.611779673411478e-06, 'learning_rate/action_model': 4.6117796734114785e-05, 'epoch': 1.13}) +[2026-06-17 13:21:04,294][starVLA.training.train_starvla][INFO] - Step 3847, Loss: {'train/loss': 0.5398935675621033, 'train/grad_norm_pre_clip': 16.486364364624023, 'timing/data': 0.005171751021407545, 'timing/model': 1.340952554019168, 'learning_rate/qwen_vl_interface': 4.609636659690588e-06, 'learning_rate/action_model': 4.609636659690587e-05, 'epoch': 1.13}) +[2026-06-17 13:21:09,037][starVLA.training.train_starvla][INFO] - Step 3848, Loss: {'train/loss': 0.6523999571800232, 'train/grad_norm_pre_clip': 16.918642044067383, 'timing/data': 0.004714397015050054, 'timing/model': 1.2882208429509774, 'learning_rate/qwen_vl_interface': 4.607493778717705e-06, 'learning_rate/action_model': 4.6074937787177055e-05, 'epoch': 1.13}) +[2026-06-17 13:21:13,697][starVLA.training.train_starvla][INFO] - Step 3849, Loss: {'train/loss': 0.665458083152771, 'train/grad_norm_pre_clip': 11.601947784423828, 'timing/data': 0.004709419910795987, 'timing/model': 1.2047344479942694, 'learning_rate/qwen_vl_interface': 4.605351030937054e-06, 'learning_rate/action_model': 4.605351030937055e-05, 'epoch': 1.13}) +[2026-06-17 13:21:18,492][starVLA.training.train_starvla][INFO] - Step 3850, Loss: {'train/loss': 0.7862018346786499, 'train/grad_norm_pre_clip': 16.267681121826172, 'timing/data': 0.004802018986083567, 'timing/model': 1.2024083270225674, 'learning_rate/qwen_vl_interface': 4.60320841679283e-06, 'learning_rate/action_model': 4.60320841679283e-05, 'epoch': 1.13}) +[2026-06-17 13:21:23,283][starVLA.training.train_starvla][INFO] - Step 3851, Loss: {'train/loss': 0.4563547968864441, 'train/grad_norm_pre_clip': 7.82895565032959, 'timing/data': 0.004817608976736665, 'timing/model': 1.2120799920521677, 'learning_rate/qwen_vl_interface': 4.6010659367291945e-06, 'learning_rate/action_model': 4.601065936729194e-05, 'epoch': 1.13}) +[2026-06-17 13:21:28,079][starVLA.training.train_starvla][INFO] - Step 3852, Loss: {'train/loss': 0.5665733814239502, 'train/grad_norm_pre_clip': 10.737290382385254, 'timing/data': 0.005031525972299278, 'timing/model': 1.2048009920399636, 'learning_rate/qwen_vl_interface': 4.598923591190289e-06, 'learning_rate/action_model': 4.598923591190289e-05, 'epoch': 1.13}) +[2026-06-17 13:21:32,854][starVLA.training.train_starvla][INFO] - Step 3853, Loss: {'train/loss': 0.6222118735313416, 'train/grad_norm_pre_clip': 11.660918235778809, 'timing/data': 0.004958512960001826, 'timing/model': 1.2065193870803341, 'learning_rate/qwen_vl_interface': 4.596781380620225e-06, 'learning_rate/action_model': 4.5967813806202245e-05, 'epoch': 1.13}) +[2026-06-17 13:21:37,609][starVLA.training.train_starvla][INFO] - Step 3854, Loss: {'train/loss': 0.7256673574447632, 'train/grad_norm_pre_clip': 12.274970054626465, 'timing/data': 0.004884375026449561, 'timing/model': 1.301176434964873, 'learning_rate/qwen_vl_interface': 4.5946393054630845e-06, 'learning_rate/action_model': 4.5946393054630845e-05, 'epoch': 1.13}) +[2026-06-17 13:21:42,289][starVLA.training.train_starvla][INFO] - Step 3855, Loss: {'train/loss': 0.6624399423599243, 'train/grad_norm_pre_clip': 10.648750305175781, 'timing/data': 0.004848201991990209, 'timing/model': 1.2060883860103786, 'learning_rate/qwen_vl_interface': 4.592497366162919e-06, 'learning_rate/action_model': 4.592497366162918e-05, 'epoch': 1.13}) +[2026-06-17 13:21:47,091][starVLA.training.train_starvla][INFO] - Step 3856, Loss: {'train/loss': 0.9019106030464172, 'train/grad_norm_pre_clip': 15.119363784790039, 'timing/data': 0.00482935004401952, 'timing/model': 1.209987638052553, 'learning_rate/qwen_vl_interface': 4.590355563163758e-06, 'learning_rate/action_model': 4.5903555631637574e-05, 'epoch': 1.13}) +[2026-06-17 13:21:51,887][starVLA.training.train_starvla][INFO] - Step 3857, Loss: {'train/loss': 0.5528842210769653, 'train/grad_norm_pre_clip': 10.391231536865234, 'timing/data': 0.00492932100314647, 'timing/model': 1.2059049520175904, 'learning_rate/qwen_vl_interface': 4.588213896909598e-06, 'learning_rate/action_model': 4.588213896909598e-05, 'epoch': 1.13}) +[2026-06-17 13:21:56,682][starVLA.training.train_starvla][INFO] - Step 3858, Loss: {'train/loss': 0.48650991916656494, 'train/grad_norm_pre_clip': 9.744729042053223, 'timing/data': 0.0050001079216599464, 'timing/model': 1.2054688020143658, 'learning_rate/qwen_vl_interface': 4.58607236784441e-06, 'learning_rate/action_model': 4.58607236784441e-05, 'epoch': 1.14}) +[2026-06-17 13:22:01,482][starVLA.training.train_starvla][INFO] - Step 3859, Loss: {'train/loss': 0.8087767362594604, 'train/grad_norm_pre_clip': 13.619206428527832, 'timing/data': 0.0054138669511303306, 'timing/model': 1.3419467179337516, 'learning_rate/qwen_vl_interface': 4.583930976412133e-06, 'learning_rate/action_model': 4.583930976412133e-05, 'epoch': 1.14}) +[2026-06-17 13:22:06,158][starVLA.training.train_starvla][INFO] - Step 3860, Loss: {'train/loss': 0.7606400847434998, 'train/grad_norm_pre_clip': 11.755477905273438, 'timing/data': 0.004977497039362788, 'timing/model': 1.2070221409667283, 'learning_rate/qwen_vl_interface': 4.5817897230566834e-06, 'learning_rate/action_model': 4.581789723056683e-05, 'epoch': 1.14}) +[2026-06-17 13:22:10,955][starVLA.training.train_starvla][INFO] - Step 3861, Loss: {'train/loss': 0.7509185075759888, 'train/grad_norm_pre_clip': 15.745698928833008, 'timing/data': 0.004929299000650644, 'timing/model': 1.2086516689741984, 'learning_rate/qwen_vl_interface': 4.579648608221943e-06, 'learning_rate/action_model': 4.579648608221943e-05, 'epoch': 1.14}) +[2026-06-17 13:22:15,768][starVLA.training.train_starvla][INFO] - Step 3862, Loss: {'train/loss': 0.7387750744819641, 'train/grad_norm_pre_clip': 12.098876953125, 'timing/data': 0.005043118959292769, 'timing/model': 1.210145112941973, 'learning_rate/qwen_vl_interface': 4.577507632351766e-06, 'learning_rate/action_model': 4.5775076323517664e-05, 'epoch': 1.14}) +[2026-06-17 13:22:20,481][starVLA.training.train_starvla][INFO] - Step 3863, Loss: {'train/loss': 0.5190020799636841, 'train/grad_norm_pre_clip': 9.872360229492188, 'timing/data': 0.004491043975576758, 'timing/model': 1.2063332030083984, 'learning_rate/qwen_vl_interface': 4.575366795889981e-06, 'learning_rate/action_model': 4.575366795889981e-05, 'epoch': 1.14}) +[2026-06-17 13:22:25,262][starVLA.training.train_starvla][INFO] - Step 3864, Loss: {'train/loss': 0.9287043809890747, 'train/grad_norm_pre_clip': 16.268543243408203, 'timing/data': 0.005022572004236281, 'timing/model': 1.214603008935228, 'learning_rate/qwen_vl_interface': 4.573226099280386e-06, 'learning_rate/action_model': 4.573226099280386e-05, 'epoch': 1.14}) +[2026-06-17 13:22:30,052][starVLA.training.train_starvla][INFO] - Step 3865, Loss: {'train/loss': 0.7924812436103821, 'train/grad_norm_pre_clip': 14.359297752380371, 'timing/data': 0.0049413410015404224, 'timing/model': 1.3381249189842492, 'learning_rate/qwen_vl_interface': 4.571085542966748e-06, 'learning_rate/action_model': 4.5710855429667485e-05, 'epoch': 1.14}) +[2026-06-17 13:22:34,801][starVLA.training.train_starvla][INFO] - Step 3866, Loss: {'train/loss': 0.7274987101554871, 'train/grad_norm_pre_clip': 13.674530982971191, 'timing/data': 0.004844804061576724, 'timing/model': 1.2131683659972623, 'learning_rate/qwen_vl_interface': 4.568945127392807e-06, 'learning_rate/action_model': 4.5689451273928066e-05, 'epoch': 1.14}) +[2026-06-17 13:22:39,588][starVLA.training.train_starvla][INFO] - Step 3867, Loss: {'train/loss': 0.6478965282440186, 'train/grad_norm_pre_clip': 15.403038024902344, 'timing/data': 0.005038580042310059, 'timing/model': 1.2051750459941104, 'learning_rate/qwen_vl_interface': 4.566804853002275e-06, 'learning_rate/action_model': 4.566804853002275e-05, 'epoch': 1.14}) +[2026-06-17 13:22:44,373][starVLA.training.train_starvla][INFO] - Step 3868, Loss: {'train/loss': 0.7916783094406128, 'train/grad_norm_pre_clip': 14.053960800170898, 'timing/data': 0.005043974961154163, 'timing/model': 1.2063036100007594, 'learning_rate/qwen_vl_interface': 4.564664720238833e-06, 'learning_rate/action_model': 4.564664720238833e-05, 'epoch': 1.14}) +[2026-06-17 13:22:49,148][starVLA.training.train_starvla][INFO] - Step 3869, Loss: {'train/loss': 0.8139790892601013, 'train/grad_norm_pre_clip': 9.274060249328613, 'timing/data': 0.0049586460227146745, 'timing/model': 1.2055097109405324, 'learning_rate/qwen_vl_interface': 4.5625247295461314e-06, 'learning_rate/action_model': 4.562524729546131e-05, 'epoch': 1.14}) +[2026-06-17 13:22:53,896][starVLA.training.train_starvla][INFO] - Step 3870, Loss: {'train/loss': 0.7687296271324158, 'train/grad_norm_pre_clip': 11.575692176818848, 'timing/data': 0.005002489080652595, 'timing/model': 1.2055513879749924, 'learning_rate/qwen_vl_interface': 4.560384881367791e-06, 'learning_rate/action_model': 4.5603848813677904e-05, 'epoch': 1.14}) +[2026-06-17 13:22:58,689][starVLA.training.train_starvla][INFO] - Step 3871, Loss: {'train/loss': 0.40717610716819763, 'train/grad_norm_pre_clip': 8.877324104309082, 'timing/data': 0.004512512939982116, 'timing/model': 1.3293032760266215, 'learning_rate/qwen_vl_interface': 4.5582451761474085e-06, 'learning_rate/action_model': 4.558245176147408e-05, 'epoch': 1.14}) +[2026-06-17 13:23:03,355][starVLA.training.train_starvla][INFO] - Step 3872, Loss: {'train/loss': 0.9816622734069824, 'train/grad_norm_pre_clip': 15.545384407043457, 'timing/data': 0.004549012053757906, 'timing/model': 1.2137335479492322, 'learning_rate/qwen_vl_interface': 4.556105614328545e-06, 'learning_rate/action_model': 4.5561056143285455e-05, 'epoch': 1.14}) +[2026-06-17 13:23:08,189][starVLA.training.train_starvla][INFO] - Step 3873, Loss: {'train/loss': 0.809497594833374, 'train/grad_norm_pre_clip': 11.946321487426758, 'timing/data': 0.005185998044908047, 'timing/model': 1.2099375480320305, 'learning_rate/qwen_vl_interface': 4.553966196354734e-06, 'learning_rate/action_model': 4.553966196354734e-05, 'epoch': 1.14}) +[2026-06-17 13:23:12,978][starVLA.training.train_starvla][INFO] - Step 3874, Loss: {'train/loss': 0.6724405288696289, 'train/grad_norm_pre_clip': 10.97899055480957, 'timing/data': 0.004758963012136519, 'timing/model': 1.2052370290039107, 'learning_rate/qwen_vl_interface': 4.55182692266948e-06, 'learning_rate/action_model': 4.551826922669479e-05, 'epoch': 1.14}) +[2026-06-17 13:23:17,823][starVLA.training.train_starvla][INFO] - Step 3875, Loss: {'train/loss': 0.4508238732814789, 'train/grad_norm_pre_clip': 9.391209602355957, 'timing/data': 0.005067992024123669, 'timing/model': 1.2084495089948177, 'learning_rate/qwen_vl_interface': 4.549687793716257e-06, 'learning_rate/action_model': 4.549687793716257e-05, 'epoch': 1.14}) +[2026-06-17 13:23:22,630][starVLA.training.train_starvla][INFO] - Step 3876, Loss: {'train/loss': 0.6058354377746582, 'train/grad_norm_pre_clip': 11.777554512023926, 'timing/data': 0.0049176180036738515, 'timing/model': 1.3457158509409055, 'learning_rate/qwen_vl_interface': 4.547548809938507e-06, 'learning_rate/action_model': 4.5475488099385075e-05, 'epoch': 1.14}) +[2026-06-17 13:23:27,304][starVLA.training.train_starvla][INFO] - Step 3877, Loss: {'train/loss': 0.6411464214324951, 'train/grad_norm_pre_clip': 12.271310806274414, 'timing/data': 0.005458997911773622, 'timing/model': 1.2121227340539917, 'learning_rate/qwen_vl_interface': 4.545409971779644e-06, 'learning_rate/action_model': 4.545409971779644e-05, 'epoch': 1.14}) +[2026-06-17 13:23:32,093][starVLA.training.train_starvla][INFO] - Step 3878, Loss: {'train/loss': 0.6843258738517761, 'train/grad_norm_pre_clip': 13.02025032043457, 'timing/data': 0.004872552934102714, 'timing/model': 1.2054727070499212, 'learning_rate/qwen_vl_interface': 4.543271279683052e-06, 'learning_rate/action_model': 4.543271279683053e-05, 'epoch': 1.14}) +[2026-06-17 13:23:36,809][starVLA.training.train_starvla][INFO] - Step 3879, Loss: {'train/loss': 0.5451164245605469, 'train/grad_norm_pre_clip': 10.443103790283203, 'timing/data': 0.004542496986687183, 'timing/model': 1.2041332700755447, 'learning_rate/qwen_vl_interface': 4.541132734092086e-06, 'learning_rate/action_model': 4.541132734092086e-05, 'epoch': 1.14}) +[2026-06-17 13:23:41,556][starVLA.training.train_starvla][INFO] - Step 3880, Loss: {'train/loss': 0.8609280586242676, 'train/grad_norm_pre_clip': 12.586773872375488, 'timing/data': 0.004964450024999678, 'timing/model': 1.208752157050185, 'learning_rate/qwen_vl_interface': 4.538994335450066e-06, 'learning_rate/action_model': 4.5389943354500664e-05, 'epoch': 1.14}) +[2026-06-17 13:23:46,364][starVLA.training.train_starvla][INFO] - Step 3881, Loss: {'train/loss': 0.8621941208839417, 'train/grad_norm_pre_clip': 13.255428314208984, 'timing/data': 0.00491311599034816, 'timing/model': 1.2047852650284767, 'learning_rate/qwen_vl_interface': 4.5368560842002836e-06, 'learning_rate/action_model': 4.5368560842002834e-05, 'epoch': 1.14}) +[2026-06-17 13:23:51,152][starVLA.training.train_starvla][INFO] - Step 3882, Loss: {'train/loss': 0.625214695930481, 'train/grad_norm_pre_clip': 11.28847599029541, 'timing/data': 0.0052197640761733055, 'timing/model': 1.3326877619838342, 'learning_rate/qwen_vl_interface': 4.534717980786005e-06, 'learning_rate/action_model': 4.5347179807860046e-05, 'epoch': 1.14}) +[2026-06-17 13:23:55,813][starVLA.training.train_starvla][INFO] - Step 3883, Loss: {'train/loss': 0.6580456495285034, 'train/grad_norm_pre_clip': 11.449175834655762, 'timing/data': 0.005000028060749173, 'timing/model': 1.2061981810256839, 'learning_rate/qwen_vl_interface': 4.532580025650457e-06, 'learning_rate/action_model': 4.5325800256504574e-05, 'epoch': 1.14}) +[2026-06-17 13:24:00,600][starVLA.training.train_starvla][INFO] - Step 3884, Loss: {'train/loss': 0.6843320727348328, 'train/grad_norm_pre_clip': 14.11934757232666, 'timing/data': 0.005122469970956445, 'timing/model': 1.206540023908019, 'learning_rate/qwen_vl_interface': 4.530442219236842e-06, 'learning_rate/action_model': 4.530442219236842e-05, 'epoch': 1.14}) +[2026-06-17 13:24:05,475][starVLA.training.train_starvla][INFO] - Step 3885, Loss: {'train/loss': 0.49937906861305237, 'train/grad_norm_pre_clip': 9.183526039123535, 'timing/data': 0.0045867570443078876, 'timing/model': 1.206835106946528, 'learning_rate/qwen_vl_interface': 4.5283045619883285e-06, 'learning_rate/action_model': 4.5283045619883285e-05, 'epoch': 1.14}) +[2026-06-17 13:24:10,224][starVLA.training.train_starvla][INFO] - Step 3886, Loss: {'train/loss': 1.1073977947235107, 'train/grad_norm_pre_clip': 14.865649223327637, 'timing/data': 0.005246335989795625, 'timing/model': 1.203797303023748, 'learning_rate/qwen_vl_interface': 4.526167054348057e-06, 'learning_rate/action_model': 4.526167054348057e-05, 'epoch': 1.14}) +[2026-06-17 13:24:15,083][starVLA.training.train_starvla][INFO] - Step 3887, Loss: {'train/loss': 0.5213358402252197, 'train/grad_norm_pre_clip': 10.915116310119629, 'timing/data': 0.004445153987035155, 'timing/model': 1.217285317950882, 'learning_rate/qwen_vl_interface': 4.524029696759135e-06, 'learning_rate/action_model': 4.5240296967591354e-05, 'epoch': 1.14}) +[2026-06-17 13:24:19,884][starVLA.training.train_starvla][INFO] - Step 3888, Loss: {'train/loss': 0.6641823649406433, 'train/grad_norm_pre_clip': 13.86853313446045, 'timing/data': 0.004833554965443909, 'timing/model': 1.344952445011586, 'learning_rate/qwen_vl_interface': 4.5218924896646354e-06, 'learning_rate/action_model': 4.521892489664635e-05, 'epoch': 1.14}) +[2026-06-17 13:24:24,547][starVLA.training.train_starvla][INFO] - Step 3889, Loss: {'train/loss': 0.600857138633728, 'train/grad_norm_pre_clip': 14.575506210327148, 'timing/data': 0.005192840006202459, 'timing/model': 1.2097224349854514, 'learning_rate/qwen_vl_interface': 4.519755433507609e-06, 'learning_rate/action_model': 4.5197554335076086e-05, 'epoch': 1.14}) +[2026-06-17 13:24:29,340][starVLA.training.train_starvla][INFO] - Step 3890, Loss: {'train/loss': 0.6803953647613525, 'train/grad_norm_pre_clip': 13.670073509216309, 'timing/data': 0.005454359925352037, 'timing/model': 1.2045481440145522, 'learning_rate/qwen_vl_interface': 4.517618528731066e-06, 'learning_rate/action_model': 4.517618528731065e-05, 'epoch': 1.14}) +[2026-06-17 13:24:34,140][starVLA.training.train_starvla][INFO] - Step 3891, Loss: {'train/loss': 0.7952889204025269, 'train/grad_norm_pre_clip': 13.284542083740234, 'timing/data': 0.005223863059654832, 'timing/model': 1.2093989219283685, 'learning_rate/qwen_vl_interface': 4.515481775777989e-06, 'learning_rate/action_model': 4.515481775777988e-05, 'epoch': 1.14}) +[2026-06-17 13:24:38,926][starVLA.training.train_starvla][INFO] - Step 3892, Loss: {'train/loss': 0.6562602519989014, 'train/grad_norm_pre_clip': 11.89528751373291, 'timing/data': 0.004940479062497616, 'timing/model': 1.2045799300540239, 'learning_rate/qwen_vl_interface': 4.51334517509133e-06, 'learning_rate/action_model': 4.5133451750913305e-05, 'epoch': 1.15}) +[2026-06-17 13:24:43,595][starVLA.training.train_starvla][INFO] - Step 3893, Loss: {'train/loss': 0.5942461490631104, 'train/grad_norm_pre_clip': 10.205222129821777, 'timing/data': 0.005228380090557039, 'timing/model': 1.2041275040246546, 'learning_rate/qwen_vl_interface': 4.511208727114011e-06, 'learning_rate/action_model': 4.51120872711401e-05, 'epoch': 1.15}) +[2026-06-17 13:24:48,311][starVLA.training.train_starvla][INFO] - Step 3894, Loss: {'train/loss': 0.610701322555542, 'train/grad_norm_pre_clip': 11.99461841583252, 'timing/data': 0.004962889943271875, 'timing/model': 1.2492427689721808, 'learning_rate/qwen_vl_interface': 4.5090724322889175e-06, 'learning_rate/action_model': 4.5090724322889174e-05, 'epoch': 1.15}) +[2026-06-17 13:24:52,980][starVLA.training.train_starvla][INFO] - Step 3895, Loss: {'train/loss': 0.7282829284667969, 'train/grad_norm_pre_clip': 11.469255447387695, 'timing/data': 0.004836162901483476, 'timing/model': 1.2095580730820075, 'learning_rate/qwen_vl_interface': 4.506936291058902e-06, 'learning_rate/action_model': 4.506936291058902e-05, 'epoch': 1.15}) +[2026-06-17 13:24:57,717][starVLA.training.train_starvla][INFO] - Step 3896, Loss: {'train/loss': 0.7558205127716064, 'train/grad_norm_pre_clip': 21.336925506591797, 'timing/data': 0.005088101956062019, 'timing/model': 1.2069266280159354, 'learning_rate/qwen_vl_interface': 4.504800303866798e-06, 'learning_rate/action_model': 4.504800303866798e-05, 'epoch': 1.15}) +[2026-06-17 13:25:02,480][starVLA.training.train_starvla][INFO] - Step 3897, Loss: {'train/loss': 0.9291408658027649, 'train/grad_norm_pre_clip': 13.609688758850098, 'timing/data': 0.004796948982402682, 'timing/model': 1.2096148339333013, 'learning_rate/qwen_vl_interface': 4.50266447115539e-06, 'learning_rate/action_model': 4.5026644711553904e-05, 'epoch': 1.15}) +[2026-06-17 13:25:07,270][starVLA.training.train_starvla][INFO] - Step 3898, Loss: {'train/loss': 0.8394930958747864, 'train/grad_norm_pre_clip': 15.523636817932129, 'timing/data': 0.00499797600787133, 'timing/model': 1.206602617050521, 'learning_rate/qwen_vl_interface': 4.500528793367444e-06, 'learning_rate/action_model': 4.5005287933674435e-05, 'epoch': 1.15}) +[2026-06-17 13:25:12,060][starVLA.training.train_starvla][INFO] - Step 3899, Loss: {'train/loss': 0.862706184387207, 'train/grad_norm_pre_clip': 15.157142639160156, 'timing/data': 0.004988825065083802, 'timing/model': 1.3342250169953331, 'learning_rate/qwen_vl_interface': 4.498393270945683e-06, 'learning_rate/action_model': 4.498393270945683e-05, 'epoch': 1.15}) +[2026-06-17 13:26:53,865][starVLA.training.train_starvla][INFO] - Step 3900, Loss: {'train/loss': 0.5538048148155212, 'train/grad_norm_pre_clip': 10.971878051757812, 'eval/loss': 0.7389782667160034, 'eval/latency_0/loss': 0.7389782667160034, 'eval/demon_fire_precision': 0.8967587034813925, 'eval/demon_fire_recall': 0.9306478405315615, 'eval/demon_fire_f1': 0.913389036070919, 'eval/demon_fire_rate_model': 0.6949388209121246, 'eval/demon_fire_rate_teacher': 0.6696329254727474, 'eval/demon_left_precision': 0.8133022170361727, 'eval/demon_left_recall': 0.7898016997167139, 'eval/demon_left_f1': 0.8013797068123024, 'eval/demon_left_rate_model': 0.47664071190211343, 'eval/demon_left_rate_teacher': 0.49082313681868744, 'eval/demon_right_precision': 0.7998929909042268, 'eval/demon_right_recall': 0.844632768361582, 'eval/demon_right_f1': 0.8216543006320418, 'eval/demon_right_rate_model': 0.5197441601779755, 'eval/demon_right_rate_teacher': 0.4922135706340378, 'eval/demon_move_f1': 0.8115170037221722, 'eval/demon_cc_f1': 0.8802393529133277, 'eval/latency_0/demon_fire_precision': 0.8967587034813925, 'eval/latency_0/demon_fire_recall': 0.9306478405315615, 'eval/latency_0/demon_fire_f1': 0.913389036070919, 'eval/latency_0/demon_fire_rate_model': 0.6949388209121246, 'eval/latency_0/demon_fire_rate_teacher': 0.6696329254727474, 'eval/latency_0/demon_left_precision': 0.8133022170361727, 'eval/latency_0/demon_left_recall': 0.7898016997167139, 'eval/latency_0/demon_left_f1': 0.8013797068123024, 'eval/latency_0/demon_left_rate_model': 0.47664071190211343, 'eval/latency_0/demon_left_rate_teacher': 0.49082313681868744, 'eval/latency_0/demon_right_precision': 0.7998929909042268, 'eval/latency_0/demon_right_recall': 0.844632768361582, 'eval/latency_0/demon_right_f1': 0.8216543006320418, 'eval/latency_0/demon_right_rate_model': 0.5197441601779755, 'eval/latency_0/demon_right_rate_teacher': 0.4922135706340378, 'eval/latency_0/demon_move_f1': 0.8115170037221722, 'eval/latency_0/demon_cc_f1': 0.8802393529133277, 'timing/data': 0.005023348960094154, 'timing/model': 1.2011533390032127, 'learning_rate/qwen_vl_interface': 4.496257904332806e-06, 'learning_rate/action_model': 4.4962579043328064e-05, 'epoch': 1.15}) +[2026-06-17 13:26:53,866][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_3900_state +[2026-06-17 13:26:53,869][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 13:26:57,432][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_3900_state/model.safetensors +[2026-06-17 13:27:05,062][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_3900_state/optimizer.bin +[2026-06-17 13:27:05,063][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_3900_state/sampler.bin +[2026-06-17 13:27:05,063][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_3900_state/sampler_1.bin +[2026-06-17 13:27:05,064][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_3900_state/random_states_0.pkl +[2026-06-17 13:27:05,122][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3900 +[2026-06-17 13:27:05,123][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 13:27:05,133][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 13:27:09,846][starVLA.training.train_starvla][INFO] - Step 3901, Loss: {'train/loss': 0.6406645774841309, 'train/grad_norm_pre_clip': 10.947257041931152, 'timing/data': 0.0047190070617944, 'timing/model': 1.3231113979127258, 'learning_rate/qwen_vl_interface': 4.494122693971478e-06, 'learning_rate/action_model': 4.4941226939714785e-05, 'epoch': 1.15}) +[2026-06-17 13:27:14,444][starVLA.training.train_starvla][INFO] - Step 3902, Loss: {'train/loss': 0.6933118104934692, 'train/grad_norm_pre_clip': 13.39118480682373, 'timing/data': 0.0049218430649489164, 'timing/model': 1.191584724932909, 'learning_rate/qwen_vl_interface': 4.491987640304327e-06, 'learning_rate/action_model': 4.4919876403043265e-05, 'epoch': 1.15}) +[2026-06-17 13:27:19,187][starVLA.training.train_starvla][INFO] - Step 3903, Loss: {'train/loss': 0.6571161150932312, 'train/grad_norm_pre_clip': 12.697307586669922, 'timing/data': 0.004928145091980696, 'timing/model': 1.1974781400058419, 'learning_rate/qwen_vl_interface': 4.489852743773955e-06, 'learning_rate/action_model': 4.489852743773955e-05, 'epoch': 1.15}) +[2026-06-17 13:27:23,888][starVLA.training.train_starvla][INFO] - Step 3904, Loss: {'train/loss': 0.6052606105804443, 'train/grad_norm_pre_clip': 11.300365447998047, 'timing/data': 0.0050814528949558735, 'timing/model': 1.193843210930936, 'learning_rate/qwen_vl_interface': 4.487718004822928e-06, 'learning_rate/action_model': 4.487718004822928e-05, 'epoch': 1.15}) +[2026-06-17 13:27:28,632][starVLA.training.train_starvla][INFO] - Step 3905, Loss: {'train/loss': 0.7952263355255127, 'train/grad_norm_pre_clip': 13.785202980041504, 'timing/data': 0.005111211095936596, 'timing/model': 1.1951733879977837, 'learning_rate/qwen_vl_interface': 4.485583423893778e-06, 'learning_rate/action_model': 4.4855834238937774e-05, 'epoch': 1.15}) +[2026-06-17 13:27:33,392][starVLA.training.train_starvla][INFO] - Step 3906, Loss: {'train/loss': 1.1252230405807495, 'train/grad_norm_pre_clip': 13.6914701461792, 'timing/data': 0.004809804959222674, 'timing/model': 1.1968484100652859, 'learning_rate/qwen_vl_interface': 4.483449001429005e-06, 'learning_rate/action_model': 4.483449001429005e-05, 'epoch': 1.15}) +[2026-06-17 13:27:38,151][starVLA.training.train_starvla][INFO] - Step 3907, Loss: {'train/loss': 0.6856881380081177, 'train/grad_norm_pre_clip': 11.771956443786621, 'timing/data': 0.0051961400313302875, 'timing/model': 1.3260850859805942, 'learning_rate/qwen_vl_interface': 4.481314737871081e-06, 'learning_rate/action_model': 4.4813147378710804e-05, 'epoch': 1.15}) +[2026-06-17 13:27:42,911][starVLA.training.train_starvla][INFO] - Step 3908, Loss: {'train/loss': 0.7051730155944824, 'train/grad_norm_pre_clip': 11.9237699508667, 'timing/data': 0.0050914550665766, 'timing/model': 1.3261346770450473, 'learning_rate/qwen_vl_interface': 4.47918063366244e-06, 'learning_rate/action_model': 4.4791806336624395e-05, 'epoch': 1.15}) +[2026-06-17 13:27:47,543][starVLA.training.train_starvla][INFO] - Step 3909, Loss: {'train/loss': 0.6748772859573364, 'train/grad_norm_pre_clip': 11.987914085388184, 'timing/data': 0.005718804895877838, 'timing/model': 1.200531289097853, 'learning_rate/qwen_vl_interface': 4.477046689245481e-06, 'learning_rate/action_model': 4.477046689245481e-05, 'epoch': 1.15}) +[2026-06-17 13:27:52,275][starVLA.training.train_starvla][INFO] - Step 3910, Loss: {'train/loss': 0.5659416317939758, 'train/grad_norm_pre_clip': 11.720414161682129, 'timing/data': 0.005494017968885601, 'timing/model': 1.1984192209783942, 'learning_rate/qwen_vl_interface': 4.474912905062575e-06, 'learning_rate/action_model': 4.474912905062575e-05, 'epoch': 1.15}) +[2026-06-17 13:27:57,038][starVLA.training.train_starvla][INFO] - Step 3911, Loss: {'train/loss': 0.68224036693573, 'train/grad_norm_pre_clip': 14.950055122375488, 'timing/data': 0.004604892921634018, 'timing/model': 1.198818371980451, 'learning_rate/qwen_vl_interface': 4.472779281556059e-06, 'learning_rate/action_model': 4.4727792815560585e-05, 'epoch': 1.15}) +[2026-06-17 13:28:01,806][starVLA.training.train_starvla][INFO] - Step 3912, Loss: {'train/loss': 0.9887630939483643, 'train/grad_norm_pre_clip': 19.185588836669922, 'timing/data': 0.00507462490350008, 'timing/model': 1.201027370058, 'learning_rate/qwen_vl_interface': 4.470645819168233e-06, 'learning_rate/action_model': 4.4706458191682324e-05, 'epoch': 1.15}) +[2026-06-17 13:28:06,526][starVLA.training.train_starvla][INFO] - Step 3913, Loss: {'train/loss': 0.6728289723396301, 'train/grad_norm_pre_clip': 10.916740417480469, 'timing/data': 0.00498825206886977, 'timing/model': 1.2767927929526195, 'learning_rate/qwen_vl_interface': 4.468512518341367e-06, 'learning_rate/action_model': 4.4685125183413674e-05, 'epoch': 1.15}) +[2026-06-17 13:28:11,188][starVLA.training.train_starvla][INFO] - Step 3914, Loss: {'train/loss': 0.65181565284729, 'train/grad_norm_pre_clip': 12.672500610351562, 'timing/data': 0.00470730394590646, 'timing/model': 1.2035462299827486, 'learning_rate/qwen_vl_interface': 4.466379379517699e-06, 'learning_rate/action_model': 4.4663793795176985e-05, 'epoch': 1.15}) +[2026-06-17 13:28:15,957][starVLA.training.train_starvla][INFO] - Step 3915, Loss: {'train/loss': 0.7531691193580627, 'train/grad_norm_pre_clip': 13.247398376464844, 'timing/data': 0.004948911024257541, 'timing/model': 1.2068791730562225, 'learning_rate/qwen_vl_interface': 4.464246403139429e-06, 'learning_rate/action_model': 4.464246403139428e-05, 'epoch': 1.15}) +[2026-06-17 13:28:20,728][starVLA.training.train_starvla][INFO] - Step 3916, Loss: {'train/loss': 0.7954954504966736, 'train/grad_norm_pre_clip': 18.860570907592773, 'timing/data': 0.004914539982564747, 'timing/model': 1.2038520800415426, 'learning_rate/qwen_vl_interface': 4.462113589648723e-06, 'learning_rate/action_model': 4.462113589648723e-05, 'epoch': 1.15}) +[2026-06-17 13:28:25,501][starVLA.training.train_starvla][INFO] - Step 3917, Loss: {'train/loss': 0.5255881547927856, 'train/grad_norm_pre_clip': 8.755669593811035, 'timing/data': 0.005677164997905493, 'timing/model': 1.2010107969399542, 'learning_rate/qwen_vl_interface': 4.459980939487718e-06, 'learning_rate/action_model': 4.4599809394877176e-05, 'epoch': 1.15}) +[2026-06-17 13:28:30,362][starVLA.training.train_starvla][INFO] - Step 3918, Loss: {'train/loss': 0.7381704449653625, 'train/grad_norm_pre_clip': 10.92585277557373, 'timing/data': 0.0050362839829176664, 'timing/model': 1.3276309240609407, 'learning_rate/qwen_vl_interface': 4.457848453098514e-06, 'learning_rate/action_model': 4.4578484530985144e-05, 'epoch': 1.15}) +[2026-06-17 13:28:35,085][starVLA.training.train_starvla][INFO] - Step 3919, Loss: {'train/loss': 0.6301808953285217, 'train/grad_norm_pre_clip': 13.685104370117188, 'timing/data': 0.004943754058331251, 'timing/model': 1.280272479983978, 'learning_rate/qwen_vl_interface': 4.45571613092318e-06, 'learning_rate/action_model': 4.4557161309231794e-05, 'epoch': 1.15}) +[2026-06-17 13:28:39,733][starVLA.training.train_starvla][INFO] - Step 3920, Loss: {'train/loss': 0.6505671739578247, 'train/grad_norm_pre_clip': 15.76892375946045, 'timing/data': 0.004787613986991346, 'timing/model': 1.2011954740155488, 'learning_rate/qwen_vl_interface': 4.453583973403746e-06, 'learning_rate/action_model': 4.453583973403746e-05, 'epoch': 1.15}) +[2026-06-17 13:28:44,514][starVLA.training.train_starvla][INFO] - Step 3921, Loss: {'train/loss': 0.41415297985076904, 'train/grad_norm_pre_clip': 12.512264251708984, 'timing/data': 0.004836648935452104, 'timing/model': 1.202716901898384, 'learning_rate/qwen_vl_interface': 4.451451980982209e-06, 'learning_rate/action_model': 4.4514519809822086e-05, 'epoch': 1.15}) +[2026-06-17 13:28:49,307][starVLA.training.train_starvla][INFO] - Step 3922, Loss: {'train/loss': 0.7995163202285767, 'train/grad_norm_pre_clip': 15.110047340393066, 'timing/data': 0.005080575938336551, 'timing/model': 1.2039131730562076, 'learning_rate/qwen_vl_interface': 4.4493201541005385e-06, 'learning_rate/action_model': 4.4493201541005384e-05, 'epoch': 1.15}) +[2026-06-17 13:28:54,109][starVLA.training.train_starvla][INFO] - Step 3923, Loss: {'train/loss': 0.6249036192893982, 'train/grad_norm_pre_clip': 12.381160736083984, 'timing/data': 0.005182249937206507, 'timing/model': 1.204985156073235, 'learning_rate/qwen_vl_interface': 4.447188493200661e-06, 'learning_rate/action_model': 4.447188493200661e-05, 'epoch': 1.15}) +[2026-06-17 13:28:58,879][starVLA.training.train_starvla][INFO] - Step 3924, Loss: {'train/loss': 0.5815121531486511, 'train/grad_norm_pre_clip': 10.35020637512207, 'timing/data': 0.004900760017335415, 'timing/model': 1.2052928800694644, 'learning_rate/qwen_vl_interface': 4.445056998724472e-06, 'learning_rate/action_model': 4.4450569987244716e-05, 'epoch': 1.15}) +[2026-06-17 13:29:03,643][starVLA.training.train_starvla][INFO] - Step 3925, Loss: {'train/loss': 0.7447556257247925, 'train/grad_norm_pre_clip': 12.16637897491455, 'timing/data': 0.004904353991150856, 'timing/model': 1.3022547470172867, 'learning_rate/qwen_vl_interface': 4.4429256711138335e-06, 'learning_rate/action_model': 4.442925671113833e-05, 'epoch': 1.15}) +[2026-06-17 13:29:08,302][starVLA.training.train_starvla][INFO] - Step 3926, Loss: {'train/loss': 0.48890236020088196, 'train/grad_norm_pre_clip': 10.38717269897461, 'timing/data': 0.004727758001536131, 'timing/model': 1.2027570569189265, 'learning_rate/qwen_vl_interface': 4.440794510810572e-06, 'learning_rate/action_model': 4.440794510810572e-05, 'epoch': 1.16}) +[2026-06-17 13:29:13,104][starVLA.training.train_starvla][INFO] - Step 3927, Loss: {'train/loss': 0.8430225253105164, 'train/grad_norm_pre_clip': 14.136919975280762, 'timing/data': 0.005013365065678954, 'timing/model': 1.2041848939843476, 'learning_rate/qwen_vl_interface': 4.438663518256481e-06, 'learning_rate/action_model': 4.438663518256481e-05, 'epoch': 1.16}) +[2026-06-17 13:29:17,917][starVLA.training.train_starvla][INFO] - Step 3928, Loss: {'train/loss': 0.7805935144424438, 'train/grad_norm_pre_clip': 13.67895793914795, 'timing/data': 0.005022950936108828, 'timing/model': 1.210095837013796, 'learning_rate/qwen_vl_interface': 4.436532693893311e-06, 'learning_rate/action_model': 4.436532693893311e-05, 'epoch': 1.16}) +[2026-06-17 13:29:22,727][starVLA.training.train_starvla][INFO] - Step 3929, Loss: {'train/loss': 0.9572604298591614, 'train/grad_norm_pre_clip': 13.17680549621582, 'timing/data': 0.005077373934909701, 'timing/model': 1.218138722004369, 'learning_rate/qwen_vl_interface': 4.434402038162793e-06, 'learning_rate/action_model': 4.434402038162793e-05, 'epoch': 1.16}) +[2026-06-17 13:29:27,526][starVLA.training.train_starvla][INFO] - Step 3930, Loss: {'train/loss': 0.9015174508094788, 'train/grad_norm_pre_clip': 12.901819229125977, 'timing/data': 0.004919053986668587, 'timing/model': 1.3406618010485545, 'learning_rate/qwen_vl_interface': 4.432271551506609e-06, 'learning_rate/action_model': 4.4322715515066094e-05, 'epoch': 1.16}) +[2026-06-17 13:29:32,191][starVLA.training.train_starvla][INFO] - Step 3931, Loss: {'train/loss': 0.8564338684082031, 'train/grad_norm_pre_clip': 14.7678804397583, 'timing/data': 0.0047341230092570186, 'timing/model': 1.2105397040722892, 'learning_rate/qwen_vl_interface': 4.4301412343664125e-06, 'learning_rate/action_model': 4.4301412343664125e-05, 'epoch': 1.16}) +[2026-06-17 13:29:36,988][starVLA.training.train_starvla][INFO] - Step 3932, Loss: {'train/loss': 0.6745548248291016, 'train/grad_norm_pre_clip': 12.834300994873047, 'timing/data': 0.005922774085775018, 'timing/model': 1.205750138964504, 'learning_rate/qwen_vl_interface': 4.428011087183819e-06, 'learning_rate/action_model': 4.428011087183819e-05, 'epoch': 1.16}) +[2026-06-17 13:29:41,777][starVLA.training.train_starvla][INFO] - Step 3933, Loss: {'train/loss': 0.642052173614502, 'train/grad_norm_pre_clip': 13.155465126037598, 'timing/data': 0.004813423031009734, 'timing/model': 1.2051431289874017, 'learning_rate/qwen_vl_interface': 4.4258811104004125e-06, 'learning_rate/action_model': 4.425881110400413e-05, 'epoch': 1.16}) +[2026-06-17 13:29:46,484][starVLA.training.train_starvla][INFO] - Step 3934, Loss: {'train/loss': 0.6106821298599243, 'train/grad_norm_pre_clip': 10.537810325622559, 'timing/data': 0.0049888629000633955, 'timing/model': 1.2068496519932523, 'learning_rate/qwen_vl_interface': 4.4237513044577394e-06, 'learning_rate/action_model': 4.4237513044577386e-05, 'epoch': 1.16}) +[2026-06-17 13:29:51,236][starVLA.training.train_starvla][INFO] - Step 3935, Loss: {'train/loss': 0.6664108037948608, 'train/grad_norm_pre_clip': 11.922245025634766, 'timing/data': 0.004833256010897458, 'timing/model': 1.2075075479224324, 'learning_rate/qwen_vl_interface': 4.421621669797309e-06, 'learning_rate/action_model': 4.421621669797308e-05, 'epoch': 1.16}) +[2026-06-17 13:29:56,039][starVLA.training.train_starvla][INFO] - Step 3936, Loss: {'train/loss': 0.5024689435958862, 'train/grad_norm_pre_clip': 10.748016357421875, 'timing/data': 0.004813881008885801, 'timing/model': 1.3403258809121326, 'learning_rate/qwen_vl_interface': 4.419492206860599e-06, 'learning_rate/action_model': 4.419492206860598e-05, 'epoch': 1.16}) +[2026-06-17 13:30:00,701][starVLA.training.train_starvla][INFO] - Step 3937, Loss: {'train/loss': 0.9038063883781433, 'train/grad_norm_pre_clip': 12.277763366699219, 'timing/data': 0.004947825917042792, 'timing/model': 1.2062087929807603, 'learning_rate/qwen_vl_interface': 4.417362916089048e-06, 'learning_rate/action_model': 4.417362916089047e-05, 'epoch': 1.16}) +[2026-06-17 13:30:05,501][starVLA.training.train_starvla][INFO] - Step 3938, Loss: {'train/loss': 0.5689545273780823, 'train/grad_norm_pre_clip': 10.43124008178711, 'timing/data': 0.005020115990191698, 'timing/model': 1.205591936944984, 'learning_rate/qwen_vl_interface': 4.415233797924059e-06, 'learning_rate/action_model': 4.415233797924059e-05, 'epoch': 1.16}) +[2026-06-17 13:30:10,292][starVLA.training.train_starvla][INFO] - Step 3939, Loss: {'train/loss': 0.5028900504112244, 'train/grad_norm_pre_clip': 10.09982967376709, 'timing/data': 0.005067607969976962, 'timing/model': 1.2044415069976822, 'learning_rate/qwen_vl_interface': 4.413104852807002e-06, 'learning_rate/action_model': 4.413104852807001e-05, 'epoch': 1.16}) +[2026-06-17 13:30:15,079][starVLA.training.train_starvla][INFO] - Step 3940, Loss: {'train/loss': 0.5113058686256409, 'train/grad_norm_pre_clip': 11.754803657531738, 'timing/data': 0.005168467992916703, 'timing/model': 1.206970865954645, 'learning_rate/qwen_vl_interface': 4.41097608117921e-06, 'learning_rate/action_model': 4.41097608117921e-05, 'epoch': 1.16}) +[2026-06-17 13:30:19,826][starVLA.training.train_starvla][INFO] - Step 3941, Loss: {'train/loss': 0.5215348601341248, 'train/grad_norm_pre_clip': 9.833666801452637, 'timing/data': 0.004870877950452268, 'timing/model': 1.205322201945819, 'learning_rate/qwen_vl_interface': 4.40884748348198e-06, 'learning_rate/action_model': 4.40884748348198e-05, 'epoch': 1.16}) +[2026-06-17 13:30:24,605][starVLA.training.train_starvla][INFO] - Step 3942, Loss: {'train/loss': 0.6920257210731506, 'train/grad_norm_pre_clip': 12.115431785583496, 'timing/data': 0.005244723986834288, 'timing/model': 1.3238988649100065, 'learning_rate/qwen_vl_interface': 4.406719060156572e-06, 'learning_rate/action_model': 4.4067190601565716e-05, 'epoch': 1.16}) +[2026-06-17 13:30:29,282][starVLA.training.train_starvla][INFO] - Step 3943, Loss: {'train/loss': 0.8351510763168335, 'train/grad_norm_pre_clip': 14.546112060546875, 'timing/data': 0.004928862093947828, 'timing/model': 1.2062072579283267, 'learning_rate/qwen_vl_interface': 4.404590811644209e-06, 'learning_rate/action_model': 4.4045908116442087e-05, 'epoch': 1.16}) +[2026-06-17 13:30:34,107][starVLA.training.train_starvla][INFO] - Step 3944, Loss: {'train/loss': 0.5715672373771667, 'train/grad_norm_pre_clip': 10.880998611450195, 'timing/data': 0.00950521300546825, 'timing/model': 1.2104775729821995, 'learning_rate/qwen_vl_interface': 4.40246273838608e-06, 'learning_rate/action_model': 4.40246273838608e-05, 'epoch': 1.16}) +[2026-06-17 13:30:38,909][starVLA.training.train_starvla][INFO] - Step 3945, Loss: {'train/loss': 0.7087441682815552, 'train/grad_norm_pre_clip': 13.71981430053711, 'timing/data': 0.004924456938169897, 'timing/model': 1.2091101920232177, 'learning_rate/qwen_vl_interface': 4.400334840823339e-06, 'learning_rate/action_model': 4.400334840823339e-05, 'epoch': 1.16}) +[2026-06-17 13:30:43,700][starVLA.training.train_starvla][INFO] - Step 3946, Loss: {'train/loss': 1.111130952835083, 'train/grad_norm_pre_clip': 19.03133773803711, 'timing/data': 0.0050811279797926545, 'timing/model': 1.2068833699449897, 'learning_rate/qwen_vl_interface': 4.398207119397098e-06, 'learning_rate/action_model': 4.398207119397098e-05, 'epoch': 1.16}) +[2026-06-17 13:30:48,485][starVLA.training.train_starvla][INFO] - Step 3947, Loss: {'train/loss': 0.7988153696060181, 'train/grad_norm_pre_clip': 12.518908500671387, 'timing/data': 0.00480583100579679, 'timing/model': 1.3334893649443984, 'learning_rate/qwen_vl_interface': 4.396079574548439e-06, 'learning_rate/action_model': 4.396079574548439e-05, 'epoch': 1.16}) +[2026-06-17 13:30:53,152][starVLA.training.train_starvla][INFO] - Step 3948, Loss: {'train/loss': 0.8036342859268188, 'train/grad_norm_pre_clip': 16.171730041503906, 'timing/data': 0.00527178100310266, 'timing/model': 1.204300865996629, 'learning_rate/qwen_vl_interface': 4.393952206718403e-06, 'learning_rate/action_model': 4.3939522067184026e-05, 'epoch': 1.16}) +[2026-06-17 13:30:57,936][starVLA.training.train_starvla][INFO] - Step 3949, Loss: {'train/loss': 0.6119730472564697, 'train/grad_norm_pre_clip': 12.24284839630127, 'timing/data': 0.004984194994904101, 'timing/model': 1.2063300899462774, 'learning_rate/qwen_vl_interface': 4.391825016347996e-06, 'learning_rate/action_model': 4.391825016347996e-05, 'epoch': 1.16}) +[2026-06-17 13:31:02,649][starVLA.training.train_starvla][INFO] - Step 3950, Loss: {'train/loss': 0.5955682396888733, 'train/grad_norm_pre_clip': 10.997885704040527, 'timing/data': 0.005141376983374357, 'timing/model': 1.2062950240215287, 'learning_rate/qwen_vl_interface': 4.3896980038781844e-06, 'learning_rate/action_model': 4.3896980038781846e-05, 'epoch': 1.16}) +[2026-06-17 13:31:07,432][starVLA.training.train_starvla][INFO] - Step 3951, Loss: {'train/loss': 0.7094547748565674, 'train/grad_norm_pre_clip': 12.071477890014648, 'timing/data': 0.004948249086737633, 'timing/model': 1.2082798649789765, 'learning_rate/qwen_vl_interface': 4.3875711697499054e-06, 'learning_rate/action_model': 4.387571169749905e-05, 'epoch': 1.16}) +[2026-06-17 13:31:12,218][starVLA.training.train_starvla][INFO] - Step 3952, Loss: {'train/loss': 0.7788999080657959, 'train/grad_norm_pre_clip': 15.935134887695312, 'timing/data': 0.005239355028606951, 'timing/model': 1.2046652439748868, 'learning_rate/qwen_vl_interface': 4.38544451440405e-06, 'learning_rate/action_model': 4.3854445144040495e-05, 'epoch': 1.16}) +[2026-06-17 13:31:17,004][starVLA.training.train_starvla][INFO] - Step 3953, Loss: {'train/loss': 0.6918387413024902, 'train/grad_norm_pre_clip': 14.889228820800781, 'timing/data': 0.005444235983304679, 'timing/model': 1.330221607000567, 'learning_rate/qwen_vl_interface': 4.383318038281475e-06, 'learning_rate/action_model': 4.383318038281475e-05, 'epoch': 1.16}) +[2026-06-17 13:31:21,660][starVLA.training.train_starvla][INFO] - Step 3954, Loss: {'train/loss': 0.863298773765564, 'train/grad_norm_pre_clip': 24.18396759033203, 'timing/data': 0.005072383908554912, 'timing/model': 1.2033124479930848, 'learning_rate/qwen_vl_interface': 4.381191741823005e-06, 'learning_rate/action_model': 4.3811917418230046e-05, 'epoch': 1.16}) +[2026-06-17 13:31:26,454][starVLA.training.train_starvla][INFO] - Step 3955, Loss: {'train/loss': 0.5094484090805054, 'train/grad_norm_pre_clip': 11.078837394714355, 'timing/data': 0.004749043029733002, 'timing/model': 1.203693607938476, 'learning_rate/qwen_vl_interface': 4.379065625469422e-06, 'learning_rate/action_model': 4.3790656254694215e-05, 'epoch': 1.16}) +[2026-06-17 13:31:31,225][starVLA.training.train_starvla][INFO] - Step 3956, Loss: {'train/loss': 0.7591471672058105, 'train/grad_norm_pre_clip': 12.57837200164795, 'timing/data': 0.004943331936374307, 'timing/model': 1.2035484119551256, 'learning_rate/qwen_vl_interface': 4.3769396896614724e-06, 'learning_rate/action_model': 4.376939689661472e-05, 'epoch': 1.16}) +[2026-06-17 13:31:35,962][starVLA.training.train_starvla][INFO] - Step 3957, Loss: {'train/loss': 0.487448126077652, 'train/grad_norm_pre_clip': 14.811891555786133, 'timing/data': 0.004666798049584031, 'timing/model': 1.2043813440250233, 'learning_rate/qwen_vl_interface': 4.374813934839862e-06, 'learning_rate/action_model': 4.374813934839862e-05, 'epoch': 1.16}) +[2026-06-17 13:31:40,746][starVLA.training.train_starvla][INFO] - Step 3958, Loss: {'train/loss': 0.49540024995803833, 'train/grad_norm_pre_clip': 9.903515815734863, 'timing/data': 0.005158197018317878, 'timing/model': 1.210201718029566, 'learning_rate/qwen_vl_interface': 4.372688361445267e-06, 'learning_rate/action_model': 4.3726883614452665e-05, 'epoch': 1.16}) +[2026-06-17 13:31:45,541][starVLA.training.train_starvla][INFO] - Step 3959, Loss: {'train/loss': 0.5824555158615112, 'train/grad_norm_pre_clip': 9.998899459838867, 'timing/data': 0.005063534947112203, 'timing/model': 1.342035380890593, 'learning_rate/qwen_vl_interface': 4.370562969918319e-06, 'learning_rate/action_model': 4.370562969918319e-05, 'epoch': 1.16}) +[2026-06-17 13:31:50,210][starVLA.training.train_starvla][INFO] - Step 3960, Loss: {'train/loss': 0.9115179181098938, 'train/grad_norm_pre_clip': 19.782939910888672, 'timing/data': 0.004937075078487396, 'timing/model': 1.2163361699786037, 'learning_rate/qwen_vl_interface': 4.368437760699613e-06, 'learning_rate/action_model': 4.3684377606996124e-05, 'epoch': 1.17}) +[2026-06-17 13:31:55,009][starVLA.training.train_starvla][INFO] - Step 3961, Loss: {'train/loss': 0.9947336316108704, 'train/grad_norm_pre_clip': 16.820642471313477, 'timing/data': 0.005203790031373501, 'timing/model': 1.2052388560259715, 'learning_rate/qwen_vl_interface': 4.366312734229706e-06, 'learning_rate/action_model': 4.366312734229706e-05, 'epoch': 1.17}) +[2026-06-17 13:31:59,802][starVLA.training.train_starvla][INFO] - Step 3962, Loss: {'train/loss': 0.9277782440185547, 'train/grad_norm_pre_clip': 16.432493209838867, 'timing/data': 0.005069207982160151, 'timing/model': 1.2069069109857082, 'learning_rate/qwen_vl_interface': 4.364187890949122e-06, 'learning_rate/action_model': 4.364187890949122e-05, 'epoch': 1.17}) +[2026-06-17 13:32:04,667][starVLA.training.train_starvla][INFO] - Step 3963, Loss: {'train/loss': 0.5792587995529175, 'train/grad_norm_pre_clip': 9.404977798461914, 'timing/data': 0.00492786499671638, 'timing/model': 1.206400783965364, 'learning_rate/qwen_vl_interface': 4.362063231298342e-06, 'learning_rate/action_model': 4.3620632312983425e-05, 'epoch': 1.17}) +[2026-06-17 13:32:09,345][starVLA.training.train_starvla][INFO] - Step 3964, Loss: {'train/loss': 0.8463191986083984, 'train/grad_norm_pre_clip': 14.457015991210938, 'timing/data': 0.004895721096545458, 'timing/model': 1.205604881979525, 'learning_rate/qwen_vl_interface': 4.359938755717807e-06, 'learning_rate/action_model': 4.359938755717807e-05, 'epoch': 1.17}) +[2026-06-17 13:32:14,054][starVLA.training.train_starvla][INFO] - Step 3965, Loss: {'train/loss': 0.4815722703933716, 'train/grad_norm_pre_clip': 10.361193656921387, 'timing/data': 0.004851803998462856, 'timing/model': 1.2521264939568937, 'learning_rate/qwen_vl_interface': 4.357814464647929e-06, 'learning_rate/action_model': 4.357814464647929e-05, 'epoch': 1.17}) +[2026-06-17 13:32:18,709][starVLA.training.train_starvla][INFO] - Step 3966, Loss: {'train/loss': 0.7741807103157043, 'train/grad_norm_pre_clip': 14.037890434265137, 'timing/data': 0.005128521006554365, 'timing/model': 1.2050688659073785, 'learning_rate/qwen_vl_interface': 4.355690358529071e-06, 'learning_rate/action_model': 4.3556903585290705e-05, 'epoch': 1.17}) +[2026-06-17 13:32:23,449][starVLA.training.train_starvla][INFO] - Step 3967, Loss: {'train/loss': 0.6793234348297119, 'train/grad_norm_pre_clip': 33.124000549316406, 'timing/data': 0.00594179704785347, 'timing/model': 1.2097846110118553, 'learning_rate/qwen_vl_interface': 4.353566437801564e-06, 'learning_rate/action_model': 4.353566437801563e-05, 'epoch': 1.17}) +[2026-06-17 13:32:28,234][starVLA.training.train_starvla][INFO] - Step 3968, Loss: {'train/loss': 0.6310909390449524, 'train/grad_norm_pre_clip': 11.605549812316895, 'timing/data': 0.004833620972931385, 'timing/model': 1.2046803120756522, 'learning_rate/qwen_vl_interface': 4.351442702905698e-06, 'learning_rate/action_model': 4.3514427029056976e-05, 'epoch': 1.17}) +[2026-06-17 13:32:33,017][starVLA.training.train_starvla][INFO] - Step 3969, Loss: {'train/loss': 0.5121247172355652, 'train/grad_norm_pre_clip': 9.733691215515137, 'timing/data': 0.005006809951737523, 'timing/model': 1.2054684760514647, 'learning_rate/qwen_vl_interface': 4.349319154281727e-06, 'learning_rate/action_model': 4.349319154281728e-05, 'epoch': 1.17}) +[2026-06-17 13:32:37,813][starVLA.training.train_starvla][INFO] - Step 3970, Loss: {'train/loss': 0.7851677536964417, 'train/grad_norm_pre_clip': 17.454345703125, 'timing/data': 0.004999950993806124, 'timing/model': 1.3345110439695418, 'learning_rate/qwen_vl_interface': 4.347195792369867e-06, 'learning_rate/action_model': 4.347195792369867e-05, 'epoch': 1.17}) +[2026-06-17 13:32:42,516][starVLA.training.train_starvla][INFO] - Step 3971, Loss: {'train/loss': 0.7607769966125488, 'train/grad_norm_pre_clip': 79.19258117675781, 'timing/data': 0.005102501017972827, 'timing/model': 1.2528909330721945, 'learning_rate/qwen_vl_interface': 4.345072617610288e-06, 'learning_rate/action_model': 4.345072617610288e-05, 'epoch': 1.17}) +[2026-06-17 13:32:47,177][starVLA.training.train_starvla][INFO] - Step 3972, Loss: {'train/loss': 0.7227309942245483, 'train/grad_norm_pre_clip': 12.684342384338379, 'timing/data': 0.004983274033293128, 'timing/model': 1.20679145201575, 'learning_rate/qwen_vl_interface': 4.342949630443129e-06, 'learning_rate/action_model': 4.342949630443129e-05, 'epoch': 1.17}) +[2026-06-17 13:32:51,915][starVLA.training.train_starvla][INFO] - Step 3973, Loss: {'train/loss': 0.791035532951355, 'train/grad_norm_pre_clip': 10.95863151550293, 'timing/data': 0.005154142971150577, 'timing/model': 1.2051324690692127, 'learning_rate/qwen_vl_interface': 4.340826831308488e-06, 'learning_rate/action_model': 4.340826831308488e-05, 'epoch': 1.17}) +[2026-06-17 13:32:56,705][starVLA.training.train_starvla][INFO] - Step 3974, Loss: {'train/loss': 0.6490715146064758, 'train/grad_norm_pre_clip': 14.479151725769043, 'timing/data': 0.005037554074078798, 'timing/model': 1.204561157966964, 'learning_rate/qwen_vl_interface': 4.338704220646424e-06, 'learning_rate/action_model': 4.338704220646424e-05, 'epoch': 1.17}) +[2026-06-17 13:33:01,497][starVLA.training.train_starvla][INFO] - Step 3975, Loss: {'train/loss': 0.7574732899665833, 'train/grad_norm_pre_clip': 12.676715850830078, 'timing/data': 0.004973403993062675, 'timing/model': 1.20521495398134, 'learning_rate/qwen_vl_interface': 4.336581798896955e-06, 'learning_rate/action_model': 4.336581798896955e-05, 'epoch': 1.17}) +[2026-06-17 13:33:06,285][starVLA.training.train_starvla][INFO] - Step 3976, Loss: {'train/loss': 0.8651748299598694, 'train/grad_norm_pre_clip': 13.188233375549316, 'timing/data': 0.00498950993642211, 'timing/model': 1.3308106230106205, 'learning_rate/qwen_vl_interface': 4.334459566500062e-06, 'learning_rate/action_model': 4.334459566500062e-05, 'epoch': 1.17}) +[2026-06-17 13:33:11,085][starVLA.training.train_starvla][INFO] - Step 3977, Loss: {'train/loss': 0.7593027353286743, 'train/grad_norm_pre_clip': 13.498322486877441, 'timing/data': 0.004982869955711067, 'timing/model': 1.3409093039808795, 'learning_rate/qwen_vl_interface': 4.332337523895688e-06, 'learning_rate/action_model': 4.3323375238956876e-05, 'epoch': 1.17}) +[2026-06-17 13:33:15,749][starVLA.training.train_starvla][INFO] - Step 3978, Loss: {'train/loss': 0.47124558687210083, 'train/grad_norm_pre_clip': 8.53635025024414, 'timing/data': 0.004670294933021069, 'timing/model': 1.2088954680366442, 'learning_rate/qwen_vl_interface': 4.330215671523731e-06, 'learning_rate/action_model': 4.330215671523731e-05, 'epoch': 1.17}) +[2026-06-17 13:33:20,511][starVLA.training.train_starvla][INFO] - Step 3979, Loss: {'train/loss': 0.8978166580200195, 'train/grad_norm_pre_clip': 15.282793998718262, 'timing/data': 0.005133531987667084, 'timing/model': 1.203995586023666, 'learning_rate/qwen_vl_interface': 4.328094009824054e-06, 'learning_rate/action_model': 4.328094009824054e-05, 'epoch': 1.17}) +[2026-06-17 13:33:25,317][starVLA.training.train_starvla][INFO] - Step 3980, Loss: {'train/loss': 0.9956551194190979, 'train/grad_norm_pre_clip': 19.01076889038086, 'timing/data': 0.005142578040249646, 'timing/model': 1.205226021935232, 'learning_rate/qwen_vl_interface': 4.3259725392364825e-06, 'learning_rate/action_model': 4.325972539236482e-05, 'epoch': 1.17}) +[2026-06-17 13:33:30,113][starVLA.training.train_starvla][INFO] - Step 3981, Loss: {'train/loss': 0.49994146823883057, 'train/grad_norm_pre_clip': 12.083088874816895, 'timing/data': 0.005687023978680372, 'timing/model': 1.2095628399401903, 'learning_rate/qwen_vl_interface': 4.323851260200797e-06, 'learning_rate/action_model': 4.323851260200796e-05, 'epoch': 1.17}) +[2026-06-17 13:33:34,857][starVLA.training.train_starvla][INFO] - Step 3982, Loss: {'train/loss': 0.628119170665741, 'train/grad_norm_pre_clip': 10.72443675994873, 'timing/data': 0.004954454023391008, 'timing/model': 1.2793333670124412, 'learning_rate/qwen_vl_interface': 4.321730173156743e-06, 'learning_rate/action_model': 4.321730173156743e-05, 'epoch': 1.17}) +[2026-06-17 13:33:39,515][starVLA.training.train_starvla][INFO] - Step 3983, Loss: {'train/loss': 0.7436084747314453, 'train/grad_norm_pre_clip': 14.959965705871582, 'timing/data': 0.0049887270433828235, 'timing/model': 1.203456198098138, 'learning_rate/qwen_vl_interface': 4.31960927854402e-06, 'learning_rate/action_model': 4.31960927854402e-05, 'epoch': 1.17}) +[2026-06-17 13:33:44,284][starVLA.training.train_starvla][INFO] - Step 3984, Loss: {'train/loss': 0.6416503190994263, 'train/grad_norm_pre_clip': 10.831668853759766, 'timing/data': 0.005236064083874226, 'timing/model': 1.2055951319634914, 'learning_rate/qwen_vl_interface': 4.317488576802298e-06, 'learning_rate/action_model': 4.3174885768022986e-05, 'epoch': 1.17}) +[2026-06-17 13:33:49,082][starVLA.training.train_starvla][INFO] - Step 3985, Loss: {'train/loss': 0.5992621183395386, 'train/grad_norm_pre_clip': 11.8572998046875, 'timing/data': 0.005052305990830064, 'timing/model': 1.2067307030083612, 'learning_rate/qwen_vl_interface': 4.315368068371197e-06, 'learning_rate/action_model': 4.315368068371197e-05, 'epoch': 1.17}) +[2026-06-17 13:33:53,877][starVLA.training.train_starvla][INFO] - Step 3986, Loss: {'train/loss': 0.7038431167602539, 'train/grad_norm_pre_clip': 16.455245971679688, 'timing/data': 0.005045262980274856, 'timing/model': 1.2063068089773878, 'learning_rate/qwen_vl_interface': 4.313247753690298e-06, 'learning_rate/action_model': 4.3132477536902985e-05, 'epoch': 1.17}) +[2026-06-17 13:33:58,662][starVLA.training.train_starvla][INFO] - Step 3987, Loss: {'train/loss': 0.575344979763031, 'train/grad_norm_pre_clip': 12.224115371704102, 'timing/data': 0.004945508087985218, 'timing/model': 1.3311594530241564, 'learning_rate/qwen_vl_interface': 4.311127633199151e-06, 'learning_rate/action_model': 4.3111276331991504e-05, 'epoch': 1.17}) +[2026-06-17 13:34:03,404][starVLA.training.train_starvla][INFO] - Step 3988, Loss: {'train/loss': 0.6274529695510864, 'train/grad_norm_pre_clip': 14.238378524780273, 'timing/data': 0.005030539003200829, 'timing/model': 1.2873930159257725, 'learning_rate/qwen_vl_interface': 4.309007707337255e-06, 'learning_rate/action_model': 4.309007707337255e-05, 'epoch': 1.17}) +[2026-06-17 13:34:08,081][starVLA.training.train_starvla][INFO] - Step 3989, Loss: {'train/loss': 0.7818328738212585, 'train/grad_norm_pre_clip': 10.716544151306152, 'timing/data': 0.005101392976939678, 'timing/model': 1.2182705838931724, 'learning_rate/qwen_vl_interface': 4.3068879765440745e-06, 'learning_rate/action_model': 4.306887976544074e-05, 'epoch': 1.17}) +[2026-06-17 13:34:12,872][starVLA.training.train_starvla][INFO] - Step 3990, Loss: {'train/loss': 0.5562816262245178, 'train/grad_norm_pre_clip': 9.581368446350098, 'timing/data': 0.00503682903945446, 'timing/model': 1.2033769579138607, 'learning_rate/qwen_vl_interface': 4.304768441259029e-06, 'learning_rate/action_model': 4.3047684412590285e-05, 'epoch': 1.17}) +[2026-06-17 13:34:17,670][starVLA.training.train_starvla][INFO] - Step 3991, Loss: {'train/loss': 0.5971164703369141, 'train/grad_norm_pre_clip': 14.30109977722168, 'timing/data': 0.0047071019653230906, 'timing/model': 1.2093780220020562, 'learning_rate/qwen_vl_interface': 4.302649101921505e-06, 'learning_rate/action_model': 4.302649101921505e-05, 'epoch': 1.17}) +[2026-06-17 13:34:22,464][starVLA.training.train_starvla][INFO] - Step 3992, Loss: {'train/loss': 0.6585831046104431, 'train/grad_norm_pre_clip': 10.302395820617676, 'timing/data': 0.004879994899965823, 'timing/model': 1.2074970249086618, 'learning_rate/qwen_vl_interface': 4.300529958970841e-06, 'learning_rate/action_model': 4.300529958970841e-05, 'epoch': 1.17}) +[2026-06-17 13:34:27,248][starVLA.training.train_starvla][INFO] - Step 3993, Loss: {'train/loss': 0.4509425461292267, 'train/grad_norm_pre_clip': 10.694847106933594, 'timing/data': 0.004713881062343717, 'timing/model': 1.212811317993328, 'learning_rate/qwen_vl_interface': 4.298411012846338e-06, 'learning_rate/action_model': 4.2984110128463376e-05, 'epoch': 1.17}) +[2026-06-17 13:34:32,000][starVLA.training.train_starvla][INFO] - Step 3994, Loss: {'train/loss': 0.7213141322135925, 'train/grad_norm_pre_clip': 13.084105491638184, 'timing/data': 0.005092615960165858, 'timing/model': 1.2987278969958425, 'learning_rate/qwen_vl_interface': 4.296292263987255e-06, 'learning_rate/action_model': 4.2962922639872556e-05, 'epoch': 1.18}) +[2026-06-17 13:34:36,667][starVLA.training.train_starvla][INFO] - Step 3995, Loss: {'train/loss': 0.8173394203186035, 'train/grad_norm_pre_clip': 16.719661712646484, 'timing/data': 0.004646937013603747, 'timing/model': 1.212582346983254, 'learning_rate/qwen_vl_interface': 4.294173712832813e-06, 'learning_rate/action_model': 4.294173712832813e-05, 'epoch': 1.18}) +[2026-06-17 13:34:41,461][starVLA.training.train_starvla][INFO] - Step 3996, Loss: {'train/loss': 1.0616847276687622, 'train/grad_norm_pre_clip': 12.207406997680664, 'timing/data': 0.005453372956253588, 'timing/model': 1.2069603230338544, 'learning_rate/qwen_vl_interface': 4.292055359822191e-06, 'learning_rate/action_model': 4.29205535982219e-05, 'epoch': 1.18}) +[2026-06-17 13:34:46,263][starVLA.training.train_starvla][INFO] - Step 3997, Loss: {'train/loss': 0.9290094375610352, 'train/grad_norm_pre_clip': 13.090332984924316, 'timing/data': 0.00485359993763268, 'timing/model': 1.2060451080324128, 'learning_rate/qwen_vl_interface': 4.289937205394519e-06, 'learning_rate/action_model': 4.2899372053945185e-05, 'epoch': 1.18}) +[2026-06-17 13:34:51,059][starVLA.training.train_starvla][INFO] - Step 3998, Loss: {'train/loss': 0.6366128325462341, 'train/grad_norm_pre_clip': 12.058579444885254, 'timing/data': 0.004936069017276168, 'timing/model': 1.2056293439818546, 'learning_rate/qwen_vl_interface': 4.287819249988901e-06, 'learning_rate/action_model': 4.2878192499889007e-05, 'epoch': 1.18}) +[2026-06-17 13:34:55,897][starVLA.training.train_starvla][INFO] - Step 3999, Loss: {'train/loss': 0.6502834558486938, 'train/grad_norm_pre_clip': 11.799628257751465, 'timing/data': 0.004878599080257118, 'timing/model': 1.3369555090321228, 'learning_rate/qwen_vl_interface': 4.285701494044387e-06, 'learning_rate/action_model': 4.285701494044387e-05, 'epoch': 1.18}) +[2026-06-17 13:45:04,032][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 4000 +[2026-06-17 13:45:04,033][starVLA.training.train_starvla][INFO] - Step 4000, Loss: {'train/loss': 0.5682976245880127, 'train/grad_norm_pre_clip': 9.226941108703613, 'eval/loss': 0.7367339134216309, 'eval/latency_0/loss': 0.7367339134216309, 'eval/demon_fire_precision': 0.920675105485232, 'eval/demon_fire_recall': 0.9061461794019934, 'eval/demon_fire_f1': 0.9133528673084973, 'eval/demon_fire_rate_model': 0.6590656284760845, 'eval/demon_fire_rate_teacher': 0.6696329254727474, 'eval/demon_left_precision': 0.820067264573991, 'eval/demon_left_recall': 0.8288951841359773, 'eval/demon_left_f1': 0.8244575936883628, 'eval/demon_left_rate_model': 0.4961067853170189, 'eval/demon_left_rate_teacher': 0.49082313681868744, 'eval/demon_right_precision': 0.8291995490417137, 'eval/demon_right_recall': 0.8310734463276837, 'eval/demon_right_f1': 0.8301354401805869, 'eval/demon_right_rate_model': 0.4933259176863181, 'eval/demon_right_rate_teacher': 0.4922135706340378, 'eval/demon_move_f1': 0.8272965169344748, 'eval/demon_cc_f1': 0.8857129970404907, 'eval/latency_0/demon_fire_precision': 0.920675105485232, 'eval/latency_0/demon_fire_recall': 0.9061461794019934, 'eval/latency_0/demon_fire_f1': 0.9133528673084973, 'eval/latency_0/demon_fire_rate_model': 0.6590656284760845, 'eval/latency_0/demon_fire_rate_teacher': 0.6696329254727474, 'eval/latency_0/demon_left_precision': 0.820067264573991, 'eval/latency_0/demon_left_recall': 0.8288951841359773, 'eval/latency_0/demon_left_f1': 0.8244575936883628, 'eval/latency_0/demon_left_rate_model': 0.4961067853170189, 'eval/latency_0/demon_left_rate_teacher': 0.49082313681868744, 'eval/latency_0/demon_right_precision': 0.8291995490417137, 'eval/latency_0/demon_right_recall': 0.8310734463276837, 'eval/latency_0/demon_right_f1': 0.8301354401805869, 'eval/latency_0/demon_right_rate_model': 0.4933259176863181, 'eval/latency_0/demon_right_rate_teacher': 0.4922135706340378, 'eval/latency_0/demon_move_f1': 0.8272965169344748, 'eval/latency_0/demon_cc_f1': 0.8857129970404907, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 5964.0, 'rl_games_eval/mid_train/mean_length': 3600.0, 'rl_games_eval/mid_train/std_reward': 258.3486016993318, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/macro_mean_reward': 5964.0, 'rl_games_eval/mid_train/macro_mean_length': 3600.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/demon_attack__latency_0/mean_reward': 5964.0, 'rl_games_eval/mid_train/demon_attack__latency_0/mean_length': 3600.0, 'rl_games_eval/mid_train/demon_attack__latency_0/std_reward': 258.3486016993318, 'rl_games_eval/mid_train/demon_attack__latency_0/std_length': 0.0, 'rl_games_eval/mid_train/demon_attack__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 5964.0, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.004861447960138321, 'timing/model': 1.2066299839643762, 'learning_rate/qwen_vl_interface': 4.283583937999991e-06, 'learning_rate/action_model': 4.2835839379999906e-05, 'epoch': 1.18}) +[2026-06-17 13:45:04,033][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_4000_state +[2026-06-17 13:45:04,036][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 13:45:07,598][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_4000_state/model.safetensors +[2026-06-17 13:45:15,356][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_4000_state/optimizer.bin +[2026-06-17 13:45:15,356][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_4000_state/sampler.bin +[2026-06-17 13:45:15,356][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_4000_state/sampler_1.bin +[2026-06-17 13:45:15,357][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_4000_state/random_states_0.pkl +[2026-06-17 13:45:15,405][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4000 +[2026-06-17 13:45:15,406][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 13:45:15,416][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 13:45:19,987][starVLA.training.train_starvla][INFO] - Step 4001, Loss: {'train/loss': 0.5580481886863708, 'train/grad_norm_pre_clip': 11.72940731048584, 'timing/data': 0.005366542958654463, 'timing/model': 1.1868069160263985, 'learning_rate/qwen_vl_interface': 4.2814665822946825e-06, 'learning_rate/action_model': 4.281466582294682e-05, 'epoch': 1.18}) +[2026-06-17 13:45:24,709][starVLA.training.train_starvla][INFO] - Step 4002, Loss: {'train/loss': 0.7156628370285034, 'train/grad_norm_pre_clip': 13.886630058288574, 'timing/data': 0.004924995941109955, 'timing/model': 1.194034883985296, 'learning_rate/qwen_vl_interface': 4.279349427367395e-06, 'learning_rate/action_model': 4.279349427367395e-05, 'epoch': 1.18}) +[2026-06-17 13:45:29,456][starVLA.training.train_starvla][INFO] - Step 4003, Loss: {'train/loss': 0.785825788974762, 'train/grad_norm_pre_clip': 14.52187728881836, 'timing/data': 0.004989319946616888, 'timing/model': 1.1953999609686434, 'learning_rate/qwen_vl_interface': 4.277232473657018e-06, 'learning_rate/action_model': 4.277232473657018e-05, 'epoch': 1.18}) +[2026-06-17 13:45:34,189][starVLA.training.train_starvla][INFO] - Step 4004, Loss: {'train/loss': 0.46718791127204895, 'train/grad_norm_pre_clip': 8.93264389038086, 'timing/data': 0.005313077941536903, 'timing/model': 1.192543516983278, 'learning_rate/qwen_vl_interface': 4.2751157216023934e-06, 'learning_rate/action_model': 4.2751157216023936e-05, 'epoch': 1.18}) +[2026-06-17 13:45:38,937][starVLA.training.train_starvla][INFO] - Step 4005, Loss: {'train/loss': 0.751372754573822, 'train/grad_norm_pre_clip': 12.683989524841309, 'timing/data': 0.004926418070681393, 'timing/model': 1.3212221909780055, 'learning_rate/qwen_vl_interface': 4.272999171642329e-06, 'learning_rate/action_model': 4.272999171642329e-05, 'epoch': 1.18}) +[2026-06-17 13:45:43,637][starVLA.training.train_starvla][INFO] - Step 4006, Loss: {'train/loss': 0.6743269562721252, 'train/grad_norm_pre_clip': 10.850404739379883, 'timing/data': 0.004593132995069027, 'timing/model': 1.19581238203682, 'learning_rate/qwen_vl_interface': 4.270882824215588e-06, 'learning_rate/action_model': 4.270882824215588e-05, 'epoch': 1.18}) +[2026-06-17 13:45:48,385][starVLA.training.train_starvla][INFO] - Step 4007, Loss: {'train/loss': 0.5882726907730103, 'train/grad_norm_pre_clip': 13.875870704650879, 'timing/data': 0.005048003979027271, 'timing/model': 1.1967464290792122, 'learning_rate/qwen_vl_interface': 4.268766679760891e-06, 'learning_rate/action_model': 4.268766679760891e-05, 'epoch': 1.18}) +[2026-06-17 13:45:53,191][starVLA.training.train_starvla][INFO] - Step 4008, Loss: {'train/loss': 0.9766374230384827, 'train/grad_norm_pre_clip': 18.05255699157715, 'timing/data': 0.005138296051882207, 'timing/model': 1.1982054689433426, 'learning_rate/qwen_vl_interface': 4.266650738716916e-06, 'learning_rate/action_model': 4.266650738716916e-05, 'epoch': 1.18}) +[2026-06-17 13:45:57,929][starVLA.training.train_starvla][INFO] - Step 4009, Loss: {'train/loss': 0.8168204426765442, 'train/grad_norm_pre_clip': 15.976749420166016, 'timing/data': 0.0049278069054707885, 'timing/model': 1.200969806057401, 'learning_rate/qwen_vl_interface': 4.264535001522302e-06, 'learning_rate/action_model': 4.2645350015223025e-05, 'epoch': 1.18}) +[2026-06-17 13:46:02,693][starVLA.training.train_starvla][INFO] - Step 4010, Loss: {'train/loss': 0.535834550857544, 'train/grad_norm_pre_clip': 14.530029296875, 'timing/data': 0.004833040991798043, 'timing/model': 1.1985877809347585, 'learning_rate/qwen_vl_interface': 4.262419468615644e-06, 'learning_rate/action_model': 4.2624194686156435e-05, 'epoch': 1.18}) +[2026-06-17 13:46:07,472][starVLA.training.train_starvla][INFO] - Step 4011, Loss: {'train/loss': 0.48449060320854187, 'train/grad_norm_pre_clip': 23.299945831298828, 'timing/data': 0.005175468046218157, 'timing/model': 1.328669710084796, 'learning_rate/qwen_vl_interface': 4.2603041404354925e-06, 'learning_rate/action_model': 4.260304140435492e-05, 'epoch': 1.18}) +[2026-06-17 13:46:12,110][starVLA.training.train_starvla][INFO] - Step 4012, Loss: {'train/loss': 0.9277112483978271, 'train/grad_norm_pre_clip': 13.352333068847656, 'timing/data': 0.005014697089791298, 'timing/model': 1.2000879170373082, 'learning_rate/qwen_vl_interface': 4.258189017420356e-06, 'learning_rate/action_model': 4.258189017420355e-05, 'epoch': 1.18}) +[2026-06-17 13:46:16,872][starVLA.training.train_starvla][INFO] - Step 4013, Loss: {'train/loss': 0.7961397767066956, 'train/grad_norm_pre_clip': 12.495243072509766, 'timing/data': 0.00492638407740742, 'timing/model': 1.1975606459891424, 'learning_rate/qwen_vl_interface': 4.2560741000087035e-06, 'learning_rate/action_model': 4.2560741000087034e-05, 'epoch': 1.18}) +[2026-06-17 13:46:21,590][starVLA.training.train_starvla][INFO] - Step 4014, Loss: {'train/loss': 0.680654764175415, 'train/grad_norm_pre_clip': 11.911454200744629, 'timing/data': 0.004772880929522216, 'timing/model': 1.2009921389399096, 'learning_rate/qwen_vl_interface': 4.253959388638961e-06, 'learning_rate/action_model': 4.2539593886389604e-05, 'epoch': 1.18}) +[2026-06-17 13:46:26,358][starVLA.training.train_starvla][INFO] - Step 4015, Loss: {'train/loss': 0.7448530197143555, 'train/grad_norm_pre_clip': 12.881636619567871, 'timing/data': 0.004811847931705415, 'timing/model': 1.200362968025729, 'learning_rate/qwen_vl_interface': 4.251844883749507e-06, 'learning_rate/action_model': 4.2518448837495066e-05, 'epoch': 1.18}) +[2026-06-17 13:46:31,125][starVLA.training.train_starvla][INFO] - Step 4016, Loss: {'train/loss': 0.4508087635040283, 'train/grad_norm_pre_clip': 8.693476676940918, 'timing/data': 0.006044493056833744, 'timing/model': 1.200322160962969, 'learning_rate/qwen_vl_interface': 4.249730585778684e-06, 'learning_rate/action_model': 4.2497305857786834e-05, 'epoch': 1.18}) +[2026-06-17 13:46:35,915][starVLA.training.train_starvla][INFO] - Step 4017, Loss: {'train/loss': 0.7477729320526123, 'train/grad_norm_pre_clip': 13.290212631225586, 'timing/data': 0.006728283013217151, 'timing/model': 1.3354487490141764, 'learning_rate/qwen_vl_interface': 4.247616495164788e-06, 'learning_rate/action_model': 4.247616495164787e-05, 'epoch': 1.18}) +[2026-06-17 13:46:40,610][starVLA.training.train_starvla][INFO] - Step 4018, Loss: {'train/loss': 0.5207748413085938, 'train/grad_norm_pre_clip': 9.535888671875, 'timing/data': 0.004707115003839135, 'timing/model': 1.2490060779964551, 'learning_rate/qwen_vl_interface': 4.24550261234607e-06, 'learning_rate/action_model': 4.2455026123460694e-05, 'epoch': 1.18}) +[2026-06-17 13:46:45,258][starVLA.training.train_starvla][INFO] - Step 4019, Loss: {'train/loss': 0.6068482398986816, 'train/grad_norm_pre_clip': 11.189630508422852, 'timing/data': 0.0050364729249849916, 'timing/model': 1.203514766995795, 'learning_rate/qwen_vl_interface': 4.2433889377607405e-06, 'learning_rate/action_model': 4.24338893776074e-05, 'epoch': 1.18}) +[2026-06-17 13:46:50,003][starVLA.training.train_starvla][INFO] - Step 4020, Loss: {'train/loss': 1.1346484422683716, 'train/grad_norm_pre_clip': 34.09101486206055, 'timing/data': 0.005070711951702833, 'timing/model': 1.2038884649518877, 'learning_rate/qwen_vl_interface': 4.241275471846968e-06, 'learning_rate/action_model': 4.241275471846968e-05, 'epoch': 1.18}) +[2026-06-17 13:46:54,784][starVLA.training.train_starvla][INFO] - Step 4021, Loss: {'train/loss': 0.8913680911064148, 'train/grad_norm_pre_clip': 14.261839866638184, 'timing/data': 0.005006495979614556, 'timing/model': 1.202516586985439, 'learning_rate/qwen_vl_interface': 4.239162215042876e-06, 'learning_rate/action_model': 4.239162215042876e-05, 'epoch': 1.18}) +[2026-06-17 13:46:59,561][starVLA.training.train_starvla][INFO] - Step 4022, Loss: {'train/loss': 0.9341949224472046, 'train/grad_norm_pre_clip': 18.73145866394043, 'timing/data': 0.0050484020030125976, 'timing/model': 1.2031061290763319, 'learning_rate/qwen_vl_interface': 4.237049167786544e-06, 'learning_rate/action_model': 4.2370491677865444e-05, 'epoch': 1.18}) +[2026-06-17 13:47:04,352][starVLA.training.train_starvla][INFO] - Step 4023, Loss: {'train/loss': 0.8643210530281067, 'train/grad_norm_pre_clip': 15.550455093383789, 'timing/data': 0.004911302006803453, 'timing/model': 1.33332616800908, 'learning_rate/qwen_vl_interface': 4.234936330516009e-06, 'learning_rate/action_model': 4.234936330516009e-05, 'epoch': 1.18}) +[2026-06-17 13:47:09,127][starVLA.training.train_starvla][INFO] - Step 4024, Loss: {'train/loss': 0.7679865956306458, 'train/grad_norm_pre_clip': 12.736748695373535, 'timing/data': 0.005639809998683631, 'timing/model': 1.3281195759773254, 'learning_rate/qwen_vl_interface': 4.2328237036692675e-06, 'learning_rate/action_model': 4.232823703669267e-05, 'epoch': 1.18}) +[2026-06-17 13:47:13,875][starVLA.training.train_starvla][INFO] - Step 4025, Loss: {'train/loss': 0.5657211542129517, 'train/grad_norm_pre_clip': 13.318092346191406, 'timing/data': 0.004907506052404642, 'timing/model': 1.204300735029392, 'learning_rate/qwen_vl_interface': 4.2307112876842645e-06, 'learning_rate/action_model': 4.2307112876842645e-05, 'epoch': 1.18}) +[2026-06-17 13:47:18,653][starVLA.training.train_starvla][INFO] - Step 4026, Loss: {'train/loss': 0.7823296189308167, 'train/grad_norm_pre_clip': 13.198945045471191, 'timing/data': 0.005233328090980649, 'timing/model': 1.2052814370254055, 'learning_rate/qwen_vl_interface': 4.228599082998908e-06, 'learning_rate/action_model': 4.228599082998908e-05, 'epoch': 1.18}) +[2026-06-17 13:47:23,439][starVLA.training.train_starvla][INFO] - Step 4027, Loss: {'train/loss': 0.6129202246665955, 'train/grad_norm_pre_clip': 11.580570220947266, 'timing/data': 0.005024365964345634, 'timing/model': 1.205385156092234, 'learning_rate/qwen_vl_interface': 4.226487090051062e-06, 'learning_rate/action_model': 4.2264870900510624e-05, 'epoch': 1.18}) +[2026-06-17 13:47:28,222][starVLA.training.train_starvla][INFO] - Step 4028, Loss: {'train/loss': 0.8936521410942078, 'train/grad_norm_pre_clip': 12.191227912902832, 'timing/data': 0.004748420906253159, 'timing/model': 1.2027364039095119, 'learning_rate/qwen_vl_interface': 4.224375309278543e-06, 'learning_rate/action_model': 4.2243753092785434e-05, 'epoch': 1.19}) +[2026-06-17 13:47:32,939][starVLA.training.train_starvla][INFO] - Step 4029, Loss: {'train/loss': 0.5819700956344604, 'train/grad_norm_pre_clip': 11.01605224609375, 'timing/data': 0.005009453976526856, 'timing/model': 1.2704171560471877, 'learning_rate/qwen_vl_interface': 4.222263741119127e-06, 'learning_rate/action_model': 4.222263741119126e-05, 'epoch': 1.19}) +[2026-06-17 13:47:37,589][starVLA.training.train_starvla][INFO] - Step 4030, Loss: {'train/loss': 0.7652628421783447, 'train/grad_norm_pre_clip': 15.547645568847656, 'timing/data': 0.00530302501283586, 'timing/model': 1.2034630849957466, 'learning_rate/qwen_vl_interface': 4.220152386010541e-06, 'learning_rate/action_model': 4.2201523860105405e-05, 'epoch': 1.19}) +[2026-06-17 13:47:42,353][starVLA.training.train_starvla][INFO] - Step 4031, Loss: {'train/loss': 0.5103017687797546, 'train/grad_norm_pre_clip': 12.31778621673584, 'timing/data': 0.005411447025835514, 'timing/model': 1.206044940976426, 'learning_rate/qwen_vl_interface': 4.218041244390476e-06, 'learning_rate/action_model': 4.218041244390475e-05, 'epoch': 1.19}) +[2026-06-17 13:47:47,142][starVLA.training.train_starvla][INFO] - Step 4032, Loss: {'train/loss': 0.7573186755180359, 'train/grad_norm_pre_clip': 36.67844772338867, 'timing/data': 0.004927512025460601, 'timing/model': 1.2043086610501632, 'learning_rate/qwen_vl_interface': 4.215930316696571e-06, 'learning_rate/action_model': 4.215930316696571e-05, 'epoch': 1.19}) +[2026-06-17 13:47:51,925][starVLA.training.train_starvla][INFO] - Step 4033, Loss: {'train/loss': 0.6427595615386963, 'train/grad_norm_pre_clip': 11.790706634521484, 'timing/data': 0.0051764450035989285, 'timing/model': 1.2052575550042093, 'learning_rate/qwen_vl_interface': 4.213819603366424e-06, 'learning_rate/action_model': 4.213819603366424e-05, 'epoch': 1.19}) +[2026-06-17 13:47:56,711][starVLA.training.train_starvla][INFO] - Step 4034, Loss: {'train/loss': 0.7011784315109253, 'train/grad_norm_pre_clip': 13.844265937805176, 'timing/data': 0.0046567689860239625, 'timing/model': 1.3313592720078304, 'learning_rate/qwen_vl_interface': 4.211709104837587e-06, 'learning_rate/action_model': 4.211709104837587e-05, 'epoch': 1.19}) +[2026-06-17 13:48:01,440][starVLA.training.train_starvla][INFO] - Step 4035, Loss: {'train/loss': 0.6821034550666809, 'train/grad_norm_pre_clip': 11.61532211303711, 'timing/data': 0.004653031937777996, 'timing/model': 1.280191463069059, 'learning_rate/qwen_vl_interface': 4.2095988215475705e-06, 'learning_rate/action_model': 4.20959882154757e-05, 'epoch': 1.19}) +[2026-06-17 13:48:06,102][starVLA.training.train_starvla][INFO] - Step 4036, Loss: {'train/loss': 0.7314791083335876, 'train/grad_norm_pre_clip': 11.880754470825195, 'timing/data': 0.004820075933821499, 'timing/model': 1.2088203089078888, 'learning_rate/qwen_vl_interface': 4.207488753933838e-06, 'learning_rate/action_model': 4.207488753933838e-05, 'epoch': 1.19}) +[2026-06-17 13:48:10,888][starVLA.training.train_starvla][INFO] - Step 4037, Loss: {'train/loss': 0.4699729084968567, 'train/grad_norm_pre_clip': 10.765599250793457, 'timing/data': 0.00478016200941056, 'timing/model': 1.2056758860126138, 'learning_rate/qwen_vl_interface': 4.205378902433807e-06, 'learning_rate/action_model': 4.205378902433807e-05, 'epoch': 1.19}) +[2026-06-17 13:48:15,692][starVLA.training.train_starvla][INFO] - Step 4038, Loss: {'train/loss': 0.7662035226821899, 'train/grad_norm_pre_clip': 13.378997802734375, 'timing/data': 0.004916697973385453, 'timing/model': 1.2039944920688868, 'learning_rate/qwen_vl_interface': 4.203269267484856e-06, 'learning_rate/action_model': 4.203269267484856e-05, 'epoch': 1.19}) +[2026-06-17 13:48:20,482][starVLA.training.train_starvla][INFO] - Step 4039, Loss: {'train/loss': 0.612841010093689, 'train/grad_norm_pre_clip': 10.662567138671875, 'timing/data': 0.004764538956806064, 'timing/model': 1.206732345977798, 'learning_rate/qwen_vl_interface': 4.201159849524313e-06, 'learning_rate/action_model': 4.201159849524312e-05, 'epoch': 1.19}) +[2026-06-17 13:48:25,287][starVLA.training.train_starvla][INFO] - Step 4040, Loss: {'train/loss': 1.1154747009277344, 'train/grad_norm_pre_clip': 18.16843605041504, 'timing/data': 0.0048171429662033916, 'timing/model': 1.2063615429215133, 'learning_rate/qwen_vl_interface': 4.19905064898946e-06, 'learning_rate/action_model': 4.199050648989459e-05, 'epoch': 1.19}) +[2026-06-17 13:48:30,032][starVLA.training.train_starvla][INFO] - Step 4041, Loss: {'train/loss': 0.7889032959938049, 'train/grad_norm_pre_clip': 12.901891708374023, 'timing/data': 0.00505960697773844, 'timing/model': 1.2904707189882174, 'learning_rate/qwen_vl_interface': 4.196941666317537e-06, 'learning_rate/action_model': 4.196941666317537e-05, 'epoch': 1.19}) +[2026-06-17 13:48:34,693][starVLA.training.train_starvla][INFO] - Step 4042, Loss: {'train/loss': 0.5821799635887146, 'train/grad_norm_pre_clip': 10.801895141601562, 'timing/data': 0.004860143992118537, 'timing/model': 1.2076771330321208, 'learning_rate/qwen_vl_interface': 4.194832901945741e-06, 'learning_rate/action_model': 4.194832901945741e-05, 'epoch': 1.19}) +[2026-06-17 13:48:39,477][starVLA.training.train_starvla][INFO] - Step 4043, Loss: {'train/loss': 0.8908361196517944, 'train/grad_norm_pre_clip': 15.424078941345215, 'timing/data': 0.004631364950910211, 'timing/model': 1.20425377599895, 'learning_rate/qwen_vl_interface': 4.192724356311221e-06, 'learning_rate/action_model': 4.19272435631122e-05, 'epoch': 1.19}) +[2026-06-17 13:48:44,265][starVLA.training.train_starvla][INFO] - Step 4044, Loss: {'train/loss': 0.4803316295146942, 'train/grad_norm_pre_clip': 12.087430000305176, 'timing/data': 0.004710067994892597, 'timing/model': 1.205439887009561, 'learning_rate/qwen_vl_interface': 4.190616029851079e-06, 'learning_rate/action_model': 4.190616029851079e-05, 'epoch': 1.19}) +[2026-06-17 13:48:49,051][starVLA.training.train_starvla][INFO] - Step 4045, Loss: {'train/loss': 0.8994264006614685, 'train/grad_norm_pre_clip': 11.94681167602539, 'timing/data': 0.004777773981913924, 'timing/model': 1.2057643089210615, 'learning_rate/qwen_vl_interface': 4.1885079230023715e-06, 'learning_rate/action_model': 4.1885079230023715e-05, 'epoch': 1.19}) +[2026-06-17 13:48:53,846][starVLA.training.train_starvla][INFO] - Step 4046, Loss: {'train/loss': 1.3627643585205078, 'train/grad_norm_pre_clip': 17.816566467285156, 'timing/data': 0.0048587649362161756, 'timing/model': 1.3359544300474226, 'learning_rate/qwen_vl_interface': 4.186400036202117e-06, 'learning_rate/action_model': 4.1864000362021175e-05, 'epoch': 1.19}) +[2026-06-17 13:48:58,502][starVLA.training.train_starvla][INFO] - Step 4047, Loss: {'train/loss': 0.7353624105453491, 'train/grad_norm_pre_clip': 10.463092803955078, 'timing/data': 0.004603918991051614, 'timing/model': 1.2044940498890355, 'learning_rate/qwen_vl_interface': 4.184292369887279e-06, 'learning_rate/action_model': 4.1842923698872785e-05, 'epoch': 1.19}) +[2026-06-17 13:49:03,286][starVLA.training.train_starvla][INFO] - Step 4048, Loss: {'train/loss': 0.8377528786659241, 'train/grad_norm_pre_clip': 12.905593872070312, 'timing/data': 0.0049104379722848535, 'timing/model': 1.2061150479130447, 'learning_rate/qwen_vl_interface': 4.182184924494778e-06, 'learning_rate/action_model': 4.1821849244947773e-05, 'epoch': 1.19}) +[2026-06-17 13:49:08,077][starVLA.training.train_starvla][INFO] - Step 4049, Loss: {'train/loss': 0.6125438213348389, 'train/grad_norm_pre_clip': 9.585455894470215, 'timing/data': 0.005095161031931639, 'timing/model': 1.204648017999716, 'learning_rate/qwen_vl_interface': 4.1800777004614915e-06, 'learning_rate/action_model': 4.180077700461492e-05, 'epoch': 1.19}) +[2026-06-17 13:49:12,787][starVLA.training.train_starvla][INFO] - Step 4050, Loss: {'train/loss': 0.8144798874855042, 'train/grad_norm_pre_clip': 13.934606552124023, 'timing/data': 0.004843933042138815, 'timing/model': 1.2049476839601994, 'learning_rate/qwen_vl_interface': 4.177970698224249e-06, 'learning_rate/action_model': 4.17797069822425e-05, 'epoch': 1.19}) +[2026-06-17 13:49:17,523][starVLA.training.train_starvla][INFO] - Step 4051, Loss: {'train/loss': 0.784948468208313, 'train/grad_norm_pre_clip': 12.542941093444824, 'timing/data': 0.005220858962275088, 'timing/model': 1.2052578009897843, 'learning_rate/qwen_vl_interface': 4.175863918219836e-06, 'learning_rate/action_model': 4.175863918219836e-05, 'epoch': 1.19}) +[2026-06-17 13:49:22,320][starVLA.training.train_starvla][INFO] - Step 4052, Loss: {'train/loss': 0.7604869604110718, 'train/grad_norm_pre_clip': 10.13455581665039, 'timing/data': 0.005115072941407561, 'timing/model': 1.3367529200622812, 'learning_rate/qwen_vl_interface': 4.173757360884986e-06, 'learning_rate/action_model': 4.173757360884986e-05, 'epoch': 1.19}) +[2026-06-17 13:49:26,993][starVLA.training.train_starvla][INFO] - Step 4053, Loss: {'train/loss': 0.8656617999076843, 'train/grad_norm_pre_clip': 13.139625549316406, 'timing/data': 0.004704833030700684, 'timing/model': 1.2153571440139785, 'learning_rate/qwen_vl_interface': 4.171651026656396e-06, 'learning_rate/action_model': 4.171651026656395e-05, 'epoch': 1.19}) +[2026-06-17 13:49:31,784][starVLA.training.train_starvla][INFO] - Step 4054, Loss: {'train/loss': 0.8344659805297852, 'train/grad_norm_pre_clip': 13.475794792175293, 'timing/data': 0.005227930028922856, 'timing/model': 1.2069373030681163, 'learning_rate/qwen_vl_interface': 4.1695449159707066e-06, 'learning_rate/action_model': 4.169544915970707e-05, 'epoch': 1.19}) +[2026-06-17 13:49:36,576][starVLA.training.train_starvla][INFO] - Step 4055, Loss: {'train/loss': 0.6415388584136963, 'train/grad_norm_pre_clip': 8.995954513549805, 'timing/data': 0.004965197993442416, 'timing/model': 1.20986548892688, 'learning_rate/qwen_vl_interface': 4.167439029264519e-06, 'learning_rate/action_model': 4.167439029264519e-05, 'epoch': 1.19}) +[2026-06-17 13:49:41,359][starVLA.training.train_starvla][INFO] - Step 4056, Loss: {'train/loss': 0.8838533759117126, 'train/grad_norm_pre_clip': 15.656272888183594, 'timing/data': 0.004731324966996908, 'timing/model': 1.2048556639347225, 'learning_rate/qwen_vl_interface': 4.165333366974385e-06, 'learning_rate/action_model': 4.165333366974385e-05, 'epoch': 1.19}) +[2026-06-17 13:49:46,102][starVLA.training.train_starvla][INFO] - Step 4057, Loss: {'train/loss': 0.5214175581932068, 'train/grad_norm_pre_clip': 10.673430442810059, 'timing/data': 0.0051237549632787704, 'timing/model': 1.2045748889213428, 'learning_rate/qwen_vl_interface': 4.163227929536812e-06, 'learning_rate/action_model': 4.163227929536812e-05, 'epoch': 1.19}) +[2026-06-17 13:49:50,877][starVLA.training.train_starvla][INFO] - Step 4058, Loss: {'train/loss': 0.8476266264915466, 'train/grad_norm_pre_clip': 10.501946449279785, 'timing/data': 0.004675514996051788, 'timing/model': 1.3229248060379177, 'learning_rate/qwen_vl_interface': 4.1611227173882596e-06, 'learning_rate/action_model': 4.1611227173882596e-05, 'epoch': 1.19}) +[2026-06-17 13:49:55,535][starVLA.training.train_starvla][INFO] - Step 4059, Loss: {'train/loss': 0.8503475785255432, 'train/grad_norm_pre_clip': 12.0982084274292, 'timing/data': 0.005028463085182011, 'timing/model': 1.2047011419199407, 'learning_rate/qwen_vl_interface': 4.1590177309651365e-06, 'learning_rate/action_model': 4.159017730965136e-05, 'epoch': 1.19}) +[2026-06-17 13:50:00,336][starVLA.training.train_starvla][INFO] - Step 4060, Loss: {'train/loss': 0.6070536375045776, 'train/grad_norm_pre_clip': 12.676376342773438, 'timing/data': 0.004842122085392475, 'timing/model': 1.2066290120128542, 'learning_rate/qwen_vl_interface': 4.156912970703815e-06, 'learning_rate/action_model': 4.1569129707038155e-05, 'epoch': 1.19}) +[2026-06-17 13:50:05,137][starVLA.training.train_starvla][INFO] - Step 4061, Loss: {'train/loss': 0.4990251958370209, 'train/grad_norm_pre_clip': 9.54240894317627, 'timing/data': 0.004981714999303222, 'timing/model': 1.206933824927546, 'learning_rate/qwen_vl_interface': 4.15480843704061e-06, 'learning_rate/action_model': 4.154808437040609e-05, 'epoch': 1.19}) +[2026-06-17 13:50:09,927][starVLA.training.train_starvla][INFO] - Step 4062, Loss: {'train/loss': 0.8953115940093994, 'train/grad_norm_pre_clip': 12.007922172546387, 'timing/data': 0.0047836219891905785, 'timing/model': 1.2054925300180912, 'learning_rate/qwen_vl_interface': 4.152704130411794e-06, 'learning_rate/action_model': 4.152704130411793e-05, 'epoch': 1.2}) +[2026-06-17 13:50:14,719][starVLA.training.train_starvla][INFO] - Step 4063, Loss: {'train/loss': 0.6635445952415466, 'train/grad_norm_pre_clip': 10.502038955688477, 'timing/data': 0.004907555063255131, 'timing/model': 1.3366268139798194, 'learning_rate/qwen_vl_interface': 4.150600051253591e-06, 'learning_rate/action_model': 4.150600051253591e-05, 'epoch': 1.2}) +[2026-06-17 13:50:19,375][starVLA.training.train_starvla][INFO] - Step 4064, Loss: {'train/loss': 0.6981644630432129, 'train/grad_norm_pre_clip': 11.291375160217285, 'timing/data': 0.004877569037489593, 'timing/model': 1.2040653560543433, 'learning_rate/qwen_vl_interface': 4.148496200002183e-06, 'learning_rate/action_model': 4.148496200002183e-05, 'epoch': 1.2}) +[2026-06-17 13:50:24,163][starVLA.training.train_starvla][INFO] - Step 4065, Loss: {'train/loss': 0.41819998621940613, 'train/grad_norm_pre_clip': 9.069308280944824, 'timing/data': 0.004915893077850342, 'timing/model': 1.2067384700058028, 'learning_rate/qwen_vl_interface': 4.146392577093697e-06, 'learning_rate/action_model': 4.1463925770936974e-05, 'epoch': 1.2}) +[2026-06-17 13:50:28,877][starVLA.training.train_starvla][INFO] - Step 4066, Loss: {'train/loss': 0.7699692249298096, 'train/grad_norm_pre_clip': 67.68490600585938, 'timing/data': 0.005068863974884152, 'timing/model': 1.206933608977124, 'learning_rate/qwen_vl_interface': 4.144289182964218e-06, 'learning_rate/action_model': 4.144289182964218e-05, 'epoch': 1.2}) +[2026-06-17 13:50:33,617][starVLA.training.train_starvla][INFO] - Step 4067, Loss: {'train/loss': 0.6910489797592163, 'train/grad_norm_pre_clip': 12.199637413024902, 'timing/data': 0.004962321021594107, 'timing/model': 1.2051400410709903, 'learning_rate/qwen_vl_interface': 4.14218601804978e-06, 'learning_rate/action_model': 4.1421860180497796e-05, 'epoch': 1.2}) +[2026-06-17 13:50:38,427][starVLA.training.train_starvla][INFO] - Step 4068, Loss: {'train/loss': 0.6677018404006958, 'train/grad_norm_pre_clip': 12.252202987670898, 'timing/data': 0.0049660170916467905, 'timing/model': 1.2043912949739024, 'learning_rate/qwen_vl_interface': 4.1400830827863725e-06, 'learning_rate/action_model': 4.1400830827863725e-05, 'epoch': 1.2}) +[2026-06-17 13:50:43,223][starVLA.training.train_starvla][INFO] - Step 4069, Loss: {'train/loss': 0.49826765060424805, 'train/grad_norm_pre_clip': 8.819725036621094, 'timing/data': 0.0048377979546785355, 'timing/model': 1.3334043129580095, 'learning_rate/qwen_vl_interface': 4.137980377609937e-06, 'learning_rate/action_model': 4.137980377609937e-05, 'epoch': 1.2}) +[2026-06-17 13:50:47,883][starVLA.training.train_starvla][INFO] - Step 4070, Loss: {'train/loss': 0.5370151400566101, 'train/grad_norm_pre_clip': 10.479532241821289, 'timing/data': 0.004790342994965613, 'timing/model': 1.2056936890585348, 'learning_rate/qwen_vl_interface': 4.135877902956364e-06, 'learning_rate/action_model': 4.1358779029563636e-05, 'epoch': 1.2}) +[2026-06-17 13:50:52,670][starVLA.training.train_starvla][INFO] - Step 4071, Loss: {'train/loss': 0.6961176991462708, 'train/grad_norm_pre_clip': 9.115706443786621, 'timing/data': 0.004674586933106184, 'timing/model': 1.2059751600027084, 'learning_rate/qwen_vl_interface': 4.133775659261503e-06, 'learning_rate/action_model': 4.133775659261502e-05, 'epoch': 1.2}) +[2026-06-17 13:50:57,457][starVLA.training.train_starvla][INFO] - Step 4072, Loss: {'train/loss': 0.6695181727409363, 'train/grad_norm_pre_clip': 10.540244102478027, 'timing/data': 0.004599769948981702, 'timing/model': 1.2096105680102482, 'learning_rate/qwen_vl_interface': 4.131673646961149e-06, 'learning_rate/action_model': 4.131673646961149e-05, 'epoch': 1.2}) +[2026-06-17 13:51:02,202][starVLA.training.train_starvla][INFO] - Step 4073, Loss: {'train/loss': 0.63047856092453, 'train/grad_norm_pre_clip': 28.630807876586914, 'timing/data': 0.004921669024042785, 'timing/model': 1.2056257829535753, 'learning_rate/qwen_vl_interface': 4.129571866491051e-06, 'learning_rate/action_model': 4.1295718664910504e-05, 'epoch': 1.2}) +[2026-06-17 13:51:06,985][starVLA.training.train_starvla][INFO] - Step 4074, Loss: {'train/loss': 0.8239495158195496, 'train/grad_norm_pre_clip': 11.187189102172852, 'timing/data': 0.0050818800227716565, 'timing/model': 1.2070340110221878, 'learning_rate/qwen_vl_interface': 4.127470318286908e-06, 'learning_rate/action_model': 4.127470318286908e-05, 'epoch': 1.2}) +[2026-06-17 13:51:11,771][starVLA.training.train_starvla][INFO] - Step 4075, Loss: {'train/loss': 0.7031696438789368, 'train/grad_norm_pre_clip': 12.1678466796875, 'timing/data': 0.005075740977190435, 'timing/model': 1.3350699950242415, 'learning_rate/qwen_vl_interface': 4.125369002784378e-06, 'learning_rate/action_model': 4.1253690027843776e-05, 'epoch': 1.2}) +[2026-06-17 13:51:16,475][starVLA.training.train_starvla][INFO] - Step 4076, Loss: {'train/loss': 0.6934802532196045, 'train/grad_norm_pre_clip': 11.999565124511719, 'timing/data': 0.0049344709841534495, 'timing/model': 1.203024231013842, 'learning_rate/qwen_vl_interface': 4.123267920419064e-06, 'learning_rate/action_model': 4.123267920419064e-05, 'epoch': 1.2}) +[2026-06-17 13:51:21,288][starVLA.training.train_starvla][INFO] - Step 4077, Loss: {'train/loss': 0.7414135336875916, 'train/grad_norm_pre_clip': 11.232169151306152, 'timing/data': 0.005267804022878408, 'timing/model': 1.220748538034968, 'learning_rate/qwen_vl_interface': 4.121167071626522e-06, 'learning_rate/action_model': 4.121167071626522e-05, 'epoch': 1.2}) +[2026-06-17 13:51:26,097][starVLA.training.train_starvla][INFO] - Step 4078, Loss: {'train/loss': 0.5517472624778748, 'train/grad_norm_pre_clip': 9.626221656799316, 'timing/data': 0.0052487150533124804, 'timing/model': 1.207996733020991, 'learning_rate/qwen_vl_interface': 4.119066456842261e-06, 'learning_rate/action_model': 4.119066456842261e-05, 'epoch': 1.2}) +[2026-06-17 13:51:30,895][starVLA.training.train_starvla][INFO] - Step 4079, Loss: {'train/loss': 0.5980838537216187, 'train/grad_norm_pre_clip': 12.389601707458496, 'timing/data': 0.00485963502433151, 'timing/model': 1.2051994210341945, 'learning_rate/qwen_vl_interface': 4.1169660765017416e-06, 'learning_rate/action_model': 4.1169660765017414e-05, 'epoch': 1.2}) +[2026-06-17 13:51:35,565][starVLA.training.train_starvla][INFO] - Step 4080, Loss: {'train/loss': 0.9166092276573181, 'train/grad_norm_pre_clip': 13.045509338378906, 'timing/data': 0.004998894990421832, 'timing/model': 1.2062715620268136, 'learning_rate/qwen_vl_interface': 4.1148659310403725e-06, 'learning_rate/action_model': 4.1148659310403723e-05, 'epoch': 1.2}) +[2026-06-17 13:51:40,285][starVLA.training.train_starvla][INFO] - Step 4081, Loss: {'train/loss': 0.7534769177436829, 'train/grad_norm_pre_clip': 11.855184555053711, 'timing/data': 0.0051998639246448874, 'timing/model': 1.2449885399546474, 'learning_rate/qwen_vl_interface': 4.112766020893516e-06, 'learning_rate/action_model': 4.1127660208935164e-05, 'epoch': 1.2}) +[2026-06-17 13:51:44,947][starVLA.training.train_starvla][INFO] - Step 4082, Loss: {'train/loss': 0.8701413869857788, 'train/grad_norm_pre_clip': 13.025861740112305, 'timing/data': 0.0050135370111092925, 'timing/model': 1.2065721990074962, 'learning_rate/qwen_vl_interface': 4.11066634649649e-06, 'learning_rate/action_model': 4.1106663464964895e-05, 'epoch': 1.2}) +[2026-06-17 13:51:49,682][starVLA.training.train_starvla][INFO] - Step 4083, Loss: {'train/loss': 0.8332151174545288, 'train/grad_norm_pre_clip': 13.469605445861816, 'timing/data': 0.00510622002184391, 'timing/model': 1.205742398975417, 'learning_rate/qwen_vl_interface': 4.1085669082845545e-06, 'learning_rate/action_model': 4.108566908284555e-05, 'epoch': 1.2}) +[2026-06-17 13:51:54,471][starVLA.training.train_starvla][INFO] - Step 4084, Loss: {'train/loss': 0.6854743361473083, 'train/grad_norm_pre_clip': 11.571539878845215, 'timing/data': 0.005022358032874763, 'timing/model': 1.2051036820048466, 'learning_rate/qwen_vl_interface': 4.10646770669293e-06, 'learning_rate/action_model': 4.106467706692929e-05, 'epoch': 1.2}) +[2026-06-17 13:51:59,264][starVLA.training.train_starvla][INFO] - Step 4085, Loss: {'train/loss': 0.7705742716789246, 'train/grad_norm_pre_clip': 12.161219596862793, 'timing/data': 0.005192131968215108, 'timing/model': 1.2092802779516205, 'learning_rate/qwen_vl_interface': 4.104368742156774e-06, 'learning_rate/action_model': 4.104368742156774e-05, 'epoch': 1.2}) +[2026-06-17 13:52:04,058][starVLA.training.train_starvla][INFO] - Step 4086, Loss: {'train/loss': 0.6376259326934814, 'train/grad_norm_pre_clip': 13.533203125, 'timing/data': 0.004760771989822388, 'timing/model': 1.331678114947863, 'learning_rate/qwen_vl_interface': 4.102270015111217e-06, 'learning_rate/action_model': 4.102270015111217e-05, 'epoch': 1.2}) +[2026-06-17 13:52:08,762][starVLA.training.train_starvla][INFO] - Step 4087, Loss: {'train/loss': 0.4391263425350189, 'train/grad_norm_pre_clip': 9.13361644744873, 'timing/data': 0.004940265906043351, 'timing/model': 1.2505934870569035, 'learning_rate/qwen_vl_interface': 4.100171525991318e-06, 'learning_rate/action_model': 4.100171525991318e-05, 'epoch': 1.2}) +[2026-06-17 13:52:13,434][starVLA.training.train_starvla][INFO] - Step 4088, Loss: {'train/loss': 0.9763301610946655, 'train/grad_norm_pre_clip': 14.845210075378418, 'timing/data': 0.004697448923252523, 'timing/model': 1.209408839000389, 'learning_rate/qwen_vl_interface': 4.098073275232098e-06, 'learning_rate/action_model': 4.098073275232098e-05, 'epoch': 1.2}) +[2026-06-17 13:52:18,206][starVLA.training.train_starvla][INFO] - Step 4089, Loss: {'train/loss': 0.7417450547218323, 'train/grad_norm_pre_clip': 12.822617530822754, 'timing/data': 0.005127704935148358, 'timing/model': 1.2072069100104272, 'learning_rate/qwen_vl_interface': 4.09597526326853e-06, 'learning_rate/action_model': 4.095975263268529e-05, 'epoch': 1.2}) +[2026-06-17 13:52:23,086][starVLA.training.train_starvla][INFO] - Step 4090, Loss: {'train/loss': 0.68763267993927, 'train/grad_norm_pre_clip': 11.56015396118164, 'timing/data': 0.0054640150628983974, 'timing/model': 1.2040786540601403, 'learning_rate/qwen_vl_interface': 4.09387749053553e-06, 'learning_rate/action_model': 4.0938774905355296e-05, 'epoch': 1.2}) +[2026-06-17 13:52:27,874][starVLA.training.train_starvla][INFO] - Step 4091, Loss: {'train/loss': 0.879934549331665, 'train/grad_norm_pre_clip': 12.41541576385498, 'timing/data': 0.005537439021281898, 'timing/model': 1.2059810399077833, 'learning_rate/qwen_vl_interface': 4.0917799574679725e-06, 'learning_rate/action_model': 4.091779957467973e-05, 'epoch': 1.2}) +[2026-06-17 13:52:32,675][starVLA.training.train_starvla][INFO] - Step 4092, Loss: {'train/loss': 0.4068302810192108, 'train/grad_norm_pre_clip': 7.930070877075195, 'timing/data': 0.004912699921987951, 'timing/model': 1.3401155750034377, 'learning_rate/qwen_vl_interface': 4.0896826645006725e-06, 'learning_rate/action_model': 4.089682664500673e-05, 'epoch': 1.2}) +[2026-06-17 13:52:37,471][starVLA.training.train_starvla][INFO] - Step 4093, Loss: {'train/loss': 0.7593449354171753, 'train/grad_norm_pre_clip': 12.477309226989746, 'timing/data': 0.004997776006348431, 'timing/model': 1.3401288239983842, 'learning_rate/qwen_vl_interface': 4.087585612068408e-06, 'learning_rate/action_model': 4.0875856120684074e-05, 'epoch': 1.2}) +[2026-06-17 13:52:42,128][starVLA.training.train_starvla][INFO] - Step 4094, Loss: {'train/loss': 0.9729738831520081, 'train/grad_norm_pre_clip': 13.85154914855957, 'timing/data': 0.005224096006713808, 'timing/model': 1.2047642600955442, 'learning_rate/qwen_vl_interface': 4.0854888006058955e-06, 'learning_rate/action_model': 4.085488800605896e-05, 'epoch': 1.2}) +[2026-06-17 13:52:46,888][starVLA.training.train_starvla][INFO] - Step 4095, Loss: {'train/loss': 0.7795142531394958, 'train/grad_norm_pre_clip': 13.221321105957031, 'timing/data': 0.004538942012004554, 'timing/model': 1.2057800240581855, 'learning_rate/qwen_vl_interface': 4.083392230547808e-06, 'learning_rate/action_model': 4.083392230547808e-05, 'epoch': 1.2}) +[2026-06-17 13:52:51,687][starVLA.training.train_starvla][INFO] - Step 4096, Loss: {'train/loss': 0.5129579901695251, 'train/grad_norm_pre_clip': 8.815363883972168, 'timing/data': 0.004921089974232018, 'timing/model': 1.2125834670150653, 'learning_rate/qwen_vl_interface': 4.081295902328765e-06, 'learning_rate/action_model': 4.081295902328765e-05, 'epoch': 1.21}) +[2026-06-17 13:52:56,474][starVLA.training.train_starvla][INFO] - Step 4097, Loss: {'train/loss': 0.8005571961402893, 'train/grad_norm_pre_clip': 13.869161605834961, 'timing/data': 0.004765910096466541, 'timing/model': 1.20400847599376, 'learning_rate/qwen_vl_interface': 4.079199816383341e-06, 'learning_rate/action_model': 4.0791998163833404e-05, 'epoch': 1.21}) +[2026-06-17 13:53:01,204][starVLA.training.train_starvla][INFO] - Step 4098, Loss: {'train/loss': 1.03178870677948, 'train/grad_norm_pre_clip': 20.97957992553711, 'timing/data': 0.0049066070932894945, 'timing/model': 1.2755609881132841, 'learning_rate/qwen_vl_interface': 4.077103973146054e-06, 'learning_rate/action_model': 4.077103973146054e-05, 'epoch': 1.21}) +[2026-06-17 13:53:05,871][starVLA.training.train_starvla][INFO] - Step 4099, Loss: {'train/loss': 0.6431436538696289, 'train/grad_norm_pre_clip': 10.953904151916504, 'timing/data': 0.004703990998677909, 'timing/model': 1.2052967820782214, 'learning_rate/qwen_vl_interface': 4.075008373051375e-06, 'learning_rate/action_model': 4.075008373051374e-05, 'epoch': 1.21}) +[2026-06-17 13:54:48,089][starVLA.training.train_starvla][INFO] - Step 4100, Loss: {'train/loss': 0.6285374164581299, 'train/grad_norm_pre_clip': 11.76479721069336, 'eval/loss': 0.755577564239502, 'eval/latency_0/loss': 0.755577564239502, 'eval/demon_fire_precision': 0.8749052312357847, 'eval/demon_fire_recall': 0.9584717607973422, 'eval/demon_fire_f1': 0.9147839873166864, 'eval/demon_fire_rate_model': 0.7335928809788654, 'eval/demon_fire_rate_teacher': 0.6696329254727474, 'eval/demon_left_precision': 0.8409915356711004, 'eval/demon_left_recall': 0.7881019830028329, 'eval/demon_left_f1': 0.8136882129277567, 'eval/demon_left_rate_model': 0.4599555061179088, 'eval/demon_left_rate_teacher': 0.49082313681868744, 'eval/demon_right_precision': 0.7987519500780031, 'eval/demon_right_recall': 0.8677966101694915, 'eval/demon_right_f1': 0.8318440292445166, 'eval/demon_right_rate_model': 0.5347608453837598, 'eval/demon_right_rate_teacher': 0.4922135706340378, 'eval/demon_move_f1': 0.8227661210861367, 'eval/demon_cc_f1': 0.8850876231535117, 'eval/latency_0/demon_fire_precision': 0.8749052312357847, 'eval/latency_0/demon_fire_recall': 0.9584717607973422, 'eval/latency_0/demon_fire_f1': 0.9147839873166864, 'eval/latency_0/demon_fire_rate_model': 0.7335928809788654, 'eval/latency_0/demon_fire_rate_teacher': 0.6696329254727474, 'eval/latency_0/demon_left_precision': 0.8409915356711004, 'eval/latency_0/demon_left_recall': 0.7881019830028329, 'eval/latency_0/demon_left_f1': 0.8136882129277567, 'eval/latency_0/demon_left_rate_model': 0.4599555061179088, 'eval/latency_0/demon_left_rate_teacher': 0.49082313681868744, 'eval/latency_0/demon_right_precision': 0.7987519500780031, 'eval/latency_0/demon_right_recall': 0.8677966101694915, 'eval/latency_0/demon_right_f1': 0.8318440292445166, 'eval/latency_0/demon_right_rate_model': 0.5347608453837598, 'eval/latency_0/demon_right_rate_teacher': 0.4922135706340378, 'eval/latency_0/demon_move_f1': 0.8227661210861367, 'eval/latency_0/demon_cc_f1': 0.8850876231535117, 'timing/data': 0.0048441889230161905, 'timing/model': 1.207026885007508, 'learning_rate/qwen_vl_interface': 4.072913016533726e-06, 'learning_rate/action_model': 4.072913016533726e-05, 'epoch': 1.21}) +[2026-06-17 13:54:48,089][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_4100_state +[2026-06-17 13:54:48,092][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 13:54:51,641][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_4100_state/model.safetensors +[2026-06-17 13:54:59,260][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_4100_state/optimizer.bin +[2026-06-17 13:54:59,260][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_4100_state/sampler.bin +[2026-06-17 13:54:59,260][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_4100_state/sampler_1.bin +[2026-06-17 13:54:59,261][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_4100_state/random_states_0.pkl +[2026-06-17 13:54:59,314][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4100 +[2026-06-17 13:54:59,315][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 13:54:59,328][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 13:55:04,053][starVLA.training.train_starvla][INFO] - Step 4101, Loss: {'train/loss': 0.8532810211181641, 'train/grad_norm_pre_clip': 15.267251014709473, 'timing/data': 0.004859281936660409, 'timing/model': 1.186248627025634, 'learning_rate/qwen_vl_interface': 4.070817904027474e-06, 'learning_rate/action_model': 4.070817904027474e-05, 'epoch': 1.21}) +[2026-06-17 13:55:08,836][starVLA.training.train_starvla][INFO] - Step 4102, Loss: {'train/loss': 0.6591288447380066, 'train/grad_norm_pre_clip': 14.596428871154785, 'timing/data': 0.004756362992338836, 'timing/model': 1.1902980710146949, 'learning_rate/qwen_vl_interface': 4.06872303596694e-06, 'learning_rate/action_model': 4.06872303596694e-05, 'epoch': 1.21}) +[2026-06-17 13:55:13,448][starVLA.training.train_starvla][INFO] - Step 4103, Loss: {'train/loss': 1.081154227256775, 'train/grad_norm_pre_clip': 16.184253692626953, 'timing/data': 0.004896898986771703, 'timing/model': 1.1921517739538103, 'learning_rate/qwen_vl_interface': 4.066628412786391e-06, 'learning_rate/action_model': 4.066628412786391e-05, 'epoch': 1.21}) +[2026-06-17 13:55:18,104][starVLA.training.train_starvla][INFO] - Step 4104, Loss: {'train/loss': 0.4506588876247406, 'train/grad_norm_pre_clip': 11.493669509887695, 'timing/data': 0.005067165940999985, 'timing/model': 1.1965071689337492, 'learning_rate/qwen_vl_interface': 4.0645340349200466e-06, 'learning_rate/action_model': 4.0645340349200464e-05, 'epoch': 1.21}) +[2026-06-17 13:55:22,845][starVLA.training.train_starvla][INFO] - Step 4105, Loss: {'train/loss': 0.6420160531997681, 'train/grad_norm_pre_clip': 13.228887557983398, 'timing/data': 0.004966195090673864, 'timing/model': 1.3163242780137807, 'learning_rate/qwen_vl_interface': 4.062439902802073e-06, 'learning_rate/action_model': 4.0624399028020725e-05, 'epoch': 1.21}) +[2026-06-17 13:55:27,465][starVLA.training.train_starvla][INFO] - Step 4106, Loss: {'train/loss': 0.6862170100212097, 'train/grad_norm_pre_clip': 15.464710235595703, 'timing/data': 0.0048538988921791315, 'timing/model': 1.1951886839233339, 'learning_rate/qwen_vl_interface': 4.060346016866584e-06, 'learning_rate/action_model': 4.060346016866584e-05, 'epoch': 1.21}) +[2026-06-17 13:55:32,226][starVLA.training.train_starvla][INFO] - Step 4107, Loss: {'train/loss': 0.36162269115448, 'train/grad_norm_pre_clip': 7.924200534820557, 'timing/data': 0.005371870007365942, 'timing/model': 1.194798463024199, 'learning_rate/qwen_vl_interface': 4.058252377547645e-06, 'learning_rate/action_model': 4.058252377547645e-05, 'epoch': 1.21}) +[2026-06-17 13:55:36,975][starVLA.training.train_starvla][INFO] - Step 4108, Loss: {'train/loss': 0.6528167724609375, 'train/grad_norm_pre_clip': 11.767330169677734, 'timing/data': 0.005144600989297032, 'timing/model': 1.1987536599626765, 'learning_rate/qwen_vl_interface': 4.056158985279272e-06, 'learning_rate/action_model': 4.0561589852792715e-05, 'epoch': 1.21}) +[2026-06-17 13:55:41,734][starVLA.training.train_starvla][INFO] - Step 4109, Loss: {'train/loss': 0.876587986946106, 'train/grad_norm_pre_clip': 14.20637321472168, 'timing/data': 0.004814091022126377, 'timing/model': 1.2011414340231568, 'learning_rate/qwen_vl_interface': 4.054065840495425e-06, 'learning_rate/action_model': 4.054065840495425e-05, 'epoch': 1.21}) +[2026-06-17 13:55:46,481][starVLA.training.train_starvla][INFO] - Step 4110, Loss: {'train/loss': 0.5892065167427063, 'train/grad_norm_pre_clip': 11.305048942565918, 'timing/data': 0.005242957966402173, 'timing/model': 1.3257764619775116, 'learning_rate/qwen_vl_interface': 4.0519729436300165e-06, 'learning_rate/action_model': 4.051972943630016e-05, 'epoch': 1.21}) +[2026-06-17 13:55:51,187][starVLA.training.train_starvla][INFO] - Step 4111, Loss: {'train/loss': 0.767257571220398, 'train/grad_norm_pre_clip': 13.984070777893066, 'timing/data': 0.004945893073454499, 'timing/model': 1.2806263420497999, 'learning_rate/qwen_vl_interface': 4.049880295116906e-06, 'learning_rate/action_model': 4.049880295116906e-05, 'epoch': 1.21}) +[2026-06-17 13:55:55,896][starVLA.training.train_starvla][INFO] - Step 4112, Loss: {'train/loss': 0.4702112376689911, 'train/grad_norm_pre_clip': 8.282007217407227, 'timing/data': 0.004779314971528947, 'timing/model': 1.1983481629285961, 'learning_rate/qwen_vl_interface': 4.047787895389902e-06, 'learning_rate/action_model': 4.0477878953899016e-05, 'epoch': 1.21}) +[2026-06-17 13:56:00,673][starVLA.training.train_starvla][INFO] - Step 4113, Loss: {'train/loss': 0.7362222671508789, 'train/grad_norm_pre_clip': 12.546097755432129, 'timing/data': 0.004989254986867309, 'timing/model': 1.2062777259852737, 'learning_rate/qwen_vl_interface': 4.0456957448827615e-06, 'learning_rate/action_model': 4.045695744882761e-05, 'epoch': 1.21}) +[2026-06-17 13:56:05,442][starVLA.training.train_starvla][INFO] - Step 4114, Loss: {'train/loss': 0.5034733414649963, 'train/grad_norm_pre_clip': 9.690573692321777, 'timing/data': 0.005087218945845962, 'timing/model': 1.1976893719984218, 'learning_rate/qwen_vl_interface': 4.043603844029188e-06, 'learning_rate/action_model': 4.0436038440291874e-05, 'epoch': 1.21}) +[2026-06-17 13:56:10,209][starVLA.training.train_starvla][INFO] - Step 4115, Loss: {'train/loss': 0.557096540927887, 'train/grad_norm_pre_clip': 10.838715553283691, 'timing/data': 0.005509547074325383, 'timing/model': 1.2002710639499128, 'learning_rate/qwen_vl_interface': 4.041512193262836e-06, 'learning_rate/action_model': 4.041512193262836e-05, 'epoch': 1.21}) +[2026-06-17 13:56:14,964][starVLA.training.train_starvla][INFO] - Step 4116, Loss: {'train/loss': 0.5567744374275208, 'train/grad_norm_pre_clip': 10.834012985229492, 'timing/data': 0.005232150084339082, 'timing/model': 1.199395125033334, 'learning_rate/qwen_vl_interface': 4.039420793017306e-06, 'learning_rate/action_model': 4.039420793017306e-05, 'epoch': 1.21}) +[2026-06-17 13:56:19,685][starVLA.training.train_starvla][INFO] - Step 4117, Loss: {'train/loss': 0.586139976978302, 'train/grad_norm_pre_clip': 11.678218841552734, 'timing/data': 0.005154842976480722, 'timing/model': 1.2887258640257642, 'learning_rate/qwen_vl_interface': 4.03732964372615e-06, 'learning_rate/action_model': 4.037329643726149e-05, 'epoch': 1.21}) +[2026-06-17 13:56:24,325][starVLA.training.train_starvla][INFO] - Step 4118, Loss: {'train/loss': 0.48897770047187805, 'train/grad_norm_pre_clip': 8.920825004577637, 'timing/data': 0.005306188017129898, 'timing/model': 1.2046721499646083, 'learning_rate/qwen_vl_interface': 4.035238745822861e-06, 'learning_rate/action_model': 4.035238745822861e-05, 'epoch': 1.21}) +[2026-06-17 13:56:29,107][starVLA.training.train_starvla][INFO] - Step 4119, Loss: {'train/loss': 0.8738572597503662, 'train/grad_norm_pre_clip': 20.299951553344727, 'timing/data': 0.00548846204765141, 'timing/model': 1.2015204199124128, 'learning_rate/qwen_vl_interface': 4.033148099740889e-06, 'learning_rate/action_model': 4.0331480997408885e-05, 'epoch': 1.21}) +[2026-06-17 13:56:33,904][starVLA.training.train_starvla][INFO] - Step 4120, Loss: {'train/loss': 0.7731762528419495, 'train/grad_norm_pre_clip': 14.289291381835938, 'timing/data': 0.005279117031022906, 'timing/model': 1.2001372709637508, 'learning_rate/qwen_vl_interface': 4.031057705913626e-06, 'learning_rate/action_model': 4.031057705913626e-05, 'epoch': 1.21}) +[2026-06-17 13:56:38,727][starVLA.training.train_starvla][INFO] - Step 4121, Loss: {'train/loss': 0.4791792035102844, 'train/grad_norm_pre_clip': 11.572996139526367, 'timing/data': 0.005323232035152614, 'timing/model': 1.199819556903094, 'learning_rate/qwen_vl_interface': 4.028967564774411e-06, 'learning_rate/action_model': 4.0289675647744106e-05, 'epoch': 1.21}) +[2026-06-17 13:56:43,521][starVLA.training.train_starvla][INFO] - Step 4122, Loss: {'train/loss': 0.9316585063934326, 'train/grad_norm_pre_clip': 12.837150573730469, 'timing/data': 0.005350193008780479, 'timing/model': 1.3442978070816025, 'learning_rate/qwen_vl_interface': 4.026877676756534e-06, 'learning_rate/action_model': 4.026877676756534e-05, 'epoch': 1.21}) +[2026-06-17 13:56:48,191][starVLA.training.train_starvla][INFO] - Step 4123, Loss: {'train/loss': 0.42431339621543884, 'train/grad_norm_pre_clip': 9.91562271118164, 'timing/data': 0.005663022049702704, 'timing/model': 1.20335005200468, 'learning_rate/qwen_vl_interface': 4.0247880422932305e-06, 'learning_rate/action_model': 4.02478804229323e-05, 'epoch': 1.21}) +[2026-06-17 13:56:52,986][starVLA.training.train_starvla][INFO] - Step 4124, Loss: {'train/loss': 0.4484153389930725, 'train/grad_norm_pre_clip': 9.89419937133789, 'timing/data': 0.005507037043571472, 'timing/model': 1.2007783629233018, 'learning_rate/qwen_vl_interface': 4.0226986618176855e-06, 'learning_rate/action_model': 4.022698661817686e-05, 'epoch': 1.21}) +[2026-06-17 13:56:57,761][starVLA.training.train_starvla][INFO] - Step 4125, Loss: {'train/loss': 0.746626615524292, 'train/grad_norm_pre_clip': 12.935516357421875, 'timing/data': 0.00539726996794343, 'timing/model': 1.2050216980278492, 'learning_rate/qwen_vl_interface': 4.0206095357630256e-06, 'learning_rate/action_model': 4.020609535763026e-05, 'epoch': 1.21}) +[2026-06-17 13:57:02,447][starVLA.training.train_starvla][INFO] - Step 4126, Loss: {'train/loss': 0.58543860912323, 'train/grad_norm_pre_clip': 13.239569664001465, 'timing/data': 0.0055172889260575175, 'timing/model': 1.2014099840307608, 'learning_rate/qwen_vl_interface': 4.018520664562334e-06, 'learning_rate/action_model': 4.018520664562334e-05, 'epoch': 1.21}) +[2026-06-17 13:57:07,172][starVLA.training.train_starvla][INFO] - Step 4127, Loss: {'train/loss': 0.7282497882843018, 'train/grad_norm_pre_clip': 15.48983097076416, 'timing/data': 0.0051845189882442355, 'timing/model': 1.2026126480195671, 'learning_rate/qwen_vl_interface': 4.016432048648635e-06, 'learning_rate/action_model': 4.016432048648635e-05, 'epoch': 1.21}) +[2026-06-17 13:57:11,962][starVLA.training.train_starvla][INFO] - Step 4128, Loss: {'train/loss': 0.7699462175369263, 'train/grad_norm_pre_clip': 18.2375431060791, 'timing/data': 0.00530969793908298, 'timing/model': 1.3418213649420068, 'learning_rate/qwen_vl_interface': 4.014343688454898e-06, 'learning_rate/action_model': 4.0143436884548976e-05, 'epoch': 1.21}) +[2026-06-17 13:57:16,622][starVLA.training.train_starvla][INFO] - Step 4129, Loss: {'train/loss': 0.6879675984382629, 'train/grad_norm_pre_clip': 15.516571998596191, 'timing/data': 0.005002298974432051, 'timing/model': 1.2044795430265367, 'learning_rate/qwen_vl_interface': 4.012255584414044e-06, 'learning_rate/action_model': 4.0122555844140436e-05, 'epoch': 1.21}) +[2026-06-17 13:57:21,394][starVLA.training.train_starvla][INFO] - Step 4130, Loss: {'train/loss': 0.6836439967155457, 'train/grad_norm_pre_clip': 12.656036376953125, 'timing/data': 0.005223011947236955, 'timing/model': 1.201973862014711, 'learning_rate/qwen_vl_interface': 4.010167736958939e-06, 'learning_rate/action_model': 4.010167736958938e-05, 'epoch': 1.22}) +[2026-06-17 13:57:26,188][starVLA.training.train_starvla][INFO] - Step 4131, Loss: {'train/loss': 0.6691157221794128, 'train/grad_norm_pre_clip': 15.308000564575195, 'timing/data': 0.005361643969081342, 'timing/model': 1.2055437159724534, 'learning_rate/qwen_vl_interface': 4.008080146522397e-06, 'learning_rate/action_model': 4.008080146522396e-05, 'epoch': 1.22}) +[2026-06-17 13:57:31,050][starVLA.training.train_starvla][INFO] - Step 4132, Loss: {'train/loss': 0.8300268650054932, 'train/grad_norm_pre_clip': 15.79313850402832, 'timing/data': 0.005902512930333614, 'timing/model': 1.2041617080103606, 'learning_rate/qwen_vl_interface': 4.005992813537176e-06, 'learning_rate/action_model': 4.005992813537176e-05, 'epoch': 1.22}) +[2026-06-17 13:57:35,811][starVLA.training.train_starvla][INFO] - Step 4133, Loss: {'train/loss': 0.8651123046875, 'train/grad_norm_pre_clip': 15.463876724243164, 'timing/data': 0.005088769015856087, 'timing/model': 1.20821290591266, 'learning_rate/qwen_vl_interface': 4.003905738435984e-06, 'learning_rate/action_model': 4.003905738435984e-05, 'epoch': 1.22}) +[2026-06-17 13:57:40,591][starVLA.training.train_starvla][INFO] - Step 4134, Loss: {'train/loss': 0.8667735457420349, 'train/grad_norm_pre_clip': 13.804407119750977, 'timing/data': 0.005044837947934866, 'timing/model': 1.3341925160493702, 'learning_rate/qwen_vl_interface': 4.001818921651475e-06, 'learning_rate/action_model': 4.0018189216514746e-05, 'epoch': 1.22}) +[2026-06-17 13:57:45,313][starVLA.training.train_starvla][INFO] - Step 4135, Loss: {'train/loss': 0.6888488531112671, 'train/grad_norm_pre_clip': 11.752096176147461, 'timing/data': 0.004831124912016094, 'timing/model': 1.2016467000357807, 'learning_rate/qwen_vl_interface': 3.999732363616244e-06, 'learning_rate/action_model': 3.999732363616244e-05, 'epoch': 1.22}) +[2026-06-17 13:57:50,128][starVLA.training.train_starvla][INFO] - Step 4136, Loss: {'train/loss': 0.7341592311859131, 'train/grad_norm_pre_clip': 18.030010223388672, 'timing/data': 0.005207597045227885, 'timing/model': 1.2013012859970331, 'learning_rate/qwen_vl_interface': 3.9976460647628394e-06, 'learning_rate/action_model': 3.997646064762839e-05, 'epoch': 1.22}) +[2026-06-17 13:57:54,900][starVLA.training.train_starvla][INFO] - Step 4137, Loss: {'train/loss': 0.7724661231040955, 'train/grad_norm_pre_clip': 14.519610404968262, 'timing/data': 0.0051361690275371075, 'timing/model': 1.2004902540938929, 'learning_rate/qwen_vl_interface': 3.995560025523755e-06, 'learning_rate/action_model': 3.995560025523754e-05, 'epoch': 1.22}) +[2026-06-17 13:57:59,691][starVLA.training.train_starvla][INFO] - Step 4138, Loss: {'train/loss': 0.4684855341911316, 'train/grad_norm_pre_clip': 11.159784317016602, 'timing/data': 0.005033096997067332, 'timing/model': 1.2023809329839423, 'learning_rate/qwen_vl_interface': 3.993474246331426e-06, 'learning_rate/action_model': 3.993474246331426e-05, 'epoch': 1.22}) +[2026-06-17 13:58:04,469][starVLA.training.train_starvla][INFO] - Step 4139, Loss: {'train/loss': 0.771787166595459, 'train/grad_norm_pre_clip': 14.308551788330078, 'timing/data': 0.005008821026422083, 'timing/model': 1.3294880450703204, 'learning_rate/qwen_vl_interface': 3.991388727618237e-06, 'learning_rate/action_model': 3.991388727618237e-05, 'epoch': 1.22}) +[2026-06-17 13:58:09,115][starVLA.training.train_starvla][INFO] - Step 4140, Loss: {'train/loss': 0.8333808779716492, 'train/grad_norm_pre_clip': 11.98176097869873, 'timing/data': 0.005166801973246038, 'timing/model': 1.201882985071279, 'learning_rate/qwen_vl_interface': 3.989303469816521e-06, 'learning_rate/action_model': 3.9893034698165205e-05, 'epoch': 1.22}) +[2026-06-17 13:58:13,882][starVLA.training.train_starvla][INFO] - Step 4141, Loss: {'train/loss': 0.8133290410041809, 'train/grad_norm_pre_clip': 14.836508750915527, 'timing/data': 0.005074990098364651, 'timing/model': 1.198859822936356, 'learning_rate/qwen_vl_interface': 3.9872184733585526e-06, 'learning_rate/action_model': 3.987218473358552e-05, 'epoch': 1.22}) +[2026-06-17 13:58:18,597][starVLA.training.train_starvla][INFO] - Step 4142, Loss: {'train/loss': 0.7890650629997253, 'train/grad_norm_pre_clip': 15.495159149169922, 'timing/data': 0.00528531102463603, 'timing/model': 1.2017932629678398, 'learning_rate/qwen_vl_interface': 3.985133738676552e-06, 'learning_rate/action_model': 3.985133738676552e-05, 'epoch': 1.22}) +[2026-06-17 13:58:23,334][starVLA.training.train_starvla][INFO] - Step 4143, Loss: {'train/loss': 0.7690410017967224, 'train/grad_norm_pre_clip': 21.396486282348633, 'timing/data': 0.00490532408002764, 'timing/model': 1.2063184489961714, 'learning_rate/qwen_vl_interface': 3.983049266202688e-06, 'learning_rate/action_model': 3.983049266202687e-05, 'epoch': 1.22}) +[2026-06-17 13:58:28,113][starVLA.training.train_starvla][INFO] - Step 4144, Loss: {'train/loss': 0.5800828337669373, 'train/grad_norm_pre_clip': 13.866242408752441, 'timing/data': 0.0055927010253071785, 'timing/model': 1.2023480159696192, 'learning_rate/qwen_vl_interface': 3.980965056369075e-06, 'learning_rate/action_model': 3.980965056369075e-05, 'epoch': 1.22}) +[2026-06-17 13:58:32,887][starVLA.training.train_starvla][INFO] - Step 4145, Loss: {'train/loss': 1.051537036895752, 'train/grad_norm_pre_clip': 17.143970489501953, 'timing/data': 0.005629118997603655, 'timing/model': 1.3296475949464366, 'learning_rate/qwen_vl_interface': 3.978881109607772e-06, 'learning_rate/action_model': 3.978881109607772e-05, 'epoch': 1.22}) +[2026-06-17 13:58:37,644][starVLA.training.train_starvla][INFO] - Step 4146, Loss: {'train/loss': 0.6007990837097168, 'train/grad_norm_pre_clip': 11.913809776306152, 'timing/data': 0.005138820968568325, 'timing/model': 1.2014022300718352, 'learning_rate/qwen_vl_interface': 3.976797426350783e-06, 'learning_rate/action_model': 3.976797426350783e-05, 'epoch': 1.22}) +[2026-06-17 13:58:42,410][starVLA.training.train_starvla][INFO] - Step 4147, Loss: {'train/loss': 0.7387256026268005, 'train/grad_norm_pre_clip': 15.38286304473877, 'timing/data': 0.0053047980181872845, 'timing/model': 1.2004526909440756, 'learning_rate/qwen_vl_interface': 3.974714007030055e-06, 'learning_rate/action_model': 3.974714007030056e-05, 'epoch': 1.22}) +[2026-06-17 13:58:47,189][starVLA.training.train_starvla][INFO] - Step 4148, Loss: {'train/loss': 0.9124645590782166, 'train/grad_norm_pre_clip': 12.442644119262695, 'timing/data': 0.005581576959230006, 'timing/model': 1.2104202409973368, 'learning_rate/qwen_vl_interface': 3.972630852077488e-06, 'learning_rate/action_model': 3.972630852077488e-05, 'epoch': 1.22}) +[2026-06-17 13:58:51,961][starVLA.training.train_starvla][INFO] - Step 4149, Loss: {'train/loss': 0.6802223324775696, 'train/grad_norm_pre_clip': 11.017290115356445, 'timing/data': 0.007223546039313078, 'timing/model': 1.216291117016226, 'learning_rate/qwen_vl_interface': 3.9705479619249195e-06, 'learning_rate/action_model': 3.9705479619249197e-05, 'epoch': 1.22}) +[2026-06-17 13:58:56,731][starVLA.training.train_starvla][INFO] - Step 4150, Loss: {'train/loss': 0.5565555691719055, 'train/grad_norm_pre_clip': 15.099451065063477, 'timing/data': 0.005152995930984616, 'timing/model': 1.2017890060087666, 'learning_rate/qwen_vl_interface': 3.9684653370041345e-06, 'learning_rate/action_model': 3.968465337004134e-05, 'epoch': 1.22}) +[2026-06-17 13:59:01,534][starVLA.training.train_starvla][INFO] - Step 4151, Loss: {'train/loss': 1.0791575908660889, 'train/grad_norm_pre_clip': 16.757123947143555, 'timing/data': 0.005151442950591445, 'timing/model': 1.3352809250354767, 'learning_rate/qwen_vl_interface': 3.966382977746866e-06, 'learning_rate/action_model': 3.9663829777468656e-05, 'epoch': 1.22}) +[2026-06-17 13:59:06,189][starVLA.training.train_starvla][INFO] - Step 4152, Loss: {'train/loss': 0.819430947303772, 'train/grad_norm_pre_clip': 15.612162590026855, 'timing/data': 0.005250234040431678, 'timing/model': 1.2066296499688178, 'learning_rate/qwen_vl_interface': 3.9643008845847845e-06, 'learning_rate/action_model': 3.964300884584785e-05, 'epoch': 1.22}) +[2026-06-17 13:59:10,998][starVLA.training.train_starvla][INFO] - Step 4153, Loss: {'train/loss': 0.7502520680427551, 'train/grad_norm_pre_clip': 13.746000289916992, 'timing/data': 0.005328524042852223, 'timing/model': 1.210368484025821, 'learning_rate/qwen_vl_interface': 3.962219057949517e-06, 'learning_rate/action_model': 3.9622190579495166e-05, 'epoch': 1.22}) +[2026-06-17 13:59:15,804][starVLA.training.train_starvla][INFO] - Step 4154, Loss: {'train/loss': 0.7230919599533081, 'train/grad_norm_pre_clip': 14.700016021728516, 'timing/data': 0.005285880994051695, 'timing/model': 1.2029154689516872, 'learning_rate/qwen_vl_interface': 3.960137498272619e-06, 'learning_rate/action_model': 3.9601374982726194e-05, 'epoch': 1.22}) +[2026-06-17 13:59:20,604][starVLA.training.train_starvla][INFO] - Step 4155, Loss: {'train/loss': 0.48095008730888367, 'train/grad_norm_pre_clip': 60.49357223510742, 'timing/data': 0.005183699075132608, 'timing/model': 1.2038461100310087, 'learning_rate/qwen_vl_interface': 3.958056205985611e-06, 'learning_rate/action_model': 3.95805620598561e-05, 'epoch': 1.22}) +[2026-06-17 13:59:25,277][starVLA.training.train_starvla][INFO] - Step 4156, Loss: {'train/loss': 0.811599612236023, 'train/grad_norm_pre_clip': 12.488761901855469, 'timing/data': 0.005305745988152921, 'timing/model': 1.2005747799994424, 'learning_rate/qwen_vl_interface': 3.95597518151994e-06, 'learning_rate/action_model': 3.955975181519939e-05, 'epoch': 1.22}) +[2026-06-17 13:59:29,978][starVLA.training.train_starvla][INFO] - Step 4157, Loss: {'train/loss': 0.7014655470848083, 'train/grad_norm_pre_clip': 17.792604446411133, 'timing/data': 0.005389651050791144, 'timing/model': 1.2385888940189034, 'learning_rate/qwen_vl_interface': 3.953894425307005e-06, 'learning_rate/action_model': 3.9538944253070055e-05, 'epoch': 1.22}) +[2026-06-17 13:59:34,619][starVLA.training.train_starvla][INFO] - Step 4158, Loss: {'train/loss': 0.6476829648017883, 'train/grad_norm_pre_clip': 10.840479850769043, 'timing/data': 0.005197206977754831, 'timing/model': 1.1997642489150167, 'learning_rate/qwen_vl_interface': 3.951813937778151e-06, 'learning_rate/action_model': 3.95181393777815e-05, 'epoch': 1.22}) +[2026-06-17 13:59:39,348][starVLA.training.train_starvla][INFO] - Step 4159, Loss: {'train/loss': 0.7096104025840759, 'train/grad_norm_pre_clip': 14.360355377197266, 'timing/data': 0.004916894948109984, 'timing/model': 1.202913889894262, 'learning_rate/qwen_vl_interface': 3.949733719364664e-06, 'learning_rate/action_model': 3.949733719364664e-05, 'epoch': 1.22}) +[2026-06-17 13:59:44,140][starVLA.training.train_starvla][INFO] - Step 4160, Loss: {'train/loss': 0.5764859318733215, 'train/grad_norm_pre_clip': 12.11669635772705, 'timing/data': 0.004772447980940342, 'timing/model': 1.2006729770218953, 'learning_rate/qwen_vl_interface': 3.947653770497778e-06, 'learning_rate/action_model': 3.947653770497778e-05, 'epoch': 1.22}) +[2026-06-17 13:59:48,924][starVLA.training.train_starvla][INFO] - Step 4161, Loss: {'train/loss': 0.704871654510498, 'train/grad_norm_pre_clip': 13.673850059509277, 'timing/data': 0.005237436038441956, 'timing/model': 1.208614124916494, 'learning_rate/qwen_vl_interface': 3.945574091608664e-06, 'learning_rate/action_model': 3.9455740916086636e-05, 'epoch': 1.22}) +[2026-06-17 13:59:53,700][starVLA.training.train_starvla][INFO] - Step 4162, Loss: {'train/loss': 0.6714708805084229, 'train/grad_norm_pre_clip': 10.424976348876953, 'timing/data': 0.005094514926895499, 'timing/model': 1.3277083229040727, 'learning_rate/qwen_vl_interface': 3.9434946831284465e-06, 'learning_rate/action_model': 3.943494683128447e-05, 'epoch': 1.22}) +[2026-06-17 13:59:58,424][starVLA.training.train_starvla][INFO] - Step 4163, Loss: {'train/loss': 0.681776225566864, 'train/grad_norm_pre_clip': 13.785828590393066, 'timing/data': 0.005307316896505654, 'timing/model': 1.2688882580259815, 'learning_rate/qwen_vl_interface': 3.941415545488187e-06, 'learning_rate/action_model': 3.941415545488187e-05, 'epoch': 1.22}) +[2026-06-17 14:00:03,080][starVLA.training.train_starvla][INFO] - Step 4164, Loss: {'train/loss': 0.7082217931747437, 'train/grad_norm_pre_clip': 14.616222381591797, 'timing/data': 0.005530296009965241, 'timing/model': 1.2022495749406517, 'learning_rate/qwen_vl_interface': 3.939336679118892e-06, 'learning_rate/action_model': 3.939336679118892e-05, 'epoch': 1.23}) +[2026-06-17 14:00:07,814][starVLA.training.train_starvla][INFO] - Step 4165, Loss: {'train/loss': 0.7343707084655762, 'train/grad_norm_pre_clip': 10.268401145935059, 'timing/data': 0.005015376955270767, 'timing/model': 1.2023811800172552, 'learning_rate/qwen_vl_interface': 3.937258084451514e-06, 'learning_rate/action_model': 3.937258084451514e-05, 'epoch': 1.23}) +[2026-06-17 14:00:12,611][starVLA.training.train_starvla][INFO] - Step 4166, Loss: {'train/loss': 0.7902703285217285, 'train/grad_norm_pre_clip': 15.542000770568848, 'timing/data': 0.005288319895043969, 'timing/model': 1.2024231409886852, 'learning_rate/qwen_vl_interface': 3.9351797619169485e-06, 'learning_rate/action_model': 3.935179761916949e-05, 'epoch': 1.23}) +[2026-06-17 14:00:17,411][starVLA.training.train_starvla][INFO] - Step 4167, Loss: {'train/loss': 0.7989451885223389, 'train/grad_norm_pre_clip': 12.824920654296875, 'timing/data': 0.005302443983964622, 'timing/model': 1.2039536940865219, 'learning_rate/qwen_vl_interface': 3.933101711946034e-06, 'learning_rate/action_model': 3.9331017119460336e-05, 'epoch': 1.23}) +[2026-06-17 14:00:22,194][starVLA.training.train_starvla][INFO] - Step 4168, Loss: {'train/loss': 0.8874666094779968, 'train/grad_norm_pre_clip': 14.224427223205566, 'timing/data': 0.005331612075679004, 'timing/model': 1.3318180160131305, 'learning_rate/qwen_vl_interface': 3.93102393496955e-06, 'learning_rate/action_model': 3.93102393496955e-05, 'epoch': 1.23}) +[2026-06-17 14:00:26,960][starVLA.training.train_starvla][INFO] - Step 4169, Loss: {'train/loss': 0.8753571510314941, 'train/grad_norm_pre_clip': 12.86765193939209, 'timing/data': 0.004869989003054798, 'timing/model': 1.3193023490021005, 'learning_rate/qwen_vl_interface': 3.928946431418223e-06, 'learning_rate/action_model': 3.9289464314182225e-05, 'epoch': 1.23}) +[2026-06-17 14:00:31,603][starVLA.training.train_starvla][INFO] - Step 4170, Loss: {'train/loss': 0.8413426280021667, 'train/grad_norm_pre_clip': 10.869747161865234, 'timing/data': 0.004989028093405068, 'timing/model': 1.2021976209944114, 'learning_rate/qwen_vl_interface': 3.926869201722723e-06, 'learning_rate/action_model': 3.926869201722723e-05, 'epoch': 1.23}) +[2026-06-17 14:00:36,366][starVLA.training.train_starvla][INFO] - Step 4171, Loss: {'train/loss': 0.5921002626419067, 'train/grad_norm_pre_clip': 11.105045318603516, 'timing/data': 0.004884428926743567, 'timing/model': 1.2064290560083464, 'learning_rate/qwen_vl_interface': 3.9247922463136615e-06, 'learning_rate/action_model': 3.924792246313662e-05, 'epoch': 1.23}) +[2026-06-17 14:00:41,145][starVLA.training.train_starvla][INFO] - Step 4172, Loss: {'train/loss': 0.6802423000335693, 'train/grad_norm_pre_clip': 10.81049633026123, 'timing/data': 0.005171997006982565, 'timing/model': 1.2064774230821058, 'learning_rate/qwen_vl_interface': 3.922715565621593e-06, 'learning_rate/action_model': 3.9227155656215926e-05, 'epoch': 1.23}) +[2026-06-17 14:00:45,939][starVLA.training.train_starvla][INFO] - Step 4173, Loss: {'train/loss': 0.9518536329269409, 'train/grad_norm_pre_clip': 14.854963302612305, 'timing/data': 0.005021186894737184, 'timing/model': 1.203344130073674, 'learning_rate/qwen_vl_interface': 3.9206391600770164e-06, 'learning_rate/action_model': 3.920639160077016e-05, 'epoch': 1.23}) +[2026-06-17 14:00:50,683][starVLA.training.train_starvla][INFO] - Step 4174, Loss: {'train/loss': 0.7391209006309509, 'train/grad_norm_pre_clip': 12.614535331726074, 'timing/data': 0.005406016018241644, 'timing/model': 1.28760007803794, 'learning_rate/qwen_vl_interface': 3.918563030110373e-06, 'learning_rate/action_model': 3.918563030110373e-05, 'epoch': 1.23}) +[2026-06-17 14:00:55,346][starVLA.training.train_starvla][INFO] - Step 4175, Loss: {'train/loss': 0.5998243093490601, 'train/grad_norm_pre_clip': 12.0493745803833, 'timing/data': 0.005618534982204437, 'timing/model': 1.2029375169659033, 'learning_rate/qwen_vl_interface': 3.916487176152045e-06, 'learning_rate/action_model': 3.916487176152045e-05, 'epoch': 1.23}) +[2026-06-17 14:01:00,101][starVLA.training.train_starvla][INFO] - Step 4176, Loss: {'train/loss': 0.5336611270904541, 'train/grad_norm_pre_clip': 10.509489059448242, 'timing/data': 0.005180811043828726, 'timing/model': 1.2001346809556708, 'learning_rate/qwen_vl_interface': 3.914411598632359e-06, 'learning_rate/action_model': 3.914411598632359e-05, 'epoch': 1.23}) +[2026-06-17 14:01:04,877][starVLA.training.train_starvla][INFO] - Step 4177, Loss: {'train/loss': 0.733188807964325, 'train/grad_norm_pre_clip': 11.548551559448242, 'timing/data': 0.005130704026669264, 'timing/model': 1.2033813679590821, 'learning_rate/qwen_vl_interface': 3.912336297981588e-06, 'learning_rate/action_model': 3.912336297981588e-05, 'epoch': 1.23}) +[2026-06-17 14:01:09,656][starVLA.training.train_starvla][INFO] - Step 4178, Loss: {'train/loss': 0.7852031588554382, 'train/grad_norm_pre_clip': 11.911474227905273, 'timing/data': 0.00494151096791029, 'timing/model': 1.2050673179328442, 'learning_rate/qwen_vl_interface': 3.910261274629941e-06, 'learning_rate/action_model': 3.910261274629941e-05, 'epoch': 1.23}) +[2026-06-17 14:01:14,435][starVLA.training.train_starvla][INFO] - Step 4179, Loss: {'train/loss': 0.6511958837509155, 'train/grad_norm_pre_clip': 12.139842987060547, 'timing/data': 0.004882637062110007, 'timing/model': 1.3317314310697839, 'learning_rate/qwen_vl_interface': 3.908186529007576e-06, 'learning_rate/action_model': 3.908186529007575e-05, 'epoch': 1.23}) +[2026-06-17 14:01:19,186][starVLA.training.train_starvla][INFO] - Step 4180, Loss: {'train/loss': 0.6300317049026489, 'train/grad_norm_pre_clip': 10.918574333190918, 'timing/data': 0.005167998024262488, 'timing/model': 1.2837715880014002, 'learning_rate/qwen_vl_interface': 3.906112061544582e-06, 'learning_rate/action_model': 3.906112061544582e-05, 'epoch': 1.23}) +[2026-06-17 14:01:23,830][starVLA.training.train_starvla][INFO] - Step 4181, Loss: {'train/loss': 0.8159466981887817, 'train/grad_norm_pre_clip': 11.32119369506836, 'timing/data': 0.005181428045034409, 'timing/model': 1.2027746860403568, 'learning_rate/qwen_vl_interface': 3.904037872671008e-06, 'learning_rate/action_model': 3.904037872671008e-05, 'epoch': 1.23}) +[2026-06-17 14:01:28,616][starVLA.training.train_starvla][INFO] - Step 4182, Loss: {'train/loss': 0.6622766852378845, 'train/grad_norm_pre_clip': 11.162152290344238, 'timing/data': 0.005432677920907736, 'timing/model': 1.2036516240332276, 'learning_rate/qwen_vl_interface': 3.90196396281683e-06, 'learning_rate/action_model': 3.9019639628168295e-05, 'epoch': 1.23}) +[2026-06-17 14:01:33,406][starVLA.training.train_starvla][INFO] - Step 4183, Loss: {'train/loss': 0.6265426874160767, 'train/grad_norm_pre_clip': 8.831184387207031, 'timing/data': 0.0052105640061199665, 'timing/model': 1.2026227299356833, 'learning_rate/qwen_vl_interface': 3.899890332411972e-06, 'learning_rate/action_model': 3.8998903324119724e-05, 'epoch': 1.23}) +[2026-06-17 14:01:38,207][starVLA.training.train_starvla][INFO] - Step 4184, Loss: {'train/loss': 0.9049611687660217, 'train/grad_norm_pre_clip': 15.740503311157227, 'timing/data': 0.00493431300856173, 'timing/model': 1.2060959510272369, 'learning_rate/qwen_vl_interface': 3.897816981886302e-06, 'learning_rate/action_model': 3.897816981886302e-05, 'epoch': 1.23}) +[2026-06-17 14:01:43,010][starVLA.training.train_starvla][INFO] - Step 4185, Loss: {'train/loss': 0.6267893314361572, 'train/grad_norm_pre_clip': 11.129840850830078, 'timing/data': 0.00686991005204618, 'timing/model': 1.2068396200193092, 'learning_rate/qwen_vl_interface': 3.895743911669627e-06, 'learning_rate/action_model': 3.895743911669627e-05, 'epoch': 1.23}) +[2026-06-17 14:01:47,760][starVLA.training.train_starvla][INFO] - Step 4186, Loss: {'train/loss': 0.7041571736335754, 'train/grad_norm_pre_clip': 10.821990966796875, 'timing/data': 0.004955557058565319, 'timing/model': 1.2916771059390157, 'learning_rate/qwen_vl_interface': 3.893671122191698e-06, 'learning_rate/action_model': 3.8936711221916974e-05, 'epoch': 1.23}) +[2026-06-17 14:01:52,467][starVLA.training.train_starvla][INFO] - Step 4187, Loss: {'train/loss': 0.6656337380409241, 'train/grad_norm_pre_clip': 10.840782165527344, 'timing/data': 0.0053542290115728974, 'timing/model': 1.2100516069913283, 'learning_rate/qwen_vl_interface': 3.8915986138822e-06, 'learning_rate/action_model': 3.8915986138822e-05, 'epoch': 1.23}) +[2026-06-17 14:01:57,274][starVLA.training.train_starvla][INFO] - Step 4188, Loss: {'train/loss': 0.6198067665100098, 'train/grad_norm_pre_clip': 11.38620662689209, 'timing/data': 0.005063239950686693, 'timing/model': 1.2084839859744534, 'learning_rate/qwen_vl_interface': 3.889526387170775e-06, 'learning_rate/action_model': 3.889526387170775e-05, 'epoch': 1.23}) +[2026-06-17 14:02:02,084][starVLA.training.train_starvla][INFO] - Step 4189, Loss: {'train/loss': 0.5666483640670776, 'train/grad_norm_pre_clip': 10.268627166748047, 'timing/data': 0.005140910041518509, 'timing/model': 1.2056215649936348, 'learning_rate/qwen_vl_interface': 3.8874544424869924e-06, 'learning_rate/action_model': 3.8874544424869926e-05, 'epoch': 1.23}) +[2026-06-17 14:02:06,899][starVLA.training.train_starvla][INFO] - Step 4190, Loss: {'train/loss': 0.5149421691894531, 'train/grad_norm_pre_clip': 7.927077293395996, 'timing/data': 0.005236907978542149, 'timing/model': 1.2069027790566906, 'learning_rate/qwen_vl_interface': 3.885382780260371e-06, 'learning_rate/action_model': 3.885382780260371e-05, 'epoch': 1.23}) +[2026-06-17 14:02:11,783][starVLA.training.train_starvla][INFO] - Step 4191, Loss: {'train/loss': 0.7736575603485107, 'train/grad_norm_pre_clip': 13.293773651123047, 'timing/data': 0.005145539063960314, 'timing/model': 1.360056301113218, 'learning_rate/qwen_vl_interface': 3.883311400920366e-06, 'learning_rate/action_model': 3.883311400920366e-05, 'epoch': 1.23}) +[2026-06-17 14:02:16,449][starVLA.training.train_starvla][INFO] - Step 4192, Loss: {'train/loss': 0.48120611906051636, 'train/grad_norm_pre_clip': 9.938703536987305, 'timing/data': 0.005198022001422942, 'timing/model': 1.2053286019945517, 'learning_rate/qwen_vl_interface': 3.88124030489638e-06, 'learning_rate/action_model': 3.88124030489638e-05, 'epoch': 1.23}) +[2026-06-17 14:02:21,253][starVLA.training.train_starvla][INFO] - Step 4193, Loss: {'train/loss': 0.7024343609809875, 'train/grad_norm_pre_clip': 11.30444622039795, 'timing/data': 0.005208939081057906, 'timing/model': 1.203306442941539, 'learning_rate/qwen_vl_interface': 3.8791694926177495e-06, 'learning_rate/action_model': 3.87916949261775e-05, 'epoch': 1.23}) +[2026-06-17 14:02:26,043][starVLA.training.train_starvla][INFO] - Step 4194, Loss: {'train/loss': 0.5777221322059631, 'train/grad_norm_pre_clip': 11.331892013549805, 'timing/data': 0.005413879989646375, 'timing/model': 1.204732671030797, 'learning_rate/qwen_vl_interface': 3.877098964513759e-06, 'learning_rate/action_model': 3.877098964513759e-05, 'epoch': 1.23}) +[2026-06-17 14:02:30,775][starVLA.training.train_starvla][INFO] - Step 4195, Loss: {'train/loss': 0.4341069757938385, 'train/grad_norm_pre_clip': 8.725729942321777, 'timing/data': 0.005013679969124496, 'timing/model': 1.2085729020182043, 'learning_rate/qwen_vl_interface': 3.875028721013633e-06, 'learning_rate/action_model': 3.875028721013632e-05, 'epoch': 1.23}) +[2026-06-17 14:02:35,537][starVLA.training.train_starvla][INFO] - Step 4196, Loss: {'train/loss': 0.702506422996521, 'train/grad_norm_pre_clip': 11.115793228149414, 'timing/data': 0.005218505975790322, 'timing/model': 1.2063729269430041, 'learning_rate/qwen_vl_interface': 3.87295876254653e-06, 'learning_rate/action_model': 3.8729587625465304e-05, 'epoch': 1.23}) +[2026-06-17 14:02:40,340][starVLA.training.train_starvla][INFO] - Step 4197, Loss: {'train/loss': 0.6515415906906128, 'train/grad_norm_pre_clip': 10.710343360900879, 'timing/data': 0.0052063820185139775, 'timing/model': 1.3371328249340877, 'learning_rate/qwen_vl_interface': 3.870889089541558e-06, 'learning_rate/action_model': 3.870889089541558e-05, 'epoch': 1.23}) +[2026-06-17 14:02:45,018][starVLA.training.train_starvla][INFO] - Step 4198, Loss: {'train/loss': 0.6409761905670166, 'train/grad_norm_pre_clip': 12.674554824829102, 'timing/data': 0.0048472919734194875, 'timing/model': 1.2077572220005095, 'learning_rate/qwen_vl_interface': 3.868819702427762e-06, 'learning_rate/action_model': 3.8688197024277616e-05, 'epoch': 1.24}) +[2026-06-17 14:02:49,816][starVLA.training.train_starvla][INFO] - Step 4199, Loss: {'train/loss': 0.6607730984687805, 'train/grad_norm_pre_clip': 12.567327499389648, 'timing/data': 0.004899130901321769, 'timing/model': 1.2055143100442365, 'learning_rate/qwen_vl_interface': 3.866750601634128e-06, 'learning_rate/action_model': 3.8667506016341277e-05, 'epoch': 1.24}) +[2026-06-17 14:04:31,105][starVLA.training.train_starvla][INFO] - Step 4200, Loss: {'train/loss': 0.6846475601196289, 'train/grad_norm_pre_clip': 12.578743934631348, 'eval/loss': 0.7370494604110718, 'eval/latency_0/loss': 0.7370494604110718, 'eval/demon_fire_precision': 0.9136363636363637, 'eval/demon_fire_recall': 0.9181893687707641, 'eval/demon_fire_f1': 0.915907207953604, 'eval/demon_fire_rate_model': 0.6729699666295884, 'eval/demon_fire_rate_teacher': 0.6696329254727474, 'eval/demon_left_precision': 0.8472816127061699, 'eval/demon_left_recall': 0.7858356940509915, 'eval/demon_left_f1': 0.8154027042915931, 'eval/demon_left_rate_model': 0.45522803114571747, 'eval/demon_left_rate_teacher': 0.49082313681868744, 'eval/demon_right_precision': 0.797514241325738, 'eval/demon_right_recall': 0.8700564971751412, 'eval/demon_right_f1': 0.832207511483383, 'eval/demon_right_rate_model': 0.5369855394883204, 'eval/demon_right_rate_teacher': 0.4922135706340378, 'eval/demon_move_f1': 0.8238051078874881, 'eval/demon_cc_f1': 0.8861843620960104, 'eval/latency_0/demon_fire_precision': 0.9136363636363637, 'eval/latency_0/demon_fire_recall': 0.9181893687707641, 'eval/latency_0/demon_fire_f1': 0.915907207953604, 'eval/latency_0/demon_fire_rate_model': 0.6729699666295884, 'eval/latency_0/demon_fire_rate_teacher': 0.6696329254727474, 'eval/latency_0/demon_left_precision': 0.8472816127061699, 'eval/latency_0/demon_left_recall': 0.7858356940509915, 'eval/latency_0/demon_left_f1': 0.8154027042915931, 'eval/latency_0/demon_left_rate_model': 0.45522803114571747, 'eval/latency_0/demon_left_rate_teacher': 0.49082313681868744, 'eval/latency_0/demon_right_precision': 0.797514241325738, 'eval/latency_0/demon_right_recall': 0.8700564971751412, 'eval/latency_0/demon_right_f1': 0.832207511483383, 'eval/latency_0/demon_right_rate_model': 0.5369855394883204, 'eval/latency_0/demon_right_rate_teacher': 0.4922135706340378, 'eval/latency_0/demon_move_f1': 0.8238051078874881, 'eval/latency_0/demon_cc_f1': 0.8861843620960104, 'timing/data': 0.005371915991418064, 'timing/model': 1.2093212659237906, 'learning_rate/qwen_vl_interface': 3.864681787589581e-06, 'learning_rate/action_model': 3.8646817875895816e-05, 'epoch': 1.24}) +[2026-06-17 14:04:31,105][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_4200_state +[2026-06-17 14:04:31,109][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 14:04:34,772][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_4200_state/model.safetensors +[2026-06-17 14:04:42,419][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_4200_state/optimizer.bin +[2026-06-17 14:04:42,419][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_4200_state/sampler.bin +[2026-06-17 14:04:42,420][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_4200_state/sampler_1.bin +[2026-06-17 14:04:42,420][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_4200_state/random_states_0.pkl +[2026-06-17 14:04:42,466][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4200 +[2026-06-17 14:04:42,466][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 14:04:42,476][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 14:04:47,195][starVLA.training.train_starvla][INFO] - Step 4201, Loss: {'train/loss': 0.6777862906455994, 'train/grad_norm_pre_clip': 10.69054889678955, 'timing/data': 0.005181845044717193, 'timing/model': 1.1908518370473757, 'learning_rate/qwen_vl_interface': 3.862613260722991e-06, 'learning_rate/action_model': 3.862613260722992e-05, 'epoch': 1.24}) +[2026-06-17 14:04:51,937][starVLA.training.train_starvla][INFO] - Step 4202, Loss: {'train/loss': 0.7910158634185791, 'train/grad_norm_pre_clip': 12.79681396484375, 'timing/data': 0.005265792948193848, 'timing/model': 1.1973917429568246, 'learning_rate/qwen_vl_interface': 3.860545021463166e-06, 'learning_rate/action_model': 3.860545021463167e-05, 'epoch': 1.24}) +[2026-06-17 14:04:56,681][starVLA.training.train_starvla][INFO] - Step 4203, Loss: {'train/loss': 0.8549841046333313, 'train/grad_norm_pre_clip': 13.417325973510742, 'timing/data': 0.004843000904656947, 'timing/model': 1.1982988329837099, 'learning_rate/qwen_vl_interface': 3.858477070238854e-06, 'learning_rate/action_model': 3.8584770702388534e-05, 'epoch': 1.24}) +[2026-06-17 14:05:01,423][starVLA.training.train_starvla][INFO] - Step 4204, Loss: {'train/loss': 0.589479923248291, 'train/grad_norm_pre_clip': 9.560420989990234, 'timing/data': 0.004872096003964543, 'timing/model': 1.3218794770073146, 'learning_rate/qwen_vl_interface': 3.856409407478741e-06, 'learning_rate/action_model': 3.856409407478741e-05, 'epoch': 1.24}) +[2026-06-17 14:05:06,049][starVLA.training.train_starvla][INFO] - Step 4205, Loss: {'train/loss': 0.6207062602043152, 'train/grad_norm_pre_clip': 11.15247631072998, 'timing/data': 0.005116668995469809, 'timing/model': 1.1940662069246173, 'learning_rate/qwen_vl_interface': 3.854342033611456e-06, 'learning_rate/action_model': 3.854342033611456e-05, 'epoch': 1.24}) +[2026-06-17 14:05:10,806][starVLA.training.train_starvla][INFO] - Step 4206, Loss: {'train/loss': 0.6855370402336121, 'train/grad_norm_pre_clip': 13.638178825378418, 'timing/data': 0.004935770994052291, 'timing/model': 1.2027499260148033, 'learning_rate/qwen_vl_interface': 3.852274949065571e-06, 'learning_rate/action_model': 3.852274949065571e-05, 'epoch': 1.24}) +[2026-06-17 14:05:15,484][starVLA.training.train_starvla][INFO] - Step 4207, Loss: {'train/loss': 0.70450758934021, 'train/grad_norm_pre_clip': 11.73404598236084, 'timing/data': 0.00505189795512706, 'timing/model': 1.2034803989809006, 'learning_rate/qwen_vl_interface': 3.850208154269593e-06, 'learning_rate/action_model': 3.8502081542695934e-05, 'epoch': 1.24}) +[2026-06-17 14:05:20,198][starVLA.training.train_starvla][INFO] - Step 4208, Loss: {'train/loss': 0.8204354047775269, 'train/grad_norm_pre_clip': 12.919445991516113, 'timing/data': 0.005074584041722119, 'timing/model': 1.1988999960012734, 'learning_rate/qwen_vl_interface': 3.8481416496519725e-06, 'learning_rate/action_model': 3.848141649651972e-05, 'epoch': 1.24}) +[2026-06-17 14:05:24,940][starVLA.training.train_starvla][INFO] - Step 4209, Loss: {'train/loss': 0.5197197794914246, 'train/grad_norm_pre_clip': 17.024986267089844, 'timing/data': 0.005208445945754647, 'timing/model': 1.2044111309805885, 'learning_rate/qwen_vl_interface': 3.846075435641093e-06, 'learning_rate/action_model': 3.846075435641093e-05, 'epoch': 1.24}) +[2026-06-17 14:05:29,713][starVLA.training.train_starvla][INFO] - Step 4210, Loss: {'train/loss': 0.5197727680206299, 'train/grad_norm_pre_clip': 11.556185722351074, 'timing/data': 0.0051987109472975135, 'timing/model': 1.3281234840396792, 'learning_rate/qwen_vl_interface': 3.844009512665289e-06, 'learning_rate/action_model': 3.844009512665289e-05, 'epoch': 1.24}) +[2026-06-17 14:05:34,383][starVLA.training.train_starvla][INFO] - Step 4211, Loss: {'train/loss': 0.7995983958244324, 'train/grad_norm_pre_clip': 13.81826114654541, 'timing/data': 0.005053600063547492, 'timing/model': 1.2050894920248538, 'learning_rate/qwen_vl_interface': 3.841943881152826e-06, 'learning_rate/action_model': 3.841943881152826e-05, 'epoch': 1.24}) +[2026-06-17 14:05:39,163][starVLA.training.train_starvla][INFO] - Step 4212, Loss: {'train/loss': 0.8701027631759644, 'train/grad_norm_pre_clip': 15.692837715148926, 'timing/data': 0.005142876994796097, 'timing/model': 1.2000137940049171, 'learning_rate/qwen_vl_interface': 3.83987854153191e-06, 'learning_rate/action_model': 3.83987854153191e-05, 'epoch': 1.24}) +[2026-06-17 14:05:43,957][starVLA.training.train_starvla][INFO] - Step 4213, Loss: {'train/loss': 0.80210280418396, 'train/grad_norm_pre_clip': 13.628471374511719, 'timing/data': 0.004868732998147607, 'timing/model': 1.2046424050349742, 'learning_rate/qwen_vl_interface': 3.83781349423069e-06, 'learning_rate/action_model': 3.83781349423069e-05, 'epoch': 1.24}) +[2026-06-17 14:05:48,676][starVLA.training.train_starvla][INFO] - Step 4214, Loss: {'train/loss': 0.6670668125152588, 'train/grad_norm_pre_clip': 11.648783683776855, 'timing/data': 0.004743375931866467, 'timing/model': 1.202043982106261, 'learning_rate/qwen_vl_interface': 3.835748739677253e-06, 'learning_rate/action_model': 3.835748739677253e-05, 'epoch': 1.24}) +[2026-06-17 14:05:53,452][starVLA.training.train_starvla][INFO] - Step 4215, Loss: {'train/loss': 0.6951134204864502, 'train/grad_norm_pre_clip': 13.1781005859375, 'timing/data': 0.004879538086242974, 'timing/model': 1.2024528729962185, 'learning_rate/qwen_vl_interface': 3.8336842782996245e-06, 'learning_rate/action_model': 3.8336842782996244e-05, 'epoch': 1.24}) +[2026-06-17 14:05:58,230][starVLA.training.train_starvla][INFO] - Step 4216, Loss: {'train/loss': 0.6760739684104919, 'train/grad_norm_pre_clip': 11.76709270477295, 'timing/data': 0.005149186938069761, 'timing/model': 1.3342517870478332, 'learning_rate/qwen_vl_interface': 3.831620110525766e-06, 'learning_rate/action_model': 3.831620110525767e-05, 'epoch': 1.24}) +[2026-06-17 14:06:02,885][starVLA.training.train_starvla][INFO] - Step 4217, Loss: {'train/loss': 0.768991231918335, 'train/grad_norm_pre_clip': 12.390192985534668, 'timing/data': 0.00479700998403132, 'timing/model': 1.2020436499733478, 'learning_rate/qwen_vl_interface': 3.829556236783588e-06, 'learning_rate/action_model': 3.8295562367835884e-05, 'epoch': 1.24}) +[2026-06-17 14:06:07,675][starVLA.training.train_starvla][INFO] - Step 4218, Loss: {'train/loss': 0.9378455281257629, 'train/grad_norm_pre_clip': 14.244673728942871, 'timing/data': 0.004945334978401661, 'timing/model': 1.2047254130011424, 'learning_rate/qwen_vl_interface': 3.827492657500931e-06, 'learning_rate/action_model': 3.8274926575009305e-05, 'epoch': 1.24}) +[2026-06-17 14:06:12,450][starVLA.training.train_starvla][INFO] - Step 4219, Loss: {'train/loss': 0.642142653465271, 'train/grad_norm_pre_clip': 11.585945129394531, 'timing/data': 0.004876479972153902, 'timing/model': 1.205934825935401, 'learning_rate/qwen_vl_interface': 3.8254293731055754e-06, 'learning_rate/action_model': 3.825429373105576e-05, 'epoch': 1.24}) +[2026-06-17 14:06:17,189][starVLA.training.train_starvla][INFO] - Step 4220, Loss: {'train/loss': 0.6454750895500183, 'train/grad_norm_pre_clip': 12.452013969421387, 'timing/data': 0.005032508983276784, 'timing/model': 1.2049864099826664, 'learning_rate/qwen_vl_interface': 3.823366384025243e-06, 'learning_rate/action_model': 3.8233663840252426e-05, 'epoch': 1.24}) +[2026-06-17 14:06:21,963][starVLA.training.train_starvla][INFO] - Step 4221, Loss: {'train/loss': 0.7605518102645874, 'train/grad_norm_pre_clip': 13.576557159423828, 'timing/data': 0.005169866024516523, 'timing/model': 1.2046665999805555, 'learning_rate/qwen_vl_interface': 3.821303690687595e-06, 'learning_rate/action_model': 3.821303690687595e-05, 'epoch': 1.24}) +[2026-06-17 14:06:26,768][starVLA.training.train_starvla][INFO] - Step 4222, Loss: {'train/loss': 0.6013206243515015, 'train/grad_norm_pre_clip': 18.149723052978516, 'timing/data': 0.0048623919719830155, 'timing/model': 1.3352539340266958, 'learning_rate/qwen_vl_interface': 3.81924129352023e-06, 'learning_rate/action_model': 3.8192412935202304e-05, 'epoch': 1.24}) +[2026-06-17 14:06:31,432][starVLA.training.train_starvla][INFO] - Step 4223, Loss: {'train/loss': 0.4446457326412201, 'train/grad_norm_pre_clip': 8.511568069458008, 'timing/data': 0.005154556012712419, 'timing/model': 1.2052298970520496, 'learning_rate/qwen_vl_interface': 3.817179192950682e-06, 'learning_rate/action_model': 3.8171791929506826e-05, 'epoch': 1.24}) +[2026-06-17 14:06:36,219][starVLA.training.train_starvla][INFO] - Step 4224, Loss: {'train/loss': 0.6351280808448792, 'train/grad_norm_pre_clip': 17.05219841003418, 'timing/data': 0.004680481972172856, 'timing/model': 1.2057850339915603, 'learning_rate/qwen_vl_interface': 3.815117389406432e-06, 'learning_rate/action_model': 3.8151173894064324e-05, 'epoch': 1.24}) +[2026-06-17 14:06:41,015][starVLA.training.train_starvla][INFO] - Step 4225, Loss: {'train/loss': 0.7633448839187622, 'train/grad_norm_pre_clip': 13.322203636169434, 'timing/data': 0.004879168933257461, 'timing/model': 1.2128390739671886, 'learning_rate/qwen_vl_interface': 3.8130558833148906e-06, 'learning_rate/action_model': 3.813055883314891e-05, 'epoch': 1.24}) +[2026-06-17 14:06:45,805][starVLA.training.train_starvla][INFO] - Step 4226, Loss: {'train/loss': 0.5622004866600037, 'train/grad_norm_pre_clip': 12.042009353637695, 'timing/data': 0.005118984961882234, 'timing/model': 1.2057960069505498, 'learning_rate/qwen_vl_interface': 3.81099467510341e-06, 'learning_rate/action_model': 3.81099467510341e-05, 'epoch': 1.24}) +[2026-06-17 14:06:50,473][starVLA.training.train_starvla][INFO] - Step 4227, Loss: {'train/loss': 0.6699094772338867, 'train/grad_norm_pre_clip': 12.197726249694824, 'timing/data': 0.005552895017899573, 'timing/model': 1.2058837729273364, 'learning_rate/qwen_vl_interface': 3.8089337651992815e-06, 'learning_rate/action_model': 3.808933765199281e-05, 'epoch': 1.24}) +[2026-06-17 14:06:55,181][starVLA.training.train_starvla][INFO] - Step 4228, Loss: {'train/loss': 0.6360605955123901, 'train/grad_norm_pre_clip': 12.674360275268555, 'timing/data': 0.005129072000272572, 'timing/model': 1.2495469700079411, 'learning_rate/qwen_vl_interface': 3.8068731540297344e-06, 'learning_rate/action_model': 3.806873154029734e-05, 'epoch': 1.24}) +[2026-06-17 14:06:59,844][starVLA.training.train_starvla][INFO] - Step 4229, Loss: {'train/loss': 0.3448975682258606, 'train/grad_norm_pre_clip': 7.389513969421387, 'timing/data': 0.004846115014515817, 'timing/model': 1.2083010689821094, 'learning_rate/qwen_vl_interface': 3.8048128420219366e-06, 'learning_rate/action_model': 3.804812842021936e-05, 'epoch': 1.24}) +[2026-06-17 14:07:04,587][starVLA.training.train_starvla][INFO] - Step 4230, Loss: {'train/loss': 0.786734938621521, 'train/grad_norm_pre_clip': 14.06517219543457, 'timing/data': 0.005040533025749028, 'timing/model': 1.2075919499620795, 'learning_rate/qwen_vl_interface': 3.80275282960299e-06, 'learning_rate/action_model': 3.80275282960299e-05, 'epoch': 1.24}) +[2026-06-17 14:07:09,415][starVLA.training.train_starvla][INFO] - Step 4231, Loss: {'train/loss': 0.8657510280609131, 'train/grad_norm_pre_clip': 13.16262149810791, 'timing/data': 0.0045963149750605226, 'timing/model': 1.2380821760743856, 'learning_rate/qwen_vl_interface': 3.8006931171999394e-06, 'learning_rate/action_model': 3.800693117199939e-05, 'epoch': 1.24}) +[2026-06-17 14:07:14,209][starVLA.training.train_starvla][INFO] - Step 4232, Loss: {'train/loss': 0.5578943490982056, 'train/grad_norm_pre_clip': 12.035247802734375, 'timing/data': 0.005067768041044474, 'timing/model': 1.2052524669561535, 'learning_rate/qwen_vl_interface': 3.798633705239766e-06, 'learning_rate/action_model': 3.7986337052397655e-05, 'epoch': 1.25}) +[2026-06-17 14:07:18,996][starVLA.training.train_starvla][INFO] - Step 4233, Loss: {'train/loss': 0.5127370357513428, 'train/grad_norm_pre_clip': 11.090935707092285, 'timing/data': 0.005262118997052312, 'timing/model': 1.3321484309853986, 'learning_rate/qwen_vl_interface': 3.796574594149387e-06, 'learning_rate/action_model': 3.796574594149387e-05, 'epoch': 1.25}) +[2026-06-17 14:07:23,780][starVLA.training.train_starvla][INFO] - Step 4234, Loss: {'train/loss': 1.1843308210372925, 'train/grad_norm_pre_clip': 15.725691795349121, 'timing/data': 0.004504902986809611, 'timing/model': 1.3264720879960805, 'learning_rate/qwen_vl_interface': 3.794515784355657e-06, 'learning_rate/action_model': 3.7945157843556575e-05, 'epoch': 1.25}) +[2026-06-17 14:07:28,441][starVLA.training.train_starvla][INFO] - Step 4235, Loss: {'train/loss': 0.7745837569236755, 'train/grad_norm_pre_clip': 13.564009666442871, 'timing/data': 0.004890325013548136, 'timing/model': 1.2064927390310913, 'learning_rate/qwen_vl_interface': 3.7924572762853727e-06, 'learning_rate/action_model': 3.7924572762853725e-05, 'epoch': 1.25}) +[2026-06-17 14:07:33,178][starVLA.training.train_starvla][INFO] - Step 4236, Loss: {'train/loss': 0.616958498954773, 'train/grad_norm_pre_clip': 11.670495986938477, 'timing/data': 0.0051857869839295745, 'timing/model': 1.2024558110861108, 'learning_rate/qwen_vl_interface': 3.7903990703652643e-06, 'learning_rate/action_model': 3.790399070365264e-05, 'epoch': 1.25}) +[2026-06-17 14:07:37,973][starVLA.training.train_starvla][INFO] - Step 4237, Loss: {'train/loss': 0.5562261939048767, 'train/grad_norm_pre_clip': 15.527417182922363, 'timing/data': 0.004933072021231055, 'timing/model': 1.210616608033888, 'learning_rate/qwen_vl_interface': 3.7883411670219976e-06, 'learning_rate/action_model': 3.788341167021998e-05, 'epoch': 1.25}) +[2026-06-17 14:07:42,772][starVLA.training.train_starvla][INFO] - Step 4238, Loss: {'train/loss': 0.6137175559997559, 'train/grad_norm_pre_clip': 9.998600959777832, 'timing/data': 0.00496019097045064, 'timing/model': 1.2062755270162597, 'learning_rate/qwen_vl_interface': 3.7862835666821795e-06, 'learning_rate/action_model': 3.78628356668218e-05, 'epoch': 1.25}) +[2026-06-17 14:07:47,559][starVLA.training.train_starvla][INFO] - Step 4239, Loss: {'train/loss': 0.8470668196678162, 'train/grad_norm_pre_clip': 15.123312950134277, 'timing/data': 0.0048830839805305, 'timing/model': 1.3333152430132031, 'learning_rate/qwen_vl_interface': 3.7842262697723543e-06, 'learning_rate/action_model': 3.784226269772354e-05, 'epoch': 1.25}) +[2026-06-17 14:07:52,348][starVLA.training.train_starvla][INFO] - Step 4240, Loss: {'train/loss': 0.6097750067710876, 'train/grad_norm_pre_clip': 11.88701343536377, 'timing/data': 0.004779983079060912, 'timing/model': 1.3321138280443847, 'learning_rate/qwen_vl_interface': 3.782169276719e-06, 'learning_rate/action_model': 3.782169276719e-05, 'epoch': 1.25}) +[2026-06-17 14:07:57,008][starVLA.training.train_starvla][INFO] - Step 4241, Loss: {'train/loss': 0.40190017223358154, 'train/grad_norm_pre_clip': 9.593114852905273, 'timing/data': 0.004786864970810711, 'timing/model': 1.2042336779413745, 'learning_rate/qwen_vl_interface': 3.7801125879485355e-06, 'learning_rate/action_model': 3.7801125879485356e-05, 'epoch': 1.25}) +[2026-06-17 14:08:01,759][starVLA.training.train_starvla][INFO] - Step 4242, Loss: {'train/loss': 0.6824498772621155, 'train/grad_norm_pre_clip': 10.78628921508789, 'timing/data': 0.0051517600659281015, 'timing/model': 1.2033719969913363, 'learning_rate/qwen_vl_interface': 3.77805620388731e-06, 'learning_rate/action_model': 3.7780562038873105e-05, 'epoch': 1.25}) +[2026-06-17 14:08:06,573][starVLA.training.train_starvla][INFO] - Step 4243, Loss: {'train/loss': 0.656487226486206, 'train/grad_norm_pre_clip': 10.910391807556152, 'timing/data': 0.0049853980308398604, 'timing/model': 1.2131801070645452, 'learning_rate/qwen_vl_interface': 3.7760001249616204e-06, 'learning_rate/action_model': 3.77600012496162e-05, 'epoch': 1.25}) +[2026-06-17 14:08:11,369][starVLA.training.train_starvla][INFO] - Step 4244, Loss: {'train/loss': 0.6316145658493042, 'train/grad_norm_pre_clip': 12.273216247558594, 'timing/data': 0.0049847320187836885, 'timing/model': 1.2087239620741457, 'learning_rate/qwen_vl_interface': 3.7739443515976898e-06, 'learning_rate/action_model': 3.7739443515976894e-05, 'epoch': 1.25}) +[2026-06-17 14:08:16,103][starVLA.training.train_starvla][INFO] - Step 4245, Loss: {'train/loss': 0.648906946182251, 'train/grad_norm_pre_clip': 12.548215866088867, 'timing/data': 0.005266685038805008, 'timing/model': 1.2791334229987115, 'learning_rate/qwen_vl_interface': 3.7718888842216822e-06, 'learning_rate/action_model': 3.771888884221682e-05, 'epoch': 1.25}) +[2026-06-17 14:08:20,768][starVLA.training.train_starvla][INFO] - Step 4246, Loss: {'train/loss': 0.575139045715332, 'train/grad_norm_pre_clip': 13.268072128295898, 'timing/data': 0.004907543072476983, 'timing/model': 1.2063665289897472, 'learning_rate/qwen_vl_interface': 3.7698337232597002e-06, 'learning_rate/action_model': 3.7698337232597e-05, 'epoch': 1.25}) +[2026-06-17 14:08:25,541][starVLA.training.train_starvla][INFO] - Step 4247, Loss: {'train/loss': 0.7854111790657043, 'train/grad_norm_pre_clip': 15.659319877624512, 'timing/data': 0.005138425040058792, 'timing/model': 1.2065453409450129, 'learning_rate/qwen_vl_interface': 3.7677788691377814e-06, 'learning_rate/action_model': 3.7677788691377817e-05, 'epoch': 1.25}) +[2026-06-17 14:08:30,336][starVLA.training.train_starvla][INFO] - Step 4248, Loss: {'train/loss': 0.7024925947189331, 'train/grad_norm_pre_clip': 11.800244331359863, 'timing/data': 0.005165875889360905, 'timing/model': 1.2064168769866228, 'learning_rate/qwen_vl_interface': 3.7657243222818978e-06, 'learning_rate/action_model': 3.7657243222818976e-05, 'epoch': 1.25}) +[2026-06-17 14:08:35,133][starVLA.training.train_starvla][INFO] - Step 4249, Loss: {'train/loss': 0.6418921947479248, 'train/grad_norm_pre_clip': 14.026224136352539, 'timing/data': 0.005336471018381417, 'timing/model': 1.2042013989994302, 'learning_rate/qwen_vl_interface': 3.7636700831179575e-06, 'learning_rate/action_model': 3.763670083117957e-05, 'epoch': 1.25}) +[2026-06-17 14:08:39,922][starVLA.training.train_starvla][INFO] - Step 4250, Loss: {'train/loss': 0.7573894262313843, 'train/grad_norm_pre_clip': 14.641165733337402, 'timing/data': 0.005071577965281904, 'timing/model': 1.3331591619644314, 'learning_rate/qwen_vl_interface': 3.7616161520718096e-06, 'learning_rate/action_model': 3.76161615207181e-05, 'epoch': 1.25}) +[2026-06-17 14:08:44,668][starVLA.training.train_starvla][INFO] - Step 4251, Loss: {'train/loss': 0.4958503842353821, 'train/grad_norm_pre_clip': 11.036201477050781, 'timing/data': 0.0047215300146490335, 'timing/model': 1.2866210129577667, 'learning_rate/qwen_vl_interface': 3.7595625295692352e-06, 'learning_rate/action_model': 3.759562529569235e-05, 'epoch': 1.25}) +[2026-06-17 14:08:49,330][starVLA.training.train_starvla][INFO] - Step 4252, Loss: {'train/loss': 0.9275078177452087, 'train/grad_norm_pre_clip': 17.92486572265625, 'timing/data': 0.005114280967973173, 'timing/model': 1.2065995659213513, 'learning_rate/qwen_vl_interface': 3.7575092160359523e-06, 'learning_rate/action_model': 3.757509216035952e-05, 'epoch': 1.25}) +[2026-06-17 14:08:54,166][starVLA.training.train_starvla][INFO] - Step 4253, Loss: {'train/loss': 0.579034686088562, 'train/grad_norm_pre_clip': 13.617291450500488, 'timing/data': 0.0050700280116871, 'timing/model': 1.2053689999738708, 'learning_rate/qwen_vl_interface': 3.7554562118976135e-06, 'learning_rate/action_model': 3.7554562118976136e-05, 'epoch': 1.25}) +[2026-06-17 14:08:58,988][starVLA.training.train_starvla][INFO] - Step 4254, Loss: {'train/loss': 0.5866127014160156, 'train/grad_norm_pre_clip': 11.570741653442383, 'timing/data': 0.004890155047178268, 'timing/model': 1.2059518119785935, 'learning_rate/qwen_vl_interface': 3.7534035175798113e-06, 'learning_rate/action_model': 3.753403517579811e-05, 'epoch': 1.25}) +[2026-06-17 14:09:03,777][starVLA.training.train_starvla][INFO] - Step 4255, Loss: {'train/loss': 0.6845596432685852, 'train/grad_norm_pre_clip': 13.491212844848633, 'timing/data': 0.005000120960175991, 'timing/model': 1.2038501281058416, 'learning_rate/qwen_vl_interface': 3.7513511335080706e-06, 'learning_rate/action_model': 3.75135113350807e-05, 'epoch': 1.25}) +[2026-06-17 14:09:08,553][starVLA.training.train_starvla][INFO] - Step 4256, Loss: {'train/loss': 0.6855288743972778, 'train/grad_norm_pre_clip': 12.607026100158691, 'timing/data': 0.00492054398637265, 'timing/model': 1.2045222199521959, 'learning_rate/qwen_vl_interface': 3.74929906010785e-06, 'learning_rate/action_model': 3.74929906010785e-05, 'epoch': 1.25}) +[2026-06-17 14:09:13,334][starVLA.training.train_starvla][INFO] - Step 4257, Loss: {'train/loss': 0.5401166677474976, 'train/grad_norm_pre_clip': 12.111119270324707, 'timing/data': 0.005189524032175541, 'timing/model': 1.2908625899581239, 'learning_rate/qwen_vl_interface': 3.747247297804553e-06, 'learning_rate/action_model': 3.7472472978045526e-05, 'epoch': 1.25}) +[2026-06-17 14:09:17,994][starVLA.training.train_starvla][INFO] - Step 4258, Loss: {'train/loss': 0.6181709170341492, 'train/grad_norm_pre_clip': 13.748878479003906, 'timing/data': 0.0049719029339030385, 'timing/model': 1.2046215240843594, 'learning_rate/qwen_vl_interface': 3.745195847023506e-06, 'learning_rate/action_model': 3.745195847023506e-05, 'epoch': 1.25}) +[2026-06-17 14:09:22,775][starVLA.training.train_starvla][INFO] - Step 4259, Loss: {'train/loss': 0.5646377205848694, 'train/grad_norm_pre_clip': 10.292390823364258, 'timing/data': 0.004928437992930412, 'timing/model': 1.2066192959900945, 'learning_rate/qwen_vl_interface': 3.7431447081899795e-06, 'learning_rate/action_model': 3.7431447081899796e-05, 'epoch': 1.25}) +[2026-06-17 14:09:27,589][starVLA.training.train_starvla][INFO] - Step 4260, Loss: {'train/loss': 0.6159343719482422, 'train/grad_norm_pre_clip': 11.155179977416992, 'timing/data': 0.00548290996812284, 'timing/model': 1.2094868560088798, 'learning_rate/qwen_vl_interface': 3.7410938817291753e-06, 'learning_rate/action_model': 3.741093881729175e-05, 'epoch': 1.25}) +[2026-06-17 14:09:32,403][starVLA.training.train_starvla][INFO] - Step 4261, Loss: {'train/loss': 0.8399531245231628, 'train/grad_norm_pre_clip': 14.630501747131348, 'timing/data': 0.005182785098440945, 'timing/model': 1.209940596949309, 'learning_rate/qwen_vl_interface': 3.739043368066234e-06, 'learning_rate/action_model': 3.739043368066234e-05, 'epoch': 1.25}) +[2026-06-17 14:09:37,205][starVLA.training.train_starvla][INFO] - Step 4262, Loss: {'train/loss': 0.6129597425460815, 'train/grad_norm_pre_clip': 13.071069717407227, 'timing/data': 0.005325139034539461, 'timing/model': 1.3366476870141923, 'learning_rate/qwen_vl_interface': 3.7369931676262304e-06, 'learning_rate/action_model': 3.7369931676262304e-05, 'epoch': 1.25}) +[2026-06-17 14:09:41,863][starVLA.training.train_starvla][INFO] - Step 4263, Loss: {'train/loss': 0.6571581363677979, 'train/grad_norm_pre_clip': 11.113934516906738, 'timing/data': 0.005021099001169205, 'timing/model': 1.2029742250451818, 'learning_rate/qwen_vl_interface': 3.7349432808341673e-06, 'learning_rate/action_model': 3.7349432808341674e-05, 'epoch': 1.25}) +[2026-06-17 14:09:46,677][starVLA.training.train_starvla][INFO] - Step 4264, Loss: {'train/loss': 0.7118794918060303, 'train/grad_norm_pre_clip': 14.07833194732666, 'timing/data': 0.0047742059687152505, 'timing/model': 1.217468820977956, 'learning_rate/qwen_vl_interface': 3.7328937081149952e-06, 'learning_rate/action_model': 3.732893708114995e-05, 'epoch': 1.25}) +[2026-06-17 14:09:51,485][starVLA.training.train_starvla][INFO] - Step 4265, Loss: {'train/loss': 0.5876412391662598, 'train/grad_norm_pre_clip': 11.822831153869629, 'timing/data': 0.004928609007038176, 'timing/model': 1.2051755799911916, 'learning_rate/qwen_vl_interface': 3.7308444498935894e-06, 'learning_rate/action_model': 3.730844449893589e-05, 'epoch': 1.25}) +[2026-06-17 14:09:56,200][starVLA.training.train_starvla][INFO] - Step 4266, Loss: {'train/loss': 0.6602152585983276, 'train/grad_norm_pre_clip': 13.069058418273926, 'timing/data': 0.005074329092167318, 'timing/model': 1.2065383029403165, 'learning_rate/qwen_vl_interface': 3.7287955065947634e-06, 'learning_rate/action_model': 3.728795506594763e-05, 'epoch': 1.26}) +[2026-06-17 14:10:00,945][starVLA.training.train_starvla][INFO] - Step 4267, Loss: {'train/loss': 0.6407433152198792, 'train/grad_norm_pre_clip': 13.356828689575195, 'timing/data': 0.0050995140336453915, 'timing/model': 1.2059097720775753, 'learning_rate/qwen_vl_interface': 3.7267468786432655e-06, 'learning_rate/action_model': 3.7267468786432654e-05, 'epoch': 1.26}) +[2026-06-17 14:10:05,740][starVLA.training.train_starvla][INFO] - Step 4268, Loss: {'train/loss': 0.5751661658287048, 'train/grad_norm_pre_clip': 22.32889747619629, 'timing/data': 0.005215291050262749, 'timing/model': 1.3344841520301998, 'learning_rate/qwen_vl_interface': 3.7246985664637787e-06, 'learning_rate/action_model': 3.724698566463778e-05, 'epoch': 1.26}) +[2026-06-17 14:10:10,544][starVLA.training.train_starvla][INFO] - Step 4269, Loss: {'train/loss': 0.5766422748565674, 'train/grad_norm_pre_clip': 11.124133110046387, 'timing/data': 0.005112905986607075, 'timing/model': 1.2051915730116889, 'learning_rate/qwen_vl_interface': 3.722650570480922e-06, 'learning_rate/action_model': 3.722650570480922e-05, 'epoch': 1.26}) +[2026-06-17 14:10:15,343][starVLA.training.train_starvla][INFO] - Step 4270, Loss: {'train/loss': 0.6773332953453064, 'train/grad_norm_pre_clip': 12.298800468444824, 'timing/data': 0.005113933002576232, 'timing/model': 1.2047223179833964, 'learning_rate/qwen_vl_interface': 3.7206028911192445e-06, 'learning_rate/action_model': 3.720602891119244e-05, 'epoch': 1.26}) +[2026-06-17 14:10:20,165][starVLA.training.train_starvla][INFO] - Step 4271, Loss: {'train/loss': 0.8244041204452515, 'train/grad_norm_pre_clip': 13.316987037658691, 'timing/data': 0.005713046062737703, 'timing/model': 1.210216002073139, 'learning_rate/qwen_vl_interface': 3.7185555288032315e-06, 'learning_rate/action_model': 3.7185555288032316e-05, 'epoch': 1.26}) +[2026-06-17 14:10:24,980][starVLA.training.train_starvla][INFO] - Step 4272, Loss: {'train/loss': 0.6588155031204224, 'train/grad_norm_pre_clip': 13.027739524841309, 'timing/data': 0.004977092030458152, 'timing/model': 1.208851268980652, 'learning_rate/qwen_vl_interface': 3.716508483957308e-06, 'learning_rate/action_model': 3.716508483957308e-05, 'epoch': 1.26}) +[2026-06-17 14:10:29,800][starVLA.training.train_starvla][INFO] - Step 4273, Loss: {'train/loss': 0.5709421634674072, 'train/grad_norm_pre_clip': 11.528812408447266, 'timing/data': 0.004822536022402346, 'timing/model': 1.210851515061222, 'learning_rate/qwen_vl_interface': 3.7144617570058245e-06, 'learning_rate/action_model': 3.7144617570058245e-05, 'epoch': 1.26}) +[2026-06-17 14:10:34,571][starVLA.training.train_starvla][INFO] - Step 4274, Loss: {'train/loss': 0.5874821543693542, 'train/grad_norm_pre_clip': 21.672693252563477, 'timing/data': 0.005019405041821301, 'timing/model': 1.317692628945224, 'learning_rate/qwen_vl_interface': 3.7124153483730706e-06, 'learning_rate/action_model': 3.712415348373071e-05, 'epoch': 1.26}) +[2026-06-17 14:10:39,227][starVLA.training.train_starvla][INFO] - Step 4275, Loss: {'train/loss': 0.7533921599388123, 'train/grad_norm_pre_clip': 12.924647331237793, 'timing/data': 0.004519053967669606, 'timing/model': 1.2068398819537833, 'learning_rate/qwen_vl_interface': 3.71036925848327e-06, 'learning_rate/action_model': 3.71036925848327e-05, 'epoch': 1.26}) +[2026-06-17 14:10:44,016][starVLA.training.train_starvla][INFO] - Step 4276, Loss: {'train/loss': 0.613747239112854, 'train/grad_norm_pre_clip': 14.979700088500977, 'timing/data': 0.005153161007910967, 'timing/model': 1.2086457379627973, 'learning_rate/qwen_vl_interface': 3.7083234877605804e-06, 'learning_rate/action_model': 3.7083234877605806e-05, 'epoch': 1.26}) +[2026-06-17 14:10:48,820][starVLA.training.train_starvla][INFO] - Step 4277, Loss: {'train/loss': 0.6008015871047974, 'train/grad_norm_pre_clip': 12.087250709533691, 'timing/data': 0.00522191496565938, 'timing/model': 1.20399919396732, 'learning_rate/qwen_vl_interface': 3.70627803662909e-06, 'learning_rate/action_model': 3.70627803662909e-05, 'epoch': 1.26}) +[2026-06-17 14:10:53,639][starVLA.training.train_starvla][INFO] - Step 4278, Loss: {'train/loss': 0.6690002083778381, 'train/grad_norm_pre_clip': 12.767033576965332, 'timing/data': 0.005364213022403419, 'timing/model': 1.2051385899540037, 'learning_rate/qwen_vl_interface': 3.7042329055128216e-06, 'learning_rate/action_model': 3.704232905512821e-05, 'epoch': 1.26}) +[2026-06-17 14:10:58,445][starVLA.training.train_starvla][INFO] - Step 4279, Loss: {'train/loss': 0.6171635985374451, 'train/grad_norm_pre_clip': 11.608988761901855, 'timing/data': 0.005217350902967155, 'timing/model': 1.3410713829798624, 'learning_rate/qwen_vl_interface': 3.7021880948357375e-06, 'learning_rate/action_model': 3.702188094835737e-05, 'epoch': 1.26}) +[2026-06-17 14:11:03,104][starVLA.training.train_starvla][INFO] - Step 4280, Loss: {'train/loss': 0.5453274250030518, 'train/grad_norm_pre_clip': 11.820353507995605, 'timing/data': 0.004983250983059406, 'timing/model': 1.2058639049064368, 'learning_rate/qwen_vl_interface': 3.700143605021726e-06, 'learning_rate/action_model': 3.7001436050217256e-05, 'epoch': 1.26}) +[2026-06-17 14:11:07,956][starVLA.training.train_starvla][INFO] - Step 4281, Loss: {'train/loss': 0.9531487822532654, 'train/grad_norm_pre_clip': 16.59539794921875, 'timing/data': 0.005038139992393553, 'timing/model': 1.2137245060876012, 'learning_rate/qwen_vl_interface': 3.6980994364946133e-06, 'learning_rate/action_model': 3.698099436494613e-05, 'epoch': 1.26}) +[2026-06-17 14:11:12,661][starVLA.training.train_starvla][INFO] - Step 4282, Loss: {'train/loss': 0.49648338556289673, 'train/grad_norm_pre_clip': 8.843103408813477, 'timing/data': 0.005085732904262841, 'timing/model': 1.2083187849493697, 'learning_rate/qwen_vl_interface': 3.696055589678156e-06, 'learning_rate/action_model': 3.6960555896781554e-05, 'epoch': 1.26}) +[2026-06-17 14:11:17,413][starVLA.training.train_starvla][INFO] - Step 4283, Loss: {'train/loss': 0.5804242491722107, 'train/grad_norm_pre_clip': 13.93443775177002, 'timing/data': 0.004664710024371743, 'timing/model': 1.2039098099339753, 'learning_rate/qwen_vl_interface': 3.694012064996049e-06, 'learning_rate/action_model': 3.694012064996049e-05, 'epoch': 1.26}) +[2026-06-17 14:11:22,207][starVLA.training.train_starvla][INFO] - Step 4284, Loss: {'train/loss': 0.8038284182548523, 'train/grad_norm_pre_clip': 12.872404098510742, 'timing/data': 0.004934650962240994, 'timing/model': 1.2081966290716082, 'learning_rate/qwen_vl_interface': 3.6919688628719153e-06, 'learning_rate/action_model': 3.691968862871915e-05, 'epoch': 1.26}) +[2026-06-17 14:11:27,009][starVLA.training.train_starvla][INFO] - Step 4285, Loss: {'train/loss': 0.8935927748680115, 'train/grad_norm_pre_clip': 15.485102653503418, 'timing/data': 0.005055734072811902, 'timing/model': 1.3367530109826475, 'learning_rate/qwen_vl_interface': 3.689925983729311e-06, 'learning_rate/action_model': 3.689925983729311e-05, 'epoch': 1.26}) +[2026-06-17 14:11:31,673][starVLA.training.train_starvla][INFO] - Step 4286, Loss: {'train/loss': 0.7242373824119568, 'train/grad_norm_pre_clip': 14.18124771118164, 'timing/data': 0.0050033219158649445, 'timing/model': 1.2080822049174458, 'learning_rate/qwen_vl_interface': 3.68788342799173e-06, 'learning_rate/action_model': 3.6878834279917294e-05, 'epoch': 1.26}) +[2026-06-17 14:11:36,478][starVLA.training.train_starvla][INFO] - Step 4287, Loss: {'train/loss': 0.7958900332450867, 'train/grad_norm_pre_clip': 14.610694885253906, 'timing/data': 0.004997391952201724, 'timing/model': 1.2094536270014942, 'learning_rate/qwen_vl_interface': 3.6858411960825953e-06, 'learning_rate/action_model': 3.685841196082595e-05, 'epoch': 1.26}) +[2026-06-17 14:11:41,260][starVLA.training.train_starvla][INFO] - Step 4288, Loss: {'train/loss': 0.7968991994857788, 'train/grad_norm_pre_clip': 12.32957935333252, 'timing/data': 0.005214307922869921, 'timing/model': 1.208425689022988, 'learning_rate/qwen_vl_interface': 3.6837992884252636e-06, 'learning_rate/action_model': 3.683799288425264e-05, 'epoch': 1.26}) +[2026-06-17 14:11:46,008][starVLA.training.train_starvla][INFO] - Step 4289, Loss: {'train/loss': 0.7178418636322021, 'train/grad_norm_pre_clip': 13.40537166595459, 'timing/data': 0.005432208999991417, 'timing/model': 1.2032377980649471, 'learning_rate/qwen_vl_interface': 3.6817577054430243e-06, 'learning_rate/action_model': 3.6817577054430244e-05, 'epoch': 1.26}) +[2026-06-17 14:11:50,790][starVLA.training.train_starvla][INFO] - Step 4290, Loss: {'train/loss': 0.5335628390312195, 'train/grad_norm_pre_clip': 11.842193603515625, 'timing/data': 0.00561427092179656, 'timing/model': 1.2085290539544076, 'learning_rate/qwen_vl_interface': 3.6797164475591003e-06, 'learning_rate/action_model': 3.6797164475591e-05, 'epoch': 1.26}) +[2026-06-17 14:11:55,585][starVLA.training.train_starvla][INFO] - Step 4291, Loss: {'train/loss': 0.8293331265449524, 'train/grad_norm_pre_clip': 13.925284385681152, 'timing/data': 0.004884939990006387, 'timing/model': 1.3423917279578745, 'learning_rate/qwen_vl_interface': 3.677675515196648e-06, 'learning_rate/action_model': 3.677675515196648e-05, 'epoch': 1.26}) +[2026-06-17 14:12:00,244][starVLA.training.train_starvla][INFO] - Step 4292, Loss: {'train/loss': 0.8569890260696411, 'train/grad_norm_pre_clip': 16.944501876831055, 'timing/data': 0.00502651697024703, 'timing/model': 1.2052101930603385, 'learning_rate/qwen_vl_interface': 3.6756349087787524e-06, 'learning_rate/action_model': 3.6756349087787524e-05, 'epoch': 1.26}) +[2026-06-17 14:12:05,034][starVLA.training.train_starvla][INFO] - Step 4293, Loss: {'train/loss': 0.9225172400474548, 'train/grad_norm_pre_clip': 12.925477027893066, 'timing/data': 0.0052780890837311745, 'timing/model': 1.2056523120263591, 'learning_rate/qwen_vl_interface': 3.6735946287284325e-06, 'learning_rate/action_model': 3.6735946287284324e-05, 'epoch': 1.26}) +[2026-06-17 14:12:09,820][starVLA.training.train_starvla][INFO] - Step 4294, Loss: {'train/loss': 0.5780870318412781, 'train/grad_norm_pre_clip': 11.43004322052002, 'timing/data': 0.00492107798345387, 'timing/model': 1.2065822910517454, 'learning_rate/qwen_vl_interface': 3.6715546754686444e-06, 'learning_rate/action_model': 3.6715546754686444e-05, 'epoch': 1.26}) +[2026-06-17 14:12:14,616][starVLA.training.train_starvla][INFO] - Step 4295, Loss: {'train/loss': 0.7830568552017212, 'train/grad_norm_pre_clip': 13.543933868408203, 'timing/data': 0.005212164018303156, 'timing/model': 1.2055590179516003, 'learning_rate/qwen_vl_interface': 3.6695150494222697e-06, 'learning_rate/action_model': 3.6695150494222695e-05, 'epoch': 1.26}) +[2026-06-17 14:12:19,305][starVLA.training.train_starvla][INFO] - Step 4296, Loss: {'train/loss': 0.599392831325531, 'train/grad_norm_pre_clip': 11.966547012329102, 'timing/data': 0.004993284004740417, 'timing/model': 1.2047660630196333, 'learning_rate/qwen_vl_interface': 3.6674757510121255e-06, 'learning_rate/action_model': 3.6674757510121255e-05, 'epoch': 1.26}) +[2026-06-17 14:12:24,019][starVLA.training.train_starvla][INFO] - Step 4297, Loss: {'train/loss': 1.0420584678649902, 'train/grad_norm_pre_clip': 15.12190055847168, 'timing/data': 0.00511046196334064, 'timing/model': 1.2462749250698835, 'learning_rate/qwen_vl_interface': 3.6654367806609615e-06, 'learning_rate/action_model': 3.665436780660961e-05, 'epoch': 1.26}) +[2026-06-17 14:12:28,681][starVLA.training.train_starvla][INFO] - Step 4298, Loss: {'train/loss': 0.5866208672523499, 'train/grad_norm_pre_clip': 12.412125587463379, 'timing/data': 0.005231766030192375, 'timing/model': 1.207381310989149, 'learning_rate/qwen_vl_interface': 3.6633981387914597e-06, 'learning_rate/action_model': 3.66339813879146e-05, 'epoch': 1.26}) +[2026-06-17 14:12:33,461][starVLA.training.train_starvla][INFO] - Step 4299, Loss: {'train/loss': 0.508059561252594, 'train/grad_norm_pre_clip': 8.120000839233398, 'timing/data': 0.005070455954410136, 'timing/model': 1.2126812760252506, 'learning_rate/qwen_vl_interface': 3.6613598258262295e-06, 'learning_rate/action_model': 3.6613598258262296e-05, 'epoch': 1.26}) +[2026-06-17 14:14:14,786][starVLA.training.train_starvla][INFO] - Step 4300, Loss: {'train/loss': 0.6565752029418945, 'train/grad_norm_pre_clip': 12.340651512145996, 'eval/loss': 0.7393378615379333, 'eval/latency_0/loss': 0.7393378615379333, 'eval/demon_fire_precision': 0.8830567348514087, 'eval/demon_fire_recall': 0.9501661129568106, 'eval/demon_fire_f1': 0.915383076615323, 'eval/demon_fire_rate_model': 0.7205228031145717, 'eval/demon_fire_rate_teacher': 0.6696329254727474, 'eval/demon_left_precision': 0.827708703374778, 'eval/demon_left_recall': 0.7920679886685552, 'eval/demon_left_f1': 0.8094962362478286, 'eval/demon_left_rate_model': 0.4696885428253615, 'eval/demon_left_rate_teacher': 0.49082313681868744, 'eval/demon_right_precision': 0.8, 'eval/demon_right_recall': 0.847457627118644, 'eval/demon_right_f1': 0.8230452674897119, 'eval/demon_right_rate_model': 0.521412680756396, 'eval/demon_right_rate_teacher': 0.4922135706340378, 'eval/demon_move_f1': 0.8162707518687702, 'eval/demon_cc_f1': 0.883210995901264, 'eval/latency_0/demon_fire_precision': 0.8830567348514087, 'eval/latency_0/demon_fire_recall': 0.9501661129568106, 'eval/latency_0/demon_fire_f1': 0.915383076615323, 'eval/latency_0/demon_fire_rate_model': 0.7205228031145717, 'eval/latency_0/demon_fire_rate_teacher': 0.6696329254727474, 'eval/latency_0/demon_left_precision': 0.827708703374778, 'eval/latency_0/demon_left_recall': 0.7920679886685552, 'eval/latency_0/demon_left_f1': 0.8094962362478286, 'eval/latency_0/demon_left_rate_model': 0.4696885428253615, 'eval/latency_0/demon_left_rate_teacher': 0.49082313681868744, 'eval/latency_0/demon_right_precision': 0.8, 'eval/latency_0/demon_right_recall': 0.847457627118644, 'eval/latency_0/demon_right_f1': 0.8230452674897119, 'eval/latency_0/demon_right_rate_model': 0.521412680756396, 'eval/latency_0/demon_right_rate_teacher': 0.4922135706340378, 'eval/latency_0/demon_move_f1': 0.8162707518687702, 'eval/latency_0/demon_cc_f1': 0.883210995901264, 'timing/data': 0.005185127025470138, 'timing/model': 1.2087014509597793, 'learning_rate/qwen_vl_interface': 3.6593218421878157e-06, 'learning_rate/action_model': 3.659321842187816e-05, 'epoch': 1.27}) +[2026-06-17 14:14:14,787][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_4300_state +[2026-06-17 14:14:14,790][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 14:14:18,356][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_4300_state/model.safetensors +[2026-06-17 14:14:25,958][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_4300_state/optimizer.bin +[2026-06-17 14:14:25,959][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_4300_state/sampler.bin +[2026-06-17 14:14:25,959][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_4300_state/sampler_1.bin +[2026-06-17 14:14:25,960][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_4300_state/random_states_0.pkl +[2026-06-17 14:14:26,013][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4300 +[2026-06-17 14:14:26,014][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 14:14:26,027][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 14:14:30,629][starVLA.training.train_starvla][INFO] - Step 4301, Loss: {'train/loss': 0.5388387441635132, 'train/grad_norm_pre_clip': 11.709403991699219, 'timing/data': 0.005256169009953737, 'timing/model': 1.185298442025669, 'learning_rate/qwen_vl_interface': 3.657284188298697e-06, 'learning_rate/action_model': 3.657284188298697e-05, 'epoch': 1.27}) +[2026-06-17 14:14:35,325][starVLA.training.train_starvla][INFO] - Step 4302, Loss: {'train/loss': 0.6264952421188354, 'train/grad_norm_pre_clip': 12.814644813537598, 'timing/data': 0.00545589707326144, 'timing/model': 1.190278523019515, 'learning_rate/qwen_vl_interface': 3.65524686458128e-06, 'learning_rate/action_model': 3.65524686458128e-05, 'epoch': 1.27}) +[2026-06-17 14:14:40,056][starVLA.training.train_starvla][INFO] - Step 4303, Loss: {'train/loss': 0.9710775017738342, 'train/grad_norm_pre_clip': 19.216032028198242, 'timing/data': 0.005896197049878538, 'timing/model': 1.1896724760299549, 'learning_rate/qwen_vl_interface': 3.653209871457904e-06, 'learning_rate/action_model': 3.653209871457904e-05, 'epoch': 1.27}) +[2026-06-17 14:14:44,832][starVLA.training.train_starvla][INFO] - Step 4304, Loss: {'train/loss': 0.4655264914035797, 'train/grad_norm_pre_clip': 10.18414306640625, 'timing/data': 0.005424931994639337, 'timing/model': 1.3638081909157336, 'learning_rate/qwen_vl_interface': 3.6511732093508376e-06, 'learning_rate/action_model': 3.651173209350838e-05, 'epoch': 1.27}) +[2026-06-17 14:14:49,442][starVLA.training.train_starvla][INFO] - Step 4305, Loss: {'train/loss': 0.9624999165534973, 'train/grad_norm_pre_clip': 23.543611526489258, 'timing/data': 0.0053958260687068105, 'timing/model': 1.1991592720150948, 'learning_rate/qwen_vl_interface': 3.6491368786822885e-06, 'learning_rate/action_model': 3.6491368786822884e-05, 'epoch': 1.27}) +[2026-06-17 14:14:54,186][starVLA.training.train_starvla][INFO] - Step 4306, Loss: {'train/loss': 0.5801810026168823, 'train/grad_norm_pre_clip': 10.229848861694336, 'timing/data': 0.005517682060599327, 'timing/model': 1.194183196988888, 'learning_rate/qwen_vl_interface': 3.647100879874386e-06, 'learning_rate/action_model': 3.6471008798743855e-05, 'epoch': 1.27}) +[2026-06-17 14:14:58,962][starVLA.training.train_starvla][INFO] - Step 4307, Loss: {'train/loss': 0.6306770443916321, 'train/grad_norm_pre_clip': 13.444089889526367, 'timing/data': 0.005680438014678657, 'timing/model': 1.1934453919529915, 'learning_rate/qwen_vl_interface': 3.645065213349195e-06, 'learning_rate/action_model': 3.645065213349195e-05, 'epoch': 1.27}) +[2026-06-17 14:15:03,726][starVLA.training.train_starvla][INFO] - Step 4308, Loss: {'train/loss': 0.6601443886756897, 'train/grad_norm_pre_clip': 10.845375061035156, 'timing/data': 0.005456151906400919, 'timing/model': 1.2030287310481071, 'learning_rate/qwen_vl_interface': 3.6430298795287135e-06, 'learning_rate/action_model': 3.643029879528713e-05, 'epoch': 1.27}) +[2026-06-17 14:15:08,487][starVLA.training.train_starvla][INFO] - Step 4309, Loss: {'train/loss': 0.8604608774185181, 'train/grad_norm_pre_clip': 13.64078426361084, 'timing/data': 0.005630586063489318, 'timing/model': 1.1963080810382962, 'learning_rate/qwen_vl_interface': 3.640994878834866e-06, 'learning_rate/action_model': 3.640994878834866e-05, 'epoch': 1.27}) +[2026-06-17 14:15:13,243][starVLA.training.train_starvla][INFO] - Step 4310, Loss: {'train/loss': 0.834673285484314, 'train/grad_norm_pre_clip': 14.254733085632324, 'timing/data': 0.005657288013026118, 'timing/model': 1.3279012360144407, 'learning_rate/qwen_vl_interface': 3.6389602116895136e-06, 'learning_rate/action_model': 3.6389602116895134e-05, 'epoch': 1.27}) +[2026-06-17 14:15:18,008][starVLA.training.train_starvla][INFO] - Step 4311, Loss: {'train/loss': 0.519110381603241, 'train/grad_norm_pre_clip': 10.886866569519043, 'timing/data': 0.005551169975660741, 'timing/model': 1.330679752980359, 'learning_rate/qwen_vl_interface': 3.6369258785144395e-06, 'learning_rate/action_model': 3.636925878514439e-05, 'epoch': 1.27}) +[2026-06-17 14:15:22,644][starVLA.training.train_starvla][INFO] - Step 4312, Loss: {'train/loss': 0.6144120693206787, 'train/grad_norm_pre_clip': 14.423735618591309, 'timing/data': 0.005573117989115417, 'timing/model': 1.1988587389932945, 'learning_rate/qwen_vl_interface': 3.6348918797313693e-06, 'learning_rate/action_model': 3.634891879731369e-05, 'epoch': 1.27}) +[2026-06-17 14:15:27,377][starVLA.training.train_starvla][INFO] - Step 4313, Loss: {'train/loss': 0.6920351386070251, 'train/grad_norm_pre_clip': 12.13498306274414, 'timing/data': 0.00580817402806133, 'timing/model': 1.1998759009875357, 'learning_rate/qwen_vl_interface': 3.6328582157619495e-06, 'learning_rate/action_model': 3.6328582157619496e-05, 'epoch': 1.27}) +[2026-06-17 14:15:32,143][starVLA.training.train_starvla][INFO] - Step 4314, Loss: {'train/loss': 0.5725816488265991, 'train/grad_norm_pre_clip': 11.27543830871582, 'timing/data': 0.005751930992119014, 'timing/model': 1.1986561439698562, 'learning_rate/qwen_vl_interface': 3.6308248870277613e-06, 'learning_rate/action_model': 3.630824887027761e-05, 'epoch': 1.27}) +[2026-06-17 14:15:36,913][starVLA.training.train_starvla][INFO] - Step 4315, Loss: {'train/loss': 0.8397159576416016, 'train/grad_norm_pre_clip': 13.06171703338623, 'timing/data': 0.005714408005587757, 'timing/model': 1.1989972100127488, 'learning_rate/qwen_vl_interface': 3.628791893950316e-06, 'learning_rate/action_model': 3.628791893950316e-05, 'epoch': 1.27}) +[2026-06-17 14:15:41,647][starVLA.training.train_starvla][INFO] - Step 4316, Loss: {'train/loss': 0.48560383915901184, 'train/grad_norm_pre_clip': 16.048091888427734, 'timing/data': 0.005628463928587735, 'timing/model': 1.269711607019417, 'learning_rate/qwen_vl_interface': 3.6267592369510563e-06, 'learning_rate/action_model': 3.6267592369510564e-05, 'epoch': 1.27}) +[2026-06-17 14:15:46,284][starVLA.training.train_starvla][INFO] - Step 4317, Loss: {'train/loss': 0.6648954153060913, 'train/grad_norm_pre_clip': 16.190963745117188, 'timing/data': 0.005409237928688526, 'timing/model': 1.199435752001591, 'learning_rate/qwen_vl_interface': 3.624726916451355e-06, 'learning_rate/action_model': 3.624726916451355e-05, 'epoch': 1.27}) +[2026-06-17 14:15:51,024][starVLA.training.train_starvla][INFO] - Step 4318, Loss: {'train/loss': 0.8409918546676636, 'train/grad_norm_pre_clip': 13.715941429138184, 'timing/data': 0.0052529009990394115, 'timing/model': 1.2009334670146927, 'learning_rate/qwen_vl_interface': 3.622694932872509e-06, 'learning_rate/action_model': 3.622694932872509e-05, 'epoch': 1.27}) +[2026-06-17 14:15:55,797][starVLA.training.train_starvla][INFO] - Step 4319, Loss: {'train/loss': 0.5410807728767395, 'train/grad_norm_pre_clip': 9.33147144317627, 'timing/data': 0.005525929038412869, 'timing/model': 1.2029064449016005, 'learning_rate/qwen_vl_interface': 3.620663286635758e-06, 'learning_rate/action_model': 3.620663286635758e-05, 'epoch': 1.27}) +[2026-06-17 14:16:00,572][starVLA.training.train_starvla][INFO] - Step 4320, Loss: {'train/loss': 0.4939548075199127, 'train/grad_norm_pre_clip': 9.321538925170898, 'timing/data': 0.005675693973898888, 'timing/model': 1.200202026986517, 'learning_rate/qwen_vl_interface': 3.618631978162259e-06, 'learning_rate/action_model': 3.618631978162259e-05, 'epoch': 1.27}) +[2026-06-17 14:16:05,338][starVLA.training.train_starvla][INFO] - Step 4321, Loss: {'train/loss': 0.6651875376701355, 'train/grad_norm_pre_clip': 18.03365707397461, 'timing/data': 0.005752834957093, 'timing/model': 1.3247877539834008, 'learning_rate/qwen_vl_interface': 3.6166010078731073e-06, 'learning_rate/action_model': 3.6166010078731075e-05, 'epoch': 1.27}) +[2026-06-17 14:16:10,071][starVLA.training.train_starvla][INFO] - Step 4322, Loss: {'train/loss': 0.8911844491958618, 'train/grad_norm_pre_clip': 13.295415878295898, 'timing/data': 0.005457040970213711, 'timing/model': 1.287542380974628, 'learning_rate/qwen_vl_interface': 3.614570376189323e-06, 'learning_rate/action_model': 3.614570376189323e-05, 'epoch': 1.27}) +[2026-06-17 14:16:14,717][starVLA.training.train_starvla][INFO] - Step 4323, Loss: {'train/loss': 0.9655458927154541, 'train/grad_norm_pre_clip': 15.2493896484375, 'timing/data': 0.005738563020713627, 'timing/model': 1.2025288039585575, 'learning_rate/qwen_vl_interface': 3.6125400835318604e-06, 'learning_rate/action_model': 3.61254008353186e-05, 'epoch': 1.27}) +[2026-06-17 14:16:19,499][starVLA.training.train_starvla][INFO] - Step 4324, Loss: {'train/loss': 0.5604082345962524, 'train/grad_norm_pre_clip': 10.208751678466797, 'timing/data': 0.005400891997851431, 'timing/model': 1.2009044429287314, 'learning_rate/qwen_vl_interface': 3.6105101303216016e-06, 'learning_rate/action_model': 3.610510130321601e-05, 'epoch': 1.27}) +[2026-06-17 14:16:24,265][starVLA.training.train_starvla][INFO] - Step 4325, Loss: {'train/loss': 0.6204873919487, 'train/grad_norm_pre_clip': 10.790762901306152, 'timing/data': 0.0054118880070745945, 'timing/model': 1.2018462710548192, 'learning_rate/qwen_vl_interface': 3.608480516979354e-06, 'learning_rate/action_model': 3.608480516979354e-05, 'epoch': 1.27}) +[2026-06-17 14:16:29,055][starVLA.training.train_starvla][INFO] - Step 4326, Loss: {'train/loss': 0.6592051386833191, 'train/grad_norm_pre_clip': 12.927813529968262, 'timing/data': 0.005327640916220844, 'timing/model': 1.2023633199278265, 'learning_rate/qwen_vl_interface': 3.6064512439258646e-06, 'learning_rate/action_model': 3.606451243925865e-05, 'epoch': 1.27}) +[2026-06-17 14:16:33,829][starVLA.training.train_starvla][INFO] - Step 4327, Loss: {'train/loss': 0.5396024584770203, 'train/grad_norm_pre_clip': 10.250568389892578, 'timing/data': 0.005995532963424921, 'timing/model': 1.1989005360519513, 'learning_rate/qwen_vl_interface': 3.6044223115818e-06, 'learning_rate/action_model': 3.6044223115818e-05, 'epoch': 1.27}) +[2026-06-17 14:16:38,580][starVLA.training.train_starvla][INFO] - Step 4328, Loss: {'train/loss': 0.583453357219696, 'train/grad_norm_pre_clip': 12.077195167541504, 'timing/data': 0.005246848100796342, 'timing/model': 1.2950988169759512, 'learning_rate/qwen_vl_interface': 3.6023937203677606e-06, 'learning_rate/action_model': 3.6023937203677604e-05, 'epoch': 1.27}) +[2026-06-17 14:16:43,230][starVLA.training.train_starvla][INFO] - Step 4329, Loss: {'train/loss': 0.774461567401886, 'train/grad_norm_pre_clip': 13.541797637939453, 'timing/data': 0.005274883005768061, 'timing/model': 1.2020032180007547, 'learning_rate/qwen_vl_interface': 3.6003654707042746e-06, 'learning_rate/action_model': 3.6003654707042745e-05, 'epoch': 1.27}) +[2026-06-17 14:16:48,010][starVLA.training.train_starvla][INFO] - Step 4330, Loss: {'train/loss': 0.5224358439445496, 'train/grad_norm_pre_clip': 9.820626258850098, 'timing/data': 0.005331555032171309, 'timing/model': 1.2040242119692266, 'learning_rate/qwen_vl_interface': 3.5983375630118033e-06, 'learning_rate/action_model': 3.598337563011803e-05, 'epoch': 1.27}) +[2026-06-17 14:16:52,788][starVLA.training.train_starvla][INFO] - Step 4331, Loss: {'train/loss': 0.6591408252716064, 'train/grad_norm_pre_clip': 12.02743148803711, 'timing/data': 0.00550432491581887, 'timing/model': 1.2044013449922204, 'learning_rate/qwen_vl_interface': 3.596309997710733e-06, 'learning_rate/action_model': 3.596309997710733e-05, 'epoch': 1.27}) +[2026-06-17 14:16:57,600][starVLA.training.train_starvla][INFO] - Step 4332, Loss: {'train/loss': 0.6348794102668762, 'train/grad_norm_pre_clip': 12.430041313171387, 'timing/data': 0.011319182929582894, 'timing/model': 1.2093858030857518, 'learning_rate/qwen_vl_interface': 3.5942827752213788e-06, 'learning_rate/action_model': 3.594282775221379e-05, 'epoch': 1.27}) +[2026-06-17 14:17:02,386][starVLA.training.train_starvla][INFO] - Step 4333, Loss: {'train/loss': 0.8034157156944275, 'train/grad_norm_pre_clip': 13.149262428283691, 'timing/data': 0.00587271002586931, 'timing/model': 1.3344319630414248, 'learning_rate/qwen_vl_interface': 3.592255895963986e-06, 'learning_rate/action_model': 3.592255895963986e-05, 'epoch': 1.27}) +[2026-06-17 14:17:07,132][starVLA.training.train_starvla][INFO] - Step 4334, Loss: {'train/loss': 0.4496209919452667, 'train/grad_norm_pre_clip': 10.234457969665527, 'timing/data': 0.005637944908812642, 'timing/model': 1.2117206930415705, 'learning_rate/qwen_vl_interface': 3.590229360358731e-06, 'learning_rate/action_model': 3.590229360358731e-05, 'epoch': 1.28}) +[2026-06-17 14:17:11,933][starVLA.training.train_starvla][INFO] - Step 4335, Loss: {'train/loss': 0.7162929177284241, 'train/grad_norm_pre_clip': 14.072708129882812, 'timing/data': 0.006046848022378981, 'timing/model': 1.2039471310563385, 'learning_rate/qwen_vl_interface': 3.5882031688257147e-06, 'learning_rate/action_model': 3.588203168825715e-05, 'epoch': 1.28}) +[2026-06-17 14:17:16,742][starVLA.training.train_starvla][INFO] - Step 4336, Loss: {'train/loss': 0.8470115065574646, 'train/grad_norm_pre_clip': 19.57053565979004, 'timing/data': 0.005320207914337516, 'timing/model': 1.2096162020461634, 'learning_rate/qwen_vl_interface': 3.5861773217849694e-06, 'learning_rate/action_model': 3.5861773217849693e-05, 'epoch': 1.28}) +[2026-06-17 14:17:21,474][starVLA.training.train_starvla][INFO] - Step 4337, Loss: {'train/loss': 0.4743332862854004, 'train/grad_norm_pre_clip': 9.597526550292969, 'timing/data': 0.005764122004620731, 'timing/model': 1.2107553359819576, 'learning_rate/qwen_vl_interface': 3.5841518196564566e-06, 'learning_rate/action_model': 3.584151819656457e-05, 'epoch': 1.28}) +[2026-06-17 14:17:26,222][starVLA.training.train_starvla][INFO] - Step 4338, Loss: {'train/loss': 0.6685577630996704, 'train/grad_norm_pre_clip': 13.440743446350098, 'timing/data': 0.005540474900044501, 'timing/model': 1.21228096797131, 'learning_rate/qwen_vl_interface': 3.582126662860065e-06, 'learning_rate/action_model': 3.582126662860065e-05, 'epoch': 1.28}) +[2026-06-17 14:17:31,043][starVLA.training.train_starvla][INFO] - Step 4339, Loss: {'train/loss': 0.6943354606628418, 'train/grad_norm_pre_clip': 13.239070892333984, 'timing/data': 0.005498645012266934, 'timing/model': 1.3435893080895767, 'learning_rate/qwen_vl_interface': 3.5801018518156096e-06, 'learning_rate/action_model': 3.5801018518156096e-05, 'epoch': 1.28}) +[2026-06-17 14:17:35,751][starVLA.training.train_starvla][INFO] - Step 4340, Loss: {'train/loss': 0.7433318495750427, 'train/grad_norm_pre_clip': 14.208494186401367, 'timing/data': 0.007857104064896703, 'timing/model': 1.2245144419139251, 'learning_rate/qwen_vl_interface': 3.5780773869428366e-06, 'learning_rate/action_model': 3.5780773869428364e-05, 'epoch': 1.28}) +[2026-06-17 14:17:40,567][starVLA.training.train_starvla][INFO] - Step 4341, Loss: {'train/loss': 0.7966837286949158, 'train/grad_norm_pre_clip': 14.785246849060059, 'timing/data': 0.005364360986277461, 'timing/model': 1.2046478090342134, 'learning_rate/qwen_vl_interface': 3.576053268661422e-06, 'learning_rate/action_model': 3.5760532686614216e-05, 'epoch': 1.28}) +[2026-06-17 14:17:45,362][starVLA.training.train_starvla][INFO] - Step 4342, Loss: {'train/loss': 0.6584234833717346, 'train/grad_norm_pre_clip': 14.44705867767334, 'timing/data': 0.005583953927271068, 'timing/model': 1.205344317946583, 'learning_rate/qwen_vl_interface': 3.5740294973909643e-06, 'learning_rate/action_model': 3.5740294973909644e-05, 'epoch': 1.28}) +[2026-06-17 14:17:50,189][starVLA.training.train_starvla][INFO] - Step 4343, Loss: {'train/loss': 0.8720748424530029, 'train/grad_norm_pre_clip': 14.737818717956543, 'timing/data': 0.005859968019649386, 'timing/model': 1.211457666940987, 'learning_rate/qwen_vl_interface': 3.5720060735509976e-06, 'learning_rate/action_model': 3.5720060735509975e-05, 'epoch': 1.28}) +[2026-06-17 14:17:54,942][starVLA.training.train_starvla][INFO] - Step 4344, Loss: {'train/loss': 0.6228815317153931, 'train/grad_norm_pre_clip': 10.64908504486084, 'timing/data': 0.005609332001768053, 'timing/model': 1.203038362087682, 'learning_rate/qwen_vl_interface': 3.569982997560974e-06, 'learning_rate/action_model': 3.5699829975609736e-05, 'epoch': 1.28}) +[2026-06-17 14:17:59,719][starVLA.training.train_starvla][INFO] - Step 4345, Loss: {'train/loss': 0.6459529399871826, 'train/grad_norm_pre_clip': 13.522552490234375, 'timing/data': 0.005470223957672715, 'timing/model': 1.3267561179818586, 'learning_rate/qwen_vl_interface': 3.567960269840286e-06, 'learning_rate/action_model': 3.567960269840286e-05, 'epoch': 1.28}) +[2026-06-17 14:18:04,392][starVLA.training.train_starvla][INFO] - Step 4346, Loss: {'train/loss': 0.6882131099700928, 'train/grad_norm_pre_clip': 13.657485961914062, 'timing/data': 0.005636769928969443, 'timing/model': 1.21528347407002, 'learning_rate/qwen_vl_interface': 3.5659378908082427e-06, 'learning_rate/action_model': 3.5659378908082426e-05, 'epoch': 1.28}) +[2026-06-17 14:18:09,219][starVLA.training.train_starvla][INFO] - Step 4347, Loss: {'train/loss': 0.6816443800926208, 'train/grad_norm_pre_clip': 13.8164644241333, 'timing/data': 0.005462670931592584, 'timing/model': 1.2134214319521561, 'learning_rate/qwen_vl_interface': 3.5639158608840862e-06, 'learning_rate/action_model': 3.563915860884086e-05, 'epoch': 1.28}) +[2026-06-17 14:18:14,021][starVLA.training.train_starvla][INFO] - Step 4348, Loss: {'train/loss': 0.4902978241443634, 'train/grad_norm_pre_clip': 12.782722473144531, 'timing/data': 0.005858627031557262, 'timing/model': 1.2041145029943436, 'learning_rate/qwen_vl_interface': 3.5618941804869866e-06, 'learning_rate/action_model': 3.561894180486987e-05, 'epoch': 1.28}) +[2026-06-17 14:18:18,807][starVLA.training.train_starvla][INFO] - Step 4349, Loss: {'train/loss': 0.5395292043685913, 'train/grad_norm_pre_clip': 14.94473934173584, 'timing/data': 0.005625012097880244, 'timing/model': 1.2039912219624966, 'learning_rate/qwen_vl_interface': 3.5598728500360395e-06, 'learning_rate/action_model': 3.55987285003604e-05, 'epoch': 1.28}) +[2026-06-17 14:18:23,595][starVLA.training.train_starvla][INFO] - Step 4350, Loss: {'train/loss': 0.5859668850898743, 'train/grad_norm_pre_clip': 16.338483810424805, 'timing/data': 0.00534970301669091, 'timing/model': 1.340232138056308, 'learning_rate/qwen_vl_interface': 3.557851869950272e-06, 'learning_rate/action_model': 3.557851869950272e-05, 'epoch': 1.28}) +[2026-06-17 14:18:28,267][starVLA.training.train_starvla][INFO] - Step 4351, Loss: {'train/loss': 0.47451406717300415, 'train/grad_norm_pre_clip': 8.905632972717285, 'timing/data': 0.0051819500513374805, 'timing/model': 1.20489734003786, 'learning_rate/qwen_vl_interface': 3.555831240648628e-06, 'learning_rate/action_model': 3.5558312406486274e-05, 'epoch': 1.28}) +[2026-06-17 14:18:33,060][starVLA.training.train_starvla][INFO] - Step 4352, Loss: {'train/loss': 0.8211297392845154, 'train/grad_norm_pre_clip': 13.785930633544922, 'timing/data': 0.0054128889460116625, 'timing/model': 1.203521162038669, 'learning_rate/qwen_vl_interface': 3.553810962549994e-06, 'learning_rate/action_model': 3.553810962549994e-05, 'epoch': 1.28}) +[2026-06-17 14:18:37,767][starVLA.training.train_starvla][INFO] - Step 4353, Loss: {'train/loss': 0.777373194694519, 'train/grad_norm_pre_clip': 18.000446319580078, 'timing/data': 0.006120883976109326, 'timing/model': 1.2053816999541596, 'learning_rate/qwen_vl_interface': 3.5517910360731723e-06, 'learning_rate/action_model': 3.551791036073172e-05, 'epoch': 1.28}) +[2026-06-17 14:18:42,507][starVLA.training.train_starvla][INFO] - Step 4354, Loss: {'train/loss': 0.5924144387245178, 'train/grad_norm_pre_clip': 10.759742736816406, 'timing/data': 0.0055981200421229005, 'timing/model': 1.2053011369425803, 'learning_rate/qwen_vl_interface': 3.549771461636896e-06, 'learning_rate/action_model': 3.549771461636896e-05, 'epoch': 1.28}) +[2026-06-17 14:18:47,299][starVLA.training.train_starvla][INFO] - Step 4355, Loss: {'train/loss': 0.801796019077301, 'train/grad_norm_pre_clip': 12.60363483428955, 'timing/data': 0.00560335093177855, 'timing/model': 1.2150170980021358, 'learning_rate/qwen_vl_interface': 3.5477522396598236e-06, 'learning_rate/action_model': 3.5477522396598235e-05, 'epoch': 1.28}) +[2026-06-17 14:18:52,110][starVLA.training.train_starvla][INFO] - Step 4356, Loss: {'train/loss': 0.5532627105712891, 'train/grad_norm_pre_clip': 12.533917427062988, 'timing/data': 0.005241189035587013, 'timing/model': 1.33561602095142, 'learning_rate/qwen_vl_interface': 3.5457333705605456e-06, 'learning_rate/action_model': 3.5457333705605456e-05, 'epoch': 1.28}) +[2026-06-17 14:18:56,777][starVLA.training.train_starvla][INFO] - Step 4357, Loss: {'train/loss': 0.718213677406311, 'train/grad_norm_pre_clip': 12.586076736450195, 'timing/data': 0.005660504917614162, 'timing/model': 1.2071731499163434, 'learning_rate/qwen_vl_interface': 3.543714854757574e-06, 'learning_rate/action_model': 3.5437148547575736e-05, 'epoch': 1.28}) +[2026-06-17 14:19:01,581][starVLA.training.train_starvla][INFO] - Step 4358, Loss: {'train/loss': 1.169238567352295, 'train/grad_norm_pre_clip': 15.556434631347656, 'timing/data': 0.0051827579736709595, 'timing/model': 1.2031327009899542, 'learning_rate/qwen_vl_interface': 3.5416966926693464e-06, 'learning_rate/action_model': 3.5416966926693465e-05, 'epoch': 1.28}) +[2026-06-17 14:19:06,367][starVLA.training.train_starvla][INFO] - Step 4359, Loss: {'train/loss': 0.8127992153167725, 'train/grad_norm_pre_clip': 16.153242111206055, 'timing/data': 0.005489965085871518, 'timing/model': 1.2045056120259687, 'learning_rate/qwen_vl_interface': 3.539678884714235e-06, 'learning_rate/action_model': 3.539678884714235e-05, 'epoch': 1.28}) +[2026-06-17 14:19:11,143][starVLA.training.train_starvla][INFO] - Step 4360, Loss: {'train/loss': 0.9771230816841125, 'train/grad_norm_pre_clip': 15.25210952758789, 'timing/data': 0.005866260966286063, 'timing/model': 1.2046962849562988, 'learning_rate/qwen_vl_interface': 3.5376614313105307e-06, 'learning_rate/action_model': 3.53766143131053e-05, 'epoch': 1.28}) +[2026-06-17 14:19:15,948][starVLA.training.train_starvla][INFO] - Step 4361, Loss: {'train/loss': 0.5207509994506836, 'train/grad_norm_pre_clip': 11.49813461303711, 'timing/data': 0.005642283940687776, 'timing/model': 1.2030690829269588, 'learning_rate/qwen_vl_interface': 3.5356443328764546e-06, 'learning_rate/action_model': 3.5356443328764546e-05, 'epoch': 1.28}) +[2026-06-17 14:19:20,753][starVLA.training.train_starvla][INFO] - Step 4362, Loss: {'train/loss': 0.7177619338035583, 'train/grad_norm_pre_clip': 11.257182121276855, 'timing/data': 0.005512341973371804, 'timing/model': 1.3404526700032875, 'learning_rate/qwen_vl_interface': 3.5336275898301515e-06, 'learning_rate/action_model': 3.5336275898301514e-05, 'epoch': 1.28}) +[2026-06-17 14:19:25,428][starVLA.training.train_starvla][INFO] - Step 4363, Loss: {'train/loss': 0.604672372341156, 'train/grad_norm_pre_clip': 11.058002471923828, 'timing/data': 0.005809845984913409, 'timing/model': 1.2097074270714074, 'learning_rate/qwen_vl_interface': 3.531611202589697e-06, 'learning_rate/action_model': 3.531611202589697e-05, 'epoch': 1.28}) +[2026-06-17 14:19:30,218][starVLA.training.train_starvla][INFO] - Step 4364, Loss: {'train/loss': 0.7983815670013428, 'train/grad_norm_pre_clip': 14.036515235900879, 'timing/data': 0.006667915964499116, 'timing/model': 1.2032650710316375, 'learning_rate/qwen_vl_interface': 3.5295951715730902e-06, 'learning_rate/action_model': 3.52959517157309e-05, 'epoch': 1.28}) +[2026-06-17 14:19:35,007][starVLA.training.train_starvla][INFO] - Step 4365, Loss: {'train/loss': 0.4655228853225708, 'train/grad_norm_pre_clip': 10.925787925720215, 'timing/data': 0.005320898024365306, 'timing/model': 1.2041700129630044, 'learning_rate/qwen_vl_interface': 3.527579497198255e-06, 'learning_rate/action_model': 3.527579497198255e-05, 'epoch': 1.28}) +[2026-06-17 14:19:39,799][starVLA.training.train_starvla][INFO] - Step 4366, Loss: {'train/loss': 0.5611622929573059, 'train/grad_norm_pre_clip': 10.353509902954102, 'timing/data': 0.0055987419327721, 'timing/model': 1.2056075559230521, 'learning_rate/qwen_vl_interface': 3.525564179883042e-06, 'learning_rate/action_model': 3.525564179883042e-05, 'epoch': 1.28}) +[2026-06-17 14:19:44,470][starVLA.training.train_starvla][INFO] - Step 4367, Loss: {'train/loss': 0.5636031627655029, 'train/grad_norm_pre_clip': 11.963589668273926, 'timing/data': 0.005663169082254171, 'timing/model': 1.203679436002858, 'learning_rate/qwen_vl_interface': 3.5235492200452308e-06, 'learning_rate/action_model': 3.523549220045231e-05, 'epoch': 1.29}) +[2026-06-17 14:19:49,159][starVLA.training.train_starvla][INFO] - Step 4368, Loss: {'train/loss': 0.4682253301143646, 'train/grad_norm_pre_clip': 10.696102142333984, 'timing/data': 0.0055190869607031345, 'timing/model': 1.2384221170796081, 'learning_rate/qwen_vl_interface': 3.521534618102525e-06, 'learning_rate/action_model': 3.5215346181025246e-05, 'epoch': 1.29}) +[2026-06-17 14:19:53,858][starVLA.training.train_starvla][INFO] - Step 4369, Loss: {'train/loss': 0.7950053215026855, 'train/grad_norm_pre_clip': 12.893167495727539, 'timing/data': 0.005431305966340005, 'timing/model': 1.2080093390541151, 'learning_rate/qwen_vl_interface': 3.5195203744725516e-06, 'learning_rate/action_model': 3.5195203744725516e-05, 'epoch': 1.29}) +[2026-06-17 14:19:58,594][starVLA.training.train_starvla][INFO] - Step 4370, Loss: {'train/loss': 0.7363744378089905, 'train/grad_norm_pre_clip': 12.385331153869629, 'timing/data': 0.005894806003198028, 'timing/model': 1.2044441839680076, 'learning_rate/qwen_vl_interface': 3.5175064895728665e-06, 'learning_rate/action_model': 3.5175064895728664e-05, 'epoch': 1.29}) +[2026-06-17 14:20:03,412][starVLA.training.train_starvla][INFO] - Step 4371, Loss: {'train/loss': 0.9425976276397705, 'train/grad_norm_pre_clip': 14.320481300354004, 'timing/data': 0.005715135019272566, 'timing/model': 1.2037575480062515, 'learning_rate/qwen_vl_interface': 3.515492963820951e-06, 'learning_rate/action_model': 3.515492963820951e-05, 'epoch': 1.29}) +[2026-06-17 14:20:08,208][starVLA.training.train_starvla][INFO] - Step 4372, Loss: {'train/loss': 0.6751003265380859, 'train/grad_norm_pre_clip': 13.446296691894531, 'timing/data': 0.005488571943715215, 'timing/model': 1.2057042949600145, 'learning_rate/qwen_vl_interface': 3.5134797976342123e-06, 'learning_rate/action_model': 3.513479797634212e-05, 'epoch': 1.29}) +[2026-06-17 14:20:12,991][starVLA.training.train_starvla][INFO] - Step 4373, Loss: {'train/loss': 0.6053142547607422, 'train/grad_norm_pre_clip': 11.119640350341797, 'timing/data': 0.005581301054917276, 'timing/model': 1.3318982800701633, 'learning_rate/qwen_vl_interface': 3.5114669914299776e-06, 'learning_rate/action_model': 3.511466991429978e-05, 'epoch': 1.29}) +[2026-06-17 14:20:17,696][starVLA.training.train_starvla][INFO] - Step 4374, Loss: {'train/loss': 0.774428129196167, 'train/grad_norm_pre_clip': 16.791322708129883, 'timing/data': 0.005569271044805646, 'timing/model': 1.249948794953525, 'learning_rate/qwen_vl_interface': 3.5094545456255065e-06, 'learning_rate/action_model': 3.509454545625506e-05, 'epoch': 1.29}) +[2026-06-17 14:20:22,359][starVLA.training.train_starvla][INFO] - Step 4375, Loss: {'train/loss': 0.4265870451927185, 'train/grad_norm_pre_clip': 35.92558288574219, 'timing/data': 0.005497051985003054, 'timing/model': 1.2055688539985567, 'learning_rate/qwen_vl_interface': 3.5074424606379826e-06, 'learning_rate/action_model': 3.5074424606379824e-05, 'epoch': 1.29}) +[2026-06-17 14:20:27,111][starVLA.training.train_starvla][INFO] - Step 4376, Loss: {'train/loss': 0.6230833530426025, 'train/grad_norm_pre_clip': 11.193825721740723, 'timing/data': 0.005075276945717633, 'timing/model': 1.2111971250269562, 'learning_rate/qwen_vl_interface': 3.505430736884511e-06, 'learning_rate/action_model': 3.5054307368845105e-05, 'epoch': 1.29}) +[2026-06-17 14:20:31,932][starVLA.training.train_starvla][INFO] - Step 4377, Loss: {'train/loss': 0.8378978371620178, 'train/grad_norm_pre_clip': 13.85262393951416, 'timing/data': 0.005306449020281434, 'timing/model': 1.2130783629836515, 'learning_rate/qwen_vl_interface': 3.503419374782123e-06, 'learning_rate/action_model': 3.503419374782123e-05, 'epoch': 1.29}) +[2026-06-17 14:20:36,726][starVLA.training.train_starvla][INFO] - Step 4378, Loss: {'train/loss': 0.5203390717506409, 'train/grad_norm_pre_clip': 12.74630069732666, 'timing/data': 0.005569330998696387, 'timing/model': 1.2061373989563435, 'learning_rate/qwen_vl_interface': 3.50140837474778e-06, 'learning_rate/action_model': 3.50140837474778e-05, 'epoch': 1.29}) +[2026-06-17 14:20:41,550][starVLA.training.train_starvla][INFO] - Step 4379, Loss: {'train/loss': 0.36097264289855957, 'train/grad_norm_pre_clip': 9.519181251525879, 'timing/data': 0.005128895048983395, 'timing/model': 1.330636905040592, 'learning_rate/qwen_vl_interface': 3.4993977371983625e-06, 'learning_rate/action_model': 3.499397737198362e-05, 'epoch': 1.29}) +[2026-06-17 14:20:46,323][starVLA.training.train_starvla][INFO] - Step 4380, Loss: {'train/loss': 0.7022565007209778, 'train/grad_norm_pre_clip': 12.225528717041016, 'timing/data': 0.0055968910455703735, 'timing/model': 1.3240271989488974, 'learning_rate/qwen_vl_interface': 3.497387462550676e-06, 'learning_rate/action_model': 3.4973874625506757e-05, 'epoch': 1.29}) +[2026-06-17 14:20:50,988][starVLA.training.train_starvla][INFO] - Step 4381, Loss: {'train/loss': 0.7031258940696716, 'train/grad_norm_pre_clip': 13.127098083496094, 'timing/data': 0.006098748999647796, 'timing/model': 1.2067998120328411, 'learning_rate/qwen_vl_interface': 3.4953775512214568e-06, 'learning_rate/action_model': 3.495377551221457e-05, 'epoch': 1.29}) +[2026-06-17 14:20:55,747][starVLA.training.train_starvla][INFO] - Step 4382, Loss: {'train/loss': 0.7032714486122131, 'train/grad_norm_pre_clip': 12.767492294311523, 'timing/data': 0.005790296010673046, 'timing/model': 1.2066734629916027, 'learning_rate/qwen_vl_interface': 3.4933680036273583e-06, 'learning_rate/action_model': 3.493368003627358e-05, 'epoch': 1.29}) +[2026-06-17 14:21:00,530][starVLA.training.train_starvla][INFO] - Step 4383, Loss: {'train/loss': 0.8740681409835815, 'train/grad_norm_pre_clip': 14.381545066833496, 'timing/data': 0.005718639004044235, 'timing/model': 1.2047169730067253, 'learning_rate/qwen_vl_interface': 3.491358820184963e-06, 'learning_rate/action_model': 3.491358820184963e-05, 'epoch': 1.29}) +[2026-06-17 14:21:05,316][starVLA.training.train_starvla][INFO] - Step 4384, Loss: {'train/loss': 0.7002675533294678, 'train/grad_norm_pre_clip': 11.692230224609375, 'timing/data': 0.005331685999408364, 'timing/model': 1.2054154919460416, 'learning_rate/qwen_vl_interface': 3.4893500013107758e-06, 'learning_rate/action_model': 3.4893500013107754e-05, 'epoch': 1.29}) +[2026-06-17 14:21:10,071][starVLA.training.train_starvla][INFO] - Step 4385, Loss: {'train/loss': 0.6818127632141113, 'train/grad_norm_pre_clip': 13.093067169189453, 'timing/data': 0.0054590420331805944, 'timing/model': 1.2966097519965842, 'learning_rate/qwen_vl_interface': 3.4873415474212287e-06, 'learning_rate/action_model': 3.487341547421228e-05, 'epoch': 1.29}) +[2026-06-17 14:21:14,728][starVLA.training.train_starvla][INFO] - Step 4386, Loss: {'train/loss': 1.059672236442566, 'train/grad_norm_pre_clip': 16.28825569152832, 'timing/data': 0.005595283000729978, 'timing/model': 1.2042137760436162, 'learning_rate/qwen_vl_interface': 3.4853334589326764e-06, 'learning_rate/action_model': 3.4853334589326765e-05, 'epoch': 1.29}) +[2026-06-17 14:21:19,503][starVLA.training.train_starvla][INFO] - Step 4387, Loss: {'train/loss': 0.6864242553710938, 'train/grad_norm_pre_clip': 14.344093322753906, 'timing/data': 0.005449217977002263, 'timing/model': 1.211014482891187, 'learning_rate/qwen_vl_interface': 3.483325736261394e-06, 'learning_rate/action_model': 3.483325736261394e-05, 'epoch': 1.29}) +[2026-06-17 14:21:24,316][starVLA.training.train_starvla][INFO] - Step 4388, Loss: {'train/loss': 0.5456722974777222, 'train/grad_norm_pre_clip': 27.054218292236328, 'timing/data': 0.005559467943385243, 'timing/model': 1.2125683049671352, 'learning_rate/qwen_vl_interface': 3.48131837982359e-06, 'learning_rate/action_model': 3.48131837982359e-05, 'epoch': 1.29}) +[2026-06-17 14:21:29,135][starVLA.training.train_starvla][INFO] - Step 4389, Loss: {'train/loss': 0.6268993020057678, 'train/grad_norm_pre_clip': 10.211159706115723, 'timing/data': 0.005723491893149912, 'timing/model': 1.209463614039123, 'learning_rate/qwen_vl_interface': 3.4793113900353885e-06, 'learning_rate/action_model': 3.479311390035389e-05, 'epoch': 1.29}) +[2026-06-17 14:21:33,934][starVLA.training.train_starvla][INFO] - Step 4390, Loss: {'train/loss': 0.6375622153282166, 'train/grad_norm_pre_clip': 12.603299140930176, 'timing/data': 0.00556436600163579, 'timing/model': 1.334361401037313, 'learning_rate/qwen_vl_interface': 3.4773047673128412e-06, 'learning_rate/action_model': 3.477304767312841e-05, 'epoch': 1.29}) +[2026-06-17 14:21:38,677][starVLA.training.train_starvla][INFO] - Step 4391, Loss: {'train/loss': 0.7498275637626648, 'train/grad_norm_pre_clip': 15.452824592590332, 'timing/data': 0.005329687031917274, 'timing/model': 1.285710294963792, 'learning_rate/qwen_vl_interface': 3.4752985120719225e-06, 'learning_rate/action_model': 3.475298512071922e-05, 'epoch': 1.29}) +[2026-06-17 14:21:43,348][starVLA.training.train_starvla][INFO] - Step 4392, Loss: {'train/loss': 0.6647123098373413, 'train/grad_norm_pre_clip': 10.742241859436035, 'timing/data': 0.006226927042007446, 'timing/model': 1.2056434369878843, 'learning_rate/qwen_vl_interface': 3.4732926247285324e-06, 'learning_rate/action_model': 3.4732926247285325e-05, 'epoch': 1.29}) +[2026-06-17 14:21:48,245][starVLA.training.train_starvla][INFO] - Step 4393, Loss: {'train/loss': 0.6413009762763977, 'train/grad_norm_pre_clip': 16.6754093170166, 'timing/data': 0.005286199972033501, 'timing/model': 1.2117243109969422, 'learning_rate/qwen_vl_interface': 3.471287105698494e-06, 'learning_rate/action_model': 3.471287105698494e-05, 'epoch': 1.29}) +[2026-06-17 14:21:53,046][starVLA.training.train_starvla][INFO] - Step 4394, Loss: {'train/loss': 0.8609187006950378, 'train/grad_norm_pre_clip': 14.501763343811035, 'timing/data': 0.006091534974984825, 'timing/model': 1.2070423508994281, 'learning_rate/qwen_vl_interface': 3.4692819553975514e-06, 'learning_rate/action_model': 3.469281955397551e-05, 'epoch': 1.29}) +[2026-06-17 14:21:57,837][starVLA.training.train_starvla][INFO] - Step 4395, Loss: {'train/loss': 1.0674188137054443, 'train/grad_norm_pre_clip': 14.222960472106934, 'timing/data': 0.005347311031073332, 'timing/model': 1.2036050300812349, 'learning_rate/qwen_vl_interface': 3.4672771742413754e-06, 'learning_rate/action_model': 3.467277174241375e-05, 'epoch': 1.29}) +[2026-06-17 14:22:02,617][starVLA.training.train_starvla][INFO] - Step 4396, Loss: {'train/loss': 0.42850616574287415, 'train/grad_norm_pre_clip': 11.008262634277344, 'timing/data': 0.005377635010518134, 'timing/model': 1.204490982927382, 'learning_rate/qwen_vl_interface': 3.46527276264556e-06, 'learning_rate/action_model': 3.46527276264556e-05, 'epoch': 1.29}) +[2026-06-17 14:22:07,390][starVLA.training.train_starvla][INFO] - Step 4397, Loss: {'train/loss': 0.4562547504901886, 'train/grad_norm_pre_clip': 9.766653060913086, 'timing/data': 0.005470722913742065, 'timing/model': 1.303052081959322, 'learning_rate/qwen_vl_interface': 3.4632687210256218e-06, 'learning_rate/action_model': 3.463268721025621e-05, 'epoch': 1.29}) +[2026-06-17 14:22:12,068][starVLA.training.train_starvla][INFO] - Step 4398, Loss: {'train/loss': 0.964873194694519, 'train/grad_norm_pre_clip': 14.378213882446289, 'timing/data': 0.005640835966914892, 'timing/model': 1.2033079819520935, 'learning_rate/qwen_vl_interface': 3.461265049796999e-06, 'learning_rate/action_model': 3.461265049796999e-05, 'epoch': 1.29}) +[2026-06-17 14:22:16,873][starVLA.training.train_starvla][INFO] - Step 4399, Loss: {'train/loss': 0.6140317320823669, 'train/grad_norm_pre_clip': 11.645275115966797, 'timing/data': 0.005580118973739445, 'timing/model': 1.208688510931097, 'learning_rate/qwen_vl_interface': 3.459261749375057e-06, 'learning_rate/action_model': 3.4592617493750567e-05, 'epoch': 1.29}) +[2026-06-17 14:23:58,645][starVLA.training.train_starvla][INFO] - Step 4400, Loss: {'train/loss': 0.5842936635017395, 'train/grad_norm_pre_clip': 11.490355491638184, 'eval/loss': 0.7301886677742004, 'eval/latency_0/loss': 0.7301886677742004, 'eval/demon_fire_precision': 0.8866615265998458, 'eval/demon_fire_recall': 0.9551495016611296, 'eval/demon_fire_f1': 0.9196321471411435, 'eval/demon_fire_rate_model': 0.721357063403782, 'eval/demon_fire_rate_teacher': 0.6696329254727474, 'eval/demon_left_precision': 0.8349001175088132, 'eval/demon_left_recall': 0.805099150141643, 'eval/demon_left_f1': 0.8197288722238246, 'eval/demon_left_rate_model': 0.4733036707452725, 'eval/demon_left_rate_teacher': 0.49082313681868744, 'eval/demon_right_precision': 0.8113917248790973, 'eval/demon_right_recall': 0.8531073446327684, 'eval/demon_right_f1': 0.8317267970256128, 'eval/demon_right_rate_model': 0.5175194660734149, 'eval/demon_right_rate_teacher': 0.4922135706340378, 'eval/demon_move_f1': 0.8257278346247188, 'eval/demon_cc_f1': 0.8892922316844442, 'eval/latency_0/demon_fire_precision': 0.8866615265998458, 'eval/latency_0/demon_fire_recall': 0.9551495016611296, 'eval/latency_0/demon_fire_f1': 0.9196321471411435, 'eval/latency_0/demon_fire_rate_model': 0.721357063403782, 'eval/latency_0/demon_fire_rate_teacher': 0.6696329254727474, 'eval/latency_0/demon_left_precision': 0.8349001175088132, 'eval/latency_0/demon_left_recall': 0.805099150141643, 'eval/latency_0/demon_left_f1': 0.8197288722238246, 'eval/latency_0/demon_left_rate_model': 0.4733036707452725, 'eval/latency_0/demon_left_rate_teacher': 0.49082313681868744, 'eval/latency_0/demon_right_precision': 0.8113917248790973, 'eval/latency_0/demon_right_recall': 0.8531073446327684, 'eval/latency_0/demon_right_f1': 0.8317267970256128, 'eval/latency_0/demon_right_rate_model': 0.5175194660734149, 'eval/latency_0/demon_right_rate_teacher': 0.4922135706340378, 'eval/latency_0/demon_move_f1': 0.8257278346247188, 'eval/latency_0/demon_cc_f1': 0.8892922316844442, 'timing/data': 0.005628135986626148, 'timing/model': 1.2044171640882269, 'learning_rate/qwen_vl_interface': 3.457258820175082e-06, 'learning_rate/action_model': 3.457258820175082e-05, 'epoch': 1.29}) +[2026-06-17 14:23:58,645][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_4400_state +[2026-06-17 14:23:58,649][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 14:24:02,213][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_4400_state/model.safetensors +[2026-06-17 14:24:09,796][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_4400_state/optimizer.bin +[2026-06-17 14:24:09,797][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_4400_state/sampler.bin +[2026-06-17 14:24:09,797][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_4400_state/sampler_1.bin +[2026-06-17 14:24:09,797][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_4400_state/random_states_0.pkl +[2026-06-17 14:24:09,846][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4400 +[2026-06-17 14:24:09,846][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 14:24:09,856][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 14:24:14,565][starVLA.training.train_starvla][INFO] - Step 4401, Loss: {'train/loss': 0.5484764575958252, 'train/grad_norm_pre_clip': 12.086503982543945, 'timing/data': 0.004880773019976914, 'timing/model': 1.1899403389543295, 'learning_rate/qwen_vl_interface': 3.4552562626122817e-06, 'learning_rate/action_model': 3.455256262612281e-05, 'epoch': 1.3}) +[2026-06-17 14:24:19,182][starVLA.training.train_starvla][INFO] - Step 4402, Loss: {'train/loss': 0.45997703075408936, 'train/grad_norm_pre_clip': 10.987628936767578, 'timing/data': 0.0048419899540022016, 'timing/model': 1.1895665600895882, 'learning_rate/qwen_vl_interface': 3.453254077101789e-06, 'learning_rate/action_model': 3.453254077101789e-05, 'epoch': 1.3}) +[2026-06-17 14:24:23,822][starVLA.training.train_starvla][INFO] - Step 4403, Loss: {'train/loss': 0.620582640171051, 'train/grad_norm_pre_clip': 11.937010765075684, 'timing/data': 0.005180376931093633, 'timing/model': 1.1898284999188036, 'learning_rate/qwen_vl_interface': 3.45125226405866e-06, 'learning_rate/action_model': 3.4512522640586595e-05, 'epoch': 1.3}) +[2026-06-17 14:24:28,506][starVLA.training.train_starvla][INFO] - Step 4404, Loss: {'train/loss': 0.5390143394470215, 'train/grad_norm_pre_clip': 12.417110443115234, 'timing/data': 0.0048686269437894225, 'timing/model': 1.2686310070566833, 'learning_rate/qwen_vl_interface': 3.449250823897872e-06, 'learning_rate/action_model': 3.449250823897872e-05, 'epoch': 1.3}) +[2026-06-17 14:24:33,113][starVLA.training.train_starvla][INFO] - Step 4405, Loss: {'train/loss': 0.7006584405899048, 'train/grad_norm_pre_clip': 11.726438522338867, 'timing/data': 0.004878874984569848, 'timing/model': 1.1947170669445768, 'learning_rate/qwen_vl_interface': 3.447249757034328e-06, 'learning_rate/action_model': 3.4472497570343276e-05, 'epoch': 1.3}) +[2026-06-17 14:24:37,854][starVLA.training.train_starvla][INFO] - Step 4406, Loss: {'train/loss': 0.694481372833252, 'train/grad_norm_pre_clip': 13.135148048400879, 'timing/data': 0.00525845808442682, 'timing/model': 1.1929254890419543, 'learning_rate/qwen_vl_interface': 3.445249063882845e-06, 'learning_rate/action_model': 3.445249063882845e-05, 'epoch': 1.3}) +[2026-06-17 14:24:42,593][starVLA.training.train_starvla][INFO] - Step 4407, Loss: {'train/loss': 0.8000744581222534, 'train/grad_norm_pre_clip': 12.837530136108398, 'timing/data': 0.0049962790217250586, 'timing/model': 1.1940674009965733, 'learning_rate/qwen_vl_interface': 3.4432487448581762e-06, 'learning_rate/action_model': 3.443248744858176e-05, 'epoch': 1.3}) +[2026-06-17 14:24:47,339][starVLA.training.train_starvla][INFO] - Step 4408, Loss: {'train/loss': 0.9525981545448303, 'train/grad_norm_pre_clip': 18.0106143951416, 'timing/data': 0.00513610802590847, 'timing/model': 1.196278083953075, 'learning_rate/qwen_vl_interface': 3.4412488003749867e-06, 'learning_rate/action_model': 3.4412488003749866e-05, 'epoch': 1.3}) +[2026-06-17 14:24:52,078][starVLA.training.train_starvla][INFO] - Step 4409, Loss: {'train/loss': 0.6485798954963684, 'train/grad_norm_pre_clip': 12.716163635253906, 'timing/data': 0.004878066014498472, 'timing/model': 1.1979074470000342, 'learning_rate/qwen_vl_interface': 3.4392492308478657e-06, 'learning_rate/action_model': 3.439249230847866e-05, 'epoch': 1.3}) +[2026-06-17 14:24:56,791][starVLA.training.train_starvla][INFO] - Step 4410, Loss: {'train/loss': 0.5830524563789368, 'train/grad_norm_pre_clip': 15.724617004394531, 'timing/data': 0.004807665012776852, 'timing/model': 1.2809592669364065, 'learning_rate/qwen_vl_interface': 3.437250036691328e-06, 'learning_rate/action_model': 3.437250036691328e-05, 'epoch': 1.3}) +[2026-06-17 14:25:01,507][starVLA.training.train_starvla][INFO] - Step 4411, Loss: {'train/loss': 0.5303335189819336, 'train/grad_norm_pre_clip': 10.823203086853027, 'timing/data': 0.005099665024317801, 'timing/model': 1.1977857779711485, 'learning_rate/qwen_vl_interface': 3.4352512183198094e-06, 'learning_rate/action_model': 3.4352512183198095e-05, 'epoch': 1.3}) +[2026-06-17 14:25:06,268][starVLA.training.train_starvla][INFO] - Step 4412, Loss: {'train/loss': 0.6260442137718201, 'train/grad_norm_pre_clip': 11.323891639709473, 'timing/data': 0.004854465019889176, 'timing/model': 1.1973378859693184, 'learning_rate/qwen_vl_interface': 3.4332527761476662e-06, 'learning_rate/action_model': 3.433252776147666e-05, 'epoch': 1.3}) +[2026-06-17 14:25:11,041][starVLA.training.train_starvla][INFO] - Step 4413, Loss: {'train/loss': 0.8006574511528015, 'train/grad_norm_pre_clip': 15.888652801513672, 'timing/data': 0.0052777399541810155, 'timing/model': 1.2116547289770097, 'learning_rate/qwen_vl_interface': 3.431254710589174e-06, 'learning_rate/action_model': 3.431254710589174e-05, 'epoch': 1.3}) +[2026-06-17 14:25:15,821][starVLA.training.train_starvla][INFO] - Step 4414, Loss: {'train/loss': 0.6476590633392334, 'train/grad_norm_pre_clip': 13.052237510681152, 'timing/data': 0.005135061917826533, 'timing/model': 1.199487859965302, 'learning_rate/qwen_vl_interface': 3.4292570220585407e-06, 'learning_rate/action_model': 3.4292570220585407e-05, 'epoch': 1.3}) +[2026-06-17 14:25:20,588][starVLA.training.train_starvla][INFO] - Step 4415, Loss: {'train/loss': 0.763850212097168, 'train/grad_norm_pre_clip': 17.070627212524414, 'timing/data': 0.004907620023004711, 'timing/model': 1.3311930379131809, 'learning_rate/qwen_vl_interface': 3.4272597109698857e-06, 'learning_rate/action_model': 3.427259710969886e-05, 'epoch': 1.3}) +[2026-06-17 14:25:25,221][starVLA.training.train_starvla][INFO] - Step 4416, Loss: {'train/loss': 0.6274168491363525, 'train/grad_norm_pre_clip': 12.934562683105469, 'timing/data': 0.00490896194241941, 'timing/model': 1.1994296789634973, 'learning_rate/qwen_vl_interface': 3.4252627777372533e-06, 'learning_rate/action_model': 3.425262777737253e-05, 'epoch': 1.3}) +[2026-06-17 14:25:29,982][starVLA.training.train_starvla][INFO] - Step 4417, Loss: {'train/loss': 1.031009554862976, 'train/grad_norm_pre_clip': 15.601221084594727, 'timing/data': 0.005175295053049922, 'timing/model': 1.1998433419503272, 'learning_rate/qwen_vl_interface': 3.4232662227746095e-06, 'learning_rate/action_model': 3.423266222774609e-05, 'epoch': 1.3}) +[2026-06-17 14:25:34,754][starVLA.training.train_starvla][INFO] - Step 4418, Loss: {'train/loss': 0.5683320164680481, 'train/grad_norm_pre_clip': 11.26763916015625, 'timing/data': 0.0053744850447401404, 'timing/model': 1.2004121370846406, 'learning_rate/qwen_vl_interface': 3.4212700464958448e-06, 'learning_rate/action_model': 3.421270046495845e-05, 'epoch': 1.3}) +[2026-06-17 14:25:39,443][starVLA.training.train_starvla][INFO] - Step 4419, Loss: {'train/loss': 0.6589210629463196, 'train/grad_norm_pre_clip': 15.524765014648438, 'timing/data': 0.004864613991230726, 'timing/model': 1.2001000619493425, 'learning_rate/qwen_vl_interface': 3.4192742493147686e-06, 'learning_rate/action_model': 3.419274249314768e-05, 'epoch': 1.3}) +[2026-06-17 14:25:44,163][starVLA.training.train_starvla][INFO] - Step 4420, Loss: {'train/loss': 0.7740519046783447, 'train/grad_norm_pre_clip': 24.222414016723633, 'timing/data': 0.0050690219504758716, 'timing/model': 1.2003127241041511, 'learning_rate/qwen_vl_interface': 3.4172788316451065e-06, 'learning_rate/action_model': 3.417278831645107e-05, 'epoch': 1.3}) +[2026-06-17 14:25:48,936][starVLA.training.train_starvla][INFO] - Step 4421, Loss: {'train/loss': 0.5230022072792053, 'train/grad_norm_pre_clip': 11.964723587036133, 'timing/data': 0.0050868220860138535, 'timing/model': 1.3326371279545128, 'learning_rate/qwen_vl_interface': 3.4152837939005177e-06, 'learning_rate/action_model': 3.415283793900518e-05, 'epoch': 1.3}) +[2026-06-17 14:25:53,575][starVLA.training.train_starvla][INFO] - Step 4422, Loss: {'train/loss': 0.6967945098876953, 'train/grad_norm_pre_clip': 37.82675552368164, 'timing/data': 0.005216834950260818, 'timing/model': 1.200283696060069, 'learning_rate/qwen_vl_interface': 3.413289136494572e-06, 'learning_rate/action_model': 3.413289136494572e-05, 'epoch': 1.3}) +[2026-06-17 14:25:58,341][starVLA.training.train_starvla][INFO] - Step 4423, Loss: {'train/loss': 0.37291136384010315, 'train/grad_norm_pre_clip': 13.676939964294434, 'timing/data': 0.00523969903588295, 'timing/model': 1.201138133998029, 'learning_rate/qwen_vl_interface': 3.411294859840765e-06, 'learning_rate/action_model': 3.411294859840764e-05, 'epoch': 1.3}) +[2026-06-17 14:26:03,121][starVLA.training.train_starvla][INFO] - Step 4424, Loss: {'train/loss': 0.8457449674606323, 'train/grad_norm_pre_clip': 15.815377235412598, 'timing/data': 0.005329512991011143, 'timing/model': 1.201302947010845, 'learning_rate/qwen_vl_interface': 3.409300964352511e-06, 'learning_rate/action_model': 3.409300964352511e-05, 'epoch': 1.3}) +[2026-06-17 14:26:07,899][starVLA.training.train_starvla][INFO] - Step 4425, Loss: {'train/loss': 0.8268470168113708, 'train/grad_norm_pre_clip': 14.726967811584473, 'timing/data': 0.005245395004749298, 'timing/model': 1.2004172450397164, 'learning_rate/qwen_vl_interface': 3.407307450443148e-06, 'learning_rate/action_model': 3.407307450443148e-05, 'epoch': 1.3}) +[2026-06-17 14:26:12,624][starVLA.training.train_starvla][INFO] - Step 4426, Loss: {'train/loss': 0.7466176748275757, 'train/grad_norm_pre_clip': 14.404012680053711, 'timing/data': 0.004968693945556879, 'timing/model': 1.1997657440369949, 'learning_rate/qwen_vl_interface': 3.4053143185259353e-06, 'learning_rate/action_model': 3.405314318525935e-05, 'epoch': 1.3}) +[2026-06-17 14:26:17,380][starVLA.training.train_starvla][INFO] - Step 4427, Loss: {'train/loss': 0.7849864363670349, 'train/grad_norm_pre_clip': 15.788914680480957, 'timing/data': 0.005041890079155564, 'timing/model': 1.3155132459942251, 'learning_rate/qwen_vl_interface': 3.4033215690140475e-06, 'learning_rate/action_model': 3.4033215690140476e-05, 'epoch': 1.3}) +[2026-06-17 14:26:22,032][starVLA.training.train_starvla][INFO] - Step 4428, Loss: {'train/loss': 0.4026549160480499, 'train/grad_norm_pre_clip': 9.813122749328613, 'timing/data': 0.0050512689631432295, 'timing/model': 1.2092606259975582, 'learning_rate/qwen_vl_interface': 3.4013292023205852e-06, 'learning_rate/action_model': 3.401329202320585e-05, 'epoch': 1.3}) +[2026-06-17 14:26:26,839][starVLA.training.train_starvla][INFO] - Step 4429, Loss: {'train/loss': 0.9244198799133301, 'train/grad_norm_pre_clip': 14.271754264831543, 'timing/data': 0.005054099019616842, 'timing/model': 1.2076750580454245, 'learning_rate/qwen_vl_interface': 3.3993372188585694e-06, 'learning_rate/action_model': 3.3993372188585694e-05, 'epoch': 1.3}) +[2026-06-17 14:26:31,642][starVLA.training.train_starvla][INFO] - Step 4430, Loss: {'train/loss': 0.5920405387878418, 'train/grad_norm_pre_clip': 10.92243480682373, 'timing/data': 0.004834786988794804, 'timing/model': 1.209362223977223, 'learning_rate/qwen_vl_interface': 3.3973456190409397e-06, 'learning_rate/action_model': 3.39734561904094e-05, 'epoch': 1.3}) +[2026-06-17 14:26:36,457][starVLA.training.train_starvla][INFO] - Step 4431, Loss: {'train/loss': 0.45718511939048767, 'train/grad_norm_pre_clip': 10.852412223815918, 'timing/data': 0.004879891988821328, 'timing/model': 1.20950813498348, 'learning_rate/qwen_vl_interface': 3.395354403280555e-06, 'learning_rate/action_model': 3.395354403280555e-05, 'epoch': 1.3}) +[2026-06-17 14:26:41,261][starVLA.training.train_starvla][INFO] - Step 4432, Loss: {'train/loss': 0.7045983672142029, 'train/grad_norm_pre_clip': 14.038469314575195, 'timing/data': 0.0047792899422347546, 'timing/model': 1.3361211370211095, 'learning_rate/qwen_vl_interface': 3.3933635719902e-06, 'learning_rate/action_model': 3.3933635719901995e-05, 'epoch': 1.3}) +[2026-06-17 14:26:45,918][starVLA.training.train_starvla][INFO] - Step 4433, Loss: {'train/loss': 0.4556184709072113, 'train/grad_norm_pre_clip': 10.168000221252441, 'timing/data': 0.005364995100535452, 'timing/model': 1.205013380968012, 'learning_rate/qwen_vl_interface': 3.3913731255825754e-06, 'learning_rate/action_model': 3.3913731255825754e-05, 'epoch': 1.3}) +[2026-06-17 14:26:50,699][starVLA.training.train_starvla][INFO] - Step 4434, Loss: {'train/loss': 0.7751001119613647, 'train/grad_norm_pre_clip': 12.247818946838379, 'timing/data': 0.00558721402194351, 'timing/model': 1.2013678909279406, 'learning_rate/qwen_vl_interface': 3.3893830644703003e-06, 'learning_rate/action_model': 3.3893830644703e-05, 'epoch': 1.3}) +[2026-06-17 14:26:55,386][starVLA.training.train_starvla][INFO] - Step 4435, Loss: {'train/loss': 0.5696791410446167, 'train/grad_norm_pre_clip': 11.631489753723145, 'timing/data': 0.00511990103404969, 'timing/model': 1.2029790410306305, 'learning_rate/qwen_vl_interface': 3.3873933890659174e-06, 'learning_rate/action_model': 3.3873933890659174e-05, 'epoch': 1.31}) +[2026-06-17 14:27:00,197][starVLA.training.train_starvla][INFO] - Step 4436, Loss: {'train/loss': 0.96373051404953, 'train/grad_norm_pre_clip': 14.617774963378906, 'timing/data': 0.0050453710136935115, 'timing/model': 1.2026740129804239, 'learning_rate/qwen_vl_interface': 3.3854040997818897e-06, 'learning_rate/action_model': 3.38540409978189e-05, 'epoch': 1.31}) +[2026-06-17 14:27:04,992][starVLA.training.train_starvla][INFO] - Step 4437, Loss: {'train/loss': 0.4556313157081604, 'train/grad_norm_pre_clip': 10.092449188232422, 'timing/data': 0.005055108922533691, 'timing/model': 1.2071890269871801, 'learning_rate/qwen_vl_interface': 3.383415197030599e-06, 'learning_rate/action_model': 3.383415197030599e-05, 'epoch': 1.31}) +[2026-06-17 14:27:09,781][starVLA.training.train_starvla][INFO] - Step 4438, Loss: {'train/loss': 0.5519208908081055, 'train/grad_norm_pre_clip': 11.421839714050293, 'timing/data': 0.005025184014812112, 'timing/model': 1.3298123419517651, 'learning_rate/qwen_vl_interface': 3.3814266812243473e-06, 'learning_rate/action_model': 3.3814266812243474e-05, 'epoch': 1.31}) +[2026-06-17 14:27:14,436][starVLA.training.train_starvla][INFO] - Step 4439, Loss: {'train/loss': 0.5940122604370117, 'train/grad_norm_pre_clip': 10.745077133178711, 'timing/data': 0.005315055954270065, 'timing/model': 1.2079424230614677, 'learning_rate/qwen_vl_interface': 3.3794385527753522e-06, 'learning_rate/action_model': 3.379438552775352e-05, 'epoch': 1.31}) +[2026-06-17 14:27:19,200][starVLA.training.train_starvla][INFO] - Step 4440, Loss: {'train/loss': 0.5463180541992188, 'train/grad_norm_pre_clip': 10.048235893249512, 'timing/data': 0.004597624996677041, 'timing/model': 1.2002709669759497, 'learning_rate/qwen_vl_interface': 3.3774508120957606e-06, 'learning_rate/action_model': 3.37745081209576e-05, 'epoch': 1.31}) +[2026-06-17 14:27:23,962][starVLA.training.train_starvla][INFO] - Step 4441, Loss: {'train/loss': 0.6527202725410461, 'train/grad_norm_pre_clip': 13.616373062133789, 'timing/data': 0.0049510960234329104, 'timing/model': 1.2031646220711991, 'learning_rate/qwen_vl_interface': 3.37546345959763e-06, 'learning_rate/action_model': 3.3754634595976295e-05, 'epoch': 1.31}) +[2026-06-17 14:27:28,741][starVLA.training.train_starvla][INFO] - Step 4442, Loss: {'train/loss': 0.8533111810684204, 'train/grad_norm_pre_clip': 13.606412887573242, 'timing/data': 0.005009007058106363, 'timing/model': 1.2073074749205261, 'learning_rate/qwen_vl_interface': 3.3734764956929407e-06, 'learning_rate/action_model': 3.37347649569294e-05, 'epoch': 1.31}) +[2026-06-17 14:27:33,513][starVLA.training.train_starvla][INFO] - Step 4443, Loss: {'train/loss': 0.7467432618141174, 'train/grad_norm_pre_clip': 11.216024398803711, 'timing/data': 0.004521830007433891, 'timing/model': 1.2034459369024262, 'learning_rate/qwen_vl_interface': 3.3714899207935937e-06, 'learning_rate/action_model': 3.371489920793594e-05, 'epoch': 1.31}) +[2026-06-17 14:27:38,301][starVLA.training.train_starvla][INFO] - Step 4444, Loss: {'train/loss': 0.5689585208892822, 'train/grad_norm_pre_clip': 11.301514625549316, 'timing/data': 0.004833125043660402, 'timing/model': 1.3383747929474339, 'learning_rate/qwen_vl_interface': 3.369503735311407e-06, 'learning_rate/action_model': 3.369503735311407e-05, 'epoch': 1.31}) +[2026-06-17 14:27:42,945][starVLA.training.train_starvla][INFO] - Step 4445, Loss: {'train/loss': 0.6641448140144348, 'train/grad_norm_pre_clip': 13.616974830627441, 'timing/data': 0.005212365067563951, 'timing/model': 1.2057279489235952, 'learning_rate/qwen_vl_interface': 3.3675179396581208e-06, 'learning_rate/action_model': 3.367517939658121e-05, 'epoch': 1.31}) +[2026-06-17 14:27:47,742][starVLA.training.train_starvla][INFO] - Step 4446, Loss: {'train/loss': 0.68504798412323, 'train/grad_norm_pre_clip': 10.212356567382812, 'timing/data': 0.004892470082268119, 'timing/model': 1.201608079019934, 'learning_rate/qwen_vl_interface': 3.365532534245389e-06, 'learning_rate/action_model': 3.3655325342453894e-05, 'epoch': 1.31}) +[2026-06-17 14:27:52,538][starVLA.training.train_starvla][INFO] - Step 4447, Loss: {'train/loss': 0.9083212614059448, 'train/grad_norm_pre_clip': 14.092154502868652, 'timing/data': 0.005162514979019761, 'timing/model': 1.2025609040865675, 'learning_rate/qwen_vl_interface': 3.3635475194847927e-06, 'learning_rate/action_model': 3.363547519484793e-05, 'epoch': 1.31}) +[2026-06-17 14:27:57,329][starVLA.training.train_starvla][INFO] - Step 4448, Loss: {'train/loss': 0.6048413515090942, 'train/grad_norm_pre_clip': 11.547491073608398, 'timing/data': 0.004541015950962901, 'timing/model': 1.2040316719794646, 'learning_rate/qwen_vl_interface': 3.361562895787825e-06, 'learning_rate/action_model': 3.361562895787825e-05, 'epoch': 1.31}) +[2026-06-17 14:28:01,992][starVLA.training.train_starvla][INFO] - Step 4449, Loss: {'train/loss': 0.5001873970031738, 'train/grad_norm_pre_clip': 10.007312774658203, 'timing/data': 0.0052095839055255055, 'timing/model': 1.2051280059386045, 'learning_rate/qwen_vl_interface': 3.3595786635659e-06, 'learning_rate/action_model': 3.3595786635659e-05, 'epoch': 1.31}) +[2026-06-17 14:28:06,697][starVLA.training.train_starvla][INFO] - Step 4450, Loss: {'train/loss': 0.6201477646827698, 'train/grad_norm_pre_clip': 13.321425437927246, 'timing/data': 0.005114265950396657, 'timing/model': 1.2548054410144687, 'learning_rate/qwen_vl_interface': 3.357594823230353e-06, 'learning_rate/action_model': 3.3575948232303525e-05, 'epoch': 1.31}) +[2026-06-17 14:28:11,425][starVLA.training.train_starvla][INFO] - Step 4451, Loss: {'train/loss': 0.5787906646728516, 'train/grad_norm_pre_clip': 24.913480758666992, 'timing/data': 0.005312392953783274, 'timing/model': 1.2009739980567247, 'learning_rate/qwen_vl_interface': 3.3556113751924357e-06, 'learning_rate/action_model': 3.355611375192436e-05, 'epoch': 1.31}) +[2026-06-17 14:28:16,147][starVLA.training.train_starvla][INFO] - Step 4452, Loss: {'train/loss': 0.7972297668457031, 'train/grad_norm_pre_clip': 14.471415519714355, 'timing/data': 0.004646020941436291, 'timing/model': 1.201764120021835, 'learning_rate/qwen_vl_interface': 3.3536283198633183e-06, 'learning_rate/action_model': 3.353628319863318e-05, 'epoch': 1.31}) +[2026-06-17 14:28:20,918][starVLA.training.train_starvla][INFO] - Step 4453, Loss: {'train/loss': 0.7954457402229309, 'train/grad_norm_pre_clip': 14.151786804199219, 'timing/data': 0.005007512983866036, 'timing/model': 1.2022159158950672, 'learning_rate/qwen_vl_interface': 3.3516456576540913e-06, 'learning_rate/action_model': 3.3516456576540914e-05, 'epoch': 1.31}) +[2026-06-17 14:28:25,697][starVLA.training.train_starvla][INFO] - Step 4454, Loss: {'train/loss': 0.7245427966117859, 'train/grad_norm_pre_clip': 13.184012413024902, 'timing/data': 0.005005875020287931, 'timing/model': 1.204232144053094, 'learning_rate/qwen_vl_interface': 3.349663388975763e-06, 'learning_rate/action_model': 3.3496633889757625e-05, 'epoch': 1.31}) +[2026-06-17 14:28:30,477][starVLA.training.train_starvla][INFO] - Step 4455, Loss: {'train/loss': 0.5708907246589661, 'train/grad_norm_pre_clip': 11.57175350189209, 'timing/data': 0.005056982045061886, 'timing/model': 1.3291938620386645, 'learning_rate/qwen_vl_interface': 3.347681514239258e-06, 'learning_rate/action_model': 3.347681514239257e-05, 'epoch': 1.31}) +[2026-06-17 14:28:35,244][starVLA.training.train_starvla][INFO] - Step 4456, Loss: {'train/loss': 0.5428764820098877, 'train/grad_norm_pre_clip': 10.987943649291992, 'timing/data': 0.005000041099265218, 'timing/model': 1.3248270569602028, 'learning_rate/qwen_vl_interface': 3.345700033855421e-06, 'learning_rate/action_model': 3.345700033855421e-05, 'epoch': 1.31}) +[2026-06-17 14:28:39,891][starVLA.training.train_starvla][INFO] - Step 4457, Loss: {'train/loss': 0.6792985796928406, 'train/grad_norm_pre_clip': 12.917505264282227, 'timing/data': 0.004783689975738525, 'timing/model': 1.2012152520474046, 'learning_rate/qwen_vl_interface': 3.3437189482350173e-06, 'learning_rate/action_model': 3.343718948235017e-05, 'epoch': 1.31}) +[2026-06-17 14:28:44,627][starVLA.training.train_starvla][INFO] - Step 4458, Loss: {'train/loss': 0.4281553626060486, 'train/grad_norm_pre_clip': 10.45840835571289, 'timing/data': 0.0049899721052497625, 'timing/model': 1.2043801780091599, 'learning_rate/qwen_vl_interface': 3.341738257788727e-06, 'learning_rate/action_model': 3.3417382577887266e-05, 'epoch': 1.31}) +[2026-06-17 14:28:49,434][starVLA.training.train_starvla][INFO] - Step 4459, Loss: {'train/loss': 0.5675184726715088, 'train/grad_norm_pre_clip': 10.887044906616211, 'timing/data': 0.005103872972540557, 'timing/model': 1.2060395689914003, 'learning_rate/qwen_vl_interface': 3.3397579629271487e-06, 'learning_rate/action_model': 3.3397579629271486e-05, 'epoch': 1.31}) +[2026-06-17 14:28:54,211][starVLA.training.train_starvla][INFO] - Step 4460, Loss: {'train/loss': 0.7643147110939026, 'train/grad_norm_pre_clip': 13.739296913146973, 'timing/data': 0.005181904067285359, 'timing/model': 1.2010022059548646, 'learning_rate/qwen_vl_interface': 3.3377780640607997e-06, 'learning_rate/action_model': 3.3377780640608e-05, 'epoch': 1.31}) +[2026-06-17 14:28:58,980][starVLA.training.train_starvla][INFO] - Step 4461, Loss: {'train/loss': 0.7001048922538757, 'train/grad_norm_pre_clip': 12.630021095275879, 'timing/data': 0.004693463910371065, 'timing/model': 1.3299205279909074, 'learning_rate/qwen_vl_interface': 3.3357985616001166e-06, 'learning_rate/action_model': 3.335798561600116e-05, 'epoch': 1.31}) +[2026-06-17 14:29:03,745][starVLA.training.train_starvla][INFO] - Step 4462, Loss: {'train/loss': 1.0194200277328491, 'train/grad_norm_pre_clip': 17.957082748413086, 'timing/data': 0.005107924924232066, 'timing/model': 1.3228508899919689, 'learning_rate/qwen_vl_interface': 3.3338194559554524e-06, 'learning_rate/action_model': 3.333819455955453e-05, 'epoch': 1.31}) +[2026-06-17 14:29:08,481][starVLA.training.train_starvla][INFO] - Step 4463, Loss: {'train/loss': 0.595706582069397, 'train/grad_norm_pre_clip': 12.120196342468262, 'timing/data': 0.00525361392647028, 'timing/model': 1.2031411700882018, 'learning_rate/qwen_vl_interface': 3.3318407475370758e-06, 'learning_rate/action_model': 3.3318407475370755e-05, 'epoch': 1.31}) +[2026-06-17 14:29:13,243][starVLA.training.train_starvla][INFO] - Step 4464, Loss: {'train/loss': 0.5888233780860901, 'train/grad_norm_pre_clip': 13.0576810836792, 'timing/data': 0.005122065078467131, 'timing/model': 1.2098983399337158, 'learning_rate/qwen_vl_interface': 3.329862436755176e-06, 'learning_rate/action_model': 3.329862436755176e-05, 'epoch': 1.31}) +[2026-06-17 14:29:18,069][starVLA.training.train_starvla][INFO] - Step 4465, Loss: {'train/loss': 0.7638387680053711, 'train/grad_norm_pre_clip': 12.522370338439941, 'timing/data': 0.005007339990697801, 'timing/model': 1.2017544860718772, 'learning_rate/qwen_vl_interface': 3.3278845240198592e-06, 'learning_rate/action_model': 3.3278845240198595e-05, 'epoch': 1.31}) +[2026-06-17 14:29:22,935][starVLA.training.train_starvla][INFO] - Step 4466, Loss: {'train/loss': 0.5567917227745056, 'train/grad_norm_pre_clip': 13.129444122314453, 'timing/data': 0.004853251972235739, 'timing/model': 1.205484538921155, 'learning_rate/qwen_vl_interface': 3.325907009741149e-06, 'learning_rate/action_model': 3.3259070097411485e-05, 'epoch': 1.31}) +[2026-06-17 14:29:27,670][starVLA.training.train_starvla][INFO] - Step 4467, Loss: {'train/loss': 0.5664323568344116, 'train/grad_norm_pre_clip': 10.163768768310547, 'timing/data': 0.00530114199500531, 'timing/model': 1.2845591580262408, 'learning_rate/qwen_vl_interface': 3.323929894328987e-06, 'learning_rate/action_model': 3.323929894328987e-05, 'epoch': 1.31}) +[2026-06-17 14:29:32,328][starVLA.training.train_starvla][INFO] - Step 4468, Loss: {'train/loss': 0.6147643327713013, 'train/grad_norm_pre_clip': 11.018900871276855, 'timing/data': 0.005066188983619213, 'timing/model': 1.2094029770232737, 'learning_rate/qwen_vl_interface': 3.321953178193228e-06, 'learning_rate/action_model': 3.321953178193228e-05, 'epoch': 1.31}) +[2026-06-17 14:29:37,101][starVLA.training.train_starvla][INFO] - Step 4469, Loss: {'train/loss': 0.5371835231781006, 'train/grad_norm_pre_clip': 11.529712677001953, 'timing/data': 0.004980586003512144, 'timing/model': 1.2035844610072672, 'learning_rate/qwen_vl_interface': 3.319976861743653e-06, 'learning_rate/action_model': 3.319976861743653e-05, 'epoch': 1.32}) +[2026-06-17 14:29:41,882][starVLA.training.train_starvla][INFO] - Step 4470, Loss: {'train/loss': 0.6217749118804932, 'train/grad_norm_pre_clip': 13.523266792297363, 'timing/data': 0.005091995000839233, 'timing/model': 1.202670076978393, 'learning_rate/qwen_vl_interface': 3.3180009453899496e-06, 'learning_rate/action_model': 3.318000945389949e-05, 'epoch': 1.32}) +[2026-06-17 14:29:46,677][starVLA.training.train_starvla][INFO] - Step 4471, Loss: {'train/loss': 0.4847414791584015, 'train/grad_norm_pre_clip': 9.619939804077148, 'timing/data': 0.00487352698110044, 'timing/model': 1.2054950529709458, 'learning_rate/qwen_vl_interface': 3.3160254295417287e-06, 'learning_rate/action_model': 3.3160254295417285e-05, 'epoch': 1.32}) +[2026-06-17 14:29:51,467][starVLA.training.train_starvla][INFO] - Step 4472, Loss: {'train/loss': 0.910663902759552, 'train/grad_norm_pre_clip': 14.082296371459961, 'timing/data': 0.005209827097132802, 'timing/model': 1.3345676850294694, 'learning_rate/qwen_vl_interface': 3.314050314608518e-06, 'learning_rate/action_model': 3.314050314608518e-05, 'epoch': 1.32}) +[2026-06-17 14:29:56,198][starVLA.training.train_starvla][INFO] - Step 4473, Loss: {'train/loss': 0.6888872385025024, 'train/grad_norm_pre_clip': 12.410423278808594, 'timing/data': 0.005099480971693993, 'timing/model': 1.2837190029677004, 'learning_rate/qwen_vl_interface': 3.3120756009997606e-06, 'learning_rate/action_model': 3.312075600999761e-05, 'epoch': 1.32}) +[2026-06-17 14:30:00,855][starVLA.training.train_starvla][INFO] - Step 4474, Loss: {'train/loss': 0.5615533590316772, 'train/grad_norm_pre_clip': 12.381098747253418, 'timing/data': 0.005276671960018575, 'timing/model': 1.206873137038201, 'learning_rate/qwen_vl_interface': 3.310101289124817e-06, 'learning_rate/action_model': 3.310101289124817e-05, 'epoch': 1.32}) +[2026-06-17 14:30:05,648][starVLA.training.train_starvla][INFO] - Step 4475, Loss: {'train/loss': 0.9962345361709595, 'train/grad_norm_pre_clip': 13.815445899963379, 'timing/data': 0.004835481056943536, 'timing/model': 1.2097660099389032, 'learning_rate/qwen_vl_interface': 3.3081273793929615e-06, 'learning_rate/action_model': 3.308127379392961e-05, 'epoch': 1.32}) +[2026-06-17 14:30:10,439][starVLA.training.train_starvla][INFO] - Step 4476, Loss: {'train/loss': 0.5712869763374329, 'train/grad_norm_pre_clip': 14.017047882080078, 'timing/data': 0.005077988957054913, 'timing/model': 1.204723178059794, 'learning_rate/qwen_vl_interface': 3.306153872213392e-06, 'learning_rate/action_model': 3.306153872213392e-05, 'epoch': 1.32}) +[2026-06-17 14:30:15,232][starVLA.training.train_starvla][INFO] - Step 4477, Loss: {'train/loss': 0.5213314294815063, 'train/grad_norm_pre_clip': 10.741169929504395, 'timing/data': 0.005037005990743637, 'timing/model': 1.2102584870299324, 'learning_rate/qwen_vl_interface': 3.304180767995215e-06, 'learning_rate/action_model': 3.3041807679952146e-05, 'epoch': 1.32}) +[2026-06-17 14:30:20,004][starVLA.training.train_starvla][INFO] - Step 4478, Loss: {'train/loss': 0.5637630820274353, 'train/grad_norm_pre_clip': 12.144071578979492, 'timing/data': 0.005337179056368768, 'timing/model': 1.2024614709662274, 'learning_rate/qwen_vl_interface': 3.3022080671474588e-06, 'learning_rate/action_model': 3.302208067147459e-05, 'epoch': 1.32}) +[2026-06-17 14:30:24,754][starVLA.training.train_starvla][INFO] - Step 4479, Loss: {'train/loss': 0.825990617275238, 'train/grad_norm_pre_clip': 14.52026653289795, 'timing/data': 0.004941791063174605, 'timing/model': 1.2983279540203512, 'learning_rate/qwen_vl_interface': 3.300235770079066e-06, 'learning_rate/action_model': 3.300235770079066e-05, 'epoch': 1.32}) +[2026-06-17 14:30:29,451][starVLA.training.train_starvla][INFO] - Step 4480, Loss: {'train/loss': 0.3693900406360626, 'train/grad_norm_pre_clip': 9.229019165039062, 'timing/data': 0.005206714035011828, 'timing/model': 1.2051281440071762, 'learning_rate/qwen_vl_interface': 3.298263877198896e-06, 'learning_rate/action_model': 3.298263877198896e-05, 'epoch': 1.32}) +[2026-06-17 14:30:34,333][starVLA.training.train_starvla][INFO] - Step 4481, Loss: {'train/loss': 0.7894428968429565, 'train/grad_norm_pre_clip': 13.068503379821777, 'timing/data': 0.005102368071675301, 'timing/model': 1.2053047110093758, 'learning_rate/qwen_vl_interface': 3.296292388915726e-06, 'learning_rate/action_model': 3.296292388915726e-05, 'epoch': 1.32}) +[2026-06-17 14:30:39,129][starVLA.training.train_starvla][INFO] - Step 4482, Loss: {'train/loss': 0.612525463104248, 'train/grad_norm_pre_clip': 13.989550590515137, 'timing/data': 0.004955595009960234, 'timing/model': 1.2049815891077742, 'learning_rate/qwen_vl_interface': 3.294321305638243e-06, 'learning_rate/action_model': 3.294321305638243e-05, 'epoch': 1.32}) +[2026-06-17 14:30:43,932][starVLA.training.train_starvla][INFO] - Step 4483, Loss: {'train/loss': 0.4696817696094513, 'train/grad_norm_pre_clip': 9.185576438903809, 'timing/data': 0.005220219027251005, 'timing/model': 1.2053669979795814, 'learning_rate/qwen_vl_interface': 3.292350627775061e-06, 'learning_rate/action_model': 3.292350627775061e-05, 'epoch': 1.32}) +[2026-06-17 14:30:48,745][starVLA.training.train_starvla][INFO] - Step 4484, Loss: {'train/loss': 0.6547860503196716, 'train/grad_norm_pre_clip': 14.148124694824219, 'timing/data': 0.00490526098292321, 'timing/model': 1.3457644960144535, 'learning_rate/qwen_vl_interface': 3.2903803557346997e-06, 'learning_rate/action_model': 3.2903803557347e-05, 'epoch': 1.32}) +[2026-06-17 14:30:53,396][starVLA.training.train_starvla][INFO] - Step 4485, Loss: {'train/loss': 0.6318943500518799, 'train/grad_norm_pre_clip': 11.236084938049316, 'timing/data': 0.005136544932611287, 'timing/model': 1.2020719819702208, 'learning_rate/qwen_vl_interface': 3.2884104899255986e-06, 'learning_rate/action_model': 3.288410489925599e-05, 'epoch': 1.32}) +[2026-06-17 14:30:58,176][starVLA.training.train_starvla][INFO] - Step 4486, Loss: {'train/loss': 0.4801284372806549, 'train/grad_norm_pre_clip': 11.948434829711914, 'timing/data': 0.005521540995687246, 'timing/model': 1.2045681199524552, 'learning_rate/qwen_vl_interface': 3.2864410307561124e-06, 'learning_rate/action_model': 3.286441030756112e-05, 'epoch': 1.32}) +[2026-06-17 14:31:02,955][starVLA.training.train_starvla][INFO] - Step 4487, Loss: {'train/loss': 0.7483674883842468, 'train/grad_norm_pre_clip': 13.222535133361816, 'timing/data': 0.005150870070792735, 'timing/model': 1.2036037599900737, 'learning_rate/qwen_vl_interface': 3.284471978634515e-06, 'learning_rate/action_model': 3.284471978634515e-05, 'epoch': 1.32}) +[2026-06-17 14:31:07,672][starVLA.training.train_starvla][INFO] - Step 4488, Loss: {'train/loss': 0.5957033634185791, 'train/grad_norm_pre_clip': 14.327975273132324, 'timing/data': 0.005028803017921746, 'timing/model': 1.2047121870564297, 'learning_rate/qwen_vl_interface': 3.2825033339689926e-06, 'learning_rate/action_model': 3.282503333968993e-05, 'epoch': 1.32}) +[2026-06-17 14:31:12,404][starVLA.training.train_starvla][INFO] - Step 4489, Loss: {'train/loss': 0.5968592762947083, 'train/grad_norm_pre_clip': 11.419233322143555, 'timing/data': 0.004981326055712998, 'timing/model': 1.2036840099608526, 'learning_rate/qwen_vl_interface': 3.2805350971676447e-06, 'learning_rate/action_model': 3.280535097167645e-05, 'epoch': 1.32}) +[2026-06-17 14:31:17,191][starVLA.training.train_starvla][INFO] - Step 4490, Loss: {'train/loss': 0.4559018611907959, 'train/grad_norm_pre_clip': 9.550960540771484, 'timing/data': 0.005552925984375179, 'timing/model': 1.3342171190306544, 'learning_rate/qwen_vl_interface': 3.2785672686384893e-06, 'learning_rate/action_model': 3.278567268638489e-05, 'epoch': 1.32}) +[2026-06-17 14:31:21,853][starVLA.training.train_starvla][INFO] - Step 4491, Loss: {'train/loss': 0.9983566999435425, 'train/grad_norm_pre_clip': 17.53936004638672, 'timing/data': 0.005061198957264423, 'timing/model': 1.2057822139468044, 'learning_rate/qwen_vl_interface': 3.2765998487894622e-06, 'learning_rate/action_model': 3.276599848789462e-05, 'epoch': 1.32}) +[2026-06-17 14:31:26,665][starVLA.training.train_starvla][INFO] - Step 4492, Loss: {'train/loss': 0.6433659791946411, 'train/grad_norm_pre_clip': 13.867474555969238, 'timing/data': 0.004951181006617844, 'timing/model': 1.2039407449774444, 'learning_rate/qwen_vl_interface': 3.2746328380284098e-06, 'learning_rate/action_model': 3.27463283802841e-05, 'epoch': 1.32}) +[2026-06-17 14:31:31,484][starVLA.training.train_starvla][INFO] - Step 4493, Loss: {'train/loss': 0.5329980254173279, 'train/grad_norm_pre_clip': 11.162086486816406, 'timing/data': 0.0049345120787620544, 'timing/model': 1.2066528629511595, 'learning_rate/qwen_vl_interface': 3.2726662367630953e-06, 'learning_rate/action_model': 3.272666236763095e-05, 'epoch': 1.32}) +[2026-06-17 14:31:36,269][starVLA.training.train_starvla][INFO] - Step 4494, Loss: {'train/loss': 0.7442435622215271, 'train/grad_norm_pre_clip': 12.834383964538574, 'timing/data': 0.005434599006548524, 'timing/model': 1.2081068800762296, 'learning_rate/qwen_vl_interface': 3.270700045401197e-06, 'learning_rate/action_model': 3.2707000454011974e-05, 'epoch': 1.32}) +[2026-06-17 14:31:41,045][starVLA.training.train_starvla][INFO] - Step 4495, Loss: {'train/loss': 0.5734289884567261, 'train/grad_norm_pre_clip': 12.902039527893066, 'timing/data': 0.004899365943856537, 'timing/model': 1.2041523039806634, 'learning_rate/qwen_vl_interface': 3.26873426435031e-06, 'learning_rate/action_model': 3.2687342643503096e-05, 'epoch': 1.32}) +[2026-06-17 14:31:45,820][starVLA.training.train_starvla][INFO] - Step 4496, Loss: {'train/loss': 0.5704142451286316, 'train/grad_norm_pre_clip': 177.397705078125, 'timing/data': 0.004875986953265965, 'timing/model': 1.327902109012939, 'learning_rate/qwen_vl_interface': 3.266768894017942e-06, 'learning_rate/action_model': 3.266768894017942e-05, 'epoch': 1.32}) +[2026-06-17 14:31:50,470][starVLA.training.train_starvla][INFO] - Step 4497, Loss: {'train/loss': 0.702731728553772, 'train/grad_norm_pre_clip': 14.445719718933105, 'timing/data': 0.0048540940042585135, 'timing/model': 1.2039781770436093, 'learning_rate/qwen_vl_interface': 3.264803934811514e-06, 'learning_rate/action_model': 3.264803934811514e-05, 'epoch': 1.32}) +[2026-06-17 14:31:55,269][starVLA.training.train_starvla][INFO] - Step 4498, Loss: {'train/loss': 0.527552604675293, 'train/grad_norm_pre_clip': 12.353261947631836, 'timing/data': 0.005216695019043982, 'timing/model': 1.2082892099861056, 'learning_rate/qwen_vl_interface': 3.2628393871383706e-06, 'learning_rate/action_model': 3.26283938713837e-05, 'epoch': 1.32}) +[2026-06-17 14:32:00,061][starVLA.training.train_starvla][INFO] - Step 4499, Loss: {'train/loss': 0.7181602716445923, 'train/grad_norm_pre_clip': 17.22547721862793, 'timing/data': 0.0050418860046193, 'timing/model': 1.2064645639620721, 'learning_rate/qwen_vl_interface': 3.2608752514057585e-06, 'learning_rate/action_model': 3.260875251405758e-05, 'epoch': 1.32}) +[2026-06-17 14:42:10,537][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 4500 +[2026-06-17 14:42:10,538][starVLA.training.train_starvla][INFO] - Step 4500, Loss: {'train/loss': 0.5629664659500122, 'train/grad_norm_pre_clip': 11.617850303649902, 'eval/loss': 0.7258113622665405, 'eval/latency_0/loss': 0.7258113622665405, 'eval/demon_fire_precision': 0.8967434471803019, 'eval/demon_fire_recall': 0.9377076411960132, 'eval/demon_fire_f1': 0.9167681688997157, 'eval/demon_fire_rate_model': 0.7002224694104561, 'eval/demon_fire_rate_teacher': 0.6696329254727474, 'eval/demon_left_precision': 0.8106476399560922, 'eval/demon_left_recall': 0.8368271954674221, 'eval/demon_left_f1': 0.8235294117647061, 'eval/demon_left_rate_model': 0.5066740823136818, 'eval/demon_left_rate_teacher': 0.49082313681868744, 'eval/demon_right_precision': 0.836027713625866, 'eval/demon_right_recall': 0.8180790960451978, 'eval/demon_right_f1': 0.826956025128498, 'eval/demon_right_rate_model': 0.48164627363737483, 'eval/demon_right_rate_teacher': 0.4922135706340378, 'eval/demon_move_f1': 0.825242718446602, 'eval/demon_cc_f1': 0.8872474987368292, 'eval/latency_0/demon_fire_precision': 0.8967434471803019, 'eval/latency_0/demon_fire_recall': 0.9377076411960132, 'eval/latency_0/demon_fire_f1': 0.9167681688997157, 'eval/latency_0/demon_fire_rate_model': 0.7002224694104561, 'eval/latency_0/demon_fire_rate_teacher': 0.6696329254727474, 'eval/latency_0/demon_left_precision': 0.8106476399560922, 'eval/latency_0/demon_left_recall': 0.8368271954674221, 'eval/latency_0/demon_left_f1': 0.8235294117647061, 'eval/latency_0/demon_left_rate_model': 0.5066740823136818, 'eval/latency_0/demon_left_rate_teacher': 0.49082313681868744, 'eval/latency_0/demon_right_precision': 0.836027713625866, 'eval/latency_0/demon_right_recall': 0.8180790960451978, 'eval/latency_0/demon_right_f1': 0.826956025128498, 'eval/latency_0/demon_right_rate_model': 0.48164627363737483, 'eval/latency_0/demon_right_rate_teacher': 0.4922135706340378, 'eval/latency_0/demon_move_f1': 0.825242718446602, 'eval/latency_0/demon_cc_f1': 0.8872474987368292, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 7476.0, 'rl_games_eval/mid_train/mean_length': 3600.0, 'rl_games_eval/mid_train/std_reward': 372.52382474145196, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/macro_mean_reward': 7476.0, 'rl_games_eval/mid_train/macro_mean_length': 3600.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/demon_attack__latency_0/mean_reward': 7476.0, 'rl_games_eval/mid_train/demon_attack__latency_0/mean_length': 3600.0, 'rl_games_eval/mid_train/demon_attack__latency_0/std_reward': 372.52382474145196, 'rl_games_eval/mid_train/demon_attack__latency_0/std_length': 0.0, 'rl_games_eval/mid_train/demon_attack__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 7476.0, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.005055390065535903, 'timing/model': 1.2029170939931646, 'learning_rate/qwen_vl_interface': 3.2589115280208473e-06, 'learning_rate/action_model': 3.258911528020847e-05, 'epoch': 1.32}) +[2026-06-17 14:42:10,538][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_4500_state +[2026-06-17 14:42:10,541][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 14:42:14,105][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_4500_state/model.safetensors +[2026-06-17 14:42:21,747][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_4500_state/optimizer.bin +[2026-06-17 14:42:21,748][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_4500_state/sampler.bin +[2026-06-17 14:42:21,748][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_4500_state/sampler_1.bin +[2026-06-17 14:42:21,749][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_4500_state/random_states_0.pkl +[2026-06-17 14:42:21,795][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4500 +[2026-06-17 14:42:21,795][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 14:42:21,830][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 14:42:26,530][starVLA.training.train_starvla][INFO] - Step 4501, Loss: {'train/loss': 0.6319579482078552, 'train/grad_norm_pre_clip': 14.669079780578613, 'timing/data': 0.004945390042848885, 'timing/model': 1.31288072001189, 'learning_rate/qwen_vl_interface': 3.2569482173907164e-06, 'learning_rate/action_model': 3.2569482173907165e-05, 'epoch': 1.32}) +[2026-06-17 14:42:31,199][starVLA.training.train_starvla][INFO] - Step 4502, Loss: {'train/loss': 0.6926668286323547, 'train/grad_norm_pre_clip': 14.425971984863281, 'timing/data': 0.0049442260060459375, 'timing/model': 1.2686888759490103, 'learning_rate/qwen_vl_interface': 3.2549853199223668e-06, 'learning_rate/action_model': 3.254985319922367e-05, 'epoch': 1.32}) +[2026-06-17 14:42:35,781][starVLA.training.train_starvla][INFO] - Step 4503, Loss: {'train/loss': 0.6575527191162109, 'train/grad_norm_pre_clip': 12.906110763549805, 'timing/data': 0.0052364569855853915, 'timing/model': 1.1870122710242867, 'learning_rate/qwen_vl_interface': 3.253022836022708e-06, 'learning_rate/action_model': 3.253022836022708e-05, 'epoch': 1.33}) +[2026-06-17 14:42:40,505][starVLA.training.train_starvla][INFO] - Step 4504, Loss: {'train/loss': 0.9556737542152405, 'train/grad_norm_pre_clip': 15.279807090759277, 'timing/data': 0.005494590965099633, 'timing/model': 1.1852182040456682, 'learning_rate/qwen_vl_interface': 3.2510607660985627e-06, 'learning_rate/action_model': 3.251060766098563e-05, 'epoch': 1.33}) +[2026-06-17 14:42:45,220][starVLA.training.train_starvla][INFO] - Step 4505, Loss: {'train/loss': 0.7679304480552673, 'train/grad_norm_pre_clip': 13.704020500183105, 'timing/data': 0.004774717963300645, 'timing/model': 1.1880409270524979, 'learning_rate/qwen_vl_interface': 3.249099110556672e-06, 'learning_rate/action_model': 3.249099110556672e-05, 'epoch': 1.33}) +[2026-06-17 14:42:49,948][starVLA.training.train_starvla][INFO] - Step 4506, Loss: {'train/loss': 0.5287674069404602, 'train/grad_norm_pre_clip': 12.38519287109375, 'timing/data': 0.004986022948287427, 'timing/model': 1.1913881349610165, 'learning_rate/qwen_vl_interface': 3.247137869803689e-06, 'learning_rate/action_model': 3.2471378698036886e-05, 'epoch': 1.33}) +[2026-06-17 14:42:54,678][starVLA.training.train_starvla][INFO] - Step 4507, Loss: {'train/loss': 0.6105870008468628, 'train/grad_norm_pre_clip': 13.171492576599121, 'timing/data': 0.004796688095666468, 'timing/model': 1.1906046569347382, 'learning_rate/qwen_vl_interface': 3.2451770442461813e-06, 'learning_rate/action_model': 3.245177044246181e-05, 'epoch': 1.33}) +[2026-06-17 14:42:59,372][starVLA.training.train_starvla][INFO] - Step 4508, Loss: {'train/loss': 0.5936991572380066, 'train/grad_norm_pre_clip': 11.606386184692383, 'timing/data': 0.0052008769707754254, 'timing/model': 1.2795386159559712, 'learning_rate/qwen_vl_interface': 3.2432166342906286e-06, 'learning_rate/action_model': 3.2432166342906284e-05, 'epoch': 1.33}) +[2026-06-17 14:43:03,982][starVLA.training.train_starvla][INFO] - Step 4509, Loss: {'train/loss': 0.7672311663627625, 'train/grad_norm_pre_clip': 12.785002708435059, 'timing/data': 0.005209053051657975, 'timing/model': 1.1934046250535175, 'learning_rate/qwen_vl_interface': 3.241256640343427e-06, 'learning_rate/action_model': 3.241256640343427e-05, 'epoch': 1.33}) +[2026-06-17 14:43:08,730][starVLA.training.train_starvla][INFO] - Step 4510, Loss: {'train/loss': 0.6793695092201233, 'train/grad_norm_pre_clip': 15.143031120300293, 'timing/data': 0.0051316419849172235, 'timing/model': 1.1946638419758528, 'learning_rate/qwen_vl_interface': 3.239297062810886e-06, 'learning_rate/action_model': 3.239297062810886e-05, 'epoch': 1.33}) +[2026-06-17 14:43:13,495][starVLA.training.train_starvla][INFO] - Step 4511, Loss: {'train/loss': 0.9276485443115234, 'train/grad_norm_pre_clip': 18.469356536865234, 'timing/data': 0.00470675399992615, 'timing/model': 1.1954400659305975, 'learning_rate/qwen_vl_interface': 3.2373379020992252e-06, 'learning_rate/action_model': 3.237337902099225e-05, 'epoch': 1.33}) +[2026-06-17 14:43:18,248][starVLA.training.train_starvla][INFO] - Step 4512, Loss: {'train/loss': 0.8255813717842102, 'train/grad_norm_pre_clip': 15.655034065246582, 'timing/data': 0.004909457988105714, 'timing/model': 1.1936040800064802, 'learning_rate/qwen_vl_interface': 3.2353791586145855e-06, 'learning_rate/action_model': 3.2353791586145855e-05, 'epoch': 1.33}) +[2026-06-17 14:43:23,004][starVLA.training.train_starvla][INFO] - Step 4513, Loss: {'train/loss': 0.7139532566070557, 'train/grad_norm_pre_clip': 15.369696617126465, 'timing/data': 0.0052617189940065145, 'timing/model': 1.3280365410028026, 'learning_rate/qwen_vl_interface': 3.2334208327630158e-06, 'learning_rate/action_model': 3.233420832763016e-05, 'epoch': 1.33}) +[2026-06-17 14:43:27,630][starVLA.training.train_starvla][INFO] - Step 4514, Loss: {'train/loss': 0.795488178730011, 'train/grad_norm_pre_clip': 17.395994186401367, 'timing/data': 0.004660201957449317, 'timing/model': 1.1945757130160928, 'learning_rate/qwen_vl_interface': 3.231462924950475e-06, 'learning_rate/action_model': 3.2314629249504746e-05, 'epoch': 1.33}) +[2026-06-17 14:43:32,452][starVLA.training.train_starvla][INFO] - Step 4515, Loss: {'train/loss': 0.9558228850364685, 'train/grad_norm_pre_clip': 14.73941707611084, 'timing/data': 0.005164990900084376, 'timing/model': 1.1957997689023614, 'learning_rate/qwen_vl_interface': 3.2295054355828396e-06, 'learning_rate/action_model': 3.22950543558284e-05, 'epoch': 1.33}) +[2026-06-17 14:43:37,200][starVLA.training.train_starvla][INFO] - Step 4516, Loss: {'train/loss': 0.46237897872924805, 'train/grad_norm_pre_clip': 12.073307991027832, 'timing/data': 0.0051006010035052896, 'timing/model': 1.197349934023805, 'learning_rate/qwen_vl_interface': 3.227548365065905e-06, 'learning_rate/action_model': 3.227548365065905e-05, 'epoch': 1.33}) +[2026-06-17 14:43:41,870][starVLA.training.train_starvla][INFO] - Step 4517, Loss: {'train/loss': 0.6360778212547302, 'train/grad_norm_pre_clip': 12.135382652282715, 'timing/data': 0.0048818549839779735, 'timing/model': 1.1954392309999093, 'learning_rate/qwen_vl_interface': 3.2255917138053695e-06, 'learning_rate/action_model': 3.2255917138053696e-05, 'epoch': 1.33}) +[2026-06-17 14:43:46,574][starVLA.training.train_starvla][INFO] - Step 4518, Loss: {'train/loss': 0.4026924669742584, 'train/grad_norm_pre_clip': 11.107666015625, 'timing/data': 0.005150430952198803, 'timing/model': 1.19578961096704, 'learning_rate/qwen_vl_interface': 3.2236354822068515e-06, 'learning_rate/action_model': 3.223635482206852e-05, 'epoch': 1.33}) +[2026-06-17 14:43:51,356][starVLA.training.train_starvla][INFO] - Step 4519, Loss: {'train/loss': 0.7953938841819763, 'train/grad_norm_pre_clip': 17.242950439453125, 'timing/data': 0.0071245989529415965, 'timing/model': 1.3473569289781153, 'learning_rate/qwen_vl_interface': 3.2216796706758757e-06, 'learning_rate/action_model': 3.221679670675876e-05, 'epoch': 1.33}) +[2026-06-17 14:43:55,987][starVLA.training.train_starvla][INFO] - Step 4520, Loss: {'train/loss': 0.5479688048362732, 'train/grad_norm_pre_clip': 11.137943267822266, 'timing/data': 0.004553406033664942, 'timing/model': 1.1989075388992205, 'learning_rate/qwen_vl_interface': 3.219724279617888e-06, 'learning_rate/action_model': 3.219724279617888e-05, 'epoch': 1.33}) +[2026-06-17 14:44:00,744][starVLA.training.train_starvla][INFO] - Step 4521, Loss: {'train/loss': 0.8983335494995117, 'train/grad_norm_pre_clip': 12.522974967956543, 'timing/data': 0.005103560979478061, 'timing/model': 1.1970260569360107, 'learning_rate/qwen_vl_interface': 3.2177693094382424e-06, 'learning_rate/action_model': 3.2177693094382424e-05, 'epoch': 1.33}) +[2026-06-17 14:44:05,500][starVLA.training.train_starvla][INFO] - Step 4522, Loss: {'train/loss': 0.3620949387550354, 'train/grad_norm_pre_clip': 11.234777450561523, 'timing/data': 0.004726859042420983, 'timing/model': 1.1986462739296257, 'learning_rate/qwen_vl_interface': 3.2158147605422052e-06, 'learning_rate/action_model': 3.215814760542205e-05, 'epoch': 1.33}) +[2026-06-17 14:44:10,249][starVLA.training.train_starvla][INFO] - Step 4523, Loss: {'train/loss': 0.6744661331176758, 'train/grad_norm_pre_clip': 12.64433479309082, 'timing/data': 0.004906377987936139, 'timing/model': 1.1992958060000092, 'learning_rate/qwen_vl_interface': 3.2138606333349576e-06, 'learning_rate/action_model': 3.213860633334957e-05, 'epoch': 1.33}) +[2026-06-17 14:44:14,963][starVLA.training.train_starvla][INFO] - Step 4524, Loss: {'train/loss': 0.6840951442718506, 'train/grad_norm_pre_clip': 14.339083671569824, 'timing/data': 0.005259569967165589, 'timing/model': 1.1991524500772357, 'learning_rate/qwen_vl_interface': 3.2119069282215913e-06, 'learning_rate/action_model': 3.211906928221591e-05, 'epoch': 1.33}) +[2026-06-17 14:44:19,718][starVLA.training.train_starvla][INFO] - Step 4525, Loss: {'train/loss': 0.7236142158508301, 'train/grad_norm_pre_clip': 15.126825332641602, 'timing/data': 0.005267885979264975, 'timing/model': 1.3224567199358717, 'learning_rate/qwen_vl_interface': 3.2099536456071118e-06, 'learning_rate/action_model': 3.209953645607112e-05, 'epoch': 1.33}) +[2026-06-17 14:44:24,357][starVLA.training.train_starvla][INFO] - Step 4526, Loss: {'train/loss': 0.4721348285675049, 'train/grad_norm_pre_clip': 14.483930587768555, 'timing/data': 0.004799248999916017, 'timing/model': 1.197664070990868, 'learning_rate/qwen_vl_interface': 3.2080007858964347e-06, 'learning_rate/action_model': 3.2080007858964345e-05, 'epoch': 1.33}) +[2026-06-17 14:44:29,121][starVLA.training.train_starvla][INFO] - Step 4527, Loss: {'train/loss': 0.5072221755981445, 'train/grad_norm_pre_clip': 11.8787202835083, 'timing/data': 0.00514451798517257, 'timing/model': 1.2000610310351476, 'learning_rate/qwen_vl_interface': 3.206048349494396e-06, 'learning_rate/action_model': 3.206048349494396e-05, 'epoch': 1.33}) +[2026-06-17 14:44:33,891][starVLA.training.train_starvla][INFO] - Step 4528, Loss: {'train/loss': 0.6718614101409912, 'train/grad_norm_pre_clip': 13.122966766357422, 'timing/data': 0.004825580050237477, 'timing/model': 1.1992845289641991, 'learning_rate/qwen_vl_interface': 3.2040963368057323e-06, 'learning_rate/action_model': 3.204096336805732e-05, 'epoch': 1.33}) +[2026-06-17 14:44:38,679][starVLA.training.train_starvla][INFO] - Step 4529, Loss: {'train/loss': 0.7576965093612671, 'train/grad_norm_pre_clip': 14.136503219604492, 'timing/data': 0.005273196962662041, 'timing/model': 1.200550744892098, 'learning_rate/qwen_vl_interface': 3.2021447482350994e-06, 'learning_rate/action_model': 3.202144748235099e-05, 'epoch': 1.33}) +[2026-06-17 14:44:43,499][starVLA.training.train_starvla][INFO] - Step 4530, Loss: {'train/loss': 0.5263433456420898, 'train/grad_norm_pre_clip': 12.619796752929688, 'timing/data': 0.006411586073227227, 'timing/model': 1.342690023011528, 'learning_rate/qwen_vl_interface': 3.200193584187061e-06, 'learning_rate/action_model': 3.200193584187061e-05, 'epoch': 1.33}) +[2026-06-17 14:44:48,137][starVLA.training.train_starvla][INFO] - Step 4531, Loss: {'train/loss': 0.8300673961639404, 'train/grad_norm_pre_clip': 16.208681106567383, 'timing/data': 0.005043857963755727, 'timing/model': 1.2040241210488603, 'learning_rate/qwen_vl_interface': 3.198242845066101e-06, 'learning_rate/action_model': 3.198242845066101e-05, 'epoch': 1.33}) +[2026-06-17 14:44:52,900][starVLA.training.train_starvla][INFO] - Step 4532, Loss: {'train/loss': 0.5239125490188599, 'train/grad_norm_pre_clip': 12.438128471374512, 'timing/data': 0.004696626914665103, 'timing/model': 1.1995924490038306, 'learning_rate/qwen_vl_interface': 3.1962925312766087e-06, 'learning_rate/action_model': 3.196292531276608e-05, 'epoch': 1.33}) +[2026-06-17 14:44:57,623][starVLA.training.train_starvla][INFO] - Step 4533, Loss: {'train/loss': 0.6041051149368286, 'train/grad_norm_pre_clip': 17.46820640563965, 'timing/data': 0.004822027054615319, 'timing/model': 1.2002286029746756, 'learning_rate/qwen_vl_interface': 3.19434264322288e-06, 'learning_rate/action_model': 3.19434264322288e-05, 'epoch': 1.33}) +[2026-06-17 14:45:02,337][starVLA.training.train_starvla][INFO] - Step 4534, Loss: {'train/loss': 0.34755781292915344, 'train/grad_norm_pre_clip': 9.853066444396973, 'timing/data': 0.004827090073376894, 'timing/model': 1.2003163089975715, 'learning_rate/qwen_vl_interface': 3.1923931813091364e-06, 'learning_rate/action_model': 3.192393181309136e-05, 'epoch': 1.33}) +[2026-06-17 14:45:07,101][starVLA.training.train_starvla][INFO] - Step 4535, Loss: {'train/loss': 0.7648720741271973, 'train/grad_norm_pre_clip': 17.654924392700195, 'timing/data': 0.005193014047108591, 'timing/model': 1.1987743970239535, 'learning_rate/qwen_vl_interface': 3.1904441459395e-06, 'learning_rate/action_model': 3.1904441459395e-05, 'epoch': 1.33}) +[2026-06-17 14:45:11,863][starVLA.training.train_starvla][INFO] - Step 4536, Loss: {'train/loss': 0.704096794128418, 'train/grad_norm_pre_clip': 12.638907432556152, 'timing/data': 0.005306396982632577, 'timing/model': 1.3281640160130337, 'learning_rate/qwen_vl_interface': 3.1884955375180088e-06, 'learning_rate/action_model': 3.1884955375180085e-05, 'epoch': 1.33}) +[2026-06-17 14:45:16,535][starVLA.training.train_starvla][INFO] - Step 4537, Loss: {'train/loss': 0.7321842908859253, 'train/grad_norm_pre_clip': 20.005807876586914, 'timing/data': 0.00489436904899776, 'timing/model': 1.1999454379547387, 'learning_rate/qwen_vl_interface': 3.18654735644861e-06, 'learning_rate/action_model': 3.1865473564486095e-05, 'epoch': 1.34}) +[2026-06-17 14:45:21,299][starVLA.training.train_starvla][INFO] - Step 4538, Loss: {'train/loss': 0.761410117149353, 'train/grad_norm_pre_clip': 16.818206787109375, 'timing/data': 0.005101494956761599, 'timing/model': 1.2000485359458253, 'learning_rate/qwen_vl_interface': 3.184599603135165e-06, 'learning_rate/action_model': 3.1845996031351646e-05, 'epoch': 1.34}) +[2026-06-17 14:45:26,103][starVLA.training.train_starvla][INFO] - Step 4539, Loss: {'train/loss': 0.5043495297431946, 'train/grad_norm_pre_clip': 10.469606399536133, 'timing/data': 0.005058557027950883, 'timing/model': 1.2006732439622283, 'learning_rate/qwen_vl_interface': 3.1826522779814453e-06, 'learning_rate/action_model': 3.182652277981445e-05, 'epoch': 1.34}) +[2026-06-17 14:45:30,831][starVLA.training.train_starvla][INFO] - Step 4540, Loss: {'train/loss': 0.5099633932113647, 'train/grad_norm_pre_clip': 11.215078353881836, 'timing/data': 0.005160691915079951, 'timing/model': 1.2003549649380147, 'learning_rate/qwen_vl_interface': 3.1807053813911317e-06, 'learning_rate/action_model': 3.180705381391132e-05, 'epoch': 1.34}) +[2026-06-17 14:45:35,605][starVLA.training.train_starvla][INFO] - Step 4541, Loss: {'train/loss': 0.849348783493042, 'train/grad_norm_pre_clip': 14.253705978393555, 'timing/data': 0.004612830933183432, 'timing/model': 1.2058219129685313, 'learning_rate/qwen_vl_interface': 3.17875891376782e-06, 'learning_rate/action_model': 3.1787589137678196e-05, 'epoch': 1.34}) +[2026-06-17 14:45:40,419][starVLA.training.train_starvla][INFO] - Step 4542, Loss: {'train/loss': 0.8592643737792969, 'train/grad_norm_pre_clip': 16.18702507019043, 'timing/data': 0.004790239967405796, 'timing/model': 1.329164318041876, 'learning_rate/qwen_vl_interface': 3.176812875515014e-06, 'learning_rate/action_model': 3.176812875515014e-05, 'epoch': 1.34}) +[2026-06-17 14:45:45,055][starVLA.training.train_starvla][INFO] - Step 4543, Loss: {'train/loss': 0.9421737194061279, 'train/grad_norm_pre_clip': 15.221944808959961, 'timing/data': 0.004958403995260596, 'timing/model': 1.200005596037954, 'learning_rate/qwen_vl_interface': 3.1748672670361304e-06, 'learning_rate/action_model': 3.17486726703613e-05, 'epoch': 1.34}) +[2026-06-17 14:45:49,815][starVLA.training.train_starvla][INFO] - Step 4544, Loss: {'train/loss': 0.8468600511550903, 'train/grad_norm_pre_clip': 13.1037015914917, 'timing/data': 0.004344155080616474, 'timing/model': 1.1992122760275379, 'learning_rate/qwen_vl_interface': 3.172922088734493e-06, 'learning_rate/action_model': 3.172922088734493e-05, 'epoch': 1.34}) +[2026-06-17 14:45:54,583][starVLA.training.train_starvla][INFO] - Step 4545, Loss: {'train/loss': 0.6027976870536804, 'train/grad_norm_pre_clip': 14.053553581237793, 'timing/data': 0.004934481927193701, 'timing/model': 1.200743678957224, 'learning_rate/qwen_vl_interface': 3.1709773410133434e-06, 'learning_rate/action_model': 3.170977341013343e-05, 'epoch': 1.34}) +[2026-06-17 14:45:59,346][starVLA.training.train_starvla][INFO] - Step 4546, Loss: {'train/loss': 0.757359504699707, 'train/grad_norm_pre_clip': 12.890771865844727, 'timing/data': 0.0051622079918161035, 'timing/model': 1.1986904030200094, 'learning_rate/qwen_vl_interface': 3.16903302427583e-06, 'learning_rate/action_model': 3.16903302427583e-05, 'epoch': 1.34}) +[2026-06-17 14:46:03,997][starVLA.training.train_starvla][INFO] - Step 4547, Loss: {'train/loss': 0.6529793739318848, 'train/grad_norm_pre_clip': 25.72031593322754, 'timing/data': 0.004865663941018283, 'timing/model': 1.2037196580786258, 'learning_rate/qwen_vl_interface': 3.1670891389250073e-06, 'learning_rate/action_model': 3.167089138925007e-05, 'epoch': 1.34}) +[2026-06-17 14:46:08,673][starVLA.training.train_starvla][INFO] - Step 4548, Loss: {'train/loss': 0.5479435920715332, 'train/grad_norm_pre_clip': 12.522913932800293, 'timing/data': 0.005085773067548871, 'timing/model': 1.23909123998601, 'learning_rate/qwen_vl_interface': 3.1651456853638464e-06, 'learning_rate/action_model': 3.165145685363846e-05, 'epoch': 1.34}) +[2026-06-17 14:46:13,347][starVLA.training.train_starvla][INFO] - Step 4549, Loss: {'train/loss': 0.5312860012054443, 'train/grad_norm_pre_clip': 14.79395580291748, 'timing/data': 0.0080188880674541, 'timing/model': 1.213606107980013, 'learning_rate/qwen_vl_interface': 3.1632026639952303e-06, 'learning_rate/action_model': 3.16320266399523e-05, 'epoch': 1.34}) +[2026-06-17 14:46:18,123][starVLA.training.train_starvla][INFO] - Step 4550, Loss: {'train/loss': 0.5458417534828186, 'train/grad_norm_pre_clip': 15.184490203857422, 'timing/data': 0.004919419996440411, 'timing/model': 1.1988728910218924, 'learning_rate/qwen_vl_interface': 3.161260075221947e-06, 'learning_rate/action_model': 3.161260075221947e-05, 'epoch': 1.34}) +[2026-06-17 14:46:23,007][starVLA.training.train_starvla][INFO] - Step 4551, Loss: {'train/loss': 0.6805269718170166, 'train/grad_norm_pre_clip': 15.968701362609863, 'timing/data': 0.0050651050405576825, 'timing/model': 1.2077536020660773, 'learning_rate/qwen_vl_interface': 3.1593179194467e-06, 'learning_rate/action_model': 3.1593179194467e-05, 'epoch': 1.34}) +[2026-06-17 14:46:27,813][starVLA.training.train_starvla][INFO] - Step 4552, Loss: {'train/loss': 0.907102108001709, 'train/grad_norm_pre_clip': 21.11417579650879, 'timing/data': 0.0049852520460262895, 'timing/model': 1.2074665470281616, 'learning_rate/qwen_vl_interface': 3.157376197072094e-06, 'learning_rate/action_model': 3.157376197072094e-05, 'epoch': 1.34}) +[2026-06-17 14:46:32,624][starVLA.training.train_starvla][INFO] - Step 4553, Loss: {'train/loss': 0.6656014919281006, 'train/grad_norm_pre_clip': 13.597497940063477, 'timing/data': 0.004562853020615876, 'timing/model': 1.3458439010428265, 'learning_rate/qwen_vl_interface': 3.1554349085006557e-06, 'learning_rate/action_model': 3.155434908500656e-05, 'epoch': 1.34}) +[2026-06-17 14:46:37,324][starVLA.training.train_starvla][INFO] - Step 4554, Loss: {'train/loss': 0.7537210583686829, 'train/grad_norm_pre_clip': 16.65498924255371, 'timing/data': 0.005298015079461038, 'timing/model': 1.2559525270480663, 'learning_rate/qwen_vl_interface': 3.1534940541348152e-06, 'learning_rate/action_model': 3.153494054134815e-05, 'epoch': 1.34}) +[2026-06-17 14:46:41,964][starVLA.training.train_starvla][INFO] - Step 4555, Loss: {'train/loss': 0.4069794714450836, 'train/grad_norm_pre_clip': 11.033255577087402, 'timing/data': 0.005150364013388753, 'timing/model': 1.2009873989736661, 'learning_rate/qwen_vl_interface': 3.1515536343769133e-06, 'learning_rate/action_model': 3.151553634376913e-05, 'epoch': 1.34}) +[2026-06-17 14:46:46,697][starVLA.training.train_starvla][INFO] - Step 4556, Loss: {'train/loss': 0.5922368168830872, 'train/grad_norm_pre_clip': 12.137199401855469, 'timing/data': 0.004784738062880933, 'timing/model': 1.198555475100875, 'learning_rate/qwen_vl_interface': 3.1496136496292e-06, 'learning_rate/action_model': 3.1496136496292e-05, 'epoch': 1.34}) +[2026-06-17 14:46:51,468][starVLA.training.train_starvla][INFO] - Step 4557, Loss: {'train/loss': 0.3461103141307831, 'train/grad_norm_pre_clip': 9.943382263183594, 'timing/data': 0.005009936052374542, 'timing/model': 1.1989649520255625, 'learning_rate/qwen_vl_interface': 3.1476741002938383e-06, 'learning_rate/action_model': 3.147674100293838e-05, 'epoch': 1.34}) +[2026-06-17 14:46:56,254][starVLA.training.train_starvla][INFO] - Step 4558, Loss: {'train/loss': 0.8328770995140076, 'train/grad_norm_pre_clip': 12.87827205657959, 'timing/data': 0.005073640961199999, 'timing/model': 1.1998493149876595, 'learning_rate/qwen_vl_interface': 3.145734986772897e-06, 'learning_rate/action_model': 3.1457349867728966e-05, 'epoch': 1.34}) +[2026-06-17 14:47:01,083][starVLA.training.train_starvla][INFO] - Step 4559, Loss: {'train/loss': 0.8279099464416504, 'train/grad_norm_pre_clip': 15.664874076843262, 'timing/data': 0.0048988849157467484, 'timing/model': 1.3511279430240393, 'learning_rate/qwen_vl_interface': 3.1437963094683566e-06, 'learning_rate/action_model': 3.1437963094683564e-05, 'epoch': 1.34}) +[2026-06-17 14:47:05,723][starVLA.training.train_starvla][INFO] - Step 4560, Loss: {'train/loss': 0.8971459865570068, 'train/grad_norm_pre_clip': 15.345252990722656, 'timing/data': 0.005105603951960802, 'timing/model': 1.2018769779242575, 'learning_rate/qwen_vl_interface': 3.1418580687821103e-06, 'learning_rate/action_model': 3.14185806878211e-05, 'epoch': 1.34}) +[2026-06-17 14:47:10,522][starVLA.training.train_starvla][INFO] - Step 4561, Loss: {'train/loss': 0.8552768230438232, 'train/grad_norm_pre_clip': 16.14451026916504, 'timing/data': 0.005168336909264326, 'timing/model': 1.203807899961248, 'learning_rate/qwen_vl_interface': 3.1399202651159532e-06, 'learning_rate/action_model': 3.1399202651159534e-05, 'epoch': 1.34}) +[2026-06-17 14:47:15,320][starVLA.training.train_starvla][INFO] - Step 4562, Loss: {'train/loss': 0.9600937366485596, 'train/grad_norm_pre_clip': 43.83567810058594, 'timing/data': 0.005010507069528103, 'timing/model': 1.2103400309570134, 'learning_rate/qwen_vl_interface': 3.1379828988715955e-06, 'learning_rate/action_model': 3.1379828988715956e-05, 'epoch': 1.34}) +[2026-06-17 14:47:20,119][starVLA.training.train_starvla][INFO] - Step 4563, Loss: {'train/loss': 0.7982578277587891, 'train/grad_norm_pre_clip': 16.226238250732422, 'timing/data': 0.0048823070246726274, 'timing/model': 1.2085574410157278, 'learning_rate/qwen_vl_interface': 3.136045970450654e-06, 'learning_rate/action_model': 3.136045970450654e-05, 'epoch': 1.34}) +[2026-06-17 14:47:24,883][starVLA.training.train_starvla][INFO] - Step 4564, Loss: {'train/loss': 0.7196195125579834, 'train/grad_norm_pre_clip': 16.679656982421875, 'timing/data': 0.005126707023009658, 'timing/model': 1.3275730179157108, 'learning_rate/qwen_vl_interface': 3.1341094802546595e-06, 'learning_rate/action_model': 3.1341094802546595e-05, 'epoch': 1.34}) +[2026-06-17 14:47:29,533][starVLA.training.train_starvla][INFO] - Step 4565, Loss: {'train/loss': 0.6882063150405884, 'train/grad_norm_pre_clip': 14.44819164276123, 'timing/data': 0.004995489027351141, 'timing/model': 1.201659233076498, 'learning_rate/qwen_vl_interface': 3.132173428685048e-06, 'learning_rate/action_model': 3.132173428685048e-05, 'epoch': 1.34}) +[2026-06-17 14:47:34,313][starVLA.training.train_starvla][INFO] - Step 4566, Loss: {'train/loss': 0.6066025495529175, 'train/grad_norm_pre_clip': 10.6947603225708, 'timing/data': 0.004992304020561278, 'timing/model': 1.2021472010528669, 'learning_rate/qwen_vl_interface': 3.13023781614316e-06, 'learning_rate/action_model': 3.13023781614316e-05, 'epoch': 1.34}) +[2026-06-17 14:47:39,077][starVLA.training.train_starvla][INFO] - Step 4567, Loss: {'train/loss': 0.7189315557479858, 'train/grad_norm_pre_clip': 12.577152252197266, 'timing/data': 0.005016458919271827, 'timing/model': 1.2017460589995608, 'learning_rate/qwen_vl_interface': 3.1283026430302553e-06, 'learning_rate/action_model': 3.128302643030255e-05, 'epoch': 1.34}) +[2026-06-17 14:47:43,916][starVLA.training.train_starvla][INFO] - Step 4568, Loss: {'train/loss': 0.43268659710884094, 'train/grad_norm_pre_clip': 9.64589786529541, 'timing/data': 0.004982020007446408, 'timing/model': 1.202504850924015, 'learning_rate/qwen_vl_interface': 3.1263679097474958e-06, 'learning_rate/action_model': 3.126367909747496e-05, 'epoch': 1.34}) +[2026-06-17 14:47:48,681][starVLA.training.train_starvla][INFO] - Step 4569, Loss: {'train/loss': 0.6201839447021484, 'train/grad_norm_pre_clip': 15.282706260681152, 'timing/data': 0.005159056046977639, 'timing/model': 1.1988918730057776, 'learning_rate/qwen_vl_interface': 3.124433616695955e-06, 'learning_rate/action_model': 3.1244336166959545e-05, 'epoch': 1.34}) +[2026-06-17 14:47:53,445][starVLA.training.train_starvla][INFO] - Step 4570, Loss: {'train/loss': 0.6943594217300415, 'train/grad_norm_pre_clip': 11.954543113708496, 'timing/data': 0.005220268038101494, 'timing/model': 1.3268432919867337, 'learning_rate/qwen_vl_interface': 3.1224997642766115e-06, 'learning_rate/action_model': 3.122499764276611e-05, 'epoch': 1.34}) +[2026-06-17 14:47:58,170][starVLA.training.train_starvla][INFO] - Step 4571, Loss: {'train/loss': 0.5025497078895569, 'train/grad_norm_pre_clip': 12.508626937866211, 'timing/data': 0.004974627983756363, 'timing/model': 1.2831403409363702, 'learning_rate/qwen_vl_interface': 3.1205663528903563e-06, 'learning_rate/action_model': 3.1205663528903565e-05, 'epoch': 1.35}) +[2026-06-17 14:48:02,816][starVLA.training.train_starvla][INFO] - Step 4572, Loss: {'train/loss': 0.6034862399101257, 'train/grad_norm_pre_clip': 10.670796394348145, 'timing/data': 0.0053077139891684055, 'timing/model': 1.2058222469640896, 'learning_rate/qwen_vl_interface': 3.118633382937987e-06, 'learning_rate/action_model': 3.118633382937987e-05, 'epoch': 1.35}) +[2026-06-17 14:48:07,604][starVLA.training.train_starvla][INFO] - Step 4573, Loss: {'train/loss': 0.5878654718399048, 'train/grad_norm_pre_clip': 10.304768562316895, 'timing/data': 0.004541298025287688, 'timing/model': 1.2032818900188431, 'learning_rate/qwen_vl_interface': 3.116700854820209e-06, 'learning_rate/action_model': 3.116700854820209e-05, 'epoch': 1.35}) +[2026-06-17 14:48:12,412][starVLA.training.train_starvla][INFO] - Step 4574, Loss: {'train/loss': 0.9988563656806946, 'train/grad_norm_pre_clip': 18.69713020324707, 'timing/data': 0.004861723980866373, 'timing/model': 1.207134563010186, 'learning_rate/qwen_vl_interface': 3.1147687689376443e-06, 'learning_rate/action_model': 3.1147687689376444e-05, 'epoch': 1.35}) +[2026-06-17 14:48:17,210][starVLA.training.train_starvla][INFO] - Step 4575, Loss: {'train/loss': 0.4659076929092407, 'train/grad_norm_pre_clip': 14.089515686035156, 'timing/data': 0.005212420015595853, 'timing/model': 1.2056369649944827, 'learning_rate/qwen_vl_interface': 3.1128371256908087e-06, 'learning_rate/action_model': 3.112837125690809e-05, 'epoch': 1.35}) +[2026-06-17 14:48:21,984][starVLA.training.train_starvla][INFO] - Step 4576, Loss: {'train/loss': 0.7280271053314209, 'train/grad_norm_pre_clip': 13.761006355285645, 'timing/data': 0.005283119040541351, 'timing/model': 1.2014738540165126, 'learning_rate/qwen_vl_interface': 3.1109059254801368e-06, 'learning_rate/action_model': 3.1109059254801366e-05, 'epoch': 1.35}) +[2026-06-17 14:48:26,753][starVLA.training.train_starvla][INFO] - Step 4577, Loss: {'train/loss': 0.932175874710083, 'train/grad_norm_pre_clip': 14.708407402038574, 'timing/data': 0.004826203919947147, 'timing/model': 1.2886255680350587, 'learning_rate/qwen_vl_interface': 3.108975168705966e-06, 'learning_rate/action_model': 3.1089751687059656e-05, 'epoch': 1.35}) +[2026-06-17 14:48:31,422][starVLA.training.train_starvla][INFO] - Step 4578, Loss: {'train/loss': 0.7699066996574402, 'train/grad_norm_pre_clip': 14.005823135375977, 'timing/data': 0.005110602010972798, 'timing/model': 1.2059299639659002, 'learning_rate/qwen_vl_interface': 3.107044855768549e-06, 'learning_rate/action_model': 3.107044855768549e-05, 'epoch': 1.35}) +[2026-06-17 14:48:36,220][starVLA.training.train_starvla][INFO] - Step 4579, Loss: {'train/loss': 0.7501882314682007, 'train/grad_norm_pre_clip': 13.981945037841797, 'timing/data': 0.004771930049173534, 'timing/model': 1.206177073996514, 'learning_rate/qwen_vl_interface': 3.1051149870680413e-06, 'learning_rate/action_model': 3.105114987068041e-05, 'epoch': 1.35}) +[2026-06-17 14:48:41,003][starVLA.training.train_starvla][INFO] - Step 4580, Loss: {'train/loss': 0.6704610586166382, 'train/grad_norm_pre_clip': 13.197080612182617, 'timing/data': 0.004699963028542697, 'timing/model': 1.2004498390015215, 'learning_rate/qwen_vl_interface': 3.103185563004503e-06, 'learning_rate/action_model': 3.103185563004503e-05, 'epoch': 1.35}) +[2026-06-17 14:48:45,772][starVLA.training.train_starvla][INFO] - Step 4581, Loss: {'train/loss': 0.3276972472667694, 'train/grad_norm_pre_clip': 9.071663856506348, 'timing/data': 0.005167334922589362, 'timing/model': 1.2020746899070218, 'learning_rate/qwen_vl_interface': 3.101256583977905e-06, 'learning_rate/action_model': 3.101256583977905e-05, 'epoch': 1.35}) +[2026-06-17 14:48:50,556][starVLA.training.train_starvla][INFO] - Step 4582, Loss: {'train/loss': 0.8527514934539795, 'train/grad_norm_pre_clip': 36.79595184326172, 'timing/data': 0.004739488009363413, 'timing/model': 1.345430689980276, 'learning_rate/qwen_vl_interface': 3.099328050388131e-06, 'learning_rate/action_model': 3.0993280503881305e-05, 'epoch': 1.35}) +[2026-06-17 14:48:55,211][starVLA.training.train_starvla][INFO] - Step 4583, Loss: {'train/loss': 0.694814920425415, 'train/grad_norm_pre_clip': 15.502193450927734, 'timing/data': 0.005113988998346031, 'timing/model': 1.205396326025948, 'learning_rate/qwen_vl_interface': 3.0973999626349667e-06, 'learning_rate/action_model': 3.097399962634966e-05, 'epoch': 1.35}) +[2026-06-17 14:49:00,001][starVLA.training.train_starvla][INFO] - Step 4584, Loss: {'train/loss': 0.5539842247962952, 'train/grad_norm_pre_clip': 10.980039596557617, 'timing/data': 0.005347648984752595, 'timing/model': 1.2033343970542774, 'learning_rate/qwen_vl_interface': 3.095472321118107e-06, 'learning_rate/action_model': 3.095472321118107e-05, 'epoch': 1.35}) +[2026-06-17 14:49:04,788][starVLA.training.train_starvla][INFO] - Step 4585, Loss: {'train/loss': 0.8017574548721313, 'train/grad_norm_pre_clip': 12.943931579589844, 'timing/data': 0.004831333993934095, 'timing/model': 1.202003534999676, 'learning_rate/qwen_vl_interface': 3.093545126237153e-06, 'learning_rate/action_model': 3.0935451262371527e-05, 'epoch': 1.35}) +[2026-06-17 14:49:09,494][starVLA.training.train_starvla][INFO] - Step 4586, Loss: {'train/loss': 0.5188822746276855, 'train/grad_norm_pre_clip': 11.390321731567383, 'timing/data': 0.005310242995619774, 'timing/model': 1.2025622599758208, 'learning_rate/qwen_vl_interface': 3.091618378391615e-06, 'learning_rate/action_model': 3.0916183783916146e-05, 'epoch': 1.35}) +[2026-06-17 14:49:14,228][starVLA.training.train_starvla][INFO] - Step 4587, Loss: {'train/loss': 0.6174924969673157, 'train/grad_norm_pre_clip': 12.803166389465332, 'timing/data': 0.005131812999024987, 'timing/model': 1.2013991479761899, 'learning_rate/qwen_vl_interface': 3.0896920779809097e-06, 'learning_rate/action_model': 3.08969207798091e-05, 'epoch': 1.35}) +[2026-06-17 14:49:19,018][starVLA.training.train_starvla][INFO] - Step 4588, Loss: {'train/loss': 0.654114305973053, 'train/grad_norm_pre_clip': 13.76892375946045, 'timing/data': 0.005034556030295789, 'timing/model': 1.342958325985819, 'learning_rate/qwen_vl_interface': 3.0877662254043605e-06, 'learning_rate/action_model': 3.0877662254043606e-05, 'epoch': 1.35}) +[2026-06-17 14:49:23,684][starVLA.training.train_starvla][INFO] - Step 4589, Loss: {'train/loss': 0.6847515106201172, 'train/grad_norm_pre_clip': 15.169672012329102, 'timing/data': 0.004705338040366769, 'timing/model': 1.2031489149667323, 'learning_rate/qwen_vl_interface': 3.0858408210612e-06, 'learning_rate/action_model': 3.0858408210612e-05, 'epoch': 1.35}) +[2026-06-17 14:49:28,492][starVLA.training.train_starvla][INFO] - Step 4590, Loss: {'train/loss': 0.5690351128578186, 'train/grad_norm_pre_clip': 13.686216354370117, 'timing/data': 0.005034209927543998, 'timing/model': 1.20168854505755, 'learning_rate/qwen_vl_interface': 3.0839158653505663e-06, 'learning_rate/action_model': 3.0839158653505664e-05, 'epoch': 1.35}) +[2026-06-17 14:49:33,286][starVLA.training.train_starvla][INFO] - Step 4591, Loss: {'train/loss': 0.5434517860412598, 'train/grad_norm_pre_clip': 12.09857177734375, 'timing/data': 0.005066363955847919, 'timing/model': 1.2054401990026236, 'learning_rate/qwen_vl_interface': 3.081991358671504e-06, 'learning_rate/action_model': 3.081991358671504e-05, 'epoch': 1.35}) +[2026-06-17 14:49:38,046][starVLA.training.train_starvla][INFO] - Step 4592, Loss: {'train/loss': 0.26824018359184265, 'train/grad_norm_pre_clip': 8.787102699279785, 'timing/data': 0.004845252027735114, 'timing/model': 1.2009200169704854, 'learning_rate/qwen_vl_interface': 3.080067301422965e-06, 'learning_rate/action_model': 3.080067301422965e-05, 'epoch': 1.35}) +[2026-06-17 14:49:42,782][starVLA.training.train_starvla][INFO] - Step 4593, Loss: {'train/loss': 0.6814707517623901, 'train/grad_norm_pre_clip': 19.92432975769043, 'timing/data': 0.005103195086121559, 'timing/model': 1.202465252019465, 'learning_rate/qwen_vl_interface': 3.07814369400381e-06, 'learning_rate/action_model': 3.07814369400381e-05, 'epoch': 1.35}) +[2026-06-17 14:49:47,538][starVLA.training.train_starvla][INFO] - Step 4594, Loss: {'train/loss': 0.5227689743041992, 'train/grad_norm_pre_clip': 12.400850296020508, 'timing/data': 0.005107274046167731, 'timing/model': 1.3166620440315455, 'learning_rate/qwen_vl_interface': 3.076220536812807e-06, 'learning_rate/action_model': 3.076220536812807e-05, 'epoch': 1.35}) +[2026-06-17 14:49:52,184][starVLA.training.train_starvla][INFO] - Step 4595, Loss: {'train/loss': 0.7435531616210938, 'train/grad_norm_pre_clip': 13.149508476257324, 'timing/data': 0.004870936973020434, 'timing/model': 1.2048684109468013, 'learning_rate/qwen_vl_interface': 3.0742978302486214e-06, 'learning_rate/action_model': 3.074297830248621e-05, 'epoch': 1.35}) +[2026-06-17 14:49:56,955][starVLA.training.train_starvla][INFO] - Step 4596, Loss: {'train/loss': 0.9118659496307373, 'train/grad_norm_pre_clip': 16.213829040527344, 'timing/data': 0.004849680932238698, 'timing/model': 1.2013897619908676, 'learning_rate/qwen_vl_interface': 3.0723755747098387e-06, 'learning_rate/action_model': 3.072375574709839e-05, 'epoch': 1.35}) +[2026-06-17 14:50:01,735][starVLA.training.train_starvla][INFO] - Step 4597, Loss: {'train/loss': 0.6447228193283081, 'train/grad_norm_pre_clip': 14.3947114944458, 'timing/data': 0.004952571936883032, 'timing/model': 1.2060517350910231, 'learning_rate/qwen_vl_interface': 3.070453770594943e-06, 'learning_rate/action_model': 3.070453770594943e-05, 'epoch': 1.35}) +[2026-06-17 14:50:06,507][starVLA.training.train_starvla][INFO] - Step 4598, Loss: {'train/loss': 0.46420490741729736, 'train/grad_norm_pre_clip': 10.606911659240723, 'timing/data': 0.005169676966033876, 'timing/model': 1.2025828559417278, 'learning_rate/qwen_vl_interface': 3.0685324183023264e-06, 'learning_rate/action_model': 3.068532418302326e-05, 'epoch': 1.35}) +[2026-06-17 14:50:11,293][starVLA.training.train_starvla][INFO] - Step 4599, Loss: {'train/loss': 0.5744057893753052, 'train/grad_norm_pre_clip': 12.28984546661377, 'timing/data': 0.005248886998742819, 'timing/model': 1.3427730050170794, 'learning_rate/qwen_vl_interface': 3.0666115182302863e-06, 'learning_rate/action_model': 3.0666115182302865e-05, 'epoch': 1.35}) +[2026-06-17 14:51:52,602][starVLA.training.train_starvla][INFO] - Step 4600, Loss: {'train/loss': 0.6252225041389465, 'train/grad_norm_pre_clip': 13.344557762145996, 'eval/loss': 0.7151200175285339, 'eval/latency_0/loss': 0.7151200175285339, 'eval/demon_fire_precision': 0.9052460349735665, 'eval/demon_fire_recall': 0.9244186046511628, 'eval/demon_fire_f1': 0.9147318676802959, 'eval/demon_fire_rate_model': 0.6838153503893215, 'eval/demon_fire_rate_teacher': 0.6696329254727474, 'eval/demon_left_precision': 0.8237986270022883, 'eval/demon_left_recall': 0.8158640226628895, 'eval/demon_left_f1': 0.8198121263877028, 'eval/demon_left_rate_model': 0.4860956618464961, 'eval/demon_left_rate_teacher': 0.49082313681868744, 'eval/demon_right_precision': 0.8206253428414702, 'eval/demon_right_recall': 0.8451977401129943, 'eval/demon_right_f1': 0.8327303089340384, 'eval/demon_right_rate_model': 0.506952169076752, 'eval/demon_right_rate_teacher': 0.4922135706340378, 'eval/demon_move_f1': 0.8262712176608706, 'eval/demon_cc_f1': 0.8862667354690593, 'eval/latency_0/demon_fire_precision': 0.9052460349735665, 'eval/latency_0/demon_fire_recall': 0.9244186046511628, 'eval/latency_0/demon_fire_f1': 0.9147318676802959, 'eval/latency_0/demon_fire_rate_model': 0.6838153503893215, 'eval/latency_0/demon_fire_rate_teacher': 0.6696329254727474, 'eval/latency_0/demon_left_precision': 0.8237986270022883, 'eval/latency_0/demon_left_recall': 0.8158640226628895, 'eval/latency_0/demon_left_f1': 0.8198121263877028, 'eval/latency_0/demon_left_rate_model': 0.4860956618464961, 'eval/latency_0/demon_left_rate_teacher': 0.49082313681868744, 'eval/latency_0/demon_right_precision': 0.8206253428414702, 'eval/latency_0/demon_right_recall': 0.8451977401129943, 'eval/latency_0/demon_right_f1': 0.8327303089340384, 'eval/latency_0/demon_right_rate_model': 0.506952169076752, 'eval/latency_0/demon_right_rate_teacher': 0.4922135706340378, 'eval/latency_0/demon_move_f1': 0.8262712176608706, 'eval/latency_0/demon_cc_f1': 0.8862667354690593, 'timing/data': 0.004763845936395228, 'timing/model': 1.202075829030946, 'learning_rate/qwen_vl_interface': 3.064691070777027e-06, 'learning_rate/action_model': 3.064691070777027e-05, 'epoch': 1.35}) +[2026-06-17 14:51:52,603][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_4600_state +[2026-06-17 14:51:52,605][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 14:51:56,167][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_4600_state/model.safetensors +[2026-06-17 14:52:03,728][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_4600_state/optimizer.bin +[2026-06-17 14:52:03,728][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_4600_state/sampler.bin +[2026-06-17 14:52:03,728][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_4600_state/sampler_1.bin +[2026-06-17 14:52:03,729][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_4600_state/random_states_0.pkl +[2026-06-17 14:52:03,777][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4600 +[2026-06-17 14:52:03,777][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 14:52:03,791][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 14:52:08,512][starVLA.training.train_starvla][INFO] - Step 4601, Loss: {'train/loss': 0.6808742880821228, 'train/grad_norm_pre_clip': 14.28300952911377, 'timing/data': 0.005449683987535536, 'timing/model': 1.3309830439975485, 'learning_rate/qwen_vl_interface': 3.062771076340662e-06, 'learning_rate/action_model': 3.0627710763406617e-05, 'epoch': 1.35}) +[2026-06-17 14:52:13,112][starVLA.training.train_starvla][INFO] - Step 4602, Loss: {'train/loss': 0.7159625887870789, 'train/grad_norm_pre_clip': 12.192911148071289, 'timing/data': 0.00476400600746274, 'timing/model': 1.1905595579883084, 'learning_rate/qwen_vl_interface': 3.0608515353192046e-06, 'learning_rate/action_model': 3.060851535319205e-05, 'epoch': 1.35}) +[2026-06-17 14:52:17,840][starVLA.training.train_starvla][INFO] - Step 4603, Loss: {'train/loss': 0.7843990325927734, 'train/grad_norm_pre_clip': 14.24763298034668, 'timing/data': 0.004545864067040384, 'timing/model': 1.1918912379769608, 'learning_rate/qwen_vl_interface': 3.0589324481105814e-06, 'learning_rate/action_model': 3.058932448110581e-05, 'epoch': 1.35}) +[2026-06-17 14:52:22,575][starVLA.training.train_starvla][INFO] - Step 4604, Loss: {'train/loss': 0.632167398929596, 'train/grad_norm_pre_clip': 10.97857666015625, 'timing/data': 0.004759373026899993, 'timing/model': 1.1921013569226488, 'learning_rate/qwen_vl_interface': 3.0570138151126185e-06, 'learning_rate/action_model': 3.057013815112618e-05, 'epoch': 1.35}) +[2026-06-17 14:52:27,318][starVLA.training.train_starvla][INFO] - Step 4605, Loss: {'train/loss': 0.5862838625907898, 'train/grad_norm_pre_clip': 17.787349700927734, 'timing/data': 0.004922752967104316, 'timing/model': 1.1954804740380496, 'learning_rate/qwen_vl_interface': 3.0550956367230506e-06, 'learning_rate/action_model': 3.05509563672305e-05, 'epoch': 1.36}) +[2026-06-17 14:52:31,952][starVLA.training.train_starvla][INFO] - Step 4606, Loss: {'train/loss': 0.4618719816207886, 'train/grad_norm_pre_clip': 11.377761840820312, 'timing/data': 0.0048399289371445775, 'timing/model': 1.194264902966097, 'learning_rate/qwen_vl_interface': 3.0531779133395184e-06, 'learning_rate/action_model': 3.053177913339518e-05, 'epoch': 1.36}) +[2026-06-17 14:52:36,615][starVLA.training.train_starvla][INFO] - Step 4607, Loss: {'train/loss': 0.7250977754592896, 'train/grad_norm_pre_clip': 13.59099006652832, 'timing/data': 0.004771897103637457, 'timing/model': 1.2323584300465882, 'learning_rate/qwen_vl_interface': 3.0512606453595708e-06, 'learning_rate/action_model': 3.0512606453595704e-05, 'epoch': 1.36}) +[2026-06-17 14:52:41,246][starVLA.training.train_starvla][INFO] - Step 4608, Loss: {'train/loss': 0.6125873327255249, 'train/grad_norm_pre_clip': 12.036789894104004, 'timing/data': 0.004910326097160578, 'timing/model': 1.2030497269006446, 'learning_rate/qwen_vl_interface': 3.049343833180658e-06, 'learning_rate/action_model': 3.049343833180658e-05, 'epoch': 1.36}) +[2026-06-17 14:52:45,946][starVLA.training.train_starvla][INFO] - Step 4609, Loss: {'train/loss': 0.6432230472564697, 'train/grad_norm_pre_clip': 13.697343826293945, 'timing/data': 0.004785360069945455, 'timing/model': 1.1993933010380715, 'learning_rate/qwen_vl_interface': 3.0474274772001355e-06, 'learning_rate/action_model': 3.0474274772001355e-05, 'epoch': 1.36}) +[2026-06-17 14:52:50,669][starVLA.training.train_starvla][INFO] - Step 4610, Loss: {'train/loss': 0.5864470601081848, 'train/grad_norm_pre_clip': 11.287631034851074, 'timing/data': 0.004768998012878001, 'timing/model': 1.1967058320296928, 'learning_rate/qwen_vl_interface': 3.045511577815265e-06, 'learning_rate/action_model': 3.045511577815265e-05, 'epoch': 1.36}) +[2026-06-17 14:52:55,432][starVLA.training.train_starvla][INFO] - Step 4611, Loss: {'train/loss': 0.6536014676094055, 'train/grad_norm_pre_clip': 13.04488754272461, 'timing/data': 0.004670549067668617, 'timing/model': 1.198393122991547, 'learning_rate/qwen_vl_interface': 3.043596135423218e-06, 'learning_rate/action_model': 3.043596135423218e-05, 'epoch': 1.36}) +[2026-06-17 14:53:00,197][starVLA.training.train_starvla][INFO] - Step 4612, Loss: {'train/loss': 0.6625280976295471, 'train/grad_norm_pre_clip': 11.018975257873535, 'timing/data': 0.0054455900099128485, 'timing/model': 1.3274078889517114, 'learning_rate/qwen_vl_interface': 3.041681150421067e-06, 'learning_rate/action_model': 3.041681150421067e-05, 'epoch': 1.36}) +[2026-06-17 14:53:04,842][starVLA.training.train_starvla][INFO] - Step 4613, Loss: {'train/loss': 0.5935923457145691, 'train/grad_norm_pre_clip': 14.478240966796875, 'timing/data': 0.004742200020700693, 'timing/model': 1.2026513440068811, 'learning_rate/qwen_vl_interface': 3.039766623205791e-06, 'learning_rate/action_model': 3.039766623205791e-05, 'epoch': 1.36}) +[2026-06-17 14:53:09,606][starVLA.training.train_starvla][INFO] - Step 4614, Loss: {'train/loss': 0.8539884686470032, 'train/grad_norm_pre_clip': 17.422103881835938, 'timing/data': 0.004950081929564476, 'timing/model': 1.2006394519703463, 'learning_rate/qwen_vl_interface': 3.037852554174269e-06, 'learning_rate/action_model': 3.0378525541742692e-05, 'epoch': 1.36}) +[2026-06-17 14:53:14,334][starVLA.training.train_starvla][INFO] - Step 4615, Loss: {'train/loss': 0.6654645204544067, 'train/grad_norm_pre_clip': 15.224853515625, 'timing/data': 0.0044540270464494824, 'timing/model': 1.2009532318916172, 'learning_rate/qwen_vl_interface': 3.035938943723295e-06, 'learning_rate/action_model': 3.035938943723295e-05, 'epoch': 1.36}) +[2026-06-17 14:53:19,109][starVLA.training.train_starvla][INFO] - Step 4616, Loss: {'train/loss': 0.6901125907897949, 'train/grad_norm_pre_clip': 15.454000473022461, 'timing/data': 0.004907267051748931, 'timing/model': 1.2020787000656128, 'learning_rate/qwen_vl_interface': 3.03402579224956e-06, 'learning_rate/action_model': 3.03402579224956e-05, 'epoch': 1.36}) +[2026-06-17 14:53:23,885][starVLA.training.train_starvla][INFO] - Step 4617, Loss: {'train/loss': 0.47357746958732605, 'train/grad_norm_pre_clip': 11.385449409484863, 'timing/data': 0.005101785063743591, 'timing/model': 1.2004124199738726, 'learning_rate/qwen_vl_interface': 3.032113100149663e-06, 'learning_rate/action_model': 3.0321131001496628e-05, 'epoch': 1.36}) +[2026-06-17 14:53:28,655][starVLA.training.train_starvla][INFO] - Step 4618, Loss: {'train/loss': 0.5435734391212463, 'train/grad_norm_pre_clip': 10.556477546691895, 'timing/data': 0.004966886946931481, 'timing/model': 1.327793598989956, 'learning_rate/qwen_vl_interface': 3.0302008678201067e-06, 'learning_rate/action_model': 3.0302008678201067e-05, 'epoch': 1.36}) +[2026-06-17 14:53:33,376][starVLA.training.train_starvla][INFO] - Step 4619, Loss: {'train/loss': 0.5400896668434143, 'train/grad_norm_pre_clip': 14.123525619506836, 'timing/data': 0.004617388010956347, 'timing/model': 1.2714780980022624, 'learning_rate/qwen_vl_interface': 3.028289095657299e-06, 'learning_rate/action_model': 3.028289095657299e-05, 'epoch': 1.36}) +[2026-06-17 14:53:38,028][starVLA.training.train_starvla][INFO] - Step 4620, Loss: {'train/loss': 0.5721479058265686, 'train/grad_norm_pre_clip': 14.418109893798828, 'timing/data': 0.005192331969738007, 'timing/model': 1.2063406989909708, 'learning_rate/qwen_vl_interface': 3.026377784057552e-06, 'learning_rate/action_model': 3.026377784057552e-05, 'epoch': 1.36}) +[2026-06-17 14:53:42,762][starVLA.training.train_starvla][INFO] - Step 4621, Loss: {'train/loss': 0.712051272392273, 'train/grad_norm_pre_clip': 13.250938415527344, 'timing/data': 0.00482034299056977, 'timing/model': 1.2024511020863429, 'learning_rate/qwen_vl_interface': 3.0244669334170822e-06, 'learning_rate/action_model': 3.0244669334170822e-05, 'epoch': 1.36}) +[2026-06-17 14:53:47,542][starVLA.training.train_starvla][INFO] - Step 4622, Loss: {'train/loss': 0.50382399559021, 'train/grad_norm_pre_clip': 10.101790428161621, 'timing/data': 0.005065025994554162, 'timing/model': 1.2041576419724151, 'learning_rate/qwen_vl_interface': 3.0225565441320156e-06, 'learning_rate/action_model': 3.0225565441320152e-05, 'epoch': 1.36}) +[2026-06-17 14:53:52,329][starVLA.training.train_starvla][INFO] - Step 4623, Loss: {'train/loss': 0.5785543918609619, 'train/grad_norm_pre_clip': 13.749175071716309, 'timing/data': 0.004770500003360212, 'timing/model': 1.2034357619704679, 'learning_rate/qwen_vl_interface': 3.0206466165983716e-06, 'learning_rate/action_model': 3.0206466165983716e-05, 'epoch': 1.36}) +[2026-06-17 14:53:57,106][starVLA.training.train_starvla][INFO] - Step 4624, Loss: {'train/loss': 0.54579758644104, 'train/grad_norm_pre_clip': 12.794148445129395, 'timing/data': 0.004855267005041242, 'timing/model': 1.3319686889881268, 'learning_rate/qwen_vl_interface': 3.0187371512120833e-06, 'learning_rate/action_model': 3.0187371512120832e-05, 'epoch': 1.36}) +[2026-06-17 14:54:01,893][starVLA.training.train_starvla][INFO] - Step 4625, Loss: {'train/loss': 0.5835058689117432, 'train/grad_norm_pre_clip': 12.366890907287598, 'timing/data': 0.005081114009954035, 'timing/model': 1.337158155045472, 'learning_rate/qwen_vl_interface': 3.0168281483689825e-06, 'learning_rate/action_model': 3.0168281483689825e-05, 'epoch': 1.36}) +[2026-06-17 14:54:06,542][starVLA.training.train_starvla][INFO] - Step 4626, Loss: {'train/loss': 0.6099666357040405, 'train/grad_norm_pre_clip': 12.814229965209961, 'timing/data': 0.004795373999513686, 'timing/model': 1.2045299790333956, 'learning_rate/qwen_vl_interface': 3.014919608464812e-06, 'learning_rate/action_model': 3.0149196084648117e-05, 'epoch': 1.36}) +[2026-06-17 14:54:11,292][starVLA.training.train_starvla][INFO] - Step 4627, Loss: {'train/loss': 0.7805789709091187, 'train/grad_norm_pre_clip': 27.277637481689453, 'timing/data': 0.004743068013340235, 'timing/model': 1.2034786120057106, 'learning_rate/qwen_vl_interface': 3.0130115318952124e-06, 'learning_rate/action_model': 3.0130115318952122e-05, 'epoch': 1.36}) +[2026-06-17 14:54:16,081][starVLA.training.train_starvla][INFO] - Step 4628, Loss: {'train/loss': 0.6790346503257751, 'train/grad_norm_pre_clip': 16.887304306030273, 'timing/data': 0.004753827000968158, 'timing/model': 1.2041576389456168, 'learning_rate/qwen_vl_interface': 3.011103919055726e-06, 'learning_rate/action_model': 3.011103919055726e-05, 'epoch': 1.36}) +[2026-06-17 14:54:20,862][starVLA.training.train_starvla][INFO] - Step 4629, Loss: {'train/loss': 0.3846093714237213, 'train/grad_norm_pre_clip': 11.04023265838623, 'timing/data': 0.004359477083198726, 'timing/model': 1.2041202059481293, 'learning_rate/qwen_vl_interface': 3.009196770341809e-06, 'learning_rate/action_model': 3.009196770341809e-05, 'epoch': 1.36}) +[2026-06-17 14:54:25,592][starVLA.training.train_starvla][INFO] - Step 4630, Loss: {'train/loss': 0.7240696549415588, 'train/grad_norm_pre_clip': 15.139488220214844, 'timing/data': 0.0048223050544038415, 'timing/model': 1.276119922986254, 'learning_rate/qwen_vl_interface': 3.0072900861488125e-06, 'learning_rate/action_model': 3.0072900861488125e-05, 'epoch': 1.36}) +[2026-06-17 14:54:30,253][starVLA.training.train_starvla][INFO] - Step 4631, Loss: {'train/loss': 0.8668068647384644, 'train/grad_norm_pre_clip': 16.012096405029297, 'timing/data': 0.004878108971752226, 'timing/model': 1.211021622060798, 'learning_rate/qwen_vl_interface': 3.0053838668719944e-06, 'learning_rate/action_model': 3.0053838668719943e-05, 'epoch': 1.36}) +[2026-06-17 14:54:35,008][starVLA.training.train_starvla][INFO] - Step 4632, Loss: {'train/loss': 0.8490049839019775, 'train/grad_norm_pre_clip': 19.158815383911133, 'timing/data': 0.005056484020315111, 'timing/model': 1.201806370052509, 'learning_rate/qwen_vl_interface': 3.003478112906515e-06, 'learning_rate/action_model': 3.0034781129065148e-05, 'epoch': 1.36}) +[2026-06-17 14:54:39,791][starVLA.training.train_starvla][INFO] - Step 4633, Loss: {'train/loss': 0.6879016757011414, 'train/grad_norm_pre_clip': 12.248042106628418, 'timing/data': 0.005067032063379884, 'timing/model': 1.2008372698910534, 'learning_rate/qwen_vl_interface': 3.0015728246474405e-06, 'learning_rate/action_model': 3.0015728246474405e-05, 'epoch': 1.36}) +[2026-06-17 14:54:44,570][starVLA.training.train_starvla][INFO] - Step 4634, Loss: {'train/loss': 0.46185532212257385, 'train/grad_norm_pre_clip': 11.303497314453125, 'timing/data': 0.004484952078200877, 'timing/model': 1.2050079250475392, 'learning_rate/qwen_vl_interface': 2.99966800248974e-06, 'learning_rate/action_model': 2.99966800248974e-05, 'epoch': 1.36}) +[2026-06-17 14:54:49,349][starVLA.training.train_starvla][INFO] - Step 4635, Loss: {'train/loss': 0.5086835622787476, 'train/grad_norm_pre_clip': 11.955694198608398, 'timing/data': 0.004736906033940613, 'timing/model': 1.3277788709383458, 'learning_rate/qwen_vl_interface': 2.9977636468282817e-06, 'learning_rate/action_model': 2.9977636468282816e-05, 'epoch': 1.36}) +[2026-06-17 14:54:54,089][starVLA.training.train_starvla][INFO] - Step 4636, Loss: {'train/loss': 0.4875682592391968, 'train/grad_norm_pre_clip': 13.374493598937988, 'timing/data': 0.004896805970929563, 'timing/model': 1.2869107209844515, 'learning_rate/qwen_vl_interface': 2.995859758057846e-06, 'learning_rate/action_model': 2.9958597580578456e-05, 'epoch': 1.36}) +[2026-06-17 14:54:58,743][starVLA.training.train_starvla][INFO] - Step 4637, Loss: {'train/loss': 0.8865916728973389, 'train/grad_norm_pre_clip': 15.580046653747559, 'timing/data': 0.0045197439612820745, 'timing/model': 1.2068123549688607, 'learning_rate/qwen_vl_interface': 2.993956336573106e-06, 'learning_rate/action_model': 2.9939563365731062e-05, 'epoch': 1.36}) +[2026-06-17 14:55:03,525][starVLA.training.train_starvla][INFO] - Step 4638, Loss: {'train/loss': 0.7611806988716125, 'train/grad_norm_pre_clip': 12.925676345825195, 'timing/data': 0.005265558953396976, 'timing/model': 1.2009397789370269, 'learning_rate/qwen_vl_interface': 2.9920533827686453e-06, 'learning_rate/action_model': 2.9920533827686453e-05, 'epoch': 1.36}) +[2026-06-17 14:55:08,302][starVLA.training.train_starvla][INFO] - Step 4639, Loss: {'train/loss': 0.7254242897033691, 'train/grad_norm_pre_clip': 15.028665542602539, 'timing/data': 0.004830498015508056, 'timing/model': 1.2041392660466954, 'learning_rate/qwen_vl_interface': 2.9901508970389477e-06, 'learning_rate/action_model': 2.9901508970389476e-05, 'epoch': 1.37}) +[2026-06-17 14:55:13,089][starVLA.training.train_starvla][INFO] - Step 4640, Loss: {'train/loss': 0.5787487030029297, 'train/grad_norm_pre_clip': 13.582852363586426, 'timing/data': 0.00497051200363785, 'timing/model': 1.2143843079684302, 'learning_rate/qwen_vl_interface': 2.988248879778401e-06, 'learning_rate/action_model': 2.988248879778401e-05, 'epoch': 1.37}) +[2026-06-17 14:55:17,859][starVLA.training.train_starvla][INFO] - Step 4641, Loss: {'train/loss': 0.7636834383010864, 'train/grad_norm_pre_clip': 28.6934871673584, 'timing/data': 0.00489614496473223, 'timing/model': 1.2029989029979333, 'learning_rate/qwen_vl_interface': 2.986347331381297e-06, 'learning_rate/action_model': 2.9863473313812967e-05, 'epoch': 1.37}) +[2026-06-17 14:55:22,609][starVLA.training.train_starvla][INFO] - Step 4642, Loss: {'train/loss': 0.5156235694885254, 'train/grad_norm_pre_clip': 15.94502067565918, 'timing/data': 0.0049222479574382305, 'timing/model': 1.2990896860137582, 'learning_rate/qwen_vl_interface': 2.984446252241826e-06, 'learning_rate/action_model': 2.984446252241826e-05, 'epoch': 1.37}) +[2026-06-17 14:55:27,266][starVLA.training.train_starvla][INFO] - Step 4643, Loss: {'train/loss': 0.7445421814918518, 'train/grad_norm_pre_clip': 12.77294635772705, 'timing/data': 0.004720913013443351, 'timing/model': 1.205667829955928, 'learning_rate/qwen_vl_interface': 2.982545642754083e-06, 'learning_rate/action_model': 2.9825456427540826e-05, 'epoch': 1.37}) +[2026-06-17 14:55:32,055][starVLA.training.train_starvla][INFO] - Step 4644, Loss: {'train/loss': 0.5758833289146423, 'train/grad_norm_pre_clip': 10.634500503540039, 'timing/data': 0.006922891945578158, 'timing/model': 1.2102470840327442, 'learning_rate/qwen_vl_interface': 2.9806455033120697e-06, 'learning_rate/action_model': 2.9806455033120696e-05, 'epoch': 1.37}) +[2026-06-17 14:55:36,841][starVLA.training.train_starvla][INFO] - Step 4645, Loss: {'train/loss': 0.5417333245277405, 'train/grad_norm_pre_clip': 11.6614351272583, 'timing/data': 0.004868053016252816, 'timing/model': 1.2029939650092274, 'learning_rate/qwen_vl_interface': 2.978745834309685e-06, 'learning_rate/action_model': 2.9787458343096846e-05, 'epoch': 1.37}) +[2026-06-17 14:55:41,626][starVLA.training.train_starvla][INFO] - Step 4646, Loss: {'train/loss': 0.6798235774040222, 'train/grad_norm_pre_clip': 15.336551666259766, 'timing/data': 0.004654887015931308, 'timing/model': 1.2033423570683226, 'learning_rate/qwen_vl_interface': 2.9768466361407335e-06, 'learning_rate/action_model': 2.9768466361407334e-05, 'epoch': 1.37}) +[2026-06-17 14:55:46,427][starVLA.training.train_starvla][INFO] - Step 4647, Loss: {'train/loss': 1.1761606931686401, 'train/grad_norm_pre_clip': 15.925567626953125, 'timing/data': 0.0054995439713820815, 'timing/model': 1.333143790019676, 'learning_rate/qwen_vl_interface': 2.97494790919892e-06, 'learning_rate/action_model': 2.9749479091989197e-05, 'epoch': 1.37}) +[2026-06-17 14:55:51,075][starVLA.training.train_starvla][INFO] - Step 4648, Loss: {'train/loss': 0.7099850177764893, 'train/grad_norm_pre_clip': 13.575155258178711, 'timing/data': 0.0043699119705706835, 'timing/model': 1.2037324419943616, 'learning_rate/qwen_vl_interface': 2.9730496538778524e-06, 'learning_rate/action_model': 2.9730496538778523e-05, 'epoch': 1.37}) +[2026-06-17 14:55:55,852][starVLA.training.train_starvla][INFO] - Step 4649, Loss: {'train/loss': 0.5773239135742188, 'train/grad_norm_pre_clip': 17.610517501831055, 'timing/data': 0.0048917949898168445, 'timing/model': 1.2034891080111265, 'learning_rate/qwen_vl_interface': 2.9711518705710422e-06, 'learning_rate/action_model': 2.9711518705710424e-05, 'epoch': 1.37}) +[2026-06-17 14:56:00,631][starVLA.training.train_starvla][INFO] - Step 4650, Loss: {'train/loss': 0.43657243251800537, 'train/grad_norm_pre_clip': 11.115675926208496, 'timing/data': 0.004690173082053661, 'timing/model': 1.2051710590021685, 'learning_rate/qwen_vl_interface': 2.9692545596719014e-06, 'learning_rate/action_model': 2.969254559671901e-05, 'epoch': 1.37}) +[2026-06-17 14:56:05,337][starVLA.training.train_starvla][INFO] - Step 4651, Loss: {'train/loss': 0.4117002785205841, 'train/grad_norm_pre_clip': 9.715848922729492, 'timing/data': 0.004720833036117256, 'timing/model': 1.2035863569471985, 'learning_rate/qwen_vl_interface': 2.9673577215737454e-06, 'learning_rate/action_model': 2.9673577215737452e-05, 'epoch': 1.37}) +[2026-06-17 14:56:10,071][starVLA.training.train_starvla][INFO] - Step 4652, Loss: {'train/loss': 0.5923128724098206, 'train/grad_norm_pre_clip': 11.271122932434082, 'timing/data': 0.004944877931848168, 'timing/model': 1.2060682760784402, 'learning_rate/qwen_vl_interface': 2.9654613566697886e-06, 'learning_rate/action_model': 2.9654613566697887e-05, 'epoch': 1.37}) +[2026-06-17 14:56:14,855][starVLA.training.train_starvla][INFO] - Step 4653, Loss: {'train/loss': 0.7394567131996155, 'train/grad_norm_pre_clip': 18.177766799926758, 'timing/data': 0.004649279988370836, 'timing/model': 1.3342062080046162, 'learning_rate/qwen_vl_interface': 2.9635654653531516e-06, 'learning_rate/action_model': 2.9635654653531515e-05, 'epoch': 1.37}) +[2026-06-17 14:56:19,607][starVLA.training.train_starvla][INFO] - Step 4654, Loss: {'train/loss': 0.9697430729866028, 'train/grad_norm_pre_clip': 16.108722686767578, 'timing/data': 0.005029045045375824, 'timing/model': 1.2089892770163715, 'learning_rate/qwen_vl_interface': 2.961670048016852e-06, 'learning_rate/action_model': 2.9616700480168518e-05, 'epoch': 1.37}) +[2026-06-17 14:56:24,393][starVLA.training.train_starvla][INFO] - Step 4655, Loss: {'train/loss': 0.6707581281661987, 'train/grad_norm_pre_clip': 14.24260425567627, 'timing/data': 0.00490109296515584, 'timing/model': 1.203644953086041, 'learning_rate/qwen_vl_interface': 2.9597751050538177e-06, 'learning_rate/action_model': 2.9597751050538176e-05, 'epoch': 1.37}) +[2026-06-17 14:56:29,168][starVLA.training.train_starvla][INFO] - Step 4656, Loss: {'train/loss': 0.4975045919418335, 'train/grad_norm_pre_clip': 12.567506790161133, 'timing/data': 0.004723549005575478, 'timing/model': 1.2020071330480278, 'learning_rate/qwen_vl_interface': 2.957880636856866e-06, 'learning_rate/action_model': 2.9578806368568655e-05, 'epoch': 1.37}) +[2026-06-17 14:56:34,007][starVLA.training.train_starvla][INFO] - Step 4657, Loss: {'train/loss': 0.8412649035453796, 'train/grad_norm_pre_clip': 17.992084503173828, 'timing/data': 0.004738318966701627, 'timing/model': 1.2019238070352003, 'learning_rate/qwen_vl_interface': 2.955986643818723e-06, 'learning_rate/action_model': 2.9559866438187228e-05, 'epoch': 1.37}) +[2026-06-17 14:56:38,755][starVLA.training.train_starvla][INFO] - Step 4658, Loss: {'train/loss': 0.8250868320465088, 'train/grad_norm_pre_clip': 13.721049308776855, 'timing/data': 0.004799048067070544, 'timing/model': 1.2073082079878077, 'learning_rate/qwen_vl_interface': 2.954093126332019e-06, 'learning_rate/action_model': 2.9540931263320194e-05, 'epoch': 1.37}) +[2026-06-17 14:56:43,541][starVLA.training.train_starvla][INFO] - Step 4659, Loss: {'train/loss': 0.6631134748458862, 'train/grad_norm_pre_clip': 16.25200080871582, 'timing/data': 0.004571308963932097, 'timing/model': 1.324309029034339, 'learning_rate/qwen_vl_interface': 2.9522000847892807e-06, 'learning_rate/action_model': 2.9522000847892807e-05, 'epoch': 1.37}) +[2026-06-17 14:56:48,201][starVLA.training.train_starvla][INFO] - Step 4660, Loss: {'train/loss': 0.957694411277771, 'train/grad_norm_pre_clip': 14.226285934448242, 'timing/data': 0.004977826960384846, 'timing/model': 1.2040821049595252, 'learning_rate/qwen_vl_interface': 2.9503075195829393e-06, 'learning_rate/action_model': 2.9503075195829393e-05, 'epoch': 1.37}) +[2026-06-17 14:56:52,983][starVLA.training.train_starvla][INFO] - Step 4661, Loss: {'train/loss': 0.7343818545341492, 'train/grad_norm_pre_clip': 17.649242401123047, 'timing/data': 0.00502286700066179, 'timing/model': 1.202109616016969, 'learning_rate/qwen_vl_interface': 2.9484154311053203e-06, 'learning_rate/action_model': 2.94841543110532e-05, 'epoch': 1.37}) +[2026-06-17 14:56:57,774][starVLA.training.train_starvla][INFO] - Step 4662, Loss: {'train/loss': 0.7516250610351562, 'train/grad_norm_pre_clip': 14.219526290893555, 'timing/data': 0.004896857077255845, 'timing/model': 1.2024814060423523, 'learning_rate/qwen_vl_interface': 2.94652381974866e-06, 'learning_rate/action_model': 2.94652381974866e-05, 'epoch': 1.37}) +[2026-06-17 14:57:02,561][starVLA.training.train_starvla][INFO] - Step 4663, Loss: {'train/loss': 0.6541241407394409, 'train/grad_norm_pre_clip': 14.188128471374512, 'timing/data': 0.00491338805295527, 'timing/model': 1.205112608964555, 'learning_rate/qwen_vl_interface': 2.9446326859050916e-06, 'learning_rate/action_model': 2.9446326859050916e-05, 'epoch': 1.37}) +[2026-06-17 14:57:07,340][starVLA.training.train_starvla][INFO] - Step 4664, Loss: {'train/loss': 0.5073983073234558, 'train/grad_norm_pre_clip': 11.0167236328125, 'timing/data': 0.005114493076689541, 'timing/model': 1.330567714991048, 'learning_rate/qwen_vl_interface': 2.9427420299666472e-06, 'learning_rate/action_model': 2.942742029966647e-05, 'epoch': 1.37}) +[2026-06-17 14:57:11,996][starVLA.training.train_starvla][INFO] - Step 4665, Loss: {'train/loss': 0.5364934802055359, 'train/grad_norm_pre_clip': 12.443222045898438, 'timing/data': 0.00464903109241277, 'timing/model': 1.2052626970689744, 'learning_rate/qwen_vl_interface': 2.940851852325263e-06, 'learning_rate/action_model': 2.9408518523252632e-05, 'epoch': 1.37}) +[2026-06-17 14:57:16,785][starVLA.training.train_starvla][INFO] - Step 4666, Loss: {'train/loss': 0.7492870688438416, 'train/grad_norm_pre_clip': 13.912023544311523, 'timing/data': 0.005000552977435291, 'timing/model': 1.2049163789488375, 'learning_rate/qwen_vl_interface': 2.9389621533727745e-06, 'learning_rate/action_model': 2.9389621533727747e-05, 'epoch': 1.37}) +[2026-06-17 14:57:21,500][starVLA.training.train_starvla][INFO] - Step 4667, Loss: {'train/loss': 0.725530743598938, 'train/grad_norm_pre_clip': 16.88976287841797, 'timing/data': 0.004762070020660758, 'timing/model': 1.2073663149494678, 'learning_rate/qwen_vl_interface': 2.9370729335009185e-06, 'learning_rate/action_model': 2.9370729335009184e-05, 'epoch': 1.37}) +[2026-06-17 14:57:26,242][starVLA.training.train_starvla][INFO] - Step 4668, Loss: {'train/loss': 0.5839630961418152, 'train/grad_norm_pre_clip': 11.41794490814209, 'timing/data': 0.004808594938367605, 'timing/model': 1.2057427999097854, 'learning_rate/qwen_vl_interface': 2.93518419310133e-06, 'learning_rate/action_model': 2.93518419310133e-05, 'epoch': 1.37}) +[2026-06-17 14:57:31,116][starVLA.training.train_starvla][INFO] - Step 4669, Loss: {'train/loss': 0.6048625111579895, 'train/grad_norm_pre_clip': 12.986612319946289, 'timing/data': 0.004742537043057382, 'timing/model': 1.2047456209547818, 'learning_rate/qwen_vl_interface': 2.9332959325655537e-06, 'learning_rate/action_model': 2.9332959325655535e-05, 'epoch': 1.37}) +[2026-06-17 14:57:35,925][starVLA.training.train_starvla][INFO] - Step 4670, Loss: {'train/loss': 0.6397735476493835, 'train/grad_norm_pre_clip': 14.184700012207031, 'timing/data': 0.005165148060768843, 'timing/model': 1.3471333120251074, 'learning_rate/qwen_vl_interface': 2.9314081522850206e-06, 'learning_rate/action_model': 2.9314081522850205e-05, 'epoch': 1.37}) +[2026-06-17 14:57:40,649][starVLA.training.train_starvla][INFO] - Step 4671, Loss: {'train/loss': 0.660693347454071, 'train/grad_norm_pre_clip': 11.805917739868164, 'timing/data': 0.004837702959775925, 'timing/model': 1.2071488719666377, 'learning_rate/qwen_vl_interface': 2.9295208526510726e-06, 'learning_rate/action_model': 2.9295208526510725e-05, 'epoch': 1.37}) +[2026-06-17 14:57:45,452][starVLA.training.train_starvla][INFO] - Step 4672, Loss: {'train/loss': 0.5640106797218323, 'train/grad_norm_pre_clip': 12.471673011779785, 'timing/data': 0.00446570897474885, 'timing/model': 1.2097811920102686, 'learning_rate/qwen_vl_interface': 2.9276340340549476e-06, 'learning_rate/action_model': 2.9276340340549473e-05, 'epoch': 1.37}) +[2026-06-17 14:57:50,313][starVLA.training.train_starvla][INFO] - Step 4673, Loss: {'train/loss': 0.587975025177002, 'train/grad_norm_pre_clip': 12.347783088684082, 'timing/data': 0.0046842669835314155, 'timing/model': 1.2055323818931356, 'learning_rate/qwen_vl_interface': 2.9257476968877875e-06, 'learning_rate/action_model': 2.9257476968877874e-05, 'epoch': 1.38}) +[2026-06-17 14:57:55,081][starVLA.training.train_starvla][INFO] - Step 4674, Loss: {'train/loss': 0.6413347125053406, 'train/grad_norm_pre_clip': 13.578413963317871, 'timing/data': 0.005284362006932497, 'timing/model': 1.2088802949292585, 'learning_rate/qwen_vl_interface': 2.923861841540631e-06, 'learning_rate/action_model': 2.923861841540631e-05, 'epoch': 1.38}) +[2026-06-17 14:57:59,885][starVLA.training.train_starvla][INFO] - Step 4675, Loss: {'train/loss': 0.4826928973197937, 'train/grad_norm_pre_clip': 9.740699768066406, 'timing/data': 0.004788075108081102, 'timing/model': 1.210190113983117, 'learning_rate/qwen_vl_interface': 2.9219764684044196e-06, 'learning_rate/action_model': 2.9219764684044193e-05, 'epoch': 1.38}) +[2026-06-17 14:58:04,681][starVLA.training.train_starvla][INFO] - Step 4676, Loss: {'train/loss': 0.535210371017456, 'train/grad_norm_pre_clip': 10.473553657531738, 'timing/data': 0.004982972051948309, 'timing/model': 1.3364854210522026, 'learning_rate/qwen_vl_interface': 2.920091577869988e-06, 'learning_rate/action_model': 2.9200915778699877e-05, 'epoch': 1.38}) +[2026-06-17 14:58:09,337][starVLA.training.train_starvla][INFO] - Step 4677, Loss: {'train/loss': 0.8408198952674866, 'train/grad_norm_pre_clip': 15.943370819091797, 'timing/data': 0.005147804971784353, 'timing/model': 1.2035523098893464, 'learning_rate/qwen_vl_interface': 2.918207170328081e-06, 'learning_rate/action_model': 2.918207170328081e-05, 'epoch': 1.38}) +[2026-06-17 14:58:14,130][starVLA.training.train_starvla][INFO] - Step 4678, Loss: {'train/loss': 0.7621506452560425, 'train/grad_norm_pre_clip': 14.546806335449219, 'timing/data': 0.004868257092311978, 'timing/model': 1.2079511319752783, 'learning_rate/qwen_vl_interface': 2.9163232461693373e-06, 'learning_rate/action_model': 2.9163232461693373e-05, 'epoch': 1.38}) +[2026-06-17 14:58:18,929][starVLA.training.train_starvla][INFO] - Step 4679, Loss: {'train/loss': 0.5723893642425537, 'train/grad_norm_pre_clip': 12.92231559753418, 'timing/data': 0.004576234030537307, 'timing/model': 1.2040910890791565, 'learning_rate/qwen_vl_interface': 2.9144398057842948e-06, 'learning_rate/action_model': 2.9144398057842948e-05, 'epoch': 1.38}) +[2026-06-17 14:58:23,728][starVLA.training.train_starvla][INFO] - Step 4680, Loss: {'train/loss': 0.5928627252578735, 'train/grad_norm_pre_clip': 10.493987083435059, 'timing/data': 0.005115655949339271, 'timing/model': 1.2073083700379357, 'learning_rate/qwen_vl_interface': 2.912556849563394e-06, 'learning_rate/action_model': 2.9125568495633943e-05, 'epoch': 1.38}) +[2026-06-17 14:58:28,412][starVLA.training.train_starvla][INFO] - Step 4681, Loss: {'train/loss': 0.7503913640975952, 'train/grad_norm_pre_clip': 13.10461711883545, 'timing/data': 0.004774080007337034, 'timing/model': 1.2031501790042967, 'learning_rate/qwen_vl_interface': 2.910674377896974e-06, 'learning_rate/action_model': 2.910674377896974e-05, 'epoch': 1.38}) +[2026-06-17 14:58:33,102][starVLA.training.train_starvla][INFO] - Step 4682, Loss: {'train/loss': 0.6902775764465332, 'train/grad_norm_pre_clip': 16.530973434448242, 'timing/data': 0.004945575958117843, 'timing/model': 1.2451588190160692, 'learning_rate/qwen_vl_interface': 2.9087923911752723e-06, 'learning_rate/action_model': 2.9087923911752725e-05, 'epoch': 1.38}) +[2026-06-17 14:58:37,756][starVLA.training.train_starvla][INFO] - Step 4683, Loss: {'train/loss': 0.5989510416984558, 'train/grad_norm_pre_clip': 12.433237075805664, 'timing/data': 0.005027499049901962, 'timing/model': 1.2036977279931307, 'learning_rate/qwen_vl_interface': 2.906910889788425e-06, 'learning_rate/action_model': 2.906910889788425e-05, 'epoch': 1.38}) +[2026-06-17 14:58:42,556][starVLA.training.train_starvla][INFO] - Step 4684, Loss: {'train/loss': 0.7989625930786133, 'train/grad_norm_pre_clip': 15.318188667297363, 'timing/data': 0.004584893933497369, 'timing/model': 1.205766384024173, 'learning_rate/qwen_vl_interface': 2.905029874126476e-06, 'learning_rate/action_model': 2.905029874126476e-05, 'epoch': 1.38}) +[2026-06-17 14:58:47,348][starVLA.training.train_starvla][INFO] - Step 4685, Loss: {'train/loss': 0.586421012878418, 'train/grad_norm_pre_clip': 12.231781959533691, 'timing/data': 0.004866122966632247, 'timing/model': 1.2060691149672493, 'learning_rate/qwen_vl_interface': 2.9031493445793558e-06, 'learning_rate/action_model': 2.9031493445793558e-05, 'epoch': 1.38}) +[2026-06-17 14:58:52,128][starVLA.training.train_starvla][INFO] - Step 4686, Loss: {'train/loss': 0.5566617250442505, 'train/grad_norm_pre_clip': 11.750354766845703, 'timing/data': 0.005194921977818012, 'timing/model': 1.2038897899910808, 'learning_rate/qwen_vl_interface': 2.9012693015369007e-06, 'learning_rate/action_model': 2.9012693015369008e-05, 'epoch': 1.38}) +[2026-06-17 14:58:56,919][starVLA.training.train_starvla][INFO] - Step 4687, Loss: {'train/loss': 0.5843116641044617, 'train/grad_norm_pre_clip': 10.477301597595215, 'timing/data': 0.00483528699260205, 'timing/model': 1.3322155389469117, 'learning_rate/qwen_vl_interface': 2.8993897453888454e-06, 'learning_rate/action_model': 2.899389745388845e-05, 'epoch': 1.38}) +[2026-06-17 14:59:01,619][starVLA.training.train_starvla][INFO] - Step 4688, Loss: {'train/loss': 0.6131628155708313, 'train/grad_norm_pre_clip': 14.779119491577148, 'timing/data': 0.005243564024567604, 'timing/model': 1.2517952530179173, 'learning_rate/qwen_vl_interface': 2.8975106765248267e-06, 'learning_rate/action_model': 2.8975106765248267e-05, 'epoch': 1.38}) +[2026-06-17 14:59:06,268][starVLA.training.train_starvla][INFO] - Step 4689, Loss: {'train/loss': 0.3653799891471863, 'train/grad_norm_pre_clip': 9.501229286193848, 'timing/data': 0.004852549056522548, 'timing/model': 1.203596935956739, 'learning_rate/qwen_vl_interface': 2.895632095334378e-06, 'learning_rate/action_model': 2.895632095334378e-05, 'epoch': 1.38}) +[2026-06-17 14:59:11,007][starVLA.training.train_starvla][INFO] - Step 4690, Loss: {'train/loss': 0.8226701617240906, 'train/grad_norm_pre_clip': 15.049320220947266, 'timing/data': 0.004700034041889012, 'timing/model': 1.203220605966635, 'learning_rate/qwen_vl_interface': 2.8937540022069267e-06, 'learning_rate/action_model': 2.8937540022069265e-05, 'epoch': 1.38}) +[2026-06-17 14:59:15,882][starVLA.training.train_starvla][INFO] - Step 4691, Loss: {'train/loss': 0.6699360013008118, 'train/grad_norm_pre_clip': 14.125467300415039, 'timing/data': 0.005279363947920501, 'timing/model': 1.2050226300489157, 'learning_rate/qwen_vl_interface': 2.8918763975318076e-06, 'learning_rate/action_model': 2.8918763975318076e-05, 'epoch': 1.38}) +[2026-06-17 14:59:20,678][starVLA.training.train_starvla][INFO] - Step 4692, Loss: {'train/loss': 0.6127247214317322, 'train/grad_norm_pre_clip': 13.41641616821289, 'timing/data': 0.004729188978672028, 'timing/model': 1.2108207290293649, 'learning_rate/qwen_vl_interface': 2.889999281698248e-06, 'learning_rate/action_model': 2.889999281698248e-05, 'epoch': 1.38}) +[2026-06-17 14:59:25,486][starVLA.training.train_starvla][INFO] - Step 4693, Loss: {'train/loss': 0.563186764717102, 'train/grad_norm_pre_clip': 10.911019325256348, 'timing/data': 0.00457567791454494, 'timing/model': 1.3417748409556225, 'learning_rate/qwen_vl_interface': 2.8881226550953772e-06, 'learning_rate/action_model': 2.888122655095377e-05, 'epoch': 1.38}) +[2026-06-17 14:59:30,271][starVLA.training.train_starvla][INFO] - Step 4694, Loss: {'train/loss': 0.45081984996795654, 'train/grad_norm_pre_clip': 9.62617301940918, 'timing/data': 0.004771467996761203, 'timing/model': 1.326166840037331, 'learning_rate/qwen_vl_interface': 2.8862465181122212e-06, 'learning_rate/action_model': 2.886246518112221e-05, 'epoch': 1.38}) +[2026-06-17 14:59:34,940][starVLA.training.train_starvla][INFO] - Step 4695, Loss: {'train/loss': 0.6995209455490112, 'train/grad_norm_pre_clip': 24.042259216308594, 'timing/data': 0.004671814036555588, 'timing/model': 1.213059718022123, 'learning_rate/qwen_vl_interface': 2.884370871137706e-06, 'learning_rate/action_model': 2.884370871137706e-05, 'epoch': 1.38}) +[2026-06-17 14:59:39,709][starVLA.training.train_starvla][INFO] - Step 4696, Loss: {'train/loss': 0.7951818108558655, 'train/grad_norm_pre_clip': 15.108173370361328, 'timing/data': 0.004678890109062195, 'timing/model': 1.2194054940482602, 'learning_rate/qwen_vl_interface': 2.8824957145606548e-06, 'learning_rate/action_model': 2.8824957145606548e-05, 'epoch': 1.38}) +[2026-06-17 14:59:44,493][starVLA.training.train_starvla][INFO] - Step 4697, Loss: {'train/loss': 0.9551298022270203, 'train/grad_norm_pre_clip': 13.617069244384766, 'timing/data': 0.004797455971129239, 'timing/model': 1.2044083509827033, 'learning_rate/qwen_vl_interface': 2.8806210487697866e-06, 'learning_rate/action_model': 2.8806210487697867e-05, 'epoch': 1.38}) +[2026-06-17 14:59:49,279][starVLA.training.train_starvla][INFO] - Step 4698, Loss: {'train/loss': 0.923369288444519, 'train/grad_norm_pre_clip': 24.819992065429688, 'timing/data': 0.004970688954927027, 'timing/model': 1.2039739009924233, 'learning_rate/qwen_vl_interface': 2.87874687415373e-06, 'learning_rate/action_model': 2.87874687415373e-05, 'epoch': 1.38}) +[2026-06-17 14:59:54,008][starVLA.training.train_starvla][INFO] - Step 4699, Loss: {'train/loss': 0.4295841157436371, 'train/grad_norm_pre_clip': 14.507574081420898, 'timing/data': 0.004735309979878366, 'timing/model': 1.2816312400391325, 'learning_rate/qwen_vl_interface': 2.876873191100995e-06, 'learning_rate/action_model': 2.8768731911009948e-05, 'epoch': 1.38}) +[2026-06-17 15:01:35,619][starVLA.training.train_starvla][INFO] - Step 4700, Loss: {'train/loss': 0.5573174357414246, 'train/grad_norm_pre_clip': 13.731864929199219, 'eval/loss': 0.7185298204421997, 'eval/latency_0/loss': 0.7185298204421997, 'eval/demon_fire_precision': 0.9003202562049639, 'eval/demon_fire_recall': 0.9339700996677741, 'eval/demon_fire_f1': 0.9168365267019974, 'eval/demon_fire_rate_model': 0.6946607341490545, 'eval/demon_fire_rate_teacher': 0.6696329254727474, 'eval/demon_left_precision': 0.8220823798627003, 'eval/demon_left_recall': 0.8141643059490085, 'eval/demon_left_f1': 0.8181041844577285, 'eval/demon_left_rate_model': 0.4860956618464961, 'eval/demon_left_rate_teacher': 0.49082313681868744, 'eval/demon_right_precision': 0.8221606648199447, 'eval/demon_right_recall': 0.8384180790960452, 'eval/demon_right_f1': 0.8302097902097902, 'eval/demon_right_rate_model': 0.5019466073414905, 'eval/demon_right_rate_teacher': 0.4922135706340378, 'eval/demon_move_f1': 0.8241569873337593, 'eval/demon_cc_f1': 0.8869154380547661, 'eval/latency_0/demon_fire_precision': 0.9003202562049639, 'eval/latency_0/demon_fire_recall': 0.9339700996677741, 'eval/latency_0/demon_fire_f1': 0.9168365267019974, 'eval/latency_0/demon_fire_rate_model': 0.6946607341490545, 'eval/latency_0/demon_fire_rate_teacher': 0.6696329254727474, 'eval/latency_0/demon_left_precision': 0.8220823798627003, 'eval/latency_0/demon_left_recall': 0.8141643059490085, 'eval/latency_0/demon_left_f1': 0.8181041844577285, 'eval/latency_0/demon_left_rate_model': 0.4860956618464961, 'eval/latency_0/demon_left_rate_teacher': 0.49082313681868744, 'eval/latency_0/demon_right_precision': 0.8221606648199447, 'eval/latency_0/demon_right_recall': 0.8384180790960452, 'eval/latency_0/demon_right_f1': 0.8302097902097902, 'eval/latency_0/demon_right_rate_model': 0.5019466073414905, 'eval/latency_0/demon_right_rate_teacher': 0.4922135706340378, 'eval/latency_0/demon_move_f1': 0.8241569873337593, 'eval/latency_0/demon_cc_f1': 0.8869154380547661, 'timing/data': 0.004985262989066541, 'timing/model': 1.2032803639303893, 'learning_rate/qwen_vl_interface': 2.8750000000000013e-06, 'learning_rate/action_model': 2.875000000000001e-05, 'epoch': 1.38}) +[2026-06-17 15:01:35,620][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_4700_state +[2026-06-17 15:01:35,631][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 15:01:39,195][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_4700_state/model.safetensors +[2026-06-17 15:01:46,818][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_4700_state/optimizer.bin +[2026-06-17 15:01:46,819][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_4700_state/sampler.bin +[2026-06-17 15:01:46,819][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_4700_state/sampler_1.bin +[2026-06-17 15:01:46,820][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_4700_state/random_states_0.pkl +[2026-06-17 15:01:46,863][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4700 +[2026-06-17 15:01:46,863][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 15:01:46,873][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 15:01:51,574][starVLA.training.train_starvla][INFO] - Step 4701, Loss: {'train/loss': 0.7110336422920227, 'train/grad_norm_pre_clip': 22.646347045898438, 'timing/data': 0.006376998964697123, 'timing/model': 1.3128150139236823, 'learning_rate/qwen_vl_interface': 2.8731273012390614e-06, 'learning_rate/action_model': 2.8731273012390615e-05, 'epoch': 1.38}) +[2026-06-17 15:01:56,177][starVLA.training.train_starvla][INFO] - Step 4702, Loss: {'train/loss': 0.3260652422904968, 'train/grad_norm_pre_clip': 8.995309829711914, 'timing/data': 0.005345159908756614, 'timing/model': 1.1939043620368466, 'learning_rate/qwen_vl_interface': 2.8712550952063912e-06, 'learning_rate/action_model': 2.871255095206391e-05, 'epoch': 1.38}) +[2026-06-17 15:02:00,899][starVLA.training.train_starvla][INFO] - Step 4703, Loss: {'train/loss': 0.5844410061836243, 'train/grad_norm_pre_clip': 13.374002456665039, 'timing/data': 0.005384887917898595, 'timing/model': 1.191061518038623, 'learning_rate/qwen_vl_interface': 2.8693833822901012e-06, 'learning_rate/action_model': 2.8693833822901012e-05, 'epoch': 1.38}) +[2026-06-17 15:02:05,654][starVLA.training.train_starvla][INFO] - Step 4704, Loss: {'train/loss': 0.5754327774047852, 'train/grad_norm_pre_clip': 12.615023612976074, 'timing/data': 0.005409107077866793, 'timing/model': 1.2087383100297302, 'learning_rate/qwen_vl_interface': 2.867512162878196e-06, 'learning_rate/action_model': 2.867512162878196e-05, 'epoch': 1.38}) +[2026-06-17 15:02:10,397][starVLA.training.train_starvla][INFO] - Step 4705, Loss: {'train/loss': 0.542592465877533, 'train/grad_norm_pre_clip': 20.50316619873047, 'timing/data': 0.005567008047364652, 'timing/model': 1.194405427086167, 'learning_rate/qwen_vl_interface': 2.8656414373585812e-06, 'learning_rate/action_model': 2.865641437358581e-05, 'epoch': 1.38}) +[2026-06-17 15:02:15,147][starVLA.training.train_starvla][INFO] - Step 4706, Loss: {'train/loss': 0.7039211988449097, 'train/grad_norm_pre_clip': 13.701683044433594, 'timing/data': 0.005559104029089212, 'timing/model': 1.3250592809636146, 'learning_rate/qwen_vl_interface': 2.8637712061190644e-06, 'learning_rate/action_model': 2.863771206119064e-05, 'epoch': 1.38}) +[2026-06-17 15:02:19,851][starVLA.training.train_starvla][INFO] - Step 4707, Loss: {'train/loss': 0.6557416915893555, 'train/grad_norm_pre_clip': 15.867341995239258, 'timing/data': 0.0047784969210624695, 'timing/model': 1.2792867799289525, 'learning_rate/qwen_vl_interface': 2.861901469547344e-06, 'learning_rate/action_model': 2.861901469547344e-05, 'epoch': 1.39}) +[2026-06-17 15:02:24,528][starVLA.training.train_starvla][INFO] - Step 4708, Loss: {'train/loss': 0.4614178538322449, 'train/grad_norm_pre_clip': 10.317148208618164, 'timing/data': 0.005824582069180906, 'timing/model': 1.1998122659279034, 'learning_rate/qwen_vl_interface': 2.8600322280310185e-06, 'learning_rate/action_model': 2.8600322280310183e-05, 'epoch': 1.39}) +[2026-06-17 15:02:29,300][starVLA.training.train_starvla][INFO] - Step 4709, Loss: {'train/loss': 0.676550030708313, 'train/grad_norm_pre_clip': 11.940231323242188, 'timing/data': 0.004948375979438424, 'timing/model': 1.2046811379259452, 'learning_rate/qwen_vl_interface': 2.8581634819575843e-06, 'learning_rate/action_model': 2.8581634819575843e-05, 'epoch': 1.39}) +[2026-06-17 15:02:34,059][starVLA.training.train_starvla][INFO] - Step 4710, Loss: {'train/loss': 0.6437956094741821, 'train/grad_norm_pre_clip': 13.006513595581055, 'timing/data': 0.005194509983994067, 'timing/model': 1.1990176369436085, 'learning_rate/qwen_vl_interface': 2.856295231714434e-06, 'learning_rate/action_model': 2.8562952317144337e-05, 'epoch': 1.39}) +[2026-06-17 15:02:38,853][starVLA.training.train_starvla][INFO] - Step 4711, Loss: {'train/loss': 0.7296977639198303, 'train/grad_norm_pre_clip': 16.63241195678711, 'timing/data': 0.005575345014221966, 'timing/model': 1.228354093967937, 'learning_rate/qwen_vl_interface': 2.854427477688857e-06, 'learning_rate/action_model': 2.854427477688857e-05, 'epoch': 1.39}) +[2026-06-17 15:02:43,608][starVLA.training.train_starvla][INFO] - Step 4712, Loss: {'train/loss': 0.6321083307266235, 'train/grad_norm_pre_clip': 15.06595516204834, 'timing/data': 0.00597708800341934, 'timing/model': 1.1981926719890907, 'learning_rate/qwen_vl_interface': 2.8525602202680424e-06, 'learning_rate/action_model': 2.8525602202680424e-05, 'epoch': 1.39}) +[2026-06-17 15:02:48,339][starVLA.training.train_starvla][INFO] - Step 4713, Loss: {'train/loss': 0.6061331629753113, 'train/grad_norm_pre_clip': 14.107612609863281, 'timing/data': 0.005575318937189877, 'timing/model': 1.2914268600288779, 'learning_rate/qwen_vl_interface': 2.8506934598390736e-06, 'learning_rate/action_model': 2.8506934598390735e-05, 'epoch': 1.39}) +[2026-06-17 15:02:52,991][starVLA.training.train_starvla][INFO] - Step 4714, Loss: {'train/loss': 0.5994449853897095, 'train/grad_norm_pre_clip': 9.927398681640625, 'timing/data': 0.005739189102314413, 'timing/model': 1.2109779919264838, 'learning_rate/qwen_vl_interface': 2.848827196788932e-06, 'learning_rate/action_model': 2.848827196788932e-05, 'epoch': 1.39}) +[2026-06-17 15:02:57,763][starVLA.training.train_starvla][INFO] - Step 4715, Loss: {'train/loss': 0.712823212146759, 'train/grad_norm_pre_clip': 16.245567321777344, 'timing/data': 0.00532219500746578, 'timing/model': 1.2012752359732985, 'learning_rate/qwen_vl_interface': 2.8469614315044975e-06, 'learning_rate/action_model': 2.8469614315044975e-05, 'epoch': 1.39}) +[2026-06-17 15:03:02,542][starVLA.training.train_starvla][INFO] - Step 4716, Loss: {'train/loss': 0.9149066805839539, 'train/grad_norm_pre_clip': 13.67570686340332, 'timing/data': 0.00561677897349, 'timing/model': 1.200507991015911, 'learning_rate/qwen_vl_interface': 2.8450961643725405e-06, 'learning_rate/action_model': 2.8450961643725405e-05, 'epoch': 1.39}) +[2026-06-17 15:03:07,320][starVLA.training.train_starvla][INFO] - Step 4717, Loss: {'train/loss': 0.636847972869873, 'train/grad_norm_pre_clip': 22.476259231567383, 'timing/data': 0.0062320230063050985, 'timing/model': 1.201618246966973, 'learning_rate/qwen_vl_interface': 2.8432313957797415e-06, 'learning_rate/action_model': 2.8432313957797414e-05, 'epoch': 1.39}) +[2026-06-17 15:03:12,105][starVLA.training.train_starvla][INFO] - Step 4718, Loss: {'train/loss': 0.637611448764801, 'train/grad_norm_pre_clip': 14.363424301147461, 'timing/data': 0.005443311063572764, 'timing/model': 1.3420079370262101, 'learning_rate/qwen_vl_interface': 2.8413671261126613e-06, 'learning_rate/action_model': 2.8413671261126613e-05, 'epoch': 1.39}) +[2026-06-17 15:03:16,746][starVLA.training.train_starvla][INFO] - Step 4719, Loss: {'train/loss': 0.5876240730285645, 'train/grad_norm_pre_clip': 14.674766540527344, 'timing/data': 0.0055862730368971825, 'timing/model': 1.2049229260301217, 'learning_rate/qwen_vl_interface': 2.8395033557577668e-06, 'learning_rate/action_model': 2.8395033557577668e-05, 'epoch': 1.39}) +[2026-06-17 15:03:21,543][starVLA.training.train_starvla][INFO] - Step 4720, Loss: {'train/loss': 1.0216400623321533, 'train/grad_norm_pre_clip': 17.916006088256836, 'timing/data': 0.005864736041985452, 'timing/model': 1.2043516979319975, 'learning_rate/qwen_vl_interface': 2.8376400851014214e-06, 'learning_rate/action_model': 2.8376400851014214e-05, 'epoch': 1.39}) +[2026-06-17 15:03:26,320][starVLA.training.train_starvla][INFO] - Step 4721, Loss: {'train/loss': 0.6051053404808044, 'train/grad_norm_pre_clip': 19.57694435119629, 'timing/data': 0.005569580942392349, 'timing/model': 1.1996055839117616, 'learning_rate/qwen_vl_interface': 2.8357773145298823e-06, 'learning_rate/action_model': 2.8357773145298823e-05, 'epoch': 1.39}) +[2026-06-17 15:03:31,011][starVLA.training.train_starvla][INFO] - Step 4722, Loss: {'train/loss': 0.4981675148010254, 'train/grad_norm_pre_clip': 14.175627708435059, 'timing/data': 0.005731783923693001, 'timing/model': 1.2030356959439814, 'learning_rate/qwen_vl_interface': 2.8339150444293055e-06, 'learning_rate/action_model': 2.8339150444293055e-05, 'epoch': 1.39}) +[2026-06-17 15:03:35,732][starVLA.training.train_starvla][INFO] - Step 4723, Loss: {'train/loss': 0.7520690560340881, 'train/grad_norm_pre_clip': 16.542219161987305, 'timing/data': 0.0057975269155576825, 'timing/model': 1.2014927840791643, 'learning_rate/qwen_vl_interface': 2.8320532751857365e-06, 'learning_rate/action_model': 2.8320532751857366e-05, 'epoch': 1.39}) +[2026-06-17 15:03:40,519][starVLA.training.train_starvla][INFO] - Step 4724, Loss: {'train/loss': 0.4912891983985901, 'train/grad_norm_pre_clip': 10.452862739562988, 'timing/data': 0.0051884769927710295, 'timing/model': 1.3400878590764478, 'learning_rate/qwen_vl_interface': 2.830192007185127e-06, 'learning_rate/action_model': 2.830192007185127e-05, 'epoch': 1.39}) +[2026-06-17 15:03:45,178][starVLA.training.train_starvla][INFO] - Step 4725, Loss: {'train/loss': 0.5469761490821838, 'train/grad_norm_pre_clip': 15.299055099487305, 'timing/data': 0.005401331000030041, 'timing/model': 1.2037551020039245, 'learning_rate/qwen_vl_interface': 2.828331240813318e-06, 'learning_rate/action_model': 2.8283312408133176e-05, 'epoch': 1.39}) +[2026-06-17 15:03:49,965][starVLA.training.train_starvla][INFO] - Step 4726, Loss: {'train/loss': 0.7941306829452515, 'train/grad_norm_pre_clip': 13.45187759399414, 'timing/data': 0.005675253923982382, 'timing/model': 1.209983311011456, 'learning_rate/qwen_vl_interface': 2.826470976456049e-06, 'learning_rate/action_model': 2.826470976456049e-05, 'epoch': 1.39}) +[2026-06-17 15:03:54,740][starVLA.training.train_starvla][INFO] - Step 4727, Loss: {'train/loss': 0.562366783618927, 'train/grad_norm_pre_clip': 10.836039543151855, 'timing/data': 0.005706474068574607, 'timing/model': 1.2018716980237514, 'learning_rate/qwen_vl_interface': 2.8246112144989548e-06, 'learning_rate/action_model': 2.8246112144989543e-05, 'epoch': 1.39}) +[2026-06-17 15:03:59,513][starVLA.training.train_starvla][INFO] - Step 4728, Loss: {'train/loss': 0.8248266577720642, 'train/grad_norm_pre_clip': 14.712717056274414, 'timing/data': 0.005109401070512831, 'timing/model': 1.2055124329635873, 'learning_rate/qwen_vl_interface': 2.8227519553275644e-06, 'learning_rate/action_model': 2.8227519553275645e-05, 'epoch': 1.39}) +[2026-06-17 15:04:04,267][starVLA.training.train_starvla][INFO] - Step 4729, Loss: {'train/loss': 0.889617919921875, 'train/grad_norm_pre_clip': 15.711104393005371, 'timing/data': 0.005205118912272155, 'timing/model': 1.2095663520740345, 'learning_rate/qwen_vl_interface': 2.8208931993273066e-06, 'learning_rate/action_model': 2.8208931993273062e-05, 'epoch': 1.39}) +[2026-06-17 15:04:09,042][starVLA.training.train_starvla][INFO] - Step 4730, Loss: {'train/loss': 0.8328592777252197, 'train/grad_norm_pre_clip': 15.78531551361084, 'timing/data': 0.005453543970361352, 'timing/model': 1.3230165789136663, 'learning_rate/qwen_vl_interface': 2.8190349468835006e-06, 'learning_rate/action_model': 2.8190349468835004e-05, 'epoch': 1.39}) +[2026-06-17 15:04:13,695][starVLA.training.train_starvla][INFO] - Step 4731, Loss: {'train/loss': 0.5022341012954712, 'train/grad_norm_pre_clip': 10.928750038146973, 'timing/data': 0.005413872073404491, 'timing/model': 1.2031227849656716, 'learning_rate/qwen_vl_interface': 2.817177198381369e-06, 'learning_rate/action_model': 2.817177198381369e-05, 'epoch': 1.39}) +[2026-06-17 15:04:18,480][starVLA.training.train_starvla][INFO] - Step 4732, Loss: {'train/loss': 0.5698941946029663, 'train/grad_norm_pre_clip': 12.449362754821777, 'timing/data': 0.005202186992391944, 'timing/model': 1.2062663310207427, 'learning_rate/qwen_vl_interface': 2.8153199542060206e-06, 'learning_rate/action_model': 2.8153199542060206e-05, 'epoch': 1.39}) +[2026-06-17 15:04:23,262][starVLA.training.train_starvla][INFO] - Step 4733, Loss: {'train/loss': 0.454315721988678, 'train/grad_norm_pre_clip': 10.22222900390625, 'timing/data': 0.005294178961776197, 'timing/model': 1.204195179976523, 'learning_rate/qwen_vl_interface': 2.813463214742465e-06, 'learning_rate/action_model': 2.813463214742465e-05, 'epoch': 1.39}) +[2026-06-17 15:04:28,047][starVLA.training.train_starvla][INFO] - Step 4734, Loss: {'train/loss': 0.561197817325592, 'train/grad_norm_pre_clip': 10.958624839782715, 'timing/data': 0.005535264033824205, 'timing/model': 1.2049632909474894, 'learning_rate/qwen_vl_interface': 2.8116069803756047e-06, 'learning_rate/action_model': 2.8116069803756047e-05, 'epoch': 1.39}) +[2026-06-17 15:04:32,827][starVLA.training.train_starvla][INFO] - Step 4735, Loss: {'train/loss': 0.5673991441726685, 'train/grad_norm_pre_clip': 45.9787483215332, 'timing/data': 0.005515411030501127, 'timing/model': 1.3322449680417776, 'learning_rate/qwen_vl_interface': 2.809751251490244e-06, 'learning_rate/action_model': 2.809751251490244e-05, 'epoch': 1.39}) +[2026-06-17 15:04:37,481][starVLA.training.train_starvla][INFO] - Step 4736, Loss: {'train/loss': 0.6410055160522461, 'train/grad_norm_pre_clip': 13.281331062316895, 'timing/data': 0.005579052958637476, 'timing/model': 1.2028943339828402, 'learning_rate/qwen_vl_interface': 2.807896028471076e-06, 'learning_rate/action_model': 2.807896028471076e-05, 'epoch': 1.39}) +[2026-06-17 15:04:42,263][starVLA.training.train_starvla][INFO] - Step 4737, Loss: {'train/loss': 0.5959762334823608, 'train/grad_norm_pre_clip': 31.861068725585938, 'timing/data': 0.005220577004365623, 'timing/model': 1.2046500579454005, 'learning_rate/qwen_vl_interface': 2.806041311702688e-06, 'learning_rate/action_model': 2.806041311702688e-05, 'epoch': 1.39}) +[2026-06-17 15:04:46,969][starVLA.training.train_starvla][INFO] - Step 4738, Loss: {'train/loss': 0.3651607930660248, 'train/grad_norm_pre_clip': 9.5445556640625, 'timing/data': 0.00589480996131897, 'timing/model': 1.2047882030019537, 'learning_rate/qwen_vl_interface': 2.8041871015695626e-06, 'learning_rate/action_model': 2.8041871015695626e-05, 'epoch': 1.39}) +[2026-06-17 15:04:51,706][starVLA.training.train_starvla][INFO] - Step 4739, Loss: {'train/loss': 0.704757571220398, 'train/grad_norm_pre_clip': 12.281255722045898, 'timing/data': 0.005481360014528036, 'timing/model': 1.2031731409952044, 'learning_rate/qwen_vl_interface': 2.802333398456086e-06, 'learning_rate/action_model': 2.8023333984560856e-05, 'epoch': 1.39}) +[2026-06-17 15:04:56,489][starVLA.training.train_starvla][INFO] - Step 4740, Loss: {'train/loss': 0.762779712677002, 'train/grad_norm_pre_clip': 16.078588485717773, 'timing/data': 0.005609382991679013, 'timing/model': 1.2006588110234588, 'learning_rate/qwen_vl_interface': 2.8004802027465296e-06, 'learning_rate/action_model': 2.8004802027465293e-05, 'epoch': 1.39}) +[2026-06-17 15:05:01,274][starVLA.training.train_starvla][INFO] - Step 4741, Loss: {'train/loss': 0.4566078782081604, 'train/grad_norm_pre_clip': 10.193381309509277, 'timing/data': 0.005335050052963197, 'timing/model': 1.3340259549440816, 'learning_rate/qwen_vl_interface': 2.7986275148250624e-06, 'learning_rate/action_model': 2.7986275148250622e-05, 'epoch': 1.4}) +[2026-06-17 15:05:05,964][starVLA.training.train_starvla][INFO] - Step 4742, Loss: {'train/loss': 0.7486206889152527, 'train/grad_norm_pre_clip': 20.737239837646484, 'timing/data': 0.005388068966567516, 'timing/model': 1.2047910830006003, 'learning_rate/qwen_vl_interface': 2.7967753350757498e-06, 'learning_rate/action_model': 2.7967753350757497e-05, 'epoch': 1.4}) +[2026-06-17 15:05:10,744][starVLA.training.train_starvla][INFO] - Step 4743, Loss: {'train/loss': 0.7120180130004883, 'train/grad_norm_pre_clip': 12.948407173156738, 'timing/data': 0.005492921103723347, 'timing/model': 1.2047338800039142, 'learning_rate/qwen_vl_interface': 2.7949236638825496e-06, 'learning_rate/action_model': 2.7949236638825495e-05, 'epoch': 1.4}) +[2026-06-17 15:05:15,519][starVLA.training.train_starvla][INFO] - Step 4744, Loss: {'train/loss': 0.5939052700996399, 'train/grad_norm_pre_clip': 14.875606536865234, 'timing/data': 0.005382461939007044, 'timing/model': 1.2037007679464296, 'learning_rate/qwen_vl_interface': 2.7930725016293155e-06, 'learning_rate/action_model': 2.793072501629315e-05, 'epoch': 1.4}) +[2026-06-17 15:05:20,276][starVLA.training.train_starvla][INFO] - Step 4745, Loss: {'train/loss': 0.31329047679901123, 'train/grad_norm_pre_clip': 8.868605613708496, 'timing/data': 0.005750717944465578, 'timing/model': 1.217937896028161, 'learning_rate/qwen_vl_interface': 2.7912218486997954e-06, 'learning_rate/action_model': 2.7912218486997954e-05, 'epoch': 1.4}) +[2026-06-17 15:05:25,051][starVLA.training.train_starvla][INFO] - Step 4746, Loss: {'train/loss': 0.8531586527824402, 'train/grad_norm_pre_clip': 15.630653381347656, 'timing/data': 0.005542889004573226, 'timing/model': 1.2048404950182885, 'learning_rate/qwen_vl_interface': 2.7893717054776327e-06, 'learning_rate/action_model': 2.7893717054776327e-05, 'epoch': 1.4}) +[2026-06-17 15:05:29,850][starVLA.training.train_starvla][INFO] - Step 4747, Loss: {'train/loss': 0.6545007824897766, 'train/grad_norm_pre_clip': 13.564635276794434, 'timing/data': 0.005313439993187785, 'timing/model': 1.3389615420019254, 'learning_rate/qwen_vl_interface': 2.7875220723463624e-06, 'learning_rate/action_model': 2.7875220723463625e-05, 'epoch': 1.4}) +[2026-06-17 15:05:34,513][starVLA.training.train_starvla][INFO] - Step 4748, Loss: {'train/loss': 0.7885327935218811, 'train/grad_norm_pre_clip': 13.674367904663086, 'timing/data': 0.0053358550649136305, 'timing/model': 1.2044016310246661, 'learning_rate/qwen_vl_interface': 2.7856729496894165e-06, 'learning_rate/action_model': 2.7856729496894162e-05, 'epoch': 1.4}) +[2026-06-17 15:05:39,319][starVLA.training.train_starvla][INFO] - Step 4749, Loss: {'train/loss': 0.6351019144058228, 'train/grad_norm_pre_clip': 9.638713836669922, 'timing/data': 0.0057124600280076265, 'timing/model': 1.2036632458912209, 'learning_rate/qwen_vl_interface': 2.783824337890118e-06, 'learning_rate/action_model': 2.783824337890118e-05, 'epoch': 1.4}) +[2026-06-17 15:05:44,115][starVLA.training.train_starvla][INFO] - Step 4750, Loss: {'train/loss': 0.6747941970825195, 'train/grad_norm_pre_clip': 15.036784172058105, 'timing/data': 0.005314692039974034, 'timing/model': 1.2034762039547786, 'learning_rate/qwen_vl_interface': 2.7819762373316916e-06, 'learning_rate/action_model': 2.7819762373316917e-05, 'epoch': 1.4}) +[2026-06-17 15:05:48,930][starVLA.training.train_starvla][INFO] - Step 4751, Loss: {'train/loss': 0.5114370584487915, 'train/grad_norm_pre_clip': 10.140401840209961, 'timing/data': 0.005289386957883835, 'timing/model': 1.2190002849092707, 'learning_rate/qwen_vl_interface': 2.7801286483972458e-06, 'learning_rate/action_model': 2.7801286483972457e-05, 'epoch': 1.4}) +[2026-06-17 15:05:53,802][starVLA.training.train_starvla][INFO] - Step 4752, Loss: {'train/loss': 0.5015888810157776, 'train/grad_norm_pre_clip': 21.193073272705078, 'timing/data': 0.005114104016683996, 'timing/model': 1.2040865150047466, 'learning_rate/qwen_vl_interface': 2.778281571469786e-06, 'learning_rate/action_model': 2.7782815714697858e-05, 'epoch': 1.4}) +[2026-06-17 15:05:58,490][starVLA.training.train_starvla][INFO] - Step 4753, Loss: {'train/loss': 0.8885238766670227, 'train/grad_norm_pre_clip': 17.87294578552246, 'timing/data': 0.005443596979603171, 'timing/model': 1.2428468270227313, 'learning_rate/qwen_vl_interface': 2.776435006932219e-06, 'learning_rate/action_model': 2.7764350069322187e-05, 'epoch': 1.4}) +[2026-06-17 15:06:03,159][starVLA.training.train_starvla][INFO] - Step 4754, Loss: {'train/loss': 0.8042157292366028, 'train/grad_norm_pre_clip': 13.569607734680176, 'timing/data': 0.005144277936778963, 'timing/model': 1.2208624969935045, 'learning_rate/qwen_vl_interface': 2.7745889551673355e-06, 'learning_rate/action_model': 2.7745889551673353e-05, 'epoch': 1.4}) +[2026-06-17 15:06:07,896][starVLA.training.train_starvla][INFO] - Step 4755, Loss: {'train/loss': 0.9493528008460999, 'train/grad_norm_pre_clip': 14.540553092956543, 'timing/data': 0.005825211061164737, 'timing/model': 1.2109257690608501, 'learning_rate/qwen_vl_interface': 2.772743416557826e-06, 'learning_rate/action_model': 2.772743416557826e-05, 'epoch': 1.4}) +[2026-06-17 15:06:12,699][starVLA.training.train_starvla][INFO] - Step 4756, Loss: {'train/loss': 0.7636526823043823, 'train/grad_norm_pre_clip': 13.720293045043945, 'timing/data': 0.005394494044594467, 'timing/model': 1.2093705580336973, 'learning_rate/qwen_vl_interface': 2.770898391486272e-06, 'learning_rate/action_model': 2.7708983914862718e-05, 'epoch': 1.4}) +[2026-06-17 15:06:17,490][starVLA.training.train_starvla][INFO] - Step 4757, Loss: {'train/loss': 0.3881043791770935, 'train/grad_norm_pre_clip': 10.9183988571167, 'timing/data': 0.005373629042878747, 'timing/model': 1.2060038700001314, 'learning_rate/qwen_vl_interface': 2.7690538803351484e-06, 'learning_rate/action_model': 2.7690538803351484e-05, 'epoch': 1.4}) +[2026-06-17 15:06:22,296][starVLA.training.train_starvla][INFO] - Step 4758, Loss: {'train/loss': 0.7461264729499817, 'train/grad_norm_pre_clip': 19.389677047729492, 'timing/data': 0.0055475760018453, 'timing/model': 1.3387380329659209, 'learning_rate/qwen_vl_interface': 2.7672098834868253e-06, 'learning_rate/action_model': 2.7672098834868253e-05, 'epoch': 1.4}) +[2026-06-17 15:06:27,088][starVLA.training.train_starvla][INFO] - Step 4759, Loss: {'train/loss': 0.44868332147598267, 'train/grad_norm_pre_clip': 13.30328369140625, 'timing/data': 0.005379495094530284, 'timing/model': 1.2529899090295658, 'learning_rate/qwen_vl_interface': 2.7653664013235655e-06, 'learning_rate/action_model': 2.7653664013235653e-05, 'epoch': 1.4}) +[2026-06-17 15:06:31,761][starVLA.training.train_starvla][INFO] - Step 4760, Loss: {'train/loss': 0.5307790040969849, 'train/grad_norm_pre_clip': 14.630680084228516, 'timing/data': 0.005451628938317299, 'timing/model': 1.2103650610661134, 'learning_rate/qwen_vl_interface': 2.763523434227524e-06, 'learning_rate/action_model': 2.763523434227524e-05, 'epoch': 1.4}) +[2026-06-17 15:06:36,516][starVLA.training.train_starvla][INFO] - Step 4761, Loss: {'train/loss': 0.4258926510810852, 'train/grad_norm_pre_clip': 11.667268753051758, 'timing/data': 0.005431915051303804, 'timing/model': 1.2123473190004006, 'learning_rate/qwen_vl_interface': 2.76168098258075e-06, 'learning_rate/action_model': 2.76168098258075e-05, 'epoch': 1.4}) +[2026-06-17 15:06:41,310][starVLA.training.train_starvla][INFO] - Step 4762, Loss: {'train/loss': 0.5407896041870117, 'train/grad_norm_pre_clip': 28.85993194580078, 'timing/data': 0.0053541610250249505, 'timing/model': 1.2034604990622029, 'learning_rate/qwen_vl_interface': 2.7598390467651865e-06, 'learning_rate/action_model': 2.7598390467651863e-05, 'epoch': 1.4}) +[2026-06-17 15:06:46,119][starVLA.training.train_starvla][INFO] - Step 4763, Loss: {'train/loss': 0.5552332997322083, 'train/grad_norm_pre_clip': 12.770783424377441, 'timing/data': 0.005529204965569079, 'timing/model': 1.2206368109909818, 'learning_rate/qwen_vl_interface': 2.757997627162665e-06, 'learning_rate/action_model': 2.7579976271626652e-05, 'epoch': 1.4}) +[2026-06-17 15:06:50,935][starVLA.training.train_starvla][INFO] - Step 4764, Loss: {'train/loss': 0.7579445838928223, 'train/grad_norm_pre_clip': 16.701496124267578, 'timing/data': 0.00555464590433985, 'timing/model': 1.347581559093669, 'learning_rate/qwen_vl_interface': 2.7561567241549203e-06, 'learning_rate/action_model': 2.7561567241549202e-05, 'epoch': 1.4}) +[2026-06-17 15:06:55,713][starVLA.training.train_starvla][INFO] - Step 4765, Loss: {'train/loss': 0.7952565550804138, 'train/grad_norm_pre_clip': 14.085593223571777, 'timing/data': 0.00517205405049026, 'timing/model': 1.3233323260210454, 'learning_rate/qwen_vl_interface': 2.7543163381235705e-06, 'learning_rate/action_model': 2.7543163381235702e-05, 'epoch': 1.4}) +[2026-06-17 15:07:00,372][starVLA.training.train_starvla][INFO] - Step 4766, Loss: {'train/loss': 0.6321970820426941, 'train/grad_norm_pre_clip': 10.932049751281738, 'timing/data': 0.005357938003726304, 'timing/model': 1.2035821300232783, 'learning_rate/qwen_vl_interface': 2.752476469450128e-06, 'learning_rate/action_model': 2.7524764694501283e-05, 'epoch': 1.4}) +[2026-06-17 15:07:05,142][starVLA.training.train_starvla][INFO] - Step 4767, Loss: {'train/loss': 0.8690934777259827, 'train/grad_norm_pre_clip': 16.557342529296875, 'timing/data': 0.005305730039253831, 'timing/model': 1.2020587020087987, 'learning_rate/qwen_vl_interface': 2.750637118515999e-06, 'learning_rate/action_model': 2.750637118515999e-05, 'epoch': 1.4}) +[2026-06-17 15:07:09,934][starVLA.training.train_starvla][INFO] - Step 4768, Loss: {'train/loss': 0.858840823173523, 'train/grad_norm_pre_clip': 18.046144485473633, 'timing/data': 0.005556644988246262, 'timing/model': 1.2031002349685878, 'learning_rate/qwen_vl_interface': 2.7487982857024875e-06, 'learning_rate/action_model': 2.7487982857024876e-05, 'epoch': 1.4}) +[2026-06-17 15:07:14,724][starVLA.training.train_starvla][INFO] - Step 4769, Loss: {'train/loss': 0.7256842851638794, 'train/grad_norm_pre_clip': 15.35568618774414, 'timing/data': 0.005655303015373647, 'timing/model': 1.20406783092767, 'learning_rate/qwen_vl_interface': 2.746959971390782e-06, 'learning_rate/action_model': 2.7469599713907817e-05, 'epoch': 1.4}) +[2026-06-17 15:07:19,466][starVLA.training.train_starvla][INFO] - Step 4770, Loss: {'train/loss': 0.632341206073761, 'train/grad_norm_pre_clip': 14.07010269165039, 'timing/data': 0.005496431956999004, 'timing/model': 1.2769232330610976, 'learning_rate/qwen_vl_interface': 2.745122175961967e-06, 'learning_rate/action_model': 2.7451221759619672e-05, 'epoch': 1.4}) +[2026-06-17 15:07:24,130][starVLA.training.train_starvla][INFO] - Step 4771, Loss: {'train/loss': 0.8365180492401123, 'train/grad_norm_pre_clip': 16.578561782836914, 'timing/data': 0.005313745932653546, 'timing/model': 1.2167800089810044, 'learning_rate/qwen_vl_interface': 2.743284899797022e-06, 'learning_rate/action_model': 2.7432848997970218e-05, 'epoch': 1.4}) +[2026-06-17 15:07:28,928][starVLA.training.train_starvla][INFO] - Step 4772, Loss: {'train/loss': 0.5392757058143616, 'train/grad_norm_pre_clip': 12.882184028625488, 'timing/data': 0.0054221199825406075, 'timing/model': 1.204040788114071, 'learning_rate/qwen_vl_interface': 2.741448143276813e-06, 'learning_rate/action_model': 2.741448143276813e-05, 'epoch': 1.4}) +[2026-06-17 15:07:33,713][starVLA.training.train_starvla][INFO] - Step 4773, Loss: {'train/loss': 0.6097058653831482, 'train/grad_norm_pre_clip': 13.208940505981445, 'timing/data': 0.005653396947309375, 'timing/model': 1.2037652500439435, 'learning_rate/qwen_vl_interface': 2.7396119067821043e-06, 'learning_rate/action_model': 2.7396119067821043e-05, 'epoch': 1.4}) +[2026-06-17 15:07:38,507][starVLA.training.train_starvla][INFO] - Step 4774, Loss: {'train/loss': 0.39309725165367126, 'train/grad_norm_pre_clip': 12.385068893432617, 'timing/data': 0.005460978019982576, 'timing/model': 1.204683059011586, 'learning_rate/qwen_vl_interface': 2.737776190693548e-06, 'learning_rate/action_model': 2.7377761906935477e-05, 'epoch': 1.4}) +[2026-06-17 15:07:43,293][starVLA.training.train_starvla][INFO] - Step 4775, Loss: {'train/loss': 0.9124230146408081, 'train/grad_norm_pre_clip': 13.5758695602417, 'timing/data': 0.005568091059103608, 'timing/model': 1.3382547500077635, 'learning_rate/qwen_vl_interface': 2.7359409953916914e-06, 'learning_rate/action_model': 2.7359409953916913e-05, 'epoch': 1.41}) +[2026-06-17 15:07:48,046][starVLA.training.train_starvla][INFO] - Step 4776, Loss: {'train/loss': 0.7374001145362854, 'train/grad_norm_pre_clip': 14.116978645324707, 'timing/data': 0.005819473997689784, 'timing/model': 1.278775576967746, 'learning_rate/qwen_vl_interface': 2.734106321256971e-06, 'learning_rate/action_model': 2.7341063212569705e-05, 'epoch': 1.41}) +[2026-06-17 15:07:52,707][starVLA.training.train_starvla][INFO] - Step 4777, Loss: {'train/loss': 0.9118639826774597, 'train/grad_norm_pre_clip': 15.950417518615723, 'timing/data': 0.005434337072074413, 'timing/model': 1.2030139099806547, 'learning_rate/qwen_vl_interface': 2.732272168669718e-06, 'learning_rate/action_model': 2.732272168669718e-05, 'epoch': 1.41}) +[2026-06-17 15:07:57,508][starVLA.training.train_starvla][INFO] - Step 4778, Loss: {'train/loss': 0.7312358617782593, 'train/grad_norm_pre_clip': 14.715873718261719, 'timing/data': 0.005415488965809345, 'timing/model': 1.2058997889980674, 'learning_rate/qwen_vl_interface': 2.7304385380101512e-06, 'learning_rate/action_model': 2.730438538010151e-05, 'epoch': 1.41}) +[2026-06-17 15:08:02,324][starVLA.training.train_starvla][INFO] - Step 4779, Loss: {'train/loss': 0.511612594127655, 'train/grad_norm_pre_clip': 9.752662658691406, 'timing/data': 0.005118191009387374, 'timing/model': 1.2132265219697729, 'learning_rate/qwen_vl_interface': 2.7286054296583904e-06, 'learning_rate/action_model': 2.72860542965839e-05, 'epoch': 1.41}) +[2026-06-17 15:08:07,110][starVLA.training.train_starvla][INFO] - Step 4780, Loss: {'train/loss': 0.5670584440231323, 'train/grad_norm_pre_clip': 11.503719329833984, 'timing/data': 0.005684546078555286, 'timing/model': 1.2042712499387562, 'learning_rate/qwen_vl_interface': 2.726772843994435e-06, 'learning_rate/action_model': 2.7267728439944352e-05, 'epoch': 1.41}) +[2026-06-17 15:08:11,888][starVLA.training.train_starvla][INFO] - Step 4781, Loss: {'train/loss': 0.4776040017604828, 'train/grad_norm_pre_clip': 12.591992378234863, 'timing/data': 0.005319847026839852, 'timing/model': 1.2042993539944291, 'learning_rate/qwen_vl_interface': 2.7249407813981834e-06, 'learning_rate/action_model': 2.724940781398183e-05, 'epoch': 1.41}) +[2026-06-17 15:08:16,643][starVLA.training.train_starvla][INFO] - Step 4782, Loss: {'train/loss': 0.8025493025779724, 'train/grad_norm_pre_clip': 15.10872745513916, 'timing/data': 0.005622120923362672, 'timing/model': 1.2900321698980406, 'learning_rate/qwen_vl_interface': 2.7231092422494255e-06, 'learning_rate/action_model': 2.7231092422494253e-05, 'epoch': 1.41}) +[2026-06-17 15:08:21,328][starVLA.training.train_starvla][INFO] - Step 4783, Loss: {'train/loss': 0.6439424753189087, 'train/grad_norm_pre_clip': 15.72158432006836, 'timing/data': 0.005582205951213837, 'timing/model': 1.216161179007031, 'learning_rate/qwen_vl_interface': 2.7212782269278417e-06, 'learning_rate/action_model': 2.7212782269278413e-05, 'epoch': 1.41}) +[2026-06-17 15:08:26,129][starVLA.training.train_starvla][INFO] - Step 4784, Loss: {'train/loss': 0.4175349175930023, 'train/grad_norm_pre_clip': 12.054609298706055, 'timing/data': 0.005843056016601622, 'timing/model': 1.2041997499763966, 'learning_rate/qwen_vl_interface': 2.719447735813004e-06, 'learning_rate/action_model': 2.719447735813004e-05, 'epoch': 1.41}) +[2026-06-17 15:08:30,916][starVLA.training.train_starvla][INFO] - Step 4785, Loss: {'train/loss': 0.5926187038421631, 'train/grad_norm_pre_clip': 12.948677062988281, 'timing/data': 0.005088325007818639, 'timing/model': 1.2048686379566789, 'learning_rate/qwen_vl_interface': 2.7176177692843697e-06, 'learning_rate/action_model': 2.7176177692843697e-05, 'epoch': 1.41}) +[2026-06-17 15:08:35,709][starVLA.training.train_starvla][INFO] - Step 4786, Loss: {'train/loss': 0.4700808525085449, 'train/grad_norm_pre_clip': 11.733914375305176, 'timing/data': 0.0053423140197992325, 'timing/model': 1.2090106239775196, 'learning_rate/qwen_vl_interface': 2.7157883277212985e-06, 'learning_rate/action_model': 2.7157883277212985e-05, 'epoch': 1.41}) +[2026-06-17 15:08:40,506][starVLA.training.train_starvla][INFO] - Step 4787, Loss: {'train/loss': 0.6924450397491455, 'train/grad_norm_pre_clip': 17.073625564575195, 'timing/data': 0.005422780988737941, 'timing/model': 1.3350365250371397, 'learning_rate/qwen_vl_interface': 2.7139594115030347e-06, 'learning_rate/action_model': 2.7139594115030348e-05, 'epoch': 1.41}) +[2026-06-17 15:08:45,168][starVLA.training.train_starvla][INFO] - Step 4788, Loss: {'train/loss': 0.5769887566566467, 'train/grad_norm_pre_clip': 11.426956176757812, 'timing/data': 0.005590433022007346, 'timing/model': 1.2031370269833133, 'learning_rate/qwen_vl_interface': 2.712131021008714e-06, 'learning_rate/action_model': 2.7121310210087137e-05, 'epoch': 1.41}) +[2026-06-17 15:08:49,973][starVLA.training.train_starvla][INFO] - Step 4789, Loss: {'train/loss': 0.6140138506889343, 'train/grad_norm_pre_clip': 12.041648864746094, 'timing/data': 0.005470982985571027, 'timing/model': 1.204746623057872, 'learning_rate/qwen_vl_interface': 2.7103031566173622e-06, 'learning_rate/action_model': 2.710303156617362e-05, 'epoch': 1.41}) +[2026-06-17 15:08:54,750][starVLA.training.train_starvla][INFO] - Step 4790, Loss: {'train/loss': 0.5702171325683594, 'train/grad_norm_pre_clip': 12.666214942932129, 'timing/data': 0.005235153948888183, 'timing/model': 1.201702581020072, 'learning_rate/qwen_vl_interface': 2.7084758187078995e-06, 'learning_rate/action_model': 2.7084758187078997e-05, 'epoch': 1.41}) +[2026-06-17 15:08:59,453][starVLA.training.train_starvla][INFO] - Step 4791, Loss: {'train/loss': 0.36400657892227173, 'train/grad_norm_pre_clip': 10.235698699951172, 'timing/data': 0.005680739996023476, 'timing/model': 1.2040917270351201, 'learning_rate/qwen_vl_interface': 2.7066490076591344e-06, 'learning_rate/action_model': 2.7066490076591343e-05, 'epoch': 1.41}) +[2026-06-17 15:09:04,250][starVLA.training.train_starvla][INFO] - Step 4792, Loss: {'train/loss': 0.6358656883239746, 'train/grad_norm_pre_clip': 10.5620698928833, 'timing/data': 0.005517498007975519, 'timing/model': 1.2040541339665651, 'learning_rate/qwen_vl_interface': 2.704822723849765e-06, 'learning_rate/action_model': 2.7048227238497648e-05, 'epoch': 1.41}) +[2026-06-17 15:09:09,029][starVLA.training.train_starvla][INFO] - Step 4793, Loss: {'train/loss': 0.7143321633338928, 'train/grad_norm_pre_clip': 13.491022109985352, 'timing/data': 0.00507677101995796, 'timing/model': 1.3328955689212307, 'learning_rate/qwen_vl_interface': 2.702996967658388e-06, 'learning_rate/action_model': 2.7029969676583876e-05, 'epoch': 1.41}) +[2026-06-17 15:09:13,686][starVLA.training.train_starvla][INFO] - Step 4794, Loss: {'train/loss': 0.33975470066070557, 'train/grad_norm_pre_clip': 9.742382049560547, 'timing/data': 0.005866873078048229, 'timing/model': 1.2043960359878838, 'learning_rate/qwen_vl_interface': 2.7011717394634767e-06, 'learning_rate/action_model': 2.701171739463477e-05, 'epoch': 1.41}) +[2026-06-17 15:09:18,481][starVLA.training.train_starvla][INFO] - Step 4795, Loss: {'train/loss': 0.5672757625579834, 'train/grad_norm_pre_clip': 12.853856086730957, 'timing/data': 0.00550951506011188, 'timing/model': 1.205149336019531, 'learning_rate/qwen_vl_interface': 2.699347039643408e-06, 'learning_rate/action_model': 2.699347039643408e-05, 'epoch': 1.41}) +[2026-06-17 15:09:23,276][starVLA.training.train_starvla][INFO] - Step 4796, Loss: {'train/loss': 0.9327154159545898, 'train/grad_norm_pre_clip': 15.491134643554688, 'timing/data': 0.005272760055959225, 'timing/model': 1.2053021730389446, 'learning_rate/qwen_vl_interface': 2.6975228685764398e-06, 'learning_rate/action_model': 2.69752286857644e-05, 'epoch': 1.41}) +[2026-06-17 15:09:28,056][starVLA.training.train_starvla][INFO] - Step 4797, Loss: {'train/loss': 0.8644301891326904, 'train/grad_norm_pre_clip': 19.431962966918945, 'timing/data': 0.005442122928798199, 'timing/model': 1.2049149059457704, 'learning_rate/qwen_vl_interface': 2.6956992266407307e-06, 'learning_rate/action_model': 2.6956992266407306e-05, 'epoch': 1.41}) +[2026-06-17 15:09:32,809][starVLA.training.train_starvla][INFO] - Step 4798, Loss: {'train/loss': 1.0406653881072998, 'train/grad_norm_pre_clip': 19.829557418823242, 'timing/data': 0.005752566969022155, 'timing/model': 1.201900671934709, 'learning_rate/qwen_vl_interface': 2.693876114214321e-06, 'learning_rate/action_model': 2.6938761142143208e-05, 'epoch': 1.41}) +[2026-06-17 15:09:37,582][starVLA.training.train_starvla][INFO] - Step 4799, Loss: {'train/loss': 0.4446384012699127, 'train/grad_norm_pre_clip': 11.939502716064453, 'timing/data': 0.005458649015054107, 'timing/model': 1.3228721109917387, 'learning_rate/qwen_vl_interface': 2.6920535316751417e-06, 'learning_rate/action_model': 2.6920535316751416e-05, 'epoch': 1.41}) +[2026-06-17 15:11:19,336][starVLA.training.train_starvla][INFO] - Step 4800, Loss: {'train/loss': 0.5891022682189941, 'train/grad_norm_pre_clip': 11.55716323852539, 'eval/loss': 0.7238571643829346, 'eval/latency_0/loss': 0.7238571643829346, 'eval/demon_fire_precision': 0.8977812995245642, 'eval/demon_fire_recall': 0.9410299003322259, 'eval/demon_fire_f1': 0.91889699918897, 'eval/demon_fire_rate_model': 0.7018909899888766, 'eval/demon_fire_rate_teacher': 0.6696329254727474, 'eval/demon_left_precision': 0.7998938428874734, 'eval/demon_left_recall': 0.8538243626062323, 'eval/demon_left_f1': 0.825979720471362, 'eval/demon_left_rate_model': 0.5239154616240267, 'eval/demon_left_rate_teacher': 0.49082313681868744, 'eval/demon_right_precision': 0.8421985815602837, 'eval/demon_right_recall': 0.8050847457627118, 'eval/demon_right_f1': 0.8232235701906412, 'eval/demon_right_rate_model': 0.47052280311457173, 'eval/demon_right_rate_teacher': 0.4922135706340378, 'eval/demon_move_f1': 0.8246016453310017, 'eval/demon_cc_f1': 0.8884190856896639, 'eval/latency_0/demon_fire_precision': 0.8977812995245642, 'eval/latency_0/demon_fire_recall': 0.9410299003322259, 'eval/latency_0/demon_fire_f1': 0.91889699918897, 'eval/latency_0/demon_fire_rate_model': 0.7018909899888766, 'eval/latency_0/demon_fire_rate_teacher': 0.6696329254727474, 'eval/latency_0/demon_left_precision': 0.7998938428874734, 'eval/latency_0/demon_left_recall': 0.8538243626062323, 'eval/latency_0/demon_left_f1': 0.825979720471362, 'eval/latency_0/demon_left_rate_model': 0.5239154616240267, 'eval/latency_0/demon_left_rate_teacher': 0.49082313681868744, 'eval/latency_0/demon_right_precision': 0.8421985815602837, 'eval/latency_0/demon_right_recall': 0.8050847457627118, 'eval/latency_0/demon_right_f1': 0.8232235701906412, 'eval/latency_0/demon_right_rate_model': 0.47052280311457173, 'eval/latency_0/demon_right_rate_teacher': 0.4922135706340378, 'eval/latency_0/demon_move_f1': 0.8246016453310017, 'eval/latency_0/demon_cc_f1': 0.8884190856896639, 'timing/data': 0.005047990940511227, 'timing/model': 1.203027916024439, 'learning_rate/qwen_vl_interface': 2.6902314794010164e-06, 'learning_rate/action_model': 2.6902314794010165e-05, 'epoch': 1.41}) +[2026-06-17 15:11:19,336][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_4800_state +[2026-06-17 15:11:19,339][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 15:11:23,007][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_4800_state/model.safetensors +[2026-06-17 15:11:30,658][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_4800_state/optimizer.bin +[2026-06-17 15:11:30,659][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_4800_state/sampler.bin +[2026-06-17 15:11:30,659][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_4800_state/sampler_1.bin +[2026-06-17 15:11:30,659][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_4800_state/random_states_0.pkl +[2026-06-17 15:11:30,713][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4800 +[2026-06-17 15:11:30,713][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 15:11:30,723][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 15:11:35,429][starVLA.training.train_starvla][INFO] - Step 4801, Loss: {'train/loss': 0.6875439882278442, 'train/grad_norm_pre_clip': 13.3368558883667, 'timing/data': 0.005054707056842744, 'timing/model': 1.3088113489793614, 'learning_rate/qwen_vl_interface': 2.6884099577696615e-06, 'learning_rate/action_model': 2.6884099577696615e-05, 'epoch': 1.41}) +[2026-06-17 15:11:40,033][starVLA.training.train_starvla][INFO] - Step 4802, Loss: {'train/loss': 0.7639690637588501, 'train/grad_norm_pre_clip': 14.733104705810547, 'timing/data': 0.005580781027674675, 'timing/model': 1.193866723915562, 'learning_rate/qwen_vl_interface': 2.686588967158679e-06, 'learning_rate/action_model': 2.6865889671586792e-05, 'epoch': 1.41}) +[2026-06-17 15:11:44,685][starVLA.training.train_starvla][INFO] - Step 4803, Loss: {'train/loss': 0.7898730039596558, 'train/grad_norm_pre_clip': 16.36418914794922, 'timing/data': 0.004519128007814288, 'timing/model': 1.1911829420132563, 'learning_rate/qwen_vl_interface': 2.6847685079455614e-06, 'learning_rate/action_model': 2.6847685079455613e-05, 'epoch': 1.41}) +[2026-06-17 15:11:49,376][starVLA.training.train_starvla][INFO] - Step 4804, Loss: {'train/loss': 0.4132736623287201, 'train/grad_norm_pre_clip': 10.032523155212402, 'timing/data': 0.0044402769999578595, 'timing/model': 1.2027568509802222, 'learning_rate/qwen_vl_interface': 2.6829485805076927e-06, 'learning_rate/action_model': 2.6829485805076925e-05, 'epoch': 1.41}) +[2026-06-17 15:11:54,082][starVLA.training.train_starvla][INFO] - Step 4805, Loss: {'train/loss': 0.6589788794517517, 'train/grad_norm_pre_clip': 13.116270065307617, 'timing/data': 0.0053391719702631235, 'timing/model': 1.190791542059742, 'learning_rate/qwen_vl_interface': 2.6811291852223452e-06, 'learning_rate/action_model': 2.6811291852223454e-05, 'epoch': 1.41}) +[2026-06-17 15:11:58,842][starVLA.training.train_starvla][INFO] - Step 4806, Loss: {'train/loss': 0.7150771021842957, 'train/grad_norm_pre_clip': 15.56987190246582, 'timing/data': 0.005050417967140675, 'timing/model': 1.3365958409849554, 'learning_rate/qwen_vl_interface': 2.679310322466683e-06, 'learning_rate/action_model': 2.6793103224666827e-05, 'epoch': 1.41}) +[2026-06-17 15:12:03,467][starVLA.training.train_starvla][INFO] - Step 4807, Loss: {'train/loss': 0.6513580679893494, 'train/grad_norm_pre_clip': 14.408174514770508, 'timing/data': 0.004868265008553863, 'timing/model': 1.1969669830286875, 'learning_rate/qwen_vl_interface': 2.6774919926177563e-06, 'learning_rate/action_model': 2.677491992617756e-05, 'epoch': 1.41}) +[2026-06-17 15:12:08,222][starVLA.training.train_starvla][INFO] - Step 4808, Loss: {'train/loss': 0.6564390063285828, 'train/grad_norm_pre_clip': 13.622512817382812, 'timing/data': 0.005159992026165128, 'timing/model': 1.1976228079292923, 'learning_rate/qwen_vl_interface': 2.675674196052507e-06, 'learning_rate/action_model': 2.6756741960525072e-05, 'epoch': 1.41}) +[2026-06-17 15:12:12,986][starVLA.training.train_starvla][INFO] - Step 4809, Loss: {'train/loss': 0.8943207263946533, 'train/grad_norm_pre_clip': 25.67020606994629, 'timing/data': 0.005120510933920741, 'timing/model': 1.1988398430403322, 'learning_rate/qwen_vl_interface': 2.6738569331477677e-06, 'learning_rate/action_model': 2.6738569331477674e-05, 'epoch': 1.42}) +[2026-06-17 15:12:17,707][starVLA.training.train_starvla][INFO] - Step 4810, Loss: {'train/loss': 0.6956889629364014, 'train/grad_norm_pre_clip': 13.995007514953613, 'timing/data': 0.004813183913938701, 'timing/model': 1.1986466980306432, 'learning_rate/qwen_vl_interface': 2.672040204280258e-06, 'learning_rate/action_model': 2.672040204280258e-05, 'epoch': 1.42}) +[2026-06-17 15:12:22,474][starVLA.training.train_starvla][INFO] - Step 4811, Loss: {'train/loss': 0.9594998359680176, 'train/grad_norm_pre_clip': 15.782657623291016, 'timing/data': 0.0050119549268856645, 'timing/model': 1.1994496489642188, 'learning_rate/qwen_vl_interface': 2.6702240098265864e-06, 'learning_rate/action_model': 2.6702240098265864e-05, 'epoch': 1.42}) +[2026-06-17 15:12:27,258][starVLA.training.train_starvla][INFO] - Step 4812, Loss: {'train/loss': 0.5442978143692017, 'train/grad_norm_pre_clip': 13.298062324523926, 'timing/data': 0.005150089971721172, 'timing/model': 1.330579828005284, 'learning_rate/qwen_vl_interface': 2.6684083501632573e-06, 'learning_rate/action_model': 2.6684083501632574e-05, 'epoch': 1.42}) +[2026-06-17 15:12:31,912][starVLA.training.train_starvla][INFO] - Step 4813, Loss: {'train/loss': 0.7860032916069031, 'train/grad_norm_pre_clip': 17.42573356628418, 'timing/data': 0.004847073927521706, 'timing/model': 1.206979454960674, 'learning_rate/qwen_vl_interface': 2.6665932256666525e-06, 'learning_rate/action_model': 2.6665932256666527e-05, 'epoch': 1.42}) +[2026-06-17 15:12:36,709][starVLA.training.train_starvla][INFO] - Step 4814, Loss: {'train/loss': 0.43946436047554016, 'train/grad_norm_pre_clip': 9.72611141204834, 'timing/data': 0.004646418965421617, 'timing/model': 1.207084321998991, 'learning_rate/qwen_vl_interface': 2.6647786367130513e-06, 'learning_rate/action_model': 2.664778636713051e-05, 'epoch': 1.42}) +[2026-06-17 15:12:41,476][starVLA.training.train_starvla][INFO] - Step 4815, Loss: {'train/loss': 0.6057873964309692, 'train/grad_norm_pre_clip': 14.191023826599121, 'timing/data': 0.004957222030498087, 'timing/model': 1.203652924974449, 'learning_rate/qwen_vl_interface': 2.6629645836786217e-06, 'learning_rate/action_model': 2.6629645836786217e-05, 'epoch': 1.42}) +[2026-06-17 15:12:46,224][starVLA.training.train_starvla][INFO] - Step 4816, Loss: {'train/loss': 0.6822503209114075, 'train/grad_norm_pre_clip': 12.955694198608398, 'timing/data': 0.004932075040414929, 'timing/model': 1.1998559230705723, 'learning_rate/qwen_vl_interface': 2.661151066939417e-06, 'learning_rate/action_model': 2.6611510669394168e-05, 'epoch': 1.42}) +[2026-06-17 15:12:50,993][starVLA.training.train_starvla][INFO] - Step 4817, Loss: {'train/loss': 0.4305776059627533, 'train/grad_norm_pre_clip': 10.38113784790039, 'timing/data': 0.005475268932059407, 'timing/model': 1.1995110280113295, 'learning_rate/qwen_vl_interface': 2.6593380868713843e-06, 'learning_rate/action_model': 2.659338086871384e-05, 'epoch': 1.42}) +[2026-06-17 15:12:55,784][starVLA.training.train_starvla][INFO] - Step 4818, Loss: {'train/loss': 0.5776606202125549, 'train/grad_norm_pre_clip': 12.972087860107422, 'timing/data': 0.005433524027466774, 'timing/model': 1.3403362920507789, 'learning_rate/qwen_vl_interface': 2.6575256438503507e-06, 'learning_rate/action_model': 2.6575256438503504e-05, 'epoch': 1.42}) +[2026-06-17 15:13:00,441][starVLA.training.train_starvla][INFO] - Step 4819, Loss: {'train/loss': 0.5383328795433044, 'train/grad_norm_pre_clip': 14.529233932495117, 'timing/data': 0.00560013996437192, 'timing/model': 1.2109120630193502, 'learning_rate/qwen_vl_interface': 2.6557137382520413e-06, 'learning_rate/action_model': 2.6557137382520412e-05, 'epoch': 1.42}) +[2026-06-17 15:13:05,237][starVLA.training.train_starvla][INFO] - Step 4820, Loss: {'train/loss': 0.8310191035270691, 'train/grad_norm_pre_clip': 14.751594543457031, 'timing/data': 0.004595098900608718, 'timing/model': 1.2130842710612342, 'learning_rate/qwen_vl_interface': 2.6539023704520666e-06, 'learning_rate/action_model': 2.6539023704520664e-05, 'epoch': 1.42}) +[2026-06-17 15:13:10,019][starVLA.training.train_starvla][INFO] - Step 4821, Loss: {'train/loss': 0.6483522057533264, 'train/grad_norm_pre_clip': 16.007688522338867, 'timing/data': 0.0049903299659490585, 'timing/model': 1.2024463010020554, 'learning_rate/qwen_vl_interface': 2.6520915408259234e-06, 'learning_rate/action_model': 2.6520915408259235e-05, 'epoch': 1.42}) +[2026-06-17 15:13:14,798][starVLA.training.train_starvla][INFO] - Step 4822, Loss: {'train/loss': 0.3084952235221863, 'train/grad_norm_pre_clip': 8.056368827819824, 'timing/data': 0.005058898939751089, 'timing/model': 1.2023596239741892, 'learning_rate/qwen_vl_interface': 2.650281249749e-06, 'learning_rate/action_model': 2.6502812497489998e-05, 'epoch': 1.42}) +[2026-06-17 15:13:19,488][starVLA.training.train_starvla][INFO] - Step 4823, Loss: {'train/loss': 0.5885964632034302, 'train/grad_norm_pre_clip': 12.870356559753418, 'timing/data': 0.005251134978607297, 'timing/model': 1.2019796359818429, 'learning_rate/qwen_vl_interface': 2.6484714975965704e-06, 'learning_rate/action_model': 2.6484714975965703e-05, 'epoch': 1.42}) +[2026-06-17 15:13:24,197][starVLA.training.train_starvla][INFO] - Step 4824, Loss: {'train/loss': 0.5441220998764038, 'train/grad_norm_pre_clip': 12.21792221069336, 'timing/data': 0.004977134987711906, 'timing/model': 1.2518521189922467, 'learning_rate/qwen_vl_interface': 2.646662284743799e-06, 'learning_rate/action_model': 2.646662284743799e-05, 'epoch': 1.42}) +[2026-06-17 15:13:28,860][starVLA.training.train_starvla][INFO] - Step 4825, Loss: {'train/loss': 0.889975368976593, 'train/grad_norm_pre_clip': 13.733232498168945, 'timing/data': 0.004929144983179867, 'timing/model': 1.2043780860258266, 'learning_rate/qwen_vl_interface': 2.6448536115657353e-06, 'learning_rate/action_model': 2.6448536115657353e-05, 'epoch': 1.42}) +[2026-06-17 15:13:33,609][starVLA.training.train_starvla][INFO] - Step 4826, Loss: {'train/loss': 0.4243229627609253, 'train/grad_norm_pre_clip': 13.307966232299805, 'timing/data': 0.005309222964569926, 'timing/model': 1.2103508239379153, 'learning_rate/qwen_vl_interface': 2.6430454784373248e-06, 'learning_rate/action_model': 2.643045478437325e-05, 'epoch': 1.42}) +[2026-06-17 15:13:38,398][starVLA.training.train_starvla][INFO] - Step 4827, Loss: {'train/loss': 0.47884824872016907, 'train/grad_norm_pre_clip': 10.63032054901123, 'timing/data': 0.004925934015773237, 'timing/model': 1.2025203360244632, 'learning_rate/qwen_vl_interface': 2.6412378857333903e-06, 'learning_rate/action_model': 2.6412378857333903e-05, 'epoch': 1.42}) +[2026-06-17 15:13:43,183][starVLA.training.train_starvla][INFO] - Step 4828, Loss: {'train/loss': 0.7108343243598938, 'train/grad_norm_pre_clip': 14.068726539611816, 'timing/data': 0.005293654045090079, 'timing/model': 1.2056720519904047, 'learning_rate/qwen_vl_interface': 2.639430833828649e-06, 'learning_rate/action_model': 2.6394308338286488e-05, 'epoch': 1.42}) +[2026-06-17 15:13:47,960][starVLA.training.train_starvla][INFO] - Step 4829, Loss: {'train/loss': 0.8422415852546692, 'train/grad_norm_pre_clip': 17.967674255371094, 'timing/data': 0.0057748600374907255, 'timing/model': 1.333812332013622, 'learning_rate/qwen_vl_interface': 2.637624323097703e-06, 'learning_rate/action_model': 2.6376243230977026e-05, 'epoch': 1.42}) +[2026-06-17 15:13:52,733][starVLA.training.train_starvla][INFO] - Step 4830, Loss: {'train/loss': 0.9348107576370239, 'train/grad_norm_pre_clip': 17.352853775024414, 'timing/data': 0.005525878048501909, 'timing/model': 1.32346518302802, 'learning_rate/qwen_vl_interface': 2.635818353915048e-06, 'learning_rate/action_model': 2.635818353915048e-05, 'epoch': 1.42}) +[2026-06-17 15:13:57,385][starVLA.training.train_starvla][INFO] - Step 4831, Loss: {'train/loss': 1.0427019596099854, 'train/grad_norm_pre_clip': 19.637004852294922, 'timing/data': 0.005184617009945214, 'timing/model': 1.201456727925688, 'learning_rate/qwen_vl_interface': 2.634012926655063e-06, 'learning_rate/action_model': 2.6340129266550628e-05, 'epoch': 1.42}) +[2026-06-17 15:14:02,147][starVLA.training.train_starvla][INFO] - Step 4832, Loss: {'train/loss': 0.6329777240753174, 'train/grad_norm_pre_clip': 13.287677764892578, 'timing/data': 0.005225604982115328, 'timing/model': 1.2109905648976564, 'learning_rate/qwen_vl_interface': 2.632208041692009e-06, 'learning_rate/action_model': 2.6322080416920088e-05, 'epoch': 1.42}) +[2026-06-17 15:14:06,941][starVLA.training.train_starvla][INFO] - Step 4833, Loss: {'train/loss': 0.7302172780036926, 'train/grad_norm_pre_clip': 18.01542854309082, 'timing/data': 0.004675069008953869, 'timing/model': 1.2060739279258996, 'learning_rate/qwen_vl_interface': 2.630403699400047e-06, 'learning_rate/action_model': 2.6304036994000468e-05, 'epoch': 1.42}) +[2026-06-17 15:14:11,744][starVLA.training.train_starvla][INFO] - Step 4834, Loss: {'train/loss': 0.4428136646747589, 'train/grad_norm_pre_clip': 12.872370719909668, 'timing/data': 0.00500312598887831, 'timing/model': 1.2078230129554868, 'learning_rate/qwen_vl_interface': 2.628599900153217e-06, 'learning_rate/action_model': 2.628599900153217e-05, 'epoch': 1.42}) +[2026-06-17 15:14:16,532][starVLA.training.train_starvla][INFO] - Step 4835, Loss: {'train/loss': 0.7237275242805481, 'train/grad_norm_pre_clip': 16.313276290893555, 'timing/data': 0.004990918911062181, 'timing/model': 1.3347978859674186, 'learning_rate/qwen_vl_interface': 2.6267966443254484e-06, 'learning_rate/action_model': 2.6267966443254483e-05, 'epoch': 1.42}) +[2026-06-17 15:14:21,310][starVLA.training.train_starvla][INFO] - Step 4836, Loss: {'train/loss': 0.8078470230102539, 'train/grad_norm_pre_clip': 17.526296615600586, 'timing/data': 0.005184639943763614, 'timing/model': 1.3306105970405042, 'learning_rate/qwen_vl_interface': 2.6249939322905576e-06, 'learning_rate/action_model': 2.6249939322905576e-05, 'epoch': 1.42}) +[2026-06-17 15:14:25,971][starVLA.training.train_starvla][INFO] - Step 4837, Loss: {'train/loss': 0.5487972497940063, 'train/grad_norm_pre_clip': 11.994685173034668, 'timing/data': 0.00469198205973953, 'timing/model': 1.2040388580644503, 'learning_rate/qwen_vl_interface': 2.6231917644222505e-06, 'learning_rate/action_model': 2.6231917644222504e-05, 'epoch': 1.42}) +[2026-06-17 15:14:30,735][starVLA.training.train_starvla][INFO] - Step 4838, Loss: {'train/loss': 0.525730311870575, 'train/grad_norm_pre_clip': 11.984658241271973, 'timing/data': 0.005155813996680081, 'timing/model': 1.2051355059957132, 'learning_rate/qwen_vl_interface': 2.621390141094116e-06, 'learning_rate/action_model': 2.621390141094116e-05, 'epoch': 1.42}) +[2026-06-17 15:14:35,518][starVLA.training.train_starvla][INFO] - Step 4839, Loss: {'train/loss': 0.5925424098968506, 'train/grad_norm_pre_clip': 12.612726211547852, 'timing/data': 0.004942507017403841, 'timing/model': 1.2051939059747383, 'learning_rate/qwen_vl_interface': 2.6195890626796345e-06, 'learning_rate/action_model': 2.6195890626796345e-05, 'epoch': 1.42}) +[2026-06-17 15:14:40,298][starVLA.training.train_starvla][INFO] - Step 4840, Loss: {'train/loss': 0.5936186909675598, 'train/grad_norm_pre_clip': 12.147300720214844, 'timing/data': 0.004957183031365275, 'timing/model': 1.2000814439961687, 'learning_rate/qwen_vl_interface': 2.6177885295521706e-06, 'learning_rate/action_model': 2.6177885295521702e-05, 'epoch': 1.42}) +[2026-06-17 15:14:45,025][starVLA.training.train_starvla][INFO] - Step 4841, Loss: {'train/loss': 0.7609156966209412, 'train/grad_norm_pre_clip': 15.835030555725098, 'timing/data': 0.00539722794201225, 'timing/model': 1.269072074093856, 'learning_rate/qwen_vl_interface': 2.6159885420849766e-06, 'learning_rate/action_model': 2.6159885420849766e-05, 'epoch': 1.42}) +[2026-06-17 15:14:49,669][starVLA.training.train_starvla][INFO] - Step 4842, Loss: {'train/loss': 0.6026871800422668, 'train/grad_norm_pre_clip': 12.499556541442871, 'timing/data': 0.004774188040755689, 'timing/model': 1.2027878819499165, 'learning_rate/qwen_vl_interface': 2.6141891006511927e-06, 'learning_rate/action_model': 2.6141891006511925e-05, 'epoch': 1.42}) +[2026-06-17 15:14:54,444][starVLA.training.train_starvla][INFO] - Step 4843, Loss: {'train/loss': 0.6568508744239807, 'train/grad_norm_pre_clip': 14.757532119750977, 'timing/data': 0.004902945016510785, 'timing/model': 1.2027909030439332, 'learning_rate/qwen_vl_interface': 2.612390205623842e-06, 'learning_rate/action_model': 2.612390205623842e-05, 'epoch': 1.43}) +[2026-06-17 15:14:59,300][starVLA.training.train_starvla][INFO] - Step 4844, Loss: {'train/loss': 0.6665363311767578, 'train/grad_norm_pre_clip': 14.979022026062012, 'timing/data': 0.00507367099635303, 'timing/model': 1.211021694005467, 'learning_rate/qwen_vl_interface': 2.6105918573758425e-06, 'learning_rate/action_model': 2.6105918573758427e-05, 'epoch': 1.43}) +[2026-06-17 15:15:04,101][starVLA.training.train_starvla][INFO] - Step 4845, Loss: {'train/loss': 0.8276214003562927, 'train/grad_norm_pre_clip': 16.082069396972656, 'timing/data': 0.0053175369976088405, 'timing/model': 1.210610853973776, 'learning_rate/qwen_vl_interface': 2.608794056279991e-06, 'learning_rate/action_model': 2.6087940562799906e-05, 'epoch': 1.43}) +[2026-06-17 15:15:08,885][starVLA.training.train_starvla][INFO] - Step 4846, Loss: {'train/loss': 0.6290844678878784, 'train/grad_norm_pre_clip': 15.668848991394043, 'timing/data': 0.004734813002869487, 'timing/model': 1.337337743025273, 'learning_rate/qwen_vl_interface': 2.6069968027089744e-06, 'learning_rate/action_model': 2.6069968027089742e-05, 'epoch': 1.43}) +[2026-06-17 15:15:13,615][starVLA.training.train_starvla][INFO] - Step 4847, Loss: {'train/loss': 0.5313183665275574, 'train/grad_norm_pre_clip': 14.835936546325684, 'timing/data': 0.005032368935644627, 'timing/model': 1.2824748329585418, 'learning_rate/qwen_vl_interface': 2.605200097035362e-06, 'learning_rate/action_model': 2.605200097035362e-05, 'epoch': 1.43}) +[2026-06-17 15:15:18,278][starVLA.training.train_starvla][INFO] - Step 4848, Loss: {'train/loss': 0.8617392778396606, 'train/grad_norm_pre_clip': 13.707514762878418, 'timing/data': 0.005198809085413814, 'timing/model': 1.2055255860323086, 'learning_rate/qwen_vl_interface': 2.6034039396316167e-06, 'learning_rate/action_model': 2.6034039396316167e-05, 'epoch': 1.43}) +[2026-06-17 15:15:23,064][starVLA.training.train_starvla][INFO] - Step 4849, Loss: {'train/loss': 0.5797064304351807, 'train/grad_norm_pre_clip': 12.488852500915527, 'timing/data': 0.005251561990007758, 'timing/model': 1.2073164749890566, 'learning_rate/qwen_vl_interface': 2.601608330870083e-06, 'learning_rate/action_model': 2.601608330870083e-05, 'epoch': 1.43}) +[2026-06-17 15:15:27,852][starVLA.training.train_starvla][INFO] - Step 4850, Loss: {'train/loss': 0.6295490264892578, 'train/grad_norm_pre_clip': 14.410077095031738, 'timing/data': 0.004948963993228972, 'timing/model': 1.2035007199738175, 'learning_rate/qwen_vl_interface': 2.5998132711229918e-06, 'learning_rate/action_model': 2.5998132711229915e-05, 'epoch': 1.43}) +[2026-06-17 15:15:32,634][starVLA.training.train_starvla][INFO] - Step 4851, Loss: {'train/loss': 0.6141090393066406, 'train/grad_norm_pre_clip': 13.989609718322754, 'timing/data': 0.004926678026095033, 'timing/model': 1.2013633380411193, 'learning_rate/qwen_vl_interface': 2.598018760762461e-06, 'learning_rate/action_model': 2.598018760762461e-05, 'epoch': 1.43}) +[2026-06-17 15:15:37,405][starVLA.training.train_starvla][INFO] - Step 4852, Loss: {'train/loss': 0.5002273917198181, 'train/grad_norm_pre_clip': 12.706462860107422, 'timing/data': 0.0050201809499412775, 'timing/model': 1.2018422100227326, 'learning_rate/qwen_vl_interface': 2.5962248001604958e-06, 'learning_rate/action_model': 2.5962248001604956e-05, 'epoch': 1.43}) +[2026-06-17 15:15:42,175][starVLA.training.train_starvla][INFO] - Step 4853, Loss: {'train/loss': 0.4425189793109894, 'train/grad_norm_pre_clip': 8.639557838439941, 'timing/data': 0.004931087023578584, 'timing/model': 1.291858106967993, 'learning_rate/qwen_vl_interface': 2.5944313896889837e-06, 'learning_rate/action_model': 2.594431389688984e-05, 'epoch': 1.43}) +[2026-06-17 15:15:46,842][starVLA.training.train_starvla][INFO] - Step 4854, Loss: {'train/loss': 0.5835275053977966, 'train/grad_norm_pre_clip': 12.56118106842041, 'timing/data': 0.00512138195335865, 'timing/model': 1.2023468789411709, 'learning_rate/qwen_vl_interface': 2.592638529719703e-06, 'learning_rate/action_model': 2.5926385297197033e-05, 'epoch': 1.43}) +[2026-06-17 15:15:51,638][starVLA.training.train_starvla][INFO] - Step 4855, Loss: {'train/loss': 0.7837277054786682, 'train/grad_norm_pre_clip': 17.492252349853516, 'timing/data': 0.005360221024602652, 'timing/model': 1.2047139599453658, 'learning_rate/qwen_vl_interface': 2.590846220624314e-06, 'learning_rate/action_model': 2.5908462206243138e-05, 'epoch': 1.43}) +[2026-06-17 15:15:56,427][starVLA.training.train_starvla][INFO] - Step 4856, Loss: {'train/loss': 0.9652220010757446, 'train/grad_norm_pre_clip': 23.116588592529297, 'timing/data': 0.005151326069608331, 'timing/model': 1.2028394009685144, 'learning_rate/qwen_vl_interface': 2.5890544627743652e-06, 'learning_rate/action_model': 2.589054462774365e-05, 'epoch': 1.43}) +[2026-06-17 15:16:01,218][starVLA.training.train_starvla][INFO] - Step 4857, Loss: {'train/loss': 0.6790269613265991, 'train/grad_norm_pre_clip': 12.589037895202637, 'timing/data': 0.005045884987339377, 'timing/model': 1.20398992090486, 'learning_rate/qwen_vl_interface': 2.58726325654129e-06, 'learning_rate/action_model': 2.5872632565412898e-05, 'epoch': 1.43}) +[2026-06-17 15:16:06,027][starVLA.training.train_starvla][INFO] - Step 4858, Loss: {'train/loss': 0.45121678709983826, 'train/grad_norm_pre_clip': 10.509775161743164, 'timing/data': 0.005474198958836496, 'timing/model': 1.334908029064536, 'learning_rate/qwen_vl_interface': 2.5854726022964043e-06, 'learning_rate/action_model': 2.5854726022964042e-05, 'epoch': 1.43}) +[2026-06-17 15:16:10,698][starVLA.training.train_starvla][INFO] - Step 4859, Loss: {'train/loss': 0.5249641537666321, 'train/grad_norm_pre_clip': 12.992022514343262, 'timing/data': 0.004932564916089177, 'timing/model': 1.2151501249754801, 'learning_rate/qwen_vl_interface': 2.5836825004109173e-06, 'learning_rate/action_model': 2.5836825004109172e-05, 'epoch': 1.43}) +[2026-06-17 15:16:15,477][starVLA.training.train_starvla][INFO] - Step 4860, Loss: {'train/loss': 0.5391533374786377, 'train/grad_norm_pre_clip': 15.242966651916504, 'timing/data': 0.005075050052255392, 'timing/model': 1.2036785639356822, 'learning_rate/qwen_vl_interface': 2.5818929512559193e-06, 'learning_rate/action_model': 2.5818929512559193e-05, 'epoch': 1.43}) +[2026-06-17 15:16:20,258][starVLA.training.train_starvla][INFO] - Step 4861, Loss: {'train/loss': 0.5583277940750122, 'train/grad_norm_pre_clip': 13.618252754211426, 'timing/data': 0.005072637111879885, 'timing/model': 1.2035985309630632, 'learning_rate/qwen_vl_interface': 2.580103955202381e-06, 'learning_rate/action_model': 2.580103955202381e-05, 'epoch': 1.43}) +[2026-06-17 15:16:24,966][starVLA.training.train_starvla][INFO] - Step 4862, Loss: {'train/loss': 0.6514188051223755, 'train/grad_norm_pre_clip': 15.881028175354004, 'timing/data': 0.004917977028526366, 'timing/model': 1.2051280400482938, 'learning_rate/qwen_vl_interface': 2.5783155126211647e-06, 'learning_rate/action_model': 2.5783155126211644e-05, 'epoch': 1.43}) +[2026-06-17 15:16:29,719][starVLA.training.train_starvla][INFO] - Step 4863, Loss: {'train/loss': 0.6653121709823608, 'train/grad_norm_pre_clip': 13.201855659484863, 'timing/data': 0.005064974073320627, 'timing/model': 1.2079135229578242, 'learning_rate/qwen_vl_interface': 2.576527623883019e-06, 'learning_rate/action_model': 2.576527623883019e-05, 'epoch': 1.43}) +[2026-06-17 15:16:34,502][starVLA.training.train_starvla][INFO] - Step 4864, Loss: {'train/loss': 0.39921334385871887, 'train/grad_norm_pre_clip': 10.41978645324707, 'timing/data': 0.005182752967812121, 'timing/model': 1.334855419001542, 'learning_rate/qwen_vl_interface': 2.5747402893585732e-06, 'learning_rate/action_model': 2.574740289358573e-05, 'epoch': 1.43}) +[2026-06-17 15:16:39,234][starVLA.training.train_starvla][INFO] - Step 4865, Loss: {'train/loss': 0.9362213015556335, 'train/grad_norm_pre_clip': 19.20686912536621, 'timing/data': 0.005159209948033094, 'timing/model': 1.214051024056971, 'learning_rate/qwen_vl_interface': 2.5729535094183445e-06, 'learning_rate/action_model': 2.5729535094183444e-05, 'epoch': 1.43}) +[2026-06-17 15:16:44,039][starVLA.training.train_starvla][INFO] - Step 4866, Loss: {'train/loss': 0.5243756175041199, 'train/grad_norm_pre_clip': 16.609655380249023, 'timing/data': 0.005121988011524081, 'timing/model': 1.2174599450081587, 'learning_rate/qwen_vl_interface': 2.571167284432734e-06, 'learning_rate/action_model': 2.5711672844327338e-05, 'epoch': 1.43}) +[2026-06-17 15:16:48,883][starVLA.training.train_starvla][INFO] - Step 4867, Loss: {'train/loss': 0.7025876045227051, 'train/grad_norm_pre_clip': 13.827606201171875, 'timing/data': 0.005214795004576445, 'timing/model': 1.209586673998274, 'learning_rate/qwen_vl_interface': 2.569381614772028e-06, 'learning_rate/action_model': 2.5693816147720278e-05, 'epoch': 1.43}) +[2026-06-17 15:16:53,678][starVLA.training.train_starvla][INFO] - Step 4868, Loss: {'train/loss': 0.749666690826416, 'train/grad_norm_pre_clip': 17.861780166625977, 'timing/data': 0.005104340962134302, 'timing/model': 1.210909848101437, 'learning_rate/qwen_vl_interface': 2.5675965008063976e-06, 'learning_rate/action_model': 2.5675965008063974e-05, 'epoch': 1.43}) +[2026-06-17 15:16:58,435][starVLA.training.train_starvla][INFO] - Step 4869, Loss: {'train/loss': 0.7424819469451904, 'train/grad_norm_pre_clip': 16.947689056396484, 'timing/data': 0.005237895995378494, 'timing/model': 1.2041396689601243, 'learning_rate/qwen_vl_interface': 2.565811942905899e-06, 'learning_rate/action_model': 2.5658119429058993e-05, 'epoch': 1.43}) +[2026-06-17 15:17:03,237][starVLA.training.train_starvla][INFO] - Step 4870, Loss: {'train/loss': 0.6424471139907837, 'train/grad_norm_pre_clip': 13.045135498046875, 'timing/data': 0.005029777064919472, 'timing/model': 1.3385456500109285, 'learning_rate/qwen_vl_interface': 2.5640279414404723e-06, 'learning_rate/action_model': 2.5640279414404722e-05, 'epoch': 1.43}) +[2026-06-17 15:17:07,919][starVLA.training.train_starvla][INFO] - Step 4871, Loss: {'train/loss': 0.8186020255088806, 'train/grad_norm_pre_clip': 15.070775032043457, 'timing/data': 0.005114067113026977, 'timing/model': 1.2033246679930016, 'learning_rate/qwen_vl_interface': 2.5622444967799444e-06, 'learning_rate/action_model': 2.5622444967799443e-05, 'epoch': 1.43}) +[2026-06-17 15:17:12,734][starVLA.training.train_starvla][INFO] - Step 4872, Loss: {'train/loss': 0.5422031283378601, 'train/grad_norm_pre_clip': 13.393421173095703, 'timing/data': 0.005089312908239663, 'timing/model': 1.2054908850695938, 'learning_rate/qwen_vl_interface': 2.5604616092940245e-06, 'learning_rate/action_model': 2.5604616092940242e-05, 'epoch': 1.43}) +[2026-06-17 15:17:17,579][starVLA.training.train_starvla][INFO] - Step 4873, Loss: {'train/loss': 0.7252163887023926, 'train/grad_norm_pre_clip': 19.102235794067383, 'timing/data': 0.004824542091228068, 'timing/model': 1.2072135749040172, 'learning_rate/qwen_vl_interface': 2.5586792793523052e-06, 'learning_rate/action_model': 2.5586792793523052e-05, 'epoch': 1.43}) +[2026-06-17 15:17:22,364][starVLA.training.train_starvla][INFO] - Step 4874, Loss: {'train/loss': 0.7465680241584778, 'train/grad_norm_pre_clip': 12.282047271728516, 'timing/data': 0.005162643035873771, 'timing/model': 1.2037652949802577, 'learning_rate/qwen_vl_interface': 2.556897507324271e-06, 'learning_rate/action_model': 2.5568975073242705e-05, 'epoch': 1.43}) +[2026-06-17 15:17:27,153][starVLA.training.train_starvla][INFO] - Step 4875, Loss: {'train/loss': 0.9134170413017273, 'train/grad_norm_pre_clip': 18.681407928466797, 'timing/data': 0.005009012063965201, 'timing/model': 1.3363327549304813, 'learning_rate/qwen_vl_interface': 2.5551162935792782e-06, 'learning_rate/action_model': 2.5551162935792784e-05, 'epoch': 1.43}) +[2026-06-17 15:17:31,815][starVLA.training.train_starvla][INFO] - Step 4876, Loss: {'train/loss': 0.8244032263755798, 'train/grad_norm_pre_clip': 16.79224967956543, 'timing/data': 0.004991048015654087, 'timing/model': 1.2032553559402004, 'learning_rate/qwen_vl_interface': 2.5533356384865777e-06, 'learning_rate/action_model': 2.5533356384865776e-05, 'epoch': 1.43}) +[2026-06-17 15:17:36,591][starVLA.training.train_starvla][INFO] - Step 4877, Loss: {'train/loss': 0.782353937625885, 'train/grad_norm_pre_clip': 17.094104766845703, 'timing/data': 0.005137274973094463, 'timing/model': 1.2057266159681603, 'learning_rate/qwen_vl_interface': 2.5515555424153017e-06, 'learning_rate/action_model': 2.5515555424153016e-05, 'epoch': 1.44}) +[2026-06-17 15:17:41,294][starVLA.training.train_starvla][INFO] - Step 4878, Loss: {'train/loss': 0.6839277744293213, 'train/grad_norm_pre_clip': 16.78670883178711, 'timing/data': 0.004747573984786868, 'timing/model': 1.2046850239858031, 'learning_rate/qwen_vl_interface': 2.5497760057344663e-06, 'learning_rate/action_model': 2.549776005734466e-05, 'epoch': 1.44}) +[2026-06-17 15:17:46,036][starVLA.training.train_starvla][INFO] - Step 4879, Loss: {'train/loss': 0.42182686924934387, 'train/grad_norm_pre_clip': 11.041509628295898, 'timing/data': 0.005277662072330713, 'timing/model': 1.2060002359794453, 'learning_rate/qwen_vl_interface': 2.5479970288129714e-06, 'learning_rate/action_model': 2.5479970288129714e-05, 'epoch': 1.44}) +[2026-06-17 15:17:50,825][starVLA.training.train_starvla][INFO] - Step 4880, Loss: {'train/loss': 0.5475525856018066, 'train/grad_norm_pre_clip': 12.645615577697754, 'timing/data': 0.004510978003963828, 'timing/model': 1.2106632989598438, 'learning_rate/qwen_vl_interface': 2.546218612019597e-06, 'learning_rate/action_model': 2.5462186120195967e-05, 'epoch': 1.44}) +[2026-06-17 15:17:55,627][starVLA.training.train_starvla][INFO] - Step 4881, Loss: {'train/loss': 0.5591405034065247, 'train/grad_norm_pre_clip': 12.402400016784668, 'timing/data': 0.004928839043714106, 'timing/model': 1.3412774429889396, 'learning_rate/qwen_vl_interface': 2.544440755723015e-06, 'learning_rate/action_model': 2.5444407557230147e-05, 'epoch': 1.44}) +[2026-06-17 15:18:00,280][starVLA.training.train_starvla][INFO] - Step 4882, Loss: {'train/loss': 0.47127318382263184, 'train/grad_norm_pre_clip': 11.475900650024414, 'timing/data': 0.005166493938304484, 'timing/model': 1.2061660940526053, 'learning_rate/qwen_vl_interface': 2.5426634602917756e-06, 'learning_rate/action_model': 2.5426634602917755e-05, 'epoch': 1.44}) +[2026-06-17 15:18:05,072][starVLA.training.train_starvla][INFO] - Step 4883, Loss: {'train/loss': 0.8223090171813965, 'train/grad_norm_pre_clip': 13.373597145080566, 'timing/data': 0.005211598007008433, 'timing/model': 1.204543280065991, 'learning_rate/qwen_vl_interface': 2.540886726094315e-06, 'learning_rate/action_model': 2.540886726094315e-05, 'epoch': 1.44}) +[2026-06-17 15:18:09,850][starVLA.training.train_starvla][INFO] - Step 4884, Loss: {'train/loss': 0.7222052812576294, 'train/grad_norm_pre_clip': 11.956364631652832, 'timing/data': 0.005026262952014804, 'timing/model': 1.206449190969579, 'learning_rate/qwen_vl_interface': 2.5391105534989503e-06, 'learning_rate/action_model': 2.53911055349895e-05, 'epoch': 1.44}) +[2026-06-17 15:18:14,603][starVLA.training.train_starvla][INFO] - Step 4885, Loss: {'train/loss': 0.7365120649337769, 'train/grad_norm_pre_clip': 11.945785522460938, 'timing/data': 0.004993589012883604, 'timing/model': 1.2051505860872567, 'learning_rate/qwen_vl_interface': 2.537334942873885e-06, 'learning_rate/action_model': 2.5373349428738846e-05, 'epoch': 1.44}) +[2026-06-17 15:18:19,385][starVLA.training.train_starvla][INFO] - Step 4886, Loss: {'train/loss': 0.46497395634651184, 'train/grad_norm_pre_clip': 10.742318153381348, 'timing/data': 0.005350847030058503, 'timing/model': 1.2053314340300858, 'learning_rate/qwen_vl_interface': 2.535559894587205e-06, 'learning_rate/action_model': 2.535559894587205e-05, 'epoch': 1.44}) +[2026-06-17 15:18:24,180][starVLA.training.train_starvla][INFO] - Step 4887, Loss: {'train/loss': 0.5349838733673096, 'train/grad_norm_pre_clip': 13.816059112548828, 'timing/data': 0.004905864014290273, 'timing/model': 1.3386537289479747, 'learning_rate/qwen_vl_interface': 2.533785409006878e-06, 'learning_rate/action_model': 2.533785409006878e-05, 'epoch': 1.44}) +[2026-06-17 15:18:28,843][starVLA.training.train_starvla][INFO] - Step 4888, Loss: {'train/loss': 0.6215267777442932, 'train/grad_norm_pre_clip': 15.539050102233887, 'timing/data': 0.004623487940989435, 'timing/model': 1.2086393459467217, 'learning_rate/qwen_vl_interface': 2.532011486500762e-06, 'learning_rate/action_model': 2.532011486500762e-05, 'epoch': 1.44}) +[2026-06-17 15:18:33,625][starVLA.training.train_starvla][INFO] - Step 4889, Loss: {'train/loss': 0.5650923252105713, 'train/grad_norm_pre_clip': 11.764395713806152, 'timing/data': 0.004502386902458966, 'timing/model': 1.2026428469689563, 'learning_rate/qwen_vl_interface': 2.530238127436588e-06, 'learning_rate/action_model': 2.5302381274365882e-05, 'epoch': 1.44}) +[2026-06-17 15:18:38,408][starVLA.training.train_starvla][INFO] - Step 4890, Loss: {'train/loss': 0.9552575945854187, 'train/grad_norm_pre_clip': 14.917706489562988, 'timing/data': 0.00471092201769352, 'timing/model': 1.2080108280060813, 'learning_rate/qwen_vl_interface': 2.528465332181977e-06, 'learning_rate/action_model': 2.5284653321819767e-05, 'epoch': 1.44}) +[2026-06-17 15:18:43,202][starVLA.training.train_starvla][INFO] - Step 4891, Loss: {'train/loss': 0.7826204299926758, 'train/grad_norm_pre_clip': 15.015607833862305, 'timing/data': 0.00515914405696094, 'timing/model': 1.2061145920306444, 'learning_rate/qwen_vl_interface': 2.526693101104429e-06, 'learning_rate/action_model': 2.526693101104429e-05, 'epoch': 1.44}) +[2026-06-17 15:18:47,882][starVLA.training.train_starvla][INFO] - Step 4892, Loss: {'train/loss': 0.4603171646595001, 'train/grad_norm_pre_clip': 11.726395606994629, 'timing/data': 0.0048965830355882645, 'timing/model': 1.2055483319563791, 'learning_rate/qwen_vl_interface': 2.524921434571333e-06, 'learning_rate/action_model': 2.524921434571333e-05, 'epoch': 1.44}) +[2026-06-17 15:18:52,582][starVLA.training.train_starvla][INFO] - Step 4893, Loss: {'train/loss': 0.6993671655654907, 'train/grad_norm_pre_clip': 15.530427932739258, 'timing/data': 0.005010220920667052, 'timing/model': 1.2430672439513728, 'learning_rate/qwen_vl_interface': 2.5231503329499585e-06, 'learning_rate/action_model': 2.5231503329499583e-05, 'epoch': 1.44}) +[2026-06-17 15:18:57,237][starVLA.training.train_starvla][INFO] - Step 4894, Loss: {'train/loss': 0.7524001598358154, 'train/grad_norm_pre_clip': 13.82280445098877, 'timing/data': 0.0046084190253168344, 'timing/model': 1.2070713270222768, 'learning_rate/qwen_vl_interface': 2.5213797966074506e-06, 'learning_rate/action_model': 2.5213797966074505e-05, 'epoch': 1.44}) +[2026-06-17 15:19:01,963][starVLA.training.train_starvla][INFO] - Step 4895, Loss: {'train/loss': 0.5497248768806458, 'train/grad_norm_pre_clip': 18.62571907043457, 'timing/data': 0.0048523800214752555, 'timing/model': 1.2020757870050147, 'learning_rate/qwen_vl_interface': 2.519609825910848e-06, 'learning_rate/action_model': 2.5196098259108482e-05, 'epoch': 1.44}) +[2026-06-17 15:19:06,746][starVLA.training.train_starvla][INFO] - Step 4896, Loss: {'train/loss': 0.5596070885658264, 'train/grad_norm_pre_clip': 11.763469696044922, 'timing/data': 0.005340800969861448, 'timing/model': 1.206860058940947, 'learning_rate/qwen_vl_interface': 2.517840421227067e-06, 'learning_rate/action_model': 2.517840421227067e-05, 'epoch': 1.44}) +[2026-06-17 15:19:11,531][starVLA.training.train_starvla][INFO] - Step 4897, Loss: {'train/loss': 0.6547710299491882, 'train/grad_norm_pre_clip': 13.314913749694824, 'timing/data': 0.005247108987532556, 'timing/model': 1.204467204050161, 'learning_rate/qwen_vl_interface': 2.5160715829229067e-06, 'learning_rate/action_model': 2.5160715829229066e-05, 'epoch': 1.44}) +[2026-06-17 15:19:16,324][starVLA.training.train_starvla][INFO] - Step 4898, Loss: {'train/loss': 0.8323527574539185, 'train/grad_norm_pre_clip': 16.437135696411133, 'timing/data': 0.005132758989930153, 'timing/model': 1.3348128000507131, 'learning_rate/qwen_vl_interface': 2.5143033113650484e-06, 'learning_rate/action_model': 2.5143033113650483e-05, 'epoch': 1.44}) +[2026-06-17 15:19:21,040][starVLA.training.train_starvla][INFO] - Step 4899, Loss: {'train/loss': 0.6010718941688538, 'train/grad_norm_pre_clip': 14.705077171325684, 'timing/data': 0.004966989043168724, 'timing/model': 1.25678054895252, 'learning_rate/qwen_vl_interface': 2.5125356069200572e-06, 'learning_rate/action_model': 2.5125356069200574e-05, 'epoch': 1.44}) +[2026-06-17 15:21:03,167][starVLA.training.train_starvla][INFO] - Step 4900, Loss: {'train/loss': 0.746935248374939, 'train/grad_norm_pre_clip': 14.525320053100586, 'eval/loss': 0.719814121723175, 'eval/latency_0/loss': 0.719814121723175, 'eval/demon_fire_precision': 0.8981371383273881, 'eval/demon_fire_recall': 0.9410299003322259, 'eval/demon_fire_f1': 0.9190833502332184, 'eval/demon_fire_rate_model': 0.7016129032258065, 'eval/demon_fire_rate_teacher': 0.6696329254727474, 'eval/demon_left_precision': 0.8419489007724302, 'eval/demon_left_recall': 0.8028328611898017, 'eval/demon_left_f1': 0.8219257540603249, 'eval/demon_left_rate_model': 0.46802002224694106, 'eval/demon_left_rate_teacher': 0.49082313681868744, 'eval/demon_right_precision': 0.8126326963906582, 'eval/demon_right_recall': 0.8649717514124293, 'eval/demon_right_f1': 0.8379857690202517, 'eval/demon_right_rate_model': 0.5239154616240267, 'eval/demon_right_rate_teacher': 0.4922135706340378, 'eval/demon_move_f1': 0.8299557615402883, 'eval/demon_cc_f1': 0.8903978408798994, 'eval/latency_0/demon_fire_precision': 0.8981371383273881, 'eval/latency_0/demon_fire_recall': 0.9410299003322259, 'eval/latency_0/demon_fire_f1': 0.9190833502332184, 'eval/latency_0/demon_fire_rate_model': 0.7016129032258065, 'eval/latency_0/demon_fire_rate_teacher': 0.6696329254727474, 'eval/latency_0/demon_left_precision': 0.8419489007724302, 'eval/latency_0/demon_left_recall': 0.8028328611898017, 'eval/latency_0/demon_left_f1': 0.8219257540603249, 'eval/latency_0/demon_left_rate_model': 0.46802002224694106, 'eval/latency_0/demon_left_rate_teacher': 0.49082313681868744, 'eval/latency_0/demon_right_precision': 0.8126326963906582, 'eval/latency_0/demon_right_recall': 0.8649717514124293, 'eval/latency_0/demon_right_f1': 0.8379857690202517, 'eval/latency_0/demon_right_rate_model': 0.5239154616240267, 'eval/latency_0/demon_right_rate_teacher': 0.4922135706340378, 'eval/latency_0/demon_move_f1': 0.8299557615402883, 'eval/latency_0/demon_cc_f1': 0.8903978408798994, 'timing/data': 0.0051067579770460725, 'timing/model': 1.2043467070907354, 'learning_rate/qwen_vl_interface': 2.5107684699543817e-06, 'learning_rate/action_model': 2.5107684699543816e-05, 'epoch': 1.44}) +[2026-06-17 15:21:03,167][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_4900_state +[2026-06-17 15:21:03,178][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 15:21:06,743][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_4900_state/model.safetensors +[2026-06-17 15:21:14,399][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_4900_state/optimizer.bin +[2026-06-17 15:21:14,399][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_4900_state/sampler.bin +[2026-06-17 15:21:14,399][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_4900_state/sampler_1.bin +[2026-06-17 15:21:14,400][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_4900_state/random_states_0.pkl +[2026-06-17 15:21:14,454][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4900 +[2026-06-17 15:21:14,454][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 15:21:14,499][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 15:21:19,123][starVLA.training.train_starvla][INFO] - Step 4901, Loss: {'train/loss': 0.6713629364967346, 'train/grad_norm_pre_clip': 16.17719268798828, 'timing/data': 0.005023250938393176, 'timing/model': 1.1884428359335288, 'learning_rate/qwen_vl_interface': 2.509001900834349e-06, 'learning_rate/action_model': 2.5090019008343492e-05, 'epoch': 1.44}) +[2026-06-17 15:21:23,864][starVLA.training.train_starvla][INFO] - Step 4902, Loss: {'train/loss': 0.4948331117630005, 'train/grad_norm_pre_clip': 12.862836837768555, 'timing/data': 0.005425188923254609, 'timing/model': 1.191729151061736, 'learning_rate/qwen_vl_interface': 2.507235899926172e-06, 'learning_rate/action_model': 2.5072358999261718e-05, 'epoch': 1.44}) +[2026-06-17 15:21:28,485][starVLA.training.train_starvla][INFO] - Step 4903, Loss: {'train/loss': 0.7023593187332153, 'train/grad_norm_pre_clip': 14.14988899230957, 'timing/data': 0.005370670929551125, 'timing/model': 1.1947622440056875, 'learning_rate/qwen_vl_interface': 2.5054704675959445e-06, 'learning_rate/action_model': 2.5054704675959445e-05, 'epoch': 1.44}) +[2026-06-17 15:21:33,123][starVLA.training.train_starvla][INFO] - Step 4904, Loss: {'train/loss': 0.5173340439796448, 'train/grad_norm_pre_clip': 10.348174095153809, 'timing/data': 0.005625026999041438, 'timing/model': 1.1946985520189628, 'learning_rate/qwen_vl_interface': 2.5037056042096418e-06, 'learning_rate/action_model': 2.5037056042096417e-05, 'epoch': 1.44}) +[2026-06-17 15:21:37,784][starVLA.training.train_starvla][INFO] - Step 4905, Loss: {'train/loss': 0.5381491184234619, 'train/grad_norm_pre_clip': 12.104521751403809, 'timing/data': 0.005122511996887624, 'timing/model': 1.196551262983121, 'learning_rate/qwen_vl_interface': 2.5019413101331212e-06, 'learning_rate/action_model': 2.5019413101331214e-05, 'epoch': 1.44}) +[2026-06-17 15:21:42,528][starVLA.training.train_starvla][INFO] - Step 4906, Loss: {'train/loss': 0.9254550337791443, 'train/grad_norm_pre_clip': 14.1767578125, 'timing/data': 0.005153508973307908, 'timing/model': 1.3149109010118991, 'learning_rate/qwen_vl_interface': 2.5001775857321254e-06, 'learning_rate/action_model': 2.5001775857321253e-05, 'epoch': 1.44}) +[2026-06-17 15:21:47,155][starVLA.training.train_starvla][INFO] - Step 4907, Loss: {'train/loss': 0.7360289692878723, 'train/grad_norm_pre_clip': 13.555737495422363, 'timing/data': 0.0052169120172038674, 'timing/model': 1.19818592001684, 'learning_rate/qwen_vl_interface': 2.4984144313722765e-06, 'learning_rate/action_model': 2.4984144313722766e-05, 'epoch': 1.44}) +[2026-06-17 15:21:51,894][starVLA.training.train_starvla][INFO] - Step 4908, Loss: {'train/loss': 0.7193456888198853, 'train/grad_norm_pre_clip': 17.039390563964844, 'timing/data': 0.004947084002196789, 'timing/model': 1.2023643079446629, 'learning_rate/qwen_vl_interface': 2.496651847419076e-06, 'learning_rate/action_model': 2.496651847419076e-05, 'epoch': 1.44}) +[2026-06-17 15:21:56,666][starVLA.training.train_starvla][INFO] - Step 4909, Loss: {'train/loss': 0.7240163087844849, 'train/grad_norm_pre_clip': 15.466927528381348, 'timing/data': 0.006207746104337275, 'timing/model': 1.198882367927581, 'learning_rate/qwen_vl_interface': 2.4948898342379076e-06, 'learning_rate/action_model': 2.4948898342379075e-05, 'epoch': 1.44}) +[2026-06-17 15:22:01,436][starVLA.training.train_starvla][INFO] - Step 4910, Loss: {'train/loss': 0.7490178942680359, 'train/grad_norm_pre_clip': 12.445831298828125, 'timing/data': 0.004796205088496208, 'timing/model': 1.2064669369719923, 'learning_rate/qwen_vl_interface': 2.4931283921940434e-06, 'learning_rate/action_model': 2.4931283921940434e-05, 'epoch': 1.44}) +[2026-06-17 15:22:06,207][starVLA.training.train_starvla][INFO] - Step 4911, Loss: {'train/loss': 0.6790914535522461, 'train/grad_norm_pre_clip': 13.246419906616211, 'timing/data': 0.0054850520100444555, 'timing/model': 1.3358220169320703, 'learning_rate/qwen_vl_interface': 2.4913675216526302e-06, 'learning_rate/action_model': 2.4913675216526304e-05, 'epoch': 1.45}) +[2026-06-17 15:22:10,936][starVLA.training.train_starvla][INFO] - Step 4912, Loss: {'train/loss': 0.6672162413597107, 'train/grad_norm_pre_clip': 11.556891441345215, 'timing/data': 0.004895539954304695, 'timing/model': 1.286400720011443, 'learning_rate/qwen_vl_interface': 2.4896072229786996e-06, 'learning_rate/action_model': 2.4896072229786995e-05, 'epoch': 1.45}) +[2026-06-17 15:22:15,596][starVLA.training.train_starvla][INFO] - Step 4913, Loss: {'train/loss': 0.6910874843597412, 'train/grad_norm_pre_clip': 14.469328880310059, 'timing/data': 0.005401883041486144, 'timing/model': 1.207990227965638, 'learning_rate/qwen_vl_interface': 2.4878474965371587e-06, 'learning_rate/action_model': 2.4878474965371586e-05, 'epoch': 1.45}) +[2026-06-17 15:22:20,378][starVLA.training.train_starvla][INFO] - Step 4914, Loss: {'train/loss': 0.7162386178970337, 'train/grad_norm_pre_clip': 13.14812183380127, 'timing/data': 0.0051869050366804, 'timing/model': 1.2012625539209694, 'learning_rate/qwen_vl_interface': 2.4860883426928053e-06, 'learning_rate/action_model': 2.4860883426928053e-05, 'epoch': 1.45}) +[2026-06-17 15:22:25,145][starVLA.training.train_starvla][INFO] - Step 4915, Loss: {'train/loss': 0.6895750761032104, 'train/grad_norm_pre_clip': 13.180410385131836, 'timing/data': 0.005177577957510948, 'timing/model': 1.2031512600369751, 'learning_rate/qwen_vl_interface': 2.4843297618103127e-06, 'learning_rate/action_model': 2.4843297618103128e-05, 'epoch': 1.45}) +[2026-06-17 15:22:29,913][starVLA.training.train_starvla][INFO] - Step 4916, Loss: {'train/loss': 0.7238498330116272, 'train/grad_norm_pre_clip': 14.834268569946289, 'timing/data': 0.005389217985793948, 'timing/model': 1.200369870988652, 'learning_rate/qwen_vl_interface': 2.482571754254237e-06, 'learning_rate/action_model': 2.4825717542542372e-05, 'epoch': 1.45}) +[2026-06-17 15:22:34,746][starVLA.training.train_starvla][INFO] - Step 4917, Loss: {'train/loss': 0.5187918543815613, 'train/grad_norm_pre_clip': 10.094644546508789, 'timing/data': 0.004964923951774836, 'timing/model': 1.2028183520305902, 'learning_rate/qwen_vl_interface': 2.4808143203890137e-06, 'learning_rate/action_model': 2.480814320389014e-05, 'epoch': 1.45}) +[2026-06-17 15:22:39,496][starVLA.training.train_starvla][INFO] - Step 4918, Loss: {'train/loss': 0.5121076107025146, 'train/grad_norm_pre_clip': 13.594826698303223, 'timing/data': 0.005377736059017479, 'timing/model': 1.3097328019794077, 'learning_rate/qwen_vl_interface': 2.4790574605789627e-06, 'learning_rate/action_model': 2.4790574605789624e-05, 'epoch': 1.45}) +[2026-06-17 15:22:44,157][starVLA.training.train_starvla][INFO] - Step 4919, Loss: {'train/loss': 0.8148337006568909, 'train/grad_norm_pre_clip': 15.462013244628906, 'timing/data': 0.005308552994392812, 'timing/model': 1.2073082299903035, 'learning_rate/qwen_vl_interface': 2.4773011751882807e-06, 'learning_rate/action_model': 2.4773011751882804e-05, 'epoch': 1.45}) +[2026-06-17 15:22:48,953][starVLA.training.train_starvla][INFO] - Step 4920, Loss: {'train/loss': 0.8167484998703003, 'train/grad_norm_pre_clip': 15.919732093811035, 'timing/data': 0.004867931944318116, 'timing/model': 1.2024092390201986, 'learning_rate/qwen_vl_interface': 2.4755454645810473e-06, 'learning_rate/action_model': 2.4755454645810473e-05, 'epoch': 1.45}) +[2026-06-17 15:22:53,756][starVLA.training.train_starvla][INFO] - Step 4921, Loss: {'train/loss': 0.6367424130439758, 'train/grad_norm_pre_clip': 12.79248046875, 'timing/data': 0.005028402898460627, 'timing/model': 1.208542385022156, 'learning_rate/qwen_vl_interface': 2.473790329121228e-06, 'learning_rate/action_model': 2.473790329121228e-05, 'epoch': 1.45}) +[2026-06-17 15:22:58,532][starVLA.training.train_starvla][INFO] - Step 4922, Loss: {'train/loss': 0.9126533269882202, 'train/grad_norm_pre_clip': 12.826111793518066, 'timing/data': 0.0051590410294011235, 'timing/model': 1.1996443499810994, 'learning_rate/qwen_vl_interface': 2.4720357691726595e-06, 'learning_rate/action_model': 2.4720357691726594e-05, 'epoch': 1.45}) +[2026-06-17 15:23:03,309][starVLA.training.train_starvla][INFO] - Step 4923, Loss: {'train/loss': 0.5682623386383057, 'train/grad_norm_pre_clip': 12.42883014678955, 'timing/data': 0.005018059979192913, 'timing/model': 1.3329274340067059, 'learning_rate/qwen_vl_interface': 2.4702817850990645e-06, 'learning_rate/action_model': 2.470281785099064e-05, 'epoch': 1.45}) +[2026-06-17 15:23:07,955][starVLA.training.train_starvla][INFO] - Step 4924, Loss: {'train/loss': 0.476725310087204, 'train/grad_norm_pre_clip': 10.862818717956543, 'timing/data': 0.005285453982651234, 'timing/model': 1.2012467950116843, 'learning_rate/qwen_vl_interface': 2.4685283772640452e-06, 'learning_rate/action_model': 2.468528377264045e-05, 'epoch': 1.45}) +[2026-06-17 15:23:12,733][starVLA.training.train_starvla][INFO] - Step 4925, Loss: {'train/loss': 0.5341401100158691, 'train/grad_norm_pre_clip': 14.908944129943848, 'timing/data': 0.0047303789760917425, 'timing/model': 1.202940539107658, 'learning_rate/qwen_vl_interface': 2.466775546031088e-06, 'learning_rate/action_model': 2.4667755460310878e-05, 'epoch': 1.45}) +[2026-06-17 15:23:17,508][starVLA.training.train_starvla][INFO] - Step 4926, Loss: {'train/loss': 0.6623360514640808, 'train/grad_norm_pre_clip': 14.72652530670166, 'timing/data': 0.0051636050920933485, 'timing/model': 1.2035191310569644, 'learning_rate/qwen_vl_interface': 2.465023291763555e-06, 'learning_rate/action_model': 2.465023291763555e-05, 'epoch': 1.45}) +[2026-06-17 15:23:22,207][starVLA.training.train_starvla][INFO] - Step 4927, Loss: {'train/loss': 0.6647605895996094, 'train/grad_norm_pre_clip': 12.969191551208496, 'timing/data': 0.005063677090220153, 'timing/model': 1.2030863410327584, 'learning_rate/qwen_vl_interface': 2.4632716148246903e-06, 'learning_rate/action_model': 2.4632716148246903e-05, 'epoch': 1.45}) +[2026-06-17 15:23:26,928][starVLA.training.train_starvla][INFO] - Step 4928, Loss: {'train/loss': 0.7333664298057556, 'train/grad_norm_pre_clip': 16.687911987304688, 'timing/data': 0.005366681958548725, 'timing/model': 1.2021835889900103, 'learning_rate/qwen_vl_interface': 2.4615205155776185e-06, 'learning_rate/action_model': 2.4615205155776188e-05, 'epoch': 1.45}) +[2026-06-17 15:23:31,724][starVLA.training.train_starvla][INFO] - Step 4929, Loss: {'train/loss': 0.7725347280502319, 'train/grad_norm_pre_clip': 15.445287704467773, 'timing/data': 0.0052678699139505625, 'timing/model': 1.3350312780821696, 'learning_rate/qwen_vl_interface': 2.4597699943853444e-06, 'learning_rate/action_model': 2.4597699943853442e-05, 'epoch': 1.45}) +[2026-06-17 15:23:36,378][starVLA.training.train_starvla][INFO] - Step 4930, Loss: {'train/loss': 0.2640850841999054, 'train/grad_norm_pre_clip': 8.076789855957031, 'timing/data': 0.005221780971623957, 'timing/model': 1.2034448459744453, 'learning_rate/qwen_vl_interface': 2.458020051610753e-06, 'learning_rate/action_model': 2.4580200516107528e-05, 'epoch': 1.45}) +[2026-06-17 15:23:41,174][starVLA.training.train_starvla][INFO] - Step 4931, Loss: {'train/loss': 0.7760714292526245, 'train/grad_norm_pre_clip': 13.311137199401855, 'timing/data': 0.005463708890601993, 'timing/model': 1.2013742829440162, 'learning_rate/qwen_vl_interface': 2.456270687616609e-06, 'learning_rate/action_model': 2.4562706876166087e-05, 'epoch': 1.45}) +[2026-06-17 15:23:45,970][starVLA.training.train_starvla][INFO] - Step 4932, Loss: {'train/loss': 0.796700656414032, 'train/grad_norm_pre_clip': 27.040180206298828, 'timing/data': 0.005341156967915595, 'timing/model': 1.2139358969870955, 'learning_rate/qwen_vl_interface': 2.4545219027655577e-06, 'learning_rate/action_model': 2.4545219027655574e-05, 'epoch': 1.45}) +[2026-06-17 15:23:50,764][starVLA.training.train_starvla][INFO] - Step 4933, Loss: {'train/loss': 0.7166914939880371, 'train/grad_norm_pre_clip': 13.062657356262207, 'timing/data': 0.004729353007860482, 'timing/model': 1.2054603809956461, 'learning_rate/qwen_vl_interface': 2.452773697420124e-06, 'learning_rate/action_model': 2.452773697420124e-05, 'epoch': 1.45}) +[2026-06-17 15:23:55,525][starVLA.training.train_starvla][INFO] - Step 4934, Loss: {'train/loss': 0.6942078471183777, 'train/grad_norm_pre_clip': 13.944125175476074, 'timing/data': 0.004922927008010447, 'timing/model': 1.2118805130012333, 'learning_rate/qwen_vl_interface': 2.4510260719427127e-06, 'learning_rate/action_model': 2.4510260719427127e-05, 'epoch': 1.45}) +[2026-06-17 15:24:00,314][starVLA.training.train_starvla][INFO] - Step 4935, Loss: {'train/loss': 0.9205031394958496, 'train/grad_norm_pre_clip': 16.515945434570312, 'timing/data': 0.00520847795996815, 'timing/model': 1.3260465849889442, 'learning_rate/qwen_vl_interface': 2.4492790266956092e-06, 'learning_rate/action_model': 2.449279026695609e-05, 'epoch': 1.45}) +[2026-06-17 15:24:04,999][starVLA.training.train_starvla][INFO] - Step 4936, Loss: {'train/loss': 0.6478742361068726, 'train/grad_norm_pre_clip': 15.37509536743164, 'timing/data': 0.00543279096018523, 'timing/model': 1.2083854470402002, 'learning_rate/qwen_vl_interface': 2.4475325620409765e-06, 'learning_rate/action_model': 2.4475325620409764e-05, 'epoch': 1.45}) +[2026-06-17 15:24:09,787][starVLA.training.train_starvla][INFO] - Step 4937, Loss: {'train/loss': 0.5866875052452087, 'train/grad_norm_pre_clip': 22.674165725708008, 'timing/data': 0.0055326740257442, 'timing/model': 1.203622052911669, 'learning_rate/qwen_vl_interface': 2.445786678340859e-06, 'learning_rate/action_model': 2.4457866783408588e-05, 'epoch': 1.45}) +[2026-06-17 15:24:14,571][starVLA.training.train_starvla][INFO] - Step 4938, Loss: {'train/loss': 0.4240522086620331, 'train/grad_norm_pre_clip': 10.087857246398926, 'timing/data': 0.005111483042128384, 'timing/model': 1.2087473660940304, 'learning_rate/qwen_vl_interface': 2.444041375957179e-06, 'learning_rate/action_model': 2.444041375957179e-05, 'epoch': 1.45}) +[2026-06-17 15:24:19,398][starVLA.training.train_starvla][INFO] - Step 4939, Loss: {'train/loss': 0.6351150274276733, 'train/grad_norm_pre_clip': 13.306877136230469, 'timing/data': 0.005106388940475881, 'timing/model': 1.2220061710104346, 'learning_rate/qwen_vl_interface': 2.4422966552517434e-06, 'learning_rate/action_model': 2.4422966552517433e-05, 'epoch': 1.45}) +[2026-06-17 15:24:24,316][starVLA.training.train_starvla][INFO] - Step 4940, Loss: {'train/loss': 0.6593574285507202, 'train/grad_norm_pre_clip': 15.752979278564453, 'timing/data': 0.00534731603693217, 'timing/model': 1.3335136149544269, 'learning_rate/qwen_vl_interface': 2.440552516586232e-06, 'learning_rate/action_model': 2.440552516586232e-05, 'epoch': 1.45}) +[2026-06-17 15:24:28,972][starVLA.training.train_starvla][INFO] - Step 4941, Loss: {'train/loss': 0.6841979026794434, 'train/grad_norm_pre_clip': 12.236576080322266, 'timing/data': 0.0050301989540457726, 'timing/model': 1.208373864996247, 'learning_rate/qwen_vl_interface': 2.4388089603222085e-06, 'learning_rate/action_model': 2.4388089603222087e-05, 'epoch': 1.45}) +[2026-06-17 15:24:33,783][starVLA.training.train_starvla][INFO] - Step 4942, Loss: {'train/loss': 0.7701671719551086, 'train/grad_norm_pre_clip': 13.02754020690918, 'timing/data': 0.005263879080303013, 'timing/model': 1.2109064169926569, 'learning_rate/qwen_vl_interface': 2.4370659868211095e-06, 'learning_rate/action_model': 2.4370659868211097e-05, 'epoch': 1.45}) +[2026-06-17 15:24:38,507][starVLA.training.train_starvla][INFO] - Step 4943, Loss: {'train/loss': 0.8647183179855347, 'train/grad_norm_pre_clip': 15.687675476074219, 'timing/data': 0.005162760033272207, 'timing/model': 1.2110765340039507, 'learning_rate/qwen_vl_interface': 2.4353235964442597e-06, 'learning_rate/action_model': 2.4353235964442598e-05, 'epoch': 1.45}) +[2026-06-17 15:24:43,249][starVLA.training.train_starvla][INFO] - Step 4944, Loss: {'train/loss': 0.45658281445503235, 'train/grad_norm_pre_clip': 11.47195816040039, 'timing/data': 0.005277617019601166, 'timing/model': 1.2032145949779078, 'learning_rate/qwen_vl_interface': 2.433581789552857e-06, 'learning_rate/action_model': 2.433581789552857e-05, 'epoch': 1.45}) +[2026-06-17 15:24:48,041][starVLA.training.train_starvla][INFO] - Step 4945, Loss: {'train/loss': 0.5617293119430542, 'train/grad_norm_pre_clip': 13.024445533752441, 'timing/data': 0.005365037941373885, 'timing/model': 1.2043183069908991, 'learning_rate/qwen_vl_interface': 2.4318405665079806e-06, 'learning_rate/action_model': 2.4318405665079805e-05, 'epoch': 1.46}) +[2026-06-17 15:24:52,825][starVLA.training.train_starvla][INFO] - Step 4946, Loss: {'train/loss': 0.9424434900283813, 'train/grad_norm_pre_clip': 17.22142219543457, 'timing/data': 0.005188022973015904, 'timing/model': 1.331554296077229, 'learning_rate/qwen_vl_interface': 2.4300999276705868e-06, 'learning_rate/action_model': 2.4300999276705868e-05, 'epoch': 1.46}) +[2026-06-17 15:24:57,477][starVLA.training.train_starvla][INFO] - Step 4947, Loss: {'train/loss': 0.9268252849578857, 'train/grad_norm_pre_clip': 15.298773765563965, 'timing/data': 0.004992853035219014, 'timing/model': 1.2045429770369083, 'learning_rate/qwen_vl_interface': 2.428359873401512e-06, 'learning_rate/action_model': 2.4283598734015124e-05, 'epoch': 1.46}) +[2026-06-17 15:25:02,251][starVLA.training.train_starvla][INFO] - Step 4948, Loss: {'train/loss': 0.8427602648735046, 'train/grad_norm_pre_clip': 16.224618911743164, 'timing/data': 0.005240702070295811, 'timing/model': 1.2031974049750715, 'learning_rate/qwen_vl_interface': 2.4266204040614722e-06, 'learning_rate/action_model': 2.4266204040614724e-05, 'epoch': 1.46}) +[2026-06-17 15:25:07,028][starVLA.training.train_starvla][INFO] - Step 4949, Loss: {'train/loss': 0.7886048555374146, 'train/grad_norm_pre_clip': 15.608675956726074, 'timing/data': 0.005710978992283344, 'timing/model': 1.2051842330256477, 'learning_rate/qwen_vl_interface': 2.424881520011059e-06, 'learning_rate/action_model': 2.424881520011059e-05, 'epoch': 1.46}) +[2026-06-17 15:25:11,785][starVLA.training.train_starvla][INFO] - Step 4950, Loss: {'train/loss': 0.6778437495231628, 'train/grad_norm_pre_clip': 18.107431411743164, 'timing/data': 0.0047993509797379375, 'timing/model': 1.2044700009282678, 'learning_rate/qwen_vl_interface': 2.4231432216107494e-06, 'learning_rate/action_model': 2.4231432216107494e-05, 'epoch': 1.46}) +[2026-06-17 15:25:16,578][starVLA.training.train_starvla][INFO] - Step 4951, Loss: {'train/loss': 0.7002203464508057, 'train/grad_norm_pre_clip': 13.279646873474121, 'timing/data': 0.005155503051355481, 'timing/model': 1.2069443720392883, 'learning_rate/qwen_vl_interface': 2.4214055092208907e-06, 'learning_rate/action_model': 2.4214055092208907e-05, 'epoch': 1.46}) +[2026-06-17 15:25:21,389][starVLA.training.train_starvla][INFO] - Step 4952, Loss: {'train/loss': 0.6384385824203491, 'train/grad_norm_pre_clip': 15.251975059509277, 'timing/data': 0.004938571015372872, 'timing/model': 1.3501257839379832, 'learning_rate/qwen_vl_interface': 2.4196683832017134e-06, 'learning_rate/action_model': 2.4196683832017135e-05, 'epoch': 1.46}) +[2026-06-17 15:25:26,053][starVLA.training.train_starvla][INFO] - Step 4953, Loss: {'train/loss': 0.6559734344482422, 'train/grad_norm_pre_clip': 14.2767915725708, 'timing/data': 0.005341875948943198, 'timing/model': 1.202585924998857, 'learning_rate/qwen_vl_interface': 2.4179318439133245e-06, 'learning_rate/action_model': 2.417931843913324e-05, 'epoch': 1.46}) +[2026-06-17 15:25:30,839][starVLA.training.train_starvla][INFO] - Step 4954, Loss: {'train/loss': 0.41966676712036133, 'train/grad_norm_pre_clip': 38.68836212158203, 'timing/data': 0.0050662909634411335, 'timing/model': 1.2049820190295577, 'learning_rate/qwen_vl_interface': 2.4161958917157126e-06, 'learning_rate/action_model': 2.4161958917157126e-05, 'epoch': 1.46}) +[2026-06-17 15:25:35,633][starVLA.training.train_starvla][INFO] - Step 4955, Loss: {'train/loss': 0.6431180834770203, 'train/grad_norm_pre_clip': 12.122904777526855, 'timing/data': 0.005096422042697668, 'timing/model': 1.2055909970076755, 'learning_rate/qwen_vl_interface': 2.4144605269687445e-06, 'learning_rate/action_model': 2.4144605269687446e-05, 'epoch': 1.46}) +[2026-06-17 15:25:40,416][starVLA.training.train_starvla][INFO] - Step 4956, Loss: {'train/loss': 0.6373448371887207, 'train/grad_norm_pre_clip': 13.76677131652832, 'timing/data': 0.0051956139504909515, 'timing/model': 1.2055934750242159, 'learning_rate/qwen_vl_interface': 2.412725750032156e-06, 'learning_rate/action_model': 2.4127257500321563e-05, 'epoch': 1.46}) +[2026-06-17 15:25:45,082][starVLA.training.train_starvla][INFO] - Step 4957, Loss: {'train/loss': 0.4323407709598541, 'train/grad_norm_pre_clip': 9.790596008300781, 'timing/data': 0.005097285029478371, 'timing/model': 1.2048213860252872, 'learning_rate/qwen_vl_interface': 2.4109915612655745e-06, 'learning_rate/action_model': 2.4109915612655745e-05, 'epoch': 1.46}) +[2026-06-17 15:25:49,785][starVLA.training.train_starvla][INFO] - Step 4958, Loss: {'train/loss': 0.4484511911869049, 'train/grad_norm_pre_clip': 10.560254096984863, 'timing/data': 0.00513126200530678, 'timing/model': 1.2482646339340135, 'learning_rate/qwen_vl_interface': 2.409257961028498e-06, 'learning_rate/action_model': 2.4092579610284978e-05, 'epoch': 1.46}) +[2026-06-17 15:25:54,447][starVLA.training.train_starvla][INFO] - Step 4959, Loss: {'train/loss': 0.7689275741577148, 'train/grad_norm_pre_clip': 14.087684631347656, 'timing/data': 0.0050675999373197556, 'timing/model': 1.2104005600558594, 'learning_rate/qwen_vl_interface': 2.4075249496803026e-06, 'learning_rate/action_model': 2.4075249496803025e-05, 'epoch': 1.46}) +[2026-06-17 15:25:59,203][starVLA.training.train_starvla][INFO] - Step 4960, Loss: {'train/loss': 0.6127594709396362, 'train/grad_norm_pre_clip': 12.656192779541016, 'timing/data': 0.005039779935032129, 'timing/model': 1.2077758320374414, 'learning_rate/qwen_vl_interface': 2.405792527580243e-06, 'learning_rate/action_model': 2.405792527580243e-05, 'epoch': 1.46}) +[2026-06-17 15:26:03,993][starVLA.training.train_starvla][INFO] - Step 4961, Loss: {'train/loss': 0.5711074471473694, 'train/grad_norm_pre_clip': 13.176668167114258, 'timing/data': 0.0052156789461150765, 'timing/model': 1.206487369025126, 'learning_rate/qwen_vl_interface': 2.404060695087454e-06, 'learning_rate/action_model': 2.4040606950874536e-05, 'epoch': 1.46}) +[2026-06-17 15:26:08,813][starVLA.training.train_starvla][INFO] - Step 4962, Loss: {'train/loss': 0.6491658687591553, 'train/grad_norm_pre_clip': 12.106729507446289, 'timing/data': 0.007330736028961837, 'timing/model': 1.2178611860144883, 'learning_rate/qwen_vl_interface': 2.4023294525609437e-06, 'learning_rate/action_model': 2.4023294525609437e-05, 'epoch': 1.46}) +[2026-06-17 15:26:13,602][starVLA.training.train_starvla][INFO] - Step 4963, Loss: {'train/loss': 0.502662718296051, 'train/grad_norm_pre_clip': 13.120199203491211, 'timing/data': 0.005377343040890992, 'timing/model': 1.3322727310005575, 'learning_rate/qwen_vl_interface': 2.400598800359604e-06, 'learning_rate/action_model': 2.4005988003596035e-05, 'epoch': 1.46}) +[2026-06-17 15:26:18,321][starVLA.training.train_starvla][INFO] - Step 4964, Loss: {'train/loss': 0.8267161846160889, 'train/grad_norm_pre_clip': 15.568370819091797, 'timing/data': 0.004911858006380498, 'timing/model': 1.2618393010925502, 'learning_rate/qwen_vl_interface': 2.398868738842198e-06, 'learning_rate/action_model': 2.3988687388421978e-05, 'epoch': 1.46}) +[2026-06-17 15:26:22,991][starVLA.training.train_starvla][INFO] - Step 4965, Loss: {'train/loss': 0.8650268912315369, 'train/grad_norm_pre_clip': 16.705860137939453, 'timing/data': 0.005631072097457945, 'timing/model': 1.2084059229819104, 'learning_rate/qwen_vl_interface': 2.3971392683673705e-06, 'learning_rate/action_model': 2.3971392683673706e-05, 'epoch': 1.46}) +[2026-06-17 15:26:27,744][starVLA.training.train_starvla][INFO] - Step 4966, Loss: {'train/loss': 0.6960839629173279, 'train/grad_norm_pre_clip': 13.535321235656738, 'timing/data': 0.005484941997565329, 'timing/model': 1.2053781520808116, 'learning_rate/qwen_vl_interface': 2.3954103892936425e-06, 'learning_rate/action_model': 2.3954103892936424e-05, 'epoch': 1.46}) +[2026-06-17 15:26:32,531][starVLA.training.train_starvla][INFO] - Step 4967, Loss: {'train/loss': 0.6373018026351929, 'train/grad_norm_pre_clip': 14.782646179199219, 'timing/data': 0.005791609990410507, 'timing/model': 1.2029893370345235, 'learning_rate/qwen_vl_interface': 2.393682101979411e-06, 'learning_rate/action_model': 2.3936821019794105e-05, 'epoch': 1.46}) +[2026-06-17 15:26:37,319][starVLA.training.train_starvla][INFO] - Step 4968, Loss: {'train/loss': 0.5571011304855347, 'train/grad_norm_pre_clip': 12.925146102905273, 'timing/data': 0.005149322096258402, 'timing/model': 1.2061321820365265, 'learning_rate/qwen_vl_interface': 2.3919544067829546e-06, 'learning_rate/action_model': 2.3919544067829545e-05, 'epoch': 1.46}) +[2026-06-17 15:26:42,145][starVLA.training.train_starvla][INFO] - Step 4969, Loss: {'train/loss': 0.5703316926956177, 'train/grad_norm_pre_clip': 11.037694931030273, 'timing/data': 0.004950560978613794, 'timing/model': 1.3390154869994149, 'learning_rate/qwen_vl_interface': 2.3902273040624267e-06, 'learning_rate/action_model': 2.390227304062427e-05, 'epoch': 1.46}) +[2026-06-17 15:26:46,807][starVLA.training.train_starvla][INFO] - Step 4970, Loss: {'train/loss': 0.42000317573547363, 'train/grad_norm_pre_clip': 13.990978240966797, 'timing/data': 0.005309909000061452, 'timing/model': 1.2062891400419176, 'learning_rate/qwen_vl_interface': 2.388500794175855e-06, 'learning_rate/action_model': 2.3885007941758548e-05, 'epoch': 1.46}) +[2026-06-17 15:26:51,590][starVLA.training.train_starvla][INFO] - Step 4971, Loss: {'train/loss': 1.0637184381484985, 'train/grad_norm_pre_clip': 33.946407318115234, 'timing/data': 0.004870631964877248, 'timing/model': 1.2066579860402271, 'learning_rate/qwen_vl_interface': 2.386774877481145e-06, 'learning_rate/action_model': 2.3867748774811452e-05, 'epoch': 1.46}) +[2026-06-17 15:26:56,388][starVLA.training.train_starvla][INFO] - Step 4972, Loss: {'train/loss': 0.5166624784469604, 'train/grad_norm_pre_clip': 11.849590301513672, 'timing/data': 0.005036488990299404, 'timing/model': 1.2057725540362298, 'learning_rate/qwen_vl_interface': 2.385049554336087e-06, 'learning_rate/action_model': 2.385049554336087e-05, 'epoch': 1.46}) +[2026-06-17 15:27:01,177][starVLA.training.train_starvla][INFO] - Step 4973, Loss: {'train/loss': 0.5972698330879211, 'train/grad_norm_pre_clip': 12.071320533752441, 'timing/data': 0.005157514940947294, 'timing/model': 1.2040660489583388, 'learning_rate/qwen_vl_interface': 2.3833248250983383e-06, 'learning_rate/action_model': 2.3833248250983383e-05, 'epoch': 1.46}) +[2026-06-17 15:27:05,971][starVLA.training.train_starvla][INFO] - Step 4974, Loss: {'train/loss': 0.8124944567680359, 'train/grad_norm_pre_clip': 16.898683547973633, 'timing/data': 0.00496145396027714, 'timing/model': 1.335462421993725, 'learning_rate/qwen_vl_interface': 2.3816006901254396e-06, 'learning_rate/action_model': 2.3816006901254396e-05, 'epoch': 1.46}) +[2026-06-17 15:27:10,645][starVLA.training.train_starvla][INFO] - Step 4975, Loss: {'train/loss': 0.36588966846466064, 'train/grad_norm_pre_clip': 8.11794376373291, 'timing/data': 0.005424401024356484, 'timing/model': 1.2041635329369456, 'learning_rate/qwen_vl_interface': 2.3798771497748012e-06, 'learning_rate/action_model': 2.379877149774801e-05, 'epoch': 1.46}) +[2026-06-17 15:27:15,445][starVLA.training.train_starvla][INFO] - Step 4976, Loss: {'train/loss': 0.6268734335899353, 'train/grad_norm_pre_clip': 37.08470916748047, 'timing/data': 0.004999189986847341, 'timing/model': 1.21099680790212, 'learning_rate/qwen_vl_interface': 2.3781542044037195e-06, 'learning_rate/action_model': 2.3781542044037193e-05, 'epoch': 1.46}) +[2026-06-17 15:27:20,213][starVLA.training.train_starvla][INFO] - Step 4977, Loss: {'train/loss': 0.545006275177002, 'train/grad_norm_pre_clip': 15.510695457458496, 'timing/data': 0.004890476004220545, 'timing/model': 1.2073448710143566, 'learning_rate/qwen_vl_interface': 2.376431854369362e-06, 'learning_rate/action_model': 2.3764318543693618e-05, 'epoch': 1.46}) +[2026-06-17 15:27:25,002][starVLA.training.train_starvla][INFO] - Step 4978, Loss: {'train/loss': 0.6145904064178467, 'train/grad_norm_pre_clip': 13.96972370147705, 'timing/data': 0.005120057030580938, 'timing/model': 1.2045623429585248, 'learning_rate/qwen_vl_interface': 2.374710100028773e-06, 'learning_rate/action_model': 2.3747101000287728e-05, 'epoch': 1.46}) +[2026-06-17 15:27:29,803][starVLA.training.train_starvla][INFO] - Step 4979, Loss: {'train/loss': 0.5491155982017517, 'train/grad_norm_pre_clip': 11.406848907470703, 'timing/data': 0.005256125004962087, 'timing/model': 1.2110549409408122, 'learning_rate/qwen_vl_interface': 2.3729889417388736e-06, 'learning_rate/action_model': 2.3729889417388737e-05, 'epoch': 1.47}) +[2026-06-17 15:27:34,590][starVLA.training.train_starvla][INFO] - Step 4980, Loss: {'train/loss': 0.7047634720802307, 'train/grad_norm_pre_clip': 16.304462432861328, 'timing/data': 0.005190755007788539, 'timing/model': 1.3297049599932507, 'learning_rate/qwen_vl_interface': 2.3712683798564628e-06, 'learning_rate/action_model': 2.3712683798564625e-05, 'epoch': 1.47}) +[2026-06-17 15:27:39,388][starVLA.training.train_starvla][INFO] - Step 4981, Loss: {'train/loss': 0.8573474287986755, 'train/grad_norm_pre_clip': 46.047672271728516, 'timing/data': 0.00484962691552937, 'timing/model': 1.2816533889854327, 'learning_rate/qwen_vl_interface': 2.3695484147382146e-06, 'learning_rate/action_model': 2.3695484147382147e-05, 'epoch': 1.47}) +[2026-06-17 15:27:44,050][starVLA.training.train_starvla][INFO] - Step 4982, Loss: {'train/loss': 0.671261191368103, 'train/grad_norm_pre_clip': 13.738917350769043, 'timing/data': 0.005118475994095206, 'timing/model': 1.2052034969674423, 'learning_rate/qwen_vl_interface': 2.367829046740678e-06, 'learning_rate/action_model': 2.3678290467406776e-05, 'epoch': 1.47}) +[2026-06-17 15:27:48,839][starVLA.training.train_starvla][INFO] - Step 4983, Loss: {'train/loss': 0.5983185768127441, 'train/grad_norm_pre_clip': 13.647820472717285, 'timing/data': 0.005518427002243698, 'timing/model': 1.2058439740212634, 'learning_rate/qwen_vl_interface': 2.3661102762202846e-06, 'learning_rate/action_model': 2.3661102762202843e-05, 'epoch': 1.47}) +[2026-06-17 15:27:53,648][starVLA.training.train_starvla][INFO] - Step 4984, Loss: {'train/loss': 0.5642208456993103, 'train/grad_norm_pre_clip': 13.288970947265625, 'timing/data': 0.005273590097203851, 'timing/model': 1.2057029319694266, 'learning_rate/qwen_vl_interface': 2.364392103533332e-06, 'learning_rate/action_model': 2.364392103533332e-05, 'epoch': 1.47}) +[2026-06-17 15:27:58,515][starVLA.training.train_starvla][INFO] - Step 4985, Loss: {'train/loss': 0.43591904640197754, 'train/grad_norm_pre_clip': 11.132648468017578, 'timing/data': 0.005155087914317846, 'timing/model': 1.204886505031027, 'learning_rate/qwen_vl_interface': 2.362674529036002e-06, 'learning_rate/action_model': 2.362674529036002e-05, 'epoch': 1.47}) +[2026-06-17 15:28:03,305][starVLA.training.train_starvla][INFO] - Step 4986, Loss: {'train/loss': 0.7102681398391724, 'train/grad_norm_pre_clip': 15.630516052246094, 'timing/data': 0.005181254004128277, 'timing/model': 1.2081552989548072, 'learning_rate/qwen_vl_interface': 2.360957553084347e-06, 'learning_rate/action_model': 2.360957553084347e-05, 'epoch': 1.47}) +[2026-06-17 15:28:08,095][starVLA.training.train_starvla][INFO] - Step 4987, Loss: {'train/loss': 0.738853931427002, 'train/grad_norm_pre_clip': 14.122642517089844, 'timing/data': 0.005584152997471392, 'timing/model': 1.2927767280489206, 'learning_rate/qwen_vl_interface': 2.359241176034302e-06, 'learning_rate/action_model': 2.359241176034302e-05, 'epoch': 1.47}) +[2026-06-17 15:28:12,753][starVLA.training.train_starvla][INFO] - Step 4988, Loss: {'train/loss': 0.8059659004211426, 'train/grad_norm_pre_clip': 18.68671226501465, 'timing/data': 0.005487215006724, 'timing/model': 1.203869724064134, 'learning_rate/qwen_vl_interface': 2.3575253982416732e-06, 'learning_rate/action_model': 2.357525398241673e-05, 'epoch': 1.47}) +[2026-06-17 15:28:17,537][starVLA.training.train_starvla][INFO] - Step 4989, Loss: {'train/loss': 0.8473207354545593, 'train/grad_norm_pre_clip': 14.085332870483398, 'timing/data': 0.005419930908828974, 'timing/model': 1.2051316819852218, 'learning_rate/qwen_vl_interface': 2.3558102200621385e-06, 'learning_rate/action_model': 2.3558102200621386e-05, 'epoch': 1.47}) +[2026-06-17 15:28:22,341][starVLA.training.train_starvla][INFO] - Step 4990, Loss: {'train/loss': 0.5156201124191284, 'train/grad_norm_pre_clip': 11.078008651733398, 'timing/data': 0.005190843949094415, 'timing/model': 1.205614056903869, 'learning_rate/qwen_vl_interface': 2.3540956418512607e-06, 'learning_rate/action_model': 2.354095641851261e-05, 'epoch': 1.47}) +[2026-06-17 15:28:27,135][starVLA.training.train_starvla][INFO] - Step 4991, Loss: {'train/loss': 0.6588788032531738, 'train/grad_norm_pre_clip': 15.277400970458984, 'timing/data': 0.00534898997284472, 'timing/model': 1.2070574309909716, 'learning_rate/qwen_vl_interface': 2.352381663964473e-06, 'learning_rate/action_model': 2.352381663964473e-05, 'epoch': 1.47}) +[2026-06-17 15:28:31,931][starVLA.training.train_starvla][INFO] - Step 4992, Loss: {'train/loss': 0.5818047523498535, 'train/grad_norm_pre_clip': 17.049457550048828, 'timing/data': 0.005269166082143784, 'timing/model': 1.3369352569570765, 'learning_rate/qwen_vl_interface': 2.3506682867570846e-06, 'learning_rate/action_model': 2.3506682867570844e-05, 'epoch': 1.47}) +[2026-06-17 15:28:36,589][starVLA.training.train_starvla][INFO] - Step 4993, Loss: {'train/loss': 0.6567949056625366, 'train/grad_norm_pre_clip': 14.568461418151855, 'timing/data': 0.005080233095213771, 'timing/model': 1.2071331199258566, 'learning_rate/qwen_vl_interface': 2.3489555105842796e-06, 'learning_rate/action_model': 2.3489555105842794e-05, 'epoch': 1.47}) +[2026-06-17 15:28:41,371][starVLA.training.train_starvla][INFO] - Step 4994, Loss: {'train/loss': 0.6120331287384033, 'train/grad_norm_pre_clip': 12.069934844970703, 'timing/data': 0.005280715995468199, 'timing/model': 1.205726915039122, 'learning_rate/qwen_vl_interface': 2.347243335801119e-06, 'learning_rate/action_model': 2.347243335801119e-05, 'epoch': 1.47}) +[2026-06-17 15:28:46,163][starVLA.training.train_starvla][INFO] - Step 4995, Loss: {'train/loss': 0.4274851083755493, 'train/grad_norm_pre_clip': 12.534875869750977, 'timing/data': 0.005667384015396237, 'timing/model': 1.2112316789571196, 'learning_rate/qwen_vl_interface': 2.3455317627625385e-06, 'learning_rate/action_model': 2.3455317627625385e-05, 'epoch': 1.47}) +[2026-06-17 15:28:50,871][starVLA.training.train_starvla][INFO] - Step 4996, Loss: {'train/loss': 0.5403525233268738, 'train/grad_norm_pre_clip': 15.230842590332031, 'timing/data': 0.005248474073596299, 'timing/model': 1.2060013900045305, 'learning_rate/qwen_vl_interface': 2.3438207918233495e-06, 'learning_rate/action_model': 2.3438207918233494e-05, 'epoch': 1.47}) +[2026-06-17 15:28:55,608][starVLA.training.train_starvla][INFO] - Step 4997, Loss: {'train/loss': 0.8607171773910522, 'train/grad_norm_pre_clip': 15.378241539001465, 'timing/data': 0.005063502001576126, 'timing/model': 1.20427609491162, 'learning_rate/qwen_vl_interface': 2.3421104233382366e-06, 'learning_rate/action_model': 2.3421104233382364e-05, 'epoch': 1.47}) +[2026-06-17 15:29:00,397][starVLA.training.train_starvla][INFO] - Step 4998, Loss: {'train/loss': 0.5839641690254211, 'train/grad_norm_pre_clip': 12.080046653747559, 'timing/data': 0.00508658797480166, 'timing/model': 1.3381540949922055, 'learning_rate/qwen_vl_interface': 2.340400657661762e-06, 'learning_rate/action_model': 2.340400657661762e-05, 'epoch': 1.47}) +[2026-06-17 15:29:05,054][starVLA.training.train_starvla][INFO] - Step 4999, Loss: {'train/loss': 0.7639926671981812, 'train/grad_norm_pre_clip': 14.352243423461914, 'timing/data': 0.005254309973679483, 'timing/model': 1.2041480999905616, 'learning_rate/qwen_vl_interface': 2.338691495148363e-06, 'learning_rate/action_model': 2.3386914951483626e-05, 'epoch': 1.47}) +[2026-06-17 15:39:13,958][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 5000 +[2026-06-17 15:39:13,958][starVLA.training.train_starvla][INFO] - Step 5000, Loss: {'train/loss': 0.793441653251648, 'train/grad_norm_pre_clip': 13.933080673217773, 'eval/loss': 0.7177402377128601, 'eval/latency_0/loss': 0.7177402377128601, 'eval/demon_fire_precision': 0.9047234557932984, 'eval/demon_fire_recall': 0.9306478405315615, 'eval/demon_fire_f1': 0.9175025588536336, 'eval/demon_fire_rate_model': 0.6888209121245829, 'eval/demon_fire_rate_teacher': 0.6696329254727474, 'eval/demon_left_precision': 0.8307426597582038, 'eval/demon_left_recall': 0.8175637393767705, 'eval/demon_left_f1': 0.8241005139920047, 'eval/demon_left_rate_model': 0.48303670745272526, 'eval/demon_left_rate_teacher': 0.49082313681868744, 'eval/demon_right_precision': 0.8211738891936369, 'eval/demon_right_recall': 0.8457627118644068, 'eval/demon_right_f1': 0.83328694684108, 'eval/demon_right_rate_model': 0.506952169076752, 'eval/demon_right_rate_teacher': 0.4922135706340378, 'eval/demon_move_f1': 0.8286937304165423, 'eval/demon_cc_f1': 0.8889235122894948, 'eval/latency_0/demon_fire_precision': 0.9047234557932984, 'eval/latency_0/demon_fire_recall': 0.9306478405315615, 'eval/latency_0/demon_fire_f1': 0.9175025588536336, 'eval/latency_0/demon_fire_rate_model': 0.6888209121245829, 'eval/latency_0/demon_fire_rate_teacher': 0.6696329254727474, 'eval/latency_0/demon_left_precision': 0.8307426597582038, 'eval/latency_0/demon_left_recall': 0.8175637393767705, 'eval/latency_0/demon_left_f1': 0.8241005139920047, 'eval/latency_0/demon_left_rate_model': 0.48303670745272526, 'eval/latency_0/demon_left_rate_teacher': 0.49082313681868744, 'eval/latency_0/demon_right_precision': 0.8211738891936369, 'eval/latency_0/demon_right_recall': 0.8457627118644068, 'eval/latency_0/demon_right_f1': 0.83328694684108, 'eval/latency_0/demon_right_rate_model': 0.506952169076752, 'eval/latency_0/demon_right_rate_teacher': 0.4922135706340378, 'eval/latency_0/demon_move_f1': 0.8286937304165423, 'eval/latency_0/demon_cc_f1': 0.8889235122894948, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 3542.0, 'rl_games_eval/mid_train/mean_length': 3600.0, 'rl_games_eval/mid_train/std_reward': 3010.3365260382434, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/macro_mean_reward': 3542.0, 'rl_games_eval/mid_train/macro_mean_length': 3600.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/demon_attack__latency_0/mean_reward': 3542.0, 'rl_games_eval/mid_train/demon_attack__latency_0/mean_length': 3600.0, 'rl_games_eval/mid_train/demon_attack__latency_0/std_reward': 3010.3365260382434, 'rl_games_eval/mid_train/demon_attack__latency_0/std_length': 0.0, 'rl_games_eval/mid_train/demon_attack__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 3542.0, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.005244542029686272, 'timing/model': 1.2059904219349846, 'learning_rate/qwen_vl_interface': 2.336982936152347e-06, 'learning_rate/action_model': 2.3369829361523468e-05, 'epoch': 1.47}) +[2026-06-17 15:39:13,959][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_5000_state +[2026-06-17 15:39:13,961][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 15:39:17,536][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_5000_state/model.safetensors +[2026-06-17 15:39:25,259][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_5000_state/optimizer.bin +[2026-06-17 15:39:25,260][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_5000_state/sampler.bin +[2026-06-17 15:39:25,260][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_5000_state/sampler_1.bin +[2026-06-17 15:39:25,261][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_5000_state/random_states_0.pkl +[2026-06-17 15:39:25,317][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 5000 +[2026-06-17 15:39:25,317][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 15:39:25,371][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 15:39:29,992][starVLA.training.train_starvla][INFO] - Step 5001, Loss: {'train/loss': 0.6667400002479553, 'train/grad_norm_pre_clip': 14.85877799987793, 'timing/data': 0.005643315962515771, 'timing/model': 1.1866405530599877, 'learning_rate/qwen_vl_interface': 2.3352749810279047e-06, 'learning_rate/action_model': 2.3352749810279047e-05, 'epoch': 1.47}) +[2026-06-17 15:39:34,724][starVLA.training.train_starvla][INFO] - Step 5002, Loss: {'train/loss': 0.6437432169914246, 'train/grad_norm_pre_clip': 14.902521133422852, 'timing/data': 0.005616529029794037, 'timing/model': 1.1864343320485204, 'learning_rate/qwen_vl_interface': 2.3335676301290972e-06, 'learning_rate/action_model': 2.3335676301290973e-05, 'epoch': 1.47}) +[2026-06-17 15:39:39,450][starVLA.training.train_starvla][INFO] - Step 5003, Loss: {'train/loss': 0.760660707950592, 'train/grad_norm_pre_clip': 15.885090827941895, 'timing/data': 0.00569644698407501, 'timing/model': 1.1909996710019186, 'learning_rate/qwen_vl_interface': 2.331860883809857e-06, 'learning_rate/action_model': 2.3318608838098565e-05, 'epoch': 1.47}) +[2026-06-17 15:39:44,188][starVLA.training.train_starvla][INFO] - Step 5004, Loss: {'train/loss': 0.74424809217453, 'train/grad_norm_pre_clip': 14.412233352661133, 'timing/data': 0.0053555070189759135, 'timing/model': 1.3234011930180714, 'learning_rate/qwen_vl_interface': 2.330154742423993e-06, 'learning_rate/action_model': 2.330154742423993e-05, 'epoch': 1.47}) +[2026-06-17 15:39:48,799][starVLA.training.train_starvla][INFO] - Step 5005, Loss: {'train/loss': 0.6427997946739197, 'train/grad_norm_pre_clip': 13.341727256774902, 'timing/data': 0.005743349087424576, 'timing/model': 1.1951170850079507, 'learning_rate/qwen_vl_interface': 2.328449206325195e-06, 'learning_rate/action_model': 2.3284492063251947e-05, 'epoch': 1.47}) +[2026-06-17 15:39:53,544][starVLA.training.train_starvla][INFO] - Step 5006, Loss: {'train/loss': 0.5117923021316528, 'train/grad_norm_pre_clip': 11.104818344116211, 'timing/data': 0.005715954932384193, 'timing/model': 1.195476923021488, 'learning_rate/qwen_vl_interface': 2.32674427586702e-06, 'learning_rate/action_model': 2.3267442758670198e-05, 'epoch': 1.47}) +[2026-06-17 15:39:58,280][starVLA.training.train_starvla][INFO] - Step 5007, Loss: {'train/loss': 0.8148047924041748, 'train/grad_norm_pre_clip': 13.464506149291992, 'timing/data': 0.005331757012754679, 'timing/model': 1.2011993819614872, 'learning_rate/qwen_vl_interface': 2.3250399514029018e-06, 'learning_rate/action_model': 2.3250399514029016e-05, 'epoch': 1.47}) +[2026-06-17 15:40:03,061][starVLA.training.train_starvla][INFO] - Step 5008, Loss: {'train/loss': 0.7198225259780884, 'train/grad_norm_pre_clip': 20.941423416137695, 'timing/data': 0.005335732013918459, 'timing/model': 1.2023043429944664, 'learning_rate/qwen_vl_interface': 2.32333623328615e-06, 'learning_rate/action_model': 2.3233362332861498e-05, 'epoch': 1.47}) +[2026-06-17 15:40:07,834][starVLA.training.train_starvla][INFO] - Step 5009, Loss: {'train/loss': 0.6878992915153503, 'train/grad_norm_pre_clip': 15.708850860595703, 'timing/data': 0.005373508902266622, 'timing/model': 1.204634407069534, 'learning_rate/qwen_vl_interface': 2.3216331218699457e-06, 'learning_rate/action_model': 2.3216331218699456e-05, 'epoch': 1.47}) +[2026-06-17 15:40:12,604][starVLA.training.train_starvla][INFO] - Step 5010, Loss: {'train/loss': 0.4902024269104004, 'train/grad_norm_pre_clip': 12.620172500610352, 'timing/data': 0.005024202051572502, 'timing/model': 1.3272234440082684, 'learning_rate/qwen_vl_interface': 2.3199306175073465e-06, 'learning_rate/action_model': 2.3199306175073467e-05, 'epoch': 1.47}) +[2026-06-17 15:40:17,361][starVLA.training.train_starvla][INFO] - Step 5011, Loss: {'train/loss': 0.6546717286109924, 'train/grad_norm_pre_clip': 14.871950149536133, 'timing/data': 0.005580155993811786, 'timing/model': 1.3167450350010768, 'learning_rate/qwen_vl_interface': 2.318228720551283e-06, 'learning_rate/action_model': 2.318228720551283e-05, 'epoch': 1.47}) +[2026-06-17 15:40:22,000][starVLA.training.train_starvla][INFO] - Step 5012, Loss: {'train/loss': 0.6526212096214294, 'train/grad_norm_pre_clip': 12.488706588745117, 'timing/data': 0.0057497849920764565, 'timing/model': 1.1979363990249112, 'learning_rate/qwen_vl_interface': 2.3165274313545606e-06, 'learning_rate/action_model': 2.3165274313545606e-05, 'epoch': 1.47}) +[2026-06-17 15:40:26,744][starVLA.training.train_starvla][INFO] - Step 5013, Loss: {'train/loss': 0.5249856114387512, 'train/grad_norm_pre_clip': 14.666760444641113, 'timing/data': 0.0051542280707508326, 'timing/model': 1.1975116080138832, 'learning_rate/qwen_vl_interface': 2.3148267502698584e-06, 'learning_rate/action_model': 2.3148267502698584e-05, 'epoch': 1.48}) +[2026-06-17 15:40:31,517][starVLA.training.train_starvla][INFO] - Step 5014, Loss: {'train/loss': 0.6402319669723511, 'train/grad_norm_pre_clip': 15.928277015686035, 'timing/data': 0.005360136041417718, 'timing/model': 1.1999549110187218, 'learning_rate/qwen_vl_interface': 2.313126677649729e-06, 'learning_rate/action_model': 2.313126677649729e-05, 'epoch': 1.48}) +[2026-06-17 15:40:36,306][starVLA.training.train_starvla][INFO] - Step 5015, Loss: {'train/loss': 0.6729004383087158, 'train/grad_norm_pre_clip': 78.09210205078125, 'timing/data': 0.00549336604308337, 'timing/model': 1.2002026619156823, 'learning_rate/qwen_vl_interface': 2.3114272138465975e-06, 'learning_rate/action_model': 2.3114272138465977e-05, 'epoch': 1.48}) +[2026-06-17 15:40:41,023][starVLA.training.train_starvla][INFO] - Step 5016, Loss: {'train/loss': 0.4089144766330719, 'train/grad_norm_pre_clip': 10.758979797363281, 'timing/data': 0.005983529961667955, 'timing/model': 1.2653151369886473, 'learning_rate/qwen_vl_interface': 2.3097283592127706e-06, 'learning_rate/action_model': 2.3097283592127706e-05, 'epoch': 1.48}) +[2026-06-17 15:40:45,662][starVLA.training.train_starvla][INFO] - Step 5017, Loss: {'train/loss': 0.5654396414756775, 'train/grad_norm_pre_clip': 12.975789070129395, 'timing/data': 0.005281109013594687, 'timing/model': 1.1988242170773447, 'learning_rate/qwen_vl_interface': 2.3080301141004165e-06, 'learning_rate/action_model': 2.3080301141004165e-05, 'epoch': 1.48}) +[2026-06-17 15:40:50,419][starVLA.training.train_starvla][INFO] - Step 5018, Loss: {'train/loss': 0.6316842436790466, 'train/grad_norm_pre_clip': 17.706436157226562, 'timing/data': 0.005250419955700636, 'timing/model': 1.20653204806149, 'learning_rate/qwen_vl_interface': 2.3063324788615843e-06, 'learning_rate/action_model': 2.306332478861584e-05, 'epoch': 1.48}) +[2026-06-17 15:40:55,214][starVLA.training.train_starvla][INFO] - Step 5019, Loss: {'train/loss': 0.7983488440513611, 'train/grad_norm_pre_clip': 17.164810180664062, 'timing/data': 0.005738916923291981, 'timing/model': 1.2028189840493724, 'learning_rate/qwen_vl_interface': 2.304635453848198e-06, 'learning_rate/action_model': 2.3046354538481983e-05, 'epoch': 1.48}) +[2026-06-17 15:40:59,989][starVLA.training.train_starvla][INFO] - Step 5020, Loss: {'train/loss': 0.6161834001541138, 'train/grad_norm_pre_clip': 13.690858840942383, 'timing/data': 0.005356221925467253, 'timing/model': 1.201681009028107, 'learning_rate/qwen_vl_interface': 2.3029390394120517e-06, 'learning_rate/action_model': 2.3029390394120513e-05, 'epoch': 1.48}) +[2026-06-17 15:41:04,824][starVLA.training.train_starvla][INFO] - Step 5021, Loss: {'train/loss': 0.5414074659347534, 'train/grad_norm_pre_clip': 13.747133255004883, 'timing/data': 0.0052424350287765265, 'timing/model': 1.3793299179524183, 'learning_rate/qwen_vl_interface': 2.301243235904813e-06, 'learning_rate/action_model': 2.3012432359048133e-05, 'epoch': 1.48}) +[2026-06-17 15:41:09,551][starVLA.training.train_starvla][INFO] - Step 5022, Loss: {'train/loss': 0.5766839385032654, 'train/grad_norm_pre_clip': 13.163557052612305, 'timing/data': 0.005565391038544476, 'timing/model': 1.2826011809520423, 'learning_rate/qwen_vl_interface': 2.299548043678026e-06, 'learning_rate/action_model': 2.299548043678026e-05, 'epoch': 1.48}) +[2026-06-17 15:41:14,197][starVLA.training.train_starvla][INFO] - Step 5023, Loss: {'train/loss': 0.6620535850524902, 'train/grad_norm_pre_clip': 12.819994926452637, 'timing/data': 0.005335111985914409, 'timing/model': 1.2009571429807693, 'learning_rate/qwen_vl_interface': 2.297853463083103e-06, 'learning_rate/action_model': 2.297853463083103e-05, 'epoch': 1.48}) +[2026-06-17 15:41:18,998][starVLA.training.train_starvla][INFO] - Step 5024, Loss: {'train/loss': 0.6645802855491638, 'train/grad_norm_pre_clip': 11.040977478027344, 'timing/data': 0.0056363679468631744, 'timing/model': 1.2174989280756563, 'learning_rate/qwen_vl_interface': 2.296159494471335e-06, 'learning_rate/action_model': 2.2961594944713347e-05, 'epoch': 1.48}) +[2026-06-17 15:41:23,792][starVLA.training.train_starvla][INFO] - Step 5025, Loss: {'train/loss': 0.671157956123352, 'train/grad_norm_pre_clip': 15.363922119140625, 'timing/data': 0.005240345955826342, 'timing/model': 1.2070786339463666, 'learning_rate/qwen_vl_interface': 2.294466138193881e-06, 'learning_rate/action_model': 2.2944661381938812e-05, 'epoch': 1.48}) +[2026-06-17 15:41:28,599][starVLA.training.train_starvla][INFO] - Step 5026, Loss: {'train/loss': 0.5748976469039917, 'train/grad_norm_pre_clip': 13.185384750366211, 'timing/data': 0.005432546022348106, 'timing/model': 1.2073706099763513, 'learning_rate/qwen_vl_interface': 2.2927733946017777e-06, 'learning_rate/action_model': 2.2927733946017778e-05, 'epoch': 1.48}) +[2026-06-17 15:41:33,375][starVLA.training.train_starvla][INFO] - Step 5027, Loss: {'train/loss': 0.7161558866500854, 'train/grad_norm_pre_clip': 15.785621643066406, 'timing/data': 0.005583305028267205, 'timing/model': 1.2023895720485598, 'learning_rate/qwen_vl_interface': 2.2910812640459314e-06, 'learning_rate/action_model': 2.2910812640459315e-05, 'epoch': 1.48}) +[2026-06-17 15:41:38,118][starVLA.training.train_starvla][INFO] - Step 5028, Loss: {'train/loss': 0.9919837117195129, 'train/grad_norm_pre_clip': 25.25115394592285, 'timing/data': 0.005255257012322545, 'timing/model': 1.2953861700370908, 'learning_rate/qwen_vl_interface': 2.289389746877123e-06, 'learning_rate/action_model': 2.2893897468771228e-05, 'epoch': 1.48}) +[2026-06-17 15:41:42,780][starVLA.training.train_starvla][INFO] - Step 5029, Loss: {'train/loss': 0.7569695711135864, 'train/grad_norm_pre_clip': 16.68226432800293, 'timing/data': 0.006164490012452006, 'timing/model': 1.209686164977029, 'learning_rate/qwen_vl_interface': 2.2876988434460046e-06, 'learning_rate/action_model': 2.2876988434460046e-05, 'epoch': 1.48}) +[2026-06-17 15:41:47,569][starVLA.training.train_starvla][INFO] - Step 5030, Loss: {'train/loss': 0.7582306265830994, 'train/grad_norm_pre_clip': 13.526768684387207, 'timing/data': 0.0056898200418800116, 'timing/model': 1.2019374889787287, 'learning_rate/qwen_vl_interface': 2.2860085541031048e-06, 'learning_rate/action_model': 2.2860085541031047e-05, 'epoch': 1.48}) +[2026-06-17 15:41:52,359][starVLA.training.train_starvla][INFO] - Step 5031, Loss: {'train/loss': 0.5756535530090332, 'train/grad_norm_pre_clip': 11.942926406860352, 'timing/data': 0.0054395420011132956, 'timing/model': 1.2058826360153034, 'learning_rate/qwen_vl_interface': 2.284318879198823e-06, 'learning_rate/action_model': 2.284318879198823e-05, 'epoch': 1.48}) +[2026-06-17 15:41:57,157][starVLA.training.train_starvla][INFO] - Step 5032, Loss: {'train/loss': 0.619918942451477, 'train/grad_norm_pre_clip': 12.135092735290527, 'timing/data': 0.005817480036057532, 'timing/model': 1.2026040810160339, 'learning_rate/qwen_vl_interface': 2.2826298190834266e-06, 'learning_rate/action_model': 2.2826298190834266e-05, 'epoch': 1.48}) +[2026-06-17 15:42:01,946][starVLA.training.train_starvla][INFO] - Step 5033, Loss: {'train/loss': 0.8061255216598511, 'train/grad_norm_pre_clip': 14.866130828857422, 'timing/data': 0.00555168092250824, 'timing/model': 1.3356124880956486, 'learning_rate/qwen_vl_interface': 2.2809413741070603e-06, 'learning_rate/action_model': 2.28094137410706e-05, 'epoch': 1.48}) +[2026-06-17 15:42:06,606][starVLA.training.train_starvla][INFO] - Step 5034, Loss: {'train/loss': 0.5810419917106628, 'train/grad_norm_pre_clip': 12.394613265991211, 'timing/data': 0.0058259989600628614, 'timing/model': 1.2083380139665678, 'learning_rate/qwen_vl_interface': 2.2792535446197435e-06, 'learning_rate/action_model': 2.2792535446197436e-05, 'epoch': 1.48}) +[2026-06-17 15:42:11,395][starVLA.training.train_starvla][INFO] - Step 5035, Loss: {'train/loss': 1.0434620380401611, 'train/grad_norm_pre_clip': 15.97980785369873, 'timing/data': 0.005579207092523575, 'timing/model': 1.201478666975163, 'learning_rate/qwen_vl_interface': 2.2775663309713643e-06, 'learning_rate/action_model': 2.2775663309713645e-05, 'epoch': 1.48}) +[2026-06-17 15:42:16,182][starVLA.training.train_starvla][INFO] - Step 5036, Loss: {'train/loss': 0.6750910878181458, 'train/grad_norm_pre_clip': 13.728126525878906, 'timing/data': 0.005827413988299668, 'timing/model': 1.202904846984893, 'learning_rate/qwen_vl_interface': 2.2758797335116843e-06, 'learning_rate/action_model': 2.2758797335116844e-05, 'epoch': 1.48}) +[2026-06-17 15:42:20,882][starVLA.training.train_starvla][INFO] - Step 5037, Loss: {'train/loss': 0.8071973323822021, 'train/grad_norm_pre_clip': 16.816099166870117, 'timing/data': 0.005814166972413659, 'timing/model': 1.2044305599993095, 'learning_rate/qwen_vl_interface': 2.2741937525903324e-06, 'learning_rate/action_model': 2.2741937525903324e-05, 'epoch': 1.48}) +[2026-06-17 15:42:25,626][starVLA.training.train_starvla][INFO] - Step 5038, Loss: {'train/loss': 0.5895208716392517, 'train/grad_norm_pre_clip': 13.117713928222656, 'timing/data': 0.005341631011106074, 'timing/model': 1.2131680770544335, 'learning_rate/qwen_vl_interface': 2.2725083885568198e-06, 'learning_rate/action_model': 2.2725083885568197e-05, 'epoch': 1.48}) +[2026-06-17 15:42:30,415][starVLA.training.train_starvla][INFO] - Step 5039, Loss: {'train/loss': 0.7940235733985901, 'train/grad_norm_pre_clip': 13.943671226501465, 'timing/data': 0.0054993960075080395, 'timing/model': 1.3367798489052802, 'learning_rate/qwen_vl_interface': 2.2708236417605217e-06, 'learning_rate/action_model': 2.2708236417605216e-05, 'epoch': 1.48}) +[2026-06-17 15:42:35,069][starVLA.training.train_starvla][INFO] - Step 5040, Loss: {'train/loss': 0.7252406477928162, 'train/grad_norm_pre_clip': 15.546640396118164, 'timing/data': 0.0057261360343545675, 'timing/model': 1.2041332089575008, 'learning_rate/qwen_vl_interface': 2.2691395125506885e-06, 'learning_rate/action_model': 2.2691395125506885e-05, 'epoch': 1.48}) +[2026-06-17 15:42:39,907][starVLA.training.train_starvla][INFO] - Step 5041, Loss: {'train/loss': 1.0539244413375854, 'train/grad_norm_pre_clip': 14.895621299743652, 'timing/data': 0.005118396016769111, 'timing/model': 1.2052498399280012, 'learning_rate/qwen_vl_interface': 2.2674560012764426e-06, 'learning_rate/action_model': 2.2674560012764423e-05, 'epoch': 1.48}) +[2026-06-17 15:42:44,690][starVLA.training.train_starvla][INFO] - Step 5042, Loss: {'train/loss': 0.5766707062721252, 'train/grad_norm_pre_clip': 14.39220142364502, 'timing/data': 0.0057173059321939945, 'timing/model': 1.2049664430087432, 'learning_rate/qwen_vl_interface': 2.2657731082867764e-06, 'learning_rate/action_model': 2.265773108286776e-05, 'epoch': 1.48}) +[2026-06-17 15:42:49,467][starVLA.training.train_starvla][INFO] - Step 5043, Loss: {'train/loss': 0.7029193639755249, 'train/grad_norm_pre_clip': 13.991026878356934, 'timing/data': 0.005442868918180466, 'timing/model': 1.2065398399718106, 'learning_rate/qwen_vl_interface': 2.264090833930555e-06, 'learning_rate/action_model': 2.2640908339305553e-05, 'epoch': 1.48}) +[2026-06-17 15:42:54,214][starVLA.training.train_starvla][INFO] - Step 5044, Loss: {'train/loss': 0.673893928527832, 'train/grad_norm_pre_clip': 15.066166877746582, 'timing/data': 0.005697900080122054, 'timing/model': 1.2034740260569379, 'learning_rate/qwen_vl_interface': 2.2624091785565165e-06, 'learning_rate/action_model': 2.2624091785565167e-05, 'epoch': 1.48}) +[2026-06-17 15:42:59,004][starVLA.training.train_starvla][INFO] - Step 5045, Loss: {'train/loss': 0.8642535209655762, 'train/grad_norm_pre_clip': 15.240751266479492, 'timing/data': 0.005716539919376373, 'timing/model': 1.334641422959976, 'learning_rate/qwen_vl_interface': 2.2607281425132728e-06, 'learning_rate/action_model': 2.260728142513273e-05, 'epoch': 1.48}) +[2026-06-17 15:43:03,723][starVLA.training.train_starvla][INFO] - Step 5046, Loss: {'train/loss': 0.6706401705741882, 'train/grad_norm_pre_clip': 12.032272338867188, 'timing/data': 0.005614444031380117, 'timing/model': 1.2050412889802828, 'learning_rate/qwen_vl_interface': 2.2590477261493e-06, 'learning_rate/action_model': 2.2590477261493e-05, 'epoch': 1.48}) +[2026-06-17 15:43:08,521][starVLA.training.train_starvla][INFO] - Step 5047, Loss: {'train/loss': 0.49450141191482544, 'train/grad_norm_pre_clip': 10.25269603729248, 'timing/data': 0.005031462060287595, 'timing/model': 1.206595059018582, 'learning_rate/qwen_vl_interface': 2.257367929812953e-06, 'learning_rate/action_model': 2.2573679298129525e-05, 'epoch': 1.49}) +[2026-06-17 15:43:13,326][starVLA.training.train_starvla][INFO] - Step 5048, Loss: {'train/loss': 0.6553289294242859, 'train/grad_norm_pre_clip': 15.132573127746582, 'timing/data': 0.0058424510061740875, 'timing/model': 1.2045859729405493, 'learning_rate/qwen_vl_interface': 2.2556887538524514e-06, 'learning_rate/action_model': 2.2556887538524512e-05, 'epoch': 1.49}) +[2026-06-17 15:43:18,125][starVLA.training.train_starvla][INFO] - Step 5049, Loss: {'train/loss': 0.507408082485199, 'train/grad_norm_pre_clip': 12.584110260009766, 'timing/data': 0.005822497070766985, 'timing/model': 1.2035115599865094, 'learning_rate/qwen_vl_interface': 2.2540101986158953e-06, 'learning_rate/action_model': 2.2540101986158953e-05, 'epoch': 1.49}) +[2026-06-17 15:43:22,925][starVLA.training.train_starvla][INFO] - Step 5050, Loss: {'train/loss': 0.7627331614494324, 'train/grad_norm_pre_clip': 12.127875328063965, 'timing/data': 0.005173329031094909, 'timing/model': 1.3406537600094453, 'learning_rate/qwen_vl_interface': 2.2523322644512506e-06, 'learning_rate/action_model': 2.2523322644512504e-05, 'epoch': 1.49}) +[2026-06-17 15:43:27,589][starVLA.training.train_starvla][INFO] - Step 5051, Loss: {'train/loss': 0.6619269847869873, 'train/grad_norm_pre_clip': 14.507627487182617, 'timing/data': 0.005575246992520988, 'timing/model': 1.2103705900954083, 'learning_rate/qwen_vl_interface': 2.2506549517063497e-06, 'learning_rate/action_model': 2.2506549517063496e-05, 'epoch': 1.49}) +[2026-06-17 15:43:32,402][starVLA.training.train_starvla][INFO] - Step 5052, Loss: {'train/loss': 0.6035056114196777, 'train/grad_norm_pre_clip': 12.24089241027832, 'timing/data': 0.005687761935405433, 'timing/model': 1.2086974040139467, 'learning_rate/qwen_vl_interface': 2.248978260728907e-06, 'learning_rate/action_model': 2.2489782607289067e-05, 'epoch': 1.49}) +[2026-06-17 15:43:37,110][starVLA.training.train_starvla][INFO] - Step 5053, Loss: {'train/loss': 0.6315211653709412, 'train/grad_norm_pre_clip': 11.825469970703125, 'timing/data': 0.0052478479919955134, 'timing/model': 1.2041286421008408, 'learning_rate/qwen_vl_interface': 2.2473021918664987e-06, 'learning_rate/action_model': 2.247302191866499e-05, 'epoch': 1.49}) +[2026-06-17 15:43:41,856][starVLA.training.train_starvla][INFO] - Step 5054, Loss: {'train/loss': 0.5166379809379578, 'train/grad_norm_pre_clip': 11.589666366577148, 'timing/data': 0.005660621915012598, 'timing/model': 1.2073544649174437, 'learning_rate/qwen_vl_interface': 2.245626745466578e-06, 'learning_rate/action_model': 2.2456267454665782e-05, 'epoch': 1.49}) +[2026-06-17 15:43:46,660][starVLA.training.train_starvla][INFO] - Step 5055, Loss: {'train/loss': 0.6442931890487671, 'train/grad_norm_pre_clip': 24.405841827392578, 'timing/data': 0.005144117050804198, 'timing/model': 1.2118009199621156, 'learning_rate/qwen_vl_interface': 2.2439519218764657e-06, 'learning_rate/action_model': 2.2439519218764656e-05, 'epoch': 1.49}) +[2026-06-17 15:43:51,460][starVLA.training.train_starvla][INFO] - Step 5056, Loss: {'train/loss': 0.6942298412322998, 'train/grad_norm_pre_clip': 16.301137924194336, 'timing/data': 0.005720590939745307, 'timing/model': 1.3327860209392384, 'learning_rate/qwen_vl_interface': 2.242277721443355e-06, 'learning_rate/action_model': 2.2422777214433545e-05, 'epoch': 1.49}) +[2026-06-17 15:43:56,115][starVLA.training.train_starvla][INFO] - Step 5057, Loss: {'train/loss': 0.47690191864967346, 'train/grad_norm_pre_clip': 12.784944534301758, 'timing/data': 0.005929382052272558, 'timing/model': 1.203789617982693, 'learning_rate/qwen_vl_interface': 2.2406041445143086e-06, 'learning_rate/action_model': 2.2406041445143086e-05, 'epoch': 1.49}) +[2026-06-17 15:44:00,903][starVLA.training.train_starvla][INFO] - Step 5058, Loss: {'train/loss': 0.6278979182243347, 'train/grad_norm_pre_clip': 14.38418960571289, 'timing/data': 0.0052983020432293415, 'timing/model': 1.2062362589640543, 'learning_rate/qwen_vl_interface': 2.2389311914362614e-06, 'learning_rate/action_model': 2.2389311914362616e-05, 'epoch': 1.49}) +[2026-06-17 15:44:05,692][starVLA.training.train_starvla][INFO] - Step 5059, Loss: {'train/loss': 0.7757066488265991, 'train/grad_norm_pre_clip': 38.109066009521484, 'timing/data': 0.005685516051016748, 'timing/model': 1.2043877680553123, 'learning_rate/qwen_vl_interface': 2.2372588625560185e-06, 'learning_rate/action_model': 2.2372588625560185e-05, 'epoch': 1.49}) +[2026-06-17 15:44:10,457][starVLA.training.train_starvla][INFO] - Step 5060, Loss: {'train/loss': 0.5855758786201477, 'train/grad_norm_pre_clip': 13.549052238464355, 'timing/data': 0.005763660999946296, 'timing/model': 1.2157528719399124, 'learning_rate/qwen_vl_interface': 2.2355871582202544e-06, 'learning_rate/action_model': 2.2355871582202544e-05, 'epoch': 1.49}) +[2026-06-17 15:44:15,252][starVLA.training.train_starvla][INFO] - Step 5061, Loss: {'train/loss': 0.4331052005290985, 'train/grad_norm_pre_clip': 14.696382522583008, 'timing/data': 0.0052800970152020454, 'timing/model': 1.2071491519454867, 'learning_rate/qwen_vl_interface': 2.2339160787755175e-06, 'learning_rate/action_model': 2.2339160787755174e-05, 'epoch': 1.49}) +[2026-06-17 15:44:20,060][starVLA.training.train_starvla][INFO] - Step 5062, Loss: {'train/loss': 0.5708126425743103, 'train/grad_norm_pre_clip': 12.302070617675781, 'timing/data': 0.005873810034245253, 'timing/model': 1.3474522759206593, 'learning_rate/qwen_vl_interface': 2.23224562456822e-06, 'learning_rate/action_model': 2.23224562456822e-05, 'epoch': 1.49}) +[2026-06-17 15:44:24,720][starVLA.training.train_starvla][INFO] - Step 5063, Loss: {'train/loss': 0.6005383133888245, 'train/grad_norm_pre_clip': 22.700969696044922, 'timing/data': 0.005474084988236427, 'timing/model': 1.2034779519308358, 'learning_rate/qwen_vl_interface': 2.2305757959446538e-06, 'learning_rate/action_model': 2.2305757959446537e-05, 'epoch': 1.49}) +[2026-06-17 15:44:29,516][starVLA.training.train_starvla][INFO] - Step 5064, Loss: {'train/loss': 0.671330988407135, 'train/grad_norm_pre_clip': 11.946237564086914, 'timing/data': 0.005547648994252086, 'timing/model': 1.2125010419404134, 'learning_rate/qwen_vl_interface': 2.228906593250976e-06, 'learning_rate/action_model': 2.228906593250976e-05, 'epoch': 1.49}) +[2026-06-17 15:44:34,309][starVLA.training.train_starvla][INFO] - Step 5065, Loss: {'train/loss': 0.9422129392623901, 'train/grad_norm_pre_clip': 17.28790855407715, 'timing/data': 0.0053472110303118825, 'timing/model': 1.2100901900557801, 'learning_rate/qwen_vl_interface': 2.2272380168332105e-06, 'learning_rate/action_model': 2.2272380168332103e-05, 'epoch': 1.49}) +[2026-06-17 15:44:39,127][starVLA.training.train_starvla][INFO] - Step 5066, Loss: {'train/loss': 1.04252028465271, 'train/grad_norm_pre_clip': 13.919719696044922, 'timing/data': 0.0057406669948250055, 'timing/model': 1.2025319029344246, 'learning_rate/qwen_vl_interface': 2.225570067037255e-06, 'learning_rate/action_model': 2.2255700670372548e-05, 'epoch': 1.49}) +[2026-06-17 15:44:43,854][starVLA.training.train_starvla][INFO] - Step 5067, Loss: {'train/loss': 0.8492394089698792, 'train/grad_norm_pre_clip': 15.283125877380371, 'timing/data': 0.005784567096270621, 'timing/model': 1.2083834480727091, 'learning_rate/qwen_vl_interface': 2.223902744208881e-06, 'learning_rate/action_model': 2.2239027442088808e-05, 'epoch': 1.49}) +[2026-06-17 15:44:48,558][starVLA.training.train_starvla][INFO] - Step 5068, Loss: {'train/loss': 0.49793490767478943, 'train/grad_norm_pre_clip': 12.509594917297363, 'timing/data': 0.005651345010846853, 'timing/model': 1.250654423958622, 'learning_rate/qwen_vl_interface': 2.2222360486937238e-06, 'learning_rate/action_model': 2.2222360486937235e-05, 'epoch': 1.49}) +[2026-06-17 15:44:53,230][starVLA.training.train_starvla][INFO] - Step 5069, Loss: {'train/loss': 0.5546397566795349, 'train/grad_norm_pre_clip': 12.364175796508789, 'timing/data': 0.005896196002140641, 'timing/model': 1.214387875981629, 'learning_rate/qwen_vl_interface': 2.220569980837292e-06, 'learning_rate/action_model': 2.220569980837292e-05, 'epoch': 1.49}) +[2026-06-17 15:44:57,976][starVLA.training.train_starvla][INFO] - Step 5070, Loss: {'train/loss': 0.7027453184127808, 'train/grad_norm_pre_clip': 11.226518630981445, 'timing/data': 0.005415617022663355, 'timing/model': 1.2102082530036569, 'learning_rate/qwen_vl_interface': 2.2189045409849625e-06, 'learning_rate/action_model': 2.2189045409849628e-05, 'epoch': 1.49}) +[2026-06-17 15:45:02,789][starVLA.training.train_starvla][INFO] - Step 5071, Loss: {'train/loss': 0.5267869830131531, 'train/grad_norm_pre_clip': 11.348828315734863, 'timing/data': 0.005847770953550935, 'timing/model': 1.210813182988204, 'learning_rate/qwen_vl_interface': 2.2172397294819843e-06, 'learning_rate/action_model': 2.2172397294819843e-05, 'epoch': 1.49}) +[2026-06-17 15:45:07,572][starVLA.training.train_starvla][INFO] - Step 5072, Loss: {'train/loss': 0.8477930426597595, 'train/grad_norm_pre_clip': 13.095515251159668, 'timing/data': 0.0056729489006102085, 'timing/model': 1.2022719280794263, 'learning_rate/qwen_vl_interface': 2.215575546673473e-06, 'learning_rate/action_model': 2.215575546673473e-05, 'epoch': 1.49}) +[2026-06-17 15:45:12,353][starVLA.training.train_starvla][INFO] - Step 5073, Loss: {'train/loss': 0.49918147921562195, 'train/grad_norm_pre_clip': 10.302412986755371, 'timing/data': 0.005838591023348272, 'timing/model': 1.32900696399156, 'learning_rate/qwen_vl_interface': 2.213911992904416e-06, 'learning_rate/action_model': 2.213911992904416e-05, 'epoch': 1.49}) +[2026-06-17 15:45:17,062][starVLA.training.train_starvla][INFO] - Step 5074, Loss: {'train/loss': 0.6394750475883484, 'train/grad_norm_pre_clip': 18.600204467773438, 'timing/data': 0.005556488991715014, 'timing/model': 1.2545336929615587, 'learning_rate/qwen_vl_interface': 2.2122490685196703e-06, 'learning_rate/action_model': 2.2122490685196704e-05, 'epoch': 1.49}) +[2026-06-17 15:45:21,730][starVLA.training.train_starvla][INFO] - Step 5075, Loss: {'train/loss': 0.6223687529563904, 'train/grad_norm_pre_clip': 11.479639053344727, 'timing/data': 0.00600691803265363, 'timing/model': 1.2169062390457839, 'learning_rate/qwen_vl_interface': 2.210586773863961e-06, 'learning_rate/action_model': 2.210586773863961e-05, 'epoch': 1.49}) +[2026-06-17 15:45:26,488][starVLA.training.train_starvla][INFO] - Step 5076, Loss: {'train/loss': 0.4997129738330841, 'train/grad_norm_pre_clip': 10.821578979492188, 'timing/data': 0.005667352001182735, 'timing/model': 1.2078483839286491, 'learning_rate/qwen_vl_interface': 2.208925109281884e-06, 'learning_rate/action_model': 2.208925109281884e-05, 'epoch': 1.49}) +[2026-06-17 15:45:31,279][starVLA.training.train_starvla][INFO] - Step 5077, Loss: {'train/loss': 0.6660151481628418, 'train/grad_norm_pre_clip': 16.34803009033203, 'timing/data': 0.00590592494700104, 'timing/model': 1.2076903430279344, 'learning_rate/qwen_vl_interface': 2.2072640751179026e-06, 'learning_rate/action_model': 2.2072640751179025e-05, 'epoch': 1.49}) +[2026-06-17 15:45:36,068][starVLA.training.train_starvla][INFO] - Step 5078, Loss: {'train/loss': 0.8495171666145325, 'train/grad_norm_pre_clip': 14.00230884552002, 'timing/data': 0.005542309023439884, 'timing/model': 1.203726557083428, 'learning_rate/qwen_vl_interface': 2.205603671716356e-06, 'learning_rate/action_model': 2.2056036717163557e-05, 'epoch': 1.49}) +[2026-06-17 15:45:40,856][starVLA.training.train_starvla][INFO] - Step 5079, Loss: {'train/loss': 0.700449526309967, 'train/grad_norm_pre_clip': 16.21299171447754, 'timing/data': 0.005786983063444495, 'timing/model': 1.3319562070537359, 'learning_rate/qwen_vl_interface': 2.2039438994214424e-06, 'learning_rate/action_model': 2.2039438994214424e-05, 'epoch': 1.49}) +[2026-06-17 15:45:45,636][starVLA.training.train_starvla][INFO] - Step 5080, Loss: {'train/loss': 0.5395320653915405, 'train/grad_norm_pre_clip': 11.399249076843262, 'timing/data': 0.005835265968926251, 'timing/model': 1.326298450003378, 'learning_rate/qwen_vl_interface': 2.202284758577235e-06, 'learning_rate/action_model': 2.2022847585772347e-05, 'epoch': 1.49}) +[2026-06-17 15:45:50,296][starVLA.training.train_starvla][INFO] - Step 5081, Loss: {'train/loss': 0.8418600559234619, 'train/grad_norm_pre_clip': 15.479301452636719, 'timing/data': 0.005578665994107723, 'timing/model': 1.2062754849903286, 'learning_rate/qwen_vl_interface': 2.2006262495276774e-06, 'learning_rate/action_model': 2.2006262495276775e-05, 'epoch': 1.5}) +[2026-06-17 15:45:55,070][starVLA.training.train_starvla][INFO] - Step 5082, Loss: {'train/loss': 0.4623652696609497, 'train/grad_norm_pre_clip': 12.732635498046875, 'timing/data': 0.005469193099997938, 'timing/model': 1.2089342720573768, 'learning_rate/qwen_vl_interface': 2.1989683726165795e-06, 'learning_rate/action_model': 2.1989683726165795e-05, 'epoch': 1.5}) +[2026-06-17 15:45:59,867][starVLA.training.train_starvla][INFO] - Step 5083, Loss: {'train/loss': 0.7040643692016602, 'train/grad_norm_pre_clip': 18.836910247802734, 'timing/data': 0.005807844921946526, 'timing/model': 1.2069412829587236, 'learning_rate/qwen_vl_interface': 2.197311128187623e-06, 'learning_rate/action_model': 2.1973111281876232e-05, 'epoch': 1.5}) +[2026-06-17 15:46:04,697][starVLA.training.train_starvla][INFO] - Step 5084, Loss: {'train/loss': 1.0839496850967407, 'train/grad_norm_pre_clip': 18.545751571655273, 'timing/data': 0.005754629033617675, 'timing/model': 1.2326715940143913, 'learning_rate/qwen_vl_interface': 2.1956545165843517e-06, 'learning_rate/action_model': 2.1956545165843517e-05, 'epoch': 1.5}) +[2026-06-17 15:46:09,462][starVLA.training.train_starvla][INFO] - Step 5085, Loss: {'train/loss': 0.7575151920318604, 'train/grad_norm_pre_clip': 13.709548950195312, 'timing/data': 0.005386156029999256, 'timing/model': 1.2826503929682076, 'learning_rate/qwen_vl_interface': 2.1939985381501866e-06, 'learning_rate/action_model': 2.1939985381501866e-05, 'epoch': 1.5}) +[2026-06-17 15:46:14,135][starVLA.training.train_starvla][INFO] - Step 5086, Loss: {'train/loss': 0.71917724609375, 'train/grad_norm_pre_clip': 14.503569602966309, 'timing/data': 0.005431775934994221, 'timing/model': 1.2042390099959448, 'learning_rate/qwen_vl_interface': 2.1923431932284135e-06, 'learning_rate/action_model': 2.1923431932284136e-05, 'epoch': 1.5}) +[2026-06-17 15:46:18,934][starVLA.training.train_starvla][INFO] - Step 5087, Loss: {'train/loss': 0.5257200002670288, 'train/grad_norm_pre_clip': 12.267958641052246, 'timing/data': 0.0054727219976484776, 'timing/model': 1.2047917790478095, 'learning_rate/qwen_vl_interface': 2.190688482162187e-06, 'learning_rate/action_model': 2.190688482162187e-05, 'epoch': 1.5}) +[2026-06-17 15:46:23,742][starVLA.training.train_starvla][INFO] - Step 5088, Loss: {'train/loss': 0.36394694447517395, 'train/grad_norm_pre_clip': 8.86597728729248, 'timing/data': 0.005662699928507209, 'timing/model': 1.2072447119280696, 'learning_rate/qwen_vl_interface': 2.189034405294529e-06, 'learning_rate/action_model': 2.189034405294529e-05, 'epoch': 1.5}) +[2026-06-17 15:46:28,552][starVLA.training.train_starvla][INFO] - Step 5089, Loss: {'train/loss': 0.6625761985778809, 'train/grad_norm_pre_clip': 13.802886962890625, 'timing/data': 0.005519283004105091, 'timing/model': 1.204724759911187, 'learning_rate/qwen_vl_interface': 2.1873809629683333e-06, 'learning_rate/action_model': 2.1873809629683333e-05, 'epoch': 1.5}) +[2026-06-17 15:46:33,338][starVLA.training.train_starvla][INFO] - Step 5090, Loss: {'train/loss': 0.5570849776268005, 'train/grad_norm_pre_clip': 11.745820045471191, 'timing/data': 0.005944555043242872, 'timing/model': 1.3295629679923877, 'learning_rate/qwen_vl_interface': 2.185728155526359e-06, 'learning_rate/action_model': 2.1857281555263588e-05, 'epoch': 1.5}) +[2026-06-17 15:46:38,072][starVLA.training.train_starvla][INFO] - Step 5091, Loss: {'train/loss': 0.8430874347686768, 'train/grad_norm_pre_clip': 12.768409729003906, 'timing/data': 0.005655605928041041, 'timing/model': 1.2812824960565194, 'learning_rate/qwen_vl_interface': 2.184075983311234e-06, 'learning_rate/action_model': 2.184075983311234e-05, 'epoch': 1.5}) +[2026-06-17 15:46:42,735][starVLA.training.train_starvla][INFO] - Step 5092, Loss: {'train/loss': 0.7599810361862183, 'train/grad_norm_pre_clip': 13.357634544372559, 'timing/data': 0.0052358550019562244, 'timing/model': 1.2057935709599406, 'learning_rate/qwen_vl_interface': 2.18242444666546e-06, 'learning_rate/action_model': 2.18242444666546e-05, 'epoch': 1.5}) +[2026-06-17 15:46:47,529][starVLA.training.train_starvla][INFO] - Step 5093, Loss: {'train/loss': 0.7278170585632324, 'train/grad_norm_pre_clip': 14.617631912231445, 'timing/data': 0.005679396097548306, 'timing/model': 1.2039440659573302, 'learning_rate/qwen_vl_interface': 2.1807735459313974e-06, 'learning_rate/action_model': 2.180773545931397e-05, 'epoch': 1.5}) +[2026-06-17 15:46:52,346][starVLA.training.train_starvla][INFO] - Step 5094, Loss: {'train/loss': 0.9283246397972107, 'train/grad_norm_pre_clip': 29.238914489746094, 'timing/data': 0.005610366933979094, 'timing/model': 1.2070482319686562, 'learning_rate/qwen_vl_interface': 2.1791232814512815e-06, 'learning_rate/action_model': 2.1791232814512815e-05, 'epoch': 1.5}) +[2026-06-17 15:46:57,165][starVLA.training.train_starvla][INFO] - Step 5095, Loss: {'train/loss': 0.7145108580589294, 'train/grad_norm_pre_clip': 13.193573951721191, 'timing/data': 0.005310171050950885, 'timing/model': 1.205101586994715, 'learning_rate/qwen_vl_interface': 2.1774736535672117e-06, 'learning_rate/action_model': 2.1774736535672115e-05, 'epoch': 1.5}) +[2026-06-17 15:47:01,944][starVLA.training.train_starvla][INFO] - Step 5096, Loss: {'train/loss': 0.5636643767356873, 'train/grad_norm_pre_clip': 12.507253646850586, 'timing/data': 0.005464868037961423, 'timing/model': 1.2046041230205446, 'learning_rate/qwen_vl_interface': 2.1758246626211617e-06, 'learning_rate/action_model': 2.1758246626211617e-05, 'epoch': 1.5}) +[2026-06-17 15:47:06,692][starVLA.training.train_starvla][INFO] - Step 5097, Loss: {'train/loss': 0.45389795303344727, 'train/grad_norm_pre_clip': 10.537726402282715, 'timing/data': 0.005793957971036434, 'timing/model': 1.291020788019523, 'learning_rate/qwen_vl_interface': 2.174176308954969e-06, 'learning_rate/action_model': 2.174176308954969e-05, 'epoch': 1.5}) +[2026-06-17 15:47:11,350][starVLA.training.train_starvla][INFO] - Step 5098, Loss: {'train/loss': 0.7061933279037476, 'train/grad_norm_pre_clip': 13.365477561950684, 'timing/data': 0.005408021970652044, 'timing/model': 1.2037050849758089, 'learning_rate/qwen_vl_interface': 2.172528592910335e-06, 'learning_rate/action_model': 2.1725285929103348e-05, 'epoch': 1.5}) +[2026-06-17 15:47:16,134][starVLA.training.train_starvla][INFO] - Step 5099, Loss: {'train/loss': 0.6949608325958252, 'train/grad_norm_pre_clip': 12.533617973327637, 'timing/data': 0.006051404052414, 'timing/model': 1.2029819990275428, 'learning_rate/qwen_vl_interface': 2.170881514828834e-06, 'learning_rate/action_model': 2.170881514828834e-05, 'epoch': 1.5}) +[2026-06-17 15:48:57,690][starVLA.training.train_starvla][INFO] - Step 5100, Loss: {'train/loss': 0.8231576085090637, 'train/grad_norm_pre_clip': 14.240399360656738, 'eval/loss': 0.7215235233306885, 'eval/latency_0/loss': 0.7215235233306885, 'eval/demon_fire_precision': 0.9024, 'eval/demon_fire_recall': 0.9368770764119602, 'eval/demon_fire_f1': 0.9193154034229829, 'eval/demon_fire_rate_model': 0.6952169076751946, 'eval/demon_fire_rate_teacher': 0.6696329254727474, 'eval/demon_left_precision': 0.8016922263352724, 'eval/demon_left_recall': 0.8589235127478754, 'eval/demon_left_f1': 0.8293216630196937, 'eval/demon_left_rate_model': 0.5258620689655172, 'eval/demon_left_rate_teacher': 0.49082313681868744, 'eval/demon_right_precision': 0.8510131108462455, 'eval/demon_right_recall': 0.8067796610169492, 'eval/demon_right_f1': 0.8283062645011601, 'eval/demon_right_rate_model': 0.46662958843159064, 'eval/demon_right_rate_teacher': 0.4922135706340378, 'eval/demon_move_f1': 0.8288139637604268, 'eval/demon_cc_f1': 0.8901555337126301, 'eval/latency_0/demon_fire_precision': 0.9024, 'eval/latency_0/demon_fire_recall': 0.9368770764119602, 'eval/latency_0/demon_fire_f1': 0.9193154034229829, 'eval/latency_0/demon_fire_rate_model': 0.6952169076751946, 'eval/latency_0/demon_fire_rate_teacher': 0.6696329254727474, 'eval/latency_0/demon_left_precision': 0.8016922263352724, 'eval/latency_0/demon_left_recall': 0.8589235127478754, 'eval/latency_0/demon_left_f1': 0.8293216630196937, 'eval/latency_0/demon_left_rate_model': 0.5258620689655172, 'eval/latency_0/demon_left_rate_teacher': 0.49082313681868744, 'eval/latency_0/demon_right_precision': 0.8510131108462455, 'eval/latency_0/demon_right_recall': 0.8067796610169492, 'eval/latency_0/demon_right_f1': 0.8283062645011601, 'eval/latency_0/demon_right_rate_model': 0.46662958843159064, 'eval/latency_0/demon_right_rate_teacher': 0.4922135706340378, 'eval/latency_0/demon_move_f1': 0.8288139637604268, 'eval/latency_0/demon_cc_f1': 0.8901555337126301, 'timing/data': 0.005665092030540109, 'timing/model': 1.2050676200306043, 'learning_rate/qwen_vl_interface': 2.1692350750519093e-06, 'learning_rate/action_model': 2.1692350750519092e-05, 'epoch': 1.5}) +[2026-06-17 15:48:57,691][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_5100_state +[2026-06-17 15:48:57,693][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 15:49:01,247][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_5100_state/model.safetensors +[2026-06-17 15:49:08,855][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_5100_state/optimizer.bin +[2026-06-17 15:49:08,856][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_5100_state/sampler.bin +[2026-06-17 15:49:08,856][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_5100_state/sampler_1.bin +[2026-06-17 15:49:08,856][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_5100_state/random_states_0.pkl +[2026-06-17 15:49:08,906][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 5100 +[2026-06-17 15:49:08,906][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 15:49:08,923][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 15:49:13,643][starVLA.training.train_starvla][INFO] - Step 5101, Loss: {'train/loss': 0.5997340679168701, 'train/grad_norm_pre_clip': 12.3825101852417, 'timing/data': 0.0053594199707731605, 'timing/model': 1.191822363063693, 'learning_rate/qwen_vl_interface': 2.167589273920869e-06, 'learning_rate/action_model': 2.167589273920869e-05, 'epoch': 1.5}) +[2026-06-17 15:49:18,357][starVLA.training.train_starvla][INFO] - Step 5102, Loss: {'train/loss': 0.9117291569709778, 'train/grad_norm_pre_clip': 13.92141342163086, 'timing/data': 0.004790017963387072, 'timing/model': 1.1897656560176983, 'learning_rate/qwen_vl_interface': 2.1659441117768888e-06, 'learning_rate/action_model': 2.1659441117768886e-05, 'epoch': 1.5}) +[2026-06-17 15:49:23,028][starVLA.training.train_starvla][INFO] - Step 5103, Loss: {'train/loss': 0.7212844491004944, 'train/grad_norm_pre_clip': 17.891489028930664, 'timing/data': 0.005110737984068692, 'timing/model': 1.1932185579789802, 'learning_rate/qwen_vl_interface': 2.164299588961013e-06, 'learning_rate/action_model': 2.1642995889610128e-05, 'epoch': 1.5}) +[2026-06-17 15:49:27,734][starVLA.training.train_starvla][INFO] - Step 5104, Loss: {'train/loss': 0.530040979385376, 'train/grad_norm_pre_clip': 12.200551986694336, 'timing/data': 0.00489945407025516, 'timing/model': 1.2835827660746872, 'learning_rate/qwen_vl_interface': 2.162655705814153e-06, 'learning_rate/action_model': 2.1626557058141527e-05, 'epoch': 1.5}) +[2026-06-17 15:49:32,356][starVLA.training.train_starvla][INFO] - Step 5105, Loss: {'train/loss': 0.6332008838653564, 'train/grad_norm_pre_clip': 11.327282905578613, 'timing/data': 0.005068772938102484, 'timing/model': 1.1963588719954714, 'learning_rate/qwen_vl_interface': 2.161012462677087e-06, 'learning_rate/action_model': 2.161012462677087e-05, 'epoch': 1.5}) +[2026-06-17 15:49:37,110][starVLA.training.train_starvla][INFO] - Step 5106, Loss: {'train/loss': 0.7188363075256348, 'train/grad_norm_pre_clip': 21.735063552856445, 'timing/data': 0.004919669008813798, 'timing/model': 1.1968085879925638, 'learning_rate/qwen_vl_interface': 2.1593698598904612e-06, 'learning_rate/action_model': 2.1593698598904614e-05, 'epoch': 1.5}) +[2026-06-17 15:49:41,874][starVLA.training.train_starvla][INFO] - Step 5107, Loss: {'train/loss': 0.5693981647491455, 'train/grad_norm_pre_clip': 14.215103149414062, 'timing/data': 0.005214715958572924, 'timing/model': 1.197458544978872, 'learning_rate/qwen_vl_interface': 2.1577278977947896e-06, 'learning_rate/action_model': 2.1577278977947897e-05, 'epoch': 1.5}) +[2026-06-17 15:49:46,670][starVLA.training.train_starvla][INFO] - Step 5108, Loss: {'train/loss': 0.552151620388031, 'train/grad_norm_pre_clip': 15.06928539276123, 'timing/data': 0.005346684018149972, 'timing/model': 1.1992017209995538, 'learning_rate/qwen_vl_interface': 2.1560865767304524e-06, 'learning_rate/action_model': 2.156086576730452e-05, 'epoch': 1.5}) +[2026-06-17 15:49:51,463][starVLA.training.train_starvla][INFO] - Step 5109, Loss: {'train/loss': 0.5569745898246765, 'train/grad_norm_pre_clip': 15.060723304748535, 'timing/data': 0.00496748392470181, 'timing/model': 1.3442089539021254, 'learning_rate/qwen_vl_interface': 2.154445897037696e-06, 'learning_rate/action_model': 2.154445897037696e-05, 'epoch': 1.5}) +[2026-06-17 15:49:56,108][starVLA.training.train_starvla][INFO] - Step 5110, Loss: {'train/loss': 0.6698598265647888, 'train/grad_norm_pre_clip': 13.788434028625488, 'timing/data': 0.004896958009339869, 'timing/model': 1.2031718289945275, 'learning_rate/qwen_vl_interface': 2.1528058590566353e-06, 'learning_rate/action_model': 2.1528058590566353e-05, 'epoch': 1.5}) +[2026-06-17 15:50:00,887][starVLA.training.train_starvla][INFO] - Step 5111, Loss: {'train/loss': 0.6271668076515198, 'train/grad_norm_pre_clip': 15.100045204162598, 'timing/data': 0.005180325941182673, 'timing/model': 1.2030381170334294, 'learning_rate/qwen_vl_interface': 2.1511664631272543e-06, 'learning_rate/action_model': 2.1511664631272545e-05, 'epoch': 1.5}) +[2026-06-17 15:50:05,673][starVLA.training.train_starvla][INFO] - Step 5112, Loss: {'train/loss': 0.5248917937278748, 'train/grad_norm_pre_clip': 13.162384033203125, 'timing/data': 0.005565212108194828, 'timing/model': 1.2057154540671036, 'learning_rate/qwen_vl_interface': 2.1495277095894015e-06, 'learning_rate/action_model': 2.1495277095894014e-05, 'epoch': 1.5}) +[2026-06-17 15:50:10,369][starVLA.training.train_starvla][INFO] - Step 5113, Loss: {'train/loss': 0.9802008867263794, 'train/grad_norm_pre_clip': 23.734018325805664, 'timing/data': 0.004977103089913726, 'timing/model': 1.2007295459043235, 'learning_rate/qwen_vl_interface': 2.147889598782787e-06, 'learning_rate/action_model': 2.147889598782787e-05, 'epoch': 1.5}) +[2026-06-17 15:50:15,107][starVLA.training.train_starvla][INFO] - Step 5114, Loss: {'train/loss': 0.7099383473396301, 'train/grad_norm_pre_clip': 13.109671592712402, 'timing/data': 0.005211998010054231, 'timing/model': 1.2028647179249674, 'learning_rate/qwen_vl_interface': 2.1462521310469987e-06, 'learning_rate/action_model': 2.1462521310469987e-05, 'epoch': 1.5}) +[2026-06-17 15:50:19,892][starVLA.training.train_starvla][INFO] - Step 5115, Loss: {'train/loss': 0.6098261475563049, 'train/grad_norm_pre_clip': 11.545828819274902, 'timing/data': 0.005152335972525179, 'timing/model': 1.3350486060371622, 'learning_rate/qwen_vl_interface': 2.144615306721484e-06, 'learning_rate/action_model': 2.144615306721484e-05, 'epoch': 1.51}) +[2026-06-17 15:50:24,554][starVLA.training.train_starvla][INFO] - Step 5116, Loss: {'train/loss': 0.6646489500999451, 'train/grad_norm_pre_clip': 12.670114517211914, 'timing/data': 0.005294175003655255, 'timing/model': 1.2059552969876677, 'learning_rate/qwen_vl_interface': 2.1429791261455575e-06, 'learning_rate/action_model': 2.1429791261455572e-05, 'epoch': 1.51}) +[2026-06-17 15:50:29,334][starVLA.training.train_starvla][INFO] - Step 5117, Loss: {'train/loss': 0.8304109573364258, 'train/grad_norm_pre_clip': 14.928187370300293, 'timing/data': 0.0051377080380916595, 'timing/model': 1.2033828620333225, 'learning_rate/qwen_vl_interface': 2.1413435896584006e-06, 'learning_rate/action_model': 2.1413435896584005e-05, 'epoch': 1.51}) +[2026-06-17 15:50:34,113][starVLA.training.train_starvla][INFO] - Step 5118, Loss: {'train/loss': 0.8012233972549438, 'train/grad_norm_pre_clip': 16.13010025024414, 'timing/data': 0.005296568037010729, 'timing/model': 1.204175482969731, 'learning_rate/qwen_vl_interface': 2.1397086975990633e-06, 'learning_rate/action_model': 2.139708697599063e-05, 'epoch': 1.51}) +[2026-06-17 15:50:38,881][starVLA.training.train_starvla][INFO] - Step 5119, Loss: {'train/loss': 0.6357875466346741, 'train/grad_norm_pre_clip': 12.175528526306152, 'timing/data': 0.005158102954737842, 'timing/model': 1.2018257839372382, 'learning_rate/qwen_vl_interface': 2.1380744503064597e-06, 'learning_rate/action_model': 2.1380744503064598e-05, 'epoch': 1.51}) +[2026-06-17 15:50:43,646][starVLA.training.train_starvla][INFO] - Step 5120, Loss: {'train/loss': 0.5925101041793823, 'train/grad_norm_pre_clip': 10.56644344329834, 'timing/data': 0.0051642999751493335, 'timing/model': 1.2042091330513358, 'learning_rate/qwen_vl_interface': 2.1364408481193714e-06, 'learning_rate/action_model': 2.1364408481193712e-05, 'epoch': 1.51}) +[2026-06-17 15:50:48,414][starVLA.training.train_starvla][INFO] - Step 5121, Loss: {'train/loss': 0.8537255525588989, 'train/grad_norm_pre_clip': 57.39339828491211, 'timing/data': 0.004926942987367511, 'timing/model': 1.316531972028315, 'learning_rate/qwen_vl_interface': 2.1348078913764454e-06, 'learning_rate/action_model': 2.1348078913764452e-05, 'epoch': 1.51}) +[2026-06-17 15:50:53,063][starVLA.training.train_starvla][INFO] - Step 5122, Loss: {'train/loss': 0.7041022181510925, 'train/grad_norm_pre_clip': 13.515012741088867, 'timing/data': 0.005362133029848337, 'timing/model': 1.2029889289988205, 'learning_rate/qwen_vl_interface': 2.1331755804161957e-06, 'learning_rate/action_model': 2.133175580416196e-05, 'epoch': 1.51}) +[2026-06-17 15:50:57,847][starVLA.training.train_starvla][INFO] - Step 5123, Loss: {'train/loss': 0.6523262858390808, 'train/grad_norm_pre_clip': 15.585830688476562, 'timing/data': 0.005368157988414168, 'timing/model': 1.2038986909901723, 'learning_rate/qwen_vl_interface': 2.131543915577002e-06, 'learning_rate/action_model': 2.131543915577002e-05, 'epoch': 1.51}) +[2026-06-17 15:51:02,692][starVLA.training.train_starvla][INFO] - Step 5124, Loss: {'train/loss': 0.38420525193214417, 'train/grad_norm_pre_clip': 80.12718200683594, 'timing/data': 0.005398905021138489, 'timing/model': 1.2044502809876576, 'learning_rate/qwen_vl_interface': 2.129912897197108e-06, 'learning_rate/action_model': 2.1299128971971084e-05, 'epoch': 1.51}) +[2026-06-17 15:51:07,487][starVLA.training.train_starvla][INFO] - Step 5125, Loss: {'train/loss': 0.8381894826889038, 'train/grad_norm_pre_clip': 16.202951431274414, 'timing/data': 0.005500635947100818, 'timing/model': 1.208826464950107, 'learning_rate/qwen_vl_interface': 2.1282825256146305e-06, 'learning_rate/action_model': 2.12828252561463e-05, 'epoch': 1.51}) +[2026-06-17 15:51:12,289][starVLA.training.train_starvla][INFO] - Step 5126, Loss: {'train/loss': 0.4754641354084015, 'train/grad_norm_pre_clip': 9.96407699584961, 'timing/data': 0.005097760003991425, 'timing/model': 1.345733963069506, 'learning_rate/qwen_vl_interface': 2.1266528011675453e-06, 'learning_rate/action_model': 2.126652801167545e-05, 'epoch': 1.51}) +[2026-06-17 15:51:16,967][starVLA.training.train_starvla][INFO] - Step 5127, Loss: {'train/loss': 0.4445539712905884, 'train/grad_norm_pre_clip': 11.856224060058594, 'timing/data': 0.005602699937298894, 'timing/model': 1.2132911790395156, 'learning_rate/qwen_vl_interface': 2.125023724193694e-06, 'learning_rate/action_model': 2.1250237241936936e-05, 'epoch': 1.51}) +[2026-06-17 15:51:21,774][starVLA.training.train_starvla][INFO] - Step 5128, Loss: {'train/loss': 0.6256726980209351, 'train/grad_norm_pre_clip': 31.580034255981445, 'timing/data': 0.005747376009821892, 'timing/model': 1.2092834610957652, 'learning_rate/qwen_vl_interface': 2.1233952950307857e-06, 'learning_rate/action_model': 2.1233952950307855e-05, 'epoch': 1.51}) +[2026-06-17 15:51:26,527][starVLA.training.train_starvla][INFO] - Step 5129, Loss: {'train/loss': 0.7246062755584717, 'train/grad_norm_pre_clip': 14.065511703491211, 'timing/data': 0.005325242062099278, 'timing/model': 1.208174719940871, 'learning_rate/qwen_vl_interface': 2.1217675140163984e-06, 'learning_rate/action_model': 2.1217675140163983e-05, 'epoch': 1.51}) +[2026-06-17 15:51:31,301][starVLA.training.train_starvla][INFO] - Step 5130, Loss: {'train/loss': 0.6623528599739075, 'train/grad_norm_pre_clip': 11.405933380126953, 'timing/data': 0.005415740073658526, 'timing/model': 1.209408638998866, 'learning_rate/qwen_vl_interface': 2.1201403814879715e-06, 'learning_rate/action_model': 2.1201403814879717e-05, 'epoch': 1.51}) +[2026-06-17 15:51:36,133][starVLA.training.train_starvla][INFO] - Step 5131, Loss: {'train/loss': 0.7284667491912842, 'train/grad_norm_pre_clip': 18.10166358947754, 'timing/data': 0.005414690938778222, 'timing/model': 1.2183213499374688, 'learning_rate/qwen_vl_interface': 2.1185138977828122e-06, 'learning_rate/action_model': 2.118513897782812e-05, 'epoch': 1.51}) +[2026-06-17 15:51:40,927][starVLA.training.train_starvla][INFO] - Step 5132, Loss: {'train/loss': 0.5629407167434692, 'train/grad_norm_pre_clip': 12.124710083007812, 'timing/data': 0.005339159048162401, 'timing/model': 1.3330069719813764, 'learning_rate/qwen_vl_interface': 2.116888063238091e-06, 'learning_rate/action_model': 2.116888063238091e-05, 'epoch': 1.51}) +[2026-06-17 15:51:45,596][starVLA.training.train_starvla][INFO] - Step 5133, Loss: {'train/loss': 0.8175023794174194, 'train/grad_norm_pre_clip': 15.220979690551758, 'timing/data': 0.005239546997472644, 'timing/model': 1.2048191719222814, 'learning_rate/qwen_vl_interface': 2.1152628781908455e-06, 'learning_rate/action_model': 2.1152628781908455e-05, 'epoch': 1.51}) +[2026-06-17 15:51:50,406][starVLA.training.train_starvla][INFO] - Step 5134, Loss: {'train/loss': 0.4512981176376343, 'train/grad_norm_pre_clip': 11.545601844787598, 'timing/data': 0.005300048040226102, 'timing/model': 1.204598237061873, 'learning_rate/qwen_vl_interface': 2.1136383429779787e-06, 'learning_rate/action_model': 2.1136383429779788e-05, 'epoch': 1.51}) +[2026-06-17 15:51:55,183][starVLA.training.train_starvla][INFO] - Step 5135, Loss: {'train/loss': 0.44774213433265686, 'train/grad_norm_pre_clip': 11.358658790588379, 'timing/data': 0.005071958061307669, 'timing/model': 1.2062826739856973, 'learning_rate/qwen_vl_interface': 2.1120144579362582e-06, 'learning_rate/action_model': 2.1120144579362582e-05, 'epoch': 1.51}) +[2026-06-17 15:51:59,934][starVLA.training.train_starvla][INFO] - Step 5136, Loss: {'train/loss': 0.766923189163208, 'train/grad_norm_pre_clip': 15.269566535949707, 'timing/data': 0.004961605998687446, 'timing/model': 1.206959028961137, 'learning_rate/qwen_vl_interface': 2.1103912234023175e-06, 'learning_rate/action_model': 2.1103912234023175e-05, 'epoch': 1.51}) +[2026-06-17 15:52:04,708][starVLA.training.train_starvla][INFO] - Step 5137, Loss: {'train/loss': 0.8525693416595459, 'train/grad_norm_pre_clip': 17.463090896606445, 'timing/data': 0.00531490205321461, 'timing/model': 1.2051394889131188, 'learning_rate/qwen_vl_interface': 2.1087686397126544e-06, 'learning_rate/action_model': 2.1087686397126545e-05, 'epoch': 1.51}) +[2026-06-17 15:52:09,496][starVLA.training.train_starvla][INFO] - Step 5138, Loss: {'train/loss': 0.48985016345977783, 'train/grad_norm_pre_clip': 18.48257827758789, 'timing/data': 0.004812453989870846, 'timing/model': 1.3358797109685838, 'learning_rate/qwen_vl_interface': 2.107146707203632e-06, 'learning_rate/action_model': 2.1071467072036323e-05, 'epoch': 1.51}) +[2026-06-17 15:52:14,157][starVLA.training.train_starvla][INFO] - Step 5139, Loss: {'train/loss': 0.5740167498588562, 'train/grad_norm_pre_clip': 12.124566078186035, 'timing/data': 0.004774478962644935, 'timing/model': 1.2088302329648286, 'learning_rate/qwen_vl_interface': 2.105525426211478e-06, 'learning_rate/action_model': 2.1055254262114778e-05, 'epoch': 1.51}) +[2026-06-17 15:52:18,985][starVLA.training.train_starvla][INFO] - Step 5140, Loss: {'train/loss': 0.49334716796875, 'train/grad_norm_pre_clip': 16.180782318115234, 'timing/data': 0.004910530988126993, 'timing/model': 1.2103178519755602, 'learning_rate/qwen_vl_interface': 2.1039047970722894e-06, 'learning_rate/action_model': 2.1039047970722893e-05, 'epoch': 1.51}) +[2026-06-17 15:52:23,823][starVLA.training.train_starvla][INFO] - Step 5141, Loss: {'train/loss': 0.5722485184669495, 'train/grad_norm_pre_clip': 12.09856128692627, 'timing/data': 0.00561191001906991, 'timing/model': 1.20982039894443, 'learning_rate/qwen_vl_interface': 2.1022848201220204e-06, 'learning_rate/action_model': 2.1022848201220204e-05, 'epoch': 1.51}) +[2026-06-17 15:52:28,648][starVLA.training.train_starvla][INFO] - Step 5142, Loss: {'train/loss': 0.6245348453521729, 'train/grad_norm_pre_clip': 13.100088119506836, 'timing/data': 0.005099826026707888, 'timing/model': 1.210265157977119, 'learning_rate/qwen_vl_interface': 2.1006654956964935e-06, 'learning_rate/action_model': 2.1006654956964936e-05, 'epoch': 1.51}) +[2026-06-17 15:52:33,457][starVLA.training.train_starvla][INFO] - Step 5143, Loss: {'train/loss': 0.46634310483932495, 'train/grad_norm_pre_clip': 11.391797065734863, 'timing/data': 0.005136036081239581, 'timing/model': 1.3383756129769608, 'learning_rate/qwen_vl_interface': 2.0990468241314004e-06, 'learning_rate/action_model': 2.0990468241314002e-05, 'epoch': 1.51}) +[2026-06-17 15:52:38,138][starVLA.training.train_starvla][INFO] - Step 5144, Loss: {'train/loss': 0.5407341122627258, 'train/grad_norm_pre_clip': 10.967634201049805, 'timing/data': 0.00505251600407064, 'timing/model': 1.210441022994928, 'learning_rate/qwen_vl_interface': 2.0974288057622905e-06, 'learning_rate/action_model': 2.0974288057622903e-05, 'epoch': 1.51}) +[2026-06-17 15:52:42,932][starVLA.training.train_starvla][INFO] - Step 5145, Loss: {'train/loss': 0.7897471189498901, 'train/grad_norm_pre_clip': 16.136999130249023, 'timing/data': 0.00508898904081434, 'timing/model': 1.2059755370719358, 'learning_rate/qwen_vl_interface': 2.095811440924584e-06, 'learning_rate/action_model': 2.0958114409245835e-05, 'epoch': 1.51}) +[2026-06-17 15:52:47,670][starVLA.training.train_starvla][INFO] - Step 5146, Loss: {'train/loss': 0.5347680449485779, 'train/grad_norm_pre_clip': 11.225407600402832, 'timing/data': 0.005066939047537744, 'timing/model': 1.2101046630414203, 'learning_rate/qwen_vl_interface': 2.0941947299535554e-06, 'learning_rate/action_model': 2.0941947299535552e-05, 'epoch': 1.51}) +[2026-06-17 15:52:52,461][starVLA.training.train_starvla][INFO] - Step 5147, Loss: {'train/loss': 0.6297388672828674, 'train/grad_norm_pre_clip': 13.56032657623291, 'timing/data': 0.005199349019676447, 'timing/model': 1.2047089230036363, 'learning_rate/qwen_vl_interface': 2.0925786731843573e-06, 'learning_rate/action_model': 2.0925786731843572e-05, 'epoch': 1.51}) +[2026-06-17 15:52:57,258][starVLA.training.train_starvla][INFO] - Step 5148, Loss: {'train/loss': 0.8146513104438782, 'train/grad_norm_pre_clip': 13.384649276733398, 'timing/data': 0.00505913607776165, 'timing/model': 1.2039741829503328, 'learning_rate/qwen_vl_interface': 2.0909632709519976e-06, 'learning_rate/action_model': 2.090963270951998e-05, 'epoch': 1.51}) +[2026-06-17 15:53:02,180][starVLA.training.train_starvla][INFO] - Step 5149, Loss: {'train/loss': 0.6802222728729248, 'train/grad_norm_pre_clip': 14.635591506958008, 'timing/data': 0.007456635939888656, 'timing/model': 1.3510894689243287, 'learning_rate/qwen_vl_interface': 2.0893485235913514e-06, 'learning_rate/action_model': 2.0893485235913514e-05, 'epoch': 1.52}) +[2026-06-17 15:53:06,861][starVLA.training.train_starvla][INFO] - Step 5150, Loss: {'train/loss': 0.7409886717796326, 'train/grad_norm_pre_clip': 17.274059295654297, 'timing/data': 0.00491288909688592, 'timing/model': 1.2077539230231196, 'learning_rate/qwen_vl_interface': 2.0877344314371575e-06, 'learning_rate/action_model': 2.0877344314371573e-05, 'epoch': 1.52}) +[2026-06-17 15:53:11,690][starVLA.training.train_starvla][INFO] - Step 5151, Loss: {'train/loss': 0.9285710453987122, 'train/grad_norm_pre_clip': 14.278592109680176, 'timing/data': 0.0049712309846654534, 'timing/model': 1.211318660993129, 'learning_rate/qwen_vl_interface': 2.086120994824018e-06, 'learning_rate/action_model': 2.0861209948240178e-05, 'epoch': 1.52}) +[2026-06-17 15:53:16,423][starVLA.training.train_starvla][INFO] - Step 5152, Loss: {'train/loss': 0.979347288608551, 'train/grad_norm_pre_clip': 19.09596824645996, 'timing/data': 0.004712164984084666, 'timing/model': 1.2026310950750485, 'learning_rate/qwen_vl_interface': 2.0845082140864012e-06, 'learning_rate/action_model': 2.084508214086401e-05, 'epoch': 1.52}) +[2026-06-17 15:53:21,213][starVLA.training.train_starvla][INFO] - Step 5153, Loss: {'train/loss': 0.500002384185791, 'train/grad_norm_pre_clip': 11.566938400268555, 'timing/data': 0.00535119092091918, 'timing/model': 1.2048175419913605, 'learning_rate/qwen_vl_interface': 2.0828960895586365e-06, 'learning_rate/action_model': 2.0828960895586364e-05, 'epoch': 1.52}) +[2026-06-17 15:53:26,005][starVLA.training.train_starvla][INFO] - Step 5154, Loss: {'train/loss': 0.63248610496521, 'train/grad_norm_pre_clip': 14.975845336914062, 'timing/data': 0.005597612005658448, 'timing/model': 1.2051044330000877, 'learning_rate/qwen_vl_interface': 2.0812846215749227e-06, 'learning_rate/action_model': 2.0812846215749227e-05, 'epoch': 1.52}) +[2026-06-17 15:53:30,794][starVLA.training.train_starvla][INFO] - Step 5155, Loss: {'train/loss': 0.5847175121307373, 'train/grad_norm_pre_clip': 11.717650413513184, 'timing/data': 0.005551871960051358, 'timing/model': 1.3369203630136326, 'learning_rate/qwen_vl_interface': 2.079673810469316e-06, 'learning_rate/action_model': 2.0796738104693158e-05, 'epoch': 1.52}) +[2026-06-17 15:53:35,494][starVLA.training.train_starvla][INFO] - Step 5156, Loss: {'train/loss': 0.8534104228019714, 'train/grad_norm_pre_clip': 14.401765823364258, 'timing/data': 0.005209550028666854, 'timing/model': 1.2506294519407675, 'learning_rate/qwen_vl_interface': 2.0780636565757392e-06, 'learning_rate/action_model': 2.0780636565757392e-05, 'epoch': 1.52}) +[2026-06-17 15:53:40,153][starVLA.training.train_starvla][INFO] - Step 5157, Loss: {'train/loss': 0.880363941192627, 'train/grad_norm_pre_clip': 20.060579299926758, 'timing/data': 0.004930410999804735, 'timing/model': 1.2048242948949337, 'learning_rate/qwen_vl_interface': 2.0764541602279777e-06, 'learning_rate/action_model': 2.0764541602279778e-05, 'epoch': 1.52}) +[2026-06-17 15:53:44,919][starVLA.training.train_starvla][INFO] - Step 5158, Loss: {'train/loss': 0.6670957207679749, 'train/grad_norm_pre_clip': 15.297990798950195, 'timing/data': 0.0050217920215800405, 'timing/model': 1.2123364170547575, 'learning_rate/qwen_vl_interface': 2.0748453217596856e-06, 'learning_rate/action_model': 2.0748453217596856e-05, 'epoch': 1.52}) +[2026-06-17 15:53:49,741][starVLA.training.train_starvla][INFO] - Step 5159, Loss: {'train/loss': 0.491372287273407, 'train/grad_norm_pre_clip': 11.13363265991211, 'timing/data': 0.005107130971737206, 'timing/model': 1.217243082006462, 'learning_rate/qwen_vl_interface': 2.0732371415043766e-06, 'learning_rate/action_model': 2.0732371415043765e-05, 'epoch': 1.52}) +[2026-06-17 15:53:54,669][starVLA.training.train_starvla][INFO] - Step 5160, Loss: {'train/loss': 0.6088200807571411, 'train/grad_norm_pre_clip': 15.678961753845215, 'timing/data': 0.005055975983850658, 'timing/model': 1.221755906008184, 'learning_rate/qwen_vl_interface': 2.071629619795425e-06, 'learning_rate/action_model': 2.071629619795425e-05, 'epoch': 1.52}) +[2026-06-17 15:53:59,474][starVLA.training.train_starvla][INFO] - Step 5161, Loss: {'train/loss': 1.0702067613601685, 'train/grad_norm_pre_clip': 19.085500717163086, 'timing/data': 0.005238940007984638, 'timing/model': 1.343801288981922, 'learning_rate/qwen_vl_interface': 2.070022756966071e-06, 'learning_rate/action_model': 2.0700227569660712e-05, 'epoch': 1.52}) +[2026-06-17 15:54:04,275][starVLA.training.train_starvla][INFO] - Step 5162, Loss: {'train/loss': 0.7527047991752625, 'train/grad_norm_pre_clip': 13.939203262329102, 'timing/data': 0.00495801295619458, 'timing/model': 1.3309526550583541, 'learning_rate/qwen_vl_interface': 2.068416553349424e-06, 'learning_rate/action_model': 2.068416553349424e-05, 'epoch': 1.52}) +[2026-06-17 15:54:08,955][starVLA.training.train_starvla][INFO] - Step 5163, Loss: {'train/loss': 0.4849710166454315, 'train/grad_norm_pre_clip': 13.009441375732422, 'timing/data': 0.00492756103631109, 'timing/model': 1.2079007138963789, 'learning_rate/qwen_vl_interface': 2.0668110092784484e-06, 'learning_rate/action_model': 2.0668110092784484e-05, 'epoch': 1.52}) +[2026-06-17 15:54:13,717][starVLA.training.train_starvla][INFO] - Step 5164, Loss: {'train/loss': 0.5482203960418701, 'train/grad_norm_pre_clip': 10.839150428771973, 'timing/data': 0.005599680007435381, 'timing/model': 1.2070838760118932, 'learning_rate/qwen_vl_interface': 2.0652061250859758e-06, 'learning_rate/action_model': 2.0652061250859756e-05, 'epoch': 1.52}) +[2026-06-17 15:54:18,503][starVLA.training.train_starvla][INFO] - Step 5165, Loss: {'train/loss': 0.6548642516136169, 'train/grad_norm_pre_clip': 12.424600601196289, 'timing/data': 0.005118750035762787, 'timing/model': 1.2094598980620503, 'learning_rate/qwen_vl_interface': 2.0636019011047005e-06, 'learning_rate/action_model': 2.0636019011047003e-05, 'epoch': 1.52}) +[2026-06-17 15:54:23,347][starVLA.training.train_starvla][INFO] - Step 5166, Loss: {'train/loss': 0.7905264496803284, 'train/grad_norm_pre_clip': 14.116296768188477, 'timing/data': 0.004690526984632015, 'timing/model': 1.2084577759960666, 'learning_rate/qwen_vl_interface': 2.0619983376671796e-06, 'learning_rate/action_model': 2.0619983376671793e-05, 'epoch': 1.52}) +[2026-06-17 15:54:28,082][starVLA.training.train_starvla][INFO] - Step 5167, Loss: {'train/loss': 0.6397324204444885, 'train/grad_norm_pre_clip': 12.012681007385254, 'timing/data': 0.005051783053204417, 'timing/model': 1.2741974839009345, 'learning_rate/qwen_vl_interface': 2.060395435105834e-06, 'learning_rate/action_model': 2.060395435105834e-05, 'epoch': 1.52}) +[2026-06-17 15:54:32,769][starVLA.training.train_starvla][INFO] - Step 5168, Loss: {'train/loss': 0.7140367031097412, 'train/grad_norm_pre_clip': 13.520169258117676, 'timing/data': 0.005300778895616531, 'timing/model': 1.211232598987408, 'learning_rate/qwen_vl_interface': 2.0587931937529465e-06, 'learning_rate/action_model': 2.0587931937529464e-05, 'epoch': 1.52}) +[2026-06-17 15:54:37,553][starVLA.training.train_starvla][INFO] - Step 5169, Loss: {'train/loss': 0.8176907896995544, 'train/grad_norm_pre_clip': 20.820749282836914, 'timing/data': 0.005046867998316884, 'timing/model': 1.220146170933731, 'learning_rate/qwen_vl_interface': 2.0571916139406648e-06, 'learning_rate/action_model': 2.0571916139406647e-05, 'epoch': 1.52}) +[2026-06-17 15:54:42,374][starVLA.training.train_starvla][INFO] - Step 5170, Loss: {'train/loss': 0.6018339991569519, 'train/grad_norm_pre_clip': 12.061007499694824, 'timing/data': 0.005264467909000814, 'timing/model': 1.2111422260059044, 'learning_rate/qwen_vl_interface': 2.055590696000996e-06, 'learning_rate/action_model': 2.0555906960009962e-05, 'epoch': 1.52}) +[2026-06-17 15:54:47,183][starVLA.training.train_starvla][INFO] - Step 5171, Loss: {'train/loss': 0.6620938181877136, 'train/grad_norm_pre_clip': 14.570372581481934, 'timing/data': 0.005361496936529875, 'timing/model': 1.2120800509583205, 'learning_rate/qwen_vl_interface': 2.0539904402658144e-06, 'learning_rate/action_model': 2.0539904402658142e-05, 'epoch': 1.52}) +[2026-06-17 15:54:52,002][starVLA.training.train_starvla][INFO] - Step 5172, Loss: {'train/loss': 0.7700256109237671, 'train/grad_norm_pre_clip': 13.15888786315918, 'timing/data': 0.0051677749725058675, 'timing/model': 1.3458978059934452, 'learning_rate/qwen_vl_interface': 2.052390847066851e-06, 'learning_rate/action_model': 2.052390847066851e-05, 'epoch': 1.52}) +[2026-06-17 15:54:56,755][starVLA.training.train_starvla][INFO] - Step 5173, Loss: {'train/loss': 0.6505247354507446, 'train/grad_norm_pre_clip': 12.461282730102539, 'timing/data': 0.0051789849530905485, 'timing/model': 1.2894932590425014, 'learning_rate/qwen_vl_interface': 2.0507919167357093e-06, 'learning_rate/action_model': 2.050791916735709e-05, 'epoch': 1.52}) +[2026-06-17 15:55:01,416][starVLA.training.train_starvla][INFO] - Step 5174, Loss: {'train/loss': 0.5951915383338928, 'train/grad_norm_pre_clip': 11.731197357177734, 'timing/data': 0.005162011017091572, 'timing/model': 1.2081887910608202, 'learning_rate/qwen_vl_interface': 2.049193649603844e-06, 'learning_rate/action_model': 2.049193649603844e-05, 'epoch': 1.52}) +[2026-06-17 15:55:06,262][starVLA.training.train_starvla][INFO] - Step 5175, Loss: {'train/loss': 0.6296754479408264, 'train/grad_norm_pre_clip': 13.692994117736816, 'timing/data': 0.004789829952642322, 'timing/model': 1.2050569140119478, 'learning_rate/qwen_vl_interface': 2.047596046002578e-06, 'learning_rate/action_model': 2.0475960460025778e-05, 'epoch': 1.52}) +[2026-06-17 15:55:11,050][starVLA.training.train_starvla][INFO] - Step 5176, Loss: {'train/loss': 0.43965116143226624, 'train/grad_norm_pre_clip': 12.796603202819824, 'timing/data': 0.0050090940203517675, 'timing/model': 1.207580866990611, 'learning_rate/qwen_vl_interface': 2.0459991062630994e-06, 'learning_rate/action_model': 2.045999106263099e-05, 'epoch': 1.52}) +[2026-06-17 15:55:15,848][starVLA.training.train_starvla][INFO] - Step 5177, Loss: {'train/loss': 0.6565147638320923, 'train/grad_norm_pre_clip': 13.113896369934082, 'timing/data': 0.00502675399184227, 'timing/model': 1.2037212660070509, 'learning_rate/qwen_vl_interface': 2.0444028307164532e-06, 'learning_rate/action_model': 2.0444028307164534e-05, 'epoch': 1.52}) +[2026-06-17 15:55:20,640][starVLA.training.train_starvla][INFO] - Step 5178, Loss: {'train/loss': 0.6961423754692078, 'train/grad_norm_pre_clip': 12.59218978881836, 'timing/data': 0.0052954599959775805, 'timing/model': 1.2072312249802053, 'learning_rate/qwen_vl_interface': 2.0428072196935516e-06, 'learning_rate/action_model': 2.0428072196935514e-05, 'epoch': 1.52}) +[2026-06-17 15:55:25,411][starVLA.training.train_starvla][INFO] - Step 5179, Loss: {'train/loss': 0.5875638127326965, 'train/grad_norm_pre_clip': 12.270574569702148, 'timing/data': 0.005443420028313994, 'timing/model': 1.300208768923767, 'learning_rate/qwen_vl_interface': 2.041212273525161e-06, 'learning_rate/action_model': 2.041212273525161e-05, 'epoch': 1.52}) +[2026-06-17 15:55:30,094][starVLA.training.train_starvla][INFO] - Step 5180, Loss: {'train/loss': 0.6881393790245056, 'train/grad_norm_pre_clip': 14.681901931762695, 'timing/data': 0.005445256945677102, 'timing/model': 1.2195674640825018, 'learning_rate/qwen_vl_interface': 2.03961799254192e-06, 'learning_rate/action_model': 2.03961799254192e-05, 'epoch': 1.52}) +[2026-06-17 15:55:34,880][starVLA.training.train_starvla][INFO] - Step 5181, Loss: {'train/loss': 0.5869065523147583, 'train/grad_norm_pre_clip': 13.572059631347656, 'timing/data': 0.0050086360424757, 'timing/model': 1.205243862932548, 'learning_rate/qwen_vl_interface': 2.0380243770743246e-06, 'learning_rate/action_model': 2.0380243770743247e-05, 'epoch': 1.52}) +[2026-06-17 15:55:39,674][starVLA.training.train_starvla][INFO] - Step 5182, Loss: {'train/loss': 0.6915827393531799, 'train/grad_norm_pre_clip': 12.588777542114258, 'timing/data': 0.005152167985215783, 'timing/model': 1.205958597944118, 'learning_rate/qwen_vl_interface': 2.0364314274527318e-06, 'learning_rate/action_model': 2.0364314274527314e-05, 'epoch': 1.52}) +[2026-06-17 15:55:44,492][starVLA.training.train_starvla][INFO] - Step 5183, Loss: {'train/loss': 0.6711666584014893, 'train/grad_norm_pre_clip': 11.863615036010742, 'timing/data': 0.0049514619167894125, 'timing/model': 1.2152845360105857, 'learning_rate/qwen_vl_interface': 2.034839144007362e-06, 'learning_rate/action_model': 2.034839144007362e-05, 'epoch': 1.53}) +[2026-06-17 15:55:49,297][starVLA.training.train_starvla][INFO] - Step 5184, Loss: {'train/loss': 0.6348803639411926, 'train/grad_norm_pre_clip': 15.67944622039795, 'timing/data': 0.005099114030599594, 'timing/model': 1.3467344550881535, 'learning_rate/qwen_vl_interface': 2.033247527068297e-06, 'learning_rate/action_model': 2.0332475270682972e-05, 'epoch': 1.53}) +[2026-06-17 15:55:53,965][starVLA.training.train_starvla][INFO] - Step 5185, Loss: {'train/loss': 0.5863070487976074, 'train/grad_norm_pre_clip': 14.944754600524902, 'timing/data': 0.0060989499324932694, 'timing/model': 1.2091132879722863, 'learning_rate/qwen_vl_interface': 2.0316565769654816e-06, 'learning_rate/action_model': 2.0316565769654816e-05, 'epoch': 1.53}) +[2026-06-17 15:55:58,759][starVLA.training.train_starvla][INFO] - Step 5186, Loss: {'train/loss': 0.6778128743171692, 'train/grad_norm_pre_clip': 15.482329368591309, 'timing/data': 0.005056528956629336, 'timing/model': 1.2053941800259054, 'learning_rate/qwen_vl_interface': 2.0300662940287193e-06, 'learning_rate/action_model': 2.0300662940287192e-05, 'epoch': 1.53}) +[2026-06-17 15:56:03,610][starVLA.training.train_starvla][INFO] - Step 5187, Loss: {'train/loss': 0.5175380706787109, 'train/grad_norm_pre_clip': 11.308652877807617, 'timing/data': 0.005165803013369441, 'timing/model': 1.2052007960155606, 'learning_rate/qwen_vl_interface': 2.0284766785876812e-06, 'learning_rate/action_model': 2.0284766785876812e-05, 'epoch': 1.53}) +[2026-06-17 15:56:08,328][starVLA.training.train_starvla][INFO] - Step 5188, Loss: {'train/loss': 0.997967004776001, 'train/grad_norm_pre_clip': 17.570772171020508, 'timing/data': 0.005193216959014535, 'timing/model': 1.2091171779902652, 'learning_rate/qwen_vl_interface': 2.026887730971894e-06, 'learning_rate/action_model': 2.0268877309718938e-05, 'epoch': 1.53}) +[2026-06-17 15:56:13,064][starVLA.training.train_starvla][INFO] - Step 5189, Loss: {'train/loss': 0.5177109837532043, 'train/grad_norm_pre_clip': 13.313745498657227, 'timing/data': 0.005470759002491832, 'timing/model': 1.2031793289352208, 'learning_rate/qwen_vl_interface': 2.0252994515107475e-06, 'learning_rate/action_model': 2.0252994515107472e-05, 'epoch': 1.53}) +[2026-06-17 15:56:17,863][starVLA.training.train_starvla][INFO] - Step 5190, Loss: {'train/loss': 0.41549283266067505, 'train/grad_norm_pre_clip': 9.671823501586914, 'timing/data': 0.005240346072241664, 'timing/model': 1.347336392966099, 'learning_rate/qwen_vl_interface': 2.023711840533493e-06, 'learning_rate/action_model': 2.023711840533493e-05, 'epoch': 1.53}) +[2026-06-17 15:56:22,541][starVLA.training.train_starvla][INFO] - Step 5191, Loss: {'train/loss': 0.685589611530304, 'train/grad_norm_pre_clip': 15.649435997009277, 'timing/data': 0.005581169039942324, 'timing/model': 1.2063772820401937, 'learning_rate/qwen_vl_interface': 2.0221248983692475e-06, 'learning_rate/action_model': 2.0221248983692474e-05, 'epoch': 1.53}) +[2026-06-17 15:56:27,353][starVLA.training.train_starvla][INFO] - Step 5192, Loss: {'train/loss': 0.6392146348953247, 'train/grad_norm_pre_clip': 15.361774444580078, 'timing/data': 0.004982684971764684, 'timing/model': 1.206285682041198, 'learning_rate/qwen_vl_interface': 2.020538625346984e-06, 'learning_rate/action_model': 2.020538625346984e-05, 'epoch': 1.53}) +[2026-06-17 15:56:32,139][starVLA.training.train_starvla][INFO] - Step 5193, Loss: {'train/loss': 0.7061948776245117, 'train/grad_norm_pre_clip': 16.257713317871094, 'timing/data': 0.005076623056083918, 'timing/model': 1.2066396380541846, 'learning_rate/qwen_vl_interface': 2.0189530217955382e-06, 'learning_rate/action_model': 2.018953021795538e-05, 'epoch': 1.53}) +[2026-06-17 15:56:36,918][starVLA.training.train_starvla][INFO] - Step 5194, Loss: {'train/loss': 0.39811500906944275, 'train/grad_norm_pre_clip': 10.84567642211914, 'timing/data': 0.004905279027298093, 'timing/model': 1.204083436052315, 'learning_rate/qwen_vl_interface': 2.0173680880436076e-06, 'learning_rate/action_model': 2.0173680880436076e-05, 'epoch': 1.53}) +[2026-06-17 15:56:41,667][starVLA.training.train_starvla][INFO] - Step 5195, Loss: {'train/loss': 0.621549129486084, 'train/grad_norm_pre_clip': 13.196575164794922, 'timing/data': 0.0052748339949175715, 'timing/model': 1.2047076660674065, 'learning_rate/qwen_vl_interface': 2.015783824419751e-06, 'learning_rate/action_model': 2.0157838244197508e-05, 'epoch': 1.53}) +[2026-06-17 15:56:46,452][starVLA.training.train_starvla][INFO] - Step 5196, Loss: {'train/loss': 0.5199081897735596, 'train/grad_norm_pre_clip': 10.733369827270508, 'timing/data': 0.005274230032227933, 'timing/model': 1.3248729510232806, 'learning_rate/qwen_vl_interface': 2.0142002312523877e-06, 'learning_rate/action_model': 2.014200231252388e-05, 'epoch': 1.53}) +[2026-06-17 15:56:51,130][starVLA.training.train_starvla][INFO] - Step 5197, Loss: {'train/loss': 0.5081521272659302, 'train/grad_norm_pre_clip': 10.905935287475586, 'timing/data': 0.004774381988681853, 'timing/model': 1.213322779047303, 'learning_rate/qwen_vl_interface': 2.0126173088697988e-06, 'learning_rate/action_model': 2.0126173088697987e-05, 'epoch': 1.53}) +[2026-06-17 15:56:55,930][starVLA.training.train_starvla][INFO] - Step 5198, Loss: {'train/loss': 0.9914472699165344, 'train/grad_norm_pre_clip': 21.59334945678711, 'timing/data': 0.0052837509429082274, 'timing/model': 1.2070486049633473, 'learning_rate/qwen_vl_interface': 2.0110350576001244e-06, 'learning_rate/action_model': 2.0110350576001245e-05, 'epoch': 1.53}) +[2026-06-17 15:57:00,727][starVLA.training.train_starvla][INFO] - Step 5199, Loss: {'train/loss': 0.6359593272209167, 'train/grad_norm_pre_clip': 13.459823608398438, 'timing/data': 0.00515782693400979, 'timing/model': 1.206539626000449, 'learning_rate/qwen_vl_interface': 2.0094534777713684e-06, 'learning_rate/action_model': 2.0094534777713684e-05, 'epoch': 1.53}) +[2026-06-17 15:58:42,690][starVLA.training.train_starvla][INFO] - Step 5200, Loss: {'train/loss': 0.577031672000885, 'train/grad_norm_pre_clip': 13.711942672729492, 'eval/loss': 0.7180200219154358, 'eval/latency_0/loss': 0.7180200219154358, 'eval/demon_fire_precision': 0.8936003140950137, 'eval/demon_fire_recall': 0.9451827242524917, 'eval/demon_fire_f1': 0.9186680121089809, 'eval/demon_fire_rate_model': 0.7082869855394883, 'eval/demon_fire_rate_teacher': 0.6696329254727474, 'eval/demon_left_precision': 0.816924778761062, 'eval/demon_left_recall': 0.8368271954674221, 'eval/demon_left_f1': 0.8267562272600055, 'eval/demon_left_rate_model': 0.5027808676307007, 'eval/demon_left_rate_teacher': 0.49082313681868744, 'eval/demon_right_precision': 0.8351335986355884, 'eval/demon_right_recall': 0.8299435028248587, 'eval/demon_right_f1': 0.8325304618872201, 'eval/demon_right_rate_model': 0.48915461624026696, 'eval/demon_right_rate_teacher': 0.4922135706340378, 'eval/demon_move_f1': 0.8296433445736129, 'eval/demon_cc_f1': 0.8900170948837457, 'eval/latency_0/demon_fire_precision': 0.8936003140950137, 'eval/latency_0/demon_fire_recall': 0.9451827242524917, 'eval/latency_0/demon_fire_f1': 0.9186680121089809, 'eval/latency_0/demon_fire_rate_model': 0.7082869855394883, 'eval/latency_0/demon_fire_rate_teacher': 0.6696329254727474, 'eval/latency_0/demon_left_precision': 0.816924778761062, 'eval/latency_0/demon_left_recall': 0.8368271954674221, 'eval/latency_0/demon_left_f1': 0.8267562272600055, 'eval/latency_0/demon_left_rate_model': 0.5027808676307007, 'eval/latency_0/demon_left_rate_teacher': 0.49082313681868744, 'eval/latency_0/demon_right_precision': 0.8351335986355884, 'eval/latency_0/demon_right_recall': 0.8299435028248587, 'eval/latency_0/demon_right_f1': 0.8325304618872201, 'eval/latency_0/demon_right_rate_model': 0.48915461624026696, 'eval/latency_0/demon_right_rate_teacher': 0.4922135706340378, 'eval/latency_0/demon_move_f1': 0.8296433445736129, 'eval/latency_0/demon_cc_f1': 0.8900170948837457, 'timing/data': 0.005031449953094125, 'timing/model': 1.2099846130004153, 'learning_rate/qwen_vl_interface': 2.007872569711394e-06, 'learning_rate/action_model': 2.007872569711394e-05, 'epoch': 1.53}) +[2026-06-17 15:58:42,691][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_5200_state +[2026-06-17 15:58:42,693][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 15:58:46,251][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_5200_state/model.safetensors +[2026-06-17 15:58:53,787][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_5200_state/optimizer.bin +[2026-06-17 15:58:53,787][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_5200_state/sampler.bin +[2026-06-17 15:58:53,788][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_5200_state/sampler_1.bin +[2026-06-17 15:58:53,788][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_5200_state/random_states_0.pkl +[2026-06-17 15:58:53,836][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 5200 +[2026-06-17 15:58:53,837][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 15:58:53,846][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 15:58:58,545][starVLA.training.train_starvla][INFO] - Step 5201, Loss: {'train/loss': 0.4461207389831543, 'train/grad_norm_pre_clip': 11.26543140411377, 'timing/data': 0.00509677198715508, 'timing/model': 1.187753485981375, 'learning_rate/qwen_vl_interface': 2.0062923337479223e-06, 'learning_rate/action_model': 2.006292333747922e-05, 'epoch': 1.53}) +[2026-06-17 15:59:03,270][starVLA.training.train_starvla][INFO] - Step 5202, Loss: {'train/loss': 0.888067364692688, 'train/grad_norm_pre_clip': 13.24041748046875, 'timing/data': 0.005590509041212499, 'timing/model': 1.189646530081518, 'learning_rate/qwen_vl_interface': 2.0047127702085437e-06, 'learning_rate/action_model': 2.0047127702085437e-05, 'epoch': 1.53}) +[2026-06-17 15:59:08,007][starVLA.training.train_starvla][INFO] - Step 5203, Loss: {'train/loss': 0.8756453394889832, 'train/grad_norm_pre_clip': 16.289703369140625, 'timing/data': 0.0048080619890242815, 'timing/model': 1.321744901011698, 'learning_rate/qwen_vl_interface': 2.003133879420698e-06, 'learning_rate/action_model': 2.003133879420698e-05, 'epoch': 1.53}) +[2026-06-17 15:59:12,691][starVLA.training.train_starvla][INFO] - Step 5204, Loss: {'train/loss': 0.5672570466995239, 'train/grad_norm_pre_clip': 18.276554107666016, 'timing/data': 0.004921532003208995, 'timing/model': 1.2076730349799618, 'learning_rate/qwen_vl_interface': 2.001555661711691e-06, 'learning_rate/action_model': 2.001555661711691e-05, 'epoch': 1.53}) +[2026-06-17 15:59:17,444][starVLA.training.train_starvla][INFO] - Step 5205, Loss: {'train/loss': 0.4847853481769562, 'train/grad_norm_pre_clip': 11.47744369506836, 'timing/data': 0.004540286026895046, 'timing/model': 1.192126972018741, 'learning_rate/qwen_vl_interface': 1.999978117408693e-06, 'learning_rate/action_model': 1.999978117408693e-05, 'epoch': 1.53}) +[2026-06-17 15:59:22,179][starVLA.training.train_starvla][INFO] - Step 5206, Loss: {'train/loss': 0.6428495049476624, 'train/grad_norm_pre_clip': 15.6775484085083, 'timing/data': 0.0047975491033867, 'timing/model': 1.1967404989991337, 'learning_rate/qwen_vl_interface': 1.9984012468387285e-06, 'learning_rate/action_model': 1.9984012468387285e-05, 'epoch': 1.53}) +[2026-06-17 15:59:26,867][starVLA.training.train_starvla][INFO] - Step 5207, Loss: {'train/loss': 0.8786266446113586, 'train/grad_norm_pre_clip': 15.99973201751709, 'timing/data': 0.004953191033564508, 'timing/model': 1.1952083919895813, 'learning_rate/qwen_vl_interface': 1.9968250503286857e-06, 'learning_rate/action_model': 1.996825050328686e-05, 'epoch': 1.53}) +[2026-06-17 15:59:31,638][starVLA.training.train_starvla][INFO] - Step 5208, Loss: {'train/loss': 0.5177623629570007, 'train/grad_norm_pre_clip': 13.664578437805176, 'timing/data': 0.0047219060361385345, 'timing/model': 1.1985739440424368, 'learning_rate/qwen_vl_interface': 1.9952495282053073e-06, 'learning_rate/action_model': 1.9952495282053073e-05, 'epoch': 1.53}) +[2026-06-17 15:59:36,419][starVLA.training.train_starvla][INFO] - Step 5209, Loss: {'train/loss': 0.6576932072639465, 'train/grad_norm_pre_clip': 13.652535438537598, 'timing/data': 0.004474977031350136, 'timing/model': 1.336575158056803, 'learning_rate/qwen_vl_interface': 1.993674680795206e-06, 'learning_rate/action_model': 1.9936746807952058e-05, 'epoch': 1.53}) +[2026-06-17 15:59:41,046][starVLA.training.train_starvla][INFO] - Step 5210, Loss: {'train/loss': 0.5432307124137878, 'train/grad_norm_pre_clip': 13.038057327270508, 'timing/data': 0.005072633968666196, 'timing/model': 1.1976114310091361, 'learning_rate/qwen_vl_interface': 1.992100508424847e-06, 'learning_rate/action_model': 1.9921005084248467e-05, 'epoch': 1.53}) +[2026-06-17 15:59:45,800][starVLA.training.train_starvla][INFO] - Step 5211, Loss: {'train/loss': 0.6383872628211975, 'train/grad_norm_pre_clip': 17.423988342285156, 'timing/data': 0.004611433949321508, 'timing/model': 1.1986852179979905, 'learning_rate/qwen_vl_interface': 1.990527011420558e-06, 'learning_rate/action_model': 1.990527011420558e-05, 'epoch': 1.53}) +[2026-06-17 15:59:50,551][starVLA.training.train_starvla][INFO] - Step 5212, Loss: {'train/loss': 0.6704031825065613, 'train/grad_norm_pre_clip': 18.081314086914062, 'timing/data': 0.005043562036007643, 'timing/model': 1.198754213983193, 'learning_rate/qwen_vl_interface': 1.9889541901085274e-06, 'learning_rate/action_model': 1.9889541901085273e-05, 'epoch': 1.53}) +[2026-06-17 15:59:55,287][starVLA.training.train_starvla][INFO] - Step 5213, Loss: {'train/loss': 0.5431255102157593, 'train/grad_norm_pre_clip': 12.96454906463623, 'timing/data': 0.004881382919847965, 'timing/model': 1.2007175849284977, 'learning_rate/qwen_vl_interface': 1.987382044814802e-06, 'learning_rate/action_model': 1.987382044814802e-05, 'epoch': 1.53}) +[2026-06-17 16:00:00,049][starVLA.training.train_starvla][INFO] - Step 5214, Loss: {'train/loss': 0.6626290082931519, 'train/grad_norm_pre_clip': 10.903958320617676, 'timing/data': 0.0044604879803955555, 'timing/model': 1.1979263689136133, 'learning_rate/qwen_vl_interface': 1.9858105758652895e-06, 'learning_rate/action_model': 1.9858105758652892e-05, 'epoch': 1.53}) +[2026-06-17 16:00:04,810][starVLA.training.train_starvla][INFO] - Step 5215, Loss: {'train/loss': 0.7998574376106262, 'train/grad_norm_pre_clip': 19.452104568481445, 'timing/data': 0.005035904003307223, 'timing/model': 1.3266317959642038, 'learning_rate/qwen_vl_interface': 1.9842397835857555e-06, 'learning_rate/action_model': 1.9842397835857554e-05, 'epoch': 1.53}) +[2026-06-17 16:00:09,446][starVLA.training.train_starvla][INFO] - Step 5216, Loss: {'train/loss': 0.6295943260192871, 'train/grad_norm_pre_clip': 12.022871971130371, 'timing/data': 0.004712691064924002, 'timing/model': 1.2004721900448203, 'learning_rate/qwen_vl_interface': 1.982669668301831e-06, 'learning_rate/action_model': 1.982669668301831e-05, 'epoch': 1.53}) +[2026-06-17 16:00:14,214][starVLA.training.train_starvla][INFO] - Step 5217, Loss: {'train/loss': 0.6601114869117737, 'train/grad_norm_pre_clip': 13.807753562927246, 'timing/data': 0.005104494048282504, 'timing/model': 1.1993012820603326, 'learning_rate/qwen_vl_interface': 1.9811002303389993e-06, 'learning_rate/action_model': 1.9811002303389993e-05, 'epoch': 1.54}) +[2026-06-17 16:00:18,982][starVLA.training.train_starvla][INFO] - Step 5218, Loss: {'train/loss': 0.782454788684845, 'train/grad_norm_pre_clip': 23.739580154418945, 'timing/data': 0.004824380972422659, 'timing/model': 1.2010470279492438, 'learning_rate/qwen_vl_interface': 1.9795314700226068e-06, 'learning_rate/action_model': 1.9795314700226066e-05, 'epoch': 1.54}) +[2026-06-17 16:00:23,809][starVLA.training.train_starvla][INFO] - Step 5219, Loss: {'train/loss': 0.7288482785224915, 'train/grad_norm_pre_clip': 18.678165435791016, 'timing/data': 0.00483252399135381, 'timing/model': 1.2018992319935933, 'learning_rate/qwen_vl_interface': 1.9779633876778587e-06, 'learning_rate/action_model': 1.9779633876778586e-05, 'epoch': 1.54}) +[2026-06-17 16:00:28,468][starVLA.training.train_starvla][INFO] - Step 5220, Loss: {'train/loss': 0.7092241048812866, 'train/grad_norm_pre_clip': 17.23792839050293, 'timing/data': 0.004603650071658194, 'timing/model': 1.2012157450662926, 'learning_rate/qwen_vl_interface': 1.9763959836298225e-06, 'learning_rate/action_model': 1.9763959836298225e-05, 'epoch': 1.54}) +[2026-06-17 16:00:33,150][starVLA.training.train_starvla][INFO] - Step 5221, Loss: {'train/loss': 0.5513628125190735, 'train/grad_norm_pre_clip': 11.846334457397461, 'timing/data': 0.004686411004513502, 'timing/model': 1.2398459929972887, 'learning_rate/qwen_vl_interface': 1.9748292582034228e-06, 'learning_rate/action_model': 1.9748292582034228e-05, 'epoch': 1.54}) +[2026-06-17 16:00:37,797][starVLA.training.train_starvla][INFO] - Step 5222, Loss: {'train/loss': 0.7231380343437195, 'train/grad_norm_pre_clip': 17.480209350585938, 'timing/data': 0.004796096938662231, 'timing/model': 1.201808285084553, 'learning_rate/qwen_vl_interface': 1.9732632117234405e-06, 'learning_rate/action_model': 1.9732632117234404e-05, 'epoch': 1.54}) +[2026-06-17 16:00:42,521][starVLA.training.train_starvla][INFO] - Step 5223, Loss: {'train/loss': 0.5945199131965637, 'train/grad_norm_pre_clip': 14.605035781860352, 'timing/data': 0.004830069025047123, 'timing/model': 1.2035058739129454, 'learning_rate/qwen_vl_interface': 1.9716978445145195e-06, 'learning_rate/action_model': 1.9716978445145193e-05, 'epoch': 1.54}) +[2026-06-17 16:00:47,307][starVLA.training.train_starvla][INFO] - Step 5224, Loss: {'train/loss': 0.9528954029083252, 'train/grad_norm_pre_clip': 15.304695129394531, 'timing/data': 0.0047144959680736065, 'timing/model': 1.20474457310047, 'learning_rate/qwen_vl_interface': 1.9701331569011634e-06, 'learning_rate/action_model': 1.9701331569011633e-05, 'epoch': 1.54}) +[2026-06-17 16:00:52,088][starVLA.training.train_starvla][INFO] - Step 5225, Loss: {'train/loss': 0.741147518157959, 'train/grad_norm_pre_clip': 13.781306266784668, 'timing/data': 0.004655154072679579, 'timing/model': 1.202678874018602, 'learning_rate/qwen_vl_interface': 1.968569149207734e-06, 'learning_rate/action_model': 1.968569149207734e-05, 'epoch': 1.54}) +[2026-06-17 16:00:56,861][starVLA.training.train_starvla][INFO] - Step 5226, Loss: {'train/loss': 0.5949602723121643, 'train/grad_norm_pre_clip': 13.274469375610352, 'timing/data': 0.004492663079872727, 'timing/model': 1.3269800679991022, 'learning_rate/qwen_vl_interface': 1.9670058217584505e-06, 'learning_rate/action_model': 1.9670058217584504e-05, 'epoch': 1.54}) +[2026-06-17 16:01:01,580][starVLA.training.train_starvla][INFO] - Step 5227, Loss: {'train/loss': 0.9341492652893066, 'train/grad_norm_pre_clip': 19.85463523864746, 'timing/data': 0.004983557970263064, 'timing/model': 1.269733974011615, 'learning_rate/qwen_vl_interface': 1.965443174877393e-06, 'learning_rate/action_model': 1.965443174877393e-05, 'epoch': 1.54}) +[2026-06-17 16:01:06,232][starVLA.training.train_starvla][INFO] - Step 5228, Loss: {'train/loss': 0.5353171825408936, 'train/grad_norm_pre_clip': 10.831055641174316, 'timing/data': 0.004777816007845104, 'timing/model': 1.206201454042457, 'learning_rate/qwen_vl_interface': 1.9638812088885003e-06, 'learning_rate/action_model': 1.9638812088885e-05, 'epoch': 1.54}) +[2026-06-17 16:01:10,974][starVLA.training.train_starvla][INFO] - Step 5229, Loss: {'train/loss': 0.5080487132072449, 'train/grad_norm_pre_clip': 11.641850471496582, 'timing/data': 0.004876145976595581, 'timing/model': 1.2095065650064498, 'learning_rate/qwen_vl_interface': 1.9623199241155684e-06, 'learning_rate/action_model': 1.9623199241155683e-05, 'epoch': 1.54}) +[2026-06-17 16:01:15,768][starVLA.training.train_starvla][INFO] - Step 5230, Loss: {'train/loss': 0.7255473732948303, 'train/grad_norm_pre_clip': 15.732823371887207, 'timing/data': 0.004796401015482843, 'timing/model': 1.2034681630320847, 'learning_rate/qwen_vl_interface': 1.9607593208822536e-06, 'learning_rate/action_model': 1.9607593208822535e-05, 'epoch': 1.54}) +[2026-06-17 16:01:20,552][starVLA.training.train_starvla][INFO] - Step 5231, Loss: {'train/loss': 0.3523915112018585, 'train/grad_norm_pre_clip': 9.342228889465332, 'timing/data': 0.0056791509268805385, 'timing/model': 1.201976499054581, 'learning_rate/qwen_vl_interface': 1.9591993995120715e-06, 'learning_rate/action_model': 1.9591993995120712e-05, 'epoch': 1.54}) +[2026-06-17 16:01:25,328][starVLA.training.train_starvla][INFO] - Step 5232, Loss: {'train/loss': 0.5318235754966736, 'train/grad_norm_pre_clip': 12.250328063964844, 'timing/data': 0.00477380701340735, 'timing/model': 1.3295323749771342, 'learning_rate/qwen_vl_interface': 1.9576401603283944e-06, 'learning_rate/action_model': 1.9576401603283944e-05, 'epoch': 1.54}) +[2026-06-17 16:01:30,102][starVLA.training.train_starvla][INFO] - Step 5233, Loss: {'train/loss': 0.6165364384651184, 'train/grad_norm_pre_clip': 13.548200607299805, 'timing/data': 0.004880910040810704, 'timing/model': 1.3199595180340111, 'learning_rate/qwen_vl_interface': 1.956081603654455e-06, 'learning_rate/action_model': 1.956081603654455e-05, 'epoch': 1.54}) +[2026-06-17 16:01:34,753][starVLA.training.train_starvla][INFO] - Step 5234, Loss: {'train/loss': 0.7001652121543884, 'train/grad_norm_pre_clip': 13.580208778381348, 'timing/data': 0.004753869026899338, 'timing/model': 1.202425671974197, 'learning_rate/qwen_vl_interface': 1.9545237298133423e-06, 'learning_rate/action_model': 1.9545237298133424e-05, 'epoch': 1.54}) +[2026-06-17 16:01:39,501][starVLA.training.train_starvla][INFO] - Step 5235, Loss: {'train/loss': 0.6118031740188599, 'train/grad_norm_pre_clip': 13.515944480895996, 'timing/data': 0.004711371031589806, 'timing/model': 1.2025538459420204, 'learning_rate/qwen_vl_interface': 1.9529665391280096e-06, 'learning_rate/action_model': 1.95296653912801e-05, 'epoch': 1.54}) +[2026-06-17 16:01:44,282][starVLA.training.train_starvla][INFO] - Step 5236, Loss: {'train/loss': 0.6312611699104309, 'train/grad_norm_pre_clip': 12.120636940002441, 'timing/data': 0.004837983986362815, 'timing/model': 1.2030763060320169, 'learning_rate/qwen_vl_interface': 1.9514100319212594e-06, 'learning_rate/action_model': 1.9514100319212593e-05, 'epoch': 1.54}) +[2026-06-17 16:01:49,070][starVLA.training.train_starvla][INFO] - Step 5237, Loss: {'train/loss': 0.6568859815597534, 'train/grad_norm_pre_clip': 14.276361465454102, 'timing/data': 0.004893637960776687, 'timing/model': 1.2031075530685484, 'learning_rate/qwen_vl_interface': 1.949854208515757e-06, 'learning_rate/action_model': 1.9498542085157573e-05, 'epoch': 1.54}) +[2026-06-17 16:01:53,788][starVLA.training.train_starvla][INFO] - Step 5238, Loss: {'train/loss': 0.6505934596061707, 'train/grad_norm_pre_clip': 14.071189880371094, 'timing/data': 0.004511165083386004, 'timing/model': 1.2737788900267333, 'learning_rate/qwen_vl_interface': 1.9482990692340305e-06, 'learning_rate/action_model': 1.9482990692340305e-05, 'epoch': 1.54}) +[2026-06-17 16:01:58,433][starVLA.training.train_starvla][INFO] - Step 5239, Loss: {'train/loss': 0.5593107342720032, 'train/grad_norm_pre_clip': 12.867382049560547, 'timing/data': 0.00501553004141897, 'timing/model': 1.2027420779922977, 'learning_rate/qwen_vl_interface': 1.94674461439846e-06, 'learning_rate/action_model': 1.9467446143984598e-05, 'epoch': 1.54}) +[2026-06-17 16:02:03,187][starVLA.training.train_starvla][INFO] - Step 5240, Loss: {'train/loss': 0.6888676881790161, 'train/grad_norm_pre_clip': 12.850335121154785, 'timing/data': 0.0052482400787994266, 'timing/model': 1.2032932169968262, 'learning_rate/qwen_vl_interface': 1.9451908443312863e-06, 'learning_rate/action_model': 1.9451908443312865e-05, 'epoch': 1.54}) +[2026-06-17 16:02:08,007][starVLA.training.train_starvla][INFO] - Step 5241, Loss: {'train/loss': 0.42753729224205017, 'train/grad_norm_pre_clip': 11.509079933166504, 'timing/data': 0.00498015305493027, 'timing/model': 1.2032133369939402, 'learning_rate/qwen_vl_interface': 1.943637759354604e-06, 'learning_rate/action_model': 1.9436377593546042e-05, 'epoch': 1.54}) +[2026-06-17 16:02:12,795][starVLA.training.train_starvla][INFO] - Step 5242, Loss: {'train/loss': 0.7339813709259033, 'train/grad_norm_pre_clip': 14.379402160644531, 'timing/data': 0.004578717984259129, 'timing/model': 1.2028426269534975, 'learning_rate/qwen_vl_interface': 1.942085359790373e-06, 'learning_rate/action_model': 1.942085359790373e-05, 'epoch': 1.54}) +[2026-06-17 16:02:17,584][starVLA.training.train_starvla][INFO] - Step 5243, Loss: {'train/loss': 0.48989278078079224, 'train/grad_norm_pre_clip': 10.491138458251953, 'timing/data': 0.00482019002083689, 'timing/model': 1.336375688901171, 'learning_rate/qwen_vl_interface': 1.940533645960407e-06, 'learning_rate/action_model': 1.940533645960407e-05, 'epoch': 1.54}) +[2026-06-17 16:02:22,331][starVLA.training.train_starvla][INFO] - Step 5244, Loss: {'train/loss': 0.7194596529006958, 'train/grad_norm_pre_clip': 17.4692440032959, 'timing/data': 0.004938059952110052, 'timing/model': 1.2862200039671734, 'learning_rate/qwen_vl_interface': 1.9389826181863767e-06, 'learning_rate/action_model': 1.9389826181863767e-05, 'epoch': 1.54}) +[2026-06-17 16:02:26,979][starVLA.training.train_starvla][INFO] - Step 5245, Loss: {'train/loss': 0.5999062061309814, 'train/grad_norm_pre_clip': 12.974257469177246, 'timing/data': 0.004378021927550435, 'timing/model': 1.2045045440318063, 'learning_rate/qwen_vl_interface': 1.937432276789813e-06, 'learning_rate/action_model': 1.9374322767898127e-05, 'epoch': 1.54}) +[2026-06-17 16:02:31,768][starVLA.training.train_starvla][INFO] - Step 5246, Loss: {'train/loss': 0.3841184675693512, 'train/grad_norm_pre_clip': 9.219331741333008, 'timing/data': 0.004888199036940932, 'timing/model': 1.2068708429578692, 'learning_rate/qwen_vl_interface': 1.935882622092103e-06, 'learning_rate/action_model': 1.935882622092103e-05, 'epoch': 1.54}) +[2026-06-17 16:02:36,610][starVLA.training.train_starvla][INFO] - Step 5247, Loss: {'train/loss': 0.7756291031837463, 'train/grad_norm_pre_clip': 20.572832107543945, 'timing/data': 0.005188958020880818, 'timing/model': 1.2034137779846787, 'learning_rate/qwen_vl_interface': 1.9343336544144907e-06, 'learning_rate/action_model': 1.9343336544144906e-05, 'epoch': 1.54}) +[2026-06-17 16:02:41,383][starVLA.training.train_starvla][INFO] - Step 5248, Loss: {'train/loss': 0.644307553768158, 'train/grad_norm_pre_clip': 34.86161422729492, 'timing/data': 0.00480322795920074, 'timing/model': 1.2004932810086757, 'learning_rate/qwen_vl_interface': 1.9327853740780787e-06, 'learning_rate/action_model': 1.932785374078079e-05, 'epoch': 1.54}) +[2026-06-17 16:02:46,166][starVLA.training.train_starvla][INFO] - Step 5249, Loss: {'train/loss': 0.5372149348258972, 'train/grad_norm_pre_clip': 15.062606811523438, 'timing/data': 0.004833711893297732, 'timing/model': 1.2100339400349185, 'learning_rate/qwen_vl_interface': 1.9312377814038314e-06, 'learning_rate/action_model': 1.9312377814038312e-05, 'epoch': 1.54}) +[2026-06-17 16:02:50,908][starVLA.training.train_starvla][INFO] - Step 5250, Loss: {'train/loss': 0.7514287233352661, 'train/grad_norm_pre_clip': 15.233940124511719, 'timing/data': 0.004663623054511845, 'timing/model': 1.287463763030246, 'learning_rate/qwen_vl_interface': 1.9296908767125612e-06, 'learning_rate/action_model': 1.929690876712561e-05, 'epoch': 1.54}) +[2026-06-17 16:02:55,561][starVLA.training.train_starvla][INFO] - Step 5251, Loss: {'train/loss': 0.6635577082633972, 'train/grad_norm_pre_clip': 15.614890098571777, 'timing/data': 0.004730431945063174, 'timing/model': 1.2040051370859146, 'learning_rate/qwen_vl_interface': 1.928144660324945e-06, 'learning_rate/action_model': 1.928144660324945e-05, 'epoch': 1.55}) +[2026-06-17 16:03:00,338][starVLA.training.train_starvla][INFO] - Step 5252, Loss: {'train/loss': 0.5371829271316528, 'train/grad_norm_pre_clip': 10.921944618225098, 'timing/data': 0.0050186620792374015, 'timing/model': 1.2032700330018997, 'learning_rate/qwen_vl_interface': 1.9265991325615135e-06, 'learning_rate/action_model': 1.9265991325615135e-05, 'epoch': 1.55}) +[2026-06-17 16:03:05,119][starVLA.training.train_starvla][INFO] - Step 5253, Loss: {'train/loss': 0.5936292409896851, 'train/grad_norm_pre_clip': 13.464397430419922, 'timing/data': 0.0043725959258154035, 'timing/model': 1.2008038039784878, 'learning_rate/qwen_vl_interface': 1.9250542937426604e-06, 'learning_rate/action_model': 1.9250542937426602e-05, 'epoch': 1.55}) +[2026-06-17 16:03:09,916][starVLA.training.train_starvla][INFO] - Step 5254, Loss: {'train/loss': 0.6107527613639832, 'train/grad_norm_pre_clip': 14.014518737792969, 'timing/data': 0.004744566977024078, 'timing/model': 1.2094846409745514, 'learning_rate/qwen_vl_interface': 1.9235101441886307e-06, 'learning_rate/action_model': 1.923510144188631e-05, 'epoch': 1.55}) +[2026-06-17 16:03:14,694][starVLA.training.train_starvla][INFO] - Step 5255, Loss: {'train/loss': 0.6383443474769592, 'train/grad_norm_pre_clip': 15.95132064819336, 'timing/data': 0.004860979039222002, 'timing/model': 1.3343611829914153, 'learning_rate/qwen_vl_interface': 1.921966684219525e-06, 'learning_rate/action_model': 1.921966684219525e-05, 'epoch': 1.55}) +[2026-06-17 16:03:19,340][starVLA.training.train_starvla][INFO] - Step 5256, Loss: {'train/loss': 0.8386437296867371, 'train/grad_norm_pre_clip': 17.430030822753906, 'timing/data': 0.005011813016608357, 'timing/model': 1.200399006018415, 'learning_rate/qwen_vl_interface': 1.920423914155308e-06, 'learning_rate/action_model': 1.920423914155308e-05, 'epoch': 1.55}) +[2026-06-17 16:03:24,161][starVLA.training.train_starvla][INFO] - Step 5257, Loss: {'train/loss': 0.7074965834617615, 'train/grad_norm_pre_clip': 15.917096138000488, 'timing/data': 0.004903028020635247, 'timing/model': 1.2022118070162833, 'learning_rate/qwen_vl_interface': 1.9188818343157962e-06, 'learning_rate/action_model': 1.9188818343157964e-05, 'epoch': 1.55}) +[2026-06-17 16:03:28,964][starVLA.training.train_starvla][INFO] - Step 5258, Loss: {'train/loss': 0.566834032535553, 'train/grad_norm_pre_clip': 13.370323181152344, 'timing/data': 0.004658404039219022, 'timing/model': 1.203620980028063, 'learning_rate/qwen_vl_interface': 1.9173404450206654e-06, 'learning_rate/action_model': 1.9173404450206654e-05, 'epoch': 1.55}) +[2026-06-17 16:03:33,660][starVLA.training.train_starvla][INFO] - Step 5259, Loss: {'train/loss': 0.5368285179138184, 'train/grad_norm_pre_clip': 12.80594539642334, 'timing/data': 0.004960461053997278, 'timing/model': 1.200868495972827, 'learning_rate/qwen_vl_interface': 1.9157997465894452e-06, 'learning_rate/action_model': 1.9157997465894454e-05, 'epoch': 1.55}) +[2026-06-17 16:03:38,402][starVLA.training.train_starvla][INFO] - Step 5260, Loss: {'train/loss': 0.5732406973838806, 'train/grad_norm_pre_clip': 14.881092071533203, 'timing/data': 0.0048875040374696255, 'timing/model': 1.2128539039986208, 'learning_rate/qwen_vl_interface': 1.9142597393415264e-06, 'learning_rate/action_model': 1.914259739341526e-05, 'epoch': 1.55}) +[2026-06-17 16:03:43,187][starVLA.training.train_starvla][INFO] - Step 5261, Loss: {'train/loss': 0.5134122967720032, 'train/grad_norm_pre_clip': 16.301429748535156, 'timing/data': 0.004755061934702098, 'timing/model': 1.3393758969614282, 'learning_rate/qwen_vl_interface': 1.9127204235961524e-06, 'learning_rate/action_model': 1.9127204235961525e-05, 'epoch': 1.55}) +[2026-06-17 16:03:47,836][starVLA.training.train_starvla][INFO] - Step 5262, Loss: {'train/loss': 0.5034410953521729, 'train/grad_norm_pre_clip': 9.964174270629883, 'timing/data': 0.004438972915522754, 'timing/model': 1.203125542961061, 'learning_rate/qwen_vl_interface': 1.911181799672426e-06, 'learning_rate/action_model': 1.9111817996724264e-05, 'epoch': 1.55}) +[2026-06-17 16:03:52,699][starVLA.training.train_starvla][INFO] - Step 5263, Loss: {'train/loss': 0.9401090145111084, 'train/grad_norm_pre_clip': 16.317962646484375, 'timing/data': 0.00463543098885566, 'timing/model': 1.2019370869966224, 'learning_rate/qwen_vl_interface': 1.909643867889305e-06, 'learning_rate/action_model': 1.909643867889305e-05, 'epoch': 1.55}) +[2026-06-17 16:03:57,475][starVLA.training.train_starvla][INFO] - Step 5264, Loss: {'train/loss': 0.7641751766204834, 'train/grad_norm_pre_clip': 14.978803634643555, 'timing/data': 0.004585616989061236, 'timing/model': 1.2039742439519614, 'learning_rate/qwen_vl_interface': 1.908106628565605e-06, 'learning_rate/action_model': 1.9081066285656047e-05, 'epoch': 1.55}) +[2026-06-17 16:04:02,239][starVLA.training.train_starvla][INFO] - Step 5265, Loss: {'train/loss': 0.5971046686172485, 'train/grad_norm_pre_clip': 13.937211990356445, 'timing/data': 0.004800271010026336, 'timing/model': 1.2019939189776778, 'learning_rate/qwen_vl_interface': 1.9065700820199965e-06, 'learning_rate/action_model': 1.9065700820199963e-05, 'epoch': 1.55}) +[2026-06-17 16:04:06,971][starVLA.training.train_starvla][INFO] - Step 5266, Loss: {'train/loss': 0.6149687170982361, 'train/grad_norm_pre_clip': 17.811065673828125, 'timing/data': 0.004665566957555711, 'timing/model': 1.2032209229655564, 'learning_rate/qwen_vl_interface': 1.9050342285710063e-06, 'learning_rate/action_model': 1.9050342285710062e-05, 'epoch': 1.55}) +[2026-06-17 16:04:11,742][starVLA.training.train_starvla][INFO] - Step 5267, Loss: {'train/loss': 0.699887216091156, 'train/grad_norm_pre_clip': 13.997907638549805, 'timing/data': 0.004804347991012037, 'timing/model': 1.3276261349674314, 'learning_rate/qwen_vl_interface': 1.9034990685370217e-06, 'learning_rate/action_model': 1.9034990685370217e-05, 'epoch': 1.55}) +[2026-06-17 16:04:16,403][starVLA.training.train_starvla][INFO] - Step 5268, Loss: {'train/loss': 0.3273257911205292, 'train/grad_norm_pre_clip': 9.203526496887207, 'timing/data': 0.005069619044661522, 'timing/model': 1.2034452270017937, 'learning_rate/qwen_vl_interface': 1.901964602236282e-06, 'learning_rate/action_model': 1.901964602236282e-05, 'epoch': 1.55}) +[2026-06-17 16:04:21,183][starVLA.training.train_starvla][INFO] - Step 5269, Loss: {'train/loss': 0.8293002843856812, 'train/grad_norm_pre_clip': 17.5632381439209, 'timing/data': 0.005038521019741893, 'timing/model': 1.2046067209448665, 'learning_rate/qwen_vl_interface': 1.900430829986882e-06, 'learning_rate/action_model': 1.900430829986882e-05, 'epoch': 1.55}) +[2026-06-17 16:04:26,043][starVLA.training.train_starvla][INFO] - Step 5270, Loss: {'train/loss': 0.48560890555381775, 'train/grad_norm_pre_clip': 36.90024948120117, 'timing/data': 0.004211953957565129, 'timing/model': 1.2037569329841062, 'learning_rate/qwen_vl_interface': 1.898897752106773e-06, 'learning_rate/action_model': 1.898897752106773e-05, 'epoch': 1.55}) +[2026-06-17 16:04:30,819][starVLA.training.train_starvla][INFO] - Step 5271, Loss: {'train/loss': 0.49127399921417236, 'train/grad_norm_pre_clip': 11.842580795288086, 'timing/data': 0.004867424955591559, 'timing/model': 1.2008878340711817, 'learning_rate/qwen_vl_interface': 1.8973653689137683e-06, 'learning_rate/action_model': 1.8973653689137684e-05, 'epoch': 1.55}) +[2026-06-17 16:04:35,596][starVLA.training.train_starvla][INFO] - Step 5272, Loss: {'train/loss': 0.789454460144043, 'train/grad_norm_pre_clip': 27.798763275146484, 'timing/data': 0.004836859996430576, 'timing/model': 1.332643467001617, 'learning_rate/qwen_vl_interface': 1.8958336807255298e-06, 'learning_rate/action_model': 1.8958336807255297e-05, 'epoch': 1.55}) +[2026-06-17 16:04:40,240][starVLA.training.train_starvla][INFO] - Step 5273, Loss: {'train/loss': 0.9560249447822571, 'train/grad_norm_pre_clip': 16.603734970092773, 'timing/data': 0.004837048938497901, 'timing/model': 1.201973402989097, 'learning_rate/qwen_vl_interface': 1.8943026878595785e-06, 'learning_rate/action_model': 1.8943026878595786e-05, 'epoch': 1.55}) +[2026-06-17 16:04:45,007][starVLA.training.train_starvla][INFO] - Step 5274, Loss: {'train/loss': 0.6129310131072998, 'train/grad_norm_pre_clip': 15.937437057495117, 'timing/data': 0.005198316997848451, 'timing/model': 1.2007134660379961, 'learning_rate/qwen_vl_interface': 1.8927723906332913e-06, 'learning_rate/action_model': 1.8927723906332912e-05, 'epoch': 1.55}) +[2026-06-17 16:04:49,703][starVLA.training.train_starvla][INFO] - Step 5275, Loss: {'train/loss': 1.1045387983322144, 'train/grad_norm_pre_clip': 19.026731491088867, 'timing/data': 0.005011245957575738, 'timing/model': 1.203590891091153, 'learning_rate/qwen_vl_interface': 1.8912427893638996e-06, 'learning_rate/action_model': 1.8912427893638996e-05, 'epoch': 1.55}) +[2026-06-17 16:04:54,440][starVLA.training.train_starvla][INFO] - Step 5276, Loss: {'train/loss': 0.46567797660827637, 'train/grad_norm_pre_clip': 12.039377212524414, 'timing/data': 0.0049906279891729355, 'timing/model': 1.2022262229584157, 'learning_rate/qwen_vl_interface': 1.8897138843684918e-06, 'learning_rate/action_model': 1.8897138843684917e-05, 'epoch': 1.55}) +[2026-06-17 16:04:59,223][starVLA.training.train_starvla][INFO] - Step 5277, Loss: {'train/loss': 0.3239280879497528, 'train/grad_norm_pre_clip': 8.895820617675781, 'timing/data': 0.004670448019169271, 'timing/model': 1.2032457519089803, 'learning_rate/qwen_vl_interface': 1.8881856759640122e-06, 'learning_rate/action_model': 1.8881856759640122e-05, 'epoch': 1.55}) +[2026-06-17 16:05:04,010][starVLA.training.train_starvla][INFO] - Step 5278, Loss: {'train/loss': 0.5203269124031067, 'train/grad_norm_pre_clip': 11.097406387329102, 'timing/data': 0.004577387007884681, 'timing/model': 1.3371183639392257, 'learning_rate/qwen_vl_interface': 1.8866581644672589e-06, 'learning_rate/action_model': 1.8866581644672586e-05, 'epoch': 1.55}) +[2026-06-17 16:05:08,658][starVLA.training.train_starvla][INFO] - Step 5279, Loss: {'train/loss': 0.903997004032135, 'train/grad_norm_pre_clip': 19.079336166381836, 'timing/data': 0.005127902957610786, 'timing/model': 1.2029696739045903, 'learning_rate/qwen_vl_interface': 1.8851313501948872e-06, 'learning_rate/action_model': 1.885131350194887e-05, 'epoch': 1.55}) +[2026-06-17 16:05:13,451][starVLA.training.train_starvla][INFO] - Step 5280, Loss: {'train/loss': 0.4827398359775543, 'train/grad_norm_pre_clip': 11.682652473449707, 'timing/data': 0.00474792905151844, 'timing/model': 1.204066384001635, 'learning_rate/qwen_vl_interface': 1.8836052334634072e-06, 'learning_rate/action_model': 1.883605233463407e-05, 'epoch': 1.55}) +[2026-06-17 16:05:18,318][starVLA.training.train_starvla][INFO] - Step 5281, Loss: {'train/loss': 0.5948281288146973, 'train/grad_norm_pre_clip': 12.656160354614258, 'timing/data': 0.004988041939213872, 'timing/model': 1.2029062079964206, 'learning_rate/qwen_vl_interface': 1.8820798145891834e-06, 'learning_rate/action_model': 1.8820798145891832e-05, 'epoch': 1.55}) +[2026-06-17 16:05:23,054][starVLA.training.train_starvla][INFO] - Step 5282, Loss: {'train/loss': 0.653810977935791, 'train/grad_norm_pre_clip': 16.71360206604004, 'timing/data': 0.004606619942933321, 'timing/model': 1.2051007489208132, 'learning_rate/qwen_vl_interface': 1.88055509388844e-06, 'learning_rate/action_model': 1.88055509388844e-05, 'epoch': 1.55}) +[2026-06-17 16:05:27,863][starVLA.training.train_starvla][INFO] - Step 5283, Loss: {'train/loss': 0.875985324382782, 'train/grad_norm_pre_clip': 14.031220436096191, 'timing/data': 0.004721964010968804, 'timing/model': 1.203722003963776, 'learning_rate/qwen_vl_interface': 1.8790310716772525e-06, 'learning_rate/action_model': 1.8790310716772524e-05, 'epoch': 1.55}) +[2026-06-17 16:05:32,693][starVLA.training.train_starvla][INFO] - Step 5284, Loss: {'train/loss': 0.6402634978294373, 'train/grad_norm_pre_clip': 13.019235610961914, 'timing/data': 0.004846546915359795, 'timing/model': 1.3438543520169333, 'learning_rate/qwen_vl_interface': 1.8775077482715496e-06, 'learning_rate/action_model': 1.8775077482715494e-05, 'epoch': 1.55}) +[2026-06-17 16:05:37,345][starVLA.training.train_starvla][INFO] - Step 5285, Loss: {'train/loss': 0.6438071727752686, 'train/grad_norm_pre_clip': 15.879751205444336, 'timing/data': 0.004907845985144377, 'timing/model': 1.2052343069808558, 'learning_rate/qwen_vl_interface': 1.8759851239871176e-06, 'learning_rate/action_model': 1.8759851239871175e-05, 'epoch': 1.56}) +[2026-06-17 16:05:42,134][starVLA.training.train_starvla][INFO] - Step 5286, Loss: {'train/loss': 0.6035150289535522, 'train/grad_norm_pre_clip': 26.199480056762695, 'timing/data': 0.005281353020109236, 'timing/model': 1.2063800770556554, 'learning_rate/qwen_vl_interface': 1.8744631991396023e-06, 'learning_rate/action_model': 1.8744631991396023e-05, 'epoch': 1.56}) +[2026-06-17 16:05:46,916][starVLA.training.train_starvla][INFO] - Step 5287, Loss: {'train/loss': 0.5011124014854431, 'train/grad_norm_pre_clip': 11.68787956237793, 'timing/data': 0.004475138965062797, 'timing/model': 1.2036625819746405, 'learning_rate/qwen_vl_interface': 1.8729419740444977e-06, 'learning_rate/action_model': 1.8729419740444977e-05, 'epoch': 1.56}) +[2026-06-17 16:05:51,704][starVLA.training.train_starvla][INFO] - Step 5288, Loss: {'train/loss': 0.7017495632171631, 'train/grad_norm_pre_clip': 12.967408180236816, 'timing/data': 0.0048784969840198755, 'timing/model': 1.2091575270751491, 'learning_rate/qwen_vl_interface': 1.8714214490171554e-06, 'learning_rate/action_model': 1.8714214490171552e-05, 'epoch': 1.56}) +[2026-06-17 16:05:56,382][starVLA.training.train_starvla][INFO] - Step 5289, Loss: {'train/loss': 0.7663406729698181, 'train/grad_norm_pre_clip': 18.390066146850586, 'timing/data': 0.004720991011708975, 'timing/model': 1.2129265769617632, 'learning_rate/qwen_vl_interface': 1.8699016243727826e-06, 'learning_rate/action_model': 1.8699016243727827e-05, 'epoch': 1.56}) +[2026-06-17 16:06:01,078][starVLA.training.train_starvla][INFO] - Step 5290, Loss: {'train/loss': 0.6507129073143005, 'train/grad_norm_pre_clip': 14.442976951599121, 'timing/data': 0.0048641510074958205, 'timing/model': 1.2469094559783116, 'learning_rate/qwen_vl_interface': 1.8683825004264402e-06, 'learning_rate/action_model': 1.86838250042644e-05, 'epoch': 1.56}) +[2026-06-17 16:06:05,731][starVLA.training.train_starvla][INFO] - Step 5291, Loss: {'train/loss': 0.9001798033714294, 'train/grad_norm_pre_clip': 17.05430793762207, 'timing/data': 0.005195140023715794, 'timing/model': 1.203230562969111, 'learning_rate/qwen_vl_interface': 1.8668640774930444e-06, 'learning_rate/action_model': 1.8668640774930442e-05, 'epoch': 1.56}) +[2026-06-17 16:06:10,464][starVLA.training.train_starvla][INFO] - Step 5292, Loss: {'train/loss': 0.8555450439453125, 'train/grad_norm_pre_clip': 16.036170959472656, 'timing/data': 0.004648188012652099, 'timing/model': 1.2043824560241774, 'learning_rate/qwen_vl_interface': 1.8653463558873651e-06, 'learning_rate/action_model': 1.8653463558873652e-05, 'epoch': 1.56}) +[2026-06-17 16:06:15,257][starVLA.training.train_starvla][INFO] - Step 5293, Loss: {'train/loss': 0.6643887758255005, 'train/grad_norm_pre_clip': 17.920005798339844, 'timing/data': 0.004714336944743991, 'timing/model': 1.204438347951509, 'learning_rate/qwen_vl_interface': 1.863829335924029e-06, 'learning_rate/action_model': 1.8638293359240288e-05, 'epoch': 1.56}) +[2026-06-17 16:06:20,072][starVLA.training.train_starvla][INFO] - Step 5294, Loss: {'train/loss': 0.3994643986225128, 'train/grad_norm_pre_clip': 11.361773490905762, 'timing/data': 0.004568760050460696, 'timing/model': 1.2045213679084554, 'learning_rate/qwen_vl_interface': 1.8623130179175148e-06, 'learning_rate/action_model': 1.8623130179175146e-05, 'epoch': 1.56}) +[2026-06-17 16:06:24,856][starVLA.training.train_starvla][INFO] - Step 5295, Loss: {'train/loss': 0.5988394618034363, 'train/grad_norm_pre_clip': 13.14737319946289, 'timing/data': 0.0049612929578870535, 'timing/model': 1.3301429740386084, 'learning_rate/qwen_vl_interface': 1.8607974021821573e-06, 'learning_rate/action_model': 1.8607974021821572e-05, 'epoch': 1.56}) +[2026-06-17 16:06:29,562][starVLA.training.train_starvla][INFO] - Step 5296, Loss: {'train/loss': 0.7499206066131592, 'train/grad_norm_pre_clip': 12.298704147338867, 'timing/data': 0.005105194984935224, 'timing/model': 1.2543405839242041, 'learning_rate/qwen_vl_interface': 1.8592824890321443e-06, 'learning_rate/action_model': 1.8592824890321442e-05, 'epoch': 1.56}) +[2026-06-17 16:06:34,211][starVLA.training.train_starvla][INFO] - Step 5297, Loss: {'train/loss': 0.6157543063163757, 'train/grad_norm_pre_clip': 14.328635215759277, 'timing/data': 0.0052406430477276444, 'timing/model': 1.204980239039287, 'learning_rate/qwen_vl_interface': 1.8577682787815223e-06, 'learning_rate/action_model': 1.8577682787815225e-05, 'epoch': 1.56}) +[2026-06-17 16:06:38,942][starVLA.training.train_starvla][INFO] - Step 5298, Loss: {'train/loss': 0.5832585096359253, 'train/grad_norm_pre_clip': 13.140960693359375, 'timing/data': 0.004708638996817172, 'timing/model': 1.2057706080377102, 'learning_rate/qwen_vl_interface': 1.856254771744185e-06, 'learning_rate/action_model': 1.856254771744185e-05, 'epoch': 1.56}) +[2026-06-17 16:06:43,732][starVLA.training.train_starvla][INFO] - Step 5299, Loss: {'train/loss': 0.8100219368934631, 'train/grad_norm_pre_clip': 14.369189262390137, 'timing/data': 0.004732982954010367, 'timing/model': 1.2074993499554694, 'learning_rate/qwen_vl_interface': 1.8547419682338838e-06, 'learning_rate/action_model': 1.8547419682338836e-05, 'epoch': 1.56}) +[2026-06-17 16:08:25,518][starVLA.training.train_starvla][INFO] - Step 5300, Loss: {'train/loss': 0.5073022246360779, 'train/grad_norm_pre_clip': 13.432863235473633, 'eval/loss': 0.706576406955719, 'eval/latency_0/loss': 0.706576406955719, 'eval/demon_fire_precision': 0.911620294599018, 'eval/demon_fire_recall': 0.925249169435216, 'eval/demon_fire_f1': 0.9183841714756802, 'eval/demon_fire_rate_model': 0.6796440489432704, 'eval/demon_fire_rate_teacher': 0.6696329254727474, 'eval/demon_left_precision': 0.8380614657210402, 'eval/demon_left_recall': 0.803399433427762, 'eval/demon_left_f1': 0.8203644778709863, 'eval/demon_left_rate_model': 0.47052280311457173, 'eval/demon_left_rate_teacher': 0.49082313681868744, 'eval/demon_right_precision': 0.8097768331562167, 'eval/demon_right_recall': 0.8610169491525423, 'eval/demon_right_f1': 0.8346111719605696, 'eval/demon_right_rate_model': 0.5233592880978866, 'eval/demon_right_rate_teacher': 0.4922135706340378, 'eval/demon_move_f1': 0.827487824915778, 'eval/demon_cc_f1': 0.8890853909789422, 'eval/latency_0/demon_fire_precision': 0.911620294599018, 'eval/latency_0/demon_fire_recall': 0.925249169435216, 'eval/latency_0/demon_fire_f1': 0.9183841714756802, 'eval/latency_0/demon_fire_rate_model': 0.6796440489432704, 'eval/latency_0/demon_fire_rate_teacher': 0.6696329254727474, 'eval/latency_0/demon_left_precision': 0.8380614657210402, 'eval/latency_0/demon_left_recall': 0.803399433427762, 'eval/latency_0/demon_left_f1': 0.8203644778709863, 'eval/latency_0/demon_left_rate_model': 0.47052280311457173, 'eval/latency_0/demon_left_rate_teacher': 0.49082313681868744, 'eval/latency_0/demon_right_precision': 0.8097768331562167, 'eval/latency_0/demon_right_recall': 0.8610169491525423, 'eval/latency_0/demon_right_f1': 0.8346111719605696, 'eval/latency_0/demon_right_rate_model': 0.5233592880978866, 'eval/latency_0/demon_right_rate_teacher': 0.4922135706340378, 'eval/latency_0/demon_move_f1': 0.827487824915778, 'eval/latency_0/demon_cc_f1': 0.8890853909789422, 'timing/data': 0.005017268005758524, 'timing/model': 1.203995011979714, 'learning_rate/qwen_vl_interface': 1.8532298685642275e-06, 'learning_rate/action_model': 1.8532298685642274e-05, 'epoch': 1.56}) +[2026-06-17 16:08:25,518][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_5300_state +[2026-06-17 16:08:25,522][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 16:08:29,081][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_5300_state/model.safetensors +[2026-06-17 16:08:36,715][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_5300_state/optimizer.bin +[2026-06-17 16:08:36,715][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_5300_state/sampler.bin +[2026-06-17 16:08:36,716][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_5300_state/sampler_1.bin +[2026-06-17 16:08:36,716][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_5300_state/random_states_0.pkl +[2026-06-17 16:08:36,770][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 5300 +[2026-06-17 16:08:36,770][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 16:08:36,783][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 16:08:41,487][starVLA.training.train_starvla][INFO] - Step 5301, Loss: {'train/loss': 0.5617499351501465, 'train/grad_norm_pre_clip': 17.370542526245117, 'timing/data': 0.005393061088398099, 'timing/model': 1.190133371972479, 'learning_rate/qwen_vl_interface': 1.8517184730486745e-06, 'learning_rate/action_model': 1.8517184730486746e-05, 'epoch': 1.56}) +[2026-06-17 16:08:46,225][starVLA.training.train_starvla][INFO] - Step 5302, Loss: {'train/loss': 0.6007702350616455, 'train/grad_norm_pre_clip': 13.179156303405762, 'timing/data': 0.0052735620411112905, 'timing/model': 1.1931364929769188, 'learning_rate/qwen_vl_interface': 1.8502077820005408e-06, 'learning_rate/action_model': 1.8502077820005407e-05, 'epoch': 1.56}) +[2026-06-17 16:08:50,974][starVLA.training.train_starvla][INFO] - Step 5303, Loss: {'train/loss': 0.49135541915893555, 'train/grad_norm_pre_clip': 17.48234748840332, 'timing/data': 0.005117986002005637, 'timing/model': 1.326259723980911, 'learning_rate/qwen_vl_interface': 1.8486977957329888e-06, 'learning_rate/action_model': 1.8486977957329886e-05, 'epoch': 1.56}) +[2026-06-17 16:08:55,592][starVLA.training.train_starvla][INFO] - Step 5304, Loss: {'train/loss': 0.48866066336631775, 'train/grad_norm_pre_clip': 11.092900276184082, 'timing/data': 0.0046961079351603985, 'timing/model': 1.1942837029928342, 'learning_rate/qwen_vl_interface': 1.847188514559045e-06, 'learning_rate/action_model': 1.847188514559045e-05, 'epoch': 1.56}) +[2026-06-17 16:09:00,355][starVLA.training.train_starvla][INFO] - Step 5305, Loss: {'train/loss': 0.7104732990264893, 'train/grad_norm_pre_clip': 15.105406761169434, 'timing/data': 0.005251509021036327, 'timing/model': 1.19850391196087, 'learning_rate/qwen_vl_interface': 1.8456799387915839e-06, 'learning_rate/action_model': 1.8456799387915838e-05, 'epoch': 1.56}) +[2026-06-17 16:09:05,110][starVLA.training.train_starvla][INFO] - Step 5306, Loss: {'train/loss': 0.4274531900882721, 'train/grad_norm_pre_clip': 14.64931583404541, 'timing/data': 0.004568109987303615, 'timing/model': 1.2004485140787438, 'learning_rate/qwen_vl_interface': 1.8441720687433345e-06, 'learning_rate/action_model': 1.8441720687433344e-05, 'epoch': 1.56}) +[2026-06-17 16:09:09,871][starVLA.training.train_starvla][INFO] - Step 5307, Loss: {'train/loss': 0.8260764479637146, 'train/grad_norm_pre_clip': 14.733114242553711, 'timing/data': 0.004476270056329668, 'timing/model': 1.1958394809626043, 'learning_rate/qwen_vl_interface': 1.8426649047268808e-06, 'learning_rate/action_model': 1.8426649047268807e-05, 'epoch': 1.56}) +[2026-06-17 16:09:14,633][starVLA.training.train_starvla][INFO] - Step 5308, Loss: {'train/loss': 0.5447288155555725, 'train/grad_norm_pre_clip': 12.238162994384766, 'timing/data': 0.004559412016533315, 'timing/model': 1.3309975989395753, 'learning_rate/qwen_vl_interface': 1.8411584470546582e-06, 'learning_rate/action_model': 1.841158447054658e-05, 'epoch': 1.56}) +[2026-06-17 16:09:19,269][starVLA.training.train_starvla][INFO] - Step 5309, Loss: {'train/loss': 0.43557408452033997, 'train/grad_norm_pre_clip': 10.327787399291992, 'timing/data': 0.005089403945021331, 'timing/model': 1.1997684719972312, 'learning_rate/qwen_vl_interface': 1.8396526960389575e-06, 'learning_rate/action_model': 1.8396526960389575e-05, 'epoch': 1.56}) +[2026-06-17 16:09:24,030][starVLA.training.train_starvla][INFO] - Step 5310, Loss: {'train/loss': 0.5982296466827393, 'train/grad_norm_pre_clip': 11.989297866821289, 'timing/data': 0.0048082900466397405, 'timing/model': 1.1998176029883325, 'learning_rate/qwen_vl_interface': 1.8381476519919215e-06, 'learning_rate/action_model': 1.8381476519919212e-05, 'epoch': 1.56}) +[2026-06-17 16:09:28,795][starVLA.training.train_starvla][INFO] - Step 5311, Loss: {'train/loss': 0.5036676526069641, 'train/grad_norm_pre_clip': 12.933347702026367, 'timing/data': 0.005151124903932214, 'timing/model': 1.1999091829638928, 'learning_rate/qwen_vl_interface': 1.8366433152255511e-06, 'learning_rate/action_model': 1.8366433152255512e-05, 'epoch': 1.56}) +[2026-06-17 16:09:33,449][starVLA.training.train_starvla][INFO] - Step 5312, Loss: {'train/loss': 1.031972050666809, 'train/grad_norm_pre_clip': 17.633089065551758, 'timing/data': 0.004961997037753463, 'timing/model': 1.2000194611027837, 'learning_rate/qwen_vl_interface': 1.8351396860516931e-06, 'learning_rate/action_model': 1.835139686051693e-05, 'epoch': 1.56}) +[2026-06-17 16:09:38,139][starVLA.training.train_starvla][INFO] - Step 5313, Loss: {'train/loss': 0.8352924585342407, 'train/grad_norm_pre_clip': 18.282114028930664, 'timing/data': 0.005175673053599894, 'timing/model': 1.200883644982241, 'learning_rate/qwen_vl_interface': 1.8336367647820527e-06, 'learning_rate/action_model': 1.8336367647820528e-05, 'epoch': 1.56}) +[2026-06-17 16:09:42,858][starVLA.training.train_starvla][INFO] - Step 5314, Loss: {'train/loss': 0.39949098229408264, 'train/grad_norm_pre_clip': 11.124041557312012, 'timing/data': 0.004917896003462374, 'timing/model': 1.2775992640526965, 'learning_rate/qwen_vl_interface': 1.832134551728186e-06, 'learning_rate/action_model': 1.832134551728186e-05, 'epoch': 1.56}) +[2026-06-17 16:09:47,507][starVLA.training.train_starvla][INFO] - Step 5315, Loss: {'train/loss': 0.8577672839164734, 'train/grad_norm_pre_clip': 15.069075584411621, 'timing/data': 0.004843123024329543, 'timing/model': 1.2010460019810125, 'learning_rate/qwen_vl_interface': 1.8306330472015051e-06, 'learning_rate/action_model': 1.8306330472015053e-05, 'epoch': 1.56}) +[2026-06-17 16:09:52,295][starVLA.training.train_starvla][INFO] - Step 5316, Loss: {'train/loss': 0.45088493824005127, 'train/grad_norm_pre_clip': 12.109787940979004, 'timing/data': 0.0051881601102650166, 'timing/model': 1.210876535042189, 'learning_rate/qwen_vl_interface': 1.8291322515132757e-06, 'learning_rate/action_model': 1.8291322515132756e-05, 'epoch': 1.56}) +[2026-06-17 16:09:57,069][starVLA.training.train_starvla][INFO] - Step 5317, Loss: {'train/loss': 0.6405584216117859, 'train/grad_norm_pre_clip': 21.615615844726562, 'timing/data': 0.005053862929344177, 'timing/model': 1.2022215030156076, 'learning_rate/qwen_vl_interface': 1.8276321649746076e-06, 'learning_rate/action_model': 1.8276321649746076e-05, 'epoch': 1.56}) +[2026-06-17 16:10:01,856][starVLA.training.train_starvla][INFO] - Step 5318, Loss: {'train/loss': 0.6196262240409851, 'train/grad_norm_pre_clip': 12.74935531616211, 'timing/data': 0.005333382054232061, 'timing/model': 1.2138969999505207, 'learning_rate/qwen_vl_interface': 1.826132787896476e-06, 'learning_rate/action_model': 1.826132787896476e-05, 'epoch': 1.56}) +[2026-06-17 16:10:06,618][starVLA.training.train_starvla][INFO] - Step 5319, Loss: {'train/loss': 0.5439701080322266, 'train/grad_norm_pre_clip': 12.676087379455566, 'timing/data': 0.0049482600297778845, 'timing/model': 1.2026530649745837, 'learning_rate/qwen_vl_interface': 1.8246341205897018e-06, 'learning_rate/action_model': 1.8246341205897017e-05, 'epoch': 1.57}) +[2026-06-17 16:10:11,378][starVLA.training.train_starvla][INFO] - Step 5320, Loss: {'train/loss': 0.4766889810562134, 'train/grad_norm_pre_clip': 12.129499435424805, 'timing/data': 0.005069173057563603, 'timing/model': 1.302599468966946, 'learning_rate/qwen_vl_interface': 1.8231361633649606e-06, 'learning_rate/action_model': 1.8231361633649606e-05, 'epoch': 1.57}) +[2026-06-17 16:10:16,037][starVLA.training.train_starvla][INFO] - Step 5321, Loss: {'train/loss': 0.4709680676460266, 'train/grad_norm_pre_clip': 12.128142356872559, 'timing/data': 0.0049859159626066685, 'timing/model': 1.2029498049523681, 'learning_rate/qwen_vl_interface': 1.8216389165327797e-06, 'learning_rate/action_model': 1.8216389165327798e-05, 'epoch': 1.57}) +[2026-06-17 16:10:20,828][starVLA.training.train_starvla][INFO] - Step 5322, Loss: {'train/loss': 1.357182502746582, 'train/grad_norm_pre_clip': 21.531023025512695, 'timing/data': 0.004728574072942138, 'timing/model': 1.2030597049742937, 'learning_rate/qwen_vl_interface': 1.8201423804035406e-06, 'learning_rate/action_model': 1.8201423804035405e-05, 'epoch': 1.57}) +[2026-06-17 16:10:25,707][starVLA.training.train_starvla][INFO] - Step 5323, Loss: {'train/loss': 0.49402326345443726, 'train/grad_norm_pre_clip': 12.220560073852539, 'timing/data': 0.004892157972790301, 'timing/model': 1.2047664270503446, 'learning_rate/qwen_vl_interface': 1.8186465552874769e-06, 'learning_rate/action_model': 1.8186465552874766e-05, 'epoch': 1.57}) +[2026-06-17 16:10:30,497][starVLA.training.train_starvla][INFO] - Step 5324, Loss: {'train/loss': 0.8284077644348145, 'train/grad_norm_pre_clip': 17.281957626342773, 'timing/data': 0.004560019006021321, 'timing/model': 1.203955593984574, 'learning_rate/qwen_vl_interface': 1.8171514414946745e-06, 'learning_rate/action_model': 1.8171514414946745e-05, 'epoch': 1.57}) +[2026-06-17 16:10:35,306][starVLA.training.train_starvla][INFO] - Step 5325, Loss: {'train/loss': 0.5186314582824707, 'train/grad_norm_pre_clip': 14.375611305236816, 'timing/data': 0.004792042076587677, 'timing/model': 1.3464598329737782, 'learning_rate/qwen_vl_interface': 1.815657039335072e-06, 'learning_rate/action_model': 1.815657039335072e-05, 'epoch': 1.57}) +[2026-06-17 16:10:40,043][starVLA.training.train_starvla][INFO] - Step 5326, Loss: {'train/loss': 0.8910639882087708, 'train/grad_norm_pre_clip': 15.501235961914062, 'timing/data': 0.005354413064196706, 'timing/model': 1.2034592899726704, 'learning_rate/qwen_vl_interface': 1.8141633491184612e-06, 'learning_rate/action_model': 1.814163349118461e-05, 'epoch': 1.57}) +[2026-06-17 16:10:44,830][starVLA.training.train_starvla][INFO] - Step 5327, Loss: {'train/loss': 0.4482520818710327, 'train/grad_norm_pre_clip': 9.559067726135254, 'timing/data': 0.004840761073864996, 'timing/model': 1.2067384829279035, 'learning_rate/qwen_vl_interface': 1.8126703711544848e-06, 'learning_rate/action_model': 1.812670371154485e-05, 'epoch': 1.57}) +[2026-06-17 16:10:49,621][starVLA.training.train_starvla][INFO] - Step 5328, Loss: {'train/loss': 0.6582631468772888, 'train/grad_norm_pre_clip': 12.320781707763672, 'timing/data': 0.005202788976021111, 'timing/model': 1.2033960989210755, 'learning_rate/qwen_vl_interface': 1.8111781057526381e-06, 'learning_rate/action_model': 1.811178105752638e-05, 'epoch': 1.57}) +[2026-06-17 16:10:54,332][starVLA.training.train_starvla][INFO] - Step 5329, Loss: {'train/loss': 0.7048666477203369, 'train/grad_norm_pre_clip': 14.036489486694336, 'timing/data': 0.0049029599176719785, 'timing/model': 1.2043388129677624, 'learning_rate/qwen_vl_interface': 1.8096865532222714e-06, 'learning_rate/action_model': 1.8096865532222714e-05, 'epoch': 1.57}) +[2026-06-17 16:10:59,066][starVLA.training.train_starvla][INFO] - Step 5330, Loss: {'train/loss': 0.5098662972450256, 'train/grad_norm_pre_clip': 11.45987319946289, 'timing/data': 0.005089855054393411, 'timing/model': 1.2040278719505295, 'learning_rate/qwen_vl_interface': 1.8081957138725861e-06, 'learning_rate/action_model': 1.808195713872586e-05, 'epoch': 1.57}) +[2026-06-17 16:11:03,870][starVLA.training.train_starvla][INFO] - Step 5331, Loss: {'train/loss': 0.6570580005645752, 'train/grad_norm_pre_clip': 14.45587158203125, 'timing/data': 0.00487995205912739, 'timing/model': 1.3349243639968336, 'learning_rate/qwen_vl_interface': 1.8067055880126311e-06, 'learning_rate/action_model': 1.806705588012631e-05, 'epoch': 1.57}) +[2026-06-17 16:11:08,541][starVLA.training.train_starvla][INFO] - Step 5332, Loss: {'train/loss': 0.8328787684440613, 'train/grad_norm_pre_clip': 17.356082916259766, 'timing/data': 0.004983906983397901, 'timing/model': 1.204279029974714, 'learning_rate/qwen_vl_interface': 1.8052161759513114e-06, 'learning_rate/action_model': 1.8052161759513115e-05, 'epoch': 1.57}) +[2026-06-17 16:11:13,362][starVLA.training.train_starvla][INFO] - Step 5333, Loss: {'train/loss': 1.0041909217834473, 'train/grad_norm_pre_clip': 19.518484115600586, 'timing/data': 0.00464407098479569, 'timing/model': 1.2054970480967313, 'learning_rate/qwen_vl_interface': 1.8037274779973868e-06, 'learning_rate/action_model': 1.8037274779973867e-05, 'epoch': 1.57}) +[2026-06-17 16:11:18,157][starVLA.training.train_starvla][INFO] - Step 5334, Loss: {'train/loss': 0.7703794240951538, 'train/grad_norm_pre_clip': 12.728931427001953, 'timing/data': 0.004986413987353444, 'timing/model': 1.2062023651087657, 'learning_rate/qwen_vl_interface': 1.8022394944594643e-06, 'learning_rate/action_model': 1.802239494459464e-05, 'epoch': 1.57}) +[2026-06-17 16:11:22,931][starVLA.training.train_starvla][INFO] - Step 5335, Loss: {'train/loss': 0.42793264985084534, 'train/grad_norm_pre_clip': 12.090254783630371, 'timing/data': 0.004776556044816971, 'timing/model': 1.202372569940053, 'learning_rate/qwen_vl_interface': 1.8007522256460062e-06, 'learning_rate/action_model': 1.8007522256460062e-05, 'epoch': 1.57}) +[2026-06-17 16:11:27,672][starVLA.training.train_starvla][INFO] - Step 5336, Loss: {'train/loss': 0.5246849656105042, 'train/grad_norm_pre_clip': 14.243477821350098, 'timing/data': 0.004981413017958403, 'timing/model': 1.2045969140017405, 'learning_rate/qwen_vl_interface': 1.7992656718653201e-06, 'learning_rate/action_model': 1.79926567186532e-05, 'epoch': 1.57}) +[2026-06-17 16:11:32,447][starVLA.training.train_starvla][INFO] - Step 5337, Loss: {'train/loss': 0.8325945734977722, 'train/grad_norm_pre_clip': 17.23248291015625, 'timing/data': 0.005012960988096893, 'timing/model': 1.327302681049332, 'learning_rate/qwen_vl_interface': 1.797779833425575e-06, 'learning_rate/action_model': 1.797779833425575e-05, 'epoch': 1.57}) +[2026-06-17 16:11:37,103][starVLA.training.train_starvla][INFO] - Step 5338, Loss: {'train/loss': 0.4846774935722351, 'train/grad_norm_pre_clip': 13.174378395080566, 'timing/data': 0.004922150983475149, 'timing/model': 1.2055453599896282, 'learning_rate/qwen_vl_interface': 1.7962947106347846e-06, 'learning_rate/action_model': 1.7962947106347845e-05, 'epoch': 1.57}) +[2026-06-17 16:11:41,913][starVLA.training.train_starvla][INFO] - Step 5339, Loss: {'train/loss': 0.6233135461807251, 'train/grad_norm_pre_clip': 15.766855239868164, 'timing/data': 0.0045559220016002655, 'timing/model': 1.2122414570767432, 'learning_rate/qwen_vl_interface': 1.794810303800818e-06, 'learning_rate/action_model': 1.794810303800818e-05, 'epoch': 1.57}) +[2026-06-17 16:11:46,734][starVLA.training.train_starvla][INFO] - Step 5340, Loss: {'train/loss': 0.6702753305435181, 'train/grad_norm_pre_clip': 14.483548164367676, 'timing/data': 0.0051801810041069984, 'timing/model': 1.2172433990053833, 'learning_rate/qwen_vl_interface': 1.7933266132313924e-06, 'learning_rate/action_model': 1.7933266132313923e-05, 'epoch': 1.57}) +[2026-06-17 16:11:51,542][starVLA.training.train_starvla][INFO] - Step 5341, Loss: {'train/loss': 0.9290317893028259, 'train/grad_norm_pre_clip': 19.970746994018555, 'timing/data': 0.004841602989472449, 'timing/model': 1.205228278064169, 'learning_rate/qwen_vl_interface': 1.7918436392340803e-06, 'learning_rate/action_model': 1.7918436392340802e-05, 'epoch': 1.57}) +[2026-06-17 16:11:56,342][starVLA.training.train_starvla][INFO] - Step 5342, Loss: {'train/loss': 0.7718634605407715, 'train/grad_norm_pre_clip': 13.011964797973633, 'timing/data': 0.0050482089864090085, 'timing/model': 1.3446576789719984, 'learning_rate/qwen_vl_interface': 1.7903613821163022e-06, 'learning_rate/action_model': 1.790361382116302e-05, 'epoch': 1.57}) +[2026-06-17 16:12:01,012][starVLA.training.train_starvla][INFO] - Step 5343, Loss: {'train/loss': 0.4967638850212097, 'train/grad_norm_pre_clip': 10.934688568115234, 'timing/data': 0.004788724007084966, 'timing/model': 1.2106310810195282, 'learning_rate/qwen_vl_interface': 1.7888798421853313e-06, 'learning_rate/action_model': 1.7888798421853313e-05, 'epoch': 1.57}) +[2026-06-17 16:12:05,803][starVLA.training.train_starvla][INFO] - Step 5344, Loss: {'train/loss': 0.7577462196350098, 'train/grad_norm_pre_clip': 14.297492027282715, 'timing/data': 0.004500273033045232, 'timing/model': 1.207562746014446, 'learning_rate/qwen_vl_interface': 1.7873990197482962e-06, 'learning_rate/action_model': 1.787399019748296e-05, 'epoch': 1.57}) +[2026-06-17 16:12:10,565][starVLA.training.train_starvla][INFO] - Step 5345, Loss: {'train/loss': 0.7522758841514587, 'train/grad_norm_pre_clip': 19.092979431152344, 'timing/data': 0.004980444093234837, 'timing/model': 1.206198436091654, 'learning_rate/qwen_vl_interface': 1.785918915112169e-06, 'learning_rate/action_model': 1.785918915112169e-05, 'epoch': 1.57}) +[2026-06-17 16:12:15,324][starVLA.training.train_starvla][INFO] - Step 5346, Loss: {'train/loss': 0.7689617872238159, 'train/grad_norm_pre_clip': 14.702253341674805, 'timing/data': 0.00480914895888418, 'timing/model': 1.2158694620011374, 'learning_rate/qwen_vl_interface': 1.7844395285837773e-06, 'learning_rate/action_model': 1.7844395285837774e-05, 'epoch': 1.57}) +[2026-06-17 16:12:20,115][starVLA.training.train_starvla][INFO] - Step 5347, Loss: {'train/loss': 0.6506164073944092, 'train/grad_norm_pre_clip': 13.142268180847168, 'timing/data': 0.00513137795496732, 'timing/model': 1.2083880220307037, 'learning_rate/qwen_vl_interface': 1.7829608604698003e-06, 'learning_rate/action_model': 1.7829608604698e-05, 'epoch': 1.57}) +[2026-06-17 16:12:24,918][starVLA.training.train_starvla][INFO] - Step 5348, Loss: {'train/loss': 0.833952009677887, 'train/grad_norm_pre_clip': 16.022069931030273, 'timing/data': 0.0050458438927307725, 'timing/model': 1.3457252429798245, 'learning_rate/qwen_vl_interface': 1.7814829110767685e-06, 'learning_rate/action_model': 1.7814829110767685e-05, 'epoch': 1.57}) +[2026-06-17 16:12:29,596][starVLA.training.train_starvla][INFO] - Step 5349, Loss: {'train/loss': 0.7277193665504456, 'train/grad_norm_pre_clip': 14.170851707458496, 'timing/data': 0.005128666991367936, 'timing/model': 1.20449839590583, 'learning_rate/qwen_vl_interface': 1.7800056807110627e-06, 'learning_rate/action_model': 1.7800056807110626e-05, 'epoch': 1.57}) +[2026-06-17 16:12:34,399][starVLA.training.train_starvla][INFO] - Step 5350, Loss: {'train/loss': 0.7397496104240417, 'train/grad_norm_pre_clip': 15.216339111328125, 'timing/data': 0.004750638036057353, 'timing/model': 1.204198861028999, 'learning_rate/qwen_vl_interface': 1.7785291696789103e-06, 'learning_rate/action_model': 1.7785291696789102e-05, 'epoch': 1.57}) +[2026-06-17 16:12:39,188][starVLA.training.train_starvla][INFO] - Step 5351, Loss: {'train/loss': 0.7870063781738281, 'train/grad_norm_pre_clip': 20.587080001831055, 'timing/data': 0.0051051509799435735, 'timing/model': 1.204611344030127, 'learning_rate/qwen_vl_interface': 1.7770533782863983e-06, 'learning_rate/action_model': 1.7770533782863983e-05, 'epoch': 1.57}) +[2026-06-17 16:12:43,976][starVLA.training.train_starvla][INFO] - Step 5352, Loss: {'train/loss': 0.5954979062080383, 'train/grad_norm_pre_clip': 13.390334129333496, 'timing/data': 0.0050680870190262794, 'timing/model': 1.2217569049680606, 'learning_rate/qwen_vl_interface': 1.7755783068394578e-06, 'learning_rate/action_model': 1.7755783068394577e-05, 'epoch': 1.57}) +[2026-06-17 16:12:48,812][starVLA.training.train_starvla][INFO] - Step 5353, Loss: {'train/loss': 0.44365811347961426, 'train/grad_norm_pre_clip': 10.950712203979492, 'timing/data': 0.004829928046092391, 'timing/model': 1.222107339068316, 'learning_rate/qwen_vl_interface': 1.7741039556438733e-06, 'learning_rate/action_model': 1.7741039556438733e-05, 'epoch': 1.58}) +[2026-06-17 16:12:53,615][starVLA.training.train_starvla][INFO] - Step 5354, Loss: {'train/loss': 0.7872338891029358, 'train/grad_norm_pre_clip': 15.07271671295166, 'timing/data': 0.004696894087828696, 'timing/model': 1.3409220710163936, 'learning_rate/qwen_vl_interface': 1.7726303250052785e-06, 'learning_rate/action_model': 1.7726303250052784e-05, 'epoch': 1.58}) +[2026-06-17 16:12:58,281][starVLA.training.train_starvla][INFO] - Step 5355, Loss: {'train/loss': 0.4040023982524872, 'train/grad_norm_pre_clip': 9.563058853149414, 'timing/data': 0.004989909008145332, 'timing/model': 1.2144939550198615, 'learning_rate/qwen_vl_interface': 1.7711574152291591e-06, 'learning_rate/action_model': 1.771157415229159e-05, 'epoch': 1.58}) +[2026-06-17 16:13:03,110][starVLA.training.train_starvla][INFO] - Step 5356, Loss: {'train/loss': 0.5387550592422485, 'train/grad_norm_pre_clip': 16.38217544555664, 'timing/data': 0.005151182995177805, 'timing/model': 1.2056659259833395, 'learning_rate/qwen_vl_interface': 1.7696852266208512e-06, 'learning_rate/action_model': 1.7696852266208512e-05, 'epoch': 1.58}) +[2026-06-17 16:13:07,896][starVLA.training.train_starvla][INFO] - Step 5357, Loss: {'train/loss': 0.6332269906997681, 'train/grad_norm_pre_clip': 14.505952835083008, 'timing/data': 0.00478882307652384, 'timing/model': 1.2048942600376904, 'learning_rate/qwen_vl_interface': 1.7682137594855408e-06, 'learning_rate/action_model': 1.768213759485541e-05, 'epoch': 1.58}) +[2026-06-17 16:13:12,697][starVLA.training.train_starvla][INFO] - Step 5358, Loss: {'train/loss': 0.760424792766571, 'train/grad_norm_pre_clip': 17.856918334960938, 'timing/data': 0.004663257976062596, 'timing/model': 1.2038665839936584, 'learning_rate/qwen_vl_interface': 1.7667430141282649e-06, 'learning_rate/action_model': 1.7667430141282648e-05, 'epoch': 1.58}) +[2026-06-17 16:13:17,407][starVLA.training.train_starvla][INFO] - Step 5359, Loss: {'train/loss': 0.4214642643928528, 'train/grad_norm_pre_clip': 11.07175350189209, 'timing/data': 0.005433471989817917, 'timing/model': 1.205146071035415, 'learning_rate/qwen_vl_interface': 1.7652729908539101e-06, 'learning_rate/action_model': 1.76527299085391e-05, 'epoch': 1.58}) +[2026-06-17 16:13:22,115][starVLA.training.train_starvla][INFO] - Step 5360, Loss: {'train/loss': 0.6573073267936707, 'train/grad_norm_pre_clip': 23.929969787597656, 'timing/data': 0.0048819780349731445, 'timing/model': 1.243521602009423, 'learning_rate/qwen_vl_interface': 1.7638036899672137e-06, 'learning_rate/action_model': 1.7638036899672136e-05, 'epoch': 1.58}) +[2026-06-17 16:13:26,787][starVLA.training.train_starvla][INFO] - Step 5361, Loss: {'train/loss': 0.5644711256027222, 'train/grad_norm_pre_clip': 14.461127281188965, 'timing/data': 0.004574652994051576, 'timing/model': 1.2039660930167884, 'learning_rate/qwen_vl_interface': 1.7623351117727628e-06, 'learning_rate/action_model': 1.7623351117727626e-05, 'epoch': 1.58}) +[2026-06-17 16:13:31,552][starVLA.training.train_starvla][INFO] - Step 5362, Loss: {'train/loss': 0.727939784526825, 'train/grad_norm_pre_clip': 16.320362091064453, 'timing/data': 0.004695511073805392, 'timing/model': 1.2146412689471617, 'learning_rate/qwen_vl_interface': 1.7608672565749973e-06, 'learning_rate/action_model': 1.760867256574997e-05, 'epoch': 1.58}) +[2026-06-17 16:13:36,371][starVLA.training.train_starvla][INFO] - Step 5363, Loss: {'train/loss': 0.8501275181770325, 'train/grad_norm_pre_clip': 17.281991958618164, 'timing/data': 0.0049749709432944655, 'timing/model': 1.2048277539433911, 'learning_rate/qwen_vl_interface': 1.7594001246782039e-06, 'learning_rate/action_model': 1.7594001246782037e-05, 'epoch': 1.58}) +[2026-06-17 16:13:41,165][starVLA.training.train_starvla][INFO] - Step 5364, Loss: {'train/loss': 0.6298734545707703, 'train/grad_norm_pre_clip': 33.51513671875, 'timing/data': 0.004959804937243462, 'timing/model': 1.205314248916693, 'learning_rate/qwen_vl_interface': 1.7579337163865221e-06, 'learning_rate/action_model': 1.757933716386522e-05, 'epoch': 1.58}) +[2026-06-17 16:13:45,961][starVLA.training.train_starvla][INFO] - Step 5365, Loss: {'train/loss': 0.9092609882354736, 'train/grad_norm_pre_clip': 15.895049095153809, 'timing/data': 0.00507415400352329, 'timing/model': 1.333509838907048, 'learning_rate/qwen_vl_interface': 1.7564680320039356e-06, 'learning_rate/action_model': 1.7564680320039358e-05, 'epoch': 1.58}) +[2026-06-17 16:13:50,678][starVLA.training.train_starvla][INFO] - Step 5366, Loss: {'train/loss': 0.6044877767562866, 'train/grad_norm_pre_clip': 12.224602699279785, 'timing/data': 0.0050205959705635905, 'timing/model': 1.2525887180818245, 'learning_rate/qwen_vl_interface': 1.7550030718342867e-06, 'learning_rate/action_model': 1.7550030718342867e-05, 'epoch': 1.58}) +[2026-06-17 16:13:55,340][starVLA.training.train_starvla][INFO] - Step 5367, Loss: {'train/loss': 0.48165416717529297, 'train/grad_norm_pre_clip': 13.390266418457031, 'timing/data': 0.004879322019405663, 'timing/model': 1.2094924190314487, 'learning_rate/qwen_vl_interface': 1.753538836181262e-06, 'learning_rate/action_model': 1.753538836181262e-05, 'epoch': 1.58}) +[2026-06-17 16:14:00,106][starVLA.training.train_starvla][INFO] - Step 5368, Loss: {'train/loss': 0.6412912607192993, 'train/grad_norm_pre_clip': 16.628644943237305, 'timing/data': 0.0048706879606470466, 'timing/model': 1.2063590340549126, 'learning_rate/qwen_vl_interface': 1.7520753253483987e-06, 'learning_rate/action_model': 1.7520753253483986e-05, 'epoch': 1.58}) +[2026-06-17 16:14:04,899][starVLA.training.train_starvla][INFO] - Step 5369, Loss: {'train/loss': 0.896801233291626, 'train/grad_norm_pre_clip': 17.55621337890625, 'timing/data': 0.004899297025986016, 'timing/model': 1.2058471100172028, 'learning_rate/qwen_vl_interface': 1.7506125396390847e-06, 'learning_rate/action_model': 1.7506125396390847e-05, 'epoch': 1.58}) +[2026-06-17 16:14:09,688][starVLA.training.train_starvla][INFO] - Step 5370, Loss: {'train/loss': 0.4160289764404297, 'train/grad_norm_pre_clip': 12.1496000289917, 'timing/data': 0.004590380005538464, 'timing/model': 1.203398928977549, 'learning_rate/qwen_vl_interface': 1.7491504793565564e-06, 'learning_rate/action_model': 1.7491504793565562e-05, 'epoch': 1.58}) +[2026-06-17 16:14:14,476][starVLA.training.train_starvla][INFO] - Step 5371, Loss: {'train/loss': 0.6880775094032288, 'train/grad_norm_pre_clip': 13.23382568359375, 'timing/data': 0.004673090064898133, 'timing/model': 1.3343300729757175, 'learning_rate/qwen_vl_interface': 1.7476891448039012e-06, 'learning_rate/action_model': 1.7476891448039013e-05, 'epoch': 1.58}) +[2026-06-17 16:14:19,271][starVLA.training.train_starvla][INFO] - Step 5372, Loss: {'train/loss': 0.7949358820915222, 'train/grad_norm_pre_clip': 18.72742462158203, 'timing/data': 0.004810184007510543, 'timing/model': 1.332586682983674, 'learning_rate/qwen_vl_interface': 1.7462285362840534e-06, 'learning_rate/action_model': 1.746228536284053e-05, 'epoch': 1.58}) +[2026-06-17 16:14:23,944][starVLA.training.train_starvla][INFO] - Step 5373, Loss: {'train/loss': 0.5648897886276245, 'train/grad_norm_pre_clip': 24.38762664794922, 'timing/data': 0.0048024560092017055, 'timing/model': 1.2072601369582117, 'learning_rate/qwen_vl_interface': 1.7447686540998027e-06, 'learning_rate/action_model': 1.7447686540998026e-05, 'epoch': 1.58}) +[2026-06-17 16:14:28,737][starVLA.training.train_starvla][INFO] - Step 5374, Loss: {'train/loss': 0.6134851574897766, 'train/grad_norm_pre_clip': 12.838357925415039, 'timing/data': 0.0049247839488089085, 'timing/model': 1.209813334979117, 'learning_rate/qwen_vl_interface': 1.743309498553781e-06, 'learning_rate/action_model': 1.743309498553781e-05, 'epoch': 1.58}) +[2026-06-17 16:14:33,569][starVLA.training.train_starvla][INFO] - Step 5375, Loss: {'train/loss': 0.7509521842002869, 'train/grad_norm_pre_clip': 21.312524795532227, 'timing/data': 0.0049113130662590265, 'timing/model': 1.205370798939839, 'learning_rate/qwen_vl_interface': 1.7418510699484728e-06, 'learning_rate/action_model': 1.7418510699484727e-05, 'epoch': 1.58}) +[2026-06-17 16:14:38,364][starVLA.training.train_starvla][INFO] - Step 5376, Loss: {'train/loss': 0.5719114542007446, 'train/grad_norm_pre_clip': 22.36372184753418, 'timing/data': 0.004856999032199383, 'timing/model': 1.2052125699119642, 'learning_rate/qwen_vl_interface': 1.7403933685862118e-06, 'learning_rate/action_model': 1.7403933685862118e-05, 'epoch': 1.58}) +[2026-06-17 16:14:43,098][starVLA.training.train_starvla][INFO] - Step 5377, Loss: {'train/loss': 1.0303938388824463, 'train/grad_norm_pre_clip': 17.473583221435547, 'timing/data': 0.005066770943813026, 'timing/model': 1.2793068579630926, 'learning_rate/qwen_vl_interface': 1.7389363947691837e-06, 'learning_rate/action_model': 1.7389363947691838e-05, 'epoch': 1.58}) +[2026-06-17 16:14:47,760][starVLA.training.train_starvla][INFO] - Step 5378, Loss: {'train/loss': 0.6395478248596191, 'train/grad_norm_pre_clip': 12.768301963806152, 'timing/data': 0.0046902099857106805, 'timing/model': 1.2055624308995903, 'learning_rate/qwen_vl_interface': 1.7374801487994198e-06, 'learning_rate/action_model': 1.7374801487994198e-05, 'epoch': 1.58}) +[2026-06-17 16:14:52,524][starVLA.training.train_starvla][INFO] - Step 5379, Loss: {'train/loss': 0.736712634563446, 'train/grad_norm_pre_clip': 14.48629093170166, 'timing/data': 0.0048150940565392375, 'timing/model': 1.2068309099413455, 'learning_rate/qwen_vl_interface': 1.7360246309787982e-06, 'learning_rate/action_model': 1.736024630978798e-05, 'epoch': 1.58}) +[2026-06-17 16:14:57,317][starVLA.training.train_starvla][INFO] - Step 5380, Loss: {'train/loss': 0.6210366487503052, 'train/grad_norm_pre_clip': 17.882400512695312, 'timing/data': 0.004847920965403318, 'timing/model': 1.2048823270015419, 'learning_rate/qwen_vl_interface': 1.7345698416090534e-06, 'learning_rate/action_model': 1.7345698416090533e-05, 'epoch': 1.58}) +[2026-06-17 16:15:02,106][starVLA.training.train_starvla][INFO] - Step 5381, Loss: {'train/loss': 0.8707273006439209, 'train/grad_norm_pre_clip': 17.665613174438477, 'timing/data': 0.004926729016005993, 'timing/model': 1.2085273410193622, 'learning_rate/qwen_vl_interface': 1.7331157809917632e-06, 'learning_rate/action_model': 1.733115780991763e-05, 'epoch': 1.58}) +[2026-06-17 16:15:06,894][starVLA.training.train_starvla][INFO] - Step 5382, Loss: {'train/loss': 0.8075241446495056, 'train/grad_norm_pre_clip': 13.66814136505127, 'timing/data': 0.0048005880089476705, 'timing/model': 1.3333037019474432, 'learning_rate/qwen_vl_interface': 1.7316624494283557e-06, 'learning_rate/action_model': 1.7316624494283557e-05, 'epoch': 1.58}) +[2026-06-17 16:15:11,643][starVLA.training.train_starvla][INFO] - Step 5383, Loss: {'train/loss': 0.890434205532074, 'train/grad_norm_pre_clip': 18.887001037597656, 'timing/data': 0.005053768982179463, 'timing/model': 1.28162464289926, 'learning_rate/qwen_vl_interface': 1.7302098472201085e-06, 'learning_rate/action_model': 1.7302098472201084e-05, 'epoch': 1.58}) +[2026-06-17 16:15:16,308][starVLA.training.train_starvla][INFO] - Step 5384, Loss: {'train/loss': 0.5075066089630127, 'train/grad_norm_pre_clip': 13.39537525177002, 'timing/data': 0.005030623986385763, 'timing/model': 1.2054673669626936, 'learning_rate/qwen_vl_interface': 1.728757974668147e-06, 'learning_rate/action_model': 1.728757974668147e-05, 'epoch': 1.58}) +[2026-06-17 16:15:21,113][starVLA.training.train_starvla][INFO] - Step 5385, Loss: {'train/loss': 0.6404592394828796, 'train/grad_norm_pre_clip': 12.199846267700195, 'timing/data': 0.004924764041788876, 'timing/model': 1.2058441389817744, 'learning_rate/qwen_vl_interface': 1.727306832073447e-06, 'learning_rate/action_model': 1.7273068320734472e-05, 'epoch': 1.58}) +[2026-06-17 16:15:25,908][starVLA.training.train_starvla][INFO] - Step 5386, Loss: {'train/loss': 0.45778778195381165, 'train/grad_norm_pre_clip': 11.077219009399414, 'timing/data': 0.004623497021384537, 'timing/model': 1.2060821619816124, 'learning_rate/qwen_vl_interface': 1.7258564197368308e-06, 'learning_rate/action_model': 1.725856419736831e-05, 'epoch': 1.58}) +[2026-06-17 16:15:30,705][starVLA.training.train_starvla][INFO] - Step 5387, Loss: {'train/loss': 0.8528504967689514, 'train/grad_norm_pre_clip': 14.414510726928711, 'timing/data': 0.00500037707388401, 'timing/model': 1.2110620479797944, 'learning_rate/qwen_vl_interface': 1.7244067379589708e-06, 'learning_rate/action_model': 1.7244067379589707e-05, 'epoch': 1.59}) +[2026-06-17 16:15:35,477][starVLA.training.train_starvla][INFO] - Step 5388, Loss: {'train/loss': 0.6597883105278015, 'train/grad_norm_pre_clip': 16.686830520629883, 'timing/data': 0.004944299929775298, 'timing/model': 1.2057221799623221, 'learning_rate/qwen_vl_interface': 1.7229577870403886e-06, 'learning_rate/action_model': 1.7229577870403886e-05, 'epoch': 1.59}) +[2026-06-17 16:15:40,240][starVLA.training.train_starvla][INFO] - Step 5389, Loss: {'train/loss': 0.7250014543533325, 'train/grad_norm_pre_clip': 15.46316146850586, 'timing/data': 0.0049507130170241, 'timing/model': 1.3055585080292076, 'learning_rate/qwen_vl_interface': 1.7215095672814517e-06, 'learning_rate/action_model': 1.7215095672814516e-05, 'epoch': 1.59}) +[2026-06-17 16:15:44,902][starVLA.training.train_starvla][INFO] - Step 5390, Loss: {'train/loss': 0.812623918056488, 'train/grad_norm_pre_clip': 14.164039611816406, 'timing/data': 0.004844383103772998, 'timing/model': 1.2115395229775459, 'learning_rate/qwen_vl_interface': 1.7200620789823769e-06, 'learning_rate/action_model': 1.7200620789823767e-05, 'epoch': 1.59}) +[2026-06-17 16:15:49,700][starVLA.training.train_starvla][INFO] - Step 5391, Loss: {'train/loss': 0.9766431450843811, 'train/grad_norm_pre_clip': 15.854243278503418, 'timing/data': 0.004959307028912008, 'timing/model': 1.2100173949729651, 'learning_rate/qwen_vl_interface': 1.7186153224432331e-06, 'learning_rate/action_model': 1.718615322443233e-05, 'epoch': 1.59}) +[2026-06-17 16:15:54,498][starVLA.training.train_starvla][INFO] - Step 5392, Loss: {'train/loss': 0.6364530920982361, 'train/grad_norm_pre_clip': 12.462821960449219, 'timing/data': 0.004724885919131339, 'timing/model': 1.2084948240080848, 'learning_rate/qwen_vl_interface': 1.7171692979639342e-06, 'learning_rate/action_model': 1.717169297963934e-05, 'epoch': 1.59}) +[2026-06-17 16:15:59,289][starVLA.training.train_starvla][INFO] - Step 5393, Loss: {'train/loss': 0.8336083292961121, 'train/grad_norm_pre_clip': 12.224655151367188, 'timing/data': 0.004487185971811414, 'timing/model': 1.2061412601033226, 'learning_rate/qwen_vl_interface': 1.7157240058442397e-06, 'learning_rate/action_model': 1.71572400584424e-05, 'epoch': 1.59}) +[2026-06-17 16:16:04,078][starVLA.training.train_starvla][INFO] - Step 5394, Loss: {'train/loss': 0.528738260269165, 'train/grad_norm_pre_clip': 12.255891799926758, 'timing/data': 0.004788240999914706, 'timing/model': 1.3387860950315371, 'learning_rate/qwen_vl_interface': 1.7142794463837608e-06, 'learning_rate/action_model': 1.714279446383761e-05, 'epoch': 1.59}) +[2026-06-17 16:16:08,729][starVLA.training.train_starvla][INFO] - Step 5395, Loss: {'train/loss': 0.8409944772720337, 'train/grad_norm_pre_clip': 13.54340648651123, 'timing/data': 0.004741154960356653, 'timing/model': 1.2032250589691103, 'learning_rate/qwen_vl_interface': 1.712835619881959e-06, 'learning_rate/action_model': 1.7128356198819588e-05, 'epoch': 1.59}) +[2026-06-17 16:16:13,513][starVLA.training.train_starvla][INFO] - Step 5396, Loss: {'train/loss': 0.3917733430862427, 'train/grad_norm_pre_clip': 9.39309024810791, 'timing/data': 0.005120546906255186, 'timing/model': 1.205556349013932, 'learning_rate/qwen_vl_interface': 1.7113925266381383e-06, 'learning_rate/action_model': 1.7113925266381383e-05, 'epoch': 1.59}) +[2026-06-17 16:16:18,312][starVLA.training.train_starvla][INFO] - Step 5397, Loss: {'train/loss': 0.7085234522819519, 'train/grad_norm_pre_clip': 16.156904220581055, 'timing/data': 0.0044167389860376716, 'timing/model': 1.212368986918591, 'learning_rate/qwen_vl_interface': 1.7099501669514569e-06, 'learning_rate/action_model': 1.7099501669514568e-05, 'epoch': 1.59}) +[2026-06-17 16:16:23,038][starVLA.training.train_starvla][INFO] - Step 5398, Loss: {'train/loss': 0.6940785646438599, 'train/grad_norm_pre_clip': 14.805355072021484, 'timing/data': 0.005560591001994908, 'timing/model': 1.2066982029937208, 'learning_rate/qwen_vl_interface': 1.7085085411209118e-06, 'learning_rate/action_model': 1.7085085411209118e-05, 'epoch': 1.59}) +[2026-06-17 16:16:27,785][starVLA.training.train_starvla][INFO] - Step 5399, Loss: {'train/loss': 0.6873961091041565, 'train/grad_norm_pre_clip': 18.744298934936523, 'timing/data': 0.004963150015100837, 'timing/model': 1.203970742993988, 'learning_rate/qwen_vl_interface': 1.7070676494453587e-06, 'learning_rate/action_model': 1.7070676494453586e-05, 'epoch': 1.59}) +[2026-06-17 16:18:09,741][starVLA.training.train_starvla][INFO] - Step 5400, Loss: {'train/loss': 0.7365301251411438, 'train/grad_norm_pre_clip': 16.83169937133789, 'eval/loss': 0.7086973786354065, 'eval/latency_0/loss': 0.7086973786354065, 'eval/demon_fire_precision': 0.9015273311897106, 'eval/demon_fire_recall': 0.9314784053156147, 'eval/demon_fire_f1': 0.9162581699346406, 'eval/demon_fire_rate_model': 0.6918798665183538, 'eval/demon_fire_rate_teacher': 0.6696329254727474, 'eval/demon_left_precision': 0.8377896613190731, 'eval/demon_left_recall': 0.7988668555240793, 'eval/demon_left_f1': 0.8178654292343387, 'eval/demon_left_rate_model': 0.46802002224694106, 'eval/demon_left_rate_teacher': 0.49082313681868744, 'eval/demon_right_precision': 0.8105656350053362, 'eval/demon_right_recall': 0.8581920903954803, 'eval/demon_right_f1': 0.8336992316136115, 'eval/demon_right_rate_model': 0.521134593993326, 'eval/demon_right_rate_teacher': 0.4922135706340378, 'eval/demon_move_f1': 0.8257823304239751, 'eval/demon_cc_f1': 0.8870999605481791, 'eval/latency_0/demon_fire_precision': 0.9015273311897106, 'eval/latency_0/demon_fire_recall': 0.9314784053156147, 'eval/latency_0/demon_fire_f1': 0.9162581699346406, 'eval/latency_0/demon_fire_rate_model': 0.6918798665183538, 'eval/latency_0/demon_fire_rate_teacher': 0.6696329254727474, 'eval/latency_0/demon_left_precision': 0.8377896613190731, 'eval/latency_0/demon_left_recall': 0.7988668555240793, 'eval/latency_0/demon_left_f1': 0.8178654292343387, 'eval/latency_0/demon_left_rate_model': 0.46802002224694106, 'eval/latency_0/demon_left_rate_teacher': 0.49082313681868744, 'eval/latency_0/demon_right_precision': 0.8105656350053362, 'eval/latency_0/demon_right_recall': 0.8581920903954803, 'eval/latency_0/demon_right_f1': 0.8336992316136115, 'eval/latency_0/demon_right_rate_model': 0.521134593993326, 'eval/latency_0/demon_right_rate_teacher': 0.4922135706340378, 'eval/latency_0/demon_move_f1': 0.8257823304239751, 'eval/latency_0/demon_cc_f1': 0.8870999605481791, 'timing/data': 0.005059738061390817, 'timing/model': 1.3526624629739672, 'learning_rate/qwen_vl_interface': 1.705627492223494e-06, 'learning_rate/action_model': 1.705627492223494e-05, 'epoch': 1.59}) +[2026-06-17 16:18:09,741][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_5400_state +[2026-06-17 16:18:09,744][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 16:18:13,252][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_5400_state/model.safetensors +[2026-06-17 16:18:20,604][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_5400_state/optimizer.bin +[2026-06-17 16:18:20,604][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_5400_state/sampler.bin +[2026-06-17 16:18:20,604][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_5400_state/sampler_1.bin +[2026-06-17 16:18:20,605][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_5400_state/random_states_0.pkl +[2026-06-17 16:18:20,651][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 5400 +[2026-06-17 16:18:20,651][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 16:18:20,666][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 16:18:25,417][starVLA.training.train_starvla][INFO] - Step 5401, Loss: {'train/loss': 0.6749938130378723, 'train/grad_norm_pre_clip': 12.63664722442627, 'timing/data': 0.004871784010902047, 'timing/model': 1.3150815339758992, 'learning_rate/qwen_vl_interface': 1.7041880697538644e-06, 'learning_rate/action_model': 1.704188069753864e-05, 'epoch': 1.59}) +[2026-06-17 16:18:30,119][starVLA.training.train_starvla][INFO] - Step 5402, Loss: {'train/loss': 0.5841367840766907, 'train/grad_norm_pre_clip': 13.863546371459961, 'timing/data': 0.005135732004418969, 'timing/model': 1.2969628280261531, 'learning_rate/qwen_vl_interface': 1.702749382334863e-06, 'learning_rate/action_model': 1.7027493823348632e-05, 'epoch': 1.59}) +[2026-06-17 16:18:34,734][starVLA.training.train_starvla][INFO] - Step 5403, Loss: {'train/loss': 0.65287184715271, 'train/grad_norm_pre_clip': 12.937661170959473, 'timing/data': 0.00494843500200659, 'timing/model': 1.1966081040445715, 'learning_rate/qwen_vl_interface': 1.7013114302647318e-06, 'learning_rate/action_model': 1.7013114302647318e-05, 'epoch': 1.59}) +[2026-06-17 16:18:39,488][starVLA.training.train_starvla][INFO] - Step 5404, Loss: {'train/loss': 0.9319041967391968, 'train/grad_norm_pre_clip': 20.076190948486328, 'timing/data': 0.005138816894032061, 'timing/model': 1.201348840026185, 'learning_rate/qwen_vl_interface': 1.69987421384156e-06, 'learning_rate/action_model': 1.69987421384156e-05, 'epoch': 1.59}) +[2026-06-17 16:18:44,264][starVLA.training.train_starvla][INFO] - Step 5405, Loss: {'train/loss': 0.5694847702980042, 'train/grad_norm_pre_clip': 12.3052396774292, 'timing/data': 0.00465341005474329, 'timing/model': 1.195117054041475, 'learning_rate/qwen_vl_interface': 1.6984377333632815e-06, 'learning_rate/action_model': 1.6984377333632813e-05, 'epoch': 1.59}) +[2026-06-17 16:18:49,011][starVLA.training.train_starvla][INFO] - Step 5406, Loss: {'train/loss': 0.6375331878662109, 'train/grad_norm_pre_clip': 26.06329345703125, 'timing/data': 0.005309924017637968, 'timing/model': 1.1969309689011425, 'learning_rate/qwen_vl_interface': 1.697001989127686e-06, 'learning_rate/action_model': 1.697001989127686e-05, 'epoch': 1.59}) +[2026-06-17 16:18:53,720][starVLA.training.train_starvla][INFO] - Step 5407, Loss: {'train/loss': 0.4970633089542389, 'train/grad_norm_pre_clip': 12.634744644165039, 'timing/data': 0.0046801280695945024, 'timing/model': 1.1978492320049554, 'learning_rate/qwen_vl_interface': 1.6955669814324003e-06, 'learning_rate/action_model': 1.6955669814324e-05, 'epoch': 1.59}) +[2026-06-17 16:18:58,473][starVLA.training.train_starvla][INFO] - Step 5408, Loss: {'train/loss': 0.5917602777481079, 'train/grad_norm_pre_clip': 15.197943687438965, 'timing/data': 0.004673412069678307, 'timing/model': 1.3262161569437012, 'learning_rate/qwen_vl_interface': 1.6941327105749034e-06, 'learning_rate/action_model': 1.694132710574903e-05, 'epoch': 1.59}) +[2026-06-17 16:19:03,133][starVLA.training.train_starvla][INFO] - Step 5409, Loss: {'train/loss': 0.7351006865501404, 'train/grad_norm_pre_clip': 14.400537490844727, 'timing/data': 0.005097936955280602, 'timing/model': 1.200618181959726, 'learning_rate/qwen_vl_interface': 1.6926991768525205e-06, 'learning_rate/action_model': 1.6926991768525205e-05, 'epoch': 1.59}) +[2026-06-17 16:19:07,908][starVLA.training.train_starvla][INFO] - Step 5410, Loss: {'train/loss': 0.6714743375778198, 'train/grad_norm_pre_clip': 13.985191345214844, 'timing/data': 0.004743720055557787, 'timing/model': 1.1996210559736937, 'learning_rate/qwen_vl_interface': 1.6912663805624278e-06, 'learning_rate/action_model': 1.6912663805624275e-05, 'epoch': 1.59}) +[2026-06-17 16:19:12,687][starVLA.training.train_starvla][INFO] - Step 5411, Loss: {'train/loss': 0.6247425675392151, 'train/grad_norm_pre_clip': 15.047404289245605, 'timing/data': 0.005080207949504256, 'timing/model': 1.2093239299720153, 'learning_rate/qwen_vl_interface': 1.6898343220016446e-06, 'learning_rate/action_model': 1.6898343220016447e-05, 'epoch': 1.59}) +[2026-06-17 16:19:17,461][starVLA.training.train_starvla][INFO] - Step 5412, Loss: {'train/loss': 0.727796196937561, 'train/grad_norm_pre_clip': 13.114313125610352, 'timing/data': 0.0044744720216840506, 'timing/model': 1.2044898649910465, 'learning_rate/qwen_vl_interface': 1.6884030014670353e-06, 'learning_rate/action_model': 1.688403001467035e-05, 'epoch': 1.59}) +[2026-06-17 16:19:22,185][starVLA.training.train_starvla][INFO] - Step 5413, Loss: {'train/loss': 0.7419353127479553, 'train/grad_norm_pre_clip': 13.763710975646973, 'timing/data': 0.004723611054942012, 'timing/model': 1.2004788399208337, 'learning_rate/qwen_vl_interface': 1.6869724192553168e-06, 'learning_rate/action_model': 1.686972419255317e-05, 'epoch': 1.59}) +[2026-06-17 16:19:26,938][starVLA.training.train_starvla][INFO] - Step 5414, Loss: {'train/loss': 0.6264652609825134, 'train/grad_norm_pre_clip': 17.285409927368164, 'timing/data': 0.004609851981513202, 'timing/model': 1.315356933977455, 'learning_rate/qwen_vl_interface': 1.6855425756630506e-06, 'learning_rate/action_model': 1.6855425756630506e-05, 'epoch': 1.59}) +[2026-06-17 16:19:31,583][starVLA.training.train_starvla][INFO] - Step 5415, Loss: {'train/loss': 0.45305755734443665, 'train/grad_norm_pre_clip': 10.96995735168457, 'timing/data': 0.0047078890493139625, 'timing/model': 1.2013448500074446, 'learning_rate/qwen_vl_interface': 1.684113470986644e-06, 'learning_rate/action_model': 1.6841134709866438e-05, 'epoch': 1.59}) +[2026-06-17 16:19:36,370][starVLA.training.train_starvla][INFO] - Step 5416, Loss: {'train/loss': 0.6290469765663147, 'train/grad_norm_pre_clip': 16.622648239135742, 'timing/data': 0.004946253960952163, 'timing/model': 1.2035635629436001, 'learning_rate/qwen_vl_interface': 1.682685105522351e-06, 'learning_rate/action_model': 1.682685105522351e-05, 'epoch': 1.59}) +[2026-06-17 16:19:41,149][starVLA.training.train_starvla][INFO] - Step 5417, Loss: {'train/loss': 0.4858856499195099, 'train/grad_norm_pre_clip': 11.249738693237305, 'timing/data': 0.004993755021132529, 'timing/model': 1.2032545990077779, 'learning_rate/qwen_vl_interface': 1.6812574795662751e-06, 'learning_rate/action_model': 1.681257479566275e-05, 'epoch': 1.59}) +[2026-06-17 16:19:46,000][starVLA.training.train_starvla][INFO] - Step 5418, Loss: {'train/loss': 0.5795515179634094, 'train/grad_norm_pre_clip': 13.85819149017334, 'timing/data': 0.004772669053636491, 'timing/model': 1.2031424660235643, 'learning_rate/qwen_vl_interface': 1.6798305934143635e-06, 'learning_rate/action_model': 1.6798305934143633e-05, 'epoch': 1.59}) +[2026-06-17 16:19:50,787][starVLA.training.train_starvla][INFO] - Step 5419, Loss: {'train/loss': 0.5599275827407837, 'train/grad_norm_pre_clip': 14.8029203414917, 'timing/data': 0.005009204964153469, 'timing/model': 1.3386772410012782, 'learning_rate/qwen_vl_interface': 1.6784044473624118e-06, 'learning_rate/action_model': 1.6784044473624118e-05, 'epoch': 1.59}) +[2026-06-17 16:19:55,438][starVLA.training.train_starvla][INFO] - Step 5420, Loss: {'train/loss': 0.6467117667198181, 'train/grad_norm_pre_clip': 14.074640274047852, 'timing/data': 0.004631996038369834, 'timing/model': 1.203481761040166, 'learning_rate/qwen_vl_interface': 1.6769790417060611e-06, 'learning_rate/action_model': 1.676979041706061e-05, 'epoch': 1.59}) +[2026-06-17 16:20:00,211][starVLA.training.train_starvla][INFO] - Step 5421, Loss: {'train/loss': 0.6612534523010254, 'train/grad_norm_pre_clip': 12.571942329406738, 'timing/data': 0.005065744975581765, 'timing/model': 1.20145680999849, 'learning_rate/qwen_vl_interface': 1.6755543767407998e-06, 'learning_rate/action_model': 1.6755543767407998e-05, 'epoch': 1.6}) +[2026-06-17 16:20:04,923][starVLA.training.train_starvla][INFO] - Step 5422, Loss: {'train/loss': 0.5625224709510803, 'train/grad_norm_pre_clip': 11.435970306396484, 'timing/data': 0.0049604319501668215, 'timing/model': 1.2132194059668109, 'learning_rate/qwen_vl_interface': 1.6741304527619624e-06, 'learning_rate/action_model': 1.674130452761962e-05, 'epoch': 1.6}) +[2026-06-17 16:20:09,666][starVLA.training.train_starvla][INFO] - Step 5423, Loss: {'train/loss': 0.5814223885536194, 'train/grad_norm_pre_clip': 12.45909309387207, 'timing/data': 0.004842223017476499, 'timing/model': 1.2056733201025054, 'learning_rate/qwen_vl_interface': 1.672707270064729e-06, 'learning_rate/action_model': 1.6727072700647288e-05, 'epoch': 1.6}) +[2026-06-17 16:20:14,448][starVLA.training.train_starvla][INFO] - Step 5424, Loss: {'train/loss': 0.5838083028793335, 'train/grad_norm_pre_clip': 12.609589576721191, 'timing/data': 0.004710969980806112, 'timing/model': 1.210284200962633, 'learning_rate/qwen_vl_interface': 1.6712848289441288e-06, 'learning_rate/action_model': 1.6712848289441287e-05, 'epoch': 1.6}) +[2026-06-17 16:20:19,229][starVLA.training.train_starvla][INFO] - Step 5425, Loss: {'train/loss': 0.8661714196205139, 'train/grad_norm_pre_clip': 15.293890953063965, 'timing/data': 0.004837887012399733, 'timing/model': 1.3349290949990973, 'learning_rate/qwen_vl_interface': 1.669863129695036e-06, 'learning_rate/action_model': 1.669863129695036e-05, 'epoch': 1.6}) +[2026-06-17 16:20:23,881][starVLA.training.train_starvla][INFO] - Step 5426, Loss: {'train/loss': 0.612855851650238, 'train/grad_norm_pre_clip': 13.810464859008789, 'timing/data': 0.004785227007232606, 'timing/model': 1.2033145519671962, 'learning_rate/qwen_vl_interface': 1.6684421726121675e-06, 'learning_rate/action_model': 1.6684421726121674e-05, 'epoch': 1.6}) +[2026-06-17 16:20:28,660][starVLA.training.train_starvla][INFO] - Step 5427, Loss: {'train/loss': 0.7804661989212036, 'train/grad_norm_pre_clip': 19.74730682373047, 'timing/data': 0.005004269885830581, 'timing/model': 1.2049468730110675, 'learning_rate/qwen_vl_interface': 1.6670219579900888e-06, 'learning_rate/action_model': 1.6670219579900887e-05, 'epoch': 1.6}) +[2026-06-17 16:20:33,443][starVLA.training.train_starvla][INFO] - Step 5428, Loss: {'train/loss': 0.44224023818969727, 'train/grad_norm_pre_clip': 10.775325775146484, 'timing/data': 0.004786340985447168, 'timing/model': 1.209862349089235, 'learning_rate/qwen_vl_interface': 1.665602486123215e-06, 'learning_rate/action_model': 1.665602486123215e-05, 'epoch': 1.6}) +[2026-06-17 16:20:38,193][starVLA.training.train_starvla][INFO] - Step 5429, Loss: {'train/loss': 0.768419623374939, 'train/grad_norm_pre_clip': 14.47852897644043, 'timing/data': 0.004608406103216112, 'timing/model': 1.2090597619535401, 'learning_rate/qwen_vl_interface': 1.6641837573058034e-06, 'learning_rate/action_model': 1.6641837573058035e-05, 'epoch': 1.6}) +[2026-06-17 16:20:42,976][starVLA.training.train_starvla][INFO] - Step 5430, Loss: {'train/loss': 0.634285569190979, 'train/grad_norm_pre_clip': 19.30010223388672, 'timing/data': 0.00447390996851027, 'timing/model': 1.207456400967203, 'learning_rate/qwen_vl_interface': 1.6627657718319568e-06, 'learning_rate/action_model': 1.6627657718319568e-05, 'epoch': 1.6}) +[2026-06-17 16:20:47,765][starVLA.training.train_starvla][INFO] - Step 5431, Loss: {'train/loss': 0.8492186665534973, 'train/grad_norm_pre_clip': 12.894761085510254, 'timing/data': 0.004912370932288468, 'timing/model': 1.3362496489426121, 'learning_rate/qwen_vl_interface': 1.6613485299956265e-06, 'learning_rate/action_model': 1.6613485299956262e-05, 'epoch': 1.6}) +[2026-06-17 16:20:52,435][starVLA.training.train_starvla][INFO] - Step 5432, Loss: {'train/loss': 0.6166771650314331, 'train/grad_norm_pre_clip': 15.141225814819336, 'timing/data': 0.00504472607281059, 'timing/model': 1.20673377008643, 'learning_rate/qwen_vl_interface': 1.6599320320906064e-06, 'learning_rate/action_model': 1.6599320320906066e-05, 'epoch': 1.6}) +[2026-06-17 16:20:57,239][starVLA.training.train_starvla][INFO] - Step 5433, Loss: {'train/loss': 0.5907763838768005, 'train/grad_norm_pre_clip': 12.886373519897461, 'timing/data': 0.005271910922601819, 'timing/model': 1.2043567090295255, 'learning_rate/qwen_vl_interface': 1.65851627841054e-06, 'learning_rate/action_model': 1.65851627841054e-05, 'epoch': 1.6}) +[2026-06-17 16:21:02,021][starVLA.training.train_starvla][INFO] - Step 5434, Loss: {'train/loss': 1.0434505939483643, 'train/grad_norm_pre_clip': 15.330795288085938, 'timing/data': 0.0050162909319624305, 'timing/model': 1.2047550519928336, 'learning_rate/qwen_vl_interface': 1.657101269248913e-06, 'learning_rate/action_model': 1.657101269248913e-05, 'epoch': 1.6}) +[2026-06-17 16:21:06,807][starVLA.training.train_starvla][INFO] - Step 5435, Loss: {'train/loss': 0.8429075479507446, 'train/grad_norm_pre_clip': 14.63249397277832, 'timing/data': 0.0049719869857653975, 'timing/model': 1.2054205790627748, 'learning_rate/qwen_vl_interface': 1.65568700489906e-06, 'learning_rate/action_model': 1.65568700489906e-05, 'epoch': 1.6}) +[2026-06-17 16:21:11,494][starVLA.training.train_starvla][INFO] - Step 5436, Loss: {'train/loss': 0.6578016877174377, 'train/grad_norm_pre_clip': 18.337785720825195, 'timing/data': 0.0045941389398649335, 'timing/model': 1.2087230320321396, 'learning_rate/qwen_vl_interface': 1.6542734856541577e-06, 'learning_rate/action_model': 1.6542734856541576e-05, 'epoch': 1.6}) +[2026-06-17 16:21:16,278][starVLA.training.train_starvla][INFO] - Step 5437, Loss: {'train/loss': 0.7714235782623291, 'train/grad_norm_pre_clip': 32.70408630371094, 'timing/data': 0.004405701998621225, 'timing/model': 1.2471606669714674, 'learning_rate/qwen_vl_interface': 1.652860711807231e-06, 'learning_rate/action_model': 1.652860711807231e-05, 'epoch': 1.6}) +[2026-06-17 16:21:20,944][starVLA.training.train_starvla][INFO] - Step 5438, Loss: {'train/loss': 0.5319757461547852, 'train/grad_norm_pre_clip': 13.972412109375, 'timing/data': 0.004774410044774413, 'timing/model': 1.2056247879518196, 'learning_rate/qwen_vl_interface': 1.6514486836511484e-06, 'learning_rate/action_model': 1.6514486836511484e-05, 'epoch': 1.6}) +[2026-06-17 16:21:25,684][starVLA.training.train_starvla][INFO] - Step 5439, Loss: {'train/loss': 0.48446133732795715, 'train/grad_norm_pre_clip': 11.889792442321777, 'timing/data': 0.004695569048635662, 'timing/model': 1.2098999659065157, 'learning_rate/qwen_vl_interface': 1.6500374014786288e-06, 'learning_rate/action_model': 1.650037401478629e-05, 'epoch': 1.6}) +[2026-06-17 16:21:30,468][starVLA.training.train_starvla][INFO] - Step 5440, Loss: {'train/loss': 0.5820774435997009, 'train/grad_norm_pre_clip': 13.707015991210938, 'timing/data': 0.004511920036748052, 'timing/model': 1.2055496200919151, 'learning_rate/qwen_vl_interface': 1.6486268655822277e-06, 'learning_rate/action_model': 1.6486268655822277e-05, 'epoch': 1.6}) +[2026-06-17 16:21:35,329][starVLA.training.train_starvla][INFO] - Step 5441, Loss: {'train/loss': 0.39874738454818726, 'train/grad_norm_pre_clip': 12.4109525680542, 'timing/data': 0.004898531944490969, 'timing/model': 1.2056606860132888, 'learning_rate/qwen_vl_interface': 1.6472170762543515e-06, 'learning_rate/action_model': 1.6472170762543514e-05, 'epoch': 1.6}) +[2026-06-17 16:21:40,124][starVLA.training.train_starvla][INFO] - Step 5442, Loss: {'train/loss': 0.4484107494354248, 'train/grad_norm_pre_clip': 11.046013832092285, 'timing/data': 0.0049486770294606686, 'timing/model': 1.341614683973603, 'learning_rate/qwen_vl_interface': 1.6458080337872534e-06, 'learning_rate/action_model': 1.6458080337872535e-05, 'epoch': 1.6}) +[2026-06-17 16:21:44,832][starVLA.training.train_starvla][INFO] - Step 5443, Loss: {'train/loss': 0.4625026285648346, 'train/grad_norm_pre_clip': 11.479628562927246, 'timing/data': 0.004829342011362314, 'timing/model': 1.2525138600030914, 'learning_rate/qwen_vl_interface': 1.644399738473029e-06, 'learning_rate/action_model': 1.644399738473029e-05, 'epoch': 1.6}) +[2026-06-17 16:21:49,495][starVLA.training.train_starvla][INFO] - Step 5444, Loss: {'train/loss': 0.6892857551574707, 'train/grad_norm_pre_clip': 12.021265029907227, 'timing/data': 0.004677603021264076, 'timing/model': 1.2069411840057, 'learning_rate/qwen_vl_interface': 1.6429921906036175e-06, 'learning_rate/action_model': 1.6429921906036176e-05, 'epoch': 1.6}) +[2026-06-17 16:21:54,252][starVLA.training.train_starvla][INFO] - Step 5445, Loss: {'train/loss': 0.5682889223098755, 'train/grad_norm_pre_clip': 27.594757080078125, 'timing/data': 0.0051269709365442395, 'timing/model': 1.210308001958765, 'learning_rate/qwen_vl_interface': 1.6415853904708068e-06, 'learning_rate/action_model': 1.6415853904708066e-05, 'epoch': 1.6}) +[2026-06-17 16:21:59,055][starVLA.training.train_starvla][INFO] - Step 5446, Loss: {'train/loss': 0.6469742059707642, 'train/grad_norm_pre_clip': 12.665815353393555, 'timing/data': 0.004799210000783205, 'timing/model': 1.2061260050395504, 'learning_rate/qwen_vl_interface': 1.6401793383662274e-06, 'learning_rate/action_model': 1.6401793383662273e-05, 'epoch': 1.6}) +[2026-06-17 16:22:03,851][starVLA.training.train_starvla][INFO] - Step 5447, Loss: {'train/loss': 0.3618645668029785, 'train/grad_norm_pre_clip': 9.887803077697754, 'timing/data': 0.004683075938373804, 'timing/model': 1.2052865780424327, 'learning_rate/qwen_vl_interface': 1.6387740345813553e-06, 'learning_rate/action_model': 1.6387740345813552e-05, 'epoch': 1.6}) +[2026-06-17 16:22:08,639][starVLA.training.train_starvla][INFO] - Step 5448, Loss: {'train/loss': 0.529516339302063, 'train/grad_norm_pre_clip': 12.149048805236816, 'timing/data': 0.005160240107215941, 'timing/model': 1.3317725009983405, 'learning_rate/qwen_vl_interface': 1.6373694794075115e-06, 'learning_rate/action_model': 1.6373694794075115e-05, 'epoch': 1.6}) +[2026-06-17 16:22:13,419][starVLA.training.train_starvla][INFO] - Step 5449, Loss: {'train/loss': 0.682694673538208, 'train/grad_norm_pre_clip': 13.168131828308105, 'timing/data': 0.004470420069992542, 'timing/model': 1.3243816189933568, 'learning_rate/qwen_vl_interface': 1.6359656731358626e-06, 'learning_rate/action_model': 1.6359656731358626e-05, 'epoch': 1.6}) +[2026-06-17 16:22:18,079][starVLA.training.train_starvla][INFO] - Step 5450, Loss: {'train/loss': 0.6117814183235168, 'train/grad_norm_pre_clip': 15.214154243469238, 'timing/data': 0.0049714320339262486, 'timing/model': 1.2056771699571982, 'learning_rate/qwen_vl_interface': 1.634562616057418e-06, 'learning_rate/action_model': 1.634562616057418e-05, 'epoch': 1.6}) +[2026-06-17 16:22:22,842][starVLA.training.train_starvla][INFO] - Step 5451, Loss: {'train/loss': 0.8576142191886902, 'train/grad_norm_pre_clip': 16.088071823120117, 'timing/data': 0.00446908304002136, 'timing/model': 1.2070615319535136, 'learning_rate/qwen_vl_interface': 1.6331603084630338e-06, 'learning_rate/action_model': 1.6331603084630338e-05, 'epoch': 1.6}) +[2026-06-17 16:22:27,637][starVLA.training.train_starvla][INFO] - Step 5452, Loss: {'train/loss': 0.5763430595397949, 'train/grad_norm_pre_clip': 14.335491180419922, 'timing/data': 0.005099270027130842, 'timing/model': 1.206733335973695, 'learning_rate/qwen_vl_interface': 1.6317587506434076e-06, 'learning_rate/action_model': 1.6317587506434076e-05, 'epoch': 1.6}) +[2026-06-17 16:22:32,443][starVLA.training.train_starvla][INFO] - Step 5453, Loss: {'train/loss': 0.5137300491333008, 'train/grad_norm_pre_clip': 43.24917984008789, 'timing/data': 0.005116023006848991, 'timing/model': 1.2049457220127806, 'learning_rate/qwen_vl_interface': 1.6303579428890881e-06, 'learning_rate/action_model': 1.630357942889088e-05, 'epoch': 1.6}) +[2026-06-17 16:22:37,166][starVLA.training.train_starvla][INFO] - Step 5454, Loss: {'train/loss': 0.5944254398345947, 'train/grad_norm_pre_clip': 47.09038162231445, 'timing/data': 0.004835910978727043, 'timing/model': 1.272022319957614, 'learning_rate/qwen_vl_interface': 1.6289578854904626e-06, 'learning_rate/action_model': 1.6289578854904625e-05, 'epoch': 1.6}) +[2026-06-17 16:22:41,819][starVLA.training.train_starvla][INFO] - Step 5455, Loss: {'train/loss': 0.7321220636367798, 'train/grad_norm_pre_clip': 13.645136833190918, 'timing/data': 0.004874931066296995, 'timing/model': 1.2052102190209553, 'learning_rate/qwen_vl_interface': 1.6275585787377627e-06, 'learning_rate/action_model': 1.6275585787377626e-05, 'epoch': 1.61}) +[2026-06-17 16:22:46,595][starVLA.training.train_starvla][INFO] - Step 5456, Loss: {'train/loss': 0.6197189688682556, 'train/grad_norm_pre_clip': 13.041985511779785, 'timing/data': 0.00485364010091871, 'timing/model': 1.2069684599991888, 'learning_rate/qwen_vl_interface': 1.6261600229210662e-06, 'learning_rate/action_model': 1.626160022921066e-05, 'epoch': 1.61}) +[2026-06-17 16:22:51,381][starVLA.training.train_starvla][INFO] - Step 5457, Loss: {'train/loss': 0.5501899123191833, 'train/grad_norm_pre_clip': 24.215314865112305, 'timing/data': 0.004841316025704145, 'timing/model': 1.2050742390565574, 'learning_rate/qwen_vl_interface': 1.6247622183302972e-06, 'learning_rate/action_model': 1.624762218330297e-05, 'epoch': 1.61}) +[2026-06-17 16:22:56,169][starVLA.training.train_starvla][INFO] - Step 5458, Loss: {'train/loss': 0.5847524404525757, 'train/grad_norm_pre_clip': 13.792485237121582, 'timing/data': 0.005396798951551318, 'timing/model': 1.2068412409862503, 'learning_rate/qwen_vl_interface': 1.6233651652552218e-06, 'learning_rate/action_model': 1.6233651652552217e-05, 'epoch': 1.61}) +[2026-06-17 16:23:00,972][starVLA.training.train_starvla][INFO] - Step 5459, Loss: {'train/loss': 0.896304726600647, 'train/grad_norm_pre_clip': 15.511775016784668, 'timing/data': 0.004904962028376758, 'timing/model': 1.3415450840257108, 'learning_rate/qwen_vl_interface': 1.6219688639854507e-06, 'learning_rate/action_model': 1.6219688639854506e-05, 'epoch': 1.61}) +[2026-06-17 16:23:05,709][starVLA.training.train_starvla][INFO] - Step 5460, Loss: {'train/loss': 0.5657749176025391, 'train/grad_norm_pre_clip': 12.082500457763672, 'timing/data': 0.004796023014932871, 'timing/model': 1.2829661139985546, 'learning_rate/qwen_vl_interface': 1.620573314810436e-06, 'learning_rate/action_model': 1.620573314810436e-05, 'epoch': 1.61}) +[2026-06-17 16:23:10,376][starVLA.training.train_starvla][INFO] - Step 5461, Loss: {'train/loss': 0.6011507511138916, 'train/grad_norm_pre_clip': 15.04458236694336, 'timing/data': 0.0055587380193173885, 'timing/model': 1.2104753060266376, 'learning_rate/qwen_vl_interface': 1.6191785180194797e-06, 'learning_rate/action_model': 1.6191785180194796e-05, 'epoch': 1.61}) +[2026-06-17 16:23:15,298][starVLA.training.train_starvla][INFO] - Step 5462, Loss: {'train/loss': 0.6487990617752075, 'train/grad_norm_pre_clip': 13.190103530883789, 'timing/data': 0.005217202007770538, 'timing/model': 1.2039752920391038, 'learning_rate/qwen_vl_interface': 1.6177844739017236e-06, 'learning_rate/action_model': 1.6177844739017236e-05, 'epoch': 1.61}) +[2026-06-17 16:23:20,109][starVLA.training.train_starvla][INFO] - Step 5463, Loss: {'train/loss': 0.5679774284362793, 'train/grad_norm_pre_clip': 13.931703567504883, 'timing/data': 0.005125603987835348, 'timing/model': 1.2205207630759105, 'learning_rate/qwen_vl_interface': 1.6163911827461546e-06, 'learning_rate/action_model': 1.6163911827461546e-05, 'epoch': 1.61}) +[2026-06-17 16:23:24,896][starVLA.training.train_starvla][INFO] - Step 5464, Loss: {'train/loss': 0.709024965763092, 'train/grad_norm_pre_clip': 13.114531517028809, 'timing/data': 0.00471043202560395, 'timing/model': 1.2038514660671353, 'learning_rate/qwen_vl_interface': 1.6149986448416032e-06, 'learning_rate/action_model': 1.614998644841603e-05, 'epoch': 1.61}) +[2026-06-17 16:23:29,712][starVLA.training.train_starvla][INFO] - Step 5465, Loss: {'train/loss': 0.7294415831565857, 'train/grad_norm_pre_clip': 15.526734352111816, 'timing/data': 0.004713764996267855, 'timing/model': 1.213133349083364, 'learning_rate/qwen_vl_interface': 1.6136068604767444e-06, 'learning_rate/action_model': 1.6136068604767442e-05, 'epoch': 1.61}) +[2026-06-17 16:23:34,549][starVLA.training.train_starvla][INFO] - Step 5466, Loss: {'train/loss': 0.6079804301261902, 'train/grad_norm_pre_clip': 12.593591690063477, 'timing/data': 0.005109913996420801, 'timing/model': 1.3071278820279986, 'learning_rate/qwen_vl_interface': 1.6122158299400962e-06, 'learning_rate/action_model': 1.612215829940096e-05, 'epoch': 1.61}) +[2026-06-17 16:23:39,215][starVLA.training.train_starvla][INFO] - Step 5467, Loss: {'train/loss': 0.598665714263916, 'train/grad_norm_pre_clip': 11.696905136108398, 'timing/data': 0.004677467048168182, 'timing/model': 1.2036323710344732, 'learning_rate/qwen_vl_interface': 1.6108255535200193e-06, 'learning_rate/action_model': 1.6108255535200193e-05, 'epoch': 1.61}) +[2026-06-17 16:23:44,033][starVLA.training.train_starvla][INFO] - Step 5468, Loss: {'train/loss': 0.6166843175888062, 'train/grad_norm_pre_clip': 14.295984268188477, 'timing/data': 0.004776113084517419, 'timing/model': 1.2144885709276423, 'learning_rate/qwen_vl_interface': 1.6094360315047236e-06, 'learning_rate/action_model': 1.6094360315047237e-05, 'epoch': 1.61}) +[2026-06-17 16:23:48,838][starVLA.training.train_starvla][INFO] - Step 5469, Loss: {'train/loss': 0.5623731017112732, 'train/grad_norm_pre_clip': 12.616086959838867, 'timing/data': 0.004926115041598678, 'timing/model': 1.2068955949507654, 'learning_rate/qwen_vl_interface': 1.608047264182255e-06, 'learning_rate/action_model': 1.608047264182255e-05, 'epoch': 1.61}) +[2026-06-17 16:23:53,624][starVLA.training.train_starvla][INFO] - Step 5470, Loss: {'train/loss': 0.7748827934265137, 'train/grad_norm_pre_clip': 24.978187561035156, 'timing/data': 0.004954618983902037, 'timing/model': 1.2031862159492448, 'learning_rate/qwen_vl_interface': 1.6066592518405077e-06, 'learning_rate/action_model': 1.6066592518405076e-05, 'epoch': 1.61}) +[2026-06-17 16:23:58,408][starVLA.training.train_starvla][INFO] - Step 5471, Loss: {'train/loss': 0.5808843374252319, 'train/grad_norm_pre_clip': 15.675806999206543, 'timing/data': 0.00503184599801898, 'timing/model': 1.3323339950293303, 'learning_rate/qwen_vl_interface': 1.605271994767216e-06, 'learning_rate/action_model': 1.605271994767216e-05, 'epoch': 1.61}) +[2026-06-17 16:24:03,069][starVLA.training.train_starvla][INFO] - Step 5472, Loss: {'train/loss': 0.5171072483062744, 'train/grad_norm_pre_clip': 11.168621063232422, 'timing/data': 0.004789859056472778, 'timing/model': 1.203216259018518, 'learning_rate/qwen_vl_interface': 1.6038854932499635e-06, 'learning_rate/action_model': 1.6038854932499633e-05, 'epoch': 1.61}) +[2026-06-17 16:24:07,964][starVLA.training.train_starvla][INFO] - Step 5473, Loss: {'train/loss': 0.7383490204811096, 'train/grad_norm_pre_clip': 17.703027725219727, 'timing/data': 0.004830444930121303, 'timing/model': 1.2038676099618897, 'learning_rate/qwen_vl_interface': 1.6024997475761727e-06, 'learning_rate/action_model': 1.602499747576173e-05, 'epoch': 1.61}) +[2026-06-17 16:24:12,764][starVLA.training.train_starvla][INFO] - Step 5474, Loss: {'train/loss': 0.6734421253204346, 'train/grad_norm_pre_clip': 14.339879989624023, 'timing/data': 0.004902712069451809, 'timing/model': 1.211115528945811, 'learning_rate/qwen_vl_interface': 1.6011147580331065e-06, 'learning_rate/action_model': 1.6011147580331066e-05, 'epoch': 1.61}) +[2026-06-17 16:24:17,496][starVLA.training.train_starvla][INFO] - Step 5475, Loss: {'train/loss': 0.6449788808822632, 'train/grad_norm_pre_clip': 11.970083236694336, 'timing/data': 0.00537265301682055, 'timing/model': 1.2094781399937347, 'learning_rate/qwen_vl_interface': 1.5997305249078788e-06, 'learning_rate/action_model': 1.5997305249078787e-05, 'epoch': 1.61}) +[2026-06-17 16:24:22,247][starVLA.training.train_starvla][INFO] - Step 5476, Loss: {'train/loss': 0.6628015637397766, 'train/grad_norm_pre_clip': 15.890872955322266, 'timing/data': 0.004908681032247841, 'timing/model': 1.2069906439865008, 'learning_rate/qwen_vl_interface': 1.5983470484874412e-06, 'learning_rate/action_model': 1.598347048487441e-05, 'epoch': 1.61}) +[2026-06-17 16:24:27,053][starVLA.training.train_starvla][INFO] - Step 5477, Loss: {'train/loss': 0.5376582145690918, 'train/grad_norm_pre_clip': 14.6425199508667, 'timing/data': 0.004757483955472708, 'timing/model': 1.3507351080188528, 'learning_rate/qwen_vl_interface': 1.5969643290585893e-06, 'learning_rate/action_model': 1.5969643290585892e-05, 'epoch': 1.61}) +[2026-06-17 16:24:31,721][starVLA.training.train_starvla][INFO] - Step 5478, Loss: {'train/loss': 0.8283726572990417, 'train/grad_norm_pre_clip': 13.797880172729492, 'timing/data': 0.0050180519465357065, 'timing/model': 1.2119889680761844, 'learning_rate/qwen_vl_interface': 1.5955823669079632e-06, 'learning_rate/action_model': 1.595582366907963e-05, 'epoch': 1.61}) +[2026-06-17 16:24:36,511][starVLA.training.train_starvla][INFO] - Step 5479, Loss: {'train/loss': 0.6432623267173767, 'train/grad_norm_pre_clip': 20.8487491607666, 'timing/data': 0.004657755955122411, 'timing/model': 1.206508764065802, 'learning_rate/qwen_vl_interface': 1.594201162322045e-06, 'learning_rate/action_model': 1.594201162322045e-05, 'epoch': 1.61}) +[2026-06-17 16:24:41,305][starVLA.training.train_starvla][INFO] - Step 5480, Loss: {'train/loss': 0.7869611978530884, 'train/grad_norm_pre_clip': 17.698593139648438, 'timing/data': 0.004799642018042505, 'timing/model': 1.2077996240695938, 'learning_rate/qwen_vl_interface': 1.592820715587159e-06, 'learning_rate/action_model': 1.5928207155871593e-05, 'epoch': 1.61}) +[2026-06-17 16:24:46,090][starVLA.training.train_starvla][INFO] - Step 5481, Loss: {'train/loss': 0.7008264064788818, 'train/grad_norm_pre_clip': 13.54958724975586, 'timing/data': 0.0048773420276120305, 'timing/model': 1.2057294190162793, 'learning_rate/qwen_vl_interface': 1.5914410269894743e-06, 'learning_rate/action_model': 1.591441026989474e-05, 'epoch': 1.61}) +[2026-06-17 16:24:50,852][starVLA.training.train_starvla][INFO] - Step 5482, Loss: {'train/loss': 0.7706700563430786, 'train/grad_norm_pre_clip': 21.566484451293945, 'timing/data': 0.0046457069693133235, 'timing/model': 1.2051307699875906, 'learning_rate/qwen_vl_interface': 1.5900620968150015e-06, 'learning_rate/action_model': 1.5900620968150015e-05, 'epoch': 1.61}) +[2026-06-17 16:24:55,633][starVLA.training.train_starvla][INFO] - Step 5483, Loss: {'train/loss': 0.9637280702590942, 'train/grad_norm_pre_clip': 16.042850494384766, 'timing/data': 0.004761017975397408, 'timing/model': 1.3261817579623312, 'learning_rate/qwen_vl_interface': 1.5886839253495944e-06, 'learning_rate/action_model': 1.5886839253495943e-05, 'epoch': 1.61}) +[2026-06-17 16:25:00,300][starVLA.training.train_starvla][INFO] - Step 5484, Loss: {'train/loss': 0.8522815704345703, 'train/grad_norm_pre_clip': 13.773486137390137, 'timing/data': 0.004820303060114384, 'timing/model': 1.210385832004249, 'learning_rate/qwen_vl_interface': 1.5873065128789495e-06, 'learning_rate/action_model': 1.5873065128789493e-05, 'epoch': 1.61}) +[2026-06-17 16:25:05,091][starVLA.training.train_starvla][INFO] - Step 5485, Loss: {'train/loss': 0.6619715690612793, 'train/grad_norm_pre_clip': 13.343749046325684, 'timing/data': 0.00450185697991401, 'timing/model': 1.2094393060542643, 'learning_rate/qwen_vl_interface': 1.5859298596886038e-06, 'learning_rate/action_model': 1.585929859688604e-05, 'epoch': 1.61}) +[2026-06-17 16:25:09,911][starVLA.training.train_starvla][INFO] - Step 5486, Loss: {'train/loss': 0.5157701969146729, 'train/grad_norm_pre_clip': 12.924095153808594, 'timing/data': 0.00527763890568167, 'timing/model': 1.2101055979728699, 'learning_rate/qwen_vl_interface': 1.5845539660639429e-06, 'learning_rate/action_model': 1.5845539660639426e-05, 'epoch': 1.61}) +[2026-06-17 16:25:14,722][starVLA.training.train_starvla][INFO] - Step 5487, Loss: {'train/loss': 0.7146816253662109, 'train/grad_norm_pre_clip': 13.977362632751465, 'timing/data': 0.004692335030995309, 'timing/model': 1.2091771180275828, 'learning_rate/qwen_vl_interface': 1.5831788322901901e-06, 'learning_rate/action_model': 1.58317883229019e-05, 'epoch': 1.61}) +[2026-06-17 16:25:19,497][starVLA.training.train_starvla][INFO] - Step 5488, Loss: {'train/loss': 0.9240432977676392, 'train/grad_norm_pre_clip': 26.798818588256836, 'timing/data': 0.004863832029514015, 'timing/model': 1.3330372780328616, 'learning_rate/qwen_vl_interface': 1.5818044586524092e-06, 'learning_rate/action_model': 1.5818044586524092e-05, 'epoch': 1.61}) +[2026-06-17 16:25:24,153][starVLA.training.train_starvla][INFO] - Step 5489, Loss: {'train/loss': 0.4842809736728668, 'train/grad_norm_pre_clip': 10.527634620666504, 'timing/data': 0.004839555942453444, 'timing/model': 1.2063511839369312, 'learning_rate/qwen_vl_interface': 1.5804308454355102e-06, 'learning_rate/action_model': 1.5804308454355102e-05, 'epoch': 1.62}) +[2026-06-17 16:25:28,946][starVLA.training.train_starvla][INFO] - Step 5490, Loss: {'train/loss': 0.7187076807022095, 'train/grad_norm_pre_clip': 15.500521659851074, 'timing/data': 0.005017156014218926, 'timing/model': 1.2096634039189667, 'learning_rate/qwen_vl_interface': 1.5790579929242475e-06, 'learning_rate/action_model': 1.5790579929242475e-05, 'epoch': 1.62}) +[2026-06-17 16:25:33,647][starVLA.training.train_starvla][INFO] - Step 5491, Loss: {'train/loss': 0.38946837186813354, 'train/grad_norm_pre_clip': 12.211054801940918, 'timing/data': 0.0047897499753162265, 'timing/model': 1.2030226610368118, 'learning_rate/qwen_vl_interface': 1.577685901403212e-06, 'learning_rate/action_model': 1.577685901403212e-05, 'epoch': 1.62}) +[2026-06-17 16:25:38,396][starVLA.training.train_starvla][INFO] - Step 5492, Loss: {'train/loss': 0.5854977965354919, 'train/grad_norm_pre_clip': 15.725605964660645, 'timing/data': 0.004711508983746171, 'timing/model': 1.2134267210494727, 'learning_rate/qwen_vl_interface': 1.5763145711568414e-06, 'learning_rate/action_model': 1.5763145711568413e-05, 'epoch': 1.62}) +[2026-06-17 16:25:43,210][starVLA.training.train_starvla][INFO] - Step 5493, Loss: {'train/loss': 0.5983129739761353, 'train/grad_norm_pre_clip': 13.636780738830566, 'timing/data': 0.0048373600002378225, 'timing/model': 1.2159371549496427, 'learning_rate/qwen_vl_interface': 1.5749440024694129e-06, 'learning_rate/action_model': 1.574944002469413e-05, 'epoch': 1.62}) +[2026-06-17 16:25:48,009][starVLA.training.train_starvla][INFO] - Step 5494, Loss: {'train/loss': 0.6576747298240662, 'train/grad_norm_pre_clip': 14.659162521362305, 'timing/data': 0.004773422027938068, 'timing/model': 1.3420625790022314, 'learning_rate/qwen_vl_interface': 1.5735741956250476e-06, 'learning_rate/action_model': 1.5735741956250475e-05, 'epoch': 1.62}) +[2026-06-17 16:25:52,677][starVLA.training.train_starvla][INFO] - Step 5495, Loss: {'train/loss': 0.7454097270965576, 'train/grad_norm_pre_clip': 16.14177703857422, 'timing/data': 0.004708236898295581, 'timing/model': 1.207325050025247, 'learning_rate/qwen_vl_interface': 1.5722051509077068e-06, 'learning_rate/action_model': 1.572205150907707e-05, 'epoch': 1.62}) +[2026-06-17 16:25:57,467][starVLA.training.train_starvla][INFO] - Step 5496, Loss: {'train/loss': 0.7706929445266724, 'train/grad_norm_pre_clip': 15.031652450561523, 'timing/data': 0.0052047360222786665, 'timing/model': 1.2057182980934158, 'learning_rate/qwen_vl_interface': 1.570836868601196e-06, 'learning_rate/action_model': 1.570836868601196e-05, 'epoch': 1.62}) +[2026-06-17 16:26:02,246][starVLA.training.train_starvla][INFO] - Step 5497, Loss: {'train/loss': 1.2262301445007324, 'train/grad_norm_pre_clip': 19.207475662231445, 'timing/data': 0.004708242020569742, 'timing/model': 1.2075364879565313, 'learning_rate/qwen_vl_interface': 1.5694693489891616e-06, 'learning_rate/action_model': 1.5694693489891617e-05, 'epoch': 1.62}) +[2026-06-17 16:26:07,011][starVLA.training.train_starvla][INFO] - Step 5498, Loss: {'train/loss': 0.6066182255744934, 'train/grad_norm_pre_clip': 14.709615707397461, 'timing/data': 0.005345817073248327, 'timing/model': 1.2057617680402473, 'learning_rate/qwen_vl_interface': 1.5681025923550914e-06, 'learning_rate/action_model': 1.568102592355091e-05, 'epoch': 1.62}) +[2026-06-17 16:26:11,792][starVLA.training.train_starvla][INFO] - Step 5499, Loss: {'train/loss': 0.7014418244361877, 'train/grad_norm_pre_clip': 12.752378463745117, 'timing/data': 0.004853728110902011, 'timing/model': 1.207336888066493, 'learning_rate/qwen_vl_interface': 1.5667365989823161e-06, 'learning_rate/action_model': 1.566736598982316e-05, 'epoch': 1.62}) +[2026-06-17 16:35:53,746][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 5500 +[2026-06-17 16:35:53,747][starVLA.training.train_starvla][INFO] - Step 5500, Loss: {'train/loss': 0.7421363592147827, 'train/grad_norm_pre_clip': 17.801700592041016, 'eval/loss': 0.7093004584312439, 'eval/latency_0/loss': 0.7093004584312439, 'eval/demon_fire_precision': 0.9114132674083231, 'eval/demon_fire_recall': 0.9186046511627907, 'eval/demon_fire_f1': 0.9149948293691831, 'eval/demon_fire_rate_model': 0.674916573971079, 'eval/demon_fire_rate_teacher': 0.6696329254727474, 'eval/demon_left_precision': 0.8389221556886227, 'eval/demon_left_recall': 0.7937677053824362, 'eval/demon_left_f1': 0.8157205240174672, 'eval/demon_left_rate_model': 0.46440489432703, 'eval/demon_left_rate_teacher': 0.49082313681868744, 'eval/demon_right_precision': 0.8058510638297872, 'eval/demon_right_recall': 0.8559322033898306, 'eval/demon_right_f1': 0.8301369863013699, 'eval/demon_right_rate_model': 0.5228031145717463, 'eval/demon_right_rate_teacher': 0.4922135706340378, 'eval/demon_move_f1': 0.8229287551594185, 'eval/demon_cc_f1': 0.8852822471257739, 'eval/latency_0/demon_fire_precision': 0.9114132674083231, 'eval/latency_0/demon_fire_recall': 0.9186046511627907, 'eval/latency_0/demon_fire_f1': 0.9149948293691831, 'eval/latency_0/demon_fire_rate_model': 0.674916573971079, 'eval/latency_0/demon_fire_rate_teacher': 0.6696329254727474, 'eval/latency_0/demon_left_precision': 0.8389221556886227, 'eval/latency_0/demon_left_recall': 0.7937677053824362, 'eval/latency_0/demon_left_f1': 0.8157205240174672, 'eval/latency_0/demon_left_rate_model': 0.46440489432703, 'eval/latency_0/demon_left_rate_teacher': 0.49082313681868744, 'eval/latency_0/demon_right_precision': 0.8058510638297872, 'eval/latency_0/demon_right_recall': 0.8559322033898306, 'eval/latency_0/demon_right_f1': 0.8301369863013699, 'eval/latency_0/demon_right_rate_model': 0.5228031145717463, 'eval/latency_0/demon_right_rate_teacher': 0.4922135706340378, 'eval/latency_0/demon_move_f1': 0.8229287551594185, 'eval/latency_0/demon_cc_f1': 0.8852822471257739, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 7866.0, 'rl_games_eval/mid_train/mean_length': 3418.0, 'rl_games_eval/mid_train/std_reward': 349.9485676495905, 'rl_games_eval/mid_train/std_length': 222.9035665932692, 'rl_games_eval/mid_train/macro_mean_reward': 7866.0, 'rl_games_eval/mid_train/macro_mean_length': 3418.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/demon_attack__latency_0/mean_reward': 7866.0, 'rl_games_eval/mid_train/demon_attack__latency_0/mean_length': 3418.0, 'rl_games_eval/mid_train/demon_attack__latency_0/std_reward': 349.9485676495905, 'rl_games_eval/mid_train/demon_attack__latency_0/std_length': 222.9035665932692, 'rl_games_eval/mid_train/demon_attack__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 7866.0, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.005124768009409308, 'timing/model': 1.339518695953302, 'learning_rate/qwen_vl_interface': 1.5653713691540053e-06, 'learning_rate/action_model': 1.5653713691540054e-05, 'epoch': 1.62}) +[2026-06-17 16:35:53,747][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_5500_state +[2026-06-17 16:35:53,749][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 16:35:57,434][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_5500_state/model.safetensors +[2026-06-17 16:36:05,129][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_5500_state/optimizer.bin +[2026-06-17 16:36:05,130][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_5500_state/sampler.bin +[2026-06-17 16:36:05,130][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_5500_state/sampler_1.bin +[2026-06-17 16:36:05,130][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_5500_state/random_states_0.pkl +[2026-06-17 16:36:05,179][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 5500 +[2026-06-17 16:36:05,179][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 16:36:05,189][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 16:36:09,790][starVLA.training.train_starvla][INFO] - Step 5501, Loss: {'train/loss': 0.7487165331840515, 'train/grad_norm_pre_clip': 15.999239921569824, 'timing/data': 0.0051700330805033445, 'timing/model': 1.1876563250552863, 'learning_rate/qwen_vl_interface': 1.5640069031531774e-06, 'learning_rate/action_model': 1.5640069031531772e-05, 'epoch': 1.62}) +[2026-06-17 16:36:14,501][starVLA.training.train_starvla][INFO] - Step 5502, Loss: {'train/loss': 0.9088284969329834, 'train/grad_norm_pre_clip': 15.356948852539062, 'timing/data': 0.005058990907855332, 'timing/model': 1.1871370550943539, 'learning_rate/qwen_vl_interface': 1.5626432012626835e-06, 'learning_rate/action_model': 1.5626432012626836e-05, 'epoch': 1.62}) +[2026-06-17 16:36:19,221][starVLA.training.train_starvla][INFO] - Step 5503, Loss: {'train/loss': 0.6304202079772949, 'train/grad_norm_pre_clip': 12.620108604431152, 'timing/data': 0.0056427710223942995, 'timing/model': 1.1890960989985615, 'learning_rate/qwen_vl_interface': 1.5612802637652202e-06, 'learning_rate/action_model': 1.5612802637652203e-05, 'epoch': 1.62}) +[2026-06-17 16:36:23,974][starVLA.training.train_starvla][INFO] - Step 5504, Loss: {'train/loss': 0.5040788650512695, 'train/grad_norm_pre_clip': 10.959671020507812, 'timing/data': 0.004708316992036998, 'timing/model': 1.2091893720207736, 'learning_rate/qwen_vl_interface': 1.5599180909433292e-06, 'learning_rate/action_model': 1.5599180909433292e-05, 'epoch': 1.62}) +[2026-06-17 16:36:28,705][starVLA.training.train_starvla][INFO] - Step 5505, Loss: {'train/loss': 0.600247859954834, 'train/grad_norm_pre_clip': 14.101998329162598, 'timing/data': 0.004899983061477542, 'timing/model': 1.3202194820623845, 'learning_rate/qwen_vl_interface': 1.5585566830793883e-06, 'learning_rate/action_model': 1.558556683079388e-05, 'epoch': 1.62}) +[2026-06-17 16:36:33,328][starVLA.training.train_starvla][INFO] - Step 5506, Loss: {'train/loss': 0.8438143134117126, 'train/grad_norm_pre_clip': 14.86531925201416, 'timing/data': 0.004702653037384152, 'timing/model': 1.1950128269381821, 'learning_rate/qwen_vl_interface': 1.5571960404556204e-06, 'learning_rate/action_model': 1.5571960404556203e-05, 'epoch': 1.62}) +[2026-06-17 16:36:38,087][starVLA.training.train_starvla][INFO] - Step 5507, Loss: {'train/loss': 0.6996008157730103, 'train/grad_norm_pre_clip': 14.830561637878418, 'timing/data': 0.004280199995264411, 'timing/model': 1.1970543389907107, 'learning_rate/qwen_vl_interface': 1.5558361633540848e-06, 'learning_rate/action_model': 1.5558361633540848e-05, 'epoch': 1.62}) +[2026-06-17 16:36:42,774][starVLA.training.train_starvla][INFO] - Step 5508, Loss: {'train/loss': 0.48309069871902466, 'train/grad_norm_pre_clip': 11.459284782409668, 'timing/data': 0.004728944040834904, 'timing/model': 1.197672970010899, 'learning_rate/qwen_vl_interface': 1.5544770520566884e-06, 'learning_rate/action_model': 1.5544770520566885e-05, 'epoch': 1.62}) +[2026-06-17 16:36:47,521][starVLA.training.train_starvla][INFO] - Step 5509, Loss: {'train/loss': 0.49543678760528564, 'train/grad_norm_pre_clip': 10.57713508605957, 'timing/data': 0.006829379010014236, 'timing/model': 1.2148549250559881, 'learning_rate/qwen_vl_interface': 1.553118706845176e-06, 'learning_rate/action_model': 1.5531187068451757e-05, 'epoch': 1.62}) +[2026-06-17 16:36:52,278][starVLA.training.train_starvla][INFO] - Step 5510, Loss: {'train/loss': 0.7098633646965027, 'train/grad_norm_pre_clip': 18.174789428710938, 'timing/data': 0.004887407063506544, 'timing/model': 1.1975691749248654, 'learning_rate/qwen_vl_interface': 1.5517611280011333e-06, 'learning_rate/action_model': 1.551761128001133e-05, 'epoch': 1.62}) +[2026-06-17 16:36:57,031][starVLA.training.train_starvla][INFO] - Step 5511, Loss: {'train/loss': 0.7142966985702515, 'train/grad_norm_pre_clip': 16.233789443969727, 'timing/data': 0.004925317014567554, 'timing/model': 1.3213061760179698, 'learning_rate/qwen_vl_interface': 1.5504043158059886e-06, 'learning_rate/action_model': 1.5504043158059885e-05, 'epoch': 1.62}) +[2026-06-17 16:37:01,661][starVLA.training.train_starvla][INFO] - Step 5512, Loss: {'train/loss': 0.6570903062820435, 'train/grad_norm_pre_clip': 13.739020347595215, 'timing/data': 0.004837030079215765, 'timing/model': 1.1989041499327868, 'learning_rate/qwen_vl_interface': 1.54904827054101e-06, 'learning_rate/action_model': 1.54904827054101e-05, 'epoch': 1.62}) +[2026-06-17 16:37:06,420][starVLA.training.train_starvla][INFO] - Step 5513, Loss: {'train/loss': 0.36903446912765503, 'train/grad_norm_pre_clip': 11.262503623962402, 'timing/data': 0.0049583109794184566, 'timing/model': 1.2009687670506537, 'learning_rate/qwen_vl_interface': 1.5476929924873074e-06, 'learning_rate/action_model': 1.5476929924873073e-05, 'epoch': 1.62}) +[2026-06-17 16:37:11,155][starVLA.training.train_starvla][INFO] - Step 5514, Loss: {'train/loss': 0.6241655945777893, 'train/grad_norm_pre_clip': 16.828536987304688, 'timing/data': 0.004945669090375304, 'timing/model': 1.2062081400072202, 'learning_rate/qwen_vl_interface': 1.5463384819258302e-06, 'learning_rate/action_model': 1.54633848192583e-05, 'epoch': 1.62}) +[2026-06-17 16:37:15,931][starVLA.training.train_starvla][INFO] - Step 5515, Loss: {'train/loss': 0.6141023635864258, 'train/grad_norm_pre_clip': 12.560798645019531, 'timing/data': 0.0051404370460659266, 'timing/model': 1.203624980058521, 'learning_rate/qwen_vl_interface': 1.544984739137373e-06, 'learning_rate/action_model': 1.544984739137373e-05, 'epoch': 1.62}) +[2026-06-17 16:37:20,781][starVLA.training.train_starvla][INFO] - Step 5516, Loss: {'train/loss': 0.5841047763824463, 'train/grad_norm_pre_clip': 14.875737190246582, 'timing/data': 0.004850090015679598, 'timing/model': 1.1996858769562095, 'learning_rate/qwen_vl_interface': 1.543631764402566e-06, 'learning_rate/action_model': 1.5436317644025658e-05, 'epoch': 1.62}) +[2026-06-17 16:37:25,606][starVLA.training.train_starvla][INFO] - Step 5517, Loss: {'train/loss': 0.8644083738327026, 'train/grad_norm_pre_clip': 23.18058967590332, 'timing/data': 0.0053128269501030445, 'timing/model': 1.3390228130156174, 'learning_rate/qwen_vl_interface': 1.5422795580018813e-06, 'learning_rate/action_model': 1.542279558001881e-05, 'epoch': 1.62}) +[2026-06-17 16:37:30,297][starVLA.training.train_starvla][INFO] - Step 5518, Loss: {'train/loss': 0.6747361421585083, 'train/grad_norm_pre_clip': 14.567970275878906, 'timing/data': 0.004763921955600381, 'timing/model': 1.250517298001796, 'learning_rate/qwen_vl_interface': 1.5409281202156329e-06, 'learning_rate/action_model': 1.540928120215633e-05, 'epoch': 1.62}) +[2026-06-17 16:37:34,943][starVLA.training.train_starvla][INFO] - Step 5519, Loss: {'train/loss': 0.6218187808990479, 'train/grad_norm_pre_clip': 12.555985450744629, 'timing/data': 0.004875306971371174, 'timing/model': 1.2055747339036316, 'learning_rate/qwen_vl_interface': 1.5395774513239777e-06, 'learning_rate/action_model': 1.539577451323978e-05, 'epoch': 1.62}) +[2026-06-17 16:37:39,681][starVLA.training.train_starvla][INFO] - Step 5520, Loss: {'train/loss': 0.7042809128761292, 'train/grad_norm_pre_clip': 19.488893508911133, 'timing/data': 0.0050990000599995255, 'timing/model': 1.2028683440294117, 'learning_rate/qwen_vl_interface': 1.5382275516069112e-06, 'learning_rate/action_model': 1.5382275516069113e-05, 'epoch': 1.62}) +[2026-06-17 16:37:44,467][starVLA.training.train_starvla][INFO] - Step 5521, Loss: {'train/loss': 0.6335703134536743, 'train/grad_norm_pre_clip': 13.993377685546875, 'timing/data': 0.004777287947945297, 'timing/model': 1.2031316000502557, 'learning_rate/qwen_vl_interface': 1.5368784213442653e-06, 'learning_rate/action_model': 1.5368784213442654e-05, 'epoch': 1.62}) +[2026-06-17 16:37:49,355][starVLA.training.train_starvla][INFO] - Step 5522, Loss: {'train/loss': 0.803726315498352, 'train/grad_norm_pre_clip': 16.426855087280273, 'timing/data': 0.005183589993976057, 'timing/model': 1.2020058429334313, 'learning_rate/qwen_vl_interface': 1.5355300608157172e-06, 'learning_rate/action_model': 1.5355300608157173e-05, 'epoch': 1.62}) +[2026-06-17 16:37:54,133][starVLA.training.train_starvla][INFO] - Step 5523, Loss: {'train/loss': 0.5333757400512695, 'train/grad_norm_pre_clip': 14.505407333374023, 'timing/data': 0.004623885033652186, 'timing/model': 1.3316747660283, 'learning_rate/qwen_vl_interface': 1.5341824703007854e-06, 'learning_rate/action_model': 1.5341824703007854e-05, 'epoch': 1.63}) +[2026-06-17 16:37:58,903][starVLA.training.train_starvla][INFO] - Step 5524, Loss: {'train/loss': 0.8458946347236633, 'train/grad_norm_pre_clip': 15.962791442871094, 'timing/data': 0.004830464953556657, 'timing/model': 1.3219547210028395, 'learning_rate/qwen_vl_interface': 1.5328356500788263e-06, 'learning_rate/action_model': 1.5328356500788264e-05, 'epoch': 1.63}) +[2026-06-17 16:38:03,548][starVLA.training.train_starvla][INFO] - Step 5525, Loss: {'train/loss': 0.6703026294708252, 'train/grad_norm_pre_clip': 30.1823673248291, 'timing/data': 0.005111112026497722, 'timing/model': 1.2023103180108592, 'learning_rate/qwen_vl_interface': 1.5314896004290366e-06, 'learning_rate/action_model': 1.5314896004290365e-05, 'epoch': 1.63}) +[2026-06-17 16:38:08,300][starVLA.training.train_starvla][INFO] - Step 5526, Loss: {'train/loss': 0.5716493129730225, 'train/grad_norm_pre_clip': 11.001522064208984, 'timing/data': 0.0050307640340179205, 'timing/model': 1.203832329949364, 'learning_rate/qwen_vl_interface': 1.530144321630454e-06, 'learning_rate/action_model': 1.530144321630454e-05, 'epoch': 1.63}) +[2026-06-17 16:38:13,072][starVLA.training.train_starvla][INFO] - Step 5527, Loss: {'train/loss': 0.9562399983406067, 'train/grad_norm_pre_clip': 16.0496826171875, 'timing/data': 0.0048877979861572385, 'timing/model': 1.204129593912512, 'learning_rate/qwen_vl_interface': 1.5287998139619564e-06, 'learning_rate/action_model': 1.5287998139619562e-05, 'epoch': 1.63}) +[2026-06-17 16:38:17,846][starVLA.training.train_starvla][INFO] - Step 5528, Loss: {'train/loss': 0.9549075365066528, 'train/grad_norm_pre_clip': 16.808944702148438, 'timing/data': 0.0051585499895736575, 'timing/model': 1.2016406919574365, 'learning_rate/qwen_vl_interface': 1.5274560777022613e-06, 'learning_rate/action_model': 1.5274560777022613e-05, 'epoch': 1.63}) +[2026-06-17 16:38:22,562][starVLA.training.train_starvla][INFO] - Step 5529, Loss: {'train/loss': 0.7879002094268799, 'train/grad_norm_pre_clip': 16.898042678833008, 'timing/data': 0.004837632994167507, 'timing/model': 1.2724194810725749, 'learning_rate/qwen_vl_interface': 1.5261131131299274e-06, 'learning_rate/action_model': 1.5261131131299272e-05, 'epoch': 1.63}) +[2026-06-17 16:38:27,332][starVLA.training.train_starvla][INFO] - Step 5530, Loss: {'train/loss': 0.976409912109375, 'train/grad_norm_pre_clip': 16.832231521606445, 'timing/data': 0.004416823037900031, 'timing/model': 1.202564149047248, 'learning_rate/qwen_vl_interface': 1.5247709205233518e-06, 'learning_rate/action_model': 1.5247709205233518e-05, 'epoch': 1.63}) +[2026-06-17 16:38:32,085][starVLA.training.train_starvla][INFO] - Step 5531, Loss: {'train/loss': 0.43312981724739075, 'train/grad_norm_pre_clip': 15.592256546020508, 'timing/data': 0.004693935974501073, 'timing/model': 1.2035302909789607, 'learning_rate/qwen_vl_interface': 1.5234295001607732e-06, 'learning_rate/action_model': 1.5234295001607732e-05, 'epoch': 1.63}) +[2026-06-17 16:38:36,878][starVLA.training.train_starvla][INFO] - Step 5532, Loss: {'train/loss': 0.7220247983932495, 'train/grad_norm_pre_clip': 12.869147300720215, 'timing/data': 0.00490091391839087, 'timing/model': 1.2170011029811576, 'learning_rate/qwen_vl_interface': 1.5220888523202687e-06, 'learning_rate/action_model': 1.5220888523202686e-05, 'epoch': 1.63}) +[2026-06-17 16:38:41,659][starVLA.training.train_starvla][INFO] - Step 5533, Loss: {'train/loss': 0.8105812072753906, 'train/grad_norm_pre_clip': 13.58755111694336, 'timing/data': 0.005346690071746707, 'timing/model': 1.2039697179570794, 'learning_rate/qwen_vl_interface': 1.5207489772797557e-06, 'learning_rate/action_model': 1.5207489772797556e-05, 'epoch': 1.63}) +[2026-06-17 16:38:46,453][starVLA.training.train_starvla][INFO] - Step 5534, Loss: {'train/loss': 0.504784107208252, 'train/grad_norm_pre_clip': 11.212484359741211, 'timing/data': 0.0049880159785971045, 'timing/model': 1.337865178938955, 'learning_rate/qwen_vl_interface': 1.5194098753169933e-06, 'learning_rate/action_model': 1.5194098753169933e-05, 'epoch': 1.63}) +[2026-06-17 16:38:51,249][starVLA.training.train_starvla][INFO] - Step 5535, Loss: {'train/loss': 0.7119072675704956, 'train/grad_norm_pre_clip': 13.710028648376465, 'timing/data': 0.0046494759153574705, 'timing/model': 1.283279285999015, 'learning_rate/qwen_vl_interface': 1.5180715467095799e-06, 'learning_rate/action_model': 1.5180715467095798e-05, 'epoch': 1.63}) +[2026-06-17 16:38:55,902][starVLA.training.train_starvla][INFO] - Step 5536, Loss: {'train/loss': 0.6194905638694763, 'train/grad_norm_pre_clip': 14.009982109069824, 'timing/data': 0.0049152340507134795, 'timing/model': 1.2037790010217577, 'learning_rate/qwen_vl_interface': 1.5167339917349482e-06, 'learning_rate/action_model': 1.5167339917349482e-05, 'epoch': 1.63}) +[2026-06-17 16:39:00,679][starVLA.training.train_starvla][INFO] - Step 5537, Loss: {'train/loss': 0.4802969992160797, 'train/grad_norm_pre_clip': 12.402527809143066, 'timing/data': 0.004779485985636711, 'timing/model': 1.2042554899817333, 'learning_rate/qwen_vl_interface': 1.5153972106703778e-06, 'learning_rate/action_model': 1.5153972106703779e-05, 'epoch': 1.63}) +[2026-06-17 16:39:05,464][starVLA.training.train_starvla][INFO] - Step 5538, Loss: {'train/loss': 0.4469310939311981, 'train/grad_norm_pre_clip': 22.07015609741211, 'timing/data': 0.005355553003028035, 'timing/model': 1.2098068239865825, 'learning_rate/qwen_vl_interface': 1.5140612037929844e-06, 'learning_rate/action_model': 1.5140612037929844e-05, 'epoch': 1.63}) +[2026-06-17 16:39:10,339][starVLA.training.train_starvla][INFO] - Step 5539, Loss: {'train/loss': 0.5889691114425659, 'train/grad_norm_pre_clip': 116.67974090576172, 'timing/data': 0.004922871943563223, 'timing/model': 1.207156240940094, 'learning_rate/qwen_vl_interface': 1.512725971379724e-06, 'learning_rate/action_model': 1.5127259713797239e-05, 'epoch': 1.63}) +[2026-06-17 16:39:15,123][starVLA.training.train_starvla][INFO] - Step 5540, Loss: {'train/loss': 0.7131277322769165, 'train/grad_norm_pre_clip': 15.25616455078125, 'timing/data': 0.004797624074853957, 'timing/model': 1.2037907559424639, 'learning_rate/qwen_vl_interface': 1.5113915137073905e-06, 'learning_rate/action_model': 1.5113915137073906e-05, 'epoch': 1.63}) +[2026-06-17 16:39:19,894][starVLA.training.train_starvla][INFO] - Step 5541, Loss: {'train/loss': 0.5933429002761841, 'train/grad_norm_pre_clip': 28.311716079711914, 'timing/data': 0.005072808009572327, 'timing/model': 1.296982147032395, 'learning_rate/qwen_vl_interface': 1.5100578310526198e-06, 'learning_rate/action_model': 1.5100578310526197e-05, 'epoch': 1.63}) +[2026-06-17 16:39:24,587][starVLA.training.train_starvla][INFO] - Step 5542, Loss: {'train/loss': 0.6757152676582336, 'train/grad_norm_pre_clip': 13.299799919128418, 'timing/data': 0.00488575897179544, 'timing/model': 1.2106791239930317, 'learning_rate/qwen_vl_interface': 1.508724923691885e-06, 'learning_rate/action_model': 1.5087249236918848e-05, 'epoch': 1.63}) +[2026-06-17 16:39:29,399][starVLA.training.train_starvla][INFO] - Step 5543, Loss: {'train/loss': 0.5810492634773254, 'train/grad_norm_pre_clip': 14.188247680664062, 'timing/data': 0.005022750934585929, 'timing/model': 1.2057229130296037, 'learning_rate/qwen_vl_interface': 1.5073927919014986e-06, 'learning_rate/action_model': 1.5073927919014985e-05, 'epoch': 1.63}) +[2026-06-17 16:39:34,174][starVLA.training.train_starvla][INFO] - Step 5544, Loss: {'train/loss': 0.7842754125595093, 'train/grad_norm_pre_clip': 15.429266929626465, 'timing/data': 0.004946150002069771, 'timing/model': 1.2023741529555991, 'learning_rate/qwen_vl_interface': 1.506061435957614e-06, 'learning_rate/action_model': 1.506061435957614e-05, 'epoch': 1.63}) +[2026-06-17 16:39:38,944][starVLA.training.train_starvla][INFO] - Step 5545, Loss: {'train/loss': 0.873557984828949, 'train/grad_norm_pre_clip': 18.03993797302246, 'timing/data': 0.004949476919136941, 'timing/model': 1.202330478001386, 'learning_rate/qwen_vl_interface': 1.5047308561362225e-06, 'learning_rate/action_model': 1.5047308561362225e-05, 'epoch': 1.63}) +[2026-06-17 16:39:43,718][starVLA.training.train_starvla][INFO] - Step 5546, Loss: {'train/loss': 0.39542171359062195, 'train/grad_norm_pre_clip': 10.600866317749023, 'timing/data': 0.004975004005245864, 'timing/model': 1.2055691259447485, 'learning_rate/qwen_vl_interface': 1.5034010527131548e-06, 'learning_rate/action_model': 1.5034010527131549e-05, 'epoch': 1.63}) +[2026-06-17 16:39:48,462][starVLA.training.train_starvla][INFO] - Step 5547, Loss: {'train/loss': 0.3962933123111725, 'train/grad_norm_pre_clip': 10.159939765930176, 'timing/data': 0.004967062035575509, 'timing/model': 1.2894608200294897, 'learning_rate/qwen_vl_interface': 1.5020720259640792e-06, 'learning_rate/action_model': 1.5020720259640792e-05, 'epoch': 1.63}) +[2026-06-17 16:39:53,124][starVLA.training.train_starvla][INFO] - Step 5548, Loss: {'train/loss': 0.6654738783836365, 'train/grad_norm_pre_clip': 13.740360260009766, 'timing/data': 0.005034258007071912, 'timing/model': 1.2081885050283745, 'learning_rate/qwen_vl_interface': 1.5007437761645064e-06, 'learning_rate/action_model': 1.5007437761645063e-05, 'epoch': 1.63}) +[2026-06-17 16:39:57,892][starVLA.training.train_starvla][INFO] - Step 5549, Loss: {'train/loss': 0.5943541526794434, 'train/grad_norm_pre_clip': 13.982905387878418, 'timing/data': 0.004825291922315955, 'timing/model': 1.2014635659288615, 'learning_rate/qwen_vl_interface': 1.4994163035897845e-06, 'learning_rate/action_model': 1.4994163035897846e-05, 'epoch': 1.63}) +[2026-06-17 16:40:02,677][starVLA.training.train_starvla][INFO] - Step 5550, Loss: {'train/loss': 0.8174956440925598, 'train/grad_norm_pre_clip': 17.294151306152344, 'timing/data': 0.004911748925223947, 'timing/model': 1.2040751420427114, 'learning_rate/qwen_vl_interface': 1.4980896085150976e-06, 'learning_rate/action_model': 1.4980896085150975e-05, 'epoch': 1.63}) +[2026-06-17 16:40:07,459][starVLA.training.train_starvla][INFO] - Step 5551, Loss: {'train/loss': 0.776516318321228, 'train/grad_norm_pre_clip': 15.570524215698242, 'timing/data': 0.005151589051820338, 'timing/model': 1.2024690380785614, 'learning_rate/qwen_vl_interface': 1.4967636912154707e-06, 'learning_rate/action_model': 1.4967636912154706e-05, 'epoch': 1.63}) +[2026-06-17 16:40:12,254][starVLA.training.train_starvla][INFO] - Step 5552, Loss: {'train/loss': 0.728834867477417, 'train/grad_norm_pre_clip': 13.082605361938477, 'timing/data': 0.00502133194822818, 'timing/model': 1.3403208010131493, 'learning_rate/qwen_vl_interface': 1.4954385519657701e-06, 'learning_rate/action_model': 1.49543855196577e-05, 'epoch': 1.63}) +[2026-06-17 16:40:16,907][starVLA.training.train_starvla][INFO] - Step 5553, Loss: {'train/loss': 0.8264095187187195, 'train/grad_norm_pre_clip': 15.053296089172363, 'timing/data': 0.0049111900152638555, 'timing/model': 1.2050923040369526, 'learning_rate/qwen_vl_interface': 1.4941141910406984e-06, 'learning_rate/action_model': 1.4941141910406984e-05, 'epoch': 1.63}) +[2026-06-17 16:40:21,689][starVLA.training.train_starvla][INFO] - Step 5554, Loss: {'train/loss': 0.8834405541419983, 'train/grad_norm_pre_clip': 15.45861530303955, 'timing/data': 0.0047996219946071506, 'timing/model': 1.2052575879497454, 'learning_rate/qwen_vl_interface': 1.4927906087147966e-06, 'learning_rate/action_model': 1.4927906087147967e-05, 'epoch': 1.63}) +[2026-06-17 16:40:26,561][starVLA.training.train_starvla][INFO] - Step 5555, Loss: {'train/loss': 0.7649016976356506, 'train/grad_norm_pre_clip': 14.651962280273438, 'timing/data': 0.004380585043691099, 'timing/model': 1.2061717009637505, 'learning_rate/qwen_vl_interface': 1.491467805262444e-06, 'learning_rate/action_model': 1.4914678052624439e-05, 'epoch': 1.63}) +[2026-06-17 16:40:31,272][starVLA.training.train_starvla][INFO] - Step 5556, Loss: {'train/loss': 0.6195098757743835, 'train/grad_norm_pre_clip': 11.896526336669922, 'timing/data': 0.0049159349873661995, 'timing/model': 1.204450513003394, 'learning_rate/qwen_vl_interface': 1.4901457809578598e-06, 'learning_rate/action_model': 1.4901457809578597e-05, 'epoch': 1.63}) +[2026-06-17 16:40:36,019][starVLA.training.train_starvla][INFO] - Step 5557, Loss: {'train/loss': 0.4698895812034607, 'train/grad_norm_pre_clip': 10.403595924377441, 'timing/data': 0.004837858956307173, 'timing/model': 1.2046527799684554, 'learning_rate/qwen_vl_interface': 1.4888245360751014e-06, 'learning_rate/action_model': 1.4888245360751013e-05, 'epoch': 1.64}) +[2026-06-17 16:40:40,835][starVLA.training.train_starvla][INFO] - Step 5558, Loss: {'train/loss': 1.0245357751846313, 'train/grad_norm_pre_clip': 13.886422157287598, 'timing/data': 0.004551071091555059, 'timing/model': 1.350127000012435, 'learning_rate/qwen_vl_interface': 1.4875040708880631e-06, 'learning_rate/action_model': 1.487504070888063e-05, 'epoch': 1.64}) +[2026-06-17 16:40:45,500][starVLA.training.train_starvla][INFO] - Step 5559, Loss: {'train/loss': 0.6258583068847656, 'train/grad_norm_pre_clip': 15.9766206741333, 'timing/data': 0.005280351964756846, 'timing/model': 1.2058136109262705, 'learning_rate/qwen_vl_interface': 1.4861843856704802e-06, 'learning_rate/action_model': 1.4861843856704801e-05, 'epoch': 1.64}) +[2026-06-17 16:40:50,288][starVLA.training.train_starvla][INFO] - Step 5560, Loss: {'train/loss': 0.38638243079185486, 'train/grad_norm_pre_clip': 9.05114459991455, 'timing/data': 0.0048960159765556455, 'timing/model': 1.2051547690061852, 'learning_rate/qwen_vl_interface': 1.4848654806959243e-06, 'learning_rate/action_model': 1.4848654806959242e-05, 'epoch': 1.64}) +[2026-06-17 16:40:55,089][starVLA.training.train_starvla][INFO] - Step 5561, Loss: {'train/loss': 0.5698093175888062, 'train/grad_norm_pre_clip': 13.15910530090332, 'timing/data': 0.004575483966618776, 'timing/model': 1.207933355937712, 'learning_rate/qwen_vl_interface': 1.483547356237805e-06, 'learning_rate/action_model': 1.4835473562378049e-05, 'epoch': 1.64}) +[2026-06-17 16:40:59,871][starVLA.training.train_starvla][INFO] - Step 5562, Loss: {'train/loss': 0.6816392540931702, 'train/grad_norm_pre_clip': 16.9931640625, 'timing/data': 0.005003594909794629, 'timing/model': 1.2058907090686262, 'learning_rate/qwen_vl_interface': 1.4822300125693709e-06, 'learning_rate/action_model': 1.4822300125693708e-05, 'epoch': 1.64}) +[2026-06-17 16:41:04,616][starVLA.training.train_starvla][INFO] - Step 5563, Loss: {'train/loss': 0.8021418452262878, 'train/grad_norm_pre_clip': 16.27838706970215, 'timing/data': 0.005060920026153326, 'timing/model': 1.2059550420381129, 'learning_rate/qwen_vl_interface': 1.480913449963712e-06, 'learning_rate/action_model': 1.480913449963712e-05, 'epoch': 1.64}) +[2026-06-17 16:41:09,397][starVLA.training.train_starvla][INFO] - Step 5564, Loss: {'train/loss': 0.7506626844406128, 'train/grad_norm_pre_clip': 14.016938209533691, 'timing/data': 0.004986181971617043, 'timing/model': 1.3254799980204552, 'learning_rate/qwen_vl_interface': 1.4795976686937488e-06, 'learning_rate/action_model': 1.4795976686937487e-05, 'epoch': 1.64}) +[2026-06-17 16:41:14,058][starVLA.training.train_starvla][INFO] - Step 5565, Loss: {'train/loss': 0.5298619270324707, 'train/grad_norm_pre_clip': 12.360762596130371, 'timing/data': 0.005146000999957323, 'timing/model': 1.2047063700156286, 'learning_rate/qwen_vl_interface': 1.4782826690322453e-06, 'learning_rate/action_model': 1.4782826690322451e-05, 'epoch': 1.64}) +[2026-06-17 16:41:18,846][starVLA.training.train_starvla][INFO] - Step 5566, Loss: {'train/loss': 0.8790731430053711, 'train/grad_norm_pre_clip': 16.438993453979492, 'timing/data': 0.004775744047947228, 'timing/model': 1.2057033920427784, 'learning_rate/qwen_vl_interface': 1.4769684512518045e-06, 'learning_rate/action_model': 1.4769684512518045e-05, 'epoch': 1.64}) +[2026-06-17 16:41:23,631][starVLA.training.train_starvla][INFO] - Step 5567, Loss: {'train/loss': 0.5326512455940247, 'train/grad_norm_pre_clip': 12.695779800415039, 'timing/data': 0.004980911035090685, 'timing/model': 1.2052465190645307, 'learning_rate/qwen_vl_interface': 1.4756550156248633e-06, 'learning_rate/action_model': 1.4756550156248632e-05, 'epoch': 1.64}) +[2026-06-17 16:41:28,433][starVLA.training.train_starvla][INFO] - Step 5568, Loss: {'train/loss': 0.8998367190361023, 'train/grad_norm_pre_clip': 15.689192771911621, 'timing/data': 0.005103657022118568, 'timing/model': 1.2080378630198538, 'learning_rate/qwen_vl_interface': 1.4743423624237e-06, 'learning_rate/action_model': 1.4743423624236998e-05, 'epoch': 1.64}) +[2026-06-17 16:41:33,233][starVLA.training.train_starvla][INFO] - Step 5569, Loss: {'train/loss': 0.8787400126457214, 'train/grad_norm_pre_clip': 16.8748779296875, 'timing/data': 0.004684471990913153, 'timing/model': 1.343049522023648, 'learning_rate/qwen_vl_interface': 1.4730304919204253e-06, 'learning_rate/action_model': 1.4730304919204252e-05, 'epoch': 1.64}) +[2026-06-17 16:41:37,915][starVLA.training.train_starvla][INFO] - Step 5570, Loss: {'train/loss': 0.62276291847229, 'train/grad_norm_pre_clip': 22.880109786987305, 'timing/data': 0.004916733945719898, 'timing/model': 1.2156106210313737, 'learning_rate/qwen_vl_interface': 1.4717194043869953e-06, 'learning_rate/action_model': 1.4717194043869953e-05, 'epoch': 1.64}) +[2026-06-17 16:41:42,710][starVLA.training.train_starvla][INFO] - Step 5571, Loss: {'train/loss': 0.6948546767234802, 'train/grad_norm_pre_clip': 14.410040855407715, 'timing/data': 0.004742126096971333, 'timing/model': 1.2041066450765356, 'learning_rate/qwen_vl_interface': 1.4704091000951982e-06, 'learning_rate/action_model': 1.4704091000951981e-05, 'epoch': 1.64}) +[2026-06-17 16:41:47,422][starVLA.training.train_starvla][INFO] - Step 5572, Loss: {'train/loss': 0.5912158489227295, 'train/grad_norm_pre_clip': 11.522809982299805, 'timing/data': 0.004817144013941288, 'timing/model': 1.204386848025024, 'learning_rate/qwen_vl_interface': 1.4690995793166615e-06, 'learning_rate/action_model': 1.4690995793166616e-05, 'epoch': 1.64}) +[2026-06-17 16:41:52,165][starVLA.training.train_starvla][INFO] - Step 5573, Loss: {'train/loss': 0.7511376142501831, 'train/grad_norm_pre_clip': 13.790802955627441, 'timing/data': 0.0049865590408444405, 'timing/model': 1.2055031219497323, 'learning_rate/qwen_vl_interface': 1.4677908423228506e-06, 'learning_rate/action_model': 1.4677908423228507e-05, 'epoch': 1.64}) +[2026-06-17 16:41:56,954][starVLA.training.train_starvla][INFO] - Step 5574, Loss: {'train/loss': 0.751602828502655, 'train/grad_norm_pre_clip': 13.514460563659668, 'timing/data': 0.004870436969213188, 'timing/model': 1.2073889049934223, 'learning_rate/qwen_vl_interface': 1.4664828893850677e-06, 'learning_rate/action_model': 1.4664828893850676e-05, 'epoch': 1.64}) +[2026-06-17 16:42:01,745][starVLA.training.train_starvla][INFO] - Step 5575, Loss: {'train/loss': 0.5734057426452637, 'train/grad_norm_pre_clip': 11.818589210510254, 'timing/data': 0.004858572036027908, 'timing/model': 1.3378409040160477, 'learning_rate/qwen_vl_interface': 1.4651757207744539e-06, 'learning_rate/action_model': 1.4651757207744537e-05, 'epoch': 1.64}) +[2026-06-17 16:42:06,402][starVLA.training.train_starvla][INFO] - Step 5576, Loss: {'train/loss': 0.45931705832481384, 'train/grad_norm_pre_clip': 12.911293983459473, 'timing/data': 0.004920257022604346, 'timing/model': 1.2066421440104023, 'learning_rate/qwen_vl_interface': 1.4638693367619837e-06, 'learning_rate/action_model': 1.4638693367619834e-05, 'epoch': 1.64}) +[2026-06-17 16:42:11,195][starVLA.training.train_starvla][INFO] - Step 5577, Loss: {'train/loss': 0.60276198387146, 'train/grad_norm_pre_clip': 14.666680335998535, 'timing/data': 0.004743361030705273, 'timing/model': 1.2075495490571484, 'learning_rate/qwen_vl_interface': 1.4625637376184768e-06, 'learning_rate/action_model': 1.4625637376184768e-05, 'epoch': 1.64}) +[2026-06-17 16:42:15,978][starVLA.training.train_starvla][INFO] - Step 5578, Loss: {'train/loss': 0.30664658546447754, 'train/grad_norm_pre_clip': 9.899948120117188, 'timing/data': 0.005158211104571819, 'timing/model': 1.2042823600349948, 'learning_rate/qwen_vl_interface': 1.4612589236145811e-06, 'learning_rate/action_model': 1.461258923614581e-05, 'epoch': 1.64}) +[2026-06-17 16:42:20,722][starVLA.training.train_starvla][INFO] - Step 5579, Loss: {'train/loss': 0.7937483787536621, 'train/grad_norm_pre_clip': 16.46117401123047, 'timing/data': 0.004939122009091079, 'timing/model': 1.2051589399343356, 'learning_rate/qwen_vl_interface': 1.4599548950207878e-06, 'learning_rate/action_model': 1.4599548950207878e-05, 'epoch': 1.64}) +[2026-06-17 16:42:25,559][starVLA.training.train_starvla][INFO] - Step 5580, Loss: {'train/loss': 0.4447413384914398, 'train/grad_norm_pre_clip': 11.969727516174316, 'timing/data': 0.005677707958966494, 'timing/model': 1.2032750570215285, 'learning_rate/qwen_vl_interface': 1.4586516521074214e-06, 'learning_rate/action_model': 1.4586516521074214e-05, 'epoch': 1.64}) +[2026-06-17 16:42:30,405][starVLA.training.train_starvla][INFO] - Step 5581, Loss: {'train/loss': 0.544705331325531, 'train/grad_norm_pre_clip': 17.32020378112793, 'timing/data': 0.0047185400035232306, 'timing/model': 1.3460513040190563, 'learning_rate/qwen_vl_interface': 1.457349195144649e-06, 'learning_rate/action_model': 1.457349195144649e-05, 'epoch': 1.64}) +[2026-06-17 16:42:35,063][starVLA.training.train_starvla][INFO] - Step 5582, Loss: {'train/loss': 0.6547617316246033, 'train/grad_norm_pre_clip': 14.09703254699707, 'timing/data': 0.004944437998346984, 'timing/model': 1.2034742049872875, 'learning_rate/qwen_vl_interface': 1.4560475244024702e-06, 'learning_rate/action_model': 1.4560475244024702e-05, 'epoch': 1.64}) +[2026-06-17 16:42:39,864][starVLA.training.train_starvla][INFO] - Step 5583, Loss: {'train/loss': 0.5292778611183167, 'train/grad_norm_pre_clip': 14.255495071411133, 'timing/data': 0.004867682931944728, 'timing/model': 1.2126952219987288, 'learning_rate/qwen_vl_interface': 1.4547466401507213e-06, 'learning_rate/action_model': 1.4547466401507211e-05, 'epoch': 1.64}) +[2026-06-17 16:42:44,757][starVLA.training.train_starvla][INFO] - Step 5584, Loss: {'train/loss': 0.7503209114074707, 'train/grad_norm_pre_clip': 14.6776762008667, 'timing/data': 0.005273960996419191, 'timing/model': 1.2039243089966476, 'learning_rate/qwen_vl_interface': 1.4534465426590763e-06, 'learning_rate/action_model': 1.4534465426590763e-05, 'epoch': 1.64}) +[2026-06-17 16:42:49,557][starVLA.training.train_starvla][INFO] - Step 5585, Loss: {'train/loss': 0.7517186999320984, 'train/grad_norm_pre_clip': 13.759171485900879, 'timing/data': 0.005270099034532905, 'timing/model': 1.212283622007817, 'learning_rate/qwen_vl_interface': 1.4521472321970496e-06, 'learning_rate/action_model': 1.4521472321970494e-05, 'epoch': 1.64}) +[2026-06-17 16:42:54,246][starVLA.training.train_starvla][INFO] - Step 5586, Loss: {'train/loss': 1.0005455017089844, 'train/grad_norm_pre_clip': 20.014822006225586, 'timing/data': 0.005071612074971199, 'timing/model': 1.2052630520192906, 'learning_rate/qwen_vl_interface': 1.450848709033988e-06, 'learning_rate/action_model': 1.4508487090339878e-05, 'epoch': 1.64}) +[2026-06-17 16:42:58,947][starVLA.training.train_starvla][INFO] - Step 5587, Loss: {'train/loss': 0.7381682395935059, 'train/grad_norm_pre_clip': 18.553020477294922, 'timing/data': 0.0050623309798538685, 'timing/model': 1.2463578230235726, 'learning_rate/qwen_vl_interface': 1.4495509734390769e-06, 'learning_rate/action_model': 1.4495509734390769e-05, 'epoch': 1.64}) +[2026-06-17 16:43:03,600][starVLA.training.train_starvla][INFO] - Step 5588, Loss: {'train/loss': 0.4485519528388977, 'train/grad_norm_pre_clip': 10.287127494812012, 'timing/data': 0.004916886915452778, 'timing/model': 1.2045626390026882, 'learning_rate/qwen_vl_interface': 1.448254025681339e-06, 'learning_rate/action_model': 1.448254025681339e-05, 'epoch': 1.64}) +[2026-06-17 16:43:08,354][starVLA.training.train_starvla][INFO] - Step 5589, Loss: {'train/loss': 0.4965835213661194, 'train/grad_norm_pre_clip': 9.585912704467773, 'timing/data': 0.004956888034939766, 'timing/model': 1.2049866230227053, 'learning_rate/qwen_vl_interface': 1.4469578660296327e-06, 'learning_rate/action_model': 1.4469578660296326e-05, 'epoch': 1.64}) +[2026-06-17 16:43:13,155][starVLA.training.train_starvla][INFO] - Step 5590, Loss: {'train/loss': 0.7670475840568542, 'train/grad_norm_pre_clip': 22.27562141418457, 'timing/data': 0.0049476929707452655, 'timing/model': 1.2053140429779887, 'learning_rate/qwen_vl_interface': 1.4456624947526524e-06, 'learning_rate/action_model': 1.4456624947526524e-05, 'epoch': 1.64}) +[2026-06-17 16:43:17,946][starVLA.training.train_starvla][INFO] - Step 5591, Loss: {'train/loss': 0.46908456087112427, 'train/grad_norm_pre_clip': 9.627408027648926, 'timing/data': 0.004604061017744243, 'timing/model': 1.2045975850196555, 'learning_rate/qwen_vl_interface': 1.4443679121189306e-06, 'learning_rate/action_model': 1.4443679121189305e-05, 'epoch': 1.65}) +[2026-06-17 16:43:22,730][starVLA.training.train_starvla][INFO] - Step 5592, Loss: {'train/loss': 0.7622020244598389, 'train/grad_norm_pre_clip': 13.484338760375977, 'timing/data': 0.005035627982579172, 'timing/model': 1.3331466959789395, 'learning_rate/qwen_vl_interface': 1.4430741183968361e-06, 'learning_rate/action_model': 1.443074118396836e-05, 'epoch': 1.65}) +[2026-06-17 16:43:27,444][starVLA.training.train_starvla][INFO] - Step 5593, Loss: {'train/loss': 0.438590943813324, 'train/grad_norm_pre_clip': 11.467069625854492, 'timing/data': 0.004667097004130483, 'timing/model': 1.2571520700585097, 'learning_rate/qwen_vl_interface': 1.441781113854573e-06, 'learning_rate/action_model': 1.4417811138545728e-05, 'epoch': 1.65}) +[2026-06-17 16:43:32,103][starVLA.training.train_starvla][INFO] - Step 5594, Loss: {'train/loss': 0.5395681262016296, 'train/grad_norm_pre_clip': 18.96500015258789, 'timing/data': 0.0050466719549149275, 'timing/model': 1.2056629480794072, 'learning_rate/qwen_vl_interface': 1.4404888987601828e-06, 'learning_rate/action_model': 1.4404888987601828e-05, 'epoch': 1.65}) +[2026-06-17 16:43:36,863][starVLA.training.train_starvla][INFO] - Step 5595, Loss: {'train/loss': 0.5773751735687256, 'train/grad_norm_pre_clip': 13.343905448913574, 'timing/data': 0.004481047973968089, 'timing/model': 1.2079694180283695, 'learning_rate/qwen_vl_interface': 1.4391974733815417e-06, 'learning_rate/action_model': 1.4391974733815416e-05, 'epoch': 1.65}) +[2026-06-17 16:43:41,656][starVLA.training.train_starvla][INFO] - Step 5596, Loss: {'train/loss': 0.5487130284309387, 'train/grad_norm_pre_clip': 14.786977767944336, 'timing/data': 0.004982006968930364, 'timing/model': 1.207354906015098, 'learning_rate/qwen_vl_interface': 1.437906837986368e-06, 'learning_rate/action_model': 1.4379068379863677e-05, 'epoch': 1.65}) +[2026-06-17 16:43:46,460][starVLA.training.train_starvla][INFO] - Step 5597, Loss: {'train/loss': 0.5696660876274109, 'train/grad_norm_pre_clip': 11.923425674438477, 'timing/data': 0.004861349007114768, 'timing/model': 1.2062199430074543, 'learning_rate/qwen_vl_interface': 1.436616992842207e-06, 'learning_rate/action_model': 1.4366169928422068e-05, 'epoch': 1.65}) +[2026-06-17 16:43:51,247][starVLA.training.train_starvla][INFO] - Step 5598, Loss: {'train/loss': 0.6533157229423523, 'train/grad_norm_pre_clip': 15.739444732666016, 'timing/data': 0.005001208046451211, 'timing/model': 1.3325836600270122, 'learning_rate/qwen_vl_interface': 1.4353279382164459e-06, 'learning_rate/action_model': 1.4353279382164458e-05, 'epoch': 1.65}) +[2026-06-17 16:43:56,030][starVLA.training.train_starvla][INFO] - Step 5599, Loss: {'train/loss': 0.4417227506637573, 'train/grad_norm_pre_clip': 10.81777572631836, 'timing/data': 0.005252850009128451, 'timing/model': 1.32766761793755, 'learning_rate/qwen_vl_interface': 1.4340396743763084e-06, 'learning_rate/action_model': 1.4340396743763085e-05, 'epoch': 1.65}) +[2026-06-17 16:45:37,991][starVLA.training.train_starvla][INFO] - Step 5600, Loss: {'train/loss': 0.77187180519104, 'train/grad_norm_pre_clip': 15.253852844238281, 'eval/loss': 0.7078637480735779, 'eval/latency_0/loss': 0.7078637480735779, 'eval/demon_fire_precision': 0.909312728751525, 'eval/demon_fire_recall': 0.9285714285714286, 'eval/demon_fire_f1': 0.9188411752619684, 'eval/demon_fire_rate_model': 0.6838153503893215, 'eval/demon_fire_rate_teacher': 0.6696329254727474, 'eval/demon_left_precision': 0.835203780271707, 'eval/demon_left_recall': 0.8011331444759207, 'eval/demon_left_f1': 0.8178137651821862, 'eval/demon_left_rate_model': 0.47080088987764185, 'eval/demon_left_rate_teacher': 0.49082313681868744, 'eval/demon_right_precision': 0.8127682403433476, 'eval/demon_right_recall': 0.8559322033898306, 'eval/demon_right_f1': 0.8337919647771052, 'eval/demon_right_rate_model': 0.5183537263626251, 'eval/demon_right_rate_teacher': 0.4922135706340378, 'eval/demon_move_f1': 0.8258028649796456, 'eval/demon_cc_f1': 0.8888004141256467, 'eval/latency_0/demon_fire_precision': 0.909312728751525, 'eval/latency_0/demon_fire_recall': 0.9285714285714286, 'eval/latency_0/demon_fire_f1': 0.9188411752619684, 'eval/latency_0/demon_fire_rate_model': 0.6838153503893215, 'eval/latency_0/demon_fire_rate_teacher': 0.6696329254727474, 'eval/latency_0/demon_left_precision': 0.835203780271707, 'eval/latency_0/demon_left_recall': 0.8011331444759207, 'eval/latency_0/demon_left_f1': 0.8178137651821862, 'eval/latency_0/demon_left_rate_model': 0.47080088987764185, 'eval/latency_0/demon_left_rate_teacher': 0.49082313681868744, 'eval/latency_0/demon_right_precision': 0.8127682403433476, 'eval/latency_0/demon_right_recall': 0.8559322033898306, 'eval/latency_0/demon_right_f1': 0.8337919647771052, 'eval/latency_0/demon_right_rate_model': 0.5183537263626251, 'eval/latency_0/demon_right_rate_teacher': 0.4922135706340378, 'eval/latency_0/demon_move_f1': 0.8258028649796456, 'eval/latency_0/demon_cc_f1': 0.8888004141256467, 'timing/data': 0.004834549035876989, 'timing/model': 1.2126556519651785, 'learning_rate/qwen_vl_interface': 1.4327522015888535e-06, 'learning_rate/action_model': 1.4327522015888534e-05, 'epoch': 1.65}) +[2026-06-17 16:45:37,991][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_5600_state +[2026-06-17 16:45:38,002][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 16:45:41,559][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_5600_state/model.safetensors +[2026-06-17 16:45:49,140][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_5600_state/optimizer.bin +[2026-06-17 16:45:49,140][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_5600_state/sampler.bin +[2026-06-17 16:45:49,140][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_5600_state/sampler_1.bin +[2026-06-17 16:45:49,141][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_5600_state/random_states_0.pkl +[2026-06-17 16:45:49,191][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 5600 +[2026-06-17 16:45:49,191][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 16:45:49,201][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 16:45:53,788][starVLA.training.train_starvla][INFO] - Step 5601, Loss: {'train/loss': 0.5062482953071594, 'train/grad_norm_pre_clip': 12.626326560974121, 'timing/data': 0.005453074933029711, 'timing/model': 1.1910900269867852, 'learning_rate/qwen_vl_interface': 1.4314655201209756e-06, 'learning_rate/action_model': 1.4314655201209757e-05, 'epoch': 1.65}) +[2026-06-17 16:45:58,535][starVLA.training.train_starvla][INFO] - Step 5602, Loss: {'train/loss': 0.6611212491989136, 'train/grad_norm_pre_clip': 16.220897674560547, 'timing/data': 0.00529315194580704, 'timing/model': 1.1917863310081884, 'learning_rate/qwen_vl_interface': 1.4301796302394019e-06, 'learning_rate/action_model': 1.4301796302394018e-05, 'epoch': 1.65}) +[2026-06-17 16:46:03,155][starVLA.training.train_starvla][INFO] - Step 5603, Loss: {'train/loss': 0.622260332107544, 'train/grad_norm_pre_clip': 11.699877738952637, 'timing/data': 0.005032998975366354, 'timing/model': 1.1931735639227554, 'learning_rate/qwen_vl_interface': 1.4288945322107025e-06, 'learning_rate/action_model': 1.4288945322107026e-05, 'epoch': 1.65}) +[2026-06-17 16:46:07,786][starVLA.training.train_starvla][INFO] - Step 5604, Loss: {'train/loss': 0.44012200832366943, 'train/grad_norm_pre_clip': 13.905097961425781, 'timing/data': 0.005281177000142634, 'timing/model': 1.1933434600941837, 'learning_rate/qwen_vl_interface': 1.4276102263012781e-06, 'learning_rate/action_model': 1.427610226301278e-05, 'epoch': 1.65}) +[2026-06-17 16:46:12,455][starVLA.training.train_starvla][INFO] - Step 5605, Loss: {'train/loss': 0.7327835559844971, 'train/grad_norm_pre_clip': 12.311027526855469, 'timing/data': 0.005258693010546267, 'timing/model': 1.195646922918968, 'learning_rate/qwen_vl_interface': 1.426326712777367e-06, 'learning_rate/action_model': 1.4263267127773669e-05, 'epoch': 1.65}) +[2026-06-17 16:46:17,210][starVLA.training.train_starvla][INFO] - Step 5606, Loss: {'train/loss': 0.7123247981071472, 'train/grad_norm_pre_clip': 14.674489974975586, 'timing/data': 0.005246603977866471, 'timing/model': 1.32484693499282, 'learning_rate/qwen_vl_interface': 1.425043991905043e-06, 'learning_rate/action_model': 1.425043991905043e-05, 'epoch': 1.65}) +[2026-06-17 16:46:21,857][starVLA.training.train_starvla][INFO] - Step 5607, Loss: {'train/loss': 0.8058169484138489, 'train/grad_norm_pre_clip': 17.121135711669922, 'timing/data': 0.0049578300677239895, 'timing/model': 1.2045538949314505, 'learning_rate/qwen_vl_interface': 1.4237620639502149e-06, 'learning_rate/action_model': 1.4237620639502147e-05, 'epoch': 1.65}) +[2026-06-17 16:46:26,626][starVLA.training.train_starvla][INFO] - Step 5608, Loss: {'train/loss': 0.6525447964668274, 'train/grad_norm_pre_clip': 14.482966423034668, 'timing/data': 0.005114407977089286, 'timing/model': 1.2110789659200236, 'learning_rate/qwen_vl_interface': 1.4224809291786281e-06, 'learning_rate/action_model': 1.4224809291786282e-05, 'epoch': 1.65}) +[2026-06-17 16:46:31,468][starVLA.training.train_starvla][INFO] - Step 5609, Loss: {'train/loss': 0.3605358302593231, 'train/grad_norm_pre_clip': 9.99025821685791, 'timing/data': 0.005320962984114885, 'timing/model': 1.199266577954404, 'learning_rate/qwen_vl_interface': 1.4212005878558622e-06, 'learning_rate/action_model': 1.4212005878558622e-05, 'epoch': 1.65}) +[2026-06-17 16:46:36,253][starVLA.training.train_starvla][INFO] - Step 5610, Loss: {'train/loss': 0.49273958802223206, 'train/grad_norm_pre_clip': 12.690953254699707, 'timing/data': 0.00482197804376483, 'timing/model': 1.2064050229964778, 'learning_rate/qwen_vl_interface': 1.4199210402473368e-06, 'learning_rate/action_model': 1.4199210402473367e-05, 'epoch': 1.65}) +[2026-06-17 16:46:41,017][starVLA.training.train_starvla][INFO] - Step 5611, Loss: {'train/loss': 0.6893037557601929, 'train/grad_norm_pre_clip': 14.865317344665527, 'timing/data': 0.004931490984745324, 'timing/model': 1.3263888970250264, 'learning_rate/qwen_vl_interface': 1.4186422866183e-06, 'learning_rate/action_model': 1.4186422866182999e-05, 'epoch': 1.65}) +[2026-06-17 16:46:45,735][starVLA.training.train_starvla][INFO] - Step 5612, Loss: {'train/loss': 0.7169220447540283, 'train/grad_norm_pre_clip': 13.385653495788574, 'timing/data': 0.005212276941165328, 'timing/model': 1.277155683026649, 'learning_rate/qwen_vl_interface': 1.41736432723384e-06, 'learning_rate/action_model': 1.41736432723384e-05, 'epoch': 1.65}) +[2026-06-17 16:46:50,374][starVLA.training.train_starvla][INFO] - Step 5613, Loss: {'train/loss': 0.5750375986099243, 'train/grad_norm_pre_clip': 12.981376647949219, 'timing/data': 0.005138718988746405, 'timing/model': 1.202438629930839, 'learning_rate/qwen_vl_interface': 1.4160871623588774e-06, 'learning_rate/action_model': 1.4160871623588772e-05, 'epoch': 1.65}) +[2026-06-17 16:46:55,156][starVLA.training.train_starvla][INFO] - Step 5614, Loss: {'train/loss': 0.6469904184341431, 'train/grad_norm_pre_clip': 11.950554847717285, 'timing/data': 0.005067870952188969, 'timing/model': 1.2021523050498217, 'learning_rate/qwen_vl_interface': 1.414810792258173e-06, 'learning_rate/action_model': 1.414810792258173e-05, 'epoch': 1.65}) +[2026-06-17 16:46:59,940][starVLA.training.train_starvla][INFO] - Step 5615, Loss: {'train/loss': 0.7727215886116028, 'train/grad_norm_pre_clip': 15.117911338806152, 'timing/data': 0.005185359972529113, 'timing/model': 1.1996295580174774, 'learning_rate/qwen_vl_interface': 1.4135352171963186e-06, 'learning_rate/action_model': 1.4135352171963185e-05, 'epoch': 1.65}) +[2026-06-17 16:47:04,743][starVLA.training.train_starvla][INFO] - Step 5616, Loss: {'train/loss': 0.5995147228240967, 'train/grad_norm_pre_clip': 15.666425704956055, 'timing/data': 0.007100899936631322, 'timing/model': 1.2202900550328195, 'learning_rate/qwen_vl_interface': 1.412260437437741e-06, 'learning_rate/action_model': 1.412260437437741e-05, 'epoch': 1.65}) +[2026-06-17 16:47:09,511][starVLA.training.train_starvla][INFO] - Step 5617, Loss: {'train/loss': 0.957641065120697, 'train/grad_norm_pre_clip': 17.197547912597656, 'timing/data': 0.00517365790437907, 'timing/model': 1.1996718070004135, 'learning_rate/qwen_vl_interface': 1.4109864532467051e-06, 'learning_rate/action_model': 1.4109864532467052e-05, 'epoch': 1.65}) +[2026-06-17 16:47:14,255][starVLA.training.train_starvla][INFO] - Step 5618, Loss: {'train/loss': 0.6335457563400269, 'train/grad_norm_pre_clip': 15.161584854125977, 'timing/data': 0.005045859958045185, 'timing/model': 1.2989111930364743, 'learning_rate/qwen_vl_interface': 1.4097132648873082e-06, 'learning_rate/action_model': 1.4097132648873083e-05, 'epoch': 1.65}) +[2026-06-17 16:47:18,914][starVLA.training.train_starvla][INFO] - Step 5619, Loss: {'train/loss': 0.9888008236885071, 'train/grad_norm_pre_clip': 17.642005920410156, 'timing/data': 0.00529675604775548, 'timing/model': 1.2028516080463305, 'learning_rate/qwen_vl_interface': 1.4084408726234838e-06, 'learning_rate/action_model': 1.4084408726234838e-05, 'epoch': 1.65}) +[2026-06-17 16:47:23,694][starVLA.training.train_starvla][INFO] - Step 5620, Loss: {'train/loss': 0.5032700300216675, 'train/grad_norm_pre_clip': 14.191426277160645, 'timing/data': 0.004870963981375098, 'timing/model': 1.2027508389437571, 'learning_rate/qwen_vl_interface': 1.407169276719e-06, 'learning_rate/action_model': 1.4071692767190001e-05, 'epoch': 1.65}) +[2026-06-17 16:47:28,491][starVLA.training.train_starvla][INFO] - Step 5621, Loss: {'train/loss': 0.6733571290969849, 'train/grad_norm_pre_clip': 13.014800071716309, 'timing/data': 0.005262207007035613, 'timing/model': 1.210879779071547, 'learning_rate/qwen_vl_interface': 1.4058984774374606e-06, 'learning_rate/action_model': 1.4058984774374607e-05, 'epoch': 1.65}) +[2026-06-17 16:47:33,284][starVLA.training.train_starvla][INFO] - Step 5622, Loss: {'train/loss': 0.5395609736442566, 'train/grad_norm_pre_clip': 12.730232238769531, 'timing/data': 0.005115695996209979, 'timing/model': 1.2007623829413205, 'learning_rate/qwen_vl_interface': 1.404628475042303e-06, 'learning_rate/action_model': 1.404628475042303e-05, 'epoch': 1.65}) +[2026-06-17 16:47:38,064][starVLA.training.train_starvla][INFO] - Step 5623, Loss: {'train/loss': 0.4566524624824524, 'train/grad_norm_pre_clip': 10.693014144897461, 'timing/data': 0.005322937970049679, 'timing/model': 1.3318404210731387, 'learning_rate/qwen_vl_interface': 1.4033592697967998e-06, 'learning_rate/action_model': 1.4033592697967998e-05, 'epoch': 1.65}) +[2026-06-17 16:47:42,735][starVLA.training.train_starvla][INFO] - Step 5624, Loss: {'train/loss': 0.7546350955963135, 'train/grad_norm_pre_clip': 16.285846710205078, 'timing/data': 0.005171992001123726, 'timing/model': 1.2038412590045482, 'learning_rate/qwen_vl_interface': 1.4020908619640602e-06, 'learning_rate/action_model': 1.4020908619640602e-05, 'epoch': 1.65}) +[2026-06-17 16:47:47,511][starVLA.training.train_starvla][INFO] - Step 5625, Loss: {'train/loss': 0.6833105683326721, 'train/grad_norm_pre_clip': 14.015843391418457, 'timing/data': 0.005064442986622453, 'timing/model': 1.2035238950047642, 'learning_rate/qwen_vl_interface': 1.4008232518070248e-06, 'learning_rate/action_model': 1.4008232518070247e-05, 'epoch': 1.66}) +[2026-06-17 16:47:52,309][starVLA.training.train_starvla][INFO] - Step 5626, Loss: {'train/loss': 0.6249727010726929, 'train/grad_norm_pre_clip': 11.294699668884277, 'timing/data': 0.005291839013807476, 'timing/model': 1.2110761030344293, 'learning_rate/qwen_vl_interface': 1.3995564395884712e-06, 'learning_rate/action_model': 1.399556439588471e-05, 'epoch': 1.66}) +[2026-06-17 16:47:57,036][starVLA.training.train_starvla][INFO] - Step 5627, Loss: {'train/loss': 0.5171964168548584, 'train/grad_norm_pre_clip': 29.345352172851562, 'timing/data': 0.005234096082858741, 'timing/model': 1.2139856070280075, 'learning_rate/qwen_vl_interface': 1.3982904255710094e-06, 'learning_rate/action_model': 1.3982904255710094e-05, 'epoch': 1.66}) +[2026-06-17 16:48:01,781][starVLA.training.train_starvla][INFO] - Step 5628, Loss: {'train/loss': 0.6433744430541992, 'train/grad_norm_pre_clip': 13.646749496459961, 'timing/data': 0.005186951952055097, 'timing/model': 1.2099392359377816, 'learning_rate/qwen_vl_interface': 1.3970252100170892e-06, 'learning_rate/action_model': 1.3970252100170892e-05, 'epoch': 1.66}) +[2026-06-17 16:48:06,577][starVLA.training.train_starvla][INFO] - Step 5629, Loss: {'train/loss': 0.5952038764953613, 'train/grad_norm_pre_clip': 15.482503890991211, 'timing/data': 0.005305621074512601, 'timing/model': 1.3346391470404342, 'learning_rate/qwen_vl_interface': 1.3957607931889885e-06, 'learning_rate/action_model': 1.3957607931889884e-05, 'epoch': 1.66}) +[2026-06-17 16:48:11,288][starVLA.training.train_starvla][INFO] - Step 5630, Loss: {'train/loss': 0.5577479600906372, 'train/grad_norm_pre_clip': 11.19388198852539, 'timing/data': 0.0050384970381855965, 'timing/model': 1.2118997770594433, 'learning_rate/qwen_vl_interface': 1.3944971753488238e-06, 'learning_rate/action_model': 1.3944971753488236e-05, 'epoch': 1.66}) +[2026-06-17 16:48:16,105][starVLA.training.train_starvla][INFO] - Step 5631, Loss: {'train/loss': 0.39069873094558716, 'train/grad_norm_pre_clip': 10.876989364624023, 'timing/data': 0.005106672993861139, 'timing/model': 1.2141191059490666, 'learning_rate/qwen_vl_interface': 1.3932343567585408e-06, 'learning_rate/action_model': 1.3932343567585407e-05, 'epoch': 1.66}) +[2026-06-17 16:48:20,894][starVLA.training.train_starvla][INFO] - Step 5632, Loss: {'train/loss': 0.4854094982147217, 'train/grad_norm_pre_clip': 11.293288230895996, 'timing/data': 0.005276773008517921, 'timing/model': 1.2037088689394295, 'learning_rate/qwen_vl_interface': 1.3919723376799268e-06, 'learning_rate/action_model': 1.3919723376799269e-05, 'epoch': 1.66}) +[2026-06-17 16:48:25,664][starVLA.training.train_starvla][INFO] - Step 5633, Loss: {'train/loss': 0.8221080303192139, 'train/grad_norm_pre_clip': 18.66644859313965, 'timing/data': 0.005067189922556281, 'timing/model': 1.2033036960056052, 'learning_rate/qwen_vl_interface': 1.3907111183745992e-06, 'learning_rate/action_model': 1.3907111183745991e-05, 'epoch': 1.66}) +[2026-06-17 16:48:30,427][starVLA.training.train_starvla][INFO] - Step 5634, Loss: {'train/loss': 0.9465581774711609, 'train/grad_norm_pre_clip': 18.726762771606445, 'timing/data': 0.005030339933000505, 'timing/model': 1.2057686540065333, 'learning_rate/qwen_vl_interface': 1.3894506991040089e-06, 'learning_rate/action_model': 1.3894506991040087e-05, 'epoch': 1.66}) +[2026-06-17 16:48:35,233][starVLA.training.train_starvla][INFO] - Step 5635, Loss: {'train/loss': 0.4389793574810028, 'train/grad_norm_pre_clip': 10.36410140991211, 'timing/data': 0.004828469012863934, 'timing/model': 1.3340500589692965, 'learning_rate/qwen_vl_interface': 1.3881910801294433e-06, 'learning_rate/action_model': 1.3881910801294431e-05, 'epoch': 1.66}) +[2026-06-17 16:48:39,907][starVLA.training.train_starvla][INFO] - Step 5636, Loss: {'train/loss': 0.6223018765449524, 'train/grad_norm_pre_clip': 17.319297790527344, 'timing/data': 0.0052459759172052145, 'timing/model': 1.2069193659117445, 'learning_rate/qwen_vl_interface': 1.3869322617120223e-06, 'learning_rate/action_model': 1.3869322617120222e-05, 'epoch': 1.66}) +[2026-06-17 16:48:44,715][starVLA.training.train_starvla][INFO] - Step 5637, Loss: {'train/loss': 0.6260219216346741, 'train/grad_norm_pre_clip': 14.864639282226562, 'timing/data': 0.005202791071496904, 'timing/model': 1.2104365149280056, 'learning_rate/qwen_vl_interface': 1.3856742441127e-06, 'learning_rate/action_model': 1.3856742441126999e-05, 'epoch': 1.66}) +[2026-06-17 16:48:49,538][starVLA.training.train_starvla][INFO] - Step 5638, Loss: {'train/loss': 0.5420018434524536, 'train/grad_norm_pre_clip': 12.84835147857666, 'timing/data': 0.0052351419581100345, 'timing/model': 1.223156260093674, 'learning_rate/qwen_vl_interface': 1.384417027592264e-06, 'learning_rate/action_model': 1.3844170275922638e-05, 'epoch': 1.66}) +[2026-06-17 16:48:54,335][starVLA.training.train_starvla][INFO] - Step 5639, Loss: {'train/loss': 0.7141636610031128, 'train/grad_norm_pre_clip': 14.047747611999512, 'timing/data': 0.005289243999868631, 'timing/model': 1.2101499709533527, 'learning_rate/qwen_vl_interface': 1.3831606124113397e-06, 'learning_rate/action_model': 1.3831606124113398e-05, 'epoch': 1.66}) +[2026-06-17 16:48:59,140][starVLA.training.train_starvla][INFO] - Step 5640, Loss: {'train/loss': 0.8800167441368103, 'train/grad_norm_pre_clip': 18.232269287109375, 'timing/data': 0.005111764068715274, 'timing/model': 1.332486851955764, 'learning_rate/qwen_vl_interface': 1.3819049988303809e-06, 'learning_rate/action_model': 1.381904998830381e-05, 'epoch': 1.66}) +[2026-06-17 16:49:03,820][starVLA.training.train_starvla][INFO] - Step 5641, Loss: {'train/loss': 0.7853059768676758, 'train/grad_norm_pre_clip': 17.231779098510742, 'timing/data': 0.005783831002190709, 'timing/model': 1.2154842640738934, 'learning_rate/qwen_vl_interface': 1.3806501871096774e-06, 'learning_rate/action_model': 1.3806501871096774e-05, 'epoch': 1.66}) +[2026-06-17 16:49:08,620][starVLA.training.train_starvla][INFO] - Step 5642, Loss: {'train/loss': 0.6236255168914795, 'train/grad_norm_pre_clip': 13.782767295837402, 'timing/data': 0.0052798070246353745, 'timing/model': 1.2034969020169228, 'learning_rate/qwen_vl_interface': 1.3793961775093528e-06, 'learning_rate/action_model': 1.3793961775093528e-05, 'epoch': 1.66}) +[2026-06-17 16:49:13,349][starVLA.training.train_starvla][INFO] - Step 5643, Loss: {'train/loss': 0.4518970847129822, 'train/grad_norm_pre_clip': 11.714881896972656, 'timing/data': 0.004805518081411719, 'timing/model': 1.2126313979970291, 'learning_rate/qwen_vl_interface': 1.378142970289367e-06, 'learning_rate/action_model': 1.378142970289367e-05, 'epoch': 1.66}) +[2026-06-17 16:49:18,099][starVLA.training.train_starvla][INFO] - Step 5644, Loss: {'train/loss': 0.678970217704773, 'train/grad_norm_pre_clip': 13.302268981933594, 'timing/data': 0.005279420991428196, 'timing/model': 1.204297918942757, 'learning_rate/qwen_vl_interface': 1.3768905657095109e-06, 'learning_rate/action_model': 1.3768905657095107e-05, 'epoch': 1.66}) +[2026-06-17 16:49:22,885][starVLA.training.train_starvla][INFO] - Step 5645, Loss: {'train/loss': 0.5470920205116272, 'train/grad_norm_pre_clip': 12.45195198059082, 'timing/data': 0.005069555016234517, 'timing/model': 1.2030735419830307, 'learning_rate/qwen_vl_interface': 1.3756389640294066e-06, 'learning_rate/action_model': 1.3756389640294064e-05, 'epoch': 1.66}) +[2026-06-17 16:49:27,690][starVLA.training.train_starvla][INFO] - Step 5646, Loss: {'train/loss': 0.45609939098358154, 'train/grad_norm_pre_clip': 14.420358657836914, 'timing/data': 0.00527080602478236, 'timing/model': 1.3327471709344536, 'learning_rate/qwen_vl_interface': 1.3743881655085131e-06, 'learning_rate/action_model': 1.3743881655085131e-05, 'epoch': 1.66}) +[2026-06-17 16:49:32,347][starVLA.training.train_starvla][INFO] - Step 5647, Loss: {'train/loss': 0.6275116205215454, 'train/grad_norm_pre_clip': 13.893043518066406, 'timing/data': 0.005205834982916713, 'timing/model': 1.2039192100055516, 'learning_rate/qwen_vl_interface': 1.3731381704061248e-06, 'learning_rate/action_model': 1.3731381704061246e-05, 'epoch': 1.66}) +[2026-06-17 16:49:37,133][starVLA.training.train_starvla][INFO] - Step 5648, Loss: {'train/loss': 0.7239383459091187, 'train/grad_norm_pre_clip': 15.351923942565918, 'timing/data': 0.005051661049947143, 'timing/model': 1.2043922350276262, 'learning_rate/qwen_vl_interface': 1.3718889789813647e-06, 'learning_rate/action_model': 1.3718889789813646e-05, 'epoch': 1.66}) +[2026-06-17 16:49:41,913][starVLA.training.train_starvla][INFO] - Step 5649, Loss: {'train/loss': 0.7839587330818176, 'train/grad_norm_pre_clip': 15.19747257232666, 'timing/data': 0.0050107690040022135, 'timing/model': 1.2084389630472288, 'learning_rate/qwen_vl_interface': 1.370640591493193e-06, 'learning_rate/action_model': 1.370640591493193e-05, 'epoch': 1.66}) +[2026-06-17 16:49:46,681][starVLA.training.train_starvla][INFO] - Step 5650, Loss: {'train/loss': 0.45510995388031006, 'train/grad_norm_pre_clip': 9.999923706054688, 'timing/data': 0.005012513021938503, 'timing/model': 1.2040715799666941, 'learning_rate/qwen_vl_interface': 1.3693930082004006e-06, 'learning_rate/action_model': 1.3693930082004006e-05, 'epoch': 1.66}) +[2026-06-17 16:49:51,477][starVLA.training.train_starvla][INFO] - Step 5651, Loss: {'train/loss': 0.7079652547836304, 'train/grad_norm_pre_clip': 14.48091983795166, 'timing/data': 0.004904469009488821, 'timing/model': 1.2050340849673375, 'learning_rate/qwen_vl_interface': 1.3681462293616131e-06, 'learning_rate/action_model': 1.3681462293616131e-05, 'epoch': 1.66}) +[2026-06-17 16:49:56,276][starVLA.training.train_starvla][INFO] - Step 5652, Loss: {'train/loss': 0.6838657259941101, 'train/grad_norm_pre_clip': 14.190078735351562, 'timing/data': 0.005027423030696809, 'timing/model': 1.3433351760031655, 'learning_rate/qwen_vl_interface': 1.3669002552352898e-06, 'learning_rate/action_model': 1.3669002552352897e-05, 'epoch': 1.66}) +[2026-06-17 16:50:00,939][starVLA.training.train_starvla][INFO] - Step 5653, Loss: {'train/loss': 0.6293281316757202, 'train/grad_norm_pre_clip': 12.370603561401367, 'timing/data': 0.005296685034409165, 'timing/model': 1.2065461579477414, 'learning_rate/qwen_vl_interface': 1.3656550860797223e-06, 'learning_rate/action_model': 1.3656550860797223e-05, 'epoch': 1.66}) +[2026-06-17 16:50:05,726][starVLA.training.train_starvla][INFO] - Step 5654, Loss: {'train/loss': 0.8278059363365173, 'train/grad_norm_pre_clip': 21.29055404663086, 'timing/data': 0.00478805007878691, 'timing/model': 1.2053432560060173, 'learning_rate/qwen_vl_interface': 1.3644107221530344e-06, 'learning_rate/action_model': 1.3644107221530345e-05, 'epoch': 1.66}) +[2026-06-17 16:50:10,507][starVLA.training.train_starvla][INFO] - Step 5655, Loss: {'train/loss': 0.703394889831543, 'train/grad_norm_pre_clip': 18.527645111083984, 'timing/data': 0.004910156014375389, 'timing/model': 1.2043370300671086, 'learning_rate/qwen_vl_interface': 1.3631671637131856e-06, 'learning_rate/action_model': 1.3631671637131855e-05, 'epoch': 1.66}) +[2026-06-17 16:50:15,329][starVLA.training.train_starvla][INFO] - Step 5656, Loss: {'train/loss': 0.39448630809783936, 'train/grad_norm_pre_clip': 12.06558895111084, 'timing/data': 0.0053958260687068105, 'timing/model': 1.2096080729970708, 'learning_rate/qwen_vl_interface': 1.3619244110179655e-06, 'learning_rate/action_model': 1.3619244110179654e-05, 'epoch': 1.66}) +[2026-06-17 16:50:20,018][starVLA.training.train_starvla][INFO] - Step 5657, Loss: {'train/loss': 0.46690833568573, 'train/grad_norm_pre_clip': 11.318988800048828, 'timing/data': 0.005125510971993208, 'timing/model': 1.2171766699757427, 'learning_rate/qwen_vl_interface': 1.360682464324998e-06, 'learning_rate/action_model': 1.360682464324998e-05, 'epoch': 1.66}) +[2026-06-17 16:50:24,733][starVLA.training.train_starvla][INFO] - Step 5658, Loss: {'train/loss': 0.9579871892929077, 'train/grad_norm_pre_clip': 17.546586990356445, 'timing/data': 0.005217512953095138, 'timing/model': 1.2526689629303291, 'learning_rate/qwen_vl_interface': 1.3594413238917425e-06, 'learning_rate/action_model': 1.3594413238917425e-05, 'epoch': 1.66}) +[2026-06-17 16:50:29,405][starVLA.training.train_starvla][INFO] - Step 5659, Loss: {'train/loss': 0.5119724869728088, 'train/grad_norm_pre_clip': 11.702301979064941, 'timing/data': 0.004896164988167584, 'timing/model': 1.2068985579535365, 'learning_rate/qwen_vl_interface': 1.3582009899754864e-06, 'learning_rate/action_model': 1.3582009899754864e-05, 'epoch': 1.67}) +[2026-06-17 16:50:34,159][starVLA.training.train_starvla][INFO] - Step 5660, Loss: {'train/loss': 0.5829602479934692, 'train/grad_norm_pre_clip': 27.421281814575195, 'timing/data': 0.005031006992794573, 'timing/model': 1.2125519030960277, 'learning_rate/qwen_vl_interface': 1.3569614628333515e-06, 'learning_rate/action_model': 1.3569614628333516e-05, 'epoch': 1.67}) +[2026-06-17 16:50:38,957][starVLA.training.train_starvla][INFO] - Step 5661, Loss: {'train/loss': 0.5712783336639404, 'train/grad_norm_pre_clip': 17.48002052307129, 'timing/data': 0.005483180051669478, 'timing/model': 1.2124440469779074, 'learning_rate/qwen_vl_interface': 1.3557227427222954e-06, 'learning_rate/action_model': 1.3557227427222954e-05, 'epoch': 1.67}) +[2026-06-17 16:50:43,760][starVLA.training.train_starvla][INFO] - Step 5662, Loss: {'train/loss': 0.5707795023918152, 'train/grad_norm_pre_clip': 13.797159194946289, 'timing/data': 0.005430527031421661, 'timing/model': 1.2067970719654113, 'learning_rate/qwen_vl_interface': 1.3544848298991056e-06, 'learning_rate/action_model': 1.3544848298991056e-05, 'epoch': 1.67}) +[2026-06-17 16:50:48,548][starVLA.training.train_starvla][INFO] - Step 5663, Loss: {'train/loss': 0.5049366354942322, 'train/grad_norm_pre_clip': 12.91015625, 'timing/data': 0.005138093023560941, 'timing/model': 1.3314603000180796, 'learning_rate/qwen_vl_interface': 1.3532477246204037e-06, 'learning_rate/action_model': 1.3532477246204036e-05, 'epoch': 1.67}) +[2026-06-17 16:50:53,254][starVLA.training.train_starvla][INFO] - Step 5664, Loss: {'train/loss': 0.576799750328064, 'train/grad_norm_pre_clip': 13.195175170898438, 'timing/data': 0.005230662995018065, 'timing/model': 1.2533302459632978, 'learning_rate/qwen_vl_interface': 1.3520114271426382e-06, 'learning_rate/action_model': 1.3520114271426381e-05, 'epoch': 1.67}) +[2026-06-17 16:50:57,926][starVLA.training.train_starvla][INFO] - Step 5665, Loss: {'train/loss': 0.6439599394798279, 'train/grad_norm_pre_clip': 12.591996192932129, 'timing/data': 0.0050612930208444595, 'timing/model': 1.2141134049743414, 'learning_rate/qwen_vl_interface': 1.3507759377221001e-06, 'learning_rate/action_model': 1.3507759377221003e-05, 'epoch': 1.67}) +[2026-06-17 16:51:02,752][starVLA.training.train_starvla][INFO] - Step 5666, Loss: {'train/loss': 0.6216802000999451, 'train/grad_norm_pre_clip': 11.446381568908691, 'timing/data': 0.004742916091345251, 'timing/model': 1.2106087590800598, 'learning_rate/qwen_vl_interface': 1.3495412566149058e-06, 'learning_rate/action_model': 1.3495412566149057e-05, 'epoch': 1.67}) +[2026-06-17 16:51:07,566][starVLA.training.train_starvla][INFO] - Step 5667, Loss: {'train/loss': 0.9100004434585571, 'train/grad_norm_pre_clip': 13.509846687316895, 'timing/data': 0.004657244076952338, 'timing/model': 1.2085061869584024, 'learning_rate/qwen_vl_interface': 1.3483073840770067e-06, 'learning_rate/action_model': 1.3483073840770066e-05, 'epoch': 1.67}) +[2026-06-17 16:51:12,384][starVLA.training.train_starvla][INFO] - Step 5668, Loss: {'train/loss': 0.8628832697868347, 'train/grad_norm_pre_clip': 18.841672897338867, 'timing/data': 0.004918882041238248, 'timing/model': 1.2113992189988494, 'learning_rate/qwen_vl_interface': 1.3470743203641846e-06, 'learning_rate/action_model': 1.3470743203641844e-05, 'epoch': 1.67}) +[2026-06-17 16:51:17,182][starVLA.training.train_starvla][INFO] - Step 5669, Loss: {'train/loss': 0.6485230326652527, 'train/grad_norm_pre_clip': 13.573936462402344, 'timing/data': 0.0051975189708173275, 'timing/model': 1.3394828309537843, 'learning_rate/qwen_vl_interface': 1.3458420657320555e-06, 'learning_rate/action_model': 1.3458420657320555e-05, 'epoch': 1.67}) +[2026-06-17 16:51:21,965][starVLA.training.train_starvla][INFO] - Step 5670, Loss: {'train/loss': 0.7661598920822144, 'train/grad_norm_pre_clip': 15.733915328979492, 'timing/data': 0.005192153970710933, 'timing/model': 1.3269126969389617, 'learning_rate/qwen_vl_interface': 1.344610620436067e-06, 'learning_rate/action_model': 1.344610620436067e-05, 'epoch': 1.67}) +[2026-06-17 16:51:26,632][starVLA.training.train_starvla][INFO] - Step 5671, Loss: {'train/loss': 0.539476215839386, 'train/grad_norm_pre_clip': 11.820955276489258, 'timing/data': 0.004991823923774064, 'timing/model': 1.20489924098365, 'learning_rate/qwen_vl_interface': 1.3433799847314989e-06, 'learning_rate/action_model': 1.3433799847314988e-05, 'epoch': 1.67}) +[2026-06-17 16:51:31,399][starVLA.training.train_starvla][INFO] - Step 5672, Loss: {'train/loss': 0.5977219939231873, 'train/grad_norm_pre_clip': 12.183770179748535, 'timing/data': 0.0049663810059428215, 'timing/model': 1.2072718889685348, 'learning_rate/qwen_vl_interface': 1.3421501588734655e-06, 'learning_rate/action_model': 1.3421501588734656e-05, 'epoch': 1.67}) +[2026-06-17 16:51:36,193][starVLA.training.train_starvla][INFO] - Step 5673, Loss: {'train/loss': 0.6947723627090454, 'train/grad_norm_pre_clip': 20.03415870666504, 'timing/data': 0.005085258046165109, 'timing/model': 1.2052581389434636, 'learning_rate/qwen_vl_interface': 1.3409211431169084e-06, 'learning_rate/action_model': 1.3409211431169083e-05, 'epoch': 1.67}) +[2026-06-17 16:51:41,001][starVLA.training.train_starvla][INFO] - Step 5674, Loss: {'train/loss': 0.5401249527931213, 'train/grad_norm_pre_clip': 13.714893341064453, 'timing/data': 0.004693855065852404, 'timing/model': 1.2147947190096602, 'learning_rate/qwen_vl_interface': 1.3396929377166042e-06, 'learning_rate/action_model': 1.3396929377166043e-05, 'epoch': 1.67}) +[2026-06-17 16:51:45,746][starVLA.training.train_starvla][INFO] - Step 5675, Loss: {'train/loss': 0.6251127123832703, 'train/grad_norm_pre_clip': 13.130562782287598, 'timing/data': 0.005029160063713789, 'timing/model': 1.2746977219358087, 'learning_rate/qwen_vl_interface': 1.338465542927161e-06, 'learning_rate/action_model': 1.3384655429271608e-05, 'epoch': 1.67}) +[2026-06-17 16:51:50,416][starVLA.training.train_starvla][INFO] - Step 5676, Loss: {'train/loss': 0.6299542188644409, 'train/grad_norm_pre_clip': 16.095125198364258, 'timing/data': 0.005151675082743168, 'timing/model': 1.2056585079990327, 'learning_rate/qwen_vl_interface': 1.3372389590030204e-06, 'learning_rate/action_model': 1.3372389590030204e-05, 'epoch': 1.67}) +[2026-06-17 16:51:55,185][starVLA.training.train_starvla][INFO] - Step 5677, Loss: {'train/loss': 0.5558998584747314, 'train/grad_norm_pre_clip': 15.144699096679688, 'timing/data': 0.0055046989582479, 'timing/model': 1.2052968950010836, 'learning_rate/qwen_vl_interface': 1.3360131861984551e-06, 'learning_rate/action_model': 1.3360131861984552e-05, 'epoch': 1.67}) +[2026-06-17 16:51:59,984][starVLA.training.train_starvla][INFO] - Step 5678, Loss: {'train/loss': 0.5452049374580383, 'train/grad_norm_pre_clip': 13.32143783569336, 'timing/data': 0.0053411280969157815, 'timing/model': 1.2097760520409793, 'learning_rate/qwen_vl_interface': 1.3347882247675662e-06, 'learning_rate/action_model': 1.3347882247675661e-05, 'epoch': 1.67}) +[2026-06-17 16:52:04,789][starVLA.training.train_starvla][INFO] - Step 5679, Loss: {'train/loss': 0.6083871722221375, 'train/grad_norm_pre_clip': 16.34624481201172, 'timing/data': 0.0050519490614533424, 'timing/model': 1.2047703670104966, 'learning_rate/qwen_vl_interface': 1.3335640749642923e-06, 'learning_rate/action_model': 1.3335640749642923e-05, 'epoch': 1.67}) +[2026-06-17 16:52:09,582][starVLA.training.train_starvla][INFO] - Step 5680, Loss: {'train/loss': 0.5193785429000854, 'train/grad_norm_pre_clip': 14.675743103027344, 'timing/data': 0.004973145900294185, 'timing/model': 1.3365763879846781, 'learning_rate/qwen_vl_interface': 1.3323407370423993e-06, 'learning_rate/action_model': 1.3323407370423994e-05, 'epoch': 1.67}) +[2026-06-17 16:52:14,319][starVLA.training.train_starvla][INFO] - Step 5681, Loss: {'train/loss': 0.6507565975189209, 'train/grad_norm_pre_clip': 15.831818580627441, 'timing/data': 0.005237825098447502, 'timing/model': 1.281478155986406, 'learning_rate/qwen_vl_interface': 1.3311182112554877e-06, 'learning_rate/action_model': 1.3311182112554876e-05, 'epoch': 1.67}) +[2026-06-17 16:52:18,990][starVLA.training.train_starvla][INFO] - Step 5682, Loss: {'train/loss': 0.8906614780426025, 'train/grad_norm_pre_clip': 15.254616737365723, 'timing/data': 0.005378043046221137, 'timing/model': 1.2111338620306924, 'learning_rate/qwen_vl_interface': 1.3298964978569873e-06, 'learning_rate/action_model': 1.3298964978569872e-05, 'epoch': 1.67}) +[2026-06-17 16:52:23,781][starVLA.training.train_starvla][INFO] - Step 5683, Loss: {'train/loss': 0.5412761569023132, 'train/grad_norm_pre_clip': 13.738748550415039, 'timing/data': 0.0053483989322558045, 'timing/model': 1.2040949660586193, 'learning_rate/qwen_vl_interface': 1.3286755971001618e-06, 'learning_rate/action_model': 1.3286755971001618e-05, 'epoch': 1.67}) +[2026-06-17 16:52:28,571][starVLA.training.train_starvla][INFO] - Step 5684, Loss: {'train/loss': 0.560181200504303, 'train/grad_norm_pre_clip': 12.331817626953125, 'timing/data': 0.005222880048677325, 'timing/model': 1.2081883640494198, 'learning_rate/qwen_vl_interface': 1.327455509238104e-06, 'learning_rate/action_model': 1.3274555092381039e-05, 'epoch': 1.67}) +[2026-06-17 16:52:33,364][starVLA.training.train_starvla][INFO] - Step 5685, Loss: {'train/loss': 0.4707251489162445, 'train/grad_norm_pre_clip': 12.464866638183594, 'timing/data': 0.005289488006383181, 'timing/model': 1.2043029649648815, 'learning_rate/qwen_vl_interface': 1.32623623452374e-06, 'learning_rate/action_model': 1.32623623452374e-05, 'epoch': 1.67}) +[2026-06-17 16:52:38,162][starVLA.training.train_starvla][INFO] - Step 5686, Loss: {'train/loss': 0.7665349245071411, 'train/grad_norm_pre_clip': 12.589066505432129, 'timing/data': 0.005292088026180863, 'timing/model': 1.2094622839940712, 'learning_rate/qwen_vl_interface': 1.325017773209827e-06, 'learning_rate/action_model': 1.325017773209827e-05, 'epoch': 1.67}) +[2026-06-17 16:52:42,942][starVLA.training.train_starvla][INFO] - Step 5687, Loss: {'train/loss': 0.8220946192741394, 'train/grad_norm_pre_clip': 15.087178230285645, 'timing/data': 0.0050524279940873384, 'timing/model': 1.3018471180694178, 'learning_rate/qwen_vl_interface': 1.323800125548953e-06, 'learning_rate/action_model': 1.323800125548953e-05, 'epoch': 1.67}) +[2026-06-17 16:52:47,603][starVLA.training.train_starvla][INFO] - Step 5688, Loss: {'train/loss': 0.6748262047767639, 'train/grad_norm_pre_clip': 15.037169456481934, 'timing/data': 0.004785403027199209, 'timing/model': 1.205886343959719, 'learning_rate/qwen_vl_interface': 1.3225832917935376e-06, 'learning_rate/action_model': 1.3225832917935377e-05, 'epoch': 1.67}) +[2026-06-17 16:52:52,401][starVLA.training.train_starvla][INFO] - Step 5689, Loss: {'train/loss': 0.4873228073120117, 'train/grad_norm_pre_clip': 11.797090530395508, 'timing/data': 0.004906833986751735, 'timing/model': 1.2053572100121528, 'learning_rate/qwen_vl_interface': 1.3213672721958309e-06, 'learning_rate/action_model': 1.321367272195831e-05, 'epoch': 1.67}) +[2026-06-17 16:52:57,200][starVLA.training.train_starvla][INFO] - Step 5690, Loss: {'train/loss': 0.5776388645172119, 'train/grad_norm_pre_clip': 14.467445373535156, 'timing/data': 0.005169217009097338, 'timing/model': 1.2087849719682708, 'learning_rate/qwen_vl_interface': 1.3201520670079172e-06, 'learning_rate/action_model': 1.3201520670079171e-05, 'epoch': 1.67}) +[2026-06-17 16:53:02,010][starVLA.training.train_starvla][INFO] - Step 5691, Loss: {'train/loss': 0.629397451877594, 'train/grad_norm_pre_clip': 11.614439010620117, 'timing/data': 0.005135745974257588, 'timing/model': 1.2069272579392418, 'learning_rate/qwen_vl_interface': 1.31893767648171e-06, 'learning_rate/action_model': 1.3189376764817099e-05, 'epoch': 1.67}) +[2026-06-17 16:53:06,803][starVLA.training.train_starvla][INFO] - Step 5692, Loss: {'train/loss': 0.5742323398590088, 'train/grad_norm_pre_clip': 30.216588973999023, 'timing/data': 0.0050209269393235445, 'timing/model': 1.3392473759595305, 'learning_rate/qwen_vl_interface': 1.3177241008689513e-06, 'learning_rate/action_model': 1.3177241008689511e-05, 'epoch': 1.67}) +[2026-06-17 16:53:11,463][starVLA.training.train_starvla][INFO] - Step 5693, Loss: {'train/loss': 0.6055501103401184, 'train/grad_norm_pre_clip': 13.949115753173828, 'timing/data': 0.005209810915403068, 'timing/model': 1.2088237470015883, 'learning_rate/qwen_vl_interface': 1.316511340421216e-06, 'learning_rate/action_model': 1.316511340421216e-05, 'epoch': 1.68}) +[2026-06-17 16:53:16,262][starVLA.training.train_starvla][INFO] - Step 5694, Loss: {'train/loss': 0.8166799545288086, 'train/grad_norm_pre_clip': 14.120611190795898, 'timing/data': 0.005057928035967052, 'timing/model': 1.213757369085215, 'learning_rate/qwen_vl_interface': 1.315299395389914e-06, 'learning_rate/action_model': 1.315299395389914e-05, 'epoch': 1.68}) +[2026-06-17 16:53:21,047][starVLA.training.train_starvla][INFO] - Step 5695, Loss: {'train/loss': 0.48884716629981995, 'train/grad_norm_pre_clip': 10.999250411987305, 'timing/data': 0.005126151023432612, 'timing/model': 1.2073862459510565, 'learning_rate/qwen_vl_interface': 1.3140882660262818e-06, 'learning_rate/action_model': 1.3140882660262818e-05, 'epoch': 1.68}) +[2026-06-17 16:53:25,796][starVLA.training.train_starvla][INFO] - Step 5696, Loss: {'train/loss': 0.8109064102172852, 'train/grad_norm_pre_clip': 17.873727798461914, 'timing/data': 0.005315654096193612, 'timing/model': 1.210002105915919, 'learning_rate/qwen_vl_interface': 1.3128779525813865e-06, 'learning_rate/action_model': 1.3128779525813864e-05, 'epoch': 1.68}) +[2026-06-17 16:53:30,539][starVLA.training.train_starvla][INFO] - Step 5697, Loss: {'train/loss': 0.6012531518936157, 'train/grad_norm_pre_clip': 13.537700653076172, 'timing/data': 0.005238894023932517, 'timing/model': 1.2089531989768147, 'learning_rate/qwen_vl_interface': 1.3116684553061288e-06, 'learning_rate/action_model': 1.3116684553061287e-05, 'epoch': 1.68}) +[2026-06-17 16:53:35,333][starVLA.training.train_starvla][INFO] - Step 5698, Loss: {'train/loss': 0.6688799262046814, 'train/grad_norm_pre_clip': 13.779708862304688, 'timing/data': 0.005250253016129136, 'timing/model': 1.3370052989339456, 'learning_rate/qwen_vl_interface': 1.310459774451238e-06, 'learning_rate/action_model': 1.3104597744512379e-05, 'epoch': 1.68}) +[2026-06-17 16:53:39,988][starVLA.training.train_starvla][INFO] - Step 5699, Loss: {'train/loss': 0.6898893713951111, 'train/grad_norm_pre_clip': 15.685161590576172, 'timing/data': 0.005278873955830932, 'timing/model': 1.2038183899130672, 'learning_rate/qwen_vl_interface': 1.3092519102672744e-06, 'learning_rate/action_model': 1.3092519102672742e-05, 'epoch': 1.68}) +[2026-06-17 16:55:22,005][starVLA.training.train_starvla][INFO] - Step 5700, Loss: {'train/loss': 0.6944265365600586, 'train/grad_norm_pre_clip': 13.853641510009766, 'eval/loss': 0.7041112780570984, 'eval/latency_0/loss': 0.7041112780570984, 'eval/demon_fire_precision': 0.9022526146419951, 'eval/demon_fire_recall': 0.9314784053156147, 'eval/demon_fire_f1': 0.9166326113608501, 'eval/demon_fire_rate_model': 0.6913236929922135, 'eval/demon_fire_rate_teacher': 0.6696329254727474, 'eval/demon_left_precision': 0.8355106888361045, 'eval/demon_left_recall': 0.7971671388101983, 'eval/demon_left_f1': 0.81588866338069, 'eval/demon_left_rate_model': 0.4682981090100111, 'eval/demon_left_rate_teacher': 0.49082313681868744, 'eval/demon_right_precision': 0.810079575596817, 'eval/demon_right_recall': 0.8627118644067797, 'eval/demon_right_f1': 0.8355677154582762, 'eval/demon_right_rate_model': 0.5241935483870968, 'eval/demon_right_rate_teacher': 0.4922135706340378, 'eval/demon_move_f1': 0.8257281894194831, 'eval/demon_cc_f1': 0.8873268651891798, 'eval/latency_0/demon_fire_precision': 0.9022526146419951, 'eval/latency_0/demon_fire_recall': 0.9314784053156147, 'eval/latency_0/demon_fire_f1': 0.9166326113608501, 'eval/latency_0/demon_fire_rate_model': 0.6913236929922135, 'eval/latency_0/demon_fire_rate_teacher': 0.6696329254727474, 'eval/latency_0/demon_left_precision': 0.8355106888361045, 'eval/latency_0/demon_left_recall': 0.7971671388101983, 'eval/latency_0/demon_left_f1': 0.81588866338069, 'eval/latency_0/demon_left_rate_model': 0.4682981090100111, 'eval/latency_0/demon_left_rate_teacher': 0.49082313681868744, 'eval/latency_0/demon_right_precision': 0.810079575596817, 'eval/latency_0/demon_right_recall': 0.8627118644067797, 'eval/latency_0/demon_right_f1': 0.8355677154582762, 'eval/latency_0/demon_right_rate_model': 0.5241935483870968, 'eval/latency_0/demon_right_rate_teacher': 0.4922135706340378, 'eval/latency_0/demon_move_f1': 0.8257281894194831, 'eval/latency_0/demon_cc_f1': 0.8873268651891798, 'timing/data': 0.005296493065543473, 'timing/model': 1.2062712810002267, 'learning_rate/qwen_vl_interface': 1.3080448630046312e-06, 'learning_rate/action_model': 1.308044863004631e-05, 'epoch': 1.68}) +[2026-06-17 16:55:22,006][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_5700_state +[2026-06-17 16:55:22,016][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 16:55:25,564][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_5700_state/model.safetensors +[2026-06-17 16:55:33,189][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_5700_state/optimizer.bin +[2026-06-17 16:55:33,189][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_5700_state/sampler.bin +[2026-06-17 16:55:33,189][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_5700_state/sampler_1.bin +[2026-06-17 16:55:33,190][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_5700_state/random_states_0.pkl +[2026-06-17 16:55:33,239][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 5700 +[2026-06-17 16:55:33,239][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 16:55:33,249][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 16:55:37,826][starVLA.training.train_starvla][INFO] - Step 5701, Loss: {'train/loss': 0.5331060886383057, 'train/grad_norm_pre_clip': 16.012466430664062, 'timing/data': 0.0047354960115626454, 'timing/model': 1.1881583930226043, 'learning_rate/qwen_vl_interface': 1.3068386329135287e-06, 'learning_rate/action_model': 1.3068386329135287e-05, 'epoch': 1.68}) +[2026-06-17 16:55:42,569][starVLA.training.train_starvla][INFO] - Step 5702, Loss: {'train/loss': 0.8527510762214661, 'train/grad_norm_pre_clip': 16.790111541748047, 'timing/data': 0.0047536930069327354, 'timing/model': 1.1902805879944935, 'learning_rate/qwen_vl_interface': 1.3056332202440211e-06, 'learning_rate/action_model': 1.305633220244021e-05, 'epoch': 1.68}) +[2026-06-17 16:55:47,299][starVLA.training.train_starvla][INFO] - Step 5703, Loss: {'train/loss': 0.7258477807044983, 'train/grad_norm_pre_clip': 14.493450164794922, 'timing/data': 0.0051025759894400835, 'timing/model': 1.1925551369786263, 'learning_rate/qwen_vl_interface': 1.3044286252459914e-06, 'learning_rate/action_model': 1.3044286252459915e-05, 'epoch': 1.68}) +[2026-06-17 16:55:52,044][starVLA.training.train_starvla][INFO] - Step 5704, Loss: {'train/loss': 0.8315220475196838, 'train/grad_norm_pre_clip': 16.191347122192383, 'timing/data': 0.004692240036092699, 'timing/model': 1.1940175590571016, 'learning_rate/qwen_vl_interface': 1.3032248481691523e-06, 'learning_rate/action_model': 1.3032248481691523e-05, 'epoch': 1.68}) +[2026-06-17 16:55:56,792][starVLA.training.train_starvla][INFO] - Step 5705, Loss: {'train/loss': 0.4889325201511383, 'train/grad_norm_pre_clip': 11.5552339553833, 'timing/data': 0.005131814978085458, 'timing/model': 1.3244475830579177, 'learning_rate/qwen_vl_interface': 1.302021889263051e-06, 'learning_rate/action_model': 1.3020218892630508e-05, 'epoch': 1.68}) +[2026-06-17 16:56:01,419][starVLA.training.train_starvla][INFO] - Step 5706, Loss: {'train/loss': 0.6370280981063843, 'train/grad_norm_pre_clip': 14.875251770019531, 'timing/data': 0.004874965059570968, 'timing/model': 1.198053425992839, 'learning_rate/qwen_vl_interface': 1.3008197487770593e-06, 'learning_rate/action_model': 1.3008197487770593e-05, 'epoch': 1.68}) +[2026-06-17 16:56:06,180][starVLA.training.train_starvla][INFO] - Step 5707, Loss: {'train/loss': 0.4585382044315338, 'train/grad_norm_pre_clip': 12.730524063110352, 'timing/data': 0.00476915598846972, 'timing/model': 1.2020793640986085, 'learning_rate/qwen_vl_interface': 1.2996184269603835e-06, 'learning_rate/action_model': 1.2996184269603833e-05, 'epoch': 1.68}) +[2026-06-17 16:56:10,854][starVLA.training.train_starvla][INFO] - Step 5708, Loss: {'train/loss': 0.6749792098999023, 'train/grad_norm_pre_clip': 15.783661842346191, 'timing/data': 0.005198318976908922, 'timing/model': 1.2015829340089113, 'learning_rate/qwen_vl_interface': 1.2984179240620567e-06, 'learning_rate/action_model': 1.2984179240620567e-05, 'epoch': 1.68}) +[2026-06-17 16:56:15,567][starVLA.training.train_starvla][INFO] - Step 5709, Loss: {'train/loss': 0.9034990072250366, 'train/grad_norm_pre_clip': 20.798429489135742, 'timing/data': 0.005079891998320818, 'timing/model': 1.1992602889658883, 'learning_rate/qwen_vl_interface': 1.2972182403309474e-06, 'learning_rate/action_model': 1.2972182403309474e-05, 'epoch': 1.68}) +[2026-06-17 16:56:20,318][starVLA.training.train_starvla][INFO] - Step 5710, Loss: {'train/loss': 0.6948901414871216, 'train/grad_norm_pre_clip': 14.917038917541504, 'timing/data': 0.004807336954399943, 'timing/model': 1.1984908040612936, 'learning_rate/qwen_vl_interface': 1.2960193760157507e-06, 'learning_rate/action_model': 1.2960193760157507e-05, 'epoch': 1.68}) +[2026-06-17 16:56:25,108][starVLA.training.train_starvla][INFO] - Step 5711, Loss: {'train/loss': 0.7198810577392578, 'train/grad_norm_pre_clip': 14.090807914733887, 'timing/data': 0.004916053032502532, 'timing/model': 1.3514344899449497, 'learning_rate/qwen_vl_interface': 1.294821331364992e-06, 'learning_rate/action_model': 1.294821331364992e-05, 'epoch': 1.68}) +[2026-06-17 16:56:29,748][starVLA.training.train_starvla][INFO] - Step 5712, Loss: {'train/loss': 0.5835079550743103, 'train/grad_norm_pre_clip': 16.055749893188477, 'timing/data': 0.004782194038853049, 'timing/model': 1.202102212002501, 'learning_rate/qwen_vl_interface': 1.2936241066270278e-06, 'learning_rate/action_model': 1.2936241066270278e-05, 'epoch': 1.68}) +[2026-06-17 16:56:34,518][starVLA.training.train_starvla][INFO] - Step 5713, Loss: {'train/loss': 0.7898353338241577, 'train/grad_norm_pre_clip': 13.051579475402832, 'timing/data': 0.005010925000533462, 'timing/model': 1.2007228570291772, 'learning_rate/qwen_vl_interface': 1.2924277020500427e-06, 'learning_rate/action_model': 1.2924277020500428e-05, 'epoch': 1.68}) +[2026-06-17 16:56:39,296][starVLA.training.train_starvla][INFO] - Step 5714, Loss: {'train/loss': 0.6459357738494873, 'train/grad_norm_pre_clip': 13.358216285705566, 'timing/data': 0.004982295911759138, 'timing/model': 1.2054454830940813, 'learning_rate/qwen_vl_interface': 1.291232117882054e-06, 'learning_rate/action_model': 1.2912321178820538e-05, 'epoch': 1.68}) +[2026-06-17 16:56:44,018][starVLA.training.train_starvla][INFO] - Step 5715, Loss: {'train/loss': 0.43626198172569275, 'train/grad_norm_pre_clip': 8.627366065979004, 'timing/data': 0.0052296079229563475, 'timing/model': 1.2023820659378543, 'learning_rate/qwen_vl_interface': 1.2900373543709065e-06, 'learning_rate/action_model': 1.2900373543709065e-05, 'epoch': 1.68}) +[2026-06-17 16:56:48,796][starVLA.training.train_starvla][INFO] - Step 5716, Loss: {'train/loss': 0.6448436379432678, 'train/grad_norm_pre_clip': 11.571639060974121, 'timing/data': 0.004951197071932256, 'timing/model': 1.202987473923713, 'learning_rate/qwen_vl_interface': 1.288843411764277e-06, 'learning_rate/action_model': 1.288843411764277e-05, 'epoch': 1.68}) +[2026-06-17 16:56:53,581][starVLA.training.train_starvla][INFO] - Step 5717, Loss: {'train/loss': 0.3291614353656769, 'train/grad_norm_pre_clip': 10.336759567260742, 'timing/data': 0.005516599980182946, 'timing/model': 1.3386835599085316, 'learning_rate/qwen_vl_interface': 1.2876502903096709e-06, 'learning_rate/action_model': 1.2876502903096708e-05, 'epoch': 1.68}) +[2026-06-17 16:56:58,235][starVLA.training.train_starvla][INFO] - Step 5718, Loss: {'train/loss': 0.6888182163238525, 'train/grad_norm_pre_clip': 15.626235961914062, 'timing/data': 0.004366911016404629, 'timing/model': 1.204938265029341, 'learning_rate/qwen_vl_interface': 1.286457990254423e-06, 'learning_rate/action_model': 1.286457990254423e-05, 'epoch': 1.68}) +[2026-06-17 16:57:03,090][starVLA.training.train_starvla][INFO] - Step 5719, Loss: {'train/loss': 0.7715849280357361, 'train/grad_norm_pre_clip': 16.278728485107422, 'timing/data': 0.004811133956536651, 'timing/model': 1.2060507100541145, 'learning_rate/qwen_vl_interface': 1.2852665118456971e-06, 'learning_rate/action_model': 1.2852665118456972e-05, 'epoch': 1.68}) +[2026-06-17 16:57:07,867][starVLA.training.train_starvla][INFO] - Step 5720, Loss: {'train/loss': 0.39585554599761963, 'train/grad_norm_pre_clip': 9.967339515686035, 'timing/data': 0.0049415279645472765, 'timing/model': 1.2042418660130352, 'learning_rate/qwen_vl_interface': 1.2840758553304927e-06, 'learning_rate/action_model': 1.2840758553304927e-05, 'epoch': 1.68}) +[2026-06-17 16:57:12,604][starVLA.training.train_starvla][INFO] - Step 5721, Loss: {'train/loss': 0.6780886054039001, 'train/grad_norm_pre_clip': 15.649588584899902, 'timing/data': 0.005282560014165938, 'timing/model': 1.2029673949582502, 'learning_rate/qwen_vl_interface': 1.2828860209556293e-06, 'learning_rate/action_model': 1.2828860209556292e-05, 'epoch': 1.68}) +[2026-06-17 16:57:17,397][starVLA.training.train_starvla][INFO] - Step 5722, Loss: {'train/loss': 0.723666787147522, 'train/grad_norm_pre_clip': 16.369155883789062, 'timing/data': 0.005292520974762738, 'timing/model': 1.2046254359884188, 'learning_rate/qwen_vl_interface': 1.2816970089677617e-06, 'learning_rate/action_model': 1.2816970089677615e-05, 'epoch': 1.68}) +[2026-06-17 16:57:22,184][starVLA.training.train_starvla][INFO] - Step 5723, Loss: {'train/loss': 0.8836162090301514, 'train/grad_norm_pre_clip': 18.612966537475586, 'timing/data': 0.00514935702085495, 'timing/model': 1.3387901700334623, 'learning_rate/qwen_vl_interface': 1.280508819613375e-06, 'learning_rate/action_model': 1.280508819613375e-05, 'epoch': 1.68}) +[2026-06-17 16:57:26,847][starVLA.training.train_starvla][INFO] - Step 5724, Loss: {'train/loss': 0.6723167896270752, 'train/grad_norm_pre_clip': 12.499406814575195, 'timing/data': 0.005052744061686099, 'timing/model': 1.2055164030753076, 'learning_rate/qwen_vl_interface': 1.2793214531387812e-06, 'learning_rate/action_model': 1.2793214531387812e-05, 'epoch': 1.68}) +[2026-06-17 16:57:31,629][starVLA.training.train_starvla][INFO] - Step 5725, Loss: {'train/loss': 0.7573632597923279, 'train/grad_norm_pre_clip': 16.2315731048584, 'timing/data': 0.005138408043421805, 'timing/model': 1.2034955070121214, 'learning_rate/qwen_vl_interface': 1.2781349097901246e-06, 'learning_rate/action_model': 1.2781349097901247e-05, 'epoch': 1.68}) +[2026-06-17 16:57:36,410][starVLA.training.train_starvla][INFO] - Step 5726, Loss: {'train/loss': 0.8192063570022583, 'train/grad_norm_pre_clip': 14.77313232421875, 'timing/data': 0.004922754014842212, 'timing/model': 1.203355217934586, 'learning_rate/qwen_vl_interface': 1.2769491898133723e-06, 'learning_rate/action_model': 1.2769491898133722e-05, 'epoch': 1.68}) +[2026-06-17 16:57:41,198][starVLA.training.train_starvla][INFO] - Step 5727, Loss: {'train/loss': 0.8850529789924622, 'train/grad_norm_pre_clip': 18.021066665649414, 'timing/data': 0.005241497070528567, 'timing/model': 1.2039989080512896, 'learning_rate/qwen_vl_interface': 1.2757642934543293e-06, 'learning_rate/action_model': 1.2757642934543293e-05, 'epoch': 1.69}) +[2026-06-17 16:57:45,873][starVLA.training.train_starvla][INFO] - Step 5728, Loss: {'train/loss': 0.6187347173690796, 'train/grad_norm_pre_clip': 14.507450103759766, 'timing/data': 0.004958522040396929, 'timing/model': 1.204229495022446, 'learning_rate/qwen_vl_interface': 1.274580220958625e-06, 'learning_rate/action_model': 1.274580220958625e-05, 'epoch': 1.69}) +[2026-06-17 16:57:50,565][starVLA.training.train_starvla][INFO] - Step 5729, Loss: {'train/loss': 0.6919398903846741, 'train/grad_norm_pre_clip': 16.555065155029297, 'timing/data': 0.00506029196549207, 'timing/model': 1.2416658110450953, 'learning_rate/qwen_vl_interface': 1.2733969725717184e-06, 'learning_rate/action_model': 1.2733969725717184e-05, 'epoch': 1.69}) +[2026-06-17 16:57:55,225][starVLA.training.train_starvla][INFO] - Step 5730, Loss: {'train/loss': 0.597118079662323, 'train/grad_norm_pre_clip': 12.105050086975098, 'timing/data': 0.0050045340321958065, 'timing/model': 1.209195084986277, 'learning_rate/qwen_vl_interface': 1.272214548538899e-06, 'learning_rate/action_model': 1.2722145485388989e-05, 'epoch': 1.69}) +[2026-06-17 16:57:59,969][starVLA.training.train_starvla][INFO] - Step 5731, Loss: {'train/loss': 0.8131862878799438, 'train/grad_norm_pre_clip': 16.14263343811035, 'timing/data': 0.005062977084890008, 'timing/model': 1.2108603509841487, 'learning_rate/qwen_vl_interface': 1.2710329491052834e-06, 'learning_rate/action_model': 1.2710329491052833e-05, 'epoch': 1.69}) +[2026-06-17 16:58:04,766][starVLA.training.train_starvla][INFO] - Step 5732, Loss: {'train/loss': 0.5527613759040833, 'train/grad_norm_pre_clip': 12.446017265319824, 'timing/data': 0.005248658009804785, 'timing/model': 1.205522559932433, 'learning_rate/qwen_vl_interface': 1.2698521745158188e-06, 'learning_rate/action_model': 1.2698521745158187e-05, 'epoch': 1.69}) +[2026-06-17 16:58:09,546][starVLA.training.train_starvla][INFO] - Step 5733, Loss: {'train/loss': 0.5930987000465393, 'train/grad_norm_pre_clip': 14.117060661315918, 'timing/data': 0.004787312005646527, 'timing/model': 1.2041457499144599, 'learning_rate/qwen_vl_interface': 1.2686722250152804e-06, 'learning_rate/action_model': 1.2686722250152806e-05, 'epoch': 1.69}) +[2026-06-17 16:58:14,332][starVLA.training.train_starvla][INFO] - Step 5734, Loss: {'train/loss': 0.35333389043807983, 'train/grad_norm_pre_clip': 9.122824668884277, 'timing/data': 0.004909459967166185, 'timing/model': 1.332737116026692, 'learning_rate/qwen_vl_interface': 1.2674931008482763e-06, 'learning_rate/action_model': 1.2674931008482763e-05, 'epoch': 1.69}) +[2026-06-17 16:58:19,042][starVLA.training.train_starvla][INFO] - Step 5735, Loss: {'train/loss': 0.7895711660385132, 'train/grad_norm_pre_clip': 17.03130531311035, 'timing/data': 0.0049604339292272925, 'timing/model': 1.2577939690090716, 'learning_rate/qwen_vl_interface': 1.2663148022592365e-06, 'learning_rate/action_model': 1.2663148022592363e-05, 'epoch': 1.69}) +[2026-06-17 16:58:23,707][starVLA.training.train_starvla][INFO] - Step 5736, Loss: {'train/loss': 0.663358211517334, 'train/grad_norm_pre_clip': 13.179766654968262, 'timing/data': 0.004969422006979585, 'timing/model': 1.2072963940445334, 'learning_rate/qwen_vl_interface': 1.2651373294924249e-06, 'learning_rate/action_model': 1.2651373294924248e-05, 'epoch': 1.69}) +[2026-06-17 16:58:28,471][starVLA.training.train_starvla][INFO] - Step 5737, Loss: {'train/loss': 0.6383746862411499, 'train/grad_norm_pre_clip': 13.504638671875, 'timing/data': 0.004961885046213865, 'timing/model': 1.20407974999398, 'learning_rate/qwen_vl_interface': 1.2639606827919313e-06, 'learning_rate/action_model': 1.2639606827919311e-05, 'epoch': 1.69}) +[2026-06-17 16:58:33,262][starVLA.training.train_starvla][INFO] - Step 5738, Loss: {'train/loss': 0.7021317481994629, 'train/grad_norm_pre_clip': 14.222711563110352, 'timing/data': 0.005110740079544485, 'timing/model': 1.2084280339768156, 'learning_rate/qwen_vl_interface': 1.2627848624016788e-06, 'learning_rate/action_model': 1.2627848624016788e-05, 'epoch': 1.69}) +[2026-06-17 16:58:38,057][starVLA.training.train_starvla][INFO] - Step 5739, Loss: {'train/loss': 0.8319864869117737, 'train/grad_norm_pre_clip': 17.341163635253906, 'timing/data': 0.005215562996454537, 'timing/model': 1.2056459980085492, 'learning_rate/qwen_vl_interface': 1.261609868565416e-06, 'learning_rate/action_model': 1.2616098685654159e-05, 'epoch': 1.69}) +[2026-06-17 16:58:42,863][starVLA.training.train_starvla][INFO] - Step 5740, Loss: {'train/loss': 0.8387314677238464, 'train/grad_norm_pre_clip': 17.006298065185547, 'timing/data': 0.004982725018635392, 'timing/model': 1.3435310969362035, 'learning_rate/qwen_vl_interface': 1.2604357015267165e-06, 'learning_rate/action_model': 1.2604357015267166e-05, 'epoch': 1.69}) +[2026-06-17 16:58:47,649][starVLA.training.train_starvla][INFO] - Step 5741, Loss: {'train/loss': 0.4297816753387451, 'train/grad_norm_pre_clip': 14.150035858154297, 'timing/data': 0.005215056939050555, 'timing/model': 1.3288150520529598, 'learning_rate/qwen_vl_interface': 1.2592623615289906e-06, 'learning_rate/action_model': 1.2592623615289906e-05, 'epoch': 1.69}) +[2026-06-17 16:58:52,313][starVLA.training.train_starvla][INFO] - Step 5742, Loss: {'train/loss': 0.474167138338089, 'train/grad_norm_pre_clip': 11.461523056030273, 'timing/data': 0.0049288959708064795, 'timing/model': 1.208176210988313, 'learning_rate/qwen_vl_interface': 1.258089848815471e-06, 'learning_rate/action_model': 1.2580898488154708e-05, 'epoch': 1.69}) +[2026-06-17 16:58:57,071][starVLA.training.train_starvla][INFO] - Step 5743, Loss: {'train/loss': 0.5770649909973145, 'train/grad_norm_pre_clip': 12.26048755645752, 'timing/data': 0.005041385069489479, 'timing/model': 1.204710702993907, 'learning_rate/qwen_vl_interface': 1.256918163629221e-06, 'learning_rate/action_model': 1.256918163629221e-05, 'epoch': 1.69}) +[2026-06-17 16:59:01,858][starVLA.training.train_starvla][INFO] - Step 5744, Loss: {'train/loss': 0.5927555561065674, 'train/grad_norm_pre_clip': 14.712187767028809, 'timing/data': 0.004956670920364559, 'timing/model': 1.2029259579721838, 'learning_rate/qwen_vl_interface': 1.2557473062131332e-06, 'learning_rate/action_model': 1.2557473062131331e-05, 'epoch': 1.69}) +[2026-06-17 16:59:06,667][starVLA.training.train_starvla][INFO] - Step 5745, Loss: {'train/loss': 0.46529677510261536, 'train/grad_norm_pre_clip': 12.414223670959473, 'timing/data': 0.004980162950232625, 'timing/model': 1.207298414898105, 'learning_rate/qwen_vl_interface': 1.2545772768099263e-06, 'learning_rate/action_model': 1.2545772768099263e-05, 'epoch': 1.69}) +[2026-06-17 16:59:11,399][starVLA.training.train_starvla][INFO] - Step 5746, Loss: {'train/loss': 0.7912794351577759, 'train/grad_norm_pre_clip': 16.69474983215332, 'timing/data': 0.005412781029008329, 'timing/model': 1.2852532109245658, 'learning_rate/qwen_vl_interface': 1.2534080756621488e-06, 'learning_rate/action_model': 1.2534080756621488e-05, 'epoch': 1.69}) +[2026-06-17 16:59:16,067][starVLA.training.train_starvla][INFO] - Step 5747, Loss: {'train/loss': 0.9985601902008057, 'train/grad_norm_pre_clip': 19.890140533447266, 'timing/data': 0.005510500050149858, 'timing/model': 1.2056702730478719, 'learning_rate/qwen_vl_interface': 1.2522397030121782e-06, 'learning_rate/action_model': 1.2522397030121783e-05, 'epoch': 1.69}) +[2026-06-17 16:59:20,848][starVLA.training.train_starvla][INFO] - Step 5748, Loss: {'train/loss': 0.7645843029022217, 'train/grad_norm_pre_clip': 14.528751373291016, 'timing/data': 0.005060605006292462, 'timing/model': 1.209084980073385, 'learning_rate/qwen_vl_interface': 1.2510721591022194e-06, 'learning_rate/action_model': 1.2510721591022192e-05, 'epoch': 1.69}) +[2026-06-17 16:59:25,630][starVLA.training.train_starvla][INFO] - Step 5749, Loss: {'train/loss': 0.47789081931114197, 'train/grad_norm_pre_clip': 11.239033699035645, 'timing/data': 0.005027072969824076, 'timing/model': 1.2047025630017743, 'learning_rate/qwen_vl_interface': 1.2499054441743047e-06, 'learning_rate/action_model': 1.2499054441743047e-05, 'epoch': 1.69}) +[2026-06-17 16:59:30,460][starVLA.training.train_starvla][INFO] - Step 5750, Loss: {'train/loss': 0.6314188838005066, 'train/grad_norm_pre_clip': 12.830766677856445, 'timing/data': 0.004711253917776048, 'timing/model': 1.2069071150617674, 'learning_rate/qwen_vl_interface': 1.2487395584702955e-06, 'learning_rate/action_model': 1.2487395584702954e-05, 'epoch': 1.69}) +[2026-06-17 16:59:35,249][starVLA.training.train_starvla][INFO] - Step 5751, Loss: {'train/loss': 0.5115883350372314, 'train/grad_norm_pre_clip': 13.541547775268555, 'timing/data': 0.004710027948021889, 'timing/model': 1.330470580025576, 'learning_rate/qwen_vl_interface': 1.2475745022318803e-06, 'learning_rate/action_model': 1.2475745022318803e-05, 'epoch': 1.69}) +[2026-06-17 16:59:39,988][starVLA.training.train_starvla][INFO] - Step 5752, Loss: {'train/loss': 0.47435295581817627, 'train/grad_norm_pre_clip': 10.297069549560547, 'timing/data': 0.005057067959569395, 'timing/model': 1.2816218769876286, 'learning_rate/qwen_vl_interface': 1.2464102757005788e-06, 'learning_rate/action_model': 1.2464102757005788e-05, 'epoch': 1.69}) +[2026-06-17 16:59:44,711][starVLA.training.train_starvla][INFO] - Step 5753, Loss: {'train/loss': 0.6653918027877808, 'train/grad_norm_pre_clip': 16.45008659362793, 'timing/data': 0.004914384917356074, 'timing/model': 1.204588424996473, 'learning_rate/qwen_vl_interface': 1.2452468791177357e-06, 'learning_rate/action_model': 1.2452468791177357e-05, 'epoch': 1.69}) +[2026-06-17 16:59:49,501][starVLA.training.train_starvla][INFO] - Step 5754, Loss: {'train/loss': 0.5970967411994934, 'train/grad_norm_pre_clip': 13.030112266540527, 'timing/data': 0.004874450038187206, 'timing/model': 1.2051911159651354, 'learning_rate/qwen_vl_interface': 1.2440843127245235e-06, 'learning_rate/action_model': 1.2440843127245235e-05, 'epoch': 1.69}) +[2026-06-17 16:59:54,314][starVLA.training.train_starvla][INFO] - Step 5755, Loss: {'train/loss': 0.7615315914154053, 'train/grad_norm_pre_clip': 15.198412895202637, 'timing/data': 0.004566399031318724, 'timing/model': 1.2071599320042878, 'learning_rate/qwen_vl_interface': 1.2429225767619427e-06, 'learning_rate/action_model': 1.2429225767619427e-05, 'epoch': 1.69}) +[2026-06-17 16:59:59,107][starVLA.training.train_starvla][INFO] - Step 5756, Loss: {'train/loss': 0.7108438014984131, 'train/grad_norm_pre_clip': 25.36053466796875, 'timing/data': 0.004994634888134897, 'timing/model': 1.2093160930089653, 'learning_rate/qwen_vl_interface': 1.2417616714708246e-06, 'learning_rate/action_model': 1.2417616714708247e-05, 'epoch': 1.69}) +[2026-06-17 17:00:03,894][starVLA.training.train_starvla][INFO] - Step 5757, Loss: {'train/loss': 0.43760576844215393, 'train/grad_norm_pre_clip': 12.562702178955078, 'timing/data': 0.005231268936768174, 'timing/model': 1.2051509049488232, 'learning_rate/qwen_vl_interface': 1.2406015970918247e-06, 'learning_rate/action_model': 1.2406015970918247e-05, 'epoch': 1.69}) +[2026-06-17 17:00:08,641][starVLA.training.train_starvla][INFO] - Step 5758, Loss: {'train/loss': 0.6210522651672363, 'train/grad_norm_pre_clip': 14.211000442504883, 'timing/data': 0.0046926140785217285, 'timing/model': 1.293766026967205, 'learning_rate/qwen_vl_interface': 1.2394423538654296e-06, 'learning_rate/action_model': 1.2394423538654296e-05, 'epoch': 1.69}) +[2026-06-17 17:00:13,314][starVLA.training.train_starvla][INFO] - Step 5759, Loss: {'train/loss': 0.5482606291770935, 'train/grad_norm_pre_clip': 15.029870986938477, 'timing/data': 0.0050063260132446885, 'timing/model': 1.2069629030302167, 'learning_rate/qwen_vl_interface': 1.2382839420319478e-06, 'learning_rate/action_model': 1.2382839420319478e-05, 'epoch': 1.69}) +[2026-06-17 17:00:18,135][starVLA.training.train_starvla][INFO] - Step 5760, Loss: {'train/loss': 0.5200728178024292, 'train/grad_norm_pre_clip': 12.143710136413574, 'timing/data': 0.0048951259814202785, 'timing/model': 1.2097637869883329, 'learning_rate/qwen_vl_interface': 1.2371263618315225e-06, 'learning_rate/action_model': 1.2371263618315224e-05, 'epoch': 1.69}) +[2026-06-17 17:00:22,945][starVLA.training.train_starvla][INFO] - Step 5761, Loss: {'train/loss': 0.5756776928901672, 'train/grad_norm_pre_clip': 13.455378532409668, 'timing/data': 0.004838093998841941, 'timing/model': 1.2038770919898525, 'learning_rate/qwen_vl_interface': 1.2359696135041202e-06, 'learning_rate/action_model': 1.2359696135041201e-05, 'epoch': 1.7}) +[2026-06-17 17:00:27,731][starVLA.training.train_starvla][INFO] - Step 5762, Loss: {'train/loss': 0.6056872010231018, 'train/grad_norm_pre_clip': 14.473462104797363, 'timing/data': 0.0050158180529251695, 'timing/model': 1.203846330055967, 'learning_rate/qwen_vl_interface': 1.2348136972895369e-06, 'learning_rate/action_model': 1.2348136972895368e-05, 'epoch': 1.7}) +[2026-06-17 17:00:32,547][starVLA.training.train_starvla][INFO] - Step 5763, Loss: {'train/loss': 0.5386130809783936, 'train/grad_norm_pre_clip': 10.73551082611084, 'timing/data': 0.03175895207095891, 'timing/model': 1.3404806070029736, 'learning_rate/qwen_vl_interface': 1.2336586134273945e-06, 'learning_rate/action_model': 1.2336586134273945e-05, 'epoch': 1.7}) +[2026-06-17 17:00:37,220][starVLA.training.train_starvla][INFO] - Step 5764, Loss: {'train/loss': 0.8635526299476624, 'train/grad_norm_pre_clip': 17.292856216430664, 'timing/data': 0.004991997964680195, 'timing/model': 1.2128436990315095, 'learning_rate/qwen_vl_interface': 1.2325043621571428e-06, 'learning_rate/action_model': 1.2325043621571427e-05, 'epoch': 1.7}) +[2026-06-17 17:00:42,012][starVLA.training.train_starvla][INFO] - Step 5765, Loss: {'train/loss': 0.6562606692314148, 'train/grad_norm_pre_clip': 15.580060005187988, 'timing/data': 0.005049134022556245, 'timing/model': 1.212070390000008, 'learning_rate/qwen_vl_interface': 1.2313509437180603e-06, 'learning_rate/action_model': 1.2313509437180604e-05, 'epoch': 1.7}) +[2026-06-17 17:00:46,830][starVLA.training.train_starvla][INFO] - Step 5766, Loss: {'train/loss': 0.7507774233818054, 'train/grad_norm_pre_clip': 14.257662773132324, 'timing/data': 0.00492141698487103, 'timing/model': 1.2057359949685633, 'learning_rate/qwen_vl_interface': 1.2301983583492502e-06, 'learning_rate/action_model': 1.2301983583492502e-05, 'epoch': 1.7}) +[2026-06-17 17:00:51,541][starVLA.training.train_starvla][INFO] - Step 5767, Loss: {'train/loss': 0.6835393905639648, 'train/grad_norm_pre_clip': 12.831305503845215, 'timing/data': 0.005265708896331489, 'timing/model': 1.210385294049047, 'learning_rate/qwen_vl_interface': 1.229046606289648e-06, 'learning_rate/action_model': 1.229046606289648e-05, 'epoch': 1.7}) +[2026-06-17 17:00:56,298][starVLA.training.train_starvla][INFO] - Step 5768, Loss: {'train/loss': 0.534553050994873, 'train/grad_norm_pre_clip': 20.992626190185547, 'timing/data': 0.004596179001964629, 'timing/model': 1.2112371970433742, 'learning_rate/qwen_vl_interface': 1.2278956877780102e-06, 'learning_rate/action_model': 1.2278956877780101e-05, 'epoch': 1.7}) +[2026-06-17 17:01:01,111][starVLA.training.train_starvla][INFO] - Step 5769, Loss: {'train/loss': 0.7657585144042969, 'train/grad_norm_pre_clip': 13.498056411743164, 'timing/data': 0.004962351988069713, 'timing/model': 1.3374686850002035, 'learning_rate/qwen_vl_interface': 1.2267456030529245e-06, 'learning_rate/action_model': 1.2267456030529245e-05, 'epoch': 1.7}) +[2026-06-17 17:01:05,776][starVLA.training.train_starvla][INFO] - Step 5770, Loss: {'train/loss': 0.4642009139060974, 'train/grad_norm_pre_clip': 11.707704544067383, 'timing/data': 0.005086604971438646, 'timing/model': 1.2157321620034054, 'learning_rate/qwen_vl_interface': 1.225596352352803e-06, 'learning_rate/action_model': 1.225596352352803e-05, 'epoch': 1.7}) +[2026-06-17 17:01:10,561][starVLA.training.train_starvla][INFO] - Step 5771, Loss: {'train/loss': 0.6033555865287781, 'train/grad_norm_pre_clip': 14.835525512695312, 'timing/data': 0.004830910940654576, 'timing/model': 1.2046342139365152, 'learning_rate/qwen_vl_interface': 1.2244479359158895e-06, 'learning_rate/action_model': 1.2244479359158895e-05, 'epoch': 1.7}) +[2026-06-17 17:01:15,345][starVLA.training.train_starvla][INFO] - Step 5772, Loss: {'train/loss': 0.7117886543273926, 'train/grad_norm_pre_clip': 13.235054969787598, 'timing/data': 0.004908471019007266, 'timing/model': 1.2037582009797916, 'learning_rate/qwen_vl_interface': 1.2233003539802512e-06, 'learning_rate/action_model': 1.2233003539802513e-05, 'epoch': 1.7}) +[2026-06-17 17:01:20,123][starVLA.training.train_starvla][INFO] - Step 5773, Loss: {'train/loss': 0.4838898181915283, 'train/grad_norm_pre_clip': 11.901765823364258, 'timing/data': 0.005167269031517208, 'timing/model': 1.2053149370476604, 'learning_rate/qwen_vl_interface': 1.2221536067837807e-06, 'learning_rate/action_model': 1.2221536067837808e-05, 'epoch': 1.7}) +[2026-06-17 17:01:24,868][starVLA.training.train_starvla][INFO] - Step 5774, Loss: {'train/loss': 0.5186529159545898, 'train/grad_norm_pre_clip': 14.936482429504395, 'timing/data': 0.005393858999013901, 'timing/model': 1.2049025499727577, 'learning_rate/qwen_vl_interface': 1.2210076945642025e-06, 'learning_rate/action_model': 1.2210076945642024e-05, 'epoch': 1.7}) +[2026-06-17 17:01:29,645][starVLA.training.train_starvla][INFO] - Step 5775, Loss: {'train/loss': 1.008255124092102, 'train/grad_norm_pre_clip': 17.68861961364746, 'timing/data': 0.0047242759028449655, 'timing/model': 1.326428701984696, 'learning_rate/qwen_vl_interface': 1.219862617559065e-06, 'learning_rate/action_model': 1.219862617559065e-05, 'epoch': 1.7}) +[2026-06-17 17:01:34,301][starVLA.training.train_starvla][INFO] - Step 5776, Loss: {'train/loss': 0.5142174959182739, 'train/grad_norm_pre_clip': 11.145216941833496, 'timing/data': 0.005023423000238836, 'timing/model': 1.2077864989405498, 'learning_rate/qwen_vl_interface': 1.2187183760057431e-06, 'learning_rate/action_model': 1.218718376005743e-05, 'epoch': 1.7}) +[2026-06-17 17:01:39,095][starVLA.training.train_starvla][INFO] - Step 5777, Loss: {'train/loss': 0.5797067284584045, 'train/grad_norm_pre_clip': 12.798269271850586, 'timing/data': 0.004946473985910416, 'timing/model': 1.2058058560360223, 'learning_rate/qwen_vl_interface': 1.2175749701414393e-06, 'learning_rate/action_model': 1.2175749701414394e-05, 'epoch': 1.7}) +[2026-06-17 17:01:43,929][starVLA.training.train_starvla][INFO] - Step 5778, Loss: {'train/loss': 0.8337334394454956, 'train/grad_norm_pre_clip': 15.80777359008789, 'timing/data': 0.0046322239795699716, 'timing/model': 1.204281484009698, 'learning_rate/qwen_vl_interface': 1.216432400203184e-06, 'learning_rate/action_model': 1.2164324002031841e-05, 'epoch': 1.7}) +[2026-06-17 17:01:48,726][starVLA.training.train_starvla][INFO] - Step 5779, Loss: {'train/loss': 0.7386519908905029, 'train/grad_norm_pre_clip': 13.448978424072266, 'timing/data': 0.00481567601673305, 'timing/model': 1.2059772030916065, 'learning_rate/qwen_vl_interface': 1.215290666427832e-06, 'learning_rate/action_model': 1.2152906664278318e-05, 'epoch': 1.7}) +[2026-06-17 17:01:53,524][starVLA.training.train_starvla][INFO] - Step 5780, Loss: {'train/loss': 0.7948042154312134, 'train/grad_norm_pre_clip': 15.50193977355957, 'timing/data': 0.004958175937645137, 'timing/model': 1.3363112029619515, 'learning_rate/qwen_vl_interface': 1.214149769052067e-06, 'learning_rate/action_model': 1.214149769052067e-05, 'epoch': 1.7}) +[2026-06-17 17:01:58,197][starVLA.training.train_starvla][INFO] - Step 5781, Loss: {'train/loss': 0.6211493611335754, 'train/grad_norm_pre_clip': 12.821590423583984, 'timing/data': 0.004894279059953988, 'timing/model': 1.2080599220935255, 'learning_rate/qwen_vl_interface': 1.2130097083123971e-06, 'learning_rate/action_model': 1.213009708312397e-05, 'epoch': 1.7}) +[2026-06-17 17:02:02,982][starVLA.training.train_starvla][INFO] - Step 5782, Loss: {'train/loss': 0.7944750785827637, 'train/grad_norm_pre_clip': 53.60540008544922, 'timing/data': 0.004847562988288701, 'timing/model': 1.2051448620622978, 'learning_rate/qwen_vl_interface': 1.2118704844451587e-06, 'learning_rate/action_model': 1.2118704844451586e-05, 'epoch': 1.7}) +[2026-06-17 17:02:07,691][starVLA.training.train_starvla][INFO] - Step 5783, Loss: {'train/loss': 0.6255001425743103, 'train/grad_norm_pre_clip': 12.479557037353516, 'timing/data': 0.005069131962954998, 'timing/model': 1.2069118910003453, 'learning_rate/qwen_vl_interface': 1.2107320976865138e-06, 'learning_rate/action_model': 1.2107320976865139e-05, 'epoch': 1.7}) +[2026-06-17 17:02:12,462][starVLA.training.train_starvla][INFO] - Step 5784, Loss: {'train/loss': 0.4497113823890686, 'train/grad_norm_pre_clip': 11.23750114440918, 'timing/data': 0.005126361967995763, 'timing/model': 1.210885068983771, 'learning_rate/qwen_vl_interface': 1.209594548272451e-06, 'learning_rate/action_model': 1.209594548272451e-05, 'epoch': 1.7}) +[2026-06-17 17:02:17,276][starVLA.training.train_starvla][INFO] - Step 5785, Loss: {'train/loss': 0.6413657665252686, 'train/grad_norm_pre_clip': 18.131433486938477, 'timing/data': 0.00526497804094106, 'timing/model': 1.2124256789684296, 'learning_rate/qwen_vl_interface': 1.2084578364387869e-06, 'learning_rate/action_model': 1.208457836438787e-05, 'epoch': 1.7}) +[2026-06-17 17:02:22,083][starVLA.training.train_starvla][INFO] - Step 5786, Loss: {'train/loss': 0.8844735026359558, 'train/grad_norm_pre_clip': 19.618074417114258, 'timing/data': 0.004913368029519916, 'timing/model': 1.3321070478996262, 'learning_rate/qwen_vl_interface': 1.2073219624211635e-06, 'learning_rate/action_model': 1.2073219624211635e-05, 'epoch': 1.7}) +[2026-06-17 17:02:26,734][starVLA.training.train_starvla][INFO] - Step 5787, Loss: {'train/loss': 0.649420440196991, 'train/grad_norm_pre_clip': 13.477834701538086, 'timing/data': 0.005239618010818958, 'timing/model': 1.203293437953107, 'learning_rate/qwen_vl_interface': 1.2061869264550458e-06, 'learning_rate/action_model': 1.2061869264550457e-05, 'epoch': 1.7}) +[2026-06-17 17:02:31,533][starVLA.training.train_starvla][INFO] - Step 5788, Loss: {'train/loss': 0.638530433177948, 'train/grad_norm_pre_clip': 15.216492652893066, 'timing/data': 0.004805164993740618, 'timing/model': 1.2070909009780735, 'learning_rate/qwen_vl_interface': 1.2050527287757288e-06, 'learning_rate/action_model': 1.2050527287757289e-05, 'epoch': 1.7}) +[2026-06-17 17:02:36,313][starVLA.training.train_starvla][INFO] - Step 5789, Loss: {'train/loss': 0.6505452394485474, 'train/grad_norm_pre_clip': 12.794785499572754, 'timing/data': 0.005180763080716133, 'timing/model': 1.204945888952352, 'learning_rate/qwen_vl_interface': 1.203919369618335e-06, 'learning_rate/action_model': 1.2039193696183348e-05, 'epoch': 1.7}) +[2026-06-17 17:02:41,136][starVLA.training.train_starvla][INFO] - Step 5790, Loss: {'train/loss': 0.431108295917511, 'train/grad_norm_pre_clip': 11.842552185058594, 'timing/data': 0.004976411000825465, 'timing/model': 1.204891995061189, 'learning_rate/qwen_vl_interface': 1.202786849217809e-06, 'learning_rate/action_model': 1.202786849217809e-05, 'epoch': 1.7}) +[2026-06-17 17:02:45,907][starVLA.training.train_starvla][INFO] - Step 5791, Loss: {'train/loss': 0.41491538286209106, 'train/grad_norm_pre_clip': 11.715107917785645, 'timing/data': 0.004883013898506761, 'timing/model': 1.202982162940316, 'learning_rate/qwen_vl_interface': 1.2016551678089247e-06, 'learning_rate/action_model': 1.2016551678089246e-05, 'epoch': 1.7}) +[2026-06-17 17:02:50,705][starVLA.training.train_starvla][INFO] - Step 5792, Loss: {'train/loss': 0.6985031366348267, 'train/grad_norm_pre_clip': 15.279230117797852, 'timing/data': 0.004916028934530914, 'timing/model': 1.345538774970919, 'learning_rate/qwen_vl_interface': 1.2005243256262803e-06, 'learning_rate/action_model': 1.2005243256262802e-05, 'epoch': 1.7}) +[2026-06-17 17:02:55,367][starVLA.training.train_starvla][INFO] - Step 5793, Loss: {'train/loss': 0.6223912239074707, 'train/grad_norm_pre_clip': 14.171430587768555, 'timing/data': 0.004867119947448373, 'timing/model': 1.2068969079991803, 'learning_rate/qwen_vl_interface': 1.1993943229043003e-06, 'learning_rate/action_model': 1.1993943229043003e-05, 'epoch': 1.7}) +[2026-06-17 17:03:00,171][starVLA.training.train_starvla][INFO] - Step 5794, Loss: {'train/loss': 0.392214834690094, 'train/grad_norm_pre_clip': 10.753796577453613, 'timing/data': 0.0050390600226819515, 'timing/model': 1.2114353610668331, 'learning_rate/qwen_vl_interface': 1.1982651598772358e-06, 'learning_rate/action_model': 1.1982651598772356e-05, 'epoch': 1.7}) +[2026-06-17 17:03:04,972][starVLA.training.train_starvla][INFO] - Step 5795, Loss: {'train/loss': 0.5479564666748047, 'train/grad_norm_pre_clip': 13.235968589782715, 'timing/data': 0.004703432088717818, 'timing/model': 1.2154178080381826, 'learning_rate/qwen_vl_interface': 1.197136836779164e-06, 'learning_rate/action_model': 1.1971368367791639e-05, 'epoch': 1.71}) +[2026-06-17 17:03:09,770][starVLA.training.train_starvla][INFO] - Step 5796, Loss: {'train/loss': 0.709689736366272, 'train/grad_norm_pre_clip': 16.8317928314209, 'timing/data': 0.004893612000159919, 'timing/model': 1.205383257009089, 'learning_rate/qwen_vl_interface': 1.1960093538439866e-06, 'learning_rate/action_model': 1.1960093538439864e-05, 'epoch': 1.71}) +[2026-06-17 17:03:14,459][starVLA.training.train_starvla][INFO] - Step 5797, Loss: {'train/loss': 0.5390151739120483, 'train/grad_norm_pre_clip': 11.060081481933594, 'timing/data': 0.005408123950473964, 'timing/model': 1.2059511800762266, 'learning_rate/qwen_vl_interface': 1.1948827113054323e-06, 'learning_rate/action_model': 1.1948827113054323e-05, 'epoch': 1.71}) +[2026-06-17 17:03:19,177][starVLA.training.train_starvla][INFO] - Step 5798, Loss: {'train/loss': 0.7041589617729187, 'train/grad_norm_pre_clip': 16.053436279296875, 'timing/data': 0.005010711960494518, 'timing/model': 1.260118011967279, 'learning_rate/qwen_vl_interface': 1.1937569093970558e-06, 'learning_rate/action_model': 1.1937569093970558e-05, 'epoch': 1.71}) +[2026-06-17 17:03:23,836][starVLA.training.train_starvla][INFO] - Step 5799, Loss: {'train/loss': 0.7184638977050781, 'train/grad_norm_pre_clip': 16.61316680908203, 'timing/data': 0.004934986005537212, 'timing/model': 1.2050674220081419, 'learning_rate/qwen_vl_interface': 1.1926319483522353e-06, 'learning_rate/action_model': 1.1926319483522353e-05, 'epoch': 1.71}) +[2026-06-17 17:05:05,846][starVLA.training.train_starvla][INFO] - Step 5800, Loss: {'train/loss': 0.6667351126670837, 'train/grad_norm_pre_clip': 17.32037925720215, 'eval/loss': 0.7057566046714783, 'eval/latency_0/loss': 0.7057566046714783, 'eval/demon_fire_precision': 0.9001597444089456, 'eval/demon_fire_recall': 0.936046511627907, 'eval/demon_fire_f1': 0.9177524429967426, 'eval/demon_fire_rate_model': 0.696329254727475, 'eval/demon_fire_rate_teacher': 0.6696329254727474, 'eval/demon_left_precision': 0.8398320335932813, 'eval/demon_left_recall': 0.7932011331444759, 'eval/demon_left_f1': 0.8158508158508159, 'eval/demon_left_rate_model': 0.4635706340378198, 'eval/demon_left_rate_teacher': 0.49082313681868744, 'eval/demon_right_precision': 0.8113107822410148, 'eval/demon_right_recall': 0.867231638418079, 'eval/demon_right_f1': 0.8383397050791916, 'eval/demon_right_rate_model': 0.5261401557285873, 'eval/demon_right_rate_teacher': 0.4922135706340378, 'eval/demon_move_f1': 0.8270952604650037, 'eval/demon_cc_f1': 0.8885349821153284, 'eval/latency_0/demon_fire_precision': 0.9001597444089456, 'eval/latency_0/demon_fire_recall': 0.936046511627907, 'eval/latency_0/demon_fire_f1': 0.9177524429967426, 'eval/latency_0/demon_fire_rate_model': 0.696329254727475, 'eval/latency_0/demon_fire_rate_teacher': 0.6696329254727474, 'eval/latency_0/demon_left_precision': 0.8398320335932813, 'eval/latency_0/demon_left_recall': 0.7932011331444759, 'eval/latency_0/demon_left_f1': 0.8158508158508159, 'eval/latency_0/demon_left_rate_model': 0.4635706340378198, 'eval/latency_0/demon_left_rate_teacher': 0.49082313681868744, 'eval/latency_0/demon_right_precision': 0.8113107822410148, 'eval/latency_0/demon_right_recall': 0.867231638418079, 'eval/latency_0/demon_right_f1': 0.8383397050791916, 'eval/latency_0/demon_right_rate_model': 0.5261401557285873, 'eval/latency_0/demon_right_rate_teacher': 0.4922135706340378, 'eval/latency_0/demon_move_f1': 0.8270952604650037, 'eval/latency_0/demon_cc_f1': 0.8885349821153284, 'timing/data': 0.00502664502710104, 'timing/model': 1.2099118669284508, 'learning_rate/qwen_vl_interface': 1.1915078284041793e-06, 'learning_rate/action_model': 1.1915078284041793e-05, 'epoch': 1.71}) +[2026-06-17 17:05:05,846][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_5800_state +[2026-06-17 17:05:05,849][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 17:05:09,398][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_5800_state/model.safetensors +[2026-06-17 17:05:16,994][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_5800_state/optimizer.bin +[2026-06-17 17:05:16,994][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_5800_state/sampler.bin +[2026-06-17 17:05:16,994][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_5800_state/sampler_1.bin +[2026-06-17 17:05:16,995][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_5800_state/random_states_0.pkl +[2026-06-17 17:05:17,048][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 5800 +[2026-06-17 17:05:17,048][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 17:05:17,081][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 17:05:21,811][starVLA.training.train_starvla][INFO] - Step 5801, Loss: {'train/loss': 0.8496142625808716, 'train/grad_norm_pre_clip': 21.24510383605957, 'timing/data': 0.004732550936751068, 'timing/model': 1.1884535050485283, 'learning_rate/qwen_vl_interface': 1.1903845497859185e-06, 'learning_rate/action_model': 1.1903845497859184e-05, 'epoch': 1.71}) +[2026-06-17 17:05:26,483][starVLA.training.train_starvla][INFO] - Step 5802, Loss: {'train/loss': 0.6433069705963135, 'train/grad_norm_pre_clip': 14.402316093444824, 'timing/data': 0.005018125055357814, 'timing/model': 1.1915946840308607, 'learning_rate/qwen_vl_interface': 1.1892621127303067e-06, 'learning_rate/action_model': 1.1892621127303066e-05, 'epoch': 1.71}) +[2026-06-17 17:05:31,280][starVLA.training.train_starvla][INFO] - Step 5803, Loss: {'train/loss': 0.6504760980606079, 'train/grad_norm_pre_clip': 13.761495590209961, 'timing/data': 0.005117584019899368, 'timing/model': 1.1904901310335845, 'learning_rate/qwen_vl_interface': 1.1881405174700292e-06, 'learning_rate/action_model': 1.1881405174700291e-05, 'epoch': 1.71}) +[2026-06-17 17:05:36,043][starVLA.training.train_starvla][INFO] - Step 5804, Loss: {'train/loss': 0.5204727053642273, 'train/grad_norm_pre_clip': 25.89604949951172, 'timing/data': 0.004791354993358254, 'timing/model': 1.201171217020601, 'learning_rate/qwen_vl_interface': 1.1870197642375935e-06, 'learning_rate/action_model': 1.1870197642375935e-05, 'epoch': 1.71}) +[2026-06-17 17:05:40,811][starVLA.training.train_starvla][INFO] - Step 5805, Loss: {'train/loss': 0.740882933139801, 'train/grad_norm_pre_clip': 15.190098762512207, 'timing/data': 0.004813265986740589, 'timing/model': 1.3418487888993695, 'learning_rate/qwen_vl_interface': 1.1858998532653326e-06, 'learning_rate/action_model': 1.1858998532653325e-05, 'epoch': 1.71}) +[2026-06-17 17:05:45,495][starVLA.training.train_starvla][INFO] - Step 5806, Loss: {'train/loss': 0.7358133792877197, 'train/grad_norm_pre_clip': 15.653124809265137, 'timing/data': 0.005255860975012183, 'timing/model': 1.2488138700136915, 'learning_rate/qwen_vl_interface': 1.1847807847854054e-06, 'learning_rate/action_model': 1.1847807847854054e-05, 'epoch': 1.71}) +[2026-06-17 17:05:50,129][starVLA.training.train_starvla][INFO] - Step 5807, Loss: {'train/loss': 0.526508629322052, 'train/grad_norm_pre_clip': 13.05569839477539, 'timing/data': 0.004638153011910617, 'timing/model': 1.1981714529683813, 'learning_rate/qwen_vl_interface': 1.1836625590297959e-06, 'learning_rate/action_model': 1.1836625590297958e-05, 'epoch': 1.71}) +[2026-06-17 17:05:54,879][starVLA.training.train_starvla][INFO] - Step 5808, Loss: {'train/loss': 0.4716733694076538, 'train/grad_norm_pre_clip': 13.580185890197754, 'timing/data': 0.0050594189669936895, 'timing/model': 1.2015475379303098, 'learning_rate/qwen_vl_interface': 1.1825451762303134e-06, 'learning_rate/action_model': 1.1825451762303133e-05, 'epoch': 1.71}) +[2026-06-17 17:05:59,641][starVLA.training.train_starvla][INFO] - Step 5809, Loss: {'train/loss': 0.663514256477356, 'train/grad_norm_pre_clip': 16.466310501098633, 'timing/data': 0.0054523099679499865, 'timing/model': 1.2017244070302695, 'learning_rate/qwen_vl_interface': 1.1814286366185923e-06, 'learning_rate/action_model': 1.1814286366185922e-05, 'epoch': 1.71}) +[2026-06-17 17:06:04,408][starVLA.training.train_starvla][INFO] - Step 5810, Loss: {'train/loss': 0.7531523108482361, 'train/grad_norm_pre_clip': 16.523618698120117, 'timing/data': 0.004811887978576124, 'timing/model': 1.1998200439848006, 'learning_rate/qwen_vl_interface': 1.1803129404260932e-06, 'learning_rate/action_model': 1.1803129404260931e-05, 'epoch': 1.71}) +[2026-06-17 17:06:09,187][starVLA.training.train_starvla][INFO] - Step 5811, Loss: {'train/loss': 0.6489839553833008, 'train/grad_norm_pre_clip': 15.22442626953125, 'timing/data': 0.004790217964909971, 'timing/model': 1.3426342100137845, 'learning_rate/qwen_vl_interface': 1.1791980878841008e-06, 'learning_rate/action_model': 1.1791980878841008e-05, 'epoch': 1.71}) +[2026-06-17 17:06:13,950][starVLA.training.train_starvla][INFO] - Step 5812, Loss: {'train/loss': 0.5395718812942505, 'train/grad_norm_pre_clip': 12.98986530303955, 'timing/data': 0.004868861986324191, 'timing/model': 1.3276234118966386, 'learning_rate/qwen_vl_interface': 1.1780840792237246e-06, 'learning_rate/action_model': 1.1780840792237246e-05, 'epoch': 1.71}) +[2026-06-17 17:06:18,627][starVLA.training.train_starvla][INFO] - Step 5813, Loss: {'train/loss': 0.5278204083442688, 'train/grad_norm_pre_clip': 14.279845237731934, 'timing/data': 0.00498836999759078, 'timing/model': 1.2023936529876664, 'learning_rate/qwen_vl_interface': 1.1769709146759002e-06, 'learning_rate/action_model': 1.1769709146759e-05, 'epoch': 1.71}) +[2026-06-17 17:06:23,367][starVLA.training.train_starvla][INFO] - Step 5814, Loss: {'train/loss': 0.40403828024864197, 'train/grad_norm_pre_clip': 9.373223304748535, 'timing/data': 0.004641255014576018, 'timing/model': 1.1993152960203588, 'learning_rate/qwen_vl_interface': 1.175858594471388e-06, 'learning_rate/action_model': 1.175858594471388e-05, 'epoch': 1.71}) +[2026-06-17 17:06:28,221][starVLA.training.train_starvla][INFO] - Step 5815, Loss: {'train/loss': 0.6147165298461914, 'train/grad_norm_pre_clip': 16.54409408569336, 'timing/data': 0.004906554939225316, 'timing/model': 1.2011459060013294, 'learning_rate/qwen_vl_interface': 1.174747118840775e-06, 'learning_rate/action_model': 1.1747471188407749e-05, 'epoch': 1.71}) +[2026-06-17 17:06:33,007][starVLA.training.train_starvla][INFO] - Step 5816, Loss: {'train/loss': 0.7685962319374084, 'train/grad_norm_pre_clip': 13.93007755279541, 'timing/data': 0.005111636943183839, 'timing/model': 1.2037876159884036, 'learning_rate/qwen_vl_interface': 1.173636488014468e-06, 'learning_rate/action_model': 1.173636488014468e-05, 'epoch': 1.71}) +[2026-06-17 17:06:37,716][starVLA.training.train_starvla][INFO] - Step 5817, Loss: {'train/loss': 0.5423502326011658, 'train/grad_norm_pre_clip': 14.169541358947754, 'timing/data': 0.005094041000120342, 'timing/model': 1.2650213829474524, 'learning_rate/qwen_vl_interface': 1.1725267022227024e-06, 'learning_rate/action_model': 1.1725267022227025e-05, 'epoch': 1.71}) +[2026-06-17 17:06:42,359][starVLA.training.train_starvla][INFO] - Step 5818, Loss: {'train/loss': 0.46739763021469116, 'train/grad_norm_pre_clip': 28.546735763549805, 'timing/data': 0.00503359199501574, 'timing/model': 1.201289874035865, 'learning_rate/qwen_vl_interface': 1.17141776169554e-06, 'learning_rate/action_model': 1.17141776169554e-05, 'epoch': 1.71}) +[2026-06-17 17:06:47,131][starVLA.training.train_starvla][INFO] - Step 5819, Loss: {'train/loss': 0.6080187559127808, 'train/grad_norm_pre_clip': 18.88348388671875, 'timing/data': 0.004655797034502029, 'timing/model': 1.2018488990142941, 'learning_rate/qwen_vl_interface': 1.170309666662864e-06, 'learning_rate/action_model': 1.170309666662864e-05, 'epoch': 1.71}) +[2026-06-17 17:06:51,905][starVLA.training.train_starvla][INFO] - Step 5820, Loss: {'train/loss': 0.4922317862510681, 'train/grad_norm_pre_clip': 11.622868537902832, 'timing/data': 0.004879048909060657, 'timing/model': 1.2022308090236038, 'learning_rate/qwen_vl_interface': 1.1692024173543858e-06, 'learning_rate/action_model': 1.1692024173543859e-05, 'epoch': 1.71}) +[2026-06-17 17:06:56,686][starVLA.training.train_starvla][INFO] - Step 5821, Loss: {'train/loss': 0.5484291911125183, 'train/grad_norm_pre_clip': 13.696097373962402, 'timing/data': 0.004873685073107481, 'timing/model': 1.2023410479305312, 'learning_rate/qwen_vl_interface': 1.168096013999635e-06, 'learning_rate/action_model': 1.168096013999635e-05, 'epoch': 1.71}) +[2026-06-17 17:07:01,466][starVLA.training.train_starvla][INFO] - Step 5822, Loss: {'train/loss': 0.4074304401874542, 'train/grad_norm_pre_clip': 10.513227462768555, 'timing/data': 0.004900724976323545, 'timing/model': 1.3315290900645778, 'learning_rate/qwen_vl_interface': 1.1669904568279742e-06, 'learning_rate/action_model': 1.1669904568279741e-05, 'epoch': 1.71}) +[2026-06-17 17:07:06,193][starVLA.training.train_starvla][INFO] - Step 5823, Loss: {'train/loss': 0.43080779910087585, 'train/grad_norm_pre_clip': 10.952337265014648, 'timing/data': 0.0048409359296783805, 'timing/model': 1.279118517995812, 'learning_rate/qwen_vl_interface': 1.165885746068585e-06, 'learning_rate/action_model': 1.1658857460685848e-05, 'epoch': 1.71}) +[2026-06-17 17:07:10,857][starVLA.training.train_starvla][INFO] - Step 5824, Loss: {'train/loss': 0.5422253012657166, 'train/grad_norm_pre_clip': 13.274882316589355, 'timing/data': 0.005146296927705407, 'timing/model': 1.210044156992808, 'learning_rate/qwen_vl_interface': 1.1647818819504754e-06, 'learning_rate/action_model': 1.1647818819504756e-05, 'epoch': 1.71}) +[2026-06-17 17:07:15,658][starVLA.training.train_starvla][INFO] - Step 5825, Loss: {'train/loss': 0.7183095216751099, 'train/grad_norm_pre_clip': 13.806042671203613, 'timing/data': 0.004727879073470831, 'timing/model': 1.2061702409991995, 'learning_rate/qwen_vl_interface': 1.1636788647024776e-06, 'learning_rate/action_model': 1.1636788647024776e-05, 'epoch': 1.71}) +[2026-06-17 17:07:20,448][starVLA.training.train_starvla][INFO] - Step 5826, Loss: {'train/loss': 0.5481348633766174, 'train/grad_norm_pre_clip': 14.259611129760742, 'timing/data': 0.005167305003851652, 'timing/model': 1.2038111039437354, 'learning_rate/qwen_vl_interface': 1.1625766945532485e-06, 'learning_rate/action_model': 1.1625766945532484e-05, 'epoch': 1.71}) +[2026-06-17 17:07:25,246][starVLA.training.train_starvla][INFO] - Step 5827, Loss: {'train/loss': 0.5705393552780151, 'train/grad_norm_pre_clip': 15.955103874206543, 'timing/data': 0.004919394035823643, 'timing/model': 1.2098569700028747, 'learning_rate/qwen_vl_interface': 1.1614753717312683e-06, 'learning_rate/action_model': 1.1614753717312681e-05, 'epoch': 1.71}) +[2026-06-17 17:07:30,027][starVLA.training.train_starvla][INFO] - Step 5828, Loss: {'train/loss': 0.6498020887374878, 'train/grad_norm_pre_clip': 13.566961288452148, 'timing/data': 0.005147306015715003, 'timing/model': 1.2039430269505829, 'learning_rate/qwen_vl_interface': 1.1603748964648423e-06, 'learning_rate/action_model': 1.1603748964648422e-05, 'epoch': 1.71}) +[2026-06-17 17:07:34,805][starVLA.training.train_starvla][INFO] - Step 5829, Loss: {'train/loss': 0.7267784476280212, 'train/grad_norm_pre_clip': 22.874656677246094, 'timing/data': 0.004919141996651888, 'timing/model': 1.2907716570189223, 'learning_rate/qwen_vl_interface': 1.1592752689821024e-06, 'learning_rate/action_model': 1.1592752689821024e-05, 'epoch': 1.72}) +[2026-06-17 17:07:39,463][starVLA.training.train_starvla][INFO] - Step 5830, Loss: {'train/loss': 0.7140801548957825, 'train/grad_norm_pre_clip': 14.278282165527344, 'timing/data': 0.005148495081812143, 'timing/model': 1.207469827029854, 'learning_rate/qwen_vl_interface': 1.1581764895110007e-06, 'learning_rate/action_model': 1.1581764895110007e-05, 'epoch': 1.72}) +[2026-06-17 17:07:44,267][starVLA.training.train_starvla][INFO] - Step 5831, Loss: {'train/loss': 0.7379473447799683, 'train/grad_norm_pre_clip': 15.905106544494629, 'timing/data': 0.004896558937616646, 'timing/model': 1.2096045030048117, 'learning_rate/qwen_vl_interface': 1.1570785582793153e-06, 'learning_rate/action_model': 1.1570785582793154e-05, 'epoch': 1.72}) +[2026-06-17 17:07:49,066][starVLA.training.train_starvla][INFO] - Step 5832, Loss: {'train/loss': 0.8911715745925903, 'train/grad_norm_pre_clip': 14.696499824523926, 'timing/data': 0.004834480001591146, 'timing/model': 1.209802035940811, 'learning_rate/qwen_vl_interface': 1.1559814755146484e-06, 'learning_rate/action_model': 1.1559814755146483e-05, 'epoch': 1.72}) +[2026-06-17 17:07:53,867][starVLA.training.train_starvla][INFO] - Step 5833, Loss: {'train/loss': 0.8072124123573303, 'train/grad_norm_pre_clip': 16.0334415435791, 'timing/data': 0.004965017084032297, 'timing/model': 1.2059623330133036, 'learning_rate/qwen_vl_interface': 1.154885241444428e-06, 'learning_rate/action_model': 1.1548852414444277e-05, 'epoch': 1.72}) +[2026-06-17 17:07:58,660][starVLA.training.train_starvla][INFO] - Step 5834, Loss: {'train/loss': 0.43784889578819275, 'train/grad_norm_pre_clip': 19.874692916870117, 'timing/data': 0.005621050018817186, 'timing/model': 1.3333496371051297, 'learning_rate/qwen_vl_interface': 1.1537898562959039e-06, 'learning_rate/action_model': 1.1537898562959037e-05, 'epoch': 1.72}) +[2026-06-17 17:08:03,321][starVLA.training.train_starvla][INFO] - Step 5835, Loss: {'train/loss': 0.8489399552345276, 'train/grad_norm_pre_clip': 17.572174072265625, 'timing/data': 0.005167823052033782, 'timing/model': 1.2107601970201358, 'learning_rate/qwen_vl_interface': 1.152695320296149e-06, 'learning_rate/action_model': 1.1526953202961488e-05, 'epoch': 1.72}) +[2026-06-17 17:08:08,125][starVLA.training.train_starvla][INFO] - Step 5836, Loss: {'train/loss': 0.679663360118866, 'train/grad_norm_pre_clip': 14.120794296264648, 'timing/data': 0.005115104955621064, 'timing/model': 1.2044111560098827, 'learning_rate/qwen_vl_interface': 1.1516016336720635e-06, 'learning_rate/action_model': 1.1516016336720633e-05, 'epoch': 1.72}) +[2026-06-17 17:08:12,909][starVLA.training.train_starvla][INFO] - Step 5837, Loss: {'train/loss': 0.7055734395980835, 'train/grad_norm_pre_clip': 18.460416793823242, 'timing/data': 0.004854945000261068, 'timing/model': 1.2048325350042433, 'learning_rate/qwen_vl_interface': 1.1505087966503697e-06, 'learning_rate/action_model': 1.1505087966503697e-05, 'epoch': 1.72}) +[2026-06-17 17:08:17,621][starVLA.training.train_starvla][INFO] - Step 5838, Loss: {'train/loss': 0.5660134553909302, 'train/grad_norm_pre_clip': 13.671899795532227, 'timing/data': 0.005088448990136385, 'timing/model': 1.2048473570030183, 'learning_rate/qwen_vl_interface': 1.1494168094576138e-06, 'learning_rate/action_model': 1.1494168094576138e-05, 'epoch': 1.72}) +[2026-06-17 17:08:22,430][starVLA.training.train_starvla][INFO] - Step 5839, Loss: {'train/loss': 0.9916831851005554, 'train/grad_norm_pre_clip': 17.545698165893555, 'timing/data': 0.005051048938184977, 'timing/model': 1.2037185139488429, 'learning_rate/qwen_vl_interface': 1.1483256723201656e-06, 'learning_rate/action_model': 1.1483256723201655e-05, 'epoch': 1.72}) +[2026-06-17 17:08:27,367][starVLA.training.train_starvla][INFO] - Step 5840, Loss: {'train/loss': 0.6676381230354309, 'train/grad_norm_pre_clip': 15.27954387664795, 'timing/data': 0.004806114011444151, 'timing/model': 1.3345009919721633, 'learning_rate/qwen_vl_interface': 1.1472353854642187e-06, 'learning_rate/action_model': 1.1472353854642187e-05, 'epoch': 1.72}) +[2026-06-17 17:08:32,025][starVLA.training.train_starvla][INFO] - Step 5841, Loss: {'train/loss': 0.8576273918151855, 'train/grad_norm_pre_clip': 15.51265811920166, 'timing/data': 0.004998835967853665, 'timing/model': 1.207567224977538, 'learning_rate/qwen_vl_interface': 1.1461459491157914e-06, 'learning_rate/action_model': 1.1461459491157914e-05, 'epoch': 1.72}) +[2026-06-17 17:08:36,834][starVLA.training.train_starvla][INFO] - Step 5842, Loss: {'train/loss': 0.6838904619216919, 'train/grad_norm_pre_clip': 14.8526611328125, 'timing/data': 0.005455996026284993, 'timing/model': 1.2057129630120471, 'learning_rate/qwen_vl_interface': 1.145057363500725e-06, 'learning_rate/action_model': 1.1450573635007251e-05, 'epoch': 1.72}) +[2026-06-17 17:08:41,683][starVLA.training.train_starvla][INFO] - Step 5843, Loss: {'train/loss': 0.4879496097564697, 'train/grad_norm_pre_clip': 14.257603645324707, 'timing/data': 0.0049615290481597185, 'timing/model': 1.2057649330236018, 'learning_rate/qwen_vl_interface': 1.1439696288446845e-06, 'learning_rate/action_model': 1.1439696288446844e-05, 'epoch': 1.72}) +[2026-06-17 17:08:46,541][starVLA.training.train_starvla][INFO] - Step 5844, Loss: {'train/loss': 0.44197899103164673, 'train/grad_norm_pre_clip': 9.85297966003418, 'timing/data': 0.005190533003769815, 'timing/model': 1.2668728770222515, 'learning_rate/qwen_vl_interface': 1.1428827453731575e-06, 'learning_rate/action_model': 1.1428827453731574e-05, 'epoch': 1.72}) +[2026-06-17 17:08:51,302][starVLA.training.train_starvla][INFO] - Step 5845, Loss: {'train/loss': 0.5463879704475403, 'train/grad_norm_pre_clip': 14.457368850708008, 'timing/data': 0.005118625005707145, 'timing/model': 1.2050177020719275, 'learning_rate/qwen_vl_interface': 1.1417967133114575e-06, 'learning_rate/action_model': 1.1417967133114573e-05, 'epoch': 1.72}) +[2026-06-17 17:08:56,075][starVLA.training.train_starvla][INFO] - Step 5846, Loss: {'train/loss': 0.6199952960014343, 'train/grad_norm_pre_clip': 19.27836799621582, 'timing/data': 0.005121047957800329, 'timing/model': 1.3212382349884138, 'learning_rate/qwen_vl_interface': 1.140711532884718e-06, 'learning_rate/action_model': 1.140711532884718e-05, 'epoch': 1.72}) +[2026-06-17 17:09:00,733][starVLA.training.train_starvla][INFO] - Step 5847, Loss: {'train/loss': 0.590705394744873, 'train/grad_norm_pre_clip': 13.691378593444824, 'timing/data': 0.004916960024274886, 'timing/model': 1.2073025039862841, 'learning_rate/qwen_vl_interface': 1.1396272043179004e-06, 'learning_rate/action_model': 1.1396272043179005e-05, 'epoch': 1.72}) +[2026-06-17 17:09:05,516][starVLA.training.train_starvla][INFO] - Step 5848, Loss: {'train/loss': 0.7383454442024231, 'train/grad_norm_pre_clip': 14.828280448913574, 'timing/data': 0.005192894954234362, 'timing/model': 1.2034002990694717, 'learning_rate/qwen_vl_interface': 1.138543727835788e-06, 'learning_rate/action_model': 1.138543727835788e-05, 'epoch': 1.72}) +[2026-06-17 17:09:10,310][starVLA.training.train_starvla][INFO] - Step 5849, Loss: {'train/loss': 0.5702910423278809, 'train/grad_norm_pre_clip': 14.209898948669434, 'timing/data': 0.004595911013893783, 'timing/model': 1.2048996950034052, 'learning_rate/qwen_vl_interface': 1.1374611036629832e-06, 'learning_rate/action_model': 1.1374611036629831e-05, 'epoch': 1.72}) +[2026-06-17 17:09:15,107][starVLA.training.train_starvla][INFO] - Step 5850, Loss: {'train/loss': 0.8316385746002197, 'train/grad_norm_pre_clip': 18.41916847229004, 'timing/data': 0.0050233660731464624, 'timing/model': 1.2039893510518596, 'learning_rate/qwen_vl_interface': 1.136379332023916e-06, 'learning_rate/action_model': 1.1363793320239161e-05, 'epoch': 1.72}) +[2026-06-17 17:09:19,896][starVLA.training.train_starvla][INFO] - Step 5851, Loss: {'train/loss': 0.6198838949203491, 'train/grad_norm_pre_clip': 16.948986053466797, 'timing/data': 0.004953038063831627, 'timing/model': 1.335362107027322, 'learning_rate/qwen_vl_interface': 1.1352984131428413e-06, 'learning_rate/action_model': 1.1352984131428413e-05, 'epoch': 1.72}) +[2026-06-17 17:09:24,551][starVLA.training.train_starvla][INFO] - Step 5852, Loss: {'train/loss': 0.6903690695762634, 'train/grad_norm_pre_clip': 15.687026023864746, 'timing/data': 0.0052311510080471635, 'timing/model': 1.204218736034818, 'learning_rate/qwen_vl_interface': 1.1342183472438329e-06, 'learning_rate/action_model': 1.1342183472438328e-05, 'epoch': 1.72}) +[2026-06-17 17:09:29,338][starVLA.training.train_starvla][INFO] - Step 5853, Loss: {'train/loss': 0.6931148171424866, 'train/grad_norm_pre_clip': 18.33387565612793, 'timing/data': 0.004789061960764229, 'timing/model': 1.2033224379410967, 'learning_rate/qwen_vl_interface': 1.1331391345507905e-06, 'learning_rate/action_model': 1.1331391345507905e-05, 'epoch': 1.72}) +[2026-06-17 17:09:34,049][starVLA.training.train_starvla][INFO] - Step 5854, Loss: {'train/loss': 0.6769707798957825, 'train/grad_norm_pre_clip': 23.55830955505371, 'timing/data': 0.005086309975013137, 'timing/model': 1.2057877879124135, 'learning_rate/qwen_vl_interface': 1.1320607752874341e-06, 'learning_rate/action_model': 1.1320607752874341e-05, 'epoch': 1.72}) +[2026-06-17 17:09:38,787][starVLA.training.train_starvla][INFO] - Step 5855, Loss: {'train/loss': 0.7135040760040283, 'train/grad_norm_pre_clip': 14.92329216003418, 'timing/data': 0.004932770039886236, 'timing/model': 1.20630452001933, 'learning_rate/qwen_vl_interface': 1.1309832696773104e-06, 'learning_rate/action_model': 1.1309832696773104e-05, 'epoch': 1.72}) +[2026-06-17 17:09:43,583][starVLA.training.train_starvla][INFO] - Step 5856, Loss: {'train/loss': 0.6601533889770508, 'train/grad_norm_pre_clip': 14.642868041992188, 'timing/data': 0.004866852075792849, 'timing/model': 1.205126688000746, 'learning_rate/qwen_vl_interface': 1.1299066179437876e-06, 'learning_rate/action_model': 1.1299066179437876e-05, 'epoch': 1.72}) +[2026-06-17 17:09:48,376][starVLA.training.train_starvla][INFO] - Step 5857, Loss: {'train/loss': 0.9232850074768066, 'train/grad_norm_pre_clip': 16.415939331054688, 'timing/data': 0.0050166279543191195, 'timing/model': 1.3304328459780663, 'learning_rate/qwen_vl_interface': 1.1288308203100565e-06, 'learning_rate/action_model': 1.1288308203100564e-05, 'epoch': 1.72}) +[2026-06-17 17:09:53,044][starVLA.training.train_starvla][INFO] - Step 5858, Loss: {'train/loss': 0.6511042714118958, 'train/grad_norm_pre_clip': 14.94668960571289, 'timing/data': 0.004998014075681567, 'timing/model': 1.2074628280242905, 'learning_rate/qwen_vl_interface': 1.12775587699913e-06, 'learning_rate/action_model': 1.12775587699913e-05, 'epoch': 1.72}) +[2026-06-17 17:09:57,842][starVLA.training.train_starvla][INFO] - Step 5859, Loss: {'train/loss': 0.48391014337539673, 'train/grad_norm_pre_clip': 20.019229888916016, 'timing/data': 0.0049053579568862915, 'timing/model': 1.2080626040697098, 'learning_rate/qwen_vl_interface': 1.126681788233847e-06, 'learning_rate/action_model': 1.1266817882338468e-05, 'epoch': 1.72}) +[2026-06-17 17:10:02,629][starVLA.training.train_starvla][INFO] - Step 5860, Loss: {'train/loss': 0.7060743570327759, 'train/grad_norm_pre_clip': 13.342193603515625, 'timing/data': 0.004678551922552288, 'timing/model': 1.2116805920377374, 'learning_rate/qwen_vl_interface': 1.1256085542368655e-06, 'learning_rate/action_model': 1.1256085542368655e-05, 'epoch': 1.72}) +[2026-06-17 17:10:07,386][starVLA.training.train_starvla][INFO] - Step 5861, Loss: {'train/loss': 1.1204264163970947, 'train/grad_norm_pre_clip': 17.333208084106445, 'timing/data': 0.004928536945953965, 'timing/model': 1.2125827280106023, 'learning_rate/qwen_vl_interface': 1.1245361752306676e-06, 'learning_rate/action_model': 1.1245361752306677e-05, 'epoch': 1.72}) +[2026-06-17 17:10:12,191][starVLA.training.train_starvla][INFO] - Step 5862, Loss: {'train/loss': 0.5904516577720642, 'train/grad_norm_pre_clip': 11.67636775970459, 'timing/data': 0.005248437984846532, 'timing/model': 1.210003218962811, 'learning_rate/qwen_vl_interface': 1.1234646514375625e-06, 'learning_rate/action_model': 1.1234646514375624e-05, 'epoch': 1.72}) +[2026-06-17 17:10:16,983][starVLA.training.train_starvla][INFO] - Step 5863, Loss: {'train/loss': 0.5995621085166931, 'train/grad_norm_pre_clip': 12.50796890258789, 'timing/data': 0.005126456031575799, 'timing/model': 1.3382149179233238, 'learning_rate/qwen_vl_interface': 1.122393983079674e-06, 'learning_rate/action_model': 1.122393983079674e-05, 'epoch': 1.73}) +[2026-06-17 17:10:21,692][starVLA.training.train_starvla][INFO] - Step 5864, Loss: {'train/loss': 0.6957294940948486, 'train/grad_norm_pre_clip': 16.052906036376953, 'timing/data': 0.005413544946350157, 'timing/model': 1.2120008479105309, 'learning_rate/qwen_vl_interface': 1.1213241703789534e-06, 'learning_rate/action_model': 1.1213241703789534e-05, 'epoch': 1.73}) +[2026-06-17 17:10:26,495][starVLA.training.train_starvla][INFO] - Step 5865, Loss: {'train/loss': 0.5385894775390625, 'train/grad_norm_pre_clip': 13.235694885253906, 'timing/data': 0.005273871938697994, 'timing/model': 1.2074754829518497, 'learning_rate/qwen_vl_interface': 1.1202552135571759e-06, 'learning_rate/action_model': 1.1202552135571759e-05, 'epoch': 1.73}) +[2026-06-17 17:10:31,281][starVLA.training.train_starvla][INFO] - Step 5866, Loss: {'train/loss': 0.3593005836009979, 'train/grad_norm_pre_clip': 10.588595390319824, 'timing/data': 0.005039692972786725, 'timing/model': 1.206213210010901, 'learning_rate/qwen_vl_interface': 1.1191871128359367e-06, 'learning_rate/action_model': 1.1191871128359365e-05, 'epoch': 1.73}) +[2026-06-17 17:10:36,092][starVLA.training.train_starvla][INFO] - Step 5867, Loss: {'train/loss': 0.6598596572875977, 'train/grad_norm_pre_clip': 15.309645652770996, 'timing/data': 0.00503533601295203, 'timing/model': 1.2063426589593291, 'learning_rate/qwen_vl_interface': 1.1181198684366548e-06, 'learning_rate/action_model': 1.1181198684366548e-05, 'epoch': 1.73}) +[2026-06-17 17:10:40,773][starVLA.training.train_starvla][INFO] - Step 5868, Loss: {'train/loss': 0.39447152614593506, 'train/grad_norm_pre_clip': 11.718550682067871, 'timing/data': 0.005074099055491388, 'timing/model': 1.2077395940432325, 'learning_rate/qwen_vl_interface': 1.117053480580568e-06, 'learning_rate/action_model': 1.1170534805805679e-05, 'epoch': 1.73}) +[2026-06-17 17:10:45,472][starVLA.training.train_starvla][INFO] - Step 5869, Loss: {'train/loss': 0.37150824069976807, 'train/grad_norm_pre_clip': 10.368943214416504, 'timing/data': 0.004991465946659446, 'timing/model': 1.2397167269373313, 'learning_rate/qwen_vl_interface': 1.1159879494887429e-06, 'learning_rate/action_model': 1.115987949488743e-05, 'epoch': 1.73}) +[2026-06-17 17:10:50,123][starVLA.training.train_starvla][INFO] - Step 5870, Loss: {'train/loss': 0.6320111155509949, 'train/grad_norm_pre_clip': 14.451017379760742, 'timing/data': 0.004850637051276863, 'timing/model': 1.2047972599975765, 'learning_rate/qwen_vl_interface': 1.1149232753820645e-06, 'learning_rate/action_model': 1.1149232753820647e-05, 'epoch': 1.73}) +[2026-06-17 17:10:54,864][starVLA.training.train_starvla][INFO] - Step 5871, Loss: {'train/loss': 0.42936980724334717, 'train/grad_norm_pre_clip': 11.977627754211426, 'timing/data': 0.005112654063850641, 'timing/model': 1.2066936990013346, 'learning_rate/qwen_vl_interface': 1.1138594584812406e-06, 'learning_rate/action_model': 1.1138594584812406e-05, 'epoch': 1.73}) +[2026-06-17 17:10:59,651][starVLA.training.train_starvla][INFO] - Step 5872, Loss: {'train/loss': 0.727445125579834, 'train/grad_norm_pre_clip': 15.078542709350586, 'timing/data': 0.004980958998203278, 'timing/model': 1.2045014719478786, 'learning_rate/qwen_vl_interface': 1.1127964990068017e-06, 'learning_rate/action_model': 1.1127964990068017e-05, 'epoch': 1.73}) +[2026-06-17 17:11:04,443][starVLA.training.train_starvla][INFO] - Step 5873, Loss: {'train/loss': 0.44112449884414673, 'train/grad_norm_pre_clip': 11.882144927978516, 'timing/data': 0.004808680037967861, 'timing/model': 1.2091751920524985, 'learning_rate/qwen_vl_interface': 1.1117343971791005e-06, 'learning_rate/action_model': 1.1117343971791006e-05, 'epoch': 1.73}) +[2026-06-17 17:11:09,242][starVLA.training.train_starvla][INFO] - Step 5874, Loss: {'train/loss': 0.6608772873878479, 'train/grad_norm_pre_clip': 15.987635612487793, 'timing/data': 0.005048300023190677, 'timing/model': 1.327788466005586, 'learning_rate/qwen_vl_interface': 1.1106731532183114e-06, 'learning_rate/action_model': 1.1106731532183114e-05, 'epoch': 1.73}) +[2026-06-17 17:11:13,943][starVLA.training.train_starvla][INFO] - Step 5875, Loss: {'train/loss': 0.44503331184387207, 'train/grad_norm_pre_clip': 10.725570678710938, 'timing/data': 0.004911677911877632, 'timing/model': 1.2491866929922253, 'learning_rate/qwen_vl_interface': 1.1096127673444309e-06, 'learning_rate/action_model': 1.109612767344431e-05, 'epoch': 1.73}) +[2026-06-17 17:11:18,603][starVLA.training.train_starvla][INFO] - Step 5876, Loss: {'train/loss': 1.0337934494018555, 'train/grad_norm_pre_clip': 19.20513343811035, 'timing/data': 0.005226360983215272, 'timing/model': 1.2052522409940138, 'learning_rate/qwen_vl_interface': 1.108553239777281e-06, 'learning_rate/action_model': 1.108553239777281e-05, 'epoch': 1.73}) +[2026-06-17 17:11:23,351][starVLA.training.train_starvla][INFO] - Step 5877, Loss: {'train/loss': 0.5313777327537537, 'train/grad_norm_pre_clip': 13.174257278442383, 'timing/data': 0.004795655026100576, 'timing/model': 1.204399668960832, 'learning_rate/qwen_vl_interface': 1.1074945707365004e-06, 'learning_rate/action_model': 1.1074945707365003e-05, 'epoch': 1.73}) +[2026-06-17 17:11:28,154][starVLA.training.train_starvla][INFO] - Step 5878, Loss: {'train/loss': 0.48850440979003906, 'train/grad_norm_pre_clip': 10.875962257385254, 'timing/data': 0.005221924977377057, 'timing/model': 1.2068358940305188, 'learning_rate/qwen_vl_interface': 1.1064367604415522e-06, 'learning_rate/action_model': 1.1064367604415521e-05, 'epoch': 1.73}) +[2026-06-17 17:11:32,948][starVLA.training.train_starvla][INFO] - Step 5879, Loss: {'train/loss': 0.6703518033027649, 'train/grad_norm_pre_clip': 12.796773910522461, 'timing/data': 0.004930748022161424, 'timing/model': 1.2060606279410422, 'learning_rate/qwen_vl_interface': 1.105379809111721e-06, 'learning_rate/action_model': 1.105379809111721e-05, 'epoch': 1.73}) +[2026-06-17 17:11:37,752][starVLA.training.train_starvla][INFO] - Step 5880, Loss: {'train/loss': 0.6986896395683289, 'train/grad_norm_pre_clip': 21.028650283813477, 'timing/data': 0.005051397019997239, 'timing/model': 1.34686029702425, 'learning_rate/qwen_vl_interface': 1.1043237169661167e-06, 'learning_rate/action_model': 1.1043237169661166e-05, 'epoch': 1.73}) +[2026-06-17 17:11:42,536][starVLA.training.train_starvla][INFO] - Step 5881, Loss: {'train/loss': 0.7200603485107422, 'train/grad_norm_pre_clip': 14.268406867980957, 'timing/data': 0.004933746997267008, 'timing/model': 1.3254490459803492, 'learning_rate/qwen_vl_interface': 1.103268484223667e-06, 'learning_rate/action_model': 1.1032684842236669e-05, 'epoch': 1.73}) +[2026-06-17 17:11:47,200][starVLA.training.train_starvla][INFO] - Step 5882, Loss: {'train/loss': 0.6251131892204285, 'train/grad_norm_pre_clip': 14.397034645080566, 'timing/data': 0.0051668069791048765, 'timing/model': 1.2084587760036811, 'learning_rate/qwen_vl_interface': 1.102214111103123e-06, 'learning_rate/action_model': 1.102214111103123e-05, 'epoch': 1.73}) +[2026-06-17 17:11:52,029][starVLA.training.train_starvla][INFO] - Step 5883, Loss: {'train/loss': 0.8617270588874817, 'train/grad_norm_pre_clip': 17.149906158447266, 'timing/data': 0.005440017092041671, 'timing/model': 1.2037884580204263, 'learning_rate/qwen_vl_interface': 1.1011605978230552e-06, 'learning_rate/action_model': 1.1011605978230551e-05, 'epoch': 1.73}) +[2026-06-17 17:11:56,811][starVLA.training.train_starvla][INFO] - Step 5884, Loss: {'train/loss': 0.6328235268592834, 'train/grad_norm_pre_clip': 16.987449645996094, 'timing/data': 0.005126517033204436, 'timing/model': 1.2041330609936267, 'learning_rate/qwen_vl_interface': 1.100107944601861e-06, 'learning_rate/action_model': 1.100107944601861e-05, 'epoch': 1.73}) +[2026-06-17 17:12:01,667][starVLA.training.train_starvla][INFO] - Step 5885, Loss: {'train/loss': 0.5911046266555786, 'train/grad_norm_pre_clip': 11.754027366638184, 'timing/data': 0.004992994014173746, 'timing/model': 1.2050445990171283, 'learning_rate/qwen_vl_interface': 1.0990561516577552e-06, 'learning_rate/action_model': 1.0990561516577552e-05, 'epoch': 1.73}) +[2026-06-17 17:12:06,401][starVLA.training.train_starvla][INFO] - Step 5886, Loss: {'train/loss': 0.6260499954223633, 'train/grad_norm_pre_clip': 12.61587142944336, 'timing/data': 0.005751168937422335, 'timing/model': 1.2770218149526045, 'learning_rate/qwen_vl_interface': 1.0980052192087753e-06, 'learning_rate/action_model': 1.0980052192087754e-05, 'epoch': 1.73}) +[2026-06-17 17:12:11,056][starVLA.training.train_starvla][INFO] - Step 5887, Loss: {'train/loss': 0.33088481426239014, 'train/grad_norm_pre_clip': 11.228235244750977, 'timing/data': 0.004706633975729346, 'timing/model': 1.2032441400224343, 'learning_rate/qwen_vl_interface': 1.096955147472782e-06, 'learning_rate/action_model': 1.0969551474727821e-05, 'epoch': 1.73}) +[2026-06-17 17:12:15,819][starVLA.training.train_starvla][INFO] - Step 5888, Loss: {'train/loss': 0.7290011644363403, 'train/grad_norm_pre_clip': 16.42046546936035, 'timing/data': 0.004804730997420847, 'timing/model': 1.2049246189417318, 'learning_rate/qwen_vl_interface': 1.095905936667455e-06, 'learning_rate/action_model': 1.0959059366674552e-05, 'epoch': 1.73}) +[2026-06-17 17:12:20,609][starVLA.training.train_starvla][INFO] - Step 5889, Loss: {'train/loss': 0.6163994073867798, 'train/grad_norm_pre_clip': 17.046728134155273, 'timing/data': 0.005003149970434606, 'timing/model': 1.208923743921332, 'learning_rate/qwen_vl_interface': 1.0948575870102977e-06, 'learning_rate/action_model': 1.0948575870102978e-05, 'epoch': 1.73}) +[2026-06-17 17:12:25,435][starVLA.training.train_starvla][INFO] - Step 5890, Loss: {'train/loss': 0.5046169757843018, 'train/grad_norm_pre_clip': 18.832962036132812, 'timing/data': 0.004990252899006009, 'timing/model': 1.2177755520679057, 'learning_rate/qwen_vl_interface': 1.0938100987186325e-06, 'learning_rate/action_model': 1.0938100987186323e-05, 'epoch': 1.73}) +[2026-06-17 17:12:30,237][starVLA.training.train_starvla][INFO] - Step 5891, Loss: {'train/loss': 0.5223468542098999, 'train/grad_norm_pre_clip': 13.289294242858887, 'timing/data': 0.005134232924319804, 'timing/model': 1.3426563979592174, 'learning_rate/qwen_vl_interface': 1.092763472009608e-06, 'learning_rate/action_model': 1.092763472009608e-05, 'epoch': 1.73}) +[2026-06-17 17:12:34,988][starVLA.training.train_starvla][INFO] - Step 5892, Loss: {'train/loss': 0.6591747999191284, 'train/grad_norm_pre_clip': 14.406423568725586, 'timing/data': 0.00522797997109592, 'timing/model': 1.2900386590044945, 'learning_rate/qwen_vl_interface': 1.0917177071001884e-06, 'learning_rate/action_model': 1.0917177071001884e-05, 'epoch': 1.73}) +[2026-06-17 17:12:39,698][starVLA.training.train_starvla][INFO] - Step 5893, Loss: {'train/loss': 0.3658375144004822, 'train/grad_norm_pre_clip': 9.866415977478027, 'timing/data': 0.004896743921563029, 'timing/model': 1.2093812960665673, 'learning_rate/qwen_vl_interface': 1.0906728042071627e-06, 'learning_rate/action_model': 1.0906728042071627e-05, 'epoch': 1.73}) +[2026-06-17 17:12:44,518][starVLA.training.train_starvla][INFO] - Step 5894, Loss: {'train/loss': 0.6111481189727783, 'train/grad_norm_pre_clip': 13.670190811157227, 'timing/data': 0.005000358913093805, 'timing/model': 1.2144705419195816, 'learning_rate/qwen_vl_interface': 1.089628763547139e-06, 'learning_rate/action_model': 1.0896287635471389e-05, 'epoch': 1.73}) +[2026-06-17 17:12:49,327][starVLA.training.train_starvla][INFO] - Step 5895, Loss: {'train/loss': 0.8126929998397827, 'train/grad_norm_pre_clip': 15.474124908447266, 'timing/data': 0.005073387990705669, 'timing/model': 1.2063371039694175, 'learning_rate/qwen_vl_interface': 1.0885855853365505e-06, 'learning_rate/action_model': 1.0885855853365503e-05, 'epoch': 1.73}) +[2026-06-17 17:12:54,133][starVLA.training.train_starvla][INFO] - Step 5896, Loss: {'train/loss': 0.4982962906360626, 'train/grad_norm_pre_clip': 15.206475257873535, 'timing/data': 0.005276573938317597, 'timing/model': 1.2024326430400833, 'learning_rate/qwen_vl_interface': 1.0875432697916492e-06, 'learning_rate/action_model': 1.087543269791649e-05, 'epoch': 1.73}) +[2026-06-17 17:12:58,928][starVLA.training.train_starvla][INFO] - Step 5897, Loss: {'train/loss': 0.4052012264728546, 'train/grad_norm_pre_clip': 13.639165878295898, 'timing/data': 0.0046199619537219405, 'timing/model': 1.207326842006296, 'learning_rate/qwen_vl_interface': 1.0865018171285047e-06, 'learning_rate/action_model': 1.0865018171285047e-05, 'epoch': 1.74}) +[2026-06-17 17:13:03,688][starVLA.training.train_starvla][INFO] - Step 5898, Loss: {'train/loss': 0.46798351407051086, 'train/grad_norm_pre_clip': 10.65888786315918, 'timing/data': 0.00504580105189234, 'timing/model': 1.2924849779810756, 'learning_rate/qwen_vl_interface': 1.0854612275630146e-06, 'learning_rate/action_model': 1.0854612275630146e-05, 'epoch': 1.74}) +[2026-06-17 17:13:08,363][starVLA.training.train_starvla][INFO] - Step 5899, Loss: {'train/loss': 0.6791748404502869, 'train/grad_norm_pre_clip': 15.72386646270752, 'timing/data': 0.005063742981292307, 'timing/model': 1.2162188389338553, 'learning_rate/qwen_vl_interface': 1.0844215013108936e-06, 'learning_rate/action_model': 1.0844215013108936e-05, 'epoch': 1.74}) +[2026-06-17 17:14:50,554][starVLA.training.train_starvla][INFO] - Step 5900, Loss: {'train/loss': 0.6145551204681396, 'train/grad_norm_pre_clip': 12.618557929992676, 'eval/loss': 0.7044247984886169, 'eval/latency_0/loss': 0.7044247984886169, 'eval/demon_fire_precision': 0.9099062372604974, 'eval/demon_fire_recall': 0.9269102990033222, 'eval/demon_fire_f1': 0.9183295618185557, 'eval/demon_fire_rate_model': 0.682146829810901, 'eval/demon_fire_rate_teacher': 0.6696329254727474, 'eval/demon_left_precision': 0.84007242003621, 'eval/demon_left_recall': 0.7886685552407932, 'eval/demon_left_f1': 0.8135593220338982, 'eval/demon_left_rate_model': 0.460789766407119, 'eval/demon_left_rate_teacher': 0.49082313681868744, 'eval/demon_right_precision': 0.8043818466353677, 'eval/demon_right_recall': 0.8711864406779661, 'eval/demon_right_f1': 0.836452400325468, 'eval/demon_right_rate_model': 0.5330923248053393, 'eval/demon_right_rate_teacher': 0.4922135706340378, 'eval/demon_move_f1': 0.8250058611796831, 'eval/demon_cc_f1': 0.8881883140182573, 'eval/latency_0/demon_fire_precision': 0.9099062372604974, 'eval/latency_0/demon_fire_recall': 0.9269102990033222, 'eval/latency_0/demon_fire_f1': 0.9183295618185557, 'eval/latency_0/demon_fire_rate_model': 0.682146829810901, 'eval/latency_0/demon_fire_rate_teacher': 0.6696329254727474, 'eval/latency_0/demon_left_precision': 0.84007242003621, 'eval/latency_0/demon_left_recall': 0.7886685552407932, 'eval/latency_0/demon_left_f1': 0.8135593220338982, 'eval/latency_0/demon_left_rate_model': 0.460789766407119, 'eval/latency_0/demon_left_rate_teacher': 0.49082313681868744, 'eval/latency_0/demon_right_precision': 0.8043818466353677, 'eval/latency_0/demon_right_recall': 0.8711864406779661, 'eval/latency_0/demon_right_f1': 0.836452400325468, 'eval/latency_0/demon_right_rate_model': 0.5330923248053393, 'eval/latency_0/demon_right_rate_teacher': 0.4922135706340378, 'eval/latency_0/demon_move_f1': 0.8250058611796831, 'eval/latency_0/demon_cc_f1': 0.8881883140182573, 'timing/data': 0.00494728097692132, 'timing/model': 1.2067389918956906, 'learning_rate/qwen_vl_interface': 1.0833826385876772e-06, 'learning_rate/action_model': 1.0833826385876771e-05, 'epoch': 1.74}) +[2026-06-17 17:14:50,554][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_5900_state +[2026-06-17 17:14:50,557][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 17:14:54,120][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_5900_state/model.safetensors +[2026-06-17 17:15:01,760][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_5900_state/optimizer.bin +[2026-06-17 17:15:01,761][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_5900_state/sampler.bin +[2026-06-17 17:15:01,761][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_5900_state/sampler_1.bin +[2026-06-17 17:15:01,762][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_5900_state/random_states_0.pkl +[2026-06-17 17:15:01,806][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 5900 +[2026-06-17 17:15:01,806][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 17:15:01,851][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 17:15:06,607][starVLA.training.train_starvla][INFO] - Step 5901, Loss: {'train/loss': 0.8996426463127136, 'train/grad_norm_pre_clip': 21.057703018188477, 'timing/data': 0.005307275918312371, 'timing/model': 1.1878226019907743, 'learning_rate/qwen_vl_interface': 1.0823446396087239e-06, 'learning_rate/action_model': 1.0823446396087239e-05, 'epoch': 1.74}) +[2026-06-17 17:15:11,233][starVLA.training.train_starvla][INFO] - Step 5902, Loss: {'train/loss': 0.5856319069862366, 'train/grad_norm_pre_clip': 14.702781677246094, 'timing/data': 0.005321946926414967, 'timing/model': 1.1931196079822257, 'learning_rate/qwen_vl_interface': 1.0813075045892116e-06, 'learning_rate/action_model': 1.0813075045892116e-05, 'epoch': 1.74}) +[2026-06-17 17:15:15,898][starVLA.training.train_starvla][INFO] - Step 5903, Loss: {'train/loss': 0.6861574053764343, 'train/grad_norm_pre_clip': 14.860642433166504, 'timing/data': 0.005055543966591358, 'timing/model': 1.1941179260611534, 'learning_rate/qwen_vl_interface': 1.0802712337441383e-06, 'learning_rate/action_model': 1.0802712337441384e-05, 'epoch': 1.74}) +[2026-06-17 17:15:20,636][starVLA.training.train_starvla][INFO] - Step 5904, Loss: {'train/loss': 0.4720970392227173, 'train/grad_norm_pre_clip': 11.585465431213379, 'timing/data': 0.005456323968246579, 'timing/model': 1.1939406340243295, 'learning_rate/qwen_vl_interface': 1.0792358272883262e-06, 'learning_rate/action_model': 1.0792358272883262e-05, 'epoch': 1.74}) +[2026-06-17 17:15:25,395][starVLA.training.train_starvla][INFO] - Step 5905, Loss: {'train/loss': 0.6444029211997986, 'train/grad_norm_pre_clip': 14.498576164245605, 'timing/data': 0.004974856972694397, 'timing/model': 1.3312106920639053, 'learning_rate/qwen_vl_interface': 1.0782012854364145e-06, 'learning_rate/action_model': 1.0782012854364145e-05, 'epoch': 1.74}) +[2026-06-17 17:15:30,026][starVLA.training.train_starvla][INFO] - Step 5906, Loss: {'train/loss': 0.6168133020401001, 'train/grad_norm_pre_clip': 12.918061256408691, 'timing/data': 0.005134537001140416, 'timing/model': 1.1981301200576127, 'learning_rate/qwen_vl_interface': 1.0771676084028652e-06, 'learning_rate/action_model': 1.0771676084028652e-05, 'epoch': 1.74}) +[2026-06-17 17:15:34,780][starVLA.training.train_starvla][INFO] - Step 5907, Loss: {'train/loss': 0.9150829315185547, 'train/grad_norm_pre_clip': 19.12078285217285, 'timing/data': 0.005522581981495023, 'timing/model': 1.1975217059953138, 'learning_rate/qwen_vl_interface': 1.0761347964019611e-06, 'learning_rate/action_model': 1.076134796401961e-05, 'epoch': 1.74}) +[2026-06-17 17:15:39,542][starVLA.training.train_starvla][INFO] - Step 5908, Loss: {'train/loss': 0.570344090461731, 'train/grad_norm_pre_clip': 13.802939414978027, 'timing/data': 0.00514981197193265, 'timing/model': 1.1983765379991382, 'learning_rate/qwen_vl_interface': 1.075102849647804e-06, 'learning_rate/action_model': 1.0751028496478038e-05, 'epoch': 1.74}) +[2026-06-17 17:15:44,297][starVLA.training.train_starvla][INFO] - Step 5909, Loss: {'train/loss': 0.5348514914512634, 'train/grad_norm_pre_clip': 12.89009952545166, 'timing/data': 0.005032907007262111, 'timing/model': 1.2026880970224738, 'learning_rate/qwen_vl_interface': 1.0740717683543193e-06, 'learning_rate/action_model': 1.0740717683543191e-05, 'epoch': 1.74}) +[2026-06-17 17:15:49,032][starVLA.training.train_starvla][INFO] - Step 5910, Loss: {'train/loss': 0.6262872815132141, 'train/grad_norm_pre_clip': 11.460944175720215, 'timing/data': 0.005596593022346497, 'timing/model': 1.1993433439638466, 'learning_rate/qwen_vl_interface': 1.0730415527352518e-06, 'learning_rate/action_model': 1.0730415527352517e-05, 'epoch': 1.74}) +[2026-06-17 17:15:53,794][starVLA.training.train_starvla][INFO] - Step 5911, Loss: {'train/loss': 0.4629930257797241, 'train/grad_norm_pre_clip': 11.942481994628906, 'timing/data': 0.005704432958737016, 'timing/model': 1.317872463958338, 'learning_rate/qwen_vl_interface': 1.0720122030041638e-06, 'learning_rate/action_model': 1.0720122030041637e-05, 'epoch': 1.74}) +[2026-06-17 17:15:58,448][starVLA.training.train_starvla][INFO] - Step 5912, Loss: {'train/loss': 0.7778118848800659, 'train/grad_norm_pre_clip': 18.00894546508789, 'timing/data': 0.005277154967188835, 'timing/model': 1.214166641002521, 'learning_rate/qwen_vl_interface': 1.0709837193744412e-06, 'learning_rate/action_model': 1.070983719374441e-05, 'epoch': 1.74}) +[2026-06-17 17:16:03,230][starVLA.training.train_starvla][INFO] - Step 5913, Loss: {'train/loss': 0.38887470960617065, 'train/grad_norm_pre_clip': 16.92731475830078, 'timing/data': 0.005187995033338666, 'timing/model': 1.2037603160133585, 'learning_rate/qwen_vl_interface': 1.069956102059291e-06, 'learning_rate/action_model': 1.069956102059291e-05, 'epoch': 1.74}) +[2026-06-17 17:16:08,012][starVLA.training.train_starvla][INFO] - Step 5914, Loss: {'train/loss': 0.4905535876750946, 'train/grad_norm_pre_clip': 14.41109848022461, 'timing/data': 0.005531427916139364, 'timing/model': 1.2010226680431515, 'learning_rate/qwen_vl_interface': 1.0689293512717387e-06, 'learning_rate/action_model': 1.0689293512717385e-05, 'epoch': 1.74}) +[2026-06-17 17:16:12,874][starVLA.training.train_starvla][INFO] - Step 5915, Loss: {'train/loss': 0.527762234210968, 'train/grad_norm_pre_clip': 12.78602409362793, 'timing/data': 0.005482148029841483, 'timing/model': 1.200886677019298, 'learning_rate/qwen_vl_interface': 1.0679034672246322e-06, 'learning_rate/action_model': 1.0679034672246322e-05, 'epoch': 1.74}) +[2026-06-17 17:16:17,639][starVLA.training.train_starvla][INFO] - Step 5916, Loss: {'train/loss': 0.5157292485237122, 'train/grad_norm_pre_clip': 15.045809745788574, 'timing/data': 0.0055637439945712686, 'timing/model': 1.3287127780495211, 'learning_rate/qwen_vl_interface': 1.0668784501306348e-06, 'learning_rate/action_model': 1.0668784501306348e-05, 'epoch': 1.74}) +[2026-06-17 17:16:22,281][starVLA.training.train_starvla][INFO] - Step 5917, Loss: {'train/loss': 0.44656962156295776, 'train/grad_norm_pre_clip': 12.784817695617676, 'timing/data': 0.005354149034246802, 'timing/model': 1.2016528840176761, 'learning_rate/qwen_vl_interface': 1.0658543002022377e-06, 'learning_rate/action_model': 1.0658543002022375e-05, 'epoch': 1.74}) +[2026-06-17 17:16:27,088][starVLA.training.train_starvla][INFO] - Step 5918, Loss: {'train/loss': 0.9709853529930115, 'train/grad_norm_pre_clip': 16.83222007751465, 'timing/data': 0.0052916500717401505, 'timing/model': 1.2016241690143943, 'learning_rate/qwen_vl_interface': 1.0648310176517454e-06, 'learning_rate/action_model': 1.0648310176517453e-05, 'epoch': 1.74}) +[2026-06-17 17:16:31,856][starVLA.training.train_starvla][INFO] - Step 5919, Loss: {'train/loss': 0.4956202208995819, 'train/grad_norm_pre_clip': 13.381817817687988, 'timing/data': 0.005145424976944923, 'timing/model': 1.2020984939299524, 'learning_rate/qwen_vl_interface': 1.0638086026912873e-06, 'learning_rate/action_model': 1.0638086026912872e-05, 'epoch': 1.74}) +[2026-06-17 17:16:36,590][starVLA.training.train_starvla][INFO] - Step 5920, Loss: {'train/loss': 0.4112253487110138, 'train/grad_norm_pre_clip': 10.703289031982422, 'timing/data': 0.005418302956968546, 'timing/model': 1.2065657449420542, 'learning_rate/qwen_vl_interface': 1.0627870555328102e-06, 'learning_rate/action_model': 1.06278705553281e-05, 'epoch': 1.74}) +[2026-06-17 17:16:41,369][starVLA.training.train_starvla][INFO] - Step 5921, Loss: {'train/loss': 0.6132277250289917, 'train/grad_norm_pre_clip': 14.785954475402832, 'timing/data': 0.005610903026536107, 'timing/model': 1.2020600730320439, 'learning_rate/qwen_vl_interface': 1.061766376388082e-06, 'learning_rate/action_model': 1.0617663763880821e-05, 'epoch': 1.74}) +[2026-06-17 17:16:46,156][starVLA.training.train_starvla][INFO] - Step 5922, Loss: {'train/loss': 0.5332021713256836, 'train/grad_norm_pre_clip': 21.654050827026367, 'timing/data': 0.0053987279534339905, 'timing/model': 1.3403379860101268, 'learning_rate/qwen_vl_interface': 1.0607465654686908e-06, 'learning_rate/action_model': 1.0607465654686907e-05, 'epoch': 1.74}) +[2026-06-17 17:16:50,820][starVLA.training.train_starvla][INFO] - Step 5923, Loss: {'train/loss': 0.5087149739265442, 'train/grad_norm_pre_clip': 12.58407974243164, 'timing/data': 0.005357422051019967, 'timing/model': 1.2030896389624104, 'learning_rate/qwen_vl_interface': 1.0597276229860434e-06, 'learning_rate/action_model': 1.0597276229860434e-05, 'epoch': 1.74}) +[2026-06-17 17:16:55,601][starVLA.training.train_starvla][INFO] - Step 5924, Loss: {'train/loss': 0.7172659635543823, 'train/grad_norm_pre_clip': 17.33858299255371, 'timing/data': 0.005470862030051649, 'timing/model': 1.2048292680410668, 'learning_rate/qwen_vl_interface': 1.0587095491513709e-06, 'learning_rate/action_model': 1.0587095491513709e-05, 'epoch': 1.74}) +[2026-06-17 17:17:00,374][starVLA.training.train_starvla][INFO] - Step 5925, Loss: {'train/loss': 0.3837825655937195, 'train/grad_norm_pre_clip': 12.142361640930176, 'timing/data': 0.005290930042974651, 'timing/model': 1.2040863930014893, 'learning_rate/qwen_vl_interface': 1.0576923441757174e-06, 'learning_rate/action_model': 1.0576923441757176e-05, 'epoch': 1.74}) +[2026-06-17 17:17:05,135][starVLA.training.train_starvla][INFO] - Step 5926, Loss: {'train/loss': 0.7663467526435852, 'train/grad_norm_pre_clip': 15.777857780456543, 'timing/data': 0.005962458089925349, 'timing/model': 1.2037305729463696, 'learning_rate/qwen_vl_interface': 1.0566760082699515e-06, 'learning_rate/action_model': 1.0566760082699514e-05, 'epoch': 1.74}) +[2026-06-17 17:17:09,912][starVLA.training.train_starvla][INFO] - Step 5927, Loss: {'train/loss': 0.7092075347900391, 'train/grad_norm_pre_clip': 16.454238891601562, 'timing/data': 0.005302015109919012, 'timing/model': 1.2065637540072203, 'learning_rate/qwen_vl_interface': 1.0556605416447626e-06, 'learning_rate/action_model': 1.0556605416447625e-05, 'epoch': 1.74}) +[2026-06-17 17:17:14,693][starVLA.training.train_starvla][INFO] - Step 5928, Loss: {'train/loss': 0.564677894115448, 'train/grad_norm_pre_clip': 15.072239875793457, 'timing/data': 0.0049844939494505525, 'timing/model': 1.3347015799954534, 'learning_rate/qwen_vl_interface': 1.054645944510656e-06, 'learning_rate/action_model': 1.054645944510656e-05, 'epoch': 1.74}) +[2026-06-17 17:17:19,358][starVLA.training.train_starvla][INFO] - Step 5929, Loss: {'train/loss': 0.40370920300483704, 'train/grad_norm_pre_clip': 14.370869636535645, 'timing/data': 0.005121793947182596, 'timing/model': 1.211096263024956, 'learning_rate/qwen_vl_interface': 1.0536322170779605e-06, 'learning_rate/action_model': 1.0536322170779606e-05, 'epoch': 1.74}) +[2026-06-17 17:17:24,147][starVLA.training.train_starvla][INFO] - Step 5930, Loss: {'train/loss': 0.660449206829071, 'train/grad_norm_pre_clip': 16.9304256439209, 'timing/data': 0.005445356015115976, 'timing/model': 1.2042473399778828, 'learning_rate/qwen_vl_interface': 1.0526193595568203e-06, 'learning_rate/action_model': 1.0526193595568203e-05, 'epoch': 1.74}) +[2026-06-17 17:17:28,922][starVLA.training.train_starvla][INFO] - Step 5931, Loss: {'train/loss': 0.8638203740119934, 'train/grad_norm_pre_clip': 17.010740280151367, 'timing/data': 0.005108718993142247, 'timing/model': 1.2005878670606762, 'learning_rate/qwen_vl_interface': 1.0516073721572043e-06, 'learning_rate/action_model': 1.0516073721572044e-05, 'epoch': 1.75}) +[2026-06-17 17:17:33,707][starVLA.training.train_starvla][INFO] - Step 5932, Loss: {'train/loss': 0.8166042566299438, 'train/grad_norm_pre_clip': 20.366670608520508, 'timing/data': 0.0050395100843161345, 'timing/model': 1.2008900770451874, 'learning_rate/qwen_vl_interface': 1.0505962550888982e-06, 'learning_rate/action_model': 1.0505962550888982e-05, 'epoch': 1.75}) +[2026-06-17 17:17:38,371][starVLA.training.train_starvla][INFO] - Step 5933, Loss: {'train/loss': 0.4604773223400116, 'train/grad_norm_pre_clip': 13.804303169250488, 'timing/data': 0.005183501052670181, 'timing/model': 1.2039740369655192, 'learning_rate/qwen_vl_interface': 1.0495860085615065e-06, 'learning_rate/action_model': 1.0495860085615064e-05, 'epoch': 1.75}) +[2026-06-17 17:17:43,141][starVLA.training.train_starvla][INFO] - Step 5934, Loss: {'train/loss': 0.8891995549201965, 'train/grad_norm_pre_clip': 21.01338005065918, 'timing/data': 0.00545116001740098, 'timing/model': 1.2505754620069638, 'learning_rate/qwen_vl_interface': 1.0485766327844558e-06, 'learning_rate/action_model': 1.0485766327844557e-05, 'epoch': 1.75}) +[2026-06-17 17:17:47,804][starVLA.training.train_starvla][INFO] - Step 5935, Loss: {'train/loss': 0.3910837173461914, 'train/grad_norm_pre_clip': 10.22842788696289, 'timing/data': 0.005349861923605204, 'timing/model': 1.2092408179305494, 'learning_rate/qwen_vl_interface': 1.0475681279669892e-06, 'learning_rate/action_model': 1.0475681279669891e-05, 'epoch': 1.75}) +[2026-06-17 17:17:52,545][starVLA.training.train_starvla][INFO] - Step 5936, Loss: {'train/loss': 0.5143236517906189, 'train/grad_norm_pre_clip': 13.098014831542969, 'timing/data': 0.005441404995508492, 'timing/model': 1.203968237969093, 'learning_rate/qwen_vl_interface': 1.0465604943181717e-06, 'learning_rate/action_model': 1.0465604943181717e-05, 'epoch': 1.75}) +[2026-06-17 17:17:57,337][starVLA.training.train_starvla][INFO] - Step 5937, Loss: {'train/loss': 0.4623599052429199, 'train/grad_norm_pre_clip': 11.51426887512207, 'timing/data': 0.005206406000070274, 'timing/model': 1.2072149500017986, 'learning_rate/qwen_vl_interface': 1.0455537320468861e-06, 'learning_rate/action_model': 1.045553732046886e-05, 'epoch': 1.75}) +[2026-06-17 17:18:02,132][starVLA.training.train_starvla][INFO] - Step 5938, Loss: {'train/loss': 0.48506593704223633, 'train/grad_norm_pre_clip': 12.486878395080566, 'timing/data': 0.005900243064388633, 'timing/model': 1.2064496589591727, 'learning_rate/qwen_vl_interface': 1.0445478413618382e-06, 'learning_rate/action_model': 1.0445478413618382e-05, 'epoch': 1.75}) +[2026-06-17 17:18:06,913][starVLA.training.train_starvla][INFO] - Step 5939, Loss: {'train/loss': 0.6422587633132935, 'train/grad_norm_pre_clip': 17.42782974243164, 'timing/data': 0.005462074070237577, 'timing/model': 1.3303107280516997, 'learning_rate/qwen_vl_interface': 1.0435428224715472e-06, 'learning_rate/action_model': 1.0435428224715471e-05, 'epoch': 1.75}) +[2026-06-17 17:18:11,688][starVLA.training.train_starvla][INFO] - Step 5940, Loss: {'train/loss': 0.4945072829723358, 'train/grad_norm_pre_clip': 12.867630004882812, 'timing/data': 0.005177087965421379, 'timing/model': 1.3254771620268002, 'learning_rate/qwen_vl_interface': 1.0425386755843547e-06, 'learning_rate/action_model': 1.0425386755843547e-05, 'epoch': 1.75}) +[2026-06-17 17:18:16,345][starVLA.training.train_starvla][INFO] - Step 5941, Loss: {'train/loss': 0.6322516798973083, 'train/grad_norm_pre_clip': 32.28232955932617, 'timing/data': 0.005303250043652952, 'timing/model': 1.204232541960664, 'learning_rate/qwen_vl_interface': 1.0415354009084219e-06, 'learning_rate/action_model': 1.041535400908422e-05, 'epoch': 1.75}) +[2026-06-17 17:18:21,101][starVLA.training.train_starvla][INFO] - Step 5942, Loss: {'train/loss': 0.7059602737426758, 'train/grad_norm_pre_clip': 15.440313339233398, 'timing/data': 0.005487169022671878, 'timing/model': 1.2062764230649918, 'learning_rate/qwen_vl_interface': 1.0405329986517301e-06, 'learning_rate/action_model': 1.0405329986517301e-05, 'epoch': 1.75}) +[2026-06-17 17:18:25,908][starVLA.training.train_starvla][INFO] - Step 5943, Loss: {'train/loss': 0.7000365853309631, 'train/grad_norm_pre_clip': 17.587488174438477, 'timing/data': 0.005025596008636057, 'timing/model': 1.2040737870847806, 'learning_rate/qwen_vl_interface': 1.0395314690220776e-06, 'learning_rate/action_model': 1.0395314690220776e-05, 'epoch': 1.75}) +[2026-06-17 17:18:30,694][starVLA.training.train_starvla][INFO] - Step 5944, Loss: {'train/loss': 0.9200252294540405, 'train/grad_norm_pre_clip': 16.027183532714844, 'timing/data': 0.00543842907063663, 'timing/model': 1.2004062390187755, 'learning_rate/qwen_vl_interface': 1.0385308122270815e-06, 'learning_rate/action_model': 1.0385308122270814e-05, 'epoch': 1.75}) +[2026-06-17 17:18:35,479][starVLA.training.train_starvla][INFO] - Step 5945, Loss: {'train/loss': 0.6715542078018188, 'train/grad_norm_pre_clip': 15.891338348388672, 'timing/data': 0.0055790189653635025, 'timing/model': 1.3349021209869534, 'learning_rate/qwen_vl_interface': 1.0375310284741787e-06, 'learning_rate/action_model': 1.0375310284741788e-05, 'epoch': 1.75}) +[2026-06-17 17:18:40,135][starVLA.training.train_starvla][INFO] - Step 5946, Loss: {'train/loss': 0.5920411348342896, 'train/grad_norm_pre_clip': 15.267168998718262, 'timing/data': 0.0054767889669165015, 'timing/model': 1.204574548988603, 'learning_rate/qwen_vl_interface': 1.0365321179706276e-06, 'learning_rate/action_model': 1.0365321179706276e-05, 'epoch': 1.75}) +[2026-06-17 17:18:44,924][starVLA.training.train_starvla][INFO] - Step 5947, Loss: {'train/loss': 0.5598584413528442, 'train/grad_norm_pre_clip': 14.677034378051758, 'timing/data': 0.005405183997936547, 'timing/model': 1.2051555520156398, 'learning_rate/qwen_vl_interface': 1.035534080923503e-06, 'learning_rate/action_model': 1.0355340809235029e-05, 'epoch': 1.75}) +[2026-06-17 17:18:49,710][starVLA.training.train_starvla][INFO] - Step 5948, Loss: {'train/loss': 0.9073013663291931, 'train/grad_norm_pre_clip': 18.33182144165039, 'timing/data': 0.005227594985626638, 'timing/model': 1.2067918800748885, 'learning_rate/qwen_vl_interface': 1.0345369175396972e-06, 'learning_rate/action_model': 1.0345369175396972e-05, 'epoch': 1.75}) +[2026-06-17 17:18:54,492][starVLA.training.train_starvla][INFO] - Step 5949, Loss: {'train/loss': 0.7004219889640808, 'train/grad_norm_pre_clip': 14.111347198486328, 'timing/data': 0.005144022987224162, 'timing/model': 1.2030488019809127, 'learning_rate/qwen_vl_interface': 1.0335406280259255e-06, 'learning_rate/action_model': 1.0335406280259254e-05, 'epoch': 1.75}) +[2026-06-17 17:18:59,277][starVLA.training.train_starvla][INFO] - Step 5950, Loss: {'train/loss': 0.46013569831848145, 'train/grad_norm_pre_clip': 12.69262981414795, 'timing/data': 0.005103548988699913, 'timing/model': 1.332066974020563, 'learning_rate/qwen_vl_interface': 1.0325452125887177e-06, 'learning_rate/action_model': 1.0325452125887178e-05, 'epoch': 1.75}) +[2026-06-17 17:19:03,931][starVLA.training.train_starvla][INFO] - Step 5951, Loss: {'train/loss': 0.5503500699996948, 'train/grad_norm_pre_clip': 15.050140380859375, 'timing/data': 0.005278661963529885, 'timing/model': 1.20451155398041, 'learning_rate/qwen_vl_interface': 1.031550671434426e-06, 'learning_rate/action_model': 1.031550671434426e-05, 'epoch': 1.75}) +[2026-06-17 17:19:08,715][starVLA.training.train_starvla][INFO] - Step 5952, Loss: {'train/loss': 0.5133984088897705, 'train/grad_norm_pre_clip': 13.668269157409668, 'timing/data': 0.005009364918805659, 'timing/model': 1.207559091039002, 'learning_rate/qwen_vl_interface': 1.0305570047692197e-06, 'learning_rate/action_model': 1.0305570047692195e-05, 'epoch': 1.75}) +[2026-06-17 17:19:13,471][starVLA.training.train_starvla][INFO] - Step 5953, Loss: {'train/loss': 0.5731061697006226, 'train/grad_norm_pre_clip': 23.314584732055664, 'timing/data': 0.005339156021364033, 'timing/model': 1.2040954750264063, 'learning_rate/qwen_vl_interface': 1.0295642127990863e-06, 'learning_rate/action_model': 1.0295642127990863e-05, 'epoch': 1.75}) +[2026-06-17 17:19:18,315][starVLA.training.train_starvla][INFO] - Step 5954, Loss: {'train/loss': 0.7375407218933105, 'train/grad_norm_pre_clip': 17.506898880004883, 'timing/data': 0.0052376569947227836, 'timing/model': 1.2042753449641168, 'learning_rate/qwen_vl_interface': 1.0285722957298326e-06, 'learning_rate/action_model': 1.0285722957298326e-05, 'epoch': 1.75}) +[2026-06-17 17:19:23,096][starVLA.training.train_starvla][INFO] - Step 5955, Loss: {'train/loss': 0.46041736006736755, 'train/grad_norm_pre_clip': 12.158759117126465, 'timing/data': 0.00519883306697011, 'timing/model': 1.2044633250916377, 'learning_rate/qwen_vl_interface': 1.027581253767085e-06, 'learning_rate/action_model': 1.027581253767085e-05, 'epoch': 1.75}) +[2026-06-17 17:19:28,102][starVLA.training.train_starvla][INFO] - Step 5956, Loss: {'train/loss': 0.527193009853363, 'train/grad_norm_pre_clip': 15.227278709411621, 'timing/data': 0.004979133955202997, 'timing/model': 1.3429693100042641, 'learning_rate/qwen_vl_interface': 1.0265910871162858e-06, 'learning_rate/action_model': 1.0265910871162858e-05, 'epoch': 1.75}) +[2026-06-17 17:19:32,837][starVLA.training.train_starvla][INFO] - Step 5957, Loss: {'train/loss': 0.5884162187576294, 'train/grad_norm_pre_clip': 13.93990421295166, 'timing/data': 0.006003884947858751, 'timing/model': 1.2801961479708552, 'learning_rate/qwen_vl_interface': 1.0256017959827008e-06, 'learning_rate/action_model': 1.0256017959827008e-05, 'epoch': 1.75}) +[2026-06-17 17:19:37,505][starVLA.training.train_starvla][INFO] - Step 5958, Loss: {'train/loss': 0.6084637641906738, 'train/grad_norm_pre_clip': 14.809590339660645, 'timing/data': 0.005196276004426181, 'timing/model': 1.2054174559889361, 'learning_rate/qwen_vl_interface': 1.0246133805714084e-06, 'learning_rate/action_model': 1.0246133805714084e-05, 'epoch': 1.75}) +[2026-06-17 17:19:42,293][starVLA.training.train_starvla][INFO] - Step 5959, Loss: {'train/loss': 0.4888981580734253, 'train/grad_norm_pre_clip': 10.000106811523438, 'timing/data': 0.0048829339211806655, 'timing/model': 1.2054641039576381, 'learning_rate/qwen_vl_interface': 1.0236258410873076e-06, 'learning_rate/action_model': 1.0236258410873076e-05, 'epoch': 1.75}) +[2026-06-17 17:19:47,079][starVLA.training.train_starvla][INFO] - Step 5960, Loss: {'train/loss': 0.7799248099327087, 'train/grad_norm_pre_clip': 16.782882690429688, 'timing/data': 0.0050481329672038555, 'timing/model': 1.2038102380465716, 'learning_rate/qwen_vl_interface': 1.0226391777351188e-06, 'learning_rate/action_model': 1.0226391777351189e-05, 'epoch': 1.75}) +[2026-06-17 17:19:51,863][starVLA.training.train_starvla][INFO] - Step 5961, Loss: {'train/loss': 1.10515558719635, 'train/grad_norm_pre_clip': 22.157779693603516, 'timing/data': 0.0054418540094047785, 'timing/model': 1.2035913249710575, 'learning_rate/qwen_vl_interface': 1.0216533907193777e-06, 'learning_rate/action_model': 1.0216533907193776e-05, 'epoch': 1.75}) +[2026-06-17 17:19:56,638][starVLA.training.train_starvla][INFO] - Step 5962, Loss: {'train/loss': 0.9361382722854614, 'train/grad_norm_pre_clip': 18.45621681213379, 'timing/data': 0.00527250103186816, 'timing/model': 1.2030923690181226, 'learning_rate/qwen_vl_interface': 1.020668480244438e-06, 'learning_rate/action_model': 1.020668480244438e-05, 'epoch': 1.75}) +[2026-06-17 17:20:01,384][starVLA.training.train_starvla][INFO] - Step 5963, Loss: {'train/loss': 0.5322988629341125, 'train/grad_norm_pre_clip': 14.348973274230957, 'timing/data': 0.005410499987192452, 'timing/model': 1.2952546619344503, 'learning_rate/qwen_vl_interface': 1.0196844465144735e-06, 'learning_rate/action_model': 1.0196844465144734e-05, 'epoch': 1.75}) +[2026-06-17 17:20:06,043][starVLA.training.train_starvla][INFO] - Step 5964, Loss: {'train/loss': 1.0730865001678467, 'train/grad_norm_pre_clip': 16.483333587646484, 'timing/data': 0.0054437280632555485, 'timing/model': 1.205582113005221, 'learning_rate/qwen_vl_interface': 1.0187012897334755e-06, 'learning_rate/action_model': 1.0187012897334755e-05, 'epoch': 1.75}) +[2026-06-17 17:20:10,833][starVLA.training.train_starvla][INFO] - Step 5965, Loss: {'train/loss': 0.752478837966919, 'train/grad_norm_pre_clip': 15.683363914489746, 'timing/data': 0.005500420928001404, 'timing/model': 1.2068333979696035, 'learning_rate/qwen_vl_interface': 1.017719010105253e-06, 'learning_rate/action_model': 1.017719010105253e-05, 'epoch': 1.76}) +[2026-06-17 17:20:15,623][starVLA.training.train_starvla][INFO] - Step 5966, Loss: {'train/loss': 0.4272313416004181, 'train/grad_norm_pre_clip': 11.556844711303711, 'timing/data': 0.004940776037983596, 'timing/model': 1.2040280210785568, 'learning_rate/qwen_vl_interface': 1.0167376078334337e-06, 'learning_rate/action_model': 1.0167376078334337e-05, 'epoch': 1.76}) +[2026-06-17 17:20:20,409][starVLA.training.train_starvla][INFO] - Step 5967, Loss: {'train/loss': 0.47868040204048157, 'train/grad_norm_pre_clip': 12.161832809448242, 'timing/data': 0.005326556973159313, 'timing/model': 1.2040091099916026, 'learning_rate/qwen_vl_interface': 1.0157570831214634e-06, 'learning_rate/action_model': 1.0157570831214634e-05, 'epoch': 1.76}) +[2026-06-17 17:20:25,272][starVLA.training.train_starvla][INFO] - Step 5968, Loss: {'train/loss': 0.5587652921676636, 'train/grad_norm_pre_clip': 13.806220054626465, 'timing/data': 0.005255953990854323, 'timing/model': 1.3347993739880621, 'learning_rate/qwen_vl_interface': 1.0147774361726066e-06, 'learning_rate/action_model': 1.0147774361726064e-05, 'epoch': 1.76}) +[2026-06-17 17:20:30,011][starVLA.training.train_starvla][INFO] - Step 5969, Loss: {'train/loss': 0.8587146401405334, 'train/grad_norm_pre_clip': 15.206025123596191, 'timing/data': 0.005224228953011334, 'timing/model': 1.2040842630667612, 'learning_rate/qwen_vl_interface': 1.013798667189944e-06, 'learning_rate/action_model': 1.013798667189944e-05, 'epoch': 1.76}) +[2026-06-17 17:20:34,801][starVLA.training.train_starvla][INFO] - Step 5970, Loss: {'train/loss': 0.7002338171005249, 'train/grad_norm_pre_clip': 16.756677627563477, 'timing/data': 0.0052782350685447454, 'timing/model': 1.2086562790209427, 'learning_rate/qwen_vl_interface': 1.0128207763763755e-06, 'learning_rate/action_model': 1.0128207763763755e-05, 'epoch': 1.76}) +[2026-06-17 17:20:39,615][starVLA.training.train_starvla][INFO] - Step 5971, Loss: {'train/loss': 0.5027500987052917, 'train/grad_norm_pre_clip': 11.485723495483398, 'timing/data': 0.005303817102685571, 'timing/model': 1.2134941909462214, 'learning_rate/qwen_vl_interface': 1.0118437639346204e-06, 'learning_rate/action_model': 1.0118437639346204e-05, 'epoch': 1.76}) +[2026-06-17 17:20:44,333][starVLA.training.train_starvla][INFO] - Step 5972, Loss: {'train/loss': 0.6163535714149475, 'train/grad_norm_pre_clip': 15.282928466796875, 'timing/data': 0.004845303017646074, 'timing/model': 1.2094034489709884, 'learning_rate/qwen_vl_interface': 1.0108676300672144e-06, 'learning_rate/action_model': 1.0108676300672144e-05, 'epoch': 1.76}) +[2026-06-17 17:20:49,082][starVLA.training.train_starvla][INFO] - Step 5973, Loss: {'train/loss': 0.7792623043060303, 'train/grad_norm_pre_clip': 13.632214546203613, 'timing/data': 0.0056505349930375814, 'timing/model': 1.2065472119720653, 'learning_rate/qwen_vl_interface': 1.0098923749765086e-06, 'learning_rate/action_model': 1.0098923749765086e-05, 'epoch': 1.76}) +[2026-06-17 17:20:53,898][starVLA.training.train_starvla][INFO] - Step 5974, Loss: {'train/loss': 0.6261048316955566, 'train/grad_norm_pre_clip': 16.999391555786133, 'timing/data': 0.005460162996314466, 'timing/model': 1.3467532079666853, 'learning_rate/qwen_vl_interface': 1.008917998864676e-06, 'learning_rate/action_model': 1.008917998864676e-05, 'epoch': 1.76}) +[2026-06-17 17:20:58,587][starVLA.training.train_starvla][INFO] - Step 5975, Loss: {'train/loss': 0.6455054879188538, 'train/grad_norm_pre_clip': 14.742962837219238, 'timing/data': 0.005418202956207097, 'timing/model': 1.2132519739679992, 'learning_rate/qwen_vl_interface': 1.0079445019337063e-06, 'learning_rate/action_model': 1.0079445019337063e-05, 'epoch': 1.76}) +[2026-06-17 17:21:03,392][starVLA.training.train_starvla][INFO] - Step 5976, Loss: {'train/loss': 0.8667418360710144, 'train/grad_norm_pre_clip': 19.665790557861328, 'timing/data': 0.0052776309894397855, 'timing/model': 1.2112163229612634, 'learning_rate/qwen_vl_interface': 1.0069718843854058e-06, 'learning_rate/action_model': 1.0069718843854058e-05, 'epoch': 1.76}) +[2026-06-17 17:21:08,186][starVLA.training.train_starvla][INFO] - Step 5977, Loss: {'train/loss': 0.6557015180587769, 'train/grad_norm_pre_clip': 13.645695686340332, 'timing/data': 0.0053014010190963745, 'timing/model': 1.207099056104198, 'learning_rate/qwen_vl_interface': 1.0060001464214007e-06, 'learning_rate/action_model': 1.0060001464214006e-05, 'epoch': 1.76}) +[2026-06-17 17:21:13,002][starVLA.training.train_starvla][INFO] - Step 5978, Loss: {'train/loss': 0.6037045121192932, 'train/grad_norm_pre_clip': 12.613039016723633, 'timing/data': 0.005172549048438668, 'timing/model': 1.2109259340213612, 'learning_rate/qwen_vl_interface': 1.00502928824313e-06, 'learning_rate/action_model': 1.00502928824313e-05, 'epoch': 1.76}) +[2026-06-17 17:21:17,755][starVLA.training.train_starvla][INFO] - Step 5979, Loss: {'train/loss': 0.8890650272369385, 'train/grad_norm_pre_clip': 14.713727951049805, 'timing/data': 0.005495569086633623, 'timing/model': 1.208392672939226, 'learning_rate/qwen_vl_interface': 1.0040593100518567e-06, 'learning_rate/action_model': 1.0040593100518566e-05, 'epoch': 1.76}) +[2026-06-17 17:21:22,543][starVLA.training.train_starvla][INFO] - Step 5980, Loss: {'train/loss': 0.4798886775970459, 'train/grad_norm_pre_clip': 13.974808692932129, 'timing/data': 0.00532322796061635, 'timing/model': 1.3244144939817488, 'learning_rate/qwen_vl_interface': 1.003090212048657e-06, 'learning_rate/action_model': 1.003090212048657e-05, 'epoch': 1.76}) +[2026-06-17 17:21:27,220][starVLA.training.train_starvla][INFO] - Step 5981, Loss: {'train/loss': 0.8543571829795837, 'train/grad_norm_pre_clip': 15.093439102172852, 'timing/data': 0.005590924993157387, 'timing/model': 1.2087484590010718, 'learning_rate/qwen_vl_interface': 1.0021219944344263e-06, 'learning_rate/action_model': 1.0021219944344262e-05, 'epoch': 1.76}) +[2026-06-17 17:21:32,035][starVLA.training.train_starvla][INFO] - Step 5982, Loss: {'train/loss': 0.6040666103363037, 'train/grad_norm_pre_clip': 16.751855850219727, 'timing/data': 0.0053000570042058825, 'timing/model': 1.210461338981986, 'learning_rate/qwen_vl_interface': 1.0011546574098773e-06, 'learning_rate/action_model': 1.0011546574098772e-05, 'epoch': 1.76}) +[2026-06-17 17:21:36,832][starVLA.training.train_starvla][INFO] - Step 5983, Loss: {'train/loss': 0.4904787242412567, 'train/grad_norm_pre_clip': 11.695876121520996, 'timing/data': 0.005277583957649767, 'timing/model': 1.2118959189392626, 'learning_rate/qwen_vl_interface': 1.0001882011755396e-06, 'learning_rate/action_model': 1.0001882011755395e-05, 'epoch': 1.76}) +[2026-06-17 17:21:41,661][starVLA.training.train_starvla][INFO] - Step 5984, Loss: {'train/loss': 0.6660436987876892, 'train/grad_norm_pre_clip': 16.99889373779297, 'timing/data': 0.005316609982401133, 'timing/model': 1.2140173140214756, 'learning_rate/qwen_vl_interface': 9.992226259317605e-07, 'learning_rate/action_model': 9.992226259317605e-06, 'epoch': 1.76}) +[2026-06-17 17:21:46,452][starVLA.training.train_starvla][INFO] - Step 5985, Loss: {'train/loss': 0.6847003698348999, 'train/grad_norm_pre_clip': 12.79049301147461, 'timing/data': 0.005126064992509782, 'timing/model': 1.3360576069680974, 'learning_rate/qwen_vl_interface': 9.982579318787044e-07, 'learning_rate/action_model': 9.982579318787044e-06, 'epoch': 1.76}) +[2026-06-17 17:21:51,132][starVLA.training.train_starvla][INFO] - Step 5986, Loss: {'train/loss': 0.4748380482196808, 'train/grad_norm_pre_clip': 13.015674591064453, 'timing/data': 0.005797295016236603, 'timing/model': 1.2154404709581286, 'learning_rate/qwen_vl_interface': 9.97294119216356e-07, 'learning_rate/action_model': 9.97294119216356e-06, 'epoch': 1.76}) +[2026-06-17 17:21:55,968][starVLA.training.train_starvla][INFO] - Step 5987, Loss: {'train/loss': 0.5777395367622375, 'train/grad_norm_pre_clip': 13.113814353942871, 'timing/data': 0.005441086948849261, 'timing/model': 1.202207648078911, 'learning_rate/qwen_vl_interface': 9.96331188144512e-07, 'learning_rate/action_model': 9.96331188144512e-06, 'epoch': 1.76}) +[2026-06-17 17:22:00,672][starVLA.training.train_starvla][INFO] - Step 5988, Loss: {'train/loss': 0.40259888768196106, 'train/grad_norm_pre_clip': 11.954081535339355, 'timing/data': 0.005737567902542651, 'timing/model': 1.2061924079898745, 'learning_rate/qwen_vl_interface': 9.953691388627887e-07, 'learning_rate/action_model': 9.953691388627886e-06, 'epoch': 1.76}) +[2026-06-17 17:22:05,404][starVLA.training.train_starvla][INFO] - Step 5989, Loss: {'train/loss': 0.6586853861808777, 'train/grad_norm_pre_clip': 15.018604278564453, 'timing/data': 0.005592313013039529, 'timing/model': 1.2035711250500754, 'learning_rate/qwen_vl_interface': 9.94407971570622e-07, 'learning_rate/action_model': 9.94407971570622e-06, 'epoch': 1.76}) +[2026-06-17 17:22:10,193][starVLA.training.train_starvla][INFO] - Step 5990, Loss: {'train/loss': 0.6399062871932983, 'train/grad_norm_pre_clip': 667.8654174804688, 'timing/data': 0.0055408329935744405, 'timing/model': 1.208750582067296, 'learning_rate/qwen_vl_interface': 9.934476864672618e-07, 'learning_rate/action_model': 9.934476864672617e-06, 'epoch': 1.76}) +[2026-06-17 17:22:15,020][starVLA.training.train_starvla][INFO] - Step 5991, Loss: {'train/loss': 0.4649364948272705, 'train/grad_norm_pre_clip': 12.292546272277832, 'timing/data': 0.0070569380186498165, 'timing/model': 1.3634220430394635, 'learning_rate/qwen_vl_interface': 9.924882837517781e-07, 'learning_rate/action_model': 9.924882837517782e-06, 'epoch': 1.76}) +[2026-06-17 17:22:19,683][starVLA.training.train_starvla][INFO] - Step 5992, Loss: {'train/loss': 0.6136883497238159, 'train/grad_norm_pre_clip': 15.966015815734863, 'timing/data': 0.0053587889997288585, 'timing/model': 1.204996209940873, 'learning_rate/qwen_vl_interface': 9.91529763623052e-07, 'learning_rate/action_model': 9.91529763623052e-06, 'epoch': 1.76}) +[2026-06-17 17:22:24,472][starVLA.training.train_starvla][INFO] - Step 5993, Loss: {'train/loss': 0.5136796832084656, 'train/grad_norm_pre_clip': 12.29952621459961, 'timing/data': 0.005263836006633937, 'timing/model': 1.206421596929431, 'learning_rate/qwen_vl_interface': 9.905721262797894e-07, 'learning_rate/action_model': 9.905721262797892e-06, 'epoch': 1.76}) +[2026-06-17 17:22:29,299][starVLA.training.train_starvla][INFO] - Step 5994, Loss: {'train/loss': 0.7625579833984375, 'train/grad_norm_pre_clip': 20.4925594329834, 'timing/data': 0.005421322071924806, 'timing/model': 1.2051181270508096, 'learning_rate/qwen_vl_interface': 9.896153719205082e-07, 'learning_rate/action_model': 9.89615371920508e-06, 'epoch': 1.76}) +[2026-06-17 17:22:34,035][starVLA.training.train_starvla][INFO] - Step 5995, Loss: {'train/loss': 0.7523412108421326, 'train/grad_norm_pre_clip': 13.678927421569824, 'timing/data': 0.00532966002356261, 'timing/model': 1.2037272830493748, 'learning_rate/qwen_vl_interface': 9.886595007435448e-07, 'learning_rate/action_model': 9.886595007435447e-06, 'epoch': 1.76}) +[2026-06-17 17:22:38,814][starVLA.training.train_starvla][INFO] - Step 5996, Loss: {'train/loss': 0.6327498555183411, 'train/grad_norm_pre_clip': 17.051549911499023, 'timing/data': 0.005980332964099944, 'timing/model': 1.2036753809079528, 'learning_rate/qwen_vl_interface': 9.877045129470524e-07, 'learning_rate/action_model': 9.877045129470524e-06, 'epoch': 1.76}) +[2026-06-17 17:22:43,605][starVLA.training.train_starvla][INFO] - Step 5997, Loss: {'train/loss': 0.8317804932594299, 'train/grad_norm_pre_clip': 17.973369598388672, 'timing/data': 0.005280288984067738, 'timing/model': 1.334086160059087, 'learning_rate/qwen_vl_interface': 9.86750408729001e-07, 'learning_rate/action_model': 9.867504087290009e-06, 'epoch': 1.76}) +[2026-06-17 17:22:48,280][starVLA.training.train_starvla][INFO] - Step 5998, Loss: {'train/loss': 0.7248509526252747, 'train/grad_norm_pre_clip': 16.274646759033203, 'timing/data': 0.0049536110600456595, 'timing/model': 1.2151799460407346, 'learning_rate/qwen_vl_interface': 9.857971882871773e-07, 'learning_rate/action_model': 9.857971882871772e-06, 'epoch': 1.76}) +[2026-06-17 17:22:53,126][starVLA.training.train_starvla][INFO] - Step 5999, Loss: {'train/loss': 0.4076463580131531, 'train/grad_norm_pre_clip': 12.21861743927002, 'timing/data': 0.0050595649518072605, 'timing/model': 1.204949166974984, 'learning_rate/qwen_vl_interface': 9.848448518191838e-07, 'learning_rate/action_model': 9.848448518191836e-06, 'epoch': 1.77}) +[2026-06-17 17:31:23,184][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 6000 +[2026-06-17 17:31:23,185][starVLA.training.train_starvla][INFO] - Step 6000, Loss: {'train/loss': 0.20191751420497894, 'train/grad_norm_pre_clip': 7.6211628913879395, 'eval/loss': 0.7047950625419617, 'eval/latency_0/loss': 0.7047950625419617, 'eval/demon_fire_precision': 0.9075425790754258, 'eval/demon_fire_recall': 0.9294019933554817, 'eval/demon_fire_f1': 0.9183422240459581, 'eval/demon_fire_rate_model': 0.685761957730812, 'eval/demon_fire_rate_teacher': 0.6696329254727474, 'eval/demon_left_precision': 0.8341094295692666, 'eval/demon_left_recall': 0.8118980169971671, 'eval/demon_left_f1': 0.8228538616135517, 'eval/demon_left_rate_model': 0.4777530589543938, 'eval/demon_left_rate_teacher': 0.49082313681868744, 'eval/demon_right_precision': 0.8183795541054921, 'eval/demon_right_recall': 0.8502824858757062, 'eval/demon_right_f1': 0.8340260459961208, 'eval/demon_right_rate_model': 0.5114015572858732, 'eval/demon_right_rate_teacher': 0.4922135706340378, 'eval/demon_move_f1': 0.8284399538048363, 'eval/demon_cc_f1': 0.8893873405898269, 'eval/latency_0/demon_fire_precision': 0.9075425790754258, 'eval/latency_0/demon_fire_recall': 0.9294019933554817, 'eval/latency_0/demon_fire_f1': 0.9183422240459581, 'eval/latency_0/demon_fire_rate_model': 0.685761957730812, 'eval/latency_0/demon_fire_rate_teacher': 0.6696329254727474, 'eval/latency_0/demon_left_precision': 0.8341094295692666, 'eval/latency_0/demon_left_recall': 0.8118980169971671, 'eval/latency_0/demon_left_f1': 0.8228538616135517, 'eval/latency_0/demon_left_rate_model': 0.4777530589543938, 'eval/latency_0/demon_left_rate_teacher': 0.49082313681868744, 'eval/latency_0/demon_right_precision': 0.8183795541054921, 'eval/latency_0/demon_right_recall': 0.8502824858757062, 'eval/latency_0/demon_right_f1': 0.8340260459961208, 'eval/latency_0/demon_right_rate_model': 0.5114015572858732, 'eval/latency_0/demon_right_rate_teacher': 0.4922135706340378, 'eval/latency_0/demon_move_f1': 0.8284399538048363, 'eval/latency_0/demon_cc_f1': 0.8893873405898269, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 5040.0, 'rl_games_eval/mid_train/mean_length': 2894.4, 'rl_games_eval/mid_train/std_reward': 2404.7868928451853, 'rl_games_eval/mid_train/std_length': 864.1802126871455, 'rl_games_eval/mid_train/macro_mean_reward': 5040.0, 'rl_games_eval/mid_train/macro_mean_length': 2894.4, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/demon_attack__latency_0/mean_reward': 5040.0, 'rl_games_eval/mid_train/demon_attack__latency_0/mean_length': 2894.4, 'rl_games_eval/mid_train/demon_attack__latency_0/std_reward': 2404.7868928451853, 'rl_games_eval/mid_train/demon_attack__latency_0/std_length': 864.1802126871455, 'rl_games_eval/mid_train/demon_attack__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 5040.0, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.005174160003662109, 'timing/model': 1.2032873480347916, 'learning_rate/qwen_vl_interface': 9.838933995224439e-07, 'learning_rate/action_model': 9.83893399522444e-06, 'epoch': 1.77}) +[2026-06-17 17:31:23,185][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_6000_state +[2026-06-17 17:31:23,191][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 17:31:26,754][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_6000_state/model.safetensors +[2026-06-17 17:31:34,381][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_6000_state/optimizer.bin +[2026-06-17 17:31:34,381][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_6000_state/sampler.bin +[2026-06-17 17:31:34,381][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_6000_state/sampler_1.bin +[2026-06-17 17:31:34,382][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_6000_state/random_states_0.pkl +[2026-06-17 17:31:34,436][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 6000 +[2026-06-17 17:31:34,436][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 17:31:34,452][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 17:31:39,170][starVLA.training.train_starvla][INFO] - Step 6001, Loss: {'train/loss': 0.8299462199211121, 'train/grad_norm_pre_clip': 74.78465270996094, 'timing/data': 0.005545741063542664, 'timing/model': 1.3264817749150097, 'learning_rate/qwen_vl_interface': 9.829428315941921e-07, 'learning_rate/action_model': 9.829428315941922e-06, 'epoch': 1.77}) +[2026-06-17 17:31:43,757][starVLA.training.train_starvla][INFO] - Step 6002, Loss: {'train/loss': 0.8177591562271118, 'train/grad_norm_pre_clip': 14.223387718200684, 'timing/data': 0.005067802965641022, 'timing/model': 1.188264488009736, 'learning_rate/qwen_vl_interface': 9.819931482314823e-07, 'learning_rate/action_model': 9.819931482314823e-06, 'epoch': 1.77}) +[2026-06-17 17:31:48,491][starVLA.training.train_starvla][INFO] - Step 6003, Loss: {'train/loss': 0.4157457947731018, 'train/grad_norm_pre_clip': 11.175421714782715, 'timing/data': 0.005155425053089857, 'timing/model': 1.1897392189130187, 'learning_rate/qwen_vl_interface': 9.81044349631185e-07, 'learning_rate/action_model': 9.81044349631185e-06, 'epoch': 1.77}) +[2026-06-17 17:31:53,224][starVLA.training.train_starvla][INFO] - Step 6004, Loss: {'train/loss': 0.597840428352356, 'train/grad_norm_pre_clip': 10.04085922241211, 'timing/data': 0.004944585030898452, 'timing/model': 1.1919192440109327, 'learning_rate/qwen_vl_interface': 9.800964359899886e-07, 'learning_rate/action_model': 9.800964359899886e-06, 'epoch': 1.77}) +[2026-06-17 17:31:57,968][starVLA.training.train_starvla][INFO] - Step 6005, Loss: {'train/loss': 0.5209664106369019, 'train/grad_norm_pre_clip': 14.308411598205566, 'timing/data': 0.005163176101632416, 'timing/model': 1.1941440240480006, 'learning_rate/qwen_vl_interface': 9.791494075043963e-07, 'learning_rate/action_model': 9.791494075043963e-06, 'epoch': 1.77}) +[2026-06-17 17:32:02,721][starVLA.training.train_starvla][INFO] - Step 6006, Loss: {'train/loss': 0.731062650680542, 'train/grad_norm_pre_clip': 17.827287673950195, 'timing/data': 0.004903865046799183, 'timing/model': 1.3246871950104833, 'learning_rate/qwen_vl_interface': 9.78203264370727e-07, 'learning_rate/action_model': 9.78203264370727e-06, 'epoch': 1.77}) +[2026-06-17 17:32:07,446][starVLA.training.train_starvla][INFO] - Step 6007, Loss: {'train/loss': 0.5160541534423828, 'train/grad_norm_pre_clip': 13.548295021057129, 'timing/data': 0.004945595981553197, 'timing/model': 1.2991962729720399, 'learning_rate/qwen_vl_interface': 9.772580067851162e-07, 'learning_rate/action_model': 9.772580067851161e-06, 'epoch': 1.77}) +[2026-06-17 17:32:12,069][starVLA.training.train_starvla][INFO] - Step 6008, Loss: {'train/loss': 0.5111322999000549, 'train/grad_norm_pre_clip': 11.506664276123047, 'timing/data': 0.005234629032202065, 'timing/model': 1.1948366480646655, 'learning_rate/qwen_vl_interface': 9.763136349435195e-07, 'learning_rate/action_model': 9.763136349435194e-06, 'epoch': 1.77}) +[2026-06-17 17:32:16,822][starVLA.training.train_starvla][INFO] - Step 6009, Loss: {'train/loss': 0.822337806224823, 'train/grad_norm_pre_clip': 17.9658260345459, 'timing/data': 0.005127697950229049, 'timing/model': 1.1977429270045832, 'learning_rate/qwen_vl_interface': 9.753701490417048e-07, 'learning_rate/action_model': 9.753701490417047e-06, 'epoch': 1.77}) +[2026-06-17 17:32:21,571][starVLA.training.train_starvla][INFO] - Step 6010, Loss: {'train/loss': 0.565008819103241, 'train/grad_norm_pre_clip': 11.313386917114258, 'timing/data': 0.0052255039336159825, 'timing/model': 1.197036604047753, 'learning_rate/qwen_vl_interface': 9.744275492752584e-07, 'learning_rate/action_model': 9.744275492752583e-06, 'epoch': 1.77}) +[2026-06-17 17:32:26,330][starVLA.training.train_starvla][INFO] - Step 6011, Loss: {'train/loss': 0.2761320471763611, 'train/grad_norm_pre_clip': 9.238947868347168, 'timing/data': 0.005091649014502764, 'timing/model': 1.1976065899943933, 'learning_rate/qwen_vl_interface': 9.734858358395808e-07, 'learning_rate/action_model': 9.734858358395807e-06, 'epoch': 1.77}) +[2026-06-17 17:32:31,044][starVLA.training.train_starvla][INFO] - Step 6012, Loss: {'train/loss': 0.5154129266738892, 'train/grad_norm_pre_clip': 17.227195739746094, 'timing/data': 0.005162891931831837, 'timing/model': 1.2003862319979817, 'learning_rate/qwen_vl_interface': 9.72545008929892e-07, 'learning_rate/action_model': 9.725450089298921e-06, 'epoch': 1.77}) +[2026-06-17 17:32:35,806][starVLA.training.train_starvla][INFO] - Step 6013, Loss: {'train/loss': 0.6502958536148071, 'train/grad_norm_pre_clip': 16.425554275512695, 'timing/data': 0.004818464047275484, 'timing/model': 1.3304996970109642, 'learning_rate/qwen_vl_interface': 9.71605068741225e-07, 'learning_rate/action_model': 9.716050687412251e-06, 'epoch': 1.77}) +[2026-06-17 17:32:40,455][starVLA.training.train_starvla][INFO] - Step 6014, Loss: {'train/loss': 0.5809041857719421, 'train/grad_norm_pre_clip': 13.853486061096191, 'timing/data': 0.005026223952881992, 'timing/model': 1.2162362929666415, 'learning_rate/qwen_vl_interface': 9.706660154684312e-07, 'learning_rate/action_model': 9.706660154684312e-06, 'epoch': 1.77}) +[2026-06-17 17:32:45,216][starVLA.training.train_starvla][INFO] - Step 6015, Loss: {'train/loss': 0.5830705761909485, 'train/grad_norm_pre_clip': 15.147130966186523, 'timing/data': 0.005084729054942727, 'timing/model': 1.198789531015791, 'learning_rate/qwen_vl_interface': 9.697278493061773e-07, 'learning_rate/action_model': 9.697278493061773e-06, 'epoch': 1.77}) +[2026-06-17 17:32:49,997][starVLA.training.train_starvla][INFO] - Step 6016, Loss: {'train/loss': 0.6750952005386353, 'train/grad_norm_pre_clip': 15.043058395385742, 'timing/data': 0.0065758839482441545, 'timing/model': 1.199675824958831, 'learning_rate/qwen_vl_interface': 9.687905704489461e-07, 'learning_rate/action_model': 9.687905704489462e-06, 'epoch': 1.77}) +[2026-06-17 17:32:54,777][starVLA.training.train_starvla][INFO] - Step 6017, Loss: {'train/loss': 0.6938710808753967, 'train/grad_norm_pre_clip': 15.960142135620117, 'timing/data': 0.00474357302300632, 'timing/model': 1.2213414990110323, 'learning_rate/qwen_vl_interface': 9.678541790910361e-07, 'learning_rate/action_model': 9.678541790910361e-06, 'epoch': 1.77}) +[2026-06-17 17:32:59,516][starVLA.training.train_starvla][INFO] - Step 6018, Loss: {'train/loss': 0.578251838684082, 'train/grad_norm_pre_clip': 15.811683654785156, 'timing/data': 0.004959971061907709, 'timing/model': 1.1997540219454095, 'learning_rate/qwen_vl_interface': 9.669186754265619e-07, 'learning_rate/action_model': 9.669186754265618e-06, 'epoch': 1.77}) +[2026-06-17 17:33:04,325][starVLA.training.train_starvla][INFO] - Step 6019, Loss: {'train/loss': 0.6784404516220093, 'train/grad_norm_pre_clip': 16.60606575012207, 'timing/data': 0.005175665020942688, 'timing/model': 1.3195920740254223, 'learning_rate/qwen_vl_interface': 9.65984059649457e-07, 'learning_rate/action_model': 9.65984059649457e-06, 'epoch': 1.77}) +[2026-06-17 17:33:08,965][starVLA.training.train_starvla][INFO] - Step 6020, Loss: {'train/loss': 0.8926307559013367, 'train/grad_norm_pre_clip': 20.147546768188477, 'timing/data': 0.005137962056323886, 'timing/model': 1.2002685209736228, 'learning_rate/qwen_vl_interface': 9.650503319534652e-07, 'learning_rate/action_model': 9.650503319534651e-06, 'epoch': 1.77}) +[2026-06-17 17:33:13,748][starVLA.training.train_starvla][INFO] - Step 6021, Loss: {'train/loss': 0.6470310688018799, 'train/grad_norm_pre_clip': 15.112144470214844, 'timing/data': 0.005274606985040009, 'timing/model': 1.2017531560268253, 'learning_rate/qwen_vl_interface': 9.641174925321494e-07, 'learning_rate/action_model': 9.641174925321494e-06, 'epoch': 1.77}) +[2026-06-17 17:33:18,618][starVLA.training.train_starvla][INFO] - Step 6022, Loss: {'train/loss': 0.607518196105957, 'train/grad_norm_pre_clip': 14.198009490966797, 'timing/data': 0.004869100986979902, 'timing/model': 1.2029137259814888, 'learning_rate/qwen_vl_interface': 9.631855415788904e-07, 'learning_rate/action_model': 9.631855415788904e-06, 'epoch': 1.77}) +[2026-06-17 17:33:23,399][starVLA.training.train_starvla][INFO] - Step 6023, Loss: {'train/loss': 0.5739622712135315, 'train/grad_norm_pre_clip': 14.104389190673828, 'timing/data': 0.0051146079786121845, 'timing/model': 1.204070859006606, 'learning_rate/qwen_vl_interface': 9.622544792868812e-07, 'learning_rate/action_model': 9.622544792868812e-06, 'epoch': 1.77}) +[2026-06-17 17:33:28,183][starVLA.training.train_starvla][INFO] - Step 6024, Loss: {'train/loss': 0.5372133851051331, 'train/grad_norm_pre_clip': 15.370800018310547, 'timing/data': 0.005701906979084015, 'timing/model': 1.3415247600059956, 'learning_rate/qwen_vl_interface': 9.613243058491334e-07, 'learning_rate/action_model': 9.613243058491334e-06, 'epoch': 1.77}) +[2026-06-17 17:33:32,837][starVLA.training.train_starvla][INFO] - Step 6025, Loss: {'train/loss': 0.6032754778862, 'train/grad_norm_pre_clip': 13.493725776672363, 'timing/data': 0.005144643015228212, 'timing/model': 1.2056409959914163, 'learning_rate/qwen_vl_interface': 9.603950214584695e-07, 'learning_rate/action_model': 9.603950214584696e-06, 'epoch': 1.77}) +[2026-06-17 17:33:37,628][starVLA.training.train_starvla][INFO] - Step 6026, Loss: {'train/loss': 0.6786683797836304, 'train/grad_norm_pre_clip': 16.1154727935791, 'timing/data': 0.005003486992791295, 'timing/model': 1.2038371169473976, 'learning_rate/qwen_vl_interface': 9.59466626307535e-07, 'learning_rate/action_model': 9.594666263075349e-06, 'epoch': 1.77}) +[2026-06-17 17:33:42,323][starVLA.training.train_starvla][INFO] - Step 6027, Loss: {'train/loss': 0.6123883724212646, 'train/grad_norm_pre_clip': 16.2663631439209, 'timing/data': 0.004853804945014417, 'timing/model': 1.2023321139859036, 'learning_rate/qwen_vl_interface': 9.585391205887851e-07, 'learning_rate/action_model': 9.585391205887851e-06, 'epoch': 1.77}) +[2026-06-17 17:33:47,050][starVLA.training.train_starvla][INFO] - Step 6028, Loss: {'train/loss': 0.6076123118400574, 'train/grad_norm_pre_clip': 14.959603309631348, 'timing/data': 0.005354612949304283, 'timing/model': 1.2016908680088818, 'learning_rate/qwen_vl_interface': 9.57612504494494e-07, 'learning_rate/action_model': 9.57612504494494e-06, 'epoch': 1.77}) +[2026-06-17 17:33:51,829][starVLA.training.train_starvla][INFO] - Step 6029, Loss: {'train/loss': 0.47335943579673767, 'train/grad_norm_pre_clip': 12.237128257751465, 'timing/data': 0.00528177700471133, 'timing/model': 1.2011340599274263, 'learning_rate/qwen_vl_interface': 9.566867782167494e-07, 'learning_rate/action_model': 9.566867782167494e-06, 'epoch': 1.77}) +[2026-06-17 17:33:56,613][starVLA.training.train_starvla][INFO] - Step 6030, Loss: {'train/loss': 0.71321702003479, 'train/grad_norm_pre_clip': 13.967377662658691, 'timing/data': 0.005011912086047232, 'timing/model': 1.3353103779954836, 'learning_rate/qwen_vl_interface': 9.55761941947455e-07, 'learning_rate/action_model': 9.557619419474551e-06, 'epoch': 1.77}) +[2026-06-17 17:34:01,258][starVLA.training.train_starvla][INFO] - Step 6031, Loss: {'train/loss': 0.531181812286377, 'train/grad_norm_pre_clip': 11.835592269897461, 'timing/data': 0.004993715090677142, 'timing/model': 1.20054288697429, 'learning_rate/qwen_vl_interface': 9.548379958783308e-07, 'learning_rate/action_model': 9.548379958783308e-06, 'epoch': 1.77}) +[2026-06-17 17:34:06,041][starVLA.training.train_starvla][INFO] - Step 6032, Loss: {'train/loss': 0.5994280576705933, 'train/grad_norm_pre_clip': 12.597360610961914, 'timing/data': 0.0053141029784455895, 'timing/model': 1.2126197520410642, 'learning_rate/qwen_vl_interface': 9.539149402009114e-07, 'learning_rate/action_model': 9.539149402009115e-06, 'epoch': 1.77}) +[2026-06-17 17:34:10,821][starVLA.training.train_starvla][INFO] - Step 6033, Loss: {'train/loss': 0.6571981310844421, 'train/grad_norm_pre_clip': 12.5176362991333, 'timing/data': 0.00517269701231271, 'timing/model': 1.203203639946878, 'learning_rate/qwen_vl_interface': 9.529927751065493e-07, 'learning_rate/action_model': 9.529927751065492e-06, 'epoch': 1.78}) +[2026-06-17 17:34:15,565][starVLA.training.train_starvla][INFO] - Step 6034, Loss: {'train/loss': 0.6215744018554688, 'train/grad_norm_pre_clip': 13.809159278869629, 'timing/data': 0.00496151705738157, 'timing/model': 1.2066512410528958, 'learning_rate/qwen_vl_interface': 9.520715007864082e-07, 'learning_rate/action_model': 9.52071500786408e-06, 'epoch': 1.78}) +[2026-06-17 17:34:20,338][starVLA.training.train_starvla][INFO] - Step 6035, Loss: {'train/loss': 0.5349774360656738, 'train/grad_norm_pre_clip': 13.349908828735352, 'timing/data': 0.005143519025295973, 'timing/model': 1.2036763059441, 'learning_rate/qwen_vl_interface': 9.511511174314689e-07, 'learning_rate/action_model': 9.511511174314688e-06, 'epoch': 1.78}) +[2026-06-17 17:34:25,124][starVLA.training.train_starvla][INFO] - Step 6036, Loss: {'train/loss': 0.5986981987953186, 'train/grad_norm_pre_clip': 14.07683277130127, 'timing/data': 0.004871829994954169, 'timing/model': 1.3324010559590533, 'learning_rate/qwen_vl_interface': 9.502316252325283e-07, 'learning_rate/action_model': 9.502316252325283e-06, 'epoch': 1.78}) +[2026-06-17 17:34:29,772][starVLA.training.train_starvla][INFO] - Step 6037, Loss: {'train/loss': 0.5576673150062561, 'train/grad_norm_pre_clip': 13.881251335144043, 'timing/data': 0.005366994999349117, 'timing/model': 1.2033561239950359, 'learning_rate/qwen_vl_interface': 9.493130243801993e-07, 'learning_rate/action_model': 9.493130243801992e-06, 'epoch': 1.78}) +[2026-06-17 17:34:34,557][starVLA.training.train_starvla][INFO] - Step 6038, Loss: {'train/loss': 0.7166858911514282, 'train/grad_norm_pre_clip': 16.058473587036133, 'timing/data': 0.005434994003735483, 'timing/model': 1.2039193579694256, 'learning_rate/qwen_vl_interface': 9.483953150649091e-07, 'learning_rate/action_model': 9.483953150649091e-06, 'epoch': 1.78}) +[2026-06-17 17:34:39,346][starVLA.training.train_starvla][INFO] - Step 6039, Loss: {'train/loss': 0.9401466846466064, 'train/grad_norm_pre_clip': 17.598176956176758, 'timing/data': 0.005476319929584861, 'timing/model': 1.2050679200328887, 'learning_rate/qwen_vl_interface': 9.474784974768982e-07, 'learning_rate/action_model': 9.474784974768982e-06, 'epoch': 1.78}) +[2026-06-17 17:34:44,125][starVLA.training.train_starvla][INFO] - Step 6040, Loss: {'train/loss': 0.4743664562702179, 'train/grad_norm_pre_clip': 20.26426887512207, 'timing/data': 0.004998889984562993, 'timing/model': 1.2057745720958337, 'learning_rate/qwen_vl_interface': 9.465625718062236e-07, 'learning_rate/action_model': 9.465625718062235e-06, 'epoch': 1.78}) +[2026-06-17 17:34:48,791][starVLA.training.train_starvla][INFO] - Step 6041, Loss: {'train/loss': 0.8596859574317932, 'train/grad_norm_pre_clip': 20.803213119506836, 'timing/data': 0.005540629965253174, 'timing/model': 1.2056435719132423, 'learning_rate/qwen_vl_interface': 9.456475382427591e-07, 'learning_rate/action_model': 9.45647538242759e-06, 'epoch': 1.78}) +[2026-06-17 17:34:53,486][starVLA.training.train_starvla][INFO] - Step 6042, Loss: {'train/loss': 0.43400508165359497, 'train/grad_norm_pre_clip': 11.7876615524292, 'timing/data': 0.005459825973957777, 'timing/model': 1.2445468279765919, 'learning_rate/qwen_vl_interface': 9.447333969761915e-07, 'learning_rate/action_model': 9.447333969761915e-06, 'epoch': 1.78}) +[2026-06-17 17:34:58,143][starVLA.training.train_starvla][INFO] - Step 6043, Loss: {'train/loss': 0.6381710767745972, 'train/grad_norm_pre_clip': 11.859457015991211, 'timing/data': 0.005579867982305586, 'timing/model': 1.2053533220896497, 'learning_rate/qwen_vl_interface': 9.438201481960235e-07, 'learning_rate/action_model': 9.438201481960234e-06, 'epoch': 1.78}) +[2026-06-17 17:35:02,897][starVLA.training.train_starvla][INFO] - Step 6044, Loss: {'train/loss': 0.4271341562271118, 'train/grad_norm_pre_clip': 11.156250953674316, 'timing/data': 0.006366329966112971, 'timing/model': 1.211277917958796, 'learning_rate/qwen_vl_interface': 9.429077920915724e-07, 'learning_rate/action_model': 9.429077920915724e-06, 'epoch': 1.78}) +[2026-06-17 17:35:07,683][starVLA.training.train_starvla][INFO] - Step 6045, Loss: {'train/loss': 0.41906479001045227, 'train/grad_norm_pre_clip': 13.269392013549805, 'timing/data': 0.004881458007730544, 'timing/model': 1.2027944199508056, 'learning_rate/qwen_vl_interface': 9.419963288519711e-07, 'learning_rate/action_model': 9.41996328851971e-06, 'epoch': 1.78}) +[2026-06-17 17:35:12,471][starVLA.training.train_starvla][INFO] - Step 6046, Loss: {'train/loss': 0.7670943737030029, 'train/grad_norm_pre_clip': 17.469797134399414, 'timing/data': 0.004836134030483663, 'timing/model': 1.2050525549566373, 'learning_rate/qwen_vl_interface': 9.410857586661659e-07, 'learning_rate/action_model': 9.410857586661659e-06, 'epoch': 1.78}) +[2026-06-17 17:35:17,263][starVLA.training.train_starvla][INFO] - Step 6047, Loss: {'train/loss': 0.5896141529083252, 'train/grad_norm_pre_clip': 14.60206127166748, 'timing/data': 0.005257771001197398, 'timing/model': 1.34227878798265, 'learning_rate/qwen_vl_interface': 9.401760817229197e-07, 'learning_rate/action_model': 9.401760817229196e-06, 'epoch': 1.78}) +[2026-06-17 17:35:21,968][starVLA.training.train_starvla][INFO] - Step 6048, Loss: {'train/loss': 0.6574887633323669, 'train/grad_norm_pre_clip': 42.12105941772461, 'timing/data': 0.005125849042087793, 'timing/model': 1.2507687140023336, 'learning_rate/qwen_vl_interface': 9.392672982108091e-07, 'learning_rate/action_model': 9.392672982108091e-06, 'epoch': 1.78}) +[2026-06-17 17:35:26,617][starVLA.training.train_starvla][INFO] - Step 6049, Loss: {'train/loss': 0.621326744556427, 'train/grad_norm_pre_clip': 14.598050117492676, 'timing/data': 0.004783764015883207, 'timing/model': 1.202885975013487, 'learning_rate/qwen_vl_interface': 9.383594083182265e-07, 'learning_rate/action_model': 9.383594083182265e-06, 'epoch': 1.78}) +[2026-06-17 17:35:31,364][starVLA.training.train_starvla][INFO] - Step 6050, Loss: {'train/loss': 0.7555787563323975, 'train/grad_norm_pre_clip': 15.084775924682617, 'timing/data': 0.004962258972227573, 'timing/model': 1.2041491119889542, 'learning_rate/qwen_vl_interface': 9.374524122333769e-07, 'learning_rate/action_model': 9.374524122333769e-06, 'epoch': 1.78}) +[2026-06-17 17:35:36,147][starVLA.training.train_starvla][INFO] - Step 6051, Loss: {'train/loss': 0.4580170810222626, 'train/grad_norm_pre_clip': 11.986064910888672, 'timing/data': 0.005473372992128134, 'timing/model': 1.2062728820601478, 'learning_rate/qwen_vl_interface': 9.365463101442842e-07, 'learning_rate/action_model': 9.365463101442842e-06, 'epoch': 1.78}) +[2026-06-17 17:35:40,931][starVLA.training.train_starvla][INFO] - Step 6052, Loss: {'train/loss': 0.680508017539978, 'train/grad_norm_pre_clip': 14.995658874511719, 'timing/data': 0.004891460062935948, 'timing/model': 1.2036874470068142, 'learning_rate/qwen_vl_interface': 9.356411022387823e-07, 'learning_rate/action_model': 9.356411022387824e-06, 'epoch': 1.78}) +[2026-06-17 17:35:45,722][starVLA.training.train_starvla][INFO] - Step 6053, Loss: {'train/loss': 0.575984537601471, 'train/grad_norm_pre_clip': 13.668978691101074, 'timing/data': 0.005018521915189922, 'timing/model': 1.3351938750129193, 'learning_rate/qwen_vl_interface': 9.347367887045236e-07, 'learning_rate/action_model': 9.347367887045236e-06, 'epoch': 1.78}) +[2026-06-17 17:35:50,501][starVLA.training.train_starvla][INFO] - Step 6054, Loss: {'train/loss': 0.76849764585495, 'train/grad_norm_pre_clip': 13.847869873046875, 'timing/data': 0.005328341969288886, 'timing/model': 1.3222699649631977, 'learning_rate/qwen_vl_interface': 9.338333697289697e-07, 'learning_rate/action_model': 9.338333697289696e-06, 'epoch': 1.78}) +[2026-06-17 17:35:55,153][starVLA.training.train_starvla][INFO] - Step 6055, Loss: {'train/loss': 0.5072504878044128, 'train/grad_norm_pre_clip': 11.512519836425781, 'timing/data': 0.005256829084828496, 'timing/model': 1.2054224549792707, 'learning_rate/qwen_vl_interface': 9.329308454994044e-07, 'learning_rate/action_model': 9.329308454994043e-06, 'epoch': 1.78}) +[2026-06-17 17:35:59,913][starVLA.training.train_starvla][INFO] - Step 6056, Loss: {'train/loss': 0.5281287431716919, 'train/grad_norm_pre_clip': 12.912726402282715, 'timing/data': 0.0050584550481289625, 'timing/model': 1.2084568060236052, 'learning_rate/qwen_vl_interface': 9.320292162029192e-07, 'learning_rate/action_model': 9.320292162029191e-06, 'epoch': 1.78}) +[2026-06-17 17:36:04,696][starVLA.training.train_starvla][INFO] - Step 6057, Loss: {'train/loss': 0.5063989162445068, 'train/grad_norm_pre_clip': 13.064397811889648, 'timing/data': 0.005460045067593455, 'timing/model': 1.205544124939479, 'learning_rate/qwen_vl_interface': 9.311284820264245e-07, 'learning_rate/action_model': 9.311284820264244e-06, 'epoch': 1.78}) +[2026-06-17 17:36:09,488][starVLA.training.train_starvla][INFO] - Step 6058, Loss: {'train/loss': 0.6576147079467773, 'train/grad_norm_pre_clip': 14.127748489379883, 'timing/data': 0.005192731041461229, 'timing/model': 1.2074124329956248, 'learning_rate/qwen_vl_interface': 9.302286431566428e-07, 'learning_rate/action_model': 9.302286431566428e-06, 'epoch': 1.78}) +[2026-06-17 17:36:14,263][starVLA.training.train_starvla][INFO] - Step 6059, Loss: {'train/loss': 0.6167126297950745, 'train/grad_norm_pre_clip': 17.015661239624023, 'timing/data': 0.0051277290331199765, 'timing/model': 1.274177129031159, 'learning_rate/qwen_vl_interface': 9.293296997801113e-07, 'learning_rate/action_model': 9.293296997801112e-06, 'epoch': 1.78}) +[2026-06-17 17:36:18,924][starVLA.training.train_starvla][INFO] - Step 6060, Loss: {'train/loss': 0.779897928237915, 'train/grad_norm_pre_clip': 14.302882194519043, 'timing/data': 0.005004617036320269, 'timing/model': 1.2043002879945561, 'learning_rate/qwen_vl_interface': 9.284316520831825e-07, 'learning_rate/action_model': 9.284316520831825e-06, 'epoch': 1.78}) +[2026-06-17 17:36:23,694][starVLA.training.train_starvla][INFO] - Step 6061, Loss: {'train/loss': 0.6525018811225891, 'train/grad_norm_pre_clip': 15.504890441894531, 'timing/data': 0.005083916010335088, 'timing/model': 1.206776890088804, 'learning_rate/qwen_vl_interface': 9.275345002520225e-07, 'learning_rate/action_model': 9.275345002520225e-06, 'epoch': 1.78}) +[2026-06-17 17:36:28,485][starVLA.training.train_starvla][INFO] - Step 6062, Loss: {'train/loss': 0.41153162717819214, 'train/grad_norm_pre_clip': 11.285812377929688, 'timing/data': 0.004820031113922596, 'timing/model': 1.2089493420207873, 'learning_rate/qwen_vl_interface': 9.266382444726117e-07, 'learning_rate/action_model': 9.266382444726117e-06, 'epoch': 1.78}) +[2026-06-17 17:36:33,271][starVLA.training.train_starvla][INFO] - Step 6063, Loss: {'train/loss': 0.8690218925476074, 'train/grad_norm_pre_clip': 14.167964935302734, 'timing/data': 0.005122689995914698, 'timing/model': 1.205681586987339, 'learning_rate/qwen_vl_interface': 9.257428849307446e-07, 'learning_rate/action_model': 9.257428849307445e-06, 'epoch': 1.78}) +[2026-06-17 17:36:38,057][starVLA.training.train_starvla][INFO] - Step 6064, Loss: {'train/loss': 0.7143805027008057, 'train/grad_norm_pre_clip': 17.178293228149414, 'timing/data': 0.004876842023804784, 'timing/model': 1.3295303849736229, 'learning_rate/qwen_vl_interface': 9.248484218120307e-07, 'learning_rate/action_model': 9.248484218120307e-06, 'epoch': 1.78}) +[2026-06-17 17:36:42,810][starVLA.training.train_starvla][INFO] - Step 6065, Loss: {'train/loss': 0.645310640335083, 'train/grad_norm_pre_clip': 13.680693626403809, 'timing/data': 0.005113895982503891, 'timing/model': 1.2839898350648582, 'learning_rate/qwen_vl_interface': 9.239548553018925e-07, 'learning_rate/action_model': 9.239548553018924e-06, 'epoch': 1.78}) +[2026-06-17 17:36:47,467][starVLA.training.train_starvla][INFO] - Step 6066, Loss: {'train/loss': 1.0960551500320435, 'train/grad_norm_pre_clip': 16.388736724853516, 'timing/data': 0.004937933990731835, 'timing/model': 1.20543136599008, 'learning_rate/qwen_vl_interface': 9.230621855855687e-07, 'learning_rate/action_model': 9.230621855855687e-06, 'epoch': 1.78}) +[2026-06-17 17:36:52,263][starVLA.training.train_starvla][INFO] - Step 6067, Loss: {'train/loss': 0.6125761866569519, 'train/grad_norm_pre_clip': 14.289912223815918, 'timing/data': 0.0053311140509322286, 'timing/model': 1.2045254219556227, 'learning_rate/qwen_vl_interface': 9.221704128481107e-07, 'learning_rate/action_model': 9.221704128481106e-06, 'epoch': 1.79}) +[2026-06-17 17:36:57,052][starVLA.training.train_starvla][INFO] - Step 6068, Loss: {'train/loss': 0.6290399432182312, 'train/grad_norm_pre_clip': 14.260027885437012, 'timing/data': 0.005238411948084831, 'timing/model': 1.2056300640106201, 'learning_rate/qwen_vl_interface': 9.212795372743815e-07, 'learning_rate/action_model': 9.212795372743813e-06, 'epoch': 1.79}) +[2026-06-17 17:37:01,839][starVLA.training.train_starvla][INFO] - Step 6069, Loss: {'train/loss': 0.36099106073379517, 'train/grad_norm_pre_clip': 12.924708366394043, 'timing/data': 0.005288598011247814, 'timing/model': 1.2056911289691925, 'learning_rate/qwen_vl_interface': 9.203895590490617e-07, 'learning_rate/action_model': 9.203895590490616e-06, 'epoch': 1.79}) +[2026-06-17 17:37:06,619][starVLA.training.train_starvla][INFO] - Step 6070, Loss: {'train/loss': 0.7246987819671631, 'train/grad_norm_pre_clip': 17.4581356048584, 'timing/data': 0.005008634063415229, 'timing/model': 1.2062563160434365, 'learning_rate/qwen_vl_interface': 9.195004783566464e-07, 'learning_rate/action_model': 9.195004783566463e-06, 'epoch': 1.79}) +[2026-06-17 17:37:11,365][starVLA.training.train_starvla][INFO] - Step 6071, Loss: {'train/loss': 0.5493453145027161, 'train/grad_norm_pre_clip': 12.899751663208008, 'timing/data': 0.005476222024299204, 'timing/model': 1.2972527910023928, 'learning_rate/qwen_vl_interface': 9.186122953814414e-07, 'learning_rate/action_model': 9.186122953814413e-06, 'epoch': 1.79}) +[2026-06-17 17:37:16,017][starVLA.training.train_starvla][INFO] - Step 6072, Loss: {'train/loss': 0.45134106278419495, 'train/grad_norm_pre_clip': 12.257856369018555, 'timing/data': 0.0054822140373289585, 'timing/model': 1.2024899800308049, 'learning_rate/qwen_vl_interface': 9.177250103075686e-07, 'learning_rate/action_model': 9.177250103075686e-06, 'epoch': 1.79}) +[2026-06-17 17:37:20,797][starVLA.training.train_starvla][INFO] - Step 6073, Loss: {'train/loss': 0.9478171467781067, 'train/grad_norm_pre_clip': 19.839757919311523, 'timing/data': 0.00541861797682941, 'timing/model': 1.202579353004694, 'learning_rate/qwen_vl_interface': 9.168386233189627e-07, 'learning_rate/action_model': 9.168386233189627e-06, 'epoch': 1.79}) +[2026-06-17 17:37:25,587][starVLA.training.train_starvla][INFO] - Step 6074, Loss: {'train/loss': 0.6834662556648254, 'train/grad_norm_pre_clip': 15.015085220336914, 'timing/data': 0.005103584029711783, 'timing/model': 1.2060757650760934, 'learning_rate/qwen_vl_interface': 9.159531345993726e-07, 'learning_rate/action_model': 9.159531345993725e-06, 'epoch': 1.79}) +[2026-06-17 17:37:30,378][starVLA.training.train_starvla][INFO] - Step 6075, Loss: {'train/loss': 0.7787573933601379, 'train/grad_norm_pre_clip': 18.37488555908203, 'timing/data': 0.005069968989118934, 'timing/model': 1.2070119570707902, 'learning_rate/qwen_vl_interface': 9.150685443323619e-07, 'learning_rate/action_model': 9.150685443323618e-06, 'epoch': 1.79}) +[2026-06-17 17:37:35,178][starVLA.training.train_starvla][INFO] - Step 6076, Loss: {'train/loss': 0.6931353211402893, 'train/grad_norm_pre_clip': 14.120416641235352, 'timing/data': 0.005295853014104068, 'timing/model': 1.3456709110178053, 'learning_rate/qwen_vl_interface': 9.141848527013059e-07, 'learning_rate/action_model': 9.141848527013059e-06, 'epoch': 1.79}) +[2026-06-17 17:37:39,838][starVLA.training.train_starvla][INFO] - Step 6077, Loss: {'train/loss': 0.7366716265678406, 'train/grad_norm_pre_clip': 23.380950927734375, 'timing/data': 0.005015459028072655, 'timing/model': 1.208019788027741, 'learning_rate/qwen_vl_interface': 9.13302059889396e-07, 'learning_rate/action_model': 9.13302059889396e-06, 'epoch': 1.79}) +[2026-06-17 17:37:44,630][starVLA.training.train_starvla][INFO] - Step 6078, Loss: {'train/loss': 0.6745041012763977, 'train/grad_norm_pre_clip': 16.661054611206055, 'timing/data': 0.005176833947189152, 'timing/model': 1.2033271960681304, 'learning_rate/qwen_vl_interface': 9.12420166079635e-07, 'learning_rate/action_model': 9.12420166079635e-06, 'epoch': 1.79}) +[2026-06-17 17:37:49,423][starVLA.training.train_starvla][INFO] - Step 6079, Loss: {'train/loss': 0.4928547143936157, 'train/grad_norm_pre_clip': 12.473891258239746, 'timing/data': 0.004720867029391229, 'timing/model': 1.2074664660030976, 'learning_rate/qwen_vl_interface': 9.115391714548418e-07, 'learning_rate/action_model': 9.115391714548418e-06, 'epoch': 1.79}) +[2026-06-17 17:37:54,249][starVLA.training.train_starvla][INFO] - Step 6080, Loss: {'train/loss': 0.4724614918231964, 'train/grad_norm_pre_clip': 14.314136505126953, 'timing/data': 0.005107032950036228, 'timing/model': 1.2064035050570965, 'learning_rate/qwen_vl_interface': 9.106590761976454e-07, 'learning_rate/action_model': 9.106590761976455e-06, 'epoch': 1.79}) +[2026-06-17 17:37:58,985][starVLA.training.train_starvla][INFO] - Step 6081, Loss: {'train/loss': 0.6565686464309692, 'train/grad_norm_pre_clip': 15.457086563110352, 'timing/data': 0.005390009959228337, 'timing/model': 1.2047161080408841, 'learning_rate/qwen_vl_interface': 9.097798804904934e-07, 'learning_rate/action_model': 9.097798804904933e-06, 'epoch': 1.79}) +[2026-06-17 17:38:03,791][starVLA.training.train_starvla][INFO] - Step 6082, Loss: {'train/loss': 0.7569786906242371, 'train/grad_norm_pre_clip': 16.24552345275879, 'timing/data': 0.00491569796577096, 'timing/model': 1.3338673330144957, 'learning_rate/qwen_vl_interface': 9.089015845156422e-07, 'learning_rate/action_model': 9.089015845156422e-06, 'epoch': 1.79}) +[2026-06-17 17:38:08,451][starVLA.training.train_starvla][INFO] - Step 6083, Loss: {'train/loss': 0.3767445683479309, 'train/grad_norm_pre_clip': 10.642914772033691, 'timing/data': 0.005405873991549015, 'timing/model': 1.2054754940327257, 'learning_rate/qwen_vl_interface': 9.080241884551627e-07, 'learning_rate/action_model': 9.080241884551627e-06, 'epoch': 1.79}) +[2026-06-17 17:38:13,236][starVLA.training.train_starvla][INFO] - Step 6084, Loss: {'train/loss': 0.7281467914581299, 'train/grad_norm_pre_clip': 15.802618026733398, 'timing/data': 0.004760632989928126, 'timing/model': 1.2045047510182485, 'learning_rate/qwen_vl_interface': 9.071476924909417e-07, 'learning_rate/action_model': 9.071476924909416e-06, 'epoch': 1.79}) +[2026-06-17 17:38:18,025][starVLA.training.train_starvla][INFO] - Step 6085, Loss: {'train/loss': 0.6149850487709045, 'train/grad_norm_pre_clip': 15.456812858581543, 'timing/data': 0.005155963008292019, 'timing/model': 1.2030415839981288, 'learning_rate/qwen_vl_interface': 9.062720968046773e-07, 'learning_rate/action_model': 9.062720968046773e-06, 'epoch': 1.79}) +[2026-06-17 17:38:22,813][starVLA.training.train_starvla][INFO] - Step 6086, Loss: {'train/loss': 0.8569520115852356, 'train/grad_norm_pre_clip': 15.52924919128418, 'timing/data': 0.005238084006123245, 'timing/model': 1.2024597079725936, 'learning_rate/qwen_vl_interface': 9.053974015778817e-07, 'learning_rate/action_model': 9.053974015778817e-06, 'epoch': 1.79}) +[2026-06-17 17:38:27,567][starVLA.training.train_starvla][INFO] - Step 6087, Loss: {'train/loss': 1.03474760055542, 'train/grad_norm_pre_clip': 16.730165481567383, 'timing/data': 0.004891825024969876, 'timing/model': 1.2055394620401785, 'learning_rate/qwen_vl_interface': 9.045236069918788e-07, 'learning_rate/action_model': 9.045236069918788e-06, 'epoch': 1.79}) +[2026-06-17 17:38:32,380][starVLA.training.train_starvla][INFO] - Step 6088, Loss: {'train/loss': 0.5658454298973083, 'train/grad_norm_pre_clip': 15.756921768188477, 'timing/data': 0.004786460078321397, 'timing/model': 1.3236273609800264, 'learning_rate/qwen_vl_interface': 9.036507132278085e-07, 'learning_rate/action_model': 9.036507132278086e-06, 'epoch': 1.79}) +[2026-06-17 17:38:37,031][starVLA.training.train_starvla][INFO] - Step 6089, Loss: {'train/loss': 0.7546201348304749, 'train/grad_norm_pre_clip': 14.770942687988281, 'timing/data': 0.005535303964279592, 'timing/model': 1.2040295009501278, 'learning_rate/qwen_vl_interface': 9.02778720466622e-07, 'learning_rate/action_model': 9.02778720466622e-06, 'epoch': 1.79}) +[2026-06-17 17:38:41,816][starVLA.training.train_starvla][INFO] - Step 6090, Loss: {'train/loss': 0.6481729745864868, 'train/grad_norm_pre_clip': 16.920501708984375, 'timing/data': 0.0050754520343616605, 'timing/model': 1.2037339910166338, 'learning_rate/qwen_vl_interface': 9.019076288890852e-07, 'learning_rate/action_model': 9.019076288890852e-06, 'epoch': 1.79}) +[2026-06-17 17:38:46,605][starVLA.training.train_starvla][INFO] - Step 6091, Loss: {'train/loss': 0.8046562075614929, 'train/grad_norm_pre_clip': 15.619297981262207, 'timing/data': 0.005147660034708679, 'timing/model': 1.20404593704734, 'learning_rate/qwen_vl_interface': 9.010374386757755e-07, 'learning_rate/action_model': 9.010374386757754e-06, 'epoch': 1.79}) +[2026-06-17 17:38:51,391][starVLA.training.train_starvla][INFO] - Step 6092, Loss: {'train/loss': 0.6668125987052917, 'train/grad_norm_pre_clip': 15.085237503051758, 'timing/data': 0.0049321039114147425, 'timing/model': 1.203273204038851, 'learning_rate/qwen_vl_interface': 9.001681500070842e-07, 'learning_rate/action_model': 9.001681500070842e-06, 'epoch': 1.79}) +[2026-06-17 17:38:56,190][starVLA.training.train_starvla][INFO] - Step 6093, Loss: {'train/loss': 0.5542401671409607, 'train/grad_norm_pre_clip': 14.458043098449707, 'timing/data': 0.0048294030129909515, 'timing/model': 1.3439571299823, 'learning_rate/qwen_vl_interface': 8.992997630632161e-07, 'learning_rate/action_model': 8.99299763063216e-06, 'epoch': 1.79}) +[2026-06-17 17:39:00,861][starVLA.training.train_starvla][INFO] - Step 6094, Loss: {'train/loss': 0.4805915951728821, 'train/grad_norm_pre_clip': 10.6398286819458, 'timing/data': 0.004785730037838221, 'timing/model': 1.2118552479660138, 'learning_rate/qwen_vl_interface': 8.984322780241879e-07, 'learning_rate/action_model': 8.984322780241878e-06, 'epoch': 1.79}) +[2026-06-17 17:39:05,647][starVLA.training.train_starvla][INFO] - Step 6095, Loss: {'train/loss': 0.4236100912094116, 'train/grad_norm_pre_clip': 10.820940017700195, 'timing/data': 0.005323552992194891, 'timing/model': 1.2038765070028603, 'learning_rate/qwen_vl_interface': 8.975656950698325e-07, 'learning_rate/action_model': 8.975656950698324e-06, 'epoch': 1.79}) +[2026-06-17 17:39:10,362][starVLA.training.train_starvla][INFO] - Step 6096, Loss: {'train/loss': 0.7078211307525635, 'train/grad_norm_pre_clip': 14.293837547302246, 'timing/data': 0.004782745963893831, 'timing/model': 1.2072602779371664, 'learning_rate/qwen_vl_interface': 8.967000143797913e-07, 'learning_rate/action_model': 8.967000143797913e-06, 'epoch': 1.79}) +[2026-06-17 17:39:15,136][starVLA.training.train_starvla][INFO] - Step 6097, Loss: {'train/loss': 0.6355735659599304, 'train/grad_norm_pre_clip': 17.142471313476562, 'timing/data': 0.005079052993096411, 'timing/model': 1.2114019440487027, 'learning_rate/qwen_vl_interface': 8.958352361335214e-07, 'learning_rate/action_model': 8.958352361335213e-06, 'epoch': 1.79}) +[2026-06-17 17:39:19,944][starVLA.training.train_starvla][INFO] - Step 6098, Loss: {'train/loss': 0.7031587362289429, 'train/grad_norm_pre_clip': 14.995505332946777, 'timing/data': 0.005070068989880383, 'timing/model': 1.2230490030488, 'learning_rate/qwen_vl_interface': 8.949713605102917e-07, 'learning_rate/action_model': 8.949713605102916e-06, 'epoch': 1.79}) +[2026-06-17 17:39:24,765][starVLA.training.train_starvla][INFO] - Step 6099, Loss: {'train/loss': 0.5049524903297424, 'train/grad_norm_pre_clip': 12.706542015075684, 'timing/data': 0.006122077931649983, 'timing/model': 1.3622436759760603, 'learning_rate/qwen_vl_interface': 8.941083876891857e-07, 'learning_rate/action_model': 8.941083876891856e-06, 'epoch': 1.79}) +[2026-06-17 17:41:06,708][starVLA.training.train_starvla][INFO] - Step 6100, Loss: {'train/loss': 0.5088828802108765, 'train/grad_norm_pre_clip': 14.021247863769531, 'eval/loss': 0.7070123553276062, 'eval/latency_0/loss': 0.7070123553276062, 'eval/demon_fire_precision': 0.9050121261115602, 'eval/demon_fire_recall': 0.9298172757475083, 'eval/demon_fire_f1': 0.9172470299057763, 'eval/demon_fire_rate_model': 0.6879866518353727, 'eval/demon_fire_rate_teacher': 0.6696329254727474, 'eval/demon_left_precision': 0.8241262683201804, 'eval/demon_left_recall': 0.828328611898017, 'eval/demon_left_f1': 0.8262220966374682, 'eval/demon_left_rate_model': 0.4933259176863181, 'eval/demon_left_rate_teacher': 0.49082313681868744, 'eval/demon_right_precision': 0.8319327731092437, 'eval/demon_right_recall': 0.8389830508474576, 'eval/demon_right_f1': 0.8354430379746834, 'eval/demon_right_rate_model': 0.49638487208008897, 'eval/demon_right_rate_teacher': 0.4922135706340378, 'eval/demon_move_f1': 0.8308325673060758, 'eval/demon_cc_f1': 0.8894923394221291, 'eval/latency_0/demon_fire_precision': 0.9050121261115602, 'eval/latency_0/demon_fire_recall': 0.9298172757475083, 'eval/latency_0/demon_fire_f1': 0.9172470299057763, 'eval/latency_0/demon_fire_rate_model': 0.6879866518353727, 'eval/latency_0/demon_fire_rate_teacher': 0.6696329254727474, 'eval/latency_0/demon_left_precision': 0.8241262683201804, 'eval/latency_0/demon_left_recall': 0.828328611898017, 'eval/latency_0/demon_left_f1': 0.8262220966374682, 'eval/latency_0/demon_left_rate_model': 0.4933259176863181, 'eval/latency_0/demon_left_rate_teacher': 0.49082313681868744, 'eval/latency_0/demon_right_precision': 0.8319327731092437, 'eval/latency_0/demon_right_recall': 0.8389830508474576, 'eval/latency_0/demon_right_f1': 0.8354430379746834, 'eval/latency_0/demon_right_rate_model': 0.49638487208008897, 'eval/latency_0/demon_right_rate_teacher': 0.4922135706340378, 'eval/latency_0/demon_move_f1': 0.8308325673060758, 'eval/latency_0/demon_cc_f1': 0.8894923394221291, 'timing/data': 0.00507665891200304, 'timing/model': 1.2102995190070942, 'learning_rate/qwen_vl_interface': 8.932463178490991e-07, 'learning_rate/action_model': 8.93246317849099e-06, 'epoch': 1.79}) +[2026-06-17 17:41:06,709][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_6100_state +[2026-06-17 17:41:06,712][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 17:41:10,271][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_6100_state/model.safetensors +[2026-06-17 17:41:17,866][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_6100_state/optimizer.bin +[2026-06-17 17:41:17,867][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_6100_state/sampler.bin +[2026-06-17 17:41:17,867][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_6100_state/sampler_1.bin +[2026-06-17 17:41:17,867][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_6100_state/random_states_0.pkl +[2026-06-17 17:41:17,912][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 6100 +[2026-06-17 17:41:17,913][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 17:41:17,951][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 17:41:22,701][starVLA.training.train_starvla][INFO] - Step 6101, Loss: {'train/loss': 0.7071368098258972, 'train/grad_norm_pre_clip': 16.771224975585938, 'timing/data': 0.004886589013040066, 'timing/model': 1.3230525039834902, 'learning_rate/qwen_vl_interface': 8.923851511687379e-07, 'learning_rate/action_model': 8.923851511687378e-06, 'epoch': 1.8}) +[2026-06-17 17:41:27,312][starVLA.training.train_starvla][INFO] - Step 6102, Loss: {'train/loss': 0.6386539340019226, 'train/grad_norm_pre_clip': 14.547723770141602, 'timing/data': 0.0044645360903814435, 'timing/model': 1.193160670925863, 'learning_rate/qwen_vl_interface': 8.915248878266228e-07, 'learning_rate/action_model': 8.915248878266227e-06, 'epoch': 1.8}) +[2026-06-17 17:41:32,046][starVLA.training.train_starvla][INFO] - Step 6103, Loss: {'train/loss': 0.7748710513114929, 'train/grad_norm_pre_clip': 16.59931182861328, 'timing/data': 0.005041107069700956, 'timing/model': 1.1918379859998822, 'learning_rate/qwen_vl_interface': 8.906655280010895e-07, 'learning_rate/action_model': 8.906655280010896e-06, 'epoch': 1.8}) +[2026-06-17 17:41:36,708][starVLA.training.train_starvla][INFO] - Step 6104, Loss: {'train/loss': 0.5340672731399536, 'train/grad_norm_pre_clip': 13.235692977905273, 'timing/data': 0.0048826070269569755, 'timing/model': 1.1928698429837823, 'learning_rate/qwen_vl_interface': 8.898070718702831e-07, 'learning_rate/action_model': 8.898070718702831e-06, 'epoch': 1.8}) +[2026-06-17 17:41:41,430][starVLA.training.train_starvla][INFO] - Step 6105, Loss: {'train/loss': 0.649774968624115, 'train/grad_norm_pre_clip': 37.78158950805664, 'timing/data': 0.005209209979511797, 'timing/model': 1.1952775419922546, 'learning_rate/qwen_vl_interface': 8.889495196121628e-07, 'learning_rate/action_model': 8.889495196121627e-06, 'epoch': 1.8}) +[2026-06-17 17:41:46,193][starVLA.training.train_starvla][INFO] - Step 6106, Loss: {'train/loss': 0.5931035876274109, 'train/grad_norm_pre_clip': 14.010465621948242, 'timing/data': 0.004888964002020657, 'timing/model': 1.2003709580749273, 'learning_rate/qwen_vl_interface': 8.880928714044994e-07, 'learning_rate/action_model': 8.880928714044993e-06, 'epoch': 1.8}) +[2026-06-17 17:41:50,966][starVLA.training.train_starvla][INFO] - Step 6107, Loss: {'train/loss': 0.5293467044830322, 'train/grad_norm_pre_clip': 12.588892936706543, 'timing/data': 0.004788308986462653, 'timing/model': 1.3250572769902647, 'learning_rate/qwen_vl_interface': 8.872371274248773e-07, 'learning_rate/action_model': 8.872371274248772e-06, 'epoch': 1.8}) +[2026-06-17 17:41:55,598][starVLA.training.train_starvla][INFO] - Step 6108, Loss: {'train/loss': 0.7616798877716064, 'train/grad_norm_pre_clip': 16.592721939086914, 'timing/data': 0.004964424995705485, 'timing/model': 1.199321851017885, 'learning_rate/qwen_vl_interface': 8.863822878506931e-07, 'learning_rate/action_model': 8.86382287850693e-06, 'epoch': 1.8}) +[2026-06-17 17:42:00,372][starVLA.training.train_starvla][INFO] - Step 6109, Loss: {'train/loss': 0.5053523182868958, 'train/grad_norm_pre_clip': 14.36782455444336, 'timing/data': 0.00497688795439899, 'timing/model': 1.2020214879885316, 'learning_rate/qwen_vl_interface': 8.855283528591566e-07, 'learning_rate/action_model': 8.855283528591565e-06, 'epoch': 1.8}) +[2026-06-17 17:42:05,133][starVLA.training.train_starvla][INFO] - Step 6110, Loss: {'train/loss': 0.6177365779876709, 'train/grad_norm_pre_clip': 13.771193504333496, 'timing/data': 0.004981112899258733, 'timing/model': 1.2043255709577352, 'learning_rate/qwen_vl_interface': 8.846753226272884e-07, 'learning_rate/action_model': 8.846753226272883e-06, 'epoch': 1.8}) +[2026-06-17 17:42:09,890][starVLA.training.train_starvla][INFO] - Step 6111, Loss: {'train/loss': 0.4586186707019806, 'train/grad_norm_pre_clip': 11.11733627319336, 'timing/data': 0.004896961967460811, 'timing/model': 1.2062417368870229, 'learning_rate/qwen_vl_interface': 8.838231973319235e-07, 'learning_rate/action_model': 8.838231973319235e-06, 'epoch': 1.8}) +[2026-06-17 17:42:14,669][starVLA.training.train_starvla][INFO] - Step 6112, Loss: {'train/loss': 0.7969356775283813, 'train/grad_norm_pre_clip': 17.129575729370117, 'timing/data': 0.004789500031620264, 'timing/model': 1.20602149493061, 'learning_rate/qwen_vl_interface': 8.829719771497072e-07, 'learning_rate/action_model': 8.829719771497071e-06, 'epoch': 1.8}) +[2026-06-17 17:42:19,447][starVLA.training.train_starvla][INFO] - Step 6113, Loss: {'train/loss': 1.0186736583709717, 'train/grad_norm_pre_clip': 18.3387508392334, 'timing/data': 0.004972712020389736, 'timing/model': 1.3331650319742039, 'learning_rate/qwen_vl_interface': 8.821216622570999e-07, 'learning_rate/action_model': 8.821216622570998e-06, 'epoch': 1.8}) +[2026-06-17 17:42:24,090][starVLA.training.train_starvla][INFO] - Step 6114, Loss: {'train/loss': 0.5583016872406006, 'train/grad_norm_pre_clip': 11.78548812866211, 'timing/data': 0.004735336056910455, 'timing/model': 1.2010977379977703, 'learning_rate/qwen_vl_interface': 8.812722528303727e-07, 'learning_rate/action_model': 8.812722528303727e-06, 'epoch': 1.8}) +[2026-06-17 17:42:28,872][starVLA.training.train_starvla][INFO] - Step 6115, Loss: {'train/loss': 0.6602585911750793, 'train/grad_norm_pre_clip': 13.296000480651855, 'timing/data': 0.004891043994575739, 'timing/model': 1.2046120719751343, 'learning_rate/qwen_vl_interface': 8.804237490456076e-07, 'learning_rate/action_model': 8.804237490456075e-06, 'epoch': 1.8}) +[2026-06-17 17:42:33,647][starVLA.training.train_starvla][INFO] - Step 6116, Loss: {'train/loss': 0.484531968832016, 'train/grad_norm_pre_clip': 12.341976165771484, 'timing/data': 0.004446285078302026, 'timing/model': 1.2050312240608037, 'learning_rate/qwen_vl_interface': 8.795761510787007e-07, 'learning_rate/action_model': 8.795761510787006e-06, 'epoch': 1.8}) +[2026-06-17 17:42:38,428][starVLA.training.train_starvla][INFO] - Step 6117, Loss: {'train/loss': 0.7812700271606445, 'train/grad_norm_pre_clip': 16.286556243896484, 'timing/data': 0.005054775043390691, 'timing/model': 1.205815396970138, 'learning_rate/qwen_vl_interface': 8.78729459105361e-07, 'learning_rate/action_model': 8.78729459105361e-06, 'epoch': 1.8}) +[2026-06-17 17:42:43,091][starVLA.training.train_starvla][INFO] - Step 6118, Loss: {'train/loss': 0.7713255286216736, 'train/grad_norm_pre_clip': 14.33611011505127, 'timing/data': 0.004895355901680887, 'timing/model': 1.2020929330028594, 'learning_rate/qwen_vl_interface': 8.778836733011079e-07, 'learning_rate/action_model': 8.778836733011079e-06, 'epoch': 1.8}) +[2026-06-17 17:42:47,784][starVLA.training.train_starvla][INFO] - Step 6119, Loss: {'train/loss': 0.7172489762306213, 'train/grad_norm_pre_clip': 18.23257827758789, 'timing/data': 0.005218711914494634, 'timing/model': 1.2434196359245107, 'learning_rate/qwen_vl_interface': 8.770387938412753e-07, 'learning_rate/action_model': 8.770387938412753e-06, 'epoch': 1.8}) +[2026-06-17 17:42:52,441][starVLA.training.train_starvla][INFO] - Step 6120, Loss: {'train/loss': 0.6645668148994446, 'train/grad_norm_pre_clip': 15.814898490905762, 'timing/data': 0.0050146509893238544, 'timing/model': 1.2031837900867686, 'learning_rate/qwen_vl_interface': 8.76194820901004e-07, 'learning_rate/action_model': 8.76194820901004e-06, 'epoch': 1.8}) +[2026-06-17 17:42:57,192][starVLA.training.train_starvla][INFO] - Step 6121, Loss: {'train/loss': 0.5824006795883179, 'train/grad_norm_pre_clip': 12.942152976989746, 'timing/data': 0.004601257969625294, 'timing/model': 1.2117794789373875, 'learning_rate/qwen_vl_interface': 8.753517546552545e-07, 'learning_rate/action_model': 8.753517546552544e-06, 'epoch': 1.8}) +[2026-06-17 17:43:01,988][starVLA.training.train_starvla][INFO] - Step 6122, Loss: {'train/loss': 0.7379515767097473, 'train/grad_norm_pre_clip': 17.047988891601562, 'timing/data': 0.004764834069646895, 'timing/model': 1.204306878033094, 'learning_rate/qwen_vl_interface': 8.74509595278793e-07, 'learning_rate/action_model': 8.74509595278793e-06, 'epoch': 1.8}) +[2026-06-17 17:43:06,790][starVLA.training.train_starvla][INFO] - Step 6123, Loss: {'train/loss': 0.5367345809936523, 'train/grad_norm_pre_clip': 15.15960693359375, 'timing/data': 0.004908294999040663, 'timing/model': 1.2064289280679077, 'learning_rate/qwen_vl_interface': 8.736683429462003e-07, 'learning_rate/action_model': 8.736683429462004e-06, 'epoch': 1.8}) +[2026-06-17 17:43:11,575][starVLA.training.train_starvla][INFO] - Step 6124, Loss: {'train/loss': 0.39965078234672546, 'train/grad_norm_pre_clip': 10.463589668273926, 'timing/data': 0.00486151606310159, 'timing/model': 1.3305057510733604, 'learning_rate/qwen_vl_interface': 8.728279978318696e-07, 'learning_rate/action_model': 8.728279978318695e-06, 'epoch': 1.8}) +[2026-06-17 17:43:16,281][starVLA.training.train_starvla][INFO] - Step 6125, Loss: {'train/loss': 0.7242581248283386, 'train/grad_norm_pre_clip': 14.65026569366455, 'timing/data': 0.004758928087539971, 'timing/model': 1.2549561989726499, 'learning_rate/qwen_vl_interface': 8.719885601100052e-07, 'learning_rate/action_model': 8.719885601100051e-06, 'epoch': 1.8}) +[2026-06-17 17:43:20,935][starVLA.training.train_starvla][INFO] - Step 6126, Loss: {'train/loss': 0.7698870301246643, 'train/grad_norm_pre_clip': 14.531889915466309, 'timing/data': 0.00517677899915725, 'timing/model': 1.203993960050866, 'learning_rate/qwen_vl_interface': 8.71150029954623e-07, 'learning_rate/action_model': 8.71150029954623e-06, 'epoch': 1.8}) +[2026-06-17 17:43:25,679][starVLA.training.train_starvla][INFO] - Step 6127, Loss: {'train/loss': 0.6974313855171204, 'train/grad_norm_pre_clip': 11.777364730834961, 'timing/data': 0.0046759420074522495, 'timing/model': 1.2084351290250197, 'learning_rate/qwen_vl_interface': 8.703124075395504e-07, 'learning_rate/action_model': 8.703124075395504e-06, 'epoch': 1.8}) +[2026-06-17 17:43:30,497][starVLA.training.train_starvla][INFO] - Step 6128, Loss: {'train/loss': 0.5789051651954651, 'train/grad_norm_pre_clip': 14.007343292236328, 'timing/data': 0.004917983082123101, 'timing/model': 1.2120210389839485, 'learning_rate/qwen_vl_interface': 8.694756930384305e-07, 'learning_rate/action_model': 8.694756930384306e-06, 'epoch': 1.8}) +[2026-06-17 17:43:35,296][starVLA.training.train_starvla][INFO] - Step 6129, Loss: {'train/loss': 0.7113008499145508, 'train/grad_norm_pre_clip': 15.83197021484375, 'timing/data': 0.004622193984687328, 'timing/model': 1.2068358180113137, 'learning_rate/qwen_vl_interface': 8.686398866247121e-07, 'learning_rate/action_model': 8.68639886624712e-06, 'epoch': 1.8}) +[2026-06-17 17:43:40,084][starVLA.training.train_starvla][INFO] - Step 6130, Loss: {'train/loss': 0.7099767327308655, 'train/grad_norm_pre_clip': 15.967122077941895, 'timing/data': 0.004743251018226147, 'timing/model': 1.3315293008927256, 'learning_rate/qwen_vl_interface': 8.6780498847166e-07, 'learning_rate/action_model': 8.678049884716599e-06, 'epoch': 1.8}) +[2026-06-17 17:43:44,855][starVLA.training.train_starvla][INFO] - Step 6131, Loss: {'train/loss': 0.6906809210777283, 'train/grad_norm_pre_clip': 16.192848205566406, 'timing/data': 0.005147140007466078, 'timing/model': 1.3215899310307577, 'learning_rate/qwen_vl_interface': 8.669709987523479e-07, 'learning_rate/action_model': 8.669709987523477e-06, 'epoch': 1.8}) +[2026-06-17 17:43:49,511][starVLA.training.train_starvla][INFO] - Step 6132, Loss: {'train/loss': 0.8102460503578186, 'train/grad_norm_pre_clip': 68.79175567626953, 'timing/data': 0.004711050074547529, 'timing/model': 1.2032160820672289, 'learning_rate/qwen_vl_interface': 8.661379176396651e-07, 'learning_rate/action_model': 8.66137917639665e-06, 'epoch': 1.8}) +[2026-06-17 17:43:54,283][starVLA.training.train_starvla][INFO] - Step 6133, Loss: {'train/loss': 0.4712865352630615, 'train/grad_norm_pre_clip': 10.099672317504883, 'timing/data': 0.004726277897134423, 'timing/model': 1.21144434902817, 'learning_rate/qwen_vl_interface': 8.653057453063092e-07, 'learning_rate/action_model': 8.653057453063091e-06, 'epoch': 1.8}) +[2026-06-17 17:43:59,069][starVLA.training.train_starvla][INFO] - Step 6134, Loss: {'train/loss': 0.7336753606796265, 'train/grad_norm_pre_clip': 14.662713050842285, 'timing/data': 0.005144237889908254, 'timing/model': 1.206779146916233, 'learning_rate/qwen_vl_interface': 8.644744819247908e-07, 'learning_rate/action_model': 8.644744819247908e-06, 'epoch': 1.8}) +[2026-06-17 17:44:03,847][starVLA.training.train_starvla][INFO] - Step 6135, Loss: {'train/loss': 0.532366156578064, 'train/grad_norm_pre_clip': 12.552160263061523, 'timing/data': 0.004809370031580329, 'timing/model': 1.2023334079422057, 'learning_rate/qwen_vl_interface': 8.636441276674308e-07, 'learning_rate/action_model': 8.636441276674308e-06, 'epoch': 1.81}) +[2026-06-17 17:44:08,567][starVLA.training.train_starvla][INFO] - Step 6136, Loss: {'train/loss': 0.5875586271286011, 'train/grad_norm_pre_clip': 15.090567588806152, 'timing/data': 0.0050794510170817375, 'timing/model': 1.2745908609358594, 'learning_rate/qwen_vl_interface': 8.628146827063632e-07, 'learning_rate/action_model': 8.628146827063633e-06, 'epoch': 1.81}) +[2026-06-17 17:44:13,228][starVLA.training.train_starvla][INFO] - Step 6137, Loss: {'train/loss': 0.693088710308075, 'train/grad_norm_pre_clip': 14.092835426330566, 'timing/data': 0.005139179993420839, 'timing/model': 1.2041613770416006, 'learning_rate/qwen_vl_interface': 8.619861472135332e-07, 'learning_rate/action_model': 8.619861472135332e-06, 'epoch': 1.81}) +[2026-06-17 17:44:18,079][starVLA.training.train_starvla][INFO] - Step 6138, Loss: {'train/loss': 0.9392414093017578, 'train/grad_norm_pre_clip': 16.92196273803711, 'timing/data': 0.005036463961005211, 'timing/model': 1.2037486500339583, 'learning_rate/qwen_vl_interface': 8.611585213606964e-07, 'learning_rate/action_model': 8.611585213606964e-06, 'epoch': 1.81}) +[2026-06-17 17:44:22,887][starVLA.training.train_starvla][INFO] - Step 6139, Loss: {'train/loss': 0.4070836305618286, 'train/grad_norm_pre_clip': 11.721755981445312, 'timing/data': 0.005189034971408546, 'timing/model': 1.2174829930299893, 'learning_rate/qwen_vl_interface': 8.603318053194206e-07, 'learning_rate/action_model': 8.603318053194205e-06, 'epoch': 1.81}) +[2026-06-17 17:44:27,681][starVLA.training.train_starvla][INFO] - Step 6140, Loss: {'train/loss': 0.6872564554214478, 'train/grad_norm_pre_clip': 15.625395774841309, 'timing/data': 0.004912368021905422, 'timing/model': 1.2108324150322005, 'learning_rate/qwen_vl_interface': 8.595059992610846e-07, 'learning_rate/action_model': 8.595059992610846e-06, 'epoch': 1.81}) +[2026-06-17 17:44:32,469][starVLA.training.train_starvla][INFO] - Step 6141, Loss: {'train/loss': 0.5723013877868652, 'train/grad_norm_pre_clip': 12.103707313537598, 'timing/data': 0.004678578930906951, 'timing/model': 1.3320610329974443, 'learning_rate/qwen_vl_interface': 8.586811033568796e-07, 'learning_rate/action_model': 8.586811033568797e-06, 'epoch': 1.81}) +[2026-06-17 17:44:37,347][starVLA.training.train_starvla][INFO] - Step 6142, Loss: {'train/loss': 0.7225815057754517, 'train/grad_norm_pre_clip': 14.20710563659668, 'timing/data': 0.0047835169825702906, 'timing/model': 1.2913335229968652, 'learning_rate/qwen_vl_interface': 8.578571177778076e-07, 'learning_rate/action_model': 8.578571177778075e-06, 'epoch': 1.81}) +[2026-06-17 17:44:42,000][starVLA.training.train_starvla][INFO] - Step 6143, Loss: {'train/loss': 0.8292991518974304, 'train/grad_norm_pre_clip': 15.408771514892578, 'timing/data': 0.0048134030075743794, 'timing/model': 1.2037609070539474, 'learning_rate/qwen_vl_interface': 8.570340426946805e-07, 'learning_rate/action_model': 8.570340426946804e-06, 'epoch': 1.81}) +[2026-06-17 17:44:46,782][starVLA.training.train_starvla][INFO] - Step 6144, Loss: {'train/loss': 0.7058948278427124, 'train/grad_norm_pre_clip': 14.08649730682373, 'timing/data': 0.004790488048456609, 'timing/model': 1.2038057980826125, 'learning_rate/qwen_vl_interface': 8.562118782781241e-07, 'learning_rate/action_model': 8.56211878278124e-06, 'epoch': 1.81}) +[2026-06-17 17:44:51,581][starVLA.training.train_starvla][INFO] - Step 6145, Loss: {'train/loss': 0.9423320293426514, 'train/grad_norm_pre_clip': 17.32320213317871, 'timing/data': 0.004911020980216563, 'timing/model': 1.2107517560943961, 'learning_rate/qwen_vl_interface': 8.553906246985722e-07, 'learning_rate/action_model': 8.553906246985721e-06, 'epoch': 1.81}) +[2026-06-17 17:44:56,389][starVLA.training.train_starvla][INFO] - Step 6146, Loss: {'train/loss': 0.5067105889320374, 'train/grad_norm_pre_clip': 14.48192310333252, 'timing/data': 0.004696197924204171, 'timing/model': 1.204485985916108, 'learning_rate/qwen_vl_interface': 8.545702821262733e-07, 'learning_rate/action_model': 8.545702821262733e-06, 'epoch': 1.81}) +[2026-06-17 17:45:01,175][starVLA.training.train_starvla][INFO] - Step 6147, Loss: {'train/loss': 0.5240978598594666, 'train/grad_norm_pre_clip': 14.171138763427734, 'timing/data': 0.0048417269717901945, 'timing/model': 1.216193511034362, 'learning_rate/qwen_vl_interface': 8.537508507312844e-07, 'learning_rate/action_model': 8.537508507312844e-06, 'epoch': 1.81}) +[2026-06-17 17:45:05,932][starVLA.training.train_starvla][INFO] - Step 6148, Loss: {'train/loss': 0.43721893429756165, 'train/grad_norm_pre_clip': 9.991293907165527, 'timing/data': 0.004864587099291384, 'timing/model': 1.3033745160792023, 'learning_rate/qwen_vl_interface': 8.52932330683476e-07, 'learning_rate/action_model': 8.529323306834759e-06, 'epoch': 1.81}) +[2026-06-17 17:45:10,587][starVLA.training.train_starvla][INFO] - Step 6149, Loss: {'train/loss': 0.9022420048713684, 'train/grad_norm_pre_clip': 20.363479614257812, 'timing/data': 0.004864510032348335, 'timing/model': 1.20456102991011, 'learning_rate/qwen_vl_interface': 8.52114722152525e-07, 'learning_rate/action_model': 8.521147221525249e-06, 'epoch': 1.81}) +[2026-06-17 17:45:15,371][starVLA.training.train_starvla][INFO] - Step 6150, Loss: {'train/loss': 0.4288170635700226, 'train/grad_norm_pre_clip': 10.97380542755127, 'timing/data': 0.004870097967796028, 'timing/model': 1.2045769130345434, 'learning_rate/qwen_vl_interface': 8.51298025307925e-07, 'learning_rate/action_model': 8.51298025307925e-06, 'epoch': 1.81}) +[2026-06-17 17:45:20,164][starVLA.training.train_starvla][INFO] - Step 6151, Loss: {'train/loss': 0.6056693196296692, 'train/grad_norm_pre_clip': 29.99323272705078, 'timing/data': 0.004987371037714183, 'timing/model': 1.205406442983076, 'learning_rate/qwen_vl_interface': 8.504822403189783e-07, 'learning_rate/action_model': 8.504822403189782e-06, 'epoch': 1.81}) +[2026-06-17 17:45:24,971][starVLA.training.train_starvla][INFO] - Step 6152, Loss: {'train/loss': 0.5235779881477356, 'train/grad_norm_pre_clip': 11.412325859069824, 'timing/data': 0.004656772944144905, 'timing/model': 1.2060716599225998, 'learning_rate/qwen_vl_interface': 8.496673673547967e-07, 'learning_rate/action_model': 8.496673673547968e-06, 'epoch': 1.81}) +[2026-06-17 17:45:29,766][starVLA.training.train_starvla][INFO] - Step 6153, Loss: {'train/loss': 0.528011679649353, 'train/grad_norm_pre_clip': 12.130668640136719, 'timing/data': 0.004900317988358438, 'timing/model': 1.3348130310187116, 'learning_rate/qwen_vl_interface': 8.488534065843046e-07, 'learning_rate/action_model': 8.488534065843045e-06, 'epoch': 1.81}) +[2026-06-17 17:45:34,420][starVLA.training.train_starvla][INFO] - Step 6154, Loss: {'train/loss': 0.7615804672241211, 'train/grad_norm_pre_clip': 14.615568161010742, 'timing/data': 0.004673319053836167, 'timing/model': 1.2049830229952931, 'learning_rate/qwen_vl_interface': 8.480403581762378e-07, 'learning_rate/action_model': 8.480403581762378e-06, 'epoch': 1.81}) +[2026-06-17 17:45:39,208][starVLA.training.train_starvla][INFO] - Step 6155, Loss: {'train/loss': 0.8465628027915955, 'train/grad_norm_pre_clip': 23.78523063659668, 'timing/data': 0.004581105895340443, 'timing/model': 1.2050984019879252, 'learning_rate/qwen_vl_interface': 8.472282222991411e-07, 'learning_rate/action_model': 8.47228222299141e-06, 'epoch': 1.81}) +[2026-06-17 17:45:43,995][starVLA.training.train_starvla][INFO] - Step 6156, Loss: {'train/loss': 0.8862183094024658, 'train/grad_norm_pre_clip': 18.161941528320312, 'timing/data': 0.004759871982969344, 'timing/model': 1.2053895379649475, 'learning_rate/qwen_vl_interface': 8.464169991213705e-07, 'learning_rate/action_model': 8.464169991213706e-06, 'epoch': 1.81}) +[2026-06-17 17:45:48,717][starVLA.training.train_starvla][INFO] - Step 6157, Loss: {'train/loss': 0.507293164730072, 'train/grad_norm_pre_clip': 13.088102340698242, 'timing/data': 0.0047150650061666965, 'timing/model': 1.2034447670448571, 'learning_rate/qwen_vl_interface': 8.456066888110964e-07, 'learning_rate/action_model': 8.456066888110964e-06, 'epoch': 1.81}) +[2026-06-17 17:45:53,480][starVLA.training.train_starvla][INFO] - Step 6158, Loss: {'train/loss': 0.4030928909778595, 'train/grad_norm_pre_clip': 18.60678482055664, 'timing/data': 0.004724520957097411, 'timing/model': 1.2047440810129046, 'learning_rate/qwen_vl_interface': 8.447972915362941e-07, 'learning_rate/action_model': 8.44797291536294e-06, 'epoch': 1.81}) +[2026-06-17 17:45:58,296][starVLA.training.train_starvla][INFO] - Step 6159, Loss: {'train/loss': 0.6141908764839172, 'train/grad_norm_pre_clip': 20.853775024414062, 'timing/data': 0.004788265912793577, 'timing/model': 1.3520832499489188, 'learning_rate/qwen_vl_interface': 8.439888074647536e-07, 'learning_rate/action_model': 8.439888074647536e-06, 'epoch': 1.81}) +[2026-06-17 17:46:02,999][starVLA.training.train_starvla][INFO] - Step 6160, Loss: {'train/loss': 0.42584431171417236, 'train/grad_norm_pre_clip': 11.436735153198242, 'timing/data': 0.0047996509820222855, 'timing/model': 1.2434298950247467, 'learning_rate/qwen_vl_interface': 8.431812367640732e-07, 'learning_rate/action_model': 8.431812367640731e-06, 'epoch': 1.81}) +[2026-06-17 17:46:07,819][starVLA.training.train_starvla][INFO] - Step 6161, Loss: {'train/loss': 0.7249435782432556, 'train/grad_norm_pre_clip': 11.955760955810547, 'timing/data': 0.005086533026769757, 'timing/model': 1.212699203984812, 'learning_rate/qwen_vl_interface': 8.423745796016659e-07, 'learning_rate/action_model': 8.42374579601666e-06, 'epoch': 1.81}) +[2026-06-17 17:46:12,615][starVLA.training.train_starvla][INFO] - Step 6162, Loss: {'train/loss': 0.9071987867355347, 'train/grad_norm_pre_clip': 14.816866874694824, 'timing/data': 0.0049917129799723625, 'timing/model': 1.203093321993947, 'learning_rate/qwen_vl_interface': 8.41568836144751e-07, 'learning_rate/action_model': 8.41568836144751e-06, 'epoch': 1.81}) +[2026-06-17 17:46:17,412][starVLA.training.train_starvla][INFO] - Step 6163, Loss: {'train/loss': 0.7193867564201355, 'train/grad_norm_pre_clip': 14.681587219238281, 'timing/data': 0.004933541989885271, 'timing/model': 1.202898174058646, 'learning_rate/qwen_vl_interface': 8.407640065603601e-07, 'learning_rate/action_model': 8.4076400656036e-06, 'epoch': 1.81}) +[2026-06-17 17:46:22,155][starVLA.training.train_starvla][INFO] - Step 6164, Loss: {'train/loss': 0.5081915259361267, 'train/grad_norm_pre_clip': 13.878096580505371, 'timing/data': 0.004976960946805775, 'timing/model': 1.2081061749486253, 'learning_rate/qwen_vl_interface': 8.399600910153336e-07, 'learning_rate/action_model': 8.399600910153336e-06, 'epoch': 1.81}) +[2026-06-17 17:46:26,939][starVLA.training.train_starvla][INFO] - Step 6165, Loss: {'train/loss': 0.8725700378417969, 'train/grad_norm_pre_clip': 18.442855834960938, 'timing/data': 0.004968016990460455, 'timing/model': 1.323096806067042, 'learning_rate/qwen_vl_interface': 8.391570896763266e-07, 'learning_rate/action_model': 8.391570896763267e-06, 'epoch': 1.81}) +[2026-06-17 17:46:31,602][starVLA.training.train_starvla][INFO] - Step 6166, Loss: {'train/loss': 0.7034468054771423, 'train/grad_norm_pre_clip': 15.67099666595459, 'timing/data': 0.004867798998020589, 'timing/model': 1.2083603329956532, 'learning_rate/qwen_vl_interface': 8.38355002709802e-07, 'learning_rate/action_model': 8.38355002709802e-06, 'epoch': 1.81}) +[2026-06-17 17:46:36,407][starVLA.training.train_starvla][INFO] - Step 6167, Loss: {'train/loss': 0.37161433696746826, 'train/grad_norm_pre_clip': 9.293245315551758, 'timing/data': 0.005160150001756847, 'timing/model': 1.2041209290036932, 'learning_rate/qwen_vl_interface': 8.375538302820314e-07, 'learning_rate/action_model': 8.375538302820315e-06, 'epoch': 1.81}) +[2026-06-17 17:46:41,241][starVLA.training.train_starvla][INFO] - Step 6168, Loss: {'train/loss': 0.6279773116111755, 'train/grad_norm_pre_clip': 15.556705474853516, 'timing/data': 0.004910286981612444, 'timing/model': 1.2040850629564375, 'learning_rate/qwen_vl_interface': 8.367535725591003e-07, 'learning_rate/action_model': 8.367535725591003e-06, 'epoch': 1.81}) +[2026-06-17 17:46:46,043][starVLA.training.train_starvla][INFO] - Step 6169, Loss: {'train/loss': 0.8116288781166077, 'train/grad_norm_pre_clip': 15.608386993408203, 'timing/data': 0.005004609934985638, 'timing/model': 1.2078970259753987, 'learning_rate/qwen_vl_interface': 8.359542297069024e-07, 'learning_rate/action_model': 8.359542297069024e-06, 'epoch': 1.82}) +[2026-06-17 17:46:50,831][starVLA.training.train_starvla][INFO] - Step 6170, Loss: {'train/loss': 0.7871584296226501, 'train/grad_norm_pre_clip': 16.71219825744629, 'timing/data': 0.005022166995331645, 'timing/model': 1.33545127697289, 'learning_rate/qwen_vl_interface': 8.351558018911423e-07, 'learning_rate/action_model': 8.351558018911423e-06, 'epoch': 1.82}) +[2026-06-17 17:46:55,503][starVLA.training.train_starvla][INFO] - Step 6171, Loss: {'train/loss': 0.6271955966949463, 'train/grad_norm_pre_clip': 23.112693786621094, 'timing/data': 0.004684593994170427, 'timing/model': 1.2019185590324923, 'learning_rate/qwen_vl_interface': 8.343582892773352e-07, 'learning_rate/action_model': 8.343582892773351e-06, 'epoch': 1.82}) +[2026-06-17 17:47:00,293][starVLA.training.train_starvla][INFO] - Step 6172, Loss: {'train/loss': 0.5475660562515259, 'train/grad_norm_pre_clip': 13.042096138000488, 'timing/data': 0.004905323032289743, 'timing/model': 1.2055572039680555, 'learning_rate/qwen_vl_interface': 8.335616920308061e-07, 'learning_rate/action_model': 8.33561692030806e-06, 'epoch': 1.82}) +[2026-06-17 17:47:05,050][starVLA.training.train_starvla][INFO] - Step 6173, Loss: {'train/loss': 0.6650326251983643, 'train/grad_norm_pre_clip': 15.249408721923828, 'timing/data': 0.005401707021519542, 'timing/model': 1.2072453630389646, 'learning_rate/qwen_vl_interface': 8.327660103166907e-07, 'learning_rate/action_model': 8.327660103166906e-06, 'epoch': 1.82}) +[2026-06-17 17:47:09,798][starVLA.training.train_starvla][INFO] - Step 6174, Loss: {'train/loss': 0.6280133128166199, 'train/grad_norm_pre_clip': 18.377758026123047, 'timing/data': 0.00501783296931535, 'timing/model': 1.204182503046468, 'learning_rate/qwen_vl_interface': 8.319712442999335e-07, 'learning_rate/action_model': 8.319712442999335e-06, 'epoch': 1.82}) +[2026-06-17 17:47:14,591][starVLA.training.train_starvla][INFO] - Step 6175, Loss: {'train/loss': 0.5910936594009399, 'train/grad_norm_pre_clip': 12.207930564880371, 'timing/data': 0.0050890629645437, 'timing/model': 1.2059866430936381, 'learning_rate/qwen_vl_interface': 8.311773941452924e-07, 'learning_rate/action_model': 8.311773941452925e-06, 'epoch': 1.82}) +[2026-06-17 17:47:19,380][starVLA.training.train_starvla][INFO] - Step 6176, Loss: {'train/loss': 0.4206438958644867, 'train/grad_norm_pre_clip': 12.342912673950195, 'timing/data': 0.004662460996769369, 'timing/model': 1.3346987259574234, 'learning_rate/qwen_vl_interface': 8.303844600173326e-07, 'learning_rate/action_model': 8.303844600173325e-06, 'epoch': 1.82}) +[2026-06-17 17:47:24,039][starVLA.training.train_starvla][INFO] - Step 6177, Loss: {'train/loss': 0.7091227769851685, 'train/grad_norm_pre_clip': 14.005760192871094, 'timing/data': 0.004858488100580871, 'timing/model': 1.2054533390328288, 'learning_rate/qwen_vl_interface': 8.295924420804297e-07, 'learning_rate/action_model': 8.295924420804296e-06, 'epoch': 1.82}) +[2026-06-17 17:47:28,846][starVLA.training.train_starvla][INFO] - Step 6178, Loss: {'train/loss': 0.6637747287750244, 'train/grad_norm_pre_clip': 13.49738883972168, 'timing/data': 0.004462215001694858, 'timing/model': 1.2241501200478524, 'learning_rate/qwen_vl_interface': 8.288013404987691e-07, 'learning_rate/action_model': 8.28801340498769e-06, 'epoch': 1.82}) +[2026-06-17 17:47:33,619][starVLA.training.train_starvla][INFO] - Step 6179, Loss: {'train/loss': 0.5218140482902527, 'train/grad_norm_pre_clip': 10.063661575317383, 'timing/data': 0.0047430540435016155, 'timing/model': 1.2030003960244358, 'learning_rate/qwen_vl_interface': 8.28011155436349e-07, 'learning_rate/action_model': 8.28011155436349e-06, 'epoch': 1.82}) +[2026-06-17 17:47:38,360][starVLA.training.train_starvla][INFO] - Step 6180, Loss: {'train/loss': 0.8739983439445496, 'train/grad_norm_pre_clip': 19.313810348510742, 'timing/data': 0.0048023570561781526, 'timing/model': 1.205858270986937, 'learning_rate/qwen_vl_interface': 8.272218870569748e-07, 'learning_rate/action_model': 8.272218870569748e-06, 'epoch': 1.82}) +[2026-06-17 17:47:43,149][starVLA.training.train_starvla][INFO] - Step 6181, Loss: {'train/loss': 0.6437387466430664, 'train/grad_norm_pre_clip': 20.36322021484375, 'timing/data': 0.00519447005353868, 'timing/model': 1.2043403110001236, 'learning_rate/qwen_vl_interface': 8.264335355242625e-07, 'learning_rate/action_model': 8.264335355242625e-06, 'epoch': 1.82}) +[2026-06-17 17:47:47,959][starVLA.training.train_starvla][INFO] - Step 6182, Loss: {'train/loss': 0.48489081859588623, 'train/grad_norm_pre_clip': 12.255258560180664, 'timing/data': 0.004676848999224603, 'timing/model': 1.3482491730246693, 'learning_rate/qwen_vl_interface': 8.256461010016376e-07, 'learning_rate/action_model': 8.256461010016376e-06, 'epoch': 1.82}) +[2026-06-17 17:47:52,631][starVLA.training.train_starvla][INFO] - Step 6183, Loss: {'train/loss': 0.5674710869789124, 'train/grad_norm_pre_clip': 16.83573341369629, 'timing/data': 0.004708949010819197, 'timing/model': 1.2144588420633227, 'learning_rate/qwen_vl_interface': 8.248595836523368e-07, 'learning_rate/action_model': 8.248595836523368e-06, 'epoch': 1.82}) +[2026-06-17 17:47:57,448][starVLA.training.train_starvla][INFO] - Step 6184, Loss: {'train/loss': 0.6022132039070129, 'train/grad_norm_pre_clip': 12.777419090270996, 'timing/data': 0.004805552074685693, 'timing/model': 1.2046360929962248, 'learning_rate/qwen_vl_interface': 8.240739836394068e-07, 'learning_rate/action_model': 8.240739836394068e-06, 'epoch': 1.82}) +[2026-06-17 17:48:02,261][starVLA.training.train_starvla][INFO] - Step 6185, Loss: {'train/loss': 0.4561249613761902, 'train/grad_norm_pre_clip': 11.436905860900879, 'timing/data': 0.005318337003700435, 'timing/model': 1.2045842410298064, 'learning_rate/qwen_vl_interface': 8.232893011257019e-07, 'learning_rate/action_model': 8.232893011257018e-06, 'epoch': 1.82}) +[2026-06-17 17:48:07,041][starVLA.training.train_starvla][INFO] - Step 6186, Loss: {'train/loss': 0.5531786680221558, 'train/grad_norm_pre_clip': 14.666472434997559, 'timing/data': 0.0048094469821080565, 'timing/model': 1.2057234039530158, 'learning_rate/qwen_vl_interface': 8.225055362738886e-07, 'learning_rate/action_model': 8.225055362738885e-06, 'epoch': 1.82}) +[2026-06-17 17:48:11,727][starVLA.training.train_starvla][INFO] - Step 6187, Loss: {'train/loss': 0.6968734860420227, 'train/grad_norm_pre_clip': 16.25809669494629, 'timing/data': 0.0051739850314334035, 'timing/model': 1.205271241022274, 'learning_rate/qwen_vl_interface': 8.217226892464411e-07, 'learning_rate/action_model': 8.21722689246441e-06, 'epoch': 1.82}) +[2026-06-17 17:48:16,425][starVLA.training.train_starvla][INFO] - Step 6188, Loss: {'train/loss': 0.4487234055995941, 'train/grad_norm_pre_clip': 16.328874588012695, 'timing/data': 0.004529498051851988, 'timing/model': 1.245961518958211, 'learning_rate/qwen_vl_interface': 8.209407602056457e-07, 'learning_rate/action_model': 8.209407602056457e-06, 'epoch': 1.82}) +[2026-06-17 17:48:21,086][starVLA.training.train_starvla][INFO] - Step 6189, Loss: {'train/loss': 0.5692104697227478, 'train/grad_norm_pre_clip': 15.771711349487305, 'timing/data': 0.004949640948325396, 'timing/model': 1.2046317480271682, 'learning_rate/qwen_vl_interface': 8.201597493135962e-07, 'learning_rate/action_model': 8.201597493135961e-06, 'epoch': 1.82}) +[2026-06-17 17:48:25,861][starVLA.training.train_starvla][INFO] - Step 6190, Loss: {'train/loss': 0.9159122705459595, 'train/grad_norm_pre_clip': 14.971187591552734, 'timing/data': 0.004845460061915219, 'timing/model': 1.2088136960519478, 'learning_rate/qwen_vl_interface': 8.193796567321982e-07, 'learning_rate/action_model': 8.193796567321982e-06, 'epoch': 1.82}) +[2026-06-17 17:48:30,669][starVLA.training.train_starvla][INFO] - Step 6191, Loss: {'train/loss': 0.7657142281532288, 'train/grad_norm_pre_clip': 14.555917739868164, 'timing/data': 0.0065841940231621265, 'timing/model': 1.213025221019052, 'learning_rate/qwen_vl_interface': 8.186004826231644e-07, 'learning_rate/action_model': 8.186004826231644e-06, 'epoch': 1.82}) +[2026-06-17 17:48:35,467][starVLA.training.train_starvla][INFO] - Step 6192, Loss: {'train/loss': 0.6231723427772522, 'train/grad_norm_pre_clip': 12.213238716125488, 'timing/data': 0.004969364032149315, 'timing/model': 1.2063446510583162, 'learning_rate/qwen_vl_interface': 8.178222271480192e-07, 'learning_rate/action_model': 8.178222271480191e-06, 'epoch': 1.82}) +[2026-06-17 17:48:40,259][starVLA.training.train_starvla][INFO] - Step 6193, Loss: {'train/loss': 0.5441248416900635, 'train/grad_norm_pre_clip': 15.219673156738281, 'timing/data': 0.004667020053602755, 'timing/model': 1.3327434010570869, 'learning_rate/qwen_vl_interface': 8.170448904680943e-07, 'learning_rate/action_model': 8.170448904680943e-06, 'epoch': 1.82}) +[2026-06-17 17:48:44,966][starVLA.training.train_starvla][INFO] - Step 6194, Loss: {'train/loss': 0.599010705947876, 'train/grad_norm_pre_clip': 16.814416885375977, 'timing/data': 0.005427105003036559, 'timing/model': 1.2540752660715953, 'learning_rate/qwen_vl_interface': 8.162684727445347e-07, 'learning_rate/action_model': 8.162684727445346e-06, 'epoch': 1.82}) +[2026-06-17 17:48:49,622][starVLA.training.train_starvla][INFO] - Step 6195, Loss: {'train/loss': 0.7532756328582764, 'train/grad_norm_pre_clip': 15.91846752166748, 'timing/data': 0.005026277038268745, 'timing/model': 1.2045745300129056, 'learning_rate/qwen_vl_interface': 8.154929741382922e-07, 'learning_rate/action_model': 8.154929741382922e-06, 'epoch': 1.82}) +[2026-06-17 17:48:54,387][starVLA.training.train_starvla][INFO] - Step 6196, Loss: {'train/loss': 0.8731040954589844, 'train/grad_norm_pre_clip': 15.356915473937988, 'timing/data': 0.00498109997715801, 'timing/model': 1.2122295059962198, 'learning_rate/qwen_vl_interface': 8.147183948101266e-07, 'learning_rate/action_model': 8.147183948101265e-06, 'epoch': 1.82}) +[2026-06-17 17:48:59,204][starVLA.training.train_starvla][INFO] - Step 6197, Loss: {'train/loss': 0.5208156704902649, 'train/grad_norm_pre_clip': 13.119649887084961, 'timing/data': 0.005133038968779147, 'timing/model': 1.2142865650821477, 'learning_rate/qwen_vl_interface': 8.139447349206111e-07, 'learning_rate/action_model': 8.139447349206111e-06, 'epoch': 1.82}) +[2026-06-17 17:49:04,004][starVLA.training.train_starvla][INFO] - Step 6198, Loss: {'train/loss': 0.6305596828460693, 'train/grad_norm_pre_clip': 13.99837589263916, 'timing/data': 0.0048064509173855186, 'timing/model': 1.2079053409397602, 'learning_rate/qwen_vl_interface': 8.131719946301255e-07, 'learning_rate/action_model': 8.131719946301255e-06, 'epoch': 1.82}) +[2026-06-17 17:49:08,791][starVLA.training.train_starvla][INFO] - Step 6199, Loss: {'train/loss': 0.6482403874397278, 'train/grad_norm_pre_clip': 14.350341796875, 'timing/data': 0.005124164978042245, 'timing/model': 1.3345167440129444, 'learning_rate/qwen_vl_interface': 8.124001740988595e-07, 'learning_rate/action_model': 8.124001740988595e-06, 'epoch': 1.82}) +[2026-06-17 17:50:50,734][starVLA.training.train_starvla][INFO] - Step 6200, Loss: {'train/loss': 0.5317220687866211, 'train/grad_norm_pre_clip': 13.912686347961426, 'eval/loss': 0.7065042853355408, 'eval/latency_0/loss': 0.7065042853355408, 'eval/demon_fire_precision': 0.8984932593180016, 'eval/demon_fire_recall': 0.9410299003322259, 'eval/demon_fire_f1': 0.9192697768762678, 'eval/demon_fire_rate_model': 0.7013348164627363, 'eval/demon_fire_rate_teacher': 0.6696329254727474, 'eval/demon_left_precision': 0.8347927612375948, 'eval/demon_left_recall': 0.8101983002832861, 'eval/demon_left_f1': 0.8223116733755031, 'eval/demon_left_rate_model': 0.47636262513904337, 'eval/demon_left_rate_teacher': 0.49082313681868744, 'eval/demon_right_precision': 0.8204989154013015, 'eval/demon_right_recall': 0.8548022598870056, 'eval/demon_right_f1': 0.8372993912562259, 'eval/demon_right_rate_model': 0.5127919911012235, 'eval/demon_right_rate_teacher': 0.4922135706340378, 'eval/demon_move_f1': 0.8298055323158645, 'eval/demon_cc_f1': 0.8904684199113656, 'eval/latency_0/demon_fire_precision': 0.8984932593180016, 'eval/latency_0/demon_fire_recall': 0.9410299003322259, 'eval/latency_0/demon_fire_f1': 0.9192697768762678, 'eval/latency_0/demon_fire_rate_model': 0.7013348164627363, 'eval/latency_0/demon_fire_rate_teacher': 0.6696329254727474, 'eval/latency_0/demon_left_precision': 0.8347927612375948, 'eval/latency_0/demon_left_recall': 0.8101983002832861, 'eval/latency_0/demon_left_f1': 0.8223116733755031, 'eval/latency_0/demon_left_rate_model': 0.47636262513904337, 'eval/latency_0/demon_left_rate_teacher': 0.49082313681868744, 'eval/latency_0/demon_right_precision': 0.8204989154013015, 'eval/latency_0/demon_right_recall': 0.8548022598870056, 'eval/latency_0/demon_right_f1': 0.8372993912562259, 'eval/latency_0/demon_right_rate_model': 0.5127919911012235, 'eval/latency_0/demon_right_rate_teacher': 0.4922135706340378, 'eval/latency_0/demon_move_f1': 0.8298055323158645, 'eval/latency_0/demon_cc_f1': 0.8904684199113656, 'timing/data': 0.005252563045360148, 'timing/model': 1.3229316259967163, 'learning_rate/qwen_vl_interface': 8.116292734868126e-07, 'learning_rate/action_model': 8.116292734868125e-06, 'epoch': 1.82}) +[2026-06-17 17:50:50,735][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_6200_state +[2026-06-17 17:50:50,738][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 17:50:54,276][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_6200_state/model.safetensors +[2026-06-17 17:51:01,979][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_6200_state/optimizer.bin +[2026-06-17 17:51:01,980][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_6200_state/sampler.bin +[2026-06-17 17:51:01,980][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_6200_state/sampler_1.bin +[2026-06-17 17:51:01,981][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_6200_state/random_states_0.pkl +[2026-06-17 17:51:02,032][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 6200 +[2026-06-17 17:51:02,032][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 17:51:02,060][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 17:51:06,779][starVLA.training.train_starvla][INFO] - Step 6201, Loss: {'train/loss': 0.6584169268608093, 'train/grad_norm_pre_clip': 18.56327247619629, 'timing/data': 0.005009649903513491, 'timing/model': 1.3195522689493373, 'learning_rate/qwen_vl_interface': 8.108592929537928e-07, 'learning_rate/action_model': 8.108592929537928e-06, 'epoch': 1.82}) +[2026-06-17 17:51:11,373][starVLA.training.train_starvla][INFO] - Step 6202, Loss: {'train/loss': 0.5920107960700989, 'train/grad_norm_pre_clip': 14.56428337097168, 'timing/data': 0.005040978896431625, 'timing/model': 1.191567045985721, 'learning_rate/qwen_vl_interface': 8.100902326594186e-07, 'learning_rate/action_model': 8.100902326594186e-06, 'epoch': 1.82}) +[2026-06-17 17:51:16,111][starVLA.training.train_starvla][INFO] - Step 6203, Loss: {'train/loss': 0.5634247064590454, 'train/grad_norm_pre_clip': 14.266287803649902, 'timing/data': 0.004651747993193567, 'timing/model': 1.1948950060177594, 'learning_rate/qwen_vl_interface': 8.093220927631171e-07, 'learning_rate/action_model': 8.09322092763117e-06, 'epoch': 1.83}) +[2026-06-17 17:51:20,737][starVLA.training.train_starvla][INFO] - Step 6204, Loss: {'train/loss': 0.5268078446388245, 'train/grad_norm_pre_clip': 11.89442253112793, 'timing/data': 0.004839707049541175, 'timing/model': 1.1965215890668333, 'learning_rate/qwen_vl_interface': 8.085548734241233e-07, 'learning_rate/action_model': 8.085548734241232e-06, 'epoch': 1.83}) +[2026-06-17 17:51:25,392][starVLA.training.train_starvla][INFO] - Step 6205, Loss: {'train/loss': 0.6267266869544983, 'train/grad_norm_pre_clip': 18.27816390991211, 'timing/data': 0.005107708042487502, 'timing/model': 1.197376579977572, 'learning_rate/qwen_vl_interface': 8.077885748014843e-07, 'learning_rate/action_model': 8.077885748014841e-06, 'epoch': 1.83}) +[2026-06-17 17:51:30,067][starVLA.training.train_starvla][INFO] - Step 6206, Loss: {'train/loss': 0.8053807616233826, 'train/grad_norm_pre_clip': 15.777215957641602, 'timing/data': 0.004951626993715763, 'timing/model': 1.1969821830280125, 'learning_rate/qwen_vl_interface': 8.070231970540532e-07, 'learning_rate/action_model': 8.070231970540532e-06, 'epoch': 1.83}) +[2026-06-17 17:51:34,769][starVLA.training.train_starvla][INFO] - Step 6207, Loss: {'train/loss': 0.9017400741577148, 'train/grad_norm_pre_clip': 18.818283081054688, 'timing/data': 0.00477659294847399, 'timing/model': 1.2734390340046957, 'learning_rate/qwen_vl_interface': 8.062587403404934e-07, 'learning_rate/action_model': 8.062587403404934e-06, 'epoch': 1.83}) +[2026-06-17 17:51:39,401][starVLA.training.train_starvla][INFO] - Step 6208, Loss: {'train/loss': 0.901238203048706, 'train/grad_norm_pre_clip': 17.985727310180664, 'timing/data': 0.00515030394308269, 'timing/model': 1.1993221020093188, 'learning_rate/qwen_vl_interface': 8.054952048192793e-07, 'learning_rate/action_model': 8.054952048192794e-06, 'epoch': 1.83}) +[2026-06-17 17:51:44,163][starVLA.training.train_starvla][INFO] - Step 6209, Loss: {'train/loss': 0.6140616536140442, 'train/grad_norm_pre_clip': 14.189684867858887, 'timing/data': 0.005261547979898751, 'timing/model': 1.197632604977116, 'learning_rate/qwen_vl_interface': 8.047325906486926e-07, 'learning_rate/action_model': 8.047325906486925e-06, 'epoch': 1.83}) +[2026-06-17 17:51:48,925][starVLA.training.train_starvla][INFO] - Step 6210, Loss: {'train/loss': 0.4865938723087311, 'train/grad_norm_pre_clip': 14.886380195617676, 'timing/data': 0.004781494033522904, 'timing/model': 1.1984325570520014, 'learning_rate/qwen_vl_interface': 8.039708979868229e-07, 'learning_rate/action_model': 8.039708979868227e-06, 'epoch': 1.83}) +[2026-06-17 17:51:53,697][starVLA.training.train_starvla][INFO] - Step 6211, Loss: {'train/loss': 1.0219223499298096, 'train/grad_norm_pre_clip': 18.341693878173828, 'timing/data': 0.004750272957608104, 'timing/model': 1.199729196028784, 'learning_rate/qwen_vl_interface': 8.032101269915687e-07, 'learning_rate/action_model': 8.032101269915687e-06, 'epoch': 1.83}) +[2026-06-17 17:51:58,451][starVLA.training.train_starvla][INFO] - Step 6212, Loss: {'train/loss': 0.7245516180992126, 'train/grad_norm_pre_clip': 14.145195007324219, 'timing/data': 0.005082758027128875, 'timing/model': 1.199898304999806, 'learning_rate/qwen_vl_interface': 8.024502778206411e-07, 'learning_rate/action_model': 8.02450277820641e-06, 'epoch': 1.83}) +[2026-06-17 17:52:03,174][starVLA.training.train_starvla][INFO] - Step 6213, Loss: {'train/loss': 0.7313188910484314, 'train/grad_norm_pre_clip': 13.52748966217041, 'timing/data': 0.005219976999796927, 'timing/model': 1.283963297959417, 'learning_rate/qwen_vl_interface': 8.016913506315566e-07, 'learning_rate/action_model': 8.016913506315565e-06, 'epoch': 1.83}) +[2026-06-17 17:52:07,822][starVLA.training.train_starvla][INFO] - Step 6214, Loss: {'train/loss': 0.7214113473892212, 'train/grad_norm_pre_clip': 14.362468719482422, 'timing/data': 0.004925508983433247, 'timing/model': 1.2049908980261534, 'learning_rate/qwen_vl_interface': 8.009333455816416e-07, 'learning_rate/action_model': 8.009333455816417e-06, 'epoch': 1.83}) +[2026-06-17 17:52:12,599][starVLA.training.train_starvla][INFO] - Step 6215, Loss: {'train/loss': 0.5711854696273804, 'train/grad_norm_pre_clip': 11.401865005493164, 'timing/data': 0.004791616927832365, 'timing/model': 1.2051799159962684, 'learning_rate/qwen_vl_interface': 8.001762628280318e-07, 'learning_rate/action_model': 8.001762628280319e-06, 'epoch': 1.83}) +[2026-06-17 17:52:17,388][starVLA.training.train_starvla][INFO] - Step 6216, Loss: {'train/loss': 0.6186372637748718, 'train/grad_norm_pre_clip': 15.59164047241211, 'timing/data': 0.005001801066100597, 'timing/model': 1.203172513982281, 'learning_rate/qwen_vl_interface': 7.994201025276708e-07, 'learning_rate/action_model': 7.994201025276708e-06, 'epoch': 1.83}) +[2026-06-17 17:52:22,170][starVLA.training.train_starvla][INFO] - Step 6217, Loss: {'train/loss': 0.8217548131942749, 'train/grad_norm_pre_clip': 18.253833770751953, 'timing/data': 0.005185391986742616, 'timing/model': 1.203404700034298, 'learning_rate/qwen_vl_interface': 7.986648648373115e-07, 'learning_rate/action_model': 7.986648648373115e-06, 'epoch': 1.83}) +[2026-06-17 17:52:26,956][starVLA.training.train_starvla][INFO] - Step 6218, Loss: {'train/loss': 0.4874250590801239, 'train/grad_norm_pre_clip': 14.114251136779785, 'timing/data': 0.004815081018023193, 'timing/model': 1.3358727840241045, 'learning_rate/qwen_vl_interface': 7.979105499135157e-07, 'learning_rate/action_model': 7.979105499135155e-06, 'epoch': 1.83}) +[2026-06-17 17:52:31,606][starVLA.training.train_starvla][INFO] - Step 6219, Loss: {'train/loss': 0.5583600997924805, 'train/grad_norm_pre_clip': 13.175572395324707, 'timing/data': 0.004914959077723324, 'timing/model': 1.2029225709848106, 'learning_rate/qwen_vl_interface': 7.971571579126533e-07, 'learning_rate/action_model': 7.971571579126534e-06, 'epoch': 1.83}) +[2026-06-17 17:52:36,400][starVLA.training.train_starvla][INFO] - Step 6220, Loss: {'train/loss': 0.5946046710014343, 'train/grad_norm_pre_clip': 15.190303802490234, 'timing/data': 0.005155315971933305, 'timing/model': 1.2047259060200304, 'learning_rate/qwen_vl_interface': 7.964046889909038e-07, 'learning_rate/action_model': 7.964046889909038e-06, 'epoch': 1.83}) +[2026-06-17 17:52:41,184][starVLA.training.train_starvla][INFO] - Step 6221, Loss: {'train/loss': 0.6984111666679382, 'train/grad_norm_pre_clip': 15.4793701171875, 'timing/data': 0.004688197979703546, 'timing/model': 1.2073884340934455, 'learning_rate/qwen_vl_interface': 7.956531433042543e-07, 'learning_rate/action_model': 7.956531433042542e-06, 'epoch': 1.83}) +[2026-06-17 17:52:45,936][starVLA.training.train_starvla][INFO] - Step 6222, Loss: {'train/loss': 0.46594032645225525, 'train/grad_norm_pre_clip': 11.78930950164795, 'timing/data': 0.004479825031012297, 'timing/model': 1.205590144963935, 'learning_rate/qwen_vl_interface': 7.94902521008501e-07, 'learning_rate/action_model': 7.94902521008501e-06, 'epoch': 1.83}) +[2026-06-17 17:52:50,682][starVLA.training.train_starvla][INFO] - Step 6223, Loss: {'train/loss': 0.6753221154212952, 'train/grad_norm_pre_clip': 13.377052307128906, 'timing/data': 0.004809169098734856, 'timing/model': 1.203744558035396, 'learning_rate/qwen_vl_interface': 7.941528222592501e-07, 'learning_rate/action_model': 7.941528222592501e-06, 'epoch': 1.83}) +[2026-06-17 17:52:55,481][starVLA.training.train_starvla][INFO] - Step 6224, Loss: {'train/loss': 0.5230492353439331, 'train/grad_norm_pre_clip': 12.220199584960938, 'timing/data': 0.005033178953453898, 'timing/model': 1.3408445559907705, 'learning_rate/qwen_vl_interface': 7.934040472119132e-07, 'learning_rate/action_model': 7.934040472119133e-06, 'epoch': 1.83}) +[2026-06-17 17:53:00,137][starVLA.training.train_starvla][INFO] - Step 6225, Loss: {'train/loss': 0.6990219354629517, 'train/grad_norm_pre_clip': 13.895769119262695, 'timing/data': 0.004947540001012385, 'timing/model': 1.2048062169924378, 'learning_rate/qwen_vl_interface': 7.926561960217127e-07, 'learning_rate/action_model': 7.926561960217126e-06, 'epoch': 1.83}) +[2026-06-17 17:53:04,925][starVLA.training.train_starvla][INFO] - Step 6226, Loss: {'train/loss': 0.7274582386016846, 'train/grad_norm_pre_clip': 14.143648147583008, 'timing/data': 0.005064164055511355, 'timing/model': 1.20581231394317, 'learning_rate/qwen_vl_interface': 7.919092688436783e-07, 'learning_rate/action_model': 7.919092688436783e-06, 'epoch': 1.83}) +[2026-06-17 17:53:09,705][starVLA.training.train_starvla][INFO] - Step 6227, Loss: {'train/loss': 0.44096291065216064, 'train/grad_norm_pre_clip': 11.782657623291016, 'timing/data': 0.005169200012460351, 'timing/model': 1.2043175500584766, 'learning_rate/qwen_vl_interface': 7.911632658326505e-07, 'learning_rate/action_model': 7.911632658326505e-06, 'epoch': 1.83}) +[2026-06-17 17:53:14,492][starVLA.training.train_starvla][INFO] - Step 6228, Loss: {'train/loss': 0.52857506275177, 'train/grad_norm_pre_clip': 15.781283378601074, 'timing/data': 0.0049879830330610275, 'timing/model': 1.204751466982998, 'learning_rate/qwen_vl_interface': 7.904181871432757e-07, 'learning_rate/action_model': 7.904181871432757e-06, 'epoch': 1.83}) +[2026-06-17 17:53:19,243][starVLA.training.train_starvla][INFO] - Step 6229, Loss: {'train/loss': 0.45265525579452515, 'train/grad_norm_pre_clip': 11.687792778015137, 'timing/data': 0.005125338095240295, 'timing/model': 1.2054170459741727, 'learning_rate/qwen_vl_interface': 7.896740329300091e-07, 'learning_rate/action_model': 7.89674032930009e-06, 'epoch': 1.83}) +[2026-06-17 17:53:24,019][starVLA.training.train_starvla][INFO] - Step 6230, Loss: {'train/loss': 0.35956189036369324, 'train/grad_norm_pre_clip': 9.821642875671387, 'timing/data': 0.0050141699612140656, 'timing/model': 1.32080682402011, 'learning_rate/qwen_vl_interface': 7.889308033471156e-07, 'learning_rate/action_model': 7.889308033471155e-06, 'epoch': 1.83}) +[2026-06-17 17:53:28,673][starVLA.training.train_starvla][INFO] - Step 6231, Loss: {'train/loss': 0.6538857817649841, 'train/grad_norm_pre_clip': 15.785243034362793, 'timing/data': 0.00483190396334976, 'timing/model': 1.2040811739861965, 'learning_rate/qwen_vl_interface': 7.881884985486662e-07, 'learning_rate/action_model': 7.881884985486662e-06, 'epoch': 1.83}) +[2026-06-17 17:53:33,467][starVLA.training.train_starvla][INFO] - Step 6232, Loss: {'train/loss': 0.6267089247703552, 'train/grad_norm_pre_clip': 14.11385726928711, 'timing/data': 0.004923618980683386, 'timing/model': 1.2059672300238162, 'learning_rate/qwen_vl_interface': 7.874471186885427e-07, 'learning_rate/action_model': 7.874471186885427e-06, 'epoch': 1.83}) +[2026-06-17 17:53:38,258][starVLA.training.train_starvla][INFO] - Step 6233, Loss: {'train/loss': 0.657494068145752, 'train/grad_norm_pre_clip': 13.64469051361084, 'timing/data': 0.005118825007230043, 'timing/model': 1.2064426470315084, 'learning_rate/qwen_vl_interface': 7.867066639204326e-07, 'learning_rate/action_model': 7.867066639204327e-06, 'epoch': 1.83}) +[2026-06-17 17:53:43,219][starVLA.training.train_starvla][INFO] - Step 6234, Loss: {'train/loss': 0.5707558393478394, 'train/grad_norm_pre_clip': 12.923178672790527, 'timing/data': 0.0049948099767789245, 'timing/model': 1.2058019930263981, 'learning_rate/qwen_vl_interface': 7.859671343978348e-07, 'learning_rate/action_model': 7.859671343978348e-06, 'epoch': 1.83}) +[2026-06-17 17:53:48,011][starVLA.training.train_starvla][INFO] - Step 6235, Loss: {'train/loss': 0.670249879360199, 'train/grad_norm_pre_clip': 13.253889083862305, 'timing/data': 0.0050052698934450746, 'timing/model': 1.3320494439685717, 'learning_rate/qwen_vl_interface': 7.852285302740528e-07, 'learning_rate/action_model': 7.852285302740527e-06, 'epoch': 1.83}) +[2026-06-17 17:53:52,678][starVLA.training.train_starvla][INFO] - Step 6236, Loss: {'train/loss': 0.9260589480400085, 'train/grad_norm_pre_clip': 17.085769653320312, 'timing/data': 0.0049147349782288074, 'timing/model': 1.2115505959372967, 'learning_rate/qwen_vl_interface': 7.844908517022005e-07, 'learning_rate/action_model': 7.844908517022005e-06, 'epoch': 1.83}) +[2026-06-17 17:53:57,467][starVLA.training.train_starvla][INFO] - Step 6237, Loss: {'train/loss': 0.8229966759681702, 'train/grad_norm_pre_clip': 17.54792594909668, 'timing/data': 0.004771906998939812, 'timing/model': 1.2046480309218168, 'learning_rate/qwen_vl_interface': 7.837540988352001e-07, 'learning_rate/action_model': 7.837540988352e-06, 'epoch': 1.84}) +[2026-06-17 17:54:02,175][starVLA.training.train_starvla][INFO] - Step 6238, Loss: {'train/loss': 0.5996047854423523, 'train/grad_norm_pre_clip': 16.899381637573242, 'timing/data': 0.004893120029009879, 'timing/model': 1.2058163529727608, 'learning_rate/qwen_vl_interface': 7.83018271825782e-07, 'learning_rate/action_model': 7.83018271825782e-06, 'epoch': 1.84}) +[2026-06-17 17:54:06,914][starVLA.training.train_starvla][INFO] - Step 6239, Loss: {'train/loss': 0.7375367879867554, 'train/grad_norm_pre_clip': 13.088963508605957, 'timing/data': 0.004987898049876094, 'timing/model': 1.2060766129288822, 'learning_rate/qwen_vl_interface': 7.822833708264821e-07, 'learning_rate/action_model': 7.822833708264821e-06, 'epoch': 1.84}) +[2026-06-17 17:54:11,792][starVLA.training.train_starvla][INFO] - Step 6240, Loss: {'train/loss': 0.5996708869934082, 'train/grad_norm_pre_clip': 15.735047340393066, 'timing/data': 0.004993503913283348, 'timing/model': 1.208789503085427, 'learning_rate/qwen_vl_interface': 7.81549395989646e-07, 'learning_rate/action_model': 7.81549395989646e-06, 'epoch': 1.84}) +[2026-06-17 17:54:16,574][starVLA.training.train_starvla][INFO] - Step 6241, Loss: {'train/loss': 0.5778475999832153, 'train/grad_norm_pre_clip': 12.406167984008789, 'timing/data': 0.0049511760007590055, 'timing/model': 1.3319382659392431, 'learning_rate/qwen_vl_interface': 7.808163474674291e-07, 'learning_rate/action_model': 7.80816347467429e-06, 'epoch': 1.84}) +[2026-06-17 17:54:21,251][starVLA.training.train_starvla][INFO] - Step 6242, Loss: {'train/loss': 0.8585677742958069, 'train/grad_norm_pre_clip': 17.853389739990234, 'timing/data': 0.005101346061564982, 'timing/model': 1.2066116100177169, 'learning_rate/qwen_vl_interface': 7.800842254117922e-07, 'learning_rate/action_model': 7.80084225411792e-06, 'epoch': 1.84}) +[2026-06-17 17:54:26,048][starVLA.training.train_starvla][INFO] - Step 6243, Loss: {'train/loss': 0.5155295133590698, 'train/grad_norm_pre_clip': 12.23674488067627, 'timing/data': 0.005740954075008631, 'timing/model': 1.2055130070075393, 'learning_rate/qwen_vl_interface': 7.793530299745061e-07, 'learning_rate/action_model': 7.793530299745061e-06, 'epoch': 1.84}) +[2026-06-17 17:54:30,836][starVLA.training.train_starvla][INFO] - Step 6244, Loss: {'train/loss': 1.017928123474121, 'train/grad_norm_pre_clip': 17.352293014526367, 'timing/data': 0.0047785540809854865, 'timing/model': 1.2120454229880124, 'learning_rate/qwen_vl_interface': 7.786227613071464e-07, 'learning_rate/action_model': 7.786227613071464e-06, 'epoch': 1.84}) +[2026-06-17 17:54:35,596][starVLA.training.train_starvla][INFO] - Step 6245, Loss: {'train/loss': 0.5828893184661865, 'train/grad_norm_pre_clip': 14.701347351074219, 'timing/data': 0.004624711931683123, 'timing/model': 1.2057559170061722, 'learning_rate/qwen_vl_interface': 7.778934195610998e-07, 'learning_rate/action_model': 7.778934195610998e-06, 'epoch': 1.84}) +[2026-06-17 17:54:40,399][starVLA.training.train_starvla][INFO] - Step 6246, Loss: {'train/loss': 0.9477201104164124, 'train/grad_norm_pre_clip': 15.798047065734863, 'timing/data': 0.004848952987231314, 'timing/model': 1.2238831670256332, 'learning_rate/qwen_vl_interface': 7.771650048875598e-07, 'learning_rate/action_model': 7.771650048875597e-06, 'epoch': 1.84}) +[2026-06-17 17:54:45,193][starVLA.training.train_starvla][INFO] - Step 6247, Loss: {'train/loss': 0.5153430104255676, 'train/grad_norm_pre_clip': 11.10560131072998, 'timing/data': 0.004920010105706751, 'timing/model': 1.339083241065964, 'learning_rate/qwen_vl_interface': 7.764375174375276e-07, 'learning_rate/action_model': 7.764375174375276e-06, 'epoch': 1.84}) +[2026-06-17 17:54:49,847][starVLA.training.train_starvla][INFO] - Step 6248, Loss: {'train/loss': 0.5744618773460388, 'train/grad_norm_pre_clip': 13.229687690734863, 'timing/data': 0.005183049943298101, 'timing/model': 1.205048140953295, 'learning_rate/qwen_vl_interface': 7.757109573618116e-07, 'learning_rate/action_model': 7.757109573618115e-06, 'epoch': 1.84}) +[2026-06-17 17:54:54,637][starVLA.training.train_starvla][INFO] - Step 6249, Loss: {'train/loss': 0.8420979380607605, 'train/grad_norm_pre_clip': 16.383758544921875, 'timing/data': 0.004979021963663399, 'timing/model': 1.203800881979987, 'learning_rate/qwen_vl_interface': 7.749853248110287e-07, 'learning_rate/action_model': 7.749853248110287e-06, 'epoch': 1.84}) +[2026-06-17 17:54:59,422][starVLA.training.train_starvla][INFO] - Step 6250, Loss: {'train/loss': 0.6044120192527771, 'train/grad_norm_pre_clip': 12.30605697631836, 'timing/data': 0.004571043071337044, 'timing/model': 1.2044070530682802, 'learning_rate/qwen_vl_interface': 7.742606199356032e-07, 'learning_rate/action_model': 7.74260619935603e-06, 'epoch': 1.84}) +[2026-06-17 17:55:04,209][starVLA.training.train_starvla][INFO] - Step 6251, Loss: {'train/loss': 0.6919273734092712, 'train/grad_norm_pre_clip': 15.150035858154297, 'timing/data': 0.005524383042939007, 'timing/model': 1.2045069399755448, 'learning_rate/qwen_vl_interface': 7.735368428857665e-07, 'learning_rate/action_model': 7.735368428857665e-06, 'epoch': 1.84}) +[2026-06-17 17:55:08,878][starVLA.training.train_starvla][INFO] - Step 6252, Loss: {'train/loss': 0.8741201758384705, 'train/grad_norm_pre_clip': 15.807040214538574, 'timing/data': 0.0052064110059291124, 'timing/model': 1.2061764160171151, 'learning_rate/qwen_vl_interface': 7.728139938115605e-07, 'learning_rate/action_model': 7.728139938115604e-06, 'epoch': 1.84}) +[2026-06-17 17:55:13,585][starVLA.training.train_starvla][INFO] - Step 6253, Loss: {'train/loss': 0.7023645639419556, 'train/grad_norm_pre_clip': 14.32665729522705, 'timing/data': 0.005039724987000227, 'timing/model': 1.2480586849851534, 'learning_rate/qwen_vl_interface': 7.720920728628306e-07, 'learning_rate/action_model': 7.720920728628306e-06, 'epoch': 1.84}) +[2026-06-17 17:55:18,247][starVLA.training.train_starvla][INFO] - Step 6254, Loss: {'train/loss': 0.6336466670036316, 'train/grad_norm_pre_clip': 14.433076858520508, 'timing/data': 0.005023461999371648, 'timing/model': 1.205674865981564, 'learning_rate/qwen_vl_interface': 7.713710801892327e-07, 'learning_rate/action_model': 7.713710801892326e-06, 'epoch': 1.84}) +[2026-06-17 17:55:22,987][starVLA.training.train_starvla][INFO] - Step 6255, Loss: {'train/loss': 0.5702432990074158, 'train/grad_norm_pre_clip': 15.855452537536621, 'timing/data': 0.00532849901355803, 'timing/model': 1.2063353140838444, 'learning_rate/qwen_vl_interface': 7.70651015940228e-07, 'learning_rate/action_model': 7.70651015940228e-06, 'epoch': 1.84}) +[2026-06-17 17:55:27,786][starVLA.training.train_starvla][INFO] - Step 6256, Loss: {'train/loss': 0.49449414014816284, 'train/grad_norm_pre_clip': 16.630952835083008, 'timing/data': 0.004812204046174884, 'timing/model': 1.2106294539989904, 'learning_rate/qwen_vl_interface': 7.699318802650891e-07, 'learning_rate/action_model': 7.69931880265089e-06, 'epoch': 1.84}) +[2026-06-17 17:55:32,592][starVLA.training.train_starvla][INFO] - Step 6257, Loss: {'train/loss': 0.8367493748664856, 'train/grad_norm_pre_clip': 16.899677276611328, 'timing/data': 0.004855131963267922, 'timing/model': 1.2073342551011592, 'learning_rate/qwen_vl_interface': 7.692136733128918e-07, 'learning_rate/action_model': 7.692136733128917e-06, 'epoch': 1.84}) +[2026-06-17 17:55:37,377][starVLA.training.train_starvla][INFO] - Step 6258, Loss: {'train/loss': 0.7618212699890137, 'train/grad_norm_pre_clip': 15.900460243225098, 'timing/data': 0.0044011210557073355, 'timing/model': 1.3317249770043418, 'learning_rate/qwen_vl_interface': 7.684963952325207e-07, 'learning_rate/action_model': 7.684963952325206e-06, 'epoch': 1.84}) +[2026-06-17 17:55:42,084][starVLA.training.train_starvla][INFO] - Step 6259, Loss: {'train/loss': 0.77508544921875, 'train/grad_norm_pre_clip': 14.382009506225586, 'timing/data': 0.00475760898552835, 'timing/model': 1.2555523570626974, 'learning_rate/qwen_vl_interface': 7.6778004617267e-07, 'learning_rate/action_model': 7.6778004617267e-06, 'epoch': 1.84}) +[2026-06-17 17:55:46,744][starVLA.training.train_starvla][INFO] - Step 6260, Loss: {'train/loss': 0.5540972948074341, 'train/grad_norm_pre_clip': 11.709799766540527, 'timing/data': 0.004674060037359595, 'timing/model': 1.2069410390686244, 'learning_rate/qwen_vl_interface': 7.670646262818387e-07, 'learning_rate/action_model': 7.670646262818387e-06, 'epoch': 1.84}) +[2026-06-17 17:55:51,496][starVLA.training.train_starvla][INFO] - Step 6261, Loss: {'train/loss': 0.7973595261573792, 'train/grad_norm_pre_clip': 19.687379837036133, 'timing/data': 0.004628365044482052, 'timing/model': 1.2066309369402006, 'learning_rate/qwen_vl_interface': 7.663501357083342e-07, 'learning_rate/action_model': 7.66350135708334e-06, 'epoch': 1.84}) +[2026-06-17 17:55:56,288][starVLA.training.train_starvla][INFO] - Step 6262, Loss: {'train/loss': 0.6197478771209717, 'train/grad_norm_pre_clip': 13.364121437072754, 'timing/data': 0.004651644034311175, 'timing/model': 1.2062524119392037, 'learning_rate/qwen_vl_interface': 7.656365746002715e-07, 'learning_rate/action_model': 7.656365746002716e-06, 'epoch': 1.84}) +[2026-06-17 17:56:01,086][starVLA.training.train_starvla][INFO] - Step 6263, Loss: {'train/loss': 0.6324899792671204, 'train/grad_norm_pre_clip': 18.235553741455078, 'timing/data': 0.004967299988493323, 'timing/model': 1.2106668399646878, 'learning_rate/qwen_vl_interface': 7.649239431055718e-07, 'learning_rate/action_model': 7.649239431055717e-06, 'epoch': 1.84}) +[2026-06-17 17:56:05,871][starVLA.training.train_starvla][INFO] - Step 6264, Loss: {'train/loss': 0.7043635249137878, 'train/grad_norm_pre_clip': 15.059165000915527, 'timing/data': 0.004770531901158392, 'timing/model': 1.3318516130093485, 'learning_rate/qwen_vl_interface': 7.642122413719652e-07, 'learning_rate/action_model': 7.642122413719651e-06, 'epoch': 1.84}) +[2026-06-17 17:56:10,647][starVLA.training.train_starvla][INFO] - Step 6265, Loss: {'train/loss': 0.7950870990753174, 'train/grad_norm_pre_clip': 18.440418243408203, 'timing/data': 0.004764872952364385, 'timing/model': 1.3240172440418974, 'learning_rate/qwen_vl_interface': 7.635014695469877e-07, 'learning_rate/action_model': 7.635014695469878e-06, 'epoch': 1.84}) +[2026-06-17 17:56:15,302][starVLA.training.train_starvla][INFO] - Step 6266, Loss: {'train/loss': 0.6322066783905029, 'train/grad_norm_pre_clip': 16.187711715698242, 'timing/data': 0.004944786080159247, 'timing/model': 1.2051680570002645, 'learning_rate/qwen_vl_interface': 7.627916277779832e-07, 'learning_rate/action_model': 7.627916277779831e-06, 'epoch': 1.84}) +[2026-06-17 17:56:20,071][starVLA.training.train_starvla][INFO] - Step 6267, Loss: {'train/loss': 0.7357699275016785, 'train/grad_norm_pre_clip': 12.522801399230957, 'timing/data': 0.0052038030698895454, 'timing/model': 1.2062992099672556, 'learning_rate/qwen_vl_interface': 7.62082716212103e-07, 'learning_rate/action_model': 7.62082716212103e-06, 'epoch': 1.84}) +[2026-06-17 17:56:24,867][starVLA.training.train_starvla][INFO] - Step 6268, Loss: {'train/loss': 0.4020189344882965, 'train/grad_norm_pre_clip': 10.21910572052002, 'timing/data': 0.004951086943037808, 'timing/model': 1.2096854160772637, 'learning_rate/qwen_vl_interface': 7.613747349963048e-07, 'learning_rate/action_model': 7.613747349963048e-06, 'epoch': 1.84}) +[2026-06-17 17:56:29,654][starVLA.training.train_starvla][INFO] - Step 6269, Loss: {'train/loss': 0.5849224328994751, 'train/grad_norm_pre_clip': 16.71537971496582, 'timing/data': 0.005111093050800264, 'timing/model': 1.2053027199581265, 'learning_rate/qwen_vl_interface': 7.606676842773531e-07, 'learning_rate/action_model': 7.60667684277353e-06, 'epoch': 1.84}) +[2026-06-17 17:56:34,432][starVLA.training.train_starvla][INFO] - Step 6270, Loss: {'train/loss': 0.7158252000808716, 'train/grad_norm_pre_clip': 15.882224082946777, 'timing/data': 0.005093975109048188, 'timing/model': 1.2815324689727277, 'learning_rate/qwen_vl_interface': 7.599615642018218e-07, 'learning_rate/action_model': 7.599615642018218e-06, 'epoch': 1.84}) +[2026-06-17 17:56:39,098][starVLA.training.train_starvla][INFO] - Step 6271, Loss: {'train/loss': 0.7314303517341614, 'train/grad_norm_pre_clip': 15.256125450134277, 'timing/data': 0.005367131088860333, 'timing/model': 1.2056727510644123, 'learning_rate/qwen_vl_interface': 7.592563749160908e-07, 'learning_rate/action_model': 7.592563749160908e-06, 'epoch': 1.85}) +[2026-06-17 17:56:43,863][starVLA.training.train_starvla][INFO] - Step 6272, Loss: {'train/loss': 0.645723819732666, 'train/grad_norm_pre_clip': 19.222793579101562, 'timing/data': 0.004898732993751764, 'timing/model': 1.2065696410136297, 'learning_rate/qwen_vl_interface': 7.585521165663445e-07, 'learning_rate/action_model': 7.585521165663445e-06, 'epoch': 1.85}) +[2026-06-17 17:56:48,748][starVLA.training.train_starvla][INFO] - Step 6273, Loss: {'train/loss': 0.5759715437889099, 'train/grad_norm_pre_clip': 15.313072204589844, 'timing/data': 0.005448601907119155, 'timing/model': 1.2111120709450915, 'learning_rate/qwen_vl_interface': 7.578487892985769e-07, 'learning_rate/action_model': 7.578487892985769e-06, 'epoch': 1.85}) +[2026-06-17 17:56:53,540][starVLA.training.train_starvla][INFO] - Step 6274, Loss: {'train/loss': 0.6218594908714294, 'train/grad_norm_pre_clip': 13.197690963745117, 'timing/data': 0.0050780740566551685, 'timing/model': 1.2074510030215606, 'learning_rate/qwen_vl_interface': 7.571463932585901e-07, 'learning_rate/action_model': 7.571463932585901e-06, 'epoch': 1.85}) +[2026-06-17 17:56:58,325][starVLA.training.train_starvla][INFO] - Step 6275, Loss: {'train/loss': 0.5705899000167847, 'train/grad_norm_pre_clip': 13.087845802307129, 'timing/data': 0.004880597000010312, 'timing/model': 1.331423411029391, 'learning_rate/qwen_vl_interface': 7.564449285919897e-07, 'learning_rate/action_model': 7.564449285919897e-06, 'epoch': 1.85}) +[2026-06-17 17:57:03,079][starVLA.training.train_starvla][INFO] - Step 6276, Loss: {'train/loss': 0.7124494314193726, 'train/grad_norm_pre_clip': 15.722539901733398, 'timing/data': 0.004929731017909944, 'timing/model': 1.2950178700266406, 'learning_rate/qwen_vl_interface': 7.557443954441924e-07, 'learning_rate/action_model': 7.557443954441923e-06, 'epoch': 1.85}) +[2026-06-17 17:57:07,777][starVLA.training.train_starvla][INFO] - Step 6277, Loss: {'train/loss': 0.4426051378250122, 'train/grad_norm_pre_clip': 11.649483680725098, 'timing/data': 0.004722513956949115, 'timing/model': 1.209199454053305, 'learning_rate/qwen_vl_interface': 7.550447939604161e-07, 'learning_rate/action_model': 7.550447939604161e-06, 'epoch': 1.85}) +[2026-06-17 17:57:12,572][starVLA.training.train_starvla][INFO] - Step 6278, Loss: {'train/loss': 0.498992383480072, 'train/grad_norm_pre_clip': 13.740087509155273, 'timing/data': 0.0048585759941488504, 'timing/model': 1.2062721980037168, 'learning_rate/qwen_vl_interface': 7.543461242856916e-07, 'learning_rate/action_model': 7.543461242856916e-06, 'epoch': 1.85}) +[2026-06-17 17:57:17,380][starVLA.training.train_starvla][INFO] - Step 6279, Loss: {'train/loss': 0.5390384197235107, 'train/grad_norm_pre_clip': 13.745356559753418, 'timing/data': 0.004970514914020896, 'timing/model': 1.2173450849950314, 'learning_rate/qwen_vl_interface': 7.536483865648531e-07, 'learning_rate/action_model': 7.536483865648531e-06, 'epoch': 1.85}) +[2026-06-17 17:57:22,188][starVLA.training.train_starvla][INFO] - Step 6280, Loss: {'train/loss': 0.575370192527771, 'train/grad_norm_pre_clip': 15.14169979095459, 'timing/data': 0.004933730000630021, 'timing/model': 1.2141678460175171, 'learning_rate/qwen_vl_interface': 7.52951580942542e-07, 'learning_rate/action_model': 7.5295158094254195e-06, 'epoch': 1.85}) +[2026-06-17 17:57:26,967][starVLA.training.train_starvla][INFO] - Step 6281, Loss: {'train/loss': 0.5599396824836731, 'train/grad_norm_pre_clip': 13.115592956542969, 'timing/data': 0.0049815920647233725, 'timing/model': 1.2093175370246172, 'learning_rate/qwen_vl_interface': 7.522557075632078e-07, 'learning_rate/action_model': 7.522557075632078e-06, 'epoch': 1.85}) +[2026-06-17 17:57:31,766][starVLA.training.train_starvla][INFO] - Step 6282, Loss: {'train/loss': 0.5916457772254944, 'train/grad_norm_pre_clip': 17.73073387145996, 'timing/data': 0.004987890948541462, 'timing/model': 1.2919098809361458, 'learning_rate/qwen_vl_interface': 7.515607665711051e-07, 'learning_rate/action_model': 7.515607665711051e-06, 'epoch': 1.85}) +[2026-06-17 17:57:36,445][starVLA.training.train_starvla][INFO] - Step 6283, Loss: {'train/loss': 0.6186525225639343, 'train/grad_norm_pre_clip': 13.440834999084473, 'timing/data': 0.0047381429467350245, 'timing/model': 1.2061896619852632, 'learning_rate/qwen_vl_interface': 7.508667581102959e-07, 'learning_rate/action_model': 7.508667581102958e-06, 'epoch': 1.85}) +[2026-06-17 17:57:41,239][starVLA.training.train_starvla][INFO] - Step 6284, Loss: {'train/loss': 0.6650454998016357, 'train/grad_norm_pre_clip': 13.814040184020996, 'timing/data': 0.004853103077039123, 'timing/model': 1.2078137340722606, 'learning_rate/qwen_vl_interface': 7.50173682324648e-07, 'learning_rate/action_model': 7.501736823246481e-06, 'epoch': 1.85}) +[2026-06-17 17:57:46,061][starVLA.training.train_starvla][INFO] - Step 6285, Loss: {'train/loss': 0.3592047095298767, 'train/grad_norm_pre_clip': 9.79703426361084, 'timing/data': 0.005042850971221924, 'timing/model': 1.2034231289289892, 'learning_rate/qwen_vl_interface': 7.494815393578394e-07, 'learning_rate/action_model': 7.494815393578393e-06, 'epoch': 1.85}) +[2026-06-17 17:57:50,873][starVLA.training.train_starvla][INFO] - Step 6286, Loss: {'train/loss': 0.6000572443008423, 'train/grad_norm_pre_clip': 12.477304458618164, 'timing/data': 0.004726296989247203, 'timing/model': 1.206088092061691, 'learning_rate/qwen_vl_interface': 7.487903293533501e-07, 'learning_rate/action_model': 7.4879032935335e-06, 'epoch': 1.85}) +[2026-06-17 17:57:55,698][starVLA.training.train_starvla][INFO] - Step 6287, Loss: {'train/loss': 0.5344699025154114, 'train/grad_norm_pre_clip': 11.69893741607666, 'timing/data': 0.00489232805557549, 'timing/model': 1.3543897300260141, 'learning_rate/qwen_vl_interface': 7.481000524544686e-07, 'learning_rate/action_model': 7.481000524544686e-06, 'epoch': 1.85}) +[2026-06-17 17:58:00,376][starVLA.training.train_starvla][INFO] - Step 6288, Loss: {'train/loss': 0.5381568670272827, 'train/grad_norm_pre_clip': 16.560243606567383, 'timing/data': 0.004676965065300465, 'timing/model': 1.21388441498857, 'learning_rate/qwen_vl_interface': 7.474107088042898e-07, 'learning_rate/action_model': 7.474107088042898e-06, 'epoch': 1.85}) +[2026-06-17 17:58:05,190][starVLA.training.train_starvla][INFO] - Step 6289, Loss: {'train/loss': 0.49910813570022583, 'train/grad_norm_pre_clip': 11.36944580078125, 'timing/data': 0.004841997055336833, 'timing/model': 1.207941055065021, 'learning_rate/qwen_vl_interface': 7.467222985457167e-07, 'learning_rate/action_model': 7.4672229854571675e-06, 'epoch': 1.85}) +[2026-06-17 17:58:10,002][starVLA.training.train_starvla][INFO] - Step 6290, Loss: {'train/loss': 0.3920004367828369, 'train/grad_norm_pre_clip': 9.74795150756836, 'timing/data': 0.0047732830280438066, 'timing/model': 1.206595716997981, 'learning_rate/qwen_vl_interface': 7.46034821821458e-07, 'learning_rate/action_model': 7.46034821821458e-06, 'epoch': 1.85}) +[2026-06-17 17:58:14,733][starVLA.training.train_starvla][INFO] - Step 6291, Loss: {'train/loss': 0.8712764978408813, 'train/grad_norm_pre_clip': 16.824586868286133, 'timing/data': 0.004923692089505494, 'timing/model': 1.2088281130418181, 'learning_rate/qwen_vl_interface': 7.453482787740255e-07, 'learning_rate/action_model': 7.4534827877402545e-06, 'epoch': 1.85}) +[2026-06-17 17:58:19,501][starVLA.training.train_starvla][INFO] - Step 6292, Loss: {'train/loss': 0.683995246887207, 'train/grad_norm_pre_clip': 13.33749771118164, 'timing/data': 0.004962179926224053, 'timing/model': 1.2123380489647388, 'learning_rate/qwen_vl_interface': 7.446626695457427e-07, 'learning_rate/action_model': 7.446626695457427e-06, 'epoch': 1.85}) +[2026-06-17 17:58:24,304][starVLA.training.train_starvla][INFO] - Step 6293, Loss: {'train/loss': 0.6547139883041382, 'train/grad_norm_pre_clip': 15.908615112304688, 'timing/data': 0.004975765012204647, 'timing/model': 1.3348539259750396, 'learning_rate/qwen_vl_interface': 7.439779942787367e-07, 'learning_rate/action_model': 7.439779942787367e-06, 'epoch': 1.85}) +[2026-06-17 17:58:28,977][starVLA.training.train_starvla][INFO] - Step 6294, Loss: {'train/loss': 0.7290023565292358, 'train/grad_norm_pre_clip': 16.27528953552246, 'timing/data': 0.004949575988575816, 'timing/model': 1.206054528011009, 'learning_rate/qwen_vl_interface': 7.432942531149409e-07, 'learning_rate/action_model': 7.432942531149409e-06, 'epoch': 1.85}) +[2026-06-17 17:58:33,771][starVLA.training.train_starvla][INFO] - Step 6295, Loss: {'train/loss': 0.5404446125030518, 'train/grad_norm_pre_clip': 13.402095794677734, 'timing/data': 0.005161028006114066, 'timing/model': 1.2098527619382367, 'learning_rate/qwen_vl_interface': 7.42611446196096e-07, 'learning_rate/action_model': 7.426114461960959e-06, 'epoch': 1.85}) +[2026-06-17 17:58:38,553][starVLA.training.train_starvla][INFO] - Step 6296, Loss: {'train/loss': 0.6560163497924805, 'train/grad_norm_pre_clip': 13.148735046386719, 'timing/data': 0.004858109052293003, 'timing/model': 1.2031540439929813, 'learning_rate/qwen_vl_interface': 7.419295736637483e-07, 'learning_rate/action_model': 7.419295736637483e-06, 'epoch': 1.85}) +[2026-06-17 17:58:43,337][starVLA.training.train_starvla][INFO] - Step 6297, Loss: {'train/loss': 0.7714589238166809, 'train/grad_norm_pre_clip': 14.30207633972168, 'timing/data': 0.005051033920608461, 'timing/model': 1.2063090429874137, 'learning_rate/qwen_vl_interface': 7.412486356592512e-07, 'learning_rate/action_model': 7.412486356592511e-06, 'epoch': 1.85}) +[2026-06-17 17:58:48,103][starVLA.training.train_starvla][INFO] - Step 6298, Loss: {'train/loss': 0.4976687431335449, 'train/grad_norm_pre_clip': 11.536444664001465, 'timing/data': 0.004895268008112907, 'timing/model': 1.2047585760010406, 'learning_rate/qwen_vl_interface': 7.405686323237623e-07, 'learning_rate/action_model': 7.405686323237623e-06, 'epoch': 1.85}) +[2026-06-17 17:58:52,886][starVLA.training.train_starvla][INFO] - Step 6299, Loss: {'train/loss': 0.9387819766998291, 'train/grad_norm_pre_clip': 15.175439834594727, 'timing/data': 0.005066776066087186, 'timing/model': 1.3227904579835013, 'learning_rate/qwen_vl_interface': 7.398895637982504e-07, 'learning_rate/action_model': 7.398895637982503e-06, 'epoch': 1.85}) +[2026-06-17 18:00:34,688][starVLA.training.train_starvla][INFO] - Step 6300, Loss: {'train/loss': 1.022308349609375, 'train/grad_norm_pre_clip': 16.091753005981445, 'eval/loss': 0.7012152671813965, 'eval/latency_0/loss': 0.7012152671813965, 'eval/demon_fire_precision': 0.9098694942903752, 'eval/demon_fire_recall': 0.9264950166112956, 'eval/demon_fire_f1': 0.9181069958847737, 'eval/demon_fire_rate_model': 0.681868743047831, 'eval/demon_fire_rate_teacher': 0.6696329254727474, 'eval/demon_left_precision': 0.8239277652370203, 'eval/demon_left_recall': 0.8271954674220963, 'eval/demon_left_f1': 0.8255583828102911, 'eval/demon_left_rate_model': 0.492769744160178, 'eval/demon_left_rate_teacher': 0.49082313681868744, 'eval/demon_right_precision': 0.8289693593314763, 'eval/demon_right_recall': 0.8406779661016949, 'eval/demon_right_f1': 0.8347826086956521, 'eval/demon_right_rate_model': 0.49916573971078976, 'eval/demon_right_rate_teacher': 0.4922135706340378, 'eval/demon_move_f1': 0.8301704957529716, 'eval/demon_cc_f1': 0.8898301948725715, 'eval/latency_0/demon_fire_precision': 0.9098694942903752, 'eval/latency_0/demon_fire_recall': 0.9264950166112956, 'eval/latency_0/demon_fire_f1': 0.9181069958847737, 'eval/latency_0/demon_fire_rate_model': 0.681868743047831, 'eval/latency_0/demon_fire_rate_teacher': 0.6696329254727474, 'eval/latency_0/demon_left_precision': 0.8239277652370203, 'eval/latency_0/demon_left_recall': 0.8271954674220963, 'eval/latency_0/demon_left_f1': 0.8255583828102911, 'eval/latency_0/demon_left_rate_model': 0.492769744160178, 'eval/latency_0/demon_left_rate_teacher': 0.49082313681868744, 'eval/latency_0/demon_right_precision': 0.8289693593314763, 'eval/latency_0/demon_right_recall': 0.8406779661016949, 'eval/latency_0/demon_right_f1': 0.8347826086956521, 'eval/latency_0/demon_right_rate_model': 0.49916573971078976, 'eval/latency_0/demon_right_rate_teacher': 0.4922135706340378, 'eval/latency_0/demon_move_f1': 0.8301704957529716, 'eval/latency_0/demon_cc_f1': 0.8898301948725715, 'timing/data': 0.004894722951576114, 'timing/model': 1.205360152060166, 'learning_rate/qwen_vl_interface': 7.392114302234838e-07, 'learning_rate/action_model': 7.3921143022348385e-06, 'epoch': 1.85}) +[2026-06-17 18:00:34,689][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_6300_state +[2026-06-17 18:00:34,692][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 18:00:38,240][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_6300_state/model.safetensors +[2026-06-17 18:00:45,611][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_6300_state/optimizer.bin +[2026-06-17 18:00:45,611][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_6300_state/sampler.bin +[2026-06-17 18:00:45,611][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_6300_state/sampler_1.bin +[2026-06-17 18:00:45,612][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_6300_state/random_states_0.pkl +[2026-06-17 18:00:45,665][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 6300 +[2026-06-17 18:00:45,665][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 18:00:45,716][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 18:00:50,446][starVLA.training.train_starvla][INFO] - Step 6301, Loss: {'train/loss': 0.39427751302719116, 'train/grad_norm_pre_clip': 11.771081924438477, 'timing/data': 0.005646339035592973, 'timing/model': 1.3147644229466096, 'learning_rate/qwen_vl_interface': 7.385342317400418e-07, 'learning_rate/action_model': 7.385342317400417e-06, 'epoch': 1.85}) +[2026-06-17 18:00:55,052][starVLA.training.train_starvla][INFO] - Step 6302, Loss: {'train/loss': 0.9889469146728516, 'train/grad_norm_pre_clip': 20.522994995117188, 'timing/data': 0.0056144180707633495, 'timing/model': 1.1917869949247688, 'learning_rate/qwen_vl_interface': 7.378579684883069e-07, 'learning_rate/action_model': 7.378579684883069e-06, 'epoch': 1.85}) +[2026-06-17 18:00:59,704][starVLA.training.train_starvla][INFO] - Step 6303, Loss: {'train/loss': 0.9420199990272522, 'train/grad_norm_pre_clip': 18.30923843383789, 'timing/data': 0.005042891018092632, 'timing/model': 1.1919381669722497, 'learning_rate/qwen_vl_interface': 7.371826406084714e-07, 'learning_rate/action_model': 7.371826406084714e-06, 'epoch': 1.85}) +[2026-06-17 18:01:04,395][starVLA.training.train_starvla][INFO] - Step 6304, Loss: {'train/loss': 0.8542975783348083, 'train/grad_norm_pre_clip': 17.626928329467773, 'timing/data': 0.0052727420115843415, 'timing/model': 1.1952362200245261, 'learning_rate/qwen_vl_interface': 7.365082482405315e-07, 'learning_rate/action_model': 7.365082482405315e-06, 'epoch': 1.85}) +[2026-06-17 18:01:09,214][starVLA.training.train_starvla][INFO] - Step 6305, Loss: {'train/loss': 0.7383795976638794, 'train/grad_norm_pre_clip': 12.07909870147705, 'timing/data': 0.005523476982489228, 'timing/model': 1.196810289984569, 'learning_rate/qwen_vl_interface': 7.358347915242877e-07, 'learning_rate/action_model': 7.358347915242876e-06, 'epoch': 1.86}) +[2026-06-17 18:01:13,970][starVLA.training.train_starvla][INFO] - Step 6306, Loss: {'train/loss': 0.9007195234298706, 'train/grad_norm_pre_clip': 16.950918197631836, 'timing/data': 0.00570500292815268, 'timing/model': 1.3263943390920758, 'learning_rate/qwen_vl_interface': 7.351622705993481e-07, 'learning_rate/action_model': 7.351622705993481e-06, 'epoch': 1.86}) +[2026-06-17 18:01:18,632][starVLA.training.train_starvla][INFO] - Step 6307, Loss: {'train/loss': 0.604375958442688, 'train/grad_norm_pre_clip': 14.827230453491211, 'timing/data': 0.0057652109535411, 'timing/model': 1.1953080120729282, 'learning_rate/qwen_vl_interface': 7.344906856051288e-07, 'learning_rate/action_model': 7.344906856051288e-06, 'epoch': 1.86}) +[2026-06-17 18:01:23,393][starVLA.training.train_starvla][INFO] - Step 6308, Loss: {'train/loss': 0.6379448175430298, 'train/grad_norm_pre_clip': 14.715279579162598, 'timing/data': 0.005583159974776208, 'timing/model': 1.1964230010053143, 'learning_rate/qwen_vl_interface': 7.338200366808493e-07, 'learning_rate/action_model': 7.338200366808492e-06, 'epoch': 1.86}) +[2026-06-17 18:01:28,167][starVLA.training.train_starvla][INFO] - Step 6309, Loss: {'train/loss': 0.5052198171615601, 'train/grad_norm_pre_clip': 11.762716293334961, 'timing/data': 0.005512727075256407, 'timing/model': 1.203660016064532, 'learning_rate/qwen_vl_interface': 7.331503239655361e-07, 'learning_rate/action_model': 7.33150323965536e-06, 'epoch': 1.86}) +[2026-06-17 18:01:32,887][starVLA.training.train_starvla][INFO] - Step 6310, Loss: {'train/loss': 0.48677584528923035, 'train/grad_norm_pre_clip': 13.863618850708008, 'timing/data': 0.005104568088427186, 'timing/model': 1.1994945710757747, 'learning_rate/qwen_vl_interface': 7.324815475980207e-07, 'learning_rate/action_model': 7.324815475980207e-06, 'epoch': 1.86}) +[2026-06-17 18:01:37,661][starVLA.training.train_starvla][INFO] - Step 6311, Loss: {'train/loss': 0.5935207009315491, 'train/grad_norm_pre_clip': 10.352529525756836, 'timing/data': 0.005713368067517877, 'timing/model': 1.200162892928347, 'learning_rate/qwen_vl_interface': 7.318137077169416e-07, 'learning_rate/action_model': 7.318137077169415e-06, 'epoch': 1.86}) +[2026-06-17 18:01:42,451][starVLA.training.train_starvla][INFO] - Step 6312, Loss: {'train/loss': 0.8288323283195496, 'train/grad_norm_pre_clip': 16.48432159423828, 'timing/data': 0.005635957000777125, 'timing/model': 1.3399293299298733, 'learning_rate/qwen_vl_interface': 7.311468044607425e-07, 'learning_rate/action_model': 7.311468044607425e-06, 'epoch': 1.86}) +[2026-06-17 18:01:47,119][starVLA.training.train_starvla][INFO] - Step 6313, Loss: {'train/loss': 0.7665315866470337, 'train/grad_norm_pre_clip': 14.596691131591797, 'timing/data': 0.005706088966690004, 'timing/model': 1.2024547139881179, 'learning_rate/qwen_vl_interface': 7.304808379676739e-07, 'learning_rate/action_model': 7.3048083796767394e-06, 'epoch': 1.86}) +[2026-06-17 18:01:51,940][starVLA.training.train_starvla][INFO] - Step 6314, Loss: {'train/loss': 0.7768077850341797, 'train/grad_norm_pre_clip': 16.670698165893555, 'timing/data': 0.005521274055354297, 'timing/model': 1.2025979070458561, 'learning_rate/qwen_vl_interface': 7.298158083757903e-07, 'learning_rate/action_model': 7.298158083757903e-06, 'epoch': 1.86}) +[2026-06-17 18:01:56,791][starVLA.training.train_starvla][INFO] - Step 6315, Loss: {'train/loss': 0.3913858234882355, 'train/grad_norm_pre_clip': 11.498638153076172, 'timing/data': 0.0055808479664847255, 'timing/model': 1.2052376479841769, 'learning_rate/qwen_vl_interface': 7.291517158229536e-07, 'learning_rate/action_model': 7.291517158229535e-06, 'epoch': 1.86}) +[2026-06-17 18:02:01,537][starVLA.training.train_starvla][INFO] - Step 6316, Loss: {'train/loss': 0.6514129638671875, 'train/grad_norm_pre_clip': 12.406668663024902, 'timing/data': 0.005501130013726652, 'timing/model': 1.2073460520477965, 'learning_rate/qwen_vl_interface': 7.284885604468312e-07, 'learning_rate/action_model': 7.284885604468311e-06, 'epoch': 1.86}) +[2026-06-17 18:02:06,316][starVLA.training.train_starvla][INFO] - Step 6317, Loss: {'train/loss': 0.8053855895996094, 'train/grad_norm_pre_clip': 21.608322143554688, 'timing/data': 0.0055545150535181165, 'timing/model': 1.2025267760036513, 'learning_rate/qwen_vl_interface': 7.278263423848948e-07, 'learning_rate/action_model': 7.278263423848948e-06, 'epoch': 1.86}) +[2026-06-17 18:02:11,107][starVLA.training.train_starvla][INFO] - Step 6318, Loss: {'train/loss': 0.6271881461143494, 'train/grad_norm_pre_clip': 14.5389404296875, 'timing/data': 0.005619789008051157, 'timing/model': 1.3354446650482714, 'learning_rate/qwen_vl_interface': 7.271650617744239e-07, 'learning_rate/action_model': 7.271650617744238e-06, 'epoch': 1.86}) +[2026-06-17 18:02:15,808][starVLA.training.train_starvla][INFO] - Step 6319, Loss: {'train/loss': 0.6556596159934998, 'train/grad_norm_pre_clip': 14.792158126831055, 'timing/data': 0.005836488911882043, 'timing/model': 1.2031253320164979, 'learning_rate/qwen_vl_interface': 7.265047187525033e-07, 'learning_rate/action_model': 7.265047187525033e-06, 'epoch': 1.86}) +[2026-06-17 18:02:20,599][starVLA.training.train_starvla][INFO] - Step 6320, Loss: {'train/loss': 0.6253132820129395, 'train/grad_norm_pre_clip': 13.603073120117188, 'timing/data': 0.005505650071427226, 'timing/model': 1.204970536986366, 'learning_rate/qwen_vl_interface': 7.258453134560212e-07, 'learning_rate/action_model': 7.258453134560211e-06, 'epoch': 1.86}) +[2026-06-17 18:02:25,401][starVLA.training.train_starvla][INFO] - Step 6321, Loss: {'train/loss': 0.8498268127441406, 'train/grad_norm_pre_clip': 17.190998077392578, 'timing/data': 0.005313442088663578, 'timing/model': 1.2027546080062166, 'learning_rate/qwen_vl_interface': 7.251868460216743e-07, 'learning_rate/action_model': 7.251868460216743e-06, 'epoch': 1.86}) +[2026-06-17 18:02:30,195][starVLA.training.train_starvla][INFO] - Step 6322, Loss: {'train/loss': 0.627501904964447, 'train/grad_norm_pre_clip': 22.742700576782227, 'timing/data': 0.0057682410115376115, 'timing/model': 1.2092988080112264, 'learning_rate/qwen_vl_interface': 7.245293165859629e-07, 'learning_rate/action_model': 7.245293165859629e-06, 'epoch': 1.86}) +[2026-06-17 18:02:34,864][starVLA.training.train_starvla][INFO] - Step 6323, Loss: {'train/loss': 0.5346457362174988, 'train/grad_norm_pre_clip': 12.110737800598145, 'timing/data': 0.005351676954887807, 'timing/model': 1.2011585670989007, 'learning_rate/qwen_vl_interface': 7.238727252851935e-07, 'learning_rate/action_model': 7.238727252851935e-06, 'epoch': 1.86}) +[2026-06-17 18:02:39,561][starVLA.training.train_starvla][INFO] - Step 6324, Loss: {'train/loss': 0.8489217758178711, 'train/grad_norm_pre_clip': 16.976856231689453, 'timing/data': 0.00551767204888165, 'timing/model': 1.2455277100671083, 'learning_rate/qwen_vl_interface': 7.23217072255479e-07, 'learning_rate/action_model': 7.23217072255479e-06, 'epoch': 1.86}) +[2026-06-17 18:02:44,216][starVLA.training.train_starvla][INFO] - Step 6325, Loss: {'train/loss': 0.3754788041114807, 'train/grad_norm_pre_clip': 11.305264472961426, 'timing/data': 0.005814612959511578, 'timing/model': 1.2042814360465854, 'learning_rate/qwen_vl_interface': 7.225623576327361e-07, 'learning_rate/action_model': 7.22562357632736e-06, 'epoch': 1.86}) +[2026-06-17 18:02:48,944][starVLA.training.train_starvla][INFO] - Step 6326, Loss: {'train/loss': 0.39809373021125793, 'train/grad_norm_pre_clip': 15.177608489990234, 'timing/data': 0.005522963008843362, 'timing/model': 1.202545264037326, 'learning_rate/qwen_vl_interface': 7.219085815526887e-07, 'learning_rate/action_model': 7.219085815526887e-06, 'epoch': 1.86}) +[2026-06-17 18:02:53,733][starVLA.training.train_starvla][INFO] - Step 6327, Loss: {'train/loss': 0.6826748847961426, 'train/grad_norm_pre_clip': 14.438315391540527, 'timing/data': 0.005424560979008675, 'timing/model': 1.203569463104941, 'learning_rate/qwen_vl_interface': 7.212557441508641e-07, 'learning_rate/action_model': 7.212557441508641e-06, 'epoch': 1.86}) +[2026-06-17 18:02:58,521][starVLA.training.train_starvla][INFO] - Step 6328, Loss: {'train/loss': 1.0783145427703857, 'train/grad_norm_pre_clip': 19.625171661376953, 'timing/data': 0.005483848042786121, 'timing/model': 1.2034045300679281, 'learning_rate/qwen_vl_interface': 7.206038455625972e-07, 'learning_rate/action_model': 7.206038455625971e-06, 'epoch': 1.86}) +[2026-06-17 18:03:03,307][starVLA.training.train_starvla][INFO] - Step 6329, Loss: {'train/loss': 0.571336567401886, 'train/grad_norm_pre_clip': 14.005622863769531, 'timing/data': 0.005433415994048119, 'timing/model': 1.3316598030505702, 'learning_rate/qwen_vl_interface': 7.199528859230271e-07, 'learning_rate/action_model': 7.199528859230272e-06, 'epoch': 1.86}) +[2026-06-17 18:03:08,017][starVLA.training.train_starvla][INFO] - Step 6330, Loss: {'train/loss': 0.7310941219329834, 'train/grad_norm_pre_clip': 13.752384185791016, 'timing/data': 0.005658323992975056, 'timing/model': 1.2549643989186734, 'learning_rate/qwen_vl_interface': 7.19302865367098e-07, 'learning_rate/action_model': 7.19302865367098e-06, 'epoch': 1.86}) +[2026-06-17 18:03:12,672][starVLA.training.train_starvla][INFO] - Step 6331, Loss: {'train/loss': 0.9300480484962463, 'train/grad_norm_pre_clip': 16.09576416015625, 'timing/data': 0.005587083054706454, 'timing/model': 1.2045092709595338, 'learning_rate/qwen_vl_interface': 7.186537840295599e-07, 'learning_rate/action_model': 7.186537840295598e-06, 'epoch': 1.86}) +[2026-06-17 18:03:17,471][starVLA.training.train_starvla][INFO] - Step 6332, Loss: {'train/loss': 0.679686427116394, 'train/grad_norm_pre_clip': 17.434106826782227, 'timing/data': 0.0057050129398703575, 'timing/model': 1.2071156309684739, 'learning_rate/qwen_vl_interface': 7.18005642044969e-07, 'learning_rate/action_model': 7.180056420449689e-06, 'epoch': 1.86}) +[2026-06-17 18:03:22,273][starVLA.training.train_starvla][INFO] - Step 6333, Loss: {'train/loss': 0.5488765239715576, 'train/grad_norm_pre_clip': 13.203201293945312, 'timing/data': 0.005343400058336556, 'timing/model': 1.2070744259981439, 'learning_rate/qwen_vl_interface': 7.173584395476852e-07, 'learning_rate/action_model': 7.173584395476851e-06, 'epoch': 1.86}) +[2026-06-17 18:03:27,058][starVLA.training.train_starvla][INFO] - Step 6334, Loss: {'train/loss': 0.42264044284820557, 'train/grad_norm_pre_clip': 9.968029975891113, 'timing/data': 0.00550681899767369, 'timing/model': 1.2039340799674392, 'learning_rate/qwen_vl_interface': 7.16712176671874e-07, 'learning_rate/action_model': 7.16712176671874e-06, 'epoch': 1.86}) +[2026-06-17 18:03:31,862][starVLA.training.train_starvla][INFO] - Step 6335, Loss: {'train/loss': 0.5164635181427002, 'train/grad_norm_pre_clip': 13.342367172241211, 'timing/data': 0.005952978041023016, 'timing/model': 1.3435580959776416, 'learning_rate/qwen_vl_interface': 7.160668535515059e-07, 'learning_rate/action_model': 7.1606685355150585e-06, 'epoch': 1.86}) +[2026-06-17 18:03:36,653][starVLA.training.train_starvla][INFO] - Step 6336, Loss: {'train/loss': 0.4300541281700134, 'train/grad_norm_pre_clip': 10.500100135803223, 'timing/data': 0.0058115069987252355, 'timing/model': 1.3320875630015507, 'learning_rate/qwen_vl_interface': 7.154224703203584e-07, 'learning_rate/action_model': 7.154224703203584e-06, 'epoch': 1.86}) +[2026-06-17 18:03:41,308][starVLA.training.train_starvla][INFO] - Step 6337, Loss: {'train/loss': 0.7301716804504395, 'train/grad_norm_pre_clip': 15.018475532531738, 'timing/data': 0.0058575100265443325, 'timing/model': 1.2040129019878805, 'learning_rate/qwen_vl_interface': 7.147790271120122e-07, 'learning_rate/action_model': 7.147790271120122e-06, 'epoch': 1.86}) +[2026-06-17 18:03:46,066][starVLA.training.train_starvla][INFO] - Step 6338, Loss: {'train/loss': 0.4817960262298584, 'train/grad_norm_pre_clip': 11.636414527893066, 'timing/data': 0.005904686055146158, 'timing/model': 1.206525300978683, 'learning_rate/qwen_vl_interface': 7.141365240598545e-07, 'learning_rate/action_model': 7.141365240598545e-06, 'epoch': 1.86}) +[2026-06-17 18:03:50,853][starVLA.training.train_starvla][INFO] - Step 6339, Loss: {'train/loss': 0.7834681272506714, 'train/grad_norm_pre_clip': 15.76385498046875, 'timing/data': 0.006335104000754654, 'timing/model': 1.2076769759878516, 'learning_rate/qwen_vl_interface': 7.134949612970749e-07, 'learning_rate/action_model': 7.134949612970748e-06, 'epoch': 1.87}) +[2026-06-17 18:03:55,643][starVLA.training.train_starvla][INFO] - Step 6340, Loss: {'train/loss': 0.6071432828903198, 'train/grad_norm_pre_clip': 15.111100196838379, 'timing/data': 0.005617519025690854, 'timing/model': 1.2052251030690968, 'learning_rate/qwen_vl_interface': 7.128543389566718e-07, 'learning_rate/action_model': 7.128543389566718e-06, 'epoch': 1.87}) +[2026-06-17 18:04:00,381][starVLA.training.train_starvla][INFO] - Step 6341, Loss: {'train/loss': 0.5704888105392456, 'train/grad_norm_pre_clip': 12.333786964416504, 'timing/data': 0.0055713370675221086, 'timing/model': 1.2757854610681534, 'learning_rate/qwen_vl_interface': 7.122146571714464e-07, 'learning_rate/action_model': 7.122146571714463e-06, 'epoch': 1.87}) +[2026-06-17 18:04:05,039][starVLA.training.train_starvla][INFO] - Step 6342, Loss: {'train/loss': 0.6895978450775146, 'train/grad_norm_pre_clip': 63.69719696044922, 'timing/data': 0.005991748999804258, 'timing/model': 1.2063353740377352, 'learning_rate/qwen_vl_interface': 7.115759160740058e-07, 'learning_rate/action_model': 7.115759160740057e-06, 'epoch': 1.87}) +[2026-06-17 18:04:09,833][starVLA.training.train_starvla][INFO] - Step 6343, Loss: {'train/loss': 0.6368051767349243, 'train/grad_norm_pre_clip': 15.937140464782715, 'timing/data': 0.005666400073096156, 'timing/model': 1.2174658389994875, 'learning_rate/qwen_vl_interface': 7.109381157967607e-07, 'learning_rate/action_model': 7.109381157967607e-06, 'epoch': 1.87}) +[2026-06-17 18:04:14,628][starVLA.training.train_starvla][INFO] - Step 6344, Loss: {'train/loss': 0.8057324886322021, 'train/grad_norm_pre_clip': 19.095199584960938, 'timing/data': 0.005681647919118404, 'timing/model': 1.2066125490237027, 'learning_rate/qwen_vl_interface': 7.103012564719289e-07, 'learning_rate/action_model': 7.103012564719288e-06, 'epoch': 1.87}) +[2026-06-17 18:04:19,426][starVLA.training.train_starvla][INFO] - Step 6345, Loss: {'train/loss': 0.5258150696754456, 'train/grad_norm_pre_clip': 12.066831588745117, 'timing/data': 0.00544211408123374, 'timing/model': 1.2062792090000585, 'learning_rate/qwen_vl_interface': 7.096653382315315e-07, 'learning_rate/action_model': 7.096653382315315e-06, 'epoch': 1.87}) +[2026-06-17 18:04:24,211][starVLA.training.train_starvla][INFO] - Step 6346, Loss: {'train/loss': 0.6296176910400391, 'train/grad_norm_pre_clip': 15.612253189086914, 'timing/data': 0.005782774998806417, 'timing/model': 1.3328612929908559, 'learning_rate/qwen_vl_interface': 7.090303612073946e-07, 'learning_rate/action_model': 7.090303612073946e-06, 'epoch': 1.87}) +[2026-06-17 18:04:28,963][starVLA.training.train_starvla][INFO] - Step 6347, Loss: {'train/loss': 0.5167186260223389, 'train/grad_norm_pre_clip': 11.84299373626709, 'timing/data': 0.005604435107670724, 'timing/model': 1.2810203260742128, 'learning_rate/qwen_vl_interface': 7.083963255311508e-07, 'learning_rate/action_model': 7.083963255311509e-06, 'epoch': 1.87}) +[2026-06-17 18:04:33,630][starVLA.training.train_starvla][INFO] - Step 6348, Loss: {'train/loss': 0.40784743428230286, 'train/grad_norm_pre_clip': 53.38914108276367, 'timing/data': 0.0054160659201443195, 'timing/model': 1.2040369421010837, 'learning_rate/qwen_vl_interface': 7.07763231334236e-07, 'learning_rate/action_model': 7.077632313342359e-06, 'epoch': 1.87}) +[2026-06-17 18:04:38,423][starVLA.training.train_starvla][INFO] - Step 6349, Loss: {'train/loss': 0.5887361764907837, 'train/grad_norm_pre_clip': 13.334149360656738, 'timing/data': 0.005643157986924052, 'timing/model': 1.2038174159824848, 'learning_rate/qwen_vl_interface': 7.071310787478902e-07, 'learning_rate/action_model': 7.071310787478902e-06, 'epoch': 1.87}) +[2026-06-17 18:04:43,220][starVLA.training.train_starvla][INFO] - Step 6350, Loss: {'train/loss': 0.6948657035827637, 'train/grad_norm_pre_clip': 24.95900535583496, 'timing/data': 0.005463079083710909, 'timing/model': 1.2051815220620483, 'learning_rate/qwen_vl_interface': 7.0649986790316e-07, 'learning_rate/action_model': 7.0649986790315995e-06, 'epoch': 1.87}) +[2026-06-17 18:04:48,013][starVLA.training.train_starvla][INFO] - Step 6351, Loss: {'train/loss': 0.6864539980888367, 'train/grad_norm_pre_clip': 13.231757164001465, 'timing/data': 0.005564228049479425, 'timing/model': 1.2038011869881302, 'learning_rate/qwen_vl_interface': 7.058695989308965e-07, 'learning_rate/action_model': 7.058695989308965e-06, 'epoch': 1.87}) +[2026-06-17 18:04:52,797][starVLA.training.train_starvla][INFO] - Step 6352, Loss: {'train/loss': 0.7930456399917603, 'train/grad_norm_pre_clip': 15.083700180053711, 'timing/data': 0.005360265029594302, 'timing/model': 1.2045998240355402, 'learning_rate/qwen_vl_interface': 7.052402719617554e-07, 'learning_rate/action_model': 7.052402719617554e-06, 'epoch': 1.87}) +[2026-06-17 18:04:57,541][starVLA.training.train_starvla][INFO] - Step 6353, Loss: {'train/loss': 0.6476151943206787, 'train/grad_norm_pre_clip': 13.91477108001709, 'timing/data': 0.005487606977112591, 'timing/model': 1.2900960700353608, 'learning_rate/qwen_vl_interface': 7.046118871261952e-07, 'learning_rate/action_model': 7.0461188712619514e-06, 'epoch': 1.87}) +[2026-06-17 18:05:02,213][starVLA.training.train_starvla][INFO] - Step 6354, Loss: {'train/loss': 0.6347060799598694, 'train/grad_norm_pre_clip': 27.173498153686523, 'timing/data': 0.005376440938562155, 'timing/model': 1.2068714059423655, 'learning_rate/qwen_vl_interface': 7.039844445544828e-07, 'learning_rate/action_model': 7.039844445544827e-06, 'epoch': 1.87}) +[2026-06-17 18:05:07,003][starVLA.training.train_starvla][INFO] - Step 6355, Loss: {'train/loss': 0.6931326389312744, 'train/grad_norm_pre_clip': 17.775175094604492, 'timing/data': 0.005525127984583378, 'timing/model': 1.2085727429948747, 'learning_rate/qwen_vl_interface': 7.033579443766864e-07, 'learning_rate/action_model': 7.033579443766864e-06, 'epoch': 1.87}) +[2026-06-17 18:05:11,797][starVLA.training.train_starvla][INFO] - Step 6356, Loss: {'train/loss': 0.4969395101070404, 'train/grad_norm_pre_clip': 11.794816017150879, 'timing/data': 0.005602278979495168, 'timing/model': 1.203700948972255, 'learning_rate/qwen_vl_interface': 7.027323867226807e-07, 'learning_rate/action_model': 7.0273238672268065e-06, 'epoch': 1.87}) +[2026-06-17 18:05:16,591][starVLA.training.train_starvla][INFO] - Step 6357, Loss: {'train/loss': 0.634290337562561, 'train/grad_norm_pre_clip': 14.302183151245117, 'timing/data': 0.005680378060787916, 'timing/model': 1.2113601319724694, 'learning_rate/qwen_vl_interface': 7.021077717221441e-07, 'learning_rate/action_model': 7.0210777172214415e-06, 'epoch': 1.87}) +[2026-06-17 18:05:21,429][starVLA.training.train_starvla][INFO] - Step 6358, Loss: {'train/loss': 0.5213902592658997, 'train/grad_norm_pre_clip': 14.238706588745117, 'timing/data': 0.005540206911973655, 'timing/model': 1.340189330978319, 'learning_rate/qwen_vl_interface': 7.014840995045602e-07, 'learning_rate/action_model': 7.014840995045602e-06, 'epoch': 1.87}) +[2026-06-17 18:05:26,114][starVLA.training.train_starvla][INFO] - Step 6359, Loss: {'train/loss': 0.688868522644043, 'train/grad_norm_pre_clip': 14.367727279663086, 'timing/data': 0.005255149910226464, 'timing/model': 1.2059256799984723, 'learning_rate/qwen_vl_interface': 7.008613701992172e-07, 'learning_rate/action_model': 7.008613701992171e-06, 'epoch': 1.87}) +[2026-06-17 18:05:30,907][starVLA.training.train_starvla][INFO] - Step 6360, Loss: {'train/loss': 0.36529192328453064, 'train/grad_norm_pre_clip': 12.038103103637695, 'timing/data': 0.0057750840205699205, 'timing/model': 1.2071256469935179, 'learning_rate/qwen_vl_interface': 7.002395839352063e-07, 'learning_rate/action_model': 7.002395839352063e-06, 'epoch': 1.87}) +[2026-06-17 18:05:35,698][starVLA.training.train_starvla][INFO] - Step 6361, Loss: {'train/loss': 0.5785326361656189, 'train/grad_norm_pre_clip': 14.569838523864746, 'timing/data': 0.005388114950619638, 'timing/model': 1.206041265046224, 'learning_rate/qwen_vl_interface': 6.996187408414268e-07, 'learning_rate/action_model': 6.996187408414268e-06, 'epoch': 1.87}) +[2026-06-17 18:05:40,411][starVLA.training.train_starvla][INFO] - Step 6362, Loss: {'train/loss': 0.7444744110107422, 'train/grad_norm_pre_clip': 15.980596542358398, 'timing/data': 0.005949362996034324, 'timing/model': 1.2056880199816078, 'learning_rate/qwen_vl_interface': 6.98998841046578e-07, 'learning_rate/action_model': 6.989988410465779e-06, 'epoch': 1.87}) +[2026-06-17 18:05:45,153][starVLA.training.train_starvla][INFO] - Step 6363, Loss: {'train/loss': 0.4768603444099426, 'train/grad_norm_pre_clip': 14.555338859558105, 'timing/data': 0.005695042083971202, 'timing/model': 1.2052033130312338, 'learning_rate/qwen_vl_interface': 6.983798846791667e-07, 'learning_rate/action_model': 6.983798846791667e-06, 'epoch': 1.87}) +[2026-06-17 18:05:49,956][starVLA.training.train_starvla][INFO] - Step 6364, Loss: {'train/loss': 0.6485971808433533, 'train/grad_norm_pre_clip': 27.37225914001465, 'timing/data': 0.005211001029238105, 'timing/model': 1.336623959010467, 'learning_rate/qwen_vl_interface': 6.977618718675023e-07, 'learning_rate/action_model': 6.977618718675024e-06, 'epoch': 1.87}) +[2026-06-17 18:05:54,623][starVLA.training.train_starvla][INFO] - Step 6365, Loss: {'train/loss': 0.6315248012542725, 'train/grad_norm_pre_clip': 14.6567964553833, 'timing/data': 0.005461832974106073, 'timing/model': 1.2076866060961038, 'learning_rate/qwen_vl_interface': 6.971448027397007e-07, 'learning_rate/action_model': 6.971448027397007e-06, 'epoch': 1.87}) +[2026-06-17 18:05:59,501][starVLA.training.train_starvla][INFO] - Step 6366, Loss: {'train/loss': 0.4754834473133087, 'train/grad_norm_pre_clip': 12.805712699890137, 'timing/data': 0.005655757035128772, 'timing/model': 1.2052764430409297, 'learning_rate/qwen_vl_interface': 6.965286774236814e-07, 'learning_rate/action_model': 6.965286774236813e-06, 'epoch': 1.87}) +[2026-06-17 18:06:04,285][starVLA.training.train_starvla][INFO] - Step 6367, Loss: {'train/loss': 0.5157396793365479, 'train/grad_norm_pre_clip': 14.823029518127441, 'timing/data': 0.005550364963710308, 'timing/model': 1.2052250750130042, 'learning_rate/qwen_vl_interface': 6.959134960471664e-07, 'learning_rate/action_model': 6.959134960471664e-06, 'epoch': 1.87}) +[2026-06-17 18:06:09,070][starVLA.training.train_starvla][INFO] - Step 6368, Loss: {'train/loss': 0.3732634484767914, 'train/grad_norm_pre_clip': 10.558595657348633, 'timing/data': 0.005540351965464652, 'timing/model': 1.2107258379692212, 'learning_rate/qwen_vl_interface': 6.952992587376836e-07, 'learning_rate/action_model': 6.952992587376836e-06, 'epoch': 1.87}) +[2026-06-17 18:06:13,819][starVLA.training.train_starvla][INFO] - Step 6369, Loss: {'train/loss': 0.6320188641548157, 'train/grad_norm_pre_clip': 47.69919204711914, 'timing/data': 0.005545528954826295, 'timing/model': 1.2058405850548297, 'learning_rate/qwen_vl_interface': 6.946859656225662e-07, 'learning_rate/action_model': 6.946859656225661e-06, 'epoch': 1.87}) +[2026-06-17 18:06:18,605][starVLA.training.train_starvla][INFO] - Step 6370, Loss: {'train/loss': 0.8359059691429138, 'train/grad_norm_pre_clip': 23.395780563354492, 'timing/data': 0.005540821002796292, 'timing/model': 1.3264410450356081, 'learning_rate/qwen_vl_interface': 6.940736168289497e-07, 'learning_rate/action_model': 6.940736168289497e-06, 'epoch': 1.87}) +[2026-06-17 18:06:23,262][starVLA.training.train_starvla][INFO] - Step 6371, Loss: {'train/loss': 0.5584965348243713, 'train/grad_norm_pre_clip': 14.980605125427246, 'timing/data': 0.0050506379920989275, 'timing/model': 1.2038302270229906, 'learning_rate/qwen_vl_interface': 6.934622124837751e-07, 'learning_rate/action_model': 6.93462212483775e-06, 'epoch': 1.87}) +[2026-06-17 18:06:28,052][starVLA.training.train_starvla][INFO] - Step 6372, Loss: {'train/loss': 0.5583788156509399, 'train/grad_norm_pre_clip': 11.330238342285156, 'timing/data': 0.0058433120138943195, 'timing/model': 1.2049263290828094, 'learning_rate/qwen_vl_interface': 6.928517527137864e-07, 'learning_rate/action_model': 6.9285175271378635e-06, 'epoch': 1.87}) +[2026-06-17 18:06:32,839][starVLA.training.train_starvla][INFO] - Step 6373, Loss: {'train/loss': 0.7559493780136108, 'train/grad_norm_pre_clip': 18.204959869384766, 'timing/data': 0.0060343139339238405, 'timing/model': 1.2055043779546395, 'learning_rate/qwen_vl_interface': 6.922422376455333e-07, 'learning_rate/action_model': 6.9224223764553326e-06, 'epoch': 1.88}) +[2026-06-17 18:06:37,625][starVLA.training.train_starvla][INFO] - Step 6374, Loss: {'train/loss': 0.841083824634552, 'train/grad_norm_pre_clip': 15.428722381591797, 'timing/data': 0.005296926945447922, 'timing/model': 1.2033414680045098, 'learning_rate/qwen_vl_interface': 6.916336674053688e-07, 'learning_rate/action_model': 6.9163366740536875e-06, 'epoch': 1.88}) +[2026-06-17 18:06:42,416][starVLA.training.train_starvla][INFO] - Step 6375, Loss: {'train/loss': 0.5227250456809998, 'train/grad_norm_pre_clip': 15.320505142211914, 'timing/data': 0.00558246998116374, 'timing/model': 1.3366004299605265, 'learning_rate/qwen_vl_interface': 6.9102604211945e-07, 'learning_rate/action_model': 6.910260421194501e-06, 'epoch': 1.88}) +[2026-06-17 18:06:47,078][starVLA.training.train_starvla][INFO] - Step 6376, Loss: {'train/loss': 0.8701398968696594, 'train/grad_norm_pre_clip': 21.846729278564453, 'timing/data': 0.005929924082010984, 'timing/model': 1.20621408498846, 'learning_rate/qwen_vl_interface': 6.904193619137381e-07, 'learning_rate/action_model': 6.904193619137381e-06, 'epoch': 1.88}) +[2026-06-17 18:06:51,910][starVLA.training.train_starvla][INFO] - Step 6377, Loss: {'train/loss': 0.529685914516449, 'train/grad_norm_pre_clip': 13.82014274597168, 'timing/data': 0.006000164081342518, 'timing/model': 1.2049189700046554, 'learning_rate/qwen_vl_interface': 6.898136269139989e-07, 'learning_rate/action_model': 6.8981362691399885e-06, 'epoch': 1.88}) +[2026-06-17 18:06:56,691][starVLA.training.train_starvla][INFO] - Step 6378, Loss: {'train/loss': 0.50621497631073, 'train/grad_norm_pre_clip': 11.51853084564209, 'timing/data': 0.0055258870124816895, 'timing/model': 1.2061326610855758, 'learning_rate/qwen_vl_interface': 6.89208837245802e-07, 'learning_rate/action_model': 6.892088372458019e-06, 'epoch': 1.88}) +[2026-06-17 18:07:01,485][starVLA.training.train_starvla][INFO] - Step 6379, Loss: {'train/loss': 0.5775662660598755, 'train/grad_norm_pre_clip': 13.883458137512207, 'timing/data': 0.005498511949554086, 'timing/model': 1.2086666401010007, 'learning_rate/qwen_vl_interface': 6.886049930345204e-07, 'learning_rate/action_model': 6.8860499303452035e-06, 'epoch': 1.88}) +[2026-06-17 18:07:06,278][starVLA.training.train_starvla][INFO] - Step 6380, Loss: {'train/loss': 0.5284202098846436, 'train/grad_norm_pre_clip': 17.684490203857422, 'timing/data': 0.005493622971698642, 'timing/model': 1.2048469959991053, 'learning_rate/qwen_vl_interface': 6.880020944053326e-07, 'learning_rate/action_model': 6.880020944053325e-06, 'epoch': 1.88}) +[2026-06-17 18:07:11,078][starVLA.training.train_starvla][INFO] - Step 6381, Loss: {'train/loss': 0.47380155324935913, 'train/grad_norm_pre_clip': 11.321871757507324, 'timing/data': 0.005526473978534341, 'timing/model': 1.3339726449921727, 'learning_rate/qwen_vl_interface': 6.87400141483219e-07, 'learning_rate/action_model': 6.87400141483219e-06, 'epoch': 1.88}) +[2026-06-17 18:07:15,774][starVLA.training.train_starvla][INFO] - Step 6382, Loss: {'train/loss': 0.5976068377494812, 'train/grad_norm_pre_clip': 13.847145080566406, 'timing/data': 0.005534962983801961, 'timing/model': 1.2051790510304272, 'learning_rate/qwen_vl_interface': 6.867991343929655e-07, 'learning_rate/action_model': 6.867991343929656e-06, 'epoch': 1.88}) +[2026-06-17 18:07:20,558][starVLA.training.train_starvla][INFO] - Step 6383, Loss: {'train/loss': 0.5655755996704102, 'train/grad_norm_pre_clip': 12.306838035583496, 'timing/data': 0.005723912036046386, 'timing/model': 1.205921986955218, 'learning_rate/qwen_vl_interface': 6.861990732591622e-07, 'learning_rate/action_model': 6.861990732591622e-06, 'epoch': 1.88}) +[2026-06-17 18:07:25,348][starVLA.training.train_starvla][INFO] - Step 6384, Loss: {'train/loss': 0.7002593874931335, 'train/grad_norm_pre_clip': 46.528446197509766, 'timing/data': 0.005586649989709258, 'timing/model': 1.2045657880371436, 'learning_rate/qwen_vl_interface': 6.855999582062017e-07, 'learning_rate/action_model': 6.855999582062017e-06, 'epoch': 1.88}) +[2026-06-17 18:07:30,095][starVLA.training.train_starvla][INFO] - Step 6385, Loss: {'train/loss': 0.5285078883171082, 'train/grad_norm_pre_clip': 14.05127239227295, 'timing/data': 0.0060846490086987615, 'timing/model': 1.202743444009684, 'learning_rate/qwen_vl_interface': 6.850017893582818e-07, 'learning_rate/action_model': 6.850017893582818e-06, 'epoch': 1.88}) +[2026-06-17 18:07:34,881][starVLA.training.train_starvla][INFO] - Step 6386, Loss: {'train/loss': 0.621890127658844, 'train/grad_norm_pre_clip': 16.332998275756836, 'timing/data': 0.00552624196279794, 'timing/model': 1.2082868210272864, 'learning_rate/qwen_vl_interface': 6.844045668394024e-07, 'learning_rate/action_model': 6.844045668394024e-06, 'epoch': 1.88}) +[2026-06-17 18:07:39,682][starVLA.training.train_starvla][INFO] - Step 6387, Loss: {'train/loss': 0.8198138475418091, 'train/grad_norm_pre_clip': 15.556262016296387, 'timing/data': 0.005629251012578607, 'timing/model': 1.336301477975212, 'learning_rate/qwen_vl_interface': 6.838082907733694e-07, 'learning_rate/action_model': 6.838082907733694e-06, 'epoch': 1.88}) +[2026-06-17 18:07:44,342][starVLA.training.train_starvla][INFO] - Step 6388, Loss: {'train/loss': 0.752996027469635, 'train/grad_norm_pre_clip': 17.020524978637695, 'timing/data': 0.005557722062803805, 'timing/model': 1.2048799000913277, 'learning_rate/qwen_vl_interface': 6.832129612837912e-07, 'learning_rate/action_model': 6.832129612837912e-06, 'epoch': 1.88}) +[2026-06-17 18:07:49,130][starVLA.training.train_starvla][INFO] - Step 6389, Loss: {'train/loss': 0.4908851981163025, 'train/grad_norm_pre_clip': 12.69253921508789, 'timing/data': 0.0053921910002827644, 'timing/model': 1.2032626039581373, 'learning_rate/qwen_vl_interface': 6.826185784940801e-07, 'learning_rate/action_model': 6.826185784940801e-06, 'epoch': 1.88}) +[2026-06-17 18:07:53,982][starVLA.training.train_starvla][INFO] - Step 6390, Loss: {'train/loss': 0.4121290445327759, 'train/grad_norm_pre_clip': 10.721331596374512, 'timing/data': 0.005829481058754027, 'timing/model': 1.2051517149666324, 'learning_rate/qwen_vl_interface': 6.820251425274526e-07, 'learning_rate/action_model': 6.820251425274527e-06, 'epoch': 1.88}) +[2026-06-17 18:07:58,877][starVLA.training.train_starvla][INFO] - Step 6391, Loss: {'train/loss': 0.43003273010253906, 'train/grad_norm_pre_clip': 14.415563583374023, 'timing/data': 0.005456627928651869, 'timing/model': 1.2055900420527905, 'learning_rate/qwen_vl_interface': 6.81432653506928e-07, 'learning_rate/action_model': 6.814326535069279e-06, 'epoch': 1.88}) +[2026-06-17 18:08:03,650][starVLA.training.train_starvla][INFO] - Step 6392, Loss: {'train/loss': 0.6036021709442139, 'train/grad_norm_pre_clip': 12.564922332763672, 'timing/data': 0.0055006350157782435, 'timing/model': 1.2058809790760279, 'learning_rate/qwen_vl_interface': 6.808411115553305e-07, 'learning_rate/action_model': 6.808411115553305e-06, 'epoch': 1.88}) +[2026-06-17 18:08:08,390][starVLA.training.train_starvla][INFO] - Step 6393, Loss: {'train/loss': 0.321990042924881, 'train/grad_norm_pre_clip': 8.60431957244873, 'timing/data': 0.005319638992659748, 'timing/model': 1.2440408339025453, 'learning_rate/qwen_vl_interface': 6.802505167952871e-07, 'learning_rate/action_model': 6.80250516795287e-06, 'epoch': 1.88}) +[2026-06-17 18:08:13,060][starVLA.training.train_starvla][INFO] - Step 6394, Loss: {'train/loss': 0.5243555307388306, 'train/grad_norm_pre_clip': 12.93757152557373, 'timing/data': 0.00539519009180367, 'timing/model': 1.2055284490343183, 'learning_rate/qwen_vl_interface': 6.796608693492293e-07, 'learning_rate/action_model': 6.7966086934922925e-06, 'epoch': 1.88}) +[2026-06-17 18:08:17,886][starVLA.training.train_starvla][INFO] - Step 6395, Loss: {'train/loss': 0.7021950483322144, 'train/grad_norm_pre_clip': 13.28527545928955, 'timing/data': 0.005329025094397366, 'timing/model': 1.207476960029453, 'learning_rate/qwen_vl_interface': 6.790721693393912e-07, 'learning_rate/action_model': 6.790721693393912e-06, 'epoch': 1.88}) +[2026-06-17 18:08:22,697][starVLA.training.train_starvla][INFO] - Step 6396, Loss: {'train/loss': 0.43208542466163635, 'train/grad_norm_pre_clip': 10.74863338470459, 'timing/data': 0.005632419022731483, 'timing/model': 1.2088564039440826, 'learning_rate/qwen_vl_interface': 6.784844168878108e-07, 'learning_rate/action_model': 6.784844168878107e-06, 'epoch': 1.88}) +[2026-06-17 18:08:27,479][starVLA.training.train_starvla][INFO] - Step 6397, Loss: {'train/loss': 0.621026337146759, 'train/grad_norm_pre_clip': 13.855548858642578, 'timing/data': 0.005616643000394106, 'timing/model': 1.2033052140614018, 'learning_rate/qwen_vl_interface': 6.778976121163289e-07, 'learning_rate/action_model': 6.778976121163288e-06, 'epoch': 1.88}) +[2026-06-17 18:08:32,267][starVLA.training.train_starvla][INFO] - Step 6398, Loss: {'train/loss': 0.5600051283836365, 'train/grad_norm_pre_clip': 14.139860153198242, 'timing/data': 0.0056550310691818595, 'timing/model': 1.328734359005466, 'learning_rate/qwen_vl_interface': 6.773117551465927e-07, 'learning_rate/action_model': 6.773117551465927e-06, 'epoch': 1.88}) +[2026-06-17 18:08:36,976][starVLA.training.train_starvla][INFO] - Step 6399, Loss: {'train/loss': 0.6810190081596375, 'train/grad_norm_pre_clip': 14.027749061584473, 'timing/data': 0.005592864006757736, 'timing/model': 1.2525712579954416, 'learning_rate/qwen_vl_interface': 6.767268461000498e-07, 'learning_rate/action_model': 6.7672684610004975e-06, 'epoch': 1.88}) +[2026-06-17 18:10:19,163][starVLA.training.train_starvla][INFO] - Step 6400, Loss: {'train/loss': 0.32140666246414185, 'train/grad_norm_pre_clip': 9.320839881896973, 'eval/loss': 0.703737735748291, 'eval/latency_0/loss': 0.703737735748291, 'eval/demon_fire_precision': 0.8929411764705882, 'eval/demon_fire_recall': 0.9455980066445183, 'eval/demon_fire_f1': 0.918515530455829, 'eval/demon_fire_rate_model': 0.7091212458286985, 'eval/demon_fire_rate_teacher': 0.6696329254727474, 'eval/demon_left_precision': 0.8424821002386634, 'eval/demon_left_recall': 0.8, 'eval/demon_left_f1': 0.8206916594013368, 'eval/demon_left_rate_model': 0.4660734149054505, 'eval/demon_left_rate_teacher': 0.49082313681868744, 'eval/demon_right_precision': 0.8128964059196617, 'eval/demon_right_recall': 0.8689265536723164, 'eval/demon_right_f1': 0.8399781540141998, 'eval/demon_right_rate_model': 0.5261401557285873, 'eval/demon_right_rate_teacher': 0.4922135706340378, 'eval/demon_move_f1': 0.8303349067077683, 'eval/demon_cc_f1': 0.8901554802471111, 'eval/latency_0/demon_fire_precision': 0.8929411764705882, 'eval/latency_0/demon_fire_recall': 0.9455980066445183, 'eval/latency_0/demon_fire_f1': 0.918515530455829, 'eval/latency_0/demon_fire_rate_model': 0.7091212458286985, 'eval/latency_0/demon_fire_rate_teacher': 0.6696329254727474, 'eval/latency_0/demon_left_precision': 0.8424821002386634, 'eval/latency_0/demon_left_recall': 0.8, 'eval/latency_0/demon_left_f1': 0.8206916594013368, 'eval/latency_0/demon_left_rate_model': 0.4660734149054505, 'eval/latency_0/demon_left_rate_teacher': 0.49082313681868744, 'eval/latency_0/demon_right_precision': 0.8128964059196617, 'eval/latency_0/demon_right_recall': 0.8689265536723164, 'eval/latency_0/demon_right_f1': 0.8399781540141998, 'eval/latency_0/demon_right_rate_model': 0.5261401557285873, 'eval/latency_0/demon_right_rate_teacher': 0.4922135706340378, 'eval/latency_0/demon_move_f1': 0.8303349067077683, 'eval/latency_0/demon_cc_f1': 0.8901554802471111, 'timing/data': 0.005577816045843065, 'timing/model': 1.2064441979164258, 'learning_rate/qwen_vl_interface': 6.761428850979542e-07, 'learning_rate/action_model': 6.7614288509795415e-06, 'epoch': 1.88}) +[2026-06-17 18:10:19,164][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_6400_state +[2026-06-17 18:10:19,174][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 18:10:22,747][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_6400_state/model.safetensors +[2026-06-17 18:10:30,355][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_6400_state/optimizer.bin +[2026-06-17 18:10:30,356][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_6400_state/sampler.bin +[2026-06-17 18:10:30,356][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_6400_state/sampler_1.bin +[2026-06-17 18:10:30,357][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_6400_state/random_states_0.pkl +[2026-06-17 18:10:30,399][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 6400 +[2026-06-17 18:10:30,399][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 18:10:30,409][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 18:10:34,992][starVLA.training.train_starvla][INFO] - Step 6401, Loss: {'train/loss': 0.724880039691925, 'train/grad_norm_pre_clip': 13.392014503479004, 'timing/data': 0.005718852975405753, 'timing/model': 1.189126661978662, 'learning_rate/qwen_vl_interface': 6.755598722613579e-07, 'learning_rate/action_model': 6.755598722613579e-06, 'epoch': 1.88}) +[2026-06-17 18:10:39,715][starVLA.training.train_starvla][INFO] - Step 6402, Loss: {'train/loss': 0.462126225233078, 'train/grad_norm_pre_clip': 11.537775993347168, 'timing/data': 0.00499107304494828, 'timing/model': 1.1892943810671568, 'learning_rate/qwen_vl_interface': 6.749778077111241e-07, 'learning_rate/action_model': 6.749778077111241e-06, 'epoch': 1.88}) +[2026-06-17 18:10:44,352][starVLA.training.train_starvla][INFO] - Step 6403, Loss: {'train/loss': 0.8198052048683167, 'train/grad_norm_pre_clip': 24.17200469970703, 'timing/data': 0.0047800010070204735, 'timing/model': 1.1943246880546212, 'learning_rate/qwen_vl_interface': 6.743966915679128e-07, 'learning_rate/action_model': 6.743966915679129e-06, 'epoch': 1.88}) +[2026-06-17 18:10:48,982][starVLA.training.train_starvla][INFO] - Step 6404, Loss: {'train/loss': 0.6394728422164917, 'train/grad_norm_pre_clip': 16.374757766723633, 'timing/data': 0.004970513051375747, 'timing/model': 1.1918328499887139, 'learning_rate/qwen_vl_interface': 6.738165239521909e-07, 'learning_rate/action_model': 6.7381652395219094e-06, 'epoch': 1.88}) +[2026-06-17 18:10:53,649][starVLA.training.train_starvla][INFO] - Step 6405, Loss: {'train/loss': 0.9375318288803101, 'train/grad_norm_pre_clip': 18.328397750854492, 'timing/data': 0.004972225986421108, 'timing/model': 1.1936137430602685, 'learning_rate/qwen_vl_interface': 6.732373049842284e-07, 'learning_rate/action_model': 6.7323730498422846e-06, 'epoch': 1.88}) +[2026-06-17 18:10:58,388][starVLA.training.train_starvla][INFO] - Step 6406, Loss: {'train/loss': 0.7346034049987793, 'train/grad_norm_pre_clip': 13.753751754760742, 'timing/data': 0.004913595039397478, 'timing/model': 1.317411977914162, 'learning_rate/qwen_vl_interface': 6.726590347840972e-07, 'learning_rate/action_model': 6.726590347840972e-06, 'epoch': 1.88}) +[2026-06-17 18:11:03,011][starVLA.training.train_starvla][INFO] - Step 6407, Loss: {'train/loss': 0.5916169285774231, 'train/grad_norm_pre_clip': 11.003128051757812, 'timing/data': 0.005221518920734525, 'timing/model': 1.1968880660133436, 'learning_rate/qwen_vl_interface': 6.720817134716736e-07, 'learning_rate/action_model': 6.720817134716736e-06, 'epoch': 1.89}) +[2026-06-17 18:11:07,748][starVLA.training.train_starvla][INFO] - Step 6408, Loss: {'train/loss': 0.5456269979476929, 'train/grad_norm_pre_clip': 13.702903747558594, 'timing/data': 0.005221973988227546, 'timing/model': 1.1970120620680973, 'learning_rate/qwen_vl_interface': 6.71505341166636e-07, 'learning_rate/action_model': 6.71505341166636e-06, 'epoch': 1.89}) +[2026-06-17 18:11:12,512][starVLA.training.train_starvla][INFO] - Step 6409, Loss: {'train/loss': 0.790849506855011, 'train/grad_norm_pre_clip': 14.97541332244873, 'timing/data': 0.005243582068942487, 'timing/model': 1.1992171689635143, 'learning_rate/qwen_vl_interface': 6.709299179884689e-07, 'learning_rate/action_model': 6.709299179884689e-06, 'epoch': 1.89}) +[2026-06-17 18:11:17,272][starVLA.training.train_starvla][INFO] - Step 6410, Loss: {'train/loss': 0.6373767256736755, 'train/grad_norm_pre_clip': 13.148680686950684, 'timing/data': 0.005116531974636018, 'timing/model': 1.200490747927688, 'learning_rate/qwen_vl_interface': 6.703554440564569e-07, 'learning_rate/action_model': 6.703554440564569e-06, 'epoch': 1.89}) +[2026-06-17 18:11:22,039][starVLA.training.train_starvla][INFO] - Step 6411, Loss: {'train/loss': 0.583503782749176, 'train/grad_norm_pre_clip': 14.055438041687012, 'timing/data': 0.00483376404736191, 'timing/model': 1.3275445710169151, 'learning_rate/qwen_vl_interface': 6.697819194896889e-07, 'learning_rate/action_model': 6.697819194896889e-06, 'epoch': 1.89}) +[2026-06-17 18:11:26,765][starVLA.training.train_starvla][INFO] - Step 6412, Loss: {'train/loss': 0.7045483589172363, 'train/grad_norm_pre_clip': 26.68714714050293, 'timing/data': 0.005071225925348699, 'timing/model': 1.280488496995531, 'learning_rate/qwen_vl_interface': 6.692093444070571e-07, 'learning_rate/action_model': 6.692093444070572e-06, 'epoch': 1.89}) +[2026-06-17 18:11:31,444][starVLA.training.train_starvla][INFO] - Step 6413, Loss: {'train/loss': 0.6490518450737, 'train/grad_norm_pre_clip': 12.388167381286621, 'timing/data': 0.005111981998197734, 'timing/model': 1.2006339929066598, 'learning_rate/qwen_vl_interface': 6.686377189272584e-07, 'learning_rate/action_model': 6.686377189272583e-06, 'epoch': 1.89}) +[2026-06-17 18:11:36,217][starVLA.training.train_starvla][INFO] - Step 6414, Loss: {'train/loss': 0.4074828624725342, 'train/grad_norm_pre_clip': 11.75648021697998, 'timing/data': 0.0053876020247116685, 'timing/model': 1.204785701003857, 'learning_rate/qwen_vl_interface': 6.680670431687898e-07, 'learning_rate/action_model': 6.680670431687898e-06, 'epoch': 1.89}) +[2026-06-17 18:11:40,994][starVLA.training.train_starvla][INFO] - Step 6415, Loss: {'train/loss': 0.6938266754150391, 'train/grad_norm_pre_clip': 14.90230941772461, 'timing/data': 0.004881680943071842, 'timing/model': 1.2015342268859968, 'learning_rate/qwen_vl_interface': 6.674973172499536e-07, 'learning_rate/action_model': 6.674973172499535e-06, 'epoch': 1.89}) +[2026-06-17 18:11:45,786][starVLA.training.train_starvla][INFO] - Step 6416, Loss: {'train/loss': 0.4911513030529022, 'train/grad_norm_pre_clip': 12.115090370178223, 'timing/data': 0.004891590913757682, 'timing/model': 1.2071712049655616, 'learning_rate/qwen_vl_interface': 6.669285412888544e-07, 'learning_rate/action_model': 6.669285412888544e-06, 'epoch': 1.89}) +[2026-06-17 18:11:50,569][starVLA.training.train_starvla][INFO] - Step 6417, Loss: {'train/loss': 0.7358139157295227, 'train/grad_norm_pre_clip': 15.569690704345703, 'timing/data': 0.0052708989242091775, 'timing/model': 1.2059001689776778, 'learning_rate/qwen_vl_interface': 6.663607154034003e-07, 'learning_rate/action_model': 6.663607154034003e-06, 'epoch': 1.89}) +[2026-06-17 18:11:55,323][starVLA.training.train_starvla][INFO] - Step 6418, Loss: {'train/loss': 0.7400302886962891, 'train/grad_norm_pre_clip': 17.278045654296875, 'timing/data': 0.005039363983087242, 'timing/model': 1.3012877349974588, 'learning_rate/qwen_vl_interface': 6.657938397113019e-07, 'learning_rate/action_model': 6.657938397113018e-06, 'epoch': 1.89}) +[2026-06-17 18:11:59,990][starVLA.training.train_starvla][INFO] - Step 6419, Loss: {'train/loss': 0.6576952338218689, 'train/grad_norm_pre_clip': 14.83056354522705, 'timing/data': 0.004919915925711393, 'timing/model': 1.2040238589979708, 'learning_rate/qwen_vl_interface': 6.652279143300734e-07, 'learning_rate/action_model': 6.652279143300734e-06, 'epoch': 1.89}) +[2026-06-17 18:12:04,777][starVLA.training.train_starvla][INFO] - Step 6420, Loss: {'train/loss': 0.5291603803634644, 'train/grad_norm_pre_clip': 12.47060489654541, 'timing/data': 0.005074706976301968, 'timing/model': 1.2004739160183817, 'learning_rate/qwen_vl_interface': 6.64662939377032e-07, 'learning_rate/action_model': 6.646629393770321e-06, 'epoch': 1.89}) +[2026-06-17 18:12:09,576][starVLA.training.train_starvla][INFO] - Step 6421, Loss: {'train/loss': 0.5962406396865845, 'train/grad_norm_pre_clip': 15.997369766235352, 'timing/data': 0.005282956990413368, 'timing/model': 1.202731690951623, 'learning_rate/qwen_vl_interface': 6.64098914969297e-07, 'learning_rate/action_model': 6.6409891496929695e-06, 'epoch': 1.89}) +[2026-06-17 18:12:14,385][starVLA.training.train_starvla][INFO] - Step 6422, Loss: {'train/loss': 0.8372318148612976, 'train/grad_norm_pre_clip': 14.943943977355957, 'timing/data': 0.005030906992033124, 'timing/model': 1.2012549699284136, 'learning_rate/qwen_vl_interface': 6.635358412237912e-07, 'learning_rate/action_model': 6.635358412237912e-06, 'epoch': 1.89}) +[2026-06-17 18:12:19,173][starVLA.training.train_starvla][INFO] - Step 6423, Loss: {'train/loss': 0.566947877407074, 'train/grad_norm_pre_clip': 15.683895111083984, 'timing/data': 0.004995147930458188, 'timing/model': 1.3335328459506854, 'learning_rate/qwen_vl_interface': 6.629737182572418e-07, 'learning_rate/action_model': 6.629737182572418e-06, 'epoch': 1.89}) +[2026-06-17 18:12:23,826][starVLA.training.train_starvla][INFO] - Step 6424, Loss: {'train/loss': 0.8031718730926514, 'train/grad_norm_pre_clip': 14.888842582702637, 'timing/data': 0.004894742975011468, 'timing/model': 1.2059531550621614, 'learning_rate/qwen_vl_interface': 6.62412546186176e-07, 'learning_rate/action_model': 6.62412546186176e-06, 'epoch': 1.89}) +[2026-06-17 18:12:28,612][starVLA.training.train_starvla][INFO] - Step 6425, Loss: {'train/loss': 0.3952414095401764, 'train/grad_norm_pre_clip': 10.657515525817871, 'timing/data': 0.004857989028096199, 'timing/model': 1.2024333380395547, 'learning_rate/qwen_vl_interface': 6.61852325126926e-07, 'learning_rate/action_model': 6.61852325126926e-06, 'epoch': 1.89}) +[2026-06-17 18:12:33,400][starVLA.training.train_starvla][INFO] - Step 6426, Loss: {'train/loss': 0.5364404320716858, 'train/grad_norm_pre_clip': 15.06114673614502, 'timing/data': 0.004996970063075423, 'timing/model': 1.2032146729761735, 'learning_rate/qwen_vl_interface': 6.61293055195625e-07, 'learning_rate/action_model': 6.6129305519562504e-06, 'epoch': 1.89}) +[2026-06-17 18:12:38,132][starVLA.training.train_starvla][INFO] - Step 6427, Loss: {'train/loss': 0.7442076206207275, 'train/grad_norm_pre_clip': 14.706514358520508, 'timing/data': 0.005127580021508038, 'timing/model': 1.2139563349774107, 'learning_rate/qwen_vl_interface': 6.607347365082127e-07, 'learning_rate/action_model': 6.607347365082127e-06, 'epoch': 1.89}) +[2026-06-17 18:12:42,879][starVLA.training.train_starvla][INFO] - Step 6428, Loss: {'train/loss': 0.946857213973999, 'train/grad_norm_pre_clip': 17.335662841796875, 'timing/data': 0.00517612905241549, 'timing/model': 1.2100480350200087, 'learning_rate/qwen_vl_interface': 6.601773691804273e-07, 'learning_rate/action_model': 6.6017736918042735e-06, 'epoch': 1.89}) +[2026-06-17 18:12:47,669][starVLA.training.train_starvla][INFO] - Step 6429, Loss: {'train/loss': 0.4871537685394287, 'train/grad_norm_pre_clip': 13.630516052246094, 'timing/data': 0.005066626006737351, 'timing/model': 1.3369392160093412, 'learning_rate/qwen_vl_interface': 6.596209533278122e-07, 'learning_rate/action_model': 6.596209533278122e-06, 'epoch': 1.89}) +[2026-06-17 18:12:52,351][starVLA.training.train_starvla][INFO] - Step 6430, Loss: {'train/loss': 0.5543842911720276, 'train/grad_norm_pre_clip': 14.830016136169434, 'timing/data': 0.00490224699024111, 'timing/model': 1.2028476659907028, 'learning_rate/qwen_vl_interface': 6.590654890657125e-07, 'learning_rate/action_model': 6.590654890657125e-06, 'epoch': 1.89}) +[2026-06-17 18:12:57,126][starVLA.training.train_starvla][INFO] - Step 6431, Loss: {'train/loss': 0.666007936000824, 'train/grad_norm_pre_clip': 15.210205078125, 'timing/data': 0.005155318998731673, 'timing/model': 1.202189142932184, 'learning_rate/qwen_vl_interface': 6.585109765092772e-07, 'learning_rate/action_model': 6.585109765092772e-06, 'epoch': 1.89}) +[2026-06-17 18:13:01,897][starVLA.training.train_starvla][INFO] - Step 6432, Loss: {'train/loss': 0.7780029773712158, 'train/grad_norm_pre_clip': 15.301373481750488, 'timing/data': 0.004954455071128905, 'timing/model': 1.2007985090604052, 'learning_rate/qwen_vl_interface': 6.579574157734575e-07, 'learning_rate/action_model': 6.579574157734575e-06, 'epoch': 1.89}) +[2026-06-17 18:13:06,655][starVLA.training.train_starvla][INFO] - Step 6433, Loss: {'train/loss': 0.7173054218292236, 'train/grad_norm_pre_clip': 15.243216514587402, 'timing/data': 0.005115782958455384, 'timing/model': 1.2009747099364176, 'learning_rate/qwen_vl_interface': 6.574048069730062e-07, 'learning_rate/action_model': 6.574048069730063e-06, 'epoch': 1.89}) +[2026-06-17 18:13:11,393][starVLA.training.train_starvla][INFO] - Step 6434, Loss: {'train/loss': 0.3437409996986389, 'train/grad_norm_pre_clip': 8.931035041809082, 'timing/data': 0.005114540923386812, 'timing/model': 1.204870712943375, 'learning_rate/qwen_vl_interface': 6.568531502224808e-07, 'learning_rate/action_model': 6.568531502224808e-06, 'epoch': 1.89}) +[2026-06-17 18:13:16,179][starVLA.training.train_starvla][INFO] - Step 6435, Loss: {'train/loss': 0.6300229430198669, 'train/grad_norm_pre_clip': 14.049432754516602, 'timing/data': 0.004745053010992706, 'timing/model': 1.3193479049950838, 'learning_rate/qwen_vl_interface': 6.563024456362395e-07, 'learning_rate/action_model': 6.563024456362395e-06, 'epoch': 1.89}) +[2026-06-17 18:13:20,839][starVLA.training.train_starvla][INFO] - Step 6436, Loss: {'train/loss': 0.7248196601867676, 'train/grad_norm_pre_clip': 14.047782897949219, 'timing/data': 0.005092567997053266, 'timing/model': 1.2045430958969519, 'learning_rate/qwen_vl_interface': 6.557526933284443e-07, 'learning_rate/action_model': 6.557526933284443e-06, 'epoch': 1.89}) +[2026-06-17 18:13:25,637][starVLA.training.train_starvla][INFO] - Step 6437, Loss: {'train/loss': 0.3859962522983551, 'train/grad_norm_pre_clip': 10.74056625366211, 'timing/data': 0.0056801720056682825, 'timing/model': 1.2046098499558866, 'learning_rate/qwen_vl_interface': 6.552038934130595e-07, 'learning_rate/action_model': 6.552038934130594e-06, 'epoch': 1.89}) +[2026-06-17 18:13:30,516][starVLA.training.train_starvla][INFO] - Step 6438, Loss: {'train/loss': 0.6543461680412292, 'train/grad_norm_pre_clip': 13.800322532653809, 'timing/data': 0.004599616979248822, 'timing/model': 1.203855923959054, 'learning_rate/qwen_vl_interface': 6.546560460038514e-07, 'learning_rate/action_model': 6.546560460038514e-06, 'epoch': 1.89}) +[2026-06-17 18:13:35,299][starVLA.training.train_starvla][INFO] - Step 6439, Loss: {'train/loss': 0.8650779128074646, 'train/grad_norm_pre_clip': 17.986108779907227, 'timing/data': 0.005030222004279494, 'timing/model': 1.202909818966873, 'learning_rate/qwen_vl_interface': 6.5410915121439e-07, 'learning_rate/action_model': 6.5410915121439e-06, 'epoch': 1.89}) +[2026-06-17 18:13:40,080][starVLA.training.train_starvla][INFO] - Step 6440, Loss: {'train/loss': 0.5653558969497681, 'train/grad_norm_pre_clip': 13.48111343383789, 'timing/data': 0.005223050015047193, 'timing/model': 1.3326196949929, 'learning_rate/qwen_vl_interface': 6.535632091580474e-07, 'learning_rate/action_model': 6.535632091580474e-06, 'epoch': 1.89}) +[2026-06-17 18:13:44,730][starVLA.training.train_starvla][INFO] - Step 6441, Loss: {'train/loss': 0.7062383890151978, 'train/grad_norm_pre_clip': 18.912580490112305, 'timing/data': 0.005103756091557443, 'timing/model': 1.204055443056859, 'learning_rate/qwen_vl_interface': 6.530182199479965e-07, 'learning_rate/action_model': 6.5301821994799645e-06, 'epoch': 1.9}) +[2026-06-17 18:13:49,507][starVLA.training.train_starvla][INFO] - Step 6442, Loss: {'train/loss': 0.5391833782196045, 'train/grad_norm_pre_clip': 13.24170970916748, 'timing/data': 0.005143898073583841, 'timing/model': 1.2025666269473732, 'learning_rate/qwen_vl_interface': 6.524741836972162e-07, 'learning_rate/action_model': 6.524741836972162e-06, 'epoch': 1.9}) +[2026-06-17 18:13:54,213][starVLA.training.train_starvla][INFO] - Step 6443, Loss: {'train/loss': 0.4191640019416809, 'train/grad_norm_pre_clip': 11.488569259643555, 'timing/data': 0.0049652960151433945, 'timing/model': 1.2032378891017288, 'learning_rate/qwen_vl_interface': 6.519311005184848e-07, 'learning_rate/action_model': 6.5193110051848476e-06, 'epoch': 1.9}) +[2026-06-17 18:13:58,953][starVLA.training.train_starvla][INFO] - Step 6444, Loss: {'train/loss': 0.849130392074585, 'train/grad_norm_pre_clip': 17.555252075195312, 'timing/data': 0.004718256066553295, 'timing/model': 1.2031671029981226, 'learning_rate/qwen_vl_interface': 6.51388970524383e-07, 'learning_rate/action_model': 6.513889705243829e-06, 'epoch': 1.9}) +[2026-06-17 18:14:03,737][starVLA.training.train_starvla][INFO] - Step 6445, Loss: {'train/loss': 0.6623634696006775, 'train/grad_norm_pre_clip': 15.431434631347656, 'timing/data': 0.005135723971761763, 'timing/model': 1.204295007046312, 'learning_rate/qwen_vl_interface': 6.508477938272963e-07, 'learning_rate/action_model': 6.508477938272962e-06, 'epoch': 1.9}) +[2026-06-17 18:14:08,520][starVLA.training.train_starvla][INFO] - Step 6446, Loss: {'train/loss': 0.35141685605049133, 'train/grad_norm_pre_clip': 10.03291130065918, 'timing/data': 0.0049400778952986, 'timing/model': 1.3332617909181863, 'learning_rate/qwen_vl_interface': 6.503075705394113e-07, 'learning_rate/action_model': 6.5030757053941125e-06, 'epoch': 1.9}) +[2026-06-17 18:14:13,170][starVLA.training.train_starvla][INFO] - Step 6447, Loss: {'train/loss': 0.46418797969818115, 'train/grad_norm_pre_clip': 10.219585418701172, 'timing/data': 0.005141237983480096, 'timing/model': 1.2048725920030847, 'learning_rate/qwen_vl_interface': 6.497683007727166e-07, 'learning_rate/action_model': 6.497683007727166e-06, 'epoch': 1.9}) +[2026-06-17 18:14:17,950][starVLA.training.train_starvla][INFO] - Step 6448, Loss: {'train/loss': 0.7822682857513428, 'train/grad_norm_pre_clip': 18.2391414642334, 'timing/data': 0.004833574057556689, 'timing/model': 1.2044080189662054, 'learning_rate/qwen_vl_interface': 6.492299846390024e-07, 'learning_rate/action_model': 6.492299846390023e-06, 'epoch': 1.9}) +[2026-06-17 18:14:22,732][starVLA.training.train_starvla][INFO] - Step 6449, Loss: {'train/loss': 0.8679277896881104, 'train/grad_norm_pre_clip': 16.575124740600586, 'timing/data': 0.004766634083352983, 'timing/model': 1.2029473870061338, 'learning_rate/qwen_vl_interface': 6.486926222498633e-07, 'learning_rate/action_model': 6.486926222498633e-06, 'epoch': 1.9}) +[2026-06-17 18:14:27,473][starVLA.training.train_starvla][INFO] - Step 6450, Loss: {'train/loss': 0.6569566130638123, 'train/grad_norm_pre_clip': 16.98367691040039, 'timing/data': 0.005188688053749502, 'timing/model': 1.2051385379163548, 'learning_rate/qwen_vl_interface': 6.481562137166952e-07, 'learning_rate/action_model': 6.481562137166952e-06, 'epoch': 1.9}) +[2026-06-17 18:14:32,256][starVLA.training.train_starvla][INFO] - Step 6451, Loss: {'train/loss': 0.6486937999725342, 'train/grad_norm_pre_clip': 14.053483009338379, 'timing/data': 0.004906467976979911, 'timing/model': 1.205489882035181, 'learning_rate/qwen_vl_interface': 6.476207591506954e-07, 'learning_rate/action_model': 6.476207591506954e-06, 'epoch': 1.9}) +[2026-06-17 18:14:37,052][starVLA.training.train_starvla][INFO] - Step 6452, Loss: {'train/loss': 0.4806176722049713, 'train/grad_norm_pre_clip': 11.560525894165039, 'timing/data': 0.005074369953945279, 'timing/model': 1.3361053729895502, 'learning_rate/qwen_vl_interface': 6.470862586628645e-07, 'learning_rate/action_model': 6.470862586628646e-06, 'epoch': 1.9}) +[2026-06-17 18:14:41,712][starVLA.training.train_starvla][INFO] - Step 6453, Loss: {'train/loss': 0.74595707654953, 'train/grad_norm_pre_clip': 16.198719024658203, 'timing/data': 0.004996518022380769, 'timing/model': 1.2050094050355256, 'learning_rate/qwen_vl_interface': 6.465527123640056e-07, 'learning_rate/action_model': 6.465527123640056e-06, 'epoch': 1.9}) +[2026-06-17 18:14:46,498][starVLA.training.train_starvla][INFO] - Step 6454, Loss: {'train/loss': 0.9665361642837524, 'train/grad_norm_pre_clip': 21.70511817932129, 'timing/data': 0.004992814036086202, 'timing/model': 1.2048982450505719, 'learning_rate/qwen_vl_interface': 6.460201203647229e-07, 'learning_rate/action_model': 6.460201203647229e-06, 'epoch': 1.9}) +[2026-06-17 18:14:51,338][starVLA.training.train_starvla][INFO] - Step 6455, Loss: {'train/loss': 0.7950472235679626, 'train/grad_norm_pre_clip': 16.493127822875977, 'timing/data': 0.004716134048067033, 'timing/model': 1.2132491909433156, 'learning_rate/qwen_vl_interface': 6.454884827754224e-07, 'learning_rate/action_model': 6.4548848277542235e-06, 'epoch': 1.9}) +[2026-06-17 18:14:56,126][starVLA.training.train_starvla][INFO] - Step 6456, Loss: {'train/loss': 0.6025750637054443, 'train/grad_norm_pre_clip': 15.221829414367676, 'timing/data': 0.005209675990045071, 'timing/model': 1.20818093593698, 'learning_rate/qwen_vl_interface': 6.449577997063156e-07, 'learning_rate/action_model': 6.449577997063155e-06, 'epoch': 1.9}) +[2026-06-17 18:15:00,805][starVLA.training.train_starvla][INFO] - Step 6457, Loss: {'train/loss': 1.2647782564163208, 'train/grad_norm_pre_clip': 18.63871192932129, 'timing/data': 0.004940166021697223, 'timing/model': 1.2063560980604962, 'learning_rate/qwen_vl_interface': 6.444280712674112e-07, 'learning_rate/action_model': 6.444280712674111e-06, 'epoch': 1.9}) +[2026-06-17 18:15:05,511][starVLA.training.train_starvla][INFO] - Step 6458, Loss: {'train/loss': 0.7101500034332275, 'train/grad_norm_pre_clip': 15.478850364685059, 'timing/data': 0.0047692639054730535, 'timing/model': 1.2419941650005057, 'learning_rate/qwen_vl_interface': 6.438992975685237e-07, 'learning_rate/action_model': 6.438992975685237e-06, 'epoch': 1.9}) +[2026-06-17 18:15:10,171][starVLA.training.train_starvla][INFO] - Step 6459, Loss: {'train/loss': 0.7667079567909241, 'train/grad_norm_pre_clip': 15.211479187011719, 'timing/data': 0.005012891022488475, 'timing/model': 1.2052912540966645, 'learning_rate/qwen_vl_interface': 6.433714787192677e-07, 'learning_rate/action_model': 6.4337147871926766e-06, 'epoch': 1.9}) +[2026-06-17 18:15:14,906][starVLA.training.train_starvla][INFO] - Step 6460, Loss: {'train/loss': 0.505007803440094, 'train/grad_norm_pre_clip': 13.338666915893555, 'timing/data': 0.004519046982750297, 'timing/model': 1.2058113550301641, 'learning_rate/qwen_vl_interface': 6.428446148290615e-07, 'learning_rate/action_model': 6.428446148290615e-06, 'epoch': 1.9}) +[2026-06-17 18:15:19,703][starVLA.training.train_starvla][INFO] - Step 6461, Loss: {'train/loss': 0.41416168212890625, 'train/grad_norm_pre_clip': 10.96200180053711, 'timing/data': 0.004596315091475844, 'timing/model': 1.2045133649371564, 'learning_rate/qwen_vl_interface': 6.423187060071248e-07, 'learning_rate/action_model': 6.423187060071248e-06, 'epoch': 1.9}) +[2026-06-17 18:15:24,507][starVLA.training.train_starvla][INFO] - Step 6462, Loss: {'train/loss': 0.6909785270690918, 'train/grad_norm_pre_clip': 14.846989631652832, 'timing/data': 0.005070299026556313, 'timing/model': 1.2059942029882222, 'learning_rate/qwen_vl_interface': 6.417937523624778e-07, 'learning_rate/action_model': 6.417937523624778e-06, 'epoch': 1.9}) +[2026-06-17 18:15:29,299][starVLA.training.train_starvla][INFO] - Step 6463, Loss: {'train/loss': 1.0133306980133057, 'train/grad_norm_pre_clip': 19.65907859802246, 'timing/data': 0.004820290021598339, 'timing/model': 1.330202266923152, 'learning_rate/qwen_vl_interface': 6.412697540039442e-07, 'learning_rate/action_model': 6.412697540039442e-06, 'epoch': 1.9}) +[2026-06-17 18:15:34,010][starVLA.training.train_starvla][INFO] - Step 6464, Loss: {'train/loss': 0.5792332291603088, 'train/grad_norm_pre_clip': 11.976799011230469, 'timing/data': 0.005041016032919288, 'timing/model': 1.2529241479933262, 'learning_rate/qwen_vl_interface': 6.407467110401505e-07, 'learning_rate/action_model': 6.407467110401505e-06, 'epoch': 1.9}) +[2026-06-17 18:15:38,718][starVLA.training.train_starvla][INFO] - Step 6465, Loss: {'train/loss': 0.40901076793670654, 'train/grad_norm_pre_clip': 11.978598594665527, 'timing/data': 0.004815252963453531, 'timing/model': 1.2019969039829448, 'learning_rate/qwen_vl_interface': 6.402246235795228e-07, 'learning_rate/action_model': 6.402246235795227e-06, 'epoch': 1.9}) +[2026-06-17 18:15:43,451][starVLA.training.train_starvla][INFO] - Step 6466, Loss: {'train/loss': 0.6232937574386597, 'train/grad_norm_pre_clip': 17.58451271057129, 'timing/data': 0.004869946977123618, 'timing/model': 1.2008023359812796, 'learning_rate/qwen_vl_interface': 6.397034917302913e-07, 'learning_rate/action_model': 6.397034917302913e-06, 'epoch': 1.9}) +[2026-06-17 18:15:48,228][starVLA.training.train_starvla][INFO] - Step 6467, Loss: {'train/loss': 0.6039057374000549, 'train/grad_norm_pre_clip': 11.17740535736084, 'timing/data': 0.0046065859496593475, 'timing/model': 1.203952502924949, 'learning_rate/qwen_vl_interface': 6.391833156004871e-07, 'learning_rate/action_model': 6.39183315600487e-06, 'epoch': 1.9}) +[2026-06-17 18:15:53,016][starVLA.training.train_starvla][INFO] - Step 6468, Loss: {'train/loss': 0.5504179000854492, 'train/grad_norm_pre_clip': 12.13619613647461, 'timing/data': 0.005261384998448193, 'timing/model': 1.2053218180080876, 'learning_rate/qwen_vl_interface': 6.386640952979427e-07, 'learning_rate/action_model': 6.386640952979426e-06, 'epoch': 1.9}) +[2026-06-17 18:15:57,799][starVLA.training.train_starvla][INFO] - Step 6469, Loss: {'train/loss': 0.556691586971283, 'train/grad_norm_pre_clip': 12.767681121826172, 'timing/data': 0.004811691935174167, 'timing/model': 1.330005310010165, 'learning_rate/qwen_vl_interface': 6.381458309302937e-07, 'learning_rate/action_model': 6.381458309302937e-06, 'epoch': 1.9}) +[2026-06-17 18:16:02,571][starVLA.training.train_starvla][INFO] - Step 6470, Loss: {'train/loss': 0.5062345266342163, 'train/grad_norm_pre_clip': 10.463464736938477, 'timing/data': 0.004873657948337495, 'timing/model': 1.3253498539561406, 'learning_rate/qwen_vl_interface': 6.376285226049762e-07, 'learning_rate/action_model': 6.376285226049762e-06, 'epoch': 1.9}) +[2026-06-17 18:16:07,249][starVLA.training.train_starvla][INFO] - Step 6471, Loss: {'train/loss': 0.5291140079498291, 'train/grad_norm_pre_clip': 12.417622566223145, 'timing/data': 0.0051090719643980265, 'timing/model': 1.2145370919024572, 'learning_rate/qwen_vl_interface': 6.371121704292292e-07, 'learning_rate/action_model': 6.371121704292292e-06, 'epoch': 1.9}) +[2026-06-17 18:16:12,006][starVLA.training.train_starvla][INFO] - Step 6472, Loss: {'train/loss': 0.6598291993141174, 'train/grad_norm_pre_clip': 17.923999786376953, 'timing/data': 0.005060096038505435, 'timing/model': 1.2040058400016278, 'learning_rate/qwen_vl_interface': 6.365967745100926e-07, 'learning_rate/action_model': 6.3659677451009256e-06, 'epoch': 1.9}) +[2026-06-17 18:16:16,810][starVLA.training.train_starvla][INFO] - Step 6473, Loss: {'train/loss': 0.4137546122074127, 'train/grad_norm_pre_clip': 11.901597023010254, 'timing/data': 0.004659609054215252, 'timing/model': 1.2079709690297022, 'learning_rate/qwen_vl_interface': 6.360823349544092e-07, 'learning_rate/action_model': 6.360823349544092e-06, 'epoch': 1.9}) +[2026-06-17 18:16:21,621][starVLA.training.train_starvla][INFO] - Step 6474, Loss: {'train/loss': 0.7603527307510376, 'train/grad_norm_pre_clip': 64.0714111328125, 'timing/data': 0.005075067048892379, 'timing/model': 1.213987949071452, 'learning_rate/qwen_vl_interface': 6.355688518688222e-07, 'learning_rate/action_model': 6.355688518688221e-06, 'epoch': 1.9}) +[2026-06-17 18:16:26,377][starVLA.training.train_starvla][INFO] - Step 6475, Loss: {'train/loss': 0.7627125382423401, 'train/grad_norm_pre_clip': 14.937644004821777, 'timing/data': 0.0049602419603616, 'timing/model': 1.2813094559824094, 'learning_rate/qwen_vl_interface': 6.350563253597781e-07, 'learning_rate/action_model': 6.350563253597781e-06, 'epoch': 1.91}) +[2026-06-17 18:16:31,055][starVLA.training.train_starvla][INFO] - Step 6476, Loss: {'train/loss': 0.36727532744407654, 'train/grad_norm_pre_clip': 11.768818855285645, 'timing/data': 0.005150604993104935, 'timing/model': 1.2119292670395225, 'learning_rate/qwen_vl_interface': 6.345447555335236e-07, 'learning_rate/action_model': 6.3454475553352345e-06, 'epoch': 1.91}) +[2026-06-17 18:16:35,824][starVLA.training.train_starvla][INFO] - Step 6477, Loss: {'train/loss': 0.4933755099773407, 'train/grad_norm_pre_clip': 12.106517791748047, 'timing/data': 0.005144826020114124, 'timing/model': 1.205112207098864, 'learning_rate/qwen_vl_interface': 6.340341424961072e-07, 'learning_rate/action_model': 6.340341424961071e-06, 'epoch': 1.91}) +[2026-06-17 18:16:40,621][starVLA.training.train_starvla][INFO] - Step 6478, Loss: {'train/loss': 0.7100706696510315, 'train/grad_norm_pre_clip': 13.749388694763184, 'timing/data': 0.004854941973462701, 'timing/model': 1.204881053068675, 'learning_rate/qwen_vl_interface': 6.335244863533802e-07, 'learning_rate/action_model': 6.3352448635338016e-06, 'epoch': 1.91}) +[2026-06-17 18:16:45,417][starVLA.training.train_starvla][INFO] - Step 6479, Loss: {'train/loss': 0.40872326493263245, 'train/grad_norm_pre_clip': 12.219074249267578, 'timing/data': 0.005023260950110853, 'timing/model': 1.2083991719409823, 'learning_rate/qwen_vl_interface': 6.330157872109953e-07, 'learning_rate/action_model': 6.330157872109953e-06, 'epoch': 1.91}) +[2026-06-17 18:16:50,212][starVLA.training.train_starvla][INFO] - Step 6480, Loss: {'train/loss': 0.6405249834060669, 'train/grad_norm_pre_clip': 14.787610054016113, 'timing/data': 0.005078681977465749, 'timing/model': 1.3305468159960583, 'learning_rate/qwen_vl_interface': 6.325080451744053e-07, 'learning_rate/action_model': 6.325080451744053e-06, 'epoch': 1.91}) +[2026-06-17 18:16:54,976][starVLA.training.train_starvla][INFO] - Step 6481, Loss: {'train/loss': 0.7364868521690369, 'train/grad_norm_pre_clip': 16.131664276123047, 'timing/data': 0.0049969840329140425, 'timing/model': 1.2898933530086651, 'learning_rate/qwen_vl_interface': 6.320012603488663e-07, 'learning_rate/action_model': 6.320012603488662e-06, 'epoch': 1.91}) +[2026-06-17 18:16:59,630][starVLA.training.train_starvla][INFO] - Step 6482, Loss: {'train/loss': 0.7198185920715332, 'train/grad_norm_pre_clip': 13.945941925048828, 'timing/data': 0.004902175976894796, 'timing/model': 1.2029440939659253, 'learning_rate/qwen_vl_interface': 6.314954328394351e-07, 'learning_rate/action_model': 6.3149543283943515e-06, 'epoch': 1.91}) +[2026-06-17 18:17:04,429][starVLA.training.train_starvla][INFO] - Step 6483, Loss: {'train/loss': 0.4676743447780609, 'train/grad_norm_pre_clip': 21.822208404541016, 'timing/data': 0.005310687003657222, 'timing/model': 1.2038667839951813, 'learning_rate/qwen_vl_interface': 6.309905627509707e-07, 'learning_rate/action_model': 6.309905627509706e-06, 'epoch': 1.91}) +[2026-06-17 18:17:09,236][starVLA.training.train_starvla][INFO] - Step 6484, Loss: {'train/loss': 0.7252559661865234, 'train/grad_norm_pre_clip': 49.91985321044922, 'timing/data': 0.004931558039970696, 'timing/model': 1.213205625070259, 'learning_rate/qwen_vl_interface': 6.304866501881327e-07, 'learning_rate/action_model': 6.304866501881327e-06, 'epoch': 1.91}) +[2026-06-17 18:17:14,054][starVLA.training.train_starvla][INFO] - Step 6485, Loss: {'train/loss': 0.7957451343536377, 'train/grad_norm_pre_clip': 16.936019897460938, 'timing/data': 0.004791316925548017, 'timing/model': 1.2248811901081353, 'learning_rate/qwen_vl_interface': 6.299836952553831e-07, 'learning_rate/action_model': 6.299836952553831e-06, 'epoch': 1.91}) +[2026-06-17 18:17:18,851][starVLA.training.train_starvla][INFO] - Step 6486, Loss: {'train/loss': 0.6488573551177979, 'train/grad_norm_pre_clip': 12.923351287841797, 'timing/data': 0.004832681966945529, 'timing/model': 1.2114788880571723, 'learning_rate/qwen_vl_interface': 6.294816980569845e-07, 'learning_rate/action_model': 6.294816980569845e-06, 'epoch': 1.91}) +[2026-06-17 18:17:23,806][starVLA.training.train_starvla][INFO] - Step 6487, Loss: {'train/loss': 0.7147995829582214, 'train/grad_norm_pre_clip': 15.71257209777832, 'timing/data': 0.004973402014002204, 'timing/model': 1.3035255080321804, 'learning_rate/qwen_vl_interface': 6.289806586970021e-07, 'learning_rate/action_model': 6.289806586970021e-06, 'epoch': 1.91}) +[2026-06-17 18:17:28,482][starVLA.training.train_starvla][INFO] - Step 6488, Loss: {'train/loss': 0.6182805299758911, 'train/grad_norm_pre_clip': 17.95276641845703, 'timing/data': 0.004919483093544841, 'timing/model': 1.2071194569580257, 'learning_rate/qwen_vl_interface': 6.284805772793011e-07, 'learning_rate/action_model': 6.284805772793011e-06, 'epoch': 1.91}) +[2026-06-17 18:17:33,274][starVLA.training.train_starvla][INFO] - Step 6489, Loss: {'train/loss': 0.6228700876235962, 'train/grad_norm_pre_clip': 16.103527069091797, 'timing/data': 0.005132378078997135, 'timing/model': 1.2065445180051029, 'learning_rate/qwen_vl_interface': 6.279814539075499e-07, 'learning_rate/action_model': 6.2798145390754985e-06, 'epoch': 1.91}) +[2026-06-17 18:17:38,071][starVLA.training.train_starvla][INFO] - Step 6490, Loss: {'train/loss': 0.7673648595809937, 'train/grad_norm_pre_clip': 16.131486892700195, 'timing/data': 0.004945421009324491, 'timing/model': 1.2040329270530492, 'learning_rate/qwen_vl_interface': 6.274832886852172e-07, 'learning_rate/action_model': 6.2748328868521724e-06, 'epoch': 1.91}) +[2026-06-17 18:17:42,860][starVLA.training.train_starvla][INFO] - Step 6491, Loss: {'train/loss': 0.6121594905853271, 'train/grad_norm_pre_clip': 15.800394058227539, 'timing/data': 0.005149961914867163, 'timing/model': 1.2046903499867767, 'learning_rate/qwen_vl_interface': 6.269860817155721e-07, 'learning_rate/action_model': 6.269860817155721e-06, 'epoch': 1.91}) +[2026-06-17 18:17:47,641][starVLA.training.train_starvla][INFO] - Step 6492, Loss: {'train/loss': 0.49216893315315247, 'train/grad_norm_pre_clip': 10.96166706085205, 'timing/data': 0.005012125009670854, 'timing/model': 1.329558106022887, 'learning_rate/qwen_vl_interface': 6.264898331016865e-07, 'learning_rate/action_model': 6.264898331016865e-06, 'epoch': 1.91}) +[2026-06-17 18:17:52,296][starVLA.training.train_starvla][INFO] - Step 6493, Loss: {'train/loss': 0.5522032976150513, 'train/grad_norm_pre_clip': 14.947906494140625, 'timing/data': 0.004926903056912124, 'timing/model': 1.2051786630181596, 'learning_rate/qwen_vl_interface': 6.259945429464343e-07, 'learning_rate/action_model': 6.259945429464342e-06, 'epoch': 1.91}) +[2026-06-17 18:17:57,127][starVLA.training.train_starvla][INFO] - Step 6494, Loss: {'train/loss': 0.5811864137649536, 'train/grad_norm_pre_clip': 14.667667388916016, 'timing/data': 0.005154325976036489, 'timing/model': 1.2087064469233155, 'learning_rate/qwen_vl_interface': 6.255002113524885e-07, 'learning_rate/action_model': 6.255002113524886e-06, 'epoch': 1.91}) +[2026-06-17 18:18:01,916][starVLA.training.train_starvla][INFO] - Step 6495, Loss: {'train/loss': 1.0424989461898804, 'train/grad_norm_pre_clip': 16.69666862487793, 'timing/data': 0.004760114010423422, 'timing/model': 1.2089585410431027, 'learning_rate/qwen_vl_interface': 6.250068384223255e-07, 'learning_rate/action_model': 6.250068384223255e-06, 'epoch': 1.91}) +[2026-06-17 18:18:06,633][starVLA.training.train_starvla][INFO] - Step 6496, Loss: {'train/loss': 0.38956770300865173, 'train/grad_norm_pre_clip': 10.06344223022461, 'timing/data': 0.004920506034977734, 'timing/model': 1.2083581339102238, 'learning_rate/qwen_vl_interface': 6.245144242582214e-07, 'learning_rate/action_model': 6.245144242582214e-06, 'epoch': 1.91}) +[2026-06-17 18:18:11,389][starVLA.training.train_starvla][INFO] - Step 6497, Loss: {'train/loss': 0.7476332783699036, 'train/grad_norm_pre_clip': 17.125083923339844, 'timing/data': 0.005058323964476585, 'timing/model': 1.2055048570036888, 'learning_rate/qwen_vl_interface': 6.240229689622546e-07, 'learning_rate/action_model': 6.240229689622545e-06, 'epoch': 1.91}) +[2026-06-17 18:18:16,190][starVLA.training.train_starvla][INFO] - Step 6498, Loss: {'train/loss': 0.4024999141693115, 'train/grad_norm_pre_clip': 12.57007884979248, 'timing/data': 0.004897723090834916, 'timing/model': 1.343612955068238, 'learning_rate/qwen_vl_interface': 6.235324726363042e-07, 'learning_rate/action_model': 6.235324726363041e-06, 'epoch': 1.91}) +[2026-06-17 18:18:20,843][starVLA.training.train_starvla][INFO] - Step 6499, Loss: {'train/loss': 0.4557293951511383, 'train/grad_norm_pre_clip': 12.989960670471191, 'timing/data': 0.004956631106324494, 'timing/model': 1.205432366929017, 'learning_rate/qwen_vl_interface': 6.230429353820503e-07, 'learning_rate/action_model': 6.230429353820503e-06, 'epoch': 1.91}) +[2026-06-17 18:28:30,846][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 6500 +[2026-06-17 18:28:30,846][starVLA.training.train_starvla][INFO] - Step 6500, Loss: {'train/loss': 0.3957905173301697, 'train/grad_norm_pre_clip': 12.626790046691895, 'eval/loss': 0.7036328911781311, 'eval/latency_0/loss': 0.7036328911781311, 'eval/demon_fire_precision': 0.9053014973694861, 'eval/demon_fire_recall': 0.9289867109634552, 'eval/demon_fire_f1': 0.9169911867185899, 'eval/demon_fire_rate_model': 0.6871523915461624, 'eval/demon_fire_rate_teacher': 0.6696329254727474, 'eval/demon_left_precision': 0.8351841028638223, 'eval/demon_left_recall': 0.8096317280453258, 'eval/demon_left_f1': 0.8222094361334867, 'eval/demon_left_rate_model': 0.47580645161290325, 'eval/demon_left_rate_teacher': 0.49082313681868744, 'eval/demon_right_precision': 0.8173025255239119, 'eval/demon_right_recall': 0.8593220338983051, 'eval/demon_right_f1': 0.8377857339575875, 'eval/demon_right_rate_model': 0.5175194660734149, 'eval/demon_right_rate_teacher': 0.4922135706340378, 'eval/demon_move_f1': 0.8299975850455371, 'eval/demon_cc_f1': 0.8890367059594504, 'eval/latency_0/demon_fire_precision': 0.9053014973694861, 'eval/latency_0/demon_fire_recall': 0.9289867109634552, 'eval/latency_0/demon_fire_f1': 0.9169911867185899, 'eval/latency_0/demon_fire_rate_model': 0.6871523915461624, 'eval/latency_0/demon_fire_rate_teacher': 0.6696329254727474, 'eval/latency_0/demon_left_precision': 0.8351841028638223, 'eval/latency_0/demon_left_recall': 0.8096317280453258, 'eval/latency_0/demon_left_f1': 0.8222094361334867, 'eval/latency_0/demon_left_rate_model': 0.47580645161290325, 'eval/latency_0/demon_left_rate_teacher': 0.49082313681868744, 'eval/latency_0/demon_right_precision': 0.8173025255239119, 'eval/latency_0/demon_right_recall': 0.8593220338983051, 'eval/latency_0/demon_right_f1': 0.8377857339575875, 'eval/latency_0/demon_right_rate_model': 0.5175194660734149, 'eval/latency_0/demon_right_rate_teacher': 0.4922135706340378, 'eval/latency_0/demon_move_f1': 0.8299975850455371, 'eval/latency_0/demon_cc_f1': 0.8890367059594504, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 8184.0, 'rl_games_eval/mid_train/mean_length': 3600.0, 'rl_games_eval/mid_train/std_reward': 1062.9270906322786, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/macro_mean_reward': 8184.0, 'rl_games_eval/mid_train/macro_mean_length': 3600.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/demon_attack__latency_0/mean_reward': 8184.0, 'rl_games_eval/mid_train/demon_attack__latency_0/mean_length': 3600.0, 'rl_games_eval/mid_train/demon_attack__latency_0/std_reward': 1062.9270906322786, 'rl_games_eval/mid_train/demon_attack__latency_0/std_length': 0.0, 'rl_games_eval/mid_train/demon_attack__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 8184.0, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.004859141074120998, 'timing/model': 1.2047060070326552, 'learning_rate/qwen_vl_interface': 6.225543573009747e-07, 'learning_rate/action_model': 6.225543573009748e-06, 'epoch': 1.91}) +[2026-06-17 18:28:30,846][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_6500_state +[2026-06-17 18:28:30,849][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 18:28:34,421][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_6500_state/model.safetensors +[2026-06-17 18:28:42,079][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_6500_state/optimizer.bin +[2026-06-17 18:28:42,080][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_6500_state/sampler.bin +[2026-06-17 18:28:42,080][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_6500_state/sampler_1.bin +[2026-06-17 18:28:42,080][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_6500_state/random_states_0.pkl +[2026-06-17 18:28:42,136][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 6500 +[2026-06-17 18:28:42,136][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 18:28:42,154][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 18:28:46,800][starVLA.training.train_starvla][INFO] - Step 6501, Loss: {'train/loss': 0.9303160905838013, 'train/grad_norm_pre_clip': 15.000837326049805, 'timing/data': 0.004801061004400253, 'timing/model': 1.1832665300462395, 'learning_rate/qwen_vl_interface': 6.2206673849436e-07, 'learning_rate/action_model': 6.2206673849436e-06, 'epoch': 1.91}) +[2026-06-17 18:28:51,475][starVLA.training.train_starvla][INFO] - Step 6502, Loss: {'train/loss': 0.5530992746353149, 'train/grad_norm_pre_clip': 11.592391967773438, 'timing/data': 0.004929424030706286, 'timing/model': 1.1833078659838066, 'learning_rate/qwen_vl_interface': 6.215800790632905e-07, 'learning_rate/action_model': 6.215800790632904e-06, 'epoch': 1.91}) +[2026-06-17 18:28:56,248][starVLA.training.train_starvla][INFO] - Step 6503, Loss: {'train/loss': 0.43463656306266785, 'train/grad_norm_pre_clip': 17.12010955810547, 'timing/data': 0.005187230068258941, 'timing/model': 1.1858702769968659, 'learning_rate/qwen_vl_interface': 6.210943791086506e-07, 'learning_rate/action_model': 6.210943791086506e-06, 'epoch': 1.91}) +[2026-06-17 18:29:00,965][starVLA.training.train_starvla][INFO] - Step 6504, Loss: {'train/loss': 0.6515584588050842, 'train/grad_norm_pre_clip': 15.179685592651367, 'timing/data': 0.005052088061347604, 'timing/model': 1.3171941619366407, 'learning_rate/qwen_vl_interface': 6.206096387311278e-07, 'learning_rate/action_model': 6.206096387311277e-06, 'epoch': 1.91}) +[2026-06-17 18:29:05,561][starVLA.training.train_starvla][INFO] - Step 6505, Loss: {'train/loss': 0.5097846388816833, 'train/grad_norm_pre_clip': 21.040037155151367, 'timing/data': 0.005000325036235154, 'timing/model': 1.1879685019375756, 'learning_rate/qwen_vl_interface': 6.201258580312078e-07, 'learning_rate/action_model': 6.201258580312077e-06, 'epoch': 1.91}) +[2026-06-17 18:29:10,284][starVLA.training.train_starvla][INFO] - Step 6506, Loss: {'train/loss': 0.6326898336410522, 'train/grad_norm_pre_clip': 14.902398109436035, 'timing/data': 0.005469659925438464, 'timing/model': 1.1899908779887483, 'learning_rate/qwen_vl_interface': 6.196430371091792e-07, 'learning_rate/action_model': 6.196430371091792e-06, 'epoch': 1.91}) +[2026-06-17 18:29:15,073][starVLA.training.train_starvla][INFO] - Step 6507, Loss: {'train/loss': 0.6581088304519653, 'train/grad_norm_pre_clip': 19.411758422851562, 'timing/data': 0.004714435082860291, 'timing/model': 1.1924834368983284, 'learning_rate/qwen_vl_interface': 6.191611760651312e-07, 'learning_rate/action_model': 6.191611760651312e-06, 'epoch': 1.91}) +[2026-06-17 18:29:19,782][starVLA.training.train_starvla][INFO] - Step 6508, Loss: {'train/loss': 0.9899486303329468, 'train/grad_norm_pre_clip': 17.97284507751465, 'timing/data': 0.004744471050798893, 'timing/model': 1.1933937480207533, 'learning_rate/qwen_vl_interface': 6.18680274998955e-07, 'learning_rate/action_model': 6.186802749989549e-06, 'epoch': 1.91}) +[2026-06-17 18:29:24,522][starVLA.training.train_starvla][INFO] - Step 6509, Loss: {'train/loss': 0.8067829608917236, 'train/grad_norm_pre_clip': 14.007394790649414, 'timing/data': 0.005044130026362836, 'timing/model': 1.1936033649835736, 'learning_rate/qwen_vl_interface': 6.182003340103413e-07, 'learning_rate/action_model': 6.1820033401034135e-06, 'epoch': 1.92}) +[2026-06-17 18:29:29,258][starVLA.training.train_starvla][INFO] - Step 6510, Loss: {'train/loss': 0.7224549651145935, 'train/grad_norm_pre_clip': 16.231578826904297, 'timing/data': 0.004600454936735332, 'timing/model': 1.32096639694646, 'learning_rate/qwen_vl_interface': 6.177213531987819e-07, 'learning_rate/action_model': 6.177213531987819e-06, 'epoch': 1.92}) +[2026-06-17 18:29:33,879][starVLA.training.train_starvla][INFO] - Step 6511, Loss: {'train/loss': 0.7215734124183655, 'train/grad_norm_pre_clip': 20.183263778686523, 'timing/data': 0.0050836720038205385, 'timing/model': 1.1950237540295348, 'learning_rate/qwen_vl_interface': 6.172433326635714e-07, 'learning_rate/action_model': 6.172433326635713e-06, 'epoch': 1.92}) +[2026-06-17 18:29:38,631][starVLA.training.train_starvla][INFO] - Step 6512, Loss: {'train/loss': 0.7415520548820496, 'train/grad_norm_pre_clip': 27.055011749267578, 'timing/data': 0.005619412986561656, 'timing/model': 1.196135681006126, 'learning_rate/qwen_vl_interface': 6.167662725038025e-07, 'learning_rate/action_model': 6.1676627250380255e-06, 'epoch': 1.92}) +[2026-06-17 18:29:43,340][starVLA.training.train_starvla][INFO] - Step 6513, Loss: {'train/loss': 0.5298163890838623, 'train/grad_norm_pre_clip': 14.147479057312012, 'timing/data': 0.00514210807159543, 'timing/model': 1.1962727040518075, 'learning_rate/qwen_vl_interface': 6.16290172818371e-07, 'learning_rate/action_model': 6.16290172818371e-06, 'epoch': 1.92}) +[2026-06-17 18:29:48,112][starVLA.training.train_starvla][INFO] - Step 6514, Loss: {'train/loss': 1.0299073457717896, 'train/grad_norm_pre_clip': 17.60664939880371, 'timing/data': 0.005107981036417186, 'timing/model': 1.2097083519911394, 'learning_rate/qwen_vl_interface': 6.158150337059725e-07, 'learning_rate/action_model': 6.158150337059725e-06, 'epoch': 1.92}) +[2026-06-17 18:29:52,885][starVLA.training.train_starvla][INFO] - Step 6515, Loss: {'train/loss': 0.7424643039703369, 'train/grad_norm_pre_clip': 12.765430450439453, 'timing/data': 0.005095183965750039, 'timing/model': 1.211063526920043, 'learning_rate/qwen_vl_interface': 6.153408552651045e-07, 'learning_rate/action_model': 6.153408552651045e-06, 'epoch': 1.92}) +[2026-06-17 18:29:57,644][starVLA.training.train_starvla][INFO] - Step 6516, Loss: {'train/loss': 0.6593593955039978, 'train/grad_norm_pre_clip': 21.832063674926758, 'timing/data': 0.00533146399538964, 'timing/model': 1.3270362389739603, 'learning_rate/qwen_vl_interface': 6.14867637594064e-07, 'learning_rate/action_model': 6.14867637594064e-06, 'epoch': 1.92}) +[2026-06-17 18:30:02,336][starVLA.training.train_starvla][INFO] - Step 6517, Loss: {'train/loss': 0.6273243427276611, 'train/grad_norm_pre_clip': 14.375249862670898, 'timing/data': 0.005211855983361602, 'timing/model': 1.2566437919158489, 'learning_rate/qwen_vl_interface': 6.143953807909497e-07, 'learning_rate/action_model': 6.143953807909496e-06, 'epoch': 1.92}) +[2026-06-17 18:30:06,968][starVLA.training.train_starvla][INFO] - Step 6518, Loss: {'train/loss': 0.4784366488456726, 'train/grad_norm_pre_clip': 14.980378150939941, 'timing/data': 0.005456459010019898, 'timing/model': 1.1998004160122946, 'learning_rate/qwen_vl_interface': 6.139240849536616e-07, 'learning_rate/action_model': 6.1392408495366156e-06, 'epoch': 1.92}) +[2026-06-17 18:30:11,684][starVLA.training.train_starvla][INFO] - Step 6519, Loss: {'train/loss': 0.4496175944805145, 'train/grad_norm_pre_clip': 10.471534729003906, 'timing/data': 0.005094780935905874, 'timing/model': 1.1984042070107535, 'learning_rate/qwen_vl_interface': 6.134537501798993e-07, 'learning_rate/action_model': 6.1345375017989925e-06, 'epoch': 1.92}) +[2026-06-17 18:30:16,467][starVLA.training.train_starvla][INFO] - Step 6520, Loss: {'train/loss': 0.40942347049713135, 'train/grad_norm_pre_clip': 13.502401351928711, 'timing/data': 0.004943727981299162, 'timing/model': 1.2034858179977164, 'learning_rate/qwen_vl_interface': 6.129843765671634e-07, 'learning_rate/action_model': 6.129843765671633e-06, 'epoch': 1.92}) +[2026-06-17 18:30:21,253][starVLA.training.train_starvla][INFO] - Step 6521, Loss: {'train/loss': 0.7394529581069946, 'train/grad_norm_pre_clip': 15.10295581817627, 'timing/data': 0.004958262899890542, 'timing/model': 1.201368154026568, 'learning_rate/qwen_vl_interface': 6.125159642127559e-07, 'learning_rate/action_model': 6.125159642127558e-06, 'epoch': 1.92}) +[2026-06-17 18:30:26,015][starVLA.training.train_starvla][INFO] - Step 6522, Loss: {'train/loss': 0.7670823335647583, 'train/grad_norm_pre_clip': 17.33987808227539, 'timing/data': 0.005060533992946148, 'timing/model': 1.3273003930225968, 'learning_rate/qwen_vl_interface': 6.120485132137793e-07, 'learning_rate/action_model': 6.120485132137794e-06, 'epoch': 1.92}) +[2026-06-17 18:30:30,785][starVLA.training.train_starvla][INFO] - Step 6523, Loss: {'train/loss': 0.5704532265663147, 'train/grad_norm_pre_clip': 12.188088417053223, 'timing/data': 0.005096976994536817, 'timing/model': 1.3322664420120418, 'learning_rate/qwen_vl_interface': 6.115820236671369e-07, 'learning_rate/action_model': 6.115820236671369e-06, 'epoch': 1.92}) +[2026-06-17 18:30:35,424][starVLA.training.train_starvla][INFO] - Step 6524, Loss: {'train/loss': 0.5567777156829834, 'train/grad_norm_pre_clip': 11.918781280517578, 'timing/data': 0.005394352017901838, 'timing/model': 1.1991724990075454, 'learning_rate/qwen_vl_interface': 6.111164956695316e-07, 'learning_rate/action_model': 6.111164956695316e-06, 'epoch': 1.92}) +[2026-06-17 18:30:40,156][starVLA.training.train_starvla][INFO] - Step 6525, Loss: {'train/loss': 0.6997517347335815, 'train/grad_norm_pre_clip': 11.55940055847168, 'timing/data': 0.005258960998617113, 'timing/model': 1.20061927405186, 'learning_rate/qwen_vl_interface': 6.106519293174684e-07, 'learning_rate/action_model': 6.106519293174684e-06, 'epoch': 1.92}) +[2026-06-17 18:30:44,926][starVLA.training.train_starvla][INFO] - Step 6526, Loss: {'train/loss': 0.6452199816703796, 'train/grad_norm_pre_clip': 13.96750545501709, 'timing/data': 0.005148489028215408, 'timing/model': 1.201619339059107, 'learning_rate/qwen_vl_interface': 6.101883247072525e-07, 'learning_rate/action_model': 6.1018832470725245e-06, 'epoch': 1.92}) +[2026-06-17 18:30:49,697][starVLA.training.train_starvla][INFO] - Step 6527, Loss: {'train/loss': 0.9154677391052246, 'train/grad_norm_pre_clip': 15.340388298034668, 'timing/data': 0.005080678965896368, 'timing/model': 1.1996157870162278, 'learning_rate/qwen_vl_interface': 6.097256819349899e-07, 'learning_rate/action_model': 6.0972568193498985e-06, 'epoch': 1.92}) +[2026-06-17 18:30:54,410][starVLA.training.train_starvla][INFO] - Step 6528, Loss: {'train/loss': 0.48670679330825806, 'train/grad_norm_pre_clip': 11.790206909179688, 'timing/data': 0.0052645569667220116, 'timing/model': 1.2697452859720215, 'learning_rate/qwen_vl_interface': 6.092640010965864e-07, 'learning_rate/action_model': 6.092640010965864e-06, 'epoch': 1.92}) +[2026-06-17 18:30:59,093][starVLA.training.train_starvla][INFO] - Step 6529, Loss: {'train/loss': 0.7914283871650696, 'train/grad_norm_pre_clip': 17.45795440673828, 'timing/data': 0.00520383205730468, 'timing/model': 1.2048710440285504, 'learning_rate/qwen_vl_interface': 6.088032822877488e-07, 'learning_rate/action_model': 6.088032822877488e-06, 'epoch': 1.92}) +[2026-06-17 18:31:03,845][starVLA.training.train_starvla][INFO] - Step 6530, Loss: {'train/loss': 0.8520141839981079, 'train/grad_norm_pre_clip': 18.566143035888672, 'timing/data': 0.0047350459499284625, 'timing/model': 1.2038759050192311, 'learning_rate/qwen_vl_interface': 6.083435256039854e-07, 'learning_rate/action_model': 6.083435256039854e-06, 'epoch': 1.92}) +[2026-06-17 18:31:08,615][starVLA.training.train_starvla][INFO] - Step 6531, Loss: {'train/loss': 0.4577408730983734, 'train/grad_norm_pre_clip': 12.209345817565918, 'timing/data': 0.0047291930532082915, 'timing/model': 1.202272992930375, 'learning_rate/qwen_vl_interface': 6.078847311406033e-07, 'learning_rate/action_model': 6.078847311406032e-06, 'epoch': 1.92}) +[2026-06-17 18:31:13,384][starVLA.training.train_starvla][INFO] - Step 6532, Loss: {'train/loss': 0.5062772035598755, 'train/grad_norm_pre_clip': 12.414241790771484, 'timing/data': 0.005054936045780778, 'timing/model': 1.203582119080238, 'learning_rate/qwen_vl_interface': 6.07426898992712e-07, 'learning_rate/action_model': 6.074268989927119e-06, 'epoch': 1.92}) +[2026-06-17 18:31:18,161][starVLA.training.train_starvla][INFO] - Step 6533, Loss: {'train/loss': 0.7139338850975037, 'train/grad_norm_pre_clip': 16.163171768188477, 'timing/data': 0.004846779978834093, 'timing/model': 1.3268526520114392, 'learning_rate/qwen_vl_interface': 6.069700292552202e-07, 'learning_rate/action_model': 6.069700292552202e-06, 'epoch': 1.92}) +[2026-06-17 18:31:22,875][starVLA.training.train_starvla][INFO] - Step 6534, Loss: {'train/loss': 0.4118868112564087, 'train/grad_norm_pre_clip': 12.004409790039062, 'timing/data': 0.004640512983314693, 'timing/model': 1.2785237139323726, 'learning_rate/qwen_vl_interface': 6.065141220228372e-07, 'learning_rate/action_model': 6.065141220228372e-06, 'epoch': 1.92}) +[2026-06-17 18:31:27,508][starVLA.training.train_starvla][INFO] - Step 6535, Loss: {'train/loss': 0.6440497636795044, 'train/grad_norm_pre_clip': 12.91470718383789, 'timing/data': 0.005089981015771627, 'timing/model': 1.1993633079109713, 'learning_rate/qwen_vl_interface': 6.060591773900732e-07, 'learning_rate/action_model': 6.060591773900733e-06, 'epoch': 1.92}) +[2026-06-17 18:31:32,278][starVLA.training.train_starvla][INFO] - Step 6536, Loss: {'train/loss': 0.6579571962356567, 'train/grad_norm_pre_clip': 18.953222274780273, 'timing/data': 0.004555371007882059, 'timing/model': 1.2004159829812124, 'learning_rate/qwen_vl_interface': 6.056051954512388e-07, 'learning_rate/action_model': 6.0560519545123875e-06, 'epoch': 1.92}) +[2026-06-17 18:31:37,044][starVLA.training.train_starvla][INFO] - Step 6537, Loss: {'train/loss': 0.519217312335968, 'train/grad_norm_pre_clip': 13.554635047912598, 'timing/data': 0.005132245016284287, 'timing/model': 1.2005481240339577, 'learning_rate/qwen_vl_interface': 6.051521763004455e-07, 'learning_rate/action_model': 6.051521763004456e-06, 'epoch': 1.92}) +[2026-06-17 18:31:41,839][starVLA.training.train_starvla][INFO] - Step 6538, Loss: {'train/loss': 0.4408794045448303, 'train/grad_norm_pre_clip': 13.051358222961426, 'timing/data': 0.004979908932000399, 'timing/model': 1.2045551519840956, 'learning_rate/qwen_vl_interface': 6.047001200316038e-07, 'learning_rate/action_model': 6.047001200316037e-06, 'epoch': 1.92}) +[2026-06-17 18:31:46,605][starVLA.training.train_starvla][INFO] - Step 6539, Loss: {'train/loss': 0.7918583750724792, 'train/grad_norm_pre_clip': 17.387889862060547, 'timing/data': 0.004885222064331174, 'timing/model': 1.199707075022161, 'learning_rate/qwen_vl_interface': 6.042490267384255e-07, 'learning_rate/action_model': 6.0424902673842545e-06, 'epoch': 1.92}) +[2026-06-17 18:31:51,364][starVLA.training.train_starvla][INFO] - Step 6540, Loss: {'train/loss': 0.619597852230072, 'train/grad_norm_pre_clip': 14.682653427124023, 'timing/data': 0.005107603967189789, 'timing/model': 1.2869014970492572, 'learning_rate/qwen_vl_interface': 6.03798896514423e-07, 'learning_rate/action_model': 6.0379889651442295e-06, 'epoch': 1.92}) +[2026-06-17 18:31:56,021][starVLA.training.train_starvla][INFO] - Step 6541, Loss: {'train/loss': 0.6982734203338623, 'train/grad_norm_pre_clip': 17.494665145874023, 'timing/data': 0.004951647017151117, 'timing/model': 1.203570958925411, 'learning_rate/qwen_vl_interface': 6.033497294529093e-07, 'learning_rate/action_model': 6.033497294529093e-06, 'epoch': 1.92}) +[2026-06-17 18:32:00,797][starVLA.training.train_starvla][INFO] - Step 6542, Loss: {'train/loss': 0.6415805220603943, 'train/grad_norm_pre_clip': 14.817408561706543, 'timing/data': 0.005029123974964023, 'timing/model': 1.2033801400102675, 'learning_rate/qwen_vl_interface': 6.029015256469967e-07, 'learning_rate/action_model': 6.0290152564699666e-06, 'epoch': 1.93}) +[2026-06-17 18:32:05,578][starVLA.training.train_starvla][INFO] - Step 6543, Loss: {'train/loss': 0.45636746287345886, 'train/grad_norm_pre_clip': 10.48770523071289, 'timing/data': 0.004725545062683523, 'timing/model': 1.2048256039852276, 'learning_rate/qwen_vl_interface': 6.024542851895976e-07, 'learning_rate/action_model': 6.0245428518959756e-06, 'epoch': 1.93}) +[2026-06-17 18:32:10,362][starVLA.training.train_starvla][INFO] - Step 6544, Loss: {'train/loss': 0.6544797420501709, 'train/grad_norm_pre_clip': 13.799817085266113, 'timing/data': 0.005193821038119495, 'timing/model': 1.2049703689990565, 'learning_rate/qwen_vl_interface': 6.020080081734267e-07, 'learning_rate/action_model': 6.020080081734266e-06, 'epoch': 1.93}) +[2026-06-17 18:32:15,143][starVLA.training.train_starvla][INFO] - Step 6545, Loss: {'train/loss': 0.5997139811515808, 'train/grad_norm_pre_clip': 14.253175735473633, 'timing/data': 0.005380952963605523, 'timing/model': 1.3321964989881963, 'learning_rate/qwen_vl_interface': 6.015626946909974e-07, 'learning_rate/action_model': 6.015626946909973e-06, 'epoch': 1.93}) +[2026-06-17 18:32:19,788][starVLA.training.train_starvla][INFO] - Step 6546, Loss: {'train/loss': 0.7774313688278198, 'train/grad_norm_pre_clip': 14.605247497558594, 'timing/data': 0.005011276924051344, 'timing/model': 1.2016017290297896, 'learning_rate/qwen_vl_interface': 6.011183448346233e-07, 'learning_rate/action_model': 6.011183448346233e-06, 'epoch': 1.93}) +[2026-06-17 18:32:24,573][starVLA.training.train_starvla][INFO] - Step 6547, Loss: {'train/loss': 0.6746454238891602, 'train/grad_norm_pre_clip': 17.639596939086914, 'timing/data': 0.0046673460165038705, 'timing/model': 1.2058378969086334, 'learning_rate/qwen_vl_interface': 6.006749586964183e-07, 'learning_rate/action_model': 6.006749586964183e-06, 'epoch': 1.93}) +[2026-06-17 18:32:29,431][starVLA.training.train_starvla][INFO] - Step 6548, Loss: {'train/loss': 0.7748265862464905, 'train/grad_norm_pre_clip': 22.336545944213867, 'timing/data': 0.004963240004144609, 'timing/model': 1.2033308650134131, 'learning_rate/qwen_vl_interface': 6.002325363682974e-07, 'learning_rate/action_model': 6.0023253636829736e-06, 'epoch': 1.93}) +[2026-06-17 18:32:34,145][starVLA.training.train_starvla][INFO] - Step 6549, Loss: {'train/loss': 0.49161794781684875, 'train/grad_norm_pre_clip': 13.100520133972168, 'timing/data': 0.005198430037125945, 'timing/model': 1.2051163569558412, 'learning_rate/qwen_vl_interface': 5.997910779419756e-07, 'learning_rate/action_model': 5.9979107794197565e-06, 'epoch': 1.93}) +[2026-06-17 18:32:38,896][starVLA.training.train_starvla][INFO] - Step 6550, Loss: {'train/loss': 0.6195522546768188, 'train/grad_norm_pre_clip': 14.287012100219727, 'timing/data': 0.005068400059826672, 'timing/model': 1.209935775026679, 'learning_rate/qwen_vl_interface': 5.993505835089661e-07, 'learning_rate/action_model': 5.993505835089661e-06, 'epoch': 1.93}) +[2026-06-17 18:32:43,678][starVLA.training.train_starvla][INFO] - Step 6551, Loss: {'train/loss': 0.4839670658111572, 'train/grad_norm_pre_clip': 12.830018043518066, 'timing/data': 0.004933672957122326, 'timing/model': 1.333022263017483, 'learning_rate/qwen_vl_interface': 5.98911053160586e-07, 'learning_rate/action_model': 5.989110531605859e-06, 'epoch': 1.93}) +[2026-06-17 18:32:48,325][starVLA.training.train_starvla][INFO] - Step 6552, Loss: {'train/loss': 0.786917507648468, 'train/grad_norm_pre_clip': 21.75101661682129, 'timing/data': 0.005068350932560861, 'timing/model': 1.2020443710498512, 'learning_rate/qwen_vl_interface': 5.984724869879494e-07, 'learning_rate/action_model': 5.984724869879493e-06, 'epoch': 1.93}) +[2026-06-17 18:32:53,118][starVLA.training.train_starvla][INFO] - Step 6553, Loss: {'train/loss': 0.6927388906478882, 'train/grad_norm_pre_clip': 19.85253143310547, 'timing/data': 0.005098034045659006, 'timing/model': 1.2035995150217786, 'learning_rate/qwen_vl_interface': 5.980348850819711e-07, 'learning_rate/action_model': 5.98034885081971e-06, 'epoch': 1.93}) +[2026-06-17 18:32:57,908][starVLA.training.train_starvla][INFO] - Step 6554, Loss: {'train/loss': 0.6912407875061035, 'train/grad_norm_pre_clip': 14.803823471069336, 'timing/data': 0.0051942659774795175, 'timing/model': 1.2099139880156144, 'learning_rate/qwen_vl_interface': 5.975982475333664e-07, 'learning_rate/action_model': 5.975982475333663e-06, 'epoch': 1.93}) +[2026-06-17 18:33:02,682][starVLA.training.train_starvla][INFO] - Step 6555, Loss: {'train/loss': 0.5653421878814697, 'train/grad_norm_pre_clip': 12.904967308044434, 'timing/data': 0.004963145940564573, 'timing/model': 1.2039904339471832, 'learning_rate/qwen_vl_interface': 5.971625744326516e-07, 'learning_rate/action_model': 5.971625744326515e-06, 'epoch': 1.93}) +[2026-06-17 18:33:07,421][starVLA.training.train_starvla][INFO] - Step 6556, Loss: {'train/loss': 0.576049268245697, 'train/grad_norm_pre_clip': 17.027177810668945, 'timing/data': 0.0050648750038817525, 'timing/model': 1.2046083169989288, 'learning_rate/qwen_vl_interface': 5.967278658701421e-07, 'learning_rate/action_model': 5.967278658701421e-06, 'epoch': 1.93}) +[2026-06-17 18:33:12,205][starVLA.training.train_starvla][INFO] - Step 6557, Loss: {'train/loss': 0.6715952157974243, 'train/grad_norm_pre_clip': 15.412239074707031, 'timing/data': 0.004967823973856866, 'timing/model': 1.3230118789942935, 'learning_rate/qwen_vl_interface': 5.962941219359527e-07, 'learning_rate/action_model': 5.962941219359527e-06, 'epoch': 1.93}) +[2026-06-17 18:33:16,866][starVLA.training.train_starvla][INFO] - Step 6558, Loss: {'train/loss': 0.7211431860923767, 'train/grad_norm_pre_clip': 16.113147735595703, 'timing/data': 0.00522826099768281, 'timing/model': 1.2047173019964248, 'learning_rate/qwen_vl_interface': 5.958613427199998e-07, 'learning_rate/action_model': 5.958613427199997e-06, 'epoch': 1.93}) +[2026-06-17 18:33:21,661][starVLA.training.train_starvla][INFO] - Step 6559, Loss: {'train/loss': 0.7244921922683716, 'train/grad_norm_pre_clip': 19.062076568603516, 'timing/data': 0.005153810023330152, 'timing/model': 1.2050284639699385, 'learning_rate/qwen_vl_interface': 5.954295283119985e-07, 'learning_rate/action_model': 5.954295283119985e-06, 'epoch': 1.93}) +[2026-06-17 18:33:26,458][starVLA.training.train_starvla][INFO] - Step 6560, Loss: {'train/loss': 0.6314767003059387, 'train/grad_norm_pre_clip': 15.944671630859375, 'timing/data': 0.0044074730249121785, 'timing/model': 1.2062204340472817, 'learning_rate/qwen_vl_interface': 5.949986788014643e-07, 'learning_rate/action_model': 5.949986788014643e-06, 'epoch': 1.93}) +[2026-06-17 18:33:31,247][starVLA.training.train_starvla][INFO] - Step 6561, Loss: {'train/loss': 0.6769357919692993, 'train/grad_norm_pre_clip': 22.029991149902344, 'timing/data': 0.004926968016661704, 'timing/model': 1.2034375129733235, 'learning_rate/qwen_vl_interface': 5.94568794277713e-07, 'learning_rate/action_model': 5.94568794277713e-06, 'epoch': 1.93}) +[2026-06-17 18:33:36,031][starVLA.training.train_starvla][INFO] - Step 6562, Loss: {'train/loss': 0.8180347084999084, 'train/grad_norm_pre_clip': 13.320897102355957, 'timing/data': 0.004820864996872842, 'timing/model': 1.3338250980013981, 'learning_rate/qwen_vl_interface': 5.941398748298603e-07, 'learning_rate/action_model': 5.941398748298603e-06, 'epoch': 1.93}) +[2026-06-17 18:33:40,696][starVLA.training.train_starvla][INFO] - Step 6563, Loss: {'train/loss': 0.4500699043273926, 'train/grad_norm_pre_clip': 10.323770523071289, 'timing/data': 0.005140500026755035, 'timing/model': 1.2032880230108276, 'learning_rate/qwen_vl_interface': 5.937119205468217e-07, 'learning_rate/action_model': 5.9371192054682165e-06, 'epoch': 1.93}) +[2026-06-17 18:33:45,472][starVLA.training.train_starvla][INFO] - Step 6564, Loss: {'train/loss': 0.677635669708252, 'train/grad_norm_pre_clip': 14.385974884033203, 'timing/data': 0.004960860009305179, 'timing/model': 1.2040723379468545, 'learning_rate/qwen_vl_interface': 5.93284931517312e-07, 'learning_rate/action_model': 5.93284931517312e-06, 'epoch': 1.93}) +[2026-06-17 18:33:50,176][starVLA.training.train_starvla][INFO] - Step 6565, Loss: {'train/loss': 0.5373150706291199, 'train/grad_norm_pre_clip': 35.518070220947266, 'timing/data': 0.004807707038708031, 'timing/model': 1.2051362870261073, 'learning_rate/qwen_vl_interface': 5.928589078298468e-07, 'learning_rate/action_model': 5.928589078298468e-06, 'epoch': 1.93}) +[2026-06-17 18:33:54,908][starVLA.training.train_starvla][INFO] - Step 6566, Loss: {'train/loss': 0.528572142124176, 'train/grad_norm_pre_clip': 12.17870044708252, 'timing/data': 0.004676925018429756, 'timing/model': 1.2054038479691371, 'learning_rate/qwen_vl_interface': 5.924338495727416e-07, 'learning_rate/action_model': 5.9243384957274155e-06, 'epoch': 1.93}) +[2026-06-17 18:33:59,708][starVLA.training.train_starvla][INFO] - Step 6567, Loss: {'train/loss': 0.7007474303245544, 'train/grad_norm_pre_clip': 13.514066696166992, 'timing/data': 0.004994094022549689, 'timing/model': 1.2046168950619176, 'learning_rate/qwen_vl_interface': 5.92009756834111e-07, 'learning_rate/action_model': 5.9200975683411094e-06, 'epoch': 1.93}) +[2026-06-17 18:34:04,507][starVLA.training.train_starvla][INFO] - Step 6568, Loss: {'train/loss': 0.8743504881858826, 'train/grad_norm_pre_clip': 24.859628677368164, 'timing/data': 0.004836451029404998, 'timing/model': 1.341069430927746, 'learning_rate/qwen_vl_interface': 5.915866297018697e-07, 'learning_rate/action_model': 5.915866297018696e-06, 'epoch': 1.93}) +[2026-06-17 18:34:09,167][starVLA.training.train_starvla][INFO] - Step 6569, Loss: {'train/loss': 0.5930131673812866, 'train/grad_norm_pre_clip': 14.365768432617188, 'timing/data': 0.004994604969397187, 'timing/model': 1.2085096889641136, 'learning_rate/qwen_vl_interface': 5.911644682637329e-07, 'learning_rate/action_model': 5.911644682637329e-06, 'epoch': 1.93}) +[2026-06-17 18:34:13,949][starVLA.training.train_starvla][INFO] - Step 6570, Loss: {'train/loss': 0.9006139039993286, 'train/grad_norm_pre_clip': 21.651859283447266, 'timing/data': 0.0046646579867228866, 'timing/model': 1.2050015630666167, 'learning_rate/qwen_vl_interface': 5.907432726072149e-07, 'learning_rate/action_model': 5.9074327260721485e-06, 'epoch': 1.93}) +[2026-06-17 18:34:18,729][starVLA.training.train_starvla][INFO] - Step 6571, Loss: {'train/loss': 0.5227414965629578, 'train/grad_norm_pre_clip': 13.616426467895508, 'timing/data': 0.005185134010389447, 'timing/model': 1.203686799039133, 'learning_rate/qwen_vl_interface': 5.903230428196305e-07, 'learning_rate/action_model': 5.9032304281963045e-06, 'epoch': 1.93}) +[2026-06-17 18:34:23,466][starVLA.training.train_starvla][INFO] - Step 6572, Loss: {'train/loss': 0.8163073658943176, 'train/grad_norm_pre_clip': 34.540069580078125, 'timing/data': 0.005069973063655198, 'timing/model': 1.207439628080465, 'learning_rate/qwen_vl_interface': 5.899037789880929e-07, 'learning_rate/action_model': 5.899037789880929e-06, 'epoch': 1.93}) +[2026-06-17 18:34:28,253][starVLA.training.train_starvla][INFO] - Step 6573, Loss: {'train/loss': 0.6467578411102295, 'train/grad_norm_pre_clip': 15.475058555603027, 'timing/data': 0.004830491030588746, 'timing/model': 1.2137198860291392, 'learning_rate/qwen_vl_interface': 5.894854811995166e-07, 'learning_rate/action_model': 5.894854811995165e-06, 'epoch': 1.93}) +[2026-06-17 18:34:33,150][starVLA.training.train_starvla][INFO] - Step 6574, Loss: {'train/loss': 0.5927378535270691, 'train/grad_norm_pre_clip': 19.058082580566406, 'timing/data': 0.005040414980612695, 'timing/model': 1.3456130399135873, 'learning_rate/qwen_vl_interface': 5.890681495406152e-07, 'learning_rate/action_model': 5.890681495406151e-06, 'epoch': 1.93}) +[2026-06-17 18:34:37,817][starVLA.training.train_starvla][INFO] - Step 6575, Loss: {'train/loss': 0.9458826780319214, 'train/grad_norm_pre_clip': 16.879980087280273, 'timing/data': 0.004553070990368724, 'timing/model': 1.2109249280765653, 'learning_rate/qwen_vl_interface': 5.886517840979015e-07, 'learning_rate/action_model': 5.886517840979016e-06, 'epoch': 1.93}) +[2026-06-17 18:34:42,632][starVLA.training.train_starvla][INFO] - Step 6576, Loss: {'train/loss': 0.5580536723136902, 'train/grad_norm_pre_clip': 15.10452938079834, 'timing/data': 0.005022134981118143, 'timing/model': 1.2068214060273021, 'learning_rate/qwen_vl_interface': 5.882363849576889e-07, 'learning_rate/action_model': 5.882363849576889e-06, 'epoch': 1.94}) +[2026-06-17 18:34:47,420][starVLA.training.train_starvla][INFO] - Step 6577, Loss: {'train/loss': 0.7142161726951599, 'train/grad_norm_pre_clip': 14.206287384033203, 'timing/data': 0.005011591943912208, 'timing/model': 1.2068729740567505, 'learning_rate/qwen_vl_interface': 5.8782195220609e-07, 'learning_rate/action_model': 5.878219522060899e-06, 'epoch': 1.94}) +[2026-06-17 18:34:52,211][starVLA.training.train_starvla][INFO] - Step 6578, Loss: {'train/loss': 0.5107001066207886, 'train/grad_norm_pre_clip': 12.07582950592041, 'timing/data': 0.005075925961136818, 'timing/model': 1.2065211889566854, 'learning_rate/qwen_vl_interface': 5.874084859290172e-07, 'learning_rate/action_model': 5.874084859290171e-06, 'epoch': 1.94}) +[2026-06-17 18:34:56,894][starVLA.training.train_starvla][INFO] - Step 6579, Loss: {'train/loss': 0.4997645318508148, 'train/grad_norm_pre_clip': 15.789164543151855, 'timing/data': 0.005046507925726473, 'timing/model': 1.2055302039952949, 'learning_rate/qwen_vl_interface': 5.869959862121822e-07, 'learning_rate/action_model': 5.869959862121822e-06, 'epoch': 1.94}) +[2026-06-17 18:35:01,594][starVLA.training.train_starvla][INFO] - Step 6580, Loss: {'train/loss': 0.9102005958557129, 'train/grad_norm_pre_clip': 21.34665298461914, 'timing/data': 0.005410131998360157, 'timing/model': 1.2469255570322275, 'learning_rate/qwen_vl_interface': 5.86584453141097e-07, 'learning_rate/action_model': 5.86584453141097e-06, 'epoch': 1.94}) +[2026-06-17 18:35:06,296][starVLA.training.train_starvla][INFO] - Step 6581, Loss: {'train/loss': 0.537177324295044, 'train/grad_norm_pre_clip': 14.994268417358398, 'timing/data': 0.005051121930591762, 'timing/model': 1.203818772919476, 'learning_rate/qwen_vl_interface': 5.861738868010723e-07, 'learning_rate/action_model': 5.861738868010723e-06, 'epoch': 1.94}) +[2026-06-17 18:35:11,210][starVLA.training.train_starvla][INFO] - Step 6582, Loss: {'train/loss': 0.5428733825683594, 'train/grad_norm_pre_clip': 14.191580772399902, 'timing/data': 0.005067583988420665, 'timing/model': 1.2047282779822126, 'learning_rate/qwen_vl_interface': 5.857642872772198e-07, 'learning_rate/action_model': 5.8576428727721986e-06, 'epoch': 1.94}) +[2026-06-17 18:35:16,010][starVLA.training.train_starvla][INFO] - Step 6583, Loss: {'train/loss': 0.7320644855499268, 'train/grad_norm_pre_clip': 18.50691795349121, 'timing/data': 0.005301231052726507, 'timing/model': 1.2072492890292779, 'learning_rate/qwen_vl_interface': 5.853556546544491e-07, 'learning_rate/action_model': 5.8535565465444904e-06, 'epoch': 1.94}) +[2026-06-17 18:35:20,803][starVLA.training.train_starvla][INFO] - Step 6584, Loss: {'train/loss': 0.38843926787376404, 'train/grad_norm_pre_clip': 10.003626823425293, 'timing/data': 0.005219436949118972, 'timing/model': 1.2047725330339745, 'learning_rate/qwen_vl_interface': 5.849479890174708e-07, 'learning_rate/action_model': 5.8494798901747075e-06, 'epoch': 1.94}) +[2026-06-17 18:35:25,595][starVLA.training.train_starvla][INFO] - Step 6585, Loss: {'train/loss': 0.548076331615448, 'train/grad_norm_pre_clip': 14.880805015563965, 'timing/data': 0.005358494003303349, 'timing/model': 1.3311465450096875, 'learning_rate/qwen_vl_interface': 5.845412904507948e-07, 'learning_rate/action_model': 5.845412904507948e-06, 'epoch': 1.94}) +[2026-06-17 18:35:30,305][starVLA.training.train_starvla][INFO] - Step 6586, Loss: {'train/loss': 0.5182628631591797, 'train/grad_norm_pre_clip': 11.843713760375977, 'timing/data': 0.00538784300442785, 'timing/model': 1.2563996309181675, 'learning_rate/qwen_vl_interface': 5.841355590387291e-07, 'learning_rate/action_model': 5.841355590387291e-06, 'epoch': 1.94}) +[2026-06-17 18:35:34,967][starVLA.training.train_starvla][INFO] - Step 6587, Loss: {'train/loss': 0.7284853458404541, 'train/grad_norm_pre_clip': 14.105494499206543, 'timing/data': 0.004920760984532535, 'timing/model': 1.205828694975935, 'learning_rate/qwen_vl_interface': 5.837307948653821e-07, 'learning_rate/action_model': 5.837307948653821e-06, 'epoch': 1.94}) +[2026-06-17 18:35:39,716][starVLA.training.train_starvla][INFO] - Step 6588, Loss: {'train/loss': 0.4938322603702545, 'train/grad_norm_pre_clip': 14.920762062072754, 'timing/data': 0.004668451962061226, 'timing/model': 1.2048655409598723, 'learning_rate/qwen_vl_interface': 5.833269980146631e-07, 'learning_rate/action_model': 5.83326998014663e-06, 'epoch': 1.94}) +[2026-06-17 18:35:44,517][starVLA.training.train_starvla][INFO] - Step 6589, Loss: {'train/loss': 0.5467846393585205, 'train/grad_norm_pre_clip': 13.55589485168457, 'timing/data': 0.005034959991462529, 'timing/model': 1.2129168950486928, 'learning_rate/qwen_vl_interface': 5.829241685702791e-07, 'learning_rate/action_model': 5.82924168570279e-06, 'epoch': 1.94}) +[2026-06-17 18:35:49,291][starVLA.training.train_starvla][INFO] - Step 6590, Loss: {'train/loss': 0.7845508456230164, 'train/grad_norm_pre_clip': 21.032047271728516, 'timing/data': 0.004657619982026517, 'timing/model': 1.20218477898743, 'learning_rate/qwen_vl_interface': 5.825223066157365e-07, 'learning_rate/action_model': 5.8252230661573645e-06, 'epoch': 1.94}) +[2026-06-17 18:35:54,070][starVLA.training.train_starvla][INFO] - Step 6591, Loss: {'train/loss': 0.9245465993881226, 'train/grad_norm_pre_clip': 33.301551818847656, 'timing/data': 0.00476505106780678, 'timing/model': 1.3295431029982865, 'learning_rate/qwen_vl_interface': 5.821214122343423e-07, 'learning_rate/action_model': 5.821214122343422e-06, 'epoch': 1.94}) +[2026-06-17 18:35:58,848][starVLA.training.train_starvla][INFO] - Step 6592, Loss: {'train/loss': 0.5187186002731323, 'train/grad_norm_pre_clip': 12.974531173706055, 'timing/data': 0.004939593025483191, 'timing/model': 1.3228264609351754, 'learning_rate/qwen_vl_interface': 5.817214855092024e-07, 'learning_rate/action_model': 5.817214855092024e-06, 'epoch': 1.94}) +[2026-06-17 18:36:03,513][starVLA.training.train_starvla][INFO] - Step 6593, Loss: {'train/loss': 0.8859941363334656, 'train/grad_norm_pre_clip': 21.75765037536621, 'timing/data': 0.005154525977559388, 'timing/model': 1.2094804799417034, 'learning_rate/qwen_vl_interface': 5.813225265232216e-07, 'learning_rate/action_model': 5.813225265232215e-06, 'epoch': 1.94}) +[2026-06-17 18:36:08,289][starVLA.training.train_starvla][INFO] - Step 6594, Loss: {'train/loss': 0.5922257304191589, 'train/grad_norm_pre_clip': 13.667476654052734, 'timing/data': 0.006148117012344301, 'timing/model': 1.206612792913802, 'learning_rate/qwen_vl_interface': 5.809245353591048e-07, 'learning_rate/action_model': 5.8092453535910485e-06, 'epoch': 1.94}) +[2026-06-17 18:36:13,081][starVLA.training.train_starvla][INFO] - Step 6595, Loss: {'train/loss': 0.5538314580917358, 'train/grad_norm_pre_clip': 15.233983993530273, 'timing/data': 0.005327844992280006, 'timing/model': 1.2063262690789998, 'learning_rate/qwen_vl_interface': 5.805275120993561e-07, 'learning_rate/action_model': 5.805275120993561e-06, 'epoch': 1.94}) +[2026-06-17 18:36:17,890][starVLA.training.train_starvla][INFO] - Step 6596, Loss: {'train/loss': 0.614499568939209, 'train/grad_norm_pre_clip': 13.878267288208008, 'timing/data': 0.004960275953635573, 'timing/model': 1.2055171229876578, 'learning_rate/qwen_vl_interface': 5.801314568262787e-07, 'learning_rate/action_model': 5.801314568262787e-06, 'epoch': 1.94}) +[2026-06-17 18:36:22,639][starVLA.training.train_starvla][INFO] - Step 6597, Loss: {'train/loss': 0.8627605438232422, 'train/grad_norm_pre_clip': 18.524423599243164, 'timing/data': 0.0052095819264650345, 'timing/model': 1.2899777919519693, 'learning_rate/qwen_vl_interface': 5.797363696219751e-07, 'learning_rate/action_model': 5.797363696219751e-06, 'epoch': 1.94}) +[2026-06-17 18:36:27,307][starVLA.training.train_starvla][INFO] - Step 6598, Loss: {'train/loss': 0.5667657852172852, 'train/grad_norm_pre_clip': 17.08668327331543, 'timing/data': 0.00500994804315269, 'timing/model': 1.2055804359260947, 'learning_rate/qwen_vl_interface': 5.793422505683474e-07, 'learning_rate/action_model': 5.793422505683474e-06, 'epoch': 1.94}) +[2026-06-17 18:36:32,086][starVLA.training.train_starvla][INFO] - Step 6599, Loss: {'train/loss': 0.6265513896942139, 'train/grad_norm_pre_clip': 11.824007987976074, 'timing/data': 0.004911049967631698, 'timing/model': 1.2076024179114029, 'learning_rate/qwen_vl_interface': 5.789490997470978e-07, 'learning_rate/action_model': 5.789490997470978e-06, 'epoch': 1.94}) +[2026-06-17 18:38:13,756][starVLA.training.train_starvla][INFO] - Step 6600, Loss: {'train/loss': 0.4778469502925873, 'train/grad_norm_pre_clip': 12.56350040435791, 'eval/loss': 0.7018105387687683, 'eval/latency_0/loss': 0.7018105387687683, 'eval/demon_fire_precision': 0.9086108854589764, 'eval/demon_fire_recall': 0.9289867109634552, 'eval/demon_fire_f1': 0.9186858316221767, 'eval/demon_fire_rate_model': 0.6846496106785317, 'eval/demon_fire_rate_teacher': 0.6696329254727474, 'eval/demon_left_precision': 0.8332378223495702, 'eval/demon_left_recall': 0.8237960339943343, 'eval/demon_left_f1': 0.8284900284900286, 'eval/demon_left_rate_model': 0.48526140155728587, 'eval/demon_left_rate_teacher': 0.49082313681868744, 'eval/demon_right_precision': 0.8282107574094402, 'eval/demon_right_recall': 0.8525423728813559, 'eval/demon_right_f1': 0.8402004454342985, 'eval/demon_right_rate_model': 0.5066740823136818, 'eval/demon_right_rate_teacher': 0.4922135706340378, 'eval/demon_move_f1': 0.8343452369621636, 'eval/demon_cc_f1': 0.891645963429606, 'eval/latency_0/demon_fire_precision': 0.9086108854589764, 'eval/latency_0/demon_fire_recall': 0.9289867109634552, 'eval/latency_0/demon_fire_f1': 0.9186858316221767, 'eval/latency_0/demon_fire_rate_model': 0.6846496106785317, 'eval/latency_0/demon_fire_rate_teacher': 0.6696329254727474, 'eval/latency_0/demon_left_precision': 0.8332378223495702, 'eval/latency_0/demon_left_recall': 0.8237960339943343, 'eval/latency_0/demon_left_f1': 0.8284900284900286, 'eval/latency_0/demon_left_rate_model': 0.48526140155728587, 'eval/latency_0/demon_left_rate_teacher': 0.49082313681868744, 'eval/latency_0/demon_right_precision': 0.8282107574094402, 'eval/latency_0/demon_right_recall': 0.8525423728813559, 'eval/latency_0/demon_right_f1': 0.8402004454342985, 'eval/latency_0/demon_right_rate_model': 0.5066740823136818, 'eval/latency_0/demon_right_rate_teacher': 0.4922135706340378, 'eval/latency_0/demon_move_f1': 0.8343452369621636, 'eval/latency_0/demon_cc_f1': 0.891645963429606, 'timing/data': 0.004910075920633972, 'timing/model': 1.2052234020084143, 'learning_rate/qwen_vl_interface': 5.785569172397254e-07, 'learning_rate/action_model': 5.785569172397254e-06, 'epoch': 1.94}) +[2026-06-17 18:38:13,756][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_6600_state +[2026-06-17 18:38:13,759][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 18:38:17,323][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_6600_state/model.safetensors +[2026-06-17 18:38:24,956][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_6600_state/optimizer.bin +[2026-06-17 18:38:24,956][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_6600_state/sampler.bin +[2026-06-17 18:38:24,956][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_6600_state/sampler_1.bin +[2026-06-17 18:38:24,957][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_6600_state/random_states_0.pkl +[2026-06-17 18:38:25,005][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 6600 +[2026-06-17 18:38:25,005][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 18:38:25,016][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 18:38:29,616][starVLA.training.train_starvla][INFO] - Step 6601, Loss: {'train/loss': 0.8534793853759766, 'train/grad_norm_pre_clip': 12.321381568908691, 'timing/data': 0.00501791900023818, 'timing/model': 1.190102425054647, 'learning_rate/qwen_vl_interface': 5.781657031275306e-07, 'learning_rate/action_model': 5.781657031275306e-06, 'epoch': 1.94}) +[2026-06-17 18:38:34,368][starVLA.training.train_starvla][INFO] - Step 6602, Loss: {'train/loss': 0.5582305788993835, 'train/grad_norm_pre_clip': 18.55327033996582, 'timing/data': 0.004901895998045802, 'timing/model': 1.1955992570146918, 'learning_rate/qwen_vl_interface': 5.777754574916129e-07, 'learning_rate/action_model': 5.7777545749161285e-06, 'epoch': 1.94}) +[2026-06-17 18:38:39,118][starVLA.training.train_starvla][INFO] - Step 6603, Loss: {'train/loss': 0.8007926344871521, 'train/grad_norm_pre_clip': 15.79849910736084, 'timing/data': 0.0055128129897639155, 'timing/model': 1.1937164099654183, 'learning_rate/qwen_vl_interface': 5.773861804128701e-07, 'learning_rate/action_model': 5.773861804128701e-06, 'epoch': 1.94}) +[2026-06-17 18:38:43,858][starVLA.training.train_starvla][INFO] - Step 6604, Loss: {'train/loss': 0.5675932765007019, 'train/grad_norm_pre_clip': 31.546249389648438, 'timing/data': 0.0049052509712055326, 'timing/model': 1.3242724490119144, 'learning_rate/qwen_vl_interface': 5.769978719720011e-07, 'learning_rate/action_model': 5.76997871972001e-06, 'epoch': 1.94}) +[2026-06-17 18:38:48,494][starVLA.training.train_starvla][INFO] - Step 6605, Loss: {'train/loss': 0.5090757608413696, 'train/grad_norm_pre_clip': 15.462597846984863, 'timing/data': 0.0051917580422014, 'timing/model': 1.2021101110149175, 'learning_rate/qwen_vl_interface': 5.766105322495002e-07, 'learning_rate/action_model': 5.766105322495002e-06, 'epoch': 1.94}) +[2026-06-17 18:38:53,348][starVLA.training.train_starvla][INFO] - Step 6606, Loss: {'train/loss': 1.043250322341919, 'train/grad_norm_pre_clip': 19.685657501220703, 'timing/data': 0.005077305948361754, 'timing/model': 1.1960404349956661, 'learning_rate/qwen_vl_interface': 5.762241613256654e-07, 'learning_rate/action_model': 5.762241613256654e-06, 'epoch': 1.94}) +[2026-06-17 18:38:58,127][starVLA.training.train_starvla][INFO] - Step 6607, Loss: {'train/loss': 0.7115753889083862, 'train/grad_norm_pre_clip': 24.25469398498535, 'timing/data': 0.004905904992483556, 'timing/model': 1.1960041489219293, 'learning_rate/qwen_vl_interface': 5.758387592805916e-07, 'learning_rate/action_model': 5.758387592805915e-06, 'epoch': 1.94}) +[2026-06-17 18:39:02,881][starVLA.training.train_starvla][INFO] - Step 6608, Loss: {'train/loss': 0.8648896813392639, 'train/grad_norm_pre_clip': 16.18172264099121, 'timing/data': 0.005044193007051945, 'timing/model': 1.197570607997477, 'learning_rate/qwen_vl_interface': 5.754543261941719e-07, 'learning_rate/action_model': 5.754543261941718e-06, 'epoch': 1.94}) +[2026-06-17 18:39:07,614][starVLA.training.train_starvla][INFO] - Step 6609, Loss: {'train/loss': 0.6202623248100281, 'train/grad_norm_pre_clip': 15.168301582336426, 'timing/data': 0.005139406071975827, 'timing/model': 1.1965736049460247, 'learning_rate/qwen_vl_interface': 5.750708621461011e-07, 'learning_rate/action_model': 5.75070862146101e-06, 'epoch': 1.94}) +[2026-06-17 18:39:12,343][starVLA.training.train_starvla][INFO] - Step 6610, Loss: {'train/loss': 0.8041384816169739, 'train/grad_norm_pre_clip': 16.71359634399414, 'timing/data': 0.005140640074387193, 'timing/model': 1.2937725658994168, 'learning_rate/qwen_vl_interface': 5.746883672158705e-07, 'learning_rate/action_model': 5.746883672158705e-06, 'epoch': 1.95}) +[2026-06-17 18:39:16,993][starVLA.training.train_starvla][INFO] - Step 6611, Loss: {'train/loss': 1.0974342823028564, 'train/grad_norm_pre_clip': 19.432079315185547, 'timing/data': 0.00527380604762584, 'timing/model': 1.1994633100694045, 'learning_rate/qwen_vl_interface': 5.743068414827726e-07, 'learning_rate/action_model': 5.743068414827726e-06, 'epoch': 1.95}) +[2026-06-17 18:39:21,821][starVLA.training.train_starvla][INFO] - Step 6612, Loss: {'train/loss': 0.9954349994659424, 'train/grad_norm_pre_clip': 30.522062301635742, 'timing/data': 0.005113965948112309, 'timing/model': 1.1990183599991724, 'learning_rate/qwen_vl_interface': 5.739262850258968e-07, 'learning_rate/action_model': 5.7392628502589684e-06, 'epoch': 1.95}) +[2026-06-17 18:39:26,606][starVLA.training.train_starvla][INFO] - Step 6613, Loss: {'train/loss': 0.714149534702301, 'train/grad_norm_pre_clip': 14.311875343322754, 'timing/data': 0.004923049011267722, 'timing/model': 1.2023191699991003, 'learning_rate/qwen_vl_interface': 5.735466979241352e-07, 'learning_rate/action_model': 5.735466979241352e-06, 'epoch': 1.95}) +[2026-06-17 18:39:31,387][starVLA.training.train_starvla][INFO] - Step 6614, Loss: {'train/loss': 0.4128648638725281, 'train/grad_norm_pre_clip': 10.946168899536133, 'timing/data': 0.005042543984018266, 'timing/model': 1.2019088689703494, 'learning_rate/qwen_vl_interface': 5.731680802561744e-07, 'learning_rate/action_model': 5.731680802561745e-06, 'epoch': 1.95}) +[2026-06-17 18:39:36,175][starVLA.training.train_starvla][INFO] - Step 6615, Loss: {'train/loss': 0.706981360912323, 'train/grad_norm_pre_clip': 14.364215850830078, 'timing/data': 0.004897497012279928, 'timing/model': 1.3470001970417798, 'learning_rate/qwen_vl_interface': 5.727904321005031e-07, 'learning_rate/action_model': 5.727904321005031e-06, 'epoch': 1.95}) +[2026-06-17 18:39:40,824][starVLA.training.train_starvla][INFO] - Step 6616, Loss: {'train/loss': 0.8075253963470459, 'train/grad_norm_pre_clip': 16.020910263061523, 'timing/data': 0.004942912026308477, 'timing/model': 1.2025106470100582, 'learning_rate/qwen_vl_interface': 5.724137535354077e-07, 'learning_rate/action_model': 5.724137535354077e-06, 'epoch': 1.95}) +[2026-06-17 18:39:45,592][starVLA.training.train_starvla][INFO] - Step 6617, Loss: {'train/loss': 0.5824130773544312, 'train/grad_norm_pre_clip': 18.142675399780273, 'timing/data': 0.005053165019489825, 'timing/model': 1.2004967000102624, 'learning_rate/qwen_vl_interface': 5.720380446389748e-07, 'learning_rate/action_model': 5.720380446389748e-06, 'epoch': 1.95}) +[2026-06-17 18:39:50,373][starVLA.training.train_starvla][INFO] - Step 6618, Loss: {'train/loss': 0.796312689781189, 'train/grad_norm_pre_clip': 15.943812370300293, 'timing/data': 0.005091830971650779, 'timing/model': 1.205252057989128, 'learning_rate/qwen_vl_interface': 5.716633054890897e-07, 'learning_rate/action_model': 5.716633054890896e-06, 'epoch': 1.95}) +[2026-06-17 18:39:55,075][starVLA.training.train_starvla][INFO] - Step 6619, Loss: {'train/loss': 0.49218764901161194, 'train/grad_norm_pre_clip': 16.394943237304688, 'timing/data': 0.005127354990690947, 'timing/model': 1.202925474033691, 'learning_rate/qwen_vl_interface': 5.712895361634345e-07, 'learning_rate/action_model': 5.712895361634345e-06, 'epoch': 1.95}) +[2026-06-17 18:39:59,808][starVLA.training.train_starvla][INFO] - Step 6620, Loss: {'train/loss': 0.7091920971870422, 'train/grad_norm_pre_clip': 18.616853713989258, 'timing/data': 0.005002161022275686, 'timing/model': 1.203147971071303, 'learning_rate/qwen_vl_interface': 5.709167367394936e-07, 'learning_rate/action_model': 5.709167367394936e-06, 'epoch': 1.95}) +[2026-06-17 18:40:04,592][starVLA.training.train_starvla][INFO] - Step 6621, Loss: {'train/loss': 0.7208985090255737, 'train/grad_norm_pre_clip': 16.84398651123047, 'timing/data': 0.005239328951574862, 'timing/model': 1.3332394430181012, 'learning_rate/qwen_vl_interface': 5.705449072945479e-07, 'learning_rate/action_model': 5.705449072945479e-06, 'epoch': 1.95}) +[2026-06-17 18:40:09,249][starVLA.training.train_starvla][INFO] - Step 6622, Loss: {'train/loss': 0.6627729535102844, 'train/grad_norm_pre_clip': 17.802165985107422, 'timing/data': 0.00546055193990469, 'timing/model': 1.2020314169349149, 'learning_rate/qwen_vl_interface': 5.70174047905678e-07, 'learning_rate/action_model': 5.701740479056779e-06, 'epoch': 1.95}) +[2026-06-17 18:40:14,034][starVLA.training.train_starvla][INFO] - Step 6623, Loss: {'train/loss': 0.8071609735488892, 'train/grad_norm_pre_clip': 16.587535858154297, 'timing/data': 0.004885411006398499, 'timing/model': 1.2015925789019093, 'learning_rate/qwen_vl_interface': 5.69804158649764e-07, 'learning_rate/action_model': 5.6980415864976405e-06, 'epoch': 1.95}) +[2026-06-17 18:40:18,817][starVLA.training.train_starvla][INFO] - Step 6624, Loss: {'train/loss': 0.7334877252578735, 'train/grad_norm_pre_clip': 15.21863079071045, 'timing/data': 0.005010835942812264, 'timing/model': 1.2033498489763588, 'learning_rate/qwen_vl_interface': 5.69435239603484e-07, 'learning_rate/action_model': 5.69435239603484e-06, 'epoch': 1.95}) +[2026-06-17 18:40:23,604][starVLA.training.train_starvla][INFO] - Step 6625, Loss: {'train/loss': 0.5179568529129028, 'train/grad_norm_pre_clip': 16.849279403686523, 'timing/data': 0.00472305400762707, 'timing/model': 1.221622785902582, 'learning_rate/qwen_vl_interface': 5.690672908433149e-07, 'learning_rate/action_model': 5.690672908433149e-06, 'epoch': 1.95}) +[2026-06-17 18:40:28,375][starVLA.training.train_starvla][INFO] - Step 6626, Loss: {'train/loss': 0.8511820435523987, 'train/grad_norm_pre_clip': 17.426803588867188, 'timing/data': 0.005010775988921523, 'timing/model': 1.2182434209389612, 'learning_rate/qwen_vl_interface': 5.68700312445534e-07, 'learning_rate/action_model': 5.687003124455339e-06, 'epoch': 1.95}) +[2026-06-17 18:40:33,151][starVLA.training.train_starvla][INFO] - Step 6627, Loss: {'train/loss': 0.6683617830276489, 'train/grad_norm_pre_clip': 18.997278213500977, 'timing/data': 0.005051856976933777, 'timing/model': 1.325298793031834, 'learning_rate/qwen_vl_interface': 5.683343044862151e-07, 'learning_rate/action_model': 5.68334304486215e-06, 'epoch': 1.95}) +[2026-06-17 18:40:37,801][starVLA.training.train_starvla][INFO] - Step 6628, Loss: {'train/loss': 0.6175404191017151, 'train/grad_norm_pre_clip': 13.757926940917969, 'timing/data': 0.005015161936171353, 'timing/model': 1.2041137230116874, 'learning_rate/qwen_vl_interface': 5.679692670412324e-07, 'learning_rate/action_model': 5.679692670412324e-06, 'epoch': 1.95}) +[2026-06-17 18:40:42,598][starVLA.training.train_starvla][INFO] - Step 6629, Loss: {'train/loss': 0.7883056402206421, 'train/grad_norm_pre_clip': 16.654617309570312, 'timing/data': 0.005157008999958634, 'timing/model': 1.2022502159234136, 'learning_rate/qwen_vl_interface': 5.67605200186259e-07, 'learning_rate/action_model': 5.67605200186259e-06, 'epoch': 1.95}) +[2026-06-17 18:40:47,395][starVLA.training.train_starvla][INFO] - Step 6630, Loss: {'train/loss': 0.47369998693466187, 'train/grad_norm_pre_clip': 13.911845207214355, 'timing/data': 0.005000006058253348, 'timing/model': 1.211075786035508, 'learning_rate/qwen_vl_interface': 5.67242103996766e-07, 'learning_rate/action_model': 5.67242103996766e-06, 'epoch': 1.95}) +[2026-06-17 18:40:52,191][starVLA.training.train_starvla][INFO] - Step 6631, Loss: {'train/loss': 0.6238097548484802, 'train/grad_norm_pre_clip': 15.678893089294434, 'timing/data': 0.0052836069371551275, 'timing/model': 1.211194687988609, 'learning_rate/qwen_vl_interface': 5.668799785480242e-07, 'learning_rate/action_model': 5.668799785480242e-06, 'epoch': 1.95}) +[2026-06-17 18:40:56,969][starVLA.training.train_starvla][INFO] - Step 6632, Loss: {'train/loss': 0.7047072052955627, 'train/grad_norm_pre_clip': 16.114803314208984, 'timing/data': 0.005084812990389764, 'timing/model': 1.3301414269953966, 'learning_rate/qwen_vl_interface': 5.665188239151019e-07, 'learning_rate/action_model': 5.665188239151018e-06, 'epoch': 1.95}) +[2026-06-17 18:41:01,619][starVLA.training.train_starvla][INFO] - Step 6633, Loss: {'train/loss': 0.9515064358711243, 'train/grad_norm_pre_clip': 16.453617095947266, 'timing/data': 0.00468303298112005, 'timing/model': 1.2036757339956239, 'learning_rate/qwen_vl_interface': 5.661586401728671e-07, 'learning_rate/action_model': 5.6615864017286705e-06, 'epoch': 1.95}) +[2026-06-17 18:41:06,418][starVLA.training.train_starvla][INFO] - Step 6634, Loss: {'train/loss': 1.0321693420410156, 'train/grad_norm_pre_clip': 20.0004825592041, 'timing/data': 0.004916848964057863, 'timing/model': 1.216095001087524, 'learning_rate/qwen_vl_interface': 5.657994273959857e-07, 'learning_rate/action_model': 5.657994273959856e-06, 'epoch': 1.95}) +[2026-06-17 18:41:11,149][starVLA.training.train_starvla][INFO] - Step 6635, Loss: {'train/loss': 0.6839644908905029, 'train/grad_norm_pre_clip': 16.225894927978516, 'timing/data': 0.007265585940331221, 'timing/model': 1.21033668005839, 'learning_rate/qwen_vl_interface': 5.654411856589233e-07, 'learning_rate/action_model': 5.654411856589233e-06, 'epoch': 1.95}) +[2026-06-17 18:41:15,894][starVLA.training.train_starvla][INFO] - Step 6636, Loss: {'train/loss': 0.7474158406257629, 'train/grad_norm_pre_clip': 15.809473991394043, 'timing/data': 0.0055002940353006124, 'timing/model': 1.2084234450012445, 'learning_rate/qwen_vl_interface': 5.650839150359444e-07, 'learning_rate/action_model': 5.650839150359444e-06, 'epoch': 1.95}) +[2026-06-17 18:41:20,682][starVLA.training.train_starvla][INFO] - Step 6637, Loss: {'train/loss': 0.6578188538551331, 'train/grad_norm_pre_clip': 15.115087509155273, 'timing/data': 0.005139610031619668, 'timing/model': 1.204981927992776, 'learning_rate/qwen_vl_interface': 5.647276156011114e-07, 'learning_rate/action_model': 5.6472761560111135e-06, 'epoch': 1.95}) +[2026-06-17 18:41:25,472][starVLA.training.train_starvla][INFO] - Step 6638, Loss: {'train/loss': 0.4897846579551697, 'train/grad_norm_pre_clip': 11.54520320892334, 'timing/data': 0.00478117901366204, 'timing/model': 1.339870213996619, 'learning_rate/qwen_vl_interface': 5.643722874282846e-07, 'learning_rate/action_model': 5.643722874282846e-06, 'epoch': 1.95}) +[2026-06-17 18:41:30,134][starVLA.training.train_starvla][INFO] - Step 6639, Loss: {'train/loss': 1.1918162107467651, 'train/grad_norm_pre_clip': 18.114049911499023, 'timing/data': 0.0049309240421280265, 'timing/model': 1.2126545289065689, 'learning_rate/qwen_vl_interface': 5.640179305911253e-07, 'learning_rate/action_model': 5.640179305911253e-06, 'epoch': 1.95}) +[2026-06-17 18:41:34,929][starVLA.training.train_starvla][INFO] - Step 6640, Loss: {'train/loss': 0.5952444076538086, 'train/grad_norm_pre_clip': 13.918949127197266, 'timing/data': 0.005075601977296174, 'timing/model': 1.2070369620341808, 'learning_rate/qwen_vl_interface': 5.636645451630908e-07, 'learning_rate/action_model': 5.6366454516309084e-06, 'epoch': 1.95}) +[2026-06-17 18:41:39,702][starVLA.training.train_starvla][INFO] - Step 6641, Loss: {'train/loss': 0.5172860026359558, 'train/grad_norm_pre_clip': 12.805978775024414, 'timing/data': 0.00518327800091356, 'timing/model': 1.204570956993848, 'learning_rate/qwen_vl_interface': 5.63312131217439e-07, 'learning_rate/action_model': 5.633121312174389e-06, 'epoch': 1.95}) +[2026-06-17 18:41:44,449][starVLA.training.train_starvla][INFO] - Step 6642, Loss: {'train/loss': 0.8684021234512329, 'train/grad_norm_pre_clip': 23.025482177734375, 'timing/data': 0.005216473946347833, 'timing/model': 1.2094450030708686, 'learning_rate/qwen_vl_interface': 5.629606888272253e-07, 'learning_rate/action_model': 5.629606888272252e-06, 'epoch': 1.95}) +[2026-06-17 18:41:49,221][starVLA.training.train_starvla][INFO] - Step 6643, Loss: {'train/loss': 0.42395302653312683, 'train/grad_norm_pre_clip': 10.43941593170166, 'timing/data': 0.005409626057371497, 'timing/model': 1.204856054042466, 'learning_rate/qwen_vl_interface': 5.626102180653046e-07, 'learning_rate/action_model': 5.626102180653045e-06, 'epoch': 1.95}) +[2026-06-17 18:41:54,021][starVLA.training.train_starvla][INFO] - Step 6644, Loss: {'train/loss': 0.7136630415916443, 'train/grad_norm_pre_clip': 16.09865379333496, 'timing/data': 0.005018264055252075, 'timing/model': 1.3375772830331698, 'learning_rate/qwen_vl_interface': 5.62260719004329e-07, 'learning_rate/action_model': 5.6226071900432895e-06, 'epoch': 1.96}) +[2026-06-17 18:41:58,679][starVLA.training.train_starvla][INFO] - Step 6645, Loss: {'train/loss': 0.6826425194740295, 'train/grad_norm_pre_clip': 16.145599365234375, 'timing/data': 0.005274121998809278, 'timing/model': 1.2031492800451815, 'learning_rate/qwen_vl_interface': 5.619121917167504e-07, 'learning_rate/action_model': 5.619121917167505e-06, 'epoch': 1.96}) +[2026-06-17 18:42:03,456][starVLA.training.train_starvla][INFO] - Step 6646, Loss: {'train/loss': 0.7476174831390381, 'train/grad_norm_pre_clip': 17.27701759338379, 'timing/data': 0.004957229015417397, 'timing/model': 1.201026126043871, 'learning_rate/qwen_vl_interface': 5.615646362748199e-07, 'learning_rate/action_model': 5.6156463627482e-06, 'epoch': 1.96}) +[2026-06-17 18:42:08,231][starVLA.training.train_starvla][INFO] - Step 6647, Loss: {'train/loss': 0.5497193336486816, 'train/grad_norm_pre_clip': 13.729397773742676, 'timing/data': 0.005103929084725678, 'timing/model': 1.2027688190573826, 'learning_rate/qwen_vl_interface': 5.612180527505854e-07, 'learning_rate/action_model': 5.612180527505854e-06, 'epoch': 1.96}) +[2026-06-17 18:42:13,088][starVLA.training.train_starvla][INFO] - Step 6648, Loss: {'train/loss': 0.5339064002037048, 'train/grad_norm_pre_clip': 16.394973754882812, 'timing/data': 0.005917489994317293, 'timing/model': 1.250220246030949, 'learning_rate/qwen_vl_interface': 5.608724412158931e-07, 'learning_rate/action_model': 5.608724412158931e-06, 'epoch': 1.96}) +[2026-06-17 18:42:17,763][starVLA.training.train_starvla][INFO] - Step 6649, Loss: {'train/loss': 0.4281632900238037, 'train/grad_norm_pre_clip': 14.78736686706543, 'timing/data': 0.005468104034662247, 'timing/model': 1.2043157760053873, 'learning_rate/qwen_vl_interface': 5.605278017423897e-07, 'learning_rate/action_model': 5.605278017423897e-06, 'epoch': 1.96}) +[2026-06-17 18:42:22,451][starVLA.training.train_starvla][INFO] - Step 6650, Loss: {'train/loss': 0.5915016531944275, 'train/grad_norm_pre_clip': 12.293663024902344, 'timing/data': 0.0052134530851617455, 'timing/model': 1.2400178030366078, 'learning_rate/qwen_vl_interface': 5.601841344015193e-07, 'learning_rate/action_model': 5.601841344015193e-06, 'epoch': 1.96}) +[2026-06-17 18:42:27,106][starVLA.training.train_starvla][INFO] - Step 6651, Loss: {'train/loss': 0.54942387342453, 'train/grad_norm_pre_clip': 13.136262893676758, 'timing/data': 0.0048455450451001525, 'timing/model': 1.2065354549558833, 'learning_rate/qwen_vl_interface': 5.598414392645239e-07, 'learning_rate/action_model': 5.598414392645239e-06, 'epoch': 1.96}) +[2026-06-17 18:42:31,905][starVLA.training.train_starvla][INFO] - Step 6652, Loss: {'train/loss': 0.5301930904388428, 'train/grad_norm_pre_clip': 13.967782974243164, 'timing/data': 0.005103594972752035, 'timing/model': 1.2034145649522543, 'learning_rate/qwen_vl_interface': 5.594997164024454e-07, 'learning_rate/action_model': 5.594997164024454e-06, 'epoch': 1.96}) +[2026-06-17 18:42:36,689][starVLA.training.train_starvla][INFO] - Step 6653, Loss: {'train/loss': 0.5246495008468628, 'train/grad_norm_pre_clip': 12.994430541992188, 'timing/data': 0.00527558196336031, 'timing/model': 1.20552029798273, 'learning_rate/qwen_vl_interface': 5.591589658861228e-07, 'learning_rate/action_model': 5.591589658861228e-06, 'epoch': 1.96}) +[2026-06-17 18:42:41,475][starVLA.training.train_starvla][INFO] - Step 6654, Loss: {'train/loss': 0.752579927444458, 'train/grad_norm_pre_clip': 15.198410987854004, 'timing/data': 0.005334838991984725, 'timing/model': 1.2029336160048842, 'learning_rate/qwen_vl_interface': 5.588191877861945e-07, 'learning_rate/action_model': 5.588191877861944e-06, 'epoch': 1.96}) +[2026-06-17 18:42:46,269][starVLA.training.train_starvla][INFO] - Step 6655, Loss: {'train/loss': 0.4900458753108978, 'train/grad_norm_pre_clip': 11.358013153076172, 'timing/data': 0.005385840893723071, 'timing/model': 1.3460974359186366, 'learning_rate/qwen_vl_interface': 5.584803821730961e-07, 'learning_rate/action_model': 5.58480382173096e-06, 'epoch': 1.96}) +[2026-06-17 18:42:50,972][starVLA.training.train_starvla][INFO] - Step 6656, Loss: {'train/loss': 0.7738115787506104, 'train/grad_norm_pre_clip': 16.43704605102539, 'timing/data': 0.004976046038791537, 'timing/model': 1.2503231789451092, 'learning_rate/qwen_vl_interface': 5.581425491170627e-07, 'learning_rate/action_model': 5.581425491170626e-06, 'epoch': 1.96}) +[2026-06-17 18:42:55,625][starVLA.training.train_starvla][INFO] - Step 6657, Loss: {'train/loss': 0.5737229585647583, 'train/grad_norm_pre_clip': 13.753800392150879, 'timing/data': 0.005289633991196752, 'timing/model': 1.2039901930838823, 'learning_rate/qwen_vl_interface': 5.578056886881274e-07, 'learning_rate/action_model': 5.578056886881274e-06, 'epoch': 1.96}) +[2026-06-17 18:43:00,361][starVLA.training.train_starvla][INFO] - Step 6658, Loss: {'train/loss': 0.40631812810897827, 'train/grad_norm_pre_clip': 10.317390441894531, 'timing/data': 0.00488613301422447, 'timing/model': 1.202797455014661, 'learning_rate/qwen_vl_interface': 5.574698009561223e-07, 'learning_rate/action_model': 5.574698009561223e-06, 'epoch': 1.96}) +[2026-06-17 18:43:05,148][starVLA.training.train_starvla][INFO] - Step 6659, Loss: {'train/loss': 0.5363022089004517, 'train/grad_norm_pre_clip': 14.475176811218262, 'timing/data': 0.005165740964002907, 'timing/model': 1.204454526072368, 'learning_rate/qwen_vl_interface': 5.571348859906763e-07, 'learning_rate/action_model': 5.571348859906764e-06, 'epoch': 1.96}) +[2026-06-17 18:43:09,930][starVLA.training.train_starvla][INFO] - Step 6660, Loss: {'train/loss': 0.9778328537940979, 'train/grad_norm_pre_clip': 21.258544921875, 'timing/data': 0.005190931959077716, 'timing/model': 1.201512568979524, 'learning_rate/qwen_vl_interface': 5.568009438612186e-07, 'learning_rate/action_model': 5.568009438612186e-06, 'epoch': 1.96}) +[2026-06-17 18:43:14,712][starVLA.training.train_starvla][INFO] - Step 6661, Loss: {'train/loss': 0.7459953427314758, 'train/grad_norm_pre_clip': 16.007686614990234, 'timing/data': 0.005190121941268444, 'timing/model': 1.3333337309304625, 'learning_rate/qwen_vl_interface': 5.564679746369755e-07, 'learning_rate/action_model': 5.5646797463697556e-06, 'epoch': 1.96}) +[2026-06-17 18:43:19,502][starVLA.training.train_starvla][INFO] - Step 6662, Loss: {'train/loss': 0.5923787951469421, 'train/grad_norm_pre_clip': 13.609713554382324, 'timing/data': 0.005212176009081304, 'timing/model': 1.2528675110079348, 'learning_rate/qwen_vl_interface': 5.561359783869712e-07, 'learning_rate/action_model': 5.5613597838697116e-06, 'epoch': 1.96}) +[2026-06-17 18:43:24,167][starVLA.training.train_starvla][INFO] - Step 6663, Loss: {'train/loss': 0.8212460875511169, 'train/grad_norm_pre_clip': 12.492982864379883, 'timing/data': 0.005169729003682733, 'timing/model': 1.20492116198875, 'learning_rate/qwen_vl_interface': 5.558049551800291e-07, 'learning_rate/action_model': 5.558049551800292e-06, 'epoch': 1.96}) +[2026-06-17 18:43:28,907][starVLA.training.train_starvla][INFO] - Step 6664, Loss: {'train/loss': 0.5220605134963989, 'train/grad_norm_pre_clip': 12.519369125366211, 'timing/data': 0.005018894909881055, 'timing/model': 1.2018209380330518, 'learning_rate/qwen_vl_interface': 5.554749050847718e-07, 'learning_rate/action_model': 5.5547490508477176e-06, 'epoch': 1.96}) +[2026-06-17 18:43:33,689][starVLA.training.train_starvla][INFO] - Step 6665, Loss: {'train/loss': 0.757965624332428, 'train/grad_norm_pre_clip': 16.912004470825195, 'timing/data': 0.004854526952840388, 'timing/model': 1.2053560690255836, 'learning_rate/qwen_vl_interface': 5.55145828169618e-07, 'learning_rate/action_model': 5.5514582816961805e-06, 'epoch': 1.96}) +[2026-06-17 18:43:38,475][starVLA.training.train_starvla][INFO] - Step 6666, Loss: {'train/loss': 0.575899064540863, 'train/grad_norm_pre_clip': 13.381216049194336, 'timing/data': 0.004825209034606814, 'timing/model': 1.2040307850111276, 'learning_rate/qwen_vl_interface': 5.548177245027859e-07, 'learning_rate/action_model': 5.548177245027859e-06, 'epoch': 1.96}) +[2026-06-17 18:43:43,265][starVLA.training.train_starvla][INFO] - Step 6667, Loss: {'train/loss': 0.7262400984764099, 'train/grad_norm_pre_clip': 15.913935661315918, 'timing/data': 0.005386960110627115, 'timing/model': 1.331482143024914, 'learning_rate/qwen_vl_interface': 5.544905941522913e-07, 'learning_rate/action_model': 5.544905941522913e-06, 'epoch': 1.96}) +[2026-06-17 18:43:48,044][starVLA.training.train_starvla][INFO] - Step 6668, Loss: {'train/loss': 0.6084918975830078, 'train/grad_norm_pre_clip': 22.609813690185547, 'timing/data': 0.0046948440140113235, 'timing/model': 1.3278342449339107, 'learning_rate/qwen_vl_interface': 5.541644371859498e-07, 'learning_rate/action_model': 5.541644371859498e-06, 'epoch': 1.96}) +[2026-06-17 18:43:52,697][starVLA.training.train_starvla][INFO] - Step 6669, Loss: {'train/loss': 0.48812681436538696, 'train/grad_norm_pre_clip': 11.409470558166504, 'timing/data': 0.005151372984983027, 'timing/model': 1.205452746944502, 'learning_rate/qwen_vl_interface': 5.53839253671373e-07, 'learning_rate/action_model': 5.5383925367137305e-06, 'epoch': 1.96}) +[2026-06-17 18:43:57,521][starVLA.training.train_starvla][INFO] - Step 6670, Loss: {'train/loss': 0.7947375178337097, 'train/grad_norm_pre_clip': 15.726432800292969, 'timing/data': 0.005275456001982093, 'timing/model': 1.2055149900261313, 'learning_rate/qwen_vl_interface': 5.535150436759725e-07, 'learning_rate/action_model': 5.535150436759725e-06, 'epoch': 1.96}) +[2026-06-17 18:44:02,302][starVLA.training.train_starvla][INFO] - Step 6671, Loss: {'train/loss': 0.6825542449951172, 'train/grad_norm_pre_clip': 15.461349487304688, 'timing/data': 0.004797277040779591, 'timing/model': 1.203923113993369, 'learning_rate/qwen_vl_interface': 5.531918072669566e-07, 'learning_rate/action_model': 5.531918072669565e-06, 'epoch': 1.96}) +[2026-06-17 18:44:07,105][starVLA.training.train_starvla][INFO] - Step 6672, Loss: {'train/loss': 0.7379604578018188, 'train/grad_norm_pre_clip': 17.4240779876709, 'timing/data': 0.005225726985372603, 'timing/model': 1.2103615299565718, 'learning_rate/qwen_vl_interface': 5.528695445113336e-07, 'learning_rate/action_model': 5.528695445113335e-06, 'epoch': 1.96}) +[2026-06-17 18:44:11,868][starVLA.training.train_starvla][INFO] - Step 6673, Loss: {'train/loss': 0.7373704314231873, 'train/grad_norm_pre_clip': 14.882970809936523, 'timing/data': 0.004967654938809574, 'timing/model': 1.283844409044832, 'learning_rate/qwen_vl_interface': 5.525482554759079e-07, 'learning_rate/action_model': 5.5254825547590795e-06, 'epoch': 1.96}) +[2026-06-17 18:44:16,548][starVLA.training.train_starvla][INFO] - Step 6674, Loss: {'train/loss': 0.8302518725395203, 'train/grad_norm_pre_clip': 16.92123794555664, 'timing/data': 0.004858525935560465, 'timing/model': 1.20440507796593, 'learning_rate/qwen_vl_interface': 5.522279402272834e-07, 'learning_rate/action_model': 5.522279402272834e-06, 'epoch': 1.96}) +[2026-06-17 18:44:21,379][starVLA.training.train_starvla][INFO] - Step 6675, Loss: {'train/loss': 0.8060710430145264, 'train/grad_norm_pre_clip': 12.338725090026855, 'timing/data': 0.004968628054484725, 'timing/model': 1.2050325879827142, 'learning_rate/qwen_vl_interface': 5.519085988318627e-07, 'learning_rate/action_model': 5.5190859883186265e-06, 'epoch': 1.96}) +[2026-06-17 18:44:26,175][starVLA.training.train_starvla][INFO] - Step 6676, Loss: {'train/loss': 0.5558767318725586, 'train/grad_norm_pre_clip': 12.311821937561035, 'timing/data': 0.005213472992181778, 'timing/model': 1.2093957649776712, 'learning_rate/qwen_vl_interface': 5.515902313558443e-07, 'learning_rate/action_model': 5.515902313558444e-06, 'epoch': 1.96}) +[2026-06-17 18:44:30,964][starVLA.training.train_starvla][INFO] - Step 6677, Loss: {'train/loss': 0.6303951740264893, 'train/grad_norm_pre_clip': 16.866107940673828, 'timing/data': 0.005402010050602257, 'timing/model': 1.2051596069941297, 'learning_rate/qwen_vl_interface': 5.512728378652266e-07, 'learning_rate/action_model': 5.512728378652266e-06, 'epoch': 1.96}) +[2026-06-17 18:44:35,758][starVLA.training.train_starvla][INFO] - Step 6678, Loss: {'train/loss': 0.8826262354850769, 'train/grad_norm_pre_clip': 17.153989791870117, 'timing/data': 0.005023836973123252, 'timing/model': 1.3376929140649736, 'learning_rate/qwen_vl_interface': 5.509564184258055e-07, 'learning_rate/action_model': 5.509564184258055e-06, 'epoch': 1.97}) +[2026-06-17 18:44:40,494][starVLA.training.train_starvla][INFO] - Step 6679, Loss: {'train/loss': 0.8037004470825195, 'train/grad_norm_pre_clip': 16.06723976135254, 'timing/data': 0.00471292098518461, 'timing/model': 1.281350475968793, 'learning_rate/qwen_vl_interface': 5.506409731031754e-07, 'learning_rate/action_model': 5.506409731031755e-06, 'epoch': 1.97}) +[2026-06-17 18:44:45,157][starVLA.training.train_starvla][INFO] - Step 6680, Loss: {'train/loss': 0.5239180326461792, 'train/grad_norm_pre_clip': 14.606616020202637, 'timing/data': 0.005189780960790813, 'timing/model': 1.2100906029809266, 'learning_rate/qwen_vl_interface': 5.503265019627289e-07, 'learning_rate/action_model': 5.503265019627289e-06, 'epoch': 1.97}) +[2026-06-17 18:44:49,949][starVLA.training.train_starvla][INFO] - Step 6681, Loss: {'train/loss': 0.8784449100494385, 'train/grad_norm_pre_clip': 16.664592742919922, 'timing/data': 0.00507594901137054, 'timing/model': 1.2057140400866047, 'learning_rate/qwen_vl_interface': 5.500130050696548e-07, 'learning_rate/action_model': 5.500130050696548e-06, 'epoch': 1.97}) +[2026-06-17 18:44:54,740][starVLA.training.train_starvla][INFO] - Step 6682, Loss: {'train/loss': 0.33428651094436646, 'train/grad_norm_pre_clip': 8.822239875793457, 'timing/data': 0.005217051017098129, 'timing/model': 1.208318349905312, 'learning_rate/qwen_vl_interface': 5.497004824889424e-07, 'learning_rate/action_model': 5.497004824889423e-06, 'epoch': 1.97}) +[2026-06-17 18:44:59,532][starVLA.training.train_starvla][INFO] - Step 6683, Loss: {'train/loss': 0.5295352339744568, 'train/grad_norm_pre_clip': 15.500715255737305, 'timing/data': 0.004878075094893575, 'timing/model': 1.2040607339004055, 'learning_rate/qwen_vl_interface': 5.493889342853779e-07, 'learning_rate/action_model': 5.493889342853778e-06, 'epoch': 1.97}) +[2026-06-17 18:45:04,311][starVLA.training.train_starvla][INFO] - Step 6684, Loss: {'train/loss': 0.5554571151733398, 'train/grad_norm_pre_clip': 12.489241600036621, 'timing/data': 0.004815045977011323, 'timing/model': 1.2042150410125032, 'learning_rate/qwen_vl_interface': 5.490783605235451e-07, 'learning_rate/action_model': 5.49078360523545e-06, 'epoch': 1.97}) +[2026-06-17 18:45:09,048][starVLA.training.train_starvla][INFO] - Step 6685, Loss: {'train/loss': 0.577210545539856, 'train/grad_norm_pre_clip': 28.59117317199707, 'timing/data': 0.005139208980835974, 'timing/model': 1.290013263002038, 'learning_rate/qwen_vl_interface': 5.487687612678265e-07, 'learning_rate/action_model': 5.4876876126782645e-06, 'epoch': 1.97}) +[2026-06-17 18:45:13,701][starVLA.training.train_starvla][INFO] - Step 6686, Loss: {'train/loss': 0.6086167097091675, 'train/grad_norm_pre_clip': 14.818904876708984, 'timing/data': 0.005170235992409289, 'timing/model': 1.205223539029248, 'learning_rate/qwen_vl_interface': 5.484601365824032e-07, 'learning_rate/action_model': 5.4846013658240315e-06, 'epoch': 1.97}) +[2026-06-17 18:45:18,487][starVLA.training.train_starvla][INFO] - Step 6687, Loss: {'train/loss': 0.45761871337890625, 'train/grad_norm_pre_clip': 12.290958404541016, 'timing/data': 0.005090927937999368, 'timing/model': 1.2041298460680991, 'learning_rate/qwen_vl_interface': 5.481524865312518e-07, 'learning_rate/action_model': 5.481524865312518e-06, 'epoch': 1.97}) +[2026-06-17 18:45:23,267][starVLA.training.train_starvla][INFO] - Step 6688, Loss: {'train/loss': 0.8441272974014282, 'train/grad_norm_pre_clip': 18.548229217529297, 'timing/data': 0.004969410947524011, 'timing/model': 1.2039223540341482, 'learning_rate/qwen_vl_interface': 5.478458111781496e-07, 'learning_rate/action_model': 5.478458111781496e-06, 'epoch': 1.97}) +[2026-06-17 18:45:28,169][starVLA.training.train_starvla][INFO] - Step 6689, Loss: {'train/loss': 0.5500903129577637, 'train/grad_norm_pre_clip': 12.779316902160645, 'timing/data': 0.005059313029050827, 'timing/model': 1.2027945310110226, 'learning_rate/qwen_vl_interface': 5.475401105866712e-07, 'learning_rate/action_model': 5.4754011058667114e-06, 'epoch': 1.97}) +[2026-06-17 18:45:32,952][starVLA.training.train_starvla][INFO] - Step 6690, Loss: {'train/loss': 0.6201114654541016, 'train/grad_norm_pre_clip': 12.505663871765137, 'timing/data': 0.005149487988092005, 'timing/model': 1.334099144092761, 'learning_rate/qwen_vl_interface': 5.47235384820187e-07, 'learning_rate/action_model': 5.47235384820187e-06, 'epoch': 1.97}) +[2026-06-17 18:45:37,609][starVLA.training.train_starvla][INFO] - Step 6691, Loss: {'train/loss': 0.5284656286239624, 'train/grad_norm_pre_clip': 13.587173461914062, 'timing/data': 0.00518238905351609, 'timing/model': 1.2047385849291459, 'learning_rate/qwen_vl_interface': 5.469316339418681e-07, 'learning_rate/action_model': 5.469316339418681e-06, 'epoch': 1.97}) +[2026-06-17 18:45:42,397][starVLA.training.train_starvla][INFO] - Step 6692, Loss: {'train/loss': 0.6415790319442749, 'train/grad_norm_pre_clip': 13.921422958374023, 'timing/data': 0.00482236803509295, 'timing/model': 1.2078165140701458, 'learning_rate/qwen_vl_interface': 5.466288580146825e-07, 'learning_rate/action_model': 5.466288580146825e-06, 'epoch': 1.97}) +[2026-06-17 18:45:47,191][starVLA.training.train_starvla][INFO] - Step 6693, Loss: {'train/loss': 0.6766887307167053, 'train/grad_norm_pre_clip': 14.891813278198242, 'timing/data': 0.007126563927158713, 'timing/model': 1.2121728990459815, 'learning_rate/qwen_vl_interface': 5.463270571013954e-07, 'learning_rate/action_model': 5.463270571013954e-06, 'epoch': 1.97}) +[2026-06-17 18:45:51,893][starVLA.training.train_starvla][INFO] - Step 6694, Loss: {'train/loss': 0.5694018602371216, 'train/grad_norm_pre_clip': 14.101273536682129, 'timing/data': 0.0054619200527668, 'timing/model': 1.2004784550517797, 'learning_rate/qwen_vl_interface': 5.460262312645709e-07, 'learning_rate/action_model': 5.460262312645709e-06, 'epoch': 1.97}) +[2026-06-17 18:45:56,639][starVLA.training.train_starvla][INFO] - Step 6695, Loss: {'train/loss': 0.6790381073951721, 'train/grad_norm_pre_clip': 14.028258323669434, 'timing/data': 0.005342043004930019, 'timing/model': 1.2039813729934394, 'learning_rate/qwen_vl_interface': 5.457263805665707e-07, 'learning_rate/action_model': 5.457263805665707e-06, 'epoch': 1.97}) +[2026-06-17 18:46:01,420][starVLA.training.train_starvla][INFO] - Step 6696, Loss: {'train/loss': 0.8188421130180359, 'train/grad_norm_pre_clip': 16.030773162841797, 'timing/data': 0.005080956965684891, 'timing/model': 1.3317560069262981, 'learning_rate/qwen_vl_interface': 5.454275050695529e-07, 'learning_rate/action_model': 5.454275050695529e-06, 'epoch': 1.97}) +[2026-06-17 18:46:06,188][starVLA.training.train_starvla][INFO] - Step 6697, Loss: {'train/loss': 0.47982653975486755, 'train/grad_norm_pre_clip': 12.469046592712402, 'timing/data': 0.0049158919136971235, 'timing/model': 1.207961139967665, 'learning_rate/qwen_vl_interface': 5.451296048354764e-07, 'learning_rate/action_model': 5.4512960483547635e-06, 'epoch': 1.97}) +[2026-06-17 18:46:10,983][starVLA.training.train_starvla][INFO] - Step 6698, Loss: {'train/loss': 0.6407453417778015, 'train/grad_norm_pre_clip': 14.079534530639648, 'timing/data': 0.0049157540779560804, 'timing/model': 1.2109958078945056, 'learning_rate/qwen_vl_interface': 5.448326799260951e-07, 'learning_rate/action_model': 5.448326799260952e-06, 'epoch': 1.97}) +[2026-06-17 18:46:15,809][starVLA.training.train_starvla][INFO] - Step 6699, Loss: {'train/loss': 0.8767543435096741, 'train/grad_norm_pre_clip': 18.629446029663086, 'timing/data': 0.0054849430453032255, 'timing/model': 1.2090938079636544, 'learning_rate/qwen_vl_interface': 5.445367304029627e-07, 'learning_rate/action_model': 5.445367304029626e-06, 'epoch': 1.97}) +[2026-06-17 18:47:57,727][starVLA.training.train_starvla][INFO] - Step 6700, Loss: {'train/loss': 0.5406001210212708, 'train/grad_norm_pre_clip': 14.6173677444458, 'eval/loss': 0.6991778016090393, 'eval/latency_0/loss': 0.6991778016090393, 'eval/demon_fire_precision': 0.9118129614438064, 'eval/demon_fire_recall': 0.9231727574750831, 'eval/demon_fire_f1': 0.9174576970697482, 'eval/demon_fire_rate_model': 0.6779755283648499, 'eval/demon_fire_rate_teacher': 0.6696329254727474, 'eval/demon_left_precision': 0.83584686774942, 'eval/demon_left_recall': 0.8164305949008499, 'eval/demon_left_f1': 0.8260246488965319, 'eval/demon_left_rate_model': 0.4794215795328142, 'eval/demon_left_rate_teacher': 0.49082313681868744, 'eval/demon_right_precision': 0.8204015192620727, 'eval/demon_right_recall': 0.8542372881355932, 'eval/demon_right_f1': 0.836977580957653, 'eval/demon_right_rate_model': 0.5125139043381535, 'eval/demon_right_rate_teacher': 0.4922135706340378, 'eval/demon_move_f1': 0.8315011149270924, 'eval/demon_cc_f1': 0.8898608382564571, 'eval/latency_0/demon_fire_precision': 0.9118129614438064, 'eval/latency_0/demon_fire_recall': 0.9231727574750831, 'eval/latency_0/demon_fire_f1': 0.9174576970697482, 'eval/latency_0/demon_fire_rate_model': 0.6779755283648499, 'eval/latency_0/demon_fire_rate_teacher': 0.6696329254727474, 'eval/latency_0/demon_left_precision': 0.83584686774942, 'eval/latency_0/demon_left_recall': 0.8164305949008499, 'eval/latency_0/demon_left_f1': 0.8260246488965319, 'eval/latency_0/demon_left_rate_model': 0.4794215795328142, 'eval/latency_0/demon_left_rate_teacher': 0.49082313681868744, 'eval/latency_0/demon_right_precision': 0.8204015192620727, 'eval/latency_0/demon_right_recall': 0.8542372881355932, 'eval/latency_0/demon_right_f1': 0.836977580957653, 'eval/latency_0/demon_right_rate_model': 0.5125139043381535, 'eval/latency_0/demon_right_rate_teacher': 0.4922135706340378, 'eval/latency_0/demon_move_f1': 0.8315011149270924, 'eval/latency_0/demon_cc_f1': 0.8898608382564571, 'timing/data': 0.00510772701818496, 'timing/model': 1.2084595260675997, 'learning_rate/qwen_vl_interface': 5.442417563274289e-07, 'learning_rate/action_model': 5.442417563274289e-06, 'epoch': 1.97}) +[2026-06-17 18:47:57,728][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_6700_state +[2026-06-17 18:47:57,731][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 18:48:01,298][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_6700_state/model.safetensors +[2026-06-17 18:48:08,992][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_6700_state/optimizer.bin +[2026-06-17 18:48:08,992][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_6700_state/sampler.bin +[2026-06-17 18:48:08,992][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_6700_state/sampler_1.bin +[2026-06-17 18:48:08,993][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_6700_state/random_states_0.pkl +[2026-06-17 18:48:09,046][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 6700 +[2026-06-17 18:48:09,046][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 18:48:09,082][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 18:48:13,806][starVLA.training.train_starvla][INFO] - Step 6701, Loss: {'train/loss': 0.6606508493423462, 'train/grad_norm_pre_clip': 31.89336585998535, 'timing/data': 0.004721197998151183, 'timing/model': 1.1869384680176154, 'learning_rate/qwen_vl_interface': 5.439477577606429e-07, 'learning_rate/action_model': 5.4394775776064295e-06, 'epoch': 1.97}) +[2026-06-17 18:48:18,535][starVLA.training.train_starvla][INFO] - Step 6702, Loss: {'train/loss': 0.8012159466743469, 'train/grad_norm_pre_clip': 17.77311897277832, 'timing/data': 0.004585040966048837, 'timing/model': 1.192640993045643, 'learning_rate/qwen_vl_interface': 5.436547347635506e-07, 'learning_rate/action_model': 5.436547347635506e-06, 'epoch': 1.97}) +[2026-06-17 18:48:23,216][starVLA.training.train_starvla][INFO] - Step 6703, Loss: {'train/loss': 0.6430559158325195, 'train/grad_norm_pre_clip': 16.34523582458496, 'timing/data': 0.0045902259880676866, 'timing/model': 1.1919061579974368, 'learning_rate/qwen_vl_interface': 5.433626873968962e-07, 'learning_rate/action_model': 5.433626873968962e-06, 'epoch': 1.97}) +[2026-06-17 18:48:27,959][starVLA.training.train_starvla][INFO] - Step 6704, Loss: {'train/loss': 0.6072536110877991, 'train/grad_norm_pre_clip': 13.416321754455566, 'timing/data': 0.004790545906871557, 'timing/model': 1.3247085129842162, 'learning_rate/qwen_vl_interface': 5.43071615721221e-07, 'learning_rate/action_model': 5.43071615721221e-06, 'epoch': 1.97}) +[2026-06-17 18:48:32,574][starVLA.training.train_starvla][INFO] - Step 6705, Loss: {'train/loss': 0.6776228547096252, 'train/grad_norm_pre_clip': 13.760730743408203, 'timing/data': 0.004680581972934306, 'timing/model': 1.1956977570662275, 'learning_rate/qwen_vl_interface': 5.427815197968647e-07, 'learning_rate/action_model': 5.4278151979686465e-06, 'epoch': 1.97}) +[2026-06-17 18:48:37,341][starVLA.training.train_starvla][INFO] - Step 6706, Loss: {'train/loss': 0.491624116897583, 'train/grad_norm_pre_clip': 27.816295623779297, 'timing/data': 0.0052864160388708115, 'timing/model': 1.1999934250488877, 'learning_rate/qwen_vl_interface': 5.424923996839647e-07, 'learning_rate/action_model': 5.4249239968396466e-06, 'epoch': 1.97}) +[2026-06-17 18:48:42,108][starVLA.training.train_starvla][INFO] - Step 6707, Loss: {'train/loss': 0.5603793263435364, 'train/grad_norm_pre_clip': 12.817956924438477, 'timing/data': 0.0047817539889365435, 'timing/model': 1.1988513820106164, 'learning_rate/qwen_vl_interface': 5.422042554424557e-07, 'learning_rate/action_model': 5.4220425544245565e-06, 'epoch': 1.97}) +[2026-06-17 18:48:46,857][starVLA.training.train_starvla][INFO] - Step 6708, Loss: {'train/loss': 0.5131937861442566, 'train/grad_norm_pre_clip': 13.856915473937988, 'timing/data': 0.004820197937078774, 'timing/model': 1.1974955529440194, 'learning_rate/qwen_vl_interface': 5.419170871320705e-07, 'learning_rate/action_model': 5.419170871320706e-06, 'epoch': 1.97}) +[2026-06-17 18:48:51,575][starVLA.training.train_starvla][INFO] - Step 6709, Loss: {'train/loss': 0.8654989004135132, 'train/grad_norm_pre_clip': 16.417369842529297, 'timing/data': 0.004728078027255833, 'timing/model': 1.1999711939133704, 'learning_rate/qwen_vl_interface': 5.41630894812339e-07, 'learning_rate/action_model': 5.41630894812339e-06, 'epoch': 1.97}) +[2026-06-17 18:48:56,341][starVLA.training.train_starvla][INFO] - Step 6710, Loss: {'train/loss': 0.5522668957710266, 'train/grad_norm_pre_clip': 13.97346019744873, 'timing/data': 0.005279375007376075, 'timing/model': 1.3154626530595124, 'learning_rate/qwen_vl_interface': 5.413456785425896e-07, 'learning_rate/action_model': 5.413456785425896e-06, 'epoch': 1.97}) +[2026-06-17 18:49:00,980][starVLA.training.train_starvla][INFO] - Step 6711, Loss: {'train/loss': 0.7579249739646912, 'train/grad_norm_pre_clip': 13.6539306640625, 'timing/data': 0.004890276002697647, 'timing/model': 1.198234414914623, 'learning_rate/qwen_vl_interface': 5.410614383819474e-07, 'learning_rate/action_model': 5.4106143838194736e-06, 'epoch': 1.97}) +[2026-06-17 18:49:05,766][starVLA.training.train_starvla][INFO] - Step 6712, Loss: {'train/loss': 0.8913062214851379, 'train/grad_norm_pre_clip': 31.36355209350586, 'timing/data': 0.006320173968560994, 'timing/model': 1.21304579207208, 'learning_rate/qwen_vl_interface': 5.407781743893364e-07, 'learning_rate/action_model': 5.407781743893364e-06, 'epoch': 1.98}) +[2026-06-17 18:49:10,539][starVLA.training.train_starvla][INFO] - Step 6713, Loss: {'train/loss': 0.6972712874412537, 'train/grad_norm_pre_clip': 17.656105041503906, 'timing/data': 0.00490044301841408, 'timing/model': 1.2037554290145636, 'learning_rate/qwen_vl_interface': 5.404958866234772e-07, 'learning_rate/action_model': 5.404958866234772e-06, 'epoch': 1.98}) +[2026-06-17 18:49:15,307][starVLA.training.train_starvla][INFO] - Step 6714, Loss: {'train/loss': 0.7614721059799194, 'train/grad_norm_pre_clip': 14.56237506866455, 'timing/data': 0.0043525430373847485, 'timing/model': 1.201046279980801, 'learning_rate/qwen_vl_interface': 5.402145751428884e-07, 'learning_rate/action_model': 5.402145751428884e-06, 'epoch': 1.98}) +[2026-06-17 18:49:20,083][starVLA.training.train_starvla][INFO] - Step 6715, Loss: {'train/loss': 0.6343636512756348, 'train/grad_norm_pre_clip': 14.81752872467041, 'timing/data': 0.005166541086509824, 'timing/model': 1.329621570999734, 'learning_rate/qwen_vl_interface': 5.399342400058856e-07, 'learning_rate/action_model': 5.399342400058856e-06, 'epoch': 1.98}) +[2026-06-17 18:49:24,738][starVLA.training.train_starvla][INFO] - Step 6716, Loss: {'train/loss': 0.5098638534545898, 'train/grad_norm_pre_clip': 12.477334976196289, 'timing/data': 0.004801734932698309, 'timing/model': 1.2012603860348463, 'learning_rate/qwen_vl_interface': 5.396548812705838e-07, 'learning_rate/action_model': 5.396548812705839e-06, 'epoch': 1.98}) +[2026-06-17 18:49:29,510][starVLA.training.train_starvla][INFO] - Step 6717, Loss: {'train/loss': 0.8025221824645996, 'train/grad_norm_pre_clip': 17.325294494628906, 'timing/data': 0.0048711010022088885, 'timing/model': 1.2014930350705981, 'learning_rate/qwen_vl_interface': 5.393764989948933e-07, 'learning_rate/action_model': 5.393764989948933e-06, 'epoch': 1.98}) +[2026-06-17 18:49:34,291][starVLA.training.train_starvla][INFO] - Step 6718, Loss: {'train/loss': 0.5097028017044067, 'train/grad_norm_pre_clip': 13.037419319152832, 'timing/data': 0.004879025975242257, 'timing/model': 1.2022099880268797, 'learning_rate/qwen_vl_interface': 5.390990932365234e-07, 'learning_rate/action_model': 5.390990932365233e-06, 'epoch': 1.98}) +[2026-06-17 18:49:39,041][starVLA.training.train_starvla][INFO] - Step 6719, Loss: {'train/loss': 0.6298946738243103, 'train/grad_norm_pre_clip': 14.79619312286377, 'timing/data': 0.005109412013553083, 'timing/model': 1.210655369097367, 'learning_rate/qwen_vl_interface': 5.388226640529808e-07, 'learning_rate/action_model': 5.388226640529808e-06, 'epoch': 1.98}) +[2026-06-17 18:49:43,828][starVLA.training.train_starvla][INFO] - Step 6720, Loss: {'train/loss': 0.3171389102935791, 'train/grad_norm_pre_clip': 8.166237831115723, 'timing/data': 0.004503612988628447, 'timing/model': 1.2049698199843988, 'learning_rate/qwen_vl_interface': 5.385472115015697e-07, 'learning_rate/action_model': 5.385472115015696e-06, 'epoch': 1.98}) +[2026-06-17 18:49:48,641][starVLA.training.train_starvla][INFO] - Step 6721, Loss: {'train/loss': 0.49198850989341736, 'train/grad_norm_pre_clip': 22.729829788208008, 'timing/data': 0.004768127924762666, 'timing/model': 1.332530353916809, 'learning_rate/qwen_vl_interface': 5.382727356393909e-07, 'learning_rate/action_model': 5.382727356393909e-06, 'epoch': 1.98}) +[2026-06-17 18:49:53,289][starVLA.training.train_starvla][INFO] - Step 6722, Loss: {'train/loss': 0.6083205938339233, 'train/grad_norm_pre_clip': 13.512330055236816, 'timing/data': 0.0043317979434505105, 'timing/model': 1.2018072970677167, 'learning_rate/qwen_vl_interface': 5.379992365233442e-07, 'learning_rate/action_model': 5.379992365233442e-06, 'epoch': 1.98}) +[2026-06-17 18:49:58,063][starVLA.training.train_starvla][INFO] - Step 6723, Loss: {'train/loss': 0.7133702039718628, 'train/grad_norm_pre_clip': 17.430530548095703, 'timing/data': 0.004840413108468056, 'timing/model': 1.2019633190939203, 'learning_rate/qwen_vl_interface': 5.377267142101261e-07, 'learning_rate/action_model': 5.37726714210126e-06, 'epoch': 1.98}) +[2026-06-17 18:50:02,837][starVLA.training.train_starvla][INFO] - Step 6724, Loss: {'train/loss': 0.630551815032959, 'train/grad_norm_pre_clip': 17.541465759277344, 'timing/data': 0.004882054985500872, 'timing/model': 1.2031356809893623, 'learning_rate/qwen_vl_interface': 5.374551687562306e-07, 'learning_rate/action_model': 5.374551687562306e-06, 'epoch': 1.98}) +[2026-06-17 18:50:07,575][starVLA.training.train_starvla][INFO] - Step 6725, Loss: {'train/loss': 0.5731970071792603, 'train/grad_norm_pre_clip': 13.660653114318848, 'timing/data': 0.004762246971949935, 'timing/model': 1.2042075650533661, 'learning_rate/qwen_vl_interface': 5.371846002179494e-07, 'learning_rate/action_model': 5.371846002179493e-06, 'epoch': 1.98}) +[2026-06-17 18:50:12,349][starVLA.training.train_starvla][INFO] - Step 6726, Loss: {'train/loss': 0.535650908946991, 'train/grad_norm_pre_clip': 11.848026275634766, 'timing/data': 0.004845579038374126, 'timing/model': 1.2059381450526416, 'learning_rate/qwen_vl_interface': 5.36915008651372e-07, 'learning_rate/action_model': 5.369150086513719e-06, 'epoch': 1.98}) +[2026-06-17 18:50:17,134][starVLA.training.train_starvla][INFO] - Step 6727, Loss: {'train/loss': 0.8211532235145569, 'train/grad_norm_pre_clip': 19.40239715576172, 'timing/data': 0.004775718902237713, 'timing/model': 1.335339845973067, 'learning_rate/qwen_vl_interface': 5.366463941123853e-07, 'learning_rate/action_model': 5.366463941123853e-06, 'epoch': 1.98}) +[2026-06-17 18:50:21,789][starVLA.training.train_starvla][INFO] - Step 6728, Loss: {'train/loss': 0.7351197600364685, 'train/grad_norm_pre_clip': 15.952259063720703, 'timing/data': 0.005000642035156488, 'timing/model': 1.2038997580530122, 'learning_rate/qwen_vl_interface': 5.36378756656672e-07, 'learning_rate/action_model': 5.36378756656672e-06, 'epoch': 1.98}) +[2026-06-17 18:50:26,574][starVLA.training.train_starvla][INFO] - Step 6729, Loss: {'train/loss': 0.6950020790100098, 'train/grad_norm_pre_clip': 19.755409240722656, 'timing/data': 0.005164149100892246, 'timing/model': 1.204842168954201, 'learning_rate/qwen_vl_interface': 5.361120963397151e-07, 'learning_rate/action_model': 5.361120963397152e-06, 'epoch': 1.98}) +[2026-06-17 18:50:31,359][starVLA.training.train_starvla][INFO] - Step 6730, Loss: {'train/loss': 0.7253661155700684, 'train/grad_norm_pre_clip': 16.306734085083008, 'timing/data': 0.0046628660056740046, 'timing/model': 1.203063823049888, 'learning_rate/qwen_vl_interface': 5.358464132167927e-07, 'learning_rate/action_model': 5.358464132167927e-06, 'epoch': 1.98}) +[2026-06-17 18:50:36,149][starVLA.training.train_starvla][INFO] - Step 6731, Loss: {'train/loss': 0.8699877858161926, 'train/grad_norm_pre_clip': 20.22201919555664, 'timing/data': 0.0047134900232777, 'timing/model': 1.204024538048543, 'learning_rate/qwen_vl_interface': 5.355817073429818e-07, 'learning_rate/action_model': 5.355817073429818e-06, 'epoch': 1.98}) +[2026-06-17 18:50:40,826][starVLA.training.train_starvla][INFO] - Step 6732, Loss: {'train/loss': 0.551479160785675, 'train/grad_norm_pre_clip': 12.390061378479004, 'timing/data': 0.004836112959310412, 'timing/model': 1.2114536579465494, 'learning_rate/qwen_vl_interface': 5.353179787731557e-07, 'learning_rate/action_model': 5.353179787731557e-06, 'epoch': 1.98}) +[2026-06-17 18:50:45,538][starVLA.training.train_starvla][INFO] - Step 6733, Loss: {'train/loss': 0.8145151138305664, 'train/grad_norm_pre_clip': 15.958125114440918, 'timing/data': 0.004748986917547882, 'timing/model': 1.2493411100003868, 'learning_rate/qwen_vl_interface': 5.350552275619857e-07, 'learning_rate/action_model': 5.350552275619857e-06, 'epoch': 1.98}) +[2026-06-17 18:50:50,200][starVLA.training.train_starvla][INFO] - Step 6734, Loss: {'train/loss': 0.9618144631385803, 'train/grad_norm_pre_clip': 25.307973861694336, 'timing/data': 0.0047404629876837134, 'timing/model': 1.2096408960642293, 'learning_rate/qwen_vl_interface': 5.347934537639405e-07, 'learning_rate/action_model': 5.347934537639405e-06, 'epoch': 1.98}) +[2026-06-17 18:50:54,976][starVLA.training.train_starvla][INFO] - Step 6735, Loss: {'train/loss': 0.6968510150909424, 'train/grad_norm_pre_clip': 18.484397888183594, 'timing/data': 0.005036780959926546, 'timing/model': 1.2125658359145746, 'learning_rate/qwen_vl_interface': 5.345326574332865e-07, 'learning_rate/action_model': 5.345326574332865e-06, 'epoch': 1.98}) +[2026-06-17 18:50:59,796][starVLA.training.train_starvla][INFO] - Step 6736, Loss: {'train/loss': 0.8189698457717896, 'train/grad_norm_pre_clip': 14.594684600830078, 'timing/data': 0.004531605984084308, 'timing/model': 1.208181353053078, 'learning_rate/qwen_vl_interface': 5.342728386240867e-07, 'learning_rate/action_model': 5.3427283862408665e-06, 'epoch': 1.98}) +[2026-06-17 18:51:04,616][starVLA.training.train_starvla][INFO] - Step 6737, Loss: {'train/loss': 0.417220801115036, 'train/grad_norm_pre_clip': 10.16515827178955, 'timing/data': 0.004851332982070744, 'timing/model': 1.2085939360549673, 'learning_rate/qwen_vl_interface': 5.340139973902013e-07, 'learning_rate/action_model': 5.340139973902013e-06, 'epoch': 1.98}) +[2026-06-17 18:51:09,418][starVLA.training.train_starvla][INFO] - Step 6738, Loss: {'train/loss': 0.5524589419364929, 'train/grad_norm_pre_clip': 14.049863815307617, 'timing/data': 0.005144073045812547, 'timing/model': 1.3374741859734058, 'learning_rate/qwen_vl_interface': 5.337561337852892e-07, 'learning_rate/action_model': 5.337561337852892e-06, 'epoch': 1.98}) +[2026-06-17 18:51:14,299][starVLA.training.train_starvla][INFO] - Step 6739, Loss: {'train/loss': 0.3694080114364624, 'train/grad_norm_pre_clip': 9.89387035369873, 'timing/data': 0.0047291090013459325, 'timing/model': 1.3324319559615105, 'learning_rate/qwen_vl_interface': 5.334992478628054e-07, 'learning_rate/action_model': 5.334992478628054e-06, 'epoch': 1.98}) +[2026-06-17 18:51:18,970][starVLA.training.train_starvla][INFO] - Step 6740, Loss: {'train/loss': 0.7836111783981323, 'train/grad_norm_pre_clip': 18.2631778717041, 'timing/data': 0.004651495022699237, 'timing/model': 1.2073635410051793, 'learning_rate/qwen_vl_interface': 5.332433396760027e-07, 'learning_rate/action_model': 5.332433396760028e-06, 'epoch': 1.98}) +[2026-06-17 18:51:23,732][starVLA.training.train_starvla][INFO] - Step 6741, Loss: {'train/loss': 0.7220058441162109, 'train/grad_norm_pre_clip': 16.435184478759766, 'timing/data': 0.004402324091643095, 'timing/model': 1.2040745780104771, 'learning_rate/qwen_vl_interface': 5.329884092779312e-07, 'learning_rate/action_model': 5.329884092779312e-06, 'epoch': 1.98}) +[2026-06-17 18:51:28,546][starVLA.training.train_starvla][INFO] - Step 6742, Loss: {'train/loss': 0.44423097372055054, 'train/grad_norm_pre_clip': 13.223034858703613, 'timing/data': 0.004627724993042648, 'timing/model': 1.2069518020143732, 'learning_rate/qwen_vl_interface': 5.327344567214386e-07, 'learning_rate/action_model': 5.327344567214385e-06, 'epoch': 1.98}) +[2026-06-17 18:51:33,373][starVLA.training.train_starvla][INFO] - Step 6743, Loss: {'train/loss': 0.5672188401222229, 'train/grad_norm_pre_clip': 13.908534049987793, 'timing/data': 0.004595251986756921, 'timing/model': 1.2203296410152689, 'learning_rate/qwen_vl_interface': 5.324814820591683e-07, 'learning_rate/action_model': 5.324814820591682e-06, 'epoch': 1.98}) +[2026-06-17 18:51:38,190][starVLA.training.train_starvla][INFO] - Step 6744, Loss: {'train/loss': 0.8088533282279968, 'train/grad_norm_pre_clip': 16.216461181640625, 'timing/data': 0.004629096016287804, 'timing/model': 1.3392014430137351, 'learning_rate/qwen_vl_interface': 5.322294853435631e-07, 'learning_rate/action_model': 5.322294853435631e-06, 'epoch': 1.98}) +[2026-06-17 18:51:42,857][starVLA.training.train_starvla][INFO] - Step 6745, Loss: {'train/loss': 0.6598065495491028, 'train/grad_norm_pre_clip': 14.153679847717285, 'timing/data': 0.004596657934598625, 'timing/model': 1.206472926074639, 'learning_rate/qwen_vl_interface': 5.319784666268626e-07, 'learning_rate/action_model': 5.319784666268626e-06, 'epoch': 1.98}) +[2026-06-17 18:51:47,695][starVLA.training.train_starvla][INFO] - Step 6746, Loss: {'train/loss': 1.0591286420822144, 'train/grad_norm_pre_clip': 19.42036247253418, 'timing/data': 0.004639104939997196, 'timing/model': 1.2129115220159292, 'learning_rate/qwen_vl_interface': 5.317284259611024e-07, 'learning_rate/action_model': 5.317284259611024e-06, 'epoch': 1.99}) +[2026-06-17 18:51:52,500][starVLA.training.train_starvla][INFO] - Step 6747, Loss: {'train/loss': 0.6033962368965149, 'train/grad_norm_pre_clip': 14.303393363952637, 'timing/data': 0.0046705229906365275, 'timing/model': 1.2089507470373064, 'learning_rate/qwen_vl_interface': 5.314793633981169e-07, 'learning_rate/action_model': 5.314793633981169e-06, 'epoch': 1.99}) +[2026-06-17 18:51:57,301][starVLA.training.train_starvla][INFO] - Step 6748, Loss: {'train/loss': 0.3962673842906952, 'train/grad_norm_pre_clip': 10.737781524658203, 'timing/data': 0.004692542017437518, 'timing/model': 1.2062748920870945, 'learning_rate/qwen_vl_interface': 5.312312789895369e-07, 'learning_rate/action_model': 5.3123127898953685e-06, 'epoch': 1.99}) +[2026-06-17 18:52:02,105][starVLA.training.train_starvla][INFO] - Step 6749, Loss: {'train/loss': 0.6218805313110352, 'train/grad_norm_pre_clip': 14.36215591430664, 'timing/data': 0.004505272023379803, 'timing/model': 1.344593318994157, 'learning_rate/qwen_vl_interface': 5.3098417278679e-07, 'learning_rate/action_model': 5.3098417278679e-06, 'epoch': 1.99}) +[2026-06-17 18:52:06,772][starVLA.training.train_starvla][INFO] - Step 6750, Loss: {'train/loss': 0.6298826336860657, 'train/grad_norm_pre_clip': 18.90228271484375, 'timing/data': 0.004618567996658385, 'timing/model': 1.208749140962027, 'learning_rate/qwen_vl_interface': 5.307380448411022e-07, 'learning_rate/action_model': 5.307380448411022e-06, 'epoch': 1.99}) +[2026-06-17 18:52:11,554][starVLA.training.train_starvla][INFO] - Step 6751, Loss: {'train/loss': 0.7064704895019531, 'train/grad_norm_pre_clip': 14.878805160522461, 'timing/data': 0.005109596997499466, 'timing/model': 1.2032854120479897, 'learning_rate/qwen_vl_interface': 5.304928952034955e-07, 'learning_rate/action_model': 5.304928952034956e-06, 'epoch': 1.99}) +[2026-06-17 18:52:16,337][starVLA.training.train_starvla][INFO] - Step 6752, Loss: {'train/loss': 0.6771215796470642, 'train/grad_norm_pre_clip': 14.718940734863281, 'timing/data': 0.004726685001514852, 'timing/model': 1.2061694050207734, 'learning_rate/qwen_vl_interface': 5.302487239247905e-07, 'learning_rate/action_model': 5.302487239247904e-06, 'epoch': 1.99}) +[2026-06-17 18:52:21,146][starVLA.training.train_starvla][INFO] - Step 6753, Loss: {'train/loss': 0.5466206669807434, 'train/grad_norm_pre_clip': 13.303327560424805, 'timing/data': 0.004748133011162281, 'timing/model': 1.2046737089985982, 'learning_rate/qwen_vl_interface': 5.300055310556034e-07, 'learning_rate/action_model': 5.300055310556034e-06, 'epoch': 1.99}) +[2026-06-17 18:52:25,957][starVLA.training.train_starvla][INFO] - Step 6754, Loss: {'train/loss': 0.46619415283203125, 'train/grad_norm_pre_clip': 14.003767013549805, 'timing/data': 0.004764078999869525, 'timing/model': 1.206495571997948, 'learning_rate/qwen_vl_interface': 5.297633166463482e-07, 'learning_rate/action_model': 5.297633166463481e-06, 'epoch': 1.99}) +[2026-06-17 18:52:30,744][starVLA.training.train_starvla][INFO] - Step 6755, Loss: {'train/loss': 0.717766523361206, 'train/grad_norm_pre_clip': 17.422550201416016, 'timing/data': 0.004795002052560449, 'timing/model': 1.332792375003919, 'learning_rate/qwen_vl_interface': 5.295220807472373e-07, 'learning_rate/action_model': 5.295220807472374e-06, 'epoch': 1.99}) +[2026-06-17 18:52:35,480][starVLA.training.train_starvla][INFO] - Step 6756, Loss: {'train/loss': 0.504100501537323, 'train/grad_norm_pre_clip': 12.509446144104004, 'timing/data': 0.004640757921151817, 'timing/model': 1.2838571219472215, 'learning_rate/qwen_vl_interface': 5.292818234082786e-07, 'learning_rate/action_model': 5.292818234082787e-06, 'epoch': 1.99}) +[2026-06-17 18:52:40,143][starVLA.training.train_starvla][INFO] - Step 6757, Loss: {'train/loss': 0.7482876777648926, 'train/grad_norm_pre_clip': 14.877286911010742, 'timing/data': 0.004621876985765994, 'timing/model': 1.206681192968972, 'learning_rate/qwen_vl_interface': 5.290425446792772e-07, 'learning_rate/action_model': 5.2904254467927716e-06, 'epoch': 1.99}) +[2026-06-17 18:52:44,960][starVLA.training.train_starvla][INFO] - Step 6758, Loss: {'train/loss': 0.7278808951377869, 'train/grad_norm_pre_clip': 16.641202926635742, 'timing/data': 0.00474654498975724, 'timing/model': 1.207365581067279, 'learning_rate/qwen_vl_interface': 5.288042446098364e-07, 'learning_rate/action_model': 5.2880424460983634e-06, 'epoch': 1.99}) +[2026-06-17 18:52:49,744][starVLA.training.train_starvla][INFO] - Step 6759, Loss: {'train/loss': 0.6265307068824768, 'train/grad_norm_pre_clip': 15.371223449707031, 'timing/data': 0.004725020960904658, 'timing/model': 1.2034413380315527, 'learning_rate/qwen_vl_interface': 5.28566923249356e-07, 'learning_rate/action_model': 5.28566923249356e-06, 'epoch': 1.99}) +[2026-06-17 18:52:54,534][starVLA.training.train_starvla][INFO] - Step 6760, Loss: {'train/loss': 0.5260076522827148, 'train/grad_norm_pre_clip': 12.829238891601562, 'timing/data': 0.004783881944604218, 'timing/model': 1.2046655470039696, 'learning_rate/qwen_vl_interface': 5.283305806470329e-07, 'learning_rate/action_model': 5.2833058064703286e-06, 'epoch': 1.99}) +[2026-06-17 18:52:59,355][starVLA.training.train_starvla][INFO] - Step 6761, Loss: {'train/loss': 0.6245359778404236, 'train/grad_norm_pre_clip': 13.330755233764648, 'timing/data': 0.004646267043426633, 'timing/model': 1.2033675550483167, 'learning_rate/qwen_vl_interface': 5.280952168518612e-07, 'learning_rate/action_model': 5.280952168518612e-06, 'epoch': 1.99}) +[2026-06-17 18:53:04,151][starVLA.training.train_starvla][INFO] - Step 6762, Loss: {'train/loss': 0.49658963084220886, 'train/grad_norm_pre_clip': 13.085664749145508, 'timing/data': 0.004910941002890468, 'timing/model': 1.2982133789919317, 'learning_rate/qwen_vl_interface': 5.278608319126318e-07, 'learning_rate/action_model': 5.278608319126318e-06, 'epoch': 1.99}) +[2026-06-17 18:53:08,817][starVLA.training.train_starvla][INFO] - Step 6763, Loss: {'train/loss': 0.6939995884895325, 'train/grad_norm_pre_clip': 37.91031265258789, 'timing/data': 0.004733137087896466, 'timing/model': 1.2074734220514074, 'learning_rate/qwen_vl_interface': 5.276274258779332e-07, 'learning_rate/action_model': 5.2762742587793315e-06, 'epoch': 1.99}) +[2026-06-17 18:53:13,614][starVLA.training.train_starvla][INFO] - Step 6764, Loss: {'train/loss': 0.5920619964599609, 'train/grad_norm_pre_clip': 14.310891151428223, 'timing/data': 0.0046635339967906475, 'timing/model': 1.209549893043004, 'learning_rate/qwen_vl_interface': 5.273949987961508e-07, 'learning_rate/action_model': 5.273949987961508e-06, 'epoch': 1.99}) +[2026-06-17 18:53:18,407][starVLA.training.train_starvla][INFO] - Step 6765, Loss: {'train/loss': 0.47365859150886536, 'train/grad_norm_pre_clip': 10.91391372680664, 'timing/data': 0.004815420019440353, 'timing/model': 1.205488782026805, 'learning_rate/qwen_vl_interface': 5.271635507154671e-07, 'learning_rate/action_model': 5.2716355071546706e-06, 'epoch': 1.99}) +[2026-06-17 18:53:23,197][starVLA.training.train_starvla][INFO] - Step 6766, Loss: {'train/loss': 0.5680119395256042, 'train/grad_norm_pre_clip': 43.78343200683594, 'timing/data': 0.004516025073826313, 'timing/model': 1.2095576280262321, 'learning_rate/qwen_vl_interface': 5.269330816838614e-07, 'learning_rate/action_model': 5.269330816838613e-06, 'epoch': 1.99}) +[2026-06-17 18:53:27,987][starVLA.training.train_starvla][INFO] - Step 6767, Loss: {'train/loss': 0.7230343222618103, 'train/grad_norm_pre_clip': 14.13325309753418, 'timing/data': 0.004834683961234987, 'timing/model': 1.3395185059634969, 'learning_rate/qwen_vl_interface': 5.267035917491095e-07, 'learning_rate/action_model': 5.267035917491094e-06, 'epoch': 1.99}) +[2026-06-17 18:53:32,648][starVLA.training.train_starvla][INFO] - Step 6768, Loss: {'train/loss': 0.8850907683372498, 'train/grad_norm_pre_clip': 18.948657989501953, 'timing/data': 0.004821783979423344, 'timing/model': 1.2060025130631402, 'learning_rate/qwen_vl_interface': 5.264750809587862e-07, 'learning_rate/action_model': 5.264750809587863e-06, 'epoch': 1.99}) +[2026-06-17 18:53:37,435][starVLA.training.train_starvla][INFO] - Step 6769, Loss: {'train/loss': 0.5697622299194336, 'train/grad_norm_pre_clip': 12.166898727416992, 'timing/data': 0.004445836995728314, 'timing/model': 1.2076281969202682, 'learning_rate/qwen_vl_interface': 5.262475493602609e-07, 'learning_rate/action_model': 5.26247549360261e-06, 'epoch': 1.99}) +[2026-06-17 18:53:42,216][starVLA.training.train_starvla][INFO] - Step 6770, Loss: {'train/loss': 0.6265691518783569, 'train/grad_norm_pre_clip': 14.53411865234375, 'timing/data': 0.005009173066355288, 'timing/model': 1.2029519350035116, 'learning_rate/qwen_vl_interface': 5.260209970007018e-07, 'learning_rate/action_model': 5.260209970007018e-06, 'epoch': 1.99}) +[2026-06-17 18:53:46,933][starVLA.training.train_starvla][INFO] - Step 6771, Loss: {'train/loss': 0.6189712882041931, 'train/grad_norm_pre_clip': 12.569748878479004, 'timing/data': 0.004544202005490661, 'timing/model': 1.2053162510273978, 'learning_rate/qwen_vl_interface': 5.257954239270732e-07, 'learning_rate/action_model': 5.257954239270732e-06, 'epoch': 1.99}) +[2026-06-17 18:53:51,672][starVLA.training.train_starvla][INFO] - Step 6772, Loss: {'train/loss': 0.5697700381278992, 'train/grad_norm_pre_clip': 14.112785339355469, 'timing/data': 0.0050309409853070974, 'timing/model': 1.2065956570440903, 'learning_rate/qwen_vl_interface': 5.255708301861367e-07, 'learning_rate/action_model': 5.255708301861367e-06, 'epoch': 1.99}) +[2026-06-17 18:53:56,465][starVLA.training.train_starvla][INFO] - Step 6773, Loss: {'train/loss': 0.4934673011302948, 'train/grad_norm_pre_clip': 13.34926986694336, 'timing/data': 0.004824885982088745, 'timing/model': 1.3399447289993986, 'learning_rate/qwen_vl_interface': 5.253472158244506e-07, 'learning_rate/action_model': 5.253472158244505e-06, 'epoch': 1.99}) +[2026-06-17 18:54:01,124][starVLA.training.train_starvla][INFO] - Step 6774, Loss: {'train/loss': 0.9000129699707031, 'train/grad_norm_pre_clip': 16.88976287841797, 'timing/data': 0.006801521056331694, 'timing/model': 1.2046580360038206, 'learning_rate/qwen_vl_interface': 5.251245808883707e-07, 'learning_rate/action_model': 5.251245808883707e-06, 'epoch': 1.99}) +[2026-06-17 18:54:05,909][starVLA.training.train_starvla][INFO] - Step 6775, Loss: {'train/loss': 0.7496129870414734, 'train/grad_norm_pre_clip': 20.222780227661133, 'timing/data': 0.005015849019400775, 'timing/model': 1.2076226270291954, 'learning_rate/qwen_vl_interface': 5.249029254240495e-07, 'learning_rate/action_model': 5.249029254240495e-06, 'epoch': 1.99}) +[2026-06-17 18:54:10,758][starVLA.training.train_starvla][INFO] - Step 6776, Loss: {'train/loss': 0.5785238146781921, 'train/grad_norm_pre_clip': 40.52397537231445, 'timing/data': 0.005040422081947327, 'timing/model': 1.2054930339800194, 'learning_rate/qwen_vl_interface': 5.246822494774364e-07, 'learning_rate/action_model': 5.246822494774363e-06, 'epoch': 1.99}) +[2026-06-17 18:54:15,534][starVLA.training.train_starvla][INFO] - Step 6777, Loss: {'train/loss': 0.5809481739997864, 'train/grad_norm_pre_clip': 14.271347999572754, 'timing/data': 0.004484385019168258, 'timing/model': 1.2089889229973778, 'learning_rate/qwen_vl_interface': 5.244625530942774e-07, 'learning_rate/action_model': 5.244625530942774e-06, 'epoch': 1.99}) +[2026-06-17 18:54:20,274][starVLA.training.train_starvla][INFO] - Step 6778, Loss: {'train/loss': 0.6365521550178528, 'train/grad_norm_pre_clip': 18.366174697875977, 'timing/data': 0.004706036997959018, 'timing/model': 1.2042039589723572, 'learning_rate/qwen_vl_interface': 5.242438363201161e-07, 'learning_rate/action_model': 5.2424383632011605e-06, 'epoch': 1.99}) +[2026-06-17 18:54:25,054][starVLA.training.train_starvla][INFO] - Step 6779, Loss: {'train/loss': 0.44873660802841187, 'train/grad_norm_pre_clip': 10.483020782470703, 'timing/data': 0.004801109898835421, 'timing/model': 1.3297522239154205, 'learning_rate/qwen_vl_interface': 5.240260992002928e-07, 'learning_rate/action_model': 5.240260992002928e-06, 'epoch': 1.99}) +[2026-06-17 18:54:29,730][starVLA.training.train_starvla][INFO] - Step 6780, Loss: {'train/loss': 0.9354004859924316, 'train/grad_norm_pre_clip': 20.142820358276367, 'timing/data': 0.004904928966425359, 'timing/model': 1.2070862499531358, 'learning_rate/qwen_vl_interface': 5.238093417799446e-07, 'learning_rate/action_model': 5.238093417799446e-06, 'epoch': 2.0}) +[2026-06-17 18:54:34,529][starVLA.training.train_starvla][INFO] - Step 6781, Loss: {'train/loss': 0.6668227910995483, 'train/grad_norm_pre_clip': 14.567206382751465, 'timing/data': 0.004759188974276185, 'timing/model': 1.2041832340182737, 'learning_rate/qwen_vl_interface': 5.235935641040048e-07, 'learning_rate/action_model': 5.235935641040048e-06, 'epoch': 2.0}) +[2026-06-17 18:54:39,329][starVLA.training.train_starvla][INFO] - Step 6782, Loss: {'train/loss': 0.505730926990509, 'train/grad_norm_pre_clip': 12.793329238891602, 'timing/data': 0.004539931076578796, 'timing/model': 1.212419643998146, 'learning_rate/qwen_vl_interface': 5.233787662172059e-07, 'learning_rate/action_model': 5.23378766217206e-06, 'epoch': 2.0}) +[2026-06-17 18:54:44,149][starVLA.training.train_starvla][INFO] - Step 6783, Loss: {'train/loss': 0.3863348960876465, 'train/grad_norm_pre_clip': 9.36138916015625, 'timing/data': 0.004484800039790571, 'timing/model': 1.2097177889663726, 'learning_rate/qwen_vl_interface': 5.231649481640741e-07, 'learning_rate/action_model': 5.231649481640741e-06, 'epoch': 2.0}) +[2026-06-17 18:54:48,949][starVLA.training.train_starvla][INFO] - Step 6784, Loss: {'train/loss': 0.7335745096206665, 'train/grad_norm_pre_clip': 15.009838104248047, 'timing/data': 0.00467152101919055, 'timing/model': 1.339358900906518, 'learning_rate/qwen_vl_interface': 5.229521099889351e-07, 'learning_rate/action_model': 5.229521099889352e-06, 'epoch': 2.0}) +[2026-06-17 18:54:53,619][starVLA.training.train_starvla][INFO] - Step 6785, Loss: {'train/loss': 0.4787693917751312, 'train/grad_norm_pre_clip': 11.247543334960938, 'timing/data': 0.004798311972990632, 'timing/model': 1.2127785490592942, 'learning_rate/qwen_vl_interface': 5.227402517359105e-07, 'learning_rate/action_model': 5.2274025173591045e-06, 'epoch': 2.0}) +[2026-06-17 18:54:58,407][starVLA.training.train_starvla][INFO] - Step 6786, Loss: {'train/loss': 0.629995584487915, 'train/grad_norm_pre_clip': 13.573001861572266, 'timing/data': 0.004579773056320846, 'timing/model': 1.2025959560414776, 'learning_rate/qwen_vl_interface': 5.225293734489182e-07, 'learning_rate/action_model': 5.225293734489181e-06, 'epoch': 2.0}) +[2026-06-17 18:55:03,113][starVLA.training.train_starvla][INFO] - Step 6787, Loss: {'train/loss': 0.5213002562522888, 'train/grad_norm_pre_clip': 12.572549819946289, 'timing/data': 0.0047678640112280846, 'timing/model': 1.2099680399987847, 'learning_rate/qwen_vl_interface': 5.223194751716733e-07, 'learning_rate/action_model': 5.223194751716733e-06, 'epoch': 2.0}) +[2026-06-17 18:55:07,892][starVLA.training.train_starvla][INFO] - Step 6788, Loss: {'train/loss': 0.5841256976127625, 'train/grad_norm_pre_clip': 13.082404136657715, 'timing/data': 0.004686745000071824, 'timing/model': 1.2053632189054042, 'learning_rate/qwen_vl_interface': 5.221105569476893e-07, 'learning_rate/action_model': 5.221105569476892e-06, 'epoch': 2.0}) +[2026-06-17 18:55:12,703][starVLA.training.train_starvla][INFO] - Step 6789, Loss: {'train/loss': 0.44391003251075745, 'train/grad_norm_pre_clip': 13.481033325195312, 'timing/data': 0.004844672977924347, 'timing/model': 1.205768223037012, 'learning_rate/qwen_vl_interface': 5.21902618820274e-07, 'learning_rate/action_model': 5.21902618820274e-06, 'epoch': 2.0}) +[2026-06-17 18:55:17,559][starVLA.training.train_starvla][INFO] - Step 6790, Loss: {'train/loss': 0.9038663506507874, 'train/grad_norm_pre_clip': 16.323413848876953, 'timing/data': 0.004882336012087762, 'timing/model': 1.341542091104202, 'learning_rate/qwen_vl_interface': 5.216956608325331e-07, 'learning_rate/action_model': 5.2169566083253304e-06, 'epoch': 2.0}) +[2026-06-17 18:55:22,249][starVLA.training.train_starvla][INFO] - Step 6791, Loss: {'train/loss': 0.6210605502128601, 'train/grad_norm_pre_clip': 11.296626091003418, 'timing/data': 0.004971889080479741, 'timing/model': 1.2109671429498121, 'learning_rate/qwen_vl_interface': 5.2148968302737e-07, 'learning_rate/action_model': 5.2148968302737005e-06, 'epoch': 2.0}) +[2026-06-17 18:55:27,044][starVLA.training.train_starvla][INFO] - Step 6792, Loss: {'train/loss': 0.5662951469421387, 'train/grad_norm_pre_clip': 11.830961227416992, 'timing/data': 0.0046971539268270135, 'timing/model': 1.2114469630178064, 'learning_rate/qwen_vl_interface': 5.212846854474834e-07, 'learning_rate/action_model': 5.212846854474833e-06, 'epoch': 2.0}) +[2026-06-17 18:55:31,858][starVLA.training.train_starvla][INFO] - Step 6793, Loss: {'train/loss': 0.444580614566803, 'train/grad_norm_pre_clip': 11.047844886779785, 'timing/data': 0.004566082963719964, 'timing/model': 1.2062480020103976, 'learning_rate/qwen_vl_interface': 5.2108066813537e-07, 'learning_rate/action_model': 5.2108066813537e-06, 'epoch': 2.0}) +[2026-06-17 18:55:36,615][starVLA.training.train_starvla][INFO] - Step 6794, Loss: {'train/loss': 0.8831372261047363, 'train/grad_norm_pre_clip': 17.7171630859375, 'timing/data': 0.004673109040595591, 'timing/model': 1.205694074044004, 'learning_rate/qwen_vl_interface': 5.208776311333223e-07, 'learning_rate/action_model': 5.208776311333223e-06, 'epoch': 2.0}) +[2026-06-17 18:55:41,392][starVLA.training.train_starvla][INFO] - Step 6795, Loss: {'train/loss': 0.451561838388443, 'train/grad_norm_pre_clip': 11.555431365966797, 'timing/data': 0.004864217946305871, 'timing/model': 1.197544916998595, 'learning_rate/qwen_vl_interface': 5.206755744834312e-07, 'learning_rate/action_model': 5.2067557448343125e-06, 'epoch': 2.0}) +[2026-06-17 18:55:46,134][starVLA.training.train_starvla][INFO] - Step 6796, Loss: {'train/loss': 0.5912018418312073, 'train/grad_norm_pre_clip': 14.066954612731934, 'timing/data': 0.004169697989709675, 'timing/model': 1.317111476091668, 'learning_rate/qwen_vl_interface': 5.20474498227582e-07, 'learning_rate/action_model': 5.2047449822758196e-06, 'epoch': 2.0}) +[2026-06-17 18:55:50,469][starVLA.training.train_starvla][INFO] - Step 6797, Loss: {'train/loss': 0.776970624923706, 'train/grad_norm_pre_clip': 17.90052604675293, 'timing/data': 0.0002895171055570245, 'timing/model': 0.9174790279939771, 'learning_rate/qwen_vl_interface': 5.202744024074583e-07, 'learning_rate/action_model': 5.202744024074584e-06, 'epoch': 2.0}) +[2026-06-17 18:55:55,538][starVLA.training.train_starvla][INFO] - Step 6798, Loss: {'train/loss': 0.39179128408432007, 'train/grad_norm_pre_clip': 10.908930778503418, 'timing/data': 0.0051819729851558805, 'timing/model': 1.20792039798107, 'learning_rate/qwen_vl_interface': 5.200752870645407e-07, 'learning_rate/action_model': 5.200752870645406e-06, 'epoch': 2.0}) +[2026-06-17 18:56:00,296][starVLA.training.train_starvla][INFO] - Step 6799, Loss: {'train/loss': 0.5577176213264465, 'train/grad_norm_pre_clip': 13.136539459228516, 'timing/data': 0.004646960063837469, 'timing/model': 1.2072392890695482, 'learning_rate/qwen_vl_interface': 5.198771522401054e-07, 'learning_rate/action_model': 5.198771522401054e-06, 'epoch': 2.0}) +[2026-06-17 18:57:42,154][starVLA.training.train_starvla][INFO] - Step 6800, Loss: {'train/loss': 0.4154024124145508, 'train/grad_norm_pre_clip': 12.240694999694824, 'eval/loss': 0.7002893686294556, 'eval/latency_0/loss': 0.7002893686294556, 'eval/demon_fire_precision': 0.9019138755980861, 'eval/demon_fire_recall': 0.9393687707641196, 'eval/demon_fire_f1': 0.9202603742880391, 'eval/demon_fire_rate_model': 0.6974416017797553, 'eval/demon_fire_rate_teacher': 0.6696329254727474, 'eval/demon_left_precision': 0.830848623853211, 'eval/demon_left_recall': 0.8209631728045326, 'eval/demon_left_f1': 0.8258763180393274, 'eval/demon_left_rate_model': 0.4849833147942158, 'eval/demon_left_rate_teacher': 0.49082313681868744, 'eval/demon_right_precision': 0.8266593527153044, 'eval/demon_right_recall': 0.8514124293785311, 'eval/demon_right_f1': 0.8388533259114945, 'eval/demon_right_rate_model': 0.506952169076752, 'eval/demon_right_rate_teacher': 0.4922135706340378, 'eval/demon_move_f1': 0.832364821975411, 'eval/demon_cc_f1': 0.8920024356967475, 'eval/latency_0/demon_fire_precision': 0.9019138755980861, 'eval/latency_0/demon_fire_recall': 0.9393687707641196, 'eval/latency_0/demon_fire_f1': 0.9202603742880391, 'eval/latency_0/demon_fire_rate_model': 0.6974416017797553, 'eval/latency_0/demon_fire_rate_teacher': 0.6696329254727474, 'eval/latency_0/demon_left_precision': 0.830848623853211, 'eval/latency_0/demon_left_recall': 0.8209631728045326, 'eval/latency_0/demon_left_f1': 0.8258763180393274, 'eval/latency_0/demon_left_rate_model': 0.4849833147942158, 'eval/latency_0/demon_left_rate_teacher': 0.49082313681868744, 'eval/latency_0/demon_right_precision': 0.8266593527153044, 'eval/latency_0/demon_right_recall': 0.8514124293785311, 'eval/latency_0/demon_right_f1': 0.8388533259114945, 'eval/latency_0/demon_right_rate_model': 0.506952169076752, 'eval/latency_0/demon_right_rate_teacher': 0.4922135706340378, 'eval/latency_0/demon_move_f1': 0.832364821975411, 'eval/latency_0/demon_cc_f1': 0.8920024356967475, 'timing/data': 0.004855632083490491, 'timing/model': 1.2049828600138426, 'learning_rate/qwen_vl_interface': 5.196799979752265e-07, 'learning_rate/action_model': 5.196799979752264e-06, 'epoch': 2.0}) +[2026-06-17 18:57:42,154][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_6800_state +[2026-06-17 18:57:42,157][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 18:57:45,711][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_6800_state/model.safetensors +[2026-06-17 18:57:53,377][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_6800_state/optimizer.bin +[2026-06-17 18:57:53,377][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_6800_state/sampler.bin +[2026-06-17 18:57:53,377][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_6800_state/sampler_1.bin +[2026-06-17 18:57:53,378][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_6800_state/random_states_0.pkl +[2026-06-17 18:57:53,422][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 6800 +[2026-06-17 18:57:53,422][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 18:57:53,432][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 18:57:58,137][starVLA.training.train_starvla][INFO] - Step 6801, Loss: {'train/loss': 0.5207133889198303, 'train/grad_norm_pre_clip': 12.612643241882324, 'timing/data': 0.005273782997392118, 'timing/model': 1.185309965047054, 'learning_rate/qwen_vl_interface': 5.194838243107742e-07, 'learning_rate/action_model': 5.194838243107742e-06, 'epoch': 2.0}) +[2026-06-17 18:58:02,818][starVLA.training.train_starvla][INFO] - Step 6802, Loss: {'train/loss': 0.8786982297897339, 'train/grad_norm_pre_clip': 16.50581932067871, 'timing/data': 0.004716990049928427, 'timing/model': 1.1869317420059815, 'learning_rate/qwen_vl_interface': 5.192886312874148e-07, 'learning_rate/action_model': 5.1928863128741475e-06, 'epoch': 2.0}) +[2026-06-17 18:58:07,550][starVLA.training.train_starvla][INFO] - Step 6803, Loss: {'train/loss': 0.6469906568527222, 'train/grad_norm_pre_clip': 15.198184967041016, 'timing/data': 0.004248270997777581, 'timing/model': 1.3254976370371878, 'learning_rate/qwen_vl_interface': 5.190944189456133e-07, 'learning_rate/action_model': 5.190944189456132e-06, 'epoch': 2.0}) +[2026-06-17 18:58:12,176][starVLA.training.train_starvla][INFO] - Step 6804, Loss: {'train/loss': 0.519432783126831, 'train/grad_norm_pre_clip': 14.311708450317383, 'timing/data': 0.005082081072032452, 'timing/model': 1.1913730209926143, 'learning_rate/qwen_vl_interface': 5.189011873256295e-07, 'learning_rate/action_model': 5.189011873256295e-06, 'epoch': 2.0}) +[2026-06-17 18:58:16,933][starVLA.training.train_starvla][INFO] - Step 6805, Loss: {'train/loss': 0.6295355558395386, 'train/grad_norm_pre_clip': 14.767539978027344, 'timing/data': 0.004701182944700122, 'timing/model': 1.1976587290409952, 'learning_rate/qwen_vl_interface': 5.187089364675195e-07, 'learning_rate/action_model': 5.187089364675194e-06, 'epoch': 2.0}) +[2026-06-17 18:58:21,693][starVLA.training.train_starvla][INFO] - Step 6806, Loss: {'train/loss': 0.7018497586250305, 'train/grad_norm_pre_clip': 15.194717407226562, 'timing/data': 0.005490768933668733, 'timing/model': 1.19635676802136, 'learning_rate/qwen_vl_interface': 5.18517666411139e-07, 'learning_rate/action_model': 5.18517666411139e-06, 'epoch': 2.0}) +[2026-06-17 18:58:26,405][starVLA.training.train_starvla][INFO] - Step 6807, Loss: {'train/loss': 0.5779810547828674, 'train/grad_norm_pre_clip': 14.329285621643066, 'timing/data': 0.004917755955830216, 'timing/model': 1.1995183089748025, 'learning_rate/qwen_vl_interface': 5.18327377196137e-07, 'learning_rate/action_model': 5.18327377196137e-06, 'epoch': 2.0}) +[2026-06-17 18:58:31,172][starVLA.training.train_starvla][INFO] - Step 6808, Loss: {'train/loss': 0.7435592412948608, 'train/grad_norm_pre_clip': 13.670273780822754, 'timing/data': 0.004966713022440672, 'timing/model': 1.2006327810231596, 'learning_rate/qwen_vl_interface': 5.181380688619617e-07, 'learning_rate/action_model': 5.181380688619617e-06, 'epoch': 2.0}) +[2026-06-17 18:58:35,939][starVLA.training.train_starvla][INFO] - Step 6809, Loss: {'train/loss': 0.5298435091972351, 'train/grad_norm_pre_clip': 13.251983642578125, 'timing/data': 0.004844500916078687, 'timing/model': 1.3336375040234998, 'learning_rate/qwen_vl_interface': 5.179497414478561e-07, 'learning_rate/action_model': 5.179497414478562e-06, 'epoch': 2.0}) +[2026-06-17 18:58:40,575][starVLA.training.train_starvla][INFO] - Step 6810, Loss: {'train/loss': 0.6563454270362854, 'train/grad_norm_pre_clip': 13.488856315612793, 'timing/data': 0.005554516101256013, 'timing/model': 1.1995346279582009, 'learning_rate/qwen_vl_interface': 5.177623949928608e-07, 'learning_rate/action_model': 5.177623949928608e-06, 'epoch': 2.0}) +[2026-06-17 18:58:45,338][starVLA.training.train_starvla][INFO] - Step 6811, Loss: {'train/loss': 0.9079345464706421, 'train/grad_norm_pre_clip': 19.654760360717773, 'timing/data': 0.004419501055963337, 'timing/model': 1.200221767066978, 'learning_rate/qwen_vl_interface': 5.175760295358137e-07, 'learning_rate/action_model': 5.175760295358136e-06, 'epoch': 2.0}) +[2026-06-17 18:58:50,117][starVLA.training.train_starvla][INFO] - Step 6812, Loss: {'train/loss': 0.4723006784915924, 'train/grad_norm_pre_clip': 12.106086730957031, 'timing/data': 0.004785900004208088, 'timing/model': 1.1986528020352125, 'learning_rate/qwen_vl_interface': 5.173906451153477e-07, 'learning_rate/action_model': 5.173906451153477e-06, 'epoch': 2.0}) +[2026-06-17 18:58:54,856][starVLA.training.train_starvla][INFO] - Step 6813, Loss: {'train/loss': 0.7130219340324402, 'train/grad_norm_pre_clip': 16.54513168334961, 'timing/data': 0.005416319007053971, 'timing/model': 1.2012418640078977, 'learning_rate/qwen_vl_interface': 5.172062417698931e-07, 'learning_rate/action_model': 5.1720624176989304e-06, 'epoch': 2.0}) +[2026-06-17 18:58:59,621][starVLA.training.train_starvla][INFO] - Step 6814, Loss: {'train/loss': 0.36437708139419556, 'train/grad_norm_pre_clip': 8.252424240112305, 'timing/data': 0.0048567489720880985, 'timing/model': 1.2026724979514256, 'learning_rate/qwen_vl_interface': 5.170228195376778e-07, 'learning_rate/action_model': 5.170228195376778e-06, 'epoch': 2.01}) +[2026-06-17 18:59:04,406][starVLA.training.train_starvla][INFO] - Step 6815, Loss: {'train/loss': 0.630753755569458, 'train/grad_norm_pre_clip': 15.265031814575195, 'timing/data': 0.005565761937759817, 'timing/model': 1.3348168020602316, 'learning_rate/qwen_vl_interface': 5.168403784567247e-07, 'learning_rate/action_model': 5.1684037845672475e-06, 'epoch': 2.01}) +[2026-06-17 18:59:09,065][starVLA.training.train_starvla][INFO] - Step 6816, Loss: {'train/loss': 0.8834552764892578, 'train/grad_norm_pre_clip': 14.976678848266602, 'timing/data': 0.004894157988019288, 'timing/model': 1.215051066945307, 'learning_rate/qwen_vl_interface': 5.166589185648543e-07, 'learning_rate/action_model': 5.1665891856485425e-06, 'epoch': 2.01}) +[2026-06-17 18:59:13,846][starVLA.training.train_starvla][INFO] - Step 6817, Loss: {'train/loss': 0.28239789605140686, 'train/grad_norm_pre_clip': 8.977882385253906, 'timing/data': 0.004635686986148357, 'timing/model': 1.2021317140897736, 'learning_rate/qwen_vl_interface': 5.164784398996836e-07, 'learning_rate/action_model': 5.164784398996836e-06, 'epoch': 2.01}) +[2026-06-17 18:59:18,633][starVLA.training.train_starvla][INFO] - Step 6818, Loss: {'train/loss': 0.6951032280921936, 'train/grad_norm_pre_clip': 14.885889053344727, 'timing/data': 0.004847460077144206, 'timing/model': 1.2065378640545532, 'learning_rate/qwen_vl_interface': 5.162989424986253e-07, 'learning_rate/action_model': 5.162989424986253e-06, 'epoch': 2.01}) +[2026-06-17 18:59:23,402][starVLA.training.train_starvla][INFO] - Step 6819, Loss: {'train/loss': 0.5065474510192871, 'train/grad_norm_pre_clip': 11.783760070800781, 'timing/data': 0.004968310939148068, 'timing/model': 1.201725721010007, 'learning_rate/qwen_vl_interface': 5.161204263988904e-07, 'learning_rate/action_model': 5.161204263988904e-06, 'epoch': 2.01}) +[2026-06-17 18:59:28,083][starVLA.training.train_starvla][INFO] - Step 6820, Loss: {'train/loss': 0.6441246867179871, 'train/grad_norm_pre_clip': 15.622054100036621, 'timing/data': 0.0046100650215521455, 'timing/model': 1.2134382999502122, 'learning_rate/qwen_vl_interface': 5.159428916374845e-07, 'learning_rate/action_model': 5.159428916374845e-06, 'epoch': 2.01}) +[2026-06-17 18:59:32,771][starVLA.training.train_starvla][INFO] - Step 6821, Loss: {'train/loss': 0.7005102634429932, 'train/grad_norm_pre_clip': 14.60150146484375, 'timing/data': 0.005253967945463955, 'timing/model': 1.2443405860103667, 'learning_rate/qwen_vl_interface': 5.157663382512118e-07, 'learning_rate/action_model': 5.157663382512119e-06, 'epoch': 2.01}) +[2026-06-17 18:59:37,445][starVLA.training.train_starvla][INFO] - Step 6822, Loss: {'train/loss': 1.0081915855407715, 'train/grad_norm_pre_clip': 16.123849868774414, 'timing/data': 0.004816776025108993, 'timing/model': 1.2049104149919003, 'learning_rate/qwen_vl_interface': 5.155907662766711e-07, 'learning_rate/action_model': 5.155907662766711e-06, 'epoch': 2.01}) +[2026-06-17 18:59:42,186][starVLA.training.train_starvla][INFO] - Step 6823, Loss: {'train/loss': 0.6611117720603943, 'train/grad_norm_pre_clip': 14.582233428955078, 'timing/data': 0.005060640047304332, 'timing/model': 1.2104401959804818, 'learning_rate/qwen_vl_interface': 5.15416175750259e-07, 'learning_rate/action_model': 5.154161757502589e-06, 'epoch': 2.01}) +[2026-06-17 18:59:46,968][starVLA.training.train_starvla][INFO] - Step 6824, Loss: {'train/loss': 0.604322612285614, 'train/grad_norm_pre_clip': 13.730362892150879, 'timing/data': 0.005120720015838742, 'timing/model': 1.2056165990652516, 'learning_rate/qwen_vl_interface': 5.15242566708168e-07, 'learning_rate/action_model': 5.15242566708168e-06, 'epoch': 2.01}) +[2026-06-17 18:59:51,749][starVLA.training.train_starvla][INFO] - Step 6825, Loss: {'train/loss': 0.4459437429904938, 'train/grad_norm_pre_clip': 12.764057159423828, 'timing/data': 0.004726651939563453, 'timing/model': 1.2033079930115491, 'learning_rate/qwen_vl_interface': 5.150699391863884e-07, 'learning_rate/action_model': 5.150699391863884e-06, 'epoch': 2.01}) +[2026-06-17 18:59:56,547][starVLA.training.train_starvla][INFO] - Step 6826, Loss: {'train/loss': 0.706718385219574, 'train/grad_norm_pre_clip': 15.450934410095215, 'timing/data': 0.004811177961528301, 'timing/model': 1.3365353250410408, 'learning_rate/qwen_vl_interface': 5.148982932207052e-07, 'learning_rate/action_model': 5.148982932207052e-06, 'epoch': 2.01}) +[2026-06-17 19:00:01,321][starVLA.training.train_starvla][INFO] - Step 6827, Loss: {'train/loss': 0.5941404700279236, 'train/grad_norm_pre_clip': 13.598103523254395, 'timing/data': 0.004934558994136751, 'timing/model': 1.3309410080546513, 'learning_rate/qwen_vl_interface': 5.147276288467008e-07, 'learning_rate/action_model': 5.147276288467008e-06, 'epoch': 2.01}) +[2026-06-17 19:00:05,978][starVLA.training.train_starvla][INFO] - Step 6828, Loss: {'train/loss': 0.415606826543808, 'train/grad_norm_pre_clip': 10.98042106628418, 'timing/data': 0.004918012069538236, 'timing/model': 1.205254698987119, 'learning_rate/qwen_vl_interface': 5.145579460997544e-07, 'learning_rate/action_model': 5.145579460997544e-06, 'epoch': 2.01}) +[2026-06-17 19:00:10,742][starVLA.training.train_starvla][INFO] - Step 6829, Loss: {'train/loss': 0.3717194199562073, 'train/grad_norm_pre_clip': 11.490766525268555, 'timing/data': 0.004441556986421347, 'timing/model': 1.2071646560216323, 'learning_rate/qwen_vl_interface': 5.143892450150411e-07, 'learning_rate/action_model': 5.143892450150411e-06, 'epoch': 2.01}) +[2026-06-17 19:00:15,546][starVLA.training.train_starvla][INFO] - Step 6830, Loss: {'train/loss': 0.792694628238678, 'train/grad_norm_pre_clip': 14.819669723510742, 'timing/data': 0.0049256510101258755, 'timing/model': 1.2119698739843443, 'learning_rate/qwen_vl_interface': 5.142215256275338e-07, 'learning_rate/action_model': 5.142215256275338e-06, 'epoch': 2.01}) +[2026-06-17 19:00:20,358][starVLA.training.train_starvla][INFO] - Step 6831, Loss: {'train/loss': 0.920577347278595, 'train/grad_norm_pre_clip': 33.673561096191406, 'timing/data': 0.005036549060605466, 'timing/model': 1.2053680839017034, 'learning_rate/qwen_vl_interface': 5.140547879719993e-07, 'learning_rate/action_model': 5.140547879719993e-06, 'epoch': 2.01}) +[2026-06-17 19:00:25,083][starVLA.training.train_starvla][INFO] - Step 6832, Loss: {'train/loss': 0.5856607556343079, 'train/grad_norm_pre_clip': 15.347940444946289, 'timing/data': 0.004555008024908602, 'timing/model': 1.2752837520092726, 'learning_rate/qwen_vl_interface': 5.138890320830038e-07, 'learning_rate/action_model': 5.138890320830038e-06, 'epoch': 2.01}) +[2026-06-17 19:00:29,759][starVLA.training.train_starvla][INFO] - Step 6833, Loss: {'train/loss': 0.48563718795776367, 'train/grad_norm_pre_clip': 13.722414016723633, 'timing/data': 0.005064386990852654, 'timing/model': 1.2126448430353776, 'learning_rate/qwen_vl_interface': 5.137242579949081e-07, 'learning_rate/action_model': 5.137242579949081e-06, 'epoch': 2.01}) +[2026-06-17 19:00:34,539][starVLA.training.train_starvla][INFO] - Step 6834, Loss: {'train/loss': 0.47511792182922363, 'train/grad_norm_pre_clip': 12.320080757141113, 'timing/data': 0.005186225054785609, 'timing/model': 1.209823218989186, 'learning_rate/qwen_vl_interface': 5.135604657418702e-07, 'learning_rate/action_model': 5.135604657418702e-06, 'epoch': 2.01}) +[2026-06-17 19:00:39,322][starVLA.training.train_starvla][INFO] - Step 6835, Loss: {'train/loss': 0.6313785910606384, 'train/grad_norm_pre_clip': 14.481599807739258, 'timing/data': 0.005102050956338644, 'timing/model': 1.203674859018065, 'learning_rate/qwen_vl_interface': 5.133976553578443e-07, 'learning_rate/action_model': 5.1339765535784425e-06, 'epoch': 2.01}) +[2026-06-17 19:00:44,119][starVLA.training.train_starvla][INFO] - Step 6836, Loss: {'train/loss': 0.7330840826034546, 'train/grad_norm_pre_clip': 17.191680908203125, 'timing/data': 0.005037733004428446, 'timing/model': 1.2046959520084783, 'learning_rate/qwen_vl_interface': 5.132358268765813e-07, 'learning_rate/action_model': 5.132358268765813e-06, 'epoch': 2.01}) +[2026-06-17 19:00:48,929][starVLA.training.train_starvla][INFO] - Step 6837, Loss: {'train/loss': 0.49704158306121826, 'train/grad_norm_pre_clip': 13.590936660766602, 'timing/data': 0.004872763995081186, 'timing/model': 1.3318630820140243, 'learning_rate/qwen_vl_interface': 5.130749803316284e-07, 'learning_rate/action_model': 5.130749803316284e-06, 'epoch': 2.01}) +[2026-06-17 19:00:53,657][starVLA.training.train_starvla][INFO] - Step 6838, Loss: {'train/loss': 0.5981509685516357, 'train/grad_norm_pre_clip': 14.267916679382324, 'timing/data': 0.0045537929981946945, 'timing/model': 1.2789209459442645, 'learning_rate/qwen_vl_interface': 5.129151157563294e-07, 'learning_rate/action_model': 5.129151157563293e-06, 'epoch': 2.01}) +[2026-06-17 19:00:58,311][starVLA.training.train_starvla][INFO] - Step 6839, Loss: {'train/loss': 0.6924231052398682, 'train/grad_norm_pre_clip': 17.967260360717773, 'timing/data': 0.004978094948455691, 'timing/model': 1.2052342910319567, 'learning_rate/qwen_vl_interface': 5.127562331838241e-07, 'learning_rate/action_model': 5.12756233183824e-06, 'epoch': 2.01}) +[2026-06-17 19:01:03,101][starVLA.training.train_starvla][INFO] - Step 6840, Loss: {'train/loss': 0.6611472964286804, 'train/grad_norm_pre_clip': 25.693660736083984, 'timing/data': 0.004579860949888825, 'timing/model': 1.206186456955038, 'learning_rate/qwen_vl_interface': 5.125983326470492e-07, 'learning_rate/action_model': 5.125983326470492e-06, 'epoch': 2.01}) +[2026-06-17 19:01:07,886][starVLA.training.train_starvla][INFO] - Step 6841, Loss: {'train/loss': 0.7554334402084351, 'train/grad_norm_pre_clip': 16.798280715942383, 'timing/data': 0.004698551958426833, 'timing/model': 1.2036670920206234, 'learning_rate/qwen_vl_interface': 5.124414141787372e-07, 'learning_rate/action_model': 5.124414141787372e-06, 'epoch': 2.01}) +[2026-06-17 19:01:12,669][starVLA.training.train_starvla][INFO] - Step 6842, Loss: {'train/loss': 0.7852495908737183, 'train/grad_norm_pre_clip': 17.246807098388672, 'timing/data': 0.004828082979656756, 'timing/model': 1.2042044239351526, 'learning_rate/qwen_vl_interface': 5.122854778114185e-07, 'learning_rate/action_model': 5.122854778114185e-06, 'epoch': 2.01}) +[2026-06-17 19:01:17,457][starVLA.training.train_starvla][INFO] - Step 6843, Loss: {'train/loss': 0.6454265713691711, 'train/grad_norm_pre_clip': 16.561792373657227, 'timing/data': 0.004704788094386458, 'timing/model': 1.2064722339855507, 'learning_rate/qwen_vl_interface': 5.12130523577418e-07, 'learning_rate/action_model': 5.12130523577418e-06, 'epoch': 2.01}) +[2026-06-17 19:01:22,198][starVLA.training.train_starvla][INFO] - Step 6844, Loss: {'train/loss': 0.6203585863113403, 'train/grad_norm_pre_clip': 13.46840763092041, 'timing/data': 0.0050363290356472135, 'timing/model': 1.2895405270392075, 'learning_rate/qwen_vl_interface': 5.119765515088579e-07, 'learning_rate/action_model': 5.1197655150885785e-06, 'epoch': 2.01}) +[2026-06-17 19:01:26,875][starVLA.training.train_starvla][INFO] - Step 6845, Loss: {'train/loss': 0.7053089737892151, 'train/grad_norm_pre_clip': 13.885144233703613, 'timing/data': 0.004529727040790021, 'timing/model': 1.207339650951326, 'learning_rate/qwen_vl_interface': 5.11823561637657e-07, 'learning_rate/action_model': 5.1182356163765705e-06, 'epoch': 2.01}) +[2026-06-17 19:01:31,666][starVLA.training.train_starvla][INFO] - Step 6846, Loss: {'train/loss': 0.3987492322921753, 'train/grad_norm_pre_clip': 10.477673530578613, 'timing/data': 0.004876684048213065, 'timing/model': 1.2086371349869296, 'learning_rate/qwen_vl_interface': 5.116715539955306e-07, 'learning_rate/action_model': 5.116715539955305e-06, 'epoch': 2.01}) +[2026-06-17 19:01:36,458][starVLA.training.train_starvla][INFO] - Step 6847, Loss: {'train/loss': 0.49010682106018066, 'train/grad_norm_pre_clip': 11.999236106872559, 'timing/data': 0.004881122033111751, 'timing/model': 1.2113980669528246, 'learning_rate/qwen_vl_interface': 5.115205286139895e-07, 'learning_rate/action_model': 5.115205286139895e-06, 'epoch': 2.01}) +[2026-06-17 19:01:41,253][starVLA.training.train_starvla][INFO] - Step 6848, Loss: {'train/loss': 0.6137250065803528, 'train/grad_norm_pre_clip': 14.532654762268066, 'timing/data': 0.0045669859973713756, 'timing/model': 1.2097848620032892, 'learning_rate/qwen_vl_interface': 5.11370485524342e-07, 'learning_rate/action_model': 5.1137048552434205e-06, 'epoch': 2.02}) +[2026-06-17 19:01:46,038][starVLA.training.train_starvla][INFO] - Step 6849, Loss: {'train/loss': 0.416550874710083, 'train/grad_norm_pre_clip': 12.503165245056152, 'timing/data': 0.004590910975821316, 'timing/model': 1.337145552970469, 'learning_rate/qwen_vl_interface': 5.112214247576915e-07, 'learning_rate/action_model': 5.1122142475769155e-06, 'epoch': 2.02}) +[2026-06-17 19:01:50,696][starVLA.training.train_starvla][INFO] - Step 6850, Loss: {'train/loss': 0.3775237202644348, 'train/grad_norm_pre_clip': 10.95154094696045, 'timing/data': 0.004941742983646691, 'timing/model': 1.206325397011824, 'learning_rate/qwen_vl_interface': 5.110733463449384e-07, 'learning_rate/action_model': 5.110733463449385e-06, 'epoch': 2.02}) +[2026-06-17 19:01:55,482][starVLA.training.train_starvla][INFO] - Step 6851, Loss: {'train/loss': 0.7359517812728882, 'train/grad_norm_pre_clip': 107.4025650024414, 'timing/data': 0.005069080973044038, 'timing/model': 1.2064728760160506, 'learning_rate/qwen_vl_interface': 5.109262503167808e-07, 'learning_rate/action_model': 5.109262503167808e-06, 'epoch': 2.02}) +[2026-06-17 19:02:00,267][starVLA.training.train_starvla][INFO] - Step 6852, Loss: {'train/loss': 0.8267291188240051, 'train/grad_norm_pre_clip': 16.015792846679688, 'timing/data': 0.0048798039788380265, 'timing/model': 1.2048818790353835, 'learning_rate/qwen_vl_interface': 5.107801367037106e-07, 'learning_rate/action_model': 5.107801367037107e-06, 'epoch': 2.02}) +[2026-06-17 19:02:04,972][starVLA.training.train_starvla][INFO] - Step 6853, Loss: {'train/loss': 0.675899863243103, 'train/grad_norm_pre_clip': 17.11581802368164, 'timing/data': 0.005590123008005321, 'timing/model': 1.2035745800239965, 'learning_rate/qwen_vl_interface': 5.106350055360173e-07, 'learning_rate/action_model': 5.106350055360173e-06, 'epoch': 2.02}) +[2026-06-17 19:02:09,734][starVLA.training.train_starvla][INFO] - Step 6854, Loss: {'train/loss': 0.5959214568138123, 'train/grad_norm_pre_clip': 15.357208251953125, 'timing/data': 0.004742122022435069, 'timing/model': 1.2037509370129555, 'learning_rate/qwen_vl_interface': 5.104908568437875e-07, 'learning_rate/action_model': 5.104908568437875e-06, 'epoch': 2.02}) +[2026-06-17 19:02:14,518][starVLA.training.train_starvla][INFO] - Step 6855, Loss: {'train/loss': 0.7042720317840576, 'train/grad_norm_pre_clip': 15.455103874206543, 'timing/data': 0.004954273928888142, 'timing/model': 1.333168598008342, 'learning_rate/qwen_vl_interface': 5.103476906569034e-07, 'learning_rate/action_model': 5.103476906569033e-06, 'epoch': 2.02}) +[2026-06-17 19:02:19,180][starVLA.training.train_starvla][INFO] - Step 6856, Loss: {'train/loss': 0.6085677742958069, 'train/grad_norm_pre_clip': 13.372340202331543, 'timing/data': 0.00486128602642566, 'timing/model': 1.2063073019962758, 'learning_rate/qwen_vl_interface': 5.102055070050428e-07, 'learning_rate/action_model': 5.102055070050427e-06, 'epoch': 2.02}) +[2026-06-17 19:02:23,969][starVLA.training.train_starvla][INFO] - Step 6857, Loss: {'train/loss': 0.6439851522445679, 'train/grad_norm_pre_clip': 15.213277816772461, 'timing/data': 0.004781599971465766, 'timing/model': 1.2051627089967951, 'learning_rate/qwen_vl_interface': 5.100643059176803e-07, 'learning_rate/action_model': 5.100643059176802e-06, 'epoch': 2.02}) +[2026-06-17 19:02:28,751][starVLA.training.train_starvla][INFO] - Step 6858, Loss: {'train/loss': 0.4728103280067444, 'train/grad_norm_pre_clip': 13.003385543823242, 'timing/data': 0.0049518420128151774, 'timing/model': 1.2057430960703641, 'learning_rate/qwen_vl_interface': 5.099240874240883e-07, 'learning_rate/action_model': 5.099240874240883e-06, 'epoch': 2.02}) +[2026-06-17 19:02:33,576][starVLA.training.train_starvla][INFO] - Step 6859, Loss: {'train/loss': 0.651017963886261, 'train/grad_norm_pre_clip': 15.503850936889648, 'timing/data': 0.004797403002157807, 'timing/model': 1.2104656640440226, 'learning_rate/qwen_vl_interface': 5.097848515533331e-07, 'learning_rate/action_model': 5.0978485155333305e-06, 'epoch': 2.02}) +[2026-06-17 19:02:38,338][starVLA.training.train_starvla][INFO] - Step 6860, Loss: {'train/loss': 0.40147390961647034, 'train/grad_norm_pre_clip': 11.090959548950195, 'timing/data': 0.005093374056741595, 'timing/model': 1.208958534989506, 'learning_rate/qwen_vl_interface': 5.096465983342791e-07, 'learning_rate/action_model': 5.096465983342791e-06, 'epoch': 2.02}) +[2026-06-17 19:02:43,126][starVLA.training.train_starvla][INFO] - Step 6861, Loss: {'train/loss': 0.817199170589447, 'train/grad_norm_pre_clip': 19.01868438720703, 'timing/data': 0.004975173040293157, 'timing/model': 1.3274675549473614, 'learning_rate/qwen_vl_interface': 5.095093277955859e-07, 'learning_rate/action_model': 5.095093277955859e-06, 'epoch': 2.02}) +[2026-06-17 19:02:47,779][starVLA.training.train_starvla][INFO] - Step 6862, Loss: {'train/loss': 0.8623479604721069, 'train/grad_norm_pre_clip': 17.450681686401367, 'timing/data': 0.0045496870297938585, 'timing/model': 1.204404029995203, 'learning_rate/qwen_vl_interface': 5.093730399657101e-07, 'learning_rate/action_model': 5.0937303996571e-06, 'epoch': 2.02}) +[2026-06-17 19:02:52,563][starVLA.training.train_starvla][INFO] - Step 6863, Loss: {'train/loss': 0.6907462477684021, 'train/grad_norm_pre_clip': 13.821131706237793, 'timing/data': 0.004872513934969902, 'timing/model': 1.205326360068284, 'learning_rate/qwen_vl_interface': 5.092377348729043e-07, 'learning_rate/action_model': 5.092377348729043e-06, 'epoch': 2.02}) +[2026-06-17 19:02:57,456][starVLA.training.train_starvla][INFO] - Step 6864, Loss: {'train/loss': 0.6906333565711975, 'train/grad_norm_pre_clip': 18.345417022705078, 'timing/data': 0.0048401569947600365, 'timing/model': 1.2232693289406598, 'learning_rate/qwen_vl_interface': 5.091034125452171e-07, 'learning_rate/action_model': 5.091034125452171e-06, 'epoch': 2.02}) +[2026-06-17 19:03:02,252][starVLA.training.train_starvla][INFO] - Step 6865, Loss: {'train/loss': 0.5985480546951294, 'train/grad_norm_pre_clip': 14.967068672180176, 'timing/data': 0.004781809984706342, 'timing/model': 1.2092887969920412, 'learning_rate/qwen_vl_interface': 5.089700730104944e-07, 'learning_rate/action_model': 5.089700730104944e-06, 'epoch': 2.02}) +[2026-06-17 19:03:07,049][starVLA.training.train_starvla][INFO] - Step 6866, Loss: {'train/loss': 0.5959057807922363, 'train/grad_norm_pre_clip': 15.973997116088867, 'timing/data': 0.0051375150214880705, 'timing/model': 1.3442240880103782, 'learning_rate/qwen_vl_interface': 5.08837716296377e-07, 'learning_rate/action_model': 5.0883771629637696e-06, 'epoch': 2.02}) +[2026-06-17 19:03:11,709][starVLA.training.train_starvla][INFO] - Step 6867, Loss: {'train/loss': 0.6383073329925537, 'train/grad_norm_pre_clip': 13.187064170837402, 'timing/data': 0.004709627944976091, 'timing/model': 1.2069825580110773, 'learning_rate/qwen_vl_interface': 5.087063424303026e-07, 'learning_rate/action_model': 5.087063424303026e-06, 'epoch': 2.02}) +[2026-06-17 19:03:16,528][starVLA.training.train_starvla][INFO] - Step 6868, Loss: {'train/loss': 0.4627523124217987, 'train/grad_norm_pre_clip': 17.542110443115234, 'timing/data': 0.004777281079441309, 'timing/model': 1.2143995210062712, 'learning_rate/qwen_vl_interface': 5.085759514395052e-07, 'learning_rate/action_model': 5.085759514395051e-06, 'epoch': 2.02}) +[2026-06-17 19:03:21,258][starVLA.training.train_starvla][INFO] - Step 6869, Loss: {'train/loss': 0.7089334726333618, 'train/grad_norm_pre_clip': 17.8219051361084, 'timing/data': 0.004793693078681827, 'timing/model': 1.2040988829685375, 'learning_rate/qwen_vl_interface': 5.084465433510151e-07, 'learning_rate/action_model': 5.0844654335101515e-06, 'epoch': 2.02}) +[2026-06-17 19:03:26,019][starVLA.training.train_starvla][INFO] - Step 6870, Loss: {'train/loss': 0.786710798740387, 'train/grad_norm_pre_clip': 16.213237762451172, 'timing/data': 0.005222963052801788, 'timing/model': 1.2064717939356342, 'learning_rate/qwen_vl_interface': 5.083181181916588e-07, 'learning_rate/action_model': 5.083181181916588e-06, 'epoch': 2.02}) +[2026-06-17 19:03:30,808][starVLA.training.train_starvla][INFO] - Step 6871, Loss: {'train/loss': 0.6489710211753845, 'train/grad_norm_pre_clip': 14.397377014160156, 'timing/data': 0.005162953049875796, 'timing/model': 1.2049303849926218, 'learning_rate/qwen_vl_interface': 5.081906759880585e-07, 'learning_rate/action_model': 5.081906759880585e-06, 'epoch': 2.02}) +[2026-06-17 19:03:35,598][starVLA.training.train_starvla][INFO] - Step 6872, Loss: {'train/loss': 0.7481725215911865, 'train/grad_norm_pre_clip': 15.860838890075684, 'timing/data': 0.004853502963669598, 'timing/model': 1.3378060760442168, 'learning_rate/qwen_vl_interface': 5.080642167666338e-07, 'learning_rate/action_model': 5.080642167666337e-06, 'epoch': 2.02}) +[2026-06-17 19:03:40,262][starVLA.training.train_starvla][INFO] - Step 6873, Loss: {'train/loss': 0.6741229295730591, 'train/grad_norm_pre_clip': 13.908905029296875, 'timing/data': 0.005011811968870461, 'timing/model': 1.2084470510017127, 'learning_rate/qwen_vl_interface': 5.079387405535995e-07, 'learning_rate/action_model': 5.079387405535995e-06, 'epoch': 2.02}) +[2026-06-17 19:03:45,058][starVLA.training.train_starvla][INFO] - Step 6874, Loss: {'train/loss': 0.6333247423171997, 'train/grad_norm_pre_clip': 13.640613555908203, 'timing/data': 0.004938623984344304, 'timing/model': 1.2048111150506884, 'learning_rate/qwen_vl_interface': 5.07814247374967e-07, 'learning_rate/action_model': 5.078142473749671e-06, 'epoch': 2.02}) +[2026-06-17 19:03:49,848][starVLA.training.train_starvla][INFO] - Step 6875, Loss: {'train/loss': 0.452000230550766, 'train/grad_norm_pre_clip': 13.619677543640137, 'timing/data': 0.004413374001160264, 'timing/model': 1.2067434419877827, 'learning_rate/qwen_vl_interface': 5.07690737256544e-07, 'learning_rate/action_model': 5.07690737256544e-06, 'epoch': 2.02}) +[2026-06-17 19:03:54,600][starVLA.training.train_starvla][INFO] - Step 6876, Loss: {'train/loss': 0.7362823486328125, 'train/grad_norm_pre_clip': 23.962135314941406, 'timing/data': 0.004716795985586941, 'timing/model': 1.2073276189621538, 'learning_rate/qwen_vl_interface': 5.075682102239336e-07, 'learning_rate/action_model': 5.075682102239335e-06, 'epoch': 2.02}) +[2026-06-17 19:03:59,373][starVLA.training.train_starvla][INFO] - Step 6877, Loss: {'train/loss': 0.42243069410324097, 'train/grad_norm_pre_clip': 11.90893840789795, 'timing/data': 0.004815916996449232, 'timing/model': 1.203996257041581, 'learning_rate/qwen_vl_interface': 5.074466663025369e-07, 'learning_rate/action_model': 5.074466663025369e-06, 'epoch': 2.02}) +[2026-06-17 19:04:04,168][starVLA.training.train_starvla][INFO] - Step 6878, Loss: {'train/loss': 0.7653700113296509, 'train/grad_norm_pre_clip': 16.425209045410156, 'timing/data': 0.004935770994052291, 'timing/model': 1.3358617060584947, 'learning_rate/qwen_vl_interface': 5.07326105517549e-07, 'learning_rate/action_model': 5.07326105517549e-06, 'epoch': 2.02}) +[2026-06-17 19:04:08,829][starVLA.training.train_starvla][INFO] - Step 6879, Loss: {'train/loss': 0.5109174847602844, 'train/grad_norm_pre_clip': 11.337331771850586, 'timing/data': 0.004851378966122866, 'timing/model': 1.2052862440468743, 'learning_rate/qwen_vl_interface': 5.072065278939628e-07, 'learning_rate/action_model': 5.072065278939627e-06, 'epoch': 2.02}) +[2026-06-17 19:04:13,628][starVLA.training.train_starvla][INFO] - Step 6880, Loss: {'train/loss': 0.6349479556083679, 'train/grad_norm_pre_clip': 14.885775566101074, 'timing/data': 0.004663006053306162, 'timing/model': 1.2114239999791607, 'learning_rate/qwen_vl_interface': 5.070879334565671e-07, 'learning_rate/action_model': 5.070879334565671e-06, 'epoch': 2.02}) +[2026-06-17 19:04:18,423][starVLA.training.train_starvla][INFO] - Step 6881, Loss: {'train/loss': 0.7868082523345947, 'train/grad_norm_pre_clip': 16.118541717529297, 'timing/data': 0.004717167932540178, 'timing/model': 1.2099527929676697, 'learning_rate/qwen_vl_interface': 5.069703222299464e-07, 'learning_rate/action_model': 5.069703222299463e-06, 'epoch': 2.02}) +[2026-06-17 19:04:23,231][starVLA.training.train_starvla][INFO] - Step 6882, Loss: {'train/loss': 0.4897676110267639, 'train/grad_norm_pre_clip': 13.114053726196289, 'timing/data': 0.004543104092590511, 'timing/model': 1.2113725889939815, 'learning_rate/qwen_vl_interface': 5.068536942384813e-07, 'learning_rate/action_model': 5.068536942384812e-06, 'epoch': 2.03}) +[2026-06-17 19:04:27,917][starVLA.training.train_starvla][INFO] - Step 6883, Loss: {'train/loss': 0.5921565890312195, 'train/grad_norm_pre_clip': 15.093668937683105, 'timing/data': 0.005446888972073793, 'timing/model': 1.2047227750299498, 'learning_rate/qwen_vl_interface': 5.067380495063494e-07, 'learning_rate/action_model': 5.067380495063494e-06, 'epoch': 2.03}) +[2026-06-17 19:04:32,611][starVLA.training.train_starvla][INFO] - Step 6884, Loss: {'train/loss': 0.9806424975395203, 'train/grad_norm_pre_clip': 21.131717681884766, 'timing/data': 0.004317743005231023, 'timing/model': 1.2408570889383554, 'learning_rate/qwen_vl_interface': 5.066233880575239e-07, 'learning_rate/action_model': 5.0662338805752395e-06, 'epoch': 2.03}) +[2026-06-17 19:04:37,269][starVLA.training.train_starvla][INFO] - Step 6885, Loss: {'train/loss': 0.7865679860115051, 'train/grad_norm_pre_clip': 22.110076904296875, 'timing/data': 0.004717558971606195, 'timing/model': 1.205344578018412, 'learning_rate/qwen_vl_interface': 5.065097099157743e-07, 'learning_rate/action_model': 5.065097099157743e-06, 'epoch': 2.03}) +[2026-06-17 19:04:42,018][starVLA.training.train_starvla][INFO] - Step 6886, Loss: {'train/loss': 0.5331745147705078, 'train/grad_norm_pre_clip': 12.320960998535156, 'timing/data': 0.004867136012762785, 'timing/model': 1.2050280750263482, 'learning_rate/qwen_vl_interface': 5.063970151046654e-07, 'learning_rate/action_model': 5.0639701510466545e-06, 'epoch': 2.03}) +[2026-06-17 19:04:46,807][starVLA.training.train_starvla][INFO] - Step 6887, Loss: {'train/loss': 0.4519933760166168, 'train/grad_norm_pre_clip': 12.198981285095215, 'timing/data': 0.004607924958691001, 'timing/model': 1.2055949290515855, 'learning_rate/qwen_vl_interface': 5.062853036475604e-07, 'learning_rate/action_model': 5.062853036475603e-06, 'epoch': 2.03}) +[2026-06-17 19:04:51,589][starVLA.training.train_starvla][INFO] - Step 6888, Loss: {'train/loss': 0.6529240012168884, 'train/grad_norm_pre_clip': 13.90299129486084, 'timing/data': 0.0046490279491990805, 'timing/model': 1.2048122199485078, 'learning_rate/qwen_vl_interface': 5.061745755676161e-07, 'learning_rate/action_model': 5.061745755676161e-06, 'epoch': 2.03}) +[2026-06-17 19:04:56,365][starVLA.training.train_starvla][INFO] - Step 6889, Loss: {'train/loss': 0.7237557768821716, 'train/grad_norm_pre_clip': 16.314199447631836, 'timing/data': 0.004826157004572451, 'timing/model': 1.3302538839634508, 'learning_rate/qwen_vl_interface': 5.060648308877866e-07, 'learning_rate/action_model': 5.0606483088778656e-06, 'epoch': 2.03}) +[2026-06-17 19:05:01,217][starVLA.training.train_starvla][INFO] - Step 6890, Loss: {'train/loss': 0.526073157787323, 'train/grad_norm_pre_clip': 12.746975898742676, 'timing/data': 0.004669017042033374, 'timing/model': 1.3294040070613846, 'learning_rate/qwen_vl_interface': 5.059560696308229e-07, 'learning_rate/action_model': 5.059560696308228e-06, 'epoch': 2.03}) +[2026-06-17 19:05:05,877][starVLA.training.train_starvla][INFO] - Step 6891, Loss: {'train/loss': 0.5236572623252869, 'train/grad_norm_pre_clip': 12.824424743652344, 'timing/data': 0.004412498907186091, 'timing/model': 1.206392896012403, 'learning_rate/qwen_vl_interface': 5.0584829181927e-07, 'learning_rate/action_model': 5.058482918192699e-06, 'epoch': 2.03}) +[2026-06-17 19:05:10,625][starVLA.training.train_starvla][INFO] - Step 6892, Loss: {'train/loss': 0.4632641077041626, 'train/grad_norm_pre_clip': 13.078465461730957, 'timing/data': 0.00472491595428437, 'timing/model': 1.2058616318972781, 'learning_rate/qwen_vl_interface': 5.05741497475472e-07, 'learning_rate/action_model': 5.05741497475472e-06, 'epoch': 2.03}) +[2026-06-17 19:05:15,413][starVLA.training.train_starvla][INFO] - Step 6893, Loss: {'train/loss': 0.9046869277954102, 'train/grad_norm_pre_clip': 16.907169342041016, 'timing/data': 0.004993035923689604, 'timing/model': 1.2052099170396104, 'learning_rate/qwen_vl_interface': 5.056356866215663e-07, 'learning_rate/action_model': 5.056356866215662e-06, 'epoch': 2.03}) +[2026-06-17 19:05:20,204][starVLA.training.train_starvla][INFO] - Step 6894, Loss: {'train/loss': 0.3846876621246338, 'train/grad_norm_pre_clip': 10.663601875305176, 'timing/data': 0.004741125972941518, 'timing/model': 1.20809586904943, 'learning_rate/qwen_vl_interface': 5.055308592794876e-07, 'learning_rate/action_model': 5.0553085927948756e-06, 'epoch': 2.03}) +[2026-06-17 19:05:24,987][starVLA.training.train_starvla][INFO] - Step 6895, Loss: {'train/loss': 0.5734636187553406, 'train/grad_norm_pre_clip': 12.692885398864746, 'timing/data': 0.004420764045789838, 'timing/model': 1.3313161220867187, 'learning_rate/qwen_vl_interface': 5.054270154709674e-07, 'learning_rate/action_model': 5.054270154709673e-06, 'epoch': 2.03}) +[2026-06-17 19:05:29,797][starVLA.training.train_starvla][INFO] - Step 6896, Loss: {'train/loss': 0.7200984358787537, 'train/grad_norm_pre_clip': 15.969480514526367, 'timing/data': 0.005077151930890977, 'timing/model': 1.3240688550285995, 'learning_rate/qwen_vl_interface': 5.053241552175322e-07, 'learning_rate/action_model': 5.053241552175321e-06, 'epoch': 2.03}) +[2026-06-17 19:05:34,456][starVLA.training.train_starvla][INFO] - Step 6897, Loss: {'train/loss': 0.5693634152412415, 'train/grad_norm_pre_clip': 15.22396183013916, 'timing/data': 0.004835283965803683, 'timing/model': 1.2065616570180282, 'learning_rate/qwen_vl_interface': 5.05222278540505e-07, 'learning_rate/action_model': 5.05222278540505e-06, 'epoch': 2.03}) +[2026-06-17 19:05:39,221][starVLA.training.train_starvla][INFO] - Step 6898, Loss: {'train/loss': 0.5596969127655029, 'train/grad_norm_pre_clip': 28.48103141784668, 'timing/data': 0.0047703610034659505, 'timing/model': 1.2069154649507254, 'learning_rate/qwen_vl_interface': 5.051213854610051e-07, 'learning_rate/action_model': 5.0512138546100514e-06, 'epoch': 2.03}) +[2026-06-17 19:05:44,013][starVLA.training.train_starvla][INFO] - Step 6899, Loss: {'train/loss': 0.6209570169448853, 'train/grad_norm_pre_clip': 15.901034355163574, 'timing/data': 0.0048110849456861615, 'timing/model': 1.2073284030193463, 'learning_rate/qwen_vl_interface': 5.050214759999478e-07, 'learning_rate/action_model': 5.050214759999478e-06, 'epoch': 2.03}) +[2026-06-17 19:07:25,702][starVLA.training.train_starvla][INFO] - Step 6900, Loss: {'train/loss': 0.48297858238220215, 'train/grad_norm_pre_clip': 12.841085433959961, 'eval/loss': 0.7020787596702576, 'eval/latency_0/loss': 0.7020787596702576, 'eval/demon_fire_precision': 0.9028, 'eval/demon_fire_recall': 0.9372923588039868, 'eval/demon_fire_f1': 0.9197229013854931, 'eval/demon_fire_rate_model': 0.6952169076751946, 'eval/demon_fire_rate_teacher': 0.6696329254727474, 'eval/demon_left_precision': 0.8243243243243243, 'eval/demon_left_recall': 0.8294617563739377, 'eval/demon_left_f1': 0.8268850607173115, 'eval/demon_left_rate_model': 0.4938820912124583, 'eval/demon_left_rate_teacher': 0.49082313681868744, 'eval/demon_right_precision': 0.8309151785714286, 'eval/demon_right_recall': 0.8412429378531073, 'eval/demon_right_f1': 0.8360471645143178, 'eval/demon_right_rate_model': 0.4983314794215795, 'eval/demon_right_rate_teacher': 0.4922135706340378, 'eval/demon_move_f1': 0.8314661126158147, 'eval/demon_cc_f1': 0.8913392760000053, 'eval/latency_0/demon_fire_precision': 0.9028, 'eval/latency_0/demon_fire_recall': 0.9372923588039868, 'eval/latency_0/demon_fire_f1': 0.9197229013854931, 'eval/latency_0/demon_fire_rate_model': 0.6952169076751946, 'eval/latency_0/demon_fire_rate_teacher': 0.6696329254727474, 'eval/latency_0/demon_left_precision': 0.8243243243243243, 'eval/latency_0/demon_left_recall': 0.8294617563739377, 'eval/latency_0/demon_left_f1': 0.8268850607173115, 'eval/latency_0/demon_left_rate_model': 0.4938820912124583, 'eval/latency_0/demon_left_rate_teacher': 0.49082313681868744, 'eval/latency_0/demon_right_precision': 0.8309151785714286, 'eval/latency_0/demon_right_recall': 0.8412429378531073, 'eval/latency_0/demon_right_f1': 0.8360471645143178, 'eval/latency_0/demon_right_rate_model': 0.4983314794215795, 'eval/latency_0/demon_right_rate_teacher': 0.4922135706340378, 'eval/latency_0/demon_move_f1': 0.8314661126158147, 'eval/latency_0/demon_cc_f1': 0.8913392760000053, 'timing/data': 0.005045488942414522, 'timing/model': 1.204267107998021, 'learning_rate/qwen_vl_interface': 5.049225501780441e-07, 'learning_rate/action_model': 5.049225501780441e-06, 'epoch': 2.03}) +[2026-06-17 19:07:25,702][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_6900_state +[2026-06-17 19:07:25,705][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 19:07:29,257][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_6900_state/model.safetensors +[2026-06-17 19:07:36,950][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_6900_state/optimizer.bin +[2026-06-17 19:07:36,951][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_6900_state/sampler.bin +[2026-06-17 19:07:36,951][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_6900_state/sampler_1.bin +[2026-06-17 19:07:36,951][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_6900_state/random_states_0.pkl +[2026-06-17 19:07:36,995][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 6900 +[2026-06-17 19:07:36,995][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 19:07:37,009][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 19:07:41,743][starVLA.training.train_starvla][INFO] - Step 6901, Loss: {'train/loss': 0.6139487028121948, 'train/grad_norm_pre_clip': 12.405842781066895, 'timing/data': 0.004570786957629025, 'timing/model': 1.187394856940955, 'learning_rate/qwen_vl_interface': 5.048246080158016e-07, 'learning_rate/action_model': 5.048246080158016e-06, 'epoch': 2.03}) +[2026-06-17 19:07:46,464][starVLA.training.train_starvla][INFO] - Step 6902, Loss: {'train/loss': 0.5128886699676514, 'train/grad_norm_pre_clip': 13.79376220703125, 'timing/data': 0.004904404981061816, 'timing/model': 1.1899397181114182, 'learning_rate/qwen_vl_interface': 5.047276495335243e-07, 'learning_rate/action_model': 5.047276495335242e-06, 'epoch': 2.03}) +[2026-06-17 19:07:51,174][starVLA.training.train_starvla][INFO] - Step 6903, Loss: {'train/loss': 0.8406455516815186, 'train/grad_norm_pre_clip': 15.679485321044922, 'timing/data': 0.005363080999813974, 'timing/model': 1.2600659639574587, 'learning_rate/qwen_vl_interface': 5.046316747513115e-07, 'learning_rate/action_model': 5.046316747513115e-06, 'epoch': 2.03}) +[2026-06-17 19:07:55,783][starVLA.training.train_starvla][INFO] - Step 6904, Loss: {'train/loss': 0.7467542290687561, 'train/grad_norm_pre_clip': 18.03078269958496, 'timing/data': 0.005079157999716699, 'timing/model': 1.1930740220705047, 'learning_rate/qwen_vl_interface': 5.045366836890585e-07, 'learning_rate/action_model': 5.045366836890584e-06, 'epoch': 2.03}) +[2026-06-17 19:08:00,502][starVLA.training.train_starvla][INFO] - Step 6905, Loss: {'train/loss': 0.4976692795753479, 'train/grad_norm_pre_clip': 13.025016784667969, 'timing/data': 0.004716756986454129, 'timing/model': 1.1953091400209814, 'learning_rate/qwen_vl_interface': 5.044426763664572e-07, 'learning_rate/action_model': 5.044426763664571e-06, 'epoch': 2.03}) +[2026-06-17 19:08:05,256][starVLA.training.train_starvla][INFO] - Step 6906, Loss: {'train/loss': 0.6159084439277649, 'train/grad_norm_pre_clip': 12.743271827697754, 'timing/data': 0.004504868062213063, 'timing/model': 1.1968750699888915, 'learning_rate/qwen_vl_interface': 5.043496528029954e-07, 'learning_rate/action_model': 5.043496528029954e-06, 'epoch': 2.03}) +[2026-06-17 19:08:10,015][starVLA.training.train_starvla][INFO] - Step 6907, Loss: {'train/loss': 0.5471144914627075, 'train/grad_norm_pre_clip': 13.02903938293457, 'timing/data': 0.005076528061181307, 'timing/model': 1.1987569109769538, 'learning_rate/qwen_vl_interface': 5.042576130179573e-07, 'learning_rate/action_model': 5.042576130179572e-06, 'epoch': 2.03}) +[2026-06-17 19:08:14,712][starVLA.training.train_starvla][INFO] - Step 6908, Loss: {'train/loss': 0.5863490700721741, 'train/grad_norm_pre_clip': 14.669081687927246, 'timing/data': 0.004920506034977734, 'timing/model': 1.264413911034353, 'learning_rate/qwen_vl_interface': 5.041665570304227e-07, 'learning_rate/action_model': 5.041665570304226e-06, 'epoch': 2.03}) +[2026-06-17 19:08:19,344][starVLA.training.train_starvla][INFO] - Step 6909, Loss: {'train/loss': 0.723300576210022, 'train/grad_norm_pre_clip': 15.143704414367676, 'timing/data': 0.004817030043341219, 'timing/model': 1.1998090490233153, 'learning_rate/qwen_vl_interface': 5.040764848592675e-07, 'learning_rate/action_model': 5.040764848592675e-06, 'epoch': 2.03}) +[2026-06-17 19:08:24,127][starVLA.training.train_starvla][INFO] - Step 6910, Loss: {'train/loss': 0.8629432916641235, 'train/grad_norm_pre_clip': 16.152856826782227, 'timing/data': 0.0046639819629490376, 'timing/model': 1.2014076709747314, 'learning_rate/qwen_vl_interface': 5.039873965231644e-07, 'learning_rate/action_model': 5.039873965231643e-06, 'epoch': 2.03}) +[2026-06-17 19:08:28,891][starVLA.training.train_starvla][INFO] - Step 6911, Loss: {'train/loss': 0.6300010681152344, 'train/grad_norm_pre_clip': 17.024015426635742, 'timing/data': 0.005170680000446737, 'timing/model': 1.2007267089793459, 'learning_rate/qwen_vl_interface': 5.0389929204058e-07, 'learning_rate/action_model': 5.0389929204058e-06, 'epoch': 2.03}) +[2026-06-17 19:08:33,653][starVLA.training.train_starvla][INFO] - Step 6912, Loss: {'train/loss': 0.6071299910545349, 'train/grad_norm_pre_clip': 14.688488960266113, 'timing/data': 0.004915339988656342, 'timing/model': 1.19878008402884, 'learning_rate/qwen_vl_interface': 5.038121714297802e-07, 'learning_rate/action_model': 5.038121714297802e-06, 'epoch': 2.03}) +[2026-06-17 19:08:38,421][starVLA.training.train_starvla][INFO] - Step 6913, Loss: {'train/loss': 0.5089709758758545, 'train/grad_norm_pre_clip': 12.759963989257812, 'timing/data': 0.004662813036702573, 'timing/model': 1.3319654850056395, 'learning_rate/qwen_vl_interface': 5.037260347088236e-07, 'learning_rate/action_model': 5.037260347088236e-06, 'epoch': 2.03}) +[2026-06-17 19:08:43,151][starVLA.training.train_starvla][INFO] - Step 6914, Loss: {'train/loss': 0.518852174282074, 'train/grad_norm_pre_clip': 11.850739479064941, 'timing/data': 0.005000121076591313, 'timing/model': 1.2767234870698303, 'learning_rate/qwen_vl_interface': 5.03640881895568e-07, 'learning_rate/action_model': 5.03640881895568e-06, 'epoch': 2.03}) +[2026-06-17 19:08:47,792][starVLA.training.train_starvla][INFO] - Step 6915, Loss: {'train/loss': 0.5911841988563538, 'train/grad_norm_pre_clip': 15.647379875183105, 'timing/data': 0.004776083980686963, 'timing/model': 1.1997558289440349, 'learning_rate/qwen_vl_interface': 5.035567130076642e-07, 'learning_rate/action_model': 5.035567130076642e-06, 'epoch': 2.03}) +[2026-06-17 19:08:52,568][starVLA.training.train_starvla][INFO] - Step 6916, Loss: {'train/loss': 0.40719231963157654, 'train/grad_norm_pre_clip': 10.629546165466309, 'timing/data': 0.004586427006870508, 'timing/model': 1.2006662570638582, 'learning_rate/qwen_vl_interface': 5.034735280625617e-07, 'learning_rate/action_model': 5.034735280625616e-06, 'epoch': 2.04}) +[2026-06-17 19:08:57,348][starVLA.training.train_starvla][INFO] - Step 6917, Loss: {'train/loss': 0.5068963170051575, 'train/grad_norm_pre_clip': 12.182661056518555, 'timing/data': 0.004697725991718471, 'timing/model': 1.204039876931347, 'learning_rate/qwen_vl_interface': 5.033913270775039e-07, 'learning_rate/action_model': 5.033913270775039e-06, 'epoch': 2.04}) +[2026-06-17 19:09:02,127][starVLA.training.train_starvla][INFO] - Step 6918, Loss: {'train/loss': 0.6115454435348511, 'train/grad_norm_pre_clip': 12.210071563720703, 'timing/data': 0.004841624991968274, 'timing/model': 1.2023040279746056, 'learning_rate/qwen_vl_interface': 5.03310110069532e-07, 'learning_rate/action_model': 5.03310110069532e-06, 'epoch': 2.04}) +[2026-06-17 19:09:06,893][starVLA.training.train_starvla][INFO] - Step 6919, Loss: {'train/loss': 0.606028139591217, 'train/grad_norm_pre_clip': 17.164213180541992, 'timing/data': 0.004748782026581466, 'timing/model': 1.20658032305073, 'learning_rate/qwen_vl_interface': 5.032298770554818e-07, 'learning_rate/action_model': 5.032298770554818e-06, 'epoch': 2.04}) +[2026-06-17 19:09:11,635][starVLA.training.train_starvla][INFO] - Step 6920, Loss: {'train/loss': 0.6943010091781616, 'train/grad_norm_pre_clip': 21.389484405517578, 'timing/data': 0.004838298074901104, 'timing/model': 1.2903307270025834, 'learning_rate/qwen_vl_interface': 5.031506280519855e-07, 'learning_rate/action_model': 5.031506280519855e-06, 'epoch': 2.04}) +[2026-06-17 19:09:16,282][starVLA.training.train_starvla][INFO] - Step 6921, Loss: {'train/loss': 0.8322471976280212, 'train/grad_norm_pre_clip': 17.196800231933594, 'timing/data': 0.0045459900284186006, 'timing/model': 1.2014580179238692, 'learning_rate/qwen_vl_interface': 5.030723630754723e-07, 'learning_rate/action_model': 5.030723630754723e-06, 'epoch': 2.04}) +[2026-06-17 19:09:21,059][starVLA.training.train_starvla][INFO] - Step 6922, Loss: {'train/loss': 0.5012927651405334, 'train/grad_norm_pre_clip': 14.426398277282715, 'timing/data': 0.00514226290397346, 'timing/model': 1.202069299062714, 'learning_rate/qwen_vl_interface': 5.02995082142166e-07, 'learning_rate/action_model': 5.029950821421659e-06, 'epoch': 2.04}) +[2026-06-17 19:09:25,839][starVLA.training.train_starvla][INFO] - Step 6923, Loss: {'train/loss': 0.6364861726760864, 'train/grad_norm_pre_clip': 13.392081260681152, 'timing/data': 0.004734831978566945, 'timing/model': 1.2042121869744733, 'learning_rate/qwen_vl_interface': 5.029187852680869e-07, 'learning_rate/action_model': 5.029187852680868e-06, 'epoch': 2.04}) +[2026-06-17 19:09:30,625][starVLA.training.train_starvla][INFO] - Step 6924, Loss: {'train/loss': 0.7333555221557617, 'train/grad_norm_pre_clip': 13.774334907531738, 'timing/data': 0.004779729060828686, 'timing/model': 1.2032268811017275, 'learning_rate/qwen_vl_interface': 5.028434724690521e-07, 'learning_rate/action_model': 5.028434724690521e-06, 'epoch': 2.04}) +[2026-06-17 19:09:35,407][starVLA.training.train_starvla][INFO] - Step 6925, Loss: {'train/loss': 0.4948848783969879, 'train/grad_norm_pre_clip': 11.636761665344238, 'timing/data': 0.0049234830075874925, 'timing/model': 1.335964241065085, 'learning_rate/qwen_vl_interface': 5.027691437606734e-07, 'learning_rate/action_model': 5.027691437606733e-06, 'epoch': 2.04}) +[2026-06-17 19:09:40,057][starVLA.training.train_starvla][INFO] - Step 6926, Loss: {'train/loss': 0.7453548908233643, 'train/grad_norm_pre_clip': 13.753779411315918, 'timing/data': 0.0046788930194452405, 'timing/model': 1.2068948850501329, 'learning_rate/qwen_vl_interface': 5.026957991583597e-07, 'learning_rate/action_model': 5.0269579915835965e-06, 'epoch': 2.04}) +[2026-06-17 19:09:44,853][starVLA.training.train_starvla][INFO] - Step 6927, Loss: {'train/loss': 0.526753306388855, 'train/grad_norm_pre_clip': 13.108284950256348, 'timing/data': 0.004857674008235335, 'timing/model': 1.2091557669918984, 'learning_rate/qwen_vl_interface': 5.026234386773145e-07, 'learning_rate/action_model': 5.026234386773145e-06, 'epoch': 2.04}) +[2026-06-17 19:09:49,640][starVLA.training.train_starvla][INFO] - Step 6928, Loss: {'train/loss': 0.4430985450744629, 'train/grad_norm_pre_clip': 12.011842727661133, 'timing/data': 0.004980406956747174, 'timing/model': 1.2044835379347205, 'learning_rate/qwen_vl_interface': 5.025520623325398e-07, 'learning_rate/action_model': 5.025520623325397e-06, 'epoch': 2.04}) +[2026-06-17 19:09:54,357][starVLA.training.train_starvla][INFO] - Step 6929, Loss: {'train/loss': 0.9736658930778503, 'train/grad_norm_pre_clip': 26.749589920043945, 'timing/data': 0.004755766945891082, 'timing/model': 1.2073992830701172, 'learning_rate/qwen_vl_interface': 5.024816701388302e-07, 'learning_rate/action_model': 5.024816701388302e-06, 'epoch': 2.04}) +[2026-06-17 19:09:59,084][starVLA.training.train_starvla][INFO] - Step 6930, Loss: {'train/loss': 0.4228222072124481, 'train/grad_norm_pre_clip': 12.894302368164062, 'timing/data': 0.0044270160142332315, 'timing/model': 1.2046707840636373, 'learning_rate/qwen_vl_interface': 5.024122621107795e-07, 'learning_rate/action_model': 5.024122621107795e-06, 'epoch': 2.04}) +[2026-06-17 19:10:03,859][starVLA.training.train_starvla][INFO] - Step 6931, Loss: {'train/loss': 0.6470484733581543, 'train/grad_norm_pre_clip': 15.266830444335938, 'timing/data': 0.0045487870229408145, 'timing/model': 1.3342966519994661, 'learning_rate/qwen_vl_interface': 5.023438382627749e-07, 'learning_rate/action_model': 5.0234383826277486e-06, 'epoch': 2.04}) +[2026-06-17 19:10:08,503][starVLA.training.train_starvla][INFO] - Step 6932, Loss: {'train/loss': 0.3794075846672058, 'train/grad_norm_pre_clip': 9.976678848266602, 'timing/data': 0.00468828494194895, 'timing/model': 1.2033001179806888, 'learning_rate/qwen_vl_interface': 5.022763986090018e-07, 'learning_rate/action_model': 5.022763986090018e-06, 'epoch': 2.04}) +[2026-06-17 19:10:13,278][starVLA.training.train_starvla][INFO] - Step 6933, Loss: {'train/loss': 0.7766841650009155, 'train/grad_norm_pre_clip': 16.127092361450195, 'timing/data': 0.004838189925067127, 'timing/model': 1.2049519179854542, 'learning_rate/qwen_vl_interface': 5.022099431634398e-07, 'learning_rate/action_model': 5.022099431634398e-06, 'epoch': 2.04}) +[2026-06-17 19:10:18,056][starVLA.training.train_starvla][INFO] - Step 6934, Loss: {'train/loss': 0.5739988088607788, 'train/grad_norm_pre_clip': 15.011964797973633, 'timing/data': 0.005069449893198907, 'timing/model': 1.207677389960736, 'learning_rate/qwen_vl_interface': 5.021444719398656e-07, 'learning_rate/action_model': 5.021444719398656e-06, 'epoch': 2.04}) +[2026-06-17 19:10:22,844][starVLA.training.train_starvla][INFO] - Step 6935, Loss: {'train/loss': 0.6766390204429626, 'train/grad_norm_pre_clip': 15.246604919433594, 'timing/data': 0.004827806958928704, 'timing/model': 1.2082468098960817, 'learning_rate/qwen_vl_interface': 5.020799849518513e-07, 'learning_rate/action_model': 5.020799849518513e-06, 'epoch': 2.04}) +[2026-06-17 19:10:27,577][starVLA.training.train_starvla][INFO] - Step 6936, Loss: {'train/loss': 0.5225746035575867, 'train/grad_norm_pre_clip': 13.604704856872559, 'timing/data': 0.004687994020059705, 'timing/model': 1.203208725899458, 'learning_rate/qwen_vl_interface': 5.020164822127652e-07, 'learning_rate/action_model': 5.020164822127652e-06, 'epoch': 2.04}) +[2026-06-17 19:10:32,381][starVLA.training.train_starvla][INFO] - Step 6937, Loss: {'train/loss': 0.49049609899520874, 'train/grad_norm_pre_clip': 12.45286750793457, 'timing/data': 0.004656801000237465, 'timing/model': 1.3302750610746443, 'learning_rate/qwen_vl_interface': 5.019539637357715e-07, 'learning_rate/action_model': 5.019539637357714e-06, 'epoch': 2.04}) +[2026-06-17 19:10:37,057][starVLA.training.train_starvla][INFO] - Step 6938, Loss: {'train/loss': 0.5471847653388977, 'train/grad_norm_pre_clip': 19.84493064880371, 'timing/data': 0.005335870082490146, 'timing/model': 1.2090520600322634, 'learning_rate/qwen_vl_interface': 5.0189242953383e-07, 'learning_rate/action_model': 5.018924295338299e-06, 'epoch': 2.04}) +[2026-06-17 19:10:41,844][starVLA.training.train_starvla][INFO] - Step 6939, Loss: {'train/loss': 0.5364778637886047, 'train/grad_norm_pre_clip': 15.616619110107422, 'timing/data': 0.004780043032951653, 'timing/model': 1.2066615760559216, 'learning_rate/qwen_vl_interface': 5.018318796196967e-07, 'learning_rate/action_model': 5.018318796196967e-06, 'epoch': 2.04}) +[2026-06-17 19:10:46,623][starVLA.training.train_starvla][INFO] - Step 6940, Loss: {'train/loss': 0.7912482023239136, 'train/grad_norm_pre_clip': 14.583009719848633, 'timing/data': 0.0049904880579560995, 'timing/model': 1.2035205779829994, 'learning_rate/qwen_vl_interface': 5.017723140059243e-07, 'learning_rate/action_model': 5.017723140059242e-06, 'epoch': 2.04}) +[2026-06-17 19:10:51,406][starVLA.training.train_starvla][INFO] - Step 6941, Loss: {'train/loss': 0.5824246406555176, 'train/grad_norm_pre_clip': 13.551795959472656, 'timing/data': 0.004772608983330429, 'timing/model': 1.2055440130643547, 'learning_rate/qwen_vl_interface': 5.017137327048606e-07, 'learning_rate/action_model': 5.017137327048606e-06, 'epoch': 2.04}) +[2026-06-17 19:10:56,199][starVLA.training.train_starvla][INFO] - Step 6942, Loss: {'train/loss': 0.2988846004009247, 'train/grad_norm_pre_clip': 8.592209815979004, 'timing/data': 0.004636093974113464, 'timing/model': 1.337364293052815, 'learning_rate/qwen_vl_interface': 5.016561357286498e-07, 'learning_rate/action_model': 5.016561357286498e-06, 'epoch': 2.04}) +[2026-06-17 19:11:00,857][starVLA.training.train_starvla][INFO] - Step 6943, Loss: {'train/loss': 0.38064044713974, 'train/grad_norm_pre_clip': 13.138916969299316, 'timing/data': 0.0048273560823872685, 'timing/model': 1.2042683380423114, 'learning_rate/qwen_vl_interface': 5.015995230892312e-07, 'learning_rate/action_model': 5.015995230892311e-06, 'epoch': 2.04}) +[2026-06-17 19:11:05,657][starVLA.training.train_starvla][INFO] - Step 6944, Loss: {'train/loss': 0.5455801486968994, 'train/grad_norm_pre_clip': 14.977700233459473, 'timing/data': 0.005437188083305955, 'timing/model': 1.2047205830458552, 'learning_rate/qwen_vl_interface': 5.015438947983409e-07, 'learning_rate/action_model': 5.015438947983409e-06, 'epoch': 2.04}) +[2026-06-17 19:11:10,380][starVLA.training.train_starvla][INFO] - Step 6945, Loss: {'train/loss': 0.4377043843269348, 'train/grad_norm_pre_clip': 13.029411315917969, 'timing/data': 0.004812801955267787, 'timing/model': 1.2049712430452928, 'learning_rate/qwen_vl_interface': 5.014892508675108e-07, 'learning_rate/action_model': 5.014892508675107e-06, 'epoch': 2.04}) +[2026-06-17 19:11:15,112][starVLA.training.train_starvla][INFO] - Step 6946, Loss: {'train/loss': 0.6069168448448181, 'train/grad_norm_pre_clip': 15.476799011230469, 'timing/data': 0.004783403011970222, 'timing/model': 1.2048874649917707, 'learning_rate/qwen_vl_interface': 5.014355913080683e-07, 'learning_rate/action_model': 5.014355913080683e-06, 'epoch': 2.04}) +[2026-06-17 19:11:19,898][starVLA.training.train_starvla][INFO] - Step 6947, Loss: {'train/loss': 0.5964193344116211, 'train/grad_norm_pre_clip': 13.456542015075684, 'timing/data': 0.00496448390185833, 'timing/model': 1.2045440609799698, 'learning_rate/qwen_vl_interface': 5.013829161311377e-07, 'learning_rate/action_model': 5.013829161311377e-06, 'epoch': 2.04}) +[2026-06-17 19:11:24,696][starVLA.training.train_starvla][INFO] - Step 6948, Loss: {'train/loss': 0.6360580325126648, 'train/grad_norm_pre_clip': 20.470252990722656, 'timing/data': 0.0045485569862648845, 'timing/model': 1.3445613470394164, 'learning_rate/qwen_vl_interface': 5.013312253476383e-07, 'learning_rate/action_model': 5.0133122534763826e-06, 'epoch': 2.04}) +[2026-06-17 19:11:29,374][starVLA.training.train_starvla][INFO] - Step 6949, Loss: {'train/loss': 0.49483999609947205, 'train/grad_norm_pre_clip': 16.631452560424805, 'timing/data': 0.004939488950185478, 'timing/model': 1.2110482869902626, 'learning_rate/qwen_vl_interface': 5.012805189682861e-07, 'learning_rate/action_model': 5.01280518968286e-06, 'epoch': 2.04}) +[2026-06-17 19:11:34,164][starVLA.training.train_starvla][INFO] - Step 6950, Loss: {'train/loss': 0.5272963047027588, 'train/grad_norm_pre_clip': 17.935876846313477, 'timing/data': 0.004957617027685046, 'timing/model': 1.2090768771013245, 'learning_rate/qwen_vl_interface': 5.012307970035915e-07, 'learning_rate/action_model': 5.012307970035915e-06, 'epoch': 2.05}) +[2026-06-17 19:11:38,945][starVLA.training.train_starvla][INFO] - Step 6951, Loss: {'train/loss': 0.7950516939163208, 'train/grad_norm_pre_clip': 17.132043838500977, 'timing/data': 0.004900136962532997, 'timing/model': 1.2040377750527114, 'learning_rate/qwen_vl_interface': 5.011820594638626e-07, 'learning_rate/action_model': 5.0118205946386256e-06, 'epoch': 2.05}) +[2026-06-17 19:11:43,774][starVLA.training.train_starvla][INFO] - Step 6952, Loss: {'train/loss': 0.6793901324272156, 'train/grad_norm_pre_clip': 16.307905197143555, 'timing/data': 0.007069639046676457, 'timing/model': 1.2190753070171922, 'learning_rate/qwen_vl_interface': 5.011343063592032e-07, 'learning_rate/action_model': 5.011343063592032e-06, 'epoch': 2.05}) +[2026-06-17 19:11:48,570][starVLA.training.train_starvla][INFO] - Step 6953, Loss: {'train/loss': 0.4379948377609253, 'train/grad_norm_pre_clip': 14.12983512878418, 'timing/data': 0.004538527922704816, 'timing/model': 1.2097572350176051, 'learning_rate/qwen_vl_interface': 5.010875376995121e-07, 'learning_rate/action_model': 5.01087537699512e-06, 'epoch': 2.05}) +[2026-06-17 19:11:53,367][starVLA.training.train_starvla][INFO] - Step 6954, Loss: {'train/loss': 0.7237712144851685, 'train/grad_norm_pre_clip': 15.961709976196289, 'timing/data': 0.004976346041075885, 'timing/model': 1.336822405923158, 'learning_rate/qwen_vl_interface': 5.010417534944839e-07, 'learning_rate/action_model': 5.010417534944838e-06, 'epoch': 2.05}) +[2026-06-17 19:11:58,025][starVLA.training.train_starvla][INFO] - Step 6955, Loss: {'train/loss': 0.5776867270469666, 'train/grad_norm_pre_clip': 17.595170974731445, 'timing/data': 0.004799740039743483, 'timing/model': 1.2088155689416453, 'learning_rate/qwen_vl_interface': 5.009969537536106e-07, 'learning_rate/action_model': 5.009969537536106e-06, 'epoch': 2.05}) +[2026-06-17 19:12:02,808][starVLA.training.train_starvla][INFO] - Step 6956, Loss: {'train/loss': 0.40440502762794495, 'train/grad_norm_pre_clip': 9.91845989227295, 'timing/data': 0.004561024019494653, 'timing/model': 1.2052083150483668, 'learning_rate/qwen_vl_interface': 5.00953138486179e-07, 'learning_rate/action_model': 5.00953138486179e-06, 'epoch': 2.05}) +[2026-06-17 19:12:07,614][starVLA.training.train_starvla][INFO] - Step 6957, Loss: {'train/loss': 0.5139051675796509, 'train/grad_norm_pre_clip': 14.154655456542969, 'timing/data': 0.004761534975841641, 'timing/model': 1.2102786769391969, 'learning_rate/qwen_vl_interface': 5.009103077012723e-07, 'learning_rate/action_model': 5.009103077012723e-06, 'epoch': 2.05}) +[2026-06-17 19:12:12,465][starVLA.training.train_starvla][INFO] - Step 6958, Loss: {'train/loss': 0.8343313932418823, 'train/grad_norm_pre_clip': 16.090539932250977, 'timing/data': 0.004539435030892491, 'timing/model': 1.2036027220310643, 'learning_rate/qwen_vl_interface': 5.008684614077684e-07, 'learning_rate/action_model': 5.008684614077684e-06, 'epoch': 2.05}) +[2026-06-17 19:12:17,141][starVLA.training.train_starvla][INFO] - Step 6959, Loss: {'train/loss': 0.4798514246940613, 'train/grad_norm_pre_clip': 12.32882308959961, 'timing/data': 0.005300019052810967, 'timing/model': 1.2041613680776209, 'learning_rate/qwen_vl_interface': 5.008275996143429e-07, 'learning_rate/action_model': 5.0082759961434295e-06, 'epoch': 2.05}) +[2026-06-17 19:12:21,854][starVLA.training.train_starvla][INFO] - Step 6960, Loss: {'train/loss': 0.5611262917518616, 'train/grad_norm_pre_clip': 11.512001037597656, 'timing/data': 0.004770076950080693, 'timing/model': 1.2433461899636313, 'learning_rate/qwen_vl_interface': 5.007877223294668e-07, 'learning_rate/action_model': 5.0078772232946675e-06, 'epoch': 2.05}) +[2026-06-17 19:12:26,511][starVLA.training.train_starvla][INFO] - Step 6961, Loss: {'train/loss': 0.48535922169685364, 'train/grad_norm_pre_clip': 11.999825477600098, 'timing/data': 0.004699566983617842, 'timing/model': 1.2057863960508257, 'learning_rate/qwen_vl_interface': 5.007488295614058e-07, 'learning_rate/action_model': 5.007488295614058e-06, 'epoch': 2.05}) +[2026-06-17 19:12:31,256][starVLA.training.train_starvla][INFO] - Step 6962, Loss: {'train/loss': 0.6197817921638489, 'train/grad_norm_pre_clip': 16.656187057495117, 'timing/data': 0.004727176972664893, 'timing/model': 1.205286356038414, 'learning_rate/qwen_vl_interface': 5.007109213182231e-07, 'learning_rate/action_model': 5.007109213182231e-06, 'epoch': 2.05}) +[2026-06-17 19:12:36,040][starVLA.training.train_starvla][INFO] - Step 6963, Loss: {'train/loss': 0.48166990280151367, 'train/grad_norm_pre_clip': 12.573004722595215, 'timing/data': 0.004473809036426246, 'timing/model': 1.2048571669729427, 'learning_rate/qwen_vl_interface': 5.006739976077769e-07, 'learning_rate/action_model': 5.0067399760777695e-06, 'epoch': 2.05}) +[2026-06-17 19:12:40,820][starVLA.training.train_starvla][INFO] - Step 6964, Loss: {'train/loss': 0.25544318556785583, 'train/grad_norm_pre_clip': 8.551338195800781, 'timing/data': 0.004832334001548588, 'timing/model': 1.2023590269964188, 'learning_rate/qwen_vl_interface': 5.006380584377213e-07, 'learning_rate/action_model': 5.006380584377213e-06, 'epoch': 2.05}) +[2026-06-17 19:12:45,602][starVLA.training.train_starvla][INFO] - Step 6965, Loss: {'train/loss': 0.5991365909576416, 'train/grad_norm_pre_clip': 13.437129974365234, 'timing/data': 0.004742933087982237, 'timing/model': 1.3378474499331787, 'learning_rate/qwen_vl_interface': 5.006031038155068e-07, 'learning_rate/action_model': 5.006031038155067e-06, 'epoch': 2.05}) +[2026-06-17 19:12:50,375][starVLA.training.train_starvla][INFO] - Step 6966, Loss: {'train/loss': 0.5425439476966858, 'train/grad_norm_pre_clip': 15.041507720947266, 'timing/data': 0.0049718749942258, 'timing/model': 1.3245712380157784, 'learning_rate/qwen_vl_interface': 5.00569133748379e-07, 'learning_rate/action_model': 5.00569133748379e-06, 'epoch': 2.05}) +[2026-06-17 19:12:55,030][starVLA.training.train_starvla][INFO] - Step 6967, Loss: {'train/loss': 0.686939001083374, 'train/grad_norm_pre_clip': 15.699053764343262, 'timing/data': 0.004878210020251572, 'timing/model': 1.2080174899892882, 'learning_rate/qwen_vl_interface': 5.005361482433811e-07, 'learning_rate/action_model': 5.005361482433811e-06, 'epoch': 2.05}) +[2026-06-17 19:12:59,777][starVLA.training.train_starvla][INFO] - Step 6968, Loss: {'train/loss': 0.6621568202972412, 'train/grad_norm_pre_clip': 16.24074935913086, 'timing/data': 0.005302538047544658, 'timing/model': 1.2031119980383664, 'learning_rate/qwen_vl_interface': 5.005041473073501e-07, 'learning_rate/action_model': 5.005041473073501e-06, 'epoch': 2.05}) +[2026-06-17 19:13:04,559][starVLA.training.train_starvla][INFO] - Step 6969, Loss: {'train/loss': 0.44098353385925293, 'train/grad_norm_pre_clip': 12.16800308227539, 'timing/data': 0.004936552024446428, 'timing/model': 1.2039025509729981, 'learning_rate/qwen_vl_interface': 5.004731309469201e-07, 'learning_rate/action_model': 5.004731309469201e-06, 'epoch': 2.05}) +[2026-06-17 19:13:09,339][starVLA.training.train_starvla][INFO] - Step 6970, Loss: {'train/loss': 0.49768662452697754, 'train/grad_norm_pre_clip': 21.494131088256836, 'timing/data': 0.004789622966200113, 'timing/model': 1.2042288390221074, 'learning_rate/qwen_vl_interface': 5.004430991685201e-07, 'learning_rate/action_model': 5.0044309916852016e-06, 'epoch': 2.05}) +[2026-06-17 19:13:14,194][starVLA.training.train_starvla][INFO] - Step 6971, Loss: {'train/loss': 0.6479743123054504, 'train/grad_norm_pre_clip': 16.080623626708984, 'timing/data': 0.004936357028782368, 'timing/model': 1.3364252530736849, 'learning_rate/qwen_vl_interface': 5.004140519783773e-07, 'learning_rate/action_model': 5.0041405197837725e-06, 'epoch': 2.05}) +[2026-06-17 19:13:18,968][starVLA.training.train_starvla][INFO] - Step 6972, Loss: {'train/loss': 0.46069034934043884, 'train/grad_norm_pre_clip': 11.269515991210938, 'timing/data': 0.005074798944406211, 'timing/model': 1.3252671610098332, 'learning_rate/qwen_vl_interface': 5.003859893825117e-07, 'learning_rate/action_model': 5.003859893825117e-06, 'epoch': 2.05}) +[2026-06-17 19:13:23,626][starVLA.training.train_starvla][INFO] - Step 6973, Loss: {'train/loss': 0.5779542326927185, 'train/grad_norm_pre_clip': 12.67432975769043, 'timing/data': 0.004993907990865409, 'timing/model': 1.2046704860404134, 'learning_rate/qwen_vl_interface': 5.003589113867419e-07, 'learning_rate/action_model': 5.003589113867419e-06, 'epoch': 2.05}) +[2026-06-17 19:13:28,390][starVLA.training.train_starvla][INFO] - Step 6974, Loss: {'train/loss': 0.6918873190879822, 'train/grad_norm_pre_clip': 16.211095809936523, 'timing/data': 0.0048054540529847145, 'timing/model': 1.203535641077906, 'learning_rate/qwen_vl_interface': 5.003328179966799e-07, 'learning_rate/action_model': 5.003328179966798e-06, 'epoch': 2.05}) +[2026-06-17 19:13:33,172][starVLA.training.train_starvla][INFO] - Step 6975, Loss: {'train/loss': 0.6410455703735352, 'train/grad_norm_pre_clip': 15.446154594421387, 'timing/data': 0.005214834935031831, 'timing/model': 1.2036973300855607, 'learning_rate/qwen_vl_interface': 5.003077092177363e-07, 'learning_rate/action_model': 5.003077092177363e-06, 'epoch': 2.05}) +[2026-06-17 19:13:37,960][starVLA.training.train_starvla][INFO] - Step 6976, Loss: {'train/loss': 0.5119893550872803, 'train/grad_norm_pre_clip': 12.471556663513184, 'timing/data': 0.005055018933489919, 'timing/model': 1.2061516599496827, 'learning_rate/qwen_vl_interface': 5.002835850551146e-07, 'learning_rate/action_model': 5.002835850551146e-06, 'epoch': 2.05}) +[2026-06-17 19:13:42,693][starVLA.training.train_starvla][INFO] - Step 6977, Loss: {'train/loss': 0.44325998425483704, 'train/grad_norm_pre_clip': 9.291096687316895, 'timing/data': 0.004593300982378423, 'timing/model': 1.2794068810762838, 'learning_rate/qwen_vl_interface': 5.002604455138174e-07, 'learning_rate/action_model': 5.002604455138173e-06, 'epoch': 2.05}) +[2026-06-17 19:13:47,352][starVLA.training.train_starvla][INFO] - Step 6978, Loss: {'train/loss': 0.9486052989959717, 'train/grad_norm_pre_clip': 17.43181610107422, 'timing/data': 0.004566713003441691, 'timing/model': 1.2049570410745218, 'learning_rate/qwen_vl_interface': 5.002382905986407e-07, 'learning_rate/action_model': 5.002382905986407e-06, 'epoch': 2.05}) +[2026-06-17 19:13:52,117][starVLA.training.train_starvla][INFO] - Step 6979, Loss: {'train/loss': 0.700844943523407, 'train/grad_norm_pre_clip': 16.347414016723633, 'timing/data': 0.004321228014305234, 'timing/model': 1.205523020005785, 'learning_rate/qwen_vl_interface': 5.002171203141771e-07, 'learning_rate/action_model': 5.0021712031417704e-06, 'epoch': 2.05}) +[2026-06-17 19:13:56,909][starVLA.training.train_starvla][INFO] - Step 6980, Loss: {'train/loss': 0.6686419248580933, 'train/grad_norm_pre_clip': 15.691611289978027, 'timing/data': 0.004660524078644812, 'timing/model': 1.2064734500600025, 'learning_rate/qwen_vl_interface': 5.001969346648154e-07, 'learning_rate/action_model': 5.001969346648154e-06, 'epoch': 2.05}) +[2026-06-17 19:14:01,695][starVLA.training.train_starvla][INFO] - Step 6981, Loss: {'train/loss': 0.4665934443473816, 'train/grad_norm_pre_clip': 12.110062599182129, 'timing/data': 0.0052838369738310575, 'timing/model': 1.2052743149688467, 'learning_rate/qwen_vl_interface': 5.0017773365474e-07, 'learning_rate/action_model': 5.001777336547399e-06, 'epoch': 2.05}) +[2026-06-17 19:14:06,482][starVLA.training.train_starvla][INFO] - Step 6982, Loss: {'train/loss': 0.7150759100914001, 'train/grad_norm_pre_clip': 27.192432403564453, 'timing/data': 0.005175242084078491, 'timing/model': 1.3307479759678245, 'learning_rate/qwen_vl_interface': 5.001595172879317e-07, 'learning_rate/action_model': 5.001595172879316e-06, 'epoch': 2.05}) +[2026-06-17 19:14:11,232][starVLA.training.train_starvla][INFO] - Step 6983, Loss: {'train/loss': 0.6899479627609253, 'train/grad_norm_pre_clip': 15.333488464355469, 'timing/data': 0.004933029063977301, 'timing/model': 1.2921318500302732, 'learning_rate/qwen_vl_interface': 5.001422855681669e-07, 'learning_rate/action_model': 5.001422855681669e-06, 'epoch': 2.05}) +[2026-06-17 19:14:15,899][starVLA.training.train_starvla][INFO] - Step 6984, Loss: {'train/loss': 0.4930356442928314, 'train/grad_norm_pre_clip': 14.177042007446289, 'timing/data': 0.004869223921559751, 'timing/model': 1.2071536020375788, 'learning_rate/qwen_vl_interface': 5.001260384990169e-07, 'learning_rate/action_model': 5.001260384990169e-06, 'epoch': 2.06}) +[2026-06-17 19:14:20,688][starVLA.training.train_starvla][INFO] - Step 6985, Loss: {'train/loss': 0.7345147132873535, 'train/grad_norm_pre_clip': 14.745848655700684, 'timing/data': 0.0049833268858492374, 'timing/model': 1.2058445389848202, 'learning_rate/qwen_vl_interface': 5.001107760838504e-07, 'learning_rate/action_model': 5.001107760838504e-06, 'epoch': 2.06}) +[2026-06-17 19:14:25,508][starVLA.training.train_starvla][INFO] - Step 6986, Loss: {'train/loss': 0.6048398613929749, 'train/grad_norm_pre_clip': 13.149224281311035, 'timing/data': 0.004881203989498317, 'timing/model': 1.2063961080275476, 'learning_rate/qwen_vl_interface': 5.000964983258306e-07, 'learning_rate/action_model': 5.000964983258306e-06, 'epoch': 2.06}) +[2026-06-17 19:14:30,289][starVLA.training.train_starvla][INFO] - Step 6987, Loss: {'train/loss': 0.422983318567276, 'train/grad_norm_pre_clip': 12.539600372314453, 'timing/data': 0.004461335949599743, 'timing/model': 1.2043251369614154, 'learning_rate/qwen_vl_interface': 5.000832052279185e-07, 'learning_rate/action_model': 5.000832052279185e-06, 'epoch': 2.06}) +[2026-06-17 19:14:35,075][starVLA.training.train_starvla][INFO] - Step 6988, Loss: {'train/loss': 0.6214250922203064, 'train/grad_norm_pre_clip': 15.54171371459961, 'timing/data': 0.004586725030094385, 'timing/model': 1.2042420690413564, 'learning_rate/qwen_vl_interface': 5.000708967928686e-07, 'learning_rate/action_model': 5.000708967928685e-06, 'epoch': 2.06}) +[2026-06-17 19:14:39,822][starVLA.training.train_starvla][INFO] - Step 6989, Loss: {'train/loss': 0.5523810982704163, 'train/grad_norm_pre_clip': 13.83761215209961, 'timing/data': 0.004791131941601634, 'timing/model': 1.2906148580368608, 'learning_rate/qwen_vl_interface': 5.00059573023233e-07, 'learning_rate/action_model': 5.0005957302323305e-06, 'epoch': 2.06}) +[2026-06-17 19:14:44,484][starVLA.training.train_starvla][INFO] - Step 6990, Loss: {'train/loss': 0.40664801001548767, 'train/grad_norm_pre_clip': 12.487932205200195, 'timing/data': 0.0048836099449545145, 'timing/model': 1.2051422629738227, 'learning_rate/qwen_vl_interface': 5.000492339213594e-07, 'learning_rate/action_model': 5.000492339213594e-06, 'epoch': 2.06}) +[2026-06-17 19:14:49,275][starVLA.training.train_starvla][INFO] - Step 6991, Loss: {'train/loss': 0.4294879734516144, 'train/grad_norm_pre_clip': 12.172850608825684, 'timing/data': 0.005042264936491847, 'timing/model': 1.2055545289767906, 'learning_rate/qwen_vl_interface': 5.000398794893907e-07, 'learning_rate/action_model': 5.000398794893906e-06, 'epoch': 2.06}) +[2026-06-17 19:14:54,062][starVLA.training.train_starvla][INFO] - Step 6992, Loss: {'train/loss': 0.5361199975013733, 'train/grad_norm_pre_clip': 12.495064735412598, 'timing/data': 0.0048936649691313505, 'timing/model': 1.205177603987977, 'learning_rate/qwen_vl_interface': 5.000315097292659e-07, 'learning_rate/action_model': 5.000315097292659e-06, 'epoch': 2.06}) +[2026-06-17 19:14:58,849][starVLA.training.train_starvla][INFO] - Step 6993, Loss: {'train/loss': 0.7739320993423462, 'train/grad_norm_pre_clip': 17.360448837280273, 'timing/data': 0.004855253966525197, 'timing/model': 1.2090626100543886, 'learning_rate/qwen_vl_interface': 5.000241246427207e-07, 'learning_rate/action_model': 5.000241246427207e-06, 'epoch': 2.06}) +[2026-06-17 19:15:03,645][starVLA.training.train_starvla][INFO] - Step 6994, Loss: {'train/loss': 0.7818724513053894, 'train/grad_norm_pre_clip': 18.176225662231445, 'timing/data': 0.004690272966399789, 'timing/model': 1.335455487947911, 'learning_rate/qwen_vl_interface': 5.000177242312855e-07, 'learning_rate/action_model': 5.000177242312855e-06, 'epoch': 2.06}) +[2026-06-17 19:15:08,306][starVLA.training.train_starvla][INFO] - Step 6995, Loss: {'train/loss': 0.41784635186195374, 'train/grad_norm_pre_clip': 12.463593482971191, 'timing/data': 0.0051365050021559, 'timing/model': 1.2066146659199148, 'learning_rate/qwen_vl_interface': 5.00012308496287e-07, 'learning_rate/action_model': 5.000123084962871e-06, 'epoch': 2.06}) +[2026-06-17 19:15:13,108][starVLA.training.train_starvla][INFO] - Step 6996, Loss: {'train/loss': 0.6218957304954529, 'train/grad_norm_pre_clip': 13.525636672973633, 'timing/data': 0.004967277985997498, 'timing/model': 1.2052395759383217, 'learning_rate/qwen_vl_interface': 5.000078774388487e-07, 'learning_rate/action_model': 5.000078774388487e-06, 'epoch': 2.06}) +[2026-06-17 19:15:17,891][starVLA.training.train_starvla][INFO] - Step 6997, Loss: {'train/loss': 0.4476378858089447, 'train/grad_norm_pre_clip': 10.08072566986084, 'timing/data': 0.005039948038756847, 'timing/model': 1.2047384369652718, 'learning_rate/qwen_vl_interface': 5.000044310598879e-07, 'learning_rate/action_model': 5.00004431059888e-06, 'epoch': 2.06}) +[2026-06-17 19:15:22,657][starVLA.training.train_starvla][INFO] - Step 6998, Loss: {'train/loss': 0.6650476455688477, 'train/grad_norm_pre_clip': 16.073902130126953, 'timing/data': 0.004875829094089568, 'timing/model': 1.219640456023626, 'learning_rate/qwen_vl_interface': 5.000019693601204e-07, 'learning_rate/action_model': 5.000019693601204e-06, 'epoch': 2.06}) +[2026-06-17 19:15:27,416][starVLA.training.train_starvla][INFO] - Step 6999, Loss: {'train/loss': 0.5857092142105103, 'train/grad_norm_pre_clip': 15.386381149291992, 'timing/data': 0.005146601935848594, 'timing/model': 1.2035861449548975, 'learning_rate/qwen_vl_interface': 5.000004923400556e-07, 'learning_rate/action_model': 5.0000049234005565e-06, 'epoch': 2.06}) +[2026-06-17 19:25:36,640][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 7000 +[2026-06-17 19:25:36,641][starVLA.training.train_starvla][INFO] - Step 7000, Loss: {'train/loss': 0.7526401877403259, 'train/grad_norm_pre_clip': 15.97942066192627, 'eval/loss': 0.7032653093338013, 'eval/latency_0/loss': 0.7032653093338013, 'eval/demon_fire_precision': 0.8988142292490119, 'eval/demon_fire_recall': 0.9443521594684385, 'eval/demon_fire_f1': 0.9210206561360875, 'eval/demon_fire_rate_model': 0.703559510567297, 'eval/demon_fire_rate_teacher': 0.6696329254727474, 'eval/demon_left_precision': 0.8275469550369948, 'eval/demon_left_recall': 0.8237960339943343, 'eval/demon_left_f1': 0.8256672345258376, 'eval/demon_left_rate_model': 0.48859844271412683, 'eval/demon_left_rate_teacher': 0.49082313681868744, 'eval/demon_right_precision': 0.8279095421952565, 'eval/demon_right_recall': 0.8480225988700565, 'eval/demon_right_f1': 0.8378453809656712, 'eval/demon_right_rate_model': 0.5041713014460512, 'eval/demon_right_rate_teacher': 0.4922135706340378, 'eval/demon_move_f1': 0.8317563077457544, 'eval/demon_cc_f1': 0.892292328196144, 'eval/latency_0/demon_fire_precision': 0.8988142292490119, 'eval/latency_0/demon_fire_recall': 0.9443521594684385, 'eval/latency_0/demon_fire_f1': 0.9210206561360875, 'eval/latency_0/demon_fire_rate_model': 0.703559510567297, 'eval/latency_0/demon_fire_rate_teacher': 0.6696329254727474, 'eval/latency_0/demon_left_precision': 0.8275469550369948, 'eval/latency_0/demon_left_recall': 0.8237960339943343, 'eval/latency_0/demon_left_f1': 0.8256672345258376, 'eval/latency_0/demon_left_rate_model': 0.48859844271412683, 'eval/latency_0/demon_left_rate_teacher': 0.49082313681868744, 'eval/latency_0/demon_right_precision': 0.8279095421952565, 'eval/latency_0/demon_right_recall': 0.8480225988700565, 'eval/latency_0/demon_right_f1': 0.8378453809656712, 'eval/latency_0/demon_right_rate_model': 0.5041713014460512, 'eval/latency_0/demon_right_rate_teacher': 0.4922135706340378, 'eval/latency_0/demon_move_f1': 0.8317563077457544, 'eval/latency_0/demon_cc_f1': 0.892292328196144, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 7320.0, 'rl_games_eval/mid_train/mean_length': 3600.0, 'rl_games_eval/mid_train/std_reward': 610.1393283505006, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/macro_mean_reward': 7320.0, 'rl_games_eval/mid_train/macro_mean_length': 3600.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/demon_attack__latency_0/mean_reward': 7320.0, 'rl_games_eval/mid_train/demon_attack__latency_0/mean_length': 3600.0, 'rl_games_eval/mid_train/demon_attack__latency_0/std_reward': 610.1393283505006, 'rl_games_eval/mid_train/demon_attack__latency_0/std_length': 0.0, 'rl_games_eval/mid_train/demon_attack__latency_0/num_episodes': 5.0, 'checkpoint/current_eval_score': 7320.0, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.004834368941374123, 'timing/model': 1.3350491529563442, 'learning_rate/qwen_vl_interface': 5e-07, 'learning_rate/action_model': 4.9999999999999996e-06, 'epoch': 2.06}) +[2026-06-17 19:25:36,641][accelerate.accelerator][INFO] - Saving current state to /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_7000_state +[2026-06-17 19:25:36,643][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-17 19:25:40,200][accelerate.checkpointing][INFO] - Model weights saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_7000_state/model.safetensors +[2026-06-17 19:25:47,798][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_7000_state/optimizer.bin +[2026-06-17 19:25:47,799][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_7000_state/sampler.bin +[2026-06-17 19:25:47,799][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_7000_state/sampler_1.bin +[2026-06-17 19:25:47,800][accelerate.checkpointing][INFO] - Random states saved in /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints/steps_7000_state/random_states_0.pkl +[2026-06-17 19:25:47,848][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 7000 +[2026-06-17 19:25:47,848][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-17 19:25:47,858][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-17 19:25:47,858][starVLA.training.train_starvla][INFO] - Training complete. Final training state is saved under /workspace/starVLA/results/Checkpoints/demon_attack_fix_latency_0_200ep/checkpoints +[2026-06-17 19:25:51,014][starVLA.training.train_starvla][INFO] - ... and that's all, folks! diff --git a/zero-latency/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_0_200ep/summary.jsonl b/zero-latency/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_0_200ep/summary.jsonl new file mode 100644 index 0000000000000000000000000000000000000000..67f4020ac3871f475553f86f0c342379cb905502 --- /dev/null +++ b/zero-latency/demon-attack/vla/starvla-qwenoft-h1/demon_attack_fix_latency_0_200ep/summary.jsonl @@ -0,0 +1,70 @@ +{"steps": 100} +{"steps": 200} +{"steps": 300} +{"steps": 400} +{"steps": 500} +{"steps": 600} +{"steps": 700} +{"steps": 800} +{"steps": 900} +{"steps": 1000} +{"steps": 1100} +{"steps": 1200} +{"steps": 1300} +{"steps": 1400} +{"steps": 1500} +{"steps": 1600} +{"steps": 1700} +{"steps": 1800} +{"steps": 1900} +{"steps": 2000} +{"steps": 2100} +{"steps": 2200} +{"steps": 2300} +{"steps": 2400} +{"steps": 2500} +{"steps": 2600} +{"steps": 2700} +{"steps": 2800} +{"steps": 2900} +{"steps": 3000} +{"steps": 3100} +{"steps": 3200} +{"steps": 3300} +{"steps": 3400} +{"steps": 3500} +{"steps": 3600} +{"steps": 3700} +{"steps": 3800} +{"steps": 3900} +{"steps": 4000} +{"steps": 4100} +{"steps": 4200} +{"steps": 4300} +{"steps": 4400} +{"steps": 4500} +{"steps": 4600} +{"steps": 4700} +{"steps": 4800} +{"steps": 4900} +{"steps": 5000} +{"steps": 5100} +{"steps": 5200} +{"steps": 5300} +{"steps": 5400} +{"steps": 5500} +{"steps": 5600} +{"steps": 5700} +{"steps": 5800} +{"steps": 5900} +{"steps": 6000} +{"steps": 6100} +{"steps": 6200} +{"steps": 6300} +{"steps": 6400} +{"steps": 6500} +{"steps": 6600} +{"steps": 6700} +{"steps": 6800} +{"steps": 6900} +{"steps": 7000} diff --git a/zero-latency/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_0_200ep/checkpoints/steps_5000_model.safetensors b/zero-latency/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_0_200ep/checkpoints/steps_5000_model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..366682e0990b47bda7a22d5a2305b2160158c61a --- /dev/null +++ b/zero-latency/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_0_200ep/checkpoints/steps_5000_model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:7e22c0b5c6118ff5bac5536d076ef7c8a51cf14cfa8b86752464321ac7311d39 +size 9138230516 diff --git a/zero-latency/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_0_200ep/config.full.yaml b/zero-latency/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_0_200ep/config.full.yaml new file mode 100644 index 0000000000000000000000000000000000000000..326023b93e0478738a4fda22095cbcdcca80cefc --- /dev/null +++ b/zero-latency/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_0_200ep/config.full.yaml @@ -0,0 +1,198 @@ +framework: + name: QwenOFT + qwenvl: + base_vlm: /workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + attn_implementation: flash_attention_2 + enable_gradient_checkpointing: true + action_model: + action_model_type: MLP + action_dim: 7 + action_hidden_dim: 2560 + future_action_window_size: 0 + past_action_window_size: 0 + loss_type: discrete_ce + state_dim: 7 + action_horizon: 1 + action_env_dim: 2 +datasets: + vla_data: + dataset_py: lerobot_datasets + include_state: true + data_root_dir: /workspace/starVLA/data/flappy_fix_latency_0_200ep + data_mix: flappy_train__bridge + eval_data_mix: flappy_train__bridge__val + custom_mixtures_path: null + action_type: discrete + sequential_step_sampling: false + shuffle: true + action_balance: + enabled: false + strategy: balanced_epoch + action_key: action_id + target_flap_fraction: 0.3 + noop_id: 0 + flap_id: 1 + per_device_batch_size: 32 + load_all_data_for_training: true + obs_image_size: null + video_backend: torchvision_av +dataset: + source_hf: '' + config_name: null + converted_name: flappy_train + single_source_hf: '' + mixed_source_hf: '' + single_converted_name: flappy_train + mixed_converted_name: flappy_mixed_latency_train + single_latency_filter: null + mixed_latency_filter: null + force_download: false + setup_force: false + verify_rows: 200 + max_episodes: null + episodes_per_latency: null + latency_filter: null + debug_subset: + enabled: false + max_episodes: 5 + suffix: debug +base_model: + repo_id: Qwen/Qwen3-VL-4B-Instruct +initialization: + checkpoint_local_dir: playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_hf_repo_id: StarVLA/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_filename: checkpoints/steps_5000_pytorch_model.pt +trainer: + max_train_steps: 5000 + num_warmup_steps: 100 + save_interval: 100 + eval_interval: 100 + eval_num_batches: 100 + eval_action_classification: true + cc_f1_tolerance: 1 + learning_rate: + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + action_model: 0.0001 + lr_scheduler_type: cosine_with_min_lr + scheduler_specific_kwargs: + min_lr: 1.0e-06 + freeze_modules: '' + loss_scale: + vla: 1.0 + vlm: 0.1 + max_grad_norm: 1.0 + weight_decay: 0.0 + logging_frequency: 1 + gradient_clipping: 1.0 + gradient_accumulation_steps: 4 + distributed_backend: none + is_resume: true + pretrained_checkpoint: /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_1300_state + resume_step: 1300 + reload_modules: null + optimizer: + name: AdamW + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + weight_decay: 1.0e-08 + fused: false + save_format: pt +workspace_dir: WORKSPACE_DIR +run_root_dir: /workspace/starVLA/results/Checkpoints +seed: 42 +wandb_entity: dongqianyu99-zhejiang-university +wandb_project: starVLA_rl_games +auth: + env_file: null + hf_token_env: HF_TOKEN + wandb_api_key_env: WANDB_API_KEY +paths: + run_root_dir: results/Checkpoints + dataset_local_dir: playground/Datasets/rl_games + dataset_cache_dir: null + base_model_dir: playground/Pretrained_models/Qwen3-VL-4B-Instruct + accelerate_config: starVLA/config/deepseeds/deepspeed_zero2.yaml +launch: + use_accelerate: true + gpus: null + num_processes: 1 + dry_run: false +conda: + enabled: true + env_name: null +rl_games: + model_alias: openvla + env_eval: + image_size: 224 + frameskip: 1 + seed: 42 + fixed_episode_seeds: true + latency_seed_stride: 0 + task_seed_stride: 0 + task_description: '' + enabled: true + distributed_mode: none + vectorized: + enabled: false + batch_size: 1 + latency: + prompt_map_path: null + mode: single + values: + - 0 + mid_train: + enabled: false + interval_steps: 100 + latencies: + - 0 + num_episodes: 5 + max_steps_per_episode: 3600 + post_train: + enabled: false + latencies: + - 0 + - 1 + - 2 + - 3 + - 4 + num_episodes: 5 + max_steps_per_episode: 3600 + task: flappy + initialization_mode: bridge + action_carrier: bridge +model: openvla +env: flappy +init: bridge +bridge_base_model: + repo_id: + openvla: Qwen/Qwen3-VL-4B-Instruct + pi0: StarVLA/Qwen2.5-VL-3B-Instruct-Action + pi05: Qwen/Qwen3-VL-4B-Instruct + gr00t: Qwen/Qwen3-VL-4B-Instruct + local_dir: + openvla: playground/Pretrained_models/Qwen3-VL-4B-Instruct + pi0: playground/Pretrained_models/Qwen2.5-VL-3B-Instruct-Action + pi05: playground/Pretrained_models/Qwen3-VL-4B-Instruct + gr00t: playground/Pretrained_models/Qwen3-VL-4B-Instruct +mode: single +checkpoint: + load: auto + hf_repo_id: null + save_best_model: false + save_pt_file: false + local: + keep_last_n: 1 + sync: + enabled: false + repo_id: null + keep_last_n: 0 + sync_every_n_checkpoints: 1 + resume_policy: local_latest +run_id: flappy_fix_latency_0_200ep +output_dir: /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep +config_yaml: null +is_debug: false +version_id: '0.21' diff --git a/zero-latency/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_0_200ep/config.yaml b/zero-latency/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_0_200ep/config.yaml new file mode 100644 index 0000000000000000000000000000000000000000..e10d6042c8a01367c5c5f6a4d12c2c2bf7216a59 --- /dev/null +++ b/zero-latency/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_0_200ep/config.yaml @@ -0,0 +1,71 @@ +checkpoint: + local: + keep_last_n: 1 + save_best_model: false + save_pt_file: false + sync: + enabled: false + keep_last_n: 0 + repo_id: null +datasets: + vla_data: + data_mix: flappy_train__bridge + dataset_py: lerobot_datasets + eval_data_mix: flappy_train__bridge__val + obs_image_size: null + per_device_batch_size: 32 +framework: + action_model: + action_dim: 7 + action_env_dim: 2 + action_hidden_dim: 2560 + action_horizon: 1 + action_model_type: MLP + loss_type: discrete_ce + name: QwenOFT + qwenvl: + attn_implementation: flash_attention_2 + base_vlm: /workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + enable_gradient_checkpointing: true +output_dir: /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep +rl_games: + env_eval: + enabled: true + mid_train: + enabled: false + interval_steps: 100 + task: flappy +run_id: flappy_fix_latency_0_200ep +run_root_dir: /workspace/starVLA/results/Checkpoints +seed: 42 +trainer: + distributed_backend: none + eval_action_classification: true + eval_interval: 100 + eval_num_batches: 100 + freeze_modules: '' + gradient_accumulation_steps: 4 + gradient_clipping: 1.0 + is_resume: true + learning_rate: + action_model: 0.0001 + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + logging_frequency: 1 + lr_scheduler_type: cosine_with_min_lr + max_train_steps: 5000 + num_warmup_steps: 100 + optimizer: + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + fused: false + weight_decay: 1.0e-08 + pretrained_checkpoint: /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_1300_state + resume_step: 1300 + save_interval: 100 + scheduler_specific_kwargs: + min_lr: 1.0e-06 +wandb_entity: dongqianyu99-zhejiang-university +wandb_project: starVLA_rl_games diff --git a/zero-latency/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_0_200ep/dataset_statistics.json b/zero-latency/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_0_200ep/dataset_statistics.json new file mode 100644 index 0000000000000000000000000000000000000000..9da22dc41af3ec6fc40d10a9daedf9f736bc7d29 --- /dev/null +++ b/zero-latency/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_0_200ep/dataset_statistics.json @@ -0,0 +1,127 @@ +{ + "new_embodiment": { + "action": { + "mean": [ + 0.813460648059845, + 0.18653932213783264, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.3898015022277832, + 0.3898015022277832, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "mask": [ + true, + true, + true, + true, + true, + true, + true + ] + }, + "state": { + "mean": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ] + }, + "num_transitions": 279882, + "num_trajectories": 180 + } +} \ No newline at end of file diff --git a/zero-latency/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_0_200ep/dataset_statistics_eval.json b/zero-latency/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_0_200ep/dataset_statistics_eval.json new file mode 100644 index 0000000000000000000000000000000000000000..4a1c3a83be372c3c5fd5770724ed11593880b782 --- /dev/null +++ b/zero-latency/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_0_200ep/dataset_statistics_eval.json @@ -0,0 +1,127 @@ +{ + "new_embodiment": { + "action": { + "mean": [ + 0.9196249842643738, + 0.08037500083446503, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.2717601954936981, + 0.2717601954936981, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "mask": [ + true, + true, + true, + true, + true, + true, + true + ] + }, + "state": { + "mean": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ] + }, + "num_transitions": 72000, + "num_trajectories": 20 + } +} \ No newline at end of file diff --git a/zero-latency/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_0_200ep/hydra/.hydra/config.yaml b/zero-latency/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_0_200ep/hydra/.hydra/config.yaml new file mode 100644 index 0000000000000000000000000000000000000000..0957da95be1d2576c41dd79a48e06612cd325115 --- /dev/null +++ b/zero-latency/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_0_200ep/hydra/.hydra/config.yaml @@ -0,0 +1,196 @@ +framework: + qwenvl: + base_vlm: /workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + attn_implementation: flash_attention_2 + enable_gradient_checkpointing: true + action_model: + state_dim: 7 + loss_type: discrete_ce + action_horizon: 1 + future_action_window_size: 0 + past_action_window_size: 0 + action_dim: 7 + action_env_dim: 2 + name: QwenOFT +datasets: + vla_data: + dataset_py: lerobot_datasets + include_state: true + data_root_dir: /workspace/starVLA/data/flappy_fix_latency_0_200ep + data_mix: flappy_train__bridge + eval_data_mix: flappy_train__bridge__val + custom_mixtures_path: null + action_type: discrete + sequential_step_sampling: false + shuffle: true + action_balance: + enabled: false + strategy: balanced_epoch + action_key: action_id + target_flap_fraction: 0.3 + noop_id: 0 + flap_id: 1 + per_device_batch_size: 32 + load_all_data_for_training: true + obs_image_size: null + video_backend: torchvision_av +dataset: + source_hf: ${dataset.single_source_hf} + config_name: null + converted_name: ${dataset.single_converted_name} + single_source_hf: '' + mixed_source_hf: '' + single_converted_name: flappy_train + mixed_converted_name: flappy_mixed_latency_train + single_latency_filter: null + mixed_latency_filter: null + force_download: false + setup_force: false + verify_rows: 200 + max_episodes: null + episodes_per_latency: null + latency_filter: ${dataset.single_latency_filter} + debug_subset: + enabled: false + max_episodes: 5 + suffix: debug +base_model: + repo_id: ${bridge_base_model.repo_id.${model}} +initialization: + checkpoint_local_dir: playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_hf_repo_id: StarVLA/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_filename: checkpoints/steps_5000_pytorch_model.pt +trainer: + max_train_steps: 5000 + num_warmup_steps: 100 + save_interval: 100 + eval_interval: 100 + eval_num_batches: 100 + eval_action_classification: true + cc_f1_tolerance: 1 + learning_rate: + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + action_model: 0.0001 + lr_scheduler_type: cosine_with_min_lr + scheduler_specific_kwargs: + min_lr: 1.0e-06 + freeze_modules: '' + loss_scale: + vla: 1.0 + vlm: 0.1 + max_grad_norm: 1.0 + weight_decay: 0.0 + logging_frequency: 1 + gradient_clipping: 1.0 + gradient_accumulation_steps: 4 + distributed_backend: none + is_resume: true + pretrained_checkpoint: /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_1300_state + resume_step: 1300 + reload_modules: null + optimizer: + name: AdamW + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + weight_decay: 1.0e-08 + fused: false + save_format: pt +workspace_dir: WORKSPACE_DIR +run_root_dir: /workspace/starVLA/results/Checkpoints +seed: 42 +wandb_entity: dongqianyu99-zhejiang-university +wandb_project: starVLA_rl_games +auth: + env_file: null + hf_token_env: HF_TOKEN + wandb_api_key_env: WANDB_API_KEY +paths: + run_root_dir: results/Checkpoints + dataset_local_dir: playground/Datasets/rl_games + dataset_cache_dir: null + base_model_dir: ${bridge_base_model.local_dir.${model}} + accelerate_config: starVLA/config/deepseeds/deepspeed_zero2.yaml +launch: + use_accelerate: true + gpus: null + num_processes: 1 + dry_run: false +conda: + enabled: true + env_name: null +rl_games: + model_alias: openvla + env_eval: + image_size: 224 + frameskip: 1 + seed: 42 + fixed_episode_seeds: true + latency_seed_stride: 0 + task_seed_stride: 0 + task_description: '' + enabled: true + distributed_mode: none + vectorized: + enabled: false + batch_size: 1 + latency: + prompt_map_path: null + mode: single + values: + - 0 + mid_train: + enabled: false + interval_steps: 100 + latencies: + - 0 + num_episodes: 5 + max_steps_per_episode: 3600 + post_train: + enabled: false + latencies: + - 0 + - 1 + - 2 + - 3 + - 4 + num_episodes: 5 + max_steps_per_episode: 3600 + task: flappy + initialization_mode: bridge + action_carrier: bridge +model: openvla +env: flappy +init: bridge +bridge_base_model: + repo_id: + openvla: Qwen/Qwen3-VL-4B-Instruct + pi0: StarVLA/Qwen2.5-VL-3B-Instruct-Action + pi05: Qwen/Qwen3-VL-4B-Instruct + gr00t: Qwen/Qwen3-VL-4B-Instruct + local_dir: + openvla: playground/Pretrained_models/Qwen3-VL-4B-Instruct + pi0: playground/Pretrained_models/Qwen2.5-VL-3B-Instruct-Action + pi05: playground/Pretrained_models/Qwen3-VL-4B-Instruct + gr00t: playground/Pretrained_models/Qwen3-VL-4B-Instruct +mode: single +checkpoint: + load: auto + hf_repo_id: null + save_best_model: false + save_pt_file: false + local: + keep_last_n: 1 + sync: + enabled: false + repo_id: null + keep_last_n: 0 + sync_every_n_checkpoints: 1 + resume_policy: local_latest +run_id: flappy_fix_latency_0_200ep +output_dir: null +config_yaml: null +is_debug: false +version_id: '0.21' diff --git a/zero-latency/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_0_200ep/hydra/.hydra/hydra.yaml b/zero-latency/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_0_200ep/hydra/.hydra/hydra.yaml new file mode 100644 index 0000000000000000000000000000000000000000..a667ada506cca593694ab5cc17f8677faeab1135 --- /dev/null +++ b/zero-latency/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_0_200ep/hydra/.hydra/hydra.yaml @@ -0,0 +1,252 @@ +hydra: + run: + dir: ${run_root_dir}/${run_id}/hydra + sweep: + dir: multirun/${now:%Y-%m-%d}/${now:%H-%M-%S} + subdir: ${hydra.job.num} + launcher: + _target_: hydra._internal.core_plugins.basic_launcher.BasicLauncher + sweeper: + _target_: hydra._internal.core_plugins.basic_sweeper.BasicSweeper + max_batch_size: null + params: null + help: + app_name: ${hydra.job.name} + header: '${hydra.help.app_name} is powered by Hydra. + + ' + footer: 'Powered by Hydra (https://hydra.cc) + + Use --hydra-help to view Hydra specific help + + ' + template: '${hydra.help.header} + + == Configuration groups == + + Compose your configuration from those groups (group=option) + + + $APP_CONFIG_GROUPS + + + == Config == + + Override anything in the config (foo.bar=value) + + + $CONFIG + + + ${hydra.help.footer} + + ' + hydra_help: + template: 'Hydra (${hydra.runtime.version}) + + See https://hydra.cc for more info. + + + == Flags == + + $FLAGS_HELP + + + == Configuration groups == + + Compose your configuration from those groups (For example, append hydra/job_logging=disabled + to command line) + + + $HYDRA_CONFIG_GROUPS + + + Use ''--cfg hydra'' to Show the Hydra config. + + ' + hydra_help: ??? + hydra_logging: + version: 1 + formatters: + simple: + format: '[%(asctime)s][HYDRA] %(message)s' + handlers: + console: + class: logging.StreamHandler + formatter: simple + stream: ext://sys.stdout + root: + level: INFO + handlers: + - console + loggers: + logging_example: + level: DEBUG + disable_existing_loggers: false + job_logging: + version: 1 + formatters: + simple: + format: '[%(asctime)s][%(name)s][%(levelname)s] - %(message)s' + handlers: + console: + class: logging.StreamHandler + formatter: simple + stream: ext://sys.stdout + file: + class: logging.FileHandler + formatter: simple + filename: ${hydra.runtime.output_dir}/${hydra.job.name}.log + root: + level: INFO + handlers: + - console + - file + disable_existing_loggers: false + env: {} + mode: RUN + searchpath: [] + callbacks: {} + output_subdir: .hydra + overrides: + hydra: + - hydra.mode=RUN + task: + - model=openvla + - env=flappy + - init=bridge + - mode=single + - run_id=flappy_fix_latency_0_200ep + - run_root_dir=/workspace/starVLA/results/Checkpoints + - seed=42 + - wandb_entity=dongqianyu99-zhejiang-university + - wandb_project=starVLA_rl_games + - rl_games.env_eval.enabled=true + - checkpoint.sync.enabled=false + - checkpoint.sync.keep_last_n=0 + - checkpoint.local.keep_last_n=1 + - checkpoint.save_best_model=false + - checkpoint.save_pt_file=false + - trainer.is_resume=true + - trainer.pretrained_checkpoint=/workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_1300_state + - trainer.resume_step=1300 + - trainer.max_train_steps=5000 + - trainer.num_warmup_steps=100 + - trainer.save_interval=100 + - trainer.eval_interval=100 + - trainer.eval_num_batches=100 + - trainer.logging_frequency=1 + - trainer.gradient_accumulation_steps=4 + - trainer.distributed_backend=none + - trainer.learning_rate.base=2e-05 + - trainer.learning_rate.qwen_vl_interface=1e-05 + - trainer.learning_rate.action_model=0.0001 + - trainer.lr_scheduler_type=cosine_with_min_lr + - trainer.scheduler_specific_kwargs.min_lr=1e-06 + - trainer.loss_scale.vla=1.0 + - trainer.loss_scale.vlm=0.1 + - trainer.max_grad_norm=1.0 + - trainer.weight_decay=0.0 + - trainer.gradient_clipping=1.0 + - trainer.optimizer.name=AdamW + - trainer.optimizer.betas=[0.9, 0.95] + - trainer.optimizer.eps=1e-08 + - trainer.optimizer.weight_decay=1e-08 + - trainer.optimizer.fused=false + - trainer.save_format=pt + - framework.name=QwenOFT + - framework.qwenvl.attn_implementation=flash_attention_2 + - framework.qwenvl.enable_gradient_checkpointing=true + - framework.action_model.action_dim=7 + - framework.action_model.action_env_dim=2 + - framework.action_model.state_dim=7 + - framework.action_model.loss_type=discrete_ce + - framework.action_model.action_horizon=1 + - framework.action_model.future_action_window_size=0 + - framework.action_model.past_action_window_size=0 + - datasets.vla_data.include_state=true + - datasets.vla_data.action_type=discrete + - datasets.vla_data.sequential_step_sampling=false + - datasets.vla_data.shuffle=true + - datasets.vla_data.action_balance.enabled=false + - datasets.vla_data.action_balance.strategy=balanced_epoch + - datasets.vla_data.action_balance.action_key=action_id + - datasets.vla_data.action_balance.target_flap_fraction=0.3 + - datasets.vla_data.action_balance.noop_id=0 + - datasets.vla_data.action_balance.flap_id=1 + - datasets.vla_data.per_device_batch_size=32 + - datasets.vla_data.load_all_data_for_training=true + - datasets.vla_data.video_backend=torchvision_av + - datasets.vla_data.data_root_dir=/workspace/starVLA/data/flappy_fix_latency_0_200ep + - datasets.vla_data.data_mix=flappy_train__bridge + - datasets.vla_data.eval_data_mix=flappy_train__bridge__val + - framework.qwenvl.base_vlm=/workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + - rl_games.task=flappy + - rl_games.model_alias=openvla + - rl_games.initialization_mode=bridge + - rl_games.action_carrier=bridge + - rl_games.env_eval.distributed_mode=none + - rl_games.env_eval.latency.mode=single + - rl_games.env_eval.frameskip=1 + - rl_games.env_eval.image_size=224 + - rl_games.env_eval.seed=42 + - rl_games.env_eval.fixed_episode_seeds=true + - rl_games.env_eval.latency_seed_stride=0 + - rl_games.env_eval.task_seed_stride=0 + - rl_games.env_eval.latency.values=[0] + - rl_games.env_eval.mid_train.enabled=false + - rl_games.env_eval.mid_train.interval_steps=100 + - rl_games.env_eval.mid_train.num_episodes=5 + - rl_games.env_eval.mid_train.max_steps_per_episode=3600 + - rl_games.env_eval.mid_train.latencies=[0] + - rl_games.env_eval.post_train.enabled=false + - rl_games.env_eval.post_train.num_episodes=5 + - rl_games.env_eval.post_train.max_steps_per_episode=3600 + - rl_games.env_eval.post_train.latencies=[0,1,2,3,4] + job: + name: train_starvla_hydra + chdir: false + override_dirname: checkpoint.local.keep_last_n=1,checkpoint.save_best_model=false,checkpoint.save_pt_file=false,checkpoint.sync.enabled=false,checkpoint.sync.keep_last_n=0,datasets.vla_data.action_balance.action_key=action_id,datasets.vla_data.action_balance.enabled=false,datasets.vla_data.action_balance.flap_id=1,datasets.vla_data.action_balance.noop_id=0,datasets.vla_data.action_balance.strategy=balanced_epoch,datasets.vla_data.action_balance.target_flap_fraction=0.3,datasets.vla_data.action_type=discrete,datasets.vla_data.data_mix=flappy_train__bridge,datasets.vla_data.data_root_dir=/workspace/starVLA/data/flappy_fix_latency_0_200ep,datasets.vla_data.eval_data_mix=flappy_train__bridge__val,datasets.vla_data.include_state=true,datasets.vla_data.load_all_data_for_training=true,datasets.vla_data.per_device_batch_size=32,datasets.vla_data.sequential_step_sampling=false,datasets.vla_data.shuffle=true,datasets.vla_data.video_backend=torchvision_av,env=flappy,framework.action_model.action_dim=7,framework.action_model.action_env_dim=2,framework.action_model.action_horizon=1,framework.action_model.future_action_window_size=0,framework.action_model.loss_type=discrete_ce,framework.action_model.past_action_window_size=0,framework.action_model.state_dim=7,framework.name=QwenOFT,framework.qwenvl.attn_implementation=flash_attention_2,framework.qwenvl.base_vlm=/workspace/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct,framework.qwenvl.enable_gradient_checkpointing=true,init=bridge,mode=single,model=openvla,rl_games.action_carrier=bridge,rl_games.env_eval.distributed_mode=none,rl_games.env_eval.enabled=true,rl_games.env_eval.fixed_episode_seeds=true,rl_games.env_eval.frameskip=1,rl_games.env_eval.image_size=224,rl_games.env_eval.latency.mode=single,rl_games.env_eval.latency.values=[0],rl_games.env_eval.latency_seed_stride=0,rl_games.env_eval.mid_train.enabled=false,rl_games.env_eval.mid_train.interval_steps=100,rl_games.env_eval.mid_train.latencies=[0],rl_games.env_eval.mid_train.max_steps_per_episode=3600,rl_games.env_eval.mid_train.num_episodes=5,rl_games.env_eval.post_train.enabled=false,rl_games.env_eval.post_train.latencies=[0,1,2,3,4],rl_games.env_eval.post_train.max_steps_per_episode=3600,rl_games.env_eval.post_train.num_episodes=5,rl_games.env_eval.seed=42,rl_games.env_eval.task_seed_stride=0,rl_games.initialization_mode=bridge,rl_games.model_alias=openvla,rl_games.task=flappy,run_id=flappy_fix_latency_0_200ep,run_root_dir=/workspace/starVLA/results/Checkpoints,seed=42,trainer.distributed_backend=none,trainer.eval_interval=100,trainer.eval_num_batches=100,trainer.gradient_accumulation_steps=4,trainer.gradient_clipping=1.0,trainer.is_resume=true,trainer.learning_rate.action_model=0.0001,trainer.learning_rate.base=2e-05,trainer.learning_rate.qwen_vl_interface=1e-05,trainer.logging_frequency=1,trainer.loss_scale.vla=1.0,trainer.loss_scale.vlm=0.1,trainer.lr_scheduler_type=cosine_with_min_lr,trainer.max_grad_norm=1.0,trainer.max_train_steps=5000,trainer.num_warmup_steps=100,trainer.optimizer.betas=[0.9, + 0.95],trainer.optimizer.eps=1e-08,trainer.optimizer.fused=false,trainer.optimizer.name=AdamW,trainer.optimizer.weight_decay=1e-08,trainer.pretrained_checkpoint=/workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/checkpoints/steps_1300_state,trainer.resume_step=1300,trainer.save_format=pt,trainer.save_interval=100,trainer.scheduler_specific_kwargs.min_lr=1e-06,trainer.weight_decay=0.0,wandb_entity=dongqianyu99-zhejiang-university,wandb_project=starVLA_rl_games + id: ??? + num: ??? + config_name: train + env_set: {} + env_copy: [] + config: + override_dirname: + kv_sep: '=' + item_sep: ',' + exclude_keys: [] + runtime: + version: 1.3.3 + version_base: '1.1' + cwd: /workspace/starVLA + config_sources: + - path: hydra.conf + schema: pkg + provider: hydra + - path: /workspace/starVLA/examples/rl_games/config + schema: file + provider: main + - path: '' + schema: structured + provider: schema + output_dir: /workspace/starVLA/results/Checkpoints/flappy_fix_latency_0_200ep/hydra + choices: + cross_task_setup: null + checkpoint: default + mode: single + init: bridge + env: flappy + model: openvla + hydra/env: default + hydra/callbacks: null + hydra/job_logging: default + hydra/hydra_logging: default + hydra/hydra_help: default + hydra/help: default + hydra/sweeper: basic + hydra/launcher: basic + hydra/output: default + verbose: false